A Comparative Study of Microcystin-LR Degradation by UV-A, Solar and Visible Light Irradiation Using Bare and C / N / S-Modiﬁed Titania

: In an endeavor to tackle environmental problems, the photodegradation of microcystin-LR (MC-LR), one of the most common and toxic cyanotoxins, produced by the cyanobacteria blooms, was examined using nanostructured TiO 2 photocatalysts (anatase, brookite, anatase–brookite, and C / N / S co-modiﬁed anatase–brookite) under UV-A, solar and visible light irradiation. The tailoring of TiO 2 properties to hinder the electron–hole recombination and improve MC-LR adsorption on TiO 2 surface was achieved by altering the preparation pH value. The highest photocatalytic e ﬃ ciency was 97% and 99% with degradation rate of 0.002 mmol L − 1 min − 1 and 0.0007 mmol L − 1 min − 1 under UV and solar irradiation, respectively, using a bare TiO 2 photocatalyst prepared at pH 10 with anatase to brookite ratio of ca. 1:2.5. However, the bare TiO 2 samples were hardly active under visible light irradiation ( < 25%) due to a large band gap. Upon UV, solar and vis irradiation, the complete MC-LR degradation (100%) was obtained in the presence of C / N / S co-modiﬁed TiO 2 with a degradation rate constant of 0.26 min − 1 , 0.11 min − 1 and 0.04 min − 1 , respectively. It was proposed that the remarkable activity of co-modiﬁed TiO 2 might originate from its mixed-phase composition, mesoporous structure, and non-metal co-modiﬁcation.


Introduction
Cyanobacteria (blue-green algae) are naturally present in aquatic environments. The increase in nutrient concentration, global water temperature, and sunlight intensity results in cyanobacterial blooms, named harmful algal blooms (HABs) [1]. It should be pointed that, besides a decrease in ecosystem stability, HABs might also cause a production of highly active toxic compounds, known as cyanotoxins, during cell lysis, which is of special concern for drinking water sources [1][2][3]. Moreover, the presence of cyanotoxins can increase the chemical oxygen demand, microbial growth, and disinfection photocatalytic efficiency compared to single-modified TiO 2 [1,40,[42][43][44][45][47][48][49][50]. However, there are only a few reports focusing on the photodegradation of MC-LR over mesoporous A/B TiO 2 nanoparticles. In our previous reports, non-metal co-modified mesoporous anatase/brookite TiO 2 was prepared and used for photocatalytic degradation of cyanotoxins and pharmaceuticals [1,40,47,48]. For example, C/N-co-modified mesoporous anatase/brookite TiO 2 photocatalysts were highly active for MC-LR degradation under vis irradiation [1]. The impacts of initial pH value, the TiO 2 content, and MC-LR concentration on the photocatalytic activity were also investigated. It was found that the complete degradation (100%) of MC-LR (10 mg L −1 ) was achieved, using co-modified TiO 2 (0.4 g L −1 ) at pH 4 under visible light irradiation. Continuously, in this study, a facile method based on tuning the phase content and surface area of bare TiO 2 nanoparticles to improve the photocatalytic degradation of MC-LR was investigated. It was found that both pristine TiO 2 (prepared at pH 10), and C/N/S co-modified TiO 2 (synthesized by a simple method in which the best TiO 2 sample was calcined with thiourea) exhibited an efficient performance for the decomposition of MC-LR (C 0 = 10 mg L −1 , pH 4) during only 15 and 60-min irradiation with UV-A and solar simulation, respectively. In addition, the non-metal co-modified TiO 2 showed 4× higher photocatalytic activity than bare TiO 2 for MC-LR degradation during 3 h-vis irradiation.

Characterization of TiO 2 Photocatalysts
Six titania samples were used in this study, i.e., five non-modified samples (named as S1, S2, S3, S4 and S5) and one C/N/S-co modified sample (S4 modified with carbon, nitrogen and sulphur; named as CNS-S4). Phase structure and morphology (X-ray diffraction (XRD), and field emission scanning electron microscopy (FE-SEM); Figures 1 and 2), textural properties (specific surface area and particle size; Table 1), absorption properties (ultraviolet-visible diffuse reflectance spectroscopy (DRS); Figure 3), surface chemical characterization (X-ray photoelectron spectroscopy (XPS); Figure 4), and the photoluminescence (PL) properties of the samples were investigated. The preparation conditions, phase composition, crystallite size, specific surface area, particle size, absorption edge, and band gap of the samples are summarized in Table 1.
Catalysts 2019, 9, x FOR PEER REVIEW 4 of 16 the band gap value decreased from 3.17 to 2.9 eV by non-metal modification (see Figure 3 and Table  1). The XPS spectroscopy revealed that the CNS-S4 sample was modified with C, N, and S (20.78% Ti, 58.47% O, 16.41% C, 2.5% S and 1.84% N), while the S4 sample was non-modified TiO2 (see Figure  4). Figure 4a shows the XPS survey spectra for S4 and CNS-S4 samples. Figure 4b displays an XPS spectrum of CNS-S4 for C 1s. Three peaks were observed with binding energies of 284.8, 286.5, and 289 eV, which were ascribed to C−O and C=O, O=C-O, Ti-O-C, and C-N bonds (289 and 286.5 eV), and C−C and C−H bonds (284.8 eV) [1,40,42,43,45,47,48]. One peak with binding energy of 401 eV was obtained for nitrogen (N 1s), which was attributed to interstitial N-doping (Ti-O-N and Ti-N-O linkage), substitutional N-doping (O-Ti-N linkage), hyponitrite species, and chemisorbed N species (NO, N2O, NO 2-, and NO 3-) (see Figure 4c) [1,28,34,[40][41][42][43][44][45]47,48,50]. Figure 4d gives XPS spectrum of CNS-S4 for sulphur (S 2p) with binding energy of 168.6 eV, which might be assigned to S 6+ 2P3/2 [40,42,45,46,48]. The substitution of Ti 4+ by S 6+ is much easier and more favorable than the replacement of O 2− with S 2− [40,48].      The XRD spectra revealed that the preparation pH value (i.e., glycine/NaOH volume ratio) controlled the phase structure. The single-phase anatase (A) and brookite (B) were formed at pH 3 and 11, respectively, whereas, with pH value ranging from 5 to 10, A/B-mixed-phase TiO 2 powders were formed with a decrease in the anatase content through increasing the pH value (see Figure 1a and Table 1). Moreover, the phase structure did not change by non-metal modification, whereas the phase composition changed (anatase content increased), as shown in Figure 1b and Table 1. The crystalline size of anatase and brookite decreased within increasing of pH values (see Table 1). The specific surface area increased with increasing pH value till pH 10, and then decreased at pH 11 ( Table 1). The co-modified TiO 2 (CNS-S4) and non-modified (S4) possessed a mesoporous structure, whereas the microporous structure appeared in all other samples, as shown in Table 1. Figure 2a-c shows FE-SEM for S1, S4, S5 and CNS-S4 samples, indicating that the single-phase anatase and brookite contain nano-quasi-spherical-like, and nano-spindle-like particles, respectively. In contrast, the mixed-phase TiO 2 contained nano-quasi-spherical-like anatase mixed with nano-rod-like brookite. As displayed by UV-Vis spectroscopy, the absorption was red-shifted for the co-modified sample, reflecting that the band gap value decreased from 3.17 to 2.9 eV by non-metal modification (see Figure 3 and Table 1). The XPS spectroscopy revealed that the CNS-S4 sample was modified with C, N, and S (20.78% Ti, 58.47% O, 16.41% C, 2.5% S and 1.84% N), while the S4 sample was non-modified TiO 2 (see Figure 4). Figure 4a shows the XPS survey spectra for S4 and CNS-S4 samples. Figure 4b displays an XPS spectrum of CNS-S4 for C 1s. Three peaks were observed with binding energies of 284.8, 286.5, and 289 eV, which were ascribed to C−O and C=O, O=C-O, Ti-O-C, and C-N bonds (289 and 286.5 eV), and C−C and C−H bonds (284.8 eV) [1,40,42,43,45,47,48]. One peak with binding energy of 401 eV was obtained for nitrogen (N 1s), which was attributed to interstitial N-doping (Ti-O-N and Ti-N-O linkage), substitutional N-doping (O-Ti-N linkage), hyponitrite species, and chemisorbed N species (NO, N 2 O, NO 2-, and NO 3-) (see Figure 4c) [1,28,34,[40][41][42][43][44][45]47,48,50]. Figure 4d gives XPS spectrum of CNS-S4 for sulphur (S 2p) with binding energy of 168.6 eV, which might be assigned to S 6+ 2P 3/2 [40,42,45,46,48]. The substitution of Ti 4+ by S 6+ is much easier and more favorable than the replacement of O 2− with S 2− [40,48].

Removal of MC-LR by Adsorption and Photolysis
The experiments of MC-LR removal in the dark (absence of light) were conducted to determine the extent of MC-LR adsorption on the TiO2 surface. Figure 5a shows the change in MC-LR concentration vs. adsorption time in the presence of non-modified TiO2 (S1, S2, S3, S4, and S5), and non-metal-co-modified TiO2 (CNS-S4) catalysts. It was indicated that 3-h stirring resulted in 41%,

Removal of MC-LR by Adsorption and Photolysis
The experiments of MC-LR removal in the dark (absence of light) were conducted to determine the extent of MC-LR adsorption on the TiO 2 surface. Figure 5a shows the change in MC-LR concentration vs. adsorption time in the presence of non-modified TiO 2 (S1, S2, S3, S4, and S5), and non-metal-co-modified TiO 2 (CNS-S4) catalysts. It was indicated that 3-h stirring resulted in 41%, 39%, 25%, 18%, 16%, and 14% adsorption of MC-LR on the surface of CNS-S4, S4, S3, S2, S1, and S5, respectively (see Figure 5). More efficient MC-LR adsorption on the surface of non-metal co-modified TiO 2 sample (CNS-S4) compared to the non-modified samples ( Figure 5) might result from either the non-metals presence or mesoporous structure, which act as active sites for pollutants adsorption and hence improves the adsorption capacity [51][52][53][54][55]. Among the non-modified TiO 2 , the S4 catalyst possessed the highest adsorption capacity because of its largest specific surface area (see Figure 6a).
(a) (b) The direct photolysis experiments of MC-LR by UV-A, solar, and visible irradiation were carried out (after obtaining adsorption equilibrium) in an aqueous solution (Figure 6b). It was found that the MC-LR degradation by photolysis was negligible, indicating that MC-LR could not easily be removed by irradiation alone, whether UV, solar, or visible. Indeed, many reports demonstrated the radiation alone (UV, solar and visible) is not effective for elimination of MC-LR [1,3,[26][27][28][56][57][58][59][60][61][62][63][64]. From the above results, it was concluded that a more efficient method should be applied for MC-LR degradation, e.g., the photocatalysis (photocatalyst + light). above results, it was concluded that a more efficient method should be applied for MC-LR degradation, e.g., the photocatalysis (photocatalyst + light).
In order to reveal the superior catalytic activity of S4 and CNS-S4 compared to the rest of the bare TiO2 samples (S1, S2, S3, and S5), the PL spectra with excitation wavelength of 300 nm, for all samples were conducted, and the results are shown in Figure 11. It was found that the PL spectra peaks of all samples are around 470 nm with different intensities, which decrease in the following order: CNS-S4 ˂ S4 ˂ S3 ˂ S2 ˂ S1 ˂ S5 (see Figure 11). The higher PL intensity of single-phase TiO2 (S1 (A), and S5 (B)) demonstrated their lower photocatalytic activity compared to mixed-phases TiO2 (S2, S3, S4, and CNS-S4). These findings could be explained by the fact that hole trapping by non-metals and synergistic effect between anatase and brookite might facilitate the electron-hole separation [1,31,39,47,48].        The remarkable photocatalytic performance of S4 (among the non-modified samples) and CNS-S4 (among all samples) might be explained by the phase structure and composition (S4 and CNS-S4), mesoporous structure (S4 and CNS-S4), high specific surface area (S4), and non-metal co-modification (CNS-S4) [1,14,28,31,34,[36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51][53][54][55]. Both S4 and CNS-S4 contained anatase/brookite mixed-phase, and this might result in efficient separation of charge carriers by their possible migration between two phases (PL; Figure 11) [1,31,[36][37][38][39][40]47,48]. The mesoporous structure of both S4 and CNS-S4 is also favorable for efficient activity because of providing more active sites on TiO 2 surface, the accumulation of hydroxyl radicals inside the mesopores, high dispersion of mesoporous TiO 2 in the aqueous solution, and rapid diffusion of MC-LR to the active sites on the surface of the mesoporous TiO 2 photocatalyst (as also proved by high adsorption ability) [1,31,39,40,47,48,65,66]. More efficient adsorption and degradation of MC-LR under UV-A and solar light over the S4 sample than that on the other non-modified TiO 2 samples could result from larger specific surface area, as shown in Figure 9. In order to correct the photocatalytic activity of bare TiO 2 samples (S1, S2, S3, S4, and S5) considering the surface area, the reaction rate constant was normalized by a specific surface area and summarized in Table 2. The highest photocatalytic degradation rate and normalized reaction rate constant, based on surface area, were observed for the S4 sample (with the highest surface area). Therefore, it was concluded that the surface area can play an important role in the photocatalytic degradation of MC-LR. Higher surface area introduces more active sites on the photocatalyst surface, enhancing the adsorption of organic pollutants, and it might also lead to a high concentration of surface hydroxyl groups, which can trap the photogenerated holes and thus decrease the electron-hole recombination, as demonstrated by PL results (see Figure 11) [1,31,39,40,47,66]. From the above results and discussion, it is reasonable to hypothesize that the preparation pH value is a key factor that affects directly the photocatalytic activity of the TiO 2 nanoparticles. Indeed, it is known that the pH value might control the surface characteristics and the size of aggregated nanoparticles resulting in higher content of formed hydroxyl radicals, and improved adsorption capacity of organic pollutants onto a photocatalysts surface [67]. Therefore, our results demonstrating the role of preparation pH value for controlling of the phase composition and boosting the photocatalytic activity of TiO 2 nanoparticles is consentient with previous reports [68,69]. It was also proposed that the highest activity of the co-modified TiO 2 photocatalyst (CNS-S4) results from a non-metals presence, which could act as active sites for efficient MC-LR adsorption and degradation [51][52][53][54][55]. It is known that non-metal modification is responsible for the increase of superficial hydroxyl groups' content and faster electron transfer, and thus higher photocatalytic activity [1,14,28,[42][43][44][45]47].  Figure 11. PL spectra of S1, S2, S3, S4, and CNS-S4 samples. Wavelength /nm S1 S2 S3 S4 S5 CNS-S4 Figure 11. PL spectra of S1, S2, S3, S4, and CNS-S4 samples.

Materials and Methods
Although the results clearly showed much higher activity of CNS-S4 than other samples under solar irradiation, the reason could not be unequivocally decided, i.e., favorable surface properties, best phase composition (A/B = 2.5) or non-metal presence. Therefore, the photodegradation experiments under visible light irradiation were performed for two samples of the highest activity (S4 and CNS-S4), and obtained results are shown in Figure 10a. It was found that the non-modified TiO 2 photocatalyst, which could absorb only UV light (λ < 400 nm), was almost inactive (<25%) under visible light irradiation, and this low MC-LR degradation might be assigned to slight light transmittance below the wavelength of 420 nm [3,28]. In contrast, MC-LR was completely degraded after 180-min stirring under visible light irradiation in the presence of C/N/S-co-modified TiO 2 (CNS-S4), as shown in Figure 10a. Similar to degradation under UV-A and solar irradiation, photocatalytic degradation of MC-LR followed first-order kinetics, as shown in Figure 10b. The rate constants and reaction rates were: K = 0.002 min −1 , r = 0.170 × 10 −4 mmol L −1 min −1 and K = 0.037 min −1 , r = 3.2 × 10 −4 mmol L −1 min −1 for S4 and CNS-S4 photocatalysts, respectively. Therefore, it was proposed that the structure and specific surface area were not the reasons for the remarkable photocatalytic activity of CNS-S4 sample under visible light irradiation. It should be pointed that these two samples possessed almost the same morphology, i.e., nano-rod-like brookite with nano-quasi-spherical-like anatase (see FE-SEM images: Figure 2c,d). In addition, although the specific surface area of CNS-S4 (30.00 m 2 g −1 ) was two times lower than that of S4 (62.3 m 2 g −1 ), the reaction rate of CNS-S4 was ca. 20× higher than that of S4. Therefore, it could be concluded that non-metal modification played a major role in enhancing the activity of CNS-S4 under the vis range of a solar spectrum, as clearly indicated from sample characteristics. It was proposed that the O 2p orbitals of TiO 2 could overlap with orbitals of C, N and S, forming mid-gap levels between conduction band (CB) and valence band (VB), and thus narrowing the bandgap of CNS-S4 [40,45,47,48,50]. Additionally, the carbonaceous species, which may act as a photo-sensitizer like organic dyes, might be formed on the surface of TiO 2 by carbon modification [70], whereas nitrogen might convert some Ti 4+ to Ti 3+ by charge compensation, and thus form donor energy levels below the conduction band [71]. The O 2p of TiO 2 could be substituted by nitrogen atoms to form isolated impurity energy levels above the valence band [70]. The photoluminescence (PL) is often a useful tool for investigating separation/recombination of photogenerated charges in semiconductors since the PL emission intensity is related directly to the electron-hole recombination rate. Lower PL emission intensities correspond to more efficient electron-hole separation and hence improve the photocatalytic activity of the photocatalyst [1,47,48]. In order to reveal the superior catalytic activity of S4 and CNS-S4 compared to the rest of the bare TiO 2 samples (S1, S2, S3, and S5), the PL spectra with excitation wavelength of 300 nm, for all samples were conducted, and the results are shown in Figure 11. It was found that the PL spectra peaks of all samples are around 470 nm with different intensities, which decrease in the following order: CNS-S4 < S4 < S3 < S2 < S1 < S5 (see Figure 11). The higher PL intensity of single-phase TiO 2 (S1 (A), and S5 (B)) demonstrated their lower photocatalytic activity compared to mixed-phases TiO 2 (S2, S3, S4, and CNS-S4). These findings could be explained by the fact that hole trapping by non-metals and synergistic effect between anatase and brookite might facilitate the electron-hole separation [1,31,39,47,48].

Preparation and Characterization of TiO 2 Photocatalysts
The preparation and characterization of non-modified TiO 2 , and C/N/S-co-modified TiO 2 were described in our previous works [39,48]. In short, to synthesize the non-modified TiO 2 , with different anatase/brookite ratios, the NaNO 3 as an oxidizing agent was added to aqueous Ti 2 (SO 4 ) 3 solution, and then stirred till formation of the colorless solution. After that, different volume ratios of glycine/NaOH were added dropwise to the colorless solution resulting in different pH values. After stirring for 25 min, the suspension was transferred into a 100-mL Teflon-lined tube and heated at 200 • C for 20 h. Ultimately; the TiO 2 catalyst was collected, rinsed several times with DI-H 2 O and alcohol, and dried at 60 • C. The samples prepared at different pH values of 3, 5, 7, 10 and 11 were named as S1, S2, S3, S4 and S5, respectively.
The N/C/S-co-modified mesoporous A/B TiO 2 photocatalyst was prepared by an ex-situ method. Shortly, the non-modified A/B mixed-phase TiO 2 powder (S4) was grounded with thiourea with a 1:1 weight ratio, and then put into a muffle furnace (Nabertherm with controller) and calcined at 450 • C for 1 h. The obtained sample was washed several times and then dried at 60 • C for 10 h. The yellow powder was signed as CNS-S4.

Photocatalytic Tests
A stock aqueous solution of MC-LR (10 mg L −1 ) was prepared (pH 6.3). Prior to the photocatalytic experiments, the standard calibration curve was made for MC-LR concentrations in the range of 5-30 mg L −1 . The photocatalytic degradation experiments were carried out in a double jacket round quartz reactor with a 50 mL volume. The temperature was maintained at 25 • C by the circulation of thermostated water around the reactor. Firstly, an aqueous solution of MC-LR (40 mL) was added to a round quartz reactor containing TiO 2 powder (0.4 g L −1 ), and sonicated to obtain a uniform suspension, and then the pH value was adjusted to be 4. The reaction solution (MC-LR, H 2 O, and TiO 2 ; pH 4) was stirred for 180 min in the dark to achieve the adsorption equilibrium for MC-LR on the catalyst surface before starting the experiments. After that, the suspension was irradiated from the top by a UV-A lamp (λ max = 365 nm, intensity = 2 mW cm −2 ), solar simulator lamp (SOL1200 lamp, intensity = 20 mW cm −2 ), and visible-LED lamp (λ max = 420 nm, intensity = 1 mW cm −2 ). The samples were taken at different times, and then filtered using a 0.22-µm filter membrane. The residual MC-LR concentration at different durations was analyzed using a high-performance liquid chromatography (HPLC, 1260, Agilent, Hamburg, Germany) with a G1311C-1260 Quat pump and a G1365D-1260 MWD UV detector (Hamburg, Germany), set at 238 nm with a C18 column (100 mm Long × 4.6 mm i.d., 3.5 µm particles) by the method reported before [1,72]. The reaction rates were estimated and fitted with the Langmuir-Hinshelwood first-order kinetic model. The degradation rate (r) was calculated using Equation (1) [1,12,31,40]: where K is the rate constant, C 0 is the initial concentration of MC-LR, and n is the order of the reaction. The MC-LR photodegradation efficiencies (%) using the prepared TiO 2 photocatalysts under UV-A, solar and visible light were estimated using Equation (2) where C 0 and C are the MC-LR concentrations before and after irradiation, respectively.

Conclusions
The MC-LR toxin was removed from aqueous solution by UV-A, solar and visible light in the presence of nanostructured TiO 2 photocatalysts (anatase, brookite, anatase-brookite, C/N/S-co-modified anatase-brookite). A simple hydrothermal method was investigated to synthesize pristine TiO 2 nanoparticles with tunable A/B ratios, which was achieved by changing the preparation pH value. In addition, the best bare TiO 2 was calcined with thiourea to obtain C/N/S-co-modified mesoporous A/B TiO 2 . The effect of the preparation pH value on the phase composition, surface area, and photocatalytic activity was investigated, and pH-dependent behavior was observed. It was found that the single-phase TiO 2 nanoparticles, anatase and brookite, were formed at high acidic and basic pH, respectively, and A/B TiO 2 samples were obtained in the pH value range of 5 to 10. Upon increasing the pH, the specific surface areas increased, leading to higher photocatalytic activity. The co-modified and non-modified mixed-phases TiO 2 exhibited a superior photocatalytic activity compared to the single-phase TiO 2 (anatase and brookite) under UV-A and solar irradiation, probably because of mixed-phase formation, mesoporous structure, and higher specific surface area (non-modified mixed-phase TiO 2 ). The non-modified TiO 2 practically was inactive under visible light irradiation (<25%), whilst the complete MC-LR degradation (100%) was achieved in the presence of C/N/S co-modified-TiO 2 . It is proposed that this improved activity for co-modified TiO 2 comes from non-metal-co-modification, resulting in bandgap narrowing. Hence, highly active photocatalysts against very toxic pollutants (MC-LR) could be efficiently applied for water/wastewater purification under natural solar radiation.