Nitrate-to-ammonia conversion with a plasmonic antenna–reactor catalyst

Abstract

Electrochemical conversion of nitrate to ammonia is an appealing route to efficiently synthesize ammonia under ambient conditions while reducing environmental nitrate pollutants. However, this approach is obstructed by the limited yield and selectivity of ammonia because the electrochemical nitrate-to-ammonia conversion involves multi-electron/proton transfer and faces competition from the hydrogen evolution reaction. Here, we demonstrate a plasmon-assisted strategy to improve the performance of nitrate-to-ammonia electrochemical conversion by constructing plasmonic antenna–reactor catalysts, where Au and Pd nanoparticles/hydrogen substituted graphdiyne (Pd/HsGDY) work as the light antenna and reaction site, respectively. Plasmonic excitation of Au–Pd/HsGDY catalysts can remarkably accelerate the nitrate reduction, with the yield rate, selectivity, and Faradaic efficiency of ammonia respectively increased by 14.3, 2.1, and 1.8 times under optimal conditions. Mechanistic investigations unveil that Au plasmon-induced hot electrons facilitate nitrate-to-ammonia reaction by regulating the adsorption of reaction intermediates on Pd/HsGDY, wherein the rate-determining step was shifted from nitrate adsorption to *NH protonation and the overall apparent activation was reduced. Moreover, hot electrons suppress the competing hydrogen evolution by enlarging the Gibbs free energy of hydrogen formation. These results open a way to develop desirable catalysts for producing value-added ammonia from environmentally hazardous nitrate by a synergistic combination of electricity and light.

---

Broader context

Converting nitrate (NO₃⁻) to ammonia (NH₃) has recently garnered significant attention as a one-stone-two-bird technology—offering an appealing alternative to the traditional energy-intensive Haber–Bosch process for NH₃ production while simultaneously mitigating NO₃⁻ contamination from industrial and agricultural wastewater. However, the challenge lies in manipulating reaction pathways to enhance the NH₃ yield and selectivity, due to complex reaction kinetics and the competitive hydrogen evolution reaction. This work presents a proof-of-concept demonstration of improving electrochemical NO₃⁻-to-NH₃ conversion using surface plasmons—collective oscillations of free electrons in noble metal nanostructures in response to incident light. In this plasmonic antenna–reactor catalyst, plasmon-induced hot electrons shift the rate-determining step from NO₃⁻ adsorption to *NH protonation by regulating the adsorption of reaction intermediates, and suppress the side reaction by raising the Gibbs free energy of hydrogen formation. As a result, the yield rate, selectivity, and faradaic efficiency of NH₃ are increased by 14.3, 2.1, and 1.8 times under optimal plasmonic excitation.

1. Introduction

Ammonia (NH₃) is an essential feedstock used for artificial fertilizers and various chemicals, with a global annual production of over 182 million tons.¹ It is also considered a promising candidate as a next-generation carbon-free energy carrier due to its high volumetric density of hydrogen, ease of liquefaction, and compatibility with the existing pipeline transport infrastructure.² More than 90% of NH₃ is produced from nitrogen (N₂) and hydrogen (H₂) by the Haber–Bosch process.^1,3 It operates under high temperatures (350 °C to 450 °C) and high pressure (100–200 bar), and the H₂ gas is industrially derived from steam reforming of methane, which generates 5.5 metric tons of carbon dioxide (CO₂) per metric ton of H₂ produced.⁴ Consequently, this process consumes approximately 2% of the world's annual energy output and is responsible for about 1.2% of greenhouse gas emissions,^1,5 highlighting the urgent need for more sustainable and efficient alternatives.

As an alternative, the electrochemical synthesis of NH₃ has been drawing increasing interest due to the various inherent advantages of electrocatalysis, such as the distributed production approach, mild conditions, and zero carbon footprint.^3,6,7 Direct electrochemical reduction of N₂ to NH₃ in aqueous solutions under an ambient atmosphere seems to provide tantalizing opportunities for producing NH₃ economically and environmentally benignly.⁸ This process can even be compatible with green solar cells.⁹ However, the strong N≡N triple bond (941 kJ mol⁻¹) and extremely limited N₂ solubility in water inevitably lead to the low yield rate of this pathway, 2–3 orders of magnitude lower than the Haber–Bosch process.¹⁰ In the search for desirable N-containing raw materials used for the electrochemical synthesis of NH₃, the nitrate anion (NO₃⁻) stands out because of the relatively low dissociation energy of the N=O bond (204 kJ mol⁻¹) and excellent solubility in aqueous solution.^11,12 Besides, NO₃⁻ ubiquitously exists in industrial wastewater (∼2000 ppm), surface water and groundwater (∼50 ppm) as a toxic contaminant caused by anthropogenic use in soil and crop fertilization.^13–16 Therefore, electrochemical conversion of NO₃⁻ to NH₃ has the potential to achieve efficient and green production of NH₃, and simultaneously to help address environmental pollution.

To this end, various strategies have been adopted to develop efficient electrocatalysts for NO₃⁻-to-NH₃ conversion, including introduction of defect sites and doping elements,^17–19 construction of single atom reaction sites,^20–23 regulation of reaction microenvironments and active hydrogen,^24–27 design of tandem catalysis,^5,16,28,29 enrichment of nitrate,^30–32 machine learning,³³ and so on.^34–36 These well-designed attempts converge on the fact that the appropriate adsorption behaviours of N-containing reaction intermediates play a pivotal role in efficiently and preferentially generating NH₃ from NO₃⁻ electrochemical reduction, which involves complex multi-steps of electron/proton transfer. Another key lesson is to suppress the undesired consumption of electron donors by H₂ generation.³⁷

Excited plasmons, which originated from the collective oscillation of free electrons in optically excited coinage metal nanostructures, can profoundly affect intermediates' adsorption/desorption behaviour,^38–40 thus changing the reaction rate and product distribution. For instance, plasmonic excitation of Au nanoparticles induced NH₃ synthesis with up to a 15× boost in activity relative to conventional electrocatalysis, attributed to energetic carriers and charged interfaces induced by the non-radioactive decay of Au plasmons.⁴¹ Unfortunately, its optimal yield rate of NH₃ is only ∼55 μg h⁻¹ cm⁻¹, significantly inferior to the reported values. This undesired result was primarily caused by the intrinsic inertness of Au toward NO₃⁻ reduction, where the Au nanoparticles served the dual role of plasmonic absorption and catalytic site. Plasmonic antenna–reactor catalysts offered the perfect solution to this issue by integrating plasmonic metals of strong light harvesting capability with catalytic elements of high activity for specific reactions in one entity.³⁸

In this vein, for the first time, we demonstrated the plasmon-assisted electrochemical conversion of NO₃⁻ to NH₃ by constructing plasmonic antenna–reactor (Au–Pd/HsGDY) catalysts. The results showed that the electrochemical reduction of NO₃⁻ on Pd/HsGDY catalysts was substantially accelerated by photoexcitation of Au plasmons. The yield rate of NH₃ was boosted by up to 14.3 times and could reach 7.6 mg h⁻¹ cm⁻² under light illumination, while its selectivity and Faradaic efficiency (FE) of NH₃ were approximately doubled. The mechanistic studies revealed that the hot electrons generated from Au plasmon relaxations optimized the adsorption energy of reaction intermediates in both NO₃⁻ reduction and H₂ formation pathways, contributing to the efficient and selective reduction of NO₃⁻ to NH₃ and simultaneously suppressing the competing hydrogen evolution side reaction. This work provides a novel way to design a superior catalyst for NO₃⁻-to-NH₃ conversion by combining electric and light energy, which is vitally important for achieving the green production of NH₃ and alleviating the environmental threat of NO₃⁻ pollution.

2. Results and discussion

A schematic representation of the plasmon-enhanced NO₃⁻-to-NH₃ conversion on the plasmonic antenna–reactor catalyst (Au–Pd/HsGDY) is shown in Fig. 1 The Pd nanoparticles serve as catalytic reaction sites where the electrochemical reaction of NO₃⁻ takes place. HsGDY is chosen as a support due to its previously demonstrated ability to facilitate the hydrogenation of nitrogen intermediates and desorption of NH₃ on Pd sites.⁴² The Au nanoparticles (Au NP) work as antennas for harvesting light energy from the excited surface plasmons, namely the collective oscillation of their free electrons in response to incident light irradiation. It has been well documented that the excited surface plasmons are relaxed by either re-emitting photons to free space (radiative decay) or yielding energetic charge carriers and heating up the crystal lattice (non-radiative decays).⁴³ Both energetic charge carriers and elevated temperature would have profound effects on the outcomes of the reaction in the vicinity of plasmonic metals, which are usually referred to as non-thermal and thermal effects in plasmon-mediated chemical reactions.⁴⁴ In the present study, non-thermal and thermal effects during non-radiative decays of Au plasmons are expected to modify the energetics and kinetics of the NO₃⁻ electrochemical reaction at Pd sites, thus mediating the reaction rate and product distribution. In the following experiments, we did observe that the yield rate of NH₃ was improved by a factor of 14.3, while both the FE and selectivity were more than doubled under optimal conditions.

Fig. 1 Plasmon-assisted electrochemical synthesis of NH₃. A schematic of the electrochemical reduction of NO₃⁻ to NH₃ executed at the Au–Pd/HsGDY plasmonic antenna–reactor under light illumination. The gold and midnight blue hemispheres denote Au NPs and Pd NPs on the HsGDY (hydrogen substituted graphdiyne, grey white) support, which act as plasmonic antenna and catalytic sites, respectively.

The Pd precursors were trapped into the molecular pore of HsGDY and in situ reduced to nanoparticles by a mild pyrosis method following our previously published protocol.⁴² The mean diameter of Pd nanoparticles is larger than the pore size of HsGDY (Fig. S1, ESI†), indicating that the Pd nanoparticles were preferentially anchored onto the surface of HsGDY during the pyrolysis process. The Au NPs with a mean diameter of 4.1 nm (Fig. S2, ESI†), were subsequently grown on the Pd/HsGDY using a magnetic sputtering technique (Fig. 2a–g), ensuring the direct contact between Au and Pd. Three sets of atomic lattice fingers with their interplane spacings equalling 2.25, 2.03, and 2.39 Å are clearly observed in the high-resolution TEM images, which can be indexed to Bragg reflections of the (111) lattice plane of face centred-cubic Pd (PDF# 65-2867), and the (200) and (111) lattice plane of cubic Au (PDF# 65-2870), respectively (Fig. 2h and Fig. S3, ESI†). This assignment was also supported by the selected area electron diffraction results (Fig. S4, ESI†). The atomic percentages of C, Pd, and Au are 96.79%, 1.57%, and 1.64%, respectively (Fig. S5, ESI†), consistent with the results of inductively coupled plasma mass spectrometry measurements (ICP-OES, Fig. S6, ESI†). Kindly note that the shade of elementary mapping colour cannot be precisely related to their contents, especially at such a low range, and may be affected by various factors.

Fig. 2 Au–Pd/HsGDY plasmonic antenna–reactor characterization. (a) Scanning electron microscopy and (b) transmission electron microscopy (TEM) images of Au–Pd/HsGDY. (c) High-angle annular dark-field scanning electron microscopy (HAADF-STEM) images of Au–Pd/HsGDY and (d)–(f) corresponding elementary dispersive spectroscopy (EDS) mapping of C, Au and Pd. (g) and (h) High resolution TEM images of Au–Pd/HsGDY, with the lattice plane distances of Au and Pd nanoparticles marked by gold and white lines, respectively. A schematic model of the Au–Pd configuration is shown in the inset. (i) and (l) Normalized Pd K and Au L₃ edge over the XANES profiles of Au–Pd/HsGDY and Au foil. (j) and (m) k³-weighted FT-EXAFS spectra of Pd K and Au L₃ edges in Au–Pd/HsGDY. (k)and (n) WT-EXAFS plots of Pd K and Au L₃ edges in Au–Pd/HsGDY.

The Pd K edges over the X-ray absorption near edge structure (XANES) profile of Au–Pd/HsGDY almost overlap with that of the Pd foil, with little features of PdO in certain localized regions (Fig. 2i), indicating that the valence state of Pd is predominantly zero. This is in line with the Pd 3d core-level XPS analysis of Au–Pd/HsGDY (Fig. S7, ESI†). The Fourier transformed extended X-ray absorption fine structure (FT-EXAFS) analysis was carried out to determine the short-range local coordination structure. The intense shell peak (R = 2.5 Å) can be safety assigned to the Pd–Pd scattering path (Fig. 2j), which is consistent with a previous publication.⁴⁵ The wavelet transform (WT) EXAFS plots of Au–Pd/HsGDY and Pd foil are nearly indistinguishable but quite different from the PdO reference (Fig. 2k and Fig. S8, ESI†), showing a maximum lobe at 9.8 Å⁻¹ from 2.0 to 2.9 Å, which can be indexed to Pd–Pd coordination.⁴⁶ The Au L₃ edges over the XANES profile of Au–Pd/HsGDY are identical with that of the Au foil, indicating its Au(0) state (Fig. 2l), which is consistent with the Au 4f XPS characterization (Fig. S9, ESI†). The FT-EXAFS spectrum of the Au L3 edge in Au–Pd/HsGDY is characterized by an intensive coordination peak attributable to the Au–Au scattering path (R = 2.7 Å), indicative of the Au–Au coordination (Fig. 2m). This indication is further supported by the WT-EXAFS data (Fig. 2n and Fig. S10, ESI†).^47,48 The Fourier transformed infrared spectrum of Au–Pd/HsGDY displays the characteristic absorption of alkynyl carbon bonds, benzenoid carbon bonds, and the C–H stretching vibrations of HsGDY (Fig. S11, ESI†). The presence of alkynyl and benzenoid carbon bonds was confirmed by Raman (Fig. S12, ESI†) and XPS characterization (Fig. S13, ESI†). Based on these results, it is clear that the present plasmonic antenna–reactor catalyst features aggregated irregular HsGDY spheres with the Au and Pd nanoparticles riveted on the surface.

The electrochemical reduction of NO₃⁻ at Au–Pd/HsGDY with and without plasmonic excitation was executed in our home-made set-up (Fig. S14, ESI†). With the absence of NaNO₃, the LSV curve of Au–Pd/HsGDY displays a small cathodic current intensity from −0.3 to −0.9 V versus the reversible hydrogen electrode (RHE, Fig. 3a). The larger cathodic current was observed after adding NaNO₃ into the electrolyte, demonstrating the catalytic activity of the Au–Pd/HsGDY for electrochemical reduction of NO₃⁻. The Raman bands of NO₃⁻ (1053 cm⁻¹), NH₃ (1103 cm⁻¹, 1402 cm⁻¹), and NO₂⁻ (1327 cm⁻¹) were observed in the in situ Raman spectra of the working electrode (Fig. S15, ESI†).^49,50 The LSV curves of Pd/HsGDY and Au–Pd/HsGDY are nearly identical (Fig. S16, ESI†). The amount of generated NH₃ and NO₂⁻ is very small and no N₂H₄ was detected when Au/HsGDY was used (Fig. S17–S19, ESI†). Therefore, this activity should not originate from Au NPs, consistent with Au's intrinsically low activity toward NO₃⁻ reduction.⁴¹ The bare HsGDY is chemically inert within the investigated voltage ranges (Fig. S20, ESI†). The Pd nanoparticles are thus the catalytic sites for NO₃⁻ reduction.

Fig. 3 Plasmon-enhanced NO₃⁻-to-NH₃ conversion on Au–Pd/HsGDY. (a) Linear sweep voltammetry (LSV) curves of the Au–Pd/HsGDY electrode in the electrolyte without NaNO₃ and the electrolyte containing NaNO₃ in the absence and presence of white laser illumination (135.1 W cm⁻²). (b) Potential-dependent yield rate of NH₃ for electrochemical reduction of NO₃⁻ executed on the Au–Pd/HsGDY electrode with the light switched on and off. (c) Proton nuclear magnetic resonance (¹H NMR) spectra of the electrolyte of electrochemical ¹⁴NO₃⁻ and ¹⁵NO₃⁻ reduction at −0.6 V for 30 min. (d) Potential-dependent yield rate of NO₂⁻ with the light switched on and off. (e) Plasmon-induced yield enhancement of NH₃ and NO₂⁻ at indicated electrode potentials. (f) Selectivity for NH₃ during electrochemical reduction of NO₃⁻ at the Au–Pd/HsGDY electrode with the light on and off. (g)–(i) FE of NH₃, NO₂⁻ and H₂ production for NO₃⁻ reduction with and without light illumination.

Compared to the Pd/HsGDY, the Au–Pd/HsGDY catalyst displays a broad plasmonic absorption from 400–850 nm (Fig. S21, ESI†), indicating the effective utilization of light energy. In response to the illumination of white light (135.1 W cm⁻²), the cathodic current of Au–Pd/HsGDY catalysts was remarkably enhanced, while that of Pd/HsGDY remained almost unperturbed (Fig. 3a and Fig. S22, ESI†). Electrochemical impedance spectroscopy measurement shows that the Au–Pd/HsGDY electrode features smaller charge transfer resistance (R_ct) and diffusional impedance (W) under light illumination, which is indicative of faster charge transfer kinetics and mass transfer (Fig. S23 and Table S1, ESI†), respectively. These observations suggest that photoexcitation of Au NP surface plasmons accelerates the electrochemical reduction of NO₃⁻ executed on the surface of Au–Pd/HsGDY catalysts.

To explore the effects of the Au NP plasmons on the reaction selectivity, we quantitatively analysed the yield rate of typical reaction products using colorimetric methods.⁸ In the absence of light illumination, NH₃ was generated with a yield rate of 0.025 mg h⁻¹ cm⁻² at −0.4 V. It was increased with enlarging electrode bias, reaching 6.338 mg h⁻¹ cm⁻² at −0.9 V (Fig. 3b and Fig. S24, S25, ESI†), because the larger overpotential signifies the greater driving force for the reaction according to the Butler–Volmer equation. Increasing the catalyst loading could further improve the NH₃ yield rate, which was not explored in this study as it falls beyond the scope of our research. The isotope labelling study ascertains that the nitrogen of NH₃ was sourced from NO₃⁻, where the proton nuclear magnetic resonance (¹H NMR) spectra of NH₃ produced from ¹⁴NO₃⁻ and ¹⁵NO₃⁻ electrolyte shows the triplet of ¹⁴NH₄⁺ and the doublet of ¹⁵NH₄⁺, respectively (Fig. 3c). Upon illumination, the yield rate of NH₃ was enhanced at all investigated electrode potentials. Of particular note, the NH₃ was produced at 3.160 mg h⁻¹ cm⁻² at −0.6 V, exceeding that at 2.694 mg h⁻¹ cm⁻² at −0.8 V with the light switched off. NO₂⁻, a two-electron intermediate in the NO₃⁻ electroreduction pathway, was determined as another prevalent product. The yield rate of NO₂⁻ got larger as the electrode potential become more negative and was also noticeably improved by light irradiation, e.g., from 0.130 to 0.535 mg h⁻¹ cm⁻² at −0.4 V and from 4.164 to 10.537 mg h⁻¹ cm⁻² at −0.9 V (Fig. 3d and Fig. S26, S27, ESI†). Kindly note that colorimetric detection of NO₂⁻ probably produced inaccurate results under certain situations. To check the reliability of our quantification method, we also measured the NO₂⁻ using ion chromatography and obtained identical results (Fig. S28 and S29, ESI†), which is possibly due to the relatively high concentration of NO₂⁻ in our case. The yield rate of both NH₃ and NO₂⁻ was improved by the same magnitude whether the Pt wire or graphite rod was used as the counter electrode (Fig. S30 and S31, ESI†), showing that the improvement was not affected by the anodic reaction on the counter electrode.⁵¹ It is noteworthy that N₂H₄ was undetectable in the above electrolytes (Fig. S32–S34, ESI†).

Even though the Au plasmons enhanced the yield rate of both NH₃ and NO₂⁻, their magnitudes of enhancement were disparate. To provide an intuitive indicator, the yield enhancement was calculated as the ratio of yield rate in the dark and under light irradiation (Fig. 3e). The yield enhancement of NH₃ manifests a volcano plot as a function of electrode potential, rising from 8.1 at −0.4 V to 14.3 at −0.6 V but then plummeting to 1.2 at −0.9 V, whereas that of NO₂⁻ is little influenced by electrode potential, fluctuating between 4.1 and 2.5. As a consequence, the product distribution of the electrochemical reduction of NO₃⁻ was significantly changed by the impingent of light. The selectivity for NH₃, calculated as the molar percentage of NH₃ in the N-containing products, is below 50% with an electrode potential of −0.4 and −0.5 V, and above 51% at the more negative potentials under dark conditions (Fig. 3f), demonstrating that the NO₂⁻ was the predominant product with the electrode close to the onset potential, while the NH₃ was preferentially produced at large electrode bias. The selectivity of products in the dark has to be decided by the Pd nanoparticles. Once light is illuminated, the percentage of NH₃ is enlarged between −0.4 and −0.7 V but diminished between −0.8 and −0.9 V, with an optimal value of 89.8% at −0.6 V. Moreover, this value exceeds 50% within the entire investigated potential range. These results unequivocally reveal that the selective NO₃⁻-to-NH₃ conversion is favoured at plasmonically excited Au–Pd/HsGDY catalysts, especially at moderate electrode biases. The selectivity change induced by the light illumination indicates a different enhancement mechanism for different products.

Under conventional electrochemical conditions, the FE of NH₃ is quite low (16.7% at −0.4 V) at Au–Pd/HsGDY and exhibits an increasing trend with electrode bias. Upon the impingement of light, it shows a notable increase with progressively enlarging magnitude when the electrode potential is adjusted from −0.4 to −0.7 V, but a slight decrease at larger electrode bias, leading to an optimal value of 75.6% at −0.7 V (Fig. 3g). The FE of NO₂⁻ is below 10% at the indicated potentials under dark conditions and becomes lower except at −0.8 V with light illumination (Fig. 3h). Note that the conversion of NO₃⁻-to-NH₃ is an eight-electron process while that of NO₃⁻-to-NO₂⁻ involves only two electron transfers, leading to the substantially lower NO₂⁻ FE despite its comparable yield rate to NH₃ at specific electrode potentials.

It should be emphasized that the sum FE of NH₃ and NO₂⁻ does not exceed 80.65%. The rest of the consumed charge could be attributed to the hydrogen generation, which was determined by the gas chromatography (Fig. S35, ESI†). The FE of hydrogen is 75.3% at −0.4 V and decreases with the electrode potential biased from −0.4 to −0.9 V under dark conditions (Fig. 3i). Under light illumination, it is remarkably reduced with the light switched on, and shows a downhill trend from −0.4 to −0.7 V but experiences certain increase at −0.8 and −0.9 V (Fig. 3g). It is signified that the Au plasmons hampered the storage of the consumed charges in hydrogen at moderately electrified Au–Pd/HsGDY catalysts.

During the plasmon-mediated electrochemical reduction of NO₃⁻-to-NH₃, TEM images, elementary mapping and XRD patterns cooperatively prove that the morphology and crystal phases of Au–Pd/HsGDY were scarcely changed (Fig. S36, ESI†). The slight change in the XRD patterns of the working electrode after the reaction was ascribed to the uneven distribution of Au–Pd/HsGDY across the entire carbon cloth (Fig. S37, ESI†). The ICP-MS measurements show the Pd and Au ions were undetectable in the electrolyte after NO₃⁻-to-NH₃ conversion (Table S2, ESI†), indicating that catalyst leaching did not occur. In addition, the Pd 3d core-level XPS remained nearly unchanged (Fig. S38a, ESI†). The temperature programmed reduction profile of the nanodendrites, prepared according to a published protocol,⁵² exhibited a negative peak at 386 K ascribable to Pd hydride decomposition,⁵³ which was absent in the profile for Au–Pd/HsGDY (Fig. S38b, ESI†). These results suggest that the Pd hydride was not formed on the Pd surface. Moreover, the yield rate of NH₃ remained nearly constant for 12 h (Fig. S39, ESI†). These data show the long-term stability of the plasmonic antenna–reactor catalyst in the present study.

The wavelength-dependent reduction rate of NO₃⁻ at the illuminated electrode matches the optical absorption spectrum of Au–Pd/HsGDY catalysts (Fig. 4a and Fig. S40, ESI†), consolidating the plasmonic origin of the observed performance enhancement. During the non-radiative decay processes of excited plasmons, both thermal effects and non-thermal effects (e.g., hot electrons and photopotentials) were claimed to mediate the nearby chemical reaction outcomes. A linear dependence of reaction rate on incident optical power generally indicates a photochemical process, whereas an exponential dependence would rather be the signature of a thermal effect.⁵⁴ In the present study, the yield rate of NH₃ and NO₂⁻ cannot be linearly or exponentially correlated to light intensity (Fig. 4b and Fig. S41, S42, ESI†), suggesting that the NH₃ and NO₂⁻ production were jointly adjusted by thermal and non-thermal effects. The disparate line shapes of rate versus light intensity for NH₃ and NO₂⁻ indicate that the thermal and non-thermal effects distinctly influenced the NH₃ and NO₂⁻ generation. Besides, the selectivity of NH₃ is 84% and 75% when the electrode was impinged by the 425 and 625 nm light at identical optical intensity, respectively (Fig. 4c). In lieu of very close optical absorption of Au–Pd/HsGDY at these two wavelengths, this difference in NH₃ selectivity should be ascribed to different thermal and non-thermal effects, which is similar to the observations in our recent studies.^39,40,55

Fig. 4 Thermal and nonthermal effects of plasmon-enhanced NO₃⁻-to-NH₃ conversion on Au–Pd/HsGDY. (a) Wavelength-dependent reduction rate of NO₃⁻ on a plasmonically excited Au–Pd/HsGDY electrode at a potential of −0.6 V. The absorption spectrum of Au–Pd/HsGDY is superimposed for comparison. (b) Light intensity-dependent yield rate of NH₃ and NO₂⁻. (c) Selectivity of NH₃ for NO₃⁻ electroreduction rate on the Au–Pd/HsGDY electrode illuminated by light of different wavelength at identical intensity. (d) Thermal images of the electrochemical cell in the presence and absence of light illumination (135.1 W cm⁻²), with the highest saturated temperature indicated by specific values. Temperature-dependent yield rate of NO₂⁻ (e) and NH₃ (f), wherein the electrolyte was preheated to the signified temperatures using a thermoelectric couple.

To examine the thermal effects in the plasmon-enhanced NO₃⁻ reduction at electrified Au–Pd/HsGDY catalysts, we first recorded the temperature evolution of the working electrode during the electrochemical reduction of NO₃⁻ under white light illumination (135.1 W cm⁻²), which resulted in a maximum temperature of 61.5 °C (Fig. 4d). Subsequently, we conducted the same electrochemical reduction in a series of electrolytes that were preheated to various temperatures using an electric heater, but without illumination. The yield rate of NO₂⁻ was remarkably improved by elevating the reaction temperature from 0.220 mg h⁻¹ cm⁻² at 22 °C to 2.281 mg h⁻¹ cm⁻² at 60 °C (Fig. 4e and Fig. S43, ESI†). The yield enhancement of 4.1 induced by the 38 °C increment in the thermal control experiment is comparable to that of 2.5–4.0 resulting from Au plasmons in the light experiment, which equivocally reveals that thermal effects play a dominant role in the plasmon-assisted conversion of NO₃⁻ to NO₂⁻. In contrast, the yield rate of NH₃ only rose by 14.6% with the electrolyte temperature increased from 22 °C to 60 °C, demonstrating that the NO₃⁻-to-NH₃ conversion is insensitive to temperature variations (Fig. 4f). Therefore, the yield enhancement of NH₃ at the Au–Pd/HsGDY electrode should be mainly ascribed to the non-thermal effects of Au plasmons.

The community has not yet reached a consensus on the exact origin of non-thermal effects in plasmonic chemistry.⁴³ Hot charge carriers generated from the non-radiative decay of plasmons can modify the energetics and kinetics of chemical transformations in multiple ways, directly reducing reactants to products,⁵⁶ modulating the adsorption/desorption behaviour of intermediates,^38,57 mediating catalyst chemical valency,⁵⁸ and so on. Photopotentials, which originate from the asymmetric charge transfer to solution-phase acceptors or the change of free carrier density under off-resonance plasmonic excitation, can facilitate chemical conversions.^40,43,59 The voltage-step-controlled experiment, in which small cathodic voltage pulses replaced the light illumination, has proved to be an effective and convenient way to explore the role of photopotentials in plasmon-mediated electrochemisty.^40,60 Both voltage pulses and photopotentials will induce an apparent spike current. Following the same strategy, we measured the chronoamperometric curves of the working electrode under the superimposition of various cathodic voltage pulses (Fig. S44, ESI†) and the impingent of white light, respectively. The apparent spike currents were clearly observed in the response current curve of the voltage-step-controlled experiment but absent in that of the light experiment. This stark comparison illustrates the negligible contribution of photopotentials in plasmon-mediated electrochemical reduction of NO₃⁻ at Au–Pd/HsGDY. In this scenario, it is plausibly postulated that hot charge carriers predominate the non-thermal effects, inducing the yield enhancement of NH₃ at the Au–Pd/HsGDY electrode.

The generation and dynamics of hot electrons in Au/HsGDY and Au–Pd/HsGDY were experimentally investigated by ultrafast transient absorption (TA) measurement. Upon pulse excitation, the transient bleach signal around 580 nm with a positive swing was observed in their optical density (Fig. S45, ESI†), which can be ascribed to the generation of hot electrons during Au plasmon decay.^61,62 The dynamics of hot electrons were tracked by the TA kinetics in picoseconds, and fitted with a two-step decay model (Fig. 5a and b). The time constant of electron–phonon scattering (τ₁) was shortened from 1.7 ps in Au/HsGDY to 0.7 ps in Au–Pd/HsGDY, which can be ascribed to the larger electron–phonon coupling constant in Pd (8.7 × 10¹⁷ W m⁻³ K⁻¹) compared with that of Au (3.0 × 10¹⁶ W m⁻³ K⁻¹).^61,63 These results provide strong evidence for the hot injection from Au to Pd nanocrystals. The injected hot electrons can diffuse to the surface of Pd nanocrystals because their transport distance can reach ∼20 nm,⁶⁴ larger than the Pd diameter (4.7 nm). In the literature, the explanation of time constant (τ₂) remains inconclusive (the charge recombination⁶³ or phonon–phonon scattering⁶²). Therefore, the τ₂ are not discussed further in the present study. Note that hot electrons may transfer to HsGDY, but have negligible effects.

Fig. 5 Mechanisms underlying the improvement in NH₃ synthesis from NO₃⁻ under plasmonic excitation. (a) and (b) Ultrafast TA signals from kinetic curves of Au/HsGDY and Au–Pd/HsGDY as a function of probe delay recorded with a 500 nm pump and probed at 580 nm. (c) Free-energy diagrams of NO₃⁻ reduction on the surface of the Pd₅Au (111) under dark conditions (ground) and light illumination (excited), with their rate-determining steps (RSD) marked by the black and red boxes, respectively. (d) The ground-state and excited-state potential energy profile of the reaction pathways for hydrogen evolution on the surface of Pd₅Au(111). The Pd, Au, N, O and H atoms are represented by cyan, yellow, blue, red and white spheres, respectively.

Density functional theory (DFT) calculations were performed to elucidate how the hot charge-carriers facilitate NH₃ production from NO₃⁻ at Au–Pd/HsGDY (see the ESI† for details). Based on the experimental results, the NO₃⁻-to-NH₃ conversion took place at Pd nanoparticles, which was improved by the hot carriers generated from the non-radiative decay of Au NP plasmons. It has been well established that the plasmons are confined within surface of Au NPs, i.e., localized surface plasmon resonance.⁶⁵ The NO₃⁻ is preferentially adsorbed at Pd surfaces rather than at Au–Pd interfaces (Fig. S46, ESI†). The Pd electronic structure of Pd₅Au is almost identical with metallic Pd (Fig. S47, ESI†). Therefore, the surface of the Pd₅Au (111) model was used in the DFT calculation, where the Pd atoms act as the reaction sites and the Au atoms beneath represent the origin of hot carriers. The small quantity of Au reflects the involvement of only surface Au atoms, and ensures that the electronic structures of Pd sites would not be affected by the Au themselves. The ground-state and excited-state potential energy profiles refer to the reaction that took place in the dark and under light illumination, respectively (Fig. 5c and Table S3, ESI†). In calculating the excited-state intermediates, we adopted an equivalent approximation approach, which has been developed in a recent study.⁶⁶ Specifically, the Pd₅Au (111) and reaction intermediates were treated as a single entity, and one electron was promoted to the lowest unoccupied molecular orbitals (LUMO) of this entity to simplify the hot electron injection from plasmonically excited Au NPs to reaction intermediates. The ground-state adsorption energy barrier of NO₃⁻ is 0.72 eV, representing the rate-determining step (RSD) for NH₃ production under dark conditions. To our delight, this value is reduced to 0.35 eV in the excited state. This substantial reduction shifts the RSD from NO₃⁻ adsorption to *NH protonation (0.43 eV). Note that the adsorption involves complex interactions beyond simple charge considerations. The net benefit of 0.29 eV in the reaction energy barrier is kinetically conducive to the NH₃ generation, which matches our experimental observations that the NO₃⁻-to- NH₃ conversion was accelerated by light illumination. The contribution of hot electrons is expected to be maximized once all of the hot electrons are energetically accessible to the reaction sites (∼−0.7 V).³⁹ Regarding NO₂⁻ production, the RSD is a desorption energy barrier of *NO₂⁻ both in the ground (2.00 eV) and excited states (2.13 eV), indicating its unfavourable generation. Only a 6.1% drop is consistent with the fact that hot electrons exert negligible effects on NO₃⁻-to-NO₂⁻ conversion. In addition, the reaction energy barrier for H₂ formation was increased from 0.49 eV at the ground state to 0.98 eV at the excited one through plasmon-induced modification of the electronic structure of the entity comprising Pd₅Au (111) and adsorbed H, explaining the decrease in FE for H₂ generation (Fig. 5d). To sum up, the injection of hot electrons of Au plasmons facilitates the electrochemical reduction of NO₃⁻ to NH₃ and suppresses the hydrogen evolution reaction at the Pd/HsGDY catalyst through optimizing the reaction intermediates’ adsorption behaviour at the catalyst.

Besides unravelling the effect of hot electrons, the DFT calculations shed some light on the disparate thermal effects of plasmon-meditated electrochemical reduction of NO₃⁻ to NO₂⁻ and NH₃ as well. On the one hand, the activation energy barrier is larger for the NO₂⁻ generation than the NH₃⁻ formation, according to the thermal Arrhenius empirical law, meaning that the same temperature increment will lead to a bigger increment in the yield rate of NO₂⁻ than that of NH₃. This prediction is supported by the experimental data that the temperature increment of 38 °C accelerates the yield rate of NO₂⁻ and NH₃ by 310% and 14.6%, respectively. On the other hand, once all of the hot electrons were energetically accessible to the reaction sites at large electrode biases, their contribution is expected to reach saturation, while thermal effects are active within the entire potential range. As a result, the thermal effects would become more prominent by facilitating the H₂ desorption at elevated temperatures, which is supported by the increased FE of H₂ and decreased FE of NH₃ and NO₂⁻ over −0.7 V. Quantification of the contribution of thermal and non-thermal effects remains a grand challenge in the plasmonic catalysis field, which is beyond the scope of our current study. It is worth noting that the maximum utilization of light energy for the NO₃⁻-to-NH₃ electrochemical conversion is 0.25% (Fig. S48, ESI†), due to the fact the majority of hot electrons would recombine with holes in plasmonic metals.^67,68 It is also worth noting that the lowest production price is US$3.2 per kilogram NH₃ (Fig. S49, ESI†), while it was US$1.15 for the Haber–Bosch process.⁶⁹ However, the treatment cost of typical NO₃⁻ wastewater is US$65 per kmol N,⁶⁹ which is equivalent to US$3.82 per kg NH₃. Therefore, the NO₃⁻-to-NH₃ conversion reported herein is still economically competitive.

3. Conclusions

In summary, a plasmonic antenna reactor, Au–Pd/HsGDY, was constructed for harvesting light energy to improve the performance of NO₃⁻ to NH₃ electrochemical conversion. Plasmonic excitation of Au–Pd/HsGDY resulted in a remarkable increase in the activity of NO₃⁻ reduction, with preferential selectivity toward NH₃ over NO₂⁻ production. Under optimal conditions, the yield rate of NH₃ was enhanced by 14.3 times, and the molar ratio of NH₃ in N-containing products was almost doubled (89.5%). We elucidated that non-thermal effects are dominantly responsible for boosting the NH₃ over NO₂⁻ at plasmonically excited Au–Pd/HsGDY. DFT calculations further reveal that hot electrons can significantly change the adsorption energy of intermediates, which shifts the rate-determining step from NO₃⁻ adsorption to the *NH protonation and lowers the apparent activation energy of NO₃⁻ reduction processes. In addition, hot electrons will suppress the hydrogen evolution by strengthening the H* adsorption at catalytic sites. The present design principle can be extended to other catalysts, offering a new avenue to design superior catalysts for efficient and sustainable NO₃⁻-to-NH₃ conversion by combining the merits of both plasmonic chemistry and electrocatalysis.

Data availability

The data supporting this article have been included as part of the ESI.† Other details are available upon request from the corresponding authors.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

This work was jointly supported by the Innovation and Technology Commission of HKSAR through the Hong Kong Branch of the National Precious Metals Material Engineering Research Centre, the Research Grants Council of Hong Kong (GRF grant no. 15304519), the National Natural Science Foundation of China (22408055), the Foundation of Basic and Applied Basic Research of Guangdong Province (2023A1515110233), the Foundation of Basic and Applied Basic Research of Guangzhou (SL2024A04J00921), and the Guangdong Province Science and Technology Plan Project (2023B1212120008). Prof. Lu thanks the support from Guangdong-Hong Kong Joint Laboratory of Modern Surface Engineering Technology, City University of Hong Kong, Hong Kong, China.

Notes and references

1. K. Li, Z. Andersen Suzanne, J. Statt Michael, M. Saccoccio, J. Bukas Vanessa, K. Krempl, R. Sažinas, B. Pedersen Jakob, V. Shadravan, Y. Zhou, D. Chakraborty, J. Kibsgaard, C. K. Vesborg Peter, K. Nørskov Jens and I. Chorkendorff, Science, 2021, 374, 1593–1597.
2. D. R. MacFarlane, P. V. Cherepanov, J. Choi, B. H. R. Suryanto, R. Y. Hodgetts, J. M. Bakker, F. M. Ferrero Vallana and A. N. Simonov, Joule, 2020, 4, 1186–1205.
3. G. Soloveichik, Nat. Catal., 2019, 2, 377–380.
4. J. G. Chen, R. M. Crooks, L. C. Seefeldt, K. L. Bren, R. M. Bullock, M. Y. Darensbourg, P. L. Holland, B. Hoffman, M. J. Janik, A. K. Jones, M. G. Kanatzidis, P. King, K. M. Lancaster, S. V. Lymar, P. Pfromm, W. F. Schneider and R. R. Schrock, Science, 2018, 360, eaar6611.
5. P. Li, R. Li, Y. Liu, M. Xie, Z. Jin and G. Yu, J. Am. Chem. Soc., 2023, 145, 6471–6479.
6. Y. Zhang, X. Zheng, X. Guo, J. Zhang, A. Yuan, Y. Du and F. Gao, Appl. Catal., B, 2023, 336, 122891.
7. Y. Zhang, F. Gao, D. Wang, Z. Li, X. Wang, C. Wang, K. Zhang and Y. Du, Coord. Chem. Rev., 2023, 475, 214916.
8. Y. Guo, T. Wang, Q. Yang, X. Li, H. Li, Y. Wang, T. Jiao, Z. Huang, B. Dong, W. Zhang, J. Fan and C. Zhi, ACS Nano, 2020, 14, 9089–9097.
9. N. C. Kani, J. A. Gauthier, A. Prajapati, J. Edgington, I. Bordawekar, W. Shields, M. Shields, L. C. Seitz, A. R. Singh and M. R. Singh, Environ. Sci. Technol., 2021, 14, 6349–6359.
10. B. H. R. Suryanto, H.-L. Du, D. Wang, J. Chen, A. N. Simonov and D. R. MacFarlane, Nat. Catal., 2019, 2, 290–296.
11. J. Long, S. Chen, Y. Zhang, C. Guo, X. Fu, D. Deng and J. Xiao, Angew. Chem., Int. Ed., 2020, 59, 9711–9718.
12. G.-F. Chen, Y. Yuan, H. Jiang, S.-Y. Ren, L.-X. Ding, L. Ma, T. Wu, J. Lu and H. Wang, Nat. Energy, 2020, 5, 605–613.
13. X. Wang, X. Wu, W. Ma, X. Zhou, S. Zhang, D. Huang, L. R. Winter, J. H. Kim and M. Elimelech, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2217703120.
14. Y. Wang, A. Xu, Z. Wang, L. Huang, J. Li, F. Li, J. Wicks, M. Luo, D. H. Nam, C. S. Tan, Y. Ding, J. Wu, Y. Lum, C. T. Dinh, D. Sinton, G. Zheng and E. H. Sargent, J. Am. Chem. Soc., 2020, 142, 5702–5708.
15. F. Y. Chen, Z. Y. Wu, S. Gupta, D. J. Rivera, S. V. Lambeets, S. Pecaut, J. Y. T. Kim, P. Zhu, Y. Z. Finfrock, D. M. Meira, G. King, G. Gao, W. Xu, D. A. Cullen, H. Zhou, Y. Han, D. E. Perea, C. L. Muhich and H. Wang, Nat. Nanotechnol., 2022, 17, 759–767.
16. S. Han, H. Li, T. Li, F. Chen, R. Yang, Y. Yu and B. Zhang, Nat. Catal., 2023, 6, 402–414.
17. Z. Gong, X. Xiang, W. Zhong, C. Jia, P. Chen, N. Zhang, S. Zhao, W. Liu, Y. Chen and Z. Lin, Angew. Chem., Int. Ed., 2023, 62, e202308775.
18. J. Xu, S. Zhang, H. Liu, S. Liu, Y. Yuan, Y. Meng, M. Wang, C. Shen, Q. Peng, J. Chen, X. Wang, L. Song, K. Li and W. Chen, Angew. Chem., Int. Ed., 2023, 62, e202308044.
19. H. Du, H. Guo, K. Wang, X. Du, B. A. Beshiwork, S. Sun, Y. Luo, Q. Liu, T. Li and X. Sun, Angew. Chem., Int. Ed., 2023, 62, e202215782.
20. M. Xie, S. Tang, Z. Li, M. Wang, Z. Jin, P. Li, X. Zhan, H. Zhou and G. Yu, J. Am. Chem. Soc., 2023, 145, 13957–13967.
21. E. Murphy, Y. Liu, I. Matanovic, M. Ruscher, Y. Huang, A. Ly, S. Guo, W. Zang, X. Yan, A. Martini, J. Timoshenko, B. R. Cuenya, I. V. Zenyuk, X. Pan, E. D. Spoerke and P. Atanassov, Nat. Commun., 2023, 14, 4554.
22. Y. Yao, L. Zhao, J. Dai, J. Wang, C. Fang, G. Zhan, Q. Zheng, W. Hou and L. Zhang, Angew. Chem., Int. Ed., 2022, 61, e202208215.
23. J. Wang, W. He, Y. Zong, Y. Tang, J. Wang and R. Ma, Chem. Commun., 2024, 60, 9444–9447.
24. D. Qin, S. Song, Y. Liu, K. Wang, B. Yang and S. Zhang, Angew. Chem., Int. Ed., 2023, 62, e202304935.
25. Y. Zhang, H. Zheng, K. Zhou, J. Ye, K. Chu, Z. Zhou, L. Zhang and T. Liu, Adv. Mater., 2023, 35, e2209855.
26. K. Fan, W. Xie, J. Li, Y. Sun, P. Xu, Y. Tang, Z. Li and M. Shao, Nat. Commun., 2022, 13, 7958.
27. Y. Feng, X. Wang, J. Ma, N. Wang, Q. Liu, K. Suenaga, W. Chen, J. Zhang, Y. Zhou and J. Wang, Adv. Energy Mater., 2024, 14, 2401501.
28. J. Zhang, W. He, T. Quast, J. R. C. Junqueira, S. Saddeler, S. Schulz and W. Schuhmann, Angew. Chem., Int. Ed., 2023, 62, e202214830.
29. H. Jiang, G. F. Chen, O. Savateev, J. Xue, L. X. Ding, Z. Liang, M. Antonietti and H. Wang, Angew. Chem., Int. Ed., 2023, 62, e202218717.
30. K. Kim, A. Zagalskaya, J. L. Ng, J. Hong, V. Alexandrov, T. A. Pham and X. Su, Nat. Commun., 2023, 14, 823.
31. W. J. Sun, H. Q. Ji, L. X. Li, H. Y. Zhang, Z. K. Wang, J. H. He and J. M. Lu, Angew. Chem., Int. Ed., 2021, 60, 22933–22939.
32. Z. Song, Y. Liu, Y. Zhong, Q. Guo, J. Zeng and Z. Geng, Adv. Mater., 2022, 34, e2204306.
33. Q. Gao, H. S. Pillai, Y. Huang, S. Liu, Q. Mu, X. Han, Z. Yan, H. Zhou, Q. He, H. Xin and H. Zhu, Nat. Commun., 2022, 13, 2338.
34. Q. Gao, B. Yao, H. S. Pillai, W. Zang, X. Han, Y. Liu, S.-W. Yu, Z. Yan, B. Min, S. Zhang, H. Zhou, L. Ma, H. Xin, Q. He and H. Zhu, Nat. Synth., 2023, 2, 624–634.
35. B. Li, F. Xia, Y. Liu, H. Tan, S. Gao, J. Kaelin, Y. Liu, K. Lu, T. J. Marks and Y. Cheng, Nano Lett., 2023, 23, 1459–1466.
36. Y. Bu, C. Wang, W. Zhang, X. Yang, J. Ding and G. Gao, Angew. Chem., Int. Ed., 2023, 62, e202217337.
37. H. Zhu, J. J. Wang, Z. Xu, Y. Tan and J. Wang, Small, 2024, 2404919.
38. L. Zhou, D. F. Swearer, C. Zhang, H. Robatjazi, H. Zhao, L. Henderson, L. Dong, P. Christopher, E. A. Carter, P. Nordlander and N. J. Halas, Science, 2018, 362, 69–72.
39. W. Ou, B. Zhou, J. Shen, T. W. Lo, D. Lei, S. Li, J. Zhong, Y. Y. Li and J. Lu, Angew. Chem., Int. Ed., 2020, 59, 6790–6793.
40. W. Ou, Y. Fan, J. Shen, Y. Xu, D. Huang, B. Zhou, T. W. Lo, S. Li, Y. Y. Li, D. Lei and J. Lu, Nano Lett., 2022, 22, 8397–8405.
41. E. Contreras, R. Nixon, C. Litts, W. Zhang, F. M. Alcorn and P. K. Jain, J. Am. Chem. Soc., 2022, 144, 10743–10751.
42. Y. Guo, J. Liu, Q. Yang, P. Khemthong, Z. Huang, Y. Zhao, Z. Chen, B. Dong, X.-Z. Fu, J.-L. Luo and C. Zhi, Nano Energy, 2021, 86, 106099.
43. W. Ou, B. Zhou, J. Shen, C. Zhao, Y. Y. Li and J. Lu, iScience, 2021, 24, 101982.
44. S. Linic, U. Aslam, C. Boerigter and M. Morabito, Nat. Mater., 2015, 14, 567–576.
45. V. Muravev, G. Spezzati, Y.-Q. Su, A. Parastaev, F.-K. Chiang, A. Longo, C. Escudero, N. Kosinov and E. J. M. Hensen, Nat. Catal., 2021, 4, 469–478.
46. Z. Wei, Z. Zhao, C. Qiu, S. Huang, Z. Yao, M. Wang, Y. Chen, Y. Lin, X. Zhong, X. Li and J. Wang, Nat. Commun., 2023, 14, 661.
47. F. Zhao, L. Zheng, Q. Yuan, X. Yang, Q. Zhang, H. Xu, Y. Guo, S. Yang, Z. Zhou, L. Gu and X. Wang, Adv. Mater., 2021, 33, e2103383.
48. X. Ma, Y. Shen, S. Yao, C. An, W. Zhang, J. Zhu, R. Si, C. Guo and C. An, J. Mater. Chem. A, 2020, 8, 3344–3350.
49. W. He, J. Zhang, S. Dieckhofer, S. Varhade, A. C. Brix, A. Lielpetere, S. Seisel, J. R. C. Junqueira and W. Schuhmann, Nat. Commun., 2022, 13, 1129.
50. W. Gao, K. Xie, J. Xie, X. Wang, H. Zhang, S. Chen, H. Wang, Z. Li and C. Li, Adv. Mater., 2023, 35, e2202952.
51. R. Chen, C. Yang, W. Cai, H.-Y. Wang, J. Miao, L. Zhang, S. Chen and B. Liu, ACS Energy Lett., 2017, 2, 1070–1075.
52. H.-Y. Sun, Y. Ding, Y.-Q. Yue, Q. Xue, F.-M. Li, J.-X. Jiang, P. Chen and Y. Chen, ACS Appl. Mater. Interfaces, 2021, 13, 13149–13157.
53. F. Cárdenas-Lizana, Y. Hao, M. Crespo-Quesada, I. Yuranov, X. Wang, M. A. Keane and L. Kiwi-Minsker, ACS Catal., 2013, 3, 1386–1396.
54. G. Baffou, I. Bordacchini, A. Baldi and R. Quidant, Light: Sci. Appl., 2020, 9, 108.
55. B. Zhou, W. Ou, J. Shen, C. Zhao, J. Zhong, P. Du, H. Bian, P. Li, L. Yang, J. Lu and Y. Y. Li, ACS Catal., 2021, 11, 14898–14905.
56. S. Mubeen, J. Lee, N. Singh, S. Kramer, G. D. Stucky and M. Moskovits, Nat. Nanotechnol., 2013, 8, 247–251.
57. Y. Dubi, I. W. Un, J. H. Baraban and Y. Sivan, Nat. Catal., 2022, 5, 244–246.
58. A. Marimuthu, J. Zhang and S. Linic, Science, 2013, 339, 1590–1593.
59. A. J. Wilson and P. K. Jain, Acc. Chem. Res., 2020, 53, 1773–1781.
60. C. Zhan, B.-W. Liu, Y.-F. Huang, S. Hu, B. Ren, M. Moskovits and Z.-Q. Tian, Nat. Commun., 2019, 10, 2671.
61. J. Guo, Y. Zhang, L. Shi, Y. Zhu, M. F. Mideksa, K. Hou, W. Zhao, D. Wang, M. Zhao, X. Zhang, J. Lv, J. Zhang, X. Wang and Z. Tang, J. Am. Chem. Soc., 2017, 139, 17964–17972.
62. X. Zhang, Y. Fan, E. You, Z. Li, Y. Dong, L. Chen, Y. Yang, Z. Xie, Q. Kuang and L. Zheng, Nano Energy, 2021, 84, 105950.
63. H. Huang, L. Zhang, Z. Lv, R. Long, C. Zhang, Y. Lin, K. Wei, C. Wang, L. Chen, Z. Y. Li, Q. Zhang, Y. Luo and Y. Xiong, J. Am. Chem. Soc., 2016, 138, 6822–6828.
64. S. C. Huang, X. Wang, Q. Q. Zhao, J. F. Zhu, C. W. Li, Y. H. He, S. Hu, M. M. Sartin, S. Yan and B. Ren, Nat. Commun., 2020, 11, 4211.
65. C. Zhan, X.-J. Chen, J. Yi, J.-F. Li, D.-Y. Wu and Z.-Q. Tian, Nat. Rev. Chem., 2018, 2, 216–230.
66. J. Zhao, Y. Bai, Z. Li, J. Liu, W. Wang, P. Wang, B. Yang, R. Shi, G. I. N. Waterhouse, X. D. Wen, Q. Dai and T. Zhang, Angew. Chem., Int. Ed., 2023, 62, e202219299.
67. H. Robatjazi, J. L. Bao, M. Zhang, L. Zhou, P. Christopher, E. A. Carter, P. Nordlander and N. J. Halas, Nat. Catal., 2020, 3, 564–573.
68. S. Linic, S. Chavez and R. Elias, Nat. Mater., 2021, 20, 916–924.
69. F.-Y. Chen, A. Elgazzar, S. Pecaut, C. Qiu, Y. Feng, S. Ashokkumar, Z. Yu, C. Sellers, S. Hao, P. Zhu and H. Wang, Nat. Catal., 2024, 7, 1032–1043.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ee03678f

---