Ultrasound-assisted oxidative desulfurization of model fuel oil using HPW/Al2O3/PANI composite: experimental and kinetic modeling Authors: Masoumeh Ezati, Mohammad Reza Khosravi-Nikou, Reza Mosayebi Behbahani Published: 2025 ABSTRACT AbstractIn the present study, a novel composite catalyst HPW/Al2O3/PANI (γ-Al2O3, phosphotungstic acid (HPW), polyaniline (PANI)) was synthesized via the ultrasonic-assisted method. It was thoroughly characterized using XRD, FTIR, BET-BJH, FE-SEM, EDS, and TGA techniques. The as-synthesized catalyst was applied for ultrasound-assisted oxidative desulfurization to remove benzothiophene and dibenzothiophene from model hydrocarbon fuel. Response surface methodology (RSM), employing the Box–Behnken design (BBD), was utilized to systematically evaluate and optimize the influence of key process variables, including temperature, oxidant/sulfur (mol/mol), and catalyst loading (g/L). It was found the maximum removal of sulfur containing compounds was achieved through oxidant/sulfur (mol/mol.): 4.15, reaction temperature: 52.85 °C and catalyst/fuel (g/L): 3.26. Kinetic analysis demonstrated that the UAOD process followed the pseudo-first-order reaction kinetic model. Furthermore, the catalyst exhibited excellent reusability over five successive cycles, demonstrating high structural stability and potential for practical applications.Keywords: Oxidative desulfurization, Ultrasound-assisted catalysis, Heteropoly acid, Polyaniline, Kinetics, Clean fuel, Response surface methodologySubject terms: Chemistry, Engineering, Materials science Introduction Stringent environmental regulations have necessitated a substantial reduction in the sulfur content of petroleum-derived fuels, often to ultra-low levels, to reduce the adverse effects of sulfur oxides on air quality and human health1–4. While this shift contributes significantly to environmental protection, it poses substantial technical and economic challenges to the refining sector. Hydrodesulfurization (HDS) remains the dominant industrial process for sulfur removal, relying on high-pressure catalytic hydrogenation to convert organosulfur compounds into hydrogen sulfide5–7. However, certain refractory sulfur species, particularly DBT and its alkylated derivatives, strongly resist HDS due to their steric hindrance and localized electron density around the sulfur atom8. To address these limitations, alternative and complementary desulfurization techniques have been actively explored, including extractive desulfurization (EDS)9,10, adsorptive desulfurization (ADS)11–13, bio-desulfurization (BDS)14–16, and oxidative desulfurization (ODS)17,18. Among these, ODS has gained prominence as a highly effective approach for deep sulfur removal, chiefly due to its capability to function under mild operational conditions without the necessity for hydrogen gas19, making it cost-effective and operationally feasible for downstream applications17,20,21. The fundamental mechanism of ODS involves the catalytic oxidation of sulfur-containing compounds into their corresponding sulfones22–24, which are more polar25 and thus more amenable to removal via extraction using polar solvents such as acetonitrile or methanol26,27 or through adsorption using solid materials like activated carbon28–30. A wide range of oxidants has been studied for use in ODS systems, including peracids31, peroxy salts32, hydroperoxides33, ozone34, and molecular oxygen35. Among these, hydrogen peroxide (H2O2) is particularly attractive due to its high oxidative potential, low cost, and environmentally benign byproducts (i.e., water and oxygen)36–39. Despite these advantages, one of the key challenges of H2O2-based ODS lies in the relatively slow oxidation rate, often attributed to the biphasic nature of the system, where limited interfacial mass transfer between the aqueous oxidant phase and the organic fuel phase hinders overall efficiency. Extensive efforts have been made to investigate effective catalysts to overcome this kinetic barrier and enhance oxidation rates40–42. Heterogeneous catalysts have emerged as a cornerstone in the advancement of ODS technologies, owing to their ability to significantly enhance mass transfer between immiscible phases and facilitate efficient post-reaction separation, thus contributing to the overall economic and operational viability of the process43. Among various strategies employed to improve catalytic performance, the immobilization of heteropoly acids (HPAs) onto metal oxides garnered considerable interest owing to their outstanding oxidative capabilities and the tunable catalytic behavior44,45. Metal oxides such as SiO2, TiO2, and Al2O3 have been frequently utilized as catalyst supports, with Al2O3 standing out due to its exceptional physicochemical characteristics. Notably, the surface of Al2O3 can interact effectively with thiophenic sulfur compounds through coordination with the sulfur atom’s lone electron pair, enhancing adsorption and catalytic conversion44,46. The extensive utilization of Al2O3 can be attributed to its excellent thermal stability, structural flexibility, economic feasibility, and advantageous surface properties, which make it an attractive material for robust heterogeneous catalytic systems47,48. In parallel, polyoxometalates (POMs), a class of discrete metal–oxygen clusters, are receiving heightened interest in scientific research, due to their exceptional molecular structures, high oxygen density, and inherent anionic nature, which collectively contribute to their superior oxidative capabilities49. Among the various POM configurations, Keggin-type structures have demonstrated outstanding catalytic activity across numerous liquid-phase oxidation reactions, positioning them as highly versatile and efficient oxidative agents50. Other notable structural forms include Dawson–Anderson51, Preisler52, and sandwich-type clusters53. However, the intrinsic solubility of many POMs in polar solvents poses a significant challenge for their recovery and reuse, limiting their practical deployment in heterogeneous catalysis54. To mitigate this drawback, immobilization of POMs on solid supports has been proposed as a viable strategy for enhancing catalyst stability and reusability. Furthermore, more recently, hybrid catalysts that combine inorganic supports with functional polymers have emerged as a promising direction in ODS research. Surface modification using conductive polymers has shown promise in augmenting the mechanical, physicochemical, and electronic attributes of the composite catalyst. PANI, in particular, has gained attention due to its favorable redox characteristics, environmental stability, economic monomer source, and facile synthesis55–58. These features make PANI an attractive scaffold for incorporating HPAs, ultimately yielding hybrid materials with enhanced catalytic efficiency, durability, and multifunctionality. Beyond the application of heterogeneous catalysts, ultrasonic irradiation has proven to be a highly effective technique for intensifying ODS processes, particularly by enhancing the interfacial contact between immiscible aqueous and organic phases59,60. The underlying mechanism involves acoustic cavitation, wherein the formation, growth, and implosive collapse of microbubbles generate both mechanical agitation and localized high-energy microenvironments61. These phenomena yield both physical and chemical catalytic benefits: physically, they facilitate the dispersion of reactants by producing microemulsions that significantly enhance interphase mass transfer; chemically, the collapse of cavitation bubbles leads to the in situ formation of reactive radical species, which drive the oxidative transformation of sulfur-containing compounds at accelerated rates62,63. In this context, the present study explores the synergistic integration of a novel organic–inorganic hybrid catalyst and ultrasonic energy to advance the efficiency of the ODS process. A polyaniline-supported heteropolyacid composite (HPW/Al2O3/PANI) was synthesized and systematically characterized for its physicochemical properties and catalytic potential. The hybrid material was subsequently utilized in the ultra-deep oxidative desulfurization (UAOD) of a model fuel containing BT and DBT, with H2O2 acting as an environmentally benign oxidant. A BBD design coupled with RSM was utilized to assess the influence of key operational parameters and their interactions on sulfur removal efficiency. BBD was specifically chosen because it efficiently explores the interaction effects among multiple variables with a relatively small number of experimental runs, reducing time and cost while providing sufficient information to develop accurate predictive models. This approach also allows the identification of optimal conditions for the UAOD process in a systematic and statistically robust manner64. Furthermore, the influence of ultrasonic irradiation on the kinetic parameters, including the apparent rate constants and Arrhenius activation energy components, was investigated to provide mechanistic insight into the enhanced reactivity observed under sonochemical conditions. Experimental methods Chemicals and reagents N-octane, BT, and DBT were supplied by Merck Company, Germany. H2O2 (30 wt.% aqueous solution) was purchased from Sumchun (Korea). Acetonitrile (99.8%) was supplied by Sigma-Aldrich. Aniline, ammonium persulfate (APS), and p-toluene sulfonic acid (PTSA) were purchased from Sinopharm Chemical Reagent Co. Ltd., China. γ-Al2O3 was purchased from Fluka Chemicals Ltd. HPW was purchased from Lancaster. All chemicals employed were of analytical grade and utilized without any additional purification. Catalyst preparation and characterization techniques Synthesis of HPW/Al2O3 The HPW/Al2O3 sample was synthesized via the incipient wetness impregnation approach. Initially, 0.6 g of HPW was dissolved carefully in calculated amounts of distilled water at room temperature. Then, 2g of γ-Al2O3 was added slowly to this solution. The obtained slurry was stirred at a constant rate of 300 rpm at room temperature for 1h. After drying at 110 °C for 17 h, followed by calcination at 400 °C for 3h, the HPW/Al2O3 sample was obtained. The synthesis steps are shown in Fig. 165. Synthesis steps of HPW/Al2O3/PANI composite. Synthesis of HPW/Al2O3 with PANI Initially, ammonium persulfate, p-toluenesulfonic acid, and a predetermined amount of HPW/Al2O3 were ultrasonically dispersed in 300 mL of distilled water at 500 W for 25 min, and the resulting suspension was transferred to a crystallizing dish placed in an ice–water bath. Subsequently, a solution of aniline (1.75 g) in 60 mL of distilled water was slowly added to the system under continuous magnetic stirring at 900 rpm for 15 min, while maintaining the temperature below 10 °C. The resulting mixture was subsequently subjected to ultrasonic irradiation (500 W) for 40 min and then allowed to react in an ice–water bath for 5 h. Finally, the product was collected by filtration through a sintered glass funnel and thoroughly washed with deionized water and ethanol until the filtrate became clear. The 0.2% HPW/Al2O3/PANI composite was obtained following a 5 h drying process in an air-forced oven at 80 °C (Fig. 1)66. The synthetic procedure to generate HPW/Al2O3/PANI composite is shown in Fig. 2. The synthetic procedure to generate HPW/Al2O3/PANI composite. Characterization tests The structure of the synthesized catalyst was verified using X-ray diffraction (XRD) patterns at 2θ angles of 10–80° by X’ Pert PRO Diffractometer. Fourier-transform infrared (FTIR) spectroscopy was utilized to identify and characterize the functional groups of the synthesized catalysts, conducted on a Spectrum Two spectrometer (PerkinElmer) within the wavenumber range of 400–4000 cm−1. Nitrogen adsorption–desorption isotherms were utilized to evaluate the specific surface areas of the samples based on the Brunauer–Emmett–Teller (BET) methodology. In addition, the pore size distribution and average pore size were calculated using the Barrett-Joyner-Halenda (BJH) method by BEISORP MINI analyzer from Microtrac Bel Corp. Surface morphological characteristics of the samples were investigated using a Zeiss SIGMA VP field emission scanning electron microscope (FESEM), and their elemental composition was determined via Energy Dispersive X-ray Spectroscopy (EDS). Ultrasound-assisted oxidative desulfurization (UAOD) method In this work, the experimental setup of the ODS process includes a 70 ml double-wall glass reactor with an internal diameter of 3.2 cm, a water circulation system, a condenser, and a thermometer (Fig. 3). Model fuel was made up of n-octane, BT (400 ppm wt.), and DBT (400 ppm wt.). HPW/Al2O3/PANI as the heterogeneous catalyst and H2O2 as oxidant were added to the mixture. Given that the investigation of operating parameters, including the molar ratio of oxidant/sulfur, reaction temperature, and catalyst/fuel, was considered, the values ​​of these parameters for each run were determined by the design of the experiments. In each experimental run, the reactor mixture was subjected to direct ultrasonic irradiation using a titanium probe with a diameter of 12 mm, operating at a frequency of 20 kHz (UIP500hd, Ultrasound Technology, Hielscher Ultrasonics, Germany) and an amplitude range of 50–100%. The probe was immersed approximately 1.2 cm into the reaction mixture during sonication58. Throughout all experiments, the reactor was placed on a heater, and the temperature of the reaction mixture was precisely controlled by a combination of a stable water circulation system and continuous temperature monitoring with a thermometer. After each oxidation test, acetonitrile solvent was added to the fuel at a 1:1 volume ratio to extract the produced sulfones. The mixture was stirred with a magnetic stirrer at 1000 rpm for 15 min, followed by a 10-min settling period to allow the aqueous and organic phases to separate. At predetermined time intervals, aliquots of the fuel phase were extracted using a microsyringe and subsequently analyzed by gas chromatography (GC) to determine the conversion efficiencies of BT and DBT. The sulfur compound removal percentage was determined based on their initial concentration (C₀) and the concentration at time t (Cₜ), calculated using the following equation: Experimental setup of UAOD process. To evaluate the contribution of individual components to the desulfurization process, a series of blank and control experiments were performed under optimized operational conditions. Initially, the effect of the oxidant (H2O2) alone was examined. Subsequently, experiments were carried out in the presence of acetonitrile as a solvent, and separately with the catalyst alone, to assess their independent roles. The catalytic activities of HPW/Al2O3 and HPW/Al2O3/PANI were then compared with those of γ-Al2O3 under identical conditions. In addition, formic acid was employed in combination with H2O2 in selected runs to investigate the role of in-situ peracid formation in facilitating the oxidation of organosulfur compounds. To evaluate the recoverability and reusability of the catalyst, after the extraction of sulfones using acetonitrile, the catalyst was first separated by centrifugation. It was then dispersed in ethanol and stirred at 60 °C for 50 min. Afterward, the catalyst was rinsed with deionized water and dried in an oven at 100 °C. Experimental design Based on the literature review58, parameters affecting the process, including temperature (°C), oxidant/sulfur ratio (mol/mol), and catalyst/fuel ratio (g/L), were selected as the key operational variables for investigation. To save time and money, as well as for a comprehensive analysis of the effect of the mentioned parameters on the efficiency of the UAOD process, an experimental design with BBD matrix was created; range and levels of independent variables in coded and actual units are given in Table 158. Table 1The levels of the independent variables in BBD.VariableSymbolLevel − 10 + 1Molar ratio of oxidant/sulfurX123.55Reaction temperature (°C)X2405060Catalyst (g)/fuel(L)X3135 The levels of the independent variables in BBD. In BBD design, the total number of experiments was determined according to Eq. (2): where N represents the total number of experiments, K denotes the number of independent variables, and C₀ is the number of central points64. Among the statistical techniques applied for data analysis, simple and multiple linear regression are the most commonly employed across a wide range of scientific fields. In regression analysis, the relationships between variables are examined to determine whether one variable can influence another. In other words, given one or more independent variables, a regression equation can be formulated and used to estimate the outcomes of the response variable based on the model inputs. In this work, a second-order model (Eq. 3) was utilized to systematically explore and quantify the complex interactions between the independent variables and the response variable: In this model, Ym represents the predicted response, k denotes the number of independent variables, and xi corresponds to the coded values of these variables. The term β₀ signifies the intercept, while bi, xii, and xii correspond to the linear, quadratic, and interaction coefficients, respectively. The symbol ε accounts for the random error term66,67. The influence of individual factors and their interactions on the response is determined by analyzing both the magnitude and the sign of the corresponding regression coefficients. The adequacy and predictive capability of the fitted model were assessed by calculating the coefficients of determination, R2, and adjusted R2 (R2_adj). The mathematical expressions used to compute these coefficients are presented in Eqs. (4) and (5). In these equations, m is the number of unknown parameters, n is the number of experiments, and SS is the sum of the squares. From a statistical point of view, the high coefficient of determination alone is not enough to diagnose the model’s accuracy. The acceptability and accuracy of the obtained models should also be investigated by analysis of variance (ANOVA). After the validity of the models was confirmed, the equation Y was applied to create two-dimensional (2D) and three-dimensional (3D) response surface plots. Results and discussion Characterization tests Figure 4a–c shows the XRD patterns of pure γ- Al2O3 and modified samples. As can be seen, in the pattern of the γ-Al2O3 sample, three peaks at 2θ = 38°, 47°, and 66° are observed, which correspond to the (311), (400), and (440) crystal planes of the sample68. The X-ray diffraction (XRD) pattern of HPW/Al2O3 exhibits a similarity to that of alumina, with no discernible peaks. This absence of characteristic reflections indicates that HPW is highly dispersed on the surface of the γ-Al2O3 sample, forming an amorphous surface compound69. After modifying HPW/Al2O3 with PANI, an additional peak at 2θ = 27° is observed, which is assigned to the characteristic peak of PANI (110)52. XRD patterns of (a) γ-Al2O3, (b) HPW/Al2O3, and (c) HPW/Al2O3/PANI samples. The functional groups and characteristic bonds of γ-Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples were investigated using FTIR in the 4000–400 cm-1 (Fig. 5a–c). The vibration bands observed at approximately 1627 and 3445 cm−1 in all FT-IR spectra are attributed to the bending and stretching vibrations of water molecules adsorbed on the surface70. The broad absorption band in the FTIR spectrum in the wave number range from 1000 cm−1 to 400 cm−1 was the characteristic vibration of γ-Al2O371. In the HPW/Al2O3 sample, the characteristic Keggin-structure peaks of HPW are also observed, despite substantial overlap with the alumina bands. Specifically, the P–O stretching vibration appears around 1080 cm−1, and the terminal W = O stretching vibration appears near 970 cm−1, confirming the presence of the Keggin structure within the composite. The FTIR spectrum of HPW/Al2O3/PANI shows all characteristic peaks of γ-Al2O3 with some slight spectral intensity changes. The absorption peaks at 1496 cm-1 and 1307 cm-1 are attributed to quinoid (Q) and benzenoid (B) ring structures of PANI, respectively; and the band located at 1131 cm−1 corresponds to the C–H bending vibrational mode72,73. FT-IR analysis of catalyst samples: (a) γ-Al2O3, (b) HPW/Al2O3, and (c) HPW/Al2O3/PANI. Nitrogen adsorption–desorption isotherm curves of γ-Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples are shown in Fig. 6a. In all cases, it is apparent that the powders exhibit classical type-IV isotherm curves with a H4 hysteresis loop, indicating the occurrence of capillary condensation in the pores74. The adsorption isotherms of these samples exhibited a gradual uptake at low to moderate relative pressures, transitioning to a rapid increase at higher relative pressures, characteristic of capillary condensation in mesoporous structures. As can be seen, this rapid increase in pure alumina occurs at approximately a relative pressure of 0.8. After modifying the sample with HPW and PANI, the inflection points have been shifted to lower values. Distribution curves of pore diameters of Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples, derived from the adsorption curves of the isotherms using the BJH method, are presented in Fig. 6b. All samples exhibited characteristic mesoporous structures. (a) N2 adsorption–desorption isotherms and (b) pore size distributions of γ-Al2O3, HPW/ Al2O3, and HPW/Al2O3/PANI samples. Table 2 presents the findings from the BET-BJH characterization of γ-Al2O3 and the modified samples. As demonstrated, the pure γ-Al2O3 sample has a surface area and mesopore volume of 156 m2/g and 1.17 cm3 g-1, respectively. The surface area of the HPW/Al2O3 sample is lower than that of the pure γ-Al2O3 due to the deposition of the HPW on the surface. After the modification of the catalyst with polyaniline, the surface area and pore volume have been significantly reduced. This observation confirms the adhesion of PANI particles within the channels58. Based on the results of the BET analysis, the pore size of the synthesized catalyst is in the meso range of 3–10 nm, which is very suitable for the ODS process, based on the results reported in the research38,75. Table 2Results of BET/BJH analysis of γ-Al2O3 and new HPW/Al2O3 and HPW/Al2O3/PANI samples.SampleSurface area from BET analysis, SBET(m2/g-1)Nitrogen pore volume from BJH analysis, Vtotal (cm3g-1)Pore diameter, from BJH analysis (nm)Al2O3155.861.1710.62HPW/Al2O31350.849.93HPW/Al2O3/PANI67.470.315.27 Results of BET/BJH analysis of γ-Al2O3 and new HPW/Al2O3 and HPW/Al2O3/PANI samples. The morphologies of γ-Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples are characterized by SEM images at two different magnifications (Fig. 7a–f). According to Fig. 7a,b, the γ-Al2O3 sample exhibits a spherical-like aggregate morphology. After the addition of HPW, the overall morphology of the particles does not show significant changes, indicating that the deposition of HPW did not alter the surface structure of the support (Fig. 7c,d). After in-situ polymerization and preparation of HPW/Al2O3/PANI composite, the SEM image (Fig. 7e,f) displays the large agglomerated structures. This phenomenon can be attributed to the formation of a polymeric layer around the HPW/Al2O3 particles. The polymer acts as a binding matrix, promoting interparticle interactions and resulting in the formation of larger agglomerates. To determine the elements present in the samples, energy dispersive X-ray analysis (EDS) was used, and the results are given in Fig. 8a–c. As can be seen, the structure of pure γ-Al2O3 (Fig. 8a) is composed of Al and O components. Following modification of the sample with HPW, the presence of phosphorus (P) and tungsten (W) elements was confirmed, as shown in Fig. 8b. The nitrogen (N) element that appeared in Fig. 8c corresponds to PANI. The elemental analysis confirms the successful fabrication of the HPW/Al2O3/PANI catalyst. SEM images of (a,b) γ-Al2O3, (c,d) HPW/Al2O3, (e,f) HPW/Al2O3/PANI samples. Elemental mapping (a) and EDS analysis (b) of HPW/Al2O3/PANI composite. The TEM images (Fig. 9) reveal that the incorporation of HPW onto γ-Al2O3 did not cause significant changes in the overall particle morphology. This is expected since HPW is mainly dispersed as nanoclusters or thin layers on the alumina surface. In contrast, after the introduction of PANI, noticeable morphological changes were observed. These modifications can be attributed to the formation of a polymeric coating layer on the surface, partial filling or reshaping of surface pores. TEM images of (a,b) γ-Al2O3, (c,d) HPW/Al2O3, (e,f) HPW/Al2O3/PANI samples. The thermal stability of the prepared samples was evaluated by thermogravimetric (TGA) and derivative thermogravimetric analysis (DTG). The results are shown in Fig. 10a,b. The TG curves of γ-Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples show a weight loss of 5.46, 8.22, and 7.19%, respectively, which demonstrates the excellent thermal stability of the samples. The DTG curves reveal that all three samples undergo their major weight loss below 400 °C, mainly due to the desorption of surface-adsorbed water and the removal of hydroxyl groups. Pure γ-Al2O3 exhibits the smallest overall weight loss with a shallow DTG minimum, consistent with its simple dehydration process. In contrast, HPW/Al2O3 and HPW/Al2O3/PANI show more pronounced DTG signals with sharper minima, indicating faster mass-loss events arising from the release of protonated HPW species and the decomposition of organic moieties in PANI in addition to water removal. Result of (a) TGA and (b) DGT analysis of γ-Al2O3, HPW/Al2O3, and HPW/Al2O3/PANI samples. Experimental tests Different operating conditions, such as the oxidant/sulfur ratio (mol/mol), reaction temperature, and catalyst/fuel ratio (g/L), influence the efficiency of the ODS process. A series of experiments was planned using the design of experiments by Minitab statistical software V. 19 to evaluate the reactivity of BT and DBT compounds in the UAOD system in the presence of HPW/Al2O3/PANI catalyst. The results are presented in Table 3. Table 3The BBD matrix and experimental results generated using Minitab software*.Experiment numberCoded factorsReal factorsRemoval (%) at 15 minX1X2X3Oxidant/sulfur (mol/mol)Temperature (°C)Catalyst/Fuel (g/L)YBTYDBT130 − 113.5405616690113.56057279140003.5503758410003.550374835 − 10 − 1250132374 − 1 − 102403263120 − 1 − 13.5401535871 − 105403727412 − 110260341481001 − 13.560163708 − 1012505434730003.5503758311101550569721510 − 155017074611056037178*All experiments were carried out under ultrasonic irradiation at a power of 500 W and a frequency of 20 kHz. The BBD matrix and experimental results generated using Minitab software*. *All experiments were carried out under ultrasonic irradiation at a power of 500 W and a frequency of 20 kHz. The pathway of sulfur compounds oxidation by H2O2 is shown in Fig. 11. As can be seen, theoretically, two moles of oxygen are needed to convert one mole of sulfur compounds into their corresponding sulfones76. Due to the limitations of mass transfer and the possibility of thermal decomposition of H2O2, using a molar ratio equal to the stoichiometric ratio is insufficient. For this reason, the oxidant/sulfur molar ratio is usually considered larger than the stoichiometric ratio77,78. Reaction pathway of oxidation of BT and DBT in the ODS process. Statistical analysis The effects of the oxidant/sulfur ratio, reaction temperature, and catalyst/fuel on BT and DBT removal in the presence of HPW/Al2O3/PANI catalyst were investigated using RSM analysis. The experimental results were analyzed using statistical analysis. The significance and accuracy of the fitted models were evaluated by analysis of variance (ANOVA) and diagnostic plots, respectively. The results of the ANOVA analysis are presented in Tables 4 and 5. As can be seen, for both models, the p-values are below 0.05, while the F-values exceed the corresponding critical values, indicating statistical significance. Table 4ANOVA results for BT removal using the HPW/Al2O3/PANI composite.SourceDFSum of squaresMean squareF-ValueP-ValueNotesModel93795.08421.6865.380.000SignificantX112450.002450.00379.840.000SignificantX21153.12153.1223.740.005SignificantX3191.1291.1214.130.013SignificantX211844.67844.67130.960.000SignificantX221150.06150.0623.260.005SignificantX231106.67106.6716.540.010SignificantX1 X2164.0064.009.920.025SignificantX1 X3136.0036.005.580.065Not significantX2 X310.250.250.040.852Not significantError532.256.45Lack-of-Fit330.2510.0810.080.092Not significantPure Error22.001.00Total143827.33 ANOVA results for BT removal using the HPW/Al2O3/PANI composite. Table 5ANOVA results for DBT removal using the HPW/Al2O3/PANI composite.SourceDFSum of squaresMean squareF-ValueP-ValueNotesModel94144.48460.5078.720.000SignificantX112278.122278.12389.420.000SignificantX21264.50264.5045.210.001SignificantX3178.1278.1213.350.015SignificantX2111196.311196.31204.500.000SignificantX221194.08194.0833.180.002SignificantX231207.69207.6935.500.002SignificantX1 X2142.2542.257.220.043SignificantX1 X3136.0036.006.150.056Not significantX2 X310.250.250.040.844Not significantError529.255.85Lack-of-fit327.259.089.080.101Not significantPure error22.001.00Total144173.73 ANOVA results for DBT removal using the HPW/Al2O3/PANI composite. The coefficients of determination, including R-squared, adjusted R-squared, and predicted R-squared, for the suggested correlations, are summarized in Table 6. The results indicate a high level of agreement between the experimental data obtained and the data predicted by the models. Moreover, the lack of fit of the models was not statistically significant at a 95% confidence level. Table 6The model fit coefficients for the suggested quadratic correlation.CompR2R2(adj.)R2 (pred.)BT99.1697.6487.24DBT99.398.0489.45 The model fit coefficients for the suggested quadratic correlation. Based on the analysis of variance results, including Fisher’s tests and p-values evaluated at a 95% confidence level, insignificant terms were excluded, experimental relationships between BT and DBT removal as responses, and the independent variables were proposed in terms of quadratic polynomial equations as follows: As it is known, the removal efficiency of BT and DBT depends on the first order (linear effects) of X1, X2, and X3, the second order (quadratic effect) of X1, X2, and X3, and the interaction effects of X1, X2. According to the coefficients of Eqs. (5) and (6), within the studied ranges, the parameters affecting BT and DBT removal are ranked in the following order of influence: oxidant/sulfur (mol/mol.) > reaction temperature > catalyst/fuel (g/L). To check the assumptions of ANOVA analysis, diagnostic plots, normal probability, and residuals-versus-predicted plots are shown in Figs. 12 and 13. According to the normal probability plots of the residuals (Fig. 12), the closeness of the data points along a straight line and the linear trend of the plots suggest that the residuals are normally distributed. This normal distribution implies that the mean difference between the residuals and the actual values is approximately zero, verifying the underlying assumptions of the model. The plots of residuals vs. predicted values show no clear pattern or trend, which implies a random distribution of the residuals (Fig. 13). Therefore, the ANOVA results and diagnostic plots confirm that the fitted models are acceptable. Normal plot of residuals. Residuals versus predicted values. Analysis of response surfaces To evaluate the interaction between independent variables affecting the removal of BT and DBT using HPW/Al2O3/PANI composite, two-dimensional contours and three-dimensional response surface plots were created by Minitab software. According to the ANOVA result, just X1X2 (oxidant/sulfur ratio * temperature) affects the response; therefore, just this interaction was investigated. As illustrated in Fig. 14, both the oxidant/sulfur molar ratio and temperature exhibit a nonlinear (curvature) effect. With an increase in these parameters up to certain levels (oxidant/sulfur molar ratio: 4.5 and T: 50 °C), the reactivity of BT and DBT initially increases. However, further increases beyond these levels lead to a reduction in the responses. The reason for this reduction is that at high temperatures, the oxidant is thermally decomposed. (b,d) Two-dimensional and (a,c) three-dimensional plots of the interaction effect of oxidant/sulfur molar ratio and reaction temperature on the BT and DBT conversion in the catalyst/fuel of 3 mg/L, ultrasonic power of 500 W, and frequency of 20 kHz. Process optimization To obtain optimal conditions for removing BT and DBT during the UAOD process, it is necessary to search for a combination of variables in which the highest efficiency of sulfur compound removal occurs. Process optimization was done by applying the RSM using Minitab software, and the results are presented in Table 7. To confirm the adequacy of the model prediction, two replicate experiments were performed in optimal conditions. A good agreement between the predicted and experimental values was observed in the averaged results, confirming the validity of the model. Figure 15a,b shows the GC analysis of the original feed and the model fuel after achieving maximum BT and DBT conversion and extraction of sulfone species. Table 7Optimal conditions for the maximum desulfurization (ultrasound power of 500 W, frequency of 20 kHz).Sulfur compoundsIndependent variablesResponse variableX1X2X3Removal (%)PredictedRealBT4.2751.73.1279.3877DBT4.1552.853.2687.786 Optimal conditions for the maximum desulfurization (ultrasound power of 500 W, frequency of 20 kHz). The GC chromatograms of (a) original model feed and (b) model fuel after reaction, (c) model fuel after extraction of sulfones. Additional and complementary experiments The effect of the blank experiment on desulfurization efficiency is assessed in this section, focusing specifically on optimized operational conditions. As presented in Table 8, the oxidant showed minimal activity, achieving only around 3–3.5% conversion of sulfur compounds. The use of acetonitrile as a solvent improved desulfurization performance, leading to approximately 32% sulfur removal58. In contrast, employing the catalyst alone resulted in insignificant adsorption of sulfur species, suggesting a negligible contribution to the desulfurization process. The results highlight the superior catalytic performance of HPW/Al2O3 and HPW/Al2O3/PANI compared to γ-Al2O3. These observations underscore the potential of immobilizing heteropoly acids within an alumina nanoparticle framework and further covering their surface with a polymeric coating, effectively harnessing the synergistic advantages of both heteropoly compounds and polymers. Although the presence of polyaniline on the catalyst surface leads to partial coverage and consequently reduces the effective specific surface area, the structural and electronic properties of polyaniline largely compensate for this drawback. Owing to its polar nature and the presence of electron-donating/electron-withdrawing groups, polyaniline facilitates charge transfer at the catalyst–substrate interface. By enhancing surface polarity and providing efficient pathways for electron transport it improves the activation of sulfur-containing molecules. Therefore, despite the reduced surface area, the rate and efficiency of sulfur compound oxidation increase, resulting in an overall enhancement of the catalyst’s performance. The results also show that if formic acid is used in the presence of H2O2, as an organic acid, it reacts with H2O2 to generate peracid, which effectively facilitates the oxidation of organic sulfur compounds to sulfones33,36. Table 8Comparative study of various catalysts in desulfurization of BT and DBT *.EntryDesulfurization systemSulfur removal %BTDBT1γ-Al2O3002HPW/Al2O3003HPW/Al2O3/PANI004H2O22.93.65Acetonitrile31336Al2O3/H2O234387HPW/Al2O3/ H2O239448HPW/Al2O3/PANI/ H2O245529HPW/Al2O3/PANI H2O2/acetonitrile ( V/V ratio: 1)566810HPW/Al2O3/PANI/ H2O2/acetonitrile (V/V ratio: 1)/ultrasound778611HPW/Al2O3/PANI/ H2O2-formic acid/acetonitrile (V/V: 1)/ultrasound8695*Condition of desulfurization process: 30 ml of model fuel (800 ppmw of BT and DBT in n-octane), catalyst/fuel = 3.2 g/L, reaction temperature = 52 °C, oxidant/sulfur molar ratio: 4.1, ultrasonic power of 500 W, and frequency of 20 kHz. Comparative study of various catalysts in desulfurization of BT and DBT *. *Condition of desulfurization process: 30 ml of model fuel (800 ppmw of BT and DBT in n-octane), catalyst/fuel = 3.2 g/L, reaction temperature = 52 °C, oxidant/sulfur molar ratio: 4.1, ultrasonic power of 500 W, and frequency of 20 kHz. Kinetics modeling Predominantly, the kinetic analyses of desulfurization processes across different catalysts have employed power-law kinetic models79. Given that oxidation reactions are typically conducted under conditions of excess oxidant, fluctuations in oxidant concentration during the reaction are often negligible. Consequently, the rate constant is commonly determined based on a pseudo-first-order kinetic model, as represented in Eqs. (8) to (11): where Here, and represent the concentrations of the sulfur-containing compound at the initial time (t = 0) and at a specific reaction time (t = t), respectively. The apparent rate constant k′ (min−1) was determined from the slope of the linear plot of against time at a constant temperature. Furthermore, k′ exhibits a temperature dependence consistent with the classical Arrhenius formulation (Eq. 9), wherein A corresponds to the pre-exponential factor, Ea signifies the apparent activation energy, R denotes the universal gas constant, with a value of 8.3145 J mol−1 K−1, and T represents the absolute temperature (K)58. The kinetic behavior of the oxidation of sulfur compounds during the UAOD and ODS processes, catalyzed by HPW/Al2O3/PANI, was examined at three distinct temperatures, 40, 50, and 60 °C, and the corresponding results are presented in Figs. 16 and 17. Results revealed a linear correlation, which substantiated the appropriateness of the pseudo-first-order kinetic model for both the ODS and UAOD processes. Plots of [C]S.C,t/[C]S.C,0i against time for the (a) BT, and (b) DBT in the ODS process. Plots of [C]S.C,t/[C]S.C,i against time for the (a) BT and (b) DBT in the UAOD process. The kinetic rate constants and correlation coefficients for BT and DBT under both systems are provided in Table 9. Table 9The kinetic parameters for the oxidative conversion of BT and DBT in both ODS and UAOD.Sulfur compoundTemperature (°C)k'ODS (min −1)k'UAOD (min −1)R2ODSR2UAODBT400.0080.0520.9540.979500.0120.0740.9910.990600.0120.0840.9920.990DBT400.0110.0700.9590.957500.0140.0860.9790.966600.0150.0920.9950.982 The kinetic parameters for the oxidative conversion of BT and DBT in both ODS and UAOD. As indicated in Table 10, BT exhibited lower reaction rates than DBT at every temperature examined. This behavior is primarily ascribed to the notably lower electron density of the sulfur atom in BT, which renders it less reactive. The results show that at all temperatures; the apparent rate constants of oxidation of BT and DBT in the UAOD process are greater than ODS. This disparity elucidates the significantly higher sulfur conversion attained more rapidly when utilizing the ultrasound irradiation, as the higher rate constant signifies a more rapid reaction rate. The Arrhenius parameters, namely the pre-exponential factor (A) and the apparent activation energy (Ea), for the oxidation of the sulfur compounds were derived from the linear plots of ln(k) against 1/T, as shown in Fig. 18a,b and summarized in Table 10. Based on the obtained results, ultrasonic irradiation did not affect the activation energy associated with the oxidation of sulfur-containing compounds in either of the studied processes. Despite the enhanced reaction rates under ultrasonic irradiation, the apparent activation energies for BT and DBT remained unchanged, indicating that the observed improvements were not due to changes in activation energy. Instead, ultrasound significantly increased the Arrhenius pre-exponential factors, elevating them from 32.14 to 189.86 min−1 and 1.26 to 7.57 for BT and DBT oxidation, respectively. The pre-exponential factor reflects the frequency of effective molecular collisions at the reactive interface, independent of the reaction energy. Therefore, the accelerated reaction rate can be mainly attributed to the physical enhancement of mass transfer at the interface through improved emulsification, rather than chemical radical pathways, which appear to contribute only marginally. This finding is consistent with previous observations that ultrasound markedly improves interfacial mass transfer between the oxidant, fuel, and catalyst phases58,80. Table 10The Arrhenius parameters for the oxidation of sulfur compounds.Sulfur compoundEa, ODS (kJ/mol)A (min-1)Ea, UAOD (kJ/mol)A (min-1)BT21.6032.1421.28189.86DBT12.251.2612.137.54 The Arrhenius parameters for the oxidation of sulfur compounds. Arrhenius plots for the oxidation of (a) BT, (b) DBT. In Fig. 19, the role of ultrasound application on the kinetics of oxidation of sulfur compounds is demonstrated. As can be seen, in the ODS process, mechanical stirring induces the formation of an intermediate layer adjacent to the interface between the aqueous and organic phases, transitioning the catalyst from a stationary to a mobile state. Under these conditions, catalysts migrate to interfacial regions where sulfur species in proximity to their surfaces undergo oxidation in the presence of H2O2. Conversely, ultrasonic irradiation promotes a homogeneous dispersion of both the oxidant and catalyst throughout the fuel phase. The observed enhancement is predominantly attributed to the generation of reactive free radicals and the formation of microemulsions, both initiated by cavitational bubble collapse under ultrasonic irradiation. These phenomena significantly intensify mass transfer and promote more efficient interaction between the oxidant, catalyst, and sulfur-containing compounds, thereby accelerating their oxidation in the model fuel. Comparison of catalyst, oxidant, and sulfur compounds dispersion in ODS and UAOD systems. Regeneration of the catalyst To assess the recoverability and reusability of the catalyst, following the extraction of sulfones using acetonitrile as the solvent, the catalyst was first separated by centrifugation. It was then dispersed in ethanol and stirred for 50 min at 60 °C. The catalyst underwent rinsing with deionized water, followed by drying in an oven at 100 °C. The results of the recovery tests (Fig. 20a,b) indicate that after five consecutive reuse cycles, the sulfur removal efficiency decreased by only 2% for BT and 3% for DBT. This slight decline is primarily attributed to minor catalyst loss during the recovery process, as well as the retention of sulfur-containing species on the catalyst surface. These residual compounds, mainly sulfones, remain adsorbed at the active sites of the catalyst despite washing. The stability of the HPW/Al2O3/PANI composite was evaluated through X-ray diffraction (XRD) analysis and Fourier-transform infrared (FTIR) spectroscopy. As shown in Fig. 21, after being used repeatedly for five runs, no significant change was observed in the spectral patterns, indicating a high degree of structural stability of the catalyst. The performance of the recycled catalyst in the oxidation of (a) BT and (b) DBT over five consecutive reaction cycles under optimal conditions, ultrasonic power of 500 W, frequency of 20 kHz, and reaction time of 15 min per cycle. The XRD patterns and FTIR spectra of fresh composite (a), after first (b), second (c), third (d), and fourth (e) rounds of recycling. Conclusion This study investigated the ultrasound-assisted oxidative desulfurization (UAOD) of model fuels using a newly synthesized HPW/Al2O3/PANI composite catalyst. Characterization confirmed the successful synthesis of the composite, which exhibited high efficiency in sulfur removal within short reaction times. Process optimization using response surface methodology (RSM) and the Box–Behnken design (BBD) revealed that the oxidant-to-sulfur molar ratio had the most significant positive effect on BT and DBT removal, while negative effects were mainly attributed to the quadratic and interaction terms of this factor with temperature. The addition of formic acid as a co-oxidant further improved DBT removal to 95% under optimized conditions. Kinetic modeling demonstrated the superior performance of the UAOD process compared to conventional ODS. Moreover, regeneration studies showed that the catalyst retained excellent activity after five reuse cycles, confirming its stability and practical potential. Acknowledgements The authors declare that no funds, grants, or other support were received during the preparation of this manuscript Author contributions M.E performed the experiments, analyzed the results, collected the data, developed the methodology, and wrote the manuscript. MR. KH and R. MB supervised the research, reviewed, and edited the manuscript. Funding The authors declare that no financial support was received for the research, authorship, or publication of this article. Data availability All data generated or analyzed during this study are included in this published article. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analyzed during this study are included in this published article.