Selective Continuous Flow Iodination Guided by Direct Spectroscopic Observation of Equilibrating Aryl Lithium Regioisomers

Article abstract of DOI:10.1021/acs.organomet.8b00538

The iodination of 4-fluoro-2-(trifluoromethyl)benzonitrile via C-H lithiation and subsequent treatment with iodine under continuous flow conditions is described. Screening identified both LDA and PhLi as effective bases, giving the desired 3-iodo regioisomer as the major product. Use of LDA results in varying amounts of the undesired 5-iodo isomer, while PhLi results in more reliable formation of the 3-iodo product. An initial flow process was developed using PhLi that produced 4-fluoro-3-iodo-2-(trifluoromethyl)benzonitrile in 63% yield on a gram scale. Process modifications to enable pilot-scale operation resulted in a yield decrease to <50%, persistent formation of a byproduct resulting from PhLi addition to the nitrile, and formation of solids during longer runs. As a result, the use of LDA was investigated under continuous flow conditions. In situ NMR and IR spectroscopy allowed observation of the 5-[Li] species and its conversion to the thermodynamically preferred 3-[Li] species. These mechanistic insights drove development of a second-generation continuous flow process using LDA that achieves 30:1 regioselectivity and an 84% solution yield of the desired product (67% isolated yield after recrystallization). Furthermore, this process increases throughput by 10-fold, providing a path to manufacturing-scale operation.

Full text of DOI:10.1021/acs.organomet.8b00538

Article
pubs.acs.org/Organometallics
Cite This: Organometallics XXXX, XXX, XXX−XXX
Selective Continuous Flow Iodination Guided by Direct
Spectroscopic Observation of Equilibrating Aryl Lithium
Regioisomers
Anna L. Dunn,*,† David C. Leitch,*,† Michel Journet,† Michael Martin,‡ Elie A. Tabet,‡
Neil R. Curtis,§ Glynn Williams,§ Charles Goss,∥ Tony Shaw,† Bernie O’Hare,⊥ Charles Wade,§
Matthew A. Toczko,‡ and Peng Liu†
†API Chemistry, GlaxoSmithKline, King of Prussia, Pennsylvania 19406, United States
‡Early Development Sciences, GlaxoSmithKline, Research Triangle Park, North Carolina 27709, United States
§API Chemistry, GlaxoSmithKline, Stevenage, U.K. SG1 2NY
∥Process Analytical Technologies, GlaxoSmithKline, King of Prussia, Pennsylvania 19406, United States
⊥Global Spectroscopy, GlaxoSmithKline, King of Prussia, Pennsylvania 19406, United States
S
* Supporting Information
ABSTRACT: The iodination of 4-fluoro-2-(trifluoromethyl)-
benzonitrile via C−H lithiation and subsequent treatment with
iodine under continuous flow conditions is described. Screen-
ing identified both LDA and PhLi as effective bases, giving the
desired 3-iodo regioisomer as the major product. Use of LDA
results in varying amounts of the undesired 5-iodo isomer,
while PhLi results in more reliable formation of the 3-iodo
product. An initial flow process was developed using PhLi that
produced 4-fluoro-3-iodo-2-(trifluoromethyl)benzonitrile in
63% yield on a gram scale. Process modifications to enable
pilot-scale operation resulted in a yield decrease to <50%,
persistent formation of a byproduct resulting from PhLi
addition to the nitrile, and formation of solids during longer
runs. As a result, the use of LDA was investigated under
continuous flow conditions. In situ NMR and IR spectroscopy allowed observation of the 5-[Li] species and its conversion to
the thermodynamically preferred 3-[Li] species. These mechanistic insights drove development of a second-generation
continuous flow process using LDA that achieves 30:1 regioselectivity and an 84% solution yield of the desired product (67%
isolated yield after recrystallization). Furthermore, this process increases throughput by 10-fold, providing a path to
manufacturing-scale operation.
While this chemistry was suitable to support discovery
chemistry efforts, affording 3 in 44% isolated yield on a gram
scale, attempts to scale up to kilogram quantities resulted in a
precipitous reduction in yield (11%). This was due to
decomposition and formation of multiple byproducts (4, 5,
and others not identified) despite considerable effort to modify
the reaction conditions from those initially developed. Our
hypothesis was that the ArLi species is unstable under the
reaction conditions and therefore decomposed during the
longer addition times and localized hot spots common to larger
scale batch reactions,5 especially given the exothermicity of the
INTRODUCTION
■
Halogenation of organometallic reagents is a powerful method
to form C−X bonds in a highly regioselective manner.1 This is
particularly true for halogenation of electron-deficient
aromatics, which will not readily undergo electrophilic
aromatic substitution but are often excellent substrates for
C−H metalation using strong bases such as lithium
diisopropylamide (LDA).2 As part of development efforts
toward a selective androgen receptor modulator (SARM)3
clinical candidate, we required large amounts of the
tetrasubstituted iodoarene 3 as an intermediate in the
preparation of the active pharmaceutical ingredient (API)
(Scheme 1). Prior work in our medicinal chemistry group
focused on iodination of an ArLi species (2a) generated by
regioselective metalation of 4-fluoro-2-(trifluoromethyl)-
benzonitrile (1) using LDA.4
Special Issue: The Roles of Organometallic Chemistry in Pharma-
ceutical Research and Development
Received: July 30, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
K), KOtBu, iPrMgCl, and PhMgBr. With the exception of LDA
and PhLi, these bases led to either no conversion of starting
material or formation of new byproducts (tentatively identified
Scheme 1. Initial Batch Synthesis of Iodoarene 3 with
Byproducts 4 and 5
n
as attack of the nitrile by nucleophilic bases such as BuLi).
Table 1 summarizes results obtained for LDA and PhLi with
amounts of 1, 3, and 4 determined by HPLC assay or 19F
Table 1. Effect of Base Identity, Amounts, and Temperature
on Yield and Regiochemistry of Iodination
step 1 time
(min)
a
b
b
b
entry base (equiv) T (°C)
1 (%) 3 (%) 4 (%)
c
1
2
3
4
LDA (0.95)
LDA (1.05)
LDA (1.05)
LDA (0.95)
LDA (1.05)
LDA (1.05)
LDA (1.05)
LDA (1.05)
LDA (1.05)
LDA (1.05)
PhLi (1.00)
PhLi (1.00)
PhLi (0.95)
PhLi (1.00)
PhLi (1.00)
−70
−45
−45
−45
−45
−45
−30
−30
−30
−10
−70
−45
−45
−45
−30
2
0.33
2
2
5
10
0.33
2
5
0.33
5
20
13
10
10
10
7
12
8
11
10
9
45
56
63
65
75
72
63
65
48
52
74
50
44
59
28
17
20
14
3
4
3
2
2
2
2
c
5
6
7
8
iodination step (∼20 °C temperature rise observed in gram-
scale runs).
Although alternative routes were evaluated, they all resulted
in a significant increase in material costs, number of synthetic
transformations, or both. Process chemistry efforts therefore
focused on developing a reliable, large-scale method to prepare
3, requiring an in-depth study of this transformation to
determine the cause of poor scalability. Herein we describe the
development of a regioselective metalation/iodination that is
amendable to multi-kilogram-scale preparation of 3, focusing
on the use of continuous flow to alleviate mixing and energy
transfer issues associated with the batch chemistry.6 Key to the
success of the iodination is generation of the correct
organometallic regioisomer (2a), which we have determined
is the thermodynamically most stable ArLi species.7 By
combining small-scale screening experiments with operando
NMR and IR spectroscopy, we identified the key factors that
lead to a selective and higher-yielding halogenation. These
insights drove development of continuous processes using both
PhLi and LDA as metalation reagents, enabling access to 3 in
up to 67% isolated yield with a throughput of 425 g L−1 h−1.
9
10
11
12
13
14
15
2
d
0.33
2
5
11
17
12
7
8
c
1
1
2
e
0.33
a
Abbreviations: LDA, lithium diisopropylamide (prepared fresh from
nBuLi/HN(iPr)2 as a 1.0 M solution in THF/hexanes); PhLi,
phenyllithium (commercial 1.8 M solution in nBu2O). Unless
otherwise noted, the base was added in one portion to a solution of
b
c
1 (0.187 mmol). Solution yield determined by HPLC assay. Base
added dropwise to a solution of 1 (0.800 mmol, 4× scale); solution
yields determined by 19F NMR spectroscopy versus internal standard.
d
Coelution of 4 and PhI (6) resulted in an inflated assay yield of 4.
e
The major product is imine 7 (Scheme 2).10
NMR spectroscopy versus internal standard. These data clearly
show that metalation of 1 is rapid with either base, even at −70
°C. Using LDA, the amount of 4 generated is dependent on
reaction time, temperature, and amount of base. At −70 °C, a
2 min deprotonation time gives 80% conversion of 1 and a
∼2.5:1 ratio of 3 to 4 (entry 1). Increasing the temperature to
−45 °C increases the conversion to ∼90% even with only a 20
s metalation time; however, the ratio of 3 to 4 is still <3:1
(entry 2). Increasing the metalation time to 2 min improves
the product ratio (entry 3), which is further improved by
reducing the LDA charge (entry 4). The highest solution yield
(75%) and best regioisomer ratio (19:1) is obtained with a 5
min metalation time at −45 °C (entry 5); increasing the
reaction time to 10 min did not result in higher solution yield
or improved regioisomer ratio (entry 6). At −30 °C, 20 s is
sufficient to achieve a 63% solution yield of 3 (entry 7), with
increased reaction times eventually leading to a drop in yield
due to decomposition (entry 9). At −10 °C, 20 s is too long to
prevent decomposition (entry 10), and attempts to metalate at
0 °C even at short reaction times led to complex mixtures with
small quantities of 3 present.
RESULTS AND DISCUSSION
■
Batch Reaction Screening and Development. Given
the scalability issues observed with the lithiation/iodination
sequence described in Scheme 1, several other iodination
approaches that would avoid the use of sensitive organo-
metallic intermediates were investigated. Unfortunately,
electrophilic iodination does not proceed using reagents such
as ICl and NIS, undoubtedly due to the electron-deficient
nature of 1.8 Catalytic C−H functionalization using Pd, Rh, or
Ir systems under conditions described in the literature also
failed to give the desired product.9 Rather than completely
redesign the synthesis of the final API, we initiated a more
thorough evaluation of metalation/iodination conditions to
determine the causes for poor scalability.
A series of small-scale screening experiments was designed
to evaluate the base, stoichiometry, temperature, and metal-
ation time. Many bases other than LDA were evaluated,
n
including PhLi, BuLi, the hexamethyldisilazide series (Li, Na,
B
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Metalation/iodination with PhLi gave similar results, with
two important distinctions. The levels of 4 observed are lower
than with LDA (compare Table 1, entries 2 and 12), generally
between only 1 and 2%, and there are two additional
byproducts generated (Scheme 2). Iodobenzene (6) results
studied previously, indicating that diisopropylamine is a
capable catalyst;2c−g unreacted 1 could also initiate this
interconversion by acting as a proton source. That 2a is
thermodynamically more stable is consistent with prior
thermodynamic studies of the directing ability of fluorine
and −CF3 groups in multiply fluorinated aromatics conducted
by Schlosser and co-workers.7b,c For trifluorotoluene, the
energy difference between the o- and p-lithium species is 2.7
kcal/mol at −75 °C in favor of the ortho regioisomer,7c which
matches well with the generally high regioselectivity observed
(for example, Table 1, entry 11: ∼35:1 ratio of 3 to 4 at −70
°C).
Scheme 2. Reaction Pathways for Lithiation/Iodination
a
Using PhLi or LDA and Origin of Byproducts 4, 6, and 7
With LDA, shorter metalation reaction times and lower
temperatures lead to an increase in the amount of 4, leading us
to suspect that 2b may be the kinetically preferred metalation
product using this base at −70 °C, likely due to a combination
of sterics and lithium speciation (vide infra).2g Unfortunately,
the typical strategy of using higher reaction temperatures and
longer times to ensure equilibration is not compatible with the
observed thermal instability of these species. In contrast, using
PhLi appears to directly generate a thermodynamic ratio of 2a/
2b; this could be through preferential metalation at the 3-
position and/or rapid equilibration.2e−g
Initial Continuous Process Development. From the
data obtained in small-scale screening experiments, it is clear
that careful control over short reaction times, maintenance of
low temperatures during exothermic processes, and precise
reaction stoichiometry are critical to ensuring the maximum
yield of 3 regardless of whether LDA or PhLi is used.
Regioisomer 4 is the most difficult contaminant to remove by
crystallization; we therefore focused our development efforts
on minimizing formation of this byproduct. Because the use of
PhLi consistently leads to high regioselectivity, and the
formation of 6 and 7 can be controlled with temperature
and stoichiometry, initial process development efforts focused
on these conditions.
Given the previous difficulties encountered in scaling this
reaction, and the need to tightly control the reaction
parameters and times, we eschewed batch chemistry develop-
ment for a flow chemistry approach.6 The rapid reaction
kinetics for both metalation and iodination are ideal for short
residence time flow chemistry processes, and maintaining low
reaction temperatures is much easier in smaller reactors with
low reactive inventory, even for exothermic reactions. Many
research groups in both academia and industry have taken
advantage of flow chemistry for effecting difficult metalations,
a
Thermally sensitive intermediates are shown in the red box.
from iodination of unreacted PhLi, which can be mitigated by
avoiding the use of excess base and ensuring a long enough
reaction time to completely consume PhLi (compare entries
12 and 14). The primary imine 7 is formed through
competitive attack of PhLi at the nitrile, analogous to
decomposition observed with other nucleophilic bases (vide
supra). This side pathway is more prevalent at elevated
temperatures, contributing to lower solution yields at −45 and
−30 °C in comparison with LDA (compare entries 5 and 14
and entries 7 and 15). In fact, with PhLi at −30 °C the major
product is 7.10
To account for the observations outlined in Table 1, we
propose the reaction pathways in Scheme 2. Using LDA or
PhLi to metalate 1 results in a mixture of equilibrating ArLi
regioisomers 2a and 2b, where 2a is thermodynamically
preferred. This equilibration must be catalyzed by a proton
source: equilibration of ArLi species using LDA has been
Scheme 3. Schematic of Lab-Scale Flow Process for the Synthesis of 3 Using PhLi
C
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
including those that generate unstable intermediates that must
be quenched quickly.11,12 The use of microreactors is one
approach to this problem;6a,11c however, our need for large-
scale production is incompatible with the throughput afforded
by microreactor systems.12 Finally, one key challenge with
organometallic chemistry in flow is the frequent generation of
solids that lead to reactor clogging and overpressurization.
While technologies exist to handle solid-containing streams,
avoiding the formation of insoluble materials is preferable.
A laboratory-scale flow reactor was assembled according to
Scheme 3. The feed lines and reactor loops (10 mL) were
made of fluoropolymer tubing, all of which were submerged in
a cold bath at −70 °C. Standard syringe or HPLC pumps were
used to deliver the reagent solutions, which were precooled in
1 mL loops before being mixed at simple T-junctions. A
targeted set of conditions based on those from entry 11 in
Table 1 were evaluated, maintaining the stoichiometry at 1.0
equiv of PhLi and 1.4 equiv of I2. In this reactor at −70 °C, a 5
min residence time for deprotonation (τ1) is not sufficient to
reach >50% conversion of 1. This apparently slower reactivity
was attributed to improved heat transfer: even in small-scale
batch experiments, the exothermic addition of PhLi could
result in an increased internal reaction temperature, giving a
faster apparent rate. Increasing τ1 to 20 min by lowering the
flow rates enables ∼90% conversion of 1, with >15:1 selectivity
for the desired regioisomer 3. Batch workup and isolation of a
150 mL fraction of flow output resulted in 63% yield after
crystallization; this corresponds to a throughput (after
purification) of 57 g L−1 h−1.
2b as a function of time, temperature, and reaction
stoichiometry was therefore critical. To confirm our hypothesis
that 2b is the initial metalation product with LDA, which
equilibrates to the thermodynamically preferred 2a, we used
low-temperature NMR spectroscopy to directly observe these
species (Figure 1). Treatment of 1 with 0.95 equiv of LDA at
Figure 1. 1H NMR spectra (500 MHz, −70 °C, d8-THF) of 1 and (a)
0.95 or (b) 1.1 equiv of LDA. Poor selectivity in (b) is likely due to
the sample warming above −70 °C during injection.
After this gram-scale demonstration, the suitability of these
conditions for multi-kilogram production of 3 was evaluated.
Three major issues with the process outlined in Scheme 3 were
identified: formation of the imine impurity 7, leading to the
need for low temperatures; the tendency of the reaction stream
to generate solids during prolonged operation; and the low
material throughput. Addressing this last point was critical: at
57 g L−1 h−1, preparation of 10 kg of 3 using our ∼150 mL
pilot-scale reactor10 would require almost 50 days of
continuous processing time.
In order to increase the reaction rate, and operate within the
temperature range of our pilot reactor, we raised the reaction
temperature to −50 °C. While this enables a reduction in total
residence time to 5 min, the isolated yield of 3 under these
conditions dropped to <50%. This is consistent with our small-
scale screening experiments, where the solution assay of 3
drops 15% when the metalation temperature is increased from
−70 °C to −45 °C (Table 1, compare entries 11 and 14); the
drop in yield is due primarily to increased formation of 7. We
were able to supply 6.85 kg of 3 using this chemistry, but
prolonged operation was not possible due to the formation of
solids. One other practical concern manifested during our
initial scale-up: THF solutions of iodine are not stable at
ambient temperatures due to polymerization of the solvent.13
We were able to mitigate this issue by holding the iodine feed
solution at −16 °C and using it within 3 days; however, use of
an alternative solvent would be preferable. All of these factors
drove us to develop a more robust process for long-term
supply of 3.
−70 °C results in one major species with spectral features
consistent with 2a (Figure 1a). In contrast, using a slight excess
of LDA produces both 2a and a second species consistent with
2b in a 1.25:1 ratio; relative integration of the signals for 2a
and 2b in the former case gives a ∼10:1 ratio.14 Quenching
these solutions with I2 generates mixtures of 3 and 4 consistent
with these assignments.10
While these NMR spectroscopic results support our
hypothesis of thermodynamic control over regioselectivity in
the metalation of 1, we were unable to observe the initial
metalation event or conversion between the two regioisomers
due to rapid metalation occurring prior to spectral acquisition.
An operando ATR-FTIR spectroscopic method capable of 10 s
acquisition time allowed observation of the kinetics of the
metalation and equilibration.15 An iterative metalation−
quench protocol was employed to correlate IR spectral features
to iodinated product distribution, revealing signals at 1105 and
1130 cm−1 that correspond to 2a and 2b, respectively. Figure 2
shows a representative reaction profile for the metalation of 1
conducted using an inverse addition (solution of 1 added
dropwise to a solution of 0.95 equiv of LDA) at −70 °C. The
profile clearly shows rapid initial formation of 2b followed by
slower equilibration to 2a. At −70 °C this equilibration takes
approximately 20 min to reach steady state; at −45 °C, both
2b and 2a are generated within 10 s, and complete
equilibration to 2a is complete in <90 s.10 The depletion of
2a upon addition of I2 is also extremely rapid and is complete
essentially upon mixing, even at −70 °C.
A titration experiment at −70 °C monitored by operando
ATR-FTIR confirms that 2b is the major regioisomer in the
presence of excess LDA at low temperature, and equilibration
occurs only once excess 1 is present (Figure 3). In fact, at the
titration end point (Figure 3, point f), equilibration
commences but stops prematurely. At low [1] and low
Revisiting LDA: Direct Observation of Equilibrating
ArLi Isomers. Given that the formation of 7 is mostly
responsible for the loss in yield, we re-evaluated the suitability
of LDA for a continuous process. Understanding the factors
that control the formation of putative ArLi regioisomers 2a and
D
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 4. Stability of 2a at a variety of temperatures as monitored by
ATR-FTIR. A THF solution of 1 was added dropwise to a solution of
LDA at the time indicated, followed by holding the solution at the
specified temperature. Note that the initial decrease in 2a signal upon
warming to −45 °C is due to temperature-induced changes in
absorbance intensity, and the apparent increase in 2b at −10 °C is
indicative of decomposition, not interconversion of isomers. Peak
heights are relative to absorbance at 1092 cm−1 (single-point
baseline).
−45 °C and holding for >20 min; a separate experiment
indicates prolonged hold times at −45 °C do result in slow
decomposition.10 Further warming to −10 °C caused multiple
changes to the IR spectral features, including a large decrease
in 2a signal, reflecting instability. Thus, efficient heat transfer is
critical to the success of this chemistry. On a small scale and in
flow reactors, maintaining an internal temperature of −45 °C is
readily achievable,6,11,12 whereas in large batch reactors, local
temperature spikes would decompose 2a.5 Furthermore, long
addition times of both the base and iodine solutions would be
required to control these highly exothermic reactions, which
are not compatible with the thermal instability of 2a over time,
even at −45 °C. The root cause of the poor yield shown in
Scheme 1 is almost certainly the highly exothermic nature of
the initial metalation and I2 addition.
Figure 2. (a) Observation of the two aryllithium species under batch
conditions via ATR-FTIR at −70 °C (2a, 1105 cm−1; 2b, 1130 cm−1).
(b) Reaction progress plot. A THF solution of 1 was added over 1
min to a solution of LDA at the time indicated, and the I2 solution
was added dropwise at 50 min. Peak heights are relative to absorbance
at 1092 cm−1 (single-point baseline).
On the basis of the data for metalation with LDA, we
redesigned our lab-scale flow system to incorporate smaller
volume reactors for shorter residence times (Figure 5). All of
the reactor tubing was composed of ethylene tetrafluoro-
ethylene (ETFE, 1/16 in. outer diameter) for efficient heat
transfer and chemical compatibility. Rather than using simple
T-junctions, we incorporated in-line stainless steel active
mixers (magnetically driven agitation in a tube, or MDAT) to
ensure effective mixing, even at low flow rates.16 Each reagent
line has a precooling loop prior to mixing, and the metalation
residence time is more than long enough for equilibration to be
complete, ensuring maximum [2a]. The reaction with I2 occurs
immediately upon mixing; therefore, we shortened the
residence time to 16 s to increase throughput. The output
stream is directly quenched into a stirred solution of aqueous
sodium sulfite to destroy the excess I2; failure to do so resulted
in formation of poly-THF, which negatively affects product
isolation.
Figure 3. Titration of a THF solution of 1 into an LDA solution at
−70 °C monitored by ATR-FTIR. At each point a−d, 0.21 equiv of 1
was added. At point e, 0.11 equiv of 1 was added, and 0.053 equiv of 1
was added at points f and g, totaling 1.06 equiv of 1 added. Note that
at point f LDA and 1 are equimolar, leading to incomplete conversion
to 2a. Peak heights are relative to absorbance at 1092 cm−1 (single-
point baseline).
[LDA], the rate of reaction between 1 and LDA is similar to
the rate of equilibration, leaving unreacted 1 available to
mediate the isomerization; as LDA finally consumes 1, there is
no longer an available proton source, and equilibration stops.
These insights were used to set the stoichiometry (0.95 equiv
of LDA) and residence time (4 min for metalation) for our
subsequent flow experiments to ensure complete equilibration
before quenching with iodine.
The optimized flow process was performed with the
following feed solutions: 0.33 M 1 in anhydrous THF (1
equiv), 0.31 M freshly prepared LDA in hexanes/anhydrous
THF (0.95 equiv), and 0.49 M I2 (1.5 equiv) in anhydrous 2-
MeTHF with 10% toluene as an HPLC internal standard. This
switch to 2-MeTHF alleviates the polymerization issue
We then assessed the stability of 2a at increasing
temperatures (Figure 4). The ATR-FTIR signal for 2a remains
steady following formation at −70 °C and after warming to
1
observed with THF: only trace polymer was observed by H
E
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
which requires short reaction times, tight control over
temperature and stoichiometry, and the intermediacy of
unstable and interconverting organometallic species. The use
of PhLi does give lower amounts of the regioisomeric impurity
4 under a wider range of conditions; however, direct attack of
the phenyl anion at the nitrile leads to a significant amount of
another byproduct (7), lowering the overall yield. In contrast,
the use of LDA as the metalation reagent leads to the
formation of two regioisomeric ArLi species. Direct observa-
tion of these ArLi intermediates, and the kinetics of their
interconversion by operando IR spectroscopy, guided our
efforts to develop the LDA-based sequence. This flow process
has a throughput of 425 g L−1 h−1, which upon scale-up will
meet our material production requirements. We have
performed the chemistry on a multigram scale with 84%
solution yield of the desired product and 67% isolated yield
after recrystallization. As demonstrated by this work, the use of
spectroscopy to observe reactive intermediates and study rapid
kinetics is a powerful means for continuous flow process
development and understanding.
EXPERIMENTAL SECTION
■
Materials. All common reagents, solvents, and compound 1 were
purchased from commercial sources and used as received.
Compounds 4, 5, and 7 were prepared and characterized as described
in the Supporting Information.
NMR Spectroscopy. Ambient-temperature NMR spectra were
acquired on a Bruker AVANCE III 400 MHz spectrometer (5 mm
BBFO probe). Low-temperature NMR experiments were performed
on a Bruker AVANCE III 500 MHz spectrometer (5 mm BBI probe).
Chemical shifts were calibrated relative to residual protio solvent.
Data were processed using TopSpin or MestreNova.
HPLC. Routine HPLC analysis was performed on Agilent 1100 or
1200 series instruments with diode array detectors. Unless otherwise
noted, analysis was typically done with traces from a single wavelength
with the following method: column, Zorbax SB-C18, 1.8 μm, 3 × 50
mm; column temperature, 60 °C; flow rate, 1.5 mL/min; solvent
gradient, H2O (0.05% TFA v/v)/ACN (0.05% TFA v/v), from 100/0
to 5/95 over 2.7 min; detection wavelength, 220 nm.
Table 1, entries 2, 3, 5−12, 14, and 15, used an Ascentis Express
C18 column (100 mm × 4.6 mm; 2.7 μm): column temperature, 40
°C; flow rate, 1.5 mL/min; solvent gradient, H2O (0.1% H3PO4)/
ACN, from 70/30 to 0/100 over 2.5 min; detection wavelength, 228
nm for detection of 1, 3, and 4 and 210 nm for detection of 7.
GC/MS. GC/MS analysis was performed on an Agilent 7890B GC/
5977B MSD EI system using an HP-5MS column (30 m length, 0.25
mm diameter, 0.25 μm film): temperature ramp to 250 °C at a rate of
15 °C/min; hold time 10 min; run time 26 min.
Operando ATR-FTIR. Reactions were monitored by attenuated
total internal reflection Fourier transform infrared spectroscopy
(ATR-FTIR) using a Mettler Toledo ReactIR iC10 FTIR
spectrometer with an MCT detector, HappGenzel apodization, and
1× gain. The diamond window ATR probe (DiComp, 9 mm
diameter, 457 mm long, Hastelloy C22) was connected to the
spectrometer by a 2 m silver halide optical fiber conduit. Prior to
reaction monitoring, the background spectrum (256 scans) was
obtained with the clean probe in air. Reaction spectra (2500−650
cm−1, 8 cm−1 resolution, 16 scans per spectrum) were collected at 10
s intervals using Mettler Toledo iCIR software v4.3.27.0 run on a
Windows laptop computer. Reaction components were tracked using
peak height profiles determined from reference spectra or assigned on
the basis of analysis of changes in the reaction spectra.
Figure 5. Multigram continuous flow run to produce 3 in 67%
isolated yield using LDA and diagram of the lab-scale flow reactor.
Everything within the box was placed in a dry ice/acetonitrile bath.
NMR spectroscopy after storing an I2/2-MeTHF solution for 1
month.10 Attempts to use more concentrated feeds resulted in
precipitation of solids, which we identified as a diisopropy-
lammonium salt. Presumably this crystalline species is the
product of diisopropylamine reacting with excess iodine prior
to the quench to generate [iPr2NHI]+I−.17
Under the conditions shown in Figure 5, we were able to
achieve an 84% solution yield of 3 over a >3 h run, with no
solid formation or loss of purity over time. The iodination
regioselectivity is 30:1 (determined by NMR spectroscopy),
confirming complete ArLi equilibration within the 4 min
residence time. Purification by crystallization from heptane
gives 3 in 67% isolated yield, corresponding to a throughput
(after purification) of 425 g h−1 L−1. This represents an
improvement of almost 10-fold over the first-generation flow
process described in Scheme 3, bringing the processing time
needed to prepare 10 kg of 3 from >1 month to <1 week using
our pilot-scale system. Work is currently underway to adapt
this new process to pilot-scale production of this key
intermediate for API synthesis.
CONCLUSIONS
■
By using a combination of small-scale screening experiments
and mechanistic understanding gleaned through in situ
spectroscopy, we have developed an efficient and selective
lithiation/iodination sequence for the synthesis of a key
pharmaceutical intermediate. The use of continuous flow
chemistry is critical to the success of the C−H metalation,
SAFETY NOTE. Cryogens can cause burns upon exposure. Wear
proper PPE and avoid contact with cryogens. Continuous flow
equipment should be thoroughly leak-checked prior to operation,
especially when reactive and potentially pyrophoric streams are used.
F
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
General Procedure for ATR-FTIR Experiments. Anhydrous
THF (10 mL) and N,N-diisopropylamine (0.7 mL, 5.0 mmol, 0.95
equiv) were charged to a dried three-neck 100 mL round-bottom flask
with a stir bar and ATR-FTIR probe under N2. The solution was
cooled in a dry ice/acetone (−70 °C) or dry ice/acetonitrile (−45
°C) bath. nBuLi in hexanes (2.5 M, 2 mL, 5.0 mmol, 0.95 equiv) was
added dropwise, followed by a solution of 1 (1.0 g, 5.3 mmol, 1 equiv)
in 5 mL of anhydrous THF dropwise. At the indicated time, a solution
of I2 (2.0 g, 7.9 mmol, 1.5 equiv) was added dropwise.
General Procedure for Small-Scale Optimization Experi-
ments. Table 1, Entries 1 and 4. Anhydrous THF (1 mL) was added
to 150 mg of 1 (0.8 mmol, 1 equiv) in an 8 mL septum-capped vial
with a stir bar under N2. In a separate 8 mL vial under N2 in a dry ice/
acetone or dry ice/acetonitrile cooling bath, an LDA solution was
prepared with anhydrous THF (1 mL) and N,N-diisopropylamine
(0.1 mL, 0.76 mmol, 0.95 equiv), followed by dropwise addition of
2.5 M nBuLi in hexanes (0.3 mL, 0.76 mmol, 0.95 equiv). The
contents of both vials were stirred for 5 min in a dry ice/acetone bath
for temperature equilibration. The LDA solution was added to the
solution of 1 dropwise over 2 min via syringe to minimize the
temperature rise from a reaction exotherm, followed by a solution of
I2 in 1 mL of anhydrous THF (303 mg, 1.19 mmol, 1.5 equiv) in one
portion. The reaction was then quenched with 4 mL of saturated
aqueous sodium sulfite (Na2SO3) solution, and 25 μL of 1,3,5-
trifluorobenzene was added as a 19F NMR internal standard. For
NMR measurement, 100 μL of organic layer was added to 600 μL of
CDCl3.
Table 1, Entry 13. As above, but PhLi (1.8 M in nBu2O, 444 μL,
0.800 mmol) was used instead of LDA.
Table 1, Entries 2, 3, and 5−10. N,N-Diisopropylamine (29.3 μL,
0.206 mmol) was dissolved in THF (90 μL) in a septum-cap-sealed
1.5 mL vial. The vial was cooled to −5 °C prior to the addition of
nBuLi (2.5 M in hexanes, 79 μL, 0.197 mmol) over 5 s. In a separate
vial containing a stirbar, 1 (35.4 mg, 0.187 mmol) was dissolved in
THF (375 μL). The vial was sealed with a septum cap and cooled to
−70 °C (dry ice/acetone) or −45 °C (dry ice/acetonitrile). A third
vial was charged with I2 (71.3 mg, 0.281 mmol) and THF (280 μL).
The entire LDA solution was added via syringe in one portion to a
stirred solution of 1. This solution was stirred for the desired
metalation time, followed by addition of the entire I2 solution via
syringe. The reaction was quenched 5 s later by the addition of 1 M
HCl (1 mL). The entire reaction mixture was diluted to a total
volume of 100 mL with MeOH and assayed by HPLC against
calibrated absorbance values.
(3.1 mL, 1.6 M in hexanes, 5.0 mmol). In an 8 mL vial, 1.0 g of 1 (5.2
mmol) was dissolved in 5 mL of anhydrous THF and flushed with N2.
While IR spectra were acquired, a solution of 1 was added portionwise
via syringe. A 1 mL portion of the 1 solution (1.05 mmol, 0.21 equiv)
was added at 9, 15, 20, and 25 min. A 0.5 mL portion of the 1 solution
(0.52 mmol, 0.11 equiv) was added at 29 min, and 0.25 mL (0.26
mmol, 0.053 equiv) was added at 33 and 41 min reaction time,
totaling 1.06 equiv of 1 added.
General Procedure for LDA Lab-Scale Flow Reaction. All
lines were primed with anhydrous THF (1 and LDA lines) or 2-
MeTHF (I2 line) at 1 mL/min, and the reactor was submerged within
a dry ice/acetonitrile bath (−45.1 °C). The following feed solutions
were prepared in 250 mL volumetric flasks: (1) 15.5 g of 1 (82 mmol,
1 equiv) filled with anhydrous THF and (2) 31.2 g of I2 (123 mmol,
1.5 equiv) and 25 mL of toluene (internal standard) filled with 2-
MeTHF. Each feed solution was added to feed bottles and placed
under a positive pressure of N2. The LDA feed solution was prepared
in the feed bottle with a stir bar: 208 mL of anhydrous THF and 10.9
mL of N,N-diisopropylamine (78 mmol, 0.95 equiv) were added
under N2 and cooled in a dry ice/acetonitrile bath, followed by
n
dropwise addition of 31.1 mL of 2.5 M BuLi in hexanes (78 mmol,
0.95 equiv). The LDA and 1 solutions, followed by the I2 solution
after 4 min, were pumped through the lines, each at a rate of 1 mL/
min. After 10 min equilibration time, 5 min fractions were collected
into 20 mL vials containing 6 mL of stirred saturated aqueous Na2SO3
to quench excess I2. After 50 min, 30 min fractions were collected into
175 mL bottles containing 50 mL of stirred saturated aqueous
Na2SO3. The fractions from 20 min to 3 h 33 min (input of 12.3 g of
1) were combined into a 1 L separatory funnel with 200 mL of EtOAc
and washed with 100 mL of saturated aqueous Na2SO3 combined
with 50 mL of H2O (2×), 0.5 M HCl (100 mL × 3), and H2O (1 ×
1
100 mL). The solvent was removed via rotary evaporation. H NMR
with 1,3,5-trimethoxybenzene added as an internal standard indicated
a purity of 78%, giving an NMR yield of 84% of 3. In a 500 mL round-
bottom flask, the product was added to 160 mL of heptane and heated
in a water bath to 70 °C until all solids dissolved. The flask was slowly
cooled to room temperature and then was cooled in an ice bath. The
resulting solids were collected by filtration and rinsed with 80 mL of
cold heptane to provide 13.0 g of 3 as an off-white solid (67% isolated
yield).
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.8b00538.
Table 1, Entries 11, 12, 14, and 15. As above, but PhLi (1.8 M in
nBu2O, 104 μL, 0.187 mmol) was used instead of LDA.
General Procedure for PhLi Lab-Scale Flow Reaction. Each
line was primed with anhydrous THF at 1 mL/min for 20 min, and
the reactor was submerged within a dry ice/acetone bath. Each feed
solution was prepared as follows: (1) 7.57 g of 1 (40.0 mmol) was
filled with anhydrous THF to a total final volume of 200 mL, (2) 15.2
g of I2 (59.9 mmol) was filled with anhydrous THF to a total final
Flow reactor information, further information on NMR
and ATR-FTIR experiments, characterization data, and
2-MeTHF polymerization data (PDF)
n
volume of 300 mL, and (3) PhLi (1.8 M in Bu2O, Sigma-Aldrich)
AUTHOR INFORMATION
Corresponding Authors
*E-mail for A.L.D.: anna.l.dunn@gsk.com.
*E-mail for D.C.L.: david.c.leitch@gsk.com; dcleitch@uvic.ca.
■
was used as received. The stoichiometry of 1:PhLi:I2 (1:1:1.4) was
achieved by using flow rates of 0.495:0.055:0.693 mL/min. After
reactor equilibration, 150 mL of product fractions was collected,
combined with 50 mL of EtOAc, and washed with saturated aqueous
Na2SO3 solution (50 mL) until complete discoloration was observed.
The layers were separated, and the aqueous layer was washed with 20
mL of EtOAc. The combined organics were evaporated to dryness
and added to 30 mL of heptane. The solution was heated to reflux to
dissolve all solids and then cooled to ambient temperature. Solids
were collected and rinsed with 5 mL of heptane, yielding 2.40 g of off-
white product 3 (yield 64% based on 150 mL flow output theoretical
yield).
ORCID
Anna L. Dunn: 0000-0003-2852-7755
David C. Leitch: 0000-0002-8726-3318
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Titration of 1 into LDA at −70 °C (Figure 3). N,N-
Diisopropylamine (0.71 mL, 5.0 mmol) and 5 mL of anhydrous
THF were charged to a dry 50 mL three-necked round-bottom flask
with stir bar and ATR-FTIR probe under N2. The solution was cooled
The authors thank all of the members of the global Continuous
Primary group, Pilot Plant Operations, and former colleagues
at Research Triangle Park for their assistance during the course
of this work.
n
to −70 °C in a dry ice/acetone bath, and BuLi was added dropwise
G
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Schlosser, M. Relative Basicities of ortho-, meta-, and para-Substituted
Aryllithiums. J. Org. Chem. 2009, 74, 222−229.
(8) (a) Merkushev, E. B. Advances in the Synthesis of Iodoaromatic
Compounds. Synthesis 1988, 1988, 923−937. (b) Rozen, S.; Zamir, D.
A Novel Aromatic Iodination Method Using F2. J. Org. Chem. 1990,
55, 3552−3555.
REFERENCES
■
(1) Larock, R. C.; Zhang, L. Halogenation of Organometallics. In
Comprehensive Organic Transformations: A Guide to Functional Group
Preparations, 3rd ed.; Larock, R. C., Ed.; Wiley: Hoboken, NJ, 2018;
pp 1467−1480.
(2) (a) Gschwend, H. W.; Rodriguez, H. R. Heteroatom-Facilitated
Lithiations. Org. React. 1979, 26, 1−360. (b) Snieckus, V. Directed
Ortho Metalation. Tertiary Amide and O-Carbamate Directors in
Synthetic Strategies for Polysubstituted Aromatics. Chem. Rev. 1990,
90, 879−933. (c) Cottet, F.; Schlosser, M. Three Chloro-
(trifluoromethyl)pyridines as Model Substrates for Regioexhaustive
Functionalization. Eur. J. Org. Chem. 2004, 2004, 3793−3798.
(d) Collum, D. B.; McNeil, A. J.; Ramirez, A. Lithium
Diisopropylamide: Solution Kinetics and Implications for Organic
Synthesis. Angew. Chem., Int. Ed. 2007, 46, 3002−3017. (e) Viciu, M.
S.; Gupta, L.; Collum, D. B. Mechanism of Lithium Diisopropyla-
mide-Mediated Substitution of 2,6-Difluoropyridine. J. Am. Chem. Soc.
2010, 132, 6361−6365. (f) Fukuda, T.; Ohta, T.; Sudo, E.; Iwao, M.
Directed Lithiation of N-Benzenesulfonyl-3-bromopyrrole. Electro-
phile-Controlled Regioselective Functionalization via Dynamic
Equilibrium between C-2 and C-5 Lithio Species. Org. Lett. 2010,
12, 2734−2737. (g) Hoepker, A. C.; Gupta, L.; Ma, Y.; Faggin, M. F.;
Collum, D. B. Regioselective Lithium Diisopropylamide-Mediated
Ortholithiation of 1-Chloro-3-(trifluoromethyl)benzene: Role of
Autocatalysis, Lithium Chloride Catalysis, and Reversibility. J. Am.
Chem. Soc. 2011, 133, 7135−7151. (h) Florio, S.; Salomone, A.
Heterocycle-Mediated ortho-Functionalization of Aromatic Com-
pounds: The DoM Methodology and Synthetic Utility. Synthesis
2016, 48, 1993−2008.
(3) Mohler, M. L.; Bohl, C. E.; Jones, A.; Coss, C. C.; Narayanan, R.;
He, Y.; Jin Hwang, D.; Dalton, J. T.; Miller, D. D. Nonsteroidal
Selective Androgen Receptor Modulators (SARMs): Dissociating the
Anabolic and Androgenic Activities of the Androgen Receptor for
Therapeutic Benefit. J. Med. Chem. 2009, 52, 3597−3617.
(4) Turnbull, P. S.; Cadilla, R. Indolecarbonitriles as Selective
Androgen Receptor Modulators. PCT Int. Appl. WO
2014013309A120140123, 2014.
(5) (a) McConville, F. X. The Pilot Plant Real Book: A Unique
Handbook for the Chemical Process Industry; FXM Engineering and
Design: Worcester, MA, 2002. (b) Stoessel, F. Thermal Safety of
Chemical Processes: Risk Assessment and Process Design; Wiley-VCH:
Weinheim, Germany, 2008. (c) Davis, E. M.; Viswanath, S. K. Heat
Transfer Based Scale-Down of Chemical Reactions. Org. Process Res.
Dev. 2012, 16, 1360−1370.
(9) (a) Kalyani, D.; Dick, A. R.; Anani, W. Q.; Sanford, M. S. A
Simple Catalytic Method for the Regioselective Halogenation of
Arenes. Org. Lett. 2006, 8, 2523−2526. (b) Mei, T.-S.; Giri, R.;
Maugel, N.; Yu, J.-Q. PdII-Catalyzed Monoselective ortho Halogen-
ation of C-H Bonds Assisted by Counter Cations: A Complementary
Method to Directed ortho Lithiation. Angew. Chem., Int. Ed. 2008, 47,
5215−5219. (c) Schroder, N.; Wencel-Delord, J.; Glorius, F. High-
Yielding, Versatile, and Practical [Rh(III)Cp*]-Catalyzed Ortho
Bromination and Iodination of Arenes. J. Am. Chem. Soc. 2012, 134,
8298−8301. (d) Partridge, B. M.; Hartwig, J. F. Sterically Controlled
Iodination of Arenes via Iridium-Catalyzed C-H Borylation. Org. Lett.
2013, 15, 140−143.
(10) See the Supporting Information for more details.
(11) (a) Kupracz, L.; Kirschning, A. Multiple Organolithium
Generation in the Continuous Flow Synthesis of Amitriptyline. Adv.
Synth. Catal. 2013, 355, 3375−3380. (b) Pieber, B.; Glasnov, T.;
Kappe, C. O. Flash Carboxylation: Fast Lithiation − Carboxylation
Sequence at Room Temperature in Continuous Flow. RSC Adv. 2014,
4, 13430−13433. (c) Westermann, T.; Mleczko, L. Heat Management
in Microreactors for Fast Exothermic Organic Syntheses − First
Design Principles. Org. Process Res. Dev. 2016, 20, 487−494.
(12) (a) Newby, J. A.; Blaylock, D. W.; Witt, P. M.; Turner, R. M.;
Heider, P. L.; Harji, B. H.; Browne, D. L.; Ley, S. V. Reconfiguration
of a Continuous Flow Platform for Extended Operation: Application
to a Cryogenic Fluorine-Directed ortho-Lithiation Reaction. Org.
Process Res. Dev. 2014, 18, 1221−1228. (b) Kopach, M. E.; Cole, K.
P.; Pollock, P. M.; Johnson, M. D.; Braden, T. M.; Webster, L. P.;
McClary Groh, J.; McFarland, A. D.; Schafer, J. P.; Adler, J. J.;
Rosemeyer, M. Flow Grignard and Lithiation: Screening Tools and
Development of Continuous Processes for a Benzyl Alcohol Starting
Material. Org. Process Res. Dev. 2016, 20, 1581−1592. (c) Thaisri-
vongs, D. A.; Naber, J. R.; McMullen, J. P. Using Flow to Outpace
Fast Proton Transfer in an Organometallic Reaction for the
Manufacture of Verubecestat (MK-8931). Org. Process Res. Dev.
2016, 20, 1997−2004. (d) Laue, S.; Haverkamp, V.; Mleczko, L.
Experience with Scale-Up of Low-Temperature Organometallic
Reactions in Continuous Flow. Org. Process Res. Dev. 2016, 20,
480−486. (e) Feng, R.; Ramchandani, S.; Ramalingam, B.; Wei
Benjamin Tan, S.; Li, C.; Khean Teoh, S.; Boodhoo, K.; Sharratt, P.
Intensification of Continuous Ortho-Lithiation at Ambient Con-
ditions − Process Understanding and Assessment of Sustainability
Benefits. Org. Process Res. Dev. 2017, 21, 1259−1271. (f) Usutani, H.;
Cork, D. G. Effective Utilization of Flow Chemistry: Use of Unstable
Intermediates, Inhibition of Side Reactions, and Scale-Up for Boronic
Acid Synthesis. Org. Process Res. Dev. 2018, 22, 741−746.
(6) (a) Yoshida, J.; Takahashi, Y.; Nagaki, A. Flash chemistry: flow
chemistry that cannot be done in batch. Chem. Commun. 2013, 49,
9896−9904. (b) Gutmann, B.; Cantillo, D.; Kappe, C. O.
Continuous-Flow TechnologyA Tool for the Safe Manufacturing
of Active Pharmaceutical Ingredients. Angew. Chem., Int. Ed. 2015, 54,
6688−6729. (c) Baumann, M.; Baxendale, I. R. The Synthesis of
Active Pharmaceutical Ingredients (APIs) using Continuous Flow
Chemistry. Beilstein J. Org. Chem. 2015, 11, 1194−1219. (d) Porta, R.;
Benaglia, M.; Puglisi, A. Flow Chemistry: Recent Developments in the
Synthesis of Pharmaceutical Products. Org. Process Res. Dev. 2016, 20,
2−25. (e) Movsisyan, M.; Delbeke, E. I. P.; Berton, J. K. E. T.;
Battilocchio, C.; Ley, S. V.; Stevens, C. V. Taming Hazardous
Chemistry by Continuous Flow Technology. Chem. Soc. Rev. 2016,
45, 4892−4928. (f) May, S. A. Flow Chemistry, Continuous
Processing, and Continuous Manufacturing: A Pharmaceutical
Perspective. J. Flow Chem. 2017, 7, 137−145. (g) Plutschack, M.
B.; Pieber, B.; Gilmore, K.; Seeberger, P. H. The Hitchhiker’s Guide
to Flow Chemistry. Chem. Rev. 2017, 117, 11796−11893.
(13) Cataldo, F. Iodine: A Ring Opening Polymerization Catalyst for
Tetrahydrofuran. Eur. Polym. J. 1996, 32, 1297−1302.
(14) The 1H NMR signals assigned to 2b appear as singlets, whereas
coupling to F should be observed (as it is for 2a). We attribute this
apparent lack of coupling to a rapid monomer/dimer equilibrium for
2b at −70 °C, analogous to that observed for PhLi. For 2a, steric
hindrance likely prevents dimerization, enabling observation of
coupling. For solution dynamics of PhLi, see: Reich, H. J.; Green,
D. P.; Medina, M. A.; Goldenberg, W. S.; Gudmundsson, B. O.;
Dykstra, R. R.; Phillips, N. H. Aggregation and Reactivity of
Phenyllithium Solutions. J. Am. Chem. Soc. 1998, 120, 7201−7210.
(15) (a) Li, X.; Leonori, D.; Sheikh, N. S.; Coldham, I. Synthesis of
1-Substituted Tetrahydroisoquinolines by Lithiation and Electrophilic
Quneching Guided by In Situ IR and NMR Spectroscopy and
Application to the Synthesis of Salsolidine, Carnegine and
Laudanosine. Chem. - Eur. J. 2013, 19, 7724−7730. (b) Keles, H.;
Susanne, F.; Livingstone, H.; Hunter, S.; Wade, C.; Bourdon, R.;
Rutter, A. Development of a Robust and Reusable Microreactor
(7) (a) Schlosser, M. The 2 × 3 Toolbox of Organometallic
Methods for Regiochemically Exhaustive Functionalization. Angew.
Chem., Int. Ed. 2005, 44, 376−393. (b) Hyla-Kryspin, I.; Grimme, S.;
Buker, H. H.; Nibbering, N. M. M.; Cottet, F.; Schlosser, M. The Gas
̈
Phase Acidity of Oligofluorobenzenes and Oligochlorobenzenes:
About the Additivity or Non-Additivity of Substituent Effects.
Chem. - Eur. J. 2005, 11, 1251−1256. (c) Gorecka-Kobylinska, J.;
H
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Employing Laser Based Mid-IR Chemical Imaging for the Automated
Quantification of Reaction Kinetics. Org. Process Res. Dev. 2017, 21,
1761−1768. (c) Li, H.; Sheeran, J. W.; Clausen, A. M.; Fang, Y.-Q.;
Bio, M. M.; Bader, S. Flow Asymmetric Propargylation: Development
of Continuous Processes for the Preparation of a Chiral β-Amino
Alcohol. Angew. Chem., Int. Ed. 2017, 56, 9425−9429. (d) Carter, N.;
Li, X.; Reavey, L.; Meijer, A. J. H. M.; Coldham, I. Synthesis and
Kinetic Resolution of Substituted Tetrahydroquinolines by Lithiation
then Electrophilic Quench. Chem. Sci. 2018, 9, 1352−1357.
(16) Dolman, S. J.; Nyrop, J. L.; Kuethe, J. T. Magnetically Driven
Agitation in a Tube Mixer Affords Clog-Resistant Fast Mixing
Independent of Linear Velocity. J. Org. Chem. 2011, 76, 993−996.
(17) Southwick, P. L.; Christman, D. R. Reactions of Unsaturated
Compounds with Iodine-Amine Complexes. I. Reactions of
Benzalacetophenone and Benzalacetone. J. Am. Chem. Soc. 1952, 74,
1886−1891.
I
DOI: 10.1021/acs.organomet.8b00538
Organometallics XXXX, XXX, XXX−XXX