Curcumin Doped SiO 2 / TiO 2 Nanocomposites for Enhanced Photocatalytic Reduction of Cr (VI) under Visible Light

: In order to further improve the photocatalytic performance of the SiO 2 / TiO 2 composite under visible light irradiation, curcumin-doped SiO 2 / TiO 2 nanocomposites were synthesized via directly incorporating it into the structure of SiO 2 / TiO 2 during the synthesis using an inexpensive and readily available natural pigment (curcumin) as doping agent. The physicochemical properties of SiO 2 / TiO 2 nanocomposites were characterized in detail by X-ray di ﬀ raction, transmission electron microscopy, Fourier transform-infrared spectroscopy, N 2 adsorption–desorption isotherms, X-ray photoelectron spectroscopy, and UV-vis di ﬀ use reﬂectance spectroscopy. The results indicate that all SiO 2 / TiO 2 nanocomposites exhibited an anatase phase with a typical mesoporous structure. It was found that the dope of curcumin in the SiO 2 / TiO 2 composite could decrease the crystal size, slightly improve the speciﬁc surface areas, signiﬁcantly enhance the visible light absorption, and e ﬀ ectively narrow the band gap energy from 3.04 to 10(eV). Compared with bare SiO 2 / TiO 2 , the curcumin-doped SiO 2 / TiO 2 resulted in enhanced photocatalytic reduction activity for Cr(VI) under visible light irradiation, and the CTS (12) sample with the appropriate content of curcumin of 12 wt % shows the photocatalytic yield reaching 100% within 2.5 hours, which is larger than CT (12) without silica. This could be attributed to the curcumin doping and the synergetic e ﬀ ects of SiO 2 and TiO 2 in SiO 2 / TiO 2 nanocomposites.


Introduction
Titanium dioxide photocatalysis is considered as a promising and effective technology for the removal of organic pollutants from aquatic systems due to its non-toxic, low cost, and resistance to photochemical and chemical corrosion [1,2]. Anatase, rutile, and brookite are three polymorphs of TiO 2 , among which anatase is the most active one in the photocatalytic process [3]. In order to improve its photocatalytic activity, novel TiO 2 -based photocatalysts were designed and synthesized through both the structural modification and the compositional change [3,4]. Compared to pure titanium oxide, SiO 2 /TiO 2 composite is found to be more conducive to improve the photocatalytic activity and the separation of products after the reaction [5][6][7][8]. For examples, Hou et al. [9] found that the dope of SiO 2 significantly changed the phase composition, specific surface area, microstructure, and the photocatalytic activity of TiO 2 . Mahyar et al. [10] reported that the specific surface area of a catalyst increased by doping the SiO 2 , which allowed an increase in adsorption of pollutant molecules, thus, improving photocatalytic activity of TiO 2 . Meng et al. [11] reported that the increase of the photocatalytic activity resulted from the addition of silica, which reduced the particle size, improved the specific surface area and thermal stability of anatase. However, like TiO 2 , SiO 2 /TiO 2 composites can only be excited under the UV light since the band gap energy of TiO 2 is 3.2 eV. Therefore, there is a growing interest in doping modification of SiO 2 /TiO 2 for improving its visible light photocatalytic performance. For examples, Ghanbari et al. [12] found that doping N and F in TiO 2 /SiO 2 nanoparticles led to the fully anatase crystalline structure, the minimized agglomeration of synthesized nanoparticles, and enhanced photocatalytic activity under solar and visible light irradiation. Liu et al. [13] prepared F/W co-doped TiO 2 -SiO 2 composite aerogels with the higher photoelectron conversion efficiency, which are advantageous to degrade the pollutant RhB in the aqueous solution under visible light irradiation.
Beside element doping, the dye sensitization is also proved to be an effective method to increase the visible light absorption of photocatalysts. This has been widely reported in the field of solar cells, the production of hydrogen and the degradation of organic pollutants [14][15][16]. Recently, natural metal-free dyes derived from flowers, leaves, and fruits have become viable alternatives to expensive or toxic organic sensitizers because they are low-cost, easy to obtain, abundant in raw materials, and have no threat to the environment [17,18]. Many colored natural products, such as chlorophylls, carotenoids, anthocyanins, flavonoids, etc., have been used as sensitizers for TiO 2 system [19][20][21]. Among the large variety of natural pigments, curcumin, a natural ingredient in commercial turmeric powder, is an orange-yellow pigment, which has been widely used in physics, chemistry, biology, and medicine due to its non-toxic, biocompatible, naturally occurring characteristics [22]. In the structure of curcumin, there are two α, β unsaturated carbonyl groups at the center, which connected two phenolic groups. The β-diketone group is one of the most prominent functional groups and has an intense driving force for chelation with a metal ion. The metal complexes formed tend to be more stable than curcumin itself, which is easily degraded [23]. Moreover, curcumin has the strong absorption of a long wavelength in the visible range from 420 to 580 nm, which makes it an ideal dye sensitizer for environmental photochemical transformations in the visible light environment. For examples, Ganesh et al. [24] reported that curcumin is an effective dye in dye-sensitized solar cells (DSSC), in which the photocurrent density is 1.66 mA/cm 2 . Lim et al. [25] reported the curcumin-sensitized TiO 2 composite showed significant activity in degrading organic pollutions and confirmed that the photochemical and photoelectrochemical activities of curcumin/TiO 2 are consistently higher than that of RuL 3 /TiO 2 over a wide pH range. Recently, curcumin is coupled with wide band gap metal oxides, such as Bi 0.5 Na 0.5 TiO 3 and MgO·Al 2 O 3 ·ZnO, to obtain visible light catalysts for the dye degradation [26,27]. However, it should be noted that curcumin along with other dyes attached to the TiO 2 or other oxides by simple adsorption of dyes can only be considered as a sensitizer, not a photocatalyst. The limited adsorption capacity of dyes and weak binding between dye molecules and catalysts are not conducive to rapidly transferring electrons from the dye molecules to the catalysts. Therefore, it is very meaningful to find a more suitable sensitization method to improve the photocatalytic efficiency.
In our previous works, TiO 2 nanoparticles with visible photocatalytic activity have been successfully synthesized by using anthocyanin dyes as a template during the synthesis [21,28]. The prepared TiO 2 exhibits much higher activity under visible light irradiation than under UV light. In this way, it is worthwhile to design and prepare a SiO 2 /TiO 2 nanocomposite with visible photocatalytic activity by using a curcumin pigment. To the best of our knowledge, there is no report on the synthesis of SiO 2/ TiO 2 nanomaterials by natural pigments at present.
In this work, a curcumin doped SiO 2 /TiO 2 nanocomposite was prepared by combining supercritical drying with solvothermal-assisted crystallization, in which the curcumin pigment is directly incorporated into the structure of SiO 2 /TiO 2 during the synthesis instead of external adsorption after the formation of the SiO 2 /TiO 2 composites. The effects of curcumin content on microstructure, optical, and surface electronic state of the SiO 2 /TiO 2 nanocomposites were observed. The photocatalytic performance was presented by reducing Cr (VI) under visible light illumination.

Structure and Morphology
The XRD pattern of SiO 2 /TiO 2 nanocomposites and CT (12) sample are showed in Figure 1 and Figure S1. The strong diffraction peaks at 25.3 • , 37.8 • , 48.0 • , 55.0 • , and 62.8 • correspond to (101), (004), (200) (211), and (116) crystal planes of anatase TiO 2 , respectively, revealing the higher crystallinity of TiO 2 . For CTS (X) samples, the crystallinity of TiO 2 tended to slightly decrease with the increase of doped curcumin amount. The SiO 2 did not appear in the XRD patterns, indicating that SiO 2 existed as an amorphous phase in the nanocomposites. The average crystalline size of TS, CTS (3), CTS (6), CTS (9), CTS (12), CTS (15), and CT (12) was calculated using the Scherrer equation and the value were 9.8, 8.2, 7.5, 6.8, 6.5, 5.8, and 11.5 nm, respectively, as shown in Table 1. Obviously, the crystalline size of SiO 2 /TiO 2 nanocomposites became smaller with the increase of curcumin concentration from 3 to 15%. This may be due to the formation of complexes between curcumin and the Ti atom, which effectively suppress the agglomeration and growth of TiO 2 particles, like a supramolecular template. In addition, the crystalline size of the CT (12) sample without SiO 2 was significantly greater than that of CTS (12), which indicated that SiO 2 was also responsible to suppress the aggregation and growth of the TiO 2 particles. Similar results have been reported by Hou [9] and Meng [11] et al. The textural properties and band gap of these samples in Table 1 will be presented later.      Figure 2 showed the SEM images of typical SiO 2 /TiO 2 nanocomposites. It can be seen that both TS and CTS (12) nanocomposites were composed of irregular particles and displayed a similar particle accumulation pore network structure.  Figure 2 showed the SEM images of typical SiO2/TiO2 nanocomposites. It can be seen that both  ratios. It is interesting to find that the TS composite without curcumin doping actually exhibited a 125 denser network structure than CTS (X). The dispersion of SiO2/TiO2 nanoparticles was obviously improved with the increase of curcumin, and highly dispersed nanoparticles were obtained when 127 the weight ratios of curcumin were more than 6%. It is reasonable that the chelation of curcumin 128 with Ti metal will suppress the agglomeration and growth of TiO2 particles to a certain extent, which 129 may be beneficial to increase the specific surface areas of the composite. Figure 3e shows the HRTEM 130 images of CTS (12) samples, which indicate the TiO2 particle from the CTS (12) sample was an 131 anatase structure and its diameter was about 6 nm. This is consistent with the result by the XRD 132 measurement. Moreover, the selected area electron diffraction (SAED) pattern in the inset of Figure   133 3e revealed that TiO2 had high crystallinity. The energy dispersive X-ray spectroscopy (EDS) spectra 134 in Figure 3f indicate that the Ti, Si, and O element existed in the SiO2/TiO2 nanocomposites.

135
Moreover, carbon element appeared in samples, which confirms that curcumin was successfully  Figure 3 shows the TEM images of SiO 2 /TiO 2 nanocomposites with different curcumin weight ratios. It is interesting to find that the TS composite without curcumin doping actually exhibited a denser network structure than CTS (X). The dispersion of SiO 2 /TiO 2 nanoparticles was obviously improved with the increase of curcumin, and highly dispersed nanoparticles were obtained when the weight ratios of curcumin were more than 6%. It is reasonable that the chelation of curcumin with Ti metal will suppress the agglomeration and growth of TiO 2 particles to a certain extent, which may be beneficial to increase the specific surface areas of the composite. Figure 3e shows the HRTEM images of CTS (12) samples, which indicate the TiO 2 particle from the CTS (12) sample was an anatase structure and its diameter was about 6 nm. This is consistent with the result by the XRD measurement. Moreover, the selected area electron diffraction (SAED) pattern in the inset of Figure 3e revealed that TiO 2 had high crystallinity. The energy dispersive X-ray spectroscopy (EDS) spectra in Figure 3f indicate that the Ti, Si, and O element existed in the SiO 2 /TiO 2 nanocomposites. Moreover, carbon element appeared in samples, which confirms that curcumin was successfully doped into the SiO 2 /TiO 2 nanocomposites.  Figure 2 showed the SEM images of typical SiO2/TiO2 nanocomposites. It can be seen that both 119 TS and CTS (12) nanocomposites were composed of irregular particles and displayed a similar 120 particle accumulation pore network structure. 123 Figure 3 shows the TEM images of SiO2/TiO2 nanocomposites with different curcumin weight 124 ratios. It is interesting to find that the TS composite without curcumin doping actually exhibited a 125 denser network structure than CTS (X). The dispersion of SiO2/TiO2 nanoparticles was obviously 126 improved with the increase of curcumin, and highly dispersed nanoparticles were obtained when 127 the weight ratios of curcumin were more than 6%. It is reasonable that the chelation of curcumin 128 with Ti metal will suppress the agglomeration and growth of TiO2 particles to a certain extent, which 129 may be beneficial to increase the specific surface areas of the composite. Figure 3e shows the HRTEM 130 images of CTS (12) samples, which indicate the TiO2 particle from the CTS (12) sample was an 131 anatase structure and its diameter was about 6 nm. This is consistent with the result by the XRD 132 measurement. Moreover, the selected area electron diffraction (SAED) pattern in the inset of Figure   133 3e revealed that TiO2 had high crystallinity. The energy dispersive X-ray spectroscopy (EDS) spectra 134 in Figure 3f indicate that the Ti, Si, and O element existed in the SiO2/TiO2 nanocomposites.
doped into the SiO2/TiO2 nanocomposites. The N 2 adsorption/desorption isotherms and pore size distribution curves of the SiO 2 /TiO 2 nanocomposites and CT (12) sample are shown in Figure 4 and Figure S2. It is found that these samples exhibited typical type-IV isotherms having an obvious hysteresis loop at the 0.6-1.0 P/P 0 range, which demonstrated that numerous mesopores exist in the bulk phase of the SiO 2 /TiO 2 nanocomposites. The pore size distribution curves in Figure 4b revealed that the pore size of the SiO 2 /TiO 2 nanocomposites had a decreasing trend with the increase of the curcumin doping amount, which could be due to excessive curcumin and its decomposition residues blocking the pores. The specific surface areas, pore volumes, and pore sizes of SiO 2 /TiO 2 nanocomposites were listed in Table 1. It could be found that specific surface areas of SiO 2 /TiO 2 nanocomposites occur a slight increase when the weight ratios of curcumin change from 0 to 15%. By contrast, the CT (12) sample shows a lower specific BET surface area of 134 m 2 g −1 than CTS (12) (287 m 2 g −1 ). Many reports [29,30] suggest that the addition of silica is able to improve the specific surface area of TiO 2 . This would provide more surface sites for the adsorption of reactants molecules and lead to more active sites, which would contribute to the improvement of photocatalytic activity [31]. In addition, the pore size of the SiO 2 /TiO 2 nanocomposites and CT (12) sample were between 15.0 and 18.6 nm in the mesoporous region.  and selected area electron diffraction (SAED) of CTS (12), and (f) EDS of CTS (12).

140
The N2 adsorption/desorption isotherms and pore size distribution curves of the SiO2/TiO2 141 nanocomposites and CT (12) sample are shown in Figure 4 and Figure S2. It is found that these 142 samples exhibited typical type-IV isotherms having an obvious hysteresis loop at the 0.6-1.0 P/P0 143 range, which demonstrated that numerous mesopores exist in the bulk phase of the SiO2/TiO2 144 nanocomposites. The pore size distribution curves in Figure 4b revealed that the pore size of the 145 SiO2/TiO2 nanocomposites had a decreasing trend with the increase of the curcumin doping amount, 146 which could be due to excessive curcumin and its decomposition residues blocking the pores. The Table 1. It could be found that specific surface areas of SiO2/TiO2 nanocomposites occur a slight 149 increase when the weight ratios of curcumin change from 0 to 15%. By contrast, the CT (12) sample 150 shows a lower specific BET surface area of 134 m 2 g −1 than CTS (12) (287 m 2 g −1 ). Many reports [29,30] 151 suggest that the addition of silica is able to improve the specific surface area of TiO2. This would 152 provide more surface sites for the adsorption of reactants molecules and lead to more active sites,

153
which would contribute to the improvement of photocatalytic activity [31]. In addition, the pore size    Figure 5 presents FTIR spectra of curcumin and SiO 2 /TiO 2 nanocomposites between 500 and 4000 cm −1 . The IR spectrum in Figure 5a shows a characteristic band of curcumin at 3520 cm −1 assigned to the phenolic O-H stretching vibration. Moreover, the band at 1633 cm −1 can be due to the stretching vibration of benzene ring of curcumin, a band at 1510 cm −1 to C=O and C=C vibration, a band at 1283 cm −1 to the aromatic C-O stretching vibration, the band at 1000-1300 cm −1 to C-O-C stretching vibration [27]. The FT-IR spectrum of CTS (9) was very similar to that of pure SiO 2 /TiO 2 , indicating that the main structure of CTS (9) was almost identical to that of pure SiO 2 /TiO 2 . The bands at 830 and 1096 cm −1 are related to the Si-O-Si symmetric and asymmetric stretching vibration, respectively. The typical Ti-O-Ti stretching vibration due to the anatase phase is located in the region of 500-590 cm −1 . Moreover, the very weak band around 952 cm −1 is associated with the Ti-O-Si linkage, which indirectly indicates the interaction between TiO 2 and SiO 2 on the molecular scale [9]. The two broad peaks at 3420 and 1635 cm −1 can be attributed to the adsorption of water and hydroxyl on the surface. However, no characteristic peak of curcumin was found in CTS (9). In the case of CTS (X) samples, the FT-IR spectra were not distinctively different from each other, as shown in Figure 5b, which indicates that curcumin in the CTS (X) nanocomposites may be decomposed during the heating under nitrogen atmosphere.  CTS (15) CTS (12) CTS (9) CTS (6) CTS (3) Transmittance (%)

FT-IR and XPS Analyses
Wavenumber (cm -1 )  Figure 6 shows the XPS spectra of the CTS (12) nanocomposite. The full survey spectrum in Figure 6a expounds that the nanocomposites contain the Si, Ti, O, and C element. Figure 6b shows the XPS spectrum of Ti2p, in which two peaks at 464.2 and 458.4eV were ascribed to Ti 2p1/2 and Ti 2p3/2, respectively. These values are in good agreement with Ti 4+ in pure anatase TiO 2 [13]. Figure 6c shows the XPS spectrum of O1s, where three peaks at 529.8 eV, 531.0 eV, and 532.3 eV were assigned to Ti-O, Si-O, and O-H groups, respectively [9]. The C1s spectrum in Figure 6d can be fitted by three curves appearing at 284.8, 286.4, and 288.7 eV due to C-C (C-H), C=O, and Ti-O-C, respectively [32][33][34], manifesting the existence of decomposition residues from curcumin on the surface of SiO 2 /TiO 2 . Moreover, the atomic mass fractions in the CTS (12) and TS samples can be obtained from the curve fitting results of XPS spectra, as shown in Table 2. It is found that the mass fractions of the C atom are 23.3% and 16.1% for CTS (12) and TS, respectively, which further verifies the successful doping of curcumin into SiO 2 /TiO 2 .        the CT (12) sample also shows significant visible-light absorption, as shown in Figure S3, and its 208 band gap energy was 2.78 eV, which was higher than that of CTS (12) Figure 7a presents the UV-Vis diffraction reflectance spectra of curcumin, TS, and CTS (X). Obviously, TS do not show significant absorption in the visible region from 400 to 750 nm, while the curcumin revealed significant absorption at the same region. After doping the curcumin, the absorption of CTS (X) samples in the visible range was significantly improved with the increase of curcumin content, and CTS (12) achieved maximum absorption strength. While curcumin content was 15%, the absorption of CTS (15) had a slight decline. The bandgap energies of the TS and CTS (X) samples were obtained from Tauc plots in Figure 7b. It could be found from Table 1 that the bandgap energies ranged from 3.04 eV of TS to 2.10 eV of CTS (12), indicating that the dope of curcumin could effectively narrow the bandgap energy of SiO 2 /TiO 2 nanocomposites. The enhanced visible light absorption and the narrowed bandgap of CTS (X) contributed to generate photoelectron-hole pairs, further leading to the improvement of photocatalytic activity. In addition, the CT (12) sample also shows significant visible-light absorption, as shown in Figure S3, and its band gap energy was 2.78 eV, which was higher than that of CTS (12), suggesting that SiO 2 helped to decrease the bandgap energy of SiO 2 /TiO 2 nanocomposites.

UV-Vis DRS Analyses
196 Figure 7a presents the UV-Vis diffraction reflectance spectra of curcumin, TS, and CTS (X).  Figure S3, and its 208 band gap energy was 2.78 eV, which was higher than that of CTS (12) (12) sample showed the best photocatalytic activity with the photocatalytic yield reaching 100% within 2.5 h. This result implies that the improved photocatalytic activities resulted from the dope of curcumin, and the introduction of a suitable amount of curcumin was beneficial to improve photocatalytic performance of SiO 2 /TiO 2 composites. In fact, owing to the much stronger absorption in the visible range of CTS (X) nanocomposites than that of the TS sample, it was reasonable that curcumin-doped CTS (X) samples showed much higher photocatalytic activity than the TS sample under visible-light irradiation. Furthermore, the carbonaceous species loaded on materials were responsible for the improved photocatalytic activity. Many investigations [31,35,36] show that carbonaceous species on TiO 2 are capable of promoting direct transfer and separation of the electron. However, the CTS (15) sample with the highest doping curcumin exhibited decreased photocatalytic reduction properties, probably owing to the fact that the addition of excessive curcumin would block the pores or destroy the structural integrity of SiO 2 /TiO 2 , thus, may produce more recombination centers for photogenerated electron-hole pairs [16,37]. In addition, it can be found that curcumin-doped TiO 2 samples showed a 69.1% photocatalytic reduction yield under a visible-light illumination. However, in spite of the same curcumin doping amount as CTS (12), the photocatalytic reduction yield of CT (12) was still lower than that of CTS (12). The result indicates that the synergy between SiO 2 and TiO 2 in the CTS (12) nanocomposites also contributed to the improvement of photocatalytic activity.

233
in spite of the same curcumin doping amount as CTS (12), the photocatalytic reduction yield of CT 234 (12) was still lower than that of CTS (12). The result indicates that the synergy between SiO2 and TiO2 235 in the CTS (12) nanocomposites also contributed to the improvement of photocatalytic activity.

240
The CTS (12) sample was used to evaluate the influence of the initial concentration on 241 photocatalytic Cr(VI) reduction by changing initial concentration of Cr(VI) from 10 to 50 ppm with 242 0.6 mL/L formic acid. Figure 9 shows that the yields of photocatalytic Cr(VI) reduction decreased 243 from 100% to 45% with increasing the initial Cr(VI) concentration after 2.

Effect of Initial Concentration on the Photocatalytic Reduction of Cr(VI)
The CTS (12) sample was used to evaluate the influence of the initial concentration on photocatalytic Cr(VI) reduction by changing initial concentration of Cr(VI) from 10 to 50 ppm with 0.6 mL/L formic acid. Figure 9 shows that the yields of photocatalytic Cr(VI) reduction decreased from 100% to 45% with increasing the initial Cr(VI) concentration after 2.5 hours. As a reason, with the increase of initial Cr(VI) concentration, more Cr(VI) ions are adsorbed on the surface of the photocatalyst. This not only prevents the absorption of light by the catalyst, but also covers the active site on the surface of the photocatalyst, thus inhibiting further photocatalytic Cr (VI) reduction [38,39].

252
The pH value of the solution is one of the main factors affecting the photocatalytic reduction of Table 3. As was observed, the adsorption ability for Cr (VI) decreased when the pH value increased.

265
The reduction yield of Cr (VI) by CTS (12) was 100% when pH was 2, while it was only 5.0% when 266 pH was 8. As is well known, the reduction of Cr (VI) mainly involves the following reactions:

272
TiO2-graphen [40], and TiO2-P25 [41] were used as photocatalysts. Considering that under strong 273 acidic conditions, such as pH 2, the adsorption of Cr (VI) by CTS (12) was relatively high and was 274 over 45%. As a result, the appropriate acidity condition for the removal of Cr (VI) was considered to 275 be pH 4 when using CTS(X).

276
Additionally, compared with the results in Figure 7, the photocatalytic reduction yield under 277 acidic conditions was much lower than that in the presence of formic acid. This means that formic

Effect of pH on the Photocatalytic Reduction of Cr(VI)
The pH value of the solution is one of the main factors affecting the photocatalytic reduction of Cr (VI), and the types of Cr (VI) and surface charges of the catalyst are different in different pH solutions. The adsorption of Cr (VI) over CTS (12) sample under different pH value was shown in Table 3. As was observed, the adsorption ability for Cr (VI) decreased when the pH value increased. Actually, the surface of CTS (12) carries more positive charges in acidic solution and has a strong affinity toward HCr 2 O 7 − anion that is the dominant species at pH 2-6. When the pH value increased, the positive charges would be less on the surface of CTS (12), which is not conducive to the attraction of HCr 2 O 7 − or CrO 4 2− with negative charges. This would lead to a lower adsorption capacity for Cr (VI), especially in the alkaline medium.  Figure 10 presents the effect of pH on photocatalytic reduction of Cr (VI) when the pH value is 2, 4, 5, 6, and 8, in which the initial Cr (VI) concentration remained the same at 20 ppm. It can be seen that the photocatalytic reduction yields were decreased with increasing the pH value of the solution. The reduction yield of Cr (VI) by CTS (12) was 100% when pH was 2, while it was only 5.0% when pH was 8. As is well known, the reduction of Cr (VI) mainly involves the following reactions: So, H + is favorable for the Cr (VI) reduction under acidic medium, while OH − shows the opposite effect in alkaline conditions. Therefore, the lower the pH is, the higher the photoreduction efficiency of Cr (VI) is. Similar phenomena had been observed when carbon dots-TiO 2 [35], TiO 2 -graphen [40], and TiO 2 -P25 [41] were used as photocatalysts. Considering that under strong acidic conditions, such as pH 2, the adsorption of Cr (VI) by CTS (12) was relatively high and was over 45%. As a result, the appropriate acidity condition for the removal of Cr (VI) was considered to be pH 4 when using CTS(X).
Additionally, compared with the results in Figure 7, the photocatalytic reduction yield under acidic conditions was much lower than that in the presence of formic acid. This means that formic acid is necessary as a sacrificial agent to capture holes and provide more electrons for Cr (VI) reduction.

Recyclability of Photocatalysts
The durability of catalysts is a key criterion for practical applications, so the photocatalytic stability of CTS (12) was evaluated under visible light irradiation by a 6 h reaction during three consecutive runs. Figure 11 exhibits that the reduction yield of photocatalytic Cr (VI) could still reach 97% after three photocatalytic cycles. The result suggests that the photocatalytic activity for CTS (12) had no obvious attenuation after the cycling tests, indicating its intense antiphotocorrosion stability.    to absorb more visible light [32]. Second, it is generally accepted that [34,42] doped carbon in TiO2

298
can act as an electron scavenging agent to reduce the recombination of the photogenerated electron-

Possible Reasons for the Enchanced Visble-light Photocatalytic Activity
On the basis of the above discussion, three reasons may account for the significant photocatalytic activity of the CTS (12) under visible irradiation. First, the dope of curcumin was responsible for the far more intense adsorption of visible light. When carbonaceous species from curcumin were introduced, the band gap of CTS (X) would be narrowed, allowing the photocatalyst to absorb more visible light [32]. Second, it is generally accepted that [34,42] doped carbon in TiO 2 can act as an electron scavenging agent to reduce the recombination of the photogenerated electron-hole pair, and enhance the charge separation effect of photocatalysts. Thus, when CTS (X) was irradiated by visible light, more electrons can be provided to reduce Cr (VI) to Cr (III), while the holes would be consumed by oxidizing formic acid. Finally, the synergetic effects of SiO 2 and TiO 2 play an important role in the improvement of photocatalytic performance. As described above, besides curcumin, SiO 2 also contributes to improve specific surface area and narrow the bandgap energy of CTS (12). So, it is reasonable that absorption towards visible light for CTS (12) is far more than that of CT (12). Moreover, CTS (12) with large specific surface area can not only improve the adsorption of reactants molecules, which is usually the first step of the photocatalytic reaction, but also provide more active sites to accelerate the diffusion and transition of reactants during photocatalysis [29,31,43,44], all which are beneficial to the improvement of photocatalytic performance. Therefore, the significantly enhanced visible photocatalytic activity may be due to the curcumin doping and the synergetic effects of SiO 2 and TiO 2 in SiO 2 /TiO 2 nanocomposites.

Synthesis
The curcumin was provided by Yunnan Tonghai Yang Natural Products Co., LTD (Yuxi, China). Titanium isopropylate and tetraethyl orthosilicate was bought from Aladdin Industrial Inc. (Shanghai, China).
Curcumin doped SiO 2 /TiO 2 nanocomposites were prepared by a modified process [21,28]. Different amounts of curcumin was dissolved in 50 mL of ethanol, 5 mL tetraethyl orthosilicate and 15 mL titanium isopropylate were added in turn under stirring, then the mixture of 0.1 mL nitric acid and 4 mL deionized water was added to the above solution drop by drop and continued to stir for 30 min. The obtained mixture was aged into a gel at room temperature for 24 h. The gel was transferred into an autoclave and dehydrated at 245 • C for 2 h under nitrogen atmosphere. Before being cooled down to room temperature, the sample was rinsed again with nitrogen for another 15 min. The final products were collected for use without further processing. A series of SiO 2/ TiO 2 samples (denoted as CTS(X), X is the amount of curcumin, and X = 3 wt %, 6 wt %, 9 wt %, 12 wt %, and 15 wt %, respectively) were prepared by changing the content of curcumin in the precursor. For comparison, the pure SiO 2 /TiO 2 nanocomposite without adding curcumin and TiO 2 sample with 12 wt % curcumin was prepared according to the same procedure, which was denoted as CTS and CT (12), respectively.

Characterizations
X-ray diffraction (XRD) data were obtained on a Rigaku TTRAX III spectrometer with Cu Kα radiation (Rigaku Co., Tokyo, Japan) with a scanning angle between 10 and 80 • at a rate of 10 • /min. Scanning electron microscopy (SEM) characterizations were carried out by a FEI Quanta 200 ESEM (FEI, Hillsboro, OR, USA) with an accelerating voltage of 10 kV. Transmission electron microscopy (TEM) images were analyzed by a JEM-2100 (Japan Electron Optics Laboratory CO., Ltd., Tokyo, Japan) operated at 200 kV. The specific surface areas were determined by a Micromeritics Tristar II Surface Area and Porosity Analyzer (Micromeritics, Norcross, GA, USA). Fourier transform infrared (FTIR) spectra were obtained on a Thermo Nicolet 8700 instrument (Thermo Fisher Scientific, Waltham, MA, USA). The light absorption spectra of the samples were recorded using a Shimadzu UV2600 spectrophotometer (Shimadzu Corp., Kyoto, Japan) over the 200-800 nm range. The X-ray photoelectron spectroscopy (XPS) experiment was performed using a PHI5500ESCA (Thermo Fisher Scientific Inc., Waltham, MA, USA) spectrophotometer with the 200 W Mg Kα source.

Photocatalytic Activity
Photocatalytic reactions were performed in a quartz photoreactor. The photocatalyst of 25 mg was added to a 50 mL 10-50 ppm Cr (VI) aqueous solution with 0.6 mL/L formic acid. The solution pH value was adjusted to 2.0-8.0 with 0.2 M of HCl or 0.2 M of NaOH. The suspensions were transferred in a quartz photoreactor and magnetically stirred kept for 6 h under dark conditions to achieve adsorption-desorption equilibrium. The photocatalytic process was initiated by exposing the suspensions under visible light using the 350 W Xe lamp with a 420 nm cut-off filter. During light exposure, cold water was circulated around the photoreactor to maintain room temperature. At scheduled time intervals, the samples were collected from the reactor during illumination and then filtered through a 0.45 mm PTFE syringe filter to remove the powder samples. The concentrations of Cr (VI) were measured using a UV-Vis spectrophotometer at 540 nm.

Conclusions
In summary, the curcumin-doped SiO 2 /TiO 2 nanocomposites were prepared by directly incorporating it into the structure of SiO 2 /TiO 2 during the synthesis using an inexpensive and readily available natural pigment (curcumin) as the doping agent, rather than conventional photosensitization by adsorbing. Curcumin could reduce the crystal sizes and slightly increase the specific surface area, by contrast, the visible absorbance of SiO 2 /TiO 2 could be enhanced remarkably, which was related with the content of doped curcumin. The appropriate curcumin in SiO 2 /TiO 2 was conducive to enhance the visible photocatalytic activity. In particular, CTS (12) nanocomposites with the content of curcumin of 12 wt % exhibited the best visible-light photocatalytic activity for the reduction of Cr (VI) in the solution. This can be ascribed to the curcumin doping and the synergetic effects of SiO 2 and TiO 2 in SiO 2 /TiO 2 nanocomposites. CTS (12) also showed excellent stability and did not lose significant activity after being reused 3 times. Furthermore, this approach is simpler and provides new thought for the design of the visible light catalysts with high efficiency for environmental applications, specifically for the treatment of polluted water applications.

Conflicts of Interest:
The authors declare no conflict of interest.