This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
An Equipment Selection Methodology for Continuous Manufacturing of Small-Molecule Drugs by
Kevin Peng B.S. Chemical Engineering Cornell University, 2009
Submitted to the MIT Sloan School of Management and the
Department of Chemical Engineering in Partial Fulfillment of the Requirements for the Degrees of
Master of Business Administration
and Master of Science in Chemical Engineering
In conjunction with the Leaders for Global Operations Program at the
MIT Sloan School of Management, Department of Chemical Engineering May 12, 2017
Certified by __________________________________________________________________________
Richard D. Braatz, Thesis Supervisor Edwin R. Gilliland Professor of Chemical Engineering
Certified by __________________________________________________________________________
Roy E. Welsch, Thesis Supervisor Eastman Kodak Leaders for Global Operations Professor of Management
Accepted by __________________________________________________________________________
Maura Herson, Director of MIT Sloan MBA Program MIT Sloan School of Management
Accepted by __________________________________________________________________________
Daniel Blankschtein, Herman P. Meissner Professor of Chemical Engineering, Graduate Officer Department of Chemical Engineering
2
This page has been intentionally left blank.
3
An Equipment Selection Methodology for Continuous Manufacturing of Small-Molecule Drugs by
Kevin Peng
Submitted to the MIT Sloan School of Management and the Department of Chemical Engineering on May 12, 2017, in Partial Fulfillment of the Requirements for the Degrees of
Master of Business Administration and Master of Science in Chemical Engineering
Abstract
Flexible, modular, continuous manufacturing small-scale plants (MCSPs) for small-molecule drugs have been recognized as potential safe and economical solutions for pharmaceutical manufacturing. However, among the variety of equipment technologies required for an MCSP platform, there are only a few technologies that have publicly available methodologies for equipment selection. In this study, a new method and tool for computer-assisted equipment selection was developed, which use key engineering correlations and design criteria to match off-the-shelf equipment with the synthesis processes of interest. Furthermore, the tool allows simultaneous equipment selection for multiple synthesis processes to allow the identification of the most flexible MCSP assets. The long-term goal of this tool is to encompass the entire span of technologies that could be used in an MCSP skid and to serve as a communal storage location for vendor-available equipment information to facilitate collaboration and design of a mainstream continuous manufacturing (CM) system. This methodology was applied to equipment selection for the continuous manufacturing of an actual Amgen small-molecule drug substance (API) as a case study. The results from this study showed that the new tool can improve the speed at which equipment is selected and can aid the process developer in decision-making for choosing the most suitable CM asset. Thesis Supervisor: Richard D. Braatz Edwin R. Gilliland Professor of Chemical Engineering Department of Chemical Engineering Thesis Supervisor: Roy E. Welsch Eastman Kodak Leaders for Global Operations Professor of Management MIT Sloan School of Management
4
This page has been intentionally left blank.
5
Acknowledgements
To my academic advisors, Professors Richard D. Braatz and Roy E. Welsch, thank you for your guidance and contributions to this work. To my industry advisor, Dr. Roger Hart, and colleagues at Amgen, thank you for your mentorship, enthusiasm, and efforts to ensure that the research and internship was a high priority and a success. To the Leaders for Global Operations (LGO) Program and Professor Richard D. Braatz, thank you for granting me this opportunity for professional and personal growth at MIT. To my professors, classmates, and roommates at MIT, thank you for providing such a rich and diverse learning environment. To my parents, my brother and his wife, my uncle, my aunt, my grandmother, and my in-laws, thank you for your endless encouragement and motivation. To our new baby Charlotte, for bringing so much joy to us during our time at MIT. Most importantly, to my wife, thank you for your unconditional support and patience.
1.1 Current Status of Continuous Manufacturing ..................................................................... 111.2 Project Drivers: Needs for Novel Equipment Selection Strategies .................................... 121.3 Problem Statement .............................................................................................................. 151.4 Amgen, Inc. ......................................................................................................................... 161.5 Project Focus ....................................................................................................................... 201.6 Project Goals ....................................................................................................................... 211.7 Thesis Overview ................................................................................................................. 21
Chapter 2: Pharmaceutical Manufacturing: Overview ................................................................. 222.1 Manufacturing Process Overview ....................................................................................... 222.2 Conventional (Batch) Manufacturing of Drug Substances ................................................. 232.3 Continuous Manufacturing of Drug Substances ................................................................. 232.4 CM of Drug Substance: Equipment Technology Overview ............................................... 25
Chapter 3: Hypothesis ................................................................................................................... 29Chapter 4: Problem Analysis ........................................................................................................ 29
4.1 State of the CM Equipment Selection Process prior to Project Implementation ................ 304.2 Improvements to the Current Equipment Selection Process ............................................... 31
Chapter 5: Redesign of the Equipment Selection Process ............................................................ 325.1 Overview of Software Tool Design .................................................................................... 325.2 Equipment Selection: General Design Requirements ......................................................... 345.3 Equipment Selection: CSTR-specific Design Requirements .............................................. 365.4 Equipment Selection: PFR-specific Design Requirements ................................................. 485.5 Equipment Selection: Packed Bed Reactor-specific Design Criteria ................................. 525.6. Equipment Selection: Mixed-Suspension Mixed-Product Removal Crystallizers ............ 565.7. Equipment Selection: Motionless Mixer-specific Design Criteria .................................... 585.8 Equipment Selection for Multiple Processes ...................................................................... 595.9 Equipment Selection: Other Equipment Information ......................................................... 605.10 Equipment Selection: Feature-based Analysis .................................................................. 60
Chapter 6: Case Study Analysis .................................................................................................... 616.1 Case Study Methodology .................................................................................................... 616.2 Results: New Tool vs. Original Equipment Selection Process ........................................... 62
Appendix ....................................................................................................................................... 70A.1: Summary of CSTR Selection Inputs and Calculations ..................................................... 70A.2: Summary of PFR Selection Inputs and Calculations ........................................................ 74A.3: Summary of PBR Selection Inputs and Calculations ....................................................... 78A.4: Summary of MSMPR Selection Inputs and Calculations ................................................. 81A.5: Summary of Motionless Mixer Selection Inputs and Calculations .................................. 85
Pharmaceutical companies have traditionally used batch manufacturing processes for drug
production since the industry’s inception. Batch manufacturing processes have associated
inefficiencies, which are estimated to cost the industry $50 billion annually [14]. In response,
pharmaceutical companies are developing continuous manufacturing (CM) processes for
synthesizing small-molecule drugs to reduce their manufacturing footprint, improve safety, and
adopt more efficient chemistries that are not cost effective or safe to use in batch processes [5].
Since 2015, significant progress has been made by both regulatory agencies and pharmaceutical
companies in implementing continuous manufacturing (CM) processes for the development and
commercial production of pharmaceutical drugs [10]. For example, in 2015 the FDA published
its draft guidance for the advancement of emerging technologies to encourage the modernization
of pharmaceutical manufacturing through technologies such as CM [11], and large
pharmaceutical companies such has Eli Lilly, Janssen, Novartis, and Vertex Pharmaceuticals
have all developed commercial-size continuous processes for drug substance and/or drug product
production.
Although there is increasing use of commercial CM facilities, these facilities have been costly,
due to heavy customization and the limitation of only producing one drug substance or one drug
product per facility. To reduce the costs associated with customized CM facilities and redundant
setups, modular continuous manufacturing platforms for small-molecule drugs (MCSPs) using
off-the-shelf equipment are being developed in both academia and industry as a potential
12
solution. In industry, initiatives such as the European FP7 Research Project are trying to
introduce flexibility and modularization to continuous manufacturing with the goal of identifying
a “fast, flexible, future” (“F3”) factory [3]. In academia, the MIT “Pharmacy on Demand”
project, funded by DARPA, investigates the possibility of producing over 16 drug substances
using one CM setup [12]. Though significant progress has been made for CM advancement to
become mainstream, there are still very few, if any, commercial-sized MCSP platforms for
manufacturing multiple drug substances and products.
1.2 Project Drivers: Needs for Novel Equipment Selection Strategies
While there is interest for MCSPs, a major challenge for designing these platforms is equipment
selection. Figure 1 depicts the major elements of the product lifecycle for the pharmaceutical
industry. Among these four general phases in the life cycle, the two steps in which equipment
selection is most important are: process development and commercial production of the drug
substance and product.
Figure 1: Pharmaceutical Product Lifecycle
For equipment selection, there are two important steps: (1) the selection of equipment technology
and (2) the selection of the actual make and model of the equipment. The selection of the proper
equipment technology is essential in the process development step in the drug development
lifecycle, because the technology is typically kept the same from lab to commercial production, it
is only scaled up with each step [15]. The selection of the right makes and models is essential
R&D(MoleculeDiscovery)
PD(Process
DevelopmentandCharacterization)
Manufacturing(CommercialProduction)
MarketingandSales
13
for process development so that the assets can provide the right chemical reactions, facilitate data
collection, and can be reused as a manufacturing platform for various synthesis steps and drug
types.
For commercial production, specific equipment selection is critical to ensure the satisfaction of
market demand, process efficiency, safety, and reliability.
In 2013, Commenge et al. [7] introduced a fast and effective method that covers the first step of
equipment selection – determining the most suitable equipment technologies based on specific
process chemistries and process limitations. Although their method filters for the appropriate
types of equipment, their method does not include the selection of real, usable equipment makes
and models for a specific synthesis process (Fig. 2).
In 2014, Krasberg et al. [6] proposed a computer-assisted method to cover the second step of
identifying appropriate off-the-shelf Plug Flow Reactors (PFRs) for a process based on: (1) a set
quantitative process inputs that characterizes the specific synthesis chemistry, (2) the production
scale, and (3) a set of quantitative selection criteria based on first principles of reaction. While
the method is rigorous and effective for determining the correct PFR model for a continuous
manufacturing process, there has been no follow-up to expand this approach for selecting makes
and models for other types of equipment technologies that are critical to drug production, such as
continuous stirred tank reactors (CSTRs), static mixers, mixed suspension, and mixed product
crystallizers (MSMPRs). Figure 2 highlights the need for additional identification methods for
specific, usable equipment models.
14
Figure 2: Equipment Selection Process Reveals Need for Make and Model Identification
Moreover, neither of the proposed equipment selection methods suggest a way to choose
equipment simultaneously for multiple synthesis processes. Thus, the development of a
standardized manufacturing platform (MSCP) can be time-consuming, inefficient, and iterative.
From a process development perspective, choosing equipment for each individual process can
also be costly because it can lead to the purchase of equipment assets that are redundant or rarely
reused.
The business driver for faster process development and faster CM equipment selection is clear.
In addition to reducing the $50 billion annual waste associated with batch manufacturing, CM
equipment platforms need to be developed as efficiently as possible because the economics
behind drug development provide strong incentive for speed and efficiency. Research indicates
that first-to-market drugs with over $100M annual sales have an average of 6% greater market
share than its competitors [1], which is a primary incentive for pharmaceutical companies to
accelerate all aspects of the drug development process for each of their pipeline drugs.
ProposedSynthesis
ProcessLimitIdentification
OperationandIntensification
StrategyFormulation
EquipmentTechnologyIdentification
EquipmentMakeandModel
Identification(?)
CoveredbyCommengeetal.[7]
OnlyPFRscoveredbyKrasbergetal.
[6]
15
Furthermore, studies have shown that annual expenditures for Phase I, II, and III clinical trials
cost $17M, $34M, and $27M per year of trial, respectively [2]. These numbers imply that there
are significant cost savings if scientists can choose equipment for manufacturing its drug
substances more quickly.
1.3 Problem Statement
While Commenge et al. [7] has introduced a well-developed process for identifying the correct
equipment technologies, there is clearly a business need to develop a tool that makes the
selection of makes and models of equipment more efficient. One approach is to build on
Krasberg’s research [6] for selecting makes and models for other equipment technologies, in
addition to PFRs. Expanding on the Krasberg research would allow specific equipment to be
selected for an entire continuous manufacturing process. Furthermore, there needs to be a
method to more-easily identify flexible CM equipment assets that can be reused for multiple
syntheses in both a process development setting and a commercial CM setting so that
manufacturing platforms can be easily developed.
Concisely, the industry needs an equipment selection tool for essential equipment
technologies that (1) chooses equipment based on first-principles design criteria and key
process inputs [6] and (2) allows the identification of the equipment models that are
appropriate for a portfolio of drug syntheses.
16
1.4 Amgen, Inc.
1.4.1 Manufacturing Strategy
As publicly stated in Amgen’s R&D Strategy,1 Amgen is a modality-independent company “with
a focus on biologics while ensuring world-class small-molecule capabilities.” Amgen’s small-
molecule products include Corlanor®, Sensipar®, and Kyprolis® with six additional small-
molecule candidates across all phases of its pipeline.2 Furthermore, since its acquisition of Onyx
Pharmaceuticals in 2013, Amgen has plans to build a facility in Singapore for commercial
manufacturing of the drug substance for Kyprolis® [13].
As a strategy to maintain “world-class small-molecule capabilities” and strength in
manufacturing, Amgen has released plans to pursue technology to support Manufacturing of the
Future (MoF) using standardized, modular, flexible manufacturing facilities with the goals of
reducing costs associated with waste and facility footprint.3
In other words, Amgen’s manufacturing goals are consistent with those in the industry,
regulatory agencies, and academia. Thus, Amgen also has a vested interest in the advancement
of MCSPs and a continuous manufacturing platform that can be used to synthesize their portfolio
of small-molecule drugs.
1. As of 9/16/16 (www.amgen.com/science/research-and-development-strategy/) 2. As of 9/16/16 (www.amgenpipeline.com) 3. Amgen Transforming Biotech Manufacturing Infographic, October 11, 2015
17
1.4.2 Stakeholder Analysis: The Value of the MIT LGO Partnership
Clearly, a selection tool for identifying production equipment aligns with Amgen’s strategy to
develop a continuous manufacturing platform. While this project is a high priority, significant
resources are required including: (1) expertise in continuous manufacturing technology, (2)
expertise in integrating new tools into business processes, and (3) a project manager to combine
the contributions from all stakeholders and ensure timely progress.
While Amgen has significant internal capabilities to develop this tool, the project benefits greatly
from external collaborations, such as with faculty at MIT who can provide additional
perspectives and information on CM technology, selection strategy, and integration of the tool
into the Process Development workflow. The stakeholder analysis (Figure 3) indicates that an
MIT Leaders for Global Operations (LGO) Research Internship is ideal for combining the
technological and process expertise of Amgen and MIT.
More specifically, the LGO Intern can frequently interact with its Group 1, 2, and 3 stakeholders
(Figure 3) to develop a tool that will be value-adding and frequently used and updated. Group 1
stakeholders are Amgen’s internal customers from the Process Development organization. These
stakeholders will be the primary users of the tool but are also the technology experts for CM at
Amgen that can provide detailed feedback on how to abstract the information that they need into
a software tool. Feedback from these users will be critical to ensure that the software developed
will include the necessary features and be used in the long term. Group 2 stakeholders include
the vast LGO Alumni network at Amgen and the LGO Program leaders at Amgen who have a
vested interest in the success of the intern (and thus, the Amgen-LGO partnership) and can also
18
provide expertise in implementation of the new tool and integrating it with existing practices.
Group 3 stakeholders include the MIT Faculty advisors who can provide the intern with deep
technical expertise and general guidance to ensure that the Intern provides a tool that covers the
most relevant equipment technologies in enough technical depth and includes the most recent
information in the field of continuous manufacturing. Finally, the LGO Intern can also serve as a
project manager to ensure that the deliverables are completed efficiently and on-schedule.
The direct link to all of these stakeholders makes the MIT LGO Internship an ideal program for
developing and implementing this project, because Amgen is provided with an Intern that is low
cost; versatily trained in project management, engineering, and implementation; and has a vast
network of experts that can contribute significantly to developing an industry-leading equipment
selection tool.
19
Figure 3: Project Stakeholder Analysis
Group1Stakeholders
CoreTechnicalTeam
Group2Stakeholders
Group3Stakeholders
1.ProjectSupervisor
MITLGOIntern
2.ProjectChampion
3.SimulationEngineer
7.ProjectControlsDirector
4.CMProcessChemists
10.AmgenLGOAlumni
8.SeniorScientists
9.LGOSponsors&
ProgramLeader
5.PDEngineers
6.PDManager
11&12.MIT
Faculty
20
1.5 Project Focus
More specifically, the Amgen organization that is sponsoring this study (Pivotal Drug Substance
Process Development) has envisioned an equipment selection tool to aid their group in the
identification of low-cost, flexible off-the-shelf CM equipment to be used for a continuous
manufacturing platform for drug substance production. This project will focus on developing a
new methodology for equipment make and model selection for the continuous manufacturing of
drug substances; however, the equipment selection approach shall be designed in a general
manner so that its use can be expanded to other steps in the drug manufacturing process (e.g.,
drug product and batch production processes).
This study builds on the Krasberg method [6] that uses five quantitative selection criteria to
determine the appropriate equipment makes for a given process. Given the six-month time
constraint, this project will only focus on applying four out of the five proposed selection steps to
other equipment technologies. These four selection criteria (e.g., operating temperature, mixing
and pressure drop criteria) are based on design parameters and the overall steady-state material
and energy balance equations around the entire reactor. This study does not incorporate the fifth
criterion presented in the paper [6], which requires numerical methods to solve differential
energy balance equations to determine potential formations of local reactor hot spots.
Furthermore, the development of the equipment selection tool will be limited to the most
commonly used reactor, mixer, and crystallizer technologies for drug substance manufacturing –
specifically: (1) plug flow and microchannel reactors, (2) continuous stirred tank reactors, (3)
packed bed reactors, (4) static mixers, and (5) mixed-product, mixed-suspension crystallizers. An
21
overview describing each equipment technology is presented in Chapter 2.
Future studies on CM equipment selection are expected to build on the methodology introduced
in this thesis and be applied to other essential CM equipment technologies, such as continuous
filtration, pumps, distillation equipment, liquid-liquid extraction equipment, and drug product
manufacturing equipment.
1.6 Project Goals
The tool resulting from this research is also applied as a proof-of-concept to identify usable
equipment for three experimental continuous synthesis steps for an actual drug substance that is
in Amgen’s pipeline. The array of equipment identified by the tool for each synthesis step is
then compared with the set of equipment that was chosen by Amgen scientists for the same
syntheses, using their current equipment selection approach.
The general approach for equipment identification using this tool is also compared with the
equipment selection approach that Amgen scientists currently use. Any relative advantages and
disadvantages using the new method are noted.
1.7 Thesis Overview
Chapter 2 is an overview of the drug manufacturing process, including a high-level qualitative
description of the equipment technologies
Chapter 3 provides a hypothesis for the study.
Chapter 4 provides the data collection methodology and analysis to understand the state of
22
equipment selection prior to this research.
Chapter 5 is a detailed discussion of the approach for improving upon the current methods used
for equipment selection, including a design of a new computer-assisted equipment selection tool.
Chapter 6 is a discussion of the results and analysis of a case study conducted at Amgen, where
the new tool is applied to select equipment for the CM of an actual pipeline drug.
Chapter 7 provides detailed recommendations based on the aforementioned analysis and a
discussion of how to integrate the tool into the current process development work flow and the
potential for long term and future applications of the new tool.
Chapter 2: Pharmaceutical Manufacturing: Overview
2.1 Manufacturing Process Overview
Pharmaceutical manufacturing can be separated into two general categories. The “upstream”
side of manufacturing is the production of the drug substance or the active pharmaceutical
ingredient (API). The “downstream” side is the manufacturing of the drug product or the final
formulation of the drug that is administered in patients. The drug product contains both the API
and excipients, which are chemicals that serve various purposes including stabilization and
bulking up the formulation.
From a manufacturing perspective, small-molecule drugs, aka synthetics, differ from large-
molecule drugs, aka biologics, in the sense that synthetic drug substances are typically produced
via traditional organic chemistry reactions, while biologic drug substances are typically proteins
produced in bioprocesses, where cellular organisms produce and secrete the drug substance.
23
2.2 Conventional (Batch) Manufacturing of Drug Substances
Traditionally, both biologics and synthetics have been produced in batch processes, where raw
materials are charged into a closed system and processed completely before the proceeding to the
next batch step. In other words, batch manufacturing is comprised of sequential, time-dependent
processing steps that transform raw materials into a final drug product.
Drug substance batch manufacturing steps typically involve a (1) chemical reaction to synthesize
the API, (2) a step to crystallize the API out of solution, and (3) a method to filter out the
solution and retain the dry solid form of the drug substance [8]. Each step is completed before
proceeding to the next step. In addition to storage vessels, typical batch manufacturing facilities
include one or two stirred vessels for chemical reaction and crystallization processes and a filter
dryer vessel to isolate the drug substance solids. Often, the same stirred vessel can be used for
the reaction and crystallization steps.
Advantages to batch manufacturing are that the equipment technology is well-established and
that the vessels can be easily scaled up or become modular and reused for various operations [9].
Drawbacks are that the batch handling between steps and the cycle time for each batch are often
significant and can require larger batch sizes and the need for larger equipment and facility
footprint. In addition to the associated costs with larger manufacturing systems, a large batch
size can also dissuade process developers from pursuing highly exothermic chemistries for safety
reasons, even if they are the most efficient option [5,8].
2.3 Continuous Manufacturing of Drug Substances
Continuous manufacturing processes are similar to batch processes in the sense that their
24
equipment performs similar unit operations (e.g., mixing, reaction, crystallization, filtration, and
drying), but differ from batch processes because, after initial startup, all continuous processes are
operating simultaneously, and material continuously flows from one step to the next. In the
context of drug substance manufacturing, in a fully continuous process, the reaction,
crystallization, and filtration steps are all operating at the same time, and the reactants, products,
solvents, and other materials are all continuously flowing through from one step to another. In
other words, raw materials are continuously consumed, while drug substance is continuously
produced [9]. After an initial startup, all of the above-listed CM operations have the capacity of
operating at quasi-steady conditions in which the inputs, outputs, and states only have small
variations over time. With some abuse of notation, such operations will be referred to as being
in “steady-state” in this thesis, although some time variations always occur during manufacturing
due to process disturbances. Some CM operations such as continuous chromatography do not
operate under quasi-steady conditions, but are more prevalent in biologic rather than synthetic
drug substances.
A major advantage of continuous processing is the reduction of handling, downtime, and setup in
between processing steps, which allows more time dedicated to production. The additional
production time relative to a batch process allows for smaller equipment and thus a smaller
factory footprint [9], which directly reduces costs. Moreover, smaller equipment and less
material holdup give process developers the option to choose potentially more efficient, highly
exothermic reactions for synthesizing their drug substance without incurring the safety risks
associated with those chemistries when using large batch equipment.
25
Though CM is promising in theory, the equipment classes used for continuous drug substance
production can vary greatly, depending on reaction and crystallization needs. The variety of CM
equipment technologies and lack of process flexibility relative to batch technologies has been
recognized as a hurdle for adopting continuous manufacturing processes as a mainstream
production strategy [3]. One of the major challenges to CM standardization and modularization
is choosing from the numerous makes and models and types of technology for a flexible
continuous manufacturing platform.
2.4 CM of Drug Substance: Equipment Technology Overview
In this section, the general characteristics and diagrams are presented for the equipment covered
in this study. The detailed design equations that govern the operations and performance of each
type of equipment are presented in Chapter 5.
2.4.1 Continuous Reactors
For the reaction step, numerous types of continuous reactors are available for the synthesis
process. The reactors covered in this study include: plug flow reactors (PFRs), microchanneled
reactors (microreactors), packed bed reactors (PBRs), and continuous stirred tank reactors
(CSTRs). Each reactor is chosen based on different chemistry needs, such as the need for solid
catalysts, or better heat transfer, mixing, conversion, and selectivity requirements. At steady
state, all continuous reactors are designed to have little to no accumulation of material or energy
in the vessels and are fed a continuous stream of reactants and have a continuous stream of crude
product mixture exiting the system.
26
2.4.1.1 Continuous Stirred Tank Reactors (CSTRs)
CSTR is a broad category of reactors with mechanical stirrers, such as impellers, inside the
vessel that maintain a well-mixed slurry or solution. The vessel is typically upright and
cylindrical with a convex (e.g., conical) bottom outlet to prevent solids settling. Other reagents
or chemicals needed to control reaction conditions are introduced to the reactor using feed lines,
and a jacket for heating and/or cooling is often present to control system conditions. For
traditional first- and second-order kinetic reactions, CSTRs, in general, have a lower overall
conversion per unit volume when compared with other reactors such as the PFR. CSTRs are the
most suitable reaction type for certain reaction types that require significant mixing or
suspension, and for autocatalytic reactions, in which CSTRs have higher overall conversion per
unit volume [15].
2.4.1.2 Plug Flow Reactors (PFR) and Microreactors
The PFR, aka tube or tubular reactor, is a category of tube-shaped reactors with continuous flow
that is often simply a tube or pipe with controlled operating conditions (e.g., temperature and
pressure) to yield the desired reaction results. Because a mechanical agitator is present, reaction
conditions are highly non-uniform throughout the tube and can vary both radially and axially,
depending on component concentrations. For classic first- and second-order reactions, PFRs are
often preferred over CSTRs when high conversion is desired, because the reactant concentrations
towards the entrance of the tube are higher, allowing a higher overall conversion per unit volume
of reactor relative to that of a CSTR of the same volume. In other words, a CSTR will often
need a larger volume to achieve the same conversion as that of a PFR for first- and second-order
reactions [15]. This well-known result can be seen mathematically from the mass balance design
27
equations for each type of reactor, which is discussed in Chapter 5.
Microreactors are reactors with volumes on the scale of microliters (µl) that often have multiple
channels and bends to reduce footprint. Microreactors are often used by process developers for
small-scale syntheses and can be approximately characterized using the design equations for
normal tube reactors, using hydraulic diameters and equivalent pipe lengths. Their advantages
include ease of use and good heat transfer, but their tortuous channel paths can cause a high
frictional pressure drop, making these reactors difficult to scale up geometrically.
2.4.1.3 Packed Bed Reactors
The PBR is a general category of reactors in which cylindrical vessels or columns are filled with
beds of solid packing that are essential for the proper reaction(s) to occur. Typically, the packing
is a catalyst that significantly aids in the kinetics and selectivity of the reaction. This type of
reactor is common for chemistries such as hydrogenation that require a solid catalyst in contact
with a gaseous and/or liquid reactant stream. These flow reactors are commonly used in the
petroleum refining industry to saturate aromatic compounds, crack large organic compounds, or
remove sulfur from organic material.
2.4.2 Static Mixers
Depending on reactant or product mixing requirements, mixers could also be an essential
equipment class for the synthesis process. This study focuses on the selection of static mixers,
which is a class of equipment that does not have any mechanically powered mixing components.
The mixing is driven mostly by the fluid velocity in contact with fixed baffles and walls within
28
the mixer that create a tortuous path. The performance of mixing is strongly correlated with
frictional pressure drop [16].
2.4.3 Crystallization
For crystallization technology, this study focuses on mixed-suspension mixed-product removal
crystallizers (MSMPRs), as that technology is the most mature in industry [31]. MSMPRs are
continuous stirred vessels that can also be used for CSTR applications. In addition to having
mechanical stirrers, MSMPRs often also have a cooling jacket and chemical feed lines to control
key crystallization variables such as pH or yield. MSMPRs are advantageous when strong
agitation or stirring is needed, but other crystallization technologies are preferred if crystals are
sensitive to shear induced by the impellers (e.g., [33]).
2.4.4 Filtration and Drying
After the crystallization step, the slurry is pumped or fed gravitationally to a set of filters and
dryers, where the solid API crystals are separated from the liquid filtrate. Several reagents are
often passed over the slurry to fully remove solvents and to ensure a dry API cake of adequate
purity. While commercial equipment for continuous filtration and drying are available, first-
principles models are not yet well validated for pharmaceutical applications, so filtration and
drying are not a focus of this study.
29
Chapter 3: Hypothesis
As seen in the previous chapter, there are numerous different CM equipment technologies and
makes and models that are governed by a variety of equations and principles. This diversity of
available CM equipment can make equipment selection very complex, especially when choosing
the most flexible equipment assets to be used for multiple synthesis steps.
It is hypothesized that a novel tool providing computer-assisted equipment selection for each
equipment technology with a feature allowing simultaneous equipment selection for multiple
processes will positively impact the equipment selection process in terms of: (1) selection speed
and (2) the identification of lower cost or more flexible, usable assets.
Chapter 4: Problem Analysis
All interviews, data collection, and analysis were conducted at Amgen, Inc. primarily at the
Process Development site in Cambridge, Massachusetts.
The data collection and analysis methodology was divided into three distinct parts: (1)
understanding the current state of equipment selection, (2) understanding the needs of key
stakeholders who will use the tool, and (3) continuously requesting feedback from users as the
tool is developed.
30
The needs of the stakeholders were summarized in the Problem Focus in Chapter 1. This chapter
discusses the current state of equipment selection, which was used as the basis for developing the
software tool to ensure that the tool improved upon the status quo.
4.1 State of the CM Equipment Selection Process prior to Project Implementation
Figure 4 is a schematic of the generalized form of the existing equipment selection process at
Amgen prior to the implementation of the tool developed from this research. First, a compatible
equipment technology is chosen for the proposed continuous synthesis being studied, and then
process information such as fluid properties and reaction kinetics are either modeled or collected
through experimental data. Then the equipment is sized using an equipment design software
suite such as DynoChem®, based on physical and chemical properties of the process and the
production scale. Then the equipment make and model were chosen by manually searching
through catalogs from common vendors or by contacting a sales representative to identify
equipment that matches the design from the software suite. Occasionally, if an off-the-shelf
model is not available, sales representatives are contacted to discuss the possibility of customized
equipment. Approximately 75% of the time, scientists have been able to reuse equipment assets
that have already been purchased, and about 25% of the time, scientists have had to purchase
new equipment assets. The feedback about this process was that it requires significant time
resources and is very tedious – too tedious to complete the design requirement calculations for
multiple processes.
31
Figure 4: Existing Equipment Selection Decision-making Process (Prior to Project)
4.2 Improvements to the Current Equipment Selection Process
As discussed in Chapter 1, the focus of this project is to improve how the equipment makes and
models are identified, assuming that the equipment technology has already been selected. For
this aspect of equipment selection, surveys indicated that there were two key areas of
improvement that are strongly desired.
First, the search for equipment models manually through numerous catalogs and the wait for
sales representatives were significant time-consuming steps in the equipment selection process.
Therefore, there was strong demand for a tool that could facilitate the search by combining
information from various vendors about makes and models of the same equipment technology
into one searchable tool or database. In other words, the goal is to create a software equipment
selector tool that provides a “one stop shop.”
Additionally, there was a request for software that could conduct all of the tedious calculations
that characterized the performance for each make and model of equipment and all of the
calculations that characterized the equipment design requirements for the given process. For
example, there is currently no software available at Amgen that estimates both the heat transfer
32
rate capabilities (performance) of each make and model of CSTRs and the required heat transfer
rate for a given chemistry (requirement). Consequently, to compare the heat transfer
requirements with a product’s capabilities, calculations, and comparisons had to be done either in
separate software suites or manually. There was no “one stop shop.”
Chapter 5: Redesign of the Equipment Selection Process
The demand for computer assistance for engineering calculations, as presented in Chapter 4, is
consistent with the strategy to build upon the Krasberg approach for selecting equipment [6].
Figure 5 is a schematic for a proposed new process for equipment selection with modifications
shown in green.
Figure 5: Proposed Equipment Selection Process with New Tool
5.1 Overview of Software Tool Design
Specifically, each equipment technology has a tool that includes five different types of
spreadsheets. First, there is a sheet for the user to input values for extensive and intensive
chemical and physical properties (“Process Inputs Sheet”) related to the synthesis process (e.g.,
33
production scale, reaction exothermicity, fluid viscosity and density). Second, there is a sheet
used as a database to store parametric information for each make and model of equipment
(“Equipment Inputs Sheet”) that is needed to characterize its size and performance (e.g.
geometric dimensions, material of construction, allowable temperatures, and maximum agitator
speed). The information from these two sheets provides inputs into the third spreadsheet
(“Calculation Sheet”), which includes the formulas that execute the design requirement
calculations and the Boolean logic that determines if the equipment make and model
performance satisfies the design requirements. If all design requirements are satisfied, the tool
indicates that the equipment model is ‘appropriate’ for the process.
Lastly, there was a fifth spreadsheet (“Results Sheet”) proposed that included pivot tables to
easily generate pivot charts to quickly and visually categorize equipment makes and models
based on specific design parameters (e.g., flexibility, cost, and pressure drop).
The first and third spreadsheets are designed to allow inputs and calculations for multiple CM
processes being studied so that appropriate equipment can be identified for each process
simultaneously. This design allows the user to easily identify the most flexible equipment assets
for his or her portfolio of chemistries and drug synthesis steps.
For each of the equipment technologies listed in the Project Focus in Chapter 1, a set of
quantitative and qualitative design requirements are chosen to determine if an equipment make
and model are appropriate for the synthesis process. The next sections explain each design
requirement and selection criterion in detail for each equipment technology. As a reminder, the
34
previous ‘Parameter Legend’ section defines each parameter in the equations displayed in this
chapter.
5.2 Equipment Selection: General Design Requirements
While each equipment technology has its specific performance design requirements, there are
three general safety and reliability requirements that can be applied to all equipment. These
three requirements are: material compatibility, and the ability for the equipment models to
withstand process temperatures and pressures.
5.2.1 Material Compatibility Requirement
Material compatibility of the equipment model with the process chemistry is important to ensure
that the solvents, reactants, products, and reagents can be operated safely within the equipment
without significant corrosion or other damage to the vessel that could cause an unplanned or
accelerated loss of containment. Furthermore, the equipment material of construction must be
inert and not react with the process material to ensure that the synthesis reactions occur as
designed.
In the Process Inputs sheet, the user specifies the required material of construction for the vessel,
based on his or her knowledge of the process chemistry and its compatibility with certain
materials. For each equipment make and model, there is a column in the Equipment Inputs
spreadsheet that documents the material of construction for each vessel, as specified by the
vendor. In the Calculations sheet, there is a column that contains Boolean Logic comparing the
material of construction of the equipment with the material of construction requirements
35
specified by the user. If the material of construction of the equipment does not meet the
requirements, the equipment model is considered inappropriate for the process chemistry.
For future versions of this tool, it is recommended that a material compatibility matrix be added
as a spreadsheet. Newer versions of the tool should include features that allow the user can
specify the specific chemical components of the process fluid and use algorithms to search
through the matrix for the materials of construction that are compatible with the components.
This extension would allow the material of construction requirements to be determined
automatically by the tool as opposed to being determined manually by the user.
5.2.2 Process Temperature Requirement
The process temperature requirement column in the “Calculations” spreadsheet is Boolean logic
that compares the desired process operating temperature (user-specified) with the allowable
temperature range for each CSTR vessel (vendor-specified) that is documented in the
“Equipment Inputs” spreadsheet. If the process temperature is not within the allowable
temperature range of the equipment, the equipment model is considered unusable for the process
being studied.
5.2.3 Process Pressure Requirement
Similar to the Process Temperature Requirement in Section 5.2.2, there is also a pressure
requirement in the “Calculations” spreadsheet. This column includes Boolean logic that
compares the desired operating pressure (user-specified) with the allowable pressure range for
each CSTR vessel (vendor-specified). If the process pressure is not within the window of
36
allowable pressures for the equipment, the equipment model is considered unusable for the
process being studied.
For PFRs, packed bed reactors, and micromixers, the pressure drop across the vessel must be
added to the maximum expected outlet pressure to estimate the maximum inlet pressure. This
inlet pressure is then compared to the maximum pressure of the vessel. Instructions for pressure-
drop calculations are detailed in the below sections.
Table A.5.3: Summary of Motionless Mixer Design Criteria (Make/Model usable if all criteria
are passed)
87
Bibliography
[1] Cha, M. and Yu, F. “Pharma’s first-to-market advantage.” Pharmaceuticals & Medical Products. McKinsey and Company, Aug. 2014.
[2] Adams, C.P. and Brantner, V.V. “Spending on new drug development.” Health Econ.
19.2: 130–141, 2010.
[3] Buchholz, S. “Future manufacturing approaches in the chemical and pharmaceutical industry.” Chemical Engineering and Processing: Process Intensification 49.10: 993–995, 2010.
[4] Singh, R., Godfrey, A., Gregertsen, B., Muller, F., Gernaey, K., Gani, R., and Woodley,
J., “Systematic substrate adoption methodology (SAM) for future flexible, generic pharmaceutical production processes.” Computers and Chemical Engineering 58: 344–368, 2013.
[5] Gutmann, B., Cantillo, D., and Kappe, C.O. “Continuous-flow technology—A tool for
the safe manufacturing of active pharmaceutical ingredients.” Angewandte Chemie International Edition 54.23: 6688–6728, 2015.
[6] Krasberg, N., Hohmann, L., Bieringer, T., Bramsiepe, C., and Kockmann, N. “Selection
of technical reactor equipment for modular, continuous small-scale plants.” Processes, 2(1), 265-292, 2014.
[7] Commenge, J.-M., and Falk, L. "Methodological framework for choice of intensified
equipment and development of innovative technologies.” Chemical Engineering and Processing: Process Intensification 84: 109–127, 2014.
C.M., Lawrence, X.Y. and Woodcock, J. “Modernizing pharmaceutical manufacturing: From batch to continuous production.” Journal of Pharmaceutical Innovation 10(3), pp.191–199, 2015.
[9] Sharratt, P.N., editor, Handbook of Batch Process Design. Springer Science & Business
Media, London, UK, 1997.
[10] Yu, L. “Continuous manufacturing has a strong impact on product quality.” FDA Voice. U.S. Food and Drug Administration,Silver Spring, Maryland, April 12, 2016. Web.
[11] Lee, S., Advancement of Emerging Technology Applications to Modernize the
Pharmaceutical Manufacturing Base Guidance for Industry DRAFT GUIDANCE. U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Silver Spring, Maryland, December 2015.
88
[12] Adamo, A., Beingessner, R.L., Behnam, M., Chen, J., Jamison, T.F., Jensen, K.F., Monbaliu, J.C.M., Myerson, A.S., Revalor, E.M., Snead, D.R. and Stelzer, T., “On-demand continuous-flow production of pharmaceuticals in a compact, reconfigurable system.” Science 351.6281: 61–67, 2016.
[13] Root, C. “Amgen opens biologics manufacturing facility in Singapore.” Bioprocess
Online, November 25, 2014. Web.
[14] Neuman, B.S. “Pharmaceutical industry wastes $50 billion a year due to inefficient manufacturing.” The Source. Washington University in St. Louis, Missouri, October 9, 2016. Web.
[15] Nauman, E.B. Chemical Reactor Design, Optimization, and Scaleup. John Wiley & Sons,
Hoboken, NJ, USA, 2008.
[16] Paul, E.L., Atiemo-Obeng, V.A., and Kresta, S.M., editors. Handbook of Industrial Mixing: Science and Practice. John Wiley & Sons, Hoboken, New Jersey, USA, 2004.
[17] Bell, K.J. “Overall heat transfer coefficient.” In A-to-Z Guide to Thermodynamics, Heat
and Mass Transfer, and Fluids Engineering, 2017. http://www.thermopedia.com/content/1007/
[18] Churchill, S.W. and Chu, H.H., “Correlating equations for laminar and turbulent free
convection from a vertical plate.” International Journal of Heat and Mass Transfer, 18.11: 1323–1329, 1975.
[19] Cullen, P.J., Abatzoglou, N. and Rielly, C.D. eds., Pharmaceutical Blending and Mixing.
John Wiley & Sons, Hoboken, NJ, USA, 2015.
[20] Sánchez Pérez, J.A., Rodríguez Porcel, E.M., Casas López, J.L., Fernández Sevilla, J.M., and Chisti, Y. “Shear rate in stirred tank and bubble column bioreactors.” Chemical Engineering Journal 124.1–3: 1–5, 2006.
[21] Atherton, J.H. and Carpenter, K.J. Process Development: Physicochemical Concepts.
pp.1–88, 1999.
[22] Lines, P.C., and Carpenter, K.J. “The effect of physical property ranges on correlations for minimum impeller speeds to disperse immiscible liquid mixtures.” IChemE Symposium Series 121: 167–182, 1990.
[23] Bergman, T.L., Incropera, F.P., DeWitt, D.P. and Lavine, A.S., Fundamentals of Heat
and Mass Transfer. John Wiley & Sons, Hoboken, New Jersey , USA, 2011.
[24] Colebrook, C.F., Blench, T., Chatley, H., Essex, E., Finniecome, J., Lacey, G., Williamson, J. and Macdonald, G., “Turbulent flow in pipes, with particular reference to the transition region between the smooth and rough pipe laws (includes plates).” Journal of the Institution of Civil Engineers, 12.8: 393–422, 1939.
89
[25] Peters, P.E., Schiffino, R.S. and Harriott, P., “Heat transfer in packed-tube reactors.” Ind. Eng. Chem. Res. 27.2: 226–233, 1988.
[26] Ergun, S. “Fluid flow through packed columns.” Chem. Eng. Prog. 48: 89–94, 1952.
[27] Perry, R.H., and Green, D.W. Perry’s Chemical Engineers’ handbook. McGraw-Hill