Recent Progress in the Photocatalytic Reduction of Carbon Dioxide.

Elimination or reduction of CO2 in the atmosphere is a serious problem faced by humankind, and it has become imperative for chemists to find ways of transforming undesirable CO2 to useful chemicals. One of the best means is the use of solar energy for the photochemical reduction of CO2. In spite of considerable efforts, discovery of stable photocatalysts which work in the absence of scavengers has remained a challenge although encouraging results have been obtained in the photocatalytic reduction of CO2 in both gas and liquid phases. Semiconductor-based catalysts, multicomponent semiconductors, metal-organic frameworks (MOFs), and dyes as well as composites involving novel composite materials containing C3N4 and MoS2 have been employed for the photoreduction process. Semiconductor heterostructures, especially those containing bimetallic alloys as well as chemical modification of oxides and other materials with aliovalent anion substitution (N3- and F- in place of O2-), remain worthwhile efforts. In this article, we provide a brief perspective of the present status of photocatalytic reduction of CO2 in both liquid and gas phases.

Introduction

Nearly 80% of the energy supply of the world is based on fossil fuels, and humankind is expected to face a serious crisis because of the fast depletion of fossil fuels and increasing levels of CO2 in the atmosphere. The present global energy consumption rate of 16.3 TW (2012) will be reaching nearly 40 TW by 2050 and nearly 60 TW by 2100.[1] The Intergovernmental Panel on Climate Change (IPCC) predicts that by the year 2100 the CO2 levels may rise up to 590 ppm, and the average global temperature may rise by 1.9 °C.[2] As per the fifth assessment report (2014) of the IPCC, the levels of green house gases (GHGs) are increasing drastically. The steady increase of +1.3% per year (1970–2000) has taken a steep rise of +2.2% per year (2000 onward).[3] Increasing atmospheric levels of CO2 can adversely affect the world, causing an increase in average sea levels and average global temperature.[3] Either developing alternative fuels or converting CO2 back to the fuel forms would balance the CO2 levels in the atmosphere. Sun’s energy reaching the earth’s surface (1.3 × 105 TW) is 10 000 times higher than the present rate of consumption or demand. Researchers are, therefore, attempting to develop photocatalysts for splitting water or reducing CO2 to fuels (H2, CO, HCOOH, HCHO, CH3OH, CH4) using solar energy.

Natural photosynthesis involves both the oxidation of water and the reduction of CO2 wherein three processes, light harvesting, generation of carriers, and separation and catalytic reactions, occur. Artificial photosynthesis attempts are being conducted to accomplish these by a two- or one-step process, the former being akin to the Z-scheme in natural photosynthesis.[4] The photocatalytic agents in artificial photosynthesis can be semiconductors or dyes. In this perspective, we discuss the present status of the approaches to reduce CO2 to the production of alternative fuels based on the recent literature as well as some of our own findings.

Basics

Photocatalytic CO2 reduction with water is akin to natural photosynthesis, a process wherein plants convert CO2 and H2O to oxygen and carbohydrates in the presence of sunlight. In this process solar energy is being converted and stored in the form of chemical bonds (carbohydrates). It is a combination of water oxidation and carbon dioxide reduction (or CO2 fixation) reactions and involves both light and dark reactions.

The process of photocatalytic CO2 reduction on a semiconductor photocatalyst is illustrated in Figure . Similar to photocatalytic water splitting,[5] this process also involves three steps. They areUpon illumination of light the photocatalyst absorbs light, and electrons in the VB get excited to the CB, leaving behind holes in the VB. Reduction of CO2 is an uphill reaction, and in order to enable these uphill reactions, CB and VB of the photocatalyst should straddle the reduction potential of CO2 and the oxidation potential of water (Figure ). A negative CB edge relative to the reduction potential CO2 facilitates the transfer of electrons from the CB to CO2. A positive VB edge relative to the water oxidation potential facilitates the transfer of holes from the VB to water. Therefore, under the irradiation of light, transfer of electrons from the CB to CO2 as well as from water to the VB are thermodynamically favorable, whereas direct reduction of CO2 using water is an uphill reaction. Electrons in the CB reduce CO2, whereas holes in the VB oxidize water to oxygen (Figure ).

Figure 1

(a) Schematic illustration of mechanism and (b) relative energy levels of photocatalytic reduction of CO2 on a semiconductor photocatalyst.

Absorption of light (hν > Eg) by the semiconductor;

Separation of charge carriers (electron–hole);

Redox reactions at the surface

The formal reduction potentials of the reactions associated with the photoreduction of CO2 and H2O are given in eqs –8 (at pH 7).[6] One-electron reduction reactions of CO2 such as formation of CO2•– are not feasible due to their large negative reduction potentials (E(CO2/ CO2•–) = −1.85 V vs SHE) relative to the conduction band (CB) edges of many of the semiconductors. Highly negative reduction potential arises from the change in hybridization of C from sp2 to sp3.[6] Reduction potentials for the formation of HCOOH, HCHO, CH3OH, and CH4 are small (−0.665, −0.485, −0.399, and −0.246 V) and positive relative to CB edges of several of the semiconductors. It is therefore preferabe to undergo proton-coupled electron transfer (PCET) wherein electron transfer to CO2 is associated with proton transfer.

The ΔG values of the reactions are obtained from the equation, ΔG = −nFEcell. The ΔG values of the above reactions are positive (nonspontaneous), being the least for the formation of HCOOH and maximum for the formation of CH4 (HCOOH < CO < HCHO < CH3OH < CH4). However, they are also more positive compared to the ΔG of splitting of water.

The mechanism of reduction of CO2 on metallic surfaces has been widely discussed.[6] Here, we briefly discuss the mechanism of CO2 reduction on the TiO2 surface.[6] There are several possible configurations for CO2 adsorbed on the photocatalyst as shown in Figure .[7] Spectroscopic studies have revealed the presence of bent CO2 on the TiO2 surface under illumination. Surface Ti3+ adsorbs CO2 and forms bent CO2•– by transferring an electron, and Ti3+ gets oxidized back to Ti4+. The H atom present on the surface assists in the further reduction of CO2•–. Gaseous CO2 is a linear molecule, whereas the surface-bound carbon species have a bent structure. The adsorbed carbonate or CO2•– species are active in the photoreduction of CO2.[8] LUMO of the surface chemisorbed CO2 species (carbonate or CO2•–) is lower compared to that of gaseous CO2, thereby making it easier to accept excited electrons form the semiconductors. CO2 is an acidic molecule; therefore, basic metal oxides have greater tendency for CO2 adsorption.

Figure 2

Schematic of the possible CO2 adsorption configurations on surfaces of photocatalysts. Reproduced with permission from ref (7).

The reduction potentials for the PCET reactions are within the range of −0.7 to −0.2 V and are close to the reduction potential of water (−0.414 V at pH 7). The reduction products of CO2 (HCHO, CH3OH, and CH4) require multielectron transfer, whereas the reduction of water is a two-electron process. Reduction of water therefore competes with the reduction of CO2. Because of this, the extent of hydrogen generation is greater in liquid-phase reactions compared to vapor-phase reactions. Yields of photocatalytic processes are often much lower than the amount of photocatalysts. It is possible to obtain the products from stoichiometric reactions involved with unstable photocatalysts. It is, therefore, necessary to present the activities in terms of TON and TOF (TON per unit time) in order to distinguish from stoichiometric reactions. Generally, for catalytic reactions TON is ≫1.

Methods of Reducing CO2

Reduction of CO2 can be carried out in the liquid phase or gas phase. In the liquid-phase reaction, reduction of CO2 is carried out with a saturated aqueous solution of CO2. Limited solubility of CO2 in water is the crucial problem to drive photocatalytic reduction of CO2 efficiently.[9] The solubility of CO2 in water can be improved by using additives such as NaOH, NaHCO3, or Na2CO3. These additives enhance the CO2 solubility, though the reduction of bicarbonate and carbonate species is more difficult. Surface adsorption of H2O is more preferable over CO2 in the liquid phase, and thus the reduction of water is favorable. On the other hand, gas-phase reactions were carried out with humidified CO2. In order to explore the effect of the method of evaluation on activities, Xie et al.[10] have employed TiO2 and Pt-TiO2 as photocatalysts and studied the reduction of CO2 in both the liquid phase and gas phase. CH4 production activity is nearly 3 times more, whereas hydrogen production is less in gas-phase compared to liquid-phase reactions. Nearly 3–4 times more selectivity in CO2 reduction over water reduction is observed in the gas phase.

Quantum dot photocatalysts have gained importance due to their high surface area and shorter charge transfer pathways.[11] In addition, they possess enlarged band gaps and shifted band positions because of quantum confinement that provides more potential energy for photochemical reactions.[12] In a recent study, CsPbBr3 quantum dots were employed for photocatalytic reduction of CO2. The QDs are more active than the bulk compound, giving CO 49.5 μmol g–1 and CH4 22.9 μmol g–1 after 12 h (100 W Xe lamp with an AM 1.5G filter).[13]

The efficiency of photocatalysts depends on the morphology, exposed facets, size and surface vacancies, etc. The activity in the liquid phase is also affected by the pH of the reaction medium, surface hydroxyl groups, solvent, and additives. The increase in pH increases the rate of the reactionThis leads to different concentrations of species (CO32–, HCO3–, and CO2) at different values of pH.[14] The different chemical species get adsorbed to different extents on the catalyst surface and have different reduction potentials.[15] Thus, addition of NaOH could improve the dissolution of CO2, thereby increasing the efficiency of photoreduction of CO2 on TiO2-supported Cu catalysts.[16] In the gas phase, photocatalytic activity is affected by the surface properties of the photocatalysts, CO2–H2O ratio, and feed pressure, temperature, etc.

It is often customary to use sacrificial reagents to consume the holes. The addition of a sacrificial reagent to the reaction mixture enhances photocatalytic reduction, but in most of the cases the reagent would contribute to the yield of products. For example, methanol is used as a sacrificial hole acceptor. The reaction mechanism for the oxidation of methanol involves a step in which an electron is injected into the conduction band of the photocatalyst. This implies that part of the CO2 reduction products are formed through the action of holes and not that of the electrons.[17] The presence of organic adsorbents (CH3CO2H, CH3OH, HCO2H, etc.) on the surface of the photocatalyst plays an important role in the photoreduction of CO2. The presence of CH3CO2H on the photocatalyst surface leads to the formation of CH4. This can occur by the conversion of CH3CO2H to CH4 via the photo-Kolbe reaction.[18] Photocatalysis without any organic adsorbents shows a negligible amount of CH4 production.[19]

Semiconductor-Based Photocatalysts

TiO2 has been employed extensively as a photocatalyst since the early report by Fujishima et al.[20] for the photochemical reduction of CO2. A systematic study on photocatalytic as well as photoelectrochemical reduction of CO2 was carried out in aqueous suspensions of semiconductor powders such as TiO2, ZnO, CdS, Sic, etc.[20] The yields of methanol are well correlated with the CB edge position of the semiconductors, indicating the necessary requirement of a negative CB edge relative to the CO2/CH3OH reduction potential. Anpo et al.[21] have shown the necessity of the presence of H2O for the photoreduction of CO2 by employing titanium oxides anchored within the micropores on zeolites.

Use of Cocatalyst

Deposition of metals such as Pt, Au, Pd, Ag, and Cu on the photocatalysts affects the photochemical reduction of CO2. The metals catalyze the reduction reactions upon receiving electrons from the photocatalyst. Ag shows superior activity for CO2 reduction due to weak binding with the CO, whereas Pt gets poisoned by CO. Pt is also highly selective to the reduction of water.[6] Iizuka et al.[22] have reported photocatalytic reduction of CO2 over Ag cocatalyst-loaded anisotropic ALa4Ti4O15 (A = Ca, Sr, and Ba) using water as a reducing reagent. Here, the Ag cocatalyst efficiently reduces CO2, the high yields of CO arising from the separate reaction sites for reduction and oxidation. Thus, coupling of semiconductors with bimetallic systems also shows promising results. Thus, Cu–Pt bimetallic cocatalysts loaded on TiO2 (Cu–Pt/TiO2) nanotubes reduce CO2 to CH4, C2H4, C2H6, etc., with 4-fold improvement in the conversion in the presence of H2O under solar irradiation.[23] Codeposition of oxidation and reduction cocatalysts such as RuO2 (1 wt %) and Pt (1 wt %) on Zn2GeO4 causes significant improvement in CH4 production.[24]

Doping or Codoping

Since the UV component in the sunlight is limited (∼4%), researchers have mainly explored visible-light (∼43%) sensitive photocatalysts. The electronic band structure of the wide-band-gap semiconductors is altered by the incorporation of foreign elements into the lattice of the semiconductors. For example, N doping or Ce doping in TiO2 shows a red-shift in the absorption onset with visible absorption and superior activity compared to undoped TiO2.[6] N doping in TiO2 as in CuO/TiO2–N enhances the conversion of CO2 to CH4 under solar irradiation.[25] Nakanishi et al.[26] have employed metal-doped NaTaO3 (NaTaO3:A, where A = Ca, Mg, Sr, Ba, or La) with Ag cocatalysts under UV–visible irradiation (Figure ) for the reduction of CO2 to CO. Ag deposited NaTaO3:Ba exhibits superior activity among all cation-doped NaTaO3, in the absence of any additives. Addition of NaHCO3 enhances CO production activity with NaTaO3:Ba being superior in the reduction of CO2 with nearly 90% selectivity toward the reduction of CO2. However, a considerable fraction of electrons is utilized in the reduction of water. Herein, water acts as a reducing agent and gets oxidized to oxygen.

Figure 3

Photocatalytic CO2 reduction on (a) Ag-NaTaO3:Ca, (b) Ag-NaTaO3:Sr, and (c) Ag-NaTaO3:Ba in the liquid phase under UV irradiation. Figure legends: hydrogen (open circle), oxygen (solid circle), and CO (triangle). Reproduced with permission from ref (26).

Solid Solutions

Electronic band structures of wide-band-gap semiconductors can be tailored by forming solid solutions. Liu et al.[27] have employed (Zn1+Ge)(N2O) nanostructures for the reduction of CO2 to CH4 under visible-light irradiation. Figure shows the electronic absorption spectra of the nitridation product of Zn2GeO4 (band gap 4.53 eV) with varying duration of nitridation at 700 °C. The absorption onset is gradually red-shifted with the increase in nitridiation period, and samples turn yellow in color. Superior CO2 reduction activity under visible-light irradiation is obtained with the product having a band gap of ∼2.4 eV, obtained after a nitridation period of 6 h (Zn 26.4%, Ge 19.6%, N 40.8%, and O 13.2%) with CH4 yield of 1.63 μmol h–1 with O2:CH4 ratio of 2:1 since the CH4 is an eight-electron reduction production and O2 is a four-electron oxidation production.

Figure 4

(a) Electronic absorption spectra of nitridation products of Zn2GeO4 and (b) corresponding photocatalytic methane product activities. (c) Curve depicting the recyclability and stablitity of the photocatalyst. Reproduced with permission from ref (27).

Multicomponent Photocatalysts

TiO2–RuH

Huang et al.[28] have developed a mononuclear C5H5–RuH complex oxo-bridged with TiO2 for photoreduction of CO2 to CH4 under visible-light irradiation with a quantum efficiency of 0.56%. In this system, C5H5–RuH serves as the photon harvester and water-oxidation site, whereas TiO2 acts as the electron collector and site of CO2 reduction (Figure ). The amount of CH4 increases linearly under visible-light irradiation with time, whereas under dark conditions, no CH4 is detected. The photocatalytic activity is maximum at Ru loading of 0.5 wt % on TiO2. The superior efficiency of the catalyst for CO2 reduction is attributed to the long-lived D+–C–A– charge-separated state.

Figure 5

(a) Schematic diagram of the proposed mechanism of light-induced charge transfer in C5H5–RuH–TiO2. (b) Electron decay kinetics CpRu0.5/TiO2 and TiO2. (c) Reduction of CO2 to CH4 on CpRu0.5/TiO2 under visible-light irradiation. (d) Dependence of CH4 evolution with varying the Ru content. Reproduced with permission from ref (28).

Semiconductor MOFs

Metal–organic frameworks have gained importance as photocatalysts in recent years since many of them possess high surface areas and good gas-adsorption properties. MOFs with a high CO2 uptake can be considered as potential candidates for the CO2 reduction. Here, we discuss a study based on a combination of TiO2 with MOFs. Li et al.[29] have developed Cu3(BTC)2@TiO2 (BTC-benzene-1,3,5-tricarboxylate) core–shell photocatalysts for the reduction of CO2. Figure shows a schematic diagram as well as TEM and SEM images of the corresponding nanostructures. There is a shell of TiO2 over the Cu3(BTC)2 microcrystals, and the BET surface areas of Cu3(BTC)2 and Cu3(BTC)2@TiO2 are 1183 m2 g–1 and 756 m2 g–1, respectively. The decrease in the surface area of the core–shell structure is attributed to the presence of a TiO2 shell which does not have a large surface area. CO2 adsorption based on the weight of the Cu3(BTC)2 component (80.75 cm3 g–1) of the core–shell structure is comparable to the bare Cu3(BTC)2 (49.17 cm3 g–1). This observation implies that the CO2 molecules can easily go through the macroporous TiO2 shells and get adsorbed at the microporous Cu3(BTC)2 cores. The photocatalytic activities of the TiO2 and Cu3(BTC)2@TiO2 core–shell structure are given in Figure . The CH4 production rate of the MOF is higher than that for TiO2, and the MOF does not generate H2. Note that the Cu3(BTC)2 itself is inactive.

Figure 6

(a) Schematic diagram, (b) TEM, and (c) SEM images of Cu3(BTC)2@TiO2 (BTC-benzene-1,3,5-tricarboxylate). (d) CO2 uptake and (d) comparison of CO2 reduction activities on Cu3(BTC)2@TiO2 (BTC-benzene-1,3,5-tricarboxylate) and Cu3(BTC)2@TiO2 (BTC-benzene-1,3,5-tricarboxylate). Reproduced with permission from ref (29).

Composites of Graphitic C3N4

Graphitic C3N4 has been used as a photocatalyst due to its high stability and visible response (Eg = 2.7 eV).[30] Due to the fast recombination of the excited electron–hole pair, C3N4 is employed in combination with other materials. Loading of noble metals on the surface of g-C3N4 facilitates the interfacial transfer of electrons from g-C3N4 to the metal particles. For example, loading of Pt on g-C3N4 increases the rate of reduction of CO2. The selectivity depends on the amount of Pt. An increase in Pt content from 0 to 1 wt % increases the yields of CH4 and CH3OH, and a Pt content of 0–0.75 wt % favors formation of HCOOH.[31]

g-C3N4 is also used to form heterojunctions to suppress the recombination of electrons and holes. The introduction of g-C3N4 in NaNbO3 nanowires increases the photocatalytic performance considerably (8-fold higher activity than naked C3N4).[32] Similarly, deposition of Ag3PO4 on g-C3N4 enhances the photocatalytic performance.[33] Ong et al.[34] have demonstrated a robust method to synthesize proton-functionalized/protonated g-C3N4 (pCN). The positively charged g-C3N4 can form a hybrid nanostructure with the negatively charged GO sheets, forming rGO/pCN which is more active than the g-C3N4 itself. Addition of rGO has a substantial effect on the photocatalytic performance of the g-C3N4 which increases with the increase in rGO content. The highest CH4 production occurs for the 15 wt % rGO/pCN (13.93 μmol g–1, 5.4 times higher than C3N4). The higher activity by the addition of rGO is due to its lower Fermi level relative to the conduction band of C3N4. The formation of this heterojunction suppresses the charge recombination.

Semiconductor Homojunctions

The dependence of the photocatalytic activity of TiO2 on the facet and shape of the catalyst particles is known. Yu et al.[35]have carried out photocatalytic CO2 reduction on TiO2 with coexposed facets of (001) and (101). DFT electronic structure calculations show that the CB and the VB of the (101) facets of TiO2 are positioned slightly below the CB and the VB of the (001) facets of TiO2, respectively. This is similar to type II band alignment in semiconductor-based heterostructures. Apparently, exposure of the (001) facets increases the amount of CH4. The facets help in the separation of charges with the electrons migrating to the (101) facet and the holes migrating to the (001) facet. A study based on TiO2 with systematically varied ratios of (001) and (101) facets shows that the photocatalytic production of CH4 increases with the increase in the fraction of (001) facets, reaching a maximum of 1.35 μmol h–1 g–1 at 58% of (001) facets (Figure ). Further increase in the (001) facet fraction decreases the photocatalytic activity. Similarly, facet-derived heterostructures formed by ZnSn(OH)6 also give rise to superior photocatalytic activity and longer charge carrier lifetimes.[36]

Figure 7

Comparison of (a) density of states and (b) relative band positions in (001) and (101) facets of TiO2. (c) Photocatalytic CO2 reduction to CH4 with different ratios of (001) to (010) facets in TiO2. Reproduced with permission from ref (35).

Cosubstitution of aliovalent anions such as N3– and F– for O2– in an oxide has a huge effect in altering the electronic band structure as well as optical and catalytic properties.[37] N and F cosubstituted TiO2 (anatase) has a band gap of 2.6 eV and is yellow in color (Figure ). N,F-TiO2 reduces CO2 to CO using water as the reducing agent under natural solar irradiation. Ag deposition on N,F-TiO2 is more effective, Pt being the least effective for the reduction of CO2. N,F-TiO2 has been explored for the construction of homojunctions with undoped TiO2. From the first-principles calculations on N,F-TiO2 it is evident that the CB of N,F-TiO2 remains the same as that of TiO2; however, the VB is dominated by the N 2p states. Band alignment of the junction of TiO2 and N,F-TiO2 can be considered as quasi-type II. Reduction of CO2 is superior with homojunctions over individual TiO2 and N,F-TiO2, showing the effectiveness of homojunctions (Table ).[22,24,27,38−44]

Figure 8

(a) Comparison of tauc plots of N,F-TiO2 (TiO1.8N0.1F0.1) and undoped TiO2. (b) Comparison of reduction of CO2 to CO on Ag, Au, and Pt deposited N,F-TiO2 with bare N,F-TiO2 under the irradiation of sunlight. (c) and (d) Comparison of reduction of CO2 to CO on TiO2:N,F-TiO2 homojunctions with individual TiO2 and N,F-TiO2 under the irradiation of sunlight.

Table 1

Photocatalytic CO2 Reduction Yields Obtained by Various Photocatalysts in Gas-Phase Reactions Using Water as the Reducing Agent

photocatalyst	light source	reaction medium	major products	yields (μmol h–1 g–1)	references
N,F-TiO2:Ag	solar	CO2 and H2O	CO	8.7	present study
TiO2:N,F-TiO2	solar	CO2 and H2O	CO	6.0	present study
TiO2 (anatase:brukite)	solar	CO2 and H2O	CO	2.1	(40)
P25 (anatase–rutile)	solar	CO2 and H2O	CO	1.3	(40)
CuO–TiO2–xNx	solar	CO2 and H2O	CO	41 ppm h–1 g–1	(41)
α-Fe2O3/Cu2O	visible	CO2 and H2O	CO	1.67	(38)
Pt:Ti3+-doped TiO2	visible	CO2 and H2O	CH4	1.6	(42)
WO3/Au/In2S3	visible	CO2 and H2O	CH4	0.42	(39)
Pt:Zn2GeO4:RuO2	UV	CO2 and H2O	CH4	6.7	(24)
anatase TiO2 (010) facet	UV	CO2 and H2O	CH4	1.2	(43)
Cu2ZnSnS4-RGO-TiO2	UV	CO2 and H2O	CH4	120 ppm h–1 g–1	(44)
Ag-loaded BaLa4Ti4O15	UV	CO2 and H2O	CH4	14.3	(22)
(Zn1+xGe)(N2Ox)	visible	CO2 and H2O	CH4	2.5	(27)

Semiconductor Heterostructures

Type II band alignment in heterostructures causes effective charge separation. Kim et al.[44] have employed Cu2ZnSnS4/TiO2 heterostructures with type II band alignment for the reduction of CO2 under solar irradiation (Figure ). CH4 yield of 119 ppm h–1 g–1 is achieved with the heterostructure, which is nearly 12 times higher compared to TiO2 alone. The superior activity in heterostructures is attributed to the improved light absorption and surface area in addition to effective charge separation. Wang et al.[45] have employed CdSe/Pt/TiO2 heterostructures for the visible-light-induced reduction of CO2. These heterostructures exhibited reduction of CO2 under visible-light irradiation, in the presence of H2O, leading to the formation of CH4 (48 ppm g–1 h–1) and methanol (3.3 ppm g–1 h–1) and H2 (trace) and CO (trace). Replacing Pt with Fe only resulted in the formation of H2 (>55 ppm g–1 h–1) which indicates that Fe catalyzes photoreduction of water rather than of CO2. Wang et al.[46] have used PbS/Cu/TiO2 and CdSe/Pt/TiO2 heterostructures for the photoreduction of CO2. Both of them suffer from photodegradation due to the oxidation of CdSe and PbS quantum dots. CsPbBr3 QD/GO shows higher activity than CsPbBr3 QDs with CO and CH4 rates of 58.7 and 29.6 μmol/g in 12 h.[13] Roy in this laboratory has been able to reduce CO2 to CO by solar radiation using Cd4P2X3 (X = Cl, Br, or I) in the absence of any sacrificial agent.

Figure 9

(a) Schematic illustration of the process of charge transfer and CO2 reduction in Cu2ZnSnS4/TiO2 heterostructures. (b) Comparison of photocatalytic activities of Cu2ZnSnS4/TiO2 with different compositions and individual Cu2ZnSnS4 and TiO2. Reproduced with permission from ref (44).

ZnO/Pt/CdS heterostructures with type II band alignment and metal nanoparticles present on the oxide surface have shown excellent photocatalytic activity for hydrogen evolution.[47] These heterostructures would be potential candidates for the efficient reduction of CO2 under visible-light irradiation. However, Pt is often poisoned by CO; therefore, it would be beneficial to use Ag in place of Pt. Cu is also effective for the reduction of CO2 but is unstable. In view of this, we have employed heterostructures containing bimetallic alloys and systematically evaluated the photocatalytic activities of ZnO/M/CdS (M = Ag, Au, Pt, Ag1–Au, Ag1–Cu) heterostructures. ZnO/Ag/CdS and ZnO/Au/CdS exhibit significant CO2 reduction activity, whereas ZnO/Pt/CdS exhibits mainly reduction of water. Bimetallic Ag0.5Au0.5 exhibits superior CO2 reduction activity compared to individual Ag and Au under visible-light irradiation. Photocatalytic reduction of CO2 under visible-light irradiation on ZnO/Ag1–Cu/CdS heterostructures is presented in Figure . ZnO/Ag0.75Cu0.25/CdS heterostructures exhibit superior activity (327 μmol h–1 g–1) with the least hydrogen evolution activity among all the compositions. The ZnO/CdS interface causes effective separation of charge carriers, and the bimetallic alloy, Ag0.75Cu0.25, is a stable and efficient cocatalyst for the reduction of CO2 (Table ).[10,48−52] A noteworthy feature of these heterostructures is that they are photocatalytically active even in the absence of any sacrificial agent in the gas phase with a CO production activity of 2 μmol h–1 g–1.

Figure 10

(a) Comparison of photocatalytic activities and (b) schematic illustration of process of charge transfer and CO2 reduction in ZnO/Ag1–Cu/CdS (x = 0–0.75) heterostructures.

Table 2

Photocatalytic CO2 Reduction Yields Obtained by Photocatalysts in Liquid-Phase Reactions

photocatalyst	light source	reaction medium	products	yields (μmol h–1 g–1)	references
ZnO/Ag0.5Cu0.5/CdS	sunlight	2-propanol·H2O	CO, CH4, H2	162 (CO)	present work
ZnO/Ag0.75Cu0.25/CdS	visible	2-propanol·H2O	CO, CH4, H2	327 (CO)	present work
Pt:TiO2	UV	H2O	H2, CH4, CO	1.4 (CH4)	(10)
ZnS:Cd	UV (Xe)	2-propanol·H2O	HCOOH, H2	10 μmol h–1 (HCOOH)	(48)
CdS	visible	2-propanol·H2O	HCOOH, CO, H2	10 (HCOOH)	(49)
ZnFe2O4/TiO2	UV	cyclohexanol·H2O	HCOO–	22 (HCOO–)	(50)
Ru(II)-complex/C3N4	visible	methanol–DMA	CO	5.7 (CO)	(51)
Ru(II)-complex/Ag/C3N4	visible	K2CO3–EDTA·H2O	HCOO–, H2	83 (HCOO–)	(52)

Z-Scheme Photocatalysts

Wang et al.[38] have recently developed an α-Fe2O3/Cu2O Z-scheme photocatalyst for the reduction of CO2 using water as the reducing agent (Figure ). α-Fe2O3 is an n-type material, while Cu2O is a p-type material. The photocatalytic activity under visible-light irradiation increases with the increasing Cu2O content reaching a maximum yield of 5 μmol g–1 in 3 h with a Cu:Fe ratio of 0.5:1. Further increase in the Cu2O content decreases the activity. Band positions estimated from electronic absorption, ultraviolet photoelectron, and X-ray photoemission spectroscopic studies show that α-Fe2O3/Cu2O forms a p–n junction. The observed photocatalytic activity arises because of the operation of the Z-scheme.

Figure 11

(a) Schematic illustration of the process of Z-scheme charge transfer and CO2 reduction in α-Fe2O3/Cu2O heterostructures. (b) Comparison of yield of CO production with varying the Cu content in these heterostructures. Reproduced with permission from ref (38).

Takayama et al.[53] have employed CuGaS2–RGO–TiO2 for the liquid-phase reduction of CO2 under UV–visible irradiation (Figure ). CuGaS2 acts as a reduction center, whereas TiO2 acts as an O2 evolution center. RGO facilitates the electron transfer from the CB of TiO2 to the VB of the CuGaS2. In the absence of RGO, the photocatalytic activities are negligible. The photocatalytic production of hydrogen (28.8 μmol h–1) is larger than that of CO (0.15 μmol h–1).

Figure 12

(a) Schematic illustration of the process of Z-scheme charge transfer and CO2 reduction and (b) photocatalytic evolution of H2, O2, and CO in CuGaS2–RGO–TiO2 under UV–visible light irradiation. Reproduced with permission from ref (53).

Outlook

While finding photocatalysts with viable photostability remains a great challenge for reducing CO2 in the absence of any hole scavenger, chemists have no option but to discover the right catalysts. Considering the small UV fraction in solar irradiation, we need to concentrate on visible-light-sensitive photocatalysis. In this regard, N and F codoped systems such as ZnO1–(N,F), TiO2–(N,F), and SrTiO3–(N,F) could be potential candidates for both visible-light sensitization as well as photostability. Recent work suggests beneficial results by using bimetallic alloys in heterostructures. There may be new catalytic materials such as Cd4P2X3 (X = Cl, Br, or I) with favorable characteristics. Recently it has been shown that cross-linked materials of C3N4 and others show remarkable properties. Exploring them for the reduction of CO2 is also important.

Bipolar membranes are employed between anodic and cathodic compartments in order to avoid the migration of reactive species. Therefore, carrying out oxidation and reduction reactions in chambers separated with bipolar membranes would be a good strategy for beneficial conversion of CO2. Since it is likely that CO will be the prominent product of photoreduction of CO2, it would require further processing to generate methanol and other products. Direct photochemical conversion of CO2 to methanol and other useful compounds continues to remain an important research target.