Optimal operation of simulated moving bed and varicol processes for bio separation

OPTIMAL OPERATION OF SIMULATED MOVING BED AND
VARICOL PROCESSES FOR BIO-SEPARATION

FALDY WONGSO

NATIONAL UNIVERSITY OF SINGAPORE
2003



OPTIMAL OPERATION OF SIMULATED MOVING BED AND VARICOL
PROCESSES FOR BIO-SEPARATION

FALDY WONGSO
(B. Tech., University of Gadjah Mada, Indonesia)

A THESIS SUBMITTED
FOR THE DEGREE OF MASTER OF ENGINEERING
DEPARTMENT OF CHEMICAL & ENVIRONMENTAL ENGINEERING
NATIONAL UNIVERSITY OF SINGAPORE
2003


To my beloved parents


Acknowledgement
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Acknowledgement
I would like to use this opportunity to salute my supervisor Prof. Ajay Kumar Ray for
allocating his valuable time through incessant guidance, brilliant idea and outstanding

patience during my 2 years of candidature in the National University of Singapore. It is
right to express my deepest gratefulness for everything we have done together. I
acknowledge the boundless thought and interest of Prof. Hidajat throughout my study and
also for sourcing the interesting chiral compounds as the scope of this work.
I would also like to thank Prof. Andrezj Krasalawski (Lappeenranta University of
Technology, Finland) for his sober assistance and advice prior and during the ESCAPE13 symposium, Prof. Iftekhar Abubakar Karimi and Prof. Gade Pandu Rangaiah for the
implied trust in my appointment as graduate tutor. My special thanks to Prof. Santosh
Kumar Gupta for the NSGA jumping genes algorithm which is a superb tool in carrying
out our work. It is a great honor to have Prof. Lakshminarayanan Samavedham’s resourceful presence and input at my PG seminar. Thanks are also due to Prof. Marc
Garland for the fruitful feedback on this thesis.
I want to express my gratitude to Mr Boey and Mr Mao Ning for the access to the
workstation and Mr Toh for the Microsoft Visual Fortran software, Mr Chun See Chong
for his assistance during my PG seminar, Mr Yeo Eng Hee and Mr Zhang Xinhuai from
the SVU team for their excellent support in the execution of my simulation program. I
always appreciate the National University of Singapore for the Research Scholarship
which made this work possible.
I will not forget the sincerity of my senior, Dr Effendi Widjaja who has encouraged
me to pursue my master degree in NUS. I’m thankful to Dr Zhang Ziyang for his
cooperative assistance and comment in all my doubts. Million thanks to all my labmates
and friends: Yu Weifang, Kanheya Mehrotra, Hari Prasad Janakiram Subramani, Chen
Saoping, Kurup Anjushri Sreedhar, Eng Yongyong, Zhang Yan, Naveen Agrawal, Lee
Yen Mei, and Paritam Kumar Dutta for the advices, jokes and time we spent together. I
also thank my flatmate, Handoko for sharing his experience on Microsoft Visual Basic
program. Foremost, my deepest gratitude goes to my beloved parents and all family
members whose understanding, support and love have always inspired me this time round.

i


Table of Contents

¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Table of Contents

Acknowledgments…………………………………………………………………………i
Table of Contents………………………………………………………………………....ii
Summary…..……………………………………………………………………………viii
Nomenclature……………………………………………………………………………..x
List of Figures…………………………………………………………………………..xiii
List of Tables…………………………………………………………………...………xvii

1.

Introduction………………………………………….……………………………1

2.

Literature Review………………………………………………………………...7
2.1.Review on Stereochemistry……………………………………………………7
2.1.1. An Overview on Chirality…..………...…………………………...7
2.1.2. Trends in Chiral Chemistry……………………………………....10
2.1.2.1. Trends in Pharmaceutical Industry………………………11
2.1.2.2. Trends in Fine Chemical Industry……………………….15
2.2.Chromatographic Separation…………………………………………………21
2.2.1. Elution Chromatography…………………………………………21
2.2.2. Continuous Crosscurrent Chromatography………………………24
2.2.3. Continuous Countercurrent Chromatography………………....…27
2.2.3.1. True Moving Bed Chromatography………………...…...28
2.2.3.2. Simulated Moving Bed Chromatography….……………32

ii

Table of Contents
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
2.3.Application of SMB Technology……………………………….…………….35
2.3.1. Petrochemical Industry…………………………….……………..36
2.3.2. Food and Flavor Industry………………………………………...38
2.3.3. Pharmaceutical and Fine Chemical Industry….………………….41
2.3.4. Protein Separation……………………………………….………..45
2.4.Optimization of Simulated Moving Bed.……………….…………………….47
2.4.1. Optimization Algorithm………………………………….………..48
2.4.2. Optimization Work on SMB…………………………….………...50
2.4.2.1. Single Objective Optimization…….……………….…….51
2.4.2.2. Multi Objective Optimization……….……….…………..55
2.5.Update on Moving Bed Technology…………………….……….…………...55
2.5.1. Reactive SMB………………………………….…….……..…….56
2.5.2. Ternary and Pseudo-SMB……………………….….…………….58
2.5.3. Supercritical Fluid-SMB Chromatography………..……………...62
2.5.4. Varicol Process……………………………………..…………….66
2.5.5. SMB Process Control…………………………………….…….....69

3. Simulated Moving Bed and Varicol Process………………………………………..72
3.1. Schematic Diagram of SMB and Varicol Process…………………………...72

4. Optimal Operation of Moving Bed Process for Chiral Drug Separation…………77
4.1. Background of Enantio-Separation ………………………………………….77
4.2. Numerical Simulation of SMB and Varicol Process……………..………….79

iii

Table of Contents
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
4.3. Calculation of Theoretical Number of Plate…………………………………82
4.4. Model Validation…………………………………………………………….85
4.5. Sensitivity Study……………………………………………………………..88
4.5.1. The Effect of Switching Time……………………………………..88
4.5.2. The Effect of Feed Flow Rate……..……………………………….91
4.5.3. The Effect of Raffinate Flow Rate…………………………………92
4.5.4. The Effect of Desorbent Flow Rate…………..……………………94
4.6. Single Objective Optimization………………………………………...…….96
4.6.1. Case 1. Single Objective Optimization: Maximization of throughput
..........................................................................................................97
4.6.2. Case 2. Single Objective Optimization: Minimization of desorbent
consumption ……………………………………………………...100
4.7. Multi-Objectives Optimization……………………………………………..105
4.7.1. Case 3. Two Objectives Optimization: Maximization of raffinate and
extract productivity………………………………....…...………..107
4.7.2. Case 4. Two Objectives Optimization: Maximization of raffinate purity and productivity…...................................................................109
4.7.3. Case 5. Two Objectives Optimization: Maximization of extract purity and productivity……..….……………...………………………112
4.7.4. Case 6. Two Objectives Optimization: Maximization of throughput
and minimization of desorbent consumption………….……...…115
4.7.5. Case 7. Three Objectives Optimization: Maximization of raffinate
and extract productivity and minimization of solid requirement...118

iv


Table of Contents
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

4.7.6. Case 8. Three Objectives Optimization: Maximization of raffinate
and extract productivity and minimization of desorbent consumption………………………………………...……………….……..120
4.8. Pure Separation Regime for Binary Separation…...………………………..122
4.9. The Effect of Sub-interval and Partial Feed Operation….............................127

5. Optimization Study of Continuous Chromatographic Separation of a Chiral
Intermediate………………………………………………………………………….134
5.1. The Application of 1,1’-bi-2-naphtol……………………………….………134
5.2. Mathematical Model of SMB and Varicol Process...………………………137
5.3. Model Validation…………………………………………………………...141
5.4. Sensitivity Analysis...………………………………………………………145
5.4.1. The Effect of Switching Time……………………………………146
5.4.2. The Effect of Feed Flow Rate………………………...…….……147
5.4.3. The Effect of Raffinate Flow Rate………………………….…….148
5.4.4. The Effect of Desorbent Flow Rate………………………………149
5.4.5. The Effect of Column Number……………………………….…..150
5.5. Optimization Study………….………………………………………...……153
5.5.1. Single Objective Optimization……………………………………154
5.5.1.1. Case 1. Single Objective Optimization: Maximize feed flow
rate….…………………….…………………………..…154
5.5.1.2. Case 2. Single Objective Optimization: Minimize desorbent
flow rate.………………….……………….…………….156

v


Table of Contents
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
5.5.2. Multi Objective Optimization…………………………………….158
5.5.2.1. Case 3. Multi Objective Optimization: Maximize raffinate

and extract purity………………………………………..160
5.5.2.2. Case 4. Multi Objective Optimization: Maximize raffinate
and extract productivity…………………………………163
5.5.2.3. Case 5. Multi Objective Optimization: Maximize raffinate
purity and productivity………………………………….165
5.5.2.4. Case 6. Multi Objective Optimization: Maximize extract purity and productivity……..………………………………168
5.5.2.5. Case 7. Multi Objective Optimization: Maximize feed and
minimize desorbent rate………………………………….170
5.5.2.6. Case 8. Multi Objective Optimization: Maximize raffinate
and extract productivity and minimize desorbent rate…..173
5.5.2.7. Case 9. Multi Objective Optimization: Maximize raffinate
and extract productivity and minimize column length…..175
5.6. Complete Separation Region……………………………………………….177
5.7. The Effect of Flow Rate in Zone 1 (Q1) on Countercurrent Separation……181

6. Conclusion and Recommendation…………………….……………………………185
6.1. Optimal Operation of SMB and Varicol Processes for Chiral Drug Separation
……………………………………………………………………………...185
6.2. Optimization Study of Continuous Chromatographic Separation of a Chiral
Intermediate in SMB and Varicol System………………………………….187

vi


Table of Contents
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
6.3. Recommendation for Future Work…………………………………………190

References…………………………………………………………….………………..192
List of Publications…………………………………………………………………….212

Appendix A……………………………………………………………….………….....213
Appendix B……………………………………………………………….………….....217
Appendix C………………………………………………………………….……….....219
Appendix D……………………………………………………………….…………….225
Appendix E……………………………………………………………….…………….226

vii


Summary
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

SUMMARY

In recent years, there has been considerable interest in the synthesis and separation of
enantiomer of organic compounds especially due to their importance in the biotechnology
and pharmaceutical industry. There are two pedestal ways of producing pure enantiomer
form of a compound, conventional chemical synthesis and asymmetric synthesis.
Conventional synthesis, even though economically attractive, might result in chiral
molecules. This method is favored for compounds where both enantiomers are of equal
importance. Asymmetric synthesis, on the other hand, can selectively produce certain
species. This method is highly preferred when one species of the optical isomer is
toxic/harmful. However, it takes not only many years to develop asymmetric synthesis for
any particular drug, but at times the process is expensive. Simulated moving bed (SMB)
technology and its modification, Varicol process, has become an alternate technology for
separation of such difficult or nearly impossible to separate mixtures with extremely low
separation factor, which is difficult to be separated using conventional extraction, and the
considerably low volatility gap makes it impractical to be purified using common
distillation. In this technology, separation is achieved by differential migration rates of the
two components, which is further enhanced by simulating countercurrent movement of

the solid and the liquid phases. The general case of simulated countercurrent process is
that successive switching of the feed and the product positions at timed interval simulates
the movement of solid. By periodically changing the feed and the product locations
sequentially along a fixed bed leads to high mass transfer driving forces. In traditional
SMB process, a synchronous switch of all the inlet and outlet and streams are performed

viii


Summary
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

while in the Varicol process a non-synchronous switch is employed to render more
flexibility to the system. High product yield and purity can be achieved by selective
adsorption in which compounds of a mixture are driven at different speeds through the
chromatographic bed. However, like all methods, the chromatographic method suffers
from certain drawbacks: requirement of expensive stationary phases, relatively higher
volume of mobile phase consumption and high dilution of the separated products.
In this thesis, two separation problems are considered. The first is the separation of
chiral drugs, SB-553261, and the second is the separation of chiral intermediates, 1,1’-bi2-naphtol. Equilibrium dispersive model coupled with lumped kinetic approximation and
non-linear equilibrium isotherm constitute the simulation model, which was verified with
published experimental results for the above two systems. Single as well as multiple
objective optimization studies were carried for many different meaningful optimization
formulations. The optimization work was carried out using Non-Dominated Sorting
Genetic Algorithm with jumping genes (NSGA II-JG). It employs the principles of
genetics and the Darwinian principle of natural selection (i.e. survival of the fittest). A
systematic optimization study is carried out involving the use of single objective as well
as several objectives, which is often conflicting with each other. The use of nondominated sorting genetic algorithm with jumping genes resulted in Pareto optimal
solutions, which can easily be explained by the Triangle theory. Optimal operating
conditions and optimal Pareto solutions clearly show that the concept of multiobjective

optimization could be very useful in the design and operation of SMB and Varicol process
for many important industrial applications. These optimal results help enhancing the
performance of the existing design and serve as an important design tool for new
developments.

ix


Nomenclature
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Nomenclature

A

more retained component, column area (cm2)

B

less retained component

C

liquid phase concentration (g/l)

CSP

chiral stationary phase

D

desorbent flow rate (ml/min)

DL

apparent axial dispersion coefficient (cm2/min)

E

extract flow rate (ml/min)

F

feed flow rate (ml/min)

H

height equivalent to theoretical plate (cm), Langmuir isotherm parameter

I

objective function index

J

theoretical number of cells

k

mass transfer coefficient (min-1)

K

adsorption constant, equilibrium constant

Lcol

length of individual column (cm)

m

flow rate parameter

N

total number of column, switching index

P

zone I in SMB unit

Pr

productivity (g/h, g/day)

Pur

purity (%)

q

solid phase concentration (g/l)

Q

liquid flow rate (ml/min), zone II in SMB unit

x


Nomenclature
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
R

raffinate flow rate (ml/min), zone III in SMB unit

Rec

recovery (%)

S

zone IV in SMB unit

SC

solvent consumption (m3 desorbent/kg product)

SMB

simulated moving bed

t

time (min)

Vcol

volume of individual column

Varicol

variable column system

Y

yield (g/h/lsolid or g/day/gsolid)

z

axial coordinate (cm)

Greek Symbols
χ

column configuration

∆P

pressure drop (psi)

ε

void fraction

δ

phase ratio

φ

zone index in SMB and varicol

σ

relative carrying capacity

ζ

pseudo solid phase velocity

Subscripts and Superscripts
col

column

D

desorbent

xi

Nomenclature
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
E

extract

F

feed

g

gas

i

component index

j

column index

k

cell index

R

raffinate

s

solid, switching

xii


List of Figures
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
List of Figures

Figure 2.1

Non-superimposable mirror image structures…….……………………….8

Figure 2.2

Schematic diagram of TMB chromatographic process…………………...29

Figure 2.3

Adsorption-desorption diagram of SMB process………………………...33

Figure 2.4

Ternary SMB with 5-zone configuration…….…………………………...59

Figure 2.5

Ternary multi zone SMB configuration..…………………………………60

Figure 2.6

Pseudo 4-zone SMB configuration (JO process)………….……………...61

Figure 2.7

Pseudo 4-zone SMB concentration profile (JO process)...….……………62

Figure 2.8

Switching sequence for 4-sub interval Varicol process……….………….67

Figure 2.9

Varicol design connection………….…………………………………….68

Figure 3.1

Schematic diagram of 4-zone 8-column SMB with 2 columns per zone...72

Figure 3.2

Switching profile of 4-zone 5-column SMB and Varicol processes.…….73

Figure 3.3

Schematic diagram of 3-zone 4-column SMB/Varicol process……….…75

Figure 4.1

Molecular structure of SB-553261……………………………………….82

Figure 4.2

Breakthrough curve during single column experiment (pulse injection at Q
= 1.5 ml/min, Vinj = 10 µl, Cinj = 1 g/l, UV detection λ = 310 nm)………83

Figure 4.3

Simulated elution profile for enantioseparation of SB-553261………….85

Figure 4.4

Concentration profile on 6-column SMB (Nplate = 120, Q1 = 15.3 ml/min,
QF = 0.3 ml/min, QR = 1.79 ml/min, QD = 8.55 ml/min, ts = 1.11 min,
1/2/2/1 setup)……………………………………………………………..87

Figure 4.5

Sensitivity study: (a) Plot of productivity vs switching time (b) Plot of
purity vs switching time……………………..……….…………..………89

xiii


List of Figures
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Figure 4.6

Sensitivity study: (a) Plot of productivity vs feed flow rate (b) Plot of
purity vs feed flow rate.………………..…………….……………...……91

Figure 4.7

Sensitivity study: (a) Plot of productivity vs raffinate flow rate (b) Plot of
purity vs raffinate flow rate..…………….…………….…………………92

Figure 4.8

Effect of raffinate flow rate on steady state concentration profile for 5column SMB. (a) QR = 1.5ml/min, (b) QR = 2.5ml/min, (c) QR = 3.5
ml/min…………………………………………………………………….93

Figure 4.9

Sensitivity study: (a) Plot of productivity vs desorbent flow rate (b) Plot of
purity vs desorbent flow rate……………..……….………………...……94

Figure 4.10

Effect of desorbent flow rate on steady state concentration profile for 5column SMB. (a) QD = 9.5 ml/min, (b) QD = 10 ml/min, (c) QD = 10.5
ml/min…………………………………………………………………….95

Figure 4.11

Pareto optimal solution and plot of decision variables (case 3) for SMB and
Varicol systems………………………………………………………….108

Figure 4.12

Pareto optimal solution and plot of decision variables (case 4) for SMB and
Varicol systems………………………………………………………….110

Figure 4.13

Pareto optimal solution and plot of decision variables (case 5) for SMB and
Varicol systems………………………………………………………….113

Figure 4.14

Pareto optimal solution and plot of decision variables (comparison between
case 3, case 4 and case 5) for SMB and Varicol systems……………….114

Figure 4.15

Pareto optimal solution and plot of decision variables (case 6) for SMB and
Varicol systems………………………………………………………….116

xiv


List of Figures
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Figure 4.16

Pareto optimal solution and plot of decision variables (case 7) for SMB and
Varicol systems………………………………………………………….119

Figure 4.17

Pareto optimal solution and plot of decision variables (case 8) for SMB and
Varicol systems………………………………………………………….121

Figure 4.18

Optimum operating regime in m2-m3 plane for enantioseparation of SB553261 racemate using 5-column SMB (case 6)………………………..123

Figure 4.19

Optimum operating regime in m2-m3 plane for enantioseparation of SB553261 racemate using 5-column Varicol (case 6)……………………..123

Figure 4.20

m operating plane for enantioseparation of SB-553261 racemate using 5column SMB (case 6)……………………………………….…………..124

Figure 4.21

m operating plane for enantioseparation of SB-553261 racemate using 5column Varicol (case 6)…………………………………………………124

Figure 4.22

The effect of subinterval for 5-column varicol (case 3)………………...128

Figure 4.23

Discrete feed operation for 5-column 4-interval SMB process…………129

Figure 4.24

Comparison between discrete feed and constant feed operation for 5column 4-interval SMB and Varicol processes (case 3)…………..……131

Figure 4.25

Feed profile for discrete feed and constant feed operation for 5-column 4interval SMB and Varicol processes (case 3)………………………...…132

Figure 5.1

Molecular structure of 1,1'-bi-2-naphtol optical isomer……..………….135

Figure 5.2

Experimental and simulated concentration profile on 8-column SMB based
on operating parameter in Pais et al., 1998 (symbol: experiment, black:
simulation by Pais et al., grey: simulation by this work)………………..143

Figure 5.3

The effect of switching time on purity and productivity……………......146

xv


List of Figures
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Figure 5.4

The effect of feed flow rate on purity and productivity………………...147

Figure 5.5

The effect of raffinate flow rate on purity and productivity……………148

Figure 5.6

The effect of desorbent flow rate on purity and productivity………......149

Figure 5.7

The effect of column number in zone I on purity and productivity.........150

Figure 5.8

The effect of column number in zone II on purity and productivity……151

Figure 5.9

The effect of column number in zone III on purity and productivity…...152

Figure 5.10

The effect of column number in zone IV on purity and productivity…..152

Figure 5.11

Multi objective optimization results (case 3) for SMB and Varicol….....161

Figure 5.12

Multi objective optimization results (case 4) for SMB and Varicol….....164

Figure 5.13

Multi objective optimization results (case 5) for SMB and Varicol….....166

Figure 5.14

Multi objective optimization results (case 6) for SMB and Varicol…....169

Figure 5.15

Multi objective optimization results (case 7) for SMB and Varicol…....171

Figure 5.16

Multi objective optimization results (case 8) for SMB and Varicol…....174

Figure 5.17

Multi objective optimization results (case 9) for SMB and Varicol…....176

Figure 5.18

Plot of binary separation plane for optimization case 3….………...…...178

Figure 5.19

Plot of binary separation plane under various purity constraints……….179

Figure 5.20

Plot of m flow rate parameter for all points in Pareto case 3….……...…180

Figure 5.21

The effect of Q1 on optimum points in case 3…………………………..182

Figure 5.22

Profile of flow rate in zone I (Q1) vs raffinate purity……………….......183

xvi


List of Tables
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
List of Tables

Table 2.1

Chiral molecules and bioactivity…………………………………………10

Table 2.2

Sepracor racemic switch project………………………………………….13

Table 2.3

Moving bed application in petrochemical industry……………………....37

Table 2.4

Moving bed application in food industry………………………...……….40

Table 2.5

Chiral separation on SMB technology………………………..…………..42

Table 3.1

Possible column configurations for SMB and Varicol processes.………..74

Table 4.1

Chiral drug global sales data……………………...………………………78

Table 4.2

Column specification and adsorption isotherm for SMB experiment used
by Ludemann-Hombourger et al. (2002)…………………………………86

Table 4.3

Performance comparison for 5-column, 4-subinterval Varicol process….87

Table 4.4

Relative velocity of each species in zone II at various switching time for
enantioseparation of SB-553261 racemate…………………………..…...90

Table 4.5

Optimization attributes used in single objective optimization………...…98

Table 4.6

Optimum column configuration for SMB and Varicol process for chiral
drug separation……………………………………………………….…..99

Table 4.7

Single objective optimization results...……………….…………………102

Table 4.8

Optimization attributes used in multi objectives optimization…....…….106

Table 5.1

Experimental and simulation process parameter for enantioseparation of
1,1'-bi-2 naphtol racemate…………………………………………….....141

xvii


List of Tables
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Table 5.2

Comparison between experimental (Pais et al., 1997a) and simulation
result for enantioseparation of 1,1'-bi-2 naphtol racemate at various
switching time…………………………………………………….…......142

Table 5.3

Comparison between experimental (Pais et al., 1997a) and simulation
result for enantioseparation of 1,1'-bi-2 naphtol racemate at various column
configuration………………………….………………………................144

Table 5.4

Comparison between experimental (Pais et al., 1997b) and simulation
result for enantioseparation of 1,1'-bi-2 naphtol racemate…...………....144

Table 5.5

Comparison between experimental (Pais et al., 1998) and simulation result
for enantioseparation of 1,1'-bi-2 naphtol racemate……………….……145

Table 5.6

Single objective optimization attributes used in the enantioseparation of
1,1'-bi-2 naphtol racemate……………………………………………….155

Table 5.7

Optimum column configuration for SMB and Varicol processes for enantioseparation of 1,1'-bi-2 naphtol racemate……….…...………………...156

Table 5.8

Single objective optimization result in the enantioseparation of 1,1'-bi-2
naphtol racemate………………………………………………………...157

Table 5.9

Multi objective Optimization Attributes used in the Enantioseparation of
1,1'-bi-2 naphtol racemate……………………………………………….159

Table 5.10

Condition for counter-currency for case 3 in each SMB zone for
enantioseparation of 1,1′-bi-2-naphtol racemate…………………….…..179

Table 5.11

Condition for counter-currency for variable Q1 in each SMB zone for
enantioseparation of 1,1′-bi-2-naphtol racemate…………………….….184

Table B.1

Experimental data on productivity and desorbent consumption………...217

xviii


Introduction
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Chapter 1
Introduction

Synthesis and isolation of enantiomers of organic compounds has been gaining steady
interest over the past few years owing to their advantages in the fine chemical and
pharmaceutical industry. A number of synthesis techniques are available for the
preparation of single enantiomer drugs. Among these methods, resolution by preparative
chiral chromatography is receiving increasing attention nowadays. Chromatographic
principle is believed to be one of the powerful tools in which components are separated by
differential migration through a system of two phases: column packing which acts as
stationary phase and liquid phase which acts as the mobile phase. Advantages such as
high separating power, selectivity, versatility, low operating cost and mild operating
conditions has led to the convenient utilization of chromatography in the field of
enantioseparation. High product yield and purity can be achieved by selective adsorption
in which compounds of a mixture are driven at different speeds through the
chromatographic bed. However, like all methods, the chromatographic method suffers
from certain drawbacks: large requirement of expensive stationary phases, relatively high
volume of mobile phase consumption and high dilution of the separated products.
Recently the Simulated Moving Bed (SMB) chromatographic technology has been
introduced for continuous separation of chemicals with extremely low separation factor,
which is difficult to be separated using conventional extraction, and considerably low
volatility gap makes it impractical to be purified using distillation. SMB chromatography
has excellent prospects in the field of pharmaceutical and fine chemicals separation and its

1


Introduction
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
development requires better understanding of the process dynamics. The concept of SMB
is based on true moving bed process, in which a stream of solids flows countercurrent to
an inert carrier fluid passing a stationary inlet, by periodically changing feed and product

locations sequentially along a fixed bed leading to high mass transfer driving forces.
Successive switching of the feed and product positions at timed interval simulates the
countercurrent movement of solid. True countercurrent motion is thus replaced by a
periodic motion, while overcoming the problems of solids handling and attrition inherent
in moving bed operations, as well as avoiding flow channeling that might occur while
scaling up to large column diameters.
Continuous chromatography based on simulated moving bed technology has been
widely used in the separation and purification of chemicals, which are difficult to be
separated using other methods. This has found ultimate interest in the area of biochemical,
pharmaceutical and fine chemicals due to its ability to meet the market needs of high yield
and high purity coupled with moderate processing conditions. This has become more
popular in recent years due to the availability of chromatographic phases, while
eliminating the drawbacks of batch chromatography, namely dilution of species and low
adsorbent utilization leading to cleaner, smaller, safer and faster processes.
The design of a SMB unit depends on appropriate choice of operating conditions such
as liquid and solid flow rates. Further investigation into the system reveals the complex
interplay between variables, i.e. high feed rate increase the productivity at the expense of
high eluent/desorbent rate and product purity making this system prone to optimization.
The optimization problem for simulated countercurrent process is to find the optimum
values of the key process parameters under certain specified constrains. In this work, the

2


Introduction
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
efficiency of the SMB technology in carrying out enantio-separation of SB-553261 and
1,1’-bi-2-naphtol racemate at various operating configuration are examined.
Modeling and simulation of SMB process towards improvement in process
performance have received great attention as this effort could yield significant savings for

industry dealing with biotechnology, pharmaceutical and fine chemistry. There are two
approaches of modeling SMB chromatographic process. The first is to model the true
moving bed process to describe the dynamics neglecting cyclic switching. The second
approach is to connect dynamic model of single chromatographic columns while
considering the cyclic port switching. In this work, a model for prediction of cyclic steady
state performance of SMB is developed based on the latter approach, direct simulation of
port switching (SMB model). Two types of SMB model are considered here, namely,
mixing cell model and continuous-flow model. The first model is known as equilibrium
stage model, which is suitable under the usual conditions of high performance preparative
chromatography. The latter model assumes axial dispersion flow for the liquid phase and
linear driving force (LDF) approximation for intra-particle mass transfer rate. Both
models take into account multi-component adsorption equilibrium.
A recent development in SMB technology is known as the Varicol process. The basic
configuration of the Varicol process is similar to SMB, only the injection and collection
lines are shifted at different times non-synchronously. Hence, there will be variations of
zone length for a multicolumn systems and the system will return to its original position at
the last periodic sub-interval. The zone lengths are continuously varying and the increase
of one column is compensated for by the decrease of the adjacent one. Varicol process

3