Biotemplated Mesoporous TiO 2 / SiO 2 Composite Derived from Aquatic Plant Leaves for Efficient Dye Degradation

The biotemplating technique is an environmental-protective high-efficiency new technology by which the resulting TiO2 may simultaneously attain the duplication of structure and self-doping elements from biotemplate materials, which is highly desirable for photocatalytic applications. In this paper, aquatic plant leaves—including reed, water hyacinth, and duckweed—were used as both templates and silicon precursors to successfully synthesize biomorphic TiO2/SiO2 composite with mesoporous structures. X-ray diffraction, scanning electron microscopy, transmission electron microscopy, N2 adsorption–desorption, and UV–visible diffuse reflectance spectra were applied to characterize the microstructures of the samples. The results show that all TiO2/SiO2 composites are mainly composed of an anatase phase with mesoporous structure and possess high specific surface area. Compared with commercial Degussa P25 TiO2, all TiO2/SiO2 samples display intensive light-harvesting efficiency, particularly in the visible light range. The activities were evaluated by using gentian violet as a target for photocatalytic degradation experiments under simulated solar irradiation. The TiO2/SiO2 samples templated by reed and water hyacinth leaves exhibit high activity, while the TiO2/SiO2 samples obtained from duckweed are inferior to P25 in the degradation of gentian violet. A synergistic effect of SiO2 incorporation and structural construction through biotemplating is proposed to be beneficial to photocatalytic activity.


Introduction
Photocatalytic processes to degrade organic pollutants in water by utilizing a catalyst under solar light illumination have been proven to be potentially advantageous and useful when we have to resolve increasingly serious environmental problems [1].As a typical environmentally benign photocatalyst, titanium dioxide (TiO 2 ) shows potential applications in the future because of its stability as well as the low cost.Unfortunately, the pure form of TiO 2 can only absorb a small fraction of solar energy, and currently used TiO 2 photocatalysts, such as commercial Degussa P25, exhibit low specific surface area and poor adsorption, which restrict its practical applications.To date, numerous studies have been carried out for improving the photocatalytic performance of TiO 2 through both structural modification and compositional change.One way is to prepare TiO 2 in the form of a porous structure by applying various chemicals like supramolecular compounds or self-assembled polymers as templates [2].Another is to dope TiO 2 with nonmetal elements (nitrogen, sulfur, carbon, etc.) [3,4], or synthesize a mixed oxide, like TiO 2 /SiO 2 , by utilizing an in situ technique or post treatment [5][6][7][8].Despite great progress being made, the application of chemical templates is uneconomical and, in turn, further increases environmental pollution.Also, the introduction of additional precursor to dope elements in photocatalysts is complex and cumbersome.More important, most studies were devoted to only one of the methods, rather than both of them, and neither of the methods by which TiO 2 photocatalysts are prepared are widely applied-although the environmental problem is more and more serious-which to some extent is owing to the poor light absorption and small specific surface area of TiO 2 [3].Therefore, it is highly desirable to develop a simple, low-cost, and environmentally friendly method to prepare efficient TiO 2 photocatalysts by concurrently modifying the structure and composition.
Nowadays, the biotemplating technique, which combines biological resources with a sol-gel synthetic method, has been widely used in templated synthesis of TiO 2 , especially when inheritance of the structure/morphology from template to product is desired.The biotemplate materials are more economical and environmentally benign, which are incomparable with traditional chemical template methods.So, the biotemplating technique has attracted more and more attention.So far, many plant tissues are used as novel biotemplates for the development of promising photocatalysts.On the one hand, the hierarchical pore structures of plant tissues, which are designed favorably for photosynthesis, endow them with unique properties, particularly great light-adsorbing capability.For example, the spotlight effect of epidermal cells in leaves and the light-scattering of plant tissues would contribute to promoting light adsorption at the interior of the leaf [9,10].When these plant tissues are pretreated with a glutaraldehyde, HCl, and TiCl 3 solution, their morphology and structural features can be well copied to the resulting TiO 2 via a sol-gel process [10,11].On the other hand, plant tissues are rich in a variety of nonmetallic elements (C, N and Si etc.), which can be applied as bioprecursors for the synthesis of nonmetal-doped photocatalysts [10].Thus, the resulting TiO 2 synthesized by employing single plant tissues as templates may simultaneously have copied the template structure and incorporated the element, which is highly desirable for photocatalytic applications.For example, Zhang et al. [11] prepared the morph-structured TiO 2 with self-doped N simultaneously by employing green leaf, which has N-rich chlorophylls, as biotemplates.The absorbance intensity of morph-TiO 2 in the visible range increases by 103%-258% compared to that of TiO 2 without a template.In addition, among the doped nonmetallic elements, silica as a dopant is typically used in TiO 2 for photocatalysis.For example, Dorian et al. [12] and Kibombo et al. [13] reported, respectively, that the introduction of SiO 2 could improve not only the specific surface area of TiO 2 photocatalysts, but also its photocatalytic performance in purifying water applications; this result was due to the improved charge carrier separation at TiO 2 -SiO 2 interfaces.Also, Yang et al. [14] and Ma et al. [15] utilized rice husk and corn plant as biotemplates, respectively, to synthesize hierarchical porous TiO 2 /SiO 2 composites.They found enhancement not only in the visible-light harvesting ability of TiO 2 /SiO 2 , but also the photocatalytic activity in the degradation of dye.Recently, Liu et al. [16] constructed hierarchical artificial leaves (TiO 2 -SiO 2 photocatalyst) by using aquatic plant leaves (Vallisneria).The obtained TiO 2 -SiO 2 replicas degraded dye 3 times better than commercial Degussa P25 TiO 2 under UV irradiation.Moreover, during the preparation process, a mesoporous directing agent as a second template and SiO 2 as a mesoporous stabilizer were simultaneously introduced.As known, the aquatic plant leaves are generally thin, soft, and even transparent, which is beneficial to allow access to the light.So far, to our knowledge, the synthesis of TiO 2 /SiO 2 photocatalysts based on a single aquatic plant has not yet been reported until now.
In this work, three typical aquatic plant leaves-including reed (Phragmites australis), water hyacinth (Eichhornia crassipes), and duckweed (Lemna minor L.)-were adopted as both biotemplates and sources of silica to fabricate biomorphic mesoporous TiO 2 /SiO 2 photocatalysts.All TiO 2 /SiO 2 samples exhibit improved light-harvesting abilities and their specific surface area is higher than P25 TiO 2 .The photocatalytic activities were evaluated by using gentian violet as the target for photocatalytic degradation experiments under simulated sunlight irradiation.The results indicate that TiO 2 /SiO 2 samples templated by reed and water hyacinth leaves exhibit high activity, while TiO 2 /SiO 2 samples obtained from duckweed is inferior to P25 in the degradation of gentian violet.A synergistic effect on their photocatalytic activity is proposed.The elemental components and content of the samples, estimated by X-ray fluorescence (XRF) measurement, are shown in Table 1.Aside from Ti and O existing in these samples, 15.66%, 4.22%, and 0.65% of Si was also found in reed (R-TiO 2 /SiO 2 ), water hyacinth (WH-TiO 2 /SiO 2 ), and duckweed (D-TiO 2 /SiO 2 ) samples, respectively.This suggests that SiO 2 in templates was preserved after calcination.Moreover, 3.98% of C was detected in D-TiO 2 /SiO 2 , which could be attributed to the residual duckweed biotemplate after calcination at 450 • C. and 0.65% of Si was also found in reed (R-TiO2/SiO2), water hyacinth (WH-TiO2/SiO2), and duckweed (D-TiO2/SiO2) samples, respectively.This suggests that SiO2 in templates was preserved after calcination.Moreover, 3.98% of C was detected in D-TiO2/SiO2, which could be attributed to the residual duckweed biotemplate after calcination at 450 °C.The wide-angle XRD patterns of the samples are shown in Figure 1.The diffraction peaks of the 2θ values 25.3°, 37.9°, 48.0°, 53.8°, 55.0°, 62.8°, 68.6°, 70.4°, and 75.1° correspond to (101), (004), (200), (105), ( 211), ( 116), ( 204), (220), and (215) crystalline phases of anatase (JCPDS card No. 84-1285), respectively.It is seen that all TiO2/SiO2 composites are mainly made of the anatase phase after calcination at 723 K for 6 h, except a weak peak at 30.8° corresponding to the crystalline phase of brookite in WH-TiO2/SiO2.There are no diffraction peaks of silica or silicates in the XRD patterns, which indicates that the silica structure in the TiO2/SiO2 composite, just as in the original plant leaves, is still amorphous.It has been reported that SiO2 particles in TiO2 grain boundaries are able to inhibit the formation of rutile by restricting the atomic rearrangement during the phase transformation [12,13,17].Thus, the enhanced stability of the anatase phase could be due to the introduction of SiO2.In addition, the stronger (101) diffraction peaks for R-TiO2/SiO2 and WH-TiO2/SiO2 indicated that they possessed a higher crystallinity than D-TiO2/SiO2.The average grain sizes of the TiO2/SiO2 samples, calculated by using the Scherrer equation according to (101) diffraction, are 7.7 nm for WH-TiO2/SiO2, 4.2 nm for R-TiO2/SiO2, and 3.5 nm for D-TiO2/SiO2 (Table 1).It is seen that all TiO 2 /SiO 2 composites are mainly made of the anatase phase after calcination at 723 K for 6 h, except a weak peak at 30.8 • corresponding to the crystalline phase of brookite in WH-TiO 2 /SiO 2 .There are no diffraction peaks of silica or silicates in the XRD patterns, which indicates that the silica structure in the TiO 2 /SiO 2 composite, just as in the original plant leaves, is still amorphous.It has been reported that SiO 2 particles in TiO 2 grain boundaries are able to inhibit the formation of rutile by restricting the atomic rearrangement during the phase transformation [12,13,17].Thus, the enhanced stability of the anatase phase could be due to the introduction of SiO 2 .In addition, the stronger (101) diffraction peaks for R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 indicated that they possessed a higher crystallinity than D-TiO 2 /SiO 2 .The average grain sizes of the TiO 2 /SiO 2 samples, calculated by using the Scherrer equation according to (101) diffraction, are 7.7 nm for WH-TiO 2 /SiO 2 , 4.2 nm for R-TiO 2 /SiO 2 , and 3.5 nm for D-TiO 2 /SiO 2 (Table 1).

Results and Discussion
The Fourier-transform infrared (FTIR) spectra of the biotemplated TiO 2 /SiO 2 samples are shown in Figure 2. A typical Ti-O-Ti stretching vibration in the range of 500-590 cm −1 corresponds to the anatase phase.The absorption peaks at about 1630 and 3430 cm −1 are attributed to the bending and stretching vibrations of surface water molecules.Finally, the bands at 470, 800, and 1096 cm −1 are associated with the Si-O-Si deformation pattern, and symmetric and asymmetric stretching vibration, respectively [13,18].These bands are stronger in the R-TiO 2 /SiO 2 sample than in the other two, indicating that the Si present in the aquatic plant leaves was introduced to the resulting TiO 2 /SiO 2 composites.In particular, a very weak band at 952 cm −1 due to Ti-O-Si linkage could be found in the R-TiO 2 /SiO 2 sample, further suggesting that the SiO 2 was bonded to the TiO 2 particles [13,19,20].The peak intensity of Ti-O-Si and Si-O-Si becomes less pronounced in WH-TiO 2 /SiO 2 and D-TiO 2 /SiO 2 , probably because the low level of SiO 2 is difficult to detect using FTIR.
Catalysts 2017, 7, 82 4 of 13 The Fourier-transform infrared (FTIR) spectra of the biotemplated TiO2/SiO2 samples are shown in Figure 2. A typical Ti-O-Ti stretching vibration in the range of 500-590 cm −1 corresponds to the anatase phase.The absorption peaks at about 1630 and 3430 cm −1 are attributed to the bending and stretching vibrations of surface water molecules.Finally, the bands at 470, 800, and 1096 cm −1 are associated with the Si-O-Si deformation pattern, and symmetric and asymmetric stretching vibration, respectively [13,18].These bands are stronger in the R-TiO2/SiO2 sample than in the other two, indicating that the Si present in the aquatic plant leaves was introduced to the resulting TiO2/SiO2 composites.In particular, a very weak band at 952 cm −1 due to Ti-O-Si linkage could be found in the R-TiO2/SiO2 sample, further suggesting that the SiO2 was bonded to the TiO2 particles [13,19,20].The peak intensity of Ti-O-Si and Si-O-Si becomes less pronounced in WH-TiO2/SiO2 and D-TiO2/SiO2, probably because the low level of SiO2 is difficult to detect using FTIR.The N2 adsorption/desorption isotherms and the pore size distribution of the R-TiO2/SiO2 sample derived from reed leaves are shown in Figure 3.It can be seen in Figure 3a that the N2 sorption isotherms show a type-IV hysteresis loop in the P/P0 range from 0.40 to 1.0, a characteristic of titania-type mesoporous materials [21].Moreover, the hysteresis loops of type H3 suggest that the mesopores may be pile-up pores resulting from particle aggregation [22].This result is further proved by the corresponding Barrett-Joyner-Halenda (BJH) pore size distribution shown in Figure 3b, in which the distribution range of primary pore size is from 3.5 to 30 nm.Similar mesoporous structures were obtained from the WH-TiO2/SiO2 and D-TiO2/SiO2 samples and are shown in Figure S1 (Supplementary Materials).The corresponding pore size distribution is between 5.9 and 30 nm for WH-TiO2/SiO2 samples, and between 6.5 and 30 nm for D-TiO2/SiO2 samples.The wide pore size distribution is due to the natural hierarchical structures of the plant templates [15].The N 2 adsorption/desorption isotherms and the pore size distribution of the R-TiO 2 /SiO 2 sample derived from reed leaves are shown in Figure 3.It can be seen in Figure 3a that the N 2 sorption isotherms show a type-IV hysteresis loop in the P/P 0 range from 0.40 to 1.0, a characteristic of titania-type mesoporous materials [21].Moreover, the hysteresis loops of type H3 suggest that the mesopores may be pile-up pores resulting from particle aggregation [22].This result is further proved by the corresponding Barrett-Joyner-Halenda (BJH) pore size distribution shown in Figure 3b, in which the distribution range of primary pore size is from 3.5 to 30 nm.Similar mesoporous structures were obtained from the WH-TiO 2 /SiO 2 and D-TiO 2 /SiO 2 samples and are shown in Figure S1 (Supplementary Materials).The corresponding pore size distribution is between 5.9 and 30 nm for WH-TiO 2 /SiO 2 samples, and between 6.5 and 30 nm for D-TiO 2 /SiO 2 samples.The wide pore size distribution is due to the natural hierarchical structures of the plant templates [15].
proved by the corresponding Barrett-Joyner-Halenda (BJH) pore size distribution shown in Figure 3b, in which the distribution range of primary pore size is from 3.5 to 30 nm.Similar mesoporous structures were obtained from the WH-TiO2/SiO2 and D-TiO2/SiO2 samples and are shown in Figure S1 (Supplementary Materials).The corresponding pore size distribution is between 5.9 and 30 nm for WH-TiO2/SiO2 samples, and between 6.5 and 30 nm for D-TiO2/SiO2 samples.The wide pore size distribution is due to the natural hierarchical structures of the plant templates [15].As shown in Table 1, the Brunauer-Emmett-Teller (BET) specific surface areas of the TiO 2 /SiO 2 samples are 235, 127, and 139 m 2 /g, corresponding to the R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , D-TiO 2 /SiO 2 samples, respectively.Particularly, the R-TiO 2 /SiO 2 sample is four times more than P25 (50 m 2 •g −1 ) for specific surface area.As known, the biomorphic mesoporous structure is believed to improve the specific surface area of TiO 2 .Besides, it is frequently reported that the introduction of SiO 2 is also capable of increasing the specific surface area of TiO 2 photocatalysts [12,13].As a result, TiO 2 /SiO 2 photocatalysts could provide more adsorption and active sites, and quicken the diffusion and transition of reactants in the photocatalytic process, ultimately benefiting the increase of photocatalytic efficiency [6,7].
Reed, water hyacinth, and duckweed are characteristic of aquatic environments.Figure 4 displays the leaf epidermal micromorphology of the three species, the hierarchical structures of the TiO 2 /SiO 2 composite, and the corresponding spectra of electron diffraction from a selected area.It can be seen in Figure 4a that the original reed leaves have a linearly arrayed skeleton of veins, and vascular bundles with narrow space display a regular location pattern.In particular, many stomata at the same level of epidermal cells are regularly arranged on leaf epiderm.It is worth noting that located between the veins are convexly shaped epidermal cells, which are able to focus light like a lens within the leaf by refraction [23,24].The original water hyacinth leaf and the original duckweed leaf, by contrast, display an irregular skeleton of veins.As seen in Figure 4c, water hyacinth leaf has an interlaced network structure formed by corrugated epidermal cells.The original duckweed leaf in Figure 4e has a disorderly venation, and epidermal cell anticlinal walls are wavy and curved.The morphologies of the prepared TiO 2 /SiO 2 samples derived from original aquatic plant leaves are revealed in Figure 4b,d,f.It can be seen that the TiO 2 /SiO 2 samples are similar to the epidermal microstructures of the original plant leaves, indicating that the morphologies of plant templates were inherited by samples through the biotemplating process.In particular, the Scanning electron microscopy (SEM) image of the white circle in Figure 4b displays clear papillae, just like in the reed leaf.In Figure 4d, the stomas could be observed as indicated by the arrows.It has been reported that the papillae and stomata in plants are helpful for the capture of light.So, it is desirable that TiO 2 /SiO 2 samples replicate the papillae and stomata from plants, therby improving the ability to catch light and providing more adsorption and active sites in the photocatalytic process [23,24].The energy dispersive X-ray spectroscopy (EDS) analyses shown in Figure 4g-i suggest the presence of Ti, Si, and O in the TiO 2 /SiO 2 composite.Particularly, the R-TiO 2 /SiO 2 sample contains more Si content than the others.This is in agreement with the result of XRF.The peaks associated with C in Figure 4g,h should be contributed to the conductive tape used in EDS experiments, while that in Figure 4i is more likely due to the sample itself.In addition, P was detected in all samples, and should be due to the residual phosphorus compounds from biotemplates.Figure 5a-c show the Transmission electron microscopy (TEM) images of the biotemplated TiO2/SiO2 derived from reed, water hyacinth, and duckweed leaves.It was found that all biotemplated TiO2/SiO2 frameworks display the honeycomb mesoporous structure composed of fine particulate, which is consistent with results obtained by N2 adsorption/desorption measurements.In particular, in Figure 5a, the linear and parallel skeleton of veins originating from reed leaf can be observed in R-TiO2/SiO2 frameworks.
The corresponding High-resolution TEM (HRTEM) images of the TiO2/SiO2 samples are displayed in Figure 5d-f, respectively.It can be found in Figure 5d,e that both the R-TiO2/SiO2 and WH-TiO2/SiO2 nanoparticles are of high crystallinity and good particle dispersion.On the contrary, D-TiO2/SiO2 nanoparticles, shown in Figure 5f, exhibit lower crystallinity, and are surrounded by an obvious organic carbon layer.Also, the average crystallite sizes, estimated from Figure 5d-f, are about 4, 8, and 4 nm for R-TiO2/SiO2, WH-TiO2/SiO2, and D-TiO2/SiO2, which is consistent with the result from by XRD analysis.Moreover, the lattice fringes of 3.50 Å shown in Figure 5g correspond to the (101) lattice planes of anatase TiO2. Figure 5a-c show the Transmission electron microscopy (TEM) images of the biotemplated TiO 2 /SiO 2 derived from reed, water hyacinth, and duckweed leaves.It was found that all biotemplated TiO 2 /SiO 2 frameworks display the honeycomb mesoporous structure composed of fine particulate, which is consistent with results obtained by N 2 adsorption/desorption measurements.In particular, in Figure 5a, the linear and parallel skeleton of veins originating from reed leaf can be observed in R-TiO 2 /SiO 2 frameworks.
The corresponding High-resolution TEM (HRTEM) images of the TiO 2 /SiO 2 samples are displayed in Figure 5d-f, respectively.It can be found in Figure 5d,e that both the R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 nanoparticles are of high crystallinity and good particle dispersion.On the contrary, D-TiO 2 /SiO 2 nanoparticles, shown in Figure 5f, exhibit lower crystallinity, and are surrounded by an obvious organic carbon layer.Also, the average crystallite sizes, estimated from Figure 5d-f

Studies on Light-Harvesting and Photocatalytic Activities of TiO2/SiO2 Photocatalysts
UV-vis diffuse reflectance spectra of the TiO2/SiO2 samples along with P25 TiO2 are shown in Figure 6.Compared with P25, the TiO2/SiO2 composites show significant enhancement in the visible-light areas.As everyone knows, neither pure SiO2 nor bare TiO2 exhibit visible-light absorption.All enhancements can be attributed to the biomorphic structure, which means "structure-introduced enhancement", as reported by Li [11] and Ma [15].Briefly, the method here All enhancements can be attributed to the biomorphic structure, which means "structure-introduced enhancement", as reported by Li [11] and Ma [15].Briefly, the method here realizes not only structure replication but also introduction of elements from biotemplates to TiO 2 /SiO 2 samples, and thus the resultant TiO 2 /SiO 2 composites inherit the enhanced visible-light-harvesting of plants.More notably, the blue shifts, as revealed in Figure 6, imply an increase in band gap energy of TiO 2 /SiO 2 samples, which is similar to TiO 2 -SiO 2 composites reported by Ma et al. [15].The corresponding band gap energies estimated by Equation ( 1) are shown in Table 2, where E g is the band gap energy (eV), and λ is the wavelength of adsorption edge (nm).
It can be seen that the band gap energies are about 3.25, 3.22, and 3.21 eV for R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and D-TiO 2 /SiO 2 samples, respectively, which are higher than that of P25 (3.16 eV) and pure anatase (3.2 eV).As known, P25 TiO 2 is composed of both rutile and anatase TiO 2 , whereas the TiO 2 /SiO 2 samples from aquatic plants are almost in pure anatase phase based on the XRD spectra.It is well known that band gap energy of pure anatase is higher than bulk rutile TiO 2 , and thus the unique anatase content of the biotemplated samples result in a raise in band gap energy [15].Besides, Jafry et al. [7] and Malinowska et al. [25] presented that the change in the band gap energy of TiO 2 -SiO 2 composites, prepared by an in situ doping of SiO 2 , results from the reduction of valence band and the increase of conduction band.In this work, the band gap energy seems to be increasing slowly with the change in Si content, which will help to improve the efficiency of electron-hole separation, thus leading to efficient photocatalysis.In short, no extra silicon source is used in our method, and the visible-light-harvesting properties of biotemplate TiO 2 /SiO 2 composites were enhanced.
Catalysts 2017, 7, 82 8 of 13 realizes not only structure replication but also introduction of elements from biotemplates to TiO2/SiO2 samples, and thus the resultant TiO2/SiO2 composites inherit the enhanced visible-light-harvesting of plants.More notably, the blue shifts, as revealed in Figure 6, imply an increase in band gap energy of TiO2/SiO2 samples, which is similar to TiO2-SiO2 composites reported by Ma et al. [15].The corresponding band gap energies estimated by Equation ( 1) are shown in Table 2, where Eg is the band gap energy (eV), and λ is the wavelength of adsorption edge (nm).
It can be seen that the band gap energies are about 3.25, 3.22, and 3.21 eV for R-TiO2/SiO2, WH-TiO2/SiO2, and D-TiO2/SiO2 samples, respectively, which are higher than that of P25 (3.16 eV) and pure anatase (3.2 eV).As known, P25 TiO2 is composed of both rutile and anatase TiO2, whereas the TiO2/SiO2 samples from aquatic plants are almost in pure anatase phase based on the XRD spectra.It is well known that band gap energy of pure anatase is higher than bulk rutile TiO2, and thus the unique anatase content of the biotemplated samples result in a raise in band gap energy [15].Besides, Jafry et al. [7] and Malinowska et al. [25] presented that the change in the band gap energy of TiO2-SiO2 composites, prepared by an in situ doping of SiO2, results from the reduction of valence band and the increase of conduction band.In this work, the band gap energy seems to be increasing slowly with the change in Si content, which will help to improve the efficiency of electron-hole separation, thus leading to efficient photocatalysis.In short, no extra silicon source is used in our method, and the visible-light-harvesting properties of biotemplate TiO2/SiO2 composites were enhanced.Photocatalytic activities were studied by degrading gentian violet (GV) dye in the presence of TiO2/SiO2 samples, P25, and TiO2-p (TiO2 particles prepared without template) under simulated sunlight irradiation.To achieve real photocatalytic degradation owing to photocatalysis in the presence of photocatalysts, the adsorption yield of dye was investigated.When adsorptiondesorption equilibrium was reached after 12 h in the dark, the adsorption yields (shown in Table 3)  Photocatalytic activities were studied by degrading gentian violet (GV) dye in the presence of TiO 2 /SiO 2 samples, P25, and TiO 2 -p (TiO 2 particles prepared without template) under simulated sunlight irradiation.To achieve real photocatalytic degradation owing to photocatalysis in the presence of photocatalysts, the adsorption yield of dye was investigated.When adsorption-desorption equilibrium was reached after 12 h in the dark, the adsorption yields (shown in Table 3) are 62.4%, 51.5%, and 43% for R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and D-TiO 2 /SiO 2 samples, respectively, which is much higher than 7.4% for P25 and 20.6% for TiO 2 -p.The strong adsorption effect is relative to their higher specific surface area, as reported by Anderson et al. [6].After deducting the decreases of dye concentration due to adsorption and direct photolysis, the real photocatalytic degradation owing to the photocatalysis in the presence of photocatalysts was obtained and is shown in Figure 7.The degradation yields of biotemplated samples after 5 h are 72.0%,61.4%, and 11.7% for the R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and D-TiO 2 /SiO 2 composites.In particular, R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 samples possess better photocatalytic activities than P25 TiO 2 and TiO 2 -p.For the D-TiO 2 /SiO 2 sample, the degradation yield increases to 15.6% in 3 h, then slowly decreases to 11.7% in 5 h, which is lower than P25 TiO 2 and TiO 2 -p.The results indicate that biotemplating may play an important role in improving photocatalytic activity, and that the activity is closely related to the plant species used as templates to duplicate photocatalysts.
Catalysts 2017, 7, 82 9 of 13 are 62.4%, 51.5%, and 43% for R-TiO2/SiO2, WH-TiO2/SiO2, and D-TiO2/SiO2 samples, respectively, which is much higher than 7.4% for P25 and 20.6% for TiO2-p.The strong adsorption effect is relative to their higher specific surface area, as reported by Anderson et al. [6].After deducting the decreases of dye concentration due to adsorption and direct photolysis, the real photocatalytic degradation owing to the photocatalysis in the presence of photocatalysts was obtained and is shown in Figure 7.The degradation yields of biotemplated samples after 5 h are 72.0%,61.4%, and 11.7% for the R-TiO2/SiO2, WH-TiO2/SiO2, and D-TiO2/SiO2 composites.In particular, R-TiO2/SiO2 and WH-TiO2/SiO2 samples possess better photocatalytic activities than P25 TiO2 and TiO2-p.For the D-TiO2/SiO2 sample, the degradation yield increases to 15.6% in 3 h, then slowly decreases to 11.7% in 5 h, which is lower than P25 TiO2 and TiO2-p.The results indicate that biotemplating may play an important role in improving photocatalytic activity, and that the activity is closely related to the plant species used as templates to duplicate photocatalysts.According to the experimental results, it can be deduced that the SiO2 incorporation and structural construction should be the primary factors devoted to photocatalytic performance enhancement.On the one hand, it should be noted that the introduction of Si has been reported to improve not only the specific surface area but also the visible-light-harvesting of TiO2/SiO2 [14,15].In this work, with Si introduced, the TiO2/SiO2 samples, particularly R-TiO2/SiO2 composites, show a large specific surface area and enhanced visible-light harvesting.Additionally, the photocatalytic activities were improved successively with Si content in the order of D-TiO2/SiO2, WH-TiO2/SiO2, and R-TiO2/SiO2 composites.On the other hand, as described above, the unique structural features derived from the natural plant leaves endow them with a great light-adsorbing capability.In particular, the papillae and stomatas in leaves were duplicated, respectively, to R-TiO2/SiO2 and According to the experimental results, it can be deduced that the SiO 2 incorporation and structural construction should be the primary factors devoted to photocatalytic performance enhancement.On the one hand, it should be noted that the introduction of Si has been reported to improve not only the specific surface area but also the visible-light-harvesting of TiO 2 /SiO 2 [14,15].In this work, with Si introduced, the TiO 2 /SiO 2 samples, particularly R-TiO 2 /SiO 2 composites, show a large specific surface area and enhanced visible-light harvesting.Additionally, the photocatalytic activities were improved successively with Si content in the order of D-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and R-TiO 2 /SiO 2 composites.On the other hand, as described above, the unique structural features derived from the natural plant leaves endow them with a great light-adsorbing capability.In particular, the papillae and stomatas in leaves were duplicated, respectively, to R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 samples.Such structures are favorable for light harvesting [11,15].Also, the highly crystalline mesoporous structure in R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 catalysts, which also originated from the natural hierarchical structures of the plant templates, endowed them larger specific surface areas, and more adsorption and active sites, which are conducive to quickening mass transport and improving the efficiency of electron-hole separation [26].Besides, the increasing band gaps in TiO 2 /SiO 2 samples as a result of the structure and introduction of Si could reduce electron-hole recombination, and help to capture more photoinduced holes during the photocatalytic process [26,27].Therefore, in this work, the photocatalytic activity of R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 samples was higher than that of P25 TiO 2 and TiO 2 -p.In the case of D-TiO 2 /SiO 2 samples, although it possesses high specific surface area and band gaps, the negligible Si content appears to dampen the effect on its photocatalytic activity.In particular, the presence of a large amount of organic carbon led to its lower crystalline structure, which is not conducive to improving the photocatalytic activity.Besides, lack of special structural characteristics for original duckweed may also be a factor that leads to the poorer photocatalytic activity compared to R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 .As a result, a synergistic effect of SiO 2 incorporation and structural construction through biotemplating is suggested to be beneficial to photocatalytic activity.

Synthesis
TiO 2 samples were synthesized by using a sol-gel process according to our previous works [28].Firstly, fresh leaves were carefully rinsed with deionized water several times to remove the plankton and other impurities adsorbed on the surface.Then, the leaves were cut into little pieces and soaked in 2% glutaraldehyde/phosphate-buffered saline (PBS; pH = 7.2) solution at 4 • C for 12 h to fix the structures of plants.Secondly, the fixed leaves were immersed in HCl solution of 5% for 3 h until the green leaves became yellow-green, then, immersed in 5% TiCl 3 solution for 1 day.The as-treated leaves were gradually dehydrated in ethanol (30%, 60%, 90%, and 100%, respectively) and then were treated by sol infiltration of TTIP/ethanol/ACAC solution (volume ratio of 1:29:0.1)for 3 days.After filtration and washing with ethanol several times to remove any superfluous titanium source completely, the infiltrated samples were hydrolyzed in air for 4 days.Finally, the hydrolyzed samples were dried at 90 • C for 12 h, calcined at 280 • C for 2 h and at 450 • C for 6 h to remove the biotemplates and crystallize the TiO 2 .Finally, the, grey-white products were formed.The samples templated by reed leaves, water hyacinth and duckweed are designated as R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and D-TiO 2 /SiO 2 , respectively.For comparison, the TiO 2 -p particles were prepared by the same sol-gel method in the absence of any template.

Characterizations
Powder X-ray diffraction (XRD) was performed on a TTR III spectrometer with Cu Kα radiation (Rigaku D/max-3B, Tokyo, Japan).Scanning electron microscopy (SEM) images of the samples were taken on a FEIQuanta 200FEG microscope (FEI, Eindhoven, Netherlands) with energy dispersive X-ray spectroscopy (EDS, Oxford INCA, Abingdon, Oxfordshire, UK) at 5 kV.X-ray fluorescence (XRF) analysis was measured on a ZSX100E spectrometer (Rigaku, Tokyo, Japan).Pore size distributions, BET surface areas, and BJH pore size distribution were determined by nitrogen adsorption-desorption using a Tristar II 3020 surface area and porosity analyzer (Micromeritics, Norcross, GA, USA).Fourier-transform infrared spectrum (FTIR) measurement was conducted on a Thermo Nicolet
Figure5a-cshow the Transmission electron microscopy (TEM) images of the biotemplated TiO 2 /SiO 2 derived from reed, water hyacinth, and duckweed leaves.It was found that all biotemplated TiO 2 /SiO 2 frameworks display the honeycomb mesoporous structure composed of fine particulate, which is consistent with results obtained by N 2 adsorption/desorption measurements.In particular, in Figure5a, the linear and parallel skeleton of veins originating from reed leaf can be observed in R-TiO 2 /SiO 2 frameworks.The corresponding High-resolution TEM (HRTEM) images of the TiO 2 /SiO 2 samples are displayed in Figure5d-f, respectively.It can be found in Figure5d,e that both the R-TiO 2 /SiO 2 and WH-TiO 2 /SiO 2 nanoparticles are of high crystallinity and good particle dispersion.On the contrary, D-TiO 2 /SiO 2 nanoparticles, shown in Figure5f, exhibit lower crystallinity, and are surrounded by an obvious organic carbon layer.Also, the average crystallite sizes, estimated from Figure5d-f, are about 4, 8, and 4 nm for R-TiO 2 /SiO 2 , WH-TiO 2 /SiO 2 , and D-TiO 2 /SiO 2 , which is consistent with the result from by XRD analysis.Moreover, the lattice fringes of 3.50 Å shown in Figure 5g correspond to the (101) lattice planes of anatase TiO 2 .

Table 1 .
Elemental content and structural characteristics of the samples.

Table 1 .
Elemental content and structural characteristics of the samples.

Table 2 .
Band gap energy calculated by UV-vis diffuse reflectance spectra.

Table 2 .
Band gap energy calculated by UV-vis diffuse reflectance spectra.

Table 3 .
Adsorption and degradation yields of gentian violet (GV) over the biotemplated samples and P25.

Table 3 .
Adsorption and degradation yields of gentian violet (GV) over the biotemplated samples and P25.