Carbon-Supported Pt-SnO₂ Catalysts for Oxygen Reduction Reaction over a Wide Temperature Range: Rotating Disk Electrode Study

Abstract

Pt/C and Pt/x-SnO₂/C catalysts (where x is mass content of SnO₂) were synthesized using a polyol method. Their kinetic properties towards oxygen reduction reaction were studied by a rotating disk electrode (RDE) technique in a temperature range from 1 to 50 °C. The SnO₂ content of catalyst samples was 5 and 10 wt.%. A quick evaluation of the catalyst activity, electrochemical behavior and average number of transferred electrons were performed using the RDE technique. It has been shown that the use of x-SnO₂ (through modification of the carbon support) in a binary system together with Pt does not reduce the catalyst activity in the temperature range of 1–30 °C. The temperature rising up to 50 °C resulted in composite catalyst activity reduction at about 30%.

1. Introduction

Increasing interest in renewable energy sources and hydrogen energy makes the research and development of polymer electrolyte membrane fuel cells (PEMFCs) an important task. A key element of PEMFCs is a membrane electrode assembly (MEA), which includes the anode and cathode catalytic layers. Since the kinetics of the hydrogen oxidation reaction on the anode catalytic layer is rather rapid, the rate-limiting process in the PEMFC is the oxygen reduction reaction (ORR) at the cathode (Figure 1), which is a slow four-electron transfer reaction. The low ORR rate calls for high-effective cathode catalysts showing high activity in this process.

Pt-based carbon-supported catalysts are commonly used at the PEMFC cathode [1,2]. Such electrocatalysts have high activity and sufficient stability under standard operating conditions of PEMFCs. However, widening of the range of external operating conditions (temperature and humidity) involves a number of challenges. First, drying of the ion-exchange polymer is observed at high temperatures and low humidity, which leads to a deterioration in proton conductivity both in the catalytic layer ionomer and in the membrane [3,4,5,6]. Second, at negative temperatures, ice crystals are formed inside MEA components, and destruction of the catalytic layers and membrane occurs during freeze-thaw cycles [7,8,9,10,11,12,13]. These problems can be solved by applying additives with high water retention ability, such as tin dioxide [14,15,16,17,18].

However, the addition of tin dioxide particles may have a negative effect on the activity of the platinum catalysts due to a possible blockage of Pt active sites or formation of large agglomerates of particles [19,20]. Therefore, a detailed study of activity of composite tin dioxide catalysts in the ORR is important for evaluation of their applicability in PEMFCs over a wide range of operating conditions.

The technique of a rotating disk electrode (RDE) in a three-electrode electrochemical cell is commonly used for the determination of electrocatalyst activity towards ORR. This technique allows for estimating kinetic and diffusion components of the reaction current. An extension of this method to low and high temperatures may aid in evaluating the efficiency of the catalysts over a wide temperature range and determining the activation energy in the ORR.

Most studies [21,22,23,24] are dedicated to the evaluation of the activity of tin dioxide composite catalysts in the alcohol oxidation reaction since Pt-SnO₂ hetero-clusters enhance oxidation of CO_ads intermediates. Fewer studies [25,26,27] have focused on investigation of the activity in the ORR of hybrid Pt-SnO₂ catalyst with a high tin dioxide loading, co-catalyst properties of SnO₂ have been reported.

As reported in our previous work [20], the composite electrocatalysts with the tin dioxide content of 5 and 10 wt.% demonstrate a significantly improved stability in the course of the accelerated stress testing and good efficiency within the PEMFC. This effect is achieved due to formation of Pt-SnO₂ hetero-clusters, providing an improvement in the electrocatalyst durability. However, the presence of tin dioxide led to a reduction in the electrochemically active surface area (EASA), and may also affect the ORR kinetics.

This paper is dedicated to further investigation of composite Pt-SnO₂ catalysts. The kinetics of different electrocatalysts in the ORR was investigated using the Koutecky–Levich approach. The study of the kinetics in the temperature range from 1 to 20 °C allows for evaluating the catalyst efficiency in conditions of starting the PEMFC at low ambient temperatures (“cold start”). The upper limit of the measurement temperature of 50 °C is explained by the increased evolution of acid vapors and accelerated corrosion of the RDE unit at higher temperatures.

2. Results and Discussion

2.1. Catalyst Characterization

High quality RDE measurements require thin and uniform films over the entire surface of the electrode [28]. Figure 2 shows micrographs of catalyst films obtained using catalyst ink of the proposed composition. The film in the left image was produced by the method of rotational drying at a certain angle, which is detailed below in Section 3.2, and the film in the right image was produced by stationary drying. Figure 2a shows the film with a uniform structure over the entire surface of the glassy carbon (GC) electrode. The catalyst film in Figure 2b is unevenly distributed over the GC electrode; defects and voids are observed. Thus, according to the micrographs, the proposed method for catalytic ink drying allows obtaining films with a high degree of uniformity.

Figure 3 shows the surface of the GC electrode coated with Pt⁴⁰/C (40 wt.% of Pt on carbon carrier) before and after the RDE measurements in the HClO₄ electrolyte. The distribution of the catalyst particles is very even, free from formation of agglomerates and defects, despite unavoidable formation of some islands (Figure 3a,c) at the sub-micrometer level during drying [29]. The image of the layer surface after the RDE measurements (Figure 3c) did not show significant changes in the surface structure in comparison with the initial layer structure (Figure 3a).

The image of the film before the RDE measurements taken at a higher magnification (Figure 3b) shows thin μm-sized patches of the ionomer layer (circled in red), which is described in [30]. The ionomer performs several important functions such as attaching catalyst particles to one another, obtaining homogenous and well-dispersed ink, and uniform application of the ink [31]. Such ionomer patches disappeared after the series of RDE measurements (Figure 3d), which points to the dissolution of the ionomer film in the electrolyte. The dissolution of the ionomer may lead to an increase in the catalyst layer activity due to better mass transport of oxygen to active sites and lower electrical resistance [31,32]. However, the amount of ionomer in the film and the number of patches were rather small, so the ionomer dissolution should not have a significant effect on the quality of measurements.

The absence of significant changes in the layer surface structure allows for obtaining high-quality results during the entire series of RDE measurements in the wide temperature range.

2.2. Electrochemical Studies

Figure 4 presents cyclic voltammograms (CVs) of different catalysts. The EASA values are 42, 83, 55 and 57 m² g⁻¹ Pt for Pt⁴⁰/C, Pt²⁰/C, Pt²⁰SnO₂⁵/C and Pt²⁰SnO₂¹⁰/C, respectively.

The CVs demonstrate well-defined hydrogen adsorption/desorption peaks in the potential range of 0.05–0.40 V vs. RHE. Two peaks in the hydrogen desorption region at 0.13 and 0.20 V correspond to platinum (110) and (100) active sites, respectively, [26,33]. The sample with SnO₂ content of 10 wt.% demonstrates a shift of the hydrogen desorption peak to more positive potentials, which may be attributed to the effect of the presence of tin dioxide on the surface structure of platinum nanoparticles and various distributions of crystal orientations [34]. The smaller EASA for composite catalysts can be explained by the more facile adsorption of OH⁻ species, which impedes hydrogen adsorption and by possible blockage of Pt sites by SnO₂ nanoparticles. A small peak at 0.73 V for the CVs of the modified catalysts may be attributed to the oxygen adsorption from the dissociation of water on the surface of tin dioxide particles [35].

Measurement of the electrolyte solution resistance is an important factor for obtaining correct values from processing the data acquired by the RDE technique. An almost two-fold decrease in resistance with increasing temperature was observed (Figure 5). The obtained values of electrolyte solution resistance were used in further calculations and plotting of polarization curves.

Figure 6 shows the polarization curves for a Pt²⁰/C-coated electrode at different electrolyte temperatures and at the electrode rotation speed of 1600 rpm. The current density on the diffusion-limited plateau of the polarization curves increases with temperature, indicating that the decrease in oxygen solubility with temperature was lower than the increase in the oxygen diffusion coefficient [36]. A shift of the half-wave potential of the polarization curves to the high-potential region is also observed with increasing temperature, which characterizes the positive dependence of the reaction rate on temperature.

Table 1. Comparison of catalyst parameters: EASA, area-specific (S_a) and mass-specific (M_a) activity of catalysts at 20 °C and at potential 0.9 V.

Catalyst	Synthesis Method	EASA, m2g−1	Sa, mA cm2	Ma, mA mg−1	Scan Rate, mV s−1
Pt40/C	Polyol method	42	0.14 ± 0.01	56 ± 1	10
Pt40/C [37]	Polyol method	49 ± 1	0.73 ± 0.03	359 ± 14	20
Pt40/C [38]	Polyol method	49 ± 1	0.48 ± 0.03	230 ± 10	20
Pt40/C [39]	Polyol method	43	0.21	90	10
Pt20/C	Polyol method	83	0.25 ± 0.04	207 ± 4	10
Pt20/C [26]	Carbonyl route	63 ± 2	0.37 ± 0.01	75 ± 1	10
Pt20/C [26]	Photo-deposition	72 ± 2	0.41 ± 0.02	94 ± 5	10
Pt20/C [38]	Polyol method	55 ± 10	0.33 ± 0.06	180 ± 6	20
Pt20/C [28]	-	66	0.20	160	5
Pt20SnO25/C	Polyol method	55	0.37 ± 0.01	205 ± 2	10
Pt20SnO210/C	Polyol method	57	0.36 ± 0.01	205 ± 5	10
Pt20SnO220/C [26]	Photo-deposition	33 ± 1	0.44 ± 0.05	49 ± 4	10
Pt20SnO220/C [27]	Polyol method	53	0.3	156	10

shows the values of EASA and activities of the catalysts under the study at 20 °C and the values obtained by the other groups. The polyol method is the most commonly used method for the synthesis of platinum catalysts.

The electrochemical surface area and activity values for the Pt⁴⁰/C catalyst proved to be lower than the values reported in literature. The relatively low values can be explained by the catalyst preparation technique [20], in which the carbon support was preliminarily impregnated with a precursor followed by reduction. At the same time, with an increase in the platinum content, the average size of platinum nanoparticles exceeded 3.5 nm, which was described in our previous work [40]. According to [41], for particles with a size less than 3.5 nm, predominant for catalysts with a lower platinum content, the crystal orientation of the surface (110) dominates, and its activity in ORR is higher than for (100) Pt sites [42].

The values of EASA and mass activity of the Pt²⁰/C are comparatively higher than the values reported in the literature. The difference in the values may be due to different techniques used for recording polarization curves and catalyst film preparation. However, a comparative analysis of the obtained values is possible.

Catalysts with hybrid support and SnO₂ loading of 20 wt.% have the lower values of EASA and mass activity, which can be attributed to the increase in the particle size with an increase in the concentration of tin dioxide [20].

Figure 7 shows the Koutecky–Levich (K-L) plots for catalysts with a platinum content of 20 wt.% at 50 °C. The plots remain almost parallel up to a potential of 0.9 V, pointing to a weak dependence of the number of transferred electrons on the potential and applicability of the K-L theory for increased temperatures. The number of transferred electrons for all samples is close to 4, which suggests high selectivity of samples up to elevated temperatures.

Table 2. Catalyst activity parameters: kinetic current density (j_k), area-specific (S_a) and mass-specific (M_a) activity of catalysts in the temperature range from 1 to 50 °C at potential of 0.9 V.

T, °C	1	10	20	30	50
Pt40/C (EASA = 42 m2 g−1)
jk, mA cm−2	3.7 ± 0.2	4.3 ± 0.3	4.6 ± 0.1	5.2 ± 0.2	5.7 ± 0.2
Sa, mA cm−2	0.12 ± 0.01	0.13 ± 0.01	0.14 ± 0.01	0.16 ± 0.01	0.17 ± 0.01
Ma, mA mg−1	47 ± 2	51 ± 4	56 ± 1	62 ± 2	69 ± 2
Pt20/C (EASA = 83 m2 g−1)
jk, mA cm−2	4.2 ± 0.2	4.8 ± 0.2	5.6 ± 0.1	6.9 ± 0.1	11.5 ± 0.4
Sa, mA cm−2	0.19 ± 0.01	0.22 ± 0.01	0.25 ± 0.04	0.31 ± 0.01	0.52 ± 0.02
Ma, mA mg−1	154 ± 7	178 ± 8	207 ± 4	256 ± 2	428 ± 15
Pt20SnO25/C (EASA = 55 m2 g−1)
jk, mA cm−2	2.9 ± 0.1	3.6 ± 0.1	4.5 ± 0.1	6.3 ± 0.1	7.9 ± 0.3
Sa, mA cm−2	0.24 ± 0.01	0.30 ± 0.05	0.37 ± 0.01	0.52 ± 0.01	0.65 ± 0.02
Ma, mA mg−1	132 ± 1	165 ± 3	205 ± 2	286 ± 6	357 ± 13
Pt20SnO210/C (EASA = 57 m2 g−1)
jk, mA cm−2	3.8 ± 0.1	5.0 ± 0.1	5.9 ± 0.2	7.1 ± 0.2	8.0 ± 0.1
Sa, mA cm−2	0.24 ± 0.04	0.31 ± 0.01	0.36 ± 0.01	0.44 ± 0.01	0.49 ± 0.01
Ma, mA mg−1	134 ± 2	175 ± 2	205 ± 5	248 ± 6	280 ± 4

shows the values of kinetic current and catalyst activity at various temperatures.

The Pt⁴⁰/C catalyst demonstrates low activity values over the whole temperature range due to the larger platinum nanoparticle size and the smaller fraction of the platinum surface involved in the ORR.

The polarization curves for catalysts with platinum content of 20 wt.% have close values of kinetic current density and activity up to 30 °C. The structure of the sample with the lowest tin dioxide content Pt²⁰SnO₂⁵/C is mainly represented by individual highly dispersed Pt and SnO₂ particles, which participate as a co-catalyst in ORR. Thus, Pt²⁰SnO₂⁵/C catalyst has close values of activity in comparison with Pt²⁰/C up to 50 °C. The Pt²⁰SnO₂¹⁰/C sample shows mass activity reduction at about 30% at the temperature of 50 °C, which can be explained by an agglomeration of tin dioxide particles, as well as the formation of Pt-SnO₂ hetero-clusters.

Based on the obtained values, the sample with tin dioxide content of 5 wt.% has activity similar to that of the standard catalyst over the entire temperature range, which together with high stability of this sample [20], makes it promising for use in PEMFCs under a wide range of operating conditions. The sample with tin dioxide content of 10 wt.% also demonstrates high activity, however, at 50 °C, the activity was comparatively lower. Nevertheless, under conditions of increased temperature and low humidity, the use of this catalyst in PEMFC is justified taking into account the high water retention ability of tin dioxide nanoparticles.

The Arrhenius plots for the Pt²⁰/C and Pt²⁰SnO₂¹⁰/C samples under study at the potential of 0.9 V are shown in Figure 8b,d. Some plots demonstrate non-linear behavior. This could be due to activation of additional platinum catalytic centers at increased temperatures and larger kinetic current error than the calculated one.

According to the slope of the Arrhenius plots, activation energies for the ORR are 20 ± 1, 25 ± 1, 30 ± 1 and 25 ± 2 kJ mol⁻¹ for Pt⁴⁰/C, Pt²⁰/C, Pt²⁰SnO₂⁵/C and Pt²⁰SnO₂¹⁰/C, respectively. The activation energies obtained in our study are −28 and 21 ± 3 kJ mol⁻¹, respectively, which agrees well with the values for the catalyst with Pt content of 20 wt.% reported in papers [43,44]. Close activation energy values for the composite catalysts and reference platinum samples is indicative of high efficiency of catalysts with tin dioxide in the ORR. Lower values of the activation energy for Pt²⁰SnO₂¹⁰/C than for Pt²⁰SnO₂⁵/C may be attributed to the presence of the Pt-SnO₂ hetero-clusters and higher activation energy of the SnO₂ active centers of the Pt²⁰SnO₂⁵/C sample.

Thus, note should be made of the low dependence of the ORR kinetics on temperature and high efficiency of the electrocatalysts down to temperatures close to 0 ℃. Such catalysts allow maintaining the PEMFC performance during the “cold start” from low ambient temperatures to standard operating conditions. Catalysts modified with tin dioxide have activity comparable to that of Pt²⁰/C, despite the partial blockage of Pt sites by SnO₂ nanoparticles. The high activity of the composite catalysts suggests that tin dioxide participates in the ORR as a co-catalyst, and high water adsorption ability of SnO₂ makes these catalysts efficient at low temperatures due to prevention of formation of ice crystals in MEA, and at increased temperatures by preventing the PEMFC components from overdrying.

3. Materials and Methods

3.1. Preparation of Catalysts

Pt/x-SnO₂/C and Pt/C catalysts were synthesized using a polyol method in ethylene glycol described in our previous works [20,24].

3.2. Electrode Preparation

Thick catalyst films on a polished GC disk electrode (0.102 cm²; Volta, Russia) were prepared by dropping catalyst ink with an Eppendorf micropipette. The catalyst ink included Nafion^® used as a binder in the amount of 7 wt.% of the specified catalyst weight.

The catalyst inks were prepared by ultrasound treatment. The mixture consisting of 12 mg of catalyst powder, 0.18 mL of isopropanol and 0.2 mL of 0.5% Nafion^® solution in DI water [45] was treated in ultrasonic homogenizer (the catalyst amount was ca. 60 mg mL⁻¹) for 1 h. Five microliters of the prepared catalyst ink were dropped with an Eppendorf micropipette onto the surface of polished GC electrode. The drying method was obtained from experiments with the used ink composition. The GC electrode was neatly tilted at an angle of 45° and evenly rotated at 60 rpm until the ink became dry. The electrode was dried in air at room temperature. The final catalyst loading was about 0.2 mg cm⁻².

3.3. Electrochemical Studies

The cyclic voltammograms (CVs) were measured in 0.1M HClO₄ at 25 °C using a three-electrode glass cell equipped with a polished GC working electrode, a Pt wire counter electrode placed in a fritted glass tube, and RHE as a reference electrode connected to the electrochemical cell by a Luggin capillary. We used the RHE because it has a low level of impurities and does not require potential correction [27].

The electrode was activated in an N₂ saturated 0.1 M HClO₄ solution at the potential range of 0.05 to 1.20 V at a 50 mV s⁻¹ sweep rate for about 30 cycles until a stable CV was obtained. The CVs registered at the potential range of 0.05 to 1.20 V and at a 20 mV s⁻¹ sweep rate were used to characterize the catalyst.

The measurements were performed using a CorrTest CS350 electrochemical workstation (Wuhan, Corrtest Instruments Corp., Ltd., Hubei, Wuhan, China). The catalyst EASA was calculated using the hydrogen adsorption-desorption peaks at 0.05–0.40 V vs. RHE as described in [24,46].

After taking the CV measurements, the background current was measured by running the ORR sweep profile in an N₂-purged electrolyte solution at the electrode rotation speed of 1600 rpm in the potential range of 0.05–1.05 V and at the 10 mV s⁻¹ sweep rate. Then the electrolyte was purged with oxygen for 1.5–2.0 h. The ORR polarization curves were recorded at the electrode rotation speeds of 500, 700, 900, 1200 and 1600 rpm at the 10 mV s⁻¹ sweep rate and in the potential range of 0.05–1.05 V. The electrolyte was purged by O₂ for at least 5 min between the measurements.

To obtain polarization curves at different temperatures (1, 10, 20, 30 and 50 °C), a thermostatically controlled chamber of the three-electrode cell was connected to a water thermostat. The upper temperature was limited by increased acid evaporation and accelerated corrosion of the RDE unit, and the lower temperature limit was determined by freezing of the water used as cooling/heating medium in the thermostat. To accurately control the electrolyte temperature, a thermocouple was installed inside the cell. The polarization curves were recorded after reaching thermal equilibrium. At each temperature the impedance spectroscopy curves were recorded to take the electrolyte solution resistance into consideration in the Koutecky–Levich calculations. The electrochemical impedance spectroscopy curves were plotted with the CorrTest CS350 electrochemical workstation at frequencies of 10⁵ to 0.1 Hz.

3.4. Catalysts Surface Structure Characterization

The uniformity of the formed films was evaluated using a Levenhuk DTX 90 digital optical microscope (USA) with 10–300× magnification.

The electrode surface morphology was observed using a Versa 3D DualBeam scanning electron microscope (SEM) (FEI, Hillsboro, OR, USA) under vacuum. Low vacuum images were obtained using a low vacuum secondary electron detector (LVSED) and a concentric backscattered detector (CBS).

3.5. Calculation Methods

The Koutecky–Levich equation was used for calculation of the kinetic and mass-transfer components of the measured currents:

$$\frac{1}{j}=\frac{1}{j_k}+\frac{1}{j_d}=-\frac{1}{\left|j_k\right|}-\frac{1}{0.62nF{D}_{O_2}^{\frac{2}{3}}{\nu }^{-\frac{1}{6}}{C}_{O_2}^b{\omega }^{\frac{1}{2}}}$$

(1)

where j (mA cm⁻²) is the measured current density, j_k (mA cm⁻²) is the kinetic current density, j_d (mA cm⁻²) is the diffusion-limited current density, n is the average number of electrons transferred per O₂ molecule, F (C mole⁻¹) is the Faraday constant, $D_{O_2}$ (cm² s⁻¹) is the diffusion coefficient of O₂ in the 0.1M HClO₄ solution, $\nu$ (cm² s⁻¹) is the kinematic viscosity,$C_{O_2}^b$ (mol cm⁻³) is the oxygen concentration in the electrolyte, and ω (rad s⁻¹) is the electrode angular rotation rate.

The area-specific (S_a) and mass-specific (M_a) catalyst activity were determined by the following equations:

$$S_a=\frac{i_k}{EASA\times m_{Pt}},$$

(2)

$$M_a=\frac{i_k}{m_{Pt}} ,$$

(3)

where m_Pt (mg_Pt) is the mass of Pt on the electrode and i_k (mA) is the measured kinetic current.

To evaluate the average number of transferred electrons n at temperatures different from room temperature, it is important to take into consideration the temperature dependence of the diffusion coefficient, kinematic viscosity and concentration of dissolved oxygen. The concentration of dissolved oxygen at each temperature was calculated according to Henry’s law:

$$C_{O_2^b}{e}^{\left[\frac{{\mathsf{\Delta }}_{sol}H}{R}\left(-\frac{1}{T}\right)\right]}=const$$

(4)

where ${\mathsf{\Delta }}_{sol}H$ is the enthalpy of solution, R is the universal gas constant and $\frac{{\mathsf{\Delta }}_{sol}H}{R}$ is a constant, equal to 1700 K for oxygen [47]. The values of kinetic viscosity at different temperatures were taken from [48], and the oxygen diffusion coefficient was calculated by the Stokes−Einstein equation [49]:

$$D_{O_2}\nu /T=const$$

(5)

The activation energy of the ORR according to the Arrhenius equation was estimated by multiplying the gas constant by the slope of the log (i_k) vs. 1/T plot. The kinetic current was normalized with the respect to the concentration of dissolved oxygen at each temperature.

4. Conclusions

The ORR kinetics on composite carbon-supported Pt-SnO₂ catalysts were investigated over a wide temperature range. The CVs show a reduction in the EASA for the composite catalysts due to possible blockage of Pt active sites by SnO₂ nanoparticles. The RDE technique was used for determination of kinetic and mass components of the ORR current. K-L plots at 50 °C remain almost parallel up to the potential of 0.9 V, which suggests that the number of transferred electrons weakly depends on the potential, and that the K-L theory can be applied for increased temperatures. The number of transferred electrons for all samples is close to 4, which points to high selectivity of samples at temperatures up to 50 °C. The catalysts modified with tin dioxide demonstrate high activity despite the partial blockage of Pt active sites by SnO₂ nanoparticles and agglomeration. The relatively high activity of the composite catalysts is explained by participation of tin dioxide in the ORR as a co-catalyst. The Pt²⁰SnO₂⁵/C catalyst has close values of activity in comparison with Pt²⁰/C up to 50 °C, but the Pt²⁰SnO₂¹⁰/C sample shows mass activity reduction at about 30% at the temperature of 50 °C, which can be explained by the agglomeration of tin dioxide particles, as well as the formation of Pt-SnO₂ hetero-clusters. According to the slope of the Arrhenius plots, the activation energies for the ORR lie in the range from 20 to 30 kJ mol⁻¹ (at temperatures ranging from 1 to 50 °C). The values of activation energies obtained in our study are in a good agreement with the values of ~28 and 21 ± 3 kJ mol⁻¹ reported in the literature. Thus, the approach used in our study to synthesize modified catalysts allows for producing hybrid catalysts having not only improved stability, but also high activity at the temperature up to 50 °C for the Pt²⁰SnO₂⁵/C and up to 30 °C for the Pt²⁰SnO₂¹⁰/C.