The optimization of the adsorption desulfurization process for dibenzothiophene in a model oil using different ratios of hybrid MOF: AC micro adsorbers

Elimination of the release of hazardous sulfur compounds into the environment is imperative, and achieving this requires the desulfurization of petroleum products using an effective method. Adsorption desulfurization is gaining as a cost-effective alternative to various other techniques. In this study, cobalt-based metal-organic framework (Co-MOF) and Hybrid MOF: AC adsorbers were prepared using various ratios through the solvothermal method. The objective was to determine the optimal ratio for maximum removal of dibenzothiophene in a model oil through batch process adsorption. The adsorbers were characterized using FTIR, XRD, FESEM-EDS, and N2 adsorption-desorption analysis. The results indicated that a ratio of 1:5 MOF: AC achieved a 98% removal (147.57 mg/g) of DBT among (MOF, AC, 1:1 MOF: AC, 1:2 MOF: AC, and 1:10 MOF: AC) due to its optimal number of active sites and surface area. Furthermore, the optimal conditions for maximum DBT adsorption were a dosage of 0.1 g and a contact time of 2 h. The kinetic study of the adsorber showed conformity with a pseudo-first-order model. GRAPHICAL ABSTRACT


Introduction
Even at very low concentrations, sulfur compounds in transportation fuel produce harmful effects on human health.The emission of SO x and acid rain are the main problems produced by fuel combustion containing sulfur compounds [1][2][3][4].Many methods have been used to eliminate sulfur compounds and decrease the risks [5][6][7], among them, adsorption desulfurization is the most common, costless promising, and environmentally friendly method to reduce high-level sulfur compounds [8,9].
Numerous adsorbers have been used for adsorption desulfurization due to the porosity existence in their compositions, such as activated charcoal, zeolite, and metal oxides [10].Although, suitable highly selective adsorbers for sulfur compounds and sound adsorptive desulfurization capacity are the most common problems in the traditional adsorbers' utilization.To reach the appropriate adsorbers, scientists have tried to make the best adsorbers [11,12], among the attempt, very recently novel adsorber metal-organic frameworks have been synthesized [13].
Metal-organic frameworks (MOFs) have emerged as a new type of porous crystalline material with exceptional properties that offer substantial benefits and ideal features when compared to traditional catalysts.These advantages include a high specific surface area, a well-organized porous structure, and a variety of functional approaches [14].Activatedcarbon adsorbers are known for their characteristic properties such as thermochemical stability, high surface area, and inexpensive [15].
Numerous researchers have tried to synthesize hybrid adsorbers to improve the adsorption capacity of the new adsorbers by combining the physicochemical properties.this modification resulted from the adsorption selectivity and textural properties [16].For instance, GO/HKUST-1 adsorber was used to remove thiophene in a model oil [13].The Cu-BTC/Gr adsorber was applied to DBT in a model oil [17].MOF-5/AC was examined to desulfurize thiophene sulfur compounds [16].
The aim of this research is to investigate the influence of MOF: AC ratios on the desulfurization process of dibenzothiophene in a model oil.The novelty of this study lies in developing a cost-effective adsorber that can efficiently remove sulfur compounds from high-sulfur fuels through an adsorption mechanism, while minimizing the required experimental time.To achieve this, Co-based MOF (Co-benzene tricarboxylic acid) and hybrid MOF: AC adsorbers were synthesized and characterized using various techniques.The synthesized hybrid adsorbers were employed to achieve a high level of sulfur removal.The study examined the impact of adsorber dosage, contact time, initial concentration, kinetics model, and adsorption isotherm to determine the optimal adsorber configuration.

Synthesis of Co-BTC MOF [18]
The reported solvothermal method was followed for the synthesis of Co-BTC MOF.An aqueous solution of metal salt Cobalt nitrate hexahydrate 95% (Biochem) was obtained by dissolving 2 g cobalt salt in 17 ml deionized water.The linker solution was prepared by dissolving 1 g benzene tricarboxylic acid (BTC) 95% (Biochem) in 34 ml (1:1 mixed solvent absolute ethanol (Scharlau): dimethylformamide (Biochem) 95%).The Homogenous mixture was attained by the dropwise addition of linker solution into metal salt solution, then the addition of dropwise ammonia 95% (Biochem) to attain alkaline pH.The mixture solution was shifted to a Teflon-lined autoclave (100 ml) for 24 hrs to complete the reaction in an oven at 100 o C.After cooling the Autoclave, the crystalline product was collected via filtration and successive washing with solvents followed by heating in a vacuum oven for 24 hrs at 65 o C.

Synthesis of Co-MOF: AC composites [16]
A homogenous mixture of linker (ethylene glycol) and metal solution (2 g cobalt nitrate hexahydrate in 17 ml deionized water) was obtained by 60 min stirring followed by 60 min sonication, after the addition of a stoichiometric amount of activated charcoal (AC) (Carlo Erba) to prepare 1:1, 1:2, 1:5, and 1:10 of MOF to AC composites.The reaction mixture was transferred into the autoclave and kept at 100°C for 24 h in the heating oven.After filtration and repeated washing, the solid product was vacuum dried at 65°C overnight.

Characterizations
The morphology, particle size, and chemical composition of adsorbers were determined by Field emission scanning electron micropores (FESEM-EDS) (TESCAN-model: MIRA, MAP: MIRA II Czech Republic).X-Ray Diffractometer (Panalytical X-Ray Diffractometer, XPERT-PRO diffractometer) was used to obtain the X-Ray Diffraction (XRD) pattern of the samples.For identification of the functional groups present on the surface of the adsorbers, the Fourier Transform Infrared Spectroscopy (FT-IR) spectra were recorded on Perkin-Elmer FT/IR spectrometer.The surface area and textural properties were determined by Brunauer-Emmett-Teller (BET) adsorption analyzer (Quantachrome instruments, asiqw in Version 4.0, USA).N 2 adsorption -desorption isotherms were carried out at -196°C.Gas chromatographic analysis (GC-PFPD) equipped with a capillary column (beifen 3420A; KB-PONA, 50m * 0.32 mm * 0.50 μm) was used to monitor the desulfurization process.
where q e (mg.g −1 ) is the amount of sulfur compound adsorbed on the adsorber, C o and C e (mgS.L −1 ) are the concentration of DBT before and after desulfurization, respectively.
V is the volume of model oil in (L), and m is the mass of adsorber in (g).
The factors that affect the adsorption process for 1:5 MOF: AC adsorber have been studied.To study the effect of adsorber dosage (0.05, 0.1, 0.15, 0.2, 0.25) g of 1:5 MOF: AC adsorber have been used to desulfurize (15 ml of 1000 ppm DBT solution, 2 h stirring at 500 RPM, 25°C).The effect of contact time and the kinetic models of the adsorption desulfurization process have been studied by monitoring the desulfurization process from initial to 180 min contact time at condition (15 ml of 1000 ppm DBT solution, 0.1 g adsorber dosage of 1:5 MOF: AC, stirring at 500 RPM, 25°C).

Fourier transform infrared spectroscopy (FT-IR)
As illustrated in the FT-IR spectrum Figure 1, the Co-BTC has been synthesized successfully.According to the FT-IR characterization, the MOF adsorber has a broad peak at 3000 to 3500 cm −1 belonging to the O-H functional group or adsorbed water molecules of the surface of the adsorber [21].Additionally, some intense peaks appeared around 1400 to 1600 cm −1 assigned to stretching vibrations C = O and C = C of the aromatic ring of BTC [22].Moreover, at the fingerprint position in a range of 400 to 700 cm −1 , the sharp peak is assigned to Co-O vibration in the structure of the MOF [23].
For activated charcoal adsorber.at wavenumber 3400 cm −1 an adsorption band assigned to the asymmetric stretching vibration of the O-H group [24], the peaks at 2920 and 2852 cm −1 are symmetric and asymmetric stretching vibration of C-H respectively [25], the peak at 1434 cm −1 appeared due to aromatic C = C ring stretching [26], the adsorption band at 1165 cm −1 C-O stretching tertiary alcohol [27], and the adsorption band at 877 cm −1 assigned to C-H bending out of the plane of aromatic rings [28].For hybrid adsorbers, the same adsorption bands with different intensities have appeared, depending on the ratio of activated charcoal and Co-MOF.By increasing the amount of MOF in the adsorber, the intensity of peaks of carbon base decreased and the characteristic peaks of MOF have been shown more [29].

X-ray diffraction (XRD)
The XRD pattern of synthesized adsorbers has been shown and compared in the Figure 2. The simulated XRD, Figure 3, [30] confirmed that the Co-MOF was successfully prepared.The XRD diffraction peaks are located at values between 5 θ and 70 θ .For AC adsorber, the peaks at 26.55 θ , 29.35 θ, and 43.55 θ are assigned to the amorphous nature of AC [20].For MOF adsorber, the sharp peaks show the crystalline structure of the particles.The peaks at theta degrees 10.1, 16, 17, 18, 22, 33, 42, 45, and 58 indicated the presence of cobalt metal oxides and hydro-oxide formation resulting from the coordination of the framework [18,31],.As can be seen from the XRD pattern in hybrid 1:1, 1:2, 1:5, and 1:10 AC/MOF adsorbers the main peaks of both AC and MOF appeared.However, a few peaks have disappeared due to the combination and overlap of the peaks.Additionally, by increasing the amount of MOF, the rough baseline and wide peaks of the amorphous AC decreased and the crystalline structure was observed.Thus, the MOF particles are composited on the surface of AC [16].

Brunauer-Emmett-Teller (BET)
The textual properties such as surface area, pore volume, and pore diameter of synthesized adsorbers have been determined by the N 2 adsorption-desorption isotherm method Figure 4.According to the shape of the isotherm, the MOF adsorber is mainly microporous because it exhibits Type II isotherm classification [32].The AC and hybrid MOF: AC adsorbers exhibit Type IV isotherm classification which suggests that the solids have micro-and mesopores and due to interactions between N 2 gas and mesopore surface capillary condensation occurred [33].Thus, the desorption process is different from the adsorption, resulting in a hysteresis loop (Type H 4 ) which is characteristic of microporous substances with narrow slit-like pores [34].The summary of textual properties of all absorbers is shown in the Table 1.The MOF has the lowest surface area, pore volume, and pore size which were equal to S BET 8.716 m 2 /g, V BJH 0.032 cc/g, and d BJH 8.2187 nm.Thus, MOF is classified as mesoporous because the pore diameter lies between 2 nm to 50 nm [35].AC has the highest surface area S BET 169.112 m 2 /g, microporous adsorbers d BJH 1.6717 nm, and average pore volume V BJH 0.189 cc/g [36,37].Moreover, for hybrid MOF:AC adsorbers as shown in the Table 1 there are variations in the textual properties.For instance, by increasing the amount of MOF composite on the AC, the surface area, pore volume   decreased and the pore diameter is approximately constant which resulted from pore blocking by MOF composite [38].The MOF: AC hybrid exhibited a wider hysteresis loop, as observed.Figure 2S displays the pore size distribution, revealing that the hybrids had a broader range of pore sizes with an increase in the average pore radius.Interestingly, the micropore region of the hybrids shared a similar pore size distribution with pure Co-BTC, indicating that the open pore structure of Co-BTC was maintained in the hybrids.The widened pore size distribution may suggest a slight modification in the porous structure of the new adsorber network comprising Co-BTC [39].

Field emission scanning electron micropores (FESEM-EDS)
The morphology of AC, Co-MOF, and hybrid MOF: AC were investigated by (FE-SEM EDS).According to the Figure 5(a), the AC adsorber has amorphous nature [40].Co-MOF has a crystalline hexagonal structure [41] and particle size in the range (0.1-07μ) as shown in the Figure 5(f).In the Figure 5(b-e), the microparticles of Co-MOF were distributed uniformly due to the amount of MOF composites and showed clearly the ratios (1:10, 1:5,  1:2, and 1:1) MOF: AC and the microparticles of Co-MOF in the 1:10 ratio was indicated in the Figure 3S.EDS was used to show the elemental analysis for the prepared adsorbers and Co-MOF are regularly distributed on the AC, the weight composition of the adsorbers are arranged in Table 2 [42].

Scanning of different adsorbers
To determine the effect of the amount of ratio of MOF to AC on the adsorption desulfurization efficiency different adsorbers Co-MOF, (1:1, 1:2, 1:5, and 1:10) MOF: AC and AC have been synthesized and applied in adsorption desulfurization at the same condition (15 ml of 1000 ppm DBT solution, 0.1 g adsorber dosage, 2 h stirring at 500 RPM, 25 o C).As shown in Figure 6, pure AC and pure MOF have the lowest sulfur removal percentage which were equal to 44.06% and 69.93% respectively.In the hybrid MOF: AC adsorber by increasing the amount of AC in adsorbers with a ratio 1:1, 1:2, and 1:5 the sulfur removal percentage increased to 94.47%, 95.33%, and 98.02% respectively.However, with a further increase in the amount of AC to 1:10 ratio the sulfur removal decreased to 86.72%.This result can be explained by the BET and FESEM-EDS analyses.By increasing the amount of AC and decreasing the amount of Co-MOF in the hybrid adsorber, the surface area increases, and the adsorption capacity increase [43].However, by decreasing the amount of Co-MOF, the amount of adsorption active sites decrease, and adsorption capacity decreases [44].For 1:5 MOF: AC adsorber which has a high surface area (98.475 m 2 /g) and optimum amount of Co-BTC MOF (0.65%Co) which are active sites for DBT adsorption [16].Thus, 1:5 MOF: AC adsorber has the maximum DBT adsorption capacity (92.42 mg/g) and sulfur removal percentage (98.02%).Moreover, by comparing the adsorption capacity (q e ) of 1:5 MOF: AC adsorber with the literature GO/HKUST-1 [13] and CU-BTC/Gr [17], it exhibited excellent adsorption capacity which were equal to 92.42, 60.67 mg/g, and 46.2 mg/g respectively.

Effect of adsorber dosage
The experiments were carried out by varying the adsorber dosage from 0.05 to 0.25 g.The percentage of DBT removal by 1:5 MOF: AC at different adsorber dosages is presented in the Figure 7. Experimental studies were carried out at 25°C with a solution of 15 ml/1000 ppm DBT, 2 h, 500 RPM in batch process.The sulfur removal percentage increased rapidly from 18.21% to 98.38% with an increase in the adsorber dose from 0.05 to 0.1 g.This enhancement resulted from accessible of more active sites for DBT uptake [42].However, with a further increase in the amount of adsorber to 0.15, 0.20, and 0.25 the percentage of sulfur removal decreased with no additional improvement seen in the elimination of DBT from the model oil.This could be due to the aggregation of adsorber particles, which obstructs some active sites from being available for DBT adsorption.As a result, a dosage of 0.1 g of the adsorber is sufficient to attain desulfurization of DBT from the model oil [4,45].

Effect of contact time and kinetic studies
The adsorption mechanism was studied by two kinetic models, Pseudo-First-Order and Pseudo-Second-Order models [46,47].To determine the kinetics of the DBT adsorption process and equilibration time to achieve maximum DBT adsorption, the effect of contact time has been studied [48].For this purpose, the adsorption desulfurization process has been studied by monitoring the desulfurization process from initial to 180 min contact time at condition (15 ml of 1000 ppm DBT solution, 0.1 g adsorber dosage of 1:5 MOF: AC, stirring at 500 RPM, 25°C).Lagergren Model (equation ( 3)) described Pseudo-First-Order Model [49] and Blanchard, Maunaye, and Martin (equation ( 4)) proposed the Pseudo-Second-Order model and describes the complete adsorption [50].
where: q e and q t (mg/g) denote the adsorption capacity at constant time t (min) at equilibrium, respectively, k 1 (min −1 ) is the pseudo-first-order rate constant, K 2 is the Pseudo-Second-Order rate constant(g/mg.min).The experimental data fitted to both models have  been suggested to estimate the conceivable mechanism of adsorption of DBT.According to the results Figure 8, the adsorption process of DBT by 1:5 MOF: AC adsorber reaches the equilibrium at 60 min and it has maximum adsorption capacity which was equal to 92.40 mg/g.Table 3 shows all calculated parameters attained by plotting two kinetic models.In the results, the adsorption process of DBT is more closely fitted with the Pseudo-First-Order kinetic model because the correlation coefficient R 2 -values for Pseudo-First-Order and Pseudo-Second-Order are equal to 0.9544 and 0.9277 respectively.Additionally, there is an agreement between the experimental q e and calculated q e in Pseudo-First-Order kinetic model [51,52].Thus, the adsorption process of DBT was physisorption which involves intermolecular forces due to adsorption in a liquid-solid state [53,54].

Effect of initial DBT concentration
The capacity of adsorption for 1:5 MOF: AC was evaluated by conducting experiments with different concentrations of DBT ranging from 250 mg.L −1 to 2000mg.L −1 .Figure 4S illustrates the impact of the initial DBT concentration on the 1:5 MOF: AC, revealing that the adsorber's capacity for adsorption increased as the DBT concentration rose up to 1000 mg.L −1 .However, beyond this point, the adsorber's capacity gradually declined.This increase in DBT adsorption can be attributed to a rise in the driving force for mass transfer, resulting from the concentration gradient between the bulk solution and the surface of the adsorbent [55].Two widely recognized mathematical models, namely Langmuir and Freundlich isotherm, have provided explanations for the mechanism of adsorption [56].The Langmuir model, represented by Eq. 5, predicts the formation of a monolayer coverage on a uniform adsorber surface.On the other hand, the Freundlich model, represented by Eq. 6, predicts the occurrence of multilayer adsorption with a diverse distribution of functional groups and interactions among the adsorbed molecules [45].q e = q max .K L .
C e 1 + C e .K L . . ...(Langmuir) ( 5) where q max is the maximum adsorption capacity, K L (Kg.mg −1 ) is the Langmuir constant and K F (mg 1−1/n /L 1/n /g) is the Freundlich constant.Based on the findings, Table 4, it was determined that the Langmuir model demonstrated a stronger correlation with the experimental data compared to the Freundlich model.This conclusion was drawn based on the higher value of R 2 associated with the Langmuir model and the calculated value of the adsorption capacity (233 mg.g −1 ) in the Langmuir model was closed to the experimental value (226 mg.g −1 ).Furthermore, the adsorption of DBT on the adsorbent (1: MOF: AC) occurred in a consistent monolayer [57].This information is depicted in Figure 9, which illustrates the equilibrium isotherms of the DBT solution adsorbed by the adsorber.

Comparison of performance with published works
The notable outcomes obtained from the 1:5 MOF:AC composite were contrasted with the results reported in other studies, as presented in the Table 5.The capacity of the 1:5 MOF:AC composite outperforms the other adsorbers.This enhanced capacity for DBT removal from model oil can be attributed to its high surface area and optimal distribution of active sites.

Conclusion
The of this study was to synthesize hybrid micro adsorbers through the solvothermal method for use in the adsorption desulfurization (ADS) of DBT in a model oil.Co-MOF was successfully prepared by reacting cobalt nitrate with benzene tricarboxylic acid, and it was utilized to create hybrid adsorbers with MOF: AC ratios of 1:1, 1:2, 1:5 and 1:10.The best adsorber was determined using the GC-PFPD technique, and the results indicated that the 1:5 MOF: AC ratio adsorbed the highest amount of DBT 98% (147.57mg/g).The contact time and adsorber dosage were also studied for the best adsorber.Overall, the adsorption process of the best adsorber was found to conform well with the pseudo-firstorder model, and it can be considered one of the most promising adsorbers for removing DBT in the ADS process.

Figure 2 .
Figure 2. XRD pattern of AC and AC: MOF of adsorbers.

Figure 3 .
Figure 3. Experimental XRD pattern and stimulated XRD pattern of synthesized Co-BTC.

Table 1 .
Textural properties of the synthesized adsorbers.
AdsorbersSurface area S BET (m 2 /g) Average pore volume V BJH (cc/g) Average pore diameter d BJH (nm)

Table 3 .
Parameters and R 2 -value of kinetic models for adsorption DBT by 1:5 MOF: AC adsorber.

Table 5 .
Comparison of various adsorbers on the capacity ADS of DBT.