A Flow Process Built upon a Batch Foundation - Preparation of a Key Amino Alcohol Intermediate via Multistage Continuous Synthesis

Article abstract of DOI:10.1021/acs.oprd.9b00478

This paper describes recent efforts to apply flow technology in the preparation of the key amino alcohol intermediate 3b so as to address manufacturability issues present in the batch process of a PRMT5 inhibitor. The continuous process, one of the first reported pharmaceutical processes to use aqueous NH4OH in flow, eliminates an isolation and the use of dichloromethane in the workup and improves reaction time >140-fold compared with the batch process to deliver multigram quantities of 3b in 60-65% isolated yield with >99 HPLC area % and >99% ee. While the flow process greatly increases the efficiency compared with the batch process, small-scale batch experiments were crucial in gaining reaction understanding to increase the kinetics and minimize impurity formation. The holistic process design underscores our belief that large-scale flow processes are built upon the knowledge gained through well-chosen small-scale batch experiments.

Full text of DOI:10.1021/acs.oprd.9b00478

Subscriber access provided by BIU Pharmacie | Faculté de Pharmacie, Université Paris V
Full Paper
A Flow Process Built Upon a Batch Foundation – Preparation of a Key
Amino-Alcohol Intermediate via Multi-Stage Continuous Synthesis
John Jin Lim, Kenneth Arrington, Anna L Dunn, David C. Leitch, Ian Andrews, Neil R Curtis, Mark J Hughes,
Daniel Tray, Charles E Wade, Matthew Peter Whiting, Charles Goss, Yangmu C Liu, and Brian Roesch
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00478 • Publication Date (Web): 10 Dec 2019
Downloaded from pubs.acs.org on December 10, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
A Flow Process Built Upon a Batch Foundation –
Preparation of a Key Amino-Alcohol Intermediate
via Multi-Stage Continuous Synthesis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
,†
†
†
1†
†
John Jin Lim* , Kenneth Arrington* , Anna L. Dunn , David C. Leitch , Ian Andrews ,
ǂ
ǂ
ǂ
ǂ
Neil R. Curtis , Mark J. Hughes , Daniel R. Tray , Charles E. Wade , Matthew P.
ǂ
§
§
§
Whiting , Charles Goss , Yangmu Chloe Liu , Brian M. Roesch
†Chemical Development, API Chemistry, GlaxoSmithKline, Upper Providence, PA, USA.
1Department of Chemistry, University of Victoria, Victoria, BC, V8P 5C2, Canada.
ǂChemical Development, API Chemistry, GlaxoSmithKline, Stevenage, UK.
§Chemical Development, Product and Process Engineering, GlaxoSmithKline, Upper
Providence, PA, USA.
Corresponding Authors
*
john-jin.x.lim@gsk.com, kenneth.x.arrington@gsk.com
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table of Contents artwork.
NH4OH
75 equiv
O
3.0 M NaOH
Cl
3
equiv
OH
N
O
OH
(
R)-ECH
equiv
Crystallization H N
2
OH
N
2
H N
N
2
Cl
N
2
OH
O
Stage 1A
Stage 2
90oC, 3-4' 
3a (free base)
90% assay yield
Stage 1B
40 C, 15'
79% assay yield
o

40 C, 15'
o
O
1
(1/2)
HN
9
0-95% assay yield
OH
THIQ
water soluble impurities
3b
60-65% over 2 steps
99% HPLC purity
99% e.e.
50 g synthesized
>
>
Phase separation
ACS Paragon Plus Environment
2

Page 3 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT. This paper describes recent efforts to apply flow technology in the
preparation of a key amino-alcohol intermediate (3b) so as to address manufacturability
issues present in the batch process of a PRMT5 inhibitor. The continuous process, one
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of the first reported pharmaceutical processes to use aqueous NH OH in flow, eliminates
4
an isolation, the use of DCM in the work-up, and improves reaction time >140-fold
compared to the batch process to deliver multi-gram quantities of 3b in 60-65% isolated
yield and >99% HPLC area % and >99% e.e. While the flow process greatly increases
efficiency compared to batch, small scale batch experiments were crucial in gaining
reaction understanding to increase kinetics and minimize impurity formation. The holistic
process design underscores our belief that large-scale flow processes are built upon the
knowledge gained through small scale, well-chosen batch experiments.
Keywords:
Continuous processing, batch chemistry, amino-alcohol, ammonium
hydroxide, kinetics, PRMT5.
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Introduction.
Continuous manufacturing of active pharmaceutical ingredients (APIs) using flow
chemistry is a significant and growing aspect of pharmaceutical process development.1
The benefits of continuous flow reactors as an enabling technology in pharmaceutical
research have been well established, with several excellent reviews available.2-6 It is
becoming common for industrial research groups to leverage these benefits by
investigating whether their current batch API processes can be converted to continuous
flow.7-13 However, the challenge of transforming a well understood batch process to a
continuous process can be daunting.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Through developing multiple continuous processes for clinical candidate APIs,12,13 our
group has found that successful continuous processes are most efficiently built upon
reaction understanding gained through traditional batch chemistry. Performing reaction
evaluation and optimization experiments in flow can be problematic due to a number of
logistical concerns. The more reagents tested, the more feed lines and vessels are
required. Testing new chemistry may cause unexpected solid formation, resulting in
clogging. Additionally, large amounts of material must be discarded in every experiment
ACS Paragon Plus Environment
4

Page 5 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
as flow conditions reach steady state. In contrast, traditional batch chemistry presents a
much lower barrier to implementation. Screening reactions and time course experiments
are often very simple to perform in batch, and provide valuable information regarding
reaction conditions and kinetics. The process understanding can then be used to choose
the critical process parameters for the flow process, such as residence time and reaction
temperature, and helps project teams make crucial equipment choices for scale-up.
As part of synthetic efforts toward protein arginine methyltransferase-5 (PRMT5)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
14
inhibitors, a class of promising oncology targets, we required a high-yielding, efficient
synthesis of amino alcohol 3b. The batch synthesis (Scheme 1) required large volumes
of ammonium hydroxide (NH OH) solution (~30 vol, ~90 equiv) at low temperature and
4
long reaction times, a seeded isolation of unstable and potentially hazardous intermediate
1, and used the non-preferred solvent dichloromethane (DCM) for the final work-up and
isolation of 3b.
Scheme 1. Batch synthesis of 3b. Undesirable attributes of this synthesis are labeled in
red.
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
Cl
(
R)-epichlorohydrin
1
) NH OH (30 vol, 90 equiv),
(
1.1 equiv)
4
o
o
18 h, EtOH, -5-0 C
6
0 h, -5-0 C
OH
HO
HO
O
O
2) K CO , DCM
1
:1 EtOH:water
2
3
H N
2
N
1/2
OH
3) Oxalic acid
(6 vol)
Cl
N
3b
HN
seed
Stage 2
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Heptane
Tetrahydroisoquinoline
(THIQ)
55-65% isolated yield over 2 stages
>98% e.e.
Potential handling hazard
Stage 1
o
Unstable above 20 C
Our group recognized that several challenges to the manufacturability of 3b could be
addressed by utilizing a continuous process. Telescoping the two-step synthesis of 1 to
3
b would eliminate a time consuming isolation of an unstable and possibly hazardous
15
intermediate 1, which is similar in structure to epichlorohydrin, a suspected carcinogen ,
and thus flagged as a potential handling risk. Furthermore, using a flow reactor at high
temperature with sufficient back pressure could remove the need for a high pressure
reactor for the ammonium hydroxide reaction while also decreasing reaction time.
Modifying our process to telescope two stages into one does remove an isolation step
that provides an opportunity to purge impurities; therefore, it is extremely important to
assess the impact of this process change to the purity of downstream intermediates and
ultimately the target API. Additionally, our laboratory experiments uncovered another
ACS Paragon Plus Environment
6

Page 7 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
potential risk to address: direct treatment of a solution of chlorohydrin 1 with excess
NH OH at room temperature in batch results in ~5% racemization of 3a, presumably
4
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
through the route shown in Scheme 2.
Scheme 2. Impurities (red) formed in Stages 1-2 from 1.
OH
NH4OH
H2N
N
N
N
HO
HO
6 (undesired enantiomer)
azetedinium mixture
(identified by LCMS, tentative structures)
X
water
heat
OH
OH
OH
base, heat
O
NH4OH
H
OH
H2N
N
N
N
N
Cl
N
N
Stage 2
3
a (free base)
5
2
1
(only observed in Stage 2)
THIQ
OH
N
N
4
(only observed in Stage 1)
Although the azetedinium mixture was not isolated or fully characterized, there is
abundant literature precedent for its formation from intramolecular cyclization following
the reaction of primary and secondary amines and epoxides such as epichlorohydrin
16
(ECH). Because we are using (R)-ECH, the intramolecular cyclization results in a 1:1
mixture of non-equivalent azetediniums. Thus, the ammonium hydroxide can react with
the mixture to form either 6 or 3a, resulting in loss of enantiopurity. The 3b synthesized
by the batch process was highly enantiopure; thus, we hypothesize that any undesired
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
enantiomer 6 that formed in the batch process was either purged during the isolation of
3
b, or the azetedinium mixture that eventually forms 6 was purged during the isolation of
. Further study revealed that the azetedinium mixture could not form from epoxide 2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Thus, our flow strategy (Scheme 3) was to form 1 from THIQ and (R)-ECH in Stage 1A
then quickly convert it to 2 by the addition of exogenous base in Stage 1B. Next, a liquid-
liquid extraction (LLE) is performed to remove water soluble impurities, including the
azetediniums which we found partitioned preferentially into the aqueous layer (details in
the Supporting Information). Aminolysis of the Stage 1B organic solution containing 2 with
NH OH in Stage 2 would furnish 3a. Reaction conditions had to be carefully controlled to
4
minimize the undesired bis-substitution products 4 and 5 that could form in Stage 1 and
2, respectively, since the purging limits of these impurities was unknown. Compound 4
was not observed in large quantities in the batch process due to the milder reaction
conditions and the purging of excess THIQ after the Stage 1 isolation.
Scheme 3. Proposed flow scheme Stage 1-2
ACS Paragon Plus Environment
8

Page 9 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
O
Cl
base
NH4OH
O
OH
(
R)-ECH
OH
N
2
H N
N
Cl
N
2
Stage 2
3a (free base)
Stage 1A
Stage 1B
1
HN
THIQ
water soluble impurities
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Phase separation
Stage 1A Development.
Numerous solvent/base screens were run in batch to evaluate Stage 1 conditions (details
in the Supporting Information). Initially in situ formation of the epoxide 2 was investigated
by adding a base to the reaction with THIQ and (R)-ECH. We quickly discovered that
addition of base in Stage 1 not only promotes the formation of 2, but also gives
unacceptably high levels of 4. Alkyl ether side products were observed when employing
straight chain alcoholic solvents such as MeOH and EtOH, presumably through a
Williamson-type process where the base deprotonates the alcohol to form the alkoxide,
which can then displace the chlorine in 1. Therefore, base is introduced after the formation
of 1. A base screen revealed t-AmOH as a suitable solvent, facilitating clean conversion
to 1 in the absence of base, albeit at a slower rate than MeOH or EtOH. The hindered
nature of t-AmOH also renders the solvent less likely to undergo substitution with (R)-
ECH in the presence of base. To increase the rate of reaction in t-AmOH, we investigated
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
water as a co-solvent. The effect of water on nucleophilic reactions with epoxides have
been well studied, with factors ranging from pH, hydrogen bonding interactions, and
entropy all relevant for any given reaction.17-20 We were pleased to see the addition of
water not only accelerated the rate of formation of 1, but likewise suppressed the
formation of 4 (details in the Supporting Information). The azetedinium mixture did not
increase significantly in the presence of water; we suspect this is due to the relatively mild
temperature and that 1 is the more kinetically favored product. Notably, hydrolysis of (R)-
ECH was not observed. The use of t-AmOH also facilitates an LLE after Stage 1B and
offered solvent continuity with subsequent stages of chemistry not shown in this paper.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
While a full kinetic model for Stage 1A has not yet been determined, several reaction
progress experiments were conducted in batch to evaluate reaction conditions using 4:1
t-AmOH/water as solvent (Figure 1). It was determined that manipulations of the THIQ
concentration and the equivalents of ECH do not significantly impact the early rate of
reaction but do affect the overall kinetics. Increasing the concentration from 5 to 2.5
volumes increases the overall rate and the use of excess (R)-ECH resulted in the
ACS Paragon Plus Environment
10

Page 11 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
o
complete consumption of THIQ after 20 min at 55 C, though significantly more
azetedinium and 4 were observed. By running Stage 1A more concentrated with an
excess of reagent we were able to reduce the temperature, improving the impurity profile.
A “same excess” kinetic experiment21-23 did not implicate (R)-ECH decomposition for the
slower reaction rate at lower loadings (details in the Supporting Information). While
excess (R)-ECH loading is not ideal, the reagent is inexpensive and NMR studies
indicated that the excess should be at least partially quenched in Stage 1B (details in the
Supporting Information). We also felt that any residual (R)-ECH or its by-products could
be removed by an LLE prior to Stage 2 or during the final isolation of 3b.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Stage 1A Time Course Data for Compound 1 (batch reactions) demonstrating
the effect on the rate of the reaction as a function of reaction concentration, equivalents
of (R)-ECH, and temperature.
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
Cl
OH
(
R)-ECH
Cl
N
HN
Stage 1A
1
THIQ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Further kinetic analysis indicated that while the reaction rate to 1 in Stage 1A is not very
sensitive to temperature, the rates of impurity formation are strongly temperature
dependent (details in the Supporting Information). These insights led to our optimized
Stage 1A conditions: THIQ dissolved in 2.5 vol 4:1 t-AmOH:water (referred hereafter as
Feed 1) treated with 2 equiv of (R)-ECH at 40°C and 15 minute residence time (). These
conditions routinely provide 1 in 90-95% assay yield with ~ 2.1% HPLC area unreacted
THIQ and ~2.6% HPLC area azetedinium mixture, with no 4 observed. A flow IR method
was developed as an in-line process analytical technique (PAT) to track the formation of
1 for real-time steady state determination of Stage 1A (see Supporting Information)
ACS Paragon Plus Environment
12

Page 13 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Stage 1B Development.
While direct treatment of the Stage 1A output with NH OH is desirable from an operational
4
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
standpoint, a batch experiment where Stage 1A output produced in flow was treated with
excess NH OH at room temperature resulted in ~5% loss of enantiopurity for 3a,
4
presumably through the pathway shown in Scheme 2. Since the (R)-ECH from our
supplier was >99% enantiopure, we assume the e.e. erosion was due to the azetedinium
formation. In an effort to ameliorate the e.e. erosion, we screened numerous bases in
small-scale batch experiments to identify those that favor formation of epoxide 2 over the
azetedinium mixture from chorohydrin 1 (details in the Supporting Information). Aqueous
3.0 M NaOH proved to be the most effective in terms of reaction time, side product
minimization, and rapid phase separation during the subsequent LLE. Achieving rapid
phase separation is critical, as the Stage 1B output stream is not stable over time at room
temperature, forming degradation products and 4.
The use of aqueous base may lead to kinetics being dependent on mass transfer through
liquid-liquid mixing. In order to interrogate such effects, we performed several reaction
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
progress experiments in batch. These experiments clearly indicated the reaction rate is
somewhat dependent on mass transfer (details in the Supporting Information). Thus, we
studied Stage 1B in flow reactors with varying mixing modes to more clearly elucidate the
extent of mass transfer effects (Table 1, see Supporting Information for a picture of the
reactor set-up). The Stage 1A conditions were kept constant to maintain a 15 min .
Because Stage 1A is homogeneous and presumably non-mass transfer limited, we used
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10-20 mL perfluoroalkoxy alkane (PFA) reactors (1/16” OD, 0.03” ID) without static mixing
elements. Feed 1 (THIQ in 2.5 vol 4:1 t-AmOH/water) and Feed 2 (neat (R)- ECH) were
passed through an MDAT (magnetically driven agitation in a tube)24 mixer before
o
introduction to the Stage 1A reactor set at 40 C. Stage 1A was monitored by on-line flow
IR and off-line HPLC. Assay yields for Stage 1A typically ranged from 90-95% with
unreacted THIQ and the azetedinium mixture as the mass balance. The Stage 1A output
was premixed with a stream of 3.0 M NaOH (2.5 equiv) in an MDAT prior to introduction
to the Stage 1B reactor. An 8 bar back pressure regulator (BPR) was attached at the
Stage 1B outlet. The results in Table 1 show the complex interaction of mixing, residence
time, and temperature.
ACS Paragon Plus Environment
14

Page 15 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Table 1 – Stage 1B Continuous Flow Results Using Tube-Based Reactors.
O
3.0 M NaOH
2.5 equiv
Cl
OH
O
(
R)-ECH
equiv
N
N
N
N
OH
N
2
Cl
N
HO
HO
azetedinium mixture
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
Stage 1A
Stage 1B
HN
1
THIQ
Condition
Stage
3.0 M
NaOH
flow rate
Total flow
Stage 1B
residence
time
Stage
1B
HPLC area
% of
azetedinium
mixture
1
A
Feed 1
flow rate
(mL/min)
(R)-ECH
flow rate
(mL/min)
HPLC
area %
of 1
HPLC
area % area %
of 2
HPLC
rate out of
Stage 1B
(mL/min)
reactor
volume
Notes
temp
of 4
o
(
mL/min)
(min)
( C)
(
mL)
1
2
3
4
5
6
7
8
10.0
10.0
10.0
10.0
20.0
20.0
20.0
20.0
0.449
0.449
0.449
0.449
0.898
0.898
0.898
0.898
0.218
0.218
0.218
0.218
0.436
0.436
0.436
0.436
0.983
0.983
0.983
0.983
1.97
1.65
1.65
1.65
1.65
3.30
3.30
3.30
3.30
7.87
7.87
7.87
7.87
3.94
3.94
3.94
3.94
35
45
55
65
35
45
55
65
44.075
34.079
20.587
4.344
49.824
60.102
73.777
87.655
50.039
55.026
62.731
69.763
0.263
0.258
0.288
0.296
0.217
0.279
0.341
0.414
1.325
1.484
1.872
2.223
1.087
1.283
1.288
1.741
Longer
residence
time in
Stage 1B
43.042
39.336
32.384
23.884
1.97
Better
mixing in
Stage 1B
1.97
1.97
no static
mixing
9
10.0
0.449
0.218
0.983
1.65
15.1
35
12.761
74.899
n/o*
2.541
o
Stage 1A: temp = 40 C,  = 15’, conversion = 90-95% for all conditions.
*n/o = not
observed
Conditions 1-8 used a custom built 13 mL Vapourtec PFA reactor (3/16” OD, 0.13” ID)
containing static mixing elements for Stage 1B (details in the Supporting Information).
Condition 9 was performed in a 25 mL PFA reactor (1/16” OD, 0.03” ID) without static
mixing. For each condition, ~ 5 mL of the Stage 1B output was collected, allowed to phase
separate, and the organic layer was sampled for HPLC. Nearly equal conversions were
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
observed in Condition 5 and Condition 1 despite 50% reduced residence time,
demonstrating the effects of better mixing at higher flow rates. Given sufficient residence
time, high conversion can be achieved even without active mixing (Condition 9); however,
more azetedinium is observed in the organic layer. At shorter residence times (Conditions
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5-8), less azetedinium is observed in the organic layer most likely due the increased
efficiency of the partitioning at higher flow rates (i.e., better mixing). In effect, the LLE is
occurring at the same time as the hydrolysis, so the better mixed conditions show more
favorable impurity partitioning. High conversions can be also be achieved by increasing
the temperature, though this leads to increased impurity formation (Conditions 4 and 8).
Very little of 2 partitioned into the aqueous layer after phase separation, so no back-
extraction was necessary.
We next examined the mixing effects on Stage 1B with a Corning Low Flow (LF) reactor
under several sets of conditions (Table 2). Each plate is comprised of heart-shaped
channels which provide intensive mixing even at relatively low flow rates (details in the
Supporting Information). Not only does the Corning reactor provide effective mixing, but
ACS Paragon Plus Environment
16

Page 17 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
lab-scale reactor characterization can be used to predict the behavior of the reaction in
25
plant-scale equipment. The lab-scale reactor is comprised of a number of glass plates,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
each with an internal volume of 0.40-0.45 mL. At the 2-3 g/min flow rates recommended
26
by the manufacturer, even biphasic reaction mixtures are effectively emulsified. Using
these reactors, we collected data on the impact of mixing on the reaction rate and impurity
formation by adjusting the flow rates for the telescoped Stage 1A-1B reaction (Diagram
of the reactor set-up in the Supporting Information).
The residence time in Stage 1B was controlled by adjusting the flow rates of the three
feeds to the process. Due to the number of glass plates available in our labs, the Stage
1A reactors could not be configured to maintain a constant 15 min . This constraint on
our reactor set-up did not affect the conversion of starting material in Stage 1A (Table 2),
but may lead to higher impurity levels than those observed with the optimal 15 min ;
therefore, these studies were used to only investigate conversion to 2, and not impurity
partitioning. We thus quenched the Stage 1B output stream directly into HPLC diluent
rather than allowing the Stage 1B output to phase separate and sampling each phase
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
separately as we had done for the experiments in Table 1. Due to the biphasic nature of
the samples, 2-3 samples were taken for each condition. The results are summarized in
Table 2 and Figure 2 (a more detailed table showing all flow rates is available in the
Supporting Information).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2 – Stage 1B Continuous Flow Results Using a Plate-Based Reactor (Corning LF)
O
3.0 M NaOH
equiv
Cl
3
(
2
R)-ECH
equiv
O
OH
N
Cl
N
Stage 1A
Stage 1B
2
1
HN
(THIQ)
Total
Stage
1B
flow
rate
Stage
3M
NaOH
flow
rate
(g/min)
1
A
Feed 1
flow rate
(g/min)
(R)-ECH
rate
(g/min)
Stage 1A
conversion
(%)
Corning
volume
(mL)
HPLC
area %
of 1
HPLC
area %
of 2
Stage 1A
conditions
Stage 1B
conditions
Condition reactor
vol
(mL)
(
g/min)
o
o
79.221a
9.719a
1
2
3
4
5
6
7
8
16
16
16
20
16
16
16
16
21’, 40 C
0.519
0.778
0.778
1.038
0.519
0.260
0.260
0.390
0.231
0.347
0.347
0.463
0.231
0.116
0.116
0.173
n/a
94
95
93
97
97
96
98
0.4’, 40 C
1.400
2.101
2.101
2.801
1.400
0.700
0.700
1.050
2.150
3.226
3.226
4.302
2.150
1.075
1.075
1.613
0.85
0.85
3.90
3.90
3.90
3.90
3.90
3.90
o
o
81.402a
6.374a
15’, 40 C
0.3’, 40 C
o
o
27.212a
55.084a
15’, 40 C
1.3’, 60 C
o
o
43.158a
45.291a
14’, 40 C
1.0’, 60 C
o
o
9.040a
73.023a
21’, 40 C
1.9’, 60 C
o
o
1.481b
74.064b
42’, 40 C
3.9’, 60 C
o
o
7.661b
77.982b
42’, 40 C
3.9’, 40 C
o
o
1.789b
83.825b
28’, 40 C
2.6’, 60 C
a
b
Average of 2 HPLC injections. Average of 3 HPLC injections
Density Feed 1 = 0.92 g/mL, Density (R)-ECH = 1.18 g/mL, Density 3.0M NaOH = 1.12 g/mL
ACS Paragon Plus Environment
18

Page 19 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Figure 2 – Stage 1B Conversion vs Residence Time in Corning LF reactor at 40 and
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
0oC
3.0 M NaOH
3
equiv
O
OH
N
Cl
N
Stage 1B
2
1
As shown in Table 2, poor conversions are observed with short residence times, even
with very good mixing (Conditions 1 and 2). As residence time is increased, conversion
improves. Even for conditions that deliver sub-optimal mixing (Conditions 6-8, total flow
rate < 2 g/min), high conversions can be achieved through a combination of increased
temperature and residence time. Encouragingly, even at the higher temperatures, HPLC
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
quantification of impurity 4 did not exceed 1.1 area %. Due to the material incompatibility
o
of the caustic Stage 1B output with the glass reactor plates, temperatures above 60 C
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
were not examined.
The results from Tables 1 and 2 illustrate that high conversions in Stage 1B are achieved
by balancing residence time, mixing, and temperature. It is concluded that the Stage 1B
reactions were not completely mass transfer limited because there is a large impact of
residence time on conversion, indicating at least a partially kinetically controlled system.
Furthermore, reactions run at lower flow rates (and therefore, less mixing) did not result
in low conversion which would be the case if the system was strictly mass transfer limited.
However, there appears to be a relatively small effect of temperature on conversion
o
between 40 and 60 C (Figure 2). If the system is kinetically limited, one would expect a
larger temperature effect; however, we note that this is a relatively small sample size. In
addition, our crystallization studies (details in the Supporting Information) do indicate that
leftover (R)-ECH is present throughout the stages, and its effect on the reaction
mechanism for Stage 1B is not currently known. While the exact reaction mechanism for
ACS Paragon Plus Environment
20

Page 21 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Stage 1B has yet to be determined, these results demonstrate that the reaction can
proceed in high conversions largely independent of reactor type, allowing greater
flexibility in reactor choice for Stage 1B. These simple flow experiments revealed an
important fact about Stage 1B - given sufficient residence time, good conversions to 2
could be achieved even with sub-optimal mixing. However, more effective partitioning of
impurities into the aqueous phase likely occurs when the reaction is better mixed.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Following Stage 1B, the output is collected in an intermediate storage vessel in an ice-
water bath. Epoxide 2 partitions exclusively into the organic layer, while the majority of
the azetediniums partition into the aqueous (details in the Supporting Information). After
phase separation, the organic layer is transferred to a feed vessel, stored cold, and then
used as-is for Stage 2.
Stage 2 Development.
The Stage 2 batch process required large excesses (90 equiv) of NH OH at extended
4
o
reaction times and low temperature (0-5 C, 18 h) to minimize formation of 5. As NH OH
4
is a poor nucleophile compared to the Stage 2 product 3a, large excesses are required
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to avoid formation of impurity 5. Increasing the temperature of the batch reaction would
require the use of specialized pressure reactors due to the amount of NH OH present,
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
further complicating the process. Thus, we focused our efforts to adapt the batch process
to flow utilizing back pressure regulators in standard flow reactors at elevated
temperatures. Surprisingly, the use of aqueous NH OH in flow organic synthesis has not
4
27
been frequently reported; we identified only one prior example in the literature. The use
of ammonia surrogates was also investigated, but we found these to be unsuitable for
flow operation.28 The use of gaseous ammonia NH in flow was considered but not
3
explored due to development timelines.
Gathering kinetic data for Stage 2 in batch posed a challenge due to the high pressures
resulting from using NH OH at elevated temperatures. We decided to explore time points
4
o
o
o
at three different temperatures (70 C, 90 C, and 100 C) by dispensing aliquots of Stage
1B organic flow output containing 2 in solution to glass vials equipped for magnetic stirring
(
Figure 3). Aqueous NH OH (24-26% w/w) was added such that very little headspace
4
remained in the vial in an attempt to mimic conditions in a flow reactor (~ 73-75
ACS Paragon Plus Environment
22

Page 23 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
equivalents based on HPLC assay of 2). The vials were then placed on a stirring block
set at the desired temperature and maximum agitation rate. Vials were removed at
designated time points, allowed to cool, and then sampled for HPLC.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3 – Stage 2 Time Course Data (Batch)
NH4OH
(~75 equiv)
OH
OH
OH
H
O
H2N
N
N
N
N
N
Stage 2
3a (free base)
5
2
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We were pleased that our initial time course data showed Stage 2 was complete in less
than 10 min at elevated temperature. Extended reactions times resulted in increased 5,
o
and temperatures >70 C were required for complete consumption of 2. It was also noted
that the solutions were homogeneous at high temperature, indicating that the kinetics
should not be mass transfer limited. Thus, intensive static mixing should not be required
for Stage 2.
Following our time course studies in batch, Stage 2 scoping studies were performed in
flow (Table 3, see Supporting Information for a picture of the reactor set-up). Stage 2 was
performed in a 10 mL PFA reactor (1/16” OD, 0.03” ID). Feed 3 (Stage 1B organic output
following phase separation) and Feed 4 (aqueous NH OH) were mixed in an MDAT
4
ACS Paragon Plus Environment
24

Page 25 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
before introduction to the Stage 2 reactor set at the desired temperature. Feed 3 was kept
o
at ~ 0-5 C in an ice water bath during the reaction. An 8 bar BPR was attached at the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Stage 2 outlet. The system pressure stabilized to 6-7 bar at steady state.
Table 3 - Stage 2 scoping in flow as monitored by offline HPLC.
NH4OH
OH
OH
OH
H
O
2
H N
N
N
N
N
N
Stage 2
3a (free base)
5
2
Condition NH
4
OH
Stage 2
residence
time (min)
Stage 2
HPLC
area %
of 2
HPLC
area %
of 3a
HPLC
area %
of 5
HPLC area
% of
unknown
impurity @
(equiv)
temperature
o
( C)
2
1 min
1
2
3
4
5
6
7
8
9
75
75
50
75
95
75
75
75
75
75
75
5
5
5
5
5
5
5
3
4
5
6
75
80
85
85
85
90
95
90
90
90
90
6.381
4.410
4.410
2.490
1.153
1.271
0.630
3.003
1.417
1.466
1.606
54.706
58.101
46.843
55.770
58.009
58.747
48.714
55.797
56.974
56.682
55.917
3.010
6.155
14.091
11.279
10.335
6.850
14.412
4.534
6.617
7.652
7.080
0.267
0.251
4.543
2.008
1.393
1.455
5.437
n/o*
0.469
0.843
1.196
1
0
1
1
*n/o = not observed
This data gave several important insights into this reaction. Using 50 equiv of NH OH is
4
not sufficient for complete conversion for the tested residence times and allows for greater
formation of 5 and an unknown non-polar impurity at a 21 min retention time on our HPLC
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
method, possibly a trimeric side product (Condition 3). However, the addition of 95 equiv
of NH OH does not appear to offer much advantage over 75 equiv (Condition 5).
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
Temperature and residence time also play very important roles, as temperatures < 75 C
lead to poor conversion while residence times > 3-4 min lead to greater impurity formation
(Conditions 1-7, 10 and 11). The optimal zone appears to be around 90°C and 3-4 minute
residence time with 75 equiv of NH OH (Conditions 8 and 9). The Stage 2 reaction in flow
4
seems slightly faster than in batch at the same temperature, most likely due to the better
heat transfer in the flow reactor.
Stage 1-2 Extended Lab Runs.
Following optimization of Stages 1-2, several extended lab runs were performed (Tables
4
& 5). Although it would be desirable use the Corning reactor for Stage 1B in these
extended runs as it is most likely the type of reactor that will be used at pilot and
manufacturing scale, we were limited with the number of Corning reactors available.
Thus, these extended runs were performed in PFA reactors. The Stage 1A conditions
o
were fixed at 40 C, 2 equiv (R)-ECH, 15’  in a 10 mL PFA reactor. After Stage 1A had
ACS Paragon Plus Environment
26

Page 27 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
reached steady state by flow IR, Stage 1B conditions were started by commencing flow
of 3.0 M NaOH (3 equiv). Stage 1B was performed in either a 25 mL PFA reactor with no
static mixing (Runs 1 & 2, Table 4) or a 13 mL PFA reactor with static mixing elements
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Run 3, Table 4). After waiting ~ 2 residence times to achieve steady state for Stage 1B,
the Stage 1B output was collected into an intermediate storage vessel. After collection,
the Stage 1B output was phase separated and the organic layer (~31% of the biphasic
output by volume) was assayed by HPLC and then stored cold. The lab equipment was
then reconfigured to perform Stage 2 using the same 10 mL PFA reactor in Stage 1A and
the Stage 1B organic output (Feed 3) and NH OH (Feed 4). Stage 2 conditions were fixed
4
o
using 75 equiv of NH OH, 3’, and 90 C. Collection of the Stage 2 output was begun
4
after ~ 2 residence times elapsed.
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 4 – Stage 1 Extended Lab Runs
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
3.0 M NaOH
3 equiv
Cl
OH
O
(
R)-ECH
equiv
N
N
N
N
OH
N
2
Cl
N
HO
HO
2
4
Stage 1A
Stage 1B
azetedinium mixture
HN
1
THIQ
Stage 1A-1B
total flow
rate
Stage 1A-1B
HPLC area %
of azetediniums
in the organic
layer
Time of
Stage 1B
collection
(h)
HPLC area %
of 4 in
organic layer
Stage 1A
conditions
Stage 1B
conditions
assay yield of 2
in organic layer
(%)
Run
1
(mL/min)
o
o
15’, 40 C, 10
mL reactor
15’, 40 C, 10
mL reactor
14.5’, 40 C,
1.72
78
79
0.88
0.78
1.36
1.41
2.5
2.0
25 mL reactor
o
o
2 (repeat
of 1)
14.5’, 40 C,
1
.72
25 mL reactor
o
7.6’, 60 C, 13
mL reactor
o
1
5’, 40 C, 10
mL reactor
3
1.72
79
2.32
1.62
1.8
(static mixing)
Table 5 – Stage 2 Extended Lab Runs
NH OH
Feed 4)
5 equiv
4
(
7
OH
OH
OH
H
O
H N
2
N
N
N
N
N
Stage 2
3a (free base)
5
2
Feed 3
Assay
Stage 2
total flow concentration
rate
mL/min)
Assay
Stage 2 concentration
Time of
Stage 2
collection
(h)
Feed 3
flow rate flow rate
mL/min) (mL/min)
Feed 4
HPLC
area %
of 5
Stage 2
conditions
assay
of 2 in Feed yield of
3 (mg/g) 3a (%)
of 3a in
Stage 2
output
Run
(
(
(
mg/g)
ACS Paragon Plus Environment
28

Page 29 of 44
Organic Process Research & Development
1
2
3
4
o
3
’, 90 C,
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
10 mL
reactor
0.414
0.414
2.71
2.71
3.10
3.10
362.5
371.9
90
58.3
57.1
5.2
7.3
2.0
2.0
o
3’, 90 C,
10 mL
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
n/d
(
repeat)
reactor
The extended lab runs showed fairly good reproducibility in terms of output quality. For
both Stage 1 and 2, no fouling or overpressurizations were observed with these extended
run times. The amount of 4 that is observed in the final Stage 1B organic layer (Runs 1-
3
, Table 4) is slightly higher than the analagous conditions observed in our Stage 1B
scoping studies (Table 1), likely due to the reaction of epoxide 2 with unreacted THIQ
remaining from Stage 1A during the longer collection times. The increased amount of 4
in Run 3 (Table 4) is probably because the receiving flask was not totally submerged in
the ice-water bath and thus the Stage 1B output had an opportunity to warm to ambient
temperature. However, 4 can be purged in the final crystallization. Somewhat surprisingly,
the amount of azetedinium in the organic layer after Stage 1B was slightly higher for Run
3 (static mixing) than the conditions without static mixing (Runs 1-2, Table 4); however,
we attribute this to the fact that the total flow rate for Stage 1A-1B was kept constant for
ACS Paragon Plus Environment
29