3.1 XRD analysis of nanocomposite
Figure
3 shows the XRD spectrum for the as-prepared GO and nanocomposite powders GO/ α-Fe
2O
3. The peaks were shown with symbol
(●) located at 23.25°, 33.18°, 35.75°, 40.89°, 49.52°, 54.14°, 57.52°, 62.49°, 64.07° correspond to standard data of pure α-Fe
2O
3 diffraction peaks (reference code 01-079-1741). These peaks ascribe to the rhombohedral crystal structure of Hematite with a lattice parameter of a = b = 5.03 A° and c = 13.74 A°. The GO with symbol (*) displays a characteristic peak at 10.37° with an index of (001) defining 0.83 nm d-spacing which is increased from 0.34 nm interlayer spacing of graphite. This significant increase in d-spacing of GO is assigned to the presence of oxygen functional groups such as carboxyl groups (COOH-), hydroxyl groups (OH-), epoxy groups (C-O-C) that can be confirmed by FTIR. The identification of GO peaks was hardly recognized from the composite pattern which might be explained by the minor amount of GO (α-Fe
2O
3 -0.05% GO) in nanocomposite and high dispersion of GO-sheets among Fe
2O
3 nanoparticles.
According to the Debye- Scherrer equation, the particle size of hematite was obtained from the XRD pattern; \(D=\frac{\text{K}\text{ƛ}}{{\beta } \text{c}\text{o}\text{s}\text{Ɵ}}\) (where K = 09, ƛ = 0.154nm, β = the reflection width (2θ) and θ = the Bragg angle). The result shows that the average particle size is about 48.27 nm.
3.2 FTIR spectrum of nanocomposite
FTIR spectra of the nanocomposites were obtained under transmission mode to identify the functional groups of compounds presented in GO- α-Fe2O3 composite. Fig. 4 shows the spectra of GO and GO-α-Fe2O3. The band located at 3400 cm-1 is assigned to the O-H stretching vibration of C–OH groups, in which this broad absorption is decreased in GO-α-Fe2O3 spectrum due to the reduction of GO during heat treatment and restoration of the conjugated aromatic system. Furthermore, the C–O stretching vibrations, C–O–C stretching modes, and C=C skeletons vibration in GO spectrum correspond to the bands located at 1038, 1215, 1615 cm-1, respectively. The C=O stretching band (COOH groups), located at the edge of GO sheets, was emerged at 1735 cm-1. The absorption bands at 468 cm-1 and 533 cm-1 are ascribed to Fe–O stretching mode, which verify the existence of α-Fe2O3 chemical compounds attached to the –COO on the edge of GO nano sheets. The next clue for the formation of monodentate and bidentate ligands in a complex between Fe and the carboxyl group is the additional vibrational band at 1385 cm-1, confirming the covalent bond formation between hematite and GO. The carbonyl group of GO at the 1640 cm-1 can be assigned as coordination with Fe on the surface of GO sheets [37].
3.3 Raman spectra of nanocomposite
Characterization for structural properties of carbon material including defect density and disorder structures is investigated by Raman spectroscopy (Fig. 5). The Raman spectroscopy of GO revealed two prominent peaks at 1304 and 1582. The D-peak that corresponds to the breathing modes of carbon sp2 rings (κ-point phonons of A1g symmetry) required defects and disordered atomic arrangement for its activation. The G-peak corresponds to the E2g phonon of carbon sp2 atoms. These two peaks of GO also could be recognized in Raman spectroscopy of GO-α-Fe2O3 as well as fundamental Raman vibration of α-Fe2O3, suggesting that the structure of GO remained in the composite. In an additional analysis of the Raman spectrum, the presence of strong peaks at 221, and 287 cm-1 (A1g symmetry), and weaker peaks at 404, 490, and 605 cm-1 (Eg symmetry) show the formation of hematite phase accurately. It can also be observed that the ID/IG ration elevated from 0.865 for GO to 1.412, which confirms penetration of Fe2O3 nanoparticles between GO layers as well as increasing disorders and defects in GO sheets [38].
3.4 Microscopy analysis of nanocomposite
The detection of morphology, size of particles, shape, and thickness of nanocomposite were done by FESEM and AFM analysis. The FESEM images of nanoparticles of α-Fe2O3 and nanocomposite α-Fe2O3- 5% GO are shown in Fig. 6. These images provide clear evidence that α-Fe2O3 particles with the average size of 63 nm were uniformly distributed on GO nano sheets. These well-dispersed hematite nanoparticles provide direct interaction with GO surface, which is desirable for photocatalytic application. The thickness of the exfoliated GO sheets was measured by AFM which have a sharp edge with a thickness of 2.5 nm. The typical thickness of monolayer GO sheets is in the range 0.7 nm. Thus, it might be a four layered GO sheet.
3.4.1 Characterization of α-Fe2O3/GO thin film nanocomposites
X-ray diffraction patterns was done for identifying the composite structures on the FTO substrate. As it is clear in Fig. 7, the identified peaks at 32.3.3°, 35.9°, 54.1° and 63.6° shown by symbol (●) are contributed to the crystallographic directions of the α-Fe2O3 phase. Diffraction peaks at 31.1°, 39.2°, 44°, 60.9° and 73.1° with symbol (*) show the presence of FTO substrate in the nanocomposite. The XRD also confirmed the existence of GO(*) on the FTO substrate at peak 10°.
Fig. 8 shows the morphology of GO/α-Fe2O3 thin films which was annealed at different temperatures of 300, 400, 500, and 700 °C. Based on the FESEM analysis of GO/ α-Fe2O3 thin films after annealing at 300, 400, 500 and 700 °C, the uniform distribution of iron oxide nanoparticles on the surface of graphene layers was occured. The corrugated and crumpled morphology of GO nanosheets is clearly seen in all synthesized GO/ α-Fe2O3 thin films. Increasing the temperature resulted in the formation of a higher rode shape structure in the layers. The perfect synthesis of α-Fe2O3 nanoparticles with GO nanosheets occurs in most part of layers, which proves the reduction of agglomeration in GO nanosheets and decrease in the aggregation of nanoparticles. The more combinations of nanoparticles and GO nanosheets on the FTO substrate may lead to higher specific surface area. According to the FESEM analysis in higher temperatures (Figs. 8Ac, and 8Ad), it seems that the nanoparticles have diffused into the sublayer resulting in the size of nanoparticles to seem smaller on the surface. The FESEM-EDS mapping of all presented elements in the thin-film layer is shown in Fig. 8 (b). The result of the analysis shows that Fe and O elements were equally diffused on the surface of GO nanosheets.
XPS analysis was done to identify the surface chemical composition of thin-film Fe2O3/GO [39]. The Fe orbitals (2p), O (1s), and C (1s) were shown in the XPS spectrum of synthesized thin-film nanocomposite at 700 °C. According to Fig. 9, the XPS peaks Fe 2p3/2 and Fe 2p1/2 are located in the binding energy of 711.1 and 725.1 eV, respectively. The satellite peak of Fe 2p3/2 is approximately 8 eV higher than the main peak. Here, the satellite peak is close to 718.7 eV. The Fe 2p3/2 peak is narrower and stronger than Fe 2p1/2 peak. Also, the area of Fe 2p3/2 is greater because of spin-orbit (j-j) coupling. XPS analysis also showed three peaks for C (1s) including non-oxygenated aromatic carbons with orbital sp2 at 284.67 eV, oxygenated functional group C—O at 286.00, and C==O at 288.5 eV. The existence of anionic oxygen in Fe2O3 and functional oxygen groups in GO sheets created the binding energy of 531.26 eV which is for O (1s) peak (Fig. 9(d)).
3.5 Photocatalytic test of α-Fe2O3/GO nanocomposites and α-Fe2O3/GO thin films
The photocatalytic properties of the nanocomposites and thin film layers were tested using irradiation of xenon lamp (185-2000 nm). As for α-Fe2O3/GO powder nanocomposites, the test was done without irradiation for the first 60 minutes to evaluate the absorption of nanocomposite in dark. Then, photodegradation was tested for 120 min irradiation. Maximum adsorption and absorption of samples are shown in Table 1 and Table 2, respectively. Different concentrations of synthesized nanocomposite were tested which showed excellent dye removal for all of them. Nanocomposites can excite more electrons to degrade dye solution due to its higher surface area. The gap energy between conduction and valence band in pure α-Fe2O3 is 2.33 eV. The recombination rate in αFe2O3 is relatively high and interaction between GO and αFe2O3 can reduce the recombination of electron hole pairs. Because of the strong reaction between negatively charged structure of GO and cationic dyes such as RhB, the photodegradation process increases significantly. Also, the strong connection of GO and αFe2O3 molecules facilitates the degradation process and reduces the recombination rate. In fact, the structure of synthesized nanocomposite delays the combination of electron-hole pairs in the system and therefore increases dye removal efficiency. Higher surface area is another reason for turning dye molecules into harmless products. More photo generated electron-hole pairs participate in degradation reactions, which finally leads to higher dye removal efficiency [40, 41]. As it is clear in Fig. 10, the pure αFe2O3 is by far lower than all powder nanocomposites especially 5 % GO/α-Fe2O3 with 64 % dye removal. The electrons were transferred through sp2 hybrid carbon channels on the GO sheets which reduced the recombination of electron-holes. Also, the presence of GO increases the light-receiving capacity which leads to the improvement of photocatalytic process.

α-Fe2O3/GO thin films, on the other hand, showed a lower removal efficiency compared to nanocomposites. The best results were for α-Fe2O3/GO 5 % with just over 47 % dye removal. The main reason for this is that the content of nanocomposites involved with dye molecules is limited and therefore lower percentage of dyes is eliminated by the photocatalytic process. Pure Fe2O3 showed minimum efficiency compared to other thin film composites. Fe2O3-GO 8% also indicated acceptable performance with 39.1% removal efficiency after 120 min irradiation. The results are shown in Fig. 11. In the photodegradation process, the recombination of photo generated electron-hole pairs has to be reduced. Here, the mixture of Fe2O3 with GO significantly prevents the rise of recombination rate on the surface of thin film. In fact, reduced GO acts as an acceptor in α-Fe2O3/GO nanocomposite layer and consequently decreases the charge recombination in the photocatalytic process. Thus, the presence of GO makes a great impact on the removal efficiency of organic dye.
In a common photocatalytic degradation system, photocatalyst releases a significant amount of electron-hole pairs in order to join in the oxidation and reduction process for degradation. Electrons react with oxygen to produce oxy radicals while holes contribute in oxidation process with hydroxyl ions to create hydroxyl radicals which decompose pollutants in degradation process. In this study, degradation process is different from common system which was mentioned above. Here, GO accepts photo generated electrons in their structures which significantly helps the suppression of recombination rate in the system. In this case, if GO receives enough electrons, they will share these electrons with free oxygens in the solution and therefore produce oxy radical to decompose pollutants in the system. The probable mechanism is that the excited electrons transfer from conduction band to valence band and will eventually become trapped on the surface of GO and fail to return to conduction band. On the other hand, the generated holes in conduction band of hematite produce enough hydroxyl radicals on the surface of layer to react with rhodamine molecules and finally turn them into harmless materials such as H2O. Fig. 12 shows the mechanism of α-Fe2O3/GO thin layer under the irradiation of xenon lamps.
Although thin film α-Fe2O3/GO showed a great performance in RhB removal, the powder α-Fe2O3/GO had more removal efficiency. It seems using powder nanocomposites in the slurry system provided more photo generated electron-hole pairs which participated in degradation reactions to decompose RhB. Despite higher recombination rate, the powder α-Fe2O3/GO contacted more with cationic dye. GO molecules covering the surface of hematite absorbed more dye molecules because they were freely in contact with all molecules in the solution. As a result, the produced electrons and holes on the surface of α-Fe2O3/GO powders decompose dye molecules more effectively. Looking at photocatalytic mechanism of powder α-Fe2O3/GO, the change of conversion and recycling between Fe(III) and Fe(II) because of free electrons caused powder nanocomposites to be more reactive. Moreover, the electrostatic interaction and π-π stacking of α-Fe2O3/GO caused α-Fe2O3/GO molecules to attract a huge number of dye molecules [40]. Consequently, the high rate of decolorization occurred as xenon lights hit the surface of nanocomposites. On the other hand, α-Fe2O3/GO also showed an acceptable removal efficiency in the photoelectrochemical reactor. The combination of reduced GO with hematite not only increased the stability of thin films for reusability process but also reduced recombination rate of generated electron-hole pairs resulting in better degradation process. However, this is an immobilized system and therefore lower amount of α-Fe2O3/GO molecules are in contact with dye molecules compared to α-Fe2O3/GO powder nanocomposites. This may lead to a considerable reduction in photodegradation efficiency for thin film layers. However, the use of both photoelectrochemical system and the presence of reduced GO prevents electron recombination, which makes it comparable with α-Fe2O3/GO in the slurry system in terms of photodegradation efficiency.
3.6 The reusability process
The recovery of nanocomposites with an external magnetic field was done from solution and they were tested again to measure dye removal in the next cycles. The thin-film composites were also tested in more cycles. Fig. 13 shows the dye removal efficiency of each cycle for both α-Fe2O3/GO powders and thin films. α-Fe2O3/GO thin films showed acceptable performance in 7 cycles with more than 46 % removal efficiency. After 4 cycles, the decline in dye removal was noticeable. A 9 % decline in thin-film removal efficiency was observed due to less surface area on the surface of films to be excited and degrade dyes. In fact, dye solutions filled the space on the surface of films, leading to the reduction of photocatalytic activity. In powder nanocomposites reusability process, the α-Fe2O3/GO nanocomposites showed higher removal efficiency in more cycles compared to thin films. In fact, synthesized nanocomposites removed up to 63 % of dyes in the solution in 6 cycles. This is attributed to the strong intermolecular structure of GO and hematite molecules. This strong structure increased the stability of nanocomposite and therefore removal efficiency remained high after 6 cycles. The hematite was well immobilized on the surface of GO which enhanced the surface area of nanocomposite to degrade more dye molecules. Also, the presence of GO helped the nanocomposite to adsorb more light for decomposition, which reinforce powder nanocomposites to remove dye molecules in 6 cycles with a high efficiency [42]. However, the change in removal efficiency was seen in the 7th cycle, with 7 % decline in the dye removal performance.
3.7 Conclusion
The photocatalytic ability of thin film α-Fe2O3-GO was compared to powder α-Fe2O3-GO nanocomposites in different concentrations. Based on the results, powder nanocomposites had better performance with above 64 % RhB removal efficiency from textile wastewater. The reusability process also worked better in the presence of powder nanocomposites which significantly removed dye with maximum removal efficiency in six cycles. Due to higher surface area, intermolecular connection of dyes and nanocomposite and absorption of more light, the powder nanocomposite showed better performance in photocatalytic activity under the irradiation of visible xenon lamps. On the other hand, thin film α-Fe2O3-GO displayed lower photocatalytic reaction because of less generated electron-hole pairs on the surface of layers. The characterization of both synthesized nanocomposites clearly supported the combination of hematite GO as a powder or as a thin film layer. The results show the great potential of the developed nanocomposites for high efficiency pohotocatalytic degradation of RhB and warrant further investigations into the application of these structures for the removal of other dyes and contaminants from wastewater.