Reactive and hybrid separations of chemicals and … and hybrid separations-A... · of chemicals...

Universität Dortmund

Reactive and hybrid separations of chemicals and bioactive substances:

modelling and optimisation

Andrzej Górak Dortmund UniversityDepartment of Biochemical and Chemical Engineering,Chair of Fluid Separation Processeswww.bci.uni-dortmund.de/tv

ESCAPE 17, Bucureşti, May, 29th, 2007

Palaeontologystudy of prehistoric life forms on Earth through the examination of plant and animal fossils

Introduction

Palaeontologistresearcher working on...

Introduction

Adaptive Palaeontologist

DistillationAbsorptionExtraction

researcher working on...

ReactivedistillationReactive absorptionReactive extractionMembrane reactor

BioextractionMembrane adsorptionDownstreamProcessing

Distillation & membranes

……


Extraction & crystallization

Distillation & crystallization


Membranes

Reactive Separations Hybrid Separations

Bioseparations Hybrid Bioseparations

(Bio)reactive and Hybrid Separations

ReactivedistillationReactive absorptionReactive extraction

Membrane reactor



……





Membranes




physical models

process simulation

theorymathem. methods

model parameter

• hydrodynamics• mass transfer• thermodynamics• reaction kinetics

experiment

Determination

Interaction of Experiment and Theory

pilot plants, industrial processes

Optimisation

Validation

Reactive Separations: Modelling Concept

Apparatus Segmentedcolumn

Stage model

black-boxshort-cutequilibrium stagemodelrate-basedapproachCFD

Gas phase

Liquid phase

Gas phase

Liquid phase

Equilibrium Stage Model Rate-Based ModelSorel E.,Rectification de l‘Alcool (1893)

Taylor R., Krishna R.,Chem. Eng. Sci. 55 (2000)

gas and liquid phase in equilibriumefficiency factor to consider the influence of column internalsPacking: HETPTray: η (Tray efficiency)Reaction in bulk phase

only interface in equilibriuminterfacial area and mass transfer coefficient for different column internalsreaction in film and bulk phaseinfluence of reaction on mass transfer rate

Reactive Separations: Modelling Concept

Short-cut-methodsEquilibrium stage model

Rate-based approach Computational Fluid

Dynamics

Model complexity

Costs

Design costs

Costs of parameterdetermination

Costs due to errors

Total costs

Optimum?

Costs of Computer Aided Process Design

Synthesis of propyl propionate

Column diameter: 50 mm

Height of packing: 5.4 mRectifying section: 2.2 m Sulzer BXStripping section: 0.5 m Sulzer BXCatalytic section: 2.7 m Sulzer

Katapak SP 11

Esterification

Catalytic Distillation

Buchaly C., Kreis P., Górak, A.,Chemical Engineering and ProcessingSep (2007) accepted

ProAc

POH

Sulzer BX2.2 m

Katapak SP112.7 m

Sulzer BX 0.5 m

ProPro

POH / Water

Results from rate-based simulations!

χ POH/ProAc 2.0; RR 2.5; feed: 2.0 kg/h; distillate 0.88 kg/h

ProAc

POH

Sulzer BX2.2 m

Katapak SP112.7 m

Sulzer BX 0.5 m

ProPro

POH / Water


Katapak SP-11 (DN50)mcatalyst,dry = 0,205 kg/m

0,75

0,78

0,81

0,84

0,87

0,90

0,0 0,5 1,0 1,5 2,0 2,5mKatalysator,trocken [kg/m]

x i,S

umpf

[mol

/mol

]

0

0,05

0,1

0,15

0,2

0,25

x i,S

umpf

[mol

/mol

]

0,205


rate based model- validated reference model- mass transport limitation & reaction kinetic


0,75

0,78

0,81

0,84

0,87

0,90


x i,S

umpf

[mol

/mol

]

0

0,05

0,1

0,15

0,2

0,25

x i,S

umpf

[mol

/mol

]

ProProPOHProAc

0,205



0,75

0,78

0,81

0,84

0,87

0,90


x i,S

umpf

[mol

/mol

]

0

0,05

0,1

0,15

0,2

0,25

x i,S

umpf

[mol

/mol

]

ProProPOHProAc

0,205 rate based model- validated reference model- mass transport limitation & reaction kinetic

EQ model with reaction kinetic- no mass transport limitation- reaction rate << mass transport rate


EQ model with reaction equilibrium- most simple model (VLE, Keq)


0,75

0,78

0,81

0,84

0,87

0,90


x i,S

umpf

[mol

/mol

]

0

0,05

0,1

0,15

0,2

0,25

x i,S

umpf

[mol

/mol

]

ProProPOHProAc

0,205 rate based model- validated reference model- mass transport limitation & reaction kinetic

EQ model with reaction kinetics- no mass transfer limitation- reaction rate << mass transport rate

+ 10

%-4

%


Equilibrium stage model is a good first approximation

Sour gases:CO2, H2S, etc.

Aqueous amines:MEA, DEA, MDEA, etc.

Sweet gas

Loaded solvent

Raw gas

Offgas

Offgas

SolventA

bsor

ber

Des

orbe

r

Flas

h

Reactive Absorption

Rate based modelling of sour gas absorption with aqueous amine solutions

Kucka L., Müller, I, Kenig E.Y., Górak A.,Chem. Eng. Sci. 58 (2003)

CO2 + OH- HCO3-

HCO3- + H2O H3O+ + CO3

2-

2 H2O H3O+ + OH-MEAH + H3O+ MEAH2

+ + H2OMEAH + HCO3

- MEACOO- + H2O

CO2 + MEAH + H2O MEACOO- + H3O+

Kinetically controlled reactions:

Instantaneous reactions:

Complex system:• Parallel reactions• Consecutive reactions• Electrolytes

Studied system: MEAH-H2O and CO2

Reactive Absorption

0.1m

1.1m

Berl saddles 1/2“

MEA 14.7w%

Experiments: Tontiwachwuthikul et al., Chem. Eng. Sci. 47 (1992)

Air streamp≈1atmCO2 19.2%

0

1

2

3

4

5

6

15 25 35 45 55T [°C]

Col

umn

heig

ht [m

] exp.sim.

0

1

2

3

4

5

6

0,00 0,05 0,10 0,15 0,20yCO2 [-]

colu

mn

heig

ht [m

]

exp.sim.

Reactive Absorption

Rate-based approach is a MUST!

RateSep Presentation

From: Chau-Chyun C. et al., Annual AIChE Meeting, San Francisco (2006)

Dewatering of heterogeneously catalyzed batch esterification by pervaporation to shift equilibrium Experimental investigation of pilot scale membrane reactorProcess analysis of membrane reactorScale up of flat membrane modules (lab scale to pilot scale)

Permeate(Water)

Alcohol + Acid ↔ Ester + Water

Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Membrane reactor equipped with 0.5 m2

membrane area from Sulzer Chemtech

Membrane Reactor

Separation of liquid (PV) / vapour (VP) mixtures

application of „dense“ polymeric membranes or inorganic membranes

mass transport: sorption, diffusion and desorption ->SDM

driving force: difference in chemical potential

feed retentate

permeate(+ sweep gas)

pP«pU pP,i«pS,i

liquid

(sweep gas)

PVVP vapour

membrane module

= ⋅ Δj j ji i iN Q DF

active layer

MEMBRANE

feed

permeate

mi

flux = permeance x driving force

Membrane Reactor

= ⋅ Δj j ji i iN Q DF

active layer

MEMBRANE

feed

permeate

mi

flux = permeance x driving force

Membrane Reactor

Bulk

FEED

jix

Film

& jiN

& j jiL ,x

− −& j 1 j 1iL ,x

Liquidphase

Activelayer

MEMBRANE

j,aM,Fix

j,aM,Pix

j,Fix

Porouslayer

pM,jiy

Vapourphase

PERMEATE

Bulk

& j jiG ,y

− −& j 1 j 1iG ,y

jiy

Rate based modelling of membrane separation

Initial alcohol to acid ratio: 2.3 : 1

Batch reaction at 83°C in fixed bed side reactor with Amberlyst 46

propionic acid conversion: 83% after 6 hours

0%

10%

20%

30%

40%

50%

60%

70%

80%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

time [h]

mas

s fr

actio

n w

reac

[kg/

kg]

propyl-propionate_exp1-propanol_exppropionic acid_expwater_exppropyl-propionate_sim1-propanol_simpropionic acid_simwater_sim

batch reaction batch reaction + pervaporation

exp. vs. sim.

Membrane Reactor

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Initial alcohol to acid ratio: 2.3 : 1

Batch reaction at 83°C in fixed bed side reactor with Amberlyst 46

propionic acid conversion: 83% after 6 hours

Batch reaction at 83°C and pervaporation at 85°C

propionic acid conversion: 99% after 13.5 hours

0%

10%

20%

30%

40%

50%

60%

70%

80%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

time [h]

mas

s fr

actio

n w

reac

[kg/

kg]

propyl-propionate_exp1-propanol_exppropionic acid_expwater_exppropyl-propionate_sim1-propanol_simpropionic acid_simwater_sim

batch reaction batch reaction + pervaporation

exp. vs. sim.

Membrane Reactor

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP

Permeate(Water)


Retentate

PV

VP




……





Membranes







……

Distillation & membranesExtraction & crystallization



Membranes




Multicomponent reacting mixturesRecovering of educts with VP

Azeotropic binary mixturesSplitting of azeotropes with VP (Bio-Ethanol)

Non-ideal multicomponent mixtures

Selective removal of one component

Membrane Assisted (Reactive) Distillation

BioethanolBioethanol production

Fermentationbroth

lab & pilot plantexperiments for model validation

Detailed modelling, simulation & optimization

procedure

experiments in lab scale

determination of themodel parameter

simulation studies in industrial scale

membrane development

120 m²

optimal membrane

0,01 m²

Bioethanol productionIntegrated material and process development

Cou

rtesy

Whi

tefo

x Te

chn.

Optimization of the module geometry- fibre diameter- membrane area

Bioethanol production

constant volume of the fibre bundle

da,F

da,Rohrda,Bündel

lF

Membrane fibres

da,F

da,Rohrda,Bündel

lF

Membra

Based on rigorous modellingof membrane separation (VP)

Bulk

FEED

jix

Vapourphase

PERMEATE

BulkFilm

Liquidphase

Activelayer

MEMBRANE

j,aM,Fixj,aM,Pix

j,Fix

Porouslayer

Film

pM,jiy F,j

iy jiy

Separation task

Process sequences

Separation of a ternary isomeric mixture

Structural and operational degrees of freedom

Distillation and Crystallization

FD1

B1

M

D2

O

D-1K-2

D-2

K-1P

R1

R2

Is that the optimal structure?

0

0,2

0,4

0,6

0,8

1

1,2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19sequence

rela

tive

ener

gy d

eman

d Separation Task

Processsequences

Assessment withShortcut-models

Selection of sequence

F

D1

M

O

D-1

D-2K-1

P

R1

B2

sequence 15


Separation of a ternary isomeric mixture

Structural and operational degrees of freedom

Sequence 16 with lowest total costs

65 % costs savings by hybrid process design

00,1

0,2

0,30,40,5

0,6

0,70,8

0,91

rela

tive

annu

al c

osts

12 13 14 15 16 17 18sequence

Separation task

Optimizedprocess

Processsequences


Selection ofsequences

MINLPoptimisation

F

D1

M

O

D-1

D-2K-1

P

R1

B2B1

R1

K-2

sequence 16





……





Membranes




Samatou J., Wentink A.E., Rosa P., Azevedo A., Aires-Barros R., Bäcker W., Górak A., ESCAPE-17 (2007)Chromatography Membrane Adsorption Extraction

© Merck KgaA © Sartorius AG © Twente University

Technologies for separation of bioactive substances

Approach for Selective Purification of Antibodies

Purification problem:How to separate the target proteinfrom impurities

Membrane adsorbersAffinity membranesIon-exchange membranes

Development of detailed mathematical models for process design and reliable scale up fordifferent geometries

Courtesy Sartorius AG

Approach for Selective Purification of Antibodies

elutionwashingloading

0.0

0.5

1.0

1.5

2.0

2.5

0 20 40 60 80 100 120 140 160 180 200

Time [min]

C/C

0 [-]

experiment BAY A2Psimulation tool DOR / ACM

Integrated material and process development

Generic process for purification of bioactive substances

ConcentrationCell removal Capture Purification

PolishingFilling

Generic modelling approach

Model structure in process model


Downstream process

Process „Downstream 1"

UO Cap1

Membran-adsorber

Chromato-graphie

Ultra/Diafiltration

Capturing

… UO Capn

Membran-adsorber

Chromato-graphie

Ultra/Diafiltration

Purification

... ...

UO Pur1

Membran-adsorber

Chromato-graphie

Ultra/Diafiltration

UO Pur2

Membran-adsorber

Chromato-graphie

Ultra/Diafiltration

... ...

Polishing

UO Pol1

Membran-adsorber

Chromato-graphie

Ultra/Diafiltration

… UO Pol n

MembranAdsorber

Chromato-graphie

Ultra/Diafiltration

... ...

C. Frerick , P. Kreis, A. Górak, A. Tappe, D. Melzner, CEP (2007), accepted

Implemented models

Transport dispersivemodel

Convection, dispersion,radial diffusion

Transport dispersivemodel

Convection, dispersion,effective mass transfer

Resistance modelMembrane & gel

resistance

Short CutYields

Short CutYields

Short CutYields

Ideal stage modelConvection,

solid & liquid equilibrium

Ideal stage modelConvection,

solid & liquid equilibrium

General rate modelConvection, dispersion,

film & pore diffusion

Chromatography(IEX, SEC)

Membrane adsorber(IEX)

Filtration(UF, DF, NF)

Mod

elco

mpl

exity


Case study

Feed

Ultrafiltration

Diafiltration

CIEX Chrom.AIEX Chrom.

Process optimisation

DSP 1Chrom. / Chrom.

15 l PurHSA > 93 %YHSA > 75 %cHSA= 50 g/l

Product: HSA Separation task

Optimizedprocess

Processsequences



optimisationmodel parametersmass transfer models

adsorption isotherms (membrane adsorber / chromatography)

Case study

Feed

Ultrafiltration

Diafiltration


AIEX Chrom.

CIEX Memb.

DSP 2Chrom. / Memb.

Separation task

Optimizedprocess

Processsequences



optimisation

Product: HSA

Case study

Feed

Ultrafiltration

Diafiltration


AIEX Memb.

CIEX Chrom.

DSP 3Memb. / Chrom.

Separation task

Optimizedprocess

Processsequences



optimisation

Product: HSA

Case study

Feed

Ultrafiltration

Diafiltration


AIEX Memb. CIEX Memb.

DSP 4Memb. / Memb.

Separation task

Optimizedprocess

Processsequences



optimisation

Product: HSA

Parent 1

Parent 2

Parent 3

Parent 4

Parent 5

Parent 6

Parent 7

Starting population

110

120

130

140

150

160

170

180

190

200

0 20 40 60

Generation [-]

Fitn

ess

[€/g

]

Parent 1Parent 2Parent 3Parent 4Parent 5Parent 6Parent 7


Results

Benchmark DSP (Chrom. / Chrom.)

Parent 1

Parent 2

Parent 3

Parent 4

Parent 5

Parent 6

Parent 7

Startpopulation

110

120

130

140

150

160

170

180

190

200

0 20 40 60

Generation [-]

Fitn

ess

[€/g

]

Parent 1Parent 2Parent 3Parent 4Parent 5Parent 6Parent 7


Results

Optimierter DSP 1 (Chrom. / Chrom.)

Parent 1, …, 7

DSP 3

End population

Feed

Ultrafiltration

DiafiltrationProdukt: HSA

AIEX Memb.AIEX Memb.

CIEX Chrom.CIEX Chrom.

> 20% cost reduction

Short-cut-methodsModel complexity

EQ stage modelCFD

Costs

Design costs

Costs of parametdetermination

Costs due to errors

Total costs

er

Recommendations: Reactive absorption

• Rate-based approach is a MUST!• Stefan-Maxwell equations are usually not necessary• Film reaction has to be included in the model

Rate-based approach

Optimum

Summary


EQ stage modelCFD

Costs

Design costs


Costs due to errors

Total costs

Recommendations: Reactive distillation

• Rate-based approach recommended but sometimes not necessary• Equilibrium reaction is NOT sufficient

Rate-based approach

Optimum

Summary


EQ stage modelCFD

Costs

Design costs

Costs of parametdetermination

Costs due to errors

Rate-based approach

Total costs

er

Recommendations: Hybrid processes

• optimal modelling depths strongly depends on investigatedunit operations and chemical system

Optimum

Summary

Short-cut-methodsSequences variants

Model complexity

Equilibrium stage model

Generation

Assessment

Optimization

Summary

Recommendations: Hybrid processes


EQ stage modelCFD

Costs

Design costs


Costs due to errors

Total costs

Recommendations: Bio(reactive)separations

• rather simple models should be used since the knowledge on biothermodynamics and hydrodynamics is limited

Rate-based approach

Optimum

Summary

Short-cut-methodsEQ stage model

CFD

Model complexity

Simul.effort

Recommendations: Modelling & Simulation

New generation of simulation tools enable user to chose

the optimal modelling depth for each special separation task

Rate-based approach

Summary

Simulation tool

new user friendly simulation tools

New generation of Simulation tools

Outlook

Outlook

Chromatography

Membranes

Adsorption/liquidAdsorption/Gas

Ion exchange

Cristallisation

Extraction

Absorption

Supercritical exctraction

Górak 1995; Goedecke 2005

Models

Experiments

Technical maturity

DistillationTe

chni

calA

pplic

atio

nIndustrial

Production

FirstApplication

Invention OptimalProcess

Classical separation processes

Outlook


Tech

nica

lApp

licat

ion

Technical Maturity

IndustrialProduction

reactive absorption

reactive distillation

reactive cristallisation reactive extraction

reactive adsorption

reactive chromatography

membrane reactors

FirstApplication

Górak 2006

Modelle

Experimente

Reactive separation processes

Tech

nica

lApp

licat

ion

FirstApplication

Technical Maturity

IndustrialProduction

separations

reactiveseparations

Bio(reactive)separations


Experiments

Models

Bio(reactive) and Hybrid Separations

Outlook

Design of a new class of interactive materials

Integrated material und process development for• monoclonal antibody purification (adsorbens/membrane adsorbers, solvents/bioextraction)

• bioethanol production (membranes/distillation)

Integrated

Incremental

Outlook

Outlook

“The innovative character of Process Intensification is in nice harmony with the objectives of Process Systems Engineering: a symbiosis between them has high potential”Moulijn, Stankiewicz,Grievink,Górak: ESCAPE 16 (2006)

Reactive and hybrid (bio)separations belong to theemerging area of Process IntensificationCharpentier, Ind.Eng.Chem.Res., (2007), 46, 4365




……





Membranes



Outlook

“The innovative character of Process Intensification is in nice harmony with the objectives of Process Systems Engineering: a symbiosis between them has high potential”Moulijn, Stankiewicz,Grievink,Górak: ESCAPE 16 (2006)

Reactive and hybrid (bio)separations belong to theemerging area of Process IntensificationCharpentier, Ind.Eng.Chem.Res., (2007), 46, 4365

The devil is in the details

•Reaction kinetics•Thermodynamics of electrolytes•Charged biomolecules•Aggregates and emulsions•Hydrodynamics of viscous system•Many more….

„No more of the old formula: Let’s make it a foot bigger in diameter and 5 ft. higher just for good luck“

- Walter G. Whitman (1924)

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=The

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=devil

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=is

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=in

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=the

http://dict.leo.org/ende?lp=ende&p=/gQPU.&search=details

Acknowledgements

Reactive and hybrid separations of chemicals and … and hybrid separations-A... · of chemicals...

Documents

Transcript of Reactive and hybrid separations of chemicals and … and hybrid separations-A... · of chemicals...