Ni₂P-Modified P-Doped Graphitic Carbon Nitride Hetero-Nanostructures for Efficient Photocatalytic Aqueous Cr(VI) Reduction

Abstract

Targeting heterostructures with modulated electronic structures and efficient charge carrier separation and mobility is an effective strategy to improve photocatalytic performance. In this study, we report the synthesis of 2D/3D hybrid heterostructures comprising P-doped graphitic carbon nitride (g-C₃N₄) nanosheets (ca. 50–60 nm in lateral size) and small-sized Ni₂P nanoparticles (ca. 10–12 nm in diameter) and demonstrate their prominent activity in the photocatalytic reduction of Cr(VI). Utilizing a combination of spectroscopic and electrochemical characterization techniques, we unveil the reasons behind the distinct photochemical performance of these materials. We show that Ni₂P modification and P doping of the g-C₃N₄ effectively improve the charge-carrier transportation and spatial separation through the interface of Ni₂P/P-doped g-C₃N₄ junctions. As a result, the catalyst containing 15 wt.% Ni₂P exhibits superior photocatalytic activity in the detoxification of Cr(VI)-contaminated effluents under UV-visible light illumination, presenting an apparent quantum yield (QY) of 12.5% at 410 nm, notably without the use of sacrificial additives. This study marks a forward step in understanding and fabricating cost-effective photocatalysts for photochemical applications.

1. Introduction

Water pollution poses a major threat not only to aquatic ecosystems but to human lifespans as well [1,2]. Quality of life has been tremendously improved through the rapid industrialization that took place to meet our commercialized needs. This extensive use of the industrial sector has a major impact on the environment and especially on the pollution of aquatic habitats. Some of the most severe chemical wastes are toxic metals, such as Pb, Hg, As and Cr [3]. In the case of chromium pollution, Cr metal ions can be predominantly found in two different oxidation states, the trivalent (Cr(III)) and hexavalent (Cr(VI)) forms. While Cr(III) compounds are not considered highly life-threatening agents, the Cr(VI) oxo-species are very harmful in the cases of ingestion through potable water [4,5,6]. Due to the extremely high toxicity and solubility of Cr(VI) in aquatic solutions, Cr(VI) has been categorized as a Group 1 carcinogen to humans with an LD₅₀ of 50–150 mg·Kg⁻¹ by the International Agency for Research on Cancer [7]. Various techniques have been applied for the remediation of Cr(VI)-polluted solutions, such as chemical precipitation, membrane filtration, adsorption, ion exchange, etc. [8,9,10,11]. Although the above techniques can be quite effective, their large-scale deployment in decontamination of Cr(VI)-polluted wastewaters is limited by high operation costs; they require excessive amounts of adsorbents or reducing reagents, such as SO₂ and FeSO₄. Alternately, the cost-effectiveness and scalable implementation of the photochemical conversion of Cr(VI) to Cr(III) makes this method one of the most sustainable solutions for environmental pollution [12,13,14]. Sunlight irradiation is the cheapest and cleanest energy source that can be utilized for chemical conversion. Also, thanks to the undemanding isolation process, the photocatalysts can be operated on freely until the full Cr(VI) reduction to Cr(III), and then they can be isolated and reused successfully in repeated catalytic cycles. An additional advantage of the photocatalytic method is that, except for heavy metals, there is a plethora of organic pollutants in industrial effluents, such as phenol, amines, dyes, ethylenediaminetetraacetic acid (EDTA), etc., which may act as sacrificial agents, accelerating the reaction rate of Cr(VI) reduction [15,16]. Those organic pollutants, which serve as electron donors during the photocatalytic process, can be degraded into their oxidized forms that are less toxic and, in some cases, can be further isolated. It is clear that photocatalytic Cr(VI) reduction is a low-cost and environmentally friendly technique, which not only can be used to decontaminate aquatic environments from toxic metals but can also simultaneously degrade organic pollutants with high reaction kinetics and efficiency.

Graphitic carbon nitride (g-C₃N₄) is a well-known 2D semiconductor material, which, due to its high chemical and thermal stability along with its unique electronic properties, presents interesting perspectives in various applications and especially in the field of photocatalysis [17,18,19]. This organic semiconductor has recently attracted the attention of the research community, since it is a non-toxic material and its synthesis requires the use of inexpensive raw materials, such as urea, thiourea and melamine. Also, high-purity g-C₃N₄ can be easily produced at a large scale for industrial use. Although g-C₃N₄ can be considered an ideal candidate for heterogeneous photocatalysis, it suffers from instantaneous electron-hole recombination, limited utilization of solar radiation (it has a band gap (E_g) of 2.7–2.8 eV) and low density of surface-active sites. These flaws guided the materials research in the direction of devising heterostructures with metals (Ag, Au, Pd, Ru, etc.) and metal oxide or chalcogenide semiconductors, doping the structure with metal (Fe, Ni, Co, Mn, La, etc.) or non-metal (B, F, O, P, S, etc.) ions and other synthetic modifications in order to further assist the photocatalytic activity of this low-cost semiconductor [20,21,22,23]. Compared to the other optimization approaches, the construction of heterojunctions between g-C₃N₄ and another material (co-catalyst) seems to be a very effective way of enhancing the photocatalytic performance of g-C₃N₄ [24,25,26]. Specifically, the close contact between the two components plays a crucial role in the electronic interfacial interactions, promoting the efficient transfer and separation of photogenerated electrons and holes at the interface, thereby improving photocatalytic efficiency. Also, the utilization of solar energy by g-C₃N₄ can be significantly improved by the proper employment of visible light-responsive co-catalysts. To this end, a diverse set of semiconductor materials, like BiVO₄, SnS₂, CdS and MoS₂, have been used for the formation of heterojunctions with g-C₃N₄ for photocatalytic reactions [27,28,29,30]. Although these co-catalysts improve interfacial charge separation and transportation, certain pitfalls in this research are the low electron transfer kinetics and poor photochemical stability. Propitious replacements of the conventional semiconductors (metal oxides, sulfides, etc.) are the transition metal phosphides (TMPs), such as Ni₂P, CoP, Cu₃P, MoP, etc. [31,32,33,34]. Despite the small progress, TMPs have already emerged with great potential in overcoming the main flaws of conventional semiconductors, demonstrating excellent multiple redox activity, enhanced visible light absorption and outstanding stability in both acidic and alkaline media [35]. Moreover, TMPs are non-toxic and inexpensive materials. Taking into consideration the above premises, the coupling of g-C₃N₄ with TMPs to form heterojunctions can provide sustainable photocatalysts with combined high efficiency and long-term stability.

In this article, we report for the first time on the synthesis of Ni₂P-modified P-doped g-C₃N₄ heterostructures via a facile two-step thermal polycondensation and hydrothermal method and demonstrate their eminence in photocatalytic detoxification of aqueous Cr(VI) solutions under UV-visible light irradiation. The resulting Ni₂P/P-g-C₃N₄ materials comprise P-doped g-C₃N₄ nanolayers and small-size Ni₂P nanoparticles. By using a combination of electrochemical and spectroscopic techniques, we provide a mechanistic understanding of charge transport and photochemical reactions in this catalytic system. Our findings indicate that Ni₂P/P-g-C₃N₄ hetero-nanostructures allow the generation of electron transfer channels with improved charge carrier dissociation across the junction, leading to superior photocatalytic performance. As a result, the catalyst with 15 wt.% Ni₂P content attains complete reduction of Cr(VI) to Cr(III) in 80 min, reaching a photon to chemical conversion yield of 12.5% at 410 nm monochromatic light, importantly, without the aid of sacrificial agents. Moreover, Ni₂P/P-g-C₃N₄ catalysts show a persistently high photochemical activity towards Cr(VI) reduction (>99% Cr(VI) reduction in 80 min) even with the coexistence of other competing ions. The results of this study demonstrate the prospective applicability of the Ni₂P-modified P-doped g-C₃N₄ hetero-nanostructures for environmental protection, including detoxification of industrial effluents with rich Cr(VI) concentration.

2. Results and Discussion

2.1. Synthesis and Structural Characterization

The Ni₂P-modified P-doped g-C₃N₄ nanosheets were synthesized via a facile two-step synthetic procedure, as described in Figure 1A. Namely, we used a conventional thermal polycondensation reaction of urea to prepare g-C₃N₄ (denoted as GCN_b). Next, the growth of Ni₂P nanoparticles on g-C₃N₄ and chemical doping of the host polymer with phosphorus atoms was achieved by reacting the as-made g-C₃N₄ with NiCl₂ and red P precursors under hydrothermal conditions, so that 2D/3D hybrid hetero-nanostructures of tightly connected P-doped g-C₃N₄ nanosheets and Ni₂P nanoparticles were obtained. In this hydrothermal process, P reacts with Ni²⁺ ions on the surface of g-C₃N₄ to give Ni₂P nanoparticles, while also substitutionally doping the g-C₃N₄ crystal lattice. By regulating the concentrations of NiCl₂ and red P in the reaction mixture, a series of Ni₂P/P-g-C₃N₄ heterostructures (denoted as x-Ni₂P/P-GCN) were produced with different weight contents of Ni₂P, i.e., x = 5, 10, 15 and 20 wt.%. For comparison reasons, pure Ni₂P and undoped (denoted as GCN) and P-doped (denoted as P-GCN) g-C₃N₄ nanosheets were also prepared using a similar hydrothermal process and thoroughly studied.

The crystal structure and chemical composition of as-prepared catalysts were assessed using energy-dispersive X-ray spectroscopy (EDS) and powder X-ray diffraction (XRD). The EDS spectra verified the presence of Ni, P, C and N elements in the Ni₂P-modified samples. On the basis of the EDS atomic ratio of Ni:N, the Ni₂P contents in these materials were found to be very close to the expected compositions by the stoichiometry of reactions, see Table S1. Results of XRD experiments on pure Ni₂P, pristine GCN and Ni₂P/P-GCN samples are shown in Figure 1B. The XRD patterns of Ni₂P/P-GCN display two diffraction peaks at ~13.2° and ~27.8°, corresponding to the (100) and (002) crystal planes of g-C₃N₄ (JCPDS card no. 87-1526), and several distinct diffraction peaks in the 2θ range of 40–60°, corresponding to the crystal planes of hexagonal Ni₂P (JCPDS card no. 74-1385). Using Scherrer’s equation, the broadening of the (111) diffraction peak of Ni₂P particles coincided with the formation of crystallites with a 15–20 nm grain size, which is consistent with the Ni₂P particle diameter obtained from TEM data (see below). These results suggest that Ni₂P nanoparticles are eventually formed on the surface of g-C₃N₄ through the employed hydrothermal reaction. By comparing the XRD patterns of GCN and P-GCN, we found that doping the structure of g-C₃N₄ with phosphorus does not alter the crystal symmetry; both GCN and P-GCN samples show similar XRD profiles (Figure S1).

The morphology of the 15 wt.% Ni₂P-modified P-doped g-C₃N₄ (15-Ni₂P/P-GCN), which is the best catalyst in this study, was observed by field emission scanning electron microscopy (FE-SEM) and transmission electron microscopy (TEM). Representative FE-SEM images in Figure 2A reveal that this sample has a flake-like morphology in which plenty of intersecting nanosheets form a layered network structure. In line with this, a low magnification TEM image of 15-Ni₂P/P-GCN, shown in Figure 2B, further confirms a nanoflake-assembled morphology of randomly interconnected nanosheets with a lateral dimension of approximately 50–60 nm. Compared to as-made g-C₃N₄ (GCN_b), the lateral size of C₃N₄ flakes of hydrothermal-treated materials is much smaller; GCN_b consists of multi-layer flakes with a lateral size of around 0.8–1.2 μm, as evidenced by FE-SEM (see Figure S2). Figure 2C is a high-resolution TEM (HRTEM) image of 15-Ni₂P/P-GCN, which shows an isolated Ni₂P nanoparticle with a diameter of 10–12 nm (appearing as dark, small areas) studded on the surface of g-C₃N₄ (appearing as light areas), in line with the EDS mapping results shown below. The very similar dimension obtained from TEM particle size and XRD grain size implies a single-crystalline nature of Ni₂P particles. The hexagonal lattice structure of Ni₂P was further verified by fast Fourier transform (FFT) analysis of the TEM image. The FFT pattern obtained from this particle reasonably correlates to the crystal planes along the [100] zone axis of the hexagonal Ni₂P (space group: P-62m) (inset of Figure 2C). In Figure 2D, the SEM-EDS elemental mapping indicates a uniform distribution of the C, N and P atoms within the detection area of the 15-Ni₂P/P-GCN surface, while Ni atoms distributed at certain regions, possibly at the edges of the g-C₃N₄ sheets [36]. Taken together, these results affirm the successful growth of Ni₂P nanoparticles on the surface of the P-doped g-C₃N₄ nanosheets, resulting in the formation of Ni₂P/P-doped g-C₃N₄ heterojunctions with, importantly, intimate contact. The strong integration between Ni₂P and P-doped g-C₃N₄ components is advantageous for photocatalysis since it allows for the smooth transfer and dissociation of charges across the Ni₂P/P-g-C₃N₄ interface.

The valence state of elements in the as-prepared materials was analyzed by X-ray photoelectron spectroscopy (XPS). Figure 3A shows the C 1s spectra for GCN, P-GCN and 15-Ni₂P/P-GCN, which is the most active catalyst in this series. For the GCN, deconvolution of the C 1s signal yields four singlet peaks at 284.8, 286.2, 288.3, 289.5 ± 0.2 eV binding energy due to the adventitious carbon, sp²-bonded (N–C=N) and sp³-bonded (C–N) carbons in triazine units, and surface COx impurities of g-C₃N₄, respectively [37,38]. Similar to this, the C 1s peaks associated with the N–C=N and C–N linkages and COx impurities of P-GCN appear at 286.2, 288.3 and 289.9 ± 0.2 eV and those peaks for 15-Ni₂P/P-GCN appear at 285.9, 288.5 and 289.8 ± 0.2 eV, respectively. The weak C 1s signal at higher binding energy indicates minor surface oxidation, possibly due to the synthesis procedure and air exposure of the samples. In the deconvoluted N 1s spectra (Figure 3B), the prominent peaks at 398.8 and 400.0 ± 0.2 eV and the broad feature at 404.6 ± 0.2 eV for GCN (398.8, 400.5 and 404.1 ± 0.2 eV for P-GCN) are assigned to the sp²-bonded (pyridinic N) and sp³-bonded (pyrrolic N) nitrogen and terminal-NH_x species of g-C₃N₄, respectively [39]. For the 15-Ni₂P/P-GCN catalyst, the N 1s XPS spectrum indicates similar chemical states at 399.1 and 400.6 and 404.7 ± 0.2 eV, while the newly appeared photoelectron peak at 396.8 ± 0.2 eV can be attributed to the metal (Ni) −N bond [40], indicating tight interfacial adhesion by covalent bonding between Ni₂P particles and the host polymer matrix. Figure 3C displays P 2p XPS spectra for P-GCN and 15-Ni₂P/P-GCN. The P 2p signal of P-GCN manifests as two singlet peaks at 129.8 and 133.2 ± 0.3 eV, in agreement with the literature data for P–P and P–N bonds, respectively [41]. This confirms the incorporation of P atoms into the g-C₃N₄ lattice during the hydrothermal synthesis, although a small amount of unreactive phosphorus is situated on the surface of the sample. In line with this, the P-GCN possesses a lower C/N atomic ratio than the pristine GCN, that is 0.72 versus 0.73, implying a carbon-deficient g-C₃N₄ structure, in which we assume C atoms are replaced by P atoms. The deviation of the C/N ratio from the ideal value (C:N = 0.75) for the GCN is likely due to the fracture of some C–N bonds in the triazine rings and, thus, the fragmentation of the g-C₃N₄ flakes during the hydrothermal process, in accordance with the TEM observations. Similar multiple valence states of P were also observed in the Ni₂P-modified sample. Specifically, the P 2p signal of 15-Ni₂P/P-GCN consists of three singlet peaks at 129.6, 133.5 and 134.7 ± 0.3 eV, which are assigned to the P–P and/or Ni–P coordination bonds as well as the P–N linkages and PO_x impurities due to the inadequate surface oxidation of P-doped GCN, respectively [42,43,44,45]. According to the XPS results, the P-doping content in the implanted samples is roughly 0.3 at.%. As for the Ni 2p_3/2 spectrum of 15-Ni₂P/P-GCN (Figure 3D), the photoelectron peak at 853.0 ± 0.2 eV points to the formation of the Ni–P coordination bonds, while the peak at 856.7 ± 0.2 eV together with the satellite signal at 862.0 ± 0.2 eV imply the presence of the paramagnetic Ni(II) ions owing to the partially oxidized Ni–PO_x/OH species [40,45,46,47]. Taking together, the results from the above studies unambiguously manifest the successful growth of Ni₂P nanoparticles and simultaneously doping of g-C₃N₄ lattice with phosphorus, highlighting the present synthetic method toward the fabrication of Ni₂P/P-doped g-C₃N₄ hetero-nanostructures with strong interfacial adhesion.

The porosity of the title materials was examined with nitrogen physisorption measurements. Figure S3 depicts the N₂ adsorption and desorption isotherms at −196 °C of the GCN_b, GCN, P-GCN and Ni₂P/P-GCN samples. All the isotherms showed type-IV curves with a H₃-type hysteresis loop, being characteristic of mesoporous systems with slit-shaped pores. Compared with the pristine GCN_b (83 m²·g⁻¹, 0.12 cm³·g⁻¹), GCN and P-GCN exhibit higher Brunauer-Emmett-Teller (BET) surface areas (113 and 97 m²·g⁻¹, respectively) and pore volumes (0.18 and 0.20 cm³·g⁻¹, respectively). The slightly larger porosity of the g-C₃N₄ samples obtained after hydrothermal treatment can be attributed to the smaller size of C₃N₄ layers that are randomly stacked together, forming an increased number of interstitial voids. The surface area and pore volume of the Ni₂P-modified catalysts were estimated to be 68–92 m²·g⁻¹ and 0.12–0.15 cm³·g⁻¹, respectively, which are slightly lower than the values of GCN and P-GCN due the incorporation of Ni₂P particles. Using the non-local density functional theory (NLDFT) analysis with slit pore geometry, the pore width in these materials was found to be ca. 3–3.3 nm (insets of the Figure S3). All the textural properties of the as-prepared materials are summarized in Table S2.

Ultraviolet-visible/near-IR (UV-vis/NIR) diffuse reflectance spectroscopy was performed to assess the optical properties of the prepared materials. As shown in Figure S4, the treated GCN exhibits an absorption onset at 2.81 eV, which corresponds to the band-to-band electronic transition of g-C₃N₄. Compared with the bulk GCN_b (it has a band gap of 2.75 eV), GCN shows a remarkable blue shift in energy gap by 60 meV, which is due to the small lateral size of the g-C₃N₄ layers, which are ca. 50–60 nm according to TEM results (Figure 2B). Meanwhile, doping the structure of g-C₃N₄ with P leads to a remarkable red shift of the optical absorption; on the basis of the UV-Vis/NIR spectrum, the energy band gap of P-GCN was estimated to be 2.76 eV, see Figure S4. Enriching the crystal lattice of g-C₃N₄ with P atoms has an immediate effect on the electronic band structure of the material due to the introduction of mid-gap electronic states near the conduction band (CB) edge. A similar effect has been previously observed in other P-doped g-C₃N₄ materials [48,49]. The Ni₂P-modified catalysts show a noteworthy red shift of the absorption spectrum and their band gap energy falls in the range of 2.70–2.74 eV. This remarkable red shift is attributed to the strong electronic interactions and charge transfer phenomena taking place between the tightly connected Ni₂P nanoparticles and P-doped g-C₃N₄ nanosheets in line with electron microscopy and XPS studies. Moreover, the Ni₂P/P-GCN catalysts exhibit an apparent absorption band below 2.7 eV (>460 nm) due to the optical absorption response of Ni₂P. The Ni₂P shows an intense absorption within the entire UV-vis/NIR region (see Figure S5), implying metallic behavior consistent with the literature data [46,50].

2.2. Photocatalytic Reduction of Cr(VI)

The photocatalytic activity of the Ni₂P/P-GCN heterostructures was evaluated in the UV-visible (λ > 360 nm) light reduction of Cr(VI) aqueous solution (50 mg·L⁻¹) without any additional hole scavenger. For comparison, we also studied the photocatalytic activity of the pristine (GCN) and P-doped (P-GCN) g-C₃N₄ samples under the same reaction conditions. The catalytic results shown in Figure 4A reveal that the deposition of Ni₂P has a notable impact on the photoactivity of g-C₃N₄, verifying that Ni₂P nanoparticles act as effective co-catalysts in the present catalytic system. Specifically, the conversion of Cr(VI) was complete (>99%) in 80 min with a 15-Ni₂P/P-GCN catalyst, whereas it took 3 h for ~44% and ~41% Cr(VI) conversion over GCN and P-GCN, respectively. The other Ni₂P-modified catalysts presented lower to moderate catalytic activity in the reduction of Cr(VI) than 15-Ni₂P/P-GCN, but still higher than those of GCN and P-GCN materials. Blank experiments revealed no obvious conversion of Cr(VI) in the absence of light or catalyst, confirming that Cr(VI) transformation is a photocatalytic process (Figure 4A). Also, when Ni₂P particles were used as a catalyst, no Cr(VI) conversion was detected under the present conditions (results not shown). The high reactivity of Ni₂P/P-GCN can be attributed to the improved charge-carrier transportation and spatial separation through the Ni₂P/P-doped g-C₃N₄ junctions due to the tight contact between Ni₂P and g-C₃N₄ components, as evidenced from electron microscopy and XPS results. To rule out this possibility, we also produced a reference catalyst by wet chemical deposition of 15 wt.% Ni₂P particles on the surface of g-C₃N₄ (denoted as 15-Ni₂P/GCN_b, see Materials and Methods for details), and we evaluated its photocatalytic activity under similar conditions. This catalyst exhibits very similar Ni₂P content, crystal structure, porosity and optical absorption abilities to those of 15-Ni₂P/GCN, as evidenced by EDS, XRD, N₂ physisorption and UV-vis/NIR spectroscopy measurements (see Table S1 and Figures S4, S6 and S7). Interestingly, 15-Ni₂P/GCN_b yielded inferior performance to 15-Ni₂P/P-GCN, even lower than that of pristine GCN and P-GCN, giving a respective ~15% Cr(VI) conversion in 3 h. The above results thus indicate that the P-doping and strong adhesion of Ni₂P particles to g-C₃N₄ have a positive effect on the Cr(VI) photoreduction performance of the resulting materials. Consequently, we focused our further catalytic studies on reactions with a 15-Ni₂P/P-GCN catalyst.

Next, we varied the catalyst loading and the solution pH to optimize the reaction conditions. Figure S8 compares the photocatalytic reduction of Cr(VI) caused by different mass loadings of 15-Ni₂P/P-GCN, i.e., from 0.4 to 1 g·L⁻¹. It can be seen that the Cr(VI) reduction efficiency increases with the catalyst concentration and reaches a maximum of 0.8 g·L⁻¹. As for the slightly inferior photoactivity at 1 g·L⁻¹ load of catalyst, it could be associated with the increased light scattering by the particle suspension. Besides catalyst concentration, the solution acidity has a prominent effect on the Cr(VI) photoreduction process. As shown in Figure S9, the 15-Ni₂P/P-GCN catalyst shows a significantly increased Cr(VI) reduction activity as the solution pH changes from 6 to 1 (keeping the catalyst load constant at 0.8 g·L⁻¹). Of particular note, in pH 1 solution, 15-Ni₂P/P-GCN eliminates >99% of Cr(VI) (50 mg·L⁻¹) in only 50 min. We interpret this behavior as the propitious adsorption of Cr(VI) oxo species (predominantly existent in the HCrO₄⁻ form at pH < 4) on the catalyst surface; g-C₃N₄ has a point of zero charge ~4–5 and, thus, at low pH environment its surface is positively charged. Also, in an acidic solution, the excessive concentration of hydronium ions (H₃O⁺) has an accelerating action on the reduction process of Cr(VI), as shown by the overall reaction in Equation (1).

2HCrO₄⁻_(aq) + 8H⁺_(aq) → 2Cr³⁺_(aq) + ³/₂O_2(g) + 5H₂O_(l)

(1)

Under optimal reaction conditions (0.8 g·L⁻¹ catalyst load), the photoreduction rate of Cr(VI) over 15-Ni₂P/P-GCN reaches 15.9 μmol·h⁻¹ at pH 2 under UV-visible (λ > 360 nm) light irradiation. Also, the apparent quantum yield (QY) of this process is calculated to be as high as 12.5 % at 410 ± 10 nm, assuming full absorption of the incident light by the catalyst. The activity of 15-Ni₂P/P-GCN is comparable and even superior to that of some of the state-of-the-art g-C₃N₄-based catalysts for Cr(VI) reduction, even if operating with electron donor additives (see Table S3).

The practical applicability of the Ni₂P/P-GCN catalytic system can be seen in the decontamination of wastewater containing Cr(VI) along with large amounts of various competitive ions. This study was conducted using the same load of 15-Ni₂P/P-GCN catalyst (0.8 g·L⁻¹) in Cr(VI) aqueous solution (50·mg·L⁻¹) without and in the presence of SO₄²⁻, CO₃²⁻, NO₃⁻, Cl⁻, Na⁺, Cu²⁺, Mn²⁺ and Zn²⁺ ions (50·mg·L⁻¹ for each ion), which is a typical composition of industrial effluents [51,52]. Commonly, the presence of interfering ions (such as SO₄²⁻, NO₃⁻, Cl⁻, Mn²⁺ and Zn²⁺) deteriorates the Cr(VI) reduction ability of the catalyst due to the competitive adsorption effects and redox reactions at the catalyst surface [53,54]. Results from Figure 4B show that 15-Ni₂P/P-GCN still demonstrates a high Cr(VI) photoreduction activity (>98% Cr(VI) reduction in 80 min), even under the presence of excessive amounts of competing ions in the solution. In addition, the photo-oxidation kinetics of 15-Ni₂P/P-GCN were also studied in the presence of EDTA, citric acid and phenol as common organic pollutants of wastewaters, which can serve as sacrificial electron donors. Since electron injection from these organic molecules can overcome the rate limitations of the water oxidation reaction (H₂O oxidation proceeds through a four-electron transfer pathway involving various sluggish absorption/dissociation steps), the overall photocatalytic reaction is expected to occur at a faster rate. Figure 4C shows that the reduction reaction of Cr(VI) occurs at a remarkably faster rate when three equivalents of the above pollutants were added to the solution, yielding >99% Cr(VI) conversion in only 5–30 min. These findings explicitly demonstrate that organic pollutants, such as EDTA, citric acid and phenol, may have a positive effect on the photoreduction kinetics of Cr(VI).

The 15-Ni₂P/P-GCN also demonstrated very good stability and reusability after multiple cycles of Cr(VI) reduction. The stability of the catalyst was examined by performing three consecutive 80 min photocatalytic tests. Before the catalytic test, the catalyst was separated from the reaction mixture through centrifugation, washed with DI water and added to a fresh Cr(VI) solution. The recycling study showed that 15-Ni₂P/P-GCN can be reused for at least three successive catalytic runs without a significant decrease of its activity, maintaining a Cr(VI) conversion efficiency as high as >98% after catalytic tests (Figure 4D). Moreover, EDS and XRD measurements revealed no obvious changes in the chemical composition and crystal structure of the reused 15-Ni₂P/P-GCN catalyst, indicating high durability (see Table S1 and Figure S10).

2.3. Effect of Ni₂P on the Photocatalytic Activity

To better understand the role of Ni₂P on the photocatalytic performance of Ni₂P/P-GCN, we performed electrochemical measurements. First, Mott-Schottky (M-S) analysis was used to assess the electronic band structure of the prepared materials (drop-casted on FTO substrates), and the respective M-S plots (1/Csc² vs. applied potential curves) are depicted in Figure 5A. It can be seen that all the examined catalysts exhibit positive slopes, suggesting n-type conductivity. By extrapolating the linear fits of the M-S slopes to 1/Csc² = 0, the flat-band (E_FB) potentials of the samples were estimated and the obtained results are listed in

Table 1. Electrochemical results deduced from EIS analysis for various catalysts.

Sample	EFB (V vs. RHE)	EVB (V vs. RHE)	Donor Density, ND (cm−3)	Rct (Ohm)
GCN_b	−1.22	1.53	1.72 × 1018	661
GCN	−1.28	1.53	3.17 × 1018	436
P-GCN	−1.15	1.61	7.12 × 1017	514
5-Ni2P/P-GCN	−1.16	1.58	5.30 × 1017	311
10-Ni2P/P-GCN	−1.14	1.60	2.28 × 1017	233
15-Ni2P/P-GCN	−1.09	1.64	1.53 × 1017	203
20-Ni2P/P-GCN	−1.07	1.63	1.71 × 1017	261
15-Ni2P/GCN_b	−1.08	1.66	6.60 × 1017	777

. All electrochemical potentials are given relative to the reversible hydrogen electrode (RHE) at pH 7. Considering that E_FB is a good proxy value of the conduction band (CB) minimum for n-type doped semiconductors (usually E_FB is 0.1–0.3 V more positive than the CB edge) [55], the valence band (VB) potential (E_VB), and thus a reasonable band-edge diagram, were obtained for each catalyst by adding the E_FB to the respective energy band gap (as estimated from the UV-vis/NIR absorption measurements). As shown in Table 1 and Figure 5B, the E_FB position of GCN is located at −1.28 V, which is cathodically shifted by ~60 meV relative to the bulk GCN_b (−1.22 V). Notably, this negative shift in E_FB of GCN correlates well with its band gap widening after hydrothermal treatment, which is associated with the confined environment in the nanometer-sized g-C₃N₄ layers. In addition, the charge donor density (N_D) of GCN (3.17 × 10¹⁸ cm⁻³), calculated from the magnitude of the M-S slope, is almost double that of GCN_b (1.72 × 10¹⁸ cm⁻³). Overall, the hydrothermal process and the consequent fragmentation of g-C₃N₄ layers endow the GCN catalyst with a higher reductive ability and increased charge donor density, resulting in an improved photocatalytic activity compared to the bulk GCN_b in agreement with the catalytic results in Figure 4A. Compared to GCN, the P-GCN shows an anodic shift in the E_FB position to −1.15 V, probably due to the mid-gap electronic states introduced by the P dopant. As for the Ni₂P/P-GCN catalysts, the growth of Ni₂P nanoparticles on the surface of P-doped g-C₃N₄ has a prominent effect on the electronic structure of the heterostructures. Specifically, the M-S results reveal a progressively anodic shift of the E_FB level from −1.16 V to −1.07 as the Ni₂P content increases from 5 to 20 wt.%. This shift in E_FB towards positive potentials can be explained by the formation of Schottky junctions between the P-doped g-C₃N₄ nanosheets and Ni₂P nanoparticles. As indicated earlier from the optical absorption spectra, the Ni₂P shows a broad absorption capacity throughout the entire UV-Vis/NIR range, signaling metallic behavior. From the M-S plot analysis, the E_FB potential of bare Ni₂P was determined at −0.1 V (ca. 4.4 eV vs. vacuum level) (inset of Figure 5A), which is very close to the literature values of 4.8–4.9 eV [45,56] and much lower than those for GCN and P-GCN. This means that, during the growth of Ni₂P on the g-C₃N₄ surface, there is an electron flow from g-C₃N₄ to Ni₂P until the Fermi levels of these components reach equilibrium. The net effect of this process is expected to cause a gradual potential drop of the g-C₃N₄ CB band-edge as the Ni₂P content increases. In line with this, the systematic decrease in N_D from 5.30 × 10¹⁷ to 1.53 × 10¹⁷ cm⁻³ as the Ni₂P content in Ni₂P/P-GCN increases from 0 to 20 wt.% attest to a progressive electron injection from P-doped g-C₃N₄ to Ni₂P, see Table 1. Comparably, the 15-Ni₂P/GCN_b undergoes an almost identical band structure deformation with the 15-Ni₂P/P-GCN (it shows an E_FB at −1.08 V), showing a significantly higher donor density (~6.59 × 10¹⁷ cm⁻³) than 15-Ni₂P/P-GCN. These results can be interpreted by a less favorable charge-transfer dynamic across the 15-Ni₂P/GCN_b interface, which probably results from the poorer electronic communication between the bulk g-C₃N₄ and Ni₂P microparticles. The inferior interfacial electron transfer kinetics of 15-Ni₂P/GCN_b were further supported by time-resolved photoluminescence measurements (see below).

The charge-transfer dynamics at the catalyst/liquid interface were investigated with electrochemical impendence spectroscopy (EIS) measurements. Figure 5C shows the Nyquist plots of the studied catalysts. In order to calculate the charge-transfer resistance (R_ct) of each catalyst, the EIS data were fitted using a Randles equivalent circuit model (inset of Figure 5C), and the calculated R_ct values are listed in Table 1. The simulated EIS results indicated a markedly lower R_ct value (~436 Ω) for the GCN, which is about 1.5 times lower than that of GCN_b (~661 Ω), implying more favorable charge-transfer kinetics for the small-sized g-C₃N₄ layers. On the contrary, P-doping seems to cause a moderate effect on the charge transfer properties of GCN, as reflected by the slightly larger R_ct value of P-GCN (~514 Ω). The above findings are in good agreement with the observed trend in the catalytic activity of GCN, P-GCN and GCN_b (Figure 4A). The Ni₂P/P-GCN catalysts, on the other hand, exhibit R_ct values of ~203–311 Ω, which are substantially lower than those of GCN and P-GCN, manifesting a more efficient charge transfer effect between the Ni₂P and P-doped g-C₃N₄ layers. Among the examined catalysts, 15-Ni₂P/P-GCN demonstrates the lowest resistance in charge transfer, which correlates well with its superior photocatalytic performance. Of note, the bulk reference catalyst (15-Ni₂P/GCN_b) showed sluggish charge transfer kinetics across the catalyst/liquid interface, judging by its higher R_ct (~777 Ω). These results confirm, once again, that Ni₂P modification and P-doping of the g-C₃N₄ structure have a positive impact on the charge transfer properties and, thus, the photochemical efficiency of Ni₂P/P-GCN heterostructures.

The dynamics of charge-carrier recombination were also studied by time-resolved photoluminescence (TR-PL) decay measurements at room temperature using an excitation laser wavelength of 375 nm. Figure 5D shows the TR-PL decay profiles for the GCN_b, GCN, P-GCN, 15-Ni₂P/P-GCN and 15-Ni₂P/GCN_b materials. To calculate the PL lifetimes, a bi-exponential function $\mathrm{I}\left(\mathrm{t}\right)={\sum }_{\mathrm{i}}{\mathsf{\alpha }}_{\mathrm{i}}{\mathrm{e}}^{–\mathrm{t}/{\mathsf{\tau }}_{\mathrm{i}}}$, where α_i is the fraction of each component (${\sum }_{\mathrm{i}}{\mathsf{\alpha }}_{\mathrm{i}}=1)$ and τ_i is the photocarrier lifetime, was used to properly fit the PL decay data. Here we used an equation with two-component (i = 1, 2) which account for the radiative charge-carrier relaxation at the surface (fast) and in bulk (slow), respectively. Through this analysis, the average lifetime (τ_av) of samples was also calculated using the Equation (2) and all the fitting parameters are listed in Table S4.

$${\mathsf{\tau }}_{\mathrm{av}}=\frac{\left({\sum }_{\mathrm{i}}{\mathsf{\alpha }}_{\mathrm{i}}{\mathsf{\tau }}_{\mathrm{i}}^2\right)}{\left({\sum }_{\mathrm{i}}{\mathsf{\alpha }}_{\mathrm{i}}{\mathsf{\tau }}_{\mathrm{i}}\right)} (\mathrm{i}=1, 2)$$

(2)

The TR-PL results showed that the τ_av of charge-carriers in GNC (4.38 ns) is slightly higher compared to the bulk GCN_b (4.20 ns), justifying a better charge dissociation and transfer caused by the low dimensionality of g-C₃N₄ layers in agreement with the EIS data. The P-GCN exhibits a τ_av value of 4.32 ns, which is similar to that obtained for GCN, suggesting that P-doping has a minor effect on the charge recombination kinetics of g-C₃N₄. Meanwhile, a prolonged carrier lifetime of 5.11 ns was obtained for 15-Ni₂P/P-GCN, reflecting a more efficient utilization of photogenerated electrons and holes for photocatalytic reactions. In agreement with the respective EIS results, the 15-Ni₂P/GCN_b catalyst showed a substantially shorter lifetime of the photogenerated carriers (3.76 ns), which agrees with the poor electronic contact and low charge-carrier dissociation in this catalyst. Therefore, the combined results from the EIS and TR-PL data clearly demonstrate the enhanced interfacial charge separation at the Schottky Ni₂P/P-GCN nanojunctions, which improves the charge transfer dynamics and catalytic reaction kinetics.

On the basis of the above results, we suggest a possible reaction scheme for the photocatalytic conversion of Cr(VI) by the Ni₂P/P-GCN catalysts. As illustrated in Scheme 1, under light illumination, the photogenerated electrons at the CB of g-C₃N₄ will transfer to Ni₂P due to the intrinsic electric field formed at the Schottky Ni₂P/P-g-C₃N₄ junctions, where they efficiently convert Cr(VI) ions (mainly in HCrO₄⁻ form) to Cr(III). Evidence for such charge transfer pathways is obtained from the EIS and TR-PL studies. In parallel, the photogenerated holes on the surface of g-C₃N₄ can oxidize the water to generate oxygen during the photocatalytic process. The energy band diagrams in Figure 5B show that the VB levels of Ni₂P/P-GCN catalysts are located below the potential of water oxidation (0.82 V at pH 7), thus demonstrating the capability of these materials for photocatalytic water splitting and oxygen evolution reaction.

3. Materials and Methods

3.1. Chemicals and Materials

Urea (CO(NH₂)₂, 99.5%), red phosphorus and nickel(II) chloride hexahydrate (NiCl₂^.6H₂O, ≥98%) were acquired from Sigma-Aldrich (Steinheim, Germany). Absolute ethanol was purchased from Fisher Scientific Company (Waltham, MA, USA). In all procedures, double-deionized (DI) water was used. All chemicals were used as received without any further purification.

3.2. Synthesis of g-C₃N₄

Porous g-C₃N₄ was synthesized by modification of previously reported procedures [57]. In a typical synthesis, 10 g of urea was inserted into a porcelain crucible enclosed with a Pyrex petri plate. The crucible was then positioned in a furnace and heated at 550 °C for 1 h (temperature rate: 30 °C min⁻¹). The final product (denoted as GCN_b) was collected and ground into a fine powder.

For comparison, g-C₃N₄ nanosheets (denoted as GCN) were also prepared by hydrothermal treatment of GCN_b (~5 mg/mL) at 150 °C for 12 h.

3.3. Synthesis of Ni₂P/g-C₃N₄ Heterostructures

The in-situ growth and deposition of Ni₂P nanoparticles on g-C₃N₄ nanosheets was achieved through a hydrothermal process. Briefly, 100 mg of as-made g-C₃N₄ and certain amounts of NiCl₂ and red phosphorus (molar ratio Ni:P = 1:5) were dispersed in 20 mL of DI water by vigorous stirring for 30 min. The suspension was then placed in a Teflon-lined autoclave reactor and heated at 150 °C for 12 h. Finally, the dark-grey colored product was collected by centrifugation, thoroughly washed with DI water and ethanol and dried at 100 °C for 12 h. The NiCl₂ and red phosphorus concentration was varied in the reaction mixture to give Ni₂P/P-g-C₃N₄ heterostructures (denoted as x-Ni₂P/P-GCN) with different content of Ni₂P, i.e., x = 5, 10, 15 and 20 wt.%.

For comparison purposes, we also prepared pure Ni₂P microparticles and P-doped g-C₃N₄ samples. The Ni₂P was obtained following the same procedure, without the addition of g-C₃N₄. For the synthesis of P-doped g-C₃N₄ (denoted as P-GCN), the same procedure as that for Ni₂P/P-g-C₃N₄ with 15% Ni₂P content was followed but without the addition of NiCl₂. We also synthesized a 15% Ni₂P/g-C₃N₄ sample by using a wet chemical deposition of Ni₂P microparticle on the g-C₃N₄ surface (denoted as 15-Ni₂P/GCN_b).

3.4. Physicochemical Characterization

Scanning electron microscopy (SEM, JEOL Ltd., Tokyo, Japan) and energy-dispersive X-ray spectroscopy (EDS, JEOL Ltd., Tokyo, Japan) were performed on a JEOL JSM-IT700HR microscope equipped with a JED-2300 detector. The transmission electron microscopy (TEM) measurements were conducted on a JEM-2100 microscope operating at 200 kV (JEOL Ltd., Tokyo, Japan). A Panalytical X’pert Pro MPD diffractometer was used to obtain the X-ray diffraction (XRD, Malvern Panalytical, Almelo, The Netherlands) patterns, using Cu-Ka radiation at λ = 1.5418 A at 45 kV voltage and 40 mA current. The porosity of the samples was examined through N₂ physisorption at −196 °C using a Quantachrome NOVA 3200e analyzer (Quantachrome Instruments, Boynton Beach, FL, USA). Prior to each measurement, all the samples were degassed at 100 °C for 12 h. The specific surface areas were calculated by applying the Brunauer-Emmett-Teller method to the adsorption data at a relative pressure (P/P₀) of 0.04–0.24 and the total pore volumes were obtained at a P/P₀ of 0.98. X-ray photoelectron spectroscopy (XPS, SPECS Surface Nanon Analysis, Berlin, Germany) was performed using a SPECS spectrometer equipped with a Phoibos 100 1D-DLD energy analyzer, using Al Ka radiation (1486.6 eV). The binding energy was corrected in regards to the signal of the adventitious carbon (284.8 eV). UV-vis diffuse reflectance spectroscopy was carried out on a Shimadzu UV-2600 spectrophotometer (UV-2600, Shimadzu Co., Kyoto, Japan), using BaSO₄ fine powder as a 100% reflectance reference. The diffuse reflectance data were converted to absorbance using the Kubelka-Munk function: α/S = (1 − R)²/(2R), where R is the reflectance, α is the absorption coefficient and S is the scattering coefficient. Time-resolved photoluminescence (TR-PL, Edinburgh Ltd., Livingston, UK) spectra were obtained using an Edinburgh FS5 spectrofluorometer equipped with a 375 nm pulsed laser.

3.5. Photocatalytic Reactions

The photocatalytic Cr(VI) reductions were carried out in a Pyrex glass flask (100 mL capacity) containing 50 mL of Cr(VI) aqueous solution and different concentrations of catalyst (0.4–1·g L⁻¹). The Cr(VI) solution (50 ppm) was prepared by dissolving K₂Cr₂O₇ in DI water, and the pH of the solution was adjusted to the desired values with dilute H₂SO₄. The mixture was stirred in the dark before the irradiation for 1 h to establish adsorption/desorption equilibrium between the photocatalyst and Cr(VI). The reaction solution was cooled to 20 ± 2 °C using a water-cooling system and irradiated at λ > 360 nm light using a 300-W Xenon lamp (Variac Cermax). During the reaction, the Cr(VI) content in the mixture was determined using the 1,5-diphenylcarbazide (DPC) colorimetric method (detection limit: 5 μg·L⁻¹) on a Perkin Elmer Lambda 25 UV–vis spectrometer (Perkin Elmer Inc., Waltham, MA, USA). The normalized concentration (C_t/C_o) of the Cr(VI) solution at different illumination times was considered in proportion to the absorbance of the DPC-Cr(VI) complex at 540 nm. The pH of the solution was adjusted by adding 2 M H₂SO₄ or 2 M NaOH.

Incident photon to energy conversion efficiency (quantum yield, QY) at a λ = 410 ± 10 nm irradiation wavelength was determined using the following equation:

$$\mathrm{QY}=\frac{3 \times \mathrm{number} \mathrm{of} \mathrm{reduced} \mathrm{Cr}\left(\mathrm{VI}\right)}{\mathrm{number} \mathrm{of} \mathrm{incident} \mathrm{photons}}$$

(3)

The number of incident photons was determined using a StarLite power meter using an FL400A-BB-50 thermal sensor (Ophir Optronics Ltd., Jerusalem, Israel).

3.6. Electrochemical Measurements

Electrochemical impedance spectroscopy (EIS, Princeton Applied Research, Princeton, NJ, USA) was carried out using a VersaSTAT 4 electrochemical station. The electrochemical three-electrode cell consisted of a working electrode (FTO coated with the catalyst), a reference electrode (Ag/AgCl in saturated KCl) and a counter electrode (Pt wire) in a 0.5 M Na₂SO₄ electrolyte (pH = 6.8). Prior to each measurement, the electrolyte was purged with Ar for at least 30 min to remove the dissolved oxygen. For the preparation of the working electrodes, 100 μL of uniform suspension containing the catalyst in 0.5 mL DMF (~10 mg·mL⁻¹) and 20 μL of 5% Nafion solution was drop-casted on the surface of fluorine-doped tin oxide (FTO, 10 Ω/sq) substrate (1 cm² effective area) and then the film was dried at 60 °C for 1 h. The space-charge capacitance (C_sc) of the samples was measured at 1 kHz with a 10 mV AC voltage amplitude. All the measured potentials were converted into the reversible hydrogen electrode (RHE) at pH = 7 using the equation:

E_RHE = E_Ag/AgCl+0.197+0.059 × pH

(4)

where E_Ag/AgCl depicts the measured potential versus Ag/AgCl.

The donor density (N_D) of the samples was calculated from the slope of the linear portion of the Mott-Schottky plots using the following equation:

$${\mathrm{N}}_{\mathrm{D}}=\frac{{\mathrm{C}}_{\mathrm{SC}}^2{·2·(\mathrm{E}–\mathrm{E}}_{\mathrm{FB}})}{{\mathrm{A}}^2{·\mathsf{\epsilon }·\mathsf{\epsilon }}_{\mathrm{o}}{·\mathrm{e}}_{\mathrm{o}}}$$

(5)

where C_sc is the space charge capacitance, E_FB is the flat band potential of the semiconductor, E is the applied potential, A is the exposed area of the sample (cm²), ε is the relative dielectric constant of the semiconductor (7.6 for g-C₃N₄) [58], ε₀ is dielectric permittivity under vacuum (8.8542 × 10⁻¹⁴ F·cm⁻¹) and e₀ is the elementary charge (1.602 × 10⁻¹⁹ C).

The Nyquist plots were recorded over a frequency range from 1 Hz to 10 KHz using an applied voltage of −1.3 V (vs. Ag/AgCl, saturated KCl) in a 0.5 M Na₂SO₄ solution (pH = 6.8). The electrochemical data were fitted using ZView Software (Scribner, Southern Pines, NC, USA).

4. Conclusions

In summary, 2D/3D hybrid heterostructures consisting of small-sized g-C₃N₄ nanosheets (ca. 50–60 nm in lateral size) and Ni₂P nanoparticles (ca. 10–12 nm in diameter) have been successfully prepared via a two-step thermal polycondensation and hydrothermal process. The chemical, structural and morphological characteristics of these materials were investigated by a combination of energy-dispersive X-ray spectroscopy (EDS), X-ray diffraction (XRD), electron microscopy (SEM/TEM), X-ray photoelectron spectroscopy (XPS) and N₂ physisorption. The Ni₂P/P-doped g-C₃N₄ hetero-nanostructures adopt a suitable electronic structure that allows efficient transfer of the photogenerated charge carriers across the interface. UV-vis/NIR absorption, electrochemical and time-resolved photoluminescence (TR-PL) spectroscopic studies prove that Ni₂P modification and P doping of the g-C₃N₄ structure leads to the formation of Schottky Ni₂P/P-doped g-C₃N₄ junctions that promote a better separation and utilization of charge carriers via a faster electron injection from g-C₃N₄ to Ni₂P. Owing to these attributes, the catalyst with 15 wt.% Ni₂P content attains outstanding photocatalytic C(VI) reduction activity (>99% Cr(VI) conversion in 80 min at pH 2), reaching an apparent quantum yield of 12.5% at 410 nm and, importantly, without the addition of any sacrificial reagents. This activity is among the highest reported for g-C₃N₄-based photocatalysts. In addition, the Ni₂P/P-doped g-C₃N₄ catalysts demonstrated persistent activity in the reduction of Cr(VI) even with the coexistence of other competing ions that commonly exist in industrial wastewater. The results of this study offer a mechanistic understanding of the charge transport dynamics at the Ni₂P/P-doped g-C₃N₄ interface and demonstrate the potential viability of the present catalytic system for photochemical applications and environmental protection.