Carbon nitrides (CN) are polymeric materials with a wide range of applications derived from carbon materials by substituting N for the C atoms, making them attractive for a diversity of applications such as membranes, adsorbents, catalytic reactions, photocatalysis, sensors, supercapacitors, solar and fuel cells, hydrogen storage devices and biomedical [
1,
2,
3]. All these applications are dependent on the exceptional electronic, optical, and chemical properties of CN in combination with its synthesis from easily available precursors and its resistance to adverse physical and chemical conditions. Like most carbon materials, CN have an extended history, dating back to 1834, when a material called “
melon” (linear polymers of tri-s-triazines linked via secondary N) was described by Liebig in 1834 [
4], although the potential value of the material in question has only been completely accepted in recent decades. This is probably due to its high thermal (stable up to 600 °C in air) [
5], hydrothermal stability (insoluble either in acidic, neutral or basic solvents) [
1,
6] and its undisclosed structure. This allows the material to work in both gaseous and liquid environments and at higher temperatures, increasing its wide range of applications [
7]. CN materials have been intensively researched since Liu and Cohen [
8] predicted that they have the potential to be ultra-hard materials. Among various analogues, graphitic carbon nitride (g-C
3N
4) as an analog of graphite, a polymeric compound constructed via tri-s-triazine units n-type semiconductor, is considered the most stable environmental allotrope and a hot topic in materials science due to its extraordinary electronic structure. Contemporary activity has been stimulated by findings indicating that the “
graphitic” materials may have helpful properties for catalysis and energy (converting or storing), as well as other possible applications [
9]. Currently, g-C
3N
4 is making a difference in many areas of chemistry. Suffice it to say that the rate of publications on CN has increased exponentially from 2015 to date, as shown in
Figure 1.
1.1. Nomenclature and Synthesis
The most suitable nomenclature to designate different classes of chemically and physically produced CN materials is the first issue to be addressed [
10]. It is increasingly common to call them g-C
3N
4, which is usually defined as a group of compounds with a general formula close to C
3N
4 and structures constructed on heptazine units. However, this definition is misleading due to different reasons. For this reason, it is also valuable to develop an effective approach to the labeling of the different classes of CN compounds. This approach should accurately describe their chemical and structural properties as they relate to their functional performance. First, these C
3N
4 forms are best described as C
xN
yH
z compounds. This is because most of the materials that have been obtained to date contain not only C and N, but also significant amounts of hydrogen (H) as a part of their constitutions. On the other hand, materials made from linked heptazine (tri-s-triazine, C
6N
7) units are questionable to be fully condensed “graphitic” structures. As an alternative, they form zig-zag polymer chains related to those present in Liebig’s “
melon”, with a composition close C
2N
3H. In addition, other compounds with planar C
3N
4 layers are also formed by polytriazine imide (PTI) linked units providing hosts for embedded ions including Li
+, Cl
– and Br
– as well as extra H
+ [
11]. The presence of H reveals that the real g-C
3N
4 is not completely condensed and that there are several surface defects, which can be valuable in catalysis. Due to the presence of H, and because N has one electron more than C, g-C
3N
4 has a variety of surface properties essential for catalysis, including basic surface functionalities, electron-rich properties, H-bonding motifs, etc. [
12].
Although g-C
3N
4 is a two-dimensional covalent structure like graphene, the latter consists of pure C, while the former contains C and N. The semiconducting properties of g-C
3N
4 are significantly different from those of graphene sheets. g-C
3N
4 is a CN type, which has seven kinds of structures with different band gaps (
Eg): alpha (
Eg = 5.5 eV), beta (
Eg = 4.8 eV), cubic phase (
Eg = 4.3 eV), quasicubic (
Eg = 4.1 eV), g-h-triazine (
Eg = 3.0 eV), g-o-triazine (
Eg = 0.9 eV), and g-h-heptazine (
Eg = 2.9 eV) [
13,
14]. Like graphite, g-C
3N
4 also has a sheet structure containing C
3N
3 and C
6N
7 rings [
15]. The rings are connected by the N at the end to form an infinite plane. g-C
3N
4 is an economical and benign material that is chemically and physically very stable. It has a narrow bandgap (
Eg = 2.7 eV) agreeing to an optical wavelength (λ) of about 460 nm because its yellow color that is appropriate for visible light (λ > 400 nm) absorption and photocatalytic activity for environmental pollution remediation, among other applications [
16]. Density functional theory (DFT) calculations shows that the
Eg of melon is 2.6 eV, decreasing from 3.5 eV in melem (the simplest heptazine-based compound), and finally decreasing to 2.1 eV with the complete formation of condensed g-C
3N
4, since the reduction and oxidation levels are related to the positions of the valence (VB) and conduction (CB) bands [
6].
Bulk g-C
3N
4 (bg-C
3N
4) can be easily obtained at high temperature by direct thermal polymerization of the precursors. Reactive nitrogen-rich and oxygen-free compounds with pre-bonded C-N core structures such as thiourea, urea, melamine, cyanamide, dicyandiamide, guanidinium chloride, guanidine thiocyanate, ammonium thiocyanate, and/or hexamethylenetetramine are the most used precursors for the chemical synthesis of g-C
3N
4 [
3,
14,
17,
18,
19,
20]. Based on planar heptazine or triazine cores, these precursors produce two-dimensional layered graphitic structures. The condensation pathways from precursors such as urea and ethylenediamine to cyanamide and dicyandiamide, as well as melamine and all other C/N materials, are easy and suitable synthetic routes to produce slightly deformed polymeric species, as shown in
Figure 2.
The methods used to prepare and characterize bg-C
3N
4 are well established, although common bulk material defects such as little specific surface area and fewer active sites hinder more advanced improvement of bg-C
3N
4. It is for this reason that mesoporous materials (MM) have been a hot topic in the last few decades. MM are known for the pore channels (< 50 nm in size) inside them, which give them superior properties to the bulk materials for various applications. In general, mesoporosity has several outstanding characteristics, including ordered pore structure, high specific surface area, more active sites, low density and high adsorptive capacity [
21,
22,
23]. The transformation of g-C
3N
4 into mesoporous material is valuable for its catalytic activity, the increase of active sites, the improvement of photon operation rate, and the promotion of further research and extensive application of g-C
3N
4 materials. Therefore, a lot of effort has been devoted to the preparation and study of MM. These include several major methods used to produce mesoporous g-C
3N
4 (mg-C
3N
4), including hard- and soft-template methods, template-free method, sol-gel method, supramolecular preorganization method, exfoliation method and others [
3,
20,
24,
25]. These methods yield products with diverse pore regularity and other properties using different strategies and precursors.
1.2. Photocatalytic Properties and Applications
mg-C
3N
4 is an effective heterogeneous catalyst for many processes due to its distinctive electronic structure and the improvements of its mesoporous structure. In addition, the catalytic performance of mg-C
3N
4 can be further enhanced or its range of applications extended by in situ or post-modification. For example, when mg-C
3N
4 is used as a photocatalyst, appropriate changes can greatly improve its photocatalytic activity. To date, reducing the
Eg to gain light absorption in the VIS region or improving the detachment of photogenerated electrons (e
-) and holes (h
+) have been the most common modification routes to improve the performance of g-C
3N
4 as a photocatalyst to reduce their recombination, which can improve the quantum efficiency of photocatalysis. The changes can cause light absorption to undergo a red shift, which broadens the absorption capacity and results in improved utilization of sunlight. Currently, several methods of modification for mg-C
3N
4 exist. Non-metal doping, noble metal loading, metal oxide loading, dye photosensitization, and polyoxometalate immobilization are the most representative and typical modification methods [
16,
18,
26,
27,
28,
29,
30]. Generally, elemental doping is assumed as an effective strategy for performance enhancement through modification of its electronic structure and surface properties of g-C
3N
4 for effective photocatalyst. To improve the separation of photogenerated e
- and to inhibit the recombination of e
-/h
+ pairs, the heterojunction structure is always used. Combining simultaneously doping and heterojunction engineering to modify g-C
3N
4 with a great potential for efficient visible light photocatalysis is expected to be very useful. The enhancement of heterojunctions based on doped g-C
3N
4 has been recently the focus of many studies [
31].
Although TiO
2 is an excellent catalyst because of its stability, non-toxicity, wide band gap (
Eg = 3.2 eV) and charge recombination, it has low catalytic efficiency as photocatalyst because of the large band gap, which limits the use of a wide spectrum of solar light (λ < 388 nm), leading to much lower quantum efficiencies when using solar spectra [
32,
33]. As compared to TiO
2, g-C
3N
4 has a suitable medium-wide
Eg for efficient absorption of visible light. In addition, flexibility in improving the photocatalytic properties by doping with metal or non-metal ions to generate active sites with generous melon moieties, design of optimized heterojunctions, and morphological modification to increase the surface area are another important point to improve its photocatalytic performance. By changing the band gap, morphology, and separation of photogenerated e
- and h
+, efficiency can be improved, making g-C
3N
4 a very promising material for use as a photocatalyst for environmental remediation and other applications [
1,
34,
35].
Heterogeneous photocatalysis (HP), one of the best studied Advanced Oxidation Processes (AOPs), involves four main processes: i) light harvesting, ii) charge excitation, iii) charge separation and transfer, and iv) surface electrocatalytic reactions [
28,
36]. Briefly, HP is based on the semiconductor (SC) irradiation particles (in this case g-C
3N
4), commonly suspended in aqueous solutions (slurry), with a wavelength energy ≥
Eg. In the course of photocatalysis, g-C
3N
4 is wholly excited by the absorption of photons with energies (
hν) greater than its
Eg. This drives an e
− into the CB, leaving an h
+ in the VB and encouraging the migration of the e
- and h
+ to the particle surface (
Figure 3). The energy is dissipated as heat and the e
- and h
+ can recombine on the particle surface in a short time. The photoexcitation process of g-C
3N
4 can be summarized as follows (Eq. 1):
(1)
Relying on the fact that the lifetime of the photogenerated charges is longer than the recombination rate of them, there is still an evident amount of e
- and h
+ available. They can be trapped by defect sites in g-C
3N
4 or transferred to its surface where energy is lower than its CB, which helps separate and transfer carriers for successful surface redox reactions. On the other hand, the undesired recombination of most of the e
-/h
+ pairs cause tremendous energy loss through the release of heat or the emission of light. In general, for the reduction of surface-adsorbed molecules to generate radicals and for the direct oxidation of organic contaminants, the surface e
- and h
+ serve to donate and accept electrons, respectively [
30].
During surface reactions, the redox aptitude of h
+ and e
− essentially depends on the VB (-1.30 eV) and CB (1.40 eV) potentials of g-C
3N
4 [
24]. Hence, photogenerated e
- can readily react with adsorbed O
2 to produce non-selective radicals (O
2•-) owing to the less negative standard redox potential of O
2•-/O
2 (- 0.13 eV vs NHE) [
37]. However, due to the redox potential of HO
•/OH
- (+1.99 eV vs NHE) is much higher than the VB potential of g-C
3N
4, the remaining h
+ is not able to be scavenged by H
2O and OH
- to form HO
• in a kinetic manner [
38]. Similarly, a small amount of HO
• is generated by the reduction of adsorbed O
2 via multi-electron reaction processes [
16]. During photocatalytic process, O
2•- and h
+ have a critical role for g-C
3N
4 in the degradation of pesticides to CO
2, H
2O and other small molecule products [
39]. In addition, the production of reactive oxygen species (ROS) from adsorbed O
2 by the capture of e
- also helps to limit the recombination of e
-/h
+ pairs.
Note that surface reactions can only happen if the reduction/oxidation potentials are more positive or negative than the CB and VB values, respectively. Compared to TiO
2 and other semiconductor materials, g-C
3N
4 has the most negative CB level (-1.3 V vs. NHE at pH = 7) and a medium band gap (2.7 eV) [
40], which facilitates its large application in VIS-light photocatalysis. Thus, the
Eg of g-C
3N
4 can be further narrowed by simple doping, defects, and other possible sensitization actions to achieve a higher utilization of VIS-light. Considering its strong reduction ability, activity in VIS-light, abundance, easy fabrication, stacked 2D layered structure, non-toxicity and high stability, its direct use in the field of sustainable chemistry as a multi-functional heterogeneous metal-free photocatalyst is possible. However, g-C
3N
4 mostly shows low photocatalytic efficiency because some serious disadvantages of the material itself, such as high e
-/h
+ recombination rate, low surface area (∼10 m
2·g
-1), insufficient visible absorption below 460 nm, high degree of monomer condensation, moderate oxidation capacity, and small active sites for interfacial photoreactions, grain boundary effects, slow surface reaction kinetics, and low charge mobility interfering with electron delocalization [
37]. Although these prominent challenges severely limit the improvement of photocatalytic performance, they also provide more chances for future synthesis of more efficient g-C
3N
4-based photocatalysts.
1.3. Pesticides and Pharmaceuticals as Emerging Pollutants
Pesticides and pharmaceuticals are widely used throughout the world, with recognized benefits to human health and well-being. However, despite their necessity, these classes of chemicals are also associated with serious risks to human and environmental health and are common micropollutants of aquatic environments. In addition, knowledge of the risks of their transformation products in the aquatic environment is very important [
41].
Pesticides (
substances intended to repel, destroy or control pests and diseases and/or prevent undesirable plant growth) have the potential to pollute surface- and groundwater, among other environmental compartments, in addition to other emerging pollutants (EPs). They can reach surface waters through soil runoff and cause groundwater pollution by leaching through the soil profile [
42]. The occurrence of pesticide residues in surface- and groundwater is increasing in OECD countries, with a significant number of samples exceeding the legal limits. A European Environment Agency (EEA) report shows that 16% of EU surface waters are of unknown chemical status, 38% are of good chemical status, while 46% are not of good chemical status [
43]. Another EEA technical report focusing on 39 European countries assessed the incidence of pesticides and their main transformation products in surface- and groundwater, showing exceeding rates ranging from 5-15% for herbicides and 3-8% for insecticides in surface water, while in groundwater the percentages were 7% and < 1%, respectively with lower frequencies for fungicides in both surface- and groundwater [
44]. In addition, for recalcitrant pesticides, the polluted water (from agricultural, industrial and urban sources) treated by standard wastewater treatment plants (WWTPs) is in some cases inadequate to achieve regulatory purity requirements. This issue is particularly important where low levels of rainfall do not offer adequate water supplies to meet the needs of the agricultural sector, which requires increased reuse of effluent from WWTPs [
45]. Furthermore, particularly in developing countries where monitoring data is often lacking and chemical analyses are often not performed due to lack of facilities or financial constraints, further research efforts are needed to better understand the presence and impacts of pesticide use in water bodies [
46,
47].
On the other hand, pharmaceuticals (
compounds used in the treatment or prevention of human and animal diseases to restore, correct, or modify organic function), are often found in high concentrations in the aquatic environment all over the world [
48]. This is probably due to their continuous release from WWTPs, which is significantly faster than their removal rates. Pharmaceuticals, which are not readily biodegradable and may persist and remain toxic, have attracted considerable attention for their frequent detection in natural and wastewater bodies as well as drinking water [
49]. Consequently, pharmaceutical residues present current and potential risks to human and environmental health.
In this context, AOPs have achieved high interest in the last years, and their applications to remove pesticide and pharmaceutical residues from water have recently increased, especially solar heterogeneous photocatalysis [
50,
51,
52,
53,
54,
55,
56]. The main benefit of these technologies is that they eliminate or at least decrease pesticide residues by mineralizing rather than transferring, as happens in conventional treatments [
57,
58,
59].
1.4. Bibliometric Study
Bibliometric reviews are now very common and consist of the analysis of scientific publications using statistical methods to provide an overview and general structure of the target research area [
60]. In science, bibliometric studies serve to reveal the bibliometric structure that comprises the system of organization among the components of a field of knowledge and that helps to reveal its intellectual dimension, based on groupings of relevant topics in that field of research [
61]. In other words, bibliometric analyses provide a qualitative and quantitative approach to a particular field of research. Thus, it uses bibliometric data about the research field, such as total number of publications, authors, citations, institutions and countries, to build a complete picture of the research area [
62]. In addition, bibliometric approaches are increasingly central to literature reviews in many fields of knowledge, as they help to analyze and visualize the status, structure, hotspots and future research trends in each discipline. Indeed, the large number of bibliometric reviews published in different scientific disciplines is striking, as a search for the keyword “
bibliometric analysis” in the title field yields almost 15,500 publications using the Web of Science (WoS
®) database.
Therefore, the aim of this work was to provide accurate and up-to-date bibliometric information on the development and future prospects of the use of g-C3N4 as a heterogeneous photocatalyst in applications for the remediation of water contaminated with pesticide and pharmaceutical residues reported in the literature to date, since, to the best of the authors’ knowledge, there is no bibliometric review of its use and could help researchers, especially those new to the field, to identify possible future lines of research by summarizing the main aspects already covered and identifying critical points of interest.