Chemical Process Research and Development in

the 21st Century: Challenges, Strategies, and

Solutions from a Pharmaceutical Industry
Perspective
HANS-JÜRGEN FEDERSEL*
Global Process R&D, AstraZeneca, 151 85 Södertälje, Sweden
RECEIVED ON DECEMBER 5, 2008

CON SPECTUS

I n process research and development (PR&D), the gen-

eration and manipulation of small-molecule drugs
ranges from bench-scale (laboratory) chemistry to pilot
plant manufacture to commercial production. A broad
range of disciplines, including process chemistry (organic
synthesis), analytical chemistry, process engineering (mass
and heat transfer, unit operations), process safety (chem-
ical risk assessment), regulatory compliance, and plant
operation, must be effectively applied. In the critical han-
dover between medicinal chemistry and PR&D, compound
production is typically scaled up from a few hundred
grams to several kilograms. Can the methodologies applied
to the former also satisfy the technical, safety, and scal-
ability aspects that come into play in the latter? Occasion-
ally, the transition might occur smoothly, but more often the situation is the opposite: much work and resources must be invested
to design a process that is feasible for manufacturing on pilot scale and, eventually, for commercial production.
Authentic examples provide enlightening illustrations of dos and don’ts for developing syntheses designed for round-flask oper-
ation into production-scale processes. Factors that are easily underestimated or even neglected in the laboratory, such as method
robustness, chemical hazards, safety concerns, environmental impact, availability of starting materials and building blocks in bulk
quantities, intellectual property (IP) issues, and the final cost of the product, will come into play and need to be addressed appro-
priately. The decision on which route will be the best for further development is a crucial event and should come into focus early
on the R&D timeline. In addition to scientific and technical concerns, the parameter of speed has come to the forefront in the phar-
maceutical arena. Although historically the drug industry has tolerated a total time investment of far more than 10 years from
idea to market, the current worldwide paradigm requires a reduction to under 10 years for the specific segment covering pre-
clinical development through launch. This change puts enormous pressure on the entire organization, and the implication for PR&D
is that the time allowed for conducting route design and scale-up has shrunk accordingly. Furthermore, molecular complexity has
become extremely challenging in many instances, and demand steadily grows for process understanding and knowledge gener-
ation about low-level byproduct, which often must be controlled even at trace concentrations to meet regulatory specifications (espe-
cially in the case of potentially genotoxic impurities). In this Account, we paint a broad picture of the technical challenges the PR&D
community is grappling with today, focusing on what measures have been taken over the years to create more efficiency and
effectiveness.

Introduction recent decades and is now regarded as a core

The design and development of syntheses and capability in the pharmaceutical industry, requir-
processes for making small molecule drugs on ing high scientific and technical skills.1 It is now
scale has experienced considerable attention over fully integrated in the value chain operating in

Published on the Web 04/01/2009 www.pubs.acs.org/acr Vol. 42, No. 5 May 2009 671-680 ACCOUNTS OF CHEMICAL RESEARCH 671
10.1021/ar800257v CCC: $40.75 © 2009 American Chemical Society
Chemical Process R&D in the 21st Century Federsel

pharma R&D (Figure 1) that starts with an idea of how a dis-

ease might be addressed and ends with the launch of a novel
drug several years later.2
The mission of a PR&D organization can be characterized
by at least two areas of responsibility. These are comprised of
(i) the design and development of scalable processes that ful-
fill criteria such as being short and atom-efficient, cost-effec-
tive, technically robust, and environmentally considerate and
(ii) the manufacture of active pharmaceutical ingredients (APIs)
to support drug projects from around the time of candidate
drug (CD) nomination and throughout the entire remaining
FIGURE 1. Overview of the value chain in pharmaceutical R&D
R&D phase of a project. Besides these core accountabilities,
with special emphasis on the interface with process R&D. Key
PR&D has the task to deliver a comprehensive documenta- activities on the timeline are indicated alongside critical milestones
tion describing the production method and how it was and trigger points.
derived, as part of the CMC (chemistry manufacturing and con-
trol) section of regulatory files. Finally, in many companies
PR&D retains the responsibility for the entire life cycle man-
agement of the commercial processes, which involves perfor-
FIGURE 2. Ebalzotan, a molecule posing considerable challenge for
mance monitoring and the replacement of “old” synthetic and
manufacturing on large scale.
process technologies with novel ones. The focus of this
Account is to address specific problems and challenges cur- volunteers). The main reasons for this is that the compound
rently facing PR&D and how they are addressed to enable the has only been made on laboratory scale (few hundred grams)
manufacture of the desired compounds using safe and effec- up to this point, which, considering the fact that not only the
tive production methods. target molecule but also some of the synthetic intermediates
are novel, means that there is no precedence on how to best
When Does It All Begin? ensure successful production. Moving fast at this early stage
In the “classical” approach to developing new drugs, the func- of PR&D work means high risk-taking. The reason being that
tions operated in isolation rather than involving the teams the big step progressing chemistry from laboratory to plant
from other departments. Such a business model was seen as scale is conducted at a point when there is still relatively poor
sufficiently effective, inasmuch as the pressure on shortening knowledge about key features of the process (that is, kinet-
time lines was rather low. Spending 15 years taking an idea ics, critical parameters, workup procedures, byproduct forma-
through to market launch was seen to be the norm, rather tion, and conditions for crystallization and isolation). An
than something that could be improved. As productivity in the illustration to this effect and at the same time a hard-learned
pharma industry has come into the limelight during the past lesson is offered by the ebalzotan case (Figure 2). This mole-
decade, this paradigm has been strongly challenged and cule, chosen as a candidate in the early 1990s, showed
today a common driver across the business is to arrange the selective 5HT1A-agonistic properties and a promising pharma-
workflow in such a way that the sequence of activities is much cological profile, which lead to its development as an antide-
faster; <10 years from start of preclinical documentation to fil- pressant (later discontinued due to side effects in healthy
ing for regulatory approval and launch is an achievable tar- volunteers). PR&D involvement did not start until the com-
get in most disease areas. pound selection had been made, and the request was to pre-
This change in the operating mode has affected PR&D a pare a first batch on a scale of about 1 kg in the shortest
great deal, because the supply of the required APIs is com- possible time. Given this minimal lead-time and being faced
pressed over a shorter time span, which requires a more expe- with a medicinal chemistry sequence of 13 linear steps, the
dient scale-up and manufacture. Experience shows that the opportunities for neither a major redesign of the route nor
delivery of the first batch of a new chemical entity (NCE) on finding the time to conduct a detailed scrutiny of the meth-
scale, normally meaning 1-5 kg, is on the critical path as this ods from a large-scale operability perspective were simply
material will have to support extended toxicological and for- available. Starting the production with provisional batch sheets
mulation studies, alongside phase I trials in humans (healthy and solving problems as they occurred (there were many of

672 ACCOUNTS OF CHEMICAL RESEARCH 671-680 May 2009 Vol. 42, No. 5
 Chemical Process R&D in the 21st Century Federsel

SCHEME 1. Formylation Chemistry: The Way in Which the

Lithiating Agent Is Added Has a Dramatic Effect on the Isolated
Yielda

a
Abbreviations: DMF, N,N-dimethylformamide.

them!), the result was that after 5 months of hard work the
material finally isolated did not mount to more than 80 g
(0.25% overall yield).1g,3 With this catastrophic outcome, an
amount delivered that was far less than expected and consid- FIGURE 3. The medicinal chemistry-process R&D interface.
erably later than planned, virtually the entire project came to Opportunities for efficiency gains, especially in speed of delivery of
first batch on scale by front-loading work into LO and
a halt. To mention but one example from this tour de force prenomination phases.
where the outcome in the pilot plant was substantially below
the yield achieved in laboratory experiments, the Li-mediated
formylation constituting the first step provides a good illustra-
tion (Scheme 1). Based on the laboratory findings, the first ver-
sion of the pilot plant method demanded the addition of BuLi
(hexane solution) to be conducted in such a way that the feed
stream was entering the vessel (600 L) above the surface of
the reaction solution. Using this procedure, the yield shrunk to
half of the expected, generating only 38% of the desired alde-
hyde compared with >70% obtained in laboratory runs. This
meant that more material had to be processed in this step
and, furthermore, that the purity profile deteriorated, which
proved to have a negative impact later in the process. In sub- FIGURE 4. Attrition is the worst enemy to the development of new
drugs, and in the phase from preclinical through first studies in
sequent studies, the poor result was traced back to the forma- humans, a statistical average of 2 out of 3 projects are terminated.
tion of local hot spots in the reaction mixture, and the solution (Source: Centre for Medicines Research International 2006/2007,
devised to avoid this from happening was to redesign the Pharmaceutical R&D Factbook).
charging device allowing the BuLi reagent to enter beneath can be gained from these early interactions that will enable a
the surface of the process solution. This trivial change ensured trained process chemist to decide when to initiate his own
a much more efficient dispersion of the reagent in the bulk of activities such as experimental work, ordering of starting mate-
the solution, which avoided formation of byproduct. rials, applying technical feasibility studies, looking for alterna-
Because time is of the essence in this early phase, a way tive synthetic routes, assessing reaction hazards, and so on.
to tackle the situation is to initiate the necessary activities ear- The change in working model thus described has reduced the
lier. This front-loading concept has been in operation in Astra- average time lag from CD nomination (Figure 1) until avail-
Zeneca’s PR&D organization for several years, and the results ability of the disired compound from typically >1/2 year to a
achieved are unambiguous: The successful delivery of the first few weeks.2d,5 It is important to stress that the chemical pro-
batch in pilot quantities is off the critical time line.4 Proper cesses applied in this early stage of a drug project are not in
management of the interface between medicinal chemistry a fully developed and optimized state. At this point, where the
and PR&D is a key to ascertain this outcome and builds to a attrition is extremely high with on the average 2 out of 3 mol-
large extent on the creation of an environment that allows a ecules being lost (Figure 4), mainly because of toxicity in
free flow of information. This way, the contacts can be pushed animals and for DMPK reasons (drug metabolism and
back into the earlier stages of lead optimization (LO), where pharmacokinetics),2d a process that meets critical safety,
there is still uncertainty about the exact structural features health, and environment (SHE) criteria and guarantees that the
characterizing the potential CD (Figure 3). However, enough API can be manufactured to the right quality attributes is good

Vol. 42, No. 5 May 2009 671-680 ACCOUNTS OF CHEMICAL RESEARCH 673
Chemical Process R&D in the 21st Century Federsel

FIGURE 5. Structural diversity as exemplified by drug molecules from the AstraZeneca portfolio.

enough. The high failure rate results in only 1 of 3 com-

pounds having to be produced a second time, and hence, it
is more rewarding to focus efforts on surviving projects rather
than spending on those that are close to becoming discontin-
ued. A question immediately comes to the fore: how to pick
the winners? In a constrained environment (money and peo-
ple), a mechanism that would identify projects standing the
highest chance of success would present an enormous asset. FIGURE 6. Source of stereogenic centers as found in a cross-
An in-depth cross-functional analysis on, for example, a dis- pharma survey6 analyzing a combined portfolio of 69 chiral
molecules.
ease area portfolio level or on individual projects will map
threats and risks on one hand and compare them with oppor- prepare a number of drug candidates.6 The survey covered
tunities, as represented by the disciplines involved. Combin- 128 synthetic sequences, representing in total 1039 discrete
ing these data and rating their relative importance will provide transformations. From this, the average number of stages per
at least some guidance on which projects to prioritize when synthesis was found to be just above 8, with a spread in the
allocating resources, with the caveat that this “cherry-pick- actual numbers from 2 to >15. Unsurprisingly, no single reac-
ing” exercise can never offer a 100%-guarantee that the right tion category is entirely dominant, and the highest score was
choice has been made. For PR&D, listening to customers and achieved by what has been described as heteroatom alkyla-
various experts will, ultimately, exert a strong influence on tion and arylation with 19%. Other areas occupying relatively
which work should be front-loaded and which can be deferred high prevalence were deprotections (15%), acylations (12%),
to a later start. C-C bond formations (11%), and interconversions of func-
tional groups (10%).
Making Molecules: An Expanding Toolbox A further lesson established by this investigation was the
The space describing the universe of small drug molecules ways in which stereogenic centers were installed7 in the 54%
(typically with molecular weight <1000 Da) shows a wide vari- of the cohort of molecules that displayed chirality (Figure 6).
ation, with new architectures being continuously added (Fig- Thus, the conclusion is that in a pharmaceutical industry sce-
ure 5). To successfully address the challenging task of nario >50% of the cases will rely on supply of the required
assembling these often highly complex compounds, it is stereochemically defined building blocks from external
essential to have access to methodologies capable of achiev- sources. With the wealth of both simple and more sophisti-
ing the synthetic transformations required. In an effort to cated enantiomerically pure compounds available com-
describe the as is situation within the pharmaceutical indus- mercially, many of them in bulk, and new ones being contin-
try, a mapping analysis was recently undertaken by three uously added, this state of the art should not come as a big
major drug companies in the UK (GlaxoSmithKline, Pfizer, and surprise. It might seem more unexpected that “classical” res-
AstraZeneca) of what specific reaction types had been used to olution (28%) still assumes a very important position. The rea-

674 ACCOUNTS OF CHEMICAL RESEARCH 671-680 May 2009 Vol. 42, No. 5
 Chemical Process R&D in the 21st Century Federsel

SCHEME 2. En route to Nexium, One of the Most Significant Drugs: The Bulk Production of the Active Ingredient Is Conducted via an
Asymmetric Ti-Catalyzed Sulfoxidation, Operating with Turnover Numbers of 4-16 and Turnover Frequencies of 3-12 h-1

son for this resides in the vast experience gathered from tal principles applied to the workflow in our PR&D organiza-
applying this technology over many decades and its reputa- tion is provided by the speed-quality-cost triangle (Figure
tion as a robust, reliable, and scale-up friendly methodology. 7).2d This model offers guidance on where to concentrate
Asymmetric synthesis, enjoying high attention and prestige in efforts as the project is progressed along the time axis. Start-
academia, is only applied in one out of 10 cases; a fact largely ing with speed, focusing on this parameter will ensure faster
attributed to the dominating perception that design of cata- cycle times, which going forward eventually brings quality into
lytic processes is complex, hugely effort consuming, and the limelight as an element to underpin a better pipeline of
tedious with uncertain outcome.5 A further contributing fac- emerging drugs. Finally, when the critical proof-of-concept
tor is the high attrition of drug projects in the early parts of the stage is successfully passed, cost will get most of the atten-
pipeline (Figure 4), leading to a reluctance to invest in the tion by virtue of its direct link to the creation of a leaner
design and development of such processes. Notwithstanding, organization.
today’s access to a whole host of catalytic procedures, be they The cost component is partly being addressed via the appli-
organometallic or enzymatic, with high specificity and capa- cation of lean methodologies and the current trend of exten-
bility to conduct selective transformations (oxidations, reduc- sive outsourcing into countries with low PR&D/manufacturing
tions, bond formations) robustly and at large scale has brought cost. In the quality area, traditional approaches (rigorous ana-
this whole area to the forefront. Thus, transition metal cata- lytical control) are now increasingly complemented by the
lyzed stereoselective synthesis has, for example, been amply concept of process analytical technology (PAT).13 Speed, as the
demonstrated and validated in numerous production scale first area to concentrate on when a new project is initiated, will
processes,8 the largest of which is the herbicide (S)-meto- most likely impact the whole portfolio of drug candidates as
lachlor9 with an annual manufacturing volume >10 000 t. they appear through the pipeline. A feature closely linked to
From the pharmaceutical sector, esomeprazole (active ingre- this parameter is the complexity presented by the target mol-
dient in AstraZeneca’s antiulcer drug Nexium) constitutes an ecule, which should not be understood purely in terms of
excellent example where a Ti-catalyzed asymmetric sulfide structural challenges but also as the assessment of the num-
oxidation is carried out on a scale of 100 t per annum ber of chemical stages required to make the product. There-
(Scheme 2).1g,10 fore, estimating the workload created by an entire portfolio of
new drugs has to be done on the basis of counting the dis-
Continuous Improvement: Adapting to crete number of steps rather than the amount of individual
Challenging Times projects, an insight often overlooked that has direct implica-
The intense drive to reduce cost in virtually all areas that the
entire pharmaceutical industry is being faced with today will
require new working practices. For PR&D, this means that
efforts will be invested only to an extent where critical API
deliveries or crucial route discovery/development work will
not be jeopardized. This approach is in line with the manage-
ment principles of lean six sigma.11 Instead of devoting efforts
on non-value-adding activities or even on what rightly can be
characterized as waste,12 the focus should be to support inno-
vation by giving priority to work that helps develop a prod-
FIGURE 7. Using the speed-quality-cost triangle as guidance for
uct or a process more quickly, with better quality, and where to put main efforts as drug projects progress along the R&D
spending less resources. A visualization of these fundamen- timeline.

Vol. 42, No. 5 May 2009 671-680 ACCOUNTS OF CHEMICAL RESEARCH 675
Chemical Process R&D in the 21st Century Federsel

SCHEME 3. An Electrophilic Fluorination on a Substituted Chroman Moiety, Using (PhSO2)2NF (NFSi) as Source of F+, a Reagent That
Contributes with Just 7% of Its Molecular Weight

tions on resource allocation. Scrutinizing a particular synthe- just one of these areas, it will not be seen as something that
sis from start to finish should, however, not stop at this point has an overall positive impact on sustainability. This raises
but include a thorough investigation of “classical” parame- complicated questions where the consequences of short- vs
ters such as yield per step, whether the route is linear or con- long-term and regional vs global decisions have to be add-
vergent, dilution in process streams, materials throughput, and ressed.
foreseeable issues with scalability. The latter is particularly cru- Introducing a fluorine substituent in an intermediate en
cial when the plan is to operate on larger scale, because it is route to the antidepressant robalzotan18 (Figure 5) provides a
absolutely necessary to take hazards and safety aspects into nice example of a chemistry that is far from being atom effi-
serious account to avoid unacceptable risks to staff and the cient. Thus, only a tiny fraction of the molecular weight (7%)
environment. of the reagent NFSi is utilized when making the product
(Scheme 3). Aiming for a more optimized procedure that
En Route to Green Processes
avoids technically complex fluorination technology (severe risk
Is it possible to apply an entirely objective judgment when try-
of corrosion), a starting material with the F-substituent in place
ing to assess the quality of chemical processes? The answer
was identified as a much more effective strategy. In a prom-
to this question is most probably no, but attempts have been
ising attempt to capture “downside” features of this sort as well
made to offer a rational approach to the assessment of routes.
as provide an overall assessment of the way in which a given
Thus, under the acronym SELECT, a cross pharma industry
molecule is synthesized, a numerical tool focusing on struc-
group in the UK has listed a number of parameters, safety,
tural intricacy has been designed that allows a revealing side-
environmental, legal, economics, control, and throughput,
by-side comparison of different approaches.19
each specified by a number of subcriteria that ultimately will
When the elegance (or lack thereof) in a synthetic route is
reveal strengths and weaknesses of a chosen process and pro-
discussed, the total number of steps required often constitutes
vide guidance on where changes are required to improve on
the measure of quality. The translation of a synthesis to a pro-
existing procedures.14 With new and improved tools becom-
ing available, for example, in the form of novel catalysts or cess will, however, not always show the same outcome in the
innovative reactor designs, processes that previously were sense that fewer steps, by default, will create a more effec-
seen as “best in class” can become obsolete or outdated. Not tive and efficient production method. Instead, technical aspects
least, the green chemistry paradigm has played an important come to the fore exemplified by the number of unit opera-
role in changing people’s minds and thoughts on how chem- tions and solvent swaps required, the level of dilution during
ical processes should be constructed in the future. The foun- reactions and workup, demands on materials of construction,
dation of green chemistry15 rests on 12 general principles, for need for special (nonstandard) equipment, and the amount of
example, elimination of waste, atom economy,16 reduction of the final material that can be produced at maximum (opti-
risk, minimization of energy consumption, renewable raw mum) capacity in a given reactor setup. Initiatives directed
materials, catalysis, and biodegradability. More recently, this toward improvements in the processing area at large have
concept has been further expanded and refined into the con- been branded process intensification, and one technology that
text of sustainability, which, besides including the “classical” has received much attention recently is reactors operating
green features, operates broadly under the definition of “meet- under continuous flow mode.20 Moving into systems of this
ing the needs of the current generation while preserving the kind will inevitably bring in a new aspect to pharmaceutical
ability of future generations to meet their needs”.17 When one production, where the tradition has been strongly in favor of
analyzes sustainability, three areas can be defined, environ- running batch-wise. However, the current view supported by
mental, economic, and societal, against which any action or experimental evidence is that continuous processing is an
change should be assessed. Thus, if an activity is advancing option in only 10-20% of the reactions normally applied in

676 ACCOUNTS OF CHEMICAL RESEARCH 671-680 May 2009 Vol. 42, No. 5
 Chemical Process R&D in the 21st Century Federsel

penultimate intermediate constituted by the hydroxy ester,

which would avoid chromatography. A possibility was identi-
fied to perform the first step in the sequence, generation of the
cyanohydrin, in an asymmetric fashion. The downside of this
FIGURE 8. Chiral tetralin-substituted R-hydroxy carboxamides, a
class of pharmacologically active molecules. idea was, however, that at the time only limited precedence
was available indicating that such a reaction should be suc-
fine chemicals and drug manufacture, and hence, both cessful with ketones. Moving to the final category, conduct-
approaches will be evaluated case by case. ing a brainstorm exercise to address how this class of target
molecules could be accessed resulted in the identification of
The Big Leap: Taking Syntheses to Scale what was characterized as a winning route (Scheme 5).
When scale up of chemical transformations is considered, a Interestingly, the starting material (7-methoxy-1-tetralone)
number of aspects need to be addressed as outlined above. is identical to the one used by medicinal chemistry, but then
An effective way is to establish mechanisms allowing PR&D to the routes diverged. In the novel process, carbon-chain exten-
interact as early as seems sensible with groups in medicinal sion was effected by a Wittig reaction where a phosphonium
chemistry. The gains obtained by this collaborative approach ylide is added to the keto function. The nonisolated exocy-
are illustrated by an authentic case,21 where the task was to clic olefin was then subjected to an oxidation under “classi-
produce 2 kg of a novel API. Thus, during the LO phase, sev- cal” Sharpless conditions, where an Os-based dihydroquinidine
eral potential candidate molecules were identified that, catalyst results in an asymmetric dihydroxylation across the
besides being chiral, shared a common structural motif in the double bond in good to excellent yields, installing the cor-
form of an R-hydroxy amide moiety (Figure 8). In a first step, rect stereochemistry at C-1. A smooth Pt-catalyzed oxidation
300 g of each of a small number of compounds was required using O2 present in the atmosphere rendered the R-hydroxy
in order to conduct further preclinical investigations that, acid as final intermediate, which after coupling with the per-
finally, would yield one CD. A synthetic strategy was con- tinent amine afforded the end product as the desired anti-
ceived in medicinal chemistry where a linear sequence led to pode in an overall yield ranging from 42-55%. The method
a racemic R-hydroxy carboxylic acid, which, subsequently, had was still a linear sequence, but with fewer steps than the orig-
to be resolved using preparative chromatography after trans- inal method and transformations that all operate at very high
forming the acid to the corresponding amide using an enan- chemical (>70%) and stereochemical yields (>98% ee). A key
tiomerically pure amine (Scheme 4). Acknowledging that some feature is the absence of chromatography, which in itself is
of the steps operate at impressive yields (90%), being forced enormously time-saving. When a side-by-side comparison of
to apply yet another chromatographic purification of a com- the two routes (Chart 1) looking at some key parameters such
plex reaction mixture further upstream and saving the stereo- as total working time spent, staff resource, and amounts
isomer separation to the very final stage offering only 25% needed of starting material (tetralone) and chiral amine is con-
yield made this process seem very unattractive. The overall ducted, it is clearly evident that a change in process is
yield of barely 2%, in combination with high cost and lim- extremely beneficial in all aspects. Specifically, the dramatic
ited bulk availability of the chiral amines of interest, prompted decrease in requirement for supply of tetralone should be
a focused search for alternatives. noted, where the original need is reduced by >90%. The les-
Several options were identified for how to handle the deliv- son provided by this example is that being open to and con-
eries, and these fell into three main categories: (i) the medic- sidering alternative solutions to a problem right at the start is
inal chemistry route is scaled up as is; (ii) the medicinal a worthwhile approach, where PR&D can make a significant
chemistry route is modified; (iii) an entirely novel sequence is contribution.
designed. The first of these was ruled out on the basis of very
long projected manufacturing times, a judgment that a chro- Intellectual Properties: A Key Interface
matographic isolation of the wanted enantiomer would not be The design and development of processes for synthesizing
workable, and fears regarding the timely supply of the chiral novel molecules is, as already alluded to, a highly innova-
amine. In the second scenario looking at the bottlenecks of tive activity. This means that results generated during these
the synthesis, there was a perception that improving on the endeavors will represent intellectual properties (IPs) to vary-
nitrile hydrolysis was entirely feasible. Moreover, an opportu- ing degrees and value for the owner and, hence, might merit
nity was seen to conduct the resolution at the stage of the being patented. In the context of pharmaceutical products, the

Vol. 42, No. 5 May 2009 671-680 ACCOUNTS OF CHEMICAL RESEARCH 677
Chemical Process R&D in the 21st Century Federsel

SCHEME 4. Entry into the Family of Chiral R-Hydroxy Carboxamides Using a Synthetic Route Devised by Medicinal Chemistry (Route A),
Requiring Two Chromatographic Stages, One for Chemical Purification of an Intermediate and One for Separation of Optical Antipodesa

a
Abbreviations: TMSCN, trimethylsilyl cyanide; TBTU, O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate.

SCHEME 5. A Significantly Improved Process (Route B) for Making R-Hydroxy Carboxamides Based on an Asymmetric Dihydroxylation
Protocola

a
Abbreviations: (DHQD)2 · PHAL, hydroquinidine 1,4-phthalazinediyl diether.

CHART 1. Side-by-Side Comparison between Routes A and B used.22 The value in protecting processes has been heavily
(Schemes 4 and 5) of a Number of Process Attributes Leading to 2 debated, and the views on pros and cons go in different direc-
kg of a Single Enantiomer of R-Hydroxy Carboxamides
tions depending on circumstances. Many claim that reveal-
ing the details of a manufacturing process in a patent
disclosure will only trigger competitors to find loopholes allow-
ing them to go outside the claims. Others say that a patented
process gives you control of how to make the compound in
question and will prevent a third party from claiming the rights
to it, which if they did would exclude you from practicing the
invention (i.e., the process). Features of processes that are pat-
entable would be, for example, the order that the individual
steps are conducted, the specific reaction conditions, and the
IP protection aims at covering various aspects of the drug, for use of a particular catalyst or solvent. Advantages over prior
example, the structure of the active ingredient, the formula- art that need to be proven to gain patent rights can be exem-
tion, and the medical indication for which it is intended to be plified by improved yields, improved chemo-, stereo-, or regi-

678 ACCOUNTS OF CHEMICAL RESEARCH 671-680 May 2009 Vol. 42, No. 5
 Chemical Process R&D in the 21st Century Federsel

oselectivity, reduced formation of byproduct, reduction in He has occupied different positions in Astra and AstraZeneca, both
effluent streams or energy usage, or telescoping of several as a scientist and manager, leading to his current roles as Direc-
tor of Science and Senior Principal Scientist. Alongside, he has
steps (one-pot protocol).
pursued a deep involvement with academia that brought him an
If full ownership represents one aspect of IP, then freedom
associate professorship in organic chemistry at the Royal Insti-
to operate is another, where an invention is made publicly tute of Technology, Stockholm, Sweden. For a more comprehen-
available (e.g., published in a journal), which guarantees every- sive bio, see Acc. Chem. Res. 2007, 40 (12), 1377.
one the access as no one can claim the rights to it. This
approach is seeing an increased use in the process arena, a FOOTNOTES
development that has been driven by the insight that polic- * To whom correspondence should be addressed. E-mail: hans-jurgen.federsel@
astrazeneca.com. Telephone: +46 8 553 270 63. Fax: +46 8 553 242 76.
ing abuse or infringement of IP is an exceedingly hard task.
Revealing that someone has copied a molecule is relatively REFERENCES
straightforward. In contrast, to successfully claim that the pro- 1 (a) Repič, O. Principles of Process Research and Chemical Development in the
Pharmaceutical Industry; Wiley and Sons: New York, 1998. (b) Process Chemistry in
cess for making a particular compound uses patented meth- the Pharmaceutical Industry; Gadamasetti, K. G., Ed.; Marcel Dekker: New York,
1999. (c) Anderson, N. G. Practical Process Research & Development; Academic
ods and procedures or proceeds via patented intermediates is Press: San Diego, CA, 2000. (d) Cabri, W.; DiFabio, R. From Bench to Market: The
very difficult, if not impossible. Nonetheless, for ethical phar- Evolution of Chemical Synthesis; Oxford University Press: Oxford, U.K., 2000. (e)
From Bench to Pilot Plant. Process Research in the Pharmaceutical Industry; Nafissi,
maceutical companies, it is crucial to prevent generic manu- M.; Ragan, J. A., DeVries, K. M., Eds.; American Chemical Society: Washington, DC,
2002. (f) Federsel, H.-J. Start small, think big - The art of process R&D. Nat. Rev.
facturers from conducting illegal competition during the time Drug Discovery 2001, 1 (12), 1013. (g) Federsel, H.-J. Logistics of process R&D:
of patent validity, and considerable efforts are devoted to Transforming laboratory methods to manufacturing scale. Nat. Rev. Drug Discovery
2003, 2 (8), 654–664. (h) Lipton, M. F.; Barrett, A. G. M., Guest Editors, Process
enforce the law in this respect. Chemistry, Chem. Rev. 2006, 106 (7), 2581-3027. (i) Rao, C. S. The Chemistry of
Process Development in Fine Chemical & Pharmaceutical Industry, 2nd ed.; Wiley &
Asian Books Private Ltd.: Chichester and New Delhi, 2006. (j) Process Chemistry in
Conclusions and Outlook the Pharmaceutical Industry, Vol. 2, Challenges in an Ever Changing Climate;
Gadamasetti, K.; Braish, T. F., Eds.; CRC Press: Boca Raton, FL, 2008. (k) Org.
In today’s pharmaceutical industry where the speed of devel- Process Res. Dev.; Laird, T., Ed.;this journal is entirely devoted to publishing work
from the process R&D arena since 1997; see journal homepage http://pubs.acs.org/
oping new drugs has become a key concern, process R&D is journal/oprdfk/.
increasingly finding itself standing in the limelight. This has 2 (a) Federsel, H.-J. The integration of process R&D in drug discovery - Challenges
and opportunities. Comb. Chem. High Throughput Screening 2006, 9 (2), 79–86.
led to a challenge of old working principles and a demand for (b) Comprehensive Medicinal Chemistry II, Vol. 2, Strategy and Drug Research;
continuous improvement and increased efficiency. The bal- Moos, W. H., Ed.; Elsevier: Oxford, U.K., 2007. (c) Woodcock, J.; Woosley, R. The
FDA critical path initiative and its influence on new drug development. Annu. Rev.
anced frontloading of work prior to the traditional starting Med. 2008, 59, 1–12. (d) Federsel, H.-J. Handing over the baton: Connecting
medicinal chemistry with process R&D. Drug News Perspect. 2008, 21 (4), 193–
point at CD nomination has generated a strong momentum, 199.
which has resulted in taking the delivery of the first API cam- 3 Federsel, H.-J. Development of a process for a chiral aminochroman antidepressant:
A case story. Org. Process Res. Dev. 2000, 4 (5), 362–369.
paign at scale off the critical path. Chemical processes oper- 4 (a) Federsel, H.-J. Drug discoverers - you need us. Drug Discovery Today 2001, 6
ated at this point will be far from the performance expected (8), 397–398. (b) Federsel, H.-J. The role of process R&D in drug discovery -
Evolution of an interface. Chim. Oggi/Chem. Today 2004, 22 (3/4), 9–12. (c)
of a fully commercially viable method, however, without com- Potoski, J. Timely synthetic support for medicinal chemists. Drug Discovery Today
2005, 10 (2), 115–120.
promising on the safety aspects. The full implementation of
5 Federsel, H.-J. In search of sustainability: Process R&D in light of current
quality by design (QbD)23 and lean six sigma philosophy, cou- pharmaceutical industry challenges. Drug Discovery Today 2006, 11 (21/22), 966–
974.
pled with application of new and innovative technologies,
6 Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Analysis of the reactions
promises to further increase the capability of PR&D to success- used for the preparation of drug candidate molecules. Org. Biomol. Chem.
2006, 4 (12), 2337–2347.
fully respond to forthcoming demands.
7 (a) Federsel, H.-J. Drug chirality - scale-up manufacturing, and control. CHEMTECH
1993, 23 (12), 24–33. (b) Federsel, H.-J. Facing chirality in the 21st century:
The author expresses his deep appreciation to colleagues, past Approaching the challenges in the pharmaceutical industry. Chirality 2003, 15 (S1),
S128-S142. (c) Federsel, H.-J. Searching for scalable processes: Addressing the
and present, in Astra and AstraZeneca, who have been instru- challenge in times of increasing complexity. Curr. Opin. Drug Discovery Dev. 2003,
mental to the successful progression of the work reported in this 6 (6), 838–847. (d) Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.;
Stürmer, R.; Zelinski, T. Industrial methods for the production of optically active
Account. A special gratitude is due to Dr. David Ennis, Process intermediates. Angew. Chem., Int. Ed. 2004, 43 (7), 788–824. (e) Federsel, H.-J.
Chiral Drug Discovery and Development - From Concept Stage to Market Launch.
R&D Avlon/Charnwood, for providing details regarding the Comprehensive Medicinal Chemistry II, Vol. 2, Strategy and Drug Research; Moos,
chiral R-hydroxy carboxamide case story. W. H., Ed.; Elsevier: Oxford, U.K., 2007; Chapter 2.17, pp 713-736.
8 (a) Blaser, H.-U.; Spindler, F.; Studer, M. Enantioselective catalysis in fine
chemicals production. Appl. Catal. A 2001, 221 (1-2), 119–143. . (b) Federsel,
H.-J. Asymmetry on large scale: The roadmap to stereoselective processes.
BIOGRAPHICAL INFORMATION Nat. Rev. Drug Discovery 2005, 4 (8), 685–697. (c) Asymmetric Catalysis on
Industrial Scale. Challenges, Approaches and Solutions; Blaser, H.-U., Schmidt,
Hans-Jürgen Federsel, born 1949 in Säter, Sweden, has E., Eds.; Wiley-VCH: Weinheim, Germany, 2004. (d) Blaser, H.-U.; Pugin, B.;
devoted his entire professional career to the field of process R&D. Spindler, F. Progress in enantioselective catalysis assessed from an industrial

Vol. 42, No. 5 May 2009 671-680 ACCOUNTS OF CHEMICAL RESEARCH 679
Chemical Process R&D in the 21st Century Federsel

point of view. J. Mol. Catal. A: Chem. 2005, 231 (1-2), 1–20. (e) Federsel, H.-J. 15 Anastas, P. T.; Kirchhoff, M. M. Origins, current status, and future challenges of
Stereoselective Synthesis of Drugs s An Industrial Perspective. In Chirality in green chemistry. Acc. Chem. Res. 2002, 35 (9), 686–694.
Drug Research; Francotte, E., Lindner, W., Eds.; Wiley-VCH: Weinheim, 16 (a) Trost, B. M. The atom economy - a search for synthetic efficiency. Science 1991,
Germany, 2006; Chapter 2, pp 29-65.
254 (5037), 1471–1477. (b) Trost, B. M. On inventing reactions for atom economy.
9 (a) Blaser, H.-U. The chiral switch of (S)-metolachlor: A personal account of an Acc. Chem. Res. 2002, 35 (9), 695–705.
industrial odyssey in asymmetric catalysis. Adv. Synth. Catal. 2002, 344 (1), 17–31.
(b) Blaser, H.-U.; Hanreich, R.; Schneider, H.-D.; Spindler, F.; Steinacher, B. The 17 (a) Anastas, P. T.; Zimmerman, J. B. The twelve principles of green engineering
chiral switch of metolachlor: The development of a large-scale enantioselective as a foundation for sustainability. In Sustainability Science and Engineering
catalytic process. Asymmetric Catalysis on Industrial Scale. Challenges, Approaches Volume 1: Defining principles; Abraham, M. A., Ed.; Elsevier: Amsterdam,
and Solutions; Blaser, H.-U., Schmidt, E., Eds.; Wiley-VCH: Weinheim, Germany, 2006; Vol 1, Chapter 2, pp 11-32. (b) Mours, M. SusChem: From vision to
2004, pp 55-70. action. ChemSusChem 2008, 1 (1/2), 59–62.
10 (a) Federsel, H.-J.; Larsson, M. An Innovative Asymmetric Sulfide Oxidation: The 18 Federsel, H.-J.; Sveno, A. To overcome the hurdles: coping with the synthesis of
Process Development History Behind the New Antiulcer Agent Esomeprazole. robalzotan, a complex chroman antidepressant. Process Chemistry in the
Asymmetric Catalysis on Industrial Scale. Challenges, Approaches and Pharmaceutical Industry, Vol. 2, Challenges in an Ever Changing Climate;
Solutions; Blaser, H.-U., Schmidt, E., Eds.; Wiley-VCH: Weinheim, Germany, Gadamasetti, K.; Braish, T. F., Eds.; CRC Press: Boca Raton, FL, 2008; Chapter 7,
2004, pp 413-436. (b) Seenivasaperumal, M.; Federsel, H.-J.; Ertan, A.; pp 111-135.
Szabó, K. J. Factors influencing the selectivity in asymmetric oxidation of
sulfides attached to nitrogen containing heterocycles. Chem. Commun. 2007, 19 Fuchs, P. L. Increase in intricacy - a tool for evaluating organic syntheses.
2187–2189. . (c) Seenivasaperumal, M.; Federsel, H.-J.; Szabó, K. J. Tetrahedron 2001, 57 (32), 6855–6875.
Mechanism of the asymmetric sulfoxidation in the esomeprazole process. 20 Wiles, C.; Watts, P. Continuous flow reactors, a tool for the modern synthetic
Effects of the imidazole backbone for the enantioselection. Adv. Synth. Catal. chemist. Eur. J. Org. Chem. 2008, 1655–1671.
2009, in press.
21 Ainge, D.; Ennis, D.; Gidlund, M.; Stefinovic, M.; Vaz, L.-M. Rapid development of an
11 George, M.; Rowlands, D.; Kastle, B. What is lean six sigma? McGraw-Hill: New enantioselective synthesis of (R)-1-hydroxy-7-methoxy-1,2,3,4-
York, 2004.
tetrahydronaphthalene-1-carboxylic acid. Org. Process Res. Dev. 2003, 7 (2), 198–
12 Sewing, A.; Winchester, T.; Carnell, P.; Hampton, D.; Keighley, W. Helping science 201.
to succeed: Improving processes in R&D. Drug Discovery Today 2008, 13 (5/6),
227–233. 22 Grubb, P. W. Patents for Chemicals, Pharmaceuticals and Biotechnology:
Fundamentals of Global Law, Practice and Strategy; Oxford University Press: Oxford,
13 Romañach, R. J. Research initiatives to support science-based pharmaceutical
manufacturing. Pharm. Technol. Eur. 2007, 19 (10), 47–51. U.K., 2005.
14 Butters, M.; Catterick, D.; Craig, A.; Curzons, A.; Dale, D.; Gillmore, A.; Green, S. P.; 23 Seibert, K. D.; Sethuraman, S.; Mitchell, J. D.; Griffiths, K. L.; McGarvey, B. The
Marziano, I.; Sherlock, J.-P.; White, W. Critical assessment of pharmaceutical use of routine process capability for the determination of process parameter
processes - A rationale for changing the synthetic route. Chem. Rev. 2006, 106 (7), criticality in small-molecule API synthesis. J. Pharm. Innovation 2008, 3, 105–
3002–3027. 112.

680 ACCOUNTS OF CHEMICAL RESEARCH 671-680 May 2009 Vol. 42, No. 5