Desulfurization Performance of Choline Chloride-Based Deep Eutectic Solvents in the Presence of Graphene Oxide

Extractive catalytic oxidative desulfurization (ECODS) is the one of the recent methods used in fuel desulfurization which involved the use of catalyst in the oxidative desulfurization of diesel fuel. This study is aimed to test the effectiveness of synthesized choline chloride (ChCl) based deep eutectic solvent (DES) in fuel desulfurization via ECODS method, with the presence of graphene oxide (GO) as catalyst and hydrogen peroxide (H2O2) as oxidant. In this study, 16 DESs based on choline chloride were synthesized using glycerol (GLY), ethylene glycol (EG), tetraethylene glycol (TEG) and polyethylene glycol (PEG). The characterization of the synthesized DES was carried out via Fourier transform infrared spectroscopy (FTIR) analysis, density, and viscosity determination. According to the screening result, ChCl-PEG (1:4) was found to be the most effective DES for desulfurization using ECODS method, with a removal of up to 47.4% of sulfur containing compounds in model oil in just 10 min per cycle after the optimization of the reaction parameters, and up to 95% desulfurization efficiency could be achieved by six cycles of desulfurization. It is found that the addition of GO as catalyst does not increase the desulfurization performance drastically; hence, future studies for the desulfurization performance of DESs made up from ChCl and PEG and its derivatives can be done simply by using extraction desulfurization (EDS) method instead of ECODS method, for cost reduction purpose and easier regulation of DES waste into environment.


Introduction
Sulfur containing compounds (SCC) are usually present in diesel fuel, which is commonly used in heavy type of vehicle or machines as a source of energy. There are a few examples of SCC in diesel fuel such as thiols, sulfides, disulfides, thiophenes, benzothiphenes (BT), and dibenzothiophenes (DBT) [1]. The combustion of diesel fuels leads to the formation of sulfur dioxide (SO 2 ) and its derivatives which are then released into the atmosphere. SO 2 is known for causing air pollution, acid rain and irritation to human skin, eyes, nose, throat, and lungs. Previous research reports that the deposition of acidic sulfur (from acid rain) into the forest soil can cause the releasing of methane gas, which is one of the greenhouse gases into the atmosphere [2]. The remnant of sulfur in diesel fuel also reduces the effectiveness of catalyst used in the emission control system which is crucial in oxidation of the carbon monoxide and hydrocarbon into relatively harmless carbon dioxide before emitting the gases into the current fuel desulfurization research. Firstly, the cost of synthesizing DES is much lower than IL, as the starting materials for DES are relatively cheaper and widely available in the laboratory [22]. Some of the IL involves the use of organic solvents as one of their starting material, and environmental problem might arise if the IL waste is not properly managed, as compared to DES, which are considered biodegradable [23,24]. DES has the ability to function as a "designer solvent", where researchers can tailor make the DES according to the requirement of the processes involved in their study [25]. DES are also immiscible in non-polar solvents typically diesel fuel making the regeneration of DES easier which is somehow favorable in petroleum refinery [26].
Generally, there are four types of DES, namely type I, type II, type III, and type IV. Type I to type III DES solvents are made up by mixing quaternary salts and metal halide (Type I), hydrated metal halide (Type II) and HBD (Type III). Type IV DES involves the mixing between metal halide and HBD. In this study, type III DES solvent is utilized, thus we will use the quaternary salt as HBA. There are some examples of HBA such as ChCl, Tetrabutylammonium bromide (TBAB), Tetrabutylammonium chloride (TBAC), Methylimidazole, Dimethyloamine and L-proline. Generally, HBD are made up of glycerol, glycol, acid, amide, and alcohol groups, which readily donate their free hydrogen to HBA. DESs based on choline chloride (ChCl) have been shown to be effective for fuel desulfurization study [25]. However, the hygroscopic nature of ChCl is known to have an effect on the desulfurization performance for the ChCl-based DES; hence, during the preparation stage, it is very important to always dry the ChCl salt prior to usage in DES.
In this study, 16 different DESs were prepared by mixing ChCl with four different types of HBDs, namely glycerol, ethylene glycol, tetraethylene glycol, and polyethylene glycol in molar ratios ranging from 1:1 to 1:4. The physiochemical properties of the synthesized DES, including the density and viscosity were measured. It is proven that DES with high density and viscosity would affect their desulfurization performance [27,28]. FTIR analysis was also carried out to characterize and compare the four different types of ChCl-based DES based on their functional groups. After the screening process, the selected DES underwent several experiments to find the optimal reaction conditions. Lastly, the desulfurization performance by selected DES on real diesel fuel was evaluated.

DES Preparation
The DESs were prepared by mixing ChCl with GLY, EG, TEG, and PEG in molar ratios ranging from 1:1 to 1:4. ChCl acted as a HBA, while GLY, EG, TEG, and PEG served as HBDs. ChCl was dried in an oven overnight at 80 • C due to its hygroscopic nature. The HBA and HBD were mixed in the desired molar ratio in a vial and stirred at either 60 or 80 • C until a homogenous solution was obtained. DES preparation conditions are summarized in Table 1.

Density and Viscosity
The density and viscosity of each DES were measured at temperatures ranging from 25 to 90 • C using a SVM 3000 viscometer (Anton Paar, Graz, Austria). The viscometer was first calibrated using a solution with a known density supplied by the manufacturer.

Fourier Transform Infrared Chromatography (FTIR)
Perkin Elmer FTIR spectrometer Frontier was used for the functional group characterization of the synthesized DESs. Background spectrum was collected prior to the procedure as to eliminate unwanted residual peak from the sample spectrum. The wavenumbers produced by the DESs and their pure components were recorded from 400 to 4000 cm −1 .

Extractive Catalytic Oxidative Desulfurization Process
The performance of the DESs for ECODS was evaluated using H 2 O 2 as an oxidant and GO as the catalyst. The n-dodecane was used to prepare model oil (MO) that contained 100 ppm dibenzothiophene (DBT) as the SCC. Each DES was mixed with the MO at a 1:5 (v/v) ratio in a reaction vial. The molar ratio of oxygen to sulfur (O/S) in each mixture was 6:1, and the GO/S mass ratio was set at 1:25. The reaction mixtures were stirred at 400 rpm for one hour at room temperature (~25 • C), after which the contents were allowed to settle for 30 min. Approximately 2 mL was removed from the upper layer of each reaction mixture for HPLC analysis to find out the DBT concentration in the MO. The HPLC instrumental details and parameters are shown in Table 2.

Extractive Catalytic Oxidative Desulfurization Parameters
The DES with the best desulfurization performance was further used in the optimization process in order to determine the optimal reaction conditions for the fuel desulfurization to occur via ECODS method.

Volume Ratio of DES to the Model Oil
The volume ratio of DES to MO was set at 2.5:1, 1:1, 1:2.5, 1:5, and 1:10, respectively, as the manipulated variables for this experiment. The DES/MO volume ratio of 2.5:1 and 1:1 were served as control of the experiment to prove directly whether the selected DES has the potential to perform well in fuel desulfurization. In this experiment, the mass ratio of GO/S was set as 1:25 and molar ratio of H 2 O 2 /S was set as 6:1. The reaction mixture was filled into a reaction vial with a magnetic stirrer by pipetting the individual components using a micropipette accordingly. The reaction mixture was then stirred at 400 rpm for one hour under room temperature. After one hour of stirring, the reaction mixture was left aside for 30 min, in order to let the DES layer to settle at bottom of the reaction vial. Then, approximately 2 mL of the upper layer of the reaction mixture (model oil layer) was carefully syringed out using a micropipette and stored in a labelled HPLC vial. The experiment procedure was repeated for another two times to obtain more reliable result. The molar ratio of H 2 O 2 to S was set at 0, 2:1, 4:1, 6:1, and 8:1 as the manipulated variables for this experiment. In this experiment, the mass ratio of GO/S was set as 1:25 and volume ratio of DES to MO was set as 1:2.5. The reaction mixture was filled into a reaction vial with a magnetic stirrer by pipetting the individual components using a micropipette accordingly. The same steps were repeated as in the previous section.

Catalyst Dosage
The mass ratio of GO to sulfur (S) was set at 0, 1:400, 1:200, 1:100, 1:50, and 1:25 as the manipulated variables for this experiment. In this experiment, the molar ratio of H 2 O 2 /S was set as 4:1 and volume ratio of DES to MO was set as 1:2.5. The reaction mixture was filled into a reaction vial with a magnetic stirrer by pipetting the individual components using a micropipette accordingly. The same steps were repeated as in the previous section.

Effect of Temperature
The reaction temperature was set at 25, 40, 50 and 60 • C as the manipulated variables for this experiment. In this experiment, the molar ratio of H 2 O 2 to S was set as 4:1, the mass ratio of GO/S was set as 1:100 and volume ratio of DES to MO was set as 1:2.5. The same steps were repeated as in the previous section.

Effect of Stirring Speed
The stirring speed was set as 200, 400, 600, 800 and 1000 rpm as the manipulated variables for this experiment. In this experiment, the molar ratio of H 2 O 2 to S was set as 4:1, the mass ratio of GO/S was set as 1:100 and volume ratio of DES to MO was set as 1:2.5. The same steps were repeated as in the previous section.

Effect of Reaction Time
The reaction time was firstly set at 5 min, followed by 10 min, after that every 10-min interval from 20 min to 1 h, and later, every one-hour interval from 1 h to 8 h. In this experiment, the molar ratio of H 2 O 2 to S was set at 4:1, the mass ratio of GO/S was set as 1:100 and volume ratio of DES to MO was set as 1:2.5. The same steps were repeated as in the previous section.

Multistage Extraction
In this experiment, the optimal reaction parameter for the selected DES was utilized. The molar ratio of H 2 O 2 to S was set at 4:1, the mass ratio of GO/S was set as 1:100, volume ratio of DES to MO was set as 1:2.5, reaction temperature was set as 25 • C, stirring speed was set at 400 rpm, and the reaction time was set at 10 min for every stage of the catalytic oxidative desulfurization process. In the first stage, 45 mL of model oil was filled into a reaction vial with a magnetic stirrer by pipetting the individual components using a micropipette accordingly. The reaction mixture was then stirred at 400 rpm for 10 min under room temperature. After that, the reaction mixture was left aside for 30 min, in order to let the DES layer to settle at bottom of the reaction vial. Then, approximately 2 mL of the upper layer of the reaction mixture (model oil layer) was carefully syringed out using micropipette and stored in a labelled HPLC vial. In the second stage, 40 mL of the reacted model oil from the first stage was pipetted to a new reaction vial, followed with the respective DES/MO volume ratio, molar ratio of H 2 O 2 /S, and mass ratio of GO/S. The same steps was taken as in the first stage. The experimental procedure was repeated for third stage by using 35 mL leftover reacted model oil from the second desulfurization stage, fourth stage by using 30 mL, fifth stage by using 25 mL, and finally the sixth stage by using 20 mL leftover reacted model oil. The experiment procedure was repeated for another two times to obtain more reliable result. Real diesel fuel was collected from Petronas fuel station in Tronoh, Iskandar. The experimental procedure was similar to Section 2.6.7, where the real diesel fuel underwent six stage of catalytic oxidative desulfurization with the selected DES. This was carried out to test the ability of the selected DES to be applied in industrial petroleum refinery.

DES Preparation
Four different types of DES, namely Choline Chloride-Glycerol (ChCl-GLY), Choline Chloride-Ethylene Glycol (ChCl-EG), Choline Chloride-Tetraethylene Glycol (ChCl-TEG), and Choline Chloride-Polyethylene Glycol (ChCl-PEG) were successfully prepared with mol ratio of HBA:HBD from 1:1 to 1:4. Table 3 describes the physical appearance of each of the prepared DES. Based on previous literatures, only the DES that are able to form a clear transparent liquid is chosen for the characterization [28][29][30]. However, in this work, all the DES were characterized using FTIR and physical properties that include density and viscosity of the prepared DESs. Figure 1 illustrates the chemical interaction between the HBA and HBD during the preparation of DES. Electrostatic force is found between the positively charged nitrogen centre of the choline and negatively charged chloride ion in the ChCl salt. From the definition, it is known that DES is formed from a HBD and a HBD. According to Figure 1, hydrogen bond between the DES is formed between the hydrogen of the hydroxyl group from the HBD with the chloride ion from the ChCl salt which acts as HBA.

Density and Viscosity
The density and viscosity are two most important physiochemical properties that can affect the effectiveness of DES on fuel desulfurization. Generally, the densities and viscosities of the DESs prepared in this study decreased with increasing temperature. As seen in Figure 2, the density of the prepared DESs decreased following the order ChCl-GLY (1: ChCl-TEG and ChCl-PEG had similar densities; however, their densities were found to decrease as the number of ethylene groups and the proportion of HBD in the DESs increased, similar observation as reported by Makoś and Boczkaj [29]. Oppositely, the densities of ChCl-GLY increased as the proportion of GLY decreased.

Density and Viscosity
The density and viscosity are two most important physiochemical properties that can affect the effectiveness of DES on fuel desulfurization. Generally, the densities and viscosities of the DESs prepared in this study decreased with increasing temperature. As seen in Figure 2, the density of the prepared DESs decreased following the order ChCl-GLY (1:4) > ChCl-GLY (1:3) > ChCl-EG (1:2) > ChCl-EG (1:3) > ChCl-EG (1:4) > ChCl-TEG (1:3) > ChCl-TEG (1:4) ≈ ChCl-PEG (1:4). ChCl-EG, ChCl-TEG and ChCl-PEG had similar densities; however, their densities were found to decrease as the number of ethylene groups and the proportion of HBD in the DESs increased, similar observation as reported by Makoś and Boczkaj [29]. Oppositely, the densities of ChCl-GLY increased as the proportion of GLY decreased. The viscosity value was determined as it is proven that there is a good correlation between fluidity and molar conductivity of the DES, whereby the fluidity can be calculated as reciprocal of viscosity [30]. Most of the DES synthesized was reported with viscosity value around 100 cP to 200 cP at room temperature, and the similar result are shown in High viscosity inhibits mass transfer as there might be presence of an extensive hydrogen-bonding network within the DES compounds itself which hinders the mobility of the free species available for fuel desulfurization process [31]. This may have been the reason why the performance of the highly viscous ChCl-GLY DESs was relatively poor in the desulfurization screening test. These results are discussed in Section 3.4.

Fourier Tansform Infrared Chromatography
In the aspect of discovering new types of DES, FTIR is important in characterizing DES by predicting the functional group present in the DES. This is done by comparing the functional group found in the pure components as shown in Figure 4. The functional groups of a compound are usually illustrated as peaks in the FTIR graph, where each peak is caused by the vibration of the functional group due to the IR radiation from the FTIR machine. Different functional groups will show different values of absorbance or transmittance at slightly deviated wavenumbers based on their respective properties. By understanding the shifting of certain functional groups between pure compound and DES, it may help in the production of DES with more desirable properties in future studies. In this section, only the characterization of the four different types of DES in mol ratio of 1:4 will be discussed, Environments 2020, 7, 0097 8 of 17 as the same type of DES will give out the same spectra as they have the same functional group and same bondings. The viscosity value was determined as it is proven that there is a good correlation between fluidity and molar conductivity of the DES, whereby the fluidity can be calculated as reciprocal of viscosity [30]. Most of the DES synthesized was reported with viscosity value around 100 cP to 200 cP at room temperature, and the similar result are shown in Figure 3. According to Figure 3 High viscosity inhibits mass transfer as there might be presence of an extensive hydrogen-bonding network within the DES compounds itself which hinders the mobility of the free species available for fuel desulfurization process [31]. This may have been the reason why the performance of the highly viscous ChCl-GLY DESs was relatively poor in the desulfurization screening test. These results are discussed in Section 3.4.    The viscosity value was determined as it is proven that there is a good correlation between fluidity and molar conductivity of the DES, whereby the fluidity can be calculated as reciprocal of viscosity [30]. Most of the DES synthesized was reported with viscosity value around 100 cP to 200 cP at room temperature, and the similar result are shown in Figure 3. According to Figure 3 4). High viscosity inhibits mass transfer as there might be presence of an extensive hydrogen-bonding network within the DES compounds itself which hinders the mobility of the free species available for fuel desulfurization process [31]. This may have been the reason why the performance of the highly viscous ChCl-GLY DESs was relatively poor in the desulfurization screening test. These results are discussed in Section 3.4.   The stretching vibration of OH group was observed around 3100-3700 cm −1 ; however, a previous study discovered that the OH peaks for glycol based deep eutectic solvent occurs around 3200-3400 cm −1 [32]. The presence of OH group stretching vibration also indicates the presence of water content in the respective pure compound or prepared DES. According to Figure 4, glycerol (GLY) and ethylene glycol (EG) is found to have abroad peak around 3281.52 and 3296.84 cm −1 , whereas the OH peak found in tetraethylene glycol (TEG) and polyethylene glycol (PEG) are less intense than that of GLY and EG, at around 3409.11 and 3449.02 cm −1 , respectively. By comparing the four HBDs, the intensity of the OH peak is decreasing in the order of GLY > EG > TEG > PEG, suggesting that, the more aliphatic chain attached to the C-OH bond, the harder it will be for the compound to form hydrogen bonding with choline chloride. properties. By understanding the shifting of certain functional groups between pure compound and DES, it may help in the production of DES with more desirable properties in future studies. In this section, only the characterization of the four different types of DES in mol ratio of 1:4 will be discussed, as the same type of DES will give out the same spectra as they have the same functional group and same bondings. The stretching vibration of OH group was observed around 3100-3700 cm −1 ; however, a previous study discovered that the OH peaks for glycol based deep eutectic solvent occurs around 3200-3400 cm −1 [32]. The presence of OH group stretching vibration also indicates the presence of water content in the respective pure compound or prepared DES. According to Figure 4, glycerol (GLY) and ethylene glycol (EG) is found to have abroad peak around 3281.52 and 3296.84 cm −1 , whereas the OH peak found in tetraethylene glycol (TEG) and polyethylene glycol (PEG) are less intense than that of GLY and EG, at around 3409.11 and 3449.02 cm −1 , respectively. By comparing the four HBDs, the intensity of the OH peak is decreasing in the order of GLY > EG > TEG > PEG, suggesting that, the more aliphatic chain attached to the C-OH bond, the harder it will be for the compound to form hydrogen bonding with choline chloride.
In Figure 5, ChCl-GLY (1:4) has the most intense OH peak around 3305.70 cm −1 , followed by ChCl-TEG (1:4), which has the OH peak around 3307.67 cm −1 . The trend is followed by ChCl-TEG (1:4) which has moderate OH peak around 3364.74 cm −1 and ChCl-PEG (1:4) having the least intensity of OH peak around 3401.80 cm −1 . The general trend for intensity of the OH peak of each type of DES mainly follows the trend of HBD which was mentioned earlier, and only a slight deviation of 10-50 cm −1 is found. For ChCl-GLY (1:4) and ChCl-EG (1:4), the wavenumber shifted to the right compared to that of GLY and EG, while for ChCl-TEG (1:4) and ChCl-PEG (1:4), the wavenumber is found to shift to the left compared to that of TEG and PEG. The shifting of frequency numbers of OH peaks in DES might be due to the energy utilization for the formation of new bonds on the C-OH bonds in the HBD and also the hydrogen bonding formed between H-Cl as illustrated previously. Overall, ChCl-  In Figure 5, ChCl-GLY (1:4) has the most intense OH peak around 3305.70 cm −1 , followed by ChCl-TEG (1:4), which has the OH peak around 3307.67 cm −1 . The trend is followed by ChCl-TEG (1:4) which has moderate OH peak around 3364.74 cm −1 and ChCl-PEG (1:4) having the least intensity of OH peak around 3401.80 cm −1 . The general trend for intensity of the OH peak of each type of DES mainly follows the trend of HBD which was mentioned earlier, and only a slight deviation of 10-50 cm −1 is found. For ChCl-GLY (1:4) and ChCl-EG (1:4), the wavenumber shifted to the right compared to that of GLY and EG, while for ChCl-TEG (1:4) and ChCl-PEG (1:4), the wavenumber is found to shift to the left compared to that of TEG and PEG. The shifting of frequency numbers of OH peaks in DES might be due to the energy utilization for the formation of new bonds on the C-OH bonds in the HBD and also the hydrogen bonding formed between H-Cl as illustrated previously. Overall, ChCl-PEG (1:4) has the highest frequency of OH stretching vibration while ChCl-GLY (1:4) has the lowest frequency of OH stretching vibration in the spectra.   It was found that all the DESs and HBDs showed the C-H stretching vibration around frequency of 2840-3000 cm −1 , where vibrations found around 2950-2975 cm −1 and 2865-2885 cm −1 are mostly contributed from the CH 3 vibration, and stretching band produced by CH 2 vibrations mainly occurred at 2915-2940 cm −1 and 2840-2875 cm −1 [33,34]. The DES have methyl group which are contributed from the choline chloride having three free methyl groups, and also -CH 2 group contributed from the respective HBD as most of their carbon center are linked to the oxygen atom. For instance, TEG and PEG obtain a single sharp peak at 2868.75 cm −1 and 2867.32 cm −1 , respectively, whereas their respective DES, namely ChCl-TEG (1:4) and ChCl-PEG (1:4) also producing one sharp peak around 2869.56 cm −1 and 2870.34 cm −1 . However, for GLY and EG, double peaks are seen at 2937.83 cm −1 and 2882.86 cm −1 for GLY, while 2929.92 cm −1 and 2875.00 cm −1 for EG; however, these two peaks are less intense compared to that of TEG and PEG. The similar double peaks are found for ChCl-GLY and ChCl-EG which occurs at 2932.15 cm −1 and 2877.00 cm −1 for ChCl-GLY, whereas 2927.88 cm −1 and 2873.72 cm −1 for the latter. However, none of the DES show similar peak found in ChCl around 3024.67 cm −1 , contributed by the N-H stretching.
Another significant peak that can be used for comparison is the C-OH and C-O-C stretching vibration, which occur in every spectrum of the DESs, including single or double positive peaks around the wavenumbers of 1033.33-1108.95 cm −1 . Overall, the C-OH and C-O-C stretching for DES is almost the same with a 5 cm −1 difference in wavenumbers. This further concludes that the internal structure of HBD will remain the same in DES, the reaction between the choline chloride and the HBD only includes the formation of hydrogen bonding between terminal hydroxyl group from HBD and chloride ion from choline chloride salt. Figure 6 shows the comparison of FTIR graph between ChCl-PEG (1:4) with its respective pure components. It is found that FTIR graph of ChCl-PEG (1:4) matches the FTIR graph of PEG rather than ChCl. In other words, the DES tends to share similar functional group with its HBD, rather than ChCl, which is their HBA. Observed in Figure 7, there is no difference in the FTIR graph among the ChCl-GLY with different mole ratio of HBA to HBD as they have the same functional group.

Fuel Desulfurization Screening Results
The desulfurization performances of the DESs are compared in Figure 8. Only clear and transparent mixture without any insoluble precipitates, which is known as DES, was selected for the fuel desulfurization screening test as any insoluble precipitates present in the DES will hinder the

Fuel Desulfurization Screening Results
The desulfurization performances of the DESs are compared in Figure 8. Only clear and transparent mixture without any insoluble precipitates, which is known as DES, was selected for the fuel desulfurization screening test as any insoluble precipitates present in the DES will hinder the DES from effectively removing SCC from the MO. 400-450 cP (Figure 3) at room temperature. The high viscosity value restricts the mobility of free species in ChCl-GLY; thus, ChCl-GLY failed to extract any SCC from the MO.
As for ChCl-EG, it is known that EG, which is the HBD component is very reactive and according to the thermogravimetric analysis by Degaldo and his colleague, ChCl-EG DES have relatively lower onset decomposition temperature (380-390 K) than other synthesized DES in their study. For instance, they found that ChCl-GLY has onset decomposition temperature range from 457 K to 500 K [35]. Low onset decomposition temperature suggests that ChCl-EG was relatively more unstable compared to other synthesized ChCl-based DESs in this study, which leads to undesired desulfurization result observed in the screening result.
The ChCl-PEG (1:4) has the highest desulfurization performance among all the ChCl-based DESs, removing 20.28% of SCC from the MO. Thus, ChCl-PEG (1:4) was selected to be further used in the optimization of the desulfurization parameters.

Volume Ratio of DES to the Model Oil
The ChCl-PEG (1:4) DES and the MO were mixed in six different volume ratios (Figure 9a), and the mixtures were stirred at 400 rpm for one hour at room temperature. Desulfurization improved as the amount of DES in the MO increased. Mixtures with DES/MO volume ratios of 2.5:1 and 1:1 served as controls to assess the desulfurization capability of ChCl-PEG (1:4), and maximum desulfurization (82%) was achieved with a DES/MO volume ratio of 2.5:1. However, given the cost of DES production, achieving moderately good results with a smaller amount of DES is preferable to using such a large quantity. In a previous study, a large DES/MO volume ratio yielded only modest improvements in desulfurization performance [36]. The authors of the study reported that multistage extraction was more effective. However, increasing the DES/MO ratio has most often been found to improve desulfurization [11,37,38]. Satisfactory desulfurization (36.75%) was achieved at a DES/MO volume ratio of 1:2.5, and this DES/MO volume ratio was selected for following experiment.

Effect of Oxidant Amount
Study was conducted for the effect of the oxidant dosage on the ECODS process by varying the molar ratio of H2O2 to SCC. Desulfurization was performed at 25 °C for six hours using reaction mixtures with a GO/S mass ratio of 1:25. The stoichiometric reaction in Equation (1) shows that two  As for ChCl-EG, it is known that EG, which is the HBD component is very reactive and according to the thermogravimetric analysis by Degaldo and his colleague, ChCl-EG DES have relatively lower onset decomposition temperature (380-390 K) than other synthesized DES in their study. For instance, they found that ChCl-GLY has onset decomposition temperature range from 457 K to 500 K [35]. Low onset decomposition temperature suggests that ChCl-EG was relatively more unstable compared to other synthesized ChCl-based DESs in this study, which leads to undesired desulfurization result observed in the screening result.
The ChCl-PEG (1:4) has the highest desulfurization performance among all the ChCl-based DESs, removing 20.28% of SCC from the MO. Thus, ChCl-PEG (1:4) was selected to be further used in the optimization of the desulfurization parameters.

Volume Ratio of DES to the Model Oil
The ChCl-PEG (1:4) DES and the MO were mixed in six different volume ratios (Figure 9a), and the mixtures were stirred at 400 rpm for one hour at room temperature. Desulfurization improved as the Environments 2020, 7, 0097 12 of 17 amount of DES in the MO increased. Mixtures with DES/MO volume ratios of 2.5:1 and 1:1 served as controls to assess the desulfurization capability of ChCl-PEG (1:4), and maximum desulfurization (82%) was achieved with a DES/MO volume ratio of 2.5:1. However, given the cost of DES production, achieving moderately good results with a smaller amount of DES is preferable to using such a large quantity. In a previous study, a large DES/MO volume ratio yielded only modest improvements in desulfurization performance [36]. The authors of the study reported that multistage extraction was more effective. However, increasing the DES/MO ratio has most often been found to improve desulfurization [11,37,38]. Satisfactory desulfurization (36.75%) was achieved at a DES/MO volume ratio of 1:2.5, and this DES/MO volume ratio was selected for following experiment.

Effect of Oxidant Amount
Study was conducted for the effect of the oxidant dosage on the ECODS process by varying the molar ratio of H 2 O 2 to SCC. Desulfurization was performed at 25 • C for six hours using reaction mixtures with a GO/S mass ratio of 1:25. The stoichiometric reaction in Equation (1) shows that two moles of H 2 O 2 reacts with one mole of DBT to form a sulfone (DBTO 2 ). Therefore, the molar ratio of H 2 O 2 to SCC was expected to have a significant impact on desulfurization. An H 2 O 2 dosage above two has been reported to improve desulfurization efficiency; however, excessive amount of oxidant will deplete the desulfurization efficiency, attributed by the non-productive thermal decomposition of the oxidant [20].
The extent of desulfurization increased incrementally from 40.86% to 41.87% when the H 2 O 2 /S molar ratio was increased from 0:1 to 2:1 and reached 43.15% when the H 2 O 2 /S molar ratio was increased to 4:1 (Figure 9b). This confirmed that the addition of H 2 O 2 enhanced the desulfurization. However, increasing the oxidant dosage further did not improve the desulfurization efficiency. Desulfurization decreased to 41.06% and 39.53% when the H 2 O 2 /S molar ratios were increased to 6:1 and 8:1, respectively, which was consistent with trends observed in previous studies [16,39]. Excessive amount of H 2 O 2 yields water as by-product. The water or moisture produced dilutes the reaction mixture, reducing the catalytic efficiency. We thus determined that the optimal H 2 O 2 /S molar ratio is 4:1.

Catalyst Dosage
The effect of varying the catalyst dosage on desulfurization by ChCl-PEG (1:4) was evaluated by preparing reaction mixtures with five different GO/S mass ratios and a H 2 O 2 /S molar ratio of 4:1. Desulfurization was then allowed to proceed for six hours at room temperature. The optimal GO/S ratio was found to be 1:100, with a removal of 48.66% (Figure 9c). Lower GO/S mass ratios of 1:400 and 1:200 demonstrated a reduction in the desulfurization with a removal of 47.18% and 47.76%, respectively. The extent of desulfurization was significantly lower in a mixture with higher GO/S mass ratio of 1:50 (46.48%) and GO/S mass ratio of 1:25 (42%). Interestingly, 46.6% desulfurization was achieved without the catalyst.
The unique layered structure of GO provides a large surface area, which promotes collisions between molecules. Increasing the GO/S ratio from 1:400 to 1:100 provided more surface area for molecules to bind, and desulfurization proceeded until the sites on the GO catalyst were saturated. GO/S ratios of 1:50 and 1:25 resulted in poorer desulfurization, because the excess GO allosterically hindered interactions between the catalyst and the substrate.

Effect of Temperature
The effect of the reaction temperature on desulfurization was also investigated. Desulfurization was performed for six hours at four different temperatures using mixtures with the optimized oxidant and catalyst dosages. Desulfurization by ChCl-PEG (1:4) decreased as the temperature increased. At reaction temperatures of 25 • C, 40 • C, 50 • C, and 60 • C, the desulfurization percentages were 48.45%, 42.51%, 39.51%, and 36.93%, respectively (Figure 9d). Similar results were obtained in previous studies, in which 25 • C was found to be the optimum temperature for IL and DES extraction [12,37,40,41]. Increasing the reaction temperature resulted in lower rates of desulfurization.

Effect of Stirring Speed
The optimal stirring speed for the ECODS process was found to be 400 rpm, with removal of 48% SCC using the ChCl-PEG (1:4) DES (Figure 9e). Desulfurization increased from 38.75 to 48.04% when the stirring speed was increased from 200 rpm to 400 rpm. This was due to the likelihood of collisions between the catalyst, the DES and DBT in the model oil being higher at 400 rpm. Increasing the number of collisions thus enabled the DES to remove more SCC. However, desulfurization decreased to 38.57%, Environments 2020, 7, 0097 14 of 17 36.73% and 34.83% when the stirring speed was increased to 600, 800 and 1000 rpm, respectively. This may have been due to the limitations of the heating block since the optimal conditions could only be achieved at a stirring speed of 400 rpm.

Effect of Reaction Time
Desulfurization was performed using the ChCl-PEG (1:4) DES for a total of eight hours. MO samples were collected every five to ten minutes during the first hour. Samples were then collected hourly until eight hours passed. The results are shown in Figure 9f. The percentage of desulfurization increased slightly from 42.5 at minute five to 45% at minute ten and reached a plateau within the first hour. These results were consistent with those of previous studies. The optimal extraction time reported for DES extraction was 10 min, although desulfurization was monitored for only 60 min [11,38]. Increasing the reaction time to eight hours was done to determine whether desulfurization continued after the first 60 min. A slight increase from 45 to 47.8% was observed between hour one and hour two. The percentage increased to 48.8% by hour three, after which no additional desulfurization was observed. It was concluded that three hours was the optimal reaction time. However, to minimize energy consumption and control costs, subsequent reactions were allowed to proceed for 10 min.

Multistage Extraction
In multistage extraction, ChCl-PEG (1:4) DES was replaced with fresh ChCl-PEG (1:4) after each cycle. A mixture containing DES/MO ratio of 1:2.5, H 2 O 2 /S ratio of 6:1 and GO/S ratio of 1:100 was used for the first desulfurization cycle. The mixture was stirred for 10 min at room temperature. The same reaction condition was set as referred to previous desulfurization cycle and the steps were repeated for a total of six desulfurization cycles. The percentage of desulfurization increased from 49.65 in the first cycle to 95.68% in the sixth cycle ( Figure 10). The S concentration in the oil was 4.72 ppm after the sixth desulfurization cycle, which satisfied the Euro 5 standard criterion (<10 ppm).

Conclusions
ECODS method was utilized in this study, where H2O2 was used as oxidant, and GO as catalyst to evaluate the fuel desulfurization performance of ChCl-based DES in 100 ppm model oil and also real diesel fuel. ChCl-PEG (1:4) was selected after the screening result as ChCl-PEG (1:4) successfully removed 20.28% SCC in MO, which is the highest desulfurization percentage among all the prepared ChCl-based DES. After rounds of experiment, the optimal reaction conditions for the ECODS method were found to be DES/MO volume ratio of 1:2.5, GO/S mass ratio of 1:100, H2O2/S molar ratio of 4:1, 10 min reaction time per cycle, stirring speed at 400 rpm and reaction temperature of 25 °C. Lastly, the desulfurization performance of ChCl-PEG(1:4) reached 32.7% in real diesel fuel and 95.28% in

Desulfurization Performance in Real Diesel Fuel
The desulfurization performance of the ChCl-PEG (1:4) DES was then evaluated using real diesel fuel obtained from the Petronas fuel station. The desulfurization performance of ChCl-PEG (1:4) in 100 ppm MO and real diesel was compared as shown in Figure 10. It was expected that the ChCl-PEG (1:4) DES would be less effective in real diesel fuel, because it contained additional sulfur species and