2D/2D Heterojunctions of Layered TiO₂ and (NH₄)₂V₃O₈ for Sunlight-Driven Methylene Blue Degradation

Abstract

Photocatalysis based on titanium dioxide (TiO₂) has become a promising method to remediate industrial and municipal effluents in an environmentally friendly manner. However, the efficiency of TiO₂ is hampered by problems such as rapid electron–hole recombination and limited solar spectrum absorption. Furthermore, the sensitization of TiO₂ through heterojunctions with other materials has gained attention. Vanadium, specifically in the form of ammonium vanadate ((NH₄)₂V₃O₈), has shown promise as a photocatalyst due to its ability to effectively absorb visible light. However, its use in photocatalysis remains limited. Herein, we present a novel synthesis method to produce lamellar (NH₄)₂V₃O₈ as a sensitizer in a supramolecular hybrid photocatalyst of TiO₂–stearic acid (SA), contributing to a deeper understanding of its structural and magnetic characteristics, expanding the range of visible light absorption, and improving the efficiency of photogenerated electron–hole separation. Materials, such as TiO₂–SA and (NH₄)₂V₃O₈, were synthesized and characterized. EPR studies of (NH₄)₂V₃O₈ demonstrated their orientation-dependent magnetic properties and, from measurements of the angular variation of g-values, suggest that the VO₂⁺ complexes are in axially distorted octahedral sites. The photocatalytic results indicate that the 2D/2D heterojunction layered TiO₂/vanadate at a ratio (1:0.050) removed 100% of the methylene blue, used as a model contaminant in this study. The study of the degradation mechanism of methylene blue emphasizes the role of reactive species such as hydroxyl radicals (^•OH) and superoxide ions (O₂^•−). These species are crucial for breaking down contaminant molecules, leading to their degradation. The band alignment between ammonium vanadate ((NH₄)₂V₃O₈) and TiO₂–SA, shows effective separation and charge transfer processes at their interface. Furthermore, the study confirms the chemical stability and recyclability of the TiO₂–SA/(NH₄)₂V₃O₈ photocatalyst, demonstrated that it could be used for multiple photocatalytic cycles without a significant loss of activity. This stability, combined with its ability to degrade organic pollutants under solar irradiation, means that the TiO₂–SA/(NH₄)₂V₃O₈ photocatalyst is a promising candidate for practical environmental remediation applications.

1. Introduction

Photocatalysis based on titanium dioxide has garnered significant attention due to its promising ability to facilitate the remediation and reuse of industrial and municipal effluents in an environmentally benign manner [1,2,3]. Nonetheless, its efficiency is hampered by a high rate of hole-electron recombination and scarce absorption of the solar spectrum [4,5]. To enhance TiO₂ performance, researchers have explored modifications to its electronic structure and morphology. These efforts involve introducing dopants, both metallic and non-metallic elements, and creating mixed phases within the same semiconductor (e.g., TiO₂-P25 and black titania) to introduce new intra-bandgap levels [6,7,8]. Recent observations indicate that molecularly thin TiO₂ sheets coated on both sides with monolayers of fatty acids outperform pristine TiO₂ [9,10]. This improvement can be attributed to various factors, such as the creation of surfaces with abundant defects, the conjugation of nano–micro dimensions in 2D structures, and the increased hydrophobicity of particle surfaces, which is similar to that of organic contaminants. Improving the photocatalytic performance of both modified anatase and lamellar derivatives is usually limited. A more effective alternative often entails sensitizing TiO₂ through heterojunctions with species that not only establish electric fields to reduce electron–hole recombination rates but also as antennas, capturing lower-energy light and transferring it to TiO₂ [11,12].

Vanadium, a member of the first transition series ([Ar] 4s² 3d³), forms stable aqueous species in various oxidation states (ranging from 2⁺ to 5⁺). Vanadium(V) oxide, V₂O₅, possesses a bandgap of 2.4–2.8 eV [13], allowing it to effectively absorb visible light [14,15]. However, its utility as a photocatalyst is limited [16]. Consequently, research has shifted its focus towards V(V)-oxosalts, especially metal metavanadates, and compounds like potassium vanadate such as K₂V₆O₁₆ and KV₃O₈ [17]. These materials have displayed photocatalytic activity, primarily due to the presence of V(IV) [17,18]. Furthermore, V₂O₅ has proven to be effective as a sensitizer for wide-bandgap semiconductors like SnO₂ [19] and TiO₂ [20], contributing to hydrogen evolution and tetracycline degradation, respectively. Layered hybrid nanostructures, exemplified as mixed-valence V(V)/V(IV) vanadium pentoxide (VO_x) nanotubes produced via hydrothermal treatment, have demonstrated their ability to sensitize ZnO for contaminant degradation under simulated sunlight [21]. Inorganic VO_x nanostructures, including microsquares belonging to the V₇O₁₆ family [22], have received reduced attention regarding their photocatalytic properties [23,24] (NH₄)₂V₃O₈, which crystallizes in the fresnoite structure with a formal vanadium ion ratio of V⁵⁺/V⁴⁺ = 2/1, has gained recognition for its magnetic properties [25,26,27] and potential use as an electrode material for lithium-ion and zinc-ion batteries [28,29,30,31]. Various studies have sought to optimize the synthesis of (NH₄)₂V₃O₈ to elucidate the distinctive structure–property relationship stemming from its 2D molecular architecture and paramagnetic d¹ electronic configuration of V(IV) [30,32,33]. Magnetic susceptibility at different temperatures and magnetic behavior has been explored using both polycrystalline [25] and oriented single-crystal samples [34]. However, to the best of our knowledge, the photocatalytic properties of (NH₄)₂V₃O₈ have not been reported thus far.

In this study, we present a new synthesis method to produce (NH₄)₂V₃O₈, X-ray diffraction (XRD) studies reveal its structure. Moreover, we investigated its orientation-dependent magnetic properties using electron paramagnetic resonance (EPR) and, evaluate its potential as a sensitizer in a TiO₂-based supramolecular hybrid photocatalyst.

Our research not only contributes to a deeper understanding of the structural and magnetic characteristics of (NH₄)₂V₃O₈, but also highlights that the TiO₂/(NH₄)₂V₂O₈ (1:0.05) heterojunction, which comprises around 5% of this vanadate can significantly improve the degradation rate of methylene blue. The degradation of this contaminant being approximately three times faster than with TiO₂–stearic acid under simulated sunlight.

2. Materials and Methods

2.1. Chemicals

All chemical reagents used in this study were of guaranteed grade and were obtained from commercial sources, Ti[OCH(CH₃)₂]₄ (titanium tetraisopropoxide), C₁₈H₃₆O₂ (stearic acid), NH₄VO₃ (ammonium metavanadate), C₁₆H₃₅N (hexadecylamine), (C₂H₄O₂) glacial acetic acid, NaOH (sodium hydroxide), C₁₆H₁₈ClN₃S (methylene blue), C₈H₆O₄ (terephthalic acid), C₆H₄O₂ (benzoquinone) and (C₂H₈N₂O₄) ammonium oxalate were purchased from Sigma-Aldrich, Saint Louis, MO, USA. (C₂H₆O) ethanol and C₃H₈O (isopropanol) were purchased from Merck KGaA, Darmstadt, Germany. All reagents were of analytical grade and were used without further purification.

2.2. Sample Preparation

Diammonium trivanadate (NH₄)₂V₃O₈ was synthesized using a straightforward hydrothermal method. In a typical synthesis, 0.936 g of NH₄VO₃ and 0.966 g of hexadecylamine (HDA) were dissolved in 40 mL of ethanol with vigorous stirring. Subsequently, 4 mL of glacial acetic acid was added to the solution. After 30 min of stirring, 40 mL of deionized water (18 MΩ) was introduced, resulting in the development of a yellow color in the solution. The solution was then sealed and kept under continuous agitation in the dark. The pH of the resulting dispersion was adjusted to 6.0 using a 0.1 M NaOH solution and transferred into a Teflon-lined stainless-steel autoclave with a 15 mL capacity. The autoclave was heated to 180 °C and maintained at this temperature for 24 h. After cooling to room temperature, the resulting dark violet crystalline product was separated by centrifugation and subjected to several washes with ethanol and acetone. Finally, the sample was dried at 60 °C overnight in the open air.

Titanium Dioxide-Stearic acid Nanocomposite (TiO₂–SA) was synthesized following a procedure detailed in our prior work [10]. Briefly, stearic acid was blended with ethanol in an argon atmosphere and stirred for 3 h at 50 °C. Subsequently, titanium tetraisopropoxide (TTIP) was added dropwise, and the mixture was continuously stirred for an additional hour at the same temperature. This suspension was then left to age at room temperature for a week. The resulting white precipitate was separated by centrifugation, thoroughly washed with ethanol, and then dried at 45 °C for a period of 72 h.

TiO₂–SA/(NH₄)₂V₃O₈ ratios were prepared in the range of 1:0.025; 1:0.05; 1:0.10 and 1:0.15 (weight ratio). These samples were mixed manually using an agate mortar. Before exposure to irradiation, the suspension was subjected to ultrasonication for 2 min to ensure the creation of a homogeneous sample, promoting intimate contact, and reducing aggregates in sample preparation [35].

2.3. Characterization Methods

A single crystal of (NH₄)₂V₃O₈ was analyzed for intensity data collection on a Nonius Kappa CCD FR590 diffractometer with mirror-monochromatized Mo-Kα radiation (λ = 0.71073 Å) to confirm the cell and resolve the structure. The structures of all compounds were determined by direct methods and refined by full-matrix least-squares procedures using the SHELXTL package. The details of the crystal parameters, data collection and refinements for all the complexes are summarized in Tables S1–S4.

Samples were subjected to powder X-ray diffraction (XRD) analyses using a D8 Advance diffractometer (Bruker AXS, Karlsruhe, Germany) instrument with Cu Kα radiation (λ = 1.5418 Å). Continuous wave (CW) electron paramagnetic resonance (EPR) experiments were conducted at temperatures of 5.5 K and 300 K utilizing a Elexsys E580 spectrometer (Bruker Corporation, Billerica, MA, USA) operating at approximately 9.476 GHz. The spectrometer was outfitted with a continuous flow liquid helium cryogenic system (Oxford ESR900) and a temperature controller Oxford ITC503 (both Oxford Instruments, Abingdon, UK). To avoid saturation effects, the microwave power was consistently attenuated by at least 6 dB below the saturation threshold. Modulation frequency is set to 100 kHz. To prevent overmodulation and passage effects, the modulation amplitude and amplifier time constant were set to at least one tenth of the narrowest line width and shortest transit time, respectively. A single tiny crystal of (NH₄)₂V₃O₈ with planar dimensions measuring approximately 1 mm × 0.5 mm is stuck with silicone vacuum grease to one of the faces of a cubic block of KCl crystal of 2 mm × 2 mm × 2 mm. The width of the crystal was very small (much smaller than one mm) and, due to the material’s delicate nature, it could not be measured. Subsequently, the KCl block and sample assembly were attached using the same grease to a base crafted from rexolite and secured at the bottom of a Wilmad EPR quartz sample tube (4 mm inner diameter) (Wilmad Glass, ATS Products, Warminster, PA, USA). For field calibration, an almost imperceptible grain of DPPH free radical (2,2–diphenyl–1–picrylhydrazil), with a g-factor of 2.0037, was carefully positioned adjacent to the sample, serving as an EPR marker. To facilitate rotation of the sample around the microwave cavity’s symmetry axis, the sample tube was connected to a ER218PG1 goniometer (Bruker Corporation, Billerica, MA, USA). The goniometer is capable of conducting accurate and automated measurements of EPR spectra while varying both the magnetic field and goniometer angle. The initial alignment of the sample relative to the applied magnetic field is achieved visually, which may introduce a minor error of a few degrees in the estimation of initial angles. Scanning electron microscopy (SEM) images were acquired using a ZEISS Gemini SEM 360 NanoVP Carl microscope (Carl Zeiss AG, Oberkochen, Germany). The images of transmission electron microscopy (TEM) were attained using a JEOL model JEM-1400 microscope (JEOL GmbH, Freising, Germany). Diffuse reflectance UV–Vis spectra (DRS) were recorded in the range of 300–800 nm using a Jasco model V-770 spectrometer (Jasco Corporation, Tokio, Japan). Reflectance measurements were renewed to absorption spectra using the Kubelka-Munk function. Fourier transform infrared (FTIR) spectra were recorded on a Nicolet iS5 Thermo Scientific FT/IR spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Photoluminescence (PL) spectra were acquired using a PerkinElmer model LS 55 fluorescence spectrophotometer (Perkin Elmer, Waltham, MA, USA). UV–Vis spectra were recorded at room temperature using a Jasco, model V-730 UV–Vis spectrometer (Jasco Corporation, Tokio, Japan).

2.4. Photocatalytic Degradation

The photocatalytic activity of the samples was assessed under simulated solar irradiation by studying the degradation of aqueous solutions containing methylene blue (MB) as a model pollutant. The characteristic absorption peak of MB was monitored using a UV–Vis spectrophotometer. The different mixtures of TiO₂–SA/(NH₄)₂V₃O₈ with ratios ranging from 1:0.025 to 1:0.15 were combined with 25 mL of a 1 × 10⁻⁵ mol L⁻¹ MB solution at a pH of 7 (buffer). Before solar irradiation measurements, the suspension was stirred magnetically in the dark for 30 min to establish an adsorption/desorption equilibrium between the dye and samples. The suspensions were exposed to a solar simulator (AM 1.5 G-1.00 kW/m² (Verasol, Newport, OR, USA). Approximately 0.5 mL of the reaction mixture was collected every 30 min, up to a total irradiation of 120 min. The samples were centrifuged to eliminate interference from suspended catalyst particles. The initial concentration of MB after adsorption/desorption equilibrium was designated as C₀, and the UV–Vis spectra were measured using nanopure water as a reference. To assess the reusability and chemical stability of the synthesized samples, we conducted five consecutive photocatalytic reactions. At the start of each cycle, a new load of MB was added and adjusted to a volume of 25 mL at a concentration of 1 × 10⁻⁵ M using the same catalyst. To verify the stability of the photocatalysts, we subjected the initial sample and the samples used in the reuse experiments to XRD analysis. This was done to confirm its photostability under UV–Vis light irradiation.

To detect the presence of hydroxyl radicals during photocatalytic reactions, we utilized terephthalic acid (TA) at a concentration of 4 mmol L⁻¹ in an alkaline medium containing NaOH 2 M (pH 8). TA undergoes reaction with ^•OH radicals, resulting in the formation of a highly fluorescent product, 2-hydroxyterephthalic acid (TAOH). At 10 min intervals following the solar irradiation, small aliquots of the reaction solution were collected and centrifuged. These samples were then analyzed using a spectrofluorometer until the irradiation duration reached 90 min. The fluorescence product was detected with a peak emission wavelength of 426 nm and an excitation wavelength of 312 nm [36].

The reuse stability of the photocatalyst was tested with four consecutive reuses. Under similar operating conditions, reuse experiments were performed with the same catalyst. After each experiment, a new methylene blue solution, calibrated to a volume of 25 mL, was added and the same absorption intensity of the 1 × 10⁻⁵ M concentration carried out during the experiments was determined by UV spectroscopy, at time intervals to evaluate the potential for reuse of the catalyst.

3. Results and Discussion

3.1. Synthesis and Characterization of the Precursors of the Photocatalytic System

In general, hydrothermal treatment of hydrogels formed via the hydrolysis of V₂O₅ or its salt precursor leads to a partial reduction of V(V) to V(IV) vanadium ions [37,38]. The extent of this reduction depends on the treatment’s conditions. In this specific case, the hydrolysis of NH₄VO₃ using hexadecylamine as a template resulted in the formation of a layered nanocomposite intercalated with the surfactant, resembling precursors of V₂O₅ hybrid structures such as nanotubes [39] and nanourchins [23]. During the alkaline hydrothermal treatment of the hydrolysis reaction product, both the partial reduction of V(V) to V(IV) and the segregation of the template occurred simultaneously, leading to the formation of ‘giant’ single-crystal thin plates with basal dimensions ranging from 2 to 10 mm (Figure S1). Single-crystal X-ray diffraction analysis of the as-prepared product confirmed its structural similarity to the (NH₄)₂V₃O₈ compound reported in the literature [40]. The crystallographic data for compounds are shown in Tables S1–S4. The (NH₄)₂V₃O₈ features a lamellar structure comprised of layers consisting of corner-sharing VO₄ tetrahedra and VO₅ square pyramids with the NH₄⁺ ions lying between the layers [25,32,41]. Figure 1a showcases a high-resolution transmission electron microscopy (HRTEM) image of the as-prepared (NH₄)₂V₃O₈, with discernible lattice planes 100 and 010 having an interplanar distance of approximately 8.9 Å [28]. The SAED patterns confirm the single-crystal nature, and the diffraction spots can be indexed to a tetragonal (NH₄)₂V₃O₈ structure (Figure 1b). The TiO₂–SA nanocomposite was synthesized by reacting titanium tetraisopropoxide (TTIP) with stearic acid in ethanol and the mixture was allowed to age for 72 h. Infrared spectroscopy was performed to confirm the presence of stearic acid in the nanocomposite. Figure S2 shows the characteristic signals attributed to the asymmetric and symmetric stretching vibrations of the CH₂ groups at 2920 and 2849 cm⁻¹, respectively, which correspond to stearic acid. Additionally, prominent signals are observed at 1538 and 1395 cm⁻¹, corresponding to the asymmetric and symmetric vibration of the C=O of the carboxylic acid, indicating its interaction with titanium atoms (COO–Ti) [42]. In the lower frequency range of the spectrum, a signal at 999 cm⁻¹ is observed, which can be attributed to the Ti–O–C bond. Figure S3 reveals the characteristic Raman scattering of linear saturated acids, appearing at around 1100 cm⁻¹, 1300 cm⁻¹, and in the range of 1500–1400 cm⁻¹. These signals are assigned to various vibrational modes, including ν(CC), CH₂ torsion, and CH₂ scissor vibrations and bending CH₃, respectively [43]. Notably, certain features are consistent with reported vibration modes reported for ultrathin TiO₂(B) nanosheets [44]. The TiO₂–stearic acid nanocomposite exhibits a layered nanostructure, with stearic acid sandwiched between inorganic sheets, as supported by previous studies [9,10,45].

3.2. Electron Paramagnetic Resonance Studies of the Electronic Structure of Vanadium in the Crystal Structure of (NH₄)₂V₃O₈

In the crystal structure of (NH₄)₂V₃O₈, it is stablished that the V=O bonds align parallel to [001] direction, while the cleavage planes are perpendicular to [001]. Consequently, the V=O bonds, pivotal in determining the EPR spectrum, are perpendicular to the principal cleavage plane. Given that the behavior of the unpaired electron in a VO²⁺ complex is predominantly dominated by the V=O bond, both its direction and the symmetry within the sample can be discerned through the analysis of the angular dependence of the EPR spectra.

Figure 2a illustrate the experimental setup employed to capture the angular dependence of the EPR spectrum for an arbitrary magnetic field orientation given by angles θ and φ. Remarkably, regardless of the field orientation, only one symmetric line is observed, which can be accurately fitted by a first derivative of a Lorentzian lineshape. Notably, when the magnetic field lies on the xy plane (θ = 90°) the EPR spectra exhibit no angular dependence on φ (results are not displayed); therefore, the angular variations presented herein are solely in relation to the angle θ. When the magnetic field is rotated within the xz plane, a pronounced dependence on θ is evident, as illustrated in Figure 2b. This figure presents three selected spectra at θ angles of 0°, 45° and 90° for both measuring temperatures of 300 K and 5.5 K. At 300 K, the data exhibit a discernable increase in line intensity from θ = 0° to θ = 90°, attributed to an increase of the linewidth, as will be elucidated subsequently. Conversely, the behavior of the line intensity at 5.5 K demonstrates a distinct pattern. Specifically, at 5.5 K, the line intensity decreases from θ = 0° to θ = 45°, followed by an increase towards θ = 90°. This distinctive behavior will be discussed later. The lack of resolved hyperfine interactions with ⁵¹V nuclei (I = 7/2, 99.8% natural abundance) on the EPR line suggests that the unpaired electron spins are influenced by exchange interactions. These interactions lead to a significant exchange narrowing effect, which effectively smears out the hyperfine structure.

The angular variation of the EPR spectra was conducted by measuring the spectra at interval ∆θ = 5°, ranging from θ = −87° to θ = +113°. Given that θ = 0°, the magnetic field is parallel to the V=O bond; this orientation is designated as the parallel direction. Consequently, θ = 90° corresponds to the perpendicular direction. Figure 3 depicts the results obtained for the angular variation of the EPR spectra of (NH₄)₂V₃O₈ at temperatures of 300 K and 5.5 K. The DPPH spectra are isotropic and are not shown in the figure. However, as observed in Figure 2b, the central field of the DPPH line undergoes slight shifts relative to the angle, particularly noticeable at 5.5 K, owing to small microwave frequency variations.

The least-square fitting of the spectra, based on the first derivative of a Lorentzian lineshape, enables the extraction of the line parameters such as the central position, peak-to-peak amplitude, and peak-to-peak width. Despite the slight asymmetry in the DPPH lineshape, a Lorentzian fitting proves sufficiently accurate for determining its central field. The g-values of the unpaired electron in the VO²⁺ complex are computed from the resonance condition.

$$h\nu =\beta g{B}_o\hspace{1em}\hspace{1em}\mathrm{and}\hspace{1em}\hspace{1em}g=g_m\frac{B_m}{B_o}$$

(1)

On the left side, β represents the Bohr magneton, ν denotes the microwave frequency and h stands for Plank’s constant. On the right side, B_m and g_m = 2.0037 are the central field and g-value of the DPPH marker, respectively. B_o and g represent the central field position and g-value of the sample spectrum, respectively. The microwave frequency is precisely locked onto the resonant frequency of the cavity. However, due to the rotation of the sample and fluctuations in room temperature, a minor change in the microwave frequency was anticipated throughout the experiment. Therefore, to account for this variability, the formula on the right was consistently employed.

Analysis of the results presented in Figure 3a,b reveals a cosine-like angular variation of the g-values as a function of θ. This variation in the g-values concerning the rotation angle can be analyzed through the expression derived from the Zeeman interaction:

$$g=\sqrt{g_{\perp }^2{\mathit{sin}}^2(\theta -{\theta }_0)+g_{\parallel }^2{\mathit{cos}}^2(\theta -{\theta }_0)}$$

(2)

where g_‖ and g_⊥ represent the parallel and the perpendicular components of the g-tensor, respectively, and θ₀ denotes an angular correction resulting from imprecision in setting the initial sample position (values are indicated in the figure). The fitting of the experimental data with this function is shown in solid red lines in Figure 3a,b, with the fitting parameters detailed in Table 1.

The principal values of the g-tensor are g_‖ =1.925 ± 0.001 and g_⊥ = 1.978 ± 0.001. These values align closely with those previously reported for (NH₄)₂V₃O₈, g_‖ = 1.9263 and g_⊥ = 1.9755 [34] and are also consistent with those documented for K₂V₃O₈, a vanadium oxide belonging to the same crystallographic group as (NH₄)₂V₃O₈: g_‖ = 1.922 and g_⊥ = 1.972 [26]. The fact that g_‖ < g_⊥ < g_e, where g_e = 2.0023 represents the free electron g-value, suggests that the VO²⁺ paramagnetic centers are located in axially distorted octahedral sites. The unpaired electron is located in the 3d_xy orbital of the transition metal ion. The relationships of g_‖, g_⊥ and g_e are observed across various vanadyl complexes and also in glasses doped with vanadium, as documented in the literature [46,47,48,49]. The tetragonality of the VO²⁺ sites are determined by the parameter (∆g_‖/∆g_⊥) where ∆g_‖ = (g_e − g_‖) and ∆g_⊥= (g_e − g_⊥). The ratio (∆g_‖/∆g_⊥) = 3.2 of (NH₄)₂V₃O₈ suggests that the vanadyl ions experience slight tetragonal distortion within the octahedral sites. It is interesting to note that this ratio aligns consistently with those reported for VO²⁺ and V⁴⁺ in several crystalline and amorphous matrices [46,49].

The linewidth of the sample spectrum is determined by measuring the field distance between the maximum and minimum points of the first-derivative spectrum, denoted by ∆B_pp. The peak-to-peak linewidth data are shown in Figure 3c,d, revealing noticeable distinctions at both temperatures. At 300 K, the curve assumes a bell-shaped form, whereas at 5.5 K, an M-shaped curve is observed for the angular dependence of the linewidth. A plausible mechanism accounting for the observed phenomena relates to the magnetic nature of the compound. Static susceptibility measurements at a high magnetic field demonstrate a paramagnetic behavior, accompanied by antiferromagnetic interactions among V⁴⁺ ions in adjacent planes oriented parallel to (001), resulting in a transition temperature of around 0.5 K. This fact is consistent with our findings of Lorentzian EPR lineshapes observed across all temperatures and angles.

The lineshapes offer some evidence of critical behavior at temperatures near the transition [50]. Within the critical region, it is anticipated that, owning to critical slowing-down, the modulation induced by exchange interactions become less efficient in averaging dipolar terms. This phenomenon should result in broader linewidth and a more Gaussian-like character for the EPR lineshapes [50]. From our analysis, we infer that the measurements presented here were conducted well above the critical region.

Given the magnetic properties of the material, it has been noted that the angular dependence of the EPR linewidth can be described by the following expression [50]:

$$\Delta B_{pp}=A+B{\mathit{sin}}^2(\theta -{\theta }_0)+C [3{\mathit{sin}}^2(\theta -{\theta }_0)-1{]}^2$$

(3)

where A, B and C are constants.

The solid lines depicted in Figure 3 represent the results of the fitting procedure applied to the experimental data using this formula, and the corresponding parameters are presented in Table 1. The sin²θ term corresponds to relaxation induced by the exchange interaction, while the (3sin²θ − 1)² term is attributed to relaxation driven by spin diffusion. Transitioning from room temperature to 5.5 K, the parameters B and C change signs, while the absolute value of C increases by more than one order of magnitude. This indicates that, with decreasing temperature, the relaxation facilitated by spin diffusion becomes increasingly significative relative to the relaxation influenced by exchange interactions.

Further evidence for the magnetic nature of the interactions between V⁴⁺ ions is the angular dependence of the integrated area of the EPR absorption spectra. The value of the area is calculated by means of the parameters derived from the Lorentzian fitting of the measured lineshapes of both the sample and marker, employing the well-known formula $S=(\pi /\sqrt{3})I_{pp}{\Delta B}_{pp}^2$, where $I_{pp}$ and $\Delta B_{pp}$ are the peak-to-peak amplitude and width of the first derivative signal, respectively. The results are shown in Figure 4. On the vertical scale, the term “relative intensity” means $100 (S-\widehat{S})/\widehat{S}$, where $\widehat{S}$ denotes the average of $S$ across all angles. In Figure 4, it is apparent that the DPPH signal amplitude does not remains constant, as expected, most probably due to a slight misalignment of the sample from the central axis of the microwave cavity. At 300 K the variations in the relative intensity are within the order of ±5% for both the sample and marker. However, at 5.5 K, while the fluctuation of the DPPH relative intensity remains at a similar amplitude, the variation in the sample signal noticeably increases in magnitude, revealing a pattern reminiscent of the previously observed M-shaped angular dependence in the linewidth.

The angular dependence behavior observed for the EPR absorption integrated area of the studied sample may have an explanation based on the dynamic of the magnetic interactions. It is well-known that a Lorentzian lineshape lacks physical meaning, as evidenced by the divergence of all its even moments. Therefore, a potential error in the calculation of the area or relative intensity, as described, is the lack of knowledge about the tails of the EPR lineshapes, which remain undetectable due to being well below the baseline noise. Consequently, it is important to emphasize that, under typical circumstances, the area of a Lorentzian lineshape cannot be measured experimentally. Indeed, according to basic EPR theory, the Lorentzian profile at the tails must exhibit a “cutoff” point, beyond which a transition to Gaussian behavior occurs. In this context, we propose that the threshold marking the shift in the dynamics of the paramagnetic system is angular-dependent; potentially providing a qualitative explanation for the observed line intensity behavior.

3.3. Characterization Heterojunction TiO₂–SA/(NH₄)₂V₃O₈

Figure 5a illustrates the X-ray diffraction (XRD) pattern of the prepared samples (NH₄)₂V₃O₈, TiO₂–SA, and TiO₂–SA/(NH₄)₂V₃O₈ (1:0.05)_. The pattern of the (NH₄)₂V₃O₈ exhibits characteristic peaks that closely match the reported values, with the two diffraction peaks at 15.9 and 31.8 in 2θ clearly indexed to the (001) and (002) planes (JCPDS 511733). In the diffractogram of TiO₂–SA, three consecutive and evenly spaced signals are observed at 3.2°, 6.2° and 9.2° in 2θ, which are attributed to the formation of structures with a preferred arrangement of stacked sheets [45]. Signals appearing between 20° and 22° in 2θ indicate the presence of a small amount of free stearic acid (JCPDS 03-0252). Finally, in the diffractogram corresponding to the TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050) sample, all the signals of both semiconductors are distinguishable. This suggests the coexistence of both structures in a single material, without any detectable changes or impurities apparent in the material.

Diffuse reflectance spectroscopy (DRS) was employed to characterize the samples. The bandgap of the samples was determined using the Tauc method, with energy (Eg) calculated from the plots of (αhʋ)^1/2 and (αhʋ)² for indirect and direct transitions, respectively, against photon energy. The bandgap values were obtained by analyzing the intercept of a tangent to the x-axis. As illustrated in the insert in Figure 5b [51]. (NH₄)₂V₃O₈, demonstrates a broad absorption within the visible light range, and indicates an indirect band gap width of 2.11 eV [41]. In contrast, TiO₂–SA displays a direct bandgap of 3.42 eV [45]. However, in the case of the TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050) composite, an indirect bandgap of 3.24 eV is observed, signifying a noticeable shift in the sample’s absorption capacity. This shift suggests a synergistic effect between the two semiconductors, potentially enhancing the composite’s light absorption capability. Figure 5c,d display the scanning electron microscopy (SEM) images for (NH₄)₂V₃O₈. These images reveal the formation of crystals with a square and uniform morphology, characterized by sizes exceeding 5 μm formed by a two-dimensional layered structure. In Figure 5e, the TiO₂–SA nanostructures are evident, showing the formation of arrangements consisting of thin sheets with rounded and stacked edges, with each having lengths of approximately 100–200 nm. This morphology is attributed to the layered nanocomposite structure [10]. EDS spectrum and elemental mapping images (Figure S4) demonstrate that TiO₂–SA/(NH₄)₂V₃O₈ comprises Ti, N, O and V, as expected. Notably, the incorporation of (NH₄)₂V₃O₈ into TiO₂–SA did not alter the morphology of the final composite TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050), where a clear and homogeneous mixture between both components is observed (Figure 5f).

3.4. Photocatalytic Properties of TiO₂–SA/(NH₄)₂V₃O₈ Composites

The adsorption capacity and photocatalytic activity of the prepared materials were assessed for removal by first adsorbing and then degrading methylene blue, a synthetic organic dye widely employed in the textile industry. The adsorption/desorption equilibration curves were obtained under dark conditions to determine the equilibration time. Figure S5 displays the results of dye adsorption using the prepared materials. It is evident that (NH₄)₂V₃O₈ exhibits a low capacity for adsorbing the contaminant. In contrast, when 0.025, 0.05, 0.10 and 0.15 moles of (NH₄)₂V₃O₈ are incorporated into TiO₂–SA, a significant decrease in dye concentration is observed within the first 30 min. This decrease remains constant over time, indicating the saturation of the active sites on the materials with adsorbed molecules. The percentage of dye removal after reaching the equilibrium is as follows: 1.74% for (NH₄)₂V₃O₈, 10.5% for TiO₂–SA, 16.4% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.025), 14.5% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050), 18.3% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.10), and 22.6 for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.15). The adsorption capacity of the prepared materials can be attributed to the surface area, primarily due to the lamellar structures of samples. Additionally, the presence of the organic component in TiO₂–SA enhances adsorption due to its affinity for the adsorbate [45]. In Figure 6a, the efficiency of MB degradation is demonstrated in the presence of TiO₂–SA/(NH₄)₂V₃O₈ samples with varying amounts of ammonium vanadium in molar ratios of 1:0.025, 1:0.050, 1:0.10, 1:0.15, as well as TiO₂–SA and (NH₄)₂V₃O₈ individually. The degradation efficiency was quantified by the ratio C/C₀, where C₀ represents the initial concentration of MB after equilibrium adsorption–desorption, and C is the concentration during the reaction.

It is notable that TiO₂–SA and (NH₄)₂V₃O₈ exhibit relatively lower degradation efficiencies, with 54% and 8% degradation, respectively, than the TiO₂–SA/(NH₄)₂V₃O₈ nanocomposites. The photocatalytic activity of the heterojunction is influenced by the concentration of (NH₄)₂V₃O₈. The TiO₂–SA/(NH₄)₂V₃O₈ percentage of dye degradation is as follows: 64.1% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.025), 100% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050), 83.1% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.10), and 68.8% for TiO₂–SA/(NH₄)₂V₃O₈ (1:0.15). The heterojunction (1:0.050) exhibits optimal efficiency in the photodegradation of MB. However, with higher concentrations of the ammonium vanadate, the photocatalytic activity decreases. This decline is likely due to excess vanadium, which increases the opacity of the sample, thus blocking light absorption and potentially affecting the absorption of visible light. The reactions follow a pseudo-first-order kinetics. The apparent rate constant (k) is determined by fitting on −Ln(C/C₀) versus reaction time, as shown in Figure 6b. The results indicate that TiO₂–SA/(NH₄)₂V₃O₈ (1:0.050) exhibits significantly enhanced photocatalytic degradation, being approximately three times faster than TiO₂–SA. This enhancement can be attributed to the two-dimensional morphology, characterized by the sheet-like arrangement of (NH₄)₂V₃O₈. This morphology promotes effective “face to face interfacial contact” [52] with the surfaces of the TiO₂–SA, facilitating absorption in the visible range and confirming the synergistic effect between both semiconductors in TiO₂–SA/(NH₄)₂V₃O₈.

3.5. Degradation Mechanism

In order to identify the species responsible for the degradation of the dye, we conducted photocatalytic tests employing the following three scavenger agents: 2-propanol (IPA), benzoquinone (BQ), and ammonium oxalate (AO). These agents selectively react with ^•OH, O₂^•− and h⁺, respectively, thereby modulating the photocatalytic efficiency of the system under investigation. The results are presented in Figure 6c. The addition of benzoquinone (BQ) resulted in a 38% decrease in photocatalytic activity. This decrease suggests the involvement of photogenerated superoxide (O₂^•−) in the photocatalytic process. Subsequently, when we repeated the experiments with the addition of 2-propanol (IPA) to the system, the concentration of the dye in the solution decreased by 42% compared to the control (no scavengers added). This finding indicates that hydroxyl radicals are also active participants in the photocatalytic process. Finally, upon incorporating ammonium oxalate, dye degradation of the dye decreased by 19%. This suggests that the primary species responsible for dye photodegradation are those generated in the transition of the electron from the valence band to the conduction band. Furthermore, the reactivity of ^•OH radicals in the degradation process was examined using photoluminescence (PL) with terephthalic acid (TA). The rates of ^•OH generation in an aqueous solution were determined under irradiation. Since TA can react with ^•OH radicals to produce fluorescent 2-hydroxyterephthalic acid (TAOH), the intensity of the PL peak indicates the quantity of ^•OH radicals produced. Figure 6d demonstrates the production of ^•OH radicals, indicating an enhancement in photocatalytic activity.

The electronic band alignment in TiO₂–SA/(NH₄)₂V₃O₈ is significant for understanding the charge transfer processes within the heterojunctions and the potential mechanisms behind the photocatalytic degradation of MB. The effectiveness of the photocatalyst depends on the effective separation of electron–hole pairs, and the migration of the photogenerated charge is closely linked to the band position in the conduction band (CB) and valence band (VB) of the semiconductors. The positions of the bands (CB and VB) for TiO₂–SA were calculated using the following equations from the literature [53]:

$$EVB=\chi -Ee-½Eg$$

(4)

$$ECB=EVB-Eg$$

(5)

Here, χ represents the absolute electronegativity of semiconductors, with values of approximately 5.81 for TiO₂ and 6.10 for (NH₄)₂V₃O₈ [53], respectively. Ee stands for the energy of free electrons on the hydrogen scale, which is approximately 4.5 eV. The bandgap energy (Eg) of the semiconductor is determined from data obtained through diffuse reflectance spectroscopy (DRS). The calculated CB and VB edge positions of the hybrid TiO₂ were determined to be −0.42 eV and 3.02 eV. In the case of ammonium vanadate, the CB and VB edge positions were determined to be 0.5 eV and 2.6 eV, based on the reference [54]. When semiconductors with varying energy levels come into contact, it typically leads to a redistribution of electric charges, resulting in a shift in the positions of the band edge. This phenomenon is known as the Fermi level alignment. This results in the equal distribution of electrons and holes locating on the TiO₂–SA and (NH₄)₂V₃O₈ species at the interface of TiO₂–SA and (NH₄)₂V₃O₈ [55]. Consequently, the CB and VB of (NH₄)₂V₃O₈ move upward [54,56,57]. The energy band diagram of the TiO₂–SA and (NH₄)₂V₃O₈ heterostructure photocatalyst next to the thermodynamic equilibrium is depicted in Figure 7. The CB potential (CB) of an n-type semiconductor has a more negative potential (approximately 0.1–0.2 V) than the flat-band potentials [58], resulting in a Fermi level of −0.32 eV in TiO₂. On the other hand, (NH₄)₂V₃O₈ is considered an intrinsic semiconductor, and its Fermi level is situated in the middle of the conduction and valence band, at approximately 1.06 eV [59]. This causes the CB and VB levels to shift from 0.5 eV to −0.96 eV and from 2.6 eV to 1.15 eV, respectively. Therefore, we proposed a mechanism under illumination in which the photogenerated electrons are excited from the VB to the CB in both (NH₄)₂V₃O₈ and TiO₂–SA, leaving holes in VB of (NH₄)₂V₃O₈ and TiO₂–SA. Consequently, the CB of (NH₄)₂V₃O₈ shifts upward (towards less negative energy), while the VB of TiO₂–SA shifts downward (more positive energy). This generates an electric field that leads to charge transfer and the recombination of electrons in the CB of TiO₂–SA with the holes located in the VB of (NH₄)₂V₃O₈ [20]. As a result, high redox carriers are produced with holes in the TiO₂–SA responsible for the oxidation. The photoinduced holes in VB of TiO₂–SA can oxidize adsorbed H₂O molecules to produce ^•OH (^•OH/H₂O is 2.34 eV vs. NHE). Photogenerated electrons in the CB of (NH₄)₂V₃O₈ produce O₂^−• from dissolved O₂ since the CB position is more positive than the potential of O₂/O₂^−• (−0.33 V vs. NHE), and the reduction potential of O₂/H₂O₂ is 0.695 eV vs. NHE. This implies that electrons can react with O₂ and H⁺ to produce H₂O₂, which indirectly generates ^•OH [60]. These radicals can efficiently degrade methylene blue.

Furthermore, the XRD analyses demonstrated the stability of the sample after undergoing consecutive photocatalytic reactions, confirming its recyclability and the ability to reuse without a significant loss of degradation efficiency (Figure 8a). Figure 8b shows the X-ray diffraction (XRD) analysis supporting the structural stability of the heterostructure catalyst.

4. Conclusions

In conclusion, this manuscript presents a novel approach for synthesizing (NH₄)₂V₃O₈ and demonstrates its potential application as a sensitizer in a TiO₂-based supramolecular hybrid photocatalyst. The research fills a significant gap in the literature by providing a thorough investigation into the structural, magnetic, and photocatalytic properties of (NH₄)₂V₃O₈, which have not been reported previously. The synthesis method resulted in the production of (NH₄)₂V₃O₈, as confirmed by X-ray diffraction. The lamellar structure of (NH₄)₂V₃O₈ was observed through transmission electron microscopy (TEM), with high-resolution TEM images revealing the lattice fringes of the material.

From measurements of the angular variation of g-values, the EPR technique applied to a sample of (NH₄)₂V₃O₈ suggests that the VO²⁺ complexes are located in axially distorted octahedral sites. EPR lineshapes are Lorentzian for all orientations of the crystal, indicating that the measurements were taken in the paramagnetic state under the regime of exchange narrowing. Lineshape parameters, such as linewidth, peak amplitude and integrated area are compatible with the paramagnetic behavior with an antiferromagnetic interaction between V⁴⁺ ions in consecutive layers.

Combining (NH₄)₂V₃O₈ with a TiO₂–stearic acid nanocomposite (TiO₂–SA) resulted in a significant improvement in photocatalytic activity for the degradation of methylene blue. This enhancement can be attributed to the 2D morphology of (NH₄)₂V₃O₈ sheets, which promotes efficient light absorption in the visible range, thereby demonstrating a synergistic effect between the two semiconductors in TiO₂–SA/(NH₄)₂V₃O₈.

The degradation mechanism of methylene blue involves the generation of various reactive species, including hydroxyl radicals (^•OH) and superoxide ions (O₂^•−), which contribute to the degradation process. The band alignment between (NH₄)₂V₃O₈ and TiO₂–SA was discussed, highlighting the efficient charge separation and transfer processes occurring at their interface. Importantly, the study also confirmed the stability and recyclability of the TiO₂–SA/(NH₄)₂V₃O₈ photocatalyst through multiple photocatalytic cycles, making it a promising candidate for practical applications in the remediation of organic contaminants under solar irradiation.

In summary, the research presented in this manuscript not only advances our understanding of the structural and photocatalytic properties of (NH₄)₂V₃O₈ but also highlights its potential for use in environmentally benign photocatalytic processes, contributing to the field of advanced materials for water treatment and pollutant degradation.