Practical Process Research
`and Development
`A guide for organic chemists
`
`Second Edition
`
`Neal G. Anderson
`Anderson’s Process Solutions
`Jacksonville, Oregon
`
`AMSTERDAM l BOSTON l HEIDELBERG l LONDON
`NEW YORK l OXFORD l PARIS l SAN DIEGO
`SAN FRANCISCO l SYDNEY l TOKYO
`
`Academic Press is an imprint of Elsevier
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 1 of 86
`
`

`

`Academic Press is an imprint of Elsevier
`The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK
`225 Wyman Street, Waltham, MA 02451, USA
`Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands
`Copyright Ó 2012 Elsevier Inc. All rights reserved.
`
`No part of this publication may be reproduced, stored in a retrieval system or transmitted in
`any form or by any means electronic, mechanical, photocopying, recording or otherwise without
`the prior written permission of the publisher
`
`Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in
`Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: permissions@
`elsevier.com. Alternatively you can submit your request online by visiting the Elsevier web site at
`http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material
`
`Notice
`No responsibility is assumed by the publisher for any injury and/or damage to persons or property
`as a matter of products liability, negligence or otherwise, or from any use or operation of any
`methods, products, instructions or ideas contained in the material herein
`
`British Library Cataloguing in Publication Data
`A catalogue record for this book is available from the British Library
`
`Library of Congress Cataloging-in-Publication Data
`Anderson, Neal G.
`Practical process research and development : a guide for organic
`chemists / Neal G. Anderson. – 2nd ed.
`p. cm.
`Rev. ed. of: Practical process research & development. c2000.
`ISBN 978-0-12-386537-3 (hardback)
`1. Chemical processes. I. Anderson, Neal G. Practical process
`research & development. II. Title. III. Title: Practical process
`research and development.
`TP155.7.A55 2000
`541’.39–dc23
`
`2011051049
`
`ISBN: 978-0-12-386537-3
`
`For information on all Academic Press publications
`visit our website at books.elsevier.com
`
`Printed and bound in the USA
`12 13 14 15 16 10 9 8 7 6 5 4 3 2 1
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 2 of 86
`
`

`

`Chapter 1
`
`Introduction
`
`Chapter Outline
`I. Introduction
`II. Equipment Considerations
`on Scale
`III. Operations Preferred
`on Scale
`
`1
`
`7
`
`12
`
`IV. Patent Considerations
`V. Summary and Perspective
`References
`
`15
`16
`19
`
`“The world doesn’t move because of idealism.. [i]t moves because of economic incentives.”
`– Fernando Canales Clariond, formerly Mexico’s secretary of the economy [1]
`
`“It is well-known that there are no technical optima in industry, only economic optima..”
`– G. Guichon et al. [2]
`
`“Today, green chemistry is simply a good business choice.”
`
`– Paul Anastas [3]
`
`I. INTRODUCTION
`
`The driving forces of the pharmaceutical industry are to develop medicines that maintain
`or improve health and the quality of life for people, and to provide a reasonable return for
`investors. The many unknowns of the business, especially our imperfect understanding
`of biology, make for a high-risk environment. Yet the rewards are high as well. In
`Table 1.1 are presented some statistics associated with developing drugs, and these
`statistics explain some of the pressures of the business.
`Significant financial gains are possible by developing drugs, as indicated by the
`penalties and fines levied against firms and people recently. For instance, in 2010
`AstraZeneca was fined $520,000,000 for promoting Seroquel for off-label uses [4] and
`GlaxoSmithKline was fined $150,000,000 for violations of current Good Manufacturing
`Practices (cGMPs) [5]. In 2009, Pfizer was fined $2,300,000,000 for illegally promoting
`
`Practical Process Research and Development. DOI: 10.1016/B978-0-12-386537-3.00001-0
`Copyright Ó 2012 Elsevier Inc. All rights reserved.
`
`1
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 3 of 86
`
`

`

`2
`
`Practical Process Research and Development
`
`TABLE 1.1 Some Statistics Relevant to the Pharmaceutical Industry
`
`Value
`
`Factor
`
`$1,300,000,000
`
`Cost to bring a drug to market (1)
`
`5e20% of drug product price
`
`CoG of API
`
`30% of the CoG for
`drug product
`
`Cost of QC (2)
`
`Millions of dollars
`
`Cost of failed drug formulation (3)
`
`As high as market will bear
`
`Price of drug product to consumer
`(higher for US than for most countries) (4)
`
`$75,000
`
`The value of one additional year of life, set in 2005 by
`health economists (5)
`
`$200,000e$300,000
`
`Annual cost of US chemist or engineer for an employer
`
`About 95%
`
`About 30%
`
`8 years
`
`20 years
`
`Portion of drug candidates that fail in pre-clinical or clinical
`studies
`
`Portion of approved drugs that recoup development costs (6)
`
`Average time of development (goal is 5 years)
`
`Period for exclusive sales of a patented drug (US)
`
`20 e Development time
`
`Years to recoup investment costs
`
`$1,000,000
`
`Sales lost for every day that a filing is delayed, if drug sales
`are $400 MM/year
`
`(1) Undoubtedly includes the cost of advancing drug candidates that failed. Jarvis, L. M. Chem. Eng. News 2010,
`88(23), 13. May be higher: Vertex developed Incivek (telaprevir) over 20 years at the cost of around
`$4,000,000,000: Jarvis, L. Chem. Eng. News 2011, 89(22), 8.
`(2) Mullin, R. Chem. Eng. News 2009, 87(39), 38.
`(3) May be higher for formulations involving spray drying. A huge cost may be incurred if the physical form of an
`API is not controlled and a dosage form fails specifications, thus interrupting clinical trials or discontinuing
`sales of a drug (Chapter 13). Thayer, A. M. Chem. Eng. News 2010, 88(22), 13.
`(4) http://www.economist.com/node/4054095 (June 16, 2005).
`(5) The Washington Post, July 17, 2005: http://www.washingtonpost.com/wp-dyn/content/article/2005/07/16/
`AR2005071600941_pf.html.
`(6) Mod. Drug Discov. 2001, 4(10), 47.
`
`Bextra and three other medications [6], and Bristol–Myers Squibb (BMS) was fined
`$2,100,000 for making agreements with Apotex to delay the launch of generic Plavix
`[7]. In 2009, the FDA stopped reviewing applications from Ranbaxy and prohibited
`Ranbaxy from importing 30 generic drugs, due to falsified QC data [8]. In 2007, a former
`top official of China’s state organization approving drugs was executed for taking bribes
`[9]. In 2004, BMS was fined $150,000,000 for “channel stuffing,” an accounting practice
`that artificially boosted the sale of drugs [10]. The reputation of the pharmaceutical
`industry has been sullied over the past few decades, but many people, including this
`author, enter this industry because they want to be able to help others.
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 4 of 86
`
`

`

`Chapter | 1 Introduction
`
`3
`
`The development and sales of drugs are influenced by an interplay of financial,
`political, governmental and personal considerations. For example, Bayer reduced the
`cost of a pill of ciprofloxacin from $1.77 to $0.75 after the horrific crashes of
`September 11, 2001 [11]. As of August 2009 the pharmaceutical lobby had 1544
`lobbyists, or almost three lobbyists per congressman in Washington DC [12]; this is
`a sizeable increase from the 625 registered lobbyists in 2001 [11]. No major US
`pharmaceutical company would develop RU-486, an abortifacient, due to backlash
`anticipated from conservative groups. The sale of anti-AIDS drugs at reduced prices to
`the third world is a nice example of philanthropy; probably some income tax write-offs
`are also involved. Philanthropic efforts exist, such as Merck’s gifts of ivermectin to
`prevent river blindness and the development of tenofovir by the Clinton Health Access
`Initiative to treat AIDS in the developing world [13]. Pharmaceutical industries may
`need support from government to continue developing drugs for the third world, such
`as compounds to treat malaria [14]. Efforts to minimize wastes and decrease impacts
`on the environment have increased, due to sensible and altruistic reasons, and due to
`penalties imposed. Although bacteria resistant to powerful antibiotics are continually
`emerging, the development of antibiotics has slowed due to the anticipated longer
`times to recoup development costs from drugs that are not taken on a chronic basis.
`The development of new chemical entities (NCEs) is becoming more difficult with
`increased scrutiny by regulatory authorities; for instance, the third or fourth entry into
`a therapeutic category may have to demonstrate superior benefits to win FDA approval
`[15], and control of potentially genotoxic impurities at the ppm level demands addi-
`tional efforts. And everything is
`influenced by people striving for personal
`advancement.
`As a result of increasing pressure to bring compounds to market, business trends within
`the pharma sector have been changing. Working smarter and faster is stressed, for
`instance, using high-throughput screening and statistically designed experiments. Phar-
`maceutical industries are being pressured to develop more efficient processes [16]. The
`FDA has advanced process analytical
`technology (PAT), and this may decrease
`manufacturing costs [17]. The importance of solid process development efforts has been
`recognized [18].
`The complexity of compounds that has emerged as active pharmaceutical ingredients
`(APIs) may be increasing, as shown in Figure 1.1, and structural complexity increases
`the cost of development. Drug development is expensive: Vertex’s first approved drug
`was developed in-house over 20 years at the cost of around $4,000,000,000. Incivek
`(telaprevir) will treat hepatitis C, at a projected price of $49,200, competing with
`Merck’s Victrelis (boceprevir) at a projected price of $31,000–$44,000. Each company
`has set up assistance programs to help with the co-payments of insured patients [19].
`The foundation of thorough processes is the work carried out by academicians.
`Some brilliant total syntheses have been described by academic chemists, such as the
`syntheses of codeine by the groups of Stork [20] and Magnus [21]. Total synthesis in an
`academic sense [22,23] is often not suitable for scale-up on an industrial scale; scaling
`up in an academic setting to millimoles or grams may not be enough to uncover and
`address processing difficulties. Prof. Hudlicky has described many practical consid-
`erations for syntheses [24] that are applicable to process R&D efforts in the phar-
`maceutical industry.
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 5 of 86
`
`

`

`4
`
`Practical Process Research and Development
`
`CO2Na
`
`O
`
`N
`
`PO O
`
`O
`
`OH
`
`3C CH3
`fosinopril sodium
`
`Ph
`H3C
`
`HS
`
`CH3
`
`N
`
`O
`captopril
`
`CO2H
`
`HN
`
`O
`
`O
`CH3
`
`HN
`
`O
`
`O
`
`N
`
`HN
`
`O
`
`NH
`
`O
`
`N N
`
`F
`
`O
`
`NH2
`
`HN
`
`O
`
`NH
`
`OH
`
`O
`
`HN
`
`O
`
`CH3
`
`H2N
`
`frakefamide
`
`telaprevir
`
`FIGURE 1.1 Structures of APIs and NCE in Table 1.2.
`
`TABLE 1.2 Details on Development of Four APIs and an NCE
`
`Captopril
`(1)
`
`Fosinopril
`(2)
`
`Frakefamide
`(3)
`
`Enfuvirtide
`(4)
`
`Telaprevir
`(5)
`
`Market date
`
`1981
`
`w1988
`
`Halted after
`Phase 2
`
`2003
`
`2011
`
`Steps
`
`Isolations
`
`Preparative
`chromatographies
`
`5
`
`5
`
`0
`
`14
`
`13
`
`0
`
`7
`
`3
`
`0
`
`109
`
`7
`
`2
`
`NA
`
`NA
`
`NA
`
`CoG
`
`NA
`
`NA
`
`$2500/kg
`
`$36,000/kg
`
`NA
`
`Therapeutic area
`
`High blood
`pressure
`
`High blood
`pressure
`
`Pain treatment
`
`AIDS
`
`Hepatitis C
`
`Patient cost/year
`
`NA
`
`NA
`
`NA
`
`$25,000
`
`$49,200
`
`(1) Anderson, N. G.; Bennett, B. J.; Feldman, A. F.; Lust, D. A.; Polomski, R. E. US Patent 5,026,873, 1991
`(to E. R. Squibb & Sons.)
`(2) Grosso, J. A. US Patent 5,162,543, 1992 (to E. R. Squibb & Sons).
`(3) Franze´n, H. M.; Bessidskaia, G.; Abedi, V.; Nilsson, A.; Nilsson, M.; Olsson, L. Org. Process Res. Dev. 2002,
`6, 788.
`(4) Enfuvirtide is a synthetic peptide made of 36 aminoacids: Bray, B. Nat. Rev. Drug Discov. 2003, 2, July, 587.
`(5) Jarvis, L. M. Chem. Eng. News 2011, 89(22), 8.
`
`Perhaps the first description of the principles of process research and development as
`applied to pharmaceutical products was written by Trevor Laird in 1988 [25]; this article
`covered a breadth of topics, including design of experiments (DoEs), automation and the
`importance of polymorphism. Caron recently detailed the responsibilities of the process
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 6 of 86
`
`

`

`Chapter | 1 Introduction
`
`5
`
`chemist [26]. In the past 11 years there have been many insightful books and many
`detailed articles on process development [27–49].
`Process chemists from AstraZeneca, GlaxoSmithKline, and Pfizer critically assessed
`the parameters for route selection, developing the SELECT acronym: for SAFETY,
`Environmental, Legal, Economics, Control, and Throughput [50]. SAFETY is of
`primary importance. The environmental impact of processing falls under the heading
`of green chemistry. Legal matters are important in the pharmaceutical business, espe-
`cially regarding the issuing of patents and freedom to operate. Economics are expressed
`in the cost of goods (CoG), as influenced by the cost of raw materials, labor, quality
`control (QC), and waste disposal. A primary goal is to minimize the cost of an API to
`$1000–$3000/kg. To minimize the CoG chemistry must be efficient and robust, and
`stringent in-process controls (IPCs) are crucial to ensuring routine operations to produce
`high-quality products. Starting materials and reagents must be subject to competitive
`bidding [51]. Multi-purpose equipment is used in development efforts in order to
`minimize capital investment; when sales of a compound seem more certain purchasing
`dedicated and specialized equipment may be justified. Throughput is reflected in not
`only isolated yields, but also in the amount of product that can be made per unit of
`volume or per unit of time (space–time yield). Manufacturing complexity is minimized;
`for instance, eliminating the use of hazardous or toxic chemicals eases SAFETY
`concerns, and by minimizing processing bottlenecks and the proportion of rejected
`batches that need to be reworked productivity is improved. Throughput and productivity
`are expressions of economics, and these are some of the key drivers for the
`manufacturing of APIs. The SELECT criteria can be considered throughout this book.
`A perspective on process research and development activities is shown in Figure 1.2, in
`which optimization of process development increases going to the right. Based on results
`of screening, a compound may be designated as an NCE and advanced on the pathway to
`becoming an API. As with any general guideline, the development of an API may not
`follow this sketch. For instance, quantities may be prepared in the kilo lab after research
`efforts have identified an optimal route and preferred reagents; in other instances supplies
`of an API suitable to meet the demand for several years may be manufactured in one 5-gm
`batch. The time to progress from an idea to an API may be 5–13 years, or longer.
`Efficient process development begins with drug discovery, and continues through
`routine manufacturing of the drug product [52]. The roles in drug discovery, process
`
`idea
`
`DISCOVERY
`
`batch
`size
`
`mg - g
`
`emphasis:
`
`expedient
`
`PROCESS
`RESEARCH
`KILO
`LAB
`
`PROCESS
`DEVELOPMENT
`
`routine
`e
`g
`manufacturing
`
`kg
`
`1 - 100 kg
`
`> 100 kg
`
`tox batches;
`Phase 1
`
`convenient
`
`Phase 2
`
`Phase 3
`
`practical
`
`efficient
`
`FIGURE 1.2 Perspective on the drug development continuum.
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 7 of 86
`
`

`

`6
`
`Practical Process Research and Development
`
`research and process development are different. Drug discovery chemists design expe-
`dient, diversity-oriented syntheses to prepare many compounds from a common inter-
`mediate. Process research chemists focus on convenient routes to prepare one or perhaps
`a few similar compounds; route selection and reagent selection are usually prime
`considerations. Process development chemists optimize the processes to prepare an NCE
`or API, focusing on minimizing impurities, streamlining work-ups, avoiding chroma-
`tography if possible, and isolating the desired final form of one product [53]. Of course
`these general descriptions may overlap. Almost inevitably process chemists change the
`route developed by the discovery chemist; for a new project one of the initial efforts of the
`process chemist is to consider possible benefits in reordering steps in a route. At some
`point process chemists consider changing the route, using different starting materials or
`taking advantage of biocatalysis. The reactions commonly encountered by researchers
`within pharmaceutical process R&D [54] and drug discovery [55] are relatively similar,
`not surprisingly. The process chemist speeds the development of an NCE by considering
`the “hows” and “whys” of processing. While carrying out the present duties one can
`consider the needs of the next step and thus smooth the development of an NCE. Time
`invested early to make material by processes that can be readily scaled up will shorten
`overall development time.
`Of course the problem with investing time in downstream efforts is that most
`compounds fail to make the marketplace; in 2003, Federsel estimated that less than
`0.03% of all active compounds prepared were approved by regulatory authorities and
`70% of drugs failed to recoup the cost of investments [56]. Perhaps 95% of compounds
`passing through process R&D operations fail. Optimizing processes too early may waste
`resources, but if more material is quickly needed, scale-up may be difficult unless some
`optimization was in place. By Phase 3 the route should be established and process
`optimization should be well underway; if not, supplies for Phase 3 may be at risk.
`Furthermore, the pressure for filing an NDA may result in advancing processes needing
`substantial optimization. Suffice it to say that not all processes are optimized before
`scale-up, and processes may be rushed into production to provide material for clinical
`trials or manufacturing. The conundrum of when to invest in process optimization often
`pivots on the GO/NO GO decision points (Table 1.3).
`“Fit for purpose” has been used and overused to describe efforts to make NCEs. All
`process R&D should be fit for purpose, but usually this phrase is applied to the early
`efforts to make sufficient supplies for toxicology and Phase 1 studies in an expeditious
`fashion. Once proof of safety in humans is established there is usually a need to rede-
`velop processes to make them more efficient and lower the CoG such efforts are also fit
`for purpose [57]. Almost inevitably processes are changed to prepare material for Phase
`3 if not for Phase 2 clinical studies.
`Serendipity should always be exploited. (As Pasteur said, “Chance favors the
`prepared mind.”) For example, the desired dibenzoylated ribonolactone epimer crys-
`tallized and was converted to gemcitabine (Figure 1.3). The benzoyl protecting group
`had been selected for ease of following the reactions by HPLC [58]. Another example
`is the fortuitous epimerization of the methine alpha to the ketone, shown in Figure 1.3.
`After condensation with the ester acid the trans:cis ratio was 35:64, but the ratio
`improved somewhat during acidification for workup. A study showed that the best
`selectivity for the trans-isomer occurred after the basic aqueous phase was held at
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 8 of 86
`
`

`

`Chapter | 1 Introduction
`
`7
`
`TABLE 1.3 Using GO/NO GO Decision Points in Process R&D (1)
`
`GO/NO GO Studies
`
`Amount of API Required Suitable Route
`
`Toxicology studies
`Bioavailability (Phase 1)
`
`Low
`Low
`
`Efficacy (Phase 3)
`
`High
`
`Expedient: consider using
`discovery route and
`preparative chromatography.
`SAFETY concerns may be addressed
`using routine laboratory operations.
`Scale-up can be grueling with only
`minimal experience
`
`SAFETY concerns must be thoroughly
`addressed for scale-up
`Cost-effective synthesis desired
`Controls needed for reliability and filing
`Productive operations desired
`
`(1) Grabowski, E. J. J., Reflections on Process Research II, in Fundamentals of Early Clinical Drug Development;
`Abdel-Magid, A. F.; Caron, S., Eds.; Wiley: Hoboken, NJ; 2006; Chapter 1.
`
`H3C
`
`CH3
`O
`
`O
`
`F
`CO2Et
`
`F
`
`OBz
`
`1) CF3CO2H,
` H2O, CH3CN
`2) Δ, -EtOH
`
`HO
`
`O
`
`BzO
`
`F
`
`1) BzCl
` pyr
` DMAP
`
`2) CH2Cl2,
` heptane
`
`O
`F
`
`4) -40 ºC, H2O
`5) MTBE
`
`organic
`phase
`
`1) (i-Pr)2NH (2.5 eq.)
` THF, -40 ºC
`2) n-hexLi (4.54 eq.)
` - 40 to 3 ºC
`
`3) (1.29 eq.)
`CO2H
` 36 ºC /
` 1.5 hr
`H3CO2C
`
`CH3
`
`HN
`
`O
`
`N
`
`BzO
`
`O
`
`O
`F
`
`BzO
`BzO
`
`+
`
`O
`
`O
`F
`
`F
`BzO
`crystallized
`
`CO2Li
`
`F
`dissolved in
`mother liquor
`
`1) 5 M HCl to
` pH 5.6 - 6.0
`2) hold 0.5 h
`3) 6 M HCl to pH 5.0
`4) H2O wash
`
`CO2H
`
`HN
`
`LiO
`
`N
`
`O
`held at 20 ºC / 8 h
`
`O
`
`HN
`
`O
`
`N
`
`(80%,
`trans:cis 95:5)
`
`FIGURE 1.3 Two examples of serendipity that were exploited.
`
`room temperature for 8 hours [59]. Researchers exploit observations such as these to
`optimize their processes.
`
`II. EQUIPMENT CONSIDERATIONS ON SCALE
`
`Basic operating differences for equipment used on scale are centered on the fact that
`the equipment is opaque and immobile, and there is limited access to the contents. Most
`multi-purpose reactors (Figure 1.4) are made of glass-lined steel, as that material is more
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 9 of 86
`
`

`

`8
`
`Practical Process Research and Development
`
`FIGURE 1.4 Section view of a cylindrical 1000 gallon multi-purpose glass-lined reactor. Heat transfer fluid
`can circulate inside the jacket for cooling/heating capability. Agitator blades are exchangeable. Courtesy of
`De Dietrich Process Systems.
`
`inert than metal. Baffles, vertical blades near the reactor walls, are used to increase the
`turbulence of the agitated mixtures. Once vessels in a pilot plant or manufacturing have
`been charged they are rarely opened; this is to ensure SAFE operations and to prevent
`contamination of the batch. All mixtures must be stirrable for transfer through a bottom
`valve. The agitator is routinely raised above the bottom surface of the reactor, and
`volumes below the agitator cannot be stirred. The minimum agitation volume (Vmin) is
`about 10% of the nominal reactor volume. The maximum volume a vessel can contain
`(Vmax) may be 100–110% of the nominal reactor volume. In a large vessel heat transfer by
`circulating heat transfer fluids through an external jacket is slower, due to the high
`volume and relatively low surface area of the reactor. A concern of using glass-lined
`reactors is that they may fracture due to thermal shock; other reactors may be preferred
`for cryogenic or high-temperature processes. Childers has presented an excellent pictorial
`perspective of large-scale syntheses from the standpoint of a chemical engineer [60].
`Most scale-up costs are associated with mass transfer and heat transfer. Heat
`transfer affects SAFETY and impurities. In semi-batch operations, heat transfer rate
`often determines the time required for operations. Continuous operations may be
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 10 of 86
`
`

`

`Chapter | 1 Introduction
`
`9
`
`are
`Most operations on scale
`possible if time is not an issue, and
`money is not an issue. Rarely are
`there excesses of either. The best
`approach to develop processes that
`can be readily scaled up is to
`mentally scale down operations to
`the laboratory, and mimic them in
`the laboratory.
`
`chosen for fast reactions due to improved
`heat transfer rates on scale. Mass transfer
`often limits the time for operations. Some
`operations cannot be readily speeded up,
`e.g., the time needed to separate immiscible
`phases during extractive work-ups, or the
`time required to transfer
`liquids through
`a 1-inch line. Efficient mixing is needed
`for heterogeneous reactions, such as liquid/
`liquid, gas/liquid, solid/liquid, and multiply
`heterogeneous processes. Efficient mixing
`may be needed for homogeneous reactions
`also; the rate and mode of addition can be crucial, and controlling micromixing and
`mesomixing may be necessary to minimize impurity formation in fast reactions.
`The limitations of heat transfer through external cooling of spherical flasks in the
`laboratory and cylindrical multi-purpose reactors are noticed with the scale-up of batch
`operations. For cylinders and spheres as the radius increases the volume increases faster
`than does the surface area, hence the ability to remove heat by circulating external fluids
`progressively decreases as the volume increases. For a 10-fold scale-up about twice the
`amount of time will be required [61]. Hence rapid additions on scale can lead to uncon-
`trollable exotherms, with degradation of the reaction mixture and possible runaway reac-
`tions. Dose-controlled additions are used to moderate exothermic response from additions.
`On scale additions typically require 20 minutes to hours; pumps are available for extended
`addition times. Longer operations are not necessarily bad for scale-up. Smaller reactors,
`frequently used for continuous operations, provide more rapid heat transfer.
`The scale-up of a Swern oxidation (Figure 1.5) illustrates the extended time required
`for heat transfer and the impact on processing. The benchmarks for the oxidation were the
`laboratory conditions with additions at 15
`
`C. Unfortunately when these temperatures
`were employed in the pilot plant the isolated
`yield was about 60% of the expected yield. In
`the pilot plant about 2 hours was required to add
`both oxalyl chloride and triethylamine, allow-
`ing time for
`the reactive intermediates to
`decompose. By making the additions at 40
`
`C
`the expected yield was achieved [62]. While the
`additions at the colder temperature may not
`have been any faster, the colder temperature
`slowed the decomposition of
`the reactive
`intermediates.
`The Swern oxidation in Figure 1.5 was successfully scaled up later through semi-
`continuous processing, affording essentially the same yield, but with improved process
`
`C, in contrast to the
`control [63] (Figure 1.6). The reaction was carried out at about 40
`colder temperatures used for batch oxidations. The higher reaction temperature was
`made possible by the extremely short residence time (s) of 0.1 seconds for the reaction
`stream in the reactor. (This is an example of the high-temperature short-time (HTST)
`practice for continuous operations, as is used for pasteurizing milk [64].) The reactor
`
`For rapid scale-up, design homoge-
`neous reaction conditions whenever
`possible. Check the stability of
`process
`streams under extended
`conditions in the laboratory before
`scale-up, because processing on
`scale will be extended.
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 11 of 86
`
`

`

`10
`
`Practical Process Research and Development
`
`O
`
`Ph
`
`N
`
`N
`
`O
`
`CO2CH3
`
`O
`
`H+
`
`- H2O
`
`CHO
`
`HN
`
`O
`
`CO2CH3
`
`O
`
`Ph
`
`O
`
`N
`
`OH
`
`1. DMSO, (COCl)2 /
` CH2Cl2
`2. alcohol
`
`3. Et3N / CH2Cl2
`CO2CH3
`
`HN
`
`O
`
`O
`
`Ph
`
`O
`
`N
`
`addition
`temperatures
`
`yield
`
`comments
`
` 2
`
` h each to add (COCl)2 and Et3N allowed
` intermediates to decompose
`faster additions gave higher yield
`
`-15 ºC
`
`-15 ºC
`
`-40 ºC
`
`55%
`
`31%
`
`52%
`
`scale
`
`laboratory
`
`pilot plant
`
`pilot plant
`
`FIGURE 1.5 Scale-up of a Swern oxidation through batch operations.
`
`+ CO + CO2
`
`CH3
`
`H
`
`Cl
`
`OS
`
`H3C
`
`Ph
`
`HO
`
`HN
`
`O
`
`CO2CH3
`
`O
`
`Ph
`
`N
`
`(~220 kg)
`
`transferred into
`Et3N (5.8 eq.)
`at -10 to 10 ºC
`
`HN
`
`O
`
`CO2CH3
`
`O
`
`N
`
`O
`
`CHO
`
`HN
`
`O
`
`CO2CH3
`
`O
`
`O
`
`Ph
`
`N
`
`O
`+ DMSO (10.8 eq.)
` [in CH2Cl2]
`
`O
`
`Cl
`
`Cl
`
`O
` (1.5 eq.)
`+ N2 sweep
`Ph
`
`O
`
`N
`
`N
`
`O
`
`O
`
`CO2CH3
`(61%)
`
`2.1 in ID pipe,
`31.5 in long,
`τ = 0.1 sec
`outlet ~ 40 ºC
`
`1) aq. washes
`2) TFA, Δ
` - H2O azeotrope
`3) Δ, - CH2Cl2
`4) + n-BuOAc
`
`O
`
`NH2
`
`NaOCl
`2 NaOH
`
`H2O
`
`Cl
`NH
`
`O
`
`NaOH
`
`O
`
`C
`
`N
`
`O
`
`H2O
`
`HN
`
`OH
`
`- CO2
`
`NH2
`
`CO2H
`
`O
`
`N
`
`F N
`
`HN
`
`O
`
`HN
`
`NH
`
`ciprofloxacin
`
`Physical separation of process
`streams minimizes urea formation
`
`X
`
`FIGURE 1.6 Large-scale manufacturing by continuous operations.
`
`
`bends in it to increase turbulence for
`was no more sophisticated than a pipe with two 90
`efficient mixing [65]. An example of large-scale manufacturing using continuous
`operations is the preparation of cyclopropylamine, a component of ciprofloxacin and
`some agrichemical products. The Hofmann rearrangement is carried out by continuous
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 12 of 86
`
`

`

`Chapter | 1 Introduction
`
`11
`
`inputs
`
`Batch and semi-batch operations
`
`elapsed
`time
`
`product
`(batch)
`
`product
`(semi-batch)
`
`Inputs
`
`Two continous processing options: CSTR and PFR
`
`continuously
`stirred tank
`reactor
`(CSTR)
`
`plug
`flow
`reactor
`(PFR)
`
`Inputs
`
`Inputs
`
`optional heat
`exchanger
`
`small reactor,
`controlling mixing &
`temperature
`
`Product
`
`Product
`(semi-continuous)
`
`FIGURE 1.7 Batch and semi-batch operations, and continuous operations.
`
`operations with either continuously stirred tank reactors [66] or static mixers [67]
`(Figure 1.7). Chemical engineers in the fine chemicals industry often have a great deal of
`experience with continuous operations.
`The two primary options for operations, batch and semi-batch (or semi-continuous)
`operations, and continuous operations, are shown schematically in Figure 1.7 [68].
`Continuous operations have been used on scale for decades and are slowly being
`incorporated into processes for APIs. Continuous operations are particularly useful
`for fast reactions, those that take place in seconds or minutes as opposed to hours.
`One attractive benefit of these operations is the SAFETY afforded by conducting
`a reaction with only a small amount of a reaction stream at any given moment
`(Chapter 14). Continuous operations are part of process intensification efforts,
`efforts to improve the space–time productivity of processing [69]. Federsel has
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 13 of 86
`
`

`

`12
`
`Practical Process Research and Development
`
`estimated that continuous operations are applied in 10–20% of the processes in the
`fine chemicals and pharmaceutical industry [70]. Roberge and coworkers feel that
`50% of the processes used in the pharmaceutical
`industry could benefit from
`continuous operations [71]. Chris Dowle, of the Center for Process Innovation, has
`stated that “Continuous processing can reduce operating expenses by at least 90%
`and capital expenses by at least 50%” [72]. Continuous operations are not appro-
`priate for every process, but sometimes are the only practical approach to make
`more material.
`
`III. OPERATIONS PREFERRED ON SCALE
`
`Perhaps the first advice that a seasoned industrial process chemist will give to
`someone from academia is that it is not necessary to dry extracts over desiccants such
`as Na2SO4 or MgSO4. This operation may be unnecessary if water is removed as an
`azeotrope during the concentration. As detailed in Table 1.4, a considerable amount
`of time may be expended in drying over desiccants. The cost of this operation can be
`calculated in several ways. The total time required of 16 operator-hours may be
`conservative; this assumes there are no problems, and no supervision is required.
`During the stir-out time of 2 hours other tasks can be undertaken, but the remaining
`steps require some attention. Considering that the cost of plant time may be as high as
`$500–750/h, a substantial amount of money may be spent unnecessarily. But
`
`TABLE 1.4 Estimated Processing Time on Scale for Na2SO4
`Drying Step
`
`Operation
`
`Set up and test filter
`
`Charge Na2SO4
`
`Stir suspension
`
`Filter off solids
`
`Deliver rinse to site
`
`Apply rinse to solids
`
`Suction dry solids
`
`Pack up solids for disposal
`
`Rinse for “SAFE to clean”
`
`Clean equipment
`
`Test equipment for cleanliness
`
`Store equipment
`
`Total
`
`Estimated Operator-hours
`
`2
`
`0.5
`
`2
`
`2
`
`0.5
`
`0.5
`
`0.5
`
`0.5
`
`0.5
`
`4
`
`2
`
`1
`
`16
`
`IPR2020-00770
`United Therapeutics EX2017
`Page 14 of 86
`
`

`

`Chapter | 1 Introduction
`
`13
`
`foremost, as mentioned earlier, this operation may be redundant if a water-solvent
`azeotrope removes water while the extract is being concentrated. The cost of an
`unnecessary step is opportunity cost, for the opportunity lost to accomplish some-
`thing else or learn something else. Hence on scale extracts are rarely dried over
`desiccants such as Na2SO4.
`Table 1.5 compares operations run in the laboratory and on scale. These are
`guidelines suggested for convenient, productive, and cost-effective operations, but most
`of the operations can be run on scale if desired.
`
`TABLE 1.5 Comparison of Process Operations in the Laboratory and on Scale
`
`In the Laboratory
`
`In Stationary
`Equipment, at Least
`50 L Volume
`
`Process
`Operation
`
`Commonly
`Carried Out
`
`Easy
`to do
`
`Commonly
`Carried Out
`
`Easy
`to do
`
`Comments
`
`Drying extracts over
`desiccants, e.g.,
`Na2SO4, MgSO4
`
`Concentrating
`to dryness (1)
`
`Triturating, lixiviating
`(4)
`
`Use of highly
`flammable solvents,
`e.g., Et2O
`
`Decanting
`
`Column
`chromatography
`
`Rapid transfers
`
`Maintain cryogenic
`temperature
`
`Extended additions
`
`Maintain constant
`pH
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X (2)
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X (5)
`
`X
`
`X
`
`X
`
`(3)
`
`X
`
`X
`
`X
`
`X
`
`X
`
`X
`
`Often redundant if
`concentrating removes
`H2O b