Alternative photocatalysts to TiO2 for the photocatalytic reduction of CO2

Applied Surface Science 391 (2017) 149–174

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Alternative photocatalysts to TiO2 for the photocatalytic reduction of CO2 Aspasia Nikokavoura, Christos Trapalis ∗ Institute of Nanoscience & Nanotechnology, NCSR Demokritos, Agia Paraskevi Attikis, 15343, Athens, Greece

a r t i c l e

i n f o

Article history: Received 20 May 2016 Received in revised form 24 June 2016 Accepted 27 June 2016 Available online 29 June 2016 Keywords: Photocatalytic reduction CO2 Photocatalysts Semiconductors Solar fuels Artificial photosynthesis

a b s t r a c t The increased concentration of CO2 in the atmosphere, originating from the burning of fossil fuels in stationary and mobile sources, is referred as the “Anthropogenic Greenhouse Effect” and constitutes a major environmental concern. The scientific community is highly concerned about the resulting enhancement of the mean atmospheric temperature, so a vast diversity of methods has been applied. Thermochemical, electrochemical, photocatalytic, photoelectrochemical processes, as well as combination of solar electricity generation and water splitting processes have been performed in order to lower the CO2 atmospheric levels. Photocatalytic methods are environmental friendly and succeed in reducing the atmospheric CO2 concentration and producing fuels or/and useful organic compounds at the same time. The most common photocatalysts for the CO2 reduction are the inorganic, the carbon based semiconductors and the hybrids based on semiconductors, which combine stability, low cost and appropriate structure in order to accomplish redox reactions. In this review, inorganic semiconductors such as single-metal oxide, mixedmetal oxides, metal oxide composites, layered double hydroxides (LDHs), salt composites, carbon based semiconductors such as graphene based composites, CNT composites, g-C3 N4 composites and hybrid organic-inorganic materials (ZIFs) were studied. TiO2 and Ti based photocatalysts are extensively studied and therefore in this review they are not mentioned. © 2016 Elsevier B.V. All rights reserved.

1. Introduction 1.1. Solar energy – solar fuels The sun constitutes the major energy source in Earth, by providing in one hour all the energy needed by humanity for one year. The basic problem for the inhabitants of the Earth, is that they are not able to exploit these huge amounts of energy [1]. The total emitted solar radiation reaching the Earth’s surface is composed of infrared radiation (52%, >700 nm), visible radiation (43%, 400–700 nm) and ultraviolet radiation (5%, <400 nm). There are several types of solar energy collectors, some of them have flat surface and collect the

Abbreviations: AC, activated carbon; ATP, adenosine-tri-phosphate; CB, conduction band; CCS, carbon capture and storage; CNTs, carbon nano-tubes; CTA, cetyl-trimethyl-ammonium; DAC, direct air capture; DHF, di-hydro-furan; DMF, di-methyl-formamide; LDHs, layered double hydroxides; NADPH, nicotinamide adenine dinucleotide phosp-hate; OFMR, optical fiber monolith reactors; PEC, photo-electrochemical reduction; RGO, reduced graphene oxide; SSR, solid state reaction; THF, tetra-hydro-furan; VB, valence band; ZIFs, zeolitic imidazolate frameworks. ∗ Corresponding author. E-mail address: [email protected] (C. Trapalis). http://dx.doi.org/10.1016/j.apsusc.2016.06.172 0169-4332/© 2016 Elsevier B.V. All rights reserved.

solar irradiation without concentrating it, so its intensity is rather low. There are also solar energy collectors that concentrate the solar irradiation (such as reflectors and luminescent solar collectors) but they are not commercially available because they are not stable and they exhibit inadequate collection efficiency [2]. For over 50 years, serious attempts have also been performed in order to produce useful compounds-fuels by utilizing solar energy. The solar fuels are compounds produced via a biomimetic approach that have captured and stored solar energy in their chemical bonds (chemical energy) [3]. The production of solar fuels is a great challenge for the scientific community. Solar fuels constitute a broad group of chemicals that can be used for electricity generation, transport and industrial purposes. The two main categories of solar fuels are (a) hydrogen and (b) carbon based fuels such as methanol (CH3 OH), carbon monoxide (CO) and methane (CH4 ). Another significant goal for the scientists is the large scale production of the solar fuels, their transportation, their storage and the opportunity to be commercially available [1].

150

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

1.2. Production of solar fuels 1.2.1. Hydrogen-water splitting Hydrogen (H2 ) is a very useful solar fuel with serious advantages and disadvantages. As major advantages are considered, its abundance in water and biomass, its convenient storage as hydride, the absence of pollutants after its combustion and the high amount of energy released when it is used as a fuel. On the other hand, hydrogen must be highly compressed during its transportation and this is an expensive process or it must be stored as a metal hydride which is an expensive and rather dangerous process too [4]. It is a transport fuel but it is also used in many industries. In nature, hydrogen is produced from water splitting via various molecular reactions involving PSII (photobiological water splitting). The water splitting mechanism includes the formation of a large number of reaction intermediates which finally produce dioxygen, electrons and protons. Advanced techniques have shown an interaction between the metal center which acts as a catalyst and its protein environment that induce the oxidation of the water molecules. The mechanism of the biological formation of hydrogen by the reaction of electrons and protons, formed from photosynthesis (PSI or PSII), is not fully understood, but it is clear that enzymes such as hydrogenases and nitrogenases play a crucial role, as they produce hydrogen from protons and electrons formed via water splitting [5]. By mimicking the nature, scientists tried to use to advantage the most abundant material in Earth, water, in order to produce hydrogen. Steam reforming, constitutes a process which involves the generation of hydrogen from CH4 and steam (H2 O) via the production of CO, is mainly used for hydrogen production. Besides, hydrogen is generated by coal gasification, a process that involves the reaction of coal with O2 and steam at high temperature and pressure and produces mainly a mixture of hydrogen, methane, CO2 and CO. Biological, pyrolysis and thermochemical processes that use the biomass are also used for the production of H2 . All these conventional methods followed for the formation of H2 require enhanced amounts of energy. Besides, the formation of pollutants such as CO2 , as byproducts, is not avoided. It is also widely known that hydrogen is produced via electrolytic or thermochemical or photobiological or photocatalytic water splitting. The thermochemical water splitting uses the heat from the sun, whereas the photocatalytic water splitting uses the sunlight irradiation and photocatalysts i.e. materials that accelerate the reaction, they are found in tiny amounts in the reaction mixture and they are not consumed at the end of the reaction. The general principle is that the energy of the absorbed photons must be higher than the band-gap energy of the photocatalyst, in order to achieve the photogeneration of electrons and holes participating in the redox reactions which produce hydrogen and oxygen. Specifically, photoinduced electrons, transported from the valence band (VB) of the photocatalyst to the conduction band (CB), reduce H+ to form H2 , on its surface. Photogenerated holes oxidize H2 O to H+ and O2 in VB. Photocatalysts such as the TiO2 analogues and other metal oxides and salt composites have been widely used [4,6]. 1.2.2. Carbon based fuels – natural/artificial photosynthesis The natural photosynthesis is the way of the biological world to produce organic fuels, such as sugars, using mainly two abundant compounds, H2 O and CO2 , via the harvest and exploitation of solar energy. The natural photosynthetic process exhibits low overall efficiency, still it constitutes an inspiration for scientists in order to produce useful fuels from CO2 , H2 O and sunlight. The natural photosynthesis involves the utilization of compounds that can be oxidized producing electrons, compounds that can be reduced from these electrons producing fuels and of course solar irradiation. Green plants, algae and cyanobacteria absorb the solar energy in order to oxidize water to O2 , protons and electrons. The pho-

togenerated electrons possess adequate negative redox potential, thus they are able to reduce photogenerated protons and CO2 in order to form hydrogen gas, carbohydrates and lipids. Plants consist of a number of subsystems combined together, so that to perform the photosynthetic process. Antenna systems are responsible for targeting the appropriate wavelengths of light and transporting that energy to the reaction centers, where the transfer of photoinduced electrons takes place and charge separated states are formed. Water oxidation and O2 formation is accomplished at the reaction center in PSII after the absorption of four photons whilst H+ and CO2 reduction is carried out by the biological molecule NADPH (nicotinamide adenine dinucleotide phosphate). The formation of ATP (adenosine triphosphate), which has a key role in the transportation of energy by several enzymes and the generation of transmembrane proton gradients is also realized by reduction mechanisms. In conclusion natural photosynthesis is a very complicated exergonic process, consisting of a large number of redox reactions which lead to the production of carbon based fuels [7,8]. On the other hand, by the term “artificial photosynthesis” is meant the utilization of CO2 , renewable (solar) energy, synthetic catalysts and water in order to produce a large range of fuels/chemicals such as hydrogen and carbon based compounds [9]. Water is used as reductant that is the electron source. In artificial photosynthesis the first step is the absorption of the ultraviolet or visible radiation. The second step is the charge generation and separation and the last step is the catalytic reaction [10]. Inspired by the natural photosynthesis, where chlorophyll molecules diffuse sunlight, artificial photosynthetic reactions also need the presence of light harvesting complexes, such as dye sensitizers. The photon capture is followed by the electron-transfer reactions that lead to charge separation. This means that electrons and positive holes move apart in order to avoid charge recombination, leading to the formation of potent oxidizing or reducing species. These species drive the desirable redox reactions such as CO2 reduction. A major difference between natural and artificial photosynthesis is the capability of biological systems to induce self repair. Unfortunately, until now there is no man-made photocatalyst with the ability to retrieve the loss of reactivity during operation [11,12]. The artificial photosynthesis products are used as synthetic fuels for transportation and storage or as useful industrial materials (plastics, fertilizers, pharmaceuticals, chemicals etc.) [1]. The greatest benefit is obtained in case that a direct conversion of solar energy into fuels is achieved. The use of water as a raw material is an important perspective too. The catalysts for the artificial photosynthesis may be either nanomaterials with nature guided design, or molecular compounds or solid state components, or combinations of them. Their common feature is the way that they participate in the multielectron redox reactions during the artificial photosynthesis process. Abundant or less abundant materials can be used as homogeneous or heterogeneous catalysts. The most common molecular catalysts are based on ruthenium and constitute model systems for the study of artificial photosynthesis [13,14]. In Fig. 1 a comparison of natural photosynthesis and artificial photosynthesis is shown [14]. One principal scientific goal currently is the enhancement of the photocatalytic efficiency in the CO2 reduction by using technical means. Methanol and methane are the most important solar fuels resulting from the photocatalytic CO2 fixation as they can store a lot of hydrogen. Other possible products are CO, HCOOH, HCHO and H2 [15]. 1.3. Why the reduction of CO2 in the atmosphere is so important? Greenhouse effect and mitigation of CO2 The burning of fossil fuels, apart from their depleting them, has raised serious environmental issues, such as the anthropogenic Greenhouse effect which is considered as one of the most seri-

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

151

Fig. 1. Comparison of natural photosynthesis and artificial photosynthesis [14].

Fig. 2. Carbon dioxide neutral cycle with renewable methanol and liquid fuel production [18].

ous environmental problems on Earth. CO2 is one of the main greenhouse gases emitted by the combustion of coal, petroleum and natural gas in stationary and mobile sources and therefore its contribution to global warming is significant. Other important longliving greenhouse gases are: CH4 , N2 O and CFCs. By that means, these gases absorb infrared light irradiation and re-emmit it in the atmosphere, where it is trapped, resulting in global warming [16]. It is estimated that the CO2 concentration in the atmosphere rises every year at a rate higher than 2% and as a result, the average temperature on the Earth’s surface has increased around 0.4–0.8 ◦ C over the last century. It is predicted that the mean global temperature may rise by 1.9 ◦ C and that the level of CO2 might climb to 590 ppm. Global warming causes the melting of the ice at the poles, the expansion of the oceans and the loss of large ice quantities from Greenland and Antarctica which cause an increase of the sea level. The extreme meteorological phenomena (abnormal winds, aridity, floods) constitute some extra consequences of the Greenhouse effect. For example, a decrease in rainfall in certain areas may diminish the required water supplies for people, ecosystems, agriculture. The enhanced precipitation in other areas constitutes an extra worrisome factor. Besides, the last 25 years the increase in

CO2 concentration is spectacular and may lead to a possible alteration of the radiative balance on the planet. On the other hand, CO2 certainly is an important gas in the atmosphere because it takes part in the Carbon cycle as a carbon source [17–20]. Generally, in order to achieve the decrease in the CO2 concentration in the atmosphere, the carbon dioxide neutral cycle, otherwise known as the net carbon neutral cycle, should be completed. Several steps are important for the maintenance of this cycle, such as the recycling of CO2 after the fuel’s combustion, the use of renewable energy (wind energy, direct solar energy etc.) for the CO2 reduction, the capturing and recycling of CO2 not only from the point source but also from the atmosphere In Fig. 2 the carbon dioxide cycle with renewable methanol and liquid fuel is presented [18]. Beyond the artificial photosynthesis way to reduce the CO2 concentration in the atmosphere, there is also the Carbon Capture and Storage (CCS) pathway. CCS comprises the collection of CO2 emissions mainly from the industry and the power stations, the generation of almost pure CO2 using a wide range of techniques, and its transportation after compression and liquification, to the seabed or other geological deposits. CCS has important drawbacks, as CO2 from the transportation, the residential and the commer-

152

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

2. Photocatalytic approaches/mechanism for the reduction of CO2 -main categories of photocatalysts

Fig. 3. CO2 emission from different sectors [21].

cial sectors cannot be exploited. Furthermore, the CO2 separation from other gases needs large amounts of energy. Direct air capture (DAC) involves an alternative process with lower energy requirements. DAC involves the reaction of the weakly acidic CO2 with an alkali component in order to separate it, in the form of salt, from other neutral gases such as nitrogen and oxygen. Following that, a regeneration process of the CO2 from the salt and the recycling of the alkali component take place. DAC may be performed anywhere, so the pressurization and transportation via pipelines of CO2 is minimized [19,21]. In conclusion, pressurization, liquification and transportation of CO2 raise the cost of these processes. Furthermore, the cost of the required energy to convert CO2 into chemicals is rather high, there are no convenient commitments in order to increase the CO2 originating chemicals and the governments do not support economically these actions [22,23]. In Fig. 3 the CO2 emission from various sectors is presented [21]. Furthermore, the reduction of the CO2 concentration in the atmosphere can be accomplished via the thermochemical, electrochemical, photocatalytic, photoelectrochemical processes, as well as the combination of solar electricity generation, water splitting and CO2 reduction. Each of them has advantages and disadvantages. The main handicap of all the thermochemical approaches is that they require high temperatures, which means that the cost of heat supply is significantly raised. In other words, by following these thermochemical processes the energy costs are very high, whereas, actual mitigation of the atmospheric CO2 does not take place [24–26]. The electrocatalysis involves the use of electrodes and an external circuit. Specifically, at the anode, water is oxidized to give O2 and electrons. These electrons arrive at the cathode via an external circuit and reduce CO2 to several products such as CO, CH3 OH etc. In the case that appropriate electrodes and catalysts are used, high product yields are obtained. Besides, the reduction and oxidation reactions are performed in different spaces and thus the interference between the two kinds of reactions and their products is totally avoided. From the opposing point of view, the electrodes may be easily deactivated, or they are not rather stable. Also, the competitive reaction, the water splitting, affects the CO2 reduction efficiency. Apparently, the combination of photovoltaics (PV) and electrocatalysis is proved to be friendlier for the environment [24,27–29].

The photocatalytic approach (artificial photosynthesis) constitutes a very convenient way to reduce the atmospheric CO2 , by using solar energy, which is the most abundant kind of energy on Earth. Photocatalytic reduction of CO2 or artificial photosynthesis, as it is called, succeeds in reducing CO2 concentration in the atmosphere and generating fuels or/and useful organics at the same time [10,30]. Certainly, although the energy of the Sun is cheap and abundant, it is impossible to be exploited in the nights or in cloudy days, thus it is important to store it in a convenient way which involves high energy densities. Besides, reduction of CO2 requires large amounts of solar energy due to CO2 thermodynamics. Still, as far as the CO2 reduction is concerned, photocatalytic methods seem to be environmental friendly and they are preferred over other methods such as thermochemical or electrochemical ones. By the term of photocatalysis is meant the direct conversion of solar energy into chemical energy. Photocatalytic conversion of CO2 into a vast variety of useful compounds such as CO, CH4 , CH3 OH, HCHO, HCOOH is conducted either in liquid or in gas phase. In case that water molecules are present, two basic reactions take place, the first one is the photoinduced water splitting to oxygen and hydrogen (1), and the second one is the combination of the photoinduced activation of CO2 and the oxidation of water (2): h␯ + 2H2 O → 2H2 + O2

(1)

h␯ + H2 O + CO2 → Fuel + O2

(2)

Generally, a photocatalytic system for the CO2 reduction must (a) capture photons of light irradiation effectively, (b) produce efficiently photoinduced electrons and holes that (c) they are transferred quickly to the photocatalyst surface (d) to the appropriate catalytic sites where CO2 is reduced (e) without recombining quickly. The combination of photochemical and electrochemical processes in order to reduce CO2 into valuable fuels constitutes the photoelectrochemical (PEC) reduction of CO2 . The disadvantage of this process is that it is electricity-dependent but its advantage is its higher efficiency due to the lowering of the recombination rate of photogenerated electrons and holes. The difference between the electrochemical and the PEC reduction of CO2 , is that in the latter, the semiconductor electrode (photoelectrode) is irradiated by UV or visible light with photon energy greater than the band-gap energy of the semiconductor. External bias voltage is also necessary for the PEC CO2 reduction. Then, photoexcited electrons are transported from the valence band to the conduction band and then they travel via an external circuit to the cathode counter electrode where they perform the reduction reactions, via the formation of CO2 −* radical anion. Still, recombination and corrosions of the photoinduced species are not avoided. Besides, electron-hole pairs travel through the electrolyte and perform oxidation and reduction reactions on the surface of the appropriate electrodes. It is found that photoelectrolysis of water and photoreduction of CO2 happen at the same time during the PEC process. Generally, PEC is performed in liquid phase solutions, either with homogeneous or with heterogeneous catalysts. Contrariwise, in electrochemical reduction of CO2 over the cathode material, the electrons and protons originate from the anode material where water is oxidized via using electric energy. A major disadvantage of the PEC process is the inevitable degradation of the photoanode and the photocathode due to their photoinstability. This disadvantage is usually cleared by using non-corroding metallic layers on the surface of the semiconductor which prevent the contact of the electrolyte and the semiconductor. Transition metal complexes of Ru, Re, bearing a large variety of geometries, have been used as homogeneous cat-

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

alysts for the CO2 reduction. Non-oxide semiconductors, such as p-GaP, n- and p- GaAs, n-Si, p-InP, GaP, CdS, SiC [28], p-CuInS2 [31] and oxide semiconductors such as WO3 , TiO2 , ZnO [28] are used as cathode materials (heterogeneous photocatalysts) and reduce CO2 to CH4 , CH3 OH, HCHO and HCOOH after several hours of illumination. In other cases, a molecular catalyst may be attached onto the surface of the photocathodic semiconductor [32]. In a PEC cell, CO2 has been reduced either by using p-InP semiconductor bearing a biological catalyst (enzyme) over its surface or by using colloidal semiconductor suspensions [24,25,28]. Gas phase PEC CO2 reduction has also been studied by using a TiO2 covered photoanode where gaseous H2 O, under light irradiation, is converted into electrons and protons which reduce gaseous CO2 on a metal loaded carbon based electrode, with high electron conversion efficiency [33]. PEC reduction of CO2 in gas-phase has several advantages such as the enhanced solubility of CO2 and the easier recovery of the products. The use of inexpensive and versatile nanocarbon materials, contributed to the enhancement of the reactivity and selectivity of the CO2 reduction because they have not only conductive properties but they also facilitate the electron mobility, the mass transfer, the dispersion of metal nanoparticles (Pt, Fe) and the formation of active hydrogen species [34]. It has also been reported the study of a solar cell driven electrochemical process, where a two-electrode system, comprising of a copper wire (cathode) and a platinum wire (anode) are involved. The CO2 reduction takes place on the cathode and CO2 exists as dilute aqueous KHCO3 solution. The production of a mixture of organic compounds, such CH3 OH, HCHO, HCOOH, CH4 , is competitive to the H2 O splitting that produces H2 . The reaction mechanism involves the excitation of the solar cell by sunlight irradiation, the photogeneration of electrons and holes, the oxidation of H2 O to O2 and H+ on the anode and the passing of the H+ through a cation exchange membrane to the cathode where H radicals are formed. When two H radicals recombine, H2 is produced via a competitive reaction to CO2 reduction [35]. Sometimes there are advantages in combining the solar water splitting process with reactions of CO2 reduction. In that way, hydrogen, produced by the water splitting, may reduce the CO2 to CH3 OH or other useful materials. For example, an electrolysis cell, with a photoelectrode which absorbs UV/visible light irradiation (photons), generates electrons and holes which electrolyse water to produce O2 and H2 . Certainly, there are several factors that must be obeyed at the same time, the band-gap energy of the photoelectrode must be low enough in order to absorb visible light irradiation and high enough in order to generate electrons capable of performing water splitting. Besides, the photoelectrode must be stable, inexpensive and resistant to corrosion. Oxides (TiO2 , CaTiO3 , SrTiO3 ), composite oxides and double layered semiconductors have been tested, but it is very challenging to satisfy all these demands [24,29]. Alternatively, electricity can be generated via the usage of solar energy, such as photovoltaics (PV) then by the usage of that electricity, water splitting is performed via an electrolysis process and the produced H2 may be used for the conversion of CO2 to useful organic compounds. This idea is theoretically promising but there are several difficulties that must be overcome, such as the high cost of the processes, the storage of H2 and the space needed for the conversion of solar energy to electricity by the use of PV [32]. Among all the above methods used for the reduction of CO2 , the most promising one seems to be the photocatalytic reduction because it is simple, friendly for the environment, nonelectricity/heat dependent and relatively costless. The most widely used photocatalyst for the CO2 reduction is TiO2 semiconductor, then, titania based photocatalysts or titanates follow [36]. The properties of TiO2 , such as the band-gap energy, were improved when it was loaded with metals or non-metals or when

153

it was modified to form for example nanotube arrays or dispersions in zeolites [26]. Indeed, functionalized titanates have been recently synthesized which exhibited rather high photocatalytic activities under visible light irradiation [37]. As Ti based photocatalysts are extensively studied, in this review, non-Ti based photocatalysts, are studied as far as it concerns their structure, composition, physical and chemical features, reactivity towards CO2 reduction and the reaction mechanisms where they participate. During the last few years, especially since 2012, a remarkable increase in the number of the published papers in the research field of the CO2 photocatalytic reduction by non-TiO2 catalysts has been remarked, as shown in Fig. 4. For that reason we believe that a wide scope review on this subject would encourage researchers to continue their efforts in this exciting and promising scientific field. The in depth study of the non-TiO2 photocalysts for the CO2 reduction is facilitated in case that they are categorized into several groups depending on their structure/chemical nature. Inorganic photocatalysts constitute the first category which is consisted of the following sub-categories of single-metal oxides, mixedmetal oxides, metal oxide composites, layered double hydroxides (LDHs) and salt composites. Carbonaceous photocatalysts constitute the second, more recent, category, which is composed of the sub-categories of graphene (GR), carbon nanotubes (CNTs) and g-C3 N4 composites. Lately, a third category that of hybrid organicinorganic photocatalytic materials appeared. In Table 1, the main groups of non-TiO2 photocatalysts for the reduction of CO2 , as well as the type of the irradiation and the experimental conditions required, are depicted. The very low sensitivity to solar irradiation which is the main disadvantage of TiO2 , was not bypassed for a large number of inorganic photocatalysts such as the single-metal oxides and mixed-metal oxides. Contrariwise, metal oxide composites, LDHs and salt composites absorb mainly in the visible part of the spectrum exhibiting rather high photocatalytic efficiencies. On the other hand, almost all the carbonaceous and hybrid photocatalytic materials absorb visible light and this fact contributes to their superiority. Additionally, they exhibit excellent electronic and physicochemical properties which facilitate the mechanism of the photocatalytic CO2 reduction, often resulting in amazing photocatalytic efficiencies. Besides, the facile synthetic processes, the low cost and the abundance of the raw materials make the preparation of carbonaceous photocatalysts more attractive and maybe the most attractive candidates for the CO2 photocatalytic reduction [38,39]. Generally, the most important factors that influence the efficiency of the photocatalytic reduction of CO2 are the degree of visible light excitation, the charge excitation and transport, the adsorption and activation of CO2 on the surface of the photocatalyst, the kinetics of the CO2 reduction and the possible undesirable reactions tha may take place during the photocatalytic reduction of CO2 , as shown in Fig. 5 [40].

3. Inorganic photocatalysts for CO2 reduction—semiconductors Semiconductors are among the first photocatalytic systems used for the performance of solar driven reactions. They possess relatively high stability and they ensure mobility of charge carriers enabling their transportation to the surface and letting the redox reaction begin. They are also inexpensive materials [41]. When semiconductors absorb light irradiation consisting of photons with energy equal or greater than their band gap, photoexcitation takes place. Electrons are excited from the valence band (VB) to the conduction band (CB) of the semiconductor leaving an unoccupied state (electron hole) in the valence band. Most of the times, elec-

154

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Fig. 4. Number of published papers per year, in the field of the photocatalytic CO2 reduction by alternative photocatalysts to TiO2 (Science Direct December 2015).

Table 1 Main categories of non-TiO2 photocatalysts for CO2 reduction. Non-TiO2 photocatalyst description

Type of irradiation

Experimental conditions

Inorganic Single-metal oxides

UV/VIS

Mixed-metal oxides Metal oxide composites Layered Double Hydroxides (LDHs)

UV/VIS VIS (mainly) UV/VIS

Salt composites

UV/VIS (mainly)

Solid-gas (catalyst-CO2 /H2 O, catalyst-CO2 /H2 , catalyst-CO2 /CH4 ) systems/Aqueous dispersions Solid-gas (catalyst-CO2 /H2 O) systems/Aqueous dispersions Solid-gas (catalyst-CO2 /H2 O) systems/Aqueous dispersions Solid-gas (catalyst-CO2 /H2 O, catalyst-CO2 /H2 ) systems/Aqueous dispersions Aqueous/organic solutions

Carbonaceous photocatalysts Composite photocatalysts based on graphene (GR) Composite photocatalysts based on carbon nanotubes (CNTs) g-C3 N4 and composite photocatalysts based on g-C3 N4

VIS VIS VIS

Aqueous dispersions Aqueous dispersions Aqueous dispersions/Solid-gas (catalyst-CO2 /H2 O) systems

VIS

Aqueous/organic solutions

Hybrid organic-inorganic materials-Zeolitic imidazolate frameworks-ZIFs)

trons and holes recombine to release the absorbed energy as heat or rarely as light and that’s the reason behind the very low photocatalytic efficiency. When they don’t recombine, they migrate to the surface of the semiconductor where they are trapped by the adsorbed reactants and the photocatalytic redox reactions begin. Conduction band electrons reduce adsorbed species on the surface of the catalyst. On the other hand, valence band holes are oxidants, evoking direct oxidation of the adsorbed species or indirect oxidation via surface-bound hydroxyl radicals formed by holes trapped at the semiconductor surface [38,42–45]. CO2 is a very stable molecule as it constitutes the combustion final product of all the carboneous fuels. The conversion process to potential fuels is endothermic and therefore demands very high sums of energy [46]. The photochemical reaction that takes place during the artificial photosynthesis is shown below: CO2 + H2 O(orH2 ) + photocatalyst + h␯ → Carbonaceous products + O2 [16]

Because of the thermodynamical inertness of the CO2 molecule, the electrons reaching the semiconductor surface are not all capable of reducing CO2 . The reducing potential of a species is a measure of its reducing power, and the higher (more negative) it is the more reductive the species is. Therefore, photoexcited electrons must have higher reduction potentials (more negative) than water reduction producing H2 . Otherwise H2 will be produced from water reduction which is a competitive reaction to the CO2 reduction. Band levels of some simple and widely used semiconductors are shown in Fig. 6 [46]. Furthermore, the reaction mechanism of CO2 reduction must have multiple stages during which electrons and protons are transferred. In case of a single electron transfer to CO2 in order to produce CO2 − , the reduction potential is too high (E◦ red = −1.9 V) making the progress of the photocatalytic reaction very difficult. If a proton is simultaneously transferred to the CO2 − , the radical anion of CO2 −* is produced, which is stabilized reducing the reduction potential (E◦ red = −1.23 V) and therefore making easier the progress

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

155

Fig. 7. Diagram of photoexcitation and electron transfer process [47].

Fig. 5. Main factors influencing the photocatalytic efficiency of the reduction of CO2 and strategies for its enhancement [40].

Fig. 6. Band levels of simple semiconductors [46].

of the reaction. Generally, when protons and electrons are transferred simultaneously in many small, rapid successive steps the CO2 photoreduction is facilitated [47,48]. Depending on the number of electrons used, the reaction product can be methane, methanol, formaldehyde, carbon monoxide or formic acid. In Fig. 7 the photoexcitation and electron transfer process for the production of CO is pictured [47]. The main reactions followed and values of reduction potentials at pH = 7 during the artificial photosynthesis are shown below: CO2 + 2e− → CO2 −∗

E◦ red = −1.90 V(VvsNHE)

CO2 + 2e− + 2H+ → HCOOH

E◦ red = −0.61 V(VvsNHE)

CO2 + 2e− + 2H+ → CO + H2 O

E◦ red = −0.53 V(VvsNHE)

CO2 + 4e− + 4H+ → HCHO + H2 O

E◦ red = −0.48 V(VvsNHE)

CO2 + 6e− + 6H+ → CH3 OH + H2 O CO2 + 8e− + 8H+ → CH4 + H2 O

E◦ red = −0.38 V(VvsNHE)

E◦ red = −0.24 V(VvsNHE)

2e− + 2H+ → H2

E◦ red = −0.42 V(VvsNHE)[46, 47]

Although CO2 photoreduction is a very challenging process, there are serious matters that hinder its worldwide use at this time. The low efficiency of the CO2 photocatalytic reduction is the major problem, which is ascribed mainly to low visible light absorption of the semiconductors. The band gap energy of a semiconductor determines which part of sunlight spectrum is absorbed. The wider the band gap energy, the shorter the sunlight wavelengths absorbed. When the band gap energy of a semiconductor is relatively high, then only the ultraviolet part of the sunlight spectrum is absorbed, meaning that only about 5% of the solar radiation reaching the Earth’s surface can be utilized and the remaining 95% remains unexploited. As a result, low efficiencies for the photocatalytic reactions are observed. Finding ways to reduce the band gap of a semiconductor is of great importance because it would allow us to take advantage of the visible part, which comprises almost half of the solar spectrum [47,49]. Usually metal ions (ferric, tangstate, etc.), or non metal ions (carbon, nitrogen etc.) doping of semiconductors has a positive effect on the photoactivity in the visible part of the solar spectrum. Besides, it is well accepted that noble metals (Au, Ag, Pd, Pt) doping contributes to the capture of CB electrons on the surface of the photocatalysts and thus lowers the possibility of recombination of the photogenerated electrons/holes; additionally it reduces the activation energy for the photocatalytic reduction of CO2 [50]. By the term “doping” is meant the assimilation of atoms or ions into the bulk of a semiconductor crystallite without altering its surface [38]. Doping must be performed at low concentrations in order to avoid recombination of holes and electrons which is fostered by the presence of metal ions. Furthermore, dopants must not distort the semiconductor lattices, so it is very important for their concentrations not to exceed the allowed limits during semiconductor synthesis. If dopants are found in the appropriate proportion into the semiconductor, their photoactivity is significantly improved. Photosensitizers, such as organic dyes or quantum dots (QDs), combined with semiconductors, improve the photocatalytic efficiency because they increase the absorption at the visible part of the spectrum. When a semiconductor absorbs visible radiation, photoexcited electrons transit from the dye to the conduction band and therefore facilitate the progress of the photocatalytic reaction. Besides, the introduction of structural defects and the surface plasmon resonance effect (SPR) increase the visible light excitation [40,43,46,47].

156

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

One other way to increase the photocatalytic efficiency of the CO2 reduction is to couple two semiconductors together providing that their reduction potentials are related and that the semiconductors interact adequately with each other. The result of this coupling is the physical separation of charge carriers following their generation and the hampering of their recombination. Furthermore, the semiconductor with the shorter band gap can act as a photosensitizer making the composite material absorbing visible light. [43,51]. It is established that bulk semiconductor photocatalysts don’t exhibit the maximum of their reactivity in the CO2 reduction, whereas, when they form nanoparticles, a significant rise to their reactivity is observed. Particle size of the photocatalyst has an impact on the photocatalytic reactivity, but not so significant. It has been shown that the particle size of a semiconductor affects the band gap energy. The smaller the particle is, the wider the band gap becomes, altering the values of reduction potentials and thus the reduction and oxidizing power. Besides, particle size affects the tendency for recombination of separated species upon their generation. The increase of the total surface area per volume of a semiconductor also increases its photocatalytic reactivity by increasing the number of reaction sites. Besides, the decrease of the particle size leads to a considerable increase in the affinity of a semiconductor to adsorb CO2 and thus enhances the photocatalytic efficiency [46,52]. Plenty of different morphologies of nanoparticles can be achieved depending on reaction conditions during the semiconductor synthesis process. It has been shown that the morphological and structural characteristics of a nanoparticle affect its photocatalytic reactivity. The separation of photoexcited electron-hole pairs and their transfer to the surface of a semiconductor is essentially assisted when hierarchical heterostructures are used. Spherical nanoparticles (zero dimensional) have the advantage of a high ratio surface/volume, so charge carriers have a short distance to traverse before reaching the surface. When different morphologies are combined, they may lead to the formation of a hierarchical heterostructure with upgraded photocatalytic efficiencies. For example when spherical nanoparticles (ZnO and Al2 O3 ) are combined with CdS nanowires, photocatalytic reactivity of the hydrogen generation reaction is enhanced because of the resulting high surface area and the efficient charge separation. Nanostructures can also assume the form of wires, tubes and belts (one dimensional) differentiating themselves from the bulk semiconductors. It has been found that as the length of a nanoparticle increases the same happens to its photocatalytic efficiency. Besides, hydrothermal and/or calcination processes usually improve the photocatalytic reactivity of heterostructures owing to increased porosity, chemical or thermal stability. Other morphologies include the two dimensional structures with a platelet shape and the placing of chemical particles on them. In this way, high surface area and remarkable electric properties are ensured for the semiconductor. Graphene or graphene-based composites are usually used as platelet supports. Three dimensional mesoporous structures have also been mentioned. For example, In(OH)3 with mesoporous structure exhibited almost 20 times higher photocatalytic efficiency towards the reduction of CO2 (0.8 ␮mol CH4/ g* h) than that of nIn(OH)3 (0.04 ␮mol CH4/ g* h) [53]. The 3D mesoporous structures possess multidimensional domains at different levels and complex/fascinating pore structures. In this case, the large number of multi-sized pores ensures the large specific area and the high rate of transfer for the species. Besides, the CO2 molecules easily diffuse into the reaction sites and the products are easily removed from them. As a consequence, the photocatalytic reactivity of the semiconductor is highly enhanced. Some typical three dimensional structures (hierarchical nanostrucrures) are the urchin-like,

Fig. 8. Mechanism for the photoreduction of CO2 with H2 O over the ordered mesoporous Fe doped CeO2 [59].

brush-like, flower-like, tree-like, dendritic and branched structures [38,51,54]. 3.1. Single-metal oxides The first photocatalytic systems, studied for the CO2 reduction were the inorganic single-metal oxide semiconductors. The photogenerated electrons are transferred from the VB to the CB, forming electron/hole pairs (e− /h+ ) which further reduce/oxidize compounds adsorbed on the surface of the semiconductor [52]. The most popular semiconductor is TiO2 , pure or modified, alone or in composites. TiO2 semiconductors exhibit some great advantages such as high stability, low cost and reduced toxicity [55]. Additionally, TiO2 exhibits enhanced photocatalytic efficiencies during photocatalytic reactions such as water splitting, detoxification etc. Nevertheless, TiO2 exhibits a major disadvantage which is the limited sensitivity to sunlight. Absorbing only the ultraviolet part of the electromagnetic solar radiation means that TiO2 absorbs only 5% of the solar energy reaching the Earth’s surface [38,43]. In order to extend the absorbance to the visible part of the spectrum several modifications of TiO2 photocatalysts have been realized as well as a lot of other single-metal oxides, with lower values of bandgap energy, have been prepared. Certainly, common single-metal oxides, such as WO3 , were found rather insufficient for the photocatalytic reduction of CO2 , due to their small value of band-gap energy which favored the recombination of the photogenerated species and their low CB edge potential [56]. 3.1.1. Solid-gas systems 3.1.1.1. H2 O as the reducing agent. In many cases, the photocatalyst in the solid state reacted with a gas mixture of the reactants. Pb3 O4 , BiOI, CeO2 reacted with a gas mixture of CO2 /H2 O, where water was the reducing compound. In the case of hydrated Pb3 O4 (hematite), the enhanced number of water molecules on its surface was crucial. It was found that the increased transfer of oxygen atoms from water molecules to the oxygen containing products, such as methanol, resulted in the increase in the amount of products (2.6 ␮mol CH3 OH/g* h/) [57]. Hierarchical structures of BiOI produced CH4 (1.58 ␮mol/g* 8 h) when they reacted with H2 O, whereas the amount of the photocatalytic product was almost the half (0.68 ␮mol/g* 8 h) when BiOI was in the bulk form [58]. Mesoporous Fe/CeO2 produced increased amounts of CO (74.3 ␮mol/g* h) and CH4 (17.3 ␮mol/g* h) because of the Fe doping which formed an extra dopant energy band and the enhanced number of chemisorbed oxygen species on its surface, as depicted in Fig. 8. The reaction mechanism involves the CO2 adsorption on the surface of the catalyst followed by the reaction of CO2 with pho-

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

togenerated electrons via the formation of free radicals CO2 −* [59]. ZnO is the most widely studied single-metal oxide photocatalyst for the reduction of CO2 . It consists the TiO2 analogue, possesses similar values of bang-gap energy with that, it is a low-cost product and it has not been regarded as a very environmentally harmful compound [60–62]. In the early 90s, ZnO in a clathrate hydrate phase reacted with CO2 gas to find that the presence of impurities such as S, Si and P atoms, lowered the excitation energy of electrons and holes and therefore shifted the absorption of ZnO to the visible part of the spectrum [63]. Studies upon Cu and N doped mesoporous ZnO materials showed that the higher value of surface area of the samples prepared by the hydrothermal method contributed to the increased photocatalytic activity compared to those prepared by other methods (SC). The insertion of Cu into the mesoporous ZnO also controlled the selectivity and increased the photocatalytic activity since its dispersion and interaction with ZnO was increased along with the absorption in the visible part of the spectrum. Still, the optimum photocatalytic efficiencies were not rather impressive (1 ␮mol CO/g* h) [60]. Sharifnia et al. synthesized calcinated ZnO immobilized on stainless steel mesh. Calcination led to the decrease of ZnO particles size, resulting in specific surface area increase and agglomeration decrease. Calcination also improved the absorbance of photocatalysts in the UV–vis region of the spectrum and shifted the absorbance slightly towards the visible area. Stainless steel mesh was selected as a support for ZnO, owing to its large specific surface area, resistance to corrosions, enhanced absorption of UV light irradiation and sufficient ventilation for gases transport [64]. Eventually, the conversion percentage of CO2 to a group of oxygenated products, on the surface of the calcinated ZnO immobilized on stainless steel mesh was satisfactory (9% after 5 h). Bartoszek et al. carried out experiments where moistened hematite photocatalysed the CO2 reduction when its surface came in contact with a mixture of CO2 -H2 O in gas phase, giving an explanation on the origin of methane in the atmosphere of Mars, which had puzzled scientists. The role of water was found to be very important; the higher the water on hematite surface, the higher was the formation of oxygenates (methanol, acetone, formaldehyde, acetaldehyde). Generally, oxygenated molecules bearing two or three carbon atoms, like acetone, might be produced via various mechanisms which include photoinduced radical reactions (such as the recombination of radicals from methanol and aldehydes) [57,65]. Recently, a 5% CdS/WO3 photocatalyst was synthesized and its photocatalytic activity for the CO2 reduction under visible light irradiation was found 1.02 ␮mol CH4 /g* h, actually not rather high. It was proved that CdS/WO3 heterostructure samples form a hierarchical Z-Scheme which contributes to the enhancement of the photocatalytic activity, compared to that of the bare samples, owing to the efficient spatial separation of the photogenerated electron/hole pairs. CdS nanoparticles were uniformly and intimately dispersed on the surface of WO3 nanospheres and the reduction reactions of CO2 took place on the CdS surface whereas the oxidation reactions took place on the WO3 surface [66].

3.1.1.2. H2 as the reducing agent. When the photocatalyst in the solid state reacted with a gas mixture of CO2 and H2 , the CO2 reduction proceeded via the dissociative adsorption of H2 on its surface. Ga2 O3 produced CO (3.6 ␮mol/g* 5 h) upon reacting with CO2 molecules, whereas MgO and ZrO2 formed slightly higher amounts of CO [67,68]. Calcinated ZnO immobilized on stainless steel mesh reacted with a CO2 /H2 gas mixture but the conversion percentage (3% after 5 h) was found to be lower compared to that obtained when gas mixtures of CO2 /H2 O (9% after 5 h) or CO2 /CH4 (11% after 5 h) were used [64]. Hydrogenation of CO2 on the surface of NiO, via a dissociative hydrogen adsorption on oxygen rich sites (surface defects), resulted in a boost to the photocatalytic activity, which

157

was further enhanced when the calcination process took place at temperatures below 400 ◦ C [69]. 3.1.1.3. CH4 as the reducing agent. In the rare case that CH4 is the reducing compound, the reaction mechanism involves the formation of free radicals CH3 * , H* , CO2−* and the reaction among them to produce mainly oxygenated compounds such as formates and acetates. The solid photocatalyst (ZrO2 ) reacted with a gas mixture of CO2 and CH4 to form little amounts of CO (almost 1 ␮mol CO/g* 6 h) [68]. In addition, calcinated ZnO immobilized on stainless steel mesh reacted with a CO2 /CH4 gas mixture displaying satisfactory conversion percentages of CO2 (11% after 5 h) [64,70]. Other studies confirmed the effect of the amount of ZnO coated on mesh on the photocatalytic activity, as it has impact on the active surface of the catalyst [70]. In Fig. 9, SEM images of ZnO photocatalysts coated on mesh are shown [70]. Moreover, efficient photosensitizers such as metallophthalocyanines (MPc), were added to ZnO coated on stainless steel mesh in order to extend the absorbance of ZnO into the visible part of the spectrum and thus enhance the photocatalytic activity (optimum conversion percentages of CO2 :23% after 5 h) [71]. 3.1.2. Aqueous dispersions The photocatalysts were dispersed in water or aqueous solutions and they reacted with CO2 in the gas phase. Interesting heterostructures, consisted of common cations, such as ZnTe/ZnO, synthesized by hydrothermal methods, produced different amounts of CH4 , depending on the ZnO nanostructures (optimum efficiency: 100 ␮mol CH4 /g* 5 h). The formation of heterojunctions via an intimate interfacing of ZnTe and ZnO, enhanced the rate of the electron transfer between ZnTe and ZnO compared to the recombination rate of the photogenerated species, and thus increased the photocatalytic activity [72]. NiO dispersed in water was found to be more efficient photocatalyst than other singlemetal oxides dispersed in aqueous solutions, such as ZnO, WO3 , Cu2 O, CuO. The enhanced photocatalytic activity of NiO (170 ␮mol CH3 OH/g* h), was attributed to the surface defects. Namely, vacant oxygen spaces on the NiO surface strengthened the interactions with CO2 molecules and facilitated the transportation of the photogenerated species [73,74]. Additionally, Yahaya et al. concluded that the CO2 photoreduction to CH3 OH does not proceed via a single mechanism, as happens in the case of the reduction of free CO2 . On the contrary, H2 CO3 formed by the CO2 dissolution in water is responsible for the mechanisms such as the one in the H2 CO3 and carbonate ions reduction [57]. Peiqiang et al. demonstrated the excellent photocatalytic property (1.8 mmol CH3 OH/L* m2* 6 h) of one dimensional wedged N/CuO owing to the N doping which contributed to a red shift at the absorption of light as it introduced an intermediate band between the VB and the CB of CuO and increased the carrier mobility [75]. Aqueous suspensions of hydrous cuprous oxide (Cu2 O·x H2 O) exhibited satisfying photocatalytic efficiency (24 ␮mol CH3 OH/L* 20 mg). Tennakone et al. presumed that quick removal of the holes in the Cu2 O·xH2 O, due to their reaction with OH− ions found on the semiconductor’s surface, is responsible for the high density of electrons and therefore the preference for multielectron processes [44]. 3.2. Mixed-metal oxides Mixed metal oxides, consisting of two or more kinds of metals and oxygen, have been widely used as photocatalysts for the CO2 reduction. They possess semiconducting properties and their aqueous suspensions, irradiated by visible light, have been mainly studied. In most cases water plays the role of the reducing agent, and it is oxidized to O2 and H+ ions. The latter, react directly with CO2 molecules and photogenerated electrons or with CO2 −* radi-

158

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Fig. 9. SEM images of photocatalysts coated on mesh: (a) uncoated mesh; (b) 5 g; (c) 6.5 g; (d) 8 g [70].

cals, in order to produce appropriate products, via many different paths, depending on the values of the reduction potential of the mixed-metal oxides.

3.2.1. Solid-gas systems 3.2.1.1. H2 O as the reducing agent. Solid-gas systems where the mixed-metal oxide photocatalyst is found in the solid state and the reaction mixture of CO2 /H2 O is found in the gas phase have been studied in a small extend. Concisely, solid Zn2 GeO4 , KNbO3 , HNbO3 and Bi2 WO6 reacted with CO2 and H2 O vapor via the oxidation of water molecules to O2 and H+ ions and the reduction of CO2 by the H+ ions and the photogenerated electrons, as it has already been referred when water plays the role of the reducing compound. Bulk Zn2 GeO4 , exhibited very small photocatalytic activity (<1 ␮mol CH4 /g* h) whereas nanoribbons of Zn2 GeO4 were found to be much more reactive photocatalysts (1.5 ␮mol CH4 /g* h, 6 ␮mol CH4 /g* 14 h). The superiority of the latter could be attributed to many factors, such as their enhanced specific surface area (28.27 m2 /g vs 0.75 m2 /g), their extremely good crystal quality, their increased lengths and ultrathin geometry which promote the separation of photoinduced species, prevent the recombination of them and promote the rapid transportation of the charge carriers. It was also demonstrated that the loading of Pt (1% wt) or RuO2 (1% wt) or even better the co-loading of Pt and RuO2 on the Zn2 GeO4 nanoribbon surface, further increased the rate of CH4 formation (25 ␮mol CH4 /g* h, 100 ␮mol CH4 /g* 16 h) [76]. Significantly lower photocatalytic efficiencies were observed for KNb3 O8 , HNb3 O8 and Bi2 WO6 . Certainly in case that they did not have bulk structures and they formed nanobelts and nanoplates, their activities were importantly enhanced (3.58 ␮mol CH4 /g* h, 1.71 ␮mol CH4 /g* h and 1.1 ␮mol CH4 /g* h, for HNbO3 , KNbO3 and Bi2 WO6 , respectively). The formation of hydrogen bonds between water molecules and the H on the HNb3 O8 surface deters the recombination of photoinduced electrons and holes because they are trapped on the interlayer surface and therefore higher values of photocatalytic effi-

ciency are achieved [77–79]. Sun et al. studied the photocatalytic activity of ball-flower-like Bi2 WO6 , prepared by a hydrothermal method without any organic precursor, under visible light. It was found that the production of CO was hindered when the water vapor was increased, due to the competitive absorption of CO2 and H2 O on the basic sites of Bi2 WO6 . The highest photocatalytic activity (333 nmol CO/g* h) was observed in the condition with little H2 O vapor, owing to the increased crystallinity, reduced recombination rate of photogenerated species and increased stability of basic sites on the surface of the photocatalyst, whereas the regeneration of the catalyst by water washing, was found to be rather effective [80]. 3.2.1.2. H2 as the reducing agent. Gas-solid systems, where the solid photocatalyst react with a gas mixture of CO2 and H2 via chemisorption of CO2 molecules onto its surface, have been sparsely studied. Generally, the decisive step of the reaction mechanism is the CO2 chemisorption, which is favored when the catalyst possesses basic sites which facilitate the absorption of CO2 by altering its linear conformation and assisting the reduction by hydrogen. As the degree of CO2 absorption enhances, the photocatalytic activity increases too. LiTaO3 exhibited low photocatalytic activity (0.42 ␮mol CO/g* 24 h), significantly lower than that of single-metal oxides such as Ga2 O3 , MgO and ZrO2 , which bear a larger number of basic sites on their surface [81]. 3.2.2. Aqueous dispersions The following mixed-metal oxides: CaFe2 O4 , BiVO4 , LaCoO3 , NaBiO3 , NaNbO3 , KNbO3 , CuGaO2 , CuAlGaO4 , Bi2 WO6 and BaCeO3 were studied either in contact with small amounts of water, or dispersed in aqueous solutions. Aqueous suspensions of ptype CaFe2 O4 , containing NaH2 PO2 and BaCO3 , saturated with CO2 gas, produced mainly methanol and formaldehyde (2 ␮mol CH3 OH/g* 4 h and 3 ␮mol HCHO/g* 4 h) via a direct mechanism which did not involve the formation of CO2 −* radicals. It was suggested that the photocatalytic reduction of CO2 took place at the

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

interface constituted of solid CaFe2 O4 , solid BaCO3 and water where CO2 and CO3 2− were more reactive [82]. Aqueous NaOH suspensions of BiVO4 produced methanol or ethanol, depending on the available amount of CO2 . The determinant role of the crystal structure to the photocatalytic activity was studied, to find that in monoclinic BiVO4 (30 ␮mol ethanol/g* 80 min) there is a higher tendency to form Bi· · ·O bonds with CO3 −2 ions than in tetragonal BiVO4 (2 ␮mol ethanol/g* 80 min) and thus the transfer of photogenerated electrons from vanadium to CO3 −2 is facilitated. The important role of the NaOH solution was partly attributed to its action as a hole scavenger which helps the separation of photoinduced species, and to its ability to dissolve more CO2 than water [83]. Aqueous NaOH suspensions of lamellar BiVO4 exhibited satisfying photocatalytic activity in methanol (35 ␮mol methanol/g* 6 h) as in the current study, the available amount of CO2 was relatively lower than in case of monoclinic/tetragonal BiVO4 and the formation of one carbon atom products was favored [83,84]. A reaction mechanism, which involves the oxidation of water molecules to O2 by the photogenerated holes and the reduction of CO2 to methanol by the photogenerated electrons on the surface of the lamellar BiVO4 was proposed [84]. The effect of the preparation method of NaBiO3 to the photocatalytic reduction of CO2 was found to be crucial when NaBiO3 nanopowders were synthesized through hydrothermal, sol-gel and dehydration methods. The diversification in the crystallinity, the specific surface area and the size and shape of the particles produced via each method resulted in different photocatalytic activities. Namely, the photocatalytic efficiency of NaBiO3 samples synthesized by the sol-gel method was several times higher than those prepared by other methods, owing to the smaller size of their crystallites, the higher value of specific area and the lower band gap energy (sol-gel: 0.6 ␮mol CH3 OH/g* h, hydrothermal: 0.37 ␮mol CH3 OH/g* h, dehydration: 0.037 ␮mol CH3 OH/g * h) [85]. Not really hopeful photocatalytic efficiencies were observed for NaNbO3 and KNbO3 , by Shi et al. (2.3 ppm CH4 /h and 7 ppm CH4 /h). However it was reported that possibly the low photocatalytic activities are attributed to the small values of specific surface area of the samples and therefore they could be easily improved [86]. Besides, Shi et al. demonstrated that NaNbO3 nanowires, prepared by a combination of hydrothermal and heating processes, exhibited tenfold increase in the photocatalytic activity (1600 ppm CH4 /g* 150 min) compared to that of bulk NaNbO3 , prepared by SSR (140 ppm CH4 /g* 150 min) [87]. In separate studies, the effect of the crystal structure and Pt doping on the photocatalytic activity, were studied concurrently. The superiority of cubic Pt/NaNbO3 (1.3 ␮mol CH4 /g* h and 4.2 ␮mol CH4 /g* h) over the orthogonal Pt/NaNbO3 (0.5 ␮mol CH4 /g* h and 1.5 ␮mol CH4 /g* h) was obvious. Apparently, the transfer of the photogenerated electrons is facilitated in the cubic crystal lattice, owing to its differentiated electronic structure [88]. Lekse et al. studied the photocatalytic activity of delafossites ABO2 (where A is a metal such as Cu, Ag, Pt, Pd, B and B is a metal such as Al, Ga, Y, Fe) which have hexagonal, rhombohedral or randomly substituted rhombohydral structure and consist of alternating layers of octahedrally B metal atoms spaced by two coordinated A metal atoms. Namely, CuGaO2 , even if it was doped with Fe, exhibited low photocatalytic activity (9 ppm CO/g* h), as the only effect of Fe doping was the shifted the absorption towards higher wavelengths (lower values of band-gap energy) [89]. On the contrary, aqueous acidic solution with Fe+2 ions (pH = 2.6) of cube-like shaped Pt/CuAlGaO4 , synthesized via the solid state fusion method, exhibited improved photocatalytic efficiency (8 ␮mol CH3 OH/g* h), when they were put in a novel twin reactor and reacted with a mixture of CO/CO2 . The singularity of the twin reactor was the division of H2 and O2 producing photocatalysts into two compartments separated by a membrane. As a result, in the one compartment H2 was generated and simultaneously reacted with CO2 to produce mainly methanol, whereas in the other

159

compartment O2 was evolved [90]. LaCoO3 displayed notable photocatalytic efficiency (40 ␮mol HCOOH/g* h), which was found to be tripled in case of the synergistic doping with C and Fe (130 ␮mol HCOOH/g* h) [91]. The modification of Bi2 WO6 microspheres with conducting polymers (PPy, PANI, PTh) increased the absorption in the visible area of the spectrum, decreased the recombination rate of the photogenerated electron-hole pairs and facilitated the charge transfer efficiency, resulting in enhancement of the photocatalytic activity by 2.8 times, compared to that of bare Bi2 WO6 . As the CB of Bi2 WO6 is found too low for the CO2 photoreduction to gas products such as CH4 and CO, the main products detected under visible light irradiation, were methanol (56.5 ␮mol CH3 OH/g* 4 h) and ethanol (20.5 ␮mol CH3 CH2 OH/g* 4 h) [92]. Various cocatalyst nanoparticles (Ag, Au, Pt, CuO and RuO2 ) enhanced the photocatalytic activity of perovskite-type BaCeO3 for the CO2 reduction, under UV light irradiation. Especially, 0.3%Ag/BaCeO3 was found to be the more effective photocatalyst among the others, but generally not rather effective (4 ␮mol CH4 /g* 7 h). The cubic crystal structure of Ag nanoparticles found on the surface of BaCeO3 promoted the close contact between Ag and BaCeO3 leading to the formation of an active center which lowered the activation energy of the reaction and facilitated the separation of the photogenerated species [93]. 3.3. Metal oxide composites The combination of two metal oxides in order to prepare photocatalysts with ameliorated photocatalytic efficiency for the reduction of CO2 has also been attempted. NiO consists one of the most widely studied metal oxides cocatalysts. In the crushing majority, the reaction mixture, either in the gas or liquid phase, composed of CO2 and H2 O. The main products of the photocatalytic reduction of CO2 with H2 O, under visible light irradiation, were CH3 OH ´ CH4 . 3.3.1. Solid-gas systems 3.3.1.1. H2 O as the reducing agent. Solid NiO/InTaO4 photocatalyst has reacted with a gas mixture of gaseous CO2 and water vapors in an optical fiber photoreactor, composed of assembled optical fibers with pure silica core and a polymeric shield (optimal efficiency: 21.0 ␮mol CH3 OH/g* h at 75 ◦ C) [94] as well as in a monolith photoreactor with internal channels covered with a layer of the catalyst, over a lower SiO2 sublayer (0.16 ␮mol CH3 OH/g* h) [95]. On the other hand, multifunctional Pt (0.5%wt)/ZnAl2 O4 /ZnGaNO with mesoporous structure, small band-gap energy, enhanced gas adsorption capacity and basicity, owing to the presence of ZnAl2 O4 , increased the light absorbance and assisted the fast migration of the photogenerated species, resulting in satisfying photocatalytic activity (38 ␮mol CH4 /g* 8 h) [96]. 3.3.2. Aqueous dispersions The effect of the nature of the supporting materials (basic or acidic) on the selectivity and the photocatalytic activity has been analyzed for CuO/ZnO supported on MgO (8 ␮mol CH4 /g* h and 0.8 ␮mol C2 H6 /g* h). It was deduced that basic supporting materials interact more thoroughly with the reactive site of the semiconductor and thus favor the formation of dimerized products [97]. Potassium bicarbonate (0.2 M) aqueous suspensions of NiO(1%wt)/InTaO samples, where NiO particles were randomly dispersed in InTaO4 agglomerates (1–2 ␮m), exhibited mediocre photocatalytic activity (1.394 ␮mol CH3 OH/g* h). NiO acted as an electron acceptor and so improved the photocatalytic activity compared to that of bare InTaO4 samples. Methanol was produced via 6 electron reduction processes where CO2 or H2 CO3 or CO3 −2 etc. were reduced by InTaO4 because of the more negative reduction potential of the conduction band of InTaO4 compared to

160

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

those of methanol yield [98]. Wu et al. demonstrated the decisive role of the photoreactor in the photocatalytic activity. NiO/InTaO4 aqueous suspensions, exhibited almost ten times lower activity (2.8 ␮mol CH3 OH/g* h) when they were put in a quartz reactor compared to that obtained when NiO/InTaO4 reacted in an optical fiber reactor (21 ␮mol CH3 OH/g* h), even if in both cases the photocatalsyts were synthesized by the same method (sol-gel). The photoreaction mechanism involves electron transportation from the valence band of InTaO4 to the NiO cluster found on its surface. Adsorbed hydrogen atoms – produced from H2 O dissociation into H+ and OH− – and electrons on NiO, initially react with CO2 and then via the formation of HCOO, H2 COO, H2 CO and CH3 O, methanol is finally formed. Holes found in the valence band of InTaO4 react with OH− ions and via the formation of OH* radicals, H2 O2 and O2 − generate O2 [94]. In order to decrease the band gap of InTaO4 , it was dopped with N, as it has been demonstrated that when the valence band constitutes of both N 2p and O 2p orbitals it is shifted upwards and therefore the band gap is shortened. Regarding the addition of NiO, it has been confirmed that it is an efficient cocatalyst because it acts as an electron acceptor and thus prevents the electron-hole recombination. However, the mismatching of the conduction bands of InTaO4 and NiO prevents the electron transfer from InTaO4 to NiO. Liu et al. used a Ni@NiO core-shell structure where the metallic Ni layer facilitated the transportation of photoinduced electrons from InTaO4 valence band to the NiO surface and thus enhanced the photocatalytic efficiency. The morphological analysis of Ni@NiO/N-InTaO4 showed that nanoparticles of the cocatalyst (Ni@NiO) were randomly deposited over the InTaO4 surface. It was found that InTaO4 was the least efficient photocatalyst with the lowest rate of CH3 OH formation (50 ␮mol CH3 OH/g* h), N-InTaO4 was more efficient (130 ␮mol CH3 OH/g* h) and Ni@NiO/N-InTaO4 was the most efficient photocatalyst (170 ␮mol CH3 OH/g* h). That was expected because N doping effectively decreased the band gap, enhanced the absorbance of visible light irradiation. Besides, Ni@NiO coreshell structure, due its higher electronegativity compared to that of Ta, assisted the migration of electrons from the InTaO4 surface to the NiO nanoparticles, where the reduction of the carbon source (CO2 , H2 CO3 and CO3 −2 ) began [99]. Aqueous potassium bicarbonate suspensions of NiO/InNbO4 and Co3 O4 /InNbO4 exhibited similar but rather low photocatalytic activities (1.6 ␮mol CH3 OH/g* h and 1.5 ␮mol CH3 OH/g* h, respectively). Still, they were slightly higher than those of pure InNbO4 [100]. More recent studies on the photocatalytic activity of aqueous KHCO3 suspensions of NiO/InVO4 demonstrated that the generation of little holes (surface defects) on the InVO4 surface after the deposition of NiO, resulted in the formation of more active sites and thus higher photocatalytic activity (1.102 ␮mol/g* h for InVO4 vs 1.526 ␮mol CH3 OH/g* h for NiO/InVO4 ). Additionally, the photocatalytic efficiency was enhanced because NiO/InVO4 absorbs more intensely in the visible part of the spectrum, owing to the formation of a subband on the upper end of the VB [101]. Studies of the photocatalytic activity of aqueous suspensions of Ni/Ni3 (BO3 )2 /NiO nanoparticles (0.41 ␮mol CH4 /g*h, 4 ␮mol CH4 /g*10 h) demonstrated that it is importantly enhanced compared to that of Ni3 (BO3 )2 (0.14 ␮mol CH4 /g*h, 1.5 ␮mol CH4 /g*10 h) and Ni3 (BO3 )2 /NiO (0.16 ␮mol CH4 /g* h, 1.6 ␮mol CH4 /g*10 h), owing to the presence of Ni which facilitated the separation of the photogenerated species and delay their recombination. Namely, the reaction mechanism of CO2 reduction on the surface of Ni/Ni3 (BO3 )2 /NiO heterostructures (shown in Fig. 10), involved (a) the transfer of the photogenerated electrons from the CB of NiO to the CB of Ni3 (BO3 )2 and then to the Fermi level of Ni, (b) the transfer of the photogenerated holes from the VB of Ni3 (BO3 )2 to the VB of NiO where they oxidize water molecules to O2 and H+ ions and c) the reduction of CO2 molecules by H+ ions and photogenerated electrons found on the

Fig. 10. Proposed mechanism for the photocatalytic reduction of CO2 on the surface of Ni/Ni3 (BO3 )2 /NiO, under visible light irradiation [102].

Fermi level of Ni, to produce CH4 . The formation of CH4 was favored as the Fermi level of Ni constitutes an electron rich region where the performance of reactions requiring a large number of electrons is promoted [102].

3.4. Layered double hydroxides (LDHs) Layered double hydroxides (LDHs) are thought to be compounds with the layered structure of brucite (Mg(OH)2 ), with a hexagonal crystal structure. They are constituted of positively charged sheets with the general type [MII 1-x MIII x (OH)2 ]x+ which indicates that some trivalent cations have taken the place of the divalent cations. Anions are placed between two cationic layers in order to ensure the neutral charge of the LDH. Besides, water is present between the cationic layers. It is confirmed that the mass of water present in a LDH is almost 50% of the total mass of cations forming the positively charged layers. MII can be one of the following metals: Mg, Mn, Fe, Co, Ni, Cu, Zn, MIII can be one of the followings Al, Cr, Mn, Fe, Ga and the value of x usually varies between 0.17–0.33. Intercalated anions can be the followings: CO3 2− , SO4 2− , NO3 − , Cl− , OH− . LDHs are efficient photocatalysts for the CO2 reduction because they exhibit a great sorption capacity for CO2 in the space between the cationic sheets. On the other hand, Teramura et al. and Tanaka et al. demonstrated that simple mixtures of the same metal hydroxides exhibit significantly lower photocatalytic activity. This fact indicates that the unique structure of LDHs facilitates the CO2 dissolution in the LDH interlayers and therefore its reaction with surrounded water molecules to produce CO and O2 via the formation of photogenerated pairs of electrons and holes. Besides, the choice of the cations present in the LDH influences its semiconductor properties, the kind of the products (usually CO and/or CH3 OH) and therefore its photocatalytic activitiy. LDHs constitute the only layered materials with photocatalytic properties which are formed by positively charged sheets, in contrast to the montmorillonites, smectites and kaolinites which are formed by negatively charged layers. In fact LDHs exhibit higher sorption capacity for CO2 than montmorillonites and similar to that of zeolites [103,104]. The evaluation of the photocatalytic activity of LDHs was performed in solid-gas systems, where the solid catalyst reacted with a gaseous mixture of CO2 /H2 O or CO2 /H2 , as well as in aqueous suspensions, under UV irradiation.

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

3.4.1. Solid-gas systems In solid-gas systems, the reaction mechanism involves the diffusion of CO2 molecules in the internal and the interlayer space of the LDH, resulting in the formation of hydrogen carbonate species, which is a reaction intermediate. 3.4.1.1. H2 O as the reducing agent. Katsumata et al. studied the photocatalytic activity of Pd or Pt or Au dopped Zn-Cr LDHs which was found satisfying (Pt (0.1%wt)/Zn-Cr LDH −7.8 ␮mol CO/g* 3 h, Pd (0.1%wt)/Zn-Cr LDH-4.3 ␮mol CO/g* 3 h and Au (0.7%wt)/ZnCr LDH-3.8 ␮mol CO/g* 3 h). It was also demonstrated that the undopped Zn-Cr LDHs exhibited very low photocatalytic activity, indicating the crucial role of Pt, Pd and Au which hinder the recombination of the photogenerated species and facilitate the progress of the photocatalytic reactions which proceed via the involvement of two electrons and produce CO [105]. 3.4.1.2. H2 as the reducing agent. As mentioned above, the selectivity of the photocatalytic reduction of CO2 was strongly depended on the LDH composition. Namely, Zn-GaIII , Zn-AlIII , Zn-Cu-GaIII and ZnCu-AlIII LDHs (shown in Fig. 11) produced CO and/or CH3 OH. It was reported that the presence of Cu in the LDH structure enhanced the production of CH3 OH, owing to the Cu-CO2 interactions. The latter promote the reaction among CO2 molecules, H+ ions and photogenerated electrons, as CO2 molecules bind to Cu sites in the form of hydrogen carbonate species. Besides, the insertion of Cu atoms in the LDH structure lowers the band-gap energy and thus enhances the photocatalytic activity. Characteristically, Zn-Cu-AlIII and ZnCu-GaIII LDHs produced 0.31 ␮mol CH3 OH/g* h and 0.49 ␮mol CH3 OH/g* h, respectively. In the second case, Cu atoms were found not only in the interlayer space but also inside the cationic layers, contributing to the faster diffusion of CO2 molecules which form the hydrogen carbonate species. Still, the photocatalytic activities were very low [103,106,107]. Au/Zn-GaIII and Ag/Zn-GaIII LDHs bearing Au or Ag nanoparticles on their surface, produced CH3 OH and CO (0.201 ␮mol CO/g* h and 0.03 ␮mol CH3 OH/g* h, 0.102 ␮mol CO/g* h and 0.118 ␮mol CH3 OH/g* h, respectively). Specifically, in Ag/Zn-GaIII LDH, Ag nanoparticles are electron donors owing to the Surface Plasmonic Resonance (SPR) effect and the methanol yield was higher, whereas, in Au/Zn-GaIII LDH, Au nanoparticles act as electron traps and CO yield was extremely limited at the expense of methanol [108]. 3.4.2. Aqueous dispersions The photocatalytic efficiencies of the aqueous dispersions of LDHs were found to be increased in comparison with those of the solid-state systems. Cu-M-Al (M = Mg, Zn, Ni) LDHs exhibited remarkable photocatalytic activity, especially Cu-Ni-Al LDH, owing to its lowered band-gap energy and its increased specific surface area (210 ␮mol CH3 OH/g* h) [109]. In other study, Ni-Al LDH did not produce CH3 OH, besides it exhibited lowered photocatalytic activity (20 ␮mol CO/g* 8 h), obviously due to the absence of Cu [110]. 3.5. Salts composites Salts have been extensively used as photocatalysts for the CO2 reduction, in order to overcome several disadvantages of metaloxides, such as the high values of band-gap energy. In most cases, their photocatalytic activity was studied in aqueous or organic suspensions, under visible light irradiation. The role of the solvent was crucial and determined the selectivity of the reaction. The formation of CO2 −* free radicals constituted the common intermediate step of all the photocatalytic reactions, where often several reduc-

161

ing agents, such as 2-propanol, Na2 S etc., were added. ZnS and CdS based photocatalysts were the most studied ones. 3.5.1. Aqueous/organic dispersions It was reported that in non polar organic suspensions of CdS and 2-propanol, CO was produced, whereas in polar ones HCOO− was also formed. Obviously, the reaction mechanisms depend on the polarity of the solvent. In non polar solvents, CO2 −* free radical is not solvated, so it is adsorbed on the surface of the catalyst, where CO and H2 O are produced, via double protonation reactions. In polar solvents CO2 −* free radical is stabilized by solvation and then HCOO− is produced via protonation and reduction reactions. Besides, the presence of thiols bound to Cd+2 ions, on the surface of CdS, led to the enhanced formation of HCOO− , owing to the decrease of the amount of adsorbed CO2 −* free radical. Still, thiol modified CdS, in DCM, exhibited low photocatalytic activity (0.5 ␮mol HCOO− /g* 7 h, 0.45 ␮mol CO/g* 7 h, 0.5 ␮mol H2 /g* 7 h and 0.5 ␮mol CH3 COCH3 /g* 7 h) [111,112]. Similar conclusions were obtained by the study of Ru/ZnS suspensions in several solvents, as shown in Fig. 12. During the use of non polar solvents, the photocatalytic activity was reduced, whereas the amount of CO was increased, as mentioned above. Contrariwise, in polar solvents and especially in water, the photocatalytic activity was enhanced, as well as the production of HCOOH, resulting in remarkable photocatalytic efficiencies ZnS (0.8 mmol HCOOH/dm3* h and 0.2 mmol CO/dm3* h, 1.2 mmol HCOOH/dm3* 5 h and 0.4 mmol CO/dm3* 5 h). The high polarity of the solvent, the great ability of Ru nanoparticles to absorb CO2 , to act as electron donors and to reduce the activation energy surely contributed to these hopeful results [112]. Yanagida et al. demonstrated that the surface characteristics of ZnS nanocrystallites (such as sulfur vacancies), which originate from their preparation method, affected the selectivity and activity of the photocatalytic CO2 reduction [113]. The addition of Cd2+ ions on the surface of CdS, leads to the formation of surface defects and the increase in the number of active sites on the surface of the catalyst. For that reason, Cd2+ /CdS suspension in DMF (and triethanolamine as electron donor) produced satisfying amounts of CO (30 ␮mol CO/g * 30 min), which finally formed condensed organic products (benzophenone, acetophenone and benzyl halides) via the formation of a Cd2+ OCOCO2 intermediate complex [114]. The use of supporting materials for the enhancement of the photocatalytic activity of ZnS and CdS salts gave remarkable results. ZnS(13%)/SiO2 suspensions in 2,5-DHF (reducing agent), produced seven times higher amounts of HCOO− (7 mmol/g* 4 h), than the unsupported ones (<1 mmol/g* 4 h). The surface dopping of supported ZnS with excess of Zn+2 ions led to the optimum photocatalytic activity (10.3 mmol/g* 4 h). It is therefore deducted that the use of porous supporting materials, the addition of excess of metal ions on the surface of the catalyst and the presence of a reducing agent in the reaction mixture, combined together, result in the optimal photocatalytic efficiencies [115]. Several studies were performed in aqueous suspensions of ZnS and CdS, supported on MMT, with CTA stabilizer which prevents the agglomeration of salt nanoparticles. They exhibited noteworthy photocatalytic production of H2 (indeed by water splitting!)(CdS/MMT/CTA: 220 ␮mol H2 /g* 25 h, ZnS/MMT/CTA: 210 ␮mol H2 /g* 25 h) and much lower production of CH4 (CdS/MMT/CTA: 30 ␮mol CH4 /g* 25 h, ZnS/MMT/CTA: 40 ␮mol CH4 /g* 25 h). The reaction mechanism for hydrogen production, involves the oxidation of water molecules to O2 and H+ ions, the formation of H* free radicals via the reaction of H+ with photogenerated electrons and finally the reaction between two radicals H* to produce H2 . CH4 is formed either, via the reaction of CO2 with H2 , or via the reaction of CH3 * free radicals (originating from CTA dissociation) with H* free radicals [116–118]. The different amount of ZnS nanoparticles on the surface of MMT (4.2,

162

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Fig. 11. (A) Structure of Zn57-x-y Cux My III (OH)72 . 38H2O cluster models: (i) x = y = 0; (ii) x = 0, y = 12, M = Al; (iii) x = 0, y = 1, M = Ga; (iv) x = 1, y = 0; and (v) x = 1, y = 12, M = Al; (B) Hydrogen carbonate species formed from the reaction of CO2 with the surface hydroxyl group of Zn57-x-y Cux My III (OH)72 . 38H2 O [103].

Fig. 12. Possible reaction mechanisms for the photocatalytic CO2 reduction on the surface of Ru/ZnS nanoparticles, in polar and nonpolar solvents [112].

3.1 ␬␣␫ 2%wt ZnS) was also found to be important as it affects the extent of the agglomeration and the uniform distribution and therefore the photocatalytic activity (optimal efficiency for the ZnS (3.1%wt)/MMT: 300 ␮mol H2 /g* 24 h, 28 ␮mol CH4 /g* 24 h, 3 ␮mol CO/g* 24 h) [119]. The prominent role of the geometric features of a photoreactor was also studied in aqueous ZnS/MMT suspensions. Between two stirred batch annular reactors the one whose geometry secured the optimum mixing of the reactants, exhibited almost 2–3 times higher photocatalytic efficiency (380 ␮mol H2 /g* h, 30 ␮mol CH4 /g* h, 4 ␮mol CO/g* h and 3.2 ␮mol CH3 OH/g* h versus 140 ␮mol H2 /g* h, 8 ␮mol CH4 /g, 0.5 ␮mol CO/g* h and 1.6 ␮mol CH3 OH/g* h) [120]. A lot of studies have been carried out in colloids of CdS and ZnS, containing sacrifice compounds-electron donors (Na2 S, NaH2 PO2 etc.) or other compounds that determine the selectivity of the photocatalytic reaction. For example, the addition of TMACl in the reaction mixture contributed to the production of dimerized products, such as ions of acetic acid or oxalic acid, owing to the formation of an aprotic layer on the surface of the cata-

lyst. Colloids of ZnS exhibited impressive photocatalytic activity (optimum: 2300 ␮mol oxalic ions/dm3* 48 h in pH = 14), whereas that of CdS was much lower (136 ␮mol formic ions/dm3* 42 h in pH = 6) [121–127]. Ni dopping of ZnS aimed to leverage the ability of Ni+2 ions to trap the photogenerated electrons and thus to reduce the probability of recombination of the photogenerated species. Besides, it was expected that the double use of methanol, as solvent and as sacrifice compound and the hydrothermal preparation method of Ni/ZnS would improve significantly the photocatalytic efficiency. Indeed, methanolic suspensios of Ni/ZnS were more efficient (121.4 ␮mol methyl formate/g* 24 h) than bare ZnS (87.6 ␮mol methyl formate/g* 24 h), but the difference was not so dramatic [128]. Contrariwise, the formation of heterojunction structures between CdS and Bi2 S3 , promoted the separation of photogenerated species, enhanced their lifetime, increased the specific surface area and the absorption in the visible part of the spectrum and therefore significantly enhanced the photocatalytic activity of their aqueous suspensions (613 ␮mol CH3 OH/g* 5 h for CdS/Bi2 S3

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

vs 201 ␮mol CH3 OH/g* 5 h for Bi2 S3 and 314 ␮mol CH3 OH/g* 5 h for CdS) [129]. Excluding ZnS and CdS salt composites, several photocatalysts based on other salts, have been studied but their photocatalytic activity was not found to be notably high. Among them, aqueous solutions of FeCl3 or Fe(OH)3 exhibited some photocatalytic activity (2.5 ␮mol HCHO/g* 30 min) owing to the action of Fe+2 ions as electron donors that facilitate the CO2 reduction via the formation of CO2 −* free radicals [130,131]. Cu2 O/SiC in the form of spherical Cu2 O nanoparticles present in SiC agglomerates, exhibited satisfying photocatalytic efficiency (191 ␮mol CH3 OH/g* 5 h), owing to the addition of the reducing agent Na2 SO3 in the reaction mixture and the enrichment with Cu2 O, which improved the separation of the photogenerated species and thus their lifetime (bare Cu2 O: 104 ␮mol CH3 OH/g* 5 h and bare SiC: 153 ␮mol CH3 OH/g* 5 h) [132]. Methanolic solutions of CdIn2 S4 exhibted extraordinary efficiencies in formic methylester (5258 ␮mol/g* 10 h), which was produced via the esterification of CH3 OH ␬␣␫ HCOOH formed at the VB of CdIn2 S4 . Besides, a correlation was performed, between the preparation process of the photocatalyst and its selectivity/activity, to find that when CdIn2 S4 was synthesized from thiouria, the photocatalytic activity was doubled compared to that exhibited in case that l-cystein was the staring material [133]. Aqueous alkaline solutions of CuInS2 and Zn-doped CuInS2 with highly ordered nanoarray structure, where the role of Zn was to reduce the recombination rate of photogenerated species and the contact resistance, were also found to be conspicuous photocatalysts for the CO2 reduction (0.012 M CH3 OH/h and 0.032 M CH3 OH/3 h) [134]. In special occasions, H2 was added to the aqueous reaction mixtures in order to enhance the photocatalytic activity. For example, aqueous suspensions of the salt based photocatalyst reacted with a gas mixture of CO2 /H2 . Namely, AgBr photocatalysts supported on paligorskite have been prepared, in order to enhance their photocatalytic activity, owing to the increased surface area and porosity and the negatively charged paligorskite surface. Besides, paligorskite has an increased gas absorption capacity for CO and H2 and it exhibits high stability towards the effect of visible light irradiation. Eventually, even if the photocatalytic efficiency of the aqueous suspensions of AgBr/paligorskite (8 ␮mol CH4 /g* h and 48 ␮mol CH4 /g* 10 h) was doubled compared to that of bare AgBr (3 ␮mol CH4 /g* h and 12 ␮mol CH4 /g* 10 h), it remained moderate [135]. Also, moderate photocatalytic activity was performed by aqueous suspensions of Cux Agy Inz Znk Sm with RuO2 or Rh1.32 Cr0.66 O3 as cocatalysts (optimum activity: 21.1 ␮mol CH3 OH/g* h) but after the addition of H2 , the activity was found to be almost five times higher (118.5 ␮mol CH3 OH/g* h for RuO2 /Cu0.30 Ag0.07 In0.34 Zn1.31 S2 ). It was reported that H2 reacts with the photogenerated holes to form H* free radicals that facilitate the production of CH3 OH and hinder its oxidation [136]. Furthermore, anomalous spherical nanoparticles of Co/MoS2 , synthesized by hydrothermal method, were used for the photoelectrocatalytic reduction of CO2 to CH3 OH under visible light (35 mmol/L). The Co-doped MoS2 exhibited enhanced photocatalytic activity compared to MoS2 , owing to the upward shift of the CB and VB redox potentials, as well as the enhanced specific surface area and the increased number of active sites [137].

3.5.2. Solid-gas system The only study that refers to the evaluation of the photocatalytic activity of a salt in a solid-gas system, did not lead to encouraging results. CoTe reacted with a gas mixture of CO2 and H2 O to produce very low amounts of CH4 (1.8 ␮mol CH4 /g* h and 4.97 ␮mol CH4 /g* 8 h). In this case the reaction mechanism proceeded via the formation of a variety of free radicals such as H* , CO2 −* , CO−* , C* , CH3 * , CH* and CH2 * [138].

163

Fig. 13. Redox potential values (eV vs NHE) of GR and several common semiconductors [140].

4. Carbonaceous photocatalysts for CO2 reduction The interest about carbonaceous photocatalysts i.e. composite photocatalysts based on carbon (graphite, active carbon, black carbon fibers, carbon nanotubes, graphene and g-C3 N4 ), that usually contain compounds with semiconducting properties, has been expressed recently. In the large majority of the studies, their photocatalytic activity for the CO2 reduction was evaluated during their reaction with H2 O, via the oxidation of water molecules and the formation of free radicals CO2 −* [139,140]. As far as the carbon-based 2D layered materials, such as graphene and g-C3 N4 , are concerned, it is evidenced that their supporting properties, their adsorption affinity for CO2 , the light utilization and the graphene/g-C3 N4 coupling determine their photocatalytic efficiency [39]. 4.1. Composite photocatalysts based on graphene (GR) GR is already known as an electron transfer medium with large specific surface area high chemical stability, excellent electronic conductivity and good optical tranmittance, widely used in photocatalysis. Heteroatom (O, N, B, P and S) doped GR exhibits improved conductive and optical properties and it favors the formation of heterojunctions which facilitate charge generation and transportation and therefore enhance its photocatalytic activity [141,142]. Furthermore, graphene (GR) has been chemically modified in order to obtain superior photocatalytic properties and it has been widely used in composite photocatalysts of the type “semiconductor/GR”. Owing to the existence of Fermi level and the even distribution of semiconductor nanoparticles on their surface, they have exhibited significant photocatalytic activity. Fermi level is found lower than the CB of most semiconductors (less negative value of reduction potential), resulting in the transfer of photogenerated electrons from the CB of the semiconductor to Fermi level of GR as shown in Fig. 13. In that way, the mean free path traversed by the photogenerated electrons is elongated and the photocatalytic reactions are facilitated. In case that Fermi level of GR is found higher than the CB of the semiconductor (more negative value of reduction potential), the photogenerated electrons can’t be transferred from the semiconductor to GR, thus, the presence of a photosensetizer is required (Fig. 13). Photosensitizers are sensitized by the absorption of photons and they produce electrons which follow the path “photosensitizer-GR-semiconductor-reactants” quickly owing to the high conductivity and the unique flat structure of GR. Therefore the mean free path of electrons is elongated, resulting in

164

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Fig. 14. Schematic diagram of the synthetic process of Cu2 O/RGO photocatalysts [139].

Fig. 15. Schematic scheme of the proposed photocatalytic CO2 reduction on the surface of ZnO/RGO photocatalyst [147].

reduction of the recombination rate between the photogenerated species and enhancement of the photocatalytic activity. Besides, owing to the formation of metal-carbon or metal-oxygen-carbon bonds between GR and semiconductors, the band gap energy diminishes and thus the photocatalyst absorbs in the visible part of the spectrtum. Generally, “Semiconductor/GR” photocatalysts exhibit higher photocatalytic activity compared to bare semiconductors as the synergistic effect plays an important role. For that reason, the preparation of “Semiconductor/GR” composites constitutes an ambitious goal for the scientific community. The main preparation processes followed for “Semiconductor/GR” photocatalysts are hydro/solvothermal, solution mixing and in situ growth methods, which apparently affect their structure, morphology, size and photocatalytic activity [39,140,143]. For example, the combination of a facile one-step green synthesis, the presence of metal-oxygen-carbon bonds and the two dimensionality (2D) of GR and GRO nanostructured composites, such as (BiO)2 CO3 /GR and (BiO)2 CO3 /GRO, resulted in high photocatalytic activities owing to their large surface areas and effective charge separation and transfer [144].

[147]. On the other hand, WO3 /RGO composite exhibited a very low photocatalytic activity (0.1 ␮mol CH4 /g* h) as a result of the lowering of its oxidizing power (shift of the VB towards lower values of oxidation potential) compared to that of WO3 , after the addition of RGO [148]. The best photocatalytic activity among the GR based photocatalysts was exhibited by hexamolybdate clusters Cs2 Mo6 Bri 8 Bra 6 /GO and (TBA)2 Mo6 Bri 8 Bra 6 /GO diluted in a mixture of H2 O/DMF (1644 ␮mol CH3OH/g* 24 h and 1294 ␮mol CH3 OH/g* 24 h, respectively versus 439 ␮mol CH3 OH/g* 24 h for bare GO). In this case, the excitation of Mo6 units via the absorption of visible light, facilitated the transfer of electrons to the CB of GO where the reduction of CO2 to methanol takes place. Besides, owing to the high mobility of charged species and its increased specific surface area, GO enhances the photocatalytic activity synergistically [149]. The photocatalytic efficiency of aqueous suspensions of carbon nanoparticles, bare or bearing metals (such as Au) on their surface, which act as electron acceptors, was also evaluated [150]. CdS nanorods/RGO composites photocatalytically reduced CO2 to CH4 under visible light irradiation (2.51 ␮mol CH4 g* h). The enhanced photocatalytic efficiency, compared to that of bare CdS nanorods, was attributed to the electron acceptor properties of RGO and to the strong adsorption and activation of CO2 molecules by RGO, through ␲-␲ conjugation interactions. The content of RGO was 0.5% wt, thus its role was that of an efficient and economic cocatalyst compared to noble metal Pt which did not actually enhance the photocatalytic activity in a higher degree [151,152].

4.1.1. Aqueous/organic suspensions ZnO/RGO, Cu2 O/RGO, BiVO4 /RGO, WO3 /RGO, Cs2 Mo6 Bri 8 Bra 6 /GO and (TBA)2 Mo6 Bri 8 Bra 6 /GO consist characteristic examples of composite photocatalysts based on GR. Aqueous alkaline suspensions of Cu2 O/RGO and BiVO4 /RGO exhibited similar photocatalytic activities in CH3 OH (41.5 ␮mol/g* 10 h) and CH3 CH2 OH (51.5 umol/g* 10 h) respectively [139,145]. The presence of RGO sheets in the Cu2 O/RGO composite, prepared by GO and Cu(AC)2 , as shown in Fig. 14, contributed to the uniform distribution of Cu2 O nanoparticles on the RGO surface. During the photocatalytic reduction of CO2 , the photogenerated electrons were transferred from the CB of Cu2 O to the surface of RGO which provided electrons to CO2 molecules [139]. It was established that ZnO/GO nanocomposites, synthesized by a solvothermal method, exhibit enhanced adsorption ability on CO2 compared to that of bare ZnO and GO, therefore they could form promising photocatalysts for the reduction of CO2 [146]. 2D ZnO/RGO composites consisting of wrinkled GO sheets with nicely distributed ZnO nanoparticles on their surface exhibited remarkable photocatalytic activity (ZnO(10%wt)/RGO:263.1 ␮mol CH3 OH/g* 3 h), 5fold higher than that of bare ZnO (52.4 ␮mol CH3 OH/g* 3 h). The strong interaction between ZnO and RGO resulted in inhibition of the fast recombination of the photogenerated holes and electrons and thus enhancement of the photocatalytic activity. In Fig. 15 is clearly demonstrated that GR acts in a similar way with RGO in the Cu2 O/RGO composite

4.2. Composite photocatalysts based on carbon nanotubes (CNTs) – aqueous suspensions CNTs have been used for the amelioration of the photocatalytic activity of carbonaceous composites, such as AgBr/CNT and plasmonic Ag/AgBr/CNT, because they constitute supporting materials and they possess enhanced ability for storing electrons which are used for the CO2 reduction. Aqueous suspensions of AgBr/CNT with KHCO3 produced significant amounts of CH4 , CH3 OH, CH3 CH2 OH and CO (156, 94, 16 and 39 ␮mol/g* 5 h, respectively). CNTs play the role of an electron sink, which gathers the photogenerated electrons originating from the CB of AgBr and thus facilitates the CO2 reduction mechanism, as shown in Fig. 16. It was also reported that as the length of CNTs was increased, the photocatalytic activity was also increased, owing to the hindering of the recombination of the photogenerated species and the binding of Ag+ ions on the CNTs surface. Besides, the effect of several carbonaceous supporting materials, such as GP, EG, AC and GP, and the effect of pH of the aqueous suspension on the photocatalytic activity, was studied. It

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

165

Fig. 16. Schematic presentation of the photocatalytic CO2 reduction on the surface of the Ag/AgBr/CNT photocatalyst, under visible light [153].

Fig. 18. Suggested schemes of photocatalytic reaction of WO3 /g-C3 N4 composites: (a) charge separation and (b) Z-Scheme type [155].

Fig. 17. Overview of the photocatalytic reduction of CO2 with g-C3 N4 [157].

was demonstrated that the use of GP and CNTs led to the optimum photocatalytic efficiencies because of their excellent electric conductivity [153,154]. 4.3. g-C3 N4 and composite photocatalysts based on g-C3 N4 g-C3 N4 constitutes the most studied carbonaceous photocatalyst for the CO2 reduction (Fig. 17), owing to its very negative value of the CB reduction potential (−1.42 V) and the small value of bandgap energy (2.67 eV), which contribute to its high reducing power and the ability of absorbing in the visible part of the spectrum, respectively. Besides, because of the formation of strong covalent bonds between carbon and nitrogen atoms, g-C3 N4 is stable under irradiation, in alkaline or acidic solutions, and it is synthesized via simple and economic methods, such as the termal polycondensation of melamine at high temperatures (450 ◦ C) [155–158]. During the last years, improved synthetic processes leading to g-C3 N4 materials with high crystallinity and drastic modifications of gC3 N4 surface (such as the introduction of heteromolecule dopants) leading to the red shift of the optical absorption range and the formation of heterostructures in g-C3 N4 , have been performed for the enhancement of the overall photocatalytic activity towards the CO2 reduction [159]. Its photocatalytic activity has been studied in gas-solid systems, in which the solid catalyst reacted with a gas mixture of CO2 /H2 O, as well as in aqueous suspensions. Still, the basic disadvantage of g-C3 N4 is its moderately low oxidizing

power, for this reason, it is usually combined with an additional material which possesses high oxidizing power and it is able to oxidize water molecules. Thus, in case that the photocatalyst is a g-C3 N4 composite, the reaction mechanism (heterojunction) is either a charge separation mechanism or a Z-Scheme mechanism. In the first case (charge separation), the photogenerated electrons are transferred from the CB of the constituent with the more negative value of reduction potential (usually this is g-C3 N4 ) to the CB of the constituent with less negative value of reduction potential (downwards), where the CO2 reduction is performed. To this effect, the photogenerated holes are transported from the VB of the constituent with more positive value of oxidation potential to the VB of the constituent with less positive value of oxidation potential (upwards), where the oxidative reactions (such as water oxidation) take place. In this way, the separation of the photogenerated species is promoted but the total redox ability of the photocatalyst is lowered. In case of the Z-Scheme mechanism, the photogenerated species are not transferred between the conduction bands and the valence bands of the two constituents and thus the redox ability of the composite photocatalyst is maintained [39,155,160,161]. The suggested charge separation and Z-Scheme mechanisms, for the photocatalytic reaction of WO3 /g-C3 N4 composites, are presented in Fig. 18 [155].

4.3.1. Solid-gas systems-H2 O as the reducing agent Outstanding photocatalytic activity (3.48 mmol CO/g* 7 h) was exhibited by mesoporous cubic structured g-C3 N4 , owing to its 3D structure and high specific surface area which enabled the activation of H2 O and CO2 molecules and the effective diffusion of CO2 in the interior space of the catalyst [162]. In other study, 2D ordered layers of g-C3 N4 with Pt nanoparticles distributed on their surface (Pt (1% wt)/g-C3 N4 ), synthesized from urea, exhibited higher photocatalytic activity compared to that of bare g-C3 N4 because Pt facilitated the transfer of the photogenerated species (0.06 umol CH4 /g* h, 0.06 umol CH3 OH/g* h, 0.02 umol HCHO/g* h and 0.12 umol CH4 /g* 7 h, 0.10 umol CH3 OH/g* 7 h, 0.022 umol HCHO/g* 7 h). Certainly, in case that the Pt percentage surpassed a certain value (1%), the efficiency was reduced, due to the re-formation of CO2 and H2 O, via the reaction of O2 with CH3 OH and HCHO [163]. Randomly ordered nanosheets of S/g-C3 N4 produced greater amounts of methanol (1.12 umol CH3 OH/g* 3 h) than pure g-C3 N4 (0.81 umol CH3 OH/g* 3 h), because S/g-C3 N4 absorbs higher amounts of

166

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Fig. 19. Proposed mechanism for the photocatalytic CO2 reduction on the surface of RGO/pCN [166].

visible irradiation (lower band-gap energy), as well as, S atoms generate structural defects that enhance the photocatalytic efficieny [164]. Halogens of Ag were deposited on the surface of protonated g-C3 N4 [AgX(X = Cl ␬␣␫ Br)/pCN], via a method which involved deposition/precipitation processes with the aid of ultrasounds. During the reduction of AgX, Ag atoms displayed surface plasma resonance effect. The enhanced photocatalytic activity of Ag/AgBr(30%wt)/pCN (10.9 ␮mol CH4 /g* 10 h) compared to that of the pure constituents was attributed to the surface plasma resonance effect and to the formation of a heterojunction between pCN and AgBr, into the hybrid photocatalyst. It was also reported that the charge separation was more effective in Ag/AgBr/pCN than in Ag/AgCl/pCN, and this was attributed to the right combination of the redox potentials of pCN and AgBr compared to the combination of redox potentials of pCN and AgCl [165]. The preparation of RGO/pCN photocatalysts via the combination of a dispersion method with ultrasounds, accumulation by electrostatic forces and ␲-␲ interactions and a reductive process with NaBH4 , led to the formation of 2D compounds with intimate contact in width of the heterojunction connection. The latter contributed to the rise of the photocatalytic activity of the hybrids (13.93 ␮mol CH4 /g* 10 h), compared to that of pure RGO and pCN. It was also reported that RGO constituted an excellent accumulation center for electrons that aided the reduction of CO2 to CH4 . Still, the high excess of RGO (>15%) occupied the active catalytic centers at pCN and thus the effective absorption of visible light and the number of photogenerated electrons from pCN to RGO, were reduced. The reaction mechanism involved the transportation of the photogenerated electrons from the VB of pCN to the CB and then to the Fermi level of RGO for the reduction of CO2 to CO2 −* as shown in Fig. 19 [166]. For the first time, it was demonstrated by Yu and Peng et al. that g-C3 N4 /ZnO composites reduced CO2 to CH3 OH (0.6 ␮mol CH3 OH/g* h) via a direct Z-Scheme mechanism (and not the conventional charge separation mechanism) favored by the intimate interfacial contact between the g-C3 N4 and ZnO phases [167].

4.3.2. Aqueous suspensions Amine-functionalized g-C3 N4 , prepared via heating urea at 500 ◦ C and treatment by monoethanolamine solution, exhibited higher CO2 adsorption capacity and thus enhanced photocatalytic activity for CO2 reduction under uv/visible light (0.34 ␮mol CH4 /g* h and 0.28 ␮mol CH3 OH/g* h) compared to that of bare g-

C3 N4 (0.26 ␮mol CH3 OH/g* h). It was inferred that HCO3 − , formed when CO2 was adsorbed on the g-C3 N4 surface, through acid-base interactions with amino groups, in the presence of H2 O, was more active than CO2 [168]. It was reported that in aqueous suspensions of WO3 (0.5%wt Au)/g-C3 N4 , WO3 (0.5%wt Ag)/g-C3 N4 and SnO2-x (42.2%wt)/g-C3 N4 composite photocatalysts, which reacted with CO2 in the gas phase, the Z-Scheme mechanism was followed. Layered structured Au or Ag/WO3 /g-C3 N4 composites, prepared via a planetary mill process, mainly produced methanol (2500 nmol/g* 24 h and 1750 nmol/g* 24 h, respectively), whereas SnO2-x /g-C3 N4 composites, prepared by a simple calcination method, principally produced CO (16 CO/g* h). Ag or Au nanoparticles in WO3 /g-C3 N4 composites participated in the reduction of CO2 , whereas the addition of SnO2-x on the g-C3 N4 surface enhanced the specific surface area and the absorption of visible light of the composite. In both cases, it was demonstrated that in heterojunctions, where the Z-Scheme mechanism is followed, an excellent mixing of the constituents is required, otherwise, the photocatalytic activity appears very low, because the mechanism can’t be evolved. Therefore, it is proved that the enhanced photocatalytic efficiency is ascribed to the effective separation of the photogenerated species, between WO3 and gC3 N4 and the retention of the high oxidizing and reducing power of WO3 and g-C3 N4 , respectively [155,169]. Effective heterojunctions between the constituents of g-C3 N4 /r-P (PCN-x), Pt (0.5%wt) In2 O3 (10%wt)/g-C3 N4 and ZnO/g-C3 N4 composites, led to high photocatalytic efficiencies (295 ␮mol CH4 /g* h, 159.2 ppm CH4 /g* 4 h and 30 ␮mol CO/g* h, respectively). It was reported that in these three cases, a charge separation mechanism was followed. The main advantage of PCN-x composites, prepared by a solid state annealing method, is that they are composed of economic and non-toxic materials which absorb visible light. g-C3 N4 nanosheets folded on the edges of r-P surface formed effective heterojunctions that facilitated the interfacing of g-C3 N4 and r-P and consequently the transfer of charges between them [170]. In similar manner, In2 O3 /g-C3 N4 composites, prepared by a solvothermal method, and composed of In2 O3 nanoparticles on the g-C3 N4 surface, as well as ZnO/g-C3 N4 composites, synthesized by an impregnation method, with chemical bonds between ZnO and g-C3 N4 , exhibited enhanced photocatalytic activity. In both cases, the photogenerated electrons are transferred from the CB of g-C3 N4 to the CB of the second constituent and the reduction of CO2 molecules by the accumulated electrons takes place [171,172].

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

167

Even though, heterojunctions were formed between their constituents, the following photocatalysts, CdMoO4 /g-C3 N4 and g-C3 N4 -BAx didn’t exhibit particularly high efficiencies. Namely, heterojunction between CdMoO4 and g-C3 N4 acted synergistically and increased the activity of the aqueous suspensions of CdMoO4 /g-C3 N4 owing to the separation of the photogenerated species, but generally, led to relative low amounts of products (18 ␮mol CO/g* h and 6 ␮mol CH3 OH/g* h) [173]. Similar, moderate photocatalytic efficiencies (56.3 ␮mol CO/g* 4 h and 12.5 ␮mol H2 /g* 4 h) were exhibited by solutions of g-C3 N4 nanosheets modified with barbitouric acid (BA) (g-C3 N4 -BAx ) in a mixture of H2 O/acetonitrile, containing Co(bpy)3 +2 and TEOA, as cocatalyst and sacrificing compound–electron donor, respectively [157].

5. Hybrid organic-inorganic materials-zeolitic imidazolate frameworks-(ZIFs) The special structure of ZIFs, resembling to that of zeolites, ensures their enhanced chemical and thermal stability, the high values of specific surface area and the high absorption ability for CO2 . Namely, ZIFs form tetrahedral networks by several transition metals such as Cu, Zn and Co which are linked via imidazolate connections. The interaction between the imidazolate connections plays a decisive role in the chemical modification and the specification of their structure. Thus, owing to their special features, recently, ZIFs were studied as catalysts or as cocatalysts of a molecular catalyst for the CO2 photocatalytic reduction. The photocatalytic activity of ZIFs was evaluated in aqueous, organic or mixtures of aqueous and organic solutions. Emphasis was given to the role of solvents-it was demonstrated that the optimum efficiencies of ZIFs were exhibited when aprotic solvents were used, because they facilitated the more sufficient solvation of CO2 [174–177]. Aqueous solutions of orthorhombic structured Cu(II)-ZIFs, prepared by hydrothermal methods, exhibited extraordinary photocatalytic activity (1712 ␮mol CH3 OH/g* 5 h), owing to synergistic effect of the orthorhombic structure and the small band-gap energy [174]. Wang et al. studied the use of a Co containing ZIF (Co-ZIF-9) as cocatalyst for the photocatalytic reduction of CO2 together with a photosensitizer compound based on Ru. It was reported that the presence of Ru increased the cost of the photocatalytic system furthermore, it wasn’t stable enough [178]. Later on, the same authors studied ZIFs (Co-ZIF-9) as cocatalysts in combination with a semiconductor (CdS). Co-ZIF-9/CdS solutions in a mixture of H2 O/acetonitrile, with bpy and TEOA, as electron mediating/donors systems, respectively, exhibited almost 45 times higher photocatalytic activity (50.4 ␮mol CO/g* h and 11.1 H2 ␮mol/g* h), compared to that of pure CdS. The accumulation of the photogenerated electrons of CdS at the active centers of Co-ZIF-9 and their reaction with the adsorbed CO2 molecules, led to a significant enhancement of the photocatalytic efficiency. Namely, it was deduced that the adsorption and diffusion of the reactants was performed on the catalytic centers of Co-ZIF9 and that the electron transfer from the CB of CdS, after the photoexcitation, was the rate determining step of the photocatalytic reaction. Besides, it was found that the role of imidazolate based ligand in the Co-ZIF-9 structure was crucial by affecting positively the adsorption and activation of CO2 and thus its conversion to CO. The effect of temperature and solvent in the photocatalytic production of CO was also studied. It was demonstrated that the temperature enhancement resulted in desorption of CO2 from the reaction system at temperatures higher than 40 ◦ C, whereas a certain amount of H2 O present in the reaction mixture is essential in order to achieve CO selectivity. A possible reaction mechanism is depicted in Fig. 20 [179]. Except of

Fig. 20. Possible reaction mechanism for the Co-ZIF-9 cocatalysed photoreduction of CO2 employing CdS as the light harvester [179].

Co-ZIF-9/CdS, the photocatalytic activity of hybrids consisting of ZIF nanoparticles distributed on the surface of a material with semiconducting properties was studied for Co-ZIF-9/g-C3 N4 and ZIF-8/ZnGeO4 composites too. The highest photocatalytic efficiencies were achieved in aprotic solvents such as MeCN, DMF, THF and DMSO, owing to the Lewis acid-base interactions that facilitated the solvation of CO2 . It was reported that the activities of Co-ZIF-9/g-C3 N4 and ZIF-8/ZnGeO4 (20.8 ␮mol CO/g* 2 h and 2.44 ␮mol CH3 OH/g* 10 h, respectively) were increased compared to those of bare g-C3 N4 and ZnGeO4 . Thus, the synergistic effect of the enhanced capability of ZIFs for CO2 adsorption, the crystallinity and the even distribution of the semiconducting material on the surface of ZIF, resulted in efficient photocatalytic hybrids [177,180,181]. 6. Comparison and assessment of the activities of the main photocatalytic systems/categories of catalysts It was made comprehensible that an ideal photosynthetic system for the CO2 reduction must absorb effectively the solar irradiation and produce large amounts of photogenerated species which do not recombine rapidly, but they are transferred to the active centers of the catalyst, where they react with CO2 molecules. The most preferred photocatalysts possess low band-gap energy and thus they absorb in the visible part of the spectrum in order to achieve their excitation. As it was previously reported, in the large majority of the studies, water was the reducing agent and the reaction mechanism involved its oxidation by the photogenerated holes to produce oxygen and H+ ions and the formation of free radicals CO2 −* intermediates [76–79]. Alternatively, but more rarely, H2 was used as reducing agent and the photocatalytic reaction proceeded via the dissociative adsorption of H2 molecules on the surface of the catalyst followed by the reaction with adsorbed CO2 molecules or free radicals CO2 −* [64,67–69]. CH4 was the least used reducing compound and in this case, the reaction mechanism involved the formation of free radicals such as CH2 * , CH3 * etc. [64,68,70]. Attempts have been performed in order to compare the photocatalytic activities for the CO2 reduction of the main categories of photocatalysts. Concisely, metal or non-metal dopped semiconductors with rather low band-gap energy, materials with special morphology and/or heterojunctions formed between their components, were found to be more efficient photocatalysts. The comparison between photocatalysts that belong to the same or different group is realized provided that it is understood that some factors such as the reaction medium, the kind of reducing compound, the physical state of the photocatalytic system, the kind and the special features of the photoreactor are different, and thus they affect the activity and the selectivity of the photocatalytic reaction.

168

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Table 2 Selected metal oxides reported for CO2 reduction. Photocatalyst description Single-metal oxides hydrated Pb3 O4 (hematite) Hierarchical structures of BiOI Mesoporous Fe/CeO2 Cu/ZnO ZnO immobilized on stainless steel mesh CdS(5%)/WO3 Ga2 O3 ZnO immobilized on stainless steel mesh ZrO2 ZnO immobilized on stainless steel mesh MPc/ZnO immobilized on stainless steel mesh ZnTe/ZnO NiO 1D wedged N/CuO Cu2 O·x H2 O

Mixed-metal oxides Zn2 GeO4 nanoribbons Pt and RuO2 /Zn2 GeO4 nanoribbon KNbO3 nanobelts HNbO3 nanobelts Bi2 WO6 nanoplates ball-flower-like Bi2 WO6 LiTaO3 CaFe2 O4 monoclinic BiVO4 tetragonal BiVO4 lamellar BiVO4 NaBiO3

NaNbO3 NaNbO3 nanowires KNbO3 Cubic Pt/NaNbO3 Orthogonal Pt/NaNbO3 Fe/CuGaO2 , cube-like shaped Pt/CuAlGaO4 C/Fe/LaCoO3 Bi2 WO6 microspheres with conducting polymers (PPy, PANI, PTh) 0.3%Ag/BaCeO3 Metal oxide composites NiO/InTaO4 NiO/InTaO4 Mesoporous Pt (0.5%wt)/ZnAl2 O4 /ZnGaNO CuO/ZnO (10%) supported on MgO NiO(1%wt)/InTaO NiO/InTaO4 Ni@NiO/N-InTaO4 NiO/InNbO4 Co3 O4 /InNbO4 NiO/InVO4 Ni/Ni3 (BO3 )2 /NiO

Efficiency

Type of irradiation

Experimental conditions

Ref. (year)

2.6 ␮mol CH3 OH/g* h 1.58 CH4 ␮mol/g* 8 h)

UV Simulated sunlight

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system

[57] (2011) [58] (2014)

74.3 ␮mol CO/g* h, 17.3 CH4 ␮mol/g* h 1 ␮mol CO/g* h 9% after 5 h

Simulated sunlight

Solid-gas (catalyst-CO2 /H2 O) system

[59] (2014)

VIS UV

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system

[60] (2013) [64] (2013)

1.02 ␮mol CH4 /g* h 3.6 CO ␮mol/g* 5 h 3% after 5 h

VIS UV–VIS UV

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 ) system Solid-gas (catalyst-CO2 /H2 O) system

[66] (2015) [67] (2010) [64] (2013)

1 ␮mol CO/g* 6 h

Solid-gas (catalyst-CO2 /CH4 ) system

[68] (1980)

11% after 5 h

high-pressure Hg-lamps or sunlight UV

Solid-gas (catalyst-CO2 /CH4 ) system

[64] (2013)

23% after 5 h

VIS

Solid-gas (catalyst-CO2 /CH4 ) system

[71] (2014)

100 ␮mol CH4 /g* 5 h 170 ␮mol CH3 OH/g* h 1.8 mmol CH3 OH/L* m2* 6 h 24 ␮mol CH3 OH/L* 20 mg

VIS UV VIS UV–VIS (200 W medium pressure mercury lamp)

Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions

[72] (2015) [73] (2004) [75] (2014) [44] (1989)

1.5 ␮mol CH4 /g* h 25 ␮mol CH4 /g* h

Solar irradiation Solar irradiation

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system

[76] (2010) [76] (2010)

1.71 ␮mol CH4 /g* h 3.58 ␮mol CH4 /g* h 1.1 ␮mol CH4 /g* h 333 nmol CO/g* h 0.42 ␮mol CO/g* 24 h 2 ␮mol CH3 OH/g* 4 h, 3 ␮mol HCHO/g* 4 h 30 ␮mol ethanol/g* 80 min 2 ␮mol ethanol/g* 80 min 35 ␮mol CH3 OH/g* 6 h 0.6 ␮mol CH3 OH/g* h (sol-gel) 0.37 ␮mol CH3 OH/g* h (hydrothermal) 2.3 ppm CH4 /h 1600 ppm CH4 /g* 150 min 7 ppm CH4 /h 1.3 ␮mol CH4 /g* h 0.5 ␮mol CH4 /g* h 9 ppm CO/g* h 8 ␮mol CH3 OH/g* h

UV UV VIS VIS UV UV

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 ) system Aqueous dispersions containing NaH2 PO2 and BaCO3

[77] (2012) [77] (2012) [79] (2011) [80] (2014) [81] (2010) [82] (1994)

VIS VIS VIS VIS

Aqueous NaOH dispersions Aqueous NaOH dispersions Aqueous NaOH dispersions Aqueous dispersions

[83] (2009) [83] (2009) [84] (2012) [85] (2014)

UV UV UV UV UV UV VIS

Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous acidic dispersions with Fe+2 ions

[86] (2012) [87] (2011) [86] (2012) [88] (2012) [88] (2012) [89] (2012) [90] (2013)

130 ␮mol HCOOH/g* h 56.5 ␮mol CH3 OH/g* 4 h, 20.5 ␮mol ethanol/g* 4 h

VIS VIS

Aqueous dispersions Aqueous dispersions

[91] (2009) [92] (2015)

4 ␮mol CH4 /g* 7 h

UV

Aqueous dispersions

[93] (2015)

21.0 ␮mol CH3 OH/g* h (optical fiber reactor) 0.16 ␮mol CH3 OH/g* h 38 ␮mol CH4 /g* 8 h

Simulated solar irradiation

Solid-gas (catalyst-CO2 /H2 O) system

[94] (2010)

VIS VIS

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system

[95] (2011) [96] (2012)

UV–VIS (mercury arc lamp source (250 mW)) VIS Simulated solar irradiation

Aqueous KHCO3 dispersions

[97] (1999)

Aqueous KHCO3 dispersions Aqueous dispersions

[98] (2007) [94] (2010)

VIS VIS VIS VIS UV

Aqueous dispersions Aqueous KHCO3 dispersions Aqueous KHCO3 dispersions Aqueous KHCO3 dispersions Aqueous dispersions

[99] (2011) [100] (2012) [100] (2012) [101] (2015) [102] (2015)

8 ␮mol CH4 /g* h, 0.8 ␮mol C2 H6 /g* h 1.394 ␮mol CH3 OH/g* h 2.8 ␮mol CH3 OH/g* h (quartz reactor) 170 ␮mol CH3 OH/g* h 1.6 ␮mol CH3 OH/g* h 1.5 ␮mol CH3 OH/g* h 1.526 ␮mol CH3 OH/g* h 0.41 ␮mol CH4 /g*h

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

169

Table 3 Selected LDHs reported for CO2 reduction. Photocatalyst description

Efficiency

Layered Double Hydroxides (LDHs) 7.8 ␮mol CO/g* 3 h Pt (0.1%wt)/Zn-Cr LDH Pd (0.1%wt)/Zn-Cr LDH 4.3 ␮mol CO/g* 3 h Au (0.7%wt)/Zn-Cr LDH 3.8 ␮mol CO/g* 3 h 0.201 ␮mol CO/g* h,0.03 ␮mol CH3 OH/g* h Au/Zn-GaIII LDH 0.102 ␮mol CO/g* h, 0.118 ␮mol CH3 OH/g* h Ag/Zn-GaIII LDH 0.31 ␮mol CH3 OH/g* h Zn-Cu-AlIII LDH 0.49 ␮mol CH3 OH/g* h Zn-Cu-GaIII LDH Cu-Ni-Al LDH 210 ␮mol CH3 OH/g* h Ni-Al LDH 20 ␮mol CO/g* 8 h

Type of irradiation

Experimental conditions

Ref. (year)

UV UV UV VIS VIS UV UV VIS UV

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 ) system Solid-gas (catalyst-CO2 /H2 ) system Solid-gas (catalyst-CO2 /H2 ) system Solid-gas (catalyst-CO2 /H2 ) system Aqueous dispersions Aqueous dispersions

[105] (2013) [105] (2013) [105] (2013) [108] (2014) [108] (2014) [107] (2014) [107] (2014) [109] (2015) [110] (2015)

Table 4 Selected LDHs photocatalysts reported for CO2 reduction. Photocatalyst description Salt composites thiol modified CdS

Efficiency

Type of irradiation

Experimental conditions

Ref. (year)

0.5 ␮mol HCOO− /g* 7 h, 0.45 ␮mol CO/g* 7 h, 0.5 ␮mol H2 /g* 7 h, 0.5 ␮mol CH3 COCH3 /g* 7 h

UV-VIS (500 W high pressure mercury arc lamp) VIS

DCM and propanol (hole scavenger)

[111] (1998)

[114] (1998)

UV

DMF and triethanolamine (as electron donor) Aqueous NaOH solutions

UV–VIS

Aqueous colloids

[121] (1998)

VIS UV

Basic aqueous dispersions Aqueous solutions

[129] (2011) [112] (2015)

UV UV

2,5-DHF Aqueous NaOH solutions

[115] (1997) [119] (2014)

UV

Aqueous NaOH solutions

[118] (2013)

UV

Aqueous solutions

[120] (2011)

Cd2+ /CdS

30 ␮mol CO/g * 30 min

CdS/MMT/CTA

220 ␮mol H2 /g* 25 h, 30 ␮mol CH4 /g* 25 h 136 ␮mol formic ions/dm3* 42 h (pH = 6) 613 ␮mol CH3 OH/g* 5 h 0.8 mmol HCOOH/dm3* h, 0.2 mmol CO/dm3* h, 1.2 mmol HCOOH/dm3* 5 h, 0.4 mmol CO/dm3* 5 h 10.3 mmol HCOO− /g* 4 h 300 ␮mol H2 /g* 24 h, 28 ␮mol CH4 /g* 24 h, 3 ␮mol CO/g* 24 h 210 ␮mol H2 /g* 25 h, 40 ␮mol CH4 /g* 25 h 380 ␮mol H2 /g* h, 30 ␮mol CH4 /g* h, 4 ␮mol CO/g* h, 3.2 ␮mol CH3 OH/g* h 2300 ␮mol oxalic ions/dm3* 48 h (pH = 14) 121.4 ␮mol methyl formate/g* 24 h 2.5 ␮mol HCHO/g* 30 min 191 ␮mol CH3 OH/g* 5 h

UV–VIS

Aqueous colloids

[121] (1998)

UV UV VIS

[128] (2013) [131] (1994) [132] (2011)

5258 ␮mol formic methylester/g* 10 h 0.012 M CH3 OH/h

VIS VIS

Methanolic dispersions Aqueous solutions Anoxic aqueous solutions with NaOH and Na2 SO3 Methanolic solutions Aqueous alkaline solutions

8 ␮mol CH4 /g* h

VIS

[135] (2012)

RuO2 /Cu0.30 Ag0.07 In0.34 Zn1.31 S2

118.5 ␮mol CH3 OH/g* h

VIS

CoTe

1.8 ␮mol CH4 /g* h

VIS

Aqueous dispersions (H2 added) Aqueous dispersions (H2 added) Solid-gas (catalyst-CO2 /H2 O) system

CdS CdS/Bi2 S3 Ru/ZnS

Zn+2 /ZnS(13%)/SiO2 ZnS (3.1%wt)/MMT ZnS/MMT/CTA ZnS/MMT ZnS Ni (0.3%wt)/ZnS Fe+2 ions (10−3 M) Cu2 O/SiC Cubic CdIn2 S4 Highly ordered nanoarray structured Zn/CuInS2 AgBr/paligorskite

The photocatalytic efficiencies of the most important metal oxides (Table 2), LDHs (Table 3), salt composites (Table 4), carbonaceous materials and ZIFs (Table 5), the kind of irradiation and the specific experimental conditions are depicted. As far as it concerns the single-metal oxides, the most widely used photocatalytic systems were the solid-gas phase systems, in which water vapor played the role of the reducing compound and the catalyst was found in the solid state. The optimal photocatalytic activity was exhibited by Fe/CeO2 (74.3 ␮mol CO/g* h) [57–65]. During the study of the aqueous suspensions of the single-metal oxide catalyst, the highest photocatalytic activity was expressed by NiO (170 ␮mol CH3 OH/g* h), and it was increased in comparison with that of Fe/CeO2 in the solid-gas phase system [44,72–75]. UV irradiation was frequently used, apparently owing to the increased band gap energy values of the semiconducting materials. On the other hand, in case of doped single-metal oxides, the photocatalytic

[116] (2011)

[133] (2014) [134] (2014)

[136] (2011) [138] (2014)

reduction often took place under visible irradiation, apparently owing to the decreased band gap energy values. Studies on the photocatalytic efficiency of the mixed-metal oxides have shown that in most cases, their aqueous suspensions were used and that the most remarkable results were displayed by the ordered monoclinic BiVO4 (35 ␮mol CH3 OH/g* 6 h) and Fe/C doped LaCoO3 (130 ␮mol HCOOH/g* h), where the synergistic effect between Fe and C contributed significantly to the enhancement of the activity [44,75,82–89,91]. Solid gas systems, where the reaction mixture consisted of H2 O and CO2 , exhibited satisfactory photocatalytic efficiencies, such that of RuO2 (1%wt)/Pt(1%wt)/ZnGeO4 (25 ␮mol CH4 /g* h, 100 ␮mol/CH4 /g* 16 h) [76–79], whereas solidgas systems where H2 was the reducing agent, didn’t produce satisfying amounts of products [81]. Almost half of the mixed-metal oxides were irradiated with UV light whereas the other halves were irradiated with visible light.

170

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

Table 5 Selected carbonaceous materials and ZIFs reported for CO2 reduction. Photocatalyst description

Efficiency

Carbonaceous photocatalysts Composite photocatalysts based on graphene (GR) Cu2 O/RGO 41.5 ␮mol/g* 10 h BiVO4 /RGO 51.5 umol ethanol/g* 10 h 263.1 ␮mol CH3 OH/g* 3 h ZnO(10%wt)/RGO WO3 /RGO 0.1 ␮mol CH4 /g* h 1644 ␮mol CH3 OH/g* 24 h Cs2 Mo6 Bri 8 Bra 6 /GO 2.51 ␮mol CH4 /g* h CdS nanorods/RGO Composite photocatalysts based on carbon nanotubes (CNTs) Ag/AgBr/CNT 30 ␮mol CH4 /g* 5 h, 16 ␮mol CH3 OH/g* 5 h, 3 ␮mol ethanol/g* 5 h, 8 ␮mol CO/g* 5 h (pH = 8) 156 ␮mol CH4 /g* 5 h, 94 ␮mol AgBr/CNT CH3 OH/g* 5 h, 16 ␮mol ethanol/g* 5 h, 39 ␮mol CO/g* 5 h (pH = 8) g-C3 N4 and composite photocatalysts based on g-C3 N4 mesoporous g-C3 N4 (cubic) 3.48 mmol CO/g* 7 h Pt (1% wt)/g-C3 N4 0.06 umol CH4 /g* h, 0.06 umol CH3 OH/g* h, 0.02 umol HCHO/g* h 1.12 umol CH3 OH/g* 3 h S/g-C3 N4 Ag/AgBr(30%wt)/pCN 10.9 ␮mol CH4 /g* 10 h RGO/pCN 13.93 ␮mol CH4 /g* 10 h Amine-functionalized g-C3 N4 0.34 ␮mol CH4 /g* h, 0.28 ␮mol CH3 OH/g* h Au/WO3 /g-C3 N4 2500 nmol CH3 OH/g* 24 h Ag/WO3 /g-C3 N4 1750 nmol CH3 OH/g* 24 h SnO2-x (42.2%wt)/g-C3 N4 16 ␮mol CH3 OH/g* h g-C3 N4 /r-P (PCN-x) 295 ␮mol CH4 /g* h 159.2 ppm CH4 /g* 4 h Pt (0.5%wt) In2 O3 (10%wt)/g-C3 N4 30 ␮mol CO/g* h ZnO/g-C3 N4 CdMoO4 /g-C3 N4 18 ␮mol CO/g* h, 6 ␮mol CH3 OH/g* h 56.3 ␮mol CO/g* 4 h, 12.5 ␮mol H2 /g* 4 h g-C3 N4 -BAx Hybrid organic-inorganic materials-Zeolitic imidazolate frameworks-ZIFs) 1712 ␮mol CH3 OH/g* 5 h Orthorhombic Cu(II)-ZIFs 50.4 ␮mol CO/g* h, 11.1 H2 ␮mol/g* h Co-ZIF-9/CdS Co-ZIF-9/g-C3 N4 20.8 ␮mol CO/g* 2 h ZIF-8/ZnGeO4

2.44 ␮mol CH3 OH/g* 10 h

The mixing of metal oxides in metal-oxide composites resulted in the amelioration of their optical properties and thus the effective photoexcitation of their VB electrons under visible irradiation. Also, metal oxide composites were studied in depth as far as it concerns their photocatalytic activity towards CO2 reduction with H2 O in solid-gas systems as well as in aqueous suspensions. N/InTaO4 was found to be the more efficient photocatalyst in a solid-gas reaction system (125 ␮mol CH3 OH/g* h and 250 ␮mol CH3 OH/g* 2 h) whereas Ni@NiO/N-InTaO4 dispersed in aqueous solution exhibited particularly high photocatalytic efficiency (170 ␮mol CH3 OH/g* h and 310 ␮mol CH3 OH/g* 2 h) owing to the presence of Ni, NiO which acted as electron donors [94–102]. The high absorption capacity for CO2 of LDHs constitutes the basic factor that contributes to the very high photocatalytic activities. It is reported that the main products are CH3 OH and CO, and the presence of Cu determines the selectivity of the reaction for methanol [103,104]. Aqueous suspensions of Cu-Ni-Al LDH exhibited the highest photocatalytic efficiencies in CH3 OH (210 ␮mol/g* h and 620 ␮mol/g* 3 h) [102], whereas solid-gas systems where the reaction mixture was H2 O/CO2 or H2 /CO2 , didn’t exhibit significant photocatalytic activities (from 0.2 ␮mol to 20 ␮mol product/g* h) [103,106–108]. The percentage of UV light absorbing LDHs is higher than that of visible light absorbing ones, owing to their high values of band-gap energy. The photocatalytic activity of salt composites was studied mainly in their aqueous or organic solutions, where usually common reducing agents were added. Besides, the effect of the solvent’s

Type of irradiation

Experimental conditions

Ref. (year)

Simulated sunlight Simulated sunlight UV–VIS VIS VIS VIS

Aqueous NaOH solutions Aqueous NaOH solutions Anoxic aqueous alkaline solutions Aqueous dispersions H2 O/DMF Aqueous acidic dispersions

[139] (2014) [145] (2015) [147] (2015) [148] (2013) [149] (2015) [152] (2014)

VIS

Aqueous suspensions with KHCO3

[153] (2013)

VIS

Aqueous suspensions with KHCO3

[154] (2014)

VIS VIS

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system

[162] (2014) [163] (2014)

VIS VIS VIS VIS

Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Solid-gas (catalyst-CO2 /H2 O) system Aqueous acidic dispersions

[164] (2015) [165] (2015) [166] (2015) [168] (2015)

VIS VIS Simulated sunlight VIS VIS Simulated sunlight VIS VIS

Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions Aqueous dispersions H2 O/acetonitrile, containing Co(bpy)3 +2 and TEOA

[155] (2014) [155] (2014) [169] (2015) [170] (2013) [171] (2014) [172] (2015) [173] (2015) [157] (2015)

VIS VIS VIS

Aqueous solutions H2 O/acetonitrile, with bpy and TEOA Aprotic solvents (MeCN, DMF, THF, DMSO) Aqueous dispersions

[174] (2013) [179] (2015) [180] (2014)

VIS

[181] (2013)

nature on the selectivity was confirmed, whereas extremely high photocatalytic activities were mentioned, higher than those of metal oxides and LDHs. Among the higher photocatalytic efficiencies are those of Zn+2 /ZnS(13%wt)/SiO2 in 2,5-DHF (10.3 mmol formats/4 h), aqueous suspensions of ZnS (3.1%wt)/MMT/CTA (300 ␮mol H2 /g* h), colloids of ZnS (2300 ␮mol oxalates/dm3 ) and methanolic solutions of CdIn2 S4 (5258 ␮mol methyl formate/10 h) [111–138]. Except of the ZnS based salt composites which mainly absorb UV irradiation, the majority of salt composites absorb visible light and this fact makes them appealing photocatalysts for the CO2 reduction when they are dissolved in aqueous or organic mediums. The most recently studied group of photocatalysts for the CO2 reduction is that of carbonaceous catalysts which exhibited very encouraging results. Except of their stability, carbonaceous materials, based on GR, CNTs or g-C3 N4 , usually combined with inorganic semiconductors, have exhibited outstanding photocatalytic activities [139,140]. In case of the composite carbonaceous photocatalysts, the enhanced photocatalytic activities were attributed to the generation of effective heterojunctions between their components (charge separation mechanism or Z-Scheme). The photocatalytic activity of carbonaceous photocatalysts has been studied either in aqueous/organic suspensions or in gassolid systems where the gaseous reaction mixture consisted of H2 O and CO2 . Aqueous suspensions of BiVO4 /RGO (51.5 ␮mol ethanol/g* 10 h), ZnO(10%wt)/RGO (263.1 ␮mol CH3 OH/g* 3 h) and solutions of Cs2 Mo6 Bri 8 Bra 6 /GO (1644 ␮mol CH3 OH/g* 24 h) in H2 O/DMF exhibited the most enhanced activities among the GR

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

based carbonaceous photocatalysts [139,145,147–150]. Among the g-C3 N4 photocatalysts, the optimal photocatalytic activities were exhibited by aqueous suspensions of g-C3 N4 /r-P (295 ␮mol CH4 /g* h) and S/g-C3 N4 (1.12 umol CH3 OH/g* 3 h) owing to its surface defects [155,157,170–173]. ZIFs have also displayed satisfying photocatalytic activities, the highest was that of the aqueous suspensions of orthorhombic structured Cu(II) ZIFs (1712 ␮mol CH3 OH/g* 5 h) [174,179–181]. Furthermore, all of them are visible light active, i.e. they satisfy one of the basic requirements of an ideal photocatalyst. In conclusion, a rough comparison of the photocatalytic activity among the main groups of photocatalysts, indicates that singlemetal oxides, mixed-metal oxides and metal oxide composites, have exhibited similar photocatalytic efficiencies (up to 170 ␮mol CH3 OH/g* h). LDHs have displayed slightly higher photocatalytic activities (up to 210 ␮mol CH3 OH/g* h), whereas salt composites, carbonaceous materials and ZIFs have exhibited even higher photocatalytic efficiencies, that reach the amount of 5258 ␮mol methyl formate/g* 10 h, 1644 ␮mol CH3 OH/g* 24 h and 1712 ␮mol CH3 OH/g* 5 h. Generally, it can be deduced that the aqueous suspensions or aqueous/organic solutions of the photocatalysts, exhibited higher photocatalytic activities than their gas-solid systems. 7. Conclusions The photocatalytic reduction of CO2 combines two great profits that take place concurrently: firstly, the lowering of the CO2 concentration in the atmosphere and thus the control of the amplification of the Greenhouse effect and secondly, the production of useful chemicals and environmentally friendly fuels. Therefore it consists one of the greatest challenges for the scientific community, even if there are many obstacles in the actualization and the extended application of this concept in the real life. The enhanced thermodynamic stability of CO2 , the incapacity of sufficient exploitation of the solar energy, transportation and storage of the produced chemicals/fuels, the economic cost of manufacturing effective photoreactors, the loss of reactivity of photocatalysts during operation, have been the major hindrances. On the other hand, some groups of photocatalysts, such as the salt composites, ZIFs and carbonaceous hybrids have exhibited outstanding photocatalytic efficiencies, whereas at the same time, they are economic materials and they have been evaluated in rather efficacious photoreactors (such as OFMRs). Therefore, as the latest studies in this field were very hopeful, scientists work intensely and promise optimal and applicable results, translated into the preparation of reactive photocatalysts for the conversion of the atmospheric CO2 wastes to environmental friendly and useful materials. Therefore, we hope that this review will motivate the scientific community to synthesize/modify alternative photocatalysts to TiO2 in order to accomplish even higher photocatalytic activities for the CO2 reduction. References [1] A.J. Heeger, Solar Fuels and Artificial Photosynthesis-Science and Innovation to Change Our Future Energy Options, 2012 www.rsc.org/solar-fuels. [2] J.A. Herron, J. Kim, A. Upadhye, G. Huber, C.T. Maravelias, A general framework for the assessment of solar fuels technologies, Energy Environ. Sci. (2014), http://dx.doi.org/10.1039/C4EE01958J. [3] K. Kalyanasundaram, M. Graetzel, Artificial photosynthesis: biomimetic approaches to solar energy conversion and storage, Curr. Opin. Biotechnol. 21 (2010) 298–310. [4] L. Chi-Hung, H. Chao-Wei, C.S. Jeffrey, Hydrogen production from semiconductor-based photocatalysis via water splitting, Catalysts 2 (2012) 490–516. [5] A. Pandit, A. Holzwarth, H.J.M. de Groot, Harnessing Solar Energy for the Production of Clean Fuel, European Science Foundation, Regensburg, 2008, ISBN 9789090239071.

171

[6] V.A. Emeline, V.A. Otroshchenko, V.K. Ryabchuk, N. Serpone, Abiogenesis and photostimulated heterogeneous reactions in the interstellar medium and on primitive earth relevance to the genesis of life, J. Photochem. Photobiol. C Photochem. Rev. 3 (2003) 203–224. [7] D. Gust, T.A. Moore, A.L. Moore, Solar fuels via artificial photosythesis, Acc. Chem. Res. 42 (12) (2009) 1890–1898. [8] J.J. Concepcion, R.L. House, J.M. Papanikolas, T.J. Meyer, Chemical approaches to artificial photosynthesis, PNAS 109 (39) (2012) 15560–15564. [9] D.C. Grills, E. Fujita, New directions for the photocatalytic reduction of CO2 : supramolecular scCO2 or biphasic ionic liquid-scCO2 systems, J. Phys. Chem. Lett. 1 (2010) 2709–2718. [10] A.D. Handoko, K. Li, J. Tang, Recent progress in artificial photosynthesis: CO2 photoreduction to valuable chemicals in a heterogeneous system, Curr. Opin. Chem. Eng. 2 (2012) 200–206. [11] A.C. Benniston, A. Harriman, Artificial photosynthesis, Mater. Today 11 (2008) 26–34. [12] I. McConnell, G. Li, G.W. Brudvig, Energy conversion in natural and artificial photosynthesis, Chem. Biol. 17 (2010) 434–447. [13] A. Thapper, S. Styring, G. Saracco, A.W. Rutherford, B. Robert, A. Magnuson, W. Lubitz, A. Liobet, P. Kurz, A. Holzwarth, S. Fiechter, H. Groot, S. Campagna, A. Braun, H. Bercegol, V. Artero, Artificial photosythesis for solar fuels-an evolving research field within AMPEA, a joint programma of the Europea Energy Research Alliance, Green 3 (1) (2013) 43–57. [14] Y. Zheng, Z. Pan, X. Wang, Advances in photocatalysis in China, Chin. J. Catal. 34 (2013) 524–535. [15] C. Avelino, G. Hermenegildo, Photocatalytic reduction of CO2 for fuel production: possibilities and challenges, J. Catal. 308 (2013) 168–175. [16] R. Richter, S. Caillol, Fighting global warming: the potential of photocatalysis against CO2 CH4 , N2 O, CFCs, tropospheric O3 , BC and other major contributors to climate change, J. Photochem. Photobiol. C: Photochem. Rev. 12 (2011) 1–19. [17] P.G. Falkowski, Biotechnology and global climate change, Curr. Opin. Biotechnol. 3 (1992) 286–290. [18] M. Tahir, N.S. Amin, Recycling of carbon dioxide to renewable fuels by photocatalysis: prospects and challenges, Renew. Sustain. Energy Rev. 25 (2013) 560–579. [19] K. Li, X. An, K.H. Park, M. Khraisheh, J. Tang, A critical review of CO2 photoconversion: catalysts and reactors, Catal. Today 224 (2014) 3–12. [20] S. Solomon, G.-K. Plattner, R. Knutti, P. Friedlingstein, Irreversible climate chang due to carbo dioxide emissions, PNAS 106 (6) (2009) 1704–1709. [21] R. Richter, T. Ming, S. Caillol, Fighting global warming by photocatalytic reduction of CO2 using giant photocatalytic reactors, Renew. Sustain. Energy Rev. 19 (2013) 82–106. [22] C. Song, Global challenges and strategies for control conversion and utilization of CO2 for sustainable development involving energy, catalysis, adsorption and chemical processing, Catal. Today 115 (2006) 2–32. [23] J.N. Armor, Catalytic solutions to reduce pollutants, Catal. Today 38 (1997) 163–167. [24] I. Ganesh, Conversion of carbon dioxide into methanol—a potential liquid fuel: fundamental challenges and opportunities, Renew. Sustain. Energy Rev. 31 (2014) 221–257. [25] G. Centi, S. Perathoner, Heterogeneous catalytic reactions with CO2 : status and perspectives, Stud. Surf. Sci. Catal. 153 (2004) 1–8. [26] G. Mul, C. Schacht, W. Swaaij, J.A. Moulijn, Functioning devices for solar to fuel conversion, Chem. Eng. Process. 51 (2012) 137–149. [27] M. Aresta, A. Dibenedetto, A. Angelini, Converting exhaust carbon into working carbon, Adv. Inorg. Chem. 66 (2014) 259–288. [28] B. Hu, C. Guild, S.L. Suib, Thermal electrochemical and photochemical conversion of CO2 to fuels and value-added products, J. CO2 Utiliz. 1 (2013) 18–27. [29] C. Graves, S.D. Ebbesen, M. Mogensen, K.S. Lackner, Sustainable hydrocarbon fuels by recycling CO2 and H2 O with renewable or nuclear energy, Renew. Sustain. Energy Rev. 15 (2011) 1–23. [30] G. Centi, S. Perathoner, Opportunities and prospects in the chemical recycling of carbon dioxide to fuels, Catal. Today 148 (2009) 191–205. [31] J. Yuan, C. Hao, Solar-driven photoelectrochemical reduction of carbon dioxide to methanol at CuInS2 thin film photocathode, Sol. Energy Mater. Sol. Cells 108 (2013) 170–174. [32] M. Aresta, A. Dibenedetto, A. Angelini, The changing paradigm in CO2 utilization, J. CO2 Utiliz. 3–4 (2013) 65–73. [33] M. Gangeri, S. Perathoner, S. Caudo, G. Centi, J. Amadou, D. Begin, C. Pham-Huu, M.J. Ledoux, J.-P. Tessonnier, D.S. Su, R. Schlogl, Fe and Pt carbon nanotubes for the electrocatalytic conversion of carbon dioxide to oxygenates, Catal. Today 143 (2009) 57–63. [34] C. Genovese, C. Ampelli, S. Perathoner, G. Centi, Electrocatalytic conversion of CO2 to liquid fuels using nanocarbon-based electrodes, J. Energy Chem. 22 (2013) 202–213. [35] Y.P. Peng, Y.T. Yeh, C.P. Wang, C.P. Huang, A solar cell driven electrochemical process for the concurrent reduction of carbon dioxide and degradation of azo dye in dilute KHCO3 electrolyte, Sep. Purif. Technol. 117 (2013) 3–11. [36] B.S. Kwak, M. Kang, Photocatalytic reduction of CO2 with H2 O using perovskite Cax Tiy O3 , Appl. Surf. Sci. 337 (2015) 138–144. [37] S. Liu, J. Xia, J. Yu, Amine-functionalized titanate nanosheet-assembled yolk@shell microspheres for efficient cocatalyst-free visible-light photocatalytic CO2 reduction, ACS Appl. Mater. Interfaces 7 (15) (2015) 8166–8175.

172

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

[38] B. Ohtani, in: S.L. Suib (Ed.), Principle of Photocatalysis and Design of Active Photocatalysts. In New and Future Developments in Catalysis: Solar Photocatalysis, 1st ed., Elsevier, Poland, 2013. [39] J. Low, B. Cheng, J. Yu, M. Jaroniec, Carbon-based two-dimensional layered materials for photocatalytic CO2 reduction to solar fuels, Energy Storage Mater. 3 (2016) 24–35. [40] X. Li, J. Wen, J. Low, Y. Fang, J. Yu, Design and fabrication of semiconductor photocatalyst for photocatalytic reduction of CO2 to solar fuel, Sci. China Mater. 57 (2014) 70–100. [41] K.D. Dubois, G. Li, Innovative photocatalysts for solar fuel generation by CO2 reduction, in: S.L. Suib (Ed.), New and Future Developments in Catalysis: Solar Photocatalysis, 1 st ed., Elsevier, Poland, 2013, http://dx.doi.org/10. 1016/B978-0-444-53872-7.00010-8. [42] L. Zhang, H.H. Mohamed, R. Dillert, D. Bahnemann, Kinetics and mechanisms of charge transfer processes in photocatalytic systems: a review, J. Photochem. Photobiol. C 13 (2012) 263–265. [43] M.D. Hernandez-Alonso, F. Fresno, S. Suarez, J.M. Coronado, Development of alternative photocatalysts to TiO2 : challenges and opportunities, Energy Environ. Sci. 2 (2009) 1231–1257. [44] K. Tennakone, Selective photoreduction of carbon dioxide to methanol with hydrous cuprous oxide, J. Photochem. Photobiol. A Chem. 49 (1989) 369–375. [45] M. Marszewski, S. Cao, J. Yu, M. Jaroniec, Semiconductor-based photocatalytic CO2 conversion, Mater. Horiz. 2 (2015) 261–278. [46] K. Li, D. Martin, J. Tang, Conversion of solar energy to fuels by inorganic heterogeneous systems, Chin. J. Catal. 32 (2011) 879–890. [47] K. Li, X. An, K.H. Park, M. Khraisheh, J. Tang, A critical review of CO2 photoconversion: catalysts and reactors, Catal. Today 224 (2014) 3–12. [48] A. Corma, H. Garcia, Photocatalytic reduction of CO2 for fuel production: possibilities and challenges, J. Catal. 308 (2013) 168–175. [49] X. Zong, L. Wang, Ion-exchangeable semiconductor materials for visible light-induced photocatalysis, J. Photochem. Photobiol. C 18 (2014) 32–49. [50] C. Wang, X. Zhang, Y. Liu, Promotion of multi-electron transfer for enhanced photocatalysis: a review focused on oxygen reduction reaction, Appl. Surf. Sci. 355 (2015) 28–45. [51] J.S. Lee, J. Jang, Hetero-structured semiconductor nanomaterials for photocatalytic applications, J. Ind. Eng. Chem. 20 (2014) 363–371. [52] H. Yoneyama, Photoreduction of carbon dioxide on quantized semiconductor nanoparticles in solution, Catal. Today 39 (1997) 169–175. [53] J. Guo, S. Ouyang, T. Kako, J. Ye, Mesoporous In(OH)3 for photoreduction of CO2 into renewable hydrocarbon fuels, Appl. Surf. Sci. 280 (2013) 418–423. [54] X. Li, J. Yu, M. Jaroniec, Hierarchical photocatalysts, Chem. Soc. Rev. 45 (9) (2016) 2603–2636. [55] B. Hauchecorne, S. Lenaerts, Unravelling the mysteries of gas phase photocatalytic reaction pathways by studying the catalyst surface: a literature review of different Fourier transform infrared spectroscopic reaction cells used in the field, J. Photochem. Photobiol. C 14 (2013) 72–85. [56] S. Girish Kumar, K.S.R. Koteswara Rao, Tungsten-based nanomaterials (WO3 & Bi2 WO6 ): modifications related to charge carrier transfer mechanisms and photocatalytic applications, Appl. Surf. Sci. 355 (2015) 939–958. [57] M. Bartoszek, M. Wecks, G. Jacobs, D. Mohlmann, Photochemically induced formation of Mars relevant oxygenates and methane from carbon dioxide and water, Planet Space Sci. 59 (2011) 259–263. [58] G. Zhang, A. Su, J. Qu, Y. Xu, Synthesis of BiOI flowerlike hierarchical structures toward photocatalytic reduction of CO2 to CH4 , Mater. Res. Bull. 55 (2014) 43–47. [59] Y. Wanga, F. Wanga, Y. Chena, D. Zhanga, B. Lib, S. Kangb, X. Lib, L. Cui, Enhanced photocatalytic performance of ordered mesoporous Fe-doped CeO2 catalysts for the reduction of CO2 with H2 O under simulated solar irradiation, Appl. Catal. B Environ. 147 (2014) 602–609. [60] J. Nuneza, V.A. Pena O’ Shea, P. Jana, J.M. Coronado, D.P. Serrano, Effect of copper on the performance of ZnO and ZnO1-x Nx oxides as CO2 photoreduction catalysts, Catal. Today 209 (2013) 21–27. [61] T. Inoue, A. Fujishima, S. Konishi, K. Honda, Nature 277 (1979) 637–638. [62] G.I. Punov, Photocatalysis of CO2 , Appl. Catal. A Gen. 99 (2) (1993) N19–N20. [63] M. Watanabe, Photosynthesis of methanol and methane from CO2 and H2 O molecules on a ZnO surface, Surf. Sci. 279 (3) (1992) L236–L242. [64] G. Mahmodi, S. Sharifnia, M. Madani, V. Vatanpour, Photoreduction of carbon dioxide in the presence of H2 , H2 O and CH4 over TiO2 and ZnO photocatalysts, Sol. Energy 97 (2013) 186–194. [65] S.K. Atreya, Z.G. Gu, Photochemistry and stability of the atmosphere of mars, Adv. Space Res. 16 (6) (1995) 57–68. [66] J. Jin, J. Yu, D. Guo, C. Cui, W. Ho, A hierarchical Z-Scheme CdS–WO3 photocatalyst with enhanced CO2 reduction activity, Small 11 (39) (2015) 5262–5271. [67] H. Tsuneoka, K. Teramura, T. Shishido, T. Tanaka, Adsorbed species of CO2 and H2 on Ga2 O3 for the photocatalytic reduction of CO2 , J. Phys. Chem. C 114 (2010) 8892–8898. [68] B. Aurian-Blajeni, M. Halmann, J. Manassen, Photoreduction of carbon dioxide and water into formaldehyde and methanol on semiconductor materials, Sol. Energy 25 (1980) 165–170. [69] E. Ivana, V. Lacatusu, M. Craiu, M. Chelu, C. Fenic, R. Dabu, A. Stratan, L. Nistor, D. Dragulinescu, I. Chis, P. Marginean, Laser enhanced H2 chemisorption on Ni1-y O catalysts for CO2 reduction, Appl. Surf. Sci. 106 (1996) 198–204.

[70] G. Mahmodi, S. Sharifnia, F. Rahimpour, S.N. Hosseini, Photocatalytic conversion of CO2 and CH4 using ZnO coated mesh: effect of operational parameters and optimization, Sol. Energy Mater. Sol. Cells 111 (2013) 31–40. [71] A. Yarahmadi, S. Sharifnia, Dye photosensitization of ZnO with metallophthalocyanines (Co, Ni and Cu) in photocatalytic conversion of greenhouse gases, Dyes Pigm. 107 (2014) 140–145. [72] M.F. Ehsan, T. He, In situ synthesis of ZnO/ZnTe common cation heterostructure and its visible-light photocatalytic reduction of CO2 into CH4 , Appl. Catal. B Environ. 166–167 (2015) 345–352. [73] A.H. Yahaya, M.A. Gondal, A. Hameed, Selective laser enhanced photocatalytic conversion of CO2 into methanol, Chem. Phys. Lett. 400 (2004) 206–212. [74] L. Yuan, Y.-J. Xu, Photocatalytic conversion of CO2 into value-added and renewable fuels, Appl. Surf. Sci. 342 (2015) 154–167. [75] P. Li, J. Xu, H. Jing, C. Wu, H. Peng, J. Lu, H. Yin, Wedged N-doped CuO with more negative conductive band and lower overpotential for high efficiency photoelectric converting CO2 to methanol, Appl. Catal. B Environ. 156–157 (2014) 134–140. [76] Q. Liu, Y. Zhou, J. Kou, X. Chen, Z. Tian, J. Gao, S. Yan, Z. Zou, High-yield synthesis of ultralong and ultrathin Zn2 GeO4 nanoribbons toward improved photocatalytic reduction of CO2 into renewable hydrocarbon fuel, J. Am. Chem. Soc. 132 (2010) 14385–14387. [77] X. Li, H. Pan, W. Li, Z. Zhuang, Photocatalytic reduction of CO2 to methane over HNb3 O8 nanobelts, Appl. Catal. A Gen. 413–414 (2012) 103–108. [78] Y. Zhou, X. Zhang, Z. Zhao, Q. Zhang, F. Wang, Y. Lin, Effects of pH on the visible-light induced photocatalytic and photoelectrochemical performances of hierarchical Bi2 WO6 microspheres, Superlattices Microstruct. 72 (2014) 238–244. [79] Y. Zhou, Z. Tian, Z. Zhao, Q. Liu, J. Kou, X. Chen, J. Gao, S. Yan, Z. Zou, High-yield synthesis of ultrathin and uniform Bi2 WO6 square nanoplates benefitting from photocatalytic reduction of CO2 into renewable hydrocarbon fuel under visible light, Appl. Mater. Interfaces 3 (2011) 3594–3601. [80] Z. Sun, Z. Yang, H. Liu, H. Wang, Z. Wu, Visible-light CO2 photocatalytic reduction performance of ball-flower-like Bi2 WO6 synthesized without organic precursor: effect of post-calcination and water vapor, Appl. Surf. Sci. 315 (2014) 360–367. [81] K. Teramura, S. Okuoka, H. Tsuneoka, T. Shishido, T. Tanaka, Photocatalytic reduction of CO2 using H2 as reductant over ATaO3 photocatalysts (A = Li, Na K), Appl. Catal. B Environ. 96 (2010) 565–568. [82] Y. Matsumoto, M. Obata, J. Hombo, Photocatalytic reduction of carbon dioxide on p-type CaFe2 O4 powder, J. Phys. Chem. 98 (1994) 2950–2951. [83] Y. Liu, B. Huang, Y. Dai, X. Zhang, X. Qin, M. Jiang, M.H. Whangbo, Selective ethanol formation from photocatalytic reduction of carbon dioxide in water with BiVO4 photocatalyst, Catal. Commun. 11 (2009) 210–213. [84] J. Mao, T. Peng, X. Zhang, K. Li, L. Zan, Selective methanol production from photocatalytic reduction of CO2 on BiVO4 under visible light irradiation, Catal. Commun. 28 (2012) 38–41. [85] I.A. Mkhalid, Preparation and characterization of NaBiO3 nanopowders by different methods for photocatalytic reduction of CO2 , Ceram. Int. 40 (2014) 5795–5800. [86] H. Shi, Z. Zou, Photophysical and photocatalytic properties of ANbO3 (A = Na, K) photocatalysts, J. Phys. Chem. Solids 73 (2012) 788–792. [87] H. Shi, T. Wang, J. Chen, C. Zhu, J. Ye, Z. Zou, Photoreduction of carbon dioxide over NaNbO3 nanostructured photocatalysts, Catal. Lett. 4 (2011) 525–530. [88] P. Li, S. Ouyang, G. Xi, T. Kako, J. Ye, The effects of crystal structure and electronic structure on photocatalytic H2 evolution and CO2 reduction over two phases of perovskite-structured NaNbO3 , J. Phys. Chem. C 116 (2012) 7621–7628. [89] J.W. Lekse, M.K. Underwood, J.P. Lewis, C. Matranga, Synthesis, characterization electronic structure and photocatalytic behavior of CuGaO2 and CuGa1-x Fex O2 (x = 0.05, 0.10, 0.15, 0.20) delafossite, J. Phys. Chem. C 116 (2012) 1865–1872. [90] Wei-H. Lee, C.-H. Liao, M.-F. Tsai, C.-W. Huang, J.C.S. Wu, A novel twin reactor for CO2 photoreduction to mimic artificial photosynthesis, Appl. Catal. B Environ. 132–133 (2013) 445–451. [91] L. Jia, J. Li, W. Fang, Enhanced visible-light active C and Fe co-doped LaCoO3 for reduction of carbon dioxide, Catal. Commun. 11 (2009) 87–90. [92] W. Dai, H. Xu, J. Yu, X. Hu, X. Luo, X. Tu, L. Yang, Photocatalytic reduction of CO2 into methanol and ethanol overconducting polymers modified Bi2 WO6 microspheres under visible light, Appl. Surf. Sci. 356 (2015) 173–180. [93] J. Wang, C. Huang, X. Chen, H. Zhang, Z. Li, Z. Zou, Photocatalytic CO2 reduction of BaCeO3 with 4f configuration electrons, Appl. Surf. Sci. 358 (2015) 463–467. [94] Z.-Y. Wang, H.-C. Chou, J.S.C. Wu, D.P. Tsai, G. Mul, CO2 photoreduction using NiO/InTaO4 in optical-fiber reactor for renewable energy, Appl. Catal. A 380 (2010) 172–177. [95] P.-Y. Liou, S.-C. Chen, J.C.S. Wu, D. Liu, S. Mackintosh, M. Maroto-Valer, R. Linforth, Photocatalytic CO2 reduction using an internally illuminated monolith photoreactor, Energy Environ. Sci. 4 (2011) 1487–1494. [96] S. Yan, H. Yu, N. Wang, Z. Li, Z. Zou, Efficient coversion of CO2 and H2O into hydrocarbon fuel over ZnAl2O4-modified mesoporous ZnGaNO under visible light irradiation, Chem. Commun. 48 (2012) 1048–1050. [97] M. Subrahmanyam, S. Kaneco, N. Alonso-Vante, A screening for the photo reduction of carbon dioxide supported on metal oxide catalysts for C1 –C3 selectivity, Appl. Catal. B Environ. 23 (1999) 169–174.

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174 [98] P.-W. Pan, Y.-W. Chen, Photocatalytic reduction of carbon dioxide on NiO/InTaO4 under visible light irradiation, Catal. Commun. 8 (2007) 1546–1549. [99] C.-W. Tsai, H.M. Chen, R.-S. Liu, K. Asakura, T.-S. Chan, Ni@NiO core-shell structure-modified nitrogen-doped InTaO4 for solar-driven highly efficient CO2 reduction to methanol, J. Phys. Chem. C 115 (2011) 10182–10186. [100] D.-S. Lee, H.-J. Chen, Y.-W. Chen, Photocatalytic reduction of carbon dioxide with water using InNbO4 catalyst with NiO and Co3 O4 cocatalysts, J. Phys. Chem. Solids 73 (2012) 661–669. [101] D.-S. Lee, Y.-W. Chen, Photocatalytic reduction of carbon dioxide with water on InVO4 with NiO cocatalysts, J. CO2 Utiliz. 10 (2015) 1–6. [102] Y. Yu, L. Guo, H.H. Cao, Y. Lv, E. Wang, Y. Cao, A new Ni/Ni3 (BO3 )2 /NiO heterostructured photocatalyst with efficient reduction of CO2 into CH4 , Sep. Purif. Technol. 142 (2015) 14–17. [103] N. Ahmed, Y. Shibata, T. Taniguchi, Y. Izumi, Photocatalytic conversion of carbon dioxide into methanol using zinc–copper–M(III) (M = aluminum, gallium) layered double hydroxides, J. Catal. 209 (2011) 123–135. [104] K. Teramura, S. Iguchi, Y. Mizuno, T. Shishido, T. Tanaka, Photocatalytic conversion of CO2 in water over layered double hydroxides, J. Appl. Crystallogr. 44 (2011) 1272–1276. [105] K. Katsumata, K. Sakai, K. Ikeda, G. Carga, N. Matsushita, K. Okada, Preparation and photocatalytic reduction of CO2 on noble metal (Pt, Pd, Au) loaded Zn–Cr layered double hydroxides, Mater. Lett. 107 (2013) 138–140. [106] N. Ahmed, M. Morikawa, Y. Izumi, Photocatalytic conversion of carbon dioxide into methanol using optimized layered double hydroxide catalysts, Catal. Today 185 (2012) 263–269. [107] M. Morikawa, N. Ahmed, Y. Yoshida, Y. Izumi, Photoconversion of carbon dioxide in zinc–copper–gallium layered double hydroxides: the kinetics to hydrogen carbonate and furtherto CO/methanol, Appl. Catal. B Environ. 144 (2014) 561–569. [108] S. Kawamura, M.C. Puscasu, Y. Yoshida, Y. Izumi, G. Carja, Tailoring assemblies of plasmonic silver/gold and zinc–gallium layered double hydroxides for photocatalytic conversion of carbon dioxide using UV–visible light, Appl. Catal. A (2016), http://dx.doi.org/10.1016/j.apcata.2014.12.042 (in press). [109] M. Lv, H. Liu, Photocatalytic property and structural stability of CuAl-based layered double hydroxides, J. Solid State Chem. 227 (2015) 232–238. [110] S. Iguchi, K. Teramura, S. Hosokawa, T. Tanaka, Photocatalytic conversion of CO2 in an aqueous solution using various kinds of layered double hydroxides, Catal. Today 251 (2015) 140–144. [111] B.-J. Liu, T. Torimoto, H. Yoneyama, Photocatalytic reduction of CO2 using surface-modified CdS photocatalysts in organic solvents, J. Photochem. Photobiol. A 113 (1998) 93–97. [112] T. Baran, S. Wojtyla, A. Dibenedetto, M. Aresta, W. Macyk, Zinc sulfide functionalized with ruthenium nanoparticles for photocatalytic reduction of CO2 , Appl. Catal. B Environ. 178 (2015) 170–176. [113] H. Fujiwara, H. Hosokawa, K. Murakoshi, Y. Wada, S. Yanagida, Surface characteristics of ZnS nanocrystallites relating to their photocatalysis for CO2 reduction, Langmuir 14 (18) (1998) 5154–5159. [114] S. Yanagida, Y. Wada, K. Murakoshi, H. Fujiwara, T. Sakata, H. Mori, Photocatalytic reduction and fixation of CO2 on cadmium sulfide nanocrystallites, Stud. Surf. Sci. Catal. 114 (1998) 183–188. [115] P. Johne, H. Kisch, Photoreduction of carbon dioxide catalysed by free and supported zinc and cadmium sulphide powders, J. Photochem. Photobiol. A 111 (1-3) (1997) 223–228. [116] P. Praus, O. Kozak, K. Koci, A. Panacek, R. Dvorsky, CdS nanoparticles deposited on montmorillonite: preparation characterizationand application for photoreduction of carbon dioxide, J. Colloid Interface Sci. 360 (2011) 574–579. [117] O. Kozak, P. Praus, K. Koci, M. Klementova, Preparation and characterization of ZnS nanoparticles deposited on montmorillonite, J. Colloid Interface Sci. 352 (2010) 244–251. [118] P. Praus, M. Reli, K. Koci, L. Obalova, Photocatalytic reactions of nanocomposite of ZnS nanoparticles and montmorillonite, Appl. Surf. Sci. 275 (2013) 369–373. [119] K. Koci, L. Matejova, O. Kozak, L. Capek, V. Vales, M. Reli, P. Praus, K. Safarova, A. Kotarba, L. Obalova, ZnS/MMT nanocomposites: the effect of ZnS loading in MMT on thephotocatalytic reduction of carbon dioxide, Appl. Catal. B Environ. 158–159 (2014) 410–417. [120] K. Koci, M. Reli, O. Kozak, Z. Lacny, D. Placha, P. Praus, L. Obalova, Influence of reactor geometry on the yield of CO2 photocatalytic reduction, Catal. Today 176 (2011) 212–214. [121] B.R. Eggins, P.K.J. Robertson, E.P. Murphy, E. Woods, J.T.S. Irvine, Factors affecting the photoelectrochemical fixation of carbon dioxide with semiconductor colloids, J. Photochem. Photobiol. A 118 (1998) 31–40. [122] B.R. Eggins, J.T.S. Irvine, E.P. Murphy, J. Grimshaw, Formation of two-carbon acids from carbon dioxide by photoreduction on cadmium sulphide, J. Chem. Soc. Chem. Comm. 16 (1988) 1123–1124. [123] J.T.S. Irvine, B.R. Eggins, J. Grimshaw, Solar energy fixation of carbon dioxide via cadmium sulphide and other semiconductor photocatalysts, Sol. Energy 45 (1) (1990) 27–33. [124] B.R. Eggins, P.K.J. Robertson, J.H. Stewart, E. Woods, Photoreduction of carbon dioxide on zinc sulfide to give four-carbon and two-carbon acids, J. Chem. Soc. Chem. Commun. 4 (1993) 349–350. [125] M. Kanemoto, T. Shiragami, C. Pac, S. Yanagida, Semiconductor photocatalysis. Effective photoreduction of carbon dioxide catalyzed by zinc

[126]

[127]

[128]

[129]

[130] [131] [132]

[133]

[134]

[135]

[136]

[137]

[138]

[139]

[140]

[141] [142] [143] [144]

[145]

[146]

[147]

[148]

[149]

[150]

[151] [152]

[153]

173

sulfide quantum crystallites with low density of surface defects, J. Phys. Chem. 96 (1992) 3521–3526. S. Kuwabata, K. Nishida, R. Tsuda, H. Inoue, H. Yoneyama, Photochemical reduction of carbon dioxide to methanol using ZnS microcrystallite as a photocatalyst in the presence of methanol dehydrogenase, J. Electrochem. Soc. 141 (6) (1994) 1498–1503. X.V. Zhang, S.P. Ellery, C.M. Friend, H.D. Holland, F.M. Michel, M.A.A. Schoonen, S.T. Martin, Photodriven reduction and oxidation reactions on colloidal semiconductor particles: implications for prebiotic synthesis, J. Photochem. Photobiol. A 185 (2007) 301–311. J. Chen, F. Xin, S. Qin, X. Yin, Photocatalytically reducing CO2 to methyl formate in methanol over ZnS and Ni-doped ZnS photocatalysts, Chem. Eng. J. 230 (2013) 506–512. X. Li, J. Chen, H. Li, J. Li, Y. Xu, Y. Liu, J. Zhou, Photoreduction of CO2 to methanol over Bi2 S3 /CdS photocatalyst under visible light irradiation, J. Nat. Gas Chem. 20 (2011) 413–417. K. Ohta, T. Ogawa, T. Mizuno, Abiological formation of formic acid on rocks in nature, Appl. Geochem. 15 (2000) 91–95. N. Getoff, Possibilities on the radiation-induced incorporation of CO2 and CO into organic compounds, Int. J. Hydrogen Energy 19 (8) (1994) 667–672. H. Li, Y. Lei, Y. Huang, Y. Fang, Y. Xu, L. Zhu, X. Li, Photocatalytic reduction of carbon dioxide to methanol by Cu2 O/SiC nanocrystallite under visible light irradiation, J. Nat. Gas Chem. 20 (2011) 145–150. W. Jiang, X. Yin, F. Xin, Y. Bi, Y. Liu, X. Li, Preparation of CdIn2 S4 microspheres and application for photocatalytic reduction of carbon dioxide, Appl. Surf. Sci. 288 (2014) 138–142. S.W. Yang, G.T. Pan, T.C.-K. Yang, C.-C. Chen, H.-C. Chiang, The photosynthesis of methanol on 1D ordered Zn: cuInS2 nanoarrays, J. Taiwan Inst. Chem. E 45 (2014) 1509–1515. X. Zhang, J. Li, X. Lu, C. Tang, G. Lu, Visible light induced CO2 reduction and Rh B decolorization over electrostatic-assembled AgBr/palygorskite, J. Colloid Interface Sci. 377 (2012) 277–283. J.-Y. Liu, B. Garg, Y.-C. Ling, Cux Agy Inz Znk Sm solid solutions customized with RuO2 or Rh1.32 Cr0.66 O3 co-catalyst display visible light-driven catalytic activity for CO2 reduction to CH3 OH, Green Chem. 13 (2011) 2029–2031. H. Peng, J. Lu, C. Wu, Z. Yang, H. Chen, W. Song, P. Li, H. Yin, Co-doped MoS2 NPs with matched energy band and low overpotential high efficiently convert CO2 to methanol, Appl. Surf. Sci. 353 (2015) 1003–1012. M.S. Khan, M.N. Ashiq, M.F. Ehsan, T. He, S. Ijaz, Controlled synthesis of cobalt telluride superstructures for the visiblelight photo-conversion of carbon dioxide into methane, Appl. Catal. A Gen. 487 (2014) 202–209. A. Wang, X. Li, Y. Zhao, W. Wu, J. Chen, H. Meng, Preparation and characterizations of Cu2 O/reduced graphene oxide nanocomposites with high photo-catalytic performances, Powder Technol. 261 (2014) 42–48. J. Chen, J. Shi, X. Wang, H. Cui, M. Fu, Recent progress in the preparation and application of semiconductor/graphene composite photocatalysts, Chin. J. Catal. 34 (2013) 621–640. K. Putri, W.-J. Ong, W.S. Chang, S.-P. Chai, Heteroatom doped graphene in photocatalysis: a review, Appl. Surf. Sci. 358 (2015) 2–14. J. Low, J. Yu, W. Ho, Graphene-based photocatalysts for CO2 reduction to solar fuel, J. Phys. Chem. Lett. 6 (2015) 4244–4251. Q. Xiang, B. Cheng, J. Yu, Graphene−based photocatalysts for solar-fuel generation, Angew. Chem. Int. Ed. 54 (2015) 11350–11366. W. Zhang, F. Dong, W. Zhang, Capture of atmospheric CO2 into (BiO)2 CO3 /graphene or graphene oxide nanocomposites with enhanced photocatalytic performance, Appl. Surf. Sci. 358 (A) (2015) 75–83. A. Wang, S. Shen, Y. Zhao, W. Wu, Preparation and characterizations of BiVO4 /reduced graphene oxide nanocomposites with higher visible light reduction activities, J. Colloid Interface Sci. 445 (2015) 330–336. W. Li, X. Jiang, H. Yang, Q. Liu, Solvothermal synthesis and enhanced CO2 adsorption ability of mesoporous graphene oxide-ZnO nanocomposite, Appl. Surf. Sci. 356 (2015) 812–822. L. Zhang, N. Li, H. Jiu, G. Qi, Y. Huang, ZnO-reduced graphene oxide nanocomposites as efficient photocatalysts for photocatalytic reduction of CO2 , Ceram. Int. 41 (5) (2015) 6256–6262. P.-Q. Wang, Y. Bai, P.-Y. Luo, J.-Y. Liu, Graphene–WO3 nanobelt composite: elevated conduction band toward photocatalytic reduction of CO2 into hydrocarbon fuels, Catal. Commun. 38 (2013) 82–85. P. Kumar, H.P. Mungse, S. Cordier, R. Boukherroub, O.P. Khatri, S.L. Jain, Hexamolybdenum clusters supported on graphene oxide: visible-light induced photocatalytic reduction of carbon dioxide into methanol, Carbon 94 (2015) 91–100. S. Sahu, L. Cao, M.J. Meziani, C.E. Bunker, K.A.S. Fernando, P. Wang, Y.-P. Sun, Carbon dioxide photoconversion driven by visible-light excitation of small carbon nanoparticles in various configurations, Chem. Phys. Lett. 634 (2015) 122–128. Q. Li, X. Li, S. Wageh, A.A. Al-Ghamdi, J. Yu, CdS/graphene nanocomposite photocatalysts, Adv. Energy Mater. 5 (2015) 1500010. J. Yu, J. Jin, B. Cheng, M. Jaroniec, A noble metal-free reduced oxide-CdS nanorod composite for the enhanced visible-light photocatalytic reduction of CO2 to solar fuel, Mater. Chem. A 2 (2014) 3407–3416. M.A. Asi, L. Zhu, C. He, V.K. Sharma, D. Shu, S. Li, J. Yang, Y. Xiong, Visible – light – harvesting reduction of CO2 to chemical fuels with plasmonic Ag@AgBr/CNT nanocomposites, Catal. Today 216 (2013) 268–275.

174

A. Nikokavoura, C. Trapalis / Applied Surface Science 391 (2017) 149–174

[154] Z. Fang, S. Li, Y. Gong, W. Liao, S. Tian, C. Shan, C. He, Comparison of catalytic activity of carbon-based AgBr nanocomposites for conversion of CO2 under visible light, J. Saudi Chem. Soc. 18 (4) (2014) 299–307. [155] T. Ohno, N. Murakami, T. Koyanagi, Y. Yang, Photocatalytic reduction of CO2 over a hybrid photocatalyst composed of WO3 and graphitic carbon nitride (g-C3 N4 ) under visible light, J.CO2 Utiliz. 6 (2014) 17–25. [156] J. Mao, T. Peng, X. Zhang, K. Li, L. Ye, L. Zan, Effect of graphitic carbon nitride microstructures on the activity and selectivity of photocatalytic CO2 reduction under visible light, Catal. Sci. Technol. 3 (2013) 1253–1260. [157] J. Qin, S. Wang, H. Ren, Y. Hou, X. Wang, Photocatalytic reduction of CO2 by graphitic carbon nitride polymers derived from urea and barbituric acid, Appl. Catal. B Environ. 179 (2015) 1–8. [158] I. Papailias, T. Giannakopoulou, N. Todorova, D. Demotikali, T. Vaimakis, C. Trapalis, Effect of processing temperature on structure and photocatalytic properties of g-C3 N4 , Appl. Surf. Sci. 358 (2015) 278–286. [159] S. Ye, R. Wang, M.-Z. Wu, Y.-P. Yuan, A review on g-C3 N4 for photocatalytic water splitting and CO2 reduction, Appl. Surf. Sci. 358 (2015) 15–27. [160] P. Zhou, J. Yu, M. Jaroniec, All-solid-state Z-Scheme photocatalytic systems, Adv. Mater. 26 (2014) 4920–4935. [161] S. Cao, J. Low, J. Yu, M. Jaroniec, Polymeric photocatalysts based on graphitic carbon nitride, Adv. Mater. 27 (2015) 215–2176. [162] Y. Wang, F. Wang, Y. Zuo, X. Zhang, L.-F. Cui, Simple synthesis of ordered cubic mesoporous graphitic carbon nitride by chemical vapor deposition method using melamine, Mater. Lett. 136 (2014) 271–273. [163] J. Yu, K. Wang, W. Xiao, B. Cheng, Photocatalytic reduction of CO2 into hydrocarbon solar fuels over g-C3 N4 -Pt nanocomposite photocatalyst, Phys. Chem. Chem. Phys. 16 (2014) 11492–11501. [164] K. Wang, Q. Li, B. Liu, B. Cheng, W. Ho, J. Yu, Sulfur-doped g-C3 N4 with enhanced photocatalytic CO2 -reduction performance, Appl. Catal. B Environ. 176–177 (2015) 44–52. [165] W.-J. Ong, K.P. Lutfi, L.L. Tan, S.-P. Chai, S.-T. Yong, HeterostructuredAgX/g-C3 N4 (X = Cl and Br) nanocomposites via a sonication-assisted deposition-precipitation approach: emerging role of halide ions in the synergistic photocatalytic reduction of carbon dioxide, Appl. Catal. B Environ. (2016), http://dx.doi.org/10.1016/j.apcatb.2015.06. 053 (in press). [166] W.-J. Ong, L.-L. Tan, S.-P. Chai, S.-T. Yong, A.R. Mohamed, Surface charge modification via protonation of graphitic carbon nitride (g-C3 N4 ) for electrostatic self-assembly construction of 2D/2D reduced graphene oxide (rGO)/g-C3 N4 nanostructures toward enhanced photocatalytic reduction of carbon dioxide to methane, Nano Energy 13 (2015) 757–770. [167] W. Yu, D. Xu, T. Peng, Enhanced photocatalytic activity of g-C3 N4 for selective CO2 reduction to CH3 OH via facile coupling of ZnO: a direct Zscheme mechanism, J. Mater. Chem. A 3 (2015) 19936–19947.

[168] Q. Huang, J. Yu, S. Cao, C. Cui, B. Cheng, Efficient photocatalytic reduction of CO2 by amine-functionalizedg-C3 N4 , Appl. Surf. Sci. 358 (2015) 350–355. [169] Y. He, L. Zhang, M. Fan, X. W.ang, M.L. Walbridge, Q. Nong, Y. Wu, L. Zhao, Z-scheme SnO2-x /g-C3 N4 composite as an efficient photocatalyst for dye degradation and photocatalytic CO2 reduction, Sol. Energy Mater. Sol. Cells 137 (2015) 175–184. [170] Y.-P. Yuan, S.-W. Cao, Y.-S. Liao, L.-S. Yin, C. Xue, Red phosphor/g-C3 N4 heterojunction with enhanced photocatalytic activities for solar fuels production, Appl. Catal. B Environ. 140–141 (2013) 164–168. [171] S.-W. Cao, X.-F. Liu, Y.-P. Yuan, Z.-Y. Zhang, Y.-S. Liao, J. Fang, S.C.J. Loo, T.C. Sum, C. Xue, Solar-to-fuels conversion over In2 O3 /g-C3 N4 hybrid photocatalysts, Appl. Catal. B Environ. 147 (2014) 940–946. [172] Y. He, Y. Wang, L. Zhang, B. Teng, M. Fan, High-efficiency conversion of CO2 to fuel over ZnO/g-C3 N4 photocatalyst, Appl. Catal. B Environ. 168–169 (2015) 1–8. [173] L. Zhao, L. Zhang, H. Lin, Q. Nong, M. Cui, Y. Wu, Y. He, Fabrication and characterization of hollow CdMoO4 coupled g-C3 N4 heterojunction with enhanced photocatalytic activity, J. Hazard. Mater. 299 (2015) 333–342. [174] J. Li, D. Luo, C. Yang, S. He, S. Chen, J. Lin, L. Zhu, X. Li, Copper(II) imidazolate frameworks as highly efficient photocatalysts for reduction of CO2 into methanol under visible light irradiation, J. Solid State Chem. 203 (2013) 154–159. [175] B. Wang, A.P. Cote, H. Furukawa, M. Keeffe, O.M. Yaghi, Colossal cages in zeolitic imidazolate frameworks as selective carbon dioxide reservoirs, Nature 453 (2008) 207–212. [176] M. Matsuoka, M. Anpo, Local structures excited states, and photocatalytic reactivities of highly dispersed catalysts constructed within zeolites, J. Photochem. Photobiol. C: Photochem. Rev. 3 (2003) 225–252. [177] C.-C. Wang, Y.-Q. Zhang, J. Li, P. Wang, Photocatalytic CO2 reduction in metal?organic frameworks: a mini review, J. Mol. Struct. 1083 (2015) 127–136. [178] S. Wang, W. Yao, J. Lin, Z. Ding, X. Wang, Cobalt imidazolate metal–organic frameworks photosplit CO2 under mild reaction conditions, Angew. Chem. Int. Ed. 53 (4) (2014) 1034–1038. [179] S. Wang, X. Wang, Photocatalytic CO2 reduction by CdS promoted with a zeolitic imidazolate framework, Appl. Catal. B Environ. 162 (2015) 494–500. [180] S. Wang, J. Lin, X. Wang, Semiconductor–redox catalysis promoted by metal–organic frameworks for CO2 reduction, Phys. Chem. Chem. Phys. 16 (2014) 14656–14660. [181] Q. Liu, Z.-X. Low, L. Li, A. Razmjou, K. Wang, J. Yao, H. Wang, ZIF-8/Zn2 GeO4 nanorods with an enhanced CO2 adsorption property in an aqueous medium for photocatalytic synthesis of liquid fuel, J. Mater. Chem. A 1 (2013) 11563–11569.