Kewen Tang *,Jingjing Luo ,Panliang Zhang ,Jianmin YiJie Hua ,Chang'an Yang
1 Department of Chemistry and Chemical Engineering,Hunan Institute of Science and Technology,Yueyang 414006,China
2 College of Chemistry,Xiangtan University,Xiangtan 411105,China
Keywords:Kinetics Amino acids Modeling Liquid-liquid extraction Enantiomers
ABSTRACT Enantioselective liquid-liquid extraction has attracted considerable attention for its potential use in largescale production.Kinetic data are needed for the reliable scale-up of the process.This paper reports the kinetic study of reactive extraction of phenylalanine(Phe)enantiomers with BINAP-copper complex(BINAP-Cu)as a chiral selector.The theory of extraction accompanied by a chemical reaction was applied.The effects of agitation speed,interfacial area,pH value of aqueous phase,initial concentration of Phe enantiomers and initial concentration of BINAP-Cu on the specific rate of extraction were investigated.The forward rate constants of the reactions in the reactive extraction process are 7.93 × 10?5 m5/2·mol?1/2·s?1 for D-Phe and 1.29 × 10?4 m5/2·mol?1/2·s?1 for L-Phe.
Phenylalanine(Phe,Fig.1)is widely used in chemical,food and medicine industries,such as synthesis of antiviral and anticancer drugs and as a raw material of a new sweetener.Phe is one of the eight essential amino acids for the human body.It is found that Phe enantiomers exhibit a large difference in pharmacological activities and biological natures in the human body[1].D-phenylalanine(DPhe)has excellent analgesic activity and can enhance the immunity of the human body.L-phenylalanine(L-Phe)is one of the essential amino acids for the human body.Therefore,it is important to obtain enantiopure Phe[1].Many methods are available to obtain enantiopure substances,for example,from natural sources,asymmetric catalysis and chiral separation[2].However,the first two methods often involve relatively high cost,are time-consuming and offer a low yield[2].Chiral separation is an attractive technique because it is simple and needs less time[3].Several separation methods such as liquid membrane[4],chromatography[5]and crystallization[6]have been developed.Due to the limited transfer rate in membrane extraction,high costs of chromatography,and low versatility and excessive solid handling in crystallization[1,7,8],their applications are limited.For chiral separation of a wide range of substances,chiral solvent extraction is considered as a promising alternative,which is expected to be cheaper and easier to scale up to commercial scale and has wide applications.
Chiral extractant(selector)plays a key role in chiral solvent extraction process.The use of chiral metal complexes has attracted much attention[9-13].Complexes of copper with hydroxyproline derivatives have been used to transport amino acids through a liquid membrane enantioselectively[9,14].Chiral BINAP-palladium complexes introduced into enantioselective extraction by Verkuijl et al.present good selectivity towards a series of amino acid enantiomers[9,15].Previous work indicates that chiral metal complexes,especially the BINAP-metal complexes,provide a promising selector for separation of amino acid enantiomers[16-18].

Fig.1.Chemical structure of phenylalanine.
Separation of optically active compounds by chiral solvent extraction has been widely studied in recent years[9,15-32],but the work focuses primarily on the synthesis of new chiral selectors and extraction equilibrium,with limited information on kinetics of reactive extraction available[33].The kinetics is important for a deep understanding of the extraction process,for the selection and design of extraction equipment,and for a reliable scale-up of the process[34,35].An economically feasible reactive extraction process demands not only high enantioselectivity,but also sufficiently fast mass transfer and chemical reaction.Therefore,it is essential to investigate the kinetics of chiral solvent extraction.
This paper reports the kinetic study on enantioselective extraction of Phe enantiomers from aqueous phase to organic phase by CuPF6-(S)-BINAP(BINAP-Cu,Fig.2).Influences of some important process variables,such as agitation speed,interface area,pH value of aqueous phase,and initial concentrations of Phe enantiomers and BINAP-Cu,on the specific rate of extraction are reported.

Fig.2.Chemical structure of CuPF6-(S)-BINAP.
Tetrakis(acetonitrile)copper(I)hexa fluorophosphate([(CH3CN)4Cu]PF6,purity≥97%)was purchased from Hewei Chemical Co.Ltd.(Guangzhou, China).(S)-(?)-2,2′-bis(diphenylphosphino)-1,1′-binaphthalene(S-BINAP)(purity≥99%)was bought from Shengjia Chemical Co.Ltd.(Shijiazhuang,China).DL-phenylalanine(DL-Phe,racemate,purity≥99%)was purchased from Nantong Chemical Co.Ltd.(Jiangsu,China).Solvent for chromatography was of HPLC grade.All other chemicals were of analytical-reagent grade.
The kinetic experiments were carried out in a modified Lewis cell,with the structure the same as shown previously[36-38].The Lewis cell was a glass cylinder,7.5 cm height and 6.8 cm inside diameter,divided into two halves by an acrylic circular disk with a circular hole in the middle.The interface of the two phases can be changed by using disks with holes of different diameters.To maintain a uniform temperature,the cell was equipped with a thermo jacket,through which water of constant temperature(5°C)was circulated.
BINAP-Cu was synthesized by dissolving[(CH3CN)4Cu]PF6and(S)-BINAP in 1,2-dichloroethane atequal concentrations,with stirring overnight.The obtained solution was diluted to the desired concentration and used as the organic phase.Phe enantiomers were dissolved in a buffer solution of 0.1 mol·L?1NaH2PO4/Na2HPO4to prepare aqueous phase(pH=7.0).In an experiment,110 mlof organic phase was introduced into the cell,and then an equal volume of aqueous phase was placed carefully on the top of the organic phase without disturbing the interface.The stirrers in the organic and aqueous phases rotated at the same speed in opposite directions.Samples of 0.1 ml were taken from the aqueous phase at prescriptive time intervals.
The quantification of Phe enantiomers was performed by HPLC(Agilent LC 1200 series apparatus,supplied by Agilent Technologies Co.Ltd.USA)with an UV detector operated at 250 nm.The column was Inertsil ODS,3.5 μm particle size of the packing material,4.6 mm×250 mm I.D.(GL Sciences Inc.Japan).The mobile phase was a 20:80(volume ratio)mixture of methanol and an aqueous solution(pH=5.0,adjusted by ammonium acetate),composed of 1.5 mmol·L?1L-proline and 0.75 mmol·L?1CuSO4·5H2O.The flow rate was set at 1 ml·min?1.
The experimental result is expressed as the concentration of extracted component in the organic phase as a function of time.With the two film model[39],the concentration of D-Phe in the cell as a function of time can be expressed as

where CD,organd CD,aqrepresent the concentrations of D-Phe in organic and aqueous phases,respectively,t represents the extraction time,RDis the extraction rate of D-Phe,A is the interfacial contact area taken as the area of the disk,and Vaqand Vorgare the volumes of bulk aqueous phase and organic phase,respectively.
To avoid the problem of reversible reaction,a reversible reaction is often studied by measuring the initial extraction rate,which is controlled by the forward reaction.For determination of the initial extraction rate,it is desirable to use the complete experimental concentration-in-time profile.The initial extraction rate RD,0is calculated from experimental data by using the following equation:

where(d CD,org/d t)t=0is the initial slope of the curve representing concentration in the organic phase(CD,org)versus time(t).The values of RD,0are determined under various experimental conditions so as to assess the effects of variables and obtain appropriate kinetic model.RL,0can be obtained by defining Eqs.(1)and(2)for L-Phe in the same way.All the fittings have the square of the correlation coefficient higher than 0.99.Each experiment was duplicated under identical conditions and the standard deviation is in the range of 2%.
Two fundamental approaches are reported in literature on reactive extraction:interfacial reaction model and homogeneous reaction model.The same experimental result may lead to different interpretations with different models.Mass transfer and chemical reaction are a serial process in the interfacial reaction model:reactants transport to the interface and are adsorbed,the reaction takes place at the interface and products transport to the bulk phase.This mechanism is frequently applied for the reactive extraction system accompanied by a ligand-exchange reaction[33].
Fig.3 depicts the mechanism of reactive extraction of Phe enantiomers by BINAP-Cu.Experiments demonstrate that BINAP-Cu and its complex with Phe enantiomers are completely insoluble in the aqueous phase;Phe enantiomers(Dand L)and its anionic form are not dissolved in organic phase.Therefore,the reaction is in the aqueous or organic phase.Interfacial reaction modelis more suitable for the reactive extraction of Phe enantiomers by BINAP-Cu.

Fig.3.Mechanism of reactive extraction.
The reaction between BINAP-Cu and solute D-Phe can be described as

It is assumed that the rate equation can be defined according to the law of mass action as a(m,n)-reaction

where BINAP-Cu-D represents the complex of BINAP-Cu with D-Phe.
In the initial reaction,the concentration of BINAP-Cu-D in the organic phase is very low,so the reaction is far from equilibrium and the reverse reaction can be neglected.When the extraction process is controlled by interfacial reaction,the rate expression for the initial extraction can be written as

Eqs.(3)-(5)can be defined for L-Phe in the same way.
Fig.4 shows the D-Phe concentration in organic phase versus time at different agitation speeds.The equilibrium time decreases with the increase of agitation speed.The initial extraction rates at different agitation speeds is obtained from the slopes of the curves at t=0.Fig.5 shows the dependence of extraction rate on the agitation speed.The initial specific rate of extraction increases with the agitation speed in the range of 30 to 60 r·min?1,because the stagnant interfacial films are thick at these agitation speeds and the extraction rate is mainly controlled by diffusion.The extraction rate is independent of agitation speed above 60 r·min?1,indicating that the extraction is controlled by kinetics.However,as the speed is higher than 90 r·min?1,extraction rate increases with the stirring speed again.It was observed that the interface became slightly unstable and high agitating speed led to morefluctuations.A higher extraction rate may be due to an increase of contact area with disturbances.Therefore,agitation speed of 75 r·min?1is chosen for most of the experiments unless specified.In this plateau region,the film resistance is negligible and chemical reactions control the rate of mass transfer.

Fig.4.D-Phe concentration in organic phase versus time at different agitation speeds.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,pH 7.0,A=9.62 cm2,T=5 °C.

Fig.5.Effect of stirring speed on initial extraction rate.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,pH 7.0,A=9.62 cm2,T=5 °C.
In a kinetic controlled regime,dependence of extraction rate on the interfacial area is different with the interfacial reaction and homogeneous reaction models.In the former case,the mass transfer rate is linear to the interfacial area A.To differentiate the two models,the effect of interfacial area on the mass transfer is studied with interfacial areas in the range of 7.06-15.91 cm2and a constant agitation speed of 75 r·min?1.The rate of mass transfer is expressed by the product of R0and A.
Fig.6 shows the D-Phe concentration versus time at different interfacial areas.The equilibrium time decreases with interfacial area.Fig.7 shows that the mass transfer is linearly proportional to the interfacial area.Therefore,the reactive extraction of Phe enantiomers is controlled by interfacial chemical reactions.
Fig.8 shows the D-Phe concentration versus time at different pH values.The equilibrium concentration of D-Phe in organic phase is higher at higher pH,increasing more considerably at pH≤7 but slightly at pH≥7.

Fig.6.D-Phe concentration in organic phase versus time at different interfacial areas.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,pH 7.0,N=75 r·min?1,T=5 °C.

Fig.7.Effect of interfacial area of two phases on initial extraction rate.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,pH 7.0,N=75 r·min?1,T=5 °C.

Fig.8.D-Phe concentration in organic phase versus time at different pH values.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,A=9.62 cm2,N=75 r·min?1,T=5 °C.
Fig.9 shows the effect of pH on the initial extraction rate,which increases rapidly with pH value at pH≤7 and keeps nearly unchanged in the pH range from 8 to 10.This can be due to the shift of association-dissociation equilibrium of Phe.According to the species fraction plot(Fig.10),the amount of anionic Phe increases rapidly with pH value while the amount of molecular Phe decreases at pH values from 5 to 7.For pH higher than 8,most of Phe is present in anionic form,the amount of anionic Phe will not increase with pH.It is Phe anion that binds BINAP-Cu,through which Phe transports into the organic phase.Therefore,the initial extraction rate first increases rapidly and then more slowly with the increase of pH.

Fig.9.Effect of pH value of aqueous phase on initial extraction rate.[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,A=9.62 cm2,N=75 r·min?1,T=5 °C.

Fig.10.Species fraction as a function of pH.
Keeping the concentration of Phe enantiomers constant at 2 mmol·L?1and varying the BINAP-Cu concentration from 0.25 mmol·L?1to 2 mmol·L?1,the effect of BINAP-Cu concentration on initial extraction rate is investigated.Fig.11 shows the DPhe concentration versus time at different concentrations of BINAP-Cu in the organic phase.The equilibrium concentration of D-Phe in organic phase is higher at a higher concentration of BINAP-Cu.

Fig.11.D-Phe concentration in organic phase versus time at different initial concentrations of BINAP-Cu in organic phase.[Phe]0=2 mmol·L?1,pH 7.0,A=9.62 cm2,N=75 r·min?1,T=5 °C.
Fig.12 shows that the relation of lg R0with lg[Cu-BINAP]is linear for each enantiomer.The slopes of the fitting lines for D-Phe and L-Phe are 0.51 and 0.48,respectively.The reaction with respect to BINAP-Cu is approximately a one-half order reaction,namely n=0.5.

Fig.12.Effect of[BINAP-Cu]0 on[R0].[Phe]0=2 mmol·L?1,A=9.62 cm2,N=75 r·min?1,pH 7.0,T=5 °C.
Fig.13 shows the D-Phe concentration versus time at different initial concentrations of D-Phe enantiomer in the aqueous phase.The equilibrium concentration of D-Phe in organic phase is higher at higher initial concentration of D-Phe.
Fig.14 shows that the relations of lg R0with lg[D-Phe]and lg[L-Phe]are linear.The reaction order with respect to D-and L-Phe is 1.12 and 0.97,respectively.The reaction with respect to Phe enantiomer is approximately a one order reaction,namely,m=1.
Take the logarithm on both sides of Eq.(5),the equation can be expressed as


Fig.13.D-Phe concentration in organic phase versus time at different initial concentrations of Phe in aqueous phase.[BINAP-Cu]0=1 mmol·L?1,pH 7.0,A=9.62 cm2,N=75 r·min?1,T=5 °C.

Fig.14.Effect of[Phe]0 on[R0].[BINAP-Cu]0=1 mmol·L?1,A=9.62 cm2,pH 7.0,N=75 r·min?1,T=5 °C.

where RD,0is the initial extraction rate of D-Phe,[D]and[BINAP-Cu]are the initial concentrations of D-Phe and BINAP-Cu,respectively.For m=1 and n=0.5,the rate constant of km,n,Dand km,n,Lcan be obtain from the intercept of Fig.14 as 7.93× 10?5and 1.29×10?4m5/2·mol?1/2·s?1,respectively.
From the forward rate constants from this work and equilibrium constants determined in our previous work[16],the backward reaction rate constants of k?m,n,Dand k?m,n,Lare estimated as 4.81 × 10?4and 2.80 × 10?4m4·mol?1·s?1,respectively.
For an reversible(1,0.5)reactive extraction,the following two rate equations for D-and L-Phe can be deduced according to Eq.(4)


A simulated extraction for Phe enantiomers can be established to describe the flux at the interface.Fig.15 shows that the predictions are in good agreement with the experimental results.Thus the model is suitable to describe the extraction course.

Fig.15.Concentrations of D-and L-Phe in aqueous phase versus time.Lines:model predictions;[Phe]0=2 mmol·L?1,[BINAP-Cu]0=1 mmol·L?1,pH 7.0,A=9.62 cm2,N=75 r·min?1,T=5 °C.
In the kinetic study for liquid-liquid reactive extraction of phenylalanine enantiomers by BINAP-Cu in a modified Lewis cell,with the theory of extraction accompanied by chemical reaction,it is found that the reactions are first order on Phe enantiomers and one-halforder with respect to BINAP-Cu.The model predictions are in good agreement with the experimental results.The results obtained in this paper will be useful for the design and operation of reactive extraction equipment.
Nomenclature
A interfacial area,cm2
BINAP-Cu complex of S-BINAP with copper
[BINAP-Cu] concentration of complex of BINAP-Cu
[BINAP-Cu-D]concentration of complex of BINAP-Cu with D-Phe
[BINAP-Cu-L]concentration of complex of BINAP-Cu with L-Phe
CL,aqconcentration of L-Phe in aqueous phase at time t
CL,orgconcentration of L-Phe in organic phase at time t
km,nrate constant
k?m,nbackward reaction rate constant
[L]aqconcentration of L-Phe in aqueous phase,mol·L?1
[L]orgconcentration of L-Phe in organic phase,mol·L?1
N stirring speed,r·min?1
RDextraction rate of D-Phe,mol·m?2·s?1
RLextraction rate of L-Phe,mol·m?2·s?1
R0initial extraction rate of Phe,mol·m?2·s?1
T temperature,°C
t time
V volume of the bulk phase,cm3
Subscripts
aq aqueous phase
D D-enantiomer
L L-enantiomer
m order of reaction with respect to Phe
n order of reaction with respect to BINAP-copper complexes
org organic phase
0 initial value
Chinese Journal of Chemical Engineering2015年1期