Liquid Extraction of Thiophene from n-Paraffins (C8, C10 or C12) Using Two Ionic Solvents [Mebupy][BF4] or [Emim][CH3SO4]

Desulfurization of gasoline and diesel models is investigated through liquid–liquid extraction of thiophene (C4H4S) from n-paraffin compounds. Ionic solvents of 4-methyl-N-butylpyridinium tetrafluoroborate [mebupy][BF4] or 1-ethyl-3-methylimidazolium methylsulfate [emim][CH3SO4] as selective solvents have been evaluated at 313.15 K and atmospheric pressure of 101.3 kPa. Experimental liquid–liquid equilibrium (LLE) data for the six ternary systems of n-octane, n-decane, or n-dodecane + thiophene + [mebupy][BF4] or [emim][CH3SO4], were used to calculate the values of the thiophene distribution ratio, selectivity and efficiency of ionic solvents. The experimental data were correlated using the UNIQUAC equations, and the binary interaction parameters have been reported. The phase diagrams for the ternary mixtures including both the experimental and calculated tie lines have been presented.


Introduction
Fuels containing sulfur compounds generate sulfur oxides (SO x ) in the combustion process, leading to severe environmental threat and serious diseases of human respiratory system [1]. The reduction of sulfur content in real diesel of 80% was achieved in a liquid-liquid extraction with chloroaluminate-containing ionic liquids, but their industrial use is undesirable because of corrosion, environmental concerns, hydrolytic stability, and regeneration aspects.
In most petroleum processes, low level of sulfur compounds such as thiophene and disulfides can be considered as the key substance to be extracted from liquid fuel oils. The maximum concentration level of total sulfur compounds content in the USA and European gasoline and diesel fuels must be at a of 10 ppm (sulfur free) [2,3]. The ionic 1 3 liquid is partly effective in removing sulfur compounds from real fuels, both diesel and gasoline [4][5][6].
Hydrodesulphurization (HDS) process is used in the removal of sulfur compounds from petroleum products and is effective in the removal of sulfur disulfides and sulfides. Also, it targets a conversion of 90% or more of gasoline/kerosene thiophene into H 2 S. Aromatic S-compounds in gasoline and diesel oil are difficult to remove completely by HDS and require extreme operating conditions of catalytic hydrodesulfurization in terms of pressure and residence time rendering the process to be uneconomic [7]. Therefore, alternative methods to HDS have been studied to remove thiophene, 2-methylthiophene, 3-methylthiophene, 2,4-dimethylthiophene and benzothiophenes including oxidation, precipitation, extraction, adsorption, distillation, and alkylation [8].
Liquid-liquid extraction is an alternative method that can be used for extractive desulfurization and denitrification of liquid fuels in light of continuous update of environmental regulations. Extraction efficiency of ionic solvents of variable cation/anion groups are investigated as economically feasible substitute.
Recent approaches employ different ionic liquids as selective solvents, due to their general immiscibility with gasoline and diesel, negligible vapor pressure, and high selectivity to sulfur-and nitrogen-containing compounds [9].
J.H. Kareem and co-authors evaluated different ionic liquids (ILs) used for the removal of thiophene, benzothiophene, and dibenzothiophene by liquid-liquid extraction and polymerization [8].
M. Rogošić et al. tested the liquid extraction of thiophene or pyridine from eight ternary systems involving one hydrocarbon using [bzmim][Tf 2 N] as an ionic solvent [14]. L. Alonso and co-authors evaluated [C 2 mim][EtSO 4 ] as green solvent for the separation of thiophene from aliphatic hydrocarbons [15]. Liquid-liquid equilibrium data for [hmmpy][NTf 2 ] + thiophene + n-hexane or n-dodecane or n-hexadecane ternary systems have been studied by M. Francisco et al. [16]. Marcin Durski et al. considered thiophene as the key substance to be separated from liquid fuel oils. The separation of thiophene from octane or hexadecane using 1-butyl-1-methylpiperidynium dicyanamide or triisobutylmethylphosphonium tosylate as a solvent has been evaluated at temperature of 308.15 K and pressure of 101 kPa [17].
R. Anantharaj 3 ] as solvents for the simultaneous separation of thiophene and pyridine from isooctane [18]. O.V. Oliveira et al. studied the desulfurization process in n-dodecane (a diesel fuel model) using the 1-butyl-3-methylimidazolium tetrafluoroborate and the model sulfur compound thiophene [19]. U. Domaǹska and M. Wlazło reported the experimental ternary LLE data for twenty-one ILs on desulfurization of gasoline and diesel models, where the fuel models consist of three sulfur aromatic compounds (thiophene, benzothiophene, and dibenzothiophene), toluene, tetralin, and heptane [20] 4 ] to act as solvent in the liquid extraction of thiophene from aliphatic hydrocarbons [24]. A.E. Gorji et al. investigated the effect of different cation structures on the thiophene distribution between the ionic liquids containing NTf 2 anion and hydrocarbon phases in the ternary systems [25]. Alonso [29]. A.A.P. Kumar and T. Banerjee evaluated different types of ionic solvents used to remove aromatic sulfur compounds from diesel oil model by liquid-liquid Extraction [30].
Although quite a few researchers preferred pyridinium-based [15] and ammoniumbased ILs [31] for extractive desulphurization. Some have been noticed to be comparable to imidazolium-based ILs if the anions matchup is just appropriate [32].
In this article, the determinations of liquid-liquid equilibria of six ternary systems, at temperature of 313.15 K and pressure of 101.3 kPa were carried out. The systems commonly involved n-octane, n-decane, or n-dodecane as characteristic compounds of gasoline or diesel models, thiophene, and ionic liquids. Two ILs, [ SO 4 ], were used to evaluate the desulfurization process. Although BF 4 ionic liquids are known of the problem of hydrolysis in presence of water [33], the ionic liquid [mebupy] [BF 4 ] was selected because pyridinium-based ionic liquids were found to be effective for the selective removal of aromatic heterocyclic sulfur compounds from fuels at room temperature. The structure and size of the pyridinium cation considerably affect the extractive capacity of the ionic solvent. The pyridinium-based ionic liquids are insoluble in the paraffinic phase and would not contaminate fuels [34]. The solute distribution ratios, selectivities, and efficiencies of IL for the liquid extraction process of thiophene from n-octane, n-decane, or n-dodecane are presented.

Chemicals
While thiophene, n-octane, n-decane, and n-dodecane were purchased from Sigma-Aldrich, [ SO 4 ]. All chemicals were stored in a cool and dry place away from sunlight in a glass container with molecular sieves to prevent moisture contamination. Prior to each experiment, the ionic liquids were degassed and dried in a rotary evaporator (WELCH model 2025) combined with a vacuum pump (WELCH model 2034) at a pressure of 12 mbar 1 3 and a temperature of below 308 K. The experiments were limited to 313.15 K to reduce the rate of any hydrolysis due to the minimal moisture content of the ionic liquids. Table 1 shows the mass fractions of purities and water contents of the pure chemicals that were used in this study.

Apparatus and Procedure
The LLE experimental systems are mixed in six 60 cm 3 glass cells heated by a water bath utilizing a Haake DC1 thermostat. The cell temperatures are measured using PT100 platinum resistance thermometer with an accuracy of ± 0.1 K. Initially, 20 g of paraffin compounds are mixed with 20 g of the ionic solvent with different amounts of thiophene. All quantities were weighed by a METTLER analytical balance accurate to ± 0.0001 g. Every mixture was stirred vigorously for one hour using a Teflon-coated magnetic stirrer before it allowed to settle for four hours to attain equilibrium. All experiments were carried out at a temperature of 313.15 K and atmospheric pressure of 101.3 kPa.

Measurements of Phase Compositions
The ternary mixtures of immiscible phases are weighed using METTLER analytical balance to experimental uncertainty of 0.0001 g. Samples of Aliphatic-rich and ionic layers were taken using a syringe and compositions of organic compounds (n-octane, n-decane, and n-dodecane) and thiophene were determined using Agilent gas chromatograph 7890B (5977A GC/MSD). The GC details are shown in Table 2. The ionic liquids cannot be analyzed by the GC because they have insignificant vapor pressures. They were collected by a pre-column to protect the primary column and to avoid any inaccuracy that can disrupt analysis which could be caused by tainting the GC with ionic liquid. A "three-point" calibration method was used to reduce systematic errors. Mole fractions of three validation samples were gravimetrically prepared and analyzed ten times using the gas chromatography. All GC analysis were repeated ten times and the average value was recorded to reduce random errors.
The  SO 4 ] were determined by material balance calculations combined with measurements of densities of samples of both layers using a density meter (Anton Paar, model DSA 5000 M) at 313.15 K with a precision of ± 10 -5 g cm 3 . The combined random uncertainty corresponding to the standard deviations of mole fraction measurements and chemical purity rectangular uncertainty applied to compositions of the order of 0.1 gives a mole fraction experimental uncertainty of 0.001.

Results and Discussion
n-decane (1) + thiophene (2)  where x is the mole fraction, while subscript 1 and 2 represent n-paraffins and thiophene compounds and superscript I and II represent the n-paraffin-rich phase and the IL-rich phase, respectively.
(2) S = x II 2 x I 1 ∕x I 2 x II 2 ,   Tables 2, 3, and 4 present the measured equilibrium mole fractions for the n-paraffin-rich phase (raffinate) and the IL-rich phase (extract) of the six ternary mixtures at 313.15 K and 101.3 kPa. Figure 1 shows the experimental tie lines of the six ternary mixtures at 313.15 K as a ternary diagram. While there was very small amount of ILs found in the n-paraffin-rich phase, more amount of n-paraffins exists in IL-rich phase as shown in the same figure.

Distribution Ratio and Selectivity
Tables 2, 3, and 4, in addition to the LLE data, include the corresponding thiophene distribution ratio and selectivity values for the six ternary systems. These parameters are widely used for evaluating the feasibility of using a solvent in a liquid extraction process.
The solvent power or capacity of an IL, which is evaluated using the distribution ratio of thiophene, was calculated for the six ternary systems. Figure 2a presents the relationship between the calculated distribution ratios and the thiophene mole fractions in the solventrich phase (x 2 ) for the six ternary systems, while Fig. 2b presents the relationship between the calculated distribution ratios and the solvent-to-feed ratio. The distribution ratio values increased with the thiophene mole fraction in the solvent-rich phase and marginally decreased with solvent-to-feed ratio. As shown in this figure, the distribution ratio values increase in the following order: n-dodecane < n-decane < n-octane. As shown in the same figure, the thiophene distribution ratio values are higher using [  selectivity of the ILs, which is a measure of the ability of the solvent to separate thiophene from paraffin compounds, was calculated.
The relationship between the calculated selectivities and the thiophene mole fractions in the solvent-rich phase (x 2 ) for the six ternary systems is presented in Fig. 3. The selectivity A comparison of the ternary systems n-octane, n-decane, or n-dodecane + thiophene + [mebupy][BF 4 ] or [emim][CH 3 SO 4 ] ILs distribution coefficients data at 313.15 K with data from previous studies using other ILs [13,15,17,24,26,27] are presented in Figs. 4, 5, and 6 for n-octane, n-decane, and n-dodecane, respectively. While Fig. 4

LLE Correlation using the UNIQUAC Model
The UNIQUAC thermodynamic model was implemented effectively to correlate the experimental (liquid + liquid) equilibrium data in several ILs in which an iterative computer program is utilized, based on flash calculation method as presented in literature [25,26]. The parameters, r i and q i , of the UNIQUAC model was predicted by Magnussen et al. [27], J.M. Seiten et al. [36], Alonso et al. [15], and Santiago et al. [17] as in Table 1. Minimizing the differences between the experimental and calculated mole fractions for each component  3 ], □, [27] determined the constituent binary parameters of UNIQUAC model over all the measured LLE data of the ternary systems.
Values of a ij and a ji parameters of the UNIQUAC model for the ternary systems at 313.15 K and atmospheric pressure are listed in Table 5 together with the root mean square deviation (RMSD) of the fit calculated as follows: where x is the mole fraction, n is the number of tie lines and subscripts exp, cal, i, j, and k represent experimental, calculated, components, phases, and tie lines, respectively.
The interaction parameters for the UNIQUAC model were used to calculate the UNIQUAC tie lines for the present systems as shown in Fig. 1. The calculations based on UNIQUAC model gave good representation of the tie line data for those systems and analysis of the mean RMSD.

Extraction Efficiency Simulation
The extraction efficiency of a 20% by mole thiophene-C 8 mixture was determined through the simulation of 4-stage extraction process. A molar solvent-to-feed ratio of 0.5 was selected for this simulation. Fig. 6 Distribution coefficient (K) against thiophene mole fraction in the solvent-rich phase (x 2 ) at 313 K for the system: n-dodecane (1) + thiophene (2) + Solvent (3) The simulation is performed using UNISIM steady-state simulator (R460.1). The ionic solvents were entered as hypothetical components and UNIQUAC interaction parameters were input in the fluid package. The extraction efficiency was calculated as follows: The influence of the number of stages on the extraction efficiency, %ε, of thiophene from n-octane using UNIQUAC model results for [ Fig. 7 showing percentage extraction efficiency of 33 to 32%, respectively, after 4 equilibrium stages. It is evident from Fig. 7 that extraction efficiency slope increases slightly as the stage number increases, indicating higher extraction efficiency at lower thiophene mole fractions at constant solvent-to-feed ratio making this process of extraction more preferable to remove thiophene at lower sulfur concentrations. Also, Fig. 7 shows comparison with literature for extraction efficiency of thiophene from n-heptane at 298.15 K and unreported solvent-to-feed ratio [37]. In spite of the variation of n-heptane instead of n-octane and the lower temperature of extraction of 298.15 K instead of 313.15 K, it is clear that the extraction efficiencies of both solvents studied in this work are comparable to that of 1,3-Dimethylimidazolium methylphosphonate [dmim][MP], while 1-ethyl-3-methylimidazolium thiocyanate [emim][SCN] showed superior extraction efficiency of about 88% after 3 stages of extraction. Further studies of ionic liquid extraction of thiophene using various alkyl chain/cation/anion groups needed before judging if these ionic liquids investigated in this study are exceptional for real applications.

Conclusion
Liquid-liquid equilibrium experimental data were studied for the ternary systems n-octane or n-decane or n-dodecane + thiophene +