1 Introduction
Efficient utilization of solar energy has been widely regarded as a promising way to solve the increasing energy demand and environmental concerns caused by the burning of fossil fuels. Because the solar energy striking on the earth surface is about 1.3 × 105 TW/a, and far exceeds the global human energy consumption of 2013 (approximately 17 TW), which is predicted to at least double by 2050, by roughly four orders of magnitude [1,2]. In 1972, Fujishima and Honda discovered the water splitting phenomenon on an ultraviolet (UV) light irradiated TiO2 electrode [3], henceforth, using solar light driven water splitting to convert and store solar energy in chemical fuel (H2) has attracted worldwide attention.
Photovoltaic-electrolysis (PVE) [4,5], photoelectrochemical (PEC) catalysis [1,6], and photocatalysis (PC) [1,7] are the three main sunlight driven approaches to produce H2 from water. Therein to, PVE has the highest solar-to-hydrogen (STH) efficiency while it suffers from the high cost competitive to nowadays industrial methane steam reforming process. Techno-economical analyses of the PEC and the PC system predict that H2 production cost would approximate 4.10–10.40 $/kg and 1.60–3.20 $/kg, respectively, if a similar STH is achieved. Obviously, PC has the lowest cost which also meets the requirement of DOE predicted target H2 price (2.00–4.00 $/kg) [8]. Moreover, the particulate PC system is a more low-cost technology with the potential to enable large-scale application, because of the ready synthesis of the associated photocatalysts as well as the simple reactor and facility designs [9]. However, the low STH efficiency of the PC system is the main factor restricting industrial application. According to the prediction by the US Department of Energy (DOE), STH should at least reach around 10% to meet the basic requirement for large-scale utilization in an acceptable cost [7,8]. Up to now, the highest STH achieved is around 1%, which is far below the lowest requirement [8,10]. Rationally designing and developing more efficient photocatalysts to promote STH efficiency is still the primary demand of a photocatalytic H2 production based PC system.
Adding sacrificial reagents into the PC system to immediately consume photo-induced holes (electrons) is effective to prolong charge carrier lifetime, thus, providing enough time for charge carrier to participate in water reduction (oxidation) reaction to generate H2 (O2). Photocatalytic H2 (O2) production in the presence of sacrificial regents is also called sacrificial half-reaction. However, photocatalytic half-reaction might be a down-hill process due to the consumption of sacrificial reagent, namely, the solar energy may be not stored in the H2 produced [8]. Meanwhile, using sacrificial reagents to trap photo-induced charge may increase the cost of the H2 produced. Hence, it is necessary to photocatalyze water splitting at a high STH efficiency without bringing in any sacrificial reagents.
2 Basis of photocatalytic overall water splitting
Photocatalytic overall water splitting is an uphill reaction (H2O= H2 + 1/2O2, DG° = 237 kJ/mol) with increased Gibbs free energy. Generally, the process from input solar energy to the final output H2 energy can be divided into three primary processes as shown in Fig. 1. Step 1, semiconductor photocatalyst absorbs incident photons with a higher energy than the band gap and generates charge carriers (electron-hole pairs). Step 2, electron-hole pairs separate and transfer to the catalyst surface, and ready to react with adsorbed reactant molecules. Step 3, electrons and holes participate in H2 evolution reaction (HER) and O2 evolution reaction (OER), respectively. In Step 1, the conduction band (CB) minimum and valence band (VB) maximum of a photocatalyst should straddle the potential of H+/H2 (0 V versus NHE) and O2/H2O (1.23 V versus NHE). From the thermodynamic viewpoint, the smallest band gap of a photocatalyst for the one-step overall water splitting should be 1.23 eV. However, in practice, due to the existence of over-potential both in reduction and oxidation half reaction, the semiconductor band gap should be larger than approximaely 1.6 eV that can drive the one-step overall water splitting. Moreover, it is worth noting that if the band gap is larger than 3.2 eV, only UV light can be absorbed to drive water splitting reaction. However, the energy of UV light only occupies approximaely 5% of the total energy of solar spectrum. Consequently, in terms of absorbing and using solar energy as more as possible to fulfill industrial STH efficiency, developing highly efficient visible light responsive photocatalysts is the primary strategy to reach this goal. To this end, GaN:ZnO [11,12], SrTiO3:Rh/BiVO4 [10,13], (oxy)sulfide [14], g-C3N4 [15,16], etc. have been developed, all of which exhibit great potential for overall water splitting. Here, this review just focuses on recent progresses in g-C3N4 based photocatalysts in this field.
3 Brief introduction to g-C3N4
3.1 g-C3N4
An ideal C3N4 nanosheet consists of C and N atoms that form the tri-s-triazine (heptazine) or triazine periodic units by sharing N atoms (Fig. 2(a)). The C3N4 nanosheet further bounds with adjacent layers by Van der Waals interaction to form a graphite-like layered structure [17–19], also known as graphitic carbon nitride (g-C3N4, hereinafter labeled as CN). Although triazine based CNs have also been reported for photocatalysis, yet, tri-s-triazine based CN is energetically favored and more stable [20]. CN can be facilely synthesized through a thermal polymerization process of N-rich molecule monomers, such as urea, melamine, cyanamide, and dicyanamide, etc [16,21–23]. Ong et al. summarized the detailed discovery and development of CN based on a series of pioneering work [16]. In the present paper, in particular, cyanamide is taken out as the typical precursor to elucidate the intermediates evolution from initial precursor to final product during the thermal polymerization process. Wang et al. combined thermogravimetric analysis (TGA) with X-ray diffraction (XRD) to monitor the decomposition and formation of intermediates related to reaction temperature [24], according to which, in the temperature range from 203°C to 335°C, cyanamide started to condense to form dicyandiamide and further form melamine accompanied with the releasement of NH3. Subsequently, melamine molecules rearranged and constructed the tri-s-triazine unit at 390°C. The polymerization of tri-s-triazine commenced around 520°C and produces the polymeric CN. The polymeric product became unstable above 600°C, and completely decomposed beyond 700°C.
3.2 Basic physical and chemical properties
The band gap of bulk CN is experimentally determined as approximately 2.7 eV, enabling visible light absorption up to approximately 460 nm. Density functional theory (DFT) calculations and experimental investigations reveal that the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) are mainly composed of N pZ orbitals and C pZ orbitals, respectively, and straddle the H+/H2 and O2/H2O potentials. Therefore, on the one hand, the energy band structure is proper for overall water splitting (Figs. 2(b) and 2(c)) [25]. On the other hand, the light-trigged electrons and holes are mainly distributed around N and C atoms, respectively, indicating that the reduction and oxidation process for splitting water occur spatial independently [26]. In addition to suitable energy band structure, low cost, environmental tolerance, and facile synthesis methodology, the graphite-like layered structure and weak Van der Waals interaction between adjacent layers make CN aggregates to be easily exfoliated into single or few layered nanosheets. The CN nanosheets with a thinner lateral thickness render the photo-induced charge carriers sustaining short migration distance, representing a higher separation and transfer efficiency from the formation sites to the reaction sites. Moreover, nanosheets can provide more reaction sites and have the potential to construct functional nanohybrids with other components [27]. In view of the above discussion, CN nanosheets preparation and subsequent usage in catalytic applications have drawn numerous interest due to the unique merits of two dimensional (2D) materials [28–31].
In spite of the tremendous fascinating advantages of CN in photocatalytic water splitting, several disadvantages seriously restrict the STH efficiency improvement. In particular, the incomplete polymerization of amino groups leads to a high bulk defect density that indicates more electron-hole recombination centers. In addition, the charge transfer between two layers, bounded by the weak Van der Waals force, needs to overcome a large potential barrier [19]. Both of the above two factors heavily decrease the possibility of photo-induced charge carriers moving from the forming sites to the surface reaction sites. Therefore, researchers have made numerous efforts to address these issues for improving the catalytic performance, and the solar-driven overall water splitting is no exception. To the best of the authors’ knowledge, the review of the advancements of CN based photocatalysts for photocatalytic overall water splitting has not yet been reported. Therefore, it is of great significance to summarize the advances, which can, in turn, contribute to the development in this field. Additionally, it is also necessary to discuss future prospects and challenges of CN based catalysts for photocatalytic overall water splitting.
4 Recent progress on overall water splitting of CN based photocatalysts
The photocatalytic overall water splitting activity of pristine CN is poor due to the lack of redox reaction sites and serious charge carrier recombination caused by the packed layers and incomplete thermal polymerization of amino groups. To overcome this problem, numerous strategies have been proposed to acquire a better performance. State-of-the-art advances are reviewed and discussed in three subsections, including loading cocatalysts, microstructure regulation and modification, and constructing nanocomposites.
4.1 Loading cocatalysts
Generally, loading cocatalyst on the surface of photocatalyst is beneficial for accelerating photocatalytic water splitting due to the synergy function of improving charge separation efficiency and acting as redox reaction sites. For example, in the presence of HER cocatalyst like noble metal Pt, photo-induced electrons are extracted by Pt nanoparticles due to the larger work function compared to the semiconductor photocatalyst, while the photo-induced holes still remain on the photocatalyst. Thus, it efficiently improves the charge carrier separation efficiency. On the other hand, Pt is an excellent H2 evolution cocatalyst. The trapped electrons by Pt particles would be ready to reduce water into H2. Eventually, the introduction of cocatalysts enhances the photocatalytic water splitting performance. Regarding pristine CN, the lack of photo-redox reaction sites and serious photo-generated charge carrier recombination heavily limit its catalytic performance. Therefore, it has a poor H2 production activity in the half reaction, and almost no activity for photocatalytic overall water splitting. Consequently, in order to acquire the pure water splitting ability under solar light irradiation, integrating cocatalyst with CN is always an inevitable and efficient strategy. Recent advancement of cocatalysts modified CN for photocatalytic overall water splitting has been tabulated in Table 1.
4.1.1 Noble metals
Noble metals, such as Pt, Au, Ir, etc. have large work functions and thus possess excellent charge carrier trapping capacities [32]. Therefore, noble metal based materials are the most widely used cocatalysts for pure water splitting. For instance, in situ formed (111) facets-oriented noble Au nanoparticles on CN nanosheets achieves overall water splitting under visible solar light (λ>420 nm) [33]. Benefiting from the extended light absorption because of the surface plasmon resonance (SPR) of Au, the Schottky barrier depressed charge recombination, and the proper band level alignment, the highest H2 evolution rate of up to 150.1 μmol/(h·g) is obtained. In another case, IrO2 quantum dots (approximately 2 nm) are deposited on protonated CN by a facile solvothermal method and play the role of oxidation reaction promotor [34]. Although the mass of IrO2 is about 0.6 wt% in the hybrid, it, with a quite small size, can be highly dispersed and form an intimate contact with CN, thus significantly promoting the electron-hole separation efficiency and concurrently increasing the absorbance capacity of visible light. After IrO2 modification, the H2 and O2 production rates can reach 45 μmol/h and 20 μmol/h under light with wavelength (l>400 nm), respectively.
In addition to Au and Ir based materials, cocatalysts containing Pt are the most widely used candidates in photocatalytic overall water splitting. However, in the sole presence of Pt, catalytic activity is obviously poorer than that in the multi-cocatalytic system, as multi-cocatalysts can facilitate H2 and O2 production separately due to the presence of respective components. Wang et al. have successfully decorated CN with Pt, PtOx, and CoOx as redox reaction sites via photo-deposition process [35]. Pt acts as HER cocatalyst, simultaneously, PtOx and CoOx facilitate O2 evolution. At λ>300 nm light illumination, 12.2 μmol/h and 6.3 μmol/h of H2 and O2 evolution rates are obtained without adding any sacrificial reagents (apparent quantum yield (AQY): 0.3% at 405 nm). Further, they constructed another multi-cocatalytic system by CoP and Pt [36]. Dual cocatalysts forming heterojunctions with CN display a two times catalytic performance improvement compared to Pt/CN. After introducing CoP, the water oxidation process is evidently accelerated by promoting the charge carrier separation efficiency. Moreover, Co3O4 and Pt nanoparticles are separately deposited on the outer and inner surface of hollow CN spheres by using mesoporous-SiO2 as hard template (Figs. 3(a) and 3(b)) [15]. The spatially separated HER and OER reactive sites accept the unidirectional migration of electrons and holes on the Janus surface, thereby decreasing charge recombination and preventing the unwanted reverse reaction of water splitting. The delicately designed Co3O4/HCNS/Pt photocatalyst has a 3.1 μmol/h H2 evolution rate and a 1.5 mmol/h O2 evolution rate. The gas evolution rate is about 10 times faster than that of pure HCNS, and it is also much better than Co3O4 and Pt co-modified HCNS without spatial separation (Figs. 3(c) and 3(d)). The selective deposition of HER and OER cocatalysts for promoting overall water splitting performance is further supported by the Rh and RhOx spatially co-loaded CN [37]. Rh-RhOx/PCN shows the accelerated reaction kinetics and boosted photo-induced charge carrier separation, compared to that of Rh or RhOx solely modified CN. Hence, it is reasonable to infer that depositing noble metal based cocatalysts or further combining them with other cocatalysts is an efficient strategy to convert solar energy into H2 through water without adding sacrificial reagents.
Pt sub-nanoparticles which are deposited on alkali-metal-oxides that are intercalated in the CN layers, and Co3O4 particle co-deposition on CN can drive overall water splitting under visible light (λ>420 nm) [38]. In the hybrid, the intercalated alkali-metal-atoms bridging two adjacent layers as charge conductor enable more efficient charge transportation and separation. In addition to accelerating H2 production, Pt sub-nanoparticles also function as H2 permeable membrane and prohibit the permeating of O2, thus supressing their back reaction. With regard to the Co3O4 particles, they significantly facilitate the water oxidation process. In the Pt/Ni(OH)2 nanowires and Pt nanoparticle co-modified system [39], abundant coordinatively unsaturated sites of Pt/Ni(OH)2 effectively boost O2 evolution, and Pt-O-Ni interactions retard O-O bond cleavage, thus inhibiting the backward H2O regeneration. Furthermore, the Schottky junction formed by the isolated Pt with CN is beneficial for electron transfer and proton reduction. Ultimately, due to the improved charge separation, prohibited backward reaction, and tuned surface kinetics, stoichiometric H2 and O2 production rates of up to 26.60 and 12.64 μmol/h (AQY: 4.2% at 420 nm), are obtained respectively. Regarding another Pt decorated β-Ni(OH)2/CN [40], β-Ni(OH)2 nanolayers tune the water oxidation kinetics by decreasing OER overpotential and promoting the charge separation/transfer efficiency. Pt nanoparticles are in situ photo-deposited on CN surface, and significantly enhance H2 evolution rate. Therefore, overall water splitting is achieved under solar light irradiation. Interestingly, the photo-induced Pt decorated Ni(OH)2 can serve as a strong synergetic cocatalyst [41]. The intimate contact of nanosized Pt with Ni(OH)2 brings up a strong Ptσ + -Oσ− interaction, which enhances H2O activation along with accelerated charge transfer/separation efficiency. This unique structure prefers to H2O adsorption and further makes H-OH bonds more easily to be activated and cleaved. This work elucidates the importance of introducing H2O activation component into the pure water splitting system.
Despite the excellent cocatalytic performance, however, the scarcity of noble metals on the earth leads to the high-cost in large-scale application. To keep the prominent cocatalytic ability and simultaneously reduce their cost, cutting large particles into smaller clusters and even isolated single atoms is an alternative and promising method. To this end, single noble metal atom modified photocatalysts have been developed and extensively investigated in overall water splitting, half reactions, and other fields [42–47]. In this paper, only the examples relating to overall water splitting over CN are summarized. Liu et al. have successfully modified CN by the atomically dispersed heterogeneous Pt1Nx-Au1Nx redox single sites. The formation of p-n coupled micro-domains as effective bifunctional electron-donor-acceptor centers evidently improves the electron-hole separation efficiency by 6–10 times relative to the individual Pt/CN and Au/CN [48]. Concurrently, a drastically reduced surface redox reaction barrier by approximately 300 mV of the bifunctional Pt-Au/CN single site photocatalyst is achieved. Under 300 W Xe-lamp irradiation, the introduction of Pt-Au single sites makes the hybrid successfully split pure water with 8.55 μmol/h of H2 production rate (AQY: 3% at 420 nm) while it is of no activity for pristine CN.
Besides developing novel cocatalysts, addition of charge conductor (GO, carbon nanotubes, etc.) to connect photocatalyst and cocatalyst is proven to be effective for improving catalytic activity by tuning the charge transportation properties. One-dimensional carbon nanotube (1D CNT) modified 2D CN serves as a non-resonant plasmonic photocatalyst and produces H2 and O2 with the co-assistance of Pt and Ru [49]. The non-resonant plasmonic effect enhances the spectral response both in UV and visible light regions. Besides, the larger specific surface area, prolonged charge life time, and strong electronic coupling effect concurrently result in the considerable H2 and O2 production rate as 2.49 and 1.14 μmol/h, respectively.
4.1.2 Non-noble cocatalysts
Although noble metal single atoms or clusters has a prominent cocatalytic ability with significantly reduced cost, it is of more value to directly use earth-abundant elements to provide the same effect in good robustness. In the metal-free system, such as C dots embedded CN (Fig. 4(a)–4(f)), which is fabricated by heating the hybrids of ammonia treated C dots with urea [50], can stimulate photocatalytic overall water splitting through the stepwise two-electron/two-electron process: (i) 2H2O → H2O2 + H2, and (ii) 2H2O2→ 2H2O+ O2. Therein to, CN is regarded to be responsible for the photocatalysis step (i) and C dots contribute to the second step (ii). Additionally, C dots are also beneficial for the light absorption. Eventually, C dots embedded CN displays a fantastic pure water splitting activity with approximately 8.4 μmol/h of H2 production rate with H2:O2 as 2.02 (Fig. 4(g)). The AQY at 420 nm reaches 16% and STH is about 2% in the absence of any noble cocatalysts (Fig. 4(h)), representing that photocatalystic water splitting is rapidly proceeding toward industrial utilization [51]. In another case, to the same C dots/CN composite, it is synthesized by a thermal pyrolysis process, in which C dots are uniformly embedded into CN matrix and interacted by chemical bonding. The introduction of C dots effectively decreases the band gap from 2.84 to 2.08 eV. Meanwhile, the photo-induced electrons are rapidly extracted by the C dots while the holes are still remained on CN, thus resulting in their improved separation efficiency and prolonged life time. Interestingly, it is found that one-third of C dots work as H2O2 decomposition catalyst and the others do as H2 production cocatalyst.
Besides the metal-free cocatalysts, most reported coatalysts are synthesized based on transition metals. Among these, transition metal oxides are an important class of non-noble candidates. For instance, by optimising the amount of NiO and calcination temperature, amorphous NiO can efficiently accept photo-induced charge carriers from CN to improve the charge separation efficiency. Meanwhile, NiO provides more redox reaction sites and simultaneously contributes to H2O2 decomposition [52]. Therefore, under visible light (420 nm<λ<700 nm), the H2 and O2 production rates of up to 1.41 and 0.71 μmol/h are achieved, respectively. Similar roles of CoO have also been reported [53]. MnO2 can accelerate the OER process because of its good activity for H2O2 decomposition. Therefore, combining MnO2 with CN can obtain overall water splitting [54]. Moreover, in the MnO2 and CNT dual cocatalysts system [55], CNTs capture photo-induced electrons from CN and act as HER cocatalyst, and MnO2 particles catalyze the decomposition of H2O2 to facilitate O2 evolution. The strengthened charge separation efficiency and bringing of cocatalytic effect concurrently trigger the overall water splitting reaction under light irradiation (380–780 nm). Dual non-noble metal cocatalyts, like MnOOH as HER cocatalyst and Mn2Co2C@C as OER cocatalyst, co-deposited CN also present a prominent photocatalytic overall water splitting activity [56]. Anchoring Mn2Co2C@C on CN can form Schottky junction to efficiently capture photo-induced electrons, thereby facilitating charge separation/migration. MnOOH is in favor of water adsorption and concurrently reduces the O2 evolution overpotential. Therefore, a H2 evolution rate of 8.876 μmol/h is achieved, which is higher than that of 1 wt% Pt loaded CN.
In a recently reported typical example, a single-site Co1-phosphide configured CN is successfully prepared which exhibits the fantastic catalytic performance [57]. The P atoms partially substitute C atoms in triazine rings and contribute to the formation of Co1-P4 structure upon phosphidation (Fig. 5). The unique micro-structure effectively suppresses charge recombination and prolongs charge carrier lifetime by 20 times compared to pristine CN. Consequently, it shows a H2 evolution rate of up to 8.206 μmol/h (AQY: 2.2% at 500 nm) with a stoichiometric O2:H2 ratio of 1:2 under simulated sunlight (λ>300 nm). This work offers the guideline to design transition metal-based single-atom grafted overall water splitting photocatalysts. Furthermore, thermally polymerizing hexacarbonyl tungsten and melamine mixture can synthesize Pt-like WC1–x modified 3D porous CN with ultrathin layers and nitrogen defects [58]. The specific surface area, light absorption, and charge carrier separation efficiency of the obtained product can be significantly improved. Density function theory (DFT) calculations reveal that the coexistence of WC1–x and N defects makes electrons densely distributed around the Fermi level of CN, thus significantly accelerating the energy flowing from the light-induced excitons into H2. Ultimately, the H2 and O2 production rates of the hybridized photocatalyst are respectively measured as 3.364 and 1.668 μmol/h under visible light (l>420 nm) in a stoichiometric ratio of 2:1.
In summary, depositing cocatalysts to enable CN to have a photocatalytic overall water splitting activity is always inevitable and effective, and of course has achieved great progress. However, the STH efficiency is still far below the industrial need. Therefore, developing more efficient, low cost, and stable cocatalyst is still the urgent task.
Tab.1 Cocatalysts for photocatalytic overall water splitting over CN based photocatalysts |
| Cocatalyst | Light source | Efficiency | HER/(mmol·h−1) | H2/O2 | Mass/mg | Ref. |
|---|---|---|---|---|---|---|
| IrO2 | 300 W Xe lamp, λ>400 nm | AQY: 2.1% not mention the light wavelength | 45 | 2.25 | 30 | [34] |
| Pt, PtOx, and CoOx | 300 W Xe lamp, λ>300 nm | AQY: 0.3% at 405 nm | 12.2 | 1.94 | 200 | [35] |
| Pt and CoP | 300 W Xe lamp, λ>420 nm | NA | 2.1 | 2.10 | 80 | [36] |
| Pt and Co3O4 | 300 W Xe lamp, λ>300 nm | NA | 3.1 | 2.10 | 20 | [15] |
| Rh-RhOx | 300 W Xe lamp, λ>400 nm | AQY: 0.1% at 420 nm | 1.4 | 2.30 | 50 | [37] |
| PtMOx and Co3O4 | 300 W Xe lamp, λ>420 nm | AQY: 4.9% at 420 nm | 2.38 | 1.99 | 50 | [38] |
| Pt/Ni(OH)2 and Pt | 300 W Xe lamp | AQY: 4.2% at 420 nm | 26.60 | 2.10 | 20 | [39] |
| Pt and Ni(OH)2 | 150 W Xe lamp, 200 nm<λ<2500 nm | AQY: 1.48% at 405 nm | 15.5 | 1.99 | 200 | [40] |
| Pt/Ni(OH)2/CN | 300 W Xe lamp | AQY: 1.8% at 420 nm | 4.257 | 2.12 | 10 | [41] |
| Pt-Au single-sites | 300 W Xe lamp | AQY: 3% at 420 nm | 8.55 | 2.00 | 30 | [48] |
| Pt and Ru | Not mention, λ>350 nm | NA | 2.49 | 2.18 | 50 | [49] |
| C dots | 300 W Xe lamp, λ>420 nm | AQY: 16% at 420 nmSTH: 2% | ~8.4 | 2.02 | 80 | [50] |
| C dots | 300 W Xe lamp, λ>420 nm | NA | 0.25 | 2.00 | 50 | [51] |
| Amorphous NiO | 300 W Xe lamp, 420 nm<λ<700 nm | NA | 1.41 | 1.99 | 50 | [52] |
| CoO | 300 W Xe lamp, 420 nm<λ<700 nm | NA | 0.46 | 2.19 | 80 | [53] |
| MnO2 | 300 W Xe lamp, λ>420 nm | AQY: 3.82% at 420 nm | 5.53 | 2.00 | 100 | [54] |
| CNT and MnO2 | 70 W metal halide lamp, 380 nm<λ<780 nm | NA | 122 | About 2.00 | 30 | [55] |
| Mn2Co2C at C and MoOOH | 300 W Xe lamp | AQY: 1.45% at 20 nm | 8.876 | 2.05 | 150 | [56] |
| Co1-phosphide single sites | 300 W Xe lamp, λ>300 nm | AQY: 2.2% at 500 nm | 8.206 | 2.00 | 20 | [57] |
| WC1–x | 300 W Xe lamp, λ>420 nm | AQY: 11.24% at 420 nm | 3.364 | 2.02 | 40 | [58] |
4.2 Microstructure regulation and modification
Regulating or modifying the microstructure of a photocatalyst is efficient to modulate the morphology and acquire desired optical and electronic properties, in particular, the flexible strucutre of CN makes the numerous investigation on microstructure modification possible. Generally, the regulation of intrinsic characteristics renders the improvement of photocatalytic performance by promoting charge carrier transfer properties. However, it is not an effective way to introduce redox reaction sites in the CN based system. In another word, despite the modification, it still has to load other cocatalysts to achieve overall water splitting for the most reported cases, except for a few instances as reviewed below. The reported cases that structure modulation enables CN to split pure water by photocatalysis have been summarized in Table 2.
4.2.1 Nanostructure modulation
Three dimensional (3D) porous CN nanosheets (3D CNNS) can be synthesized by a self-assembly of cyanuric acid with melamine, following the freeze drying and thermal polymerization processes [59]. The 3D CNNS has an interconnected open-framework assembled by highly crystalline ultrathin nanosheets, thus exhibiting a quite large specific surface area and providing a pathway for faster charge transport and separation. It achieves a H2 and O2 production of up to 5.07 and 2.455 μmol/h (AQY: 1.4% at 420 nm), respectively. Moreover, via successive sealed condensation and hydrothermal treatment, a 3D sea-urchin-like CN has been synthesized with numerous defects (C≡N), resulting in a narrowed band gap of approximately 2.0 eV [60]. The 3D sea-urchin-structure enlarges the specific surface area, provides more reaction sites, and enables more efficient photo-induced charge separation and migration. DFT calculation confirms that the existence of -OH, -C= O, and -C≡N tunes the d-band center of Pt atoms, which is beneficial for a better stabilization of adsorbates and a higher catalytic ability. After loading 3 wt% of Pt, H2 and O2 production rates respectively reach 1.04 and 0.51 μmol/h under visible light. The thermal treatment of alkali and CN mixture is effective in creating in-plane pores due to the alkali etching effect [61]. The alkali etched CN has a significant improvement over specific surface area and separation efficiency of photo-induced electron-hole pairs relative to the pristine CN. However, excessive alkali usage would, in turn, decrease visible light absorbance and further the photocatalytic activity. After loading Pt (2 wt%), the optimal sample splits pure water to generate reduction product H2 and oxidation products O2 and H2O2. Without loading OER cocatalytst (C dots), it prefers to produce H2O2 instead of O2 because the two-electron process is much more thermodynamic preferable than that of four-electron process. With the co-assistance of C dots and Pt, visible light (λ>420 nm) driven photocatalytic H2 and O2 production rates reach 9.34 and 4.68 μmol/h, respectively, which is about 24.6 times that of CN. Regarding 2D nanosheets, recently, femtosecond pulsed laser was successfully used to exfoliate graphitic CN [62]. During the preparation process, incident photons weak the Van der Waals interactions to achieve single or few layered nanosheets. Simultaneously, they break strong C-N covalent bonds and cleave s-triazine units, thus introducing abundant -C≡N defects on the newly formed surface. Interestingly, -C≡N defects contribute to stabilizing Pt atoms, modulating the electronic properties, shifting down the band edge position. The highly dispersed Pt atoms provide more reaction sites and meanwhile supress the backward reaction of H2 and O2, whose production rates respectively reach 0.426 and 0.187 μmol/h in the presence of 1.4 wt% Pt. To the best of the authors’ knowledge, 1 D and 0 D CN based catalysts have not yet been applied for photocatalytic overall water splitting.
Further proceeding the modification into the molecular level, heterostructural carbon rings (Cring) have been successfully planted within CN layers (Fig. 6) [63]. The electron-hole pair separation and photoelectrons transportation are therefore effectively expedited because of the local in-plane π-conjugated electric field. The photo-carrier diffusion length and lifetime are synergistically elongated by 10 times relative to the untreated CN. Eventually, the (Cring)-CN gives a prominent H2 production rate of up to 11.13 μmol/h under full simulated solar light and a notable AQY of 5% at 420 nm. In addition, carbon-rich CN nanosheet (Cco-CN) is prepared by a combination processes, including the hydrothermal co-crystallization of terephthalic acid (TPA) with melamine (MA), and the following two-step thermal treatment and liquid exfoliation processes [64]. In the hydrothermal co-crystallization step, the orderly molecular self-assembly of the precursor tailors the chemical composition and electronic structure of the finally obtained Cco-CN. It exhibits an excellent photocatalytic overall water splitting ability by simultaneous producing H2 and O2 of up to 15.9 and 7.65 μmol/h in the absence of sacrificial reagents and cocatalysts. The improved separation efficiency, shortened transfer distance, and prolonged lifetime of photogenerated charge carrier are attributed to the promoted performance compared to that of pristine CN. Furthermore, the co-polymerization of MA, 2,4,6-triaminopyrimidine, and 1,3,5-triaminobenzene, containing few-carbon (pyrimidine) to all-carbon (benzene) aromatic rings, can substitute triazine rings by benzene aromatic rings [65]. The formation of N vacancies extends the π-electron conjugation in the 2D planar system and reduces the band gap to 2.1 eV with the negative shifts of energy band edges. Localized charge densities over valence band maximum (VBM) and conduction band minimum (CBM) in different regions of heptazine rings effectively decrease the charge carrier recombination rate. Ultimately, benzene ring incorporated CN directly splits pure water and generates about 7 μmol/h of H2 with an AQY of 1.6% at 450 nm. Additionally, CN deprotonation by a small fraction of Na+ renders the overall splitting water ability under visible light (λ>420 nm) [66]. The substitution of protons by Na+ completely prohibits the formation of H2O2, which is the dominant oxidation product in the case of pristine CN. The H2 and O2 production rates of Na-CN are 31.5 and 15.2 μmol/h (AQY: 1.45% at 420 nm; STH: 0.28%).
Moreover, elemental doping modulates the electronic structure of a material by introducing impurity energy levels into the band gap. It can effectively extend light absorption and tune chemical/physical properties, eventually endow desired properties for a specified photocatalyst. Consequently, elemental doping has been extensively investigated in photocatalysis [67–69]. However, it may not be a very useful method to make CN possess an overall water splitting ability because it has been rarely reported. Cui et al. synthesized B/K co-doped CN by using a recyclable molten-salt method and investigated its overall water splitting performance under visible light (λ>400 nm) [70]. B/K co-doping significantly modulates the HOMO and the lowest unoccupied molecular orbitals (LUMO), and thus leads to an enhanced photo-induced carrier separation efficiency. Additionally, co-doping also promotes the resistance of H2O2 toxicity of CN. Compared to CN and K doped CN, B/K co-doped CN exhibits the highest water vapor splitting performance with a H2 and O2 production rate of 1.18 μmol/h and 0.58 μmol/h, respectively.
4.2.2 Crystallinity regulation
Insufficient crystallinity and accompanying negative effects are the well-known limitations of CN in photocatalysis. Bulk defects can act as recombination centers and heavily restrict photo-induced charge separation and migration. Improving crystallinity is a useful strategy to decrease defect density and tune electronic structure, thus reducing the carrier recombination and modulating the light harvesting ability [71]. Crystalline CN has been exploited in half photocatalytic H2 production reaction [71–74]. The thermally calcining of the mixture of KCl, LiCl, and CN is frequently employed to improve crystallinity associating with enhanced electron-hole separation and migration efficiency. Wang et al. have comprehensively investigated the effect of pre-treated temperature on the crystallinity and catalytic activity of the final product [72]. Furthermore, exfoliating the good crystalline bulk into nanosheets can significantly improve the half H2 evolution rate to 53 μmol/h under visible light (λ>420 nm), which is 66 times that of pristine CN [74]. Other progresses on crystallinity CN can be referred to in Ref. [71]. Recently, crystallinity CN is developed for photocatalytic overall water splitting. For instance, Wang et al. successfully used surface kinetic control to photocatalyze pure water splitting by a covalent CN semiconductor with a crystalline poly(triazine imide) (PTI) framework (Fig. 7) [75]. The loading of a Pt is inevitable in water reduction and the co-deposition of a cobalt species can further increase the H2 and O2 production, as well enhancing the stability. Recently, the surface area ratios have been tuned and the reactive facets of crystalline CN have been identified [76]. The photogenerated electron-hole pairs can easily migrate along the conjugated layers to the {} facets, and the co-catalysts are mostly photodeposited on the {} facets due to the more-available electrons or holes on the prismatic surfaces. The above two factors confirm that the prismatic surfaces are the primary reactive facets. Moreover, PTI-550 gives a 189 μmol/h of H2 production rate under full spectrum light irradiation (AQY: 8% at 365 nm).
Tab.2 Nanosctructure modification of CN for photocatalytic overall water splitting |
| Sample | Light source | Efficiency | HER/(μmol·h−1) | H2/O2 | Mass/mg | Other cocatalysts | Ref. |
|---|---|---|---|---|---|---|---|
| 3D CNNS | 300 W Xe lamp, λ>420 nm | AQY: 1.4% at 420 nm | 5.07 | 2.07 | 50 | 1 wt% Pt 3 wt% IrO2 | [59] |
| 3D Sea-urchin-like CN | 300 W Xe lamp, λ>420 nm | AQY: 0.43% at 420 nm | 1.04 | 2.04 | 25 | 3 wt% Pt | [60] |
| Alkali etched CN | 300 W Xe lamp, λ>420 nm | NA | 9.34 | 2.03 | 10 | 2 wt% Pt CQDs | [61] |
| Laser exfoliated CN | 300 W Xe lamp, λ>420 nm | NA | 0.426 | 2.28 | 10 | 1.4 wt% Pt | [62] |
| (Cring)-CN | 300 W Xe lamp, λ>420 nm | AQY: 5% at 420 nm | 11.13 | 2.00 | 30 | about 3 wt % Pt | [63] |
| Cco-CN | 300 W Xe lamp | AQY: 5.28% at 400 nm | 15.9 | 2.08 | 30 | None | [64] |
| Benzene ring incorporated CN | 300 W Xe lamp, λ>450 nm | AQY: 2.1% at 450 nm | 7 | NA | 5 | 1 wt% Pt | [65] |
| Na-CN | 300 W Xe lamp, λ>420 nm | AQY: 1.45% at 420 nm STH: 0.28% | 31.5 | 2.07 | 100 | 1 wt% Pt | [66] |
| K/B co-doped CN | 300 W Xe lamp, λ>400 nm | NA | 1.18 | 2.03 | NA | None | [70] |
| PTI∙HCl | 300 W Xe lamp | AQY: 2.1% at 380 nm | about 65 | about 2.00 | 100 | 1% Pt 9 wt% CoOx | [75] |
| PTI-550 | 300 W Xe lamp | AQY: 8% at 365 nm | 189 | 2.08 | 120 | 1wt% Pt 0.5 wt% Co | [76] |
4.3 Constructing nanocomposites
Combining CN with other semiconductors can remarkably improve charge separation/migration efficiency and extend light absorption. To the authors’ knowledge, three types of junctions (Type II, Z-scheme, and p-n junction, shown in Fig. 8) have been successfully constructed over CN to photocatalyze pure water splitting. The three heterojunctions demonstrate a staggered band structure between two components. The CN composited photocatalysts for overall water splitting have been reviewed in Table 3.
4.3.1 Type II
In Type II heterojunction (Fig. 8(a)), photo-induced electrons in the more negative CB of the OER component are injected into the CB of the HER component, while the holes in the VB of the HER component would be transferred to the VB of the OER component. As a result, electrons and holes are spatially remained in respective reservoirs, therefore improving their separation efficiency and further the water splitting performance. However, the reduction ability of photo-induced electrons and the oxidation ability of holes would be weakened to some extent after separation, representing a smaller driving force for each reaction.
Metal oxide semiconductors are potential candidates to construct heterojunctions with CN for photocatalytic overall water splitting. For example, in the CoO/CN heterojunction, synthesized by the one-step solvothermal method, the 2D CN nanosheet offers a large specific surface area to load CoO particles to prevent their aggregation [77]. The intimate contact of two components results in the formation of Type II energy band alignment and notably enhances charge utilization efficiency. Importantly, the introduction of CoO keeps CN from being poisoned by the H2O2 formed. Therefore, complementary advantages of two components enable durable stability after combination. Identically, the CoO nanorod/CN composite also demonstrates overall water splitting in the presence of Pt under visible light [78]. Moreover, integrating two layered materials to fabricate large-area 2D/2D heterojunction with desired functions may realize the photocatalytic overall water splitting. To this end, the 2D/2D Co3(PO4)2/CN heterojunction, assembled by Coulomb electrostatic interaction, broadens the light absorption and concurrently accelerates the interfacial charge transfer, thus promoting the conversion of solar energy to H2 [79]. The optimal sample has a 18.78 and 8.86 μmol/h of H2 and O2 production rate, respectively, and exhibits a prominent stability and recyclability. Corresponding AQY is 1.32% at 420 nm. Recently, it is reported that in order to keep the equilibrium of Fermi level differences of LaOCl and CN (Fig. 9), charge redistribution would happen after tight contact of two components [80]. The built-in interfacial electric field related to charge redistribution accelerates charge separation and migration. Ultimately, with the co-assistance of Pt and CoOx, 22.3 and 10.7 μmol/h of H2 and O2 production rates are achieved under full-spectrum simulated light irradiation, respectively. Regarding the composites containing CN and anatase/rutile TiO2 (P25), the P25/CN multi-heterostructure displays a full-spectrum photocatalytic overall water splitting activity with a H2 evolution rate of 3.742 μmol/h (AQY: 0.96% at 350 nm and 0.71% at 400 nm), which is 8 and 4 times that of CN and P25, respectively [81]. The rutile TiO2 CB accepts electrons from the anatase TiO2 and CN, while holes accumulate on the CN VB. The strengthened charge separation promises more carriers to participate in surface chemical redox reactions. Furthermore, with the co-decoration of MnOx and Au, the Type II heterojunction of CN/TiO2 also splits pure water splitting under solar light [82].
Besides oxides, chalcogenides are another kind of potential materials to construct Type II photocatalysts with CN. For example, the binary Type II heterojunction of CdS/CN shows a pure water splitting capacity under visible light. Further, intercalating Ni2P between CdS and CN to function as a charge bridge can significantly accelerate the charge transfer from the CB of CN to that of CdS [83]. Therefore, a 0.778 and 0.388 μmol/h of H2 and O2 production rate over CdS/Ni2P/CN ternary junction is respectively obtained with a good stability. The activity is 4.02 times that of binary CdS/CN. In the hollow core/shell MnOx/CN/CdS/Pt heterostructure, prepared by a continuous chemical-annealing-photoreduction method, the hollow microstructure provides abundant sites for photocatalytic redox reactions. Besides, the Type II energy band alignment between CN and CdS prominently promotes the photo-generated charge carrier separation/transport efficiency. Moreover, Pt particles deposited on CdS quickly accept the light-induced electrons, simultaneously, MnOx excites hole-related species and accelerates H2O2 decomposition. Therefore, this delicately designed composite exhibits a remarkable overall water splitting performance with a 65.17 and 32.08 μmol/h of H2 and O2 evolution rate [84]. Similarly, the Pt and MnOx dual cocatalysts deposited CN/CdS system fabricated via in situ photo-deposition method also decomposes pure water under visible light (λ>400 nm), with H2 production rate up to 9.244 μmol/h and O2 production rate of 4.60 μmol/h [85]. Additionally, the coupling P doped CN with CdSe quantum dots (QDs) could significantly extend light absorption of up to 700 nm, thus producing more charge carriers under the same light irradiation compared to the pristine CN [86]. P doping shifts CN CBM to a lower position than that of CdSe QDs, enabling photo-induced electrons to flow from CdSe to P doped CN and therefore improving the charge carrier separation/migration efficiency. Extended solar spectrum utilization and more efficient charge separation make CdSe QDs/P doped CN photocatalyst possess an overall water splitting ability under visible light (λ>420 nm), where the H2 and O2 evolution rate are 5.65 and 2.78 μmol/h, respectively. In spite of the good overall water splitting performance of chalcogenides/CN photocatalysts, the chemical instability and serious photo-corrosion of chalcogenides are the two main constrains in large scale utilization. Moreover, the leakage of highly toxic heavy-metals (Cd, Pb, etc.) into environment would be detrimental to the environment and human health.
4.3.2 Z-scheme
Compared to the Type II system, the Z-scheme system enables a high charge-separation efficiency and simultaneously keeps a strong redox ability. As shown in Fig. 8(b), the photo-produced electrons in the OER catalyst CB are transferred to recombine with the holes in the HER catalyst VB via a redox mediator or at the interface. With respect to the CN based Z-scheme system, CN is always employed as the HER component due to its suitable potential for water reduction. In another word, a Z-scheme system can be constructed over CN by introducing an OER catalyst. Potential OER photocatalysts are Fe2O3, WO3, BiVO4, TiO2, MnO2, etc. At the beginning, Fe2+/Fe3+ and I–/IO3–, etc. are generally employed as liquid redox mediators. Then, all-solid-state Z-scheme with a better physical contact is developed to keep the highly efficient charge transfer of two components for better catalytic performance.
Inspired by the natural photosynthesis system, Tang et al. designed a photosynthesis-like composited overall water splitting system by combining CN with metal oxides (BiVO4 and WO3), with the assistance of redox mediators (FeCl2, FeCl3, and NaI) [87], in which, the best catalytic performance can be achieved when WO3 is used as the OER component. In detail, under the full spectrum irradiation of a 300 W Xe lamp, CN (3 wt% Pt) – NaI – WO3 (0.5 wt% Pt) displays a H2 and O2 production rate of 22.2 and 11.1 mmol/h, respectively. Under light with l>395 nm, the H2 and O2 production rate are 6.36 and 3.30 mmol/h, respectively. In addition to the effect of OER component, redox mediators also have notable influences on activity. Both in the cases of NaI and FeCl2 solution, water splitting reaction happens directly at the start of light irradiation. When FeCl3 is used as the redox mediator, no H2 could be detected until a 15 h irradiation due to the production of Fe2+/Fe3+ couple. In another case, the 3D mesoporous CN prepared by SiO2 hard template, BiVO4, and Fe2+/Fe3+ are respectively used as HER component, OER component, and redox mediator in a Z-scheme system [88]. Porous structure enlarges the specific surface area, hence providing more reaction sites and shortening the charge transfer distance. The porous structure is also beneficial for the Fe2+/Fe3+ pairs adsorption to collect the two components of a Z-scheme. Under full spectrum light irradiation, a 81.6 and 40.4 mmol/h of H2 and O2 evolution rate is respectively obtained.
Unlike the nature inspired photosynthesis II process, all-solid-state Z-scheme heterostructures are constructed without liquid redox mediators. Up to the present, all-solid-state CN based Z-scheme photocatalysts have been extensively investigated in the overall water splitting field. For example, the composite integrated by 2D α-Fe2O3 and 2D CN works as Z-scheme overall water splitting catalyst [89]. The 2D structure keeps short transfer distance of photo-induced charge carriers during the migration process from photocatalyst to reactant. At the same time, the intimate contact of two layered components enables fast and efficient electron transfer from α-Fe2O3 CB to CN VB, thus supressing the recombination possibility. In spite of the Z-scheme band alignment between Fe2O3 and CN, doping Mn into Fe2O3 crystal structure can evidently improve the conductivity, which is beneficial for charge transportation [90]. Accordingly, after combining Mn doped Fe2O3 nanocube with CN, the charge carrier density and the corresponding transfer efficiency are significantly improved, leading to the good overall water splitting performance. On the other hand, inserting an excellent charge mediator into HER and OER components is desired to thoroughly modulate charge transportation properties of a solid-state Z-scheme photocatalytic system [91]. As a proof of concept, Wang et al. selected reduced graphene oxide (rGO) as a solid conductor and successfully fabricated an all-solid-state Z-scheme system of CN/rGO/Fe2O3 [92]. Particularly, rGO is a 2D carbon material with a high specific surface area, an excellent electron transport mobility, and a chemically stable surface. The rGO nanosheets link the Fe2O3 nanoparticles and the CN nanosheets by chemical bonds and π-π stacking. The improved photocurrent density and evident PL quenching confirm that the rGO nanosheets remarkably accelerate interfacial charge transfer. Compared to CN, the CN/rGO/Fe2O3 has a 2-fold improvement of water splitting performance with a 43.6 μmol/h of H2 evolution. Generally, the water oxidation process is the rate-determining step of the overall reaction. Reasonably, adding cocatalyt to accelerate the OER process should be effective for improving the overall performance. In the CN/α-Fe2O3 Z-scheme system, depositing Co-Pi on α-Fe2O3 to collect and store holes successfully realizes the overall water splitting under visible light (λ>420 nm) [93]. However, the H2 production rate of 0.196 μmol/h is fairly slow. Then, the hierarchical Fe2O3@MnO2 core/shell nanocube was used to combine with CN to form a Z-scheme photocatalyst, which exhibits a 124 and 60 μmol/h of H2 and O2 production rate, respectively [94]. In this ternary heterojunction, Fe2O3 and MnO2, first forming a Type II structure, leads to the rapid transfer of photo-excited holes from MnO2 VB to Fe2O3 VB. Then, the Z-scheme band structure between Fe2O3@MnO2 and CN makes the electron in MnO2 CB to combine with the hole from CN. Eventually, this ternary composite gives a preferable photocatalytic H2 production of 124 μmol/h. In addition to Fe2O3, BiFeO3 as OER catalyst has been successfully integrated with CN to form a Z-scheme heterojunction [95]. The binary composite achieves H2 and O2 production from pure water under full spectrum light illumination due to the facilitated electron-hole separation and migration.
The redox cycling properties of multivalent metal cation can work as an electron mediator in the solid-state Z-scheme photocatalytic system. For instance, in the case of Mn3+ defective MnO2 modified CN heterojunction, in addition to the OER catalyst role of MnO2 (Fig. 10), Mn3+ cations containing in MnO2 on the one hand promote H2O adsorption and on the other hand simultaneously intrigue redox cycle reaction with Mn4+ [96]. The redox cycle reaction of Mn3+/Mn4+ enables a notably strengthened charge separation/migration efficiency. Finally, the H2 and O2 generation rate is realized as 1.212 and 0.578 mmol/h, respectively. The WO3∙H2O nanosheets can absorb sufficient visible light due to the approximate 2.4 eV band gap, and also is responsible for OER because of the suitable valence band edge potential of nearly 3.45 V [97]. Accordingly, the Van der Waals heterojunction of WO3∙H2O/CN has been facilely prepared by using the hydrothermal method to split pure water. In this Z-scheme system, the electrons in the CB of WO3∙H2O and the holes in the VB of CN recombine at the interface of the two components. On the other hand, the photo-induced holes of WO3∙H2O would accumulate on its VB, and the photo-induced electrons of CN would be retained on its CB. The evidently reduced charge carrier recombination therefore renders a considerable photocatalytic overall water splitting performance, in which the average H2 and O2 generation rates reach roughly 48.2 and 23.2 μmol/h, respectively. Compared to the CN/WO3 binary system, introducing rGO between the two components can further enhance the activity [98]. Furthermore, the 2D Z-scheme BiVO4/CN photocatalytst has been prepared by using a two-step hydrothermal/calcination method [99]. In this work, the electron spin resonance (ESR) analysis, by comparing the signal intensities of hydroxyl radical (∙OH) and superoxide radical (∙O2–) in the case of CN, BiVO4, and BiVO4/CN, was used to verify the Z-scheme mechanism. Finally, as aforementioned, TiO2 and CN can form Type II heterojunction to improve the photocatalytic performance [81,100]. Further, Liu and coworkers used Type II TiO2/CN heterostructure as the HER component to couple with an OER catalyst (WO3 and BiVO4) to successfully construct a Z-scheme multi-heterostructures for overall water splitting [101]. Compared to TiO2 or CN, TiO2/CN gives a more efficient H2 production activity in the half reaction. Eventually, the hybrid system, integrated TiO2/CN with β-Ni(OH)2/WO3, achieves an overall water splitting performance with a 50.2 and 24.3 μmol/h of H2 and O2 evolution rate (AQY: 2.06% at 425 nm), respectively, in the presence of HER cocatalyst (Pt) and redox mediator (5 mmol/L NaI).
4.3.3 p-n junction
Type II heterojunction can ideally separate electron-hole pairs, but the enhancement is always not sufficient to overcome the ultrafast electron-hole recombination. Consequently, in addition to the Z-scheme heterojunction, the p-n heterojunction concept has also been proposed to accelerate the electron-hole migration across the heterojunction by forming an interfacial electric field. Because of the Fermi level difference of p and n semiconductor, before light irradiation, the electrons on the n-type semiconductor tend to diffuse into the p-type semiconductor, meanwhile, the holes on the p-type semiconductor tend to diffuse into the n-type semiconductor, therefore respectively leaving a positively and negatively charged spaces at each side. As a result, the region close to p-n interface is charged. Under light irradiation and internal electric field, photo-generated electrons and holes will migrate directionally to the CB of n-type semiconductor and the VB of p-type semiconductor, respectively. Therefore, the electron-hole separation efficiency in p-n heterojunction is faster than that of Type II heterojunction due to the synergy effect of the strong internal electric field and band alignment [102,103].
In the research conducted in the past, p-n junctions for photocatalytic application are mainly developed over inorganic materials. Although CN is a typical n-type organic polymeric semiconductor, using it to construct p-n junction has attracted little concentration. Ye et al. found that the introduction of cyano terminal groups (C≡N) in situ transform CN from n-type to p-type, and thus forming the p-n homojunction [104]. The p-n homojunction of CN demonstrates better charge separation and transportation properties, and gives a 5 fold promoted photocatalytic H2 production performance. In addition, CN based p-n heterojunctions for half photocatalytic H2 production can refer to these work [105–107]. However, CN based p-n junctions are rarely reported for overall water splitting.
Recently, one piece of work relating to CN based p-n junction has been reported for photocatalytic overall water splitting [108], in which, the doping of P in CN crystal structure changes the n-type characteristic to p-type, and obviously decreases the surface energy by replacing original atoms. Furthermore, energy level redistribution makes the local charge carriers to be dominant by the holes. After integrating CN p-n homojunction with Ti3C2, the internal electric field at the p-n region and the secondary driving force between CN and Ti3C2 significantly prohibit the recombination and prolong the life time of charge carriers. Therefore, photocatalytic overall water splitting is achieved with a H2 and O2 production of 6.271 and 3.054 μmol/h, respectively.
Tab.3 CN based heterostructured photocatalysts for overall water splitting |
| Sample | Conditions | Efficiency | HER/(μmol·h−1) | H2/O2 | Mass/mg | Other cocatalysts | Ref. |
|---|---|---|---|---|---|---|---|
| CoO/CN | LED, λ>400 nm | NA | 2.51 | 1.81 | 50 | None | [77] |
| CoO Nanorod/CN | 300 W Xe lamp, λ>400 nm | AQY: 2.9% not mention the light wavelength | 54 | 2.16 | 30 | 3 wt% Pt | [78] |
| 2D/2D Co3(PO4)2/CN | 300 W Xe lamp, λ>400 nm | AQY: 1.32% at 420 nm | 18.78 | 2.12 | 50 | None | [79] |
| LaOCl-Coupled CN | 300 W Xe lamp | AQY: 0.4% at 380 nm | 22.3 | 2.08 | 50 | 0.5 wt% Pt and 0.2 wt% CoOx | [80] |
| P25/CN | Xe lamp | AQY: 0.96% at 350 nm | 3.742 | About 2.00 | 10 | 2 wt% Pt | [81] |
| MnOx/CN/TiO2/Au | 300 W Xe lamp, λ>420 nm | NA | 4.8 | 2.04 | 100 | None | [82] |
| CdS/Ni2P/CN | 300 W Xe lamp, λ>420 nm | NA | 0.778 | 2.01 | 50 | None | [83] |
| MnOx/CN/CdS/Pt | 300 W Xe lamp | NA | 65.17 | 2.03 | 50 | None | [84] |
| MnOx/CN/CdS/Pt | 300 W Xe lamp, λ>400 nm | AQY: 3.39% at 400 nm | 9.244 | 2.01 | 10 | None | [85] |
| CdSe QDs/P-CN | 300 W Xe lamp, l>420 nm | NA | 5.65 | 2.04 | 50 | 1 wt% Pt | [86] |
| CN (3 wt% Pt) – NaI – WO3 (0.5 wt% Pt) | 300 W Xe lamp | NA | 22.2 | 2.00 | 300 | None | [87] |
| CN (3 wt% Pt) – NaI – WO3(0.5 wt% Pt) | 300 W Xe lamp, λ>395 nm | NA | 6.36 | 1.93 | 300 | None | [87] |
| CN/BiVO4, Fe2+/Fe3+ | 300 W Xe lamp | AQY: 1.8% at 420 nm | 81.6 | 2.02 | 50 | 2 wt% Pt | [88] |
| 2D α-Fe2O3/CN | 300 W Xe lamp, λ>400 nm | NA | 1.91 | 2.00 | 50 | 3 wt% Pt 0.1 wt% RuO2 | [89] |
| Mn doped Fe2O3/CN | 300 W Xe lamp, λ>400 nm | NA | 51 | 2.05 | 30 | 1 wt% Pt | [90] |
| Fe2O3/rGO/CN | 300 W Xe lamp, λ>400 nm | NA | 43.6 | 2.06 | 40 | Pt | [92] |
| CN@a-Fe2O3/Co-Pi | 300 W Xe lamp, λ>420 nm | NA | 0.196 | NA | 20 | None | [93] |
| a-Fe2O3@MnO2/CN | 300 W Xe lamp, λ>190 nm | NA | 124 | 2.07 | 30 | None | [94] |
| CN/BiFeO3 | 125 W Hg lamp with an UV filter | NA | 0.933 | 2.01 | 40 | None | [95] |
| Mn defective MnO2/Monolayer CN | 300 W Xe lamp, λ>400 nm | NA | 1.212 | 2.10 | 20 | 3 wt% Pt | [96] |
| WO3∙H2O/CN | 300 W Xe lamp, λ>420 nm | AQY: 6.2% at 420 nm | 48.2 | 2.07 | 100 | None | [97] |
| WO3/rGO/CN | 250 W halide lamp, λ>420 nm | AQY: 0.9% at 420 nm | 2.84 | 1.95 | 200 | 1 wt% Pt | [98] |
| BiVO4/CN | 300 W Xe lamp, λ>400 nm | NA | 15.6 | 2.14 | 100 | 3 wt% Pt | [99] |
| TiO2/CN/WO3/β-Ni(OH)2/ | 150W Xe lamp; NaI as redox mediator | AQY: 2.04% at 425 nm | 50.2 | 2.07 | 100 | 1 wt% Pt PtOx | [101] |
| CN p-n homojunction/Ti3C2 | 300 W Xe lamp, λ>420 nm | AQY: 8.7% at 350 nm | 6.271 | 2.05 | 10 | 3 wt% Pt | [108] |
5 Summary and outlook
In summary, up-to-date progress of CN based photocatalysts for overall water splitting by modification methods as following three aspects have been reviewed, including loading cocatalysts to bring in water reduction and oxidation sites and improve charge carrier utilization, regulating and modifying microstructure to positively modulate intrinsic properties relating to water splitting reaction, and constructing multicomponent nanocomposites to use more sun light and improve separation/transportation of the electron-hole pairs. Of course, each strategy may affect the photocatalytic reaction process from different aspects. In the future, more attention should be paid from the following perspectives to further improve the STH conversion efficiency.
Improving light harvesting ability. The theoretical maximum of the STH efficiency of a photocatalyst highly relates to the band gap value. For the one-step photo-excitation system, the STH efficiency should reach 10% at least in terms of producing H2 with an acceptable cost. However, it is impossible to meet this target for a photocatalyst with an absorption edge of around 500 nm, even with an AQY of 100% [8]. Unfortunately, the gap value of CN is about 2.7 eV, corresponding the absorption edge of around 460 nm, which does not fulfill the basic requirement of light absorption. Consequently, it is of great value to extend the light absorption and further study the photocatalytic water splitting on the light absorption strengthened CN. In addition to self-modification, integrating with broader light absorption ability materials, such as plasmonics and small band gap semiconductors, etc. may be available to improve the solar light absorption capacity of CN based photocatalysts.
Developing efficient cocatalysts. Although many cocatalysts have been developed to facilitate overall water splitting, there still exists much space for further improvement. In the studies at present, noble metals are almost contained both in HER and OER cocatalysts. Therefore, designing noble metal free cocatalysts with an excellent cocatalytic performance and stability is crucial for large-scale application. On the other hand, the H2O2 formed during pure water splitting sometimes poison the catalysts, thus decreasing catalytic activity versus reaction time. That is to say, designing and fabricating a cocatalyst with the ability to rapidly decompose H2O2 and concurrently prohibit the back reaction of water splitting should be of great significance. In recent years, numerous researches have been conducted on pure water splitting for H2 and H2O2 production [109–110], where H2 separates automatically with H2O2. Therefore, this novel system realizes pure water splitting under solar light and concurrently avoids the separation of H2 and O2 relative to conventional system. Furthermore, compared to O2 production, H2O2 formation is more thermodynamically preferable because of the two-electron process. As a consequence, designing H2O2 formation cocatalysts may be applicable for STH improvement.
Theoretical prediction and in situexperimental investigation. Using DFT calculation to screen potential materials that endow more efficient charge separation/transportation and improved light absorption after combing with CN should be valuable. Of course, the theoretical method is also suitable for designing novel cocatalysts and predict property changes after modulating of a photocatalyst. Moreover, photocatalytic overall water splitting is still in a state of chaos, restricting the efficient photocatalyst design. Constructing in situ experimental apparatus to monitor the transient progress of water splitting should be useful but challenging, and theoretical calculation is a strong method to further explore the keys of mechanism beyond experiments.