Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Electrogeneration of hydrogen peroxide in gas diffusion electrodes: Application of iron (II) phthalocyanine as a modifier of carbon black Fernando L. Silva, Rafael M. Reis, Willyam R.P. Barros, Robson S. Rocha, Marcos R.V. Lanza ⇑ Instituto de Química de São Carlos, Universidade de São Paulo, Avenida Trabalhador São Carlense 400, São Carlos 13566-590, SP, Brazil
a r t i c l e
i n f o
Article history: Received 13 December 2013 Received in revised form 5 March 2014 Accepted 6 March 2014 Available online 15 March 2014 Keywords: Hydrogen peroxide Gas diffusion electrodes Iron (II) phthalocyanine Electrochemical reduction of oxygen
a b s t r a c t Hydrogen peroxide (H2O2) is commonly produced by redox reactions involving organic compounds in organic medium, but such processes present several limitations including the need to extract and concentrate a highly active product. Hydrogen peroxide can be generated in situ by the electrochemical reduction of oxygen in aqueous medium, and the process is particularly efficient when gas diffusion electrodes (GDEs) are employed. A key challenge in the development of such electrodes is the choice of the catalytic particles. This paper describes the evaluation of iron (II) phthalocyanine as a modifier of pigment carbon black used in the construction of GDEs. After 90 min of electrolysis at constant potential, a GDE containing 5% of modifier generated 240 mg L1 of H2O2 (rate constant 7 mg L1 min1; energy consumption 165 kW h kg1 H2O2) while the unmodified GDE produced only 175 mg L1 of H2O2 (3 mg L1 min1; 300 kW h kg1 H2O2) under the same experimental conditions. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Hydrogen peroxide (H2O2) is a potent oxidizing agent and is used widely in organic synthesis, in bleaching paper and in the treatment of effluents containing organic contaminants [1–8]. Hydrogen peroxide may be synthesized on an industrial scale by a process involving autoxidation of a substituted anthrahydroquinone in organic medium, followed by reduction of the anthraquinone so-formed by hydrogen gas in the presence of a catalyst. While the anthraquinone method accounts for approximately 85% of commercially produced H2O2 [9], the process is subject to a number of restrictions since it involves extraction of the organic phase followed by concentration and purification of the active agent. A number of alternative methods for the industrial synthesis of H2O2 have been proposed, such as direct synthesis from molecular O2 catalyzed by Au/Ti [10] or Pd/Au [11] and the application of microbial fuel cells [12]. However, these methods suffer from various limitations including poor production efficiencies, elevated costs and low concentrations of H2O2 formed. Furthermore, manufacture of H2O2 at a site that is distant from its point of use necessitates the storage and transportation of a somewhat unstable and highly active agent [9,13,14]. In this context, electrochemical technology offers the possibility of generating H2O2 in situ, in aqueous medium at a range of ⇑ Corresponding author. Tel.: +55 16 33738659; fax: +55 16 33739903. E-mail address:
[email protected] (M.R.V. Lanza). http://dx.doi.org/10.1016/j.jelechem.2014.03.007 1572-6657/Ó 2014 Elsevier B.V. All rights reserved.
temperatures and pH values, and at concentrations reaching up to hundreds of milligrams per liter [15–21]. The primary reactant in the electrochemical process is O2 and this should be dissolved in the reaction medium [22]. The rate of the electrochemical process depends on the efficient replacement of O2 in the electrolyte present at the electrode surface. However, the solubility of O2 in water is low and decreases with increasing temperature of the medium, a factor that determines the operational limit for the synthesis of high concentrations of H2O2 when conventional electrodes are employed. On the other hand, gas diffusion electrodes (GDEs) comprise a porous and hydrophobic structure that enables O2 to be supplied directly to the electrode/electrolyte interface without limit [23–27]. In this manner, the GDE removes the limitations of mass transportation imposed by the low solubility of O2 [27], and allows high yields of H2O2 to be generated in both acidic [28] and alkaline [29] media. A key challenge in developing GDEs specifically for the generation of H2O2 is the choice of catalytic particles. Typically, GDEs are constructed with pigment carbon black, the particles of which have a graphite-type crystalline structure. In order to maximize the efficiency of H2O2 generation and to achieve higher concentrations of the peroxide, particularly at less negative potentials, it is necessary to incorporate modifiers into the graphitic material. Studies have been conducted with various types of modifiers including organic [25] and organometallic [24] compounds and rare earth oxide nanoparticles [26]. Such investigations are particularly important since the results obtained can provide an indication of the mechanism
F.L. Silva et al. / Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
of the oxygen reduction reaction (ORR), which may involve a 2-electron transfer for the formation of H2O2 [30,31] or a 4-electron transfer with application in fuel cells [32,33]. In this research review article, we describe an investigation of the formation of H2O2 in GDEs prepared using carbon black modified with different amounts of iron (II) phthalocyanine (FePc; CAS number 132-16-1). 2. Experimental 2.1. Evaluation of percentage modifier required for the efficient generation of H2O2 An initial study of the efficiency of H2O2 generation on carbon black with or without modifier was performed using the microporous layer technique and a rotating ring-disc electrode (RRDE). A Pine Instruments model E7R9-GC/Pt RRDE, comprising a central disc of glassy carbon and a Pt ring (coefficient of collection N = 0.37), formed the working electrode, with Pt as the counter electrode and Ag/AgCl as the reference electrode. A portion (3 mg) of a mixture containing Printex 6L carbon (Evonik Co.), hereinafter referred to as carbon black, and 0, 0.1%, 1.0%, 3.0%, 5.0% or 10.0% (w/w) of FePc (Aldrich # 379549; dye content 90%) was suspended in 3 mL of water in an ultrasonic bath, and a 20 lL aliquot transferred by pipette onto the glassy carbon disc. A constant flow of N2 was applied to the disc in an open environment in order to expedite evaporation and formation of a single microporous layer of FePc on the surface of the disc. An aqueous alcoholic solution of NafionÒ perfluorinated resin (Aldrich # 510211; resin content 5% by weight) was diluted 1:100 with water, and a 20 lL aliquot was placed on the microporous layer and dried under a N2 flow as described above. Samples containing carbon black with and without modifier were characterized physically using an X-ray fluorescence (XRF) spectrometer (PANalytical BV) equipped with a Rh tube. The electrochemical experiments were performed using an AutoLab 302N PGSTAT bi-potentiostat controller coupled to a Pine Instruments AFMSRCE rotator control box. Carbon black without modifier was employed as reference for the 2-electron transfer ORR, while carbon supported Pt/C electrocatalyst (E-TEK) served as reference for ORR by a 4-electron transfer [18,20,24–26]. In each case, the electrode system was characterized electrochemically by linear voltammetry (LV) in an aqueous electrolyte containing H2SO4 (0.1 mol L1) and K2SO4 (0.1 mol L1). The electrolyte was bubbled with N2 for 20 min prior to the initial determination of LV, following which O2 was bubbled through the electrolyte for 40 min and LV was performed again with the O2 flow maintained. All measurements were made at room temperature (25 °C) with a scan rate of 5 mV s1 and an RRDE rotation of 1600 rpm.
33
area of 20 cm2, while a pure Pt screen (30.0 mm Ø; constructed with 0.25 mm Ø wires spaced 4.0 mm apart) formed the counter electrode and Ag/AgCl was the reference electrode [24–26]. Electrolyses were performed with an aqueous electrolyte (400 mL) containing H2SO4 (0.1 mol L1) and K2SO4 (0.1 mol L1). Samples (0.5 mL) of electrolyte were collected during electrolysis and added to aliquots (4 mL) of a solution of (NH4)6Mo7O24 (2.4 103 mol L1) in H2SO4 (0.5 mol L1) to form yellow solutions, the color intensities of which were proportional to the amount of H2O2 present. Quantitative analysis of H2O2 was performed by comparison of the absorbance of the electrolyte at 350 nm (measured in a Varian Cary-50 UV–Vis spectrometer) with values from a standard calibration curve [24–26]. 3. Results and discussion 3.1. Characterization of rotating ring-disc electrodes With the aim of determining the percentage of FePc that would be most appropriate as GDE modifier for the efficient electrogeneration of H2O2, RRDEs were prepared with carbon black containing different amounts of additive. Samples of the preparations were analyzed by XRF spectroscopy to determine the effectiveness of the modifier in dispersing the carbon black. A shown in Fig. 1A, the characteristic Ka and Kb emission lines of the Fe atom (at 6.4 and 7.1 keV, respectively) were readily detectable in the samples, and the signals were proportional to the FePc present (Fig. 1A insert). A plot of peak area of the Ka signal vs. percentage modifier added to the carbon black revealed a linear relationship (Fig. 1B), a finding that is consistent with the sample preparation process. The additional signals in the region between 18 and 21 keV that were present in all of the spectra (Fig. 1A) were related to emissions of the metals Ru and Rh. These signals, which derived from the emission tube of the spectrometer, were not completely retained by the filter employed in the analysis but, because of their greater energy, they did not interfere with the detection of Fe.
2.2. Generation of H2O2 at the modified gas diffusion electrode Based on the results obtained by LV in the RRDE experiments, a quantitative study was performed of H2O2 generation during electrolysis at constant potential with GDEs comprising unmodified carbon black and carbon black modified with 5.0% of FePc. The GDEs were prepared by the hot pressing (320 °C e 7.5 ton) method using catalytic masses consisting of carbon black, with or without FePc, and 20% (w/w) of a PTFE fluoropolymer dispersion TE-3893 (Dupont; 60% dispersion PTFE.). Preparation of the catalytic masses and construction of the GDEs followed published procedures [7,23–27]. Experiments were carried out in a cylindrical, single compartment, polypropylene electrochemical cell with a total capacity of 450 mL. The working electrode was a GDE with exposed reactional
Fig. 1. (A) X-ray fluorescence spectra of carbon black: (a) without modifier ( ), (b) with 0.1% FePc ( ), (c) with 1.0% FePc ( ), (d) with 3.0% FePc ( ), (e) with 5.0% FePc ( ), or (f) with 10.0% FePc ( ). (B) Area of Ka peak as a function of the percentage of FePc in carbon black.
34
F.L. Silva et al. / Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
RRDE discs bearing microporous layers of carbon black containing different amounts of FePc were characterized electrochemically by LV analysis. The initial LV measurements were recorded while the electrolyte was bubbled with N2 in order to determine the current in the acidic electrolyte. Subsequently, the total current of the experiment, including the contribution to ORR, was measured while the electrolyte was bubbled with O2. This sequence of experiments allowed the current associated with ORR to be determined as the difference between the total current and the currents of redox reactions of water and H+. All of the LV data presented in this article relate to values obtained in the presence of O2 minus the initial values measured in the presence of N2. [18,20,24,25,34,35]. In the RRDE experiments, the Pt ring was polarized at a potential of +1.0 V (vs. Ag/AgCl) and, under this condition, the H2O2 detected at the ring derived by diffusion from the glassy carbon disc [18,21,24,25]. It should be possible, therefore, to correlate the current values at the ring (referring to the formation of H2O2) with those at the disc (referring to ORR) provided that the following assumptions are valid: (i) all species that are produced at the disc by ORR pass to the surface of the ring by virtue of the rotation applied to the electrode, (ii) there is a direct relationship between the active areas of the ring and the disc as indicated by the coefficient of collection (N = 0.37) determined by the manufacturer of the electrode, and (iii) the current detected at the ring originates only from the oxidation of H2O2 [33,35,36]. Based on the above, and considering the hydrodynamic modeling of the RRDE, the current observed at the ring polarized at +1.0 V vs. SCE should be representative of the amount of H2O2 formed at the disc. Fig. 2 displays the LVs obtained using electrodes bearing microporous layers of carbon black containing different amounts of FePc. It is of note that for RRDEs prepared using lower percentages of modifier, i.e. 0.1%, 1.0% and 3.0%, the values of the ring current were close to those recorded with unmodified carbon black. Increasing the amount of FePc to 5.0% resulted in a higher ring current although the potential at the start of H2O2 detection remained unchanged. Interestingly, further augmentation of modifier to 10% produced a decrease in ring current and a shift in ORR potential of
Fig. 2. Linear voltammograms (adjusted for the current in the acidic electrolyte) obtained with RRDE using discs onto which had been deposited microporous layers of carbon black: (a) without modifier ( ), (b) with 0.1% FePc ( ), (c) with 1.0% FePc ( ), (d) with 3.0% FePc ( ), (e) with 5.0% FePc ( ), or (f) with 10.0% FePc ( ). The aqueous electrolyte contained H2SO4 (0.1 mol L1) and K2SO4 (0.1 mol L1); measurements were made at room temperature (25 °C) with a scan rate of 5 mV s1, an RRDE rotation of 1600 rpm, and a potential range of 0 to 1.0 V vs. Ag/AgCl.
approximately 70 mV to more negative values compared with unmodified carbon black, This behavior may be associated with increased resistivity of the microlayer with the increase in the amount of FePc. With regard to the disc current, however, increases in the amount of modifier gave rise to only minor alterations such that the disc current profiles were close to the maximum values and the ORR potential was little changed. This difference in behavior of disc and ring currents was clearly associated with the activity of the modifier. Thus, while FePc did not change the total current of ORR significantly, it appeared to favor ORR current for H2O2 production leading to a higher current at the disc. Similar results have been observed in other studies in which disc currents remained constant in the presence of modifiers, but ring currents exhibited significant variation. In these cases, the authors associated this differential behavior with the presence of modifier in the carbon, which favored the formation of H2O2 by the ORR [18,20,21,24–26,29]. The current efficiency for the generation of H2O2 [I(H2O2)%] and the number of electrons involved in the reaction (nt) can be calculated from the equations [18]:
IðH2 O2 Þ% ¼
nt ¼
200 Ir=N Id þ Ir=N
ð1Þ
4jId j jId j þ INr
ð2Þ
where Ir is the ring current, Id is the disc current and N is the coefficient of collection for the RRDE. The values of these parameters were determined for RRDEs prepared with carbon black containing different amounts of FePc modifier (Table 1). The results reveal that the formation of H2O2 from ORR generally remained unchanged, irrespective of the percentage of modifier present in the carbon black, except for the electrode modified with 5% FePc, which showed a 10% increase in H2O2 formation compared with unmodified carbon black. Additionally, 2.4 electrons were exchanged in the ORR when 5% of modifier was present in the carbon black whereas for all other modified electrodes, and for unmodified carbon black itself, 2.5 electrons were transferred, these values show that the modification in the carbon did not cause significant changes in the total number of exchanged electrons, it can be associated with the mechanism of O2 reduction, where the FePc is electrochemically reduced and then back to the oxidized state chemically reducing the O2, thus the chemical O2 reduction does not promote change in the current system and not promoting changes in the number of electrons. Based on the results presented in Table 1, it can be proposed that the formation of H2O2 follows a mechanism that tends towards a 2-electron exchange, and that the addition of FePc does not promote significant modifications in the process of electron transfer. According to Wang and Hu [36], when the ring/disc current relation approaches 100% efficiency (i.e. N 1) the ORR follows a 2-electron mechanism with negligible contribution from 4-electron exchange. In the present study, although the addition of modifier exerted a small influence on H2O2 formation (increasing from 71.9% in the absence of modifier to 78.2% with 5.0% of FePc),
Table 1 Values of current efficiency of H2O2 formation [I(H2O2)%] and total number of electrons transferred (nt) both calculated from the disc currents at 0.5 V vs. Ag/AgCl in the LV (Fig. 2). Parameter
I(H2O2)% nt
Carbon black
71.9 2.5
Carbon black with FePc modifier
Pt/C
0.1%
1.0%
3.0%
5.0%
10%
71.3 2.5
72.4 2.5
72.0 2.5
78.2 2.4
71.2 2.5
1.5 3.9
35
F.L. Silva et al. / Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
the total number of electrons exchanged in the ORR did not change markedly. The results shown in Table 1 suggest that the mechanism of H2O2 formation in the presence of FePc-modified carbon black is electrochemical/chemical, in which the electrochemical reduction of Fe2+ in phthalocyanine is followed by auto-oxidation to Fe2+ with chemical reduction of O2 to H2O2. Such a mechanism would contribute to the increased formation of H2O2 (the chemical step), as observed in the present study, but should not influence the number of electrons exchanged because the electrochemical reduction of FePc likely involves a 2-electron transfer [18,20,26,34,36,37]. The value for nt calculated according to Eq. (2) relates to the global number of electrons transferred during electrolysis rather than to the exact number involved in the formation of H2O2, since additional reactions occur in parallel to ORR. Therefore, it is not possible to determine the exact contribution of each of the mechanisms involved. An alternative evaluation of the number of electrons exchanged can be performed using the Koutecky´–Levich (KL) equation (Eq. (3)), which relates the current in the disc (i) to the rotation of the electrode (x):
1 1 1 ¼ þ i ik 0:6nFAD23 v 16 C 0 x12
ð3Þ
where n is the number of electrons transferred in the half reaction, F is the Faraday constant, A is the electrode area, D is the diffusion coefficient, t is the kinematic viscosity and C0 is the concentration of analyte [18,34,37]. Fig. 3 shows KL plots of the ORR for C/Pt and for carbon black with different amounts of FePc. Unmodified carbon black, which is a reference material for the reduction of O2 involving a 2-electron transfer [20], produced a linear KL plot with a slope of 53, while C/Pt, the reference for the reduction of O2 involving a 4-electron transfer [21], gave a KL plot with a slope of 33. These results are as expected and predicted in the literature [20,21,37,38]. The straight lines in Fig. 3 relating to carbon black modified with FePc exhibit slopes that are close to 51 ± 2.4. This behavior provides further evidence that the addition of FePc to carbon black maintains the tendency toward a 2-electron transfer mechanism for ORR [21,27,38]. Another factor to be noted in the KL plots relates to the contributions of the kinetic and diffusion currents. All of the
8
materials studied showed intercepts on the y-axis close to zero, indicating that the major contribution was from the diffusion current while that of the kinetic current, although present, was small and could be ignored for the purpose of evaluating tendencies of the ORR mechanism. In the case of carbon black modified with 5% FePc, for example, the total current in the disc was 237 lA at a potential of 0.4 V (vs. Ag/AgCl), while the kinetic current obtained from the KL equation was 0.7 lA. This small contribution of kinetic current to O2 reduction suggests that electron transfer (charge transfer) was very fast while the reposition of O2 to the surface of electrode (mass transport) was slow, thereby limiting the reduction reactions [18,20]. 3.2. Controlled potential experiments with GDE In order to conduct a more extensive study of ORR in H2O2 generation, it was necessary to scale up the process so that the formation of H2O2 could be analyzed quantitatively. To meet this requirement, GDEs were constructed with unmodified and modified carbon black, and the reduction of O2 to H2O2 was quantified by UV–Vis spectrophotometry. The results obtained with RRDE using microporous layers revealed that carbon black modified with 5% of FePc produced the highest current in the ring (Fig. 2) and the highest value of current efficiency for H2O2. According to Barros [24], tendencies observed in experiments with microporous layers at the RRDE can be extrapolated to the generation and quantification of H2O2 in GDEs. On this basis, GDEs comprising carbon black and carbon black containing 5% FePc were subjected to LV in the range 0.4 V to 4.0 V (vs. Ag/AgCl). The LV plots shown in Fig. 4 were obtained by subtracting the currents measured with the GDE under N2 pressure from those measured under O2 pressure, as described by Barros [24], in order to display current profiles adjusted for the component of current in the electrolyte. The variation of ORR current in the GDE with 5% FePc displayed two sections with different profiles in which a larger increase in ORR current could be observed in the region up to 2.0 V. These regions may represent two distinct sectors for the reduction of O2, as described in the literature in general, with the preferential formation of H2O2 (2-electron mechanism) at less negative potentials, and the preferential formation of H2O (4-electron mechanism) at more negative potentials [18,20,26,34,36,37]. For both unmodified and modified GDEs, the current continued to increase with applied potential (Fig. 4), and it was not possible
7 0.0 6
-0.2 4
i/A
i -1 / mA
-0.1 5
3
-0.3 -0.4
2 -0.5 1 0.02
0.04
0.06
0.08
0.10
-1/2
ω
-0.6 -4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
E vs Ag/AgCl / V Fig. 3. Koutecky´–Levich plots for ORR with RRDE using discs onto which had been deposited microporous layers of C/Pt ( ) or carbon black: without modifier ( ), with 0.1% FePc ( ), with 1.0% FePc ( ), with 3.0% FePc ( ), with 5.0% FePc ( ), or with 10.0% FePc ( ). The aqueous electrolyte contained 1 1 H2SO4 (0.1 mol L ) and K2SO4 (0.1 mol L ). Koutecky´–Levich plots was calculated from the currents at 0.5 V vs. Ag/AgCl in the LV (Fig. 2).
Fig. 4. Linear voltammograms (adjusted for the current in the acidic electrolyte) obtained with GDEs constructed with carbon black ( ) and carbon black with 5% FePc ( ). The aqueous electrolyte contained H2SO4 (0.1 mol L1) and 1 K2SO4 (0.1 mol L ): measurements were made at room temperature (25 °C) with a scan rate of 20 mV s1, and a potential range of 0.4 to 4.0 V vs. Ag/AgCl.
36
F.L. Silva et al. / Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
Fig. 6. (A) Global kinetic constants for the formation of H2O2 (determined during the first 30 min of electrolysis) plotted as a function of the potential applied to GDEs constructed with carbon black ( ) and carbon black with 5% FePc ( ). (B) Energy consumption (kW h kg1) for the final concentration of H2O2 after 90 min of electrolysis plotted as a function of the potential applied to GDEs constructed with carbon black ( ) and carbon black with 5% FePc ( ).
Fig. 5. Production of H2O2 as a function of electrolysis time with GDEs constructed with (A) unmodified carbon black, and (B) carbon black with 5.0% FePc under constant applied potentials of 0.4 V ( ), 0.5 V ( ), 0.6 V ( ), 0.7 V ( ), 0.8 V ( ), 0.9 V ( ), 1.0 V ( ), 1.1 V ( ), 1.2 V ( ), 1.3 V ( ), and 1.4 V ( ). Panel (C) shows the final concentration of H2O2 as a function of the potential applied to GDEs constructed with carbon black ( ) and carbon black with 5% FePc ( ).
to determine the best potential for H2O2 generation. Therefore, separate experiments were performed at constant potentials in the range 0.4 V to 1.4 V (vs. Ag/AgCl), and the concentration of H2O2 generated at each potential was evaluated as a function of time of electrolysis. For the GDE with unmodified carbon black
(Fig. 5A), the variation in H2O2 generation was almost linear with respect to time over the 90 min of the experiment for all applied potentials. However, for the GDE comprising carbon black with 5% FePc (Fig. 5B), the variation in H2O2 concentration was linear over the full 90 min for potentials up to 0.9 V (vs. Ag/AgCl), but for more negative potentials linearity occurred only within the first 60 min, after which H2O2 concentration either stabilized or showed a decrease with respect to time. This change in profile was associated with the geometry of the electrochemical cell employed in which H2O2 generated in the GDE could be oxidized at the anode or reduced at the cathode. Thus, the concentration of H2O2 observed in the last 30 min of electrolysis reflected the balance between formation of the peroxide and its degradation at the anode and cathode. In Fig. 5C, the final concentrations of H2O2 generated after 90 min of electrolysis are plotted as functions of the potential applied to the GDEs studied. The GDE without modifier showed a tendency to increase the final H2O2 concentration with increasing applied potential up to 1.0 V (maximum H2O2 concentration 175 mg L1), while at more negative potentials the final H2O2 concentration decreased. In the case of the GDE modified with FePc, the final H2O2 concentration increased with increasing applied potential up to 1.0 V (maximum H2O2 concentration 240 mg L1), following which H2O2 generation stabilized such that that increases in the applied potential did not give rise to significant increases in the final concentration of H2O2. It is clear from the plots shown in Fig. 5A and B that the change in H2O2 concentration was practically linear during the first minutes of the experiment, probably indicating zero order kinetics.
F.L. Silva et al. / Journal of Electroanalytical Chemistry 722-723 (2014) 32–37
However, the change observed in H2O2 concentration with respect to time is determined by the cumulative effects of all of the reactions that occur in parallel at the cathode and the anode. Thus, the electrosynthesis of H2O2 in a GDE generally follows global pseudo zero order kinetics [6,7,27]. Values of the kinetic rate constant for the formation of H2O2 at different applied potentials were calculated from the angular coefficients of plots of H2O2 concentration (mg L1) vs. time (min), considering only the first 60 min of each experiment (Fig. 6A). For the GDE without modifier, increasing the applied potential promoted an increase in the rate of H2O2 formation, which attained 3 mg L1 min1 at 0.8 V (vs. Ag/AgCl): at more negative potentials the rate of H2O2 formation was observed to decrease. For the GDE modified with 5% FePc, the maximum rate of formation of H2O2 was 7 mg L1 min1 and this was achieved at 1.0 V (vs. Ag/AgCl): at more negative potentials the rate of H2O2 formation tended to stabilize. The observed rate constants are close to the values reported by Barros [24], who obtained 4.7 mg L1 min1 for the best modifier studied, and Valim [25], who achieved 5.9 mg L1 min1 for a GDE modified with tert-butyl anthraquinone in acidic medium. A further important factor in the electrogeneration of H2O2 is the energy consumption (EC) required to generate 1 kg of H2O2. In the present study, EC values were calculated, according to the method described in the literature [23–25,27,28], on the basis of the final concentrations of H2O2 presented in Fig. 5C. As shown in Fig. 6, EC values increased as the applied potential increased, as expected, and were similar for unmodified and modified GDEs at both low and high applied potentials. However, at a potential of 1.0 V (vs. Ag/AgCl), where the concentrations of H2O2 and the rate constants for peroxide formation were maximal for both electrodes, the unmodified GDE consumed 300 kW h kg1 H2O2 whereas the GDE modified with 5% FePc consumed only 165 kW h in the formation of 1 kg of H2O2. These results indicate that the best potential for the generation of H2O2 in the GDE is 1.0 V (vs. Ag/AgCl), under which conditions the highest concentrations of H2O2 were formed with the lowest energy consumption. 4. Conclusions A study of O2 reduction in aqueous solution using RRDEs comprising microporous layers of carbon black modified with FePc revealed that the presence of modifier induced an increase in ring current indicating an increased formation of H2O2. Moreover, current efficiency for the generation of H2O2 was 78.2% for carbon black with 5% FePc compared with 71.9% for carbon black without modifier, although the number of electrons exchanged in the ORR was not significantly altered by the presence of modifier. Quantitative
37
analysis of the electrogeneration of H2O2 under conditions of constant potential applied to a GDE comprising carbon black modified with 5% FePc showed that H2O2 formation was linear with respect to time during the first 60 min of electrolysis, after which the rate decreased with time. The maximum concentration of H2O2 (240 mg L1) attained at the end of 90 min of electrolysis with a potential of 1.0 V (vs. Ag/AgCl) applied to the modified GDE was greater than that obtained with the unmodified GDE (175 mg L1). Under these conditions, the rate constant (determined for the first 60 min of electrolysis) for the formation of H2O2 in the modified GDE was 7 mg L1 min1, and the energy consumed in the generation of 1 kg of H2O2 was 165 kW h. These results demonstrate the feasibility of using GDEs modified with FePc in the efficient electrogeneration of H2O2. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]
E. Brilas, E. Mur, J. Casado, J. Electrochem. Soc. 143 (1996) L49–L53. E. Guinea et al., Water Res. 42 (2008) 499–511. E. Guinea et al., Electrochim. Acta 55 (2010) 2101–2115. A. Gallegos-Alverez, D. Pletcher, Electrochim. Acta 44 (1999) 2483–2492. A.G.F. Beati et al., Quim. Nova 32 (2009) 125–130. J.C. Forti et al., J. Appl. Electrochem. 37 (2007) 527–531. J.C. Forti et al., Electroanal. Chem. 601 (2007) 63–67. T. Harrington, D. Pletcher, J. Electrochem. Soc. 146 (1999) 2983–2989. A. Huissoud, P. Tissot, J. Appl. Electrochem. 28 (1998) 653. M. Teranishi, S. Naya, H. Tada, J. Am. Chem. Soc. 132 (2010) 7850–7851. Y. Nomura et al., ChemSusChem 1 (2008) 619–621. L. Fu et al., J. Chem. Technol. Biotechnol. 85 (2010) 715–719. A. Huissoud, P. Tissot, J. Appl. Electrochem. 29 (1999) 11–16. A. Huissoud, P. Tissot, J. Appl. Electrochem. 29 (1999) 17–25. B. Gözmen et al., J. Hazard. Mater. 168 (2009) 129–136. S.-H. Chang et al., J. Hazard. Mater. 166 (2009) 1279–1288. S. Hanaa et al., J. Hazard. Mater. 175 (2010) 858–865. R.M. Reis, R.S. Rocha, M.R.V. Lanza, ECS Trans. 43 (2012) 103–109. M.L. Calegaro, F.H.B. Lima, E.A. Ticianelli, J. Power Sources 158 (2006) 735–739. M.H.M.T. Assumpção et al., Carbon 49 (2011) 2842–2851. F.H.B. Lima, M.L. Calegaro, E.A. Ticianelli, Electrochim. Acta 52 (2007) 3732– 3738. C.A.R. Rabnini, R.A. Di Iglia, R. Bertazzoli, Quim. Nova 24 (2001) 252–256. R.S. Rocha et al., Quim. Nova 35 (2012) 1961–1966. W.R.P. Barros et al., Electrochim. Acta 104 (2013) 12–18. R.B. Valim et al., Carbon 61 (2013) 236–244. M.H.M.T. Assumpção et al., Appl. Catal. A Gen. 411–412 (2012) 1–6. R.M. Reis et al., Ind. Eng. Chem. Res. 51 (2012) 649–654. R.S. Rocha et al., Quim. Nova 32 (2009) 354–358. M.H.M.T. Assumpção et al., Int. J. Electrochem. Sci. 6 (2011) 158615–158696. T.S. Olson et al., J. Electrochem. Soc. 157 (2010) B54–B63. N. Guillet et al., J. Appl. Electrochem. 36 (2006) 863–870. M. Bron et al., J. Appl. Electrochem. 32 (2002) 211–216. L. Jiang et al., J. Electroanal. Chem. 629 (2009) 87–93. I. Roche et al., J. Phys. Chem. C 111 (2007) 1434–1440. Y. Feng, T. He, N. Alonso-Vante, Electrochim. Acta 54 (2009) 5252–5256. F. Wang, S. Hu, Electrochim. Acta 51 (2006) 4228–4235. P. Hernandez-Fernandez et al., Appl. Catal. B Environ. 77 (2007) 19–28. S. Marcotte et al., Electrochim. Acta 50 (2004) 179–188.
