! !! !! !!! ! Al-Khwarizmi!! Engineering!!! Journal! ! Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010)! ! ! ! Extraction of Monocyclic Aromatic Hydrocarbons From Petroleum Products Using Sulfolane as Industrial Solvent Khalid F. Chasib Al-Jiboury* Moayad abd al-Hassan** * Department of Chemical Engineering/ University of Technology Email: khalid_farhod@uotechnology.edu.iq ** Department of Radiology/ College of Medical & Health Technical (Received 26 Sebtember 2009 ; Accepted 22 September 2010) Abstract Liquid – liquid equilibria data were measured at 293.15 K for the pseudo ternary system (sulfolane + alkanol) + octane + toluene. It is observed that the selectivity of pure sulfolane increases with cosolvent methanol but decreases with increasing the chain length of hydrocarbon in 1-alkanol. The nonrandom two liquid (NRTL) model, UNIQUAC model and UNIFAC model were used to correlate the experimental data and to predict the phase composition of the systems studied. The calculation based on NRTL model gave a good representation of the experimental tie-line data for all systems studied. The agreement between the correlated and the experimental results was very good. Keywords: Liquid -liquid equilibria, extraction of aromatic, activity coefficient, sulfolane 1. Introduction Solvent extraction is one of the most important methods to produce high-purity aromatic extracts from catalytic reformates. In recent years, sulfolane or tetraethylene glycol has been employed more and more in new or improved extraction processes. Therefore, it is necessary to have complete thermodynamic data for these systems. The selection of a solvent for extraction study depends on the solvent power measured by the solute distribution coefficient and also on its selectivity. In the case of recovery of aromatics from reformats, a solvent with largest possible capacity and highest selectivity toward aromatics is preferred. Sulfolane is an important industrial solvent having the ability to extract monocyclic aromatic hydrocarbons from petroleum products. The efficient separation of ring containing compounds (e.g., cyclic ethers, cyclic alcohols, or hydrocarbons) from petroleum products is an important concept in the chemical industry where many solvents have been tested to improve such recovery. Sometimes it may be desirable to use a low-boiling solvent that has to be distilled for a recycling process. Three major factors have been found to influence the equilibrium characteristics of solvent extraction of cyclic aromatic from petroleum products (i.e., the nature of the solute, the concentration of the solute, and the type of organic solvent). Liquid-liquid equilibria (LLE) data and thermophysical properties of mixtures containing an aromatic and sulfolane with other solvents have been reported by several authors [1-3]. The quaternary system sulfolane + alkanol + octane + toluene is treated as pseudo ternary system, component 1 is ( sulfolane + methanol (MeOH), ethanol (EtOH), 1-propanol (1-PrOH), 1-butanol (1-BuOH) or 1-pentanol (PeOH)). 2. Experimental Section 2.1. Materials Sulfolane (> 99.5%, GC), octane (> 99.8%, GC), toluene (> 99.0%, GC), methanol (> 99.5%, GC), ethanol (> 99.8%, GC), 1-propanol (> 99.5%, GC), 1-butanol (> 99.5%, GC), 1-pentanol (> 99.0%, GC), were supplied by Fluka. All This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!2 chemicals were used without further purification but were kept over freshly activated molecular sieves of type 4A (Union Carbide) for several days and filtered before use. Mass fractions of impurities detectable by GC were found to be <0.0020. Deionized and redistilled water was used throughout all experiments. Refractive indices were measured through an Abbe-Hilger refractometer with an uncertainty of ±5 10-4. Densities were measured using an Anton Paar DMA 4500 density meter. The estimated uncertainty in the density was ±10-4 g/cm3. 2.2. Procedure The binodal (solubility) curves were determined by the cloud-point method in an equilibrium glass cell with a water jacket to maintain isothermal conditions. The temperature in the cell was measured by a certified Fischer thermometer within an accuracy of ±0.1 K and was kept constant by circulating water from a water bath equipped with a temperature controller. The major central part of the solubility curves was obtained by titrating heterogeneous mixtures of octane + toluene with sulfolane until the turbidity had disappeared. For the octane side and solvent side limited regions in which the curve and the sides of the triangle are close and exhibit similar slopes, binary mixtures of either (octane + sulfolane) or (toluene + sulfolane) were titrated against the third component until the transition from homogeneity to cloudiness was observed. All mixtures were prepared by weight with a Mettler scale accurate to within ±10-4 g. The transition point between the homogeneous and heterogeneous regions was determined visually. The reliability of the method depends on the precision of the Metrohm microburet with an uncertainty of ±0.005 cm3 and is limited by the visual inspection of the transition across the apparatus. Concentration determinations were made with a mass fraction uncertainty of ±0.002. End-point determinations of the tie lines were based upon the independent analysis of the conjugate phases that were regarded as being in equilibrium. The tie-lines were determined using the refractive index method the experimental procedures are described by Briggs and Comings [4]. 3. Results & Discussion 3.1. Liquid-Liquid Equilibria of the ternary systems sulfolane/co-solvent +n- Octane + Toluene Liquid – liquid equilibrium for the ternary systems 1. sulfolane + n-octane + toluene 2. (sulfolane+ 5% water)+ n – octane + toluene. 3. (sulfolane+5% methanol)+n–octane+toluene. 4. (sulfolane+ 5% ethanol)+ n–octane+ toluene. 5. (sulfolane+5% 1-propanol)+n–octane+toluene. 6. (sulfolane+5% 1-butanol)+n–octane+toluene. 7. (sulfolane+5% 1-pentanol)+n–octane+toluene. were studied at 293.15 K. 3.2. Mutual Solubility The compositions of mixtures on the binodal curve for the above seven systems at 293.15 K are plotted as triangular diagrams, Figures 1-7. The minimum concentration (in mole fraction) for the solubility of toluene, over the whole composition range, in the mixture (n- octane + solvent), was found to be 0.693, 0.703, 0.702, 0.691, 0.687, 0.679, and 0.680 for sulfolane, sulfolane + 5% water, sulfolane + 5% methanol, sulfolane + 5% ethanol, sulfolane + 5% 1- propanol, sulfolane + 5% 1-butanol, and sulfolane + 5% 1-pentanol, respectively. This reflects the magnitude of the area of the two- phase region. The two-phase region increases in the order sulfolane + 5% water > sulfolane + 5% methanol > pure sulfolane > sulfolane + 5% ethanol > sulfolane + 5% 1- propanol > sulfolane + 5% 1- butanol  sulfolane + 5% 1- pentanol. The maximum solubility of sulfolane, sulfolane + water or sulfolane + alcohols in n- octane is less than 0.014 mole fraction, and the maximum solubility of n- octane in sulfolane, sulfolane + water or sulfolane + alcohol is less than 0.020 mole fraction at 293.15K. It was observed that, the two-phase area decreases as the chain length of alcohol increases, this reflects the increase in the solubility of n- octane in sulfolane + alcohols (maximum solubility of n-octane in sulfolane + alcohols is 0.008, 0.018, 0.020, and 0.028 mole fraction for sulfolane + methanol + ethanol, + 1- propanol, and + 1- pentanol, respectively). Therefore, less n- octane miscible in solvent or solvent- co- solvent, these solvents is selective for toluene. In addition the area of the two- phase region is large, it is therefore expected that one mixture containing This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!3 large proportions of toluene + n-octane extracted with sulfolane, sulfolane + water or sulfolane + alcohols, toluene will be selectively extracted by these solvents. Fig.1. Binodal Curve and Tie Lines for Pure Sulfolane + n- octane + Toluene at 293.15 K. Fig.2. Binodal Curve and Tie Lines for (Sulfolane + 5% Water ) + n- octane + Toluene at 293.15K. Fig.3. Binodal Curve and Tie Lines for ( Sulfolane + 5% MeOH ) + n- octane + Toluene at 293.15K. Fig.4. Binodal Curve and Tie Lines for ( Sulfolane + 5% EtOH ) + n- octane + Toluene at 293.15K. Fig.5. Binodal Curve and Tie Lines for ( Sulfolane + 5% 1- PrOH ) + n- octane + Toluene at 293.15K. Toluene n-Octane Sulfolane 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Toluene n-Octane Sulfolane + water 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Toluene n-Octane Sulfolane + MeOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Toluene n-Octane Sulfolane + EtOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Toluene n-Octane Sulfolane + 1-PrOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!4 Fig.6. Binodal Curve and Tie Lines for ( sulfolane + 5% 1- BuOH ) + n- octane + Toluene at 293.15K. Fig.7. Binodal Curve and tie Lines for ( sulfolane + 5% 1-Pentanol ) + n- octane + Toluene at 293.15K. 3.3. Tie Line Data Tie line data for the seven systems at 293.15K are plotted on triangular diagrams according to a method of Francies [5], Figures 1-7. The tie line data indicating the composition of the two phases (solvent- rich phase and n- octane- rich phase ). These data are observed to fit well in the smoothed binodal curves, indicating the accuracy of the experimental tie line data. From the slope of the tie lines, it can be seen that, in all cases, toluene is more soluble in n- octane – rich phase than in solvent- rich phase with a large skewing toward the solvent axis, but the selectivity is greater than 1; thus, the extraction is possible. 3.4. Evaluation of the Consistency of the Experimental Tie Lines The accuracy of the experimental data for the seven ternary systems at 293.15K were checked by the Bachman, Othmer-Tobias, Hand, and selectivity methods [4]. Bachman method        22 11 1111 x x bax …(1) Othmer- Tobias method               22 22 22 11 11 x x1 logba x x1 log …(2) Hand method             22 32 33 11 31 x x logba x x log …(3) Selectivity method             1221 2211 44 3112 1132 x.x x.x logba x.x x.x log …(4) Experimental data are plotted using these coordinates, and the plots are shown in Figures 8- 11. The parameters aj and bj (j= 1-4) of Eqs 1-4 are obtained by using maximum likelihood principle method. The parameters and the correlation coefficients, Rj, are given in Table 1. Since the data show little scattering from a straight line, they are judged acceptable on an empirical basis, indicating internal consistency of the experimental data. The estimation of plait points for the systems is also presented in Figure 10 by the use of Treybal’s method. Toluene n-Octane Sulfolane + 1-BuOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Toluene n-Octane Sulfolane + 1-Pentanol 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!5 Fig.8. Bachman Correlation for Solvent (1) + n- Octane (2) + Toluene (3) at 293.15K. Fig.9. Othmer- Tobias Correlation for Solvent (1) + n- octane (2) + Toluene (3) at 293.15 K. Fig.10. Hand Correlation and plait Point Determination for Solvent (1)+ n-octane (2) + Toluene (3) at 293.15 K. Fig.11. Selectivity Correlation for Solvent (1) + n- octane (2) + Toluene (3) at 293.15K. As can be seen from Table 1 all methods gave good correlation for the equilibrium distribution data, the largest correlation coefficient (R) being found for all systems with selectivity method. The values of the coefficient of correlation (R) are close to unity. The goodness of the fit confirms the reliability of the results. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.40 0.50 0.60 0.70 0.80 0.90 1.00 x solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol 1 1 _______ x11 x22 -1.0 -0.5 0.0 0.5 1.0 -1.5 -1.4 -1.3 -1.2 -1.1 -1.0 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol lo g ( ) __ __ __ __ __ x1 -x 11 log ( ________ x 1-x22 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 log ( x / x ) , log ( x / x ) -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 lo g ( x / x ) , lo g ( x / x ) solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol 3 1 3 1 11 3 2 32 22 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 4.4 log ( x x / x x ) 0.8 1.2 1.6 2.0 2.4 2.8 lo g ( x x / x x ) solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol 32 11 12 31 11 22 21 12 This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!6 Table 1, Results of the Maximum Likelihood Principle Method for Solvent + n- Octane + Toluene at 293.15K. Solvent Correlation Bachman Othmer - Tobias a1 b1 R1 a2 b2 R2 Sulfolane -0.0710 0.9358 0.9937 0.5050 -0.6252 0.9970 Sulfolane + water -0.0897 1.0325 0.9944 0.6045 -0.7850 0.9977 Sulfolane + Me OH -0.2285 1.1264 0.9908 0.5222 -0.5282 0.9987 Sulfolane + EtOH -0.2299 1.0939 0.9914 0.5289 -0.4367 0.9976 Sulfolane + 1-PrOH -0.3210 1.2177 0.9897 0.7134 -0.4427 0.9982 Sulfolane + 1-BuOH -0.1818 1.0965 0.9915 0.6903 -0.5542 0.9985 Sulfolane + 1-pentanol -0.3072 1.2439 0.9908 0.7215 -0.5497 0.9971 Solvent Hand Selectivity a3 b3 R3 a4 b4 R4 Sulfolane 0.6103 -0.5869 0.9975 0.4934 0.2625 0.9984 Sulfolane + water 0.6284 -0.7816 0.9975 0.7134 -0.4650 0.9985 Sulfolane + Me OH 0.6147 -0.5208 0.9989 0.4963 0.3291 0.9987 Sulfolane + EtOH 0.6281 -0.4354 0.9981 0.4852 0.3054 0.9994 Sulfolane + 1-PrOH 0.7970 -0.4597 0.9991 0.5231 0.3784 0.9993 Sulfolane + 1-BuOH 0.6966 -0.5508 0.9945 0.4660 0.3496 0.9994 Sulfolane + 1-pentanol 0.8350 -0.5289 0.9990 0.6092 0.1508 0.9998 3.5. Distribution Coefficient and Selectivity The effectiveness of the solvent for the extraction can be expressed in terms of the distribution coefficient (k1) and (k2) of the toluene and n- octane, respectively, and the selectivity (S) of the solvent. Distribution coefficients of toluene and n-octane are represented by the formula: k1= Toluene mole fraction (or mass fraction) in solvent layer = x31 …(5) Toluene mole fraction (or mass fraction) in n-octane layer x32 k2= n-Octane mole fraction (or mass fraction) in solvent layer = x21 …(6) n-Octane mole fraction (or mass fraction) in n-octane layer x22 The selectivity (S) which is a measure of the ability of solvent to separate toluene from n- octane is given by the formula: 2 1 k k S  …(7) Figure 12 shows the comparison of distribution coefficients of toluene and Figure 13 the selectivity of solvents. As can be seen from Figures 12 and 13, The selectively vary in the following order: sulfolane + 5% water > sulfolane + 5% MeOH > pure sulfolane > sulfolane + 5% 1- BuOH > sulfolane + 5% EtOH > sulfolane + 5% 1-PrOH > sulfolane + 5% 1-Pentanol, and capacity in the order sulfolane + 5% EtOH > sulfolane + 5% 1-PrOH > sulfolane + 5% MeOH > Sulfolane + 5% 1-BuOH > sulfolane + 5% 1- Pentanol > pure sulfolane > sulfolane + 5% water. This indicates the solvent power (capacity) and its selectivity. It is apparent that increasing the water content in the modified solvent increases the selectivity and reduces the hydrocarbon solubility, while increasing the alcohol content reduces selectivity and increases the hydrocarbon solubility. In multistage, countercurrent extraction (using sulfolane) of toluene from n-octane + toluene mixture the extract purity can evidently be increased to any desired level by using a water- modified solvent. This is achieved at the expense of some increase in the solvent throughput owing to the reduced hydrocarbon solubility in the extract solvent. High selectivity for a desired capacity or solvent power is the primary requirement for a This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!7 good solvent. However, an increase in the solvent capacity of a solvent leads to a decrease in its selectivity or vice versa. To choose the optimum values of selectivity and capacity is therefore a compromise between the two values which can be adjusted here by the amount of co-solvent being added to sulfolane . On balance, considering both capacity and selectivity of solvents, with the systems studied better results were obtained for sulfolane + methanol as compared with pure sulfolane or sulfolane + water, for this reason it can be used for higher recovery of aromtics at lower solvent to feed ratios and temperatures. Fig.12. Comparison of Distribution Coefficient of Toluene with Solvents -n-octane Systems at 293.15K. Fig.13. Selectivity Curves for Solvent (1) + n- octane (2) + Toluene (3) at 293.15K. 3.6. Estimation of the Plait Point The compositions of the plait points as determined by construction and Treybal methods for the seven systems are listed in Table 2. It is apparent that the plait points are located in the region of mixtures containing more solvent. Although sulfolane + water have higher selectivity and plait point composition but its capacity is very poor. On the other hand, sulfolane + methanol have higher selectivity, capacity and plait point composition compared with the solvents studied. Thus, sulfolane + methanol can be considered to be one of the most powerful solvents for the toluene extraction. Table 2, Compositions of the plait points for solvent (1) + n- octane (2) + toluene (3) at 293.15K. Solvent Construction method Treybal method x1 x2 x3 x1 x2 x3 Sulfolane 0.394 0.070 0.536 0.391 0.076 0.533 Sulfolane + water 0.312 0.078 0.610 0.315 0.077 0.608 Sulfolane + MeOH 0.327 0.091 0.582 0.324 0.090 0.586 Sulfolane + EtOH 0.355 0.090 0.555 0.352 0.091 0.557 Sulfolane + 1- PrOH 0.360 0.090 0.550 0.363 0.092 0.545 Sulfolane + 1- BuOH 0.374 0.084 0.542 0.376 0.080 0.544 Sulfolane + 1- pentanol 0.382 0.078 0.540 0.385 0.073 0.542 0.10 0.20 0.30 0.40 0.50 0.60 0.70 x 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 D is tr ib ut io n C oe ff ic ie nt solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol       + 32 0.50 0.60 0.70 0.80 0.90 1.00 x 0 10 20 30 40 50 60 S el ec tiv it y solvents Sulfolane Sulfolane + water Sulfolane + MeOH Sulfolane + EtOH Sulfolane + 1-PrOH Sulfolane + 1-BuOH Sulfolane + 1-Pentanol       + 11 This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!8 4. General Discussion The selection of a solvent for extraction study depends on the solvent power measured by the solute distribution coefficient and also on its selectivity. In the case of recovery of aromatics from reformats, a solvent with largest possible capacity and highest selectivity toward aromatics is preferred. Combinations of sulfolane + MeOH solvent have higher capacity, selectivity, and plait point compositions compared with pure sulfolane or sulfolane + water solvent systems. Thus, sulfolane + MeOH can be considered to be one of the most powerful solvents for the toluene extraction. Moreover, viscosity of the combination of sulfolane + MeOH system is very low relative to the viscosity of pure sulfolane (sulfolane = 10.286 cP, MeOH = 0.538 at 30 oC), which should improve the extraction efficiency. Thus this combination solvent system appears to be attractive for extraction of aromatics from naphtha reformate. It is worth while to mention that, the liquid- liquid equilibria in the presence of water and alcohols are determined by intermolecular forces, predominantly hydrogen bonds. The addition of water or alcohols as co-solvent to sulfolane enhances the formation of hydrogen bonded system, which a result of greater dipole-dipole interactions between sulfolane and the co-solvent molecules. The polarity difference between the (sulfolane +co-solvent) molecules and the aromatic compound increases as the polarity of the co-solvent increases. In the aromatic series, benzene, toluene, and xylene (ortho and meta) polarity increases as the molecular weight of the aromatic member increases [6] due to the greater amount of  electrons which are subject to electromeric shifts within the ring (inductive effect of the methyl groups). Rawat [7] found that the solvent power for many extractive solvents was always greater for benzene than for toluene or xylene. Other factors such as smaller molecular size and lower molecular weight also help in the association of the benzene with the solvent molecule, making benzene more effectively extracted. The polarity difference between the solvent and an aromatic compound should not be too high for effective extraction [8]. A low polarity difference between the solvent and the aromatic compound results in attractive forces between the different molecules, and as a result the aromatic molecules are preferentially pulled toward the solvent [7]. The selectivity of (sulfolane + co-solvent) decreases in the order sulfolane + H2O > sulfolane + MeOH > sulfolane + EtOH > sulfolane + 1-PrOH > sulfolane + 1-pentanol. Indeed the hydrogen bonds system formation and the polarity difference between the solvent and the aromatic compound decreasing in the same order, supporting the above arguments. 5. Prediction and Correlation of Experimental Data If a liquid mixture of a given composition and at a known temperature is separated into two phases (i.e. at equilibrium), the composition of the two phases can be calculated from the following equations: II i II i I i I i xx  …(8) II i I ii zzz  …(9) where iz , I iz and II iz are the number of moles of component i in the system and in phases I and II, respectively, and I i and II i are the corresponding activity coefficients of component i in phases I and II, as calculated from the equilibrium equations, NRTL and UNIQUAC. The generated binary and ternary–component equilibria data are used to determine interaction parameters between paraffinic/aromatic hydrocarbons and solvent; these in turn are used to estimate the activity coefficients from the NRTL and the UNIQUAC equations. In a similar fashion the interaction parameters between parffinic/aromatic hydrocarbon groups and solvent groups were used to predict the activity coefficients form the UNIFAC model. Interaction parameters between certain groups pairs have already been reported in the literature [9], and these values have been used where required. The Ri and Qi values for the UNIFAC groups and the ri and qi for the UNIQUAC compounds are shown in Table 3. Equations 8 and 9 were solved for the mole fraction (or mass fraction) xi of component i in each liquid phase. This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!9 Table 3, The Ri /ri and Qi/qi Values for the Groups/Components Resent in the Systems. UNIFAC Model [10] UNIQUAC Model [9] Group Ri Qi Component ri qi Sulfolane 3.7220 2.936 Water 3.190 2.400 H2O 0.9200 1.400 Toluene 3.922 2.968 CH3OH 1.4311 1.432 Methanol 4.502 3.856 CH3CH2OH 2.1055 1.972 Ethanol 5.175 4.396 CH2CH2OH 1.8788 1.664 n-Octane 5.847 4.936 CH3 0.9011 0.848 Sulfolane 4.034 3.200 CH2 0.6744 0.540 1-Propanol 3.026 2.752 CH3O 1.1450 1.088 1-Butanol 3.698 3.292 CH2O 0.9183 0.780 1-Pentanol 3.471 3.638 ACH 0.5313 0.400 ACCH3 1.2663 0.968 Optimal interaction parameters between compounds for NRTL and UNIQUAC and between functional groups for the UNIFAC were found by using optimized computer program using maximum likelihood principle method developed by Sorensen [10]. The objective function (F) in this case was minimized by minimizing the square of the difference between the mole fractions (or mass fractions) predicted by the respective method and these experimentally measured.          n 1i 3 1j 2 1L 2 jLjL ]i,caledxi,tlexpx[minF …(10)  i,tlexpx jL is the experimental mole fraction,  i,caledx jL is the calculated mole fraction . The subscripts and superscripts are i for the tie lines (1,2,..,n) , j for the components (1,2,3), and L for the phase (I,II). The values of the parameters that minimized this objective function were sought, using both the UNIQUAC model and the NRTL model. The values of the six parameters for the UNIQUAC model U11, U22, U33, U12, U13, U23 (J mol -1) were calculated. The values of the nine parameters for the NRTL model g11, g22 , g33 , g12 , g13 , g23 , 11 , 12 , 13 for the ternary systems were calculated by using maximum likelihood principle method [11]. The parameters calculated in this way are shown in Tables 4 and 5. The root mean square deviation (RMSD) are calculated from the results of each method according to the following equation      2 1 2 1L 2 jLjL 3 1j n 1i n6 i,calcdxi,tlexpx RMSD             …(11) The RMSD is a measure of the agreement between the experimental data and the calculated values. The calculated tie lines using the three models for all systems studied are compared with the experimental data in Figures 14-15. Fig.14. Experimental and Calculated Tie Lines for the System (Sulfolane + 5% EtOH) + n-octane + Toluene at 293.15 K. Toluene n-Octane Sulfolane + EtOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0 .0 0 0 .2 5 0 .5 0 1 .0 0 0 .7 5 Experimental NRTL UNIQUAC UNIFAC This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!10 Toluene n-Octane Sulfolane + 1-BuOH 0.00 0.25 0.50 0.75 1.00 0. 25 0. 50 0. 75 1. 00 0. 00 0.00 0.25 0.50 1.00 0.75 Experimental NRTL UNIQUAC UNIFAC Fig.15. Experimental and Calculated Tie Lines for the System (Sulfolane + 5% 1-BuOH) + n-octane + Toluene at 293.15 K. Table 4, NRTL Parameters (gij (J mol -1)) and (ij) for the Systems Solvent (1) + n-alkane (2) + Aromatic Hydrocarbons (3) at 293.15 K. System No. g11 g22 g33 g12 g13 g23 12 13 23 1 1076.000 810.669 1878.000 5607.000 8824.000 9699.000 0.292 0.412 0.401 2 1077.000 1117.000 4676.000 5547.000 7220.000 5026.000 0.367 0.332 0.345 3 776.118 189.003 3075.000 5604.000 7895.000 6176.000 0.266 0.402 0.336 4 1634.000 74.661 1196.000 5468.000 7716.000 7712.000 0.302 0.425 0.426 5 1745.000 14.885 4543.000 5465.000 6872.000 7940.000 0.290 0.310 0.460 6 1188.000 961.639 418.311 5598.000 8461.000 9525.000 0.206 0.392 0.359 7 1878.000 76.276 2431.000 5401.000 7798.000 6068.000 0.261 0.411 0.435 Table 5, UNIQUAC Parameters (Uij (J.mol -1)) for the Systems Solvent (1) + n-alkane (2) + Aromatic Hydrocarbons (3) at 293.15 K. System No. U11 U22 U33 U12 U13 U23 1 897.569 1214.000 259.684 2842.000 1957.000 1480.000 2 1294.000 1741.000 4654.000 3571.000 1855.000 1423.000 3 1718.000 1765.000 938.861 2632.000 2746.000 2167.000 4 827.113 837.306 724.378 2519.000 2040.000 1395.000 5 955.544 1797.000 176.807 3122.000 1886.000 1674.000 6 1226.000 1569.000 210.087 3222.000 1977.000 1529.000 7 769.481 1225.000 310.311 2664.000 1808.000 1289.000 The average RMSD values for the three methods for all system studied are 0.165, 0.491, and 1.304 for NRTL, UNIQUAC, and UNIFAC, respectively. The calculations based on both the UNIQUAC model and the NRTL model gave a good representation of the tie line data. However, the NRTL model, fitted to the experimental data, is more accurate than the UNIQUAC model. The UNIFAC model has also predicted the overall composition with a reasonable error, though its This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!11 average RMSD value is higher than those of the NRTL and UNIQUAC models, as would be expected. It is therefore considered to be less accurate than the NRTL and the UNIQUAC models in correlating the phase equilibria of the studied systems. 6. Conclusions From the results of the present study, it can be concluded that: 1) In combination solvent systems (sulfolane + water or + alcohols), water acts as an antisolvent, increasing the size of the two- phase region. Conversely, alcohols decrease the size of the two-phase region and may be described as prosolvents. 2) On balance, considering both capacity and selectivity of (sulfolane + water or +alcohols), with the systems studied better results were obtained with sulfolane+ methanol as compared with pure sulfolane. 3) As a result of phase diagrams produced, the addition of alcohol to sulfalone in (n-octane+ toluene) mixture leads to a decrease in the two-phase area and reflects the increase in the solubility of n-octane in the solvent mixture. 4) In multistage, counter current extraction (using sulfolane) of toluene from the (n-octane + toluene) mixture, the extract purity can evidently be increased to any desired level by using a water-modified solvent. 5) The consistency of the data was tested by the Bachman, Othmer-Tobias, Hand, and selectivity methods. All methods gave good correlations for the equilibrium distribution data. 6) The NRTL, UNIQUAC, and UNIFAC models were used to correlate the experimental data and to predict the phase compositions of the ternary systems. The agreement between the predicted and the experimental results was good with the three models. However, the calculated values based on the NRTL model are found to be better than those based on the UNIQUAC and the UNIFAC models. Abbreviations GC Gas chromatography LLE Liquid – Liquid Equilibrium NRTL Non-Random Two Liquid activity coefficient model RMSD Root mean square deviation UNIFAC UNIQUAC Functional Group Activity Coefficients model UNIQUAC Universal Quasi-Chemical Activity Coefficient model Symbols F Objective function Ki Distribution coefficient S Selectivity z I i number of mole of component i in the system in the I phase z II i number of mole of component i in the system in the II phase Greek Litters  Activity coefficient Superscript I Phase I II Phase II Subscript i component i j component j 7. References [1] Letcher, T. M.; Redhi, G. G.; Radloff, S. E.; Domanska, U., "Liquid Liquid Equilibria of the Ternary Mixtures with Sulfolane at 303.15 K", J. Chem. Eng. Data, 41, 634-638, 1996. [2] Chen, J. M.; Fei, W.; Li, Z., "Liquid Liquid Equilibria of Quaternary Systems including Cyclohexane, 1-Heptane, Benzene, Toluene This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ Khalid F. Chasib Al-Jiboury Al-Khwarizmi Engineering Journal, Vol. 6, No. 3, PP 1-13 (2010) !!12 and Sulfolane at 298.15 K", J. Chem. Eng. Data, 45,689-692, 2007. [3] Lee, S.; Kim, H., "Liquid Liquid Equilibria of the Ternary Systems Sulfolane + Octane + Benzene, Sulfolane + Octane + Toluene and Sulfolane + Octane + p-Xylene at Elevated Temperatures", J. Chem. Eng. Data , 43, 358-361 , 1998. [4] Briggs, S. W.; Comings, E. W., "Tie-Line Correlation and Plait Point Determination", Ind. Eng. Chem., 35, 411-415 ,1993. [5] Francies, A. T. "Algebraic Representation of Thermodynamic Properties and Classification of Solutions". Ind. Eng. Chem., 40, 345-348, 2004. [6] Riddick, J. A.; Bunger, W. B.; Sakano, T. K. Organic Solvents: Physical Properties and Methods of Purification; Wiley-Interscience: New York, 2006. [7] Rewat, E. R., Elements of Extraction, 4th ed., McGraw-Hill Book Co., New York, 2001. [8] Wisniak, J.; Tamir, A. Liquid-Liquid Equilibrium and Extraction: A Literature Source Book; Elsevier: Amsterdam, 2000. [9] Gmehling, J.; Rasmussen, P.; Fredenslund, Aa. "Vapor-Liquid Equilibria by UNIFAC Group Contribution: Revision and Extension 2". Ind. Eng. Chem. Process Des. Dev., 21, 118-127, 1982. [10] Sorensen, M., Hoen, S., and Nagahama, K., Computer Aided Data Book of Vapor- Liquid Equilibria, Kodansha Limited, Tokyo, 2005. [11] Anderson, T. F., Abrams, D. S., Grens, E. A., “Evaluation of parameters for Nonlinear Thermodynamic Models”, AIChE J., 24, 20, 1998. !! This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/ ود د فرھ د خال یة المجل وارزمي الھندس ة الخ دد ٦مجل فحة 3، الع 13 -1، ص )٢٠١٠( !!13 أستخالص الھیدروكاربونات األروماتیة ذات الحلقة األحادیة من المنتجات النفطیة بأستخدام السلفولین كمذیب صناعي !! **مؤید عبد الحسن *خالد فرھود الجامعة التكنولوجیة/ قسم الھندسة الكیمیاویة* ! !!!! !! !!! !! !!!!!khalid_farhod@uotechnology.edu.iq كلیة التقنیات الطبیة والصحیة/ تقنیات األشعة قسم ** الخالصة اذب ٢٩٣.١٥سائل قد تم قیاسھا عند درجة حرارة –بیانات إتزان سائل ي الك لفولین ) (pseudo ternary system(كلفن للنظام الثالث ا + س ) + نولالك اعد . تلوین+ اوكتان ذیب المس ع الم زداد م لة ) cosolvent(لقد تم مالحظة ان األنتقائیة للسلفولین النقي ت ول السلس ادة ط ع زی ل م ن تق انول ولك chain(المیث length ( ب ي المرك دروكاربونات ف انول-١للھی ـ . الك ي ل ل الریاض ل) The nonrandom two liquid (NRTL)(المودی ـ و المودی ي ل الریاض )UNIQUAC( و المودیل الریاضي لـ)UNIFAC ( ة ة المدروس وار لالنظم ز األط ؤ بتراكی ة و للتنب ات العملی ع البیان ات م ربطھم بعالق . قد تم استخدامھم ل ة ) NRTL(الحسابات المبنیة على المودیل الریاضي لـ ات العملی ة ) experimental tie-line data(تعطي تمثیل جید للبیان ة المدروس ل األنظم ق . لك التواف .بین النتائج المستحصلة من العالقات الریاضیة والبیانات العملیة كان جیدًا جدًا !! !! . This page was created using Nitro PDF trial software. To purchase, go to http://www.nitropdf.com/ http://www.nitropdf.com/