Kangning Xiong ,Shuai Shen ,Youchang Wang ,Yun Chen ,2,*

1 State Key Laboratory of Pulp and Paper Engineering,Department of Chemistry and Chemical Engineering,South China University of Technology,Guangzhou 510640,China

2 Department of Chemical&Biomolecular Engineering,North Carolina State University,Raleigh,NC27695,USA

Keywords:Liquid-liquid equilibrium Methyl tert-butyl ketone Benzenediol NRTLmodel UNIQUACmodel

ABSTRACT In this work,liquid-liquid equilibria(LLE)data for the ternary system methyl tert-butyl ketone(MTBK)+o,m,p-benzenediol+water were investigated at 333.2 K,343.2 K and 353.2 K under 101.3 k Pa.The performance of MTBK to extract o,m,p-benzenediol from wastewaters w as estimated by partition coef fi cients and separation factors.The Hand and Bachman equations w ere both applied to check the reliability of the experimental LLE data.Furthermore,the Non-Random Tw o-Liquid(NRTL)and Universal Quasi Chemical(UNIQUAC)models w ere applied to correlate the measured LLEdata.The results showed a good agreement with the determined ternary LLEdata w ith the root-mean-square error(RMSE)values below 0.71%.MTBKw as proved to be a promising extracting agent in extracting benzenediols from ef fl uents.

1.Introduction

Coal is one of the most signi fi cant energy sources in China,w hich has been used w idely in a variety of chemical industrial applications:e.g.,coal gasi fi cation,coking coal,and coal liquefaction[1,2].However,a large number of ef fl uents w ith high concentration phenol exist in these industrial processes,including cresols,phenol and benzenediols,w hich w ould bring about severe harm to human beings and the environment even though its concentration is very low[3].Moreover,phenolic compounds as vital chemical raw materials are w idely used in variousindustries,for example(i)phenol can beused to manufacture chemical herbicides,synthetic,colorant,drugs,explosives,additives,etc.;(ii)cresol can be applied to produce plastics,pesticides,spices,dyes,etc.;(iii)benzenediols can be used to make accelerator,dyestuff,synthetic fi ber,chemical intermediate,etc.Thus,removing phenolic compounds from w astew ater has become a crucial topic.The phenolic w astewater treatment process is mainly to remove the phenols in the ef fl uents in order to make the w astew ater meet the discharge standard[4].Many techniques,such as adsorption[5,6],membrane separation[7],burning,chemical oxidation[8],biochemistry[9]and solvent extraction[10,11],havebeen implemented to deal w ith phenolic w astew ater[12-19].Compared to other technologies,liquid-liquid extraction(or solvent extraction)has show n many advantages,such as high commercial ef fi ciency,simple operation and low energy consumption,w hen applied to treat phenolic w astew ater[20].

Selecting a suitable solvent has become a key step in the liquid-liquid extraction processbecause excellent solventscan enhance the separation ef fi ciency.Previousresearcheson treatment of phenolic w astewater w ith liquid-liquid extraction have studied some various types of solvents,including methyl tert-butyl ether[21,22],butyl alcohol[23],methyl acetate[24]and tributyl phosphate.Short chain ketones,such asmethylbutylketone(MBK)[25],methyl isobutyl ketone(MIBK)[26]and methyl tertbutyl ketone(MTBK)[27]are a suitable class of extractants because they have some advantages comprising higher distribution coef fi cients for polyphenols and monophenols,moderate boiling point,azeotropic point with water,smaller solubility in water and so on.Many researches on liquid-liquid equilibrium(LLE)including the ternary systems of various solvents+phenol,cresol+w ater have been done previously[17,28-31],but the related research of benzenediols is relatively less.

Benzenediols have three isomers including o,m,p-benzenediol in which two hydroxylgroups are replaced onto a benzene ring.When the concentration of benzenediolsishigh,such asup to several hundred ppm in the actual process of industry,the subsequent biochemical process is dif fi cult to deal w ith them.Therefore,it is important to decrease the benzenediols as low as possible through extraction unit.The ketones have higher distribution coef fi cients than ethers.Chen et al.[32]have studied the methyl isopropyl ketone as the extraction agent to remove(resorcinol or hydroquinone)from w astew aters;Yang and Jiang[33]and Wang et al.[34]reported the extraction of(catechol,resorcinoland hydroquinone)with methylisobutyl ketoneat lower and higher temperatures,respectively.How ever,due to extraction temperature was less than 333 K,amounts of binders containing fl y ash and coal tar from wastew ater w ould stick and deposit into heat exchange[12].So MTBK is selected as an appropriate extraction solvent,and the LLE data of MTBK+o,m,p-benzenediol+water not been report in literature were determined at the higher temperature of 333.15 K,343.15 K and 353.15 Kin this work.The coherence of the experimental LLEdata were ascertained by the Hand[35]and Bachman equations[36],and the tieline data were interrelated w ith activity coef fi cients models[37,38].

2.Experimental

2.1.Chemicals

In this w ork,Table 1 shows the detailed information of the experimental chemicals.The distilled water in these experiments was prepared from ultra-pure w ater equipment in our laboratory.The purities of chemicals w ere checked by gas chromatography.All chemicals were applied without further puri fi ed due to their purities over 98%.

Table 1 Suppliers and purity of chemicals used in this work

2.2.Experimental procedure and analysis

A glass equilibrium kettle(approximately 100 ml)w as applied to measure experimental LLEdata,as show n in Fig.1.The temperature of the kettle w as maintained by a thermostatic bath(HWS-24,Shanghai Heng Technology Co.,Ltd.)w ith a fl uctuation of±0.1 K.All chemicals w ere w eighed by a chemical balance(AUW220D,Shimadu)correct to±0.0001 g.A certain amount of MTBK,benzenediols and w ater w ere added into the kettle,stirred rigorously for at least 2 h and then stood for a minimum 20 h to make sure a thorough clari fi cation of the equilibria phase.After reaching the equilibria state betw een the tw o liquid phases,samples of the tw o phases w ere collected by different syringes and checked by using an Agilent GC-6820 gaschromatography(Agilent Technologies,USA)equipped w ith a fl ame ionization detector(FID).The capillary column w as DB-5MS(Hew lett Packard-Agilent,5%phenyl-methyl polysiloxane,30 m×0.32 mm×0.50μm).The mass fractions of MTBK and o,m,p-benzenediol w ere investigated by using internal standard method.The mass fraction of water in the aqueous phase w as obtained by subtracting those of benzenediols and MTBK from 1.Moreover,thew ater content in theorganic phasew asmeasured w ith Karl-Fischer Volumetric Moisture Titrator(New 870 Titrino plus,Metrohm).In this w ork,n-butyl acetate w as as the internal standard for MTBK and n-octanol for benzenediols.Methanol w as decided as the solvent for GCanalysis.The temperature of the injection and the detector were set at 523.2 K and 543.2 K separately.The column temperature w as programmed as follow s:kept up 313.2 K for 2 min,and then subsequently increased 463.2 K at the speed of 30 K·min-1.The carrier gas w as nitrogen w ith a speed of 30 ml·min-1around the column.Each sample wasrepeated at least three timeswith a standard deviation of smaller than 0.2%,thusthemean value w asreported in this w ork.

Fig.1.Equilibria kettle:1.Sample outlet of organic phase;2.Sample outlet of aqueous phase;3.Organic phase;4.Aqueous phase.

3.Results and Discussion

3.1.Experimental LLEdata

The investigated LLE data of the ternary system MTBK+o,m,p-benzenediol+w ater at 333.2,343.2 and 353.2 K under 101.3 kPa are show n in Tables 2-4.All concentrations in these tables are expressed in mass fraction.The tie-lines correlated data by NRTLand UNIQUACmodels for the corresponding triangle phase diagrams were plotted in Figs.2-4.As demonstrated in Figs.2-4,the slopes of tie-lines in this w ork are positive,w hich indicate that benzenediols are more soluble in the organic phase than in the aqueous phase.

In order to estimate the capacity of MTBK to extract o,m,pbenzenediol from w astew aters,the partition coef fi cients(P)and separation factors(S)w ere calculated in Tables 2-4 by the follow ing formulas:

in which w2Iand w2IIare the mass fractions of benzenediols in the organic phase and the aqueous phase,respectively;the mass fractions of w ater in above tw o phases w ere separately denoted by w3Iand w3II.The partition coef fi cients(P)and separation factors(S)for benzenediolsat each temperature w ere presented in Tables 2-4 and Figs.5,6.Show n in these tables,valuesof the partition coef fi cients(≥6.3)and separation factors(≥144)are greater than unity.It indicates that the extraction of benzenediols by MTBK is quite feasible.The in fl uence of thetemperature on the partition coef fi cients(P)and the separation factors(S)values show n in Figs.5,6 are found that the P and S obviously decrease w ith the temperature increasing.These advantages indicate that MTBK has a pow erful capability to extract benzenediols from ef fl uents.

Table 2 Experimental LLEdata(massfraction),the partition coef fi cients(P)and separation factors(S)for the ternary system,MTBK(1)+o-benzenediol(2)+w ater(3)at 333.15 K,343.15 Kand 353.15 Kunder 101.3 kPa

Table 3 Experimental LLEdata(massfraction),the partition coef fi cients(P)and separation factors(S)for the ternary system,MTBK(1)+m-benzenediol(2)+w ater(3)at 333.15 K,343.15 Kand 353.15 Kunder 101.3 kPa

Table 4 Experiment LLEdata(massfraction),valuesof the partition coef fi cients(P)and separation factors(S)for the ternary system,MTBK(1)+p-benzenediol(2)+w ater(3)at 333.15 K,343.15 Kand 353.15 Kunder 101.3 k Pa

Fig.2.Ternary diagram for experimental LLE data,correlation data of MTBK(1)+o-benzenediol(2)+water(3)at(a)333.15 K,(b)343.15 K and(c)353.15 K:▲-▲,experimental data;□,the NRTLmodel;×,the UNIQUACmodel.

Fig.3.Ternary diagram for experimental LLE data,correlation data of MTBK(1)+m-benzenediol(2)+w ater(3)at(a)333.15 K,(b)343.15 K and(c)353.15 K:▲-▲,experimental data;□,the NRTLmodel;×,the UNIQUACmodel.

3.2.Reliability of LLEdata

It is important of phase equilibrium data to design and optimize the industrial extraction process.Up to present,the reliability of the experimentally obtained LLEdata w as evaluated by the Hand and Bachman equations as follows:

where a1,b1and a2,b2are fi tting parameters of the Hand and Bachman equations.w3IIand w1Iare the mass fractions of w ater in the aqueous phase and MTBK in the organic phase,w hile w2Iand w2IIare mass fractions of benzenediols in the organic phase and in the aqueous phase.The parameters of these correlations w ere show n in Table 5.All R2values are close to 1,and it indicates that the experimental LLEdata have a good consistency.

Fig.4.Ternary diagram for experimental LLEdata,correlation data of MTBK(1)+p-benzenediol(2)+w ater(3)at(a)333.15 K,(b)343.15 K and(c)353.15 K:▲-▲,experimental data;□,the NRTLmodel;×,the UNIQUACmodel.

3.3.Data correlation

The NRTL and UNIQUAC models w ere applied to correlate the experimentally tie-line data by using Aspen Plus v8.4.The adjustable parameters of the NRTLmodel are expressed as:

w here i,j,and k refer to each component in the measured system.αijis the non-randomness parameter,w hichαijw as fi xed to 0.2 or 0.3;the parameters bijand bjiare the binary interaction parameters,w hich need to be regressed from experimental tie-line data;τij(≠τji)is the interaction parameter.

Fig.5.Partition coef fi cients(P)for ternary systems of MTBK(1)+benzenediol(2)+w ater(3)at(○)333.15 K,(▲)343.15 K and(□)353.15 K:a,o-benzenediol;b,m-benzenediol;and c,p-benzenediol.

The equation for the NUIQUACmodel is described as:

Fig.6.Separation factors(S)for ternary systemsof MTBK(1)+benzenediol(2)+water(3)at(○)333.15 K,(▲)343.15 K and(□)353.15 K:a,o-benzenediol;b,m-benzenediol;and c,p-benzenediol.

Table 5 Parameters of the Hand and Bachman equations of MTBK(1)+o,m,p-benzenediol(2)+w ater(3)at 333.15 K,343.15 K and 353.15 K

Hence z is the coordination number and is set to 10.φiandθiare the segment fraction and the area fraction of component i,respectively.riand qiare the pure component structural parameters;the parameters aijand ajiare the binary interaction parameters;τij(≠τji)is the interaction parameter.

Thebinary interaction parametersof the NRTLand UNIQUACmodels were calculated at each temperature by minimizing the function(OF):

Table 6 Binary parameters of NRTLand UNIQUACmodelsfor MTBK(1)+o,m,p-benzenediol(2)+water(3)at 333.15 K,343.15 Kand 353.15 K

Accordingly,n is the total of the tie-lines;weijxkpand wcijaklrefer to the experimental and calculated mass fraction.Subscripts i,j,and k represent the components,the phases and the tie-lines,respectively.Table 6 show s the binary parameters from Aspen Plus simulation.The root-mean-square-error(RMSE)values show n in Eq.(12)w ere usually employed to assess the agreement betw een calculated and experimental data.

The RMSEvalues of the tw o models are listed in Table 6.Seen from Table 6,all RMSEvalues are less than 0.01,w hich indicated that tw o models have good agreement of the experimental LLEdata.Moreover,the comparison results of the experimental and calculated data show n in Figs.2-4 further con fi rm the above conclusion.

4.Conclusions

The LLEdata for ternary system MTBK+o,m,p-benzenediol+water w ere investigated at 333.15 K,343.15 Kand 353.15 Kand 101.3 kPa.The partition coef fi cients(P)and separation factors(S)valuesare very high.It indicates that MTBK is a suitable solvent for extraction of benzenediols from ef fl uents.The reliability of the tie-line data of the ternary systems w aschecked by using the Hand and Bachman equations,and high regression coef fi cients w ere accomplished.The NRTLand UNIQUACmodels w ere applied to correlate the LLEdata for the ternary systems MTBK+o,m,p-benzenediol+water,w ith RMSEvalues less than 0.71%.The calculated resultsby two modelsshowed a good agreement w ith the experimental tie-line data.Additionally,the obtained binary interaction parameterscan beapplied on thedesigning and optimization of industrial extraction process.