A Recent Review on Photocatalytic CO₂ Reduction in Generating Sustainable Carbon-Based Fuels

Abstract

Emissions of greenhouse gases from industrial activity, traffic, and solid waste landfills contribute directly to the air pollution crisis. To keep the quantities of carbon dioxide in the atmosphere at a safe level, air pollution must be reduced. Many studies have reported techniques for converting the main greenhouse gas, carbon dioxide, into viable fuels as a method of reducing air pollution. The trending topic of photocatalytic conversion of CO₂ to fuels has shown high potential. This process is environment-friendly due to its capability to be performed at ambient temperature and pressure and hence utilize less energy. This review focuses on the thermodynamic and reaction kinetics of photocatalytic CO₂ reduction. Various types of photocatalysts used for CO₂ reduction are discussed excessively by considering metallic, non-metallic, and composite systems. In addition, necessary experimental and product analysis parameters are covered and mentioned in detail.

1 Introduction

A significant amount of carbon dioxide (CO₂) has been produced and released into the atmosphere as a result of the development of industries. CO₂ levels are expected to rise roughly up to twice their current quantity [1–5]. CO₂ exhaust, being a highly disruptive greenhouse gas, can unbalance the atmosphere and the biosphere accordingly. As a result, encouraging CO₂ reduction and usage limitations is critical. Because CO₂ is a linear molecule with high ionization energy and a low electron affinity, it is easier to reduce than to oxidize [1–5]. To reduce CO₂ and enhance the reaction progress, various approaches have been used. Chemical reduction, photochemical reduction, electrochemical reduction, and biological transformations, for example, are investigated. The CO₂ photoreduction reaction is the most promising among these approaches.

Photosynthesis combines the creation of O₂ and carbohydrates with the fixation of CO₂ using solar light energy. This process creates a readily available carbon source as well as an aerobic environment that can support practically all life forms on the planet [2–5]. With an extra input of energy from photosystem I, photosystem II photoinduces water oxidation, which supplies a key supply of reducing equivalents (water-derived electrons and protons) to transform CO₂ into biomass, food, and fuel [1].

For more than 30 years, the idea of imitating the overall natural photosynthetic cycle of chemical conversion of CO₂ into hydrocarbon fuels has piqued interest. Including biological conversion [1], thermochemical conversion [2], electrochemical conversion [3], and photocatalytic reduction of CO₂ [4]. Due to CO₂’s thermodynamic stability, a large amount of energy is required to break the C = O bond during the conversion process [5]. Compared to other technologies, photocatalytic CO₂ reduction into hydrocarbon fuels is a difficult but promising path. It can be a source of a sustainable alternative to traditional fossil fuels, according to the following benefits: (i) It can be done under relatively mild settings, such as at ambient temperature and pressure; (ii) This process starts with a mass of abandoned CO₂ and is powered by unlimited and pure solar energy; (iii) CO₂ photoreduction may directly generate short-chain hydrocarbon fuels like CH₄, CH₃OH, C₂H₆, and so on, [3–5] alleviating the world's growing energy crisis; (iv) the commercialization of this technique will allow CO₂ to replace fossil fuels as a source of carbon in the chemical industry [5].

In this review, we focus on the reduction of CO₂ using irradiated photons and spotlighting different drawbacks as well. Section 2 depicts the fundamental mechanisms of photocatalytic CO₂ reduction by deeply exploiting the thermodynamics and kinetics of photocatalytic CO₂ reduction (PCR). In Sect. 3, the experimental parameters affecting the process are discussed in detail. Photocatalysts types, selection, and preparations are discussed in Sect. 4. Then, product characterization, analysis, and selectivity are explained in Sect. 5. Section 6 states the main challenges facing PCR and the promising opportunities of this topic ahead. Finally, the conclusion in Sect. 7 sums up the main points and declares the critical parameters to overcome the previously mentioned obstacles facing this research work.

2 Fundamentals, Mechanisms, and Kinetics of PCR

2.1 Thermodynamics of PCR

Table 1 illustrated the standard redox potential ΔE₀ and the Gibb free energy ΔG₀ of the multi-electron water splitting and CO₂ reduction, respectively [6]. The positive ΔG₀ substantiated that the CO₂ reduction process is the endothermic one and it is a high challenge to carry out at room temperature. It also proved that the CO₂ reduction reaction can archive much more energy than the water splitting reaction [6, 7].

Table 1 ΔE₀ and ΔG₀ of the multi-electron water splitting and CO₂ reduction

The multi-electronic processes are more highly captivating than the mono-electronic process because the required energy for electron transfer is smaller. Besides that, the reaction by one electron possesses a bigger reduction potential of −1.9 V versus normal hydrogen electrode (NHE). Therefore, it calls for larger kinetics (overvoltage) [7, 9, 10].

A Latimer-Frost diagram shown in Fig. 1 depicted the multi-proton and multi-electron reduction of CO₂ in a water solution at pH 7 [11]. Table 1 also listed a summary of the standard reduction potentials of CO₂ for the half-cell reactions (at pH 7 in water solution versus NHE) [7, 9, 10, 12–14]. From Table 2 and Fig. 1, we can observe that the thermodynamic barrier was lowered significantly by a multi-electron and proton-assisted approach to the reduction of CO₂ [11]. On the other hand, the conversion from CO₂ to CO, then to H₂CO, and then to alcohols or hydrocarbons also has a lower kinetic barrier if it is collated into the mono-electron process [11, 15]. Thus, the proton-assisted multi-electron transfer is a promising candidate to reduce CO₂.

Fig. 1

(Copyright 2014, Springer. License Number 5311300338819)

Latimer–Frost photograph of the multi-proton and multi-electron processes of CO₂ in water solution at pH = 7.

Table 2 Reduction potentials of CO₂

Via the multi-electron transfer process, several compounds like CH₃OH, HCOOH, and HCHO can be produced from the reduction of H₂CO₃ and the carbonate ions in the solution. The potentials of both H₂CO₃ and CO₃²⁻ mentioned in Table 2 are conclusive that the pathway of methanol formation from H₂CO₃ or CO₃²⁻ is more thermodynamically favorable than that from CO₂ [12, 13].

The thermodynamics of the CO₂ chemical change hindering increases due to the low energy class and high stability of CO₂. Consequently, the targeted conversions need much more energy. Furthermore, due to its inertness, the utilization of catalysts is a considerable requirement [10, 15]. Because the reduction of CO₂ is sophisticated and the transformation of CO₂ into hydrocarbon fuels utilizing assisted catalyst has been expansively reported with different aspects. Some of these approaches are thermochemical, electrochemical, photo-electrochemical (PEC), and photochemical reductions [15, 16]. The lack of releasing more greenhouse effect gas, eco-friendly and economical reduction of CO₂ into value chemicals will become a trending topic if the renewable energies are used widely. Meanwhile, the new routes of the artificial photosynthetic system (APS) e.g. PEC or photochemical reduction of CO₂ into solar fuel, are the urgent aims in the reduction of CO₂ [17]. For instance, Halmann was a pioneer to explore the p-type of GaP which was able to transform CO₂ to CH₃OH via the photo-electrochemical process in 1978 []. Then, Inoue and coworkers reported several chemical products such as HCHO, CH₃OH, and HCOOH by PCR utilizing TiO₂, ZnO, and CdS, GaP, and SiC aqueous sedimentations [14]. Thanks to APS systems, humans can use solar light or CO₂ as chemical energy. In the past decade, photoreduction of CO₂ was intensively studied because of the urgency to find out solutions for environmental pollution problems and their high applicability.

According to the semiconductor materials the production of the photogenerated (PG) charge carriers (electrons and holes), as shown in Fig. 2, are created by the absorption of photons with energy larger than or equal to its bandgap (Eg). The bandgap (energy difference between the valence band (VB) and the conduction band (CB)) plays a crucial key in forming photocatalytic behaviors of the semiconductors [18, 19]. On the surface of the semiconductors, the PG holes diffused from the VB react with water to induce either O₂ or hydroxyl radicals (•OH). Then, •OH radicals oxidize the surrounding organic contaminants on the semiconductors’ surface [18, 19]. On the other hand, the electrons of the CB take part in the reduction process, which reacts either with water to yield H₂ or with CO₂ to fabricate hydrocarbon fuels [18–20]. To attain CO₂ photoreduction, there are many requirements for an elite photo-catalyst. The PCR have to be positioned more positively than the lowermost CB of the photocatalyst. The oxidation potentials of H₂O to O₂ ought to be located at a more negative potential than the uppermost VB. The redox reaction can be carried out under irradiation at a specific energy, which is equal to or greater than the Eg of the photocatalysts [21]. In 1979, the first published article in which the CB of the semiconductor approaches significantly more negative potential compared with the redox potential to perform a certain reaction of CO₂ reduction. Hence, the efficiency of obtaining products from CO₂ reduction escalated [14].

Fig. 2

(Copyright 2014, Springer. License Number 5311300338819)

Bandgap of some semiconductor photocatalysts and the redox potential of PCR at pH 7 in water solution.

Many promising semiconductors shown in Fig. 3 such as TiO₂ [22, 23], ZnO [14, 24–27], ZnS [17, 28–31], SrTiO₃ [32–35], SiC [14, 36–38], Cu₂O [39–42], CdS [14, 43–48], GaP [14, 49], TaON [50–53], C3N4 [32, 54–56], BiVO₄ [57–61] and Ta₃N₅ [62–65] are appropriate for PCR. Among them, TiO₂ is a highly highlighted material, which was studied severally. TiO₂ attracts attention due to its many excellent properties such as being non-toxic and cheap, made up of abundant elements, and resistant to photo-corrosion. Even possessing a lot of good behaviors, TiO₂ is poor visible light absorption because of a wide bandgap, so it ought to be enhanced. Conversely, Cu₂O, CdS, GaP, TaON, C₃N₄, and Ta₃N₅ are promising materials for PCR under visible light irradiation. But, they are quite sensitive because of the weak photostability, so they also should be improved. From Fig. 3, on the right side, many photocatalysts are possessing more negative CB levels suitable for the PCR.

Fig. 3

Steps containing in PCR on a heterogeneous photo-catalyst

2.1.1 Processes of PCR

Not only do the appropriate Eg and CB potential play a salient role, but variously several other factors also significantly affect the yield of the PCR, for example, the photocatalytic process and CO₂ reduction kinetics. According to a typical route of PCR over a semiconductor photo-catalyst, 4 main steps are composing the excitation, the transport, the segregation, and the electro-catalytic reduction of CO₂ and the water oxidation. In detail, the initial process (1) in Fig. 3, which indicates the formation of electron–hole pairs as the semiconductor absorbs the photon energy larger than or equal to the bandgap of the material, is the excitation. Hence, the materials with a narrow bandgap (Eg < 3.0 eV or λ > 415 nm) are high potential candidates for rising the excitation yields of electron–hole pairs by visible light. From these fundamentals, many new novels to synthesize the innovative photocatalysts utilizing the visible light region were intensively studied.

The second process (2) in Fig. 3 describes the segregation of the PG charge carriers and their relocation on the surface for specific chemical reactions. In the third process (3) in Fig. 3, the number of excited charge carriers is decreased dramatically because the holes and electrons recombine and form heat, as a consequence, the efficiency is decreased. This process can call by the deactivation one. To avoid these unwanted phenomena and improve the yield, the transfer of the PG charge carriers to the surface necessitates enhancing and prohibiting the recombination in the bulk plays a vital key. The segregation and the recombination process are dominated by structural and electronic behaviors or photocatalysts. Hence, several beneficial strategies are studied intensively to maximize the segregation and minimize the recombination such as the fabrication of semiconductor/nano-carbon heterojunctions or nanostructured semiconductors or semiconductor heterojunctions.

After the PG electrons come to the surface, the fourth process (4) in Fig. 3 or the electro-catalytic reduction of CO₂, which is frequently a multi-electron and multi-step process related to a cascade of reactions. Electron and photon transfer, C–O bond breaking, C–H/C–C bone creation, and numerous chemical compounds, will be taken place by the PG electrons trapped in the CO₂ reduction co-catalysts (CRC) or the surface active sites [66–71]. Typically, to result in a high yield, the surface of semiconductors ought to be fulfilled by the CRC. While the photocatalytic reduction of CO₂ occurs, does the formation of some kinds of stable products be to need a minimum of two electrons due to some intermediates being easily changed or hard to find out and enumerate. Due to the complicated multi-step mechanism, any approach which enhances CO₂ reduction kinetics ought to be a probable route to improve the yield. The synthesis of mesoporous photocatalysts loading CRC is a good example for such a purpose [9].

Simultaneously with the fourth process (4), the electro-catalytic oxidation of water via PG holes caught in the water oxidation co-catalysts (or the surface active sites) will happen once the PG holes approach the surface and are referred to as the fifth process (5). Enhancing water oxidation can encourage the segregation of PG charge carriers on the surface of semiconductors, hence inducing developed activity of CO₂ photoreduction. Additionally, the sixth step (6) in Fig. 3 shows the surface charge recombination process of the holes and electrons. The surface charge recombination and the efficiency of the photocatalytic CO₂ reduction process are inversely proportional. To reduce recombination, surface trapping ought to be enhanced by increasing the surface properties of the photocatalysts [9].

Besides that, the seventh process (7) and the eighth process (8) in Fig. 3 show the electro-catalytic H₂ evolution and reduction products of the electro-catalytic oxidation, respectively. In the seventh process (7), the consumption rate of PG electrons for CO₂ reduction will be decreased steadily by the H₂ evolution effect. In the eighth process (8), the oxidation of the CO₂ reduction products by the PG holes is detriment for both water oxidation and CO₂ reduction. The quantum yield will be unwantedly affected by these aforementioned two processes. Hence, the efficient approaches to enhance the PCR yield need to comprehensively consider such unfavorable factors [9].

2.2 Possible Mechanisms of PCR

In photocatalytic CO₂ reduction, there are many reduction products formed with the joining of several protons and electrons, thus, the photocatalysts need to meet extraordinarily specific requirements. Table 3 depicts several related reactions containing variable numbers of protons and electrons and their recorded reduction potentials [14, 18].

Table 3 PCR reaction and corresponding redox potentials (versus NHE at pH 7)

Furthermore, the adsorption of CO₂ on the photocatalysts as well as various reaction pathways also will affect the final products significantly. Nevertheless, existing of some ambiguous intermediate species has resulted in some challenges in the reaction mechanism study. Figure 4a and b show two possible CO₂ reduction routes through formaldehyde and carbene pathways [73].

Fig. 4

(Copyright 2020, Royal Society Chemistry License Number 1222250–1)

Schematic demonstration for the a formaldehyde pathway and b carbene pathway for CO₂ reduction. (Reproduced with permission [73].

In detail, the formaldehyde pathway is the fast hydrogenation pathway related to the simultaneous hydrogenation and deoxygenation reactions, namely displayed by the route: CO₂⁻ → HCOOH → HCHO → CH₃OH → CH₄. Whereas, HCOOH is generated by the combination of CO₂⁻, protons, and electrons. Then, they are transformed into HCHO, dihydroxymethyl, two protons, and CH4 in the final step. This pathway is evidenced to be thermodynamically feasible [73]. On the other hand, the CO₂ reduction reaction, which is the fast deoxygenation, can also occur following the carbene pathway (CO₂⁻ → CO → C → CH₃ → CH₃OH/CH₄). Namely, there is the deoxygenation reaction in the initial step and then the hydrogenation reaction in the second one [8, 26].

In the carbene pathway, the CO intermediate usually is transformed easily into the final products because of the weak affinity of the CO products and the surface of the photocatalysts. The CO intermediate can interact with the protons or electrons to generate the CH₃OH or CH₄, the surface of the photo-catalyst possesses a CO strong adsorption capacity. According to the two possible mechanisms, C–O is attacked by protons in both routes. On the other hand, there are some highlighted feature characterizations belonging to each mechanism. For example, in the carbene pathway, the carbon or mixed coordination mode is probably to break the C–O linkage. Likewise, in the formaldehyde pathway, the CO₂⁻ will adapt to oxygen coordination mode [24, 27].

2.3 Kinetics of PCR

Based on the various mechanisms, many types of kinetic equations were generated. For modeling the photocatalytic reduction of carbonate by the TiO₂ under UV light (carried out in an aqueous solution), a Langmuir–Hinshelwood (L–H) equation was expressed. The results show that the photocatalytic reduction rate of carbonate is adsorption-controlled [74]. Whereas, a one-site L–H equation was used further to simulate the reduction rate of carbon dioxide to carbon monoxide (or hydrogen) and methane utilizing water over TiO₂ [74–77]. Additionally, the mechanism of Anpo, in which carbon monoxide is defined as the primary intermediate, was strongly supported by this model [74, 75, 77].

In conclusion, because of the complicated processes of PCR and their kinetics, the yield of PCR on the photo-catalyst surface is so low. The reduction potentials of carbon dioxide conversion to hydrogen are less feasible than to methane or/and methanol in order of thermodynamic aspects. Nevertheless, because of the multi-electron reduction process leading to the kinetics of CO₂ reduction, it is unfavorable. Thus, the enhancement reduction kinetics of CO₂ makes a significant contribution to improving the efficiency. Whereas there are many complicated processes in PCR. Water oxidation, CO₂ activation, and CO₂ reduction kinetics processes also contribute important roles in the overall efficiency. Hence, to improve the overall efficiency, the modification of the kinetics and photocatalytic processes of CO₂ reduction is critical. These can be achieved by the synthesis of high-efficient photocatalysts with significant surface areas [21, 78, 79]. Figure 5 demonstrates all of the critical factors which can influence the overall PCR efficiency.

Fig. 5

Keys influencing photocatalytic yield and corresponding design strategies

3 Experimental Parameters Consideration

Photocatalytic conversion of CO₂ into value‐added hydrocarbon fuels or useful chemical products has been the focus of active research. The photocatalytic conversion rate of CO₂ to methanol depends on the photocatalyst used, the photoreactor design, and experimental parameters. Here, we introduce several considerations for photocatalytic CO₂ reduction experiments. Preparation of catalysts, light source, and type of photoreactors as the PCR experiments preparation. Experimental parameters include reaction temperature, CO₂ pressure, and presence of impurities adsorbed on the photocatalyst.

3.1 Preparation of Catalysts

There are numerous articles have been reported, including pure, doped, metal–organic framework based and composite photocatalysts, synthesized by using various methods, and used for the photocatalytic conversion of CO₂ into fuels [80–87]. They are discussed excessively in section four.

3.2 Source of Light

Among the important criteria for photocatalytic CO₂ reduction are light intensity and irradiation nature [88]. The number of photons falling on a unit area in a unit of time determines the intensity of an illumination source [89]. According to the literature, most photocatalysts work better in the UV region, hence they can only capture a small portion of the solar spectrum irradiation [90]. To get over this limitation, semiconductors are being modified to harvest a wider range of sunlight spectrum.

Aside from that, light can be focused and diverted to increase irradiation photon flow [91]. By focusing light over TiO₂ and Pt/TiO₂, Han et al. were able to reduce CO₂. The authors tested CO₂ photoreduction with various concentrating ratios (CRs). CRs are defined as the ratio of concentrated light flux (amount of energy per unit time per unit area) on the photocatalyst surface to ambient flux (under non-concentrated conditions).

By varying the distance between the Fresnel lens (placed between the light source and the photocatalyst) and the photocatalyst surface, the light irradiation is concentrated and modulated, resulting in variable light intensities with different light concentrated focal regions. The optimum concentration ratio (CR) increased dramatically, according to their findings [91]. Based on these findings, better light interaction with the photocatalyst under ideally concentrated light will result in a significant improvement in the yield [92]. Employing a greater intensity radiation source, the geometrical design of the photoreactor, the lamp to photoreactor distance, and the use of fiber optics are all appropriate techniques for maximizing the radiation intensity on the reaction medium [89].

3.3 Geometry and Design of Photoreactors

Photo-reactor geometry and design also encourage maximum photon flux distribution, allowing for a large active surface area with a high mass transfer rate and minimal light diffusion effects. However, in most reactor geometries when light impinges on the photocatalyst's surface from the center or side, a shadow is created on the other side, preventing a significant amount of the photocatalyst from being activated. Fabricating suitable reactor designs to achieve a consistent distribution of light and greater photocatalyst dispersion could improve light-photocatalyst contact [93]. A variety of ways have been documented in the literature to achieve this, including the use of various reactor geometries and catalyst supports [93–95].

Monoliths have been used as a photocatalyst support material in innovative photoreactors, attempting to overcome mass transfer restrictions and limited light distribution efficiency seen in immobilized photocatalysts. Monoliths have consistent structures, supported with parallel channels that come in a variety of shapes and sizes depending on how they are extruded. These materials have a large surface area per unit volume and other appealing characteristics such as minimal pressure drop, high mechanical strength, and thermal stability, making them superior to traditional catalyst arrangements (powders and pellets). By comparing the performance of TiO₂ coated micro channel monolith and cell type support, Tahir et al. investigated the effect of photocatalyst dispersion (dispersed as a single layer over stainless steel cell). The TiO₂-coated monolith showed a considerable increase in CO output in their research. This increase was mostly due to a larger exposed photocatalyst surface that was available for photocatalytic activity. This improvement was attributed to the monolith's increased illuminated surface allowing for more efficient photon usage [97].

Fiber optic reactors have an advantage over packed bed reactors in terms of photocatalyst dispersion and light spreading across a vast surface area. Nguyen et al. investigated the yield of photocatalytic CO₂ reduction using photocatalyst coated on optical fiber against photocatalyst coated on a glass plate. For the same amount of photocatalysts, their research showed 15.2 times CH₄ yield and increased C₂H₄ yield 11.6 times more. The synergistic effects of catalyst dispersion and effective light utilization could explain this [97].

Wang et al. used a fiber optic reactor to perform CO₂ photoreduction and ascribed the increased output to the progressive and uniform distribution of light during irradiation. Optical fibers have the advantages of catalyst support and effective light distribution, but they also have the disadvantages of limited reactor capacity utilization and shorter light transit distance from the point of incidence. They occupy 20–30% of the reactor capacity, but the effective use of incident light is limited because of the limited catalyst-coated area [99]. Ola et al. fabricated an internally illuminated monolith reactor by combining the mutual effects of greater monolith surface area and effective light distribution of fiber optics and compared the CO₂ reduction performance of this system to that of a slurry reactor. Due to the higher surface area of the monolith and the equal dispersion of light by optical fibers, it was discovered that internal illumination of the monolith reactor by optical fibers increased quantum efficiency by 23 times [97]. Optical fibers constructed of carved polymethylmethacrylate were placed into a NiO/InTaO₄ coated monolith (honeycomb structure) by Liou et al. When used for photocatalytic CO₂ reduction, this reactor increased product yield (methanol and acetaldehyde). Increased surface area, better photocatalyst loading, and effective light use are all factors that contribute to a higher yield [99].

3.4 Effect of Temperature Variation

Due to long-wavelength irradiation, concentrating solar light raises the temperature depending on the CR [12, 28, 29]. Photocatalytic CO₂ reduction at high temperatures is promising because it bypasses the thermal barriers that cause slow reaction rates and low yields [96]. The effectiveness of temperature rise in photoreaction can be demonstrated by increased effective collisions between photogenerated charges and reactants, which are directly proportional to the reaction rate [97]. Furthermore, increasing temperature increases the desorption of products, allowing CO₂ to adsorb on unoccupied sites, resulting in a faster reaction rate [98].

The reaction temperature influences photocatalytic CO₂ to methanol conversion. Thus, determining the ideal temperature is difficult since temperature affects the methanol production rate in four different ways, as shown in Table 4. First, the amounts of CO₂ that can dissolve in water is affected by temperature, therefore decreasing the temperature increases CO₂ solubility in water. The increasing samount of dissolved CO₂ in water can speed up the synthesis of methanol. When water is chilled from 25 to 0 °C, the solubility of CO₂ in water increases by around 2.5 times, resulting in a 2.5-fold increase in dissolved CO₂ [89, 97, 98]. Second, the temperature affects the ease with which reactants adhere to the surface of the catalyst.

Table 4 Effect of temperature on various factors indirectly affecting the photocatalytic CO₂ reduction rate [89]

Because there is less thermal agitation at low temperatures, reactants soak more quickly onto the catalytic surface. The photocatalytic reaction rate naturally increases with increasing the amounts of reactants adsorbed on the catalyst surface. Third, temperature influences the rate of product desorption, affecting catalyst poisoning and, as a result, catalyst availability. Catalyst poisoning occurs when reactive intermediates and products are more likely to remain adsorbed on the catalyst surface at lower temperatures, preventing further catalytic activity. As a result of the scarcity of vacant catalytic adsorption sites on the surface, fresh reactants are unable to adsorb on the surface, slowing the photocatalytic CO₂ reduction [89, 97, 98]. Fourth, the decreased temperature has an undesirable influence on the photocatalytic CO₂ reduction process because it reduces the diffusion rates and collision frequencies of the reactants, resulting in a lower reaction rate constant for methanol generation. These four temperature impacts of the photocatalytic reaction show that the negative effects of high and low temperatures can be avoided without sacrificing the favorable effects [89]. When the temperature was raised from 25 to 75 ℃, Wang et al. discovered that the production rate nearly doubled [99]. However, the reaction temperature should not rise too high, as this may cause the CO₂ to desorb, slowing down the photoreduction process.

3.5 Flowing CO₂ Gas Pressure

The CO₂ gas pressure in the reaction chamber, like reaction temperature, plays an essential role in controlling the rate of CO₂ reduction in product production. Increased CO₂ solubility in water with pressure causes an increase in reaction rate and, as a result, an increase in reaction product generation rate. Because cooling down the solvent can lower the product desorption rate, clog the catalyst surface, and slow down the reduction process, boosting CO₂ solubility by raising pressure is preferred over reducing the temperature for the goal of enhancing photocatalytic CO₂ reduction. Aside from enhanced reaction products, increasing CO₂ concentration caused by increased CO₂ pressure in aqueous media has also been shown to improve product selectivity. The increase in methanol formation rate, on the other hand, does not increase endlessly with pressure. It increases to a maximum value at optimum pressure and then begins to drop as pressure is increased further. Mizuno et al. investigated the effect of CO₂ gas pressure on the rate of reaction product formation and found that methanol was produced at ambient pressure [100]. However, as CO₂ pressure increased, the rate of methanol formation increased sharply, peaking at 1.0 MPa, and then declining significantly with further CO₂ pressure increases. When choosing CO₂ reduction reactor designs [101], the expense of producing advanced high-pressure systems must be taken into consideration.

3.6 Effect of Contaminant

Organic contaminants in the catalyst or on the catalyst surface might cause a falsely positive result since they are often reduced more efficiently than CO₂ to create diverse reduction products. This is especially true when the concentration of products obtained from photocatalytic CO₂ reduction is low, as it has been in almost all photocatalysis experiments.

Small particle size and wide surface area of the TiO₂ photocatalyst (and this also applies to other photocatalysts) are crucial elements in adsorbing airborne organic pollutants, which are more reactive than even CO₂, according to Neatu et al. [102]. As a result, they are more easily reduced than CO₂ and become the source of photoproducts, which are sometimes mistaken for CO₂ reduction reaction products [103]. When the photocatalyst amount is large but the produced yield is low this can be attributed to the adsorbed contaminants. This condition can be avoided by calcining the photocatalyst before the photocatalytic experiment to remove any carbon-containing impurities that may be present on its surface. Because CO₂ is more stable than other carbon-containing organic contaminants, organic impurities linked to photocatalyst surfaces degrade more quickly, resulting in an overestimation of photocatalytic activity and yield. In summary, photocatalyst contamination can lead to false results, hence rigorous surface cleaning is required to obtain accurate and precise results.

4 Photocatalyst Types, Selection, and Preparation

4.1 Metal System

CO₂ photocatalysts are divided into five types: metal, mixed oxides, metal sulfides, polymeric materials, metal–organic framework, and others. Metal and mixed oxides, metal sulfides, and the metal–organic framework are included in the metallic system, which reduces CO₂ into several forms, such as carbon monoxide, methane, methanol, and formic acid for fuel. Some examples included in the metal and mixed oxides are titanium oxide, iron oxide, tantalum oxide, copper oxide, niobium oxide, and strontium oxide. Furthermore, some examples of metal sulfides included are cadmium sulfide, zinc sulfide, MoS₂, SnS₂, Bi₂S₃, In₂S₃, and Znln₂S₄. Meanwhile, MIL 101, PMOF-AI, Co-ZIF9, and MOF 525-Co are examples of metals included in the metal–organic framework (MOF) group [104]. Activity and selectivity are essential properties in the selection of photocatalysts. To obtain a photocatalyst with high activity and selectivity, several things must be considered, including band structure, surface state, and photoreaction conditions. Increasing the activity and selectivity of photocatalysts can be done by optimizing carbon dioxide absorption, optimizing light harvesting, charge separation effectiveness, and synergistic effects [104].

4.1.1 Metal and Mixed Oxides

As previously described, TiO₂ is the most widely used CO₂ photocatalyst of metal oxides and mixed oxides. This is due to its non-toxicity, stability, slightly corrosive nature, and low cost. This metal is widely used to convert CO₂ to methane and carbon monoxide. However, TiO₂ is known to have a large bandgap (3.2 eV); due to this large bandgap, TiO₂ only shows photocatalytic activity under UV radiation. In contrast, sunlight as a source of radiation is still not qualified. Therefore, the development of this photocatalyst to have a narrower bandgap is still being carried out [104]. Several strategies to increase the ability of TiO₂ are doping using Ag and Cu or using co-doping [104, 105]. This doping can reduce bandgap energy, improve interfacial charge transfer, trap electrons, and allow the use of visible light [104, 105]. Another strategy is to use heterostructured crystal growth. The next modification is to insert the defect chemistry into the forbidden gap on the catalyst surface by thermal treatment [104, 105]. Modification through the formation of nanomaterials on TiO₂ also shows advantages in the catalytic behavior. Modifying the nanostructure can increase the diffusion rate and surface area to increase the catalytic activity [105]. Furthermore, to increase the efficiency of the excitation process, modifications using a dye sensitizer can be made by increasing the wavelength so that the catalytic activity increases [104, 105]. There are several methods for TiO₂ synthetizing including sol–gel, hydrothermal, impregnation, one-pot, and co-precipitation. Sol–gel process, hydrothermal synthesis, and one-pot synthesis were implemented using TiO₂ and multi-walled carbon nanotube composites to improve the photocatalytic performance [104, 105]. A co-precipitation method can narrow the bandgap to absorb large amounts of energy from visible light. Synthesis using the impregnation method was performed by increasing irradiation and doping with metal oxides such as CuO, CoO, and Fe₂O₃ [105] (Fig. 6).

Fig. 6

TiO₂ synthesis methods [105]

Silver, rhodium, gold, palladium, and platinum are metals that are often used as co-catalysts with TiO₂ to increase TiO₂ occupation. The rate of methane formation is reported to be increased due to the use of noble metal co-catalysts. One of the noble metals, Palladium, acts as a co-catalyst on the TiO₂ surface and provides active sites for CO₂ adsorption and activity [106]. In addition, the size of metal nanoparticles is a critical factor to determine the activity and the rate of methane formation [106].

Copper is a transition metal that is often used as a TiO₂ co-catalyst. This is because they are abundant, non-toxic, low in cost, and potential alternatives to the use of noble metals. The use of this co-catalyst showed an increase in the formation of methane. In addition, the use of Cu doping on TiO₂ under UV-rich illumination can convert CO₂ into formic acid [106]. The use of Cu-TiO₂ was also reported to increase the efficiency of light harvesting under UV conditions to produce carbon monoxide and methane from CO₂. Copper has also recently been reported to depress hydrogen development in CO₂ photoreduction, thereby selectively producing CO [106].

Subsequently, the use of binary co-catalysts (Cu₂O-Pt/TiO₂ and MgO-Pt/ TiO₂ systems) was investigated. Platinum in TiO₂ is intended to promote electron capture and inhibit charge pair recombination. However, the use of Platinum increased H₂, so it takes Cu₂O or MgO to suppress the formation of H₂. As a result, CH₄ will be obtained with high selectivity. Recently, multi-heterojunctions were fabricated on TiO₂-MnOx-Pt films. This multi-heterojunction can efficiently separate charged pairs to produce three times higher CH₄ and methanol than pure TiO₂ [106].

Au shows the effect of surface plasmon resonance on TiO₂. Synergistic mixing of the plasmonic effects of Au and Pt nanoparticles as electron absorbers with TiO₂ nanofibers was reported to increase visible light harvesting and inhibit the recombination of photoexcited TiO₂. The reduction of CO₂ under UV light and visible light for the Au-Cu alloy on TiO₂ as a photocatalyst showed excellent performance. The electron selectivity for CH₄ evolution can reach 97% under visible light irradiation. This high light harvesting ability comes from the plasmonic effect of Au [106].

Cu is one of the most commonly used metals for CO₂ reduction photocatalyst. This is because Cu has low bandgap energy and a high conductive band. CuO and Cu₂O are reported to have small band gaps of 1.7 and 2.2 eV. CuO nanomaterials are known to absorb visible light effectively and produce photogenerated electrons and holes. Cu-based photocatalysts have three pathways for CO₂ reduction. There is the formaldehyde, carbinol, and glyoxal pathway. The Formaldehyde pathway can produce formic acid, methane, and methanol from Cu-based photocatalysts for CO₂ reduction. The carbene pathway is usually used to produce methane and methanol. While the glyoxal pathway is usually used to produce formic acid [107].

The reduction activity and selectivity of CO₂ can be affected by the morphology, particle size, and dispersibility of Cu when different methods are used. CuO which has a high density with poor dispersion will affect light absorption due to masking between particles. To address this problem, it is possible to utilize CuO thin film to enhance the catalytic effect. In addition, the use of glass fiber mesh coating CuO can also be done to increase the production of CH₄. The photocatalytic activity and selectivity of CuO products can also be increased through size modification or create quantum dots of CuO. Modification of the size of CuO to be smaller will increase light harvesting and charge transfer separation due to the risen uniform distribution [106, 107]. The modification with CuO quantum dots plays a significant role in the CO₂ adsorption and activation [107]. There are several methods to synthesize CuO. Among them, the solvothermal method and impregnation are quite famous [107].

The use of pure Cu is also carried out to minimize semiconductor energy bandgap. This reduced band gap energy can make full use of visible light to improve light utilization and encourage the practical application of photocatalysis. Cu can be prepared by microwave hydrothermal method and secondary calcination. Cu₂O was studied as a photocatalyst to reduce CO₂. This metal oxide is known to have high photocatalytic activity. Doping using Cu₂O is known to improve its energy band structure and increase photocatalytic activity. Cu₂O can be prepared by hydrothermal deposition-reduction method and microwave-assisted in situ reduction chemistry [107] (Fig. 7).

Fig. 7

Cu synthesis methods

Perovskite is a metal oxide that is also used as a photocatalyst for CO₂ reduction [108]. Perovskite is used as a photocatalyst because it provides a broad spectrum for CO₂ conversion. In addition, perovskite also has good stability, flexible composition, efficient catalytic activity, long charge diffusion, low cost, and easy preparation [108]. Perovskite converts CO₂ through light harvesting, which then separates electrons and transfers them from VB to CB. Then charge photogeneration is carried out before the redox reaction occurs on the catalyst surface until finally the product is formed [108]. There are several classifications of perovskite oxides, including ABO₃, halide, layered, ruddlesden-popper phase, aurivillius phase, (110) layered, (111) layered, Dion-Jacobson phase, and oxynitrides [108].

ABO₃ perovskite consists of alkali metal cations at site A and transition metal cations at site B. These transition metal cations regulate the perovskite catalytic activity and electron mobility. One way to narrow the perovskite band gap can be done by forming a solid solution. This is because the solid solution has a lower CB with a higher VB. In addition, the formation of this solid solution can also reduce the CO₂ induced due to some perovskites having a band gap not suitable for CO₂ reduction under visible light. This solid solution formation also offers advantages such as band gap control, charge transfer, and chemical stability [108].

Perovskite halides (ABX₃) are another type of perovskite group where the cation at site A is usually Cs⁺ or Rb⁺, while the cation at site B is Pb, Ge, or Sn and the halide is located at site X. Perovskite halides are usually used to produce CH₄ and CO at the surface of the catalyst. There is also layered perovskite which has flexibility in its structure, effectiveness in charge transport, and attractive optoelectronic characteristics [108].

The Aurivillius phase is a perovskite with cations such as Na, K, Ca, Sr, Ba, and Bi located at site A while, cations such as Fe, Cr, Ti, Ga, Nb, V, Mo, and W are located at site B. On the other hand, the Dion-Jacobson phase (A_n1B_nO_3n+1) is a perovskite with cations such as Rb, K, Ag, and Cs at site A while Pr, Sm, Nd, and La at site B. This perovskite has a band gap of about 3.8–4.3 eV which allows it to work under UV light. Ruddlesden-Popper phase (A₂A_n−1B_nO_3n+1) is a perovskite consisting of alkali metals at site A and transition metals at site B. Finally, perovskite oxynitrides (ABO_2−xN_1+x) is a perovskite consisting of alkali metals at site A with transition metal at site B [108] (Fig. 8).

Fig. 8

(Copyright Elsevier License Number 5326340974814)

Perovskite classification [108].

Besides being used as the main catalyst, several metal oxides can also be used as co-catalysts including Cu₂O, CuO, NiO, MgO, CO₃O₄, and Fe₂O₃. The use of copper oxide as a co-catalyst can intensify CO₂ adsorption on the surface and suppress charge recombination. Cu is also able to increase the electron density at the active site when it is excited and facilitates CO₂RR which requires multi-electron transfer. Photocatalyst activity can be increased through the formation of oxygen vacancies at the active sites of Cu₂O. Not only that, but this formation is also able to reduce the rate of charge recombination. CuO is usually used as a co-catalyst together with SiC, TiO₂, K₂Ti₆O₁₃, and NaTaO₃ as the main catalyst. Magnesium oxide itself is usually used as a co-catalyst to increase the activation and adsorption of CO₂ because of its ability to interact with CO₂ strongly. This metal oxide is usually used with TiO₂ to catalyze the reduction of CO₂. Meanwhile, nickel oxide is usually used as a co-catalyst to inhibit the reverse reaction and increase charge carrier separation. This co-catalyst can increase the production of methanol, methane, and carbon monoxide. Nickel oxide is often used as a co-catalyst along with InTaO₄, InNbO₄, KTaO₃, K₂Ti₆O₁₃, and InVO₄ as the main catalyst. Last but not least, is the use of cobalt oxide as a co-catalyst for the promotion of the oxygen evolution reaction and suppressing corrosiveness due to the buildup of photogenerated holes [109].

4.1.2 Metal Sulfides

On the other hand, metal sulfides such as ZnS and CdS have also been reported. Similar to TiO₂, ZnS only absorbs visible light due to its large bandgap (3.66 eV). In contrast to CdS which has a smaller bandgap (2.4 eV), this metal is preferred for CO₂ reduction. The catalytic performance of metal sulfide can be improved by combining metal sulfide photocatalysts with other photocatalysts. Incorporating metal sulfide photocatalysts with different photocatalysts will provide more advantages in bandgap regulation. The use of metal sulfides as photocatalysts is known to have great research potential. Unfortunately, metal sulfides have less stability during photocatalytic processes leading to structural damage [104].

Molybdenum disulfide (MoS₂) is a layered binary sulfide widely applied in photocatalytic due to its excellent optical/electrical properties and flexible electronic band structure. The layer on MoS₂ has a high d electron density, so it has potential in gas–solid CO₂ photoreduction systems for methanol production. Bismuth sulfide (Bi₂S₃) is known to have a narrow bandgap (≈1.3 eV), so it is also getting attention in its role as a photocatalyst. In addition, this photocatalyst is also known to have low toxicity, high biocompatibility, good performance in reducing CO₂, and an absorption coefficient in the visible light region (>105 cm⁻¹). Like Bi₂S₃, In₂S₃ is also widely used for CO₂ reduction. This metal has an expansive light response range due to its narrow bandgap (2.0–2.3 eV), and low toxicity [72] (Fig. 9).

Fig. 9

Metal-sulfide band gap

One method that can be used to synthesize ZnS is heat-free synthesis with Cu+ and Cd²⁺ co-catalysts doping to maximize light utilization [72]. There is also the synthesis of ZnS with zinc hydroxide and thiourea using the surface of Ru nanoparticles to form formic acid, which is known to provide high product selectivity [72]. Metal sulfide type MoS₂ can be synthesized by hydrothermal method and Chemical vapor deposition (CVD) synthesis. Likewise, with the synthesis of In₂S₃ through the hydrothermal cationic exchange method [72]. SnS₂/SnO₂ type metal sulfide photocatalyst can be synthesized by hydrothermal method and sequential template to maximize CO formation [72]. The ion-exchange method between Bi and Cd₂S in ethylene glycol solution can be carried out to synthesize Bi₂S₃/CdS/FeTCPP photocatalysts [72] (Fig. 10).

Fig. 10

Metal-sulfide synthesis

Besides being used as the main catalyst, metal sulfide can also be used as a co-catalyst because it can increase charge separation and add more active sites. The co-catalysts commonly used in this group are MoS₂ and NiS₂ [9, 109]. MoS₂ is usually used together with Bi₂WO₆ under visible light to produce both ethanol and methanol. One method that can be used for the preparation of this co-catalyst is impregnation-calcination method. In addition, MoS₂ is also used with TiO₂ to produce methanol using the in situ grown method [109]. On the other hand, NiS₂ is used as a co-catalyst with graphite carbon nitride (g-C₃N₄) to produce CO due to its ability to accelerate photogenerated electron–hole pair separation. In addition, NiS₂ can also be used with ZnO to produce CO and CH₄. However, the use of metal sulfide as a co-catalyst is also reported to have drawbacks due to its poor stability [109].

4.1.3 Metal–Organic Framework

Another metal system that is widely used is MOF. MOF can be used as co-catalysts as well as stand-alone photocatalysts. The MOF structure is potentially a photocatalyst because the network is porous, regular, and heterogeneous so that ions and CO₂ can move freely into the matrix. MOF is also known to increase the reactivity of other catalysts. However, as with metal sulfides, MOFs have poor stability and problems in economic viability on an industrial scale. The use of MOFs as photocatalysts can still be improved by modifying the structure and adjusting the reactive functional groups. The metal-added MOF photocatalyst has a bandgap from 1.52–2.4 eV. Meanwhile, if MOF is used as a composite, the bandgap variation is 1.6–3.1 eV [104].

MOF can be synthesized through solvothermal, slow evaporation, microwave assistance, electrochemistry, sonochemistry, and mechanochemistry. These five methods generally combine three main components: metal salts, ligands, and solvents. The solvothermal method is the most frequently used method for MOF synthesis by involving high boiling point heating between organic linking ligands and metal salts in a solvent. Slow evaporation synthesis is preferred because it does not require external energy and uses only room temperature, even though it takes a long time. Microwave-assisted synthesis involves heating a solution with microwaves to produce nano-sized crystals. The electrochemical method is carried out by adjusting the pH of the solvent at room temperature without metal salts and consisting only of a mixture of organic linkages and electrolytes to provide metal ions. Synthesis using the sonochemical method is based on molecular changes due to ultrasonic wave radiation to produce fine crystalline materials. Finally, the mechanochemical synthesis method is based on applying a mechanical force without a solvent to form a porous MOF [110].

Metal systems have several drawbacks, including some metals toxicity, being not environmentally friendly, and have low selectivity. However, metal can still be used, such as doping carbon material as a photocatalyst. In addition, the use of bimetallic MOF is also known to be more efficient than the monometals (Fig. 11).

Fig. 11

(Copyright Elsevier License Number 5326341140976)

MOF synthesis [110].

Besides being used as the main catalyst, MOF can also be used as a co-catalyst for CO₂ reduction in nanocomposites. MOF acts as a kinetic process driver in the catalytic reaction and increases CO₂ adsorption while other components act as light harvesters. For example the use of Co-Zif-9 as a co-catalyst together with [Ru(bpy)₃]Cl₂.6H₂O as a photosensitizer, and TEOA as an electron donor under 380–700 nm irradiation. Although the use of this co-catalyst is known to have high catalytic efficiency, the selectivity of each product is quite low. Not only that, but photobleaching also causes a decrease in catalytic activity. So, it is necessary for coupling with semiconductors such as g-C₃N₄ and CdS as light harvesters. In addition, MOF is also used as a co-catalyst with TiO₂. This system shows that the photocatalytic activity for methane production increases up to five times [111].

Homogeneous metal complexes have been widely used for CO₂ reduction photocatalysts. However, this metal complex does not have long-term stability, is difficult to separate from the reaction mixture, the product will be contaminated, and it is difficult to recycle. Therefore, the use of MOF as a host to support homogeneous metal complexes is considered the right choice. This is based on the nature of the MOF which has a high surface area and uniform pores, so it can be adapted for the diffusion of reactants. MOFs also have well-defined and isolated sites for anchoring catalytic species, so they can be used to construct single-site catalysts [111].

4.1.4 Metal Complex

Transition metal complex ions have been widely used as photocatalysts. Some examples of metal complexes used as photocatalysts include complexes of ruthenium, osmium, rhenium, cyclometallation iridium, metalloporphyrins, and organic dyes. This photocatalyst has several properties, including selective light absorption, relatively inefficient reverse electron transfer rate, fast decomposition, sufficient reducing power for electron transfer, stability in ground and excited states, and reacts selectively to CO₂. The metal complex-based homogeneous photocatalytic system consists of a photosensitizer and a CO₂ reduction catalyst. The Ru(II)-Re(I) photocatalyst is reported to reduce CO₂ efficiently when installed on solid materials. Photocatalysts made with photosensitizers and metal complex catalysts showed efficient, selective, and long-lasting results. Building a photocatalyst with a hybrid system can be done by connecting a photosensitizer and a catalyst to a solid material so that electron transfer will take place faster [111].

The Re dynamic complex has CO₂ reduction activity and high product selectivity. The Re(bpy)(CO)₃Cl complex has a stable and adjustable structure so that it has the potential as a CO₂RR photocatalyst. When photocatalytic CO₂RR is carried out in a mixed solution of DMF/TEOA, the addition of TEOA can capture CO₂ even at the atmospheric level. Re(bpy-)(CO)₃ has also been reported to react with CO₂ in the dark. In addition, the use of a photosensitizer compatible with the Re complex can also increase the efficiency and durability of the catalyst [112]. Unlike the Re complex, which only reduces CO₂ to CO, the Ru photocatalyst can catalyze the conversion of CO₂ to CO and HCOOH. This reaction occurs through the capture of one CO₂ molecule by Ru(bpy⁻)₂(CO) to form Ru(bpy⁻)₂(CO)(CO₂) so that one H⁺ can be added to create the complex [Ru(bpy⁻)₂(CO)(COOH)]⁺. This complex will then produce HCOOH through the acceptance of two electrons. However, it doesn't stop at this point, the [Ru(bpy⁻)₂(CO)(COOH)]⁺ complex will then be protonated in the CO cycle so that water molecules are released and form CO molecules [112].

4.1.5 MXenes

MXenes are transition metal carbides, nitrides, or carbon nitrides that have great potential as photocatalysts. These metals consist of transition metals such as Scandium, Titanium, and Vanadium, elements from groups III A and IV A, and carbon or nitride elements. The general formula for MXenes is M_n+1X_n or M_n+1X_nT_x where T is a surface functional group such as oxygen, fluorine, and hydroxyl. Despite their potential as photocatalysts, MXenes cannot be used directly due to their non-semiconducting nature. Therefore, most of the MXenes are used as co-catalysts. MXene is commonly used in CO₂ reduction to improve photogeneration of charge carriers, photogenerated species separation, photo-corrosion inhibition, enhance CO₂ adsorption and activation, enhance light absorption, and photothermal effects. MXene can be applied in reactions of gas and liquid phase systems [114].

The photocatalytic activity and photostability of the semiconductor can be enhanced through the combination of the metallic conductivity of MXene with a suitable band structure. This combination allows for efficient electron migration from the semiconductor to the MXene. In addition, the regulation of the surface functional group of MXene can maximize CO₂ adsorption and activation. This is because the surface functional groups act as active sites for the photocatalytic reactions. MXenes can also improve light and photothermal harvesting for metal nanoparticles. MXenes as co-catalysts have a significant role in regulating product selectivity so diesel fuel production can be enhanced [114].

MXene as a co-catalyst is usually used with nitrides, metal oxides, metal salts, perovskite, and MOF as the main catalyst. One example of the use of an MXene/Nitride catalyst is the use of g-C₃N₄/Ti₃C₂. This heterojunction is reported to improve CO₂ adsorption, photogenerated charge carrier separation, and hybrid stability. There is also the use of MXene with metal oxides such as TiO₂ which is known to accelerate the efficiency of electron–hole separation. This combination will enhance the photocatalytic reduction reaction with 3.7 times higher methane production than commercial TiO₂. Other combinations such as perovskite CsPbBr₃ with MXene Ti₃C₂ were reported to produce higher CO and CH₄ than CsPbBr₃ NCs. This is due to MXene's ability to increase the photocatalytic activity [115] (Fig. 12).

Fig. 12

MXene synthesis

Generally, the basic synthesis of MXenes goes through the etching process. There are three etching approaches, namely fluorination etching, fluorine-free etching, and electrochemical etching. In the fluorination etching process, materials derived from hydrofluoric acid (HF) or salt solutions containing fluorine are used. Meanwhile, in the fluorine-free etchant method, Ti₃C₂Tx is fluorine-free in NaOH or KOH solutions. This method is considered to have a higher capacity for electrochemical properties. The electrochemical etching method is carried out through a redox reaction between the anode and cathode in an electrolytic cell through the application of voltage. Wet etching using HF is a method that is often used to synthesize MXenes. This is because HF-based etching has high selectivity and can break M-A metal bonds [115].

The MXenes as a photocatalytic co-catalyst synthesis is divided into three methods, namely mechanical mixing, self-assembly, and in situ oxidation. Mechanical mixing is a simple method for catalyst synthesis. This method includes advantages such as saving energy and low cost. Mechanical mixing is carried out through the process of grinding solid powder and mixing the substances in solution to deposit MXene on the surface of the photocatalyst. The self-assembly method is considered to have advantages over the mechanical mixing method, due to its ability to have closer contact and more uniform dispersion. Lastly, there is the direct in situ oxidation method of MXene. This method produces MXene in the form of MO/MXene or MO/MXene/C, where C is an amorphous carbon that can be used as a co-catalyst. Unfortunately, this method can only be used if the MO is a semiconductor [116] (Fig. 13).

Fig. 13

(Copyright Elsevier License Number 5326340653842)

MXenes synthesis schematic [116].

4.1.6 Metal Phosphide

Metal phosphides are widely used because of their abundance, low prices, and high electrical conductivity. Some examples of catalysts belonging to this group include Co_xP, Ni₂P, Fe₂P, Cu₃P, WP, and InP. The presence of this P group moderates the strength of the metal phosphide bond for the catalytic product conversion process, in which the active site facilitates product desorption. One example of a metal phosphide catalyst for CO₂ conversion is InP which has a band gap of 1.35 eV. Due to this large band gap, some modifications need to be made. One of them is to minimize the InP particle size to the quantum dot (QD) scale. The advantage of this QD modification is an increase in the redox potential so that the photocatalytic activity will also increase. This modification is also able to provide a more active site for the photocatalytic reaction [117].

In addition, metal phosphides are also commonly used as co-catalysts. This material can create sufficient and strong contact with the photocatalyst to effectively separate the electron–hole pairs. Metal phosphides are also considered to have an electronic structure like Pt so that they can be used as a substitute for precious metal co-catalysts. The bond between metal and phosphide can induce a small ligand effect so that the product will separate from the catalyst surface. Not only that, but this bond is also able to capture reaction intermediates so that the catalytic activity will increase [109].

Some examples of metal phosphide co-catalysts commonly used are FeP with polymeric carbon nitride as the main catalyst and WP with g-C₃N₄. The use of FeP co-catalyst with carbon nitride polymer was reported to be able to produce five times higher CO than the use of pure carbon nitride. The use of metal phosphide-based co-catalysts also has good stability. In addition, there was a narrowing in the band gap from 2.77 to 2.40 eV due to a change in the position of the VB resulting in increased light absorption under UV light. WP which was used as a co-catalyst with g-C₃N₄ also showed satisfactory results. CO production is reported to have increased considerably due to the use of this co-catalyst. Life cycle stability has also been reported to improve. This is due to the promotion of electron transfer due to the presence of P-N bonds at the WP and g-C₃N₄ interfaces [109].

4.1.7 Bismuth-Based p-Block Semiconductors

Most photocatalysts such as TiO₂ and ZnO use a d-block semiconductor. In this discussion, p-block elements will be used to construct a CO₂ reduction photocatalyst. This is because p-blocks have the potential to increase their photocatalytic activity in visible light. The p electrons in the hybridized state can decrease the CB state and increase the VB state so that the band gap becomes narrower. Not only that, another advantage that can also be obtained is the mobility of the photoexcited charge carriers and the charge separation will be increased [118].

Research conducted by Cui et al. [118] discusses the use of bismuth as a semiconductor by utilizing p electrons as a CO₂ reduction photocatalyst. According to them, there are three strategies to increase photoreduction activity, constituent adjustment, vacancy engineering, and heterostructure construction. The adjustment of the constituents is carried out through the adjustment of the halogen component and the ratio of Bi:O:X. Meanwhile, vacancy engineering is carried out through the engineering of oxygen, bismuth, and halogen vacancies. Finally, the heterostructure construction was carried out through the Bi_xO_yX_z construction of nanomaterial heterostructures and 2D semiconductors. Through this strategy, it is hoped that the band structure can be controlled so that the catalytic reduction conditions are met along with the introduction of high-level active sites on the photocatalyst surface. Improvements in energy conversion efficiency are also expected through the promotion of charge separation and transportation [118].

4.2 Non-metallic Photocatalyst

Photocatalysts in application to reduce CO₂ by using non-metallic materials exhibit rapid growth for photoconversion of CO₂. In addition, this material also shows tremendous potential as an alternative to photocatalysts owing to its abundance on earth, cost-effectiveness, high electrical conductivity, as well as environmental friendliness. Moreover, they also exhibit outstanding catalytic activity, durability, and exceptional selectivity in transforming CO₂ into solar fuel. In this section, we will explain and elaborate the non-metallic photocatalysts for CO₂ reduction [119].

4.2.1 Graphene

Graphene has been studied in-depth and has shown considerable interest in its application in photocatalytic CO₂ reduction. In addition, many experiments and studies have tried to design, develop, and apply graphene as a photocatalyst to reduce CO₂ [119]. The structure of graphene-based catalyst is unique. Aside from its large specific surface area, graphene also has many other desirable properties, including excellent electron transfer, transparency, high resistance to temperature changes, flexibility, and good CO₂ adsorption, which have made it ideal to be used as CO₂ photocatalysts [120]. The separation of charge carriers in photocatalytic CO₂ reduction can also be facilitated by graphene. In addition, graphene has a large contact area and interacts strongly with other catalysts, so the photocatalytic reaction and performance for reducing CO₂ can be promoted and improved by graphene [121].

Some common graphene includes pure graphene, graphene oxide (GO), and reduced graphene oxide. The three materials possess different functional groups, which gives them different properties as well [122]. A modified Hummers’ method is the common method to synthesize these materials using graphite as their raw material. The synthesis process with this technique begins with the chemical oxidation of graphite. The next step for preparing GO is exfoliation. GO typically has abundant oxygen-carrying functional groups, such as carboxyl, epoxy, hydroxyl, and carbonyl groups. The bandgap value of GO is about 2.2 eV while graphene pristine has a bandgap of 0 eV and does not have surface functional groups. The reduction of graphene oxide will produce rGO [123]. The reduction techniques are thermal treatment and chemical methods. The bandgap of rGO is variable, depending on the degree of reduction, from ~1.00 to 1.69 eV. The surface functional groups of rGO are the same as GO. The performance of graphene oxide can be improved by surface modification, doping, morphology control, and defect introduction. Tan et al. [124] investigated TiO₂ doped GO. The wet chemical impregnation technique is the technique used to combine GO with TiO₂ in this experiment. A modified Hummers’ technique was used to prepare graphite oxide powder. Followed by the exfoliation process through ultrasonication for 1.5 h and separation of the graphite layer. Within 8 h of reaction time, GO-TiO₂ successfully yielded 3.45 mol g⁻¹ h⁻¹ of CH₄. This performance represents a 14-fold performance increase over commercially available TiO₂ P25 [124].

The reaction scheme of the CO₂ reduction through GO-TiO₂ is shown in Fig. 14. The photocatalytic CO₂ reduction mechanism is suggested followed Langmuir–Hinshelwood (L–H) model. First, reactant molecules, in the form of carbon dioxide and water molecules, adsorb onto the photocatalyst surface. The next process is the surface reaction. At this stage where ∙CO₂⁻ radicals are generated from photogenerated electrons. Then, these ∙CO₂⁻ radicals will be transferred to the adsorbed CO₂. Meanwhile, on the other side, the hole reacts with the adsorbed H₂O molecules. The product of this reaction is ∙OH radicals and H⁺ ions. Then, it will form O₂. To produce intermediate radicals and hydrocarbon products, carbon radicals will react with H radicals formed by proton reduction on the photocatalyst surface [124].

Fig. 14

(Copyright Elsevier License Number 5321211075597)

The Langmuir–Hinshelwood mechanism [124].

4.2.2 g-C₃N₄

Graphite carbon nitride is a layered semiconductor material. g-C₃N₄ is a non-metallic material in the form of a polymer material and is composed of a polymer combination of tri-s-triazine/heptazine (C₆N₇) or triazine (C₃N₃) [120]. According to its wavelength of 460–430 nm, the g-C₃N₄ band gap value is narrow, ~2.7–2.9 eV. g-C₃N₄ also shows a CBat a negative position of about 1.1 V versus NHE. This property meets the thermodynamic requirements of photocatalytic CO₂ reduction reactions [125]. Therefore, under visible light and sunlight CO₂ will be activated and reduced successfully. The easy fabrication process, metal-free composition, durability, low cost, environmental friendliness, and good chemical and thermal stability are the superiority as a semiconductor material for the photocatalyst of g-C₃N₄. Furthermore, the conduction band of 1.14 eV and the valence band of 1.57 eV, make it a polymeric semiconductor suitable for visible light. However, g-C₃N₄ has drawbacks namely low exposed surface area, insufficient use of visible light, and fast photo-induced electron–hole pair recombination [126, 127].

Song et al. [128] developed the use of g-C₃N₄ as a photocatalyst in CO₂ reduction applications. They obtained g-CN as a catalyst. g-CN was incorporated with C and O simultaneously. The donor–acceptor structure was developed using hexa-methyl melamine (HMM) as a doping agent that provides C and O. Simple thermal condensation method of melamine was used to fabricate g-CN-coupled Samples C and O with HMM at elevated temperatures. The catalyst exhibits efficient photocatalytic CO₂ reduction activity and durability. Their experiment resulted in a production rate of CO at 34.97 µmol·g_cat⁻¹ in a 4-h-reaction-time. This result is 4.3-fold greater than that obtained by pristine g-CN [128] (Fig. 15).

Fig. 15

(Copyright Elsevier License Number 5321220110079)

Photocatalytic reduction of CO₂ mechanism through g-C3N4 [128].

4.2.3 Boron Nitride (BN)-Based Photocatalyst

Boron nitride has similarities to graphene, in particular the boron nitride with a hexagonal structure or h-BN. h-BN is a 2D material whose structure imitates a honeycomb, consisting of boron (B) and nitrogen (N) atoms with the same amounts and has excellent thermal conductivity and a large surface area. This material shows great potential in catalysis applications, especially in photocatalytic CO₂ reduction. h-BN has a broad bandgap of 5.2–5.5 eV. Similar to h-BN, porous BN also has a wide bandgap (~5.5 eV) [131].

Cao et al. [130] successfully fabricated the h-BN (O/BN) nanosheets. This nanosheet is modified with ultra-thin oxygen-containing a single bond between the B atom and O atom. When O is associated with the B atom, it has a greater electronegativity, so the introduction of the O atom into the h-BN aims to form a single bond between the B atom and the O atom. These nanosheets exhibit a great performance for CO₂ conversion to H₂ at a production rate of 3.3 µmol·g_cat⁻¹·h⁻¹ and CO at 12.5 µmol·g⁻¹·h⁻¹ [130].

4.2.4 Black Phosphorus (BP)

Black phosphorus, a 2D layered material, is an emerging remarkable photocatalytic semiconductor, especially for CO₂ reduction. BP nanosheets have advantages such as a tailorable band gap value of about 0.3–2.0 eV as the layer decreases from bulk to monolayer, suitable electronic structure, excellent optical properties, and high mobility of charge carriers, which is about 1000 cm²·V⁻¹·s⁻¹ [131].

For example, a previous study succeeded in making a CO₂ reduction photocatalyst from the fabrication of a monolayer of black phosphorus with a modified hydroxyl functional group, also known as M-BP-OH. In situ preparation of the photocatalyst can be achieved through the green exfoliation technique. In the synthesis using this technique, nitrogen is needed. From this research, the M-BP-OH exhibits an excellent photocatalytic activity performance. 112.6 mol·g⁻¹·h⁻¹ of carbon monoxide (CO) can be produced by this photocatalyst by reducing CO₂. Furthermore, the M-BP-OH exhibit excellent stability under humid condition (90% humidity) for 24 h, ambient conditions for a month, and cycling test using a Xe lamp at 300 W for 60 h [132].

Recently, a hybrid material constructed of BP QDs and g-C₃N₄ was prepared to convert CO₂ to CO. As a result of the research, the material can catalyze the CO₂ photochemical conversion reaction to produce CO. The CO product produced at high production levels was 6.54 mol·g⁻¹·h⁻¹ [125]. Compared to whole g-C₃N₄ which was only able to produce 2.65 μmol per gram of catalyst used for a 1-h basis, the yield achieved was much higher. Chen et al. (2021) fabricated a heterojunction with an S-scheme in which a stable ambient-based 2D/2D material (Pt/BP-Bi₂WO₆) for photocatalytic CO₂ reduction to syngas. The Pt/BP-OvMBWO heterojunction is electrostatically assembled. Pt/BP nanosheets are negatively charged and the OvMBWO nanosheets are positively charged. These two nanosheets were electrostatically assembled to obtain the heterojunction. As a result, this material can generate CO and H₂ as high as 20.5 µmol·g⁻¹·h⁻¹ and 16.8 µmol·g⁻¹·h⁻¹, respectively [131].

4.2.5 Silicon Carbide (SiC)

Silicon carbide is a kind of environmentally friendly semiconductor. Great chemical and thermal stability are other advantages of SiC. The band gap of SiC is wide, about 2.4–3.2 eV, and has a conduction band potential of −1.40 eV. Due to these properties, SiC has become a promising semiconductor material as a photocatalyst for CO₂ reduction [119]. Compared to the CO₂ reduction potential of some products, the potential of SiC is more negative. Owing to its properties, SiC is a suitable material for the reduction of CO₂ that requires a high potential. SiC has about 250 polytypes. However, in general, there are 2 kinds of SiC polytype, β-SiC, and α-SiC. Cubic 3C-SiC, which belongs to the β-SiC polytype is SiC. Meanwhile, the hexagonal SiC, such as 2H-, 4H-, 6H-SiC, and others belong to the α-SiC polytype [133]. Among these polytypes, the polytype that has the highest bandgap value of 2H-SiC is 3.3 eV while the lowest band gap of 3C-SiC is 2.4 eV [134].

Wang et al. [135] developed a β-SiC polytype of SiC. The developed SiC is a hollow-Sphere with an open mouth and is synthesized using environmentally friendly techniques. Glucose is used as a carbon source in this fabrication. In the application of photocatalytic CO₂ reduction with purified water, the prepared SiC showed a very active performance for converting CO₂ to methane due to the uniqueness of its electronic structure. Besides that, it is also supported by its hollow structure and its large BET surface area (28 m²/g). The performance of β-SiC Hollow-Sphere with Pt charge is greatly improved. Under sunlight, 2.0 wt% of Pt showed that the conversion of CO₂ to CH₄ reached 67.2 mol·g⁻¹. While the results with bare SiC only reached 28.1 mol/g, which means the results using β-SiC Hollow-Sphere with Pt loading were 2 times higher. Moreover, for the evolution of CH₄ from CO_2, the performance shown by this material is also more than that of many metal oxides. Figure 16 illustrates the charge transfer process for reducing CO₂ with water to produce methane under simulated sunlight in a Pt/SiC hollow sphere [135].

Fig. 16

(Copyright Elsevier License Number 5321211504925)

Photocatalytic CO₂ reduction mechanism in the Pt/SiC hollow spheres [135].

The development of polytype β-SiC, in the form of 3C-SiC, was also carried out by Li and Sun [136] on a gram scale. 3C-SiC is produced in the form of nanoparticles by using the ball milling method with a top-down approach. A cheap 3C-SiC crystalline powder is used for the fabrication of 3C-SiC nanoparticles. Ball-milling machine with high energy is used to grow nanoparticles by chemical vapor deposition. A 420 nm wavelength filter light from a Xe lamp with a power of 300 W is used in photocatalytic CO₂ reduction. 3C-SiC nanoparticles were dispersed into a glass reactor for the reduction of CO₂. 3C-SiC is capable of harvesting visible light. The band gap value is relatively small, about 2.36 eV, so 3C-SiC has an ideal band position straddling CO₂ reduction. In addition, 3C-SiC has high stability, thermal conductivity, and relatively low price. The achieved selectivity for reducing CO₂ to CH₄ is as high as 90% combined with efficient water oxidation. Therefore, 3C-SiC has great potential and deserves consideration in photocatalytic CO₂ reduction applications [136].

Liu et al. [133] fabricated a nanocage with a heterophase junction of 2H/3C-SiC. They prepared this nanocage by a simple magnesiothermic reduction technique mediated by low-temperature molten salt. SiO₂ was used as a template. Uniform heterophase junctions in the SiC nanocage will be formed by the 2H and 3C crystalline phases. This accelerates photogenerated electron transfer effectively and is very important in improving the performance to generate CO through the photocatalytic reaction of CO₂ reduction. CO produced by SiC nanocage was recorded to be 4.68 mol·g⁻¹·h⁻¹. This result is 3.25-fold more than commercially available SiC [133].

4.2.6 Covalent Organic Frameworks (COFs)

Known as COFs, covalent organic frameworks contain crystalline organic pores with orderly crystal structures connected with organic compounds by covalent bonds. COFs are polymeric materials. COF is built by elements that tend to be lightweight, such as carbon, hydrogen, oxygen, nitrogen, boron, and silica. As a new material class, COFs have some advantages. The advantages of COF include tailorable pore size, large specific area, tailored functionalities, and good thermal stability [119].

For example, Lu and coworkers designed a sequence of COFs in the form of crystalline 2D porphyrin-tetrathiafulvalente, TTCOF-M. The notation M indicates other elements used in the material, such as 2H, zinc (Zn), nickel (Ni), and copper (Cu). This material was synthesized as a photocatalyst for artificial photosynthesis including reducing CO₂. In this study, no sacrificial agents, precious metal co-catalysts, or supplementary photosensitizers were added to the photocatalyst system. Schiff base condensation technique between 5,10,15,20-tetrakis (4-aminophenyl)-porphinato] (TAPP-M) which is metalized is the technique used in this study for the fabrication of TTCOF- M. Meanwhile, the solvothermal method was used for the preparation of 2,3,6,7-tetra (4-formylphenyl)-tetrathiafulvalene [137].

The mechanism of the CO₂ reduction reaction process with H₂O oxidation is shown in Fig. 17b. Visible light irradiation encourages the CO₂ reduction process with H₂O oxidation. Before the photo-induced electron transfer process from the tetrathiafulvalene (TTF) part to the TAPP part, the photons will be absorbed first. After the electrons are excited, they travel to the active catalytic site in TAPP, which is Zn/Cu. The excited electrons that have been transferred are used to reduce CO₂. Meanwhile, TTF can generate O₂ through H₂O oxidation by the act of its photogenerated holes, in which electrons are obtained by the photocatalytic system from H₂O to maintain the charge balance. This research was able to reduce CO₂ with TTCOF-Zn/Cu COF using water as an electron donor. The highest performance is shown by TTCOF-Zn with CO production as high as 12.33 mmol and with high selectivity which is also combined with excellent durability [137].

Fig. 17

(Copyright John Wiley and sons License Number 5319200443008)

a TTCOF-M synthesis, b CO₂ reduction reaction process on TTCOF-M with H₂O oxidation [137].

Fu et al. [138] successfully fabricated a crystalline bipyridine-containing sp²c-COF as a photocatalyst to reduce CO₂. Rhenium complex [Re(CO)₅Cl] was used to modify the COF and form a homogeneous catalyst Re-Bpy-sp²c-COF or [Re(bpy)(CO)₃Cl]. This COF photocatalyst showed outstanding photocatalytic performance to convert CO₂ into CO with a production rate of CO 1040 mol·g⁻¹·h⁻¹ and selectivity of 81% (H₂) during 17.5 h of illumination. With dye sensitization, the photocatalytic performance was increased to reach 84% selectivity for more than 5 h of illumination and produce CO as much as 1400 mol·g⁻¹·h⁻¹ and 86% selectivity (CO/H₂) [138].

Recently, Cui et al. [139] developed a ketoenamine-linked 2D COF, also known as TpBb-COF. TpBb-COF was prepared under solvothermal conditions by an acid-catalyzed Schiff-base condensation (Fig. 18a). There are two reactants used in this condensation. The first reactant is 2,6-diaminobenzo[1,2-d:4,5-d’]bisthiazole or Bb-NH2, the second reactant is 1,3,5-triformylphloroglucinol or Tp-CHO. The reactant is in the binary solvent mesitylene and dioxane. TpBb-COF was used for photocatalytic reaction to convert CO₂ to produce CO. This material has a bandgap value of 1.72 eV. As a result, TpBb-COF showed excellent photocatalytic performance where as much as 52.8 mol·g⁻¹·h⁻¹ CO was able to be generated in pure CO₂. Meanwhile, in conditions of 30% CO₂ at 80 °C, the amount of CO that can be produced is 89.9 mol·g⁻¹·h⁻¹ [139].

Fig. 18

(Copyright Royal society of chemistry License Number 1226810–1)

a TpBb-COF Synthesis b Stacking mode of TpBb-COF from the top and side view [139].

4.3 Composites System

Low et al. [140] designed nanoparticles of TiO₂ on the MXene Ti₃C₂ for photocatalytic CO₂ reduction. The nanoparticles grew in situ on MXene Ti₃C₂ through a simple calcination technique. This composite has a unique structure like rice crust. Due to the unique multilayer structure of TiO₂ to Ti₃C₂, there is a large specific surface area which in turn provides a large active site for photocatalytic reduction. By using the optimized TiO₂/Ti₃C₂ composite, the performance to reduce CO₂ in photocatalytic reaction to produce methane was 3.7-fold more than commercial TiO₂ or about 0.22 μmol·h⁻¹. Ti3C2 has superior conductivity to facilitate the photogenerated electron transfer. Another consequence is that the photogenerated holes will also be suppressed by the effect of this property. TiO₂ has a conduction band of about −0.25 eV against a normal hydrogen electrode. The composite band gap value is −0.04 eV concerning normal hydrogen electrodes, a more negative value than the Fermi Ti₃C₂ level (Fig. 19a) [140].

Fig. 19

(Copyright Elsevier License Number 501735419)

a The migration of charge carrier on TiO₂/Ti₃C₂ composite scheme b Photocatalytic CO₂ reduction mechanism in TiO₂/Ti₃C₂ composites [140].

Figure 19b illustrates the photocatalytic process to reduce CO₂ in TiO₂/Ti₃C₂ composites. First, with an interesting morphology in which a surface with a large specific area will increase the ability of this composite to adsorb CO₂. Second, every component involved contributes to the outstanding performance. The black color properties of Ti₃C₂ increase the light absorption of the composite and create a photothermal effect. Third, on conductive Ti₃C₂ 2D, a heterogeneous interface will be formed from in situ growth, where this interface can change the properties of the original material. The tight and strong contact facilitates the migration of charge carriers so that electron holes separate efficiently. In addition, photogenerated electrons also accumulate in Ti₃C₂ so it is very supportive of the multi-electron CO₂ reduction reaction [140].

Wang et al. [126] developed a photocatalyst to reduce CO₂ in the form of a nanocomposite z scheme between g-C₃N₄ and α-Fe₂O₃. The two materials are linked by Al-O. The first step to make this photocatalyst is to make a g-C₃N₄/α-Fe₂O₃ nanocomposite with an optimal ratio. The hydrolysis-solvothermal method with phase separation was used in the synthesis of nanocrystalline α-Fe₂O₃. To prepare g-C₃N₄, urea was heated in an alumina combustion vessel. The mass ratio of the two components was adjusted to obtain the optimal nanocomposite. Then, the photocatalyst composite of g-C₃N₄ and α-Fe₂O₃ which is linked by the Al-O bridge was fabricated by a two-step wet chemical technique. As the source of the Al-O bridge in the nanocomposite, an aqueous solution of AlCl₃ was used. Compared to using pristine g-C₃N₄ and pristine Fe₂O₃, nanocomposite photocatalysts exhibit higher photocatalytic activity for enhanced CO₂ conversion. After irradiation for 1 h, α-Fe₂O₃ showed low photocatalytic activity in reducing CO₂ to several products, such as CO, CH₄, and O₂ with production rates of ~ 5.7 mol·g⁻¹·h⁻¹, ~ 0.3 mol·g⁻¹·h⁻¹, and ~ 6.2 mol·g⁻¹·h⁻¹ respectively. However, the addition of a large amount of g-C₃N₄, especially 15CN-F, give a positive contribution to the photocatalytic activity where the production rate for each product increased to ~ 15.8 mol CO g⁻¹·h⁻¹, ~ 3.1 mol CH4 g⁻¹·h⁻¹, and ~ 18.5 mol O2 g⁻¹·h⁻¹. This photocatalytic activity performance can be further enhanced by building a link or bridge between g-C₃N₄ and α-Fe₂O₃ (15CN-6Al-F) on 15CN-F nanocomposite in the form of Al-O. As a result, the amount of CO produced is 4 times higher than α-Fe₂O₃, or about 24 mol·g⁻¹·h⁻¹. The charge transfers and schematic of the photo-induced charge separation process for a g-C₃N₄/α-Fe₂O₃ nanocomposite linked by an Al-O bridge are illustrated in Fig. 20. Initially, the photoinduced electrons of α-Fe₂O₃ will join the g-C3N4 photoinduction holes in the g-C₃N₄/α-Fe₂O₃ nanocomposite when these two materials are excited simultaneously under a wavelength of 460 nm. In this case, a redox reaction will be induced because the holes in α-Fe₂O₃ and electrons in g-C₃N₄ are spatially separated and have thermodynamically sufficient energy. This reaction will result in a promoted charge separation which will increase the photocatalytic activity. Al-O bridges in nanocomposites will promote charge transfer from the α-Fe₂O₃ side to the g-C₃N₄ side and lead to increased photocatalytic performance for CO₂ reduction. Photo-induction of electrons in g-C₃N₄ stimulates a reduction reaction with the H₂O component to produce hydrogen atoms. The resulting hydrogen atoms attack the CO₂ molecules to produce products in the form of CO and CH₄. Meanwhile, another product, namely the radical hydroxyl group, could evolve O₂ through the photoinduced hole reaction of α-Fe₂O₃ with H₂O [126].

Fig. 20

(Copyright Elsevier License Number 501735385)

Mechanism of photoinduction charge transfer and separation in g-C₃N₄/α-Fe₂O₃ nanocomposites linked by Al-O bridges [126].

Photocatalyst made from g-C₃N₄ was also recently fabricated by Xu and coworkers. In their research, they developed a photocatalyst nanocomposite with a ternary system between g-C₃N₄, Ag₂CrO₄, and graphene oxide having a Z-scheme mechanism. This nanocomposite photocatalyst system with g-C₃N₄ was developed because its ability to absorb visible light is low and its charge recombination is so fast. In this photocatalyst system, silver chromate (Ag₂CrO₄) nanoparticles are used which act as photosensitizers to drive charge transfer and harvest light. Silver chromate (Ag₂CrO₄) was used as a photosensitizer due to its narrow bandgap (-1.8 V). This is favorable for the g-C₃N₄ light absorption so that the amount of absorbed light by g-C₃N₄ can be increased. In addition, Ag₂CrO₄ may also pair with g-C₃N₄ to form binary heterojunctions to promote photogeneration charge separation. Graphene oxide (GO) was used as a co-catalyst. Apart from being a co-catalyst, GO is also an electron collector and provides a reactive site for CO₂ conversion. A composite photocatalyst with a ternary system was recently prepared to reduce CO₂ in photocatalytic reactions under simulated light irradiation to produce methane and methanol. The precipitation self-assembly technique was used for the fabrication of Ag₂CrO₄/g-C₃N₄/GO composites (Fig. 21) [141].

Fig. 21

(Copyright Elsevier License Number 501735474)

Schematic of Ag₂CrO₄/g-C₃N₄/GO composite fabrication [141].

Under full spectrum sunlight simulation, the photocatalytic activity of the ternary composite Ag₂CrO₄/g-C₃N₄/GO in reducing CO₂ was very good, as high as 1.03 mol·g⁻¹. This result is 2.3-fold more than using the whole g-C₃N₄. The direct Z-scheme mechanism is a charge transfer process that occurs between Ag₂CrO₄ and g-C₃N₄ (Fig. 22). Apart from promoting charge separation, this charge transfer process enhances the reduction–oxidation capability of the photocatalytic system as well [141].

Fig. 22

(Copyright Elsevier License Number 501735474)

Photocatalytic process in Ag₂CrO₄/g-C₃N₄/GO ternary composite through Z-scheme [141].

Crake et al. [142] developed a composite photocatalyst of inorganic and organic materials for CO₂ reduction applications. In this study, they used titanium oxide and carbon nitride. Carbon nitride is used in the form of nanosheets (CNNS). The two materials, namely TiO₂ and CNNS were prepared by hydrothermal in situ growth technique (Fig. 5a). This photocatalyst system does not require an expensive co-catalyst and CO₂ reduction takes place in a single process. First, the preformed carbon nitride was exfoliated by sonication in water to form carbon nitride (CNNS) nanosheets. This CNNS was used to obtain TiO₂ and CNNS composites. In the PTFE autoclave, titanium butoxide, hydrofluoric acid, and CNNS were added and mixed. The process was continued by heating at 180 °C for 24 h. The addition of hydrofluoric acid aims to control the formation of crystal facets. This photocatalyst exhibits an excellent photocatalytic CO₂ reduction performance compared to its constituent materials, approximately an increase of more than tenfold. In addition, this composite also increases the capacity to adsorb CO₂ [142] (Fig. 23).

Fig. 23

(Copyright Elsevier License Number 5321210450514)

TiO₂/Ti₃C₂ composite a Schematic of the fabrication process of TiO₂/CNNS composites b Photocatalytic CO₂ reduction reaction pathway [142].

4.3.1 TiO₂-MOF Composite

The combination of TiO₂ metal with MOF covers the shortcomings of each material as a CO₂ reduction photocatalyst. These metals have their respective roles; TiO₂ acts as a primary photocatalyst. Meanwhile, MOF acts as a photosensitizer for efficient light harvesting and electron transfer. Combining the two improves light absorption, electron–hole pair separation, and photocatalytic efficiency. The results obtained can be 5 times higher than pure metal TiO₂ even though CO₂ adsorption is reduced [143].

TiO₂-MOF composites with TiO₂ particles graphed onto (Cu₃(BTC)₂) microcrystals (TiO₂ in HKUST-1) were found to have a bandgap of 1.6–3.1 eV under UV light. The use of this composite reduces the adsorption capacity of CO₂, but the production of CH₄ is five times higher than the use of pure TiO₂ due to electron–hole separation [144].

Crake et al. [145] performed the synthesis of TiO₂ through the solvothermal method. Then the in situ method was used to produce TiO₂-MOF composites under UV–Vis irradiation. This composite prevents TiO₂ aggregation by maintaining a high surface area. Furthermore, the porosity of the MOF results in a sizeable CO₂ adsorption capacity so that the amount of adsorbed CO₂ on the surface increases. Wang et al. [143] explain four types to produce TiO₂-MOF composites. The first type uses organic ligands and metal templates of different morphology to form MOFs. The second type uses a precursor to produce TiO₂ into a prepared MOF. The third type is the in situ method, as done by Crake et al. [145], and the last one uses a partial organic linker from MOF.

4.3.2 Graphene-MOF Composite

Another MOF combination, graphene, makes up for the lack of electrical conductivity in MOF and improves the photothermal stability. This composite can also reduce aggregation at high-temperature calcination and increase the catalytic activity of the active sites. The MOF in the composite acts as a substrate for the diffusion of the substrate because of its porous structure so that there is more room for additional components. In addition, MOF also has high crystallinity and stability during the photocatalytic process. Graphene is a conductor for light harvesting and electron transfer [110].

Briefly, there are two methods for making MOF-Graphene composites. The first way is physical mixing with pH adjustment to increase electrostatic adsorption. The second method is in situ growth to increase the interaction between graphene and MOF. The performance of these two methods can be improved in several ways, namely MOF aminase, polydopamine, heteroatom doping, use of graphene aerogel, and constructing heterogeneous structures [146].

4.3.3 Titanium-Based Perovskites

Modification of TiO₂ catalyst with perovskite as host material was used to adjust optical properties and induce visible light absorption. This composite is also intended to form a catalyst that has good photo-corrosion resistance and thermal stability. In addition, this composite is applied as a candidate for producing hydrogen and hydrocarbons as well as water separation processes through photocatalytic reduction of CO₂ [108].

Strontium titanate oxide (SrTiO₃) and calcium titanate oxide (CaTiO₃) are the most widely used perovskite oxides. SrTiO₃ perovskite composites with TiO₂ are known to have the ability to selectively separate CO₂ reduction reactions and enhance charge transfer synergism. In addition, this perovskite can escalate the selectivity of CH₄ and CO production. The addition of Pt and Pd nanoparticles to the composite material increased the photoconversion efficiency of CO₂ to CH₄. As a result, the production of CH₄ and CO increased by 4.7 and 1.6 times, respectively, compared to ordinary composites. This is because Pt contributes to increase the efficiency of charge separation [108] (Fig. 24).

Fig. 24

(Copyright Elsevier License Number 5326340974814)

Photocatalytic Mechanism over SrTiO₃/TiO₂ Composite. Reproduced with permission [108].

4.3.4 Graphene-Based Perovskites

Graphene is one of the most widely used photocatalysts for CO₂ reduction. However, this material has a cypher bandgap which makes it incapable to perform charge separation. However, graphene offers high charge mobility and electron–hole pair separation through its interface. The use of composites with graphene enables increasing its light absorption, charge mobility, and catalytic activity in visible light through their synergistic effect [108].

The use of perovskite Bi₂MoO₆ composite with graphene for photocatalytic CO₂ reduction was carried out to produce CH₃OH and C₂H₅OH. Bi₂MoO₆ electrons will migrate from VB to CB, then the Mo atoms in the perovskite will trap photogenerated electrons to attract CO₂. These electrons will then move towards graphene which has an electron-rich interface. Finally, CO₂ absorption occurs due to the large surface area of graphene [147].

4.3.5 Graphitic Carbon Nitride-Based Perovskites

KNbO₃ is a perovskite used as a composite with g-C₃N₄ and Pt. g-C₃N₄ functions as a sensitizer while Pt acts as a co-catalyst. The electrons will be migrated to the perovskite which is then switched to the Pt co-catalyst to reduce CO₂ into CH₄. The use of this composite system produces four times higher methane product than the use of pure g-C₃N₄, under visible light radiation. In addition, the use of perovskite as a composite material is proficient to increase the catalytic activity along with the increase in the concentration of perovskite. In another study, the use of perovskite NaNbO₃ composites with g-C₃N₄ was able to increase methane production eightfold compared to the use of pure g-C₃N₄-Pt or NaNbO₃. This is because there is a faster separation and charge transfer so that the photocatalytic activity also increases. In addition, g-C₃N₄ CB (−1.13 eV) was located at more negative potential than that of NaNbO₃ (−0.77 eV) [108] (Fig. 25).

Fig. 25

(Copyright Elsevier License Number 5326340974814)

Photocatalytic Mechanism over g-C₃N₄/ KNbO₃ Composite. Reproduced with permission [108].

4.3.6 MXenes-Based Perovskites

MXenes are transition metal carbides/nitrides and carbonitrides that have the general formula M_n+1X_n, where M is a transition metal such as scandium, titanium, zirconium, hafnium, vanadium, niobium, tantalum, and molybdenum while X is carbon or nitrogen. Perovskite Bi₂WO₆ composite with MXene Ti₃C₂ yielded 4.6 times higher CH₄ and CH₃OH than the use of pure Bi₂WO₆ perovskite. This is because CB perovskite is more negative than Ti₃C₂ so there is a sudden electron migration as a result of photo induction to the heterojunction interface. In another study, CsPbBr₃ NC/MXene perovskite nanocomposites were able to selectively produce CO and CH₄ [148].

5 Product Characterization, Analysis, and Selectivity

5.1 Product Characterization and Analysis

CO₂ photoreduction involves multi-electron reactions that can result in a wide range of products in the gas phase, including CO, CH₄, and higher hydrocarbons, as well as numerous, oxygenates in the liquid phase, including alcohols, aldehydes, and carboxylic acids. Table 3 depicts a few reaction examples that result in gas and liquid products. According to a recent analysis by Izumi, certain reduction products with reasonably high yields that are not thoroughly verified may not be entirely made from CO₂. While our study is focused on producing effective photocatalyst systems, there are a few critical challenges that need to be addressed in terms of CO₂ reduction product analysis: (1) Is it possible to reliably detect the most likely reduction products? After accurate quantification of the products and elimination of the contribution from carbon contamination, a reliable conclusion on photocatalytic activity can be reached. (2) Will organic additions have an impact on the product analysis? Organic additives are frequently utilized in homogenous molecular systems as sacrificial reagents, solvents, photocatalysts, or photosensitizers. It is necessary to evaluate their impact on the analysis of CO₂ reduction products. (3) Is it true that the measured products are the result of CO₂ reduction? Verification techniques include 13C nuclear magnetic resonance (NMR) and gas chromatography-mass spectrometry (GC–MS).

5.1.1 Analysis of Gas Phase Products

The principal gaseous products of CO₂ photoreduction are CH₄ and CO, while H₂ and/or O₂ can also be formed as byproducts of water splitting [149]. Varghese et al. also found higher alkanes as photoreduction products of CO₂ in their study [150]. Aside from the products, CO₂ must be quantified as the primary species in the gas phase due to limited conversion to estimate the reaction progress and/or the extent of CO₂ dissolution in the liquid phase [151]. The use of infrared spectroscopy (IR) or diffuse reflectance infrared Fourier transform spectroscopy (DRIFT) to check CO₂ consumption and generation has been done on occasion [152]. The most widely used method for quantifying these gas species is gas chromatography (GC) [152]. GC–MS was also employed for routine CH₄ and CO analysis in addition to GC [153] or to confirm the source of carbon [154].

5.1.2 Analysis of Liquid Phase Products

Although GC with TCD/FID is still the most commonly used technology for analyzing various forms of liquid oxygenates (mainly alcohols) [155], Other approaches have been used, such as MS, high-performance liquid chromatography (HPLC) [156], ion-exchange chromatography (IEC) [157], ultraviolet–visible (UV–Vis) spectroscopy (colorimetric assay) after reacting with chromotropic acid [158] or Nash reagent [159] and NMR [160]. Table5 summarizes the substances that can be evaluated and their detection limits utilizing various techniques, as well as the accompanying constraints.

Table 5 Comparison of techniques used for liquid phase analysis in CO₂ photoreduction

5.2 Product Selectivity

Photocatalytic CO₂ reduction can directly convert CO₂ and H₂O into hydrocarbon solar fuels utilizing endless solar energy as the only source of energy [161]. CO₂, as the highest oxidation state of carbon, can be reduced into a variety of products by obtaining different numbers of electrons and protons, such as CO, HCOOH, HCHO, CH₃OH, and CH₄ [162]. Besides these C1 products, some C2 products (C₂H₄, C₂H₆, CH₃CH₂OH) derived from the C–C coupling reaction also can be detected in the photocatalytic CO₂ reduction reactions [163].

In thermodynamics and kinetics, Fig. 26 summarizes some key influencing factors. To begin, light must be absorbed by a semiconductor photocatalyst to produce effective photogenerated electrons and holes. Photon energy and light intensity are the most common light-excitation attributes. The former determines whether a semiconductor can absorb excited photons and thus influence the product selectivity of thermodynamic reactions. The latter impacts the number of photogenerated electrons and holes generated by excitation, and subsequently affects the reaction rate and product selectivity of multiple electron processes kinetically.

Fig. 26

(Copyright Elsevier License Number 5327490662508)

Important thermodynamic and kinetic influence factors in product selectivity of photocatalytic CO₂ reduction reactions [163].

5.2.1 Product Pathways and Selectivity

Figure 27 depicts the various routes of complex photocatalytic CO₂ reduction processes, which include electron and proton transfer, hydrogenation and deoxygenation, C–O bond breaking, and C–H bond formation [165]. CO₂ adsorption on the surface of the photocatalyst initiates photocatalytic CO₂ reduction processes. CO₂ adsorption mechanisms in general include carbon and oxygen coordination with photocatalyst surface atoms [166]. The adsorption/desorption of reactants/intermediates, as well as the sequencing of hydrogenation and deoxygenation processes, are all affected by the C, H, and O affinities of photocatalysts, resulting in diverse product paths and selectivity [164].

Fig. 27

(Copyright Elsevier License Number 5327490662508)

Possible pathways of photocatalytic CO₂ reduction to CH₄ on the surface of photocatalysts [164].

5.2.2 Main Strategies for Controlling Product Selectivity

Complex multi-step mechanisms are involved in photocatalytic CO₂ reduction reactions. In terms of reaction activity and product selectivity, each procedure is critical. The photocatalytic CO₂ reduction process can be broken down into the following phases in general: (1) Photocatalysts are activated by the right amount of sunlight and produce a large number of electrons and holes. (2) Charge carriers generated by photocatalysts separate and transfer from the bulk to the surface. (3) On the surface of photocatalysts, reactants (CO₂ and H₂O) are simultaneously captured and activated. (4) Photogenerated holes are consumed by H₂O or additional sacrificial reagents to form oxidation products, whereas effective photogenerated electrons reach the surface active sites and catalytic reduction processes occur. (5) The resulting products desorb from the photocatalyst surface, and the re-exposed surface active sites engage in the subsequent catalytic events.

By adjusting the key phases of the photocatalytic CO₂ reduction process, the product selectivity of photocatalytic CO₂ reduction reactions can be controlled. The primary engineering strategies for enhancing photocatalytic CO₂ reduction product activity and selectivity are given in Fig. 28.

Fig. 28 

(Copyright Elsevier License Number 5327490662508)

Main factors influencing the product selectivity of photocatalytic CO₂ reduction reactions.

5.2.3 Light-Excitation Attributes

In general, the wavelength requirement of incident light that can excite semiconductors is determined by the band gaps of classic semiconductor photocatalysts. By altering the light-excitation properties, it is difficult to manage the product selectivity of CO₂ reduction processes. Plasmonic photocatalysts have different modes of light absorption and excitation than standard semiconductor-based photocatalysts. The collective free electron oscillations induced by light irradiation, known as localized surface plasmon resonances (LSPRs), can produce strong light absorption in plasmonic nanoparticles (NPs) [167].

Adsorbed reactants on the surface of plasmonic NPs can be activated by the energetic (hot) electrons obtained. Excess hot electrons stored on the surface can participate in CO₂ reduction processes with the help of an effective hole scavenger, and oxidation products result from the oxidation of the hole scavenger. The distribution of photoexcited hot electrons is highly influenced by the energy and density of photons, resulting in distinct reaction paths and rates. It should be possible to tune light-excitation properties to alter the product selectivity of plasmonic photocatalysts.

5.2.4 Band Structure of Photocatalysts

The redox abilities of photogenerated electrons and holes are highly dependent on the locations of photocatalysts’ CB and VB during photocatalytic processes. CO₂ conversion to various reduction products necessitates certain reduction potentials of photogenerated electrons, as shown in Fig. 29. Regulating the band structure of photocatalysts to regulate the CB positions allows efficient management of product selectivity with respect to thermodynamics by obtaining designated reduction potentials of electrons.

Fig. 29

(Copyright Elsevier License Number 5327490662508)

Reduction potentials of photocatalytic CO₂ reduction to different products (versus SHE, pH = 7) [168].

5.2.5 Separation of Photogenerated Electrons and Holes

CO₂ reductions by photocatalysis are multi-electron processes. The density of photogenerated electrons on photocatalyst surfaces can alter the rate of surface catalytic reactions, which in turn affects the overall product selectivity of the reaction. The separation efficiency of photogenerated electrons and holes in photocatalysts determines the density of surface photogenerated electrons [169]. As a result, increasing the effectiveness of charge separation of photocatalysts should be a key technique for controlling product selectivity [170].

5.2.6 Adsorption and Activation of Reactants

In a catalytic reaction, the adsorption and activation of reactant molecules is a crucial step. The only two reactant molecules in photocatalytic CO₂ reduction processes are H₂O and CO₂. Their adsorption and activation capabilities on the photocatalyst surface have a significant impact on activity and product selectivity. CO₂ is a nonpolar linear molecule that has a bond energy of 799 kJ·mol⁻¹. CO₂ is difficult to activate and diminish due to these properties [171]. CO₂ adsorption on photocatalyst surfaces can change their linear molecular conformation, allowing the CO₂ molecule to be activated for photocatalytic reduction processes [172]. CO₂ adsorption and activation are usually synchronized. CO₂ adsorption on the surface of photocatalysts is indicated by typical carbon coordination, oxygen coordination, and mixed coordination modes.

A partially charged species CO₂ radical dot will arise as a result of the interaction between CO₂ and the surface atoms of photocatalysts, which is the typical activation intermediate. The process of further reduction by acquiring electrons can occur smoothly after activation. The surface atomic states and characteristics of photocatalysts influence the coordination modes of CO₂ radical dots. Better CO₂ adsorption and activation increase CO₂ usage and lower the reaction barrier, affecting the product selectivity of photocatalytic CO₂ reduction processes [173].

H₂O is a key reactant in the photocatalytic CO₂ reduction process [174]. To complete the oxidation half-reaction, H₂O consumes photogenerated holes. H₂O also serves as a proton source for the CO₂ hydrogenation process. The adsorption and dissociation of H₂O to protons on the surface of photocatalysts can boost the photocatalytic CO₂ reduction processes by further reducing the hydrocarbon products. Pt is a good candidate for dissociating H₂O to H adatoms, and it is beneficial in the treatment of HER. Hydrocarbon compounds with a high reducing state can be obtained by further hydrogenation of CO₂ with H adatoms obtained through H₂O dissociation.

5.2.7 Control Reaction Active Sites

The product selectivity of photocatalytic CO₂ reduction processes is influenced by the quantity and species of surface reactive sites. Different active sites are appropriate for various surface catalytic processes, yielding various products. High-index facets can bring more surface low-coordinated atoms than low-index facets, which can serve as good active sites for CO₂ activation and catalytic processes.

The product selectivity of CO₂ photocatalytic reduction was heavily impacted by the size of Pt NPs. The smaller Pt NPs resulted in better H₂O reduction selectivity for H₂ production. The bigger Pt NP-modified HTSO can achieve a better CH₄ selectivity. Surface terrace sites were more common in bigger Pt NPs, while low-coordinated locations (corner and edge) were more common in smaller Pt NPs (Fig. 30a). The DFT simulations also revealed that terrace sites can operate as active sites for CH₄ production, while low-coordinated areas are better for converting H₂O to H₂.

Fig. 30

(Copyright Elsevier License Number 5327490662508)

a Relationship between selectivity of CH4, the ratio of terrace/surface, and size of Pt particles. b Process of partial CO-modified 1.8PHTSO by CO stepwise adsorption/desorption. c CO-TPD results of 1.8PHTSO after CO pulse adsorption and partial CO-modified 1.8PHTSO by He flow desorption at 280 °C. d Photocatalytic CO₂ reduction performance of 1.8PHTSO and CO-1.8PHTSO [164].

The partial CO-modified 1.8PHTSO (1.8 nm Pt modified HTSO) may cover the corner and edge sites of Pt particles, as shown in Fig. 30b, c. The photocatalytic performance of 1.8PHTSO and CO-1.8PHTSO is compared in Fig. 30d, demonstrating that the terrace sites were favored for CH₄ production. The size of Pt modified the geometric feature's surface site percentage, which varied the product selectivity of photocatalytic CO₂ reduction processes [164].

5.2.8 Intermediates/Side Products of Adsorption and Desorption

Good reactant adsorption and simple product desorption are required to yield targeted products with high selectivity. Many different intermediates or side products can be generated during the photocatalytic CO₂ reduction process. Intermediate and side products differ in their strong interactions with photocatalysts. The species are difficult to desorb from the surface of the photocatalyst if the contact was too strong. It is more likely that a further hydrogenation reduction reaction will be carried out [9, 164]. Intermediaries is the name given to these species. The weak interaction, on the other hand, will make desorption and release from photocatalyst surfaces easier, and these species act as key side products. In this regard, the adsorption and desorption of intermediates/side products can have a significant impact on photocatalytic CO₂ reduction reaction product selectivity [9, 164, 173].

6 Challenges and Opportunities

Energy scarcity and environmental issues are becoming increasingly serious global issues and major challenges for humanity in the twenty-first century. Solar-driven CO₂ reduction to value-added chemical fuels is a viable option for influencing the global carbon balance, with tremendous potential rewards. The current goal is to create photoactive materials that can chemically tie these light-driven redox processes together, achieving conversion efficiencies and selectivity that are superior to natural photosynthesis. Thermodynamically upward, multi-electron, multihole, and multi-proton reactions occur on a multicomponent photocatalyst in the photoconversion of CO₂. In the fields of catalysis, energy science, semiconductor physics and engineering, and green chemistry, long-standing difficulties are posed. Several key considerations must be balanced to develop effective CO₂ fixation and conversion, including:

• (i) A thorough understanding of the processes that take place on the surface of photocatalysts during CO₂ reduction, such as CO₂ and intermediate product adsorption and desorption, as well as the role of adsorbed water. In this regard, the research would combine experimental and computational methods to assess catalyst activity for CO₂ photoreduction, advancing fundamental knowledge in the fields of material science, surface chemistry, photocatalysis, and nanotechnology. STM would be used to determine CO₂ adsorption locations and explore molecular orbitals and transfer of charge. The local density of states, lowest-energy molecule structure, and vibrational modes would theoretically be clarified using density functional theory (DFT) computations. STM and DFT working together would give a natural framework for future research targeted at gaining an atomic-scale knowledge of CO₂ photoreduction.

• (ii) A significant increase in the lifespan of the charge-separated state could result in a significant increase in CO₂ photoreduction efficiency. This electron–hole recombination is two to three orders of magnitude faster than other electron transfer processes in terms of time scale. As a result, any approach that prevents electron–hole recombination will considerably improve CO₂ photoreduction efficiency and rates. To decrease electron–hole pair recombination, the photocatalyst's electrical conductivity and diffusion length should be as high as possible. Charge-carrier transfer to surface reaction sites for reductive chemistry may be aided by ultrathin nanostructures. The electron–hole recombination can be decreased by spatial separation of photoexcited electrons and holes, which can be done via co-catalysts, Z-schemes, or heterostructures linking two semiconductors with properly aligned band structures.

• (iii) Light absorption by the photocatalyst affects the kinetics of CO₂ photoreduction. Many optical approaches, such as structuring for multiple light scattering to increase the effective optical path length and up-conversion to convert nonabsorbed infrared light to absorbed visible light, could be useful for harvesting light to improve efficiency.

• (iv) To maximize the adsorption, transport, and desorption of reactants, intermediates, and products, a large surface area and porosity are necessary.

• (v) Because defects like oxygen vacancies regulate the majority of the chemistry at many metal oxide surfaces, oxygen vacancies are thought to play a critical role in electron trapping and CO₂ activation. These defects can be discovered by in situ electron paramagnetic resonance spectroscopy, ultraviolet photoemission spectroscopy, and metastable impact electron spectroscopy, among other techniques.

• (vi) A high-efficiency process must be based on photoactive materials that are abundant on the planet, non-toxic, light-stable, scalable, and affordable in cost. Furthermore, with prolonged irradiation, the efficiency of photocatalytic CO₂ reduction may be deactivated.

7 Conclusion

It may be argued that environmental issues are no longer localized issues, but have evolved into global issues concerning climate change. We are still far from having a superb as well as a cost-effective photocatalyst for photocatalytic CO₂ reduction with H₂O to overcome these challenges. As previously stated, photocatalytic processes are quite intricate in terms of both physical mechanism and product distribution, which limits their usefulness. The current state of research in this field is perplexing, and evaluating the efficiency of various photocatalysts is challenging due to the wide range of influencing factors and reaction circumstances. Many of the studies neglected some details like mass balance (moles of CO₂ transformed into the specified product), product distribution, and the amount of reducing agent, as well as time requirements, solution pH, temperature, CO₂ pressure, light power, and activity decay with time. Comprehensive research into this method is required before it can be put into practice on a large scale. A single, uniform and widely agreed standard format should be created.

The major challenges responsible for the current significantly low rate of average productivity in photocatalytic reduction of CO₂ with H₂O are low photocatalytic efficiency, low response to sunlight, inefficient electron transport between reduction and oxidation catalysts, and a high recombination rate of photogenerated species. The short lifetimes of one electron-reduced species and the photoexcited state in the presence of O₂ produced by H₂O oxidation are considered also more drawbacks. Even though UV light has more energy than visible light, visible light harvesting photocatalysts are the most desirable for this process due to the plentiful visible light from natural sunlight. Plasmonic photocatalysts, which are still in their inception, have already shown promise in overcoming the first two flaws listed above. Because the boundaries of these extremely effective noble metal photocatalysts are quickly growing, it is reasonable to expect plasmonic photocatalysts to play a substantial role in future environmental solutions.

Fabrication of ideal structured visible light-responsive photocatalysts with a wide bandgap, high rate of photogenerated electron–hole transport, 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 should be the focus of future research.