Design rules for catalysis in single-particle plasmonic nanogap reactors with precisely aligned molecular monolayers

Plasmonic gaps that confine light at the nanoscale enable direct study of catalytic reactions with high sensitivity, but achieving controlled and efficient reaction kinetics remains difficult when using common plasmonic metals (Au, Ag) that are chemically inert. The antenna–reactor strategy combines plasmonic and catalytic metals, yet imprecision of active sites has hindered mechanistic understanding. This study asks how precisely constructed nanogap reactors with atomic monolayer palladium (Pd) coatings on selected Au facets influence kinetics and pathways of the Suzuki–Miyaura C–C cross-coupling of surface-bound reactants. The purpose is to decouple and quantify the roles of facet identity and molecular alignment within single, well-defined NP-on-mirror nanogaps, leveraging strong near fields for both driving and in situ SERS tracking. The work aims to establish design rules that relate nanogap configuration, catalytic surface placement (NP vs mirror), and mass transport to reaction efficiency and selectivity.

Prior work has shown plasmonic gaps enable sub-diffraction confinement and hot carriers that can promote photocatalysis, with Au/Ag widely used for field enhancement but limited intrinsic catalytic activity. Antenna–reactor heterostructures integrating catalytic metals (e.g., Pd, Pt, Cu) have improved activity, though active surface sites and molecular configurations were often ill-defined. Atomic monolayer coatings on Au via chemical reduction or underpotential deposition (UPD) are established and preserve plasmonic resonances while altering surface chemistry, unlike thicker (>1 nm) catalytic layers that damp optics. The Suzuki–Miyaura reaction is a prototypical Pd-catalyzed C–C coupling; previous plasmonic studies focused on boosting efficiency using various assemblies but provided limited mechanistic insight into how surface orientation, molecular alignment, and gap environment control pathways. The present work addresses these gaps by employing precision NPOM constructs aligning molecular monolayers and catalytic monolayers on specific facets.

Nanogap reactor platform: A nanoparticle-on-mirror (NPOM) construct is used. A flat template-stripped Au mirror is functionalized with a self-assembled monolayer (SAM) of 4-bromothiophenol (4-BTP), forming the immobilized reactant layer. Au nanoparticles (NPs, 80 nm diameter) serve as the upper facet. Four nanogap reactor (NR) types are assembled by combining Au or Au@Pd (monolayer Pd-coated) NPs with Au or Au@Pd mirrors: Au-on-Au, Au-on-Pd, Pd-on-Au, Pd-on-Pd. The SAM is sandwiched between the NP and mirror forming a sub-nm gap hotspot.

Pd monolayer deposition: Pd monolayers on Au NPs are made by chemical reduction of Pd2+ (H2PdCl4) onto citrate-capped Au NPs at low temperature using ascorbic acid, following iterative additions to achieve atomically thin conformal growth (core–shell Au@Pd). Homogeneous Pd coverage is verified by TEM/EDS and cyclic voltammetry (distinct monolayer vs multilayer oxidation features). Pd monolayers on Au mirrors are prepared by electrochemical underpotential deposition (UPD) in 0.1 M H2SO4 + 0.1 mM H2PdCl4, sweeping 0.2 to 0.02 V vs Pd at 1 mV s−1; a reduction peak around 0.05 V indicates monolayer deposition. The integrated charge corresponds to one Pd(111) monolayer.

Nanogap assembly: Au (or Au@Pd) mirrors are incubated in 2 mM 4-BTP ethanol solution for 16 h, rinsed, and dried. Cleaned Au or Au@Pd NP solutions are drop-cast onto the SAM with 10 mM NaNO3 to promote adsorption, followed by rinsing and drying, yielding single NPOMs with the four surface combinations.

Control aggregates (MLaggs): Monolayer aggregates of Au or Au@Pd NPs are formed at a water–chloroform interface by adding 4-BTP and NaNO3 to induce aggregation, then transferring the interfacial NP monolayer onto glass and rinsing. These MLaggs contain many nanogaps within the laser spot and preserve sub-nm gaps.

Characterization: Dark-field microscopy and white-light scattering determine coupled mode wavelengths; small redshifts (2–14 nm) occur upon Pd monolayer addition. FDTD simulations confirm negligible optical impact of sub-nm Pd layers. SERS fingerprints of reactant (4-BTP) and product (4-cyanobiphenylthiol, NC-BPT) are recorded and compared with DFT-simulated Raman spectra.

Reaction and SERS measurements: Reactions are initiated in a flow cell under aqueous solutions of 4-cyanophenylboronic acid (CPBA, 6.7 mM) and K2CO3 (100 mM, hole scavenger) mixed 1:1. A 633 nm laser (typ. 100 µW unless varied) excites and probes SERS via a 0.5 µm spot through a 0.9 NA objective. Time-series SERS are collected (300 frames, 1 s integration) to monitor kinetics. At higher laser powers (>0.2 mW), polymerization side products can appear. Excitation at 785 nm is also tested as a control.

Spectral analysis and kinetics: Key SERS modes are assigned: for 4-BTP, peaks at 502, 1037, 1532 cm−1; for NC-BPT, 1246, 1550, 1572, 2255 cm−1 (C≡N). Peaks are Lorentzian-fitted; intensity ratios R21 = I(1572)/I(1532) and R31 = I(2255)/I(1532) track product formation and CPBA diffusion to the top facet. A two-component kinetic model is fit: early-time exponential (fast surface-driven component) and late-time linear growth (diffusion-limited component), extracting initial and final relative yields and slow rate constants (k_low). Power dependence is measured. Diffusion rates are estimated by counting molecules from SERS enhancement assumptions, yielding an inferred CPBA influx of about 2–3 molecules s−1 for Pd-on-Au.

Computations: DFT Raman spectra are computed with Gaussian 09 (PBE, 6-311++G(d,p); Los Alamos ECP for Au), and broadened to compare with experiment. Electronic structure and projected density of states (PDOS) for Au(111) and Pd monolayer on Au are calculated with VASP (PBE, PAW, 230 eV cutoff), using slab models to assess electron-donating character of Pd ML on Au near EF and inform gap dipole considerations.

Instrumentation: Olympus BX51 microscope, 633 nm diode laser, Triax 320 spectrometer with Andor CCD; dark-field condenser for locating NPOMs; custom flow cell (~300 µm liquid layer). SEM and TEM/EDS characterize structures.

• Optical resonances: Adding a single-atom Pd monolayer to the NP or mirror redshifts the NPOM coupled mode by only 2–14 nm; plasmonic properties remain Au-like because the field penetrates ~5 nm into metal while the Pd layer is <1 nm.
• SERS validation: Clear fingerprints of 4-BTP (502, 1037, 1532 cm−1) and NC-BPT (1550, 1572, 2255 cm−1) match DFT. Slight broadening near 1532 and 2255 cm−1 with Pd surfaces is consistent with Pd–adsorbate interactions.
• Pd is necessary: Au-on-Au NRs show no coupling product under 633 nm; all Pd-containing NRs produce NC-BPT, confirming Pd-catalyzed Suzuki–Miyaura coupling in the nanogap.
• Immediate product formation: Initial product yield at t=1 s is significant when Pd is present and is roughly doubled when both facets are Pd (Pd-on-Pd), indicating sub-second, surface-mediated steps enabled by precise Br–Pd alignment.
• Distinct kinetics by facet placement: R21 dynamics show two regimes. With Au mirrors (Pd-on-Au), R21 grows nonlinearly then transitions to slower linear growth after ~2 min. With Pd mirrors (Au-on-Pd, Pd-on-Pd), growth is linear from the start, implying different balances between immediate surface conversion and diffusion-limited steps.
• Diffusion tracking via CN mode: R31 (2255/1532) increases linearly across all NRs with rate ranking Pd-on-Au > Pd-on-Pd > Au-on-Pd ≈ Au-on-Au, directly probing CPBA arrival at the top facet.
• Superior performance of single-particle NPOMs vs MLaggs: Monolayer NP aggregates show >10-fold lower catalytic activity and slower kinetics than NPOMs, with both R21 and R31 saturating after ~20 min and much smaller t=0 product peaks, indicating higher barriers and less favorable molecular alignment in aggregates.
• Power and wavelength dependence: Higher laser power increases both rates and yields; 785 nm excitation produces almost no reaction compared to 633 nm, indicating efficiency increases with photon energy rather than direct dependence on the exact coupled plasmon wavelength.
• Mechanistic insights: Pd on the NP enhances oxidative addition via direct Br–Pd contact in the gap; Pd on the mirror polarizes the gap and promotes diffusion of CPBA and base (CO3 2−). Gap dipole considerations (including Au–Pd electronegativity differences) rationalize mass-transport control of the slower, 0th-order regime.
• Quantified transport: For Pd-on-Au, ~2–3 CPBA molecules s−1 diffuse into the gap, leading to an estimated ~24 min to consume all reactants in the hotspot.
• Side reactions threshold: At >0.2 mW laser power, broad C–C adduct bands appear due to polymerization; at low power no side reactions or photodegradation are detected over 5 min.

The study demonstrates that precise nanogap construction and molecular alignment fundamentally alter Pd-catalyzed Suzuki–Miyaura coupling kinetics under plasmonic excitation. Pd must be present for catalysis; Au-on-Au gaps with identical plasmonic enhancement do not react, indicating hot carriers or heating alone are insufficient without appropriate catalytic surface chemistry. The immediate product signal at t=1 s in Pd-containing NRs shows that aligning the 4-BTP bromine directly beneath a Pd site enables rapid oxidative addition and transmetallation. The subsequent linear regime reflects diffusion-limited supply of CPBA and base into the gap, with rates modulated by the effective nanogap dipole. Pd on the mirror increases electron donation relative to Au, which, together with molecular dipole orientation, modulates the net gap polarization, facilitating or hindering diffusion. This framework explains why Pd-on-Au yields the fastest R31 growth (maximized gap dipole) and why Pd-on-Pd exhibits high initial conversion but broader product signatures and different subsequent evolution. Comparison to MLaggs highlights that random molecular orientation and dual-sided thiol binding suppress both initial yields and slow rates, emphasizing the unique benefit of single-particle NPOM reactors for mechanistic dissection. Overall, the results clarify how surface identity (NP vs mirror), molecular alignment, and gap polarization collectively determine photocatalytic efficiency beyond generic plasmonic effects.

The work establishes design rules for plasmonic nanogap photocatalysis by integrating atomic Pd monolayers into single-particle NP-on-mirror reactors with precisely aligned 4-BTP monolayers. Key contributions include: (1) decoupling catalytic roles of Pd when placed on the NP versus the mirror, (2) identifying a two-step kinetic pathway with an ultrafast surface-driven step followed by diffusion-limited growth, (3) demonstrating that direct Br–Pd contact at the NP facet accelerates oxidative addition while Pd on the mirror modulates gap polarization to enhance mass transport, and (4) showing that precision NPOMs vastly outperform nanoparticle aggregates. Future work should target quantitative mechanistic separation of hot-carrier versus thermal contributions, explore other catalytic monolayers and reactants, engineer gap dipoles via materials and molecular design, and translate single-gap insights to metasurface architectures for scalable catalysis and electrocatalysis.

• The precise mechanism of photocatalytic enhancement (hot carriers versus thermal or mixed effects) is not fully resolved; further controlled studies are needed.
• Kinetic analysis relies on SERS intensity ratios and a two-component model that yields relative, not absolute, rates and conversions.
• Spectral overlap in certain configurations (e.g., Pd-on-Pd near 1530–1570 cm−1) complicates peak deconvolution and can affect relative trend comparisons.
• High laser powers (>0.2 mW) induce polymerization side reactions, limiting the usable operating window.
• Wavelength dependence shows minimal activity at 785 nm under the tested conditions; broader spectral mapping would help generalize the findings.