Dibenzothiophene Removal from Fuel Oil by Metal-Organic Frameworks: Performance and Kinetics

Desulfurization of organic sulfur in the fuel oil is essential to cut down the emission of sulfur dioxide, which is a major precursor of the acid rain and PM2.5. Currently, hydrodesulfurization is regarded as a state-of-art technology for the desulfurization of fuel oil. However, due to the stringent legislation of the fuel oil, the deep desulfurization technology is urgent to be developed. Adsorptive desulfurization method is promising due to the high selectivity and easy operation. The development of efficient adsorbent is important to advance this technology into industrial application. In this work, the five types of metal-organic frameworks (MOFs), including Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were synthesized for the adsorption of dibenzothiophene (DBT), a typical organic sulfur compound in the fuel oil. The experimental results revealed that the adsorption capacity of the five MOFs followed the order of Cu-ABTC, UMCM-150, Cu-BTC, MIL-101(Cr), and UIO-66, which adsorption capacities were 46.2, 34.2, 28.3, 26.3, and 22.0 mgS/g, respectively. The three types of Cu-based MOFs such as Cu-ABTC, UMCM-150, and Cu-BTC outperformed the Cr-based MOFs, MIL-101, and Zr-based MOFs, UIO-66. Since the surface area and pore volumes of the Cu-based MOFs were not the greatest among the tested five MOFs, the physical properties of the MOFs were not the only limited factor for the DBT adsorption. The π-complexation between DBT and linkers/metal in the MOFs was also important. Kinetic analysis showed that the DBT adsorption onto the five tested MOFs follows the pseudo-second-order kinetics, confirming that the chemical π-complexation was also contributed to the DBT adsorption. Furthermore, the operation parameters such as oil-adsorbent ratio, initial sulfur concentration and adsorption temperature for the DBT adsorption onto Cu-ABTC were optimized to be 100:1 g/g, 1000 mgS/L and 30 °C, respectively. This work can provide some insights into the development of efficient adsorbent for the organic sulfur adsorption.


Introduction
The organic sulfur contaminants in fuel oil, such as dibenzothiophenes (DBT), benzothiophene (BT), thiophenes (TH), mercaptans, and their derivatives, generate sulfur dioxides (SO 2 ) after combustion. SO 2 is a main precursor of the acid rain and PM 2.5 , which causes soil acidification, building corrosion, and other ecosystems damages [1][2][3]. Moreover, exposure to SO 2 also causes respiratory illnesses, trigger asthma, and aggravate heart disease [4]. Therefore, stringent sulfur content limitation of fuel oil was set by China and European Union [5]. The research on efficient organic sulfur compounds removal from fuel oil is urgent to the oil refining industry.
The technology of hydrodesulfurization (HDS) is mature and intensively used to remove organic sulfur from the fuel oil. Typically, the HDS was operated at high temperatures of 300-450 • C and high pressures of 3.0-5.0 MPa with an intensive consumption

Synthesis of the Five MOFs
For the Cu-BTC synthesis, 1.925 g Cu(NO 3 ) 2 ·2.5H 2 O was dissolved in 17 mL of DI water to form solution A. 1.000 g H 3 BTC was dissolved in 34 mL of the DMF and ethanol (1:1, v/v) mixture to obtain solution B. The solutions A and B were mixed and ultrasonic treated for 20 min and then transferred to a 100 mL Teflon-lined stainless steel autoclave at 100 • C for 10 h. For the UMCM-150 preparation, 0.05 g H 3 BHTC and 0.098 g Cu(NO 3 ) 2 ·2.5H 2 O were dissolved in 15 mL of the DMF-dioxane-H 2 O (4:1:1, v/v) mixture and then transferred to a 100 mL Teflon-lined stainless steel autoclave at 85 • C for 6 h. The MIL-101(Cr) was obtained by mixing 2.0 g Cr(NO 3 ) 3 ·9H 2 O and 0.83 g p-phthalic acid in 20 mL DI water for the 10 min sonication and 20 min stirring. The resulted solution was transferred to a 100 mL Teflon-lined stainless steel autoclave at 218 • C for 18 h. As for the UIO-66, 0.19 g ZrCl 4 and 0.133 g p-phthalic acid were dissolved in 82 mL DMF and then 4.8 g acetic acid was added. After the 20 min sonication and 20 min stirring, the mixture was transferred to a 100 mL Teflon-lined stainless steel autoclave at 120 • C for 24 h. The Cu-ABTC was synthesized by adding 0.18 g H 4 ABTC and 0.24 g Cu(NO 3 ) 2 ·3H 2 O into the 45mL DMF-ethanol-H 2 O (5:3:1, v/v) mixture. Then, the 2 mol/L nitric acid was added dropwise at a rate of 2 drops/sec for 5 s under stirring, and finally transferred to a 100 mL Teflon-lined stainless steel autoclave at 60 • C for 48 h. All the resulted solid-liquid mixture of the five MOFs, such as Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC, were centrifuged and successively washed by DMF and methanol for 2 h (at least 3 times). The five MOFs were dried at 150 • C for 6 h and stored in room temperature for further usage.

DBT Adsorption Procedure
The adsorption performance of the five synthesized MOFs was evaluated in a thermostatic shaker using 2 mL serum bottle. In a typical test, 10 mg of each MOFs was added into 1.42 mL simulated oil with various sulfur concentrations (100-1500 mg S/L) at 30 • C and adsorption process was operated for 7 h. The samples were collected at 10, 20, 40, 60, 120, 240, 360, and 420 min, respectively, and filtered by 0.22 µm filter for further GC analysis. To optimize the adsorption conditions, the Cu-ABTC with a high adsorption capacity was selected as the representative MOFs. The influence of initial sulfur concentrations (100-3000 mg S/L), oil-adsorbent ratios (60-120 g/g), and operation temperatures (15-40 • C) on the adsorption performance were investigated. After sampling, the MOFs and simulated oil mixture was filtered and separated by 0.22 µm filter. The remained solvent was analyzed by the gas chromatograph (Shimadzu, GC-2014, Japan) equipped with the hydrogen flame ionization detector (FID). The concentration of sulfur in simulated oil is calculated as follow: C e = m·M S ·1000/M·V (1) where m (g) is the mass of DBT; M s and M (g/mol) are the molecular weights of elemental sulfur and DBT, respectively; V (L) represents the volume of simulated oil. The kinetics of DBT adsorption onto Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were analyzed by the pseudo-first-order and pseudo-second-order models [26].
where q e and q t (mg S/g) are adsorption quantities at equilibrium and sampling time t; k 1 (min −1 ) is the adsorption rate constant of pseudo-first-order model and k 2 [g/(mg S·min)] is the adsorption rate constant of pseudo-second-order model; t presents the sampling time, s.

Characterization and Analytical Method
The crystallinity of the five MOFs was qualitatively analyzed by XRD (Empyrean, PANlytical B.V., Heracles Almelo, The Netherlands) with primary monochromatic high intensity Cu-Kα (λ = 0.154056 nm) at a scanning rate of 0.02 • /min. The tube voltage and current were 45 kV and 40 mA, respectively. The average crystallite size of the five MOFs were calculated by the Scherrer equation in JADE. FTIR measurements were carried out using a Thermo Scientifific Nicolet iN10 (Waltham, MA, USA). The five MOFs was vacuum dried at 150 • C for 2 h. The resulted MOFs was mixed with KBr and compressed to tablets. These samples were characterized under 4000-400 cm −1 . The surface area and pore size of the five MOFs were measured by N 2 -temperature programmed desorption at 77 K by ASAP 2460 (Micromeritics Instrument, Norcross, GA, USA). The samples were degassed under vacuum at 150 • C for 6 h before test. The specific surface area was determined based on the BJH model. The pore size distribution was analyzed by the NL-DFT model.
The concentration of sulfur in simulated oil was measured by gas chromatograph equipped with capillary column (RTX-1, 30 m × 0.32 mm × 0.25 µm), the temperature of injection, column, and FID detector were 290, 260, and 260 • C, respectively. N 2 was used as the carrier gas at a flow rate of 24 mL/min. The volume of the injected sample was 0.6 µL. The adsorbed capacity (q e , mg S/g) was calculated as: where C 0 and C e are the initial and equilibrium concentrations (mg S/g), respectively; V is the volume of solution (mL) and m is the mass of MOFs (mg). The adsorption thermodynamics of Cu-ABTC was determined based on the following equations: where K d is the distribution coefficient, ∆G o is Gibbs free energy (KJ/mol), ∆H o is enthalpy (KJ/mol), ∆S o is entropy (J/mol·K), T is the temperature (K), R is the constant of 8.314 J/(mol·K). The values of ∆H o and ∆S o were calculated based on the intercept and slope of the Van't Hoff plot of lnK d versus 1/T, respectively.

Characterization of Five MOFs
XRD spectra of Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were shown in Figure 1. The characteristic diffraction peaks of the five MOFs were well matched with the simulated one and literature data [27][28][29], indicating the dedicated MOFs were successfully synthesized. The diffraction peaks of Cu-BTC are at 6.5 • , 9.  Figure 1, the intensities of the characteristic peaks of the five MOFs was weaker than those in the simulation, which may be due to the low crystallinity of the synthesized MOFs. The average crystallite size of Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were 976, 371, 68, 364, and >1000 Å, respectively, which may result in the variation of the surface area and pore size of five MOFs and thus impacted the adsorption performance.
11.5°, and 13.4°. The diffraction peaks of UMCM-150 are at 6.8°, 8.7°, 10.6°, and 14.1°. The diffraction peaks of MIL-101 are at 2.8°, 3.3°, 5.2°, and 9.0°. The diffraction peaks of UIO-66 are at 7.3°, 8.5°, and 25.7°. The diffraction peaks of Cu-ABTC are at 6.2°, 7.7°, 9.9°, and 12.5°. As shown in Figure 1, the intensities of the characteristic peaks of the five MOFs was weaker than those in the simulation, which may be due to the low crystallinity of the synthesized MOFs. The average crystallite size of Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were 976, 371, 68, 364, and >1000 Å , respectively, which may result in the variation of the surface area and pore size of five MOFs and thus impacted the adsorption performance. The FTIR spectra of Cu-BTC, UMCM-150, MIL-101(Cr), UIO-66, and Cu-ABTC were analyzed to reveal thier surface functional groups ( Figure 2). As shown in Figure 2, the prepared Cu-based MOFs such as Cu-BTC, UMCM-150, and Cu-ABTC displayed a typical stretching vibration band of Cu-O at about 725 cm −1 , illustrating the successful introduction of CuO in the MOFs [30]. For the Cu-BTC, the vibration bands at 1643 cm −1 and 1373 cm −1 were detected and matched with the symmetry and asymmetry COOof H3BTC. Whereas, the COO − band was shifted to 1550 cm −1 and 1401 cm −1 in UMCM-150. As for Cu-ABTC, the COO − band was detected at 1371 cm −1 , and the N=N vibration band was observed at 1612 cm −1 [31,32]. The vibration bands at 1400 cm −1 and 1579-1623 cm −1 in MIL-101(Cr) and UIO-66 are attributed to COO − . The bands at 672 cm −1 and 559 cm −1 were assigned to the characteristic stretching vibration of Cr-O and Zr-O in MIL-101(Cr) and UIO-66, respectively. Furthermore, a broad band at 3000-3500 cm −1 was observed in all the five MOFs, which can be assigned for H2O and OH -. The detected oxygen-containing groups  [30]. For the Cu-BTC, the vibration bands at 1643 cm −1 and 1373 cm −1 were detected and matched with the symmetry and asymmetry COO − of H 3 BTC. Whereas, the COO − band was shifted to 1550 cm −1 and 1401 cm −1 in UMCM-150. As for Cu-ABTC, the COO − band was detected at 1371 cm −1 , and the N=N vibration band was observed at 1612 cm −1 [31,32]. The vibration bands at 1400 cm −1 and 1579-1623 cm −1 in MIL-101(Cr) and UIO-66 are attributed to COO − . The bands at 672 cm −1 and 559 cm −1 were assigned to the characteristic stretching vibration of Cr-O and Zr-O in MIL-101(Cr) and UIO-66, respectively. Furthermore, a broad band at 3000-3500 cm −1 was observed in all the five MOFs, which can be assigned for H 2 O and OH − . The detected oxygen-containing groups can enhance the interaction between the thiophene sulfur and linkers [33], which can contribute to the improvement of thiophene sulfur adsorption.
As show in Figure 3a, the N 2 adsorption-desorption isotherms of the five MOFs were measured. All the five MOFs displayed a type I isotherm with a sharp uptake at P/P 0 < 0.1 and a weak hysteresis loop at P/P 0 = 0.3-1.0, indicating the domination of micropores in the MOFs [34]. The NL-DFT model was used to further analyze the pore size distribution (Figure 3b). The pore sizes of all the MOFs were mainly located at 1-2 nm. The Cu-BTC, UIO-66, and Cu-ABTC possessed the micropores smaller than 1 nm, which may be helpful to improve the sulfur adsorption [17]. The BET surface area and pore volume of the five MOFs were caculated and listed in Table 1. The order of surface area and pore volumes followed the same trend and can be summarized as follows: MIL-101(Cr) > UMCM-150 > Cu-ABTC > Cu-BTC > UIO-66. The higher the surface area possessed, the greater the pore volume achieved. can enhance the interaction between the thiophene sulfur and linkers [33], which can contribute to the improvement of thiophene sulfur adsorption. As show in Figure 3a, the N2 adsorption-desorption isotherms of the five MOFs were measured. All the five MOFs displayed a type I isotherm with a sharp uptake at P/P 0 < 0.1 and a weak hysteresis loop at P/P 0 = 0.3-1.0, indicating the domination of micropores in the MOFs [34]. The NL-DFT model was used to further analyze the pore size distribution (Figure 3b). The pore sizes of all the MOFs were mainly located at 1-2 nm. The Cu-BTC, UIO-66, and Cu-ABTC possessed the micropores smaller than 1 nm, which may be helpful to improve the sulfur adsorption [17]. The BET surface area and pore volume of the five MOFs were caculated and listed in Table 1. The order of surface area and pore volumes followed the same trend and can be summarized as follows: MIL-101(Cr) > UMCM-150 > Cu-ABTC > Cu-BTC > UIO-66. The higher the surface area possessed, the greater the pore volume achieved.   As show in Figure 3a, the N2 adsorption-desorption isotherms of the five MOFs were measured. All the five MOFs displayed a type I isotherm with a sharp uptake at P/P 0 < 0.1 and a weak hysteresis loop at P/P 0 = 0.3-1.0, indicating the domination of micropores in the MOFs [34]. The NL-DFT model was used to further analyze the pore size distribution (Figure 3b). The pore sizes of all the MOFs were mainly located at 1-2 nm. The Cu-BTC, UIO-66, and Cu-ABTC possessed the micropores smaller than 1 nm, which may be helpful to improve the sulfur adsorption [17]. The BET surface area and pore volume of the five MOFs were caculated and listed in Table 1. The order of surface area and pore volumes followed the same trend and can be summarized as follows: MIL-101(Cr) > UMCM-150 > Cu-ABTC > Cu-BTC > UIO-66. The higher the surface area possessed, the greater the pore volume achieved.

Performances of DBT Adsorption
The adsorption performances of DBT onto the five MOFs were investigated with a mass ratio of 70 g/g between simulated oil and MOFs at 30 • C. The dosed sulfur concentrations ranged from 10 to 2000 mg S/L. As displayed in Figure 4a, the adsorption capacities of the five MOFs increased with the dosed sulfur concentration. At the same dosed sulfur concentration, the adsorption capacity followed the order of Cu-ABTC > UMCM-150 > Cu-BTC > MIL-101(Cr) > UIO-66, which corresponding adsorption capacities were 46.2, 34.2, 28.3, 26.3, and 22.0 mg S/g, respectively (Figure 4a). For all the five MOFs, the adsorption equilibrium was achieved within 120 min (Figure 4b). The Cu-ABTC exhibited a high adsorption capacity with the highest crystallite size of >1000 Å. MIL-101(Cr) possessing the lowest crystallite size and highest surface area and pore volume (see Table 1) exhibited the lowest adsorption capacity. Cu-BTC with a similar surface area as UIO-66 had nearly twice the absorption quantity of UIO-66. empty outermost sorbit and electrons-contained DBT could also enhance the adsorption onto the Cu-based MOFs [11]. Among the three Cu-based MOFs, the adsorption capacity of Cu-BTC was far below that of Cu-ABTC and UMCM-150, which was mainly attributed by its low surface area and pore size. However, the DBT removal by the UMCM-150 with a high surface was inferior to that of Cu-ABTC. It can be explained by the fact that the linker of Cu-ABTC contained not only two benzene rings but also N=N bond, which significantly increased the π-complexation interaction between the MOFs and DBT.

Kinetics of DBT Adsorption
The kinetics of DBT adsorption onto the five MOFs were analyzed by both the pseudo-first-order model and the pseudo-second-order model as shown in Figure 5. The values of kinetic parameters are listed in Table 2. Compared with the pseudo-first-order model, the pseudo-second-order model plots exhibited a wonderful linearity correlation with all the goodness (R 2 ) over 0.99. Moreover, the calculated values of qe,cal by pseudosecond-order model exhibited the closest value to the experimental qe,exp, indicating that the DBT adsorption kinetics by the five MOFs could be described by the pseudo-secondorder model. This result also confirmed that the chemisorption was involved in the DBT adsorption onto the tested MOFs [37]. The order of adsorption rate constant, k2, was MIL-101 (Cr) > UIO-66 > Cu ABTC > UMCM-150 > Cu BTC. The MIL-101 (Cr) with the highest adsorption rate constant may be contributed by its largest pore diameter, which reduced the DBT mass transfer limitation and thus enhanced the DBT adsorption [38]. Since the The three Cu-based MOFs displayed a higher adsorption capacity compared with the MIL-101 and UIO-66. Based on the hard/soft-acid/base principle [35], the sulfur in the DBT belongs to soft base and prefers to interact with Cu-based soft acid rather than the hard acid of Cr-based and Zr-based MOFs [36]. Additionally, the interaction between Cu with empty outermost sorbit and electrons-contained DBT could also enhance the adsorption onto the Cu-based MOFs [11]. Among the three Cu-based MOFs, the adsorption capacity of Cu-BTC was far below that of Cu-ABTC and UMCM-150, which was mainly attributed by its low surface area and pore size. However, the DBT removal by the UMCM-150 with a high surface was inferior to that of Cu-ABTC. It can be explained by the fact that the linker of Cu-ABTC contained not only two benzene rings but also N=N bond, which significantly increased the π-complexation interaction between the MOFs and DBT.

Kinetics of DBT Adsorption
The kinetics of DBT adsorption onto the five MOFs were analyzed by both the pseudofirst-order model and the pseudo-second-order model as shown in Figure 5. The values of kinetic parameters are listed in Table 2. Compared with the pseudo-first-order model, the pseudo-second-order model plots exhibited a wonderful linearity correlation with all the goodness (R 2 ) over 0.99. Moreover, the calculated values of q e,cal by pseudo-secondorder model exhibited the closest value to the experimental q e,exp , indicating that the DBT adsorption kinetics by the five MOFs could be described by the pseudo-secondorder model. This result also confirmed that the chemisorption was involved in the DBT adsorption onto the tested MOFs [37]. The order of adsorption rate constant, k 2 , was MIL-101 (Cr) > UIO-66 > Cu ABTC > UMCM-150 > Cu BTC. The MIL-101 (Cr) with the highest adsorption rate constant may be contributed by its largest pore diameter, which reduced the DBT mass transfer limitation and thus enhanced the DBT adsorption [38]. Since the pore size of Cu-ABTC was similar to the DBT, its adsorption rate was limited due to the diffusion resistance. pore size of Cu-ABTC was similar to the DBT, its adsorption rate was limited due to the diffusion resistance.

Adsorption Condition Optimization
Since the Cu-ABTC exhibited a high DBT adsorption capacity, it was used as the representative adsorbent to optimize the operation conditions. The effects of oil-adsorbent ratio, adsorption time, operation temperature, and the dosed sulfur concentration on the DBT adsorption were investigated. As shown in Figure 6a, the adsorption capacity was increased with the oil-adsorbent ratio ranging from 60 g/g to 100 g/g. When the oil-adsorbent ratio further increased to 120 g/g, the adsorption capacity slightly decreased. Therefore, the oil-adsorbent ratio of 100 g/g was recommended for the DBT adsorption onto the Cu-ABTC. The DBT adsorption onto the Cu-ABTC quickly occurred with a sulfur concentration of 1000 mgS/L and an oil-adsorbent ratio of 100 g/g. ~90% of DBT was adsorbed within 60 min and the adsorption saturation was reached within 250 min (Figure 6b). The recommended adsorption time was 60 min. As shown in Figure 6c, the DBT adsorption efficiency was gradually increased with the temperature rising from 15 to 30 °C and then remarkably decreased from 30 to 40 °C. Typically, a high temperature was not conducive to physical adsorption and thus the DBT adsorption capacity declined [39]. The temperature of 30 °C was regarded as the optimal condition. As shown in Figure 6d, the adsorption capacity was increased with the dosed sulfur concentration. When the dosed sulfur concentration reached 1000 mg S/L, the highest adsorption capacity achieved. Further increase in the dosed sulfur concentration did not boost the adsorption capacity. Overall, the optimized adsorption conditions of DBT onto Cu-ABTC were a temperature of 30 °C, an adsorption time of 60 min, an oil-adsorbent ratio of 100 g/g, and a sulfur concentration of 1000 mg S/L.

Adsorption Condition Optimization
Since the Cu-ABTC exhibited a high DBT adsorption capacity, it was used as the representative adsorbent to optimize the operation conditions. The effects of oil-adsorbent ratio, adsorption time, operation temperature, and the dosed sulfur concentration on the DBT adsorption were investigated. As shown in Figure 6a, the adsorption capacity was increased with the oil-adsorbent ratio ranging from 60 g/g to 100 g/g. When the oiladsorbent ratio further increased to 120 g/g, the adsorption capacity slightly decreased. Therefore, the oil-adsorbent ratio of 100 g/g was recommended for the DBT adsorption onto the Cu-ABTC. The DBT adsorption onto the Cu-ABTC quickly occurred with a sulfur concentration of 1000 mgS/L and an oil-adsorbent ratio of 100 g/g.~90% of DBT was adsorbed within 60 min and the adsorption saturation was reached within 250 min (Figure 6b). The recommended adsorption time was 60 min. As shown in Figure 6c, the DBT adsorption efficiency was gradually increased with the temperature rising from 15 to 30 • C and then remarkably decreased from 30 to 40 • C. Typically, a high temperature was not conducive to physical adsorption and thus the DBT adsorption capacity declined [39]. The temperature of 30 • C was regarded as the optimal condition. As shown in Figure 6d, the adsorption capacity was increased with the dosed sulfur concentration. When the dosed sulfur concentration reached 1000 mg S/L, the highest adsorption capacity achieved. Further increase in the dosed sulfur concentration did not boost the adsorption capacity. Overall, the optimized adsorption conditions of DBT onto Cu-ABTC were a temperature of 30 • C, an adsorption time of 60 min, an oil-adsorbent ratio of 100 g/g, and a sulfur concentration of 1000 mg S/L. Thermodynamic parameters such as ∆G o , ∆H o , and ∆S o of the DBT adsorption onto Cu-ABTC were calculated and listed in Table 3. The positive values of ∆G o at 20-40 • C demonstrated the feasibility and spontaneous nature of the DBT adsorption onto Cu-ABTC. The amount of ∆G o decreased with the increasing temperature, indicating that the adsorption of DBT is positively correlated with the temperature. The value of ∆H o is positive and smaller than 42 kJ/mol, suggesting that the reaction is endothermic. The positive value of ∆S o indicates that the adsorption is irreversible [40].  Thermodynamic parameters such as ∆ , ∆ , and ∆ of the DBT adsorption onto Cu-ABTC were calculated and listed in Table 3. The positive values of ∆ at 20-40 °C demonstrated the feasibility and spontaneous nature of the DBT adsorption onto Cu-ABTC. The amount of ∆ decreased with the increasing temperature, indicating that the adsorption of DBT is positively correlated with the temperature. The value of ∆ is positive and smaller than 42 kJ/mol, suggesting that the reaction is endothermic. The positive value of ∆ indicates that the adsorption is irreversible [40].

Conclusions
In summary, the five MOFs with Cu, Cr and Zr as metal centers were successfully synthesized and proven by the evidence of XRD and FRIT spectra. The adsorption test revealed that the three Cu-based MOFs, such as Cu-ABTC, UMCM-150 and Cu-BTC, exhibted a high DBT adsorption performance compared with the Cr-based MIL-101 and Zrbased UIO-66. It was confirmed that the adsorption of DBT onto the MOFs not only depended on the physisorption pertained to the surface area and pore volume but also relied on the chemisorption by the metal center and organic linkers of the MOFs. Additionally, the pseudo second-order kinetics exhibited a good correlation with the experimental values. Overall, this work can provide some guidance to the design of efficient MOFs for deep desulfurization from the fuel oil.

Conclusions
In summary, the five MOFs with Cu, Cr and Zr as metal centers were successfully synthesized and proven by the evidence of XRD and FRIT spectra. The adsorption test revealed that the three Cu-based MOFs, such as Cu-ABTC, UMCM-150 and Cu-BTC, exhibted a high DBT adsorption performance compared with the Cr-based MIL-101 and Zr-based UIO-66. It was confirmed that the adsorption of DBT onto the MOFs not only depended on the physisorption pertained to the surface area and pore volume but also relied on the chemisorption by the metal center and organic linkers of the MOFs. Additionally, the pseudo second-order kinetics exhibited a good correlation with the experimental values. Overall, this work can provide some guidance to the design of efficient MOFs for deep desulfurization from the fuel oil.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.