Platinum Nanoparticles Supported on Atomic Layer Deposited SnO₂ Decorated Multiwalled Carbon Nanotubes as the Electrocatalyst for the Oxygen Reduction Reaction

Abstract

Tin(IV) oxide (SnO₂) was deposited onto acid-washed multiwalled carbon nanotubes (MWCNTs) to be used as a support for platinum nanoparticles (PtNPs). The effect of the SnO₂–CNT support on the electrocatalytic activity of the PtNPs for the oxygen reduction reaction (ORR) in 0.1 M HClO₄ solution was investigated. The physical characterization of the catalyst confirms the presence of Pt, Sn and C on the catalyst as well as the presence of the PtNPs on SnO₂. The synthesized catalyst possesses a specific activity of 0.15 mA cm⁻² at 0.9 V, while the commercial Pt/C catalyst showed a specific activity of 0.05 mA cm⁻². Accelerated durability testing (ADT) was performed on both catalysts, with the synthesized PtNP/SnO₂–CNT catalyst retaining over 50% of its initial electrochemically active surface area (ECSA). Thus, the results obtained in this study confirm the positive influence of SnO₂-decorated CNTs on the overall electrocatalytic activity of PtNPs and their stability toward the ORR.

1. Introduction

The warming of the Earth’s temperature, resulting from the climate crisis, has given rise to an increase in the frequency of natural disasters [1]. To help combat this, the 28th UN Framework Convention on Climate Change has called for a transition away from fossil fuels [1]. To achieve the goal of net zero CO₂ emissions by 2050, the International Energy Agency suggests ending all sales of internal combustion engines (ICEs) by 2035 [2]. However, some European governments continue to subsidize fossil fuel use [1,3]. The proximity of the transition timeline further indicates the need for the rapid advancement of reliable and economically viable green energy within the transportation industry. One way this could be achieved is through the implementation and further optimization of polymer-electrolyte membrane fuel cells (PEMFCs). In contrast to ICEs, PEMFCs have only one byproduct: water [4]. Additionally, PEMFCs can operate at both a heavy and a light usage load, making them an optimal green technology for transportation and residential sectors [5]. Despite their environmental benefits, their economic viability is reduced due to the use of Pt-based catalysts in the cathode, which facilitates the oxygen reduction reaction (ORR). Thus, to improve the overall cost effectiveness of PEMFCs, Pt usage must be minimized.

One way Pt minimization can be achieved is through optimization of Pt particle size. As particle size decreases, the total amount of Pt in the fuel cell decreases, while the overall surface area of Pt will increase. The particle size effect states that, as the nanoparticle (NP) size decreases, there is an increase in the overall electrocatalytic activity of the PtNP [6]. However, there is a limit to how small the PtNP can be while still maintaining a high level of activity [7,8]. Shao et al. found that, when the PtNP size decreased from 2.2 to 1.3 nm, the mass activity decreased 2-fold, while the specific activity decreased 4-fold. Yet, specific activity remained similar as the PtNP increased in size from 2 to 4 nm [7], thus confirming that there is an optimal size of the PtNP for the ORR. However, the PtNP size alone is not sufficient to increase the overall efficiency of the PEMFC; the support onto which the PtNP is attached must also be optimized.

The support material onto which the PtNPs are anchored influences the overall electrocatalytic activity of the PtNPs toward the ORR [9,10,11]. Carbon-based supports are of particular interest because of their strong mechanical stability, good electrical conductivity and large specific surface area [10]. However, different carbon-based materials offer distinct advantages [12,13]. Ruiz-Camacho et al. studied PtNPs supported on Vulcan carbon, graphene oxide, reduced graphene oxide and graphene oxide/Vulcan composites and found that each support yielded different half-wave potentials (E_1/2), mass activities and specific activities for the ORR [11]. Due to the strong interaction between PtNPs and the support [10], the support onto which the PtNPs are attached can also aid in hindering the PtNP agglomeration [9,10]. Sakthivel and Drillet found that, due to the ordered micropores present on the support, the agglomeration of PtNPs supported on ordered mesoporous carbon (OMC) material was hindered as compared to PtNPs on Vulcan carbon [9]. However, some carbon supports, such as graphene, can restack, causing the overall activity of the electrocatalyst to decrease [10].

Carbon nanotubes (CNTs) have gained interest due to their high surface area and convex surface, which allow for a uniform distribution of PtNPs, and for their ability to increase the overall stability of the electrocatalyst [14,15,16,17]. CNTs have a unique structure that enhances their fluid transfer properties [16]. Kongkanand et al. investigated PtNPs supported on single-walled carbon nanotubes (SWCNTs) and found that, as compared to Pt/C, the PtNPs on SWCNTs had a higher electron-transfer rate constant [14]. However, the PtNPs on SWCNTs also had a lower electrochemically active surface area (ECSA) indicative of Pt agglomeration [14], which can result in a lower overall electrocatalytic activity. Kim et al. deposited PtNPs onto multiwalled carbon nanotubes (MWCNTs) and observed a three-fold higher intrinsic activity as compared to commercial Pt/C [18]. In PEMFCs, carbon corrosion occurs during times of fuel starvation. Lee et al. have shown that, when PtNPs are anchored onto a CNT support, the CO₂ generation was 400 ppm, whereas when PtNPs are anchored onto a Vulcan carbon support, the CO₂ generation was 4000 ppm [19], confirming the carbon corrosion resistance of CNT supports.

Functionalization of CNTs can further aid in electrochemical stability [20,21]. Li et al. linked thiol groups to CNTs and found that the PtNPs supported on thiol-functionalized CNTs were less likely to aggregate than those on hydroxyl-functionalized CNTs [20]. Additionally, the introduction of metal oxides, such as WO₃, SnO₂ and TiO₂, into the electrocatalytic system has gained interest as they inhibit carbon corrosion, aid in anchoring PtNPs and are stable in harsh environments [22,23,24,25,26,27,28,29,30,31,32]. SnO₂ is of particular interest as it is stable in acidic environments [23]. However, it does not have as high an electrical conductivity as carbon-based supports [33,34,35]. One way to overcome this is through the incorporation of a core–shell structure of the PtNPs themselves [36] and through the use of nanoalloys [37] and high-entropy alloy catalysts [38,39]. Another way to incorporate SnO₂ into the electrocatalytic system is through the integration of stable SnO₂ with the conductive and mechanically strong CNTs. This induces a strong metal–support interaction (SMSI) through both a chemical and electronic interaction [40,41]. Hoque et al. have shown that PtNPs supported on SnO₂-decorated mesoporous carbon exhibited a higher ORR onset potential than commercial Pt/C catalysts [42]. After stability testing, Hussain et al. showed that PtNPs on SnO₂-coated MWCNTs retained 88% of their original ECSA, while the PtNP/MWCNT catalyst maintained 86% of its original ECSA [21]. In PEMFC testing, Cao et al. found that PtNPs supported on W-SnO₂–carbon retained 89.2% of the initial peak power density after 50,000 cycles in a potential window of 0.60–0.95 V [43]. The addition of SnO₂ into a fuel cell catalyst could be beneficial as it does not need as humid of conditions as pure carbon supports, and thus, the overall power density of the system is increased [44], thus demonstrating the advantage that decorating SnO₂ onto carbon material provides on the stability of the electrocatalyst.

To investigate the stability advantage that the addition of SnO₂ onto the support has, this work focuses on PtNPs attached to MWCNTs coated with SnO₂ via atomic layer deposition (ALD). The synthesized catalyst here within is referred to as Pt/SnO₂–CNT. The synthesized catalyst was shown to be more stable than the commercial 20% Pt/C.

2. Results and Discussion

2.1. Physico-Chemical Characterization of Catalyst Materials

The X-ray photoelectron spectroscopy (XPS) analysis was conducted on the Pt/SnO₂–CNT catalyst to confirm its compositional makeup (Figure 1). An intense carbon peak (C 1s) is observed at 284 eV [45]. A prominent oxygen peak (O 1s) appeared at 540 eV, which originates from the carbon–oxygen groups on the CNTs and SnO₂. The presence of Sn was confirmed by the appearance of Sn 3d peaks at 485 and 495 eV [46]. The presence of Pt was confirmed with the doublet of Pt 4f peaks at 71 and 75 eV [47,48]. In this sample, the most negative Pt 4f peak is centered at 71.2 eV, a negative shift as compared to the Pt 4f peak measured by Li et al. that was 71.8 eV [26]. This negative shift is indicative of electron transfer between SnO₂ and Pt [26]. The electron transfer is evidence of an electronic metal–support interaction, which shifts the d-band center of Pt and can enhance the adsorption energy of O intermediates [26]. The concentration of Pt on the electrode surface was analyzed using XPS and found to be 0.6 at%, while X-ray fluorescence (XRF) analysis determined the Pt content to be 15 ± 3 wt%. The difference in values between XPS and XRF suggests a discrepancy between the platinum content within the bulk of the catalyst and its surface. The Pt mass determined via XRF was consistent with the energy dispersive X-ray spectroscopy (EDS) measurement, which determined the Pt content of the catalyst to be 13 ± 6%. The STEM-EDS element mapping (Figure 2b) confirms the uniform distribution of Sn on the CNTs. The scanning transmission electron microscopy (STEM) image (Figure 2a) of the PtNP/SnO₂–CNT catalyst further confirms the distribution of PtNPs throughout the sample. The Pt particle size was analyzed using ImageJ 1.54g software, and it was determined that the average particle size was 4.1 ± 1.5 nm. This is consistent with our previous studies and literature reports that used the sodium citrate method for the synthesis of PtNPs [13,49,50].

2.2. Electrochemical Characterization of Catalyst Materials

Cyclic voltammograms (CV) of the catalyst materials were measured in a potential range of 0.05–1.45 V_RHE with a scan rate of 50 mV s⁻¹ in an Ar-saturated 0.1 M HClO₄ solution. The CVs are presented in Figure 3a, which compares Pt/SnO₂–CNT with Pt/C. The presence of the hydrogen adsorption and desorption peaks is observed in the potential window of 0.05–0.35 V_RHE [51] and the Pt oxide peak at 0.75 V_RHE [51,52]. The oxide peak is less pronounced on the Pt/SnO₂–CNT catalyst than on the commercial Pt/C, which indicates the SMSI between Pt and SnO₂. Both the synthesized and commercial catalysts exhibit a double peak in the hydrogen adsorption/desorption region; these peaks appear at approximately 0.15 and 0.23 V, indicating the presence of Pt(110) and Pt(100) facets [53]. Additionally, the Pt/SnO₂–CNT catalyst exhibits a small anodic peak at approximately 0.34 V, which is attributed to Pt (111) sites [54]. Gasteiger et al. reported a shift in the Pt oxide peaks between Pt/C, Pt/carbon black and a pure Pt disk and attributed the shift to OH_ads species blocking sites for the ORR process to occur [52]. A more positive shift in the Pt oxide peak is an indicator for suppression of oxidized Pt [23]. Further, Spasov et al. showed that SnO₂ facilitates OH adsorption, therefore hindering adsorption of hydrogen onto the PtNPs [55]. The difference in the double-layer capacitance between Pt/SnO₂–CNT and Pt/C is due to the interactions that occur between the CNT_S and water [53]. This effect is enhanced with lower Pt loadings [56] and could also be due to the difference in the Pt d-band provided by electronic interactions [26]. The electrochemically active surface area (ECSA) of the PtNPs was calculated through the integration of the charge under the hydrogen desorption peak, assuming that 210 μC cm⁻² is the charge density required for the desorption of a monolayer of hydrogen from the Pt surface [57]. The calculated ECSA values for both Pt/SnO₂–CNT and Pt/C catalysts are presented in Table 1. Pt/SnO₂–CNT had an ECSA of 0.96 cm², while Pt/C had an ECSA of 4.7 cm². This is comparable to what has been found in the literature [21,52,55,58]. Sun et al. measured the ORR activity of Pt nanowires supported on tin oxide-coated CNTs in a H₂SO₄ solution; the synthesized catalyst has 50% lower ECSA as compared to the commercial Pt on carbon black [58]. Li et al. found similar results in that the total amount of PtNPs on SnO₂-decorated CNTs was lower than that on pristine CNTs [26]. The decreased loading of PtNPs could be due to the metal–support interactions occurring between SnO₂ and CNTs, causing a difference in the attachment of the PtNPs.

The electro-oxidation of pre-adsorbed CO was conducted in an Ar-saturated 0.1 M HClO₄ solution, in a potential range of 0.05–1.0 V_RHE with a scan rate of 20 mV s⁻¹. The CO-stripping profiles for Pt/SnO₂–CNT and Pt/C are provided in Figure 3b. The difference in ECSA between Pt/SnO₂–CNT and Pt/C can also be seen in the difference in height of the CO-stripping peak. Pt/SnO₂–CNT has a broader and shallower CO-stripping profile compared to that of Pt/C, which shows two peaks: a smaller sharp peak and a taller sharp peak—these differences are further indication of the SMSI occurring between the PtNPs and SnO₂. The shallower CO-stripping profile of the Pt/SnO₂–CNT catalyst could be due to a restriction of contact between the PtNPs and the CO molecules [59]. The prepared Pt/SnO₂–CNT catalyst has a CO-stripping profile that occurs at a lower potential than that of the commercial Pt/C. The shoulder formed at 0.3–0.7 V could be a result of the CO oxidation on the PtNPs adjacent to the Sn atoms as a result of electronic interaction [60,61]. This negative shift of the CO-stripping profile could be indicative of the bifunctional mechanism that occurs in response to the addition of SnO₂ [53], in that SnO₂ provides oxygen-containing species for the PtNPs in the proximity through the creation of oxygen vacancies, thus lowering the CO oxidation overpotential [60,62]. To a lesser extent, the difference of the CO-stripping profiles could also be a result of the different Pt facets on the catalyst [63].

2.3. Oxygen Reduction Reaction Study

The ORR measurements were conducted in an O₂-saturated 0.1 M HClO₄ solution in a potential range of 0.05–1.1 V_RHE and a scan rate of 10 mV s⁻¹. Figure 4a presents the rotating disk electrode (RDE) polarization curve for Pt/SnO₂–CNT at various electrode rotation rates where the diffusion-limited current density is constantly reached, thus further confirming typical Pt electrocatalytic behavior for a four-electron pathway of the ORR [64]. A slight decrease in current density is seen at the lower potentials. This could be indicative of the blocking of the ORR-active sites due to hydrogen adsorption [65].

Figure 5a shows a comparison of the Pt/SnO₂–CNT and the Pt/C catalysts at 1900 rpm. Pt/SnO₂–CNT shows a half-wave potential (E_1/2) of 0.83 V, while Pt/C possesses an E_1/2 of 0.88 V. The 50 mV difference could be a result of SnO₂ blocking some of the reaction sites [52]. SnO₂ also affects the efficiency of the oxidation of Pt [23], which could result in the lower half-wave potential of the Pt/SnO₂–CNT catalyst. These effects on Pt due to the decoration of SnO₂ onto the CNTs are a result of the influence that SnO₂ has on the redistribution of the d-band center of Pt [26]. However, this effect can be mitigated through optimization of the d-band center, as there is an optimal relation between the ORR activity and the tuning of the d-band center [26]. These results are similar to those reported by Ruiz Camacho et al. (E_1/2 = 0.83 V) for Pt supported on SnO₂-decorated carbon [66]. Hussain et al. also found that PtNPs on tin oxide photo-deposited onto CNTs had a lower half-wave potential than Pt/C [53]. This is in opposition to a study by Jia et al., which explored PtNPs on tin oxide chemically deposited onto nitrogen-doped carbon that outperformed the Pt/C catalyst [67]. These results could also be due to the amount of SnO₂ in the electrocatalytic system, as Gu et al. and Li et al. both showed that there is an optimal relationship between the amount of SnO₂ and the reactivity of Pt toward the ORR [26,38].

The Koutecký–Levich (K–L) equation (Equation (1)) [64] was used to analyze the electrochemical data of the ORR:

$${\displaystyle \frac{1}{j}}={\displaystyle \frac{1}{j_k}}+{\displaystyle \frac{1}{j_d}}=-{\displaystyle \frac{1}{nFk{C}_{O_2}^b}}-{\displaystyle \frac{1}{0.62nF{D}_{O_2}^{2/3}{v}^{-1/6}{C}_{O_2}^b{\omega }^{1/2}}}$$

(1)

where j is the measured current density, j_k is the kinetic current density, j_d is the diffusion-limited current density, n is the number of electrons transferred per O₂ molecule, k is the ORR rate constant, F is the Faraday constant (96,485 C mol⁻¹), ω is the electrode rotation rate, $C_{O_2}^b$ is the concentration of O₂ in the bulk solution (1.22 × 10⁻⁶ mol cm⁻³) [68], $D_{O_2}$ is the diffusion coefficient of O₂ (1.93 × 10⁻⁵ cm² s⁻¹) [68], and v is the kinematic viscosity of the electrolyte solution (0.01 cm² s⁻¹) [68].

Figure 4b shows the K–L plot for Pt/SnO₂–CNT, with an inset presenting the dependence of n vs. E. As the extrapolated trendline passes the origin of the K–L plot, it can be confirmed that the ORR process on the Pt/SnO₂–CNT catalyst is mass-transfer limited in a wide potential range; the K–L plot further confirms that the reaction follows a first-order kinetics toward O₂ concentration [69]. The ORR process proceeds by a four-electron pathway, as is evident from the K–L plot. The number of electrons transferred in the ORR process is slightly above four, but this can be attributed to the electrode. The four-electron pathway is also confirmed by the attainment of the diffusion-limited current density at various electrode rotation rates in Figure 4a. The ORR with a four-electron pathway favors the formation of water over the formation of H₂O₂, which is harmful for fuel cells and can lead to cell degradation and shorten the lifetime of the PEMFC [4].

The Tafel slopes for both Pt/SnO₂–CNT and Pt/C are −61 mV, as shown in Figure 5b and Table 1. This confirms that the transfer of the first electron to the O₂ molecule is the rate-determining step of the ORR [70]. In previous literature, Tafel slopes for other PtNPs supported on oxide-coated carbon materials are around −60 mV [53,71,72]. However, Zhang et al. measured a Tafel slope of −80 mV for a Pt catalyst on SnO₂-decorated carbon fibers [73]. Again though, the Tafel slope value is highly dependent on the roughness of the electrode surface [70], and some deviations are to be expected.

The specific and mass activities for the Pt/SnO₂–CNT and Pt/C catalysts were calculated at 0.9 V. The specific activity (SA) was calculated using the following equation:

$$\mathrm{S}\mathrm{A}= {\displaystyle \frac{I_k}{\mathrm{E}\mathrm{C}\mathrm{S}\mathrm{A}}}$$

(2)

where I_k is the kinetic current at 0.9 V and the ECSA is the electrochemical surface area calculated using the hydrogen adsorption area shown in Table 1. The SA values for Pt/SnO₂–CNT and Pt/C are 0.15 and 0.05 mA cm⁻², respectively. This is a 3-fold increase in the ORR activity as compared to the commercial Pt/C catalyst. The improved SA is a result of the electronic interaction occurring due to the addition of SnO₂ into the electro-catalytic system. The electronic interaction is further confirmed in Figure 3b, where the bifunctional mechanism occurs in which oxygen vacancies are created due to SnO₂ being adjacent to PtNPs [53,60,62], thus confirming that, when SnO₂ is added to CNTs via ALD to be used as a support for PtNPs, the overall ORR activity is increased.

The mass activity (MA) was calculated using the following equation:

$$\mathrm{M}\mathrm{A}= {\displaystyle \frac{I_k}{m_{Pt}}}$$

(3)

where I_k is the kinetic current at 0.9 V and m_Pt is the mass of Pt on the electrode. The mass activity of Pt/SnO₂–CNT is 44 A g⁻¹ and was calculated using the weight percentage of Pt from the XRF measurements, as is presented in Table 1. The MA of Pt/C was 61 A g⁻¹. Hussain et al. reported similar findings in that the PtNPs supported on photo-deposited SnO₂ on CNTs had a lower MA as compared to commercial Pt/C; their synthesized catalyst has an MA of 72 A g⁻¹ as compared to the commercial Pt/C, with an MA of 90 [21]. Rivera Rocabado et al. suggested that the lower activity could be due to the PtNPs being less negatively charged and reacting with nucleophilic species, thus blocking sites for which the ORR could occur [74]. PtNPs supported on tin oxide-decorated carbons studied by Jia et al. followed a different trend; their synthesized catalyst of tin oxide on nitrogen-doped carbon had a mass activity that was 3.5 times more active than Pt/C [67]. The higher specific activity of our prepared Pt/SnO₂–CNT confirms that, when SnO₂ is added to CNTs via ALD, the overall electrocatalytic activity of the PtNPs is improved through the electronic interaction.

Table 1. Pt content, ECSA and kinetic parameters for oxygen reduction of Pt/SnO₂–CNT and Pt/C catalysts in 0.1 M HClO₄ solution.

Catalyst	Pt (wt%)	ECSA (cm2)	E1/2 (V)	Tafel Slope (mV)	SA @ 0.9 V (mA cm−2)	MA @ 0.9 V (A g−1)
Pt/SnO2–CNT	15 ± 3 1	0.96 ± 0.02	0.83 ± 0.01	−61 ± 2	0.15 ± 0.03	44 ± 9
Pt/C	20	4.7 ± 0.2	0.88 ± 0.01	−61 ± 1	0.05 ± 0.01	61 ± 6

¹ from XRF analysis.

Table 1. Pt content, ECSA and kinetic parameters for oxygen reduction of Pt/SnO₂–CNT and Pt/C catalysts in 0.1 M HClO₄ solution.

Catalyst	Pt (wt%)	ECSA (cm2)	E1/2 (V)	Tafel Slope (mV)	SA @ 0.9 V (mA cm−2)	MA @ 0.9 V (A g−1)
Pt/SnO2–CNT	15 ± 3 1	0.96 ± 0.02	0.83 ± 0.01	−61 ± 2	0.15 ± 0.03	44 ± 9
Pt/C	20	4.7 ± 0.2	0.88 ± 0.01	−61 ± 1	0.05 ± 0.01	61 ± 6

¹ from XRF analysis.

2.4. Accelerated Durability Testing

The accelerated durability testing (ADT) was conducted in an Ar-saturated 0.1 M HClO₄ solution by applying 12,500 potential scans between 0.05 and 1.2 V using a scan rate of 100 mV s⁻¹. CV and RDE measurements were conducted prior to starting and following the ADT experiment and are shown in Figure 6 and Figure 7, respectively. Figure 6b shows a drastic decrease in the hydrogen adsorption/desorption and Pt oxide regions, indicating the dissolution and Oswald ripening of Pt on the catalyst; there is also a decrease in the carbon characteristic area, which could indicate carbon corrosion occurring during the stability testing, whereas this is not as present on the synthesized Pt/SnO₂–CNT catalyst. Changes also appear in the Pt/SnO₂–CNT catalyst; however, these are not as drastic, which could indicate the strong stability enhancements provided to the catalyst due to the SnO₂ coating, which induces a stronger metal–support interaction. The ECSA of both the prepared Pt/SnO₂–CNT and Pt/C catalysts was calculated before and after the durability testing and was found to have decreased. Pt/SnO₂–CNT retained 58% of its ECSA, while Pt/C only retained 31%, further emphasizing the superior stability afforded to the support as a result of the inclusion of SnO₂ onto the CNTs. This could be due to the anti-corrosion properties that SnO₂ enacts onto the CNTs—which ultimately results in lower dissolution of PtNPs [26]. Following the ADT, the addition of a peak around 0.6 V on the Pt/SnO₂–CNT catalyst was observed. This peak could be attributed to the two-electron transfer of quinone and hydroquinone [26,75,76]. Hussain et al. also concluded that a Pt–Sn alloy was present on the CVs of a photo-deposited SnO₂–CNT catalyst [53], which could indicate why the total ECSA of Pt/SnO₂–CNT did not decrease to the same extent as the Pt/C catalyst.

Figure 7 compares the RDE results before and after the ADT experiments. In Figure 7a, there is a change in the electrocatalytic behavior of the Pt/SnO₂–CNT catalyst after the ADT experiments, showing an effect of SnO₂ on the overall ORR electrocatalytic activity of the catalyst. Both catalysts exhibit similar shifts in E_1/2, with both shifting by 30 mV. This increased stability could be a result of the anchoring properties that SnO₂ provides to the PtNPs due to strong electronic interactions as a result of the Pt–O–Sn bonds, which strengthen the anchoring of the PtNPs and reduce their coalescence [23,77]. The introduction of SnO₂ onto the CNTs induces a synergetic effect by redistributing charges in the carbon 2p-states and the valence and conductive bands of SnO₂ [23], which increases the overall stability of the support. The onset potential of the synthesized Pt/SnO₂–CNT catalyst changes by 4 mV throughout the entirety of the ADT, while that of Pt/C decreases by 10 mV, showing the superior stability afforded to the synthesized catalyst through the electronic interaction with SnO₂. The addition of a metal oxide may be able to explain this, as Promsawan et al. mentioned that there is a decreased amount of carbonaceous intermediate species [78]. The carbonaceous intermediates could lead to further degradation of the carbon support. The results herein confirm that depositing SnO₂ via ALD onto the CNT surface results in strong support for PtNPs and produces an active, stable electrocatalyst.

3. Materials and Methods

3.1. Synthesis of Catalyst

Multiwalled carbon nanotubes (MWCNTs, purity > 95%, NanoLab Inc., Waltham, MA, USA) were acid-treated using the procedure reported in previous studies by Hussain et al. [21]. Briefly, MWCNTs were purified through reflux in a solution containing H₂SO₄ (95%, Fluka, Charlotte, NC, USA) and HNO₃ (65%, Sigma-Aldrich, St. Louis, MO, USA). A dimethylformamide (99.8%, Aldrich, St. Louis, MO, USA): carbon nanotube suspension (4:1 w/v) was then drop-cast onto silicon plates, followed by drying in an oven (Memmert, Schwabach, Germany) set at 115 °C so that the plate contained 0.2 mg of CNTs per cm².

Tin dioxide thin films were grown from tetrakis(dimethylamido)tin(IV) (TDMASn) and ozone (O₃) at 200 °C in a low-pressure, flow-type, hot-wall atomic layer deposition (ALD) reactor, described in more detail elsewhere [79]. The metal precursor TDMASn (99% (99.99%-Sn), Strem Chemicals, Inc., Boston, MA, USA) was vaporized from a semi-open glass boat held at 28–30 °C inside of the reactor. The oxygen precursor O₃ was produced from O₂ (99.999%, Linde, Dublin, Ireland) using a BMT Messtechnik 802 N generator. The ozone concentration, measured by a BMT Messtechnik 964 analyzer, was 210–230 g m⁻³ in our experiments. The precursors were pulsed into the reactor separately and alternately by computer-controlled solenoid valves using dry inert gas (N₂, 99.999%, Linde, Dublin, Ireland) flow as the precursor carrier, switching and reactor purging media. During depositions, the total gas pressure ranged from 210 to 240 Pa in the reaction zone. In order to optimize the deposition parameters and allow reliable self-limited (ALD-type) growth of the films, real-time quartz crystal microbalance (QCM) studies were carried out beforehand. On the basis of the QCM data obtained, ALD cycle times of 1–2–2–5 s for the metal precursor supply, first N₂ purge, the oxygen precursor supply and a second N₂ purge, respectively, appeared to be sufficient for the film growth on planar substrates. To enhance the precursor infiltration into the porous CNT-loaded samples and the subsequent removal of reaction byproducts, the ALD cycle times were increased to 10–20–20–40 s.

Then, PtNPs were chemically deposited using the sodium citrate method reported in our previous study [13]. Briefly, in the presence of 2.5 × 10⁻⁴ M sodium citrate (99.0%, Fluka), 2.5 × 10⁻⁴ M of H₂PtCl₆ (99.9%, Aldrich) was reduced by 0.01 M of NaBH₄ (99%, Sigma-Aldrich) in ice cold water. The SnO₂-coated MWCNTs were then added such that the final catalyst would contain about 20 wt% Pt. After 1.5 h of mixing and sonication, KOH (89.9%, Lack-Ner, Neratovice, Czech Republic) was added, and the suspension was allowed to rest overnight. The suspension was then washed with MilliQ water (Millipore, Burlington, MA, USA) and filtered. The synthesized catalyst is promptly referred to as Pt/SnO₂–CNT.

3.2. Preparation of Electrode

A glassy carbon (GC) electrode (GC-20SS, Tokai Carbon, Tokyo, Japan) with a 5 mm diameter was polished with aqueous slurries made from 1.0 and 0.3 μm alumina (Buehler, Uzwil, Switzerland). The polishing residues were then cleaned from the electrode surface via sonication. Following the cleaning procedure, a 1 mg mL⁻¹ ink was prepared such that, for every mg of catalyst, there was 0.7 mL of water, 0.2 mL of isopropanol (≥99.8%, Honeywell, Charlotte, NC, USA) and 0.1 mL of 5% Nafion solution (5 wt% in lower alcohols and water, Aldrich). Prior to coating the electrode, the ink was sonicated until the suspension was homogenous. The ink was then drop-casted onto the GC electrode in 5 µL aliquots under electrode rotation (500 rpm) until 20 µL in total was on the electrode such that the catalyst loading was 20 µg_Pt cm⁻². The electrode was under a constant flow of N₂ gas during the entire drop-casting process. Commercial 20% Pt/C (E-TEK, Inc., Milan, Italy) was used for the comparison.

3.3. Material Characterization

The catalyst materials were characterized via X-ray photoelectron spectroscopy (XPS) analysis using a SCIENTA SES-100 spectrometer equipped with a Mg Kα X-ray source (incident energy 1253.6 eV). Their morphology was characterized using a Helios Nanolab 600 (FEI) scanning electron microscope (SEM) with an INCA Energy 350 (Oxford Instruments, Abingdon, UK) system for energy-dispersive X-ray spectroscopy (EDS). Their morphology was further characterized by a Titan Themis 200 (FEI, Hillsboro, OR, USA) scanning transmission electron microscope (STEM) with a Super-X^TM X-ray detection system for EDS (Bruker, Billerica, MA, USA). The STEM analysis was performed at 200 kV. For the STEM analysis, the suspension of the catalyst material in isopropanol was drop-casted on holey carbon film-coated copper TEM grids (Agar Scientific Ltd., Rotherham, UK). The particle size distribution was assessed using ImageJ software. The Pt content was determined via an X-ray fluorescence (XRF) spectrometer ZSX400 (Rigaku, Tokyo, Japan).

The electrochemical characterizations were made in a conventional three-electrode cell. A counter electrode of Pt wire separated by a glass frit and a reference electrode of a reversible hydrogen electrode (RHE) were used. The electrochemical measurements were performed in a 0.1 M HClO₄ (70%, Sigma-Aldrich) solution using an Autolab PGSTAT30 potentiostat/galvanostat (Metrohm Autolab, Utrecht, The Netherlands). Cyclic voltammograms (CVs) from 0.05 to 1.45 V using a sweep rate of 50 mV s⁻¹ under Ar (99.999%, Linde) saturation were used as a pre-treatment for the electrodes. The electrodes were then cleaned using the protocol outlined by Solla-Gullon et al. [80]. Briefly, the electrode was held at 0.1 V, while CO (99.97%, Linde) was bubbled into the solution until hydrogen adsorption and desorption peaks were no longer visible on the CV curves, indicating complete coverage of PtNPs by adsorbed CO. Then, the solution was saturated for 30 min with Ar to remove the excess CO from the solution. Then, the CO-stripping profiles were recorded by scanning the potential from 0.5 to 1.0 V at a potential scan rate (v) of 20 mV s⁻¹.

The oxygen reduction reaction was investigated using the rotating disk electrode (RDE) technique with an EDI101 rotator and a CTV101 speed control unit (Radiometer). In an O₂ (99.999%, Linde)-saturated 0.1 M HClO₄ solution, the potential was scanned from 0.05 to 1.1 V using a scan rate of 10 mV s⁻¹ and the electrode rotation rates (ω) 360, 610, 960, 1900, 3100 and 4600 rpm.

The stability of the electrocatalysts was investigated via accelerated durability testing (ADT) in an Ar-saturated 0.1 M HClO₄ solution; 12,500 cyclic voltammograms were recorded by cycling the potential between 0.6 and 1.2 V with a scan rate of 100 mV s⁻¹.

4. Conclusions

In this work, acid-washed multiwalled carbon nanotubes were decorated with SnO₂ via atomic layer deposition. The SnO₂–CNTs were used as a support for PtNPs synthesized via the sodium citrate method. The prepared Pt/SnO₂–CNT catalyst underwent a thorough physico-chemical and electrochemical characterization. The addition of SnO₂ onto the CNT material allowed for a highly durable catalyst because of the electronic interactions between Pt and Sn. The prepared catalyst has a three-fold higher specific activity for the ORR as compared to commercial Pt/C. The addition of SnO₂ onto the CNT support also further increased the stability of the Pt/SnO₂–CNT catalyst, as shown during the ADT, where Pt/SnO₂–CNT retained more than 50% of the original ECSA, whereas Pt/C retained 31% of the original ECSA. It can be concluded that the addition of SnO₂ to the CNT material for PtNPs can enhance their overall durability.