Mechanochemical Defect Engineering of Nb₂O₅: Influence of LiBH₄ and NaBH₄ Reduction on Structure and Photocatalysis

1. Introduction

1. Transition metal oxides (TMOs), such as Nb₂O₅, are notable for their diverse oxidation states and advantageous electronic and optical properties, making them suitable for applications in energy storage and photocatalysis. Hereby, Nb₂O₅ is a stable wide-band gap n-type semiconductor, whose properties strongly depend on the type of polymorph. Among these, monoclinic H-Nb₂O₅ is the thermodynamically stable variant and consists of a ReO₃ block-type structure of 3 × 5 and 3 × 4 blocks of corner sharing NbO₆ octahedra within the blocks and edge sharing between blocks [1]. With a band gap ranging from 3.1 to 3.9 eV [2,3] and an electrical conductivity of 10^-13 - 10^-6 S cm^-1 [4], Nb₂O₅ has been primarily explored for sensor applications and various electronic devices [5,6], with some studies also investigating its potential in dye-sensitized solar cells [7] and photocatalytic processes [8].

Upon partial reduction of TMOs and creation of oxygen vacancies, their electronic properties can be enhanced compared to the pristine oxides, yielding materials with improved chemical and physical characteristics [9]. For instance, Chen et al. synthesized so-called black titania through hydrogenation of TiO₂, which introduced defects and partially reduced Ti⁴⁺ cations to Ti³⁺, resulting in increased photocatalytic activity relative to the pristine oxide [10]. This approach has spurred interest in defect engineering within TMOs, leading to reports on partially reduced oxides such as Nb₂O_5-x [11,12], ZrO_2–x [13], WO_3–x [14], V₂O_5–x [15], and MoO_3–x [16], which also exhibit enhanced light absorption [13], as well as improved photocatalytic [15] and photoelectrochemical (PEC) activities [11,12].

Despite the potential of black Nb₂O₅, its synthesis has not been extensively studied. Common synthesis methods involve high-temperature reductions using aluminum [11,12] or NaBH₄ [17] as reducing agents or the creation of oxygen vacancies by the diffusion-reduction method [18], resulting in the formation of high-performance materials for photoelectrochemical water-splitting, sodium-ion batteries, and photocatalytic CO₂ reduction, respectively. However, these methods often require energy-intensive annealing steps under inert gas [19], hydrogen [20,21,22], or vacuum [23] conditions to induce partial reduction and formation of Nb₂O_5−x.

In contrast, mechanochemical approaches offer a more energy-efficient alternative by inducing chemical reactions through the absorption of mechanical energy at room temperature [24]. While energy consumption is not only heavily reduced with this method, the continuous impact of milling balls also results in the creation of additional defects and an increase in reactivity. Literature on mechanochemical reduction of TMOs is limited, using either highly reactive alkaline metals, such as Na [25] or Li [26] or non-conventional hydrides such as TiH₂ [27]. In a recent study, we were able to show that alkali metal hydrides LiH and NaH represent readily available and powerful reducing agents for the partial reduction in a ball mill of rutile type TiO₂ and monoclinic H-Nb₂O₅ resulting in the formation of black TiO₂ and black Nb₂O₅ with enhanced photocatalytic activities for the degradation of methylene blue [28].

Building on this knowledge, we now systematically investigate the effectiveness of LiBH₄ and NaBH₄, which has already been used in high temperature reductions, as reducing agents in mechanochemical reduction processes. More specifically, the effects of concentration and milling duration on the resulting materials' photocatalytic activity for methylene blue degradation are examined.

2. Materials and Methods

Materials

Nb₂O₅ (ChemPur, Karlsruhe, Germany, 99.98%), NaBH₄ (Merck, Hohenbrunn, Germany, ≥ 98%), and LiBH₄ (Sigma Aldrich, Steinheim, Germany, ≥ 95%) were used without further purification and stored in a glovebox under argon atmosphere. 1,2-dimethoxyethane (ABCR GmBH, Karlsruhe, Germany, 99%) was dried in a solvent purifying system SPS 5 (MBRAUN, Garching, Germany). All solids have been characterized by X-ray diffraction before use.

Synthesis

The syntheses were performed in a planetary ball mill Pulverisette 7 (Fritsch, Idar-Oberstein, Germany) using ZrO₂ grinding jars with a volume of 45 mL and 180 ZrO₂ milling balls with a diameter of 5 mm. All syntheses were carried out under an argon atmosphere using glovebox technique. In a typical experiment, 3.00 g of Nb₂O₅ (11.29 mmol, 1 eq.) were milled with n eq. of ABH₄ (n = 0.25, 0.5, 1 and 2; A= Li, Na). The milling speed was set to 300 rpm, while the milling time was equal to 10, 30, and 60 min. To prevent cementation during the milling process with 1 or 2 eq. of ABH₄, 200 µL of dry 1,2-dimethoxyethane (DME) were added.

Afterwards, the samples were washed with water and MeOH several times to remove unreacted alkali metal hydrides as well as side products. Both solvents were degassed with argon for 1 h beforehand. After centrifugation, samples were dried in a vacuum oven at 80 °C and stored in an argon-filled glovebox.

Testing of the photocatalytic activity

The photocatalytic experiments were performed in an EvoluChem^TM PhotoRedOx Box (HepatoChem, Beverly, USA) equipped with 2 x 20 ml sample holder, an EvoluChem 365PF lamp (365 nm) for testing the photocatalytic activity in the UV region and an Evoluchem 6200PF lamp (cold white) for the Vis region.

In a typical methylene blue (MB) degradation experiment, 15 mg of catalyst (1 mg/ml) were added to 15 mL of an aqueous MB solution (20 ppm). After stirring for 30 min in the dark, the suspensions were irradiated. After certain time intervals, 0.4 ml of suspension were periodically sampled and centrifuged to separate the photocatalyst from the solution. The MB solution was then diluted by a factor of 2 before the concentration of MB was measured by UV-Vis spectroscopy.

Characterization

Powder X-ray diffraction (PXRD) patterns were recorded on a D8-A25-Advance diffractometer (Bruker-AXS, Karlsruhe, Germany) under ambient conditions in BraggBrentano θ-θ-geometry (goniometer radius 280 mm) with Cu K_α-radiation (λ = 154.0596 pm). A 12 µm Ni foil served as a K_β filter at the primary beam side. At the primary beam side, a variable divergence slit was mounted and a LYNXEYE detector with 192 channels at the secondary beam side. Experiments were carried out in a 2θ range of 7 to 120° with a step size of 0.013° and a total scan time of 2 h. Rietveld refinement of the recorded diffraction patterns was performed using TOPAS 5.0 software [29]. Crystallographic structure and microstructure were refined, while instrumental line broadening was included in a fundamental parameters approach [30]. The mean crystallite size <L> was calculated as the mean volume weighted column height derived from the integral breadth. Crystal structure data were obtained from the Pearson’s Crystal database [31].

EPR spectra were recorded using an Elexsys E580 X-band spectrometer (Bruker, Ettlingen, Germany) with an ER 4118X-MD5 resonator (Bruker, Ettlingen, Germany). All shown EPR spectra were recorded at 80 K using a closed cycle cryostat (Cryogenic CF VTC).

For the acquisition of the Raman spectra, a Raman microscope LabRAM HR Evolution HORIBA Jobin Yvon A (Horiba, Longmujeau, France) with a 633 nm HeNe Laser (Melles Griot, IDEX Optics and Photonics, Albuquerque, USA) and an 1800 lines/mm grating was used.

UV-Vis diffuse reflectance spectra were performed on a Perkin Elmer Lambda 750 spectrometer (PerkinElmer Inc., Shelton, USA) equipped with a 100 mm integration sphere from 290 to 1500 nm with a 2 nm increment and an integration time of 0.2 s. BaSO₄ was used as the reference.

⁷Li and ¹¹B single-pulse excitation magic angle spinning (SPE MAS) NMR spectra were recorded at 155.57 MHz and 128.43 MHz, respectively, on a Bruker AV400WB spectrometer (Bruker, Billerica, USA) at 298 K in standard ZrO₂ rotors with a diameter of 4 mm. A spinning rate of 13 kHz and a relaxation delay of 3 s were applied. Solid LiCl and NaBH₄ were used as an external reference with a chemical shift of 0 and -42 ppm, respectively. Spectra were recorded using the TopSpin software [32]. Fitting of the spectra was performed using the DMFit software program package [33]. The extracted data is compiled in Table 1.

3. Results

3.1. Reduction Process During Ball Milling

For the mechanochemically induced partial reduction of Nb₂O₅, niobia was milled with n equivalents of LiBH₄ and NaBH₄ (n = 0.25, 0.5, 1 and 2 eq.) for 10, 30, and 60 min at a constant milling speed of 300 rpm. To prevent cementation and ensure a homogeneous reaction mixture, 200 µl DME were added during milling with 1 and 2 eq. of MBH₄ for 60 min.

After just 10 minutes of mechanochemical treatment at room temperature, a color change from white Nb₂O₅ to light gray and dark blue-black is observed depending on the milling conditions (Supporting Information, Figure S1). This change in color can be attributed to the reduction of Nb⁵⁺ to Nb^{4+ [34]} accompanied by the intercalation of alkali metal ions and/or the presence of color centers [35] due to the formation of oxygen vacancies [36] to maintain electroneutrality. The formation of unpaired electrons, which confirms a successful reduction, [12,37] was proven by EPR spectroscopy even for the lighter colored NaBH₄ reduced samples (Figure 1a). Hereby, LiBH₄ produced darker (anthracite to blackish blue) and more reduced samples than NaBH₄ (light gray to darker gray). To monitor the evolution of the pressure inside the milling jar, the EASY GTM system by Fritsch was used. As shown in Figure 1b, a continuous pressure increase up to 0.4 and 1.6 bar is observed with NaBH₄ and LiBH₄, respectively, which was also observed due to hydrogen formation during the mechanochemical reduction of TiO₂ and Nb₂O₅ with NaH and LiH [28]. In theory, the formation of volatile borane species as side-products is also conceivable, but not further investigated, as the focus lays on the characterization of the as-prepared reduced oxides. The higher pressure observed with LiBH₄ aligns with its greater efficiency in achieving reduction compared to NaBH₄. The milling process also elevated the temperature but remained below 35 °C for both hydrides.

Figure 1: (a) Normalized continuous wave (CW) EPR spectra of Nb₂O₅ reduced with 0.25 eq. NaBH₄ (black) and 0.25 eq. LiBH₄ (red) for 60 min. (b) Evolution of the pressure inside the milling jar for the reaction of Nb₂O₅ with 2 eq. NaBH₄ (black) and LiBH₄ (red).

3.2. PXRD Analysis and Influence of the Hydride Concentration on the Reduction

Powder X-ray diffraction (PXRD) analysis was conducted to assess structural modifications induced by partial reduction during milling. The diffraction patterns of all reduced samples remained consistent with monoclinic H-Nb₂O₅ (space group P2/m), indicating no major structural transformation, regardless of the reducing agent, its concentration, or the milling duration (Figure 2a; Supporting Information, Figures S2–S8). However, a distinct shift toward lower diffraction angles was observed for all LiBH₄ reduced samples (Figure 2b). Rietveld refinement revealed an expansion of the unit cell volume of by up to ~6·10^-3 nm³ in these samples compared to pristine Nb₂O₅, whereas no significant increase was detected for NaBH₄ reduced samples (Figure 2c). Generally, higher LiBH₄ concentrations resulted in increased cell volumes, while the milling time has a rather low influence. As previously reported for the mechanochemical reduction of transition metal oxides using LiH and NaH [28], the observed volume increase is attributed to Li⁺ intercalation, given its smaller ionic radius of (76 pm) compared to Na⁺ (102 pm) [38]. Based on the unit cell expansion, the estimated composition of the LiBH₄ reduced samples ranged from Li_0.06(1)Nb2O₅ to Li_0.13(1)Nb2O₅ (Figure 2c). In contrast, previously reported LiH-reduced samples exhibited a broader composition range (Li_0.02(1)Nb2O₅ to Li_0.25(1)Nb2O₅) [28], consistent with the higher reactivity of LiH compared to LiBH₄.

Figure 2: (a) PXRD patterns and (b) corresponding zoom of pristine and reduced niobia. Orange ticks indicate the Bragg positions of monoclinic Nb₂O₅ (P2/m). The vertical line is meant for easier visualization of the shift of reflections. (c) Evolution of the cell volume of pristine Nb₂O₅ compared to LiBH₄ and NaBH₄ reduced samples and (d) the estimated Li content based on Rietveld refinement assuming a sum formula of Li_xNb₂O₅ for the LiBH₄ reduced samples.

Although the microstructure parameters crystallite size and strain cannot be refined simultaneously (Supporting Information, Figure S9), an overall trend can be observed in the influence of the reaction parameters on the microstructure. With increasing milling times and decreasing hydride concentrations, the niobia crystallites exhibit larger defects, which are characterized by smaller crystallite size and/or larger strain. Longer milling times tend to damage the material due to the sustained mechanical stress. [39,40,41] On the other hand, the decrease in damage with decreasing NaBH₄ and LiBH₄ suggests a mechanical buffering effect of the respective hydride, meaning that the boron hydrides can absorb some of the mechanical energy from collisions with the milling balls without decomposing. The transfer of mechanical energy is thus hindered, and less energy is available to induce the mechanochemical reduction, which also explains nicely the lighter coloration of samples prepared with high NaBH₄ concentrations. In this case, the excess of hydride has the opposite effect of an inert grinding auxiliary such as inorganic salts, which can for example be added to sticky reaction mixtures to achieve a better texture and efficient mixing [42].

While clear differences in the sample coloration are observed in the NaBH₄ reduced samples (the darker the samples are, the higher their degree of reduction), color differences in the LiBH₄ reduced samples are more subtle. This raises the question of why NaBH₄ reduced samples show larger differences in color. In fact, the higher molar mass of NaBH₄ necessitates nearly doubles the mass compared to LiBH₄ to achieve the same Nb₂O₅-to-hydride ratio; for instance, 0.8539 g of NaBH₄ is required for a 1:2 ratio, while only 0.4920 g of LiBH₄ is needed. Consequently, it is very likely that LiBH₄ shows overall the same behavior, but it is less obvious due to the lower used mass. Theoretically, the use of an even greater excess of LiBH₄ should therefore lead to a lower degree of reduction.

To test this hypothesis, Nb₂O₅ was milled with an excess of 5 eq. LiBH₄ (similar weight ratio of the reactants than Nb₂O₅/2 eq. NaBH₄) for 10 min at 300 rpm, with 200 µl of DME added to prevent cementation. The resulting light blue oxide was significantly lighter than the sample obtained with only 2 equivalents of LiBH₄ under identical milling conditions. After washing and removal of unreacted LiBH₄, a cell volume of 1364.46 ų was determined by Rietveld refinement. The lighter coloration and reduced cell volume suggest that a lower degree of reduction was achieved with the excess LiBH₄. Consequently, the above hypothesis is confirmed and higher concentrations of the two alkali metal borohydrides tested lead to an overall lower reduction as the excess hydride acts as a buffer material.

3.3. Solid State NMR Spectroscopy

Solid state NMR spectroscopy was applied to study the effectiveness of the washing process, which is needed to remove unreacted hydrides and other possibly formed boron-containing side products such as alkali metal borates [43,44], which can also be envisioned in addition to the formation of volatile boron-containing side products. Moreover, the local environment of intercalated lithium can be investigated.

As shown in the ¹¹B and ⁷Li solid state MAS NMR spectra of the pristine LiBH₄ starting material, an intense central |+1/2⟩↔|–1/2⟩ transition was observed at -42.0 and -0.7 ppm respectively, as well as a spinning sideband manifold originating from the outer satellite transitions |±1/2⟩↔|±3/2⟩ (Figure 3a, Figure 4a), which is consistent with the literature [45,46]. With the help of the DMFit software package [33] for both spectra the quadrupolar parameter C_Q was extracted for the cases of η_Q = 0 and 1, which gives a hint on the local site asymmetry and coordination environment (Supporting Information, Table S1). The ¹¹B solid-state NMR spectra show the broad signal of the probe head, which contains a B-containing stator, so the signals can only be interpreted to a limited extent.

After ball milling Nb₂O₅ reduced with 0.25 eq. LiBH₄ for 60 min, the ¹¹B spectrum shows mainly the signal of the probe head of the NMR device and no signal of pristine LiBH₄ at -42.0 ppm is detected. In addition, one can clearly see the structured signal of small spinning sidebands centered around 0.2 ppm and a broader shoulder at around 10 ppm, indicating the presence of an unknown, possibly partially oxidized boron species in the sample but only in very small amounts (Figure 3b). This becomes clear when comparing the intensities to the spectrum of LiBH₄ shown in (Figure 3a) in which the signal of the probe head is almost completely suppressed. However, as shown in Figure 3c, the LiBH₄ starting material exhibits signals in the same region, which would match with tetrahedral BO₄ (around 0 ppm) and trigonal BO₃ (around 10 ppm) units [43,47]. Therefore, it cannot indefinitely be concluded that they form during the reduction process or are a simply present in the starting material. In any way, their overall concentration is very low which agrees with the low degree of reduction.

After washing, no spinning sidebands are visible anymore, meaning that all boron species were successfully removed by washing with water (Figure 3c). Regarding the ⁷Li spectrum of Nb₂O₅ reduced with 0.25 eq. LiBH₄ for 60 min, all features discussed above in line with Li incorporation into the Nb₂O₅ are visible in the spectrum (Figure 4b). The obtained C_Q parameter differs from the one of LiBH₄ (Supporting Information, Table S1) which is qualitatively already visible by the differences in the range and the intensity profile of the spinning sideband manifold underlines the different crystallographic environment of the Li atom in LiBH₄ and the reduced Nb₂O₅. Ball milling Nb₂O₅ with 2 eq. LiBH₄ leads to almost identical looking ⁷Li spectra (Supporting Information, Figures S10 and S11).

The examples in Figure 4 are representative of the study. Two other ⁷Li spectra are shown in the Supporting Information (Figures S10 and S11) looking almost identical to the one present here, meaning that the milling conditions appear to have no significant influence on the chemical environment of the lithium nucleus after reduction that can be detected with NMR spectroscopy.

Regarding the ¹¹B solid-state MAS NMR of pristine NaBH₄, a featureless intense central transition at -42 ppm is observed (Supporting Information, Figure S12a) in accordance with literature [46]. Again, an additional signal at ~3 ppm shows the presence of unknown oxidic boron-containing species already in the starting material [44,48]. In contrast to the LiBH₄ niobia reduced sample, large amounts of unreacted NaBH₄ are still detected in the ¹¹B MAS spectrum after ball milling Nb₂O₅ and 0.25 eq. NaBH₄ for 60 min, which is successfully removed during the washing process (Supporting Information, Figure S12b). The presence of unreacted NaBH₄ in the as-milled sample indicates a lower conversion compared to using LiBH₄ as a reducing agent.

Overall, no significant amounts of solid boron-containing side-products are detected that can clearly be identified as side-products issuing from the reaction. Suspending the reaction mixture in deuterated acetonitrile after milling also gives no clear evidence for the formation of higher boranes or other soluble boron-containing species. On the other hand, the degree of reduction is rather low, especially with NaBH₄ as reducing agent, meaning that only small amounts of side products are to be expected. The low degree of reduction, the observed buffering effect and the rather complicated question of the formation of boron-containing by-products are less favorable compared to the use of alkali metal hydrides as reducing agents [28].

Figure 3: (a) ¹¹B MAS NMR spectra (black) of LiBH₄ and fitted with a single Gaussian-Lorentz line (red). (b) Normalized ¹¹B MAS NMR of the probe head measured with an empty rotor (black) in comparison to the washed (yellow) and unwashed (blue) Nb₂O₅ reduced with 0.25 eq. LiBH₄ for 60 min. (c) Comparison of the magnified oxidic boron-containing side-phases present in pristine LiBH₄ (black) and in Nb₂O₅ reduced with 0.25 eq. LiBH₄ for 60 min (blue). For better comparison, both spectra are normalized to their respective intensity value at 0.12 ppm.

3.4. Raman and UV-Vis Absorbance Spectroscopy

Raman spectroscopy was additionally employed to study structural changes during milling. As shown in Figure 5a, all typical bands of monoclinic H-Nb₂O₅ remain detectable for both NaBH₄ and LiBH₄ reduced and washed samples. These include Nb-O-Nb angle-deformations between 160 to 300 cm^-1, transverse optic modes (TO) originating from symmetric stretching of NbO₆ octahedra between 600 and 700 cm^-1 and the longitudinal optic mode (LO) of NbO₆ edge-shared octahedra at around 990 cm^-1 [42]. After mechanochemical reduction, a significant decrease in Raman activity and broadening of the observed bands were noted (Figure 5a), particularly for LiBH₄ reduced samples, which also exhibited a red shift of the LO mode to up to 885 cm^-1. In line with PXRD and NMR analysis, it can therefore be concluded that the use of LiBH₄ as reducing agent leads to higher structural disorder caused by Li intercalation [49] or formation of other defects during milling [50] due to the higher degree of reduction.

Similarly, the optical properties were also affected by the mechanochemical reduction process. While the diffuse reflectance UV-Vis (DRS-UV-Vis) spectrum of pristine Nb₂O₅ shows only UV absorption with negligible absorbance in the visible and near-infrared (NIR) regions, all reduced samples demonstrated significantly enhanced absorption in these regions (Figure 5b). This enhancement is attributed to the successful partial reduction and introduction of defects [26], leading to a decrease in the optical band gap of up to 0.4 eV as determined using the Kubelka-Munk function (Supporting Information, Figure S13, Table S2).

3.5. Photocatalytic Degradation of Methylene Blue

The enhancement of absorption properties across a broader wavelength range and the reduction of the optical band gap often improve the (photogenerated) charge carrier properties, leading to increased photocatalytic activity [26], which we previously observed for our LiH reduced samples [28]. Consequently, the photocatalytic activity of the borohydride reduced samples was evaluated through the degradation of methylene blue (MB), a common pollutant in textile wastewater [51]. The photocatalytic degradation was tested under UV irradiation (365 nm) and visible light (400-700 nm, with the strongest irradiance being at around 450 and 550 nm). Control experiments without catalysts only showed minimal MB photobleaching (Figure 6), while stirring MB with the photocatalyst in the absence of light also did not significantly reduce its concentration (Figure 6b).

Under UV light, both pristine and reduced niobia show almost no activity as only 10 to 30% of MB are degraded after 60 min, depending on the synthesis conditions of the catalyst (Figure 6a). While there is a slight improvement in the photocatalytic activity of the oxides prepared with 2 eq. LiBH₄ compared to the pristine material, both pristine and black niobia are less effective as UV photocatalysts compared to our previously synthesized black titania samples. [28] In contrast, all tested black niobia samples show an enhanced performance under visible light compared to the pristine material. While pristine Nb₂O₅ degrades only 27% of MB in five hours, the degradation was improved to 37% when using the samples prepared with 0.25 eq. NaBH₄/60 min (blue) and 2 eq. LiBH₄/10 min (purple), as shown in Figure 6b. Further milling with 0.25 (green) and 2 eq. (yellow) LiBH₄ resulted in 49 and 53% degradation, respectively, which is comparable to our LiH reduced material [28].

These findings indicate that the partial mechanochemical reduction using alkali metal boron hydrides effectively enhances the photocatalytic properties of Nb₂O₅. Overall, the use of LiBH₄ as the more effective reducing agent, higher initial hydride concentrations and longer milling times favor the faster photocatalytic degradation of MB. This can be attributed to the partial reduction to Nb⁴⁺ and the increased presence of defects and structural disorder, resulting in an enhanced light absorption across a wide range of wavelengths, reduced the optical band gap, and likely decreased recombination rate of photogenerated electron-hole pairs [26,52].

Figure 6: Photocatalytic degradation of methylene blue (MB) (a) under UV (365 nm) and (b) under visible light without catalyst (black), with pristine niobia (red) and with reduced niobia (blue: 0.25 eq. NaBH₄, 60 min; green: 0.25 eq. LiBH₄, 60 min; purple: 2 eq. LiBH₄, 10 min; yellow: 2 eq. LiBH₄, 60 min). In (b), MB and Nb₂O₅ (turquoise), as well as LiBH₄ reduced Nb₂O₅ (brown) are also stirred without illumination for comparison to stirring with illumination. Note the different time scales between (a) and (b).

5. Conclusions

In summary, the mechanochemical partial reduction of H-Nb₂O₅ using LiBH₄ and NaBH₄ as reducing agents at room temperature was systematically investigated, revealing a strong dependence on the type of the reducing agent. Due to a buffering effect regarding the transferred mechanical energy, large boron hydride concentrations do not benefit the reduction process. Structural analysis by PXRD measurements confirmed the retention of the Nb₂O₅ lattice, with Li⁺ intercalation leading to an increase in cell volume, as also supported by ⁷Li NMR spectroscopy. All reduced samples demonstrated enhanced visible light absorption and a band gap reduction of up to 0.4 eV compared to pristine niobia. Raman spectroscopy revealed increasing structural disorder due to the formation of defects, particularly for the stronger reduced LiBH₄ samples. These defects contribute to significantly improved photocatalytic performance of the synthesized materials compared to pristine oxides under visible light illumination.