27.01.2015 Views

Design of Simulated Moving Bed Chromatography - CPAC

Design of Simulated Moving Bed Chromatography - CPAC

Design of Simulated Moving Bed Chromatography - CPAC

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Simulated</strong> <strong>Moving</strong> <strong>Bed</strong> <strong>Chromatography</strong><br />

for Insulin Purification:<br />

<strong>Design</strong> Methods and Simulation Tools to<br />

Achieve High Product Purity and High<br />

Yield<br />

George Weeden and N-H. Linda Wang<br />

School <strong>of</strong> Chemical Engineering<br />

Purdue University


Overview<br />

• <strong>Chromatography</strong> and SMB<br />

– Principles<br />

– Key Barriers<br />

• Purdue SMB Technologies<br />

– Standing Wave <strong>Design</strong> & Optimization<br />

– VERSE Dynamic Simulations<br />

– Versatile SMB Equipment (V SMB)<br />

– Knowledge- Based <strong>Design</strong> Method<br />

• Examples<br />

– Tandem SMB for insulin purification<br />

– Five-zone SMB to recover six sugars from biomass<br />

hydrolysate


CONVENTIONAL CHROMATOGRAPHY<br />

A Batch Process<br />

Eluent<br />

Feed<br />

Pre-equilibration<br />

Adsorbent<br />

Particles<br />

3


<strong>Chromatography</strong>-Advantages<br />

• High selectivity – 1 in 10,000.<br />

• Compact volume compared to extraction.<br />

• Lower solvent usage than extraction.<br />

• Mild operating conditions – suitable for<br />

fragile compounds.<br />

• Versatile mechanisms <strong>of</strong> separation –<br />

adsorption, ion exchange, size exclusion,<br />

etc.


Mechanisms <strong>of</strong> Adsorption


Limitations <strong>of</strong> Batch <strong>Chromatography</strong> for<br />

Production<br />

• Difficult to achieve both high purity<br />

and high yield<br />

• Low adsorbent utilization<br />

• High solvent consumption<br />

• Product dilution<br />

• Not a continuous process<br />

• High cost


Solution I- Binary Separation in a<br />

Continuous <strong>Moving</strong> <strong>Bed</strong>


Solution II–<strong>Moving</strong> Port or SMB<br />

Desorbent<br />

1<br />

ZONE I<br />

2<br />

Extract<br />

8<br />

Solute<br />

Movement<br />

3<br />

ZONE IV<br />

7<br />

Step N<br />

ZONE III<br />

4<br />

ZONE II<br />

Raffinate<br />

6<br />

5<br />

Feed


Advantages <strong>of</strong> SMB over Batch<br />

• Only partial separation <strong>of</strong> solutes is required to<br />

obtain high purity.<br />

• Higher yield than batch – 10% more than batch.<br />

• High purity achievable without sacrificing yield.<br />

• High productivity– 5 to 10 fold increase.<br />

• Less solvent – 5 to 10 fold less.<br />

• Reduced environment impact.<br />

• Reduced footprint, equipment size, and manpower.<br />

• Continuous process-high throughput.<br />

• Economical for large scale separations.


Commercial Applications <strong>of</strong> SMB<br />

• Hydrocarbons – UOP (Molex, Parex)<br />

(1970’s)<br />

• Sugars – Tate & Lyle (1980’s)<br />

• Chiral Drugs (1990’s)<br />

• All binary separations <strong>of</strong> low MW<br />

chemicals in four-zone SMB


Key Barriers in SMB<br />

• A four-zone SMB has 9 design parameters - 2 9 to 3 9<br />

trials.<br />

• Many system & operating parameters to be optimized.<br />

• Method for multi-component separation not well<br />

developed.<br />

• Complex transient and cyclic steady state phenomena.<br />

• Equipment more complex than batch chromatography.<br />

• No commercial equipment for multi-component<br />

separation.<br />

• Control <strong>of</strong> batch identity and residence time<br />

• Experiments expensive and time-consuming


Purdue’s Platform SMB Technologies<br />

1. Standing Wave <strong>Design</strong> and<br />

Optimization.<br />

2. VERSE simulations.<br />

3. Versatile SMB equipment.<br />

4. A systematic knowledge-based<br />

method for process design and<br />

optimization.


Conc.<br />

Standing Wave <strong>Design</strong>: Binary, Ideal, Linear Systems<br />

Desorbent<br />

Feed<br />

I II III IV<br />

Extract<br />

<strong>Bed</strong> Position<br />

Raffinate<br />

=<br />

= u<br />

= u<br />

= u<br />

F<br />

u<br />

Feed<br />

I<br />

w,slow<br />

II<br />

w,fast<br />

III<br />

w,slow<br />

IV<br />

w,fast<br />

=<br />

Sε<br />

b<br />

<br />

u<br />

III<br />

0<br />

u<br />

II<br />

0<br />

<br />

u w,i<br />

a<br />

= u 0<br />

a<br />

/ ( 1 + Pd i )<br />

= wave velocity <strong>of</strong> component i in zone a<br />

= L c / t s = port movement velocity<br />

F feed = feed flow rate<br />

u 0<br />

a<br />

= interstitial velocity in zone a<br />

e b = inter-particle void fraction<br />

P = (1 – e b ) / e b = bed phase ratio<br />

d i = distribution coefficient <strong>of</strong> i<br />

S = column cross-sectional area


Conc.<br />

Standing Wave <strong>Design</strong>- Linear, Non-ideal Systems<br />

Desorbent<br />

Desorption wave <strong>of</strong> Low affinity<br />

solute confined in Zone II<br />

Feed<br />

Adsorption wave <strong>of</strong> Low affinity<br />

solute confined in Zone IV<br />

Zone I Zone II Zone III Zone IV<br />

Extract<br />

<strong>Bed</strong> Position<br />

Raffinate<br />

•Ideal System: Matching wave velocities with port<br />

velocity to ensure high purity and yield<br />

•Non-ideal Systems: Difference in wave velocity and<br />

port velocity allows wave focusing in each zone


Concentration<br />

Standing Wave <strong>Design</strong>: Multi-Component System<br />

with Mass Transfer Effects<br />

Desorbent<br />

Feed<br />

(1, 2, 3)<br />

I II III IV<br />

*<br />

1 Fast Solute<br />

2 Intermediate<br />

3 Slow Solute<br />

Extract<br />

(2, 3)<br />

<strong>Bed</strong> Position<br />

*<br />

Raffinate<br />

(1)<br />

*<br />

Focused Standing Wave<br />

Pinched Wave


Standing Wave <strong>Design</strong> & Optimization-<br />

Advantages<br />

• Four zone flow rates and port velocity are<br />

solved easily.<br />

• Little computation time required.<br />

• Easy extension to nonlinear, multicomponent,<br />

multi-zone separations.<br />

• Achieving maximum productivity and<br />

minimum solvent consumption for a given<br />

configuration.<br />

• Reduce search space by five dimensions or<br />

more for optimization.


VErsatile Reaction SEparation Simulator (VERSE)<br />

Mechanisms<br />

.<br />

Fluid<br />

Bulk<br />

Fluid<br />

Bulk Convection<br />

Adsorption<br />

Sites<br />

Reaction<br />

Adsorption<br />

Film<br />

Diffusion<br />

Interference<br />

Ion<br />

Exchange<br />

Pore<br />

Diffusion<br />

Sorbent<br />

Particle<br />

Denaturation<br />

Size<br />

Exclusion<br />

Surface<br />

Diffusion<br />

Fluid<br />

Solid<br />

Phase<br />

Reaction<br />

Pore<br />

Fluid


VERSE<br />

VErsatile Reaction SEparation Simulator<br />

• Simulation based on a detailed dynamic rate model<br />

(transient differential mass balances)<br />

• Thousands <strong>of</strong> equations solved at each instance<br />

• To predict column pr<strong>of</strong>iles and histories<br />

• To understand transient phenomena in batch<br />

chromatography, carousel, and SMB<br />

• To understand residence time and batch integrity<br />

issues in SMB<br />

• To reduce the number <strong>of</strong> costly experiments


Versatile SMB<br />

To implement complex processes for multicomponent<br />

separations<br />

Valve Controls<br />

Valves<br />

Sephadex<br />

G50 Columns<br />

Pumps


Features <strong>of</strong> Versatile SMB<br />

• Dedicated valves and manifolds for high purity<br />

and high yield<br />

• Allows zone bypass, open loop, multizone,<br />

multisolvent ,and multicomponent separation.<br />

• Easy column expansibility.<br />

• <strong>Moving</strong> port chromatography.<br />

• Online decoupled regeneration (Animation)<br />

• Independent column switching for variable step<br />

time


Knowledge-Based <strong>Design</strong> & Optimization<br />

Engineering<br />

Parameters<br />

Mobile Phase, pH,<br />

Stationary Phase,<br />

Temperature<br />

Process <strong>Design</strong> &<br />

Optimization<br />

VERSE<br />

Simulator<br />

Experimental<br />

Verification<br />

Batch or<br />

V-SMB Exp<br />

Solubility,<br />

Capacity,<br />

Selectivity<br />

Elution Pr<strong>of</strong>ile,<br />

Purity, Yield,<br />

Concentration,<br />

Productivity<br />

Elution Pr<strong>of</strong>ile,<br />

Purity, Yield,<br />

Concentration,<br />

Productivity<br />

Pulse,<br />

Frontal<br />

Voidage,<br />

Isotherm,<br />

Mass Transfer,<br />

Numerical<br />

Parameters<br />

SWD,<br />

Scale-Up,<br />

Optimization<br />

Sensitivity,<br />

Trade-<strong>of</strong>f<br />

Optimal<br />

<strong>Design</strong><br />

Optimal<br />

Process<br />

(Hirtzko et al., 1999; Mun et al., 2005; Wu et al., 1997; Xie et al., 2002)


Novel Purdue SMB Processes<br />

• Pilot Scale Separations<br />

– Insulin from zinc chloride and protein<br />

impurities<br />

– Six sugars from biomass hydrolysate (>10<br />

components)<br />

• Lab Scale<br />

– Antibiotic Drug<br />

– Amino Acids and Derivatives<br />

– Lactic Acid<br />

– Paclitaxel (anticancer drug )<br />

– Chiral Drugs


Knowledge-Based <strong>Design</strong> <strong>of</strong><br />

Tandem SMB for Insulin Purification


Biosynthetic Human Insulin (BHI) Process<br />

Recombinant E. coli.<br />

Trp proinsulin<br />

Proinsulin<br />

Crude insulin<br />

Homogenization<br />

Cleavage, oxidative sulfitolysis, folding<br />

Enzymatic transformation<br />

Ion-exchange<br />

Reversed-Phase HPLC<br />

Size exclusion chromatography (SEC)<br />

Purified BHI<br />

Crystallization<br />

Isolation & purification steps take 10 to 12 weeks.


Conventional Size Exclusion <strong>Chromatography</strong><br />

1<br />

0.8<br />

Insulin<br />

(Product)<br />

C/CF<br />

0.6<br />

0.4<br />

Proinsulin /<br />

HMWP<br />

Fronting<br />

Zinc<br />

Chloride<br />

0.2<br />

0<br />

Proinsulin / HMWP<br />

Insulin (Product)<br />

Zinc Chloride<br />

Animation<br />

0 100 200 300<br />

Volume (L)<br />

BHI exp. HPI exp. HPI sim.<br />

• Existing SEC batch process<br />

ZnCl2 exp. BHI sim. ZnCl2 sim.<br />

• Yield < 90%<br />

• Purity >99%<br />

• Resin throughput = 0.23 L / hr/100 L BV


Tandem SMB for Insulin Purification<br />

ELUENT<br />

EXTRACT ( )<br />

ELUENT<br />

EXTRACT ( )<br />

FEED ( )<br />

FEED ( )<br />

RAFFINATE ( )<br />

RAFFINATE ( )<br />

HMWP<br />

Insulin<br />

Zinc Chloride


SMB Experimental Validation : Ring I<br />

C/CF<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

(b) Effluent histories at the extract port<br />

HMWP exp.<br />

HMWP sim.<br />

Insulin exp.<br />

Insulin sim.<br />

ZnCl2 exp.<br />

ZnCl2 sim.<br />

C/CF<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

(a) Effluent histories at the raffinate port<br />

Insulin exp.<br />

Insulin sim.<br />

HMWP exp.<br />

HMWP sim.<br />

ZnCl2 exp.<br />

ZnCl2 sim.<br />

0<br />

0 10 20 30 40 50<br />

Switching Step<br />

0<br />

0 10 20 30 40 50<br />

Switching Step<br />

(c) Mid-step column pr<strong>of</strong>iles at cyclic steady state<br />

Desorbent Extract Feed Raffinate<br />

1.2<br />

0.9<br />

Zone I<br />

Zone II<br />

Zone III<br />

Zone IV<br />

Insulin exp.<br />

Insulin sim.<br />

C/CF<br />

0.6<br />

HMWP exp.<br />

HMWP sim.<br />

0.3<br />

ZnCl2 exp.<br />

ZnCl2 sim.<br />

0<br />

0 1 2 3 4 5 6 7 8 9 10<br />

Column number


SMB Experimental Validation: Ring II<br />

C/CF<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

(b) Effluent histories at the extract port<br />

0 20 40 60<br />

Switching Step<br />

Insulin exp.<br />

Insulin sim.<br />

ZnCl2 exp.<br />

ZnCl2 sim.<br />

(c) Mid-step column pr<strong>of</strong>iles at cyclic steady state<br />

Desorbent Extract Feed Raffinate<br />

1.2<br />

Zone I<br />

Zone II<br />

Zone III<br />

Zone IV<br />

C/CF<br />

0.9<br />

0.6<br />

0.3<br />

Insulin exp.<br />

Insulin sim.<br />

ZnCl2 exp.<br />

ZnCl2 sim.<br />

0<br />

0 1 2 3 4 5 6 7 8 9 10<br />

Column number


Parameter (5,000 kg insulin /<br />

year)<br />

Batch Xie (2002) Overall Optimized SMB<br />

Overall Yield (%) 89 99 99<br />

Column<br />

Configuration<br />

Ring A<br />

2-2-4-2 2-2-4-2<br />

12 in series<br />

Ring B 2-3-3-2 2-2-2-2<br />

Feed Flowrate (mL/min) 11 (each) 115 115<br />

Column Length (cm)<br />

Diameter (cm)<br />

Ring A<br />

13.7 9.34<br />

15<br />

Ring B 13.7 7.02<br />

Ring A 45<br />

47.9 27.2<br />

Ring B (12 units) 58.6 42.6<br />

Throughput (L/h/100 L BV) .23 1.19 5.36<br />

Solv. consumption (L/kg insulin) 150 42 55.6<br />

Equip. Cost ($/kg insulin) 35.71 29.76 29.17<br />

Solv. Cost ($/kg insulin) 15 4.20 5.56<br />

Resin Cost ($/kg insulin) 40.07 7.41 1.61<br />

Total Cost ($/kg) 90.78 41.50 36.35


SEC-SMB for Insulin Purification<br />

• Lab-scale tandem SMB designed and validated<br />

with experiments<br />

• 99% yield, 5X throughput, and 1/3 solvent<br />

consumption<br />

• Robust and optimal process-scale operation<br />

designed to increase throughput by 20X.<br />

• Novel strategies created to reduce protein<br />

residence time and maintain batch identity<br />

31


<strong>Design</strong> &<br />

Optimization<br />

(Mun et al., IEC Res 2003)<br />

Startup &<br />

Shutdown<br />

(Xie et al.,<br />

IEC Res 2003)<br />

Key Issues Solved<br />

Splitting<br />

Strategies<br />

(Hritzko et al., AICHE J 2002)<br />

Experimental<br />

Validation<br />

(Xie et al., Biotech Prog 2002)<br />

SMB<br />

High Purity<br />

High Yield<br />

Low Cost<br />

Robust<br />

Operation<br />

(Mun et al.,<br />

IEC Res 2003)<br />

Insulin<br />

Aggregation<br />

(Yu et al., JCIS 2006)<br />

Periodic<br />

Regeneration<br />

(Xie et al., WCCBME 2003)<br />

Residence<br />

Time<br />

(Mun et al.,<br />

AICHE J 2003)<br />

Versatile<br />

SMB<br />

(Chin & Wang,<br />

Sep Purif Rev 2004)<br />

Batch<br />

Identity<br />

(Mun et al.,<br />

Biotech Prog 2004)


Conclusions<br />

• Purdue SMB technologies<br />

–S<strong>of</strong>tware tools for design, simulation,<br />

and optimization.<br />

–Versatile SMB equipment for multicomponent<br />

separation.<br />

• Useful for developing high purity, high<br />

yield separations <strong>of</strong> chemicals,<br />

biochemicals, and pharmaceuticals.


Acknowledgements<br />

• Eli Lilly Co.<br />

• 21 st Century Indiana Research and Technology<br />

Fund (28)<br />

• Department <strong>of</strong> Energy (DE-FC36-01GO11071)<br />

• Purdue TTI-IRL (NSF IGERT 9987576)<br />

• SWD: Z. Ma, Y. Xie, K.B. Lee<br />

• VERSE: R. Whitely, Y. Xie<br />

• V-SMB equipment: C. Chin<br />

• Insulin Purification: Y. Xie, S. Mun, C. Chin

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!