Preparation of TiO 2 Grafted on Graphene and Study on their Photocatalytic Properties

. TiO 2 has potential application prospects in the ﬁ elds of environmental pollution control and energy conversion, while its characteristics including proneness to agglomeration and low activity a ﬀ ect practical applications. In view of this, P25 TiO 2 and graphene were taken as raw materials to prepare the modi ﬁ ed TiO 2 catalyst. The modi ﬁ cation mechanism of TiO 2 was explained by characterizing the phase composition, microscopic morphology, and functional groups and analyzing the speci ﬁ c surface area and particle size. Then, the photocatalytic performance of the material was explored by taking methyl orange as the target to be degraded. Results show that TiO 2 is well grafted to the laminated structure of graphene using physicochemical interactions among the introduced functional groups after modifying P25 TiO 2 and graphene. This way, the 10.00mg · L -1 methyl orange solution is almost completely degraded within 12 min at a reaction rate that is 1.81 times higher than that of P25 TiO 2 . Even after being used for 10 times, the performance still remains stable. The modi ﬁ cation process is simple, and the method is reliable. Results can promote the practical application of TiO 2 photocatalytic technology.


Introduction
To reach the carbon peaking and carbon neutrality goals, developing new materials has become a major strategic target of the Made in China 2025 Initiative. TiO 2 , as a new material that receives extensive concern, produces carriers (cavities and electrons) after being excited. The carriers may induce a slew of chemical reactions and play the catalytic action in environmental pollution control [1][2][3][4][5] and energy conversion [6,7]. However, electrons and cavities are very likely to be recombined, which lowers the catalytic efficiency. TiO 2 is prone to agglomeration, which further weakens its catalytic performance, so it still cannot be used in practical applications. Graphene, also as a new material, plays a role in electron transfer [8] and is a popular material in energy storage and catalysis fields. Previous research has demonstrated that graphene can produce lots of pores on the nanosheets after modification [9]. Such special features are expected to solve the bottleneck problem in the application of TiO 2 .
TiO 2 shows catalytic activity only when its carriers are captured by and reacted with graphene, which requires TiO 2 and graphene both to have special morphologies and form structurally stable compounds. To this end, numerous scholars have attempted to use many methods to prepare compounds, including the sol-gel method [10], hydrothermal treatment [11], and in situ synthesis [12]. However, the prepared compounds generally have poor dispersity and nonuniform particle sizes and shape, or they call for complicated processes, harsh conditions, and high costs. Therefore, these compounds cannot be scaled up, realize economic efficiency, and used in practical applications. In view of this, the research took P25 TiO 2 and graphene as raw materials to perform surface modification. Based on the physicochemical interactions among the functional groups introduced, TiO 2 is grafted on the surface of graphene, thus obtaining a modified TiO 2 catalyst. Afterwards, the catalytic activity and stability were tested, and the structure-function relationship of TiO 2 in the modification process was discussed.
Catalysts were characterized and identified by means of attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR, Thermo Scientific Nicolet iS50), X-ray diffraction (XRD, Bruker D8 Advance), full automatic specific surface and aperture analyzer (Micromeritics ASAP2460), and scanning electron microscopy (SEM, Coxem Oxford, EM-30 N).  [2] and then preserved for later use and testing. A certain amount of treated P25 TiO 2 was weighed and experienced silanization [13], washing, filtration, and vacuum drying successively to obtain samples (marked as @TiO 2 ), which were preserved for later use and testing.

Graphene Modification.
A certain amount of graphene diluent was weighed at first, then washed, filtrated, and freeze-dried to obtain granular graphene (marked as GR), which was uniformly dispersed in a mixed solution of concentrated sulfuric acid and concentrated nitric acid with a matching ratio of 1 : 1 (v·v -1 ). Then, a certain amount of potassium permanganate was added in the solution to allow the reaction for 1 h. Afterwards, the products were treated via ultrasound, washed, filtrated, and freeze-dried for later use. A certain amount of the products was weighed, in which slight hydrazine hydrate was added to react for 30 min. The sample obtained after washing, filtration, and freeze-drying was termed as RGO, which was preserved for later use and testing.

Preparation of Modified TiO 2 .
Certain amounts of P25 TiO 2 , @TiO 2 , and RGO were weighed, mixed according to a ratio of 1 : 1 (m·m -1 ), and then sufficiently ground in an agate mortar. The resulting products were parceled with a tinfoil and kept in the dark for 24 h, thus preparing the modified TiO 2 (marked as @TiO 2 /RGO). The composite samples of P25 TiO 2 and RGO (marked as P25 TiO 2 /RGO) were preserved for later use and testing.
2.3. Photocatalytic Experiment. The photocatalytic activity and stability of modified TiO 2 were characterized by taking 10.00 mg·L -1 methyl orange (initial pH, 7:80 ± 0:2) as the target to be degraded. Under a 300 W high-pressure mercury lamp (the average light intensity in solutions was about 8 MW/cm 2 ) which served as a light source and a catalyst dosage of 5 g·L -1 , the reaction was performed in a stirring reactor [13]. The temperature of the reactor was regulated at 26 ± 0:5°C by a circulating water system. After 12 min irradiation, the final pH values of samples were determined. The concentration of methyl orange was analyzed by using a 722 visible light spectrophotometer in a cuvette with the thickness of 1 cm at the wavelength of 465 nm. The degradation efficiency for methyl orange can be calculated by the following equation: where η is moment degradation efficiency, C 0 and C t are the initial concentration and the moment concentration of methyl orange in mg/L, respectively.     Figure 1 shows that P25 TiO 2 is a mixed crystal of anatase and rutile. Through hydrothermal synthesis (high-temperature and high-pressure reaction), a small amount of rutile is transformed into      International Journal of Photoenergy anatase. Previous research demonstrates that the transformation is conducive to the improvement of the activity [14]. In comparison, the silanization does not significantly influence the crystal form of TiO 2 but still remains the mixed crystal structure. As shown in Figure 2, P25 TiO 2 does not have any absorption peak, while a strong absorption peak appears at 3389.9 cm -1 after hydrothermal synthesis, which is a stretching vibration peak of O-H. This indicates that hydroxyl groups are introduced on the TiO 2 surface after hydrothermal synthesis. For @TiO 2 , an absorption peak of Ti-O-Si appears at 918.7 cm -1 [15], suggesting the presence of a condensation reaction between silanol (the hydrolysis product of the silane coupling agent) and hydroxyl groups on the TiO 2 surface. This way, aminopropyl siloxane containing amidogen groups is introduced to the surface of TiO 2 . Because amidogen groups carry positive charges, the modified TiO 2 is positively charged [16]. Peaks at 2848.6, 2916.8, and 2960.6 cm -1 are stretching vibration peaks of methene (-CH 2 ). It can be seen in Figure 3(a) that P25 TiO 2 is nonuniformly distributed and agglomerated; Figure 3(b) depicts that TiO 2 prepared through hydrothermal synthesis is an acicular crystal, which is then coarsened and becomes rod-like or tubular after silanization, with protuberant ends. This is probably because aminopropyl siloxane containing amidogen groups is grafted on the TiO 2 surface, as displayed in Figures 3(c) and 3(d). According to Table 1, P25 TiO 2 is transformed from a granular state with the average particle size of 21 nm to rod-like or tubular @TiO 2 with the particle size of 20~200 nm and length of 1~4 μm. Under the condition, the specific surface area of TiO 2 enlarges by 31.5%, and TiO 2 becomes less prone to agglomeration. Theoretically, these phenomena are all favorable for improving the catalytic performance.  Table 2. It can be seen in Figure 4 that compared with GR, RGO has a higher and unstable baseline with widened peaks, which is probably because the substitution effect of oxygen atoms is introduced  International Journal of Photoenergy due to oxidation. As a result, the carbon atoms combined with the oxygen atoms differ from other carbon atoms. Such a difference impairs the original stress equilibrium of carbon atoms, causes lattice distortion, and forms lattice defects, thus leading to widened and weakened diffraction peaks. Formation of defects is favorable for improving the electron transfer capacity of graphene [8]. Graphene is usually stored in the diluent, which mainly contains pennyroyal tea and castor leaf. After washing before usage, there are still probably some residual impurities. Therefore, it can be seen in Figure 5 that the peak at 1791.3 cm -1 is a stretching vibration peak of C=O in the molecular structures of pennyroyal tea and castor leaf; the peak at 2117.1 cm -1 is probably ascribed to the residual impurities; these peaks all disappear after oxidation. Oxygen-containing groups such as hydroxyl and carboxy groups are interpenetrated on the surface of RGO.
Owing to the presence of these functional groups, the RGO surface is negatively charged [16]. As displayed in Figures 6(a) and 6(b), GR is indeed agglomerated, while tightly bonded laminated structures are also observed, with small pores distributed on the surface. According to Table 2, the pore size is distributed in the range of 2~31 nm, which mainly are mesopores that are formed due to the accumulation of particles. As shown in Figures 6(c) and 6(d), the modification did not destroy the lamellar structure of GR which is more seriously stripped; leading to the laminated structure becomes increasingly obvious. Sheet-like lamellas become larger, thinner, and transparent, without agglomeration. As shown in Figure 7(a) and Table 2, the modification changes the specific surface area, pore volume, and structure of the material, making RGO and GR present obvious differences. It may be due to the difference between carbon atoms combined with oxygen and other carbon atoms, stress nonequilibrium occurs, and several irregularly distributed folds are formed on the surface. It can be clearly seen in Figure 6(d) that new channels are formed by these folds, which is the main cause for the gradual increase in the num-ber of macropores and micropores in the modification process of graphene. However, as shown in Figure 7(b) and Table 2, the total number of micropores is still small and the channels are mainly composed of macropores and mesopores that are formed because of the accumulation of lamellas. The pore sizes are mainly in the ranges of 2~47 nm and 80~207 nm, and the specific surface area enlarges by 45.2%. Theoretically, all these changes are conducive to the improvement of the properties of graphene.

Characterization of Composite Materials.
Two composite samples, P25TiO 2 /RGO and @TiO 2 /RGO, were characterized through FTIR and SEM, and the results are displayed in Figures 8 and 9. It can be seen in Figure    International Journal of Photoenergy very approximated to those of RGO in Figure 5 (3206.9, 1622.0, and 1041.5 cm -1 ). No new absorption peaks appear, which means that no physicochemical interactions occur even if P25 TiO 2 is mixed with RGO that contains abundant oxygencontaining groups, as evinced by the agglomeration and nonuniform distribution in Figure 9(a). Figure 9(b) shows that TiO 2 particles are simply accumulated on the surface of RGO. When @TiO 2 and RGO are mixed, a new peak appears at 2323.9 cm -1 , which belongs to the absorption peak of ester groups, indicative of the condensation reaction between RGO and TiO 2 . According to the analysis in Section 3.2, lots of negatively charged oxygen-containing groups are interpenetrated on the surface of RGO. As mentioned in Section 3.1, amidogen groups with positive charges are introduced on the @TiO 2 surface. The two also have electrostatic attraction. Therefore, TiO 2 is grafted on the laminated structure of RGO under the dual effects of electrostatic attraction and condensation reaction, thus obtaining the modified TiO 2 . It can be seen in Figure 9(c) that the sample is uniformly distributed and not agglomerated. As illustrated in Figure 9(d), rod-like or tubular TiO 2 is uniformly distributed on the surface of RGO.

Effect of Solution pH on the Degradation Kinetics.
The initial and final pH values of samples were summarized as displayed in Table 3. The results show that the pH of the sys-tem will increase with the photodegradation process of methyl orange. As a result, the system's alkalinity will gradually be increased. Such a phenomenon can be attributed to hydrogen generation by the following reaction:    International Journal of Photoenergy indicating that H + would be reduced by electrons which may be transferred from TiO 2 to graphene surface. The hydrogen ions are mainly generated from the hydrolysis of methyl orange which will be inhibited by the rise of pH. Moreover, the increase of pH will affect the forms and adsorption of ionizable compounds [17]. Specifically, it also greatly influences the surface properties of the catalyst, which can be described through Equation (3), showing that pH will affect the redox ability of the photocatalyst Apparently, the increase of solution pH was not conducive for the removal of methyl orange.

Photocatalytic Performance of Modified TiO 2 .
To characterize the catalytic activity of modified TiO 2 (@TiO 2 /RGO), the photocatalytic degradation experiments on methyl orange were conducted following the steps in Section 2.3 taking P25TiO 2 (P25TiO 2 /RGO) as the control. Because the reaction at a low concentration is approximated to the first-order catalytic reaction [(2)], ln ðC 0 /C t Þ was used to draw plots about the reaction time ( Figure 10). Kinetic parameters are calculated according to the first-order reaction, as listed in Table 4. It shows that the degradation rate of methyl orange by modified TiO 2 is 1.81 times, 0.52 times higher than that by P25 TiO 2 and immobilized PU/TiO 2 [18], respectively. The removal efficiency of methyl orange by modified TiO 2 within 12 min is also improved by 0.42 times, suggesting the better catalytic activity of modified TiO 2 .
In general, catalytic activity will be declined when the catalyst is repeatedly used. In response to this situation, the photocatalytic stability was investigated. Experiments on photocatalytic degradation of methyl orange were carried out for ten times repeatedly. Then, the catalyst was centrifugalized and freeze-dried. It was weighted to be 0.0786 g. Changes in the removal efficiency with the times of repetition are shown in Figure 11. As shown in the figure, the degradation efficiency of methyl orange reduces slightly after each use, while the reduction of removal efficiency is closely correlated with the loss of catalyst (R 2 = 0:998). This indicates that the reduction of catalytic efficiency results from the loss of the catalyst, rather than the falling off and inactivation of the catalyst, which indicates the high stability of the modified TiO 2 .
As mentioned in Section 3.4, increasing pH is unconducive to the photocatalytic degradation of methyl orange. However, the composite catalyst still exhibits good catalytic activity. As shown in Figures 9(c) and 9(d), a complex with intimate interface was obtained by integrating graphene with TiO 2, which invests the sample with prominent photocata-lytic abilities. Such a phenomenon can be explained by the fact that the modification imparts better catalytic performance to the material (for example, the introduced hydroxyl groups also have catalytic activity, enlarged specific surface area, and are not easily agglomerated, as shown in Section 3.1). Apart from this, graphene itself is characterized by π-π stacking, which provides stronger adsorption advantages for materials containing benzene rings such as methyl orange [1]. And adsorption is the primary condition for the occurrence of catalytic reactions [13].
Moreover, graphene itself has excellent mechanical, thermal, optical, and electrical properties, and its performance is further optimized after modification: (1) it has a larger specific surface and better dispersity, which avoids TiO 2 shield effective absorption of photons due to proneness to agglomeration and enlarges the contact interface between graphene and TiO 2 , as shown in Section 3.2. (2) Graphene alters the energy band structure of TiO 2 in the complex, reducing the band gap width from 3.20 eV to 2.64 eV [19]. Similarly, graphene in the compound can change the energy band structure of TiO 2 , which has been verified in N-TiO 2 /graphene, decreasing the width from 2.92 eV to 2.73 eV [20]. With the decrease in the band gap energy of TiO 2 , more photogenerated electrons reach their escape work and then are transferred to graphene surface via the composite interface [8,21]. Mott-Schottky diagram and DOS images can verify the increment in density of carriers [22,23]. 3) As can be seen in Section 3.2, there are abundant narrow holes distributed on graphene surface. This is because modification can introduce oxygen atoms into graphene, inducing the graphene to generate lattice defects. As a consequence, k a is the first-order kinetic constant obtained using ln ðC 0 /C t Þ = kt, RSQ b = R-squared.

100
The degradation rate of methyl orange (%)  International Journal of Photoenergy multichannel structures were formed on graphene surface, which change [24] and shorten [25] the transfer paths of charges and speed up the charge transfer rate.
In summary, multiple factors collaboratively promote interfacial charge separation. The reaction as shown in Equation (2) occurs to partial electrons transferred, while some electrons are subject to the reaction as presented in Equation (4), for the more holes remained, which are mainly consumed through Equation (5). Equations (2), (4), and (5) are beneficial to inhibit the recombination of electron-hole pairs. This makes the photocarriers to generate more active species [22].
Bell et al. [26] consider that these active substances, mainly ·O 2 − and ·OH, have high activity, significantly improving the photocatalytic activity of the composite catalyst. In order to identify the active oxidation species, the quenching experiment of isopropyl alcohol was conducted [27] (eliminating hydroxyl radical), as shown in Figure 12. The results showed that modified TiO 2 produced a large amount of ·OH under ultraviolet light. In summary, as discussed above, various factors synergize to reinforce the catalytic activity of TiO 2 . The improvement mechanism of the catalytic performance is illustrated in Figure 13.

Conclusions
In a nutshell, it can be concluded that (1) by modifying P25 TiO 2 , hydroxyl, and amidogen groups are successively introduced to the surface, and the material turns from a granular one to a rod-like or tubular one, with a small amount of rutile transformed into anatase. After modifying graphene, oxygen-containing groups are introduced to the surface, and the laminated structure is stripped off sufficiently. At the same time, lattice defects appear, and new multichannel structures are formed. Through physicochemical interactions, TiO 2 is grafted on the laminated structure of graphene, thus obtaining the modified TiO 2 catalyst. (2) The modified TiO 2 is able to almost completely degrade the 10.00 mg·L -1 methyl orange solution within 12 min. The reaction rate is improved by 1.81 times compared with P25 TiO 2 , and the catalytic performance still remains stable even after repetitious usage. This solves the problems of proneness to agglomeration and low activity of TiO 2 in practical applications effectively.

Data Availability
Data are available on request from the authors. The raw/ processed data for these findings cannot be shared at this time as the data are also part of an ongoing study.

Conflicts of Interest
The authors declare that they have no conflicts of interest.    Figure 13: Schematic diagram of TiO 2 photocatalytic performance improvement mechanism. 8 International Journal of Photoenergy