Multiscale Analysis of Electrochemical Dealloying of Bimetallic Nanoparticles to Tune Catalytic Activity
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | ACS Applied Materials & Interfaces |
| Published | 2026-04-13 |
| Authors | Johanna Angona, Dimitrios Valavanis, Daniel Houghton, Fengli Yang, Thais Schroeder Rossi |
| DOI | 10.1021/acsami.6c00288 |
Electrochemical dealloying increased the mass-normalized HER current of Ag60Au40 nanoparticles by about 35 times after 10.5 cycles on a glassy carbon macroelectrode in 1.0 M HClO4. Local SECCM treatment found the highest average activity after 0.5 cycle. Identical-location STEM and EDX tracked more than 70 entities through 100.5 cycles and showed shrinking and coalescence. Macroelectrode cycling also recorded catalyst deactivation.
Scale changes the dealloying optimum
Section titled “Scale changes the dealloying optimum”Macroelectrode testing placed the optimum at 10.5 cycles, whereas local SECCM testing placed it at 0.5 cycle. The cycle counts are not interchangeable because the SECCM droplet provides faster mass transport than the beaker cell.
| Configuration | Dealloying condition | Measurement readout | Main result |
|---|---|---|---|
| Glassy carbon macroelectrode | 1.0 M HClO4, 10.5 cycles | -1.1 V vs MSE sat., 2 mV s⁻¹ in 0.5 M H2SO4 | About 35-fold mass-specific current increase |
| SECCM, 400 nm tip | 1.0 M HClO4, 0.5 cycle | -0.65 V vs RHE, 0.5 V s⁻¹ in 50 mM HCl | -6 ± 3 to -12 ± 4 A mg⁻¹ |
| IL-STEM/EDX on BDD | 0 to 100.5 cycles | Ag/Au EDX and particle diameter | Ag: 60 ± 9 to 18 ± 9 at.%; diameter: 11 ± 3 to 6 ± 2 nm |
Rurainsky et al. reported about 50 times higher catalytic current density for Ag80Au20 at -0.3 V vs RHE after potentiodynamic dealloying. That study used a different composition and potential, so it is a literature reference rather than a direct match for Ag60Au40 at -1.1 V vs MSE sat.
Macroelectrode protocol
Section titled “Macroelectrode protocol”Reverse-micelle electrodeposition produced homogeneous Ag60Au40 alloy particles directly on glassy carbon. HAuCl4 and AgNO3 were confined in PS-P2VP micelles in toluene. Deposition was held at -1.3 V vs Pt for 30 min. The particles had an average diameter of 11 ± 3 nm and a composition of about 60% Ag and 40% Au. Direct electrodeposition ensured electrical contact with the substrate.
Dealloying used cyclic voltammetry in 1.0 M HClO4 from 0 to 1.25 V vs Ag|AgCl at 1 V s⁻¹. Each treatment ended at the anodic limit to prevent Ag redeposition. HER activity was measured by linear sweep voltammetry in 0.5 M deaerated H2SO4 at 2 mV s⁻¹. The reported comparison used current at -1.1 V vs MSE sat. Mass was estimated from Ag and Au stripping peaks in 0.1 M HCl at 25 mV s⁻¹.
The as-deposited Ag60Au40 particles had poorer HER mass activity than monometallic Au particles made from the same reverse micelles. After 10.5 dealloying cycles, the mass-specific HER current was about 35 times its as-deposited value. Extending treatment to 500.5 cycles left only about 27% of the 10.5-cycle activity. Surface-area normalization produced the same optimum, so the activity trend did not depend on the chosen normalization.
SECCM accelerates local dealloying
Section titled “SECCM accelerates local dealloying”Local treatment used a 65 µm HClO4-filled SECCM pipet to create four areas on glassy carbon. The areas received 0.5, 1.5, 10.5, or 100.5 cycles at 1 V s⁻¹. A 400 nm pipet filled with 50 mM HCl then collected data at 625 positions across an area of about 500 µm per side under Ar. More than 20 measurements covered each treated area, while about 400 covered untreated surroundings. SEM and footprint controls indicated that each landing contacted about 5 to 10 entities.
Each measurement included three HER cycles between -0.5 and -1 V at 0.5 V s⁻¹, followed by anodic stripping. Activity was taken from the third cycle at -0.65 V vs RHE. Ag and Au peak charges supplied relative mass and composition estimates. The lower electrolyte concentration and faster scans prevented direct comparison with the macroelectrode values.
After 0.5 cycle, the SECCM estimate fell to about 8 ± 5% Ag, with no significant further Ag loss. Random-point EDX confirmed the low Ag content. Mean mass-related current changed from -6 ± 3 A mg⁻¹ to -12 ± 4 A mg⁻¹. The authors attributed the accelerated removal to faster Ag+ ion flux in the micropipet.
Statistical distributions expose heterogeneity
Section titled “Statistical distributions expose heterogeneity”The SECCM distributions showed that ensemble averages hide broad particle-to-particle variation. The as-synthesized Ag content had a standard deviation of 11.6 at.%, while the dealloyed Ag distributions were narrower. Activity variation instead increased slightly after 0.5 cycle.
Lower Ag content tended to accompany higher activity, consistent with Au being more active than Ag for HER. The broad populations showed that composition was not the only property controlling activity. After 100.5 cycles, average activity decreased and its distribution shifted. The difference between 0.5 and 100.5 cycles was not statistically significant by the Mann-Whitney U test.
SECCM also compared separate particle sets at each treatment state. It therefore could not assign a continuous dealloying history to an individual particle.
Identical-location imaging tracks structural change
Section titled “Identical-location imaging tracks structural change”Identical-location analysis used an electron-beam-transparent BDD electrode to image the same entities after 0, 0.5, 10.5, and 100.5 cycles. STEM operated at 200 kV, and EDX quantified Ag and Au through their Lα1 lines. More than 70 entities were followed through the treatment sequence.
Primary particles stayed spherical or oval but shrank from 11 ± 3 nm to 6 ± 2 nm after 100.5 cycles. This was a 43 ± 13% size reduction. Mean Ag content changed from 60 ± 9 at.% initially to 38 ± 11 at.% after 0.5 cycle and 18 ± 9 at.% after 100.5 cycles. The first cycle produced an average relative Ag loss of 40 ± 16%. Further cycling caused Ag loss, but the loss rate per cycle decreased. The authors explain this trend by reorganization of Au surface atoms, which stabilizes the remaining Ag.
For one counted set of 25 primary particles, 23 remained after both 0.5 and 10.5 cycles. Only 18 remained after 100.5 cycles. The missing entities may have dissolved or detached. Aggregates coalesced, with the largest change occurring between 10.5 and 100.5 cycles. No pores were resolved at the magnifications used, and EDX still showed a homogeneous alloy structure throughout the treatment.
Deactivation mechanisms remain partly unresolved
Section titled “Deactivation mechanisms remain partly unresolved”The measured particle shrinkage exceeded the change predicted from Ag removal alone. A calculation assuming only Ag loss gave about 9 nm at 100.5 cycles, compared with 6 ± 2 nm from STEM. The authors therefore identified Au dissolution as the likely dominant contributor. A density increase alone could not explain the difference, but crystal structure and surface orientation were not measured. This conclusion therefore remains an inference.
Coalescence may reduce the surface-area-to-volume ratio and contribute to deactivation, while shrinking tends to increase that ratio. Au passivation was also linked to the activity loss. The evidence has limits. SECCM currents were lower after dealloying, but post-measurement SEM showed no extensive coverage change, leaving some particle loss unresolved. Pore absence was limited by microscopy resolution, and detailed composition histories covered seven particles and five aggregates.
Implications for catalyst development
Section titled “Implications for catalyst development”The results support distribution-based catalyst development rather than optimization against one ensemble value. Local screening can identify the first-cycle composition and activity distribution quickly. The optimum cycle count must still be recalibrated for each cell because mass transport changes Ag removal. Identical-location microscopy can then determine whether structural changes accompany the activity changes.
The paper demonstrates this workflow for Ag60Au40 nanoparticles. Its extension to other bi- and multimetallic catalyst systems remains a proposal rather than a demonstrated result.
Materials and support
Section titled “Materials and support”The material requirement here is a boron-doped diamond electrode for identical-location STEM/EDX during repeated electrochemical dealloying of Ag60Au40 nanoparticles. The paper needs an electron-beam-transparent BDD support with electrochemical and mechanical stability and a patterned central hole. The 6CCVD diamond materials catalog lists boron-doped diamond, and the engineering support link covers custom fabrication support for that electrode geometry.
View Original Abstract
High Resolution Image Download MS PowerPoint Slide Electrochemical dealloying is a promising technique to tune the activity and material utilization of electrocatalysts. Studying this process in bimetallic nanoparticles remains challenging since their synthesis yields ensembles with varied properties. To understand the dealloying behavior of individual nanoparticles, we present a combined characterization method that provides statistical distributions in composition, morphology, and catalytic activity. Ag x Au y alloy nanoparticles synthesized via reverse micelles were dealloyed, by potential cycling, to tune their activity for the hydrogen evolution reaction. In a multiscale approach, we performed macroscale measurements on a glassy carbon electrode and scanning electrochemical cell microscopy (SECCM) on small groups of particles/aggregates within the confines of the scanned droplet. The SECCM pipet was used to create different states of dealloying within a single sample, followed by higher-resolution SECCM activity mapping. The faster mass transport in the SECCM tip, compared to the macroelectrode configuration, led to accelerated Ag dissolution from the Ag x Au y nanoparticles. While the particles showed the highest activity and largest Ag content decrease after only 0.5 cycles, heterogeneous behavior was still evident among individual particle groups. Further, using a boron-doped diamond electrode, identical-location scanning transmission electron microscopy and energy-dispersive X-ray spectroscopy studies allowed Ag leaching and particle shrinking to be followed at the single-particle level, as a function of cycle number, for a statistically relevant sample number ( n > 70). Increased cycling induced particle deactivation, shrinking, and coalescence. Overall, our results showcase an efficient and versatile route to combine systematic tuning and high-throughput screening of nanocatalyst property-activity relationships.
Boron Doped Diamond On Titanium Bdd Electrode
View Specs →