Journal of Polytechnc, 2017; 20 (1) : 129-135 Journal of Polytechnc, 2017; 20 (1) : 129-135 Modelng Of Lqud-Lqud Equlbrum Data and Estmaton of New Bnary Interacton Parameters for NRTL Model for the Quaternary System / /1- Butanol / KCL at 298.15K Bourayou Euch. N 1, Mena A-H 2, ouaoura A. 3 1 Laboratore de L Ingénere des Procédés de L Envronnement (LIPE). Department of Process Engneerng Unversty Blda, Blda 9000, Algera 2 Department of Process Engneerng, Unversty Constantne, Constantne 25000, Alegra. 3 Department of Process Engneerng, Unversty Constantne3, Constantne 25000, Alegra. ( Receved : 26.09.2016 ; Accepted : 03.11.2016 ) ABSTRACT The present study concerns epermental measurements of the saltng-out effect on lqud-lqud phase equlbrum (LLE) of partally mscble systems such as water/ ethanol/1- butanol /Potassum chlorde at 298.15K. The salt KCl was used at dfferent mass percentages of 5, 10 and 15. The bnodal and te lne data were determned by the cloud pont method. The addton of the salt showed to be effectve n alterng the LLE of water/ ethanol/1- butanol system n favor of the solvent etracton of ethanol from the aqueous soluton partcularly at hgh salt concentratons. The epermental data were correlated consderng the modfed verson of the NRTL model for the actvty coeffcent, wth the estmaton of new bnary nteracton parameters correspondng to salt-solvent and solvent-solvent pars. The devatons between epermental and calculated compostons n both phases for the ternary system usng these new nteracton parameters showed the ablty of NRTL to determne lqud-lqud equlbrum data of the studed systems n the presence of the salts. Keywords: Salt Effect, Solublty, LLE, NRTL Model, Interacton Parameter. 1. INTRODUCTION The presence of a salt or a non-volatle solute n a solvent mture, s of ncreasng nterest due to ts nfluence on separaton processes n chemcal engneerng. The effect of salts addton n dstllaton had been studed by several nvestgators. The results have shown that t s possble by the addton of a sutable salt to alter the relatve volatlty of the system and also avod the formaton of an azeotrope. Smlarly, the separaton by solvent etracton also becomes ncreasngly more dffcult as the te-lne becomes parallel to the solvent as leadng to a solutropy whch has also to be avoded [1-4]. The effect of the addton of a sutable salt or a nonvolatle solute can be seen n two ways: graphcally, by the varaton n the sze of the two-phases regon as well as by sgnfcant changes n the te-lne slopes for a lqud-lqud equlbrum (LLE) mture, to an etent whch elmnates the solutropy, and nduces quanttatve changes of the solute dstrbuton coeffcent and the solvent selectvty. In recent years, lqud-lqud equlbra (LLE) have ganed great nterest n chemcal technology. Due to the energy rsng cost, new separaton processes based on etracton are gettng more attractve. Also t may be feasble to operate known processes at new condtons necesstatng checks for lqud phase *Sorumlu Yazar (Correspondng Author) e-posta: bourayou.nabla@gmal.com Dgtal Object Identfer (DOI) : 10.2339/2017.20.1 129-135 stablty at varous pont of the process hence the need for calculatng and predctng (LLE) compostons whch may be calculated usng any ecess bbs energy model. Santos et al [4] estmated new nteracton parameters between bnary pars of salt-solvent and solvent-solvent usng NRTL model to correlate the ternary system of water-ethanol-1pentanol n the presence of Potassum bromde (KBr) at 298.15K and 313.15K, n agreement wth Stragevtch and Avla [5]. These nteractons parameters are functons of the temperature. Tan s modfed NRTL model [6] has shown ts ablty to predct farly accurately the effect of salts on the actvty coeffcent of a solvent component n a soluton usng the solvent-salt nteracton parameters obtaned from approprate bolng pont data. A Tan s modfed NRTL model has been used to predct the salt effect on (LLE) as n (VLE). The equaton can be used to correlate the (LLE) data usng the solvent-solvent nteracton parameters of the same mture wthout salt, at the same temperature. The salt-solvent nteracton parameters can be assumed as a functon of the salt concentraton snce the bolng pont of a salt-solvent mture vares wth the concentraton of the dssolved salt [7]. In ths study, the epermental (LLE) of water, ethanol, 1-butanol and potassum chlorde at varous salt concentratons and at 298.15K were reported. The data were correlated by Tan s modfed NRTL model [6], due to ts smplcty and ts proven qute good predcton of the LLE data 129
Bourayou EUCHİ N, Mena A-H, ouaoura A. / Journal of Polytechnc, 2017; 20 (1) : 129-135 2. EXPERIMENTAL 2.1 Materals and 1-butanol were suppled by Merck Eurolab wth cerfed purtes hgher than 99.5 % and 98.5 %, respectvely. Ths was checked by refractve nde measurements at T=293.15K wth an accuracy of ±0.0004n D. Denstes were also measured usng an Anton Paar (model DMA 5000) densmeter wth a certfed precson of (±10-5 gcm -3 ). The results were compared to lterature data [8]. 2.2. Epermental Procedure Standard curves were determned by usng the 'Cloud- Pont' method whch s descrbed n detal by Letcher and Sswana [9]. The te lnes were determned by the method of the refractve nde descrbed n detal by Brggs and Comng [10]. The refractve nde of the two phases at equlbrum, represented, respectvely, by each etremty of the te- lnes was measured n order to determne ther compostons, usng the obtaned calbraton curve and then to draw pont by pont the equlbrum curves of the consdered system. Bnodal curves for each system were obtaned by preparng bnary mtures of known volumes of the two partally mscble consttuents and then by addng gradually the thrd consttuent by means of a burette. The obtaned ternary mture was shaken at room temperature, by means of an agtator provded wth a lamp to allow a good observaton of the dsappearance of the turbdty pont, after whch a small quantty of the soluton was taken to measure the refractve nde of each prepared sample by means of the refractometer (Type EUROMEX RD 645) wth an accuracy of ±0.0004n D. After the turbdty dsappearance the soluton was sampled by means of a mcroppette to measure the refractve nde of each sample. The bnodal curves of the ternary systems were used for the determnaton of the composton of the phases at equlbrum for varous te-lnes for the water/ethanol /1- butanol system at 298.15K. In order to determne epermental lqud-lqud equlbrum data ternary mtures of solvent, soluton and solute of volume equal to 10 ml but wth dfferent concentratons were prepared. The mture was placed n an equlbrum cell, wheret was agtated n order to allow an ntmate contact between the dfferent phases, and the thermodynamc equlbrum was fnally reached by lettng the mture at rest for 24 hours. The complete process was carred out at constant temperature, by usng a thermostatc bath. After a necessary rest tme, the mture spltted nto two clear and transparent lqud phases wth a well defned nterface, both phases were then separated by settlng. Samples of the organc phase and the aqueous phase were subjected to the measure of the refractve nde. The solublty of each consttuent n each phase was determned by etrapolaton from the bnodal curves of the refractve nde accordng to the fracton known for the consttuent n the soluton. 3. RESULTS AND DISCUSSION 3.1. Epermental Lqud-Lqud Equlbrum Data The compostons of the dfferent components of / / ternary system at 298.15K are presented n Table 2, where X denotes the mole fracton of the th component. Fgures 1 to 4 show the correspondng bnodal curves of the studed ternary mtures. The effect of the added salt amounts on the bnodal curves are also shown n these fgures where t can be seen that an ncrease n KCl amount led to an enlargement of the two phases regon, nducng an ncrease of and mscblty. Table 3 shows the epermental te-lne compostons of the equlbrum phases of the ternary system water/ ethanol/1-butanol at 298.15K. The superscrpts I and II denote the organc and the aqueous phases, respectvely. The comparson between the equlbrum data obtaned epermentally for the ternary system water/ ethanol/1- butanol at 298. 15K and those prevously reported n the lterature by Sorensen and Arlt [11] are presented n fgure 5. An ecellent concordance can be notced between both results, confrmng the relablty of the epermental procedure used n ths work as well as the good accuracy. Table 1. Pure components physcal propertes at 293.15K and 1 atm Component ρ/kgm -3 Refractve nde InR Ths work from Ref [8] Ths work from Ref[8] 0.7890 0.7894 1.3610 1.3605 0.8100 0.8098 1.3982 1.3981 0 Fg. 1. Bnodal curve of // system at 298.15K 130
MODELIN OF LIQUID-LIQUID EQUILIBRIUM DATA AND ESTIMATION OF NEW BI Journal of Polytechnc, 2017; 20 (1) : 129-135 0 0 Fg. 2. Bnodal curve of ///5% KCl system at 298.15K Fg. 4. Bnodal curve of //1Butanol/15% KCl system at 298.15K 0 Epermental 298.15K Sorensen and Arlt 298.15K 0 Fg. 3. Bnodal curve of //1Butanol/10% KCl system at 298.15K 0 0 Fg. 5. Epermental LLE of the / / ternary system at 298.15K Table 3. Epermental te-lne data for / / at 298.15K Organc phase I Aqueous phase II X X X X X X 0.5360 0.0560 0.4080 0.6000 0.0960 0.3040 0.6040 0.0970 0.2990 0.6610 0.0900 0.2490 0.7020 0.0860 0.2110 0.9600 0.0170 0.0230 0.9480 0.0240 0.0280 0.9420 0.0270 0.0310 0.9260 0.0360 0.0380 0.9140 0.0420 0.0440 131
Bourayou EUCHİ N, Mena A-H, ouaoura A. / Journal of Polytechnc, 2017; 20 (1) : 129-135 Table 4. Epermental and predcton te-lne data for / //KCl n a free-salt bass at 298.15K Organc phase I Aqueous phase II X X X X X X 5% KCl (XS=1.205%) Epermental data 0.4646 0.1580 0.3774 0.5219 0.2068 0.2713 0.5688 0.2020 0.2292 0.5908 0.1937 0.2155 0.6340 0.1732 0.1928 Predcted usng NRTL model 0.4601 0.1519 0.3880 0.5527 0.1905 0.2568 0.5667 0.1989 0.2343 0.5690 0.2007 0.2340 0.5669 0.1990 0.2302 0.9503 0.0313 0.0184 0.9224 0.0506 0.027 0.9070 0.0601 0.0329 0.8904 0.0695 0.0401 0.8878 0.0710 0.0412 0.9004 0.0511 0.0485 0.8994 0.0653 0.0352 0.8985 0.0686 0.0329 0.8983 0.0693 0.0324 0.8985 0.0686 0.0328 10% KCl (XS=2.410%) Epermental data 0.4320 0.1730 0.3953 0.4714 0.2316 0.2970 0.5281 0.2569 0.2148 0.5976 0.2420 0.1602 0.6314 0.2260 0.1425 0.9567 0.0278 0.0154 0.9374 0.0411 0.0214 0.9131 0.0568 0.0301 0.8771 0.0787 0.0441 0.8452 0.0975 0.0571 Predcted usng NRTL model 0.4129 0.1714 0.4158 0.4731 0.2223 0.3045 0.5413 0.2446 0.2141 0.5939 0.2517 0.1544 0.5987 0.2529 0.1485 0.9606 0.034 0.0540 0.9357 0.0513 0.0130 0.8997 0.07 0.0303 0.8506 0.0882 0.0612 0.8478 0.0901 0.0621 15% KCl (XS=3.620%) Epermental data 0.4986 0.1397 0.3617 0.6030 0.1117 0.2853 0.6396 0.1015 0.2589 0.6473 0.0994 0.2533 0.6742 0.0918 0.2340 Predcted usng NRTL model 0.5965 0.1077 0.3173 0.5987 0.1203 0.2935 0.5982 0.1239 0.2815 0.5975 0.1248 0.2786 0.5980 0.0862 0.2772 0.9702 0.0272 0.0260 0.8809 0.0895 0.0296 0.8371 0.1195 0.0434 0.8256 0.1274 0.0470 0.8141 0.1354 0.0505 0.8618 0.0852 0.0530 0.8521 0.0996 0.0483 0.8453 0.1077 0.0470 0.8430 0.1101 0.0469 0.8430 0.1106 0.0464 132
MODELIN OF LIQUID-LIQUID EQUILIBRIUM DATA AND ESTIMATION OF NEW BI Journal of Polytechnc, 2017; 20 (1) : 129-135 Fg. 6. Epermental LLE of the //1- Butanol/KCl system n a free-salt bass at 298.15K Fgure 6 shows the representaton of the effect of the addton of the salt on the water/ethanol/1-butanol system n a free-salt bass at 298.15K, through a pseudoternary dagram. It can be seen from fgure 6 that there s an ncrease n the area of mmscblty on addton of potassum bromde. The salt enlarged the area of the twophase regon; decreased the mutual solublty s of water margnally decreased the concentratons of ethanol and 1-butanol n the aqueous phase whle sgnfcantly ncreased the concentratons of the same components n the organc phase. These effects essentally ncreased the heterogenety of the system; whch s an mportant consderaton n desgnng a solvent etracton process. One way to analyze the salts effects s to compute the dstrbuton curve of the organc, say ethanol n the present case, between the aqueous and organc phases. Also n order to assess the salt affnty for the aqueous and organc phase s affnty, ts dstrbuton curve has been computed. The dstrbuton curves represented n Fgure 7 confrm that the addton of potassum chlorde actually dsrupted the dstrbuton of the ethanol between the 1-butanol and the water. We can see that the dstrbuton curves for each percentage of salt added are always above the frst bsector what ndcates that the coeffcent of dstrbuton (the report enters the quantty of the n the organc phase on that n the aqueous phase) s always upper to 1. 3.2. Parameters Estmaton And Data Correlaton New bnary nteracton parameters for the modfed verson of NRTL model whch takes nto account the presence of the salt [7], were obtaned from the epermentally measured data for the water/ethanol/1- butanol and potassum chlorde system at varous salt concentratons, by the mnmzaton of the objectve functon of the squared actvty dfferences for all components and over all the te-lnes, epressed as follows: I II 2 F j where and represent the mole fracton and the actvty coeffcent of component, respectvely for te lne, I and II denote the frst and second phases n equlbrum respectvely. The mnmzaton was performed accordng to the Smple search technque modfed by Nelder and Mead [12].The actvty coeffcent of component n the absence of salt s 0 calculate from the followng NRTL equaton [13] ln 0 j3 j1 k 3 k 1 k k j j3 k 3 j1 k 1 Where: g g RT / j k K j m3 m m1 k 3 k 1 kj m k and ep In whch g s the energy of nteracton between components j and I and the non-randomness factor. All they were assumed to be 0.2 [11]. In a salts- solvents mture, the actvty coeffcent S component s gven by accordng to [7] as follows: ln Fg. 7. Dstrbuton curves of obtaned n dfferent concentraton of salt for the ///KCl system n a free-salt bass at 298.15K s ln 0 s where s s the salt-solvent nteracton parameter of the The epermental lqud-lqud equlbrum data were correlated, performng lqud-lqud flash calculaton. For a ternary system, twelve nteracton parameters are necessary and must be calculated. It s about parameters: I, 12,, 13 21, 23,, 31 32 ( ), ( I 1s ), ( I ), 2s 3s ( II ), ( II 1s ) et II 2s ( ). 3s The naccuracy of the predcton was estmated between the measured and calculated mole fractons usng the root-mean square devaton (RMSD) defned as: RMSD N 2 3 Ep Cal 2 jk jk 1 1 1 2N N k j c 133
Bourayou EUCHİ N, Mena A-H, ouaoura A. / Journal of Polytechnc, 2017; 20 (1) : 129-135 Where N and N c are the numbers of te-lnes and components respectvely, Ep and Cal are the epermental and calculated mole fracton, respectvely, and, j and k =1, 2 N (te-lnes) for components, phases and te lnes, respectvely. roup nteracton parameters for par solvent-solvent for the system water/ethanol/1-butanol n the presence of KCl appear n Table 5. Interacton parameters between par solvent-salt are calculated n every phase organc and aqueous phases n the varous concentratons of potassum chlorde KCl, appear n Table 6. The nteracton parameters between pars salt-salt were not calculated because med-salt. The correlated results are gven n Table 4. In fgures 8-10, comparatve results between the calculated and epermental data are plotted together, for the quaternary system water/ethanol/1-butanol/kcl at 298.15K There s a good agreement between the epermental and calculated values. As well as the root-mean square salt. Ths study confrmed the great performance of NRTL model. Fg. 8. Comparng the epermental LLE of //1- Butanol/ 5%KCl wth the equlbra calculated from Table 5. NRTL bnary nteracton parameters j for the system // /KCl at 298.15K Concentraton (Salt-free bass)(g/100g) 5% KCl (X S =1.205%) 10% KCl (X S =2.410%) 15% KCl (Xs =3.620%) Component (j) j NRTL parameters - -1.503-0.832-1Butanol 3.327 4.649-1Butanol 17.48-4.985 - -0.783 3.285-1Butanol 7.835 0.675-1Butanol -2.611-2.717 - -0.311 0.201-1Butanol 4.514 0.158-1Butanol 4.043 5.743 RMSD 0.0208 0.013 0.0411 Table 6. NRTL bnary nteracton parameters s for the system water/ethanol/1-butanol /KCl at 298.15K Concentraton (Salt-free bass)(g/100g) 5% KCl (X S=1.205%) 10% KCl(X S =2.410%) 15% KCl (Xs =3.620%) Component (s) NRTL parameters -KCl 1.265 0.859 -KCl -0.069 1.017 -KCl 1.566-2.576 -KCl 0.847 0.236 -KCl 2.398 3.71 -KCl 1.467-1.372 -KCl 4.577 1.232 -KCl 0.584 1.243 -KCl 1.545 1.89 s s RMSD 0.0208 0.013 0.0411 devaton (RMSD) for all components are (0.0411, 0.0208). These values show that the NRTL model was able to represent better the equlbrum n the presence of NRTL model at 298.15K 134
MODELIN OF LIQUID-LIQUID EQUILIBRIUM DATA AND ESTIMATION OF NEW BI Journal of Polytechnc, 2017; 20 (1) : 129-135 and water and also decreased the concentraton of 1- butanol and ethanol n the aqueous phase whle sgnfcantly ncreased ther concentratons n the organc phase. The epermental data were correlated consderng the modfed verson of the NRTL model for the actvty coeffcent, wth the estmaton of new bnary nteracton parameters solvent-solvent and solvent-salt. The devatons between epermental and calculated compostons n both phases for the ternary system usng ths model were reasonable showng the ablty of model NRTL for the determnaton of LLE of solvents-salt system usng the solvent-solvent nteracton parameters gven on salt-free bass. Fg. 9. Organc phase Fg. 10. Comparng the epermental LLE of / / / 10%KCl wth the equlbra calculated from NRTL model at 298.15K Aqueous phase Comparng the epermental LLE of / / / 15%KCl wth the equlbra calculated from NRTL model at 298.15K 4. CONCLUSION Ths study concerned the saltng-out effect as appled to solvent etracton systems. Lqud-lqud equlbrum data for the partally systems of / /1- Butanol /Salts were epermentally measured at 298.15K and atmospherc pressure. The salt used was potassum chlorde at dfferent weght percentages of 10 and 15%. The bnodal and te-lne data results were determned by cloud pont measurements method. The addton of salt showed to be effectve n modfyng the LLE of / /1- Butanol system n favor of the solvent etracton of ethanol from aqueous soluton wth 1- butanol, partcularly at hgh salt concentratons. Potassum chlorde ncreased the area of the two-phase regon and decreased the mutual solubltes of 1-butanol REFERENCES 1. Ramana, R. and Subba Reddy. Salt Effect on Solute Dstrbuton n Lqud-Lqud Equlbra, 13:9, 79-81, (1978). 2. Narayana, A.S.; Nschal, R.; Patel, R.; Parkh, K.. and Sngh, R.K. Salt Effect n Lqud-lqud Equlbra of Acetc Acd--Benzene System. J. Chem. Eng. Comm, 95: 41-46, (1990). 3. Edwn,O.; Esen and Joseph Joffe.. Salt effects n Lqud- Lqud Equlbra, J. Chm. Eng., 11: 4, (1996). 4. Santos,.R.; d'avla, S.. and Aznar. Brazlan Journal of Chemcal Engneerng.17: (04-07), 721-734, (2000). 5. Stragevtch, L. and S.. d'avla. Applcaton of a generalzed Mamum Lkelhood Method n the Reducton of Multcomponent Phase Equlbrum Data. Braz. J. Chem. Eng, 41-52, (1997). 6. Tan, T.C.. Trans. Inst. Chem. Eng., Part A, 68: 93-103, (1990). 7. Tan,T.C., Aravnth,S. Lqud-Lqud equlbra of water/acetc acd/1-butanol system effect of sodum potassum) chlorde and correlaton. Flud Phase Equlbra, 163: 243-257, (1999). 8. Ytzhak, M.. The Propertes of Solvents., John Wley and Sons,4: 239, (1999). 9. Letcher, T. M.; Szwana, P. M.. Lqud-lqud equlbra for mtures of alkanol +water + methyl substtuted benzene at 25 C, Flud Phase Equlbra. 74: 203-217, (1992). 10. Brggs, S. W.; Comngs, E. W. Te-lne correlatons and plat pont determnaton, Ind. Eng. Chem., 35: 411-415, (1943). 11. Sorensen, J.M.; Arlt, W. Lqud-lqud equlbrum data collecton. Dechema. Chemstry data seres,5(2): (1987). 12. Nelder, J.A.; Mead, R..Comp, 7: 7-308, (1965). 13. Renon, H., Prausntz, J. M. Local Compostons n Thermodynamc Ecess Functons for Lqud Mtures. AlChE J. USA, 14(1): 135-44l, (1968). 135