Sustainable chemistry with plasmonic photocatalysts

3.1 Ammonia synthesis (N₂ reduction)

The Haber-Bosch Process is the traditional method of producing ammonia for fertilizer and emblematic of the challenge researchers must address to move the chemical industry into a sustainable future. Global population growth has been enabled specifically because of the invention of the ammonia synthesis process, contributing to ∼50 % of world population. Historically, around 78 % of the ammonia produced by the Haber-Bosch process is used in agriculture, as fertilizer to feed the world [45, 46]. Yet, despite its societal impact, the Haber-Bosch process is also responsible for a considerable carbon footprint. Traditionally, Fe-based catalysts supported on oxide are used under high temperature (300–500 °C) and pressure (100 atm) to convert N₂ and H₂ into NH₃ [47, 48]. The process accounts for around 2 % of total energy from the chemical industry, and consumes 50 % of hydrogen production within the sector, leading to 3 % of annual global greenhouse gas emissions [49, 50]. There are few chemical reactions as important to the world which so critically necessitate innovation.

The fixation of dinitrogen species onto the surface of active centers is a crucial step in the reaction. The electronic structure of transition metals such as Os, Ru, and Fe makes them promising candidates to provide the active surface for this step. For instance, the first and second generations of the Haber-Bosch catalysts, which are based on Fe and Ru, respectively, rely on the Fe C7 and Ru B5 active sites for the activation of dinitrogen species [14, 51]. Plasmonic antennas offer a viable approach for improving the activation and fixation of dinitrogen species through the strong optical near-field enhancement, facilitating electronic transitions of the adsorbates by hot carriers, and traditional equilibrium and thermodynamic effects from photothermal conversion of light energy.

Early photocatalytic studies of ammonia synthesis show that light irradiation can activate N₂ at milder conditions than the industrial Haber-Bosch process. Zeng, H., et al. demonstrated a proof of concept for photochemical ammonia synthesis by using Au-Os nanoparticles created via sputtering on a glass substrate. Gas-phase ammonia synthesis (N₂ + 3H₂ → 2NH₃) is achieved at room temperature and pressure using 200 mW/cm² irradiation from a solar simulator with a 450 nm cutoff filter. The study proposes that plasmons are influential to the process, since the extinction spectra of the sample matches with wavelength-dependent reactivity measurements (Figure 2A). This study represents one of the initial demonstrations of using a thin layer of Os nanoparticles as the active catalyst to drive the reaction photochemically. It is noteworthy that Os was the first transition metal historically used for traditional ammonia synthesis [52]. In an aqueous phase (N₂ + 6H⁺ + 6e⁻ → 2NH₃), Hu, C., et al. showed that Au-Ru nanoparticles with multiple bimetallic ratios under Xenon lamp illumination at 400 mW/cm², 2 bar pressure, and room temperature can facilitate the associative alternating hydrogenation of dinitrogen species when activating the surface plasmon resonance. The research team demonstrated the plasmon-driven reactivity of the system through the linear relationship between the reactivity and the input power (Figure 2B). Moreover, their first-principle model provided mechanistic insights into the necessity of a strong electric near-field for the activation of dinitrogen species onto the Ru center [53]. Au nanoparticles on various semiconductors, such as TiO₂, CeO₂, and g-C₃N₄, achieve enhanced visible light absorption as well as charge separation at the metal-semiconductor interface for nitrogen fixation. Semiconductor materials with appropriate band structure and suitable active surfaces are inherently effective catalysts for nitrogen fixation under UV illumination. Incorporating plasmon-mediated visible light absorption with active oxygen or nitrogen defects can augment ammonia fixation, providing an additional avenue for promoting dinitrogen fixation that is distinct from traditional transition metal-based catalysts [54–57]. Supporting Au nanoparticles in the metal-organic framework (MOF) UiO-66 increases permeability owing to the large surface area of the matrix, better enabling the Au nanoparticles to activate N₂ on their active sites. The MOF support provided excellent stability for over 24 h, a significant improvement compared to only 8 h when the same Au was supported on other metal oxides (Figure 2C). Traditionally, the Au surface of the MOF is a poor active center for the activation of dinitrogen species. However, the unique Au-MOF interface, which exhibits an enhanced surface area, displayed extraordinary reactivity and a decrease in light-induced activation [58]. There are also several studies utilizing Ru catalysts which are similar to industrial catalysts in components and structure, but now the plasmonic photothermal effect is synergistically employed to drive gas-phase ammonia synthesis. Li, X., et al. reported the LED illumination of a conventional Ru-Cs/MgO catalyst and utilized the temperature gradient caused by the plasmonic photothermal effect (Figure 2D) to achieve a reaction rate of 4464 μmol g⁻¹ h⁻¹ at 4.7 W/cm², detecting the gas-phase conversion of N₂ to NH₃ in a mass spectrometer [59]. Mao, C., et al. reported a K/Ru/TiO_2−x H _x catalyst at 360 °C under 300 W Xe lamp irradiation and reaching 112.6 μmol g⁻¹ h⁻¹(Figure 2E) [60]. The same research team also investigated a Fe-based nano-necklace/TiO_2−x H _x and achieved ammonia concentration of 1939 ppm at 1 atm and 19,620 ppm at 10 atm [61]. Hou, T., et al. developed a porous Cu₉₆Fe₄ nanostructure and achieved 342 μmol g⁻¹ h⁻¹ at 250 mW/cm² with the full spectrum of their Xenon lamp, and 242 μmol g⁻¹ h⁻¹ with a >400 nm filter at 200 mW/cm². The catalyst also displays good stability during 10 cycles (Figure 2F) [62]. Table 1 summarizes the recent progress on N₂ reduction. Taken together, these results show that plasmons can drive ammonia synthesis at milder temperatures and pressures than the Haber-Bosch process. Yet, further work is needed to regulate the dinitrogen activation towards the dissociative pathways and therefore boost its efficiency towards industrial scales. To identify avenues for controlling the associative hydrogenation of dinitrogen to distal and dissociation pathways, it is necessary to perform in-situ measurements of intermediate steps. Furthermore, high-resolution, and even atomic-resolution electron microscopy is essential to determine the active centers and their interaction evolution during the reaction. Theoretical studies are also crucial for gaining insights into the design of nanostructures, Particularly for understanding how excited-states can potentially lower the activation barrier of dinitrogen species.

Figure 2:

Plasmonic nitrogen fixation. (A) The wavelength-dependent reactivity (blue open dots and dashed curve) is well-correlated to the extinction spectrum (pink solid line) on a Au-Os nanoparticles for gas-phase ammonia synthesis (N₂ + 3H₂ → 2NH₃). Adapted from ref. [52] with permission. Copyright 2014 Elsevier B.V. (B) Ammonia synthesis in presence of water (N₂ + 6H⁺ + 6e⁻ → 2NH₃) on a AuRu core-antenna nanostructure under full-spectrum irradiation at 2 atm N₂. Adapted from ref. [53] with permission. Copyright 2019 American Chemical Society. (C) Stability test of ammonia synthesis on Au in presence of water on UiO-66 MOF. Adapted from ref. [63] with permission. Copyright 2019 American Chemical Society. (D) The plasmonic gas-phase ammonia synthesis with a commercial Ru-Cs/MgO catalyst driven by the thermal gradient. Adapted from ref. [59]. Copyright 2019 American Chemical Society. (E) Plasmonic photothermal gas-phase ammonia synthesis with a K/Ru/TiO_2−x H _x catalyst. Adapted from ref. [60] with permission. Copyright 2017 Elsevier B.V. (F) Stability of a porous plasmonic CuFe for nitrogen fixation with 1 atm N₂ in water. Adapted from ref. [62] with permission. Copyright 2020 American Chemical Society.

3.2 H₂ production

H₂ is not only a ubiquitous reducing agent for a multitude of reactions in heterogeneous catalysis and organic/pharmaceutical synthesis, but also a promising carbon-free fuel for vehicles and a high-density storage technology [64]. Many renewable applications try to split water into hydrogen and oxygen via photocatalysis, electrocatalysis, or photoelectrocatalysis [65–68]. The first instance of hydrogen evolution through direct light-driven water splitting dates back to 1971, when Fujishima and Honda discovered that UV irradiation on the surface of TiO₂ enabled the direct photocatalysis of H₂O into H₂ and O₂ [69]. Since then, scientists have explored various strategies to engineer semiconductor materials to achieve high yields of green hydrogen from solar energy. To achieve solar-to-hydrogen efficiency that is competitive with other hydrogen generation methods, a theoretical energy efficiency of 10 % would require a single absorber with a band gap smaller than approximately 530–600 nm [70]. The first milestone was achieved by Moskovits and co-workers, who used a Au-TiO₂-Pt/Co thin-film device for photoelectrochemical water splitting to demonstrate the proof of concept that plasmon-induced hot carriers can be extracted to drive water splitting [71, 72]. Recent advancements have focused mainly on the fundamental aspects of extracting hot electrons or hot holes to facilitate both quantum efficiency and energy efficiency [73–75]. Many demonstrations of utilizing composite heterostructure semiconductors with plasmonic antennas to enhance visible light absorption are discussed in detail in ref. [15, 76]. Challenges such as constructing appropriate multiple Schottky junctions between metal and semiconductor for efficient charge separation and understanding the multiple electron coupling process in aqueous-phase reactions need to be further investigated.

Several reactions are used to produce hydrogen in industry, including ammonia decomposition, cracking of fossil fuels, methane reforming, and the Claus process. These methods generally require harsh operating conditions (high temperatures and pressures with transition metal catalysts) [77]. Recent progress in plasmonic photocatalysis suggests that non-thermal hot-carrier-driven reactions can be used for the production of hydrogen from a variety of sources with improved selectivity and lower activation barriers at mild conditions under visible light irradiation.

Zhou, L., et al. reported Cu-Ru surface alloy antenna-reactors for ammonia decomposition (2NH₃ → 2N₂ + 3H₂) [78]. Photocatalytic reactivity under white light illumination on the surface of Cu-Ru can be 1–2 orders of magnitudes higher than the reactivity from thermocatalysis at comparable temperatures. The energy efficiency for converting from light to chemical energy can be as high as 18 %. The apparent activation energies of the reaction under different wavelengths and power densities of light can decrease from 1.2 eV to 0.2 eV (Figure 3A), suggesting that hot carriers can help tailor the rate-determining step of this reaction, which shifts from the associative desorption of N₂ in thermocatalyis to the N–H bond scission steps, therefore increasing the activity and reducing the activation barrier. Using the same synthetic techniques, the research team also compared Cu-Fe and Cu-Ru under similar light illumination conditions and found comparable performance. Originally Fe is famous for the strong adsorption with nitrogen species and can form FeN _x , thereby deactivating the active centers. However, hot carriers can remove the surface N _x species and facilitate the desorption of NH₃, turning Fe into an active center for photocatalytically decomposing the NH₃. Additionally, the authors were able to change the light source from a laser to LED while scaling the reaction by 3 orders of magnitude (Figure 3B) and maintaining an energy efficiency of more than 10 % [79].

Figure 3:

Plasmonic gas-phase catalysis for H₂ generation. (A) Apparent activation energies under different wavelengths and power densities on a Cu-Ru surface alloy antenna reactor for NH₃ decomposition (2NH₃ → 2N₂ + 3H₂). Adapted from ref. [78] with permission. Copyright 2018 The American Association for the advancement of science. (B) 3D reactor to scale up the Cu-Ru and Cu-Fe catalysts in gram scale for ammonia decomposition with LED light. Adapted from ref. [79] with permission. Copyright 2022 The American Association for the advancement of science. (C) Stability test of CuRu for methane dry reforming (CH₄ + CO₂ → 2CO + 2H₂). Adapted from ref. [80] with permission. Copyright 2020 Springer Nature. (D) Production rate and selectivity evolution with varying temperature of Au-Pt/P25 for photothermal catalysis. Adapted from ref. [81] with permission. Copyright 2022 Elsevier Inc. (E) Reaction mechanism of Au on SiO₂ for direct photocatalytic and thermocatalytic H₂S decomposition. Adapted from ref. [31] with permission. Copyright 2022 American Chemical Society. (F) In-situ FTIR measurement to detect the intermediate species during photochemical transformation on CuZn alloy for methanol steam reforming (CH₃OH + H₂O → 3H₂ + CO₂). Adapted from ref. [10] with permission. Copyright 2021 American Chemical Society.

Methane can be utilized to extract synthesis gas (carbon monoxide and hydrogen) and has gained increased attention in light of abundant methane sourced from shale gas and tight oil [82]. Methane enables hydrogen production from direct cracking, steam reforming by reacting with water, dry reforming with CO₂, and oxidation. Dry reforming is the most environmentally-friendly strategy as it reforms two greenhouse emission gases into CO and H₂ but it generally requires high operating temperatures (700–1000 °C). Zhou., L., et al. incorporated a single atom Ru into a Cu nanoparticle by the co-precipitation method and were able to drive methane dry reforming (CH₄ + CO₂ → 2CO + 2H₂) under visible light irradiation with decent stability over 50 h without coking [80]. Several ratios of Cu-Ru were tested. However, most of them were deactivated due to coking on the surface, as shown in Figure 3C. Only Cu_19.8Ru_0.2 and Cu_19.9Ru_0.1 exhibited both reactivity and selectivity for over 20 h. The authors further confirmed that Ru existed in the form of a single-atom feature by utilizing infrared CO adsorption measurements combined with a DFT + D3 calculation. By comparing the reactivity and selectivity of thermocatalysis, they suggested that the plasmon-induced desorption assists in tailoring the pathway, and the single-atom Ru center feature prevents the adjacent carbon landing sites, thereby reducing coking. Furthermore, Zhang., Z., et al. reported a reaction rate of 85.38 mmol g_cat ⁻¹ h⁻¹ and 201.92 mmol g_cat ⁻¹ h⁻¹ for H₂ and CO, respectively, using the plasmonic photothermal effect on their Pt-Au/P25 composite catalysts [81]. The results presented in this study demonstrate a 3-fold increase in reactivity compared to the reaction in the absence of light. Furthermore, the reactivity and selectivity were observed to be stable for over 15 h. The selectivity was higher at higher temperatures for both photo- and thermo-catalysis, indicating the dominance of the photothermal effect in the reaction (Figure 3D). Additionally, the authors used in-situ FTIR measurements to identify intermediate species of CH _x , HCOO^*, and H^* on the surface of their catalysts. The HCOO^* species was found to be the most important in driving the reaction towards the production of CO and H₂ by decomposition. Based on these findings, a detailed mechanism was proposed in which plasmon-induced hot carriers on the surface of TiO₂ help convert CH₄ into CH _x , while the Pt surface cleaves CO₂ and spillovers the oxygen to the interface to facilitate the formation of HCOO^*. Takami., D., et al. reported the activation of the reaction at the low temperature of 200 °C over a plasmonic Ni/Al₂O₃ catalyst, maintaining a high conversion of CH₄ and CO₂ (20 %) [83]. They recognize that the metallic Ni that support LSPR is the key to maintain the high conversion at low temperatures.

Other avenues for sustainable hydrogen production from plasmonic photocatalysis include hydrogen sulfide decomposition (H₂S → H₂ + S) and methanol steam reforming (CH₃OH + H₂O → 3H₂ + CO₂). Lou., M., et al. used plasmonic Au nanoparticles supported on silica without any external heating source to catalyze one-step direct H₂S decomposition, a potential alternative for the industrial two-steps Claus process to obtain sulfur and H₂ (Figure 3E) [31]. The results of ground and excited-state first-principle calculations, combined with partial-pressure dependent reactivity and micro-kinetic analysis, revealed that plasmon-induced hot electrons on the surface of Au nanoparticles shift the rate-determining steps from H–S^* bond scission in dark conditions to S₂ associative adsorption under light irradiation. Luo., S., et al. reported a H₂ production rate of 328 mmol g_cat ⁻¹ h⁻¹ for methanol steam reforming on a plasmonic CuZn alloy, with a light energy conversion efficiency of 1.2 % at 7.9 suns irradiation from a solar simulator [10]. The in-situ diffuse reflective infrared Fourier transform spectra (DRIFTS) confirmed the key intermediate species on the copper-based catalysts are CH _x O and HCOOH. The light illumination helps increase the concentration of the intermediates, thereby improving reactivity (Figure 3F). Gas-phase H₂ production publications are summarized in Table 2. Incorporation of these state-of-art experiments into actual industrial production requires parameter optimization and large-scale photoreactor design. Indeed, several industries and start-ups are now addressing 3D design of reactors to better utilize all the photons for photochemical transformations and to improve flow of the gas to efficiently penetrate the reactor. Although the high-efficiency conversion of H₂ generation developed so far in the field seems to facilitate the possibility of industrial-level production, fundamental studies on understanding the dynamics of hot carriers and how to extract them into chemistry are essential to increase quantum and energy efficiency. Moreover, renewable light sources such as solar energy have a rather complex wavelength distribution, making the study on extracting them more challenging than the exciting results achieved using monochromatic light sources such as lasers and LEDs.

3.3 CO₂ reduction

Reaction pathways with multiple electron transfer processes can potentially be used to reduce CO₂ in the presence of water or H₂ to turn CO₂ into value-added chemical products [84]. The metal-molecule interface of the chemical transformation can be modified via optical near-fields, hot carrier generation and multiplication [85], localized photothermal effects and gradients [86, 87], and the construction of active centers. Control of these parameters allows for coupling optical energy into different products with high selectivity, allowing for tailor-made catalysts that target particular chemical reaction pathways.

Yu., S., et al. investigated the mechanism of C–C coupling on the surface of Au with CO₂ in the presence of H₂O. The selectivity was highly tailored by both the wavelength and power density of the incident continuous wave (CW) lasers: High-intensity resonant illumination of interband transitions at 488 nm showed preferred selectivity to ethane (corresponding to a multi-electron harvesting process) while excitation at 532 nm pushed the reaction towards methane production [7]. The team found that the observed selectivity trend can be attributed to the nature of higher light intensity and photon energy, which can result in a multi-carrier process. They also observed that the reactivity and selectivity follow the Poisson statistics of electron harvesting. The team also investigated the same reaction on Ag plasmonic nanoparticles at a single-particle level by employing in-situ Raman spectroscopy. They detected C₁–C₄ product evolutions with multiple intermediate species with visible light irradiation at mild conditions [6, 88]. Trace amounts of rich catalog of multiple carbon species were generated through a multi-photon excitation process during plasmon decay. However, the detailed mechanism of this process and how to regulate the selectivity towards long carbon-chain products require further investigation and development. Abundant reports have been published of hybrid plasmonic nanoparticles with semiconductor materials or MOFs to enhance the light–matter interaction and regulate the selectivity of the major products towards CH₄ [89, 90], CO [91–93], CH₃OH [94], HCOOH [63, 95, 96], and C₂H₆ [97], among others, by taking advantages of the unique electronic structure and vacancy sites on the metal oxide semiconductor or the large surface area of MOF. Understanding the mechanism and manipulating multi-electron processes through nanostructure engineering remains a future challenge of the field given the complex possible reaction pathways.

Gas-phase CO₂ hydrogenation is another approach to use renewable light energy to produce valuable chemical products. Rather than the multiple electron transfer process as in the presence of water, gas-phase CO₂ hydrogenation mainly has two reaction pathways under room pressure: CO₂ methanation (CO₂ + 4H₂ → CH₄ + 2H₂O) and reverse water-gas shift reaction (CO₂ + H₂ → CO + H₂O). Zhang., X., et al. observed the non-thermal effect on Rh/Al₂O₃ catalysts under visible LED illumination driving reaction towards CH₄ with selectivities >86 % and >98 % under illuminations of blue (400 nm) and ultraviolet (365 nm) light, respectively [9]. Compared to thermocatalysis, plasmonic photocatalysis can selectively activate the C-O bond in CHO* intermediates by hot electrons with an apparent activation energy decrease (Figure 4A). The authors also reported a mechanistic study of a Rh/TiO₂ catalyst, showing the selective methanation of CO₂ through synergistic photothermal and hot-carrier effects (Figure 4B) [98]. The photothermal effect and the plasmon-mediated non-thermal effect (Figure 4B, iii) can be distinguished and quantified by measuring the top and bottom temperatures of the catalyst bed and combining them with a model. In addition, their DFT calculation found that CHO^* is a critical intermediate for CO₂ methanation in the presence of H₂. However, increasing the hydrogen partial pressure would inhibit the injection of hot electrons into the orbital of CHO^*. Robatjazi., H., et al. designed a sustainable aluminum-based core–shell structure of Al@MIL-53 and Al@Cu₂O to drive the reverse water-gas shift reaction by increasing the physisorption of the gas, and extracting hot electrons from Al, respectively (Figure 4C and D) [5, 99]. Al@MIL-53 enables up to eight times larger CO₂ uptake than Al nanocrystals, resulting in a four times higher reactivity towards the reverse water-gas shift reaction (shown in Figure 4C). In Al@Cu₂O, the small Schottky barrier between Al and Cu₂O (shown in Figure 4D) facilitates the separation of hot carriers generated from the Al core. This results in a regulation of selectivity towards the reverse water-gas shift reaction under visible-light illumination. Additionally, this boosts the EQE to two orders of magnitude higher than bare Cu₂O and one order of magnitude higher than Al nanoparticles. Fu., G., et al. designed a Rh/Al photothermal catalyst from colloidal synthesis with Al powders and RhCl₃ precursor, capable of regulating CO₂ hydrogenation towards methanation with nearly 100 % selectivity under a concentrated simulated solar irradiation at 11.3 W/cm², resulting in a 550 mmol g⁻¹ h⁻¹ reaction rate [8]. The in-operando Temperature-Programmed FTIR analysis identified the surface CO species on Rh is the key intermediate steps to drive the reaction towards the methanation process (Figure 4E). Table 3 outlines the relevant citations for gas-phase CO₂ hydrogenation reactions. In our opinion, further research should focus on resolving the multiphoton and electron processes under high light intensity and photon energy in both aqueous and gas-phase CO₂ reduction. Additionally, constructing active centers with different electronic structures for the various carbon-related species to accumulate and desorb can help regulate the selectivity of CO₂ conversion into valuable molecules. Even for extensively studied gas-phase reactions such as reverse water-gas shift and methanation, understanding the time resolution of the intermediate steps is important to determine how to optimize the lifetime of plasmon excitation and decay to improve conversion.

Figure 4:

Plasmonic gas-phase CO₂ hydrogenation. (A) Activation energies and selectivity of CO₂ hydrogenation towards methane on Rh/Al₂O₃ catalysts are tailored by light compared to the reaction in the dark. Adapted from ref. [9] with permission. Copyright 2017 Springer Nature. (B) CO₂ methanation reaction on Rh/TiO₂ catalysts utilizing both thermal and non-thermal effects with the temperature gradient in the catalyst bed. Adapted from ref. [98] with permission. Copyright 2018 American Chemical Society. (C) Al@MIL-53 for enlarged surface area for gas uptake and enhanced reverse water-gas shift reaction. Adapted from ref. [99] with permission. Copyright 2019 The American Association for the advancement of science. (D) Schematic of hot-electron induced chemistry at Al-Cu₂O interface for reverse water-gas shift reaction. Adapted from ref. [5] with permission. Copyright 2017 Springer Nature. (E) In-operando FTIR measurement of reaction steps under different temperatures for photothermal CO₂ methanation on a Rh/al catalyst. Adapted from ref. [8] with permission. Copyright 2021 American Chemical Society.

4.1 Wavelength-dependent photochemical transformations

For intermediate-sized nanoparticles (>5 nm and <100 nm), Mie theory and Gans theory are often employed to model the dielectric constant of a metal to analytically determine the absorption, scattering, and extinction spectra of metal nanoparticles with varying shapes, sizes, compositions, and dielectric environments. These methods are described in detail in several excellent review papers [28, 100]. Numerical methods such as finite-difference time-domain (FDTD), boundary element method (BEM), and finite element methods (FEM), can also be used to simulate and calculate the optical properties of the plasmonic nanostructure with high accuracy by solving Maxwell’s equations in three dimensions, depending on the dimension, geometry, symmetry, etc. In addition to simulating spectra and correlating these with experimental extinction, diffuse reflectance, dark-field scattering, etc., classical or semi-classical electromagnetic calculations offer us a tool to distinguish the mechanism of various reaction pathways and to interpret wavelength-dependent reactivity of plasmonic photocatalysts. An early example of utilizing this tool to distinguish the hot electrons in metal includes Zheng, B. et al., in which they fabricated a Au-TiO₂ device and measured the photocurrent with the Schottky contact and Ohmic contact between Au and TiO₂ interface [101]. The authors exploit the fact that only the plasmon-induced hot electrons have sufficient energy to overcome the Schottky barrier. The study confirmed that hot carrier generation within a plasmonic metal is directly correlated to the volume integration of |E|² within the electron mean free path of the metal. The same idea can also be applied to Al@Cu₂O core–shell nanoparticles for reverse water-gas shift, Al-Pd antenna-reactors for H₂ dissociation and C–F activation, and solvated electrons from Au and Ag nanoparticles [5, 102–104]. There are many publications which correlate absorption or extinction cross-sections with wavelength-dependent reactivity, especially when the classic absorption cross-section is very close to its extinction cross-section in Au nanoparticles [31, 53, 58, 62]. However, the portion of the energy or charge transfer process directly taking part in the chemistry is not clear in this paradigm. Yuan., L., et al. developed a model based on the semi-classical hot electron theory applied to aluminum plasmonic nanocrystals from a previous report [34, 105]. The interband hot electrons were accounted for by including the contribution of the interband transition to the imaginary part of the permittivity in the |E|² integration. Besides the peak originating from the plasmon modes, only the hot-electron cross-section follows the increasing trend at 700–800 nm, which matches well with the experiment. The strong quantum confinement due to the small size or features of nanoparticles also give rise to phonon-assisted Landau damping (intraband transition) hot carriers. These are usually considered a bonus to the classic absorption cross-section, but one can consider adding the corresponding term into the dielectric constant of the metal just like the Drude electron term and Lorentzian oscillators for interband transition (or another model such as Brendal–Bormann, and quantum mechanical model, etc.). [106, 107]. The semi-classical theory can easily decouple the hot electron cross-section from direct photoexcitation (interband transition), Landau damping (intraband transition), and photothermal effects by comparing the spectra with the wavelength-dependent reactivity or quantum efficiency.

Numerous challenges remain in electromagnetic calculations, such as taking into account the detailed picture of the charge transfer of hot electrons to the anti-bonding orbitals of the molecule (or hot holes to bonding orbitals). An illustrative example is the utilization of a Lorentzian function in hot carrier integration. This function effectively characterizes the energy position distribution of a molecular adsorbate, enabling hot carriers produced at specific energy levels to be injected into the adsorbates, consequently prompting the desired chemical reactions [79, 108]. Whether the reaction rate or external quantum efficiency should be chosen to correlate the calculated hot electron cross-section (or rate) depends on how the hot carriers impacts the conversion of molecules and whether it is one hot electron corresponding to one molecule or if carrier multiplication also contributes to the reaction [85, 109]. Moving forward, further developments in high-accuracy quantum chemistry calculations, ultrafast pump-probe measurements, and in-operando measurements of reaction intermediates will be needed to obtain accurate information on these processes.