4th International Conference on Process Analytical Technologies in Organic Process R&D Brussels 2009

PAT and Process Control: Hybrid Real-Time
Technologies for Enhanced Chemical
Development

Dominique Hebrault
Sr. Technology & Application
Consultant

Brussels, March 17-18, 2009

Presentation Outline

 Introduction

- PAT
- Process Control
 Case Studies

- Real Time In Situ Reaction Monitoring with ReactIR™
- Kinetics, Scale-up, and Process Safety with RC1e, and ReactIR™
- Crystallization with FBRM®, PVM® and ReactIR™
- Experiment Design, Data Acquisition, Analysis with Enhanced
Software Tools

1

Introduction

Organic Synthesis Lab at
the turn of the century…

…which century?

2

Introduction
FDA’s View of Process Analytical Technologies

 Process Analytical Technology (PAT)
- A system for designing, analyzing, and controlling manufacturing
- Through timely measurements of critical quality and performance
attributes of raw and in-process materials and processes

- With the goal of ensuring final product quality
 PAT Fundamental Tenets
- Quality cannot be tested into the product; it should be built-in or should
be by design

 PAT Goals
- Enhance understanding and control of processes

3

Introduction

 PAT tools can be categorized as:
- Process analyzers
- Process control tools
- Multivariate tools for design, data acquisition and analysis
- Continuous improvement and knowledge management tools
 PAT tools are used:
- Process development
 Process monitoring to develop mechanistic understanding
 Statistical DOE and model building to enhance process
understanding
 Use of risk analysis in establishment of design space
- Manufacturing

4

Introduction
API Development Scale up Production API
Understanding Optimization Scale-up Production
ReactIR™ iC10 ReactIR™ 45 MonARC

FBRM® in the lab PVM® in the lab FBRM® in the plant

RTCal™ for real-time
reaction calorimetry at
lab scale

Introduction
 Poor temperature control
– Side reactions, slow kinetics
– Supersaturation control issues → broad distribution, impurity, polymorph
 Manual addition
– High reagent concentration → by-products
– Supersaturation spikes → oiling out
 Poor mixing
– Slow reaction
– Concentration gradient→ side-reaction
– Solid breakage, attrition

 Reduce risk of experimental error
 Reproducibility, traceability, data
logging, modeling


 Introduction

- PAT
- Process Control
 Case Studies

Software Tools

7

Case Study: FTIR, PAT Tool in Pharma Development
Development of a Safe and Scalable
Oxidation Process for the Preparation of
6-Hydroxybuspirone

 Introduction

Active metabolite of Buspirone,
manufactured and marketed as Buspar,
employed for the treatment of anxiety
disorders and depression

Multi Kg amount needed for clinical dev.

Process lack of ruggedness and
unreliable product quality

Source: Daniel J. Watson,* Eric D. Dowdy, Jeffrey S. DePue, Atul S. Kotnis, Simon Leung, and Brian C. O’Reilly, Bristol-Myers Squibb
Pharmaceutical Research Institute, NJ, USA, Organic Process Research and Development, 2004, 8, 616-623; Mettler Toledo Real Time Analytics
Users’ Forum 2004, London, UK


 Challenges KHMDS

Monitor deprotonation of 1 for:
- More precise determination of endpoint
to minimize bis-deprotonation
- Allow for variations in the base titer,
water content, and phosphite quality

 Observations 1627cm-1

- Deprotonation complete within 5’
1677cm-1
- Enolate anion 3 stable at -25⁰C for 12h
- Addition of P(OEt)3 before addition of
the base → no impact on IR signal
- Kinetics of enolate degradation



 Challenges KHMDS

Monitor deprotonation of 1 for:
- More precise determination of endpoint
to minimize bis-deprotonation
- Allow for variations in the base titer,
water content, and phosphite quality

 Observations
- Deprotonation complete within 5’
- Enolate anion 3 stable at -25⁰C for 12h
- Addition of P(OEt)3 before addition of
the base → no impact on IR signal
- Kinetics of enolate degradation


 Improved process KHMDS

- Charged the base to 1 until complete
consumption → Stable signal (1677cm-1)

- 1 / THF charged back to the vessel
until the signal increased → 1-3%
excess of the starting material
(quantified FTIR)

- Result: Impurity 8 is minimized



 Conclusion
- ReactIR™ allowed titration of the
correct amount of base, prevented
accidental overcharge due to
ambiguous concentration

- Implementation to the pilot plant (13Kg)
- 69% yield and >99 area %, need for
recrystallization eliminated Buspirone enolate

- Robust, superior process &
crystallization thanks to the successful Buspirone

use of PAT

Case Study: FTIR as PAT Tool for Continuous Process
Development and Scale-up of Three
Consecutive Continuous Reactions for
Production of 6-Hydroxybuspirone

 Introduction

Control base / buspirone stoichiometry is
critical to product quality

Optimization based on offline analysis is
time consuming and wasteful

Actual feed rate adjusted based on the
feedback from inline FTIR: Flow cell and
ReactIR™ DiComp probe

Source: Thomas L. LaPorte,* Mourad Hamedi, Jeffrey S. DePue, Lifen Shen, Daniel Watson, and Daniel Hsieh, Bristol-Myers Squibb
Pharmaceutical Research Institute, NJ, USA, Organic Process Research and Development, 2008, 12, 956-966; Mettler Toledo Real Time
Analytics Users’ Forum 2005 - New York

Case Study: FTIR as PAT tool for Continuous Process

KHMDS
 Implemented startup strategy

- Start with slight undercharge of base
(feed rate) to reduce diol 8

- Flow rate increased at 1% increments
until no decrease of Buspirone 1 signal
is observed

- Base feed rate was reduced 1-3%

- Works well because enolization fast,
equilibrium reached within minutes


Case Study: FTIR as PAT Tool for Continuous Process
 Outcome
- Ensure product quality via proper ratio
and base feed rate
- Minimize waste of starting material
- Faster reach of steady state via real-
time detection of phase transitions
- FTIR also used for enolization
monitoring during steady state

 Scale-up
- Lab reactor: Over 40 hours at steady
state
- Pilot-plant reactor: Successful
implementation (3-batch, 47kg/batch)


 Introduction

- PAT
- Process Control
 Case Studies

Software Tools

16

Case Study: Calo for Reaction Kinetics Screening
An Integrated Approach Combining Type A: Very fast, t1/2< 1 s, controlled by
Reaction Engineering and Design of mixing
Experiments for Optimizing Reactions

 Introduction Type B: Rapid, 1 s < t1/2< 10 min, mostly
kinetically controlled
Early phase RC1e experiments to obtain
a basic understanding of: Type C: Slow, t1/2 > 10 min, safety issue
in a batch mode
- Enthalpy
- Kinetics
- Mass Balance
- Type of phases
50% of reactions in the
fine/pharmaceutical industry could
benefit from a continuous process
(microreactors)
Source: D.M. Roberge, Department of Process Research, Lonza, Switzerland, Organic Process Research and Development, 2004, 8, 1049-1053;
Mettler Toledo 15th International Process Development Conference 2008, Annapolis, USA; Chem. Eng. Tech., 2005, 28, No. 3, 318-323


RC1e allows precise measurement of Type A: Very fast, t1/2< 1 s
controlled by mixing
reaction enthalpy

Instantaneous reaction heat is related to

reaction rate

 Results: Very fast reaction

- No heat accumulation

- Dosing controlled
C=C double bond oxidized / cleaved by
aqueous NaOCl catalyzed by Ru

Source: D.M. Roberge, Organic Process Research and Development, 2004, 8, 1049-1053; Mettler Toledo 15th International Process Development
Conference 2008, Annapolis, USA; Chem. Eng. Tech., 2005, 28, No. 3, 318-323


Type B: Rapid, 1 s < t1/2< 10 min, mostly
 Results: Rapid reaction kinetically controlled

- Heat signal function of dosing rate

- Reagent accumulates and reacts
after the end of the dosage

- Lower temperatures favor high
accumulation

- Higher temperatures favor formation
of side products
Quench of ozonolysis into methanol /
dimethyl sulphide



 Results: Slow reaction Type C: Slow, t1/2 > 10 min, safety
issue in a batch mode
- Accumulation of energy > 70%
- Most of the heat potential evolves
after the end of addition

- Typically initiated by temperature
increase or catalyst addition

- Autocatalytic reaction and / or
induction period

 Conclusion
Real time RC1e calorimetry also for early Knoevenagel-type reaction catalyzed by NaOH:
on kinetics and safety assessment intramolecular aromatic ring condensation


Case Study: Integrated PAT for Industrial Scale-Up
Thorough Examination of a Wittig-Horner
Reaction Using Reaction Calorimetry
(RC-1), LabMax®, and ReactIR™

 Introduction
Process not ready for industrial
development: Lack of robustness due to
poor understanding of water effect, base
form, kinetics, and thermo-dynamics

- RC1™ used for kinetic and heat
information Side-reaction: Benzyl phosphonate hydrolysis

- ReactIR™ and LabMax® used for
quantitative kinetic simulation under
well controlled conditions

Source: Michael Grabarnick and Sharona Zamir*, Makhteshim Chemical Works Ltd., Israel, Organic Process Research and Development, 2003, 7,
237-243, Mettler Toledo 2001 RXE User Forum


 Results
Heat flow from RC1™ used as a real-time
monitoring technique

Initial RC1™ and DOE study results
showed reaction is fast and yield
sensitive to base addition



- Eahydrolysis > Eastilbene_formation → Validation Run
temperature↓

- Stilbene formation more sensitive to
H2O than hydrolysis → [H2O] ↓. Impact
on reaction rate constant

- Stilbene formation 2nd order versus
[BA] → [BA] ↑

- Stilbene formation 1st order versus
[KOH]

- Validation experiment under improved
conditions in RC1™


Process simulation

- BA/stilbene concentration

- Plant reactor temperature (Cp, heat
data from RC1)

Validation of the simulation process with
ReactIR™ and LabMax® - Reactants dissolution at 50⁰C
- Tj: -10⁰C
- Real-time concentration data, under - Once 30 < Tr < 40⁰C, aq. KOH added
well controlled scaled-down conditions

- Comparison to simulated profiles: good
fit, confirmed 94% yield

- Model tested: 8 m3 production reactor



 Conclusion
Use of real-time analytics (FTIR, heat)
and modeling to study the mechanism of
a Wittig-Horner reaction via a thorough
kinetic and thermodynamic research

Improved large scale conditions were
obtained

Preparation of mathematical model to
plan industrial equipment: Need for a
more effective heat exchanger for yield
improvement


 Introduction

- PAT
- Process Control
 Case Studies

Software Tools

26

FBRM® PVM®

FBRM® Technology
PVM® Technology
Focused Beam Reflectance Measurement Particle Video Microscope
Track real-time changes in particles and Microscope quality images, in-process and in
droplets as they naturally exist in the process real-time
Characterize particle systems from 0.5μm to 3mm Characterize particle systems from 2μm to 1mm

27

Case Study: Integrated PAT for Crystallization
Process Design and Scale-Up
Elements for Solvent Mediated
Polymorphic Controlled Tecastemizole
Crystallization

 Introduction
Tecastemizole: Active metabolite of
histamine H1-receptor antagonist
Astemizole, 2 polymorphic forms A
(stable) and B
800L scale process history shows high
risk of not obtaining desirable polymorph:
Seed controlled to solvent mediated
interconversion
Raman spectroscopy and DSC
unsuitable for interconversion monitoring

Source: Kostas Saranteas,* Roger Bakale, Yaping Hong, Hoa Luong, Reza Foroughi, and Stephen Wald Chemistry and Pharma Sciences,
Sepracor Inc., MA, USA, Organic Process Research and Development, 2005, 9, 911-922, Mettler Toledo 2001 Lasentec® Users' Forum


 Challenges
Water added
Methodical study under well
controlled conditions
IR Peak Area, 1513 cm-1
- Need for computer control of
batch temperature, agitation
rate, and dose control of the
antisolvent addition (LabMax® ) Tr
Particle # / sec
- Real time supersaturation Seeding
determination (ReactIR™)

- In situ particle count and size
measurements (FBRM®)  Results
Interconversion rate influenced by
temperature, mixing, B agglomerate size,
initial seed composition


- Significant effect on interconversion - Interconversion rate-limiting step is
rate of hold and cooling temperature growth of form A
profile following addition of water

- Nonlinear cooling profile takes advantage of the temperature rate
effect on form interconversion (4 h above 70⁰C)

 Summary and conclusion
Scale-down version of the crystallization
step performed at lab scale under well
controlled conditions with LabMax®,
FBRM®, and ReactIR™

Tr

New profile
After better crystallization understanding,
Old profile
and optimization, scale-up successfully
validated at both pilot (1200L) and full-
scale manufacturing (6000L)

Time


 Introduction

- PAT
- Process Control
 Case Studies

Software Tools

32

Software for Design, Data Acquisition and Analysis
 Process Analyzers  Multivariate analysis
– Control lab reactor based on trend data – ConcIRT™ live algorithm:

– Live drag/drop data exchange with reactor Converts in situ FTIR/Raman
data into concentration profiles

– iC Quant™ determines
component concentrations in
an unknown mixture

 Automated Lab Reactor  Data to information software tools

– Initiate and control PAT experiments – iC SafetyTM converts reaction
calorimetry data into process
– Live drag/drop data exchange with reactor
safety information

Summary
Process chemistry challenges: ReactIR™, calorimetry and automated reactors
- Did the reaction work?
- Understand selectivity and reactivity
- Identify intermediates or by-products
- How long did it take?
- Endpoint, initiation-point, stall-point
- Can this process be scaled-up?
- Identify key control parameters
- Understand reaction kinetics
- Will it be safe?
- Measure reaction heat/enthalpy
- Determine heat capacity, heat
transfer coefficient
- Worst case scenario estimation
- Thermal accumulation and
conversion

Summary
Crystallization development: ReactIR™, FBRM®, PVM®, automated reactors

- Do the particles have the right
dimensions, distribution,
morphology?
- Do product scale-up consistently
meet specifications? Does it
require rework?
- How is filtration rate? How about
drying time? Is it consistent?
- Measure solubility and screen MSZ
- Understand, monitor, and control
supersaturation
- Track nucleation and growth kinetics of
crystallization
- Identify and control critical parameters
- Scale-down experiments in the lab

4th International Conference on Process Analytical Technologies in Organic Process R&D Brussels 2009

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (20)

En vedette

En vedette (20)

Similaire à 4th International Conference on Process Analytical Technologies in Organic Process R&D Brussels 2009

Similaire à 4th International Conference on Process Analytical Technologies in Organic Process R&D Brussels 2009 (20)

Plus de dominev

Plus de dominev (6)

4th International Conference on Process Analytical Technologies in Organic Process R&D Brussels 2009