Lianzheng Zhang ,Jie Wang ,Lin Yang ,Dongmei Xu ,Yixin Ma ,Jun Gao, *,Yinglong Wang
1 College of Chemical and Biological Engineering,Shandong University of Science and Technology,Qingdao 266590,China
2 College of Chemical Engineering,Qingdao University of Science and Technology,Qingdao 266042,China
Keywords:Vapor-liquid equilibrium Ionic liquids NRTL Alcohol-ester azeotrope COSMO-SAC model
ABSTRACT Based on the COSMO-SAC model,1-ethyl-3-methylimidazolium acetate and 1-ethyl-3-methylimidazolium p-toluenesulfonate were selected from 30 ILs as entrainers to investigate the separation of the isopropyl alcohol+isopropyl acetate azeotrope.Two screening indicators,σ-profile and infinite dilution selectivity (S∞),were adopted as the basis.The isobaric vapor-liquid equilibrium experiments for isopropyl alcohol+isopropyl acetate binary system and isopropyl alcohol+isopropyl acetate+confirmed ILs ternary systems were performed at the pressure of atmospheric pressure.The experimental measurement demonstrated that the adopt ILs enhanced the relative volatility of the above alcohol-ester azeotrope,leading to the elimination of the azeotropic point with a certain amount ILs.Meanwhile,the thermodynamic correlation for two systems containing ILs was explored with the NRTL model,which also reflects the extensive applicability of that by comparing the deviation between experimental and calculated data.And its binary interaction parameters were regressed,which can provide a basis for its simulation process.
In many fields of chemical production,the alcohol-ester azeotropic mixture belongs to a kind of system which is virtually impossible to separate efficaciously by general distillation,and this leads to the trouble in the acquisition and recovery of high-purity alcohol and ester.The isopropyl alcohol(IPA)and isopropyl acetate(IAC)mixture is a representative alcohol-ester azeotrope generated extensively during the preparation of cosmetics,plastics,spices,coatings,pharmaceuticals and daily chemical industry[1,2].Moreover,it is inevitably to form the azeotrope with the lowest constant boiling point during the industrial esterification process with IPA as raw material[3].In view of the above,the urgent requirements are put forwarded for the separation of alcohol-ester type azeotropes,such as IPA+IAC(the azeotropic composition in mole fraction is 0.6631:0.3369 with a minimum-boiling azeotropic at 353.59 K[4]),with effective and simple method.In recent decades,extractive distillation with the addition of ionic liquids (ILs) as entrainers for breaking the azeotrope of alcohol-ester systems has getting far-ranging attention and increasingly in-depth discussion [5-9].
ILs,as a hot research field with fascinating prospects,have been received widespread concern for the desirable features such as negligible vapor pressure,excellent stability and strong designability.It has incrementally replaced the conventional organic solvents in part of separation and purification field [10-12].To explore the action principles of ILs breaking azeotropes,also make an evaluation of the influence obtained by the common adopted imidazolium-based cations with different anions,the summarization about separating alcohol-ester systems by those ILs entrainers were finished.Such as different length alkyl substituted imidazolium cations with diethylphosphate (DEP),dibutylphosphate(DBP),tetrafluoroborate (BF4),trifluoromethansulfonate (OTf),chloride (Cl),bromide (Br),disulfate (HSO4),etc.anions [6,13-17],and analyzed the role of imidazolium-based cations and anions for its corresponding azeotrope mixture separation.The results demonstrated that the above alkyl substituted imidazolium-based ILs produced a remarkable salting-out effect,which significantly changed the relative volatility of those azeotropes and eliminated its azeotropic point.Furthermore,noticeable results can be obtained that the polarity anions have a stronger role in interaction with azeotropes,while the alkyl substituent length of the imidazolium cations led to little effect.And this was also verified by comparing the VLE results of (ethyl acetate+2-propanol) system with 1-Butyl-3-methylimidazolium hexafluorophosphate [BMIM][PF6] and 1-Hexyl-3-methylimidazolium hexafluorophosphate [HMIM][PF6] as entrainers [18].Meanwhile,three alkylimidazolium-based ILs with acetate anion were selected as entrainers to separate (acetonitrile +isopropanol) azeotropes,the relative volatility change was difference but not large [19].
After consulting the literatures,little works were reported by adding ILs to achieve the separation of IPA+IAC azeotrope.Andreattaet al.successively investigated the densities,refractive indices,dynamic viscosities and vapor-liquid equilibrium (VLE)data of binary and ternary mixtures composed of IPA,IAC,and[C4/C8MIM][NTf2]) at atmospheric pressure [20,21].Zhanget al.[22]adopt three different length of alkyl(hexyl/octyl/decyl)substituted imidazolium-based ILs with acetate anion as azeotrope breakers,determined the VLE data and explored its separation performance (followed the following order: [HMIM][Ac] >[OMIM][A c] >[DMIM][Ac]),and their work indicated that the shorter the alkyl substitution chain length,the better the separation effect.From the previous works,the good results of alkyl substituent imidazolium-based ILs with polarity anions can change the relative volatility of the alcohol-ester azeotropes were verified.And the common alkyl substituent the imidazolium cations shown little contribution to the change of relative volatility of azeotropes,while polarity anions play the decisive role.
Therefore,ILs entrainers with more efficient polarity anions need to be selected,which can break the azeotropic point and maximize the relative volatility to effective separate IPA+IAC azeotrope.Fortunately,the COSMO-SAC model [23,24] has been proved to be an effective method to select possible ILs as entrainers,and this can be used to design or select the appropriate functional structures of ILs for the separation of azeotrope mixtures.Hence,1-ethyl-3-methylimidazolium acetate ([EMIM][Ac]) and 1-ethyl-3-methylimidazoliump-toluenesulfonate ([EMIM][Tos])were selected and confirmed as the IPA+IAC azeotrope breakers from 30 kinds of ILs,randomly composed of 6 anions and 5 cations,according to the σ-profile analysis and the infinite dilution selectivity (S∞) obtained by referring COSMO-SAC model.Moreover,in order to verify the separation effect of the selected ILs,the isobaric VLE experiment for IPA+IAC+[EMIM][Ac]/[EMIM][Tos] ternary systems were carried out at the pressure of atmospheric pressure.Further,the NRTL model was employed for data correlation and regression.And its corresponding binary interaction parameters were regressed,which can provide a significant reference for the separation of IPA+IAC azeotrope mixture.
The experimental chemicals involved were purchased directly,of which the purity was examined by gas chromatography (GC,Scion 456C)or Karl Fischer titration before using and no impurities with obvious influence were discovered.Specific information about chemicals related to this work is displayed in Table 1.
The experimental VLE behavior of binary and ternary systems containing [EMIM][Ac]/[EMIM][Tos] under atmospheric pressure were explored by a modified Rose type recirculating vapor-liquid equilibrium still.The concrete introduction for the operating procedures has been published in related literatures [25-27].The required equilibrium temperatures were determined by an accurate mercury thermometer.The specific experimental device is drawn in Fig.1.
For VLE measurement of ternary system containing ILs,the same IL mole fraction of the mixture,IPA+IL and IAC+IL,were produced firstly.The prepared IAC+IL mixture was transferred into the VLE still (On account of the partial miscibility of IAC+IL system as well as the high viscosity of IL,a small amount of IPA was added into IAC+IL mixture to obtain an IPA+IAC+IL homogeneous system).Then,the mixing solution of IPA+IL was decanted into IAC+IL mixtures sequentially,so as to ensure that the mole fraction of IL remained constant.The vapor and liquid phases were continuously circulating to get the establishment of equilibrium and ensure adequate contact between the two phases.After 50 min circulating,the equilibrium conditions were occurred,then,appropriate amount samples were extracted from condensed vapor and liquid phases at the same time for analysis by gas chromatography (GC).
The GC (Scion 456C) was applied to analyze the equilibrium compositions of the extracted samples in VLE measurements with the area calibration normalization method.The specific operation requirements of GC are presented in Table 2.On account of the negligible vapor pressure and minimal volatility,ILs only exists in liquid phases,while only IPA and IAC can be detected in vapor phases.Hence,the content of [EMIM][Ac]/[EMIM][Tos] was obtained with the gravimetric method according to the quality change of liquid phase samples before and after vacuum drying.Furthermore,to reduce the deviation possibly,all the them were determined at least three times and average the results with the deviations less than 0.001.
Table 1 Specific information of the experimental chemicals
Table 2 The specific operating requirements of GC
Table 3 Measured results (equilibrium temperature T,liquid phase mole fraction x1,vapor phase mole fraction y1)of isobaric VLE experiment for the IPA(1)+IAC(2)system at 101.3 kPa.
After consulting the references for the separation of alcoholester azeotrope using ILs as entrainers as well as analyzing the physicochemical properties of (IPA+IAC) system,30 kinds of ILs randomly composed of 5 cations and 6 anions were tried as candidates for further screening.By analyzing the σ-profiles,the polarity of molecules or ions can be gotten,by which a suitable evaluationof hydrogen bond donator or accepter of that can be made.The van der Waals surface electrostatic potentials were also obtained by Multiwfn 3.8 software in combination with VMD 1.9.3 software.Accordingly,in order to explore the mechanism of ILs breaking the azeotrope of IPA and IAC,the polarity analysis of cations and anions with (IPA+IAC) azeotrope system were plotted respectively,which are displayed in Figs.2 and 3.
Fig.1.VLE experimental device diagram: (1) mercury thermometer,(2) glass VLE chamber,(3) liquid phase sampling port,(4) heating rod,(5) vapor phase sampling port,(6) condenser.
In addition,the open source and available COSMO-SAC model was adopted to obtain the related COSMO files and σ-profiles that acquiring the polarity results for the molecules or ions.The details of the COSMO-SAC model have been described in the previous works [24,28].Then,the detailed calculations settings using the COSMO-SAC model are described in detail and expressed in the Supplementary Material.From Fig.2,it can be observed that σprofiles are divided into three regions by two vertical dotted lines(σ=±0.0084 e·?-2(1 ?=0.1 nm) as cut-off value),which from left to right are hydrogen bond donor region,nonpolar region and hydrogen bond acceptor region.The surface screening charge density of IAC extends from the nonpolar region to the hydrogen bond acceptor region and reaches to +0.014 e·?-2,which illustrates the greater hydrogen bond accepting ability of IAC to serve as an acceptor.While that of IPA extends from the nonpolar region to both sides with the range -0.015 e·?-2<σ <+0.016 e·?-2,for which IPA can act as both receptor and donor.As a result,it is required that the determined anion has the stronger polarity as well as the greater hydrogen bond accepting ability,so as to weaken the interactions between IPA and IAC.In the meantime,the polar regions of [EMIM]+,[BMIM]+,[MMIM]+,[OMIM]+and[HMIM]+were almost the same,so the sample and easy prepared[EMIM]+and [BMIM]+were confirmed as the cations of entrainers preliminarily according to the reviewed literature.
Fig.2.Polarity analysis for IPA,IAC and cations ([EMIM]+,[BMIM]+,[MMIM]+,[OMIM]+,[HMIM]+).
Fig.3.Polarity analysis for IPA,IAC and anions ([Ac]-,[OTf]-,[HSO4]-,[SCN]-,[NTf2]-,[Tos]-).
It is evident from Fig.3,all of the selected anions have a great hydrogen bond accepting ability with the partial peaks in the σ >+0.0084 e·nm2region.A substantial part of peaks for the six anions ([Ac]-,[OTf]-,[HSO4]-,[SCN]-,[NTf2]-,[Tos]-) were at-0.005 e·?-2<σ <+0.025 e·?-2,-0.000 e·?-2<σ <+0.015 e·?-2,-0.018 e·nm2<σ <+0.016 e·?-2,-0.000 e·?-2<σ <+0.018 e·?-2,-0.002 e·?-2<σ <+0.012 e·?-2and-0.007 e·?-2<σ <+0.018 e·?-2,respectively.In addition,the polarization charge density of anions was positively correlated with the hydrogen bonds accepting ability [29].The [Ac]-and [Tos]-anions peaks were only distributed in the nonpolar region and hydrogen bond acceptor region,for which the [Ac]-and [Tos]-anions can only serve as hydrogen bond acceptors.Compared with the other four anions,[Ac]-and [Tos]-have greater polarity,which reflects the stronger hydrogen bond accepting ability of that.
Meanwhile,the van der Waals surface electrostatic potentials were adopted as supplement with the σ-profiles.As shown in Figs.2 and 3,the minima surface of IPA appeared above O atom with blue regions,however,darker red regions appeared above H atom on the hydroxyl group which appeared as maxima.Moreover,the minima surface of IAC appeared above two O atoms with two blue regions.Compared the red region of IAC with IPA,it is less capable of binding to molecules or ions containing negative electrostatic potential.This was well in accordance with the σ-profile analysis mentioned above,which indicates that ions or molecules with hydrogen acceptor distribution can form hydrogen bonds with IPA,and thus selectively separate IPA from the azeotropic system of IPA and IAC.
It is noticed that the electrostatic potentials of [EMIM]+and[BMIM]+are red,while[Ac]-and [TOS]-are blue,and the minima of[Ac]-and[TOS]-are located above the O atoms.It was observed that the shielding charge distribution of both cations was mainly in the hydrogen donor region,while the anions were mainly in the hydrogen acceptor region.The shielding charge of [Ac]-was more divided in the hydrogen acceptor region compared with [TOS]-.Therefore,it is mainly the anion that plays a role in the separation of IPA,and the anion increases the polarity of the IL.The cationic alkyl chain growth also increases the polarity of the IL little.When forming complexes with IPA,molecules with stronger polarity can produce stronger intermolecular interaction,which can better break azeotropy.As a result,[Ac]-and[Tos]-act as hydrogen bond acceptors were easier to form strong hydrogen bond with IPA as hydrogen bond donor.Thus,[EMIM]+etc.imidazolium cations with[Ac]-/[Tos]-two anions were verified as extractant candidates for further study.
With the COSMO-SAC model,the infinite dilution selectivity(S∞) was chosen as the primary screening index of ILs,which was calculatedviathe comparison of infinite dilution activity coefficients (γ∞) of IPA and IAC in [EMIM][Ac]/[EMIM][Tos].The formula for calculation is shown in Eq.(1):
hereAandBrepresents IPA and IAC,refer to the infinite dilution activity coefficients of IPA and IAC in [EMIM][Ac]/[EMIM][Tos],respectively.In general,the greater the infinite dilution selectivity,the stronger the interaction capacity of ILs.As is shown in Fig.4,[EMIM][Ac]/[BMIM][Ac] and [EMIM][Tos]/[BMIM][Tos] have higher infinite dilution selectivity compared with the other eight ILs,which were more favorable for extractive distillation process.In sum,the infinite dilution selectivity of the investigated ILs varies with the increase of alkyl side chains in cations a little.Nevertheless,for ILs composed of the same cation and different anions,it can be seen that the influence of infinite dilution selectivity by anions was great.Combined with the polarity analysis above,[Ac]-and [Tos]-have stronger hydrogen bond accepting ability,which can act as hydrogen bond acceptors that easier to form strong hydrogen bond with IPA as hydrogen bond donor,so as to weaken the interactions between IPA and IAC.Hence,the easy prepared[EMIM][Ac] and [EMIM][Tos] were chosen to separate the IPA+IAC azeotrope in this work.
Fig.4.Infinite dilution selectivity of IPA to IAC based on ILs.
Fig.5.The x1-y1 diagram for the binary system of IPA(1)+IAC(2)at 101.3 kPa:●,experimental values;○,literature values [21].
The isobaric VLE experiment for (IPA+IAC) system was also performed firstly to verify the stability as well as the accuracy of the used VLE still,and the measured results are given in Table 3.Meanwhile,experimental values and numerical results from the reported literature [21] were compared and plotted in Fig.5.As demonstrated,the VLE data measured were essentially in agreement with the reported literature,which reflect the reliability of the VLE still adopted in this work.
The isobaric VLE experiment for the ternary systems IPA(1)+IAC(2)+[EMIM][Ac]/[EMIM][Tos](3),[EMIM][Ac]molar content of 4%,5% and 8% and [EMIM][Tos] molar content of 5%,10%and 15%,were determined at atmospheric pressure,and all of the measured data are presented in Tables 4 and 5,respectively,wherex3stands for the mole fraction of IL[EMIM][Ac]/[EMIM][Tos]in liquid phase,represents the mole fraction of IPA in liquid phase with the absence of [EMIM][Ac]/[EMIM][Tos],γ1is the activity coefficient of IPA,a12refers to the relative volatility of IPA (1) to IAC (2),of which the corresponding calculation formula is as Eq.(2):
Table 4 VLE results (equilibrium temperature T,liquid phase mole fraction x1, ,vapor phase mole fraction y1) of IPA (1)+IAC (2)+[EMIM][Ac] (3) ternary system at atmospheric pressure.
Table 4 VLE results (equilibrium temperature T,liquid phase mole fraction x1, ,vapor phase mole fraction y1) of IPA (1)+IAC (2)+[EMIM][Ac] (3) ternary system at atmospheric pressure.
Note: Standard uncertainties u are u(T)=0.01 K, u(p)=0.1 kPa, u(x’)= u(x)= u(y)=0.001.
Table 5 VLE results (equilibrium temperature T,liquid phase mole fraction x1,,vapor phase mole fraction y1) of IPA (1)+IAC (2)+[EMIM][Tos] (3) ternary system at atmospheric pressure.
Table 5 VLE results (equilibrium temperature T,liquid phase mole fraction x1,,vapor phase mole fraction y1) of IPA (1)+IAC (2)+[EMIM][Tos] (3) ternary system at atmospheric pressure.
Note: Standard uncertainties u are u(T)=0.01 K, u(p)=0.1 kPa, u(x′)= u(x)= u(y)=0.001.
Only high volatile components in vapor phase can be detected in ILs containing systems when the equilibrium state was established.Thus,the content of ILs in vapor phase can be considered ignorable due to the inappreciable vapor pressure as well as thelow volatility.Moreover,the VLE measurements were explored at atmospheric pressure.In consequence,the activity coefficient calculation for componenti(γi)is defined using the following expression [30]:
wherexiandyirefer to the mole fraction of componentiin liquid phase and vapor phase,Prepresents the system atmospheric pressure (101.3 kPa),andis the saturation vapor pressure of pure componenti,which was roughly estimated by the extended Antoine equation:
All parameters in the equation were gotten from the Aspen databank,which are listed in Table 6.Moreover,the values of activity coefficient(γi)estimated by Eq.(3)are also presented in Tables 4 and 5,respectively.
Table 6 Parameters for the Antoine equation
Table 7 Regressed parameters for the NRTL model
The accepted NRTL model was applied to correlate the determined VLE data for IPA (1)+IAC (2)+[EMIM][Ac]/[EMIM][Tos](3) ternary systems in this work,its corresponding binary nonrandomness and interaction parameters were correlated based on the determined VLE data.The non-ideal liquid phase activity coefficient of constituentiis calculated as the following equations:where γistands for the activity coefficient of componenti,xi,xjandxkrefer to the molar content in liquid phase of componenti,jandk,respectively,τijandaijwere the interaction parameter and the nonrandom parameter between componentiandj,Δgijwas the interaction energy parameter,RandTrepresent the ideal gas constant and equilibrium temperature of the system,respectively.Meanwhile,the root mean square deviation(RMSD)was adopted to evaluate the quality of the correlation results using the NRTL model.The calculation equation is listed as below:
wherenrepresents the number of measured data points,andrepresent the experimental values and calculated values of activity coefficient,respectively.The corresponding NRTL interaction and non-randomness parameters are listed in Table 7.
As illustrated in Table 7,the RMSD value calculated using the NRTL model were 1.84% and 1.65% respectively,which reflects the good consistency between the experimental data and the regressed results.It can be notice that the NRTL model was suitable for the correlation of IPA+IAC system containing the adopted two ILs.Furthermore,the predictive UNIFAC-Lei model [32-34] was applied to predict the alcohol-ester binary VLE data for comparison.All the comparisons are expressed in Fig.S1 in Supplementary Material,which showed less deviations between the predicted and the measured values.This indicates that UNIFAC-Lei model can provide better predictive VLE results for the above system.
The influence of [EMIM][Ac]/[EMIM][Tos] with different molar content on the VLE behavior of (IPA+IAC) azeotropic system are expressed in Figs.6 and 7.As in the-y1diagrams,the remarkable salting-out effect on the (IPA+IAC) system was shown.With the increase of[EMIM][Ac]/[EMIM][Tos]mole content,the azeotropic point shifts to the left part,and the IAC content in the liquid phase decreases.When the amount of [EMIM][Ac]/[EMIM][Tos]reached to a certain concentration,the azeotropic phenomenon for (IPA+IAC) mixture will disappear.It is evident that [EMIM][Ac]with a molar content of 5%and[EMIM][Tos]with a molar content of 15% can completely break the azeotrope between IPA and IAC.
Fig.6.The -y1 diagram for IPA(1)+IAC(2)+[EMIM][Ac](3)ternary system at atmospheric pressure: ○,IL free;■, x1 ≈4%;●, x2 ≈5%;▲, x3 ≈8%;solid line,correlated by NRTL model.
Fig.7.The -y1 diagram for IPA(1)+IAC(2)+[EMIM][Tos](3)ternary system at atmospheric pressure: ○,IL free;■, x1 ≈5%;●, x2 ≈10%;▲, x3 ≈15%;solid line,correlated by NRTL model.
Fig.8.Relative volatility of IPA(1) to IAC (2) with different [EMIM][Ac] (3)fraction at atmospheric pressure: ○,IL free;■, x1 ≈4%;●, x2 ≈5%;▲, x3 ≈8%;solid line,correlated by NRTL model.
Fig.9.Relative volatility of IPA(1)to IAC(2)with different[EMIM][Tos](3)fraction at atmospheric pressure:○,IL free;■,x1 ≈5%;●,x2 ≈10%;▲,x3 ≈15%;solid line,correlated by NRTL model.
As demonstrated in Figs.8 and 9,the addition of [EMIM][Ac]/[EMIM][Tos] with different molar content exhibits a satisfactory effect on the relative volatility of the IPA+IAC azeotrope due to its greater affinity to IAC and less to IPA.The relative volatility of IPA to IAC decreases along with the increase of [EMIM][Ac]/[EMIM][Tos]caused by the salting-out effect.When the molar content of that reached about 5%and 15%respectively,the azeotropic point of the studied system was eliminated.In consequence,both of [BMIM][Ac] and [EMIM][Tos] were shown outstanding separation performance in full-concentration region for breaking the azeotropes.[BMIM][Ac] was little better than [EMIM][Tos].
Two sample and easy prepared ILs,[EMIM][Ac] and [EMIM][Tos],were selected from 30 ILs candidates to explore the separation of (IPA+IAC) azeotrope mixture according to the σ-profile analysis and the infinite dilution selectivity based on the COSMO-SAC method firstly.Then,the isobaric VLE experiments for IPA+IAC binary system and IPA+IAC+[EMIM][Ac]/[EMIM][Tos] ternary systems were performed at atmospheric pressure.The determined VLE results illustrated that the addition of IL entrainers exhibits an obvious salting-out effect on IAC.The relative volatility of IPA to IAC was enhanced and with molar content of [EMIM][Ac]/[EMIM][Tos] reaches to 5%/15%,respectively,the azeotropic point was completely eliminated.Additionally,the experimental VLE data of the ternary systems were correlated by applying the NRTL model,with which the calculated RMSD value were 1.84% and 1.65%,indicating the good agreement between determined data and regressed results.Meanwhile,its binary non-randomness and interaction parameters were acquired by the NRTL model regression,which can provide the basis for the separation of IPA+IAC azeotropes.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors are grateful to the support of the National Natural Science Foundation of China(21908131),the Natural Science Foundation of Shandong Province (ZR2019BB066),and A Project of Shandong Province Higher Educational Science and Technology Program (J18KA072),and Open Project of Qingdao University of Science and Technology Chemistry Department (QUSTHX202007).
Supplementary Material
Supplementary material to this article can be found online at https://doi.org/10.1016/j.cjche.2022.07.022.
Chinese Journal of Chemical Engineering2022年10期