Tuning the nanoarchitectonics of NaNbO₃ for overall photocatalytic nitrogen fixation: insights into defect engineering and heteroatom doping

Abstract

Over the past few years, defect engineering has emerged as a powerful method to enhance the photocatalytic activities of semiconductor materials. This approach has been utilized to provide an enhanced number of active sites on the catalyst surface. Vacancy engineering and heteroatom doping are the most common methods used to generate defects in semiconductors. However, these methods have never been compared in terms of their effect on the photocatalytic activity of semiconductors. In this work, we have compared the nitrogen fixation activity of oxygen-vacancy-rich NaNbO₃ and V-doped NaNbO₃ photocatalysts. Both oxygen-vacancy-rich and V-doped NaNbO₃ photocatalysts performed better than their pristine counterpart (bare NaNbO₃). The overall highest activity was observed for V-doped NaNbO₃ (NVNO0.25) having an ammonia generation rate of 36 µmol g⁻¹ h⁻¹ as compared to the best oxygen-vacancy-rich NaNbO₃ sample (DNNO2), wherein the ammonia generation rate was 26 µmol g⁻¹ h⁻¹ under visible light irradiation. Furthermore, the light to chemical conversion efficiency and turnover frequency for NVNO0.25 were found to be 0.052% and 0.095 h⁻¹, respectively, under visible light irradiation. Besides, the nitrogen fixation activity of the photocatalysts was also explored under direct sunlight irradiation. In addition to ammonia, the formed nitrate and nitrite species were also quantified. Thus, this work provides a one-on-one comparison of the nitrogen fixation activity of two types of NaNbO₃ photocatalysts. The insights provided by this work can help researchers to design and develop efficient materials for photocatalytic nitrogen fixation.

---

1. Introduction

In recent years, a lot of pioneering developments in the field of photocatalytic nitrogen fixation has been made but still the technology is far behind from the industry requirements due to very low ammonia yields.^1–3 This can be due to the sluggish kinetics of the adsorption followed by the activation of nitrogen molecules onto the surface-active sites of the semiconductor catalysts, which results in low quantum yields and light to chemical conversions.^4–6 This creates a requirement for some modification in the catalytic system to achieve efficiency close to the atmospheric enzyme nitrogenase system.^7,8 Therefore, the current field of photocatalytic nitrogen fixation is focused on the use of defect engineering as a potential approach for controlling and modifying the local coordination environment of metal centers in semiconductor photocatalysts.⁹ It is well known that the current industrial ammonia synthesis process, i.e., the Haber–Bosch process, is not sustainable as it is very high energy intensive (400–500 °C, 150–250 atm), takes hydrogen supply from steam methane reforming or coal gasification, which releases 1.87 tons of CO₂ per ton of NH₃ produced, and increases the overall carbon footprint.^10,11 This signifies the importance of the continuous studies on defect engineered photocatalysts for nitrogen fixation under ambient conditions using water as a hydrogen source.

The redox potential of a photocatalyst is set by the band positions of the valence band (VB) and the conduction band (CB), which is quite challenging to adjust as compared to electrocatalysis where the electron energy may be tuned by adjusting the potential.^12,13 In the case of photocatalysis, the redox potentials of photoexcited CB electrons and VB holes therefore must be suitable for N₂ reduction and H₂O oxidation, respectively.⁴ Furthermore, to maximize the use of solar energy, the semiconductor band gap should be maintained as optimum as feasible. Therefore, while developing semiconductor photocatalysts for NH₃ production, it is imperative to meticulously evaluate and enhance the pertinent redox processes as well as the semiconductor's capacity to capture light. To tackle these challenges, defect engineering offers potential solutions by increasing the electron transfer, enhancing the adsorption and activation of N₂ on the catalyst surface, and tuning the band positions of the photocatalysts.^14,15 It includes various dimensions and types of defects like vacancy, doping, void, disorder, strain, twin boundary, grain boundary, edge dislocation, screw dislocation, etc.^9,16 Furthermore, the size, amount of dopant, synthesis technique, and post-synthesis treatment all affect the kinds of defects that can be found in semiconductor photocatalysts.^4,17 Among these types of defects, the 1-D defects including vacancy and doping are the most explored defects in various photocatalytic applications, such as nitrogen fixation,^18,19 CO₂ reduction,^20,21 wastewater treatment,^22,23 plastic waste upcycling,²⁴etc. as they provide enhanced electron transfer kinetics and substrate adsorption and activation, and tune the local coordination structure. It is to be noted that the anion vacancy and metal ion doping in metal oxide photocatalysts like TiO₂, ZnO, Bi₂MoO₆, BiWO₆, etc. create various active sites for the adsorption and activation of nitrogen molecules, resulting in boosted nitrogen fixation activity.^25–28

In recent research on photocatalytic nitrogen fixation, perovskite oxides have been widely used due to their excellent tunability and promising structural, optical, electronic and physicochemical properties.²⁹ The general formula for a simple perovskite oxide is ABO₃, where A is a large-sized metal cation in the +2 oxidation state, B is a small-sized metal cation in the +4 oxidation state, and the oxygen exists as O²⁻.³⁰ The A and B positions can be easily satisfied by various metals in the periodic table and the availability of different sites allows the high tunability of these materials with doping and vacancies to develop new photocatalysts.²⁹ The nanoarchitectonics term is basically used for the construction of various materials systems at the nanoscale level.^31,32 Tuning the structure of perovskite oxide nanoarchitectonics using vacancies and doping for the enhancement of photocatalytic activity has been explored in some previous works.⁵

In this regard, Kumar et al.³³ tailored oxygen vacancies (O_V) in SrTiO₃ using ascorbic acid as a reducing agent and it was found that one of the catalysts showed the highest ammonia formation owing to the optimal amount of O_V confirmed by electron paramagnetic resonance studies. Another perovskite oxide photocatalyst, sodium niobate (NaNbO₃), has proven its excellent photocatalytic activity in hydrogen evolution and wastewater treatment but its potential in photocatalytic nitrogen fixation is not explored much.^34,35 As per our literature survey, only one report by Zhang et al. studied the enhancement in nitrogen fixation (293.3 µmol g⁻¹ h⁻¹) of NaNbO₃ by O_V and the incorporation of Pt nanoparticles.³⁶

In this work, we have synthesized pristine NaNbO₃ (NNO) nanosheets using a facile hydrothermal method. To introduce defects in NNO nanosheets, oxygen-vacancy-rich NaNbO₃ (DNNO1, DNNO2 and DNNO3) and V-doped NaNbO₃ (NVNO0.25, NVNO0.50 and NVNO0.75) samples were synthesized. For oxygen-vacancy-rich samples, varying amounts of ascorbic acid (5, 10, and 15 mg) were mixed with pristine NNO and annealed under an argon atmosphere. Furthermore, V-doped NaNbO₃ samples were synthesized using a one-step facile hydrothermal method, wherein varying amounts of Nb₂O₅ and V₂O₅ were used. These synthesized samples were studied using several techniques to understand their structure and electronic and optical characteristics. With the incorporation of defects in NNO nanosheets, the specific surface area was found increase. Also, in the V-doped samples, O_V was introduced, which played a significant role in the photocatalytic activity. The photocatalytic activity of pristine NNO, oxygen-vacancy-rich samples (DNNO1, DNNO2 and DNNO3) and V-doped samples (NVNO0.25, NVNO0.50 and NVNO0.75) were compared for nitrogen fixation. Among the O_V-rich samples and V-doped samples, DNNO2 and NVNO0.25 showed higher photocatalytic nitrogen fixation. Overall, the NVNO0.25 sample showed the highest photocatalytic activity among all the defect-rich samples. The enhanced photocatalytic activity can be attributed to the synergistic effect of V-doping and O_V. Specifically, this work is expected to unfold different methods to synthesize defect-rich materials and provide a comparison of the activity of oxygen-vacancy-rich and V-doped NaNbO₃ photocatalysts.

2. Experimental details

2.1. Chemicals

Niobium oxide (Nb₂O₅) and vanadium oxide (V₂O₅) were purchased from Sigma-Aldrich, India. Ascorbic acid (C₆H₈O₆) was purchased from Merck, India. Ethylene glycol (C₂H₆O₂) was purchased from SRL chemicals, India. HgCl₂ was purchased from Qualikems, India. Sodium hydroxide (NaOH) was obtained from SRL Chemicals, India. Sodium nitrate (NaNO₃) was purchased from Loba Chemicals, India. Sulfanilamide and N-(1-naphthyl) ethylenediamine dihydrochloride were obtained from SRL Chemicals, India. Ethanol (C₂H₅OH) was purchased from Analytical CSS Reagent, India. DI water was obtained from an ELGA PURELAB Option-R7 double-stage water purifier. All the reagents were used as received without any further purification. ¹⁴N₂ gas ≥99% purity was procured from Chandigarh Gases Private Limited, India and was used for all experiments.

2.2. Synthesis of catalysts

2.2.1. Synthesis of pristine sodium niobate (NaNbO₃–NNO). Sodium niobate (NaNbO₃) was synthesized using sodium hydroxide (NaOH) and niobium oxide (Nb₂O₅) as precursors via a facile solvothermal process as earlier reported in the literature with a slight modification.³⁷ In brief, 600 mg of Nb₂O₅ was added to 8 mL of 5 M NaOH solution. Then, 56 mL of ethylene glycol was added to the mixture. This reaction mixture was sonicated for 60 min. Subsequently, the reaction mixture was transferred to a Teflon-lined autoclave of 100 mL volume capacity and heated at 150 °C for 12 h. Upon the completion of the reaction, the sample was allowed to cool down to room temperature. Finally, the sample was collected via centrifugation and washed thoroughly with ethanol and water to remove the unreacted compounds and impurities, then finally dried overnight at 80 °C. This bare NaNbO₃ sample was denoted as NNO.

2.2.2. Synthesis of oxygen-vacancy-rich sodium niobate samples (DNNO). After the successful synthesis of pristine NNO, 300 mg of the sample was taken in a mortar pestle and varying amounts of ascorbic acid were added (5, 15, and 25 mg) each time. The resultant mixture was mixed homogeneously and transferred to an alumina boat. The alumina boat was kept in a tube furnace for annealing at 400 °C for 2 h under an argon atmosphere. After the completion of the reaction, the sample was allowed to cool down to room temperature. The sample was then washed with ethanol and water to remove the impurities or unreacted compounds. The washed sample was then dried overnight at 80 °C. The samples were named DNNO1, DNNO2 and DNNO3, respectively.

2.2.3. Synthesis of vanadium-doped sodium niobate samples (NVNO). Vanadium-doped sodium niobate samples (V-doped NaNbO₃) were synthesized using NaOH, Nb₂O₅ and V₂O₅ as precursors via a solvothermal process. In brief, different amounts of Nb₂O₅ (450 mg, 300 mg, and 150 mg) and different molar ratios of V₂O₅ (0.25 mmol = 45.5 mg, 0.50 mmol = 91.0 mg and 0.75 mmol = 136.5 mg) were added to 8 mL of 5 M NaOH solution. Then, 56 mL of ethylene glycol was added and the reaction mixture was sonicated for 60 min. Subsequently, the reaction mixture was transferred to a Teflon-lined autoclave of 100 mL volume capacity and heated in an oven at 150 °C for 12 h. After the completion of the reaction, the sample was allowed to cool down to room temperature, followed by the collection of the samples via centrifugation. The obtained samples were washed repeatedly with ethanol and water to remove the unreacted compounds and then dried overnight at 80 °C. These samples were denoted as NVNO0.25, NVNO0.50 and NVNO0.75, respectively.

2.3. Photocatalytic nitrogen fixation

A specially designed photoreactor with three 45 W white LED bulbs (intensity of light = 48 900 lux) was used to perform photocatalytic N₂ fixing studies in a 100 mL two-necked round bottom flask under visible light irradiation. A standard reaction involved ultrasonically dispersing 25 mg of the photocatalyst in 50 mL of ultrapure deionized water. After that, high-purity N₂ gas was continuously stirred for 60 minutes while bubbling through the reaction mixture at a flow rate of 50 mL min⁻¹. Prior to bubbling, the N₂ gas was passed through 0.01 M KMnO₄ and 0.01 M H₂SO₄ traps to eliminate any N-containing contaminants. Afterwards, the reaction was carried out by irradiating visible light for 2 h. After the completion of the reaction, the reaction mixture was centrifuged at 12 000 rpm to remove the solid catalyst and the obtained reaction mixture was subjected to various quantification studies for determining the concentrations of NH₃, N₂H₄, NO₃⁻ and NO₂⁻ species as mentioned in detail in the SI.

3. Results and discussion

3.1. Synthesis and structural and optical studies

The preparation of pristine NaNbO₃ and V-doped NaNbO₃ as catalysts for photocatalytic nitrogen fixation was carried out by using hydrothermal treatment. On the other hand, the O_V-rich NaNbO₃ catalysts were synthesized by annealing the pristine NaNbO₃ catalyst using different amounts of ascorbic acid under an argon atmosphere. The synthesis of NaNbO₃,O_V-rich NaNbO₃ and V-doped NaNbO₃ is shown in Scheme 1(a-c). The powder X-ray diffraction (PXRD) measurements were performed for the NNO, NVNO0.25, NVNO0.50, NVNO0.75, DNNO1, DNNO2, and DNNO3 samples, as shown in Fig. 1a and c. The diffraction peaks of NNO at 2θ = 22.73°, 32.41°, 46.41°, 52.38°, 57.80°, 67.77°, 72.49°, and 77.08° correspond to the (100), (110), (200), (210), (211), (220), (221) and (310) lattice planes, respectively, and match well with the standard data (JCPDS card no. 74-2348), suggesting its successful synthesis in the monoclinic crystal system.³⁸ For the DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75 samples, similar diffraction peaks were observed. After the introduction of O_V into DNNO1, DNNO2 and DNNO3, the PXRD peaks were shifted to higher angles, as shown in Fig. 1d.³³ Also, in the V-doped samples (NVNO0.25, NVNO0.50 and NVNO0.75), the PXRD peaks were shifted to higher angles (Fig. 1b).³⁷ This shift in the peaks to higher angles signifies the lattice contraction due to the introduction of O_V and V-doping. The lattice parameters of all the samples were calculated as shown in Table S1 (refer the SI). Furthermore, the lattice properties like strain and crystallite size for all the samples (NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75) have also been determined based on literature reported procedures^39,40 and are shown in Table S1 (refer the SI).

Scheme 1 Schematic illustration depicting the synthetic procedure of the (a) pristine NNO, (b) oxygen-vacancy-rich NaNbO₃ and (c) V-doped NaNbO₃ samples.

Fig. 1 (a and c) PXRD patterns; (b and d) Bragg angle shifts in the PXRD patterns of NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50, and NVNO0.75, respectively.

The Raman spectroscopic measurements were performed for all the prepared samples using a 532 nm laser and the obtained data are shown in Fig. 2a and b. Pristine NNO shows strong peaks at 139, 181, 220, 250, 277, 433, 602 and 870 cm⁻¹.⁴¹ All the peaks in the Raman spectra of NNO from 100 to 1000 cm⁻¹ pertain to NbO₆ octahedron internal vibrational modes. The peak at 277 cm⁻¹ corresponds to the bending modes of NbO₆. The peak at 602 cm⁻¹ corresponds to the stretching modes of NbO₆ octahedra, whereas peak at 870 cm⁻¹ is a combinational peak. For the O_V samples (DNNO1, DNNO2, and DNNO3) and V-doped samples (NVNO0.25, NVNO0.50, and NVNO0.75), broadening of all the peaks was observed. This could be due to the introduction of defects in the samples (O_V, doping), which lead to a change in the crystallinity and distortion of the lattice, as shown in Table S1.⁴² The functional groups and surface chemical bonds of the as-synthesized pristine NNO, oxygen-vacancy-rich (DNNO1, DNNO2, and DNNO3) and V-doped (NVNO0.25, NVNO0.50 and NVNO0.75) samples were investigated by Fourier transform infrared (FTIR) spectroscopy (Fig. 2c and d). The absorption peaks of the NbO₆ octahedron are shown by the peaks in the 500–1000 cm⁻¹ region.⁴³ For materials that are rich in O_V (DNNO1, DNNO2, and DNNO3) and pristine NNO, the peak at 607 cm⁻¹ is attributed to O–Nb–O stretching vibration (Fig. 2c).⁴³ In the case of the V-doped NaNbO₃ samples (Fig. 2d), the peak observed at 564 cm⁻¹ corresponds to the O–Nb–O/O–V–O stretching vibrations. Another peak observed at 1061 cm⁻¹ could be due to carbonate ions (CO₃²⁻) adsorbed at the surface of the samples. The FTIR spectra complement the Raman spectra and provide more details on the vibrational modes observed in these samples.

Fig. 2 (a and b) Raman spectra; (c and d) FTIR spectra NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50, and NVNO0.75, respectively.

The optical properties of the pristine NNO, O_V-rich samples (DNNO1, DNNO2 and DNNO3) and the V-doped samples (NVNO0.25, NVNO0.50 and NVNO0.75) were investigated. A change in the color was observed after the incorporation of O_V and V-doping, as shown in Fig. S1a–g (refer the SI).⁴⁴ As the concentration of O_V was increased, the color of the samples (DNNO1, DNNO2 and DNNO3) changed from pale grey to dark grey. For the V-doped samples (NVNO0.25, NVNO0.50 and NVNO0.75), with increasing V-doping, the color of the samples changed from pale brown to dark brown. The effect of O_V and V-doping in the NaNbO₃ crystal structure was studied using UV-vis diffuse reflectance spectroscopy (DRS). The absorbance spectra of all the samples are shown in Fig. S1h and i. Furthermore, the band gaps of all the samples were calculated using the Kubelka–Munk function vs. The energy plots, as shown in Fig. S1j and k (refer the SI).⁴⁵ The calculated band gaps were found to be 3.17, 2.95, 2.94, 2.92, 2.81, 2.71, and 2.66 eV for NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75, respectively. A decrease in the band gap was observed after the incorporation of O_V and V-doping in the lattice structure of NaNbO₃.²⁰

Furthermore, reflection electron energy loss spectroscopy (REELS) studies were conducted to verify the band gaps of NNO, DNNO2 and NVNO0.25 (representative samples) and the obtained results are presented in Fig. 3a–c.⁴⁶ The band gaps of NNO, DNNO2 and NVNO0.25 were found to be 3.30 eV, 2.68 eV and 2.63 eV, respectively. The calculated band gaps were in accordance with the values calculated using the DRS spectra. In addition, the positions of the flat bands for the samples (NNO, DNNO2 and NVNO0.25) were determined using the Mott–Schottky plots at 100 Hz and 200 Hz (Fig. 3d–f).⁴⁷ The flat band potentials (E_FB) for NNO, DNNO2 and NVNO0.25 were found to be −0.77, −0.67, and −0.74 V, respectively, relative to Ag/AgCl (pH = 6.8). The flat band potentials of all these samples are relative to the reversible hydrogen electrode (RHE) and were calculated using the following equation:

Fig. 3 REELS spectra of (a) NNO, (b) DNNO2, and (c) NVNO0.25; Mott–Schottky plots of (d) NNO, (e) DNNO2, and (f) NVNO0.25, respectively.

Here, the pH of the electrolyte solution is 6.8 and is 0.197 eV. So, the values of E_FB relative to the RHE were determined to be −0.17, −0.07, and −0.14 V for NNO, DNNO2 and NVNO0.25, respectively. Also, the positive slope obtained for all the samples indicates that the samples behave as n-type semiconductors. In the case of n-type semiconductors, the E_FB and E_CB values are close to each other (negative potential −0.3 V is added to E_FB).⁴⁸ So, the resulting values of E_CB for all the samples come out to be −0.47, −0.37, and −0.44 V for NNO, DNNO2 and NVNO0.25, respectively. Based on the CB values obtained from the Mott–Schottky plots and band gap values from DRS, the VB values of NNO, DNNO2 and NVNO0.25 were found to be 2.70, 2.57, and 2.37 V, respectively. The VB-CB diagram of all the prepared samples is shown in Fig. S2 (refer the SI).

3.2. Morphological, thermal, compositional, and surface area studies

The morphological investigations on the pristine NNO, oxygen-vacancy-rich (DNNO2) and V-doped (NVNO0.25) samples were performed using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The morphology of pristine NNO was found to be agglomerated nanosheets, as shown in Fig. 4a and b. No apparent change in the morphology of DNNO2 and NVNO0.25 could be seen after incorporating O_V and V-doping (Fig. 4c–f). The TEM image of NNO (Fig. 5a) also confirms the agglomerated nanosheet morphology. Furthermore, the high-resolution (HR) TEM image of NNO shows the structural features of the nanosheets (Fig. 5b), indicating its crystalline nature. The lattice spacing was found to be 0.278 nm, corresponding to the (011) lattice plane. In addition, the inverse fast Fourier transform (IFFT) of the (011) lattice plane is shown in Fig. 5c. In addition, the TEM images of DNNO2 are shown in Fig. 5d, indicating the nanosheet morphology. The HRTEM image of DNNO2 (Fig. 5e) shows the lattice plane (100) of 0.392 nm. The IFFT of the lattice plane (100) is shown in Fig. 5f. Furthermore, the TEM image of NVNO0.25 also indicates the agglomerated nanosheet morphology (Fig. 5g). The HRTEM image of NVNO0.25 shows a lattice plane with 0.267 nm spacing, corresponding to the (011) plane (Fig. 5h). The IFFT of the (011) lattice plane is shown in Fig. 5i.

Fig. 4 SEM images of (a and b) NNO, (c and d) DNNO2, and (e and f) NVNO0.25.

Fig. 5 (a–c) TEM image, HRTEM image and IFFT of NNO; (d–f) TEM image, HRTEM image and IFFT of DNNO2; and (g–i) TEM image, HRTEM image and IFFT of NVNO0.25, respectively.

To study the oxidation state and changes in the electronic environment after the incorporation of O_V and V-doping in the NaNbO₃ crystal lattice, XPS measurements were carried out. The survey spectra show the presence of Na, Nb, O and V in the NNO, DNNO2 and NVNO0.25 samples, as shown in Fig. 6a. The high-resolution Na 1s XPS spectra of NNO, DNNO2 and NVNO0.25 are shown in Fig. 6b. The peak for Na 1s was observed at 1071.45, 1071.50 eV and 1071.39 eV for NNO and DNNO2 and NVNO0.25, respectively.⁴⁹ The binding energy of Na indicates the +1 oxidation state and a slight increase in the binding energy for DNNO2 could be due to presence of O_V, which are electronegative in nature. While for NVNO0.25, the Na 1s peak was observed at a slightly lower value than pristine NNO, which could be due to the transfer of electrons among the host atoms and doped atoms.⁵⁰ The high resolution Nb 3d XPS spectra of NNO show two peaks at 207.05 and 209.75 eV, corresponding to 3d_5/2 and 3d_3/2, respectively (Fig. 6c). The binding energies of these peaks indicate the +5 oxidation state of Nb.⁵¹ For DNNO2, the peaks corresponding to 3d_5/2 and 3d_3/2 were observed at 207.15 and 209.80 eV, which indicate the +5 oxidation state of Nb.⁵¹ In addition, the two other peaks at 206.31 (3d_5/2) and 209.05 eV (3d_3/2) indicate the +4 oxidation state of Nb. The presence of the +4 oxidation state in the Nb 3d XPS spectra further confirms the presence of O_V in the DNNO2 sample. In the Nb 3d XPS spectra of NVNO0.25, the peaks with binding energies 206.91 and 209.58 eV were deconvoluted, corresponding to 3d_5/2 and 3d_3/2, respectively. These peaks indicate the +5 oxidation state of Nb.⁵² In addition, the peaks at 206.35 and 208.10 eV also reveal the presence of Nb⁴⁺ species, which in turn indicates the presence of O_V. The increased intensity of Nb⁴⁺ peaks in NVNO0.25 as compared to DNNO2 suggests the higher concentration of O_V in NVNO0.25 than the DNNO2 sample.

Fig. 6 (a) XPS survey spectra; (b–e) high-resolution XPS spectra of Na 1s, Nb 3d, O 1s and V 2p, respectively; (f) TGA curves of NNO, DNNO2 and NVNO0.25.

The high-resolution O 1s XPS spectra of pristine NNO were deconvoluted in two peaks, i.e., 529.67 and 531.10 eV (Fig. 6d). These peaks can be ascribed to the lattice oxygen and adsorbed –OH species. For DNNO2, the O 1s XPS spectra were deconvoluted to two peaks, which are located at 530.01 and 530.77 eV. The peaks at 530.01 and 530.77 eV correspond to lattice oxygen (O_L) and O_V, respectively.³⁶ The intensity ratio of these peaks calculated based on the area under these peaks was found to be 0.49. The same two peaks for NVNO0.25 were observed at 529.78 (O_L) and 530.50 eV (O_V). The intensity ratio of these peaks calculated based on the area under these peaks was found to be 0.60. The O_V/O_L ratio for DNNO2 (0.49) and NVNO0.25 (0.60) suggests that the concentration of O_V is higher in NVNO0.25 than in DNNO2, which agrees with the results discussed earlier. Furthermore, the high-resolution V 2p_3/2 XPS spectrum was recorded to confirm V-doping in the NaNbO₃ crystal structure. The spectrum was deconvoluted into two peaks, i.e., 515.10 and 516.60 eV. These peaks correspond to the +4 and +5 oxidation states of V in NVNO0.25.⁵² The presence of V⁴⁺ indicates the presence of O_V in the NVNO0.25 crystal structure (Fig. 6e). The XPS results obtained from the high-resolution data for every element in the sample corroborate well with each other.

The thermal stability of pristine NNO, DNNO2 and NVNO0.25 was examined using thermogravimetric (TGA) studies up to 800 °C and the obtained data are shown in Fig. 6f. For NNO, a mass loss of 13% up to 800 °C was observed, which could be ascribed to loss of adsorbed hydrocarbons.⁴³ In DNNO2, a mass loss of only 4% was seen up to 800 °C, which could again be attributed to the loss of adsorbed hydrocarbons. Also, a mass loss of only 6% was observed for NVNO0.25.

Furthermore, the chemisorption of N₂ over the surface of NNO, DNNO2 and NVNO0.25 was analyzed by temperature programmed desorption (TPD) and the obtained data are presented in Fig. S3a–c. The adsorption of N₂ was measured over the surfaces of NNO, DNNO2 and NVNO0.25 using an N₂ probe at 30 °C. The peaks below 300 °C indicate the physisorption of N₂ over the catalyst surface. Subsequently, the desorption of N₂ was checked up to 800 °C. For NNO, the two peaks centered at 129 °C and 366 °C which represent physisorption and chemisorption, respectively. For DNNO2, peaks at 126 °C, 454 °C and 718 °C were observed and for NVNO0.25, three peaks at 121 °C, 351 °C and 480 °C were observed. The amount of N₂ adsorbed on the surfaces of NNO, DNNO2 and NVNO0.25 was found to be 0.343, 0.337 and 0.379 mmol g⁻¹, respectively. This result is crucial for the investigation of N₂ fixation over these catalysts.^53,54 Furthermore, electron paramagnetic resonance (EPR) studies were performed on the NNO, DNNO2 and NVNO0.25 samples. The maximum intensity of the EPR signal was found in NVNO0.25, indicating the highest intensity of O_V in the sample, followed by the DNNO2 and NNO samples (Fig. S4).

Brunauer–Emmett–Teller (BET) analysis was performed to measure the specific surface areas of the NNO, DNNO2 and NVNO0.25 samples. The N₂ adsorption–desorption isotherms of all the samples are shown in Fig. S5a–c. The pore size distribution curves of the samples are also shown in Fig. S5d–f (refer the SI). The specific surface areas of NNO, DNNO2 and NVNO0.25 were found to be 3, 15 and 13 m² g⁻¹, respectively. In addition, the pore size measured was between 3 and 50 nm, indicating the mesoporosity.³⁶

3.3. Photocatalytic nitrogen fixation studies

The photocatalytic nitrogen fixation studies were performed in ultrapure water, which were purged with N₂ (99%) gas under a visible light source for 2 h at a maintained temperature of 25 °C. The use of sacrificial agents was avoided in this study. In addition, to prevent any interference from the feeding gas, a trap system was built to ensure that the impurities, such as NO₃⁻, NO₂⁻ and NH₄⁺, present in the N₂ gas do not interfere with the measurements. Specifically, 0.1 M of KMnO₄ was used to trap NO₃⁻/NO₂⁻ impurities and 0.1 M of H₂SO₄ was used to trap NH₄⁺ impurities. The Nessler's reagent (NR) method was used to determine the amount of ammonia produced during this study.^55,56 The calibration curve of NH₃ concentration vs. absorbance for the NR technique was used to calculate the concentration of NH₃ produced photocatalytically (Fig. S6a and b).

In Fig. 7a, the absorption spectra of all the reaction mixtures containing photocatalysts, NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75 were compared for 120 min using the Nessler's reagent method. The amount of NH₃ was found to be 3, 23, 26, 18, 36, 22 and 15 μmol for NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75, respectively, as shown in Fig. 7c. The low yield of NH₃ for pristine NNO in visible light could be ascribed to its large band gap (3.17 eV), which is in the UV region.³⁶ Among the oxygen-vacancy-rich samples (DNNO1, DNNO2 and DNNO3), DNNO2 was found to show the best activity. In addition, among the V-doped NaNbO₃ samples (NVNO0.25, NVNO0.50 and NVNO0.75), NVNO0.25 was found to be the best. Among all the samples, the concentration of photocatalytically generated NH₃ was the highest in NVNO0.25 (36 μmol). Based on the above results, the yield rates were obtained for all the catalysts. The calculations related to the yield (μmol g⁻¹ h⁻¹) can be seen in the SI (S5). The histogram comparing the yields of all the photocatalysts is shown in Fig. 7d. The yield rate was found to be 3, 23, 26, 18, 36, 22 and 15 μmol g⁻¹ h⁻¹ for NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75, respectively.

Fig. 7 (a) UV-vis absorption spectra of the NH₃ samples obtained using Nessler's reagent; (b) UV-vis absorbance spectra of the N₂H₄ samples obtained using the Watt–Chrisp method; (c) histogram of the NH₃ concentration produced in μmol; (d) histogram of the NH₃ concentration produced in μmol g⁻¹ h⁻¹; and (e) histogram of the N₂H₄ concentration produced and (f) histogram of the NO₃⁻ concentration produced in NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75, respectively.

During the photocatalytic reduction of N₂, hydrazine (N₂H₄) as a side product can also be formed along with NH₃. The formation of N₂H₄ can significantly decrease the production of NH₃. Therefore, it is important to quantify the concentration of N₂H₄ in the reaction mixture. The concentration of N₂H₄ formed in the reaction mixture was determined using the Watt–Chrisp method.⁵⁷ The calibration curve was made using the N₂H₄ concentration vs. The absorbance (Fig. S6c and d). From the UV-vis data (Fig. 7b), it can be concluded that N₂H₄ is absent in reaction mixture of all the different photocatalysts (Fig. 7e). Furthermore, the absence of N₂H₄ in the reaction mixture suggests that the associative distal mechanism was followed. This mechanism involves the hydrogenation of distal nitrogen first and leads to the formation of NH₃. Subsequently, after the release of one NH₃ molecule, the second nitrogen atom gets hydrogenated and releases another NH₃ molecule.

In a photocatalytic system, both electrons (e⁻) and holes (h⁺) are present. As there are no sacrificial agents used in this study, it is necessary to identify and quantify the oxidized species, such as nitrates (NO₃⁻) and nitrites (NO₂⁻).⁵⁸ In this regard, NO₃⁻ species were identified and quantified using ion chromatography (Fig. S7 and S8a–c). The concentration of NO₃⁻ ions in the reaction mixture for all the catalysts (NNO, DNNO1, DNNO2, DNNO3, NVNO0.25, NVNO0.50 and NVNO0.75) was were found to be 1, 0, 0, 1, 2, 3, and 6 μmol g⁻¹ h⁻¹, respectively (Fig. 7f). The calibration curve of the concentration vs. the area (μS cm⁻¹) is shown in Fig. S8d. In addition, the calibration curve of NO₂⁻ ions was plotted against concentration vs. absorbance (Fig. S9a and b). The associated absorption spectra of all the catalysts are shown in Fig. S9c. However, no presence of NO₂⁻ ions was found under any conditions (Fig. S9d).

Furthermore, nuclear magnetic resonance (NMR) was also employed to determine the presence the NH₃ in the reaction mixture. The ¹H-NMR spectrum of the reaction mixture is given in Fig. S10, where three distinct peaks in the region from 6.76 ppm to 7.02 ppm can be seen with a J value of 52 Hz. These peaks signify the presence of NH₃ from the purged nitrogen gas in the experiment.⁵⁹ In addition, the control studies were performed using the best activity catalyst, NVNO0.25. The reaction was first performed without the catalyst, then under an Ar gas atmosphere and finally under dark conditions. The associated spectroscopic data of these experiments are shown in Fig. S11. There was no formation of NH₃ detected in these experiments, indicating the importance of the photocatalyst, N₂ gas and light. These results also indicate that there was no involvement of atmospheric impurities and the impurities present in feeding N₂ gas.

Subsequently, the effect of sunlight on the photocatalytic yield of NH₃ was studied using pristine NNO, DNNO2 and NVNO0.25. It was found that the NH₃ concentration was increased in the presence of sunlight, and the obtained values were 12, 36, and 44 μmol g⁻¹ h⁻¹, respectively. The associated UV-vis absorption spectra are shown in Fig. 8a and the obtained NH₃ yield was compared with the visible light, as shown in Fig. 8b. The increase in the NH₃ yield can be attributed to the higher intensity of the sunlight (63 000 Lux), as compared to the visible light used (48 900 Lux). The light to chemical conversion (LCC) efficiency was also calculated for the NH₃ yield in the presence of visible light and sunlight for NNO, DNNO2 and NVNO0.25, respectively. The formula used and the detailed calculations are shown in section S6 in the SI.⁵³ For visible light, LCCs was calculated to be 0.005%, 0.033% and 0.052%, respectively. Similarly for sunlight, it was 0.013%, 0.039% and 0.048% for NNO, DNNO2 and NVNO0.25, respectively. The comparison of calculated LCC in visible light and sunlight is shown in Fig. 8c. Besides, the turnover frequency (TOF) was calculated for NNO, DNNO2 and NVNO0.25 in the presence of visible light and sunlight. The formula and the detailed calculations are provided in section S7 in the SI.⁶⁰ So, the TOF in visible light was calculated to be 0.009, 0.077, and 0.095 h⁻¹ and in sunlight, and the TOF was calculated to be 0.035, 0.110, and 0.120 h⁻¹ for NNO, DNNO2 and NVNO0.25, respectively. The comparison of calculated TOF in visible light and sunlight is shown in Fig. 8d.

Fig. 8 (a) UV-vis absorption spectra under sunlight irradiation; (b) histogram comparing the yield of NH₃ in visible and sunlight; (c) LCC efficiency measured in visible light and sunlight; (d) TOF calculated in visible and sunlight; (e) UV-vis absorption spectra of NH₃ obtained in each hour up to 5 h using the NVNO0.25 photocatalyst; and (f) NH₃ concentration obtained in each hour up to 5 h using the NVNO0.25 photocatalyst.

In addition, a long duration experiment (5 h) was conducted with NVNO0.25 to study the changes in product yields over time. The associated absorbance spectra are shown in Fig. 8e. The concentration of NH₃ in each hour (1 h, 2 h, 3 h, 4 h and 5 h) was found to be 11, 33, 41, 22, and 19 μmol, respectively, as shown in Fig. 8f. A decrease in the NH₃ yield was observed after 3 h, which could possibly be attributed to the back conversion of the formed NH₃ into its oxidized products, such as NO₃⁻ or NO₂⁻ species.⁵⁸ It is also to be mentioned that pure N₂ gas purged only once in the beginning of the experiment and no continuous gas supply was provided.

Based on all the results obtained, the associative distal pathway, as shown in Scheme 2a, has been proposed to understand the mechanism of NH₃ formation using NVNO0.25 as a photocatalyst. The first step in the activation of N₂ molecules is its chemisorption on the surface of the catalyst. The O_V present on the surface helps in the chemisorption of N₂ molecules and electrons from surrounding metal atoms can further reduce it. Furthermore, H₂O used in the reaction will provide H⁺ ions, which will react with activated N₂ molecules and undergo intermediate states, resulting in the formation of NH₃ molecules.¹ As discussed earlier, the absence of hydrazine (N₂H₄) from the reaction mixture indicates the formation of NH₃via an associative distal pathway. Besides, the photocatalytic mechanism of NH₃ formation is given in Scheme 2b. In pristine NNO, the valence band (VB) and the conduction band (CB) were found to be at 2.70 and −0.47 V, respectively, with a band gap of 3.17 eV. However, with the incorporation of O_V in the DNNO2 catalyst, band gap narrowing was observed with VB shifting upward to 2.57 V and the CB has moved downward to −0.37 V. In addition, with V doping in NaNbO₃ lattice (NVNO0.25) has resulted in the upward shifting of VB to 2.37 V and the CB has shifted downward to −0.44 V. Also, evident from the XPS spectra, there could be defect states present below the CB of NVNO0.25. These defect states will further help in the reduction of N₂ molecules. The position of CB for NNO, DNNO2 and NVNO0.25 is more negative than the water reduction potential (H⁺/H₂; 0.0 V vs. RHE) and the VB position is more positive than the water oxidation potential (O₂/H₂O; 1.23 V vs. RHE). These values of the CB and the VB suggest that all three photocatalysts (NNO, DNNO2 and NVNO0.25) are suitable for photocatalytic overall water splitting. Furthermore, the standard reduction potential of N₂ to NH₃ is +0.0577 V vs. RHE. The more negative CB potentials of NNO, DNNO2 and NVNO0.25 (−0.47, −0.37, and −0.44 V) make them suitable for N₂ reduction into NH₃. Also, the standard oxidation potential of N₂ to NO₃⁻ is 1.32 V vs. RHE. The more positive VB potentials of NNO, DNNO2 and NVNO0.25 (2.70, 2.57, and 2.37 V) make them suitable for oxidation of N₂ to NO₃⁻. Hence, all three photocatalysts (NNO, DNNO2 and NVNO0.25) can both reduce and oxidize N₂ molecules. So, the simultaneous production of NH₄⁺ and NO₃⁻ was observed in the reaction mixtures. The order of photocatalytic activity is observed as: NVNO0.25 > DNNO2 > NNO. The better photocatalytic activity could be explained by the smaller band gap of NVNO0.25, which results in the better absorption of light. Hence, the V-doped NaNbO₃ (NVNO0.25) sample shows better activity than the oxygen-vacancy-rich (DNNO2) sample and pristine NNO. The comparison of the photocatalytic activity of our catalysts with that of similar materials reported in literature is provided in Table S2.

Scheme 2 (a) A pictorial representation of the proposed associative distal pathway for the formation of NH₃ over the NVNO0.25 surface and (b) schematic illustration of the proposed mechanism of NH₃ and NO₃⁻/NO₂⁻ formation.

In addition, the recyclability and structural stability of the photocatalyst are two most important aspects. The recyclability studies were performed up to 4 cycles, wherein the yield of NH₃ was decreased from 36 μmol to 26 μmol h⁻¹ g⁻¹ (Fig. S12a), which could be ascribed to the loss of the catalyst between the different cycles, as the photocatalyst was washed and dried before being subjected to the next recyclability cycle. Due to a decrease in the catalyst amount with each cycle, the photocatalytic amount of NH₃ produced also decreased. The recovered photocatalyst was subjected to PXRD analysis, as shown in Fig. S12b, which confirmed its structural integrity.

3.4. Photoluminescence, photocurrent, and lifetime decay studies

To analyze charge recombination in the catalysts, photoluminescence (PL) studies were employed (Fig. 9a). The emission spectra of all the samples were measured at an excitation wavelength of 320 nm. The intensity for the PL spectra was maximum for the pristine NNO. The order for PL intensity was observed as follows: NNO > NVNO0.50 > DNNO1 > NVNO0.75 > DNNO3 > NVNO0.25 > DNNO2. The lowest intensity of NVNO0.25 and DNNO2 indicates the slowest recombination of charge carriers. However, the photocatalytic activity of NVNO0.25 was found to be the highest among all the samples, which is due to the synergistic effect of O_V and V-doping in the crystal lattice.

Fig. 9 (a) PL spectra (λ_ex = 320 nm) and (b) photocurrent plots of NNO, DNNO2 and NVNO0.25, respectively.

Photocurrent studies were also performed to find out the density of charge carriers under dark and light conditions, as shown in Fig. 9b. The electrolyte used in this study was 0.1 M Na₂SO₄ and the potential applied was 0 V vs. Ag/AgCl in all cases. It was found that the change in current density was almost similar for NVNO0.25 and DNNO2 from dark to light conditions. The change in the current density is enhanced by the enhanced separation of charge carriers (e⁻ and h⁺). These results corroborate well with the PL studies. The order for the current density in light was observed as follows: NVNO0.25 > DNNO2 > NNO. The photocurrent densities of NNO, DNNO2 and NVNO0.25 are recorded as 0.029, 0.047 and 0.054 µA cm⁻². The highest current density of NVNO0.25 clearly indicates its prominence for photocatalytic NH₃ production.

Furthermore, fluorescence spectroscopy was employed to study the lifetime decay of NNO, DNNO2 and NVNO0.25. Fig. S13 shows the decay of the samples that were fitted with the triexponential function. The average lifetime of the samples (NNO, DNNO2, and NVNO0.25) was observed to be 0.5 ns, 0.9 ns and 0.9 ns, respectively. The lifetimes of DNNO2 and NVNO0.25 were increased as compared to their pristine counterpart (NNO). This enhancement of the lifetime could be attributed to the introduction of defects via O_V and doping in the lattice system of DNNO2 and NVNO0.25. Furthermore, this increased lifetime of DNNO2 and NVNO0.25 indicates the increased photocatalytic activity.

4. Conclusion

In this work, we report a simple reaction procedure using O_V and V-doping engineered NaNbO₃ nanosheets produced by varying proportions (vacancy and doping) for boosting the overall photocatalytic nitrogen fixation to ammonia and nitrates. Through a variety of detailed characterization methods, the impact of vacancies and V-doping on the morphology, structure, optical properties and photocatalytic activity of NaNbO₃ was investigated. PXRD and Raman spectroscopy confirmed lattice distortion and defect creation in the oxygen-vacancy-rich (DNNO) and V-doped NaNbO₃ (NVNO) samples while morphological analysis using SEM and TEM revealed no appreciable changes. Additionally, according to BET and TPD measurements, the inclusion of defects and the enhanced surface area improved the chemical interaction between the catalyst surface and the chemisorbed nitrogen molecules. Among all the oxygen-vacancy-rich (DNNO) and V-doping (NVNO) catalysts, NVNO0.25 exhibited the highest ammonia production, which can be ascribed to the synergistic effect between the induced O_V and V-doping that enhances the light absorption by tuning the band gap in the visible region. Ammonia quantification was carried out by the Nessler's reagent method along with the quantification of other reduced and oxidized products by specific methods. Subsequently, photocatalytic activity under visible light and sunlight was provided along with light to chemical conversion and turnover frequency comparison. Furthermore, the catalyst showed promising recyclability up to four cycles without losing its structural integrity. In summary, this study compares the effects of vacancies and heteroatom doping in NaNbO₃ semiconductor photocatalysts to open new avenues for catalyst design in photocatalytic nitrogen fixation research.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

The data supporting this article are included in the main text and the supplementary information (SI). Supplementary information: materials characterization and quantification of NH₃ using different methods; quantification of N₂H₄ using the Watt–Chrisp method; quantification of NO₃⁻ and NO₂⁻ species; preparation of samples for PL and photocurrent studies; calculated lattice parameters; crystallite size and lattice strain; FTIR spectra; images of the samples; Tauc plots; band diagrams of NNO, DNNO2 and NVNO0.25; TPD plots; EPR spectra; multi-point BET and pore size distribution curves; calibration curve for the Nessler's reagent method; calculated yields for all the samples; calibration curve for N₂H₄; calibration curve for NO₃⁻; ion chromatograms obtained for different photocatalysts for the determination of NO₃⁻ ions; calibration curve for NO₃⁻ and calculated yield of NO₂⁻; NMR spectrum of NH₃ obtained using NVNO0.25; UV-vis and absorption spectra of controlled studies; calculation of light to chemical conversion (LCC); calculation of turnover frequency (TOF); comparison table of photocatalytic activities of different materials; UV-vis spectra for recyclability and histogram; PXRD pattern of the recycled sample; lifetime decay curves; and references. See DOI: https://doi.org/10.1039/d5nr03473f.

Acknowledgements

We are thankful to the Advanced Material Research Centre (AMRC), IIT Mandi for laboratory and characterization facilities. VK acknowledges DST-SERB (CRG/2022/003559) for funding this research. AP thanks ANRF (CRG/2021/008526) for funding. MS thanks the Ministry of Education (MoE), India, for her doctoral research fellowship. BPM thanks the Department of Science and Technology, Government of India, for the National Postdoctoral Fellowship (NPDF).

References

1. M. Cheng, C. Xiao and Y. Xie, J. Mater. Chem. A, 2019, 7, 19616–19633.
2. A. Kumar and V. Krishnan, Adv. Funct. Mater., 2021, 31, 2009807.
3. X. Chen, J.-Y. Li, Z.-R. Tang and Y.-J. Xu, Catal. Sci. Technol., 2020, 10, 6098–6110.
4. R. Shi, Y. Zhao, G. I. Waterhouse, S. Zhang and T. Zhang, ACS Catal., 2019, 9, 9739–9750.
5. M. Sharma, A. Kumar, D. Sajwan, K. Kumari, B. P. Mishra and V. Krishnan, Adv. Sustainable Syst., 2025, 9, 2400903.
6. A. Kumar and V. Krishnan, Adv. Sustainable Syst., 2024, 8, 2400173.
7. R. Li, Chin. J. Catal., 2018, 39, 1180–1188.
8. X.-H. Li, H. Li, S.-L. Jiang, L. Yang, H.-Y. Li, Q.-L. Liu, W. Bai, Q. Zhang, C. Xiao and Y. Xie, ACS Catal., 2023, 13, 7189–7198.
9. S. Bai, N. Zhang, C. Gao and Y. Xiong, Nano Energy, 2018, 53, 296–336.
10. S. Yang, A. Verdaguer-Casadevall, L. Arnarson, L. Silvioli, V. Colic, R. Frydendal, J. Rossmeisl, I. Chorkendorff and I. E. Stephens, ACS Catal., 2018, 8, 4064–4081.
11. A. Kumar, M. Kumar, V. N. Rao, M. V. Shankar, S. Bhattacharya and V. Krishnan, J. Mater. Chem. A, 2021, 9, 17006–17018.
12. J. Xiong, J. Di, J. Xia, W. Zhu and H. Li, Adv. Funct. Mater., 2018, 28, 1801983.
13. D. Maarisetty and S. S. Baral, J. Mater. Chem. A, 2020, 8, 18560–18604.
14. C. Mao, J. Wang, Y. Zou, H. Li, G. Zhan, J. Li, J. Zhao and L. Zhang, Green Chem., 2019, 21, 2852–2867.
15. W. Ding, S. Yuan, Y. Yang, X. Li and M. Luo, J. Mater. Chem. A, 2023, 11, 23653–23682.
16. Z. Zafar, S. Yi, J. Li, C. Li, Y. Zhu, A. Zada, W. Yao, Z. Liu and X. Yue, Energy Environ. Mater., 2022, 5, 68–114.
17. M. Sharma, S. Suthar, M. Sarkar, A. Patra and V. Krishnan, ACS Appl. Nano Mater., 2025, 8, 1338–1351.
18. Y. Sun, H. Ji, Y. Sun, G. Zhang, H. Zhou, S. Cao, S. Liu, L. Zhang, W. Li and X. Zhu, Angew. Chem., 2024, 136, e202316973.
19. Q. Meng, C. Cao, J. Wang, L. Tian, Y. Huang, M. Yang, M. Wang, B. Cong and Y. Zhang, ACS Appl. Nano Mater., 2024, 7, 15808–15818.
20. Q. Yang, Y. Wang, Q. Tian, X. Li, A. Pan, M. Zhao, Y. Zhu, T. Wu and G. Fang, J. Mater. Chem. A, 2024, 12, 7207–7214.
21. X. Liu, J. Wang, F. Zhu, Y. Li, W. Tian, W. Wang, R. Guo, L. Liu and J. Shi, J. Mater. Chem. A, 2024, 12, 9661–9671.
22. M. Sharma, D. Sajwan, A. Gouda, A. Sharma and V. Krishnan, Photochem. Photobiol., 2024, 100, 830–896.
23. M. Sharma, R. Singh, A. Sharma and V. Krishnan, Dalton Trans., 2024, 53, 6731–6746.
24. D. Sajwan, A. Sharma, M. Sharma and V. Krishnan, ACS Catal., 2024, 14, 4865–4926.
25. M. Sharma, A. Kumar and V. Krishnan, Nanotechnology, 2022, 33, 275702.
26. A. Semwal, D. Sajwan, J. Rawat, L. Gambhir, H. Sharma and C. Dwivedi, Environ. Sci. Pollut. Res., 2023, 30, 45827–45839.
27. A. Ranjbari, J. Yu, J. Kim, J. Kim, M. Park, K.-H. Kim and P. M. Heynderickx, Appl. Surf. Sci., 2024, 659, 159867.
28. D. Sajwan, M. Sharma, S. Kumar, J. Dahiya, B. P. Mishra and V. Krishnan, ChemCatChem, 2025, 17, e00415.
29. A. Kumar, A. Kumar and V. Krishnan, ACS Catal., 2020, 10, 10253–10315.
30. M. A. Peña and J. L. Fierro, Chem. Rev., 2001, 101, 1981–2018.
31. J. Song, K. Kawakami and K. Ariga, Adv. Colloid Interface Sci., 2025, 103420.
32. V. Kumar, S. Kumar, D. Sharma and V. Krishnan, Macromol. Rapid Commun., 2025, 46, e00497.
33. A. Kumar, M. Sharma, S. Sheoran, S. Jaiswal, A. Patra, S. Bhattacharya and V. Krishnan, Nanoscale, 2023, 15, 11667–11680.
34. L. Guo, S. You, C. Wu, F. Liu, R. Zhang and X. Wang, Inorg. Chem., 2024, 63, 11832–11841.
35. X. Yan, S. Zhang, L. Pan, T. Ai, Z. Li and Y. Niu, Inorg. Chem. Commun., 2023, 158, 111510.
36. J. Zhang, L. Yue, Z. Zeng, C. Zhao, L. Fang, X. Hu, H. Lin, L. Zhao and Y. He, J. Colloid Interface Sci., 2023, 636, 480–491.
37. J. Wu, W. Li, S. Guan, X. Chen, H. Gao and X. Liu, Inorg. Chem. Commun., 2021, 131, 108669.
38. B. Yang, J. Bian, L. Wang, J. Wang, Y. Du, Z. Wang, C. Wu and Y. Yang, Phys. Chem. Chem. Phys., 2019, 21, 11697–11704.
39. A. V. Abega, C. Marchal, M.-A. Dziurla and D. Robert, J. Photochem. Photobiol., A, 2024, 115850.
40. D. Sajwan, A. Semwal, J. Rawat, H. Sharma and C. Dwivedi, Mater. Today: Proc., 2023, 73, 180–188.
41. L. Li, S. Lu, W. Cao, Q. Zhu, R. Li, Y. Wei, S. Yang and C. Wang, Inorg. Chem., 2024, 63, 11745–11756.
42. R. Nain and R. K. Dwivedi, J. Solid State Chem., 2024, 329, 124378.
43. M. R. Vicente, G. T. Xavier, W. A. Carvalho, J. M. De Carvalho and J. S. Souza, Appl. Surf. Sci., 2024, 664, 160232.
44. L. Wang, Y. Cai, B. Liu and L. Dong, Solid State Ionics, 2021, 361, 115564.
45. M. I. Prakoso, K. Y. Maharani, M. R. Ariq, A. N. Indah, A. Apriansyah, A. Iryani, L. J. Kusumawardani and Y. Syahputri, Helium, 2024, 4, 01–07.
46. K. Spilarewicz, K. Mróz, M. Kobielusz and W. Macyk, Curr. Opin. Chem. Eng., 2024, 45, 101041.
47. F. Beigi, A. R. Mahjoub and A. H. C. Khavar, Appl. Surf. Sci., 2023, 637, 157972.
48. Y. Ren, Y. Li, X. Wu, J. Wang and G. Zhang, Chin. J. Catal., 2021, 42, 69–77.
49. H. Shi, G. Chen, C. Zhang and Z. Zou, ACS Catal., 2014, 4, 3637–3643.
50. R. Jana, A. Gupta, R. Choudhary and O. Pandey, J. Sol-Gel Sci. Technol., 2020, 96, 405–415.
51. Y. Yi, M. Wu, R. Wang, Q. He, P. Sun, B. Zhou and X. Dong, Chem. Eng. J., 2024, 485, 149885.
52. Y. Li, L. Li, F. Liu, B. Wang, F. Gao, C. Liu, J. Fang, F. Huang, Z. Lin and M. Wang, Nano Res., 2022, 15, 7986–7993.
53. S. Zhang, Y. Zhao, R. Shi, C. Zhou, G. I. Waterhouse, L. Z. Wu, C. H. Tung and T. Zhang, Adv. Energy Mater., 2020, 10, 1901973.
54. G. Zhang, X. Yang, C. He, P. Zhang and H. Mi, J. Mater. Chem. A, 2020, 8, 334–341.
55. L. Huang, T. Peng, Y. Gong, R. Wang, B. He, J. Jin and H. Wang, Dalton Trans., 2024, 53, 12291–12300.
56. P. Zhang, L. Chen, D.-H. Kuo, B. Wu, Z. Su, D. Lu, Q. Wu, J. Li, J. Lin and X. Chen, J. Mater. Chem. A, 2024, 12, 7163–7177.
57. N. K. M Safeer, C. Sathiskumar and N.S. John, ChemistrySelect, 2023, 8, e202203344.
58. T. Ma, C. Yang, L. Guo, R. A. Soomro, D. Wang, B. Xu and F. Fu, Appl. Catal., B, 2023, 330, 122643.
59. Y. Zhao, Y. Zhao, G. I. Waterhouse, L. Zheng, X. Cao, F. Teng, L. Z. Wu, C. H. Tung, D. O'Hare and T. Zhang, Adv. Mater., 2017, 29, 1703828.
60. M. Sharma, A. Kumar, D. Gill, S. Jaiswal, A. Patra, S. Bhattacharya and V. Krishnan, ACS Appl. Mater. Interfaces, 2023, 15, 55765–55778.

---