- Email: [email protected]
Uniform distribution of TiO2 nanocrystals on reduced graphene oxide sheets by the chelating ligands
Journal of Colloid and Interface Science 367 (2012) 139–147
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Uniform distribution of TiO2 nanocrystals on reduced graphene oxide sheets by the chelating ligands Thuy-Duong Nguyen-Phan, Viet Hung Pham, Hyukmin Kweon, Jin Suk Chung, Eui Jung Kim, Seung Hyun Hur, Eun Woo Shin ⇑ School of Chemical Engineering and Bioengineering, University of Ulsan, Daehakro 93, Nam-gu, Ulsan 680-749, South Korea
a r t i c l e
i n f o
Article history: Received 29 June 2011 Accepted 10 October 2011 Available online 23 October 2011 Keywords: TiO2 Reduced graphene oxide Hybrids Chelating agent Adsorption Photocatalysis
a b s t r a c t Reduced graphene oxide–TiO2 hybrids were successfully prepared by the hydrothermal approach using triethanolamine and acetylacetone as the chelating agents. Without any additive, large aggregated TiO2 clusters were randomly distributed dominantly at the edge and less on the basil plane of coagulated reduced graphene oxide (RGO) layers. The presence of chelating ligands remarkably facilitated the selective growth and regular spread of TiO2 nanocrystals onto individually exfoliated RGO sheet. Such sandwich-like structure with stronger coupling and chemical interaction resulted in the surface area increase, the rearrangement of energy level, the enhanced concentration of oxygen vacancies, leading to much higher adsorbability and photocatalytic degradation of Rhodamine B under both UV and visible irradiations. These RGO–TiO2 hybrid systems are potentially beneficial for widely practical applications in air/water purification, electronic devices, batteries, solar cells or supercapacitors. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Graphene (GP), a single atom thick planar sheet of sp2-bonded carbon atoms, has attracted tremendous attention with its unique electrical and structural properties since its discovery in 2004 [1–3]. More recently, the combination of GP or reduced graphene oxide (RGO) with nanostructured metal and metal oxide has become one of the most prominent materials for wide potential applications in electronic devices, drug delivery and energy conversion [4–6]. Numerous metal oxides have been utilized such as TiO2, ZnO, MnO2, SnO2, SiO2, Co3O4, and Fe2O3. Among several metal oxide-based systems, GP-TiO2 and RGO–TiO2 hybrids/composites have been the most well known, being appropriate for solar cells, photocatalysis, batteries, and photovoltaics [4,7–15]. The distribution and coverage of TiO2 onto GP or RGO nanosheets plays an important role in the physical and chemical properties as well as the performance of the hybrids. It induces the chemical interaction and bonding between GP/RGO and TiO2, promoting the coupling effect of the hybrid. However, TiO2 nanoparticles have usually been observed primarily at the edge and more randomly on the basal planes of GP or RGO sheets where abundant oxygen-containing functional groups were observed [2,4,5–9]. This fact usually suppresses the intrinsic characteristics and activity of the hybrid or composite materials. One of the main barriers that
⇑ Corresponding author. Fax: +82 52 259 1689. E-mail address: [email protected] (E.W. Shin). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.10.021
make the synthesis of TiO2-based materials more difficult than other systems is the high reactivity of the titanium alkoxide precursor, which rapidly hydrolyzes, condenses and easily aggregates in aqueous solution [16,17]. The prevention of TiO2 aggregation and the achievement of heavy TiO2 coverage on GP or RGO layers have been described in a few studies in which the syntheses were mostly conducted in nonaqueous media, using inorganic precursors or where the quantity of water was strictly controlled [9,11–15]. By initial functionalization of GO by poly(acrylic acid) and consequently, employing the nonaqueous environment containing both ethylene glycol and acetic acid, Wang et al. successfully decorated well-dispersed TiO2 on GP from which the photovoltaic response and mean life time of electron–hole pairs were significantly improved [9]. Otherwise, Liang et al. reported the direct growth of TiO2 on GO sheets by a two-step method, in which TiO2 was first coated on GO by hydrolysis and subsequently, crystallized into anatase nanocrystals by hydrothermal treatment [11]. They slowed down the hydrolysis reaction by both adding H2SO4 and controlling ethanol/water ratio (15:1, 10:1, 3:1, v/v) in the first step, as well as underwent the crystallization in mixed water/DMF solvent at 200 °C. Zhou et al. also presented a simple solvothermal method to prepared GP-TiO2 composite with TiO2 particles uniformly grown on carbon basal planes [12]. Very small amount of water (1 mL) was added into the mixture of tetrabutyl titanate (TBOT), GO, and isopropanol. Recently, Zhang et al. obtained the whole coverage of RGO sheets by TiO2 nanoparticles using TiCl3 precursor in aqueous HCl solution [15]. Such composite in layered structure greatly enhanced the photocatalytic degradation of dye compounds
140
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Fig. 1. (a) Tapping mode AFM topographic image and height profile of starting GO nanosheets, (b and c) FE-SEM images of RGO–TiO2 hybrid prepared without any chelating agent.
Fig. 2. (a) FE-SEM, (b and c) HR-TEM (inset of b is the SAED pattern), and (d) AFM image of TG-AA hybrid.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
under visible light irradiation due to the effective charge separation and transfer. In the present study, in order to face these challenges, we report a novel approach to synthesize RGO–TiO2 hybrids with well-dispersed nanocrystals from an aqueous solution. Using triethanolamine (TEA) and acetylacetone (AA) as the chelating agent, we successfully controlled the hydrolysis–condensation rate, overcame aggregation, as well as simply induced selective crystal growth and wide distribution of TiO2 nanocrystals on both sides of RGO nanosheets. The hybrids in the sandwich-like structure consists of mono- or bi-layered RGO and several layers of TiO2 exhibited much better adsorptivity and photocatalytic performance than that prepared in only water medium. 2. Materials and methods 2.1. Preparation of hybrid materials Graphene oxide (GO) nanosheets were prepared from expanded graphite by a modified Hummers method [18]. After microwaveheating a desirable amount of expandable graphite (grade 1721, Asbury Carbon), the expanded graphite underwent the oxidation by concentrated H2SO4, KMnO4, and H2O2. The yellow suspension was washed with aqueous HCl solution (10%) and subsequently deionized water until the pH of GO dispersion reached 6. The as-synthesized GO dispersion was in the paste form (12 mg mL1). The detailed procedure was described in the Supplementary material. For the synthesis of reduced graphene oxide–TiO2 hybrids, GO dispersion was re-dispersed in deionized water and pH 10 was gradually adjusted by NaOH. The mixture was ultrasonically treated for 1 h and then stirred at 40 °C for 2 h. Subsequently, a mix-
141
ture of titanium butoxide (TBOT) and the chelating agent with molar ratio of 1:1 was added dropwise. The solution was vigorously stirred at 40 °C for 2 h and following at 80 °C for 10 h. The precipitate was rinsed by warm ethanol, centrifuged and dried at 80 °C. The RGO–TiO2 composite was obtained by annealing at 550 °C for 4 h under N2 flow (15 mL min1). All the materials containing 1 wt.% GO were denoted as TG-TEA, TG-AA, and TG (without the chelate ligand). 2.2. Characterization techniques The topography of the hybrids was observed by atomic force microscope (AFM) using a Veeco Dimension 3100 SPM (USA) with a silicon cantilever operated in the tapping mode in which the specimens were prepared by ultrasonically dispersing the as-prepared suspension in ethanol, drop-drying onto silicon wafer and then annealed at 550 °C for 4 h under N2 flow (15 mL min1). The structures of the materials were analyzed by field-emission scanning electron microscope (FE-SEM, JEOL, JSM-600F, Japan) and high-resolution transmission electron microscope (HR-TEM, JEOL, JEM-2100F, Japan). XRD patterns were analyzed on Rigaku D/MAZX 2500V/PC high power diffractometer (Japan) using Cu Ka radiation (k = 1.5418 Å). The UV–Visible diffuse reflectance spectra (UV–VisDRS) were recorded in the range from 200 to 800 nm on SPECORD 210 Plus spectroscope (Analytikjena, Germany). FT-IR spectra were recorded on Nicolet 380 spectrometer (Thermo Electron Co., USA) using KBr pellet technique, whereas Raman spectra were obtained at room temperature from DXR Raman microscope (ThermoFisher Scientific, USA) with 633 nm laser excitation and CCD detector. The thermogravimetric analysis (TGA) was conducted on a TA Q50 instrument (USA) using Universal Analysis 2000 under N2/air flow
Fig. 3. (a) FE-SEM, (b and c) HR-TEM and (d) AFM micrographs of TG-TEA hybrid.
142
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
of 60/40 mL min1 at a heating rate of 10 °C min1. The textural properties were measured by means of N2 sorption analysis on a Micromeritics ASAP 2010 apparatus (USA). 2.3. Adsorption and photocatalytic activity measurements The adsorptivity and photocatalytic activity of the hybrids was evaluated through the degradation of Rhodamine B (RhB). Twenty milligram of the photocatalyst were immersed into 100 mL of aqueous RhB solution with initial concentration of 2 105 M under constant stirring. After 2 h dark adsorption, four surrounding 20 W black light lamps (kmax = 365 nm) or 20 W daylight lamps (kmax = 545 nm) (Alim Industry Co., Korea) was switched on. The changes in RhB concentration under dark adsorption and UV-A irradiation were measured at kmax = 552 nm by a UV–Vis absorbance microplate spectrophotometer (Spectra MaxÒ Plus 384). 3. Results and discussion As seen in Fig. 1a, an AFM topography image illustrates the starting GO having a flat, smooth sheet-like structure with few wrinkles or crumpled layers. The height profile measurement shows that the average thickness of a single layer is 1.0–1.2 nm. When the TG hybrid was prepared in an aqueous system, the large TiO2 aggregates in sizes of 300–800 nm randomly decorate the surface and the edge of the corrugated RGO sheets in TG composite (Fig. 1b and c). These positions are preferable and have been widely observed in similar composite types prepared in precursor solutions containing higher water content [7,8]. Metal oxide particles have been mainly concentrated on the periphery and less on the basal plane of GO or RGO because of the presence of different oxygen-containing functional groups at these positions [2,4,5,8,9,19–21]. In this study, GO was shown to be highly oxygenated with different functional groups, such as hydroxyl, epoxy, ketone, and carboxylic acids groups (see Fig. S1 – FT-IR spectra in Supplementary material). Interestingly, when using AA as the chelating agent, both sides of RGO sheets were fully covered by TiO2 nanoparticles in TG-AA hybrids as shown in Fig. 2a and b. A small portion of free TiO2 clusters was also found, possibly because of a low percentage of RGO compared with TiO2 (1 wt.%, experimentally measured by TGA analysis – Fig. S2 in Supplementary material). The selected area electron diffraction (SAED) pattern in the inset of Fig. 2b reveals numerous diffraction spots that are ascribed to the lattice fringes of anatase (1 0 1), (0 0 4), (2 0 0), (1 0 5) TiO2, accompanied with the weak ring patterns characteristic of hexagonal symmetry of RGO layer (see the enlarged patterns in Fig. S3 in Supplementary material). The weak intensities for both components can be attributed to the good dispersion of TiO2 nanocrystals onto RGO. HRTEM micrograph in Fig. 2c shows highly crystalline TiO2 nanoparticles having average diameter of 10 nm that are well dispersed onto the RGO surface in the TG-AA hybrid. The lattice spacing of 0.35 nm is representative of the (1 0 1) reflection of the anatase phase, which is consistent with XRD patterns (see Fig. S4 in Supplementary material). The similar sandwich-like microstructure was obtained in the case of using TEA instead of AA as described in Fig. 3. The uniform growth and whole distribution of highly crystalline TiO2 nanocrystals in mean sizes of 10–12 nm onto RGO layers are proven by FE-SEM and HR-TEM images in Fig. 3a–c. Although the fringe lattice of 0.35 nm for (1 0 1) plane of anatase TiO2 was also observed as shown in Fig. 3c, small portion of rutile phase, 16.5%, was recognized in the XRD analysis (see Fig. S4 in Supplementary material). The AFM studies of both TG-AA and TG-TEA hybrids in Figs. 2d and 3d showed good exfoliation of RGO sheets that were successfully anchored by TiO2 nanocrystals. The dimension of the hybrid
sheets on which TiO2 was nucleated, grown and fully distributed is quite large, up to a few micrometers. The surface of each plate is much rougher than that of GO, and its thickness remarkably increased to 25–50 nm (see the height profiles in Fig. S5 in Supplementary material). Further, AFM images show that the TG-AA and TG-TEA hybrids are in a sandwiched geometry that are possibly composed of mono- or bi-layered RGO and several layers of TiO2 nanocrystals on both faces. The sandwich-like structure remained unaffected after sonication for AFM specimen preparation, illustrating the strong coupling and chemically bounded interaction between metal oxide and graphitic layers. It may occur by means of hydrogen bonding or electrostatic interaction between hydroxyl groups of TiO2 and functional groups of GO. The presence of RGO sheets in TG-AA and TG-TEA hybrids was also confirmed by Raman spectra in Fig. 4a. Two characteristic bands of GO, the G band at 1591 cm1 and a D band at 1330 cm1, which are representative of the first-order scattering of E2g vibration mode observed for sp2 domain as well as the edge planes and disordered structures, respectively [22], remained in all the hybrids. The Raman shift toward the higher wavenumber region and the broadening obviously demonstrate the reduction of GO and the charge transfer between TiO2 and RGO sheets. The intensity ratio of D band over G band (ID/IG) of pure GO, 1.12, is much higher than those of the composite and/or hybrids, approximately 0.8–1.0, indicating the increase in average size of sp2 domain as well as the interaction between nanocrystals and graphitic component. Additionally, the five representative peaks
Fig. 4. (a) Raman spectra and (b) UV–Vis-DRS spectra of GO, free TiO2 and RGO– TiO2 hybrids.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
143
Scheme 1. Growth mechanism of TiO2 nanocrystals on RGO nanosheets without and with chelating agents.
of anatase TiO2, including 148 (Eg), 198 (Eg), 394 (B1g), 511 (A1g + B1g) and 632 cm1 (Eg) were also detected. It can be seen that the peak intensities in TG were greater than the others due to the existence of large TiO2 aggregates on RGO. The absence of rutile phase in the TG-TEA compared with the result from XRD patterns can be attributed to either the sensitivity limitation due to its low intensity or the existence of several types of fluorescent organic species, so that this fluorescent background severely limits the application of Raman spectroscopy, hiding the features of rutile phase in the spectrum [23]. The changes in optical properties of the hybrids prepared by different conditions were evidenced by the UV–Vis-DRS spectra in Fig. 4b. The onset absorption edge of TG composite is found around 390–400 nm, quite similar to that of pure TiO2 (see Fig. S6 in Supplementary material). However, with the assistance of AA and TEA,
not only the significant red shift toward higher photoactive wavelength region of the edges but also the strong, broad shoulder absorption throughout the visible light range is observed. It can be concluded that the presence of such chelating agent in the preparation method remarkably rearranges the energy levels of TiO2. The band gap narrowing can be assigned to the chemically bounding interaction between TiO2 and RGO in the form of Ti–O–C bond in TG-AA and TG-TEA hybrids. The appearance of the visible absorption band obviously implicates the role as a photosensitizer of RGO that induces the possible working ability under visible light irradiation of the sandwich-like microstructured materials. In addition, it can be explained by the generation of oxygen vacancies in the vacuum environment due to the interactions between free carbon and oxygen in TiO2 [14,24]. The residual chelating agents that might be trapped in the lattice of TiO2 or reduced TiOx were
144
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Fig. 5. (a) RhB uptake and (b) pseudo-second-order kinetics model for RhB adsorption over RGO–TiO2 materials.
Table 1 Textural and kinetic parameters of RGO–TiO2 hybrids without and with chelating agents. Sample SBET Vt qe kads (gcat/mg min) kapp (min1) (103)c (m2/g)a (cm3/g)a (mg/gcat)b (105)b UV Vis TG 23.52 TG-AA 85.51 TG-TEA 88.97
0.03101 18.51 0.1215 35.55 0.1265 38.48
17.32 5.82 10.01
2.41 10.96 14.66
0.57 2.83 3.04
a BET surface area and total pore volume, respectively, determined from N2 adsorption–desorption measurement. b Equilibrium adsorption capacity and pseudo-second-order adsorption rate constant, determined from the slope and intercept of the plot of (t/qt) = f(t). c Apparent first-order degradation rate constant, evaluated from the slope of plot of ln(C/Co) = kappt.
removed during the calcination along with the thermal reduction of GO, possibly leading to the higher concentration of oxygen vacancies in TG-AA and TG-TEA. These vacancies may induce some localized donor states within the band gap ranging from 0.75 to 1.18 eV below the conduction band minimum, resulting in the photoresponse to visible region [15,25,26]. The influence of chelating agent on the distribution of TiO2 on RGO sheets is described in Scheme 1. Using water as the only solvent, the rapid hydrolysis–condensation occurs through the easy cleavage of butoxy groups by water molecules to form hydroxotitanium butoxides. Subsequently, the alcoholation and dehydration take place, resulting in titanium oxo-butoxide, hydroxides
Fig. 6. (a) UV-photodegradation of RhB on TG, TG-AA and TG-TEA hybrids and (b) plots of apparent first-order linear transform ln(C/Co) = f(t).
and sometimes hydroxyl–butoxides [16]. It leads to the formation of precipitation and aggregation of TiO2 particles on GO sheet (pathway A). The mechanism is consistent with the microscope observations of the TG sample aforementioned. The growth of the hybrids in the presence of AA and TEA is differently illustrated in pathways B and C, respectively. AA, belonging to the b-diketones ligand group, has been widely used in the alkoxide precursor system [16]. Herein, AA has multiple functions; it not only increases the coordination number of Ti atoms but also easily substitutes the hydrolyzable butoxy groups by strong chelating AA ligands allowing one butanol group to unbind [16,27–29]. Therefore, the less reactive Ti-AA groups act as a poison toward the hydrolysis– condensation (pathway B), inhibiting the spatial extension of the solid phase and preventing the aggregation of TiO2 [28]. This also slows the nucleation, direct growth and the interaction of TiO2 with GO layers, resulting in the whole coverage of small nanocrystals on each GO sheet (sample TG-AA). On the other hand, TEA, belonging to the alkanoamine group, is also a potential chelating ligand [16,27,30]. TEA acts as a neutral nitrogen-donor, forming amine-trialkoxo complexes [30] or replacing the –OBu groups by alkoholate groups [31] to form bi- or tridentate coordination, as shown in pathway C. Therefore, similar to the case of AA, TEA efficiently stabilizes the Ti precursor, retards the hydrolysis rate and controls the inorganic polymerization of TBOT. This inhibits the growth and agglomeration of free TiO2 in the solution and induces the full distribution of nanoparticles onto both sides of the GO layers (sample TG-TEA).
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
145
The assistance of chelating ligands plays very important role in the slow hydrolysis–condensation of titanium precursors, facilitating crystal growth, nucleation and distribution of TiO2 nanocrystals on the two-dimensional graphitic nanosheets. The advantage of such a sandwich-like structure for the hybrids with stronger chemically bounded interaction can be demonstrated through removal efficiency of dye wastewater. Herein, the dark adsorption behavior of RhB solution was investigated, followed by the photocatalytic degradation of under both UV-A and visible irradiations. Firstly, for the evaluation of the adsorptivity, the adsorbed quantity of RhB, qt (mg g1 cat ), at the interval time t was calculated by the following equation [32,33]:
qt ¼
ðC 0 C t Þ 1000 M W V 0 wcat
ð1Þ
where C0 and Ct (M) are the initial concentration and concentration at t of RhB, respectively; MW = 479.03 (g mol1) is the molecular weight of RhB; V0 (L) is the volume of aqueous dye solution; and wcat (g) is the mass of catalyst. Fig. 5a illustrates the adsorption behaviors of RhB during 120 min dark adsorption. The hybrids prepared by using the chelating ligands exhibited good adsorbability of the dye, retaining 25% RhB content in the aqueous solution. The dye uptakes over these materials are much higher than that of TG composite which immediately attained the plateau after 15 min and removed only 40% of organic molecules in total. It can be concluded that the adsorption amount significantly increased in the hybrids prepared under the assistance of AA and TEA. Based on the adsorption kinetic data, the equilibrium adsorption capacity, qe, and adsorption rate constant, kads, were estimated by applying the pseudo-first-order and second-order models [32–34]. The appropriate values are chosen according to the higher correlation coefficient (R2). In the present study, the adsorption kinetics were well-fitted with the pseudo-second-order model as below [32,34]:
dqt ¼ kads ðqe qt Þ2 dt
ð2Þ
where qe and qt are the adsorption capacity at equilibrium and at time t, respectively; and kads (gcat mg1 min1) is the rate constant of pseudo-second-order sorption. By integrating Eq. (2) for the boundary conditions (t = 0 to t = t and qt = 0 to qt = qt), the following linear form can be written as follows:
t 1 1 ¼ þ t qt kq2e qe
ð3Þ
Herein, the experimental data evaluated from the linearity transform (t/qt) versus t were entirely consistent with the pseudo-second-order kinetics as expressed in Fig. 5b. The qe and kads values were evaluated from the slope and intercept of the plot, respectively, and were summarized in Table 1. TG-AA and TGTEA display the dye adsorption capacities of 35.55 and 38:48 mg g1 cat , respectively, whereas qe value is quite low for TG composite, approximately 18:51 mg g1 cat . Correspondingly, the composite containing large TiO2 aggregates on the surface or at the edge of corrugated RGO layers shows the adsorption rate constant kads of 17.3 gcat mg1 min1, much faster than those of the materials consist of wholly covered TiO2 nanoparticles. In consequence, the significant improvement in the adsorptivity of the sandwiched hybrids can be attributed to the giant two-dimensional sheet structure of graphitic component, the p–p conjugation with face-to-face orientation between RhB and aromatic regions of RGO [4] and the remarkable increase in BET specific surface areas, 86 and 89 m2/g, respectively, compared with that using only water as solvent (see Table 1).
Fig. 7. (a) Photocatalytic degradation of RhB under visible illumination and (b) corresponding linear transform ln(C/C0) = f(t).
Subsequently, the photocatalytic performance was evaluated by the exposure of the materials at the equilibrium state under UV-A irradiation. As shown in Fig. 6a, TG-AA and TG-TEA show much higher degradation efficiencies, about 80–90%, than those of pure TiO2 and TG composite upon UV irradiation after 120 min (see Fig. S9 in Supplementary material). Their photocatalytic capacities 1 follow an order: TG-TEA (46:27 mg g1 cat Þ > TG-AAð44:52 mg gcat Þ > TGð27:06 mg g1 ), which is identical to the adsorptivities. It is cat markedly due to the significant increase in the accessible surface area of RhB over the prepared materials. Accordingly, the apparent reaction rate constants were quantitatively derived for the evaluation of photocatalytic activity. The reaction kinetics obeyed a Langmuir–Hinshelwood pseudo-first-order model (when C0 is very small) [33,35,36]:
ln
C C0
¼ kapp t
ð4Þ
where kapp (gcat mg1 min1) is the apparent first-order rate constant that is determined from the slope of the linearity plot as seen in Fig. 6a. kapp values of TG-AA and TG-TEA photocatalysts are much higher than that for TG composite composed of aggregated TiO2 clusters on RGO. TG-TEA hybrid shows the fastest rate, typically 14.66 103 min1. The enhanced UV-photocatalytic performance of such sandwich-like hybrids compared with the others can be also ascribed to the strong coupling effect with chemical interactions between well-adhered TiO2 nanocrystals and individual RGO sheets, possibly accelerating the interfacial electron-transfer. Furthermore,
146
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Scheme 2. Photodegradation of Rhodamine B over sandwich-like reduced graphene oxide–TiO2 hybrid under UV and visible irradiation.
RGO may act as an electron acceptor of photogenerated electrons from the conduction band of TiO2 nanocrystals and subsequently as a transporter to its two-dimensional p–p conjugational plane due to the higher work function of RGO (illustrated in Scheme 2) [15]. The well-dispersed TiO2 onto mono- or bi-layered RGO sheets which are in the form of giant conducting network are much beneficial for the long-lived electron–hole pairs. Additionally, as discussed above, the presence of oxygen defects might be found and they act as electron traps. These phenomena efficiently hinder the recombination of charge carriers, accelerate the interface charge transfer and therefore improve the UV-photodegradable efficiency of RhB. On the other hand, due to the significant changes in the optical properties, the visible-illuminated photocatalytic activity was also observed after reaching the adsorption saturation. As seen in Fig. 7a, TG composite displays a negligible reactivity, whereas in contrast, TG-AA and TG-TEA hybrids reveal quite good performance under the visible light with the decomposition of 30–35% adsorbed RhB. By adopting the similar kinetics mechanisms in Fig. 7b, the apparent photodegradation rate constants for these sandwiched-like materials are threefold greater than that of the others as given in Table 1. It is noteworthy that such performance promotion of the sandwich-like hybrids is entirely consistent with the UV–Vis-DRS results mentioned above. The removal efficiency of RhB in the extended light absorption range can be ascribed to the chemical interaction between TiO2 and graphitic component as well as the role of individual RGO sheet as a photosensitizer to enhance the utilization of visible light. More importantly, another reason is the assistance of AA and TEA as the chelating agents, which enrich the concentration of oxygen vacancies in the hybrids as mentioned previously. The presence of vacancy-induced band of electronic states facilitates the transport of the photocarriers to the active sites on the surface as well as reduces the energy needed for photoexcitation [25]. These facts were elucidated to be responsible for the large photodegradation activity in the visible range. The results are expected to explore widely practical applications in the air/water purification and energy conversion. 4. Conclusions We have successfully synthesized RGO–TiO2 hybrid materials by one-step hydrothermal method in which the growth, nucleation
and full coverage of TiO2 nanocrystals on both sides of large RGO sheets were investigated for the first time by controlling the chelating agents. Without chelation, aggregated TiO2 clusters were formed and randomly distributed onto RGO. In the presence of chelating ligands, such a sandwich-like structure with strong chemical interaction between metal oxides and graphitic layers exhibited much better removal efficiency of dye molecules under both UV and visible light illuminations. The RGO–TiO2 system may serve a variety of applications such as photocatalysis, electronic devices, batteries, solar cells or supercapacitors. Acknowledgment This work was supported by the development program of local science park funded by the Ulsan Metropolitan City and The MEST (Ministry of Education, Science and Technology). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcis.2011.10.021. References [1] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666. [2] A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183. [3] D. Li, M.B. Muller, S. Gilje, R.B. Kaner, G.G. Wallace, Nat. Nanotechnol. 3 (2008) 101. [4] H. Zhang, X. Lv, Y. Li, Y. Wang, J. Li, ACS Nano 4 (2010) 380. [5] S. Chen, J. Zhu, X. Wu, Q. Han, X. Wang, ACS Nano 4 (2010) 2822. [6] I.V. Lightcap, T.H. Kosel, P.V. Kamat, Nano Lett. 10 (2010) 577. [7] O. Akhavan, E. Ghaderi, J. Phys. Chem. C 113 (2009) 20214. [8] Y.H. Ng, I.V. Lightcap, K. Goodwin, M. Matsumura, P.V. Kamat, J. Phys. Chem. Lett. 1 (2010) 2222. [9] P. Wang, Y. Zhai, D. Wang, S. Dong, Nanoscale 3 (2011) 1640. [10] N. Yang, J. Zhai, D. Wang, Y. Chen, L. Jiang, ACS Nano 4 (2010) 887. [11] Y. Liang, H. Wang, H.S. Casalongue, Z. Chen, H. Dai, Nano Res. 3 (2010) 701. [12] K. Zhou, Y. Zhu, X. Yang, X. Jiang, C. Li, New J. Chem. 35 (2011) 353. [13] J. Shen, B. Yan, M. Shi, H. Ma, N. Li, M. Ye, J. Mater. Chem. 21 (2011) 3415. [14] X.-Y. Zhang, H.-P. Li, X.-L. Cui, Y. Lin, J. Mater. Chem. 20 (2010) 2801. [15] J. Zhang, Z. Xiong, X.S. Zhao, J. Mater. Chem. 21 (2011) 3634. [16] D.C. Bradley, R.C. Mehrotra, D.P. Gaur, Metal Alkoxides, Academic Press, 1978. [17] X. Chen, S.S. Mao, Chem. Rev. 107 (2007) 2891. [18] W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147 [19] T. Szab, O. Berkesi, P. Forg, K. Josepovits, Y. Sanakis, D. Petridis, I. Dkny, Chem. Mater. 18 (2006) 2740. [20] W. Gao, L.B. Alemany, L. Ci, P.M. Ajayan, Nat. Chem. 1 (2009) 403. [21] D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, Chem. Soc. Rev. 39 (2010) 228. [22] F. Tuinstra, J.L. Koenig, J. Chem. Phys. 53 (1970) 1126. [23] Y. Djaoued, S. Badilescu, P.V. Ashrit, D. Bersani, P.P. Lottici, J. Robichaud, J. SolGel Sci. Technol. 24 (2002) 255. [24] P.W. Chou, Y.S. Wang, C.C. Lin, Y.J. Chen, C.L. Cheng, M.S. Wong, Surf. Coat. Technol. 204 (2009) 834. [25] I. Justicia, P. Ordejón, G. Canto, J.L. Mozos, J. Fraxedas, G.A. Battiston, R. Gerbasi, A. Figueras, Adv. Mater. 14 (2002) 1399. [26] D.-W. Wang, D.-M. Tang, C. Liu, X. Ma, G.Q. Lu, H.-M. Cheng, Nanotechnology 19 (2008) 25606.
147
[27] D.C. Bradley, R.C. Mehrotra, I.P. Rothwell, A. Singh, Alkoxo and Aryloxo Derivatives of Metals, Academic Press, 2001. [28] F. Ribot, P. Toledano, C. Sanchez, Chem. Mater. 3 (1991) 759. [29] A. Ponton, S. Barboux-Doeuff, C. Sanchez, Colloids Surf. A 162 (1999) 177. [30] F. Wang, J. Sun, C. Ma, Stud. Surf. Sci. Catal. 165 (2007) 305. [31] M. Bockmeyer, P. Löbmann, J. Sol-Gel Sci. Technol. 45 (2008) 251. [32] R.A. Shawabkeh, M.F. Tutunji, Appl. Clay Sci. 24 (2003) 111. [33] W. Kangwansupamonkon, W. Jitbunpot, S. Kiatkamjornwong, Polym. Degrad. Stab. 95 (2010) 1894. [34] Y.S. Ho, G. McKay, Trans. Inst. Chem. Eng. 76B (1998) 332. [35] X. Li, J. Ye, J. Phys. Chem. C 111 (2007) 13109. [36] K. Yu, S. Yang, H. He, C. Sun, C. Gu, Y. Ju, J. Phys. Chem. A 113 (2009) 10024.
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Uniform distribution of TiO2 nanocrystals on reduced graphene oxide sheets by the chelating ligands Thuy-Duong Nguyen-Phan, Viet Hung Pham, Hyukmin Kweon, Jin Suk Chung, Eui Jung Kim, Seung Hyun Hur, Eun Woo Shin ⇑ School of Chemical Engineering and Bioengineering, University of Ulsan, Daehakro 93, Nam-gu, Ulsan 680-749, South Korea
a r t i c l e
i n f o
Article history: Received 29 June 2011 Accepted 10 October 2011 Available online 23 October 2011 Keywords: TiO2 Reduced graphene oxide Hybrids Chelating agent Adsorption Photocatalysis
a b s t r a c t Reduced graphene oxide–TiO2 hybrids were successfully prepared by the hydrothermal approach using triethanolamine and acetylacetone as the chelating agents. Without any additive, large aggregated TiO2 clusters were randomly distributed dominantly at the edge and less on the basil plane of coagulated reduced graphene oxide (RGO) layers. The presence of chelating ligands remarkably facilitated the selective growth and regular spread of TiO2 nanocrystals onto individually exfoliated RGO sheet. Such sandwich-like structure with stronger coupling and chemical interaction resulted in the surface area increase, the rearrangement of energy level, the enhanced concentration of oxygen vacancies, leading to much higher adsorbability and photocatalytic degradation of Rhodamine B under both UV and visible irradiations. These RGO–TiO2 hybrid systems are potentially beneficial for widely practical applications in air/water purification, electronic devices, batteries, solar cells or supercapacitors. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Graphene (GP), a single atom thick planar sheet of sp2-bonded carbon atoms, has attracted tremendous attention with its unique electrical and structural properties since its discovery in 2004 [1–3]. More recently, the combination of GP or reduced graphene oxide (RGO) with nanostructured metal and metal oxide has become one of the most prominent materials for wide potential applications in electronic devices, drug delivery and energy conversion [4–6]. Numerous metal oxides have been utilized such as TiO2, ZnO, MnO2, SnO2, SiO2, Co3O4, and Fe2O3. Among several metal oxide-based systems, GP-TiO2 and RGO–TiO2 hybrids/composites have been the most well known, being appropriate for solar cells, photocatalysis, batteries, and photovoltaics [4,7–15]. The distribution and coverage of TiO2 onto GP or RGO nanosheets plays an important role in the physical and chemical properties as well as the performance of the hybrids. It induces the chemical interaction and bonding between GP/RGO and TiO2, promoting the coupling effect of the hybrid. However, TiO2 nanoparticles have usually been observed primarily at the edge and more randomly on the basal planes of GP or RGO sheets where abundant oxygen-containing functional groups were observed [2,4,5–9]. This fact usually suppresses the intrinsic characteristics and activity of the hybrid or composite materials. One of the main barriers that
⇑ Corresponding author. Fax: +82 52 259 1689. E-mail address: [email protected] (E.W. Shin). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.10.021
make the synthesis of TiO2-based materials more difficult than other systems is the high reactivity of the titanium alkoxide precursor, which rapidly hydrolyzes, condenses and easily aggregates in aqueous solution [16,17]. The prevention of TiO2 aggregation and the achievement of heavy TiO2 coverage on GP or RGO layers have been described in a few studies in which the syntheses were mostly conducted in nonaqueous media, using inorganic precursors or where the quantity of water was strictly controlled [9,11–15]. By initial functionalization of GO by poly(acrylic acid) and consequently, employing the nonaqueous environment containing both ethylene glycol and acetic acid, Wang et al. successfully decorated well-dispersed TiO2 on GP from which the photovoltaic response and mean life time of electron–hole pairs were significantly improved [9]. Otherwise, Liang et al. reported the direct growth of TiO2 on GO sheets by a two-step method, in which TiO2 was first coated on GO by hydrolysis and subsequently, crystallized into anatase nanocrystals by hydrothermal treatment [11]. They slowed down the hydrolysis reaction by both adding H2SO4 and controlling ethanol/water ratio (15:1, 10:1, 3:1, v/v) in the first step, as well as underwent the crystallization in mixed water/DMF solvent at 200 °C. Zhou et al. also presented a simple solvothermal method to prepared GP-TiO2 composite with TiO2 particles uniformly grown on carbon basal planes [12]. Very small amount of water (1 mL) was added into the mixture of tetrabutyl titanate (TBOT), GO, and isopropanol. Recently, Zhang et al. obtained the whole coverage of RGO sheets by TiO2 nanoparticles using TiCl3 precursor in aqueous HCl solution [15]. Such composite in layered structure greatly enhanced the photocatalytic degradation of dye compounds
140
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Fig. 1. (a) Tapping mode AFM topographic image and height profile of starting GO nanosheets, (b and c) FE-SEM images of RGO–TiO2 hybrid prepared without any chelating agent.
Fig. 2. (a) FE-SEM, (b and c) HR-TEM (inset of b is the SAED pattern), and (d) AFM image of TG-AA hybrid.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
under visible light irradiation due to the effective charge separation and transfer. In the present study, in order to face these challenges, we report a novel approach to synthesize RGO–TiO2 hybrids with well-dispersed nanocrystals from an aqueous solution. Using triethanolamine (TEA) and acetylacetone (AA) as the chelating agent, we successfully controlled the hydrolysis–condensation rate, overcame aggregation, as well as simply induced selective crystal growth and wide distribution of TiO2 nanocrystals on both sides of RGO nanosheets. The hybrids in the sandwich-like structure consists of mono- or bi-layered RGO and several layers of TiO2 exhibited much better adsorptivity and photocatalytic performance than that prepared in only water medium. 2. Materials and methods 2.1. Preparation of hybrid materials Graphene oxide (GO) nanosheets were prepared from expanded graphite by a modified Hummers method [18]. After microwaveheating a desirable amount of expandable graphite (grade 1721, Asbury Carbon), the expanded graphite underwent the oxidation by concentrated H2SO4, KMnO4, and H2O2. The yellow suspension was washed with aqueous HCl solution (10%) and subsequently deionized water until the pH of GO dispersion reached 6. The as-synthesized GO dispersion was in the paste form (12 mg mL1). The detailed procedure was described in the Supplementary material. For the synthesis of reduced graphene oxide–TiO2 hybrids, GO dispersion was re-dispersed in deionized water and pH 10 was gradually adjusted by NaOH. The mixture was ultrasonically treated for 1 h and then stirred at 40 °C for 2 h. Subsequently, a mix-
141
ture of titanium butoxide (TBOT) and the chelating agent with molar ratio of 1:1 was added dropwise. The solution was vigorously stirred at 40 °C for 2 h and following at 80 °C for 10 h. The precipitate was rinsed by warm ethanol, centrifuged and dried at 80 °C. The RGO–TiO2 composite was obtained by annealing at 550 °C for 4 h under N2 flow (15 mL min1). All the materials containing 1 wt.% GO were denoted as TG-TEA, TG-AA, and TG (without the chelate ligand). 2.2. Characterization techniques The topography of the hybrids was observed by atomic force microscope (AFM) using a Veeco Dimension 3100 SPM (USA) with a silicon cantilever operated in the tapping mode in which the specimens were prepared by ultrasonically dispersing the as-prepared suspension in ethanol, drop-drying onto silicon wafer and then annealed at 550 °C for 4 h under N2 flow (15 mL min1). The structures of the materials were analyzed by field-emission scanning electron microscope (FE-SEM, JEOL, JSM-600F, Japan) and high-resolution transmission electron microscope (HR-TEM, JEOL, JEM-2100F, Japan). XRD patterns were analyzed on Rigaku D/MAZX 2500V/PC high power diffractometer (Japan) using Cu Ka radiation (k = 1.5418 Å). The UV–Visible diffuse reflectance spectra (UV–VisDRS) were recorded in the range from 200 to 800 nm on SPECORD 210 Plus spectroscope (Analytikjena, Germany). FT-IR spectra were recorded on Nicolet 380 spectrometer (Thermo Electron Co., USA) using KBr pellet technique, whereas Raman spectra were obtained at room temperature from DXR Raman microscope (ThermoFisher Scientific, USA) with 633 nm laser excitation and CCD detector. The thermogravimetric analysis (TGA) was conducted on a TA Q50 instrument (USA) using Universal Analysis 2000 under N2/air flow
Fig. 3. (a) FE-SEM, (b and c) HR-TEM and (d) AFM micrographs of TG-TEA hybrid.
142
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
of 60/40 mL min1 at a heating rate of 10 °C min1. The textural properties were measured by means of N2 sorption analysis on a Micromeritics ASAP 2010 apparatus (USA). 2.3. Adsorption and photocatalytic activity measurements The adsorptivity and photocatalytic activity of the hybrids was evaluated through the degradation of Rhodamine B (RhB). Twenty milligram of the photocatalyst were immersed into 100 mL of aqueous RhB solution with initial concentration of 2 105 M under constant stirring. After 2 h dark adsorption, four surrounding 20 W black light lamps (kmax = 365 nm) or 20 W daylight lamps (kmax = 545 nm) (Alim Industry Co., Korea) was switched on. The changes in RhB concentration under dark adsorption and UV-A irradiation were measured at kmax = 552 nm by a UV–Vis absorbance microplate spectrophotometer (Spectra MaxÒ Plus 384). 3. Results and discussion As seen in Fig. 1a, an AFM topography image illustrates the starting GO having a flat, smooth sheet-like structure with few wrinkles or crumpled layers. The height profile measurement shows that the average thickness of a single layer is 1.0–1.2 nm. When the TG hybrid was prepared in an aqueous system, the large TiO2 aggregates in sizes of 300–800 nm randomly decorate the surface and the edge of the corrugated RGO sheets in TG composite (Fig. 1b and c). These positions are preferable and have been widely observed in similar composite types prepared in precursor solutions containing higher water content [7,8]. Metal oxide particles have been mainly concentrated on the periphery and less on the basal plane of GO or RGO because of the presence of different oxygen-containing functional groups at these positions [2,4,5,8,9,19–21]. In this study, GO was shown to be highly oxygenated with different functional groups, such as hydroxyl, epoxy, ketone, and carboxylic acids groups (see Fig. S1 – FT-IR spectra in Supplementary material). Interestingly, when using AA as the chelating agent, both sides of RGO sheets were fully covered by TiO2 nanoparticles in TG-AA hybrids as shown in Fig. 2a and b. A small portion of free TiO2 clusters was also found, possibly because of a low percentage of RGO compared with TiO2 (1 wt.%, experimentally measured by TGA analysis – Fig. S2 in Supplementary material). The selected area electron diffraction (SAED) pattern in the inset of Fig. 2b reveals numerous diffraction spots that are ascribed to the lattice fringes of anatase (1 0 1), (0 0 4), (2 0 0), (1 0 5) TiO2, accompanied with the weak ring patterns characteristic of hexagonal symmetry of RGO layer (see the enlarged patterns in Fig. S3 in Supplementary material). The weak intensities for both components can be attributed to the good dispersion of TiO2 nanocrystals onto RGO. HRTEM micrograph in Fig. 2c shows highly crystalline TiO2 nanoparticles having average diameter of 10 nm that are well dispersed onto the RGO surface in the TG-AA hybrid. The lattice spacing of 0.35 nm is representative of the (1 0 1) reflection of the anatase phase, which is consistent with XRD patterns (see Fig. S4 in Supplementary material). The similar sandwich-like microstructure was obtained in the case of using TEA instead of AA as described in Fig. 3. The uniform growth and whole distribution of highly crystalline TiO2 nanocrystals in mean sizes of 10–12 nm onto RGO layers are proven by FE-SEM and HR-TEM images in Fig. 3a–c. Although the fringe lattice of 0.35 nm for (1 0 1) plane of anatase TiO2 was also observed as shown in Fig. 3c, small portion of rutile phase, 16.5%, was recognized in the XRD analysis (see Fig. S4 in Supplementary material). The AFM studies of both TG-AA and TG-TEA hybrids in Figs. 2d and 3d showed good exfoliation of RGO sheets that were successfully anchored by TiO2 nanocrystals. The dimension of the hybrid
sheets on which TiO2 was nucleated, grown and fully distributed is quite large, up to a few micrometers. The surface of each plate is much rougher than that of GO, and its thickness remarkably increased to 25–50 nm (see the height profiles in Fig. S5 in Supplementary material). Further, AFM images show that the TG-AA and TG-TEA hybrids are in a sandwiched geometry that are possibly composed of mono- or bi-layered RGO and several layers of TiO2 nanocrystals on both faces. The sandwich-like structure remained unaffected after sonication for AFM specimen preparation, illustrating the strong coupling and chemically bounded interaction between metal oxide and graphitic layers. It may occur by means of hydrogen bonding or electrostatic interaction between hydroxyl groups of TiO2 and functional groups of GO. The presence of RGO sheets in TG-AA and TG-TEA hybrids was also confirmed by Raman spectra in Fig. 4a. Two characteristic bands of GO, the G band at 1591 cm1 and a D band at 1330 cm1, which are representative of the first-order scattering of E2g vibration mode observed for sp2 domain as well as the edge planes and disordered structures, respectively [22], remained in all the hybrids. The Raman shift toward the higher wavenumber region and the broadening obviously demonstrate the reduction of GO and the charge transfer between TiO2 and RGO sheets. The intensity ratio of D band over G band (ID/IG) of pure GO, 1.12, is much higher than those of the composite and/or hybrids, approximately 0.8–1.0, indicating the increase in average size of sp2 domain as well as the interaction between nanocrystals and graphitic component. Additionally, the five representative peaks
Fig. 4. (a) Raman spectra and (b) UV–Vis-DRS spectra of GO, free TiO2 and RGO– TiO2 hybrids.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
143
Scheme 1. Growth mechanism of TiO2 nanocrystals on RGO nanosheets without and with chelating agents.
of anatase TiO2, including 148 (Eg), 198 (Eg), 394 (B1g), 511 (A1g + B1g) and 632 cm1 (Eg) were also detected. It can be seen that the peak intensities in TG were greater than the others due to the existence of large TiO2 aggregates on RGO. The absence of rutile phase in the TG-TEA compared with the result from XRD patterns can be attributed to either the sensitivity limitation due to its low intensity or the existence of several types of fluorescent organic species, so that this fluorescent background severely limits the application of Raman spectroscopy, hiding the features of rutile phase in the spectrum [23]. The changes in optical properties of the hybrids prepared by different conditions were evidenced by the UV–Vis-DRS spectra in Fig. 4b. The onset absorption edge of TG composite is found around 390–400 nm, quite similar to that of pure TiO2 (see Fig. S6 in Supplementary material). However, with the assistance of AA and TEA,
not only the significant red shift toward higher photoactive wavelength region of the edges but also the strong, broad shoulder absorption throughout the visible light range is observed. It can be concluded that the presence of such chelating agent in the preparation method remarkably rearranges the energy levels of TiO2. The band gap narrowing can be assigned to the chemically bounding interaction between TiO2 and RGO in the form of Ti–O–C bond in TG-AA and TG-TEA hybrids. The appearance of the visible absorption band obviously implicates the role as a photosensitizer of RGO that induces the possible working ability under visible light irradiation of the sandwich-like microstructured materials. In addition, it can be explained by the generation of oxygen vacancies in the vacuum environment due to the interactions between free carbon and oxygen in TiO2 [14,24]. The residual chelating agents that might be trapped in the lattice of TiO2 or reduced TiOx were
144
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Fig. 5. (a) RhB uptake and (b) pseudo-second-order kinetics model for RhB adsorption over RGO–TiO2 materials.
Table 1 Textural and kinetic parameters of RGO–TiO2 hybrids without and with chelating agents. Sample SBET Vt qe kads (gcat/mg min) kapp (min1) (103)c (m2/g)a (cm3/g)a (mg/gcat)b (105)b UV Vis TG 23.52 TG-AA 85.51 TG-TEA 88.97
0.03101 18.51 0.1215 35.55 0.1265 38.48
17.32 5.82 10.01
2.41 10.96 14.66
0.57 2.83 3.04
a BET surface area and total pore volume, respectively, determined from N2 adsorption–desorption measurement. b Equilibrium adsorption capacity and pseudo-second-order adsorption rate constant, determined from the slope and intercept of the plot of (t/qt) = f(t). c Apparent first-order degradation rate constant, evaluated from the slope of plot of ln(C/Co) = kappt.
removed during the calcination along with the thermal reduction of GO, possibly leading to the higher concentration of oxygen vacancies in TG-AA and TG-TEA. These vacancies may induce some localized donor states within the band gap ranging from 0.75 to 1.18 eV below the conduction band minimum, resulting in the photoresponse to visible region [15,25,26]. The influence of chelating agent on the distribution of TiO2 on RGO sheets is described in Scheme 1. Using water as the only solvent, the rapid hydrolysis–condensation occurs through the easy cleavage of butoxy groups by water molecules to form hydroxotitanium butoxides. Subsequently, the alcoholation and dehydration take place, resulting in titanium oxo-butoxide, hydroxides
Fig. 6. (a) UV-photodegradation of RhB on TG, TG-AA and TG-TEA hybrids and (b) plots of apparent first-order linear transform ln(C/Co) = f(t).
and sometimes hydroxyl–butoxides [16]. It leads to the formation of precipitation and aggregation of TiO2 particles on GO sheet (pathway A). The mechanism is consistent with the microscope observations of the TG sample aforementioned. The growth of the hybrids in the presence of AA and TEA is differently illustrated in pathways B and C, respectively. AA, belonging to the b-diketones ligand group, has been widely used in the alkoxide precursor system [16]. Herein, AA has multiple functions; it not only increases the coordination number of Ti atoms but also easily substitutes the hydrolyzable butoxy groups by strong chelating AA ligands allowing one butanol group to unbind [16,27–29]. Therefore, the less reactive Ti-AA groups act as a poison toward the hydrolysis– condensation (pathway B), inhibiting the spatial extension of the solid phase and preventing the aggregation of TiO2 [28]. This also slows the nucleation, direct growth and the interaction of TiO2 with GO layers, resulting in the whole coverage of small nanocrystals on each GO sheet (sample TG-AA). On the other hand, TEA, belonging to the alkanoamine group, is also a potential chelating ligand [16,27,30]. TEA acts as a neutral nitrogen-donor, forming amine-trialkoxo complexes [30] or replacing the –OBu groups by alkoholate groups [31] to form bi- or tridentate coordination, as shown in pathway C. Therefore, similar to the case of AA, TEA efficiently stabilizes the Ti precursor, retards the hydrolysis rate and controls the inorganic polymerization of TBOT. This inhibits the growth and agglomeration of free TiO2 in the solution and induces the full distribution of nanoparticles onto both sides of the GO layers (sample TG-TEA).
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
145
The assistance of chelating ligands plays very important role in the slow hydrolysis–condensation of titanium precursors, facilitating crystal growth, nucleation and distribution of TiO2 nanocrystals on the two-dimensional graphitic nanosheets. The advantage of such a sandwich-like structure for the hybrids with stronger chemically bounded interaction can be demonstrated through removal efficiency of dye wastewater. Herein, the dark adsorption behavior of RhB solution was investigated, followed by the photocatalytic degradation of under both UV-A and visible irradiations. Firstly, for the evaluation of the adsorptivity, the adsorbed quantity of RhB, qt (mg g1 cat ), at the interval time t was calculated by the following equation [32,33]:
qt ¼
ðC 0 C t Þ 1000 M W V 0 wcat
ð1Þ
where C0 and Ct (M) are the initial concentration and concentration at t of RhB, respectively; MW = 479.03 (g mol1) is the molecular weight of RhB; V0 (L) is the volume of aqueous dye solution; and wcat (g) is the mass of catalyst. Fig. 5a illustrates the adsorption behaviors of RhB during 120 min dark adsorption. The hybrids prepared by using the chelating ligands exhibited good adsorbability of the dye, retaining 25% RhB content in the aqueous solution. The dye uptakes over these materials are much higher than that of TG composite which immediately attained the plateau after 15 min and removed only 40% of organic molecules in total. It can be concluded that the adsorption amount significantly increased in the hybrids prepared under the assistance of AA and TEA. Based on the adsorption kinetic data, the equilibrium adsorption capacity, qe, and adsorption rate constant, kads, were estimated by applying the pseudo-first-order and second-order models [32–34]. The appropriate values are chosen according to the higher correlation coefficient (R2). In the present study, the adsorption kinetics were well-fitted with the pseudo-second-order model as below [32,34]:
dqt ¼ kads ðqe qt Þ2 dt
ð2Þ
where qe and qt are the adsorption capacity at equilibrium and at time t, respectively; and kads (gcat mg1 min1) is the rate constant of pseudo-second-order sorption. By integrating Eq. (2) for the boundary conditions (t = 0 to t = t and qt = 0 to qt = qt), the following linear form can be written as follows:
t 1 1 ¼ þ t qt kq2e qe
ð3Þ
Herein, the experimental data evaluated from the linearity transform (t/qt) versus t were entirely consistent with the pseudo-second-order kinetics as expressed in Fig. 5b. The qe and kads values were evaluated from the slope and intercept of the plot, respectively, and were summarized in Table 1. TG-AA and TGTEA display the dye adsorption capacities of 35.55 and 38:48 mg g1 cat , respectively, whereas qe value is quite low for TG composite, approximately 18:51 mg g1 cat . Correspondingly, the composite containing large TiO2 aggregates on the surface or at the edge of corrugated RGO layers shows the adsorption rate constant kads of 17.3 gcat mg1 min1, much faster than those of the materials consist of wholly covered TiO2 nanoparticles. In consequence, the significant improvement in the adsorptivity of the sandwiched hybrids can be attributed to the giant two-dimensional sheet structure of graphitic component, the p–p conjugation with face-to-face orientation between RhB and aromatic regions of RGO [4] and the remarkable increase in BET specific surface areas, 86 and 89 m2/g, respectively, compared with that using only water as solvent (see Table 1).
Fig. 7. (a) Photocatalytic degradation of RhB under visible illumination and (b) corresponding linear transform ln(C/C0) = f(t).
Subsequently, the photocatalytic performance was evaluated by the exposure of the materials at the equilibrium state under UV-A irradiation. As shown in Fig. 6a, TG-AA and TG-TEA show much higher degradation efficiencies, about 80–90%, than those of pure TiO2 and TG composite upon UV irradiation after 120 min (see Fig. S9 in Supplementary material). Their photocatalytic capacities 1 follow an order: TG-TEA (46:27 mg g1 cat Þ > TG-AAð44:52 mg gcat Þ > TGð27:06 mg g1 ), which is identical to the adsorptivities. It is cat markedly due to the significant increase in the accessible surface area of RhB over the prepared materials. Accordingly, the apparent reaction rate constants were quantitatively derived for the evaluation of photocatalytic activity. The reaction kinetics obeyed a Langmuir–Hinshelwood pseudo-first-order model (when C0 is very small) [33,35,36]:
ln
C C0
¼ kapp t
ð4Þ
where kapp (gcat mg1 min1) is the apparent first-order rate constant that is determined from the slope of the linearity plot as seen in Fig. 6a. kapp values of TG-AA and TG-TEA photocatalysts are much higher than that for TG composite composed of aggregated TiO2 clusters on RGO. TG-TEA hybrid shows the fastest rate, typically 14.66 103 min1. The enhanced UV-photocatalytic performance of such sandwich-like hybrids compared with the others can be also ascribed to the strong coupling effect with chemical interactions between well-adhered TiO2 nanocrystals and individual RGO sheets, possibly accelerating the interfacial electron-transfer. Furthermore,
146
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147
Scheme 2. Photodegradation of Rhodamine B over sandwich-like reduced graphene oxide–TiO2 hybrid under UV and visible irradiation.
RGO may act as an electron acceptor of photogenerated electrons from the conduction band of TiO2 nanocrystals and subsequently as a transporter to its two-dimensional p–p conjugational plane due to the higher work function of RGO (illustrated in Scheme 2) [15]. The well-dispersed TiO2 onto mono- or bi-layered RGO sheets which are in the form of giant conducting network are much beneficial for the long-lived electron–hole pairs. Additionally, as discussed above, the presence of oxygen defects might be found and they act as electron traps. These phenomena efficiently hinder the recombination of charge carriers, accelerate the interface charge transfer and therefore improve the UV-photodegradable efficiency of RhB. On the other hand, due to the significant changes in the optical properties, the visible-illuminated photocatalytic activity was also observed after reaching the adsorption saturation. As seen in Fig. 7a, TG composite displays a negligible reactivity, whereas in contrast, TG-AA and TG-TEA hybrids reveal quite good performance under the visible light with the decomposition of 30–35% adsorbed RhB. By adopting the similar kinetics mechanisms in Fig. 7b, the apparent photodegradation rate constants for these sandwiched-like materials are threefold greater than that of the others as given in Table 1. It is noteworthy that such performance promotion of the sandwich-like hybrids is entirely consistent with the UV–Vis-DRS results mentioned above. The removal efficiency of RhB in the extended light absorption range can be ascribed to the chemical interaction between TiO2 and graphitic component as well as the role of individual RGO sheet as a photosensitizer to enhance the utilization of visible light. More importantly, another reason is the assistance of AA and TEA as the chelating agents, which enrich the concentration of oxygen vacancies in the hybrids as mentioned previously. The presence of vacancy-induced band of electronic states facilitates the transport of the photocarriers to the active sites on the surface as well as reduces the energy needed for photoexcitation [25]. These facts were elucidated to be responsible for the large photodegradation activity in the visible range. The results are expected to explore widely practical applications in the air/water purification and energy conversion. 4. Conclusions We have successfully synthesized RGO–TiO2 hybrid materials by one-step hydrothermal method in which the growth, nucleation
and full coverage of TiO2 nanocrystals on both sides of large RGO sheets were investigated for the first time by controlling the chelating agents. Without chelation, aggregated TiO2 clusters were formed and randomly distributed onto RGO. In the presence of chelating ligands, such a sandwich-like structure with strong chemical interaction between metal oxides and graphitic layers exhibited much better removal efficiency of dye molecules under both UV and visible light illuminations. The RGO–TiO2 system may serve a variety of applications such as photocatalysis, electronic devices, batteries, solar cells or supercapacitors. Acknowledgment This work was supported by the development program of local science park funded by the Ulsan Metropolitan City and The MEST (Ministry of Education, Science and Technology). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcis.2011.10.021. References [1] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666. [2] A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183. [3] D. Li, M.B. Muller, S. Gilje, R.B. Kaner, G.G. Wallace, Nat. Nanotechnol. 3 (2008) 101. [4] H. Zhang, X. Lv, Y. Li, Y. Wang, J. Li, ACS Nano 4 (2010) 380. [5] S. Chen, J. Zhu, X. Wu, Q. Han, X. Wang, ACS Nano 4 (2010) 2822. [6] I.V. Lightcap, T.H. Kosel, P.V. Kamat, Nano Lett. 10 (2010) 577. [7] O. Akhavan, E. Ghaderi, J. Phys. Chem. C 113 (2009) 20214. [8] Y.H. Ng, I.V. Lightcap, K. Goodwin, M. Matsumura, P.V. Kamat, J. Phys. Chem. Lett. 1 (2010) 2222. [9] P. Wang, Y. Zhai, D. Wang, S. Dong, Nanoscale 3 (2011) 1640. [10] N. Yang, J. Zhai, D. Wang, Y. Chen, L. Jiang, ACS Nano 4 (2010) 887. [11] Y. Liang, H. Wang, H.S. Casalongue, Z. Chen, H. Dai, Nano Res. 3 (2010) 701. [12] K. Zhou, Y. Zhu, X. Yang, X. Jiang, C. Li, New J. Chem. 35 (2011) 353. [13] J. Shen, B. Yan, M. Shi, H. Ma, N. Li, M. Ye, J. Mater. Chem. 21 (2011) 3415. [14] X.-Y. Zhang, H.-P. Li, X.-L. Cui, Y. Lin, J. Mater. Chem. 20 (2010) 2801. [15] J. Zhang, Z. Xiong, X.S. Zhao, J. Mater. Chem. 21 (2011) 3634. [16] D.C. Bradley, R.C. Mehrotra, D.P. Gaur, Metal Alkoxides, Academic Press, 1978. [17] X. Chen, S.S. Mao, Chem. Rev. 107 (2007) 2891. [18] W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339.
T.-D. Nguyen-Phan et al. / Journal of Colloid and Interface Science 367 (2012) 139–147 [19] T. Szab, O. Berkesi, P. Forg, K. Josepovits, Y. Sanakis, D. Petridis, I. Dkny, Chem. Mater. 18 (2006) 2740. [20] W. Gao, L.B. Alemany, L. Ci, P.M. Ajayan, Nat. Chem. 1 (2009) 403. [21] D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, Chem. Soc. Rev. 39 (2010) 228. [22] F. Tuinstra, J.L. Koenig, J. Chem. Phys. 53 (1970) 1126. [23] Y. Djaoued, S. Badilescu, P.V. Ashrit, D. Bersani, P.P. Lottici, J. Robichaud, J. SolGel Sci. Technol. 24 (2002) 255. [24] P.W. Chou, Y.S. Wang, C.C. Lin, Y.J. Chen, C.L. Cheng, M.S. Wong, Surf. Coat. Technol. 204 (2009) 834. [25] I. Justicia, P. Ordejón, G. Canto, J.L. Mozos, J. Fraxedas, G.A. Battiston, R. Gerbasi, A. Figueras, Adv. Mater. 14 (2002) 1399. [26] D.-W. Wang, D.-M. Tang, C. Liu, X. Ma, G.Q. Lu, H.-M. Cheng, Nanotechnology 19 (2008) 25606.
147
[27] D.C. Bradley, R.C. Mehrotra, I.P. Rothwell, A. Singh, Alkoxo and Aryloxo Derivatives of Metals, Academic Press, 2001. [28] F. Ribot, P. Toledano, C. Sanchez, Chem. Mater. 3 (1991) 759. [29] A. Ponton, S. Barboux-Doeuff, C. Sanchez, Colloids Surf. A 162 (1999) 177. [30] F. Wang, J. Sun, C. Ma, Stud. Surf. Sci. Catal. 165 (2007) 305. [31] M. Bockmeyer, P. Löbmann, J. Sol-Gel Sci. Technol. 45 (2008) 251. [32] R.A. Shawabkeh, M.F. Tutunji, Appl. Clay Sci. 24 (2003) 111. [33] W. Kangwansupamonkon, W. Jitbunpot, S. Kiatkamjornwong, Polym. Degrad. Stab. 95 (2010) 1894. [34] Y.S. Ho, G. McKay, Trans. Inst. Chem. Eng. 76B (1998) 332. [35] X. Li, J. Ye, J. Phys. Chem. C 111 (2007) 13109. [36] K. Yu, S. Yang, H. He, C. Sun, C. Gu, Y. Ju, J. Phys. Chem. A 113 (2009) 10024.