Facile Synthesis of a Novel Ni-WO3@g-C3N4 Nanocomposite for Efficient Oxidative Desulfurization of Both Model and Real Fuel.

The current study comprises the successful synthesis of a Ni-WO3@g-C3N4 composite as an efficient and recoverable nanocatalyst for oxidative desulfurization of both model and real fuel oils. The physiochemical characterization of the synthesized composite was confirmed via Fourier transform infrared spectroscopy, X-ray diffraction, scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy, and thermogravimetric analysis. SEM results showed that Ni-WO3 particles were well-decorated on the g-C3N4 surface with an interesting morphology as appeared on the surface like spherical particles. The obtained findings revealed that 97% dibenzothiophene (DBT) removal can be achieved under optimized conditions (0.1 g of the catalyst, 1 mL of an oxidant, 100 mg/L DBT-based model fuel, a time duration of 180 min, and a temperature of 40 °C). Additionally, the catalytic activity for real fuel was also investigated in which 89.5 and 91.2% removal efficiencies were achieved for diesel and kerosene, respectively, as well as fuel properties following ASTM specifications. A pseudo first-order kinetic model was followed well for this reaction system, and the negative value of ΔG was due to the spontaneous process. Additionally, the desulfurization study was optimized via a response surface methodology (RSM/Box-Behnken design) for predicting optimum removal of sulfur species by drawing three-dimensional RSM surface plots. The Ni-WO3@g-C3N4 proved to be a promising catalyst for desulfurization of fuel oil by exhibiting reusability of five times with no momentous decrease in efficiency.

Introduction

Globally, researchers are focusing on the production of green fuels because of the environmental protection and increasing strict legislations regarding the limit of sulfur in hydrocarbons not to be more than 10 mg/L.[1−4] The major challenge is the removal of sulfur compounds such as thiophenes (Th), benzothiophene (BT), dibenzothiophene (DBT), and their derivatives due to stringent regulations in oil refineries. As we all know, the traditional desulfurization technology (hydrodesulfurization (HDS)) requires hydrogen under severe operating conditions (high temperatures and pressures) as well as it is unable to remove aromatic sulfur compounds such as DBT.[5,6] Considering the above-mentioned reasons, alternative technologies are under development, and they are adsorption,[7,8] bio-desulfurization (BD),[9] extractive desulfurization (ExD),[10] and oxidative desulfurization (ODS).[11] Among them, oxidative desulfurization could be a viable process for the removal of sulfur species from fuel oil owing to its high efficiency, economic aspect, eco-friendliness, low energy cost, handiness, and resistance to harsh operating conditions. In light of these striking properties, researchers are paying prime importance to this technique (ODS), which is a great challenge to produce green fuels.[12,13] In this technology, sulfur compounds oxidized into sulfoxides and sulfones are considered under a catalytic oxidation reaction with an oxidant such as H2O2.

Desulfurization assisted by an oxidative desulfurization (ODS) process is a prospective methodology for eliminating sulfur-based aromatic compounds from liquid fuels (gasoline, diesel, light diesel oil, and kerosene). During the past few decades, oxidative catalysis has enlarged substantial research interests in different applications such as water splitting, pollutant degradation, and desulfurization of fuel oil. Based on experimental evidence, above 90% of thiophene, dibenzothiophene, and their derivatives can be easily oxidized under optimized conditions.[14] Moreover, ultradeep sulfur removal by ODS can now be carried out with H2O2 (oxidant)-assisted inorganic–organic hybrid catalysts when compared with traditional desulfurization methods.

Nowadays, researchers are using graphitic carbon nitride (g-C3N4) for the desulfurization process because of its high efficiency and some of its properties including the fact that it is composed of 2D layers of its monomer, heptazine. g-C3N4 belongs to that class of compounds that have a high level of nitrogen and can be synthesized by using one major type of polymerization, that is, polycondensation.[15] The monomers used for its synthesis are organic precursors, which further include cyanamide, dicyandiamide, melamine, and urea. g-C3N4 has its vast applications in environmental decontamination and artificial photosynthesis only because of these properties: it is abundant on earth, has a distinctive electronic configuration, and shows great physiochemical stability. The first reason behind the great chemical stability of g-C3N4 is the presence of carbon and nitrogen atoms that are sp2 hybridized, and the second is the presence of a high level of nitrogen (∼55–62%).[16,17] These two described properties make g-C3N4 a worthy and suitable material for synthesis of hybrid materials. Additionally, transition metals such as nickel (Ni) and tungsten (W) are expected to be promising candidates to substitute precious metals (Pt, Rh, and Ir) for different catalytic activities owing to their low cost, high stability, and outstanding redox capabilities. Due to their unique physicochemical properties, nickel and tungsten are widely used in refractory alloys and as nanoparticles of metals in the form of oxides, carbides, and sulfides. Researchers are trying to develop nickel-, copper-, cobalt-, iron-, and tungsten-based catalysts supported onto thin sheets (g-C3N4 and graphene) for electrocatalytic (water splitting (HER and OER)) and photocatalytic applications (pollutant degradation), sensors, and many organic syntheses.[18−23]

The facts concerning oxidative desulfurization (ODS) of liquid fuels using hybrid materials are still vague. Various studies have been published on photodegradation of organic pollutants, energy production, and many other applications using hybrid materials. However, there are limited studies on the usage of a hybrid composite as a catalyst in the oxidative desulfurization (ODS) of liquid petroleum. Regarding the overhead remarks, the objective of this work was to synthesize a (Ni-WO3@g-C3N4) nanocomposite and study its performance for the first time for oxidative desulfurization of model fuel (DBT) and real fuel. The prepared catalyst was characterized via Fourier transform infrared spectroscopy (FT-IR), powder X-ray diffraction (XRD), scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDX), and thermogravimetric analysis (TGA) and discussed in detail. The effects of leading factors (temperature, time, DTB concentration, and the catalyst amount) on the sulfur conversion during the ODS reaction were studied in detail. As a final point, the reusability of the catalyst and applicability of the oxidation system for real fuel desulfurization are also reconnoitered.

Materials and Methods

Materials

All the chemicals and reagents used in this study were of analytical grade; sodium tungstate dihydrate (Na2WO4·2H2O), melamine (C3H6N6), and dibenzothiophene (C12H8S) were acquired from Sigma Aldrich Co., with hydrazine hydrate (N2H4·H2O) and nickel nitrate (Ni(NO3)2) from Merck. Regarding the ODS activity, model fuel (dibenzothiophene (DBT)) and real fuel (kerosene and diesel oil) were used in this study.

Methods

Synthesis of Materials (g-C3N4, Ni-WO3, and Ni-WO3@g-C3N4 Composite)

The preparation of g-C3N4 was done via a simple decomposition method as provided in a previous study with some minor modifications.[24] Five grams of melamine was put into an aluminum crucible and heated at 550 °C in a muffle furnace (air atmosphere) for 3 h. The obtained g-C3N4 support material was ground via a pestle and mortar.

A solid solution of Ni-WO3 was synthesized via hydrothermal treatment with a precursor molar ratio (1:1) as described subsequently. Typically, 5 mmol of sodium tungstate dihydrate (Na2WO4·2H2O) and nickel nitrate (Ni(NO3)2) was added into 100 mL of deionized water with vigorous stirring for 20 min at 30 °C. After that, 500 μL of hydrazine hydrate (N2H4·H2O) was added dropwise in the above solution and taken into a Teflon-assisted stainless-steel autoclave. The mixture was heated at 150 °C for 6 h and cooled down to room temperature, and the resulting product was centrifuged, washed with water (three times), and dried in an oven at 110 °C.

Different percentages of Ni-WO3@g-C3N4 were prepared by varying the amount of Ni-WO3 via a hydrothermal method. Weight percentages (20%) of the Ni-WO3@g-C3N4 composite were synthesized by mixing 200 mg of prepared g-C3N4 with a specific amount of Ni-WO3 in a Teflon-assisted stainless-steel autoclave and heated at 150 °C for 3 h. The prepared materials were calcined at 400 °C in a muffle furnace with a heating rate of 5 °C/min. The synthesized composite (Ni-WO3@g-C3N4) was cooled down to room temperature and ground with a pestle and mortar. Moreover, we also used Ni-WO3 and g-C3N4 for oxidative desulfurization activity for comparison purposes.

Oxidative Desulfurization Process (ODS)

Model fuel was prepared by dissolving a specific quantity of DBT (e.g., 100 mg/L) into n-hexane (50 mL) as a source of sulfur; acetonitrile (30 mL as an extractant) and 0.1 g of (20% Ni-WO3@g-C3N4) were added into a flask and placed in the dark for 20 min for adsorption–desorption equilibrium. Afterward, a known amount of (e.g., 1 mL of 30%) H2O2 was injected to the flask and stirred vigorously at a specified temperature and time (e.g., 40 °C for 60 min). The dibenzothiophene sulfone and sulfoxide formation was extracted in the acetonitrile phase that was investigated by thin-layer chromatography (TLC), the remaining concentration of DBT in the n-hexane phase was analyzed via a UV–vis spectrophotometer (at λmax = 283 nm, n–π*), and total sulfur was analyzed by PETRA X-ray fluorescence (ASTM D-4294, wt %) at different time intervals. Moreover, the specific gravity (ASTM D-1298 at 15.6 °C), the salt in fuel oil (ASTM D-3230 at ptb), the water content (ASTM D-4006 in wt %), and distillation (ASTM D-86 in °C) were also determined. The percentage efficiency for the removal of the sulfur content from both model and real fuel oil was determined via the given equation (eq )

Instrumentation

XRD patterns of g-C3N4, Ni-WO3, and the 20% synthesized composite Ni-WO3@g-C3N4 were analyzed via powder X-ray diffraction by using Cu Kα of 1.54 Å in the 2θ range of 3–70° with a 2° per min scan rate. For the vibrational band study, FT-IR spectra of prepared materials were collected by an IR-TRACER-100 (4000–400 cm–1), with the surface morphology via scanning electron microscopy (SEM) (NOVA NANO) and elemental analysis via energy-dispersive X-ray spectroscopy (EDX). Moreover, the mass loss (wt %) was also evaluated via thermogravimetric analysis (TGA) in an inert atmosphere in the range of 40–800 °C. To quantify the sulfur components in both model and real fuel, samples were analyzed via PETRA X-ray fluorescence (PETRA-XRF, ppm, ASTM D-4294). Additionally, other fuel properties (water content, specific gravity, salt in fuel, and distillation) were determined via a water content tester (China PT-D4006-8929A, vol %, ASTM D-4006), a hydrometer (g/mL at 15.6 °C, ASTM D-1298), and a distillation tester (PMD 110, PAC, ASTM D-86).

Results and Discussion

FT-IR

FT-IR studies were carried out to investigate the functional groups present in synthesized g-C3N4 and the Ni-WO3@g-C3N4 composite in the range of 4000–400 cm–1 as shown in Figure . The observed vibrational bands at 1630 and 1230 cm–1 indicate the presence of C=N and C–N stretching in g-C3N4, respectively. The characteristic bands observed in the range of 3500–3200 and 1575 cm–1 correspond to N–H stretching and the bending vibrations in g-C3N4, respectively. The other peaks observed at 860, 1320, and 1409 cm–1 are due to the deformation mode of N–H and the bending vibration as reported previously.[25,26] All the peaks remain identical in both (g-C3N4 and the Ni-WO3@g-C3N4 composite) except one peak that is observed only in the Ni-WO3@g-C3N4 composite in the range of 600–450 cm–1 owing to the M–O peak. All the observed characteristic vibrational bands give confirmation of the successful synthesis of g-C3N4 and Ni-WO3@g-C3N4.

Figure 1

IR vibrational analysis of graphitic carbon nitride (g-C3N4) (a) and nanocomposite Ni-WO3@g-C3N4 (b).

XRD Pattern

XRD patterns of g-C3N4, Ni-WO3, and the synthesized composite Ni-WO3@g-C3N4 were analyzed via powder X-ray diffraction by using Cu Kα of 0.154 nm, in the 2θ range of 3–70° with a 2° per min scan rate. In the XRD spectra of g-C3N4 (Figure a), there are two characteristic peaks that appeared at 2θ = 13.43 and 27.40° corresponding to the basal planes of (100) and (002), respectively, as matched with JCPDS card no. 01-087-1526 as well as in good agreement with a reported study.[27] The basal plane (002) is attributed to the interlayer stacking of conjugated aromatic systems in g-C3N4, which corresponds to the flake-like structure with a d-spacing of 0.33 nm, and the (001) plane is due to ordering of tri-s-triazine units. In the XRD pattern of Ni-WO3 (Figure b), the characteristic peak positions with the crystal planes are as follows: 2θ = 14.19 (100), 22.78 (001), 24.78 (110), 29.13 (200), and 36.5° (201) are in good agreement with JCPDS card no. 033-1387 due to the phase of WO3.[28] Similarly, the XRD pattern of Ni-WO3@g-C3N4 (Figure c) exhibits 2θ = 14.21 (100), 22.77 (001), 24.65 (110), 29.23 (200), and 36.61° (201). Both (g-C3N4 and Ni-WO3) peaks are observed in the spectra of Ni-WO3@g-C3N4, which revealed that the Ni-WO3@g-C3N4 composite was synthesized successfully.[29] The XRD patterns of both pristine Ni-WO3 and Ni-WO3@g-C3N4 samples indicate no peaks of the new phase that can be detected. However, the intensity of planes becomes lower in the Ni-WO3@g-C3N4 composite pattern due to the presence of Ni -WO3. The average crystallite size was calculated via the Scherrer equation (eq ):where D is the average crystallite size, k is the Scherrer constant (0.9), λ equals 0.154 nm, β is the full width at half-maximum (FWHM), and θ is the angle of reflection. The calculated average crystallite sizes of g-C3N4, Ni-WO3, and Ni-WO3@g-C3N4 are 14.2, 55, and 50.4 nm, respectively.

Figure 2

XRD spectra of (a) g-C3N4, (b) Ni-WO3, and (c) Ni-WO3@g-C3N4.

SEM and EDX

The surface morphology of g-C3N4 and Ni-WO3@g-C3N4 was analyzed via a NOVA NANO SEM. It is evident from the SEM image in Figure a that the surface of the g-C3N4 material is noticed to be a wrinkled sheet and solid agglomerates, which formed a stacked structure. It can be seen that Ni-WO3 particles are well-decorated on the g-C3N4 surface with an interesting morphology (Figure c–f). The Ni-WO3 particles appeared on the surface with a proper geometry (spherical particles) rather than wrinkled sheets. Figure a shows the elemental analysis of g-C3N4 and the Ni-WO3@g-C3N4 composite collected from energy-dispersive X-ray spectroscopy (EDX). Figure b indicates the presence of Ni and W loaded onto g-C3N4 as absent in the EDX spectra of simple g-C3N4. The elemental composition in g-C3N4 is 34.85 wt % C and 65.15 wt % N and in Ni-WO3@g-C3N4 is 37.34 wt % C, 30.79 wt % N, 13.28 wt % O, 4.93 wt % Ni, and 13.66 wt % W. The elemental mappings of C, O, N, Ni, and W in Ni-WO3@g-C3N4 are shown in light green, red, dark green, purple, and yellow colors, respectively. These EDX peaks confirm the loading of Ni-WO3 onto g-C3N4 as the morphology is also different after loading.

Figure 3

SEM micrographs of g-C3N4 (a,b) and Ni-WO3@g-C3N4 (c–f).

Figure 4

EDX of g-C3N4 (a) and Ni-WO3@g-C3N4 composite (b) and EDX mapping of C, O, N, Ni, and W in the composite.

Thermogravimetric Analysis (TGA)

To quantify the mass loss (%), the prepared materials have been examined by TGA (at 40–800 °C) under an inert atmosphere. It can be seen from Figure that there is a slight decrease in weight loss (%) in g-C3N4 below 200 °C (18%) due to volatilization of adsorbed water or other impurities. The significant weight loss observed after 600 °C (72%) corresponds to the lower thermal stability of g-C3N4 as compared to the nanocomposite. In the Ni-WO3@g-C3N4 composite curve, a 5% mass loss occurred below 200 °C due to dehydration of water. Condensation of melamine and release of ammonia gas lead to a 9% loss after 400 °C in the nanocomposite. Pure g-C3N4 has a larger weight loss than the composite in the range of 500–700 °C that revealed that the nanocomposite is more efficient for catalytic activity owing to thermal stability.[30]

Figure 5

TGA mass loss curves of g-C3N4 and Ni-WO3@g-C3N4.

Optimization of DBT Removal from Model Fuel Oil

In order to check the catalytic activity of the prepared catalyst (20% Ni-WO3@g-C3N4), multiple parameters have been optimized. Temperature also played a crucial role in the oxidative desulfurization system. The effect of time and temperature on DBT removal (%) is shown in Figure a at different times (30–180 min) and temperatures (25–40 °C) by taking 0.1 g of the synthesized catalyst, 1 mL of an oxidant, 50 mL of DBT solution (100 mg/L), and 30 mL of acetonitrile as an extractant. More than 90% DBT removal was achieved using 180 min of time. It was assumed that DBT removal might be kinetically limited at low temperatures and increased with an increase in the temperature until 40 °C.[31] Beyond 40 °C, no significant change has been observed, which might be due to decomposition of H2O2 at high temperatures. The maximum efficiency (97%) was attained at 40 °C in 180 min, and this time was optimized for further studies.

Figure 6

Effect of the temperature (a), catalyst amount (b), DBT concentration (c), and amount of H2O2 (d) on desulfurization.

The effect of different catalysts (bare g-C3N4, Ni-WO3, and Ni-WO3@g-C3N4) were investigated keeping other parameters constant (0.1 g of the catalyst, 50 mL of DBT solution (100 mg/L in n-hexane), 1 mL of oxidant (H2O2), a time duration of 180 min, and a temperature of 40 °C). It can be seen that g-C3N4 shows an about 30% removal efficiency, which might be closely related to the adsorption phenomenon to some extent. Ni-WO3 and 5, 10, and 20% Ni-WO3@g-C3N4 activities for DBT removal were compared in which Ni-WO3 supported onto g-C3N4 showed better results than Ni-WO3. Overall, 20% Ni-WO3 loaded onto g-C3N4 shows the maximum desulfurization efficiency as shown in Figure b providing maximum active sites for ODS. Additionally, the catalyst amount has direct relation with DBT removal (%).[32]

The effect of DBT concentration on desulfurization of fuel oil was monitored by varying the DBT concentration (50, 100, 200, and 400 mg/L) with other parameters kept constant (0.1 g of the catalyst, a volume of 50 mL, 1 mL of the oxidant (H2O2), a time duration of 180 min, and a temperature of 40 °C). Figure c shows that as we increased the DBT concentration from 50 to 400 mg/L, the removal efficiency decreased. The removal efficiency is affected, as it limits the available catalytic site of Ni-WO3@g-C3N4 at a high DBT concentration, but it is still able to work with a lower efficiency.[32] When a lower concentration of DBT is present, the DBT removal is faster as a higher number of sites are present. It was concluded that the greater is the presence of catalytic sites, the lower is the concentration of DBT molecules and the more effective will be the desulfurization done at optimized conditions.

The effect of H2O2 as an oxidant on desulfurization has been investigated by varying the amount of the oxidant (0.5, 1, 2, and 3 mL) keeping other parameters constant. It was observed that >90% DBT removal can be achieved by using 1 mL of H2O2 as an oxidant as shown in Figure d. Moreover, no significant increase in desulfurization has been observed by increasing the H2O2 amount. Hence, 0.1 g of the catalyst (Ni-WO3@g-C3N4) can be effectively used to remove (>97%) 100 mg/L DBT at 40 °C within 180 min.

Kinetics, Thermodynamics, and Mechanism of Oxidative Desulfurization via Ni-WO3@g-C3N4

The kinetic study of DBT removal via the ODS process was examined by applying a pseudo first-order kinetic model.[31,32] The plot of CF/Ci vs time (min) at different temperatures is shown in Figure a, and the regression coefficients (R2) were found to be in the range of 0.75–0.85, where Ci and CF are the initial and final concentrations of DBT. The results depict that DBT removal by using Ni-WO3@g-C3N4 followed a pseudo first-order kinetic model with rate constants (k, min–1) of 0.0074, 0.0078, 0.0082, and 0.0082 at 25, 30, 35, and 40 °C, respectively. Additionally, the activation energy (Ea) of the ODS process was calculated via the Arrhenius plot (1/T vs ln k, Figure b) as previously reported, and the Ea for the current study is 6.14 kJ/mol.

Figure 7

Plot of CF/Ci vs time (min) for DBT removal via the ODS process (a) and 1/T vs ln k (b).

To understand the temperature effect on DBT oxidation, standard Gibbs free energy (ΔG),standard entropy (ΔS), and standard enthalpy (ΔH°) were determined by using the Eyring equation[33,34] as given below:where Kc is the equilibrium constant and T is the temperature in Kelvin. The standard Gibbs free energy was calculated using the following equation:[35]

The obtained results showed that positive ΔH and positive ΔS from the plot of 1/T vs ln Kc correspond to the endothermic process and randomness in reaction media, respectively, as given in Table . The negative value of ΔG revealed that DBT removal via Ni-WO3@g-C3N4 is a spontaneous and thermodynamically feasible ODS process, and a more negative value has been observed at 313 K.

Table 1

Thermodynamic Parameters for DBT Removal via ODS

	ΔG (kJ mol-1)
catalytic system	298.5 K	303.5 K	308.5 K	313.5 K	ΔH (kJ mol-1)	ΔS (kJ mol-1 K-1)
Ni-WO3@g-C3N4/H2O2	–0.014	–0.024	–0.044	–0.071	354	0.004

Overall, the remarkable activity of the catalyst for DBT removal is the result of morphological effects such as the crystallite size and the surface area. Based on experimental evidence and the optimization study, the proposed mechanism for ODS of DBT-based model fuel is shown in Figure . It is a three-step mechanism as follows: (a) formation of metal oxo-peroxo species by attack of H2O2 nucleophiles, (b) loss of water molecules, and (c) formation of first sulfoxides and then sulfones. In the presence of a graphitic carbon nitride-based nanocomposite and hydrogen peroxide, a simple chemical process implied that oxygen radicals from hydrogen peroxide attack the electron-rich sulfur atom of DBT molecules resulting in the formation of sulfoxides and then sulfones. Additionally, the metal oxo-peroxo species are generated by the reaction of an oxidizing agent with the catalyst. The nucleophilic attack of peroxo species on sulfur changed it into sulfoxides and corresponding sulfones.[36,37]

Figure 8

Possible mechanism of ODS for DBT.

Desulfurization Optimization through the Response Surface Methodology (RSM)

For the purpose of theoretical optimization, the RSM was used for optimizing the optimum reaction parameters and mathematically predicting the efficiency of desulfurization of fuel oil. The four input variables include the (a) catalyst concentration (in g), (b) DBT concentration (in mg/L), (c) oxidant concentration (in mL), and (d) time duration (in min), and the efficiency of desulfurization was taken as an output variable. The Box–Behnken design (BBD) technique was followed for the optimization purpose, and in total, 29 experimental runs were carried out for varied combinations of input variables.

The above equation is the coded equation of the prediction model for calculating the efficiency of desulfurization. From the equation, it can be inferred that the catalyst concentration, oxidant concentration, and time duration had a positive effect, while the DBT concentration had a negative effect on the overall efficiency. From surface plots, it can be noted that efficiency increased with an increase in the catalyst concentration, oxidant concentration, and time duration (explain this phenomenon). On the other hand, the efficiency decreased with an increase in DBT concentration (Figure ). The maximum efficacy achieved 99% for the optimum parameters. From ANOVA (Table ), it can be noted that the prediction model is highly significant with its lack of fit being not significant, thereby confirming the model to be highly reliable. Supporting this, the R2 of the model was calculated to be 0.965 and its adjusted R2 to be 0.9301, which signifies that the accuracy of the developed model will be nearly 97%. In other words, the average standard deviation between the experimental and predicted values will be around 1.3%. Figure (predicted vs actual) illustrates the plot between the experimental and theoretical values of the developed model.

Figure 9

RSM 3D graphs presenting the effect of various factors on desulfurization of fuel oil.

Table 2

Analysis of ANOVA from Obtained Results

source	sum of squares	Df	mean	F-value	p-value	comments
model	656.9119	14	46.92228064	27.6046	9.09 × 10–8	significant
A-catalyst	275.8084	1	275.8084083	162.2594	4.33 × 10–9
B-DBT conc.	45.04688	1	45.046875	26.50129	0.000148
C-oxidant conc.	12.04003	1	12.04003333	7.083209	0.018609
D-time duration	33.60053	1	33.60053333	19.76735	0.000554
AB	0.1936	1	0.1936	0.113896	0.740758
AC	3.8809	1	3.8809	2.283152	0.153025
AD	0.055225	1	0.055225	0.032489	0.859541
BC	1.199025	1	1.199025	0.705392	0.415091
BD	24.8004	1	24.8004	14.59019	0.001876
CD	2.512225	1	2.512225	1.477954	0.244198
AÂ2	167.7335	1	167.7335353	98.67844	1.01 × 10–7
BÂ2	31.75715	1	31.75714883	18.68288	0.000702
CÂ2	84.5989	1	84.5988995	49.76994	5.73 × 10–6
DÂ2	94.8476	1	94.84760221	55.7993	3.01 × 10–6
residual	23.79719	14	1.699799167
lack of fit	21.55111	10	2.155110833	3.837995	0.103388	not significant
pure error	2.24608	4	0.56152
cor total	680.7091	28
std. dev.	1.303763		RÂ2	0.965041
mean	92.10552		adjusted RÂ2	0.930081
C.V. %	1.415511		predicted RÂ2	0.812484
			acceptable precision	16.57495

Reusability of the Catalyst

We have also tested the prepared materials for reusability and regeneration processes under similar conditions (0.1 g of the catalyst, 50 mL of 100 mg/L DBT solution, 30 mL of an extractant, 1 mL of an oxidant, and 180 min at 40 °C). At the end of each oxidation reaction, the catalyst was separated by a simple filtration method, washed with dichloromethane (CH2Cl2), and dried in an oven at 90 °C. The obtained findings revealed that no significant decrease in efficiency has been observed after two cycles (only 1%) and it decreased to some extent (about 5–6%) after further three cycles as shown in Figure . It means that the prepared material shows promising reusability and high stability. The catalytic activity might be decreased after many cycles due to a decrease in active sites as well as the mass loss of the catalyst.

Figure 10

Desulfurization efficiency of Ni-WO3@g-C3N4 in multiple cycles.

Oxidative Desulfurization (ODS) of Real Fuel Oil

The oxidative desulfurization (ODS) of commercially accessible fuel samples (diesel and kerosene) was also done by using the Ni-WO3@g-C3N4 nanocomposite and H2O2 as an oxidant under optimized conditions (0.2 g of the catalyst, 3 mL of an oxidant, 50 mL of the fuel sample, and a time duration of 180 min at 40 °C). PETRA X-ray fluorescence (XRF total sulfur analyzer, ASTM D-4294 in ppm) was employed to quantify the amount of sulfur before and after ODS, and the total sulfur contents in diesel oil and kerosene were 4630 and 1635 ppm, respectively. Additionally, other fuel properties were also examined such as the water content (ASTM D-4006), specific gravity (ASTM D-1298), and distillation (ASTM D-86) as given in Table . The obtained findings revealed that 89.5 and 91.2% sulfur elimination was achieved for diesel and kerosene, respectively, and other fuel properties are almost similar. Moreover, pure g-C3N4 was also used for the same purpose as it also shows efficiency to some extent, which might be due to the adsorption phenomenon, as well as oxidative properties and is not comparable with the composite.

Table 3

ODS of Kerosene and Diesel Oil via Ni-WO3@g-C3N4

		diesel oil		kerosene oil
tests	method no.	before ODS	after ODS	before ODS	after ODS
specific gravity (g/mL at 15.6 °C)	ASTM D-1298	0.879	0.875	0.833	0.831
total sulfur by PETRA X-ray fluorescence (ppm)	ASTM D-4294	4630	474	1635	143
water content by distillation (vol %)	ASTM D-4006	nil	nil	nil	nil
distillation (°C)
50%	ASTM D-86	285	281	237	236
90%		345	343	303	301

Comparison with Other Reported Methods

Considering inorganic–organic hybrid materials, limited studies are available for DBT removal through the oxidative desulfurization route. We cannot make comparison effectively for ODS of DBT and real fuel oil, but some of the available studies have been compared and are given in Table .

Table 4

Comparison with Previously Reported Studies (g-C3N4-Based Catalysts)

material	sulfur species	oxidant	DBT concentration (mg/L)	efficiency (%)	ref.
g-C3N4	DBT	TBHP	500	<10	(38)
MoO2/g-C3N4			500	100
Fe-CN-NHχ	H2S		5000	∼100	(15)
WO3/g-C3N4	DBT	H2O2	500	91.2	(39)
P-doped g-C3N4 nanosheet-supported V-based nanocomposites (V/P-CN)	DBT	TBHP	500	∼100	(40)
C14PPMo/g-C3N4	DBT	H2O2	500	∼100	(41)
	4,6-DMDBT		250	94.8
X-MoO2/g-C3N4	DBT	H2O2	500	96	(42)
NiW/rGO	DBT	H2O2	500	>90	(43)
CoW/rGO			500	∼100
CoO/g-C3N4	DBT	H2O2		86.7	(44)
Ni-WO3@g-C3N4	DBT	H2O2	100	97	present study
	diesel		4630	89.5
	kerosene		1635	91.5

Conclusions

A highly efficient and eco-friendly material has been synthesized and used for the first time for oxidative desulfurization of DBT model fuel as well as real fuel (diesel and kerosene oil). XRD results depict that the average crystallite size of Ni-WO3@g-C3N4 is 50.36 nm and it has a monoclinic phase. The SEM morphology of prepared materials indicates that spherical particles are well-decorated on the surface of g-C3N4. The prepared materials showed outstanding performance for DBT removal from model fuel with 97% and real fuel (diesel with 89.5% and kerosene with 91.5%) via ODS. Multiple factors have been applied to optimize the process in which the time, the catalyst amount, and the oxidant amount have direct relation and DBT concentration has indirect relation with removal efficiency. This ODS process followed pseudo first-order kinetics, and a negative value of ΔG showed the spontaneity in the reaction system. Moreover, response surface methodology (RSM) 3D plots based on the Box–Behnken design were used to optimize the desulfurization of fuel oil. The nanocomposite showed promising reusability up to five times with no significant change in desulfurization efficiency. Thus, the entire study confirms the prominence of the prepared nanocomposite for efficient production of sulfur-free oil.