Developing Recyclable Magnetic TiO₂-Fe₃O₄ Loading on Carbon Microtube Photocatalyst for Efficient Photodegradation of Microcystin-LR Under Visible Light

Abstract

Microcystins (MCs) are produced by cyanobacteria blooms in eutrophic water and can cause acute and chronic toxicity and even mortality to animals and humans. Previous MC removal strategies concernedonly highly contaminated water, in which the concentration of the pollutant was considerably larger than that in the natural world. In this study, we developed a composite of TiO₂-coated magnetic carbon microtube (C-TiO₂-Fe₃O₄) and used it as a photocatalyst to efficiently remove microcystin-LR (MC-LR) from water under visible light from water. And the huge surface of the carbon microtube dramatically boosted the adsorbability and charge mobility, which lowered the recombination rate of electron–hole pairs, and hence systematically enhanced photocatalytic activity. The combination of adsorption and photodegradation endowed the composite with a better performance in the removal of trace amounts of MC-LR than the C-TiO₂. It was found that increasing the contact time and catalyst dosage, acidic environment, and lower initial MC-LR concentration had positive effects on MC-LR removal. The optimum reaction conditions of C-TiO₂-Fe₃O₄ was a reaction time of 12.68 min, a catalyst dosage of 0.39 g·L⁻¹, and a pH of 7.72. The C-TiO₂-Fe₃O₄ (surface area normalized apparent reaction rate constants K/S_BET = 1.2 × 10⁻⁴) presented a higher reaction rate than C-TiO₂ (K/S_BET = 8.4 × 10⁻⁵). Moreover, the stable removal capability of C-TiO₂-Fe₃O₄ was confirmed over multiple cycles. Finally, the ecological safety performance was also evaluated after visible light illumination. This work paves the way for the development of more efficient and easily separable purifiers for the removal of pollutants and toxins from contaminated water.

1. Introduction

Microcystins (MCs), which is a kind of cyclic heptapeptide, are the most common cyanotoxins in eutrophic water, particularly their derivatives known as microcystin-LR (MC-LR) [1]. These toxins can damage the nervous system or liver and thus cause acute and chronic toxicity and even death in humans and animals [2]. Various methods have been used to remove MCs, but traditional water treatment processes (e.g., adsorption, flocculation, and advanced oxidation processes) have proven to be less effective in removing these toxins due to their chemical stability and tolerance to pH. Among the methods, adsorption has the advantages of simple operation, low cost, and high efficiency [3]. However, adsorbent regeneration and the secondary pollution are the challenges in using activated carbon [4], zeolite [5], graphene [6], fly ash [7], clay minerals [8], and so on. Advanced oxidation processes (AOPs) are a reasonable choice for the removal of organic pollutants in water, especially the heterogeneous photocatalysis degradation mediated by titanium dioxide (TiO₂) [9]. TiO₂ has been inspired in water purification with high photocatalytic efficiency and chemical stability [10]. However, the application of TiO₂ is limited by its low quantum efficiency, poor adsorptive capacity, narrow light absorption range, and difficulty in separating from water.

Over the last decade, various methods have been developed to suppress the recombination of photogenerated electron–hole pairs in TiO₂. TiO₂ nanoparticles were also loaded on carbon nanotubes (CNTs) to improve the removal of organic pollutants. Chen et al. [11] found that the microwave-induced preparation of CNTs presented high removal efficiency of methyl orange, methyl parathion, sodium dodecyl benzene sulfonate, and bisphenol A. Carbon microtubes (CMTs) have a tubular structure similar to CNTs and better electrochemical performance and stability than that of CNTs. The CMT also makes up for problems such as a small diameter, surface defects, and the serious agglomeration of CNTs. Biomass has been widely used as raw material to prepare CMTs, such as loofah sponge [12], cattail fibers [13], waste wood [14], and corncob [15]. Corncob presents a certain adsorption capacity for heavy metals and organics in sewage water [16,17]. The micron-sized tubular structure of corncob displayed good protection and stability for TiO₂, making TiO₂ distribute more uniformly with enhanced degradation efficiency. Therefore, the CMT prepared by corncob may be a superior carrier for TiO₂, which also could be used to realize the resource utilization of agricultural waste.

An additional filtration step was required to remove the catalyst in traditional TiO₂ water treatment [18]. In order to address this issue, magnetic separation has been developed into one of the most effective methods for separating and recovering TiO₂ nanoparticles from water [19]. Fe₃O₄ is one of the most widely used magnetic catalysts [20]. Jiao et al. [21] prepared core–shell Fe₃O₄ @ SiO₂/TiO₂, which could be quickly recovered under an external magnetic field. Another TiO₂-@ignocellulosic biomass @ Fe₃O₄ composites, showed high magnetic response, adsorption capacity, and charge mobility, thus reducing the recombination of electron–hole pairs and enhancing the photocatalytic performance [22]. So Fe₃O₄ could be used to enhance the photocatalytic performance and recovery rate of TiO₂.

In this study, a new type of magnetic photocatalyst that TiO₂ and Fe₃O₄ were loaded on the CMT was prepared by corncob powder (C-TiO₂-Fe₃O₄) for the degradation of MC-LR under visible light. The prepared photocatalysts were characterized by thermal gravimetric analyzer (TGA), X-ray diffraction (XRD), scanning electron microscope (SEM), Fourier transform infrared spectrometer (FTIR), Brunauer–Emmett–Teller (BET), and photocatalytic performance. The effect of pH, catalyst dosage, initial MC-LR concentration, and reaction time on the MC-LR degradation was studied. The response surface methodology (RSM) was used to optimize operating conditions of the catalytic process, and the mechanism of C-TiO₂-Fe₃O₄ that photodegrades MC-LR was studied.

2. Materials and Methods

2.1. Material

Corncob powder within 300 mesh size was purchased from Lianyungang Surui Straw Processing Plant, China. Nitric acid (HNO₃, CAS: 7697-37-2) was purchased from Shanghai Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China. Tetrabutyl titanate (CAS: 5593-70-4), FeCl₃ (CAS: 7705-08-0), FeSO₄ (CAS: 10028-22-5), methanol (CAS: 67-56-1), ethanol (CAS: 64-17-5), acetonitrile (CAS: 75-05-8), acetic acid (CAS: 64-19-7), trifluoroacetic acid (CAS: 76-05-1), isopropanol (CAS: 67-63-0), hydroquinone (CAS: 123-31-9), sodium oxalate (CAS: 62-76-0), silicon dioxide (SiO₂, CAS: 14808-60-7), humic acid (HA, CAS: 1415-93-6), fulvic acid (FA, CAS: 479-66-3), sodium hydroxide (NaOH, CAS: 1310-73-2), and hydrochloric acid (HCl, CAS: 7647-01-0) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., Shanghai, China. MC-LR and microcystin-RR (MC-RR, CAS: 111755-37-4) were purchased from Shanghai Yuanye Biotechnology Co., Ltd., Shanghai, China. Polystyrene with particle size of 70 μm (PS-70 μm) was purchased from Jiangsu Zhichuan Technology Co., Ltd., Nantong, China.

2.2. Preparation of Photocatalysts

2.2.1. Preparation of CMT

The pretreatment of corncob powder was performed according to the method in our previous work [23]. Corncob powder (8 g) was mixed with 1 mL of 4 wt.% CTAB solution and 1 mL of 2 wt.% Ni(NO₃)₂ solution to form a paste. The paste was then dried at 100 °C in an oven to yield a dry corncob paste. Meanwhile, an adhesion agent was prepared by adding 3 mL of 10 wt.% ethylene glycol to 7 mL of 66 wt.% phenolic resin ethanol solution. Five grams of the dry corncob paste was then added to 5 mL adhesion agent followed by thorough mixing to yield a ‘wet cake’ and calcined by heating for 3 h in a 160 °C oven. The product was marked as CC.

The corncob (10 g) was first mixed with 200 mL of 0.1 mol L⁻¹ HNO₃ and stirred with 24 h. After being washed with water three times, the product was dried at 100 °C in an oven to obtain the acidified corncob powder. Then, the acidified corncob powder (8 g) was mixed with 1 mL of 4 wt.% CTAB solution and 1 mL of 2 wt.% Ni(NO₃)₂ solution to form a paste, respectively. The paste was then dried at 100 °C in an oven to yield a dry corncob paste. Meanwhile, an adhesion agent was prepared by adding 3 mL of 10 wt.% ethylene glycol to 7 mL 66 wt.% phenolic resin ethanol solution. Five grams of the dry corncob paste was then added to 5 mL adhesion agent followed by thorough mixing to yield a ‘wet cake’ and dried by heating for 3 h in a 160 °C oven. The product was marked as AC.

The mixture solution of AC with 0.01 mol·L⁻¹ mixed with cetyltrimethyl ammonium bromide with a solid–liquid ratio of 1:1 was fully stirred and then dried in an oven at 60 °C for 6 h and calcined at 600 °C, 700 °C, and 800 °C for 3 h with increasing rate of 5 °C min⁻¹ in high-purity nitrogen in a muffle furnace. The corncob-based microtube (CMT) was then synthesized, collected, and marked as CMT 600, CMT 700, and CMT 800, respectively.

2.2.2. Preparation of C-TiO₂

0.01 mol of tetrabutyl titanate was added into 30 mL ethanol solution, and, then, 3 mL of 10% nitric acid was dropwise added. Then, 3 g of CMT was added and stirred for 1 h. The mixture was dried in an oven at 80 °C for 12 h and heated to 500 °C, 700 °C, and 900 °C for 3 h in the muffle furnace to obtain a double-layer photocatalyst C-TiO₂ with a loading ratio of 4%.

After determining the optimum temperature, 1 wt.%, 2 wt.%, and 8 wt.% C-TiO₂ were obtained by changing the amount of tetrabutyl titanate to 0.0025, 0.005, 0.02 mol at the optimum temperature for 3 h.

2.2.3. Preparation of C-TiO₂-Fe₃O₄

A total of 0.46 g C-TiO₂ (900 °C, 2 wt.%) was added to 30 mL of deionized water for ultrasonication for 30 min; then, 0.002 mol of FeCl₃·6H₂O and 0.001 mol of FeSO₄·7H₂O were added. The pH was adjusted to 10 with 0.1 M of NaOH, and, then, the mixture was stirred for 30 min. The obtained mixture was ultrasonicated for 30 min and then put into a hydrothermal reactor. The mixture was calcined at 150 °C for 12 h in the muffle furnace with increasing rate of 5 °C min⁻¹ and then washed with deionized water. The C-TiO₂-Fe₃O₄ was obtained after being dried at 60 °C for 6 h in the oven.

2.3. Photocatalytic Process

The prepared photocatalyst with concentration of 0.2 g·L⁻¹ was added to 10 mL of MC-LR with concentration of 500 μg·L⁻¹ and put into the photocatalytic reactor (Figure S1 in Supplement Information, SI). First, the photocatalytic system was stirred for 30 min at a speed of 500 rpm in the dark, and, then the 200 W xenon lamp was turned on for 30 min. The samples were taken every 5 min. After filtration by 0.22 μm filter membrane, the concentration of MC-LR was analyzed by high-performance liquid chromatography (HPLC, HITACHI, Chromaster, Japan). The samples were separated on a C₁₈ column (Waters, 150 × 4.6 mm). The mobile phase was acetonitrile and 0.1% trifluoroacetic acid micro-pure water. The volume ratio was eluted from the initial 70:30 to 60:40, and the wavelength was set to 238 nm. The total operation time was 15 min (flow rate: 1 mL·min⁻¹). The degradation rate of MC-LR was determined by the following equation:

$$\mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\left(\mathrm{\%}\right)={\displaystyle \frac{{C_0-C}_t}{C_0}}\times 100$$

(1)

where C₀ and C_t represent the initial concentration and residual concentration of MC-LR (μg·L⁻¹), respectively. The identification of MC-LR and its intermediates were carried out in positive ion model using high performance liquid chromatography-mass spectrum (HPLC-MS, Agilent HPLC 1260, Santa Clara, CA, USA) under the above conditions.

The effects of reaction time, pH, dosage, and initial MC-LR concentration on the adsorption and photocatalytic degradation of MC-LR were studied. The effect of HA, FA, MC-RR, SiO₂, and PS-70 μm in natural water on the photocatalytic degradation of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ were investigated, and the reaction mechanism of photocatalytic process was studied by quenching experiment using isopropanol, p-benzoquinone, and sodium oxalate.

Finally, the stability of C-TiO₂ and C-TiO₂-Fe₃O₄ in the photocatalytic process was evaluated by reusing the catalyst five times. After each operation, C-TiO₂ and C-TiO₂-Fe₃O₄ were recovered by centrifugation and magnetic separation, respectively.

2.4. Characterization Methods

The thermal stability and composition of photocatalyst were studied by TGA of catalyst, which was tested by NETZSCH TG 209 with temperature range from room temperature to 800 °C with heating rate of 2 °C min⁻¹. The phase of the prepared photocatalyst was tested by XRD with D/max 2550 X-ray diffractometer (Riguaku Smartlab, Tokyo, Japan) with Cu Kα (λ = 0.15406 nm) and diffraction angle from 3° to 70°. The microstructure of catalyst was analyzed by SEM with Hitachi, Japan, model SU 8010 with working voltage of 3 kV. Surface hydroxyl content of catalyst was determined by Boehm titration [24]: 0.5 g photocatalyst was dispersed in a mixture solution of 100 mL of ethanol and NaCl with fully stirred at 25 °C for 4 h. The pH value of the obtained suspension was adjusted to 4.0 with 0.1 M HCl or NaOH standard solution, and, then, it was further titrated from 4.0 to 9.0 with 0.1 M NaOH standard solution. The surface hydroxyl content was calculated as follows:

$$D={\displaystyle \frac{CV{N}_A\times {10}^{-3}}{W}}$$

(2)

where D (×10²⁰·g⁻¹) was the content of hydroxyl on the surface of the sample, C was the concentration of NaOH standard solution (mol·L⁻¹), V was the volume of NaOH solution (mL) consumed when the pH value was adjusted from 4.0 to 9.0, N_A was the Avogadro constant, and W was the mass of the samples (g).

The isopropanol (IPA), sodium oxalate (SO), and p-benzoquinone (BQ) were used to quench the radicals of hydroxyl radical (·OH), photogenerated holes (h⁺), and superoxide radical (·O₂⁻) in the photocatalytic reaction system. The generation of ·OH and ·O₂⁻ was detected by electron spin-resonance spectroscopy (ESR, Brucker EMXplus ER-4119-HS, Munich, Germany). The ESR of X-band at 9 GHz was obtained on a high-sensitivity vertical mode cavity. The conditions were as follows: the central field is 3510 G; scan width is 100 G; modulation frequency is 100 kHz; and microwave power is 20 mW. Reliable signal-to-noise ratio was obtained after 5–10 scans. 5, 5-dimethyl-1-pyrrolin-N-oxide (DMPO) was used as the radical trapping agent. After adding 100 μL of DMPO to photocatalyst, the catalytic reaction began for 60 min to prepare the sample. The sample was placed in a cylindrical quartz reaction unit to measure the ESR of C-TiO₂-Fe₃O₄.

The specific surface area of the photocatalyst was tested by BET (AUTOSORB-IQ-MP, US). All data were analyzed by Quantachrome ASiQWin (version 3.01). Functional groups of catalyst was analyzed by FTIR (Thermo Scientific Nicolet 6700, Waltham, MA, USA). The sample was dried at 100 °C for 10 h before analysis, and the wave number range was 400–4000 cm⁻¹. UV-3600 spectrophotometer (Shimadzu, Kyoto, Japan) was employed for UV–VIS diffuse reflection spectroscopy (DRS) at wavelengths of 200–900 nm. The concentration of leached metal ions of the photocatalyst after the degradation was measured by inductively coupled plasma atomic emission spectroscop (ICP-AES, Optima 2000, PerkinElmer Co., Ltd., USA). Then, RSM was used to optimize the operating conditions (Tables S1 and S2).

2.5. Ecological Safety Method

The rotifers were collected from the Wenruitang River (Wenzhou University, Wenzhou, China) in the school. The frozen concentrated chlorella nutrient solution was provided by the light macroalgae species. The rotifers were purified and expanded, and they were fed with the concentrated chlorella nutrient solution. Feed 1 mL of nutrient solution per liter, feed once a day, change the seawater every three days, change the water by 30%, and carry out stable continuous culture under laboratory conditions.

The experiment was divided into 6 groups, which were blank control group (CK group) only fed with concentrated algae solution, and 5 different concentration treatment groups (100 μg/L group, 200 μg/L group, 300 μg/L group, 400 μg/L group, and 500 μg/L group) of MC-LR after photocatalytic treatment by C-TiO₂-Fe₃O₄ with 100 μg/L, 200 μg/L, 300 μg/L, 400 μg/L, and 500 μg/L, respectively. Each group was set up with 3 parallels experiment. The culture container was a 2 L conical glass bottle with a water temperature of 23–25 °C, and the whole process was cultured in indoor natural light.

The 5 mL of rotifer culture solution from each conical flask was drawn each 24 h, and methanol solution for fixation was added. Then, the number of adults and larvae in the culture solution was counted and repeated 3 times.

2.6. Statistical Analysis

Statistical analysis was carried out using the SPSS 17.0 package. One-way analysis of variance (ANOVA) with a multiple post hoc (Tukey) test (significance level is 0.05) was used to check the significant difference among the different treatments. All data were shown as the mean of 3 replicates.

3. Results

3.1. Characterization

3.1.1. Chemical Composition Analysis

The thermal stability of corncob can be analyzed by TGA (Figure S2a). The mass of corncob powder decreased by 8% from room temperature to 150 °C, indicating that the adsorbed water content in corncob powder was 8%. And, then, the quality of corncob powder began to decline slowly to 190 °C, representing that the organic matter in corncob powder began to ash, and rapidly decreased to 33% in the range of 230–400 °C. In the subsequent temperature, the mass of corncob powder gradually tends to be stable and then stabilizes at about 23%, indicating that the content of organic matter in corncob powder was about 69%, and the remaining 23% was the inorganic matter that could not be ashed. According to the differential thermal analysis (DTG) curve (Figure S2b), there were two endothermic peaks in the range from room temperature to 300 °C and then four exothermic peaks with the increase in temperature to 800 °C, indicating that the ashing was an endothermic and then exothermic process. This result showed that corncob powder could remove 8% of the adsorbed water above 150 °C, and ash 69% of the organic matter above 450 °C.

The XRD patterns of the fabricated composites are shown in Figure 1a. The CC presented the amorphous phase of carbon. As SiO₂ could not react with nitric acid, the SiO₂ crystal was retained and increased significantly after the acidification in AC. There were sharp SiO₂ characteristic diffraction peaks in the XRD, which were corresponding to 2θ = 20.86° (d₁₀₀), 26.64° (d₀₁₁), 50.14° (d₂₁₁), and 59.96° (d₁₂₁) (JCPDF # 74-1811), respectively. According to the TGA and DTG results, the corncob powder was stable when the temperature was above 500 °C. Therefore, 600 °C, 700 °C, and 800 °C were selected for CMT preparation.

The characteristic diffraction peak of SiO₂ was retained in C-TiO₂ and C-TiO₂-Fe₃O₄. C-TiO₂ contained the characteristic diffraction peak of TiO₂ that 2θ = 25.30° (d₁₀₁), 37.79° (d₀₀₄), 48.04° (d₂₀₀), 53.89° (d₁₀₅), and 62.11° (d₂₁₃) (JCPDF # 86-1157), and C-TiO₂-Fe₃O₄ contained the characteristic diffraction peak of Fe₃O₄ that 2θ = 35.43° (d₃₁₁), 43.05° (d₄₀₀), and 56.94° (d₂₂₂) (JCPDF # 19-629). However, the main peak was shifted to higher 2θ when the CMT was loaded with TiO₂ or Fe₃O₄, indicating that the crystal interplanar distance of SiO₂ was reduced. In addition, the decrease in diffraction peak intensity may be due to the dilution of the crystal structure of SiO₂ by doping with TiO₂ or Fe₃O₄. The results showed that TiO₂-Fe₃O₄ were successfully loaded on the CMT, respectively.

3.1.2. Brunauer–Emmett–Teller Analysis

Figure 1b,c show the nitrogen adsorption–desorption isotherms and pore size distribution curves of fabricated materials. The specific surface area, average pore size, and total pore volume were calculated (Table S3). The specific surface areas of CC and AC samples were comparable at 0.73 and 1.63 m²∙g⁻¹, respectively. The specific surface areas of CC800 and CMT were 225.18 and 286.95 m²∙g⁻¹, which increased rapidly under high-temperature calcination, thus providing more sites for photocatalytic materials. The specific surface area decreased with the TiO₂ ratio increase from 1 wt.% to 8 wt.% (Figure 1d), which suggested that the destruction of pore structures were induced by excess TiO₂, which could also be evidenced by the decrease in pore volume. Among them, C-TiO₂ (900, 1 wt.%) and C-TiO₂ (900, 2 wt.%) had a high specific surface area, which were 349.52 and 320.96 m²∙g⁻¹, respectively. After loading with Fe₃O₄, the specific surface area of C-TiO₂-Fe₃O₄ was reduced to 224.62 m²∙g⁻¹.

3.1.3. Analysis of Surface Functional Groups

The surface functional groups of the prepared catalyst were analyzed by FTIR spectra (Figure 1f and Figure S3). For all samples, the wavenumber of 1632 cm⁻¹ corresponded to the bending vibration spectrum of H₂O [25], and the peak at 3436 cm⁻¹ indicated the bending and stretching vibrations of the hydroxyl group (-OH) [26]. According to the FTIR spectra of TiO₂, the absorption peaks at a wavenumber of 600–1000 cm⁻¹ corresponded to Ti-O-Ti bands [27]. All samples displayed absorption bands of around 710 cm⁻¹ except TiO₂, which is a peak of carbon substance. The absorption bands in the range of 875–1035 cm⁻¹ was related to the C-O stretching and ·OH deformation of the aliphatic carbon chains. The absorption band in the range of 1400–1425 cm⁻¹ can be ascribed to the C-C/C=C skeletal vibrations. The C=O stretching vibration absorption peak of AC appeared at 1635 cm⁻¹, which was caused by O-H bending vibration after the acidification of CC. C-TiO₂-Fe₃O₄ has an absorption peak at 619 cm⁻¹, which is the stretching vibration of Fe-O in the octahedral cell of Fe₃O₄. FTIR spectra confirmed the presence of TiO₂, Fe₃O₄, and carbon substance in photocatalysts.

3.1.4. DRS Analysis

A DRS device was used to determine the reflectance spectra to obtain the optical response of the C-TiO₂-Fe₃O₄ (Figure 1f). The bandgap of the samples was calculated by Equation (3):

$$E_g=hc/\lambda$$

(3)

where E_g is the bandgap energy (eV), h is the plank constant, c is the light velocity (m s⁻¹), and λ is the wavelength (nm). The DRS spectra of TiO₂, Fe₃O₄, C-TiO₂, and C-TiO₂-Fe₃O₄ are presented in Figure 1f. The DRS spectra of the nanocomposite indicated significant absorptions of C-TiO₂ and C-TiO₂-Fe₃O₄ at 420 nm and 385 nm, respectively. The bandgap energies of C-TiO₂ and C-TiO₂-Fe₃O₄ were 2.02 and 2.37 eV, respectively. According to the results, the bandgap energy of C-TiO₂-Fe₃O₄ was lower than that of TiO₂ (3.52 eV), which shows a red shift in the spectra following the C-TiO₂-Fe₃O₄ formation, extending the range of UV light to UV–VIS light [25]. Moreover, the electron–hole recombination in the C-TiO₂-Fe₃O₄ was prevented, causing an increase in the electron density. Therefore, the visible light absorption of C-TiO₂-Fe₃O₄ is improved.

3.1.5. Micromorphology Analysis

The surface morphology of CC, AC, and CMT prepared at different temperatures was analyzed. CC had a micron tubular structure with smooth inner walls (Figure 2a). After acidification, the tubular structure of AC became rough (Figure 2b) due to the organic matter that was oxidized and dehydrated. The surface folds of CC 800 were increased after high-temperature treatment (Figure 2c). However, the tubular structure was not completely destroyed, which indicates that acidification and calcination had little damage on the tubular microstructure of corncob powder. The surface of the CMT (Figure 2d–f) became rougher with the increase in temperature. Therefore, the CMT prepared at 800 °C was selected to load TiO₂ and TiO₂-Fe₃O₄ for subsequent adsorption and catalysis process.

There were many spherical particles present on the surface of the CMT, indicating that C-TiO₂ and C-TiO₂-Fe₃O₄ were successfully loaded on the CMT. When loaded with 4 wt.% TiO₂, the surface of C-TiO₂ became much rougher with the increase in temperature (Figure S4a–c), which was consistent with the morphological changes in CMTs prepared at different temperatures. When the preparation temperature was maintained at 900 °C, the spherical particles on the surface of C-TiO₂ increased with the increase in the loading of TiO₂ (1–8 wt.%). It can be seen that the particles covered the surface of the CMT with the loading of 4 wt.% (Figure S4c) and 8 wt.% (Figure S4f) but may reduce the specific surface area of C-TiO₂. However, there were less spherical particles on the surface of C-TiO₂ loaded with 1 wt.% (Figure S4d), which had no effect on the specific surface area of the CMT. Pure TiO₂ (Figure 2g) showed a smooth crystalline surface. At the same time, spherical particles and strip-like aggregates could be seen on the surface of C-TiO₂-Fe₃O₄ that was loaded with TiO₂ and magnetic Fe₃O₄ (Figure 2i and Figure S5).

3.1.6. Determination of Surface Hydroxyl Content

The surface hydroxyl content played an important role as active sites in the photocatalytic reaction. During the photocatalytic degradation, oxygen and organics were adsorbed at the active sites first. At the same time, photogenerated holes reacted with OH^- ions to generate hydroxyl radicals with strong oxidability. Therefore, the surface hydroxyl content of the catalyst directly affected the photocatalytic efficiency [28]. It can be seen that the surface hydroxyl content of AC increased greatly after acidification treatment (Table S4), and that of the CMT further increased to 2.7692 × 10²⁰ ·g⁻¹. When the CMT was loaded with TiO₂, the surface hydroxyl content increased greatly from 0.9632 × 10²⁰ ·g⁻¹ of pure TiO₂ to 4.214 × 10²⁰ ·g⁻¹ of C-TiO₂ to 3.2508 × 10²⁰ ·g⁻¹ of C-TiO₂-Fe₃O₄. It can be speculated that the C-TiO₂ may present the best photocatalytic performance due to the highest surface hydroxyl content of the catalyst.

3.2. Effect of Prepared Composites on MC-LR Degradation

The effect of prepared composites on MC-LR degradation is shown in Figure 3. MC-LR concentration remained unchanged without photocatalysts, so MC-LR could not be degraded under xenon lamp alone. In total, 7.5% of MC-LR was adsorbed by CC. The adsorption capacity of MC-LR could be further enhanced to 18% and 25% by AC and the CMT, respectively. Holes were generated after treatment by acidification and carbonization, and the increased specific surface area of the corncob resulted in enhanced adsorption performance [29]. Wu et al. [30] used ZnCl₂ to modify corncob, which obtained the dye adsorption rate of 95%. Chen et al. [31] used KOH for the chemical dehydration and carbonization of corncob, which showed that the removal rate of Cu²⁺ in simulated wastewater could reach 99.62% by the treated corncob. At the same time, the adsorption equilibrium of MC-LR was achieved after stirring for 30 min under dark conditions, and there was no significant change in the MC-LR concentration within 30 min irradiation, indicating that CC, AC, and CMTs had no photocatalytic activity.

The xenon lamp excited a short wavelength light that could not stimulate the photoactivity of TiO₂ that MC-LR concentration was adsorbed by 15% using pure TiO₂ and remained unchanged after the light was turned on. The adsorption of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ was most obvious in the first 30 min (the linear plots of MC-LR adsorption by Freundlich and Langmuir adsorption models and the kinetic linear equations by C-TiO₂ and C-TiO₂-Fe₃O₄, respectively, can be seen in Figures S7 and S8 and Tables S6 and S7 [32,33]), and 30 min was selected as the dark treatment time for subsequent experiments. After 30 min irradiation, the MC-LR degradation rate of C-TiO₂ (900 °C, 4 wt.%) was 70.1%, much higher than 46.1% of C-TiO₂ (500 °C, 4 wt.%) and 49.5% of C-TiO₂ (700 °C, 4 wt.%). When the TiO₂ loading was further increased to 4 wt.% and 8 wt.%, the degradation rate decreased to 70.2% and 59.3%, respectively, which might be due to the excessive TiO₂ particles that inhibited the absorption and light excitation by the prepared composite, which was also confirmed in SEM. The C-TiO₂ (900 °C, 2 wt.%) with the best adsorption rate and photocatalytic performance could degrade 92.4% of MC-LR, which was used in the subsequent experiments.

Compared with TiO₂, the visible light photocatalytic efficiency of C-TiO₂ was enhanced by the hybridization between the CMT and TiO₂. This was mainly due to the CMT that could be used as a photosensitizer or electron acceptor of TiO₂, which has further increased the visible light response, surface area, and pore volume of C-TiO₂ [34]. Compared with C-TiO₂, the adsorption rate of MC-LR by C-TiO₂-Fe₃O₄ decreased by 16%, and the photocatalytic rate decreased by 12% within 30 min, because the combination of TiO₂ and Fe₃O₄ brought unfavorable heterojunction that increased the recombination of electron/hole charges [35].

The HPLC chromatogram of the MC-LR removed by the CMT, C-TiO₂, and C-TiO₂-Fe₃O₄ is shown in Figure 3d. The peak intensities of MC-LR were decreased with the increase in retention time (RT), indicating that the content of MC-LR decreased continuously. Only the characteristic peak of MC-LR appeared in the HPLC chromatogram of MC-LR removed by the CMT, while other peaks appeared in C-TiO₂ and C-TiO₂-Fe₃O_4, indicating that intermediate by-products were produced in the process of photocatalytic degradation. Some peaks representing different by-products were observed: RT = 10.6, 14.1, and 14.7 min. The reduction in peak intensities showed that intermediate products were further degraded, which corresponds to the decrease in the degradation efficiency of MC-LR in the subsequent stage of photocatalysis, because MC-LR competed with its intermediate product for active sites.

The main intake routes of microcystins are through skin contact and daily drinking water. The non-carcinogenic potential risks of MC-LR content in two different intake pathways before and after the photocatalytic degradation of C-TiO₂ and C-TiO₂-Fe₃O₄ were calculated to evaluate the health risk reduction rate of the two materials. This is because there was a lack of data to support the carcinogenicity of MC-LR that was still in the preliminary exploration. The noncarcinogenic risk assessment was performed according to the USEPA health risk assessment model. The non-carcinogenic risk was the ratio of long-term daily intake (CDI) to the reference dose (RfD), which was expressed by the risk index (HI) and calculated according to Equations (S1)–(S4). The TDI (0.04 μg·(kg·d)⁻¹) of MC-LR recommended by the World Health Organization (WHO) was used as RfD [36,37]. The HI of MC-LR after degradation via human drinking water and skin was reduced by 92.4% and 99.24% by C-TiO₂, respectively (Figure S6 and Table S5). The HI of MC-LR after degradation via human drinking water and skin was reduced by 80.4% and 98.04% by C-TiO₂-Fe₃O₄, respectively. Therefore, C-TiO₂ and C-TiO₂-Fe₃O₄ can effectively reduce the health risk of MC-LR through drinking water or skin exposure. The HI value through drinking water was much higher than that through skin, indicating that drinking water containing MC-LR would increase the non-carcinogenic risk for humans. At the same time, it was found that children had the highest HI value and women had the lowest HI value. This may be due to the poor physical resistance of children and the difficulty in resisting the risk of various pollutants. Therefore, the HI value of children was much higher than that of adults under the same dose exposure conditions.

3.3. Photodegradation Dynamics

The Langmuir–Hinshelwood model (L-H model) was used to describe the relationship between the concentration of organics and chemical reaction rate in heterogeneous catalysis. The reactant molecules are first adsorbed on the active sites on the surface of the catalyst, then, the surface chemical reaction is carried out, and the final product is desorbed. The reaction mechanism was based on adsorption–surface reaction–desorption. The L-H model is given in Equation (3) [38].

$$r={\displaystyle \frac{dC}{dt}}={\displaystyle \frac{K_r{K}_{a}C}{1+K_{a}C}}$$

(4)

$$\mathrm{The}\mathrm{above}\mathrm{can}\mathrm{be}\mathrm{simplified}\mathrm{to}\ln ({\displaystyle \frac{C_0}{C_t}})=Kt$$

(5)

where t is the reaction time (min), C₀ is the initial MC-LR concentration (mg·L⁻¹), C_t is the MC-LR concentration at time of t (mg·L⁻¹), K is the reaction rate constant (min⁻¹), K_a is the catalyst surface rate constant (mg·L⁻¹·min), K_r is the adsorption equilibrium constant (mg·L⁻¹), and r is the reaction rate in the reaction system.

The degradation kinetics of C-TiO₂ and C-TiO₂-Fe₃O₄ are shown in Figure S9. With the increase in the initial MC-LR concentration, the first-order kinetic reaction rate K decreased gradually. As the initial MC-LR concentration increased from 500 to 1000 μg·L⁻¹, the K value of C-TiO₂ decreased from 0.027 min⁻¹ to 0.022 min⁻¹, and that of C-TiO₂-Fe₃O₄ decreased from 0.027 min⁻¹ to 0.019 min⁻¹. However, the surface area normalized apparent reaction rate constants (K/S_BET) were calculated to present the effect of internal structure [39]. The K/S_BET of C-TiO₂-Fe₃O₄ (1.2 × 10⁻⁴) were much higher than that of C-TiO₂ (8.4 × 10⁻⁵) in the 500 μg L⁻¹ MC-LR, which proved that C-TiO₂-Fe₃O₄ has a higher reaction rate than C-TiO₂, which may have resulted from the smaller surface area of C-TiO₂-Fe₃O₄.

3.4. Influence of Parameters on MC-LR Degradation

The parameters of pH, catalyst dosage, initial MC-LR concentration, and reaction time were selected for the single-factor experiment of C-TiO₂ and C-TiO₂-Fe₃O₄. The removal rate of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ increased in the acidic condition, which could reach 100% after 30 min of illumination with a pH less than 7 (Figure 4a,b). Lawton et al. [40] confirmed that the acidic condition was most conducive to the photocatalytic degradation of MC-LR when the pH > 2.1, which might be related to the change in surface charge of the photocatalyst, the hydrophobicity, and the net charge of MC-LR. Feng et al. [41] found that the surface hydroxyl groups of TiO₂ were protonated when the pH = 4, which could be charged with electronegative MC-LR that benefitted the photocatalytic performance of TiO₂.

The MC-LR degradation rate was increased with the catalyst dosage (Figure 4c,d). The degradation rate of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ could reach 100% after 30 min of irradiation when the dosage was more than 0.3 g·L⁻¹, which was reduced by 16% and 11% with the dosage reduced to 0.1 g·L⁻¹, respectively. This is because the increase in catalyst dosage led to more active substances (such as h⁺, e⁺, ·OH, and ·O₂⁻) formed. A shading effect could result from too much TiO₂, and the photon energy will not be fully utilized, which will result in a lower degradation rate [42].

With the increase in the initial MC-LR concentration, the MC-LR degradation rate decreased gradually, and there was competitive adsorption on the surface of C-TiO₂ and C-TiO₂-Fe₃O₄ in high MC-LR concentrations (Figure 4e,f). When the initial MC-LR concentration was 250 μg·L⁻¹, MC-LR rapidly degraded, which may be due to the rich surface active sites of C-TiO₂ and C-TiO₂-Fe₃O₄ [43]. When the initial MC-LR concentration was in the range of 500–1000 μg·L⁻¹, the MC-LR degradation rate was from 77.5% to 85.2% by C-TiO₂ and from 71.5% to 81.3% by C-TiO₂-Fe₃O₄. Until the initial MC-LR concentration reached 1250 μg·L⁻¹, the MC-LR degradation rate by the C-TiO₂ and C-TiO₂-Fe₃O₄ was reduced to 67.2% and 64.4%, respectively. With the increase in MC-LR concentration, the proportion of oxidized MC-LR to total MC-LR decreased due to the limitation active sites. The MC-LR degradation rate increased with the increase in reaction time. The degradation rate of MC-LR was increased from 82.6% to 99.1% after 30 min when the pH was 9, the initial MC-LR concentration was 500 μg·L⁻¹, and the dosage of C-TiO₂ was 0.2 g·L⁻¹. Rahimi et al. [27] found that the degradation rate of 20 mg·L⁻¹ of acid orange dye was increased from 0% to 94% when illumination time increased to 150 min. This is due to the increase of reaction timeincreases the possibility of interaction between MC-LR and electron–hole pairs [44]. In summary, C-TiO₂ and C-TiO₂-Fe₃O₄ had the same single-factor effect. The addition of Fe₃O₄ could not change the excellent photocatalytic performance of C-TiO₂, but it did improve its recovery rate.

3.5. Study on Interaction Between Parameters

The pH, catalyst dosage, and reaction time were selected for RSM analysis, which had great influence on MC-LR photocatalytic degradation. Based on the Box–Behnken model (Tables S1 and S2), the influence of the interaction between different variables on MC-LR degradation is shown in Figure 5.

The MC-LR degradation rate by C-TiO₂ and C-TiO₂-Fe₃O₄ was enhanced with increasing the catalyst dosage and decreasing the pH (Figure 5a,b). The results showed that the interaction between the catalyst dosage and pH was significant (C-TiO₂: p = 0.007 < 0.05; C-TiO₂-Fe₃O₄: p = 0.0007 < 0.05). MC-LR removal by C-TiO₂ and C-TiO₂-Fe₃O₄ was enhanced due to the decrease in pH value and the increase in reaction time (Figure 5c,d). The results showed that the p value of the interaction between reaction time and initial pH was not significant (C-TiO₂: p = 0.1591 > 0.05; C-TiO₂-Fe₃O₄: p = 0.1391 > 0.05). The increase in reaction time and catalyst dosage promoted the degradation of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ (Figure 5e,f). The results showed that the p value of the interaction between catalyst dosage and contact time was not significant (C-TiO₂: p = 0.1659 > 0.05; C-TiO₂-Fe₃O₄: p = 0.7587 > 0.05).

The RSM showed that the optimum conditions for C-TiO₂ were a reaction time of 11.06 min, a catalyst dosage of 0.34 g·L⁻¹, and a pH of 7.69; the optimum conditions for C-TiO₂-Fe₃O₄ were a reaction time of 12.68 min, a catalyst dosage of 0.39 g·L⁻¹, and a pH of 7.72. The optimum conditions were validated when the MC-LR removal rate by C-TiO₂ and C-TiO₂-Fe₃O₄ could reach above 99% (Figure S10).

3.6. Effects of Natural Water Bodies

The representative natural organic matter (NOM) in surface water mainly includes HA and FA [45]. With the increase in HA and FA dosage, the degradation rate of MC-LR decreased (Figure S11). However, the MC-LR degradation rate was less affected when the concentration of HA and FA was less than 8 mg·L⁻¹. When the concentration of HA increased to 10 mg·L⁻¹, the MC-LR degradation rate by C-TiO₂ and C-TiO₂-Fe₃O₄ decreased by 24% and 31.5%, respectively, while the degradation rates decresed by 32.9% and45.6% with an FA of 10 mg·L⁻¹, respectively. Therefore, the inhibitory effect of FA was more obvious than that of HA. It was confirmed that, when NOM reaches a certain concentration, the photocatalytic degradation of organic pollutants was inhibited due to the competition of HA and FA for the surface active sites of the photocatalysts [46]. In addition, the dissolved HA and FA also reduced the visible light transmitted to the surface of the photocatalysts [47]. At the same time, HA and FA were reactive oxygen species (ROS) free radical scavengers that could effectively stop the photocatalytic process [48].

In this study, MC-RR was added to simulate the MC-LR degradation in the presence of multiple toxins (Figure S11e,f). The degradation of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ was enhanced by 0.9% and 0.4% when MC-RR concentration was 100 μg·L⁻¹, respectively, which might be due to the synergistic removal effect of MC-LR and MC-RR in the low MC-RR concentration. The MC-LR degradation rate by C-TiO₂ and C-TiO₂-Fe₃O₄ decreased to 81.8% and 77.1% with the increase in MC-RR concentration to 400 μg·L⁻¹, respectively, and to 79.3% and 58.8% when the MC-RR concentration increased to 500 μg·L⁻¹, respectively. This indicated that the presence of high MC-RR concentration inhibited the removal of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ compared with the degradation process without MC-RR. This might be related to the competitive reaction between MC-RR and MC-LR.

Then, SiO₂ and general purpose polystyrene with a particle size of 70 μm (PS-70 μm) were added to simulate the effect of particles presented in natural waters on the photocatalysis of C-TiO₂ and C-TiO₂-Fe₃O₄ (Figure S12). SiO₂ can reduce the MC-LR degradation rate of C-TiO₂ from 97.5% to 72.5% and that of C-TiO₂-Fe₃O₄ from 85.4% to 40.6%. The relative inhibition value of PS-70 μm was lower, which decreased to 80.9% for C-TiO₂ and 62.8% for C-TiO₂-Fe₃O₄. The reason for it is that the presence of particles in the water will have a scattering and shielding effect on light, reducing the light energy utilization of C-TiO₂ and C-TiO₂-Fe₃O₄ and thereby affecting the degradation rate of the target degradation product [49]. On the other hand, PS had a small particle size and could be photodegraded, so its shading effect was lower than that of SiO₂ [50].

In summary, NOM, MC-RR, and particles in natural water could inhibit the MC-LR degradation by C-TiO₂ and C-TiO₂-Fe₃O₄. In the low NOM concentrations (<10 mg·L⁻¹), the MC-LR degradation rate by C-TiO₂ and C-TiO₂-Fe₃O₄ was less affected. Synergistic degradation between MC-LR and MC-RR occurred in the C-TiO₂ and C-TiO₂-Fe₃O₄ photocatalytic system in low MC-RR concentrations, but the inhibition effect occurred in the high MC-RR concentration. The addition of particles also inhibited the photocatalytic effect of C-TiO₂ and C-TiO₂-Fe₃O₄.

3.7. Quenching Experiment

The formation of active substances such as hydroxyl radical (·OH), electron hole (h⁺), and superoxide anion radical (·O₂⁻) plays an important role in the photocatalytic process. Therefore, the main radical species of the C-TiO₂ and C-TiO₂-Fe₃O₄ photocatalytic degradation of MC-LR were determined by the active species quenching experiment (Figure 6a,b). IPA (0.01 mol·L⁻¹), SO (0.01 mol·L⁻¹), and BQ (0.01 mol·L⁻¹) were used as quenchers of ·OH, h⁺, and ·O₂⁻, respectively.

After adding IPA, the MC-LR degradation rate decreased from 97.85% to 29.69% by C-TiO₂ and from 85.22% to 28.59% by C-TiO₂-Fe₃O₄, respectively. After adding SO and BQ, the MC-LR degradation rate was slightly decreased trend. DMPO was further used as an electron-trapping agent for radicals of C-TiO₂-Fe₃O₄ in the ESR test. There is a quadruple peak with an intensity ratio of 1:2:2:1 (Figure S13), which is strong evidence for ·HO, and there is also a characteristic peak of ·O₂⁻ with weaker intensity than that of ·HO. Therefore, the results showed that ·OH may be the main active substance for MC-LR photocatalytic degradation by C-TiO₂-Fe₃O₄, not h⁺ and ·O₂⁻. Moreover, a CMT with high adsorption capacity could make MC-LR molecules closer to the surface of the catalyst and improve the reaction rate of generated ·OH and MC-LR.

3.8. Catalyst Stability and Reusability

The stability of MC-LR degradation by C-TiO₂ and C-TiO₂-Fe₃O₄ was evaluated by repeating experiments. It was found that both two catalysts showed stable activity (Figure 6c,d). After the fifth operation, the MC-LR degradation rate by C-TiO₂ decreased slightly from 85.23% to 84.32% and from 97.87% to 96.11% by C-TiO₂-Fe₃O₄. In addition, XRD patterns showed the similar structures of C-TiO₂ and C-TiO₂-Fe₃O₄ before and after photocatalytic degradation (Figure 6e,f). The peak positions of C-TiO₂ and C-TiO₂-Fe₃O₄ were identical, and only slight shifts in peak intensities were observed, which further demonstrated that C-TiO₂ and C-TiO₂-Fe₃O₄ possessed reusable and sustainable removal performance in multiple removal processes. The XRD peak shifts may be related to the following two reasons: after being adsorbed on the surface of the photocatalyst, the water molecules formed hydrogen bonds with the surface hydroxyl groups, which would cause a certain degree of relaxation of the surface lattice, or the catalyst surface may form some carbon-containing intermediate products in the photocatalytic process. These intermediate products would change the structure of the catalyst surface, which would eventually lead to a slight shift in the XRD peak position. In addition, the ICP-AES analysis of the recovered catalysts after five cycles showed that the loss of metal elements from the catalyst was negligible (Figure S14a). Thus, the slight changes after the degradation of MC-LR indicated that both catalysts showed good stability. Due to the magnetic response of C-TiO₂-Fe₃O₄, it could be easily recovered by an external magnetic field, which showed the advantages of a high recovery rate in practical application (Figure S14b). The above results showed that C-TiO₂ and C-TiO₂-Fe₃O₄ presented good stability, and no photocorrosion occurred during the MC-LR degradation by C-TiO₂-Fe₃O₄, which has broad application prospects in water treatment.

3.9. MC-LR Intermediates and Potential Pathways

The reaction intermediates of MC-LR were analyzed by HPLC-MS. The possible intermediates, oxidative groups, and bond energy sites were studied. After 20 min and 30 min of light irradiation, at least 28 intermediate products can be identified in the total ion chromatogram (TIC), as shown in Figures S15 and S16. Among them, 13 intermediates were produced by the hydrolysis of peptide bonds, while the other 15 intermediates were typical photocatalytic oxidation products. So the possible degradation pathways of MC-LR in the C-TiO₂-Fe₃O₄ photocatalytic procedure were proposed (Figure 7).

The hydrolysis of the MC-LR peptide bond was rapid. First, the intermediate product m/z = 856 was produced by the loss of the D-Arg fragment (Path 1), and, then, it was hydrolyzed by the peptide bond connecting Ala and D-MeAsp-Leu to form product m/z = 618. The main hydrolysate m/z = 618 contained residues of Glu, Mdha, and Ala. The conjugated diene in the Adda side chain was pronged to oxidative hydroxylation but remained intact in the products of m/z = 468 and 452. This is because the photocatalytic system preferentially hydrolyzes the peptide bond, followed by carboxyl oxidation. The product m/z = 860 contains the hydroxylated benzene ring of Adda (Path 2). The products m/z = 622 and m/z = 468 were produced by the continuous catalytic hydrolysis of peptide bonds connecting D-Leu-MeAsp with Arg and D-Arg with Glu.

The oxidation products in the photocatalytic process, such as m/z = 844, 840, 325, and 228 were also detected without the hydrolysis of peptide bonds. The product of m/z = 1029 was produced by the dihydroxylation of the unsaturated bond of Adda by ·OH attack [51]. However, dihydroxylation is not important for the degradation of MC-LR in this studied system. The product of m/z = 683 was not hydrolyzed by peptide bonds but was an oxidation product formed by the hydroxyl carbonylation of the Arg in the product of m/z = 699 (Path 3). Most of the oxidation products, such as m/z = 844 and 699, were not similar to the oxidation products produced by typical photocatalysis, so the degradation of this system involves alternating degradation pathways. The intermediate product fragments that did not appear in the ·OH-dominated MC-LR degradation process can help explain the key role of catalytic hydrolysis in the effective utilization of visible light irradiation by C-TiO₂-Fe₃O₄. The final products of m/z = 336 and 147 produced by the hydrolysis of peptide bonds in MC-LR were more difficult to remove than MC-LR. This indicates that the catalyst preferentially interacts with the complete ring structure of MC-LR, which is composed of seven amino acids connected by peptide bonds. The first 20 min of C-TiO₂-Fe₃O₄ photocatalysis is mainly through hydrolysis to open the macrocycle, and with further decomposition in 10 min, resulting in a significant increase in small molecular fragments in 30 min. Opening the ring structure of MC-LR by photocatalysis can significantly reduce its toxicity. Therefore, the selectivity of the cyclic structure in the first 20 min of C-TiO₂-Fe₃O₄ raises good application prospects for the photocatalytic system.

3.10. Mechanism

The photocatalytic activity is affected by many factors such as the surface basic site, band gap, oxidation potential of the photogenerated holes, and separation efficiency of photogenerated electrons and holes [52]. Wang et al. [51] have shown that the carbon microtube has much larger interlayer spacing than that of commercial graphite, which might facilitate ion transport. Therefore, the higher photocatalytic rate of C-TiO₂ and C-TiO₂-Fe₃O₄ might be due to the higher mass transfer rate of carbon microtubes. In the C-TiO₂-Fe₃O₄ system, the surface basic site was determined to be 3.251 × 10²⁰ ·g⁻¹, which is smaller than that of C-TiO₂ (4.214 × 10²⁰ ·g⁻¹), so C-TiO₂ could provide more active sites for the photocatalytic reaction process.

Furthermore, the conduction (E_CB) and valence band (E_VB) position of the photocatalyst could be calculated through the following equation [53]:

$$E_{CB}=X-E_e+0.5E_g$$

(6)

$$E_{VB}=E_{CB}+E_g$$

(7)

where X is the Mulliken’s electronegativities, E_e is the energy of free electrons on the hydrogen scale (≈4.5 eV), and E_g is the band gap.

Chen et al. [54] estimated the E_VB of TiO₂-Fe₃O₄ to be 2.78 eV, which is 0.12 eV lower than that of TiO₂ (2.9 eV) [55]. The E_VB of a photocatalyst can promote the reaction of the generated holes with OH⁻ and produce active ·OH radicals. This may be the reason for the high activity of C-TiO₂ and C-TiO₂-Fe₃O₄. The difference between the E_VB potential of C-TiO₂ and C-TiO₂-Fe₃O₄ is 0.525 eV, indicating that the activity of C-TiO₂ is higher than that of C-TiO₂-Fe₃O₄. Among other wide band gap photocatalysts with good photocatalytic performance, such as In(OH)₃ (E_VB~4.2 eV) and CaSb₂O₅(OH)₂ (E_VB~4.1 eV) [32], the difference between In(OH)₃ and CaSb₂O₅(OH)₂ in E_VB is much larger. On the other hand, the Fe²⁺ generated by Fe₃O₄ can consume the electrons generated by photoexcitation and reduce the recombination of e⁻/h⁺ pairs [33,56]. Under the acidic condition in water, Fenton-like catalytic activity also contributes to the degradation of pollutants [57,58]. Moreover, the incorporation of carbon materials can improve the regeneration of Fe₃O₄ nanoparticles. According to the classic photocatalytic theory, it is reasonable to believe that the loading of Fe₃O₄ reduces the separation efficiency of electron–hole pairs due to the lower photocatalytic rate of C-TiO₂-Fe₃O₄.

Based on the above discussion, the possible mechanism of the photocatalytic degradation of MC-LR by C-TiO₂ and C-TiO₂-Fe₃O₄ was proposed, as shown in the abstract figure. When exposed to the simulated sunlight, the valence band electrons of TiO₂ and TiO₂-Fe₃O₄ can be excited into the conduction band. The electrons in the uppermost valence band of TiO₂ or TiO₂-Fe₃O₄ can be excited to the conduction band. The electrons in the upper valence band will jump to the conduction band of TiO₂ or TiO₂-Fe₃O₄. However, the holes in the valence band will quickly participate in the MC-LR degradation. As a result, the combination of electrons and holes will be reduced. Therefore, the improvement of photocatalytic performance could be attributed to the ability of the carbon microtube to hinder the recombination of electron–hole pairs. The loading of Fe₃O₄ would also reduce the recombination of electron–hole pairs, thus enhancing the photocatalytic efficiency of C-TiO₂-Fe₃O₄.

3.11. Ecological Safety

The effect of the solution after photocatalytic MC-LR by C-TiO₂-Fe₃O₄ on the population density of rotifers is shown in Figure 8a. The results presented that the population density of each experimental group was basically unchanged. When cultured for 1 d, there was a significant difference between the CK group and 100 μg/L group (p > 0.05), and the population density was similar. When cultured for 3 days, there was no significant difference between the CK group and the 100 μg/L group, 200 μg/L group, 300 μg/L group, 400 μg/L group, and 500 μg/L group (p > 0.05), and the average population density of the five concentration treatment groups was slightly higher than that of the CK group. When cultured for 7 days, there was still no significant difference between the CK group and the five treatment groups (p > 0.05). The overall population density was not less than 35 ind./mL within 7 days of sampling.

Figure 8b shows that there was no significant difference in the number of births of rotifers within 7 days compared with the CK group (p > 0.05). By calculating the ratio of the number of offspring to the number of survivals, the rate of birth (population fertility) can be obtained as shown in Figure 8c. On the first day, the number of births decreased with the increase in concentration in the five treatment groups, indicating that the high concentration showed a stronger inhibitory effect on the number of births of rotifers, while the inhibitory effect at low concentration was weak, but it showed no significant difference compared with the CK group (p > 0.05). At 3 d and 7 d, the birth rate did not show significant changes between the groups and tended to be stable at about 19%.

The population density and reproductive capacity of rotifers in each treatment group at sampling time points did not show significant differences from the CK group, which proved that the photocatalytic degradation products of MC-LR by C-TiO₂-Fe₃O₄ had no ecological toxicity effect on rotifers and would not affect their reproduction.

4. Conclusions

In this study, C-TiO₂-Fe₃O₄ was prepared from corncob to the photodegradation of MC-LR. SEM and XRD confirmed that TiO₂-Fe₃O₄ was successfully loaded on the CMT. RSM results showed that the optimum conditions were as follows: a reaction time of 12.68 min, a catalyst dosage of 0.39 g·L⁻¹, and a pH of 7.72 for C-TiO₂-Fe₃O₄. The radial of ·OH was the main active species in its photocatalytic reaction that could attack the conjugated double bond of the Adda moiety of MC-LR first. The results showed that C-TiO₂-Fe₃O₄ with high surface area normalized apparent reaction rate constants and photocatalytic efficiency could be used to effectively degrade MC-LR in polluted water, and it can be easily recovered and reused by applying external magnetic field. The photocatalytic degradation products of MC-LR by C-TiO₂-Fe₃O₄ had no ecological toxicity effect on rotifers. So the C-TiO₂-Fe₃O₄ was an efficient and safety recyclable magnetic photocatalyst that could photodegrade MC-LR under visible light.