Printed Liquid Metal–Solid Metal Hybrid Electrodes for Stabilizing Liquid Platinum–Gallium Droplets During Electrocatalysis

ABSTRACT

Dispersing catalytically active gallium (Ga)‐based liquid metal (LM) nano‐ and micro‐droplets onto solid metallic mesh substrates remains challenging due to their strong tendency to either agglomerate or migrate away from the substrate. Another challenge is the reactivity of Ga with support metals, often forming discrete solid intermetallic phases. Here, we develop a 3D matrix formed by fused tungsten nanoparticles (W NPs) to physically entrap Ga‐based LM nano‐droplets, which effectively prevent their agglomeration and stabilize their reactive surface. Printing these W/LM composites onto a porous substrate (e.g., molybdenum (Mo) mesh) further promoted their dispersion and brought about enhanced electrochemical reactivity. To verify the efficiency of this strategy, we printed Pt‐in‐Ga droplets mixed with W NPs onto porous Mo substrate and evaluated their performance by two model reactions‐CO₂RR (CO₂ reduction reaction) and HER (Hydrogen evolution reaction). The printed showed remarkable stability and reactivity toward both reactions. In addition, computational investigations revealed distinct active motifs for these two reactions, supporting the notion that adaptive LM catalysts can facilitate diverse reaction pathways depending on the targeted reactions.

A printed liquid‐metal catalyst platform is developed by confining Pt‐in–Ga liquid metal droplets within a tungsten nanoparticle matrix on porous Mo substrates. This architecture suppresses liquid metal droplet agglomeration and leaching, enabling durable and highly active electrocatalysis for both HER and CO₂RR, with Pt‐dependent active motifs revealed by experimental and computational studies.

1. Introduction

Electrocatalysis underpins a broad range of reactions that are vital for sustainable energy applications and chemical technologies [1, 2, 3, 4]. Ga‐based LMs have emerged as a new class of electrocatalysts that challenge the orthodoxy of well‐defined reaction sites, instead offering unique reaction pathways through their adaptive interfaces. The dynamic surface properties enhance resistance to coking and deactivation, while also allowing reactive sites to adapt to the surface‐adsorbed reactants through a process of structural superpositioning akin to enzymatic processes [5, 6, 7, 8]. These effects have been particularly shown in alcohol oxidation reactions, where the amount of active noble metal could be reduced by orders of magnitude [9, 10, 11]. This high activity has been attributed in part to the dissolved noble metal being able to migrate within the molten metal towards the surface, minimizing unutilized atoms in the bulk while also activating surrounding Ga atoms.

One of the challenges of using LMs as catalysts is that their fluidic nature renders them highly sensitive to changes in surface tension under electrochemical conditions. An applied potential can drastically tune the surface tension of a LM from >700 mN/m down to near zero, leading to morphological instabilities, droplet migration, and ultimately coalescence and surface area loss [12]. The supported catalytically active LM solutions (SCALMS) strategy has been effectively employed to distribute LM droplets on porous supports, such as SiO₂ or Al₂O₃, to enhance catalytic activity and stability [13, 14]. However, some limitations associated with these catalytic systems include the potential detachment of droplets from supports during reaction, which results in catalyst loss, as well as the limited ability of the traditional porous supports to provide optimal electron transport pathways. Furthermore, some support materials may lack the necessary durability for sustained operational performance. Addressing these issues is crucial for realizing the long‐term operational stability of LM‐based electrodes.

Previous studies have demonstrated the stability and structural robustness of Mo‐ and W‐based support materials in electrocatalytic applications [15, 16, 17]. Unlike most transition and noble metals, Mo and W remain chemically compatible with Ga and do not form alloys or intermetallic phases. Therefore, in this work, we developed a strategy to confine the LM droplets using the fused W NPs as the “cage” and then print this LM/W mixture onto porous Mo substrate for a higher accessible surface area. The efficiency of this design was validated using two model reactions: CO₂RR and HER with PtGa droplets as the active electrocatalysts. Electrochemical evaluation revealed that printed electrodes with Pt‐in‐Ga droplets containing 0.35 and 0.045 wt.% Pt exhibited more favorable catalytic activity towards HER and CO₂RR, respectively, i.e., a high mass activity of 2.43 A mg^{− 1} _Pt at −0.100 V vs. RHE for HER, and 3.02 A mg^{− 1} _{P t} at −1.140 V vs. RHE for CO₂RR. Computational modeling revealed the mechanism behind the seemingly paradoxical trend: high Pt content promotes HER, while low Pt favors CO₂RR. The introduction of Pt created localized electron‐rich sites that enhance H* coverage but weaken H* binding, thus facilitating H₂ evolution. In contrast, lower Pt content better modulated the oxygen binding strength on Ga surfaces, making it more favorable for CO₂ reduction. The electrode architecture developed in this work is readily adaptable to LM catalysts with diverse compositions, offering a robust platform for assessing their catalytic activity and long‐term stability.

2. Results and Discussion

As illustrated in Figure 1, the printed electrode was fabricated by slurry‐coating a Mo mesh with a mixed slurry of Pt‐Ga droplets and W NPs (30 nm in size), followed by high‐temperature annealing under Ar for 2 h to form a robust, porous structure suitable for subsequent electrochemical reactions (see the Experimental Section for details). High‐temperature annealing is crucial for promoting the fusion of W NPs, which effectively entrapped the Pt‐Ga droplets and ensured strong adhesion to the Mo substrate.

FIGURE 1.

Schematic illustration of liquid Pt‐in‐Ga‐based electrodes synthesis and electrochemical setup for HER and CO₂RR.

Figure 2a shows the TEM image and individual elemental mapping of Pt‐in‐Ga droplets with 0.35 wt.% Pt in Ga (before annealing), which confirms the uniform dispersion of Pt within the Ga matrix. The absence of crystalline peaks and the presence of a broad diffuse scattering signal in the XRD pattern, as presented in Figure 2b, confirm the liquid state of Pt‐in‐Ga droplets and uniform dispersion of Pt. The ability to form well‐dispersed Pt within the Ga matrix at room temperature—without generating intermetallic—appears to contradict the classical Ga–Pt phase diagrams, which predict only minimal Pt solubility [11, 13, 18, 19]. This discrepancy is attributed to nanoscale size effects, which are known to influence the melting point and phase behaviour of materials significantly [20].

FIGURE 2.

Surface and elemental analysis of Pt‐in‐Ga droplets and electrode before and after annealing; (a) TEM imaging of an individual droplet and elemental mapping of Ga, and Pt for 0.35 wt.% Pt‐in‐Ga, (b) XRD pattern of 0.35 wt.% Pt‐in‐Ga droplets before annealing, (c) XPS spectrum of the Ga 3d for 0.35 wt.% Pt‐in‐Ga electrode after annealing in Ar at 850°C, (d) XPS spectrum of Pt 4f, and (e) SEM image of the electrode having 0.35 wt.% Pt‐in‐Ga and elemental mapping of individual elements (Left to right).

XPS was conducted on the electrodes before and after annealing, as presented in Figure 2c,d. The surface was dominated by Ga³⁺, which is due to Ga oxide formation on the surface upon exposure to air, which is consistent with literature [21]. The XPS spectrum of Ga 3d indicates the presence of strong Ga⁰ as presented in Figure 2c, which proves that even after annealing in Ar at 850°C, the core of Ga droplets remains metallic.

Peaks in the Pt region are discernible but noisy due to the low concentration of Pt approaching the detection limit of XPS. The Pt 4f_7/2 and Pt 4f_5/2 peaks observed at 72.6 and at 75.5 eV, as shown in Figure 2d, are in good agreement with zero‐valent Pt. The previously identified complex structure of Pt‐rich Ga alloys is absent, suggesting the absence of solid Pt‐in‐Ga intermetallics and the liquid state of Pt in the Ga matrix [22]. Furthermore, the binding energies of Pt and Ga are higher in the liquid state as compared to the solid state, indicating lower electron density, enhanced degree of freedom which emerges during solid to liquid phase transition, higher atomic spacing, and reduced atomic coordination, which makes Pt atoms appear as single atoms or clusters, providing more active sites for reduction reactions [23]. Figure 2e (left to right) shows SEM images after annealing, illustrating the entrapment of the Pt‐in‐Ga droplets in W network. The XPS and SEM results of Mo mesh and Mo mesh coated with only W powder after annealing are presented in Figures S1–S3. Collectively, these characterization techniques confirmed the caging of W networks without affecting the dispersion and liquid state of the Pt‐in‐Ga droplets.

The HER is a widely studied process and can thus serve as a model reaction for evaluating the activity of the as‐designed electrode architectures. With a similar strategy, we prepared catalytic electrodes with Pt loadings ranging from 0.004 to 7.8 wt.% in Ga and tested their HER activity. We also evaluated (i) bare Mo mesh, (ii) Mo mesh coated with W powder, and (iii) Mo mesh coated with W powder and Ga droplets in the absence of Pt, which all showed negligible catalytic activity toward HER. In contrast, when Pt was introduced into the Ga droplets and incorporated into the W/Mo support structure, a significant HER activity was observed, confirming Pt as the primary active site for HER. (Figure S4). As shown in Figure 3a, the mass activity for Pt in Pt‐Ga containing 0.35 wt.% Pt was 7 times higher than that from commercial Pt/C, highlighting the crucial role of Ga in dispersing Pt and enhancing its catalytic activity.

FIGURE 3.

Catalytic activities of different catalysts for HER in an acidic medium, (a) Comparison of current densities normalized to the geometric surface area of the electrodes (n=5) in 0.5 M H₂SO₄, (b) Comparison of mass activities normalized to the mass of Pt at −0.100 V vs. RHE, (c) Comparison of chronoamperometry analysis of 0.35 wt.% W/Pt‐in‐Ga electrode at different voltages vs. RHE for HER, and (d) Comparison of HER performance of 0.35 wt.% W/Pt‐in‐Ga between first and 700th cycles.

It is also worth noting that HER reactivity increased with increasing Pt concentration from 0.004 to 7.8 wt.%. However, increasing the Pt loading beyond 0.35 wt.% only brought about a marginal increase in the activity. For example, the electrodes prepared with Pt‐in‐Ga droplets containing 0.35 wt.% Pt exhibited nearly identical onset potential to that with 7.8 wt.% Pt. The specific mass activity of 7.8 wt.% Pt (0.14 A mg⁻¹ _Pt at −0.100 V vs. RHE) was approximately 15 times lower than that of 0.35 wt.% Pt (2.43 A mg⁻¹ _Pt at −0.100 V vs. RHE) as shown in Figure 3b. This is likely attributed to Pt aggregation within the Ga matrix at higher loadings. This hypothesis was supported by SEM, TEM elemental mapping, and DSC analysis (Figures S5 and S6, Supplementary data), which indicate the formation of intermetallic phases and clustering of Pt at higher concentrations. In contrast, HRTEM analysis did not reveal any crystalline Pt nanoparticles or intermetallic domains (Figure S7a), supporting the presence of a highly dispersed Pt state. This is further confirmed by the DSC results, where no melting transition associated with Pt–Ga intermetallics was detected (Figure S7b).

Chronoamperometry measurements were carried out at different potentials (from 0.024 to −0.376 V vs. RHE) for 1 h to assess the electrochemical stability of the printed 0.35 wt.% Pt‐in‐Ga electrode. As shown in Figure 3c, stable current densities were observed at all applied potentials, with no noticeable activity decay during the measurement period. Based on these results, long‐term durability was further evaluated by continuous operation at −0.276 V vs. RHE for 5 h, during which the electrode maintained stable HER activity over time (Figure S8). Notably, a comparison of the HER performance recorded before and after the prolonged chronoamperometry test revealed a slight enhancement in activity (Figure S9). This improvement is attributed to electrochemical activation of the electrode during extended operation, likely arising from improved wetting of the catalyst surface, increased accessibility of catalytically active Pt sites, and interfacial restructuring under HER conditions, collectively confirming the excellent durability and robustness of the printed electrode.

The durability of the electrode was further validated by linear sweep voltammetry (LSV) measurements conducted at a scan rate of 5 mV s⁻¹, with the first, 500th, and 700th cycles compared in Figure 3d. The electrode showed enhanced performance from the first to the 500th cycle and remained stable up to the 700th cycle. The improved HER performance can be attributed to surface activation or better contact between the electrode and electrolyte during prolonged HER.

The Tafel slope for 0.35 wt.% Pt‐in‐Ga electrode was calculated to be ∼ 31.3 mV dec⁻¹ (Figure S10), which is very close to commercial Pt/C catalyst with 30 wt.% Pt (∼30 mV dec⁻¹) [24]. Furthermore, the HER kinetics of the 0.35 wt.% Pt‐in‐Ga catalyst were evaluated by Tafel analysis in 0.5 M H₂SO₄. At this cathodic potential, proton reduction is the only thermodynamically favorable reaction, while other possible reactions, such as the reduction of sulfate or W, are negligible. The measured Tafel slope of 31.3 mV dec^{− 1} indicates that the HER follows a Volmer–Tafel mechanism with the recombination of adsorbed hydrogen atoms (Tafel step) as the rate‐limiting process [25]. This low Tafel slope, together with the high current density, confirms the high intrinsic activity of the PtGa/W electrode toward HER. A comparison of overpotentials, Tafel slopes, and current densities at −0.100 V vs. RHE for different Pt‐based catalysts in acidic media is further presented in (Figure S23, Table S1, supplementary data), which demonstrates the efficacy of our Pt‐in‐Ga electrodes with ultra‐low Pt loading by leveraging the liquid‐metal‐based Pt‐in‐Ga system and the unique Mo/W scaffold electrode architecture, which together enable efficient atomic utilization of Pt while ensuring long‐term stability towards HER [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39].

Moreover, the mechanical robustness of the printed W/Pt‐in‐Ga electrode was evaluated by testing its activity after the bending and scratching test (see the Experimental Section for full details). Specifically, the catalytic activity of the electrode was tested for HER after each bending cycle (Figures S11 and S12, supplementary data). LSV data showed that moderate bending (particularly around 1.6 cm led to an enhancement in HER activity, potentially due to increased exposure of active sites or improved catalyst‐support interfaces. However, a minor decrease in HER activity was observed under more severe deformation conditions (0.9 cm diameter, 50 cycles). Overall, the electrode exhibited excellent mechanical robustness with minimal activity loss even under significant mechanical stress.

CO₂RR is another widely studied reaction, which involves multi‐electron, multi‐step pathways with more sluggish kinetics and product complexity. Generally, Pt‐based catalysts show inferior CO₂RR activity in comparison to their HER activity [40]. Recent studies suggest that modifying the local environment of Pt through alloying or nano‐structuring can improve CO₂RR selectivity [3, 41]. To evaluate the efficiency of our electrode, we systematically tuned the Pt concentrations in Ga across a range from 0.004 to 0.4 wt.% and assessed their performance towards CO₂RR. The catalytic electrodes showed negligible current in the N₂‐saturated electrolyte. A significantly increased current was observed when the electrolyte was saturated with CO₂, as shown in Figure 4a, indicating their activity towards CO₂RR. W and W‐Ga without Pt were also tested and showed negligible catalytic activity toward CO₂RR under identical electrochemical conditions (Figure S13), confirming the activity originated from Pt.

FIGURE 4.

Electrochemical reduction of CO₂, (a) Comparison of current densities normalized to the geometric surface area of electrodes, (inset) Comparison of mass activities normalized to the mass of Pt for CO₂ reduction, (b) Tafel plot for CO₂ reduction over 0.045 wt.% Pt‐in‐Ga electrode, and (inset) CA analysis for CO₂ reduction at −1.190 V vs. RHE for 4 h.

Again, increasing the Pt concentration from 0.004 to 0.045 wt.% led to a steady increase in current. However, further increase in Pt concentration beyond 0.045 to 0.35 wt.% resulted in a decline in current density. This is not due to intermetallic formation; our earlier results confirmed no intermetallics at 0.35 wt.% Pt. A comparison of the specific mass activities of the catalysts, normalized to Pt content, is presented as an inset of Figure 4a. For Pt‐in‐Ga droplets containing 0.045 wt.% Pt, the Tafel slope was found to be ∼23 mV dec⁻¹ as given in Figure 4b, which is ∼5 times lower than the catalysts reported in the literature [42, 43]. A lower Tafel slope implies that the reaction rate increases more rapidly with applied overpotential, indicating more efficient charge transfer at the electrode–electrolyte interface. This suggests that the reaction kinetics on the W/Pt‐in‐Ga surface are significantly enhanced, likely due to the optimized electronic environment provided by the Pt–Ga interface, which facilitates intermediate adsorption and charge transfer during CO₂ reduction. Long‐term stability testing of the 0.045 wt.% Pt‐in‐Ga was conducted using CA at −1.190 V vs. RHE for 4 h to investigate the stability of the electrode. The stable mass activity observed during CA further confirmed the structural robustness and electrochemical durability of the electrode as presented in the inset of Figure 4b. This stability is attributed to the strong interfacial confinement of Pt‐in‐Ga droplets by W particles, which prevent agglomeration or loss of active sites, while the liquid Ga matrix preserves the dispersion and accessibility of Pt.

The leaching and dissolution of Ga into the solution remains a major concern for LM‐based systems. ICP‐MS analysis of the electrolyte after HER at different applied potentials and durations was used to quantify dissolution of W and Ga during the reaction. Specifically, the concentration of W remained relatively constant and low after testing under different voltages and time intervals. On the other hand, a small amount of Ga dissolved into the electrolyte at the beginning of the test; however, with prolonged operation, the dissolution stabilized, and the Ga concentration remained essentially constant at higher voltages (Figure S14). A similar trend was observed from the ICP‐MS analysis of the electrolyte following CO₂RR at −1.190 V vs. RHE for 1, 4, and 24 h, revealing that the concentration of Ga remained nearly constant and low during the test (Figure S15), while W kept dissolving after the long‐term test. Despite this increase, the amount of W detected after 24 h corresponds to approximately 1% of the total W content on the electrode before electrolysis (Figure S15). This behaviour reflects the intrinsic properties of W, and it may be possible to further suppress its dissolution through appropriate surface modification strategies.

We also analyzed the products from CO₂RR. The formation of gas bubbles was observed during the reduction reaction. The compositional analysis of the gas produced confirms the presence of CH₄ (0.09%) and CO (0.07%), whereas the amounts of O₂ and N₂ were detected due to their presence in background air. However, they do not interfere with the interpretation of the analytical results. Liquid products were analyzed by ¹HNMR and ¹³CNMR after CO₂RR. The ¹HNMR peaks at 1.03, 3.48, and 4.64 ppm can be attributed to the presence of water, D₂O, and DMF in the electrolyte [44, 45]. The peaks in the ¹H NMR spectrum at ∼5.26 and ∼8.67 ppm confirm the presence of formate. Furthermore, the peaks in the ¹³C NMR spectrum at ∼73.7 and ∼164.4 ppm also confirmed the formation of ethanol and formate, respectively (Figures S16 and S17, supplementary data) FTIR spectroscopy analysis of the fresh and after‐reaction electrolytes was also performed to identify the liquid products which confirmed the formation of formate (∼1180 cm⁻¹) and ethanol (∼1045 cm⁻¹) during CO₂ reduction (Supporting Information) compared with the FTIR spectrums of reference 1 M ethanol and 1 M formic acid solutions (Figure S18). Raman spectroscopy was conducted to investigate potential solid products on the electrode surface after CO₂RR. However, no deposit of carbon was detected, suggesting that the main products of CO₂ reduction are liquids, namely formate and ethanol. These findings underscore the potential of the catalyst system for efficient and selective CO₂‐to‐liquid conversion, especially toward value‐added C₁ and C₂ products.

The surface chemistry, morphology, and elemental composition analysis of the W/Pt‐in‐Ga electrode after HER and CO₂RR are presented in Figure 5. As shown in Figure 5a, no significant change was observed in the surface morphology of the 0.35 wt.% W/ Pt‐in‐Ga electrode after CA testing, as the electrode retained the cage‐like structure, showing the liquid Pt‐in‐Ga droplets entrapped by the W particles, and similar for 0.045 wt.% Pt‐in‐Ga electrode after CO₂RR as well (Figure S19). The XPS analysis obtained from the sample surface is shown in Figures 5b,c. The XPS spectrum of Ga 3d for 0.35 wt.% W/Pt‐in‐Ga after HER was positioned approximately at 20.4 eV can be attributed to Ga oxide, and at 18.48 eV indicates Ga⁰, as shown in Figure 5b. The peaks for Pt 4f_7/2 and Pt 4f_5/2 for 0.35 wt.% W/PtGa electrodes were approximately at 72.6 and 75.5 eV, respectively, attributed to the presence of Pt as shown in Figure 5c. The minor shift toward positive binding energy of Pt 4f is attributed to the local environment of the Pt atoms, which is surrounded by Ga atoms, consistent with a previously reported study [20]. The peak positions overall remained unchanged, indicating the stability of both Ga and Pt during the test.

FIGURE 5.

Surface morphology and elemental analysis of the electrode having 0.35 wt.% Pt‐in‐Ga after HER, (a) Electron image of 0.35 wt.% W/Pt‐in‐Ga electrode, EDS elemental mapping of W, Ga, Pt, and O (Left to right), (b) XPS spectrum of the Ga 3d, (c)XPS spectrum of the Pt 4f region.

As observed in both the HER and CO₂RR experiments, variations in Pt content within the electrodes influence the reactions differently. For HER, the activity increases with Pt concentration, provided that no intermetallics are formed. In contrast, for CO₂RR, increasing the Pt content from 0.045 to 0.35 wt.% leads to a decrease in current, and this effect is not attributable to intermetallic formation. We therefore attribute this contrasting behaviour to differences in the underlying reaction mechanisms. To gain further insight into the HER and CO₂RR reactions using our printed electrode loaded with different concentrations of Pt, ab initio molecular dynamics (AIMD) simulations were performed. As the liquid metal itself is dynamic [46], and adsorbed species can behave dynamically as well [47], it is impossible to sample all possible configurations. Instead, we focused on a single configuration derived from AIMD simulations and examined the effects of alchemically transforming Pt to Ga without altering the atomic configuration in any other way.

For the HER reaction, in one case, we observed H₂ spontaneously form from two surface H* in the Pt‐in‐Ga system. Therefore, we used snapshots from just before and after the reaction for comparison. As expected, the reaction was found to be more energetically favorable in the high‐Pt system (2 Pt, 498 Ga) with a value of −0.31 eV, compared to the low‐Pt system (1 Pt, 499 Ga) with a value of −0.18 eV as presented in Figure 6a,b. In the Ga‐only system (500 Ga), the reaction was energetically unfavorable by 0.29 eV (Figure S21). Other studies have found that in systems of H* adsorbed on Ru (0001) surfaces, increasing the H* coverage leads to an increase in electron delocalization across the surface, which disfavors additional adsorption [48].

FIGURE 6.

Top view and side view (cross‐section) for systems (a) HER with 2 Pt atoms in the system, (b) HER with 1 Pt atom in the system.

Because Pt is negatively charged in Ga‐Pt [11], 4 additional Pt in the system provides additional reaction sites with localized electron density around the Pt centers, resulting in greater H* coverage but weaker H* binding to the surface, thereby favoring H₂ formation.

For CO₂RR, the established mechanism on solid‐state catalysts involves the adsorption of CO₂ in a bent conformation as the first step [49]. However, we note that this is an acid‐catalyzed mechanism, whereas our system is weakly basic. In our AIMD simulations, we did not observe CO₂ adsorption; therefore, we selected a snapshot where CO₂ was found close to the Pt as a starting point for further investigation. Attempts to force the CO₂ into a bent adsorbed configuration failed, as the CO₂ relaxed to a linear conformation during geometry optimization. We, therefore, investigated a model reaction where one O atom was removed from CO₂ and placed on the surface. This reaction was energetically unfavorable in all model systems but more favorable (0.64 eV) with 2 Pt in the system than with one Pt (0.80 eV) or pure Ga (1.71 eV). However, as this reaction occurs at the cathode, we investigated the effects of adding an electron to the system. In this case, the reaction with two Pt rises to 0.82 eV. In comparison, with one Pt, it lowers to 0.67 eV, thus rendering the low‐Pt system more energetically favorable. The results are presented in Figure 7a,b.

FIGURE 7.

Top view and side view (cross‐section) for systems (a) CO₂RR with 2 Pt atoms in the system, (b) CO₂RR with 1 Pt atom in the system.

This is reasonable, as in a negatively charged system such as a cathode, adding an element that would further localize negative charge near the surface would disfavor the addition of a negatively charged oxygen to the surface. For the Pt‐in‐Ga system, the Pt atom is found below the atomic layer and imparts a partial positive charge to the Ga atoms above it [50]. In contrast, for the pure Ga system, surface Ga atoms have fluctuating partial charges of ± 0.08 e [51], with adjacent Ga atoms on the surface having opposite charges. This results in configurations where the O atom is bonded to three Ga atoms above the Pt atom in systems containing Pt, while in the pure Ga system, the O atom is only bonded to two Ga atoms, despite all systems starting with identical atomic coordinates (Figure S22).

3. Conclusions

In this study, we developed a robust and efficient liquid metal‐based catalytic electrode by printing the mixture slurry formed from Pt‐in‐Ga droplets with W powder onto a Mo mesh. The fused W powder then served as a matrix to entrap the Pt‐in‐Ga droplets, which effectively addresses the instability and leaching issues of liquid metals under electrochemical conditions, thereby enabling long‐term durability and reproducibility. The engineered electrode exhibited outstanding catalytic performance, achieving a Pt mass activity nearly 7 times higher than that of commercial Pt catalysts for HER. Density functional theory calculations further revealed the favorable energetic and active sites of the printed electrodes towards different reactions. Overall, this work underscores the importance of rational structural design in stabilizing liquid metal catalysts and offers a promising pathway for advancing electrochemical CO₂ conversion and hydrogen production technologies.

4. Experimental Sections

4.1. Materials

All the materials used including Pt, Ga, W micro and nano powder, ethanol, Mo mesh, isopropanol, nitric acid (HNO₃), sulfuric acid (H₂SO₄), deuterium oxide (D₂O), N, N‐dimethylformamide (DMF), and tetrabutylammonium hexafluorophosphate (TBAPF₆) of analytical grades (99.9% purity) were purchased from Sigma Aldrich, Australia. Commercial Pt on carbon (Pt/C, 30% Pt) was purchased from the Fuel Cell Store, USA. Ar gas (99.999%) bottle was purchased from BOC, Australia. All chemicals were of analytical grade and used without further purification. Deionized (DI) water was obtained from an ELGA ultrapure water system.

4.2. Pt‐in‐Ga Alloys & Nano Droplets Synthesis

Different Pt‐in‐Ga alloys were prepared by mixing a known amount of Pt‐black nano powder with Ga. The mixture was alloyed at 350°C inside a glovebox to avoid oxidation. The temperature for the alloying was selected based on the binary phase diagram of the Pt‐in‐Ga [52]. The metals were alloyed inside a glove box using a mortar and pestle until complete homogeneity was observed after about 1 h. Later, the alloys were allowed to cool down to room temperature and stored in a freezer at −80°C for solidification. After solidification, alloys were cut into small chunks and stored in an airtight container at 4°C.

Nanodroplets of the prepared alloys as described above were synthesized at 400°C by using 1.8 g of individual alloy with a probe sonicator in molten sodium acetate solvent (15.8 g). Sonication was performed for 30 min with a SCIENTZ‐IID probe sonicator (6 mm titanium tip, 300 W). The reaction temperature was controlled on a hotplate in a custom aluminum heating block using a thermocouple [21]. After sonication, the mixture was cooled to room temperature and washed with deionized water and ethanol to remove impurities. The droplets were dried in a vacuum oven at 25°C for 4 h. and stored at 4°C for later use.

To accurately quantify the Pt content in the catalysts, inductively coupled plasma mass spectrometry (ICP–MS) was employed. For sample preparation, 10 mg of the Pt–in–Ga droplets were completely dissolved in 20 mL of freshly prepared aqua regia (HCl/HNO₃, 3:1 v/v), ensuring full digestion of both Pt and Ga. From this stock solution, a 10 µL aliquot was extracted and diluted to a final volume of 10 mL using 2% HNO₃ as the matrix for ICP–MS analysis.

The ICP–MS provided the Pt concentration in the diluted solution. The total Pt mass in the original 10 mg of droplets was calculated by applying the dilution factor as follows:

$$m_{Pt,total}=C_{Pt,ICP}\times V_{diluted}\times \frac{V_{stock}}{V_{aliquot}}$$ (1)

where C_Pt, ICP = Pt concentration measured by ICP–MS (mg L⁻¹), V_diluted = 0.010 L (final diluted volume), V_stock = 20 mL (initial digestion volume), V_aliquot = 10 µL = 1.0 × 10⁻⁵ L.

This calculation yielded the total Pt mass contained in the 10 mg of Pt–in–Ga droplets, which was then expressed as wt.% relative to the total droplet mass. This approach ensured accurate and reproducible quantification of Pt loading in the catalyst.

4.3. Electrode Fabrication

Different electrodes were prepared using Pt‐in‐Ga nanodroplets having varying concentrations of Pt. Different wt.% of droplets, i.e., 1, 5, and 10 wt.%, were mixed with W nanopowder. The 2.7 g of W nano powder and 0.3 g of Pt‐in‐Ga droplets were mixed using 400 µL of isopropanol as a solvent to make a paste. The paste was then slurry‐coated onto a Mo mesh as a thin film using a glass slide. The material was then dried in a vacuum oven at 120°C for 2 h. After drying, the material was sintered in a tube furnace at 850°C with a heating rate of 5°C min⁻¹ for 2 h. under the continuous flow of Ar at a flow rate of 0.1 L min⁻¹.

4.4. Catalyst Characterization

SEM was conducted using the FEI Verios 460L XHR‐SEM at a pressure of 5 × 10⁻⁶ mbar, an acceleration voltage of 3 kV, a current of 100 pA, and approximately 12 000 magnification. Magnification varied slightly with focus, but the scale remained standard. SEM‐EDS images and data were collected with an Oxford Instruments Xmax EDS detector, and the data were processed with Aztec software. TEM analysis was conducted on a JEOL microscope at 200 kV. EDS was performed using Aztec software with an Oxford X‐Maxn 80T spectrometer. XPS analysis was conducted with a Thermo Scientific K‐alpha spectrometer using a monochromatic Al K‐alpha source (hv ≈ 1486.6 eV) and a concentric hemispherical analyzer. For XPS analysis, the samples were scratched off from the Mo mesh in ethanol using a bath sonicator and drop‐casted onto a silicon wafer. Samples were etched using Ar+ for 100 s using 3 keV at high current setting to remove reactants and electrolytes, allowing catalyst characterization. Data was analyzed using Casa‐XPS software (version 2.3.25PR1.0) and plotted with Origin Pro 2023/2024. DSC analysis was conducted with a DSC 8000 calorimeter (TA Instruments) using ≈10 mg samples in nitrogen. A heating rate of 10°C min⁻¹ was applied from −20 to 350°C, with data presented from the second run.

¹H and ¹³C NMR analyses were performed on a Bruker Avance III 300 MHz spectrometer using 540 µL electrolytes mixed with 60 µL D₂O. ICP‐MS analysis of the electrolytes was conducted on an Agilent 7700 ICP‐MS after HER and CO₂RR. An electrolyte sample of 20 µL was diluted with 2% nitric acid before analysis. Raman spectroscopy was conducted with a 532 nm laser using a HORIBA HR Evolution (12912a). The gaseous products were analyzed using headspace gas chromatography. FTIR was done using Perkin‐Elmer's Frontier FTIR/FIR Spectrometer. The spectra of the electrolytes were obtained in the range of 4000–400 cm⁻¹ with a resolution of 4 cm⁻¹ and an average of 64 scans.

4.5. Electrochemical Reduction of CO₂

Catalytic activities of different electrodes prepared by varying concentrations of Pt for electrochemical reduction of CO₂ were tested in an aqueous electrolyte, using W/Pt‐in‐Ga (0.045 wt.% Pt) as the working electrode, Pt mesh as the counter electrode, and Ag/AgCl as a reference electrode. LSV analysis was performed by using a Vertex C.EIS potentiostat (350 mA/13 V/1 MHz). CO₂RR experiments were conducted following the same methodology as reported previously [53, 54]. LSV tests were conducted at room temperature in a potential range of −1.05 to −1.75 V vs. Ag/AgCl at a scan rate of 10 mV s⁻¹ in an aqueous electrolyte having 2 M water, 0.1 M TBAPF_6, and DMF using prepared electrodes (1 cm²). Prior to LSV scans, CO₂ was bubbled in the electrolytes for 30 min, and a continuous flow of CO₂ (0.1 L min⁻¹) was maintained during electrochemical reduction. To exclude contributions from the electrode substrate or support materials, a series of control experiments was conducted in N₂‐saturated electrolyte. Bare Mo mesh, Mo mesh coated with W powder, Mo mesh coated with W powder and Ga droplets, as well as W powder mixed with Pt‐in–Ga droplets, were systematically tested under identical electrochemical conditions. In all cases, negligible current densities were observed, and no detectable products were formed, indicating that these materials are inactive toward CO₂RR in the absence of CO₂. In contrast, when the electrolyte was saturated with CO₂, a clear increase in current density was observed only for the Pt–Pt‐containing catalyst, accompanied by the formation of CO₂ reduction products. These findings confirm that the catalytic activity originates exclusively from Pt in the Pt‐in–Ga droplets. Catalytic activities of the electrodes were tested for HER using 0.5 M H₂SO₄ as an electrolyte in a three‐electrode cell setup with W/Pt‐in‐Ga (0.35 wt.% Pt) as the working electrode, Pt mesh as the counter electrode, and Ag/AgCl as a reference electrode. LSVs were carried out in a potential range of 0.1 to −0.45 V vs. Ag/AgCl at a scan rate of 5 mV s⁻¹ at room temperature. The electrolyte was bubbled with N₂ for 30 min to remove the dissolved oxygen before electrochemical testing. For HER and CO₂RR, ten consecutive LSV scans were conducted to stabilize the electrode surface. The data presented corresponds to the final scan, where a stable current response was observed.

The tafel slope was calculated using the following equation:

$$\eta =a+b\log \left(j\right)$$ (2)

where η is the overpotential, j is the current density, a is a constant, and b is the Tafel slope.

Unless otherwise specified, all LSV results in this study are reported vs. the RHE, using the following conversion equation:

$$E_{RHE}=E_{AG}{/}_{AgCl}+0.224+0.059\left(pH\right)$$ (3)

4.6. Mechanical Stability

The mechanical robustness of the W/Pt‐in‐Ga electrode under physical deformation was tested by conducting bending tests by wrapping the electrode around cylindrical mandrels of varying diameters, i.e., 1.6 and 0.9 cm [55]. The electrode was subjected to 5, 25, 50, and 100 bending cycles for each bending diameter. After each bending test, the electrocatalytic activity of the electrode for HER was evaluated to assess any performance degradation associated with mechanical stress. This procedure was designed to simulate mechanical stress conditions and evaluate the adhesion, integrity, and overall mechanical stability of the Pt‐in‐Ga droplets in the W matrix.

4.7. Statistical Analysis

Electrochemical measurements were performed using five independently prepared electrodes for each reaction (HER and CO₂RR). For each electrode, LSV was repeated for ten consecutive cycles, and the final stabilized cycle was used for data analysis. The LSV responses obtained from different electrodes showed negligible variation; therefore, representative curves and values from the final cycle are reported. No data transformation or normalization was applied beyond standard geometric area normalization for current density and Pt mass normalization for mass activity. No outliers were excluded. Statistical hypothesis testing was not applied unless explicitly stated, as the study primarily focuses on reproducibility and performance trends.

4.8. Molecular Dynamics Simulations

All AIMD simulations and geometry optimizations were performed using the Vienna ab initio Simulation Package (VASP) [56, 57]. Initial equilibrated systems with a vacuum spacer of 15 Å in the z dimension and the following compositions: 500 Ga atoms, 499 Ga atoms and 1 Pt atom, and 498 Ga atoms and 2 Pt atoms, were taken from our previous work [58]. For HER simulations, 6 H atoms were randomly added to the surface of each system, while for CO₂RR simulations, 1 CO₂ molecule was added above the surface of each system. The systems were equilibrated using AIMD simulations for at least 2 ps with a 1 fs timestep at 303.15 K in the particles (N), volume (V), and temperature (T) (NVT) ensemble with a Nose–Hoover thermostat, projector‐augmented wave (PAW) [59] method, the Perdew‐Burke‐Ernzerhof (PBE) exchange‐correlation functional [60], an energy cutoff of 500 eV, and the gamma point only for the k‐point grid. Geometry optimizations were performed on equilibrated AIMD snapshots with a 4 × 4 × 1 k‐point grid. Visualization was performed using Visual Molecular Dynamics (VMD) 1.9.3.[61].

Author Contributions

Muhammad Hamza Nazir: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Imtisal Zahid: Formal analysis, Data curation. Caiden J. Parker: Formal analysis, Data curation. Ali Zavabeti: Visualization, Conceptualization. Karma Zuraiqi: Review & editing. Vaishnavi Krishnamurthi: Formal analysis, Data curation. Andrew J. Christofferson: Writing – original draft, Visualization, Validation, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Nastaran Meftahi: Writing – review & editing, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Kourosh Kalantar‐Zadeh: Writing – review & editing, Visualization, Validation, Conceptualization. Ken Chiang: Writing – review & editing, Supervision, Resources, Project administration, Conceptualization. Dan Yang: Review & editing, Supervision, Conceptualization, Visualization, Validation. Torben Daeneke: Writing – review & editing, Visualization, Validation, Supervision, Resources, Project administration, Conceptualization.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll72404‐sup‐0001‐SuppMat.docx

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.