Enhanced Catalytic Ozonation for Water Treatment via TiO₂-Modified LaMnO₃ Undergoing Efficient Mn³⁺/Mn⁴⁺ Redox Cycle

Abstract

The TiO₂-modified LaMnO₃ catalyst demonstrated outstanding catalytic performance across a broad pH range (4.2 to 10.0) and under various complex water conditions. It achieved complete degradation of the ibuprofen parent compound, attaining an 85.9% mineralization rate. The efficacy stems from the reversible Mn³⁺/Mn⁴⁺ redox couple. The ratio of Mn³⁺/Mn⁴⁺ was 3.9 for TiO₂-modified LaMnO₃, significantly higher than 1.2 for nanocast LaMnO₃. Experimental results and density functional theory revealed that La and Ti did not actively participate in the catalytic ozone reaction. Notably, the Mn³⁺/Mn⁴⁺ pair emerged as key drivers in the involvement of HO•, O₂•⁻, and ¹O₂ in the reactive oxygen species pathway. Notably, ozone exhibited preferential adsorption and activation on the (010) crystal face of the catalyst. A moderated reduction in interaction forces facilitated the Mn³⁺/Mn⁴⁺ redox cycle, resulting in increased production of reactive oxygen species. These findings contributed to the development of more efficient catalysts for environmental remediation.

1. Introduction

Heterogeneous catalytic ozonation has garnered significant attention in the water and wastewater treatment field due to its enhanced ozonation efficiency for pollutant removal. A range of catalysts have found use in catalytic ozonation, including carbonaceous materials, minerals, and metal-based systems [1]. The success of catalytic ozonation hinges upon the development of novel catalysts with exceptional activity. Perovskite, in particular, has become a favored catalyst in catalytic ozonation over the years. This is attributed to its remarkable design flexibility, the ability to create controllable defects for enhancing ionic or electronic conductivities, adjustable surface properties, and inherent stability [2,3]. Perovskites adopt the general formula ABO₃, with A and B denoting cations of distinct sizes [4]. La-containing (a rare-earth element) perovskite was reported to address the removal of ozone-resistant acids and pharmaceuticals and personal care products [2,5,6,7]. Manganese, chosen as the transitional metal at the B site, is favored due to its multiple oxidation states and low environmental toxicity [8,9]. The synergy among structural defects, oxygen vacancies, and abundant Mn³⁺ sites underpins the high efficacy of this catalytic ozonation system [7,10]. However, it is noteworthy that limited existing literature delves into the interconversion of Mn³⁺ and Mn⁴⁺ oxidation states within the catalytic ozone process.

Enhanced catalytic ozonation efficiency hinges on the tailored modification of perovskite. Reported methods included tailoring both the A and B perovskite sites, notably using oxalic acid to improve its performance [9]. Additionally, applying nano-casting as a surface engineering technique significantly increases the surface area of LaMnO₃ compared to the conventional sol–gel method [10]. Furthermore, Al₂O₃ and TiO₂ are suitable support materials for LaFeO₃, enabling control over morphology and enhancing stability in peroxymonosulfite activation and photo-assisted chemical wet peroxide oxidation [11,12,13]. TiO₂, in particular, exhibits excellent modification capabilities by regulating crystal growth, increasing surface area, and reducing charge recombination [14,15]. Considerable attention has been devoted to TiO₂/LaFeO₃ heterojunctions, whereas the potential of TiO₂-modified LaMnO₃ in catalytic ozonation remains unexplored. The robustness of such a modified catalyst under various treatment conditions has yet to be investigated. By combining TiO₂ with LaMnO₃, synergistic effects of the composite catalyst may result in improved pollutant degradation rates, enhanced catalyst stability, and more efficient utilization of available adsorption sites, thereby intensifying the ozonation process. However, the specific roles of individual components within the TiO₂-modified LaMnO₃ in the adsorption-catalysis ozone process remain unclear, necessitating further research to elucidate the regulation of crucial control steps and adsorption forces.

Density Functional Theory (DFT) simulations are a potent tool, unveiling intricate details in various chemical systems [16,17]. DFT calculations provide profound insights into electronic structures, exposing the nature of chemical bonds and electron density distribution [18,19]. Interaction analysis using DFT allows for a detailed examination of how molecules or materials interact. It can elucidate the energetics of chemical reactions, the binding affinities between molecules, and the steric hindrance that may affect reaction pathways [20]. The mechanism underlying the enhancement of catalytic ozone by metal atom pairs in TiO₂-modified LaMnO₃ remains unclear. The role of metal atoms needs to be revealed by elemental valence characterization and DFT simulations. Detailed bonding, orbitals, charge, and interaction analysis are lacking, especially in the adsorption-catalytic process of the O atom of O₃ and active sites. Resolving these issues shows essential implications for the adsorption-catalytic ozone processes on TiO₂-modified LaMnO₃.

Ibuprofen (IBP), a widely used non-steroidal anti-inflammatory drug for pain, fever, and inflammation, was selected as a model emerging organic pollutant with low persistence [21]. Conventional water and wastewater treatment processes such as solar photolysis, biodegradation, and activated sludge can only remove up to 85% of IBP from the influent load due to its refectory properties [22,23]. In this study, we synthesized TiO₂-modified LaMnO₃ using a sol–gel method, which served as a catalytic agent for ozone in removing IBP. Our research focused on three critical aspects: (I) the synthesis and comprehensive analysis of the catalysts; (II) the evaluation of catalytic ozone performance under various operational conditions and in complex water matrices; and (III) the elucidation of the adsorption-catalysis mechanism based on experimental results and DFT calculations.

2. Materials and Methods

2.1. Chemicals and Reagents

IBP (Chemical Abstracts Service (CAS) No. 58560-75-1, its physicochemical property is listed in Table S1), TiO₂ (CAS No. 13463-67-7), citric acid (CAS No. 77-92-9), humic acid (HA, CAS No. 68131-04-4), La(NO₃)₃·6H₂O (CAS No. 10277-43-7), and Mn(NO₃)₂·xH₂O (CAS No. 15710-66-4) were obtained from Sigma-Aldrich (Singapore). Additional chemicals and reagents (e.g., tert-butanol (TBA), HCl (36.0–38.0) wt%, NaOH, 4-chlorobenzoic acid, 1-butanol, p-benzoquinone (p-BQ), and methanol) of analytical grade were provided by Sigma-Aldrich (Singapore). Ultrapure water (>18.25 MΩ/cm) was employed for the preparation of standard stock solutions and simulated IBP-containing wastewater.

2.2. Preparation and Characterization

Following a standard citric acid-mediated sol–gel procedure [24], a mixture of ethanol and ultrapure water (1:1, 50 mL total) was prepared, followed by the addition of 10 mmol of both La(NO₃)₃·6H₂O and Mn(NO₃)₂·xH₂O. Subsequently, 20 mmol citric acid was slowly introduced with stirring. A pale-yellow gel was obtained after heating the mixed solution in a water bath at 85 °C. The pale-yellow gel was collected, transferred to a ceramic crucible, and subjected to calcination at 750 °C for 4 h (ramping rate of 5 °C min⁻¹) under air atmosphere. The resulting sample was labeled as LaMnO₃. TiO₂-modified LaMnO₃ was prepared in the same method, using TiO₂ powders as an additional precursor alongside the metal salts [13].

The prepared catalysts underwent characterization using various techniques, including N₂ sorption, X-ray diffraction (XRD), electron microscopy (SEM and TEM), thermos-gravimetric analysis (TGA), Fourier transform infrared (FT-IR) spectroscopy, X-ray photoelectron spectroscopy (XPS), and electron spin resonance (ESR). A modified pH drift method was employed to determine the point of zero charge (PZC) of the catalyst [25]. Detailed characterization methods can be found in the Supporting Information.

2.3. Experimental Set-Up and Conditions

A schematic representation of the catalytic ozonation setup is depicted in Figure S1. The experiments were conducted in a Duran bottle with 500 mL of simulated wastewater. The typical initial IBP concentration was 1.0 mg/L, and the catalyst loading was set at 500 mg/L. Ozone gas was uniformly distributed within the reactor using a porous diffuser at a flow rate of 300 mL/min and a concentration of 10 mg/L, ensuring efficient gas–liquid mass transfer. The solution temperature was maintained between 28–30 °C, and the initial pH was adjusted to approximately 6.4 using either 0.1 mol/L HCl or NaOH solution. At predetermined time intervals, samples were collected, filtered using a 0.45 μm syringe filter, and quenched with 0.2 mol/L Na₂S₂O₃ [16]. The one-factor-at-a-time method was adopted to evaluate the impact of key operational variables (catalyst and ozone dosages, initial IBP, pH) and coexisting water constituents (Cl⁻, SO₄²⁻, humic acid) [19,26,27]. To ensure statistical reliability, each experiment was performed in triplicate.

2.4. Analytical Methods

IBP concentrations were determined via LC-MS/MS (LCMS-8030, Shimadzu, Kyoto, Japan) using a Zorbax SB C18 column (2.1 × 150 mm, particle size 5 μm, Agilent, Santa Clara, CA, USA) at 40 °C. A flow rate of 0.25 mL/min was applied for the mobile phase, which consisted of 99% methanol and deionized water [28]. IBP mineralization was evaluated via dissolved organic carbon (DOC) analysis using a TOC-L total organic carbon analyzer (Shimadzu, Kyoto, Japan). Before analysis, all samples were filtered through a 0.45 μm filter paper or syringe filter. The DOC values obtained were taken as an indicator of mineralization efficiency. The investigation of ROS was conducted using electron spin resonance (ESR) with a Bruker EMX PLUS spectrometer (Bruker BioSpin, Ettlingen, Germany). For detailed ESR methods, please refer to Text S3 in the Supplementary Information.

2.5. Density Functional Theory Calculations

A cutoff energy of 550 eV and a 2 × 2 × 1 supercell were adopted for structural optimization, with convergence achieved when all residual forces fell below 0.0001 eV/Å. Total energy calculations utilized a KPT-Resolved 4 × 3 × 1 K-Mesh. Structural optimizations and electronic calculations were performed using the Vienna Ab initio Simulation Package (VASP), employing the projector augmented wave (PAW) method and the Perdew–Burke–Ernzerhof (PBE) functional. [29,30,31]. A Hubbard correction of U = 3.9 eV was added to treat d orbitals [32]. To visualize the results, electron localization function (ELF) maps, partial density of states (PDOS) curves, and crystal orbital overlap population (COOP) curves were displayed using VESTA, P4VASP, and wxDragon, respectively [33,34]. The ELF value ranges from 0 to 1, where 1 indicates complete electron localization, and 0 represents complete electron delocalization [35]. Additional features were obtained using VASPKIT [36]. For the catalytic ozonation process, Nudged Elastic Band calculations were performed with nine replicas, including the initial and final states [37]. Independent gradient model (IGM) analysis based on the Hirshfeld partition was performed with Multiwfn 3.8 [38].

3. Results and Discussion

3.1. Characterization of Catalysts

As shown in the TEM results of Figure 1a, the d-spacing of TiO₂-modified LaMnO₃ (110) was determined to be 0.277 nm, which closely matches the d-spacing value of the single-phase LaMnO₃ (110) at 0.2755 nm, as reported in JCPDS 50-0298. SEM images (Figure S2) revealed diverse particle shapes of TiO₂-modified LaMnO₃, exhibiting both regular cubic morphology and irregular forms. Energy dispersive X-ray spectrometry analysis demonstrated the presence of La, Mn, Ti, and O elements on the sample’s surface, with relative contents of 8.31%, 7.75%, 7.10%, and 76.74%, respectively (Figure S3 and Table S2). The TiO₂-modified LaMnO₃ exhibited a specific surface area of 24.43 m²/g, a pore volume of 0.059 cm³/g, and an average pore diameter of 4.8 nm (Figure S4 and Table S3).

FT-IR analysis (Figure 1b) indicated peaks below 1000 cm⁻¹, which were associated with metal–oxygen bonds [39]. Specifically, the absorption bands at 561, 609, and 621 cm⁻¹ were assigned to Mn-O or Mn-O-Mn bonds [40], while the absorption band at 491 cm⁻¹ was attributed to Ti-O bonds [41]. The firm diffraction peaks observed at (012), (110), (202), (024), (214), (208), and (128) were identified as facets of LaMnO₃ based on JCPDS 50-0298 (Figure 1c) [42]. Additionally, the diffraction peaks at (110) and (111) were associated with facets of TiO₂, as reported in JCPDS 21-1276 [43]. As presented in Text S2 and Table S4, the calculated crystallite sizes for TiO₂-modified LaMnO₃ particles fell within 19.6 to 21.2 nm, with average values of 20.6 nm, respectively. Meanwhile, its corresponding calculated crystallinity was found to be 75.5%. TGA curves (Figure S5) demonstrated that the mass loss of the sample was lower than 1.5% when the decomposition temperature reached 800 °C, highlighting the stability of TiO₂-modified LaMnO₃.

3.2. Degradation Performance and Reusability

As shown in Figure 2a,b, TiO₂-modified LaMnO₃ attained the most efficient IBP removal (98.0%) and the highest ratio of DOC removal (85.9%). The degradation kinetics were modeled using pseudo-first-order reactions, resulting in R² values exceeding 0.97. In the absence of ozone, TiO₂, LaMnO₃, and TiO₂-modified LaMnO₃ exhibited IBP removal efficiencies of 13.2%, 8.05%, and 9.80%, respectively. TiO₂-modified LaMnO₃ achieved significantly higher IBP degradation rate constants (K_obs) of 0.1974 min⁻¹ compared to ozonation, TiO₂, and LaMnO₃. Correspondingly, DOC removal for these processes was 28.5%, 54.8%, and 85.9%, respectively (Figure 2b). Figure S6 illustrates the pH variation during different processes, providing additional evidence that the inadequate mineralization observed during single ozonation was attributed to the accumulation of acid-resistant intermediates impervious to ozone degradation [10]. This illustrates that the catalytic effect of ozone was mainly attributed to LaMnO₃ rather than TiO₂, and the modification of TiO₂ contributed to enhanced adsorption-catalytic performance.

After 5 cycles, the DOC removal decreased from 85.9% to 72.2%, representing a relative reduction of 13.7% (Figure 2c). The overall stability was found to be comparable to previous studies utilizing perovskite in visible light-induced photocatalysis [44] and catalytic ozonation, where removal reductions ranged from 3% to 23% after 5 cycles [7,45]. This stability was attributed to the modification that increased the overall stability of LaMnO₃ by enhancing specific surface features. Consequently, it accelerated the IBP degradation rate in the bulk solution and facilitated the rapid removal of IBP bound to the surface via active sites. The crystallinity and crystal structures remained well-preserved, as depicted in Figure 1c and Table S4, by comparing fresh and used catalysts. After five uses (Figure 2d), the leaching of Mn and La was minimal, measuring less than 0.1 mg/L. Less Mn leaching was observed compared to the reported non-modified perovskite (0.2 mg/L) [7,9,46]. These findings demonstrate that TiO₂-modified LaMnO₃ exhibited exceptional removal efficiency and reusability.

3.3. Effect of the Operational Parameters

As shown in Figure 3a, at a catalyst dosage of 1000 mg/L, IBP was completely removed, with the observed rate constant (K_obs) reaching 0.2189 min⁻¹. Notably, even when the catalyst dosage was increased from 500 to 1000 mg/L, the observed rate constant (K_obs) increased by only 0.02 min⁻¹. This can be attributed to the fixed ozone dosage, causing an excess of the catalyst. The surplus catalyst contributes to escalated treatment costs. Figure 3b illustrates the effect of ozone dosage on IBP removal. With ozone concentration elevated across the range of 10–60 mg/L, K_obs rose from 0.1974 min⁻¹ to 0.8763 min⁻¹. This notable enhancement in K_obs can be linked to more efficient gas–liquid ozone transport [47]. A significant surge in K_obs was observed upon stepping up the ozone input from 40 to 60 mg/L. As the primary oxidant in the catalytic ozonation system, the increasing ozone dosage led to an increase in the generation of secondary oxidants such as reactive oxygen species (ROS). Considering the relative reactivity of ozone (k_O3 = 9.6 M⁻¹ s⁻¹) and hydroxyl radicals (HO•, k_HO• = 7.4 × 10⁹ M⁻¹ s⁻¹) toward IBP, this was likely due to the improvement in excessive generation of ROS [48].

As depicted in Figure 3c, K_obs decreased from 0.3880 min⁻¹ to 0.0832 min⁻¹ with an increase in the initial IBP concentration across 0.5–5 mg/L. Even when the IBP concentration was elevated to 5 mg/L, complete removal could still be achieved within a 40-min treatment, suggesting the enduring robustness of the catalyst even under high initial IBP concentration. As observed in Figure 3d, when the initial pH was increased within the range of 3 to 10, a drastic reduction in K_obs, from 0.7115 to 0.0039 min⁻¹, was consistent with the inhibitory effect of acidic conditions on IBP removal. At initial pH values of 3.0 and 4.2, IBP removal was notably inhibited compared to conditions close to neutral and alkaline. On the one hand, acidic conditions restrain ozone decomposition into ROS. Hydroxide ions promote ROS generation and act as initiators for ozone decomposition. On the other hand, PZC was measured as 5.75 (Figure S7) [49]. When the solution pH was below 4.9, protonation of both the catalyst surface and IBP rendered them mutually repulsive, thereby suppressing their interaction. This limitation reduced the surface reactions between IBP and active sites, further decreasing K_obs. Nevertheless, it is also worth noting that, apart from pH = 3.0, all the other pH conditions favored complete degradation of IBP within 40 min, indicating a wide pH tolerance range of TiO₂-modified LaMnO₃.

3.4. Effect of the Water Matrix

In the presence of 2000 mg/L chloride ions, K_obs decreased from 0.1974 to 0.1063 min⁻¹ (Figure 4a). However, IBP could still be degraded entirely within 30 min. This is consistent with findings in existing studies [50,51]. When chloride ions were introduced into the system, they directly reacted with ozone and HO• [52]. Chloride ions affected ozone dissolution in water, reacted with ozone, competed for absorption onto the catalyst surface active site, and reacted with HO•. Considering the relatively small reaction rate constant between ozone and chloride ions (3.0 × 10⁻³ M⁻¹ s⁻¹), the significant inhibition came from the reaction with HO• and the generation of less reactive chlorine radicals.

When the SO₄²⁻ concentration ranged from 100 to 1000 mg/L, both K_obs and IBP removal remained unaffected (Figure 4b). Even with the SO₄²⁻ concentration increased to 2000 mg/L, IBP could still be completely removed within 30 min, although K_obs slightly decreased to 0.1542 min⁻¹. Excessive SO₄²⁻ tends to adsorb on the catalyst surface, occupying adsorption sites and slowing down the adsorption-catalytic degradation process [25]. Furthermore, HO• can react with K_obs to produce sulfate radicals with a constant reaction rate to IBP.

The addition of HA at concentrations of 0.5, 1, 2, and 5 mg/L led to a steady decline in K_obs from 0.1974 to 0.1768, 0.1546, 0.1076, and 0.0898 min⁻¹ (Figure 4c). The polyfunctional nature of HA, bearing carboxylic and phenolic moieties, imparts a negative surface charge at pH 6.4. This negatively charged species readily competes for the Lewis acid centers on the catalyst surface, accounting for the observed suppression of catalytic activity [53]. Simultaneously, ozone could be depleted as it oxidizes the HAs, thereby diminishing the oxidation efficacy of the process.

3.5. Adsorption-Catalysis Mechanism of O₃ on TiO₂-Modified LaMnO₃ Surface

3.5.1. ROS Identification and Reaction Pathways

Different scavengers exhibit distinct reaction rates with radicals (Table S6), allowing for an indirect deduction of the ROS type in the catalytic system by EPR. A noticeable reduction in K_obs was observed when various radical scavengers were introduced into the catalytic ozone system. TBA and HCO₃⁻ act as scavengers for HO• [54]. Specifically, upon addition of 10 mmol/L TBA and 300 mg/L HCO₃⁻ (Figure 5a), K_obs decreased to 0.0547 and 0.0911 min⁻¹, respectively. This reflects the pivotal contribution of HO• to IBP degradation, wherein the extent of removal mirrors the level of radical generation. Furthermore, adding 10 mmol/L 1-butanol and 0.5 mmol/L p-BQ resulted in a significant decrease in IBP removal and K_obs, providing further evidence that, besides HO•, O₂•⁻ is also involved in IBP removal. Simultaneously, NaN₃ addition suppressed IBP removal in both the ozonation and catalytic ozonation processes at pH 3 (Figure 5b), suggesting that the inhibition arose not only from scavenging ¹O₂ but also HO• [55]. In the solution, the primary action is by ozone, while on the catalyst surface, catalytic ozone reactions generate a substantial amount of ROS, influencing IBP removal and DOC reduction. The dominance of bulk solution reactions in IBP degradation was established using TBA and HCO₃⁻ as radical probes. Meanwhile, O₂•⁻ and ¹O₂ were found to drive mineralization through a combination of HO• generation via chain reactions and direct oxidative destruction.

The intense DMPO-HO• signal (Figure 5c) underscores the superior capacity of the TiO₂-modified LaMnO₃ system for HO• generation during catalytic ozonation. Stable DMPO-O₂•⁻ signals were observed in both LaMnO₃ and TiO₂-modified LaMnO₃ systems (Figure 5d). During the single ozonation process, the absence of DMPO-signals suggested the limited involvement of O₂•⁻ in IBP removal. Figure 5e illustrates the significantly high signal intensity of TMEP-¹O₂ in catalytic ozonation using modified LaMnO₃ compared to single ozonation and catalytic ozonation using non-modified LaMnO₃. This indicates the presence of singlet oxygen in the catalytic system, consistent with the quenching experiment results. Further analysis was conducted to elucidate the ozone decomposition pathways and intermediates on distinct catalyst surfaces. The reaction energy for these pathways is depicted in Figure 5f. The initial step involved the adsorption of O₃ on the catalysts, characterized by an exothermic process with negative ΔG values. This energy release overcame the reaction barrier and reached the transition state. The third step (*O + O₂) transitioning to the fourth step (*O + O₃) was the rate-limiting step. In the final step, the energy barrier for TiO₂-modified LaMnO₃ was significantly minimal, implying that the composite catalyst facilitated the desorption of O₂²⁻. The significant enhancement of key steps significantly improved adsorption-catalysis efficiency, thereby promoting the removal of IBP and DOC.

3.5.2. Redox Pair of Mn³⁺/Mn⁴⁺

As shown in Figure 6a and Figure S8, lattice oxygen signals diminished, whereas hydroxyl and adsorbed water features intensified, suggesting that oxygen vacancies are abundant in the TiO₂-modified LaMnO₃ and play a key role in ozone decomposition. The increase in adsorbed H₂O and hydroxyl on TiO₂-modified LaMnO₃ indicated the occurrence of redox reactions on the catalyst surface [42]. The difference in binding energy between Ti 2p _3/2 and 2p _1/2 difference was measured at 5.6 eV (Figure 6b), confirming the presence of TiO₂ in the catalyst both before and after use [56]. However, the relative area of Ti 2p _3/2 and 2p _1/2 remained unchanged after the reaction. The spin–orbit splitting of the Mn 2p level was consistently 11.4 eV for both fresh and used TiO₂-modified LaMnO₃ (Figure 6c). Two spin–orbit doublets were resolved from the Mn 2p spectra, corresponding to Mn³⁺ (642.1 eV and 653.5 eV) and Mn⁴⁺ (642.9 eV and 654.4 eV). A reduction in Mn³⁺ content from 79.7% to 62.9% was observed after the reaction, coupled with an increase in Mn⁴⁺ from 20.3% to 37.1%. The Mn³⁺/Mn⁴⁺ ratio in TiO₂-modified LaMnO₃ was 3.9, in stark contrast to the ratios of 1.2 and 2.2 typically observed for nanocast LaMnO₃ and LaMn₄O_x, respectively [7,10]. Regarding La 3d binding energies, the difference between La 3d _5/2 and 3d _3/2 was 16.8 eV (Figure 6d), confirming the presence of La³⁺ in both fresh and used TiO₂-modified LaMnO₃ [57]. However, there was only a slight change after the reaction. In summary, Ti and La did not actively participate in the catalytic reaction, indicating no catalytic performance from these elements. On the other hand, Mn³⁺/Mn⁴⁺ was key to the catalytic reaction process.

3.5.3. Bonding, Orbitals, Charge, and Interaction Analysis

In TiO₂-modified LaMnO₃, LaMnO₃ played a pivotal role in adsorption-catalysis processes. We focused on elucidating the interaction between O₃ and LaMnO₃ while assessing the impacts of different components on adsorption forces. We devised three distinct adsorption configurations for O₃ on the 110, 102, and 010 crystal facets (Figure S9). The calculated adsorption energy (E_ads) suggests that O₃ exhibited the highest favorability for adsorption on the 010 crystal facet (Figure S10), with an E_ads value of −2.24 eV. Significantly, excessively high or low E_ads could adversely affect the catalytic process [58]. A more negative adsorption energy reflects greater instability of the adsorption complex, which hampers the overall catalytic performance [59]. In the optimized catalyst structure (Figure S11), Mn and La atoms constitute the primary adsorption-catalysis centers, with O atoms from the O₃ molecule forming bonds with these sites at bond lengths ranging from 1.954 to 2.461 Å. Following adsorption, the O-O bonds within the O₃ molecule elongated to 1.365 and 1.511 Å, greater than the original bond length of 1.284 Å. This observation aligned with the catalytic mechanism of O₃, as adsorption on the (010) face facilitated the breaking of O-O bonds.

As the d-band center (ɛ_d) was closely linked to catalytic activity, variations between Mn and La atoms resulted in distinct reactivities towards O₃. The calculated ɛ_d, within a range spanning from −0.289 to −1.289 eV, was in proximity to the Fermi energy (E_f, −2.830 eV), thereby reinforcing the interaction between the oxygen and metal (La, Mn) centers. Notably, the ɛ_d of the Mn atom was closer to E_f than that of the La atom, pointing to preferential O₃ adsorption at the Mn site. Spin polarization of the M 3d orbitals shaped the spin-dependent DOS of the M-O bonds (Figure 7). The up-spin DOS was dominated by O 2p and M 3d orbitals, whereas the down-spin DOS arose primarily from O 2s 2p and M 3d contributions. The DOS near E_f reflects a hybrid character stemming from the coupling of O 2p and M 3d orbitals, rather than pure M 3d states, with this coupling arising from a 5σ bond formed between the two orbitals. The 2px and 2py orbitals contributed to forming a 1π bond, while the 2s and 2py orbitals were responsible for the 3σ bond. The 2π* antibonding state underwent a dispersion effect, whereas the 4π* antibonding energy exhibited a slight decline, a behavior stemming from the interaction between the 2π bond of O and the σ bond of M atoms. The coefficients of the orbital combination in the partial density of states (PDOS) exhibited notable differences, with electrons tending to be localized on the O side rather than being shared equally between O and M atoms. Particularly significant were the interactions between the 3dxy, 3dxz, and 3d2z orbitals of M atoms with the 2pz orbital of the O atom. Negative values in the COOP curves represent antibonding interactions, while positive values signify bonding interactions. These COOP curves showed antibonding characteristics at higher energies levels and bonding characteristics in the low-energy region. The COOP curves traversed through the E_f (Figure S12), indicating a transition of interactions from antibonding to bonding.

The ELF maps (Figure 8a,b) offered additional insights into the charge density overlaps. Following adsorption, electron density within LaMnO₃ shifted toward the adsorbed O atoms, disrupting the initially homogeneous ELF field. Stronger ELF intensity at the Mn-O interface relative to La-O suggests more robust electronic coupling at Mn sites [60]. Bader charges (Figure 8c) exemplified that the charge transfer occurred from the surface M atoms to the adsorbed O atoms. Additionally, the adsorbed O atoms expedited the transfer of internal electrons to the vicinity around them, leading to alterations in the charge distribution. Compared to La atoms, the greater charge transfer suggested that exposed Mn atoms predominantly served as mediators for bonding the adsorbed O atoms. The charge density difference maps (Figure 8d) further underscored the increase in charge density surrounding the adsorbed O atoms, accompanied by a decrease in charge density around the corresponding M atoms. For a more in-depth examination of the interaction between O₃ and the catalyst surface, we conducted a reduced density gradient (RDG) analysis, as illustrated in Figure S13 and Figure 8e,f. This analysis involved the evaluation of Sign(λ₂)ρ values and electron density values at the bond critical point (BCP) for various systems, which were summarized in Table S7. RDG analysis stands as a potent method for characterizing chemical bonding interactions [20]. The values surrounding the reference point indicated pronounced attractive interactions (represented in blue), and this heightened attraction can be ascribed to O of O₃ positioned between the M atom of the catalyst. Furthermore, the presence of a BCP at the electron density level served as additional confirmation for the existence of the M-O bond. Moderate reduction in interaction forces proves advantageous for improving the adsorption-catalysis process of O₃ on the surface of TiO₂-modified LaMnO₃. Simultaneously, prompt detachment facilitates the swift cycling of Mn³⁺/Mn⁴⁺, resulting in the increased production of ROS, which can be effectively utilized for IBP removal.

4. Conclusions

In summary, the synthesized TiO₂-modified LaMnO₃ catalyst displayed exceptional efficiency in the catalytic ozonation process for removing IBP under various operational parameters and water matrices (with DOC removal increasing from 28.5% to 85.9%). Experimental evidence, including ESR and quenching experiments, unequivocally confirmed the greater production of ROS by the TiO₂-modified LaMnO₃, which was responsible for its superior IBP removal compared to the non-modified LaMnO₃. Furthermore, the Mn³⁺/Mn⁴⁺ ratio increased to 3.9, contributing significantly to ozone decomposition. DFT simulations offered valuable insights into the enhanced ROS generation, primarily attributed to the activation of surface-active Mn sites. A moderate reduction in interaction forces has proven advantageous for improving the adsorption-catalysis process of O₃ over TiO₂-modified LaMnO₃. Rapid detachment facilitated the swift cycling of Mn³⁺/Mn⁴⁺, resulting in an increased production of ROS, which can be effectively harnessed for IBP removal. This study deepens our understanding of electron behaviors and bond dynamics during ozone interaction with perovskite surface metal ions and underscores the significance of structural modifications in perovskite materials to enhance ROS generation in catalytic ozonation. Overall, this catalyst emerges as a promising candidate for advanced water treatment applications, effectively eliminating IBP from diverse and complex water sources.