g-C₃N₄-Based Nanomaterials for Visible Light-Driven Photocatalysis

Abstract

Graphitic carbon nitride (g-C₃N₄) is a promising material for photocatalytic applications such as solar fuels production through CO₂ reduction and water splitting, and environmental remediation through the degradation of organic pollutants. This promise reflects the advantageous photophysical properties of g-C₃N₄ nanostructures, notably high surface area, quantum efficiency, interfacial charge separation and transport, and ease of modification through either composite formation or the incorporation of desirable surface functionalities. Here, we review recent progress in the synthesis and photocatalytic applications of diverse g-C₃N₄ nanostructured materials, and highlight the physical basis underpinning their performance for each application. Potential new architectures, such as hierarchical or composite g-C₃N₄ nanostructures, that may offer further performance enhancements in solar energy harvesting and conversion are also outlined.

1. Introduction

1.1. Background

Future energy production, storage and security, and combating anthropogenic environmental pollution, represent key global challenges for both developed and emerging nations [1,2]. Sunlight, an essentially limitless source of clean energy, has the potential to address both these challenges [3,4], and its utilization to this end entered mainstream science following breakthroughs in semiconductor light harvesting for photocatalysis by Honda and Fujishima in the 1970s [5,6,7]. This discovery led to extensive research into titania semiconductor photocatalysts, principally for water splitting and the degradation of aqueous or airborne organic pollutants under UV light irradiation [8,9,10,11,12,13]. However, efficient harnessing of visible light (the major component of solar radiation that reaches the Earth’s surface) by photocatalysts to drive chemical transformations remains problematic [14,15,16] due to identifying suitable materials that possess narrow band gaps, high quantum yields, efficient charge carrier transport, and low rates of charge carrier recombination, and good thermo-, photo-, and chemical stability. The development of such low cost photocatalysts from earth abundant, and ideally non-toxic elements for visible light harvesting would unlock opportunities for their large-scale application to supplement existing renewable energy networks and pollution control systems.

1.2. Semiconductor Photocatalysis

Semiconductor photocatalysis refers to the acceleration of chemical transformations (most commonly oxidations and reductions) brought about through the activation of a catalyst, comprising a semiconductor either alone or in combination with metal/organic/organometallic promoters, through light absorption, with subsequent charge and/or energy transfer to adsorbed species. Note that the direct activation of reactants and intermediates through light absorption is the realm of photochemistry; in establishing whether a transformation is truly photocatalytic it is therefore crucial to establish that photons are absorbed by the catalyst rather than adsorbates [17,18]. In the photocatalytic production of so-called ‘solar fuels’, photoexcited charge carriers drive the conversion of water and CO₂ into H₂, CO, CH₄, CH₃OH and related oxygenates and hydrocarbons [19,20,21]. Such processes parallel those in nature wherein sunlight absorbed by chlorophyll in plants promotes starch and oxygen production from carbon dioxide and water), and are hence termed artificial photosynthesis (Figure 1). Photoexcited charge carriers can also either induce the total oxidation (mineralization) of organic pollutants such as those encountered in aquatic environments, either directly, or through the creation of potent oxidants such as hydroxyl radicals [22].

Figure 1. (a) Natural, and (b) artificial photosynthesis through water splitting and CO₂ reduction, and (c) photodegradation of aqueous organic pollutants.

1.3. Photocatalytic Mechanisms

Semiconductor photocatalysis is initiated by exciton formation following photon absorption and the excitation of electrons from the valence band into the conduction band (Step I). The resulting electron–hole pairs may recombine in either the bulk of the semiconductor, or at the surface, with the associated energy released through either fluorescence or thermal excitation of the lattice (Step II); recombination is the primary process that limits photocatalyst efficiency after photon capture. Electrons (and holes) that migrate to the surface of the semiconductor and do not undergo rapid recombination may participate in various oxidation and reduction reactions with adsorbates such as water, oxygen, and other organic or inorganic species (Steps III and IV) [9,10,23,24]. These steps are summarized below and illustrated in Figure 2:

Figure 2. Principal photophysical processes for a semiconductor (SC) under light irradiation.

Step I

Light absorption $\mathrm{SC}+\mathrm{h}v\to \mathrm{SC}\times \left({\mathrm{e}}_{\mathrm{CB}}^{-}+{\mathrm{h}}_{\mathrm{VB}}^{+}\right)$.

Step II

Recombination ${\mathrm{e}}_{\mathrm{CB}}^{-}+{\mathrm{h}}_{\mathrm{VB}}^{+}\to \mathrm{h}v+\mathrm{heat}$.

Step III

Reduction $\mathrm{Adsorbate}+{\mathrm{e}}_{\mathrm{CB}}^{-}\to {\mathrm{Adsorbate}}^{-}$.

Step IV

Oxidation $\mathrm{Adsorbate}+{\mathrm{h}}_{\mathrm{VB}}^{+}\to {\mathrm{Adsorbate}}^{+}$.

Oxidation and reduction reactions are fundamental to photocatalytic environmental remediation and solar fuel production, and are ultimately limited by the reduction potential of photoexcited electrons in the conduction band and oxidation potential of photogenerated holes in the valence band. The redox potential, band energies and gap of a semiconductor therefore largely determine the likelihood and rate of charge transfer, and hence are key design parameters for photocatalysts [12,25]. Although the underlying physics of space charge carriers and surface-electronic structure of photocatalysts varies between materials and applications, in essence, semiconductor photocatalysis represents interfacial reactions between electrons and holes generated through band gap excitation.

2. Photocatalytic Materials

The discovery of photocatalytic water splitting over titania electrodes under UV irradiation [5] has led to intensive research into explored H₂ production through this approach. Similarly, the first report on the photocatalytic oxidation of cyanide ions over TiO₂ powder [26] prompted a rapid expansion in environmental purification research and technologies, particularly for aqueous environments. In both cases, recent research has focused on identifying and developing alternative semiconductors to titania, offering superior performance under solar (rather than UV irradiation) [25]. Numerous semiconductors, including ZnO [27], Fe₂O₃ [28], WO₃ [29], SrTiO₃ [30], NaTaO₃ [31], CdS [32], Ag₃PO₄ [29], BiPO₄ [33], and g-C₃N₄ [34] are known photocatalysts, with their application dependent on their band gap (Figure 3). Despite a large body of literature, the practical utilization of such photocatalysts for solar fuels production or the degradation of organic pollutants remains a huge challenge due to poor visible light harvesting or efficient conversion of light energy to achieve chemical transformations [13,16,35].

Figure 3. Band gap energy and band edge energies of different semiconductors.

3. Graphitic Carbon Nitride (g-C₃N₄)

Solar energy output reaching the Earth’s surface is dominated by three regions (Figure 4) of the electromagnetic spectrum, UV (~5%), visible (~45%), and IR (~50%) [36]; visible light photocatalysis therefore offers the best opportunity to obtain maximum solar energy. However, most photocatalysts possess relatively wide band gaps, such as TiO₂ (3.0–3.2 eV) and are hence primarily active under UV irradiation (<385 nm) [8]. The quest for high performance visible light counterparts is reflected in the rapid growth of associated scientific papers and patents [8,10,15] for water splitting, CO₂ reduction and pollutants degradation [35]. Graphitic carbon nitride (g-C₃N₄) is a promising metal-free, polymeric semiconductor (Figure 5a) with a narrow band gap suited to visible light absorption (Figure 5b) [34], and amenable to large-scale synthesis. g-C₃N₄ may also be readily doped or chemically functionalized, permitting tuning of its photophysical properties, and in contrast to many other organic semiconductors, graphitic carbon nitride also exhibits high thermal and chemical stability to oxidation, even at temperatures of 500 °C. There is an extensive literature describing the synthesis of g-C₃N₄ and its derivatives for various applications [37,38,39,40,41,42,43]. This Review focuses on applications in photocatalytic environmental remediation and solar fuel generation, with an emphasis on emerging synthetic strategies to improve the photoactivity of g-C₃N₄-based nanostructures through controlling size, morphology, light absorption, charge separation, and ultimately surface reactions. Future research directions are also highlighted.

Figure 4. Spectral irradiance of solar light. Reproduced with permission from [36]. Copyright Royal Society of Chemistry, 2015.

Figure 5. (a) Graphitic carbon nitride structure comprising melem units, and (b) UV-vis diffuse reflectance spectrum and image (inset) of g-C₃N₄. Reprinted with permission from [34]. Copyright Springer Nature, 2009.

4. g-C₃N₄ Nanostructures: Size and Shape

Engineering materials at the nanoscale is critical to the development of devices for the electronics [44], catalysis [45], biomedical [46], sensing [47], and smart materials [48] sectors, with nanoparticles now in widespread use across science and engineering [48,49,50,51]. A number of key aspects differentiate nanomaterials from their bulk analogues. Nanomaterials possess a high surface: bulk atom ratio, which heavily influences their thermodynamic properties resulting in, e.g., melting temperature depression, and elevated solid–solid phase transition temperature. Quantum confinement effects, which influence the electrical and optical properties of nanomaterials, arise from their evolving band structure and the emergence of atomistic like behaviour. Many heterogeneous catalysts exhibit strong size-dependencies due to quantum confinement [52], notably gold [53,54], high surface areas, and the exposure of low-coordination, high energy sites [45,55,56,57]. In concert, these aspects may enhance the rate of interfacial charge transfer from a photocatalyst surface to an adsorbate [58,59]. The use of nanostructured g-C₃N₄ is a fast growing area of photocatalysis research, with nanoparticles, nanorods, nanowires, nanotubes, nanospheres, and particularly nanosheets, demonstrating unique features as components of photocatalyst systems [39].

2-dimensional g-C₃N₄: 2D-based materials offer an exceptionally high specific surface area, good crystallinity, rich options for host-guest interactions, maximal light absorption, and improved charge-carrier separation over their 3D analogues [60]. Numerous 2D nanomaterials have been reported as heterogeneous catalysts in recent years, with g-C₃N₄ emerging as one of the most promising photocatalysts. Ping and co-workers developed a facile method to prepare g-C₃N₄ nanosheets by direct thermal oxidative ‘etching’ of bulk g-C₃N₄ under air (Figure 6) [61]. In this method, the hydrogen-bond strands of polymeric melon units which form the interlayers, are gradually removed by oxidation such that the thickness of bulk g-C₃N₄ can be reduced to the desired nanoscale by controlling the etch time, and hence represents a simple, low-cost, and scalable synthesis. The resultant nanosheets exhibit a blue shift of the intrinsic absorption edge in their UV-vis spectra relative to the bulk. The increase in band gap of nanosheets (2.97 eV; Figure 7A) relative to their bulk counterpart (2.77 eV) is further confirmed by a blue shift in the fluorescence emission spectrum of 20 nm (Figure 7B). This widening of the band gap reflects quantum confinement which raises and lowers the conduction and valence band edges respectively [62]. Electronic properties of the nanosheets were determined from the corresponding I−V curve, semiconducting characteristics observed for single g-C₃N₄ nanosheets, suggesting electron transport within the nanosheet plane. In contrast, no current was detected for bulk particle under an applied bias spanning −10–+10 V, evidencing extremely poor electronic conductivity for bulk g-C₃N₄. The lifetime of charge carriers in the nanosheets from time-resolved fluorescence decay spectra also exceeded that of bulk g-C₃N₄.

Figure 6. Thermal exfoliation as a low-cost and green method to prepare ultrathin g-C₃N₄ nanosheets from bulk g-C₃N₄ in water. Reproduced with permission from [61]. Copyright John Wiley & Sons Inc., 2012.

Figure 7. (A) Diffuse reflectance UV-Vis spectra (DRUVS), and (B) photoluminescence (PL) spectra of bulk (a) and exfoliated nanosheet (b) g-C₃N₄. Reproduced with permission from [61]. Copyright John Wiley & Sons Inc., 2012.

Xiaodong and co-workers developed a different liquid exfoliation strategy as a low-cost and green route to ultrathin g-C₃N₄ nanosheets from bulk g-C₃N₄ in water, illustrated in Figure 8 [63]. From a range of solvents, water effectively exfoliated the g-C₃N₄ into ultrathin nanosheets, possibly reflecting its high polarity. The morphology of the exfoliated g-C₃N₄ showed free-standing nanosheets 120 nm across that were almost transparent, and displayed a well-defined Tyndall effect in solution (Figure 8 inset) indicating the presence of monodisperses ultrathin nanosheets. These g-C₃N₄ nanosheets were very stable in acidic and alkaline environments, but exhibited pH-dependent photoluminescence. The g-C₃N₄ nanosheets show superior photoabsorption to the bulk counterpart, resulting in an extremely high PL quantum yield of up to 19.6%. Liquid exfoliation of g-C₃N₄ in isopropanol [64] and methanol [65] resulted in nanosheets with improved photocatalytic performance for the degradation of organic pollutants relative to bulk g-C₃N₄.

Figure 8. Liquid exfoliation route as a low-cost and green method to prepare the ultrathin g-C₃N₄ nanosheets from bulk g-C₃N₄ in water. Reprinted with permission from [63]. Copyright American Chemical Society, 2013.

1-dimensional g-C₃N₄: In recent years, 1D nanostructures have attracted interest as photocatalysts due to their unique morphology and photophysical properties [66,67], and hence there is interest in preparing 1D g-C₃N₄. 1D g-C₃N₄ nanorods with different aspect ratios were prepared by the reflux of g-C₃N₄ nanoplates as a function of solvent and reflux time [68]. The transformation from g-C₃N₄ nanoplates to nanorods reflects an exfoliation and subsequent re-growth process, which results in ‘rolling-up’ of individual nanosheets into rods (Figure 9a). The photocatalytic activity of the as-prepared nanorods for methylene blue (MB) degradation in water was explored under visible light (λ > 420 nm) and simulated solar irradiation (λ > 290 nm). The resulting photocatalytic activity and photocurrent response of g-C₃N₄ nanorods under visible and solar light were about 50–100% greater than the g-C₃N₄ nanoplates.

Figure 9. (a) Synthesis, and (b) Transmission electron microscopy (TEM) images of g-C₃N₄ nanorods. Reprinted with permission from [68]. Copyright American Chemical Society, 2013.

Zhihong and co-workers demonstrated a large-scale synthesis of well-aligned g-C₃N₄ nanorods via the reactive thermolysis of mechanically activated molecular precursors, C₃N₆H₆ and C₃N₃C_l3, under heat treatment [69]. These nanorods exhibit peculiar optical properties, evidenced by PL emission and UV-vis absorption. Uniform g-C₃N₄ nanorods were also synthesized via a template of monodispersed, chiral, mesostructured silica nanorods, which were easily prepared via ammonia-catalyzed hydrolysis of tetraethyl orthosilicate with F127 and cetyltrimethylammonium bromide (CTAB) surfactants [70]. The one-dimensional, hexagonal mesostructure of the porous silica nanorods enabled carbon nitride condensation within the pores. The resulting g-C₃N₄ nanorods demonstrated a high photocatalytic activity in hydrogen evolution from water in the presence of triethanolamine and 1 wt % Pt as a co-catalyst compared to that obtained with a conventional g-C₃N₄ [71]. Porous g-C₃N₄ nanorods were also prepared by direct calcination of hydrous melamine nanofibers, precipitated from an aqueous solution of melamine [72]. Porosity provided an enhanced interfacial area for catalysis. Oxygen atoms doped into the g-C₃N₄ matrix altered the band structure, resulting in more effective separation of electron/hole pairs and a corresponding excellent visible light photocatalytic activity for hydrogen evolution in the presence of triethanolamine as a hole quencher. A simple wet-chemical route was also reported for the preparation of nanofiber-like g-C₃N₄ structures with an average diameter of several nm and 100 nm in length [73]. The g-C₃N₄ nanofibers exhibited a high surface area, and low density of crystalline defects, with a slight blue shift of 0.13 eV compared to bulk g-C₃N₄, possibly due to more perfect packing, electronic coupling, and quantum confinement effects. The catalytic activity of g-C₃N₄ nanofibers for Rhodamine B photodegradation was much higher than that of bulk g-C₃N₄, with the nanofibers also exhibiting superior stability. An alternative approach to the synthesis of g-C₃N₄ nanotubes adopted the direct heating of melamine, packed into a compact configuration to favour tubular structures (Figure 10a–d) [74]. This route was advantageous since it required no additional organic templates, facilitating commercial, low-cost and large-scale application. The resulting g-C₃N₄ showed intense fluorescence around 460 nm, and hence has potential application as a blue light fluorescence material. These g-C₃N₄ nanotubes exhibited better visible light photocatalytic activity for MB degradation than either bulk g-C₃N₄ or a p25 TiO₂ reference (the latter is unsurprising since pure titania is a UV band gap material). Muhammad and co-workers also prepared tubular g-C₃N₄ by pre-treating melamine with HNO₃ before thermal processing [75]. The g-C₃N₄ nanotubes were again active for MB and methylene orange (MO) degradation under visible light, and were more stable than bulk g-C₃N₄; the superior activity attributed to the higher surface area (182 m²·g⁻¹) of the tubes and improved light absorption and charge separation/transfer of electron–hole pairs. g-C₃N₄ nanotubes can also be obtained through rolling-up nanosheets via a simple water-induced morphological transformation [76], avoiding the use of organic solvents and hence promoting green chemical principles.

Figure 10. (a–c) Synthetic strategy, and corresponding (d) TEM image of g-C₃N₄ nanotubes. Reproduced with permission from [74]. Copyright Royal Society of Chemistry, 2014.

Ribbon-like g-C₃N₄ nanostructures have been prepared employing dicyandiamide (DCDA) and NaCl crystals as structure-directing agents [77], with a possible mechanism shown in Figure 11. These ribbon-like g-C₃N₄ nanostructures exhibit interesting optical and electronic properties, including a large blue shift in their absorption spectrum corresponding to an increased band gap from 2.7 eV to 3.0 eV. The latter may reflect the incorporation of some Na⁺ ions within the nitride pores, and functionalization by cyano groups. The ribbon-like g-C₃N₄ emits blue light at around 440 nm under 365 nm excitation, whereas bulk g-C₃N₄ exhibited a broad emission spanning 460–520 nm, i.e., yellow-green light. Unfortunately, these ribbon-like g-C₃N₄ nanostructures have not yet been tested for as photocatalysts.

Figure 11. Synthesis strategy (A) Ribbon-like g-C₃N₄ nanostructures (B) TEM image. Reproduced from with permission from [77]. Copyright Royal Society of Chemistry, 2014.

0-dimensional g-C₃N₄: 0D materials such as quantum dots are of great interest in photocatalysis [78]. g-C₃N₄ quantum dots have been prepared from bulk g-C₃N₄ by thermochemical etching [74]. This tunable multi-step preparation involves thermal exfoliation of 3D bulk g-C₃N₄ into 2D nanosheets, followed by acid etching with concentrated H₂SO₄ and HNO₃ to produce 1D nanoribbons. In this second step, some C–N bonds which connect the tri-s-triazine units are oxidized, resulting in the introduction of oxygenate functional groups, such as carboxylates, at edges and on the basal plane. Cleavage of the nanosheets along preferential orientations yields nanoribbons with diameters <10 nm and several tens of nm in length. In a final step, nanoribbons are converted to 0D quantum dots of 5–9 nm across by hydrothermal treatment (Figure 12) that are highly soluble in water, and stable in solution under ambient conditions for almost eight months. These quantum dots exhibited light ‘up-conversion’ when excited by long wavelength light, for example, irradiation with 705–862 nm light resulted in 350–600 nm emission, encompassing a large portion of the visible-light spectrum. This up-conversion was proposed to occur via a multiphoton process involving anti-Stokes photoluminescence. The ability of g-C₃N₄ quantum dots to convert NIR to visible light renders them a promising universal energy transfer component in a photocatalytic system, able to harness long wavelength solar energy. This was demonstrated for water splitting, in which quantum dots were added to promote photocatalytic H₂ production by platinized bulk g-C₃N₄ and P25, with dramatic rate-enhancements (up to 52-fold) observed for the latter under visible light irradiation in the presence of a methanol sacrificial hole scavenger. Single layered g-C₃N₄ quantum dots were also prepared by Guoping and co-workers, although in this instance for two-photon fluorescence imaging of cellular nucleus [42]. They again adopted a multi-step synthesis involving acid treatment of bulk g-C₃N₄ to form a porous material and subsequently ultrathin nanosheets, with subsequent ammonia addition, hydrothermal treatment, and ultrasonication of the porous g-C₃N₄ nanosheets liberating aqueous suspensions of g-C₃N₄ quantum dots.

Figure 12. (A) Synthesis, and (B) TEM images of g-C₃N₄ quantum dots. Reproduced from with permission from [74]. Copyright Royal Society of Chemistry, 2014.

3D-dimensional g-C₃N₄: 3D nanomaterials unlock a vast and complex design space for constructing novel and efficient photocatalytic systems [79], such as hierarchical 3D nanoporous g-C₃N₄ microspheres using a template-free solvothermal synthesis [80]. In this example, a two-step synthesis was adopted: (i) amorphous and nanoporous g-C₃N₄ microspheres were prepared from melamine and cyanuric chloride in acetonitrile; and (ii) subsequently subjected to thermal processing at 550 °C under argon to transform the amorphous microspheres into hierarchical g-C₃N₄ microspheres (Figure 13). Surprisingly, the hierarchical g-C₃N₄ microspheres exhibited a red-shift relative to the bulk counterpart, and uncalcined microspheres, attributed to the high degree of condensation and packing between the layers within the microspheres. The photoluminescence emission intensity of hierarchical g-C₃N₄ microspheres was low compared to bulk and uncalcined g-C₃N₄ microspheres indicating that calcination suppresses radiative charge recombination in the hierarchical structure. These porous g-C₃N₄ microspheres also exhibit a narrowed band gap (2.42 eV), lower electrical resistance and a higher photoresponse than the bulk material, facilitating visible-light harvesting and more efficient transport and separation of photo-induced charge carriers. Hierarchical g-C₃N₄ nanospheres, comprised of nanosheet assemblies, were also prepared by Jinshui and co-workers, but employing high area silica nanospheres as sacrificial templates [81]. The silica template offered efficient cyanamide adsorption, and a framework for the formation of interconnected 2D g-C₃N₄ nanosheets during self-polymerization on heating. The excellent thermal and mechanical stability of silica spheres enabled high temperature construction of the hierarchical g-C₃N₄ nanospheres, and could subsequently be removed through etching by NH₄HF₂ solution, with the hierarchical g-C₃N₄ retaining a spherical morphology. These hierarchal nanospheres are constructed of flat nanosheets emanating from the center (sphere surface) and then interconnecting to form a mesoporous shell (Figure 14), this structure may favour both charge separation and mass transport in photocatalysis. The nanospheres had a wider band gap than bulk g-C₃N₄, possibly due to quantum size effects, but superior light harvesting across the optical spectrum, especially between 430–590 nm. This may arise from multiple reflections (and hence opportunities for absorption) of incident light within the hierarchical architectures, and/or presence of a high density of defective sites associated with exposure of low-coordination sites at the ‘sharp’ edges of the constituent nanosheets.

Figure 13. (a) Synthetic strategy, (b) TEM image, and (c) room-temperature photoluminescence spectra of porous g-C₃N₄ microspheres. Reprinted with permission from [80]. Copyright Elsevier, 2015.

Figure 14. (a) Synthetic strategy, and (b,c) Scanning electron microscopy and TEM images of hierarchical g-C₃N₄ microspheres. Reproduced with permission from [81]. Copyright John Wiley & Sons Inc., 2014.

Hollow g-C₃N₄: Hollow nanostructures are another promising morphology for energy storage and conversion applications, with significant research efforts devoted to the design and synthesis of hollow nanostructures with high complexity through manipulating their geometry, chemical composition, building blocks, and interior architecture to, e.g., enhance their electrochemical performance [82,83]. Hollow g-C₃N₄ nanospheres have been synthesized using silica nanoparticles as templates [83,84]. Careful control over the shell thickness of such polymeric g-C₃N₄ hollow nanospheres prevents deformation of the core–shell arrangement (Figure 15), even after 400 °C processing. Although a blue shift in the band gap accompanying their synthesis is undesirable, and attributed to either quantum effects or enhanced H-type interlayer packing, further chemical methods, such as extending the pi system by anchoring aromatic motifs, exist to improve visible light absorption, for example, through co-polymerization; such chemical modification and extended p-conjugation can red-shift optical absorption, and improve charge separation in the shell, without damaging the hollow polymeric architectures [85]. This strategy has been adopted to tune the semiconductor properties of the shell in the hollow g-C₃N₄ nanospheres to enhance photocatalytic activity for hydrogen evolution under visible light. A simple, molecular cooperative assembly of low cost triazine molecules into hollow g-C₃N₄ is also reported by Young-Si et al. [86], with this precursor enabling simultaneous optimization of the textural and photophysical properties of g-C₃N₄.

Figure 15. Synthetic strategy for fabricating hollow g-C₃N₄ nanospheres. Reprinted with permission from [84]. Copyright Springer Nature, 2012.

Mesoporous g-C₃N₄: Mesoporous photocatalysts have attracted attention for their (comparatively) high quantum efficiency associated with high surface areas, superior molecular mass transport in-pore [87], and opportunities for enhanced light harvesting through the internal scattering of incident light. An atomically thin mesoporous mesh of g-C₃N₄ nanosheets was recently prepared by solvothermal synthesis (Figure 16) which exhibits outstanding photocatalytic activity for H₂ production [88].

Figure 16. (a) SEM images, (b) cartoon of photocatalytic H₂ from water splitting, and (c) TEM image of atomically thin, mesoporous g-C₃N₄ nanosheets. Reprinted with permission from [88]. Copyright American Chemical Society, 2016.

5. Photocatalytic Applications of g-C₃N₄ Nanostructures

g-C₃N₄ nanostructures have proven excellent catalysts in diverse applications [39,85,89,90] including hydrogen production from water splitting [88,91], CO₂ reduction to fuels and chemicals [92], environmental remediation [39], fuel cells [93], and organic synthesis [89]. Here we focus on photocatalytic applications.

5.1. Solar Fuel Generation

Solar fuels production from CO₂ and water via artificial photosynthesis is one of the promising strategies to deliver H₂, syngas and hydrocarbons as sustainable energy and chemical feedstocks [19]. g-C₃N₄ offers the promise of metal-free and scalable photocatalysts for visible light use.

5.1.1. H₂ Evolution

Hydrogen is one of the most promising alternative energy sources to fossil fuels; however, the large energy barrier to water splitting still presents a challenge to practical photocatalytic systems [35]:

2H₂O (l) → 2H₂ (g) + O₂ (g), ΔG = +474 kJ/mol

Advanced materials are hence sought that are amenable to harnessing sunlight for either direct photochemical, or photoelectrochemical water splitting. For photocatalytic water splitting, the conduction band (CB) energy must be sufficiently negative (relative to normal hydrogen electrode (NHE)) such that photoexcited electrons are sufficiently energetic to reduce water [94,95]:

$$2{\mathrm{H}}_2\mathrm{O}\to 2{\mathrm{H}}_2+{\mathrm{O}}_2$$

(1)

$${\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{OH}}^{-}$$

(2)

$$2{\mathrm{H}}_2\mathrm{O}+2{\mathrm{e}}^{-}\to {\mathrm{H}}_2+2{\mathrm{OH}}^{-}$$

(3)

$$2{\mathrm{H}}_2\mathrm{O}\to {\mathrm{O}}_2+4{\mathrm{H}}^{+}+4{\mathrm{e}}^{-}$$

(4)

The redox potential for the overall reaction at pH = 7, E_H = −1.23 V (NHE), with the corresponding half-reactions of −0.41 V (Equation (4)) and 0.82 V, giving an overall ΔG⁰ = +237 kJ·mol⁻¹.

Most single component photocatalysts exhibit poor activity for visible light-driven H₂ production. However, the combination of g-C₃N₄ with a metal co-catalyst and hole scavenger can afford high visible-light photoactivity. Shubin and co-workers [64] prepared g-C₃N₄ nanosheets by thermal exfoliation which demonstrated a superior hydrogen production from a water/triethanolamine solution relative to the bulk nitride; nanosheets with a thickness as low as 2 nm were optimal, achieving rate-enhancements of 5.5- and 3-fold under UV-vis and visible light irradiation respectively. Single atomic layer g-C₃N₄ nanosheets prepared by a chemical exfoliation [96] also display better photogenerated charge transport and separation than bulk g-C₃N₄, presumably due to the improved H₂ evolution. Atomically thin, mesoporous g-C₃N₄ nanomesh prepared by solvothermal routes exhibits an exceptional photocatalytic activity for H₂ evolution [88] of 8510 μmol·h⁻¹·g⁻¹ (with an apparent quantum efficiency of 5.1% at 420 nm), far higher than the 1560 μmol·h⁻¹·g⁻¹ achieved over non-porous 2D g-C₃N₄ nanostructures or 350 μmol·h⁻¹·g⁻¹ observed for bulk g-C₃N₄ (apparent quantum efficiency 3.75% at 420 nm); the porous g-C₃N₄ nanomesh possessed a high surface area and better alignment of conduction and valence band edges. g-C₃N₄ nanorods also show high photocatalytic activity for hydrogen production from water in the presence of triethanolamine (TEOA) and a 1 wt % Pt co-catalyst [68] wherein the platinum nanoparticles uniformly decorate the g-C₃N₄ nanorods. Such materials are also superior to mesoporous analogues [70]. TEOA is the most common hole scavenger for g-C₃N₄ photocatalysts wherein it confers superior activity to methanol (a 14-fold rate enhancement); although the origin of this difference remains poorly understood, Jones and co-workers speculated that the nitrogen lone pair is responsible for the enhanced activity [62]. P25 also exhibited superior activity for photocatalytic hydrogen production when TEOA was employed as a hole scavenger (versus methanol), albeit to a lesser extent than for carbon nitride. g-C₃N₄ nanotubes synthesized through a rolling-up mechanism by water-induced morphological transformation also display superior visible-light H₂ production bulk g-C₃N₄ or g-C₃N₄ nanosheets [76]. g-C₃N₄ quantum dots [77] prepared from bulk g-C₃N₄ by thermochemical etching were three times more active than bulk g-C₃N₄ under visible irradiation when promoted by 1 wt % Pt and using 10% triethanolamine as a sacrificial agent, possibly due to up-conversion of NIR to visible light and concomitant increased light harvesting. Tuning of the electronic band structure of g-C₃N₄ quantum dots [97] to optimize their visible or NIR light response, further enhances photocatalytic H₂ evolution. P-doped g-C₃N₄ nanosheets also exhibit promising visible-light photocatalytic H₂ productivity of 1596 mmol·h⁻¹·g⁻¹ (apparent quantum efficiency of 3.56% at 420 nm) superior to other metal-free g-C₃N₄ nanosheet photocatalysts [98]. The excellent photocatalytic activity originates from P-doped macroporous analogues arises from empty mid-gap states (−0.16 V vs. NHE) which extend light harvesting up to 557 nm. Macropores also increased the surface area to 123 m²·g⁻¹, and shortened the charge-to-surface migration length to only 5–8 nm.

Hierarchically 3D nanoporous g-C₃N₄ microspheres [80] have also been exploited for water splitting in aqueous solution with 15 triethanolamine and 3 wt % Pt as a co-catalyst under visible light. These g-C₃N₄ microspheres showed H₂ productivity 2.5 times higher than that of bulk g-C₃N₄, and good stability over five consecutive recycles. Hierarchical g-C₃N₄ nanospheres [81] comprised of nanosheets with 3 wt % Pt co-catalyst showed significant improvements in H₂ production, with an apparent quantum yield of 9.6% at 420 nm, far superior to that for individual g-C₃N₄ nanosheets of 3.75%. Monodispersed, hollow g-C₃N₄ nanospheres are also reported to exhibit high photoactivity for water splitting, and a high apparent quantum yield of 7.5% [84]. H₂ evolution over these hollow g-C₃N₄ spheres was significantly enhanced by addition of a MoS₂ co-catalyst, with the formation of the MoS₂/g-C₃N₄ heterojunctions (Figure 17) improving light-harvesting, and fast charge separation [99].

Figure 17. (a) Synthetic strategy and (b,c) TEM images and (d) Energy dispersive X-ray spectroscopy (EDX) elemental maps of MoS₂@hollow g-C₃N₄. Reprinted with permission from [99], copyright Elsevier, 2016.

g-C₃N₄ has also been coupled with semiconductors and metal nanoparticles that exhibit visible light surface plasmon resonances to extend their spectral range. Such heterojunction materials offer enhanced separation of photoexcited charge carriers, and hence suppressed recombination and energy loss through fluorescence [39,91]. Noble metal-promoted g-C₃N₄ offers improved UV and visible light harvesting, fast molecular diffusion, and a high density of photoactive sites [100,101,102]. TiO₂/g-C₃N₄ heterojunctions have been fabricated by a two-step hydrothermal-calcination route from melamine, followed by an in-situ solid-state reaction [103]. The resulting TiO₂/g-C₃N₄ heterostructures possess a narrow band gap and good photoactivity (556 μmol⁻¹·g⁻¹) for H₂ evolution under visible-light irradiation compared to pure g-C₃N₄ (108 μmol·h⁻¹·g⁻¹) or TiO₂ (130 μmol·h⁻¹·g⁻¹. Core@shell heterojunction nanocomposites have additional advantages due to a high interfacial contact area between the shell and core components [104]. For example, CdS@g-C₃N₄ core/shell nanowires [104] with different g-C₃N₄ contents were prepared by a combined solvothermal and chemisorption method (Figure 18) in which g-C₃N₄ uniformly adsorbs over CdS nanowires resulting in enhanced improved photocatalytic H₂ production of 4152 μmol·h^–1·g^–1 for 2 wt % g-C₃N₄. A one-step self-assembly route was recently developed to fabricate core–shell architecture comprising carbon spheres decorated by g-C₃N₄. These composites showed extended light absorption and high mechanical stability, with enhanced conductivity for charge transport [105], delivering hydrogen evolution rates of 129 mol·h⁻¹, and 8-fold improvement over pristine g-C₃N₄ (16 mol·h⁻¹). Other g-C₃N₄ nanocomposites were investigated with a range of materials and morphologies [39,82,91,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128], to access different charge transfer mechanisms between g-C₃N₄ and the other components. These include a g-C₃N₄-based type II heterojunction [103], g-C₃N₄-based p-n heterojunction [91,129], g-C₃N₄-based Z-scheme heterojunction [113,130], g-C₃N₄/metal heterojunction [100,102], and a g-C₃N₄/carbon heterojunction [131]. The design of g-C₃N₄ heterojunction photocatalysts is an attractive strategy to tune the electronic structure and redox potentials for visible-light absorption photocatalytic H₂ generation. Table 1 compares the performance of different g-C₃N₄ photocatalysts.

Figure 18. (a,b) TEM and (c) HR TEM image of core–shell CdS@g-C₃N₄ heterojunction nanocomposite. Reprinted with permission from [104]. Copyright 2013 American Chemical Society.

Table 1. Photocatalytic H₂ production over g-C₃N₄ nanostructured catalysts.

Entry	Photocatalyst	Co-Catalyst (Loading)	Experimental Details	H2 Productivity/μmol·g−1·h−1	Reference Material/μmol·g−1·h−1	Enhancement Relative to Conventional g-C3N4	Apparent Quantum Efficiency/%	Reference
1	g-C3N4 nanosheets (thermal exfoliation)	Pt (6 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 400 nm)	170	bulk g-C3N4 31.48	5.4		[61]
2	g-C3N4 nanosheets (liquid exfoliation)	Pt (3 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	93 μmol	bulk g-C3N4	10		[64]
3	g-C3N4 nanosheets (thermal treatment)	Pt (0.5 wt %)	15 vol% TEOA 300 W Xe (λ > 420 nm)	1400	g-C3N4 450	3	2.6 (420 nm)	[132]
4	g-C3N4 nanosheets	Pt (1 wt %)	10 vol% TEOA full sunlight and λ > 400 nm	1395	bulk g-C3N4 250	5.6		[133]
5	Single layer g-C3N4	Pt (3 wt %)	10 vol% TEOA 500 W Xe (λ > 420 nm)	230	bulk g-C3N4 90	2.5		[96]
6	Urea derived g-C3N4	Pt (3 wt %)	~10 vol% TEOA 300 W Xe (λ ≥ 395 nm)	3327.5	DCDA derived g-C3N4 thiourea derived g-C3N4	7 10	26.5 (400 nm)	[134]
7	Nano Spherical-g-C3N4	Pt (3 wt %)	10 wt % TEOA 300 W Xe (λ > 420 nm)	14,350	Pt/bulk g-C3N4 318	45	9.6 (420 nm)	[81]
8	g-C3N4 nanostructure	Pt (3 wt %)	15 wt % TEOA 300 W Xe (λ > 420 nm)	689	bulk g-C3N4 8	8.6		[135]
9	Porous g-C3N4 microspheres	Pt (3 wt %)	15 wt % TEOA 300 W Xe (λ > 420 nm)	180	bulk g-C3N4 7.8	2.3	1.62 (420 nm)	[80]
10	Silica templated mesoporous g-C3N4	Pt (3 wt %)	10 vol% TEOA λ > 420 nm	237.4 (µmol−1 m−2)	g-C3N4 9.16 (μmol·h−1·m−2)	25.8		[136]
11	macroscopic 3D porous g-C3N4 monolith	Pt (3 wt %)	10 vol% TEOA 300 W Xe (λ > 420 nm)	29	g-C3N4 10.2	2.8		[137]
12	hollow g-C3N4 nanospheres	Pt (3 wt %)	10 wt % TEOA 300 W Xe	15,000	pure g-C3N4 5000	3		[85]
13	Iodine doped-g-C3N4	Pt (3 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	890	bulk g-C3N4 98.8	9		[138]
14	P doped-g-C3N4	Pt (3 wt %)	10 wt % TEOA 300 W Xe	2082	pure g-C3N4 226.3	9.2		[139]
15	O-doping supramolecular porous g-C3N4	Pt (3 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	1204	bulk g-C3N4 3D porous g-C3N4	6.1 3.1	7.8 (420 nm)	[140]
16	K-g-C3N4	Pt (0.5 wt %)	10 vol% TEOA 300 W Xe (λ > 400 nm)	1028	pure g-C3N4 73.4	14		[141]
17	AuPd/g-C3N4	Au and Pd	10 vol% TEOA 300 W Xe (λ ≥ 400 nm)	326	Au/g-C3N4 Pd/g-C3N4	3.5 1.6		[142]
18	Hydrogenated g-C3N4	Pt (3 wt %)	10 vol% TEOA 350 W mercury arc lamp (λ > 420 nm)	900	bulk g-C3N4 132.3	6.8		[143]
19	Surface alkalization of g-C3N4	Pt (1 wt %)	20 vol% aq. methanol 300 W Xe	2230	urea derived g-C3N4 159.3	14	6.67 (400 nm)	[144]
20	dye sensitized g-C3N4 nanosheets	Pt	5 vol% TEOA 300 W Xe (λ > 420 nm)	6525	Pt/g-C3N4 466	14	33.4 (460 nm)	[145]
21	2-Aminobenzonitrile-mp-g-C3N4	Pt (3 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	229	mp-g-C3N4 127	1.8		[146]
22	PPy-g-C3N4	Pt (3 wt %)	No sacrificial reagent 350 W Xe (λ > 400 nm)	154	Pt-g-C3N4	49.3		[147]
23	Cu2O@g-C3N4 core@shell		10 vol% TEOA 300 W Xe	202.28	Cu2O 35.08	5.7		[148]
24	CdS/g-C3N4 core/shell	Pt (0.6 wt %)	0.35 M Na2S and 0.25 M Na2SO3 300 W Xe (λ ≥ 420 nm)	4152	pure CdS 2001	2.1	4.3 (420 nm)	[104]
25	Core–shell Ni/NiO-decorated g-C3N4	Ni/NiO	10 vol% TEOA 300 W Xe	10	pure g-C3N4 1.01	10		[149]
26	MoS2/g-C3N4	N/A	10 vol% TEOA 300 W Xe (λ > 400 nm)	252	pure g-C3N4 31.5	8		[150]
27	CdS QD/g-C3N4	Pt (0.5 wt %)	0.1 M l-ascorbic acid (pH = 4) 300 W Xe (λ > 420 nm)	4494	pure g-C3N4 299	15		[151]
28	CdS nanorods/g-C3N4	NiS	10 vol% triethanolamine 300 W Xe (λ ≥ 420 nm)	2563	pure g-C3N4 1582	1.6		[152]
29	CaIn2S4/g-C3N4	Pt (1 wt %)	0.5 M Na2S and 0.5 M Na2SO3 300 W Xe	102	CaIn2S4 34	3		[153]
30	BiPO4/P-g-C3N4	N/A	Na2S (0.1 M) 300 W Xe (λ ≥ 420 nm)	1110	P-g-C3N4 676	1.6		[154]
31	AgQCs/g-C3N4	Pt (1 wt %)	25 vol% methanol simulator AM 1.5 G	5.59	pure g-C3N4 3.29	1.7		[155]
32	Al2O3/g-C3N4	Pt (1 wt %)	25 vol% TEOA 300 W Xe (λ ≥ 420 nm)	52.10	pure g-C3N4 20.75	2.5		[156]
33	MoS2/mp-g-C3N4	Pt	10 vol% lactic acid 300 W Xe (λ ≥ 420 nm)	1030	Pt/mp-g-C3N4 239.5	4.3	2.7 (420 nm)	[157]
34	carbon black/g-C3N4	Pt (3 wt %)	25 vol% methanol λ > 420 nm	689	pure g-C3N4 215	3.2		[158]
35	graphene/g-C3N4	Pt (1.5 wt %)	25 vol% methanol 350 W Xe (λ > 400 nm)	451	g-C3N4 150	3		[107]
36	carbon black/NiS/g-C3N4	NiS	15 vol% TEOA 300 W Xe (λ ≥ 420 nm)	992	g-C3N4/NiS 396	2.5		[159]
37	N,S-TiO2/g-C3N4	N/A	10 vol% methanol 125 W Hg lamp	317	g-C3N4 125	2.5		[160]
38	N-CeOx/g-C3N4	Pt (1 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	292.5	g-C3N4 134.5	2		[161]
39	g-C3N4 (2D)/CdS (1D)/rGO (2D)	Pt (1 wt %)	10 vol% TEOA 300 W Xe (λ ≥ 420 nm)	4800	pure g-C3N4 g-C3N4/rGO g-C3N4/CdS	44 11 2.5		[122]
40	Au/(P3HT)/Pt/g-C3N4	Au and Pt	10 vol% TEOA 300 W Xe (λ > 420 nm)	320	g-C3N4/Au; 73 and g-C3N4/Pt; 82	4		[162]

5.1.2. CO₂ Reduction

Rising atmospheric levels of carbon dioxide and the depletion of fossil fuel reserves raise serious concerns about the continued reliance on the use of fossil fuels for both energy and chemicals production [3,163], to which the photocatalytic reduction of CO₂ to light oxygenates and hydrocarbons could provide a sustainable solution. CO₂ reduction involves multi-electron transfer and hence the reaction kinetics for, e.g., formic acid, carbon monoxide, formaldehyde, methanol and methane production are intrinsically slower than for H₂ production. CO₂ photoreduction begins with molecular adsorption at the catalyst surface, wherein the anion radical is generated by the transfer of electrons photoexcited across the semiconductor band gap following light absorption. In the case of aqueous phase CO₂ reduction, charge-compensation occurs through concomitant water splitting and the transfer of photoexcited holes in the valence band onto hydrogen atoms, with the resulting protons migrating to the CO₂ anion. The reduction potentials for CO₂ photoreduction with water to various products are described below (relative to NHE at pH = 7) [11,164]:

$${\mathrm{CO}}_2+{\mathrm{e}}^{-}\to {\mathrm{CO}}_2^{•-}{E}^0=-1.90\mathrm{eV}$$

(5)

$${\mathrm{CO}}_2+{\mathrm{H}}^{+}+2{\mathrm{e}}^{-}\to {\mathrm{HCO}}_2^{•-}{E}^0=-0.49\mathrm{eV}$$

(6)

$${\mathrm{CO}}_2+2{\mathrm{H}}^{+}+2{\mathrm{e}}^{-}\to \mathrm{CO}+{\mathrm{H}}_2\mathrm{O}{E}^0=-0.53\mathrm{eV}$$

(7)

$${\mathrm{CO}}_2+4{\mathrm{H}}^{+}+4{\mathrm{e}}^{-}\to \mathrm{HCHO}+{\mathrm{H}}_2\mathrm{O}{E}^0=-0.48\mathrm{eV}$$

(8)

$${\mathrm{CO}}_2+6{\mathrm{H}}^{+}+6{\mathrm{e}}^{-}\to {\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2\mathrm{O}{E}^0=-0.38\mathrm{eV}$$

(9)

$${\mathrm{CO}}_2+8{\mathrm{H}}^{+}+8{\mathrm{e}}^{-}\to {\mathrm{CH}}_4+2{\mathrm{H}}_2\mathrm{O}{E}^0=-0.24\mathrm{eV}$$

(10)

$${\mathrm{CO}}_2+10{\mathrm{H}}^{+}+10{\mathrm{e}}^{-}\to {\mathrm{C}}_2{\mathrm{H}}_4+4{\mathrm{H}}_2\mathrm{O}{E}^0=-0.22\mathrm{eV}$$

(11)

$${\mathrm{CO}}_2+12{\mathrm{H}}^{+}+12{\mathrm{e}}^{-}\to {\mathrm{C}}_2{\mathrm{H}}_5\mathrm{OH}+3{\mathrm{H}}_2\mathrm{O}{E}^0=-0.33\mathrm{eV}$$

(12)

Key factors influencing CO₂ photocatalytic reduction include band energy matching, efficient charge-carrier separation, kinetic of e- and hole transfer to CO₂ and the reductant, and the basicity of the photocatalyst and hence strength and coverage of CO₂ adsorption [164]. In recent years, the g-C₃N₄ nanostructured materials have been studied for CO₂ photoreduction [92,165], due to their excellent stability, sufficiently negative CB energy and narrow band gap. Many strategies are reported to promote g-C₃N₄ with condensed matter and molecular sensitizers [166,167], such as doping with metals [168,169] and non-metal [170,171,172], heterojunction construction [173,174,175,176] and Z-scheme composites employing co-catalysts [165,166,167,173,175]. Pengfei et al. reported ultrathin C₃N₄ nanosheets for enhanced photocatalytic CO₂ reduction [177] in which surface functionalization and textural modification by NH₃-mediated thermal exfoliation enhanced light harvesting, charge-carrier redox potentials, and the surface area for CO₂ adsorption (to 0.2 mmol·g⁻¹), resulting in CH₄ and CH₃OH productivities of 1.39 and 1.87 μmol·h⁻¹·g⁻¹ respectively, a five-fold increase over bulk g-C₃N₄. Jiaguo and co-workers [168] demonstrated that Pt promotion significantly influenced both the activity and selectivity of g-C₃N₄ for CO₂ photoreduction to CH₄, CH₃OH, and HCHO; Pt nanoparticles improved charge separation across the metal/semiconductor interface, and lowered the overpotential for CO₂ reduction. Qingqing et al. reported Pd nanoicosahedrons with twin defects promoted CO₂ reduction into CO and CH₄ over C₃N₄ nanosheets [126]. CO₂ conversion reached 61.4%, with an average CO productivity of 4.3 μmol·g⁻¹·h⁻¹ and average CH₄ productivity of 0.45 μmol·g⁻¹·h⁻¹, indicating the presence of highly reactive sites for CO₂ adsorption and activation.

Hierarchical, porous O-doped g-C₃N₄ nanotubes prepared via successive thermal oxidation exfoliation and condensation of bulk g-C₃N₄ also show promise for photocatalytic CO₂ reduction under visible light [171]. As-prepared O-doped g-C₃N₄ nanotubes comprise interconnected, multi-walled nanotubes with uniform diameters of 20–30 nm, which evolve methanol at 0.88 µmol·g⁻¹·h⁻¹, five times faster than bulk g-C₃N₄ (0.17 µmol·g⁻¹·h⁻¹). Heterojunction composites of g-C₃N₄/ZnO synthesized by a one-step calcination route [165] are also superior to bulk g-C₃N₄ (2.5-fold enhancement), ascribed to a direct Z-scheme mechanism reflecting efficient ZnO → g-C₃N₄ electron transfer occurring the interface. Zhongxing et al. reported that CeO₂-modified C₃N₄ photocatalysts produced by a simple hydrothermal route were effective for the selective photocatalytic reduction of CO₂ to CH₄ [178], with a CH₄ productivity of 4.79 mmol·g⁻¹·h⁻¹, about 3.44 times that of g-C₃N₄. Wang et al. prepared a 2D-2D MnO₂/g-C₃N₄ heterojunction photocatalyst by an in-situ redox reaction between KMnO₄ and MnSO₄ adsorbed at the surface of g-C₃N₄ [179] for photocatalytic CO₂ reduction to CO (9.6 mmol·g⁻¹), in which band matching facilitated efficient separation of photogenerated charge-carriers_. Photocatalytic CO₂ reduction reaction is also reported over a direct Z-scheme g-C₃N₄/SnS₂ catalyst [180] which yielded both CH₃OH (2.3 µmol·g⁻¹) and CH₄ (0.64 µmol·g⁻¹), with electrons in SnS₂ combining with holes in g-C₃N₄. Another Z-scheme mechanism is invoked for a MoO₃/g-C₃N₄ composite [181]. Ryo and co-workers adopted a different approach, attaching Ru(bipy)complexes to g-C₃N₄ nanostructures; these displayed improved activity for CO₂ photoreduction to formic acid, with a high apparent quantum yield of 5.7% at 400 nm under visible light (Figure 19). Anchoring of polyoxometalate clusters to C₃N₄ also creates active photocatalysts for CO₂ reduction [179]. Here, noble-metal-free Co₄ polyoxometallates were used to achieve a staggered band alignment, with the Co4@g-C₃N₄ hybrid photocatalysts achieving 107 μmol·g⁻¹·h⁻¹ and 94% selectivity for CO production under visible light (λ ≥ 420 nm); cumulative CO production reached 896·μmol·g⁻¹ after 10 h irradiation, far exceeding that for unpromoted g-C₃N₄.

Figure 19. CO₂ reduction using a Ru complex/C₃N₄ hybrid photocatalyst, and structures of the Ru complexes. CB = conduction band, VB = valence band. Reproduced with permission from [166]. Copyright John Wiley & Sons Inc., 2015.

A multicomponent heterostructure, termed an intercorrelated superhybrid, comprising AgBr supported on g-C₃N₄ decorated in turn on N-doped graphene (prepared by wet-chemical synthesis) has also shown excellent activity for the photocatalytic reduction of CO₂ to methanol and ethanol (Figure 20) [174]. Oluwatobi et al. reported g-C₃N₄/(Cu/TiO₂) [182] nanocomposites prepared by pyrolysis and impregnation for enhanced photoreduction of CO₂ to CH₃OH and HCOOH under UV-vis irradiation wherein maximum productivities of CH₃OH and HCOOH under visible light were 2574 and 5069 mmol·g⁻¹ respectively. Enhanced photoactivity was attributed to the location of the metal within the composite and consequent distribution of photoexcited electrons. Hailong et al. also studied g-C₃N₄/Ag-TiO₂ hybrid photocatalysts [183], wherein CO and CH₄ were preferentially formed, with a maximum CO₂ conversion of 47 µmol·g⁻¹, and product yields of 28 µmol·g⁻¹ CH₄ formation and 19 µmol·g⁻¹ CO. Enhanced activity was proposed to arise from the transfer of photoexcited electrons across the g-C₃N₄/TiO₂ heterojunction, and subsequently from TiO₂ → Ag nanoparticles due to the lower Fermi level; this spatial separation of charge greatly suppressed the electron–hole recombination, with electrons accumulating on the Ag nanoparticles on the TiO₂ surface.

Figure 20. (a) Synthetic strategy, and (b,c) photocatalytic performance for CO₂ reduction of intercorrelated superhybrid g-C₃N₄ nanocomposites under visible light and corresponding apparent quantum efficiencies. Reproduced with permission from [174]. Copyright John Wiley & Sons Inc., 2015.

Table 2 compares the performance of different g-C₃N₄ photocatalysts for photocatalytic CO₂ reduction.

Table 2. Photocatalytic CO₂ reduction over g-C₃N₄ nanostructured catalysts.

Entry	Photocatalyst	Experimental Details	Productivity/μmol·g−1·h−1	Reference Material/μmol·g−1·h−1	Enhancement Relative to Conventional g-C3N4	Apparent Quantum Efficiency/%	Reference
1	g-C3N4 nanosheets	300 W Xe (l > 420 nm), 15 °C and 25 kPa CO2, catalyst in 80 mL of H2O	CH4: 0.94	Bulk g-C3N4: 0.30	3.1		[184]
2	g-C3N4 nanosheets	300 W Xe (400 nm), 200 mW/cm2. 20 mg catalyst in 0.1 mL H2O, CO2 bubbled to 0.06 MPa	CH4: 1.2 CH3OH: 0.2	Bulk g-C3N4 CH4: 0.28 CH3OH: 0.24	CH4: 4.3		[185]
3	Ultrathin g-C3N4 nanosheets	300 W Xe, 100 mg catalyst, 0.084 g NaHCO3 + H2SO4 to release CO2	CH4: 1.39 and CH3OH: 1.87	Bulk g-C3N4 CH4: 0.14 and CH3OH: 0.35	CH3OH: 5.34		[177]
4	Thiourea and urea derived g-C3N4	300 W Xe/420 nm, 40 mg catalyst	Urea derived g-C3N4 CO: 0.56, CH3CHO: 0.44, CH4: 0.04 thiourea derived g-C3N4 CO: 0.36, CH3CHO: 0.26, CH4 = 0.025	N/A	N/A		[186]
5	Melamine and urea derived g-C3N4	300 W Xe (420 nm), 0.2 g and 1.0 M NaOH solution (100 mL)	Urea derived g-C3N4 CH3OH: 6.28, C2H5OH: 4.51, O2: 21.33 melamine derived g-C3N4 CH3OH: TRACE, C2H5OH: 3.64, O2: 10.29	N/A	N/A	Urea derived g-C3N4: 0.18, melamine derived g-C3N4: 0.08	[172]
6	Thiourea, urea and DCDA derived g-C3N4	300–795 nm KG1 filter, 40 mW cm2 illumination, 0.5 mg catalyst per mL in CH3CN/TEOA/H2O (3:1:1), t = 2 h, [Co(bpy)n]2+ as a co-catalyst	Urea derived g-C3N4 CO: 460, H2: 138 μmol thiourea derived g-C3N4 CO: 22, H2: 86 μmol DCDA derived g-C3N4 CO: 92, H2: 94 μmol	N/A	N/A		[167]
7	Sulfur-doped g-C3N4	300 W simulated solar Xe and 200 mL Pyrex reactor, 100 mg 1 wt % Pt co-catalyst, 0.12 g NaHCO3 and 0.25 mL 4 M HCl solution	CH3OH: 0.37	Bulk g-C3N4 CH3OH: 0.27	1.37		[170]
8	Pd/g-C3N4	300 W Xe/UV420 cut-off filter	CO: 0.5, CH4: 0.05, CH3OH: 1 μmol·g−1	Bulk g-C3N4 CO: 4, CH4: 0.15, CH3OH: 2.5 μmol·g−1			[187]
9	Pt-loaded g-C3N4	15 W energy-saving daylight bulb, flow rate of CO2 fixed at 5 mL·min−1	CH4: 1.3	Bulk g-C3N4 CH4: 0.25	5.2		[188]
10	Pt-g-C3N4	200 mL Pyrex reactor, 300 W simulated solar Xe, 100 mg catalyst, NaHCO3 (0.12 g) and HCl aq. solution (0.25 mL, 4 M)	CH4: 0.25, CH3OH: 0.25, HCHO: 0.125	Bulk g-C3N4 CH4: 0.07, CH3OH: 0.11, HCHO: 0.06	CH4: 3.57		[168]
11	Amine-functionalized g-C3N4	300 W Xe, Pyrex 200 mL, 100 mg catalyst, 0.084 g NaHCO3 + 0.3 mL of 2 M H2SO4	CH4: 0.34 CH3OH: 0.28	Bulk g-C3N4 CH3OH: 0.26 CH4: trace	CH4: 1.3		[189]
12	SnO2-coupled B and P co-doped g-C3N4	300 W Xe (420 nm), 0.2 g catalyst in 3 mL water/100 mL NaOH purged with CO2	CH4: 30	Bulk g-C3N4 CH4: 3.5	8.57	2.02 (420 nm)	[190]
13	g-C3N4-Ru complex	400 W Hg lamp (400 nm) 11 mL reactor containing 4 mL 20 vol % TEA in acetonitrile and 8 mg catalyst purged with CO2	HCOOH: 4.6	Bulk g-C3N4 HCOOH: trace	N/A		[191]
14	Ag3PO4/g-C3N4	500 W Xe/420 nm, stainless-steel reactor 132 mL, 10 mg in 4 mL H2O, 0.4 MPa CO2 at 80 °C	CO: 44, CH3OH: 9, CH4: 0.2, C2H5OH: 0.1	Bulk g-C3N4 CO: 4, CH3OH: 0.35, CH4: 0.09, C2H5OH: 0.01	CO: 11		[175]
15	AgX/g-C3N4 (X = Cl and Br)	15 W energy-saving daylight lamp, 100 mg catalyst, CO2 flow of 5 mL/min	CH4: 1.282	Bulk g-C3N4 CH4: 0.388	3.3		[192]
16	B4C/g-C3N4	300 W Xe (UV/IR filter), 100 mL photoreactor, 6 mg catalyst, CO2	CH4: 0.84	Bulk g-C3N4 CH4: 0.14	6		[193]
17	BiOI/g-C3N4	300 W Xe (400 nm), 0.10 g catalyst, CO2 bubbled through water.	CO: 3.58, O2: 1.96, H2: 0.4, CH4: 0.2	Bulk g-C3N4 CO: 0.2, O2: 0.56, H2: 0.92	CO: 17.9		[194]
18	g-C3N4/C	500 W Xe lamp, 0.1 g catalyst, CO2 + H2O mixture flow 20 mL min−1, 30 °C and 110 KPa CO2	CO: 2.5 CH4: 1.4	Bulk g-C3N4 CO: 1.1 CH4: 0.72	CO: 2.27		[195]
19	CeO2/g-C3N4	300 W Xe, reactor volume 500 mL, 50 mg catalyst, CO2 bubbled through water	2 wt % CO: 11.8 and CH4: 9.08 3 wt % CO: 10.16 and CH4: 13.88	Bulk g-C3N4 CO: 6.78 CH4: 0.2	CH4: 69.4		[196]
20	Graphene/g-C3N4	15 W energy saving daylight bulb, CO2 5 mL min−1	CH4: 0.59 μmol·h−1	Bulk g-C3N4 CH4: 0.25 μmol·h−1	2.36		[197]
21	g-C3N4/NaNbO3	300 W Xe, reaction volume 230 mL, 50 mg catalyst, reactor purged with CO2, then 2 mL H2O injected	CH4: 6.4	Bulk g-C3N4 CH4: 0.8	8		[173]
22	g-C3N4/N-TiO2	300 W Xe lamp, reaction system vol 780 mL, 0.1 g catalyst, flow rate of CO2 15 mL min−1	CO: 14.73 μmol	Bulk g-C3N4 CO: 4.20 μmol; P25: 3.19 μmol	3.5		[198]
23	rGO/g-C3N4	15 W energy-saving daylight lamp, CO2 at a flow rate of 5 mL/min, 100 mg catalyst	CH4: 14	Bulk g-C3N4 CH4: 2.5	5.6	0.56 (420 nm)	[199]
24	g-C3N4 and a Ru(II) complex	400 W high-pressure Hg lamp, 8 mg catalyst, DMA (containing 20 vol % TEOA) 4.0 mL	CO: 2.9 μmol·h−1, HCOOH: 1.5 μmol·h−1; H2: 0.13 μmol·h−1	Bulk g-C3N4 Only trace	N/A		[200]
25	Ru complex/mp g-C3N4	450 W Xe lamp, 8.0 mg catalyst, acetonitrile and triethanolamine (4:1 v/v) 4 mL mix in 11 mL Pyrex test tube	CO: 0.6, H2: 0.25, HCOOH: 4 μmol·h−1	Bulk g-C3N4 HCOOH: trace	N/A		[201]
26	SnO2/g-C3N4	500 W Xe, 20 mg catalyst, 4 mL water injected into the bottom of the reactor, 0.3 MPa CO2, 80 °C	CO: 19, CH4: 2, CH3OH: 3	Bulk g-C3N4 CO: 2.4, CH4: trace, CH3OH: 2.8, P25: CO: 3.5, CH3OH: 1	CO: 7.9		[202]
27	Brookite TiO2/g-C3N4	300 W Xe, 60 mg catalyst, CO2 produced from reaction of NaHCO3 (1.50 g) and H2SO4 solution (5.0 mL, 4 M)	CO: 0.84, CH4: 5.21	Bulk g-C3N4 CO: 7.10, CH4: 1.84	CH4: 2.83		[203]
28	TiO2/g-C3N4	8 W Hg lamp (λ = 254 nm; intensity = 0.5 mW/cm2), vol of SS reactor 355 cm3, 0.1 g catalyst, 140 kPa CO2	CO: 2.8, CH4: 8.5, H2:41	Bulk g-C3N4 CO: 0.93, CH4: 4.75, H2: 16.25	CO: 3		[204]
29	g-C3N4/WO3	LED (λ = 435 nm) at 3.0 mW cm2, 3 mg catalyst in 5 mL ion-exchanged water	CH3OH: 1.1 μmol, 0.5 wt % Au and Ag 2.5 and 1.5 μmol, resp.	Bulk g-C3N4 CH3OH: 0.6 μmol	1.83		[205]
30	g-C3N4/ZnO	300 W Xe lamp, 200 mL Pyrex reactor, 100 mg catalyst CO2 and H2O vapor produced by NaHCO3 (0.12 g) and HCl (0.25 mL, 4 M)	CH3OH: 0.6	Bulk g-C3N4: CH3OH: 0.26 Pure ZnO: CH3OH: 0.37	2.3		[165]
31	ZnO/g-C3N4	500 W Xe/420 nm, steel reactor 132 mL, 10 mg catalyst in 4 mL H2O, 0.4 MPa CO2 and 80 °C	CO: 29, CH3CHO: 9, CH4: 3.5, C2H5OH: 1.5	Bulk g-C3N4 CO: 4.5, CH3CHO: 4.3, CH4: 0.5, C2H5OH: trace P25CO: 4.5, CH3CHO: 3, CH4: 2, C2H5OH: trace	CO: 6.4		[206]
32	Co-porphyrin/g-C3N4	300 W Xe (UV/IR cut-off filter), 1 mL of TEOA and 4 mL of MeCN were mixed and injected into the cell, 80 kPa CO2	CO: 17	Bulk g-C3N4 CO: 1.4	12.14	0.80 (420 nm)	[207]
33	Co-(bpy)3Cl2/g-C3N4	300 W Xe lamp with a 420 nm cut-off, 50 mg catalyst, MeCN (4 mL), TEOA (2 mL), CO2 (1 bar), 60 °C	CO: 37 H2: 6		N/A		[176]
34	g-C3N4/Bi2WO6	300 W Xe/420 nm cut-off filter, reactor 500 mL, 0.1 g catalyst, CO2 and H2O vapour mixer	CO: 5.19	pure g-C3N4 CO: 0.23 Bi2WO6CO: 0.81	22		[208]
35	g-C3N4/Bi4O5I2	300 W Xe lamp with 400 nm cut-off filter, 0.10 g catalyst, Pyrex glass 350 mL, 5 mL H2SO4 (4 M) with NaHCO3 to achieve 1 bar CO2, 15 °C	CO: 45.6	Bulk g-C3N4 CO: 5.8	7.86		[209]
36	Core–shell LaPO4/g-C3N4 nanowires	300 W Xe lamp, reactor volume 500 mL, 30 mg catalyst, CO2 and water vapor	CO: 14.43	0.41	10		[210]
37	CdIn2S4/mp g-C3N4	300 W Xe lamp with 420 nm cut-off filter, 0.1 g catalyst in 100 mL water containing 0.1 M NaOH, ultrapure CO2 was continuously bubbled through	CH3OH: 42.7	pure CdIn2S4 CH3OH: 23.1	1.84	0.14 (420 nm)	[211]
38	Mesoporous phosphorylated g-C3N4	300 W Xe lamp, Pyrex glass 350 mL, 0.2 g catalyst, 5 mL of 4 M H2SO4 with NaHCO3 (1.0 g) to give 1 bar CO2 10 °C	CO: 20, CH4: 40, H2: 3, O2: 10	CO: 4.5, CH4: 4, H2: 0.5, O2: 1.75	CH4: 10	0.85 (420 nm)	[212]
39	Pt-g-C3N4/KNbO3	300 W Xe lamp with 420 nm cut-off filter, 0.1 g catalyst, CO2, 2 mL of H2O	CH4: 2.37	CH4: 0.62	3.8		[213]
40	g-C3N4/BiOBr/Au	300 W Xe lamp (λ = 380 nm), 350 mL Pyrex glass, 0.1 g catalyst, 5 mL H2SO4 (4 M) + 1.3 g NaHCO3 to give 1 bar CO2	CO: 6.67 CH4: 0.92	N/A	N/A		[214]
41	g-C3N4/Ag-TiO2	300 W Xe, 50 mg catalyst, CO2 flow rate of 3 mL·min−1, 45 °C	CH4: 9.33 and CO: 6.33	N/A	N/A		[183]

5.2. Environmental Remediation

Many large-scale processes operated by the petrochemical, textile and food industries discharge polluted water into the aquatic environment [215]. Organic dyes are often used in textile, printing, and photographic industries, and a sizable fraction of these are lost during the dying process into effluent wastewater streams. Even low concentrations of such dyes pose serious risks to human and animal health, and their bio- or chemical degradation is challenging [216,217], hence the development advanced oxidation processes (AOPs) to treat contaminated drinking ground and surface waters, and wastewaters containing toxic or non-biodegradable compounds are sought [218,219]. Semiconductor photocatalysis offer an effective and economic approach to the treatment of recalcitrant organic compounds at low concentrations in wastewater [220,221,222,223]. Photoexcited holes are the key active species in such photocatalytic environmental remediation, being powerful oxidants in their own right, or reacting with water to produce hydroxyl radicals (^•OH) which are themselves powerful oxidants with an oxidation potential of 2.8 eV (NHE). Reactively-formed ^•OH can rapidly attack adsorbed pollutants at the surface of photocatalysts or in solution, to achieve their mineralization as CO₂ and water. Mechanisms for the photocatalytic oxidation of organic pollutants in water are widely discussed in the literature [4,221,222]. Briefly:

$$\mathrm{SC}+\mathrm{h}\nu \to \mathrm{SC}\times \left({\mathrm{e}}_{\mathrm{CB}}^{-}+{\mathrm{h}}_{\mathrm{VB}}^{+}\right)$$

(13)

$${\mathrm{h}}_{\mathrm{VB}}^{+}+{\mathrm{H}}_2\mathrm{O}\to {}^{•}\mathrm{O}\mathrm{H}+{\mathrm{H}}^{+}$$

(14)

$${\mathrm{O}}_2+{\mathrm{e}}_{\mathrm{CB}}^{-}\to {\mathrm{O}}_2^{•-}$$

(15)

$${\mathrm{O}}_2^{•-}+{\mathrm{H}}^{+}\to {\mathrm{HO}}_2^{•}$$

(16)

$${\mathrm{HO}}_2^{•}+{\mathrm{HO}}_2^{•}\to {\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{O}}_2$$

(17)

$${\mathrm{O}}_2^{•-}+{\mathrm{HO}}_2^{•}\to {\mathrm{O}}_2+{\mathrm{HO}}_2^{-}$$

(18)

$${\mathrm{HO}}_2^{-}+{\mathrm{H}}^{+}\to {\mathrm{H}}_2{\mathrm{O}}_2$$

(19)

$${\mathrm{H}}_2{\mathrm{O}}_2+\mathrm{h}\nu \to 2{}^{•}\mathrm{O}\mathrm{H}$$

(20)

$${\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{O}}_2^{•-}\to {}^{•}\mathrm{O}\mathrm{H}+{\mathrm{OH}}^{-}+{\mathrm{O}}_2$$

(21)

$${\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{e}}_{\mathrm{CB}}^{-}\to {}^{•}\mathrm{O}\mathrm{H}+{\mathrm{OH}}^{-}$$

(22)

$$\mathrm{Organic}\mathrm{Compound}+{}^{•}\mathrm{O}\mathrm{H}\to \mathrm{degradation}\mathrm{products}$$

(23)

$$\mathrm{Organic}\mathrm{Compound}+\mathrm{SC}\left({\mathrm{h}}^{+}\right)\to \mathrm{degradation}\mathrm{products}$$

(24)

$$\mathrm{Organic}\mathrm{Compound}+\mathrm{SC}\left({\mathrm{e}}^{-}\right)\to \mathrm{degradation}\mathrm{products}$$

(25)

A variety of active radicals, including O₂^•−, ^•OH, HO₂^•, in addition to H₂O₂ have been invoked as the oxidants responsible for mineralization, with ^•OH the most likely candidate Equation (23). Direct oxidation of carboxylic acids by photoexcited holes to generate CO₂ Equation (24) has also been evidenced, termed the ‘photo-Kolbe reaction’. Reductive pathways involving photoexcited electrons Equation (25) are considered unimportant in dye degradation; however, thermodynamic requirements for semiconductor photocatalysts dictate that the VB and CB should be positioned such that the oxidation potential of hydroxyl radicals $E_{\left({\mathrm{H}}_2\mathrm{O}/{}^{•}\mathrm{O}\mathrm{H}\right)}^0=+2.8\mathrm{eV}\left(\mathrm{NHE}\right)$ and reduction potential of superoxide radicals $E_{\left({\mathrm{O}}_2/{\mathrm{O}}_2^{•-}\right)}^0=-0.3\mathrm{eV}\left(\mathrm{NHE}\right)$ lie well within the band gap. In other words, the redox potential of photoexcited holes must be sufficiently positive to generate ^•OH radicals, and that of photoexcited electrons sufficiently negative to generate O₂^•−.

Considerable efforts have been devoted to developing photocatalysts for water purification under solar irradiation. g-C₃N₄ based nanostructures are potential photocatalysts for the degradation of various pollutants [39,42], with photophysical properties of the parent nitride modified through doping with heteroatoms, heterojunction formation with other materials, and textural improvements to enhance surface area and porosity. For example, ultrathin g-C₃N₄ nanosheets derived from bulk g-C₃N₄ by exfoliation in methanol exhibit enhanced photocatalytic performance for methylene blue (MB) degradation [65]. g-C₃N₄ nanotubes show superior photoactivity under visible light for MB degradation than bulk g-C₃N₄ or P25 [74]. Tahir and co-workers also employed tubular g-C₃N₄ for MB and methyl orange (MO) photocatalytic degradation under visible light, observing better stability and activity than bulk g-C₃N₄, attributed to the high surface area (182 m²·g⁻¹) and improved light absorption and charge separation/transfer [75]. 1D g-C₃N₄ nanorods with different aspect ratios have been screened for MB degradation under visible light (λ > 420 nm) and simulated solar irradiation (λ > 290 nm) [68]. The resulting photocatalytic activity and photocurrent response of g-C₃N₄ nanorods under visible light were 1.5–2.0 times that of g-C₃N₄ nanoplates. A simple chemical route was reported for preparing nanofiber-like g-C₃N₄ structures which showed promising activity for Rhodamine B (RhB) photodegradation [73].

g-C₃N₄ doping is a common strategy to broaden spectral utilization and band alignment to drive separate photogenerated charge carriers. Doping by metals such as Cu and Fe [224,225,226], non-metals such as B, C, O, or S [224,227,228,229,230,231], and co-doping [232,233,234] have all been employed for environmental depollution applications. For example, S and O co-doped g-C₃N₄ prepared by melamine polymerization and subsequent H₂O₂ activation prior to trithiocyanuric acid functionalization (Figure 21a) enhanced the photocatalytic degradation of RhB (Figure 21b) 6-fold relative to the parent g-C₃N₄ nanosheet [235]. Doping resulted in a strongly delocalized HOMO and LUMO that increased the number of active sites and improved the separation of photogenerated electrons and holes.

Figure 21. (a) Synthetic strategy, and (b) photocatalytic activity of S and O co-doped g-C₃N₄ for RhB degradation. Reproduced with permission from [235]. Copyright Royal Society of Chemistry, 2017.

Plasmonic photocatalysts have also been exploited for environmental remediation, for example, 7–15 nm Au and Pt nanoparticles photodeposited on g-C₃N₄ are promising for the photocatalytic degradation of tetracycline chloride as a representative antibiotic whose uncontrolled release is of concern [236]. The Au surface plasmon resonance broadens the optical adsorption range, while Pt acts as a sink for photoexcited electrons. The combination of noble metals and g-C₃N₄ enables tunable heterojunctions with improved charge transport than traditional nanocomposites [237,238,239,240,241,242,243], and such multicomponent heterostructures are a promising solution to environmental depollution [39,40,42], for example g-C₃N₄/Ag₃PO₄ systems for MO degradation [242,243]. Ag₃PO₄@g-C₃N₄ core–shell photocatalysts have also been applied to MB degradation under visible light, achieving 97% conversion in 30 min compared with only 79% for a physical mixture of the Ag₃PO₄ and g-C₃N₄ components, and 69% for pure Ag₃PO₄. The g-C₃N₄ shell may protect Ag₃PO₄ from dissolution in the composite, conferring superior stability. Core–shell g-C₃N₄@TiO₂ photocatalysts synthesized by a sol–gel and in situ re-assembly route and subsequently applied to phenol removal under visible light were seven times more photoactive than bulk g-C₃N₄. Increasing the g-C₃N₄ shell thickness from 0 to 1 nm increased the photodegradation rate constant from 0.0018 to 0.0386 h⁻¹; however, thicker shells slowed charge transport to the external photocatalyst surface, lowering activity. Z-scheme N-doped ZnO/g-C₃N₄ hybrid core–shell nanostructures (Figure 22Aa,b) were successfully prepared via a facile, low-cost, and eco-friendly ultrasonic dispersion method [244]. The g-C₃N₄ shell thickness was tuned by varying the g-C₃N₄ loading. Direct contact between the N-doped ZnO core and g-C₃N₄ shell introduced a new energy level into the N-doped ZnO band gap, effectively narrowing the band gap. Consequently, these hybrid core–shell nanostructures showed greatly enhanced visible light photocatalysis for RhB degradation compared to pure N-doped ZnO surface or g-C₃N₄ components (Figure 22Ac) [240]. A facile, reproducible, and template-free synthesis has also been demonstrated to prepare magnetically separable g-C₃N₄−Fe₃O₄ nanocomposites (Figure 22Ba) [37]. Monodispersed Fe₃O₄ nanoparticles with 8 nm diameter were uniformly deposited over g-C₃N₄ sheets (Figure 22Bb) and exhibited enhanced charge separation and photocatalytic activity for RhB degradation under visible light irradiation (Figure 22Bc). These g-C₃N₄−Fe₃O₄ nanocomposites showed good stability with negligible loss in photocatalytic activity even after six recycles, and facilitated magnetic catalyst recovery (Figure 22Bd). Xiao et al. demonstrated that the excellent stability of g-C₃N₄ towards photocatalytic oxidation in the presence of organic pollutants reflects strong competition of the latter for ^•OH radicals under practical working conditions, resulting in preferential decomposition of the pollutants rather than the carbon nitride [245].

Figure 22. (A) (a,b) TEM images of N-ZnO-g-C₃N₄ core–shell nanoplates, and associated (c) Z-scheme mechanism. Reproduced from with permission from [244]. Copyright 2014 Royal Society of Chemistry. (B) (a) Synthetic strategy, (b) TEM image, and (c) photodegradation mechanism for g-C₃N₄−Fe₃O₄ nanocomposite, and (d) magnetic separation of photocatalyst post-reaction. Reprinted with permission from [37]. Copyright 2013 American Chemical Society.

Several multicomponent nanocomposites based on g-C₃N₄ nanosheets such as Au@g-C₃N₄–PANI [246], Au-NYF/g-C₃N₄ [105], g-C₃N₄/CNTs/Al₂O₃ [247], AgCl/Ag₃PO₄/g-C₃N₄ [248], and g-C₃N₄/Zn_0.11Sn_0.12Cd_0.88S_1.12 [249] are also reported; the performance of different g-C₃N₄ photocatalysts for the photodegradation of representative aqueous pollutants is summarized in Table 3.

Table 3. Photocatalytic degradation of aqueous pollutants over g-C₃N₄ nanostructured catalysts.

Entry	Photocatalyst	Organic Molecule	Experimental Details	Removal Efficiency/%	Reference Material Efficiency/%	Enhancement Relative to Conventional g-C3N4	Reference
1	g-C3N4@TiO2 core–shell structure	Phenol	5 mg·L−1 phenol with 25 mg catalyst. 500 W Xe lamp with 420 nm cut-off filter, 23 mW/cm2.	30	4.2	7.2	[250]
2	Ag-decorated S-doped g-C3N4	Bisphenol A (BPA)	50 mL of 10 mg·L−1 of BPA, catalyst loading of 0.60 g·L−1. Light source, 155 W Xe arc lamp with the solar region of 280–630 nm.	95	31.66	3	[233]
3	Ultrathin urea-derived g-C3N4 nanosheets	p-Nitrophenol (PNP)	100 mg catalyst, aqueous PNP (10 mg L−1, 100 mL). 300 W Xe lamp equipped with an IR cut filter and a 400 nm cut filter.	95	60	1.58	[251]
4	Mesoporous g-C3N4/TiO2	Decomposition of dinitro butyl phenol (DNBP)	25 mg catalyst added to DNBP aqueous solution (20 mg·L−1) with 500 W xenon lamp with λ < 420 nm using cut-off filter.	98.5	65	1.5	[252]
5	C3N4-nanosheets	Methylene blue (MB)	10 mg catalyst in 50 mL of 10 mg·L−1 MB solution. 150 W Xe lamp as the simulated sunlight source.	98	7.9	12.4	[253]
6	Z-scheme graphitic-C3N4/Bi2MoO6	Methylene blue	30 mL of 10 mg·L−1 MB solution, 0.03 g catalyst. 50 W LED light with of 410 nm emission.	90	18.75	4.8	[254]
7	Sm2O3/S-doped g-C3N4	Methylene blue	100 mL of MB solution (8 mg·L−1), 300 W halogen lamp with UV-stop feature.	93	27	3.5	[255]
8	Porous CeO2/sulfur-doped g-C3N4	Methylene blue	0.06–0.12 g catalyst in 6–14 mg L−1 MB, visible light (λ > 400 nm) 300 W Halogen lamp with UV stop.	91.4	25	3.65	[256]
9	ZnS/g-C3N4	Methylene blue	200 mL MB (6 mg·L−1), 30 mg catalyst under visible light source, 100 W halogen lamp.	90	34.6	2.6	[257]
10	Mesoporous Carbon Nitride Decorated with Cu Particles	Methyl orange (MO)	0.07 g catalyst in 100 mL of MO (11 mg L−1) solution under visible-light, 300 W halogen lamp with UV-stop feature.	100	28	3.57	[258]
11	Plasmonic Ag–AgBr/g-C3N4	Methyl orange	MO solution (100 mL, 10 mg L−1), 50 mg catalyst, 300 W Xe lamp with 400 nm cut-off filter.	90	14.3	6.3	[259]
12	ZnFe2O4 nanoparticles on g-C3N4 sheets	Methyl orange	100 mL of 10 mg·L−1 MO solution, 25 mg catalyst. 500 W Xe lamp with cold filter.	98	15.31	6.4	[260]
13	AgNPs/g-C3N4 nanosheets	Methyl orange	50 mL 0.02 mmol/L MO solution, 25 mg catalyst. 300 W Xe lamp with a visible light reflector (350 nm < l < 780 nm) and a 420 nm longwave-pass cut-off filter (l > 420 nm).	95.2	13.8	7	[261]
14	BiOCl/C3N4 hybrid nanocomposite	Methyl orange	15 mL of 10 mg L−1 MO solution, 10 mg catalyst. 300 W Xe lamp equipped with 420 nm cut-off filter.	84.28	14	6	[262]
15	g-C3N4/GO aerogel	Methyl orange	50 mL of 20 mg L−1 MO solution. 300 W Xe lamp with a cut off filter (λ > 420 nm).	91.1	33	2.76	[263]
16	g-C3N4 nanocrystals decorated Ag3PO4 hybrids	Methyl orange	80 mL MO, 80 mg catalyst. 500 W halogen lamp equipped with cut-off filters (420 nm < λ < 800 nm).	92	44	2	[264]
17	g-C3N4-NS/CuCr2O4 nanocomposites	Rhodamine B (RhB)	250 mL of 2.5 × 10−5 M RhB solution, 0.1 g of catalyst. 50 W LED lamp.	98.9	30	3.3	[265]
18	Porous Mn doped g-C3N4	Rhodamine B	100 mL of 10 mg·L−1 RhB solution, 50 mg catalyst. 300 W Xe lamp equipped with ultraviolet cut-off filter (>400 nm).	88.9	18	4.9	[266]
19	Mesoporous carbon nitride (mpg-C3N4/SnCoS4)	Rhodamine B	100 mL of 20 mg·L−1 RhB solution, 20 mg catalyst. 300 W Xe lamp equipped with an UV cut-off filter (λ ≥ 420 nm).	70	13	5.4	[267]
20	Iron oxyhydroxide/ultrathin g-C3N4 nanosheets	Rhodamine B	50 mL of 10 mg·L−1 RhB solution, 50 mg catalyst. 500 W Xe lamp equipped with a cut-off filter (λ ≥ 420 nm).	98	5.5	17.8	[268]
21	Two-dimensional g-C3N4/Bi2WO6	Rhodamine B	100 mL of 10 mg L−1 RhB solution, 100 mg catalyst. 300 W Xe lamp with UV cut-off filter.	80	23.5	3.4	[269]
22	Ultrathin g-C3N4 nanosheets	Rhodamine B	100 mL of 20 mg L−1 RhB solution, 100 mg catalyst. 300 W Xe lamp (>420 nm).	99	16.2	6.1	[270]
23	Z-scheme g-C3N4/TiO2 nanotube	Rhodamine B	20 mL of 5 mg·L−1 RhB solution, 2 cm × 2 cm catalyst film. 300 W Xe lamp with UV cut-off filter.	67	47.85	1.4	[271]
24	WO3@g-C3N4	Rhodamine B	50 mL of 10 mg L−1 RhB solution, 10 mg catalyst. Xe lamp with 400 nm cut-off filter, 100 mW cm−2.	90	25.7	3.5	[272]
25	Mesoporous graphitic carbon nitride modified PbBiO2Br	Rhodamine B	100 mL of 10 mg·L−1 RhB solution, 30 mg catalyst. 300 W Xe lamp with UV cut-off filter (>400 nm).	98	N/A	N/A	[273]
26	g-C3N4/CuS p-n heterojunctions	Rhodamine B	30 mL of 10 mg L−1 RhB solution, 10 mg catalyst. 300 W Xe lamp with 420 nm cut-off filter.	93	27	3.5	[274]
27	g-C3N4/kaolinite composites	Rhodamine B	100 mL of 10 ppm RhB solution, 200 mg catalyst. 500 W Xenon lamp with 400 nm cut-off filter.	90	21.8	4.1	[275]
28	Hexagonal boron nitride (h-BN) decorated g-C3N4	Rhodamine B	100 mL of 20 mg L−1 RhB solution, 50 mg catalyst. 300 W Xe lamp with 420 nm cut-off filter.	99.5	13.63	7.3	[276]
29	ZnO/g-C3N4	Rhodamine B	50 mL of 10 mg L−1 RhB solution, 50 mg catalyst. 500 W Xe lamp equipped with 420 nm cut-off filter.	51.3	24.43	2.1	[277]
30	Ag/AgO loaded g-C3N4 microspheres	Acid Violet-7 (AV-7)	100 mL of 20 mg·L−1 AV-7 solution, 100 mg catalyst. 12 × 100 W fluorescent lamps (mainly visible light, with only 3% UV).	98	48	2	[278]
31	g-C3N4/TiO2/kaolinite composite	Ciprofloxacin (CIP) antibiotic	100 mL of 10 ppm CIP solution, 200 mg catalyst. Xe lamp (90 mW/cm2) with 400 nm cut-off filter.	92	14.48	6.4	[279]
32	Z-scheme CdS/Fe3O4/g-C3N4	Ciprofloxacin	100 mL of 20 mg L−1 CIP, 50 mg photocatalyst. 300 W Xe lamp with UV filter (λ > 420 nm).	92	3.53	26	[280]
33	Carbon-Doped g-C3N4	Tetracycline (TC)	80 mL of 10−4 M TC, 40 mg catalyst. Sunlight (07/10/2015, Trivandrum, India, between 11 pm and 1 pm, 78,000–80,000 lux).	95	50	1.9	[231]
34	Phosphorous-doped ultrathin graphitic carbon nitride nanosheets	Tetracycline	100 mL of 10 mg·L−1 TC solution, 100 mg catalyst. 300 W Xe lamp equipped with UV cut-off filter (>420 nm).	96.95	71.78	1.35	[281]
35	Hierarchical WO3/g-C3N4	Tetracycline hydrochloride (TC-HCl)	100 mL of 25 mg·L−1 TC-HCL solution, 50 mg catalyst. 300 W Xe lamp with 420 nm cut-off filter.	82	48	1.7	[282]
36	Co3O4 modified g-C3N4	Diclofenac sodium (DCF)	100 mL of 10 mg·L−1 DCF solution, 50 mg catalyst. 300 W Xe lamp with 420 nm cut-off filter.	100	17	5.9	[283]
37	silver and carbon quantum dots co-loaded with ultrathin g-C3N4	Naproxen NPX	50 mL of 4 mg·L−1 NPX solution, 50 mg catalyst. 350 W Xe lamp with 420 nm and 290 nm light for visible and simulated sunlight sources.	87.5	8.75	10	[284]
38	g-C3N4	Decabromodiphenyl ether (BDE209)	20 mL of 1 × 10−3 mol/L BDE209 solution, 20 mg catalyst. 300 W Xe lamp for UV-visible irradiation (>360 nm).	65	N/A	N/A	[285]
39	Metal-free sulfur doped g-C3N4	UO22+ removal	200 mL of 0.12 mM UO22+ solution, 100 mg catalyst. 350 W Xe lamp with a 420 nm cut-off filter.	95	71	1.3	[286]

6. Conclusions

g-C₃N₄ nanostructures offer tunable textural, electronic and optical properties that are amenable to tailoring for solar energy harvesting and subsequent photocatalytic transformations for energy and environmental applications. Diverse synthetic methods are available to prepare pure g-C₃N₄ nanostructures of different dimensionality and porosity, and to integrate these within multi-functional nanocomposites with enhanced solar spectral utilization, apparent quantum yields, charge separation and transport, and ultimately photocatalytic activity and stability. The sustainable production of H₂ as an energy vector from water splitting is perhaps the most promising application, although issues remain regarding the use of sacrificial reagents and a lack of interdisciplinary efforts to improve photoreactor design. Photocatalytic reduction of CO₂ is at a more preliminary stage, with improvements in both activity, and the ability to select specific products for either energy (e.g., CO, CH₄, methanol, and formic acid) or chemicals (e.g., >C₂ olefins or alkanes) pre-requisites to bench scale demonstrations. Wastewater treatment using g-C₃N₄-based photocatalysts appears promising; however, a lack of standardization in either reactor design or experimental protocols hampers quantitative comparisons due to issues such as decoupling adsorption versus reaction, and photocatalysis from direct photochemical activation of chromophores.