Development of a Scalable Synthesis of trans-4-Fluorocyclohexylamine via Directed Hydrogenation

Article abstract of DOI:10.1021/acs.oprd.0c00444

Herein, a scalable and practical process to prepare trans-4-fluorocyclohexylamine hydrochloride (1a) is described. By exploitation of the embedded gem-difluoride motif in the commercially available 4,4-difluorocyclohexanecarboxylic acid, a derived orthoester-masked acid underwent dehydrofluorination to provide the requisite vinyl fluoride for a directed hydrogenation event, enabling selective access to the trans-configuration of 1a.

Full text of DOI:10.1021/acs.oprd.0c00444

pubs.acs.org/OPRD
Communication
Development of a Scalable Synthesis of trans-4-
Fluorocyclohexylamine via Directed Hydrogenation
Joyce C. Leung,*,# Thach T. Nguyen,# Mariusz Krawiec, Donghong A. Gao, and Jonathan T. Reeves#
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00444
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Herein, a scalable and practical process to prepare trans-4-fluorocyclohexylamine hydrochloride (1a) is described. By
exploitation of the embedded gem-difluoride motif in the commercially available 4,4-difluorocyclohexanecarboxylic acid, a derived
orthoester-masked acid underwent dehydrofluorination to provide the requisite vinyl fluoride for a directed hydrogenation event,
enabling selective access to the trans-configuration of 1a.
KEYWORDS: vinyl fluoride, directed hydrogenation, dehydrofluorination, trans-4-fluorocyclohexylamine
INTRODUCTION
■
Classical protocols to access fluorinated compounds typically
involve the use of costly fluorinating reagents and building
blocks accompanied by stoichiometric byproducts, rendering
these methods difficult for implementation in large-scale
production.1 Efforts have been devoted to simplify and
Figure 1. trans-4-fluorocyclohexylamine (1).
improve the quality of these C−F bond-forming trans-
formations to accommodate the increasing demand for
fluorinated compounds in materials science,2 agricultural,3
Following known literature procedures,6 most method-
ologies involve inverting the stereogenic center via deoxy-
fluorination of the corresponding N-protected cis-alcohol.
and pharmaceutical industries.1,4 In the past decade, both
academic and industrial researchers sought to develop more
user- and environmental-friendly methods to expand the
However, these processes are typically low yielding, require
current toolbox for incorporating fluorine into architecturally
the use of fluorinating reagents that are undesirable for scale-
complex molecules. Due to the unique chemical properties of
fluorinated medicinally relevant compounds in the pharma-
ceutical industry, synthetic strategies to access these molecules
remains an active area of investigation. Stereoselective
synthesis of these fluorinated compounds is of particular
interest owing to the limited methodologies.5 From the
industrial perspective, rapid access to these chiral fluorinated
building blocks selectively is vital in accelerating drug discovery
programs to expedite the development of a potential drug
candidate. Additionally, reagents involved in the synthesis pose
direct impact on process safety and production cost. Recently,
a comprehensive review by Caron highlighted the challenges in
accessing fluorinated molecules in the pharmaceutical industry,
providing insights into the unmet synthetic challenges of
incorporating fluorine into active pharmaceutical ingredients
(APIs).1
As part of a recent drug discovery program, trans-4-fluoro-
cyclohexylamine (1), featuring a trans-configuration between
the fluorine and the amino group, was a building block of
interest (Figure 1). Although this compound is commercially
available, it is extremely costly and materials provided by
external vendors are predominantly cis/trans-isomeric mix-
tures, which required additional chromatographic purification
to remove the major, undesired cis-isomer.
up, and occur with significant amounts of elimination
byproduct which can be difficult to separate from the desired
product (Scheme 1). Therefore, a practical and scalable
synthesis to supply 1 with a high level of trans-selectivity
needed to be developed.
RESULTS AND DISCUSSION
■
Our initial investigations to develop a practical and scalable
route to 1 focused on establishing the trans-configuration by
selective reduction of an N-sulfinylimine, which could be
prepared by condensation with 4-fluorocyclohexanone 7.
Unfortunately, all attempts to access 7 from the corresponding
alcohol 8 or from ketal protected derivatives of 8 by SN2
displacement of derived tosylates or other leaving groups with
fluoride or by deoxyfluorination of the alcohol were not
successful and resulted almost exclusively in products of
elimination (Scheme 2).
Special Issue: Celebrating Women in Process Chem-
istry
Received: September 30, 2020
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Scheme 1. Literature Reports of Deoxyfluorination6
various cyclohexanones under mild conditions to the
corresponding vinyl fluorides (Scheme 5).10 Our initial
attempts focused on exploring amino derivatives as potential
directing groups to promote the desired selectivity. With the
vinyl fluorides in hand, a variety of amino derivatives were
evaluated for their respective directing effects under the
reported hydrogenation conditions. Unfortunately, all deriva-
tives offered either no reactivity (R = NHBoc, NHAc) or
furnished only the undesired des-fluoro product (R = NH2),
suggesting that a carboxylic acid derivative might be required
at the 4-position of the cyclohexene.
The observed lack of reactivity prompted the investigation of
substrates containing carboxylic acid directing groups. Vinyl
fluoride 14 was prepared by the aforementioned deoxyfluori-
nation of the commercially available cyclohexanone 13. One
advantage of having the ester handle in 14 was that simple
elaborations gave access to various derivatives, i.e., carboxylic
acid 15 and carboxamide 16, to enable a quick screen of their
directing abilities in the hydrogenation event (Scheme 6).
Additionally, these functional handles could be readily
converted to the required amine via known rearrangement
methods.
With the available substrates for our study, we set out to
examine the selectivity of the hydrogenation employing the
literature conditions. Interestingly, moderate reactivity and
stereoselectivity was observed in the case of ester 14, while
acid 15 furnished <5% product (Table 1). In contrast, primary
amide 16 provided excellent trans/cis-selectivity, resulting in
predominantly the desired trans-isomer (≥95:5 d.r.). Notably,
during the hydrogenation screening, olefin migration was not
detected by 1H NMR. Guided by this encouraging result, more
in-depth process optimization was initiated to supply high-
quality 1 for our discovery program.
Development of Practical Vinyl Fluoride Synthesis.
Having established the directed hydrogenation of the
carboxamide 16, production of vinyl fluoride 14 was scaled
up to over the 100 g scale. Unfortunately, the deoxyfluorina-
tion method used to access the vinyl fluoride moiety was
accompanied by formation of a gem-difluoride byproduct.
Formation of this impurity significantly increased with scale up
to 30%, and it was not possible to remove the impurity by
chromatographic purification of the ester 14, which was an oil.
The acid 15 and the amide 16 were both crystalline solids, but
recrystallization of these intermediates also failed to purge the
corresponding gem-difluoride impurities. In addition, the
fluorination reaction itself posed operational and safety
challenges due to the highly corrosive nature of the reaction
mixture which contained 3 equiv of XtalFluor-E and 2 equiv of
Et3N·3HF. Any glass-on-glass contacts in the reaction setup
would become fused during the course of the reaction. On
larger scale in the lab, quenching and workup of the reaction
mixture proved to be laborious and volume intense, and the
The observed dominant elimination pathway in all attempts
to prepare 7 prompted an alternative strategy to exploit readily
available fluorinated building blocks. Recently, Glorius and co-
workers reported a direct access to all cis-(multi)fluorinated
cycloalkanes via rhodium−carbene catalyzed hydrogenation of
fluoroarenes.7 As described in the report, TBS-protected
phenol 9 could readily undergo hydrogenation to the
corresponding cis-isomer 10 in >10:1 cis/trans-selectivity and
near quantitative yield on the gram scale (Scheme 3).
Theoretically, the corresponding cyclohexanol, upon TBS
deprotection, could be displaced with a masked amine
equivalent to access the desired trans-isomer of 1. Unfortu-
nately, further evaluation led to the finding that neither the
rhodium complex nor the carbene ligand employed in this
transformation was commercially available, rendering the
strategy impractical for our purposes.
While the fluorophenol hydrogenation approach was not
feasible due to the nonavailability of the catalyst and ligand, a
new strategy utilizing hydrogenation of a vinyl fluoride
emerged. It is known that carboxamide and carboxylate
substituents exhibited directing effects when Crabtree’s catalyst
was employed in the hydrogenation of cyclohexenes.8 As
hypothesized in the report, the amide or ester carbonyl is
capable of coordinating to the cationic iridium center, which
allows for hydrogen delivery from the same face with respect to
the carboxyl moiety, enabling direct access to cyclohexane
derivatives with high trans-selectivity (Scheme 4).
Taking advantage of this concept, we postulated that
directed hydrogenation of a vinyl fluoride could allow for a
highly diastereoselective synthesis of 1. While recent advances
in methodologies for vinyl fluoride synthesis offer creative
approaches to access this moiety,9 many of these operations
are substrate specific and require cryogenic conditions or a
transition metal catalyst. To enable quick access to the
requisite vinyl fluoride motif to test our hypothesis, we were
attracted to a report on the use of XtalFluor-E to convert
Scheme 2. Proposed Retrosynthetic Analysis of 1 via Selective Reduction
B
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Scheme 3. Proposed Synthetic Approach to 1 Using Rh-Catalyzed Hydrogenation Reported by Glorius7
gram scale, the reaction resulted in a mixture of 22 and a major
side product, later confirmed as the t-amyl vinyl ether 23.
Presumably, an excess amount of base promoted displacement
of the initially formed vinyl fluoride that led to formation of
vinyl ether 23.
Scheme 4. Iridium-Catalyzed Stereoselective
Hydrogenation8
Although vinyl ether byproduct 23 could be purged in the
subsequent orthoester deprotection step, further evaluation of
the reaction conditions was desired to avoid or minimize
formation of this byproduct. Reaction parameters were
extensively studied. Overall, potassium alkoxide bases were
proven crucial for promoting dehydrofluorination. More
importantly, reaction concentration and temperature were
critical in avoiding vinyl ether formation. Although near
complete consumption of gem-difluoride 21 could be achieved
with excess base, i.e., 3 equiv, a 3:1 mixture of vinyl fluoride/
vinyl ether was obtained (Table 3, entry 2). Extended reaction
time was required for lower base loading as the reaction
appeared to be more sluggish (entry 3).
Fine-tuning of base stoichiometry and reaction time led to
the optimal conditions of 2 equiv of KOt-Amyl for 96−120 h.
It is also worth noting that commercial KOt-Amyl is available
as a 25% solution in toluene and can be used directly without
further dilution with additional solvent. Not only was this
process scalable and easy to operate, it provided a safer and
more practical alternative to synthesize the vinyl fluoride motif
required for the subsequent chemistry. With the dehydro-
fluorination method in place, vinyl fluoride 22 was readily
prepared on an over 100 g scale in excellent yields (Scheme 9).
Deprotection of the OBO-ester under mild conditions revealed
the carboxylic acid, which was further converted to
carboxamide 16 for the directed hydrogenation. Importantly,
residual imidazole (as low as 0.07 wt %) from the amidation
mediated by CDI could affect the diastereoselectivity of the
subsequent hydrogenation step, along with formation of a des-
fluoro impurity. To our delight, isolated amide could be
slurried in a minimal amount (1−1.5 vol) of cold water to
ensure complete removal of residual imidazole while
minimizing loss of water-soluble 16.
waste generated from this process required special segregation
due to the high fluoride content. Further scale-up beyond the
lab was not pursued due to these issues.
The challenges in removing the major gem-difluoride
impurity along with the cost and operational difficulties
associated with use of XtalFluor-E necessitated the develop-
ment of a practical and scalable synthesis of vinyl fluoride 14.
During our investigation on methods to recycle the gem-
difluoride byproducts from deoxyfluorination, a strategy
exploiting the gem-difluoride emerged. We proposed that an
appropriate base could potentially promote an elimination
pathway to access the requisite vinyl fluoride moiety. To test
this, elimination of commercially available 4,4-difluorocyclo-
hexanecarboxylic acid 19 was attempted and was unsuccessful.
To ensure that the acid moiety was not interfering with the
elimination, acid 19 was subsequently masked as Corey’s
OBO-ester through a rearrangement of ester 20 mediated by
BF3·OEt2, setting the stage for investigating the feasibility of
the subsequent elimination (Scheme 7).
While preparation of the vinyl fluoride motif by way of
dehydrofluorination of gem-difluorides is known, conditions
are typically harsh or are limited to activated substrates.11 The
initial screen using metal bis(trimethylsilyl)amide bases did not
provide any desired product. Of the bases screened for the
elimination reaction, alkoxides at elevated temperature
afforded the desired vinyl fluoride 22, albeit in low conversions
(Table 2, entries 7−9). Switching to potassium t-amylate
(KOt-Amyl) led to a remarkable increase in conversion to
desired product (entries 10 and 11).
Directed Hydrogenation of Vinyl Fluoride 16. As
previously mentioned, the directed hydrogenation employing
Crabtree’s catalyst revealed that a primary amide at the 4-
position of the cyclohexene was critical in achieving high trans-
diastereoselectivity. Unfortunately, defluorination was ob-
served on scale, suggesting that a different transition metal
With this encouraging result, further optimization led to
complete consumption of the gem-difluoride 21 in the presence
of 4 equiv of KOt-Amyl in xylene at reflux for 24 h (Scheme 8).
Surprisingly, when dehydrofluorination was performed on the
Scheme 5. Ir-Catalyzed Hydrogenation of Vinyl Fluorides Bearing Amino Groups
C
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Scheme 6. Synthesis of Vinyl Fluorides with Different Carboxylic Acid Derivatives
a
Table 1. Ir-Catalyzed Hydrogenation of Vinyl Fluorides
Containing Carboxylic Acid Derivatives
Table 2. Screening of Elimination of gem-Difluoride 21
b
entry
base
LiHMDS
NaHMDS
KHMDS
KHMDS
LDA
n-BuLi
NaOt-Amyl
KOt-Bu
KOt-Bu/t-BuOH
KOt-Amyl
KOt-Amyl
DBU
solvent
temp (°C)
% conversion
1
2
3
4
5
6
7
8
THF
THF
THF
PhMe
THF
THF
PhMe
THF
PhMe
PhMe
PhMe
NMP
−20 to rt
−20 to rt
−20 to rt
110
−78 to rt
−78 to 0
110
65
110
110
120
NR
NR
NR
NR
NR
5
a
a
a
entry
R
% conversion
% yield
17:18
1
2
3
OEt (14)
OH (15)
NH2 (16)
77
22
95
18
<5
69
72:28
N/D
>95:5
b
a
b
1
Conversions and yields were determined by H NMR. 2 mol %
c
[Ir(cod)py(PCy3)]PF6 was used.
3
5
10
c
Scheme 7. Preparation of OBO-Ester 21
9
c
10
11
12
97.5
99.5
5
120
a
b
Reactions were conducted on 100 mg scale. Conversions were
c
obtained by GCMS and were confirmed by 19F-NMR. 40 h.
Scheme 8. Byproduct Vinyl Ether Formation
might be required to circumvent this problem. This
observation was in agreement with the report by Krska et
al.12 Thus, we turned our attention to cationic rhodium
catalysts to minimize the undesired defluorination pathway.
Both chiral and achiral bidentate phosphine ligands
delivered moderate to good reactivity and excellent trans-
selectivity in all cases except for DPPB. Unlike previously
reported conditions by Krska et al., use of methanol as solvent
led to a significant amount of des-F-17c (Table 4, entry 6),
while <1% des-F-17c was observed with dichloromethane
(entry 3).12 Due to the urgent request for >100 g of 1 in a
short time, the nonproprietary DPDBF ligand was selected for
further optimization on the basis of its clean reaction profile
and lower cost compared to SL-J002-1 without further
screening.
At the 40 g scale, it was found that 1.0 mol % Rh was
sufficient to achieve complete conversion of vinyl fluoride to
obtain the isolated product in 82−86% yield. However, the
process suffered from several drawbacks such as the use of
toxic dichloromethane, prolonged reaction time (12−24 h),
and the need for a solvent switch from dichloromethane to
hexanes to isolate the product (Table 5, entry 1 and 2).
Further process optimization studies revealed that the same
reaction efficiency could be achieved with lower rhodium
loading (0.5%) and in a much shorter time (4 h) when THF
was used as solvent. In addition, the product can be
precipitated in >99 wt % purity simply by adding heptane to
the crude product mixture. Attempts to reduce the catalyst
loading to 0.2% led to incomplete conversion.
Although a reliable and reproducible directed hydrogenation
process was demonstrated, the intermittent commercial
availability of the DPDBF ligand might pose potential
problems if 1 was needed in kilogram quantities. Continued
screening identified the more accessible DTBPF as an equally
effective replacement for DPDBF. Interestingly, DTBPF
delivered complete conversion and consistently >90% yield
even at 0.1% rhodium. It is worth mentioning that imidazole
1
present in the vinyl fluoride in trace amount (<0.1% by H
NMR) diminished reaction efficiency and lowered trans-
selectivity (Table 6, entry 5), which was most likely a result
D
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
a
access to the vinyl fluoride was made feasible without the use
of fluorinating reagents on scale. This route eliminated the
concerns for associated gem-difluoride byproduct, rendering
the approach greener and more efficient. Subsequent
implementation of a highly diastereoselective directed hydro-
genation allowed for clean delivery of carboxamide 17c as a
single trans-isomer. Finally, a mild Hofmann rearrangement
converted the amide group in 17c to the required amino group
in 1. Overall, scalability and practicality of this synthesis was
demonstrated by its ease of product isolation and purification
by way of recrystallization, enabling production of 1a of high
quality and purity.
Table 3. Optimization of Base Stoichiometry
b
entry
base
equiv
% conversion
18
97.2, 3:1 21:22
1
2
3
KOt-Bu (solid)
KOt-Amyl
KOt-Amyl
1.5
3.0
1.5
EXPERIMENTAL SECTION
c
■
92.9
a
b
General Information. Commercially available reagents
and solvents were used without further purification unless
stated otherwise. NMR spectra (1H, 13C, 19F) were recorded
on a Bruker DRX-500 using CDCl3 or DMSO-d6 as solvent
unless stated otherwise. Chemical shifts (δ) are reported in
parts per million (ppm) and coupling constants (J) are given in
Hertz (Hz) and referenced relative to TMS (0 ppm). The
following abbreviations are used to indicate signal multiplicity:
s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, dd =
doublet of doublets, dq = doublet of quartets, td = triplet of
doublets, qd = quartet of doublets, ddt = doublet of doublet of
triplets, m = multiplet, app = apparent, and br = broad
resonance. Analysis by gas chromatography−mass spectrome-
try (GC−MS) was performed on an Agilent 6890N GC system
equipped with MS (70 eV electron Impact (EI, positive)) with
the following conditions: Agilent HP-5MS column (ID 0.25
mm, length 30 m, film thickness 0.25 μm); run time 12 min,
temperature hold at 100 °C for 2 min, ramp 25 °C/min to 300
°C, hold at 300 °C for 2 min; inlet split 30:1, inlet temperature
280 °C; carrier gas helium at a constant flow mode at a flow
rate 1.0 mL/min. Differential scanning calorimetry was
performed on a Mettler Toledo DSC 3+ featuring sensor
FRS 6+. High-resolution mass spectrometry (HRMS) data
were obtained on Thermo LTQ FT Ultra mass spectrometer at
100 000 resolving power using direct analysis in real time
(DART) source ionization in the positive ion mode. Analysis
by ultraperformance liquid chromatography−mass spectrome-
try (UPLC−MS) was performed on a Waters UPLC-MS
(TQD) equipped with MS (electrospray positive/negative
ionization) with the following conditions: Acquity UPLC BEH
C18 column (ID 2.1 mm, length 5 cm, particle diameter 1.7
μm); run time 3 min, temperature 30 °C, injection volume 0.5
μL, flow rate 0.6 mL/min and post run 0.5 min; mobile phase
A, water with 0.1% formic acid (HPLC grade); mobile phase
B, acetonitrile (HPLC grade); gradient starts at 15% B at 0
min, ramps to 95% B over 2.1 min, is held at 95% B from 2.1 to
2.7 min, and ends at 15% B at 3.0 min.
Reactions were conducted on 1 g scale. Conversions were
c
determined by GCMS. 72 h.
of competitive binding of imidazole to the cationic rhodium
center.
End Game to 1a: Hofmann Rearrangement. With
carboxamide 17c in hand, the conversion to the amine by way
of a Hofmann rearrangement was explored. Although many
reagents may be employed to effect the Hofmann rearrange-
ment, we were particularly attracted to a report using
PhI(OAc)2 due to the simplicity and mild reaction
conditions.13a Notably, use of PhI(OAc)2 in Hofmann
rearrangement has been demonstrated on the kilogram scale
production,13 which further demonstrates the safety and
scalability of the reaction. Simply stirring a solution of the
carboxamide and PhI(OAc)2 in a mixture of MeCN and water
at rt resulted in complete conversion, but a significant amount
(15%) of symmetrical urea byproduct 24 was also formed
(Scheme 10). To minimize urea formation, HCl and TFA were
examined as acid additives in the reaction. Addition of HCl
provided little conversion to 1a and led to formation of other
unknown impurities. Pleasingly, in the presence of TFA, a
clean reaction profile was obtained with only minor amounts of
urea byproduct that could readily be removed upon
recrystallization from MTBE/MeCN.
To confirm the established trans-geometry of the fluoro and
amino groups of 1a, corresponding 4-bromobenzamide
derivative was prepared to obtain a single-crystal X-ray (Figure
2). The crystal structure determined was in agreement with our
assigned trans-configuration of 1a.
CONCLUSIONS
■
In summary, a seven-step synthesis of 1a was developed
starting from the widely available and inexpensive 4,4-
difluorocyclohexanecarboxylic acid 19 (Scheme 11). By
exploitation of an orthoester to protect the carboxylic acid,
Scheme 9. Preparation of Hydrogenation Precursor 16
E
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
a
Table 4. Screening of Cationic Rhodium Catalysts with Bidentate Phosphine Ligands
entry
ligand
DPPB
SL-J002-1
SL-J001-1
SL-J001-1
SL-W003-2
rac-BINAP
DPDBF
time (h)
A% 16
A% 17c
A% des-F-17c
1
2
3
4
5
6
7
8
12
12
4
4
4
12
12
12
100
14
29
18
11
20
14
19
0
86
71
54
36
74
86
77
0
0
0
28
53
6
b
0
4
DPPF
a
b
A% was determined by UPLC-MS. MeOH as solvent.
Table 5. Directed Hydrogenation of 16 with DPDBF
Scheme 10. Hofmann Rearrangement of Carboxamide 17c
to 1a
H2
% Rh
entry loading
% DPDBF pressure
temp (°C),
time (h)
%
yield
(3-Methyloxetan-3-yl)methyl 4,4-difluorocyclohex-
ane-1-carboxylate (20). To a solution of 4,4-difluorocyclo-
hexane-1-carboxylic acid (19) (200.0 g, 1.22 mol, 1.0 equiv) in
2-MeTHF (1.6 L, 8 vol) at 0 °C was charged 1,1′-
carbonyldiimidazole (CDI) (217.3 g, 1.34 mol, 1.1 equiv) in
six portions. (Note: rapid gas evolution was observed.) The
resulting solution mixture was warmed to rt and was stirred at
rt for 1 h, at which point 3-methyl-3-oxetanemethanol (182.3
mL, 1.83 mol, 1.5 equiv) was charged. The resulting reaction
mixture continued to stir at rt overnight. Upon reaction
completion, the reaction mixture was quenched with H2O (1.0
L, 5 vol). The organic layer was washed with water and
saturated NaHCO3 (1.0 L, 5 vol) to remove unreacted acid,
followed by a wash with 1 N HCl (1.0 L, 5 vol) and a final
wash with H2O (1.0 L, 5 vol) to remove residual CDI. The
organic layer was subsequently concentrated under reduced
pressure to an oil, at which point PhMe (200.0 mL, 1 vol) was
added and the resulting solution was further concentrated
under reduced pressure to afford 328.6 g of crude 20 as a
yellow oil (92.5% assay yield).
a
loading
(psi)
solvent
1
2
3
1.0
1.0
0.5
1.1
1.1
0.55
400
400
100
CH2Cl2
CH2Cl2
THF
40, 24
40, 12
60, 4
86
82
86
a
Isolated yields.
Table 6. Directed Hydrogenation of 16 with DTBPF
a
entry
scale (g)
% conversion
trans:cis
% yield
1
2
3
4
5
55
40
40
40
42
>95
>95
>95
>95
>95
71
>98:2
>98:2
>98:2
>98:2
>98:2
83:17
86
96
96
92
1-(4,4-Difluorocyclohexyl)-4-methyl-2,6,7-
trioxabicyclo[2.2.2]octane (21). To a solution of crude 20
(328.6 g, 1.13 mol, 1 equiv) in CH2Cl2 (1.6 L, 5 vol) at 0 °C
was charged BF3·OEt2 (34.9 mL, 0.28 mol, 25 mol %). The
resulting reaction mixture was slowly warmed to rt over 30 min
and continued to stir at rt for 12−14 h. When 20 was
consumed, Et3N (157.8 mL, 1.13 mol, 1 equiv) was charged.
94
b
6
4
b
N/D
a
Isolated yields. <0.1 mol % imidazole present in the vinyl fluoride
starting material.
F
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Figure 2. Thermal ellipsoid plot of 25 in the crystal structure.
Scheme 11. Optimized Synthesis of 1a
The reaction mixture continued to stir at the same temperature
for 30 min, at which point 1 N NaOH (657.2 mL, 2 vol) was
added to quench the reaction mixture. The organic layer was
washed with H2O (657.2 mL, 2 vol) and was concentrated
under reduced pressure to afford an oil, to which 2-MeTHF
(821.5 mL, 2.5 vol) was charged. The resulting slurry was
heated to 60−65 °C to reach homogeneity and was
subsequently cooled to rt over 1 h. Heptane (2.5 L, 7.5 vol)
was charged slowly. The resulting slurry was stirred for 1 h at
rt, further cooled to 0 °C over 30 min, and stirred for an
additional hour. Solids were collected by filtration, rinsed with
heptane (328.6 mL, 1 vol), and dried at 30 °C under vacuum
with a nitrogen bleed for 12 h to afford 205.8 g of 21 as an off-
white solid (crop 1). Mother liquor was concentrated under
reduced pressure to give crude solids, which were dissolved in
2-MeTHF (1 vol), and the solution was heated to 60−65 °C
to obtain a homogeneous solution. The solution was cooled to
rt to result in a slurry, to which was charged heptane (3 vol)
slowly. The slurry continued to stir at rt for 30 min, was cooled
to 0 °C over 30 min, and was stirred for an additional hour.
Solids were collected by filtration, rinsed with heptane (1 vol),
and dried at 30 °C under vacuum with a nitrogen bleed for 12
h to afford another 58.1 g of 21 as an off-white solid (crop 2),
providing a total of 263.9 g of 21 (93.9% yield); mp 133 °C
1
(from DSC). H NMR (CDCl3, 500 MHz): δ 3.87 (s, 6 H),
2.11−2.06 (m, 2 H), 1.91−1.88 (m, 2 H), 1.70−1.47 (m, 5 H),
0.79 (s, 3 H). 13C NMR (CDCl3, 125 MHz): δ 123.5 (dd, J =
238.4, 1.9 Hz), 109.3 (d, J = 2.2 Hz), 72.6, 42.3 (dd, J = 1.7
Hz), 33.2 (dd, J = 22.3, 2.8 Hz), 30.2, 22.9 (d, J = 9.8 Hz),
14.5. 19F NMR (CDCl3, 470 MHz): δ −96.7 (d, J = 235.7 Hz),
−101.9 (d, J = 237.6 Hz). HRMS, m/z: [M + H]+ calcd for
+
(C12H19F2O3 ), 249.12968; found, 249.12984.
1-(4-Fluorocyclohex-3-en-1-yl)-4-methyl-2,6,7-
trioxabicyclo[2.2.2]octane (22). A solution of 21 (166.6 g,
0.67 mol, 1 equiv) in KOt-Amyl as a 25% solution in PhMe
(753.0 mL, 1.34 mol (2 equiv) was heated to about 110−115
°C and was stirred at this temperature for 120 h. The reaction
was cooled to rt, quenched with H2O (333.2 mL, 2 vol), and
then diluted with EtOAc (666.4 mL, 4 vol). The organic layer
was washed with H2O (333.2 mL, 2 vol) and was subsequently
G
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
concentrated to about 2 vol under reduced pressure. Heptane
(333.2 mL, 2 vol) was charged and 2 vol of solvent was
removed under reduced pressure at 55 °C. This process was
repeated twice until a slurry was obtained. The resulting slurry
was cooled to rt, filtered, and rinsed with heptane (333.2 mL, 2
vol) to afford 134.4 g of 22 as a white solid (81.3% yield); mp
mL, 1.5 vol). The resulting slurry was cooled to about 10−15
°C and was stirred at the same temperature for 30 min. The
slurry was subsequently filtered. Solids were dried on filter for
45−60 min and were further dried at 55 °C under reduced
pressure with a nitrogen bleed for 12 h to afford 130.0 g of 16
1
as a white solid (93.2% yield). H NMR (DMSO-d6, 500
1
70 °C (from DSC). H NMR (CDCl3, 500 MHz): δ 5.16−
MHz): δ 7.29 (s, 1 H), 6.79 (s, 1 H), 5.22−5.19 (s, 1 H),
2.31−2.12 (m, 5 H), 1.89−1.86 (m, 1 H), 1.70−1.62 (m, 1 H).
13C NMR (DMSO-d6, 125 MHz): δ 176.5 (d, J = 2.1 Hz),
159.2 (d, J = 250.8 Hz), 101.5 (d, J = 15.6 Hz), 39.3 (d, J = 2.2
Hz), 25.7 (t, J = 9.3 Hz), 25.1, 24.9. 19F NMR (DMSO-d6, 470
MHz): δ −101.9. HRMS, m/z: [M + H]+ calcd for
(C7H11FNO+), 144.08192; found, 144.08204.
5.11 (m, 1 H), 3.88 (s, 6 H), 2.26−2.12 (m, 3H), 2.07−1.98
(m, 2 H), 1.82−1.76 (m, 1 H), 1.55−1.47 (1 H), 0.80 (s, 3 H).
13C NMR (CDCl3, 125 MHz): δ 159.4 (d, J = 251.6 Hz),
109.6, 101.1 (d, J = 15.9 Hz), 72.6, 40.0 (d, J = 2.2 Hz), 30.2,
25.3 (d, J = 23.9 Hz), 23.0 (d, J = 8.2 Hz), 22.7 (d, J = 9.1 Hz),
14.5. 19F NMR (CDCl3, 470 MHz): δ −104.2. HRMS, m/z:
+
[M + H]+ calcd for (C12H18FO3 ), 229.12345; found,
trans-4-Fluorocyclohexane-1-carboxamide (17c). To a
2-dram vial equipped with a stir bar were charged [Rh-
(nbd)2]BF4 (102 mg, 0.1 mol %), 1,2′-bis(di-tert-
butylphosphino)ferrocene (142 mg, 0.11 mol %), and 1 mL
CH2Cl2. The resulting dark red solution was stirred at rt for 15
min and then transferred by a pipet to a stainless steel reactor
containing 16 (40.2 g, 28 mmol, 1 equiv) and anhydrous THF
(400 mL, 10 vol). The reaction mixture was purged with
nitrogen three times followed by hydrogen three times and
then placed under 200 psi hydrogen. The mixture was heated
to 60 °C and stirred for 4 h and then cooled to rt. The product
mixture was transferred to a RB flask and concentrated to 5
vol. To the slurry was charged hexanes (200 mL, 5 vol). The
slurry was stirred at rt for 1 h. The solid was collected on a
filtration funnel and washed with 1:1 THF/hexanes (2 × 40
mL, 2 vol). The solid was dried at 50 °C under reduced
pressure with a nitrogen bleed for 12 h to afford 36.4 g of 17c
229.12362.
4-Fluorocyclohex-3-ene-1-carboxylic acid (15). To a
solution of 22 (134.4 g, 0.589 mmol, 1 equiv) in THF (537.6
mL, 4 vol) was charged 1 N HCl (176.6 mL, 0.177 mmol, 30
mol %). The resulting reaction mixture was stirred at rt. After
30−40 min, 22 was consumed and H2O (537.6 mL, 4 vol) was
charged, followed by addition of LiOH·H2O (74.1 g, 1.77 mol,
3 equiv). The resulting reaction mixture continued to stir at rt
for an additional 1.5−2 h, at which point MTBE (672 mL, 5
vol) was charged. The aqueous layer acidified with 6 N HCl
until pH 1−2 was reached and was subsequently extracted with
EtOAc (537.6 mL, 4 vol). The organic layer was washed with
H2O (268.8 mL, 2 vol). Three volumes of solvent was distilled
at 40 °C under reduced pressure. At this point heptane (268.8
mL, 2 vol) was charged and 2 vol of solvent was distilled at 40
°C under reduced pressure. This process was repeated once;
then heptane (537.6 mL, 1 vol) was charged to obtain a slurry.
The resulting slurry was cooled to 5−10 °C and was stirred at
the same temperature for 30−40 min. Solids were collected by
filtration, rinsed with heptane (537.6 mL, 1 vol), and dried at
40 °C under reduced pressure with a nitrogen bleed for 6 h to
afford 80.8 g of 15 as a white solid (93.3% yield); mp 59 °C
1
as a white solid (91.9% yield). H NMR (DMSO-d6, 500
MHz): δ 7.22 (s, 1 H), 6.71 (s, 1 H), 4.49 (d, 1 H, J = 48.9
Hz), 2.15−1.93 (m, 3 H), 1.86−1.68 (m, 2 H), 1.50−1.28 (m,
4 H). 13C NMR (DMSO-d6, 125 MHz): δ 176.9 (d, J = 2.6
Hz), 91.7 (d, J = 170.3 Hz), 42.4 (d, J = 1.5 Hz), 31.9 (d, J =
18.5 Hz), 26.9 (d, J = 11.4 Hz). 19F NMR (DMSO-d6, 470
MHz): trans-isomer, δ −168.0 (d, J = 44.5 Hz); cis-isomer, δ
−148.3 (d, J = 28.1 Hz). HRMS, m/z: [M + H]+ calcd for
(C7H13FNO+), 146.09757; found, 146.09769.
1
(from DSC). H NMR (DMSO-d6, 500 MHz): δ 12.26 (s, 1
H), 5.22−5.19 (m, 1 H), 2.48−2.43 (m, 1 H), 2.25−2.14 (m, 4
H), 2.00−1.96 (m, 1 H), 1.76−1.68 (m, 1 H). 13C NMR
(DMSO-d6, 125 MHz): δ 176.2 (d, J = 1.9 Hz), 159.3 (d, J =
250.9 Hz), 101.22 (d, J = 15.9 Hz), 38.1 (d, J = 2.2 Hz), 25.0
(dd, J = 8.2, 2.3 Hz), 24.6, 24.4. 19F NMR (DMSO-d6, 470
MHz): δ −101.6. HRMS, m/z: [M + H]+ calcd for
trans-4-Fluorocyclohexan-1-amine Hydrochloride
(1a). To a solution of PhI(OAc)2 (215.5 g, 669.1 mmol, 1.1
equiv) in MeCN (618.1 mL, 7 vol) was charged H2O (264.9
mL, 3 vol) and TFA (232.9 mL, 3.04 mol, 5 equiv). The
resulting reaction mixture was stirred at rt for 15−20 min, at
which point 17c (88.3 g, 608.2 mmol, 1 equiv) was charged in
one portion as solids. The resulting reaction mixture continued
to stir at rt overnight. Solvent was distilled to minimum
stirrable volume under reduced pressure. The mixture was
azeotroped with PhMe (883.0 mL, 10 vol) five times at 55 °C
under reduced pressure to remove residual water from the
mixture. To the mixture was charged MTBE (883.0 mL, 10
vol), and solvent was distilled to a minimum stirrable volume.
This process was repeated twice. To the mixture was charged
MTBE (883.0 mL, 10 vol), followed by 4 M HCl in dioxane
(228.1 mL, 912.3 mmol, 1.5 equiv) at a rate that maintained
the internal temperature below 30 °C. The resulting mixture
was stirred at rt for 1 h. Solids were collected by filtration and
rinsed with MTBE (176.6 mL, 2 vol). Crude solids were
suspended in MTBE (176.6 mL, 2 vol) and MeCN (353.2 mL,
4 vol). The resulting mixture was heated to 65 °C and was
stirred at the same temperature for 15−20 min, at which point
it was cooled to rt and stirred for an additional 15 min. Solids
were collected by filtration and rinsed with 1:1 MTBE:MeCN
+
(C7H10FO2 ), 145.06593; found, 145.06606.
4-Fluorocyclohex-3-ene-1-carboxamide (16). To a
solution of 15 (153.0 g, 950.9 mmol, 1 equiv) in MeCN
(765.0 mL, 5 vol) was charged 1,1′-carbonyldiimidazole (169.6
g, 162.1 mmol, 1.1 equiv) in six portions at room temperature.
(Note: gas evolution was observed.) The resulting mixture was
stirred at rt for 1 h, at which point the reaction mixture was
cooled to 0−5 °C. To the cooled stirring reaction mixture was
charged NH4OH (595.0 mL, 4.8 mol, 4.5 equiv) at a rate to
maintain the internal temperature below 25 °C. The reaction
mixture was subsequently warmed to rt and continued to stir at
the same temperature for 30−40 min. Five volumes of solvent
was distilled at 55 °C under reduced pressure followed by
addition of heptane (765.0 mL, 5 vol). This process was
repeated to result in a slurry, at which point a solution of 15%
NaCl in 3 N HCl (900.0 mL, 6.3 vol) was charged at a rate
that maintained the internal temperature below 40 °C. The
resulting solution mixture was stirred for 2 h, at which point
solids were collected by filtration and rinsed with heptane
(306.0 mL, 2 vol). Solids were suspended in cold H2O (229.5
H
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
(b) Vincent, J.-M. Recent advances of fluorous chemistry in material
sciences. Chem. Commun. 2012, 48, 11382−11391.
(176.6 mL, 2 vol) followed by MTBE (176.6 mL, 2 vol).
Filtered solids were dried at 30 °C under reduced pressure
with a nitrogen bleed for 12 h to afford 88.4 g of 1a as an off-
white solid (94.6% yield). 1H NMR (DMSO-d6, 500 MHz): δ
8.13 (s, 3 H), 4.61−4.46 (m, 1 H), 3.05−3.00 (m, 1 H), 2.06−
1.95 (m, 4 H), 1.55−1.38 (m, 4 H). 13C NMR (DMSO-d6, 125
MHz): δ 90.6 (d, J = 169.6 Hz), 48.1 (d, J = 1.4 Hz), 29.8 (d, J
= 19.9 Hz), 27.4 (d, J = 11.3s Hz). 19F NMR (DMSO-d6, 470
MHz): δ −172.1. HRMS, m/z: [M−HCl+ H]+ calcd for
(C6H13F+), 118.10265; found, 118.10275.
(3) For selected reviews: (a) Jeschke, P. The unique role of halogen
substituents in the design of modern agrochemicals. Pest Manage. Sci.
2010, 66, 10−27. (b) Jeschke, P. In Modern Methods in Crop
Protection Research; Jeschke, P., Kramer, W., U. Schirmer, U.,
Witschel, M., Eds.; Wiley-VCH: Weinheim, Germany, 2012; pp
73−128. (c) Fujiwara, T.; O’Hagan, D. Successful fluorine-containing
herbicide agrochemicals. J. Fluorine Chem. 2014, 167, 16−29.
(4) (a) Kirk, K. L. Fluorination in Medicinal Chemistry: Methods,
Strategies, and Recent Developments. Org. Process Res. Dev. 2008, 12,
́
́
305−321. (b) Wang, J.; Sanchez-Rosello, M.; Acena, J. L.; del Pozo,
̃
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00444.
C.; Sorochinsky, A. E.; Fustero, S.; Soloshonok, V. A.; Liu, H.
Fluorine in Pharmaceutical Industry: Fluorine-Containing Drugs
Introduced to the Market in the Last Decade (2001−2011). Chem.
Rev. 2014, 114, 2432−2506. (c) Zhou, Y.; Wang, J.; Gu, Z.; Wang, S.;
■
sı
Zhu, W.; Acena, J. L.; Soloshonok, V. A.; Izawa, K.; Liu, H. Next
̃
1H, 13C, and 19F NMR spectra (PDF)
Generation of Fluorine-Containing Pharmaceuticals, Compounds
Currently in Phase II−III Clinical Trials of Major Pharmaceutical
Companies: New Structural Trends and Therapeutic Areas. Chem.
Rev. 2016, 116, 422−518.
Crystallographic data for 25 (CIF)
AUTHOR INFORMATION
Corresponding Author
Joyce C. Leung − Chemical Development, Boehringer
Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut
06877-0378, United States; orcid.org/0000-0003-2246-
1599; Email: joyce.leung@boehringer-ingelheim.com
(5) (a) Ma, J.-A.; Cahard, D. Asymmetric Fluorination, Trifluor-
omethylation, and Perfluoroalkylation Reactions. Chem. Rev. 2004,
104, 6119−6146. (b) Yang, X.; Wu, T.; Phipps, R. J.; Toste, F. D.
Advances in Catalytic Enantioselective Fluorination, Mono-, Di-, and
Trifluoromethylation, and Trifluoromethylthiolation Reactions. Chem.
Rev. 2015, 115, 826−870. (c) Zhu, Y.; Han, J.; Wang, J.; Shibata, N.;
Sodeoka, M.; Soloshonok, V. A.; Coelho, Jaime A. S.; Toste, F. D.
Modern Approaches for Asymmetric Construction of Carbon−
Fluorine Quaternary Stereogenic Centers: Synthetic Challenges and
Pharmaceutical Needs. Chem. Rev. 2018, 118, 3887−3964 and
references therein..
(6) (a) Nielsen, M. K.; Ahneman, D. T.; Riera, O.; Doyle, A. G.
Deoxyfluorination with Sulfonyl Fluorides: Navigating Reaction Space
with Machine Learning. J. Am. Chem. Soc. 2018, 140, 5004−5008.
(b) Shoup, T. M.; Bonab, A. A.; Vasdev, N. Synthesis and preclinical
evaluation of [18F]FCHC for neuroimaging of fatty acid amide
hydrolase. Mol. Imaging Biol. 2015, 17, 257−263.
■
Authors
Thach T. Nguyen − Chemical Development, Boehringer
Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut
06877-0378, United States
Mariusz Krawiec − Material and Analytical Sciences,
Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield,
Connecticut 06877-0378, United States
Donghong A. Gao − Material and Analytical Sciences,
Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield,
Connecticut 06877-0378, United States
Jonathan T. Reeves − Chemical Development, Boehringer
Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut
06877-0378, United States; orcid.org/0000-0002-3051-
3304
(7) Wiesenfeldt, M. P.; Nairoukh, Z.; Li, W.; Glorius, F.
Hydrogenation of fluoroarenes: Direct access to all-cis-(multi)-
fluorinated cycloalkanes. Science 2017, 357, 908−912.
(8) Schultz, A. G.; McCloskey, P. J. Carboxamide and carbalkoxy
group directed stereoselective iridium-catalyzed homogeneous olefin
hydrogenations. J. Org. Chem. 1985, 50, 5905−5907.
(9) (a) Hara, S. In Stereoselective Alkene Synthesis; Wang, J., Eds.;
Springer: Berlin, Heidelberg, 2012; pp 59−86. (b) Drouin, M.;
Hamel, J.-D.; Paquin, J.-F. Synthesis of Monofluoroalkenes: A Leap
Forward. Synthesis 2018, 50, 881−955 and references therein..
(10) Vandamme, M.; Paquin, J.-F. Eliminative Deoxofluorination
Using XtalFluor-E: A One-Step Synthesis of Monofluoroalkenes from
Cyclohexanone Derivatives. Org. Lett. 2017, 19, 3604−3607.
(11) (a) Huang, X.; Li, H.; Ma, L.; Xu, X. Preparation of
halogenated hydrocarbons by splitting-off hydrogen halides from
halogenated hydrocarbons. Patent CN 106316777A, 2017. (b) Bur-
don, J.; Ezmirly, S. T.; Huckerby, T. N. The fluorination of Butane
over cobalt trifluoride. J. Fluorine Chem. 1988, 40, 283−318.
(c) Achilefu, S.; Mansuy, L.; Selve, C.; Thiebaut, S. Synthesis of
2H,2H-perfluoroalkyl and 2H-perfluoroalkenyl carboxylic acids and
amides. J. Fluorine Chem. 1995, 70, 19−26. (d) Strobach, D. R.;
Boswell, G. A., Jr. Synthesis of 1-fluorocycloalkenes. J. Org. Chem.
1971, 36, 818−820. (e) Anilkumar, R.; Burton, D. J. An efficient
dehyrohalogenation method for the synthesis of α,β,β-trifluorostyr-
enes, α-chloro-β,β-difluorostyrenes and E-1-arylperfluoroalkenes. J.
Fluorine Chem. 2005, 126, 1174−1184.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00444
Author Contributions
#The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript. These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Mr. Max Sarvestani and Mr. Scott Pennino
for assistance in DSC and HRMS analysis, respectively.
REFERENCES
■
(1) Caron, S. Where Does the Fluorine Come From? A Review on
the Challenges Associated with the Synthesis of Organofluorine
Compounds. Org. Process Res. Dev. 2020, 24, 470−480 and references
therein..
(12) Krska, S. W.; Mitten, J. V.; Dormer, P. G.; Mowrey, D.;
Machrouhi, F.; Sun, Y.; Nelson, T. D. Enantioselective synthesis of a
chiral fluoropiperidine via asymmetric hydrogenation of a vinyl
fluoride. Tetrahedron 2009, 65, 8987−8994.
(2) (a) Berger, R.; Resnati, G.; Metrangolo, P.; Weber, E.; Hulliger,
J. Organic fluorine compounds: a great opportunity for enhanced
materials properties. Chem. Soc. Rev. 2011, 40, 3496−3508.
I
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
(13) (a) Zhang, L.-H.; Kauffman, G. S.; Pesti, J. A.; Yin, J.
Rearrangement of Nα-Protected L-Asparagines with Iodosobenzene
Diacetate. A Practical Route to β-Amino-l-alanine Derivatives. J. Org.
Chem. 1997, 62, 6918−6920. (b) Abrecht, S.; Adam, J.-M.;
Bromberger, U.; Diodone, R.; Fettes, A.; Fischer, R.; Goeckel, V.;
Hildbrand, S.; Moine, G.; Weber, M. An Efficient Process for the
Manufacture of Carmegliptin. Org. Process Res. Dev. 2011, 15, 503−
514.
J
https://dx.doi.org/10.1021/acs.oprd.0c00444
Org. Process Res. Dev. XXXX, XXX, XXX−XXX