Hollow SiO₂/MnO₂ structures for efficient and stable degradation of phenol and tetracycline under low-oxidant conditions

Abstract

Although MnO₂-based materials have been widely employed for the degradation of organic contaminants, their efficacy typically hinges on elevated oxidant concentrations, often as high as 2.0 g L⁻¹. In this study, we engineered a sophisticated hollow SiO₂/MnO₂ composite, wherein the MnO₂ outer layer orchestrates rapid and efficient pollutant degradation through peroxymonosulfate (PMS) activation, while the SiO₂ inner core acts as a robust scaffold, enhancing MnO₂ dispersion and preserving structural stability. The hollow, porous morphology of the SiO₂/MnO₂ composite optimizes mass transfer and strengthens catalyst-reactant interactions by minimizing diffusion limitations, thereby markedly elevating catalytic efficiency. Consequently, this hollow SiO₂/MnO₂ catalyst outperforms commercial MnO₂, delivering an eightfold enhancement in phenol degradation rate at a reduced PMS concentration of 0.5 g L⁻¹, alongside a threefold increase in tetracycline degradation efficiency. Liquid chromatography-mass spectrometry (LC-MS) analysis further unraveled the degradation mechanisms of phenol and tetracycline, pinpointing critical intermediates. Collectively, this work presents a pragmatic and highly effective approach to curtailing oxidant reliance in MnO₂-based systems, offering significant implications for the design of next-generation materials tailored for organic pollutant abatement.

---

1. Introduction

Water pollution has become an increasingly pressing global concern, primarily due to the persistence and bioaccumulation of residual organic compounds, which pose severe threats to both human health and environmental systems.^1,2 Even at trace concentrations, these pollutants exhibit significant carcinogenicity and toxicity, leading to long-term ecological risks and potential human exposure hazards.^3–5 Consequently, there is a growing imperative to develop efficient and economically viable technologies for the removal of such contaminants from wastewater, particularly in addressing refractory organic pollutants like phenol and tetracycline. Among the various strategies, advanced oxidation processes (AOPs) have garnered considerable attention owing to their exceptional capacity to decompose a broad spectrum of organic compounds.^6,7 AOPs operate by generating highly reactive species, such as hydroxyl and sulfate radicals, which effectively break down persistent pollutants into harmless end-products.^8,9

Among various oxidants, peroxymonosulfate (PMS) has attracted considerable attention due to its ability to generate sulfate radicals (SO₄˙⁻), which are highly reactive oxidizing species capable of efficiently degrading a wide range of organic pollutants.¹⁰ Enhancing the activation of PMS has therefore become a central focus in advancing advanced oxidation processes. Among the reported catalysts, manganese dioxide (MnO₂) stands out owing to its low toxicity, strong oxidation potential, and remarkable ability to activate PMS for sulfate radical production, facilitating the effective degradation of phenolic and other refractory compounds.¹¹ Extensive research has been devoted to tailoring the morphology and structure of MnO₂-based catalysts to improve their catalytic performance.^12,13 Various architectures, including three-dimensional hierarchical structures, one-dimensional α-MnO₂ nanostructures, and Fe₃O₄–MnO₂ core–shell composites, have demonstrated promising efficiency in phenol degradation.^14–16 However, a common limitation among these systems is their reliance on relatively high concentrations of PMS (typically ranging from 2.0 to 4.0 g L⁻¹) to achieve satisfactory pollutant removal, which increases operational costs and may raise environmental concerns.^2,9,14,15,17 Additionally, the intrinsically low surface areas (<66.0 m² g⁻¹) and non-porous nature of many MnO₂ structures restrict the exposure of active sites and limit mass transfer, thereby constraining their catalytic efficiency. To address the inherent limitations of conventional MnO₂-based catalysts—such as low surface area, limited active site accessibility, and poor structural stability—the strategic design of hollow and porous architectures has emerged as a promising approach to enhance catalytic performance in advanced oxidation processes (AOPs). Among various supporting materials, hollow silica (SiO₂) stands out due to its intrinsic properties, including chemical inertness, thermal stability, environmental compatibility, and ease of surface functionalization.^18,19 The unique structural characteristics of hollow SiO₂, such as high specific surface area, tunable porosity, and internal voids, not only facilitate uniform catalyst dispersion but also enhance mass transfer and reactant adsorption, thereby improving overall catalytic efficiency. For instance, MnO₂ uniformly anchored on SiO₂ nanofibers demonstrated excellent photo-thermo catalytic activity under simulated sunlight, achieving complete mineralization of toluene.²⁰ This performance was attributed to the synergistic effect of tortuous porous networks and interfacial photoactivation,²¹ which promoted reactive oxygen species (ROS) generation and pollutant diffusion. Similarly, MnO₂@nano-hollow carbon spheres (MnO₂@NHCS) achieved rapid removal of bisphenol A (95.3% within 10 min), further confirming that hollow confinement enhances active site exposure and accelerates ROS-mediated degradation kinetics. The structural tunability of SiO₂/MnO₂ composites also supports multifunctional applications. For example, SiO₂@MnO₂ nanocomposites synthesized via an ultrasonic method exhibited excellent dispersion and stability, enabling their integration into fluorescence quenching platforms for simultaneous detection of glutathione and antibiotic residues in complex matrices.²² These findings underscore the versatility and adaptability of hollow SiO₂-based architectures across environmental and biochemical domains. Collectively, these features elevate catalytic efficiency, making hollow SiO₂/MnO₂ composites a promising platform for advanced pollutant degradation.

In this study, we propose the design of a hollow SiO₂/MnO₂ composite catalyst, wherein MnO₂ serves as the active outer shell for PMS activation, and SiO₂ functions as a structural support to enhance dispersion and stability. The incorporation of a hollow and porous architecture is expected to facilitate reactant diffusion, increase active site exposure, and alleviate mass transfer limitations—factors that are critical for efficient catalytic oxidation. This study focuses on the rational construction of such a hybrid structure and explores its potential in activating PMS for the degradation of representative refractory pollutants, such as phenol and tetracycline. Furthermore, the underlying reaction mechanism and associated transformation pathways are systematically investigated to gain deeper insight into the catalytic behavior of the SiO₂/MnO₂ system.

2. Experimental and simulation section

2.1 Materials

Tetraethoxysilane (TEOS, ≥98%) and 2-(methacryloyloxy)ethyltrimethylammonium chloride (DMC, 75 wt% in water) were purchased from J&K Scientific Ltd (China) and used as received without further purification. Manganese(II) nitrate tetrahydrate (Mn(NO₃)₂·4H₂O, AR grade), sodium borohydride (NaBH₄, ≥96%), styrene (≥99%), and potassium persulfate (K₂S₂O₈, ≥99%) were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Prior to use, styrene was purified by washing with 5 wt% sodium hydroxide solution to remove inhibitors, followed by water rinsing and drying over anhydrous calcium chloride. All other reagents were used as received without further treatment, and deionized water was used throughout all synthesis procedures.

2.2 Preparation of cationic polystyrene spheres (CPS)

Cationic polystyrene (CPS) spheres were synthesized via free-radical emulsion polymerization following a previously reported procedure with slight modifications.^20–22 Briefly, styrene monomer (7.5 g) was added to deionized water (72 mL) in a 100 mL four-necked round-bottom flask equipped with a mechanical stirrer, a condenser, and a nitrogen inlet. After purging with N₂ for 30 minutes to remove dissolved oxygen, potassium persulfate (K₂S₂O₈, ∼0.16 g) was introduced as an initiator. The mixture was stirred at 350 rpm and maintained at 70 °C under a nitrogen atmosphere. Subsequently, 2-(methacryloyloxy)ethyltrimethylammonium chloride (DMC, 1.1 g) was added dropwise over 30 minutes using a syringe pump to ensure uniform functionalization. The reaction was allowed to proceed for 6 hours under continuous stirring. The resulting CPS product was collected by centrifugation at 8000 rpm for 10 minutes, washed thoroughly with deionized water several times to remove unreacted monomers and initiators, and then dried at 50 °C under vacuum for further use.

2.3 Preparation of hollow SiO₂ spheres (HS)

Hollow silica (SiO₂) spheres were synthesized using the CPS particles as a removable template. In a typical procedure, the as-prepared cationic polystyrene spheres (CPS, 2.0 g) were dispersed in 50 mL of absolute ethanol under magnetic stirring at 350 rpm. Tetraethyl orthosilicate (TEOS, 1.0 g) was then rapidly added to the dispersion. Subsequently, aqueous ammonia solution (NH₃·H₂O, 28–30 wt%) was introduced dropwise to initiate the hydrolysis and condensation of TEOS. The reaction mixture was maintained at 50 °C for 12 hours to allow for the electrostatic self-assembly of SiO₂ onto the CPS surface, forming CPS@SiO₂ core–shell structures. The resulting CPS@SiO₂ microspheres were collected by centrifugation, washed thoroughly with ethanol and deionized water, and dried at 60 °C. Finally, the dried samples were calcined in air at 450 °C for 2 hours (heating rate: 2 °C min⁻¹) to remove the polymer core, yielding hollow SiO₂ spheres (HS).

2.4 Preparation of SiO₂/MnO₂ (HSM) spheres

Hollow SiO₂ spheres (1.0 g) were dispersed in 50 mL of absolute ethanol under continuous magnetic stirring. A solution of manganese(II) nitrate tetrahydrate (Mn(NO₃)₂·4H₂O, 2.0 g) dissolved in 20 mL of deionized water was slowly added dropwise to the suspension via a syringe to ensure uniform adsorption of Mn species onto the silica surface. The mixture was stirred at room temperature for 12 hours to allow adequate precursor deposition. Subsequently, the solvent was evaporated by heating the mixture at 70 °C in a water bath, following a previously reported method with slight modification.²³ The resulting precursor was dried overnight at 60 °C and then calcined in air at 450 °C for 2 hours (heating rate: 2 °C min⁻¹) to obtain the final SiO₂/MnO₂ (HSM) hollow composite spheres.

2.5 Characterization

The surface morphology and internal structure of the SiO₂ and SiO₂/MnO₂ (HSM) spheres were characterized using field emission scanning electron microscopy (FE-SEM, JEOL JSM-7001F, Japan) at 10 kV and high-resolution transmission electron microscopy (HR-TEM, JEOL JEM-2100, Japan) at 200 kV. The specific surface area and pore size distribution were assessed via nitrogen adsorption–desorption isotherms on a Micromeritics ASAP 3020 II analyzer, with surface area calculated by the Brunauer–Emmett–Teller (BET) method and pore size distribution derived from the Barrett–Joyner–Halenda (BJH) model. The crystalline structure of the HSM composite was determined by X-ray diffraction (XRD) on a Bruker D8 Advance diffractometer (Germany) using Cu-Kα radiation (λ = 1.5406 Å) at 40 kV and 40 mA, scanning from 20° to 80° 2θ. The elemental composition and chemical states of the HSM spheres were investigated using X-ray photoelectron spectroscopy (XPS, Thermo Scientific ESCALAB 250Xi, USA) with Al-Kα radiation (hv = 1486.6 eV), calibrated to the C 1s peak at 284.8 eV. The mass ratio of Mn in the sample was measured by inductively coupled plasma-optical emission spectroscopy (ICP-OES). Additionally, a trace TOC instrument was used to obtain total organic carbon (TOC) for determine the mineralization of the materials.

2.6 Catalytic activity evaluation

The catalytic performance of the SiO₂/MnO₂ (HSM) composite was evaluated for the degradation of phenol and tetracycline (TC), using commercial MnO₂ and pure SiO₂ as reference catalysts. Note that the commercial MnO₂ (CM, ≥99%, CAS No. 1313-13-9) was purchased from Shanghai Macklin Biochemical Co., Ltd. In a typical experiment, 20 mg of catalyst was added to 50 mL of aqueous phenol solution (5 mg L⁻¹), and the mixture was stirred at 25 °C for 30 minutes to establish adsorption–desorption equilibrium. Subsequently, 0.025 g of peroxymonosulfate (PMS), corresponding to a concentration of 0.5 g L⁻¹, was added to initiate the oxidation reaction. tert-Butanol (TBA), methanol (MeOH), L-histidine (L-his) and p-benzoquinone (PBQ) have been chosed as the scavangers fo the radical-trapping experiment. At specified time intervals, 5 mL of the reaction mixture was withdrawn using a syringe, filtered through a 0.22 µm membrane, and analyzed by UV-vis spectroscopy at a wavelength of 270 nm.¹⁰ All experiments were conducted in triplicate, and the reported results represent the average values of three independent runs. For the recyclability test, the used HSM catalyst was recovered after each cycle by centrifugation, thoroughly washed with deionized water, and dried at 60 °C for 12 hours before reuse in the subsequent run.^24–27 No additional activation treatment was applied between cycles.

3. Results and discussion

3.1 Preparation and structural characterization of SiO₂/MnO₂ (HSM) spheres

The synthesis strategy of the hollow SiO₂/MnO₂ (HSM) spheres is illustrated in Fig. 1a. Initially, cationic polystyrene (CPS) microspheres were prepared via emulsion polymerization and subsequently coated with a silica layer through the hydrolysis and condensation of TEOS, forming core–shell CPS@SiO₂ structures. Calcination at 450 °C removed the CPS template, yielding hollow SiO₂ (HS) spheres. Finally, manganese ions were adsorbed onto the HS surface, and the material was subjected to thermal treatment at 450 °C to form the final SiO₂/MnO₂ (HSM) hollow composite. Field emission scanning electron microscopy (FE-SEM) images (Fig. 1b and c) show that both HS and HSM samples possess uniform spherical morphology with smooth and intact surfaces. Compared with the pristine HS spheres (Fig. 1b), the HSM samples exhibit a roughened surface decorated with numerous small nanoparticles (Fig. 1c), suggesting the successful deposition of MnO₂. This structural evolution is further confirmed by high-resolution transmission electron microscopy (HR-TEM, Fig. 1d), which clearly reveals the hollow interior and surface-nanoparticle features of the HSM microspheres.

Fig. 1 (a) Schematic illustration of the synthesis process for the hollow SiO₂/MnO₂ (HSM) composite; (b) SEM image of hollow SiO₂ spheres showing uniform morphology; (c) SEM image of HSM spheres revealing surface roughness due to MnO₂ deposition; (d) TEM image confirming the hollow interior and surface nanoparticle distribution of HSM; (e) high-resolution TEM (HR-TEM) image highlighting the crystalline MnO₂ domains anchored on the SiO₂ shell. For full experimental conditions, see Section 2. Experimental and simulation.

Moreover, the HSM spheres maintain their spherical integrity after Mn loading, indicating that the deposition process does not compromise the overall structure. High-magnification TEM images (Fig. 1e) show distinct contrast between the silica framework and MnO₂ nanoparticles, supporting the formation of a well-dispersed MnO₂ outer layer. Lattice fringes with an interplanar spacing of 0.31 nm were observed, corresponding to the (hkl) planes of MnO₂, further confirming its crystalline nature. The presence of manganese species on the surface was further validated by energy-dispersive X-ray spectroscopy (EDX), confirming the successful incorporation of Mn into the composite structure.

X-ray diffraction (XRD) patterns of the SiO₂/MnO₂ (HSM) composite, as depicted in Fig. 2, reveal critical insights into its structural composition. Four prominent diffraction peaks are observed at 2θ values of 37.12° (100), 42.40° (101), 56.03° (102), and 66.76° (110), which align precisely with the tetragonal phase of MnO₂ (JCPDS No. 30-0820).²⁹ This correspondence unequivocally confirms the successful integration of crystalline MnO₂ onto the SiO₂ surface, underscoring the efficacy of the synthesis approach in achieving a well-defined catalytic layer. Additionally, a broad amorphous hump spanning the 2θ range of 15°–30° is evident, a signature feature attributable to amorphous silica.³⁰ This dual-phase characteristic highlights the composite's hybrid nature, blending the ordered crystallinity of MnO₂ with the disordered, yet structurally supportive, amorphous SiO₂ framework.

Fig. 2 X-ray diffraction (XRD) pattern of the HSM composite. The characteristic diffraction peaks confirm the presence of crystalline MnO₂ and amorphous SiO₂. For full experimental conditions, see Section 2. Experimental and simulation.

The porous architecture of the SiO₂/MnO₂ (HSM) composite was comprehensively characterized through nitrogen adsorption–desorption analysis, as illustrated in Fig. 3. The resulting isotherm displays a classic type IV profile accompanied by a distinct hysteresis loop, a hallmark of mesoporous materials. The Brunauer–Emmett–Teller (BET) method yielded a specific surface area of approximately 22.31 m² g⁻¹ and 74.78 m² g⁻¹ for the MnO₂ and HSM composite, respectively. It is evident that the HSM composite has a value that notably surpasses those of numerous MnO₂-based catalysts documented in prior studies.^9,11 Complementary Barrett–Joyner–Halenda (BJH) analysis disclosed a predominant pore size distribution ranging from 5 to 10 nm, corroborating the mesoporous framework templated by the hollow SiO₂ structure with the specific surface area of ∼156.48 m² g⁻¹. We also evaluated the pore-size distribution by the NLDFT method; the results are shown in Fig. S1. Both the BJH and NLDFT curves exhibit virtually identical hierarchical meso-/macroporous features, confirming that the conclusions derived from the BJH analysis are reliable. This elevated surface area, coupled with a well-defined porosity, significantly enhances the diffusion efficiency of reactant molecules and maximizes the exposure of catalytic active sites—attributes pivotal to accelerating oxidative degradation processes. These structural features play a key role in boosting the degradation efficiency of organic pollutants such as phenol and tetracycline.

Fig. 3 Nitrogen adsorption–desorption isotherms and corresponding pore size distribution curves of (a) HSM and (b) hollow SiO₂ spheres. The results demonstrate the mesoporous structure and enhanced surface area of the HSM composite. For full experimental conditions, see Section 2. Experimental and simulation.

Recent literature further contextualizes these findings within the evolving landscape of porous catalyst design. For instance, Wei et al.²³ synthesized hollow mesoporous silica nanoreactors encapsulating manganese oxide nanoparticles (MnxOy@HMSNs), which exhibited greatly enhanced dye degradation efficiency. The performance was attributed to the synergistic effect of the mesoporous silica shell, providing diffusion channels and structural protection, and the ultrasmall MnO_x nanoparticles, which offered abundant accessible active sites—mechanistic principles that closely parallel our HSM system. Similarly, a recent study demonstrated that hollow-structured mesoporous silica spheres loaded with bimetallic Au–Pt catalysts exhibited significantly enhanced catalytic activity due to the synergistic effects of the engineered hollow cavity and porous shell, which collectively facilitated reactant transport and increased catalytic efficiency—findings that resonate with the structural advantages observed in our HSM composite system.²⁴

3.2 Crystal structure and bonding environment of HSM spheres

X-ray photoelectron spectroscopy (XPS) was employed to investigate the surface elemental composition and chemical states of the SiO₂/MnO₂ (HSM) composite, providing insights into the interaction between MnO₂ and the SiO₂ support. The survey spectrum (Fig. 4a) confirms the presence of C 1s, Si 2p, O 1s, and Mn 2p peaks, indicating that carbon, silicon, oxygen, and manganese elements are successfully incorporated into the HSM structure. The O 1s spectrum shows two components at ∼530.1 eV and ∼531.5 eV, which can be attributed to lattice oxygen and surface hydroxyl groups, respectively, further supporting the existence of MnO₂ and adsorbed oxygen species. The high-resolution Mn 2p spectrum (Fig. 4b) reveals two characteristic peaks located at binding energies of 643.1 eV and 654.2 eV, corresponding to Mn 2p_3/2 and Mn 2p_1/2, respectively. The spin–orbit splitting between the two peaks is approximately 11.0 eV, which is in good agreement with the reported values for MnO₂.³¹ Notably, the Mn 2p_3/2 peak position is consistent with the Mn⁴⁺ oxidation state, suggesting that manganese predominantly exists in the form of MnO₂ in the HSM composite. These results confirm the successful incorporation of Mn⁴⁺ species onto the SiO₂ framework and validate that MnO₂ is the dominant manganese phase, which plays a crucial role in PMS activation during catalytic oxidation reactions.

Fig. 4 (a) X-ray photoelectron spectroscopy (XPS) survey spectrum of the SiO₂/MnO₂ composite, confirming the presence of C, O, Si, and Mn elements; (b) high-resolution XPS spectrum of Mn 2p showing spin–orbit splitting and confirming the Mn⁴⁺ oxidation state. For full experimental conditions, see Section 2. Experimental and simulation.

3.3 Catalytic activity of HSM spheres

The catalytic activity of the hollow SiO₂/MnO₂ (HSM) spheres was evaluated using phenol as a model contaminant, with hollow SiO₂ (HS) and commercial MnO₂ (CM) serving as reference catalysts. As shown in Fig. 5a, phenol degradation is negligible in the presence of PMS alone, indicating that thermal activation of PMS is insufficient to generate reactive sulfate radicals under these conditions.⁴ Similarly, the addition of HS to the PMS system does not enhance phenol removal, further confirming that HS lacks the ability to activate PMS effectively. In contrast, the CM/PMS system exhibits a modest improvement in phenol degradation (∼18%), attributed to the limited catalytic ability of bulk MnO₂. Remarkably, the HSM/PMS system achieves a phenol removal efficiency of approximately 80% under identical conditions, representing a fourfold increase over CM. This enhanced catalytic activity can be ascribed to the synergistic effects of the high surface area and the hollow/porous architecture of HSM, which facilitate efficient PMS activation and increase the accessibility of active sites. Control experiments confirmed that both PMS and catalyst are required to initiate effective degradation.

Fig. 5 (a) Comparative catalytic performance of different materials (HSM, CM, HS) for phenol degradation in the presence of PMS; (b) kinetic fitting curves for phenol degradation over various catalysts, based on pseudo-first-order kinetics. For full experimental conditions, see Section 2. Experimental and simulation.

The degradation kinetics of phenol were further analyzed using the pseudo-first-order Langmuir–Hinshelwood model, described by the equation: −ln(C/C₀) = kt, where C₀ and C are the initial and time-dependent concentrations of phenol, respectively, and k is the apparent rate constant. As illustrated in Fig. 5b, the reaction rate constant for the HSM catalyst is calculated to be 0.0337 min⁻¹. This surpasses values reported for various structured MnO₂ materials. For instance, Huang et al.²⁵ observed pseudo-first-order rate constants ranging from 0.009 to 0.026 min⁻¹ for bisphenol A degradation via PMS activation, depending on MnO₂ crystallinity and oxidation state. Their findings highlighted that higher Mn(III) content, surface conductivity, and tunnel structures significantly enhanced catalytic efficiency—trends that align with the superior performance of the HSM catalyst in this study. Consistent catalytic trends were also observed in tetracycline degradation (Section 3.4), further underscoring the versatility and robustness of the HSM composite in addressing diverse organic pollutants.

To further evaluate the catalytic performance of the HSM catalyst under varying operational conditions, a series of batch experiments were conducted by altering the concentrations of the oxidant (PMS, 0.2–4.0 g L⁻¹) and the pollutant (phenol, 5–25 mg L⁻¹), as shown in Fig. 6a. The results reveal that both insufficient and excessive amounts of PMS are detrimental to the degradation efficiency. Notably, the optimal phenol removal was achieved at a PMS concentration of approximately 0.5 g L⁻¹, which is significantly lower than that typically required for MnO₂-based catalytic systems reported in the literature.^9–11 This demonstrates the superior PMS utilization efficiency of the HSM composite. At low PMS concentrations, the limited generation of reactive sulfate radicals (SO₄˙⁻) restricts the oxidation process. Conversely, an excessive amount of PMS can lead to radical–radical recombination or self-scavenging reactions, which reduces the availability of active species for phenol degradation. The decline at higher PMS concentrations may also be attributed to the excessive accumulation of SO₅²⁻ or HSO₅⁻, which act as radical scavengers. Here an additional comparison table summarizing several recent MnO₂-based strategies (structured MnO₂ or MnO₂ on structured supports, e.g., SiO₂) for PMS activation toward phenol or tetracycline degradation is shown in Table S2. By contrast, our hollow SiO₂/MnO₂ (HSM) achieves ∼80% phenol removal at a low PMS dosage of 0.5 g L⁻¹, and 91.2% tetracycline removal at the same PMS level—demonstrating efficient PMS utilization at reduced oxidant input.

Fig. 6 (a) Effect of PMS concentration (0.2–4.0 g L⁻¹) on phenol degradation efficiency using the HSM catalyst; (b) degradation performance of HSM in phenol solutions with varying initial concentrations (5–25 mg L⁻¹). The results highlight the optimal PMS dosage and the catalyst's tolerance to high pollutant loadings. For full experimental conditions, see Section 2. Experimental and simulation.

In addition, the effect of the initial phenol concentration on catalytic performance was investigated (Fig. 6b). As expected, an increase in phenol concentration from 5 to 25 mg L⁻¹ resulted in a slight decline in degradation efficiency, likely due to active site saturation and mass transfer limitations. Nevertheless, the HSM catalyst maintained a high degradation rate of 76.2% even at 25 mg L⁻¹ phenol, indicating excellent catalytic robustness and high tolerance toward elevated pollutant loads. TOC results reveal the HMS catalyst achieved 49.5% mineralization of phenol. These results collectively demonstrate that the HSM catalyst not only operates effectively under low oxidant conditions but also retains high degradation efficiency across a wide concentration range of phenolic contaminants, highlighting its potential for practical wastewater treatment applications. A similar concentration-dependent trend was observed in the degradation of tetracycline, as discussed in the following section. This concentration resilience aligns with recent advancements in catalytic material design. The HSM catalyst's ability to perform effectively under varying pollutant concentrations not only enhances its practical utility but also distinguishes it within the evolving landscape of advanced oxidation technologies for environmental remediation.

3.4 Catalytic degradation of tetracycline (TC)

To further assess the catalytic performance and universality of the HSM composite, tetracycline (TC) was selected as a representative antibiotic pollutant. As shown in Fig. 7a, all tested materials—HSM, CM, and HS—exhibited a certain degree of TC adsorption within the initial 30 minutes, likely due to surface interactions. In the absence of any catalyst, PMS alone led to moderate TC degradation, achieving a removal efficiency of 53.1% after 140 minutes. A comparable result was obtained for the HS/PMS system, indicating that hollow SiO₂ alone does not contribute to PMS activation and thus has negligible catalytic activity in TC degradation. In contrast, the CM/PMS system achieved a significantly higher degradation rate of 77.0%, attributed to the intrinsic redox activity of commercial MnO₂. Notably, the HSM/PMS system exhibited the highest TC removal efficiency, reaching 91.2% within 140 minutes. TOC results reveal the HMS catalyst achieved 61.2% mineralization of TC. This superior performance highlights the synergistic effect between the high surface area, mesoporosity, and the enhanced dispersion of MnO₂ on the SiO₂ support, which collectively promote efficient PMS activation and reactive species generation.

Fig. 7 (a) Comparative catalytic performance of HSM, CM, and HS toward tetracycline (TC) removal in the presence of PMS (0.5 g L⁻¹); (b) effect of PMS concentration on TC degradation efficiency using the HSM catalyst. The HSM demonstrates high catalytic activity at low oxidant dosage. For full experimental conditions, see Section 2. Experimental and simulation.

The impact of PMS dosage on TC degradation by the HSM catalyst was further explored, as depicted in Fig. 7b. The results reveal a pronounced dependence on oxidant concentration, with optimal TC degradation achieved at a PMS dosage of 0.025 g (equivalent to 0.5 g L⁻¹), balancing efficiency and resource utilization. Higher PMS levels failed to yield proportional enhancements, likely due to radical scavenging or reactive species saturation, as also supported by Lin et al.,²⁶ who demonstrated that low PMS concentrations can yield high degradation efficiency when using reactive site-rich catalysts like FSBC800. For instance, Wang et al.²⁷ demonstrated that increasing PMS concentration beyond an optimal threshold (6 mM) led to a decline in TC degradation efficiency in O–C₃N₄-based systems, attributing this to free radical quenching—where excessive SO₄˙⁻ species undergo self-reaction or recombination, reducing the availability of active species for pollutant oxidation. Recent work by Liu et al.²⁸ on MnO@Co/C catalysts also reported that increasing PMS dosage beyond an optimal threshold failed to proportionally enhance bisphenol A degradation, likely due to radical self-quenching and electron transfer saturation effects, highlighting the necessity of PMS optimization in advanced oxidation systems. These findings demonstrate that the HSM composite achieves highly efficient degradation of both phenol and tetracycline at a relatively low PMS concentration, underscoring its practical viability for advanced oxidation processes (AOPs) under environmentally sustainable and economically viable conditions. This dual-pollutant efficacy positions HSM as a robust contender in the evolving landscape of catalytic wastewater treatment technologies.

To further assess the catalytic behavior and environmental adaptability of the HSM catalyst, its degradation performance toward tetracycline (TC) was evaluated under varying conditions, including temperature, pH, and the presence of coexisting ions (Fig. 8). As shown in Fig. 8a, the HSM catalyst exhibited the highest degradation efficiency at 25 °C, indicating that ambient temperature favors PMS activation and radical generation. A slight reduction in activity was observed at elevated temperatures, which could be attributed to the thermal instability of reactive radicals such as SO₄˙⁻ and ˙OH, or increased rates of radical recombination. Fig. 8b illustrates the impact of initial pH on catalytic performance. The HSM system displayed effective TC removal under both acidic and alkaline conditions, with enhanced performance observed at pH 3 and pH 9. This behavior can be ascribed to the increased production of sulfate radicals (SO₄˙⁻) under extreme pH conditions, which accelerates oxidative degradation. The catalyst's broad pH applicability highlights its suitability for treating diverse wastewater matrices.

Fig. 8 Chemical oxidation activity of different conditions for TC removal. (a) Temperature, (b) pH, (c) cation, (d) anion. For full experimental conditions, see Section 2. Experimental and simulation.

The influence of typical cations is depicted in Fig. 8c. Transition metal ions (e.g., Fe³⁺ and Cu²⁺) enhanced the catalytic activity, possibly through synergistic redox interactions or secondary activation of PMS. In contrast, monovalent ions such as Na⁺ and K⁺ inhibited TC degradation, likely due to ionic shielding effects or competition for active adsorption sites on the catalyst surface. Fig. 8d shows that the addition of nitrate (NO₃⁻) promoted TC degradation, potentially by stabilizing reactive intermediates or facilitating radical formation. However, when an excess of sulfate ions (SO₄²⁻) was introduced, the mineralization efficiency of TC was markedly reduced. This suppression effect is likely due to radical quenching or competition with the target pollutant for reactive oxygen species, especially under high PMS concentrations. These results collectively demonstrate that the HSM catalyst maintains robust activity under a wide range of environmental conditions, with optimal performance at room temperature, mildly acidic or alkaline pH, and in the presence of beneficial ionic species. Such adaptability underscores its practical potential for application in complex wastewater treatment scenarios.

3.5 Recyclability and structural stability of HSM

Catalyst recyclability is a critical factor in evaluating its potential for practical wastewater treatment applications. To assess the durability of the HSM composite, repeated catalytic experiments were performed using phenol as the model pollutant. After each reaction cycle, the catalyst was recovered by centrifugation, thoroughly washed with deionized water, and dried at 60 °C prior to reuse. As shown in Fig. 9, the HSM catalyst maintained excellent catalytic activity over three successive cycles. After the third run, the phenol degradation efficiency of HSM remained at 71.3%, indicating only a modest decline from the initial performance. In contrast, the catalytic activity of commercial MnO₂ (CM) declined dramatically, with the degradation efficiency dropping to only 5% after three cycles, reflecting poor operational stability.

Fig. 9 (a) Reusability assessment of HSM and CM over three successive phenol degradation cycles, highlighting the superior stability of HSM; (b) BET surface area analysis of HSM before and after the third cycle, confirming minimal surface area loss; (c) TEM image of the used HSM catalyst showing the retained hollow structure and morphological integrity after repeated use. For full experimental conditions, see Section 2. Experimental and simulation.

The exceptional recyclability of the HSM composite can be ascribed to the robust structural scaffold provided by the hollow SiO₂ core, which effectively anchors the MnO₂ phase, mitigating leaching of active species and preventing aggregation during catalytic cycles. Post-reaction characterization, including TEM, further corroborated that the spherical morphology and hierarchical porous architecture of HSM remained largely preserved after multiple uses, underscoring its mechanical and chemical durability. These attributes align with recent advances in catalyst design; for instance, a 2024 study by Zhang et al.²⁹ demonstrated that yolk–shell CoN/N–C@SiO₂ nanoreactors exhibited excellent reusability in PMS-based tetracycline degradation, attributing their structural robustness and catalytic durability to the hydrophilic SiO₂ shell and confined microenvironment. Similarly, Boyjoo et al.³⁰ synthesized hollow MnO₂ spheres using SiO₂-templated redox precipitation, and found that the δ-MnO₂ structure maintained its spherical morphology and catalytic performance over long-term formaldehyde oxidation cycles, while γ-MnO₂ underwent structural collapse—underscoring the importance of hollow architecture and phase selection in enhancing catalyst stability.

3.6 Degradation pathways of phenol and tetracycline

To elucidate the degradation mechanisms of phenol and tetracycline (TC) over the SiO₂/MnO₂ (HSM) composite, liquid chromatography-mass spectrometry (LC-MS) was employed to identify the intermediate products. The corresponding molecular ion peaks and proposed degradation pathways are presented in Fig. 10 (phenol) and Fig. 11 (TC), respectively.

Fig. 10 (a) LC-MS spectrum of phenol degradation intermediates collected during the reaction using HSM/PMS system; (b) proposed phenol degradation pathway involving hydroxylation, ring-opening, and oxidation into short-chain organic acids before mineralization into CO₂ and H₂O. For full experimental conditions, see Section 2. Experimental and simulation.

Fig. 11 (a) LC-MS spectrum of intermediate products generated during tetracycline degradation over HSM/PMS; (b) proposed multi-pathway mechanism for TC degradation, including demethylation, hydroxylation, ring cleavage, and subsequent mineralization into NH₄⁺, CO₂, and H₂O. For full experimental conditions, see Section 2. Experimental and simulation.

(1) Tetracycline degradation pathway

For tetracycline, two parallel degradation pathways were proposed based on the detected intermediates. In pathway I, the C–N bond in the dimethylamino group is initially cleaved via demethylation, producing P1 (m/z = 431). P1 subsequently undergoes dehydration to form P2 (m/z = 410), followed by ring cleavage to generate P3 (m/z = 279). In pathway II, multiple hydroxylation steps occur on the aromatic rings, forming P4 (m/z = 477). Dehydration and deamidation of P4 result in the formation of P5 (m/z = 399), which further undergoes ring-opening to yield P6 (m/z = 282). Continued oxidation of these intermediates leads to smaller organic compounds (P7–P10), including carboxylic acids and amides. Ultimately, these byproducts are further mineralized into NH₄⁺, CO₂, and H₂O, confirming the near-complete decomposition of the parent TC molecule. Furthermore, it can be seen from the radical-trapping experiment results shown in Fig. S2 that during TC degradation, the efficiency dropped to 31.7% upon TBA addition and 42.6% with MeOH, whereas it remained at 80.2% (L-his) and 75.4% (PBQ). This multi-step degradation—involving hydroxylation, ring-opening, decarboxylation, and mineralization—is propelled by an array of reactive oxygen species (ROS) generated via PMS activation, consistent with recent mechanistic insights from ultra-high resolution mass spectrometry studies,³¹ which revealed that ˙OH preferentially initiates degradation via hydrogen abstraction and radical addition, while SO₄˙⁻ proceeds more slowly through single electron transfer and decarboxylation—supporting the multi-pathway ROS involvement observed in our HSM/PMS system. These mechanistic insights affirm the HSM composite's capacity to orchestrate sophisticated oxidative pathways, leveraging its structural advantages to enhance ROS production and pollutant accessibility. Recent literature further supports this; for instance, Ma et al.³² developed Co₃O₄ hollow multi-shelled nanoreactors that significantly enhanced PMS activation and achieved complete degradation of carbamazepine (100% within 30 min), demonstrating the critical role of hierarchical hollow architecture in promoting radical-mediated oxidation and minimizing catalyst agglomeration. Collectively, these findings position the HSM catalyst as a highly effective platform for the comprehensive degradation of diverse organic contaminants.

(2) Phenol degradation pathway

Phenol undergoes rapid hydroxylation in the presence of active radicals (˙OH and SO₄˙⁻), resulting in the formation of hydroquinone (product I, m/z = 110). Subsequently, hydroquinone is oxidized to p-benzoquinone (product II, m/z = 108) via a dehydrogenation process. The aromatic ring of p-benzoquinone is then cleaved through electrophilic attack by singlet oxygen (¹O₂), forming short-chain organic acids such as fumaric acid (product III, m/z = 116) and oxalic acid (product IV, m/z = 90 [M–H]⁻). Further oxidation of oxalic acid leads to the formation of acetic acid (product V, m/z = 59 [M–H]⁻) and formic acid, which are eventually mineralized into carbon dioxide (CO₂) and water (H₂O). Moreover, for phenol degradation, the efficiencies were 49.3% (TBA) and 51.7% (MeOH), fell to 34.6% with L-his, and were 67.8% with PBQ. This stepwise degradation pathway underscores the dual role of radical-mediated and oxygen-mediated mechanisms in achieving efficient phenol breakdown, highlighting the essential contribution of ¹O₂ in the degradation of aromatic antibiotics via Mn-based PMS activation systems, as confirmed by LC–MS and EPR characterization.³³

4. Conclusion

In this study, a hollow SiO₂/MnO₂ (HSM) composite was rationally designed and synthesized to enhance the catalytic performance of MnO₂ for the degradation of organic pollutants. The incorporation of hollow SiO₂ not only served as a structurally stable support but also significantly improved MnO₂ dispersion and reactive site exposure, owing to its high specific surface area and well-defined porous architecture.

As a result, the HSM catalyst exhibited superior catalytic activity toward the degradation of both phenol and tetracycline, outperforming commercial MnO₂ (CM) under identical conditions. Notably, efficient pollutant removal was achieved with a low dosage of PMS (0.5 g L⁻¹), thereby addressing one of the key limitations of conventional MnO₂-based oxidation systems that typically require higher oxidant concentrations. In addition to its enhanced reactivity, the HSM catalyst demonstrated excellent recyclability and structural stability over multiple catalytic cycles, attributed to the mechanical integrity and confinement effect of the hollow SiO₂ framework. Mechanistic investigations revealed that the degradation process involved multi-step oxidative pathways, including hydroxylation, ring-opening, and mineralization into CO₂ and H₂O. Overall, this work provides a promising and environmentally benign strategy for constructing high-efficiency, low-oxidant-demand catalysts for advanced oxidation processes. The SiO₂/MnO₂ composite offers strong potential for practical applications in wastewater treatment, particularly in the removal of refractory organic contaminants.

Author contributions

Yanting Zhang: data curation & writing – original draft; Manni Li: formal analysis & writing – original draft & funding acquisition; Rui Zhao: methodology; Zhengliang Yin: conceptualization; KunZhang: validation; Qingchao Liu: investigation; Zeyu Wang: funding acquisition & supervision & writing – review & revision.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Supplementary information is available. See DOI: https://doi.org/10.1039/d5nj02554k.

Data will be made available on request.

Acknowledgements

The authors appreciate the financial support from the Natural Science Foundation of the Jiangsu Higher Education Institutions (Grant No. 23KJB430008 & 24KJB430011), the Natural Science Foundation of Jiangsu Province (Grant No. BK20230531), the Open Fund of the Key Laboratory of Advanced Energy Materials Chemistry (Ministry of Education) as well as the National Natural Science Foundation of China (Grant No. 52405363).

References

1. S. Lal, P. Singh, A. Singhal, S. Kumar, A. P. Singh Gahlot, N. Gandhi and P. Kumari, Correction: Advances in metal-organic frameworks for water remediation applications, RSC Adv., 2024, 14, 7640.
2. H. Li, J. Robichaud and Y. Djaoued, A simple way to fabricate pure anatase 2D TiO2IO monolayer: structure, color control and its application in electrochromism, RSC Adv., 2021, 11, 8065–8072.
3. X. Zhou, J. Liao, Z. Lei, H. Yao, L. Zhao, C. Yang, Y. Zu and Y. Zhao, Nickel-based nanomaterials: a comprehensive analysis of risk assessment, toxicity mechanisms, and future strategies for health risk prevention, J. Nanobiotechnol., 2025, 23, 211.
4. B. Ogoh-Orch, P. Keating and A. Ivaturi, ACS Omega, 2023, 8, 43556–43572.
5. Z. Bahmani, R. Nabizadeh, K. Yaghmaeian and M. Yunesian, Evaluation of potentially toxic elements and pharmaceutical compounds in leachate and exhaust air from non-incineration medical waste treatment devices, Sci. Rep., 2025, 15, 6395.
6. G. A. Jamali, S. K. Devrajani, S. A. Memon, S. S. Qureshi, G. Anbuchezhiyan, N. M. Mubarak, S. Z. M. Shamshuddin and M. T. H. Siddiqui, Holistic insight mechanism of ozone-based oxidation process for wastewater treatment, Chemosphere, 2024, 359, 142303.
7. N. T. A. Nguyen and H. Kim, Ag3PO4-Deposited TiO2@Ti3C2 Petals for Highly Efficient Photodecomposition of Various Organic Dyes under Solar Light, Nanomaterials, 2022, 12, 2464.
8. N. N. Roslan, H. L. H. Lau, N. A. A. Suhaimi, N. N. M. Shahri, S. B. Verinda, M. Nur, J. W. Lim and A. Usman, Recent Advances in Advanced Oxidation Processes for Degrading Pharmaceuticals in Wastewater–A Review, Catalysts, 2024, 14, 189.
9. F. Yang, P. Hu, F. F. Yang, B. Chen, F. Yin, R. Sun, K. Hao, F. Zhu, K. Wang and Z. Yin, Emerging Enhancement and Regulation Strategies for Ferromagnetic 2D Transition Metal Dichalcogenides, Adv. Sci., 2023, 10, 2300952.
10. M. Wang, C. Li, B. Liu, W. Qin and Y. Xie, Facile Synthesis of Nano-Flower β-Bi2O3/TiO2 Heterojunction as Photocatalyst for Degradation RhB, Molecules, 2023, 28, 882.
11. J. A. Pinedo-Escobar, J. Fan, E. Moctezuma, C. Gomez-Solís, C. J. C. Martinez and E. Gracia-Espino, Nanoparticulate double-heterojunction photocatalysts comprising TiO2(Anatase)/WO3/TiO2(Rutile)with enhanced photocatalytic activity toward the degradation of methyl orange under near-ultraviolet and visible light, ACS Omega, 2021, 6, 11840–11848.
12. Q. Wan, J. Yang and Z. Sun, Adsorbed hydroxyl radicals dominated peroxymonosulfate activation in acetaminophen degradation using modified electrolytic manganese anode slime: Catalyst evaluation and free radical identification, Chem. Eng. J., 2024, 486, 150386.
13. H. Luo, H. Du, M. Jiang, C. Yang, T. Weng, Z. Chen, F. Jiang and H. Chen, Crystal phase-driven performance of MnO2 in aqueous phase low-temperature thermal catalysis: Synergistic interactions between Mn3+ and surface lattice oxygen, J. Hazard. Mater., 2024, 476, 135209.
14. J. D. García-Espinoza, I. Robles, A. Durán-Moreno and L. A. Godínez, Photo-assisted electrochemical advanced oxidation processes for the disinfection of aqueous solutions: A review, Chemosphere, 2021, 274, 129957.
15. S. Dasarathan, M. Ali, T. J. Jung, J. Sung, Y. C. Ha, J. W. Park and D. Kim, Vertically aligned binder-free TiO2 nanotube arrays doped with Fe, S and Fe-S for li-ion batteries, Nanomaterials, 2021, 11, 2924.
16. L. B. Zhu and S. N. Ding, Enhancing the Photocatalytic Performance of Antibiotics Using a Z-Scheme Heterojunction of 0D ZnIn2S4 Quantum Dots and 3D Hierarchical Inverse Opal TiO2, Molecules, 2023, 28, 7174.
17. D. Han, B. Li, S. Yang, X. Wang, W. Gao, Z. Si, Q. Zuo, Y. Li, Y. Li, Q. Duan and D. Wang, Engineering charge transfer characteristics in hierarchical Cu2S QdS @ ZnO nanoneedles with p–n heterojunctions: Towards highly efficient and recyclable photocatalysts, Nanomaterials, 2019, 9, 16.
18. J. Wang, H. Zhang, L. Wang, K. Yang, L. Cang, X. Liu and W. Huang, Highly Stable and Efficient Mesoporous and Hollow Silica Antireflection Coatings for Perovskite Solar Cells, ACS Appl. Energy Mater., 2020, 3, 4484–4491.
19. S. Xuejun, S. Ren, X. Du and Y. Han, Sound insulation and mechanical properties of epoxy/hollow silica nanospheres composites, Mater. Res. Express, 2022, 9, 095006.
20. J. Su, Q. Ke and H. Xu, Hierarchically MnO2/SiO2 nanocomposites with high solar light-driven photo-thermo catalysis activity for efficient toluene removal, Appl. Surf. Sci., 2023, 608, 155177.
21. Y. Zhang, F. Wang, P. Ou, H. Zhu, Y. Lai, Y. Zhao, W. Shi, Z. Chen, S. Li and T. Wang, High efficiency and rapid degradation of bisphenol A by the synergy between adsorption and oxidization on the MnO2@nano hollow carbon sphere, J. Hazard. Mater., 2018, 360, 223–232.
22. T. Peng, X. Dai, Y. Zhang, P. Zheng, J. Wang, S. Wang, Z. Wang, Y. Zeng, J. Li and H. Jiang, Facile synthesis of SiO2@MnO2 nanocomposites and their applications on platforms for sensitively sensing antibiotics and glutathione, Sens. Actuators, B, 2020, 304, 127314.
23. J. Wei, K. Li, H. Yu, H. Yin, M. A. Cohen Stuart, J. Wang and S. Zhou, Controlled Synthesis of Manganese Oxide Nanoparticles Encaged in Hollow Mesoporous Silica Nanoreactors and Their Enhanced Dye Degradation Activity, ACS Omega, 2020, 5, 6852–6861.
24. H. Tian, J. Zhao, X. Wang, L. Wang, H. Liu, G. Wang, J. Huang, J. Liu and G. Q. Lu, Construction of hollow mesoporous silica nanoreactors for enhanced photo-oxidations over Au-Pt catalysts, Natl. Sci. Rev., 2020, 7, 1647–1655.
25. J. Huang, Y. Dai, K. Singewald, C. C. Liu, S. Saxena and H. Zhang, Effects of MnO2 of different structures on activation of peroxymonosulfate for bisphenol A degradation under acidic conditions, Chem. Eng. J., 2019, 370, 906–915.
26. L. Lin, W. Fang, Q. Liang, Y. Xing, M. Sun and H. Luo, Synthesis of Fe-doped sludge biochar from Fenton sludge for efficient activation of peroxymonosulfate in tetracycline hydrochloride degradation, J. Environ. Chem. Eng., 2024, 12, 112590.
27. L. Wang, R. Li, Y. Zhang, Y. Gao, X. Xiao, Z. Zhang, T. Chen and Y. Zhao, Tetracycline degradation mechanism of peroxymonosulfate activated by oxygen-doped carbon nitride, RSC Adv., 2023, 13, 6368–6377.
28. M. Liu, W. Zhang, R. Ni, Z. Wang, H. Zhao, X. Zhong, Y. Wang, D. Shang, Z. Guo, E. H. Ang and F. Yang, Construction of phase-separated Co/MnO synergistic catalysts and integration onto sponge for rapid removal of multiple contaminants, Mater. Horiz., 2024, 11, 3316–3329.
29. S. Zhang, H. Gao, X. Xu, R. Cao, H. Yang, X. Xu and J. Li, MOF-derived CoN/N-C@SiO2 yolk–shell nanoreactor with dual active sites for highly efficient catalytic advanced oxidation processes, Chem. Eng. J., 2020, 381, 122670.
30. Y. Boyjoo, G. Rochard, J. M. Giraudon, J. Liu and J. F. Lamonier, Mesoporous MnO 2 hollow spheres for enhanced catalytic oxidation of formaldehyde, Sustainable Mater. Technol., 2019, 20, e00091.
31. S. Zhang, Z. Hao, J. Liu, L. Gutierrez and J. P. Croué, Molecular insights into the reactivity of aquatic natural organic matter towards hydroxyl (˙OH) and sulfate (SO4˙⁻) radicals using FT-ICR MS, Chem. Eng. J., 2021, 425, 130622.
32. Q. Ma, Y. Zhang, X. Zhu and B. Chen, Hollow multi-shelled Co3O4 as nanoreactors to activate peroxymonosulfate for highly effective degradation of Carbamazepine: A novel strategy to reduce nano-catalyst agglomeration, J. Hazard. Mater., 2022, 427, 127890.
33. L. Li, Y. Wang, G. Qu, N. Ren, Y. Liu, P. Lu, J. Wang, M. Cheng, Y. Cai, X. Li, D. Kong and R. Xu, Study on the Mechanism of Efficient PMS Activation by Surface Oxygen Vacancy-Enriched Mn-Doped Copper Slag Particle Electrodes for Rapid Degradation of Ofloxacin, Available at SSRN: https://ssrn.com/abstract=5164649.

---

Footnote

† These authors contributed equally.

---