All starting materials and reagents utilized in the research were of analytical grade, and the experiments were conducted with deionized water that had been subjected to double distillation. The reagents utilized include Ethanol (C₂H₅OH), ethylene glycol (HOCH₂CH₂OH), sodium hydroxide (NaOH), nitric acid (HNO₃), polyvinylpyrrolidone (PVP), 2,4-dichlorophenol (C₆H₄Cl₂O), and urea (CO(NH₂)₂) were purchased from Aladdin Group Co., Ltd. Iron(III) nitrate nonahydrate (Fe(NO₃)₃·9H₂O), neodymium(III) nitrate hexahydrate (Nd(NO₃)₃·6H₂O), sodium silicate (Na₂SiO₃), and methanol (CH₃OH) were purchased from Sinopharm Chemical Reagent Co., Ltd, Sigma Aldrich Merck Ltd, China with standard purity and utilized without additional purification.

NFONS were synthesized by a hydrothermal method; the detailed synthesis scheme is shown in Supplementary Figure 1A. In the standard experiment, Nd(NO₃)₃·6H₂O (4.82 g, 11.0 mmol) and Fe(NO₃)₃·9H₂O (4.44 g, 11.0 mmol) were dissolved in a mixed solvent of 55 mL DI water, 25 mL ethanol, and 30 mL methanol. Next, an equal molar solution of PVP was added, and the mixture was stirred at 400 rpm for 1 h to yield a clear reddish solution. After this, the solution was divided equally between two 100 mL Teflon-lined stainless-steel autoclaves (55 mL each) and heated at 180 °C for 12 h in an oven to complete the hydrothermal reaction. After cooling to room temperature, the product was well washed 3 times with ethanol and deionized water. Next, the as-received product was dried in an oven at 80 °C overnight and then calcined for 2 h in a muffle furnace at 500, with a heating rate of 4 °C per min^[21].

We obtained CNNS by calcination, as shown in Supplementary Figure 1B. In a standard experiment, 15 g of urea was placed in a ceramic crucible and heated in a muffle furnace at 550 °C for 4 h at a rate of 5 °C min^-1 (ramp time ≈ 105 min). The furnace was then allowed to cool naturally to room temperature. The collected samples were treated 5 times with 0.2 M HNO₃, and then rinsed 3 times with water and absolute ethanol. The purified samples were then dried in an oven at 80 °C for 12 h. Finally, the resulting yellowish product was ground to a fine powder^[22].

We synthesized the g-C₃N₄ nanosheet/NdFeO₃ nanosheet Z-scheme nanocomposites by a wet chemical method. In the standard experiment, we mixed different amounts of CNNS with the as-prepared NFONS. In a typical procedure, 1 g of NFONS was dispersed in a 100 mL ethanol-water mixture (1:1, v/v). Subsequently, a defined mass of CNNS corresponding to the desired mass ratios (1%, 3%, 5%, and 7%) was added to the NFONS dispersion. The resulting mixture was ultrasonicated for 1 h to obtain a homogeneous dispersion and enhance interfacial contact between the CNNS and NFONS. After ultrasonication, the mixture was stirred continuously at 400 rpm for 4 h at 85.0 °C. Once the solvent was evaporated completely, the powder-form composite was then dried in an oven at 80 °C to remove residual solvents. Finally, the dried material was calcined at 400 °C for 120 min in air, at a heating rate of 5 °C min^-1, to remove impurities, improve interfacial bonding, and strengthen the structural features of the composite^[23]. Remarkably, the resulting products were denoted as XCNNS/NFONS, where X denotes the mass percentage ratio of CNNS to NFONS.

Si-Bridge-mediated g-C₃N₄/NdFeO₃ Z-scheme heterojunction was synthesized in two steps; the comprehensive synthesis scheme is given in Supplementary Figure 1C. In the 1st step, 0.99 g of as-synthesized NFONS were dispersed in 100 mL of 1.0 M aqueous sodium silicate (Na₂SiO₃) solution (1.0 M sodium silicate solution was prepared by dissolving 12.21 g of Na₂SiO₃ in 100 mL of distilled water). The suspension was vigorously stirred for 1 h to facilitate silicate-ion adsorption onto the NFONS surface, forming a silicate bridge. After stirring, the mixture was filtered, and the collected product was dried in an oven at 85 °C for 12 h. The resultant silicate-modified NFONS was labeled as YSi/NFONS, where Y denotes the mole ratio of Si bridge to NFONS. In the 2nd step, 1.0 g of silicate bridge-modified NFONS (YSi/NFONS) was mixed with 0.05 g of previously prepared CNNS in an ethanol-water mixture (1:1, v/v). The mixture was stirred at 85 °C for 4 h. The resulting products were dried in an oven at 80 °C for several hours to remove impurities. Finally, the dried products were calcined in air at 400 °C for 2 h at a ramp rate of 5 °C/min^[15]. Additionally, other experimental methods are described in the Supplementary Materials, and a list of abbreviations and acronyms used throughout this manuscript is provided in Supplementary Table 1.

Here, we successfully prepared NFONS, which have a unique structure, high surface area, enhanced charge separation, and improved multifunctional activities as compared to NFONP. The transmission electron microscopy (TEM) image of NFONP [Figure 1A] and the scanning electron microscopy (SEM) image of NFONP [Supplementary Figure 2A] reveal that these structures show aggregation and that the nanoparticles are packed, which hinders their overall performance. In contrast, the TEM image of NFONS [Figure 1B] and the SEM image of NFONS [Supplementary Figure 2B] show well-exfoliated nanosheets with plenty of empty, active space and well-arranged structures, which increase surface area and, accordingly, overall performance. The XRD patterns [Figure 1C] reveal that NFONS exhibit well-defined diffraction peaks, indicating good crystallinity and a well-arranged crystal structure as compared to NFONP. The peaks located at 22.8°, 25.1°, 27.4°, 34.0°, 40.1°, 41.9°, 46.5°, 48.0°, 53.1°, 58.4°, 64.0°, 68.2°, 73.4°, and 77.7° corresponding to (110), (111), (121), (210), (031), (022), (202), (040), (131), (123), (133), (242), (143), and (420) crystal planes, which confirm the successful synthesis of NdFeO₃ phase (JCPDS card no. 01-082-2421)^[24]. In addition, the ultraviolet-visible diffuse reflectance spectroscopy (UV-Vis DRS) spectra [Figure 1D] illustrate that the NFONS band edge is located within the visible light region as compared to NFONP, which means that NFONS have a narrow band gap as compared to NFONP. Furthermore, the photoluminescence (PL) spectra [Figure 1E] and fluorescence spectra (FS) related to •OH radicals [Figure 1F] confirm that NFONS possess superior charge separation as compared to NFONP.

In practice, comparative multifunctional photocatalytic activity studies showed that NFONS delivered 2.19 times higher CO₂ conversion [Figure 1G], 1.50-fold higher 2,4-DCP degradation [Figure 1H], and a 2.9-fold increase in H₂ generation [Figure 1I] compared with NFONP^[25,26]. Based on these findings, NFONS show well-defined morphology, good crystal structure, enhanced charge separation, more visible-light activity, and improved photocatalytic activity compared with NFONP^[27].

In this phase of our research, we successfully synthesized CNNS by heating urea; remarkably, slow calcination (5 °C min^-1) yielded well-exfoliated, high-surface-area CNNS. Our study shows that these CNNS have a markedly larger surface area, better morphology, more pores, and improved photocatalytic performance than CNNP^[28]. In short, our experimental results demonstrated that CNNS offer exceptional performance over CNNP. The TEM image of CNNP [Figure 2A] shows aggregation and that the nanoparticles are packed, which hinders their overall performance, while the TEM image of CNNS [Figure 2B] indicates well-exfoliated nanosheets, plenty of empty and active space, and well-arranged structures, which will increase their surface area and, accordingly, their overall performance. The XRD patterns for CNNS and CNNP are shown in Figure 2C and exhibit similar peaks, suggesting that the two materials have similar crystal structures. In addition, the UV-Vis DRS spectroscopy displayed in Figure 2D illustrates that the CNNS band edge is located within the visible light region as compared to CNNP, which means that CNNS have a narrow band gap as compared to CNNP. Furthermore, the PL spectra [Figure 2E] and FS spectra related to •OH radicals [Figure 2F] confirm that CNNS possess superior charge separation as compared to CNNP^[29]. Compared to CNNP, the CNNS delivered higher multifunctional photocatalytic activities for CO₂ conversion [Figure 2G], 2,4-DCP degradation [Figure 2H], and H₂ evolution [Figure 2I]. To put it briefly, the findings of these studies confirmed that CNNS demonstrate a large surface area, more porosity, enhanced charge separation, well-exploited sheets, and improved multifunctional photocatalytic activities as compared to CNNP^[30].

As discussed above, the visible light activity of the existing NFONS is still not satisfactory, mainly due to small surface area, low charge separation, and a positive CB. To address these issues, we modified the as-synthesized NFONS by coupling them with CNNS. This Z-scheme heterojunction increased reactive sites, expanded the specific surface area, improved the surface-to-volume ratio, adjusted band gaps for photogenerated carrier transfer, provided more robust thermodynamic stabilization, and finally improved the multifunctional photocatalytic activity. XRD patterns were recorded to check the crystal nature and structural properties of the prepared sample. It can be observed from XRD patterns given in Figure 3A that the peaks located at 22.8°, 25.1°, 27.4°, 34.0°, 40.1°, 41.9°, 46.5°, 48.0°, 53.1°, 58.4°, 64.0°, 68.2°, 73.4°, and 77.7° corresponding to (110), (111), (121), (210), (031), (022), (202), (040), (131), (123), (133), (242), (143), and (420) crystal planes, which confirms the successful synthesis of NdFeO₃ phase (JCPDS card no. 01-082-2421)^[24]. The identified peaks align with the respective phases^[31,32]. Remarkably, CNNS exhibits distinct diffraction peaks at approximately 12.8° and 27.4°, corresponding to the (100) and (002) crystallographic planes of g-C₃N₄, respectively^[33]. Notably, the XRD patterns of the resulting heterojunction retain the primary diffraction peaks of NFONS as well as the characteristic diffraction peaks of g-C₃N₄, confirming that both crystalline phases coexist within the CNNS/NFONS composite without substantial structural distortion of either component. In addition, UV-Vis DRS [Figure 3B] shows a slight shift of NFONS absorption toward the visible region upon CNNS incorporation, suggesting enhanced light-absorption properties. It can be seen that after CNNS incorporation, the band edges shift to longer wavelengths and into the more visible region, and accordingly the band gap decreases, which will not enhance charge separation, but the resulting samples will also absorb more light and, finally, photocatalytic activities will increase. PL spectra serve as a valuable tool for evaluating the recombination rates of photogenerated charge carriers. Generally, PL signals arise from charge recombination upon light excitation, thereby establishing a direct correlation between the PL response intensity and the enhancement of charge separation. In short, a more intense PL signal generally indicates low charge separation, while a less intense PL signal implies enhanced charge separation. Figure 3C shows that the PL intensity of NFONS gradually decreases as CNNS content increases. Notably, the 5CNNS/NFONS nanocomposite exhibits the lowest PL intensity, indicating the highest charge-separation efficiency and suggesting the most effective photocatalytic performance. These results are further verified by FS associated with the •OH generation; it can be seen from Figure 3C inset that the FS response is increased after adding XCNNS, and the 5CNNS/NFONS showed the highest response, confirming the best sample^[34]. For practical applications, we checked the photocatalytic capability of NFONS and XCNNS/NFONS nanocomposites for CO₂ reduction, 2,4-DCP degradation, and H₂ generation under visible-light irradiation in an aqueous medium, without co-catalysts.

It can be observed from Figure 3D that NFONS have low photocatalytic activity for CO₂ conversion into CH₄ (0.45 µmol g^-1h^-1) and CO (1.30 µmol g^-1h^-1). However, photocatalytic performance improved markedly when coupled with different amounts of CNNS. As compared to other prepared samples, the 5CNNS/NFONS nanocomposite showed a 6.80-fold improvement in photocatalytic activity by producing CH₄ (3.10 µmol g^-1h^-1) and CO (8.80 µmol g^-1h^-1). Similarly, the as-prepared samples were evaluated for photocatalytic 2,4-DCP degradation [Figure 3E]. The 5CNNS/NFONS showed the highest activity (59.0%). Similarly, as shown in Figure 3F, 5CNNS/NFONS generated 95.7 µmol g^-1h^-1 hydrogen, which is 1.87 times higher than NFONS (51.20 µmol g^-1h^-1). The findings unequivocally indicate that integrating CNNS markedly improves the photocatalytic efficiency of pristine NFONS. In light of these findings, it is confirmed that the coupling of CNNS with NFONS and forming a Z-scheme heterojunction fruitfully increased reactive sites, expanded the specific surface area, improved the surface-to-volume ratio, adjusted band gaps for photogenerated carrier transfer, provided more robust thermodynamic stabilization, enhanced charge separation and finally improved the multifunctional photocatalytic activities for CO₂ conversion, 2,4-DCP degradation and H₂ generation.

The interfacial mismatch between n-type g-C₃N₄/p-type NFONS was solved by introducing a (Si-O) bridge, which acts as an interfacial electronic channel and electronic bridge that promotes charge migration, serving both as a linker and a charge-transfer facilitator. In short, this approach resolves the lattice-mismatch issue and finally improves photocatalytic efficiency^[35].

Aiming to investigate the composite system, various types of characterization were carried out. The XRD patterns displayed in Figure 4A and the UV-Vis spectra given in Figure 4B demonstrate that the incorporation of the Si-O bridge does not alter the crystallinity and the band edges, respectively. In addition, the lowest PL intensity [Figure 4C] and highest FS associated with the •OH generation (Figure 4C inset) confirm that the incorporation of the Si-O bridge effectively enhances charge separation, and among the samples, 5CNNS-1.5Si/NFONS showed the highest response. Figure 4D, with standard-deviation error bars, shows that the 5CNNS-1.5Si/NFONS composite exhibited 14.17-fold and 2.08-fold increases in activity as compared to NFONS and 5CNNS/NFONS, respectively. Similarly, the as-prepared samples were checked for photocatalytic 2,4-DCP degradation [Figure 4E] (as standard deviation error bars). The most active 5CNNS-1.5Si/NFONS showed a 7.25-fold increase. Additionally, the 5CNNS-1.5Si/NFONS generated 2.64-fold more hydrogen compared with NFONS, as standard deviation error bars are shown [Figure 4F]. Notably, if the amount of introduced Si-O or CNNS content is further increased, a decline in photocatalytic performance occurs, which means that the 5CNNS-1.5Si/NFONS is an optimized sample^[36].

The reproducibility, stability, and recyclability of the photocatalytic performance were further verified through 4 independent measurements. The consistent CH₄ production [Supplementary Figure 3A], CO production [Supplementary Figure 3B], 2,4-DCP degradation [Supplementary Figure 3C], and H₂ evolution rates [Supplementary Figure 3D] were observed. To confirm reproducibility, four independent runs (n = 4) were conducted for all samples; the mean ± standard deviation values are given in Supplementary Table 2^[37].

Catalyst stability strongly affects photocatalytic performance. With this aim, XRD patterns, photocatalytic activities, stability, and reusability were evaluated for 5CNNS-1.5Si/NFONS^[38]. Supplementary Figure 4A and B, Supplementary Figure 5A and B show that the sample recovered after different characterizations and catalytic reactions maintained a similar physical appearance in terms of color as well as weight after four consecutive cycles. In addition, thermogravimetric analysis (TGA) shown in Supplementary Figure 5C confirms that before and after different characterizations and catalytic reactions, the corresponding TGA curves exhibit nearly identical weight-loss profiles and decomposition patterns. These results confirmed that the composite structure remained stable. Furthermore, UV-Visible absorption spectra [Supplementary Figure 5D] confirm that the recovered catalyst showed no notable changes before and after different characterizations and catalytic reactions. Thus, it is confirmed that our samples have excellent structural and physicochemical stability during the photocatalytic process.

Based on the aforementioned findings, it has been confirmed that, among the prepared samples, the 5CNNS-1.5Si/NFONS is the most active and amount-optimized sample. Similarly, 5CNNS-1.5Si/NFONS shows enhanced charge separation, a larger surface area, a well-designed structure, and improved multifunctional photocatalytic activity. So, the morphology of this sample is further investigated. The TEM images are depicted in Figure 5A and B, showing that NFONS and CNNS exhibit a sheet-like morphology, and these closely assembled structures give an enlarged surface area and active sites for multifunctional applications. In addition, the high-resolution transmission electron microscopy (HRTEM) [Figure 5C and D] shows lattice fringes with interplanar spacings of 0.32, 0.29, and 0.26 nm, corresponding to g-C₃N₄ (002), NdFeO₃ (121), and Si-O bridge (100), respectively. Remarkably, the coexistence of these lattice features, together with the observed interfacial contact, supports the successful assembly of the composite structure. Notably, the Si-O bridge may facilitate interfacial charge transfer and photoelectron modulation between g-C₃N₄ and NdFeO₃^[39].