Abstract
Osmium electrocatalysts for the oxygen reduction reaction (ORR) were prepared by microwave irradiation of at different experimental conditions. The materials obtained were structurally characterized by FT-IR, micro-Raman spectroscopy and X-ray diffraction. Their chemical compositions were obtained by EDS. The electrocatalytic properties for the oxygen reduction reaction were evaluated by rotating disk electrode measurements in 0.5 mol L−1 , in the absence and presence of aqueous methanol. The kinetic parameters, such as Tafel slope, exchange current density, and charge transfer coefficient are reported.
1. Introduction
The PEMFC and DMFC have much in common, in particular their MEAs (Membrane Exchange Assembly).However, Direct Methanol Fuel Cells (DMFCs) show some advantages in comparison to Proton Exchange Membrane Fuel Cells (PEMFCs), mainly because methanol is an inexpensive, readily available, easily storable, and transportable liquid, whereas the power density and efficiency are considerably higher in a PEMFC (~60%) than in a DMFC (~35%). Another disadvantage is that the DMFC anode is limited by a poor electrochemical activity (kinetic loss) which can account for a voltage loss in the cell of more than 0.3 V at 500 mA cm−2 (at 90°C). On the other hand, the crossover effect, that is, permeation of methanol from the anode to the cathode, causes depolarization of the latter and hence severe cell power losses [1].
There are two main ways to minimize this crossover effect, one of them is to use a membrane with a higher thickness; and the second one is using an oxygen electroreduction selective cathode catalyst. At present, four main classes of oxygen-reduction catalysts are known [2]. The most well known are the noble metals, particularly platinum; a second class of electrocatalysts is made up of macrocyclic complexes of a wide range of transition metals; a third class of catalysts is derived from metallic oxides; related to the oxides are the members of the fourth class of electrocatalysts, which are based on transition metal chalcogenide compounds. Some authors have reported the use of osmium and some of its compounds and alloys as methanol tolerant catalysts for the oxygen reduction reaction [3–8]. On the other hand, osmium is a good candidate for alloying with Pt-Ru for the methanol oxidation reaction, because Os is expected to modify the adsorptive properties of Pt-Ru given its ability to adsorb water, and thereby oxygen, in acidic solutions at potentials slightly more negative than for Ru [9, 10].
Many of these catalysts have been synthesized using a conventional heating method, that is, using organic solvents at their refluxing temperature or by pyrolysis of some metal precursor. However, these methods employ relatively high synthesis times (10–24 hours). Recently, microwave energy has been used in many chemical reaction studies and has been found to change the kinetics and selectivity, often in favorable ways. These reactions include organic and inorganic syntheses, selective sorption, oxidations/reductions, and polymerizations, among many other processes [11]. Our research group has reported the synthesis of ruthenium catalysts using microwave irradiation, the materials obtained showing electrocatalytic properties similar to those of materials synthesized by a conventional method [12]. In this paper, a rapid synthesis method based on microwave heating for preparing osmium electrocatalysts with potential as DMFC cathodes is presented.
2. Experimental
2.1. Synthesis and Structural Characterization of the Catalysts
The osmium electrocatalysts were synthesized using 0.063 mmol of triosmium dodecacarbonyl (, Aldrich), which was mixed with 5 mL of 1,2-dichlorobenzene (b.p. 178–180°C, Aldrich) and treated thermally using microwave irradiation on a Discover-BenchMate at different power (80, 100 watts) and time (30 and 60 minutes) conditions, at 180°C. The products obtained were washed with isopropyl alcohol (J. T. Baker) and dried at room temperature. The electrocatalysts synthesized were structurally characterized using reflectance FT-IR spectroscopy on a Perkin-Elmer-GX3 spectrometer, with the samples dissolved in FT-IR grade KBr. The micro-Raman spectra were recorded at room temperature in a Dilor LabRam microspectrometer, using a He-Ne laser (632.8 nm). The XRD studies were performed on a Rigaku D/max-2100 diffractometer, with Cu Kα1 irradiation (1.5406 Å). A Philips XL30ESEM microscope was used to obtain energy-dispersive X-ray spectra (EDS) of the catalysts.
2.2. Electrochemical Experiments
2.2.1. Electrode Preparation
The working electrode for the rotating disk electrode (RDE) studies was prepared by mixing 1.7 mg of Vulcan XC-72 (Cabot) and 0.3 mg of the catalyst with 10 μL of a 5% Nafion solution (ElectroChem) in an ultrasonic bath; 2 μL of the resulting mixture was deposited on a glassy carbon disk electrode and dried at room temperature. The cross-section (geometrical) area of the disk electrode was 0.072 cm2.
2.2.2. Equipment
Measurements were carried out at 25°C in a conventional electrochemical cell with a three-electrode arrangement. A mercury sulfate electrode (Hg//0.5 mol L−1 ; abbreviated as MSE) was used as reference (MSE = 0.680 V/NHE), which was connected to the cell through a bridge with a Luggin capillary. All potential values, however, are referred to the normal hydrogen electrode (NHE). The counter electrode was a graphite rod and 0.5 mol L−1 was used as electrolyte, which was prepared with 98% sulfuric acid (J. T. Baker) and deionized water (18.2 MΩ-cm). A potentiostat/galvanostat (Solartron 1287) and a PC with CorreWare software were used for the electrochemical measurements. A Radiometer Analytical BM-EDI101 glassy carbon rotating disk electrode (with a CTV101 speed control unit) was used for the voltammetry studies.
2.2.3. Cyclic Voltammetry (CV)
Prior to all measurements, the electrolyte was purged with nitrogen (Infra; UHP). The activation of the electrode was done by scanning (cyclic voltammetry) between 0 and 0.98 V/NHE, at 20 mV/s, until no change on the voltammograms was observed (30 cycles). A 30% Pt/Vulcan XC-72 electrode was used for comparison, which was scanned between 0 and 1.58 V/NHE, at a 50 mV s−1 rate. The temperature of the system was maintained at 25°C.
2.2.4. Linear Sweep Voltammetry (LSV)
The electrolyte was saturated with pure oxygen (Infra; UHP) for 15 minutes. Polarization curves were obtained in the presence of oxygen in the () to 0 V/NHE range for the new materials and in the to 0.2 V/NHE range for the 30% Pt/Vulcan XC-72 electrode, at a 5 mV s−1 scan rate. Rotation rates ranged from 100 to 900 rpm. CV and LSV curves were also performed in the presence of methanol (2.0 mol L−1), under the same conditions described above.
3. Results and Discussion
3.1. Structural Characterization
Figure 1 shows the FT-IR spectra of the precursor , as well as those of the osmium electrocatalysts prepared. The precursor shows strong carbonyl stretching vibration bands around 2040 cm−1, as well as a group of bands around 570 cm, which have been assigned to carbonyl deformation modes, [13]. The stretching bands are also present in the infrared spectra of the osmium catalysts synthesized, which indicates that carbonyl groups are still present in these products. This was confirmed by the micro-Raman spectra (Figure 2), where small signals around 2040 cm−1, corresponding to carbonyl groups, are observed.
On the other hand, Uribe-Godínez et al. [6] reported that when is treated at temperatures higher than 120°C, the products obtained basically consist of osmium nanoparticles, that is, a virtually complete decarbonylation is assumed. This effect is not observed when microwave irradiation was used in the present work. Other FT-IR bands can be observed at ca. 2850 and 2930 cm−1, as well as weaker signals around 1250 and 1450 cm−1, which fall within the C–H stretching and C–C ring vibrations ranges of substituted benzenes, respectively [14]. This behavior was also observed when osmium electrocatalysts were obtained in o-xylene under reflux conditions [7]. Hence, it could be indicative of a possible coordination of aromatic molecules of the solvent to metal centers.
The presence of carbonyl groups and aromatic molecules in the structure of these osmium catalysts was corroborated by their chemical composition (Table 1), which shows the presence of carbon, oxygen, and chlorine. Figure 3 shows the corresponding XRD patterns of precursor and the osmium catalysts synthesized. The products show some very broad crystallographic peaks at low angular values (12–35), indicating the small size of the particles.
3.2. Electrochemical Characterization
3.2.1. Cyclic Voltammetry
The cyclic voltammograms of the different osmium electrodes in the absence and presence of 2 mol L−1 methanol are shown in Figure 4, as well as that of 30% Pt/Vulcan XC-72 for comparison. The osmium materials show anodic-cathodic peaks in the 0.6–0.8 V (NHE) region, which are ascribed to the Vulcan support. Hydrogen evolutions peaks are observed in the cathodic 0-0.1 V/NHE region, while the beginning of the oxygen evolution zone appears in the anodic 0.85–0.98 V/NHE region. Although the materials do not show methanol oxidation peaks, it can be observed a slight increase in the current density, due probably to the methanol adsorption on the surface. On the other hand, the platinum electrode shows very sharp methanol oxidation peaks, indicating its high activity for this reaction [15, 16].
(a)
(b)
(c)
(d)
3.2.2. Linear Sweep Voltammetry: Oxygen Reduction Reaction
Figure 5 shows typical curves for the electrochemical reduction of molecular oxygen in 0.5 mol L−1 of the osmium electrocatalysts in the absence and presence of 2 mol L−1 methanol, along with those of 30% Pt/Vulcan XC-72 as reference.It can be observed that the material with the highest time and power conditions used during the microwave assisted synthesis (Figure 5(c)), shows the highest current density; a slight current decrease is observed with methanol in the electrolyte. The presence of this contaminant also induces a decrease in the open circuit potential of ca. 0.16 V/NHE in all cases. However, these features are far more pronounced in the case of the Pt electrode, for which methanol caused the net cathodic current onset to shift negatively by ca. 0.5 V/NHE (Figure 4(d)).
(a)
(b)
(c)
(d)
The Koutecky-Levich equation at a given potential is where i is the measured disk current, ik is the kinetic current, ω the electrode rotation speed in rpm, and B is a constant given by [17]: where n is the number of electrons exchanged per mol of O2, F the Faraday constant, A is the effective catalytic surface area, ν is the kinematic viscosity of the electrolyte (0.01 cm2 s−1), is the oxygen diffusion coefficient (1.4 × 10−5 cm2 s−1) and the bulk oxygen concentration in the electrolyte (1.1 × 10−6 mol cm−3) [18]. Figure 6 shows the theoretical (with n = 2 and 4; A = electrode geometric area, 0.072 cm2) and experimental Koutecky-Levich plots in the absence and presence of 2 mol L−1 methanol, at a given potential value (0.4 V/NHE) for the osmium electrocatalysts, along with those of the 30% Pt/Vulcan electrode as reference. The experimental plots are closer to those calculated for a four-electron process, as is observed for the platinum electrode, that is, the oxygen molecules are most likely reduced directly to water by the osmium catalysts. It can be observed a slight separation between plots in the absence and presence of methanol, a behavior expected from the linear sweep voltammograms. The differences between the experimental and theoretical Koutecky-Levich plots may result from the exposed catalytically active area of the materials, which might be higher than the geometric ones [19]. The current-potential curves were corrected by the procedure described by Gojković et al. [20]; these curves, in turn, yielded the mass-corrected Tafel plots for the ORR in the absence and presence of methanol presented in Figure 7.
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
Table 2 summarizes the kinetic parameters (obtained from the Tafel Plots) and open circuit potentials () of the osmium electrocatalysts, compared with those of 30% Pt/Vulcan XC-72 and some values reported in the literature.It can be observed that in the absence of methanol all the osmium catalysts show ≥ 0.75 V/NHE, while in its presence such values decrease by approximately 0.16 V/NHE. On the other hand, although the Pt electrode shows the highest in the absence of methanol (~1.0 V/NHE), it shows the lowest value in the presence of this contaminant due to a mixed potential phenomenon [21].
Tafel slope is a parameter related to the reaction mechanism [22]; the osmium materials show in general Tafel slopes higher than platinum, which suggests that they follow a different ORR mechanism than this electrode [23]. In the absence of methanol, platinum shows the highest charge transfer coefficient (), that is, it is the catalyst with the largest decrease of reaction free energy [24]. However, in the presence of the contaminant platinum becomes inactive, while the osmium electrocatalysts do show electrocatalytic activity for the ORR, with α values around 0.2.
In general, the electrocatalysts synthesized show similar exchange current densities (~10−4 mA cm−2) in the absence of methanol, of the same order of Pt. However, in the presence of this compound, the exchange current densities of the Os catalysts decreased only one order of magnitude, while Pt lost its activity for the reaction.
All these parameters are in agreement with those of osmium electrocatalysts synthesized by conventional thermolysis methods (Table 2) [6, 7]. Therefore, an improvement of the synthetic method for this kind of materials has been achieved in this work, in the sense of employing shorter reaction times with similar results.
4. Conclusions
Osmium electrocatalysts for the oxygen reduction reaction in the absence and presence of methanol were synthesized using microwave irradiation. All the materials are constituted by metal carbonyl particles, and are probably coordinated to aromatic molecules from the solvent. These electrocatalysts show an activity comparable to other osmium catalysts reported in the literature. In addition, they show a significant tolerance to the presence of 2 mol L−1 methanol during the ORR, unlike traditional platinum catalysts. With this microwave method it is possible to reduce the relatively long times used in the conventional thermolysis processes, from 20 hours to only 30–60 minutes, obtaining materials with similar electrocatalytic properties for the ORR.
Acknowledgments
This work was supported by CONACYT through project 100212 and by DGAPA-UNAM through Project IN103410. The authors wish to thank R. A. Mauricio-Sánchez, M. A. Hernández-Landaverde, and J. E. Urbina-Alvarez (CINVESTAV-Querétaro) for valuable technical assistance. Postdoctoral scholarships from CONACYT (L. Magallón-Cacho) and UNAM (E. Borja-Arco) are also acknowledged.