Separation of Bovine Serum Albumin Using Chromatographical Column: Parameters and Simulation

Transcription

1 Iraqi Journal of Chemical and Petroleum Engineering Iraqi Journal of Chemical and Petroleum Engineering Vol.11 No.1 (March 010) 1-9 ISSN: University of Baghdad College of Engineering Searation of Bovine Serum Albumin Using Chromatograhical Column: Parameters and Simulation Cecilia Khoshaba Haweel *, Ahmed J. Ali and Samar N. Alias * Chemical Engineering Deartment - College of Engineering - University of Baghdad Iraq Abstract A liquid-solid chromatograhy of Bovine Serum Albumin (BSA) on (diethylaminoethyl-cellulose) DEAE-cellulose adsorbent is worked exerimentally, to study the effect of changing the influent concentration of (0.15, 0.5, 0.5, and 1 mg/ml) at constant volumetric flow rate Q=1ml/min. And the effect of changing the volumetric flow rate (1, 3, 5, and 10 ml/min) at constant influent concentration of Co=0.15mg/ml. By using a glass column of (1.5cm) I.D and (50cm) length, acked with adsorbent of DEAE-cellulose of height (7cm). The influent is introduced in to the column using eristaltic um and the effluent concentration is investigated using UV-sectrohotometer at 30oC and 80nm wavelength. A sread (steeer) break-through curve is gained at lean feed concentration of 0.15mg/ml, while the flow rate greater than (3ml/min) is almost the same. So it is butter to work at low volumetric flow rate between (1-3) ml/min. The equilibrium-disersion model of liquid-solid chromatograhy for a nary mixture, related with Langmuir isotherm correlation is used in the modeling of this work. The resulting model is solved numerically by using MATLAB V.6.5 rogram. Introduction Chromatograhy is a very secial searation rocess for a multitude of reasons! First of all, it can searate comlex mixtures with great recision. Even very similar comonents, such as roteins that may only vary by a single amino acid. Second, chromatograhy can be used to searate delicate roducts since the conditions under which it is erformed are not tyically serving (Hubble, 004). A mixture of various comonents enters a chromatograhy rocess and is allowed to come into contact with two hases, the stationary hase is contained in a column where the mole hase moves in a controlled manner relative to the stationary hase, carrying with it any material that may refer to mix with it. To determine the differences in the concentration of mole hase (solute), a detector is used which can give a curve of the searated substance. So, the name chromatograhy stems from this, where each searated substance has different color on the stationary hase and gives different curve from the detector (Stock, 1974, Encycloedia, 199 and Hubble, 004). Frontal chromatograhy analysis has widely used for large scale searations of industry. It is made in which conditions are chosen that the desired comonent very quickly saturates the column and emerges from the column shortly after the feed is saturated. Imurities, on the other hand, are strongly retained by the adsorbent and feeding is continued until the front of the adsorbed imurities aroaches the column outlet (Simson, 001). Ion-exchange technique of chromatograhy is used to reresent the frontal analysis. That was because of the simlicity of this technique where searation take lace according to charge of both solid adsorbent and the searated molecules. There can be two tyes of functional grous of exchangers, covalently attached to the suort beads. These are called anion exchangers (resin with ositive functional grous) or cation 1

2 Searation of Bovine Serum Albumin Using Chromatograhical Column: Parameters and Simulation exchangers (resin with negative functional grous) (Raly and Walker, 1998, and Freifelder, 1998). Reresenting the frontal analysis with Ion-exchange chromatograhy by a mathematical model is an imortant feature not only for exerimental data but, also for the rocess design and otimization. Exeriments need to be carried out to get the mass-transfer arameters for modeling and simulation and comarisons with exerimental results indicated that the rediction was accetable. Exerimental Work A system of Bovine Serum Albumin (BSA) rotein adsorbed onto DEAE-cellulose was chosen to study the frontal analysis and comare it with the theoretical work (Alias, 007). Bovine Serum Albumin rotein of molecular weight from Sigma Chemical Comany dissolved in (5mg/ml) Tris-HCl buffer of 98% urity mead by BBH Comany to roduce the mole hase of the chromatograhy system at (H 7). DEAE-cellulose anion-exchanger from Sigma Chemical Comany was used as stationary hase of the chromatograhy system. Five samles of different weights of DEAE-cellulose was reared to evaluate the equilibrium isotherm, each one of them is ut in a beaker with (30ml) of BSA solution of concentration (0.1mg/ml) and H of 7. The samles were shaked at (30oC) for about (6 hours) then the concentrations were calculated using UVsectrohotometer. The amount of solute adsorbed into DEAE-cellulose q in (mg/g DEAE-cellulose) was found according to the equilibrium relation: BSA-solution was reared at the desired concentration for each run using Tris-HCl buffer. The solution was introduced to the to of the column using eristaltic um at a certain flow rate. Samles were collected directly from the bottom of the column using tubes for every (1ml) and then the concentration was measured using UV-sectrohotometer at 80 nm wavelength as in (fig. 1). The resulting breakthrough curves were determined by lotting effluent concentration (C/Co) against dimensionless time τ. (Alias, 007). Fig.1 the exerimental aaratus of BSA adsorbed in DEAE-cellulose bed. q ( Co C) V W A L (1) Modeling and Simulation The adsortion isotherm was obtained by lotting q vs. C A glass beaker of (50ml) caacity was reared for batch exeriment. A (50ml) of BSA-solution was added to the beaker containing (3.37mg) of DEAE-cellulose. At zero time, DEAE-cellulose of certain weight was added. Samles were taken every 10 minutes during the exeriment and concentration is measured using UVsectrohotometer. Frontal chromatograhy exeriments were carried out to find the breakthrough curve of BSA adsorbed to DEAEcellulose bed. To study the effect of changing the initial concentration (0.15, 0.5, 0.5, and 1 mg/ml) at constant volumetric flow rate of (1ml/min) and changing the volumetric flow rate (1, 3, 5, and 10 ml/min) at constant initial concentration of (0.15mg/ml). The rocedures of all run exeriments are as follows: DEAE-cellulose was oured into the column at (7cm) length. Batch Adsorber Batch adsorber analyses are imortant to calculate the external mass transfer coefficient k f (m/s) of batch exeriment and ore diffusivity D P (m /s). The formulation of batch adsorber based on (Ahmed, 006). For batch adsorber, the mathematical model with ore diffusion model is: V L dc dt i Mass balance in fluid-bulk hase: 3W P A R P k f ( C C Pr R P ) 0 Where: V L = volume of fluid in the batch adsorber W A = mass of adsorbent ()

3 Cecilia Khoshaba Haweel and Ahmed J. Ali and Samar N. Alias Mass balance inside article hase: C dq 1 ( C P P D r ) t dt r r r (3) Initial conditions and boundary conditions t 0 Cb Co : C 0 q 0 r 0 : C dq 0 0 r dr r0 r0 r R C D k f [ Cb C ] r, r R rr Continuity equation in the bulk-fluid hase: In case of using Langmuir isotherm correlation: q q m C b 1 C b Where C b and C are the solute concentration in the fluidbulk and article hase, resectively. The external mass transfer coefficient for the solute adsorbed at certain article size and otimum agitation seed, can be obtained by the analytical solution for equation (.9) where at t = 0, C,r=R = 0 and C b = C o, hence: k f R V 3W t A L C ln C t o Where C o, C t are the solute concentration at time zero and time (t) and obtained from the tyical concentration decay curve. (4) (5) Modeling of Liquid-Solid Chromatograhical Column Using C i, the concentration in the stagnant fluid-hase (in the macroores), and writing the exression of interfacial flux leads to: qi t D 3k R fi C Z C C i, RR 1 b C C 0 (7) C 3 C k fi i, RR Z t R (8) Continuity equation inside the macroores: The article hase continuity equation in sherical coordinates is: b The formulation of Chromatograhical column is Based on (Volker, 1999, and Eggers, 003), the governing equations can be obtained from differential mass balances of the bulk-fluid hase and the article hase, resectively, for comonent i. Initial and boundary conditions The initial and boundary conditions may be reresented by the following equations: Initial condition (t = 0): C C ( 0, Z) 0 (10) C C ( 0, R, Z) 0 i i (11) 3

4 Searation of Bovine Serum Albumin Using Chromatograhical Column: Parameters and Simulation Boundary conditions: C Z D Z = 0: C C oi (1) Z =L: C 0 Z (13) R = 0: C i 0 R (14) R = R : C i k fi C C i, RR R D (15) i Model Parameters of Fixed-Bed Chromatograhy Column a) Correlations to estimate (molecular diffusivity) D m : Polson roosed the equation for high molecular weight solutes: D m 15 T ( ) (16) 1/3. M w Where: T is temerature (K), µ is the viscosity of the solution (Pa s) and M w is the molecular weight of the solute (Polson, 1950). batch adsorber or chromatograhy column. The bulkfluid hase and the article equations are first discretized using the FE (finite element) and the OC (orthogonal collocation) methods, resectively. The resulting ODE system is solved using an existing ODE solver rovided by MATLAB v-6.5.the code of MATLAB m-files for both batch adsorber and chromatograhy column are resented in (Ahmed, 006). This rogram is used for both single and multicomonent systems after changing some arameters. The model arameters used in the simulation rogram are listed in tables 1 and. Results and Discussion Isotherm Determinations and Constants The adsortion isotherm of BSA adsorbed onto DEAEcellulose of 0.35mm-article size at (30 o C and 5mg/ml Tris-HCl buffer H 7) is shown in (fig. ), were the initial concentration of BSA-solution was (0.1 mg/ml). By fitting the equilibrium data with Langmuir isotherm correlation, the results were good (q m =.0 mg/g) and (b = ml/mg). b) Correlation to estimate mass transfer coefficient k f in fixed-bed adsortion: 1/3 Sh Sc Re (Re 100) (17) Where: Sh=R k f /D m (Sherwood number), Sc = µ/(ρd m ) (Schmidt number), and Re=R ρν/µ (Reynolds number). µ is the fluid (water) viscosity, and ρ is the fluid (water) density (Jian, 004, and Montesinos, 005). 0.5 c) Correlation to estimate D b the axial disersion coefficient can be estimated from the following correlation: Where: R D b (Re 10) (18) Re ν is the interstitial velocity along the column = 4Q/επd (Q is the volumetric flow rate) and Re=R ρν/µ (Jian, 004, and Montesinos, 005). Simulation Program The model equations must be solved numerically. The numerical method of lines (MOL) is used in order to solve the couled system of PDE obtained ether from Fig. : Adsortion isotherm of BSA onto DEAEcellulose at (30 o C) Intraarticle Diffusivity coefficient D Estimations In order to find Intraarticle Diffusivity coefficient D, it was needed to find the mass transfer coefficient k f in batch adsortion rocess and then the numerical solution of batch adsorber model is alied. The external mass transfer k f, coefficient in batch adsorber was comuted from initial rate data, (i.e. from the concentration decay curve fig. 3) using equation (5): R VL C t k f ln (19) 3W At Co 4

5 Cecilia Khoshaba Haweel and Ahmed J. Ali and Samar N. Alias Where C o, C t are the solute concentration at time zero and time t. For accurate estimation of k f, samles were taken after 10, 0, and 30 minutes and analyzed immediately. The average calculated value of k f was m/s. The ore diffusivity coefficient was derived from the tyical concentration decay curve by iterative search technique redicted on the minimization of the difference between exerimental and redicted data from ore diffusion model (Ahmed, 006). The results are shown in (fig. 3). The ore diffusivity coefficient for BSA solute according to (fig 3) is m /s. Fig. 5 Break-through curve of BSA on DEAE-cellulose at C o = 0.5mg/ml Fig. 3 Comarison of the measured concentration-time data at (30 o C) with that redicted by ore diffusion model in batch adsorber. Chromatograhy Exeriments Figures (4 to 7) were obtained from the adsortion of BSA solution on DEAE-cellulose adsorbent at different concentrations of influent C o that at Q = 1ml/min. As follows: Fig. 6 Break-through curve of BSA on DEAE-cellulose at C o = 0.5mg/ml Figure 8 reresents the effect of changing the initial concentration at (0.15, 0.5, 0.5, and 1 mg/ml). From this figure, the change in inlet solute concentration markedly affects the shae and osition of the breakthrough curve. The higher the solute concentration the faster the break-through. However, the quantity of solution to be sulied is large when the solute concentration is low to arrive the otimum value. Fig. 4 Break-through curve of BSA on DEAE-cellulose at C o = 0.15mg/ml Fig. 7 Break-through curve of BSA on DEAE-cellulose at C o = 1mg/ml 5

6 Searation of Bovine Serum Albumin Using Chromatograhical Column: Parameters and Simulation Fig. 8 Break-through curve of BSA on DEAE-cellulose at different initial concentrations Fig. 11 Break-through curve of BSA on DEAE-cellulose at Q = 5ml/min Figures (9 to 1) were obtained from the adsortion of BSA solution on DEAE-cellulose adsorbent at different volumetric flow rates Q that at C o = 0.15mg/ml initial concentration. As follows: Fig. 1 Break-through curve of BSA on DEAE-cellulose at Q = 10ml/min Fig. 9 Break-through curve of BSA on DEAE-cellulose at Q = 1ml/min Figure 13 reresents the effect of changing the volumetric flow rate of the mole hase for (1, 3, 5, and 10 ml/min). Flow rate affects the film mass transfer coefficient. If R is unchanged, for correlation 17, k f is roortional to ν 0.5. The flow rate also affects the axial disersion coefficient according to correlation 18. For higher flow rates, Break-through is faster and oor adsortion efficiencies will result. When the flow rate is low, an increase in the sreading of the break-through curve will occur and an increase in the time that the solute is in contact with the stationary hase, allowing more time for adsortion and ermitting near-local equilibrium conditions. That the break oint is unchanged which is about 0.6 of initial concentration. Fig. 10 Break-through curve of BSA on DEAE-cellulose at Q = 3ml/min 6

7 Cecilia Khoshaba Haweel and Ahmed J. Ali and Samar N. Alias Fig. 13 Break-through curve of BSA on DEAE-cellulose at different volumetric flow rates All the resulting break-through curves obtained to the adsortion of BSA on DEAE-cellulose bed, are almost shar and the break oint occur about from the initial concentration for high volumetric flow rates, and the adsortion rocess is faster at about (30-40) minutes. From the observations of fig. 13, the break-through curves obtained for volumetric flow rates of (3-10 ml/min) are almost the same (i.e. the break-through curves are almost comatible). Table 1 Batch adsorber of BSA on DEAEcellulose system at 30 o C. Simulation Data Parameter Value R m ε 0.6 b m 3 /kg q m kg/kg D m /s k f m/s ρ kg/m 3 V L m 3 W A kg C o 0.1 kg/m 3 Table Data used in the simulation rogram of chromatograhy column Parameter Value Column length L 0.07 m Column diameter d m Particle diameter R m Column orosity ε Particle orosity ε 0.6 Axial disersion D b Correlation (18) Pore diffusivity D m /s Molecular diffusivity D m Correlation (16) External mass transfer Correlation (17) coefficient k f Maximum adsortion kg/kg caacity q m Langmuir isotherm m 3 /kg constant b Bed density ρ kg/m 3 Conclusions 1. Comarisons were made between the results obtained from the model rogram and that gained from exerimental work data, the agreement was very good, because all the arameters used in the model rogram were evaluated ether exerimentally or by emirical relations of grate accuracy. This rogram is alied in all cases of liquid-solid chromatograhy systems with a very good accuracy.. For the exerimental work data or literature data, Langmuir isotherm correlation gives good results. This correlation is a good assumtion for the system of very low concentrations of influent. 3. From the exerimental worked data it was concluded that it is referable to work with lean feed and low volumetric flow rate. In those conditions, a steeer or sread break-through curve will be gained which lead to increase the solute recovery efficiency. Also the effect of increasing the volumetric flow rat more than (3ml/min) is not efficient. 4. Whenever the break-through curve is sread (steeer), the break-through time oint will occur with less C/C o ratio 5. Axial disersion coefficient D b, which is affected by the flow rate of the system is more significant at lower flow rate and is less noticeable as the rate is increased, but the ore diffusivity D and external mass transfer coefficient k f is not that influential arameter because it reresent the diffusivity inside the ore of article, which its size is very small. 7

8 Searation of Bovine Serum Albumin Using Chromatograhical Column: Parameters and Simulation Nomenclature Greek Notation a i Constant from Langmuir correlation (q m. b) b i Langmuir isotherm constant (m 3 /kg) Bi i Biot number of mass transfer for comonent i, (k f R /ε D i ). ( - ) C oi Concentration used for nondimensionalization, max{cfi(t)}. C t Concentration of solute at any time. (kg/m 3 ) C Bulk-Fluid hase concentration of comonent i, (kg/m 3 ) C fi C i Feed concentration rofile of comonent i a time deendant variable (kg/m 3 ) Concentration of comonent i in the stagnant fluid hase inside thearticle macroores (kg/m 3 ) C* i Concentration of comonent i in the solid hase of article (kg/m 3 ) c = C /C oi ( - ) c i = C i /C oi ( - ) c* i = C* i / C oi ( - ) D Axial disersion coefficient of comonent i (m /s) D i Pore diffusivity coefficient of comonent i (m /s) D m Molecular diffusivity (m /s) d Inner diameter of a column (m) k fi, External mass transfer coefficient of comonent i (m/s) L Column length (m) M W Molecular weight. Pe Li Peclet number of axial disersion for comonent i, νl/d. Q Volumetric flow rate of mole hase (m 3 /s) q Concentration of solute in stationary hase bed (kg solute/kg sorbent) q m Maximum adsortion caacity (kg solute/kg sorbent) R Radial coordinates for article Re Reynolds number for shere, (R ρν/μ). R Particle radios (m) r = R/R Sc Schmidt number, (μ/ρdm). Sh Sherwood number for shere, (R k f /D m ). t Time (s) T Temerature (C) V L the final volume of the samle ut in beaker (ml) z Dimensionless axial coordinate Z/L. ε b Column or bed orosity. ε P Particle orosity. ν Interstitial velocity, (4Q/ε b πd ). τ Dimensionless time. μ Fluid (water) viscosity (Pa.s). ρ Fluid (water) density (Kg/m 3 ). ρ Particle density (Kg/m 3 ). Subscrits and suerscrit P Particle hase. m Mono layer. i i-th comonent. j j-th comonent. * Particle hase concentration Abbreviations BSA Bovine Serum Albumin. FDM Finite Difference Method. FEM Finite Element Method. OCM Orthogonal Collocation Method. ODE Ordinary Differential Equation. PDE Partial Differential Equation. References 1. Alias S. N., 007, "Searation of Biological Materials Using Chromatograhical Column", M Sc thesis submitted to Baghdad University.. Ahmed Kawther W., 006, Exerimental and modeling for the removal of multi ollutants by adsortion. A PhD thesis submitted to Baghdad university. 3. Eggers R., 003, Simulation of Frontal Adsortion. REPORT submitted to the Technische Universitat Hamburg-Harburg (TUHH). 4. ENCYCLOPEDIA OF CHEMICAL TECHNOLOGY, 199. Forth edition, volume 6, A Wily-Interscience Publication. 5. Freifelder David, 1998, Physical Biochemistry, second edition, W. H. Freeman and Comany, New York. 6. Jian-gang L. U., Feb. 8, 004, Mathematical modeling of salt-gradient ion-exchange simulated moving bed chromatograhy for rotein searations. Journal of Zhejiang University SCIENCE (JZUS). htt:// jzus@zju.edu.cn 7. Hubble John, 004, Adsortion and Chromatograhic searations. Literature (lecture) htt://eole.bath.ac.uk/cesjh/adsorb.htm E- mail: mailto:cesjh@bath.ac.uk 8

9 Cecilia Khoshaba Haweel and Ahmed J. Ali and Samar N. Alias 8. Montesinos-Cisneros R. M., Guzmلn-Zamudio R. Ortega-L ez J., Tejeda-Mansir Y. A., Marzo 005, FRONTAL CHROMATOGRAPHY OF PLASMIDS: PARAMETERS ESTIMATION AND SIMULATION. Revista Mexicana De Ingenieria QUIMICA Vol. 4 (005) Raley Ralh and walker John M., 1998, Molecular Biomethods Handbook. Humana Press, Totowa, NJ.s 10. Simson John M., 001, PROTEIN PURIFICATION PROCESS ENGINEERING, Marcel Dekker, Inc. 11. Stock R. and Rice C. B. F., 1974, Chromatograhic Methods, Halsted Press New York. 1. Volker Kasche, 1999, Simulation of Liquid Chromatograhy and Simulated Moving Bed (SMB) Systems. A PhD Thesis submitted to Hamburg University. 13. Polson, A. (1950). Some asects of diffusion in solution and a definition of a colloidal article. J. Phys. Colloid Chem., 54, فصل مصل الذم الحيواني باستخذام عمود الكروماتوكراف: 1 المعامالت والمحاكات طيظيهيا خ شاتا ا يم, وح ذ خ اد ػهي ط ز داذ انياص 1. كهيح ان ذطح, خايؼح تغذاد, لظى ان ذطح انكي يا يح. كهيح ان ذطح, خايؼح تغذاد, لظى ذطح انكي ياء انحياذيح 3. كهيح ان ذطح, خايؼح تغذاد, لظى ذطح انكي ياء انحياذيح الخالصة كز ياذ كزاف انصهة-انظائم ن صم انذو انحي ا ي Albumin) (Bovine Serum ػه يادج (Diethyleaminoethiel- cellulose) ا ( DEAE-cellulose ) لذ ذى اشرغان ػ هيا. نذراطح ذأثيز انرغيز انثرزكيش انذاخم ان انؼ د,0.5,0.15),0.5,1),3,5 and ثى دراطح ذأثيز انرغيز ت ؼذل اندزيا انحد ي )1ml/min( تثث خ يؼذل اندزيا انحد ي and 1mg/ml) 10ml/min) ترث خ انرزكيش انذاخم (.)0.15mg/ml انؼ م انردزيثي لذ ذى تاطرخذاو ػ د سخاخي صف ع ل لغز انذاخهي ) 1.5 طى( ع ن ( 50 طى(, يحش ت ادج (diethylaminoethyl-cellulose) DEAE-cellulose تغ ل ( 7 طى(. انظائم انذاخم يذخم إن انؼ د تاطرخذاو يضخح ي ع, eristaltic um ذزكيش انظائم انخارج ذى لياط تاطرخذاو خ اس األشؼح ف ق انث فظديح UV-sectrohotometer تذرخح حزارج (C 30) o تغ ل ي خي nm) 80). كا ذزكيش انظائم انذاخم )0.15mg/ml( األفضم ال يؼغي break-through oint لهيهح يغ break-through curve ي ثظظ. ذا انرزكيش )0.15mg/ml( اطرخذو في ذغثيك ذدارب أخز ن ؼذالخ خزيا يخرهفح. ا, يؼذالخ اندزيا انحد ي انري ي الم ي )3ml/min( ذك ذمزيثا يرغاتمح ن فض انظز ف انرشغيهيح. ن ذا ي األفضم ن ذا ان ظاو أ يؼ م ػه يؼذل خزيا حد ي لهيم يرزا ذ تي ( )1-3ml/min ت فض انظز ف. ي ديالخ انر اس ان رشرد نهكز ياذ غزاف انظائم_انصهة نخهيظ ث ائي ان زت ط تؼاللح ال ك ر ايش ثيزو لذ اطرخذو في اشرماق ان ديم انزياضي ن ذا انؼ م. ان ديم ان اذح لذ حم ػذديا تاطرخذاو تز ايح. MATLAB v.6.5 9