Enhanced Photo-Fenton Activity Using Magnetic Cu_0.5Mn_0.5Fe₂O₄ Nanoparticles as a Recoverable Catalyst for Degrading Organic Contaminants

Abstract

Interest in using various nanoparticle catalysts to activate H₂O₂ with light for organic contaminant and wastewater treatment is steadily increasing. We successfully synthesized magnetically recoverable Cu_0.5Mn_0.5Fe₂O₄ nanoparticles using a simple co-precipitation method followed by melamine-assisted calcination. Material characterization revealed that melamine acted as a coordinating agent during the calcination process that promoted a ferrite structure. Copper (Cu)-substitution effectively decreased material aggregation and promoted catalytic activities. Cu_0.5Mn_0.5Fe₂O₄ nanoparticles showed outstanding catalytic performance on several organic contaminants (87.6–100.0% removal within 2 h). Using oxytetracycline (OTC) as a surrogate wastewater constituent, we found that the hydroxyl radical (^•OH) and superoxide anions (^•O₂⁻) were the active radical species involved in OTC degradation. Cu_0.5Mn_0.5Fe₂O₄ nanoparticles exhibited excellent photo-Fenton catalytic ability in real wastewater and demonstrated high material stability, even after four consecutive uses (i.e., fourth cycle). In a pilot-scale experiment (10 L), we provide proof that our rigorous treatment system was able to remove remnant OTC, TOC, and also any available colloidal particles to only 1 NTU. Ecotoxicity studies using an aquatic plant (Hydrilla verticillata) and zooplankton revealed that treated water could be reused in various ratios. Furthermore, at 5% of treated water, rapid leaf recovery and a significant increase in rotifer numbers were reported. These observations support the use of Cu_0.5Mn_0.5Fe₂O₄/H₂O₂/light as an efficient and environmentally friendly catalytic system for treatment of organic contaminants, and a radical generating mechanism is proposed.

1. Introduction

Organic contaminants such as pharmaceuticals, personal care products, agrochemicals, and industrial dyes are ubiquitous in aquatic environments due to their widespread use and general recalcitrance during conventional wastewater treatment [1,2,3]. While chlorination is commonly used to remove pathogenic microorganisms and several organic contaminants, the generation of unwanted disinfection byproducts (DBPs) such as trihalomethane (THM) and haloacetic acid (HAA) can occur from the chlorine reaction with the organic precursors, i.e., natural organic matter (NOM). Therefore, an efficient and practical approach to removing recalcitrant wastewater contaminants using environmentally friendly treatment technologies is of paramount importance.

Advanced oxidation processes (AOPs) are recognized as established technologies that can degrade persistent contaminants by producing highly reactive oxygen species (ROS) such as ^•OH, SO₄^•−, ¹O₂, and ^•O₂⁻ [4,5]. Among AOPs, heterogeneous photo-Fenton-like processes have received considerable attention due the reusability of the catalyst and their capability to operate over a wide pH [6,7]. Although many powerful catalysts such as BUC-21 (Fe) [8], LaFeO₃/lignin-biochar [9], and Fe/SCN [10] have been created, the recovery of these non-magnetic catalysts could hinder the practical application of the technology. Recently, researchers have used several manganese ferrite nanoparticles (MnFe₂O₄) as heterogeneous catalysts to enhance oxidation via electron transfer reactions with ozone [11], persulfate [5], peroxymonosulfate [4], and heterogeneous photo-Fenton reagents [12,13]. The superiority of heterogeneous catalysts is attributable to the existence of various valences in the nanocomposites (Mn⁴⁺, Mn³⁺, Mn²⁺, Fe³⁺, and Fe²⁺) that initiate the synergistic effect between the Mn and Fe redox cycles, and result in better degradation performance [4,14]. In addition, these materials are not only biocompatible but also provide excellent intrinsic magnetism, which facilitates recovery and material reusability from the treated water. Although several researchers have proposed various approaches for synthesizing the MnFe₂O₄ nanoparticles using co-precipitation or sol-gel methods, these methods also produce a hematite (α-Fe₂O₃) impurity that reduces the catalyst magnetic properties [15,16]. By using another method such as hydrothermal synthesis [4,17], the obtained bare MnFe₂O₄ particles, in turn, possessed too high a level of magneticity that facilitates agglomeration, which reduces the active surface area. To overcome some of these unexpected shortcomings, changing the catalyst physicochemical properties by adding non-metal or metal dopants, or coupling with other semiconductors, or substituting Mn with other metals to construct multicomponent ferrite have been attempted [6,11,18,19].

Copper (Cu) ions appear to be a promising choice to substitute Mn ions to construct Cu_XMn_1-XFe₂O₄ nanoparticles because of their excellent oxygen exchangeability and outstanding capability for electron transfer through the Cu²⁺/Cu¹⁺ redox couple [20]. Because Cu has a smaller ionic radius than Mn, this substitution may create lattice distortions in the spinel structure, which can change the nanoparticle’s magnetic and catalytic properties and result in greater oxidative reactivity for pollutant removal [21,22]. These Cu-substitution dosages, including the cation distribution between the octahedral and tetrahedral sites of multicomponent ferrite, are controlling factors for morphological properties and material magnetism [19].

In this work, we synthesized Cu_0.5Mn_0.5Fe₂O₄ nanoparticles by a simple co-precipitation technique followed by melamine-assisted calcination methods. We then documented the nanoparticle’s photo-Fenton catalytic activity in degrading several organic pollutants frequently found in treated wastewater effluent, such as oxytetracycline (OTC), enrofloxacin (ENR), sulfadimethoxine (SDM), sulfamethoxazole (SMX), alachlor (ALA), 2,4 dichlorophenoxyacetic acid (2,4 D), atrazine (ATZ), parabens, methylene blue (MB), methyl orange (MO), and crystal violet (CV). The nanoparticle’s physicochemical characteristics and reusability were also quantified. Using oxytetracycline as a representative wastewater constituent, we show the effects of organic matter on degradation efficacy. Finally, to test the technology readiness of the nanoparticle, we scaled up experiments and determined the ecotoxicity of the treated water on an aquatic plant (Hydrilla verticillata) and zooplankton. Overall results provided proof that our ternary metal ferrite catalyst has great potential as a recoverable catalyst for degrading various organic contaminants and is compatible with conventional wastewater treatment systems.

2. Materials and Methods

Chemicals, vendors, and chemical analyses performed are provided in detail in Supplementary Information (Sections S1–S3 and Table S1).

2.1. Preparation of Catalysts

In an effort to avoid high pressure and temperature, the Cu_XMn_1-XFe₂O₄ (x = 0, 0.2, 0.33, 0.5, and 0.66) nanoparticles were synthesized by the co-precipitation method followed by melamine-assisted calcination. In brief, at x = 0.5, the CuSO₄ and MnSO₄⋅H₂O concentrations and volumes were equal. For example, to synthesize the Cu_0.5Mn_0.5Fe₂O₄ nanoparticles, we continuously stirred the mixture of 125 mL CuSO₄ (0.2 M), 125 mL MnSO₄⋅H₂O (0.2 M), 250 mL FeCl₃∙6H₂O (0.4 M), and 500 mL deionized water (DI) at 80 °C for 1 h. Then, we separately prepared 8 M NaOH (80 °C) and slowly dripped the base into the mixture for ~1 h to raise the pH to 10.5. After stirring for another hour, we used vacuum filtration to collect the precipitates, washed them with DI water and ethanol several times, and dried at 95 °C for 15 h. To obtain the final product, we mixed 2 g dry particles with 1 g melamine in a crucible and calcined at 550 °C for 3 h.

2.2. Photo-Fenton Catalytic Performance Evaluation

The synthesized catalysts (Cu_0.5Mn_0.5Fe₂O₄) were used in the photo-Fenton catalytic experiments, which were designed to treat organic contaminants such as antibiotics, herbicides, parabens, and industrial dyes. Initially, we investigated the photo-Fenton catalytic activity of Cu_XMn_1-XFe₂O₄ nanoparticles (x = 0, 0.2, 0.33, 0.5, and 0.66) using methylene blue (MB) (80 mg/L) as a starting substrate. To obtain the best catalyst composition, we chose the nanoparticle catalyst that performed the best in degrading methylene blue. Unless stated otherwise, all experiments were performed in a 250-mL beaker with 50 mg Cu_XMn_1-XFe₂O₄ nanoparticles, which were dispersed in 100 mL of target contaminant (C₀ = 0.1 mM) and stirred in the dark for 30 min to achieve adsorption-desorption equilibrium between the contaminant and catalyst. Then, all flasks were placed in a specifically designed reactor that housed a simulated sunlight source (~12 cm above the water surface) using a commercial 75 W halogen lamp at visible wavelength in the 400–700 nm region. The experiments started immediately after 1 mL of H₂O₂ was injected into the flask following the light irradiation. Temporal 1-mL samples were collected and passed through a 0.45 µm PTFE syringe filter. The 0.7 mL of filtered solution was transferred to an HPLC vial, which held 0.875 mL of methanol as a quencher. The samples were kept in a refrigerator at 4 °C prior to HPLC analysis. For colored contaminants, filtered samples were immediately transferred into a cuvette without quenching and quantified for residual concentration using a spectrophotometer. We measured total organic carbon (TOC) before and after the reaction to quantify TOC removal (%) at 120 min and ensure the complete contaminant degradation.

2.3. Material Adsorption Capacity and pH of Zero Point of Charge (pH_pzc)

Because most reactions occur on the catalyst surface, the prepared Cu_0.5Mn_0.5Fe₂O₄ nanoparticle adsorption capacity may reveal a key role in the degradation process. In brief, we mixed 30 mL of each contaminant (C_o = 0.1 mM) with 0.12 g of the catalysts in a 40-mL amber glass screw-cap vial and agitated at 190 rpm for up to 3 h using a reciprocating shaker. Temporal samples were periodically collected.

As pH can influence ionic species in solution as well as on the Cu_0.5Mn_0.5Fe₂O₄ nanoparticle surface, we determined the nanoparticle point of zero charge (pH_pzc) using a modified method [23]. Firstly, we adjusted the solution pH of each 50 mL sodium chloride solution (0.01 M) to a pH ranging from 2 to 12 using HCl (0.1 M) or NaOH (0.1 M). Then, 0.12 g of Cu_0.5Mn_0.5Fe₂O₄ nanoparticles was added, and the pH was initially measured prior to applying agitation for 48 h. The final pH of each vial was measured and plotted against the designated pH. The intersection of this curve and the initially measured pH curve is the sample pH_pzc values.

2.4. Quenching Experiments

To elucidate the catalytic mechanisms and determine the radicals involved during the degradation processes, radical scavenging experiments were performed. Using generalized procedures, oxytetracycline (OTC) degradation was measured but additional radical scavengers were added at a concentration of 100 mM. Specifically, we used tertbutyl alcohol (TBA) as the scavenger for hydroxyl radicals (^•OH), dimethyl sulfoxide (DMSO) for photogenerated electrons (e⁻), and p-benzoquinone (pBQ) as the scavenger for superoxide anions (^•O₂⁻). Once the experiment started, we monitored changes in OTC concentration and compared results with the control (i.e., without scavengers).

2.5. Recycling Experiments

To determine the robustness of the nanoparticle catalyst, we reused the catalyst in four consecutive experiments under identical conditions. Four-cycles of OTC degradation experiments were performed. Each cycle consisted of a 1 h treatment of OTC with the photo-Fenton treatment. After one cycle (60 min), the catalyst was retrieved magnetically from the batch reactor using a neodymium magnet and then washed with water prior to drying at 80 °C for 8 h. The OTC and TOC removal percentages and leaching metal contents (i.e., Cu, Mn, and Fe) from each cycle were recorded. After the fourth cycle, these materials were characterized using XPS for determining changes in metal oxidation states.

2.6. Effect of Real Water

To investigate the applicability of the system to treat OTC in actual water samples, we spiked OTC to an initial concentration of 0.1 mM in the waters collected from the municipal wastewater treatment plant at Kasetsart University (Bangkok, Thailand) (13.855360, 100.570124), actual discharge of domestic wastewater (Wastewater Collection Station #1: 13.854984, 100.570327, and Wastewater Collection Station #2: 13.822234, 100.471106), and canal water (13.820820, 100.469383). The basic water quality parameters were determined immediately after collection, and again after adding OTC, and finally, after treating (Table S2). Then, the same sampling protocol for OTC treatment and analysis were followed.

2.7. Pilot-Scale Experiment

To determine whether our technology could treat contaminated water more efficiently and potentially be scaled up for commercial use, we conducted a pilot-scale experiment. Here, the photocatalytic experiment was scaled up 100-fold (i.e., 100 mL to 10 L). The experiment was performed in a glass container (44 × 19 × 21 cm; W × L × H) equipped with three 75 W halogen lamps, which were aligned perpendicular to the lamp light direction and symmetrically located along the water surface. Initially, we added 2 g of catalyst into 10 L of the collected wastewater (Wastewater Collection Station #3: 13.854592, 100.570231) and continuously stirred for 30 min to reach catalyst adsorption–desorption equilibrium. Fifty mL of H₂O₂ was added to the solution followed by light irradiation. At the predetermined intervals, 40 mL of solution was transferred into an EPA vial and the catalyst was magnetically extracted to allow solution turbidity measurement. After treating for 6 h, we removed all catalysts and rested the treated water to remain stabilized at room temperature for 72 h for successive evaluations of the toxicological impact on an aquatic plant (Hydrilla verticillata) and zooplankton.

2.8. Toxicological Evaluation

To test the ecotoxicity of the treated wastewater, Hydrilla verticillata was selected as a representative aquatic submerged plant. Using water collected at Wastewater Collection Station #3, we filtered the water with a 65 µm plankton net to remove zooplankton and used this water as natural receiving water. Treated water from the previous experiments was split into either 250, 500, or 750 mL aliquots and mixed with the filtered wastewater for a total of 5 L. These mixtures of treated and untreated water were placed in 10 L water containers, and labeled as 5%, 10%, and 15%. The experiments were performed in duplicate. These water mixtures were used in the following experiment. Changes in Hydrilla verticillata and the zooplankton community in the water mixtures were compared with the untreated water (Control; 0%). Approximately 10 g of Hydrilla verticillata (wet weight) was carefully selected and raised in simulated natural water for 3 d at room temperature to ensure that these plants were maintained in healthy conditions. Then, these acclimated plants were exposed to each water mixture for up to 20 d. At 5 d, representative leaves were collected from the containers, rinsed gently with water, and visually observed using a stereomicroscope.

All zooplankton related experiments conducted in this study followed the animal research ethical approval obtained from Kasetsart University (ID# ACKU64-VTN-016). To explore changes in the zooplankton community after exposure to the water mixtures (i.e., 0, 5, 10, and 15%, v/v), we filtered natural zooplankton using a plankton net from 160 L of Wastewater Collection Station #3 collected at <45 cm below the surface water. These zooplankton were divided into eight portions and placed in water tanks containing different treated water mixtures (0–15%). Each experimental unit was performed in duplicates. After 3 d of exposure, we used a plankton net to collect zooplankton from each container, transferred to a 100-mL bottle, and added 10 mL of ethanol to preserve them for further analysis. Changes in zooplankton numbers were assessed by taking repeated observations on 1 mL sub-sample and counted using a Sedgewick–Rafter counting chamber. Identification of zooplankton was classified up to the genus level.

3. Results and Discussion

3.1. Material Characterization

Two synthesized catalysts’ purity and crystallinity were investigated using XRD (i.e., with and without melamine addition). By adding melamine, the Cu_0.5Mn_0.5Fe₂O₄ nanoparticle diffraction peaks located at 18.3°, 30.0°, 35.4°, 37.0°, 43.0°, 53.3°, 56.8°, and 62.4° were well indexed to the (111), (220), (311), (222), (400), (422), (511), and (440) crystal plane of the Cu_0.5Mn_0.5Fe₂O₄ spinel structure (JCPDS No. 74-2072) (Figure 1a). The X-ray mapping images and the elemental composition analysis using energy dispersive spectrometry (EDS) revealed that the Cu elements were distributed in the samples and the atomic composition of Cu and Mn was relatively close (Figure S1). However, the crystallite size calculation based on Debye–Scherrer’s equation revealed Cu_0.5Mn_0.5Fe₂O₄ nanoparticles in a smaller size compared to MnFe₂O₄ (35.7 vs. 29.8 nm). This was because Cu²⁺ ions had distorted the MnFe₂O₄ spinel structures, resulting in decreasing crystallite size and more active surface area. With melamine addition, XRD peaks showed much higher Cu_0.5Mn_0.5Fe₂O₄ intensity with no hematite (α-Fe₂O₃) impurity, confirming the importance of adding melamine during Cu_0.5Mn_0.5Fe₂O₄ nanoparticle synthesis (Figure 1a). To further explain the melamine role, we used FTIR and XPS to determine the surface chemical composition and compare the catalyst spectral changes with and without melamine addition, which are provided in detail in the Supplementary Information (Section S4 and Figure S2).

To verify the catalyst magnetic property, we examined their M-H hysteresis loop. Results showed that the Cu_0.5Mn_0.5Fe₂O₄ saturation magnetization (Ms) values significantly increased by 442% compared to the prepared catalyst without melamine addition and almost mirrored that of MnFe₂O₄ (Figure 1b). Although the Cu_0.5Mn_0.5Fe₂O₄ and MnFe₂O₄ Ms values were slightly different, we observed much less Cu_0.5Mn_0.5Fe₂O₄ agglomeration (Figure 1b and its inset photos before applying magnetization). This can be explained by (1) a smaller crystallite size, and (2) the Cu distribution in the spinel ferrite that reduced magnetic interaction between particles. However, the Cu_0.5Mn_0.5Fe₂O₄ showed rapid recovery using a magnet within a few minutes (Figure 1b inset photo). The magnetism improvement was mainly from the α-Fe₂O₃ disappearance after recrystallization as the α-Fe₂O₃ oriented in the antiparallel position, which exhibited weakly ferromagnetic behavior at room temperature [24] (Figure 1a).

Furthermore, the UV-Vis diffuse reflectance spectra revealed that the Cu_0.5Mn_0.5Fe₂O₄ exhibited a broad absorption in all regions of UV and visible light and, by plotting (αhν)² versus photon energy (hν) (i.e., Tauc’s equation), the Cu_0.5Mn_0.5Fe₂O₄ bandgap energy (E_g) was found to be 1.7 eV, confirming that the catalyst can be activated under visible light irradiation (Figure S3a,b).

The Cu_0.5Mn_0.5Fe₂O₄ nanoparticle morphological structures, surface textural, and size were characterized using SEM and TEM. Most particles had octahedral morphology (Figure 2a–c). Compared with other synthesis approaches (e.g., solvothermal, Sun et al. [21]; sol-gel autocombustion, Velinov et al. [16]; and co-precipitation followed by annealing treatment, Yang et al. [25]); our Cu_0.5Mn_0.5Fe₂O₄ had much higher crystalline morphology. Noticeably in both SEM and magnified TEM images, small particles were well-distributed among larger particles, which further confirmed the impeccable melamine-assisted calcination, allowing rapid electron transfer and efficient photocatalytic activity (Figure 2a–c). The HRTEM image also revealed lattice fringe spacing of 0.25 nm corresponded to the (311) crystal plane (Figure 2d).

3.2. Degradation Performance

Because the Cu-substitution dosages for synthesizing Cu_XMn_1-XFe₂O₄ ferrites can alter their physiochemical characteristic, which can directly affect catalyst applicability, we prepared four Cu_XMn_1-XFe₂O₄ catalysts at different compositions (x = 0.2, 0.33, 0.5, and 0.66) and determined the photo-Fenton efficiency on methylene blue (MB). Results showed that Cu-Mn ferrites have better catalytic performance as compared to MnFe₂O₄ while Cu_0.5Mn_0.5Fe₂O₄ provided the highest degradation performance (Figure S4). Besides decreasing the material agglomeration (Figure 1b), Cu-substitution can also provide active Cu ion species that can promote the photo-Fenton reactions [Equations (1) and (2)] [1], and can improve its catalytic activity by synergistic effects between metal valences (i.e., Mn, Fe, and Cu) [Equations (3)–(5)] [7,21].

Cu²⁺ + hν + e⁻ → Cu⁺

(1)

Cu⁺ + H₂O₂ → Cu²⁺ + ^•OH + OH⁻

(2)

Mn³⁺ + Fe²⁺ → Mn²⁺ + Fe³⁺ E₀ = 0.74 V

(3)

Fe³⁺ + Cu⁺ → Fe²⁺ + Cu²⁺ E₀ = 0.60 V

(4)

Mn³⁺ + Cu⁺ → Mn²⁺ + Cu²⁺ E₀ = 1.34 V

(5)

The results showed that the differences in methylene blue degradation efficiency were minimal between x = 0.33 and x = 0.66 (Figure S4). Moreover, further adding Cu would decrease the recovery efficiency as it reduces the catalyst Ms values [19]. Therefore, we selected Cu_0.5Mn_0.5Fe₂O₄ for all successive experiments.

One way to further investigate the Cu_0.5Mn_0.5Fe₂O₄ catalytic performance is through both contaminant and TOC removal under different treatment systems. By using oxytetracycline (OTC) as a contaminant surrogate, we found that only 50% of OTC concentration and 8% of TOC content were removed by H₂O₂ alone due to its low redox potential (E₀ = 1.76 V) [26] (Figure 3a,b). Interestingly, the OTC transformed faster in the H₂O₂/light system but no changes in TOC removal were observed, indicating that the generated ^•OH was able to degrade OTC but not mineralize (i.e., → CO₂). When we used Cu_0.5Mn_0.5Fe₂O₄ to activate H₂O₂, the OTC removal was not as strong as compared to the H₂O₂/light system, but the TOC removal efficiency increased by 16%. This increase was mainly due to the metal ions from Cu_0.5Mn_0.5Fe₂O₄ that were able to initiate other reactive oxygen species (ROS) such as superoxide radicals (^•O₂⁻) and singlet oxygen (¹O₂) [Equations (6)–(8)] [13,27,28].

M^{(n + 1) +} + H₂O₂ → Mⁿ⁺ + ^•OOH + H⁺

(6)

^•OOH → ^•O₂⁻ + H⁺

(7)

^•OOH + ^•O₂⁻ → ¹O₂ + OOH⁻

(8)

Notably, a much faster OTC removal efficiency was observed from the Cu_0.5Mn_0.5Fe₂O₄/H₂O₂ system under solar light irradiation. This system was able to completely degrade OTC within 45 min and removed up to 58% of TOC. Such results indicated that Cu_0.5Mn_0.5Fe₂O₄ can provide dual beneficial effects from both light irradiation and H₂O₂ activation in the photo-Fenton reaction, thereby promoting reactive oxygen species during oxidation processes [Equations (1) and (9)–(12)] [1,13,29] (Figure S3b).

Cu_0.5Mn_0.5Fe₂O₄ + hν → e⁻ + h⁺

(9)

e⁻ + O₂ → ^•O₂

(10)

e⁻ + H₂O₂ → ^•OH + OH⁻

(11)

Fe³⁺ + H₂O + hν → Fe²⁺ + ^•OH + H⁺

(12)

To ensure that the Cu_0.5Mn_0.5Fe₂O₄/H₂O₂/light system is well-suited to work with various organic pollutants, we conducted batch experiments using an individual pollutant as a starting substrate under neutral pH. Temporal changes in pollutant concentrations were also monitored to determine degradation rates, which fit pseudo-first order kinetics with results ranging from 0.009 min⁻¹ (methyl orange, MO) to 0.086 min⁻¹ (OTC) (Figure 4a). Our treatment system removed pollutants with 87.6% to 100.0% efficiency within 120 min while TOC removal efficiencies ranged between 28.2% and 67.9% (Figure 4a). The one exception to these findings occurred with methylene orange (MO)—an anionic dye widely used in the textile industry—where the MO reduction was only 71.4% with a 16.1% TOC removal, indicating that differences in pollutant formal charge versus catalyst surface charge can influence removal efficiencies.

The effect of charge interactions between the catalyst and contaminants was further investigated by determining the Cu_0.5Mn_0.5Fe₂O₄ nanoparticles pH of zero-point charge (pH_pzc) as well as the adsorption potential of the various contaminants following 60 min of contact. Results showed that up to 55% OTC was removed while MO and other pollutants were removed <5% except the enrofloxacin (ENR) (10%) (Figure S5). The higher adsorption efficiency can be expected as the adsorbed OTC molecules were close to the generated radicals (Figure 4a, Figure S5a). Although the adsorption efficiency depends on the adsorbent amount, previous research has shown that copious adsorbent (>1 g/L) can lead to self-aggregation and lead to a decrease in surface area active sites [30]. Furthermore, our results confirmed that using a 0.5 g/L catalyst dose was sufficient for removing OTC under photo-Fenton process (Figure 4a). The MO reduction, on the other hand, was difficult to remove either from degradation or adsorption due to the adsorbate–adsorbent charge interaction. At this working condition (pH = 7), the Cu_0.5Mn_0.5Fe₂O₄ (pH_pzc = 6.0) and the MO molecule (MO_pKa = 3.47) possessed the same negative charge, which sustained a repulsive electrostatic interaction (Figure S5b). Unlike the OTC molecule, which is always present in neutral form (OTC_pKa2 = 7.49), the adsorption favorably occurs, confirming the OTC fast degradation rate as the adsorption simultaneously occurs followed by the photo-Fenton process [31] (Figure 4a).

By using humic acid (HA) as a representative of natural organic matter, we tested our system under more environmentally relevant conditions. In the treatment of HA, TOC content and color were monitored (Figure 4b). Results showed that, up to 57.41% HA was removed and the solution color turned clear within 120 min while, in the adsorption study, up to 35% was removed, which was second to the OTC removal (Figure 4b and Figure S5a). This can be explained by the natural HA form that is mostly deprotonated. Therefore, in the HA abundant solution, the solution pH was lower than the Cu_0.5Mn_0.5Fe₂O₄ pH_pzc, making the catalyst surface positively charged, enhancing both adsorption behavior and consequently increasing photo-Fenton performance (Figure 4b and Figure S5a,b). Moreover, besides the carboxylic groups that can form the surface-OH bonding with the catalyst surface, numerously located phenolic groups on the HA surface can initiate the hydrophobic forces between the catalyst surface and the unprotonated phenolic groups. A further explanation for the fast degradation was these organic radical formations (i.e., R^•, ROO^•) [Equations (13)–(16)] [29] which participated in generating reactive radicals (i.e., ^•O₂⁻, ^•OH), which could be involved with further degradation of these organic contaminants [Equations (7) and (17)] [1].

RH + ^•OH → R^• + H₂O

(13)

R^• + O₂ → ROO^•

(14)

ROO^• + RH → ROOH + R^•

(15)

ROO^• + H₂O → ROH + ^•OOH

(16)

H₂O₂ + ^•OOH → O₂ + ^•OH + H₂O

(17)

3.3. Photo-Fenton Mechanism

The photo-Fenton process is known to generate reactive radical species during pollutant degradation. By using various radical scavengers, the roles of these species were investigated. Compared to the control, results showed that these radical scavengers inhibited OTC removal in the following order: pBQ > TBA > DMSO; this indicates that ^•O₂⁻ and ^•OH were the predominant active species (Figure 5a). Moreover, OTC degradation was impeded in the presence of DMSO, an effective scavenger for photogenerated electrons (e⁻). This result clearly confirmed that the e⁻ could also act as a precursor during the photo-Fenton process, which would be one additional route for generating subsequence ^•O₂⁻ and ^•OH as discussed earlier [Equations (9)–(11)].

To obtain in-depth information on the catalytic mechanism, we used the XPS spectra to confirm the Cu_0.5Mn_0.5Fe₂O₄ elemental composition and metal oxidation states before and after treatment in the photo-Fenton process. Mn 2p spectra were composed of two main binding energy peaks at ~641.0 (Mn 2p_3/2) and ~652.5 eV (Mn 2p_1/2). The coexistence of Mn ion multi-valence states was evidenced from the three binding energy peaks at ~639.1, ~641.2, and ~643.1 eV in Mn 2p_3/2 and ~650.8, ~652.9, and ~654.8 eV in Mn 2p_1/2, corresponding to Mn²⁺, Mn³⁺, and Mn⁴⁺ [18,32] (Figure 5b). After use in the treatment, the Mn³⁺ decreased from 63.1% to 55.5%, the Mn²⁺ increased from 16.5% to 18.6%, and the Mn⁴⁺ increased from 20.4% to 25.9%, confirming the Mn³⁺ ions having a key role in the ^•OOH and ^•OH generation (Figure 5b). The continuous interconversion among Mnⁿ⁺ states provided proof that H₂O₂ can be constantly supplied with Mn electron transfer, in which H₂O₂ would be decomposed to ^•OH, thereby enhancing the pollutant degradation efficiency [Equations (6) and (18)–(23)] [27,33,34].

Mn²⁺ + H₂O₂ → Mn³⁺ + ^•OH + OH⁻

(18)

Mn³⁺ + H₂O₂ → Mn²⁺ + ^•OOH + H⁺

(19)

Mn³⁺ + H₂O₂ → Mn⁴⁺+ ^•OH + HO⁻

(20)

Mn⁴⁺ + H₂O₂ → Mn³⁺ + ^•OOH + H⁺

(21)

Mn³⁺ + ^•OOH → Mn²⁺ + O₂ + H⁺

(22)

Mn³⁺ + ^•OH → Mn⁴⁺ + OH⁻

(23)

Fe 2p spectra revealed two binding energy peaks at ~710.5 and ~711.7 eV (Fe 2p_3/2) and 723.9 eV, 725.5 eV (Fe 2p_1/2), corresponding to Fe³⁺ while the deconvoluted peaks of ~708.5 and ~722.3 eV were ascribed to the Fe²⁺ [35] (Figure 5c). Results revealed the decrease in Fe²⁺ and the increase in Fe³⁺ after use in the treatment, indicating that Fe²⁺ ions were oxidized to Fe³⁺ ions via the Fenton reaction and induced ^•OH formation [Equation (24)] [18]. Although stable Fe³⁺ can form the surface oxide layer on Cu_0.5Mn_0.5Fe₂O₄, which can impede further H₂O₂ activation, with light irradiation, Fe³⁺ can still proceed to regenerate Fe²⁺, ^•OH, and ^•OOH [Equations (12) and (25)] [25,36].

Fe²⁺+ H₂O₂ → Fe³⁺+ ^•OH + HO⁻

(24)

Fe³⁺ + H₂O₂ → Fe²⁺+ ^•OOH + H⁺

(25)

The Cu 2p_3/2 peaks at binding energy ~931.6 were identified as Cu¹⁺ while the peaks at ~933.9 were identified as Cu²⁺ [7,37] (Figure 5d). Here, a 11.5% decrease in Cu²⁺ composition after treatment was observed, signifying that the Cu¹⁺ oxidation had occurred. This change can be inferred from the continuous cyclic transformation among these two Cu states once activated with H₂O₂, resulting in ^•OH generation [Equation (2)]. Moreover, the electron transfer between these Mn, Fe, and Cu redox reactions [Equations (3)–(5)] can always favorably occur, which improves electron transfer and redox cycling of metal ions. This can lead to an oxygen vacancy and promote degradation via ^•OH generation [20,21,38].

In the O 1s spectra, three peaks at binding energy ~528.2, ~529.1 and ~530.2 eV were assigned to the surface lattice oxygen of metal oxides (i.e., Fe–O, Mn–O and Cu–O; O_latt) [12,37] (Figure 5e). The deconvoluted peak at 531.9 eV was identified as the oxygen vacancy (O_V) while the peak at 533.2 eV was attributed to the adsorbed oxygen (O_ads) and/or adsorbed molecular H₂O [13,21]. After use in the treatment, a decrease in O_latt (84.8% to 77.1%) coincided with an increase in O_V (11.8% to 16.8%), indicating that O_latt was released, possibly to maintain the electric neutrality of transformed metal ions [39,40] (Figure 5e). Once these phenomena happened, the increase in O_V could result in multiplying nanoparticle active sites for activating chemisorbed oxygen, which improved the H₂O₂ utilization rate and later facilitated the ^•OH and ¹O₂ generation [41,42].

Based on the above investigation, we proposed the possible catalytic mechanisms by pinpointing the active species generation (Figure 5f). The available H₂O₂ was initially catalyzed by metal ions to produce ^•OH, ^•OOH, and ^•O₂⁻ [Equations (2), (6), (7) and (18)–(21), (24) and (25)], followed by releasing O_latt and generating O_V. Then, the O_V and subsequence ^•OOH generation decomposed H₂O₂ molecules and yielded O_ads that could eventually convert to large numbers of oxygen bubbles being released in the solution [Equations (17), (26) and (27)] [43].

O_V + H₂O₂ → H₂O + O_ads

(26)

2O_ads → O₂

(27)

Because of its intense O₂ formation, it was transferred by the photogenerated electrons (e⁻) to produce ^•O₂⁻ [Equations (9) and (10)]. Notably, the ^•OOH was generated along the catalytic mechanisms, but its oxidation potential on pollutant molecules was doubtful, rather it may only act with available H₂O₂ as precursor to generate ^•OH, ^•O₂⁻, and ¹O₂ [Equations (7), (8) and (17)]. During the treatment, the synergistic effect of redox reactions between transformed metal ions occurred [Equations (3)–(5]) and light irradiation and/or photogenerated e⁻ also support the metal ion coexistence, which later maintained active catalytic reaction to further ROS generation [Equations (1) and (12)]. Finally, the ^•OH and ^•O₂⁻ can oxidize the pollutant molecules into intermediates and mineralization products [Equation (28)].

^•OH/^•O₂⁻ + pollutants → intermediates + CO₂ + H₂O

(28)

3.4. Reusability

To examine the Cu_0.5Mn_0.5Fe₂O₄ reusability and stability, we conducted a four-cycle experiment using OTC as a starting substrate. Results showed that the system can maintain its high OTC degradation efficiency (>98%) through all four cycles while the TOC removal efficiency was 48% at the first cycle and continued to decrease to 23% at the last cycle (Figure 6a). This signified that Cu_0.5Mn_0.5Fe₂O₄ nanoparticle photocatalytic activity and magnetization were still intact after several uses. To take a full advantage of the material magnetism, in a separate tank after the oxidation process, a magnetic catalyst collection system could be designed using a magnetic field applied underneath or adjacent to the tank wall.

After each cycle, we monitored the Mn and Cu residuals and found that both metal concentrations were high in the first cycle, which was expected for the first-use catalyst, and continued to decrease in the following cycles (Figure 6b). This coincided with the higher TOC removal efficiencies in the first two cycles. We believed that metal speciation in the solution promoted the radical formations from homogeneous Fenton reaction. However, the TOC removal efficiencies for the last two cycles remained unchanged, implying the catalyst high stability, which was able to maintain its catalytic ability by the surface metal ions.

3.5. Effect of Real Wastewater

Because the metal leaching concentration from the previous recycling experiments were slightly higher than the USEPA National Secondary Drinking Water Regulations (NSDWRs), we reduced the catalyst mass to 0.2 g/L, which is less than typically reported in other references (e.g., TiO₂, ZnO; >0.5 g/L) [44]. An advantage of using less catalyst is that it would allow deeper light penetration, reduce H₂O₂ demand, and reduce possible secondary contamination. Five different water matrices spiked with OTC were used as received, and the same experimental procedure and collecting protocol were followed (Table S2). Results showed that, at this reduced catalyst dosage, almost 100% of OTC concentrations were removed within 60 min regardless of the wastewater types (Figure 7a). The TOC removal at 60 min in each wastewater revealed that TOC can only be removed by 22% to 38% while, in the control, TOC can be removed up to 52% (Figure 7a). This indicated that the wastewater had no impact on OTC degradation as it did on the TOC. The TOC removal inhibition was possibly from the presence of turbidity in the wastewater. However, changes in turbidity were obvious after treatment, supporting our treatment system as a strong oxidation technology that is also capable of decomposing natural organic matter.

3.6. Pilot-Scale Experiment and Toxicological Evaluation

To move the technology toward commercial use, we scaled up the experimental unit by 100-fold (10 L). Pilot-scale tests were run with actual wastewater, less catalyst (minimized the catalyst loading amount by 60%) and used the treated water for further ecotoxicity testing. Results showed that turbidity was dramatically decreased within the first 2 h (85% removal) and the turbidity plateau had reached 1 nephelometric turbidity unit (NTU), which was much lower than the drinking water standard (~5 NTU) (Figure 7b). As organic matter partially corresponds to the turbidity, with this strong oxidizing process, the turbidity and dissolved organic material can then be easily removed (Figure 7c–e). We also believe that metal ion speciation in the solution and on the nanoparticle surfaces were responsible for this decrease as they existed as metal constituent ion sources (Figure 5f). These ion-exchange ions can significantly increase not only ^•OH radicals but also promote the coalescence phenomenon [45]. In addition, the iron hydroxide that may be present following our photo-Fenton catalytic process can also act as a coagulant [46]. Lastly, the external magnetic force for catalyst removal may consequently facilitate the agglomeration process between the catalyst itself and other colloidal particles, thus decreasing the turbidity. While the pre-treatment is usually responsible for excess colloids and turbidity, we provided proof that using Cu_0.5Mn_0.5Fe₂O₄ nanoparticles would also be beneficial for removing organic pollutants as well as the available turbidity.

Because only a minimal amount of treated water is usually mixed with a large amount of receiving water, we assessed the toxicological effects on the submerged macrophytes (Hydrilla verticillata) and the changes in the zooplankton community. The Hydrilla verticillata leaves were more sensitive to treated water than the stem in all conditions while the control of healthy Hydrilla verticillata revealed its intact leaves (Figure 8a). After 5 d of exposure, microscopic observation revealed thin cuticles on the epidermis layer, which likely resulted from the first contact with active radicals (Figure 8b). Change in the leaf surface morphology, loss of chlorophyll pigments, and increase in the air pocket in the epidermal cells, are even more obvious at higher treated water composition, signaling the plant stress from the residual H₂O₂ and possibly from some of the reactive oxygen species (ROS) [47]. Notably in 5% treated water, after only 10 d of exposure, we observed leaf regeneration along the entire stem length, and all old leaves were detached and substituted with the new ones. This can be explained by the fact that the available ROS and H₂O₂ can only harm the plants in the beginning phase, supporting our catalytic system as an environmentally friendly treatment technology.

Because zooplanktons are representative bio-indicators for the aquatic ecosystem and can be used to evaluate the water pollution levels, we further investigated changes in the zooplankton community and acclimatized them as a starting substrate. From the natural habitat from the wastewater (Wastewater Collection Station #3), the control (i.e., no treated water) existed four genera, three genera, and one genus from three phyla (i.e., Arthropoda, Rotifera, and Protozoa) (Figure 9a). Among these, Arthropoda was the dominant phylum, and the Cyclopoid copepod (99 Ind/L) was numerically dominant among the other zooplankton in the control experiment (Figure 9b). After exposure to 5% treated water, the Arthropoda total amount decreased by 41% and continued to decrease at higher mixed ratios (10%, 15%). Moreover, Daphnia and Moina were the sensitive genera as they drastically decreased from the community, corresponding to the previous report that the H₂O₂ concentration at >0.04 g/L can cause tissue damage to Daphnia and Moina with acute exposure [48]. Conversely, Rotifera was the dominant phylum after exposure, constituting 73% of the zooplankton abundance, which also accounted for a dramatic increase in rotifers (12 Ind/L; 0% vs. 399 Ind/L; 5%) (Figure 9a,b). Brachionus and Anuraeiopsis were the two genera with the highest tolerance to the foreign chemicals which can be explained from their persistent behavior in drastic environmental conditions as compared to other general genera in the Arthropoda phylum such as Moina, and Copepod [49,50]. As Rotifera are known to exist in the Limnosaprobity quadrant, a slight increase in dissolved oxygen from the available H₂O₂ can support the growing phase of these local rotifers. In addition, the residual H₂O₂ may be beneficial to these rotifers as it may be capable of decomposing available organic matter that was suitable for the rotifer uptake, yet not sufficient to harm rotifers. This also coincided with the sudden decrease of Brachionus and Anuraeiopsis genera as well as the smaller number of zooplankton and their number of genera at higher exposure to treated water (>5%) (Figure 9a–c).

Overall results showed that although our photocatalytic system was able to remove organic residuals, the treated water still retains some oxidative species that can have a slight impact on small organisms in aquatic ecosystems. However, this unstable H₂O₂ may not be too problematic to the environment as long as treated water is retained in the maturation tank to ensure the H₂O₂ dissipation under sunlight. Most importantly, in the real-world application, the treated water composition in the natural water body may be much less than 5% as demonstrated in the toxicological assessment in this study.

4. Conclusions

In summary, magnetic recoverable Cu_0.5Mn_0.5Fe₂O₄ nanoparticles were successfully synthesized by coprecipitation followed by melamine-assisted calcination methods. Melamine addition during the calcination process served as a coordinating agent that reduced α-Fe₂O₃ impurity and promoted the spinel ferrite structures. Cu-substitution in the MnFe₂O₄ structure effectively reduced agglomeration, thus providing a higher active surface area and also promoting photo-Fenton catalytic activities through surface Cu ions and the synergistic effect between Mn-Cu-Fe valences. By using Cu_0.5Mn_0.5Fe₂O₄, several organic pollutant concentrations decreased by 87.6% to 100% while TOC content decreased by 28.2% to 67.9% within 120 min. Quenching experiments revealed that ^•OH and ^•O₂⁻ were predominant reactive species and the photogenerated e⁻ also promoted the active radical formations. XPS analysis of fresh and used Cu_0.5Mn_0.5Fe₂O₄ nanoparticles confirmed the metal ion transformation roles in the radical formation mechanism. Cu_0.5Mn_0.5Fe₂O₄ nanoparticles showed excellent photo-Fenton catalytic ability in different water samples and high stability after use for many cycles. Our developed system can decompose not only soluble pollutants but also effectively remove colloids in water. Treated water at different composition ratios (5–15%) had a significant impact on the Hydrilla verticillata leaves at the beginning phase and leaf regeneration was observed after 10 d of exposure. The number of surviving zooplankton significantly reduced when exposed to 10% and 15% of the treated water, but rotifer numbers increased when exposed to 5% of the treated water. Overall results reveal that the Cu_0.5Mn_0.5Fe₂O₄/H₂O₂/light system is an efficient and environmentally friendly system for degrading a variety of organic contaminants.