Alkali metal-doped g-C₃N₄: a multifunctional photocatalytic platform for solar-induced energy conversion and environmental restoration

Abstract

Alkali metal doping has emerged as a powerful strategy to unlock the full photocatalytic potential of graphitic carbon nitride (g-C₃N₄) by modulating the structural and electronic parameters. Notably, incorporating alkali metal ions into the crystal lattice of g-C₃N₄ serves as an effective strategy to modulate its crystalline structure, enhance photon-harvesting capabilities, improve surface area, and optimize photocatalytic performance, all while retaining environmental compatibility. This review summarizes recent developments in alkali metal doping of g-C₃N₄, focusing on design strategies, structural modifications, and the resulting alterations in electronic and physicochemical properties. DFT calculations, theoretical modelling, and various characterization techniques were systematically reviewed for structural elucidation and optical modification after alkali metal ion doping in g-C₃N₄. We also provide concrete case studies by reviewing the structure–activity relationships of various alkali metal-doped g-C₃N₄-based photocatalysts having potential applications in artificial photosynthesis, ranging from solar fuel production, biomass conversion, organic transformation, and pollutant degradation to other miscellaneous applications, mainly supported by the in-depth photocatalytic mechanism. Finally, we conclude by highlighting the research gap and future perspectives. We believe this review is intended to serve as a comprehensive reference for both academic researchers and industrial practitioners engaged in the rational design of alkali-doped g-C₃N₄-based photocatalytic systems.

---

Pratikshya Dash

Pratikshya Dash completed her MSc degree in Chemistry at Sambalpur University, Sambalpur, Odisha, India. She qualified GATE 2021. She is currently pursuing her PhD under the supervision of Prof. Kulamani Parida at the Centre for Nano Science and Nano Technology (CNSNT), Institute of Technical Education and Research (ITER), Siksha ‘O’ Anusandhan (Deemed to be University). Her current research interests are focused on the development of mesoporous graphitic carbon nitride-based nanomaterials for solar energy conversion applications.

Sulagna Patnaik

Dr Sulagna Patnaik completed her master's degree in chemistry at NIT, Rourkela. Then, she joined as a lecturer in chemistry at Rayagada Autonomous College, Odisha. She is presently working as a reader at the Maharshi College of Natural Science, Odisha, and completed her PhD at Siksha 'O' Anusandhan University, Bhubaneswar, Odisha. Her research work is focused on the development of modified g-C3N4-based photocatalysts for hydrogen energy production and pollution abatement.

Sonali Panda

Sonali Panda holds an MSc degree in Chemistry from Utkal University, Bhubaneswar and is a University Gold Medalist. She has qualified with both the GATE and CSIR-UGC NET. Currently, she is a research scholar under the supervision of Prof. Kulamani Parida at the Centre for Nano Science and Nano Technology (CNSNT), Institute of Technical Education and Research (ITER), Siksha ‘O’ Anusandhan (Deemed to be University). Her research focuses on the design and development of multi-metallic MOFs for sustainable energy conversion applications.

Bhagyashree Priyadarshini Mishra

Dr Bhagyashree Priyadarshini Mishra received her MSc degree from Ravenshaw University, Odisha, India. She completed her PhD at Siksha 'O' Anusandhan University, Odisha, where she worked in the field of photocatalytic hydrogen evolution and H2O2 production. She joined the research group of Prof. Venkata Krishnan in 2024 as a postdoctoral researcher under the National Postdoctoral Fellowship (NPDF) 2024. Her current research focuses on the development of effective photocatalysts for nitrogen fixation applications.

Kulamani Parida

Prof. Kulamani Parida is a globally distinguished scientist in the areas of materials science, catalysis and nanotechnology. He superannuated as a Chief Scientist and a Professor of Academy of Scientific and Innovative Research (AcSIR) at CSIR-IMMT, Bhubaneswar in 2014. His adoration for chemistry and passion for research couldn't let Prof. Parida to sit idle after retirement. With the same devotion and dedication, Prof. Parida (Director) started the ‘Centre for Nano Science and Nano Technology’, a research unit of Siksha ‘O’ Anusandhan (deemed to be university), Bhubaneswar. His research interests include the design and development of advanced materials encompassing fine chemicals, energy generation/storage, and environmental applications. He has published more than 550 research articles and review articles in renowned international journals. He has a citation count of 39593 to his credit with an h-index of 108 and i10-index of 489. For more details: https://kmparidaimmt.weebly.com/.

---

Broader context

The growing global energy demand and concerns over future energy security necessitate the transition toward sustainable alternatives such as semiconductor photocatalysis. This emerging approach enables solar-to-sustainability by harnessing artificial photosynthesis for next generation clean fuel production and environmental detoxification. Thus, the development of efficient, robust, and economically viable semiconductor photocatalysts derived from earth-abundant precursors is of paramount importance. In this review, we present a systematic evaluation of the emergence of alkali doping in g-C₃N₄, which has attracted considerable attention owing to its low cost, non-toxicity and sustainable and straight-forward synthesis approach. We critically examine and provide a comprehensive assessment of the incorporation of various alkali metal ions such as, Li⁺, Na⁺, K⁺ etc., into the interlayer cavities or lattice sites of the g-C₃N₄ framework, which induces significant structural, electronic, and optical modulation to accelerate charge carrier separation for efficient solar fuel production and environmental detoxification. Besides, this review also highlights the existing research gaps and future prospects in the advancement of semiconductor photocatalysis, ultimately contributing to the development of sustainable energy technologies for energy and environment security. We believe this review systematically addresses the synthesis–structure–activity correlations and can be an effective guiding tool for dopant selection in the fabrication of g-C₃N₄-based next-generation photocatalysts.

1. Introduction

The global energy demand is anticipated to increase exponentially, as a consequence of rapid advances in industrialization and population expansion, with projections indicating a 2-fold rise by 2050 relative to the present consumption scenario.^1–3 Concurrently, CO₂ emissions due to the overexploitation of fossil fuels pose a serious threat to global warming and need immediate intervention. Achieving carbon neutrality depends on promoting eco-sustainability and advancing renewable energy technologies. In the contemporary era, achieving sustainable development necessitates the adoption of clean and renewable energy technologies to mitigate escalating environmental challenges. Among diverse energy alternatives, solar energy has become a leading contender owing to its inexhaustibility and renewable nature for meeting future global energy needs.^4–6 Due to its potential as an environmentally sustainable solution for global energy challenges, solar energy production has attracted considerable research attention.^7,8 Solar fuel production, a critical aspect of solar energy utilization, focuses on converting abundant feedstocks into valuable chemical fuels. More broadly, the utilization of solar energy towards various artificial photosynthesis applications represents a broader framework by harnessing the renewable precursors to drive essential photochemical reactions. This encompasses solar fuel production (mainly emphasizing processes like H₂ production, CO₂ reduction, H₂O₂ synthesis, and N₂ fixation), alongside other ancillary applications such as organic and inorganic pollutant degradation, biomass conversion, organic transformations, and NO_x removal.

Semiconductor-based heterogeneous photocatalysis has garnered widespread research interest in the current era due to its potential to harness solar energy efficiently.^9–11 Considerable efforts have been made by the researchers to design advanced semiconductor-based photocatalysts to enhance their potential applications.^12–14 An ideal semiconductor photocatalyst must possess (i) a strong absorption of light across a broad spectrum, (ii) abundant surface-active sites to facilitate reactions, (iii) suppressed recombination of charge carriers to prolong their lifetime, (iv) appropriate CB and VB potentials to drive redox reactions, and (v) low-cost fabrication along with long-term photostability.^15,16 To date, a good number of semiconductor-based photocatalytic materials have emerged to address energy issues and environmental threats.^17–20 Amongst these photocatalysts, g-C₃N₄ stands out as a prominent choice and has been extensively studied due to its advantageous features such as cost-effectiveness, metal-free, easy and facile fabrication, robustness, visible light responsiveness, and suitable band edge potential for surface redox reactions.^21,22 Despite meeting several essential criteria for an efficient photocatalyst, pristine g-C₃N₄ often suffers from intrinsic drawbacks such as sluggish charge separation, limited surface-active sites, low specific surface area, poor electrical conductivity, and modest redox potential, which restrict its performance in photocatalytic and electrocatalytic processes. As a result, rational modification strategies have become central to unlocking its potential in sustainable energy conversion and environmental remediation. Among the various modification pathways, such as elemental doping, defect engineering, heterojunction construction, surface functionalization, nano-structuring by increasing porosity and surface area, non-metal/carbon doping, vacancy engineering, and single-atom loading, alkali-metal doping (Li⁺, Na⁺, K⁺, Rb⁺, and Cs⁺) has gained particular attention.^23–25 In this regard, we have recently reported a review article highlighting that the strategic nano-structuring can expose the isolated active sites for better photocatalytic activity. Moreover, the structural modification strategies, such as elemental doping, further aid the promotional impact by optimizing the microstructure and optoelectronic properties.^26,27 Doping serves as an effective strategy to introduce foreign elements into the g-C₃N₄ lattice, enabling modulation of its surface and electronic features. The introduction of metallic or non-metallic dopants alters the band gap energy and tunes the band edge potential, thereby broadening the light absorption, enhancing photocurrent generation, and modifying photo-reactivity.^28–30 Due to its adaptability for flexible in situ doping approaches and favorable biocompatibility, ion doping has been extensively studied to engineer g-C₃N₄-based heterostructures.^31–33 In the case of anionic doping, the introduced species typically occupy either substitutionally or interstitially, which can modulate both the textural and electronic structures. In contrast, cationic dopants induce orbital hybridization between the dopant's atomic orbital and the g-C₃N₄ molecular orbital, which leads to achieving the optimized VB and CB positions.^34–37 Cationic dopants can further modify the electronic properties by band gap narrowing, improving light harvesting efficiency and accelerating the charge carriers’ mobility. Common cationic dopants include alkali, transition, and noble metals, which impart distinctive physicochemical functionalities and strengthen surface interactions in the doped system. This doping is readily achieved through thermal polymerization or condensation, wherein the g-C₃N₄ precursors are simply mixed with the appropriate metal salts, either in the solid-state mixture or in solution-based systems, to obtain uniformly doped systems.^31–37

Alkali metal cations interact strongly with the heptazine nitrogen framework, creating local electronic perturbations that are difficult to achieve through heavier dopants or structural modifications alone. Their incorporation can widen interlayer spacing, enhance π-electron delocalization, and promote charge polarization, collectively leading to improved charge separation and more efficient migration of photo-generated carriers. Moreover, alkali-metal doping frequently induces subtle structural distortions and introduces shallow donor states, which modulate the band structure without significantly compromising chemical integrity. These effects can shift the conduction and valence band positions to better match key redox reactions relevant to water splitting, CO₂ reduction, nitrogen fixation, pollutant oxidation, and antibacterial disinfection. In some cases, small cations (e.g., Li⁺) also facilitate proton or ion transport at the catalyst interface, which is beneficial for hydrogen-evolution and proton-coupled reactions. The presence of alkali metal ions can also tune surface basicity and adsorption behaviour, thereby facilitating the activation of molecules such as O₂, H₂O, and CO₂, while the modulated band potentials promote the formation of reactive intermediates during the photocatalytic processes.^26,38

Alkali metal doping offers several advantages, such as environmental benignity and low toxicity. Incorporating these ions introduces structural vacancies or defects within the lattice, which strategically alter electronic properties. The specific choice of dopant and its concentration further influences charge distribution dynamics, mitigates photo-corrosion, and governs overall catalytic performance. Furthermore, alkali-doped g-C₃N₄-based intercalated systems induce non-uniform charge distribution, thereby suppressing electron–hole recombination and amplifying charge separation efficiency. They are also extensively recognised as effective electronic promoters in various catalytic applications due to their exclusive properties.^39–44

Collectively, these structural and electronic modifications explain why alkali-treated g-C₃N₄ often shows a higher photocatalytic activity than the pristine material. For instance, Sun et al. demonstrated a facile thermal polymerization method to fabricate K⁺ doped g-C₃N₄ for efficient rhodamine B degradation, which is attributed to the small grain size, dispersed active sites, intensified optical response in the visible spectrum, and lowered band gap due to the K⁺ doping.⁴² Similarly, a systematic investigation of various alkali metal ion (Li⁺, Na⁺, or K⁺) doped g-C₃N₄ systems revealed that they exhibited stronger absorption in the visible region, larger surface area for exposed active sites, and reduced band gap, which cumulatively optimized the microstructure and optoelectronic features for efficient hydrogen production.⁴⁵ Haripriya et al. demonstrated that the appropriate alkali metal (Li, Na, and K) doped g-C₃N₄ can control the defect densities, achieving an optimized band structure for efficient photocatalytic H₂O₂ production.⁴⁶ Moreover, the co-doping of alkali metals with other photoactive species can also lead to promotional activity in photocatalysis. For instance, Liang et al. developed Bi and K co-doped g-C₃N₄, which demonstrated enhanced visible light harvesting capability and improved band alignment by shifting the conduction band towards more negative energy levels, facilitating the photocatalytic oxygen reduction reaction.⁴⁷ Moreover, the co-doping reduces the energy barrier for redox reactions by improving the charge transfer process. Alkali metal-containing g-C₃N₄ systems can also form effective composite photocatalysts with other photoactive species and regulate the crystallinity, microstructure, and charge dynamics in the composite photocatalyst. Also, dual and multi-functional behaviour can be achieved, which is beneficial for the complex photocatalytic reactions.^24,38,48 The effective interfacial contact in the alkali-doped g-C₃N₄-based composite can significantly boost the charge transfer process to enable efficient redox reaction by complete utilization of maximum charge carriers.^49,50 The synthesis of alkali metal-doped catalysts neither requires post-synthesis purification steps nor relies on hazardous chemicals; therefore, it is regarded as an inherently sustainable process. The impact of alkali metal ion doping on the structural, morphological, and electronic properties in the case of other photocatalytic materials, such as titanates, is well established in the literature. Bem et al. demonstrated that the substitution of H⁺ ions by Na⁺ results in contraction of the inter-layer distance of titanate nanotubes.⁵¹ Furthermore, Qamar and co-workers elucidated the stabilizing effect of studied Na⁺ ions on the titanate framework.⁵² Even doping of sodium ions has also been reported to drive a structural phase transformation from layered titanate to anatase.⁵³

In the literature, several review articles mainly focus on the prevailing aspects of synthesis methodologies, structural modifications, and various photocatalytic applications of doped g-C₃N₄ systems. For instance, a recent review article written by Khan et al. focused on the recent advancements in metal and non-metal doped g-C₃N₄, comprising the fundamentals, photocatalytic mechanisms, and factors affecting photocatalysis due to the doping and co-doping.²⁶ Similarly, another recent review article also demonstrated the effect of s-block, p-block, and transition metal dopants on the crystal structure of g-C₃N₄ to achieve the optimal band alignment for better light harvesting phenomena.²⁹ Our group also demonstrated the role of anion doping in modulating the band edge potential and simultaneously highlighted the various effective synthesis strategies, nature of dopants, and doping sites in g-C₃N₄ for energy- and environment-related applications.⁵⁴ Similarly, a few more reviews have been reported in the literature, mostly focused on the overall doping phenomena in the g-C₃N₄ structure.^26,28,55,56 Nonetheless, systematic and comprehensive studies that explicitly rationalize the effect of alkali metal doping on the photocatalytic activity of the g-C₃N₄ crystal lattice still remain underexplored to date. The advent of alkali metal-doped g-C₃N₄ has revolutionized the field of g-C₃N₄ structural modifications by imparting novel physicochemical properties. Therefore, this review systematically summarizes the comprehensive assessment of how alkali metals influence the physicochemical, optoelectronic, and structural attributes of g-C₃N₄ photocatalysts for various efficient artificial photocatalytic applications. We further discuss the criteria for selecting appropriate alkali metals to rationally design high-performance alkali-doped g-C₃N₄ photocatalysts. The performance of these alkali-doped g-C₃N₄ systems across diverse artificial photosynthesis applications, such as solar fuel production, biomass conversion, organic transformation, pollutant degradation, and other miscellaneous applications, has been systematically reviewed using various characterization techniques, DFT calculations, and in-depth mechanistic analyses. Fig. 1 presents various alkali metal dopants incorporated in the g-C₃N₄ lattice matrix, developed via an economical, sustainable, and facile approach leading to enhanced structural and optical modulation, improved photostability, and superior redox reactivity for diverse energy harvesting and environmental remediation applications. Overall, this review aims to provide a key fundamental reference for researchers who are focused on advancing alkali metal-doped g-C₃N₄ materials for efficient and sustainable solar energy conversion.

Fig. 1 Schematic illustration of alkali-doped g-C₃N₄ highlighted with characteristic features towards energy conversion and environmental remediation.

2. Alkali metal-doped g-C₃N₄

Alkali metal-ion doping is a facile, non-interfering strategy that upgrades nearly any g-C₃N₄ modification approach and can significantly resolve the limitations of the bulk g-C₃N₄ (Fig. 6). It provides a unique combination of cost-effectiveness, ease of synthesis, enhanced interlayer spacing, precise electronic tuning, and internal-field-driven charge separation. It is completely distinguishable from more complex and expensive methods like heterojunction construction, non-metal doping, or incorporation of a single-atom catalyst (Table 1).

Table 1 Comparative assessment of alkali doping over heterojunction construction, non-metal doping, and single-atom catalysis

Criteria	Alkali-metal doping	Heterojunction	Non-metal doping	Single-atom catalysts
Synthesis complexity	Very simple, (impregnation/thermal treatment)	High (requires precise interface engineering)	Moderate	Very high (atom trapping, complex supports)
Cost effectiveness	Low-cost alkali salts; inexpensive processing	High (multiple materials + step-wise fabrication)	Low-moderate	Very high (noble metals, advanced synthesis)
Electronic structure tuning	Strong modulation of charge density via ionic/electrostatic effects	Strong but interface-dependent	Moderate; depends on substitution sites	Excellent, but limited to isolated atoms
Defect/active site creation	Easily generates oxygen vacancies, cyano defect and surface basicity	Depends on junction quality	Creates substitutional defects	Creates highly active isolated sites but fewer in number
Scalability	Highly scalable industrially	Limited scalability; complex architectures	Good	Challenging, delicate synthesis
Stability under reaction conditions	Good thermal/chemical stability	Interface may degrade	Generally stable	Stability of isolated atoms is often problematic (migration/sintering)
Environmental concerns	Minimal	Depends on materials used	Minimal	Possible issues with metal leaching

---

Various alkali metal ions (Li⁺, Na⁺, K⁺, etc.) preferably intercalate into the nitrogen pot of the g-C₃N₄ framework, cause a distinct modification in the electronic band structure, create defects and vacancies, and improve the crystallinity for efficient photocatalysis. The charge redistribution in the interlayer of pristine g-C₃N₄ is absolutely necessary to achieve improved photocatalysis, which can be carried out by intercalating the alkali metal ions into the intrinsic graphitic sp²-hybridized network composed of tri-s-triazine units.^39,57,58 For example, Li et al. unraveled the crucial role of the interlayer electron transfer mechanism for alkali metals (K, Rb, and Cs) intercalated in carbon nitride (CN).⁵⁹ The doping of alkali metals could inhibit random charge transfer across the CN planes in order to enable the directional flow of electrons (e⁻) between the adjacent layers. Consequently, the alkali metal intercalation facilitates the charge redistribution within the crystal lattice, thereby effectively enhancing the charge carrier separation to execute efficient redox reaction. Nguyen and co-workers demonstrated that incorporating single alkali metal dopants into g-C₃N₄ can reduce its vertical ionization potential and increase the electron affinity and Lewis acidity. Furthermore, the alkali-doped g-C₃N₄ systems possessed a lower band gap relative to the undoped g-C₃N₄, which is due to the electron donation from the metal atom to g-C₃N₄.⁶⁰ We first discuss the theoretical modelling and DFT calculations, which elucidate the preferred doping sites of alkali metals within the g-C₃N₄ lattice and the resulting changes in electronic band structure. Additionally, we systematically reviewed the design strategies to incorporate alkali metals and the characterization techniques used to probe the induced structural modifications. Moreover, the influence of doping on the electronic properties is also comprehensively reviewed. The doping can create localized charge densities to promote the superior charge carrier mobility; thus, the CB and VB levels adjust and realign for efficient photo redox reaction. The appropriate alkali doping also tends to increase the surface basicity of the g-C₃N₄, promoting the easy activation and adsorption of the organic molecules/reactants for the efficient photocatalytic reactions. Furthermore, the beneficial aspects of co-doping can further induce the defect-mediated active sites by promoting N vacancies in g-C₃N₄.^61–63 Overall, alkali metal doping can introduce additional electronic states, reduce the band gap energy, and induce a redshift in the absorption edge, thereby enhancing light-harvesting efficiency in the visible-light region. Therefore, this section delves into a more detailed analysis of the crystal and lattice parameters of alkali doped g-C₃N₄ in terms of DFT and theoretical investigations, reviews the promising and economic design strategies, highlights the optoarchitecture behavior of g-C₃N₄, and the effect of co-doping (Fig. 2).

Fig. 2 Schematic illustration of the effect of alkali-doping in the g-C₃N₄ matrix.

2.1. DFT calculations and theoretical modelling to unveil the effect of alkali metal doping within the g-C₃N₄ matrix

As discussed above, the incorporation of alkali metals in the g-C₃N₄ framework induces notable variations in its micro-electronic structure, which was analyzed using the plane-wave ultrasoft pseudopotential simulations. The energy states of the configurations were estimated by the generalized gradient approximation (GGA) and local density approximation (LDA) methods.⁶⁴ The alkali metal ions are preferably incorporated in the large cavity, designated as the F position over other possible binding sites designated as A, B, C, D and E in g-C₃N₄ layers (Fig. 3a). Furthermore, the DFT calculations estimated the formation energy, charge transfer, bandgap energy and related work function values, which suggested the creation of the n-type Schottky junctions at the interface of g-C₃N₄ and the alkali metals (Table 2). Moreover, the band edge potential and the corresponding band structure were significantly altered due to the redistribution of HOMO and LUMO energy levels after the alkali metal doping.

Fig. 3 (a) Top views of the six doped configurations of alkali metals adsorbed on g-C₃N₄. Reproduced with permission from ref. 64. Copyright 2016, Elsevier. Crystal structures of CN (b), CN-K2 (c), and CN-Na2 (d). (e) and (f) Top views of CN-K2 and CN-Na2. Reproduced with permission from ref. 39. Copyright 2016, American Chemical Society.

Table 2 Calculated formation energy, charge transfer, theoretical band gap, and work functions for different alkali metal-doped g-C₃N₄ photocatalysts

Dopant	Formation energy (eV)	Charge transfer Δq (e)	Band gap (eV)	Work function (eV)
Li–g-C3N4	−4.477	+0.09	0.74 (GGA-PW91)	5.94
			0.59 (LDA-CAPZ)
Na–g-C3N4	−4.144	+0.75	0.78 (GGA-PW91)	5.57
			0.61 (LDA-CAPZ)
K–g-C3N4	−3.603	+1.26	0.82 (GGA-PW91)	5.32
			0.65 (LDA-CAPZ)

---

Alkali metals preferably adopt interstitial doping rather than substitutional doping, due to their comparatively larger ionic radii relative to C and N atoms. For example, Xiong et al. presented a DFT calculation using the “Vienna ab initio simulation package” (VASP5.3), and the most stable configuration of K and Na-doped g-C₃N₄ (CN-K2 and CN-Na2, respectively) was determined using the PBE exchange–correlation functional, used within the spin-polarized GGA method.³⁹ Despite the analogous chemical nature of K and Na atoms, they impart distinct charge distribution within the CN framework. Moreover, the interplanar spacing in CN is sufficient to host alkali metal atoms, enabling their incorporation via ion–dipole interaction. DFT calculations suggested that the Na-atom has an almost negligible effect on the crystal structure due to its small atomic radius (1.86 Å) and confined within the planar cavity of CN; however, K-atoms with a relatively larger atomic radius (2.32 Å) are intercalated within the interlayer of CN, forming a bridging connection that enhances the interlayer charge transport efficiency. Moreover, the distance of the nitride cavities oriented at the centre of the nearest heptazine moieties and the interlayer spacing of CN were found to be 7.11 and 3.14 Å, respectively. Therefore, the most stable optimized crystal structure of CN is that obtained by the interstitial doping of Na in CN-Na2 and the interlayer insertion of K in CN-K2, as shown in Fig. 3b–f.

Thereafter, Nnadiekwe et al. demonstrated an in-depth theoretical investigation of the band structure of alkali-doped g-C₃N₄ systems.⁶⁵ Optimization of the bulk g-C₃N₄ and the alkali metal ion-doped g-C₃N₄ (Li⁺–g-C₃N₄, Na⁺–g-C₃N₄, and K⁺–g-C₃N₄) sheets was carried out using the B3LYP hybrid functional. Furthermore, structural geometries and electronic interactions between g-C₃N₄ and the alkali metal ions were optimized using Pople's split valence 6-31G(d) and the LANL2DZ basis sets, respectively. It is obvious that the alkali metal cation, when doped in g-C₃N₄, can lead to a change in the structural properties and experience a change in the chemical and electronic features. The optimized geometries of the pristine g-C₃N₄, M⁺(Li⁺, Na⁺, and K⁺)–g-C₃N₄ are presented in Fig. 4a–d. It is understood that the unit cell having central voids is appropriate to host the alkali metal ion. Various structural properties, such as the bond lengths and bond angles in the N–M⁺–N vicinity of alkali metal ion-doped g-C₃N₄, were estimated to comprehend the types of interaction between the alkali metal ion and electron-rich nitrogen atoms. However, the calculation results revealed that the M⁺–N distance increased in the order Li⁺–g-C₃N₄ < Na⁺–g-C₃N₄ < K⁺–g-C₃N₄, but the N–M⁺–N angles of Li⁺–g-C₃N₄ and Na⁺–g-C₃N₄ complexes did not differ remarkably. In contrast, K⁺–g-C₃N₄ exhibited different characteristics with a longer M⁺–N bond, which caused a noticeable distortion in the planarity of g-C₃N₄ sheets. The electronic configuration was analyzed by calculating and optimizing the frontier orbital arrangements of the HOMO and the LUMO. It was reported that the K⁺ ion could slightly delocalize the LUMO due to the different structural properties. In contrast, the electron-deficient nature of Li⁺–g-C₃N₄ can be understood from the charge distribution maps, suggesting the lowest HOMO–LUMO energy gap with maximum charge potential, compared to Na⁺–g-C₃N₄ and K⁺–g-C₃N₄, respectively (Fig. 4e–j).

Fig. 4 Illustration of the optimized configuration of (a) g-C₃N₄, (b) Li⁺–g-C₃N₄, (c) Na⁺–g-C₃N₄, and (d) K⁺–g-C₃N₄, calculated frontier molecular orbitals and electrostatic charge surface distribution of Li⁺–g-C₃N₄ (e) and (h), Na⁺–g-C₃N₄ (f) and (i), and K⁺–g-C₃N₄ (g) and (j). Reproduced with permission from ref. 65. Copyright 2023, Elsevier.

Zhang et al. demonstrated an in-depth DFT calculation to estimate the electronic structure of bulk g-C₃N₄ (CN) and K⁺ doped g-C₃N₄ (KCN).⁶⁶ Band structures, density of states, work functions, and Fermi levels of CN and KCN were determined by the CASTEP module analysis method using the Materials Studio software. The structure optimization was achieved by generalized gradient approximation (GGA) and the Perdew–Burke–Ernzerhof (PBE) functional analysis. An optimized structure of CN and KCN is demonstrated in Fig. 5a, suggesting the intercalation of K⁺ in the heptazine cavity of the cyano-bonded CN structure. Furthermore, the total and partial density of states (TDOS and PDOS) of CN and KCN are presented in Fig. 5b and c, suggesting that the VB maxima of CN and KCN are predominantly contributed by N 2p orbitals, whereas the CB minima are constructed by the contribution of C 2p and N 2p orbitals. However, the K doping in KCN modified the CB minima, leading to a noticeable variation in the band gap. Furthermore, the work function and Fermi level of CN and KCN were estimated by GGA-PBE functional simulation. The lower wavefunction and higher Fermi level value of KCN relative to CN suggested that the intercalation of K⁺ facilitates the formation of abundant photoexcited charge carriers for better photocatalytic reaction (Fig. 5d and e). Zhu et al. synthesized triazine-based Li-doped g-C₃N₄ derived via the thermal condensation of melamine.⁶⁷ Owing to the relatively small dimension of the nitride cavities, Li-atoms more preferably accommodate as single atoms rather than forming clusters. Furthermore, GCMC simulations were employed to obtain the optimized configuration of Li-doped g-C₃N₄. DFT simulation revealed Li-N bond lengths ranging from 1.82 to 2.01 Å. However, the co-existence of N–N and Li–N bonds exerted repulsive as well as attractive interactions, respectively, thereby suffering the crystal distortion within the g-C₃N₄ framework.

Fig. 5 Theoretical representation of (a) interlayer and in-plane crystal structures, electronic density of states (b) and (c), and electrostatic potential distribution (d) and (e) of CN and KCN. Reproduced with permission from ref. 66. Copyright 2024, Elsevier.

2.2. Design strategies of alkali metal-doped g-C₃N₄

The design principle of g-C₃N₄ with alkali metal-doping includes a cost-effective, sustainable, single-step co-condensation approach, generally utilizing hydroxides or halides of corresponding alkali metal ions as primary dopant sources.^{39–41,68–70} Conventionally, the N-rich precursor of g-C₃N₄ and the chosen alkali metal source are calcined together under controlled heating to synthesize the alkali metal-doped g-C₃N₄. In some instances, disodium ethylenediaminetetraacetate (2Na-EDTA) can also be used to achieve both the Na doping and the additional C, N source to modify the CN framework.⁷¹ The doping mechanism is strongly governed by the ionic radius of the dopant. Small alkali cations (Li⁺ and Na⁺) preferentially occupy intralayer cavities, forming strong interactions with heptazine units. This interaction generates localized polarons, enhancing charge carrier mobility and electronic conductivity. In contrast, larger ions (K⁺ and Cs⁺) are incorporated into interlayer spaces, acting as structural bridges that facilitate interlayer charge transport, promoting exfoliation and intercalation. The incorporation of alkali metal ions within the g-C₃N₄ framework can occur via two possible ways, as shown in Fig. 6: (i) co-condensation of alkali metal salts (KCl, KHCO₃, KBr, LiOH, etc) with N-rich precursors of g-C₃N₄ during thermal polymerization and (ii) post-synthesis treatment, which involves the intercalation via molten salt treatment (e.g., heating g-C₃N₄ with alkali metal precursors).^26,72–77 For example, Dong et al. synthesized alkali metal-doped g-C₃N₄ (CN–M, M = Li, Na, K) via co-pyrolysis of melamine with the respective metal chlorides.⁷² The structural and compositional features confirmed the successful alkali doping. XRD exhibited weakened and shifted diffraction peaks, indicating interlayer intercalation and framework disruption upon alkali metal incorporation. FT-IR spectra revealed the emergence of a –C≡N band and decreased N–H stretching, confirming substitution and defect generation. SEM and TEM demonstrated that Na and K doping produced nanosheets with increased roughness and network-like morphology, while Li showed a minimal morphological change. The BET adsorption isotherm demonstrated a reduced surface area and enlarged pore sizes for CN–M, suggesting structural disorder induced by doping. XPS confirmed the presence of dopant elements, surface N vacancies, and increased –C≡N functionalities, while shifts in C 1s and N 1s binding energies indicated electron redistribution from host N atoms to metal ions.

Fig. 6 Direct and step-wise synthesis of alkali metal-doped g-C₃N₄.

In contrast to the one-step co-condensation process of alkali metal doping, the post-synthetic modification strategy follows thermal annealing, enabling simultaneous structural exfoliation and alkali metal doping within the g-C₃N₄ matrix. The synergistic effects of exfoliation and doping endow more accessible active sites and improve the textural features, leading to superior photoresponsive performance. For instance, Li-doped g-C₃N₄ (Li/g-C₃N₄) was fabricated via a simple two-step impregnation process from melamine and LiOH as precursors.⁶⁹ Lithium incorporation levels at 2.7, 5.0, and 9.0 wt% were confirmed by ICP-MS. The XRD analysis revealed an increase in the interlayer spacing upon Li incorporation, indicating reduced crystallinity and structural distortion. FT-IR spectra further confirmed Li doping; it was found that the amine-related band at nearly 3344 cm⁻¹ was reduced after Li doping, indicating H substitution by Li atoms along with the enhanced band at 2174 cm⁻¹ arising from the formation of controlled defect sites. Morphological changes by Li introduction were evident from SEM and HRTEM analyses, where pristine g-C₃N₄ transformed from smooth nanosheets to wrinkled and rock-like structures at higher Li loadings. In a separate study, an ultrathin porous g-C₃N₄ (MCN) was constructed by thermally polymerizing melamine and urea. Urea served as a porogen to create a porous framework. Subsequently, K-doped MCN (denoted as KMCN) was then prepared by hydrothermal recalcination of MCN and KOH as precursors.⁷⁴ The successful synthesis of KMCN was evidenced by PXRD, XPS, and FT-IR spectroscopy. XRD data revealed that the porous structure of MCN remained intact after K incorporation. Moreover, a noticeable decrease in the intensities of (100) and (002) diffraction peaks of MCN was observed, suggesting the defect creation due to incomplete polymerization. Furthermore, the FT-IR analysis revealed that the peak around 3300 cm⁻¹ was somewhat reduced, indicating the substitution of H by K⁺ in –NH₂ groups, resulting in the creation of N–K linkage. Again, XPS analysis supported the successful doping of K⁺. The disappearance of two N 1s peaks near 400.9 and 399.5 eV in KMCN and the appearance of an additional peak around 400.7 eV (C-N3) demonstrated the deprotonation effect caused by K-doping. Additionally, K 2p peaks at 295.2 eV and 292.8 eV suggest the successful doping of K⁺ and N–K bond formation (Fig. 7a–c). The morphological analysis revealed that after K doping and recalcination, the smooth surface of g-CN becomes rough, porous, and a stacked pimple-like structure (Fig. 7d–g). The BET surface area and the pore volume of KMCN were found to be 50.1 m² g⁻¹ and 0.32 cm³ g⁻¹, suggesting the existence of mesopores in the structure.

Fig. 7 XPS spectra of (a) C 1s, (b) N 1s, and (c) K 2p of g-C₃N₄, MCN, and KMCN, SEM (d) and (e) and TEM (f) and (g) images of MCN and KMCN. Reproduced with permission from ref. 74. Copyright 2018, American Chemical Society.

Similarly, Hu et al. reported a post-synthesis strategy for fabricating gradient K-doped g-C₃N₄ (KCN) through sequential thermal treatment.⁷⁵ Initially, bulk g-C₃N₄ (CN) was synthesized by pyrolysis of melamine at 550 °C. This CN material was then thoroughly mixed with KCl and subjected to secondary calcination at 550 °C under a N₂ atmosphere (Fig. 8a). The initial polymerization at moderate temperature led to the formation of a defective layer structure with –NH₂ and –NH groups (Fig. 8b and e). Upon subsequent high temperature treatment, the hydrogen bonds were removed, the heptazine units were distorted, and the n–π* electronic transition was activated (Fig. 8c and f). Furthermore, K⁺ incorporation introduced –C≡N functionalities (Fig. 8d and g) and created a compositional gradient, leading to variation in band structure. Structural analysis revealed that K⁺ intercalation into the heptazine framework reduced the basal spacing, as evidenced by a distinct shift in the (002) XRD diffraction peak from 27.8° (pristine CN) to 28.2° (KCN). Morphological characterization further demonstrated that the introduction of K⁺ ions fragmented the simple plate-like particles of CN to smaller, irregular plate-like particles of KCN, resulting in a reduced crystallinity. HRTEM analysis confirmed the absence of discernible lattice fringes in KCN, indicating its lower crystallinity. N₂ physisorption analysis indicated improved textural properties in KCN, showing type IV isotherms with increased surface area relative to CN.

Fig. 8 (a) Schematic representation of the formation mechanism of CN samples. Structural evolution viewed from the [002] direction (b)–(d) and the [100] direction (e)–(g). Reproduced with permission from ref. 75. Copyright 2025, Royal Society of Chemistry.

The supramolecular self-assembly of g-C₃N₄ precursors is a template-free method, which leads to the production of highly porous g-C₃N₄. This efficient synthesis approach involves the formation of supramolecular aggregates via the self-assembly of g-C₃N₄ precursors such as melamine, cyanuric acid, urea, etc., through hydrogen bonding. The resultant supramolecular aggregates coupled with alkali precursors to form alkali-doped porous g-C₃N₄-based photocatalysts.^76,77 For instance, Lu et al. fabricated K-doped g-C₃N₄ porous nanosheets (CM-K) through supramolecular self-assembly, followed by a chemical cross-linking copolymerization method.⁷⁶ At first, the supramolecular precursor (MCA) was formed by treating melamine and cyanuric acid in ethanol via an H-bonded self-assembly process. The resultant MCA precursors were copolymerized with potassium hydrogen phthalate (KHP) through the cross-linking phenomenon, followed by acid–base interaction. Finally, the resultant precursor was calcined to obtain CM-K (Fig. 9a), having a more pronounced BET surface area of 127 m² g⁻¹ with abundant mesoporous channels. The resultant CM-K material exhibited a hollow box structure with enriched porosity and higher pore volumes, which can be attributed to the gas bubble formation during the thermal decomposition of the supramolecular complex, and the addition of KHP led to inadequate crystal growth, which formed a more porous structure with numerous adsorptive sites (Fig. 9b–e). HRTEM EDX elemental mapping confirmed the uniform nanoscale distribution of K throughout the g-C₃N₄ framework (Fig. 9g–j), and the magnifying aberration-corrected high-angle annular darkfield scanning transmission electron microscopy (AC HAADF-STEM) recorded the bright spots, which correspond to the existence of K single atoms in the periphery of the crystalline triazine motif in CM-K₁₂, resulting in successful K-single atom doping in g-C₃N₄ (Fig. 9e and f). Moreover, the resultant porous CN framework provided an appropriate micro-environment to ideally host the K atoms and acted as a suitable support matrix to stabilize K-single atoms in g-C₃N₄ interlayers via forming K–N coordination. The K-single atom doping was further confirmed from the K 2p XPS and Fourier-transform extended X-ray absorption fine structure (FT-EXAFS), which suggested that the K–N binding sites existed in CM-K. Furthermore, FT-EXAFS also confirmed that the oxidation state of K is between 0 and 1, which suggests the movement of e⁻ from ‘K’ to the g-C₃N₄ matrix. As a result, a greater number of e⁻ incorporated into the π-conjugated structure of CM-K₁₂; therefore, it exhibited a stronger Lorentzian line intensity than CM in EPR analysis (Fig. 18a). Thereafter, another example reported the construction of K-doped tubular g-C₃N₄ (TCN-K) via supramolecular self-assembly of melamine, cyanuric acid, and potassium acetate as precursors.⁷⁷ XRD analysis revealed that there is a minor shift in the (002) peak towards a higher angle in TCN-K, indicating a decreased interlayer stacking distance due to K⁺ doping. TCN-K exhibited a uniform porous tubular morphology with a diameter of ∼80 nm, and K⁺ incorporation did not disrupt the tubular structure, as confirmed by morphological investigation by SEM and TEM. The BET isotherm of TCN-K possessed a significantly higher surface area (62.1 m² g⁻¹) and pore volume (0.378 cm³ g⁻¹), offering more abundant active sites, compared to TCN. XPS peaks of C and N exhibited no characteristic shifts after K⁺ doping, suggesting interstitial doping rather than substitutional doping. Moreover, a distinct doublet peak for K 2p_3/2 at 293.1 eV and K 2p_1/2 at 295.3 eV was observed in TCN-K, which provided direct evidence for the successful incorporation of K into the TCN lattice.

Fig. 9 (a) Diagrammatic overview of the CM-K sample preparation route, TEM images of (b) CN (bulk g-C₃N₄), (c) CM (porous g-C₃N₄), and (d) CM-K, (e) HRTEM image of CM-K with the corresponding SAED present in the inset, (f) HAADF-STEM image of CM-K along with the EDX elemental distribution maps for (g) C, (h) N, (i) K, and (j) their composite overlay. Reproduced with permission from ref. 76. Copyright 2024, American Chemical Society.

Considering all the discussed examples in this section, g-C₃N₄ with successful alkali metal doping exhibits multifunctional properties, including electronic tuning through electron redistribution and defect creation, along with increased active-site accessibility and structural optimization that promotes porosity.

2.3. Alkali metal-induced structural and optoelectronic modulation in g-C₃N₄

The structural and optoelectronic modifications induced in g-C₃N₄ upon alkali metal doping are discussed in this section. The incorporation of alkali metal cations (Li⁺, Na⁺, K⁺, etc.) within the interlayer cavities of g-C₃N₄ can interrupt its ideal stacking order, leading to modifications of the crystal parameters and significantly altering the structural configuration. Consequently, more lattice defect sites and structural distortion arise, leading to higher surface and exposure of abundant active sites. Moreover, the doping of alkali metals modulates the electronic parameters by optimizing the charge distribution to create favorable electron transport channels that promote efficient charge transfer process. Due to the charge redistribution, shallow donor states are created, resulting in band gap narrowing with redshift in light absorption for better light harvesting.^78–82 Jiang et al. demonstrated that these alkali metal atoms are introduced as dopants; they modulate the electronic structure by shifting the CB and VB positions and inducing intra-band defects. This electronic modulation arises from the interaction between the isolated atomic orbitals of the dopant and molecular orbitals of g-C₃N₄, resulting in tuning of band edge alignments, depicted in Fig. 10a.⁴⁵ The optical and electronic characteristics of g-C₃N₄ differ according to the nature and concentration of the dopant. Several theoretical studies revealed that the introduction of Li atoms into the crystal structure of g-C₃N₄ significantly influences its electronic structure as Li atoms preferably occupy open-hollow sites in g-C₃N₄, forming three Li–N bonds with adjacent N-atoms having bond lengths of 1.82–2.01 Å.⁸² This Li-doping induced structural distortions due to the close proximity of N–Li, which enhances the N–N repulsion effect. Notably, it was also indicated that introducing an odd number of Li atoms into the g-C₃N₄ framework induces metallic behaviour, whereas doping with an even number of Li atoms alters the electronic configuration by broadening the band gap. Furthermore, Li incorporation enhances the optical absorption but reduces the thermodynamic stability of g-C₃N₄. Additionally, with an increase in Li concentration, the work function gradually decreased in Li-doped g-C₃N₄. In a separate study, Ma et al. developed Ag-loaded, Li-doped g-C₃N₄ via a straightforward and economical calcination approach by using melamine, CH₃COOLi, and AgNO₃ as precursors.⁸³ Li⁺ ions with different concentrations were doped into the interlayer voids of g-C₃N₄ (Li(x)–CN, where x = 0.05, 0.10, 0.15, 0.25%), forming a Li–N bond, as confirmed by XPS analysis. The Li deposition led to inhibition of the long-range crystal growth as evidenced by the PXRD analysis. Moreover, Li doping narrowed the band gap from 2.45 eV to 1.99 eV and extended visible light absorption with a red shift, accompanied by deepening of the color of the synthesized samples with increasing Li concentration (Fig. 10b and c). Furthermore, Li(x)–CN demonstrated a significant PL quenching, indicating highly suppressed charge recombination and improved charge separation. Additionally, Ag deposition caused improved visible light harvesting via the SPR phenomena. This dual incorporation of Li and Ag in g-C₃N₄ successfully tuned its optical features and ultimately ameliorated the photocatalytic activity.

Fig. 10 Illustration of (a) the band structure of g-CN and alkali metal-doped g-CN. Reproduced with permission from ref. 45. Copyright 2017, Elsevier. (b) UV-vis absorption spectrum and (c) band gaps of all the synthesized samples with different Li concentrations. Reproduced with permission from ref. 83. Copyright 2017, Elsevier. (for interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

Similarly, Zhang and co-workers synthesized Na-doped g-C₃N₄ samples (Na(x)–CN, where x = 0.02, 0.05, 0.1, and 0.2 M of NaOH) via calcination of dicyanamide and NaOH, where Na atoms occupied interstitial regions.⁸⁴ SEM analysis revealed partial fragmentation of the in-plane structure of g-C₃N₄, resulting in smaller crystallites by Na-doping. The XRD patterns indicated that there is a slight shift in the (002) diffraction peak towards the lower angle in Na(0.05)–CN, consistent with an increase in interlayer distance due to interstitial Na-doping, resulting in a decrease in grain size. BET isotherm analysis of Na(0.05)–CN demonstrated a type IV isotherm, signifying the existence of mesoporous architecture further confirmed from the hysteresis loop in the low-pressure range (0.4 < P/P₀ < 0.9) originated due to the intra-aggregate porosity and an additional loop at higher pressure in the range of 0.9 < P/P₀ < 1, corresponding to the presence of larger mesopores in Na(0.05)–CN (Fig. 11a). The SSA of Na(0.05)–g-CN increased to 33.5 m² g⁻¹ and provided a larger number of active sites for light absorption to carry out photocatalytic reactions. This increased SSA might be due to Na-doping, hindering the g-CN crystal growth. The XPS analysis further confirmed interstitial Na incorporation, evidenced by peak shifting towards lower binding energy for N 1s in Na–g-CN, mediated by robust interaction between N atoms and Na atoms (Fig. 11b and c). Due to the electronegativity of the N atom, a shift of e⁻ occurred from Na towards the N atom. Moreover, the Na 1s binding energy suggested that Na ions are coordinated within the C–N framework through N-bridging interactions. Hence, in Na(0.05)–g-CN, the availability of electron pairs on the N-atom contributes significantly to the redistribution of electron density and modulates the electronic structure. UV-vis spectral analysis suggested a red shift in the absorption edge, accompanied by a band gap reduction from 2.65 eV for bulk g-CN to 2.58 eV for Na(0.05)–g-CN. With increasing Na concentration, the CB and VB edges were gradually altered (Fig. 11d). Also, PL and electrochemical analyses confirmed the Na doping and subsequent structural modifications and optimized band alignment exhibited an enhancement in the separation kinetics of photogenerated charge carriers. In another study, Sudrajat et al. fabricated Na-doped g-C₃N₄ through solid-state thermal polycondensation of a mixture of urea, NaCl, and NH₄Cl.⁸⁵ It has been reported that during synthesis, exfoliation occurred and several edge defects were found. The XRD analysis revealed that the (002) and (100) peak intensities decreased, indicating lower crystallinity and exfoliation of g-C₃N₄, confirming an increase in interlayer spacing by Na doping. Moreover, an increase in Na content was found to inhibit the polymeric condensation process. By Na doping, the morphology of undoped g-C₃N₄ was disrupted to smaller, thinner nanosheets, evidenced by SEM and TEM analyses. These structural changes resulted in an increase of SSA to 27.6 m² g⁻¹ for the highest Na-doped sample and also led to a reduction in band gap, altering band edge potentials. The incorporation of Na tuned the band alignment by reducing the band gap and facilitating a shorter charge carrier migration path within Na–g-C₃N₄. Due to this, there was a redshift in the PL emission and a marked quenching of the PL intensity, signifying the effective charge carrier separation.

Fig. 11 (a) BET adsorption isotherm of g-C₃N₄ and Na(0.05)–CN, (b) N 1s, (c) Na 1s XPS spectra (possible doping site of Na⁺ (inset)) of Na(0.05)–CN, and (d) band gap structures of the prepared materials. Reproduced with permission from ref. 84. Copyright 2014, Royal Society of Chemistry.

A similar example of Na-doped g-C₃N₄ was reported by Guo and co-workers. They synthesized a deprotonated Na-doped g-C₃N₄ (Na–CN) material via solid-state reaction involving mechanochemical blending of melamine and NaCl in a weight ratio of 1 : 10, followed by small tablet creation and thermal treatment within a quartz tube.⁸⁶ This process facilitated a strong interaction between precursors, enabling Na⁺ insertion into the g-C₃N₄ crystal lattice. PXRD analysis suggested an expansion of the lamellar spacing due to the larger radius of Na⁺ ions, indicating that the Na⁺ ions could potentially replace the terminal protons and interact with the surrounding N atoms in Na–CN (Fig. 12a and b). TEM investigation showed that the lamellar morphology assembled into a stacked structure with an interplanar distance of 1.180 nm in Na–CN (Fig. 12c). After deprotonation, the remaining e⁻ were delocalized over adjacent C-atoms and resulted in the origination of defect levels within the band gap. The UV-vis spectra revealed a redshifted absorption edge (483 nm) after modification, with a decreased band gap to 2.57 eV, which is 0.12 eV less than pristine g-C₃N₄ (Fig. 12d). Mott–Schottky analysis further elucidated the electronic structure of Na–CN, indicating that the valence band maxima (VBM) position after deprotonation remained unaltered, while the conduction band minima (CBM) potential of Na–CN shifted to −1.19 V, compared to pristine CN (−1.31 V), which is a declination by ca. 0.12 V. Moreover, a marked suppression of the PL signal suggested an enhanced localization of e⁻ at the terminal surface sites of Na–CN. Additionally, the observed redshifts of the emission peak indicated the existence of intermediate electronic states within the band structure (Fig. 12e). The intensified transient photocurrent (TPC) generation indicated a superior separation efficiency of photoinduced carriers, which further supported the PL observations.

Fig. 12 (a) XRD patterns of g-C₃N₄ and Na–CN, (b) schematic depiction of synthesis of Na–CN, (c) TEM of Na–CN, (d) UV-vis DRS, and (e) PL spectra of the synthesized samples. Reproduced with permission from ref. 86. Copyright 2016, Royal Society of Chemistry.

K⁺ doping, analogous to Na⁺ incorporation, modifies the g-C₃N₄ framework and enhances its optoelectronic features. The introduction of K⁺ ions leads to reduced crystallite size, diminished recombination rate, enhanced charge kinetics, and improved surface properties and structural modification. Considering the favorable aspects of K⁺ doping, numerous research groups have investigated K-doped g-C₃N₄ systems for diverse photocatalytic applications.^87–94 For example, Li et al. developed an eco-friendly K-doped g-C₃N₄ composite supported on biochar via a calcination-impregnation approach employing dicyanamide, KOH, and biochar as precursors.⁸⁷ K-doping disrupted crystallinity and reduced the particle size, while biochar provided a high surface area and induced porosity. K-doping caused a red shift in the absorption edge, resulting in band gap narrowing and modulation of electronic structure, while improving charge carrier separation efficiency. Wang and coworkers reported a series of K-doped g-C₃N₄ (denoted as K–CN-10) photocatalysts, fabricated through thermal polymerization using thiourea and KBr. The formation mechanism of K–CN-10 is depicted in Fig. 13a.⁸⁸ Upon K⁺ intercalation, the –NK bonds were formed with the tri-s-triazine units, resulting in weakening of the (002) XRD peak intensity, with peak shifting from 27.4° to 28.3°, indicating a reduction in the interplanar distance. The strong interaction of K⁺ ions with the g-C₃N₄ ring improved the π-conjugation by lowering the interlayer energy barrier and simultaneously promoting a higher extent of polymerization. K⁺ incorporation in g-C₃N₄ altered its morphology. Both SEM and TEM images unveiled that K–CN-10 exhibited an anisotropic 2D lamellar structure with mesoporous channels (Fig. 13b and c). The formation of this mesoporous framework was further confirmed from N₂ sorption isotherms, which displayed a type-IV isotherm accompanied by H3 hysteresis loops, suggesting that the synthesized material possesses slit-like pores arising from the aggregation of plate-shaped particles. K⁺ modification in g-C₃N₄ reduced its band gap, thereby improving its light absorption ability and simultaneously prolonging the lifetime of photoexcitons, leading to a restrained recombination rate. Thereafter, Zhang and coworkers fabricated a K-doped g-C₃N₄ photocatalyst via a thermal condensation process employing dicyanamide and potassium iodide (KI) as starting materials.⁸⁹ Upon K-doping, the resultant material exhibited a redshift with a reduced band gap of 2.64 eV (K-C₃N₄). Band structure analysis suggested that the VB position of K-C₃N₄ exhibited an anodic shift due to the K-doping. Furthermore, the reduced PL intensity of K-C₃N₄ suggested the slow recombination of the photogenerated charge carriers, thereby enhancing charge separation. The EIS Nyquist plot of K-C₃N₄ exhibited a smaller arc radius than g-C₃N₄ due to the charge redistribution and enhanced interfacial charge transfer induced by K-doping. Consequently, K-C₃N₄ indicated a photocurrent response that was nearly four times higher relative to undoped g-C₃N₄. In a separate study, K-doped g-C₃N₄ (K–CN) was constructed by Hu et al. by calcining dicyanamide and KOH.⁹⁰ The interplanar spacing (0.71 nm) in g-C₃N₄ can accommodate K⁺ ions through ion–dipole interaction, as depicted in Fig. 13d. The introduction of K⁺ induces structural distortion in g-C₃N₄, causing an enhancement in surface energy. K⁺ doping caused a shift of the optical absorption edge towards a longer wavelength, resulting in band gap reduction from 2.65 eV to 2.52 eV in g-C₃N₄. Moreover, K-doping also led to changes in the electronic state upon the hybridization of the K's atomic orbital and the C₃N₄'s molecular orbital, enabling tuning of VB and CB potentials. At an optimum K⁺ doping concentration (0.05%), the VB potential became more positive (from + 1.56 eV to + 2.21 eV), while the CB potential became less negative (shifting from −1.09 to −0.31 eV) (Fig. 13e). XPS analysis revealed that K doping caused the change in the electron density on the N atom; as a result, a minor negative shift was observed in N 1s and K 2p binding energy, which evidences the presence of a strong –NK interaction in K–CN. Because of the electron transfer from K to N atoms, the electron density is more localized on the N-atom, thereby leading to a modification in the g-C₃N₄ band structure.

Fig. 13 (a) Schematic representation of the stepwise synthesis and (b) and (c) SEM and TEM pictures of K–CN-10. Reproduced with permission from ref. 88. Copyright 2018, Elsevier. (d) Illustration showing the possible doping sites of K⁺ in K(x)–CN and (e) band structures of g-C₃N₄ and K(x)–CN (where x corresponds to the concentration of KOH, x = 0.02, 0.05, 0.07, and 0.09 M). Reproduced with permission from ref. 90. Copyright 2015, Royal Society of Chemistry.

As mentioned in the DFT section, Xiong et al. investigated the comparative electronic properties of K and Na-doped g-C₃N₄ synthesized through a calcination process. In this system, K atoms were incorporated within the interlayer spaces, acting as bridges between the layers and extending π-electron delocalization, while Na atoms were introduced into the planar framework of g-C₃N₄, resulting in an enhancement of charge density in the graphitic plane.³⁹ Although both dopants enhanced the light absorption by narrowing the band gap, their structural differences led to divergent photocatalytic activity. Both K and Na doping induced a downshift in both the VB and CB levels. The more positive VB potential in K-doped g-C₃N₄ improved the oxidation capability of the photogenerated holes (h⁺), thereby imparting superior photocatalytic efficiency compared to Na-doping. This variation in activity was attributed to the distinct atomic radii of the respective alkali metal dopants. High-temperature treatment induced structural deformation, leading to the formation of –NH₂ groups at the edge of the g-C₃N₄ sheets. The accommodation of alkali metal ions was facilitated by these electron-rich –NH₂ groups. DFT calculations of CN-Na-2 and CN-K-2 samples further confirmed stable doping configurations, with K occupying the interlayer space and Na in the π-conjugated planes. Moreover, Na-doping preserved the crystal structure, while K-doping caused crystal lattice contraction due to its larger size. The electronic location function (ELF) profiles of both K-and Na-doped synthesized samples, illustrated in Fig. 14a and b. In the case of CN-K-2, the ELF values for K–C and K–N bonds were below 0.15, suggesting a weaker covalent bonding. Meanwhile, the Na–N interaction experienced an even lower ELF value, supporting ionic interaction. These distinctions in charge localization suggested that variations in bonding nature influence the photocatalytic performance of the two doped materials. Furthermore, the compositional analysis of the band energies (VB and CB) of g-C₃N₄, CN-Na2, and CN-K2 was estimated by density of states (DOS) analysis, which underscored that the CBM and VBM of all three photocatalysts were constructed by the combination of C 2p and N 2p (Fig. 14c and d). Although Na and K did not contribute to the construction of the band structure, considering their electron-donating ability to the g-C₃N₄ framework, they narrowed down the band gap energy from 1.1 eV for g-C₃N₄ to 0.84 eV and 0.98 eV for CN-Na2 and CN-K2, respectively. The intercalated K atoms facilitated the charge transport channels between the layers, extending the π-conjugated system, reducing electronic localization, and significantly enhancing the charge carrier separation and mobility. This was confirmed by increased photocurrent, smaller EIS arc radius, quenched PL intensity, and prolonged charge carrier lifetime. However, Na doping increased the in-plane electron density but did not create interlayer bridges, which resulted in a higher recombination of photogenerated charge carriers.

Fig. 14 (a) and (b) ELF analysis of CN-K2 and CN-Na, (c) DOS of CN, CN-K2, and CN-Na2, and (d) contribution of Na and K atoms to DOS (inset: partial charge density around the doping K 3p in CN-K2). Reproduced with permission from ref. 39. Copyright 2016, American Chemical Society. (e)–(i) Optimized structural configuration of bulk g-C₃N₄, Li–g-C₃N₄, Na–g-C₃N₄, K–g-C₃N₄ and Rb–g-C₃N₄, respectively. Reproduced with permission from ref. 96. Copyright 2020, Elsevier.

Xu et al. fabricated different alkali metal-doped g-C₃N₄ (GCN-M, M = Na, K, Rb, Cs) photocatalysts through a simple calcination method.⁹⁵ Structural characterization revealed that alkali dopant incorporation disrupted the in-plane long-range order and decreased the interlayer distance and crystallinity, as evidenced by the disappearance of the (100) characteristic diffraction peak in the PXRD pattern. It has been reported that the alkali cations interfere with the formation of hydrogen bonds that typically connect the polymeric melon strands within the plane. The introduction of –C≡N and hydroxyl (–OH) groups originated due to the alkali metal doping, as evidenced by FT-IR and XPS analyses, resulting in the optimization of the electronic structure. All the alkali-doped synthesized materials possessed low SSA, including Na-doped g-C₃N₄ (GCN-Na = 1.92 m² g⁻¹), K-doped g-C₃N₄ (GCN-K = 7.43 m² g⁻¹), and Cs-doped g-C₃N₄ (GCN-Cs = 6.15 m² g⁻¹), except Rb-doped g-C₃N₄ (GCN-Rb = 38.77 m² g⁻¹). UV-vis absorption spectra revealed a significant red shift in the absorption edges for the alkali metal modified samples (GCN-Na, GCN-K, GCN-Rb, and GCN-Cs), indicating enhanced visible light harvesting ability with the corresponding optical band gaps of 2.73 eV, 2.69 eV, 2.59 eV, and 2.72 eV, respectively. The effect was most pronounced for Rb⁺ and K⁺, demonstrating that the larger cations have a more pronounced effect on the structural variation, which altered the band edge potential, i.e., both CB and VB positions shifted upwards. The PL intensity was drastically quenched in the alkali metal-doped samples and followed the trend GCN-K > GCN-Rb > GCN-Cs > GCN-Na, which indicated an effective decline of radiative recombination. Furthermore, the EIS Nyquist plot demonstrated that GCN-K has a shorter arc radius, indicating a reduced charge transfer resistance. In contrast, Zhang and colleagues demonstrated that Rb-doped g-C₃N₄ exhibited the highest photoactivity compared to Li, Na, and K-doped g-C₃N₄.⁹⁶ This superior photocatalytic performance can be ascribed to the fine-tuned structural variation and optimized band alignment achieved due to Rb doping. Although all the synthesized alkali metal-doped samples exhibited a significantly enhanced photo-absorption response in UV-vis spectroscopy, Rb-doped g-C₃N₄ demonstrated an extended absorption range of 500–1100 nm. Furthermore, the band gap energy of Rb-doped g-C₃N₄ was markedly decreased to 2.0 eV compared to the bulk g-C₃N₄ (2.65 eV). Computational modeling and calculated doping energies further revealed the most favorable carbon substitution sites in alkali metal-doped g-C₃N₄, as presented in Fig. 14e–i. Both the experimental observation and DFT study also demonstrated that the extended absorption edges were led by K and Rb dopants, compared to Li and Na.

2.4. Effect of co-doping

Additionally, co-doping with synergistic elements presents a powerful strategy to engineer the photophysical properties of g-C₃N₄ relative to the mono-doping systems. Combining alkali metal ions with other metals or non-metals has emerged as a sophisticated approach for altering the electronic structure, optical properties, and surface chemistry of g-C₃N₄.^58,97–104 The synergy arises from the complementary roles where alkali metals primarily enhance the textural features of g-C₃N₄, thereby facilitating the incorporation of secondary dopants and modulating the electronic environment. This synergistic interplay leads to improved charge separation dynamics, broadened spectral response, and superior photocatalytic performance. Various alkali metal ions containing functionalized co-doped g-C₃N₄ systems are known to improve interlayer van der Waals interaction and enhance charge separation.^98,104–106 Even some metal cations, through interlayer intercalation, accelerate charge transfer between the layers, resulting in the aggregation of localized e⁻ and thus favor the formation of active species for various photo-redox reactions. Therefore, the influence of alkali metal co-doped g-C₃N₄ systems has been investigated in this section. For instance, K, P co-doped g-C₃N₄ grafted with the –C≡N group (MPCN) was synthesized by a dual-step synthesis method involving copolymerization of urea and hexachlorocyclotriphosphazene (HCCP), followed by subsequent treatment in molten KCl medium (Fig. 15a).¹⁰⁷ The co-incorporation of P and K shifted the (100) XRD peak of MPCN from 13.1° to 11.5°, indicating an expansion of in-plane structural periodicity and a shift in the (002) plane, which suggested an increase in the interlayer spacing (Fig. 15b). The successful K⁺ intercalation and –C≡N group grafting was evidenced by FTIR and XPS analyses (Fig. 15c–f). From FT-IR spectra, the peaks at 1000 cm⁻¹, 1153 cm⁻¹, and 2180 cm⁻¹ correspond to the N–K bond, asymmetric vibration of the K–CN₂ group, and the –C≡N group of MPCN, respectively (Fig. 15c). From XPS analysis, the peak intensity of C 1s spectra at 286.4 eV significantly enhanced after the K doping, demonstrating the influence of the alkali group in the creation of –C≡N linkages in MCN and MPCN. Additionally, the C 1s peak shifted to a higher binding energy in MPCN compared to CN and MCN, indicating reduced charge density at the carbon atom due to P substitution (Fig. 15d, h and i). The absence of a discernible P–C signal at nearly 131.5 eV confirmed that P predominantly replaced the C atom rather than the N atom in the CN matrix. The 2p of P exhibited two peaks around 132.8 eV and 133.8 eV corresponding to P–N and P=N binding sites, respectively, suggesting the successful P doping (Fig. 15g). The K⁺ intercalation reduced the band gap in MPCN, and the conjugation between the chromophoric –C≡N group and C=C bonds of the heptazine ring in MPCN extended the visible light absorption, causing a red shift in UV-vis spectra. PL analysis demonstrated that P, K co-doping, along with terminal –C≡N grafting, significantly quenched the PL intensity by inhibiting the charge recombination process. Similar to the above example, the synergistic incorporation of B and K into g-C₃N₄ nanosheets reduced the band gap, extended visible-light absorption, and adjusted the charge transport channels.¹⁰⁸ These changes improved the electron distribution and transfer, which effectively inhibits the charge carrier recombination.

Fig. 15 (a) Diagrammatic sketch of MPCN formation, (b) XRD, (c) FT-IR of CN, PCN, MCN, and MPCN, XPS spectra of (d) C 1s, (e) N 1s, (f) K 2p, and (g) P 2p of CN, MCN and MPCN samples, (h) and (i) optimized atomic model of PCN illustrating two possible substitutional sites. Reproduced with permission from ref. 107. Copyright 2023, Elsevier.

Thereafter, Liang et al. fabricated K, Bi co-doped g-CN (designated as Bi/K–g-CN) via a one-step thermal polycondensation approach.⁴⁷ The synthesized Bi/K–g-CN possessed a layered porous morphology with higher SSA, larger pore diameter, and availability of numerous active sites, confirmed from SEM (Fig. 16a–d) and BET sorption analyses. EDS mapping revealed the homogeneous distribution of both K and Bi dopants in the g-CN structure (Fig. 16e–h), and the incorporation of these dopants caused N-defects, which led to a change in its electronic structure. N₂ sorption analysis identified a type IV adsorption isotherm and H3-type hysteresis loops for Bi/K–g-CN, suggesting the formation of a porous structure, and the SSA was found to be more than that of undoped g-CN. XPS analysis indicated a higher C/N ratio in Bi/K–g-CN (1.25) relative to pristine g-CN (0.73), attributed to a higher number of N-defects. Furthermore, the DFT calculations revealed the optimized configuration of g-CN, K–g-CN, Bi–g-CN, and Bi/K–g-CN (Fig. 16i–l). The larger ionic radii of Bi and K caused significant lattice distortion in g-CN, which led to the removal of N atoms to form N-defect sites. Moreover, the optical absorption of Bi/K–g-CN exhibited a redshift and reduced band gap energy compared to all the synthesized photocatalysts. The band structure analysis further confirmed the dual doping, resulting in the more negative CB and VB edge potential compared to g-CN, and thereby minimizing the photocatalytic reaction threshold for enhanced photoactivity. The synergistic effect of this co-doping created charge heterogeneity, optimized the band position, and improved charge separation.

Fig. 16 (a)–(d) SEM pictures, (e)–(h) EDX profile of Bi/K–g-CN, and (i)–(l) theoretical structural models of all the prepared materials. Reproduced with permission from ref. 47. Copyright 2024, Elsevier.

In another study, Wang et al. reported the fabrication of Na, Cl-co-doped g-C₃N₄ via a simple calcination route employing dicyanamide and NaCl as precursors.¹⁰⁹ The incorporation of an optimal NaCl (0.5) content resulted in a highly crystalline material characterized by sharp PXRD peaks and a slim lamellar morphology. XPS analysis indicated the Na 1s signal at 1071.8 eV, corresponding to Na⁺ species, and the Cl 2p peaks at 197.9 and at 199.6 eV, suggesting the coordination of Cl⁻ with the C and N-atoms, confirming the co-existence of Na and Cl atoms in the g-C₃N₄ framework. After co-doping, the absorption edge was found to be red-shifted, and the band gap energy was reduced from 2.76 eV to 2.67 eV, attributed to the structural and textural modification. Mott–Schottky analysis verified n-type behavior with a positively shifted band potential responsible for enhanced oxidizing ability. The co-doping exhibited a remarkably diminished PL emission intensity, suggesting a suppressed charge carrier recombination, while the photocurrent responses also corroborated with PL results, demonstrating improved charge separation and migration efficiency. In a related study, Na and S-co-doped g-C₃N₄ was successfully constructed through a simple one-step pyrolysis approach.¹¹⁰ The incorporation of Na and S atoms induced structural disruption in the layered structure of g-C₃N₄, as evidenced by PXRD analysis. The presence of Na and S in the synthesized material was corroborated from FT-IR and XPS analysis, attributed to the Na–N, C–S, and C=S bond formation. Furthermore, the synthesized material exhibited a red shift in the optical absorption edge, suggesting narrowing of the band gap and thereby enhancing the light-harvesting capability.

To conclude, all alkali metal-doped g-C₃N₄ systems have been proven to exhibit significantly higher efficiency as visible-light-active photocatalysts relative to pristine g-C₃N₄. However, the K-doped g-C₃N₄ photocatalysts have been extensively investigated in the literature for various artificial photosynthesis applications due to their tailored electronic and optical characteristics. The enhanced photoresponsive features in alkali metal-doped systems arise from their easily tunable band structure, which generates suitable redox potentials to facilitate diverse photo-redox reactions. Furthermore, the co-doping represents a promising approach that deliberately introduces structural disorder, irregular stacking of adjacent g-C₃N₄ layers, and high density of defects, leading to local charge redistribution and optimization of the electronic properties. Table 3 summarizes the synthesis methods, precursors, type of dopant, BET surface area, and band gap of some reported alkali metal-doped g-C₃N₄ photocatalysts.

Table 3 List of alkali metal-doped g-C₃N₄ systems synthesized using various precursors

Photocatalysts	Precursors	Dopant	Synthesis method	Surface area (m^2 g^−1)	Bandgap energy (eV)	Ref.
KCNN	Urea, KCl	K	Thermal polymerization	68	2.50	61
Li/g-C3N4	Melamine, LiOH	Li	Impregnation method	—	2.51–2.68	69
Rb–CN	Melamine, RbNO3	Rb	Calcination	—	2	96
CN-BK	Melamine, KI, H3BO3	K, B	Thermal polymerization	81	2.51	108
10% K-CCN	Dicyandiamide, KOH	K	Thermal polymerization	17.7	2.54	111
AMCN/RF	Urea, KCl, NaCl, resorcinol, formaldehyde	K, Na	Calcination, sonication	—	2.07–2.13	112
Na(0.3)–CN	Melamine, Na2CO3	Na	One-step calcination	7.8	2.34	113
g-C3N5:K0.01	3-Amino-1,2,4-triazole, KBr	K	Calcination	—	1.9	114
RbGCN10	Dicyanamide, RbCl	Rb	Calcination	10.2	2.71	115
CoPc/K/Na/PCN-2	Dicyandiamide, KCl, NaCl	K, Na	Ionothermal polycondensation followed by a hydrothermal method	11.5	2.63	116
CN-80PK	Dicyandiamide, K2HPO4	K, P	Thermal polymerization	38	2.62	117
Na1.2–P0.2–CN	Melamine, NaOH, NH4H2PO4	Na, P	Thermal polymerization	—	2.61	118
Li/P–g-C3N4	Melamine, LiCl, ADP	Li, P	Calcination	14.1	2.19	119
Pd/K–gC3N4	Cyanamide, KCl, PdCl2	K, Pd	Calcination	—	2.7	120
Ag/OKCN-6	Urea, K2C2O4·H2O, Ag(acac)	K, O, Ag	Thermal polymerization	—	2.54	121
CeO2/K/Na/CN	Urea, KNO3, NaNO3	K, Na	Two-step calcination	60	—	122
GCNML	Urea, LiCl, Mg(NO3)2·6H2O	Li, Mg	Calcination	84	2.58	123

---

3. Applications of alkali metal-doped g-C₃N₄

Alkali metal-modified g-C₃N₄ demonstrates versatile applicability in photocatalysis, with several notable applications described below.

3.1. Solar fuel production

Solar fuel production via artificial photosynthesis over a suitable photocatalyst proceeds through a defined sequence of fundamental steps. The initial step requires adsorption of reactant molecules on the photocatalyst surface, then, upon irradiation of incident photons, electron–hole pairs are generated in the photocatalyst, where photon energy (hν) must meet or exceed its bandgap energy (E_g). Subsequently, the photogenerated e⁻ get promoted from the VB to the CB, leading to their effective separation and migration to the semiconductor surface. The final step involves the surface redox reactions to generate sustainable fuels. However, some of the charge carriers may undergo recombination, including both bulk and surface recombination.^124–126 Consequently, the overall quantum yield critically depends on the efficacy of charge separation and the successful migration of charge carriers to surface-active sites for productive redox chemistry.

Alike g-C₃N₄, alkali metal-doped g-C₃N₄ is a two-dimensional polymeric semiconductor predominantly comprising earth-abundant carbon and nitrogen, featuring a framework constructed from interconnected tri-s-triazine units linked via planar tertiary amine linkages. Possessing a narrow optical band gap, alkali metal-doped g-C₃N₄ exhibits significant absorption within the visible light spectrum, extending to higher wavelengths. Its appropriate structural framework and electronic band structure provide sufficient thermodynamic driving force for various types of solar fuel generation, which includes photocatalytic water splitting, CO₂ reduction, H₂O₂ generation, and N₂ fixation as discussed below. Moreover, the structure–activity relationship to underline the mechanistic investigations was also accomplished by in-depth DFT calculations along with theoretical modellings and in situ characterization (Fig. 17).

Fig. 17 Schematic depiction of alkali metal-doped g-C₃N₄ with modified structural and optoelectronic features for multi-dimensional applications.

3.1.1. Photocatalytic H₂ production. The photocatalytic H₂ production via the H₂O splitting reaction associates with a huge thermodynamic barrier, which requires a Gibbs free energy of ΔG° = 237.13 kJ mol⁻¹ (eqn (1)) and a photon energy of 1.23 eV to accomplish the reaction. This process comprises two half-cell reactions, photocatalytic H₂ and O₂ evolution, as shown in eqn (2) and (3):

Overall water splitting (OWS) reaction

2H₂O + hν → 2H₂ + O₂, ΔG° = 237.13 kJ mol⁻¹ (1)

Photocatalytic H₂ evolution

4H⁺ + 4e⁻ → 2H₂, ΔE° = 0 V (2)

Oxygen evolution reaction (OER)

2H₂O → 4H⁺ + 4e⁻ + O₂, ΔE° = 1.23 V (3)

As indicated by eqn (2) and (3), the band edge potentials of the photocatalyst must align to satisfy the thermodynamic requirement of the OWS reaction. i.e., the CB potential and VB potential are essentially more negative than 0 V vs. NHE, and more positive than 1.23 V vs. NHE, respectively. During photocatalysis, the H₂O oxidation is carried out by photogenerated h⁺ at the VB, generating protons (H⁺), which subsequently get reduced to H₂ by photogenerated e⁻ at the CB of the photocatalyst. Although the VB position of many photocatalysts thermodynamically satisfies the requirement for water oxidation, the OER remains kinetically sluggish owing to its 4e⁻ and 4H⁺ transfer mechanism involving high-energy surface intermediates (eqn (3)). The large overpotential of 1.23 V (eqn (3)) and slow interfacial hole transfer lead to pronounced charge recombination and limited OER efficiency.¹²⁷ These challenges can be addressed by using suitable sacrificial agents that selectively trap the h⁺ and liberated H⁺ at the VB. This enhances charge separation, prolongs the lifetime of CB e⁻ and facilitates H⁺ reduction for H₂ evolution. This process selectively suppresses the high-energy-demanding H₂O oxidation at the VB, while the e⁻ and H⁺ donor nature of the sacrificial agent promotes the efficient H₂ production at the CB via the photocatalytic H₂ evolution process.^128–131

Alkali metal-doped g-C₃N₄-based photocatalytic systems are suitable since they promote efficient charge transfer and effective utilization of photogenerated carriers for enhanced H₂ production relative to pristine g-C₃N₄.^{45,76,132–135} Moreover, the change in the free energy for the water splitting reaction is reduced by alkali doping in the g-C₃N₄ crystal lattice. Towards this, Sun et al. reported K-doped g-C₃N₄ nanosheets (CNK), prepared by thermal calcination of melamine pre-treated with KOH, which exhibited an enhanced photocatalytic H₂ generation of 919.5 mmol h⁻¹ g⁻¹, approximately 13 times superior to bulk CN.¹³² The K-doping caused a red shift in visible light absorption, narrowing the band gap of CNK, which modified the band structure by up-shifting the CB edge. Incorporation of K atoms in g-C₃N₄ interlayers suppressed the photogenerated charge carriers' recombination, confirmed by PL spectra. The enhanced photocatalytic performance was ascribed to the increased surface area, which offers abundant surface-active sites, and optimized band edges that favor efficient photoreduction processes. Coupling K-doped g-C₃N₄ with TiO₂ nanobelts (TCN-K) substantially boosted the hydrogen evolution efficiency, primarily owing to optimized charge transport facilitated by the interfacial linkages of K-atoms.¹³³ Analogous to the previous example, a similar observation in the reduction of band gap, accompanied by improved light harvesting, was reported after K-doping. The incorporated K atoms were chemically bonded to both C and N atoms in the g-C₃N₄ framework, improving interlayer charge transfer. Furthermore, the K-intercalation and 1D TiO₂ nanobelts in TCN-K synergistically improved light absorption, separation and migration of photoinduced charges, reduced electron localization, and extended π-conjugation. As a result, the TCN-K composite exhibited a H₂ evolution rate of 50 µmol h⁻¹ g⁻¹, approximately 2.5 times higher than that of CN-K (without TiO₂, 19 µmol h⁻¹ g⁻¹).

Similarly, Lu et al. demonstrated that effective H₂ production could be accomplished by fabricating the K single atom in g-C₃N₄ (CM-K₁₂).⁷⁶ UV-vis analysis and total density of states (TDOS) calculation revealed that the incorporation of K within the g-C₃N₄ framework resulted in a red shift of light absorption and a narrower band-gap with a more optimized band structure, respectively (Fig. 18b and c). The efficient separation and migration of photogenerated charge carriers were dominated in CM-K₁₂, as evident from the PL and EIS Nyquist plot. Furthermore, DFT calculation also revealed that the CM-K surface possessed higher electron density compared to CN and CM because of K-doping, thereby dramatically enhancing the H₂O adsorption ability and also lowering the activation energy barrier of H⁺ adsorption to facilitate better H₂ evolution (Fig. 18d and e). All these findings signified CM-K as a superior photo-responsive photocatalyst which satisfies the thermodynamic requirement for the H₂ evolution and therefore exhibited an excellent hydrogen production of 127.78 µmol h⁻¹ under visible light irradiation, representing nearly 23-fold higher compared to pristine g-C₃N₄ (Fig. 18f).

Fig. 18 (a) EPR spectra, (b) UV-vis DRS absorption spectra, (c) corresponding band structure, (d) calculated TDOS of CM and CM-K₁₂, (e) Gibbs free energy profile for photocatalytic H₂ production through the generation of H* over CM and CM-L_x. Charge density difference maps for CM-K₁₂, and (f) H₂ evolution performance of the prepared photocatalysts, where CN-K_x denotes K-single atom doped porous g-C₃N₄ with x = 6, 12, 18 corresponding to the amount of KHP precursor added. Reproduced with permission from ref. 76. Copyright 2024, American Chemical Society.

Thereafter, Guo et al. reported that Na-doped, deprotonated g-C₃N₄ (Na–CN) exhibited significantly enhanced photocatalytic performance, achieving 31.5 µmol of H₂, delivering a higher AQY of 1.45% and 15.2 µmol of O₂ evolution within 1 hour under visible light illumination in the absence of scavengers.⁸⁶ This performance markedly surpassed that of pristine g-C₃N₄, which showed a very low H₂ evolution rate (an AQY of 0.05% at 420 ± 15 nm). The significant boost in photoactivity was attributed to improved optoelectronic properties arising from deprotonation and Na doping, which facilitated more efficient charge transfer and prevented the formation of the kinetically unfavorable H₂O₂ intermediate that typically hinders photocatalytic performance (Fig. 19a). The enhanced H₂ evolution highlights the pronounced effect of deprotonation on overall water-splitting reaction under sacrificial-agent-free conditions, along with outstanding photostability and reproducibility. This improvement in the photocatalytic activity can be attributed to the accelerated transfer of photogenerated h⁺ to surface-adsorbed –OH species, facilitating O₂ evolution. Similarly, the photocatalytic H₂ evolution behavior of K-doped g-C₃N₄ using Pt as a cocatalyst and triethanolamine as the sacrificial agent was investigated by Wang and coworkers.⁸⁸ The optimized K-doped material exhibited the hydrogen evolution of 1337.2 µmol g⁻¹ h⁻¹ under visible light irradiation, markedly surpassing the pristine g-C₃N₄ (239.8 µmol g⁻¹ h⁻¹) (Fig. 19b). The band edge potential was altered by varying the doping percentage. The optimized K doping in the K–CN-10 sample improved the band structure and achieved the required CB potential, appropriate for efficient H₂ evolution. The photocatalyst also exhibited excellent stability and reusability, maintaining consistent activity and no photo-electronic structural change over four successive cycles (Fig. 19c).

Fig. 19 (a) Cyclic test for H₂ and O₂ evolution over Na–CN. Reproduced with permission from ref. 86. Copyright 2016, Royal Society of Chemistry. (b) Photocatalytic activity towards H₂ production from water, and (c) photostability test over K–CN-10. Reproduced with permission from ref. 88. Copyright 2018, Elsevier.

In a systematic comparative study on alkali metal-doped g-C₃N₄ for photocatalytic hydrogen production under visible-light illumination, Jiang et al. evaluated the impact of Li, Na, and K doping on the photocatalytic H₂ evolution of g-C₃N₄.⁴⁵ Among all the alkali metals tested, Na-doped g-C₃N₄ demonstrated the best performance, achieving a hydrogen evolution rate of 18.7 µmol h⁻¹, nearly 3.7 times higher relative to pristine g-C₃N₄ and approximately twice that of K-doped g-C₃N₄. This outstanding performance of Na–g-C₃N₄ was ascribed to its finely tuned microstructure, improved light-harvesting ability, and optimized band structure, which together enabled superior charge carrier separation and enhanced charge transfer ability. Electrochemical analyses confirmed that Na doping led to the highest photocurrent response and the lowest charge transfer resistance, while photoluminescence (PL) spectra revealed suppressed radiative recombination. Collectively, these results demonstrate that Na is the most effective alkali dopant in g-C₃N₄ for boosting the photocatalytic H₂ generation under visible light.

Strategic co-doping of alkali metals with complementary metal or non-metal dopants into the g-C₃N₄ host framework tailors its electronic configuration, leading to band gap modulation and boosting charge separation kinetics, which can lead to superior photocatalytic efficiency. For instance, Guo et al. reported K, I co-doped mesoporous g-C₃N₄ (MP–CN–KI) for photocatalytic H₂ production, synthesized by one-pot thermal polymerization of dicyandiamide, SBA-15 hard-template, and KI (Fig. 20a).¹³⁵ The successful formation of mesoporous g-C₃N₄ co-doped with K and I was validated through the combined characterization of BET, XRD, SEM, TEM, and XPS analyses. The resultant MP–CN–KI exhibited a type IV isotherm with H1 and H3 hysteresis loops and a BET surface area of 71.48 m² g⁻¹, which was around 8.6 times higher than that of the bulk g-C₃N₄. FT-IR analysis suggested the existence of the azide group, which could be due to the SBA-15-mediated synthesis and K, I doping. Furthermore, the SBA-15 driven mesopores led to the formation of numerous uncondensed surface exposed amine groups as confirmed from the broadening of the N–H stretching frequency in the range of 2900–3500 cm⁻¹. The microstructure analysis from SEM and TEM measurements suggested a rod-like morphology with numerous porous channels decorated on the surface of MP–CN–KI, suggesting better charge diffusion by reducing the paths and easing the migration of charges from the bulk to the surface. Moreover, the K and I incorporation led to the improved charge conductance; hence, the combination of the mesoporous framework with appropriate K and I doping resulted in a significant quenching in the PL-emission peak for MP–CN–KI compared to the pristine CN. MP–CN–KI exhibited improved light harvesting with a narrowed band gap energy and a red shift in light absorption, due to the synergistic effect of mesoporous creation along with the K and I co-doping. Overall, the ease of charge diffusion, better charge separation, improved optoelectronic features, and enhanced charge carriers’ lifetimes due to the mesoporous framework, along with the K, I doping, synergistically enhanced the photocatalytic H₂ evolution rate to 81 µmol h⁻¹ for MP–CN–KI, which is 10-fold greater than bulk CN (Fig. 20b–e). The robust framework of MP–CN–KI exhibited an impressive photostability even after 5 consecutive cycles.

Fig. 20 (a) Diagrammatic sketch for the formation of MP–CN–KI, (b) EIS Nyquist plots, (c) transient photocurrent response, (d) photocatalytic H₂ evolution of all the synthesized samples, and (e) cyclic stability of MP–CN–KI under visible light illumination. Reproduced with permission from ref. 135. Copyright 2018, Elsevier.

Again, Fang et al. demonstrated Na, O-co-doped g-C₃N₄ synthesized via the solvothermal method, followed by the calcination to amplify the H₂ production rate.¹³⁶ The resultant Na, O-co-doped g-C₃N₄ is a highly crystalline material having enhanced optoelectronic features relative to bulk g-C₃N₄. The successful incorporation of the O-atom modified the band gap energy level by creating structural defect states due to the interference of the O 2p orbitals, which have a potential lower than the N 2p orbitals. Furthermore, the substitutional doping of N atoms by O atoms liberated additional electrons, which were restricted within the range of π-delocalization and contributed to the defect states. Na,O-co-doped g-C₃N₄ exhibited a visible-light-mediated H₂ production rate of 810 µmol h⁻¹ with a quantum yield of 22.3%. Similarly, another report of co-doping by Cao et al. demonstrated the synthesis of P and Na co-doped g-C₃N₄ (Na–P co-doped g-C₃N₄) through pyrolysis of melamine with a very low amount of sodium tripolyphosphate as the precursor for Na and P.¹³⁷ The controlled doping (0.01 wt%) of Na and P resulted in an interstitial doping of Na and P, which can be evidenced by DFT calculation (Fig. 21a). The two non-equivalent C-atoms and three non-equivalent N-atoms of the tri-s-triazine of g-C₃N₄ can accommodate the Na and P atoms in two interstitial sites. Although trace amounts of doping (below 0.1%) didn’t show any remarkable change in absorption edge, with increasing the amount of doping (1–3%), the absorption edge was progressively red-shifted. XPS analysis comprises a peak at 133.7 eV, attributed to the P–N bonding, while another peak at 1070.33 eV, representing the Na 1s species, existed as Na–N binding sites in Na–P co-doped g-C₃N₄. A slight negative shift in the binding energy of the Na–N bond was observed, which was due to the strong coordination of the lone pair of the N-atom and the Na atoms during thermal treatment, confirming the interstitial doping of Na in Na–P co-doped g-C₃N₄. Furthermore, the in-depth electronic structure was estimated by DFT calculations, where the TDOS analysis revealed that the valence band maxima of both g-C₃N₄ and Na–P co-doped g-C₃N₄ were composed of N atoms, while the conduction band minima of g-C₃N₄ was composed of the contributions of C and N atoms. However, the conduction band minima of Na–P co-doped g-C₃N₄ were constructed by P, C, and N atoms, resulting in the extended π delocalization of the photoexcited e⁻ via a charge transport channel due to ‘P’ bridging (Fig. 21b and c). In contrast, Na doping facilitated the increase in the charge density due to its electron donation ability to adjacent N atoms in Na–P co-doped g-C₃N₄. Overall, the controlled dual doping of Na and P in Na–P co-doped g-C₃N₄ led to exposed surface-active sites, excellent charge migration, better charge separation, and high surface charge density, caused the directional flow of photogenerated e⁻ and minimized the reaction barrier for H₂ evolution, thereby displaying an enhanced H₂ production of 191.0 µmol h⁻¹ (2.2 times higher than that of g-C₃N₄) under visible light.

Fig. 21 (a) Optimized model of g-C₃N₄ demonstrating the interstitial doping sites (positions 1 and 2 represent the doping sites for P and Na, respectively) and the calculated density of states (DOS) for (b) g-C₃N₄ and (c) Na–P co-doped g-C₃N₄. Reproduced with permission from ref. 137. Copyright 2017, Elsevier. (d) CO generation rate of pure g-C₃N₄ and various alkali metal-doped g-C₃N₄ photocatalysts, (e) cycling stability performance of Rb–CN, and (f) a schematic figure showing how the charge transfer pathway in Rb–CN for CO₂ photoreduction. Reproduced with permission from ref. 96. Copyright 2020, Elsevier.

Zhu et al. synthesized a series of K and halogen (F, Cl, Br, and I) co-doped g-C₃N₄ photocatalysts by a simple single-pot thermal condensation to evaluate their hydrogen evolution performance.¹³⁸ Among all the synthesized doped systems, K, F co-doped g-C₃N₄ (F–K–C₃N₄) demonstrated the highest photocatalytic activity, with a H₂ evolution rate of 1039 µmol h⁻¹, approximately 8.5 times more than bulk g-C₃N₄. This amplification in the photoactivity was primarily attributed to the combined effect of K⁺ and F⁻ dopants. The construction of robust C–F and C–K bonds not only induced a red shift in the light absorption but also effectively tuned the band structure and improved charge carrier separation. Furthermore, the introduction of K⁺ ions created an internal electric field within the g-C₃N₄ matrix, which promotes the directional charge transport in the resultant material.

3.1.2. Photocatalytic CO₂ reduction. The photo-assisted CO₂ reduction to produce a wide range of sustainable hydrocarbon fuels presents a significant strategy to mitigate various environmental issues, including global warming.^139–141 However, the inherent thermodynamic stability of the CO₂ molecule is attributed to its strong C=O bond, with its high bond strength of 750 kJ mol⁻¹, which impedes adsorption and activation on the photocatalytic surface. Hence, high energy is required to activate the C=O bond for respective hydrocarbon conversion.^{25,78,140–144} The efficacy of a semiconductor photocatalyst in this process is governed by band structure alignment, i.e., the CB minima position must be more negative to thermodynamically drive the CO₂ reduction reaction, and the VB maxima must be more positive to promote hole scavenging. The photocatalytic CO₂ reduction can be either a single-electron or a multi-electron reduction process. In the former case, i.e., the single-electron CO₂ reduction process (CO₂ + e⁻ → CO₂˙⁻) is thermodynamically infeasible, as this process is associated with a highly negative redox potential of 1.85 V vs. NHE.^141–145 However, a viable photocatalytic pathway typically involves a multi-electron, proton-coupled reaction that requires relatively less negative reduction potential. The product selectivities for CO₂ photoreduction and conversion efficiency are further complicated by competing kinetic and thermodynamic factors. For instance, the photoreduction of CO₂ to CH₄ by 8e⁻ reduction is thermodynamically favorable and is kinetically constrained. Conversely, reduction pathways involving a lower number of e⁻, such as CO or HCOOH production, often exhibit faster kinetics but may be less thermodynamically favorable.^143–145 Various photoreduced products of CO₂ photoreduction with respect to the number of photogenerated e⁻ and reduction potential values are given below (eqn (4)–(12)),

CO₂ + e⁻ → CO₂˙⁻, E° = −1.85 V (4)

CO₂ + 2H⁺ + 2e⁻ → HCOOH, E° = −0.61 V (5)

CO₂ + 2H⁺ + 2e⁻ → CO + H₂O, E° = −0.53 V (6)

CO₂ + 4H⁺ + 4e⁻ → HCHO + H₂O, E° = −0.48 V (7)

CO₂ + 6H⁺ + 6e⁻ → CH₃OH + H₂O, E° = −0.38 V (8)

CO₂ + 8H⁺ + 8e⁻ → CH₄ + 2H₂O, E° = −0.24 V (9)

2H₂O + 4h⁺ → O₂ + 4H⁺, E° = +0.81 V (10)

H₂O + h⁺ → OH˙ + H⁺, E° = +2.31 V (11)

2H⁺ + 2e⁻ → H₂, E° = −0.42 V (12)

It can be seen that the relative energies of key intermediates significantly influence the product selectivity. Moreover, the optimum photocatalyst surface dictates the CO₂ adsorption extent and the respective band structure favors the specific multi-electron reduction process to achieve better selectivities for the CO₂ photoreduction process.^146–149 The g-C₃N₄ surface facilitates optimum adsorption and activation of CO₂, but due to its unsuitable CB position, the photoreduction is not that effective. Therefore, the rational design of g-C₃N₄ via alkali metal doping precisely engineers its band edge positions to provide sufficient overpotential for the desired multi-electron CO₂ reduction, which is also considered a low-cost and environmentally benign technique. Several studies from the literature are discussed over alkali metal-doped g-C₃N₄-based photocatalysts.^{80,91,96,114,150} For example, Zhang and coworkers demonstrated the photoelectric properties of pristine and alkali-doped (Li, Na, K, Rb) g-C₃N₄ and conducted a comparative study of photocatalytic CO₂ reduction.⁹⁶ Among all synthesized alkali-doped materials, Rb–g-C₃N₄ (Rb–CN) formed thin mesoporous nanosheets and showed red-shifted absorption with reduced band gap. Furthermore, Rb–CN displayed maximum photoconductivity and lowest electron transport resistance (confirmed from the EIS study), compared to other synthesized photocatalysts. The CO yields were 3.6, 5.8, 7.4, 9.8, and 12.1 µmol g⁻¹ for bulk g-C₃N₄ and all the synthesized alkali-doped g-C₃N₄ (Fig. 21d). Among all, Rb–CN exhibited the maximum photoactivity and maintained excellent catalytic stability towards CO₂ photoreduction (Fig. 21e). The photoreduction process was accomplished with multiple reaction steps, where the CO₂ molecules got adsorbed onto the Rb-doped–g-C₃N₄, activated by the photoexcited charge carriers on the surface of Rb-doped–g-C₃N₄. As a result, the CO₂ molecule converted to a surface adsorbed *CO₂ active intermediate, which underwent a 2e⁻ reduction process to generate CO, as shown in Fig. 21f.

Similarly, to enhance CO₂ photoreduction activity, flaky g-C₃N₄ with Na-doped and N-defects, designated as bmw-DCN-x (where x = 10, 20, 30, 40, NaCl: DCDA mass ratio), was fabricated using NaCl and DCDA as precursors (Fig. 22a).¹⁵¹ NaCl acted both as a Na source and a structure-directing agent to generate N defects at the interlayer edges while creating a flaky morphology of g-C₃N₄. FT-IR revealed that bmw-DCN-x exhibited an additional band at 2177 cm⁻¹, corresponding to the existence of –C≡N groups. In contrast, this peak was absent in bulk g-CN. The formation of Na-doping and N-defect was evidenced by XPS and EPR analyses. The combinational effect of Na doping and N-defect led to optimization of the microstructure of bmw-DCN-30, which further endowed a narrower band gap, improved light harvesting, efficient charge separation, and excellent photoexciton migration. CO₂ adsorption capacity (both physical adsorption and chemical adsorption) substantially increased in bmw-DCN-30, compared to bulk-CN (BCN), attributed to the increase in surface basic sites on the photocatalyst. Similar to Lewis acid adsorption, CO₂ adsorbed on these basic sites with Na atoms and N defects likely serve as the active basic position, thereby improving the CO₂ adsorption and activation on the bmw-DCN-30 photocatalyst. This resulted in an optimal CO₂ photoreduction performance for bmw-DCN-30, achieving CO and CH₄ production rates of 30.6 µmol g⁻¹ h⁻¹ and 5.4 µmol g⁻¹ h⁻¹, respectively, which were 15 and 11 times higher than that of BCN (Fig. 22b and c).

Fig. 22 (a) Diagrammatic illustration of the synthesis of Na-doped flaky g-C₃N₄ with N-defects, (b) CO and (c) CH₄ evolution rates of BCN and bmw-DCN-x. Reproduced with permission from ref. 151. Copyright 2021, Elsevier.

In another example, Debnath and coworkers presented a photocatalytic CO₂ conversion using K-doped g-C₃N₅ (g-C₃N₅):K(0.01) fabricated by thermal polymerization of 3-AT and KBr (Fig. 23a).¹¹⁴ K-doped g-C₃N₅ exhibited an enhanced CO₂ photoreduction in visible light irradiation with a Pt cocatalyst. The CH₄ production rate reached 0.46 µmol g⁻¹, approximately 1.7 times and 2.7 times higher than that of g-C₃N₅ and g-C₃N₄, respectively. The azo linkage and K-doping reduced the band gap energy to about 1.9 eV, resulting in an enhanced visible light harvesting, as confirmed by UV-vis DRS analysis and DFT analysis (Fig. 23b). Additionally, K-doped g-C₃N₅ showed PL quenching, indicating suppressed e⁻/h⁺ recombination, facilitating better charge carrier separation and transfer (Fig. 23c and d). Furthermore, doping of K⁺ in g-C₃N₅ provided more active sites for the CO₂ adsorption, contributing to higher photocatalytic CO₂ reduction activity, and maintained its performance stability over four consecutive cycles. According to the structure optimization analysis, as shown in Fig. 23e–g, the larger radius of K⁺ didn’t favor substitutional doping but resulted in an ion–dipole interaction between the N atom of g-C₃N₅ and K⁺. This interaction augments optimized electronic adjustment and better charge transfer efficiency to boost the enhanced charge separation. As a result, K-doped g-C₃N₅ exhibited photoreduction of CO₂ to CH₄ with 100% selectivity. Likewise, a K⁺ intercalated ultrathin g-C₃N₄ nanosheet (denoted as CNS K) was reported for CO₂ reduction.¹⁵⁰ The photocatalyst was synthesized via H-bonding self-assembly followed by thermal condensation of DCDA and KOH. CNS K exhibited an enhanced photocatalytic performance, achieving CO and CH₄ at a production rate of 1.65 µmol g⁻¹ h⁻¹ and 0.15 µmol g⁻¹ h⁻¹, respectively. The introduction of K⁺ into the g-C₃N₄ interlayers broadens the optical absorption, thereby modifying the band energy structure and also shortening the charge carrier migration to accelerate e⁻ transfer. Additionally, K⁺ intercalation created built-in electric fields that restrained the charge recombination and increased surface amino groups, enhancing CO₂ adsorption. The in situ DRIFT analysis evidenced that the infrared absorption peaks correspond to the intermediates such as b-CO₃²⁻, m-CO₃²⁻, CO₃²⁻, and HCO₃⁻, confirming stronger CO₂ adsorption and activation on the CNS K surface.

Fig. 23 (a) Schematic depiction of the synthesis of g-C₃N₅ and g-C₃N₅:K(0.01), (b) UV-vis DRS, (c) PL spectra, (d) EIS Nyquist plots of the synthesized samples, and (e)–(g) crystal structure models of g-C₃N₄, g-C₃N₅, and g-C₃N₅:K(0.01), respectively. Reproduced with permission from ref. 114. Copyright 2022, American Chemical Society.

3.1.3. Photocatalytic H₂O₂ production. The photocatalytic production of H₂O₂ from H₂O and O₂ is an endergonic process, requiring a Gibbs free energy of 117 kJ mol⁻¹, which proceeds via two photochemical pathways: (a) the water oxidation reaction (WOR) and (b) the oxygen reduction reaction (ORR). The WOR is a thermodynamically challenging process because it is associated with more complex kinetics and requires a much higher positive oxidation potential to achieve. In contrast, the photocatalytic H₂O₂ production is usually achieved via the ORR pathway over an alkali metal-doped g-C₃N₄ photocatalyst. The ORR pathway can be accomplished by two distinct routes: (i) a two-electron single-step reduction process and (ii) a one-electron two-step reduction process. In the first pathway, O₂ is converted directly to H₂O₂ by accepting 2e⁻ and 2H⁺ in a single step known as the concerted 2e⁻ ORR (eqn (13)). In the other way, the second pathway involves the formation of the superoxide radical (O₂˙⁻) by the reduction of O₂ with one e⁻, as shown in eqn (14). Finally, the generated O₂˙⁻ active species reacts with 2H⁺ and e⁻ to produce H₂O₂ (eqn (15)).^25,152–155

Two-electron single-step ORR:

O₂ + 2H⁺ + 2e⁻ → H₂O₂ (E° = +0.68 V vs. NHE) (13)

One-electron two-step ORR:

O₂ + e⁻ → O₂˙⁻  (E° = −0.33 V vs. NHE) (14)

O₂˙⁻ + 2H⁺ + e⁻ → H₂O₂ (E° = +1.44 V vs. NHE) (15)

The presence of appropriate surface-active sites and a necessary band structure in alkali-doped g-C₃N₄ is crucial to accomplish both the reaction pathways for efficient H₂O₂ production.^43,156–161 For instance, the Na-doped g-C₃N₄ photocatalyst (BNCN) was reported by Gou et al., which acted as a three-dimensional charge carrier transport channel to facilitate efficient H₂O₂ production.¹⁵⁶ BNCN was fabricated by a double pyrolysis method where dicyandiamide and NH₄Cl were first calcined to produce g-C₃N₄, which was subsequently pyrolyzed under an Ar atmosphere with NaBH₄ and sodium acetate to form BNCN (Fig. 24a). The resultant BNCN, having Na intercalation and –C≡N anchoring moieties, possessed a lamellar curly-like structure, as analyzed from the SEM analysis. The successful –C≡N anchoring was understood from the FT-IR, solid state NMR, and XPS analyses. The Na intercalation was estimated by PXRD analysis, which observed the shifting of the (002) plane towards a lower angle, suggesting the positioning of Na within the interlayers of g-C₃N₄. As a whole, the respective modification could form a 3D charge carrier transport channel in BNCN, as demonstrated in Fig. 24b. Due to this, an enhanced degree of electronic delocalization and promotional effects of interlayer charge transportation can be experienced. The synergistic effect of the Na doping and –C≡N grafting led to modifying the band structure of BNCN, due to which it exhibited an extended light absorption ability with a narrower band gap. The modified band structure in BNCN has a more negative CB potential, which suitably facilitates the reduction of O₂ to H₂O₂. The 3D transport channel, due to the –C≡N moiety and Na intercalation, directed the charge density to form an internal electric field (IEF), which promoted the charge separation efficiency. Furthermore, O₂-TPD analysis suggested that BNCN exhibited better O₂ adsorption efficiency, which is a key step of the O₂ reduction. Various trapping experiments and controlled reactions confirmed that the O₂˙⁻ was the active intermediate in H₂O₂ formation, which was obtained by the reduction of O₂ by photogenerated e⁻. Fig. 24c demonstrates that the efficient H₂O₂ production was dominated by the 3D transport channel of BNCN, where the effective IEF could efficiently separate the charge pairs and the intercalated Na atoms, further triggering the charge transportation longitudinally to promote further charge separation. As a result, more photoexcited e⁻ were involved in the O₂ reduction and exclusively formed H₂O₂.

Fig. 24 (a) Diagrammatic sketch of BNCN formation, (b) a schematic showing the 3D charge transport pathway, and (c) photocatalytic H₂O₂ production in 12 h. Reproduced with permission from ref. 156. Copyright 2025, Elsevier.

Another example demonstrated that the ‘K’ doping promotes faster space charge transmission, followed by the efficient charge carrier separation to exhibit better H₂O₂ production.¹⁵⁷ KBr guided calcination of thiourea led to the formation of K and Br doped g-C₃N₄ (K–CN). The K⁺ was reported to be retained in the interstitial cavity of the heptazine ring, which was confirmed from the PXRD analysis. KBr-mediated synthesis modified the framework structure, which can be understood from the FT-IR and XPS analyses. FT-IR analysis demonstrated that after the K doping, a stretching frequency at 2177 cm⁻¹ corresponding to –C≡N linkage was observed in the K–CN framework. Moreover, the bending frequency at 805 cm⁻¹ corresponding to the heptazine ring gradually diminished due to the increasing content of K, suggesting that the heptazine ring transformed into –C≡N moieties. The C 1s spectra of K–CN exhibited a new peak at 286.4 eV, which was attributed to –C≡N and was absent in PCN, without K-doped material. This further clarifies that the –C≡N linkages were formed due to K doping. Furthermore, the peak position of C 1s and N 1s of K–CN was noticeably changed to that of the PCN, suggesting that the electronic structure got altered due to the K doping. K–CN exhibited a redshift light absorption because of the electron-withdrawing nature of –C≡N linkages present in the K–CN framework. The PL analysis suggested that the K doping significantly reduced the PL intensity by efficiently inhibiting the charge recombination process. The K present in the interlayers of K–CN promotes the charge separation by facilitating the flow of e⁻ among the interlayers of K–CN, while the –C≡N moieties, being electronegative in nature, could trap the photoexcited e⁻ and enhance the charge separation process. As a result, K–CN exhibited a smaller EIS radius and higher photocurrent density than PCN, indicating that K–CN possessed better photogenerated electron mobility with efficient charge separation (Fig. 25a and b). The linear sweep voltammetry (LSV) curve using rotating disk electrode (RDE) test was carried out to reveal that the photocatalytic H₂O₂ process involved a 2e⁻ reduction process. Furthermore, the scavenging test confirmed that the photocatalytic H₂O₂ production was achieved by a two-step one-electron ORR process in which O₂˙⁻ and ¹O₂ were the sole reactive oxygen species (ROS). It was also reported that the Br⁻ in K–CN could promote the formation of ¹O₂ to expedite the H₂O₂ production rate. Based on the control experiment, the photocatalytic H₂O₂ efficiency was governed by the O₂ adsorption ability, which was used as a feedstock. K–CN exhibited a comparatively stronger O₂-TPD signal and possessed a lower ΔG value of K–CN from DFT calculations, suggesting a better O₂ adsorption ability of KCN than PCN. The photocatalytic mechanism of H₂O₂ production over the K–CN photocatalyst involved a two-step single-electron oxygen reduction pathway and energy transfer pathway, which was initiated by the formation of photoinduced charge carriers upon light irradiation. As a result, the photogenerated h⁺ oxidized the adsorbed ethanol to produce H⁺, while the photogenerated e⁻ reduced O₂ to O₂˙⁻. The resultant H⁺ and O₂˙⁻ reacted to produce ˙OOH reactive species, which were reduced to HO₂⁻ species and subsequently interacted with H⁺ to produce H₂O₂. Additionally, the energy transfer process facilitated the conversion of O₂ to ¹O₂ through intersystem crossing (ISC), which also participated in the H₂O₂ production, as shown in Fig. 25c.

Fig. 25 (a) EIS curves, (b) transient photocurrent response of PCN and K–CN, respectively, (c) schematic representation of the proposed mechanism for photocatalytic H₂O₂ production. Reproduced with permission from ref. 157. Copyright 2025, Wiley-VCH.

Mandal et al. developed a photocatalyst (denoted as CKCN) by the simultaneous incorporation of K⁺ ions and amorphous carbon in g-C₃N₄, significantly enhancing the photocatalytic production ability.⁴³ This dual-mode structural modification enhanced porosity, induced intra-band defects, and facilitated the distribution of exposed active sites, resulting in a H₂O₂ production of 9000 µmol g⁻¹ h⁻¹ with approximately 30% AQY under visible light irradiation. The elemental mapping analysis confirmed the homogeneous distribution of C, N, O, and K atoms, indicating successful doping of K atoms. The XRD analysis revealed the disordering of the CN framework and defect formation due to the dual incorporation effect. Furthermore, the EPR analysis suggested that CKCN exhibited an intensified signal of Lorentzian line compared to CN, which evidenced the formation of unpaired e⁻ due to N-vacancy and the amorphous structure of CKCN. Again, CKCN exhibited a reduced band gap (2.68 eV), broader optical absorption, superior charge separation, and migration in CKCN, as evidenced through optoelectronic and electrochemical analyses. Moreover, the hole scavenger and spin trapping experiments confirmed that the H₂O₂ production over CKCN was dominated via the O₂ reduction pathway to form O₂˙⁻ as an active intermediate. Furthermore, the band structure analysis found that the formation of CKCN was accomplished with a more negative CB level than g-C₃N₄, appropriate for the reduction of O₂/O₂˙⁻, which augments the efficient H₂O₂ production.

3.1.4. Photocatalytic N₂ reduction. Photocatalytic N₂ fixation offers a promising alternative pathway for NH₃ synthesis under ambient conditions compared to the energy-intensive Haber–Bosch process. The reduction of N₂ to NH₃ is a challenging reaction due to its high thermodynamic barrier, primarily stemming from the molecule's inherent stability.^162–164 The exceptionally high bond dissociation energy of the N≡N bond (941 kJ mol⁻¹) restrains its adsorption and subsequent activation on the photocatalyst surface, thereby causing a potential barrier in N₂ photoreduction. In contrast, photocatalytic N₂ reduction offers a promising alternative pathway for NH₃ production under ambient conditions (ΔG° = 339 kJ mol⁻¹).^165–167 The mechanism of N₂ photoreduction over a photocatalyst surface to synthesize NH₃ includes two simultaneous processes. Upon light irradiation, the photogenerated h⁺ in the VB splits H₂O to H⁺ and O₂; this H⁺ migrates to the CB to produce H₂. Again, at the CB, this H⁺ can be utilized by the photogenerated e⁻ and N₂ to produce NH₃. The photoreduction of N₂ to NH₃ is a multi-electron reduction process that requires 6 e⁻ to be reduced, which is kinetically unfavorable. It is more crucial to drive the N₂ photoreduction over H₂O splitting. The redox potential of N₂/NH₃ is −0.922 V, which can be theoretically achieved due to the favorable CB position (approx. ∼0.9 V) of alkali-doped g-C₃N₄. Besides, the competitive photocatalytic process of H₂O splitting and N₂ reduction usually brings complexity to efficient N₂ reduction.^165–169 Moreover, the recombination of the exciton pair further slows down the N₂ reduction kinetics. However, the optimized microstructure and modulated defect engineering via alkali metal doping offer better N₂ photoreduction by overcoming the above bottlenecks.^170–172 The alkali-doped g-C₃N₄ photocatalytic systems also extend to the photocatalytic N₂ fixation due to the conducive band alignment and acceptable optoelectronic features.

Photoexcited electron–hole recombination can significantly hinder the rate of N₂ photofixation. By tailoring the microstructure environment and introducing alkali metal-induced defect sites, charge separation is exceptionally improved by mitigating their rapid recombination, which accelerates the photoreduction kinetics of N₂. Therefore, alkali-doped g-C₃N₄-based photocatalysts exhibit favorable band alignment and enhanced optoelectronic characteristics that support their ability to drive efficient photocatalytic N₂ fixation. For example, Gu et al. reported that three-dimensional microporous g-C₃N₄ with Li doping (Li-3D-GCN) was synthesized via an efficient template-free method.¹⁷³ The resultant Li-3D-GCN was characterized by PXRD analysis, suggesting the phase purity of the material. The (002) diffraction plane exhibited a noticeable shift towards a lower 2-theta value due to the expansion of the interlayer spacing after the Li doping (Fig. 26a). Furthermore, the Li doping caused a significant red shift in the optical absorption, as illustrated in Fig. 26b. The Li doping, along with porous and high surface area in Li-3D-GCN, reduced the band gap compared to bulk g-C₃N₄ and Li doped in bulk g-C₃N₄ (Li-GCN). The SSA of Li-3D-GCN was 59.5 m² g⁻¹, which was very high compared to 7.9 m² g⁻¹ for bulk g-C₃N₄ and 8.4 m² g⁻¹ for Li-GCN, suggesting the unlocking of easy diffusion and accessible charge transportation in the Li-3D-GCN framework (Fig. 26c). The porous structure of Li-3D-GCN was further revealed by the morphology analysis. The formation of Li–N bonding sites was analyzed by the XPS analysis, which acted as the active center for the photocatalytic N₂ reduction. The N 1s peak was distinctly shifted after the Li doping, suggesting that the unsaturated e⁻ on the N atom is more localized towards the Li atom due to the Li–N coordinate interaction. N₂-TPD analysis and DFT calculation confirmed the potential N₂ adsorption ability of Li-3D-GCN (Fig. 26d and e). The porous framework in Li-3D-GCN allowed more exposed Li–N bonding sites, which efficiently adsorbed N₂ molecules. Furthermore, the Hirshfeld charge analysis showed that the Li doping can significantly increase the Hirshfeld charge of C and N atoms, which facilitates the efficient charge carrier flow in Li-3D-GCN to facilitate the more intense charge carrier separation. The charge density analysis also showed that the flow of e⁻ occurred from the Li atom of the Li–N bonding sites to N₂ to accomplish the photoreduction process (Fig. 26f). As a result, Li-3D-GCN exhibited a superior NH₄⁺ production of 4.8 mg L⁻¹ h⁻¹ g_cat⁻¹, which was 2.5 and 20 times higher than that of Li-GCN and bulk g-C₃N₄, respectively.

Fig. 26 (a) XRD patterns, (b) UV-vis absorption spectra, (c) N₂ adsorption–desorption isotherms, and (d) N₂-TPD of the synthesized catalysts, schematic representation of (e) optimized configuration of N₂ adsorption on bulk g-C₃N₄ and Li-3D-GCN, (f) charge density distribution for N₂ adsorption on Li-3D-GCN, where the yellow and blue regions denote charge accumulation and depletion, respectively. Reproduced with permission from ref. 173. Copyright 2019, Royal Society of Chemistry.

Similarly, another example demonstrates that the effect of appropriate co-doping can significantly boost the N₂ photoreduction process.¹⁷⁴ A series of alkali-doped g-C₃N₄ crystalline materials, such as LK-HTCN (Li–K co-doping), KN-HCN (K–Na co-doping), LKN-HTCN (Li–K–Na co-doping), and LN-HTCN (Li–Na co-doping), were synthesized using the systematic combination of appropriate alkali metal precursors (Fig. 27a). The phase purity and crystallinity of the synthesized material were determined by the PXRD analysis, which suggested that Li doping promoted the construction of the triazine phase, whereas K and Na promoted the development of the heptazine phase (Fig. 27b). FT-IR analysis confirmed the existence of –C≡N linkages, and its peak intensity was comparatively higher in KN-HCN than that in LK-HTCN, LKN-HTC, and LN-HTCN, corroborating the influential nature of K⁺ optimizing the microstructure and promoting more –C≡N linkages for enhanced optical properties and better charge carrier separation (Fig. 27c). Furthermore, the N vacancy was attributed to the strong EPR signal of KN-HCN, associated with a greater number of free e⁻ (Fig. 27d). The morphological analysis suggested that KN-HCN possessed a bulk structure, while LK-HTCN, LKN-HTCN, and LN-HTCN possessed nanorod structures. Moreover, the (002) in KN-HCN was wider than that of LK-HTCN, LKN-HTCN, suggesting the presence of heptazine phases due to the incomplete condensation process. The XPS analysis was employed to further confirm the existence of the –C≡N moieties as well as the formation of K–N bonding sites through ion–dipole interactions in the CN interlayers and in KN-HCN, which could act as charge transport channels and bridging layers for promoting the transportation and separation of charge carriers. The effective charge separation was further understood from the lower PL intensity of KN-HCN. The light absorption was significantly enhanced after the alkali doping and the formation of –C≡N linkages with crystalline phases in all the samples; however, KN-HCN exhibited a strong absorption edge with noticeable red shift in the optical absorption, which could be due to the relatively higher content of –C≡N linkages in its framework. The potential role of K and Na co-doping endowed KN-HCN with a stronger photocurrent density and better charge separation ability. As a result, among all the synthesized photocatalysts, KN-HCN exhibited the highest ammonia yield of 17.01 µmol g⁻¹ h⁻¹ while maintaining the photostability even after the fourth cycle, suggesting the robustness of KN-HCN in photocatalytic ammonia production (Fig. 27e). Furthermore, the variation of wavelength led to production of different amounts of NH₃, coinciding with the UV-vis analysis (Fig. 27f). The combinational effect of the –C≡N linkage and K–Na co-doping for the efficient ammonia production is demonstrated in Fig. 27g. In the photocatalytic N₂ reduction process, the photogenerated h⁺ oxidized H₂O to O₂ and H⁺, while the photoexcited e⁻, upon coupling with H⁺, reduced the surface-adsorbed N₂ to produce NH₃.

Fig. 27 (a) Diagrammatic sketch for the synthesis routes for the obtained samples, (b) XRD, (c) FT-IR spectra of LK-HTCN, KN-HCN, LKN-HTCN, and LN-HTCN, respectively, (d) EPR spectra of BCN and KN-HCN, (e) NH₃ production rate of all the prepared samples, (f) NH₃ production of KN-HCN at different wavelengths, and (g) illustration of the proposed photocatalytic N₂ reduction over KN-HCN. Reproduced with permission from ref. 174. Copyright 2022, Royal Society of Chemistry.

The demonstration of a wide range of the above-discussed case studies shows the promising aspects of the alkali metal doping approach in the g-C₃N₄ matrix for efficient solar energy conversion for sustainable fuel production. Besides these examples, Table 4 further summarizes a list of alkali-doped g-C₃N₄-based photocatalysts governing the production of a wide range of solar fuels.

Table 4 List of alkali-doped g-C₃N₄-based photocatalysts employed for various solar fuel production

Photocatalysts	Precursors	Synthesis method	Type of solar fuel production	Light source	Yield	Ref.
NaK20-CN	Melamine, NaCl, KCl	Calcination	H2 evolution	300 W Xe lamp	4502 µmol h^−1 g^−1	58
KCN-75	Melamine, KCl flux	Thermal polymerization	H2 evolution	300 W Xe lamp	382 mmol h^−1	75
CM-K12	Melamine, cyanuric acid, potassium hydrogen phthalate	Supramolecular self-assembly, cross-linking polymerization	H2 evolution	300 W Xe lamp (λ ≥ 420 nm)	127.78 µmol h^−1	76
K–CN/ZCS	Urea, KCl, Cd (CH3COO)2·2H2O, Zn (CH3COO)2·2H2O, thioacetamide	Calcination, hydrothermal	H2 evolution	300 W Xe lamp	9.606 mmol h^−1 g^−1	93
2.5CCNC	Melamine, cesium acetate	Thermal polymerization	H2 evolution	Xe lamp, 300 mW cm^−2	9.02 mmol g^−1 h^−1	134
K(0.05)–CN	Melamine, C2H3KOS	Calcination	H2 evolution	300 W Xe lamp (λ ≥ 420 nm)	1319 µmol h^−1 g^−1	175
K–CN/DFNS-2	Melamine, KOH, DFNS	Calcination	CO2 reduction	300 W Xe lamp	1.7 mmol g^−1 (CH4)	80
g-C3N5:K0.01	3-Amino-1,2,4-triazole, KBr	Calcination	CO2 reduction	300 W Xe lamp (λ = 385–740 nm)	0.752 µmol g^−1 (CH4)	114
CNS K	DCDA, KOH, polyethylene glycol	Calcination	CO2 reduction	—	1.65 µmol g^−1 h^−1 (CO)	150
bmw-DCN-30	DCDA, NaCl	Calcination	CO2 reduction	300 W Xe arc lamp (λ > 400 nm)	30.6 µmol g^−1 h^−1	151
3 Na–CN	Dicyandiamide, NaHSO3	Thermal polymerization	CO2 reduction	300 W Xe lamp	21.5 µmol g^−1 h^−1 (CO)	176
K-gCN	DCDA, KOH	Calcination	CO2 conversion	5 W LED bulb	1192 mg L^−1 (formate)	177
CKCN	Urea, KNO3	Calcination	H2O2 production	300 W Xe lamp	9000 µmol g^−1 h^−1	43
10 wt% K–g-C3N4	DCDA, KI	Thermal polymerization	H2O2 production	Direct sunlight	200 µmol g^−1 h^−1	46
KCN/ZnO	Melamine, KCl, LiCl, ZIF-8	Calcination	H2O2 production	AM 1.5	4.76 mM	50
MPCN	Urea, HCCP, KCl, LiCl	Thermal polymerization	H2O2 production	300 W Xe lamp (λ > 420 nm)	4424 µmol g^−1 h^−1	107
0.05 K–CN	Thiourea, KBr	Calcination	O2 reduction	10 W blue lamp (λ = 420–430 nm)	26 mmol g^−1 h^−1	157
Li–C3N4	Urea, Li2CO3	Thermal polymerization	O2 reduction	300 W Xe lamp, AM	920 mmol g^−1	158
KCT5-5	Urea, KCl, Ti3AlC2	Calcination, hydrothermal	H2O2 production	500 W Xe (λ = 420–780 nm)	3401 mmol L^−1	159
KCN-6	Melamine, KCl	Thermal polycondensation	H2O2 production	300 W Xe lamp (λ > 420 nm)	818.9 µmol g^−1 h^−1	178
CN-KI3-KI-MV	Melamine, KI	Oil bath reflux calcination	H2O2 production	300 W Xe lamp (λ > 420 nm)	46.40 mmol g^−1 h^−1	179
K–g-C3N4−x	Melamine, potassium oxalate	Calcination	H2O2 production	300 W Xe lamp (λ ≥ 420 nm)	706.9 µmol L^−1	180
Li-3D-GCN	Urea, LiCl	Thermal polymerization	N2 fixation	250 W sodium lamp (λ = 400–800 nm)	4.8 mg L^−1 h^−1 gcat^−1	173
KN-HCN	Melamine, KCl, NaCl	Calcination	N2 fixation	300 W Xe lamp	17.01 µmol g^−1 h^−1	174

---

3.2. Applications of alkali metal-doped g-C₃N₄ towards degradation and removal of contaminants

Over recent decades, the escalating contamination of water, soil and air has become a critical global challenge, driven significantly by industrial discharge and domestic wastewater. Various organic and inorganic contaminants, such as dyes (for example, methyl orange, methylene blue, RhB, etc.), antibiotics, insecticides, and other chemical pollutants, pose widespread concern to the ecosystem. Conventional methods, such as adsorption, membrane-based filtration, enzymatic degradation, and coagulation, are often incompetent to eradicate these potential pollutants or may generate harmful secondary by-products. Photocatalysis offers a promising and sustainable approach for degrading such pollutants. This process involves the acceleration of a photoreaction over a semiconductor photocatalyst under UV or visible light. g-C₃N₄ has emerged as a prominent photocatalyst for the degradation of pollutants.^181–184 When g-C₃N₄ absorbs photons possessing energy equal to or greater than its band gap, it generates photoexcitons, and these charge carriers initiate a series of redox reactions, forming reactive oxygen species (ROS) such as OH˙ and O₂˙⁻. These ROS are highly active and are capable of derivatizing the complex pollutants into simpler, non-toxic compounds. Precisely, the critical aspect of heterogeneous photocatalysis is photo-assisted oxidation of pollutants to benign substances. The underlying mechanism encompasses the generation of excitons and the subsequent pollutant degradation, depicted by the following equations (eqn (16)–(22)):

GCN + hν → e⁻+ h⁺ (16)

h⁺ + OH⁻ → OH˙ (17)

h⁺+ H₂O → OH˙ + H⁺ (18)

e⁻+ O₂ → O₂˙⁻ (19)

h⁺+ pollutants → Oxidation products (20)

e⁻ + pollutants → Reduction products (21)

OH˙ + pollutants → Degradation products (22)

As mentioned in the earlier section, the pristine structure of g-C₃N₄ lacks structural demerits, causing limited optoelectronic features, but the recent studies demonstrate that the alkali metal ion incorporation can effectively provide more exposed active sites, suppress the charge recombination rates, and create transport channels for photoexcited e⁻ to boost the photocatalytic activity.^185–188

As demonstrated by Ma et al., the Li-doping, along with variable concentration of Ag nanoparticles deposition, effectively tuned the optical absorption and altered the band alignment of g-C₃N₄.⁸³ Li-doping and Ag deposition exhibited significantly higher photocatalytic performance, degrading RhB around 85% within 30 min while maintaining high photostability compared to pristine g-C₃N₄, Li-doped g-C₃N₄, and Ag/g-C₃N₄. The enhanced photodegradation efficiency of RhB was primarily ascribed to the improved charge separation induced by the modified band alignment due to Li incorporation, causing a positive shift of the valence band potential (from +1.57 V to +1.99 V), thereby promoting a more favorable oxidation process. Ag nanoparticles shrank the photogenerated e⁻ from the CB level of g-C₃N₄ by forming a Schottky barrier at the Ag/g-C₃N₄ interface. These e⁻ readily reacted with adsorbed O₂ to generate O₂˙⁻ radicals, contributing to pollutant degradation.

Another example of photodegradation and mineralization of RhB by utilizing visible light active Na-doped g-C₃N₄ was reported by Zhang et al.⁸⁴ Their study includes Na-doped g-C₃N₄ exhibiting a promotional effect in the photoactivity, achieving a RhB degradation rate 3.5 times higher than the pristine g-C₃N₄. This could be ascribed to the incorporation of Na-doping, which facilitated enhanced surface area, a shift in band edge potential, reduced band gap, and the photogenerated charge separation. The mineralization efficiency was determined by estimating the total organic carbon (TOC). The improved mineralization efficiency was attributed to the tuning of the band structure, altering the optical energy level and expanding π-delocalization upon Na doping. As a result, Na-doped g-C₃N₄ exhibited a 3.5-fold enhancement in the RhB degradation rate over pristine g-C₃N₄.

In the next study, Sudrajat et al. reported the visible-light-driven photocatalytic mineralization of 17α-ethynylestradiol (EE2) over a highly photo-active Na-doped g-C₃N₄.⁸⁵ The Na-doped g-C₃N₄ displayed a remarkable enhancement in the photoactivity compared to the neat g-C₃N₄, achieving 95.7% EE2 degradation with 89.4% mineralization efficiency in the first cycle. This enhancement resulted from electronic modification in g-C₃N₄ induced by Na-doping, which generated a new energy state in between the CB and VB, thereby altering the band structure, facilitating the separation of photogenerated charge carriers, and reducing their recombination. These combined effects drove the superior degradation and mineralization of EE2. The CB of Na-doped g-C₃N₄ favored the conversion of O₂ into O₂˙⁻ radicals as the active species. This demonstrated the beneficial role of Na in enhancing EE2 mineralization. Another case study suggested that the incorporation of K in g-C₃N₄ effectively reduced the recombination tendency of excitons and enabled the degradation of naphthalene efficiently.⁸⁷ This advancement in photoactivity was primarily attributed to the cooperative effect of K doping and the integration of biochar. Biochar often acts as a cocatalyst, offering effective charge transportation and open active sites for adsorption of pollutants, providing a promotional effect of dispersion in polar solvents due to its hydrophilic nature. It also functioned as an electron acceptor to mitigate the charge recombination. An optimized mass ratio of K-doped g-C₃N₄/biochar (1 : 0.8) achieved 82.19% naphthalene degradation within 180 min, demonstrating excellent photostability for more than five cycles. Similarly, Zhang et al. demonstrated a K-doped g-C₃N₄ photocatalyst for the photo-assisted degradation of MB and phenol.⁸⁹ It was found that the K-doping altered the electronic structure as well as improved the overall photocatalytic performance, resulting in a 3.3 and 5.8-fold higher photodegradation activity of MB and phenol, respectively, than pristine g-C₃N₄ (Fig. 28a). By the K incorporation in K-doped g-C₃N₄, the flat band potential was uplifted by 0.2 V, exhibiting an anodic shift in the valence level, which enhanced the oxidizability nature of the photocatalyst for efficient photodegradation activity. Moreover, MB degradation activity increased with increasing ‘K’ content up to 22%. Later, the photodegradation activity got reduced upon further increasing the K content as more defective sites were generated, which might cause the rapid charge recombination and hinder the condensation process. The smaller arc radius observed in the EIS Nyquist plots (Fig. 28b) indicated the improved charge distribution and transfer efficiency in K-doped g-C₃N₄. This is corroborated by the photocurrent response along with reduced PL intensity (Fig. 28c).

Fig. 28 (a) Rate constants of photocatalytic degradation of phenol, (b) EIS Nyquist plots, and (c) photocurrent responses of pure g-C₃N₄ and K-C₃N₄ with varying KI concentrations. Reproduced with permission from ref. 89. Copyright 2015, Elsevier. (d) Photocatalytic degradation of RhB over the synthesized catalysts, (e) UV-vis DRS spectra, and (f) PL spectra of GCN-0, GCN-Na, GCN-K, GCN-Rb and GCN-Cs samples. Reproduced with permission from ref. 95. Copyright 2017, Royal Society of Chemistry.

K-doped g-C₃N₄ also exhibited efficient RhB degradation activity under visible light irradiation.⁹⁰ The photo-degradation activity followed first-order kinetics with a rate constant of approximately 6.4 times greater than that of neat g-C₃N₄. The enhanced photo-degradation was due to the synergistic effect of K-doping, which increased the surface area, decreased the band gap energy, and promoted easy charge separation. Similarly, Xu et al. compared the photodegradation of RhB using various alkali metal-doped (Na, K, Rb, and Cs) g-C₃N₄ samples.⁹⁵ Under identical experimental conditions, neat g-C₃N₄ decomposed 45% of RhB in 40 min; in contrast, GCN-Na, GCN-K, GCN-Rb, and GCN-Cs exhibited degradation efficiencies of 16%, 83%, 79%, and 52%, respectively (Fig. 28d). GCN-K demonstrated the most pronounced enhancement in activity among all alkali metal-doped samples. From the analysis of band gap energy and UPS spectra, a negative shift in both CB and VB potentials was detected, which favored the photoreduction process of O₂, generating superoxide radicals (O₂˙⁻), along with excellent stability. In contrast, GCN-Na showed the least efficiency, attributed to the negligible structural, optical, and electronic modifications by Na doping (Fig. 28e and f). Another example demonstrated the utilization of an Na, Cl-co-doped photocatalyst in the photodegradation of RhB.¹⁰⁹ The resultant Na, Cl-co-doped g-C₃N₄ (CN–NaCl-0.5) photocatalyst was synthesized by one-step thermal condensation of dicyandiamide and NaCl. The dual incorporation of Na and Cl led to several promotional effects, such as improved crystallinity of the material, increased inter-spacing distance of the adjacent layers, red-shift optical absorption, a narrow band gap with positive shifting of the CB and VB band positions, reduced charge recombination, and better charge separation for maximum photoresponse. As a result, CN–NaCl-0.5 exhibited 97.5% of visible light-mediated RhB degradation in 50 min, whereas neat g-C₃N₄ exhibited a degradation rate of 55.5% only. The next study demonstrated the co-doping of both K and Na into the lattice of g-C₃N₄ for the photodegradation and mineralization of RhB.¹⁸⁹ In this work, melamine was mixed with the different contents of the eutectic mixture of NaCl and KCl (at their melting point ∼476 °C) and underwent thermal condensation at 520 °C for 2 h. The resultant K, Na co-doped g-C₃N₄ showed excellent textural properties, due to the template effect of the employed molten solution of NaCl–KCl. Both Na and K were coordinated into the g-C₃N₄ lattice via ion–dipole interaction between the alkali metals and triazine moieties of g-C₃N₄. The ionic radii of Na (100 pm) and K (150 pm), which are higher than that of C (70 pm) and N (65 pm), didn’t facilitate the substitutional doping; however, the nitrogen pot having an in-planar distance of 0.71 nm facilitated the interstitial doping of Na and K by forming Na–N and K–N binding sites via ion–dipole interaction. Furthermore, the band structure was optimized due to the co-doping of Na and K, resulting in a favorable CB and VB edge potential for smooth generation of O₂˙⁻ and ˙OH active intermediates to catalyze the oxidative degradation of RhB, with 10 times higher photoreactivity than pristine g-C₃N₄.

Sun et al. reported alkali metal-doped g-C₃N₄ nanosheets (M–CN, M = L, Na, K) synthesized through copolymerization of melamine and alkali metal hydroxides for photocatalytic degradation of organic pollutants.⁴² PXRD confirmed the increased interlayer spacing and reduced crystallinity due to alkali ion intercalation. SEM and TEM analyses revealed that doping reduced the grain boundaries and introduced a porous structure, which alters the morphology of g-C₃N₄, thereby enhancing the SSA. This increased SSA of M–CN provides more active sites to facilitate diffusion and mass transfer of reactants. XPS analysis revealed a strong M–N coordination within the g-C₃N₄ framework, altering the electronic structure and charge distribution. UV-vis DRS spectra (Fig. 29a) indicated that M–CN exhibited enhanced visible light absorption and a higher absorbance and tuned the band structure by reducing the band gap and charge carrier recombination, which was demonstrated by PL and TRPL spectra. Transient photocurrent and EIS data further confirmed that K–CN has a superior conductivity and lower charge-transfer resistance. The integration of M⁺ into the g-C₃N₄ framework increased the local negative charge density, attracting h⁺ and repelling e⁻, thereby impeding the photoexcited charge carrier's recombination. These structural, electronic, and optical properties collectively improve the photoactivity of M–CN. The RhB degradation efficiency of all the synthesized catalysts under visible light irradiation followed the order: K–CN (90%), Na–CN (83%), Li–CN (79%), with K–CN exhibiting an enhanced photocatalytic performance and retaining 85% efficiency after 5 cycles, with structural stability (Fig. 29b–d).

Fig. 29 (a) UV-vis absorption spectra of CN and M–CN samples, (b) photocatalytic degradation kinetics of RhB, (c) comparison of RhB degradation rates by CN and M–CN, and (d) evaluation of catalytic stability of K–CN over five consecutive degradation cycles. Reproduced with permission from ref. 42. Copyright 2024, Royal Society of Chemistry.

Taking advantage of oxygen functionalization, enhanced charge transport through K-intercalation was investigated in the activity of K, O co-doped g-C₃N₄ samples towards photocatalytic NO removal, investigated by Li and coworkers.¹⁹⁰ Due to oriented charge flow across the layers, there is an accumulation of photoelectrons required for the generation of active species for photocatalytic redox reactions. The intercalated K-atom as a mediator exhibited oriented charge transfer to strengthen the interlayer interaction (Fig. 30a). Based on DFT calculations, the band structure of g-C₃N₄ (CN) and O-doped g-C₃N₄ (OCN), as shown in Fig. 30b and c, indicated the role of O atoms as adjusters and K-atoms as a mediator, resulting in a favorable band offset and spatial charge carrier separation. The potential energies of neat g-C₃N₄ and K, O co-doped g-C₃N₄, as shown in Fig. 30d and e, indicated an increase after the incorporation of O and K atoms. This difference in potential energy provides a driving force to transfer e⁻ between the layers through the K-atom as a channel. Furthermore, the role of O-adjuster atoms was further studied using Bader effective charge analysis, suggesting the change in the local charge density in OCN layers due to the continuous flow of charge from adjacent C and N to the O atom (Fig. 30f and g). This electron accumulation near the O atom and the promotional role of the intercalated K atom could alter the internal vdW force to strengthen the interlayer charge transfer for better charge separation (Fig. 30h and i). Hence, the availability of a greater number of e⁻ on K, O co-doped g-C₃N₄ samples favored the generation of O₂˙⁻ for the photooxidation of NO. Furthermore, according to ref. 39, K-doping in g-C₃N₄ experienced a more positive shifting of the VB edge potential position than Na-doping, and thereby the VB level of K-doped g-C₃N₄ possessed a stronger oxidizability, resulting in a higher NO removal efficiency than Na-doped g-C₃N₄.

Fig. 30 (a) Schematic representation of the internal van der Waals heterostructure (IVDWH): illustrating the “Cake Model” and structural framework of OCN-K-CN, (b) calculated total density of states (TDOS), (c) band alignment diagram of the OCN-K-CN IVDWH, (d) and (e) layered electrostatic potential distribution, (f) and (g) Bader charge distribution for pristine CN and OCN-K-CN, (h) and (i) charge density difference mapping of K-CN and OCN-K-CN, where blue, green, red, and gold spheres correspond to N, C, K, and O atoms, respectively. Reproduced with permission from ref. 190. Copyright 2018, American Chemical Society.

One of the major problems in the food industry and drinkable water bodies is the contamination of nitrite (NO₂⁻), which usually occurs due to the reckless dumping of various inorganic and organic pollutants into the environment. Moreover, its over-concentration can lead to various potential threats and health hazards to mankind. Therefore, it is absolutely necessary to detect the NO₂⁻ via a facile route. For instance, Na and K-doped g-C₃N₄ was utilized for the electrochemical sensing of NO₂⁻.¹⁹¹ The experimental outcomes suggested that Na-doping created abundant electroactive sites, which were responsible for the highest redox current response, facilitated high conductivity, and very good electron transfer ability that enabled efficient electrochemical oxidation of NO₂⁻. Similarly, uranium(VI) contamination is currently becoming a global threat due to the enormous increase in nuclear activity, mining, industrialization, etc. Among various decontamination approaches, photocatalytic reduction of a highly soluble U(VI) contaminant to the less soluble U(VI) is a relatively promising approach to mitigate the uranium contamination. For instance, a K-doped, –C≡N functionalized defective g-C₃N₄ (KHCN) photocatalyst was fabricated via a facile alkali etching process with KOH for the removal of U(VI).⁴⁴ The alkali etching enhanced the SSA of KHCN, which improved the adsorption of uranium, significantly accelerating the photocatalytic process. The XRD analysis revealed a notable attenuation in the intensity of the (002) plane of KHCN, suggesting a reduction in the interplanar stacking distance and diminished crystallinity. Additionally, the (100) plane became Table 5 weaker and broader, indicating structural disorder of tri-s-triazine units and the presence of defects in the CN framework. Furthermore, the structural modification due to doping and defect creation was confirmed by SEM, FT-IR, and XPS analyses. Compared to CN, KHCN samples exhibited surface corrosion and abundant voids due to –C≡N induced defects. KHCN-2.5 showed a ladder-like thin-layered edge structure, implying partial layer exfoliation, confirmed by TEM. This suggested that –C≡N groups interrupted heptazine growth, forming fragmented nanosheets or debris. PL analysis revealed that KHCN exhibited a reduced intensity compared to pristine CN, indicating suppressed recombination of charge carriers due to the introduction of K⁺ ions and –C≡N groups. KHCN also exhibited a red-shift absorption edge compared to CN, extending visible-light absorption due to a narrow band gap (2.28 eV). Furthermore, from XPS-VB spectra, the more negative CB potential of KHCN facilitated the O₂ reduction to O₂˙⁻, which favors the formation of an immobilized (UO₂)O₂·2H₂O adduct on the catalyst surface to execute the photocatalytic removal of U(VI). Hence, synergistic incorporation of K⁺ ions and –C≡N groups in the CN framework effectively augmented the electronic structure, improved charge dynamics, and boosted visible light absorption, enabling superior U(VI) removal efficiency.

In addition to the above examples, several studies on alkali metal-doped–g-C₃N₄-based photocatalysts have demonstrated excellent activity in pollutant degradation and removal of hazardous species as outlined in Table 5.

Table 5 List of alkali metal-doped g-C₃N₄-based photocatalysts for efficient pollutant degradation

Photocatalysts	Precursors	Synthesis method	Reaction condition	Pollutants	% of Degradation	Ref.
K–CN	Melamine, KOH	Thermal polymerization	500 W Xe lamp	RhB	90%	42
LiK15:5	Urea, LiCl, KCl	Calcination	500 W Xe lamp	Methylamine	93.1%	79
KTCN/ZnO NWs	Melamine, KCl, 2,4,6-triaminopyrimidine (C4H7N5)	Calcination	Xe lamp	MB	99.25%	192
ZnO/NaTCN	Melamine, NaCl, 2,4,6-triaminopyrimidine (C4H7N5), Zn(CH3COO)2·2H2O	Calcination	Xe lamp	MB	98.54%	193
CsGCN1	DCDA, CsCl	Thermal condensation	20 W CFL bulbs	AR 94 dye	93.4%	115
0.4-PCN	Melamine, K2CO3	Thermal polymerization	300 W Xe lamp	RhB, CFH	RhB (93%), CFH (33.4%)	185
K–g-C3N4	Melamine, KCl	Thermal polymerization	8 W VUV lamp	TMP	91%	187
3-CN	Urea, (Fe(NO3)3·9H2O), NaNO3	Calcination	Dark	MB	100%	188
Ni-MOF/Na–CN-4@MS	Melamine, NaCl, Ni(NO3)2·6H2O, H2BDC	Calcination	300 W Xe lamp	MB	99%	194
KPCN/GO/ZnFe2O4	Melamine, (NH4)2HPO4, KCl	Thermal polymerization, hydrothermal	500 W halogen lamp	RhB, MB, TC	RhB (96%), MB (93%), TC (87%)	195

---

3.3. Miscellaneous applications

This review also demonstrates how the influence of the alkali metal in g-C₃N₄ can stretch the opportunities in various exciting applications other than solar fuel production and contaminant degradation to justify its practicality. Several research groups have already demonstrated the advantageous aspects of g-C₃N₄ in the photocatalytic oxidation process.^196–203 However, alkali doping further exhibits a promotional impact due to the appropriate band alignment and electron transport channel, which efficiently facilitates the photocatalytic oxidation of various biomass compounds and organic molecules. A binary composite of K intercalated g-C₃N₄ and Bi₂MoO₆ (BMO/K–CN) forming a Z-scheme heterojunction was reported for the oxidation of para-methoxybenzyl alcohol under simulated solar irradiation.²⁰⁴ The successful composite formation was confirmed from the PXRD observation (Fig. 31a). Furthermore, the FESEM analysis suggested the hexagonal Bi₂MoO₆ (BMO) deposited on the lamellar-type sheets of K-doped g-C₃N₄ (K–CN). The successful interfacial contact was confirmed by the XPS analysis, suggesting that K doping occurred in the interstitial position beset with the heptazine unit and coordinated with N as N–K bonding sites, due to which K–CN experienced charge delocalization, encompassing better charge separation. Comparing the O 1s XPS profiles of K–CN and BMO/K–CN, the latter experienced a down shift, which clearly suggested the chemical interaction of BMO and K–CN. Similarly, after the composite formation, the up shift in the binding energies of C, N, and the down shift in the binding energies of Bi, Mo, and O of BMO/K–CN strongly suggest the flow of e⁻ from K–CN to BMO in BMO/K–CN. As a result, an appropriate band structure corresponding to the Z-scheme heterojunction of BMO/K–CN was constructed. The Z-scheme heterojunction was further confirmed by the terephthalic acid and DMPO test. BMO/K–CN exhibited red-shift light absorption with respect to BMO but maintained a similar type of light absorption trend with K–CN, as confirmed by the DR UV-vis analysis (Fig. 31b). The large interfacial contact area between BMO and K–CN promoted enhanced electron transfer ability as well as better charge separation efficiency. As a result, BMO/K–CN exhibited a one-pot conversion of para-methoxybenzyl alcohol oxidation and CO₂ reduction without any additional oxidant or co-catalyst. BMO/K–CN could be able to achieve 31% para-methoxybenzaldehyde as the oxidation product and 21 mmol g⁻¹ h⁻¹ of CO as the CO₂ reduction product in 6 h under simulated solar light irradiation. All spectroscopic evidence and control experiments suggested that, upon photoexcitation, the photoexcited e⁻ in the CB of K–CN moved towards the h⁺ in the VB of BMO; therefore, the remaining e⁻ in the CB of BMO utilized in the reduction of CO₂, while the h⁺ in the VB of K–CN oxidized para-methoxybenzyl alcohol (Fig. 31c).

Fig. 31 (a) XRD profile and (b) UV-vis DRS spectra of K–CN, BMO, and 5BMO/K–CN, and (c) picture illustrating the mechanism for CO₂ photoreduction coupled with p-MeOBA oxidation over the 5BMO/K–CN photocatalyst. Reproduced with permission from ref. 204. Copyright 2025, American Chemical Society.

Another example of efficient photooxidation of substituted benzylamine and simultaneous production of H₂O₂ was demonstrated by Liu et al.¹¹⁶ The K and Na-doped g-C₃N₄ nanosheet (K/Na/PCN) was first synthesized via ionothermal polycondensation of KCl, NaCl, and DCDA, in which K and Na ions were stabilized by the N atoms of heptazine units. Being a promising photoactive material with appropriate CB and VB edge potential, cobalt phthalocyanine (CoPc), having a planar structure, efficiently formed a Z-scheme heterojunction with K/Na/PCN through π–π stacking and H-bonding between the hydroxyl units of K/Na/PCN and CoPc (Fig. 32a). The resultant CoPc/K/Na/PCN, having an effective interfacial Z-scheme heterojunction interaction, experienced a strong interfacial intrinsic electric field, which significantly improved the charge carrier separation efficiency and thereby possessed enhanced redox ability of separated charge carriers to execute the efficient photoredox reaction. The photogenerated e⁻ of CoPc and the h⁺ of K/Na doped g-C₃N₄ combined, while the h⁺ of VB of CoPc and e⁻ of CB of the K/Na doped g-C₃N₄ remained separated to be utilized in the reduction of O₂ to O₂˙⁻ and abstraction of H⁺ from benzyl amine (PhCH₂NH₂) to PhCH₂NH₂˙⁺ respectively. The O₂˙⁻ coupled with H⁺ to form H₂O₂, and the resultant PhCH₂NH₂˙⁺ further reacted with the h⁺ to release another H⁺ and formed PhCH=NH as an intermediate species. The PhCH=NH underwent an addition reaction with another molecule of PhCH₂NH₂ to form N-benzylidenebenzylamine (N-BDBA) and NH₃ as the final product (Fig. 32b). Similarly, Dong et al. reported a highly efficient alkali-doped g-C₃N₄ photocatalyst, which was enriched with triune active sites such as alkali metal doping, –C≡N linkage, and N-vacancies for efficient ozone activation and atrazine degradation.⁷² Na, K and Li doped g-C₃N₄, designated as CN–M, M = Na, K and Li, facilitated the ozone adsorption and exhibited an enhanced electron transfer phenomenon due to the intercalation of alkali metals, which were reported to reduce the layer distance and acted as an electronic bridge. Moreover, the alkali doping created the –C≡N defect and N vacancies in the M–CN framework. The electron-withdrawing effect due to the –C≡N linkage facilitated the light harvesting by modulating the electronic structure, and the N vacancies promoted optimized charge distribution in M–CN. Therefore, the combinational effect of alkali metal doping, –C≡N linkage, and N-vacancies stimulated the M–CN framework for better ozone oxidation and atrazine degradation than undoped g-C₃N₄.

Fig. 32 (a) Schematic depiction of the preparation process and structural configuration of CoPc/K/Na/PCN Z-scheme heterojunction composites and (b) illustration of the proposed mechanism of the photooxidation of benzyl amine to N-BDBA. Reproduced with permission from ref. 116. Copyright 2025, Elsevier.

Furfural is a useful biomass-derived platform chemical that acts as a potential precursor for the production of various sustainable chemicals and fuels. The selective oxidation of furfural to furoic acid is an interesting and important transformation due to the wide range of applicability of furoic acid as a building block in the food industry, pharmaceuticals, and biofuel additives. Wang et al.²⁰⁵ and Liu et al.²⁰⁶ demonstrated the photocatalytic approach for furfural oxidation reaction using a composite of K-doped g-C₃N₄ and a photoactive NH₂-UIO-66 MOF (K_0.1-CN/UN-20). The former used Er doping as the promoter, and the latter used an aqueous solution of Na₂CO₃ externally as the mediator and promoter for the photooxidation process. Er and K doped g-C₃N₄ (Er@K–C₃N₄/UIO-66NH₂) exhibited 90.4% furfural oxidation with 63.4% furoic acid selectivity. The successful formation of the heterostructure and the K and Er doping not only enhanced the light harvesting ability but also promoted the electron transfer process to boost the charge carrier separation efficiently for efficient oxidation reaction (Fig. 33a). Similarly, the type II heterostructure between the K-doped g-C₃N₄ and NH₂-UIO-66 provided a strong interfacial contact in the heterostructure, accelerating excellent electron transfer ability. Again, the strategically tuned band gap width of the resultant heterostructure, when combined with the aqueous solution of Na₂CO₃, efficiently generated the desired reactive oxygen species (ROS) to facilitate the photocatalytic oxidation process of furfural. The photocatalytic process began with the excitation of charge carriers upon light irradiation, and the formation of the interfacial heterojunction promoted the charge separation by exhibiting band bending phenomena due to the generation of an interfacial intrinsic electric field (Fig. 33b). The photoexcited e⁻ in the CB of K–CN reduced O₂ present in air to O₂˙⁻ and subsequently formed H₂O₂ by reacting with aqueous H⁺. The in situ generated H₂O₂ preferentially oxidized furfural to furoic acid with 70.7% yield, while the Na₂CO₃ avoided the formation of unwanted ROS by selectively quenching the h⁺ in the VB of NH₂-UIO-66.

Fig. 33 Schematic illustration of the proposed reaction mechanism of furfural photo-oxidation over (a) Er@K–C₃N₄/UIO-66NH₂. Reproduced with permission from ref. 205. Copyright 2024, Elsevier and (b) K_0.1-CN/UN-20. Reproduced with permission from ref. 206. Copyright 2024, Elsevier.

Being a nitrogen-rich material, g-C₃N₄ possesses appropriate basicity and active sites for various types of biomass oxidation reactions and glucose isomerization reaction. The fabrication of alkali-doped g-C₃N₄-based active catalysts can be achieved by a molten salt-assisted synthesis method, which can induce the appropriate basicity in the g-C₃N₄ structure, to potentially enable the glucose to fructose isomerization reaction.²⁰⁷ The glucose to fructose isomerization was generally achieved in the presence of the Brønsted basic sites via a proton transfer mechanism. Based on this, several alkali metal-doped g-C₃N₄ materials were synthesized via the thermal condensation of urea and respective alkali/alkaline metal chlorides, as shown in Fig. 34a. The resultant materials were evaluated for the glucose to fructose isomerization reaction, and the K-doped (K–C₃N₄) and Mg-doped g-C₃N₄ (Mg–C₃N₄) exhibited the best activity in the glucose to fructose transformation. The PXRD and FT-IR analyses indicated the successful doping of alkali metal in g-C₃N₄ (Fig. 34b and c). The CO₂-TPD analysis confirmed that the basicity of the alkali/alkaline-doped material was significantly higher than that of the undoped g-C₃N₄ (Fig. 34d). Among various samples, Mg–C₃N₄ and K–C₃N₄ exhibited a broad CO₂ desorption peak at higher temperature (370–480 °C), suggesting that their stronger basic sites and higher overall basicity chemisorbed the CO₂ molecules. The time and temperature dependent catalytic activity test suggested that the Mg–C₃N₄ exhibited 32.5% fructose yield after 240 min at 90 °C, while Na–C₃N₄, K–C₃N₄, and Ca–C₃N₄ afforded 23.2%, 28.7%, and 28.4%, respectively.

Fig. 34 (a) Diagrammatic illustration of the synthesis of M-C₃N₄, (b) XRD patterns, (c) FT-IR spectra, and (d) CO₂-TPD of all the prepared samples. Reproduced with permission from ref. 207. Copyright 2022, American Chemical Society.

Similarly, another work achieved glucose to fructose isomerization via a photocatalytic approach using the Li and K co-doped g-C₃N₄ (CM-molten) synthesized via the molten salt-assisted thermal condensation of a melamine-cyanuric acid (MCA) supramolecular precursor (Fig. 35a).²⁰⁸ For comparison, the MCA supramolecular precursor was also pyrolyzed without the molten salt and designated as CM-air. The effect of the molten salt-driven pyrolysis was studied through several characterization processes. CM-air exhibited small-sized nanosheet morphology, whereas CM-molten exhibited a curled tubular shape consisting of micro-sized nanosheets with high crystallinity, confirmed by the existence of a well-defined (002) crystalline phase (Fig. 35b). High crystallinity of CM-molten was analyzed by PXRD analysis, suggesting the existence of various diffraction planes corresponding to poly(triazine imide). Additionally, the successful intercalation of Li⁺ in CM-molten was also validated by the PXRD analysis. In contrast, such crystalline phases were absent in the PXRD pattern CM-air, which demonstrated that the molten-assisted pyrolysis induced a high degree of crystallinity in the material (Fig. 35c). The molten-assisted synthesis was also responsible for the existence of –C≡N linkages, confirmed by the FT-IR and XPS analyses. This resulted in the enhanced light harvesting ability with red shift light absorption and a narrow bandgap energy compared to the CM-air (Fig. 35d). Moreover, these structural changes led to the existence of N vacancies and excess unpaired e⁻ in CM-molten as confirmed by the EPR analysis (Fig. 35e). As a result of the combinational effect of better light harvesting, improved charge separation, and extended functionalities due to the –C≡N linkages and N vacancies, the transient Lewis acid–base active sites could be created in CM-molten to accomplish the photocatalytic glucose to fructose isomerization (more than 25% conversion and ∼70% fructose selectivity) under visible light while CM-air exhibited very negligible activity of the glucose isomerization. As a result of the combinational effect of better light harvesting, improved charge separation, extended functionalities due to the –C≡N linkages and N vacancies, an efficient charge transfer pathway was achieved, as demonstrated in Fig. 35f. As a consequence, the transient Lewis acid–base active sites could be created in CM-molten to accomplish the photocatalytic glucose to fructose isomerization (more than 25% conversion and ∼70% fructose selectivity) under visible light, while CM-air exhibited very negligible activity of the glucose isomerization. The overall mechanism involved the activation of the aldehyde group by the Lewis acid sites, while the O atom at C-2 was activated by the Lewis base sites, resulting in the formation of an H-bonding intermediate, which further underwent deprotonation and H⁺ transfer pathways to produce fructose (Fig. 35g).

Fig. 35 (a) Schematic representation of CM-molten synthesis via supramolecular self-assembly assisted by a molten-salt approach for photocatalytic glucose isomerization, (b) FESEM image of CM-molten, (c) XRD, (d) UV-vis absorption spectra, (e) EPR spectra of CM-air and CM-molten, (f) picture illustrates the enhanced charge separation in CM-molten with oriented N-vacancies and –C≡N functionalities, and (g) glucose-to-fructose isomerization mechanism over CM-molten. Reproduced with permission from ref. 208. Copyright 2021, American Chemical Society.

Exceptionally, Li, among all the alkali metals, when doped in g-C₃N₄, exhibited excellent hydrogen adsorption efficiency and was suitable for hydrogen storage applications. Thus, Li-doped g-C₃N₄ sheets suitably drove efficient hydrogen adsorption. For instance, based on first-principles calculations on density functional theory, Zhu et al. reported the application of Li-doped g-C₃N₄ towards hydrogen storage under ambient thermodynamic conditions.⁶⁷ Their study revealed that porous open hollow sites of g-C₃N₄ sheets are found to be favorable to dope Li-atoms by coordinating with the N-atoms. The high electronegativity value of N-atoms not only induces a positive charge on Li-atoms but also polarizes H₂, enabling a strong binding (3.26 eV) with the substrate. The energy of hydrogen adsorption by Li-doped g-C₃N₄ was calculated as 3.48 kcal mol⁻¹ and studied as a good system having potential in energy storage. Similarly, another report by Song et al. proposed the first principles calculations for the reversible storage of H₂ over a super alkali NLi₄ cluster decorated g-C₃N₄ sheet ((NLi₄)₂@g-C₃N₄).²⁰⁹ The calculation results demonstrated that two NLi₄ ‘superalkali’ clusters stably anchored on a 2 × 2 g-C₃N₄ supercell in (NLi₄)₂@g-C₃N₄, having a total binding energy of −4.39 eV, and exhibited a total of 18 H₂ molecules adsorption with an adsorption energy of −0.29 eV at 200 K and 1 atm pressure. Charge analysis indicated that the optimized structure containing the NLi₄ cluster, which accommodated Li as Li⁺, temporarily polarized the H₂ and enhanced non-dissociative adsorption without forming strong chemical interactions. However, the temperature-dependent simulations further demonstrated that about 83.3% of the adsorbed H₂ got desorbed at 300 K and 1 atm, demonstrating a temperature-responsive, reversible storage behavior of H₂ that favored uptake at lower temperatures and promoted effective release near room temperature at atmospheric pressure, fulfilling a desirable combination for reversible H₂ storage applications. Besides, Li-doped graphitic carbon nitrides such as (g-C₃N₄ and g-C₄N₃) are also explored in Li ion batteries.²¹⁰ Melamine-derived triazine-based g-C₃N₄ having a single-layer nanosheet with a small pore size distribution can accommodate Li single atoms rather than cluster accumulation. According to the report, mono-layered Li-doped g-C₃N₄ acted as an anode material, while the bulk Li-doped g-C₄N₃ acted as a cathode material. Optimized Li-doped g-C₃N₄ can adsorb hydrogen in quasi-molecular form, having the H–H distance of 0.76 Å. The distance between the hydrogen molecule and the Li-atom was 2.5 and 2.18 Å in Li₂C₃N₄ and Li₂C₄N₃ systems, respectively, while the average distance between the H₂ molecule and the Li-atom was 2.86 Å and 2.43 Å. Every Li-atom has a tendency to adsorb three hydrogen molecules. Gravimetric densities of H₂ adsorption over Li₂C₃N₄ and Li₂C₄N₃ are 10.2 and 10.3 wt%, respectively, whereas the volumetric density is 100.2 g L⁻¹ in both cases.

4. Conclusion and future perspectives

This review consolidates current advancements in the design and functional tuning of alkali-metal-modified g-C₃N₄, emphasizing how these dopants reshape its structural, electronic, and photocatalytic behaviour (Fig. 36). The discussion highlights how mono-doping and co-doping have proven highly effective in tailoring the electronic structure, interfacial characteristics, and photoresponse behavior of g-C₃N₄-based materials. Smaller ions tend to substitute into the lattice, while larger ions more commonly intercalate into the layers or occupy cavity sites, generating increased porosity, disturbed stacking, and reduced crystallite size due to modified polycondensation paths. An in-depth demonstration of theoretical modellings and DFT studies is also provided in this review to underscore the charge density distribution, band structure analysis, and charge transfer mechanism, confirming that alkali ion incorporation leads to structural reorganizations that extend the optical absorption to red shift, narrowing the band gap, and optimizing the VB and CB levels for efficient utilization of charge carriers. Simultaneously, nitrogen vacancies are often formed during doping, acting as complementary active sites, enabling more efficient electron migration. By leveraging these synergistic effects, alkali-metal-doped g-C₃N₄ shows considerable promise for various sustainable and environmentally friendly photocatalytic processes, including solar fuel production, organic pollutant decomposition, biomass conversion, and energy-related applications. Insights drawn from case studies integrating morphological analysis, spectroscopic characterization, and mechanistic evaluation underscore the versatility and practical potential of these materials for efficient photocatalysis and energy harvesting. The insights summarized here provide a foundation for rational design of high-performance, energy-efficient photocatalytic systems based on alkali metal-engineered g-C₃N₄ nanostructures. Overall, we trust this review can become a reference tool for the fabrication of advanced and highly efficient alkali-doped g-C₃N₄-based energy harvesting materials.

Fig. 36 Schematic depiction of the key outcomes and future prospects of alkali metal-doped g-C₃N₄.

In contrast, several unsolved hitches and research challenges are highlighted as future prospects for advancing the alkali-doped g-C₃N₄-based photocatalysts in solar energy harvesting applications (Fig. 36). The respective research gaps need to be well addressed by making continuous efforts through extensive research in the domain of alkali doping in g-C₃N₄. Some of the points are highlighted below.

4.1. Rationalization of dopant behaviour

A deeper mechanistic insight into dopant location, bonding environment, and framework stability is needed. Since the appropriate doping of alkali metals provides a polarized charge density distribution for enhanced charge separation efficiency, a thorough investigation of advanced theoretical modellings, DFT calculations, simulations, and machine-learning-assisted design can help to predict the structure–property relationships, guide ideal dopant selection, and minimize trial-and-error experimentation.

4.2. Development of doping databases

Integrating such rationalized results from computational predictions with detailed experimental metadata (selection of precursors, synthesis parameters, defect states, morphology difference, structural outcomes, and photoactivity metrics) into a standardized database platform would accelerate the data-driven rational design of alkali-doped g-C₃N₄. Combining the database with an advanced high-throughput simulation process enables molecular-level estimation of control of dopant sites, structural configurations, photocatalytic performances, and product selectivity while maintaining very high reproducibility.

4.3. Device-level integration

The introduction of alkali metal causes substantial electronic, structural, and interfacial modifications. Moreover, considering the non-toxic, economic, and easily synthesizable nature of alkali metal-doped g-C₃N₄-based materials, it will be advantageous to expand their practical implementation in terms of progressive device design for different types of energy-related applications, such as PC and PEC electrodes, hybrid solar-driven reactors, ion storage and electrochemical energy-storage devices, and membranes, owing to their facile fabrication of device-grade films and composites using various promising techniques (powder suspensions, immobilized thin films and heterostructure formation).

4.4. Multi-parameter structural tuning

Currently, many research groups are continuously working on advancing the structural features of g-C₃N₄-based materials to achieve highly efficient, multifunctionality-integrated energy materials. However, more research is needed to fine-tune the structure and optoelectronic properties through multiple modifications and tailor the morphology to achieve highly efficient alkali-doped g-C₃N₄-based photocatalytic materials with the inherent quality of best photostability. In this regard, it is also crucial to develop a comprehensive database and advanced simulation methodologies that can serve as useful tools for future researchers to engineer at the molecular level of alkali-doped g-C₃N₄ materials, thereby precisely predicting and controlling their structural configurations, photoactivity, and product selectivity.

4.5. Multifunctional active site installation

Semiconductor photocatalysis has the potential to mimic natural photosynthesis; however, the major solar energy conversion process, such as CO₂ reduction, N₂ reduction, and water splitting reaction, often requires co-catalysts or sacrificial agents, which create complexities and difficulties, causing industrial infeasibility. Therefore, it is essential to modulate the structure and optoelectronic parameters in alkali metal-doped g-C₃N₄ to install multifunctional active sites by rationalizing the co-doping, precise defect management, and modifying physico-chemical properties for efficient solar fuel production with a co-catalyst and sacrificial agent-free condition, which is a more facile, sustainable, and economic pathway.

4.6. Scaling up sustainable synthesis

The development of alkali metal-doped g-C₃N₄ materials via a facile molten salt-assisted synthesis approach is one of the most straightforward, cheap, and sustainable approaches to advance the bulk g-C₃N₄. Therefore, it is crucial to optimize and advance this straightforward approach by rationalizing the experimental parameters (experimental conditions, types of precursors, etc.), process designing, and reactor modelling for the large-scale synthesis of alkali metal-doped g-C₃N₄-based materials in pilot plants with high reproducibility.

The precursors (melamine, urea, and dicyandiamide) used for the preparation of alkali metal-doped g-C₃N₄ are abundantly available and are of low-cost. Likewise, the synthesis process involves an energy-efficient one-pot thermal co-polymerization method, which involves simple equipment and a continuous feed process, demonstrating the high scalability. But the homogeneous distribution of alkali metals and the formation of ammonia gas during synthesis on a large scale is a challenge. Compared to the other doping systems, such as non-metal doping associated with control synthesis, transition metal doping attributed to the toxicity and leaching issues, alkali-metal-doped g-C₃N₄ ranks the best among various modification strategies, for industrial scale-up.

4.7. Collaboration and cross-disciplinary

Furthermore, it is essential to perform multi-collaborative research to strengthen future research by fabricating these structurally modified functionally rich alkali metal-doped g-C₃N₄ based materials using solar cell technologies for sustainable energy production applications.

Author contributions

Pratikshya Dash and Sulagna Patnaik share equal contribution for conceptualization, draft writing – review and editing; Sonali Panda and Bhagyashree Priyadarshini Mishra: review and editing: Kulamani Parida: supervision and review.

Conflicts of interest

There are no conflicts to declare.

Data availability

This review article only utilizes data assessment of the published articles by various researchers in this domain, which can be found from the sources referenced in the manuscript. No new data were generated during the preparation of this review. The required copyright permission has also been taken.

Acknowledgements

All the authors are thankful to Siksha ‘O’ Anusandhan (Deemed to be University) management for providing the necessary facilities and their constant support and motivation.

References

1. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
2. A. L. Linsebigler, G. Lu and J. T. Yates Jr, Chem. Rev., 1995, 95, 735–758.
3. C. Gao, J. Low, R. Long, T. Kong, J. Zhu and Y. Xiong, Chem. Rev., 2020, 120, 12175–12216.
4. J. L. Holechek, H. M. Geli, M. N. Sawalhah and R. Valdez, Sustainability, 2022, 14, 4792.
5. S. Parida and D. P. Dutta, in Biorenewable Nanocomposite Materials, Vol. 1: Electrocatalysts and Energy Storage, ed. D. Pathania and L. Singh, ACS Symposium Series, American Chemical Society, Washington, DC, 2022, vol. 1410, pp. 307–366.
6. A. Behera, A. K. Kar and R. Srivastava, Mater. Horiz., 2022, 9, 607–639.
7. M. Ashraf, M. Ayaz, M. Khan, S. F. Adil, W. Farooq, N. Ullah and M. Nawaz Tahir, Energy Fuels, 2023, 37, 6283–6301.
8. A. D. A. Bin Abu Sofian, H. R. Lim, H. Siti Halimatul Munawaroh, Z. Ma, K. W. Chew and P. L. Show, Sustain. Dev., 2024, 32, 3953–3978.
9. B. Weng, M.-Y. Qi, C. Han, Z.-R. Tang and Y.-J. Xu, ACS Catal., 2019, 9, 4642–4687.
10. S.-K. Lee, A. Mills and C. O’Rourke, Chem. Soc. Rev., 2017, 46, 4877–4894.
11. F. Wang, Q. Li and D. Xu, Adv. Energy Mater., 2017, 7, 1700529.
12. K. Mishra, N. Devi, S. S. Siwal, V. K. Gupta and V. K. Thakur, Adv. Sustainable Syst., 2023, 7, 2300095.
13. P. Kumari, N. Bahadur, L. Kong, L. A. O’Dell, A. Merenda and L. F. Dumée, Mater. Adv., 2022, 3, 2309–2323.
14. S. S. Mani, S. Rajendran, T. Mathew and C. S. Gopinath, Energy Adv., 2024, 3, 1472–1504.
15. A. Balapure, J. R. Dutta and R. Ganesan, RSC Appl. Mater. Interfaces, 2024, 1, 43–69.
16. F. Chen, T. Ma, T. Zhang, Y. Zhang and H. Huang, Adv. Mater., 2021, 33, 2005256.
17. M. Liras and M. Barawi, Chem. Soc. Rev., 2019, 48, 5454–5487.
18. L. Ding, Z.-D. Yu, X.-Y. Wang, Z.-F. Yao, Y. Lu, C.-Y. Yang, J.-Y. Wang and J. Pei, Chem. Rev., 2023, 123, 7421–7497.
19. P. Reiss, E. Couderc, J. De Girolamo and A. Pron, Nanoscale, 2011, 3, 446–489.
20. S. Ma, L. Jiang, W. Yang, F. Liu, D. Wang, F. Wang and J. Huang, Mater. Today Sustain., 2025, 29, 101023.
21. J. Fu, J. Yu, C. Jiang and B. Cheng, Adv. Energy Mater., 2018, 8, 1701503.
22. B. P. Mishra and K. Parida, J. Mater. Chem. A, 2021, 9, 10039–10080.
23. J. Wang and S. Wang, Coord. Chem. Rev., 2022, 453, 214338.
24. X. Qian, W. Li, X. Wang, H. Guan, Q. Bao, B. Zhao, B. Wulan, S. Liu, D. Zhu and X. Feng, Adv. Funct. Mater., 2025, 35, 2416946.
25. P. Dash, A. K. Kar, R. Srivastava and K. Parida, Mater. Horiz., 2025, 12, 6485–6557.
26. M. A. Khan, S. Mutahir, I. Shaheen, Y. Qunhui, M. Bououdina and M. Humayun, Coord. Chem. Rev., 2025, 522, 216227.
27. F. He, Z. Wang, Y. Li, S. Peng and B. Liu, Appl. Catal., B, 2020, 269, 118828.
28. X. Liu, R. Ma, L. Zhuang, B. Hu, J. Chen, X. Liu and X. Wang, Crit. Rev. Environ. Sci. Technol., 2021, 51, 751–790.
29. M. Rizwan, A. Zada, S. Azizi, P. Fazil, M. Naeem, G. Murtaza, N. A. Ibrahim, N. S. Basher, Z. Liu and M. H. Eisa, Int. J. Hydrogen Energy, 2025, 97, 1126–1152.
30. W. Zhang, D. Xu, F. Wang and M. Chen, Nanoscale Adv., 2021, 3, 4370–4387.
31. Y. Pang, P. Li, X. Ma, L. Sun, Y. Liu, D. Qu, L. An and Z. Sun, EES Catal., 2023, 1, 810–831.
32. Y. Zhang, J. Hou, J. Hu, H. Yu, M. Xu, Q. Chen, L. Liu, W. Yan and L. Shi, Appl. Surf. Sci., 2025, 164223.
33. C. Ma, Z. Hu, T. Zhao, Z. Gu, Q. Yuan and B. Chen, J. Mater. Chem. B, 2025, 13, 7528–7553.
34. X. Zhang, T. Li, C. Hu, X. Yan, K. Qiao and Z. Chen, Nanomaterials, 2022, 12, 2929.
35. Z. Gu, M. Jin, X. Wang, R. Zhi, Z. Hou, J. Yang, H. Hao, S. Zhang, X. Wang and E. Zhou, Catalysts, 2023, 13, 192.
36. A. Akhundi, A. Zaker Moshfegh, A. Habibi-Yangjeh and M. Sillanpää, ACS EST Engg., 2022, 2, 564–585.
37. J. Chen, S. Fang, Q. Shen, J. Fan, Q. Li and K. Lv, Catalysts, 2022, 12, 962.
38. H. M. Tofil, R. Ghazi, F. A. Ghaib, R. N. Dara, I. Kebaili, I. Boukhris, H. Ding and Z. Rehman, Sustainable Energy Fuels, 2025, 9, 2900–2927.
39. T. Xiong, W. Cen, Y. Zhang and F. Dong, ACS Catal., 2016, 6, 2462–2472.
40. L. Chen, S. Ning, R. Liang, Y. Xia, R. Huang, G. Yan and X. Wang, Int. J. Hydrogen Energy, 2022, 47, 14044–14052.
41. X. Zhong, Y. Zhu, M. Jiang, Q. Sun and J. Yao, Energy Fuels, 2023, 37, 5448–5456.
42. X. Sun, G. Zhao, S. Li, J. Liu and M. Huang, New J. Chem., 2024, 48, 9231–9242.
43. R. Mandal and S. Bhattacharyya, J. Phys. Chem. C, 2024, 128, 17342–17352.
44. Z. Dai, L. Chen, W. Zhang, B. Liang, Y. Gao and L. Li, Sep. Purif. Technol., 2024, 338, 126537.
45. J. Jiang, S. Cao, C. Hu and C. Chen, Chin. J. Catal., 2017, 38, 1981–1989.
46. P. Haripriya, T. Anjana, K. Sreelakshmi, N. T. Madhu, M. Anjana, P. V. Suneesh and D. V. R. Kumar, Catal. Commun., 2024, 187, 106909.
47. H. Liang, A. Wang, R. Cheng, F. Chen, P. Kannan, C. Molochas and P. Tsiakaras, Chem. Eng. J., 2024, 489, 151145.
48. J. Wu, J. Yu, F. Fan, R. Li, M. Wang, G. Li, Y. Wang, Y. Cui, D. Liu and Y. Wang, Green Chem. Eng., 2025, 6, 388–397.
49. N. Bai, J. Yin, X. Huo, Y. Ma, D. Guo and A. Wang, ScienceAsia., 2022, 48, 202.
50. H. Liang, J. Zhao, A. Brouzgou, A. Wang, S. Jing, P. Kannan, F. Chen and P. Tsiakaras, J. Colloid Interface Sci., 2025, 677, 1120–1133.
51. V. Bem, M. C. Neves, M. R. Nunes, A. J. Silvestre and O. C. Monteiro, J. Photochem. Photobiol. Chem., 2012, 232, 50–56.
52. M. Qamar, C. R. Yoon, H. J. Oh, N. H. Lee, K. Park, D. H. Kim, K. S. Lee, W. J. Lee and S. J. Kim, Catal. Today, 2008, 131, 3–14.
53. A. Turki, H. Kochkar, C. Guillard, G. Berhault and A. Ghorbel, Appl. Catal., B, 2013, 138, 401–415.
54. S. Patnaik, D. P. Sahoo and K. Parida, Carbon, 2021, 172, 682–711.
55. Y. Yang, W. Niu, L. Dang, Y. Mao, J. Wu and K. Xu, Front. Chem., 2022, 10, 955065.
56. D. Liu, A. Barbar, T. Najam, M. S. Javed, J. Shen, P. Tsiakaras and X. Cai, Appl. Catal., B, 2021, 297, 120389.
57. B. Li, T. Tian, Y. Zheng, D. Jiang, G. Xu, Y. Sun, Z. Li and Y. Yuan, Chem. – Eur. J., 2025, 31, e202500297.
58. J. Li, F. Xing, S. Liu, Z. Xiong, C. Wang, N. Li, H. Jin, Y. Su, C. Feng and J. Li, J. Colloid Interface Sci., 2025, 138565.
59. J. Li, W. Cui, Y. Sun, Y. Chu, W. Cen and F. Dong, J. Mater. Chem. A, 2017, 5, 9358–9364.
60. T. T. H. Nguyen, M. C. Le and N. N. Ha, Mol. Simul., 2021, 47, 10–17.
61. Y. Xia, H. Yang, W. Ho, B. Zhu and J. Yu, Appl. Catal., B, 2024, 344, 123604.
62. X. Wei, J. Shen, Z. Chen, H. Li, K. Zong, L. Yang, M. Jin, Z. Chen, X. Wang and Z. Chen, Int. J. Hydrogen Energy, 2025, 161, 150465.
63. H. Yu, R. Shi, Y. Zhao, T. Bian, Y. Zhao, C. Zhou, G. I. Waterhouse, L.-Z. Wu, C.-H. Tung and T. Zhang, Adv. Mater., 2017, 29, 1605148.
64. L. Zhu, X. Ma, N. Liu, G. Xu and C. Huang, Acta Phys.-Chim. Sin., 2016, 32, 2488–2494.
65. C. C. Nnadiekwe, U. Mustapha, I. Abdulazeez, K. Alhooshani and A. A. Al-Saadi, Surf. Interfaces, 2023, 38, 102852.
66. Q. Zhang, B. Wang, H. Miao, J. Fan, T. Sun and E. Liu, Chem. Eng. J., 2024, 482, 148844.
67. G. Zhu, K. Lü, Q. Sun, Y. Kawazoe and P. Jena, Comput. Mater. Sci., 2014, 81, 275–279.
68. Z. Shu, Y. Wang, W. Wang, J. Zhou, T. Li, X. Liu, Y. Tan and Z. Zhao, Int. J. Hydrogen Energy, 2019, 44, 748–756.
69. N. Jasni, A. Iqbal, M. N. Ahmad, H. Pauzi and M. H. Hussain, Mater. Today Proc., 2022, 57, 1154–1161.
70. B. Liu, X. Nie, Y. Tang, S. Yang, L. Bian, F. Dong, H. He, Y. Zhou and K. Liu, Langmuir, 2021, 37, 4859–4868.
71. Y. Zhang, S. Zong, C. Cheng, J. Shi, X. Guan, Y. Lu and L. Guo, Int. J. Hydrogen Energy, 2018, 43, 13953–13961.
72. S. Dong, L. Liu, W. Xu, H. Cheng, Z. He, J. Tang, D. Wang, L. Wang, S. Song and J. Ma, ACS EST Engg., 2024, 4, 3105–3117.
73. H. Li, Q. Wei, K. Lu, R. Tang, Z. Xiao, X. Liu, X. Cheng, P. Han, Z. Wang and N. Shang, Appl. Surf. Sci., 2024, 643, 158712.
74. W. Wang, P. Xu, M. Chen, G. Zeng, C. Zhang, C. Zhou, Y. Yang, D. Huang, C. Lai and M. Cheng, ACS Sustainable Chem. Eng., 2018, 6, 15503–15516.
75. J. Hu, X. Wu, X. Kong, S. Uemura, T. Kusunose, Y. Tanaka and Q. Feng, J. Mater. Chem. A, 2025, 13, 11605–11616.
76. X.-J. Lu, C.-Z. Yuan, S. Chen, J.-H. Li, I. Ullah, M. Qi and A.-W. Xu, Langmuir, 2024, 40, 11067–11077.
77. Q. Qi, Y. Li, H. Liu, B. Li, H. Wang, Y. Lu, W. Gao, Y. Tian, B. Guo and X. Jia, Appl. Catal., B, 2022, 319, 121864.
78. J. Tang, C. Guo, T. Wang, X. Cheng, L. Huo, X. Zhang, C. Huang, Z. Major and Y. Xu, Carbon Neutralization, 2024, 3, 557–583.
79. P. Xiong, Q. Li, Q. Tang, H. Wang and Z. Wu, RSC Adv., 2023, 13, 31820–31834.
80. S. A. Rawool, Y. Kar and V. Polshettiwar, Mater. Adv., 2022, 3, 8449–8459.
81. L. Yuan, W. Liu and W. Zhang, Environ. Sci. Pollut. Res., 2023, 30, 58276–58281.
82. L. Ruan, G. Xu, L. Gu, C. Li, Y. Zhu and Y. Lu, Mater. Res. Bull., 2015, 66, 156–162.
83. W. Ma, X. Wang, F. Zhang, X. Fei, X. Zhang, H. Ma and X. Dong, Mater. Res. Bull., 2017, 86, 72–79.
84. J. Zhang, S. Hu and Y. Wang, RSC Adv., 2014, 4, 62912–62919.
85. H. Sudrajat, J. Solid State Chem., 2018, 257, 26–33.
86. F. Guo, J. Chen, M. Zhang, B. Gao, B. Lin and Y. Chen, J. Mater. Chem. A, 2016, 4, 10806–10809.
87. F. Li and M. Lin, Int. J. Environ. Res. Public Health, 2020, 17, 2065.
88. Y. Wang, S. Zhao, Y. Zhang, J. Fang, Y. Zhou, S. Yuan, C. Zhang and W. Chen, Appl. Surf. Sci., 2018, 440, 258–265.
89. M. Zhang, X. Bai, D. Liu, J. Wang and Y. Zhu, Appl. Catal., B, 2015, 164, 77–81.
90. S. Hu, F. Li, Z. Fan, F. Wang, Y. Zhao and Z. Lv, Dalton Trans., 2015, 44, 1084–1092.
91. S. Wang, J. Zhan, K. Chen, A. Ali, L. Zeng, H. Zhao, W. Hu, L. Zhu and X. Xu, ACS Sustainable Chem. Eng., 2020, 8, 8214–8222.
92. J. Fan, H. Wang, G. Liao, C. Song and J. Zou, Sci. China: Technol. Sci., 2025, 68, 1620206.
93. C. Ye, P. Guo, X. Chen, Z. Zhang, Y. Guo, Z. Chen, H. Yang, D. Luo and X. Liu, Battery Energy, 2025, 4, 20240033.
94. B. Aoudi, J. Kahkeci, I. Sánchez-Montes, Y. Boluk and M. G. El-Din, Sep. Purif. Technol., 2025, 364, 132523.
95. Y. Xu, Y. Gong, H. Ren, W. Liu, L. Niu, C. Li and X. Liu, RSC Adv., 2017, 7, 32592–32600.
96. H. Zhang, Y. Tang, Z. Liu, Z. Zhu, X. Tang and Y. Wang, Chem. Phys. Lett., 2020, 751, 137467.
97. X. Leng, Y. Li, G. Xu, R. Ramaraghavulu, S. V. P. Vattikuti, N. Nguyen Dang and J. Shim, J. Alloys Compd., 2025, 1039, 182808.
98. G. Fu, W. Zhen, H. Wang, X. Zhou, L. Yang and J. Zhang, Molecules, 2024, 29, 5339.
99. W. Guo, J. Zhang, G. Li and C. Xu, Appl. Surf. Sci., 2019, 470, 99–106.
100. Y. Wu, W. Xu, Z. Wang, Y. Wang, Y. Zhang, W. Zhong, H.-L. Cai and X. S. Wu, Appl. Surf. Sci., 2021, 542, 148620.
101. Y. Wu, D. Yang, W. Xu, R. Song, M. Li, Y. Wang, B. Zhou, N. Wu, W. Zhong and H. Cai, Appl. Catal., B, 2020, 269, 118848.
102. K.-L. Chen, S.-S. Zhang, J.-Q. Yan, W. Peng, D.-P. Lei and J.-H. Huang, Int. J. Hydrogen Energy, 2019, 44, 31916–31929.
103. J.-M. Lee, Y.-J. Lee, J. H. Lee, R. Tan, J.-S. Kim, I. S. Cho, S.-J. Park and C.-G. Lee, Environ. Res., 2025, 264, 120314.
104. F. Bi, G. Ba, J. Yu, H. Hu, J. Ye and D. Wang, Nanomaterials, 2025, 15, 1021.
105. J. Gao, X. Lin, B. Jiang, H. Zhang and Y. Li, Hydrogen, 2025, 6, 45.
106. Q. Qian, X. Chen, G. Li, C. Zhang, Q. Bi and S. Xu, Appl. Surf. Sci., 2025, 709, 163839.
107. J. Yuan, N. Tian, Z. Zhu, W. Yu, M. Li, Y. Zhang and H. Huang, Chem. Eng. J., 2023, 467, 143379.
108. X. Xia, C. Xie, Q. Che and P. Yang, Langmuir, 2023, 39, 1250–1261.
109. K.-L. Wang, Y. Li, T. Sun, F. Mao, J.-K. Wu and B. Xue, J. Mater. Sci.: Mater. Electron., 2019, 30, 4446–4454.
110. J. Duhan, A. Bisht, H. Kumar and S. Obrai, Surf. Interfaces, 2024, 52, 104927.
111. D. Wang, S. Pu, Y. Chen, K. Lei, Y. Duan, L. Mao, X. Zeng, X. Luo, Y. Zhang, Y. Dong, J. Z. Zhang and Y. Sun, Sci. China: Chem., 2025, 68, 192–200.
112. N. Wahab, M. Shahab, M. U. Haq, A. U. Rahman, M. Israr, Z. Iqbal, R. Khan and A. Ali, Mater. Res. Bull., 2026, 194, 113801.
113. L. Yan, H. Gao and Y. Chen, ACS Appl. Nano Mater., 2021, 4, 7746–7757.
114. B. Debnath, S. Singh, S. M. Hossain, S. Krishnamurthy, V. Polshettiwar and S. Ogale, Langmuir, 2022, 38, 3139–3148.
115. A. Kumar, S. Kashyap, M. Sharma and V. Krishnan, Chemosphere, 2022, 287, 131988.
116. Y.-A. Liu, J.-F. You, Q.-X. Shi, X.-Z. Li, J.-R. Xia and Y.-H. Yang, Appl. Catal. Gen., 2025, 689, 120005.
117. H. Shen, M. Li, W. Guo, G. Li and C. Xu, Appl. Surf. Sci., 2020, 507, 145086.
118. Y. Wang, M. Wang, X. Zhang, X. Pan, Y. Cui, D. Liu, Y. Wang and W. Yao, J. Catal., 2024, 440, 115807.
119. D. K. Gorai, S. K. Kuila, A. Kumar, Md. I. Ahmad and T. K. Kundu, Appl. Surf. Sci., 2023, 623, 157031.
120. M. Fronczak, E. Tálas, Z. Pászti, G. P. Szijjártó, J. Mihály, A. Tompos, P. Baranowski, S. K. Tiwari and M. Bystrzejewski, Diamond Relat. Mater., 2022, 125, 109006.
121. H. Lu, X. Li, F. Li, X. Xu, R. Zhao, C. Xiong, Q. Hu, Z. Miao and M. Tian, J. Mol. Liq., 2022, 352, 118655.
122. J. Zheng, Z. Xu, S. Xin, B. Zhu and L. Nie, Dalton Trans., 2022, 51, 12883–12894.
123. D. R. Paul, R. Sharma, S. Singh, P. Singh, P. Panchal, A. Sharma, P. Devi and S. P. Nehra, Int. J. Hydrogen Energy, 2023, 48, 37746–37761.
124. H. Yang, Y. Zhou, Y. Wang, S. Hu, B. Wang, Q. Liao, H. Li, J. Bao, G. Ge and S. Jia, J. Mater. Chem. A, 2018, 6, 16485–16494.
125. B. Zhu, B. Cheng, L. Zhang and J. Yu, Carbon Energy, 2019, 1, 32–56.
126. M. Z. Rahman and C. B. Mullins, Acc. Chem. Res., 2019, 52, 248–257.
127. X. Li, C. Zhang, J. Geng, S. Zong and P. Wang, Molecules, 2025, 30, 630.
128. Y. Miseki and K. Sayama, Adv. Energy Mater., 2019, 9, 1801294.
129. A. Mishra, A. Mehta, S. Basu, N. P. Shetti, K. R. Reddy and T. M. Aminabhavi, Carbon, 2019, 149, 693–721.
130. V. Hasija, A. Kumar, A. Sudhaik, P. Raizada, P. Singh, Q. Van Le, T. T. Le and V.-H. Nguyen, Environ. Chem. Lett., 2021, 19, 2941–2966.
131. A. Augustin, P. Ganguly, S. Shenoy, C. Chuaicham, S. C. Pillai, K. Sasaki, A. F. Lee and K. Sekar, Adv. Sustainable Syst., 2024, 8, 2400321.
132. S. Sun, J. Li, J. Cui, X. Gou, Q. Yang, Y. Jiang, S. Liang and Z. Yang, Int. J. Hydrogen Energy, 2019, 44, 778–787.
133. J. Ma, W. Zhou, X. Tan and T. Yu, Nanotechnology, 2018, 29, 215706.
134. M. Ma, J. Li, X. Zhu, K. Liu, K. Huang, G. Yuan, S. Yue, Z. Wang and S. Qu, Carbon Energy, 2024, 6, e447.
135. Y. Guo, Q. Liu, Z. Li, Z. Zhang and X. Fang, Appl. Catal., B, 2018, 221, 362–370.
136. W. Fang, J. Liu, L. Yu, Z. Jiang and W. Shangguan, Appl. Catal., B, 2017, 209, 631–636.
137. S. Cao, Q. Huang, B. Zhu and J. Yu, J. Power Sources, 2017, 351, 151–159.
138. Q.-H. Zhu, Z. Chen, L.-N. Tang, Y. Zhong, X.-F. Zhao, L.-Z. Zhang and J.-H. Li, Int. J. Hydrogen Energy, 2019, 44, 27704–27712.
139. N. S. Lewis, Science, 2016, 351, aad1920.
140. K. Zhong, P. Sun and H. Xu, Small, 2025, 21, 2310677.
141. S. A. Fors and C. A. Malapit, ACS Catal., 2023, 13, 4231–4249.
142. D. Li, K. Wang, J. Li, Z. Li, H. Wang and Y. Wang, Nano Energy, 2025, 133, 110460.
143. R. J. Detz, J. N. H. Reek and B. C. C. Van Der Zwaan, Energy Environ. Sci., 2018, 11, 1653–1669.
144. X. Chang, T. Wang and J. Gong, Energy Environ. Sci., 2016, 9, 2177–2196.
145. Z. Kovacic, B. Likozar and M. Hus, ACS Catal., 2020, 10, 14984–15007.
146. K. Lai, Y. Sun, N. Li, Y. Gao, H. Li, L. Ge and T. Ma, Adv. Funct. Mater., 2024, 34, 2409031.
147. T. Wang, J. Lu, J. Chen, C. Wang, K. Li and Y. Mei, Catal. Sci. Technol., 2025, 15, 2938–2949.
148. H. Li and J. Sun, ACS Appl. Mater. Interfaces, 2021, 13, 5073–5078.
149. S. Liu, J. Yu, X. Chen, N. Li and Q. Zhou, Nanomaterials, 2025, 15, 1294.
150. R. Fang, Z. Yang, Z. C. Kadirova, Z. He, Z. Wang, J. Ran and L. Zhang, Appl. Surf. Sci., 2022, 598, 153848.
151. J. Wang, Y. Wang, G. Li, Y. Xiong, M. Zhang, S. Zhang and Q. Zhong, J. Colloid Interface Sci., 2021, 603, 210–219.
152. Z. Yong and T. Ma, Angew. Chem., 2023, 135, e202308980.
153. L. Wang, J. Zhang, Y. Zhang, H. Yu, Y. Qu and J. Yu, Small, 2022, 18, 2104561.
154. Y. An, T. Cai, W. Jiang, T. Lei and H. Pang, Green Chem., 2025, 27, 10478–10509.
155. J. Song, C. Li, X. Zhang, P. Ma and K. Zheng, ACS Catal., 2025, 15, 9058–9088.
156. S. Gou, C. Wang, Z. Gong, Q. Wang, X. Li, X. Chen, J. Wang, J. Ma and Y. Zhu, Sep. Purif. Technol., 2025, 354, 128829.
157. Z. Liu, X. Wang, P. Zhang, H. Zhang, S. Wang, G. Jiao, W. Wu and M. Wu, Chem. – Eur. J., 2025, 31, e202404003.
158. C. Zhuang, Y. Chang, W. Li, S. Li, P. Xu, H. Zhang, Y. Zhang, C. Zhang, J. Gao, G. Chen, T. Zhang, Z. Kang and X. Han, ACS Nano, 2024, 18, 5206–5217.
159. S. Zhou, S. Cheng, J. Han and M. Huang, New J. Chem., 2023, 47, 19063–19076.
160. S. Wu, H. Yu, S. Chen and X. Quan, ACS Catal., 2020, 10, 14380–14389.
161. Y. Pan, X. Liu, W. Zhang, B. Shao, Z. Liu, Q. Liang, T. Wu, Q. He, J. Huang and Z. Peng, Chem. Eng. J., 2022, 427, 132032.
162. M. Saniya, S. Shaheen, I. Sadiq and T. Ahmad, Small, 2025, 21, 2505991.
163. T. Žibert, B. Likozar and M. Huš, ChemSusChem, 2024, 17, e202301730.
164. S. L. Foster, S. I. P. Bakovic, R. D. Duda, S. Maheshwari, R. D. Milton, S. D. Minteer, M. J. Janik, J. N. Renner and L. F. Greenlee, Nat. Catal., 2018, 1, 490–500.
165. G. Zhang, Y. Li, C. He, X. Ren, P. Zhang and H. Mi, Adv. Energy Mater., 2021, 11, 2003294.
166. K. Ithisuphalap, H. Zhang, L. Guo, Q. Yang, H. Yang and G. Wu, Small Methods, 2019, 3, 1800352.
167. G. Qing, R. Ghazfar, S. T. Jackowski, F. Habibzadeh, M. M. Ashtiani, C.-P. Chen, M. R. Smith III and T. W. Hamann, Chem. Rev., 2020, 120, 5437–5516.
168. S. Lin, X. Zhang, L. Chen, Q. Zhang, L. Ma and J. Liu, Green Chem., 2022, 24, 9003–9026.
169. Z. Wang, X. Hu, Z. Liu, G. Zou, G. Wang and K. Zhang, ACS Catal., 2019, 9, 10260–10278.
170. P. Praus, ChemistrySelect, 2023, 8, e202204511.
171. L. Zhang, S. Hou, T. Wang, S. Liu, X. Gao, C. Wang and G. Wang, Small, 2022, 18, 2202252.
172. L. Li, P. Zhang, N. Li, T. Reyila, Y. Yu, X. Su, C. Peng and L. Han, J. Environ. Chem. Eng., 2024, 12, 112142.
173. G. Gu, K. Wang, N. Xiong, Z. Li, Z. Fan, S. Hu and X. Zou, Dalton Trans., 2019, 48, 5083–5089.
174. Y. Li, B. Wang, Q.-J. Xiang, Q. Zhang and G. Chen, Dalton Trans., 2022, 51, 16527–16535.
175. X. Chang, H. Fan, S. Zhu, L. Lei, X. Wu, C. Feng, W. Wang and L. Ma, Ceram. Int., 2023, 49, 6729–6738.
176. W. Li, Z. Hu, L. Song, Y. Lv, J. Liu, C. Lu and C. Xu, ChemistryOpen, 2025, 14, e202400431.
177. A. Singh, U. Alam, P. Chakraborty, B. Sundararaju and N. Verma, Chem. Eng. J., 2023, 454, 140303.
178. X. Tantai, Q. Zhou, L. Shi, M. Wu, P. Sun and X. Dong, ACS Appl. Nano Mater., 2025, 8, 19001–19011.
179. C.-W. Bai, L.-L. Liu, J.-J. Chen, F. Chen, Z.-Q. Zhang, Y.-J. Sun, X.-J. Chen, Q. Yang and H.-Q. Yu, Nat. Commun., 2024, 15, 4718.
180. L. Xu, L. Li, Z. Hu and J. C. Yu, J. Catal., 2023, 418, 300–311.
181. X. Yang, Y. Ye, J. Sun, Z. Li, J. Ping and X. Sun, Small, 2022, 18, 2105089.
182. Y. Luo, Y. Zhu, Y. Han, H. Ye, R. Liu, Y. Lan, M. Xue, X. Xie, S. Yu, L. Zhang, Z. Yin and B. Gao, Carbon Res., 2023, 2, 14.
183. Z. Tong, D. Yang, T. Xiao, Y. Tian and Z. Jiang, Chem. Eng. J., 2015, 260, 117–125.
184. J. Xu, Z. Wang and Y. Zhu, ACS Appl. Mater. Interfaces, 2017, 9, 27727–27735.
185. B. He, M. Feng, X. Chen, Y. Cui, D. Zhao and J. Sun, Appl. Surf. Sci., 2022, 575, 151695.
186. L. Wang, H. Yang, Q. Wu and R. Ma, Int. J. Hydrogen Energy, 2024, 80, 495–506.
187. W. Lu, A. Wang, Y. Zhang, S. Ren and Z. Zhang, Chem. Eng. J., 2024, 495, 153192.
188. S. Meng and Z. Nan, Sep. Purif. Technol., 2023, 309, 123026.
189. J. Zhao, L. Ma, H. Wang, Y. Zhao, J. Zhang and S. Hu, Appl. Surf. Sci., 2015, 332, 625–630.
190. J. Li, Z. Zhang, W. Cui, H. Wang, W. Cen, G. Johnson, G. Jiang, S. Zhang and F. Dong, ACS Catal., 2018, 8, 8376–8385.
191. P. Jigyasa and J. K. Rajput, J. Electroanal. Chem., 2020, 878, 114605.
192. J. Xing, Z. Liu, Z. Ruan, J. Lou, X. Yang and B. Huang, J. Mol. Struct., 2025, 1348, 143536.
193. Z. Liu, Z. Ruan, X. Yang, Y. Huang and J. Xing, Molecules, 2024, 29, 3240.
194. Z. Cao, H. Wu, S. Zhang, B. Zhai, Z. Qian, T. Wang, X. Zhou and C. Wang, J. Colloid Interface Sci., 2026, 704, 139312.
195. R. Kumar, A. Sudhaik, Sonu, P. Raizada, V.-H. Nguyen, Q. Van Le, T. Ahamad, S. Thakur, C. M. Hussain and P. Singh, Chemosphere, 2023, 337, 139267.
196. M. Bellardita, E. I. García-López, G. Marcì, I. Krivtsov, J. R. García and L. Palmisano, Appl. Catal., B, 2018, 220, 222–233.
197. X. Luo, B. Li, T. Su, X. Xie, Z. Qin and H. Ji, ACS Catal., 2025, 15, 1097–1111.
198. C. Xing, G. Yu, T. Chen, S. Liu, Q. Sun, Q. Liu, Y. Hu, H. Liu and X. Li, Appl. Catal., B, 2021, 298, 120534.
199. C. Xiao, L. Zhang, H. Hao and W. Wang, ACS Sustainable Chem. Eng., 2019, 7, 7268–7276.
200. M. J. Lima, P. B. Tavares, A. M. T. Silva, C. G. Silva and J. L. Faria, Catal. Today, 2017, 287, 70–77.
201. X. Sun, D. Jiang, L. Zhang and W. Wang, Appl. Catal., B, 2018, 220, 553–560.
202. Q. Yang, T. Wang, Z. Zheng, B. Xing, C. Li and B. Li, Appl. Catal., B, 2022, 315, 121575.
203. Y. Geng, D. Chen, N. Li, Q. Xu, H. Li, J. He and J. Lu, Appl. Catal., B, 2021, 280, 119409.
204. A. Verma, D. K. Chauhan and K. Kailasam, ACS Sustainable Chem. Eng., 2025, 13, 46–55.
205. H. Wang, L. Liu, J. Bian and C. Li, J. Fuel Chem. Technol., 2024, 52, 1617–1628.
206. L. Liu, L. Wang, J. Tan, C. Yu, J. Bian and X. Wang, J. Photochem. Photobiol. Chem., 2024, 451, 115484.
207. M. Laiq Ur Rehman, Q. Hou, X. Bai, Y. Nie, H. Qian, T. Xia, R. Lai, G. Yu and M. Ju, ACS Sustainable Chem. Eng., 2022, 10, 1986–1993.
208. J. Wang, H. Zhao, B. Zhu, S. Larter, S. Cao, J. Yu, M. G. Kibria and J. Hu, ACS Catal., 2021, 11, 12170–12178.
209. Y. Song, M. Hu, X. Liu, H. Cheng and Y. Li, Int. J. Hydrogen Energy, 2024, 83, 512–519.
210. M. Wu, Q. Wang, Q. Sun and P. Jena, J. Phys. Chem. C, 2013, 117, 6055–6059.

---