Recent Progress in Plasmonic Hybrid Photocatalysis for CO 2 Photoreduction and C–C Coupling Reactions

: Plasmonic hybrid nanostructures have been investigated as attractive heterogeneous photocatalysts that can utilize sunlight to produce valuable chemicals. In particular, the efﬁcient photoconversion of CO 2 into a stable hydrocarbon with sunlight can be a promising strategy to achieve a sustainable human life on Earth. The next step for hydrocarbons once obtained from CO 2 is the carbon–carbon coupling reactions to produce a valuable chemical for energy storage or ﬁne chemicals. For these purposes, plasmonic nanomaterials have been widely investigated as a visible-light-induced photocatalyst to achieve increased efﬁciency of photochemical reactions with sunlight. In this review, we discuss recent achievements involving plasmonic hybrid photocatalysts that have been investigated for CO and CO 2 photoreductions to form multi-carbon products and for C–C coupling reactions, such as the Suzuki–Miyaura coupling reactions.


Introduction
Global population growth and industrial development have been continuously causing the consumption of fossil fuels, resulting in environmental pollution and energy shortages. Among various alternative energy sources, sun light is an eco-friendly, clean, and sustainable energy source that can produce more than tens of thousands of terawatts from the Earth's surface [1,2]. The sun light can be utilized in various ways in producing electricity, photochemical synthesis in plants, and a giant heat source to maintain biological systems on Earth [3][4][5][6][7]. Therefore, utilizing sun light as an energy source has been an attractive research topic in chemistry and materials sciences. Among many applications with light, mimicking the photosynthetic system with a catalyst and sun light will be greatly attracting and challenging areas.
In this regard, the efficient conversion of CO 2 into hydrocarbon with sun light is one plausible way to reduce the amount of CO 2 at atmosphere by producing stable chemicals, which can be utilized as an energy source when necessary [8,9]. To achieve this challenging goal, the first step is an efficient conversion of CO 2 into hydrocarbon, either C 1 or C 2 . The next step is carbon-carbon coupling to produce multi-carbon products. Although these two types of reactions have different aspects, the reactions have the commonality of not only forming carbon-to-carbon bonds, but also their utility in the production of materials that can be used in other fields.
The first developed method for CO 2 conversion is the electrochemical approach, traced back to the 19th century [10]. The electrochemical method has the advantages of flexibility in the design of devices and individual optimization of components, but still has the disadvantage of requiring external energy (electricity). Subsequently, the photocatalytic approach was proposed, which was traced back to the 1970s [11][12][13][14]. Inspired The conversion of CO2 into C2 hydrocarbon is one type of a C-C coupling reaction [9,18]. However, the reaction requires a multi-step reaction involving multiple electrons. It is more difficult to conduct than the simple CO2 conversion reaction. Therefore, the understanding and development of suitable catalysts, reaction conditions, etc. are essentially required to improve the efficiency of CO2 conversion to multi-carbon products.
Carbon-carbon cross-couplings and related reactions, which are the other types of C-C coupling, present an important research direction in the field of chemistry [19][20][21][22][23]. After the reports of C-C coupling reactions in the 1970s [24][25][26][27], the reactions have been regarded as very powerful methods for forming C-C and C-heteroatom bonds. In recognition of their developments and applications conducted in the 1990s [28][29][30][31], in 2010, R. F. Heck, E.-I. Negishi, and A. Suzuki were awarded the Nobel Prize in Chemistry [32]. with other materials, such as carbon nanoparticles and reduced graphene oxide to convert CO 2 into formic acid, but not to produce multi-carbon products [69,74].
The formation of more valuable multi-carbon products requires multiple electronhole pairs involved in CO 2 reduction [75]. To generate abundant charge carriers, Tu et al. introduced AuNPs, showing a strong LSPR effect, into a TiO 2 hollow shell, which were called Au@TiO 2 yolk-shell hollow spheres [76]. According to this report, bare TiO 2 reduced CO 2 to produce CH 4 (1.33 µmoL g −1 h −1 ), but not C 2 H 6 , while the Au@TiO 2 yolk-shell generated both CH 4 and C 2 H 6 , indicating the significant role of AuNPs in enhancing the photocatalytic yield and the generation of multi-carbon species with rates of 2.52 µmoL g −1 h −1 to produce CH 4 and 1.67 µmoL g −1 h −1 to produce C 2 H 6 , respectively, accelerating multiple e − /h + reactions [76].
Yu et al. investigated product selectivity tuned by a light excitation attribute, which is a type of plasmonic control with polyvinylpyrrolidone (PVP)-capped AuNPs (11.8 ± 2.3 nm) [77]. They determined that higher photon energies (shorter wavelength) and flux (light intensity) tend to produce C 2 H 6 rather than CH 4 , showing C 2 H 6 selectivity ( Figure 2) [77]. Specifically, CH 4 production rates increased proportionally at higher photon energies, that is, shorter excitation wavelengths, but C 2 H 6 was produced only at higher photon energies at a fixed laser intensity of 150 mW·cm −2 (Figure 2b). In addition, CH 4 production rates increased almost linearly with increasing photon flux, i.e., light intensity (~0.5 NP −1 h −1 of turnover frequency (TOF) under 532 nm, >0.9 NP −1 h −1 of turnover frequency under 488 nm), irrespective of the wavelength of light, and C 2 H 6 generation was observed only at 488 nm of wavelength and above 300 mW cm −2 (~0.6 NP −1 h −1 of TOF at 750 mW cm −2 ), indicating the presence of a threshold intensity for C 2 H 6 production (Figure 2c,d).
ers to use AuNPs as photocatalysts [72,73]. There have been attempts to use AuNPs together with other materials, such as carbon nanoparticles and reduced graphene oxide to convert CO2 into formic acid, but not to produce multi-carbon products [69,74].
The formation of more valuable multi-carbon products requires multiple electronhole pairs involved in CO2 reduction [75]. To generate abundant charge carriers, Tu et al. introduced AuNPs, showing a strong LSPR effect, into a TiO2 hollow shell, which were called Au@TiO2 yolk-shell hollow spheres [76]. According to this report, bare TiO2 reduced CO2 to produce CH4 (1.33 μmoL g −1 h −1 ), but not C2H6, while the Au@TiO2 yolkshell generated both CH4 and C2H6, indicating the significant role of AuNPs in enhancing the photocatalytic yield and the generation of multi-carbon species with rates of 2.52 μmoL g −1 h −1 to produce CH4 and 1.67 μmoL g −1 h −1 to produce C2H6, respectively, accelerating multiple e − /h + reactions [76].
Yu et al. investigated product selectivity tuned by a light excitation attribute, which is a type of plasmonic control with polyvinylpyrrolidone (PVP)-capped AuNPs (11.8 ± 2.3 nm) [77]. They determined that higher photon energies (shorter wavelength) and flux (light intensity) tend to produce C2H6 rather than CH4, showing C2H6 selectivity ( Figure  2) [77]. Specifically, CH4 production rates increased proportionally at higher photon energies, that is, shorter excitation wavelengths, but C2H6 was produced only at higher photon energies at a fixed laser intensity of 150 mW·cm −2 (Figure 2b). In addition, CH4 production rates increased almost linearly with increasing photon flux, i.e., light intensity (~0.5 NP −1 h −1 of turnover frequency (TOF) under 532 nm, >0.9 NP −1 h −1 of turnover frequency under 488 nm), irrespective of the wavelength of light, and C2H6 generation was observed only at 488 nm of wavelength and above 300 mW cm −2 (~0.6 NP −1 h −1 of TOF at 750 mW cm −2 ), indicating the presence of a threshold intensity for C2H6 production (Figure 2c,d).  Zhao et al. also conducted plasmonic control of CO2 conversion using metal/Zn photocatalysts ( Figure 3) [78]. They revealed that the production of higher levels of h drocarbons, such as C2H6, is closely related to the coupling of the surface plasmon res nance (SPR) field with the intrinsic inner electric field, enabling the separation of electron hole pairs and the polarization and activation of absorbed substrates. They found that A interacts more strongly with semiconductors than Ag and Pd, which alters the molecula pathway of CO2 conversion, resulting in a tremendous change in the selectivity of prod ucts by density functional theory (DFT) calculations, electron paramagnetic resonanc spectroscopy, and Raman spectroscopy. By putting their results all together, they dete mined that only Au/ZnO can produce C2H6 with a rate of ~25 μmol g −1 h −1 (Figure 3b while Ag/ZnO and Pd/ZnO produce CH4 and CO without multi-carbon products (Figu 3c,d) [78]. (a) Schematic illustration of the photocatalytic conversion of CO2 to C2H6 using Au/Zn nanosheets, and the hydrocarbon production rates from CO2 with loading plasmonic metals of ( Au, (c) Ag, and (d) Pd to ZnO. Reproduced with permission from Reference [78]. Copyright 201 Elsevier.
Nguyen et al. used metal-organic frameworks (MOFs) with AuNPs for photocat lytic CO2 conversion to methanol and ethanol [79]. They observed the effect of Au loadin on the Aux@zeolitic imidazolate framework (ZIF-67) for the reaction in which Au10@ZIF 67 showed the highest performance of CH3OH production at a rate of 1623 μmol g −1 h while Au20@ZIF-67 showed that C2H5OH production occurred at a rate of 495 μmol g −1 h (Figure 4a,b), and both CH3OH and C2H5OH products decreased with Au30@ZIF-67, po sibly due to the agglomeration of AuNPs. Because the reaction to form C2H5OH require more electrons than CH3OH, these phenomena are thought to be the result of higher h electrons that enable the photoreduction and support C2H5OH generation due to high Au concentrations [79]. (a) Schematic illustration of the photocatalytic conversion of CO 2 to C 2 H 6 using Au/ZnO nanosheets, and the hydrocarbon production rates from CO 2 with loading plasmonic metals of (b) Au, (c) Ag, and (d) Pd to ZnO. Reproduced with permission from Reference [78]. Copyright 2019 Elsevier. Figure 4. Photocatalytic activity of Aux@ZIF-67 for CO2 reduction to (a) CH3OH and (b) C2H5OH, and (c) the mechanism of CO2 photoreduction. Reproduced with permission from Reference [79]. Copyright 2020 American Chemical Society.

Silver Nanoparticle (AgNP)-Assisted Plasmonic Photocatalysts for CO2 Reduction to Multi-Carbon Products
The silver nanoparticle (AgNP) is also one of the plasmonic materials that can exhibit unique optical properties for photocatalysts under visible light with LSPR, like AuNP [80]. Cai et al. prepared AgClxBr1-x alloy nanocrystals and found that the conduction band level could be affected by varying compositions [81]. The substitution of Cl with Br leads to a narrower band gap due to a negative shift in the conduction band minimum. Accordingly, AgCl0.75Br0.25 exhibited higher photocatalytic efficiency for CO2 reduction into both CH3OH (181 μmol g −1 ) and C2H5OH (362 μmol g −1 ) than any other AgClxBr1-x with different compositions. Furthermore, the amplified electric field due to the LSPR of Ag 0 species also contributes to the light enhancement by encouraging the photocatalytic reaction. Cai et al. further developed an Ag/AgCl photocatalyst system with a coaxial tri-cubic morphology, called red Ag/AgCl [82]. The enhancement of light harvesting property with red Ag/AgCl was observed, compared to normal AgCl, due to the synergistic effect between metallic AgNPs and the n-type AgCl semiconductor, which is the featured LSPR, Schottky junction, and polarization effect induced by surface plasmons. As a result, the CH3OH and C2H5OH yields and apparent quantum efficiency for the red Ag/AgCl catalysts were 146 and 223 μmol g −1 for 5 h, respectively, in which both are higher than those for normal AgCl catalysts ( Figure 5) [82].  . Photocatalytic activity of Au x @ZIF-67 for CO 2 reduction to (a) CH 3 OH and (b) C 2 H 5 OH, and (c) the mechanism of CO 2 photoreduction. Reproduced with permission from Reference [79]. Copyright 2020 American Chemical Society.

Silver Nanoparticle (AgNP)-Assisted Plasmonic Photocatalysts for CO 2 Reduction to Multi-Carbon Products
The silver nanoparticle (AgNP) is also one of the plasmonic materials that can exhibit unique optical properties for photocatalysts under visible light with LSPR, like AuNP [80]. Cai et al. prepared AgCl x Br 1-x alloy nanocrystals and found that the conduction band level could be affected by varying compositions [81]. The substitution of Cl with Br leads to a narrower band gap due to a negative shift in the conduction band minimum. Accordingly, AgCl 0.75 Br 0.25 exhibited higher photocatalytic efficiency for CO 2 reduction into both CH 3 OH (181 µmol g −1 ) and C 2 H 5 OH (362 µmol g −1 ) than any other AgCl x Br 1-x with different compositions. Furthermore, the amplified electric field due to the LSPR of Ag 0 species also contributes to the light enhancement by encouraging the photocatalytic reaction. Cai et al. further developed an Ag/AgCl photocatalyst system with a coaxial tricubic morphology, called red Ag/AgCl [82]. The enhancement of light harvesting property with red Ag/AgCl was observed, compared to normal AgCl, due to the synergistic effect between metallic AgNPs and the n-type AgCl semiconductor, which is the featured LSPR, Schottky junction, and polarization effect induced by surface plasmons. As a result, the CH 3 OH and C 2 H 5 OH yields and apparent quantum efficiency for the red Ag/AgCl catalysts were 146 and 223 µmol g −1 for 5 h, respectively, in which both are higher than those for normal AgCl catalysts ( Figure 5) [82].

Silver Nanoparticle (AgNP)-Assisted Plasmonic Photocatalysts for CO2 Reduction to Multi-Carbon Products
The silver nanoparticle (AgNP) is also one of the plasmonic materials that can exhibit unique optical properties for photocatalysts under visible light with LSPR, like AuNP [80]. Cai et al. prepared AgClxBr1-x alloy nanocrystals and found that the conduction band level could be affected by varying compositions [81]. The substitution of Cl with Br leads to a narrower band gap due to a negative shift in the conduction band minimum. Accordingly, AgCl0.75Br0.25 exhibited higher photocatalytic efficiency for CO2 reduction into both CH3OH (181 μmol g −1 ) and C2H5OH (362 μmol g −1 ) than any other AgClxBr1-x with different compositions. Furthermore, the amplified electric field due to the LSPR of Ag 0 species also contributes to the light enhancement by encouraging the photocatalytic reaction. Cai et al. further developed an Ag/AgCl photocatalyst system with a coaxial tri-cubic morphology, called red Ag/AgCl [82]. The enhancement of light harvesting property with red Ag/AgCl was observed, compared to normal AgCl, due to the synergistic effect between metallic AgNPs and the n-type AgCl semiconductor, which is the featured LSPR, Schottky junction, and polarization effect induced by surface plasmons. As a result, the CH3OH and C2H5OH yields and apparent quantum efficiency for the red Ag/AgCl catalysts were 146 and 223 μmol g −1 for 5 h, respectively, in which both are higher than those for normal AgCl catalysts ( Figure 5) [82].   Li et al. suggested a system with a plasmonic photocatalyst other than the C-C coupling of CO 2 alone by adopting the oxidative coupling of methane using CO 2 as the oxidant [83]. In the Ag/TiO 2 system, AgNPs can absorb visible light and generate hot electrons and holes, and the hot electrons are injected into TiO 2 while hot holes should be captured by CH 4 . Otherwise, it will lead to the accumulation of Ag(I), which can be reduced back to Ag(0) under UV light. At the same time, TiO 2 can generate photoexcited electron-hole pairs, and photoexcited holes can combine with hot electrons from Ag, while photoexcited electrons can reduce CO 2 adsorbed on TiO 2 . This synergistic effect enhances the photocatalytic activity of the reaction (1149 µmoL g −1 h −1 for CO and 686 µmol g −1 h −1 for C 2 H 4 ) and contributes to the high stability of the catalyst (Figure 6b) [83]. They also demonstrated that other types of support materials with Ag did not present any synergistic effect due to the formation of an unsuitable Schottky barrier and CO 2 adsorption sites (Figure 6c), and other precious metals with TiO 2 showed lower activity for the photocatalytic reaction due to the poor SPR effect of the metals (Figure 6d) [83]. Li et al. suggested a system with a plasmonic photocatalyst other than the C-C coupling of CO2 alone by adopting the oxidative coupling of methane using CO2 as the oxidant [83]. In the Ag/TiO2 system, AgNPs can absorb visible light and generate hot electrons and holes, and the hot electrons are injected into TiO2 while hot holes should be captured by CH4. Otherwise, it will lead to the accumulation of Ag(I), which can be reduced back to Ag(0) under UV light. At the same time, TiO2 can generate photoexcited electron-hole pairs, and photoexcited holes can combine with hot electrons from Ag, while photoexcited electrons can reduce CO2 adsorbed on TiO2. This synergistic effect enhances the photocatalytic activity of the reaction (1149 μmoL g −1 h −1 for CO and 686 μmol g −1 h −1 for C2H4) and contributes to the high stability of the catalyst (Figure 6b) [83]. They also demonstrated that other types of support materials with Ag did not present any synergistic effect due to the formation of an unsuitable Schottky barrier and CO2 adsorption sites (Figure 6c), and other precious metals with TiO2 showed lower activity for the photocatalytic reaction due to the poor SPR effect of the metals (Figure 6d) [83].

Copper Nanoparticle (CuNP)-Assisted Photocatalytic CO 2 Reduction to Multi-Carbon Products
Metallic Cu species, including CuNPs, have been used as cocatalysts due to their known effectiveness in generating not only C 1 products, but also multi-carbon organic compounds, not as light absorbers with the LSPR effect [84][85][86][87]. Therefore, Cu has been used as an electron acceptor and suppressor of recombination of photoexcited electron-hole pairs generated from semiconductors [88][89][90]. Shown et al. prepared CuNP-decorated graphene oxide because of the large work function of Cu compared to that of GO, producing methanol and acetaldehyde [88]. They controlled the production rates and the ratio of and between both products by adjusting the work function of Cu/GO hybrids [88]. Park et al. used trititanate nanotubes (TNTs) decorated with Cu and CdS quantum dots (CdS/Cu-TNTs) for the production of C 1 -C 3 hydrocarbons [89]. When irradiated with light, an efficient reduction of CO 2 to C 1 -C 3 hydrocarbons was observed by the transport of photogenerated electrons to the Cu part through the TNTs while photogenerated holes oxidize water, which is similar to artificial photosynthesis [89]. Chen et al. carried out photocatalytic CO 2 reduction with benzyl alcohol oxidation to benzyl acetate using Cu 2 O/Cu nanocomposites due to the narrow direct band gap and the position of the conduction band of Cu 2 O [90]. Electrons and holes are generated over Cu 2 O with visible light irradiation, and the electrons transfer to the surface of Cu due to their lower work function, while the holes can react with benzyl alcohol to form benzaldehyde, followed by the subsequent coupling reaction to form benzyl acetate [90]. CuNPs have great potential for photocatalytic C-C coupling, but more research is needed to utilize CuNPs for light harvesting materials.
There have been several attempts to produce value-added fuels through CO 2 conversion using plasmonic hybrid photocatalysts, even though only a few studies have been conducted to generate single carbon products. The hybrid structures of the photocatalysts, reaction conditions, and yields are summarized in Table 1.

Plasmonic Hybrid Photocatalysts for C-C Cross-Coupling
The Suzuki-Miyaura coupling, which is one of the most powerful methods for carboncarbon cross-coupling, is the reaction between an organoboron species and aryl halide in the presence of a palladium (Pd) catalyst and base [21,91]. Although most of this reaction involves a Pd catalyst, which has been regarded as a catalytically active site for the reaction, Pd is difficult to use as a photocatalyst because of its low absorption of visible light [92]. In recent years, to remedy this shortcoming, plasmonic nanoparticles and semiconductors that can absorb visible light have been used. In particular, plasmonic NPs (Au, Ag, Cu) are useful materials that can help the reaction of the Pd catalyst because they interact with visible light to show a strong LSPR phenomenon [93][94][95]. In addition, semiconductors are introduced into plasmonic materials, so-called plasmonic hybrid structures, to help plasmonic NPs extend the lifetime of hot electrons excited by light, or to form electronhole pairs [96,97]. The hybrid structures classified by each type of plasmonic NPs for photocatalytic C-C cross-coupling reactions are described below.

AuNP-Assisted Plasmonic Photocatalysts for C-C Cross-Coupling
For C-C cross-coupling reactions, Au-Pd nanocomposites can be utilized, where the Au part absorbs visible light and transfers hot electrons into Pd, and the Pd part acts as electron acceptors and active sites. Some studies using Au-Pd alloys without nonplasmonic materials, such as AuPd nano-wheels [98] and AuPd nanotriangles [99] for photocatalytic C-C cross-coupling, have been reported (Figure 7). Huang et al. prepared AuPd nano-wheels, in which Pd encircles an Au core, with a controllable edge length and tunable SPR using a facile wet-chemical reduction method. In this work, the photocatalytic efficiency of the nano-wheel-catalysts for benzyl alcohol conversion and Suzuki coupling was confirmed, and the yield of products was 65.8% at 50 • C for 1 h under visible light (Figure 7a) [98]. Gangishetty et al. synthesized AuPd bimetallic nanotriangles consisting of an Au nanotriangle core with an unevenly distributed Pd shell, which is similar in morphology with AuPd nano-wheels [99]. The nano-catalysts showed >80% yield of Suzuki coupling between p-iodobenzoic acid and phenylboronic acid for 5 h under a green LED, accompanied by an increase in the temperature from 25 • C to 37 • C while the dark reaction showed only ca. 35% lower conversion compared to the light reaction ( Figure 7b). Compared to the yield under dark reaction at 37 • C (ca. 75%), the pure photocatalytic effect is not significant, while the photothermal effect generated from non-radiative plasmon decay is the primary factor (Figure 7b) [99].
Sarina, Xiao et al. prepared Au-Pd alloy NPs embedded on ZrO 2 , which has a band gap of approximately 5 eV, exhibiting negligible visible light absorption above 400 nm, so that the ZrO 2 support does not contribute to the photocatalytic activity of C-C crosscoupling ( Figure 8) [100][101][102][103]. Several C-C coupling reactions such as Suzuki-Miyaura, Sonogashira, Stille, Hiyama, Buchwald-Hartwig, and Ullmann coupling, were conducted to study the effects of the wavelength and intensity of the light, the Au/Pd molar ratio of the alloy NPs on the photocatalytic activity, and the photocatalytic mechanism for the C-C coupling reactions. The researchers described the photocatalytic process in view of reaction kinetics. The reduced activation energy for the C-C coupling reaction is possible by visiblelight absorption of photocatalyst, indicating the low activation energy in the photocatalytic process compared with that of a thermal reaction process (Figure 8b,c) [101,102]. In terms of energy levels of the molecular orbital, there are two light absorption mechanisms such as inter-band excitation and LSPR absorption at 530 nm for AuNPs. Inter-band excitation is possible with short-wavelength (e.g., 400 nm) absorption via a single-electron excitation. When irradiated with short-wavelength, single-photon excitation generates hot electrons to be injected into the lowest unoccupied molecular orbital (LUMO). In the case irradiated with longer wavelengths, the hot electrons generated by the collective excitation of LSPR can only induce reactions with lower energy thresholds (Figure 8d) [103]. In spite of the low TOF and the number of conversion reactions, they provided a comprehensive insight for the photocatalytic reactions for various C-C cross-coupling reactions and explained the kinetics and mechanisms of the reactions.
Other researchers have used a variety of support materials such as semiconductors (e.g., CeO 2 , graphitic carbon nitride (g-C 3 N 4 )) [104,105], metal-organic frameworks (MOF, UiO-66-NH 2 ) [106], polymers (e.g., perylene bisimide (PBI), polystyrene) [107,108], wide band gap semiconductors (TiO 2 ) [109][110][111][112], and silica [113]. Semiconductors that have narrower band gaps than 3.1 eV (i.e., 2.7 eV for g-C 3 N 4 ) can absorb visible light to generate electron-hole pairs, which are transferred to metal NPs. By itself, a semiconductor can help the Pd catalyst in photocatalytic C-C cross-coupling reactions [114][115][116], but even the synergy of plasmonic hybrid with semiconductors can be expected to have greater photocatalytic efficiency for a C-C cross-coupling reaction. In these systems, the hot electrons generated from AuNPs (due to the strong LSPR effect of AuNPs) using irradiation can be injected into the attached Pd, in which the electrons transfer to aryl halide molecules, while the photogenerated electron-hole pairs of semiconductors by irradiation can be separated, causing electrons to recover into the Au 0 state and holes to transfer into solvent or phenylboronic acid to be activated ( Figure 9) [104,105]. Moreover, semiconductors such as g-C 3 N 4 are two-dimensional materials with large surface areas and unique electronic/optical properties, and Schottky junctions form at the interface of the metal and semiconductor, relying on the band alignment and work function [114][115][116], leading to a positive effect on catalytic efficiency (Figure 9b) [105]. Therefore, the AuPd/g-C 3 N 4 nanohybrid showed a very high TOF of 7920 h −1 in the C-C cross-coupling reaction.  Other researchers have used a variety of support materials such as semiconductors (e.g., CeO2, graphitic carbon nitride (g-C3N4)) [104,105], metal-organic frameworks (MOF, UiO-66-NH2) [106], polymers (e.g., perylene bisimide (PBI), polystyrene) [107,108], wide band gap semiconductors (TiO2) [109][110][111][112], and silica [113]. Semiconductors that have narrower band gaps than 3.1 eV (i.e., 2.7 eV for g-C3N4) can absorb visible light to generate electron-hole pairs, which are transferred to metal NPs. By itself, a semiconductor can help the Pd catalyst in photocatalytic C-C cross-coupling reactions [114][115][116], but even the  MOFs such as UiO-66-NH2 are similar to semiconductors in that they have a band gap (approximately 2.67 eV) capable of absorbing visible light. The key differences are that the energy transfer occurs from ligands to a metal. The pore volume and surface area can be controlled by using the functional group. Noble metals can be introduced into MOFs, and they exhibit high dispersion stability owing to the ultra-high surface area [106]. MOFs such as UiO-66-NH 2 are similar to semiconductors in that they have a band gap (approximately 2.67 eV) capable of absorbing visible light. The key differences are that the energy transfer occurs from ligands to a metal. The pore volume and surface area can be controlled by using the functional group. Noble metals can be introduced into MOFs, and they exhibit high dispersion stability owing to the ultra-high surface area [106]. In general, wide band gap semiconductors such as TiO2 display great light at wavelengths below 400 nm, making it difficult to expect visible light abs catalytic reactions [37]. Furthermore, the high energy of UV light might oxid duce organic reactants, which can result in low yield and selectivity. Rohani et a the morphology of TiO2 into urchin-like yolk-shell structures with unique such as a high surface area and visible-light harvesting ( Figure 10) [112]. The ated urchin-like yolk@shell TiO2 structure (HUY@S-TOH) was decorated with Au/Pd NPs for photocatalytic Suzuki coupling. The structure showed absorptio for visible light because of the Ti 3+ species on the surface of the structure an enhancing the light harvesting efficiency and the inhibition of the recombinat togenerated electron-hole pairs by decreasing the band gap of TiO2 to the visi In addition to the strong interaction between noble metals and TiO2-x, HUY@S-T showed a high catalytic efficiency with a TOF value of 7095 h −1 . In general, wide band gap semiconductors such as TiO 2 display great light absorption at wavelengths below 400 nm, making it difficult to expect visible light absorption for catalytic reactions [37]. Furthermore, the high energy of UV light might oxidize and reduce organic reactants, which can result in low yield and selectivity. Rohani et al. changed the morphology of TiO 2 into urchin-like yolk-shell structures with unique properties, such as a high surface area and visible-light harvesting ( Figure 10) [112]. The hydrogenated urchinlike yolk@shell TiO 2 structure (HUY@S-TOH) was decorated with plasmonic Au/Pd NPs for photocatalytic Suzuki coupling. The structure showed absorption property for visible light because of the Ti 3+ species on the surface of the structure and AuNPs, enhancing the light harvesting efficiency and the inhibition of the recombination of photogenerated electron-hole pairs by decreasing the band gap of TiO 2 to the visible region. In addition to the strong interaction between noble metals and TiO 2-x , HUY@S-TOH/AuPd showed a high catalytic efficiency with a TOF value of 7095 h −1 .

R PEER REVIEW
Au/Pd NPs for photocatalytic Suzuki coupling. The structure showed absorption property for visible light because of the Ti 3+ species on the surface of the structure and AuNPs, enhancing the light harvesting efficiency and the inhibition of the recombination of photogenerated electron-hole pairs by decreasing the band gap of TiO2 to the visible region. In addition to the strong interaction between noble metals and TiO2-x, HUY@S-TOH/AuPd showed a high catalytic efficiency with a TOF value of 7095 h −1 .  Graphene and its derivatives are 2D materials with an exceptional electron mobility of 2 × 10 5 cm 2 ·v −1 s −1 and variable band gap that depends on the oxidation state of graphene [69,117,118]. Hybrid structures containing graphene or its oxide have been reported for photocatalytic C-C cross-coupling to extend the lifetime of hot electrons of plasmonic NPs generated by light absorption or to prevent recombination of electron-hole pairs [119,120]. Moreover, graphene or slightly oxidized graphene on its own can also be a good support material not only for combination with Pd 2+ or Pd 0 NPs owing to its functional group, but also for transferring electrons into Pd with facilities to enhance catalytic activity [121]. Kang et al. studied the effect of interfaces on Pd-nanodot-decorated AuNPs with a graphene layer with different oxidation states for photocatalytic Suzuki coupling ( Figure 11) [119]. They prepared Pd-cys-AuNPs, Pd-GO-AuNPs, and Pd-rGO-AuNPs with different oxidation states, and found that Pd-rGO-AuNPs exhibited the fastest reaction progress (66.4% with a thermal effect and 54.4% without a thermal effect), while the Pd-cys-AuNPs exhibited the slowest reaction progress (30% with a thermal effect and 6.7% without a thermal effect) for 2 h (Figure 11b,c). The contribution of the electron transfer mechanism from plasmonic NPs (AuNPs) to Pd nanodots is significant through the graphene interface, preventing the relaxation of hot electrons of plasmonic NPs induced by light [119].
Plasmonic properties can be tuned by size and shape by utilizing varying wavelengths of light [46,122]. Gold nanorods (AuNRs), like spherical AuNPs and other noble metal NPs, have the ability to interact with light of varying wavelengths through LSPR [46,[123][124][125]. They display two SPR bands of transverse and longitudinal bands due to their anisotropic shape. The transverse mode is located near 500 nm, while the longitudinal mode varies widely depending on the aspect ratio and the overall size of AuNRs, generally located in the NIR region. Therefore, AuNRs have been used as NIR-responsive photocatalysts [122,126]. [119]. They prepared Pd-cys-AuNPs, Pd-GO-AuNPs, and Pd-rGO-AuNPs with different oxidation states, and found that Pd-rGO-AuNPs exhibited the fastest reaction progress (66.4% with a thermal effect and 54.4% without a thermal effect), while the Pd-cys-AuNPs exhibited the slowest reaction progress (30% with a thermal effect and 6.7% without a thermal effect) for 2 h (Figure 11b,c). The contribution of the electron transfer mechanism from plasmonic NPs (AuNPs) to Pd nanodots is significant through the graphene interface, preventing the relaxation of hot electrons of plasmonic NPs induced by light [119]. Plasmonic properties can be tuned by size and shape by utilizing varying wavelengths of light [46,122]. Gold nanorods (AuNRs), like spherical AuNPs and other noble metal NPs, have the ability to interact with light of varying wavelengths through LSPR [46,[123][124][125]. They display two SPR bands of transverse and longitudinal bands due to their anisotropic shape. The transverse mode is located near 500 nm, while the longitudi- AuNR-based photocatalysts have been applied to C-C coupling reactions [127][128][129][130][131]. Wang et al. synthesized Au-Pd nanostructures where Pd NPs were located at the tip of the AuNRs to a large extent as well as on the whole surface of the AuNRs (Figure 12a). They conducted photocatalytic Suzuki coupling under an 809-nm laser in the NIR region, yielding biphenyl products, where they demonstrated the two origins of the catalytic activity of Au-Pd nanostructures, which is the plasmonic heating process and plasmon-excitationinduced hot electrons [127]. Guo et al. developed Au-Pd nanostructures, called Au@Pd superstructures formed from different directing agents, investigating the effect of the shapes of Pd, which are superstructures, nano-dendrites, and shell structures (Figure 12b). Different from AuNRs, the Au@Pd core@shell exhibited weak electric field enhancement due to the plasmon shielding effect of Pd shells. In the case of Au@Pd nano-dendrites with discrete Pd surfaces, the |E|/|E 0 | around their surface was largely enhanced by 16 of the maximum value of |E max |/|E 0 |, which allows the inside of the AuNR core to be excited partially. The Au@Pd superstructures displayed further improvement of |E max |/|E 0 | up to 23 because of the ordered open structure of the Pd nano-arrays and their strong plasmonic antenna effect. Accordingly, the plasmon-enhanced photocatalytic activity of Au@Pd superstructures was >4 times higher than that of the Au@Pd core@shell (TOF ≈ 2880 h −1 ), and approximately two times higher than that of the Au@Pd nanodendrites [129]. Yoshii (Figure 12c). Almost all aspects of Pd-graphene-AuNR nanocomposites are similar to Pd-rGO-Au, and the only difference is that AuNR was used to absorb NIR light [131].
(TOF ≈ 2,880 h ), and approximately two times higher than that of the Au@Pd nanodendrites [129]. Yoshii et al. designed Pd-graphene-AuNR nanocomposite catalysts similar to Pd-rGO-Au by Kang et al., facilitating the transfer of SPR-induced hot electrons by AuNR to the catalytic active metals (Pd) through the graphene layer (Figure 12c). Almost all aspects of Pd-graphene-AuNR nanocomposites are similar to Pd-rGO-Au, and the only difference is that AuNR was used to absorb NIR light [131].

AgNP-Assisted Plasmonic Photocatalysts for C-C Cross-Coupling
AgNP-based hybrid photocatalysts with various support materials such as graphene oxide with AgBr [132], silica [113,133], supramolecular ensemble with Cu 2 O [134,135], and TiO 2 [136] for C-C cross-coupling have been reported. Verma et al. used mesoporous silica as a support material (SBA-15, Aldrich), of which the channel can contain Pd/Ag NPs with stability for photocatalytic Suzuki coupling [113,133]. They observed the effect of the shape of Ag or type of plasmonic NPs (Au or Ag), affecting plasmonic optical properties and catalytic activity ( Figure 13). They found that longer aspect ratios of the Ag nanostructure result in higher photocatalytic activity efficiency due to the light absorption property (<30% for Pd/Ag/SBA-15(Y),~40% for Pd/Ag/SBA-15(R), 53% for Pd/AgSBA-15(B)) ( Figure 13b) [133]. Although the photocatalytic activity for Suzuki coupling with Pd/Au/SBA-15 was better than that for Suzuki coupling with Pd/Ag/SBA-15 (~70% for Pd/Au/SBA-15, and~40% for Pd/Ag/SBA-15) (Figure 13d), the activity with Pd/Ag/SBA-15 for the dehydrogenation reaction was better than that with Pd/Au/SBA-15, indicating that it is difficult to simply compare the effects of those reactions [113]. Putting it all together, it is meaningful that Ag and Au can work complementarily with each other in terms of light absorption. ure 13b) [133]. Although the photocatalytic activity for Suzuki coupling with Pd/Au/SBA-15 was better than that for Suzuki coupling with Pd/Ag/SBA-15 (~70% for Pd/Au/SBA-15, and ~40% for Pd/Ag/SBA-15) (Figure 13d), the activity with Pd/Ag/SBA-15 for the dehydrogenation reaction was better than that with Pd/Au/SBA-15, indicating that it is difficult to simply compare the effects of those reactions [113]. Putting it all together, it is meaningful that Ag and Au can work complementarily with each other in terms of light absorption.  Bhalla's group used supramolecular ensembles as both reactors and stabilizers of NPs to a higher extent through electron-rich assemblies and introduced Cu 2 O as a shell around the AgNP for its affordable price, stability, and property. Cu 2 O is a p-type semiconductor and has been used as an efficient catalyst for C-C, C-N, and C-O cross-coupling reactions ( Figure 14) [134,135]. These reports are not studies for the development of conventional Suzuki coupling, but they are significant in that they were performed without palladium even though they showed relatively low photocatalytic efficiency.

CuNP-Assisted Plasmonic Photocatalysts for C-C Cross-Coupling
Copper nanoparticles (CuNPs) are plasmonic materials like AuNPs and AgNPs, which possess unique optical properties [137][138][139]. However, their instability and tendency to undergo surface oxidation make it difficult for many researchers to utilize CuNPs [140]. Nevertheless, there have been some attempts to overcome the instability in order to use CuNPs for photocatalysts because of their low cost [141][142][143].
NPs to a higher extent through electron-rich assemblies and introduced Cu2O as a shell around the AgNP for its affordable price, stability, and property. Cu2O is a p-type semiconductor and has been used as an efficient catalyst for C-C, C-N, and C-O cross-coupling reactions (Figure 14) [134,135]. These reports are not studies for the development of conventional Suzuki coupling, but they are significant in that they were performed without palladium even though they showed relatively low photocatalytic efficiency.

CuNP-Assisted Plasmonic Photocatalysts for C-C Cross-Coupling
Copper nanoparticles (CuNPs) are plasmonic materials like AuNPs and AgNPs, which possess unique optical properties [137][138][139]. However, their instability and tendency to undergo surface oxidation make it difficult for many researchers to utilize CuNPs [140]. Nevertheless, there have been some attempts to overcome the instability in order to use CuNPs for photocatalysts because of their low cost [141][142][143].
For C-C cross-coupling with CuNPs, CuPd bimetallic alloy NPs with other light-absorbing materials such as silicon carbide (SiC) as a semiconductor and NH2-UiO-66(Zr) as MOF were used under visible light ( Figure 15) [144,145]. Wang et al. prepared PdCu/SiC using a sol-gel and carbon thermal reduction process for photocatalytic Sonogashira reaction under visible light. They proposed a mechanism in which photogenerated electrons transfer to the Pd part, facilitating the cleavage of aryl halides, while photogenerated holes in the CuNPs react with phenylacetylene to form a phenylethynylcopper(I) compound ( Figure 15a) [144]. , with encapsulated bimetallic CuPd nanoclusters inside the cavities of NH2-UiO-66(Zr) via double-solvent impregnation followed by chemical reduction for a Suzuki coupling reaction. The transfer of the photogenerated electrons from the photoexcited NH2-UiO-66(Zr) to the Pd part to form electron-rich Pd was facilitated by metallic Cu acts as an electron mediator, which results in the superior activity of photocatalytic Suzuki coupling, since Cu has a higher Fermi energy level as compared to Pd and lower Fermi energy level as compared to NH2-UiO-66(Zr) (Figure 15b) [145]. These reports do not state the plasmonic light-harvesting property of CuNP itself as a photocatalyst, but attempts to introduce copper into Pd. For C-C cross-coupling with CuNPs, CuPd bimetallic alloy NPs with other lightabsorbing materials such as silicon carbide (SiC) as a semiconductor and NH 2 -UiO-66(Zr) as MOF were used under visible light ( Figure 15) [144,145]. Wang et al. prepared PdCu/SiC using a sol-gel and carbon thermal reduction process for photocatalytic Sonogashira reaction under visible light. They proposed a mechanism in which photogenerated electrons transfer to the Pd part, facilitating the cleavage of aryl halides, while photogenerated holes in the CuNPs react with phenylacetylene to form a phenylethynylcopper(I) compound (Figure 15a) [144]. Sun et al. prepared CuPd@NH 2 -UiO-66(Zr), with encapsulated bimetallic CuPd nanoclusters inside the cavities of NH 2 -UiO-66(Zr) via double-solvent impregnation followed by chemical reduction for a Suzuki coupling reaction. The transfer of the photogenerated electrons from the photoexcited NH 2 -UiO-66(Zr) to the Pd part to form electron-rich Pd was facilitated by metallic Cu acts as an electron mediator, which results in the superior activity of photocatalytic Suzuki coupling, since Cu has a higher Fermi energy level as compared to Pd and lower Fermi energy level as compared to NH 2 -UiO-66(Zr) (Figure 15b) [145]. These reports do not state the plasmonic light-harvesting property of CuNP itself as a photocatalyst, but attempts to introduce copper into Pd.
In order to stabilize CuNPs, Cui et al. introduced graphene to Cu because of the possible change of the electronic structure of Cu by the carbon vacancies or dangling bond in graphene [146,147]. Due to the LSPR effect of CuNPs, the electron density in Cu is polarized, causing charge heterogeneity at the surface of CuNPs with both relatively electron-rich sites and positively charged sites. The electron-rich sites can easily adsorb imidazole molecules and inject into the molecules, facilitating the cleavage of N-H bonds, while positively charged sites can assist to cleave C-B bonds in phenylboronic acid molecules, and, as a result, C-N bonds are formed (Figure 16a). Furthermore, graphene can also absorb light, generating a strong photocurrent, and the work function of graphene, which is lower than that of Cu, causes hot electrons to transfer to Cu from graphene easily, which can result in the collection of energetic electrons at the Cu sites to accelerate the reaction [147]. Bhalla et al. used supramolecular ensembles as reactors and stabilizers of CuNPs, which is similar to supramolecular ensemble-based Ag@Cu 2 O core@shell NPs [148]. They confirmed the existence of the LSPR band of CuNPs for plasmonic photocatalysts, and the hybrid systems showed efficient photocatalytic efficiency for photocatalytic C(sp 2 )-H alkynylation (another type of C-C coupling reaction) (Figure 16b). These reports are not about the named C-C cross-coupling reactions, but they are meaningful in that they make good use of the plasmonic properties of CuNPs as a photocatalyst. In order to stabilize CuNPs, Cui et al. introduced graphene to Cu because of the possible change of the electronic structure of Cu by the carbon vacancies or dangling bond in graphene [146,147]. Due to the LSPR effect of CuNPs, the electron density in Cu is polarized, causing charge heterogeneity at the surface of CuNPs with both relatively electronrich sites and positively charged sites. The electron-rich sites can easily adsorb imidazole molecules and inject into the molecules, facilitating the cleavage of N-H bonds, while positively charged sites can assist to cleave C-B bonds in phenylboronic acid molecules, and, as a result, C-N bonds are formed (Figure 16a). Furthermore, graphene can also absorb light, generating a strong photocurrent, and the work function of graphene, which is lower than that of Cu, causes hot electrons to transfer to Cu from graphene easily, which can result in the collection of energetic electrons at the Cu sites to accelerate the reaction [147]. Bhalla et al. used supramolecular ensembles as reactors and stabilizers of CuNPs, which is similar to supramolecular ensemble-based Ag@Cu2O core@shell NPs [148]. They confirmed the existence of the LSPR band of CuNPs for plasmonic photocatalysts, and the hybrid systems showed efficient photocatalytic efficiency for photocatalytic C(sp 2 )-H alkynylation (another type of C-C coupling reaction) (Figure 16b). These reports are not about the named C-C cross-coupling reactions, but they are meaningful in that they make good use of the plasmonic properties of CuNPs as a photocatalyst. Many attempts have been made to introduce Pd NPs into solid supports to stabilize or assist Pd for catalytic efficiency of C-C cross-coupling, in addition to the examples mentioned above. The hybrid structures of the plasmonic photocatalysts used, reaction conditions, and yields with TOF values for C-C cross-coupling are classified and summarized in Table 2. Many attempts have been made to introduce Pd NPs into solid supports to stabilize or assist Pd for catalytic efficiency of C-C cross-coupling, in addition to the examples mentioned above. The hybrid structures of the plasmonic photocatalysts used, reaction conditions, and yields with TOF values for C-C cross-coupling are classified and summarized in Table 2.

Summary and Outlook
Herein, we reviewed the research and development of plasmonic hybrid nano-catalysts for two types of photocatalytic C-C coupling reactions: C-C coupling in CO 2 reduction to hydrocarbon fuels and C-C cross-coupling in organic chemistry. The C-C coupling in CO 2 reduction into multi-carbon fuels can be promoted with the aid of plasmonic NPs as both light absorbers and affinity sites for reactants. Likewise, Suzuki coupling can also be represented in C-C cross-coupling reactions, including Heck, Sonogashira, Stille, Negishi, and other reactions, typically using Pd as an active catalyst and plasmonic NPs with semiconductors as light-responsive materials to accelerate photocatalytic reactions. As mentioned above, the performances of the photocatalytic CO 2 reduction and C-C cross-coupling reactions are summarized in Tables 1 and 2.
Although plasmonic hybrid nano-catalysts for photocatalytic C-C cross-coupling in CO 2 reduction to hydrocarbon and C-C coupling in organic chemistry have been intensively investigated for their ability to contribute to the fine-chemical industry, energy sector, and environmental fields, and for solving challenges that still exist in synthesizing suitable photocatalysts by reaching high quantum yields for commercialization, controlling photothermal effects may affect reactions, and help understand the mechanism. As for C-C coupling in CO 2 reduction to value-added products, multi-carbon production is still much more difficult than C 1 production because of the requirement of multiple electrons, steps, and low selectivity. In addition, Cu has not been properly utilized as a plasmonic material that is cheaper than gold and silver. Thus, developing new hybrid materials with high efficiency could be one of the possible solutions. As for C-C cross-coupling in organic chemistry, the photocatalytic reaction, even Suzuki coupling, has no unified reaction system, where the efficiency varies widely depending on the reaction conditions, reactants, etc., and integrated mechanism. In addition, only a few studies on other photocatalytic C-C coupling reactions including Heck, Sonogashira, Stille, and Negishi coupling, except Suzuki coupling, have been conducted. Moreover, from an economic point of view, C-C cross-coupling reactions typically require expensive Pd active catalysts, which makes commercialization more difficult, so the Pd-free reaction needs to be actively studied.