Salt effect on the liquid–liquid equilibrium of the ternary(water+phenol+methyl isobutyl ketone)system:Experimental data and correlation☆

Yun Chen*,Kangning Xiong,Shuai Shen,Huimin Wang,Shaoming Zhou,Libo Li*

Department of Chemistry and Chemical Engineering,South China University of Technology,Guangzhou 510640,China

State Key Laboratory of Pulp and Paper Engineering,South China University of Technology,Guangzhou 510640,China

Keywords:Liquid–liquid equilibrium Methyl isobutyl ketone Salt effect Phenol

ABSTRACT The effects of NaCl,KCl and Na2SO4 on the liquid–liquid equilibrium(LLE)data for the ternary system,water+phenol+methyl isobutyl ketone,were determined at 0.101 MPa and 333.15 K and 343.15 K.The nonrandom two-liquid(NRTL)model was used to correlate the experimentaldata and to yield corresponding binary interaction parameters for these salt containing systems.The Hand and Othmer–Tobias equations were used to confirm the dependability of the determined LLE data in this work.Distribution coefficient and selectivity were used to evaluate the extraction performance of methyl isobutyl ketone with the existence of salt.The magnitude of salt effect on the water+phenol+methyl isobutyl ketone(MIBK)system is in the following order:Na2SO4＞NaCl＞KCl.

1.Introduction

Phenol,a highly toxic chemical substance,often exists in wastewaters discharged from a wide range ofindustrialprocesses such as coking operations,coal gasification,refineries and petrochemical manufacture[1].The removal of phenol from wastewater has attracted broad attention over decades.Solventextraction is a very efficientmethod to treatphenolic wastewater[2],and liquid–liquid equilibrium(LLE)data are essential for extraction process design and optimization[3–5].In recent years,many LLE data for various ternary systems containing extractant,phenol and water have been published.Lei et al.reported LLE of 2-Methoxy-2-methylpropane as the solvent to extract phenol from aqueous systems from 298.15 K to 323.15 K[1,6].Liu et al.studied a ternary system containing methyl tert-butyl ketone,phenol and water from 298.15 K to 323.15 K[7].Mafra and Kr?henbühl also investigated this system at T=(323.15,333.15)K[8].Chen et al.measured tie line data and calculated distribution coefficients for two ternary systems,(methyl butyl ketone+phenol+water)and(methyl butyl ketone+hydroquinone+water),at 298.15 K and 323.15 K[9].Methyl isobutyl ketone has attracted extensive research interests as a newly developed extractant for phenols due to its excellent chemical or physical properties and high distribution coefficients forphenolic compounds(e.g.phenoland hydroquinone).Yang etal.studied a ternary system,containing MIBK,phenol and water,from 298.15 K to 323.15 K[10],and a quaternary system,containing MIBK,phenol,hydroquinone and water at 298.15 K[11].However,salt effect on the liquid–liquid equilibrium system(water+phenol+methyl isobutyl ketone)has not been reported yet.In addition,the practical extraction temperature with ketone as the extraction solvent is usually between 333 K and the azeotropic temperature with water to avoid the paraf fin(melting point:328.15–335.15 K)in the wastewater clogging extraction device[12].Thus to study MIBK's extraction performance for phenol with the existence of salt at temperatures above 333 K is of great significance in terms of both chemical engineering and environmental protection.

Inorganic salts can significantly affect the equilibrium behavior[13–16]with the so-called salting-in(increase a solute's solubility)or salting-out(decrease the solubility)effects[17–19].It is usually used to enhance separation processes and increase the separation efficiency.A straightforward explanation of salting out effect is based on ion's“hydration”nature:when some salt is added to the aqueous solution of an organic compound,the stronger af finity between salt ions and water molecules will decrease the available water molecules for the organic compound.In other words,the salt will increase the “effective”concentration of phenol in the aqueous solution,thus improve its extraction efficiency[20].In an industrial extraction process with salt,the salt concentration or used amount is usually optimized based on the balance of a range of considerations,such as the extraction ef ficiency,the recovery of salt and relevant economic efficiency,etc.

In this work,the objective is to determine MIBK as the extractant to remove phenol from water in the presence of NaCl,KCl or Na2SO4(5 wt%ofthe totalwater mass in the system)attemperatures of333.15 Kand 343.15 K under 0.103 MPa.The experimental LLE data were compared with those predicted by the NRTL model.The quality of experimental tie line data was verified by the Hand and Othmer–Tobias equations.

2.Experimental

2.1.Materials

The chemicals used in this work withoutfurtherpurification,and the purity were detected by gas chromatography.More details of those chemicals are listed in Table 1.Deionized water was used throughout all experiments in this work.

2.2.Apparatus and procedure

The experimental tie-line data were determined at 333.15 K and 343.15 K under 0.101 MPa.A 5 ml mixture solution of phenol and MIBK was added to another 50 ml aqueous solution of an inorganic salt(5 wt%ofNaCl,KClor Na2SO4)in a 100 mlcustomized glass equilibriumvessel.A thermostatic bath,with a fluctuation of±0.1 K,was used to controlthe vesseltemperature constant.This prepared mixture in the vessel was agitated intensively for more than 2 h and then settled for at least 20 h to reach the phase equilibrium completely.After the phase equilibrium was achieved,which leads to two liquid phases forming from the mixture,each phase was sampled by a syringe.A gas chromatograph(GC6820,Agilent Technologies)equipped with a flame ionization detector(FID)and a DB-5MS capillary column(30 m×0.32 mm× 0.25 μm)was used to analyze the compositions of the samples.The mass fraction of MIBK and phenol in both phases was determined by the internal standard method,where the internal standards of MIBK and phenol are n-propyl acetate and n-octanol,respectively.The content of water in the aqueous phase was determined by subtracting the mass fraction ofMIBKand phenolfrom 1,and the contentofwaterin organic phase was measured by Karl Fischer titration.Methanol was used as the organic solvent for the GC analysis,because it separated different analytes well(the retain times are,tMethanol=0.8 min,tMIBK=2.0 min and tphenol=4.4 min),and showed low toxicity.The temperature of the injector and detector of GC was maintained constant at 533.15 K and 523.15 K,respectively.The temperature of the GC oven was first set at 313.15 K for 2 min and then increased with a rate of 30 K·min-1for 4 min.Nitrogen was chosen as the carrier gas and the flow rate was set at 30 ml·min-1.To guarantee the quality of the experimental data,each sample was GC analyzed at least 3 times with standard deviations below 0.002 and the mean value was reported in this work.The saltconcentration in the organic phase was found to be negligible after it had been evaporized.

3.Results and Discussion

3.1.Experimental LLE data

The LLE data expressed in mass fraction for the system MIBK+pheenol+water in the presence of different salts at 333.15 K and 343.15 K are presented in Tables 2–4.The saltconcentration w3in the LLE system was the same like that in the initial aqueous solution,5 wt%,since the volume of the organic phase was less than 1/10 of the aqueous phase,which took in very little water(the water mass fraction in the organic phase was below 0.1,see Tables 2–4).The LLE data for the studied ternary system at two different temperatures are also depicted in Figs.1–6 as the triangular diagrams.

Table 1 The source and purities of the chemicals used

Table 2 Experimental LLE data(mass fraction)for ternary system MIBK(1)+phenol(2)+water(3)①in the presence of NaCl at 333.15 K and 343.15 K under 0.101 MPa

Table 3 Experimental LLE data(mass fraction)for ternary system MIBK(1)+phenol(2)+water(3)①in the presence of KCl at 333.15 K and 343.15 K under 0.101 MPa

Table 4 Experimental LLE data(mass fraction)for ternary system MIBK(1)+phenol(2)+water(3)①in the presence of Na2SO4 at 333.15 K and 343.15 K under 0.101 MPa

The distribution coefficient(D)and selectivity(S)are calculated to estimate the extraction efficiency ofMIBKto remove phenol from aqueous systems and are expressed as follows:

where w21and w23denote the mass fraction of phenol in the organic phase and aqueous phase,respectively,the mass fraction of water in the organic phase and aqueous phase is presented as w31and w33,respectively.Tables 2–4 listed the value of distribution coefficient and selectivity of MIBK for phenol at the studied systems.These results indicate that,MIBK's miscibility with wateris negligible and its distribution coefficients for phenol are quite high.The distribution coefficients for different salts are in the order of:Na2SO4＞NaCl＞KCl.For the same salt,the distribution coefficient at 333.15 K is relatively large than at 343.15 K.In addition,the distribution coefficients in the presence of salts are considerably higher than those without salt(D＜90)[21].

Fig.1.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 333.15 K in the presence of NaCl:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

Fig.2.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 343.15 K in the presence of NaCl:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

Fig.3.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 333.15 K in the presence of KCl:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

Fig.4.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 343.15 K in the presence of KCl:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

Fig.5.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 333.15 K in the presence of Na2SO4:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

Fig.6.Ternary diagram for experimental LLE data of{MIBK(1)+phenol(2)+water(3)}at 343.15 K in the presence of Na2SO4:■-■,experimental data;△,calculated from NRTL model.The error bars of experimental data less than 0.002.

The Hand[22]and Othmer–Tobias[23]equations were used to check the reliability of the experimental LLE data,and the expressions of those equations are shown as follows.

The fitting parameters a1,b1,a2and b2for the Hand and Othmer–Tobias equations and corresponding coefficients R2are listed in Table 5.The straight lines plotted by these two equations are also presented in Figs.7,8.All coefficients(R2)are close to 1,which indicates a high reliability of the experimental tie-line data.

Figs.9 and 10 compare the distribution coefficients in the presence of NaCl,KCl and Na2SO4with the salt free system[21]at 333.15 and 343.15 K.These results show that,the existence of salts improves the extraction efficiency forphenol,agreeing with other LLE studies[24,25].The salting out efficiency for different salts is in the order of:Na2SO4＞NaCl＞KCl,within phenol concentration investigated in the work.In addition,the distribution coefficients increase with the decrease the temperature or phenol concentration.

Table 5 Fitting parameters and correlation factors(R2)in Hand and Othmer–Tobias equation for studied ternary systems

Fig.7.Hand plots for{MIBK+phenol+water}in the presence of salts at different temperatures:■,NaCl at 333.15 K;□,NaCl at 343.15 K;●,KCl at 333.15 K;○,KCl at 343.15 K;▲,Na2SO4 at 333.15 K;△,Na2SO4 at 343.15 K.

Fig.8.Othmer–Tobias plots for{MIBK+phenol+water}in the presence of salts at different temperatures:■,NaCl at 333.15 K;□,NaCl at 343.15 K;●,KCl at 333.15 K;○,KCl at 343.15 K;▲,Na2SO4 at 333.15 K;△,Na2SO4 at 343.15 K.

Fig.9.Distribution coefficient for phenol in the presence of different salts at 333.15 K:■,NaCl;△,KCl;▲,Na2SO4;□,salt free.The error bars of experimental data less than 0.002.

3.2.Data correlation

The NRTL model was used to correlate the experimental LLE data,with non-randomness parameters(αij)presented in Table 6.In the NRTL model,the binary interaction parameters(bijand bji)were calculated by minimizing the objective function(OF):

where n refers to the total of the tie lines,the superscripts of exp and cal denote the experimental LLE data and the NRTL calculated data,respectively.wijkrepresents the mass fraction of the component i in j phase of k th tie line.The standard deviation of temperature and mass fraction is denoted as σTand σw,respectively.

The root-mean-square-deviation(RMSD)was used to assess the agreement between the experimental data and the calculated results,defined as:

Fig.10.Distribution coefficient for phenol in the presence of different salts at 343.15 K:■,NaCl;△,KCl;▲,Na2SO4;□,salt free.The error bars of experimental data less than 0.002.

Table 6 NRTL binary interaction parameters(bij and bji)and RMSD values for studied systems

The corresponding binary interaction parameters(bijand bji),together with corresponding RMSD values,are shown in Table 5.We note that,the in fluence of salt(e.g.NaCl,KCl)is implicitly incorporated into the binary interaction parameters,as we did in developing implicit solvent models,e.g.Field-SEA,for salt solutions[26].All RMSD values in Table 6 are less than 1%,indicating the NRTL model represents the tieline data for the salt containing systems very accurately.

The experimental data and calculated results from the NRTL model for each salt and each temperature are shown in Figs.1–6.The excellent agreement between experimental data and calculated results indicates NRTL model is suitable for simulating phenol extraction.

4.Conclusions

Tie-line data for the quaternary systems,containing MIBK,phenol,water and salt,were investigated at 333.15 K and 343.15 K under 0.101 MPa.The high values of distribution coefficients and selectivity show the high separation efficiency ofMIBK as the extractant to recover phenol from wastewater.The experimental data were proven to be highly consistent by the Hand and Othmer–Tobias equations.They also agree very well with results predicted by the NRTL model.The experimental LLE data and distribution coefficients show the magnitude of salt effect follows the order:Na2SO4＞NaCl＞KCl.