Enhanced methanol oxidation activity and durability of Pt–Pd alloy nanoclusters synthesised by gas-diffusion electrocrystallisation (GDEx)

Abstract

Gas-diffusion electrocrystallisation (GDEx) provides an electrified route to synthesise Pt–Pd alloy nanostructures with tunable composition. Here, we synthesised Pt–Pd alloy nanoclusters (NCs) with defined Pt/Pd molar ratios (Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, Pt₂₅–Pd₇₅), along with monometallic Pt (Pt₁₀₀) and Pd (Pd₁₀₀) NCs. The morphology, structure, and composition of the NCs were characterised by transmission electron microscopy (TEM), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS). TEM revealed that the NCs consist of aggregates of smaller primary nanoparticles (NPs), with both NC and primary NP size increasing with Pd content. XRD and XPS confirmed alloy formation and predominantly metallic character for the Pt–Pd NCs. The electrocatalytic activity of the resulting NCs towards methanol oxidation in acid media was investigated. After anodic activation, the mass activity followed the order Pt₅₀–Pd₅₀ > Pt₁₀₀ > Pt₇₅–Pd₂₅ > commercial Pt/C > Pt₂₅–Pd_75, while the specific activity decreased in the order Pt₁₀₀ > Pt₅₀–Pd₅₀ > Pt₇₅–Pd₂₅ > Pt₂₅–Pd₇₅ > commercial Pt/C. Pd₁₀₀ was inactive under these conditions. The Pt–Pd alloy NCs also exhibited enhanced durability, with Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅ retaining 40%, 70%, and 82% of their highest activity after 4000 cycles, respectively, compared with 16% for Pt₁₀₀. The improved performance is attributed to (i) dynamic changes in the oxidation states of Pt and Pd, leading to an increased fraction of metallic Pt at the catalyst surface, as evidenced by XPS, and (ii) enhanced oxidative removal of CO-derived intermediates and CO₂ formation, as demonstrated by in situ FTIR. Overall, GDEx provides a simple and efficient strategy for synthesising Pt–Pd alloy electrocatalysts with high activity and improved stability for direct methanol fuel cells.

---

Introduction

In recent years, direct methanol fuel cells (DMFCs) have gained attention as energy conversion devices due to their inherent safety, portability, and low environmental impact.¹ DMFCs operate via the methanol oxidation reaction (MOR) at the anode and the oxygen reduction reaction (ORR) at the cathode. The high volumetric (4.82 kWh L⁻¹) and gravimetric (6.1 kWh kg⁻¹) energy densities of methanol, together with its low cost and ease of transport, make DMFCs a promising technology for efficient and portable energy conversion.^2,3 However, the widespread deployment of DMFCs is hindered by the sluggish kinetics of MOR, which results in large anodic overpotentials. These overpotentials reduce the cell voltage and, consequently, the overall energy efficiency of the system, thereby limiting DMFC performance at practical current densities.⁴

Although nonprecious metal catalysts based on earth-abundant transition metals (i.e., Ni- and Cu-based) have been explored as lower-cost alternatives for alcohol oxidation, their most promising performance is generally achieved under alkaline conditions.⁵ In acidic media, however, Pt-based catalysts remain the benchmark for the methanol oxidation reaction (MOR), with Pt being the most active single-metal catalyst owing to its unmatched ability to adsorb and activate methanol.⁶ However, pure Pt catalysts suffer from limited long-term durability due to nanoparticle aggregation and poisoning by intermediates, such as adsorbed CO (CO_ads), formed during the MOR, leading to a progressive loss of catalytic performance.⁷ In addition, the limited availability and high cost of Pt further hinder the large-scale commercialisation of DMFCs.⁶ To overcome these limitations, various strategies have been explored to ameliorate the efficiency of Pt-based catalysts, including controlling the size and shape of Pt nanostructures,^8,9 strengthening Pt-support interactions,¹⁰ and alloy engineering.¹¹ Among these approaches, alloying has attracted particular attention due to its effectiveness in mitigating poisoning effects and improving MOR activity and stability.¹²

Alloying Pt with a second metal (i.e., Pd,¹¹ Ru,¹³ Au,¹⁴ Co,¹⁵ Sn,¹⁶ Ni¹⁷) can significantly modify the electronic structure and d-band centre of Pt, thereby enhancing methanol adsorption while weakening the binding strength of poisoning intermediates.¹⁸ Moreover, when the secondary metal is oxophilic, it can facilitate water activation at lower potentials, generating oxygen-containing species (e.g., OH_ads) that assist the oxidative removal of CO_ads and regenerate Pt active sites, thus improving both activity and catalyst stability.¹⁸ Among these metals, Pd is particularly attractive for alloying with Pt due to its mild oxophilicity, combined with an identical face-centred cubic (fcc) crystal structure as Pt and a closely matched lattice parameter (lattice mismatch of 0.77%).¹⁹ This high degree of structural and chemical compatibility enables the formation of stable Pt–Pd alloys, which have been extensively reported as efficient MOR electrocatalysts in various nanostructured forms, including nanocubes,²⁰ nanowires,²¹ nanodendrites,²² nanoflowers,²³ and nanosheets.²⁴

Beyond their chemical compatibility, the use of Pd in Pt-based catalysts has also been motivated by economic considerations, since Pd has historically been less expensive than Pt. Nevertheless, Pd is also subject to pronounced price volatility, and fluctuations in precious-metal markets can significantly alter this balance, underscoring the need for alternative strategies to secure a stable and sustainable supply of platinum group metals (PGMs).²⁵ One promising approach is the recovery of Pt and Pd from end-of-life PGM-containing products, such as automotive catalytic converters, thereby reducing reliance on primary mining, alleviating supply risks associated with geopolitical constraints, and limiting the depletion of finite natural resources.⁶ In that context, our research group has developed a sustainable electrochemical process, gas-diffusion electrocrystallisation (GDEx), for the efficient and selective recovery of Pt, Pd, and other noble metals, including Rh and Au.^26–28 More recently, this approach was extended to an integrated microwave-assisted leaching-GDEx flowsheet, enabling the recovery of PGMs from spent automotive catalysts and their direct conversion into alloy nanoparticles with electrocatalytic activity toward MOR.²⁹

In the GDEx process, CO₂ and H₂O are electrochemically reduced at a gas-diffusion electrode, generating CO and H₂, which act as reducing agents for metal ions with sufficiently high reduction potentials, such as Pt and Pd, under ambient conditions. In addition, CO strongly adsorbs on the surface of noble metals, suppressing particle growth and promoting the precipitation of metal ions as metallic NPs.²⁶ Owing to its high selectivity towards noble metals, GDEx enables their efficient recovery from complex solutions with high purity.²⁸

Furthermore, beyond its application as a recovery technology, GDEx can be tailored to enable the controlled synthesis of noble-metal nanostructures suitable for electrocatalytic applications. Unlike conventional wet-chemical synthesis routes, GDEx provides an electrochemically controlled pathway for the formation of Pt–Pd nanoclusters (NCs), which may result in distinct surface and compositional characteristics. While we have previously demonstrated the feasibility of producing Pt, Pd and Pt–Pd NCs via GDEx and provided initial insights into their formation mechanisms,³⁰ the influence of alloy composition on electrocatalytic performance and stability under realistic operating conditions remains insufficiently understood. In particular, a systematic correlation between Pt/Pd ratio, surface chemical state, and methanol oxidation behaviour has not yet been established for bimetallic GDEx-derived catalysts.

In this work, we address this gap by investigating a series of Pt–Pd alloy NCs with well-defined compositions (Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, Pt₂₅–Pd₇₅), alongside the corresponding monometallic NCs (Pt₁₀₀ and Pd₁₀₀), synthesised via GDEx. The study focuses on elucidating how alloy composition and electrochemical activation influence MOR activity, stability, and reaction pathways in acidic media, using a combination of electrochemical testing and surface-sensitive spectroscopic techniques. By doing so, this work aims to clarify the structure–composition–activity relationships of GDEx-derived Pt–Pd catalysts and to further establish GDEx as a versatile platform for the synthesis of high-performance electrocatalysts.

Experimental

Materials

Hexachloroplatinic acid (H₂PtCl₆) solution (24.0–26.0 %w/w Pt content) and palladium chloride (PdCl₂, 59.6–60.0 %w/w Pd content) were supplied by Johnson Matthey. Analytical grade polyvinylpyrrolidone (PVP, M_w ≈ 55 000), sodium hydroxide (NaOH), perchloric acid (HClO₄, 60%), copper sulfate pentahydrate (CuSO₄·5H₂O), and 20 wt% Pt/C were obtained from Sigma-Aldrich. Nafion® suspension (5 wt%), sodium chloride (NaCl), and methanol (MeOH) were bought from VWR. Carbon dioxide (CO₂, 99.998%) and argon (Ar, 99.99%) were procured from Air Liquide. All the solutions were prepared with ultrapure water (18 MΩ cm) obtained using a Milli-Q® purification system.

Pt–Pd alloy NCs synthesis procedure

A previously described GDEx process^26,27,30 was employed to synthesise the Pt–Pd alloy NCs with different Pt/Pd ratios (Pt₁₀₀, Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, Pt₂₅–Pd₇₅, and Pd₁₀₀) following a reported procedure using PVP as a stabiliser.³¹ Briefly, the GDEx electrochemical reactor consists of a three-compartment EC Micro Flow® cell. A VITO CoRE® gas-diffusion electrode (GDE) was used as a cathode, while a Pt-coated Ta plate served as the anode. A leak-free Ag/AgCl reference electrode (3 M KCl) was employed. The anodic and cathodic compartments were separated using a ZIRFON® PEARL ion-permeable porous separator. CO₂ was supplied at the hydrophobic (back) side of the GDE at a flow rate of 200 mL min⁻¹ and an overpressure of 20 mbar(g). The synthesis was carried out under chronopotentiometric (CP) conditions at −30 mA cm⁻². The catholyte consisted of the Pt- and Pd-containing solutions prepared in 0.5 M NaCl and 1 g L⁻¹ of PVP, while 0.5 M NaCl was utilised as the anolyte.

After the synthesis, the Pt–Pd alloy NCs suspensions were treated with NaOH to remove PVP from the NC surface following a previously reported procedure.³² NaOH pellets were added until the pH reached approximately 14, and the suspensions were allowed to settle naturally. The Pt–Pd alloy NCs were then collected by centrifugation at 15 000 rpm, and the supernatant was discarded. The particles were washed repeatedly with ultrapure water until the pH was neutral (∼7). Subsequently, one portion of the product was dried at 60 °C for 12 h for material characterisation, while the remaining portion was resuspended in ultrapure water for electrocatalytic testing. The Pt and Pd concentrations in the PVP-free Pt–Pd alloy NCs suspensions were measured with an inductively coupled plasma-optical emission spectrometer (ICP-OES, Varian 750 ES) after digestion with aqua regia. All Pt–Pd alloy NCs suspensions and powders were stored in closed vessels and kept in the dark until further use.

Material characterisation

Transmission electron microscopy (TEM) images were acquired using an EFTEM JEOL 2200 FS microscope operated at an acceleration voltage of 200 kV. Nanoparticle dispersions were drop-cast onto carbon-coated copper grids prior to imaging. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) elemental maps and energy-dispersive X-ray spectroscopy (EDS) spectra were obtained using the same microscope equipped with a silicon drift detector (SDD) X-MaxN 80 TS (Oxford Instruments).

Scanning electron microscopy (SEM) images were obtained using a Philips XL30 FEG scanning electron microscope with secondary electrons and an acceleration voltage of 30 kV. The samples were prepared by placing two drops of the NCs suspensions on an aluminium foil mounted on a sample holder.

Powder X-ray diffraction (XRD) measurements were performed in a PANanalytical X'Pert Pro diffractometer with Cu Kα radiation (λ = 1.5406 Å). Dried samples were gently ground using a mortar and pestle and mounted on standard silicon single-crystal sample holders. Data were collected using a spinning stage at 40 kV and 40 mA, with a step size of 0.04° (2θ), a counting time of 4 s per step, and a scan range of 20–120° (2θ). Profile fitting of the diffraction patterns was carried out using Highscore Plus (Malvern PANanalytical) with reference data from the Inorganic Crystal Structure Database (ICSD). Crystallite sizes were calculated using the Scherrer eqn (1):

τ = κλ/(β cos θ) (1)

where τ is the crystallite size, κ is the shape factor, β is the line broadening (full width at half maximum), λ is the X-ray wavelength, and θ is the Bragg angle.

X-ray photoelectron spectroscopy (XPS) analyses were conducted using a PHI TFA-XPS spectrometer based on the PHI 5601 platform, equipped with a monochromatic Al Kα X-ray source and a multichannel hemispherical electron analyser. The analysed area had a diameter of 0.4 mm, and spectra were recorded at a photoelectron emission angle of 45° without charge neutralisation. Powder samples were mounted on conductive carbon tape. Survey spectra were collected at a pass energy of 187 eV, while high-resolution spectra were acquired at pass energies of 29 or 58 eV. For each sample, two to three different regions were analysed to assess sample homogeneity.

Electrochemical characterisation

Electrocatalytic activity of the Pt–Pd alloy NCs was evaluated at room temperature using a conventional three-electrode configuration connected to a Bio-Logic (VMP3) multichannel potentiostat. All measurements were repeated at least three times to ensure reproducibility. A Pt coil (∼3.6 cm²) was used as the counter electrode, an Ag/AgCl electrode (3 M KCl) served as the reference electrode, and a glassy carbon electrode (GCE, 3.0 mm diameter) acted as the working electrode. Prior to catalyst deposition, the GCE was polished with alumina slurry (0.05 µm) and thoroughly rinsed with ultrapure water. The PVP-free Pt–Pd alloy NCs suspensions were sonicated for 30 min, after which 10–20 µL of the suspension was drop-cast onto the GCE and dried at room temperature. Subsequently, 5 µL of a Nafion suspension (0.5 wt%) was applied as a binder and allowed to dry for 1 h to ensure catalyst adhesion. The Pt loading was controlled at approximately 50 µg cm⁻².

The electrochemically active surface area (ECSA) was determined using the copper underpotential electrodeposition (CuUPD) method.³³ Initially, cyclic voltammetry (CV) was performed in Ar-saturated 0.1 M HClO₄ at a scan rate of 50 mV s⁻¹ between 0.05 and 1.1 V vs. RHE until a stable voltammogram was obtained (∼30 cycles), followed by a CV recorded at 10 mV s⁻¹. The electrodes were then transferred to an Ar-saturated 0.1 M HClO₄ solution containing 5 mM CuSO₄. The CuUPD was carried out by holding the electrode at 0.350 V vs. RHE for 100 s, followed by a linear sweep voltammetry (LSV) from 0.350 to 1.1 V vs. RHE at 10 mV s⁻¹. The charge associated with Cu stripping was determined after subtracting the background charge measured under identical conditions in Cu-free 0.1 M HClO₄. A specific charge of 0.42 mC cm⁻² was used to calculate the ECSA according to eqn (2), where Q_Cu is the Cu stripping charge, and L is the catalyst loading.

ECSA = Q_Cu/(0.42L) (2)

MOR measurements were performed after cleaning the electrode surface, by cycling between 0.05 and 1.1 V vs. RHE for 30 cycles at 50 mV s⁻¹ in Ar-saturated 0.1 M HClO₄. The electrolyte was then replaced with Ar-saturated 0.1 M HClO₄ containing 1.0 M MeOH. Prior to MOR measurements, the catalysts were anodically activated by holding the electrode at 1.0 V vs. RHE for 30 min. CVs were subsequently recorded between 0.2 to 1.2 V vs. RHE at a scan rate of 50 mV s⁻¹. Chronoamperometry (CA) experiments were conducted to assess short-term stability by monitoring the current for 3600 s while holding the potential at the peak potential determined from CV. Accelerated durability tests (ADT) were performed by continuous CV cycling for 4000 cycles. For the ADT measurements, the NCs were deposited onto Toray carbon paper with an exposed surface area of 1.5 cm². The Toray carbon paper was used as the electrode substrate/current collector, rather than a catalyst support. It was selected to facilitate bubble release during prolonged cycling and to enable post-electrochemical XPS analysis of the catalyst layer.

In all electrochemical measurements, ohmic losses were compensated in EC-Lab using the solution resistance determined by the current-interrupt method (85% iR compensation). Potentials measured versus the Ag/AgCl reference electrode were converted to the reversible hydrogen electrode (RHE) using eqn (3).

E_RHE = E_Ag/AgCl + 0.059 pH + E⁰_Ag/AgCl (3)

where E⁰_Ag/AgCl is the potential of the Ag/AgCl (KCl 3 M) electrode vs. RHE (0.210 V).

Electrochemical in situ FTIR reflection spectroscopy study

In situ Fourier transform infrared (FTIR) reflection spectroscopic measurements were performed using a Nicolet iS50 FT-IR Advanced Spectrometer with a liquid nitrogen-cooled MCT-A detector. N₂ gas was continuously purged through the spectrometer to minimise atmospheric contributions from water vapour and CO₂. Atmospheric suppression and attenuated total reflection (ATR) correction were applied using the OMNIC 9 software. A Polaris mid-infrared source and a KBr beamsplitter were employed. The spectrometer was fitted with a VeeMAX III specular reflection accessory using a 60° Si ATR prism.

A tailor-made glass electrolysis cell was mounted on top of the VeeMAX III accessory. The electrochemical potential was controlled using a PalmSens4 potentiostat. The working electrode consisted of a thin layer of Pt–Pd alloy NCs drop-cast onto the totally reflecting plane of the Si prism. Electrical contact to the catalyst layer was ensured by pressing three gold-plated pins onto the film. A leak-free Ag/AgCl (3 M KCl) and a Pt coil were used as reference and counter electrodes. The electrolyte consisted of a 0.1 M HClO₄ solution containing 1.0 M MeOH.

In situ FTIR spectra were collected using a multistep FTIR spectroscopy (MS-FTIR) procedure.³⁴ The reference spectrum, R(E_r), was recorded at 0.03 V vs. RHE, while sample spectra, R(E_s), were acquired following successive 0.05 V steps from 0.05 V to 1.20 V vs. RHE in the anodic direction. For each spectrum, 32 interferograms were co-added, at a spectral resolution of 4 cm⁻¹. The spectra are presented as relative changes in reflectivity, calculated according to eqn (4):

where R(E_w) is the single-beam reflection spectrum recorded with the IR window alone, used to subtract the IR reflection contribution of the Si prism.³⁵

Results and discussion

Synthesis and characterisation of the Pt–Pd alloy nanoclusters

Pt–Pd alloy NCs, together with the corresponding monometallic NCs, were synthesised using the GDEx process. The synthesis employed NaCl as the supporting electrolyte and PVP as a stabiliser. The formation of NCs is intrinsic to the GDEx mechanism, as schematically illustrated in Fig. 1. In this process, CO and H₂, which act as reducing agents for Pt(IV) and Pd(II) species in solution, are generated in situ via the simultaneous electrochemical reduction of CO₂ and H₂O at the gas diffusion electrode (eqn (5) and (6)).²⁶

CO₂ + 2H₂O + 2e⁻ ⇌ CO + 2OH⁻ E⁰ = −0.934 V_SHE (5)

2H₂O + 2e⁻ ⇌ H₂ + 2OH⁻ E⁰ = −0.828 V_SHE (6)

Fig. 1 Schematic illustration of Pt–Pd alloy NCs formation via GDEx. Dissolved Pt and Pd ions are reduced to yield primary NPs, concurrent galvanic displacement between Pt and Pd may also occur. The presence of CO is critical for suppressing particle growth and obtaining small primary NPs. These primary NPs subsequently agglomerate into larger NCs, which are stabilised by the presence of PVP.

H₂ drives the reduction of Pt(IV) and Pd(II) species (eqn (7)–(9)), while CO additionally contributes to the reduction of Pd(II) under the synthesis conditions (eqn (10)). The reduction of Pt(IV) proceeds via a two-step pathway involving Pt(II) as an intermediate, whereas Pd(II) is reduced directly to metallic Pd in a single step. As a consequence, Pd(II) reduction occurs more rapidly than Pt(IV) reduction. When both metal precursors are present in solution, the initial formation of Pd⁰ catalyses the subsequent reduction of Pt(II) through a galvanic displacement reaction (eqn (11)), leading to the simultaneous depletion of Pt and Pd species and the formation of Pt–Pd alloy NCs.³¹

[PtCl₆]²⁻ + H₂ → [PtCl₄]²⁻ + 2H⁺ + 2Cl⁻ (7)

[PtCl₄]²⁻ + H₂ → Pt⁰ + 2H⁺ + 4Cl⁻ (8)

[PdCl₄]²⁻ + H₂ → Pd⁰ + 2H⁺ + 4Cl⁻ (9)

[PdCl₄]²⁻ + CO + H₂O → Pd⁰ + CO₂ + 2H⁺ + 4Cl⁻ (10)

Pd⁰ + [PtCl₄]²⁻ → Pt⁰ + [PdCl₄]²⁻ (11)

In addition to its role in reducing, CO plays a crucial role in controlling particle growth. Strong chemisorption of CO on the nanoparticle surface suppresses growth kinetics, favouring the formation of small, monodisperse primary nanoparticles (NPs).

In parallel with these reduction and growth control processes, unreacted CO₂ also helps regulate the local reaction environment. Since both CO₂ and H₂O reduction generate OH⁻, the catholyte becomes locally alkaline during polarisation. However, dissolved CO₂ can consume OH⁻ by forming bicarbonate and carbonate species, thereby buffering the increase in pH. This CO₂/HCO₃⁻/CO₃²⁻ helps to prevent the non-selective precipitation of dissolved metal ions, particularly Pd species, as hydroxides. Under GDEx operating conditions, the high ionic strength and the buffered alkaline environment promote the coalescence of these primary NPs into larger NCs. The presence of PVP in the reaction media stabilises the resulting NCs and prevents diffusion-limited aggregation into larger agglomerates.³⁰

The TEM images (Fig. 2) show similar overall morphologies for all materials, consisting of NCs formed by agglomerates of smaller primary NPs, in agreement with previous observations.³⁰ Considering first the primary NPs, higher-magnification TEM images (Fig. S1) indicate a progressive increase in primary particle size with increasing Pd molar ratio, from approximately 2–5 nm for pure Pt (Pt₁₀₀) to 5–10 nm for pure Pd (Pd₁₀₀), with intermediate sizes observed for the Pt–Pd alloy NCs. Owing to the aggregation of the primary NPs, an accurate statistical size distribution could not be obtained from TEM.

Fig. 2 TEM micrographs of the GDEx-made NCs. (a) Pt₁₀₀, (b) Pt₇₅–Pd₂₅, (c) Pt₅₀–Pd₅₀, (d) Pt₂₅–Pd₇₅, and (e) Pd₁₀₀. Insets show representative HRTEM images highlighting lattice fringes.

At the NC level, the TEM images (Fig. 2) qualitatively suggest that the aggregates become larger and more compact as the Pd content increases. This tendency is most evident in the Pd-rich samples, where selected micrographs reveal relatively large, dense clusters. However, because TEM probes only local regions of these heterogeneous materials, individual micrographs may overrepresent either larger or smaller aggregates; therefore, a reliable statistical NC size distribution cannot be derived from these images alone. In this context, previous characterisation of the same materials is informative, revealing cluster size distributions of 20 ± 4 nm for Pt₁₀₀, 20 ± 6 nm for Pt₇₅–Pd₂₅, 22 ± 9 nm for Pt₅₀–Pd₅₀, 24 ± 12 nm for Pt₂₅–Pd₇₅, and 26 ± 16 nm for Pd₁₀₀, together with broader size distributions for the Pd-rich compositions.³¹

This behaviour can be rationalised by the combined roles of CO and PVP during the synthesis. CO strongly chemisorbs on Pt and Pd surfaces and acts as the main capping agent, thereby suppressing particle growth.³⁶ In the case of Pd, however, CO also participates directly in the reduction of Pd(II),³⁷ allowing continued growth of the primary NPs and resulting in larger particle sizes. Under these conditions, PVP primarily acts as a colloidal stabiliser, limiting diffusion-driven aggregation of the NCs rather than governing primary NP growth.³⁸ As a result, the PVP concentration used here appears sufficient to stabilise Pt-rich NCs but becomes progressively less effective for Pd-rich compositions, where larger primary NPs and NCs are formed. This interpretation is consistent with previous reports showing that Pd nanocrystal size strongly depends on the PVP monomer-to-Pd ratio.³⁹ Consequently, insufficient stabilisation at higher Pd contents leads to the formation of larger clusters, as also evidenced by SEM images of Pd₁₀₀ (Fig. S2).

High-resolution TEM images (insets in Fig. 2 and S3) reveal well-defined and continuous lattice fringes across the primary NPs, indicating a high degree of crystallinity. The measured lattice spacings range from 0.227 nm to 0.231 nm, consistent with the (111) lattice planes of fcc Pt, Pd, and Pt–Pd alloys.²⁴ The corresponding selected area electron diffraction (SAED) patterns (Fig. S4) display concentric diffraction rings, confirming the polycrystalline nature of the NCs and the presence of fcc crystal structure across all compositions. Furthermore, energy-dispersive X-ray spectroscopy (EDS) mapping of Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅ (Fig. S5) shows a homogeneous spatial distribution of Pt and Pd within the NCs, supporting the formation of alloyed nanostructures rather than phase-segregated domains.

XRD was used to investigate the crystal structure of the synthesised materials. As shown in Fig. 3, all samples exhibit five broadened diffraction peaks, which can be indexed to the (111), (200), (220), (311), and (222) planes of an fcc crystal structure. The measured peak positions, calculated lattice parameters, and crystallite size are summarised in Table S1. The pronounced peak broadening is consistent with the small crystallite size of the primary NPs, in good agreement with the dimensions observed by TEM.

Fig. 3 XRD diffractograms of the GDEx-made NCs.

A lattice contraction of −0.33% is observed for Pt₁₀₀ relative to bulk Pt (3.9242 Å), which is commonly reported for Pt nanomaterials of comparable size and is generally attributed to size-induced lattice strain arising from a high surface-to-volume ratio and the presence of surface defects. In metallic NPs, undercoordinated surface atoms and internal microstrain can reduce the average lattice parameter relative to the bulk.⁴⁰

In contrast, Pd₁₀₀ exhibits a lattice expansion of +0.33%, which can be assigned to the well-known ability of Pd to absorb hydrogen and form Pd–H solid solutions or hydride-like phases under reductive conditions. Hydrogen incorporation into octahedral interstitial sites expands the Pd lattice and counteracts the size-induced lattice contraction typically expected for small Pd nanoparticles.⁴¹

For the Pt–Pd alloy NCs (Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅), the calculated parameters fall between those of bulk Pt (3.9242 Å) and bulk Pd (3.8907 Å), with a progressive lattice contraction as the Pd molar fraction increases (Fig. S6). Such behaviour is consistent with alloy formation and reflects the combined effects of atomic size mismatch and Pt–Pd metal–metal interactions within the fcc lattice. Importantly, the lattice parameters of the Pt–Pd alloy NCs do not vary linearly with Pd content. This deviation from a simple Vegard-type behaviour indicates that the lattice response is governed by competing effects, including Pt–Pd alloying interactions, size- and strain-induced lattice contraction, and the reduced propensity for hydrogen incorporation in mixed Pt–Pd lattices compared to pure Pd. As a result, hydrogen-induced lattice expansion is progressively suppressed upon alloying with Pt, explaining the non-linear dependence of the lattice parameter on Pd molar ratio. It should be noted that XRD provides an average structural description and does not resolve short-range chemical ordering within the alloy lattice.

XPS was used to examine the surface chemical composition and oxidation states of the NCs. The Pt 4f and Pd 3d spectra (Fig. 4) confirm the presence of both elements at the surface of the Pt–Pd NCs (Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅). The characteristic Pt 4f_7/2 and Pt 4f_5/2 components are observed at binding energies of ∼71.3 and ∼74.5 eV, respectively, while the Pd 3d_5/2 and Pd 3d_3/2 components are located at ∼335.3 and ∼340.4 eV. The Pt 4f envelope was deconvoluted into contributions assigned to Pt⁰, Pt²⁺, and Pt⁴⁺, and the Pd 3d envelope into Pd⁰, Pd²⁺, and Pd⁴⁺ species (Table S2).

Fig. 4 Pt 4f and Pd 3d XPS spectra of the GDEx-prepared NCs.

Quantitative analysis indicates that the NCs surfaces contain a mixture of metallic and oxidised species, with the metallic fraction remaining substantial but decreasing for Pd-rich compositions. This trend is consistent with the higher susceptibility of Pd-containing surfaces to oxidation and hydroxylation upon exposure to air. For Pt₇₅–Pd₂₅, the Pd signal intensity was comparatively low due to the pronounced Pt surface enrichment (Table S3), which limits reliable deconvolution of minor oxidised Pd contributions and results in an apparently fully metallic component in the XPS fitting. Importantly, the detection of oxidised surface species by XPS does not contradict the metallic fcc reflections observed by XRD, as XPS probes only the near-surface region (a few nanometres), whereas XRD predominantly reflects the bulk crystalline structure and is comparatively insensitive to thin and/or amorphous surface oxide layers.⁴²

Notably, the Pt⁰ 4f_7/2 component of the Pt–Pd alloy NCs is shifted ∼0.2–0.3 eV to lower binding energy relative to Pt₁₀₀, whereas the Pd⁰ 3d_5/2 component is shifted by ∼0.1–0.2 eV to higher binding energy relative to Pd₁₀₀. Such opposite shifts are characteristic of electronic interactions in Pt–Pd alloys and have been reported for related Pt–Pd nanostructures.^19,21 The binding energy shifts suggest electron density transfer toward Pt, consistent with the electronegativity difference between Pt (2.28) and Pd (2.20), supporting alloy formation and in agreement with the STEM-EDS observations.^19,21 Besides, the downshift of the Pt binding energy, commonly associated with a downshift of the Pt d-band centre, is known to weaken the adsorption strength of poisoning intermediates (e.g., CO_ads) during MOR.¹⁹

Table S3 compares the bulk Pt/Pd composition (ICP-OES) with the surface composition obtained by XPS. For all alloy NCs, the surface is enriched in Pt relative to the bulk. This Pt surface enrichment may arise from the coupled reaction pathway during GDEx, where rapid Pd nucleation followed by Pt deposition (including galvanic displacement on Pd⁰) promotes Pt accumulation at the particle surface.

Electrocatalytic performance of the Pt–Pd alloy nanoclusters for the MOR

The electrocatalytic activity of Pt₁₀₀ and the Pt–Pd alloy NCs (Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, Pt₂₅–Pd₇₅) toward MOR was studied in acidic media (0.1 M HClO₄ + 1.0 M MeOH). The GDEx-synthesised NCs were evaluated as unsupported catalysts; this approach was selected to assess the electrocatalytic behaviour of the as-synthesised GDEx-products and to avoid additional effects arising from support properties or nanoparticle-support interactions. Pd₁₀₀ was not included, as Pd alone is inactive for MOR under acidic conditions (Fig. S7). For comparison, a commercial 20 wt% Pt/C catalyst was used as a benchmark.

Fig. 5a shows cyclic voltammograms (CVs) recorded in Ar-saturated 0.1 M HClO₄ at a scan rate of 50 mV s⁻¹. All currents are normalised to the Pt mass. All materials display the characteristic features of Pt-based electrodes, including a hydrogen adsorption/desorption region between 0.03 and 0.35 V vs. RHE, a double-layer region from 0.35 to 0.70 V vs. RHE, and hydroxyl adsorption/desorption at potentials above 0.7 V vs. RHE. The electrochemically active surface area (ECSA) values are summarised in Table 1 (normalised to Pt loading) and Table S4 (normalised to total metal loading).

Fig. 5 (a) CVs of GDEx-formed NCs in 0.1 M HClO₄ at 50 mV s⁻¹. (b) Mass activity and (c) specific activity towards MOR in 0.1 M HClO₄ + 1.0 M MeOH at 50 mV s⁻¹. Arrows indicate the scan direction. (d) Chronoamperometric responses in 0.1 M HClO₄ + 1.0 M MeOH recorded at the peak potential identified from the CVs. (e) ADT results, normalised to the highest mass activity value.

Table 1 Characteristic electrocatalytic parameters for methanol oxidation normalised to Pt loading^a

Material	ECSACuUPD (m^2 gPt^−1)	Mass activity (mA mgPt^−1)	Specific activity (mA cm^−2)
a The values reported are the average and standard deviation of at least three measurements.
Pt100	32.8 ± 1.2	666 ± 15	2.03 ± 0.05
Pt75–Pd25	36.1 ± 2.7	520 ± 9	1.44 ± 0.03
Pt50–Pd50	42.5 ± 1.7	721 ± 14	1.70 ± 0.03
Pt25–Pd75	56.1 ± 2.9	486 ± 31	0.87 ± 0.06
20% Pt/C	73.3 ± 3.9	503 ± 18	0.70 ± 0.02

---

The ECSA of the Pt–Pd alloy NCs was determined using the copper underpotential deposition (CuUPD) method. This approach was selected based on previous studies showing that CuUPD reliably estimates the electrochemically active surface area of Pt- and Pd-based nanoparticles and alloy electrocatalysts. In Pt–Pd systems, hydrogen underpotential deposition (HUPD) can be less suitable because Pd not only adsorbs hydrogen on its surface, similar to Pt, but also absorbs hydrogen into its bulk lattice, which may lead to an overestimation of the ECSA when hydrogen-based methods are used.⁴³ CO stripping is also widely employed for ECSA determination and is particularly relevant for probing CO adsorption on Pt-group metals; however, its quantitative interpretation can be affected by variations in CO coverage, adsorption configuration, surface orientation, and surface restructuring induced by strongly adsorbed CO. Comparative studies on Pt- and Pd-based nanocatalysts have shown that CuUPD provides a more accurate estimate of the electrochemically active surface area than HUPD or CO stripping, particularly for alloy nanoparticles where the measured stripping charges can be influenced by composition and surface structure.^33,44,45 Therefore, CuUPD was used here as the primary method to determine the ECSA of the Pt–Pd alloy NCs.

For bimetallic Pt–Pd catalysts, ECSA and electrocatalytic activity can be normalised either to Pt loading or to the total noble-metal loading. Pt-normalised values are useful because Pt is the primary active metal for MOR in acidic media, whereas total-metal-normalised values provide a more direct comparison of the overall utilisation of the noble-metal content in the catalyst. Therefore, both normalisations are reported here to avoid ambiguity in the interpretation of composition-dependent trends.

Integration of CuUPD stripping charges (Fig. S8) yields ECSA values of 32.8 ± 1.2 m² g_Pt⁻¹ for Pt₁₀₀, 36.1 ± 2.7 m² g_Pt⁻¹ for Pt₇₅–Pd₂₅, 42.5 ± 1.7 m² g_Pt⁻¹ for Pt₅₀–Pd₅₀, and 56.1 ± 2.9 m² g_Pt⁻¹ for Pt₂₅–Pd₇₅. Thus, when normalised to Pt content, the ECSA increases as the Pt molar fraction decreases. This effect is consistent with preferential Pt surface enrichment observed by XPS for the alloy NCs. Nevertheless, this trend should be interpreted as an apparent increase in Pt utilisation rather than as a direct increase in the total accessible metal surface area, since the CuUPD-associated stripping charge is normalised by a progressively smaller Pt mass as the Pt fraction decreases. In contrast, when the ECSA is normalised to the total metal loading (Pt + Pd), the opposite trend is observed (Table S4), with the ECSA decreasing from 32.8 ± 1.2 m² g_Pt+Pd⁻¹ for Pt₁₀₀, 30.5 ± 2.3 m² g_Pt+Pd⁻¹ for Pt₇₅–Pd₂₅, 28.9 ± 1.2 m² g_Pt+Pd⁻¹ for Pt₅₀–Pd₅₀, and 26.1 ± 1.1 m² g_Pt+Pd⁻¹ for Pt₂₅–Pd₇₅. This decrease is consistent with the larger and more compact nanocluster morphology observed at higher Pd contents, which reduces the proportion of electrochemically accessible metal surface and increases the fraction of electrochemically inactive bulk metal.

It is worth noting that the ECSA values of the GDEx-synthesised NCs are significantly lower than those of commercial Pt/C (73.3 ± 3.9 m² g_Pt⁻¹) determined using the HUPD method. For pure Pt systems, HUPD and CuUPD typically yield comparable ECSA values, and the difference does not affect the qualitative comparison presented here. Nevertheless, the substantially higher ECSA of Pt/C is primarily attributed to the high dispersion of Pt NPs on a carbon support, which promotes uniform particle distribution and maximises metal accessibility. In contrast, the GDEx-synthesised NCs investigated in this study are unsupported, and partial aggregation limits the fraction of electrochemically accessible metal surface. Despite this, unsupported catalysts offer important advantages, including reduced electrical resistance compared to carbon-supported materials and enhanced electrochemical stability by eliminating carbon corrosion, which can otherwise lead to catalyst degradation, aggregation, and metal dissolution.^46,47

The MOR performance of Pt₁₀₀, the Pt–Pd alloy NCs, and Pt/C was further evaluated by CV in Ar-saturated 0.1 M HClO₄ + 1.0 M MeOH at 50 mV s⁻¹. HClO₄ was selected as the supporting electrolyte because perchlorate is a weakly adsorbing anion on Pt-based surfaces, thereby minimising specific anion adsorption effects and enabling a clearer comparison of the intrinsic composition-dependent MOR activity of the catalysts.⁴³ While the present work focuses on elucidating the influence of Pt/Pd alloy composition on MOR activity and reaction intermediates under a fixed electrolyte composition, a comprehensive kinetic analysis involving systematic variation of methanol and supporting-electrolyte concentrations falls outside the scope of this study. Such an analysis could nevertheless provide additional insight into mass-transport and charge-transfer effects and will be considered in future work.

Since the catalytic activity of Pt-based alloys can be enhanced by anodic treatment,⁴⁸ the electrodes were anodically activated prior to MOR measurements by holding the potential constant at 1.0 V vs. RHE for 30 min. Such treatment is known to induce surface restructuring, remove weakly bound surface species (e.g., stabilisers or adventitious carbon), and modify the surface electronic and geometric structure through oxidation–reduction processes, thereby increasing the density of catalytically active Pt sites and improving reaction kinetics.⁴⁸

The resulting currents were normalised to Pt loading and ECSA to determine the mass activity (MA) (Fig. 5b) and specific activity (SA) (Fig. 5c). The CVs display the two characteristic methanol oxidation features: a forward-scan peak (i_f), associated with methanol oxidation on predominantly metallic Pt sites, and a backwards-scan peak (i_b). In line with recent mechanistic studies, the backward oxidation peak is not attributed to the oxidation of residual carbonaceous intermediates formed during the forward scan, but rather methanol oxidation occurring on oxygen-modified Pt surfaces (Pt–Ox) generated at high anodic potentials.^49,50

The forward peak potential shifts progressively to lower values with increasing Pd content from 0.997 V vs. RHE for Pt/C and 0.986 V vs. RHE for Pt₁₀₀, to 0.955 V vs. RHE for Pt₇₅–Pd₂₅, 0.939 V vs. RHE for Pt₅₀–Pd₅₀, and 0.912 V vs. RHE for Pt₂₅–Pd₇₅. This trend indicates that methanol oxidation kinetics are facilitated upon Pd incorporation, consistent with electronic and bifunctional effects in Pt–Pd alloys.

The MA and SA values are summarised in Table 1 (normalised to Pt content) and Table S4 (normalised to the total metal content). The MA values are 666 ± 15 mA mg_Pt⁻¹ for Pt₁₀₀, 520 ± 9 mA mg_Pt⁻¹ for Pt₇₅–Pd₂₅, 721 ± 14 mA mg_Pt⁻¹ for Pt₅₀–Pd₅₀, 486 ± 31 mA mg_Pt⁻¹ for Pt₂₅–Pd₇₅, and 503 ± 18 mA mg_Pt⁻¹ for Pt/C. The corresponding SA values are 2.03 ± 0.05 mA cm⁻² for Pt₁₀₀, 1.44 ± 0.03 mA cm⁻² for Pt₇₅–Pd₂₅, 1.70 ± 0.03 mA cm⁻² for Pt₅₀–Pd₅₀, 0.87 ± 0.06 mA cm⁻² for Pt₂₅–Pd₇₅, and 0.70 ± 0.02 mA cm⁻² for Pt/C. Overall, Pt₁₀₀ and the Pt–Pd alloy NCs outperform commercial Pt/C, with Pt₅₀–Pd₅₀ exhibiting the highest MA and Pt₁₀₀ the highest SA among the investigated catalysts.

The difference between the MA and SA trends arises from their different normalisation basis. MA reflects the current generated per mass of Pt, whereas SA reflects the current generated per electrochemically accessible surface area. Thus, Pt₅₀–Pd₅₀ exhibits the highest MA because it delivers the highest current per Pt loading. However, Pt₅₀–Pd₅₀ also has a higher Pt-normalised ECSA than Pt₁₀₀, 42.5 ± 1.7 m² g_Pt⁻¹ compared with 32.8 ± 1.2 m² g_Pt⁻¹, meaning that its current is normalised over a larger accessible surface area. As a result, its SA is lower than that of Pt₁₀₀. Therefore, Pt₅₀–Pd₅₀ provides the highest Pt-mass-normalised activity, whereas Pt₁₀₀ displays the highest surface-area-normalised activity.

When the activity is instead normalised to the total noble-metal loading (Table S4), the MA values of the Pt–Pd alloy NCs are lower than the Pt-normalised values because the Pd mass is also included in the denominator. This distinction is important because Pt is the primary metal responsible for methanol adsorption and dehydrogenation in acidic media, whereas Pd primarily contributes via electronic and bifunctional effects that enhance reaction kinetics and tolerance to poisoning. Therefore, Pt-normalised MA highlights Pt utilisation, while total-metal-normalised MA provides a more conservative measure of the overall utilisation of the noble-metal content. In contrast, the SA values are unaffected by whether the activity is first expressed per mass of Pt or per mass of total noble metal. This is because SA is calculated as the ratio of MA to ECSA; therefore, the loading term cancels out when both quantities are expressed consistently, either per mass of Pt or per mass of total noble metal.

These results show that Pt₅₀–Pd₅₀ offers the most favourable Pt-mass-normalised performance, whereas Pt₁₀₀ exhibits the highest surface-area-normalised activity. However, the lower MA values obtained for the Pt–Pd alloy NCs, after normalisation to the total Pt + Pd loading, indicate that further optimisation is needed to improve utilisation of the overall noble-metal content while preserving the beneficial alloying effects of Pd.

Notably, the anodic activation step has a markedly stronger effect on the Pt–Pd alloy NCs than on Pt₁₀₀, resulting in activity increases of 18% for Pt₇₅–Pd₂₅, 47% for Pt₅₀–Pd₅₀, and 38% for Pt₂₅–Pd₇₅ compared with only a 3% increase for Pt₁₀₀ (Fig. S9 and Table S5). This enhanced activation effect is attributed to the greater sensitivity of Pt–Pd surfaces to oxidative treatment, which promotes favourable surface composition changes, modifies Pt–Pd electronic interactions, and facilitates the formation of oxygen-containing species on Pd that assist in the oxidative removal of poisoning intermediates.⁴⁸ Table S6 compares the MOR activity of the present Pt–Pd alloy NCs with previously reported Pt–Pd nanomaterials, showing that the GDEx-derived catalysts perform comparably to, or better than, many unsupported and some supported systems.

To assess short-term stability, chronoamperometric curves (Fig. 5d) were recorded at the i_f peak potential of each material for 3600 s. All catalysts exhibit a gradual decrease in current over time. This decay is commonly associated with the accumulation of strongly adsorbed MOR intermediates (e.g., CO-like species) and/or gradual changes in the electrochemically accessible area during operation, including surface oxidation/reconstruction, partial dissolution, or particle detachment/aggregation. When the currents are normalised to their initial values (Fig. S10), the decay is more pronounced for Pt-only catalysts (Pt₁₀₀ and Pt/C) than in the Pt–Pd alloy NCs, indicating lower susceptibility of the Pt–Pd surfaces to deactivation under these conditions, consistent with reduced poisoning by strongly adsorbed intermediates and enhanced intermediate removal.

Accelerated durability tests (ADT) were further performed to compare the long-term stability of Pt₁₀₀ and the Pt–Pd alloy NCs (Fig. 5e) over 4000 CV cycles. To account for differences in initial activity, the MA was normalised to the maximum MA obtained for each material. After 4000 cycles, Pt₁₀₀ retained only 16% of its maximum MA, whereas the Pt–Pd alloy NCs retained 40%, 70%, and 82% for Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅, respectively, demonstrating substantially improved durability relative to Pt₁₀₀. This higher retention, particularly for Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅, is also associated with an activation period during cycling in which MA initially increases before reaching a maximum. This behaviour is especially pronounced for Pt₂₅–Pd₇₅, for which the maximum MA is only reached after ∼2500 cycles.

The lower durability of Pt₁₀₀ can be attributed to the well-known susceptibility of Pt-rich surfaces to poisoning by strongly adsorbed CO-derived intermediates formed during methanol oxidation. In the absence of a secondary metal that promotes the formation of oxygen-containing species, the oxidative removal of Co_ads is less efficient, leading to progressive blockage of Pt active sites and rapid loss of activity during cycling.

The improved durability of the Pt–Pd alloy NCs is attributed to the combined electronic and bifunctional effects introduced by Pd. Alloying with Pd modifies the electronic structure of Pt, weakening the adsorption of CO-derived intermediates, while Pd sites can facilitate the generation of oxygen-containing species that assist the oxidative removal of CO_ads from neighbouring Pt sites. This effect becomes more pronounced as the Pd fraction increases, explaining the higher activity retention of Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅ compared with Pt₁₀₀. The intermediate behaviour of Pt₇₅–Pd₂₅ indicates that, although a small amount of Pd is sufficient to partially mitigate poisoning relative to Pt₁₀₀, the low surface Pd fraction does not provide enough Pd-assisted CO removal to sustain the activity over prolonged cycling.

To further investigate the activation behaviour observed for the Pt–Pd alloy NCs, ex situ XPS analysis was performed on the carbon paper electrode preparations after (i) the CV cleaning step in 0.1 M HClO₄, (ii) the anodic activation step, (iii) 1000 CV cycles, and (iv) 4000 CV cycles, using a fresh electrode for each condition. The corresponding XPS spectra and the relative surface fractions of Pt and Pd species for Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅ are presented in Fig. S11 and Table S7.

The XPS results reveal that anodic activation induces changes in surface composition for all three Pt–Pd alloy NCs, as reflected by the increase in the Pt⁰/(Pt²⁺ + Pt⁴⁺) ratio. This indicates a higher fraction of metallic Pt exposed at the catalyst surface after activation, consistent with the enhanced MOR activity observed after anodic activation in the electrochemical measurements. A similar activation-induced surface reorganisation has been reported for other Pt-based catalysts, such as Pt–Ru systems.⁴⁸ Upon continued cycling, the Pt⁰/(Pt²⁺ + Pt⁴⁺) ratio further increases for Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅, accompanied by systematic changes in Pd speciation. For Pt₅₀–Pd₅₀, the relatively low Pd surface content favours Pd oxidation, resulting in the emergence of Pd⁴⁺ contributions. In contrast, for Pt₂₅–Pd₇₅, the higher Pd surface availability favours partial reduction back to Pd⁰. These results indicate that the surface composition and oxidation state distribution of the Pt–Pd alloy NCs evolve dynamically under electrochemical cycling conditions.⁴⁸

In contrast, Pt₇₅–Pd₂₅ does not show the same sustained increase in surface Pt⁰ upon cycling; instead, the Pt⁰ fraction decreases after extended cycling, consistent with its more pronounced activity decay during ADT. For this composition, the weak Pd signal limits a reliable discussion of Pd speciation, but the limited surface Pd availability likely reduces the extent of Pd-mediated CO removal and dynamic redox stabilisation. Therefore, the durability trend reflects a balance between Pt active-site availability and Pd-assisted poisoning tolerance. Pt₁₀₀ provides abundant Pt sites but poor resistance to CO-mediated deactivation. Pt₇₅–Pd₂₅ partially mitigates this limitation but remains insufficiently protected, whereas Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅ provide enough Pd to promote sustained surface restructuring and improved long-term stability.

This dynamic surface restructuring provides a mechanistic basis for the progressive increase in catalytic activity observed during ADT, particularly for Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅. At the same time, these results highlight that changes in activity during stability measurements can arise not only from poisoning by strongly adsorbed intermediates but also from the continuous evolution of surface composition, oxidation state, and active-site availability.

To assess whether selective metal dissolution contributes to the observed activity evolution, the bulk Pt : Pd molar ratios were measured by ICP-OES after each treatment step (Table S8). No significant changes in bulk composition were detected for any of the Pt–Pd alloy NCs, indicating that the activity enhancement upon cycling primarily arises from surface restructuring and dynamic redox processes rather than from selective Pd dissolution and consequent Pt enrichment. Therefore, the evolution of activity is likely governed by near-surface processes, including changes in the oxidation states of Pt and Pd, increased availability of metallic Pt sites, and removal of adsorbed poisoning intermediates. Such dynamic surface restructuring may contribute to the improved durability of the Pt–Pd alloy NCs by maintaining or restoring favourable active-site configurations during cycling. However, because dedicated deactivation–reactivation experiments were not performed, electrochemical regeneration cannot be claimed as an established strategy at this stage. Furthermore, post-ADT electron microscopy was not performed in this study; therefore, possible changes in nanocluster size, agglomeration state, and particle detachment during cycling cannot be excluded.

In situ FTIR spectroscopic studies

In situ FTIR spectroscopy was used to gain further insight into the MOR mechanism on the Pt–Pd alloy NCs. FTIR spectra of Pt₁₀₀, Pt₇₅–Pd₂₅, Pt₅₀–Pd₅₀, and Pt₂₅–Pd₇₅ NCs recorded in 0.1 M HClO₄ + 1.0 M MeOH are shown in Fig. 6. For Pt₁₀₀, a band at around 2032 cm⁻¹ is assigned to the linearly adsorbed CO (CO_L) on Pt sites. This band shifts to 2021 cm⁻¹ for Pt₇₅–Pd₂₅ and Pt₅₀–Pd₅₀. For Pt₂₅–Pd₇₅, an additional band at 1978 cm⁻¹, attributed to bridge-bonded CO (CO_B), appears at low potentials, while the CO_L band becomes dominant at potentials above 0.50 V vs. RHE. This behaviour has been reported previously for Pt–Pd alloys with high Pd content.⁵¹

Fig. 6 In situ FTIR spectra acquired during methanol oxidation on the GDEx-produced NCs in 0.1 M HClO₄ + 1.0 M MeOH. The reference spectrum was collected at 0.03 V vs. RHE. Spectra were recorded after successive 0.05 V potential steps while increasing the potential from 0.05 to 1.20 V vs. RHE.

The observed redshift for the CO_L band in the Pt–Pd alloy NCs suggests that alloying modifies the electronic environment of Pt, consistent with electronic (ligand) effects arising from Pt–Pd interactions. Increased electron density at the Pt site weakens the Pt–CO bond, resulting in a lower C–O stretching frequency.⁵¹

In addition to changes in CO adsorption, Pd incorporation has a marked effect on the formation of CO₂. A band at 2340 cm⁻¹ assigned to CO₂ becomes visible at potentials above 0.70 V vs. RHE for Pt₇₅–Pd₂₅ and Pt₅₀–Pd₅₀ and above 0.80 V for Pt₂₅–Pd₇₅, albeit with lower intensity. For Pt₁₀₀, the corresponding band is much less pronounced under the same conditions, appearing only as a weak feature close to the detection limit. This indicates that Pt–Pd alloy NCs promote the full oxidation of methanol or the more efficient oxidative removal of CO-derived adsorbates, consistent with their improved poisoning tolerance and durability observed in ADT.⁵²

MOR on Pt-based catalysts is generally described by parallel pathways, including an indirect route involving the formation of strongly adsorbed CO and a direct route proceeding through oxygenated intermediates. In both cases, CO₂ is the final oxidation product. In the indirect pathway, methanol dehydrogenation produces CO adsorbates that accumulate at low potentials and block active sites, whereas the oxidation of CO to CO₂ requires oxygen-containing surface species generated by water activation.^53,54 Alloying Pt with Pd can promote a bifunctional mechanism, where Pd facilitates water activation at lower potentials, thereby increasing the availability of reactive oxygen-containing species that assist the oxidative removal of CO adsorbates from neighbouring Pt sites.¹¹

Overall, the FTIR results indicate that Pt–Pd alloy NCs promote the oxidation of CO-derived intermediates and CO₂ formation relative to Pt₁₀₀, supporting a synergistic mechanism in which Pd enhances the supply of oxygen-containing surface species, while Pt offers sites for methanol adsorption and dehydrogenation. This synergy enhances the resistance to poisoning and improves the stability of Pt–Pd alloy NCs.²³

Conclusions

Gas-diffusion electrocrystallisation (GDEx) constitutes an electrified route to synthesise Pt–Pd alloy NCs that exhibit MOR activity comparable to, or higher than, monometallic Pt NCs, with further enhancement after anodic activation. In addition to high activity, Pt–Pd alloy NCs show improved stability, with Pt₅₀–Pd₅₀ and Pt₂₅–Pd₇₅ displaying a gradual increase in MOR activity during CV cycling. This activation behaviour is associated with dynamic changes in the surface oxidation states of Pt and Pd, leading to an increased fraction of metallic Pt at the catalyst surface. In situ FTIR measurements indicate that Pt–Pd alloy NCs promote the oxidative removal of CO-derived intermediates and boost CO₂ formation relative to Pt, contributing to improved poisoning tolerance and durability. These results highlight the potential of GDEx as a simple and efficient strategy for producing Pt–Pd alloy electrocatalysts for direct methanol fuel cells.

Beyond the catalyst-performance trends established in this work, these findings also support the broader use of GDEx as an electrified platform for the production of noble-metal nanostructures. Recent work from our group has shown that GDEx can be coupled with microwave-assisted leaching to recover PGMs from multielement leachates of spent automotive catalysts and directly convert them into functional electrocatalytic nanoparticles.²⁹ Therefore, although the present study uses synthetic Pt and Pd precursor solutions to isolate composition–activity relationships, it further supports the potential of GDEx-derived Pt–Pd nanoclusters as value-added catalytic materials within circular PGM-recovery strategies. This perspective is particularly relevant for reducing dependence on primary Pt and Pd resources and mitigating the impact of precious-metal supply and price volatility.

Author contributions

O. M.-M., conceptualisation, data curation, formal analysis, investigation, methodology, writing – original draft. L. F. L.-F., conceptualisation, investigation. Z. S., and S. M., resources. Z. C., methodology. J. F., conceptualisation, supervision. X. D.-B., conceptualisation, supervision, and funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ta03246j.

Acknowledgements

This research was funded by the European Union's Horizon 2020 research and innovation program, under grant agreement No. 958302 (PEACOC) and the European Union's Horizon Europe research and innovation program, under grant agreement No. 101091715 (FIREFLY). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Health and Digital Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. O. M.-M. thanks SECIHTI-Mexico (former CONACYT-Mexico) for doctoral scholarship No. 766618. We thank Wilfred Brusten for his support in the ICP-OES measurements, Myrjam Mertens for her support with the XRD measurements and Charlotte Zborowski for her support with the XPS analysis. The authors also thank Alena Michalcová for TEM and STEM-EDS imaging.

References

1. X. Ren, Q. Lv, L. Liu, B. Liu, Y. Wang, A. Liu and G. Wu, Sustain. Energy Fuels, 2020, 4, 15–30.
2. A. S. Aricò, V. Baglio and V. Antonucci, Electrocatalysis of Direct Methanol Fuel Cells: from Fundamentals to Applications, 2009, vol. 1, pp. 1–78.
3. Z. Xia, X. Zhang, H. Sun, S. Wang and G. Sun, Nano Energy, 2019, 65, 104048.
4. X. Zhao, M. Yin, L. Ma, L. Liang, C. Liu, J. Liao, T. Lu and W. Xing, Energy Environ. Sci., 2011, 4, 2736–2753.
5. S. Dai, M. Li, H. Li, Y. Shi, H. Zhang, D. Wang, K. Xiang, J. Zou and G. Luo, ChemSusChem, 2025, 18, e202402767.
6. N. Kakati, J. Maiti, S. H. Lee, S. H. Jee, B. Viswanathan and Y. S. Yoon, Chem. Rev., 2014, 114, 12397–12429.
7. Y. Zuo, W. Sheng, W. Tao and Z. Li, J. Mater. Sci. Technol., 2022, 114, 29–41.
8. H. Tian, Y. Yu, Q. Wang, J. Li, P. Rao, R. Li, Y. Du, C. Jia, J. Luo, P. Deng, Y. Shen and X. Tian, Int. J. Hydrogen Energy, 2021, 46, 31202–31215.
9. C. Koenigsmann and S. S. Wong, Energy Environ. Sci., 2011, 4, 1161–1176.
10. A. Bin Yousaf, M. Imran, N. Uwitonze, A. Zeb, S. J. Zaidi, T. M. Ansari, G. Yasmeen and S. Manzoor, J. Phys. Chem. C, 2017, 121, 2069–2079.
11. Y. Liu, M. Chi, V. Mazumder, K. L. More, S. Soled, J. D. Henao and S. Sun, Chem. Mater., 2011, 23, 4199–4203.
12. J. Wang, B. Zhang, W. Guo, L. Wang, J. Chen, H. Pan and W. Sun, Adv. Mater., 2023, 35, 2211099.
13. E. A. Franceschini, M. M. Bruno, F. J. Williams, F. A. Viva and H. R. Corti, ACS Appl. Mater. Interfaces, 2013, 5, 10437–10444.
14. C. Du, X. Gao, Z. Zhuang, C. Cheng, F. Zheng, X. Li and W. Chen, Electrochim. Acta, 2017, 238, 263–268.
15. J.-N. Zheng, L.-L. He, C. Chen, A.-J. Wang, K.-F. Ma and J.-J. Feng, J. Power Sources, 2014, 268, 744–751.
16. W. Chen, Z. Lei, T. Zeng, L. Wang, N. Cheng, Y. Tan and S. Mu, Nanoscale, 2019, 11, 19895–19902.
17. W. Gong, Z. Jiang, R. Wu, Y. Liu, L. Huang, N. Hu, P. Tsiakaras and P. K. Shen, Appl. Catal., B, 2019, 246, 277–283.
18. J. Xu, S. Guo, F. Hou, J. Li and L. Zhao, Int. J. Quantum Chem., 2018, 118, e25491.
19. W. Li, N. Bhuvanendran, H. Liu, W. Zhang, K. Hooshyari, S. Y. Lee, Q. Xu and H. Su, Int. J. Hydrogen Energy, 2024, 50, 1496–1506.
20. Y. Lu, Y. Jiang, H. Wu and W. Chen, J. Phys. Chem. C, 2013, 117, 2926–2938.
21. K. Dong, H. Dai, H. Pu, T. Zhang, Y. Wang and Y. Deng, Int. J. Hydrogen Energy, 2023, 48, 35240–35249.
22. L. Sun, H. Wang, K. Eid, S. M. Alshehri, V. Malgras, Y. Yamauchi and L. Wang, Electrochim. Acta, 2016, 188, 845–851.
23. Y. Zhang, G. Shu, Z. Shang, K. Ma, L. Song, C. Wang, C.-A. Zhou and H. Yue, ACS Sustain. Chem. Eng., 2023, 11, 8958–8967.
24. Y. W. Lee, S. W. Han and K. Y. Lee, Mater. Chem. Phys., 2018, 214, 201–208.
25. A. E. Hughes, N. Haque, S. A. Northey and S. Giddey, Resources, 2021, 10, 93.
26. O. Martinez-Mora, G. Pozo, L. F. Leon-Fernandez, J. Fransaer and X. Dominguez-Benetton, RSC Sustain., 2023, 1, 454–458.
27. L. F. Leon-Fernandez, A. Caballero-Ortiz, O. Martinez-Mora, J. Fransaer and X. Dominguez-Benetton, Electrochim. Acta, 2023, 460, 142592.
28. G. Nicol, E. Goosey, D. Ş. Yıldız, E. Loving, V. T. Nguyen, S. Riaño, I. Yakoumis, A. M. Martinez, A. Siriwardana and A. Unzurrunzaga, Johnson Matthey Technol. Rev., 2021, 65, 127–147.
29. L. F. Leon-Fernandez, F. Kukurugya, O. Martinez-Mora, S. Mourdikoudis, L. Zeng, M. M. Parascanu, A. Saez, J. Fransaer, J. Spooren and X. Dominguez-Benetton, RSC Sustain., 2026 10.1039/D6SU00190D.
30. O. Martinez-Mora, L. F. Leon-Fernandez, M. Velimirovic, F. Vanhaecke, K. Tirez, J. Fransaer and X. Dominguez-Benetton, Mater. Adv., 2023, 4, 6183–6191.
31. O. Martinez-Mora, K. Tirez, F. Beutels, W. Brusten, L. F. Leon-Fernandez, J. Fransaer, X. Dominguez-Benetton and M. Velimirovic, Nanomaterials, 2023, 13, 2610.
32. F. J. Vidal-Iglesias, R. M. Arán-Ais, J. Solla-Gullón, E. Herrero and J. M. Feliu, ACS Catal., 2012, 2, 901–910.
33. C. L. Green and A. Kucernak, J. Phys. Chem. B, 2002, 106, 1036–1047.
34. J. Zhang, X. Qu, Y. Han, L. Shen, S. Yin, G. Li, Y. Jiang and S. Sun, Appl. Catal., B, 2020, 263, 118345.
35. J.-Y. Ye, Y.-X. Jiang, T. Sheng and S.-G. Sun, Nano Energy, 2016, 29, 414–427.
36. D. Yi, B. Chaudret and K. Soulantica, in Reducing Agents in Colloidal Nanoparticle Synthesis, ed. S. Mourdikoudis, The Royal Society of Chemistry, 2021.
37. E. Groppo, G. Agostini, E. Borfecchia, L. Wei, F. Giannici, G. Portale, A. Longo and C. Lamberti, J. Phys. Chem. C, 2014, 118, 8406–8415.
38. K. M. Koczkur, S. Mourdikoudis, L. Polavarapu and S. E. Skrabalak, Dalton Trans., 2015, 44, 17883–17905.
39. J. Xian, Q. Hua, Z. Jiang, Y. Ma and W. Huang, Langmuir, 2012, 28, 6736–6741.
40. X. Yu, J. Rong, Z. Zhan, Z. Liu and J. Liu, Mater. Des., 2015, 83, 159–163.
41. E. G. Ciapina, M. L. dos Santos, R. M. I. S. Santos, J. Palombarini, O. P. Almeida Júnior, J. C. C. de Castro Santana, D. A. Modesto, A. J. C. Lanfredi and S. F. Santos, J. Alloys Compd., 2021, 881, 160628.
42. S. Mourdikoudis, R. M. Pallares and N. T. K. Thanh, Nanoscale, 2018, 10, 12871–12934.
43. X. Chen, L. P. Granda-Marulanda, I. T. McCrum and M. T. M. Koper, Chem. Sci., 2020, 11, 1703–1713.
44. E. V Gerasimova, S. A. Kleinikova, N. V Talagaeva, K. V Gor’kov, M. G. Levchenko and E. V Zolotukhina, Int. J. Hydrogen Energy, 2023, 48, 34396–34409.
45. M. Shao, J. H. Odell, S.-I. Choi and Y. Xia, Electrochem. Commun., 2013, 31, 46–48.
46. E. Antolini and J. Perez, J. Mater. Sci., 2011, 46, 4435–4457.
47. S. M. Alia, G. Zhang, D. Kisailus, D. Li, S. Gu, K. Jensen and Y. Yan, Adv. Funct. Mater., 2010, 20, 3742–3746.
48. M. V Martínez-Huerta, N. Tsiouvaras, M. A. Peña, J. L. G. Fierro, J. L. Rodríguez and E. Pastor, Electrochim. Acta, 2010, 55, 7634–7642.
49. A. M. Hofstead-Duffy, D.-J. Chen, S.-G. Sun and Y. J. Tong, J. Mater. Chem., 2012, 22, 5205–5208.
50. D. Y. Chung, K.-J. Lee and Y.-E. Sung, J. Phys. Chem. C, 2016, 120, 9028–9035.
51. C. Wang, B. Peng, H.-N. Xie, H.-X. Zhang, F.-F. Shi and W.-B. Cai, J. Phys. Chem. C, 2009, 113, 13841–13846.
52. H.-X. Liu, N. Tian, M. P. Brandon, Z.-Y. Zhou, J.-L. Lin, C. Hardacre, W.-F. Lin and S.-G. Sun, ACS Catal., 2012, 2, 708–715.
53. M. Neurock, M. Janik and A. Wieckowski, Faraday Discuss., 2009, 140, 363–378.
54. S. X. Liu, L. W. Liao, Q. Tao, Y. X. Chen and S. Ye, Phys. Chem. Chem. Phys., 2011, 13, 9725–9735.

---

Footnotes

† Current affiliation: Applied Electrochemistry & Catalysis (ELCAT), University of Antwerp, Universiteitsplein 1, 2610, Wilrijk, Belgium.

‡ Current affiliation: CINBIO, Universidade de Vigo, Materials Chemistry and Physics Group, Department of Physical Chemistry, Campus Universitario Lagoas Marcosende, 36310 Vigo, Spain.

§ Current affiliation: Equinor Energy Belgium, Kortenberglaan 120, 1000, Brussels, Belgium. E-mail: xochitldb@gmail.com

---