Isotype heterojunction: tuning the heptazine/triazine phase of crystalline nitrogen-rich C₃N₅ towards multifunctional photocatalytic applications

Abstract

Photocatalytic technology has been well studied as a means to achieve sustainable energy generation through water splitting or chemical synthesis. Recently, a low C/N atomic ratio carbon nitride allotrope, C₃N₅, has been found to be highly prospective due to its excellent electronic properties and ample N-active sites compared to g-C₃N₄. Tangentially, crystalline g-C₃N₄ has also been a prospective candidate due to its improved electron transport and extended π-conjugated system. For the first time, our group successfully employed a one-step molten salt calcination method to prepare novel N-rich crystalline C₃N₅ and elucidate the effect of calcination temperature on the heptazine/triazine phase. Calcination temperatures of 500 °C (CC3N5-500) and 550 °C (CC3N5-550) lead to crystalline carbon nitride with both heptazine and triazine phases, forming an intimate isotype heterojunction for robust interfacial charge separation. An excellent photocatalytic hydrogen evolution rate (359.97 μmol h⁻¹; apparent quantum efficiency (AQE): 12.86% at 420 nm) was achieved using CC3N5-500, which was 15-fold higher than that of pristine C₃N₅. Furthermore, CC3N5-500 exhibited improved activity for simultaneous benzyl alcohol oxidation and hydrogen production, as well as H₂O₂ production (AQE: 9.49% at 420 nm), signifying its multitudinous photoredox capabilities. Moreover, the recyclability tests of the optimal CC3N5-500 on a 3D-printed substrate also showed a 92% performance retention after 4 cycles (16 h). This highlights that crystalline C₃N₅ significantly augmented the reaction performance for diverse multifunctional solar-driven applications. As such, these results serve as a guide toward the structural tuning of 2D metal-free carbon nanomaterials with tunable crystallinity toward achieving boosted photocatalysis.

---

Wee-Jun Ong

Wee-Jun Ong – Spearheading materials science, Materials Horizons has been a remarkable platform for reporting high-quality and ground-breaking developments in the past decade. Therefore, it has been a great honor to contribute this article to Materials Horizons on the structural engineering of novel two-dimensional crystalline nitrogen-rich carbon nitride C3N5 isotype heterojunctions for solar-driven multifunctional photocatalytic applications. Moreover, I am privileged to receive recognition to report an inaugural Materials Horizons Outstanding Review article in 2018. As one of the advisory boards for the journal, I would like to continue my long-term cooperation and my unwavering commitment to this excellent journal. Congratulations on its 10th anniversary and more excitements to be uncovered!

---

New concepts

Structural engineering of 2D nanomaterials is an appealing strategy for boosting electron transport along the in-plane direction attributed to their extended π-conjugated system and delocalized π-electrons. Since 2009, g-C₃N₄ has been widely studied as a low-cost and chemically stable photocatalyst. However, in recent two years, the N-rich C₃N₅ allotrope has gained momentum owing to its ample active sites and improved visible-light harvesting ability. For the first time, this study explores a novel approach for the C₃N₅ allotrope through a facile one-pot molten salt synthesis to engineer 2D porous crystalline C₃N₅ with a heptazine/triazine isotype heterojunction. Notably, the calcination temperature strongly affects the heptazine/triazine phase. At 500 and 550 °C, both heptazine and triazine peaks were observed in the X-ray diffraction pattern, indicating the presence of a robust isotype heterojunction that facilitates interfacial electron transfer. Furthermore, crystalline C₃N₅ exhibits markedly improved activity for the half-reaction of the hydrogen evolution reaction (HER), simultaneous photoredox HER and benzyl alcohol oxidation, and H₂O₂ production, ascertaining its multifunctional photocatalytic performances, outperforming various carbon nitride catalysts. Fascinatingly, the recyclability of crystalline C₃N₅ has been tested on a 3D-printed substrate over 16 h with 92% performance retention after 4 cycles. Hence, this is a testament to the feasibility of its application in multifarious catalytic as well as potential scale-up reactions.

Introduction

The conversion of solar energy to renewable hydrogen energy has been an important topic due to ongoing global warming and climate concerns.^1,2 Traditional non-renewable fossil fuel reserves are also rapidly depleting, which has led to a global search for green power generation to reduce our reliance on fossil fuels.^3–9 Currently, hydrogen production relies majorly on fossil fuels, thereby contributing to around 830 million tonnes of CO₂ emissions per annum.¹⁰ Therefore, global support for hydrogen energy research is on the rise due to the declining cost of renewable energy.

As non-metallic semiconductor photocatalysts, 2D carbon nitrides have attracted attention due to their superior stability, high surface-area-to-volume ratio and tunability of surface engineering.^11–14 Discovered in 2009 by Wang et al.,¹⁵ graphitic carbon nitride (g-C₃N₄) has been highly regarded for its high thermal and chemical stability, non-toxicity, and affordability. Therefore, it has been a popular choice for solar-driven water splitting,^16–19 CO₂ reduction,^20–24 and other environmental applications.^25–28 Despite its unprecedented characteristics, pristine 2D g-C₃N₄ suffers from much of the same pitfalls of other photocatalysts, that is a narrow visible light absorption region (with an onset wavelength of ca. 440–450 nm),²⁹ which indicates limited light-harvesting abilities in a broad extended visible light region. Furthermore, other reasons for low activity of bare g-C₃N₄ include insufficient reactant adsorption and reactive sites as well as a rapid electron–hole recombination rate.³⁰

In light of that, a low C/N atomic ratio carbon nitride allotrope, C₃N₅, has emerged as a promising successor to overcome the drawbacks of g-C₃N₄. This 2D nitrogen-rich C₃N₅ possesses excellent electronic properties and ample N-active sites, and induces more electrons in the π–π conjugation compared to g-C₃N₄, which is attributed to the electron contribution of the sp² hybridized N on the triazole moiety.³¹ As such, the emergence of a carbon nitride allotrope represents a new paradigm of nanoengineering through chemical composition tuning. C₃N₅ has shown a remarkable performance in photo- and electrocatalytic CO₂ reduction,^32,33 hydrogen evolution,^31,34 and other environmental applications.^35,36

The typical bulk condensation of N-rich precursors (i.e. urea,³⁷ thiourea,^38,39 cyanamide,⁴⁰ dicyandiamide,⁴¹ and melamine⁴²) leads to their incomplete polymerization with large amounts of non-condensed surface amino groups that result in sluggish charge carrier transport.^18,43 The thermal shock during high-temperature calcination has also been linked to agglomeration and a disordered arrangement of homo-triazine units.⁴⁴

Therefore, there has been increasing interest in the synthesis of fully condensed crystalline carbon nitride, which has an improved electron transport along the in-plane direction ascribed to the extended π-conjugated system and delocalized π-electrons and reduced recombination centers for charge carriers.⁴⁵ In addition, its high crystallinity has been known to improve its charge transfer kinetics and enhance its light harvesting efficiency.⁴⁶ As such, molten-salt synthesis is a green and cost-effective method for structural modification. Molten-salts act as the reaction medium for synthesis and flux for crystal growth. The molten-salt synthesis method is also more facile than a soft/hard templating approach, which requires toxic ammonium bifluoride and hydrofluoric acid to remove the silicon. In contrast, salt can be easily removed with warm water.⁴⁷

Unfortunately, purely poly(triazine imide) (PTI)-based carbon nitride displayed poorer photocatalytic performance compared to heptazine-based carbon nitrides with an extended and fully condensed conjugation structure.⁴⁸ Thus, Wang et al. have combined different carbon nitride phases in g-C₃N₄ with the incorporation of 2,4-diamino-6-hydroxypyrimidine as an effective strategy towards regulating its band structure and improving its photocatalytic hydrogen evolution activity (569.5 μmol g⁻¹ h⁻¹).⁴⁹ Thus, achieving an isotype heterojunction through heptazine/triazine phase tuning can effectively realize an intimate interface contact with a low resistance interface due to the interfacial electric field.⁵⁰

Because the C₃N₅ allotrope has been reported more recently, to the best of our knowledge, there has been a lack of research on crystalline C₃N₅ phase tuning. Therefore, a one-step molten salt method was used in this study to prepare crystalline C₃N₅ nanosheets with a heptazine/triazine isotype heterojunction (Fig. 1). Alkali salts LiCl and KCl were chosen as they favor the formation of triazine and heptazine structures, respectively.⁴⁸ The units of heptazine form a π electron system, which can enhance photocatalytic stability and enhance the transfer of charge.⁵¹ Meanwhile, the incorporation of the triazine phase with the heptazine phase leads to an isotype heterojunction that synergistically improves the photogenerated electron–hole separation.⁴⁹ An increase in calcination temperature and the salt/carbon precursor ratio results in the transition of the heptazine phase to the PTI phase, allowing for tunable phase control with a moderate calcination temperature and salt ratio.^52,53 The effects of temperature on the properties of crystalline C₃N₅ as well as its versatility in photocatalytic reactions will be analyzed. This includes multifunctional photocatalytic applications toward H₂ production and H₂O₂ production as well as benzyl alcohol oxidation.

Fig. 1 This work presents a one-step molten salt synthesis approach to achieve a tunable isotype heterojunction in N-rich crystalline C₃N₅. The crystalline C₃N₅ successfully improved solar-driven hydrogen production and simultaneous hydrogen and benzaldehyde production, as well as H₂O₂ production.

Results and discussion

Crystalline C₃N₅ was synthesized through thermal calcination of 3-amino-1,2,4-triazole with LiCl/KCl molten salts (Fig. S1, ESI†). As seen in the X-ray diffraction (XRD) patterns (Fig. 2a), pristine C₃N₅ only has two main peaks at 2θ = 12.6° and 27.53°, which were ascribed to the in-plane repeating unit (100) and the interlayer stacking unit (002), respectively. Notably, the increase of synthetic temperature (CC3N5-600) leads to the complete change of heptazine-based carbon nitrides to PTI with higher crystallinity. This is attributed to the higher thermal stability of PTI compared to the heptazine phase at higher temperature.

Fig. 2 (a) XRD patterns of pristine C₃N₅ and crystalline C₃N₅ samples. (b) HRTEM images of CC3N5-500 and (c) zoomed-in image of the red box. (d) TEM image of CC3N5-500 (yellow dotted lines indicate the presence of pores). (e) HRTEM image of CC3N5-500.

At a moderate synthesis temperature (CC3N5-500), the co-existence of triazine-based carbon nitride can be seen through the diffraction peaks at 21.04°, 26.63°, 29.15°, and 32.41°, which can be assigned to the (110), (200), (102) and (210) planes of PTI, while the peaks at 12.04° and 27.64° correspond to the (100) and (002) planes of the heptazine phase CN, respectively. At 500 °C, the heptazine and triazine phases are calculated to be 55 and 45%, respectively. The triazine phase increases to 74% and 93% for the CC3N5-550 and CC3N5-600 samples, respectively. Additionally, the full width at half-maximum (FWHM) values of the (002) peak of the CC3N5-500 (0.65°) and CC3N5-550 (0.15°) samples were narrowed compared to that of the pristine C₃N₅ sample (1.05°), which indicates their improved crystallinity over bulk C₃N₅. This is consistent with the Jin et al.'s observations on crystalline g-C₃N₄.⁵²

Furthermore, the high-resolution transmission electron microscopy (HRTEM) image (Fig. 2b and c) demonstrates the coexistence of heptazine- and triazine-based crystalline carbon nitrides in CC3N5-500, whereby the lattice fringe with d = 0.32 nm corresponds to the (002) plane of heptazine-based carbon nitride and another lattice fringe of 0.35 nm corresponds to the (200) plane of triazine-based carbon nitride.⁵⁴ The presence of pores (ca. 15–20 nm) can also be clearly seen in Fig. 2d, which is an indication of a high surface area. It can also be seen in the HRTEM image in Fig. 2e that CC3N5-500 is around 4 layers thick.

The chemical structure of the as-prepared crystalline carbon nitride samples was investigated using Fourier transform infrared (FTIR) measurements (Fig. 3a). The pristine g-C₃N₄ and C₃N₅ samples have a broad peak in the range of ∼3000–3500 cm⁻¹, which is attributed to the N–H stretching vibrations of the terminal amine group or the bridging amine group between s-triazine units. The pristine carbon nitrides also possess typical stretching modes of aromatic CN heterocycles in the region of 1100–1700 cm⁻¹. A peak at ∼850 cm⁻¹ is due to the heptazine ring-bending vibrations.⁵³ As for the crystalline C₃N₅ samples, a new weak peak emerged at ∼2100 cm⁻¹ attributed to the asymmetric contraction vibrations of the cyano group (C≡N), due to the loss of ammonia.⁴⁷ A reduced broad peak at ∼3000–3500 cm⁻¹ for the crystalline C₃N₅ samples indicates a reduced concentration of the terminal amino group (N–H) and adsorbed H₂O. This can be correlated with the molten-salt induced deamination in the formation of a tri-s-triazine-based crystalline structure.⁵⁵ A band at ∼640 cm⁻¹ for CC3N5-500, CC3N5-550 and CC3N5-600 is attributed to the stretching vibrations of C–Cl formed during the molten salt process (Fig. 3b) and confirms the successful intercalation of chloride into carbon nitrides.^47,54 As such, all these observations resulted from the one-pot molten-salt synthesis for varying the as-synthesized heptazine/triazine phases of crystalline C₃N₅ with temperature control.

Fig. 3 (a) FTIR measurements for carbon nitride photocatalysts with (b) zoomed-in shaded region. (c) UV-Vis absorption spectra and (d) corresponding Tauc plots of g-C₃N₄, C₃N₅, CC3N5-400, CC3N5-450, CC3N5-500, CC3N5-550 and CC3N5-600.

The UV-Vis spectra of all the samples are displayed in Fig. 3c. g-C₃N₄ shows a characteristic absorption peak between 300 and 400 nm,⁵⁶ while C₃N₅ has a broad absorption band due to the π → π* transition in the conjugated network and is red-shifted with a band tail up to 650 nm. This difference is due to the increased azo-bridge (–N=N–) between the heptazine units, which extends the π conjugated network due to the overlap between N 2p orbitals of bridging azo moieties and N 2p in the heptazine π conjugated system.⁵⁷ The π → π* transition also facilitates the absorption in the visible light spectrum and results in the orange color of the C₃N₅ sample.⁵⁸

As for the crystalline C₃N₅ systems, CC3N5-400 and CC3N5-450 show the lowest absorption intensity in the range of 300–350 nm and an absorption edge of ∼520 nm, while CC3N5-550 and CC3N5-500 have an absorption edge at 613 and 595 nm, respectively. Furthermore, the absorption edge of CC3N5-600 is at 462 nm. The optical bandgaps were determined with a Tauc plot, exemplifying that the band gap of C₃N₅ (2.67 eV) is narrowed compared to that of g-C₃N₄ (2.98 eV), which facilitates an enhanced light utilization and harvesting. Meanwhile, the crystalline C₃N₅ samples have band gaps of 2.78, 2.79, 2.65, 2.64 and 2.72 eV for CC3N5-400, CC3N5-450, CC3N5-500, CC3N5-550 and CC3N5-600, respectively. Notably, the triazine phase of crystalline C₃N₅ samples has downshifted band positions compared to the heptazine phase (Fig. S4, ESI†), which enables tuning the band structure through controlling the calcination temperature in the salt medium. As such, the isotype heterojunction formed between the heptazine/triazine units with interleaved band gaps aids in the interfacial charge transfer.⁵⁹

X-ray photoelectron spectroscopy (XPS) analysis was conducted on CC3N5-500 to further confirm its surface compositions and elemental chemical state (Fig. 4). The C 1s spectra have four peaks at 284.8, 286.5, 288.4, and 289.8 eV, which are attributed to the sp² C–C bond, the C–O bond, the C≡N and the N–C=N species, respectively. The N 1s spectra also show a peak assigned to C≡N for CC3N5-500 (400.2 eV), along with the existence of the C=N–C peak (398.8 eV),⁶⁰ which is also in line with the FTIR results. The valence state of K was analyzed through the K 2p XPS, which shows two binding energy peaks at 293.6 and 295.5 eV assigned to the K⁺ ion, which are different from that of metallic K (294.7 eV).⁶¹ The presence of Cl⁻ ions can also be observed through XPS analysis (0.63%), as evidenced by the Cl 2p_1/2 and Cl 2p_3/2 peaks at 200.3 and 199.1 eV, respectively.⁶² The elemental ratios calculated from XPS are shown in Table S1 (ESI†). Trace Li, Cl and K elements were estimated to be less than 10% according to the XPS results. As for CC3N5-600 (Fig. S5a–d, ESI†), the O 1s spectra have two peaks located at 529 and 531 eV, corresponding to C–O and the surface adsorbed H₂O, respectively. The C≡N peak was also shifted towards a slightly higher binding energy (400.71 eV) in the N 1s spectra of CC3N5-600. For pristine C₃N₅ (Fig. S5e–h, ESI†), the C 1s peak is deconvoluted into three peaks with binding energies of 284.8, 287.9 and 288.2 eV, corresponding to the sp² C–C bond, the C≡N and the N–C=N species, respectively. Notably, the at% of C≡N species in the N 1s spectra of pristine C₃N₅ (44%) was lower than those of CC3N5-500 (87%) and CC3N5-600 (78%). There are no Li, Cl or K peaks for pristine C₃N₅ according to the survey spectra.

Fig. 4 High-resolution XPS spectra of (a) C 1s and K 2p, (b) N 1s, (c) O 1s, and (d) Cl 2p of CC3N5-500. (e) SEM and (f) EDX analysis of the CC3N5-500 sample.

Scanning electron microscopy (SEM) combined with energy-dispersive X-ray (EDX) analysis was also performed for CC3N5-500 (Fig. 4e and f). The EDX spectrum demonstrates the presence of corresponding C, N, O, Cl, and K elements in the material. The C/N ratio calculated from SEM-EDX analysis (C:N = 3 : 4.93) is also similar to that calculated from XPS (C:N = 3 : 5.51) (Table S1, ESI†).

The photocatalytic activities of the various carbon nitride catalysts for hydrogen production are shown in Fig. 5a. The hydrogen evolution rate of pristine C₃N₅ is 2.3-fold higher than that of g-C₃N₄, in the presence of a sacrificial agent (TEOA), which can be attributed to the higher C/N atomic ratio that facilitates more catalytically active sites. As for the crystalline C₃N₅ samples, the CC3N5-400 and CC3N5-600 have negligible HER activity, while the CC3N5-500 catalyst has the highest HER rate of ∼360 μmol h⁻¹ (20 mg), which outperforms those of other carbon nitride catalysts (Table S2, ESI†). This stems from the poor light absorption and excess content of PTI for CC3N5-600 (93%) and incomplete polymerization of CC3N5-400.⁴⁹ This highlights that the heptazine/PTI ratio resulting from the calcination temperature plays a predominant role in its photocatalytic performance. This is based on the XRD results that show CC3N5-500 possesses a heptazine/triazine ratio of 55 : 45. Such a trend is similar to that of studies done on crystalline g-C₃N₄, whereby the temperature range of 500 to 550 °C contributes to the highest photocatalytic performance.⁵⁴

Fig. 5 (a) HER production for all catalysts and (b) AQE measurements of CC3N5-500. (c) Photocatalytic H₂ and benzaldehyde production of pristine C₃N₅ and CC3N5-500. (d) Photocatalytic H₂O₂ production of pristine C₃N₅ and crystalline C₃N₅ (CC3N5) catalysts. (e) Recyclability test for the optimal CC3N5-500 photocatalyst on a 3D printed substrate.

The solar-to-hydrogen (STH) efficiency of CC3N5-500 measured using an AM1.5 light filter was calculated to be 1.31%. The apparent quantum efficiency (AQE) was measured at various wavelengths (380, 400, 420, 450, and 500 nm) to better evaluate its photocatalytic activity (Fig. 5b and Table S3, ESI†). At 420 nm, CC3N5-500 had an AQE of 12.86%, which is higher than that of other single materials or doped carbon nitride photocatalysts reported to date.^63–66 The AQE action spectrum of the CC3N5-500 sample indicates that it is able to generate H₂ in the visible light range up to 450 nm, consistent with the UV-Vis spectrum.

Apart from the HER half-reaction, the crystalline carbon nitride semiconductor also demonstrates a significant improvement toward simultaneous photocatalytic HER and benzyl alcohol oxidation. As shown in Fig. 5c, the CC3N5-500 sample produced 3 times more hydrogen and 50 times more benzaldehyde than pristine C₃N₅ when the sacrificial agent (TEOA) was switched to benzyl alcohol (3 vol%). Furthermore, the benzaldehyde consumption rate of CC3N5-500 (223.37 μmol h⁻¹) is also higher than that of pristine C₃N₅ (161.54 μmol h⁻¹). Therefore, it can be deduced that the improved crystallinity of CC3N5-500 leads to better utilization of both electrons and holes in order to partake in simultaneous HER and organic synthesis. This is an indication of its versatility towards photoredox chemical synthesis. Additionally, these results are also better than those of similar carbon nitride research works published in recent years (Table S4, ESI†), ascertaining the vital role of interfacial electron–hole transfer in phase tuning.

The application of crystalline C₃N₅ can also be extended toward H₂O₂ production (Fig. 5d). The most optimal sample was identified to be CC3N5-550 with a H₂O₂ evolution rate of 619.42 μmol L⁻¹ h⁻¹, whereas pristine C₃N₅ displayed negligible H₂O₂ production (Table S5, ESI†). This is followed by CC3N5-600, CC3N5-500, and CC3N5-450. There exists an optimal trend in the photocatalytic performance, which indicates that the presence of both heptazine and triazine phases aids in the oxygen reduction reaction towards H₂O₂ due to the interfacial electron transfer across the junction. Additionally, as in the control experiments, the photocatalytic H₂O₂ production activity of CC3N5-550 was studied in purified air and N₂ atmospheres to ensure the O₂ dependency of the reaction (Fig. S6c, ESI†). Compared to the experiment conducted with O₂ purging, only 364.9 and 71.35 μmol L⁻¹ h⁻¹ H₂O₂ was produced with purified air and N₂ purging, respectively, demonstrating that O₂ is the contributing factor in the reaction.⁶⁷

The reusability of a photocatalyst is a crucial measure for its practical applications. The CC3N5-500 catalyst was then tested for its recyclability in conjunction with 3D printing technology (see Methodology, Video in the ESI†). It was found that the catalyst retained 92% of photocatalytic H₂ production even after 16 h (4 cycles) (Fig. 5e), thus proving its potential to be applied in scaled-up HER setups such as in a continuous flow reactor. From Fig. S7 (ESI†), it can be seen that the catalyst has largely been retained on the 3D substrate after each cycle. Thus, this demonstrates the importance of integrating additive manufacturing processes such as 3D printing to narrow the gap between lab-scale and commercial-scale photocatalytic green fuel and chemical production.

Electrochemical analysis has also been carried out to further unveil the underlying electronic properties. Electrochemical impedance spectroscopy (EIS) was conducted to uncover the charge carrier recombination processes of the different carbon nitride catalysts. From the Nyquist plot in Fig. 6a, the arc radius of CC3N5-500 is smaller than that of the other crystalline C₃N₅ samples and pristine C₃N₅. This decreased radius demonstrates that the photogenerated carrier separation and fast surface interfacial charge transfer in CC3N5-500 are crucial factors for its high HER performance. The charge transfer resistances (Rp.R) for the different samples are shown in Table S6 (ESI†), whereby CC3N5-500 has the lowest Rp.R. This result is in line with the photocatalytic HER results and supports the interfacial electron–hole transfer hypothesis.

Fig. 6 (a) EIS with an equivalent circuit model (inset), (b) LSV, (c) transient photocurrent (yellow shaded region indicates light on), and (d) solid-state EPR spectra of C₃N₅ and CC3N5-500 catalysts. Proposed photocatalytic mechanism for (e) HER half-reaction and (f) simultaneous HER and benzyl alcohol oxidation (BA: benzyl alcohol; BAD: benzaldehyde; CBM: conduction band minimum; and VBM: valence band maximum).

The photocatalyst samples were also studied through linear sweep voltammetry (LSV) (Fig. 6b). Pristine C₃N₅ only generates a low current density of −0.069 mA cm⁻² at an applied potential of −0.8 V (vs. Ag/AgCl), while the CC3N5-500 photoelectrode exhibits a current density of −0.8 mA cm⁻² at the same applied potential. The ameliorated current density is due to the well separation of electron–hole pairs in the CC3N5-500 sample.⁶⁸ The charge transfer ability was also confirmed through transient photocurrent responses (Fig. 6c). The order of photocurrent response intensities is CC3N5-500 > CC3N5-550 > CC3N5-600 > CC3N5-450 > CC3N5-600 > CC3N5-400 > C₃N₅. This is further evidence of the high separation efficiency of charge carriers in crystalline carbon nitride,⁶⁹ particularly CC3N5-500, which enables its superb photocatalytic performance due to the heptazine/triazine phase junction.

Electron paramagnetic resonance (EPR) analysis was conducted for g-C₃N₄, C₃N₅ and CC3N5-500 samples (Fig. 6d). All three samples show an obvious EPR signal at a g-value of 2.0034, which corresponds to the defects in the carbon nitride framework.⁷⁰ The EPR intensity of C₃N₅ is higher than that of CC3N5-500, signifying that the molten salt calcination treatment leads to a lower defect concentration due to the improved polymerization degree of CC3N5-500.⁷¹ Hence, this could lead to a reduced charge recombination of crystalline C₃N₅ and excellent photocatalytic results.

The active radical species involved in photocatalytic H₂ evolution were further investigated using 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) spin-trapped EPR spectroscopy to detect the hydroxyl (˙OH⁻) and the superoxide radicals (˙O₂⁻). In Fig. S8a (ESI†), no obvious DMPO-˙OH signal can be found for pristine C₃N₅, which implies that no ˙OH radicals were generated. In comparison, 4 characteristic peaks of the DMPO-˙OH adducts can be found for the CC3N5-500 sample, indicating that these radicals are generated during photocatalytic reactions.⁷² The signals for active radicals for ˙O₂⁻ were also observed for CC3N5-500 (Fig. S8b, ESI†), demonstrating that they are part of the photocatalytic mechanism. As such, the EPR results revealed that crystalline C₃N₅ can produce more ˙OH⁻ and ˙O₂⁻ radicals under visible-light irradiation, which could be due to the reduced electron–hole recombination rate of crystalline C₃N₅ compared to its pristine form.⁷³

The Mott–Schottky curves for all samples are consistent with that of a typical n-type semiconductor (Fig. S9a–g, ESI†). In addition, the Mott–Schottky curves were extrapolated to identify the flat band potential and subsequently their conduction band (CB). The CB potential of an n-type semiconductor is ∼0.1 V more negative than its flat band potential.⁷⁴ Therefore, the CB potentials of all catalysts are listed in Table S7 (ESI†), and their band positions were calculated with the E_g data obtained from UV-Vis diffuse reflectance spectroscopy (DRS) analysis. Notably, the CB positions of all crystalline C₃N₅ samples were more positive than that of pristine C₃N₅. By combining the CB position obtained through the Mott–Schottky curves and the band gap positions obtained through UV-Vis analysis, the band positions of the heptazine phase and triazine phase were proposed.

Based on the aforementioned various characterization analyses and the superior photocatalytic activity, the possible photocatalytic mechanism was proposed and shown in Fig. 6e and f. Under light irradiation, electrons are excited from the VB to the CB of the catalyst. Due to the presence of the isotype heterojunction, the electrons flow from the CB of the heptazine to the triazine phase, while the holes migrate from the VB of the triazine to the heptazine phase.⁷⁵ This is in line with previous studies that indicate the phase junctions between heptazine-based carbon nitride and triazine-based carbon nitride belong to the type-II heterojunction.^54,76 The electrons on the triazine-based carbon nitride then react with H⁺ to produce H₂, whereas the holes oxidize TEOA to TEOA⁺. For the simultaneous HER and benzyl alcohol oxidation, the holes will oxidize benzyl alcohol to benzaldehyde at an oxidation potential of 0.68 V (vs. the normal hydrogen electrode, NHE).⁷⁷ The H₂O₂ production mechanism shown in Fig. S9h (ESI†) is also similar to the HER half-reaction mechanism, whereas the ORR reaction takes place with a reduction potential of 0.68 V (vs. NHE). As a whole, this proposed photocatalytic mechanism for the isotype heterojunction crystalline C₃N₅ facilitates charge carrier separation, which markedly accelerates the photocatalytic redox performance.

Conclusion

In this research work, we have successfully synthesized crystalline C₃N₅ with a heptazine/triazine isotype heterojunction with a molten salt synthesis route. Notably, temperature has a great influence on the phase engineering of the carbon nitride, with 500 and 550 °C displaying both heptazine and triazine phases. Notably, a low or high synthesis temperature leads to a more dominant heptazine or triazine phase, respectively, which leads to good interfacial charge transfer. The temperature–performance relationship has a strong influence on the physicochemical properties of the crystalline carbon nitrides, which in turn affects the photocatalytic performance. CC3N5-500 has a heptazine/triazine ratio of 55 : 45, and the triazine phase increases to 74% and 93% for CC3N5-550 and CC3N5-600, respectively. From the UV-Vis results, the light absorption of C₃N₅ improved in the visible light region, with a red-shifted absorption band edge at 603 nm compared to g-C₃N₄ (427 nm). XPS analysis also revealed that the C:N ratio of CC3N5-500 is 3 : 5.5, compared to that of CC3N5-600 (3 : 6.1). Furthermore, the CC3N5-500 sample exhibited the highest HER performance of 359.97 μmol h⁻¹ and an AQE of 12.86% at 420 nm, which were 15-fold higher than those of pristine C₃N₅. The recyclability tests on a 3D printed substrate also exemplified a 92% performance retention after 4 cycles (16 h), which indicates its promising potential for its use in large-scale photocatalytic reactions. The crystalline C₃N₅ sample also outperformed pristine C₃N₅ in benzyl alcohol oxidation. Meanwhile, CC3N5-550 exhibited the highest performance in photocatalytic H₂O₂ production with an AQE of 9.49% at 420 nm. This indicates that crystalline C₃N₅ improves the reaction kinetics for multitudinous solar-driven photoredox reactions due to the boosted charge transfer kinetics as evidenced by the electrochemistry results, resulting in enhanced interfacial electron–hole transfer in the crystalline heptazine/triazine carbon nitride heterojunction compared to the pure heptazine or triazine phase. In short, these results provide guidance for the structural and phase tuning of carbon nitride to aid future studies on its crystallinity for its photocatalyic applications.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The authors would like to acknowledge the financial support provided by the Ministry of Higher Education (MOHE) Malaysia under the Fundamental Research Grant Scheme (FRGS) (Ref no: FRGS/1/2020/TK0/XMU/02/1). The authors would like to thank the Ministry of Science, Technology and Innovation (MOSTI) Malaysia under the Strategic Research Fund (SRF) (S.22015). This research was supported by the National Natural Science Foundation of China (Ref no: 22202168) and Guangdong Basic and Applied Basic Research Foundation (Ref no: 2021A1515111019). We would also like to acknowledge the financial support from the State Key Laboratory of Physical Chemistry of Solid Surfaces, Xiamen University (ref. no.: 2023X11). This work was also funded by Xiamen University Malaysia Investigatorship Grant (Grant no: IENG/0038), Xiamen University Malaysia Research Fund (ICOE/0001, XMUMRF/2021-C8/IENG/0041 and XMUMRF/2019-C3/IENG/0013) and Hengyuan International Sdn. Bhd. (Grant no: EENG/0003).

References

1. Assessing the Global Climate in 2021, https://www.ncei.noaa.gov/news/global-climate-202112).
2. L. Ai, S.-F. Ng and W.-J. Ong, ChemSusChem, 2022, 15, e202200857.
3. H. Mai, T. C. Le, D. Chen, D. A. Winkler and R. A. Caruso, Chem. Rev., 2022, 122, 13478–13515.
4. A. K. Singh, K. Mathew, H. L. Zhuang and R. G. Hennig, J. Phys. Chem. Lett., 2015, 6, 1087–1098.
5. V. B.-Y. Oh, S.-F. Ng and W.-J. Ong, EcoMat, 2022, 4, e12204.
6. J. J. Foo, S.-F. Ng and W.-J. Ong, Nano Res., 2023, 16, 4310–4363.
7. C. Yang, R. Zhao, H. Xiang, J. Wu, W. Zhong, W. Li, Q. Zhang, N. Yang and X. Li, Adv. Energy Mater., 2020, 10, 2002260.
8. C.-F. Fu, X. Wu and J. Yang, Adv. Mater., 2018, 30, 1802106.
9. G. Z. S. Ling, S.-F. Ng and W.-J. Ong, Adv. Funct. Mater., 2022, 32, 2111875.
10. The Future of Hydrogen, https://www.iea.org/reports/the-future-of-hydrogen).
11. W.-J. Ong, L.-L. Tan, Y. H. Ng, S.-T. Yong and S.-P. Chai, Chem. Rev., 2016, 116, 7159–7329.
12. A. Naseri, M. Samadi, A. Pourjavadi, A. Z. Moshfegh and S. Ramakrishna, J. Mater. Chem. A, 2017, 5, 23406–23433.
13. B. Huang, Y. Wu, B. Chen, Y. Qian, N. Zhou and N. Li, Chin. J. Catal., 2021, 42, 1160–1167.
14. G. Z. S. Ling, S. H. W. Kok, P. Zhang, J. S. Tan, C. Y. Haw, L.-L. Tan, B. H. Chen and W.-J. Ong, J. Mater. Chem. A, 2023 10.1039/D3TA04204A.
15. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2008, 8, 76–80.
16. C. Wu, S. Xue, Z. Qin, M. Nazari, G. Yang, S. Yue, T. Tong, H. Ghasemi, F. C. R. Hernandez, S. Xue, D. Zhang, H. Wang, Z. M. Wang, S. Pu and J. Bao, Appl. Catal., B, 2021, 282, 119557.
17. X. Liu, Y. Deng, L. Zheng, M. R. Kesama, C. Tang and Y. Zhu, ACS Catal., 2022, 12, 5517–5526.
18. L. Lin, Z. Yu and X. Wang, Angew. Chem., Int. Ed., 2019, 58, 6164–6175.
19. M. Volokh, G. Peng, J. Barrio and M. Shalom, Angew. Chem., Int. Ed., 2019, 58, 6138–6151.
20. R. Kuriki, K. Sekizawa, O. Ishitani and K. Maeda, Angew. Chem., Int. Ed., 2015, 54, 2406–2409.
21. J. Liang, Z. Jiang, P. K. Wong and C.-S. Lee, Solar RRL, 2021, 5, 2000478.
22. Y. Li, B. Li, D. Zhang, L. Cheng and Q. Xiang, ACS Nano, 2020, 14, 10552–10561.
23. L. Cheng, P. Zhang, Q. Wen, J. Fan and Q. Xiang, Chin. J. Catal., 2022, 43, 451–460.
24. M. Sayed, B. Zhu, P. Kuang, X. Liu, B. Cheng, A. A. A. Ghamdi, S. Wageh, L. Zhang and J. Yu, Adv. Sustainable Syst., 2022, 6, 2100264.
25. A. Yuan, H. Lei, Z. Wang and X. Dong, J. Colloid Interface Sci., 2020, 560, 40–49.
26. T. Uekert, H. Kasap and E. Reisner, J. Am. Chem. Soc., 2019, 141, 15201–15210.
27. T. He, C. Zhang, L. Zhang and A. Du, Nano Res., 2019, 12, 1817–1823.
28. X.-Q. Tan, S.-F. Ng, A. R. Mohamed and W.-J. Ong, Carbon Energy, 2022, 4, 665–730.
29. J. Fu, J. Yu, C. Jiang and B. Cheng, Adv. Energy Mater., 2018, 8, 1701503.
30. X. Yu, S.-F. Ng, L. K. Putri, L.-L. Tan, A. R. Mohamed and W.-J. Ong, Small, 2021, 17, 2006851.
31. S.-F. Ng, X. Chen, J. J. Foo, M. Xiong and W.-J. Ong, Chin. J. Catal., 2023, 47, 150–160.
32. Y. Wang, T. Ngoc Pham, Y. Tian, Y. Morikawa and L. Yan, J. Colloid Interface Sci., 2021, 585, 740–749.
33. J. Zhang, Z. Li, J. He, H. Tao, M. Chen, Y. Zhou and M. Zhu, ACS Catal., 2023, 13, 785–795.
34. C. Peng, L. Han, J. Huang, S. Wang, X. Zhang and H. Chen, Chin. J. Catal., 2022, 43, 410–420.
35. Q. Li, S. Song, Z. Mo, L. Zhang, Y. Qian and C. Ge, Appl. Surf. Sci., 2022, 579, 152006.
36. C. Hu, Y.-H. Lin, M. Yoshida and S. Ashimura, ACS Appl. Mater. Interfaces, 2021, 13, 24907–24915.
37. Y. Yang, L. Geng, Y. Guo, J. Meng and Y. Guo, Appl. Surf. Sci., 2017, 425, 535–546.
38. H. Wang, Z. Sun, Q. Li, Q. Tang and Z. Wu, J. CO2 Utiliz., 2016, 14, 143–151.
39. F. Dong, Y. Sun, L. Wu, M. Fu and Z. Wu, Catal. Sci. Technol., 2012, 2, 1332–1335.
40. J. Liu, J. Yan, H. Ji, Y. Xu, L. Huang, Y. Li, Y. Song, Q. Zhang, H. Xu and H. Li, Mater. Sci. Semicond. Process., 2016, 46, 59–68.
41. L. Yang, X. Liu, Z. Liu, C. Wang, G. Liu, Q. Li and X. Feng, Ceram. Int., 2018, 44, 20613–20619.
42. N. Tian, Y. Zhang, X. Li, K. Xiao, X. Du, F. Dong, G. I. N. Waterhouse, T. Zhang and H. Huang, Nano Energy, 2017, 38, 72–81.
43. S.-F. Ng, J. J. Foo and W.-J. Ong, InfoMat, 2022, 4, e12279.
44. Y. Shao, X. Hao, S. Lu and Z. Jin, Chem. Eng. J., 2023, 454, 140123.
45. H. Li, B. Zhu, S. Cao and J. Yu, Chem. Commun., 2020, 56, 5641–5644.
46. G. Chen, F. Wei, Z. Zhou, B. Su, C. Yang, X. F. Lu, S. Wang and X. Wang, Sustainable Energy Fuels, 2023, 7, 381–388.
47. Y. Zhou, C. Liao, Y. Fan, S. Ma, M. Su, Z. Zhou, T.-S. Chan, Y.-R. Lu and K. Shih, Environ. Sci.: Nano, 2019, 6, 3324–3335.
48. Y. Li, B. Wang, Q.-J. Xiang, Q. Zhang and G. Chen, Dalton Trans., 2022, 51, 16527–16535.
49. W. Wang, X. Kou, T. Li, R. Zhao and Y. Su, J. Photochem. Photobiol., A, 2023, 435, 114308.
50. C. Cheng, L. Mao, X. Kang, C.-L. Dong, Y.-C. Huang, S. Shen, J. Shi and L. Guo, Appl. Catal., B, 2023, 331, 122733.
51. J. Liu, W. Fu, Y. Liao, J. Fan and Q. Xiang, J. Mater. Sci. Technol., 2021, 91, 224–240.
52. A. Jin, Y. Jia, C. Chen, X. Liu, J. Jiang, X. Chen and F. Zhang, J. Phys. Chem. C, 2017, 121, 21497–21509.
53. L. Lin, W. Ren, C. Wang, A. M. Asiri, J. Zhang and X. Wang, Appl. Catal., B, 2018, 231, 234–241.
54. Z. Zhao, Z. Shu, J. Zhou, T. Li, K. Wang and J. Song, J. Alloys Compd., 2023, 938, 168484.
55. W. Wang, Z. Shu, J. Zhou, D. Meng, Z. Zhao and T. Li, J. Mater. Chem. A, 2020, 8, 6785–6794.
56. Y. Zhou, L. Zhang, W. Huang, Q. Kong, X. Fan, M. Wang and J. Shi, Carbon, 2016, 99, 111–117.
57. K. M. Alam, C. E. Jensen, P. Kumar, R. W. Hooper, G. M. Bernard, A. Patidar, A. P. Manuel, N. Amer, A. Palmgren, D. N. Purschke, N. Chaulagain, J. Garcia, P. S. Kirwin, L. C. T. Shoute, K. Cui, S. Gusarov, A. E. Kobryn, V. K. Michaelis, F. A. Hegmann and K. Shankar, ACS Appl. Mater. Interfaces, 2021, 13, 47418–47439.
58. P. Kumar, E. Vahidzadeh, U. K. Thakur, P. Kar, K. M. Alam, A. Goswami, N. Mahdi, K. Cui, G. M. Bernard, V. K. Michaelis and K. Shankar, J. Am. Chem. Soc., 2019, 141, 5415–5436.
59. W. Shi, L. Cao, Y. Shi, W. Zhong, Z. Chen, Y. Wei, F. Guo, L. Chen and X. Du, J. Mol. Struct., 2023, 1280, 135076.
60. P. Xia, M. Antonietti, B. Zhu, T. Heil, J. Yu and S. Cao, Adv. Funct. Mater., 2019, 29, 1900093.
61. G. Zhang, Y. Xu, C. He, P. Zhang and H. Mi, Appl. Catal., B, 2021, 283, 119636.
62. G. Zhang, Y. Xu, M. Rauf, J. Zhu, Y. Li, C. He, X. Ren, P. Zhang and H. Mi, Adv. Sci., 2022, 9, 2201677.
63. S. Mondal, G. Mark, L. Abisdris, J. Li, T. Shmila, J. Tzadikov, M. Volokh, L. Xing and M. Shalom, Mater. Horiz., 2023, 10, 1363–1372.
64. W. Dai, R. Wang, Z. Chen, S. Deng, C. Huang, W. Luo and H. Chen, J. Mater. Chem. A, 2023, 11, 7584–7595.
65. L. Cao, Y. Shi, Z. Chen, Y. Tian, L. Chen, F. Guo and W. Shi, Mater. Today Commun., 2022, 33, 104354.
66. X. Wang, K. Wu, W. Cao, K. Rui, W. Wang, R. Zhu, J. Zhu and Z. Yan, Adv. Mater. Interfaces, 2023, 10, 2201627.
67. B. P. Mishra, L. Acharya and K. Parida, Catal. Sci. Technol., 2023, 13, 1448–1458.
68. C. Hegde, T. Rosental, J. M. R. Tan, S. Magdassi and L. H. Wong, Mater. Horiz., 2023, 10, 1806–1815.
69. B. Wu, T. Sun, N. Liu, L. Lu, R. Zhang, W. Shi and P. Cheng, ACS Appl. Mater. Interfaces, 2022, 14, 26742–26751.
70. G. Chen, Z. Zhang, W. Zhang, L. Xia, X. Nie, W. Huang, X. Wang, L. Wang, C. Hong, Z. Zhang and Y. You, Mater. Horiz., 2021, 8, 2018–2024.
71. Y. Lin, X. Wang, X. Fu and W. Su, J. Mater. Chem. A, 2022, 10, 8252–8257.
72. X. Luan, Z. Yu, J. Zi, F. Gao and Z. Lian, Adv. Funct. Mater., 2023, 2304259.
73. X. Zhang, Z. Li, N. Ta and H. Han, ACS Sustainable Chem. Eng., 2022, 10, 4059–4064.
74. S. Kaowphong, W. Choklap, A. Chachvalvutikul and N. Chandet, Colloids Surf., A, 2019, 579, 123639.
75. F. Li, X. Yue, Y. Liao, L. Qiao, K. Lv and Q. Xiang, Nat. Commun., 2023, 14, 3901.
76. G. Zhang, Y. Xu, J. Zhu, Y. Li, C. He, X. Ren, P. Zhang and H. Mi, Appl. Catal., B, 2023, 338, 123049.
77. X. Xiang, B. Zhu, J. Zhang, C. Jiang, T. Chen, H. Yu, J. Yu and L. Wang, Appl. Catal., B, 2023, 324, 122301.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3mh01115a

---