Advances in Photocatalytic CO₂ Reduction with Water: A Review.

In recent years, the increasing level of CO₂ in the atmosphere has not only contributed to global warming but has also triggered considerable interest in photocatalytic reduction of CO₂. The reduction of CO₂ with H₂O using sunlight is an innovative way to solve the current growing environmental challenges. This paper reviews the basic principles of photocatalysis and photocatalytic CO₂ reduction, discusses the measures of the photocatalytic efficiency and summarizes current advances in the exploration of this technology using different types of semiconductor photocatalysts, such as TiO₂ and modified TiO₂, layered-perovskite Ag/ALa₄Ti₄O15 (A = Ca, Ba, Sr), ferroelectric LiNbO₃, and plasmonic photocatalysts. Visible light harvesting, novel plasmonic photocatalysts offer potential solutions for some of the main drawbacks in this reduction process. Effective plasmonic photocatalysts that have shown reduction activities towards CO₂ with H₂O are highlighted here. Although this technology is still at an embryonic stage, further studies with standard theoretical and comprehensive format are suggested to develop photocatalysts with high production rates and selectivity. Based on the collected results, the immense prospects and opportunities that exist in this technique are also reviewed here.

1. Introduction

Global warming is viewed to be one of the vital environmental concerns that humankind is dealing with [1]. CO2 contributes mostly to the worldwide climate change because it is more than 64% effective than other greenhouse gasses in the atmosphere [2]. This chemically stable gas contributes to the increase in global temperature through absorption and re-emission of infrared radiation. In the past century, the temperature of the Earth’s surface increased by roughly 0.6 K; the warming trend reveals more significant changes in last 20 years, according to the Intergovernmental Panel on Climate Change (IPCC) [3]. The consequences of the greenhouse effect are global and severe, such as ice melting at the Earth’s poles, the quick rising of sea level, and growing precipitation across the globe [4]. To deal with these issues, numerous studies have been conducted over the last few decades applying various strategies to control CO2 emission or convert it into other products.

There are at least three routes of lowering the amount of CO2 in the atmosphere: (i) direct reduction of CO2 emission; (ii) CO2 capture and storage (CCS); and (iii) CO2 utilization [5,6,7]. Lowering the CO2 emission may seem quite unrealistic because of the present human lifestyle and emergent use of fossil fuel. The potential of CCS technology can be restrained because of the environmental risk of leakage and the energy requirement for fuel compression and transportation. Among the renewable resources, solar energy is the most exploitable one by making available more energy to the Earth for every hour than the total amount of energy humans consume in a year [8].

Harvesting this abundant sunlight in solving environmental problems is a promising approach and one of the ultimate goals for sustainability of global development. In the long term, photocatalytic conversion of CO2 utilizing solar energy is the most appealing route for CO2 reduction [9,10]. In addition, CO2 reduction advances recycling of carbon source [8]. The interest in this field of research has begun with the work of Fujishima and Honda in 1972 [11]. The advancements in nanotechnology, particularly the synthesis of nanomaterials with different structures and morphologies [12,13], and the most recent approach of using noble metals, such as Au or Ag, with surface plasmon resonance (SPR) to enhance the photocatalytic efficiency of TiO2 or other semiconductors [14,15,16] have facilitated the progress.

For real life application, a photocatalytic system must be capable of working under daylight even when the sun is not directly overhead and show both long-time consistency and efficiency. The reduction process has to be promoted while suppressing any side reaction that can occur during the reaction, and H2O should be used as an electron source [17]. Unfortunately, a photocatalyst that satisfies all these requirements has not been reported yet. A considerable number of review papers on this emerging topic have already been published. Some papers focus on the advances in developing novel photocatalysts with high photocatalytic activity [18,19,20,21,22], while others on studying the enhancement mechanisms and the influences of co-catalysts [23], the applications by highlighting on the reaction conditions, reactor design and analysis methods [18,24] and comprehensive discussion on general considerations that apply specifically to CO2 reduction [25,26]. Furthermore, extensive studies on TiO2-based photocatalysts [27,28,29] and noble metal nanoparticles dispersed plasmonic photocatalysts have been published as well [30,31,32]. Nevertheless, the basic insight of photocatalytic CO2 reduction in presence of H2O and comparison among the photocatalytic efficiency of different photocatalysts in this reaction has not been clearly documented to date. This review paper covers the basic aspects of photocatalytic CO2 reduction process with H2O, concentrating on recently reported semiconductor photocatalysts with high photoactivity, particularly on plasmonic photocatalysts.

2. Photocatalysis and Photocatalytic Reduction of CO2 with H2O

The word photocatalysis consists of two parts: photo and catalysis, “photo” means light and “catalysis” is the performance of a substance during the chemical transformation of the reactants to modify the reaction rate without being changed ultimately [33]. In practice, the word photocatalysis refers to the acceleration of a photoreaction in the presence of a catalyst [34]. In photocatalytic CO2 reduction system with water, both photo-reduction of CO2 and photo-oxidation of H2O occur simultaneously under sunlight irradiation using a suitable photocatalyst. A variety of reaction conditions intensely affects the product distribution of this reaction, such as reactor geometry, catalyst type, sacrificial reagents, and even illumination type. Thus, predicting the product distribution of a particular photocatalytic reaction is very challenging [35].

The photocatalytic CO2 reduction is a very effective method considering that no additional energy is needed and no negative effect on the environment is produced. The use of cheap and abundant sunlight to transform this major greenhouse gas into other carbon containing products is also an ideal approach because of its low cost. Here, the high activation energy to break very stable CO2 molecule is provided by solar energy [35]. To date, many photocatalysts, including oxides and non-oxides, e.g., TiO2, ZnO, Fe2O3, ZrO2, SnO2, BiWO3, Ti-MCM-41, CdS, TNTs, ZnS, GaN, and SiC, have been studied for the photocatalytic reduction of CO2 with H2O. A summary of different photocatalytic systems employed in this technology since 2010 are given in Table 1.

Table 1

Advances in photocatalytic systems for CO2 reduction with water since the year 2010.

Photocatalyst	Radiation Source	Major Products	Comments	References
0.5 wt % Cu/TiO2-SiO2	Xe lamp (2.4 mW cm−2, 250–400 nm)	CO and CH4	The synergistic combination of Cu deposition and high surface area of SiO2 support enhanced CO2 photoreduction rates.	[36]
ZnGa2O4	300 W Xe arc lamp	CH4	Strong gas adsorption and large specific surface area of the mesoporous ZnGa2O4 photocatalyst contribute to its high photocatalytic activity for converting CO2 into CH4.	[37]
(RuO + Pt)-Zn2GeO4	300 W Xe arc lamp	CH4	In the presence of water, ultra-long and ultrathin geometry of the Zn2GeO4 nano-ribbon promotes CO2 photo-reduction, which was significantly enhanced by loading of Pt or RuO2.	[38]
Ag/ALa4Ti4O15 (A = Ca, Ba and Sr)	400 W Hg lamp	CO, HCOOH, and H2	On the optimized Ag/BaLa4Ti4O15 photocatalyst, CO was the reported as the main product. The molar ratio of O2 production (H2 + CO:O2 = 2:1) demonstrated that water was consumed as a reducing reagent in the photocatalytic process.	[39]
I-TiO2 nanoparticles	450 W Xe lamp	CO	High photocatalytic activity was observed under visible light and the efficiency of CO2 photoreaction was much greater than undoped TiO2 due to the extension in the absorption spectra of TiO2 to the visible light region and facilitated charge separation.	[40]
LiNbO3	Natural sunlight or Hg lamp (64.2 mW cm−2)	HCOOH	The MgO-doped LiNbO3 showed an energy conversion efficiency rate of 0.72% which was lower than that for the gas–solid catalytic reaction of LiNbO3 (2.2%).	[41]
G-Ti0.91O2 hollow spheres	300 W Xe arc lamp	CH4, CO	The presence of G nanosheets compactly stacking with Ti0.91O2 nanosheets allows the rapid migration of photo-generated electrons from Ti0.91O2 nanosheets into G and improves the efficiency of the photocatalytic process.	[42]
Graphene oxides (GOs)	300 W commercial halogen lamp	CH3OH	Among all GOs, GO-3 exhibited the highest efficiency as a photocatalyst for CO2 reduction under visible light, and the conversion rate of CO2 to CH3OH on modified GO (GO-3) was 0.172 mmol g−1 cat h−1, which is six-fold higher than that of pure TiO2.	[43]
W18O49	300 W Xe lamp	CH4	The oxygen-vacancy-rich ultrathin W18O49 nanowires can be used to design materials with extraordinary photochemical activity because it displayed high CO2 reduction capability in presence of water.	[44]
Zn1.7GeN1.8O	300 W Xe arc lamp	CH4	Zn1.7GeN1.8O loaded with co-catalysts showed significantly higher conversion rate of CO2 into CH4.	[45]
Pt-, Au-, or Ag-loaded mesoporous TiO2	350 W Xe lamp	CH4	The mesoporous TiO2 showed higher efficiency towards CO2 reduction when loaded with noble metal particles, and the order of enhanced photocatalytic activity was Pt > Au > Ag. The optimum loading amount of Pt was 0.2 wt %.	[16]
0.5 wt % Pt loaded ZnAl2O4-modified mesoporous ZnGaNO	300 W Xe lamp (λ = 420 nm)	CH4	The high photocatalytic activity of this photocatalyst was attributed to the improved gas adsorption of the mesoporous structure, the chemisorption of CO2 on the photocatalyst and the narrow bandgap of ZnAl2O4-modified ZnGaNO to extend the light absorption.	[46]
Ga2O3 with mesopores and macropores	300 W Xe lamp (500 mW cm−2)	CH4	Ga2O3 with mesopores and macropores showed high photocatalytic activity due to its higher CO2 adsorption capacity (300%) and increased surface area (200%) compared to the bulk nanoparticles.	[47]
Pt-TiO2 thin nanostructured films	400 W Xe lamp	CO and CH4	The catalyst can be produced at an industrial scale for commercial application and showed high efficiency for selective CH4 formation.	[48]
HNb3O8	350 W Xe lamp	CH4	KNb3O8 and HNb3O8 were synthesized by the conventional solid-state reaction and performed more effectively in photocatalytic CO2 reduction than commercial TiO2.	[49]
ZnO-based materials	8 W fluorescent tube (average intensity 7 mW cm−2)	CO, CH4, CH3OH, H2	N-doping did not show any important influence on the photocatalytic behavior of ZnO-based photocatalysts. The mesoporous structure of ZnO favored CO and H2 production, but catalysts with Cu showed an enhancement in the hydrocarbon production, mainly CH3OH.	[50]
Ag, Pt, bimetallic Ag–Pt and core–shell Ag@silica (SiO2) nanoparticles with TiO2	100 W Hg lamp (330 nm)	CH4	The use of a reactor with three optical windows, a combination of both bimetallic co-catalysts, and Ag@SiO2 nanoparticles increased the product formation significantly compared to bare TiO2.	[51]
Carbon nanotubes Ni/TiO2 Nano-composites	75 W visible daylight lamp (λ > 400 nm)	CH4	Compared to Ni/TiO2 and pure anatase TiO2, Ni/TiO2 incorporated with carbon nanotubes demonstrated maximum CH4 product yield of 0.145 mmol h−1 g−1 catalysts after 4.5 h of irradiation under visible light.	[52]
Pt/Cu/TiO2	200 W Xe lamp	CH4, CO, H2	The addition of co-catalyst Pt decreases the selectivity for CO2 photo-reduction; however, loading Cu onto TiO2 increases the selectivity from 60 to 80%.	[53]
Au/Pt/TiO2	500 W Xe lamp	CH4, CO	Plasmonic photocatalyst Au/Pt/TiO2 provided a more effective way to harvest solar energy by consuming a high-energy photon in the solar spectrum (UV region) and using it for charge carrier generation. Moreover, it also utilized visible light to enhance the photocatalytic activity.	[54]
20 wt % montmorillonite modified TiO2	500 W Hg lamp (365 nm)	CH4	Loading of montmorillonite on TiO2 enhanced the surface area and reduced particle size, thus improving charge separation, resulting in maximum yield for CH4 (441.5 mmol·g·cat−1 h−1).	[55]
0.5 wt % Pt/NaNbO3	300 W Xe lamp (λ > 300 nm)	CH4, CO, H2	The cubic-orthorhombic surface-junctions of mixed-phase NaNbO3 enhanced the charge separation, thereby improving its photoactivity.	[56]
Ag supported on AgIO3 (Ag/AgIO3 particles)	500 W Xe arc lamp	CH4 and CO	In the conversion of CO2 to CH4 and CO using water vapor, Ag/AgIO3 particles showed high and stable activity because of the surface plasmon resonance effect of Ag particles.	[57]
g-C3N4/NaNbO3 nanowires	300 W Xe arc lamp	CH4	An intimate interface formation was suggested between the C3N4 and NaNbO3 nanowires in g-C3N4/NaNbO3 heterojunction photocatalyst, resulting in almost eight-fold higher CO2 reduction than individual C3N4 under visible light irradiation.	[58]
In2O3/g-C3N4	500 W Xe lamp	CH4	The addition of In2O3 nanocrystals onto g-C3N4 surface improved the photocatalytic CO2 reduction process significantly due to the interfacial transfer of photo-generated electrons and holes between g-C3N4 and In2O3.	[59]
SnO2−x/g-C3N4 composite	500 W Xe lamp	CO, CH3OH, and CH4	Enhancement in the surface area of g-C3N4 was observed by introducing SnO2−x. Improve photocatalytic performance was attributed to the increased light absorption and accelerated the separation of electron–hole pairs.	[60]
AgX/g-C3N4 (X = Cl and Br) nanocomposites	15 W energy-saving daylight bulb.	CH4	Under ambient condition and low-power energy-saving lamps, the optimal 30 AgBr/pCN (protonated graphitic carbon nitride photocatalyst) sample showed highest photocatalytic activity with significant enhancement in CH4 formation compared to individual AgBr and pCN photocatalyst.	[61]
Ag supported on Ag2SO3 (Ag/Ag2SO3)	500 W Xe lamp	CH4 and CO	Plasmonic photocatalyst Ag/Ag2SO3 was stable towards CO2 photoreduction after 10 repetitive catalytic cycles with high efficiency under visible light irradiation.	[62]

One of the major obstacles to this research progress is that most of the CO2 reducing photocatalysts are not visible light responsive [63]. In this context, numerous types of photocatalysts have been developed. A few of these catalysts performed under visible light irradiation with high conversion rate and selectivity, whereas other catalysts were weakly responsive under visible light and showed a low rate of reaction yield [64]. The introduction of plasmonic metal onto semiconductor materials to enhance photocatalytic activity has been demonstrated to be very attractive in the visible region.

In the following sections, the basic mechanisms and principles of measuring the efficiency of a photocatalyst in photocatalytic CO2 reduction with H2O are discussed.

2.1. Theoretical Approach

Photocatalysis means activating a semiconductor using sunlight or artificial light. When a semiconductor material absorbs photons of sufficient energy, its electrons are excited from the valence band (VB) to the conduction band (CB), creating electron–hole pairs. VB is the highest energy band occupied by electrons and CB is the lowest energy band in which there is no electron at the ground state [65]. These photo-generated electrons can move to the surface of a semiconductor and react with the adsorbed species on the surface. Meanwhile, electron–hole recombination is also possible [66]. The efficiency of the photocatalytic reaction depends on the competition between these two processes [67].

The basic photocatalytic process can be summarized as follows:

Absorption of photons with suitable energy and generation of electron–hole pairs;

Separation and transportation of electron–hole pairs (charge carriers); and

The chemical reaction of surface species with charge carriers [68,69].

This process is illustrated in Figure 1. As the charge recombination process (~10−9 s) is usually much faster than the reaction process (~10−3–10−8 s), acceleration of the electron–hole separation step remarkably affects the reaction yield [22].

Figure 1

Schematic diagram of photo-excitation and electron transfer process (adapted from [63]).

Apart from the direct photon-excited charge carrier generation process in semiconductors Figure 1, collisions, photon-electron interaction [70,71,72] or electron transfer from the SPR-excited metal nanoparticle [73,74] can also generate electron–hole pairs. However, all of the photo-excited electrons reaching the surface cannot reduce thermodynamically inert and very stable CO2 compound. This reduction reaction is endergonic and requires both hydrogen and energy [19]. Thus, photocatalytic CO2 reduction using sunlight and water has the potential to be the most feasible means to remove atmospheric CO2.

The reduction potential for the various products of CO2 reduction at pH 7 is presented in Table 2. On the one hand, single-electron CO2 reduction reaction requires a highly negative potential of −1.9 eV, which makes the one-electron reduction process very unfavorable. On the other hand, the proton assisted multi-electron CO2 reduction reaction requires comparatively low redox potential (Table 2) and are more favorable. Photocatalysts can facilitate these reduction processes with lower potential. For this purpose, an ideal photocatalyst generally requires two characteristics: (i) the redox potential of the photo-excited VB hole must be sufficiently positive so that the hole can act as an electron acceptor; and (ii) the redox potentials of the photo-excited CB electron must be more negative than that of the CO2/reduced-product redox couple.

Table 2

Reduction potentials for the CO2 reduction process. E0: Standard reduction potential.

Reactions	E0/eV
CO2 + e− → CO2	≥−1.9
CO2 + 2e− + 2H+ → HCOOH	−0.61
CO2 + 2e−+ 2H+ → CO + H2O	−0.53
CO2 + 4e− + 4H+ → HCHO + H2O	−0.48
CO2 + 6e− + 6H+ → CH3OH + H2O	−0.38
CO2 + 8e− + 8H+→ CH4 + 2H2O	−0.24

Upon absorbing radiation from the light source, photo-generated holes in the VB of the photocatalyst oxidize H2O. In addition, the photo-generated electrons in its CB form products such as HCOOH, HCHO, CH3OH, and CH4, by reducing CO2. Here, the relation between the energy levels of the photocatalyst and the redox agent determines the type of reaction that takes place. Figure 2 shows the CB, VB potentials, and bandgap energies of various semiconductor photocatalysts and relative redox potentials of compounds involved in CO2 reduction. The final carbon containing products are determined by the specific mechanism to conduct the reaction. The number and rate of transferred electrons from the photo-generated carriers to the reaction species in the reaction system also contribute in this process [26].

Figure 2

Schematic representation of conduction band, valence band potentials, and band gap energies of various semiconductor photocatalysts and relative redox potentials of the compounds involved in CO2 reduction at pH 7 (Adapted from [22]).

The most commonly used light source for photocatalysis is ultraviolet (UV) light. The high energy content of UV light can effectively excite most photocatalysts. Thus, the majority of publications on photocatalytic CO2 reduction processes are still based on using artificial UV light from high-power lamp [75,76,77]. Only about 4% of solar energy is used by UV light where 43% of solar energy is occupied by visible light; thus, a photocatalyst with a narrow bandgap that can use visible light is in high demand [65,78]. At present, a significant number of studies focus on the direct use of visible light both from artificial and natural sources. Using visible light is more favorable than using UV light because visible light is readily available from sunlight. However, the energy content of visible light is less competitive compared to UV light. Thus, in photocatalytic reduction, the visible light might not provide for an adequate amount of energy for photo-excitation of the catalysts. As such, photocatalysis using visible light and sunlight faces a great challenge [79].

2.2. Measures of Photocatalytic Efficiency

The photocatalytic CO2 reduction efficiency is generally measured by the yield of the product. Here, the general unit for R is mol·h−1·g−1 of catalyst and for the product either in molar units (μmol) or in concentration units (ppm).

In the catalyst-based measurements, the efficiency of the photocatalyst usually depends on the amount of photocatalyst, the intensity of the light, lighting area, etc., so under the irradiation of light, the amount of product formed by per gram of photocatalyst within a certain time period can be measured by its apparent quantum yield. It is calculated by using the amount of product and the incident photon number as shown in the following equations [19,26]. When the photocatalytic reduction reaction gives complex products, then the number of reacted electrons in the equation denotes the sum of the reacted electron to form each product [80,81]. Thus, in light-based measurements, the quantum yield of CO2 photo-reduction into different products can be calculated using following equations:

3. Recent Photocatalysts for CO2 Reduction with H2O

The first step towards enhancing the photocatalytic activity is the selection of a proper photocatalyst. It is a subject of considerable importance both for practical application of photocatalysts and understanding their mechanism. Photocatalysts could be categorized into two basic groups based on their structures: homogeneous and heterogeneous photocatalysts.

The seminal work by Lehn et al. demonstrated the selective CO2 reduction into CO by using Re(I) diimine complexes [82]; since then, the use of metal complexes in photocatalysis has been greatly studied for both CO2 reduction [83,84,85,86] and H2O oxidation [87,88,89]. CO2 is efficiently reduced to form CO when homogeneous photocatalysts, such as Re complexes, are used in the presence of electron donors, such as triethanolamine [80,90,91]. However, CO2 reduction and H2O oxidation processes require distinct reaction conditions.

As a result, carrying out both of the reaction simultaneously using a metal complex catalyst in a single system is a very difficult task. Reverse oxidation of organic products generated from the reduction of CO2 and the reverse reduction of O2 generated from the oxidation of H2O terminate the continuity of the reaction. Figure 3 summarizes these cases briefly [8]. Figure 3a shows the advantages of H2O oxidation of a metal complex catalyst (H2O oxidation site) with a sacrificial electron acceptor (SA). Figure 3b shows the advantages of CO2 reduction for a metal complex catalyst (CO2 reduction site) with a sacrificial electron donor (SD). Figure 3c shows the problems encountered when combining H2O oxidation site and CO2 reduction site: (I) reverse oxidation of products such as organic compounds; (II) electron transfer from H2O oxidation site to CO2 reduction site; (III) need to be electron storage; (IV) need to be active in H2O; (V) easier reduction of O2 than CO2; and (VI) stability in H2O [8]. A number of challenges are encountered in constructing a homogeneous metal complex system for CO2 reduction along with H2O oxidation. The inefficient electron transport between reduction and oxidation catalysts is one of the major difficulties in this process. Another drawback is the short lifetimes of the one-electron-reduced species and the photo-excited state in the presence of O2 generated by H2O oxidation.

Figure 3

Advantages and disadvantages of metal complex catalysts for CO2 reduction with H2O oxidation (adapted from [8]). (a) The advantages of H2O oxidation of a metal complex catalyst (H2O oxidation site) with a sacrificial electron acceptor (SA); (b) the advantages of CO2 reduction for a metal complex catalyst (CO2 reduction site) with a sacrificial electron donor (SD); (c) the problems encountered when combining H2O oxidation site and CO2 reduction site.

Since the pioneering work of Fujishima, Honda, and their co-workers, where they reported the photocatalytic reduction of CO2 to organic compounds, such as HCOOH, CH3OH, and HCHO, in the presence of various semiconductor photocatalysts, such as TiO2, ZnO, CdS, SiC, and WO3 [92], many heterogeneous semiconductor compounds, including metal oxides, oxynitrides, sulfides, and phosphides, had been investigated for this purpose [10,20]. TiO2, BaLa4Ti4O15, SrTiO3, WO3 nanosheet, NaNbO4, KNbO4, Sr2Nb2O7, Zn2GeO4, and Zn2SnO4 are the leading compounds in this list of photocatalysts and the list is increasing enormously in the last five years [1,9,10,18,19,20,28,64,65,93,94,95,96,97,98]. Activation of an inert molecule such as CO2 requires contributions of both incident photons and effectively excited electrons. Thus, the presence of reducing agents can assist the CO2 activation process. It takes advantage of H2O oxidation and CO2 fixation when H2O is used as the reducing agent. Appropriate incident light and suitable semiconductor materials have an important role in attaining this process. Moreover, intensified processing and sensibly engineered strong catalyst with great accessibility are essential to activate the very small molecules under ambient conditions [99]. Some of the desirable properties of an efficient heterogeneous photocatalyst are a high surface area, single site structure, light absorption, an efficient and long lifetime of charge separation, the high mobility of charge carriers, and product selectivity [25].

Table 1 shows the studies on photocatalytic CO2 reduction with H2O to obtain good efficiency and selectivity for specific products. However, this approach is still far from practical implementation. Application of photocatalysis in the environmental and energy industries on a large scale is still limited. Among several difficulties in the heterogeneous photocatalysis, the two major ones are low photocatalytic efficiency and the lack of suitable visible-light-responsive photocatalyst [100,101]. The first one is mostly because of the recombination of photo-generated electrons and holes. For example, the most widely used semiconductor photocatalyst, i.e., TiO2, is well known for its low cost, nontoxicity, and stability with outstanding optical and electronic properties [102,103], but the high recombination rate of photoexcited electron–hole pairs in TiO2 hinders its advanced application [104,105]. Another difficulty is that most of the commonly used photocatalysts like TiO2 and ZnO have large band-gaps, so they can only absorb sunlight in the near UV region. Thus, only a small percent of the solar spectrum is utilized, where many low-bandgap photocatalysts, such as CdS and Fe2O3, show low stability [30]. To resolve these drawbacks, new and more efficient visible-light-active photocatalysts have been studied to satisfy the necessity of future environmental and energy technologies driven by solar power [106]. The development of latest technological advances [82], application of modern synthesis methods to form high-surface-area catalyst nanostructures [83], studies on new co-catalysts to coupled with existing photocatalysts, and investigation on the visible-light-responsive plasmonic photocatalysts are some of the progressing ways of enhancing the photocatalytic activity.

In this section, we present a brief and necessary description on several recently reported semiconductor photocatalysts that exhibit high catalytic activity towards CO2 reduction with H2O. We limit our discussion here to extensively studied TiO2 and modified TiO2 photocatalysts, layered-perovskite photocatalyst ALa4Ti4O5, and ferroelectric photocatalyst LiNbO3 and presented an overview of visible-light-active novel plasmonic photocatalysts in the next section.

3.1. TiO2 and Modified TiO2

TiO2 and modified TiO2 composites are the most commonly used photocatalysts worldwide. In TiO2-based materials CO2 reduction with H2O involves these basic six steps: (i) adsorption of the reactants on the photocatalyst; (ii) activation of the adsorbed reactants by photo-generated charge carriers; (iii) surface intermediates formation; (iv) intermediates to products conversion; (v) desorption of the products from the catalyst surface; and (vi) catalyst regeneration. The dynamics of the reaction process and final products from CO2 reduction are determined by each of these steps. Previous literature has demonstrated that activation and dissociation process of CO2 on TiO2 surface can be increased by creating defect on the catalyst surface (e.g., Ti3+ and oxygen vacancy). By tailoring the crystal phase of TiO2 (e.g., a mixture of anatase/brookite or anatase/rutile), engineering the defects in TiO2 and incorporating modifiers with TiO2 (e.g., metals, metal oxides, graphene, quantum dot sensitizers); the rate of charge separation and transfer can be enhanced [29].

Most studies in this field adopted a solid–liquid interface reaction mode. In such case, particles of a photocatalyst are dispersed or suspended in the aqueous solution, which dissolves CO2. A limited reduction of CO2 and preferential adsorption of H2O on catalyst surface could occur due to limited solubility of CO2 in H2O and direct contact of liquid H2O with the photocatalyst [22]. These limitations could be overcome by using solid–gas or solid–vapor mode of reaction, which can also increase the reduction of CO2. For example, Xie et al. showed that the rate of hydrocarbon product formation increases by more than three times along with decreasing H2 production from H2O when TiO2 (P25) or Pt-TiO2 photocatalyst is placed on a holder surrounded by gaseous CO2 and H2O instead of dispersing the photocatalyst in liquid water (Table 3) [107]. The CO2 reduction selectivity increased pronouncedly from 11–19% to 40–56%. Thus, the solid–vapor reaction mode is better for preferential reduction of CO2 in the presence of H2O. The microstructure of the photocatalysts and the ratio of gaseous CO2 and H2O in gas mode reaction influence the photoactivity and selectivity. For example, Zhang and co-workers found an increased CH4 formation on Pt-loaded TiO2 nanotubes with increasing concentration of H2O molecules surrounding the TiO2 nanotubes, as well as a high concentration of –OH groups on the surface. However, the ratio of gaseous CO2/H2O displayed little effect on product formation over Pt-TiO2 nanoparticles [108]. These results indicated that the adsorption of H2O molecules on the photocatalyst surface can affect the photoreduction in gas mode. Optimizing and modulating the microstructure and surface property of the semiconductor is a very effective way to improve the activity and selectivity of photocatalytic CO2 reduction in gas mode [19].

Table 3

Influence of reaction phase on photocatalytic reduction a of CO2 with H2O using TiO2 and 0.5 wt % Pt-TiO2 photocatalyst [107].

Reaction Mode	Photocatalyst	Formation Rate (μmol·g−1h−1)			R (Electron) (μmol·g−1h−1)	Selectivity for CO2 Reduction (%)
		CO	CH4	H2
Solid–gas	TiO2	1.2	0.38	2.1	10	56
solid–liquid	TiO2	0.80	0.11	5.3	13	19
solid–gas	Pt-TiO2	1.1	5.2	33	110	40
solid–liquid	Pt-TiO2	0.76	1.4	55	123	11

a Reaction conditions: catalyst, 0.020 g; CO2 pressure, 0.2 MPa; H2O, 4.0 mL; irradiation time, 10 h.

The light source also has a strong impact on this reduction process. Varghese et al. reported that the rate of product formation from CO2 reduction is at least 20 times higher under outdoor sunlight than previously published reports, where photocatalytic reductions were carried out using UV illumination [109]. Even though TiO2 is the most widely studied and used photocatalyst, in spite of its high conversion rate, the overall quantum yield is considerably low for the reactions that have been studied. Certainly as low as 10% for most processes [110]. Pure TiO2 shows a lower efficiency towards the reduction reaction due to its high rate of charge-carriers recombination and a shorter lifetime of photo-generated charges. So far, many efforts are been made to utilize this photocatalyst more efficiently including nanostructured TiO2 synthesis, single crystal TiO2, metal or non-metal doped TiO2, dye-sensitized TiO2 etc. The majority of these techniques are expensive at the same time very complex.

3.2. Ag co-Catalyst Loaded ALa4Ti4O5 (A = Ca, Sr, and Ba)

In recent years, a new set of materials unrelated to TiO2 emerged in the photocatalysis study. Layered-perovskite ALa4Ti5O15 (A = Ca, Sr, and Ba) photocatalysts with 3.79–3.85 eV of bandgaps, had been previously reported for effective water splitting [111] and later on also employed for the CO2 reduction by Iizuka et al. In this process, HO− was used as a reducing reagent. They also discussed the factors affecting the photoactivity on the basis of the examination and characterization of the co-catalysts [39]. They found that Ag co-catalyst-loaded ALa4Ti4O15 (A = Ca, Sr, and Ba) reduces CO2 into CO using H2O as an electron donor. For this purpose, Ag co-catalyst-loaded BaLa4Ti4O15 was the most effective photocatalyst.

Although large amounts of reacted electrons and holes are present (in the order of Au > Cu > Ru > NiOx > Ag), Ag is the most active co-catalyst for CO2 reduction, and its photocatalytic activity showed dependence on the loading methods (Table 4 and Table 5). In addition to the size of Ag particles, the unique location of Ag nanoparticles on the working photocatalyst is also important. By a chemical reduction method, Ag particles could be loaded both on the edge and the basal planes of BaLa4Ti4O15, which had a plate morphology with ~100 nm thickness and ~1 μm width. This liquid-phase chemical reduction method is the best for loading fine Ag particles, where the condition of the Ag co-catalysts is changed at the beginning stage of the photocatalytic reaction.

Table 4

CO2 reduction over ALa4Ti4O15 (A = Ca, Sr and Ba) photocatalysts with various co-catalysts a [39].

Photo-Catalyst	Band Gap/eV	Co-Catalyst (wt %)	Loading Method	Activity/μmol·h−1
				H2	O2	CO	HCOOH
BaLa4Ti4O15	3.9	none	-	5.3	2.4	0	0
BaLa4Ti4O15	3.9	NiOx b (0.5)	impregnation	58	29	0.02	0
BaLa4Ti4O15	3.9	Ru (0.5)	photodeposition	84	41	0	0
BaLa4Ti4O15	3.9	Cu (0.5)	photodeposition	96	45	0.6	0
BaLa4Ti4O15	3.9	Au (0.5)	photodeposition	110	51	0	0
BaLa4Ti4O15	3.9	Ag (1.0)	photodeposition	10 c	7.0 c	4.3 c	0.3 c
CaLa4Ti4O15	3.9	none	-	1.3	0.6	0.07	0
CaLa4Ti4O15	3.9	Ag (1.0)	photodeposition	5.6	2.1	2.3	1.3
SrLa4Ti4O15	3.8	none	-	0.8	0.5	0.06	0
SrLa4Ti4O15	3.8	Ag (1.0)	photodeposition	2.7	1.8	1.8	0.5

a Catalyst 0.3 g, water 360 mL, CO2 flow system (15 mL·min−1), a 400 W high-pressure mercury lamp, an inner irradiation quartz cell. b Pretreatment: Reduced at 673 K and subsequently oxidized at 473 K after impregnation (543 K for 1 h). c Initial activity.

Table 5

Effect of loading method of Ag co-catalyst on the photocatalytic activity for CO2 reduction over ALa4Ti4O15 (A = Ca, Sr, and Ba) a [39].

Photocatalyst	Loading Amount/wt %	Loading Method	Activity/μmol·h−1
			H2	O2	CO	HCOOH
BaLa4Ti4O15	1.0	Impregnation b	8.2	5.7	5.2	0.2
BaLa4Ti4O15	1.0	Impregnation b + H2 red c	5.6	8.7	8.9	0.3
BaLa4Ti4O15	0.5	Liquid-phase reduction	4.5	6.8	11	0.03
BaLa4Ti4O15	1.0	Liquid-phase reduction	5.6	12	19	0.4
BaLa4Ti4O15	2.0	Liquid-phase reduction	10	16	22	0.7
BaLa4Ti4O15	3.0	Liquid-phase reduction	9.7	14	19	0.1
BaLa4Ti4O15	5.0	Liquid-phase reduction	4.8	6.6	12	0.02
BaLa4Ti4O15	1.0	Liquid-phase reduction	20 d	11 d	0 d	0 d
SrLa4Ti4O15	1.0	Liquid-phase reduction	4.8	5.8	7.1	0.8
CaLa4Ti4O15	1.0	Liquid-phase reduction	3.2	6.6	9.3	0.4

a Catalyst 0.3 g, water 360 mL, CO2 flow system (15 mL·min−1), a 400 W high-pressure mercury lamp, an inner irradiation quartz cell. b 723 K for 1 h, c 473 K for 2 h, d Ar flow.

In the photo-deposition process, Ag particles of 30–40 nm size are loaded on the edge of the plate predominantly. The photo-generated holes could dissolve the Ag particles on the basal plane, which are then re-photo-deposited on the edge during the photocatalytic reaction. In this case, the re-photo-deposited Ag particles on the edge are smaller than 10 nm and more uniform than the direct photo-deposited Ag particles [39]. The Ag-loaded BaLa4Ti4O15 prepared by impregnation followed by H2 reduction also shows the re-photo-deposition, and the sizes of the re-photo-deposited Ag particles are within 10–20 nm. The CO formation rate in BaLa4Ti4O15 under working conditions shows a change in an opposite sequence of the order of Ag particles size on its edge plane. Thus, smaller Ag particles on the photocatalyst result in higher activity of CO formation. The CO2 reduction mainly occurred on the Ag nanoparticle loaded edge of the BaLa4Ti4O15 plate, whereas H2O oxidation occurred on the basal plane. The unique location of Ag nanoparticles on the photocatalyst can separate the plane of oxidation and reduction reaction, thus increasing the activity of photocatalytic CO2 reduction [107]. This Ag-doped system shows high selectivity for CO2 reduction as indicated by the ratio of CO/H2 (~2.0), but solar energy conversion rate is very low due to the large bandgap of this catalyst [8].

3.3. Ferroelectric LiNbO3

The use of substances with a dipole, which separates the photogenerated electrons and holes, is an important part of surface photochemistry that has not been largely addressed. These substances are ferroelectric materials. The selective oxidation and reduction reactions, which take place on the surface of BaTiO3, was demonstrated by an early work of Giocondi and Rohrer [112]. Subsequent work on the ferroelectric methods PbZr0.3Ti0.7O3 [113,114] and LiNbO3 [115] indicated that the dipole in the ferroelectric material determines the space charge layer structure because of the spontaneous polarization associated with lattice distortions. Ferroelectric LiNbO3 is a promising photocatalyst for CO2 reduction due to its comparatively strong remnant polarization of 70 μC/cm2 [116] than other materials, such as KNbO3 (30 μC/cm2) [117] and lead zirconate titanate [Pb(ZrxTi1−x)O3 (PZT)] (25 μC/cm2) [118]. In spite of the wide bandgap of LiNbO3, which is 3.78 eV, its high remnant polarization was exploited to achieve products from CO2 conversion either under high-pressure mercury lamp illumination or natural sunlight [119]. In the case of solid–liquid reactions, LiNbO3 shows low efficiency in CO2 reduction [120] but in a solid–gas reaction scheme, this ferroelectric material produces seven times the product formed by TiO2 under UV light; under visible light, 36 times more product are produced compared with that by TiO2. This high rate of product formation by LiNbO3 in CO2 reduction with H2O can be explained by its strong remnant polarization, which is absent in TiO2.

Remnant polarization creates an electric field in ferroelectric materials, like LiNbO3, which is similar to usual p–n junction electric field. This electric field separates the photo-excited electrons and holes, leading to an enhanced lifetime of carriers. Thus, photo-excited carriers participate more in redox reactions because there is less chance of charge recombination [121]. In LiNbO3, the decay time of polaron photo-luminescence is very high (9 μs) [122], thus confirming the controlled charge recombination and longer lifetime of carriers. Remnant polarization also causes a charge experienced at the interface of LiNbO3, which interacts with the species in contact with the surface, thereby creating a strongly bound layer [123] and altering the bonding nature in physisorbed materials. Matt et al. suggested that in previous liquid–solid reaction schemes, particularly this tightly bound layer hinders high level of product formation. Another reason of high product formation rate by LiNbO3 could be the more energetically favorable reaction pathway availability than those from lower-energy photo-excited electrons of semiconductors [121]. For these characteristics, LiNbO3 is considered as a promising photocatalyst in concrete construction to improve air quality [124]. An experimental demonstration of this effect was carried out by Nath et al., in which the addition of LiNbO3 to concrete materials reduces CO2 in the presence of water and forms O2 [125]. This relatively new compound has already been used in electronic instruments in place of TiO2 to achieve artificial photosynthesis [126]. These studies clearly showed that ferroelectric LiNbO3 is effective even under weak solar energy in the ambient atmosphere for CO2 reduction with H2O. Further studies are required for the large-scale use of this water-insoluble, chemically inert photocatalyst.

4. Plasmonic Photocatalyst

For highly efficient photocatalysis process, plasmonic photocatalysts have become a topic of increasing interest, in recent years [101,103,127,128,129,130,131]. Nanoparticles of noble metals like Au, Ag, Pt exhibit strong absorption in the UV-visible region due to their surface plasmon resonance (SPR) [132]. SPR simply means the collective oscillations of conduction band electrons in a metal particle and it is driven by the electromagnetic field of incident light [133,134]. It is also known as localized surface plasmon resonance (LSPR).

Dispersal of noble metal nanoparticles of size 10 to 100 nm into semiconductor photocatalyst shows significant enhancement in photocatalytic activity under UV and visible range irradiation. In a conducting material, plasmons are the collective oscillation of the free charge (Figure 4). The oscillations confined to the surfaces of conducting materials are called surface plasmons, which strongly interact with light. When the real part of the dielectric function goes to zero at the plasmon frequency, a resonance in the absorption occurs.

Figure 4

The schematic diagram is representing surface plasmon resonance in a spherical metal particle induced by the electric field component of incident light (adapted from [138]).

The strong interaction with the resonant photons through an excitation of SPR is the characteristic of plasmonic metal nanoparticles. As such, SPR can be defined as the collective oscillation of valence electrons induced by the resonant photon. Au, Ag and Cu nanoparticles show resonant behavior when irradiated by UV and visible photons. As a large fraction of the solar energy consist of UV-vis photons, these noble materials become more promising [103]. This resonance frequency can be turned by manipulating the size, shape, material, and proximity of the nanoparticles [135,136,137].

For example, the plasmon resonance of silver lies in the UV range but can be shifted to the visible range by minimizing the size of its nanoparticles; in the case of the gold, a smaller size of the nanoparticle can shift the plasmon resonance from the visible range to the infrared range [31]. Plasmonic metal nanoparticles exhibit the exceptional capability of concentrating electromagnetic fields, scattering electromagnetic radiation, and converting the energy of photons into heat, which is useful for different applications [103].

4.1. Fundamental of Plasmonic Photocatalyst

The photocatalytic reaction itself is a very complex process; the addition of the plasmonic resonance of noble metal nanoparticles makes it more complicated. The understanding of the physical mechanism of plasmonic photocatalysis is progressing steadily but has not reached unanimity. It is generally accepted that the vital role is played by the energy transferred from the metal nanoparticles to the semiconductors. However, the difference lies in the detailed approach of energy transfer in exciting more number of electrons and holes [30].

The presence of noble metal nanoparticles benefits photocatalysis in different ways. The two very distinct characteristics of plasmonic photocatalysts are SPR or LSPR and a Schottky junction. Even though Schottky junction is not a plasmonic or resonance effect but it is considered as an intrinsic feature while discussing plasmonic photocatalysts. When noble metal nanoparticles come in contact with the semiconductor photocatalysts, it results in the Schottky junction, which builds an internal electric field in a region (space–charge region) inside the photocatalyst, closer to the metal–semiconductor interface. Once the electrons and holes are generated near the Schottky junction, this internal electric field would force them to move in a different direction [128]. Moreover, a fast lane for charge transfer is provided by the metal part [139]; its surface also acts as a charge trap center and can host more active sites for the light-induced reaction. Both the Schottky junction and the fast-lane charge transfer help to minimize the electron–hole recombination process [30].

The surface plasmon resonance (SPR) of the noble metal nanoparticles in response to the incident light is the major attribute of the plasmonic photocatalyst. It brings enhancement in the photocatalytic activity. Depending on the size, the shape, and the surrounding environment, the resonance frequency of noble metal nanoparticles (like Au/Ag) can be tuned to fall in the visible or near UV range [138]. When it falls in the visible light range, the large bandgap photocatalyst such as TiO2 becomes visible-light responsive. Again, for low-bandgap photocatalysts like Fe2O3 [140], SPR can significantly enhance the visible light absorption and UV absorption for large bandgap photocatalysts [127]. This feature is very useful for weakly absorbing materials. Due to the strong absorption, the larger portion of the incident light is absorbed by the photocatalyst surface in a thin layer (~10 nm) providing a shorter distance between the photo-generated electrons and/or holes and the surface, thus making it comparable to shorter carrier diffusion length [73,103,140]. This effect helps materials with poor electron transport. It also contributes in exciting more number of electrons and holes [70,71,72], increasing the rate of redox reaction and the mass transfer by heating up the surrounding environment [141,142,143] and enhancing adsorption by polarizing non-polar materials [142].

In general, the photocatalysis process consists of five individual steps [144,145], starts with reactants transfer to the photo-reactant surface, adsorption of the reactants, followed by redox reaction in this adsorbed phase, then product desorption from the surface, and finally, transfer of product away from the surface. Plasmonic photocatalysts contribute to all these steps. The enhancement in the creation and separation of excited electrons and holes increases the redox reaction rate, the SPR, the Schottky junction, the metal’s fast transfer, charge carrier trapping, and large contact surface has a significant influence here [127,139,140,146,147,148,149]. The heating effect also increases the redox reaction rate [141,142,143,150,151], and benefits the reactant transfer, product desorption and enhancing fluid mixing by product boosting. The polarization enhances the adsorption of reactants [142]. These are the major impacts of plasmonic photocatalysts that had been identified and verified so far [30], which explain how plasmonic photocatalysts mostly show great enhancement in the photocatalytic activity.

4.2. Reduction of CO2 with H2O by Plasmonic Photocatalyst

Noble metal nanoparticles in plasmonic photocatalysts generally coupled with substrates having a larger surface area and active sites. Thus in a co-operative way both the noble metal nanoparticle and the substrate work to enhance the photocatalytic activity [32]. The size, shape, and distribution of noble metal nanoparticles do have an effect on the plasmonic oscillation [152,153,154]. In this section, some of the recently studied plasmonic photocatalysts (Au/Ag), that demonstrate a significant enhancement in the photocatalytic CO2 reduction with H2O are being discussed.

4.2.1. Au Deposited TiO2

In a comprehensive study of photocatalytic CO2 reduction with water, Hou et al. found that depositing Au nanoparticles on top of TiO2 results in plasmonic enhancement [155]. In the visible range (532 nm) of light, the photon energy matches the plasmon resonance of the Au nanoparticles; at this wavelength, a 24-fold enhanced photocatalytic activity was reported.

The strong electric fields created by SPR of the Au nanoparticles locally excite the electron–hole pairs in the TiO2 at a rate several times higher than that by usual incident light; this phenomenon is considered to be responsible for this plasmonic enhancement. The mechanisms of photocatalytic CO2 reduction by Au nanoparticle/TiO2 were investigated under two visible and two UV range wavelengths to separate the contribution of plasmon resonance from the effect of electronic transitions in Au on the overall process. Three basic types of sample were used in this study: (1) bare TiO2; (2) Au nanoparticles deposited on top of TiO2; and (3) bare Au nanoparticles. A quantitative study of the reaction products was conducted to determine the mechanism behind the higher photocatalytic activity. Hou et al. suggested an interband transition hypothesis for the contribution of Au nanoparticles in TiO2 for the increase in photoactivity. This hypothesis was based on the comparative energies of the electrons and holes of the solid material with the redox potentials of the reaction product. Figure 5b shows that for all the three types of sample, CH4 is the only detected product. The product formation in photocatalytic CO2 reduction with H2O under visible light (532 nm) irradiation on Au/TiO2 is significantly high.

Figure 5

(a) Schematic diagrams of bare Au, Au/TiO2 and the bare TiO2 photocatalysts, (b) amount of CH4 formed on these photocatalyst surfaces after 15 h and (c) the relevant redox potentials of CO2 and H2O under visible light and energy band positions of anatase TiO2 and Au (Adapted from [155]).

The explanation for this reaction is obtained by comparing the reduction potentials of the possible products from CO2 reduction with the energies of VB and CB of TiO2 [92,156], as shown in Figure 5c. The reduction potential of CO2/CH4 lies under the CB of TiO2 [157], but for other possible products, i.e., HCOH and CH3OH, it lies above the CB potential of TiO2 [92,156]. This fact indicates the CH4 formation is energetically favorable for photocatalytic reduction of CO2 by TiO2. The reaction scheme for this photocatalytic process is as follows, where symbolizes an electron in the CB and symbolizes a hole in the valance band (VB) [92]:

Oxidation reaction

Reduction reaction

To initiate the reduction process for CH4 formation, electrons from the CB of TiO2 are transferred to CO2 [158]. In the case of bare TiO2-photocatalyzed CO2 reduction, the product yield is very low because the energy of 532 nm wavelength light (2.41 eV) is considerably lower than the band-gap of TiO2 (3.2 eV). For the third type of sample (i.e., bare Au), the amount of product formation is almost negligible, thus suggesting the importance of photocatalytic TiO2 surface in this reduction process. These results correlate with their previous publications [159,160]. Under visible-light irradiation, the sub-bandgap transitions in TiO2 generate electron–hole pairs not that in Au. When the photon energy is sufficiently high to excite d-band electrons of Au to its CB, which lies above the CB of TiO2 in the UV range (254 nm), a different mechanism takes place, resulting in the formation of additional products, including C2H6, CH3OH, and HCHO. As the energy of d-band excited electrons lies above the redox potentials of CO2/C2H6, CO2/CH3OH, and CO2/HCHO, these additional products are formed [155].

4.2.2. Ag Supported on AgIO3

In the photocatalytic conversion of CO2 using H2O, the plasmonic photocatalyst Ag supported on AgIO3 (Ag/AgIO3 particles) displays high activity and stability. In a longitudinal study, He et al. reported the synthesis, characteristics, and application of this plasmonic photocatalyst in CO2 reduction with H2O, where CH4 and CO are produced under visible-light irradiation (>400 nm wavelength) [57]. It was found that Ag plasma induced photo-excited electrons in AgIO3 facilitate the reduction reaction. The comparative photocatalytic activities towards the CO2 reduction of a different photocatalyst in the presence of water under visible light were evaluated by the amount of carbon containing products. Figure 6 shows the increasing amount of CH4 and CO formation with time under visible light range.

Figure 6

Schematic diagram showing time dependence yields of CH4 and CO yields under visible light irradiation over Ag/AgIO3 particles and over N doped-TiO2. The inset shows the time dependence of CH4 and CO yields over AgIO3 under UV-vis light (Reproduced from [57]).

In comparison to N-TiO2, Ag/AgIO3 particles display higher photocatalytic reaction rate for CH4 and CO production. The result of this 240 min reaction indicates the significance of Ag nanoparticles on AgIO3 in photocatalytic CO2 reduction. Under visible light irradiation, the estimated quantum yield is 0.19% for CO2 reduction on Ag/AgIO3 catalysts. The turnover number (TON) are 1367 and 167, respectively, for CH4 and CO formation at 240 min. This information leads to the assumption that each Ag atom exposed to visible light is a potential active site. In photocatalytic CO2 reduction, photoexcitation of Ag electrons to higher energy state is the initial step as AgIO3 itself cannot be excited by visible light. The free electrons in Ag are either promoted by the intraband transitions from the half-filled s band below the Fermi level via surface plasmon excitation to unfilled s band states above the Fermi level or by the interband transition from the d-band to unfilled s-band states [57]. The first interband excitation occurs in Ag nearly at 3.8 eV energy band [161]. However, the energy of light in the visible range (400 nm wavelength) is 3.1 eV, which is less than the required amount of energy for interband transition in Ag; thus, electrons cannot be excited to the ECB (the CB edge potential) from the d band. Thus, the interband transition is not possible. This signifies that the SPR of Ag nanoparticles is the cause of photocatalytic reduction of CO2 in this process. The plasmonic electrons and holes cannot drive the oxidation and reduction half-reactions because the plasmonic charges exist in the Fermi energy of the metal [31]. Therefore, in this process, both oxidation and reduction half-reactions occur on the AgIO3 surface.

The contribution of SPR in CO2 reduction by using Ag/AgIO3 under visible light irradiation was established by studying the wavelength dependence quantum yield. When the light-excited plasmon produced energy or charge was transferred to AgIO3 to drive the photocatalysis, only at that time photocatalytic activity at the SPR wavelength was reported [162]. In AgIO3 the electron–hole pairs are generated by dipole-dipole interaction during resonant energy transfer from Ag to AgIO3 and direct electron transfer between Ag (donor) and AgIO3 (acceptor) [162,163]. These photo-excited electrons lead to the formation of CO by photocatalytic reduction of CO2, where the photo-excited holes lead to the formation of O2 by reacting with H2O. He et al. also ran 10 repeated reaction cycle under visible light irradiation with Ag/AgIO3 plasmonic photocatalyst to examine its stability, where the catalyst showed almost constant photocatalytic activity each time [45]. Their study demonstrated that “Ag/AgIO3 particles manifest high and stable photocatalytic activity in the conversion of CO2 to CH4 and CO using water vapor.”

4.2.3. Ag Supported on Ag2SO3

The plasmonic-semiconductor structure of Ag supported on Ag2SO3 was also mentioned as an effective photocatalyst for photocatalytic reduction of CO2 with water vapor under visible light irradiation by the seminal work of Wang et al. [62]. The major carbon-containing product from this CO2 reduction is CH4, with a small amount of CO. The quantum yield is 0.126%, with an energy returned on energy invested of 0.156%. The study suggested that the energy conversion from incident photons to SPR oscillations of Ag nanoparticles initiates the photocatalytic activity of the catalysts. Ag2SO3 obtained this plasmonic energy by either one or both direct electron transfer and resonant energy transfer. The energy transfer results in separation of photogenerated electron–hole pairs, thereby increasing electron density and transferring the SPR electrons from Ag to the CB of Ag2SO3 by a direct electron transfer process as it lifts the Fermi level of Ag. Moreover, by dipole–dipole interaction, resonant energy transfer from Ag could also generate electron–hole pairs in the Ag2SO3. Thus, for the photocatalytic CO2 reduction, the light-induced sites are provided by Ag nanoparticles and the reaction sites are provided by Ag2SO3.

Figure 7 reveals the yield of products (CH4 and CO) in the presence of Ag2SO3, 1-Ag/Ag2SO3, 5-Ag/Ag2SO3, and 10-Ag/Ag2SO3 as a function of time. The CO yield on 1-Ag/Ag2SO3, 5-Ag/Ag2SO3, and 10-Ag/Ag2SO3 reached 3.06, 4.94, and 2.44 μmol/g, respectively, and the amount of CH4 formation reached 10.55, 12.05, and 7.42 μmol/g, respectively. Therefore, the optimal catalyst is 5-Ag/Ag2SO3. This result can be explained in terms of surface coverage of Ag2SO3 by Ag nanoparticles, where more Ag nanoparticles provide more light-induced sites, whereas fewer reaction sites result in a decrease in the rate of CO2 reduction. For practical applications, evaluating the stability of photocatalyst is an important concern. In this case, Ag/Ag2SO3 is considered as a stable photocatalyst for CO2 reduction because it performed consistently under visible light irradiation even after 10 repeated catalytic cycles. The XRD patterns and surface atomic compositions of new and used catalyst are also quite indistinguishable [62], thus confirming its stability

Figure 7

Formation of the product over TiO2 (P25), Ag2SO3, and Ag/Ag2SO3 photocatalysts under visible light irradiation as a function of time (Reproduced from [62]).

5. Conclusions

From this review, it can be concluded that the environmental challenges are no longer confined issues but become global problems involving climate changes. To solve these problems, we are still far from having a scientific as well as a cost-effective photocatalyst for photocatalytic CO2 reduction with H2O. As previously mentioned, both physical mechanism and product distribution in photocatalytic reactions are very complicated which confines its applicability. The present situation in this area of research is quite confusing, and comparing the efficiency of the different photocatalysts is also difficult due to the high variability of influencing factors and reaction conditions. Details such as mass balance (moles of CO2 converted into the specific product), product distribution, and the amount of reducing agent, time requirement, solution pH, temperature, CO2 pressure, light power, and activity decay over time were not discussed in many of the studies. Comprehensive studies on this process for further movement towards its practical implementation are necessary. An all-in-one standard format that will be unified and widely accepted should be established.

The low photocatalytic efficiency, low response to sunlight, inefficient electron transport between reduction and oxidation catalysts, and high recombination rate of photogenerated species are the major difficulties responsible for the current considerably low rate of average productivity in photocatalytic reduction of CO2 with H2O. Another drawback is the short lifetimes of one-electron-reduced species and the photo-excited state in the presence of O2 generated by H2O oxidation. Despite the fact that UV light can supply more energy than visible light, the availability of visible light from the abundant natural sunlight makes visible-light-harvesting photocatalysts the most desired ones for this process. In its infancy, plasmonic photocatalysts already showed promising performance to overcome the first two shortcomings of the above list. The boundaries of these highly efficient noble-metal photocatalysts are expanding rapidly that it is realistic to expect that plasmonic photocatalysts will contribute significantly in future environmental remedies. Further research should focus on the fabrication of optimal structured visible-light-responsive photocatalysts with a wide bandgap, high rate of photogenerated electron–hole transport, and low rate of recombination, to increase the possibility for practical application of photocatalytic CO2 reduction with H2O.