Effects of the Modification of Gas Diffusion Electrodes by Organic Redox Catalysts for Hydrogen Peroxide Electrosynthesis

Este trabalho apresenta um estudo comparativo da eficiência eletroquímica na eletrogeração de peróxido de hidrogênio in situ usando eletrodos da difusão gasosa modificados com os catalisadores orgânicos redox: 2-etilantraquinona, 2-terc-butilantraquinona e azobenzeno em meio de 0,1 mol L H 2 SO 4 + 0,1 mol L K 2 SO 4 , pH = 1. A produção de peróxido de hidrogênio está diretamente relacionada ao potencial aplicado e a concentração dos catalisadores adicionados. A modificação dos eletrodos resultou em um aumento significativo no rendimento de H 2 O 2 (30%) alcançando 850 mg L e o sobrepotencial da reação de redução do oxigênio foi deslocado para valores menos negativos (400 mV vs Ag/AgCl para os eletrodos modificados com quinonas e 300 mV vs. Ag/AgCl para os eletrodos modificados com azobenzeno) comparado ao eletrodo de difusão gasosa não modificado, reduzindo o consumo de energia de 596,5 para 232,4 kWh kg. Os resultados indicaram que o melhor eletrodo para a eletrogeração do H 2 O 2 é o eletrodo de difusão gasosa modificado com 10% de 2-etilantraquinona, o qual apresentou a melhor relação custo/benefício.


Introduction
2][3][4][5] In wastewater treatment, hydrogen peroxide can oxidize organic pollutants to CO 2 , offering numerous advantages in terms of treatment rate, cost and availability in comparison to traditional biodegradation methods. 6Electrochemical technology can be an efficient means of hydrogen peroxide electrosynthesis.The problem of the low solubility of oxygen in aqueous solutions, which limits the reduction reaction by mass transfer rate and, hence, to low limiting current values, can be overcome by the use of threedimensional electrodes such as gas diffusion electrodes (GDE), which have large areas available for reaction and allow for higher mass transfer rates. 6,7n recent times, much interest has focused on the modification of electrodes with organic catalysts for hydrogen peroxide electrosynthesis.Investigations with catalysts of the anthraquinone class to help hydrogen peroxide production have become common (see Table 1).The O 2 reduction reaction mechanism on the surface of anthraquinonemodified electrodes is represented by reactions: 8,9 Q (ads) + 2e − + 2H + (aq) → H 2 Q (ads) (1)   H 2 Q (ads) + O 2(aq) → Q (ads) + H 2 O 2(aq) (2)   Oxygen reduction reaction and hydrogen peroxide formation was studied using glassy carbon electrodes modified by physical adsorption of quinone derivatives. 8,14lectrode modifications by quinones resulted in a shift of the oxygen reduction overpotential to less negative values 8 and a significant increase in current values due to the formation of hydrogen peroxide. 14egrand 17 studied the catalytic reduction of O 2 to H 2 O 2 in different pH values using a glassy carbon electrode modified by the adsorption of polymeric-anthraquinone, and observed that the catalytic current value increased with pH.
Oxygen reduction on carbon paste electrodes modified with 1,4-naphthoquinone was studied by Golabi and Raoof 15 at different pH values.These authors observed a gradual increase of the peak cathodic current and a decrease of the peak anodic current as the pH value increased; pH = 8 was adopted as the best value for that study.The optimal condition for O 2 reduction on the surface of carbon paste electrodes modified with 1,4-naphthoquinone was observed to occur at potentials 350 to 550 mV less negative than those of noncatalyzed carbon paste electrodes. 15It is known that an alkaline medium is more widely used for H 2 O 2 production because of OH -in solution; however, the literature reports that the O 2 reduction peak potential shifts to less negative overpotential values when pH < 2. 8,16,20 A lesser amount than quinones, some researchers have studied the oxygen reduction reaction with azobenzene modified electrodes, as indicated in Table 2.
Sljukic et al. 21studied the oxygen reduction reaction using glassy carbon electrodes or pyrolytic graphite modified with azobenzene and its derivatives.Cyclic voltammetry of an azobenzene-modified pyrolytic graphite electrode, recorded in phosphate buffer with pH = 2, showed a reduction peak at −0.6 V vs. ECS, which corresponds to the reduction of azobenzene to hydroazobenzene.The corresponding oxidation peak was observed at +0.4 V vs. ECS.The authors also found that when the potential was scanned more negatively, the reduction of hydroazobenzene to aniline at −1.2 V vs. ECS was irreversible. 21,27,28yclic voltammetry of an azobenzene-modified gold electrode showed a relation between redox peaks and scan rate. 22At low scan rates, only one oxidation peak and one reduction peak were observed, involving 2 protons and 2 electrons in a single step, characterizing trans-azobenzene 29 .At high scan rates (≥ 300 mV s -1 ), two reversible peaks were visible, corresponding to a two-step cis-azobenzene reaction, i.e., one proton and one electron in each step. 22lassy carbon electrodes modified with azobenzene shifted the oxygen reduction reaction to less negative potential values (−0.37 V vs. ECS) compared with the noncatalyzed glassy carbon electrode (−0.53 V vs. ECS). 21his indicates that azobenzene-modified electrodes have good electrocatalytic activity for H 2 O 2 production. 21he incorporation of solid catalysts into the EDG graphitic mass is a novel application and the first results obtained by our group started to be reported last year.One of our papers 30 introduces the subject with the phrase: "Quinones have already proved their efficiency in the synthesis of hydrogen peroxide" and, although traditional they are still defeating our knowledge about how do they work and how to place them in the reaction ambient for obtaining the best catalytic effects and reaction rates.
Thus, the objective of the present work was to perform a comparative study among graphite pigment based gas diffusion electrodes (GDE) modified by 2-ethylanthraquinone, 2-terc-butylanthraquinone and azobenzene.These organic compounds were incorporated into the graphitic mass before the electrode pressing and their performance as a catalyst was compared by following the oxygen reduction currents and the yield during hydrogen peroxide synthesis.

Electrode preparation
Precursor mass for the modified gas diffusion electrode (MGDE) was prepared from Degussa Printex 6L conductive carbon-black graphite pigment.A 60% polytetrafluoroethylene dispersion (Dyneon TF 5035 PTFE) was used as hydrophobic binder.The ratio of Printex to PTFE was 8/3.3, which is equivalent to 20% of PTFE.The mixture was homogenized in a 4:1 water:isopropanol bidistilled solution.The selected amounts of organic redox catalysts 2-ethylanthraquinone (EAQ), azobenzene (AZO) and 2-terc-butylanthraquinone (BAQ) from 3%, 5% and 10% (m/m), relative to the carbon pigment, were incorporated into the MGDE precursor mass, which was dried at 110 °C for 24 h.A 200 mesh AISI 304 stainless steel screen current collector was placed at the bottom of a 60 mm diameter pressing tool, which was then filled with 8 g of the precursor mass.Sintered 3 mm thick MGDE was obtained after 1.5 h at 310 °C, under a load of 18 MPa.

MGDE behavior
For voltammetric and electrolytic experiments, an electrochemical cell (one compartment, 250 mL) was used (Figure 1).GDE prepared with different concentrations of EAQ, BAQ and AZO (3, 5 and 10%) was used as cathode.
The MGDE was placed at the bottom of the cell with an exposed area of 19 cm 2 , and the electrode was oxygen-backfed.The reference was Ag/AgCl (KCl sat.), and platinum foil was used as the counter electrode (A = 24 cm 2 ).The cell was thermostatted at 20 °C.The supporting electrolyte was 0.1 mol L -1 H 2 SO 4 plus 0.1 mol L -1 K 2 SO 4 , pH = 1.Experiments were performed under mechanical stirring.Linear voltammetry measurements were recorded from −0.2 V to −1.0 V vs. Ag/AgCl at 20 mV s -1 .The supporting electrolyte was previously saturated with nitrogen.Afterwards, i/E responses were also recorded in the presence of oxygen.
Subsequently, controlled potential electrolysis was employed to optimize the H 2 O 2 electrogeneration rate relative to the applied potential in the range of −0.4 ≤ E ≤ −0.9 V vs. Ag/AgCl for MGDE, and −0.6 ≤ E ≤ −1.0 V vs. Ag/AgCl for GDE.During electrolysis an oxygen pressure of 0.16 bar was kept through the reverse side of the electrode.The electrolyte was sampled at 5 min intervals for the first half hour, and every 10 minutes thereafter.Electrolysis experiments were conducted in duplicate.
The hydrogen peroxide concentration was determined with a UV-Vis spectrophotometer (Lambda 40, Perkin Elmer Instruments), recording the spectra over 200 to 500 nm.A solution of 2.4 mmol L -1 (NH 4 ) 6 Mo 7 O 24 .4H 2 O in 0.5 mol L -1 H 2 SO 4 was added to the samples, resulting in a yellow color. 31The absorbance was determined at 350 nm.Calibration plots based on Beer-Lambert's law were established relating absorbance to concentration.
To verify whether redox catalysts dissolve in the highly acidic medium used as supporting electrolyte (pH = 1), the solutions after electrolysis were analyzed by the High Performance Liquid Chromatography technique (Shimadzu, model 20A), using an SPD-20A UV/Visible detector, an automatic sampler, and a DGU-20A5 degasser coupled to a PC microcomputer.The stationary phase was a 250 x 4.6 mm Shimadzu Shim Pack CLC -ODS (M) column (column C18), and the mobile phase was acetonitrile (100%).The mobile phase flow rate was 1.0 mL min -1 , and wavelengths analyzed by the UV/Visible detector were 254 and 325 nm for quinones 32 and 365 nm for azobenzene. 33

Voltammetric experiments
To study the electrochemical behavior of electrodes, linear voltammograms were recorded in the cathode region in a medium of 0.1 mol L -1 H 2 SO 4 plus 0.1 mol L -1 K 2 SO 4 .
Figure 2 shows linear voltammetries (LV) recorded for electrodes with different concentrations of (A) 2-ethylanthraquinone, (B) 2-terc-butylanthraquinone and (C) azobenzene and a constant O 2 flow (P = 0.16 bar).Figure 3  The curves in Figures 2 and 3 exhibit two current plateaus not well defined associated with the O 2 reduction reaction to H 2 O 2 (reaction 3), and its subsequent reduction to water (reaction 4).
The first plateau is observed between −0.3 and −0.7 V vs. Ag/AgCl (reaction 3), and the second step starting from −0.7 V vs. Ag/AgCl (reaction 4).Separation between the two steps of the oxygen reduction reaction is 400 mV, allowing these electrodes to be used in the production of hydrogen peroxide.
Regardless of the nature of the catalyst added to the GDE, higher current values were consistently observed; indicating that the oxygen reduction reaction on modified GDEs occurred at less negative overpotentials compared to a noncatalyzed GDE, indicating that GDE modification by organic redox catalysts improved the efficiency of hydrogen peroxide electrogeneration.
Current values increased with catalyst concentration in all three cases, probably as a result of two simultaneous processes: oxygen reduction on the graphite surface (electrochemical step), and the catalyst redox reaction that reduces O 2 (chemical step).When the catalyst concentration is increased, the current also increases in response to the reduction in the overpotential required for the oxygen reduction reaction.

Electrolysis at controlled potential
To monitor the hydrogen peroxide electrogeneration, electrolyses were performed at constant potential, and the contents were monitored by collecting samples at regular intervals and analyzing them by UV-Vis spectrophotometry.
The production H 2 O 2 in function of time electrolysis for GDE modified by 10% of catalysts is show in the Figure 4.In every case, the electrogenerated H 2 O 2 concentration increased linearly over time.
The Figures 5-7 show electrogenerated H 2 O 2 content after one hour of electrolysis in function of the applied potential.H 2 O 2 electrogeneration by noncatalyzed GDE increased with the applied potential, reaching its best yield at potential of −1.0 V vs. Ag/AgCl (567 mg L -1 of H 2 O 2 ).
On the other hand, the results of electrolyses using GDEs modified with EAQ or BAQ (Figures 5 and 6) reached higher yields at the potential of −0.6 V vs. Ag/AgCl, when the electrogenerated H 2 O 2 concentration reached its maximum.The behavior observed in the LVs experiments was confirmed, i.e., an initial increase up to −0.6 V vs. Ag/AgCl related with the oxygen reduction reaction (reaction 3), and from −0.7 V vs. Ag/AgCl the H 2 O 2 the efficiency of the production process began to decline because of the decomposition of the electrogenerated hydrogen peroxide with the formation of H 2 O via 4 electrons transfer (reaction 4).
Our results indicated that GDEs modified with EAQ or BAQ are more efficient than noncatalyzed GDE: in addition to increasing the H 2 O 2 yield by about 30%, the H 2 O 2 electrogeneration overpotential was reduced by 400 mV, reducing the consumption of energy.GDE modification with AZO also increased the catalytic activity for H 2 O 2 electrogeneration (Figure 7) and a higher yield was observed compared with the noncatalyzed GDE.As for the applied potential, the maximum yield was obtained at −0.7 V vs. Ag/AgCl, displacing the overpotential for H 2 O 2 generation by 300 mV to less negative values.

Comparison of the efficiency of catalysts added to GDE
Figure 8 shows a comparison of the electrogenerated H 2 O 2 after 1 hour of electrolysis applying −0.6 V as function of the concentration of each catalyst used to modify GDEs.In general, GDE modification by redox organic catalysts resulted in more efficient H 2 O 2 electrogeneration, with lower energy consumption compared to noncatalyzed GDE due to a lower overpotential of the oxygen reduction reaction.In every case, the production of H 2 O 2 increased along with the catalyst concentration.
Comparing the results of H 2 O 2 production as a function of the catalyst, under −0.6 V, BAQ was the most efficient for concentration higher than 5% reaching a yield of 850 mg L -1 of H 2 O 2 .However, modifying the GDE modification with just 3% AZO resulted in H 2 O 2 contents equal to one half of that production volume.
An important electrochemical parameter is energy consumption (EC), which was computed from the cell potential monitored during electrolyses.Keeping the catalyst concentration at 10%, the values of energy consumption (kWh kg -1 of produced H 2 O 2 ) were compared, as shown in Table 3.
Table 3 shows that the modification by quinones was more efficient than by azobenzene.In terms of electrogenerated H 2 O 2 and required overpotential, GDE modification by BAQ was more efficient, but in terms of    energy consumption, GDE modification by EAQ was more efficient.When the cost of catalysts is taken into account, the efficiency of EAQ modified GDEs is even higher: its modification costs 4-fold less than by BAQ.In conclusion, modification of GDEs by EAQ offers a better cost/benefit ratio than by other catalysts.
To check whether the catalysts dissolve in the highly acidic medium used as supporting electrolyte (pH = 1), analyses were done by High Performance Liquid Chromatography.These analyses involved standards of 2-ethylanthraquinone and azobenzene (both dissolved in acetonitrile to 140 ppm), the supporting electrolyte (0.1 mol L -1 H 2 SO 4 plus 0.1 mol L -1 K 2 SO 4 ) and samples collected at the end of electrolyses.Chromatograms showed peaks of EAQ standard solutions at 4.6 min, and of AZO at 4.7 min.No peaks associated with the catalysts were detected in the solution after electrolyses, and it was concluded that organic compounds added to the GDE do not dissolve in the supporting electrolyte (0.1 mol L -1 H 2 SO 4 plus 0.1 mol L -1 K 2 SO 4 , pH = 1).

Conclusions
Our results indicated a great potential for the use of gas diffusion electrodes modified by organic redox catalysts 2-ethylanthraquinone, 2-terc-butylanthraquinone and azobenzene for in situ H 2 O 2 electrogeneration.The results of constant potential electrolyses showed that H 2 O 2 electrogeneration was efficient, and strongly dependent on the applied potential.
In addition to a considerable amount of H 2 O 2 electrogenerated in acidic medium, these electrodes required less energy, as the H 2 O 2 electrogeneration reaction overpotential shifted to less negative values compared to noncatalyzed GDEs.Our findings indicated that the best electrode for H 2 O 2 electrogeneration is the GDE modified with 10% of 2-ethylanthraquinone, offering the best cost/ benefit ratio.
compares the LVs recorded for MGDE (10%) under N 2 and O 2 flows.The (LV O2 -LV N2 ) curves were obtained by subtracting LV current values recorded with O 2 from those recorded with N 2 .

Figure 8 .
Figure 8. H 2 O 2 electrogeneration after 1 hour of electrolysis at −0.6 V vs. Ag/AgCl as function of the concentration of each catalyst used as GDE modifier.

Table 1 .
Quinone-modified electrodes used in oxygen reduction reactions