Recent advancements in g-C₃N₄-based photocatalysts for photocatalytic CO₂ reduction: a mini review

Abstract

Carbon dioxide (CO₂) is a very important micro-molecular resource. Using CO₂ captured from the atmosphere for high-output synthesis of chemicals as raw materials has great significance and potential for various industrial applications. Since the industrial revolution in the 18^th century, manmade CO₂ emission has increased by 45%, which negatively impacts the planetary climate by the so-called greenhouse effect. Therefore, high-efficiency photocatalysis and photocatalysts for CO₂ conversion have become the most important challenges and milestones throughout the world. In consideration of this, various catalysts have been explored. Among these, graphitic carbon nitride (g-C₃N₄) as a semiconductor is emerging as a highly promising photocatalyst for removing CO₂ from the atmosphere. Moreover, due to its excellent chemical stability and unique band structure, g-C₃N₄ has exhibited significant application potential for photocatalysis. This review summarizes the advancements that have been made in the synthesis and photocatalytic applications of g-C₃N₄-based catalysts for CO₂ reduction in recent years and explains the future challenges and prospects in this vital area of research.

---

Runlu Liu

Mr Runlu Liu received his Bachelor degree in Materials Science and Engineering from Shanghai Jiao Tong University, China in 2019. He is presently a postgraduate student at the School of Materials Science and Engineering in Shanghai Jiao Tong University, under the supervision of Prof. Shenmin Zhu. He is currently working on the research of g-C3N4-based nanocomposites and their applications in photocatalysis.

Yao Li

Dr Yao Li received his PhD in Material Science from Shanghai Jiao Tong University, China in 2015 under the supervision of Prof. Di Zhang and Prof. Shenmin Zhu. He joined as a research associate at Shanghai Jiao Tong University at 2015. He is working on the research of bioinspired material and carbon matrix material, as well as their application in environmental treatment and energy storage.

Shenmin Zhu

Prof. Shenmin Zhu received her PhD degree from Shanghai Jiao Tong University in 2001. She is presently a professor at the School of Materials Science and Engineering, State Key Lab of Metal Matrix Composites, Shanghai Jiao Tong University. Her current fields of interest are graphene-based functional materials, porous carbon, and bioinspired photonic crystals with stimuli-responsive properties.

---

1. Introduction

Carbon dioxide (CO₂) is a major greenhouse gas that contributes approximately 20% to the greenhouse effect. Its emissions are increasing and are causing a significant negative influence on the overall climate and life on our planet. It was reported that the annual global carbon dioxide emissions have exceeded 30 billion tons, and only 1% of this amount is being removed annually.¹ However, these emissions are not considered to be solely waste because CO₂ is a very important micro-molecular resource for many industrial chemicals, such as methane, methyl alcohol, and dimethylformamide (DMF). Artificial synthesis of industrial-scale chemicals with CO₂ captured from the atmosphere not only removes the harmful CO₂, but also produces valuable chemical products for industries.

Since 1970, it has been a challenge worldwide to develop a facile and efficient method for the photocatalytic reduction of CO₂. Three types of CO₂ photoreduction systems have been developed, i.e., the semiconductor photocatalytic system, the metal complex photocatalytic system, and the enzyme photocatalytic system.² In 1979, Fujishima et al. reported photocatalytic CO₂ reduction using semiconductors TiO₂ and CdS as photocatalysts, and these were among the earliest photocatalytic systems for CO₂ reduction.³ In 1983, Lehn et al. reported for the first time a CO₂ reduction system using fac-Re(bpy)(CO)₃Cl as a photocatalyst, and it exhibited high selectivity and high quantum yield.⁴ In 1986, Willner et al. reported a photocatalytic CO₂ reduction system using enzymes as catalysts.⁵

Since their discoveries, all three types of CO₂ photoreduction systems have been significantly developed. Due to the high substrate specificity, high conversion rates, excessive sensitivity, high cost, and narrow applicability of the enzymatic catalysts, their broad utilization has been limited.⁶ Thus, most of the interest has been focused on semiconductor or metal complex-based catalysts. In general, metal complex systems are based on homogeneous organic photocatalysts. Because all the metal atoms can serve as active sites, the catalytic efficiency and selectivity of the metal complex systems are sensitive to energy distribution. Moreover, due to the homogeneous structure, there are few side reactions accompanying metal complex photocatalysis.^7,8 However, it is difficult to separate reactants from products because of the homogeneous structure, which limits the cyclic utilization and can be disadvantageous for sustainable development. In comparison, semiconductor-based photocatalysts are mostly heterogeneous catalysts, and with their use, it is easier to separate products from the reaction system.⁹ Furthermore, the photocatalytic performance can be improved through various structural and chemical modifications of the photocatalyst, e.g., porous semiconductor photocatalysts, doping^10–14 or compositing of a semiconductor photocatalyst with other materials.^15–17 Thus, there has been great interest and widespread usage of semiconductor photocatalysis during recent years.

Among the various types of semiconductors, graphitic carbon nitride (g-C₃N₄) is considered promising because it is composed of triazine rings and tri-s-triazine (heptazine) rings as its basic unit, and it forms a two-dimensional lamellar structure (Fig. 1).¹⁸ Because of the structure, it is insoluble in most solvents such as water, ethanol, ether, toluene, DMF, and tetrahydrofuran (THF).¹⁹ g-C₃N₄ is a typical polymer semiconductor with a band gap of 2.7 eV that can absorb blue-violet light with a wavelength less than 475 nm in the solar spectrum,²⁰ and thus, it is a potential candidate for photocatalysis.^21,22 In 2009, Wang et al. for the first time reported photocatalytic hydrogen production using g-C₃N₄ as a photocatalyst.²³ In 2013, Peng et al. used g-C₃N₄ in CO₂ photoreduction.²⁴ This review will focus on the present pathways for the applications of g-C₃N₄ in the field of photocatalytic CO₂ reduction. Finally, the conclusion of the article addresses the current challenges and future development of g-C₃N₄-based photocatalysts.

Fig. 1 s-Triazine (left) and tri-s-triazine (right) as tectons of g-C₃N₄.¹⁸ Copyright 2012 WILEY-VCH.

2. Fundamentals of photocatalytic CO₂ reduction on semiconductors

Photocatalytic CO₂ reduction is a process where light energy is used as the input energy, and the product can be controlled by changing the reaction conditions and types of catalysts.^25,26 The common reaction pathways are listed as follows.

CO₂ + 2H⁺ + 2e⁻ → HCOOH, E⁰ = −0.61 V (I)

CO₂ + 2H⁺ + 2e⁻ → CO + H₂O, E⁰ = −0.53 V (II)

CO₂ + 4H⁺ + 4e⁻ → HCHO + H₂O, E⁰ = −0.48 V (III)

CO₂ + 6H⁺ + 6e⁻ → CH₃OH + H₂O, E⁰ = −0.38 V (IV)

CO₂ + 8H⁺ + 8e⁻ → CH₄ + 2H₂O, E⁰ = −0.24 V (V)

The principle of photocatalytic CO₂ reduction consists of five main steps: light adsorption (i), charge separation (ii), CO₂ adsorption (iii), surface redox reaction (iv), and product desorption (v).²⁷ To trigger the CO₂ reduction reaction, the conduction band edge of the photocatalysts should be more negative than the reaction potential, and they are listed above (vs. NHE).²⁸ To achieve efficient generation and transfer of photo-excited electrons and holes, certain modifications of photocatalysts are required. To increase the effectiveness of the process, a semiconductor-based photocatalyst should be designed considering these issues.

First, a semiconductor with a band gap of 2.0–3.0 eV is ideal for visible light adsorption.²⁷ Second, the separation of electron–hole pairs is sufficiently large so that the recombination of electrons and holes is limited.³⁰ Third, to enhance the adsorption of CO₂, the surface area of photocatalysts should be enlarged, which can be accomplished by providing more active sites or introducing alkaline components into the photocatalysts.³¹ Fourth, to improve the efficiency of surface redox reactions, one can introduce a cocatalyst or prepare composites to enhance the charge transfer rate. Finally, accurate timing of the release of the products is necessary to increase the conversion and reaction rate.

3. g-C₃N₄ as photocatalyst

A good photocatalyst requires a suitable energy band structure to ensure light harvesting ability and efficient charge transfer. Among varieties of photocatalysts, oxide semiconductors usually have too large band gap energies to sufficiently utilize incident light, while the chalcogenide semiconductors often have a narrow band gap, which results in easy recombination of the photogenerated charge carriers (Fig. 2). In comparison, g-C₃N₄ is a non-metal semiconductor with a band gap of 2.7 eV and suitable conduction band (CB) and valence band (VB) positions, enabling it to absorb visible light. Moreover, g-C₃N₄ can be prepared by facile methods, and the microstructure of g-C₃N₄ can be easily adjusted to create pores as active sites or enlarge the surface area for more optimal light absorption. Because of these advantages, g-C₃N₄ is a superior photocatalyst for visible-light-driven CO₂ reduction compared to TiO₂ and other common photocatalysts. However, because the 2.7 eV band gap of g-C₃N₄ is insufficiently large, the separated electron–hole pairs tend to recombine, thus limiting the effective separation of charges and subsequent redox reactions. Although it has numerous Lewis base sites, e.g., Brønsted base sites and nucleophile sites that are favourable for CO₂ adsorption, bulk g-C₃N₄ prepared by direct calcination has a relatively small surface area. Therefore, several modifications including structural tuning,^31–37 elemental doping,^38–45 addition of co-catalyst,^46–60 and compositing^61–70 have been used to increase the adsorption and obtain more effective charge separation.

Fig. 2 Band gap energies and CB (green) and VB (red) edge positions of the selected photocatalysts with respect to the vacuum level and NHE. The two dashed lines indicate the water redox reaction potentials.²⁹ Copyright 2016 WILEY-VCH.

3.1 g-C₃N₄ with nanostructures and/or defects

It is expected that g-C₃N₄ has good potential for CO₂ fixation and activation. However, the bulk form obtained from some precursors has a small surface area, and thus, has low adsorption efficiency and catalytic activity. To alleviate this drawback, researchers have taken specific measures to increase the surface area and improve the functionality of g-C₃N₄. In this regard, Zhang et al. applied porous structured g-C₃N₄ to a CO₂ photocatalytic reduction to CO under visible light irradiation.³¹ Moreover, the g-C₃N₄ was synthesized by calcination, using either melamine or melamine hydrochloride as the precursor. It was found that the latter has a surface area 39 times greater than bulk g-C₃N₄ due to the porous structure, but the performance of CO₂ photoreduction was not obviously improved. This occurred because the porous structure not only endows g-C₃N₄ with a higher surface area but also enlarges the band gap, which resulted in more difficult electron excitation and the lower yield of CO₂. Although it failed to improve the performance of CO₂ photoreduction, there was still direct significance to this pioneering research in that the structural tuning of g-C₃N₄ resulted in great improvement of the photoreactivity.

Thus far, different varieties of g-C₃N₄ nanostructures have been developed, such as porous structures, 1D nanorods, nanowires, 2D nanosheets, and 3D nanostructures.^32–37 As a typical example, Zheng and Wang et al. imprinted helical g-C₃N₄ nanorods with chiral silicon dioxide as templates (Fig. 3).³² This helical structure caused multiple reflections of incident light, leading to an improved light-harvesting capability across the entire optical spectrum. Although the band gap increased from 2.66 eV to 2.75 eV, the helical g-C₃N₄ still exhibited an overall enhanced optical absorption compared to bulk g-C₃N₄. A lower photoluminescence (PL) intensity and a stronger PL quenching was measured for the helical g-C₃N₄ nanorods as compared to the bulk g-C₃N₄, which indicated that the recombination of photogenerated charge carriers in the helical g-C₃N₄ nanorods was suppressed.⁷¹ The CO production and selectivity of the helical g-C₃N₄ system during visible-light CO₂ reduction reached 8.9 μmol and 96.7%, respectively. In recent years, there has been much experimentation with 2D g-C₃N₄ nanosheets because of their large specific surface area, which is advantageous for light harvesting and gas adsorption. Cao et al. prepared ultra-thin g-C₃N₄ nanosheets through a stepwise NH₃-mediated thermal exfoliation approach.³³ The resulting nanosheets were approximately 3 nm thick and were produced with a hierarchical structure due to the amine-induced assembly. This structure endowed g-C₃N₄ nanosheets with a much higher specific surface area, abundant active sites, shorter diffusion distance of charge carriers, and increased CO₂ adsorption. Therefore, the yields of CH₄ and CH₃OH with g-C₃N₄ nanosheets as catalyst were 9.93 and 5.34 times higher than that with bulk g-C₃N₄ as catalyst, respectively.

Fig. 3 Synthetic process for helical nanorod-like graphitic carbon nitride (HR-CN) based on chiral mesoporous silica (CMS) as the template and cyanamide (CY) as the precursor.³² Copyright 2014, Wiley-VCH.

Apart from structural tuning, the introduction of defects is also a useful method for promoting the photoactivity of g-C₃N₄. Zhang et al. introduced carbon vacancies into synthesized g-C₃N₄ with urea and heat treatment under an NH₃ atmosphere to enhance CO photogeneration.³⁴ The g-C₃N₄ with enriched C vacancies exhibited a CO evolution rate of 4.18 μmol g⁻¹ h⁻¹, which was 2.3 times higher than the evolution rate of pristine g-C₃N₄. The improved CO₂ reduction performance was ascribed to greater CO₂ adsorption and activation, an upshifted conduction band, increased charge carrier concentration, and prolonged lifetime. Furthermore, the reactive oxygen species (ROS) and electrochemical impedance spectroscopy (EIS) indicated that the C vacancies could weaken the exciton effect and elevate the charge carrier generation.

In another study, Chai et al. introduced N vacancies into ultrathin g-C₃N₄ nanosheets using NH₄Cl as a dynamic gas template.³⁵ The NH₄Cl decomposed into NH₃ and HCl gases upon heating, which formed a bubble film on the surface of the polymerized g-C₃N₄ nanosheets. As the π–π interlayer interactions in the g-C₃N₄ nanosheets became diminished with the expansion of the bubble film, an ultrathin bubble g-C₃N₄ film formed, and N vacancies were introduced afterwards by a post-treatment under a reducing atmosphere. The few-atomic-layered structure enlarged the exposed surface area, and this benefited charge carrier transportation. Additionally, N vacancies induced a mid-gap state, which facilitated multi-electron excitation in a two-step process and extended the optical absorption of the g-C₃N₄ nanosheets to the near infrared region, and thus, the utilization of solar energy was enhanced (Fig. 4). Furthermore, N vacancies could also result in an aftereffect where the decay lifetime was prolonged because the mid-gap state retarded the radiative recombination rate. The g-C₃N₄ nanosheets with N vacancies exhibited a 5.14-fold higher CH₄ evolution rate than pristine g-C₃N₄. This pioneering work was intriguing and inspiring because it highlighted the synergistic effect of morphology control and vacancy modulation on photocatalytic CO₂ reduction.

Fig. 4 Schematic illustration of nitrogen defect-modified g-C₃N₄ atomic layers for photocatalytic CH₄ evolution from CO₂.³⁵ Copyright 2019, Royal Society of Chemistry.

Both structural adjustment and defect tuning are meant to change the energy band structure of g-C₃N₄ to enhance the light-harvesting ability and charge transfer efficiency, and they are the most convenient methods to improve the photocatalytic performance of g-C₃N₄.

3.2 Non-metal doped g-C₃N₄

Pristine g-C₃N₄ exhibits a poor performance due to low CO₂ absorption and poor utilization of photo-excited charge carriers because of the deficiency of electron donation sites and unoptimized electronic structure.³⁷ To address these challenges, heteroatom doping was developed to tune the electronic structure of g-C₃N₄. The doped elements included mainly B, C, O, P, and S.^38–41

In g-C₃N₄, N atoms are active sites for photocatalytic reactions, and the electrons are mainly localized around N atoms. Under light irradiation, transfer of excited electrons from N atoms to C atoms is difficult due to the localized electron structure.^72–74 Moreover, the excited electrons are expected to transfer back for redox reaction, which increases the chances of recombination of the charge carriers. To address these problems, Liu et al. prepared boron-doped g-C₃N₄ by a one-step calcination using boric acid and urea as precursors.³⁸ B atoms were doped into cavities between adjacent tri-s-triazine units through coordination with N atoms. The density functional theory (DFT) studies demonstrated that doping of B introduced a new electron excitation pathway from N (2p_x, 2p_y) to B (2p_x, 2p_y), which improved the charge transfer and localization, and thus the reaction dynamics. Furthermore, the doped B atoms altered the gas adsorption on the surface of g-C₃N₄ so that they could subsequently act as active sites. Consequently, the optimal B-doped g-C₃N₄ resulted in a 32-fold higher yield of CH₄ than that of pure g-C₃N₄.

In another work, Tian et al. synthesized phosphorus-doped g-C₃N₄ nanotubes by calcinating a mixture of melamine and sodium hypophosphite monohydrate.³⁹ The nanotube structure attained a much larger surface area than that of g-C₃N₄ nanosheets. The doping of phosphorus resulted in a shifting down of the conduction band and valence band of g-C₃N₄, and the band gap was narrowed by 0.12 eV (Fig. 5), which enhanced the optical absorption. The phosphorus doping also created additional amino groups in the surface of g-C₃N₄, which increased CO₂ adsorption due to the acid–base interaction. As a result, the gas yields were 9.48 μmol g⁻¹ for CO and 7.24 μmol g⁻¹ for CH₄, which were 3.12 and 13.9 times the yield of pristine g-C₃N₄, respectively. Similarly, Yu et al. fabricated sulfur-doped g-C₃N₄ (TCN) by a facile calcinating method using thiourea as the precursor.⁴⁰ According to UV-Vis diffuse reflection spectra (UV-Vis DRS), the band gap of TCN was 2.63 eV, which was 0.07 eV narrower than that of un-doped g-C₃N₄ (MCN). This narrow band gap enabled TCN to absorb more solar energy and produce more charge carriers, thus improving the photocatalytic performance. According to the theoretical calculation, however, the band gaps of TCN and MCN were the same, but an impurity level was introduced due to sulfur doping on TCN. The photoelectrons could jump from the VB to the impurity level or from the impurity level to the CB. As a result, the CH₃OH yield of TCN was 2.5 times greater than that of MCN.

Fig. 5 The electronic structure and all reaction reduction potentials of hydrogen evolution and CO₂ conversion into CO and CH₄.³⁹ Copyright 2018, American Chemical Society.

It is worth noting that many studies have been recently conducted on metal and non-metal co-doped g-C₃N₄ for CO₂ photoreduction. Metal elements such as K, Na, and Co can improve the photocatalytic performance through a similar mechanism.^42–45 Zheng et al. prepared N vacancy-rich g-C₃N₄ co-doped with B and K atoms through a one-step method.⁴² In this system, K functioned as an electron donor and promoted interlayer electron transfer, B assisted in maintaining a high reduction potential and counteracted the drawbacks of K, and N vacancies lowered the conduction band minimum (CBM) and promoted CO₂ absorption. The synergistic effect of the multiple modifications significantly improved the CO₂ reduction performance. The CH₄ and CO produced in 5 h were 5.93 μmol g⁻¹ and 3.16 μmol g⁻¹, respectively, and 161% and 527% of the production of the pristine g-C₃N₄ occurred under the same conditions. It further suggested that the combination of doping and vacancies enhanced the photocatalytic performance of g-C₃N₄, and provides a new strategy for photocatalyst modification.

Overall, the non-metal doping increased the CO₂ photoreduction activity of g-C₃N₄ by tuning the energy band structure, which manifested as a narrower band gap that provided more effective light absorption in a wider range of spectra, especially in the visible light region. The non-metal doping could also introduce an impurity level or a new electron excitation pathway to g-C₃N₄, which facilitated the charge transfer. Moreover, the doping process might introduce surface groups, which improved CO₂ adsorption due to the acid–base interaction.

3.3 g-C₃N₄ with co-catalyst

In photocatalysis, suitable co-catalysts can serve as active sites for surface reactions and improve charge separation and charge utilization.⁷⁵ For a semiconductor such as g-C₃N₄, frequently used co-catalysts include metal, carbon materials, reduction co-catalyst, and double co-catalyst.⁷⁶

3.3.1 Metal co-catalyst. Generally, a metal co-catalyst refers to noble metals such as Au, Ag, and Pt. By depositing the noble metals on a photocatalyst, a Schottky barrier can be formed on the interface, and the metal nanoparticles can become electron traps to facilitate the separation of photogenerated charges.⁷⁷ Pt is one of the most common co-catalysts used in CO₂ reduction because of its great methane selectivity in a gas–solid system.⁷⁸ Yu et al. prepared g-C₃N₄ by direct calcination with thiourea as a precursor and Pt nanoparticles that were deposited for CO₂ photoreduction.⁴⁶ The deposited Pt nanoparticles absorbed nearly all the incident light, and thus increased the visible light absorption without changing the band gap of g-C₃N₄.⁷⁹ Moreover, the introduction of Pt nanoparticles promoted the generated photoelectrons to transfer from g-C₃N₄ to Pt and react with CO₂. The yield of CH₄ initially increased as the amount of deposited Pt increased from 0 to 1%, and then decreased as the loading of Pt increased to 2% or higher. Similar observations have also been reported by Chai et al. (Fig. 6).⁴⁷ This relative reduction in the production of CH₄ was attributed to the excess Pt that increased the negative charges on the surface of the g-C₃N₄ and promoted the recombination of electrons and holes. Another possibility is that excessive or redundant Pt covers the active sites on the surface of g-C₃N₄ and shields them from light. In another study, Du et al. substituted Pt nanoparticles with individual Pt atoms and found that the Pt/g-C₃N₄ catalyst preferred to reduce CO₂ to CH₄, with a rate-determining barrier of 1.16 eV and greatly enhanced visible light absorption.⁴⁸

Fig. 6 Photogenerated charge transfer process in the Pt/CN system for the reduction of CO₂ to CH₄ under visible light irradiation.⁴⁷ Copyright 2014, Royal Society of Chemistry.

Other noble metals are also frequently employed to modify g-C₃N₄.^49,50 Currently, there are few CO₂ reduction systems using non-noble metals.^51,52 However, according to a DFT calculation, Mo could be an efficient co-catalyst on g-C₃N₄ for CO₂–CO conversion.⁵³

3.3.2 Carbon materials. Carbon materials, e.g., carbon nanodots, graphene, graphene oxide, and other similar materials, have also been employed as co-catalysts to facilitate the electron transfer on g-C₃N₄. Chai et al. fabricated sandwich-like graphene-g-C₃N₄ nanocomposites (GCN) by one-pot impregnation-thermal reduction.⁵⁴ The covalent bonds formed between the graphene and g-C₃N₄ endow the GCN sample with a slight redshift of the absorption band edge. The GCN sample exhibited a 2.3-fold enhancement of CH₄ yield over pure g-C₃N₄ due to the inhibited electron–hole pair recombination and increased charge transfer.

In another study, 3D porous g-C₃N₄/carbon nanosheets were prepared by simple pyrolysis and subsequent carbothermal activation.⁵⁵ Gas-phase CO₂ photoreduction experiments under simulated solar irradiation showed that the photocatalytic activity of the 3D g-C₃N₄ was 2 times that of the bulk g-C₃N₄ due to the introduction of carbon and the 3D porous structure. First, the hierarchical 3D porous architecture promoted CO₂ transportation to active sites and accelerated the diffusion of products. Second, the 3D porous structure could have increased the light-harvesting ability. Furthermore, the photogenerated electrons could have effectively drifted from g-C₃N₄ to the hybrid carbon under the action of the inner electric field due to the quantum confinement effect, which inhibited the electron–hole pairs from combining and prolonged the lifetime of the photogenerated charge carriers. Finally, according to their X-ray photoelectron spectroscopy (XPS) analysis, there were partial electron-withdrawing carbon–oxygen groups on the surface of this composite, e.g., C–O, C=O, and O–C=O groups, which could have facilitated the electron flow on the external surface. Thus, the recombination of electron–hole pairs was suppressed.

Carbon quantum dots (CQDs) doped onto g-C₃N₄ can also offer many advantages in CO₂ photoreduction, such as reducing the band-gap, the electron-withdrawing effect, and the up-conversion effect.⁵⁶ Recently, Tang et al. reported the hole-accepting effect of carbon dots in promoting selective CO₂ reduction to methanol.⁵⁷ The microwave-synthesized carbon dots have a pure graphite structure that facilitates efficient hole transfer to carbon dots and favours electron accumulation on the surface of g-C₃N₄. Moreover, the hole-accepting carbon dots can prevent the surface adsorption of methanol, which subsequently prevents the re-oxidation of produced methanol. As a result, the carbon dot-decorated g-C₃N₄ produced methanol from water and CO₂ with nearly 100% selectivity to methanol. This work exhibited the great potential of carbon quantum dots in efficient and highly selective CO₂ photoreduction.

3.3.3 Metal oxide/sulfide/phosphide materials. Apart from carbon materials and metals, other co-catalysts are emerging, e.g., metal oxide, metal sulfide, and metal phosphide. Ye et al. loaded chainmail Ni/NiO co-catalyst onto g-C₃N₄ to increase the photocatalytic CO₂ reduction.⁵⁸ The NiO shell over the surface of Ni particles protects Ni from being oxidized to ensure increased stability and high photoactivity of elemental Ni. Additionally, the XPS spectra of g-C₃N₄ and Ni/NiO/g-C₃N₄ indicated that g-C₃N₄ interacted with Ni/NiO via π–π interactions to facilitate charge carrier transfer at the interface, which can also be demonstrated by photocurrent responses, electrochemical impedance spectra, and PL spectra. The surface photovoltage (SPV) spectra and transient photovoltage (TPV) spectra of the samples indicated that the introduction of Ni/NiO endowed the catalyst with stronger generation and more efficient separation of the photogenerated charge carriers. With an adequate amount of Ni/NiO loading, the CO yield on Ni/NiO/g-C₃N₄ reached 13.95 μmol g⁻¹ h⁻¹, which was approximately 9-fold higher than that on pristine g-C₃N₄.

Zhou et al. loaded molybdenum phosphide (MoP) onto g-C₃N₄.⁵⁹ The introduction of MoP facilitated the separation and transfer of electron hole pairs, and increased light absorption throughout the UV and visible light spectrum without changing the absorption edge of g-C₃N₄, thus endowing the photocatalyst with a 4.5-fold higher production of CO compared with pristine g-C₃N₄. Pan et al. prepared NiS₂ quantum dot-modified g-C₃N₄ photocatalysts by a hydrothermal method.⁶⁰ The lower Fermi level and surficial metallicity resulted in the ability of NiS₂ to become an ideal electron acceptor, which accelerated the separation of electron–hole pairs and inhibited their recombination. Additionally, the introduction of NiS₂ quantum dots provided more active sites and contributed to close interface contact. Therefore, the CO yield on NiS₂/g-C₃N₄ reached 10.68 μmol g⁻¹ h⁻¹, which was 3.88 times higher than that on pristine g-C₃N₄.

Co-catalysts significantly improve the photocatalytic performance of g-C₃N₄ by facilitating the charge carrier transfer and suppressing the recombination of electron–hole pairs. It is worth noting that some metal oxides, metal phosphides, and metal sulfides can form heterojunctions together with g-C₃N₄, which is further discussed below.

3.4 g-C₃N₄-based composites

Combining g-C₃N₄ with other semiconductors to form heterojunctions can facilitate electron transfer, light utilization, and even CO₂ adsorption. Hence, it is a highly recommended strategy for CO₂ photocatalytic reduction using g-C₃N₄.

3.4.1 g-C₃N₄ absorbent composites. It is well known that the adsorption/activation of CO₂ is the first step in photocatalytic CO₂ reduction. However, pristine g-C₃N₄ has a limited surface area and a small number of active sites due to its stacked structure, which impedes the adsorption of g-C₃N₄. One strategy to change this is to combine g-C₃N₄ with CO₂ adsorbents.

Layered double hydroxides (LDHs) are a class of compounds made up of positively charged brucite-like layers with an interlayered region containing charge-compensating anions and solvation molecules (Fig. 7).⁸¹ By using a general formula [MII_1−xMIII_x(OH)₂]^x+·(Aⁿ⁻)_x/n·mH₂O (M = metal, A = interlayer anion), it was determined that LDHs exhibited some physicochemical properties that are in favour of CO₂ adsorption, e.g., large surface areas, positive surface charges, and compositional flexibilities.⁷⁶ Xu et al. used Mg–Al-LDH nanosheets to enrich CO₂ on the surface of g-C₃N₄ by the exploitation of its anion-exchange capacity.⁶¹ LDH/g-C₃N₄ was assembled via electrostatic interaction combined with Pd co-catalyst, and it showed a remarkably enhanced CH₄ yield compared to the yield without LDH. Moreover, enriched CO₂ existed in the form of interlayered CO₃²⁻ in the LDH, which could have been more efficiently reduced. Apart from LDH, metal organic frameworks (MOFs) have unique structural properties such as high surface area, high porosity, and low crystal density, as well as high thermal and chemical stability, and thus, they can also be utilized to enhance CO₂ adsorption on g-C₃N₄.⁶²

Fig. 7 Schematic illustration of layered double hydroxide structure and chemical.⁸⁰ Copyright 2017, Elsevier.

3.4.2 g-C₃N₄-based heterojunctions. From previous studies, it is known that the most frequently used g-C₃N₄-based photocatalysts are composites of g-C₃N₄ with other semiconductors, and are called g-C₃N₄-based heterojunctions.⁷⁶
(a) g-C₃N₄-based type II heterojunction. The heterojunction is a type of p–n junction. It refers to two semiconductor materials that have been combined through surface assembly or internal crystal interface crosslinks, which accelerates the separation of electron–hole pairs by the construction of a built-in electric field.

In terms of the energy band, heterojunctions fall into three categories: (i) type I heterojunction, where the forbidden gap of one component is completely covered by that of the other component; (ii) type II heterojunction, where the forbidden bands of the two components are staggered, and the semiconductor has a more negative CB and a less positive VB; (iii) type III heterojunction, where the forbidden bands of the two components are completely separated. The most investigated photocatalysis systems are usually type II heterojunctions. When a photocatalytic reaction takes place, the photogenerated electrons transfer from the more negative CB to the CB of the other semiconductor and generate a reduction reaction, while the holes transfer from the more positive VB to VB of the other semiconductor and generate an oxidation reaction.

One can construct a type II heterojunction system with g-C₃N₄ and another semiconductor with a more negative CB position or a more positive VB position (Fig. 8). Commonly, there are few semiconductors whose CB position is more negative than that of g-C₃N₄, and thus, it is easy to construct a g-C₃N₄-based heterojunction, as illustrated in Fig. 8b. Many metal oxides, sulfides, halides, and other semiconductors are used for such a system.²¹ The construction of the heterostructure plays a significant role in its photocatalytic performance and stability.^82,83

Fig. 8 Charge transfer in the conventional type II g-C₃N₄-based heterojunction systems composed of g-C₃N₄ and another semiconductor (a) with more negative CB and (b) with more positive VB.²¹ Copyright 2017, Wiley-VCH.

(b) g-C₃N₄-based Z-scheme heterojunction. The Z-scheme heterojunction, mimicking the photosynthesis of plants, was first proposed to solve the problem that a single photocatalyst could not split water to simultaneously produce both hydrogen and oxygen. A classic Z-scheme photocatalytic system consists of two photoluminescence processes and a series of redox reactions, and the electron transfer takes place in two optical systems (SI and SII).⁸⁴ Different from type II heterojunctions, the electrons excited on SI transfer to SII through an electronic mediator and combine with holes generated on SII (Fig. 9a). Consequently, the electron–hole pairs generated in each semiconductor are separated, while the high reduction and oxidation abilities of the electrons and holes, respectively, are maintained.^21,85 This mechanism involved with a mediator is known as the indirect Z-scheme. There is another type of mediator-free charge transfer mechanism, namely, the direct Z-scheme mechanism. In the direct Z-scheme mechanism, the photogenerated electrons on SI are directly transferred to SII and combined with the holes in SII (Fig. 9b).^86,87 This results in direct contact between the two semiconductors and accelerates the charge transfer between the two components, which not only increases the reaction efficiency but also reduces the cost.

Fig. 9 Z-scheme charge transfer between semiconductors (a) with or (b) without an electron mediator.²¹ Copyright 2017, Wiley-VCH.

Li et al. synthesized an indirect Z-scheme BiOI/g-C₃N₄ heterojunction using a simple deposition method.⁶³ The composite at an optimal fraction of 7.4 wt% BiOI exhibited a significantly enhanced CO yield of 17.9 μmol g⁻¹ in a CO₂ photoreduction test. The indirect Z-scheme charge transfer mechanism was confirmed through a series of characterizations including UPS, XPS, and DRS. A contrast test was executed to confirm the roles of intermediate I₃⁻/I⁻ pairs. In another study, an Al–O-bridged g-C₃N₄/α-Fe₂O₃ heterojunction was prepared using a two-step wet chemical method with AlCl₃ aqueous solution as the Al–O source.⁶⁴ The Al–O bridges significantly promoted charge transfer and separation and consequently enhanced the CO₂ photoreduction performance (Fig. 10). The composite exhibited an approximately 4-fold enhancement of CO production (24 μmol g⁻¹ h⁻¹). Due to the existence of a charge transfer mediator, this type of Z-scheme heterojunction requires well-designed structures to achieve high charge separation and transfer efficiencies. In contrast, the charge transfers in direct Z-scheme heterojunctions are accelerated by the interface contacts, and there are high reaction efficiencies and low cost with direct Z-scheme heterojunctions. Therefore, there has been increasing interest in direct Z-scheme heterojunctions in recent years.

Fig. 10 Schematic showing the transfer and separation of photo-induced charges in the fabricated Al–O-bridged g-C₃N₄/α-Fe₂O₃ nanocomposites and the induced photochemical reactions.⁶⁴ Copyright 2018, Elsevier.

Wong et al. also fabricated a direct Z-scheme heterojunction consisting of urchin-shaped α-Fe₂O₃ and a g-C₃N₄ nanosheet.⁶⁵ Without using any sacrificial agent or co-catalyst, the Z-scheme hybrid had a 2.2-fold higher CO evolution rate as compared to pristine g-C₃N₄, consisting of up to 27.2 μmol g⁻¹ h⁻¹, which is also higher than that of an Al–O-bridged g-C₃N₄/α-Fe₂O₃ heterojunction. The introduction of 3D α-Fe₂O₃ not only narrows the band gap, increases the light absorption, and results in a high usage of optical energy, but it also increases the binding energy and CO₂ adsorption so that additional CO₂ molecules can participate in the reaction over a longer time span. Additionally, the direct Z-scheme structure facilitates the electron–hole separation in both α-Fe₂O₃ and g-C₃N₄, thus improving the reduction ability of photoelectrons in the conduction band of g-C₃N₄. The improved light harvesting, CO₂ adsorption, and separation of charge carriers all contributed to the improved CO₂ photoreduction performance.

There are many other materials that can be compounded with g-C₃N₄ to fabricate direct Z-scheme heterojunctions for photocatalytic CO₂ reductions, e.g., ZnO, NiAl-LDH, and Bi₂S₃. Most of these are also well designed in terms of structure for further enhancements in CO₂ photoreduction yields.^66–68 For instance, Ogale et al. synthesized a 2D g-C₃N₄/NiAl-LDH heterojunction with strong electrostatic interactions between positively charged NiAl-LDH and negatively charged g-C₃N₄.⁶⁶ The enhancement of the photocatalytic activity was mainly ascribed to the excellent interfacial contacts at the 2D/2D interface. The large contact area between the NiAl-LDH and g-C₃N₄ subsequently suppressed charge carrier recombination and improved the transfer and separation of photogenerated charge carriers. Guo et al. synthesized Bi₂S₃ quantum dots/g-C₃N₄ composites, where the Bi₂S₃ quantum dots were uniformly dispersed on g-C₃N₄, and this increased the light absorption and the separation of electron–hole pairs. A series of characterizations demonstrated the Z-scheme mechanism. The heterojunction exhibited 4-fold higher CO yield than the pristine g-C₃N₄.⁶⁷

The photocatalytic mechanism of Z-scheme heterojunctions can be revealed by PL analysis and DFT calculations. Yu et al. prepared a g-C₃N₄/ZnO binary composite for photocatalytic CO₂ reduction to CH₃OH.⁶⁸ The photocatalytic activity exhibited 2.3-fold enhancement compared with pure g-C₃N₄. The possible photocatalytic mechanism was verified by carrying out a PL analysis of hydroxyl radicals (˙OH) produced on pure g-C₃N₄, pure ZnO, and the composite. In terms of thermodynamics, the ˙OH and ˙O₂⁻ can only be produced by photogenerated holes from the VB of the ZnO and electrons from the CB of the g-C₃N₄, respectively. According to the double-transfer mechanism, neither ˙OH nor ˙O₂⁻ could be produced in the composite system, which implied that it was likely that the charge transfer followed the Z-scheme mechanism. This observation was supported by calculating the charge carrier effective mass and separating tendency of photogenerated electron–hole pairs using a DFT calculation. It was found that ZnO has a much higher relative effective mass of electrons and holes indicating that the excited electrons of ZnO have a higher tendency to transfer, and thus, the charge transfer is more likely to originate from the ZnO to g-C₃N₄ instead of the opposite direction.⁸⁸ These findings support the rationality of the Z-scheme mechanism.

3.4.3 Photocatalyst based on g-C₃N₄ and metal complexes. As mentioned before, compared to traditional semiconductor catalysts, metal complexes possess higher efficiency and selectivity, but poor oxidation ability and recycling ability, which limit their utilization.^89,90 Therefore, g-C₃N₄ and metal complexes are combined in some studies, where the metal complex and g-C₃N₄ act as active sites and light-harvesting units, respectively, to combine their advantages.

Maeda et al. synthesized a heterogeneous photocatalyst consisting of a ruthenium (Ru) complex and g-C₃N₄. The former served as the catalytic unit and the latter as a light-harvesting unit, respectively. The electronic interactions between the two components enable the electrons to transfer from g-C₃N₄ into ruthenium.⁶⁹ With trans-(Cl)-[Ru[4,4′-(CH₂PO₃H₂)₂-2,2′-bipyridine](CO)₂Cl₂] (RuP) as catalyst and a DMA : TEOA mixture (4 : 1, v/v) as solvent, CO₂ was reduced to HCOOH and CO with a selectivity of 80% for HCOOH, and the turnover number (TON) reached 1061 after 20 h of irradiation. The apparent quantum yield was 5.7% at 400 nm, which is higher than the performance reported in earlier studies. Jain et al. combined g-C₃N₄ with cobalt(II) phthalocyanine tetracarboxylate (CoPc-COOH) in various ratios to form numerous complex-semiconductor hybrid photocatalysts.⁷⁰ The CO₂ photoreduction activities of the synthesized CoPc-COOH, g-C₃N₄, and g-C₃N₄/CoPc-COOH were tested in a water/DMF environment using triethylamine (TEA) as a sacrificial agent. After a 24 h period of visible light irradiation, the CH₄ yield of the g-C₃N₄/CoPc-COOH was 12.9 mmol g⁻¹, which is 7.3 and 10.9 times higher than that of CoPc-COOH and g-C₃N₄, respectively. The improved photocatalytic efficiency was ascribed to the higher CO₂ concentration resulting from the binding ability of CoPc-COOH to CO₂.

For the g-C₃N₄/metal complex photocatalytic system, the interaction between g-C₃N₄ and metal complex plays an important role in the enhancement of charge transfer migration and suppression of charge recombination. To maximize the interaction, either π–π interaction between the tri-s-triazine unit of g-C₃N₄ and organic ligands or hydrogen bonds/covalent bonds between the functional groups can make positive contributions.

4. Conclusions

We reviewed the recent progress on g-C₃N₄-based materials for photocatalytic reduction of CO₂. Furthermore, the basic properties of g-C₃N₄, and the principles and main steps of CO₂ photoreduction have been critically elaborated. Additionally, different types of g-C₃N₄-based composites were compared, e.g., pristine g-C₃N₄, non-metal doped g-C₃N₄, g-C₃N₄ with co-catalysts, g-C₃N₄-based semiconductors, and g-C₃N₄/metal complex. Among the g-C₃N₄-based photocatalysts for CO₂ reduction, g-C₃N₄-based semiconductors are most frequently used and have attracted the most interest. The Z-scheme structure, especially direct Z-scheme heterojunctions, realize charge transfer through interface contact, which accelerates charge transfer and leads to increased reaction efficiency and reduced cost. Hence, direct Z-scheme heterojunctions have attract more interest in recent years. Moreover, combined with other treatments such as elemental doping and structural fabrication, the photocatalytic efficiency of g-C₃N₄-based materials can be further increased. However, compared to g-C₃N₄/metal complex systems, a heterogeneous photocatalysis system based on semiconductor photocatalysts exhibits relatively lower utilization of light and depressed photocatalytic efficiency. By improving the interaction between the active sites and light-harvesting unit and optimizing the reaction conditions, g-C₃N₄/metal complex photocatalysts can increase the conversion efficiency greater than 10 fold, but this can complicate separation and recycling.

In view of the above challenges, significant efforts have been made to combine homogeneous and heterogeneous photocatalysts to obtain high efficiency and high oxidation ability. For the heterogeneous photocatalytic system, strategies were mainly focused on three approaches: (i) tuning the surface properties of g-C₃N₄ to enhance the adsorption or activation of CO₂, (ii) tuning the band structure for more optimal light utilization and photogenerated charge transfer, and (iii) reasonable design of a reaction pathway to accelerate the interface reaction. With laboratory-grade reactions, greater than 90% yield and selectivity of a product could be attained, which is much better than an industrial grade CO₂ conversion. Due to financial constraints and industrial scale, these systems are not widely applicable to industrial production. However, future recommended pathways will enable the development of more efficient and sustainable technologies for the g-C₃N₄ photoreduction of CO₂ so that a cleaner environment will exist for future generations.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by National Key R&D Program of China [2016YFA0202900, 2016YFC1402400], National Natural Science Foundation of China [51672173, U1733130], MOE Joint Foundation [6141A02022264], Shanghai Science and Technology committee [17JC1400700, 18520744700, 18JC1410500], Science and Technology Planning Project of Guangdong Province [2016A010103018]. The authors gratefully acknowledge the Shanghai Synchrotron Radiation Facility (SSRF).

References

1. M. Aresta, A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114, 1709–1742.
2. X. Liu, S. Inagaki and J. Gong, Angew. Chem., Int. Ed., 2016, 55, 14924–14950.
3. T. Inoue, A. Fujishima, S. Konishi and K. Honda, Nature, 1979, 277, 637–638.
4. J. Hawecker, J. M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., 1983, 9, 536–538.
5. I. Willner, D. Mandler and A. Riklin, J. Chem. Soc., Chem. Commun., 1986, 13, 1022–1024.
6. W. Wang, J. Soulis, Y. J. Yang and P. Biswas, Aerosol Air Qual. Res., 2014, 14, 533–549.
7. J. Qin, V. A. Larionov, K. Harms and E. Meggers, ChemSusChem, 2019, 12, 320–325.
8. Y. Guo, P. Yang, S. Zhang, B. Jiang, A. Khan, L. Zhu, Z. Fu and Z. Fan, Iran. Polym. J., 2018, 27, 153–159.
9. F. Chen, X. Jiang, L. Zhang, R. Lang and B. Qiao, Chin. J. Catal., 2018, 39, 893–898.
10. W. Zhou, W. Li, J. Wang, Y. Qu, Y. Yang, Y. Xie, K. Zhang, L. Wang, H. Fu and D. Zhao, J. Am. Chem. Soc., 2014, 136, 9280–9283.
11. N. Tian, Y. Zhang, X. Li, K. Xiao, X. Du, F. Dong, G. I. N. Waterhouse, T. Zhang and H. Huang, Nano Energy, 2017, 38, 72–81.
12. C. Li, G. Chen, J. Sun, J. Rao, Z. Han, Y. Hu, W. Xing and C. Zhang, Appl. Catal., B, 2016, 188, 39–47.
13. Q. Zhang, D. Q. Lima, I. Lee, F. Zaera, M. Chi and Y. Yin, Angew. Chem., Int. Ed., 2011, 50, 7088–7092.
14. L. Jiang, X. Yuan, G. Zeng, X. Chen, Z. Wu, J. Liang, J. Zhang, H. Wang and H. Wang, ACS Sustainable Chem. Eng., 2017, 5, 5831–5841.
15. T. Xu, L. Zhang, H. Cheng and Y. Zhu, Appl. Catal., B, 2011, 101, 382–387.
16. S. D. Perera, R. G. Mariano, K. Vu, N. Nour, O. Seitz, Y. Chabal and K. J. Balkus, ACS Catal., 2012, 2, 949–956.
17. C. Chen, W. Cai, M. Long, B. Zhou, Y. Wu, D. Wu and Y. Feng, ACS Nano, 2010, 4, 6425–6432.
18. Y. Wang, X. Wang and M. Antonietti, Angew. Chem., Int. Ed., 2012, 51, 68–89.
19. E. G. Gillan, Chem. Mater., 2000, 12, 3906–3912.
20. J. Zhang, B. Wang and X. Wang, Acta Phys.-Chim. Sin., 2013, 29, 1865–1876.
21. J. Fu, J. Yu, C. Jiang and B. Cheng, Adv. Energy Mater., 2018, 8, 1701503.
22. L. Jiang, X. Yuan, Y. Pan, J. Liang, G. Zeng, Z. Wu and H. Wang, Appl. Catal., B, 2017, 217, 388–406.
23. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
24. J. Mao, T. Peng, X. Zhang, K. Li, L. Ye and L. Zan, Catal. Sci. Technol., 2013, 3, 1253–1260.
25. Y. Li, R. Ma, L. He and Z. Diao, Catal. Sci. Technol., 2014, 4, 1498–1512.
26. F. D. Meylan, V. Moreau and S. Erkman, J. CO₂ Util., 2015, 12, 101–108.
27. J. Wu, Y. Huang, W. Ye and Y. Li, Adv. Sci., 2017, 4, 1700194.
28. J. Lee, D. C. Sorescu and X. Deng, J. Am. Chem. Soc., 2011, 133, 10066–10069.
29. Q. Lu, Y. Yu, Q. Ma, B. Chen and H. Zhang, Adv. Mater., 2016, 28, 1917–1933.
30. Y. S. Seo and S. G. Oh, Korean J. Chem. Eng., 2019, 36, 2118–2124.
31. G. Dong and L. Zhang, J. Mater. Chem., 2012, 22, 1160–1166.
32. Y. Zheng, L. Lin, X. Ye, F. Guo and X. Wang, Angew. Chem., Int. Ed., 2014, 53, 11926–11930.
33. P. Xia, B. Zhu, J. Yu, S. Cao and M. Jaroniec, J. Mater. Chem. A, 2017, 5, 3230.
34. M. Shen, L. Zhang, M. Wang, J. Tian, X. Jin, L. Guo, L. Wang and J. Shi, J. Mater. Chem. A, 2019, 7, 1556–1563.
35. J. Tang, X. Kong, B. J. Ng, Y. H. Chew, A. R. Mohamed and S. P. Chai, Catal. Sci. Technol., 2019, 9, 2335–2343.
36. X. Wu, H. Ma, W. Zhong, J. Fan and H. Yu, Appl. Catal., B, 2020, 271, 118899.
37. X. Wu, D. Gao, H. Yu and J. Yu, Nanoscale, 2019, 11, 9608.
38. J. Fu, K. Liu, K. Jiang, H. Li, P. An, W. Li, N. Zhang, H. Li, X. Xu, H. Zhou, D. Tang, X. Wang, X. Qiu and M. Liu, Adv. Sci., 2019, 6, 1900796.
39. B. Liu, L. Ye, R. Wang, J. Yang, Y. Zhang, R. Guan, L. Tian and X. Chen, ACS Appl. Mater. Interfaces, 2018, 10, 4001–4009.
40. K. Wang, Q. Li, B. Liu, B. Cheng, W. Ho and J. Yu, Appl. Catal., B, 2015, 176–177, 44–52.
41. S. Samanta, R. Yadav, A. Kumar and A. K. Sinha, Appl. Catal., B, 2019, 259, 118054.
42. K. Wang, J. Fu and Y. Zheng, Appl. Catal., B, 2019, 254, 270–282.
43. S. Wang, J. Zhan, K. Chen, A. Ali, L. Zeng, H. Zhao, W. Hu, L. Zhu and X. Xu, ACS Sustainable Chem. Eng., 2020, 8, 8214–8222.
44. H. Zhang, Y. Tang, Z. Liu, Z. Zhu, X. Tang and Y. Wang, Chem. Phys. Lett., 2020, 751, 137467.
45. Z. Zhu, X. Tang, W. Fan, Z. Liu, P. Huo, T. Wang, Y. Yan and C. Li, J. Alloys Compd., 2019, 775, 248–258.
46. J. Yu, K. Wang, W. Xiao and B. Cheng, Phys. Chem. Chem. Phys., 2014, 16, 11492–11501.
47. W. J. Ong, L. L. Tan, S. P. Chai and S. T. Yong, Dalton Trans., 2015, 44, 1249–1257.
48. G. Gao, Y. Jiao, E. R. Waclawik and A. Du, J. Am. Chem. Soc., 2016, 138, 6292–6297.
49. X. Lan, Y. Li, C. Du, T. She, Q. Li and G. Bai, Chem.–Eur. J., 2019, 25, 8560–8569.
50. H. Li, Y. Gao, Z. Xiong, C. Liao and K. Shih, Appl. Surf. Sci., 2018, 439, 552–559.
51. F. Gonell, A. V. Puge, B. Julian-Lopez, H. Garcia and A. Corma, Appl. Catal., B, 2016, 180, 263–270.
52. S. Neatu, J. A. Macia-Agullo, C. Patricia and H. Garcia, J. Am. Chem. Soc., 2014, 136, 15969–15976.
53. P. Li, F. Wang, S. Wei, X. Li and Y. Zhou, Phys. Chem. Chem. Phys., 2017, 19, 4405–4410.
54. W. J. Ong, L. L. Tan, S. P. Chai and S. Yong, Chem. Commun., 2015, 51, 858–861.
55. Y. Wang, Q. Xia, X. Bai, Z. Ge, Q. Yang, C. Yin, S. Kang, M. Dong and X. Li, Appl. Catal., B, 2018, 239, 196–203.
56. H. Feng, Q. Guo, Y. Xu, T. Chen, Y. Zhou, Y. Wang, M. Wang and D. Shen, ChemSusChem, 2018, 11, 4256–4261.
57. Y. Wang, X. Liu, X. Han, R. Godin, J. Chen, W. Zhou, C. Jiang, J. F. Thompson, K. B. Mustafa, S. A. Shevlin, J. R. Durrant, Z. Guo and J. Tang, Nat. Commun., 2020, 11, 2531.
58. C. Han, R. Zhang, Y. Ye, L. Wang, Z. Ma, F. Su, H. Xie, Y. Zhou, P. K. Wong and L. Ye, J. Mater. Chem. A, 2019, 7, 9726.
59. J. Tang, D. Yang, W. Zhou, R. Guo, W. Pan and C. Huang, J. Catal., 2019, 370, 79–87.
60. H. Qin, R. Guo, X. Liu, X. Shi, Z. Wang, J. Tang and W. Pan, Colloids Surf., A, 2020, 600, 124912.
61. J. Hong, W. Zhang, Y. Wang, T. Zhou and R. Xu, ChemCatChem, 2014, 6, 2315–2321.
62. Y. S. Bae and R. Q. Snurr, Angew. Chem., Int. Ed., 2011, 50, 11586–11596.
63. J. Wang, H. Yao, Z. Fan, L. Zhang, J. Wang, S. Zang and Z. Li, ACS Appl. Mater. Interfaces, 2016, 8, 3765–3775.
64. J. Wang, C. Qin, H. Wang, M. Chu, A. Zada, X. Zhang, J. Li, F. Raziq, Y. Qu and L. Jing, Appl. Catal., B, 2018, 221, 459–466.
65. Z. Jiang, W. Wan, H. Li, S. Yuan, H. Zhao and P. Wong, Adv. Mater., 2018, 30, 1706108.
66. S. Tonda, S. Kumar, M. Bhardwaj, P. Yadav and S. Ogale, ACS Appl. Mater. Interfaces, 2018, 10, 2667–2678.
67. R. Guo, X. Liu, H. Qin, Z. Wang, X. Shi, W. Pan, Z. Fu, J. Tang, P. Jia, Y. Miao and J. Gu, Appl. Surf. Sci., 2020, 500, 144059.
68. W. Yu, D. Xu and T. Peng, J. Mater. Chem. A, 2015, 3, 19936–19947.
69. R. Kuriki, K. Sekizawa, O. Ishitani and K. Maeda, Angew. Chem., Int. Ed., 2015, 127, 2436–2439.
70. A. Kumar, P. Prajapati, K. Pankaj, M. S. Aathira, A. Bansiwal, R. Boukherroub and S. L. Jain, J. Colloid Interface Sci., 2019, 543, 201–213.
71. G. Zhang, M. Zhang, X. Ye, X. Qiu, S. Lin and X. Wang, Adv. Mater., 2014, 26, 805–809.
72. G. Zhang, G. Li, T. Heil, S. Zafeiratos, F. Lai, A. Savateev, M. Antonietti and X. Wang, Angew. Chem., Int. Ed., 2019, 58, 3433–3437.
73. D. Masih, Y. Ma and S. Rohani, Appl. Catal., B, 2017, 206, 556–588.
74. S. Cao, H. Li, T. Tong, H. Chen, A. Yu, J. Yu and H. Chen, Adv. Funct. Mater., 2018, 28, 1802169.
75. N. Sagara, S. Kamimura, T. Tsubota and T. Ohno, Appl. Catal., B, 2016, 192, 193–198.
76. Z. Sun, H. Wang, Z. Wu and L. Wang, Catal. Today, 2018, 300, 160–172.
77. J. Hong, W. Zhang, Y. Wang, T. Zhou and R. Xu, ChemCatChem, 2014, 6, 2315–2321.
78. Q. Zhai, S. Xie, W. Fan, Q. Zhang, Y. Wang, W. Deng and Y. Wang, Angew. Chem., Int. Ed., 2013, 52, 5776–5779.
79. Q. Gu, J. Long, H. Zhuang, C. Zhang, Y. Zhou and X. Wang, Phys. Chem. Chem. Phys., 2014, 16, 12521–12534.
80. M. Zubair, M. Daud, G. McKay, F. Shehzad and M. A. Al-Harthi, Appl. Clay Sci., 2017, 143, 279–292.
81. Q. Wang and D. O'Hare, Chem. Rev., 2012, 112, 4124–4155.
82. H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, Z. Wang, J. Liu and X. Wang, Chem. Soc. Rev., 2014, 43, 5234–5244.
83. H. Li, Y. Zhou, W. Tu, J. Ye and Z. Zou, Adv. Funct. Mater., 2015, 25, 998–1013.
84. K. Maeda, ACS Catal., 2013, 3, 1486–1503.
85. H. Tada, T. Mitsui, T. Kiyonaga, T. Akita and K. Tanaka, Nat. Mater., 2006, 5, 782–786.
86. J. Low, C. Jiang, B. Cheng, S. Wageh, A. A. Al-Ghamdi and J. Yu, Small Methods, 2017, 1, 1700080.
87. P. Zhou, J. Yu and M. Jaroniec, Adv. Mater., 2014, 26, 4920–4935.
88. H. Zhang, L. Liu and Z. Zhou, Phys. Chem. Chem. Phys., 2011, 14, 1286–1292.
89. B. Kumar, M. Llorente, J. Froehlich, T. Dang, A. Sathrum and C. P. Kubiak, Annu. Rev. Phys. Chem., 2012, 63, 541–569.
90. T. Yui, Y. Tamaki, K. Sekizawa and O. Ishitani, Top. Curr. Chem., 2011, 303, 151–184.

---