Recent Advances of Constructing Metal/Semiconductor Catalysts Designing for Photocatalytic CO2 Hydrogenation

With the development of the world economy and the rapid advancement of global industrialization, the demand for energy continues to grow. The significant consumption of fossil fuels, such as oil, coal, and natural gas, has led to excessive carbon dioxide emissions, causing global ecological problems. CO2 hydrogenation technology can convert CO2 into high-value chemicals and is considered one of the potential ways to solve the problem of CO2 emissions. Metal/semiconductor catalysts have shown good activity in carbon dioxide hydrogenation reactions and have attracted widespread attention. Therefore, we summarize the recent research on metal/semiconductor catalysts for photocatalytic CO2 hydrogenation from the design of catalysts to the structure of active sites and mechanistic investigations, and the internal mechanism of the enhanced activity is elaborated to give guidance for the design of highly active catalysts. Finally, based on a good understanding of the above issues, this review looks forward to the development of future CO2 hydrogenation catalysts.


Introduction
The world's consistent and large-scale dependence on fossil fuels during the past hundred years has had serious repercussions, including energy crises and the deterioration of the global climate, which was caused primarily by increasing atmospheric concentrations of CO 2 [1][2][3][4][5][6][7][8][9]. Though it has historically been discharged directly into the atmosphere as waste, increasing global interest in transitioning away from the use of fossil fuels now forces us to change our previous conceptions by employing carbon dioxide as a recyclable feedstock for the production of renewable synthetic fuels [10][11][12][13][14][15][16][17]. In recent decades, several methods have developed for carbon dioxide reduction, for example, electrocatalytic [18][19][20][21], photocatalytic [22][23][24][25][26][27][28], photoelectrochemical (PEC) [29][30][31][32], thermocatalytic [33][34][35][36][37], and photothermal [38][39][40] reduction, etc. These mentioned methods can be driven by clean electric energy or solar energy; therefore, they are considered the most promising strategies. In the case of electrocatalytic reduction [41][42][43][44], the two key issues facing electrocatalytic CO 2 reduction are obtaining an inexpensive and a renewable source of electrical energy and the development of highly active CO 2 reduction electrocatalysts. As a result, photovoltaic cells are typically used as the electricity source (thereby allowing for more direct energy conversion from sunlight), and carefully designed electrocatalysts form the cathodes on which CO 2 reduction takes place. Though much research has been conducted to address these issues, high overpotentials, poor selectivity, low electric current densities, catalyst instability, and relatively high costs of renewable electric energy continue to limit the large-scale application of electrocatalytic CO 2 reduction [45]. Photocatalytic CO 2 reduction offers another appealing strategy for reducing CO 2 into fuels by also directly absorbing solar energy as a source of energy while employing water as an electron donor [46][47][48][49]. Due to the limitations of low separation efficiency for photogenerated carriers and slow reaction kinetics in both CO 2 reduction and H 2 O oxidation, the highest product formation rates reach only micromoles per gram of catalyst per hour (µmol g cat −1 h −1 ) [50]. To overcome those limitations, many studies have been trying to embed the donor-acceptor scheme into the primary photosensitizing module [51][52][53][54][55][56]. By comparison, thermocatalytic CO 2 reduction has proven an efficient approach for the hydrogenation of CO 2 to form methanol (CH 3 OH), methane (CH 4 ), carbon monoxide (CO), and more at high rates using heterogeneous catalysts [57][58][59][60][61][62]. However, these processes require higher reaction temperatures (e.g., 200 to 250 • C) and high pressures (5 to 10 MPa) than the previously mentioned methods. Such high temperatures can be disadvantageous, though, when synthesizing products, such as methanol, whose high reaction exothermocity results in diminishing maximum theoretical yields at higher temperatures [63,64]. A related technology, solar thermal catalysis, has also emerged as a promising approach for the synthesis of renewable fuels from CO 2 , combining the efficacy of thermocatalysis with a clean source of energy (focused sunlight).
Compared to the electrocatalytic, photocatalytic, and pure thermocatalytic approaches introduced previously, solar thermal reduction methods utilize energy from a large range of the solar spectrum (making use of materials exhibiting intense, broadband light absorption), while also offering high product selectivity relative to photochemical and electrochemical paths [65]. In light of these advantages, solar thermal catalysis has demonstrated its potential for use in industrially relevant chemical reactions, such as reducing nitrogen to ammonia (Haber process) [66], methane reforming [67], and preparing synthesis gas [68,69]. Though there have been many recent discoveries and developments regarding solar thermal catalytic reactions, these specific examples are highlighted for being significantly endothermic; as such, they benefit greatly from the high temperatures and thermal input generated by solar thermal systems.

Application of Metal/Semiconductor Catalysts for Photocatalytic CO 2 Hydrogenation
The solar energy utilization efficiency in photocatalytic processes is determined by both the incident quantum absorption capability of the catalyst and its internal quantum yield. As a result, expanding the range of wavelengths absorbed by a catalyst has been proven to be one of the most effective ways of enhancing catalytic activity, especially for the semiconductor-based materials. Though great attention has been paid to broadening optical absorption ranges, such as introducing dopants, defects, sensitizers, or plasmonic metal, these strategies often result in unintended side effects (weakened redox capacity, decreased internal quantum yield, etc.). Thus, it is necessary to further explore novel catalytic systems and new mechanisms that can maximize both the light absorption efficiency and internal quantum yield [70]. Loading noble metal nanoparticles onto photocatalysts can enhance energy harvesting from visible-light photons, thanks to high-energy ("hot") electrons generated by localized surface plasmon resonance (LSPR) [71,72]. The design of metal/semiconductor catalysts has been one potential strategy for the preparation of efficient photocatalytic CO 2 hydrogenation catalysts.

Pd-Based Catalysts
To lower the temperature needed in a thermocatalytic CO 2 reduction and increase solar energy utilization, Ni et al., fabricated a palladium-nanoparticle-loaded TiO 2 (PNT) catalyst employed using a photothermochemical cycle (PTC) method [73]. In this system, the uniformly dispersed Pd nanoparticles achieved LSPR absorption in the visible region. As a result, the PNT exhibited enhanced light utilization thanks to its red-shifted absorption range in the visible spectrum ( Figure 1a). Absorbed visible IR light could then be converted to thermal energy and incorporated into a chemical reaction, allowing the 1.0 PNT catalyst to realize a stable CO production rate of 11.05 µmol g cat −1 h −1 (more than eight times the CO production rate of P25 using the PTC) ( Figure 1c). The structure of the PNT catalyst also allowed the electron-hole pairs generated from the photoexcitation of TiO 2 to be more efficiently separated (Figure 1b), thereby increasing the number of available charge carriers. Thus, the reduction rate of CO 2 production can be accelerated by an advantage of photothermal synergy. Complementary density functional theory (DFT) calculations indicated that Pd can play a further role in promoting CO 2 adsorption by forming Pd-CO 2 − and Pd-CO 2 − -V O at defect sites during the thermal catalytic progress (Figure 1d).
Molecules 2023, 28, 5693 3 times the CO production rate of P25 using the PTC) ( Figure 1c). The structure of the catalyst also allowed the electron-hole pairs generated from the photoexcitation of to be more efficiently separated (Figure 1b), thereby increasing the number of ava charge carriers. Thus, the reduction rate of CO2 production can be accelerated by a vantage of photothermal synergy. Complementary density functional theory (DFT) c lations indicated that Pd can play a further role in promoting CO2 adsorption by for Pd-CO2 − and Pd-CO2 − -VO at defect sites during the thermal catalytic progress (Figure  Another promising series of photothermal catalysts was developed by Ozin e using Pd to provide deeper insight into the synergistic mechanisms responsible fo highly selective reduction in CO2. In the first such work, nanostructured Pd@Nb2O5 n rod catalysts were prepared via an impregnation method and tested in the gas-phas drogenation of CO2 to CO [74]. The use of ethanol as a reducing agent was instrum in forming uniformly small Pd nanoparticles on the Nb2O5 nanorods, as it allowe reduction reaction rate of the palladium precursor to be controlled, thereby suppre particle formation. Experiments proved that the Pd nanoparticles were well disperse the surface of the Nb2O5 nanorods and had diameters ranging from 2 to 10 nm. D their size, these Pd nanocrystals were able to absorb shorter wavelengths of light (v and near-infrared (NIR) light) by exciting inter-band and intra-band electronic excitat helping to initiate CO2 reduction. This increased amount of thermal energy was bene as it was also proven that the reaction rate was significantly enhanced by the gener of Nb 4+ and oxygen vacancies in situ (on reduction by H2). The Nb2O5 nanorods also tured a large surface area (as measured via BET gas adsorption), which provided a number of CO2 adsorption sites for the reaction. Wavelength-and intensity-depen Another promising series of photothermal catalysts was developed by Ozin et al., using Pd to provide deeper insight into the synergistic mechanisms responsible for the highly selective reduction in CO 2 . In the first such work, nanostructured Pd@Nb 2 O 5 nanorod catalysts were prepared via an impregnation method and tested in the gas-phase hydrogenation of CO 2 to CO [74]. The use of ethanol as a reducing agent was instrumental in forming uniformly small Pd nanoparticles on the Nb 2 O 5 nanorods, as it allowed the reduction reaction rate of the palladium precursor to be controlled, thereby suppressing particle formation. Experiments proved that the Pd nanoparticles were well dispersed on the surface of the Nb 2 O 5 nanorods and had diameters ranging from 2 to 10 nm. Due to their size, these Pd nanocrystals were able to absorb shorter wavelengths of light (visible and near-infrared (NIR) light) by exciting inter-band and intra-band electronic excitations, helping to initiate CO 2 reduction. This increased amount of thermal energy was beneficial, as it was also proven that the reaction rate was significantly enhanced by the generation of Nb 4+ and oxygen vacancies in situ (on reduction by H 2 ). The Nb 2 O 5 nanorods also featured a large surface area (as measured via BET gas adsorption), which provided a large number of CO 2 adsorption sites for the reaction. Wavelength-and intensity-dependent catalytic activity measurements performed on the nanostructured Pd@Nb 2 O 5 catalyst ( Figure 2) revealed that the reverse water-gas shift (RWGS) reaction was activated photothermally, depending primarily on the conversion on light to thermal energy rather than the formation of electron-hole pairs. By virtue of this combination of features, the RWGS reaction can be catalyzed by the Pd@Nb 2 O 5 catalyst under visible and NIR light even without external heating, achieving an impressive reaction rate of 1.8 mmol g cat −1 h −1 .
Molecules 2023, 28, 5693 4 catalytic activity measurements performed on the nanostructured Pd@Nb2O5 catalyst ( ure 2) revealed that the reverse water-gas shift (RWGS) reaction was activated photot mally, depending primarily on the conversion on light to thermal energy rather than formation of electron-hole pairs. By virtue of this combination of features, the RWGS action can be catalyzed by the Pd@Nb2O5 catalyst under visible and NIR light even w out external heating, achieving an impressive reaction rate of 1.8 mmol gcat −1 h −1 . Silicon-based support materials have also been used in conjunction with ultrasm (19-atom) Pd nanoparticles, which were deposited onto two-dimensional silicon-hyd nanosheets (Pd@SiNS) and used to catalyze the RWGS reaction. Due to the innova synthesis method, Pd precursors were reduced in situ on the surface of the Si nanoshe resulting in highly dispersed and well-immobilized Pd nanoparticles while also retain the large surface area (197 m 2 g −1 ) of the two-dimensional Si nanosheets. Several cha terization techniques were also employed, including Fourier transform infrared spect copy, in situ isotopic labeling, and theoretical calculations to understanding the inte catalytic mechanism. [40] The Pd@SiNS sample showed a 13 CO production rate of 10 µ gcat -1 h -1 , which was much better than that of the pristine SiNS sample (Figure 3a), retained its good stability under repeated cycling (relative to pristine SiNSs). A cru supporting experiment showed that Pd nanoparticles on fully oxidized Si nanosh (Pd@oxSiNS) had no activity in CO2 reduction (Figure 3b), demonstrating the limited tivity of Pd nanoparticles alone and proving the synergetic effect between Pd and th nanosheets. By studying these experimental results, a clear and fundamental understa ing was reached, in that the H-transfer process from Pd to the oxidized SiNS sur Silicon-based support materials have also been used in conjunction with ultrasmall (19-atom) Pd nanoparticles, which were deposited onto two-dimensional silicon-hydride nanosheets (Pd@SiNS) and used to catalyze the RWGS reaction. Due to the innovative synthesis method, Pd precursors were reduced in situ on the surface of the Si nanosheets, resulting in highly dispersed and well-immobilized Pd nanoparticles while also retaining the large surface area (197 m 2 g −1 ) of the two-dimensional Si nanosheets. Several characterization techniques were also employed, including Fourier transform infrared spectroscopy, in situ isotopic labeling, and theoretical calculations to understanding the internal catalytic mechanism. [40] The Pd@SiNS sample showed a 13 CO production rate of 10 µmol g cat −1 h −1 , which was much better than that of the pristine SiNS sample (Figure 3a), and retained its good stability under repeated cycling (relative to pristine SiNSs). A crucial supporting experiment showed that Pd nanoparticles on fully oxidized Si nanosheets (Pd@oxSiNS) had no activity in CO 2 reduction (Figure 3b), demonstrating the limited activity of Pd nanoparticles alone and proving the synergetic effect between Pd and the Si nanosheets. By studying these experimental results, a clear and fundamental understanding was reached, in that the H-transfer process from Pd to the oxidized SiNS surface represented two mechanisms. First, one H atom adsorbed onto a Pd nanoparticle interacted with a surface Si-O-Si site of the SiNSs. Resulting in the formation of a surface Si-OH site. Second, another active H from the Pd nanoparticle forms a bond with the generated Si-OH, leading to the desorption of a molecule of water and the formation of a highly reactive surface Si radical site capable of reacting with CO 2 and enabling a catalytic reaction cycle. The strain induced in the SiNS surface by the Si-O-Si bonds is believed to be responsible for the enhanced reactivity of the oxidized SiNS surface toward catalytic hydrogenation of CO 2 under mild conditions. This work also highlights the potential role of Si surface chemistry in designing and developing new heterogeneous CO 2 catalysts in the future.
represented two mechanisms. First, one H atom adsorbed onto a Pd nanoparticle in acted with a surface Si-O-Si site of the SiNSs. Resulting in the formation of a surface OH site. Second, another active H from the Pd nanoparticle forms a bond with the ge ated Si-OH, leading to the desorption of a molecule of water and the formation of a hi reactive surface Si radical site capable of reacting with CO2 and enabling a catalytic r tion cycle. The strain induced in the SiNS surface by the Si-O-Si bonds is believed t responsible for the enhanced reactivity of the oxidized SiNS surface toward catalytic drogenation of CO2 under mild conditions. This work also highlights the potential ro Si surface chemistry in designing and developing new heterogeneous CO2 catalysts in future. Similar synergistic mechanisms between metal co-catalysts and metal oxide supp were also presented by Li et al., in which a catalyst consisting of palladium nanopart deposited into tungsten-bronze nanowires (Pd@HyWO3−x) was used to efficiently cata the reduction in CO2 to CO at a rate of 3.0 mmol gcat −1 hr −1 with a selectivity exceed 99%. [75] WO3 nanowires loaded with Pd nanocrystals provided the synthetic pathwa obtaining Pd@HyWO3−x. Through various characterization techniques, it was dem strated that HyWO3−x was formed via spillover of activated H from the Pd nanoparti The existence of Brønsted acid hydroxyls (-OH), W(V) sites, and oxygen vacancies (V HyWO3−x all contributed to the CO2 capture and reduction reactions.

Ru-Based Catalysts
Ye et al., prepared a series of Group VIII metals (Ru, Rh, Ni, Co, Pd, Pt, Ir, and loaded onto Al2O3 or TiO2 to form classical metal/metal oxide nanocatalysts and stu their photothermal effects on the catalytic reduction in CO2 into CH4 and CO (using [39]. The loading of Group VIII metals onto Al2O3 or TiO2 resulted in highly active m rials that had a dark grey or black color (Figure 4a) due largely to their nanoscale struct A high product selectivity (over 95%) for either CH4 or CO could be achieved by adjus the metal and support materials used in a catalyst (Figure 4b). The outstanding CO2 version ability of the prepared catalyst was attributed to two main features: highly cient utilization via photothermal processes and the unique ability of the Group VIII als to activate CO2 hydrogenation. Compared to pure photocatalytic pathways, photot mal CO2 conversion is not restricted to utilizing ultraviolet irradiation, as it can rea absorb energy from visible and infrared wavelengths (this range contains 96% of the s energy) can utilized by photothermal CO2 conversion. In related research from the s group, Ru-loaded ultrathin, layered double hydroxides (Ru@FL-LDHs) were also pared, achieving efficient photothermal CO2 methanation in a flow reactor system These catalysts were composed of well-dispersed Ru nanoparticles embedded in trathin, exfoliated LDH sheets. Under light irradiation, a maximum CH4 production Similar synergistic mechanisms between metal co-catalysts and metal oxide supports were also presented by Li et al., in which a catalyst consisting of palladium nanoparticles deposited into tungsten-bronze nanowires (Pd@H y WO 3−x ) was used to efficiently catalyze the reduction in CO 2 to CO at a rate of 3.0 mmol g cat −1 hr −1 with a selectivity exceeding 99% [75]. WO 3 nanowires loaded with Pd nanocrystals provided the synthetic pathway to obtaining Pd@HyWO 3−x . Through various characterization techniques, it was demonstrated that H y WO 3−x was formed via spillover of activated H from the Pd nanoparticles. The existence of Brønsted acid hydroxyls (-OH), W(V) sites, and oxygen vacancies (V O ) in HyWO 3−x all contributed to the CO 2 capture and reduction reactions.

Ru-Based Catalysts
Ye et al., prepared a series of Group VIII metals (Ru, Rh, Ni, Co, Pd, Pt, Ir, and Fe) loaded onto Al 2 O 3 or TiO 2 to form classical metal/metal oxide nanocatalysts and studied their photothermal effects on the catalytic reduction in CO 2 into CH 4 and CO (using H 2 ) [39]. The loading of Group VIII metals onto Al 2 O 3 or TiO 2 resulted in highly active materials that had a dark grey or black color (Figure 4a) due largely to their nanoscale structure. A high product selectivity (over 95%) for either CH 4 or CO could be achieved by adjusting the metal and support materials used in a catalyst (Figure 4b). The outstanding CO 2 conversion ability of the prepared catalyst was attributed to two main features: highly efficient utilization via photothermal processes and the unique ability of the Group VIII metals to activate CO 2 hydrogenation. Compared to pure photocatalytic pathways, photothermal CO 2 conversion is not restricted to utilizing ultraviolet irradiation, as it can readily absorb energy from visible and infrared wavelengths (this range contains 96% of the solar energy) can utilized by photothermal CO 2 conversion. In related research from the same group, Ruloaded ultrathin, layered double hydroxides (Ru@FL-LDHs) were also prepared, achieving efficient photothermal CO 2 methanation in a flow reactor system [76]. These catalysts were composed of well-dispersed Ru nanoparticles embedded in ultrathin, exfoliated LDH sheets. Under light irradiation, a maximum CH 4 production rate of 277 mmol g cat −1 h −1 or 230.8 mol h −1 m 2 (irradiation area) was attained, which was significantly higher than most reported LDH-based catalysts [77][78][79]. These experiments proved that ultrathin LDH sheets can provide an abundance of active sites for the adsorption and activation of CO 2 molecules, allowing for the use of a high flow rate of feedstocks without decreasing the percent conversion of CO 2 to products. Meanwhile, the highly dispersed Ru nanoparticles contributed by providing higher local surface temperatures under light irradiation (via photothermal heating) and activating H 2 molecules, thereby promoting the hydrogenation of CO 2 . In this example, the noble metal particles and supporting LDH matrix activated the CO 2 and H 2 reagents separately, highlighting the potential benefits of designing catalysts by targeting the activation of individual reactant molecules.
of 277 mmol gcat −1 h −1 or 230.8 mol h −1 m 2 (irradiation area) was attained, which was significantly higher than most reported LDH-based catalysts [77][78][79]. These experiments proved that ultrathin LDH sheets can provide an abundance of active sites for the adsorption and activation of CO2 molecules, allowing for the use of a high flow rate of feedstocks without decreasing the percent conversion of CO2 to products. Meanwhile, the highly dispersed Ru nanoparticles contributed by providing higher local surface temperatures under light irradiation (via photothermal heating) and activating H2 molecules, thereby promoting the hydrogenation of CO2. In this example, the noble metal particles and supporting LDH matrix activated the CO2 and H2 reagents separately, highlighting the potential benefits of designing catalysts by targeting the activation of individual reactant molecules. In a related work by Dai et al., a Ru/TiO(2-x)Nx catalyst was prepared by loading Ru onto nitrogen-doped TiO2 support via an impregnation-reduction method [80]. The catalytic CO2 methanation performance of this catalyst was evaluated under both visible light irradiation (435 nm < λ < 465 nm) and under dark conditions in a fixed-bed flow reactor. With the illumination of visible light, the turnover frequency (TOF) of the catalyst increased to 15.0 h −1 (compared to the 7.0 h −1 under dark condition), with the CH4 selectivity remaining stable at about 80% throughout the duration of the reaction. A combination of chemisorption, surface analysis, and photocurrent measurements revealed that visible light contributed to the catalytic reaction in two ways. First, a greater number of oxygen vacancies were formed due to the stronger visible light response of TiO2-xNx, allowing CO2 to be more efficiently adsorbed, activated, and converted into the CO. Second, Ru nanoparticles can accept photogenerated electrons from TiO(2-x)Nx, increasing the surface electron density of Ru nanoparticles and promoting the adsorption and activation of CO2.
In another study from the same group, Ru/silicon nanowire catalysts (Ru/SiNW) (Figure 5a,b) were prepared by forming 10 nm Ru on silicon nanowires (SiNWs) via sputtering. In this case, the coexistence of Ru and SiNWs resulted in the possibility of activating the Sabatier reaction (CO2 + H2  CH4 + H2O) both thermochemically and photochemically [81]. Detailed investigations were performed in the presence and absence of light irradiation (Figure 5c), with varying atmospheric pressure, (Figure 5d,e) and using varying wavelengths of light (Figure 5f) to better understand the internal mechanisms responsible for the observed catalytic activity. The results of these investigations revealed that the observed CO2 reduction mechanism could be predominantly photochemical or thermochemical, depending on the wavelengths of irradiating light. When the photon energy was less than the bandgap energy of silicon, the Ru/SiNW catalyst relied upon photothermal heating to activate the Sabatier reaction thermochemically. When the incident photon energy is sufficient to overcome the bandgap of Si, however, electron-hole pairs generated in the Ru/SiNW catalyst assisted the formation of R-H bonds. Eventually, a proportion of these excited charge carriers eventually thermalize and recombine, generating heat and In a related work by Dai et al., a Ru/TiO (2−x) N x catalyst was prepared by loading Ru onto nitrogen-doped TiO 2 support via an impregnation-reduction method [80]. The catalytic CO 2 methanation performance of this catalyst was evaluated under both visible light irradiation (435 nm < λ < 465 nm) and under dark conditions in a fixed-bed flow reactor. With the illumination of visible light, the turnover frequency (TOF) of the catalyst increased to 15.0 h −1 (compared to the 7.0 h −1 under dark condition), with the CH 4 selectivity remaining stable at about 80% throughout the duration of the reaction. A combination of chemisorption, surface analysis, and photocurrent measurements revealed that visible light contributed to the catalytic reaction in two ways. First, a greater number of oxygen vacancies were formed due to the stronger visible light response of TiO 2−x N x , allowing CO 2 to be more efficiently adsorbed, activated, and converted into the CO. Second, Ru nanoparticles can accept photogenerated electrons from TiO (2−x) N x , increasing the surface electron density of Ru nanoparticles and promoting the adsorption and activation of CO 2 .
In another study from the same group, Ru/silicon nanowire catalysts (Ru/SiNW) (Figure 5a,b) were prepared by forming 10 nm Ru on silicon nanowires (SiNWs) via sputtering. In this case, the coexistence of Ru and SiNWs resulted in the possibility of activating the Sabatier reaction (CO 2 + H 2 → CH 4 + H 2 O) both thermochemically and photochemically [81]. Detailed investigations were performed in the presence and absence of light irradiation (Figure 5c), with varying atmospheric pressure, (Figure 5d,e) and using varying wavelengths of light (Figure 5f) to better understand the internal mechanisms responsible for the observed catalytic activity. The results of these investigations revealed that the observed CO 2 reduction mechanism could be predominantly photochemical or thermochemical, depending on the wavelengths of irradiating light. When the photon energy was less than the bandgap energy of silicon, the Ru/SiNW catalyst relied upon photothermal heating to activate the Sabatier reaction thermochemically. When the incident photon energy is sufficient to overcome the bandgap of Si, however, electron-hole pairs generated in the Ru/SiNW catalyst assisted the formation of R-H bonds. Eventually, a proportion of these excited charge carriers eventually thermalize and recombine, generating heat and supporting the thermochemical promotion of the Sabatier reaction. In this way, the prepared Ru/SiNW catalysts achieved a CH 4 production rate of 0.74 mmol g Ru −1 h −1 with a simulated solar light intensity of 14.5 suns, a 4:1 H 2 /CO 2 ratio, and an atmospheric pressure of 15 psia. The concepts and the experimental approaches introduced in this work are significantly helpful for understanding the synergistic contributions of photochemical and thermochemical mechanisms to the overall process of photothermal CO 2 reduction.
supporting the thermochemical promotion of the Sabatier reaction. In this way, the prepared Ru/SiNW catalysts achieved a CH4 production rate of 0.74 mmol gRu −1 h −1 with a simulated solar light intensity of 14.5 suns, a 4:1 H2/CO2 ratio, and an atmospheric pressure of 15 psia. The concepts and the experimental approaches introduced in this work are significantly helpful for understanding the synergistic contributions of photochemical and thermochemical mechanisms to the overall process of photothermal CO2 reduction. In yet another study, the photothermal reduction in gaseous CO2 over Ru/silicon photonic crystal photocatalyst (Figure 6a,b) was studied at ambient temperature [82]. Ru/i-Sio catalysts showed increasing activity as light intensity increased, achieving peak rates of up to 4.4 mmol gcat −1 h −1 (Figure 6c). The high photomethanation rates over the Ru/i-Si-o catalyst are attributed to the excellent light-harvesting properties of the silicon photonic crystal, as proven using control experiments (Figure 6d). Complementary simulation results further indicated that the charged Ru surfaces are beneficial for CO2 destabilizationadsorption progress; additionally, the charged Ru surfaces are crucial for the adsorbing and dissociating of H2, which can react with CO2 to accelerate the Sabatier reaction.  [81].
In yet another study, the photothermal reduction in gaseous CO 2 over Ru/silicon photonic crystal photocatalyst (Figure 6a,b) was studied at ambient temperature [82]. Ru/i-Si-o catalysts showed increasing activity as light intensity increased, achieving peak rates of up to 4.4 mmol g cat −1 h −1 (Figure 6c). The high photomethanation rates over the Ru/i-Si-o catalyst are attributed to the excellent light-harvesting properties of the silicon photonic crystal, as proven using control experiments (Figure 6d). Complementary simulation results further indicated that the charged Ru surfaces are beneficial for CO 2 destabilizationadsorption progress; additionally, the charged Ru surfaces are crucial for the adsorbing and dissociating of H 2 , which can react with CO 2 to accelerate the Sabatier reaction.

Other Metal-Based Catalysts
Recently, the plasmonic properties of Rh and Au in photocatalytic CO 2 reduction were investigated by Liu et al. [83]. A Rh/Al 2 O 3 catalyst was prepared via an impregnation method. Rh nanocube colloids were dropped onto Al 2 O 3 nanoparticles, and the formed solid was ground and calcined in air ( Figure 7a) and then tested in a fixed-bed reactor (Figure 7b). It was found that the CH 4 and CO production rates for Rh/Al 2 O 3 were similar at 350 • C without light irradiation. When ultraviolet light (3 W cm −2 ) was introduced, a seven-fold increase in the CH 4 production rate was detected, while the CO production rate changed only very slightly (Figure 7c). In contrast, plasmonic Au nanoparticles could catalyze only CO production, regardless of the presence of irradiation (Figure 7d). It is also worth noting that the CH 4 selectivity of unheated Rh nanoparticles was over~86% or 98% compared to the dark condition or thermo condition, respectively. And, the reaction rate was double that of the thermocatalytic reaction rate at 350 • C (Figure 7d,e). These experimental results together indicated the selective production of CH 4 via a photothermal catalytic process. Further kinetic studies, together with DFT calculations, revealed that photo-excited hot electrons generated through the LSPR effect of Rh nanocubes would selectively be transferred into the antibonding orbital of the CHO intermediate, which can lower the activation energy of CH 4 production by~35% relative to the thermal activation energies (Figure 7f). This study well demonstrates the idea of tuning product selectivity in CO 2 reduction by using light to excite plasmonic electrons and activate key reaction intermediates.

Other Metal-Based Catalysts
Recently, the plasmonic properties of Rh and Au in photocatalytic CO2 reduction were investigated by Liu et al. [83]. A Rh/Al2O3 catalyst was prepared via an impregnation method. Rh nanocube colloids were dropped onto Al2O3 nanoparticles, and the formed solid was ground and calcined in air ( Figure 7a) and then tested in a fixed-bed reactor (Figure 7b). It was found that the CH4 and CO production rates for Rh/Al2O3 were similar at 350 °C without light irradiation. When ultraviolet light (3 W cm −2 ) was introduced, a seven-fold increase in the CH4 production rate was detected, while the CO production rate changed only very slightly (Figure 7c). In contrast, plasmonic Au nanoparticles could catalyze only CO production, regardless of the presence of irradiation (Figure 7d). It is also worth noting that the CH4 selectivity of unheated Rh nanoparticles was over ~86% or ~98% compared to the dark condition or thermo condition, respectively. And, the reaction rate was double that of the thermocatalytic reaction rate at 350 °C (Figure 7d,e). These experimental results together indicated the selective production of CH4 via a photothermal catalytic process. Further kinetic studies, together with DFT calculations, revealed that photo-excited hot electrons generated through the LSPR effect of Rh nanocubes would selectively be transferred into the antibonding orbital of the CHO intermediate, which can lower the activation energy of CH4 production by ~35% relative to the thermal activation energies (Figure 7f). This study well demonstrates the idea of tuning product selectivity in CO2 reduction by using light to excite plasmonic electrons and activate key reaction intermediates.  In order to achieve an efficient dry reforming reaction (CO2 + CH4  2CO + 2H2), et al., used plasmonic nanoparticles to promote the activity of their noble metal-bas catalyst (Figure 8) [9]. In this case, a Rh-Au/SBA-15 bimetallic catalyst was synthesized a subsequent impregnation pathway, with the prepared catalyst showing a strong LS corresponding to Au nanoparticles (Figure 8a) [11]. In a flow reactor system with visi light illumination of 420 to 800 nm and a temperature of 500 °C, the activity catalyst w enhanced 1.7 times compared to that observed in the dark. Additionally, the catalytic In order to achieve an efficient dry reforming reaction (CO 2 + CH 4 → 2CO + 2H 2 ), Ye et al., used plasmonic nanoparticles to promote the activity of their noble metal-based catalyst (Figure 8) [9]. In this case, a Rh-Au/SBA-15 bimetallic catalyst was synthesized via a subsequent impregnation pathway, with the prepared catalyst showing a strong LSPR corresponding to Au nanoparticles (Figure 8a) [11]. In a flow reactor system with visible light illumination of 420 to 800 nm and a temperature of 500 • C, the activity catalyst was enhanced 1.7 times compared to that observed in the dark. Additionally, the catalytic activity of Rh-Au/SBA-15 was much higher than that of either Rh/SBA-15 or Au/SBA-15 (Figure 8b). UV-Vis spectra and electromagnetic field simulation results confirmed the corresponding relationship between the enhanced dry reforming activity and the LSPR band of Au (Figure 8c,d). This demonstrated the photothermal activity of the dry reforming reaction over the Rh-Au/SBA-15 catalyst and the role of energetic "hot" electrons excited by LSPR in Au, contributing both thermal energy and polarization/activation of CO 2 and CH 4 to the catalytic reaction. In another work from the same group, TaN was used as a support for loading Pt nanoparticles, chosen due to its intense light absorption over a range of wavelengths (Figure 9a) [84]. This Pt/TaN catalyst was prepared via a simple impregnation method, resulting in an average Pt nanoparticle diameter of approximately 2.7 nm (Figure 9b). The Pt/TaN catalyst was also tested in the dry reforming reaction under visible light irradiation at 500 °C, which revealed that the CO2 conversion rate of the catalyst was 2.7 times that observed in the dark. In a control experiment employing Pt/Ta2O5 as a catalyst, the catalytic activity remained nearly constant even under visible light illumination (Figure 9c). Using DFT calculation results, the polarity of TaN was revealed to be a critical parameter for the efficient separation of electron-hole pairs generated in the photothermal excitation. In conclusion, this study emphasized the importance of a support material's optical absorption properties in photothermal reactions (Figure 9d). In another work from the same group, TaN was used as a support for loading Pt nanoparticles, chosen due to its intense light absorption over a range of wavelengths (Figure 9a) [84]. This Pt/TaN catalyst was prepared via a simple impregnation method, resulting in an average Pt nanoparticle diameter of approximately 2.7 nm (Figure 9b). The Pt/TaN catalyst was also tested in the dry reforming reaction under visible light irradiation at 500 • C, which revealed that the CO 2 conversion rate of the catalyst was 2.7 times that observed in the dark. In a control experiment employing Pt/Ta 2 O 5 as a catalyst, the catalytic activity remained nearly constant even under visible light illumination (Figure 9c). Using DFT calculation results, the polarity of TaN was revealed to be a critical parameter for the efficient separation of electron-hole pairs generated in the photothermal excitation. In conclusion, this study emphasized the importance of a support material's optical absorption properties in photothermal reactions (Figure 9d). The aforementioned plasmonic properties of Pd, Rh, and Au demonstrate the promotion of photocatalytic activity by extending the light absorption range, suppressing the charge-carrier recombination, and forming highly active surface reaction sites. However, the relatively high cost and low abundance of noble metals place a hard limit on their practical application on a global scale. With this in mind, Ozin et al., have recently developed plasmonic cobalt superstructures capable of achieving nearly 100% sunlight absorption while acting as photocatalysts for efficient CO2 reduction. [70] This cobalt plasmonic superstructure was composed of a nanoporous, needle-like structure containing Co metal cores and an enclosing silica shell, Co-PS@SiO2 (Figure 10a). For comparison, Co@SiO2 and Co/FTO samples were also prepared via traditional impregnation methods. The diffuse reflectance spectra of different catalysts were characterized as shown in Figure 9b. It was demonstrated that the silica shell can prevent the nanoneedle structure from collapsing and confine the growth of Co nanocrystal. As a result, more active sites will be preserved, and the photon path length can be simultaneously increased through multiple reflection. Through detailed experimental comparisons, this study concluded that the strong light absorption of Co metal combined with the antireflection of the nanoarray structure contributes to the broadband optical absorption across the entire solar spectrum. (Figure 10b). The photocatalytic activity was evaluated in a batch reactor system operating at atmospheric pressure with a 1:1 feed ratio of H2/CO2. The catalytic testing results of the three aforementioned catalysts are presented in Figure 10c,d. The CO2 conversion rate of Co-PS@SiO2 was about 0.6 mol gCo −1 h −1 (normalized to the mass of Co), which is 6 times that of Co@SiO2 and 277 times that of Co/FTO (Figure 9a). When normalized by surface area, these respective activity increases were 20 and 151 times greater (Figure 10b). Moreover, the CO selectivity of Co-PS@SiO2 was higher than that of Co@SiO2, which is also consistent The aforementioned plasmonic properties of Pd, Rh, and Au demonstrate the promotion of photocatalytic activity by extending the light absorption range, suppressing the charge-carrier recombination, and forming highly active surface reaction sites. However, the relatively high cost and low abundance of noble metals place a hard limit on their practical application on a global scale. With this in mind, Ozin et al., have recently developed plasmonic cobalt superstructures capable of achieving nearly 100% sunlight absorption while acting as photocatalysts for efficient CO 2 reduction. [70] This cobalt plasmonic superstructure was composed of a nanoporous, needle-like structure containing Co metal cores and an enclosing silica shell, Co-PS@SiO 2 (Figure 10a). For comparison, Co@SiO 2 and Co/FTO samples were also prepared via traditional impregnation methods. The diffuse reflectance spectra of different catalysts were characterized as shown in Figure 9b. It was demonstrated that the silica shell can prevent the nanoneedle structure from collapsing and confine the growth of Co nanocrystal. As a result, more active sites will be preserved, and the photon path length can be simultaneously increased through multiple reflection. Through detailed experimental comparisons, this study concluded that the strong light absorption of Co metal combined with the antireflection of the nanoarray structure contributes to the broadband optical absorption across the entire solar spectrum. (Figure 10b). The photocatalytic activity was evaluated in a batch reactor system operating at atmospheric pressure with a 1:1 feed ratio of H 2 /CO 2 . The catalytic testing results of the three aforementioned catalysts are presented in Figure 10c,d. The CO 2 conversion rate of Co-PS@SiO 2 was about 0.6 mol g Co −1 h −1 (normalized to the mass of Co), which is 6 times that of Co@SiO 2 and 277 times that of Co/FTO (Figure 9a). When normalized by surface area, these respective activity increases were 20 and 151 times greater (Figure 10b). Moreover, the CO selectivity of Co-PS@SiO 2 was higher than that of Co@SiO 2 , which is also consistent with the higher local catalyst temperature caused by the better light absorption ability. The ability of Co-PS@SiO 2 to fully harvest solar energy demonstrates the applicability of base metal plasmonic superstructures in other solar energy harvesting systems.

Summary and Outlook
In this review, we have focused on recent developments in solar-energy-driven CO2 hydrogenation reactions, from the design of catalysts to the structure of active sites and mechanistic investigations. To date, several strategies have been developed to improve the catalytic efficiency and overcome the challenges facing CO2 reduction. For the purpose of utilizing light energy, localized surface plasmon resonances of metal particles (Pd, Rh, Ni, Co, etc.) and vacancies have been proven as two simple and efficient approaches to widening the light absorption range of light-driven catalyst materials. The presence of metal particles not only enhances the optical absorption capability but also provides active sites for the activation of H2 and CO2. As a surface catalytic reaction, the presence of anion vacancies can both improve the adsorption of CO2 molecules and lower the activation energy associated with their reduction. Specialized nanostructures have also been constructed to increase the utilization of light, including photonic crystals that can enhance light trapping. Indium-based oxides are found to have excellent performance in CO2 hydrogenation because of their flexibility in terms of morphologies, phases, and surface-active sites. Though good progress has been made, several challenges must still be overcome before practical applications of these catalysts becomes possible.
First, the catalytic activity of many reported catalysts is still limited; therefore, improving nanoscale architectural design, bandgap engineering, and broad spectral adsorption appears to be particularly important. Many experimental and theoretical studies have been undergoing to study the relationship between spectral adsorption properties and

Summary and Outlook
In this review, we have focused on recent developments in solar-energy-driven CO 2 hydrogenation reactions, from the design of catalysts to the structure of active sites and mechanistic investigations. To date, several strategies have been developed to improve the catalytic efficiency and overcome the challenges facing CO 2 reduction. For the purpose of utilizing light energy, localized surface plasmon resonances of metal particles (Pd, Rh, Ni, Co, etc.) and vacancies have been proven as two simple and efficient approaches to widening the light absorption range of light-driven catalyst materials. The presence of metal particles not only enhances the optical absorption capability but also provides active sites for the activation of H 2 and CO 2 . As a surface catalytic reaction, the presence of anion vacancies can both improve the adsorption of CO 2 molecules and lower the activation energy associated with their reduction. Specialized nanostructures have also been constructed to increase the utilization of light, including photonic crystals that can enhance light trapping. Indium-based oxides are found to have excellent performance in CO 2 hydrogenation because of their flexibility in terms of morphologies, phases, and surface-active sites. Though good progress has been made, several challenges must still be overcome before practical applications of these catalysts becomes possible.
First, the catalytic activity of many reported catalysts is still limited; therefore, improving nanoscale architectural design, bandgap engineering, and broad spectral adsorption appears to be particularly important. Many experimental and theoretical studies have been undergoing to study the relationship between spectral adsorption properties and catalytic performance; however, the lack of enough in-situation or operando activities make it a challenge to discover the internal mechanism.
Second, the cost of many of these catalysts is prohibitively high. Most of the developed catalysts contain precious metals; however, increasing atomic efficiency through alloying or the development of single atom catalysts may be offering some improvement. Moreover, it is also important to design some non-noble metals and noble metals (bimetal)/semiconductor catalysts in the process of photocatalytic CO 2 reduction, such as Cu, Bi, et al.
Finally, the industrial-scale preparation of these catalysts and the developments of specialized equipment are still relatively scarcely reported. In a word, with the continuous development of broadband, highly absorbing materials, and increased understanding of photochemical and thermal properties in nanostructured catalysts, the future of photothermal catalysis offers many interesting opportunities for CO 2 capture and conversion technologies. The combining of photocatalytic and thermocatalytic approaches will doubtlessly provide new development opportunities for efficient CO 2 hydrogenation.