Development and optimization of polyoxometalate-modified metal-organic frameworks for efficient removal of sulfur and nitrogen compounds from diesel fuel under visible light

Abstract

As global fuel standards tighten, stripping sulfur and nitrogen compounds from diesel has emerged as a pivotal hurdle for both environmental protection and industrial innovation. This study explores an innovative route to synthesize metal–organic frameworks (MOFs) infused with Keggin-type polyoxometalates (Mo₆W₆), crafting next-generation nanocomposites tailored for catalytic oxidative desulfurization (ODS) and denitrogenation (ODN). Zirconium- and copper-based MOFs (NUiO-66 and HKUST) were effectively functionalized with Mo₆W₆, a modification verified through FTIR and XRD characterization. Their catalytic performance in eliminating dibenzothiophene (DBT) and quinoline (Qu) from model diesel was assessed under gentle reaction conditions, employing tert-butyl hydroperoxide (TBHP) as the oxidant and harnessing visible LED light for activation. Results obtained from the present study indicated that embedding polyoxometalates within MOFs markedly boosted catalytic efficiency for the concurrent removal of sulfur and nitrogen species. These nanocomposites offer a cost-efficient, high-performance pathway to cleaner diesel production, eliminating the need for the severe conditions typical of hydrodesulfurization. This study highlights the promise of polyoxometalate-engineered MOFs as sustainable, scalable catalysts, paving the way for cleaner fuel production in modern petroleum refining.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-34559-2.

Introduction

The increasing global energy demands, coupled with tightening vehicle fuel regulations and the steady decline of oil reserves, are creating significant challenges for the petroleum refining sector. One of the most pressing issues is the presence of sulfur and nitrogen compounds in diesel, whose toxicity and environmental harm are well-documented. Beyond hastening engine wear and fostering deposit buildup, these contaminants release sulfur oxides (SOₓ) and nitrogen oxides (NO_x) upon combustion—key culprits in urban air pollution and acid rain formation^1,2. Lowering sulfur and nitrogen levels in diesel to comply with strict international standards remains a pivotal objective for the industry³. At present, hydrodesulfurization (HDS) and hydrodenitrogenation (HDN) stand as the primary commercial methods for eliminating these pollutants. These methods generally demand harsh operating conditions, with temperatures above 300 °C and elevated pressures, relying on Co–Mo or Ni–Mo catalysts dispersed on alumina support. Although HDS is effective for broad sulfur removal, it proves less efficient against stubborn sulfur species like dibenzothiophene and its derivatives, as well as nitrogen-containing compounds such as quinoline and acridine, which further hinder the process⁴. Considering the substantial capital and operational costs of traditional hydrogenation methods, coupled with their restricted effectiveness in achieving deep desulfurization and denitrogenation, there’s an urgent demand for alternative or supplementary approaches. Various approaches, including modified hydrogen processes, adsorptive and extractive techniques, biological treatments, and oxidative desulfurization (ODS) or denitrogenation (ODN), have been explored to address these challenges^5–9. Recent research, supported by pilot-scale trials, spotlights oxidative desulfurization (ODS) and oxidative denitrogenation (ODN) as promising techniques that effectively target stubborn sulfur and nitrogen compounds under gentler conditions—usually below 100 °C and at atmospheric pressure—eliminating the need for costly hydrogen gas or high-pressure equipment^5,10. Moreover, ODS and ODN allow for the simultaneous elimination of sulfur and nitrogen compounds, providing a distinct edge over traditional hydroprocessing methods that typically tackle these contaminants separately^11–13. Due to their reduced energy demands and straightforward operation, these oxidative approaches present a compelling sustainable solution for the refining sector. In energy-dependent economies such as Iran’s, where petroleum products play a central economic role, the development of effective catalytic systems for diesel fuel purification is critical to increase product value and meet environmental mandates. Polyoxometalates (POMs) are a fascinating family of metal-oxygen clusters characterized by customizable architectures, featuring early transition metal atoms intricately bonded to oxygen within polyhedral frameworks. Their distinctive physicochemical traits enable finely tuned structural designs customized to meet specific catalytic needs¹⁴. While γ-Al₂O₃ has long served as a conventional support to anchor catalytic species, materials boasting uniform, molecular-sized pores—like metal-organic frameworks (MOFs)—are gaining momentum thanks to their vast surface area, adjustable pore structures, and flexible functionalization options^15,16. Given the urgent economic and environmental need to efficiently eliminate sulfur and nitrogen from diesel—and acknowledging the limitations of traditional hydrotreating—the fusion of MOFs and POMs presents an innovative and promising catalytic solution. MOFs adorned with POMs merge the strengths of both materials, offering improved access to active sites, superior redox behavior, and heightened catalytic performance—all achieved under gentle, eco-friendly conditions. This groundbreaking strategy streamlines fuel purification by cutting energy use and lowering operational costs, all while meeting tightening emissions regulations—delivering advantages to the petroleum refining industry and supporting broader environmental sustainability. Based on current literature, this study is the first to methodically synthesize and explore Keggin-type Mo₆W₆ polyoxometalate-functionalized zirconium-based (NUiO-66) and copper-based (HKUST) MOFs as photocatalysts for the simultaneous elimination of sulfur and nitrogen compounds from model diesel fuels under visible LED light. The cooperative interplay between Mo₆W₆ clusters and the MOF’s functional groups accelerates electron transfer and boosts active-site accessibility, leading to markedly superior catalytic activity under gentle conditions compared to earlier MOF-POM composites.

Materials and methods

Chemicals

All chemicals, including sodium molybdate dihydrate, disodium hydrogen phosphate, sulfuric acid, hydrochloric acid, hydrogen peroxide, ethanol, 1,3,5-benzene tricarboxylic acid (BTC), dibenzothiophene (DBT), sodium tungstate, diethyl ether, copper nitrate, tert-butyl hydroperoxide (TBHP), copper sulfate, quinoline (Qu), hexane, methanol, sodium nitrate, dimethylformamide (DMF), 2-aminoterephthalic acid, and ZrCl₄ were purchased from Merck (Germany) and used without further purification.

Synthesis of catalyst

Catalyst synthesis was divided into three main stages:¹ preparation of multi-metal oxide nanoparticles with Keggin structures², fabrication of copper- and zirconium-based metal-organic frameworks (MOFs), and³ stabilization of polyoxometalate species on these substrates to enhance catalytic activity.

Synthesis of multiple metal oxide nanoparticles with keggin structure (Mo₆W₆, Mo₁₂, W₁₂)

Polyoxometalate compounds with general formula H₃[PMo_{12 − y}W_yO₄₀]·nH₂O, having different molybdenum and tungsten ratios, were synthesized following a modified method based on Huizhang et al.¹⁷. To synthesize H₃[PMo₆W₆O₄₀], 2.2 g of Na₂HPO₄ was dissolved in 12.5 mL of distilled water and mixed with 25 mL sodium molybdate aqueous solution. The resulting solution was stirred for 30 min at 90 °C. Next, 8.8 g of sodium tungstate salt was first dissolved in 12.5 mL of distilled water and then added to the resultant solution. After cooling, the solution’s pH was adjusted between 1.5 and 2 using sulfuric acid. The mixture was then refluxed for eight hours to complete the reaction. The resulting heteropolyacid was extracted using diethyl ether and recrystallized, and this product was designated as Mo₆W₆. To evaluate comparative performance, H₃PMo₁₂O₄₀ (Mo₁₂) and H₃PW₁₂O₄₀ (W₁₂) catalysts were also synthesized using the same method, while adjusting the ratio of molybdenum and tungsten metal salts according to the reported reference.

Synthesis of CuO and NH₂-UiO-66 (NUiO-66)

A 0.1 M solution of copper sulfate was first made. Then, the concentrated sodium hydroxide was was gradually added to the solution until the pH balanced at pH 7, where it trigger the dissolution the initially formed precipitate. Upon further increasing the pH to 8, a deep blue copper hydroxide precipitate formed. This precipitate was subsequently filtered and washed four times with deionized water to remove impurities. To obtain black copper oxide, the filtered precipitate was calcined in an oven at 400 °C for one hour^18,19. To synthesize NUiO-66, 0.384 g of zirconium chloride was first dissolved in 50 mL of DMF and stirred gently for half an hour. Next, 0.274 g of 2-aminoterephthalic acid was added to the solution. The resulting solution was heated at 120 °C for 24 h. The final product was thoroughly rinsed with DMF and methanol before undergoing a 24-hour activation at 130 °C²⁰.

Synthesis of HKUST nanoparticles

Initially, 100 mL of of copper nitrate (0.2 moles) were dispersed in a solvent mixture containing water, ethanol, and DMF in equal proportions (1:1:1). Subsequently, 1,3,5-benzene tricarboxylic acid was added to the solution.

The mixture was gently stirred for 20 min, then heated at 85 °C for 24 h. Once cooled, the solid was carefully washed with DMF and ethanol to remove any residual impurities. Next, a vacuum pump was used to extract solvents trapped within the material’s pores. To complete the process, the sample underwent a 24-hour activation at 150 °C, resulting in a vibrant turquoise-hued solid²¹.

Synthesis of NUiO-66/Mo₆W₆ and Mo₆W₆/HKUST

To synthesize the NUiO-66/Mo₆W₆ nanocomposite, 100 mg of NUiO-66 was dispersed in a solution containing 20 mL ethanol and 40 mL water, followed by sonication for 20 min.

Subsequently, 10 mg of Mo₆W₆ was introduced into the mixture, which was continuously stirred for 2 h. The resultant suspension was then transferred to a sealed Teflon-lined autoclave and subjected to hydrothermal treatment at 120 °C for 12 h. After the reaction, the product was thoroughly washed with water and ethanol, and the obtained black precipitate was dried under vacuum. The preparation of the Mo₆W₆/HKUST composite was conducted using identical procedures, substituting HKUST for NUiO-66.

Characterization

XRD analyses (PANalytical X’Pert Pro powder diffractometer) were used to determine the crystalline phase of synthesized catalysts. Fourier transforms infrared spectroscopy (FTIR; Nicolet IR 100) was conducted in the range of 400 to 4000 cm^− 1 to determine functional groups on the catalyst’s surface. UV–Vis spectroscopy technique was used to analyze the optical properties of the samples. The nitrogen adsorption-desorption technique was used to determine the specific surface area (SBET) and pore structure of the prepared nanocomposite.

Oxidative denitrification and desulfurization of synthesis nanocomposite

To explore the efficiency of synthetic catalysts in the processes of oxidative denitrification and desulfurization, two separate solutions were prepared: one containing 250 ppm of dibenzothiophene (a sulfur pollutant) dissolved in hexane, and another with 250 ppm of quinoline (a nitrogen pollutant) also in hexane. Then, 25 mL of petroleum, 0.18 mL of tertiobutyl hydrofuran (TBHP) 70 wt%, 0.6 mL of n-dodecane as internal standard, and 0.1 g of catalyst were prepared in a 100 mL flask. The photocatalytic performance of the prepared catalysts was evaluated using a circular Pyrex reactor with an internal diameter of 10 cm and a volume of 150 mL. The reactor was equipped with an oxygen pump with a flow rate of 3.5 L/min. A light-emitting diode (LED) lamp, emitting light at wavelengths between 500 and 555 nm (in the visible spectrum), serves as the light source. The reactor was covered with a quartz sheet and the the LED regulator was placed in the middle (Fig.S1). The distance between the diodes and the surface of the solution was adjusted to be 15 cm. In each set of photocatalytic experiments, 100 mL of the pollutant solution, prepared with specific concentrations, was transferred to the reactor. Samples were collected at various stages: at the beginning and at the end of the reaction, and were then injected into the gas chromatography device (GC-FID) (Hewlett-Packard HP 5890 using an Rtx-5 column) for further analysis.

Results and discussion

Characterization

Figure 1 shows the FTIR distinctive peaks of W_12, Mo_12, and Mo₆W₆. In case of W₁₂ spectrum, the peaks at 1060–1080, 967–978, 870–880 and 770–780 cm^{− 1} were associated with central oxygen atom, terminal oxygen atom, cornersharing oxygen atom and edgesharing oxygen atom, respectivly²². The wide peak spanning 3200–3600 cm⁻¹ corresponds to the O-H groups’ symmetric stretching vibrations observed across all spectra. Meanwhile, the absorption features between 800 and 900 cm⁻¹ are linked to the bending motions that occur out of the O-H plane For the polyoxometallic sample, four characteristic bands were observed in the regions (νP-O) 1051 cm^{− 1}, (νM = O) 951 cm^{− 1}, (νM-O-M) 877 cm^{− 1} and (M-O-Mν) 741 cm^{− 117}. The observed peak in the region of 1400–1900 cm^{− 1} correspond to bending vibration of the OH group.

Fig. 1.

FTIR spectrum of W₁₂, Mo₁₂, and Mo₆W₆.

The FT-IR spectrum of CuO and HKUST is shown in Fig.S2. In the spectrum of the CuO, the vibrations at 1480, 532, and 1076 cm^{− 1} are associated with the vibration mode of the oxygen bridge connecting the copper atom¹⁹. In addition, the broad peaks observed at 3300–3600 cm^{− 1} are related to stretching vibration of the hydroxy group. For the HKUST sample, the bands observed at wavelengths below 1200 cm^{− 1} are related to the vibrations of the BTC (benzene 1,3,5-tricarboxylic acid) ligand. The band at 1300–1700 cm^{− 1} is related to the carboxylate functional group, which refers to the coordination of the BTC ligand to copper metal sites. The peak at 1646 cm^{− 1} is related to the asymmetric stretching vibrations of the carboxylate group in the BTC ligand. The observed peaks at 1448 cm^{− 1} and 1371 cm^{− 1} are related to the symmetric vibrations of the carboxylate group²¹.

As shown in Fig. 2, the FTIR spectra peaks of UiO-66 at 1370 cm⁻¹ and 1568 cm⁻¹ are attributed to the C = C vibrations of aromatic ring and the O-C-O stretching mode. The observed 437–750 cm⁻¹ peak is also associated with the combined vibrational modes of Zr-O, OH, and CH bending²⁰. The FTIR spectra of the composite sample demonstrate that the structure of the Keggin polyoxometalate is fully preserved within the composite (See Fig. 2). Furthermore, the presence of organometallic frameworks alongside Mo₆W₆ is confirmed. Mo₆W₆ interacts with the functional groups on the surfaces of NUiO-66 and HKUST, including amine and carboxyl groups. In aqueous solutions, carboxyl groups are weakly acidic, but in the presence of Mo₆W₆ (with higher acidity), it acts like a weak base.There is an electrostatic interaction between the positively charged carboxyl group and anions of Mo₆W₆, which is created with a negative charge.

Fig. 2.

FTIR spectrum of NUiO-66, Mo₆W₆/UiO-66, and Mo₆W₆/HKUST.

XRD patterns of W₁₂, Mo₁₂, and Mo₆W₆ are shown in Fig. 3. In case of W₁₂, the peaks at 14.61(200), 20.71(220), 27.31(222), 29.41(400), 32.11(300), 34.41(332), 42.11(521), 45.41(532), 63.71(554), and 64.91(660) are in fully consistent with with JCPDS No. 50–0304²³. The findings demonstrated the successful formation of samples, which are in consistent with the related references²⁴. All three polyoxometalate compounds have the same crystal phase and differ only in replacing the metal atom, which does not change the crystal phase. Therefore, the XRD spectra of the three compounds are similar.

Fig. 3.

XRD pattern of W₁₂, Mo₁₂, and Mo₆W₆.

The XRD pattern of the CuO sample confirms the formation of a monoclinic crystalline phase with high crystallinity. As can be seen in Fig.S3, there is no diffraction related to impurity in the pattern²⁵.

Figure 4 shows the XRD pattern of the HKUST sample. The observed diffraction peaks at 2θ = 6.9, 9.5, 11.6, 13.4, 17.5, and 19.5 belong to the crystal planes of (440), (511), (400), (222), (220), and (221), respectively²⁶. As can be seen in Fig. 4, after loading Mo₆W₆, all the diffractions are well preserved. However, the broadening and intensity change in some diffractions is observed, which is related to the interaction between the composite components.

Fig. 4.

XRD pattern of HKUST, and Mo₆W₆/HKUST.

In case of NUiO-66 spectrum, peaks observed at 9.31, 12.51, 21.91, 27.51, 32.11 and 35.71 are assigned to (002), (022), (115), (006), (117) and (008) planes, respectively. The XRD diffraction pattern related to the NUiO-66 sample is shown in Fig. 5. The results indicate the successful synthesis of the metal-organic framework^27,28. After stabilization (Fig. 5), the diffractions related to Mo6W6 are also observed. Although, in many angles, they overlap with the diffractions of NUiO-66. The diffraction of indices for NUiO-66 at the angles of 10.85 and 12.43 are well visible in the composite sample, indicating the preservation of the organometallic framework structure in the synthesized composite.

Fig. 5.

XRD diffraction pattern of NUiO-66, Mo₆W₆/UiO-66.

The nitrogen gas adsorption and desorption analysis results are presented in Table 1. The surface area of HKUST was estimated to be 330.6 m²/g and decreased to 257.1 m²/g for Mo₆W₆/HKUST. The NUiO-66 surface area was reached 626.5 m²/g, which the surface area of Mo₆W₆/NUiO-66 decreased to 447.47 m²/g. This decline is attributed to partial pore filling and blockage caused by the Mo₆W₆ clusters, which limit the accessibility of nitrogen molecules during adsorption. In addition, the presence of the inorganic clusters can induce slight structural modifications or constriction of the pore channels, effectively reducing the overall porosity^28,29.

Table 1.

The results of nitrogen gas absorption- desorption analysis.

Sample	BET surface area (m2/g)	Average pore volume (m3/g)
NUiO-66	626.5	8.94
HKUST	330.6	8.79
Mo6W6/NUiO-66	447.47	1.21
Mo6W6/HKUST	257.1	1.32

The optical properties of the composites were analyzed using UV-Vis spectroscopy. The extrapolation of absorbance spectra allowed for the determination of band gaps for Mo₆W₆ and its composites based on organometallic frameworks. As shown in Fig. S4, the absorbance spectra for Mo₆W₆, Mo₆W₆@HKUST, and Mo₆W₆@UiO-66 were observed at 310 nm, 400 nm, and 480 nm, respectively. These wavelengths correspond to band gaps of 3.7 eV for Mo₆W₆, 3.1 eV for Mo₆W₆@HKUST, and 2.5 eV for Mo₆W₆@UiO-66. The results indicated that stabilizing the metal oxide on the modified substrate reduces the band gap and extends absorption into the visible region³⁰. This characteristic can enhance the photooxidation process: a lower band gap leads to higher visible light absorption, which is expected to improve oxidation efficiency^31,32.

Oxidative desulfurization and denitrogenation

Determination of the optimal metal oxide catalyst (CuO, W₁₂, Mo₁₂ and Mo₆W₆)

To prevent material wastage, we first assessed the efficiency of oxidative desulfurization and denitrogenation among the metal oxide photocatalysts in order to identify the catalyst with the highest oxidation efficiency. Figure 6 illustrates the effect of time on sulfur removal during the oxidative desulfurization (ODS) process using a metal oxide catalyst with no modofication. As shown in Fig. 6, the sulfur removal increased over time across all synthetic catalysts. In the optimal Mo₆W₆ sample, sulfur removal efficiency reached 62% after 150 min. These findings are consistent with the results reported by Habimana et al.³³. To explain the mechanism, the high adsorption rate occurred initially, and the desulfurization and adsorption capacity increased over the time. However, adsorption saturation was achieved after 240 min³³. Baia et al. reported the similar results in removal of sulfur and nitrogen compounds from diesel oil using clays. They noted that the rate of adsorption for sulfur compounds was rapid during the first 10 min, however, afterwards it decreased³⁴. Similarly, Srivastav and Srivastava reported comparable behavior in the adsorptive removal of dibenzothiophene (DBT) in the presence of alumina³⁵. They noted that many vacant surface sites were available at the initial adsorption stage. Over time, however, further adsorption on the remaining vacant surface sites became more complex due to repulsive forces between the solute molecules in the solid and bulk phases^35,36.

Fig. 6.

DBT removal rate in the presence of different metal oxide catalysts in desulfurization process.

The impact of time on nitrogen removal during the denitrification process with various catalysts is illustrated in Fig. 7. As shown in Fig. 7, the nitrogen removal increased within the reaction times across all synthesized catalysts. In the optimal sample, after 120 min, the nitrogen removal in the diesel model reached 71%. The efficiency of desulfurization and denitrogenation with synthetic metal oxide catalysts were as follows: Mo₆W₆ > Mo₁₂ > W₁₂ > CuO. Based on these results, Mo₆W₆ was utilized in subsequent stages to create composites based on organometallic frameworks. An effective adsorptive denitrogenation process was only achievable after extended reaction times (24 h). Using Hf-based metal-organic frameworks (MOFs), adsorptive denitrogenation improved from 19% at 5 h to 79% at 24 h, while UiO-66(Zr) reached 76% after 24 h, with no adsorption capacity observed at 5 h. UiO-66(Hf) and UiO-66(Hf)-NH₂ exhibited superior oxidative catalytic denitrogenation compared to UiO-66(Zr), achieving 97% efficiency instead of 80%. The removal of nitrogen-containing compounds (NCs) by MOFs has been attributed to hydrogen bonding, which is enhanced by functional groups such as -NH₂, -OH, -COOH, and -SO₃H³⁷. However, this behavior was not observed for quinoline (QUI) due to potential repulsion between the basic nitrogen compound and the basic amino groups of the MOF³⁸.

Fig. 7.

QU removal rate in the presence of different metal oxide catalysts in denitrogenation process.

Silva et al. employed a composite based on MIL-100(Fe) encapsulating phosphomolybdic acid (PMo₁₂) to remove sulfur and nitrogen from fuel. Their results demonstrated that complete fuel desulfurization is achieved only after 30 min of reaction, using 0.7 mmol of 50% H₂O₂, with efficiencies exceeding 90%. Increasing the amount of oxidant from 0.7 to 0.9 mmol did not lead to faster desulfurization. Additionally, the findings indicated that a higher water concentration in the desulfurization process could negatively impact the oxidative desulfurization (ODS) system. Water in the extraction phase likely hinders the transfer of sulfur compounds from the model diesel to the [BMIM]PF₆ extraction phase³⁹. Mondol et al. utilized MIL-101(Cr) to remove nitrogenous compounds from fuel in a separate study. The adsorption experiments were conducted for 6 h to determine the adsorption equilibrium. The quantity of adsorption for indole and quinoline increased over the time, reaching near equilibrium after 3 h⁴⁰. Similarly, Emam et al. synthesized composite films from polysulfone and metal-organic frameworks (MOFs) to purify liquid fuel from nitrogenated compounds. The adsorption values increased rapidly at the beginning of the contact time, regardless of the film type or the molecules being adsorbed, but gradually approached equilibrium after 8 h of contact. This initial rapid adsorption can be attributed to the higher availability of free active sites on the film that are ready for molecular adsorption. Notably, the adsorption of nitrogenous compounds onto the composite films was significantly higher than that for the polysulfone alone⁴¹.

The optimal mass ratio

The oxidation reaction was conducted using four different mass percentages of Mo₆W₆ in relation to the NUiO-66 and HKUST metal-organic frameworks in order to determine the optimal mass ratio (Fig.S5 (a and b)). The efficiency improved for two of the samples as the mass percentage increased from 10% to 30%. However, the efficiency began to decrease when the mass loading percentage was further increased. This decline is likely due to the agglomeration of particles at high concentrations, which reduces the available active sites^41,42. Consequently, all subsequent stages of this reaction were performed using the optimal mass ratio of the desired photocatalyst. In the denitrogenation process, the Mo₆W₆/NUiO-66 sample demonstrated the higher efficiency compared to the Mo₆W₆/HKUST sample, achieving a 99% denitrogenation rate within 100 min. These results are similar to those obtained with PMA-MIL-101⁴³.

Investigating the effect of the amount of catalyst in the desulfurization process

As shown in Fig. S6, increasing catalyst dosages from 0.05 g to 0.1 g leads to a significant increase in efficiency, rising from 61% to 99%. However, as the further increase of catalyst dosagae beyond 0.1 g led to decreases in efficiency. Therefore, 0.1 g of catalyst dosage was considered to be the optimal dosage for the further steps of the denitrogenation process. Baia et al. reported that the ability of attapulgite to remove sulfur compounds increased with the adsorbent dose, reaching its peak at 0.4 g.mL⁻¹, after which the efficiency remained constant. This trend may be attributed to the aggregation of adsorbent particles, which limits access to the active sites on the adsorbent³⁴.

In contrast, the removal of sulfur compounds using bentonite showed a continuous increase with higher adsorbent doses. The increased availability of specific areas and adsorbent sites for the adsorption of sulfur compounds can explain this behavior³⁴. Similar results were observed for nitrogen compounds, where removal efficiency increased steadily with increasing adsorbent doses.

Investigating the effect of pollutant concentration in the desulfurization process

The desulfurization and denitrogenation processes were conducted at three concentration levels: 250, 300, and 350 mg/L. Figures S7 and 8 indicate that the highest efficiency was achieved at the lowest concentration (250 mg/L). It is assumed that the higher concentrations lead to contaminantion more and significant portion of the catalyst’s surface and consequently decrease the efficiency. This trend is consistent with the findings from the desulfurization and denitrogenation processes, where the highest yield was also observed at 250 mg/L. As the concentration increases, the contamination of the catalyst’s surface rises and reduce the efficiency⁴⁴.

The effect of pH in the denitrification process

One important factor in the desulfurization process is the pH level, as it affects the surface charge and influences both oxidation and absorption processes. The desulfurization process was studied in three different environments: acidic (pH = 3), neutral (pH = 6), and alkaline (pH = 9) (Fig. S9). The results indicated that optimal efficiency of 99% was achieved in 100 min within the alkaline environment (pH = 9). At lower pH levels, the interaction between sulfur-free electron pairs and positive hydrogen ions increases, which ultimately reduces the adsorption of dibenzothiophene (DBT) on the surface of the pollutant⁴⁵. In the denitrogenation process, the highest efficiency was observed in the alkaline environment, achieving results in just 75 min. Similar to the desulfurization process, the increase in interaction between free electron pairs of sulfur and positive hydrogen ions at lower pH levels leads to a reduction in the adsorption of quinoline (QU) on the surface of the pollutant³⁸.

Effect of electron scavengers

Three phenomena of electron-hole pair generation, separation, and consumption are involved in the photocatalytic process. Three types of inhibitors were selected to determine the role of active species participating in the photocatalytic process. Benzoquinone (BQ), triethanolamine (TEA), and formic acid (FA) act as inhibitors for the oxide radical, hydroxide radical, and hole, respectively⁴⁶. As shown in Fig. 8, the presence of all inhibitors reduces the photocatalytic efficiency. However, benzoquinone has a larger negative effect than the other scavengers; therefore, peroxide radicals are the main active species in the photocatalytic process⁴⁷.

Fig. 8.

Photocatalytic efficiency in the presence of different radical scavengers for (a) DBT, (b) Nitrogen removal.

Catalyst recovery

Due to the significance of catalyst recovery and reuse in the ODS process, the optimized catalyst was collected, washed for multiple times with methanol, and dried at 110 °C for 12 h after each reaction. When reused, the desulfurization percentage of the catalyst slightly decreases (Fig. 9). Therefore, from an economic perspective, the optimized synthetic catalyst remains in a favorable condition. After the fourth cycle, the photocatalyst was analyzed using FTIR spectroscopy and XRD analysis (Fig. S10). As shown in Fig. S10a, the XRD pattern of the recycled photocatalyst closely matches that for the fresh sample. Furthermore, the primary characteristic vibration peaks of the photocatalyst remained intact in the FTIR spectra (Fig. S10b). The results showed that the catalyst has good stability.

Fig. 9.

Sulfur removal after four recovery cycles for (a) DBT, (b) Nitrogen removal.

Comparative of Metal-Organic frameworks (MOFs) for efficient sulfur and nitrogen compounds removal from diesel fuel

A comparative summary of catalytic performance from this study and previous works is provided in Table 2. Key parameters including catalyst type, reaction conditions, removal efficiency, and notable features are compared.

Table 2.

Comparison of catalytic performance for sulfur and nitrogen compounds removal in diesel fuel using various MOF-Based Catalysts.

Study / catalyst	Target compounds	Operational conditions	Sulfur removal efficiency	Nitrogen removal efficiency	Highlights / remarks	Ref.
POM@MOF	Sulfur and nitrogen compounds	Mild conditions, visible light	Very high (95–98%)	Very high, selective	Optimal performance, sustainable conditions	Present study
H4SiW12O40 encapsulated in amide-functionalized MOF	Sulfur compounds in diesel fuel	Mild conditions, room temperature or mild heating, possibly with oxidant (e.g.,	> 95%	-	improved stability and reusability, selective oxidative desulfurization of diesel fuel	48
UiO-66-NO2	DBT, 2-MB, 4,6-DMDBT	Mild to moderate temperature, oxidant (e.g., H2O2),	> 95%	-	Improved performance by nitrate functional group	49
P2W18Co4@ZnFe2O4@PVA	oxidative desulfurization (PODS) of fuels	room temperature, air atmosphere, UV–visible light	> 94%	-	green and sustainable fuel desulfurization	50
FWF@NMF	Oxidative desulfurization	0.10 g of catalyst at 35 °C for 60 min	98%	-	nanocatalyst exhibited a remarkable desulfurization efficiency	51
V-SPM@PANI@CH	Desulfurization of fuels catalyzed	50 mL of model/real gasoline, 0.1 g of V-SPM@PANI@CH, reaction time of 60 min, and reaction temperature of 35 °C	96%	-	High catalytic activity	9
Mixed-valence copper oxide composites derived from MOFs	Nitrogen compounds	Visible light irradiation, mild temperature, oxidative conditions	-	High removal efficiency under visible light	-	52
MIL-101(Cr)	Nitrogen and sulfur compounds	Room temperature	Good removal for both	Decreased N to 80 ppm	Recyclable and reusable	53
Step-scheme CdS/Pt/Bi2MoO6 composite photocatalyst	Nitrogen- compounds	Visible light irradiation; mild temperature; photocatalytic oxidation	-	> 85–90%	high catalytic stability and recyclability - Suitable for sustainable fuel	54
Polyoxometalate + Ionic liquid	Sulfur and nitrogen	Stable conditions, H2O2	Very high sulfur removal efficiency (often > 95%)	High nitrogen removal efficiency	Suitable for deep desulfurization and denitrogenation of fuels.	55
CTA-SiPOM-TiO2)/H2O2	Sulfur compounds	At 60℃			scavenging of sulfur compounds in heavy crude oil	56
PMo11Cd@MnFe2O4	Desulfurization of model fuel/real gasoline	At a temperature-controlled two-necked round-bottom flask, fuel and different oxidation systems (acetic acid, benzoic acid, oxalic acid, sulfuric acid, carbonic acid)	97%	-	-	57

Conclusion

This study successfully synthesized and characterized novel modified metal-organic frameworks (MOFs) based on carboxylate ligands with zirconium and copper metals, incorporating polyoxometalates (POMs) for the oxidative desulfurization (ODS) and oxidative denitrogenation (ODN) of model diesel fuel. The synthesized catalysts, including NUiO-66/Mo6W6 and Mo6W6/HKUST demonstrated the successful integration of Keggin-structured polyoxometalates within the MOF structures, as confirmed by FTIR and XRD analyses. The presence of electrostatic interactions between Mo₆W₆ and the functional groups of NUiO-66 and HKUST suggests a synergistic effect that could enhance catalytic activity. This research highlights the potential of these novel nanocomposites as cost-effective and efficient catalysts for removal of sulfur and nitrogen compounds from diesel fuel under mild conditions. It can be considered as a promising alternative to the conventional hydrodesulfurization methods. Further studies are suggested to focus on optimizing the catalyst composition, exploring different MOF structures, and evaluating the performance of these catalysts with real diesel fuel samples to validate their industrial applicability.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Haneiyeh Fakhri: Investigation/Writing Original draft, Methodology. Masoumeh Hasham Firooz: Methodology, Investigation, Validation. Asiyeh Moteallemi: Methodology, Investigation, Validation. Ali Esrafili: Conceptualization, Writing—Reviewing and Editing. Mojtaba Yeganeh: Methodology, Investigation, Validation. Hasan Pasalari: Writing—Reviewing and Editing. Mahdi Farzadkia: Visualization, Writing original Draft/Review & Editing, Supervision.

Funding

A special thanks goes out to Iran National Science Foundation, for its financial support (Grant Number: 98010869).

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.