Kinetics of Sulfur Removal from Tehran Vehicular Gasoline by g-C3N4/SnO2 Nanocomposite.

The graphitic carbon nitride/tin oxide (g-C3N4/SnO2) nanocomposite synthesized under microwave irradiation was used for adsorptive removal of sulfur-containing dibenzothiophene (DBT) from Tehran vehicular gasoline. High-resolution transmission electron microscopy, X-ray powder diffraction, energy dispersive X-ray spectroscopy, Brunauer-Emmett-Teller, Fourier-transform infrared spectroscopy, and field emission scanning electron microscopy techniques determined the adsorbent characteristics, and gas chromatography with a flame ionization detector determined the DBT concentration of the samples. Application of the experimental data into the solid/fluid kinetic models indicated a chemisorption control regime that increased the removal of sulfur from the commercial samples used. A pseudo-second-order reaction with the rate constant of 0.015 (g mg-1 min-1) and total conversion time of 316 min described the adsorption process. Based on the real fuel results, the adsorption capacity of the g-C3N4/SnO2 adsorbent reached 10.64 mg S g-1 adsorbent at equilibrium conditions. This value was the highest adsorption capacity obtained so far for a commercial gasoline sample. The g-C3N4/SnO2 nanocomposite could, therefore, be introduced as an inexpensive, easily obtainable adsorbent that can significantly remove the sulfur from the vehicular gasoline fuels.

Introduction

Environmental concerns have compelled fuel refineries to remove the sulfur-containing compounds from transportation fuels.[1,2] Removal of sulfur compounds leads to increase of combustion engine lifetime and positive environmental attitudes. Research on clean fuel has been considered for many years to reduce air pollution and prevent the introduction of pollutants into the environment.[3,4] Existing sulfur compounds in fuel cause emission of SO, which increases the possibility of acid rain.[4−6] Accordingly, many strict regulations have been established in many countries to reduce sulfur content in public transportation fuels such as diesel and gasoline.[7] Recent regulations for the reduction of the current sulfur content to less than 10 ppm have been proposed.[8]

Improvement of the fuel cell technology is now a hot issue in research and development units of the industries. Fuel cells consume gasoline and diesel as the direct or indirect primary fuels. Since a solid oxide fuel cell works at high temperatures, the sulfur content should be less than 10 ppm. Furthermore, deep desulfurization below 1 ppm for proton-exchange-membrane fuel cell is also required.[2,9,10]

Traditional hydrodesulfurization has been used to lower sulfur content as a conventional process.[11] This process is complicated and costly especially for the removal of some sulfur compounds like dibenzothiophene (DBT) and its derivatives that are present in a few tens of ppm when a profound desulfurization goal is usually required.[5,12−14] Besides, about 90% of the sulfur content of the studied gasoline in this research is in DBT. The removal of these compounds from transportation fuels (primarily diesel and gasoline) has been a controversial topic in the industry.[15,16] In this regard, alternative or complementary desulfurization processes for ultralow sulfur fuel production have been evolved such as oxidative desulfurization,[17−20] biodesulfurization,[21] extraction,[22−25] and adsorptive desulfurization.[26−29] Investigations have shown that adsorptive desulfurization is a promising procedure at the ambient temperature and pressure because of its low-energy consumption and investment costs.

Various adsorbents have been studied before. Yang and co-workers used π-complexation-based sorbents that were obtained by ion-exchanging faujasite-type zeolites with Cu+, Ni2+, or Zn2+ cations. They reported Cu(I)-Y (VPIE) as the best sorbent with adsorption capacities of 0.395 and 0.278 (mmol S g–1 sorbent) for commercial jet fuel and diesel, respectively.[30] Over past 2 decades, several studies have been conducted on using zeolite,[31] alumina,[27] mesoporous silica,[32] and activated carbon[33] as supports for various sorbents to remove sulfur-containing compounds. For instance, Sarda et al. prepared adsorbents by varying the Ni/Cu loadings onto ZSM-5 (Si/Al = 20) and the activated alumina for removal of sulfur from diesel fuel. Their results showed that sulfur removal strongly depends on the nature and amount of the metal and the support material. Their results showed that the activated alumina has a higher sulfur removal power than ZSM-5 due to the more facile diffusion of the large sulfur molecules to the larger pores of the activated alumina than ZSM-5.[27]

Tin oxide has recently been used in a variety of applications such as gas sensors, dye-based solar cells, optoelectronic devices, electrode materials, and catalysis.[34] The presence of active sites on the surface of SnO2 particles results in a suitable catalytic behavior.[35] In a previous study, the authors presented the synthesis and characteristics of the porous graphitic carbon nitride (g-C3N4) nanocomposite,[36] which seemed usable for the construction of the support substance needed for sulfur removal from the hydrocarbon fuels. According to our knowledge, the g-C3N4/SnO2 nanocomposite has not been used as the adsorbent of sulfur compounds in any studies until now. This study aims to find a high-capacity adsorbent for the removal of sulfur from a highly used commercial petroleum product. Effects of contact time and adsorbent dose on sulfur-containing DBT removal from the regular gasoline of a local Iranian gas station are investigated. Different mathematical correlations are examined to find out the appropriate kinetic model that governs the process. The implication of the devised technique leads to the maximum adsorption capacity of 10.64 mg S g–1 adsorbent, which is obtainable after reaching the equilibrium conditions.

Results and Discussion

The X-ray powder diffraction (XRD) patterns of the g-C3N4/SnO2 nanocomposite before and after calcination are shown in Figure . The design of the calcined powder represents the significant formation of the tetragonal-phase SnO2 according to JCPDS card no. 88-0287 with the characteristic peaks located at 2θ angles of 26.50, 33.86, and 51.84°. The magnified portion of the figure indicates the existence of a peak at 26.74°. This peak belongs to the g-C3N4 (002) crystallographic plane.[37] g-C3N4 has another peak at 13.66°, which is also observed in the XRD pattern shown in Figure . Overlapping of the characteristic SnO2 (110) peak with the g-C3N4 (002) peak has been the source of confusion in analyzing the XRD patterns of the g-C3N4/SnO2 composites, as reported in the previous papers.[36,38] The current study resolved the issue by nanocomposite powder calcination at 550 °C. Based on the JCPDS Card No. 26-1076, the third peak seen in the magnified insets of Figures and 9 attributes to the carbon hexagonal phase. During powder calcination at 550 °C, a meager amount of g-C3N4 decomposes to this hardly observable tiny phase, plus evolving gaseous nitrogen. The crystallite average size of different nanocomposite samples was evaluated from the Scherrer equation.[39] It was ∼14.59 nm before heating and 18.83 nm after calcination, which confirmed particle size growth during calcination.

Figure 1

XRD patterns of the g-C3N4/SnO2 nanocomposite (a) before calcination (b) after calcination at 550 °C for 1 h.

Figure 9

XRD pattern of the adsorbent after the adsorption test.

XRD patterns of the g-C3N4/n class="Chemical">SnO2 nanocomposite (a) before calcination (b) after calcination at 550 °C for 1 h.

Field-emission scanning electron microscope (FESEM) and high-resolution transmission electron microscopy (HRTEM) were used to determine the morphology of the g-C3N4/SnO2 nanocomposite. In Figure a, the stacked layers of graphitic carbon nitride with a dimension of about 34.12 μm and highlighted pores are observable. The structure of the sample is less porous than the g-C3N4/SnO2 nanocomposite produced in our previous research.[36] During irradiation with microwave, some urea molecules decomposed to gaseous products that tended to exit from the graphitic layers of C3N4 toward the root to the porous structure.[36] Consequently, the lower amount of urea (the precursor for the fabrication of the g-C3N4/SnO2 nanocomposite) could result in fewer porosities produced in this study sample.

Figure 2

FESEM (a–c) and HRTEM images (d, e) of the g-C3N4/SnO2 nanocomposite: (a) graphitic carbon nitride sheet, (b) SnO2 nanoparticles scattered on the g-C3N4 sheets, (c) SnO2 nanoparticles’ shape and size, (d) SnO2 nanoparticles accumulated on the g-C3N4 sheets, and (e) lattice fingers of the SnO2 nanoparticles.

FEn class="Chemical">SEM (a–c) and HRTEM images (d, e) of the g-C3N4/SnO2 nanocomposite: (a) graphitic carbon nitride sheet, (b) SnO2 nanoparticles scattered on the g-C3N4 sheets, (c) SnO2 nanoparticles’ shape and size, (d) SnO2 nanoparticles accumulated on the g-C3N4 sheets, and (e) lattice fingers of the SnO2 nanoparticles.

Figure b shows the SnO2 nanoparticles scattered on g-C3N4 sheets. From image analysis, the average diameter of these particles is ∼23.58 nm. Figure d displays the HRTEM image of the SnO2 nanoparticles (the darker area) accumulated on a carbon nitride sheet. An average diameter of ∼14 nm was obtained from the image. This value agreed well with the crystallite size obtained from the XRD by the Scherrer equation. Hindering the particle growth together with a hefty gas evolution due to urea decomposition caused the decrease in average size to ∼14 nm. The lattice fingers of the sample are illustrated in Figure e. The lattice distance measured is ∼0.26 nm, which is indexed as the (101) plane of the SnO2 as depicted in the XRD pattern in Figure .

Figure shows the elemental maps, the energy dispersive X-ray spectroscopy (EDS) spectrum, and the analysis of the g-C3N4/SnO2 nanocomposite produced in this research. According to the maps, all elements (carbon, nitrogen, tin, and oxygen) are uniformly distributed in the sample. The overlap of the elemental maps supports a well-coupled SnO2 structure on the g-C3N4 layers.[40] The atomic ratio of the Sn–O in the studied area of Figure b is not far from 1:2 of a stable SnO2 molecule. The ratio higher than 3:4 for carbon/nitrogen is due to the decomposition of the stray hydrocarbon complexes that may have contaminated the production chamber.

Figure 3

(a) Elemental maps of C, N, Sn, and O in the g-C3N4/SnO2 and (b) EDS spectrum of the nanocomposite sample showing the content of the elements present in the g-C3N4/SnO2 produced in this research.

(a) Elemental maps of C, N, n class="Chemical">Sn, and O in the g-C3N4/SnO2 and (b) EDS spectrum of the nanocomposite sample showing the content of the elements present in the g-C3N4/SnO2 produced in this research.

Figure a illustrates nitrogen adsorption/desorption for the g-C3N4/SnO2. The shape of the curves and hysteresis loop suggests type IV isotherms for the sample, which indicates the existence of mesopores.[41] The Brunauer–Emmett–Teller (BET) surface area of the sample was calculated to be 66.36 m2 g–1. Figure b shows the pore size distributions of the g-C3N4/SnO2 via the BJH (Barrett–Joyner–Halenda) model. Accordingly, the BJH pore volume and the mean pore diameter were measured to be 0.43 cm3 g–1 and 18.76 nm, respectively.

Figure 4

(a) N2 adsorption/desorption and (b) pore size distributions curves of the g-C3N4/SnO2.

(a) N2 adsorption/desorption and (b) pore n class="Chemical">size distributions curves of the g-C3N4/SnO2.

Figure illustrates the effect of contact time on the local gasoline DBT adsorption at 298 K. The original DBT content of the gasoline was 358.51 mg DBT L–1 (62.38 mg S L–1), and its final content after full adsorption was 168.89 mg DBT L–1 (28.68 mg S L–1). Figure indicates that fast initial adsorption is followed by a slow-down rate that tends to reach the equilibrium state after 180 min. The equilibrium amount is, therefore, ∼3.37 mg S g–1 adsorbent. This change is due to the reduction of the active adsorbent sites by adsorbing sulfur atoms.

Figure 5

Effect of contact time on S removal from gasoline by g-C3N4/SnO2 (adsorbent mass = 0.1 g; T = 25 °C).

Effect of contact time on S removn class="Chemical">al from gasoline by g-C3N4/SnO2 (adsorbent mass = 0.1 g; T = 25 °C).

Effect of Adsorbent Dose on Sulfur Removal

Different doses of the adsorbent were added to each sample at 298 K to remove sulfur from gasoline of the local gas station. The data obtained after stirring the mixture for 360 min were plotted against the adsorbent weight (see Figure ). The best-fit curve plotted in Figure was used to determine the equilibrium data tabulated in Table . These data showed that the sulfur removal tends to 75.45% of the initial value by an adsorbent amount approaching to 1 g in weight. This increase indicated more favorable adsorption conditions with a higher number of particles of larger specific surface areas.

Figure 6

Effect of the adsorbent dose on the removal of the sulfur (V = 10 mL, T = 25 °C).

Table 1

Equilibrium Data of Sulfur Removal by the g-C3N4/SnO2

adsorbent mass (g)	Ce (mg S L–1)	Se removal (%)	qe (mg g–1)
0.025	40.88	34.46	8.60
0.05	35.46	43.16	5.38
0.1	28.68	54.02	3.37
0.5	18.47	70.39	0.88
1	15.31	75.45	0.47

Effect of the adsorbent don class="Chemical">se on the removal of the sulfur (V = 10 mL, T = 25 °C).

The gas chromatography with a flame ionization detector (GC-FID) chromatogram peaks of the DBT compound obtained from the initial and the desulfurized gasoline samples are compared in Figure . The noticeable decline of the areas under the curves by the adsorbent amount represents the rising of desulfurization by the adsorbent amount. This rising trend supports the data plotted in Figure .

Figure 7

Peaks of DBT obtained from the GC-FID chromatogram for (a) initial gasoline and (b) gasoline treated with 0.1 g, (c) 0.5 g, and (d) 1 g of adsorbent (V = 10 mL; T = 25 °C).

Peaks of DBT obtained from the GC-FID chromatogram for (a) initin class="Chemical">al gasoline and (b) gasoline treated with 0.1 g, (c) 0.5 g, and (d) 1 g of adsorbent (V = 10 mL; T = 25 °C).

Adsorption Isotherms

Removal of sulfur was due to the adsorption of DBT by g-C3N4/SnO2 that was modeled by two sorption isotherms, Freundlich and Dubinin–Radushkevitch. The characteristic parameters of the models were evaluated from the slope and the intercept of the plots using regression analysis. The linear form of the Freundlich isotherm is as follows[42]where qe is the removed amount of sulfur after stirring the mixture for 360 min (mg S g–1 adsorbent), Ce is the equilibrium sulfur concentration (mg S L–1), and kf [in (mg g–1)(mg–1 L)1/] and n are the Freundlich constants representing the adsorption capacity and rate, respectively.[43]

The linear form of the Dubinin–Radushkevitch (D–R) isotherm is as follows[43]where qm (mg g–1) is the theoretical saturation capacity, β (mol2 J–2) is the constant related to the mean free energy of adsorption, and ε is the Polany potential that is equal to[44]In this equation, R (J mol–1 K–1) and T (K) are the ideal gas constant and the absolute temperature, respectively. The adsorption average free energy (E, J mol–1) is related to the change of free energy when 1 mole of the adsorbate transfers from infinity to the surface of the adsorbent in the solution. E can be evaluated from constant β using the following relationship[45]The Freundlich constants kf and n were evaluated from the data plotted in Figure a and are summarized in Table . The values of β and qm of the the D–R isotherm were also obtained from Figure b and are tabulated in Table . As can be seen in Figure and also in Table , the experimental data for removal of sulfur by g-C3N4/SnO2 fits better with the Freundlich isotherm (R2 = 0.997) than with the D–R isotherm (R2 = 0.973). The reason for better fitting of the former is the heterogeneous surface bindings assumed in the Freundlich isotherm.[46] The low value of E (60.86 J mol–1) indicates that the physical adsorption dominates the system.[43,45] As a result, the main factor contributing in the mechanism of adsorption is van der Waals interaction forces between adsorbates and adsorbents.[43]

Figure 8

(a) Freundlich and (b) Dubinin–Radushkevitch adsorption isotherms for the removal of sulfur by g-C3N4/SnO2.

Table 2

Two Sets of the Isotherm Constants Related to Sulfur Adsorption on the g-C3N4/SnO2 Nanocomposite Produced in This Research

model	Freundlich			Dubinin–Radushkevitch
parameter	kf (mg g–1) (mg–1 L)1/n	n	R2	β (mol2 J–2)	qm (mg g–1)	E (J mol–1)	R2
g-C3N4/SnO2	1.72 × 10–4	0.343	0.997	1.35 × 10–4	10.64	60.86	0.973

(a) Freundlich and (b) Dubinin–Radushkevitch adsorption isotherms for the removal of n class="Chemical">sulfur by g-C3N4/SnO2.

In this study, the maximum adsorption capacity of g-C3N4/SnO2 was obtained to be 8.60 mg S g–1 adsorbent, which is 80.83% of the theoretical saturation capacity generated from the D–R isotherm model. The desulfurization value obtained in this research for the local commercial gasoline, which contained diverse aromatic hydrocarbons, was the highest amount achieved so far. Previous researchers had obtained a relatively low value of 3.5 mg S g–1 by graphenelike boron nitride,[26] 2.18 mg S g–1 by the Ni/SiO2-Al2O3 adsorbent and 1.57 mg S g–1 by activated alumina,[47] 0.32 mg S g–1 by zeolite as an adsorbent,[27] 1.2 mg S g–1 by 15%TiO2/C–SiO2,[48] and 7.3 mg S g–1 by utilization of 55 wt % metallic Ni with silica–alumina as support of the adsorbent.[49] The authors focused on the removal of sulfur from the DBT because it was the highest sulfur-containing compound of the selected gasoline.

Removal Mechanism

The XRD pattern of the adsorbent after adsorption was surveyed to confirm the adsorption of DBT on the g-C3N4/SnO2. As shown in Figure , all of the characteristic peaks of the SnO2 can be seen according to JCPDS card no. 88-0287. Also, the characteristic g-C3N4 peaks were located at 2θ angles of 27.04 and 13.54°. Therefore, the results indicate that the adsorbent has been stable after the adsorption test. Besides, the peaks at 2θ angles of 28.06, 23.19, and 17.40° belong to the DBT compound according to JCPDS card no. 34-1687.

The Fourier-transform infrared spectroscopy (FTIR) pattern was studied to characterize the structure of the adsorbent (g-C3N4/SnO2) before and after the adsorption test. As depicted in Figure a, the infra red (IR) peak at around 603.87 cm–1 corresponds to the Sn–O stretching vibration, which is related to the characteristic peak of SnO2. The peaks at 1403, 1630, and 1722 cm–1 are attributed to the stretching mode of C–N heterocycles. The peaks at 806 and 1047 cm–1 could be ascribed to the tri-s-triazine ring.[36,50] All of the reported IR peaks from 806 to 1722 cm–1 are characteristic of g-C3N4. Also, the peaks at 2858 cm–1 and in the region from 300 to 3500 cm–1 are attributed to the stretching vibrations of −CH groups, the N–H bond, and the surface-adsorbed OH groups. These results demonstrate the synthesized adsorbent in which the g-C3N4/SnO2 nanocomposite contains both g-C3N4 and SnO2. According to Figure b, the characteristic peaks of the g-C3N4/SnO2 still can be seen in the FTIR spectrum of the utilized adsorbent that represents the stability of the adsorbent after the adsorption test. Besides, the main peaks corresponding to DBT could be distinguished. The bands at 1376 and 2924 cm–1 could be allocated to the thiophene ring and C–H in the benzene ring.[51] Also, the band at 1376 cm–1 along with the weak band at 1413 cm–1 suggests that thiophene coordinates via its sulfur atom to unoccupied metal sites (Sn),[52] as depicted in Figure . Some mechanisms involving π-complexation, which is the formation of bonds between a pair of electrons of sulfur and thin empty s-orbitals, could occur in addition to van der Waals interactions. While for g-C3N4, the most DBT adsorption can probably rely on the physical adsorption of pore channels.

Figure 10

FTIR spectrum of (a) initial adsorbent and (b) utilized adsorbent.

Figure 11

Schematic representation of the sulfur chemisorption mechanism by the g-C3N4/SnO2 nanocomposite.

Schematic representation of the n class="Chemical">sulfur chemisorption mechanism by the g-C3N4/SnO2 nanocomposite.

Table illustrates the variation of the sulfur concentration against time. Fractional conversion is evaluated from the following correlationHeterogeneous models for two-phase reactions can be examined with the experimental data for an explanation of the process. The model assumed two steps comprising of mass transfer from gasoline (ext) and adsorption reaction at the gasoline/nanocomposite interface (ads). According to the additive stage times concept and considering the necessary corrections for external mass transfer, the time (t) to achieve a certain degree of conversion (X) can be found as follows.[53]where t is the overall adsorption time, τext and τads are time constants for external mass transfer and chemisorption of DBT, respectively, and q(X) and q(X) are mathematical functions describing individual conversion steps for spherical nanoparticles[53]All other combinations, including flat and cylindrical shapes and mixed kinetic regimes, were mathematically examined but did not show closer match with the experimental data. Best-fit examination indicated a total conversion time of τads = 316 min with the possible chemisorption mechanism illustrated in Figure .

Table 3

Kinetic Data of Sulfur Adsorption on 0.1 g of g-C3N4/SnO2

time (min)	Ct (mg S L–1)	qt (mg g–1)	X
0	62.38	0	0
15	48.58	1.38	0.41
30	40.28	2.21	0.66
60	34.08	2.83	0.84
120	30.68	3.17	0.94
240	29.18	3.32	0.98
360	28.68	3.37	1

Pseudo-first and pseudo-second-order rate equations were also tested for best-fit determination[54,55]where k1 (min–1) and k2 (g mg–1 min–1) are the pseudo-first and second-order rate constants, respectively. Variations of the left-hand terms of eqs and 10 were plotted against time. From the plots, the rate constants and the coefficients of determination were obtained by the regression analysis evaluation (Figure ). The results are listed in Table . The pseudo-first-order model did not show a straight line for ln(qe – qt) vs t, while the best-fit result belonged to the pseudo-second-order model (Figure ). The adsorption capacity gained from this model was 3.56 (mg g–1), which was close to the value of 3.37 (mg g–1) obtained from Figure . Thus, it was concluded that pseudo-second-order surface adsorption controls the rate of sulfur removal by g-C3N4/SnO2.

Figure 12

Pseudo-second order verification of the experimental data for sulfur removal from gasoline by g-C3N4/SnO2 (adsorbent mass = 0.1 g; T = 25 °C).

Table 4

Kinetic Parameters Obtained for Sulfur Removal from Gasoline Based on First- and Second-Order Adsorption Reaction Assumptions

model	pseudo-first-order			pseudo-second-order
parameter	qe(mg g–1)	k1 (min–1)	R2	qe (mg g–1)	k2 (g mg–1 min–1)	R2
value	2.19	0.017	0.952	3.56	0.015	0.999

Pseudo-n class="Chemical">second order verification of the experimental data for sulfur removal from gasoline by g-C3N4/SnO2 (adsorbent mass = 0.1 g; T = 25 °C).

Conclusions

The previously introduced g-C3N4/SnO2 nanocomposite having the BET surface area of 66.36 (m2 g–1) was used as a new sorbent for sulfur removal from the commercial gasoline samples obtained from a domestic gas station in Tehran. The experimental results showed the increase in sulfur removal with the adsorbent dose, consistent with the well-known Freundlich isotherm equation. Kinetic model calculations showed a single-step mechanism consisting of a pseudo-second-order chemical reaction with the rate constant of 0.015 g mg–1 min–1, which explained the chemisorption of sulfur on the g-C3N4/SnO2 nanocomposite sample. The nanocomposite loaded with the sulfur was then removed by high-speed centrifugation. Due to both ease in production and environmentally friendly process, the g-C3N4/SnO2 nanocomposite was therefore found to be a desirable and promising material for petroleum sulfur elimination applications.

Experimental Methods

Preparation of the Adsorbent

Urea (Merck) as the precursor of carbon nitride, 37% analytical-grade hydrochloric acid (HCl) (density of 1.19 g cm–3), and 25% ammonium hydroxide were all obtained from Merck. Commercially available pure tin, ethanol, and deionized (DI) water were purchased from a local chemical supplier. The g-C3N4/SnO2 nanocomposite was synthesized following the procedure described in a previous paper.[36] Microwave and urea were used to save energy and reduce the adsorbent cost.

It took 3 h to mix and dissolve 0.175 g of n class="Chemical">tin in 10 mL of hydrochloric acid. The pH of the solution was adjusted to 12 by adding 300 mL of ammonium hydroxide. One gram of urea was added, and the beaker was placed in a microwave oven operated at 2.45 GHz and 630 W for 30 min. The product was then washed with DI water and ethanol and centrifuged for 5 min at 4000 rpm. The product was dried at 180 °C for 48 h and calcined at 550 °C for 1 h.

Characterization

X-ray diffraction (XRD) patterns of the adsorbents were recorded using a Philips X’Pert Pro MPD with a Cu Kα radiation source. The BET surface areas of the adsorbents were measured by the Brunauer–Emmett–Teller (BET) method (BELSORP MINI II) through N2 adsorption/desorption. The Fourier transform infrared (FTIR) spectra of the samples were recorded on AVATAR (Thermo). The morphology of the adsorbents was investigated by using a field-emission scanning electron microscope (FESEM, Tescan Mira 3-XMU). High-resolution transmission electron microscopy (HRTEM) was performed using a JEM-2100F device equipped with a Gatan Orius SC1000 CCD camera, and the accelerating voltage was 80 kV. The sample powder was ultrasonically dispersed in acetone for approximately 10 min, and the suspensions were deposited on a carbon-coated copper grid. EDS mapping and EDS spectrum were carried out in scanning TEM mode using an Oxford XMax-80T EDS detector and the mentioned FESEM equipped with a SAMX EDS detector, respectively.

Adsorption Process

Gasoline used in the present study was obtained from a local petrol pump, and its total sulfur was determined by a HORIBA SLFA-2800 X-ray fluorescence analyzer (71.6 mg S L–1). Gasoline contained 358.51 mg DBT L–1 sulfur-containing DBT (62.38 mg S L–1). Different amounts of adsorbent’s dosage (0.025–1 g) were mixed with 10 mL of gasoline in a shaker for 15–360 min at ambient conditions. Gas chromatography with a flame ionization detector (GC-FID, Agilent 7890A GC) equipped with a DB-5 column (L = 30 m; id = 0.25 mm) was utilized for detecting the remaining sulfur concentration in the gasoline. The amount of sulfur compound adsorbed onto the sorbent (qt, mg g–1) is calculated fromwhere C0 and C (mg L–1) are sulfur concentrations at the beginning and at time t, respectively, V (L) is the volume of the gasoline, and m (g) is the mass of the adsorbent. The removal percentage of the sulfur can be obtained fromAfter adsorption, the nanocomposite material containing sulfur species was eliminated from the purified gasoline by centrifugation for 10 min at 6000 rpm.