Titanium dioxide (TiO₂, 98.50%), sodium hydroxide (NaOH, 96.00%), light magnesium oxide (MgO, 98.50%), nickel nitrate [Ni(NO₃)₂·6H₂O] and ethanol (C₂H₅OH, 99.70%) were obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Ultra-high purity CO₂ (>99.99%) was supplied by Xuzhou Hongfeng Gas Co., Ltd. The distilled water employed in our experiments exhibits a conductivity of 0.82 μS/cm (±0.1 μS/cm measurement uncertainty) at 25 °C, meeting the requirements for analytical grade purity.

X-ray Diffraction (XRD) analysis was performed on a D/Max-2550 diffractometer (Cu Kα, λ=0.15 nm) from 10°–80° 2θ at 40 kV and 100 mA. The sample morphology was examined using an FEI Nova Nano SEM 450 field emission scanning electron microscope (SEM) at 10.00 kV accelerating voltage. The morphology and crystal structure of samples were investigated using field emission high-resolution transmission electron microscopy (FE-HRTEM) with SAED attachment (G2F20). The X-ray photoelectron spectroscopy (XPS) measurements were performed on a VG ESCALAB MKII (VG Scienta USA) system (Mg Kα source), with energy calibration referenced to C 1s at 284.60 eV. Data analysis was performed with DigitalMicrograph software and optical properties were characterized by UV-vis spectroscopy (Lambda 950) across the 200−800 nm spectral range. The specific surface area (S_BET) was determined by N₂ adsorption-desorption isotherms measured at −196 °C (77 K) using a Micromeritics ASAP 2010 analyzer (USA). The photoluminescence (PL) spectra were acquired on a Cary-Eclipse 500 system (325 nm excitation) to characterize the separation efficiency of photoinduced electron-hole pairs. The CHI 660E electrochemical workstation was employed to record both transient photocurrent and Nyquist plots under simulated solar irradiation. The elemental composition of the samples was determined by inductively coupled plasma optical emission spectrometry/mass spectrometry (ICP-OES/MS) analysis.

A series of nano-catalysts comprising NNMTO-x solid solutions with varying molar ratios of NNTO and NMTO were synthesized via a one-step hydrothermal method. Typically, MgO and Ni(NO₃)₂·6H₂O were dissolved in a 1.20 mol/L NaOH aqueous solution, while maintaining a constant mass of TiO₂. The mixture was subjected to ultrasonication for 10 min followed by magnetic stirring for 30 min, transferred into a 130 mL stainless steel autoclave and hermetically sealed. The autoclave was gradually heated to 300 °C at a controlled heating rate of 3 °C/min and maintained isothermally for 80 min to ensure complete reaction. After natural cooling to room temperature, the samples were collected. The NNMTO-x solid solution nanocrystals were isolated through sequential washing with distilled water and ethanol (three times), then dried at 60 °C for 8 h in oven.

The photocatalytic activity for CO₂ reduction was evaluated under xenon lamp irradiation (300 W) without the addition of sacrificial reagents. In a typical procedure, 15 mg of catalyst was uniformly dispersed on a quartz sand plate and subsequently transferred to a 100 mL static reactor equipped with a quartz window. Prior to irradiation, 2 mL of deionized water was injected into the reactor, followed by purging with high-purity CO₂ (99.99%) at a constant flow rate (20 mL/min) for 30 min to ensure complete removal of atmospheric gases. The photocatalytic reaction was conducted under full-spectrum xenon lamp irradiation (350−780 nm) for 6 h, with the light intensity being enhanced by an internal reflector system. The gaseous products were quantitatively analyzed at 1 h intervals using gas chromatography. The quantification of CO and CH₄ products was performed by gas chromatography (GC-7920, AuLight, China) equipped with a flame ionization detector.

Due to the absence of a reference XRD pattern for NNTO in the standard powder diffraction database, its theoretical diffraction pattern was computationally simulated employing the Vienna Ab-initio Simulation Package (VASP) based on density functional theory (DFT) calculations. The atomic structure of NNTO, as shown in Figure 1(a), was constructed based on the isostructural analog of NMTO^[25]. The simulated XRD pattern of NNTO, presented in Figure 1(b), demonstrates excellent agreement with the diffraction pattern of NMTO (PDF No. 44-0007), confirming their structural similarity.

Figure 2(a) presents the XRD patterns of NNMTO-x samples with varying mass concentrations. The diffraction peaks of NNMTO-x show excellent agreement with the standard pattern of NMTO^[24], indicating consistent crystal structures across the different compositions. The structures of NMTO and NNTO are similar and their XRD diffraction patterns almost coincide, the XRD diffraction patterns remain nearly unchanged with increasing Ni content. The characteristic diffraction peak of NNMTO-x, as indicated within the circled region, exhibits modification, presumably attributed to the interstitial incorporation of Ni atoms into the NMTO crystal lattice. The magnified view of the strongest peak at approximately 24.5° provides conclusive evidence for the successful synthesis of NNMTO-x continuous solid solutions. As illustrated in Figure 2(b), the diffraction peaks of NNMTO-x solid solutions exhibit a slight angular shift compared to pure NMTO, which attributable to the interstitial incorporation of Ni ions within the crystal lattice.

Figure 3(a) presents the scanning electron microscopy (SEM) image of NMTO, revealing well-defined hexagonal plate-like morphology with smooth surfaces.

The observed morphological characteristics are in excellent agreement with those of pure NMTO^[24], demonstrating the preservation of the distinctive hexagonal structure. The morphological characteristics of NNMTO-x solid solutions remain virtually identical to those of pure NMTO when the NNTO molar ratio is maintained below 40% (Figure 3(b)−(d)). However, a significant morphological transformation is observed when the Ni molar ratio exceeds 40% (Figure 3(e)−(f)). This structural evolution can be attributed to the interstitial incorporation of Ni atoms into the NMTO crystal lattice, which presumably alters the surface energy and growth kinetics of the crystals.

Figure 4(a) shows the TEM image of NNMTO-30 at 200 kV, with Figure 4(b) presenting the HRTEM image from the red rectangular area. The measured lattice spacings of 1.73, 2.39 and 1.99 Å correspond to the (113), (401) and ($ 40\bar 3 $) planes of NMTO, respectively. As evidenced by Figure 4(c), (d), the interplanar spacing remains virtually unchanged, demonstrating that NNMTO-x (x<40) retains the structural characteristics of pure NMTO. Figure 4(f) presents the corresponding HRTEM image obtained from the region delineated by the yellow rectangle in Figure 4(e). The crystal fringes of d=3.63, 3.63 and 5.87 Å are determined as (110), ($ 1\bar 1 0$) and (200) crystal planes of NNMTO-40. Measurable variations in interplanar spacing were observed among the samples, resulting from localized strain induced by the incorporation of Ni²⁺ ions into the NMTO lattice through ionic radius mismatch with host cations. The chemical composition analysis, as determined by the EDX spectrum in Figure 4(i), (j), confirms the presence of Na, Mg, Ti, Ni and O elements in the NNMTO-x samples.

Moreover, the intensity of the Ni characteristic signal exhibits a proportional increase with the rising molar concentration of Ni, demonstrating the successful incorporation of nickel into the crystal structure. The ICP-OES/MS analysis results are summarized in Table 1, the measured mass fractions of all constituent elements (Na, Ni, Mg, Ti, O) show relatively agreement with the theoretical stoichiometric ratios, confirming the successful formation of the target solid solution phase.

XPS measurements were conducted for both NMTO and NNMTO-x samples.

The XPS spectra in Figure 5(a)−(c) reveal characteristic binding energies corresponding to the primary oxidation states of constituent elements in NMTO: Na 1s at 1071.20 eV (Na⁺), Mg 2p at 49.20 eV (Mg²⁺) and Ti 2p doublet at 458.20 and 464.00 eV (Ti⁴⁺). The O 1s spectrum presented in Figure 5(d) can be fitted to three distinct components with binding energies at 529.20, 531.30 and 533.10 eV, corresponding to lattice oxygen (O²⁻) in the crystal structure, surface-adsorbed water molecules (H₂O), and surface hydroxyl groups (−OH) of NMTO, respectively. The Ni 2p XPS spectrum in Figure 5(e) was deconvoluted into four characteristic peaks at binding energies of 855.20 eV (Ni 2p_3/2), 86.80 eV (satellite), 873.40 eV (Ni 2p_1/2) and 879.50 eV (satellite), confirming the presence of Ni²⁺ species within the solid solution.

Figure 6 presents the UV-visible absorption spectra of NMTO and NNMTO-x samples, recorded using a UV-visible spectrophotometer in the wavelength range of 200−800 nm. The pristine NMTO exhibits the highest absorption intensity, which can be attributed to its characteristic gray coloration.

Notably, the absorption edge demonstrates a progressive red shift with increasing NNTO content, indicating enhanced visible light absorption capability in the NNMTO-x solid solutions. The optical band gap energies were determined using the Tauc plot method based on the following equation^[26]:

The physical significance of the parameters α, ν, B and E_g in the Tauc equation is summarized in Table 2. The parameter n, which determines the transition type, is fundamentally related to the electronic band structure of the semiconductor. Specifically, a value of n=1 corresponds to a direct bandgap transition, where the minimum conduction band energy aligns with the maximum valence band energy in momentum space. Conversely, a value of n=4 indicates an indirect bandgap transition, requiring phonon assistance due to the misalignment of the conduction band minimum and valence band maximum in the Brillouin zone. Figure 6(a) demonstrates that NNTO exhibits significantly higher optical absorbance compared to the NNMTO-x samples, primarily attributed to its intrinsic gray coloration resulting from enhanced light absorption characteristics.

Figure 6(b) illustrates the Tauc plot representation of (αhν)² versus photon energy (hν), where the optical band gap is determined by extrapolating the linear region of the curve to intersect with the hν axis. Also, this indicates the prepared NNMTO-x is a direct bandgap transition semiconductor. Notably, the absorption edges of NNMTO-x samples exhibit a significant red shift compared to pristine NMTO, indicating an extended visible light absorption range. This enhanced optical property is particularly advantageous for photocatalytic applications, as it facilitates more efficient utilization of solar energy.

Figure 7 presents the nitrogen adsorption-desorption isotherms of NMTO and NNMTO-x, measured using the Brunauer-Emmett-Teller (BET) method for specific surface area determination. As summarized in Table 3, the formation of solid solutions leads to a noticeable enhancement in specific surface area. Among the synthesized samples, NNMTO-30 demonstrates optimal textural properties, exhibiting the maximum specific surface area (31.77 m²/g), the highest total pore volume (0.12 cm³/g), and the smallest average pore diameter (3.86 nm). However, a gradual decrease in specific surface area is observed with increasing NNTO molar ratio. It is mainly due to the proper ratio of NNTO to NMTO ensures the dispersion of Ni species on the surface of NMTO, forming a stable NNTO-NMTO interface.

However, excessive NNTO, due to agglomeration or coverage of the surface sites and mesopores of NMTO, reduces both the specific surface area and the number of effective NNTO-NMTO interfaces. Nevertheless, the overall variation in specific surface area across the NNMTO-x series remains relatively modest compared to pristine NMTO.

The photoelectrochemical performance of the samples was evaluated through chronoamperometric measurements under ultraviolet-visible light irradiation (λ=365 nm) using a 0.5 mol/L Na₂SO₄ aqueous solution as the electrolyte. Figure 8(a) demonstrates a rapid photocurrent response upon illumination, characterized by an immediate increase when the light source is activated and a sharp decrease upon light termination. This reproducible on-off photocurrent behavior clearly indicates the efficient generation and separation of charge carriers under light irradiation, followed by their rapid recombination when the illumination ceases. A photocurrent spike was found in NMTO and NNMTO-x photocatalyst when the light was on, which is due to the recombination of holes and electrons. The NNMTO-30 sample exhibits significantly enhanced photocurrent density compared to pristine NMTO, demonstrating superior charge separation efficiency and reduced electron-hole recombination rates, which can be attributed to its optimized electronic structure and improved light absorption characteristics^[27].

The formation of solid solutions significantly suppresses the recombination kinetics of photogenerated electron-hole pairs, thereby enhancing charge carrier utilization efficiency and consequently improving the overall photocatalytic performance. Figure 8(b) reveals that NNMTO-30 exhibits the smallest semicircle diameter in the Nyquist plot, indicating the lowest charge transfer resistance among the samples. This reduced impedance can be attributed to the effective suppression of electron-hole recombination, which is in excellent agreement with the enhanced photocurrent response observed in the transient photocurrent measurements.

Figure 9(a) demonstrates that NNMTO-30 exhibits the smallest arc radius in the electrochemical impedance spectroscopy (EIS) Nyquist plot, indicating superior charge transfer characteristics resulting from effective suppression of electron-hole recombination. The interfacial charge transfer properties can be further analyzed through Mott-Schottky measurements, where the relationship between the space charge layer capacitance (C) and the applied potential (V) is described by the following equation:

The physical significance and corresponding units of the parameters in the Mott-Schottky equation are summarized in Table 4.

The thermal voltage term ($\dfrac{{T{k_{\mathrm{B}}}}}{q} $), which has a nominal value of approximately 25 mV at room temperature (298 K), can be neglected in the calculation. The correlation between the donor density (ND), conduction band edge potential (ECB), and Fermi level (EF) can be expressed by the following relationship:

As shown in Figure 9(b), the positive slope of the Mott-Schottky plot confirms the n-type semiconductor characteristics of NNMTO-x. Furthermore, a distinct positive shift in the flat band potential is observed for the NNTO-x solid solutions compared to pristine NMTO, suggesting an upward movement of the Fermi level towards the conduction band, which corresponds to an increase in electron energy within the material system.

The photocatalytic performance for CO₂ conversion was evaluated under UV-visible light irradiation without the addition of any sacrificial reagents. As demonstrated in Figure 10, the NNMTO-x series exhibits remarkable CO₂ photoreduction activity, suggesting efficient utilization of photogenerated charge carriers for the photocatalytic conversion process. As presented in Figure 10(a), (b), the photocatalytic CO₂ reduction process demonstrates significantly higher CO production compared to CH₄ generation during the initial hour of illumination across all catalyst samples.

Notably, the formation of solid solutions substantially enhances the CO₂ reduction performance, which can be primarily attributed to the effective suppression of electron-hole recombination and the consequent improvement in charge carrier utilization efficiency. Among the synthesized catalysts, NNMTO-30 exhibits optimal photocatalytic performance, achieving remarkable CO₂ reduction rates of 10.26 μmol/(g·h) for CO and 0.33 μmol/(g·h) for CH₄ during the initial hour of illumination. As illustrated in Figure 10(c), (d), the extended photocatalytic evaluation over 6 h of continuous irradiation reveals that the NNMTO-x series maintains superior CO₂ conversion efficiency compared to both pristine NMTO and other reference catalysts, demonstrating its exceptional photocatalytic stability and performance. The photocatalytic performance exhibits a gradual degradation when the Ni molar ratio surpasses 40%, primarily due to the excessive coverage of active surface sites on NNMTO-x. This surface modification leads to diminished light absorption capacity and consequently reduces the generation of photogenerated charge carriers, ultimately impairing the overall photocatalytic efficiency. As evidenced by the control experiments presented in Figure 10(e), (f), no detectable CO or CH₄ production was observed in the absence of any essential photocatalytic components (light irradiation, catalyst, or CO₂), confirming that the reduction products are exclusively generated through photocatalytic processes. Furthermore, the significant decrease in CO and CH₄ yields when employing a 400 nm cutoff filter demonstrates the crucial role of ultraviolet light in driving the photocatalytic reactions, as the energy of visible light photons is insufficient to excite the catalyst effectively. Figure 10(g) presents the cyclic stability evaluation of NNMTO-30 through five consecutive CO₂ photocatalytic reduction tests, maintaining consistent production rates of 42.38 μmol/g for CO and 1.47 μmol/g during the 6 h irradiation period in the final cycle. As demonstrated in Figure 10(h)−(i), both the XRD patterns and morphological characteristics remain virtually unchanged after five reaction cycles, confirming the exceptional photostability and structural integrity of NNMTO-30 in photocatalytic CO₂ reduction applications. Comparative analysis with literature-reported catalysts revealed a modest performance gap in our system (Table 5), primarily attributable to the differences in experimental conditions (e.g., the quality of the photocatalyst, sacrificial agent, et al.) and variations in co-catalyst loading, et al. In subsequent studies, we propose to construct a g-C₃N₄ or pyridine-nitrogen modified NMTO heterojunction system, where g-C₃N₄ or pyridine-nitrogen will serve as cocatalyst to enhance CO₂ reduction performance through improved charge separation and visible-light absorption.

Figure 11 illustrates the proposed mechanism for photocatalytic CO₂ reduction over NNTO-x. The fundamental principle governing photocatalytic reactions requires that: the semiconductor's conduction band minimum should exceed the reaction's reduction potential, while its valence band maximum must surpass the oxidation potential. These energetic alignments facilitate the efficient transfer of photogenerated electrons from the conduction band to CO₂ molecules while enabling hole migration from the valence band to electron donors. Consequently, the appropriate alignment between the semiconductor's band edge positions and the redox potentials of reactants represents a crucial prerequisite for effective photocatalytic reactions.

The oxygen atom of CO₂ supply electrons to the Lewis acid center, while the carbon atom receives electrons from the Lewis base center^[28]. The adsorption of CO₂ molecules on the catalyst surface initiates the formation of $ * {\text{CO}} $ radical anions, which subsequently undergo a series of proton-coupled electron transfer processes to generate key reaction intermediates^[29]. The surface-bound intermediates may undergo two distinct pathways: (1) desorption from the catalyst surface to form CO as the final product, or (2) further reduction through sequential proton-electron transfer processes to generate C−H bond-containing intermediates, which ultimately lead to hydrocarbon formation^[30]. The CO₂ reduction mechanism proceeds through a series of well-defined steps: Initially, CO₂ undergoes single-electron reduction to form an activated CO₂·- radical anion^[31]. This reactive intermediate is subsequently protonated to generate the carboxyl radical (COOH·).

The carboxyl radical then undergoes further protonation to form formic acid (HCOOH), which dehydrates to produce the formyl radical (HCO·). Through sequential two-electron reduction and protonation, HCO· is converted to formaldehyde (HCOH). Subsequent hydrogenation reactions transform formaldehyde into methanol (CH₃OH)^[32]. The final step involves methanol reduction followed by dehydration, yielding methane (CH₄) as the end product.