Synthesis of MRGO Nanocomposites as a Potential Photocatalytic Demulsiﬁer for Crude Oil-in-Water Emulsion

: Oil-in-water (O/W) emulsion has been a major concern for the petroleum industry. A cost-effective magnetite-reduced graphene oxide (MRGO) nanocomposite was synthesized to study the demulsiﬁcation process of emulsion using said nanocomposite under solar illumination. Characterization data show that the magnetite was successfully deposited on reduced graphene oxide through redox reaction at varying loading amounts of magnetite. Demulsiﬁcation of the O/W emulsion using MRGO nanocomposite shows that in general the demulsiﬁcation efﬁciency was dependent on the loading amount of Fe 3 O 4 on the RGO sheet. It was proposed that the surfactant hydroxyl groups have an afﬁnity towards Fe 3 O 4 , which the loading amount was directly proportionate to available active site in Fe 3 O 4 . As the loading amount increases, charge recombination centers on the RGO sheet would increase, effectively affecting the charge distribution within MRGO structure.


Introduction
The oil and gas industry is one of the sectors with the heaviest industrialized resources in the world and the petroleum industry has long been faced with multitude of problems to improve the profitability and production. One of the issues is that the oily wastewater in the industries contains emulsion of water and oil liquids which was formed during the crude oil extraction process. Emulsions are classified into two classes, and the one focused in this work is oil-in-water (O/W) emulsion. O/W emulsion contains oil droplets encapsulated by the water matrix or dispersed water or surfactant forming an interfacial film at the oil and water interface which emulsifies the two immiscible liquids [1]. Emulsions have been a major concern in the petroleum industry as it affects the production and the pipeline system in the production [2]. Hence, in this work, a magnetite-carbon-based nanocomposite is studied for its efficiency and efficacy in demulsifying O/W emulsion.
Graphene, a revered material for its adsorption properties and in-plane conductivity, has shown excellent results in photocatalytic effect when composited with other materials [3]. Graphene oxide (GO) sheets has a π-π stacking within its sp 2 hybridized network and due to their large surface area and broad sp 2 network, they are promising material for organic compounds photodegradation [4] and demulsification [5]. However, separation of the material and recyclability has been an issue for graphene-based nanocomposites. To solve the separation issue, magnetite was suggested as Fe 3 O 4 has shown excellent photocatalytic properties in organic compounds degradation [6] and the recyclability of the material has been promising [7]. Some previous studies [8][9][10] have shown that magnetite

Synthesis of GO
GO was prepared and the magnetite-reduced graphene oxide was synthesized using an in situ chemical synthesis method [20]. An amount of 3 g of graphite flakes was oxidated with 400 mL of H 2 SO 4 , 40 mL of H 3 PO 4 and 18 g of KMnO4 by using simplified Hummers' method [21]. The chemicals were mixed by using a magnetic stirrer and stirred for 5 min. However, the mixture was stirred for 3 days to ensure complete oxidation of graphite. During the oxidation, the color of mixture changed from dark purplish green to dark brown. H 2 O 2 solution was added into the mixture to stop the oxidation process and the color of mixture changed to bright yellow, indicating a high degree of oxidation of graphite. The mixture solution was washed with 1 M of HCl aqueous solution and deionized water alternatively by using centrifugation techniques. The graphite oxide solids were obtained and dried.
The synthesis route for all MRGO nanocomposites was strictly maintained to be the same and only deionized water was used during the whole synthesis and experimental procedure.
Firstly, 0.5 g of graphene oxide was added into 200 mL of deionized water, and the GO aqueous suspension was ultrasonicated for 10 minutes or until all GO was finely dispersed in the aqueous suspension. Once done, the GO suspension was placed onto a magnetic stirrer and the pH of the GO suspension was adjusted to about 11-12 using 25% ammonia solution added dropwise. Secondly, for MRGO-5, 2.5 g of FeCl 2 was dissolved in 100 mL of deionized water and added dropwise into the GO suspension. For MRGO-10 and MRGO-20, 5.0 g and 10.0 g of FeCl 2 were used, respectively.
The mixture solutions were then left for 12 h of stirring to ensure proper deposition of magnetite on GO to form MRGO nanocomposites. After 12 h, the black precipitation of the mixture solutions was extracted and washed with deionized water using a centrifuge. The centrifuge setting was 7000 rotations per minute with the rotation period of 10 minutes. The washing procedure was repeated for 3 times to wash out the extra ammonia present in the solution. The slurry was dried at 40 • C for 12 h, and the dried precipitation was grinded to obtain the powdered form of MRGO nanocomposites.
The entire process is repeated without GO to synthesize magnetite only through redox reaction in alkaline solution.

Preparation O/W Emulsion
The O/W emulsion is prepared by mixing 10 wt% of surfactant Tween 60, 10 wt% of Tapis crude oil and 80 wt% of deionized water. The physical and chemical properties of Tapis crude oil are tabulated in Tables 1 and 2, whereas Figure 1 shows the chemical formula of surfactant Tween 60. The mixed solution is then placed in a high-speed mixer for 10 utes to form the O/W emulsion.  The synthesis route for all MRGO nanocomposites was strictly maintained to be the same and only deionized water was used during the whole synthesis and experimental procedure.
Firstly, 0.5 g of graphene oxide was added into 200 mL of deionized water, and the GO aqueous suspension was ultrasonicated for 10 minutes or until all GO was finely dispersed in the aqueous suspension. Once done, the GO suspension was placed onto a magnetic stirrer and the pH of the GO suspension was adjusted to about 11-12 using 25% ammonia solution added dropwise. Secondly, for MRGO-5, 2.5 g of FeCl2 was dissolved in 100 mL of deionized water and added dropwise into the GO suspension. For MRGO-10 and MRGO-20, 5.0 g and 10.0 g of FeCl2 were used, respectively.
The mixture solutions were then left for 12 h of stirring to ensure proper deposition of magnetite on GO to form MRGO nanocomposites. After 12 h, the black precipitation of the mixture solutions was extracted and washed with deionized water using a centrifuge. The centrifuge setting was 7000 rotations per minute with the rotation period of 10 minutes. The washing procedure was repeated for 3 times to wash out the extra ammonia present in the solution. The slurry was dried at 40 °C for 12 h, and the dried precipitation was grinded to obtain the powdered form of MRGO nanocomposites.
The entire process is repeated without GO to synthesize magnetite only through redox reaction in alkaline solution.

Preparation O/W emulsion
The O/W emulsion is prepared by mixing 10 wt% of surfactant Tween 60, 10 wt% of Tapis crude oil and 80 wt% of deionized water. The physical and chemical properties of Tapis crude oil are tabulated in Tables 1 and 2, whereas Figure 1 shows the chemical formula of surfactant Tween 60. The mixed solution is then placed in a high-speed mixer for 10 utes to form the O/W emulsion.

Characterization of MRGOs
An X-ray diffractometer (XRD Brucker D8 advance diffractometer) was used to verify the complete synthesis of GO and MRGOs with Cu Kα radiation (λ = 1.5406 A) in the range of 2θ = 5 • -80 • with a step size 0.01 • /step and exposure time 200 s/step. Then, Fieldemission Scanning Electron Microscope (ZEISS Leo Supra 55, FESEM) was used to study the surface morphology of MRGO and verified the attachment of magnetite on the RGO sheets with an accelerating voltage of 5 kV. X-ray Photoelectron Spectroscopy (Thermo Scientific K-Alpha instrument, XPS) was used to study the composition information of GO and MRGOs. Measurements were used Al K α (hv = 1468.6 eV) as the radiation source. Further on, Fourier-transform Infra-red Spectrometer (Perkin Elmer Spectrum BX, FTIR) was used to determine the surface bonding of the synthesized GO and MRGOs with the wavelength range of 400-4000 cm −1 . Raman spectroscopy (Horiba Jabin Yvon HR800, RAMAN) was used to characterize the structural properties of synthesized MRGOs with 514.5 nm laser. Lastly, a vibrating sample magnetometer (DMS Model 10, VSM) was used to evaluate the magnetic properties of the synthesized MRGOs.

Demulsification Test
The test samples were prepared by mixing 1 mL of prepared O/W emulsion and 0.05 g of MRGOs into 19.4 mL of deionized water filled in a capped bottle. A standard solution without MRGO was prepared as well. The prepared test bottle samples were sonicated at a high intensity for 10 mins; where the MRGOs settled down at bottom of the bottle, the bottle should be swirled by hand to ensure proper distribution and mixing of the materials. After sonication, the bottles were mounted onto a vertical bottle shaker with 500 oscillation/min, under solar lamp illumination with varying intensities (0, 200, 400, 600, 800 and 1000 W/m 2 ) for 30 mins. The setup was as shown in Figure 2. Then, the bottles were removed from the shaker and placed on magnets for rapid settling of the MRGOs. The test samples were then left overnight on magnets for the mixture to stabilize at ambient conditions. J. Compos. Sci. 2021, 5, x FOR PEER REVIEW 4 of 13

Characterization of MRGOs
An X-ray diffractometer (XRD Brucker D8 advance diffractometer) was used to verify the complete synthesis of GO and MRGOs with Cu Kα radiation (λ = 1.5406 A) in the range of 2θ = 5°-80° with a step size 0.01°/step and exposure time 200 s/step. Then, Fieldemission Scanning Electron Microscope (ZEISS Leo Supra 55, FESEM) was used to study the surface morphology of MRGO and verified the attachment of magnetite on the RGO sheets with an accelerating voltage of 5 kV. X-ray Photoelectron Spectroscopy (Thermo Scientific K-Alpha instrument, XPS) was used to study the composition information of GO and MRGOs. Measurements were used Al Kα (hv = 1468.6 eV) as the radiation source. Further on, Fourier-transform Infra-red Spectrometer (Perkin Elmer Spectrum BX, FTIR) was used to determine the surface bonding of the synthesized GO and MRGOs with the wavelength range of 400-4000 cm −1 . Raman spectroscopy (Horiba Jabin Yvon HR800, RAMAN) was used to characterize the structural properties of synthesized MRGOs with 514.5 nm laser. Lastly, a vibrating sample magnetometer (DMS Model 10, VSM) was used to evaluate the magnetic properties of the synthesized MRGOs.

Demulsification test
The test samples were prepared by mixing 1 mL of prepared O/W emulsion and 0.05 g of MRGOs into 19.4 mL of deionized water filled in a capped bottle. A standard solution without MRGO was prepared as well. The prepared test bottle samples were sonicated at a high intensity for 10 mins; where the MRGOs settled down at bottom of the bottle, the bottle should be swirled by hand to ensure proper distribution and mixing of the materials. After sonication, the bottles were mounted onto a vertical bottle shaker with 500 oscillation/min, under solar lamp illumination with varying intensities (0, 200, 400, 600, 800 and 1000 W/m 2 ) for 30 mins. The setup was as shown in Figure 2. Then, the bottles were removed from the shaker and placed on magnets for rapid settling of the MRGOs. The test samples were then left overnight on magnets for the mixture to stabilize at ambient conditions. The demulsification capability of MRGOs was determined by the residual oil content in the extracted water above the settled MRGO in the test samples. The oil concentration in standard and demulsified samples were obtained and compared from the standard curve of UV-visible absorption spectra by measuring the absorbance intensity at 228 nm. The demulsification efficiency under varying solar intensities was calculated by the following equation: where Co is the initial concentration of oil content in the test sample and Ci represents the final concentration of oil content in the test sample. The demulsification capability of MRGOs was determined by the residual oil content in the extracted water above the settled MRGO in the test samples. The oil concentration in standard and demulsified samples were obtained and compared from the standard curve of UV-visible absorption spectra by measuring the absorbance intensity at 228 nm. The demulsification efficiency under varying solar intensities was calculated by the following equation: where C o is the initial concentration of oil content in the test sample and C i represents the final concentration of oil content in the test sample.  [22]. Additionally, peak intensities of Fe 3 O 4 were observed to strengthen with the increase in the loading ratio of Fe 3 O 4 on GO sheets, suggesting the successful deposition of Fe 3 O 4 on RGO sheet through redox reaction at varying ratios. At high GO loading ratio, GO may hinder the crystallization of Fe 3 O 4 , while at a high Fe 3 O 4 loading ratio, the XRD patterns are similar to only nanoparticles, as seen in MRGO-20. Figure 3 displays the XRD patterns of GO, MRGO-5, MRGO-10 and MRGO-20; it shows the diffraction peak profiles of GO and MrGO from 5° to 75°. The characteristic peak of GO was observed at 10.3° corresponding to (001). However, for the MRGO nanocomposites, there are six main characteristic peaks were observed at 20°, 30.2°, 43.4°, 53.7° and 62.8°, which are corresponding to diffraction indices of (220), (311), (400), (422), (511) and (440) crystal planes of Fe3O4 [22]. Additionally, peak intensities of Fe3O4 were observed to strengthen with the increase in the loading ratio of Fe3O4 on GO sheets, suggesting the successful deposition of Fe3O4 on RGO sheet through redox reaction at varying ratios. At high GO loading ratio, GO may hinder the crystallization of Fe3O4, while at a high Fe3O4 loading ratio, the XRD patterns are similar to only nanoparticles, as seen in MRGO-20.   Figure 4a, GO sheets were associated with a rough and irregular wrinkled structure. This might be due to the oxidation of graphite where different oxygen functional groups were being introduced on the surface using the improved Hammer method. Figure 4b-d illustrate that all the Fe3O4 deposited on RGO sheets was spherical in shape and well dispersed across the sheets. By introducing Fe3O4 nanoparticles on the RGO sheets, the restacking of GO sheets was prevented [23] and the layering of RGO sheets can be well observed in Figure 4b-d. While Fe3O4 tends to aggregate, and the tendency would strengthen with increasing loading ratio [24], the deposition on RGO sheets significantly reduced the conglomeration of Fe3O4 nanoparticles in all MRGOs samples.   Figure 4a, GO sheets were associated with a rough and irregular wrinkled structure. This might be due to the oxidation of graphite where different oxygen functional groups were being introduced on the surface using the improved Hammer method. Figure 4b-d illustrate that all the Fe 3 O 4 deposited on RGO sheets was spherical in shape and well dispersed across the sheets. By introducing Fe 3 O 4 nanoparticles on the RGO sheets, the restacking of GO sheets was prevented [23] and the layering of RGO sheets can be well observed in Figure 4b-d. While Fe 3 O 4 tends to aggregate, and the tendency would strengthen with increasing loading ratio [24], the deposition on RGO sheets significantly reduced the conglomeration of Fe 3 O 4 nanoparticles in all MRGOs samples. Figure 3 displays the XRD patterns of GO, MRGO-5, MRGO-10 and MRGO-20; it shows the diffraction peak profiles of GO and MrGO from 5° to 75°. The characteristic peak of GO was observed at 10.3° corresponding to (001). However, for the MRGO nanocomposites, there are six main characteristic peaks were observed at 20°, 30.2°, 43.4°, 53.7° and 62.8°, which are corresponding to diffraction indices of (220), (311), (400), (422), (511) and (440) crystal planes of Fe3O4 [22]. Additionally, peak intensities of Fe3O4 were observed to strengthen with the increase in the loading ratio of Fe3O4 on GO sheets, suggesting the successful deposition of Fe3O4 on RGO sheet through redox reaction at varying ratios. At high GO loading ratio, GO may hinder the crystallization of Fe3O4, while at a high Fe3O4 loading ratio, the XRD patterns are similar to only nanoparticles, as seen in MRGO-20.   Figure 4a, GO sheets were associated with a rough and irregular wrinkled structure. This might be due to the oxidation of graphite where different oxygen functional groups were being introduced on the surface using the improved Hammer method. Figure 4b-d illustrate that all the Fe3O4 deposited on RGO sheets was spherical in shape and well dispersed across the sheets. By introducing Fe3O4 nanoparticles on the RGO sheets, the restacking of GO sheets was prevented [23] and the layering of RGO sheets can be well observed in Figure 4b-d. While Fe3O4 tends to aggregate, and the tendency would strengthen with increasing loading ratio [24], the deposition on RGO sheets significantly reduced the conglomeration of Fe3O4 nanoparticles in all MRGOs samples.  Moreover, as recorded in Table 3, it was observed that as the loading ratio of Fe3O4 increases, the weight percentage of Fe increases proportionately as well. The atomic ratio of Fe to C for MRGO-5, MRGO-10 and MRGO-20 was 0.37, 0.71 and 1.27, respectively, increasing in an approximately two-fold manner. As the Fe content is more than C in MRGO-20, and as observed in Figure 4c, the synthesized MRGO was similar to magnetite in appearance which agreed with the XRD results.   [25]. As observed in Figure 5 and recorded in Table 4, the ID/IG ratio increased from 0.9625 (GO) to 1.0193 (MRGO-20) after the in situ chemical synthesis. The result indicated the presence of sp 3 reduces within the sp 2 carbon network upon the reduction of GO [26], thereby confirming the formation of RGO sheets in MRGO nanocomposites. It was suggested that the D-band is in fact slightly higher than the G band signal for all MRGOs. Deposition of magnetite through reduction of GO has altered the structure of GO, resulting in higher number of sp 3 defects in the RGO sheets [27]. Furthermore, a magnetite signature was observed around 200-450 cm −1 for MRGO-20 [28], specifically at 217 cm −1 , 286 cm −1 and 393 cm −1 , at which the signal at 393 cm −1 was due to [OH] groups linked to Fe 3+ on the magnetite surface [29].

Raman Analysis
The result strongly suggested the successful deposition of magnetite through redox reaction on the RGO sheets. Moreover, as recorded in Table 3, it was observed that as the loading ratio of Fe 3 O 4 increases, the weight percentage of Fe increases proportionately as well. The atomic ratio of Fe to C for MRGO-5, MRGO-10 and MRGO-20 was 0.37, 0.71 and 1.27, respectively, increasing in an approximately two-fold manner. As the Fe content is more than C in MRGO-20, and as observed in Figure 4c, the synthesized MRGO was similar to magnetite in appearance which agreed with the XRD results.   [25]. As observed in Figure 5 and recorded in Table 4, the I D /I G ratio increased from 0.9625 (GO) to 1.0193 (MRGO-20) after the in situ chemical synthesis. The result indicated the presence of sp 3 reduces within the sp 2 carbon network upon the reduction of GO [26], thereby confirming the formation of RGO sheets in MRGO nanocomposites. It was suggested that the D-band is in fact slightly higher than the G band signal for all MRGOs. Deposition of magnetite through reduction of GO has altered the structure of GO, resulting in higher number of sp 3 defects in the RGO sheets [27]. Furthermore, a magnetite signature was observed around 200-450 cm −1 for MRGO-20 [28], specifically at 217 cm −1 , 286 cm −1 and 393 cm −1 , at which the signal at 393 cm −1 was due to [OH] groups linked to Fe 3+ on the magnetite surface [29].

Raman Analysis
The result strongly suggested the successful deposition of magnetite through redox reaction on the RGO sheets.   Figure 6 shows the FTIR spectra for GO, MRGO-5, MRGO-10 and MRGO-20. The FTIR spectrum of GO had five characteristic peaks, namely, OH out-of-plane bending at 560 cm −1 , CO-O-CO at 1074 cm −1 , epoxy C-O at 1218 cm −1 , aromatic C=C stretching at 1577 cm −1 and hydroxyl (-OH) at 3400 cm −1 . The presences of oxygenated functional groups at GO indicated that the graphite underwent oxidation. For MRGO composites, absorption spectrum observed at 3400 cm −1 was attributed to the O-H stretching vibration of C-OH group in RGO, while the spectrum at 1577 cm −1 was related to the C-OH bending in the RGO sheets [30]. Moreover, a peak was observed at 1220 cm −1 suggesting the C-O stretching vibration of epoxy groups or rings in the RGO sheets, and the peak at 618 cm −1 was the in-plane deformation vibration of aromatic aldehyde [31]. Notice that the 618 cm −1 peak became more significant as the loading ratio of Fe3O4 increases, suggesting the deposition of magnetite disrupted the structure of the RGO plane. Lastly, the adsorption peak at 560 cm −1 was depicted as the Fe-O vibrations in Fe3O4 [32], which increases in absorption intensity as the loading ratio of magnetite increases. This confirmed the interaction between magnetite and RGO sheets. Table 5 shows the comparison of functional groups between GO and MRGOs.    Figure 6 shows the FTIR spectra for GO, MRGO-5, MRGO-10 and MRGO-20. The FTIR spectrum of GO had five characteristic peaks, namely, OH out-of-plane bending at 560 cm −1 , CO-O-CO at 1074 cm −1 , epoxy C-O at 1218 cm −1 , aromatic C=C stretching at 1577 cm −1 and hydroxyl (-OH) at 3400 cm −1 . The presences of oxygenated functional groups at GO indicated that the graphite underwent oxidation. For MRGO composites, absorption spectrum observed at 3400 cm −1 was attributed to the O-H stretching vibration of C-OH group in RGO, while the spectrum at 1577 cm −1 was related to the C-OH bending in the RGO sheets [30]. Moreover, a peak was observed at 1220 cm −1 suggesting the C-O stretching vibration of epoxy groups or rings in the RGO sheets, and the peak at 618 cm −1 was the in-plane deformation vibration of aromatic aldehyde [31]. Notice that the 618 cm −1 peak became more significant as the loading ratio of Fe 3 O 4 increases, suggesting the deposition of magnetite disrupted the structure of the RGO plane. Lastly, the adsorption peak at 560 cm −1 was depicted as the Fe-O vibrations in Fe 3 O 4 [32], which increases in absorption intensity as the loading ratio of magnetite increases. This confirmed the interaction between magnetite and RGO sheets. Table 5 shows the comparison of functional groups between GO and MRGOs.   Figure 6 shows the FTIR spectra for GO, MRGO-5, MRGO-10 and MRGO-20. The FTIR spectrum of GO had five characteristic peaks, namely, OH out-of-plane bending at 560 cm −1 , CO-O-CO at 1074 cm −1 , epoxy C-O at 1218 cm −1 , aromatic C=C stretching at 1577 cm −1 and hydroxyl (-OH) at 3400 cm −1 . The presences of oxygenated functional groups at GO indicated that the graphite underwent oxidation. For MRGO composites, absorption spectrum observed at 3400 cm −1 was attributed to the O-H stretching vibration of C-OH group in RGO, while the spectrum at 1577 cm −1 was related to the C-OH bending in the RGO sheets [30]. Moreover, a peak was observed at 1220 cm −1 suggesting the C-O stretching vibration of epoxy groups or rings in the RGO sheets, and the peak at 618 cm −1 was the in-plane deformation vibration of aromatic aldehyde [31]. Notice that the 618 cm −1 peak became more significant as the loading ratio of Fe3O4 increases, suggesting the deposition of magnetite disrupted the structure of the RGO plane. Lastly, the adsorption peak at 560 cm −1 was depicted as the Fe-O vibrations in Fe3O4 [32], which increases in absorption intensity as the loading ratio of magnetite increases. This confirmed the interaction between magnetite and RGO sheets. Table 5 shows the comparison of functional groups between GO and MRGOs.    Figure 7a shows the wide scan XPS survey of MRGOs where C1s and O1s peaks at binding energies of 285 eV and 530 eV, respectively, were exhibited. As for the peak characteristic of Fe 3 O 4 , Fe2p 3/2 and Fe2p 1/2 peaks were observed at 711 eV and 725 eV, respectively [33]. In Figure 7c,d, deconvolution peaks of C1s show C-C (aromatic) bonds, C-O bonds, C=O bonds and π-π* satellite bonds signal at 285 eV, 287 eV, 289 eV and 291 eV, respectively [34,35]. A signal of π-π* satellite bonds at 291 eV indicates the successful reduction of GO sheets to RGO sheets, as RGO has more sp 2 hybridized zone due to the restoration from the reduction process [35]. The intensity of C1s peak decreases as the loading ratio of Fe 3 O 4 increases and is deposited on the RGO sheets. Meanwhile, in Figure  7e,f, deconvolution peaks of O1s indicated Fe-O bonds, C=O bonds and C-O bonds at 530 eV, 531 eV and 533 eV, respectively. The intensity of Fe-O peak at 530 eV was observed to increase significantly for MRGO-20 as more Fe 3 O 4 was deposited. Additionally, in both C1s and O1s deconvoluted peaks, C-O peaks were observed to decrease significantly in MRGO-20, which implied an increase in conductivity within the RGO sheets as more sp 2 bonds were restored, thereby enhancing its charge conductivity [36].    Figure 8 shows the VSM magnetization characteristic of MRGO-5 and MRGO-20. MRGO-5 and MRGO-20 exhibit magnetic saturation of 40.7 emu/g and 63.5 emu/g, respectively, inferring the susceptibility to magnetic field is proportionate to the concentration of Fe3O4 deposited on the RGO sheets. Further confirming the successful deposition of magnetite on GO sheets through redox reaction at varying ratios. Moreover, the very low coercivity value of the samples indicates that the MRGOs show soft magnetic material features, exhibiting superparamagnetic response and well isolated single domain particle properties [37]. Both have a significant degree of magnetization to allow them to be separated from the emulsion by a magnet after the demulsification process.

Demulsification Mechanism
Effective demulsification was proposed to be the disruption and breaking up of the interfacial film of Tween 60 between the oil and water interface. The Tween 60 structure has more hydroxyl functional groups, which have adsorptive affinity on Fe3O4 nanoparticles [7]. Hence, the demulsification efficiency of MRGOs was seen to increase with the loading ratio of magnetite on RGO sheet when there is no solar irradiance, indicating that Fe3O4 nanoparticles played an important role in demulsifying the prepared emulsion.

Demulsification Mechanism
Effective demulsification was proposed to be the disruption and breaking up of the interfacial film of Tween 60 between the oil and water interface. The Tween 60 structure has more hydroxyl functional groups, which have adsorptive affinity on Fe 3 O 4 nanoparticles [7]. Hence, the demulsification efficiency of MRGOs was seen to increase with the loading ratio of magnetite on RGO sheet when there is no solar irradiance, indicating that Fe 3 O 4 nanoparticles played an important role in demulsifying the prepared emulsion.
In Figure 9a, the demulsification efficiency of pure magnetite nanoparticles was about the same across the varying solar intensities, at 42-44% efficiency with a slight increment in line with the intensity. Pure Fe 3 O 4 nanoparticles do not react with visible light wavelength range, hence it is postulated that the slight increment may be due to the following two reasons. (a) Photothermic properties of Fe 3 O 4 due to visible light-near infrared spectrum range, which involves micro-heating at the Fe 3 O 4 active sites [38]; (b) Fenton-like mechanism due to visible light-ultraviolet spectrum range, which involves the production of oxygen peroxide radicals, O 2 and hydroxyl radicals, OH, from the photocatalytic effect of MRGOs under the influence of solar irradiance [39,40], of which both radicals have high oxidative potential on organic compounds and enhance the demulsification efficiency.
hydroxyl group to produce hydroxyl radicals which would further disrupt the oil-water interfacial film of Tween 60 through oxidative reaction [26].
Meanwhile, in Figure 9c, it was expected that the increment in Fe3O4 loading amount would improve the demulsification efficiency by increasing the radicals' production [41], but there was an optimum loading amount of Fe3O4 for the effective photocatalytic effect to occur [6]. While the adsorption capability of MRGO-10 has shown significant improvement, excess content of Fe3O4 introduced new charge recombination at Fe3O4, thus reducing the availability of charges within MRGOs [24,[42]. That effectively disrupted the affinity of Fe3O4 toward hydroxyl group of Tween 60, decreasing the demulsification efficiency. Figure 9d shows the demulsification efficiency of MRGO-20, at which it was not reactive toward solar irradiance; as magnetite excessively covered most of the RGO sheets, the MRGO nanocomposites behaved similarly to pure magnetite nanoparticles. However, due to the synergy with RGO which prevented the aggregation of Fe3O4 nanoparticles, it is suggested that there are more active sites of Fe3O4 available for adsorption as compared to pure Fe3O4 nanoparticles only. Hence, an enhanced demulsification capability independent of the presence of solar irradiance was observed.

Conclusions
Magnetite-reduced graphene oxide nanocomposites were successfully synthesized as verified in the characterization of MRGOs. In the demulsification test, the demulsification efficiency of MRGOs was observed to increase proportionately with the loading amount of Fe3O4 on RGO sheets. A positive photocatalytic effect under solar irradiance was observed only with a low loading amount of Fe3O4 on RGO sheets, such as in MRGO-5. Hence, it was concluded that while the demulsification efficiency was dependent on the loading amount of Fe3O4 on RGO sheets, only a low and optimal loading amount of Fe3O4 would show a positive photocatalytic effect.   In Figure 9b, the demulsification efficiency shows an increasing trend with the solar intensity. While the adsorptive capability of MRGO-5 was weak due to a low Fe 3 O 4 loading amount, the MRGO nanocomposites show a photocatalytic effect in the demulsification process. The synergy of Fe 3 O 4 with RGO sheets augmented the absorption wavelength range of the overall material and facilitated the photocatalytic properties of MRGOs [26]. At the optimum loading ratio of Fe 3 O 4 , magnetite reacted with the adsorbed hydroxyl group to produce hydroxyl radicals which would further disrupt the oil-water interfacial film of Tween 60 through oxidative reaction [26].
Meanwhile, in Figure 9c, it was expected that the increment in Fe 3 O 4 loading amount would improve the demulsification efficiency by increasing the radicals' production [41], but there was an optimum loading amount of Fe 3 O 4 for the effective photocatalytic effect to occur [6]. While the adsorption capability of MRGO-10 has shown significant improvement, excess content of Fe 3 O 4 introduced new charge recombination at Fe 3 O 4 , thus reducing the availability of charges within MRGOs [24,42]. That effectively disrupted the affinity of Fe 3 O 4 toward hydroxyl group of Tween 60, decreasing the demulsification efficiency. Figure 9d shows the demulsification efficiency of MRGO-20, at which it was not reactive toward solar irradiance; as magnetite excessively covered most of the RGO sheets, the MRGO nanocomposites behaved similarly to pure magnetite nanoparticles. However, due to the synergy with RGO which prevented the aggregation of Fe 3 O 4 nanoparticles, it is suggested that there are more active sites of Fe 3 O 4 available for adsorption as compared to pure Fe 3 O 4 nanoparticles only. Hence, an enhanced demulsification capability independent of the presence of solar irradiance was observed.

Conclusions
Magnetite-reduced graphene oxide nanocomposites were successfully synthesized as verified in the characterization of MRGOs. In the demulsification test, the demulsification efficiency of MRGOs was observed to increase proportionately with the loading amount of

Conflicts of Interest:
The authors declare no conflict of interest.