【作者】 黄笃树；

【导师】 黄可龙；

【作者基本信息】 中南大学 ， 生物化工， 2008， 博士

【摘要】 药物对映体的液膜手性分离及其传质模型的研究为当前最热门的课题之一。本文阐述了手性药物的基本概念,综述了目前主要的手性拆分方法及其研究进展,对液膜手性拆分过程的传质机理及模型研究现状和意义进行了阐述。对乳状液膜和中空纤维支撑液膜手性分离苯丙氨酸对映体的传质机理及模型进行了比较详细的研究与探讨,主要研究内容如下:研究了以N-癸基-L-羟基脯氨酸和铜离子配合物为手性载体,Span-80为表面活性剂,煤油为溶剂的乳状液膜体系手性分离苯丙氨酸对映体的分离性能。考察了各因素等对选择性分离性能的影响,从而确定了合适的萃取条件为:表面活性剂Span-80和有机溶剂煤油体积比8:92;苯丙氨酸外消旋体浓度为1.0 mmol/L;手性萃取途径是从外水相到内水相;外水相pH 4.8;Cu²⁺浓度5 mmol/L。通过对乳状液膜手性分离过程中传质机理的研究以及各影响因素的分析,得到了可以分析和预测乳状液膜手性拆分过程传质的反应-扩散模型,该模型同时考虑了外水相边界层的扩散阻力和手性对映体与载体的反应阻力。外水相苯丙氨酸对映体的浓度以及分离因子α表示为:当外相对映体浓度远小于液膜相手性载体浓度时,该手性传质过程中外相边界层的扩散为控制步骤,利用数据拟合得到了外水相与液膜相界面边界层扩散的传质系数k_e。将常数表达式中引入表观分配系数的概念,可以得到改进的反应-扩散模型,改进后模型对外水相D-苯丙氨酸对映体的浓度分率的模拟值与实验结果的最大偏差大大减小,由20.4%降为10.7%。模型中的常数表达式为:研究了在中空纤维液膜器中采用N-癸基-L-羟基脯氨酸和铜离子配合物为手性载体的支撑液膜体系手性拆分苯丙氨酸对映体。通过对比不同的流体流速对手性分离过程的影响,选择最佳的流体流速为3mm/s。研究了对映体浓度和分离因子与缓冲溶液pH的变化关系,选择缓冲溶液的pH为4.8。从总传质阻力的角度出发,通过对中空纤维支撑液膜过程中各影响因素的分析及理论推导,得到了接收相中各手性对映体浓度以及手性药物拆分过程分离因子对时间的总传质阻力模型。该数学模型同时考虑了手性对映体在料液相、接收相与液膜相间的分配行为和料液相边界层、接收相边界层以及支撑液膜相的扩散阻力。接收相中苯丙氨酸对映体的浓度及分离因子α可表示为:利用实验结果,拟合得到了模型中的未知参数,为以后进一步的研究及实验装置的放大提供了参考。进一步研究了中空纤维支撑液膜手性分离过程的传质机理,得到了描述接收相中各手性对映体浓度、液膜相中各手性对映体浓度以及手性药物拆分过程分离因子对时间变化的偏微分方程组。通过对某些条件的假设与简化处理,得到了偏微分方程组的解析解,即反应一扩散模型。该数学模型同时考虑了手性对映体与载体的反应和料液相边界层、接收相边界层以及支撑液膜相的扩散阻力。接收相和中空纤维支撑液膜相中苯丙氨酸对映体的浓度以及手性分离因子可表示为:在描述手性对映体与载体反应的各因素中,药物对映体与载体反应的正向速率常数,反向速率常数对浓度和分离因子的影响很小,故可以考虑将其忽略,从而得到该过程的快速反应--扩散模型。模型中的常数λ_j表达式化为:将反应-扩散模型中手性对映体与载体的反应行为用手性对映体在料液相、接收相与液膜相间的表观分配行为取代,同样可以得到描述接收相中各手性对映体浓度、液膜相中各手性对映体浓度以及手性药物拆分过程分离因子对时间变化的偏微分方程组,其解析解为分配-扩散模型。该数学模型同时考虑了手性对映体的表观分配行为和料液相边界层、接收相边界层以及支撑液膜相的扩散阻力。接收相和支撑液膜中苯丙氨酸对映体的浓度以及手性分离因子可表示为:将所得到的传质模型对手性拆分过程中苯丙氨酸对映体的浓度和分离因子进行预测。结果表明:模型的模拟结果与实验数据能够很好地吻合;总传质阻力模型、反应-扩散模型与快速反应-扩散模型以及分配-扩散模型的模拟结果基本一致。对料液相与接收相的pH值,边界层传质阻力以及液膜传质阻力等因素对手性拆分过程中对映体的浓度和分离因子的影响进行分析。结果表明这些模型能够应用于苯丙氨酸对映体的中空纤维支撑液膜手性拆分过程中各操作条件的优化。更多还原

【Abstract】 Research on the enantioseparation of drug enantiomers using liquid membrane and its mass transfer model has a great interesting potential. The basic concepts about chiral drug were expatiated, the importance of the determination of chiral drug was analyzed, and the present main methods of chiral separation and their research progress, the importance of the principle and mathematical model of mass transfer process were reviewed in this paper. The principle and mathematical model of mass transfer process using the chiral emulsion liquid membrane and the chiral supported liquid membrane were investigated and discussed in detail, the main content and results can be summarized as following:Chiral extraction of phenylalanine enantiomers in the emulsion liquid membranes system using copper (Ⅱ) N-decyl-L-hydroxyproline as chiral selector, Span-80 as surfactant and kerosene as organic solvent was studied. The effects of initial phenylalanine concentration, the direction of chiral extraction, the volume ratio of organic solvent and surfactant, the pH gradient from external to internal phase and the pH of external aqueous phase, on performances of selective extraction, were discussed, respectively. Consequently, appropriate extractive conditions were established: 8:92 (v/v) Span-80: kerosene; Initial concentration of phynelalanine is 1.0 mmol/L; External phase pH 4.8; Copper concentration 5 mmol/L; Chiral extraction from external phase to internal phase.The reaction-diffusion model was developed to analyze the concentration of enantiomers and the separation factor of the enantioseparation process of chiral emulsion liquid membrane by studying the principle and some factors of mass transfer process. The mass transfer resistance of boundary layer in outer aqueous phase and the interfacial chemical reactions at the liquid membrane interfaces were taken into account in the model equations. The concentration of D-phenylalanine in the external phase and the separation factorαcan be expressed as: The diffusion of boundary layer in outer aqueous phase is the controlled process while the concentration of D-phenylalanine in the external phase far less than the concentration of chiral carrier in the emulsion liquid membrane phase, then the diffusion coefficient k_e of boundary layer in outer aqueous phase can be achieved by data fitting.The improved reaction-diffusion model was developed by substituting the partition coefficient and the reaction equilibrium constant for the observed partition coefficient in the expression ofζ_j. The deviation of the computational results from the experimental data decreased from 20.4% to 10.7%. The expression ofζ_j can be deduced as:Resolution of racemicα-cyclohexyl-mandelic acid containing copper(Ⅱ) N-dodeecyl-(L)-hydroxyproline (CuN2) as a chiral carrier across hollow fiber supported liquid membrane was carried out successfully. The effects of velocity of feed phase and stripping phase and the pH of external and internal aqueous phase, on performances of selective extraction, were discussed, respectively. Consequently, appropriate extractive conditions which the velocity of feed phase and stripping phase is 3 mm/s and the pH of external and internal aqueous phase is 4.8 were established.The overall mass transfer model was developed to analyze the concentration of enantiomers in the stripping phase and the separation factor of the enantioseparation process. The observed partition coefficient between the feed phase and the membrane phase, the stripping phase and the membrane phase, mass transfer resistance of boundary layer in strip phase inside the hollow fibers, boundary layer in feed phase and mass transfer resistance of the membrane phase were taken into account in the model equations. The concentration of enantiomers in the stripping phase and the separation factorαcan be expressed as:Using the experimental results of the enantiomers concentration, several parameters of the proposed model had been achieved by nonlinear fitting method.The mechanism of mass transfer process across hollow fiber supported liquid membranes containing chiral carrier was investigated more detailedly. Partial differential equations describing the concentrations of enantiomers in the stripping phase, concentrations of enantiomers in the membrane phase and the separation factor of the enantioseparation process were deduced. The reaction-diffusion model has been developed according to some hypothesis and predigestion. The mass transfer resistance of boundary layer in the strip phase inside the hollow fiber and boundary layer in the feed phase, the diffusion in the membrane phase and the interfacial chemical reactions at the liquid membrane interfaces are taken into account comprehensively. The concentration of enantiomers in the stripping phase and the hollow fiber supported liquid membrane and the enantioseparation factor can be expressed as:The forward reaction rate and the backward reaction rate could be ignored because they have little influence on the concentrations of phenylalanine enantiomers and the separation factor of the enantioseparation process by the prediction of the reaction-diffusion model, so the rapid reaction-diffusion model of the enantioseparation process can be deduced. The expression ofλ_j can be deduced as:Substituting the reaction behavior between the enantiomers and the chiral carrier of the reaction-diffusion model for the observed partition behavior between the feed phase, the stripping phase and the liquid membrane phase, partial differential equations describing the concentrations of enantiomes in the stripping phase and the membrane phase were also deduced. The partition-diffusion model has been developed according to some hypothesis and predigestion. The observed partition coefficient between the feed phase and the membrane phase, the stripping phase and the membrane phase, mass transfer resistance of boundary layer in strip phase inside the hollow fibers, boundary layer in feed phase and the diffusion in the membrane phase are taken into account in the model equations. The concentration of enantiomers in the stripping phase and the hollow fiber supported liquid membrane and the enantioseparation factor can be expressed as: The mathematical model was used to predict the enantiomers concentration of (D/L)-phenylalanine and the separation factor of the enantioseparation process. The results indicate that the computational results of mathematical models were in good agreement with the experimental data and that the computational results of the overall mass transfer model, the reaction-diffusion model and the partition-diffusion model are almost the same. The effects of some factors such as pH of the feed phase and the stripping phase, mass transfer resistance of boundary layer and mass transfer resistance of liquid membrane on the enantiomers concentration of (D/L)-phenylalanine and the separation factor of the enantioseparation process were analyzed by those models. The results show that those mathematical models can be easily used to predict the concentration of (D/L)-phenylalanine and the separation factor of the hollow fiber liquid membrane resolution process.更多还原