Introduction

Graphene, a monolayer of carbon atoms with a tight packing of honeycomb lattice, has attracted immense research interest in nanotechnology in recent years (Geim and Novoselov 2007). Graphene exhibits outstanding properties such as very low resistivity, high mobility of charge carriers, optical, and electrical properties, and new quantum effect (Mukherjee and Kaloni 2012; Novoselov et al. 2005); which has been found to be useful in a variety of technological applications, including biotechnology (Dong et al. 2010; Guo et al. 2010; Liu et al. 2008; Yang et al. 2009), photocatalysis (Liang et al. 2010), and many energy storage devices (Pumera 2011) such as solar cells (Miao et al. 2012), supercapacitors (Chen et al. 2010b; Wang et al. 2010b), and lithium-ion batteries (LIBs) (Chou et al. 2010; Lee et al. 2010; Shen et al. 2011; Wang et al. 2009).

Graphene oxide (GO) was proven to be an effective and reliable precursor to produce graphene nanosheets owing to its low cost and massive scalability (Park and Ruoff 2009). Moreover, functionalized graphene-based materials can be prepared via chemical modification, which show unique electronic and optical properties as well as good biocompatibility (Akhavan et al. 2011; Sun et al. 2011). Since the creative work of Zhang et al. (2010a), who prepared a chemically bonded TiO2 (P25)–graphene composite photocatalyst by a hydrothermal method, many efforts have been made to utilize the unique properties of graphene in order to increase the efficiency of photo conversion (Liu et al. 2010; Wang et al. 2010c; Yoo et al. 2011). Graphene and GO-based nanosheets could not only assist the dispersion of the metal oxide in aqueous media but also direct the self-assembly of metal oxides into nanostructures (Park and Ruoff 2009; Wang et al. 2010a; Wu et al. 2010). Notably, enhanced cycling performance for Li-ion battery has been observed with GO-supported Co3O4 nanoparticles or manganese oxide nanosheets (Wu et al. 2010; Zhang et al. 2011).

As an important n-type semiconductor, hematite (α-Fe2O3) has been widely used for gas sensors (Chauhan et al. 1999; Chen et al. 2005), water splitting (Cesar et al. 2006), photodetectors (Hsu et al. 2009), solar cells (Beermann et al. 2000), field-emission devices (Hsu et al. 2008; Yu et al. 2006), lithium-ion batteries (Liu et al. 2009; Nuli et al. 2008; Reddy et al. 2007; Wu et al. 2008), water purification (Song et al. 2012), and field-effect transistors (Fan et al. 2005). The α-Fe2O3 has also been selected as the support material for noble metal (Chen et al. 2005; Gou et al. 2008; Hu et al. 2007; Zhang et al. 2010b). Our group recently reported the photoelectric characteristics of α-Fe2O3 particles with hollow spindle shape (Li et al. 2012).

Considering that extensive research efforts have demonstrated that intriguing properties for various applications could be obtained by coupling functional nanoparticles with graphene derivatives (Chen et al. 2010a; Chou et al. 2010; Ji et al. 2011; Lee et al. 2010; Lian et al. 2010; Liang et al. 2010; Liu et al. 2010; Shen et al. 2011; Wang et al. 2009, 2010a, 2010b, 2011a, 2011b; Wu et al. 2010; Yang et al. 2009; Zhang et al. 2010a, 2011; Zhou et al. 2010; Zhu et al. 2011a, 2011b), we believe that it is interesting to explore more physical and chemical properties of different graphene-iron oxide composites. A number of recent reports have focused on the reduced GO-coupled Fe2O3 (Zhu et al. 2011a, 2011b) and graphene/Fe3O4 composites (Ji et al. 2011; Lian et al. 2010; Wang et al. 2011b; Zhang et al. 2010c; Zhou et al. 2010) for lithium-ion batteries. So far, only a few reports have been focused on the photo-induced performance of GO and α-Fe2O3 based composites. Meng et al. (2013) studied the hematite/reduced GO composites for photocatalytic water oxidation. They found that coupling the hematite nanoparticles with the reduced GO could greatly increase the photocurrent and reduce the charge recombination rate, which led to a higher O2 evolution rate than α-Fe2O3 by a factor of two. Kim et al. (2012) reported an effective conducting scaffold of graphene–carbon nanotube composite for enhancing the photoelectrochemical water oxidation activity of hematite film. The delicate 3D-like architecture deriving from 1D carbon nanotube and 2D graphene formed a more open and highly exposed structure and effectively reduced the self-agglomeration of the carbon nanosheets, which could thus provide higher contact area between the conducting scaffold and Fe2O3 particles and also retain the intrinsic conducting ability of graphene and carbon nanotube for photoelectrochemical application. The above two instances reveal the attractive prospects of the photo-induced activity of graphene derivatives and α-Fe2O3-based composites.

In this work, we report the fabrication and the photocatalytic property of GO sheets-supported α-Fe2O3 microparticles in degrading a typical organic air pollutant. It is demonstrated that the α-Fe2O3/GO composite exhibited enhanced photo activity in terms of light absorption and photocatalytic degradation of toluene compared to the pure α-Fe2O3 counterpart.

Experimental

Synthesis of GO

GO was synthesized from graphite powder by a modified Hummers’ method (Hummers and Offeman 1958), in which pre-oxidation of graphite was followed by oxidation according to Hummers’ method. For the pre-oxidation, the graphite powder (4.0 g) was added with stirring into the concentrated H2SO4 (40 mL) solution in which K2S2O8 (2.0 g) and P2O5 (2.0 g) were completely dissolved at 80 °C. The mixture, in a beaker, was kept at 80 °C for 4.5 h using an oil bath. Then, the mixture was cooled down and was diluted with 1 L of distilled deionized water. The pretreated product was filtered and washed until the pH of filtrate water became neutral. The shiny, dark-gray, pre-oxidized graphite was dried in air overnight. Then, it was dispersed by stirring into chilled H2SO4 (92 mL) in a flask in an ice bath. KMnO4 (12 g) was added slowly with stirring to keep the temperature of reaction mixture below 20 °C. The resulting thick, dark green paste was allowed to react at 35 °C for 3 h followed by addition of H2O (184 mL) to give a dark brown solution. After additional stirring for 2 h, the dark brownish solution was further diluted with distilled water (560 mL) and then H2O2 (30 %, 40 mL) was added slowly and the color of the mixture turned into brilliant yellow. The mixture was allowed to settle overnight and the supernatant was decanted. The remaining product was washed with 10 % HCl solution with stirring and the brownish solution was allowed to settle overnight. The supernatant was decanted and the remaining product was centrifuged and washed with deionized water. The washing process was repeated until the pH of the solution became neutral. After filtration and drying under vacuum, GO was obtained as a gray powder.

Synthesis of spindle-like α-Fe2O3

Spindle-shaped α-Fe2O3 particles were obtained via a hydrothermal route. FeCl3·6H2O (18 mmol) was added into distilled water (60 mL) to form a clear solution. Then 4.32 g urea (CH4N2O) and 9 mL ethylene glycol, a protective agent were added into the solution. Subsequently after magnetic stirring for 30 min at room temperature, the resultant solution was put into a Teflon-lined stainless steel autoclave of 100 mL capacity, which was sealed and maintained at 180 °C for 12 h. After cooling the autoclave to room temperature naturally, the obtained precipitate was collected by filtration, and then washed with absolute ethanol and distilled water in sequence for several times. The α-Fe2O3 product was dried in a vacuum box at 80 °C for 12 h.

Synthesis of the α-Fe2O3/GO composites

The α-Fe2O3/GO composites were synthesized as follows. 0.05 g GO was dispersed into 50 mL distilled water and the solution was ultrasonicated for 3 h. 4.8654 g FeCl3·6H2O was dissolved in the solution with vigorous stirring for 0.5 h. Afterward, the aqueous solution of 10 mL HCl (1 M) was added into the above solution, then 12 mmol urea and 2.5 mL ethylene glycol were added into the solution with continuous stirring for 0.5 h. After that, the resultant solution was put into a Teflon-lined stainless steel autoclave of 100 mL capacity, which was sealed and maintained at 180 °C for 12 h. After cooling to room temperature naturally, the resulting product was separated by suction filtration and rinsed with distilled water and ethanol several times to remove the excess FeCl3, and then dried at 60 °C for 5 h under vacuum, resulting in the formation of the final products.

Characterizations

The crystalline phase structures of the prepared α-Fe2O3 and α-Fe2O3/GO samples were determined by X-ray diffraction (XRD, RIGAKU, Dmax22000) with Cu Kα radiation (λ = 0.15418 nm) over the 2θ range of 20°–80°. The morphology of the samples was investigated by transmission electron microscopy (TEM, FEI Tecnai G220). Moreover, high-resolution transmission electron microscopy (HRTEM, JEOL-2010 system at 200 kV), and selected area electron diffraction (SAED, Tecnai G220) were used to characterize the structure of nanocrystals. Surface imaging of the GO sheets was carried out using NT–MDT Solver PRO atomic force microscope (AFM) in tapping mode. In order to take images, samples were initially dispersed in ethanol solution and spray-coated on a freshly cleaved mica surface using a fine atomizer. The amount of GO species was determined by inductive coupling plasma (ICP) IPS-AEC (Optima 2000 DV, Perkin Elmer, USA). Surface area of the samples was determined by Brunauer-Emmet-Teller (BET). N2 gas adsorption–desorption isotherms were obtained at 77 K on a Micromeritics ASAP-2000 equipment. Samples were precedently degassed at 400 °C. Pore size distributions were calculated by the BJH model. Light absorption property was examined using a UV–Vis diffuse reflectance spectrophotometer (JASCO, UV-550). Raman spectra of synthesized samples were recorded using Raman spectroscopy (Renishaw, InVia) at 632.8 nm laser radiations. The XPS data were recorded using an ESCALAB250 electron spectrometer using achromatic Al Kα radiation (1,486.6 eV) with Ar+ sputtering to remove the surface layer of the sample.

Photocatalytic performance study of the α-Fe2O3 and α-Fe2O3/GO samples

In situ FT-IR spectra were collected with a Fourier Transform Infrared Spectrophotometer (BRUKER VERTEX 70 Optics) and a home-made in situ IR quartz photoreaction cell (Zheng et al. 2009). The photocatalyst was illuminated by an 500 W xenon lamp. The distance between the lamp and sample was about 15 cm. The light intensity at the sample holder was about 40.5 mW cm−2. Two pellets were prepared in parallel.

The reaction cell was purged by dry air for 1 h. After 1 h, the flux of dry air was set at 20 mL/min. Spectra of the clean catalyst surface were collected after this process and utilized as the background. Subsequently, toluene species was fed at the flow rate of 2 μL h−1 for ca. 30 min using a syringe pump to a mixing tee where it was vaporized and mixed with the dry air. The reactant mixture then flowed through the reaction cell and allowed to equilibrate at room temperature (293 K). We determined whether the reactant concentration was stabilized by keeping on collecting its infrared spectrum every 5 min until a stable peak line on the infrared spectrum was obtained. Once the reactant concentration was stabilized, the inlet and outlet ports were shut off and the lamps were turned on. The infrared spectra were continuously collected during the reaction. The infrared spectra were collected with a resolution of 1 cm−1 and 20 scans in the region of 4,000–1,000 cm−1.

The reactant concentration was measured by gas chromatography (Aligent 7890A, USA). During the reaction process, samples of approximately 1 μL at the outlet every 0.5 h were collected and the change in concentration of toluene was analyzed by a gas chromatography equipped with FID (HP-5 capilary column (30 m × 320 μm × 0.25 m)) and TCD (Porapak Q). When the peak intensity of toluene did not obviously change, the xenon lamp was turned off and the cell was purified with nitrogen.

Results and discussion

XRD analysis

Figure 1 shows the XRD patterns of the natural graphite, GO, α-Fe2O3, and α-Fe2O3/GO composite. The natural graphite shows a diffraction at 2θ = 25.6°, corresponding to the crystal planes of graphene (d-spacing of 3.35 Å) (Wang et al. 2011a). The pattern of GO has a peak centered at 2θ = 10.65°, corresponding to the (002) inter-planar spacing of 9.235 Å (Xu et al. 2008). The α-Fe2O3/GO composite and α-Fe2O3 possess sharp diffraction peaks at 2θ of 24.18°, 33.15°, 35.75°, 40.93°, 49.43°, 54.02°, 57.56°, 62.51°, 64.05°, 71.94°, and 75.48° corresponding to (012), (104), (110), (113), (024), (116), (018), (214), (300), (1010), and (220) crystallographic nucleation planes of hematite phase, respectively, which can be indexed to rhombohedral structure α-Fe2O3 (JCPDS 84-0311) (Li et al. 2011). These results indicate that α-Fe2O3/GO composite with relatively high crystal purity is synthesized. It can be observed that diffraction peaks of bare α-Fe2O3 have stronger intensity than those of α-Fe2O3/GO composite.

Fig. 1
figure 1

The X-ray diffraction patterns of the α-Fe2O3, graphite, GO, and α-Fe2O3/GO samples

Full size image

As calculated from the XRD line broadening of the (104) peak using the Scherrer equation D = 0.89λ/βcosθ, the average crystallite sizes of the α-Fe2O3 and α-Fe2O3/GO samples are ca. 27 and 24 nm, respectively. The broader peaks of α-Fe2O3/GO composite further indicate their smaller average crystallite size, although it is not obvious.

AFM, TEM, and ICP analysis

An AFM image of the GO sample is shown in Fig. 2a. The image of GO shows two pieces of monolayer with thickness of about 2.0 nm. The TEM image in Fig. 2b shows the morphology of the pure α-Fe2O3 sample. The α-Fe2O3 sample is composed of “spindle-like” microparticles with the dimension sizes around one micrometer. Statistical analysis of dimensional distribution based on low-resolution TEM images indicates that the length and the width are about 1.5 and 0.5 μm on average, respectively.

Fig. 2
figure 2

An AFM image of the GO sample (a), the TEM images of the α-Fe2O3 sample (b) and α-Fe2O3/GO sample (c) and (d), SAED image (inset) and HRTEM image of the α-Fe2O3/GO sample (e)

Full size image

Figure 2c shows a low-resolution TEM image of α-Fe2O3/GO composite. It can be seen that α-Fe2O3 particles are closely anchored on the surface of GO nanosheets. The agglomeration of α-Fe2O3 particles is not observed, which confirms that GO nanosheets are uniformly coated with α-Fe2O3 particles. The α-Fe2O3 particles grew onto the GO nanosheets during the synthesis process and appeared as microballs with sizes of 500–1,000 nm in the produced composite. Even after a long duration of sonication during the preparation of the TEM specimen, the GO nanosheets could act as stable supporting materials for anchoring of α-Fe2O3 particles.

The TEM image of Fig. 2d with a higher resolution indicates that the α-Fe2O3 microparticles grown on GO were formed through the agglomeration of numerous nanoparticles. The apparent sizes of these constituent α-Fe2O3 nanoparticles are in the range of about 20–100 nm. This result is consistent with the XRD-determined average crystallite size of ca. 24 nm. The selected area electron diffraction (SAED) pattern (inset of Fig. 2e) of the particles reveals ring patterns corresponding to hematite Fe2O3 with rhombohedral crystal structure (JCPDS 89-2810) (Zhu et al. 2011a), while the HRTEM observation (Fig. 2e) of the crystals confirms the same crystal structure. In the HRTEM image (Fig. 2e) of the α-Fe2O3/GO sample, the spacings between the lattice fringes are 0.25, 0.27, and 0.37 nm, which is consistent with hematite phase 110, 104, and 012 plane spacing (Reddy et al. 2007), respectively. From the ICP analysis, we know that the composite contains 0.256 wt% GO around the α-Fe2O3 particles.

BET analysis

The specific surface area and porosity of the α-Fe2O3/GO and Fe2O3 samples have been determined and displayed in Table 1. The specific surface areas of the α-Fe2O3/GO and α-Fe2O3 samples are 70.79 and 12.62 m2 g−1, respectively. It shows that the GO constituent makes the specific surface area of the sample increase a lot. Analysis of porosity was carried out following the BJH method. The α-Fe2O3 and α-Fe2O3/GO samples display monomodal pore size distribution with the most frequent pore radii of about 1.67 and 2.18 nm, respectively. The prepared samples are all mesoporous materials according to the distribution of pore sizes. The total pore volumes (P/P 0 = 0.99) of the α-Fe2O3 and α-Fe2O3/GO samples are 0.018 and 0.059 cm3 g−1, respectively. The sample α-Fe2O3/GO has a higher specific surface area and monomodal pore size distribution. These features determining the accessibility of reactant molecules are important for the catalytic performance, which implies that the α-Fe2O3/GO composite may have higher catalytic activity than the α-Fe2O3 particles.

Table 1 Specific surface area, pore volume data, and pore size distribution of the α-Fe2O3/GO and α-Fe2O3 samples
Full size table

DRS analysis

The UV–Vis absorption spectra of the α-Fe2O3 and α-Fe2O3/GO samples are shown in Fig. 3a. A stronger visible light absorption with the same peak at about 574 nm as α-Fe2O3 was found for the α-Fe2O3/GO composite, which indicates that the carbon sheets could help absorb much more UV–visible light. The introduction of GO into the α-Fe2O3 leads to a better utilization of UV–visible light without changing the spectral pattern of α-Fe2O3.

Fig. 3
figure 3

a The UV–Vis absorption spectra of the α-Fe2O3 and α-Fe2O3/GO samples, and b the Raman spectra of the prepared α-Fe2O3, GO, and α-Fe2O3/GO samples

Full size image

Raman spectra analysis

Figure 3b shows the Raman spectra of GO, spindle-like α-Fe2O3, and α-Fe2O3/GO composite. Two characteristic peaks of the D and G bands from GO could be observed at around 1,330 and 1,580 cm−1. The intensity ratios of D and G band (ID: IG) are 1.21, 1.47, and 1.32, respectively. The more disordered carbon structure in the α-Fe2O3/GO composites compared with bare GO may be due to the possible insertion of α-Fe2O3 into GO layers (Pumera 2011).

XPS analysis

The electronic structure of spindle-like α-Fe2O3 and α-Fe2O3/GO composite was investigated by XPS. Figure 4 shows the representative XPS spectra of α-Fe2O3 and α-Fe2O3/GO composite. The survey spectrum of α-Fe2O3/GO reveals the spectroscopic signature of carbon, oxygen, and iron. The Fe 2p3/2 peak is centered at 721 eV and the Fe 2p1/2 peak at 707.4 eV. This indicates the oxidation state of Fe3+ in the sample (Grosvenor et al. 2004). The oxygen peak with O1s at 526.7 eV and the carbon peak with C 1s at 281.22 eV on the surface are present in the α-Fe2O3/GO composite sample. A quantitative analysis of the data in the Fe2p3/2, Fe2p1/2, and C1s regions allows to estimate the atom ratio of Fe:O:C = 1:2.8:0.22 (Table 2). The above results further indicate the successful preparation of α-Fe2O3/GO composite.

Fig. 4
figure 4

The whole XPS spectra of the α-Fe2O3 and the α-Fe2O3/GO samples (a), Fe 2p XPS spectrum (b), C 1s XPS spectrum (c), and O 1s XPS spectrum of the α-Fe2O3/GO sample (d)

Full size image
Table 2 The results of XPS analysis for the α-Fe2O3/GO sample
Full size table

Photocatalytic capacity study on α-Fe2O3 and α-Fe2O3/GO samples

In situ FTIR study provides real-time monitoring of the transient events occurring on the catalyst during the reaction. The spectral information on the adsorption and the photodegradation of toluene over the synthesized α-Fe2O3/GO sample and α-Fe2O3 sample are shown in Fig. 5. Upon toluene entering the reaction cell, the intensities of the characteristic peaks for toluene increased quickly. After 25 min, the intensities of peaks for toluene over α-Fe2O3/GO sample decreased slowly and had a steady trend (Fig. 5a). For the α-Fe2O3 sample, the adsorption/desorption processes equilibrated in 60 min (Fig. 5b). Then the inlet and outlet ports were shut off. The initial concentration of the toluene was determined as about 750 mg m−3 by gas chromatography.

Fig. 5
figure 5

The infrared transmittance spectra of toluene adsorbed onto the α-Fe2O3/GO (a) and α-Fe2O3 samples (b) under dark condition, the infrared transmittance spectra for toluene degradation over the α-Fe2O3 (c) and the α-Fe2O3/GO samples (d, e) under 5 h irradiation by xenon lamp, and the comparison on the toluene degradation results using the catalysts and without any catalyst (f)

Full size image

During xenon-lamp illumination, a set of infrared transmission spectra were collected during the photocatalytic degradation of toluene over the α-Fe2O3/GO sample and α-Fe2O3 sample. Figure 5c–e displays the spectra for toluene photodegradation over α-Fe2O3/GO sample and α-Fe2O3 sample for 5-h illumination. Prior to xenon-lamp illumination (t = 0), the spectrum displays the characteristic toluene bands at 2,979 and 2,890 cm−1 (Fig. 5c, d). Upon irradiation, the intensity of these bands began to decrease slowly. The bands at 2,360 cm−1 and 2,340 cm−1 corresponding to CO2 (Fig. 5c, d) and H2O (1,640 cm−1) (Fig. 5c, d) increased as the reaction proceeded.

The color of catalysts was observed to change from brown to black. This degeneration of catalysts may be related to some new surface species, which could be explained by the FTIR spectra in Fig. 5e. Some bands are related to the adsorbed water and hydroxyl groups in the range of 1,300–1,700 cm−1 (bands around 1,657, 1,495, 1,451, 1,417, and 1,312 cm−1) (Fig. 5e). Similar bands have been reported for benzaldehyde adsorption on nanostructured TiO2 and Degussa P25 (Mendez-Roman and Cardona-Martinez 1998; Martra 2000). These results indicate that weakly adsorbed benzoic acid was formed as the reaction proceeded (Morterra 1988). Benzoic acid molecules resulting from the photo-oxidation of toluene interact with the catalyst surface weakly, so that they can be released to the gas phase. By contrast, hydroxyl groups on the surface of α-Fe2O3 microparticles are able to react with benzoic acid (Marcì et al. 2003), which is adsorbed on the surface of α-Fe2O3 microparticles, leading to the progressive deactivation of the catalyst in the gas–solid system.

Meanwhile, we detected the concentration change of toluene by gas chromatography during xenon-lamp illumination. Toluene was sampled each 0.5 h. The initial concentration of toluene measured by gas chromatography was 750 mg m−3. Figure 5f shows the comparison between the concentration changes of toluene by 5-h illumination over α-Fe2O3, α-Fe2O3/GO sample, and without any catalyst. After 5-h prolonged irradiation, degradation percentages of toluene over the α-Fe2O3, α-Fe2O3/GO sample, and without any catalyst reached to 70, 75, and 36 %, respectively. The α-Fe2O3/GO composite with sparser distribution of α-Fe2O3 particles could exhibit a little higher photocatalytic activity than the α-Fe2O3 sample, even though the α-Fe2O3 content in α-Fe2O3/GO composite is diluted by the two nanometer-thick carbon sheets.

The stability of the catalyst α-Fe2O3/GO

The stability test about the catalyst α-Fe2O3/GO has been done under the same in situ FT-IR test condition for four reaction cycles. The photodegradation ratio of the substrate was measured as a function of time. The results are shown in Fig. 6. It was found that the photodegradation ratio of toluene reached to 71 % after the fourth repetition over the catalyst α-Fe2O3/GO, which was slightly lower than that (75 %) with the fresh catalyst. This result shows that the α-Fe2O3/GO catalyst has good stability and reusability.

Fig. 6
figure 6

Cyclic photodegradation of toluene by the α-Fe2O3/GO sample

Full size image

Conclusions

In summary, α-Fe2O3/GO composite was prepared via a simple hydrothermal and Hummers’ method. The possible growth of α-Fe2O3 onto GO layers led to a higher absorbance capacity for visible light by α-Fe2O3/GO composite than α-Fe2O3 microparticles. Comparing the photocatalytic oxidation of gaseous toluene over the samples, the prepared α-Fe2O3/GO composite was more active than α-Fe2O3. Weakly adsorbed benzaldehyde and benzoic acid could be formed on the catalyst surface. The hydroxyl groups played a key role during the photocatalytic process, and the dehydroxylation should be responsible for the deactivation of the composite in a long-term reaction. The composite of α-Fe2O3/GO could be promisingly utilized in photo-aided air purification as a green catalyst.