Recent Advances of Graphitic Carbon Nitride-Based Structures and Applications in Catalyst, Sensing, Imaging, and LEDs

Abstract

The graphitic carbon nitride (g-C₃N₄) which is a two-dimensional conjugated polymer has drawn broad interdisciplinary attention as a low-cost, metal-free, and visible-light-responsive photocatalyst in the area of environmental remediation. The g-C₃N₄-based materials have excellent electronic band structures, electron-rich properties, basic surface functionalities, high physicochemical stabilities and are “earth-abundant.” This review summarizes the latest progress related to the design and construction of g-C₃N₄-based materials and their applications including catalysis, sensing, imaging, and white-light-emitting diodes. An outlook on possible further developments in g-C₃N₄-based research for emerging properties and applications is also included.

Highlights

• The g-C3N4-based materials have excellent electronic band structures, electron-rich properties, basic surface functionalities, high physicochemical stabilities and are “earth-abundant.”

• Recent progress of g-C3N4-based nanostructures in catalyst, sensing, imaging and LEDs have been discussed in details.

• An outlook on possible further developments in g-C3N4-based research for emerging properties and applications is also included.

Introduction

Graphitic carbon nitride (g-C₃N₄), one of the oldest reported polymers in the literature, has a general formula of (C₃N₃H)_n. The history of development could be traced back to 1834 [1]. Research work has been inspired in the 1990s due to a theoretical prediction that diamond-like β-C₃N₄ could have extremely high hardness values [2]. At ambient conditions, g-C₃N₄ is regarded as the most stable allotrope. Similar to graphite, g-C₃N₄ is a layered material in which van der Waals force holds the stacking layers (covalent C–N bonds) and each layer is composed of tri-s-triazine units connected with planar amino groups [3]. The tri-s-triazine ring structure provides the polymer a high thermal stability (600 °C in air) and chemical stability in both acidic and alkaline environments [4].

Utilization of g-C₃N₄ in the heterogeneous catalysis arena started around a decade ago [5, 6]. The discovery of g-C₃N₄ polymer as a metal-free conjugated semiconductor photocatalysis for water splitting was first reported by Wang et al. [7] due to its appealing electronic structure, i.e., having a modulated bandgap and being an indirect semiconductor. Since then, these unique properties of g-C₃N₄ make it a promising candidate for visible-light photocatalytic applications utilizing solar energy. Solar energy is attracting worldwide attention by providing about 120,000 TW annually to the earth as one of the green, clean, and sustainable energy resources. Solar-induced chemical processes would be able to greatly extend the applications of g-C₃N₄. Since the landmark discovery of photocatalytic water splitting using TiO₂ electrodes by Fujishima in 1972, photocatalytic technology has been regarded as one of the most important strategies to address global energy and environmental issues [8]. Since then, there have been numerous developments in the fabrication of highly efficient semiconductor-based photocatalysts such as metal-based oxides and sulfides [9–12].

Notably, g-C₃N₄ has become a new family of next generation, non-toxic, metal-free, earth-abundant, and visible-light-driven polymeric semiconductor for applications in the degradation of organic pollutants, hydrogen evolution from water, sensing, imaging, and energy conversion [5, 9, 13–144]. Many reviews can be found mainly focusing on synthesis and catalytic applications of g-C₃N₄ [39–45, 145, 146]. However, a systematic description of the catalyst (photo and organic), bio-imaging, (chemical and bio-) sensing, devices, and energy-related applications (batteries, supercapacitors, white-light-emitting diodes, and oxygen reduction reaction) of g-C₃N₄ has not been presented until now. In this review, we give an overview of the porosity, luminescence, conductivity, and catalytic properties of g-C₃N₄, as well as their bio-imaging, photodynamic therapy, chemical sensing, and white-light-emitting diode applications. We believe this is the dawn for the development of g-C₃N₄. There are still new physical properties yet to be discovered based on g-C₃N₄ nanostructures. We are at a critical time to highlight the progress and provide a good source of references for this booming research topic.

g-C₃N₄-Based Structures

First-principle calculations predicted seven phases of g-C₃N₄, namely α-C₃N₄ (bandgap of 5.5 eV), β-C₃N₄ (bandgap of 4.85 eV), cubic C₃N₄ (bandgap of 4.3 eV), pseudocubic C₃N₄ (bandgap of 4.13 eV), g-h-triazine (bandgap of 2.97 eV), g-o-triazine (bandgap of 0.93 eV), and g-h-heptazine (bandgap of 2.88 eV) [4]. Figure 1 shows the primary tectonic units, triazine and tri-s-triazine ring structures, for forming the allotropes of g-C₃N₄. The structure can be viewed as graphite whose carbon lattice is partially substituted with nitrogen atoms in a regular fashion. g-C₃N₄ is an n-type semiconductor [5–7], which intrinsically possesses a very high nitrogen content dominated by a pyridinic and graphitic nitrogen while supported by a two-dimensional (2D) graphene sheet or three-dimensional (3D) porous graphitic carbon. The structure of g-C₃N₄ can be controlled by a variety of synthetic routes, including different condensation temperature, different ratio of precursors, porosity induced by hard/soft templating strategies, and exfoliation and doping. Synthetic routes and various morphologies including bulk, mesoporous, 3D, 2D, one-dimensional (1D), and zero-dimensional (0D) g-C₃N₄ will be discussed in the following.

Fig. 1.

a Triazine and b tri-s-triazine (heptazine) structures of g-C₃N₄. Reprinted with permission from Ref. [40]. Copyright 2016 American Chemical Society

Synthetic Routes of g-C₃N₄

g-C₃N₄ can be synthesized by thermal polymerization of abundant nitrogen-rich and oxygen-free compound precursors containing pre-bonded C–N core structures (triazine and heptazine derivatives) such as urea [46], melamine [47–49], dicyandiamide [50–54], cyanamide [14, 35, 55, 56], thiourea [57, 58], guanidinium chloride [59–61, 125], guanidine thiocyanate [126], and thiourea oxide [127]. The condensation pathways from above C/N precursors are facile and efficient routes to form the polymeric g-C₃N₄ network [39]. Many reports discussed that different types of precursors and treatments can strongly influence the physicochemical properties of the resulting g-C₃N₄, including surface area, porosity, absorption, photoluminescence, C/N ratio, and nanostructures.

Various surface modifications and functionalities have been employed to obtain desired structures such as 3D bulks, 2D nanosheets, 2D films, 1D nanorods, 1D nanotubes, 1D nanowires, and 0D quantum dots. For example, the urea precursor can be transformed to g-C₃N₄ at ca. 550 °C, as confirmed by X-ray diffraction (XRD, Fig. 2a, b). The C₃N₄ powders are usually yellow under the visible light. The optical properties are shown in Fig. 2c. The polymeric g-C₃N₄ is unstable at above 600 °C. Beyond 700 °C, g-C₃N₄ produces nitrogen and cyano fragments.

Fig. 2.

Crystal structure and optical properties of graphitic carbon nitride. a Schematic diagram of a perfect graphitic carbon nitride sheet constructed from melem units. b Experimental XRD pattern of the polymeric carbon nitride, revealing a graphitic structure with an interplanar stacking distance of aromatic units of 0.326 nm. c UV-visible diffuse reflectance spectrum and image (inset) of g-C₃N₄. Reprinted with permission from Ref. [5]. Copyright 2009 Nature Publishing Group. Representation of the β-C₃N₄ (d), α-C₃N₄ (e), graphite-C₃N₄ (f), pseudocubic-C₃N₄ (g), and cubic-C₃N₄ (h). The carbon and nitrogen atoms are depicted as gray and blue spheres, respectively, from Ref. [4]. Copyright 1996 American Association for the Advancement of Science. (Color figure online)

General analytical techniques for confirming the presence of g-C₃N₄ include X-ray photoelectron spectroscopy (XPS), XRD, and Fourier transform infrared (FTIR) spectroscopy [40]. The density function theory (DFT) calculations are used to reveal the characteristics of the valence and conduction bands. For example, g-C₃N₄ is found to mainly composed of the nitrogen p _Z orbitals and carbon p _Z orbitals, respectively [5].

Porosity of g-C₃N₄

High surface area and continuous porosity (as active centers) are important requirements for catalysis, gas, and energy storage technologies. Introduction of porosity in g-C₃N₄ can significantly increase their exposed surface area and accessible channels, and active edges, thus promoting the charge separation, molecular mass transfer, light harvesting, and surface reactions [145]. All these advantageous features can benefit the enhancement of photocatalytic efficiency. It is well known that the porous carbons have outstanding properties with respect to their use in energy applications, including as electrode materials for supercapacitors and as materials for solid-state hydrogen and carbon dioxide storage. The attractive attributions of porous carbons include low cost, environmental friendliness, chemical and thermal stability, easiness of processing, and low framework density. The activated carbons have been traditionally employed as absorbents or catalyst supports [62]. Compared with activated carbons, g-C₃N₄ has a moderate nitrogen content and ideal stoichiometry. The nitrogen content induces their unique surface properties such as semiconducting character, mechanical stability, thermal and chemical stability, which are superior to all other carbon nanomaterials (Table 1).

Table 1.

Recent reports on g-C₃N₄-based photocatalysts

Catalysts composition	Precursors of g-C3N4	Photocatalyst applications	Ref. (year)
g-C3N4	Cyanamide	Hydrogen production	[5] (2009)
g-C3N4/Graphene/NiFe	Urea	Photoelectrochemical	[13] (2016)
g-C3N4 nanocapsules	Cyanamide	Hydrogen production	[14] (2017)
g-C3N4/Co–N	Urea	Hydrogen production	[16] (2016)
g-C3N4/Graphene	Dicyandiamide	Hydrogen production	[18] (2014)
g-C3N4/PDA	Melamine	Hydrogen production	[20] (2015)
Alkalinized-C3N4/Fe	Melamine	Photo degradation	[21] (2016)
g-C3N4/Fe3O4	Melamine	Photo degradation	[22] (2013)
g-C3N4 Film	Melamine	Photoelectrochemical	[24] (2015)
g-C3N4/AgBr	Melamine	Photo degradation	[25] (2015)
g-C3N4 nanofibers	Melamine	Photo degradation	[30] (2013)
g-C3N4/PNA	Melamine	Photo degradation	[31] (2013)
g-C3N4 Film	Cyanamide	Photoelectrochemical	[32] (2015)
P-doped g-C3N4	Melamine	Hydrogen production	[33] (2015)
Amorphous g-C3N4	Dicyandiamide	Hydrogen production	[34] (2015)
g-C3N4	Cyanamide	Hydrogen Peroxide production	[35] (2014)
g-C3N4/Ag/TiO2	Melamine	Photo degradation	[36] (2014)
g-C3N4/Bi	Urea	NO Purification	[37] (2015)
g-C3N4/TiO2	Melamine	Photoelectrochemical	[47] (2016)
g-C3N4/ZIF	Melamine	CO2 Reduction	[48] (2015)
N-doped g-C3N4	Melamine	Hydrogen production	[49] (2015)
Iodine modified g-C3N4	Dicyandiamide	Hydrogen production	[50] (2014)
Holey g-C3N4	Dicyandiamide	Hydrogen production	[51] (2015)
Phosphorylation g-C3N4	Dicyandiamide	Hydrogen production	[53] (2015)
Porous g-C3N4	Dicyandiamide	Photo degradation	[54] (2015)
g-C3N4	Cyanamide	NO decomposition	[55] (2010)
Mesoporous g-C3N4	Cyanamide	Hydrogen peroxide production	[56] (2015)
Porous g-C3N4	Thiourea	Photo degradation	[57] (2016)
S-doped g-C3N4	Thiourea and Melamine	CO2 reduction	[58] (2015)
g-C3N4/bismuth-based oxide	Melamine or guanidine hydrochloride	Photo degradation	[61] (2016)
g-C3N4/Graphene	Urea	Hydrocarbon oxidation	[66] (2016)
3D porous g-C3N4	Melamine	Photo degradation	[67] (2016)
g-C3N4 nanoplatelets	Melamine	Water splitting	[73] (2015)
Graphene-like g-C3N4 nanosheets	Dicyandiamide	Hydrogen production	[75] (2012)
Crystalline g-C3N4	Dicyandiamide	Hydrogen production	[76] (2014)
Sulfur-mediated g-C3N4	Trithiocyanuric acid	Water oxidation	[77] (2011)
Nanotube g-C3N4	Melamine	Photo degradation	[79] (2014)
Helical g-C3N4	Cyanamide	Hydrogen production	[80] (2014)
Nanorod g-C3N4	Cyanamide	Hydrogen production and photoenzymatic catalysis	[81] (2014)
Mesoporous g-C3N4 nanorods	Cyanamide	Hydrogen production and reduction of nitrophenol	[82] (2012)
PAN/g-C3N4	Melamine	Hydrogen production	[83] (2016)
g-C3N4/ZIF	Urea	Photo degradation	[84] (2017)
g-C3N4	Dicyandiamide	Photo degradation	[91] (2014)
g-C3N4/Pd	Cyanamide	Organic catalyst	[92] (2015)
Oxidized g-C3N4	Melamine	Organic synthesis	[95] (2016)
g-C3N4/GO	Melamine	Photo degradation	[98] (2014)

Cavities can be introduced to form mesoporous g-C₃N₄ frameworks [29]. Qiao’s group reported the preparation of flexible films by integrating 2D mesoporous g-C₃N₄ (SiO₂ template method) with graphene sheets. A commonly used mesoporous silica, SBA15 was employed as a template to fabricate mesoporous g-C₃N₄ [6, 56]. However, hydrogen fluoride or other acid that are usually used to remove the host matrices of SiO₂ is hazardous. Besides hard template methods [128–130], soft templates [131, 132] and bubble templates [133–136] are also utilized to fabricate porous g-C₃N₄. Liu et al. reported the preparation of porous g-C₃N₄ by a simple co-pyrolyzation of co-precursors melamine and NH₄HCO₃. NH₄HCO₃ not only enhanced the specific area by bubbles, but it also shifted the conduction band and promoted the separation of charge carriers [63]. Calcium salts have also been utilized as a template for the synthesis of porous g-C₃N₄ with enhanced surface properties [143]. Metal–organic framework (MOF) is a typical high surface area materials and has also been utilized as templates for porous g-C₃N₄. Pandiaraj et al. [64] have reported MOF-derived g-C₃N₄ porous nanostructures.

Bulky g-C₃N₄

Bulky g-C₃N₄ can be synthesized by thermal condensation of a variety of precursors such as cyanamide, dicyandiamide, melamine, thiourea or urea between 400 and 600 °C. For example, g-C₃N₄ was synthesized from cyanamide into a combination of addition and polycondensation, in which case the cyanamide molecules were condensed to dicyandiamide and melamine at 203 and 234 °C, respectively. Next, the condensed dicyandiamide was removed. Essentially, all melamine-based products were found when the temperature was about 335 °C. Further heating to about 390 °C resulted in the rearrangement of tri-mesotriazine units via melamine. Finally, the polymer g-C₃N₄ occurred at about 520 °C by further condensation of the unit, which is thermally unstable at temperatures above 600 °C. During the calcination, the color changed from white (precursor) to light yellow (500 °C) and then dark orange (650 °C) [5, 40]. However, g-C₃N₄ obtained by such methods usually possesses a relatively low surface area (10 m² g⁻¹) and poor water solubility. Furthermore, the bulk g-C₃N₄ does not exhibit any photoluminescence characteristics when dissolved in solvents.

It has recently been found that urea is an excellent precursor for the synthesis of flaky g-C₃N₄ having a high specific surface area and a high porosity. g-C₃N₄ was synthesized at various calcination temperatures between 450 and 600 °C in a muffle furnace for 2 h at a heating rate of 15 °C min⁻¹ from oxygen-containing urea. During the thermal treatment process, the generated gases such as NH₃ at a low temperature (<200 °C) and CO₂ at a high temperature play a leading role in processing highly porous g-C₃N₄. The advantage of this method includes simplicity, convenience, and the absence of introduction of impurities during the synthesis of nanostructures. Compared to urea-derived g-C₃N₄, comprising of wrinkled porous 2D nanosheets, both thiourea-derived and dicyandiamide-derived g-C₃N₄ samples showed large sheets without porous structure. The specific surface area and crystallinity of g-C₃N₄ were marginally improved with increasing calcination temperatures. Generally, the heating rate is slower; the porosity of g-C₃N₄ could be improved [40].

Exfoliation methods such as sonication have been employed as a typical top-down route to obtain ultrathin g-C₃N₄ with excellent photoluminescence properties [65]. Bulk materials can be swelled and then exfoliated in the pure water. The dispersed ultrathin g-C₃N₄ nanosheets were negatively charged (zetapotential of about ~30.3 mV).

Three-Dimensional g-C₃N₄-Based Micro/Nanostructures

Three-dimensional architectures fabricated using nano-scale building blocks (0D, 1D, and 2D) are hot topics due to the desirable combination of high internal reactive surface area and straightforward molecular transport. However, the fabrication of 3D porous g-C₃N₄ has been a big challenge up to now. Recently, Liu’s group developed an efficient chemical vapor deposition growth strategy for 3D g-C₃N₄/graphene nanocomposites [66]. They found that g-C₃N₄ can be grown along the surface of graphene (see Fig. 3a, b, c). The C–C bonds (284.7 eV) and N–C=N bonds (288.3 eV) can be detected from the high-resolution C 1 s spectrum (Fig. 3d). Furthermore, sp² aromatic C=N–C (398.8 eV), tertiary N-(C)₃ (400.1 eV), and amino group C–N–H (401.3 eV) can be fitted in the high-resolution N 1 s spectrum (Fig. 3e). UV–vis spectroscopy (Fig. 3f) reveals the optical properties of the 3D composites to be similar to those of the pure g-C₃N₄.

Fig. 3.

a Schematic 3D g-C₃N₄/Graphene structures. b, c SEM images of g-C₃N₄/graphene nanocomposites. Inset TEM image. d C 1s, e N 1s XPS and f UV–vis spectra of g-C₃N₄/graphene nanocomposites. Insets: photographs of the powder samples. Reprinted with permission from Ref. [66]. Copyright 2016 American Chemical Society

Yuan et al. [67] reported the 3D porous g-C₃N₄ network assembled by exfoliated ultrathin nanosheets interconnected in large quantity via H₂SO₄ intercalation and subsequent thermal treatment.

Two-Dimensional g-C₃N₄ Nanostructures

Two-dimensional materials have received tremendous attention in the past decade because their ultimate structure was reported by Geim [69]. The topic about graphene has been cited over 30,000 times up to now (Google Scholar), and graphene has been employed widely in energy applications (such as lithium-ion batteries and supercapacitors [70–72]). Similar to graphene, g-C₃N₄ also has a typical sp² network (graphite-like layer structure) with weak van der Waals interactions across the layers. Inspired by the successful exfoliation of graphene from bulk graphite, Xie et al. [65] firstly demonstrate that ultrathin g-C₃N₄ nanosheets could be prepared by a green liquid exfoliation from bulk g-C₃N₄ in water. From bulk to ultrathin nanosheets (several layers), g-C₃N₄ nanosheets show an obvious increase in density of states (DOS) at the conduction band edge with respect to the bulk counterpart by first-principle density-functional calculations. Global effects have been launched to synthesize single or few-layer g-C₃N₄ nanosheets due to their attractive physicochemical properties [52, 73, 74]. Figure 4 shows characterization of the g-C₃N₄ nanosheets studied by transmission electron microscopy (TEM) and atomic force microscopy (AFM), illustrating the similarity of ultrathin layers of exfoliated g-C₃N₄ to that of graphene.

Fig. 4.

a-c TEM images and d AFM image of g-C₃N₄ nanosheets. Inset of c SAED pattern. Reprinted with permission from Ref. [52]. Copyright 2014 John Wiley and Sons. e Schematic illustration of top-down and bottom-up synthetic strategies for g-C₃N₄ nanosheets. Reprinted with permission from Ref. [68]. Copyright 2015 The Royal Society of Chemistry. (Color figure online)

There are two other synthetic approaches of 2D g-C₃N₄ nanosheets: one is known as the top-down approach, and the other one is the bottom-up approach (Fig. 4e). For the top-down approach, chemical etching and ultrasonication-assisted liquid exfoliation are the two main technologies involved [75, 76]. Template-assisted method and heteroatom-mediated method have been commonly employed for the bottom-up approaches [77, 78]. The big atomic size of sulfur can influence the conformation and the connectivity of the resultant g-C₃N₄ and hence offer a template tool to tune the texture and electronic structure. The first report of template-mediate synthesis of 2D g-C₃N₄ nanosheets involved the use of GO-derived silica [78]. Zhang et al. [23] reported the sol processing for the fabrication of g-C₃N₄ thin films with HNO₃ as a strong oxidizing acid. Liu et al. [54] developed a method to grow g-C₃N₄ thin films directly on conductive substrates.

One-Dimensional g-C₃N₄ Nanostructure (Nanorods, Nanotube, and Nanofibers)

One-dimensional g-C₃N₄ nanostructures hold good promise for electronic and electrochemical performances due to their high surface area, and light harvesting and mass transfer properties. Wang et al. described a facile method to fabricate g-C₃N₄ nanotubes (Fig. 5a) by directly heating melamine without any templates. The resulting nanotubes exhibited a blue fluorescence and excellent photocatalytic properties [79]. Hollow 1D g-C₃N₄ nanostructures have been developed using a sulfur-mediated self-templating method by Liu’s group [83], as shown in Fig. 5b.

Fig. 5.

1D g-C₃N₄ nanostructures. a SEM image of nanotubes [79]. b TEM image of nanotubes [83]. c TEM image of nanorods [81]. d TEM image of porous nanorods [82]. Inset: pore size distribution. e SEM image of nanofibers [30]. f SEM image of tubular structures [85]. Reprinted with permission from Ref. [79] (Copyright 2014 The Royal Society of Chemistry), with permission from Ref. [83] (Copyright 2016 American Chemical Society), with permission from Ref. [81] (Copyright 2014 American Chemical Society), with permission from Ref. [82] (Copyright 2012 The Royal Society of Chemistry), with permission from Ref. [30] (Copyright 2013 American Chemical Society), with permission from Ref. [85] (Copyright 2016 Wiley–VCH Verlag GmbH), respectively

Liu et al. [81] also reported the synthesis of g-C₃N₄ nanorods by using chiral silica nanorods as templates for practical enzymatic applications. In addition, Li et al. [82] described a one-step method to fabricate mesoporous g-C₃N₄ nanorods with the template SBA-15. Tahir et al. [30] reported the synthesis of g-C₃N₄ nanofibers for energy storage and Photo degradation applications. Recently, Tong et al. combined the hydrothermal and condensation techniques to obtain a tubular g-C₃N₄ isotype heterojunction with excellent photocatalytic property [85]. Figure 5c–f depicts these tubular nanostructures.

Zheng et al. [80] developed a nanocasting technique to fabricate twisted hexagonal rod-like C₃N₄ by using chiral silicon dioxide as templates. The helix is an important template in nature, as one finds it in DNA, RNA, and proteins. Synthesis of chiral inorganic nanostructures has recently gained considerable attention. The chiral nanostructures shown in Fig. 6a–f reveal the helical morphology and ordered channels winding around the rod centers. This is the first report about both left- and right-handed chiral nanostructures with unique optical activities.

Fig. 6.

Morphology characterization of the HR-CN sample. a SEM, b TEM, and c, d corresponding elemental mapping images of g-C₃N₄. Reproduced from Ref. [80] by permission of the John Wiley & Sons Ltd. e A g-C₃N₄ layer and f A single g-C₃N₄ nanotube formed by rolling the g-C₃N₄ layer. Reproduced from Ref. [79] by permission of the Royal Society of Chemistry

Zero-Dimensional g-C₃N₄ Nanostructure

When the size of the nanostructure is less than 10 nm, g-C₃N₄ nanostructures (typically contain a few thousand atoms) usually show significant quantum confinement effects and possess excellent properties like bright fluorescence, water solubility, and above all, non-toxicity. This is in contrast to the fact that most essential elements in semiconductor quantum dots (QDs) (for example, Se in CdSe, Pb in PbTe, and Te in CdTe) all have risks of long-term toxicity and potential environmental hazards. Similar to the synthetic routes of 2D g-C₃N₄ nanosheets, g-C₃N₄ QDs have been mainly synthesized by top-down and bottom-up approaches.

g-C₃N₄ QDs were prepared first by the top-down approach. Wang et al. was the first to prepare g-C₃N₄ QDs by a thermal-chemical etching process from bulk g-C₃N₄, as illustrated in Fig. 7e [86]. Xie et al. demonstrated an exfoliation strategy for the preparation of single-layered QDs. When these QDs passed through the cell membranes, they exhibited an excellent two-photon absorption behavior as compared with the double-layered QDs [87] (Fig. 7a–d). Recently, Wu et al. developed a doping method (phosphorus as dopant) to control the emission wavelength of g-C₃N₄ QDs to be in the whole visible-light regime [88]. This was the first report that phosphorus doping could change the direct bandgap of g-C₃N₄ QDs.

Fig. 7.

a HOMO-n and b LUMO + n orbitals of the single-layered g-C₃N₄, respectively. c HOMO-n and d LUMO + n orbitals of the double-layered g-C₃N₄, respectively. Reproduced from Ref. [87] by permission of John Wiley & Sons Ltd. e Schematic illustration of the controllable synthesis of g-C₃N₄ nanosheets, nanoribbons, and quantum dots. Reproduced from Ref. [86] by permission of the Royal Society of Chemistry

Hydrothermal and microwave heating methods are commonly utilized for the synthesis of g-C₃N₄ QDs as the bottom-up approach [89, 90]. Lu et al. [90] reported the synthesis of g-C₃N₄ QDs with strong blue photoluminescence by hydrothermal heating of the citric acid and thiourea. Cao et al. [89] developed a facile microwave-assisted fabrication of g-C₃N₄ QDs. Compared with the majority of carbon materials such as graphene QDs and carbon QDs, g-C₃N₄ QDs which possess both nitrogen-rich and electron-rich properties and basic surface functionalities represent a new family of luminescent QDs.

Multifunctional Applications

As the most stable allotrope of carbon nitride, g-C₃N₄ is versatile materials with unique semiconducting, excellent photocatalytic properties. They are also environmental friendly, low cost, and metal-free, which make them attractive for a range of applications beyond catalysis, including sensing, bio-imaging, photodynamic therapy, and energy conversion.

g-C₃N₄ Catalysts

Polymeric g-C₃N₄ semiconductors are widely used as catalysts due to their excellent chemical stability and unique electronic band structure. In the formation of the g-C₃N₄ network, the C-p _z orbit composes the lowest unoccupied molecular orbital (LUMO), and N-p _z orbital composes the highest occupied molecular orbital (HOMO), with a 2.7 eV bandgap between these two orbitals [91]. This suitable bandgap can absorb the solar electromagnetic energy with wavelength less than 475 nm. It was found that pyrrolic nitrogen has the strongest role in acetylene hydrochlorination among all nitrogen species [17].

The rapid recombination rate of electron–hole pairs results in low efficiency, thus limiting the practical applications of g-C₃N₄. Sun et al. [92] developed a homogeneous catalyst, Pd/g-C₃N₄, for a Suzuki–Miyaura coupling reaction with superior catalytic activity under mild conditions. It is well known that the Suzuki–Miyaura coupling reaction is of primary importance for the construction of C–C bonds. The uniform Pt nanoparticles deposited on the surface of g-C₃N₄ networks can result in a high yield of 97% biphenyl and 100% bromobenzene. Kumar et al. reported a nanocomposite of an iron(II) bipyridine with carbon nitride as a photocatalyst for the oxidative coupling of benzylamines under mild reaction conditions, resulting in excellent activity and effective recycling ability [93]. A photoactive catalyst Ru/g-C₃N₄ was developed by Sharma et al. [94] for efficient photocatalytic transfer hydrogenation of aldehydes and ketones under mild conditions. Xie’s group explored the generation of singlet oxygen in oxidized g-C₃N₄ [95].

g-C₃N₄ can be used as a new kind of metal-free photocatalysts. Wang et al. [5] was among the first to use g-C₃N₄ as a photocatalyst for hydrogen production from water. However, the quantum efficiency of the catalyst is only 0.1% with the irradiation of 420–460 nm due to its fast recombination of electron–hole pairs. To solve this problem, modified 2D g-C₃N₄ materials with a redshift absorption were produced and a quantum efficiency of 8.8% was achieved at 420 nm by the same group [96]. Liu et al. [97] developed a carbon dots/g-C₃N₄ nanocomposite as a metal-free photocatalyst with high quantum yield and excellent stability. The overall evolution of H₂ and O₂ is shown in Fig. 8a with a molar ratio of 2.02 (a value of 2 is identified for water splitting). Absorbance and quantum efficiency (QE) of the carbon dots/g-C₃N₄ nanocomposite were measured and are shown in Fig. 8b. The catalyst composition was optimized by measuring QE for different concentrations of carbon dots in a fixed mass of composite, as shown in Fig. 8c. With the increase in carbon dots, the QE can reach as high as 16% (Fig. 8d).

Fig. 8.

a Typical H₂ and O₂ production from water under visible-light irradiation. b Wavelength-dependent QE (red dots) of water splitting by composite catalyst. c QE for different concentrations of carbon dots/g-C₃N₄ catalysts in a fixed mass of composite catalyst. d QE for different catalyst loads with a constant carbon dot concentration in 150 ml of ultra-pure water. Reproduced with permission from Ref. [95]. Copyright 2015 American Association for the Advancement of Science. (Color figure online)

Recently, nanohybrids (van der Waals heterostructures) which compose of different 2D nanolayers exhibit much improved catalytic activities. For example, porous g-C₃N₄/graphene films have been fabricated as electrodes for efficient hydrogen evolution by Qiao et al. [29]. Dai et al. [98] obtained graphene oxide/g-C₃N₄ nanosheets by a sonication method with reinforced photocurrent. Ma et al. demonstrated the fabrication of hybrid g-C₃N₄/graphene quantum dot nanocomposites. The hybrid nanocomposites have an excellent efficiency for water splitting due to decreased bandgap. Experimental results and DFT calculations revealed that the chemical bonding of two different layered materials can improve the catalytic activity [18]. Xu et al. [137] fabricated AgBr/g-C₃N₄ hybrid materials with synergistic visible-light photocatalytic activity, and the uniform AgBr nanoparticles were well dispersed on the g-C₃N₄ nanosheets which enhanced the optical activity. What is more, many hybrid materials like g-C₃N₄/Ag₂CO₃ [138], g-C₃N₄/ZnWO₄ [139], g-C₃N₄/Ag₂O [140], and g-C₃N₄/Ag [141] were explored to enhance photocatalytic activity. Sun group obtained mesoporous carbon/g-C₃N₄ with remarkably enhanced photocatalytic activity due to enhanced visible light and dye adsorption [142]. Mesoporous g-C₃N₄ has been fabricated in the presence of SBA-15 with enhanced photocatalytic activity for methyl orange degradation by Gao et al. [144].

Crystalline carbon nitride nanosheets have been prepared by Lotsch et al. [76] to enhance visible-light hydrogen evolution. The authors stated that morphology and surface area are the two crucial factors governing the photocatalytic performances. An efficient deposition method of growing g-C₃N₄ on different electrodes has been developed by Shalom et al. [26]. The successful deposition technique enables the fabrication of many electronic devices based on g-C₃N₄. Wu et al. studied the effect of defects in g-C₃N₄ on hydrogen evolution and photovoltage. Controlling different types of defects is the key to improve the catalytic performance [27].

He et al. [20] utilized polydopamine/g-C₃N₄ composites to produce hydrogen from water with superior activity. DFT calculations were used to obtain the band structure of g-C₃N₄ with nonmetal element doping, including boron, oxygen, phosphorous, and others. Phosphorus was predicted to be suitable as a doping element in g-C₃N₄ because it can decrease the bandgap of g-C₃N₄ from 2.7 to 2.31 eV without any mid-gap states [28].

Shiraishi et al. [35] reported a g-C₃N₄ photocatalyst to reduce O₂ to H₂O₂ via a two-electron route under visible light. Fang et al. [49] demonstrated that the nitrogen self-doped g-C₃N₄ can significantly enhance the catalytic activity (1.8 times) of hydrogen evolution and modify the optical and electronic properties with respect to the un-doped g-C₃N₄. Sun’s group reported ultrathin g-C₃N₄ nanosheets/graphene nanocomposites as a highly efficient electrocatalyst for oxygen evolution reaction. They revealed that the high oxygen evolution reaction activities resulted from pyridinic-N-related active sites [121]. They also developed 3D porous supramolecular architecture based on g-C₃N₄ nanosheets/graphene oxide as a highly efficient electrocatalyst for oxygen reduction reaction [122].

Besides metal-free materials such as graphene or carbon dots, various metal oxides and sulfides have been coupled with g-C₃N₄ for enhancing photocatalytic performances. For example, Chen et al. demonstrated that the g-C₃N₄/Ag/TiO₂ heterostructure microspheres were successfully achieved with enhanced photocatalysis performances [36]. Gu et al. [100] obtained the g-C₃N₄/TiO₂ nanosheets with high reactive (001) facets by a hydrothermal method, accompanied by a remarkable enhancement of photocatalytic capability in degradation of organic molecules under visible and UV light. Wang et al. successfully fabricated g-C₃N₄/BiPO₄ g-C₃N₄ photocatalyst. The hybrid structure has been proved by Dong et al. to be a novel photocatalyst for the application of NO purification via an in situ deposition method [37]. Wang et al. explored the enhanced photocatalytic mechanism for the hybrid g-C₃N₄/MoS₂ nanocomposites by DFT calculations. DFT calculations show these hybrid catalytic nanocomposites indeed have a higher absorption (used in the treatment of methylene orange (MO) [101]), with a decent photocatalytic performance and better separation of photo-generated carriers [38]. Ye’s group has demonstrated a zirconium-based metal–organic framework (Uio-66)-g-C₃N₄ nanosheet compound via a facile self-assembly method. Photocatalytic CO₂ reduction activities were greatly enhanced. The photo-generated electron can be transferred from the g-C₃N₄ nanosheets to Uio-66 for better reduction of CO₂ [48].

Ye’s group was inspired by the natural photosynthesis process and explored an environmental “phosphorylation” strategy to improve the photocatalytic performance [48]. A hydrogen generation rate of 947 µmol h⁻¹ and a quantum yield of 26.1% at 425 nm were achieved. This is the highest record for g-C₃N₄-based photocatalyst [53]. Recently, Tong et al. were inspired by the function of thylakoids in a natural photosynthesis system as shown in Scheme 1a. They successfully fabricated a multishell g-C₃N₄ nanocapsule photocatalysts with enhanced light harvesting and electron transfer properties. The overall synthesis procedure is illustrated in Scheme 1b. The triple-shell g-C₃N₄ can produce hydrogen as much as 630 µmol h⁻¹. This success potentially produces a new generation of solar devices for hydrogen production.

Scheme 1.

a Natural photosystem with green leafs, and the enlarged figure (right) depicts the light conversion in the stacked thylakoids. b Schematic illustration for the preparation of MSCN nanocapsules. Adapted with permission from Ref. [14]. Copyright 2017 American Chemical Society

Molecular doping not only plays an indispensable role in regulating the bandgap and electron structure of g-C₃N₄, but also controls the physical and chemical properties of g-C₃N₄. The bottom-up copolymerization method allows a large selection of structurally matched organic anchoring groups to be integrated into the g-C₃N₄ tri-triazine backbone to design highly efficient photocatalysts with the desired chemical composition and bandgap. It is expected that the modified g-C₃N₄ nanocrystals will provide an insightful view of the sustainable use of solar energy in chemistry because of the interesting new features that provide a new avenue for the study of heterogeneous catalysis.

g-C₃N₄ has a moderate bandgap of 2.7 eV, which makes it active in the visible region. Thermodynamic losses and overpotentials are usually considered in the photocatalytic process, the bandgap of 2.7 eV lying in between 2 eV (water splitting with enough endothermic driving forces) and 3.1 eV (visible-light absorption). What is more, g-C₃N₄ has a suitable conduction band position for various reduction reactions. The CB of g-C₃N₄ is more negative than H₂-evolution, CO₂-reduction, and O₂-reduction reactions, revealing that the photo-generated electrons in g-C₃N₄ possess a large thermodynamic driving force to reduce small molecules like H₂O, CO₂, and O₂. Therefore, the suitable electronic band structures of g-C₃N₄ are favorable for its applications as catalyst, such as water splitting, CO₂ reduction, oxygen reduction reaction, oxygen evolution reaction, pollutant degradation, and organic synthesis.

Advantages and disadvantages Bandgap is crucial for photocatalytic applications. g-C₃N₄ has a direct moderate bandgap of 2.7 eV which has visible-light activity besides its advantages of low cost, metal-free, 2D layered structure, easy fabrication, and high stability. Unfortunately, the bulk g-C₃N₄ generally exhibits the low photocatalytic efficiency due to some serious drawbacks of g-C₃N₄ material, such as the high electron–hole recombination rate, low surface area, insufficient visible absorption, few active sites for interfacial reactions and low charge. Among various design strategies, heterojunction construction (especially for Z-scheme construction) and porosity design (microporous, mesoporous, and macroporous) have been employed. Based on present g-C₃N₄ materials, photocatalytic CO₂ reduction seems to be more promising in developing practical g-C₃N₄-based photocatalysts.

g-C₃N₄ Sensing

It is well known that the polymeric g-C₃N₄ exhibits photoluminescence (PL) properties similar to many semiconductors materials. g-C₃N₄ emits a blue PL around 450 nm when dissolved in solvents under UV light irradiation due to its direct bandgap of 2.7 eV, which can be explained as the transition of the s-triazine ring [39]. Luminescent g-C₃N₄ (nanosheets and nanodots) can simply be regarded as nitrogen-rich carbon dots, although the quantum yield of g-C₃N₄ (up to 40%) is lower than carbon dots (up to 90%) [102]. g-C₃N₄ nanostructures exhibit a higher stability than the carbon dots, therefore potentially providing more practical applications. However, we are still far from the success of improving the quantum yield and understanding the precise PL mechanism of g-C₃N₄.

Based on the unique PL property of g-C₃N₄, g-C₃N₄ nanosheets have a strong response to copper ions [74] as turn-off chemical sensors. Since the chemical reduction of Cu²⁺ to Cu⁺ lies between the conduction band and valence band of g-C₃N₄, the PL of g-C₃N₄ can be quenched with a low detection limit of 0.5 nM [103]. A similar mechanism has been explained in previous work [104]. In that work, a relatively low detection limit of 0.04 nM has been achieved by CdTe QDs. Besides copper ions, PL of g-C₃N₄ can also be quenched by other metal ions like Fe³⁺, Ag⁺, Hg²⁺, and Cr²⁺ [89, 90, 105–110]. Huang et al. reported the fabrication of g-C₃N₄ nanosheets for the selective detection to Cu²⁺ and Ag⁺. Zhang et al. successfully prepared the g-C₃N₄ QDs as effective fluorescent probes for the detection of Fe³⁺ and Cu²⁺. Cao et al. developed the g-C₃N₄ nanodots via a microwave-assisted approach. The produced nanodots were utilized as turn-off sensors for mercury ions with a detection limit of 0.14 μM [89]. Shao’s group successfully synthesized the oxygen- and sulfur-co-doped g-C₃N₄ nanodots via a hydrothermal method, which has a lower detection limit of mercury ions (0.37 nM) [90]. Sun’s group successfully fabricated Fe-doped g-C₃N₄ nanosheets with a highly sensitive optical detection of glucose due to the chelation of Fe³⁺ with N [123]. Moreover, they demonstrated the ultrathin g-C₃N₄ nanosheets can be utilized as electrochemical glucose biosensing with a detection limit of 11 μM [124]. Rong et al. explored turn-off–turn-on sensors using g-C₃N₄ nanosheets, Cr and ascorbic acid. After Cr quenches the PL signal of g-C₃N₄, the addition of ascorbic acid can recover the PL signal due to the oxidation–reduction between Cr and ascorbic acid. Yang’s group fabricated a g-C₃N₄ nanosheet/MnO₂ sandwich nanocomposite via a one-step approach. The fabricated composites could turn-on the fluorescent probes of glutathione (GSH) with a high selectivity due to fluorescence resonance energy transfer (FRET), as shown in Scheme 2a [111]. GSH is possibly a suitable recovering agent of the PL signal of g-C₃N₄ as well. Xu et al. reported a g-C₃N₄/Hg system without PL signal, which can be switched on by an introduction of GSH. The system may be employed for detection of GSH in various food samples [112]. Based on FRET, g-C₃N₄ was found to detect riboflavin (RF) because g-C₃N₄ acts as a donor of energy transfers and RF as an acceptor [113]. A turn-on g-C₃N₄-based long-persistent luminescent probe for detection of biothiols was reported by Tang et al. [115] for the first time. This long-persistent luminescence allows the detection without external illumination. Qiao’s group has successfully prepared the proton-functionalized ultrathin g-C₃N₄ nanosheets which can interact with heparin, therefore achieving the lowest detection limit of 18 ng mL⁻¹ [52].

Scheme 2.

a Schematic representation of g-C₃N₄/MnO₂ nanocomposite for sensing of GSH. Reproduced with permission from Ref. [107]. Copyright 2014 American Chemical Society. b Schematic illustration of the dual-wavelength ratiometric ECL-RET biosensor configuration strategy. Reproduced with permission from Ref. [114]. Copyright 2016 American Chemical Society

Feng et al. successfully fabricated Au-nanoparticle-functionalized g-C₃N₄ nanosheets coupled with Ru(bpy)²⁺₃. The coupled g-C₃N₄ nanosheets can be employed for RNA detection based on dual-wavelength electrochemiluminescence (ECL), as shown in Scheme 2b. Au/g-C₃N₄ composites exhibit an emission at 460 nm and g-C₃N₄/Ru(bpy)²⁺₃ at 620 nm. Here, g-C₃N₄ acts as a donor of energy transfer and Ru(bpy)²⁺₃ as an acceptor for highly sensitive and selective detection of target miRNA (ECL signals quenching at 460 nm and increasing at 620 nm) [114].

In addition to detecting metal ions and bio-molecules by g-C₃N₄, g-C₃N₄ can also respond to temperature [116]. Debanjan et al. reported a temperature sensor based on the PL of g-C₃N₄. They found that as the temperature increased, the intensity of PL decreased.

g-C₃N₄ Imaging

Non-toxicity, metal-free, high stability, and high PL quantum yield enable g-C₃N₄ nanosheets and nanodots to be promising candidates for cell imaging. Xie’s group demonstrated the preparation of ultrathin g-C₃N₄ nanosheets for bio-imaging applications [65]. They found that g-C₃N₄ nanosheets have no significant effects on the HeLa cell viability even at a high concentration. The same group further developed the single-layered g-C₃N₄ QDs for both one-photon and two-photon cell imaging as long as the QDs can pass through the nuclear pore and enter into the nuclei [87]. Singlet oxygen, being one of the most important reactive oxygen species (ROS), could be generated in the presence of g-C₃N₄ as a photosensitizer [95].

Recently, Yang’s group reported the NIR-driven g-C₃N₄/up-conversion nanoparticle (UCNP) composite for efficient bio-imaging and photodynamic therapy (PDT) [117]. In this application, g-C₃N₄ functions as a sensitizer to absorb the UV light converted by UCNPs from 980 nm NIR light. The generated ROS causes the tumor to shrink or disappear effectively without any side effects from the irradiation [117]. However, long-time irradiation of 980 nm NIR light can cause overheating of tissues; therefore, an 808 nm laser light is more suitable for the PDT.

Feng et al. [118] fabricated a novel core–shell structure (UCNP/g-C₃N₄) for phototherapy and imaging applications. Mesoporous g-C₃N₄ was coated outside of the shell of UCNPs, generating a large amount of ROS due to the large surface area of g-C₃N₄ under the 808 nm laser irradiation. In vivo experiments were conducted as shown in Fig. 9. The factors affecting the efficiency of sensing and imaging are mainly the quantum yield, functionalization, PL and optical properties, stability, and toxicity.

Fig. 9.

a In vivo tumor volume growth curves of mice in different groups after various treatments. b Body weight changes of Balb/c mice versus treated time under different conditions. c Photographs of excised tumors from representative Balb/c mice after 14 day treatment and d the corresponding digital photographs of mice in the control group and “UCNPs@g-C3N4 − PEG with 808 nm laser” group after 14 day treatment. e H&E stained tumor sections after 14 day treatment from different groups. Reproduced with permission from Ref. [118] Copyright 2016 American Chemical Society

g-C₃N₄-Based LED

Although the PL properties of g-C₃N₄ have been investigated in the past fifteen years, the solid-state lighting of g-C₃N₄-based materials is still at an infancy stage. Various investigations about the g-C₃N₄-based solid-state lighting such as white-light-emitting diodes (WLEDs) have been carried out. Wang et al. fabricated the g-C₃N₄/silica gels for WLEDs application (Fig. 10a–d). In this work, using a one-step heat treatment approach confirmed by the FTIR, g-C₃N₄ was found to be covalently bonded with silica gels (Fig. 10e, f). The g-C₃N₄/silica gels obtained possess emerging properties with respect to bare g-C₃N₄, including water-resistance, high transparency, high flexibility, and white light emission under UV irradiation. The mechanism for white light emission can be ascribed to surface passivation by silica [119]. Bayan et al. [120] prepared a g-C₃N₄ sheet/ZnO nanorod hybrid for WLEDs by combining the emissions of g-C₃N₄ and ZnO nanorods to achieve a broad emission. Last but not the least, Gan et al. studied the origins of broad PL of g-C₃N₄ for the WLEDs applications. They demonstrated that the broad PL from g-C₃N₄ is attributed to band-to-band transitions in the tri-s-triazine rings. This novel work can help the understanding of the PL mechanism and accelerate the progress of WLEDs-based g-C₃N₄.

Fig. 10.

a Photoluminescence spectrum of the g-C₃N₄/silica gels excited at 365 nm displaying four peaks (430, 480, 580, and 627 nm) in the visible regime. b, c Photographs of a free-standing g-C₃N₄/silica-gel membrane, displaying both good transparency b and flexibility c. d CIE-1931 chromaticity diagram showing the emission from the typical g-C₃N₄/silica gels (marked by the black cross) excited at 365 nm. e FTIR spectra of AEATMS, g-C₃N₄-silica gels, and g-C₃N₄ particles. f Schematic drawing of an AEATMS-capped g-C₃N₄ particles in the g-C₃N₄/silica gels. Reproduced with permission from Ref. [119]. Copyright 2016 Wiley

Conclusion and Outlook

This review summarizes the recent advances of g-C₃N₄-based structures and applications including catalyst, chemical and biosensing, imaging, and LEDs. The performances of g-C₃N₄ are mainly based on their surface state (defects, function groups, and doping) and structures (porosity, thickness, and morphology). Although a significant advancement has been made for the development of highly efficient g-C₃N₄-based photocatalysts, there are still considerable problems that require further investigations, including the catalytic rate and design.

2D polymeric g-C₃N₄ materials featuring low cost, metal-free, environmental friendly, moderate bandgap, high chemistry activity, and high stability have only been studied for the past few years (from fundamental research to practical applications). We believe that more emerging properties and applications of g-C₃N₄ are around the corner. Integrations between experimental research and theoretical approaches will advance the research progress of g-C₃N₄ to a large extent. As the exploration of g-C₃N₄ that is still in its infancy, there are several remaining key challenges that must be met in near future, including porous nanostructures for the drug loading and delivery, improvement of electronic conductivity, memory device fabrication, solid-state lighting, energy conversion, and wearable sensors.