ReactNMR and reactIR as reaction monitoring and mechanistic elucidation tools: The NCS mediated cascade reaction of α-thioamides to α-thio-β-chloroacrylamides

Article abstract of DOI:10.1021/jo201212p

On-flow ReactIR and ¹H NMR reaction monitoring, coupled with in situ intermediate characterization, was used to aid in the mechanistic elucidation of the N-chlorosuccinimide mediated transformation of an α-thioamide. Multiple intermediates in t

Full text of DOI:10.1021/jo201212p

Article
pubs.acs.org/joc
ReactNMR and ReactIR as Reaction Monitoring and Mechanistic
Elucidation Tools: The NCS Mediated Cascade Reaction of α-
Thioamides to α-Thio-β-chloroacrylamides
David A. Foley,† Christopher W. Doecke,† Jonas Y. Buser,† Jeremy M. Merritt,‡ Linda Murphy,§
Marie Kissane,§ Stuart G. Collins,§ Anita R. Maguire,§,∥ and Andreas Kaerner*,†
‡
†Analytical Sciences Research and Development and Chemical Product Research and Development, Eli Lilly and Company, Lilly
Corporate Center, Indianapolis, Indiana 46285, United States
§Department of Chemistry and ∥School of Pharmacy, Analytical and Biological Chemistry Research Facility, University College Cork,
Cork, Ireland
S
* Supporting Information
1
ABSTRACT: On-flow ReactIR and H NMR reaction monitoring, coupled
with in situ intermediate characterization, was used to aid in the mechanistic
elucidation of the N-chlorosuccinimide mediated transformation of an α-
thioamide. Multiple intermediates in this reaction cascade are identified and
characterized, and in particular, spectroscopic evidence for the intermediacy of
the chlorosulfonium ion in the chlorination of α-thioamides is provided.
Further to this, solvent effects on the outcome of the transformation are discussed. This work also demonstrates the utility of
using a combination of ReactIR and flow NMR reaction monitoring (ReactNMR) for characterizing complex multicomponent
reaction mixtures.
took advantage of specific cases where reactions were run neat
or at increased concentrations in protonated solvents. Another
method of overcoming the issue of large solvent proton
resonances is to monitor the progress of the reaction via an
alternative nucleus.12−14 The use of NMR as a medium for
monitoring biocatalyzed reactions in aqueous systems is well
established in the biotechnology field.15,16 The issues associated
with swamping of analyte resonances by that of water in the
reaction medium is circumvented by the use of presaturation
techniques. The use of NMR in protonated solvents to monitor
reactions has been reported.9 Work by Fyfe et al. in the 1970s
used low-temperature flow NMR studies to investigate the
reactions of amines with 2,4,5-trinitroanisole, conducted in a
DMSO/MeOH solvent mixture.17,18 Recently, flow NMR was
utilized to identify an intermediate in formation of an imidazole
species and hence enhance the mechanistic understanding to
optimize the process.19 2D NMR techniques, including
HSQC,20 HMBC,21 TOCSY,22 and DOSY,23,24 have also
been exploited as reaction monitoring tools. Magic angle
spinning NMR has been utilized to study adsorption and
reaction kinetics in nanocages.25
INTRODUCTION
■
The use of high-resolution NMR spectroscopy enables the
characterization of organic compounds in a nondestructive and
noninvasive way, using relatively small quantities of material.
Under the appropriate experimental conditions, qualitative and
quantitative data can be mined from multicomponent mixtures.
However, this technique traditionally employs the use of
expensive deuterated solvents, which has been prohibitive in
the expansion of this application to reaction monitoring,
particularly on a medium to large scale. With the advent of
NMR solvent suppression techniques such as WET and low
drift magnets the restrictions placed by the necessary use of
deuterated solvents has been greatly reduced.1,2 These
advances, in tandem with the progress made in flow NMR
and ReactIR technologies, provide a unique opportunity in the
field of reaction monitoring and process analytical technology
(PAT), particularly in the development of flow and continuous
chemistry.3,4
Much of the early work in the area of flow NMR was focused
on coupling this technology with LC separation as a means of
identifying and characterizing compounds.5 Its application to
multisample analysis6 and combinatorial chemistry7 has also
been reported.
The advances outlined above paved the way for increased
utility of NMR in the area of reaction monitoring. Varying
experimental setups have been published to date.8−10 Some
have used traditional deuterated solvent media, while others
have attempted to move away from these expensive solvents, by
employing alternative methods. A report on no-D NMR
spectroscopy in reaction monitoring appeared in 2004.11 This
While employing high-resolution NMR as a probe into
reaction mechanisms and monitoring is well established, the
vast majority of the work conducted to date involves small scale
reactions in deuterated solvents, which are carried out in
standard NMR tubes. This has obvious limitations, not only in
terms of scale but also the types of reactions that can be
Received: June 22, 2011
Published: October 26, 2011
© 2011 American Chemical Society
9630
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
Scheme 1. Mechanism Proposed by Maguire et al.
investigated. Reactions normally run under high temperature
and pressure conditions, as well as heterogeneous systems,
present challenges to traditional NMR monitoring in tubes.
Another issue associated with reactions conducted in conven-
tional NMR tubes is the delay in acquiring the initial spectrum
in a series due to the necessity of locking, shimming, etc., which
can be restrictive in the case of rapid onset of reaction.
Although physical solutions to this problem have been
devised,26−28 some drawbacks still remain. A number of flow
NMR reaction monitoring instrumental setups have been
reported in recent years, which couple the reaction vessel to the
spectrometer via transfer lines.9,29−31
Here we present the use of a combination of traditional tube
based NMR, online flow NMR, and IR reaction monitoring, to
aid the mechanistic elucidation of a multicomponent reaction
cascade, identifying and monitoring reactive intermediates
involved in the conversion of α-thioamides to α-thio-β-
chloroacrylamides.
isolated from the reaction mixture by tailoring the reaction
conditions to enhance the production of a particular compound
in the cascade, which was then characterized offline. Some of
the transient species (e.g., 3) in the mechanistic sequence could
not be isolated or observed spectroscopically in the original
study. This reaction provided a potential multicomponent
process with which to demonstrate the unique capabilities that
reactNMR and reactIR provide, both in terms of reaction
monitoring and structural elucidation of reactive intermediates.
NMR spectroscopy is of particular use in monitoring the
progress of reactions (in particular this reaction cascade), as not
only does this method provide inherent quantitation, but it also
provides detailed structural identification of the intermediates
involved. A number of distinct resonances displayed by each
component in solution also allows for excellent selectivity,
especially when the reaction mixture contains a complex
mixture of starting material, intermediates, and product. These
features mean that reactNMR has a number of advantages over
other reaction monitoring techniques, including IR, HPLC, and
UV, which provide a dearth of structural information and
require calibration in order to achieve truly quantitative data.
However, the authors do acknowledge that the lifetime of
certain reactive intermediates means that their detection cannot
be observed on the NMR time-scale. Also a lack of sensitivity is
associated with the use of NMR, but monitoring of flowing
solutions means that increased concentration of species is
possible as a method of addressing this issue.
RESULTS AND DISCUSSION
■
A report by Maguire et al. on the generation of α-phenylthio-β-
chloroacrylamides via the chlorination of α-thioamide pre-
cursors32 was followed by a subsequent publication that
included a proposed mechanism for the formation of the β-
chloroacrylamide product 1 from thioamide 2, involving
multiple intermediates shown in Scheme 1.33 Some of these
intermediates (5, as well as the labile compounds 4 and 6) were
9631
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
1
Scheme 2. H NMR Resonances Used for ReactNMR Reaction Monitoring in CCl4 and MeCN
Reaction monitoring using this technique does not require
the use of isotope labeled compounds or deuterated solvents,
which has cost and time benefits for the synthetic organic
chemist, allowing increased mechanistic information to be
gained using regular unlabeled materials. Thus the need to
revisit the synthesis with expensive labeled material to obtain
this information is obviated. This also has application in the
area of process development, where advanced synthetic
intermediates involving multiple steps would need to be
synthesized with isotopic labels, delaying progress; again,
reactNMR eliminates this inconvenience. With the reactor
and flow NMR experimental setup outlined here, reactions may
be monitored on kilogram scale, without prohibitive reagent
costs.
overcome this constraint a different solvent system was used in
subsequent investigations.
In the publication on this transformation by Maguire et al.,33
carbon tetrachloride was used as an alternative solvent to
toluene. This has a distinct advantage from an NMR
standpoint, as the use of a 1H NMR transparent solvent
would greatly enhance the clarity of the spectra under
investigation, allowing for any subtle changes in the
composition of the reaction mixture to be observed, and
therefore it was decided to conduct further investigations in this
solvent. In the current work, initial characterization of the
starting α-thioamide 2 in carbon tetrachloride was carried out
in a 5-mm NMR tube, with one drop of 1,2-dichloroethane
(DCE) added as a reference signal, which was set at δH 3.68
ppm.
Preliminary work was conducted in toluene-d8 using 1.95
equiv of N-chlorosuccinamide (NCS), as these were the
optimum conditions described by Maguire et al.33 Prior to
performing this reaction in flow, an initial investigation in an
NMR tube enabled characterization of the starting materials in
toluene-d8, as well as the viability of observing components of
the reaction mixture. A number of intermediates were
identified, and distinct characteristic resonances were assigned
to these compounds; however, many of these overlapped with
those of the reaction solvent, particularly in the aromatic region
of the spectrum. With the ultimate goal of conducting this
reaction in nondeuterated solvents on a larger scale, this
overlap proved to be too restrictive. While multiple resonance
WET solvent suppression could be achieved, this would have
resulted in decreased signal intensity in these regions, giving
nonquantitative data. Therefore, the choice of reaction solvent
has implications for which areas of the spectrum become
unusable as analytical handles for reaction monitoring, and to
In an attempt to observe all of the components of the
reaction cascade, the reaction was conducted at 25 °C in carbon
tetrachloride. The progress of the reaction was monitored by
1H NMR at 10 s intervals, with an acquisition of 4 scans. Once
the species were observed in the reaction mixture, each entity
was characterized by conducting a number of homo- and
heteronuclear 2D NMR experiments (1H COSY, 1H−13C
1
HSQC, H−13C HMBC) offline in a 5-mm NMR tube. The
component of interest was formed to its maximum level, at
which time the 2D experiments were recorded. The character-
ization of each observable compound is reported in the
Experimental Section, and the individual characteristic
resonances used for reaction monitoring in carbon tetrachloride
and subsequently in acetonitrile are shown in Scheme 2.
1
A stacked plot of the H NMR spectra recorded as the
reaction proceeded is shown in Figure 1. This shows that the α-
thioamide starting material 2 began to be consumed within 16
min of NCS addition to the amide solution; this was tracked by
9632
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
As the α-thioamide 2 was consumed, appearance of a new set
of signals was observed. Two separate doublets at δH 1.28 and
1.67 ppm emerged simultaneously as the reaction progressed.
These have been assigned as the diastereomeric methyl signals,
resulting from the formation of a new stereogenic center in the
chlorosulfonium ion 3, by chlorination of sulfur. The
corresponding methine protons, alpha to the chlorosulfonium
moiety, were observed at δH 3.33 and 3.67 ppm, while two
diastereomeric NH resonances were also present at δH 8.70
and 8.92 ppm in the proton NMR spectrum (Figure 3). During
Figure 1. 1H NMR reaction profile in carbon tetrachloride at 25 °C. α-
Thioamide 2, chlorosulfonium ion 3, α-chlorosulfide 4, Dichloride 6,
dichlorosulfonium ion 7.
monitoring the doublet for the methyl group of 2 at δH 1.58
ppm. A possible explanation for this delay will be outlined later.
As can be seen in Figure 2, the solubility of NCS in CCl4
reaches a maximum at 15 min and during this time no reaction
is observed. The consumption of α-thioamide 2 occurs rapidly
once the NCS level in solution decreases and the reaction
cascade begins. Following this the level of NCS in solution
increases slowly due to a slow down in its consumption. The
signals pertaining to the methylene signal of NCS were
observed as a singlet at δH 2.80 ppm, but due to the low
solubility of NCS in CCl4 at room temperature this signal
represented the amount of NCS in solution. The succinimide
byproduct of the reaction, also of low solubility, appeared at δH
2.63 ppm as the reaction proceeded.
1
Figure 3. H NMR spectrum of chlorosulfonium ion intermediate 3.
characterization of this intermediate, a 1H−13C HSQC
experiment revealed the corresponding carbon-13 signals for
the methine carbons at δC 61.6 ppm, which correlated to a
proton at δH 3.67 ppm. A diastereomeric resonance at δC 62.8
ppm, with the corresponding proton at δH 3.33 ppm was also
observed. These values compare favorably with those reported
Figure 2. Reaction profile in carbon tetrachloride at 25 °C. Experimental data are displayed as symbols and the corresponding computationally
optimized data are displayed as solid lines.
9633
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
Scheme 3. Revised Mechanism
by Liu and Vederas, who report a downfield shift in
chlorosulfonium ion methylene carbons adjacent to sulfur to
δC ∼66 ppm.34 The methyl signals at δH 1.28 and 1.67 ppm
corresponded to carbon signals at δC 9.3 and 14.3 ppm,
respectively. Chlorosulfonium ion species have been reported
previously in the literature as intermediates in the NCS
mediated chlorination of sulfides35 and was hence proposed as
an intermediate in the cascade by Maguire. This current work
provides spectroscopic evidence for the intermediacy of this
species in the reaction pathway. This intermediate was detected
the presence of 3 coincides with the appearance of the methyl
resonance for 4 at 2.11 ppm. A decrease in the level of this
methyl signal related to 4 is observed in the 1H NMR spectrum
once the chlorosulfonium ion is consumed, demonstrating that
this ion is a precursor to α-chlorosulfide 4 formation, and 4
undergoes subsequent transformations. This is demonstrated
graphically in Figures 1 and 2. Because 4 proved to be labile on
isolation,33 its characterization was incomplete, and therefore in
situ “stopped flow” 2D NMR experiments, as well as offline
1
NMR tube spectra, aided in the complete H and 13C NMR
1
by H NMR only when the reaction was conducted in carbon
structural assignment.
tetrachloride, while experiments using toluene or acetonitrile as
the solvent showed no detectable evidence for this transient
species. This will be discussed below.
At steady state, the chlorosulfonium intermediate 3 is present
only at low levels in the reaction mixture (∼22 mol %); its
rapid conversion through to the α-chlorosulfide 4 is evident, as
The next intermediate in the cascade, the acrylamide 5, had
been isolated and structurally characterized in the original
study,33 employing just 1 equiv of NCS. In the current in situ
study, elimination of HCl from α-chlorosulfide 4 provided
acrylamide 5, although interestingly, this could only be detected
at very low levels in the 1H NMR spectrum. From the
9634
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
Chart 1
experiments conducted in this work, it was observed that the
chlorination steps occur rapidly, while the rate of elimination to
form vinylic moieties is much slower. This explains why the
acrylamide 5 does not build to significant levels in the reaction
mixture, as it is quickly transformed via the second NCS
mediated chlorination to the dichloride 6. The acrylamide
intermediate 5, which is formed preferentially when 1 equiv of
NCS is employed, could also be detected and characterized
with 2D NMR experiments.
complete; as conversion to the acrylamide 5 is a rate-
determining step, therefore the second equivalent of NCS
can react with the chlorosulfonium ion 3. Upon isolation of the
reaction mixture, labile intermediate 7 was not observed.
The experiment conducted in carbon tetrachloride provided
definitive evidence for the early part of the reaction cascade,
from sulfide 2 to chlorosulfide 4, to acrylamide 5, to dichloride
6, with the chlorination steps proceeding more rapidly than the
elimination steps. Elevated temperatures (40 °C) in carbon
tetrachloride were used in attempts to form and monitor the
final compound in the cascade, the β-chloroacrylamide 1.
However, this increase in temperature resulted in significant
broadening of signals in the recorded proton NMR spectra,
therefore data at this temperature could not be used.
The data generated from the experiment conducted in
carbon tetrachloride was subjected to reaction modeling
analysis, which showed that while the chlorosulfonium ion 3
is converted to the α-chlorosulfide 4, a competing pathway was
the formation of the dichlorosulfonium ion 7. The delayed
onset of 7 suggests that an experimentally unobserved
intermediate 8 is involved in its formation, and this is
represented in the model.
When carbon tetrachloride was employed as the reaction
solvent, a byproduct formed as the reaction progressed. This
species, tentatively identified as dichlorosulfonium ion 7
(Scheme 3), which coincided with the total consumption of
starting amide, appeared to be diastereomeric in nature. As was
described earlier for the chlorosulfonium ion 3, a doubling of
1
1
isolated resonances was noted in the H NMR and H−13C
HMBC spectra. Two methyl singlets at δH 1.93 and 1.94 ppm
displayed two bond 1H−13C coupling to separate tetra-
substituted carbons at δC 82.8 and 84.4 ppm. Long-range
coupling to the carbonyl carbon at δC 161.9 ppm was also
observed, although it could not be determined definitively if
this was composed of two unique correlations due to resolution
limitations in the HMBC experiment.
A computationally optimized model was fit to the
experimental data (Figure 2) to calculate rate constants for
each of the steps involved in the reaction cascade detailed in
Chart 1.
On examination of the experimental reaction profile (Figure
2), it appears that this dichlorosulfonium species 7 is derived
from 3. Signals for the chlorosulfonium ion 3 and
dichlorosulfonium ion 7 were detected only in carbon
tetrachloride and were not seen in toluene-d8 or acetonitrile.
This implies that the lifetime of the chlorosulfonium ion 3 is
much shorter in either toluene or acetonitrile than in carbon
tetrachloride. As the characteristic resonance of the dichlor-
osulfonium ion 7 would be suppressed in experiments
conducted in protonated acetonitrile (both at δH 1.94 ppm),
a control experiment was carried out in acetonitrile-d3, where
no evidence for the chlorosulfonium 3 or dichlorosulfonium
ion 7 was observed. Thus, the relative rates of the elimination
process from the chlorosulfonium ion 3 and the second
chlorination to the dichlorosulfonium ion 7 are solvent-
dependent, presumably due to differences in polarity. Once
formed, the dichlorosulfonium ion 7 is not transformed further
through the cascade and therefore results in a reduction in
overall yield. In the publication by Maguire et al. the
transformation was conducted at various temperatures.33
There, it was reported that at 20 °C reaction progress was
sluggish, producing only mixtures of the acrylamide 5 and
dichloride 6 after 48 h. This indicates that in carbon
tetrachloride at lower temperatures a second chlorination of
the chlorosulfonium ion 3 is a significant competing pathway. A
combination of factors support this hypothesis: byproduct
formation begins once starting material consumption is
A second experiment was conducted in an attempt to
observe conversion of the dichloride 6 to the β-chloroacryla-
mide product 1. In order to achieve this final transformation, an
elevated temperature (>40 °C) was required to promote HCl
elimination at a rate sufficient for effective monitoring. The
progress of the reaction was monitored by 1H NMR (Figure 4)
and now included simultaneous in-line IR monitoring to
provide a mobile analytical tool that can easily be moved from
lab to lab for subsequent reaction monitoring. The IR bands
were assigned to the individual reaction components using a
combination of the ConcIRT software and the trend plot
assignments obtained from the NMR. A typical IR trend plot is
shown in Figure 5, and while the IR was not capable of
detecting all of the components in the reaction mixture, it
provided sufficient data to monitor the reaction progression. A
number of stopped flow solvent suppressed 2D NMR
experiments were also recorded throughout the experiment.
In an attempt to circumvent the issues outlined previously with
loss of line shape in carbon tetrachloride, the process was
repeated in acetonitrile. A fully protonated solvent was used on
this occasion, therefore WET solvent suppression was applied
to each 1D and 2D spectrum acquired. Figure 6 illustrates the
1
trend plots obtained from the on-flow H NMR spectra of the
9635
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
to the dichloride 6. The acrylamide 5 was identified by
characteristic terminal alkene resonances at δH 5.77 and 6.37
ppm. It was also noted that the level of acrylamide increased
toward the end of reaction monitoring, as the remaining NCS is
consumed, thereby reducing the rate of further chlorination.
As was outlined earlier, no evidence for the formation of the
dichlorosulfonium ion 7 was found when acetonitrile was used
as the reaction solvent.
A delay in the reaction onset was observed in the
transformations conducted in CCl4 and acetonitrile, which
suggests that some form of activation of the NCS is involved,
possibly activated by HCl that is produced as the reaction
proceeds. This hypothesis was supported by a slow pH drop
once NCS was added, followed by a rapid pH drop once
consumption of the amide starting material commenced. Since
HCl is liberated as part of the reaction mechanism, this pH
drop was expected. In another experiment conducted to
confirm that HCl decreased the induction period, it was noted
that addition of 1 drop of conc HCl to the reaction mixture
immediately triggered the reaction cascade. In separate
experiments conducted in both CCl4 and acetonitrile, a reverse
addition (stirring NCS in the reaction solvent for 25 min prior
to addition of the amide 2) also resulted in a faster reaction
initiation (see Supporting Information).
Formation of the dichloride 6 was observed at 25 °C and
reached maximum concentration after 5 h. The integral for the
AB quartet signal for the diastereotopic methylene protons at
δH 4.07 and 4.41 ppm was utilized to monitor its concentration
in the reaction medium over time. The flow was stopped
temporarily to record a 1H−13C HSQC of the reaction mixture
(Figure 7), providing in situ proton−carbon correlations for the
dichloride 6, which could not be isolated for complete
characterization in previous studies.33 A second HSQC
provided proton−carbon correlations for the acrylamide 5,
which formed at a minor level in the reaction mixture. The flow
was restarted once the 2D acquisitions were complete, and
reaction monitoring was continued.
Figure 4. 1H NMR reaction profile in acetonitrile. The solvent
suppressed region of the spectra has been removed for clarity. α-
Thioamide 2, α-chlorosulfide 4, acrylamide 5, dichloride 6, β-
chloroacrylamide 1.
reaction. Acquisition of the 2D data during the reaction
1
resulted in gaps in the H NMR derived spectral trend plots.
As was used in the experiment conducted in carbon
tetrachloride, 1.95 equiv of NCS was added to an acetonitrile
solution of the α-thioamide 2 at 25 °C, and again a delayed
reaction onset was observed. The reaction mixture was stirred
at this temperature for 5 h, before being heated to 45 °C for the
latter stages of the process. While the cascade proceeded in a
similar manner as outlined in carbon tetrachloride, a number of
differences were identified. As the starting amide 2 was
consumed, the α-chlorosulfide 4 was generated, although the
chlorosulfonium ion precursor 3 that was observed previously
in carbon tetrachloride was not detected in acetonitrile. The
doublet representing the methyl signal of 2 at δH 1.48 ppm was
utilized for reaction monitoring purposes. The methyl
resonance for the α-chlorosulfide 4 could not be integrated,
as it was affected by the solvent suppression applied to the
acetonitrile singlet at δH 1.94 ppm. Instead, the aromatic
proton resonance at δH 7.61 ppm was tracked to monitor its
progression. The chlorosulfide formed to a maximum of
approximately 50 mol %, limited by its conversion to the
acrylamide 5. The elimination of HCl from α-chlorosulfide 4 to
form acrylamide 5 was detected at low levels for an initial
period, before further chlorination resulted in rapid conversion
In order to push the reaction cascade to completion, the
temperature of the system was increased to 45 °C. The
aromatic resonance at δH 7.19 ppm was used to monitor the
formation of the final product in the cascade, 1. The
temperature rise increased the rate of HCl elimination from
the dichloride 6 furnishing the β-chloroacrylamide 1, which was
complete after 12 h at elevated temperature, mirroring the
Figure 5. Plot of IR data showing reaction components over time (hours).
9636
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
Figure 6. Plot of component concentration versus time for the reaction of NCS with thioamide 2 in acetonitrile. Experimental data are displayed as
symbols, and the corresponding computationally optimized data are displayed as solid lines. Gap in data between 650 and 950 min due to 2D NMR
acquisitions.
observation from the original synthetic work. Again, a
computationally optimized model was fit to the experimental
(Figure 6) to calculate rate constants for each of the steps
involved in the reaction cascade detailed in Chart 2.
It should be noted that the lack of agreement for the
experimental and calculated trends for the β-chloroacrylamide
product 1 in the later time points is due to tuning changes in
the NMR.
years of research. While this intermediate is widely accepted in
chlorination of sulfides, there have been very few reports of
spectroscopic characterization of these intermediates. Further-
more this NMR study provided evidence for the first time of
the formation of the dichlorosulfonium ion as a “cul-de-sac”
side reaction in the overall reaction cascade, having clear
implications in terms of yield reduction.
The use of data rich detection was critical in making
structural assignments, and the quantitative nature of NMR
spectroscopy provided accurate concentrations used to
determine the reaction rates. The quantitative and qualitative
results were combined from the reaction in acetonitrile and
carbon tetrachloride, giving a full picture of the mechanism
described in Scheme 3.
EXPERIMENTAL SECTION
■
The apparatus used (Figure 8) was a multiport 100-mL HEL AutoLAB
jacketed reaction vessel, which was fitted with an overhead stirrer,
temperature probe, nitrogen flow line, and inlet and outlet ports for
the sample transfer lines. The transfer lines from the reactor to the
spectrometers consisted of 1/16 in. o.d. polytetrafluoroethylene
(PTFE) tubing with 0.2 mm i.d. These were contained within
thermostatically controlled temperature regulation lines, containing
Syltherm as the heat transfer fluid. A hastelloy frit was attached to the
base of the outlet transfer line to filter any solids from the reaction
mixture, preventing them from reaching the flow cell. The reaction
mixture was pumped through the transfer line by a Knauer pump
(Preparative Pump 1800) to the flow cell and returned to the reactor
vessel, thus maintaining the composition of the reactor. The flow rate
was 1.5 mL min−1 with a total volume of less than 3 mL removed from
the reactor at any time. It was determined that this flow rate was
CONCLUSION
■
This study clearly illustrates the power of monitoring complex
multistep reaction sequences via ReactNMR and ReactIR,
enabling direct observation of each of the intermediates
including those that are too labile to isolate. Mechanistic
insight is gained through the reaction profile, which shows the
relative rates at which each of the intermediates is formed and
consumed. Specifically in relation to this reaction cascade,
direct determination of the spectroscopic features of the
chlorosulfonium ion was enabled for the first time in over 15
1
appropriate for this flow probe configuration to ensure H integral
intensities were not biased by inadequate polarization. Parameters such
as temperature, pressure, stir rate, and rate of addition of reagents may
be controlled through the HEL software.
9637
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
1
Figure 7. Stopped flow H−13C HSQC spectra of dichloride 6 and acrylamide 5 as a minor component in the reaction mixture.
Chart 2
Proton NMR and stopped flow WET suppressed standard 2D
NMR experiments on the reaction mixture were acquired using a
Bruker 500 MHz Avance III spectrometer, with a custom flow BBFO
probe containing a 240-μL flow cell. Tube reactions and material
characterization were carried out in 5 mm NMR tubes, using a Bruker
600 MHz Avance III equipped with cryoprobe. NMR spectra were
recorded unlocked, and acetonitrile was injected into the flow cell to
obtain a shim set prior to beginning the experiment. NMR spectra
recorded in carbon tetrachloride were spiked with 1−2 drops of
dichloroethane to provide a chemical shift reference, which was set at
δH 3.68 ppm and δC 43.2 ppm. NMR spectra recorded in acetonitrile-
d3 were referenced at δH 1.94 ppm and δC 118.7 ppm. Topspin 2.1
was used to acquire and process the NMR data. Shimming was
conducted using Topshim. IR spectra were obtained using Metler
Toledo ReactIR 4000 and processed using iC IR, with ConcIRT
enabled to assist in monitoring the change in concentration of
components (version 4.4.26). A custom hastelloy and PEEK flowcell
was attached to the ReactIR lightpipe to provide IR spectra on the
flowing reaction stream.
9638
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
Figure 8. Schematic representation of ReactNMR configuration.
NMR data was processed and analyzed using Mnova NMR
(Mestrelab Research) data analysis package, and the data produced
was then imported into DynoChem (version 3.3.0.0) for computa-
tional analysis and Microsoft Excel to produce graphical output.
Acetonitrile (99.9%, Acros) and carbon tetrachloride (99.5%,
Aldrich) were used as the reaction solvents. The sample of N-4′-
methylphenyl-2-(phenylthio)propanamide was supplied by the
Maguire group.
In Flow Reaction NMR and IR Monitoring of the NCS
Chlorination of N-4′-Methylphenyl-2-(phenylthio)-
propanamide 2. Carbon Tetrachloride as Reaction Solvent. N-
4′-Methylphenyl-2-(phenylthio)propanamide 2 (0.50 g, 1.85 mmol)
was dissolved in carbon tetrachloride (70 mL) and was stirred at room
temperature under an atmosphere of nitrogen. The solution was
pumped to the NMR spectrometer at a flow rate of 1.5 mL/min, and a
1H NMR spectrum was recorded of the α-thioamide 2. NCS (0.49 g,
Chlorosulfonium Ion 3. Isomer A. 1H NMR (600 MHz, CCl4,
δ) 1.66 [d, J = 7.4 Hz, 3H, C(3)H3], 3.33 [q, J = 7.4 Hz, 1H, C(2)H];
13C NMR (151 MHz, CCl4, δ) 14.3 (C-3), 62.6 (C-2), 164.7 (CO).
Isomer B. 1H NMR (600 MHz, CCl4, δ) 1.28 [d, J = 7.2 Hz, 3H,
C(3)H3], 3.67 [q, J = 7.2 Hz, 1H, C(2)H]; 13C NMR (151 MHz,
CCl4, δ) 9.3 (C-3), 61.5 (C-2), 163.1 (CO).
N-4′-Methylphenyl-2-chloro-2-(phenylthio)propanamide
4. 1H NMR (600 MHz, CCl4, δ) 2.11 [s, 3H, C(3)H3], 2.31 [s, 3H,
C(4′)CH3], 7.01 [d, J = 8.1 Hz, 2H, 2 × C(3′)H], 7.11 [d, J = 8.3 Hz,
2H, 2 × C(2′)H], 7.28 (m, 2H, 2 × ArH), 7.37 (t, 1H, J = 7.3 Hz,
ArH), 7.56 (d, J = 8.1 Hz, 2 × ArH), 7.59 (br s, 1H, NH); 13C NMR
(151 MHz, CCl4, δ) 21.6 [C(4′)CH3], 29.9 (C-3), 79.9 (C-2), 120.0
(2 × C-2′), 129.7 (2 × aromatic CH), 129.8 (aromatic CH), 130.0 (2
× C-3′), 137.6 (2 × aromatic CH), 133.5 (quaternary aromatic C),
134.1 (C-4′), 135.9 (C-1′), 164.5 (CO). Note: Some overlap of
signals with those of the α-thioamide 2 was observed in the NMR
spectra of the reaction mixture characterized.
3.60 mmol) was then added to the solution in one portion. The
1
N-4′-Methylphenyl-2-(phenylthio)propenamide 5. 1H NMR
(600 MHz, CCl4, δ) 2.28 [s, 3H, C(4′)CH3], 6.08, 6.86 [2 × s, 2 × 1H,
C(3)H3]; 13C NMR (151 MHz, CCl4, δ) 21.4 [C(4′)CH3], 133.9 (C-
3), 135.9 (C-2), 160.3 (CO). Note: Aromatic resonances
overlapping with those of other compounds in the NMR spectra of
the reaction mixture characterized.
progress of the reaction was monitored by H NMR at 10 s intervals
with an acquisition of 4 scans. The reaction mixture was stirred at 25
°C for 1.5 h.
Acetonitrile as Reaction Solvent. N-4′-Methylphenyl-2-
(phenylthio)propanamide 2 (0.71 g, 2.62 mmol) was dissolved in
acetonitrile (88 mL) and was stirred at room temperature under an
atmosphere of nitrogen. The solution was pumped to the NMR
N-4′-Methylphenyl-2,3-dichloro-2-(phenylthio)-
propanamide 6. 1H NMR (600 MHz, CCl4, δ) 2.32 [s, 3H,
C(4′)CH3], 3.84−4.44 [ABq, J = 11.6 Hz, 2H, C(3)H2], 7.04 [d, J =
8.3 Hz, 2H, 2 × C(3′)H], 7.18 [d, J = 8.4 Hz, 2H, 2 × C(2′)H], 7.31
(t, J = 7.6 Hz, 2H, ArH), 7.39 (t, J = 7.4 Hz, 1H, ArH), 7.58 (d, J = 7.4
Hz, 2H, ArH), 7.88 (br s, 1H, NH); 13C NMR (151 MHz, CCl4, δ)
21.6 [C(4′)CH3], 50.4 (C-3), 82.1 (C-2), 120.4 (2 × C-2′), 128.4
(quaternary aromatic C), 129.6 (2 × aromatic CH), 129.8 (2 × C-3′),
131.1 (aromatic CH), 134.6 (C-4′), 134.8 (C-1′), 137.9 (2 × aromatic
CH), 162.2 (CO).
1
spectrometer at a flow rate of 1.5 mL/min and a H NMR spectrum
was recorded of the α-thioamide 2. NCS (0.69 g, 5.10 mmol) was then
added to the solution in one portion. The progress of the reaction was
monitored by 1H NMR at 30 s intervals with an acquisition of 8 scans.
The reaction mixture was stirred at 25 °C for 5 h. The temperature of
the system was then increased to 45 °C, and the progress of the
reaction was again monitored by 1H NMR at 10 min intervals, with an
acquisition of 8 scans.
N-4′-Methylphenyl-2-(phenylthio)propanamide 2. 1H NMR
(600 MHz, CCl4, δ) 1.58 [d, J = 7.3 Hz, 3H, C(3)H3], 2.29 [s, 3H,
C(4′)CH3], 3.76 [q, J = 7.2 Hz, 1H, C(2)H], 7.01 [d, J = 7.6 Hz, 2H,
2 × C(3′)H], 7.11−7.38 (m, 7H, ArH), 8.28 (br s, 1H, NH); 13C
NMR (151 MHz, CCl4, δ) 18.7 (C-3), 21.6 [C(4′)CH3], 48.0 (C-2),
119.9 (2 × C-2′), 127.8 (aromatic CH), 129.8 (2 × C-3′), 129.9, 130.7
(4 × aromatic CH), 133.5 (C-4′), 134.3 (quaternary aromatic C),
136.0 (C-1′), 168.5 (CO).
Dichlorosulfonium Ion 7. Isomer A. 1H NMR (600 MHz,
CCl4, δ) 1.93 [s, 3H, C(3)H3], 13C NMR (151 MHz, CCl4, δ) 19.5
(C-3), 82.8 (C-2), 161.9 (CO).
Isomer B. 1H NMR (600 MHz, CCl4, δ) 1.94 [s, 3H, C(3)H3], 13
C
NMR (151 MHz, CCl4, δ) 19.5 (C-3), 84.4 (C-2), 161.9 (CO).
Other resonances obscured by those of other components in the
reaction mixture.
9639
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640

The Journal of Organic Chemistry
Article
N-4′-Methylphenyl-Z-3-chloro-2-(phenylthio)propenamide
1. 1H NMR (500 MHz, CH3CN, δ) 2.25 [s, 3H, C(4′)CH3], 7.02−
7.41 (m, 9H, ArH), 7.61 [s, 1H, C(3)H].
(24) Nilsson, M.; Khajeh, M.; Botana, A.; Bernstein, M. A.; Morris,
G. A. Chem. Commun. 2009, 1252−1254.
(25) Xu, S.; Zhang, W.; Liu, X.; Han, X.; Bao, X. J. Am. Chem. Soc.
2009, 131, 13722−13727.
ASSOCIATED CONTENT
(26) McGarrity, J. F.; Prodolliet, J.; Smyth, T. Org. Magn. Reson.
1981, 17, 59−65.
■
S
* Supporting Information
(27) Mix, A.; Jutzi, P.; Rummel, B.; Hagedorn, K. Organometallics
2010, 29, 442−447.
1H−13C long-range correlation data of individual components
available. This material is available free of charge via the
Internet at http://pubs.acs.org.
(28) Christianson, M. D.; Tan, E. H. P.; Landis, C. R. J. Am. Chem.
Soc. 2010, 132, 11461−11463.
(29) Maiwald, M.; Fischer, H. H.; Hasse, H. Chem. Ing. Tech. 2004,
76, 965−969.
AUTHOR INFORMATION
Corresponding Author
*E-mail: andreask@lilly.com.
■
(30) Maiwald, M.; Fischer, H. H.; Kim, Y. K.; Albert, K.; Hasse, H. J.
Magn. Reson. 2004, 166, 135−146.
(31) Bernstein, M. A.; Stefinovic, M.; Sleigh, C. J. Magn. Reson. Chem.
2007, 45, 564−571.
ACKNOWLEDGMENTS
■
(32) Maguire, A. R.; Murphy, M. E.; Schaeffer, M.; Ferguson, G.
Tetrahedron Lett. 1995, 36, 467−470.
The authors would like to acknowledge Mestrelab Research for
the development of reaction monitoring software, Bruker for
developing the custom flow BBFO probe, and IRCSET for
funding to M.K. and L.M.
(33) Murphy, M.; Lynch, D.; Schaeffer, M.; Kissane, M.; Chopra, J.;
O’Brien, E.; Ford, A.; Ferguson, G.; Maguire, A. R. Org. Biomol. Chem.
2007, 5, 1228−1241.
(34) Liu, Y.; Vederas, J. C. J. Org. Chem. 1996, 61, 7856−7859.
(35) Tuleen, D. L.; Stephens, T. J. Org. Chem. 1969, 34, 31−35.
REFERENCES
■
(1) Smallcombe, S. H.; Patt, S. L.; Keifer, P. A. J. Magn. Reson., Ser. A
1995, 117, 295−303.
(2) Zheng, G.; Price, W. S. Prog. Nucl. Magn. Reson. Spectrosc. 2010,
56, 267−288.
(3) Bernstein, M. A. Spectrosc. Eur. 2007, 19, 11−13.
(4) Carter, C. F.; Lange, H.; Ley, S. V.; Baxendale, I. R.; Wittkamp,
B.; Goode, J. G.; Gaunt, N. L. Org. Process Res. Dev. 2010, 14, 393−
404.
(5) Keifer, P. A. Annu. Rep. NMR Spectrosc. 2007, 62, 1−47.
(6) Leo, G. C.; Krikava, A.; Caldwell, G. W. Anal. Chem. 2003, 75,
1954−1957.
(7) Keifer, P. A. Curr. Opin. Chem. Biol. 2003, 7, 388−394.
(8) Fyfe, C. A.; Cocivera, M.; Damji, S. W. H. Acc. Chem. Res. 1978,
11, 277−282.
(9) Maiwald, M.; Fischer, H. H.; Kim, Y. K.; Hasse, H. Anal. Bioanal.
Chem. 2003, 375, 1111−1115.
(10) Gomez, M. V.; Verputten, H. H. J.; Diaz-Ortiz, A.; Moreno, A.;
de la Hoz, A.; Velders, A. H. Chem. Commun. 2010, 46, 4514−4516.
(11) Hoye, T. R.; Eklov, B. M.; Ryba, T. D.; Voloshin, M.; Yao, L. J.
Org. Lett. 2004, 6, 953−956.
(12) Chan, S. J.; Howe, A. G.; Hook, J. M.; Harper, J. B. Magn. Reson.
Chem. 2009, 47, 342−347.
(13) Csihony, S.; Bodor, A.; Rohonczy, J.; Horvath, I. T. J. Chem.
Soc., Perkin Trans. 1 2002, 2861−2865.
(14) Sicre, C.; Braga, A. A. C.; Maseras, F.; Cid, M. M. Tetrahedron
2008, 64, 7437−7443.
(15) Brecker, L.; Ribbons, D. W. Trends Biotechnol. 2000, 18, 197−
202.
(16) Weber, H.; Brecker, L. Curr. Opin. Chem. Biol. 2000, 11, 572−
578.
(17) Fyfe, C. A.; Damji, S. W. H.; Koll, A. J. Am. Chem. Soc. 1979,
101, 956−960.
(18) Fyfe, C. A.; Damji, S. W. H.; Koll, A. J. Am. Chem. Soc. 1979,
101, 951−955.
(19) Barrios Sosa, A. C.; Williamson, R. T.; Conway, R.; Shankar, A.;
Sumpter, R.; Cleary, T. Org. Process Res. Dev. 2011, 15, 449−454.
(20) Gasparini, G.; Vitorge, B.; Scrimin, P.; Jeannerat, D.; Prins, L. J.
Chem. Commun. 2008, 3034−3036.
(21) Herrera, A.; Fernandez-Valle, E.; Gutierrez, E. M.; Martinez-
Alvarez, R.; Molero, D.; Pardo, Z. D.; Saez, E. Org. Lett. 2010, 12,
144−147.
(22) Herrera, A.; Fernandez-Valle, E.; Martinez-Alvarez, R.; Molero,
D.; Pardo, Z. D.; Saez, E.; Gal, M. Angew. Chem., Int. Ed. 2009, 48,
6274−6277, , S6274-1.
(23) Khajeh, M.; Botana, A.; Bernstein, M. A.; Nilsson, M.; Morris,
G. A. Anal. Chem. 2010, 82, 2102−2108.
9640
dx.doi.org/10.1021/jo201212p|J. Org. Chem. 2011, 76, 9630−9640