Novel two-dimensional crystalline carbon nitrides beyond g-C₃N₄: structure and applications

Abstract

Two-dimensional (2D) crystalline carbon nitrides (C_xN_y) with graphene-like atomic structures but semiconducting nature are new appealing materials, and increasing interest has been focused on their synthesis, properties, and applications. Apart from the well-known graphitic carbon nitride (g-C₃N₄), other types of 2D C_xN_y nanocrystals including C₂N, C₃N and C₅N₂ are also successfully fabricated in recent years and their properties and potential uses in various fields have been widely investigated. Meanwhile, g-CN, C₂N₂, C₂N₃ and C₅N have been theoretically predicted, and their properties and potential applications have also been researched. Until now, 2D crystalline C_xN_y compounds have demonstrated promising applications in photocatalysis, electrocatalysis, gas adsorption/separation, and energy storage devices, from both experimental and theoretical aspects, stemming from their unique geometric structures, tunable electronic properties, and high specific surface area. Herein, recent advances in the synthesis of 2D C_xN_y nanocrystals and their applications in various fields are summarized. Moreover, future challenges and opportunities for developing 2D C_xN_y nanocrystals for more broad applications including environmental remediation are also discussed.

---

Li Tan

Miss Li Tan obtained her Master's degree in 2018 from Yanshan University (China) and she is now pursuing a PhD degree under the supervision of Prof. Zhimin Ao at the School of Environmental Science and Engineering, Institute of Environmental Health and Pollution Control, Guangdong University of Technology. Her current research interests focus on the synthesis of novel 2D carbon-nitride-based catalysts for the adsorption of volatile organic compounds and their catalytic remediation.

Chunyang Nie

Dr Chunyang Nie obtained her Master's degree in 2013 from East China Normal University (China) and PhD degree in 2016 from Universite Paul Sabatier (France) under the supervision of Dr Emmanuel Flahaut and Dr Marc Monthioux. She joined Prof. Ao's group at the Guangdong University of Technology in 2016 as a Postdoctoral Fellow. Her current research interests focus on the synthesis of novel carbon-based catalysts for aqueous organic pollutant remediation and catalytic mechanisms.

Zhimin Ao

Prof. Zhimin Ao received his PhD in Materials Science from Jilin University, China in 2008. He is currently a full professor at the Institute of Environmental Health and Pollution Control, School of Environmental Science and Engineering, Guangdong University of Technology. His research interests mainly focus on the development of environmental catalysts and corresponding mechanism investigation. He has published over 130 SCI papers with citations over 4800 times and an H index of 31. He serves as an Associate Editor of Frontiers in Nanotechnology.

Hongqi Sun

Prof. Hongqi Sun received his PhD degree in Chemical Engineering in 2008 from Nanjing University of Technology, China. He joined Edith Cowan University (ECU) in March 2016 and became a full Professor of Chemical Engineering at ECU in November 2017. His research focuses on novel catalysis (photocatalysis, photo-thermal catalysis, photo-electro-chemical catalysis, and membrane catalysis) and advanced oxidation processes (AOPs) over nanostructured materials for solar-to-chemical conversion and environmental remediation. He serves as the Associate Editor of RSC Advances, and Speciality Chief Editor of Frontiers of Nanotechnology, assessor of ARC, committee member of international conferences, and referee of international journals.

Taicheng An

Prof. Taicheng An received his PhD in Environmental Science from Sun Yat-sen University, China in 2002. He is currently a full professor and the director of the Institute of Environmental Health and Pollution Control, Guangdong University of Technology. His research interests mainly focus on: (1) synthesis, characterization, optimization and application of nanostructured catalysts and natural mineral materials in environmental remediation; (2) applications of advanced oxidation processes (AOPs); (3) the transportation mechanisms, fate prediction and health effect of emerging organic contaminants. He has been awarded over 70 patents and published over 380 refereed research papers. He serves as an Associate Editor of Appl. Catal. B:Environ. and Ctr. Rev. Env. Tech.

Shaobin Wang

Prof. Shaobin Wang obtained his PhD in Chemical Engineering from the University of Queensland, Australia. He is now a Professor at the School of Chemical Engineering and Advanced Materials, The University of Adelaide, Australia. His research interests focus on nanomaterial synthesis and application for adsorption and catalysis, fuel and energy conversion and environmental remediation. He has published more than 400 refereed journal papers with citations over 40 000 and an H-index of 113. He is a Global Highly Cited Researcher in Engineering in 2016–2019.

---

1. Introduction

Graphene, as a two-dimensional (2D) atomic crystal available to the scientific community, was successfully discovered by Novoselov in 2004.¹ An ideal graphene structure is composed of a single layer of sp²-hybridized carbon atoms.² Ever since, the applications of this new material have been developed rapidly in many fields such as electronics, photonics, energy storage and conversion, and environmental remediation, arising from its superior electrical and thermal conductivity, chemical stability, mechanical stiffness, and other physical and chemical properties.³ For instance, graphene has been widely studied as an appealing adsorbent for hazardous gases like volatile organic compounds (VOCs), CO₂ and aqueous organic pollutants due to its high theoretical specific surface area (SSA, 2630 m² g⁻¹).⁴ Additionally, the unique 2D planar structure, large SSA, good electrical conductivity and chemical stability also make graphene an excellent catalyst and support in various catalytic reactions for environmental remediation.⁴ Moreover, the discovery of graphene has triggered research interest in many other 2D materials.

Regardless of the above merits, the zero-bandgap feature of graphene impeded its further applications in electronic devices like switches and light-emitting diodes, and photocatalysis.⁵ In order to open the bandgap of graphene, diverse structural modifications involving cutting graphene into narrow nanoribbons and doping with heteroelements have been attempted.⁶ Among them, doping with N atoms is a good option due to the similar relative atomic mass and electron configuration of the N element to the C element, enabling the feasible fitting of N atoms into the covalent carbon matrix.⁷ However, this chemical modification can inevitably lead to certain damage to the 2D carbon lattice of pristine graphene and the stoichiometry between C and N cannot be easily controlled via the top-down approach.⁸ With the motivation to supplement the electronic properties of the graphene system with other 2D materials that show similar atomic crystalline structures and intrinsic semiconductivity, bottom-up approaches with C-rich and N-rich precursors to fabricate 2D atomic carbon nitride (C_xN_y) crystals have been given much attention.⁹

Research on C_xN_y was mainly stimulated by the theoretical work of Lu and Cohen in 1989, which postulated that a carbon(IV) nitride (sp³-hybridized) crystal should show comparable or higher hardness than diamond.¹⁰ Since then, an exponentially growing number of publications exploring novel C_xN_y materials from both experimental and theoretical aspects have been reported, while most of them focused on the 3D C_xN_y-phases and few of them on the 2D crystalline C_xN_y materials. It was not until 2009 that polymeric graphitic-C₃N₄ (g-C₃N₄) was demonstrated as an excellent visible-light-responsive photocatalyst,¹¹ since then more and more efforts have been devoted to the synthesis of 2D g-C₃N₄ nanosheets via exfoliation of bulk g-C₃N₄ in order to further enhance the photocatalytic performance.¹² Recently, the experimental successes in 2D C₂N and C₃N crystals pushed the research of 2D C_xN_y materials to a new high level.¹³ A large number of attempts have been made on the preparation and potential applications of these new 2D materials as well as prediction of other types of 2D C_xN_y atomic crystals such as C₂N₂ and C₅N₂.¹⁴

Currently, 2D C_xN_y materials show great potential in field-effect transistors (FETs) owing to their semiconducting character, which complements the uses of graphene and insulating 2D boron nitride crystals.¹³ Meanwhile, the inherent bandgap of 2D C_xN_y materials makes them potential environment-friendly photocatalysts for different reactions like water splitting, degradation of organic pollutants, CO₂ reduction reaction (CO₂RR) and so on. Additionally, the introduction of N atoms into the carbon matrix can also break the inertness of pristine graphene and create Lewis basic sites, providing 2D C_xN_y with good catalytic activity in different reactions. Analogous to graphene, 2D C_xN_y materials show high theoretical SSAs, good chemical stability and other similar physicochemical properties as well.¹⁵ Therefore, they are considered as outstanding adsorbents for gas adsorption and catalyst supports for loading uniformly dispersed nanoparticles.^9c,16 In particular, the N atoms in the C_xN_y network can serve as anchoring sites for isolated metal atoms or dimers, leading to the formation of single atom catalysts (SACs) or double-atom catalysts as a hot topic in recent years.¹⁷

To date, a diversity of 2D C_xN_y materials with different stoichiometric ratios between C and N have been documented in the literature, which include g-C₃N₄, CN, C₂N, C₃N, C₂N₂, C₂N₃, C₅N and C₅N₂. For these C_xN_y compounds, the arrangements of C and N atoms in a basic benzene ring and the corresponding structural motif in the sp²-hydridized basal plane differ from each other (the structures of these C_xN_y materials are shown in Fig. 1). As a result, they exhibit distinct physicochemical properties with respect to the bandgap and texture. Among them, g-C₃N₄, C₂N and C₃N have aroused much interest in the field of photocatalysis because they have been already synthesized in laboratories and possess a suitable bandgap and good visible-light adsorption, while the rest of the 2D C_xN_y compounds are still in the theoretical research stage.

Fig. 1 The structures of different 2D C_xN_y compounds, including g-CN, g-C₃N₄, C₂N, C₃N, c-C₂N₂, h-C₂N₂, C₂N₃, C₅N and C₅N₂.

In this work, we present a comprehensive overview of the recent progress in 2D C_xN_y research. Considering a large quantity of reviews on 2D g-C₃N₄ in the literature,^12a herein, we will focus on the other members of 2D C_xN_y except g-C₃N₄, including g-CN, C₂N, C₃N, C₂N₂, C₂N₃, C₅N and C₅N₂. In Sections 2–4, we describe the structural characteristics, experimental synthesis, intrinsic properties and applications (practical and simulated) of C₂N, C₃N and g-CN, respectively. In Section 5, we will discuss the crystalline structure, electronic properties, and possible applications of C₂N₂, C₂N₃, C₅N and C₅N₂. Finally, we go beyond the fundamental properties of these 2D C_xN_y materials and give an outlook of the promising applications of such materials in more areas including environmental remediation in future.

2. C₂N

As early as 1998, Komatsu and Samejima reported the fabrication of amorphous C_xN_y with a random-layer structure and a chemical composition of C₂N through the shock-wave compression of a C- and N-containing polymer.¹⁸ In the later decade, little effort was devoted to the synthesis of the 2D C₂N crystal. In 2015, Baek et al. demonstrated the successful preparation of a 2D C₂N crystal and its crystalline structure with the assistance of advanced characterization tools. Afterwards, a large quantity of experimental and computational studies on 2D C₂N have been published due to its peculiar structure and physicochemical properties. In this section, the crystalline structure, synthetic methods, electronic properties, and applications of 2D C₂N are summarized.

2.1 Crystalline structure of C₂N

A C₂N monolayer can be considered as a nitrogenated holey graphene layer, in which the benzene rings are connected by nitrogen atoms to constitute large six-membered aromatic rings with two nitrogen atoms facing each other, leading to the formation of periodically uniform holes (Fig. 2).¹⁹ All the N atoms in the 2D C₂N structure belong to ‘pyridine-type’ or ‘pyrazine-type’ N.^9a The primitive cell of the C₂N monolayer is shown in Fig. 2, in which three bonds with different lengths are labeled.²⁰ The optimized lattice parameter (the inter-hole distance) is determined to be 8.30 Å by density functional theory (DFT) calculations. However, the lengths of C–N, C–C (1) and C–C (2) bonds vary slightly in different theoretical studies. For instance, Du et al.²¹ reported that the lengths of C–N, C–C (1) and C–C (2) bonds were calculated to be 1.336 Å, 1.429 Å and 1.470 Å, respectively, whereas Gong et al.²⁰ proposed that the lengths of the three bonds were estimated to be 1.339 Å, 1.432 Å, and 1.469 Å, respectively.

Fig. 2 The monolayer structure of the C₂N cell.

2.2 Synthesis of C₂N

The first synthesis report of the 2D C₂N crystal was given by Baek and co-workers in 2015^9a using the wet-chemical reaction of hexaaminobenzene (HAB) trihydrochloride and hexaketocyclohexane (HKH) octahydrate in N-methyl-2-pyrrolidone (NMP) with sulphuric acid or trifluoromethanesulphonic acid. Due to the spontaneous polycondensation between HAB and HKH, a layered crystalline 2D network structure was formed (Fig. 3a), referred to as the ‘C₂N holey 2D crystal’ or ‘C₂N-h2D crystal’. The 2D C₂N crystalline structure was validated by a series of characterization tools including X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), elemental analysis and electron microscopy (Fig. 3b–h).^9a

Fig. 3 The schematic representation of the C₂N-h2D crystal and its characterization: (a) the schematic representation of the C₂N-h2D crystal; (b) XRD pattern, (c) XPS survey spectrum, (d) SEM image, and (e) SEM image of heat-treated C₂N-h2D flakes; (f) TEM image, (g) high-resolution TEM image, and (h) STM topography image of the C₂N-h2D crystal, top-left inset is the structure of the C₂N-h2D crystal superimposed on the image. The scale bars and the inset are 2.0 nm, 2.0 nm⁻¹. Reprinted with permission from ref. 9a copyright 2015, Springer Nature.

The XRD pattern (Fig. 3b) revealed that the C₂N-h2D crystal was well-crystallized with an interlayer distance of 0.328 nm. Notably, this interspace is slightly lower than that of graphite (0.335 nm),²² attributed to the stronger interlayer interactions. XPS analysis indicated the presence of sp²-hybridized N atoms in the resultant crystals (Fig. 3c). With scanning electron microscopy (SEM, Fig. 3d), the as-prepared C₂N-h2D crystals were observed to be constituted by large and thick grains with sizes of a few hundred micrometers. Notably, a post-annealing treatment at 700 °C on the as-prepared C₂N-h2D crystals can reduce the agglomeration of the grains and thinner flakes were obtained (Fig. 3e). Meanwhile, wrinkled and transparent sheets similar to graphene were observed in the transmission electron microscopy (TEM) image of C₂N-h2D crystals (Fig. 3f) and highly crystallized microstructures of those sheets were demonstrated by the high-resolution TEM image (Fig. 3g). The interlayer distance of the C₂N-h2D crystal was measured to be 0.327 nm from the high-resolution TEM image, consistent with the XRD results. Furthermore, the evenly distributed holey structure in hexagonal arrays of a single C₂N layer as proposed by the theoretical model was unambiguously displayed by the scanning tunnelling microscopy (STM) image (Fig. 3h). The inter-hole distance of the as-prepared crystal was estimated to be ∼8.24 ± 0.96 Å according to the STM characterization, agreeing with the theoretical calculations. In Baek's subsequent research, different metal-based chlorides (CoCl₂, NiCl₂, RuCl₃, etc.) were introduced into the NMP solvent and 2D C₂N supported metallic nanoparticle composites were obtained.²³

More recently, Shinde and co-workers reported the preparation of C₂N aerogels (C₂NA) via a similar synthetic strategy to that of the C₂N-h2D crystal, which utilized hexaaminobenzene and chloroanilic acid as precursors.¹⁹ In their work, the polymerization of the two precursors in 1-methyl-2-pyrrolidinone solution was initiated by adding L-alanine and sulfuric acid and the mixture was heated at 120 °C to form a hydrogel. Afterwards, the hydrogel was freeze dried for 20 h and then was subjected to an annealing treatment in a N₂ atmosphere at 200 °C for 30 min to eliminate the oxygenated groups in the C₂NA (Fig. 4a). Meanwhile, the authors also fabricated sulfur-modulated C₂NA (S-C₂NA) and phosphorus-modulated C₂NA (P-C₂NA) via a similar procedure with the extra addition of L-cysteine and ammonium peroxydisulfate (for the S-C₂NA case) or L-phosphoserine and ammonium phosphate (for the P-C₂NA case) for the polymerization. The good crystallinity of C₂NA and doped C₂NA is also supported by the XRD patterns (Fig. 4b). Furthermore, successful doping of P or S atoms into C₂NA was validated by XPS analysis (Fig. 4c). The as-prepared aerogels show a dark grey and sponge-like appearance and low mass density (Fig. 4d). SEM and TEM micrographs reveal that the as-prepared aerogels are composed of interweaved, twisted and crystallized C₂N nanoribbons, which exhibit widths ranging from 150–200 nm (Fig. 4e and f). Meanwhile, the STM image confirms the holey C₂N structure of the nanoribbons in aerogels and the inter-hole spacing is measured to be ∼8.27 ± 1 Å (Fig. 4g), agreeing with the calculated studies.²⁰ Remarkably, the SSAs of S-C₂NA (1943 m² g⁻¹), P-C₂NA (1468 m² g⁻¹) and C₂NA (1145 m² g⁻¹) are much higher than that of the C₂N crystals prepared by Baek et al. (26 m² g⁻¹) (Fig. 4h).

Fig. 4 (a) Schematic illustration depicting the development of the S-C₂NA, (b) XRD patterns of graphite, and the pristine C₃N₄, C₂NA, P-C₂NA, and S-C₂NA, and (c) survey scans of the XPS spectra of pristine C₃N₄, C₂NA, P-C₂NA, and S-C₂NA. (d) Optical image of the fabricated S-C₂NA; the inset shows the ultralight aerogel suspended on a dandelion, (e) SEM image of S-C₂NA, (f) TEM images of the S-C₂NA (inset shows the HRTEM image), and (g) atomic-resolution inverted STM image of S-C₂NA on Cu (111). Inset shows the superimposition of the S-C₂NA structure (left), atom positions (top-right), and the topographic height profile along the dark red line (down-right). (h) N₂ adsorption–desorption isotherms of C₃N₄, C₂NA, P-C₂NA, and S-C₂NA. Reprinted with permission from ref. 19 copyright 2017 American Chemical Society.

As is known, good control over chemical reactions is highly important for tuning the local elemental distribution and structure at the atomic level of nanomaterials. In this context, Fechler and co-workers²⁴ proposed a eutectic synthesis approach using precursors containing pre-organized nitrogen atoms to realize a controlled polycondensation process instead of a spontaneous process for achieving 2D C₂N nanocrystals (Fig. 5a–c). A supramolecular pre-organized mixture of cyclohexanehexone and urea was adopted as precursors, in which the former served as the structure donating motif and the latter could reduce the melting point of the mixture and also provided a nitrogen source. Through a controlled condensation–polymerization reaction and a carbonization process at suitable temperature, graphene-like 2D C₂N with a stoichiometric ratio of 2 : 1 was fabricated. Based on this approach, other pre-organized precursors were also employed to synthesize graphene-like C₂N with a stoichiometric C/N ratio of 2 : 1, but yet disordered. For instance, Jordan et al.²⁵ successfully prepared laminar C₂N crystals starting from squaric acid and urea (Fig. 5d–f). Recently, Walczak et al.^9b and Li et al.²⁶ reported the synthesis of porous C₂N materials (SSA > 600 m² g⁻¹) via carbonizing a hexaazatriphenylene precursor at 550 °C while the resultant C₂N (C-HAT-CN-550) was rather thick and disorganized (Fig. 5g and h).

Fig. 5 (a) Reaction process scheme of C₂N by Fechler et al. (cyclohexanehexone reacted with urea reactants to produce a cross-linked intermediate, which was then converted stepwise to nitrogen functionalized carbons containing primarily pyridinic and pyrazinic nitrogens at temperatures above 500 °C (T₁ < T₂)) and (b) SEM and (c) TEM images of the resultant product. (a–c) Reprinted with permission from ref. 24 copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (d) Reaction process scheme of C₂N by Jordan et al. (squaric acid (1) reacted with urea (2), heated to form squaramide (3) by transamidation and squaramide was then converted into a C₂N material (4) via condensation and ring rearrangement at elevated temperature (T₁ < T₂)). (e and f) SEM images of the primary supramolecular crystals formed from squaric acid (SA) and urea (U) in a ratio of (a and c) 1 : 1 (SA/U-1) and (b and d) 1 : 2 (SA/U-2), respectively. The two kinds of complex were treated after thermal condensation of 550 °C. (d–f) Reprinted with permission from ref. 25 copyright 2016 Curtin University of Technology and John Wiley & Sons, Ltd. (g) Reaction progress of porous CN materials by Li et al. (h) TEM image of the resultant product. (g and h) Reprinted with permission from ref. 26 copyright 2019 Elsevier Ltd.

C₂N can be prepared by wet-chemical methods^9a,19 and solid phase methods.^9b,24–26 Compared to the solid phase method, the synthetic procedure of the wet-chemical method is more complex, and the reaction process of the wet-chemical method is more difficult to control. In addition, with different precursors and reaction conditions, the synthesized C₂N can have different morphologies, including nanosheets,¹³ aerogels,²⁷ zeolite-like,²⁶ shale-like and hollow-tube.²⁵ Accordingly, these products possess distinct structural characteristics and physicochemical properties with respect to SSA, defective level, carrier mobility, etc., which enable them to be applied in different fields.

2.3 Electronic properties of C₂N

Baek et al. examined the electronic properties of C₂N-h2D crystals by fabricating C₂N-h2D-based FET devices.^9a The semiconducting character of C₂N-h2D crystals was validated and a maximum on/off current ratio of 4.6 × 10⁷ was observed with the electron mobility and hole mobility of 13.5 cm² V⁻¹ s⁻¹ and 20.6 cm² V⁻¹ s⁻¹, respectively. Moreover, the bandgap was determined to be 1.96 eV, which is close to the value derived from DFT calculations (1.70 eV). Compared to g-C₃N₄ (2.7 eV),^15a the narrower bandgap of C₂N-h2D crystals provides them with better visible light absorption, which is beneficial to their photocatalytic applications. Similar to other 2D semiconducting materials, the bandgap of C₂N also varies with the crystal thickness. Unfortunately, most of the relevant research studies are theoretically conducted without support from experimental evidence. For instance, Zhang et al. performed a first principles study on the electronic properties of few-layer C₂N-h2D crystals with DFT calculations.³⁵ The authors reported that the band structure, especially splitting of the bands, and bandgap of this system depend on the stacking order between the layers, whereas the bandgap exhibits a monotonically decreasing behavior as the layer number increases. A recent theoretical study suggested that the direct bandgap of C₂N-h2D can be decreased by 34% with the evolution of structure from monolayer to bulk.³⁶ Therefore, these studies offer a strategy to modulate the bandgap of C₂N from the near-infrared to green visible wavelength by controlling the stacking order and layer number of the crystal.

The electronic properties of C₂N can also be modulated by doping with metal and non-metal elements (e.g. Fe, Co, Ga, Au, F, Cl, Si, Ge, P and As)^27–30 based on diverse computational studies. For example, Ma et al.³⁴ theoretically investigated the atomic configurations and electrical properties of C₂N monolayers embedded with a single atom of V, Cr and Mn. The local-density-of-states indicated that all three composites show a non-bandgap feature, similar to that of conductors. Du et al.³⁰ reported that the Fermi level of C₂N increased and the bandgap decreased with the increase of P/As doping concentration.

Additionally, combining C₂N with another semiconductor (e.g. GaTe,³¹ C₃B,³² g-C₃N₄,³³ BCN³⁸) to construct a heterojunction is also a feasible option to modulate the electronic properties of C₂N. For example, the bandgap of C₂N can be tuned from direct to indirect by forming a monolayer C₂N/GaTe heterostructure, as suggested by the calculation results of Bai et al.³¹ Moreover, the recombination of charge carriers in C₂N can be suppressed due to the formation of a built-in electric field in the heterojunction, which is desirable for its applications in optoelectronic devices and photocatalysis.

In summary, engineering the electronic properties of C₂N can be realized by thickness control, elemental doping and construction of heterostructures. For elemental doping, both the metal and non-metal dopants can modify the Fermi level and reduce the bandgap of C₂N, while the metal doping can significantly narrow the bandgap even down to zero, which makes the metal-doped C₂N not suitable for photocatalysis but as an electrocatalyst. Non-metal doping and constructing heterojunctions are effective strategies to modulate the band structure of C₂N while maintaining its semiconducting nature. Especially for heterostructured C₂N-based composites, both the light absorption and electron–hole pair separation efficiency are improved, making them appealing photocatalysts. The current theoretical progress on the electronic properties of C₂N has guiding significance for promoting its practical applications in areas like electronics, catalysis, and photocatalysis.

2.4 Applications of C₂N

With the intrinsic semiconducting properties and tunable electronic structure, C₂N holds great potential in photocatalysis. Therefore, many studies have demonstrated the photocatalytic performance of C₂N and C₂N-based materials in different reactions. Meanwhile, the 2D planar and nanoporous structure of C₂N make it an outstanding support for loading metal atoms to obtain metal@C₂N composite catalysts for various reactions. Especially, the edge N atoms in the nanopores of C₂N can effectively trap metal atoms, promoting the bonding between the metals and support. Furthermore, the nanoporosity of C₂N also makes it an efficient adsorbent for various gases. Overall, an increasing interest has been focused on the applications of C₂N-based materials in different areas.

2.4.1 Photocatalysts. As a semiconductor, the suitable bandgap of C₂N also enables it to serve as a promising visible light-activated photocatalyst. Zhang and Yang theoretically assessed the photocatalytic performances of 2D C₂N with different layers (1–5 layers) for water splitting.³⁷ Their calculations suggested that these few-layer C₂N nanocrystals have a good response to visible light and may serve as potential nonmetal photocatalysts for water splitting, while bilayer C₂N is the most promising one. Furthermore, heterostructured C₂N composite photocatalysts are proposed aiming at enhancing the photocatalytic activity of C₂N.³⁸ For instance, modelling of van der Waals heterostructures of BCN/C₂N and C₃N₄/C₂N was conducted by Zhang et al.³⁸ and Yang et al.,³³ respectively. The two studies both revealed that the photocatalytic activity of C₂N for water splitting is enhanced by constructing heterostructures due to the increased separation efficiency of charge carriers. Unfortunately, relevant experimental studies have barely been reported in the literature.

2.4.2 Catalyst support of electrochemical catalysis for water splitting. In light of the 2D planar structure and high SSA of C₂N, it is also considered as a suitable support for loading metals to synthesize heterogeneous catalysts for various reactions. Particularly, the N atoms in the aromatic rings of C₂N can serve as anchoring sites for metal atoms. Baek and co-workers investigated the catalytic activity of a cobalt nanoparticle-doped C₂N composite (Co@C₂N) for the hydrogen evolution reaction (HER) in alkaline sodium borohydride solution and observed that such a catalyst exhibited a better reactivity than the reported noble metal catalysts.²³ The authors stated that the enhanced catalytic activity and good stability originate from the strong interaction between the metallic nanoparticles and the substrate. In another study, they compared the HER performances of a diversity of metal nanoparticle-doped C₂N composites (Co@C₂N, Ru@C₂N, Ni@C₂N, Pd@C₂N and Pt@C₂N).⁴⁰ Ru@C₂N outperformed the other four composites as well as the typical Pt/C catalyst in both acidic and alkaline solutions. S-C₂NA was also demonstrated to possess excellent catalytic activities in both oxygen evolution reaction (OER) and oxygen reduction reaction,¹⁹ deriving from its high SSA and a high proportion of pyridinic plus graphitic N catalytic sites in the total N atoms (55%). With the superior bifunctional performance, the Zn–air and Li–O₂ batteries based on the S-C₂NA electrode show high efficiency in energy conversion and storage. Additionally, the HER and OER activities of C₂N supported SACs (metal₁@C₂N) are also assessed by calculating the overpotential with first principles computations.^38,41 Zhang's group conducted a systematic study on the catalytic properties of a wide range of C₂N supported transition metal (TM) catalysts (TM₁@C₂N).⁴² The pure C₂N was discovered to be inactive for the HER due to the strong bonding between H₂ and N atoms, but the doping with TMs significantly modified the catalytic activity. By comparison, Ti₁@C₂N manifested the best HER activity but poor OER activity, whereas Mn₁@C₂N was found to be a suitable bifunctional electrocatalyst for both HER and OER because it exhibited an overpotential of only 0.67 V in OER and the Gibbs free energy changes for H₂ adsorption on Mn₁@C₂N were slightly above zero.

2.4.3 Catalyst support for CO oxidation. The catalytic activity of C₂N supported metal catalysts in CO oxidation has also been investigated, while the relevant studies are mainly performed by theoretical calculations.^16b,34,39,43 For instance, He et al.^43a predicted the stable presence of a single Fe atom trapped in the cavity of a C₂N monolayer by calculating the adsorption energy and diffusion barrier and demonstrated the excellent CO oxidation ability of the Fe-embedded C₂N SAC at different temperatures. Ma et al.³⁴ used the first-principles calculation to study CO oxidation catalyzed by various 3d TM-embedded C₂N SACs (TM = Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn). Their results illustrated that C₂N monolayers are promising substrates for 3d TM SACs, while Cr- and Mn-embedded C₂N catalysts show the best catalytic CO oxidation performances. Recently, Li and Chen^16b systemically investigated the reaction mechanism of CO oxidation and the gas adsorption with the single-atom and bi-atom catalysts of copper. The C₂N embedded with a copper dimer (Cu₂@C₂N) catalyst was revealed to show a superior catalytic performance to the SAC (Cu₁@C₂N). Meanwhile, different mechanisms of CO oxidation by Cu₁@C₂N and Cu₂@C₂N were proposed. The Eley–Rideal mechanism is responsible for the Cu₂@C₂N-catalyzed CO oxidation rather than the Langmuir–Hinshelwood mechanism, which is contrary to the metal-doped (metal = Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, etc.) graphene,⁴⁴ but similar to the Ni-embedded divacancy graphene.⁴⁵ A tri-molecular Eley–Rideal (T-E–R) mechanism accounts for the Cu₁@C₂N-catalyzed CO oxidation in which the adsorption of two CO molecules can assist the breaking of the O–O bond in O₂ (Fig. 6). Moreover, a doped C₂N catalyst with oxygen was also proposed to be capable of improving the CO oxidation performance due to the extra electrons brought by oxygen.³⁹

Fig. 6 Atomic configurations of S2–S4 and the key structural parameters for CO oxidation via the T-E–R mechanism on Cu₂@C₂N. Color scheme: C, gray; N, blue; O, red; Cu, orange. Reprinted with permission from ref. 16b, copyright 2018 The Royal Society of Chemistry.

2.4.4 Catalyst supports for the CO₂ reduction reaction (CO₂RR) and nitrogen reduction reaction (NRR). C₂N supported metals have also been verified to be efficient catalysts for the CO₂RR and NRR. For example, Cui et al.²⁷ tested the CO₂RR activity of twelve types of SACs supported by C₂N through calculations. The results implied that the product selectivity (methane or methanol) can be tuned by changing the embedded metal atom species. The CO₂ electro-reduction performances of Cu₁@C₂N and Cu₂@C₂N as reported by Zhao et al.⁴⁶ were also compared and the former was still inferior to the latter. Wang et al.⁴⁷ firstly reported that Mo@C₂N shows a low onset potential of −0.17 V in NRR via DFT calculations, indicating that it may be a low-cost and high-efficient catalyst for electrochemical conversion of N₂ to NH₃. Recently, Zhou's group performed a systematic investigation on the NRR activity of nine C₂N supported TM SACs or bi-atom catalysts.⁴⁸ This comprehensive study predicted that the Mo bi-atom catalyst supported by C₂N is the most suitable one for the NRR.

2.4.5 Other applications. Apart from catalyst support, the uniformly distributed pores, special pore size and outstanding chemical stability of C₂N make it a good candidate for membrane materials for gas separation, H₂ storage, CO₂ adsorption, seawater pervaporation and volatile organic compound (VOC) adsorption.⁴⁹

The width of the pores in C₂N is estimated to be 3.0 Å, smaller than that of O₂, N₂, CO, and CO₂, but larger than the kinetic diameter of H₂ (2.89 Å) and He (2.6 Å).^49f Thus, C₂N may serve as a promising H₂ and He gas separation membrane material. In addition, the unique porous structure of C₂N also makes it a potential adsorbent for gases like H₂, CO₂ and VOCs. Hashmi et al.^49b explored the H₂ adsorption of C₂N and Li decorated C₂N using the DFT method, and found that the Li decoration significantly improved H₂ adsorption on C₂N. A theoretical gravimetric density of 13 wt% was reached for Li decorated C₂N, which is much higher than that reported for other 2D materials decorated with alkali metals. Walczak and co-workers^49d reported that CO₂ adsorption uptake of the prepared C₂N sample (C-HAT-CN-550) reached a high value of 3.24 mmol g⁻¹ (at 273 K and p = 0.15 bar). Lu's group²⁸ theoretically studied CO₂ capture properties on metal-embedded C₂N. Incorporation of open metal sites into the C₂N framework can enhance the bonding of CO₂ on the adsorbent along with low-temperature desorption capacity. The authors screened a series of metal-embedded C₂N including Ca and the early 3d TMs as CO₂ adsorbents and found that Ca-embedded C₂N shows the most promising reversible CO₂ capture performance. From DFT calculations, Ao et al.^49e found that pure C₂N can adsorb some typical VOCs like formaldehyde, benzene and trichloroethylene. Moreover, the adsorption ability of C₂N can be improved by Al doping. In their later study, the Al-doped C₂N was found to co-adsorb H₂O and O₂, implying the possible generation of hydroxyl and superoxide radicals, showing promising applications for VOC degradation as the hydroxyl and superoxide radicals are highly oxidizing species.⁵⁰

Besides the aforementioned applications, Xu and co-workers also applied C₂N as an electrode material for lithium-ion batteries (LIBs) and a high initial discharge capacity and high stability of C₂N-based LIBs were observed.^13a

2.5 Summary of C₂N

Thanks to the efforts of Baek and co-workers, the 2D C₂N crystal has attracted wide attention in the last five years, due to its high stability, hole structure, high SSA, and other physical and chemical properties. Although there are several methods to synthesize C₂N, the complex synthesis process and the expensive and rare precursors limit its large-scale production and practical applications. Therefore, it is urgent to develop low-cost and easy-operation methods to fabricate the 2D C₂N crystal. As a kind of semiconductor, the 2D C₂N crystal has a tunable bandgap, high carrier mobility, and special porous structure, which enable it to perform well in areas like FET,^9a HER²³ and LIBs.^13a In addition, based on DFT calculations, researchers reported the promising use of 2D C₂N materials in the OER,⁴² CO oxidation,³⁴ CO₂RR,²⁷ NRR,⁴⁸ and gas adsorption.⁴⁹ Moreover, Mahmood et al.^9a predicted that C₂N could be used in electrodes for energy storage, gas storage and some other areas. The applications of C₂N are summarized in Fig. 7.

Fig. 7 Potential applications of the C₂N-h2D crystals.

3. C₃N

The possible crystal structure of 2D C₃N was predicted as an indirect-bandgap semiconductor a few decades ago and it aroused much interest from the scientific community.⁵¹ Therefore, a considerable amount of effort on its synthesis was attempted, but to achieve amorphous products in a C₃N stoichiometry without a 2D structure, which could not testify the theoretical calculations.⁵² Until 2016, Baek and co-workers^13a,53 successfully synthesized a 2D layered C₃N nanocrystal, whose structure was in good agreement with the predictions. Then, this new member of the 2D C_xN_y family has been widely applied as a catalyst, electrode material, and adsorbent, due to its peculiar electrical and chemical properties.

3.1 Crystalline Structure of C₃N

A C₃N monolayer can be regarded as a N-doped graphene structure, where the molar ratio of C to N is 3 : 1 and all the N atoms belong to graphitic N. Similar to graphene, the atoms in C₃N also exhibit sp² hybridization.⁵⁴ Apparently, the C₃N framework is quite different from the C₂N framework, where periodic holes are not included. The monolayer C₃N has a lattice parameter of 4.863 Å, with the C–N bond length of 1.403 Å and the C–C bond length of 1.404 Å.

3.2 Synthesis of C₃N

Compared to C₂N, there are less available methods for the synthesis of C₃N. King et al.⁵⁵ obtained a mixture of amorphous C₃N flakes with a thickness of 1–3 μm and graphitic carbon microspheres through annealing a nitrogen-rich polymer precursor (m-phenylenediamine) in a vacuum-sealed quartz tube at 800 °C for 7–9 days. To improve the crystallinity of C₃N, Baek and co-workers^13a,53 then used organic hexaaminobenzene trihydrochloride (HAB, C₆N₆H₁₅Cl₃) crystals as precursors and successfully obtained 2D layered C₃N nanocrystals by decomposing the HAB into C₆N₂, NH₄Cl, and NH₃ at 500 °C for 2 h under an argon atmosphere (Fig. 8a). The resultant C₃N nanocrystal is multilayered and represents a similar hexagonal morphology to the HAB crystals.⁵³ The interlayer distance of the as-prepared C₃N nanocrystal is estimated to be 3.40 Å according to the XRD characteristic peak at 2θ of 26.12° (Fig. 8b). In addition, another two major peaks at 2θ of 12.8° and 44.7° are also present in the XRD pattern. Meanwhile, the authors confirmed the C₃N chemical composition, molecular structure and bond nature of the as-prepared nanocrystal via XPS (Fig. 8c).⁵³ Furthermore, the 2D C₃N framework at the atomic level was unambiguously illustrated with the assistance of the STM method and the dot-to-dot distance in the well-ordered triangular structure of C₃N was found to be 442 ± 16 pm, which was in good match with the DFT modelling (Fig. 8e).

Fig. 8 (a) Schematic representation of C₃N, (b) experimental PXRD pattern from C₃N (dark green), simulated XRD after Pawley refinement (dark red) and the difference (dark blue), (c) XPS survey spectrum of C₃N, (d) SEM image of C₃N, and (e) STM image of a 2D PANI framework (2.5 × 2.5 nm², V_s. = −1.1 V, I_t = 1.0 nA). Inset structure represents C₃N repeating units with carbon atoms (gray balls) and nitrogen atoms (blue balls). Reprinted with permission from ref. 53 copyright 2016 National Academy of Sciences.

Recently, fabrication of C₃N quantum dots (QDs) was reported by Yang et al.^13b In their work, a precursor of 2,3-diaminophenazine (DAP) is polymerized by a typical hydrothermal method, forming C₃N QDs and C₃N nanosheets (Fig. 9a). The layered C₃N was formed at a DAP concentration of 0.8 mM and 250 °C. Especially, by prolonging the polymerization reaction time to 200 h, 2D C₃N nanocrystals containing a high proportion of large, single-layer and single crystalline nanosheets can be obtained (Fig. 9b and c). For QDs, the size can be controlled in the range of 1.8–5.5 nm by varying the concentration of DAP solution and hydrothermal temperature (Fig. 9d and e).

Fig. 9 (a) Schematic of the vertical and horizontal polymerization of DAP forming C–N bonds. The polymerization advances via abstraction of H from C–H and N–H bonds followed by formation of C–N bonds. (b) TEM image of a mixture of single- and multilayer C₃N sheets, (c) spherical aberration-corrected HRTEM image showing the honeycomb structure of a single-layer sheet marked in (b), and (d) spherical aberration-corrected HRTEM image of a single C₃N QD with a C–N zigzag edged structure. The yellow balls represent the carbon and nitrogen atoms on the edge, while the red sticks represent the chemical bonds between atoms on the edge. (e) The molecular structure scheme of the C₃N QD in (d). Reprinted with permission from ref. 13b copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

3.3 Electronic properties of C₃N

C₃N is an indirect bandgap semiconductor with the valence and conduction bands located at the M and Γ points, respectively.⁵⁶ Determination of the theoretical bandgap of a C₃N monolayer is still debatable since different computation software gives discrepant values while convincing experimental evidence is not available thus far.^51b,54 For instance, Zhou et al.^56a derived a bandgap of 1.04 and 0.39 eV for a C₃N monolayer by using HSE06 software and the generalized gradient approximation (GGA) method, respectively.

Experimentally, Mahmood et al.⁵³ measured the bandgap of an as-prepared 2D C₃N framework to be 2.67 eV by using a lock-in detection technique with scanning tunnelling spectroscopy and electrochemical methods. This result is far away from the theoretical value, possibly because the thickness can affect the bandgap of C₃N nanocrystals. Surprisingly, Mahmood and co-workers found that doping the as-prepared multilayered C₃N with HCl could remarkably reduce its bandgap since the conductivity was improved from 0.72 ± 0.04 S cm⁻¹ (undoped) to 1.41 × 10³ S cm⁻¹ (doped). In fact, the HCl-doped 2D C₃N almost behaves as a conductor, which is more coincident with the calculations.

Yang et al.^13b measured the bandgap of C₃N QDs based on ultraviolet photoelectron spectroscopy and UV-vis spectroscopy. It was found that the bandgap value of C₃N QDs decreases with increasing diameters of the QDs, e.g., the bandgap varies from 2.74 to 0.39 eV with the size of QDs ranging from 1.8 to 30 nm. A further increase of the QD size larger than 30 nm barely affects the bandgap. Such a feature of size-dependent bandgaps provides C₃N QDs with tunable photoluminescence, favoring their applications in bioimaging. Notably, the unique bandgap structure of C₃N QDs also enables them to have a much higher quantum yield and life-time than graphene QDs. In addition, Yang et al.^13b also investigated the electrical properties of a monolayer C₃N nanosheet with FET devices. The average I_ON/I_OFF ratio for the C₃N-based FETs reached 5.5 × 10¹⁰ and the hole and electron mobilities of the C₃N monolayer were measured to be 1.2 and 1.5 cm² V⁻¹ s⁻¹ respectively. Notably, the I_ON/I_OFF ratio of C₃N is three orders of magnitude higher while the carrier mobility is lower than that of C₂N-h2D crystals. Interestingly, theoretical calculations reported that the carrier mobility of C₃N is anisotropic and the hole mobility along the armchair direction of the C₃N monolayer can reach up to 1.08 × 10⁴ cm² V⁻¹ s⁻¹, much higher than that of C₂N-h2D crystals.^56b These results demonstrate that the intrinsic conductivity and carrier mobility of C₃N are better than those of C₂N, indicative of the promising applications of C₃N in FET devices and electrode materials. The hole and electron mobility of C₃N could be further improved to 180 and 220 cm² V⁻¹ s⁻¹, respectively, by hydrogenation in H₂, while the intrinsic mechanism is unknown yet. Interestingly, the hydrogenated C₃N nanosheet even exhibits a ferromagnetic state according to both the DFT calculations and magnetic performance measurements.

Several theoretical studies demonstrated that the electronic properties of C₃N can also be modulated by forming heterojunctions with another semiconductor (e.g. g-C₃N₄,⁵⁷ C₃B⁵⁸). Due to the narrow bandgap and good carrier mobility of C₃N, it can act as a photosensitizer to enhance the visible light absorption and also retard the charge recombination of the heterostructure, which can improve the photocatalytic activity.

3.4 Application of C₃N

As a semiconductor with a small bandgap, C₃N shows a broadband solar light absorption, which indicates its potential utilization in photocatalysis. Meanwhile, because of the structural similarity to N-doped graphene which exhibits good electrocatalytic activities for reactions like the ORR, HER, etc., the potential application of C₃N in electrocatalysis was investigated. Furthermore, C₃N is also considered as a promising electrode material for batteries and as a gas adsorbent owing to its peculiar electrical properties and planar structure. The applications of C₃N in photocatalysis, electrocatalysis, batteries and gas adsorption are summarized in this section.

3.4.1 Photocatalysts. Due to the wide spectral responsive range and high carrier mobility, C₃N is typically combined with other semiconductors to construct heterostructured photocatalysts for enhanced photocatalytic performance. For instance, Wang et al.⁵⁷ designed C₃N/g-C₃N₄ heterojunctions and evaluated their ability for photocatalytic water splitting via DFT calculations. The authors found that the strong interlayer interaction between C₃N and g-C₃N₄ nanosheets drives the charge transfer in the heterojunction to follow a Z-scheme pathway, which significantly improves the carrier separation efficiency. Meanwhile, the integrated C₃N/g-C₃N₄ system has a wider light absorption range from the ultraviolet to the near infrared region compared to the single g-C₃N₄ system that hardly absorbs visible light. With the improved charge separation, stronger visible light absorption ability and suitable valence band position, the C₃N/g-C₃N₄ heterojunction can serve as an excellent photocatalyst for overall water splitting under visible light irradiation. Very recently, Niu et al.⁵⁸ computationally demonstrated that the C₃B/C₃N heterojunction also functions as a promising metal-free direct Z-scheme photocatalyst for water splitting under visible light irradiation. Due to the p-type semiconducting feature of C₃B and n-type feature of C₃N, they interacted with each other intensely to result in the formation of a built-in electric field from C₃N to C₃B in the heterojunction, which remarkably retards the electron–hole recombination rate and promotes the redox ability of photoinduced charges (electrons and holes) for H₂O molecules. Additionally, the Lewis acid B atoms in the C₃B lattice were found to be capable of adsorbing H₂O molecules efficiently, which also benefits the photocatalytic water splitting reaction.

3.4.2 Electrocatalysts. Several theoretical studies have evaluated the electrocatalytic water splitting performance of C₃N-based catalysts.⁵⁹ Pristine C₃N is found to exhibit a poor electrocatalytic activity for water splitting, while chemical modifications of C₃N can significantly improve its catalytic properties. He et al.^59b reported that the overpotential of B-doped C₃N in the ORR (0.6 V) was calculated to be slightly higher than that of the commonly used Pt electrocatalyst (0.45 V), indicating the comparable electrocatalytic activity of B-doped C₃N to Pt for the ORR. The excellent electrocatalytic performance of B-doped C₃N is attributed to the reason that the electron-deficient B atoms can effectively adsorb O₂ molecules and the B–N and B–C bonds are quite active towards the ORR. In addition, Xu et al.^59a showed that the B-doped C₂N/C₃N coplanar heterojunction is a promising electrocatalyst for the HER because of the good electrical conductivity, accelerated charge transfer and the low Tafel barrier of 0.86 eV which favor the HER at room temperature. Recently, Zhou et al.^59c studied the OER and ORR properties of the C₃N monolayer embedded with different types of metal atoms (Mn, Fe, Co, Ni, Rh, Pd, Ir and Pt) based on DFT calculations. All the metal-embedded C₃N monolayers exhibit metallic properties and the Rh-embedded C₃N shows the best OER and ORR activities with the overpotentials of 0.35 V and 0.27 V, respectively.

3.4.3 Electrode materials for batteries. Owing to the merits of large net positive charge densities, outstanding structural stability and enhanced electronic/ionic conductivity from the intrinsic 2D polyaniline framework of C₃N, this material is regarded as a promising alternative to graphite and graphene/carbon-based materials for the anode of metal ion batteries. Xu et al.^13a evaluated the performance of a multilayered C₃N anode in LIBs and also made a comparison with C₂N, heat-treated C₂N at 450 °C (C₂N-450) and graphite (Fig. 10a). Interestingly, C₂N-450 manifests a higher reversible capacity but still lower rate capacity than C₂N, possibly because the heating treatment can improve the crystallinity, SSA and electrical conductivity of C₂N. C₃N displays a voltage hysteresis of 0.15 V (Fig. 10b) lower than that of C₂N-450 (Fig. 10c), which results from the lower SSA (C₃N: 54.4 m² g⁻¹, C₂N-450: 280.5 m² g⁻¹) and higher electrical conductivity. Furthermore, C₃N showed better cycling stability and a more efficient operating voltage window (nearly 85% of full discharge capacity secured at above 0.45 V) than C₂N-450 (nearly 65% of full discharge capacity secured at above 0.45 V) (Fig. 10d and e). Therefore, in terms of capacity, voltage hysteresis, efficient operating voltage window and cycling stability, which are the key factors for practical applications of electrode materials in LIBs, C₃N is more promising than C₂N-450 as well as graphene.

Fig. 10 (a) Rate capability measurements for C₂N, C₂N-450, C₃N, and graphite at various C-rates from 0.1 to 10C. Representative cyclic voltammograms of (b) C₃N and (c) C₂N-450 for the five cycles at a scan rate of 0.1 mV s⁻¹. Charge/discharge profiles of (d) C₃N and (e) C₂N-450 at 1C. Reprinted with permission from ref. 13a copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

In the above work,^13a a significant capacity loss in the first several cycles was witnessed (Fig. 10a), which limits its application in LIBs. Liu and co-workers⁶⁰ performed first-principles simulations on the intercalation process of Li ions into C₃N and suggested that the behavior is a result of N excess C₃N (C_2.67N) in Xu's experiment.^13a With the N excess in the C₃N framework, N-doping and carbon-vacancy defects are present, which strengthen the intercalation of Li ions into C_2.67N and the energy barriers of the deintercalation process. It was proposed that C_3.27N maintained high reversible capacity in LIBs. Later, Preeti et al.⁶¹ demonstrated that the C₃N monolayer also serves as a promising anode material for Na/K-ion batteries with a storage capacity of up to 1072 mA h g⁻¹via first-principles calculations.

3.4.4 Gas adsorbent. Graphene has been widely applied as an adsorbent for various gases due to its extraordinary surface-volume ratio.⁶² As a graphene-like 2D material, the gas adsorption capacities of C₃N have also been widely studied.^16a Particularly, the substitution of N atoms with C atoms not only breaks the chemical inertness of the graphene framework but also lowers the work function, which enables C₃N to probably show better chemical activity towards gas molecules. For instance, Ma et al.⁶³ simulated the adsorption configuration of a series of gas molecules involving NO₂, NO, CO₂, and CO onto a C₃N monolayer. Compared to graphene,⁶⁴ the adsorption energy and charge transfer between the four molecules and C₃N are significantly enhanced. Meanwhile, the calculations showed that weak chemical adsorption with a large adsorption energy of ∼0.8 eV occurred between NO₂ and C₃N, while the adsorption between the other gas molecules and C₃N belongs to physisorption with a lower adsorption energy.

Gas adsorption capacities of the C₃N monolayer were found to be improved by doping with inorganic or metal elements. Pashangpour and Peyghan⁶⁵ assessed the adsorption of CO onto B-doped and Al-doped C₃N nanosheets by DFT calculations. The introduction of B dopants barely promoted the weak physisorption of CO onto the pristine C₃N nanosheet, whereas the Al doping significantly enhanced the interaction between CO and C₃N nanosheets. Moreover, CO adsorption onto the Al-doped C₃N nanosheet induces noticeable changes of electrical conductivity of the sheet. These findings implied the potential application of Al-doped C₃N nanosheets as both the CO adsorbent and sensors. Recently, several computational studies demonstrated the excellently reversible CO₂ capture capacity on the negatively polarized C₃N monolayer.⁶⁶ It was shown that the negative polarization could turn the physisorption of CO₂ into chemisorption. In addition, the negatively charged C₃N monolayer was a promising material to separate CO₂ from the mixtures of CO₂/N₂, CO₂/H₂, CO₂/CH₄, and CO₂/C₂H₂, and to separate C₂H₂ from the mixtures of C₂H₂/CH₄ and C₂H₂/H₂ through manipulating the negative charge density.^66b

Multilayered C₃N crystals are also reported to be highly efficient for gas adsorption,⁶⁷ similar to multilayered graphene.⁶⁸ Li et al.^16a constructed a periodic 3D pore model containing two C₃N monolayers (denoted as C₃N pores) with a pore size of 0.75 nm (Fig. 11a) and showed that the C₃N pores manifest an extremely high CO₂ adsorption capacity up to 3.99 mmol g⁻¹ in air at 23 °C. Moreover, the C₃N pores show preferential adsorption of CO₂ over CO, H₂, and CH₄ and the CO₂ selectivity could be further boosted by regulating the size of C₃N pores from 0.75 to 0.65 nm (Fig. 11b–d). The high selectivities in CO₂/CO, CO₂/H₂, and CO₂/CH₄ are ascribed to the larger polarizability and quadrupole moment of CO₂ molecules than the other three gas molecules, which enables stronger interaction between CO₂ molecules and the C₃N surface. Furthermore, the strong interaction between CO₂ molecules and the C₃N surface causes the CO₂ adsorption to slightly decrease in the presence of water (3.80 mmol g⁻¹). Recently, Ma et al.⁶⁹ used ZIF-8 and urea as precursors to synthesize C₃N-type N-rich carbon layers and applied them in methanol adsorption. A sample (CN950) with a SSA of 1496 m² g⁻¹ reached the highest methanol adsorption capacity of 13.2 mmol g⁻¹.

Fig. 11 (a) Molecular model of C₃N pores with a pore size of 0.75 nm. (b) CO₂/CO, (c) CO₂/H₂, and (d) CO₂/CH₄ selectivity of C₃N pores with different pore sizes versus the adsorption pressure at 23 °C. Reprinted with permission from ref. 16a copyright 2017 American Chemical Society.

3.5 Summary of C₃N

C₃N has a graphene-like crystalline structure, good chemical stability and an indirect bandgap, enabling plentiful applications in electronics, photocatalysis, electrocatalysis, energy storage devices, and gas adsorption. Multilayered C₃N can be synthesized via a carbonization process of precursors containing C and N, which is much simpler than that for C₂N layers. However, available methods for preparing single-layered or few-layered C₃N nanosheets and controllable modifications of C₃N are quite limited. Due to the much narrower bandgap than that of the C₂N monolayer, the C₃N monolayer itself possesses limited photocatalytic activity and it usually combines with other wide-bandgap semiconductors to form heterojunctions for photocatalysis. However, the better electrical conductivity and chemical stability of C₃N than C₂N enables it to have more promising applications in LIBs. In addition, the electrocatalytic activity of C₃N can be improved by rational chemical modifications. Furthermore, the gas adsorption capacity of C₃N can be tailored by doping, polarization and regulation of C₃N pores, making it a good candidate for a gas adsorbent. Definitely, the physicochemical and structural characteristics of C₃N are not fully understood thus far and there are still many potential applications for C₃N in the areas of environment and energy (as shown in Fig. 12). It is hoped that more efforts will be devoted to advance our knowledge in C₃N.

Fig. 12 Potential applications of C₃N.

4 Graphite-like CN (g-CN)

The presence of g-CN nanocrystals has been reported for a long time, but little progress on its synthesis and applications has been made. In 2006, Cao et al.⁷⁰ proposed a simple solvothermal approach with cyanuric chloride and sodium precursors to obtain 1D CN nanostructures, which involves an initial formation of graphite-like carbon nitride (g-CN) layers followed by self-assembly into tubes, ribbons or microspheres (Fig. 13a–c). So far, the successful preparation of CN is absent in the literature. On the other hand, many theoretical studies have been conducted on the properties and potential applications of g-CN.

Fig. 13 (a) TEM image of a bundle of g-CN nanotubes and the embedded SAED, (b) TEM image of a bundle of g-CN nanoribbons (SAED embedded), and (c) SEM image of g-CN microspheres, reprinted with permission from ref. 70 copyright 2006 Elsevier Ltd.

4.1 Crystalline structure of g-CN

The g-CN monolayer structure is composed of uniform holes and aromatic benzene rings containing three N atoms and three C atoms alternatively arranged in the ring (Fig. 1).⁷¹ The benzene rings are bridged by C–C bonds to form nanopores in g-CN, and all the N atoms belong to pyridine N. The lattice parameter is derived to be 7.12 Å, and the bond lengths of C–C and C–N are estimated to be 1.51 and 1.34 Å, respectively.

4.2 Electronic properties of g-CN

g-CN is predicted to behave as a semiconductor. To calculate the bandgap of the g-CN monolayer, Srinivasu et al.⁷² used different methods and obtained different values. They stated that 2.89 eV should be more plausible based on Cao's work,⁷⁰ which observed broad photoluminescence peaks at 430–450 nm (corresponding to a bandgap of ∼2.73 eV) for the as-prepared g-CN nanotubes, nanoribbons and microspheres. Unluckily, experimental measurements in the bandgap of 2D g-CN crystals are unavailable yet. Meanwhile, Srinivasu et al.^72b also studied the bandgap of multilayered g-CN and suggested that the bandgap decreased gradually from 2.89 to 2.72 eV with the layer number in the range of 1–6, which originates from the interlayer coupling. Furthermore, they found that doping the g-CN monolayer with phosphorus atoms (P@g-CN) could further reduce its bandgap to 2.31 eV, which corresponds to an appreciable visible light, whereas boron, oxygen and sulfur dopants are less efficient. Recently, Li et al.⁴¹ reported that embedding a TM atom (TM = Pt, Pd, Co, Ni and Cu) into the g-CN framework could transform g-CN from a direct bandgap into an indirect bandgap with values of 0.41–0.52 eV. With the maneuverable band structure, g-CN is also expected to show great potential in photocatalysis.

4.3 Applications of g-CN

Unlike C₂N and C₃N, only a few examples of the applications of g-CN have been presented from computations in the literature. For instance, Srinivasu et al.^72b evaluated the potential of g-CN, B@g-CN and P@g-CN for photocatalytic water splitting. P@g-CN is supposed to exhibit the best HER performance under visible light among the three samples due to its smaller bandgap, suitable conduction band minimum (CBM) position above the water oxidation potential and valence band minimum (VBM) position below the water reduction (Fig. 14a). In addition, the similar nanoporosity of g-CN to that of the C₂N framework also makes it an excellent support of SACs for various reactions because the sp²-hybridized N atoms at the vacancy hole edges can afford coordination sites to anchor the metal atoms. In this regard, Li et al.⁴¹ designed a series of g-CN supported TM (TM = Pt, Pd, Co, Ni and Cu) SACs and studied their OER performances by DFT calculations (Fig. 14b). The computational results displayed that all the TM@CN materials function as good electrocatalysts for the OER and Co@CN is the most ideal one due to its lowest overpotential of ∼0.16 V. Although the TM@g-CN materials are not efficient to drive the water splitting under visible light irradiation due to their narrow bandgap (0.41–0.52 eV), the stable Pd-(OH)₂@g-CN complex with modified band structure exhibits a promising photocatalytic activity for visible-light-driven OER. Recently, Chen et al.⁷¹ demonstrated that a Li decorated g-CN monolayer presents a high efficiency in hydrogen storage systems, because of sufficient adsorption sites for hydrogen molecules. With the high Li loading, the Li decorated g-CN shows a good hydrogen adsorption performance with a theoretical gravimetric density of 10.81 wt%.

Fig. 14 (a) Calculated VBM and CBM positions of g-C₃N₄ and g-CN along with their boron and phosphorus doped counter parts, reprinted with permission from ref. 72b copyright 2014 American Chemical Society. (b) The energy band alignment of TM–(OH)₂@CN. The vacuum level is set at 0 eV, and the chemical reaction potentials for H⁺/H₂ and O₂/H₂O are plotted with dotted lines. Reprinted with permission from ref. 41 copyright 2016 The Royal Society of Chemistry.

5. Other C_xN_y

Besides the aforementioned 2D C_xN_y structures, several other 2D C_xN_y compounds with different stoichiometries including C₂N₂, C₂N₃, C₅N, and C₅N₂ have also been reported in the literature, while less attention has been focused on them. Hence, we will give a brief introduction in this section.

5.1 Diamond-like CN (c-C₂N₂) and hexagonal CN (h-C₂N₂)

Recently, two new types of stable 2D C_xN_y were predicted by Dong et al.¹⁴via the DFT method, c-C₂N₂ and h-C₂N₂. It is worth mentioning that the crystalline structures of the two C₂N₂ materials are quite different from the above typical monolayer-structured C_xN_y, since there are two atomic layers in the unit cells of c-C₂N₂ and h-C₂N₂. As shown in Fig. 15, both of them are constituted by two g-CN layers connected with C–C bonds, while the stacking modes of the layers are different. As a result, C atoms exhibit sp³ hybridization and N atoms are sp² hybridized in both the c-C₂N₂ and h-C₂N₂ frameworks. According to the calculation results, c-C₂N₂ and h-C₂N₂ are wide-band semiconductors with the bandgaps of 4.43 and 3.70 eV, respectively. Meanwhile, good mechanical strength and high thermal conductivity of the two C_xN_y are also predicted. Unfortunately, the synthesis of the two structures has not been realized yet in the laboratory to provide experimental confirmation.

Fig. 15 Lattice structures of c-C₂N₂ (a) and h-C₂N₂ (b). Carbon atoms are colored with black in the bottom layer and red in the top layer. Nitrogen atoms are colored with green in the bottom layer and blue in the top layer. Reprinted with permission from ref. 14 copyright 2018 Elsevier B.V.

5.2 C₂N₃

Recently, Shi et al.⁷³ firstly studied the structural characteristics of C₂N₃ and its structural stability and strain behavior by DFT calculations. The C₂N₃ monolayer is a kind of carbon nitride with a C/N ratio of 2/3, which has the calculated lattice contents of a = 7.846 Å and c = 6.549 Å with a P6/mcm space group (Fig. 1). The bond lengths are 1.366 Å, 1.360 Å, and 1.340 Å for the N(1)–C(1), N(2)–C(1) and N(3)–C(1), respectively. The C₂N₃ monolayer exhibits a typical p-type semiconductor behavior and good structural stability with Poisson's ratios of 0.39 and 0.27 in zigzag and armchair forms, respectively.

5.3 C₅N

Wu et al.⁷⁴ used the ab initio evolutionary algorithm method to calculate four kinds of thermodynamically stable C₅N phases (P62m (Fig. 1), R3m1̄, R3m2̄ and C2/m). To explore the stability of the new phases, they carried out simulations under a high pressure of 100 GPa and the mechanical properties and evaluated the hardness of these phases. Moreover, the electronic properties, energy bandgap and band structure of the C₅N were calculated. R3m1̄, R3m2̄ and C2/m phases are semiconducting with the bandgaps of 5, 5 and 2.15 eV, respectively, while P62m shows a metallic character.

5.4 C₅N₂

The C₅N₂ monolayer has a similar porous crystalline structure to the C₂N monolayer, except that the primitive cell of the C₅N₂ is constituted by five fused benzene-like rings (three fused rings for C₂N) bridged by pyrazine N (Fig. 16).⁷⁵ Accordingly, the primitive cell of C₅N₂ contains 30 C atoms and 12 N atoms (CN stoichiometric ratio of 5 : 2) and the optimized lattice parameter and the pore diameter are derived to be 16.59 Å and 11.52 Å, respectively, larger than those of C₂N.⁷⁵ Due to the nanoporosity of C₅N₂, it is also called holey C₅N₂ (h-C₅N₂). Theoretical calculations suggested that C₅N₂ is also a semiconductor and that the bandgap of a C₅N₂ monolayer is computed to be 1.10 eV, lower than that of a C₂N monolayer (1.72 eV).⁷⁶

Fig. 16 Schematic representation of the synthesis of C₂N and C₅N₂ and graphic representation of a one-layer-conjugated skeleton. Reprinted with permission from ref. 75 copyright 2019 Elsevier B.V.

Very recently, the successful preparation of 2D multilayered C₅N₂ was reported by Kim et al.⁷⁷ In their work, 1,2,4,5-benzenetetramine and hexaketocyclohexane were used as precursors. An efficient reversible imine condensation reaction was triggered by exposing the mixture under simulated sunlight irradiation for 3 h. Meanwhile, the electrical property measurement on the pellet of as-prepared 2D C₅N₂ crystals showed that 2D C₅N₂ crystals are semiconducting with a conductivity of 2.22 × 10⁻³ S m⁻¹. Furthermore, the 2D C₅N₂ film was observed to exhibit a rapid photo-response to white light irradiation, suggesting its potential in photocatalysis.

6. Summary and outlook

Over the last few years, 2D C_xN_y nanocrystals have attracted great interest arising from their graphene-like structure, open bandgap, large SSA, good chemical stability, and tunable surface chemistry, which enable plentiful applications in the areas of electronics, photocatalysis, electrocatalysis, energy storage devices, and gas adsorption/separation. In the 2D C_xN_y family, C₂N, C₃N and C₅N₂ have been successfully synthesized, while the other members including g-CN, c-C₂N₂, h-C₂N₂ and C₅N are only theoretically predicted thus far. Among them, C₂N has been mostly studied because of its suitable bandgap and unique nanoporous structure, which make it a potential photocatalyst, catalyst support for SACs and double-atom catalysts and membrane materials.⁷⁹ With the semiconducting electronic properties, face termination for electron localization, high SSA, 2D C_xN_y almost possess all the prerequisites as nonmetal heterogeneous catalysts. Additionally, the abundantly distributed Lewis basic functions and Brönsted basic functions originating from the electronegativity differences between N and C atoms in the C_xN_y framework can serve as catalytic sites for diverse reactions. In fact, g-C₃N₄ has already been reported to be an excellent metal-free catalyst for organic reactions and cyclisation of functional nitriles.⁸⁰ Thereby, 2D C_xN_y materials are also promising metal-free heterogeneous catalysts and relevant studies deserve to be conducted in future. Hence, a vast amount of efforts has been devoted to the applications of C₂N in photocatalytic water splitting, multiple electrocatalytic reactions involving the ORR, HER, CO₂RR, and NRR, as well as gas adsorption/separation. Compared to C₂N, C₃N possesses a narrower bandgap but better electronic conductivity, which makes C₃N promising for not only photocatalysis and electrocatalysis but also battery electrodes. As for g-CN, c-C₂N₂, h-C₂N₂, C₅N and C₅N₂, their possible applications are rather limited in the literature. A summary of the physicochemical properties, synthesis methods and potential applications of 2D C_xN_y is presented in Table 1. Albeit with the considerable progress on 2D CNs, there are still challenges that should be addressed in further work:

Table 1 Summary of the physicochemical properties, synthesis methods and potential applications of 2D C_xN_y^a

CxNy	Bandgap/eV	Hole size/Å	Synthesis methods	Potential applications	Ref.
a —: unknown.
g-C3N4	2.70	6.8	Thermal reactionsSolvothermal reactionsChemical vapor depositionSol–gel synthesisMicrowave heating	Water splittingCO2 reductionOrganic contaminant purificationFuel cells	15a and 78
g-CN	2.73	—	Solvothermal approach	Water splitting	41 and 72b
				Hydrogen storage	71
C2N	1.96	8.3	Wet-chemical reactionSolid phase methods	Water splitting	19, 23, 38 and 39–41
				CO oxidation	16b, 34, 39 and 43
				CO2RR	46 and 47
				NRR	48
				Gas adsorption/separation	49
				Fuel cells	13a
				FET	9a
C3N	2.67	Non-holey	Thermal reactionsHydrothermal method	Water splitting	57 and 58
				Fuel cells	59–61
				Gas adsorbent	62–66
c-C2N2	4.43	Non-holey	—	—	14
h-C2N2	3.70	Non-holey	—	—
C2N3	3.26	—	—	—	73
C5N (P62m)	0	Non-holey	—	—	74
C5N2	1.1	11.52	Reversible imine condensation reaction	Photocatalysis	76 and 77

---

(i) The currently used precursors for C_xN_y synthesis are relatively expensive and some of the available synthetic strategies are complex and require strict experimental conditions, which limit the large-scale production and practical applications. This is also the reason that a majority of the 2D C_xN_y compounds are theoretically investigated. Consequently, it is urgent to develop low-cost and easy-operation methods to fabricate 2D C_xN_y crystals.

(ii) Most of the obtained 2D C_xN_y crystals are in multilayered structure, while the layer number greatly affects the physicochemical properties of C_xN_y and thus their applications. Hence, preparation of single-layered and few-layered C_xN_y is of importance.

(iii) Surface modifications like creating defects, functionalizing with oxygenated groups, doping with heteroatoms, etc. are believed to have a vital influence on the electronic and catalytic properties of C_xN_y, while relevant experimental and theoretical studies are very limited so far. Therefore, rational surface engineering of 2D C_xN_y should also be paid attention for promoting their practical applications.

(iv) In light of the good gas adsorption/separation efficiency of 2D C_xN_y for small molecular gases including H₂, He, CO, and CO₂, adsorption of VOC molecules is more important. Especially, the nanopores in C₂N, g-CN and C₅N₂ may be efficient for VOC separation. Doping C_xN_y with specific elements or building a composite material can enhance their selective adsorption towards gas molecules. Therefore, further attempts on the potential applications of 2D C_xN_y in adsorption are very promising.

(v) Since the good performance of C₂N- and C₃N-based photocatalysts for water splitting and the wide bandgap of g-CN, c-C₂N₂, h-C₂N₂, C₅N and C₅N₂ have been theoretically illustrated, 2D C_xN_y materials are also expected to hold potential in photocatalytic degradation of both aqueous organic pollutants and VOCs for environmental remediation. Moreover, large SSA and planar structures of 2D C_xN_y can possibly facilitate more exposure of active sites and adsorptive photocatalysis, beneficial to the photocatalytic degradation processes.

In summary, research on various 2D C_xN_y structures and their potential applications in different fields is still in its infancy. The challenges and various opportunities are huge for the scientific community to explore. We wish that the information provided in this review can attract increasing interest in the area of 2D C_xN_y.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Li Tan and Chunyang Nie contributed equally to this work. This work was supported by the National Natural Science Foundation of China (No. 21777033 and 21806024), China Postdoctoral Science Foundation (No. 2017M622637), Science and Technology Planning Project of Guangdong Province (No. 2017B020216003), Innovation Team Project of Guangdong Provincial Department of Education (No. 2017KCXTD012), and Australian Research Council (DP170104264).

References

1. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666.
2. X. Du, I. Skachko, A. Barker and E. Y. Andrei, Nat. Nanotechnol., 2008, 3, 491.
3. (a) D. Sun, D. Ye, P. Liu, Y. Tang, J. Guo, L. Wang and H. Wang, Adv. Energy Mater., 2018, 8, 1702383; (b) J. Mao, J. Iocozzia, J. Huang, K. Meng, Y. Lai and Z. Lin, Energy Environ. Sci., 2018, 11, 772; (c) K.-Q. Lu, X. Xin, N. Zhang, Z.-R. Tang and Y.-J. Xu, J. Mater. Chem. A, 2018, 6, 4590.
4. K. C. Kemp, H. Seema, M. Saleh, N. H. Le, K. Mahesh, V. Chandra and K. S. Kim, Nanoscale, 2013, 5, 3149.
5. J. B. Wu, M. L. Lin, X. Cong, H. N. Liu and P. H. Tan, Chem. Soc. Rev., 2018, 47, 1822.
6. (a) M. Terrones, A. R. Botello-Méndez, J. Campos-Delgado, F. López-Urías, Y. I. Vega-Cantú, F. J. Rodríguez-Macías, A. L. Elías, E. Muñoz-Sandoval, A. G. Cano-Márquez and J.-C. Charlier, Nano Today, 2010, 5, 351; (b) A. Chuvilin, E. Bichoutskaia, M. C. Gimenez-Lopez, T. W. Chamberlain, G. A. Rance, N. Kuganathan, J. Biskupek, U. Kaiser and A. N. Khlobystov, Nat. Mater., 2011, 10, 687; (c) L. Jia, D.-H. Wang, Y. X. Huang, A. W. Xu and H. Q. Yu, J. Phys. Chem. C, 2011, 115, 11466.
7. (a) W. Dacheng, L. Yunqi, W. Yu, Z. Hongliang, H. Liping and Y. Gui, Nano Lett., 2009, 9, 1752; (b) Z. M. Ao and F. M. Peeters, J. Phys. Chem. C, 2010, 114, 14503.
8. H. Wang, T. Maiyalagan and X. Wang, ACS Catal., 2012, 2, 781.
9. (a) J. Mahmood, E. K. Lee, M. Jung, D. Shin, I. Y. Jeon, S. M. Jung, H. J. Choi, J. M. Seo, S. Y. Bae, S. D. Sohn, N. Park, J. H. Oh, H. J. Shin and J. B. Baek, Nat. Commun., 2015, 6, 6486; (b) R. Walczak, B. Kurpil, A. Savateev, T. Heil, J. Schmidt, Q. Qin, M. Antonietti and M. Oschatz, Angew. Chem., Int. Ed., 2018, 57, 10765; (c) J. Li, Y. Li, Z. Xiong, G. Yao and B. Lai, Chin. Chem. Lett., 2019, 30, 2139.
10. A. Y. Liu and M. L. Cohen, Science, 1989, 245, 841.
11. X. Wang, K. Maeda, X. Chen, K. Takanabe, K. Domen, Y. Hou, X. Fu and M. Antonietti, J. Am. Chem. Soc., 2009, 131, 1680.
12. (a) M. Z. Rahman, K. Davey and C. B. Mullins, Adv. Sci., 2018, 5, 1800820; (b) F. Yu, L. Wang, Q. Xing, D. Wang, X. Jiang, G. Li, A. Zheng, F. Ai and J. P. Zou, Chin. Chem. Lett., 2020, 31, 1648; (c) M. Wu, H. Lv, T. Wang, Z. Ao, H. Sun, C. Wang, T. An and S. Wang, Catal. Today, 2018, 315, 205.
13. (a) J. Xu, J. Mahmood, Y. Dou, S. Dou, F. Li, L. Dai and J. B. Baek, Adv. Mater., 2017, 29, 1702007; (b) S. Yang, W. Li, C. Ye, G. Wang, H. Tian, C. Zhu, P. He, G. Ding, X. Xie, Y. Liu, Y. Lifshitz, S. T. Lee, Z. Kang and M. Jiang, Adv. Mater., 2017, 29, 1605625.
14. Y. Dong, C. Zhang, M. Meng, M. M. Groves and J. Lin, Carbon, 2018, 138, 319.
15. (a) S. Ye, R. Wang, M.-Z. Wu and Y.-P. Yuan, Appl. Surf. Sci., 2015, 358, 15; (b) M. Wu, X. He, B. Jing, T. Wang, C. Wang, Y. Qin, Z. Ao, S. Wang and T. An, J. Hazard. Mater., 2020, 384, 121323.
16. (a) X. Li, L. Zhu, Q. Xue, X. Chang, C. Ling and W. Xing, ACS Appl. Mater. Interfaces, 2017, 9, 31161; (b) F. Li and Z. Chen, Nanoscale, 2018, 10, 15696.
17. (a) O. Y. Podyacheva, D. A. Bulushev, A. N. Suboch, D. A. Svintsitskiy, A. S. Lisitsyn, E. Modin, A. Chuvilin, E. Y. Gerasimov, V. I. Sobolev and V. N. Parmon, ChemSusChem, 2018, 11, 3724; (b) Y. Chen, S. Ji, Y. Wang, J. Dong, W. Chen, Z. Li, R. Shen, L. Zheng, Z. Zhuang, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2017, 56, 6937; (c) G. Liu, J. Zhou, W. Zhao, Z. Ao and T. An, Chin. Chem. Lett., 2020, 31, 1966.
18. T. Komatsu and M. Samejima, J. Mater. Chem., 1998, 8, 193.
19. S. S. Shinde, C. H. Lee, J. Y. Yu, D. H. Kim, S. U. Lee and J. H. Lee, ACS Nano, 2018, 12, 596.
20. S. Gong, W. Wan, S. Guan, B. Tai, C. Liu, B. Fu, S. A. Yang and Y. Yao, J. Mater. Chem. C, 2017, 5, 8424.
21. J. Du, C. Xia, W. Xiong, X. Zhao, T. Wang and Y. Jia, Phys. Chem. Chem. Phys., 2016, 18, 22678.
22. X. Xie, Z. Ao, D. Su, J. Zhang and G. Wang, Adv. Funct. Mater., 2015, 25, 1393.
23. J. Mahmood, S.-M. Jung, S. J. Kim, J. Park, J.-W. Yoo and J.-B. Baek, Chem. Mater., 2015, 27, 4860.
24. N. Fechler, N. P. Zussblatt, R. Rothe, R. Schlogl, M. G. Willinger, B. F. Chmelka and M. Antonietti, Adv. Mater., 2016, 28, 1287.
25. T. Jordan, M. Shalom, M. Antonietti and N. Fechler, Asia-Pac. J. Chem. Eng., 2016, 11, 866.
26. D. Li, X.-a. Ning, Y. Huang and S. Li, Electrochim. Acta, 2019, 312, 251.
27. X. Cui, W. An, X. Liu, H. Wang, Y. Men and J. Wang, Nanoscale, 2018, 10, 15262.
28. Y. Liu, Z. Meng, X. Guo, G. Xu, D. Rao, Y. Wang, K. Deng and R. Lu, Phys. Chem. Chem. Phys., 2017, 19, 28323.
29. J. Zhu, M. Zulfiqar, S. Zeng, Y. Zhao and J. Ni, AIP Adv., 2018, 8, 055333.
30. J. Du, C. Xia, T. Wang, W. Xiong and J. Li, J. Mater. Chem. C, 2016, 4, 9294.
31. Y. Bai, Q. Zhang, N. Xu, K. Deng and E. Kan, J. Phys. Chem. C, 2018, 122, 15892.
32. C. Zhang, Y. Jiao, T. He, S. Bottle, T. Frauenheim and A. Du, J. Phys. Chem. Lett., 2018, 9, 858.
33. H. Wang, X. Li and J. Yang, Chemphyschem, 2016, 17, 2100.
34. D. W. Ma, Q. Wang, X. Yan, X. Zhang, C. He, D. Zhou, Y. Tang, Z. Lu and Z. Yang, Carbon, 2016, 105, 463.
35. R. Zhang, B. Li and J. Yang, Nanoscale, 2015, 7, 14062.
36. R. Longuinhos and J. Ribeiro-Soares, Phys. Rev. B, 2018, 97, 195119.
37. R. Zhang and J. Yang, 2015, arXiv:1505.02768.
38. R. Zhang, L. Zhang, Q. Zheng, P. Gao, J. Zhao and J. Yang, J. Phys. Chem. Lett., 2018, 9, 5419.
39. M. Makaremi, S. Grixti, K. T. Butler, G. A. Ozin and C. V. Singh, ACS Appl. Mater. Interfaces, 2018, 10, 11143.
40. J. Mahmood, F. Li, S. M. Jung, M. S. Okyay, I. Ahmad, S. J. Kim, N. Park, H. Y. Jeong and J. B. Baek, Nat. Nanotechnol., 2017, 12, 441.
41. X. Li, P. Cui, W. Zhong, J. Li, X. Wang, Z. Wang and J. Jiang, Chem. Commun., 2016, 52, 13233.
42. X. Zhang, A. Chen, Z. Zhang, M. Jiao and Z. Zhou, J. Mater. Chem. A, 2018, 6, 11446.
43. (a) B. L. He, J. S. Shen and Z. X. Tian, Phys. Chem. Chem. Phys., 2016, 18, 24261; (b) S. Chakrabarty, T. Das, P. Banerjee, R. Thapa and G. P. Das, Appl. Surf. Sci., 2017, 418, 92.
44. (a) Q. Jiang, J. Zhang, Z. Ao, H. Huang, H. He and Y. Wu, Front. Chem., 2018, 6, 187; (b) C. Li, Y. Xu, W. Sheng, W. J. Yin, G. Z. Nie and Z. Ao, Phys. Chem. Chem. Phys., 2020, 22, 615.
45. Q. Jiang, J. Zhang, H. Huang, Y. Wu and Z. Ao, J. Mater. Chem. A, 2020, 8, 287.
46. J. Zhao, J. Zhao, F. Li and Z. Chen, J. Phys. Chem. C, 2018, 122, 19712.
47. Z. Wang, Z. Yu and J. Zhao, Phys. Chem. Chem. Phys., 2018, 20, 12835.
48. X. Zhang, A. Chen, Z. Zhang and Z. Zhou, J. Mater. Chem. A, 2018, 6, 18599.
49. (a) B. Xu, H. Xiang, Q. Wei, J. Q. Liu, Y. D. Xia, J. Yin and Z. G. Liu, Phys. Chem. Chem. Phys., 2015, 17, 15115; (b) A. Hashmi, M. U. Farooq, I. Khan, J. Son and J. Hong, J. Mater. Chem. A, 2017, 5, 2821; (c) Z. Hu, B. Liu, M. Dahanayaka, A. W. Law, J. Wei and K. Zhou, Phys. Chem. Chem. Phys., 2017, 19, 15973; (d) W. Ralf, K. Bogdan, S. Aleksandr, H. Tobias, S. Johannes, Q. Qing, A. Markus and O. M. J. A. Chemie, Angew. Chem., Int. Ed., 2018, 57, 10765; (e) Y. Su, Z. Ao, Y. Ji, G. Li and T. An, Appl. Surf. Sci., 2018, 450, 484; (f) L. Zhu, Q. Xue, X. Li, T. Wu, Y. Jin and W. Xing, J. Mater. Chem. A, 2015, 3, 21351.
50. Y. Su, W. Li, G. Li, Z. Ao and T. An, Chin. J. Catal., 2019, 40, 664.
51. (a) É. Sandré, C. J. Pickard and C. Colliex, Chem. Phys. Lett., 2000, 325, 53; (b) Q. Hu, Q. Wu, H. Wang, J. He and G. Zhang, Phys. Status Solidi B, 2012, 249, 784.
52. (a) G. Sun, Z. Y. Liu and J. L. He, Chin. Phys. Lett., 2007, 24, 1092; (b) L. L. Pang, J. Q. Bi, Y. J. Bai, Y. X. Qi, H. L. Zhu, C. G. Wang, J. W. Wu and C. W. Lu, Mater. Chem. Phys., 2008, 112, 1124.
53. J. Mahmood, E. K. Lee, M. Jung, D. Shin, H. J. Choi, J. M. Seo, S. M. Jung, D. Kim, F. Li, M. S. Lah, N. Park, H. J. Shin, J. H. Oh and J. B. Baek, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 7414.
54. S. Kumar, S. Sharma, V. Babar and U. Schwingenschlögl, J. Mater. Chem. A, 2017, 5, 20407.
55. T. C. King, P. D. Matthews, J. P. Holgado, D. A. Jefferson, R. M. Lambert, A. Alavi and D. S. Wright, Carbon, 2013, 64, 6.
56. (a) X. Zhou, W. Feng, S. Guan, B. Fu, W. Su and Y. Yao, J. Mater. Res., 2017, 32, 2993; (b) X. Wang, Q. Li, H. Wang, Y. Gao, J. Hou and J. Shao, Phys. B, 2018, 537, 314.
57. J. Wang, X. Li, Y. You, X. Yang, Y. Wang and Q. Li, Nanotechnology, 2018, 29, 365401.
58. X. Niu, X. Bai, Z. Zhou and J. Wang, ACS Catal., 2020, 10, 1976.
59. (a) W. Xu, C. Chen, C. Tang, Y. Li and L. Xu, Sci. Rep., 2018, 8, 5661; (b) H. Bingling, S. Jiansheng, M. Dongwei, L. Zhansheng and Y. Zongxian, J. Phys. Chem. C, 2018, 122, 20312; (c) Y. Zhou, G. Gao, J. Kang, W. Chu and L. W. Wang, J. Mater. Chem. A, 2019, 7, 12050.
60. Q. Liu, B. Xiao, J. B. Cheng, Y. C. Li, Q. Z. Li, W. Z. Li, X. F. Xu and X. F. Yu, ACS Appl. Mater. Interfaces, 2018, 10, 37135.
61. P. Bhauriyal, A. Mahata and B. Pathak, J. Phys. Chem. C, 2018, 122, 2481.
62. C. H. Yu, C. H. Huang and C. S. Tan, Aerosol Air Qual. Res., 2012, 12, 745.
63. D. Ma, J. Zhang, X. Li, C. He, Z. Lu, Z. Lu, Z. Yang and Y. Wang, Sens. Actuators, B, 2018, 266, 664.
64. I. Choudhuri, N. Patra, A. Mahata, R. Ahuja and B. Pathak, J. Phys. Chem. C, 2015, 119, 24827.
65. M. Pashangpour and A. A. Peyghan, J. Mol. Model., 2015, 21, 116.
66. (a) G. Qin, Q. Cui, W. Wang, P. Li, A. Du and Q. Sun, Chemphyschem, 2018, 19, 2788; (b) X. Li, T. Guo, L. Zhu, C. Ling, Q. Xue and W. Xing, Chem. Eng. J., 2018, 338, 92.
67. L. B. Shi, Y. Y. Zhang, X. M. Xiu and H. K. Dong, Carbon, 2018, 134, 103.
68. H. W. Kim, H. W. Yoon, S. M. Yoon, B. M. Yoo, B. K. Ahn, Y. H. Cho, H. J. Shin, H. Yang, U. Paik and S. J. S. Kwon, Science, 2013, 342, 91.
69. X. Ma, L. Li, Z. Zeng, R. Chen, C. Wang, K. Zhou, C. Su and H. Li, Chem. Eng. J., 2019, 363, 49.
70. J. Li, C. Cao, J. Hao, H. Qiu, Y. Xu and H. Zhu, Diamond Relat. Mater., 2006, 15, 1593.
71. Y. D. Chen, S. Yu, W. H. Zhao, S. F. Li and X. M. Duan, Phys. Chem. Chem. Phys., 2018, 20, 13473.
72. (a) K. Srinivasu and S. K. Ghosh, J. Mater. Chem. A, 2015, 3, 23011; (b) K. Srinivasu, B. Modak and S. K. Ghosh, J. Phys. Chem. C, 2014, 118, 26479.
73. L. B. Shi, S. Cao, J. Zhang, X. M. Xiu and H.-K. Dong, Phys. E, 2018, 103, 252.
74. Q. Wu, Q. Hu, Y. Hou, H. Wang, A. Zhou and L. Wang, J. Phys.: Condens. Matter, 2018, 30, 385402.
75. C. Yang, Z.-D. Yang, R. Zhang and G. Zhang, Chem. Phys., 2019, 517, 104.
76. Z. D. Yang, W. Wu and X. C. Zeng, J. Mater. Chem. C, 2014, 2, 2902.
77. S. Kim and H. C. Choi, Commun. Chem., 2019, 2, 60.
78. W. K. Darkwah and Y. Ao, Nanoscale Res. Lett., 2018, 13, 388.
79. W. Yao, Y. Zhang, H. Zhu, C. Ge and D. Wang, Chin. Chem. Lett., 2020, 31, 701.
80. T. Wang, C. Nie, Z. Ao, S. Wang and T. An, J. Mater. Chem. A, 2020, 8, 485.

---

Footnote

† These authors contributed equally to this work.

---