Stirred Tank Scale-Up
An (ex-) Practitioners View
MemTide Process Development Course
London, 23-24 February 2012
Dr. Clemens Brechtelsbauer
Principal Teaching Fellow
Department of Chemical Engineering
Imperial College London
[email protected]
Contents
Further Reading
The Design Environment
Manufacturing steps
Unit operations
The design base
Mixing
Vessels, impellers and baffles
Flow regimes
Power draw
Scale-Up
Similarity concept
Scale-up rules
Multiphase systems
Dr. Clemens Brechtelsbauer Slide 2 Stirred Tank Scale-Up
Further Reading
Edward L. Paul, Victor A. Atiemo-Obeng, Suzanne M. Kresta (Eds.)
Handbook of Industrial Mixing
John Wiley & Sons, 2004
Ekato Ruehr- und Mischtechnik GmbH
Ekato Handbook of Mixing Technology
Ekato GmbH, 2000
John H. Atherton, Keith J. Carpenter
Process Development: Physicochemical Concepts
Oxford University Press, 1999
Dr. Clemens Brechtelsbauer Slide 3 Stirred Tank Scale-Up
Steps in API/Fine Chem Manufacture
Risk
Scale
Dr. Clemens Brechtelsbauer Slide 4 Stirred Tank Scale-Up
Unit Operations
Chemistry: Reactions and Separations Particle Forming Unit Operations
Reactions Separations Environmental Crystallisation Isolation Drying
Kinetics Distillation Mass Balance Scale-up Filtration Filter-
Multi- Extraction Recovery Centri- Dryers
phase fuges
Reaction
Extraction
Distillation
Crystallisation
Filtration
Drying
Dr. Clemens Brechtelsbauer Slide 5 Stirred Tank Scale-Up
The Nature of the Beast
Process = Chemistry + Physics
Equipment
Or
chemical rate constants are scale independent,
whereas physical parameters are not
John H. Atherton
Author of
Process Development: Physicochemical Concepts
Simply put:
Ye cannae change the
laws of physics, Jim!
Dr. Clemens Brechtelsbauer Slide 6 Stirred Tank Scale-Up
Standard Vessel Geometry
H=T B
=
T/10
D C=T/3
Dr. Clemens Brechtelsbauer Slide 7 Stirred Tank Scale-Up
Ships Wheel vs. Cartwheel
Low shear High shear
Good for solid suspension Good for dispersions
Dr. Clemens Brechtelsbauer Slide 8 Stirred Tank Scale-Up
Beware of the Dork Side
Minimum Stir and Mixing Volumes
1200
V(stir, min) / V(mix, min) [L]
1000
800
600
V stir min [L]
400
V mix min [L]
200
D
0
0 500 1000 1500 2000 2500 3000
Nominal vessel volume [L]
D
V static 1% V nominal Minimum stirred volume
V stir, min 5 % V nominal
does not mean
V dish 7% V nominal
V mix, min 30-40 % V nominal
Minimum mixed volume
Dr. Clemens Brechtelsbauer Slide 9 Stirred Tank Scale-Up
Impeller Types
Radial flow impeller
Discharges liquid radially outwards
towards vessel walls
Transitional & turbulent regime
Good for dispersing
Axial flow impeller
Discharges liquid axially towards base
or liquid surface depending on
rotation direction
Transitional & turbulent regime
Good for blending & suspending
Mixed flow impeller
Flow predominantly in axial direction
with also a radial component
Transitional & turbulent regime
Good for blending, suspending &
dispersing
Close clearance impeller
Ensures good motion near vessel walls
Laminar regime
Good for blending
Dr. Clemens Brechtelsbauer Slide 10 Stirred Tank Scale-Up
Understanding the Force
In a stirred vessel, the presence of baffles is essential to convert the vortexing motion of
the impeller into top to bottom mixing.
No baffles is worst. Partial baffling is better. Full baffling is best.
Baffles are needed to promote the flow pattern
characteristic of the impeller type
4 Baffles, typical width T/12 (US) or T/10 (EU)
with wall gap to prevent solids build-up
Beavertail and finger baffles are often used in
glass-lined vessels
Dr. Clemens Brechtelsbauer Slide 11 Stirred Tank Scale-Up
Power Draw
The power drawn by an
impeller is expressed through a
power number equation:
P = Po N3 D5
Power number
Po, or Newton number, Ne E1/V
Depends on A1, v1, h1, p1
Impeller type
Impeller and vessel DE/V
dimensions
Properties of the phases A2, v2, h2, p2
present E2/V
Must be measured!
Dr. Clemens Brechtelsbauer Slide 12 Stirred Tank Scale-Up
Flow Regimes
Different parts of a vessel can
experience different flow conditions
Assess by Reynolds number
N D2 Inertial Force
Re
Frictional Force
The power curve Po vs Re can be used
to evaluate the flow regime for the
whole process
Laminar: Re < 10
Po Re-1
Transitional: 10 < Re < 103
Po = f(Re)
Ekato Handbook of Mixing
Turbulent: Re > 103 Technology, p. 15
Po = constant
Dr. Clemens Brechtelsbauer Slide 13 Stirred Tank Scale-Up
Geometric Similarity
A single scale ratio, s, defines the relative magnitude of all
linear dimensions between the large and small scale:
D2 T2 H 2 C2
H1 s H2
D1 C1
D1 T1 H1 C1
T1 D2 C2
T2
Dr. Clemens Brechtelsbauer Slide 14 Stirred Tank Scale-Up
Kinematic & Dynamic Similarity
Kinematic Similarity Dynamic Similarity
Velocities at Ratio of forces
geometrically similar (dimensionless groups)
positions remain remain constant at
constant different scales
Constant tip speed Beware:
Constant superficial The relationship
gas velocity between process
performance and the
Constant maximum dimensionless group
liquid velocity in may not be linear!
impeller discharge
Dr. Clemens Brechtelsbauer Slide 15 Stirred Tank Scale-Up
Process Requirements
A process may be controlled by one or more of:
Liquid blending
reaction
homogenisation
Solid-liquid mixing
solid catalysed reaction
dispersion
Gas-liquid mixing
fermentation
hydrogenation
Dispersing immiscible liquid
reaction
emulsions
Heat transfer
Defining controlling duty is key to successful scale-up!
Dr. Clemens Brechtelsbauer Slide 16 Stirred Tank Scale-Up
Scale-Up Rules
Geometrically similar vessels
Turbulent regime
Process Rule Constant Parameter
Liquid blending Equal tip speed ND
Solid suspension Zwietering Njs D0.85
Solid distribution Equal energy input P/V
Gas-liquid Equal mass transfer P / V (= kLa)
Heat transfer Equal Re N D2
Fast reactions Equal mixing time N
Dr. Clemens Brechtelsbauer Slide 17 Stirred Tank Scale-Up
Scale-Up Decisions
(1) Constant Mixing Time
(2) Constant P/V
(3) Just suspension
(4) Tip speed
(5) Reynolds number
Note:
Criteria are mutually exclusive
On scale-up, impeller speed
goes down
Scale-up experiments and modelling essential!
Dr. Clemens Brechtelsbauer Slide 18 Stirred Tank Scale-Up
Future Multi-Purpose Plants
Mixing design by function
Principles A
Focused diversity by unit
operation B C
Multi-flight agitation
Vessel types
A: Reactor with work-up
B: Reactor / crystalliser
C: Crystalliser
D: Reactor / hydrogenator D
E: Hydrogenator
Physical properties determine
process design envelope E
Implemented at GSK
Dr. Clemens Brechtelsbauer Slide 19 Stirred Tank Scale-Up
Multiphase Systems Scale-Up
Scale-down Study
Continuous
Process Assessment Scale-up Projection
Improvement
Plant Supplies Campaign
Dr. Clemens Brechtelsbauer Slide 20 Stirred Tank Scale-Up
Process Assessment
Respiratory portfolio, final stage re-crystallisation
Important for process success:
Minimisation of variability on scale-up of particle forming step
Reproducible, narrow particle size distribution
Homogeneous growth conditions to promote particle uniformity
Low shear to prevent attrition
Initial hydro-dynamic simulation of flow profile by CFD for pilot and plant vessels
1500 L Pilot Plant 4000 L Manufacturing
Dr. Clemens Brechtelsbauer Slide 21 Stirred Tank Scale-Up
Scale-Down Study
To determine the effect of
shear & suspension on
particle size & distribution,
which cannot be predicted
through CFD
2 L conical base lab reactor
set up to scale-down
manufacturing reactor
Effect of improved suspension:
No effect on particle size
Effect of shear:
No effect on particle size or
distribution (PSD)
Dr. Clemens Brechtelsbauer Slide 22 Stirred Tank Scale-Up
Scale-Up Projection
Just suspension speed determined experimentally for 2L lab reactor
CFD multi-phase Lab, 2 L, 400 RPM Pilot Plant, 1500 L, 60-70 RPM
simulation based on
experimental stirrer
speed:
Plant operating impeller speeds recommended to maintain
homogeneous suspension
Stevenage pilot plant (1500 L): 60-70 RPM
prevent risk of settling
provide homogeneous growth conditions
verify negligible shear effect
Dr. Clemens Brechtelsbauer Slide 23 Stirred Tank Scale-Up
Plant Supplies Campaign
Analysis of lab and pilot plant Shear and energy input have
results by different scale-up no effect on particle size
parameters Increased suspension on scale-up:
solid suspension no effect on particle size
shear slightly narrower PSD
overall energy input
Penney Diagram
1000
Scale-up by constant
100 mixing time
[microns]
energy input
(P/V) l arge / (P/V) l ab
solid suspension
10 shear
heat transfer
0.1
0.01
1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04
Vplant / Vlab
Dr. Clemens Brechtelsbauer Slide 24 Stirred Tank Scale-Up
Process Assessment
Extrapolation to Manufacturing: Jurong V460 (4000 L), 48 RPM
Dr. Clemens Brechtelsbauer Slide 25 Stirred Tank Scale-Up
The Voice of Scale-Up Experience
Do Dont
Collaborate Add solids to a reaction
Log Evaporate to dryness
Sample Use all in & heat
Safety test & review Rely on critical timing
Use test Do hot filtrations
Keep it simple Risk all in one batch
McConville, F. X., CEP 103 (2007) 18-19
Dr. Clemens Brechtelsbauer Slide 26 Stirred Tank Scale-Up
As Clear as Mud?
"No one will believe you solved this problem in one day!
We've been working on it for months.
Now, go act busy for a few weeks and I'll let you know
when it's time to tell them."
(R&D supervisor, Minnesota Mining and Manufacturing/3M Corp.)
Dr. Clemens Brechtelsbauer Slide 27 Stirred Tank Scale-Up
