Photocatalytic Degradation of Deoxynivalenol Using Cerium Doped Titanium Dioxide under Ultraviolet Light Irradiation

Deoxynivalenol (DON) is a major mycotoxin with high toxicity that often contaminates grains, foods and feeds. The traditional approaches for DON removal are difficult to meet industry and agriculture demands due to the high stability of the DON molecule. Therefore, there is an urgent need to develop green and effective strategies for DON degradation. In this study, a batch of photocatalytic nanomaterials of cerium (Ce) doped titanium dioxide (TiO2) were successfully prepared by sol-gel method. The catalysts were systematically characterized by XRD, HRTEM, FT-IR, UV-Vis and XPS. The catalyst 0.5Ce-TiO2 showed superior photocatalytic activity for DON degradation in aqueous solution under ultraviolet light irradiation, better than that of traditional photocatalyst pure TiO2, and 96% DON with initial concentration of 5.0 mg/L could be degraded in 4 h. In addition, the two possible degradation intermediate products C5H8O3 and C17H18O6 were identified, the photocatalytic degradation mechanism and degradation pathway were studied. The results indicate that Ce doped TiO2 photocatalyst can be used to reduce DON effectively.


Introduction
Deoxynivalenol (DON), a high-toxicity secondary metabolite produced by Fusarium graminearum, is one of the most common mycotoxins in grains [1], foods and feeds [2]. This mycotoxin poses a serious threat to human health and animals [3,4].
Some technologies have been employed to eliminate DON. Physical methods such as washing and grinding can reduce the content of DON in contaminated grains, but DON is transferred to some other by-products. Heat treatment for DON removal is very limited, due to the strong thermal stability of DON molecule [5]. Bretz et al. reported that the DON molecule could be decomposed under alkaline conditions and then neutralized by chemical agents [6]. Biological detoxification methods could efficiently reduce the toxicity of DON under mild conditions [7]. Yin et al. found a strain A16 isolated from wheat fields, and identified as Devosia sp., which could survive and degrade DON in several conditions [8]. Although biological methods have good detoxification effects, the disadvantage is relatively long treatment time, which limits its further application. Therefore, developing a safe and efficient strategy for DON degradation is the trend in food and other related industries and agricultures.
In recent years, more and more researchers have paid increasing attention in the field of photocatalytic technology with high efficiency, low energy consumption and mild reaction conditions for DON degradation [9]. Oxide semiconductor modified doping materials can effectively degrade DON under light conditions, such as ZnO, TiO 2 , etc. Wang et al. prepared photocatalytic materials dendritic-like α-Fe 2 O 3 , which could reduce 90.3% DON with initial concentration of 4.0 µg/mL in an aqueous solution [10]. Bai et al. found the photocatalyst graphene/ZnO hybrids GZ0.3 prepared by a simple one-step hydrothermal method showed good ability to degrade DON in an aqueous suspension under ultraviolet light (UV) light irradiation. Zhou et al. reported a nanoparticles photocatalyst Tm@TiO 2 composite (UCNP@TiO 2 ), which was employed to reduce DON under the simulated sunlight [11]. Additionally, Wu et al. utilized upconversion nanoparticles@TiO 2 composites UCNP@TiO 2 to degrade DON in wheat.
Compared to other oxide semiconductor photocatalysts, TiO 2 as a non-toxic and environmentally friendly nanomaterial with good catalytic degradation of harmful substances has received extensive attention [12]. However, the wide band gap energy of TiO 2 with about 3.2 eV reduces the efficiency of light utilization and weakens its photocatalytic degradation ability [13]. Metal doping is one of the effective control strategies to enhance the photocatalytic activity by reducing the band gap energy of TiO 2 [14], which also could inhibit rapid recombination of photogenerated electron-hole pairs and broaden light absorption range redshift due to doping ions in TiO 2 , making it possible to form complexes with Lewis bases such as organic acids or alcohols [15]. Cerium (Ce), as a rare earth element, is relatively nontoxic and cheaper than other rare earth metals. It was doped with TiO 2 to form a new composite catalysts and can effectively reduce the band gap energy of materials and have better catalytic degradation performance [16].
In this study, we successfully synthesized photocatalytic nanomaterials Ce doped TiO 2 by sol-gel method. The obtained 0.5Ce-TiO 2 showed superior photocatalytic activity for DON removal under ultraviolet light irradiation (λ = 254 nm) in an aqueous solution compared with traditional photocatalyst pure TiO 2 . In addition, the possible degradation intermediate products were identified, and the photocatalytic degradation mechanism and degradation pathway were studied.

HRTEM Images and XRD Analysis
In order to study the microstructure and crystal phase composition of Ce doped TiO 2 catalytic materials, High Resolution Transmission Electron Microscope (HRTEM) were used to characterize the prepared catalytic materials. X-ray diffraction (XRD) was conducted to determine the phases and crystal structures of pure TiO 2 , xCe-TiO 2 (x = 0. 5,1,5,10,20,40) and CeO 2 . The results are shown in Figure 1. Figure 1a shows the morphology of the pristine TiO 2 particles with tiny spherical shapes. Ce doped TiO 2 shapes appearance becomes uniform and regular with doping Ce, shown in Figure 1b-h. Figure 1j displays the clear lattice fringes of 0.5Ce-TiO 2 with the particle size of approximately 24 nm. The interplanar distance of approximately 0.35 nm corresponds to the (101) crystal plane of anatase TiO 2 , and the lattice fringe spacing is about 0.19 nm, corresponding to the (220) crystal plane of anatase TiO 2 . Then, the particle size of TiO 2 doped with 0.5% Ce is about 24 nm, as seen in Figure 1i. Figure 1i shows the morphology of the pristine CeO 2 particles. Figure 1i shows that the distinct diffraction peaks at 25 dominated by anatase, and no other phases can be observed, which is consistent with the HRTEM images. The results are different from previous reports, which showed the formation of the cerium titanate Ce 2 Ti 2 O 7 phase with Ce:Ti = 50% in molar ratio [17]. Most of the Ce ions cannot enter the TiO 2 lattice because the radius of Ce 4+ is larger than that of Ti 4 , and Ce dopant have been well dispersed on the surface of TiO 2 as the form of cerium oxide [18], which may be the reason why CeO 2 phases cannot be observed. In addition, the Ce doping expands the TiO 2 lattice, causing a large lattice distortion and strain energy. In order to compensate the lattice stress, the oxygen atoms on the TiO 2 surface escape the lattice to form oxygen vacancy, playing a role of trapping holes. The effect may contribute to reduce the probability of recombination of holes and electrons in TiO 2 and to increase the photocatalytic activity. The main peak (101) broadens with the increase of Ce content, and even disappears, suggests Ce doping inhibits the crystal growth and decreases the crystallinity of the synthesized materials.

XPS Analysis
The high resolution XPS spectra were performed to analyze and determine the chemical state of the synthesized Ce-doped TiO 2 photocatalyst. All the obtained spectra were calibrated to the C 1s electron peak at 284.6 eV. The results are shown in Figure 2.

FT-IR and UV-Vis DRS Analysis
Functional groups on the surface of the TiO2, CeO2 and Ce-TiO2 catalyst were analyzed by Fourier Transform Infrared Spectrometer (FT-IR). The UV-vis diffuse reflectance spectra of the materials in the range of 200-800 nm were recorded. The results are shown in Figure 3.
We can see from Figure 3(a) that the three main characteristic absorption peaks of TiO2 are at 490, 1634 and 3425 cm −1 , respectively, and the three characteristic absorption peaks correspond to the stretching vibration of the Ti-O-Ti bond in the TiO2, and the bending vibration and stretching vibration of the O-H bond in the water molecules present on the surface [26][27][28][29]. Comparing the infrared absorption curves of TiO2 doped with Ce, the TiO2 characteristic absorption peaks in the composites doped with Ce are all present, and the absorption strength of the peaks at 3425 cm −1 is enhanced. The results are consistent with the previous reports [29]. Moreover, in the infrared absorption curve of the Ce-TiO2, the characteristic absorption peak corresponding to the TiO2 at 490 cm −1 is shifted, and the position of the peak appears at about 513 cm −1 . This may be because Ce enters the lattice of TiO2 to form Ti-O-Ce bonds, causing effect on the stretching vibration of the original Ti-O-Ti bond. From the figure, we can also see that there is an absorption peak around 2345 cm −1 corresponding to CO2. The characteristic absorption peak indicates that ethanol did not completely volatilize during the synthesis of the material, and the remaining small amount of inorganic carbon was oxidized to CO2 during drying.  . O and Ti are derived from the synthesized photocatalyst, while C 1s is derived from the anhydrous ethanol solvent added during the synthesis process, which fails to volatilize completely during drying and calcination and remains on the surface of the sample. However, it also shows that since there is low doping Ce content, the characterized peak of the Ce element is very weak. Figure 2b shows the high resolution XPS spectra of Ti 2p for the synthesized sample 0.5Ce-TiO 2 and pure TiO 2 . The TiO 2 exhibits two peaks at binding energies 458.58 eV (Ti 2p3/2) and 464.28 eV (Ti 2p1/2), confirming Ti mainly in the Ti 4+ chemical state [19]. Compared with pure TiO 2 , the two binding energies of Ti 2p after doping Ce show a slight shift for both the two spin orbitals Ti 2p3/2 (458.68 eV) and Ti 2p1/2 (464.48 eV), which indicates a strong interaction of Ti and Ce species [20]. The results suggest the changes in the Ti oxidation states (from Ti 4+ to Ti 3+ ) and confirm that the Ce element is successfully doped into the TiO 2 structure. Figure 2c shows the O 1s high resolution spectra. The characteristic peak at 529.88 eV in pure TiO 2 is assigned to the crystal lattice oxygen O 2− . The main O 1s peak position for 0.5Ce-TiO 2 is slightly shifted to lower binding energy around 530.08 eV. The reason for peak shifts in O 1s and Ti 2p can be explained by the fact that transferring electrons from O 1s and Ti 2p orbitals to Ce 4f orbitals causes change in the charge densities of the O and Ti atoms [21]. Figure 2d shows the high resolution XPS spectrum of Ce 3d for the synthesized sample 0.5Ce-TiO 2 . The spin orbitals coupling states of 3d5/2 and 3d3/2 are labelled with v and u, respectively. The XPS spectrum of the Ce 3d is relatively complex, mainly due to the hybridization of O 2p and Ce 4f orbital electrons and the partial occupation of the 4f [21]. Hence, the spectrum is categorized into ten constituents. The binding energies of Ce 3d5/2 at 880.06, 881.67, 882.58, 885.66 and 899.39 eV are labelled with v 0 , v, v , v and v , while Ce 3d3/2 at 901.22, 903.62, 908.08, 904.12 and 913.74 eV are labelled with u 0 , u, u , u and u [22]. The peaks at v 0 , v , u 0 and u are characteristic binding energies of Ce 3+ configurations between the O 2p level and Ce 4f level. v /u is related to the Ce(3d 9 4f 2 ) (O 2p 5 ) final state, and v 0 /u 0 is assigned to the Ce(3d 9 4f 1 ) (O 2p 6 ) final state [23]. The peaks at v, v , v , u, u and u" are attributed to Ce 4+ . v /u is related to the primary photoemission from Ce(3d 9 4f 0 ) (O 2p 6 ) final state, and v/u is related to the Ce(3d 9 4f 1 ) (O 2p 5 ) final state [24]. The v /u is from the transfer of two electrons from the O 2p orbital to an empty Ce 4f orbital with the Ce(3d 9 4f 1 ) (O 2p 5 ) final state [25]. It is deduced that the surface of 0.5Ce-TiO 2 is not fully oxidized due to the presence of Ce 4+ /Ce 3+ , and the Ce-O-Ti bond may be formed at the interstitial sites or interfaces between CeO 2 and Ce 2 O 3 , though their contents are too small to be detected by XRD [25].

FT-IR and UV-Vis DRS Analysis
Functional groups on the surface of the TiO 2 , CeO 2 and Ce-TiO 2 catalyst were analyzed by Fourier Transform Infrared Spectrometer (FT-IR). The UV-vis diffuse reflectance spectra of the materials in the range of 200-800 nm were recorded. The results are shown in Figure 3. According to the spectra, TiO2 mainly absorbs ultraviolet (UV) light, and its maximum wavelength is 393 nm. The maximum absorption wavelength of all the Ce-TiO2 catalyst samples is shifted to the visible range of 400-600 nm, which expands the range of its absorption spectrum. Moreover, with the increase of Ce content, the red shift of the catalyst samples is more obvious. Among all Ce-TiO2 catalyst samples, the red shift of doped metal Ce content is the largest when the content of Ce is 40%. The absorbance of pure TiO2 is from the electron transition from O 2p to Ti 3d state. The red shift of Ce-doped TiO2 may be caused by the formation a new electronic state, which reduces the distance of charge transfer between 4f electrons of Ce ions and the conduction or valence band in the TiO2 bandgap, enhancing the photocatalytic activity [30]. Li et al. have ascribed to Ce 4+ /Ce 3+ ions as electron scavengers to trap the electrons of TiO2 and the Ce 4f level as an interfacial charge transfer and elimination of electron-hole recombination [24].
Comparing with all catalyst samples in the UV region of 200-400 nm, it is found that the doping of Ce can improve the absorbance of TiO2, and the absorbance of TiO2 increases with the increase of the doping amount of Ce. Band gaps estimated from Tauc transformations of the absorbance spectra are shown in Figure 3(c). The band gap (Eg) is calculated mainly by the following formula [31]: where hν represents photon energy, α represents absorption coefficient, Eg represents band gap energy and K is a constant. It can be seen from Figure 3(c) that the band gap width of the synthesized 0.5Ce-TiO2 sample is 2.9 eV, less than that of TiO2 (Eg = 3.2 eV).
The results indicate that doping Ce in TiO2 significantly reduces the band gap energy and We can see from Figure 3a that the three main characteristic absorption peaks of TiO 2 are at 490, 1634 and 3425 cm −1 , respectively, and the three characteristic absorption peaks correspond to the stretching vibration of the Ti-O-Ti bond in the TiO 2 , and the bending vibration and stretching vibration of the O-H bond in the water molecules present on the surface [26][27][28][29]. Comparing the infrared absorption curves of TiO 2 doped with Ce, the TiO 2 characteristic absorption peaks in the composites doped with Ce are all present, and the absorption strength of the peaks at 3425 cm −1 is enhanced. The results are consistent with the previous reports [29]. Moreover, in the infrared absorption curve of the Ce-TiO 2 , the characteristic absorption peak corresponding to the TiO 2 at 490 cm −1 is shifted, and the position of the peak appears at about 513 cm −1 . This may be because Ce enters the lattice of TiO 2 to form Ti-O-Ce bonds, causing effect on the stretching vibration of the original Ti-O-Ti bond. From the figure, we can also see that there is an absorption peak around 2345 cm −1 corresponding to CO 2 . The characteristic absorption peak indicates that ethanol did not completely volatilize during the synthesis of the material, and the remaining small amount of inorganic carbon was oxidized to CO 2 during drying.
According to the spectra, TiO 2 mainly absorbs ultraviolet (UV) light, and its maximum wavelength is 393 nm. The maximum absorption wavelength of all the Ce-TiO 2 catalyst samples is shifted to the visible range of 400-600 nm, which expands the range of its absorption spectrum. Moreover, with the increase of Ce content, the red shift of the catalyst samples is more obvious. Among all Ce-TiO 2 catalyst samples, the red shift of doped metal Ce content is the largest when the content of Ce is 40%. The absorbance of pure TiO 2 is from the electron transition from O 2p to Ti 3d state. The red shift of Ce-doped TiO 2 may be caused by the formation a new electronic state, which reduces the distance of charge transfer between 4f electrons of Ce ions and the conduction or valence band in the TiO 2 bandgap, enhancing the photocatalytic activity [30]. Li et al. have ascribed to Ce 4+ /Ce 3+ ions as electron scavengers to trap the electrons of TiO 2 and the Ce 4f level as an interfacial charge transfer and elimination of electron-hole recombination [24].
Comparing with all catalyst samples in the UV region of 200-400 nm, it is found that the doping of Ce can improve the absorbance of TiO 2 , and the absorbance of TiO 2 increases with the increase of the doping amount of Ce. Band gaps estimated from Tauc transformations of the absorbance spectra are shown in Figure 3c. The band gap (E g ) is calculated mainly by the following formula [31]: where hν represents photon energy, α represents absorption coefficient, E g represents band gap energy and K is a constant. It can be seen from Figure 3c that the band gap width of the synthesized 0.5Ce-TiO 2 sample is 2.9 eV, less than that of TiO 2 (E g = 3.2 eV). The results indicate that doping Ce in TiO 2 significantly reduces the band gap energy and effectively inhibits the recombination of photogenerated electrons-holes.

Evaluation of the Effectiveness of DON Degradation
The effectiveness of Ce-TiO 2 photocatalytic degradation DON under UV light (λ = 254 nm) and the total organic carbon (TOC) changing trends during the reaction of photocatalytic degradation DON in aqueous solution using 0.5Ce-TiO 2 are shown in Figure 4. by Ce doping play a co-catalytic effect on DON degradation. The removal effects of 5Ce-TiO2 and 10Ce-TiO2 are equivalent to pure TiO2. However, the degradation rate of 10Ce-TiO2 and 40Ce-TiO2 are lower than that of TiO2. The main reason may be that excessive Ce destroy the TiO2 lattice structure and decrease the photoactivity of the nanomaterials, although the light absorption range redshifts [32]. The results indicate that 0.5Ce-TiO2 has superior photocatalytic activity for DON removal under UV light (λ = 254nm) irradiation, which is even better than that of the traditional photocatalyst TiO2. In the course of the degradation, the measurement of TOC is an important embodiment of the mineralization reaction and the mineralization degree. DON aqueous solution with initial concentrations of 1 mg/L and 5 mg/L were selected here. It can be seen from Figure(b) that the TOC values are gradually decreasing with prolonging the photocatalytic reaction, which means that the mineralization rate of the DON aqueous solution is increasing with the reaction time increasing. The results indicate that more and more DON molecules are further oxidized to H2O and CO2 completely. The TOC/TOC0 ratio of 5 mg/L is less than that of 1 mg/L resulted from more molecules involved in the degradation reaction in the high concentration of the DON aqueous solution.

Free Radical Trapping Experiments and Photocatalytic Degradation Mechanism
To explore the main active substances in the process of DON degradation, the active material capture experiments were carried out. We choose EDTA-2Na as hole (h + ) scavenger [33], tert-butanol as the hydroxyl radical (•OH) [34] and nitrogen (N2) bubbling was used to superoxide radicals (•O2 -) trapping agents. The results are shown in Figure 5(a). The rapid photocatalytic degradation of DON over 0.5Ce-TiO 2 nanomaterials was clearly seen by using High Performance Liquid Chromatography (HPLC), as shown in Figure 4a. The HPLC chromatograms show decreasing DON peaks at retention time 2.74 min with prolonging irradiation time. The Figure 4c shows the different photocatalytic degradation effects of Ce-TiO 2 with different Ce doping content and pure TiO 2 . The optimum doping amount of Ce is 0.5% in our study. The degradation rate of the DON aqueous solution at 5 mg/L can reach 96% using 0.5Ce-TiO 2 under UV light irradiation after 240 min. The photocatalytic degradation effect is higher than 85% under the same conditions. The results indicate that the small CeO 2 particles produced on the TiO 2 particles caused by Ce doping play a co-catalytic effect on DON degradation. The removal effects of 5Ce-TiO 2 and 10Ce-TiO 2 are equivalent to pure TiO 2 . However, the degradation Toxins 2021, 13, 481 7 of 13 rate of 10Ce-TiO 2 and 40Ce-TiO 2 are lower than that of TiO 2. The main reason may be that excessive Ce destroy the TiO 2 lattice structure and decrease the photoactivity of the nanomaterials, although the light absorption range redshifts [32]. The results indicate that 0.5Ce-TiO 2 has superior photocatalytic activity for DON removal under UV light (λ = 254 nm) irradiation, which is even better than that of the traditional photocatalyst TiO 2 .
In the course of the degradation, the measurement of TOC is an important embodiment of the mineralization reaction and the mineralization degree. DON aqueous solution with initial concentrations of 1 mg/L and 5 mg/L were selected here. It can be seen from Figure 4b that the TOC values are gradually decreasing with prolonging the photocatalytic reaction, which means that the mineralization rate of the DON aqueous solution is increasing with the reaction time increasing. The results indicate that more and more DON molecules are further oxidized to H 2 O and CO 2 completely. The TOC/TOC 0 ratio of 5 mg/L is less than that of 1 mg/L resulted from more molecules involved in the degradation reaction in the high concentration of the DON aqueous solution.

Free Radical Trapping Experiments and Photocatalytic Degradation Mechanism
To explore the main active substances in the process of DON degradation, the active material capture experiments were carried out. We choose EDTA-2Na as hole (h + ) scavenger [33], tert-butanol as the hydroxyl radical (•OH) [34] and nitrogen (N 2 ) bubbling was used to superoxide radicals (•O 2 − ) trapping agents. The results are shown in Figure 5a. The photocatalytic degradation mechanism for DON removal using Ce-TiO2 nanomaterials synthesized in this experiment for DON are shown in Figure 5(b). The electrons in the valence band (VB) of the Ce-TiO2 sample are excited to the conduction band (CB) under the irradiation of UV light. The number of holes on VB is the same as the number of electrons on CB. Under general conditions, photogenerated electrons-hole carriers are easily inclined to recombine, resulting in only a small part of the electrons involved in the catalytic degradation process [37]. The band position of doped samples is mainly calculated by the following formulas [38]: where EVB, ECB, χ and the Eg are the VB edge potential, CB edge potential, Sanderson electronegativity and the band gap of the photocatalysts. The value of χ for TiO2 is 5.81 eV, and E e represents the free electron energy on the hydrogen scale, with a value of 4.5 eV. According to the above formula, the ECB of Ce-TiO2 is -0.14 eV and the EVB is 2.76 eV. In this material, Ce dopants into the TiO2 lattice introduce new impurity levels (empty Ce 4f) with a smaller band gap close to the Ti 3d conduction band of TiO2 [32]. Under light illumination, the distance of the excited charge carrier transfer from Ti 3d of TiO2 to Ce 4f level is narrowed [39], which can reduce the charge carrier's recombination rate [40]. In addition, the photogenerated electrons on the conduction band react with O2 to form •O2 -, and holes in VB react with DON to form CO2 and H2O. The reaction formula is: From Figure 5a, the photocatalytic activity of 0.5Ce-TiO 2 decreases largely by the addition of hole scavenger (EDTA-2Na), while no significant decrease was observed by the addition of •OH scavengers, indicating that •OH are not the main oxidative species affecting catalytic degradation. That is, the hole plays a more key role in the photocatalytic degradation reaction than •OH in the UV light irradiation [35]. In addition, the degradation efficiency of DON using 0.5Ce-TiO 2 is obviously reduced with the anoxic solution, indicating that O 2 is another more important role in the photodegradation reaction that produces more •O 2 − , which is consistent with the previous study [36]. Based on all the results above, we can conclude that the photooxidation mechanism occurring on the surface of 0.5Ce-TiO 2 may involve in the direct oxidizing reaction of DON with •O 2 − and holes. The photocatalytic degradation mechanism of the 0.5Ce-TiO 2 sample may involve the oxidation process of holes and •O 2 − . The photocatalytic degradation mechanism for DON removal using Ce-TiO 2 nanomaterials synthesized in this experiment for DON are shown in Figure 5b catalytic degradation process [37]. The band position of doped samples is mainly calculated by the following formulas [38]: where E VB , E CB , χ and the E g are the VB edge potential, CB edge potential, Sanderson electronegativity and the band gap of the photocatalysts. The value of χ for TiO 2 is 5.81 eV, and E e represents the free electron energy on the hydrogen scale, with a value of 4.5 eV. According to the above formula, the E CB of Ce-TiO 2 is −0.14 eV and the E VB is 2.76 eV. In this material, Ce dopants into the TiO 2 lattice introduce new impurity levels (empty Ce 4f) with a smaller band gap close to the Ti 3d conduction band of TiO 2 [32]. Under light illumination, the distance of the excited charge carrier transfer from Ti 3d of TiO 2 to Ce 4f level is narrowed [39], which can reduce the charge carrier's recombination rate [40].
In addition, the photogenerated electrons on the conduction band react with O 2 to form •O 2 − , and holes in VB react with DON to form CO 2 and H 2 O. The reaction formula is: . We speculate that P1 and P2 structures are from a possible unstable intermediates (IP1) six-membered ring compound via opening loop with C1-C2 bond fractures and by removing the five-membered ring. The five-membered ring with a 12,13-epoxy group is one of the main toxic functional groups in DON [41]. The reaction can eliminate the toxicity of DON. However, we did not find that the carbonyl-containing intermediate product ion (m/z 303.09) formed by the epoxy ring group destroyed, which is different from the previous study [42] due to the strong oxidizing ability. Then, IP1 dehydrates to form the relatively stable compound IP2. Two possible reaction pathways may exist. One is that IP2 continue to be oxidized to form P1. The other one is that the two IP2 molecules are coupled to form P2. The intensity changing trends of the molecular ion peak of [DON+H] + , [P1+H] + and [P2+H] + with the reaction time are shown in Figure 6c. Because TOC content gradually decreases with the reaction time increasing, DON is further oxidized and mineralized to CO 2 and H 2 O completely. The possible pathway that intermediate products P1 and P2 generated from DON molecular break, continued to react and finally disappeared with the photocatalytic degradation reaction is speculated and shown in Figure 6d. molecules are coupled to form P2. The intensity changing trends of the molecular ion peak of [DON+H] + , [P1+H] + and [P2+H] + with the reaction time are shown in Figure 6(c). Because TOC content gradually decreases with the reaction time increasing, DON is further oxidized and mineralized to CO2 and H2O completely. The possible pathway that intermediate products P1 and P2 generated from DON molecular break, continued to react and finally disappeared with the photocatalytic degradation reaction is speculated and shown in Figure 6(d).

Conclusions
In conclusion, TiO 2 photocatalytic nanomaterials doped with Ce were successfully prepared by the sol-gel method. In these synthesized materials, 0.5Ce-TiO 2 shows superior photocatalytic activity for DON removal in aqueous solution under UV light irradiation (λ = 254 nm). The free radical trapping experiments indicate that the photogenerated h + and •O 2 − are the two main active substances for DON photocatalytic degradation. The two possible degradation intermediate products C 5 H 8 O 3 (m/z 117.07) and C 17 H 18 O 6 (m/z 319.12) were identified, which indicates that the main toxic groups in the DON molecule were destroyed. This work provides an efficient and mild method to reduce DON contamination. In order to further evaluate the feasibility of this method, the study on toxicity of DON degradation products is ongoing.

Synthesis of Ce-TiO 2 , TiO 2 and CeO 2
Ce-dropped TiO 2 (Ce-TiO 2 ) based catalysts were prepared using the sol-gel method. In the study, tertbutyl titanate (TBOT) and cerium nitrate hexahydrate (Ce(NO 3 ) 3 •6H 2 O) were used as the reaction precursors. The typical synthetic procedure was as follows: first, 15 mL TBOT and little amounts of acetic acid were dissolved in 30 mL absolute ethanol to make the solution A. The solution B was cerium nitrate hexahydrate aqueous solution. Solution B was dropped into solution A slowly with vigorous stirring. The mixed solution continued to stir, and the solution became sol and gel. The gel stood for 12 h, and then was dried at 120 • C in an oven; finally, it was calcined at 550 • C for 3 h. The product was Ce-TiO 2 . Using the same synthesis procedure, the product with different proportion was obtained by only changing the content of cerium nitrate (the Ce contents were varied as Ce: Ti = 0.5, 1, 5, 10, 20 and 40% in molar ratio) and was marked as 0.5Ce-TiO 2 , 1Ce-TiO 2 , 5Ce-TiO 2 , 10Ce-TiO 2 , 20Ce-TiO 2 and 40Ce-TiO 2 , respectively. In addition, pure TiO 2 nanomaterials were synthesized without adding cerium nitrate as a control. The CeO 2 catalyst was also prepared by sol-gel method [43].

Characterization of Catalysts
The crystal structure of the sample was determined by X-ray diffraction (Bruker D8 ADVANCE, Germany), with Cu Kα radiation as the X-ray source, and operated at 40 kV and 40 mA. The 2θ scan range was 20-85 • with a step size of 0.02 • . Morphology and structure of catalysts were observed using high-resolution transmission electron microscopy instruments (JEM-2100F, Japan). X-ray photoelectron spectroscopy (XPS) studies were performed with a Escalab 250Xi spectrometer (ThermoFisher, MA, USA), using a monochromatic Al Kα source. The infrared spectra were recorded on the Nicolet iS5 FT-IR Spectrometer (Thermo Scientific, WI, USA). The UV-Vis diffuse reflectance spectra of 200-800 nm were recorded on the Shimadzu UV 3600plus (Shimadzu, Japan).

DON Photocatalytic Tests
Photocatalytic tests were carried out in photochemical reaction apparatus, and the tests were performed using a UV lamp (254 nm). The initial concentration of DON solution was 5 mg/L. Before the light irradiation, 2.5 mg catalyst was added to 20 mL of DON solution and kept in the dark for 30 min to reach the adsorption-desorption equilibrium. After different irradiation times (30 min, 60 min, 90 min, 120 min, 180 min and 240 min), 1 mL of suspension was collected and centrifuged at high speed. The upper liquid was filtered through a 0.22 µm filter membrane and was analyzed. The concentration of DON was determined by Acquity Ultra Performance LC (Waters, Milford, MA, USA) equipped with a Waters Acquity BEH C18 column (1.7 µm, 2.1 × 100 nm) and with an isocratic mobile phase composed of methanol-water (20:80) at a flow rate of 0.25 mL/min. The DON degradation rate was calculated by the following formula. η(%) = (C 0 − C t )/C 0 × 100 (8) where C 0 represents the initial concentration of DON, and C t represents the concentration of DON after photocatalytic degradation. Seven DON aqueous solutions of 20 mL with the same initial concentration 1 mg/L and 5 mg/L were, respectively, irradiated at different irradiation times (0 min, 30 min, 60 min, 90 min, 120 min, 180 min and 240 min) and directly submitted to TOC analysis. The TOC content was determined by Total Organic Carbon Analyzer TOC-L CPH Basic System (Shimadzu Co. Ltd., Kyoto, Japan).
To investigate the photocatalytic mechanism of DON degradation, the active species trapping experiment was performed with three different active substance capture agents