Structure and photocatalytic properties of TiO2-Graphene Oxide intercalated composite

TiO2-Graphene Oxide intercalated composite (TiO2-Graphene Oxide) has been successfully prepared at low temperature (80°C) with graphite oxide (GO) and titanium sulfate (Ti(SO4)2) as initial reactants. GO was firstly exfoliated by NaOH and formed single and multi-layered graphite oxide mixture which can be defined as graphene oxide, [TiO]2+ induced by the hydrolysis of Ti(SO4)2 diffused into graphene oxide interlayer by electrostatic attraction. The nucleation and growth of TiO2 crystallites took place at low temperature and TiO2-Graphene Oxide composite was successfully synthesized. Furthermore, the photocatalytic properties of TiO2-Graphene Oxide under the irradiation of UV light were also studied. The results show that the degradation rate of methyl orange is 1.16 mg min−1 g−1(refer to the efficiency of the initial 15 min). Compared with P25 powder, this kind of intercalation composite owns much better efficiency. On the other hand, the reusable properties and stable properties of TiO2-Graphene Oxide intercalated composite are also discussed in this paper. At last, crystalline structure, interface status, thermal properties and microscopic structure of TiO2-Graphene Oxide were characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), field emission scanning electron microscopy (FESEM) and high-resolution Transmission Electron Microscopy (HRTEM). Also, we have analyzed major influencing factors and mechanism of the composite structures which evidently improve the photocatalytic properties.

Nano titanium dioxide (TiO 2 ) has attracted great attention these years and it is widely used in photocatalytic materials [1,2], solar cells [3], gas sensors [4], opto-electronics devices [5] and other areas. The photocatalytic activity of TiO 2 depends on its crystal phase, crystallite size and crystallinity. It is known that the anatase-form of TiO 2 crystallite with small crystallite size and high crystallinity exhibits better photocatalytic property [6,7]. The overall efficiency of photon utilization by TiO 2 is however limited by electron-hole recombination, photon scattering and so on [8].
In recent years, great efforts have been made to develop carbon nanotubes (CNTs)/TiO 2 or C 60 /TiO 2 nanocomposite *Corresponding author (email: heyunqiu@mail.tongji.edu.cn) [9−13]. The coupling of TiO 2 with CNTs or C 60 has been shown to provide a synergistic effect which can enhance the overall efficiency of a photocatalytic process. Gao et al. [10] prepared CNTs/TiO 2 nanocomposite by a surfactant wrapping sol-gel method, and evaluated its activity towards the photocatalytic degradation of a model organic substance (methylene blue) in aqueous suspensions. The results show that CNTs/TiO 2 nanocomposite owns a twice rate enhancement with respect to TiO 2 alone. Oh et al. [11] synthesized two kinds of nanocarbon/TiO 2 composite photocatalysts, employing CNTs and C 60 as nanocarbon sources and titanium (IV) n-butoxide (TNB) as a titanium dioxide source. CNTs/TiO 2 nanocomposite owns much better efficiency, for the reason that it showed homogenous distribu-tions with only individual CNTs, covered with TiO 2 and without any aggregates between each other. Gao et al. summarized several causes for the superior photocatalytic properties of nanocarbon/TiO 2 composite: (i) the higher surface area of the composite providing a higher adsorption capacity of reactive species; (ii) the formation of nanocarbon/TiO 2 heterojunctions which could reduce the rate of the recombination of photoinduced electrons and holes; (iii) the possible shift of the apparent Fermi level to more positive values of the nanocarbon/TiO 2 nanocomposite as compared to pure TiO 2 , therefore allowing the utilization of longer wavelength photons; and (iv) the absorption of photons by the bare nanocarbon surface injecting electrons into the TiO 2 conduction band and triggering the formation of reacting radicals (superoxide and hydroxyl radicals).
Graphite Oxide (GO) is a non-stoichiometric material in which the lamellar structure of graphite is conserved. It is obtained through the reaction of graphite with strong oxidants such as potassium permanganate in concentrated sulfuric acid [14−18]. The process of strong oxidation makes many oxygen-containing groups such as carboxyl (C=O), hydroxyl (C−OH), epoxide (C−O−C) covalently attached to its layers [19−21]. In addition, the spatial distribution of functional groups on the carbon skeleton is also different. The epoxide and hydroxyl groups are near each other, and the carboxyl is more likely to be located at the edges of graphite oxide. These oxygen-containing groups embedded in carbon sheets in GO lamellar make the interlayer distance increase from 0.335 nm of the starting graphite to 0.7−1.0 nm and the distance depends on the degree of oxidation and the interlayer water content of GO [22]. Furthermore, oxygen-containing groups also enable GO to exhibit excellent hydrophilic properties and rich intercalation chemistry [23,24]． Much interest has been focused on nano-scale materials based on layered GO and conducting polymers [25−27] since the resulting nanocomposite could possess physical properties derived synergistically from both components, such as enhancement in electrical conductivity and mechanical strength [28]. Higashika et al. [26] firstly prepared aniline-GO composite, and polyaniline/GO composite was successfully synthesized by using an in-situ polymerization method. However, Rabin et al. [28] demonstrated that polyanilines could be directly inserted into GO without preparation of precursor intercalation compounds. The synthetic methodology for the insertion could be extended to other conductive polymers such as polypyrrole [27]. These nanomaterials can be used in rechargeable battery electrode materials.
In this study, TiO 2 -Graphene Oxide composite was successfully prepared at low temperature (80°C) with single and multi-layered graphite oxide mixture (graphene oxide) and titanium sulfate (Ti(SO 4 ) 2 ) was used as initial reactants. Plenty of oxygen-containing groups(such as C=O) located in the interlayer of graphene oxide were consumed during the nucleation and growth of TiO 2 crystallites, which means graphene oxide has been partly reduced. In addition, the activity of TiO 2 -Graphene Oxide towards the photocatalytic degradation of methyl orange in aqueous suspensions is much more efficient than that of Degussa P25 powder. In the end, we discussed the reusable and stable properties and the main reasons responsible for the superior photocatalytic properties of TiO 2 -Graphene Oxide.
1 Experimental section

Materials
Reagents included concentrated sulfuric acid and sodium hydroxide were of analytic purity except titanium sulfate (chemical pure), purchased from Sinopharm Chemical Reagent Co, Ltd(SCRC) and without further purification before utilization. High purity deionized water made by ourselves was used as solvent. GO was prepared by modified Hummer method, the previous work carried out by our group [29]. P25 powders were purchased from Evonik Degussa. GO was exfoliated by sodium hydroxide solution with different concentrations. A colloidal solution of graphene oxide was obtained by mixing 10 mL of GO with 10 mL of sodium hydroxide solution and sonicating for 30 min.

Preparation of TiO 2 -Graphene Oxide
Graphene oxide solution was added immediately into Ti(SO 4 ) 2 acid solution which was preheated to 80°C in hot water bath. The resulting solution was stirred at 80°C for 2 h, and the dark precipitate of TiO 2 -Graphene Oxide was filtered out, washed with deionized water repeatedly and dried at 80°C for 12 h.

Photocatalytic experiments
The photocatalytic activity of the as-prepared samples was characterized under UV light irradiation using methyl orange (MO) as a target. The process was as follows: 50 mg of the sample was dispersed in 50 mL of aqueous MO solution with the concentration of 20 mg L −1 . Fresh air was bubbled for the purpose of stirring and oxygen supply. High pressure Hg lamp (250 W, λ ≥ 365 nm) was used as a UV source. Suspensions were collected every 15 min to monitor the degradation of MO. The transmittance (T i ) of the centrifuged suspensions was measured at 483 nm. Absorbance (A i ) was obtained by Lambert-Beer law (A i =2-lgT i ). And then photocatalytic rate (D i ) was available according to the following equation: where A 0 is the absorbance of MO solution at 483 nm before exposing under UV light.

Reusable performance and stable performance tests
The reusable performance test of the sample proceeds as follows. Wash the sample which experienced a degradation process for 60 min several times with a certain amount of deionized water. Dry and recomplete another degradation process, repeat the procedure three times. P25 powders are used as the comparison sample under the same conditions. The stable performance test of the sample proceeds as follows. Expose the sample at room temperature for 0−8 weeks. Test photocatalytic performance of the sample respectively after stored 1, 2, 4 and 8 weeks. P25 powders are used as the comparison sample under the same conditions.

Characterization
Powder X-ray diffraction (XRD) patterns were recorded on a Rigaku D/Max-rB12 kW X-ray diffractometer in the diffraction angle range 2θ=5°−60° at a rate of 5°/min using Cu Kα radiation (λ=0.15418 nm). Chemical bonding between the functional groups and carbon atoms was confirmed by X-ray photoelectron spectroscopy (XPS, PHI 5000C ESCA System). The thermal performance of samples was characterized by the thermogravimetric analysis (TGA, STA 449 C Jupiter). The morphology of the as-prepared samples was observed using a field emission scanning electron microscope (FE-SEM, Quanta 200 FEG) and high resolution transmission electron microscopy (HRTEM, JEM2010).

XRD analysis
Figure 1(a) shows the XRD data of TiO 2 -Graphene Oxide nanocomposite, in which graphene oxide was used as an intercalation matrix and it was obtained by exfoliating GO with 0.2 M NaOH. TiO 2 crystallites in the intercalated structure consisting of mixed anatase and rutile phase, in which anatase phase shows more content. (101), (004), (200) crystal surface diffraction peaks that are marked with square shapes are assigned to anatase, and the other four characteristic diffraction peaks of (110), (101), (111), (211) which are marked with round shapes are assigned to rutile phase. The average crystal size of TiO 2 in TiO 2 -Graphene Oxide under different conditions is about 9 nm. Furthermore, a very slight diffraction peak which is marked with a pentagon shape at 2θ=11° belongs to (002) crystal of GO. Compared to (002) diffraction peak of GO in Figure 1(b), the diffraction peak intensity of TiO 2 -Graphene Oxide at 2θ=11° is obviously reduced, and the X ray diffraction peak of GO(100) crystal at 2θ=42° totally disappears. These results explain that the layered GO pretreated with sodium hydroxide solution has been exfoliated to a great extent and formed graphene oxide, [TiO] 2+ groups in the solution adsorbed on the surface of graphene oxide and then the nucleation and growth of TiO 2 crystallites took place in the lamellar structure of graphene oxide, and finally formed TiO 2 -Graphene Oxide nanocomposite. The faint diffraction peak of GO which still existed in the XRD pattern which shows that a small quantity of GO was not exfoliated thoroughly, which may not play a part in the intercalating process of TiO 2 -Graphene Oxide.  Table 1 shows the XPS data of GO and TiO 2 -Graphene Oxide such as binding energies and relative area percentages. Compared to GO, the area percentages of C=O in TiO 2 -Graphene Oxide reduced sharply. To TiO 2 -Graphene Oxide(c), with the decline of C=O, the content of C−C evidently increased. This can be interpreted by the nucleation and growth of TiO 2 in graphene oxide layer, where C=O was consumed abundantly and partly reduced to C−C. Additionally, to TiO 2 -Graphene Oxide(d), the content of  C=O decreased much more evidently while COOH increased slightly. We speculate that as GO was exfoliated by 0.2 M NaOH, GO lamella broke into smaller flakes and the edge area of the formed graphene oxide flakes increased, which caused much more content of C=O and COOH. In addition, OH − introduced from the alkaline solution can further react with graphene oxide. Therefore, the area percentages of C−OH enhanced obviously while the content of C−C decreased. These results show that C=O has the priority to react with [TiO] 2+ during the formation of TiO 2 which makes it decrease sharply, while COOH and C−OH react with [TiO] 2+ more difficultly, but the presence of these groups can improve the dispersion of graphene oxide. Figure 3(a)−(b) describe XPS Ti2p of TiO 2 -Graphene Oxide (GO exfoliated by 0 M NaOH(a) and 0.2 M NaOH (b)). The Ti2p region can be decomposed into several contributions corresponding to the different oxidation states of titanium. Each contribution consists of a doublet between the 2p 3/2 and 2p 1/2 peaks. For each doublet, the ratio of the area of the two peaks A(Ti2p 1/2 )/A(Ti2p 3/2 ) is equal to 0.5 and the binding energy difference, ∆E b =E b (Ti2p 1/2 )-E b (Ti2p 3/2 ) is always 5.7 eV, as previously reported [30,31].

XPS analysis
In Figure 3, it is characterized by a main doublet composed of two symmetric peaks situated at E b (Ti2p 1/2 )= 464.7 eV and E b (Ti2p 3/2 )=459.1 eV, this main doublet is assigned to Ti IV . From the data of Table 2, we can find that titanium in these two kinds of composite are almost four valence state, the molar contents are accounted for 96.37% and 96.99% respectively. In addition, the symmetric peaks of the corresponding doublet situated at E b (Ti2p 1/2 )=463.1 eV and E b (Ti2p 3/2 )=457.4 eV are assigned to Ti 3+ and the molar contents are much lower (3.63% and 3.01% respectively). The binding energy position of Ti 2+ situated at E b (Ti2p 1/2 )= 461.7 eV and E b (Ti2p 3/2 )= 456 eV, but we have not found the XPS peaks of Ti 2+ in both composites.   obvious weight loss stages. The first stage is from room temperature 25°C to 135°C which can be attributed to the dehydration process of water molecules in the interlayer of graphene oxide. The second is from 135°C to 280°C, it is due to the burning decomposition of oxygen-containing groups attached to graphene oxide layers. And the third stage is from 280°C to 740°C which can be assigned to the burning decomposition of carbon skeleton. The mass ratio of TiO 2 particles in the TiO 2 -Graphene Oxide composite is about 57.0% which is consistent with the high temperature heat-treatment results listed in Table 3. According to the TGA curve and heat-treatment results of TiO 2 -Graphene Oxide, we can speculate the mass ratio of H 2 O, graphene oxide and TiO 2 in composite is 1 : 3.5 : 6.4.  5(c), perpendicular to graphene oxide layer). In Figure 5(a), we can find that large amounts of fine TiO 2 nano-particles are located at graphene oxide flakes which own an area of about 200 nm. The average crystal size of TiO 2 is about 9 nm. To the samples treated by 0.2 M NaOH in Figure  5(b) and (c), the size of graphene oxide flakes reduced to about 100 nm. At the same time, large-scale pieces of graphene oxide can also be observed at which TiO 2 particles are located. These broader graphene oxide layers might be due to the connection of the graphene oxide flakes during the reaction between C=O located at the edge of graphene oxide and [TiO] 2+ induced by the hydrolysis of Ti(SO 4 ) 2 . Compared to the stack structure of graphene oxide flakes in Figure 5(a), the generation of broader graphene oxide layers provides more positions for the nucleation and growth of TiO 2 grains.

FE-SEM analysis
According to the above results, the formation mechanism of TiO 2 -Graphene Oxide can be elaborated as follows. When GO was exfoliated by NaOH, large amounts of hydrated sodium ions and water molecules diffused into the GO interlayer, GO lamella was exfoliated and broke into flakes which have better dispersion performances. These formed graphene oxide flakes make [TiO] 2+ induced by the hydrolysis of Ti(SO 4 ) 2 diffuse from the edge of graphene oxide and the substitute the alkali metal ions existed in the interlayer. During the nucleation and growth of TiO 2 grains, a proportion of [TiO] 2+ groups may be consumed due to the reaction with conjugated carboxyl (C=O) located at the edges of GO layers and connect the exfoliated broken graphene oxide flakes together into broader graphene oxide layers. In addition, the XPS results show that COOH and C−O groups cannot easily react with [TiO] 2+ , while these oxygen-containing groups evidently improve the inertia state of graphite surface. And many TiO 2 grains nucleate and grow in COOH and C-OH sites by the way of heterogeneous nucleation. However, most of the TiO 2 grains formed through a heterogeneous nucleation way in which [TiO] 2+ groups defined TiO 2 crystals firstly generated on the surface of graphene oxide as interface. Finally these kinds of TiO 2 -Graphene Oxide composites with layered structure were successfully synthesized. Figure 6 gives the high-resolution TEM images of TiO 2 -Graphene Oxide(GO exfoliated by 0.2 M NaOH). In Figure  6(a), the diffraction spots in TiO 2 -Graphene Oxide composite correspond to different crystal plane of tetragonalstructure TiO 2 crystal cell such as (101), (004) crystal surface assigned to anatase and (301) crystal surface assigned to rutile. However, in Figure 6(b), the diffraction spots correspond to (100) crystal plane of hexagonal-structure GO crystal cell and the scattered diffraction spots below the hexagonal lattice correspond to (101) crystal surface of tetragonal-structure TiO 2 crystal cell. These results indicate the existence of TiO 2 particles beneath the graphene oxide layers though the number of TiO 2 grains is relatively small.

HR-TEM analysis
HRTEM tests confirmed that the TiO 2 grains can nucleate and grow in the interlayer and outer layer of graphene oxide, which agrees with the SEM analysis that TiO 2 particles uniformly distributed in both sides of graphene oxide layers. considerably lower than that of TiO 2 -Graphene Oxide composite. This may be caused by two factors. First, TiO 2 -Graphene Oxide has a better adsorption capacity than P25 powder, which makes TiO 2 particles have higher probabilities to contact with MO. Second, the average crystal size of TiO 2 in TiO 2 -Graphene Oxide under different conditions is about 9 nm while the grain size of Degussa P25 powders comes up to 21 nm. Contrasted with P25, larger  specific surface area of TiO 2 in intercalation composite certainly improves the photodegradation efficiency.

Photocatalytic properties
In addition, the properties of TiO 2 -Graphene Oxide composite (GO exfoliated by 0.2 M NaOH) are superior compared to that of the composite in which GO was exfoliated by 0 M NaOH. These results can be explained as follows, when GO was exfoliated by NaOH, hydrated sodium ions and water molecules diffused into the GO interlayer, GO lamella broke into flakes and the formed graphene oxide flakes owned better dispersion performances which provided more positions for the nucleation and growth of TiO 2 grains. In this case, when the composite is used to degrade methyl orange, more MO molecules can contact with the homogeneously distributed TiO 2 nanoparticles. Under UV light irradiation, more methyl orange molecules will be degraded and the photodegradation efficiency of the sample will be raised effectively. Figure 8 depicts the reusable performance test of TiO 2 -Graphene Oxide (a) and P25 powders (b). In Figure 8(a), TiO 2 -Graphene Oxide composite has been experienced successive degradation process (3 times) to MO and the photodegradation efficiency in 15 min is 1.16 mg min −1 g −1 , 1.02 mg min −1 g −1 and 0.83 mg min −1 g −1 respectively. Compared to Figure 8(b), when P25 powders were used 3 times successively to degrade MO, the degradation rates maintained at about η=0.51 mg min −1 g −1 which indicated that TiO 2 -Graphene Oxide has better photodegradation efficiency. However, with the increase of repetitions, the efficiency of TiO 2 -Graphene Oxide declined gradually. This phenomenon can be attributed to the agglomeration of TiO 2 grains formed on both sides of the graphene oxide layers which made the contact area between MO molecules and TiO 2 particles decrease and resulted in the reduction of photodegradation efficiency. Figure 9 describes the stable performance test of TiO 2 -Graphene Oxide (a) and P25 powders (b). In order to test the impact of powder agglomeration on photocatalytic properties, we exposed these two kinds of photocatalysts at room temperature for 0−8 weeks. In Figure 9(a), when the composite stored at room temperature for 4 weeks, TiO 2 -Graphene Oxide powders still own great stability properties and the agglomeration is not very evident which leads to almost no decline of photocatalytic efficiency (from η=1.16 mg min −1 g −1 to η=1.13 mg min −1 g −1 ). However, when the composite was exposed at room temperature for 8 weeks, the degradation rates of TiO 2 -Graphene Oxide decreased to η=0.91 mg min −1 g −1 which means that the agglomeration exacerbated and the stability became poor in this stage. In Figure  Figure 8 Reusable performance of (a) TiO 2 -Graphene Oxide(GO exfoliated by 0.2 M NaOH) and (b) P25 powders.

Figure 9
Stable performance of (a) TiO 2 -Graphene Oxide(GO exfoliated by 0.2 M NaOH) and (b) P25 powders. 9(b), the agglomeration of P25 powders is also obvious when the powders were exposed at room temperature for 8 weeks.
The efficiency of P25 decreased from η=0.51 mg min −1 g −1 to η=0.40 mg min −1 g −1 . Compared with the two photocatalysts, both the degradation rates of TiO 2 -Graphene Oxide and P25 powders are about 78% of the initial efficiency when exposed at room temperature for 8 weeks, which indicates that although the average grain size of TiO 2 in TiO 2 -Graphene Oxide is only nearly half of the P25, the agglomeration does not multiply. The photocatalytic properties of TiO 2 -Graphene Oxide are always superior to that of P25 powder.

Conclusions
(1) The TiO 2 -Graphene Oxide composite was successfully prepared at low temperature (80°C) with graphite oxide (GO) and titanium sulfate (Ti(SO 4 ) 2 ) as initial reactants. GO was firstly exfoliated by NaOH and formed single and multi-layered graphite oxide mixture which can be defined as graphene oxide, [TiO] 2+ induced by the hydrolysis of Ti(SO 4 ) 2 diffused into graphene oxide interlayer by electrostatic attraction, the nucleation and growth of TiO 2 crystallites took place at low temperature and TiO 2 -Graphene Oxide composite was successfully synthesized.
(2) Under the irradiation of UV light, the photodegradation efficiency of TiO 2 -Graphene Oxide to methyl orange in 15 min is up to 87.2%, η=1.16 mg min −1 g −1 . Compared with Degussa P25 powders, 38.4% in 15 min, η=0.51 mg min −1 g −1 , these intercalation composites own much better efficiency. This is due to two main reasons. The first is the better adsorption capacity of TiO 2 -Graphene Oxide composite, and the second is the smaller average grain size of TiO 2 crystallites (about 9 nm) which may be due to the limit of the unique layered structure of TiO 2 -Graphene Oxide.
(3) The photodegradation efficiency of TiO 2 -Graphene Oxide composite (GO exfoliated by 0.2 M NaOH) is superior compared to that of the sample treated by 0 M NaOH. These results can explain that when GO was exfoliated by NaOH, hydrated sodium ions and water molecules diffused into the GO interlayer, GO lamella broke into flakes and formed graphene oxide flakes owned better dispersion performances which provided more positions for the nucleation and growth of TiO 2 grains.
(4) The reusable performance test of TiO 2 -Graphene Oxide shows that the photocatalytic efficiency of the composite slightly declined but still maintained a high level when the pholocatalysts had been experienced successive degradation process (3 times) to MO. Moreover, the stable performance test confirmed that when the composite was stored at room temperature for 4 weeks, the TiO 2 -Graphene Oxide powders still owned great stability properties and the agglomeration was not very evident which resulted in almost no decline of photocatalytic efficiency. However, when the composite was exposed at room temperature for 8 weeks, both the degradation rates of TiO 2 -Graphene Oxide and P25 powders were about 78% of the initial efficiency which indicated that although the average grain size of TiO 2 is only nearly half of the P25, the agglomeration did not multiply. The photocatalytic properties of TiO 2 -Graphene Oxide are always superior to that of P25 powder.