Investigation of the stability of the Corey-Kim intermediate

Full text of DOI:10.1021/op0602122

Organic Process Research & Development 2007, 11, 270−274
Investigation of the Stability of the Corey-Kim Intermediate
Russell D. Cink,* Gilles Chambournier,† Herman Surjono, Zhenglong Xiao, Steve Richter, Marius Naris, and
Ashok V. Bhatia
Global Pharmaceutical Research & DeVelopment, Process Research & DeVelopment, Abbott Laboratories, 1401 Sheridan
Road, North Chicago, Illinois 60064, U.S.A.
Scheme 1. Preparation of the Corey-Kim intermediate
Introduction
The Corey-Kim oxidation1 of alcohols is an important
reaction used by synthetic chemists due to its reliability and
the use of inexpensive reagents. Despite the more recent
development of alternate oxidation methodologies, such as
o-iodoxybenzoic acid (IBX),2 Dess-Martin periodinane,3 and
tetrapropylammonium perruthenate (TPAP),4 the Corey-
Kim oxidation demonstrates selective reactivity5 that con-
tinues to find application in synthesis. The Corey-Kim
intermediate (1) is prepared by the reaction of N-chlorosuc-
cinimide (NCS) with dimethyl sulfide (DMS) as shown in
Scheme 1. The sequential addition of the starting alcohol
and a tertiary amine, typically triethylamine, completes the
oxidation to the ketone or aldehyde. The mechanism for the
Corey-Kim oxidation was evaluated by McCormick6 and
found to be consistent with the mechanism of dimethyl
sulfoxide oxidations.7
The Corey-Kim oxidation was essential for the large-
scale production of the ketolide antibiotic ABT-773 (4,
cethromycin).5b As shown in Scheme 2, the oxidation of the
C-3 hydroxyl of compound 2 provided the ketone 3 as the
final isolated intermediate. Final deprotection of the benzoate
group in 3 completed the synthesis of ABT-773.
For the application of the Corey-Kim oxidation at large
scale, the stability of the intermediate 1 was a significant
concern because Vilsmaier8 has shown that intermediate 1
undergoes decomposition via two pathways. As shown in
Scheme 3, intermediate 1 was converted into succinimide
(5) and methylthiomethyl chloride (6) upon warming, and
addition of triethylamine resulted in the formation of
methylthiomethyl succinimide adducts 7a and 7b.
Because of the temperature sensitivity of intermediate 1
and the exothermic nature of the reaction of DMS with NCS,
control of the reaction temperature was considered critical
to the successful scale-up of the Corey-Kim oxidation.
Additionally, the time required to complete the oxidation was
an important factor because the processing time would
increase upon scale-up and intermediate 1 was expected to
undergo decomposition during that time. In order to deter-
mine acceptable ranges for temperature and time, the stability
of intermediate 1 was investigated at different temperatures
by continuous FT-IR spectroscopy. A kinetic analysis on
the rate of decomposition of intermediate 1 was conducted
using a model system to quantify the concentration of 1
versus time. Because intermediate 1 was shown to undergo
decomposition with triethylamine, the presence of a tertiary
* Author to whom correspondence should be addressed. E-mail:
russell.cink@abbott.com. Telephone: 847-937-0463. Fax: 847-938-2258.
† Current address: Cayman Chemical Company, Ann Arbor, Michigan, U.S.A.
(1) (a) Corey, E. J.; Kim, C. U. J. Am. Chem. Soc. 1972, 94, 7586. (b) Corey,
E. J.; Kim, C. U. J. Org. Chem. 1973, 38, 1233. (c) Corey, E. J.; Kim, C.
U. Tetrahedron Lett. 1974, 287. (d) Corey, E. J.; Kim, C. U.; Takeda, M.
Tetrahedron Lett. 1972, 4339.
Scheme 2. Final Steps in the Synthesis of ABT-773
(2) Frigerio, M.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
(3) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
(4) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis 1994,
639.
(5) For recent applications of the Corey-Kim oxidation, see: (a) Kuwajima,
I.; Tanino, K. Chem. ReV. 2005, 105, 4661. (b) Plata, D. J.; Leanna, M.
R.; Rasmussen, M.; McLaughlin, M. A.; Condon, S. L.; Kerdesky, F. A.
J.; King, S. A.; Peterson, M. J.; Stoner, E. J.; Wittenberger, S. J. Tetrahedron
2004, 60, 10171. (c) Tanino, K.; Onuki, K.; Asano, K.; Miyashita, M.;
Nakamura, T.; Takahashi, Y.; Kuwajima, I. J. Am. Chem. Soc. 2003, 125,
1498. (d) Chandra, K. L.; Chandrasekhar, M.; Singh, V. K. J. Org. Chem.
2002, 67, 4630. (e) Maeng, J. H.; Funk, R. L. Org. Lett. 2002, 4, 331. (f)
Katayama, S.; Fukuda, K.; Watanabe, T.; Yamauchi, M. Synthesis 1988,
178.
(6) McCormick, J. P. Tetrahedron Lett. 1974, 1701.
(7) Epstein, W. W.; Sweat, F. W. Chem. ReV. 1967, 67, 247.
(8) (a) Vilsmaier, E.; Spru¨gel, W. Liebigs Ann. Chem. 1971, 747, 151. (b)
Vilsmaier, E.; Dittrich, K. H.; Spru¨gel, W. Tetrahedron Lett. 1974, 3601.
(c) Vilsmaier, E.; Spru¨gel, W. Tetrahedron Lett. 1972, 625.
270
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development
10.1021/op0602122 CCC: $37.00 © 2007 American Chemical Society
Published on Web 01/25/2007

Scheme 3. Decomposition pathways of the Corey-Kim
intermediate
could be observed. The analysis was based on the absorbance
of the band at 1011 cm-1 due to its increased response
relative to the 1036 cm-1 band. The absorbance of the band
at 1011 cm-1 versus time was plotted in Figure 2 for each
temperature evaluated.
Evaluation of Figure 2 shows that intermediate 1 was
present after 20 h for only the reactions run at -8 and 0 °C.
This was further confirmed by adding isopropanol followed
by triethylamine to the aged Corey-Kim intermediate, and
by analyzing the reactions by GC-MS. For the reactions
run at -8 and 0 °C, products corresponding to a Corey-
Kim oxidation were detected. In this model system, the
methylthiomethyl ether of isopropanol was the major product,
and acetone was formed in small quantity. The propensity
to form the methylthiomethyl ether over the carbonyl
compound has already been documented1a and was attributed
to a solvent as well as a substrate effect. For reactions run
at and above 5 °C, no oxidation products of isopropanol were
detected. Instead, the mass spectrum revealed the presence
of 5, 6, and 7a,b, consistent with the decomposition pathways
shown in Scheme 3.
amino group in compounds 2 and 3 was considered
potentially problematic. The decomposition of intermediate
1 in the presence of triethylamine was monitored by FT-
IR spectroscopy in order to estimate the rate of decomposi-
tion and to serve as a control experiment for the oxidation
of compound 2. The Corey-Kim oxidation of compound 2
was monitored by FT-IR spectroscopy, and the results are
discussed.
Results and Discussion
A kinetic analysis on the decomposition of intermediate
1 using the data in Figure 2 directly was not possible due to
the fact that serial dilution revealed that intermediate 1
exhibited a nonlinear absorbance with respect to concentra-
tion. This nonlinearity was presumed to be due to the biphasic
nature of intermediate 1 in methylene chloride. The nonlin-
earity of the absorbance, coupled with the instability of 1,
made quantifying the concentration of 1 by FT-IR not
sufficiently accurate for a kinetic analysis.
Instead, the kinetic analysis was based on measuring the
extent of a reaction mediated by intermediate 1 versus time.
The reaction selected for the kinetic analysis was the
conversion of benzyl alcohol to benzyl chloride, which has
been shown to proceed in high yield.1d The concentration of
1 was determined by charging benzyl alcohol, one equivalent
relative to the NCS charge, and measuring the amount of
unreacted benzyl alcohol by HPLC. The concentration of 1
was equal to the amount of benzyl alcohol consumed. The
process was repeated at multiple temperatures and NCS
concentrations, [C]NCS, to generate the decomposition profiles
shown in Figure 3.
The Corey-Kim intermediate was prepared according to
the reaction shown in Scheme 1 at temperatures ranging from
-8 to 30 °C. The reactions were conducted in a jacketed
Mettler RC1 calorimeter equipped with a ReactIR probe and
spectra of the mixture were recorded over a 20 h period.
The FT-IR spectra of intermediate 1 were evaluated to
identify bands that were distinct and well resolved from the
bands of NCS, DMS and dichloromethane. Vilsmaier8a
reported the carbonyl peak for intermediate 1 at 1735 cm-1,
but this was not well resolved from the carbonyl peak of
NCS. It was observed that the bands at 1036 and 1011 cm-1
were increasing in intensity versus time when DMS was
added to NCS in dichloromethane as shown in Figure 1.
These two bands could not be attributed to either NCS or
DMS, and were considered diagnostic for the presence of
intermediate 1. The bands at 1036 and 1011 cm-1 were most
consistent with a CH3 rocking vibration of the CH3-S bond
of intermediate 1.9
Figure 1. FT-IR bands of the Corey-Kim intermediate at
-8 °C at 2-min intervals.
By monitoring the intensity of the diagnostic bands over
time at different temperatures the decay of intermediate 1
Figure 2. Decay of the Corey-Kim intermediate versus time
(9) Socrates, G. Infrared and Raman Characteristic Group Frequencies: Tables
and Charts, 3rd ed.; John Wiley & Sons: New York, 2004; p 212.
for the 1011 cm-1 band.
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
271

Table 2. Test experiments on the predictive model
temp (°C) [C]NCS (M) time (h) [C] (M) calcd [C] (M) exp
-12
-4
5
0.254
0.360
0.895
4
4.5
4.5
0.221
0.287
0.641
0.207
0.284
0.626
eq 4 below. Using the empirical relationship between [C]o
and [C]NCS, transforming eq 1 into eq 5, and then inserting
the Arrhenius equation resulted in the model for the
concentration of the Corey-Kim intermediate described
by eq 6.
Figure 3. Decomposition of the Corey-Kim intermediate
versus time.
[C]o ) 0.9079[C]NCS
[C] ) [C]o - kt
(4)
(5)
Table 1. Kinetic results summary
temp (°C)
[C]NCS (M)
[C]o (M)
kobs (1/h)
k (M/h)
[C] ) 0.9079[C]NCS - {A e-E /RT}t
a
(6)
-8
-8
0
0
0
0.535
0.271
0.793
0.535
0.271
0.492
0.239
0.725
0.480
0.235
0.0104
0.0160
0.0275
0.0377
0.0587
0.00513
0.00383
0.0199
0.0181
0.0138
One of the major drawbacks to a kinetic model that
incorporates an empirical factor is the accuracy of the
prediction is poor when extrapolated outside the parameter
ranges used to develop the model. In order to test the
accuracy of the model, several experiments were run outside
the concentration and temperature range, along with a control
experiment within the range. The results of these test
experiments, shown in Table 2, demonstrate the calculated
concentration being within 7% of the experimental concen-
tration. Comparing the model to the data shown in Figure 3
revealed that the accuracy of the model decreases signifi-
cantly at [C]/[C]o values less than 0.3.
The two main conclusions drawn from the experimental
results were that in order to maximize the concentration of
the Corey-Kim intermediate the temperature should be
maintained as low as possible, as expected, and that the
concentration should be as high as possible. There was a
practical upper limit for the concentration used in the
preparation of 1; a sufficient solvent volume was required
to act as a heat sink due to the exothermic nature of the
reaction. The time required to complete the Corey-Kim
oxidation step at the manufacturing scale was estimated at
4 h or less, from the addition of dimethyl sulfide to the
addition of triethylamine. Based on the model developed for
the decay of intermediate 1, at [C]NCS of 0.535 M less than
25% degradation would occur during the 4 h required to
complete the process when the temperature was maintained
at or below 0 °C.
For this analysis, the data was plotted as [C]/[C]o versus
time, where [C]o was the measured concentration at time
zero determined by adding the benzyl alcohol immediately
after the DMS addition. Ideally the concentration at time
zero would be equivalent to the starting NCS concentration.
However, the measured concentration at time zero was
approximately 90% of the NCS concentration due to the
less than quantitative yield for the formation of 1 and the
reaction with benzyl alcohol. Figure 3 shows a linear
decomposition profile for intermediate 1 with the rate of
decomposition increasing with temperature, consistent with
the FT-IR analysis. The data also show that the rate of
decomposition, kobs
, increases as the concentration
decreases. These three findings were consistent with zero-
order kinetics, as described by eqs 1 and 2 below, where
the observed rate constant kobs was inversely related to the
initial concentration.
[C]/[C]o ) 1 - kobs
t
(1)
(2)
kobs ) k/[C]o
k ) A e-E /RT
(3)
a
The data from Figure 3 were summarized in Table 1 in
terms of the time zero concentration and observed rate
constant for each temperature and NCS concentration
combination examined. Equation 2 above was used to
calculate k for evaluation of the Arrhenius equation, eq 3
above, to account for the temperature dependence of the rate
constant. A plot of ln k versus 1/T (K) provided the activation
energy, Ea, at 101.5 kJ/mol from the slope of the line, and
the pre-exponential factor A at 4.31 × 1017 M/h as the
intercept.
With the evaluation of the decomposition of the inter-
mediate 1 versus temperature complete, the decomposition
in the presence of tertiary amines was evaluated by FT-IR.
Intermediate 1 was generated at -8 °C as shown in Scheme
1, followed by addition of triethylamine at the same
temperature. As shown by the disappearance of the 1011
cm-1 band in Figure 4, tertiary amines destroy intermediate
1 even at low temperatures in the absence of an alcohol
substrate. GC-MS analysis of the reaction mixture con-
firmed the decomposition of the intermediate, showing the
main product to be 7a,b, as shown in Scheme 3.
In order to develop a predictive model for the concentra-
tion of 1, a relationship between the NCS concentration and
the time zero concentration was required. A plot of [C]o
versus [C]NCS resulted in a the linear relationship shown in
For alcohol substrates containing a tertiary amino group,
such as 2, two competing pathways would exist; the amine
272
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development

Figure 4. Peak intensity of the 1011 cm-1 band upon addition
of triethylamine to intermediate 1.
Figure 5. Peak intensity of 1011 cm-1 band of the Corey-
Kim oxidation of compound 2.
could cause decomposition of intermediate 1 or act as the
base and cause oxidation to the desired product. If the amino
group caused decomposition, the reaction would likely
require an excess of intermediate 1 to counter the negative
effect of the amine and allow the oxidation to go to
completion. If the amino group of the substrate was
functioning as a base in the reaction, product would be
formed prior to the addition of triethylamine and a substo-
ichiometric amount of triethylamine would be required.
The Corey-Kim oxidation of the C-3 hydroxyl of
compound 2, as shown in Scheme 4, was evaluated using
FT-IR spectroscopy. The Corey-Kim intermediate was
prepared at -8 °C using 1.6 equiv of N-chlorosuccinimide
relative to 2. The addition of compound 2 to the Corey-
Kim intermediate did not result in a change in FT-IR
spectrum that could be attributed to the formation of
intermediate A in Scheme 4. It was possible that both the
Corey-Kim intermediate and intermediate A exhibited bands
at 1036 and 1011 cm-1.
Upon addition of 1.9 equiv of triethylamine, the FT-IR
band at 1011 cm-1 completely disappeared within minutes.
HPLC analysis showed the reaction was complete within 1
h after the addition of the triethylamine. It should be noted
that, while the addition of triethylamine is necessary to
complete the oxidation step, the use of excess triethylamine
relative to N-chlorosuccinimide consumed the excess Corey-
Kim intermediate. In order to maintain active Corey-Kim
intermediate after the addition of triethylamine, the amount
of triethylamine was reduced to the minimum required to
complete the oxidation. Range-finding experiments revealed
that 1.0 equiv of triethylamine resulted in complete oxidation
within 2 h, and as little as 0.5 equiv of triethylamine caused
complete oxidation after 19 h. When 1.0 equiv of triethy-
lamine was used in the oxidation of compound 2 and after
the reaction was complete by HPLC analysis, an additional
10 mol % of 2 could be added to the reaction and undergo
oxidation to 3. This result showed that active Corey-Kim
intermediate was still present when 1.0 equiv of triethylamine
was used. These results indicated that for a substrate
containing a tertiary amine, the Corey-Kim oxidation can
be completed with substoichiometric amounts of base;
however, the reaction rate was slower, and stability of the
intermediate might be problematic. The tertiary amino group
of 2 was not likely causing significant degradation of the
Corey-Kim intermediate because the oxidation could be
completed with less N-chlorosuccinimide (1.4 equiv)5b than
described in the original Corey-Kim procedure (1.5 equiv).1a
In conclusion, the Corey-Kim intermediate was found
to decompose over time, and the rate of decomposition was
related to temperature and concentration. A model was
developed to predict the concentration of the Corey-Kim
intermediate as a function of time, temperature, and NCS
concentration. The temperature dependence of the decom-
position rate was defined using the Arrhenius equation. Based
on the model developed, less than 25% decomposition of
the Corey-Kim intermediate would occur under typical
reaction conditions when the temperature was maintained
at or below 0 °C. The addition of triethylamine to complete
the oxidation process, typically in excess relative to NCS,
also destroyed the excess Corey-Kim intermediate. In order
to maintain active Corey-Kim intermediate, the amount of
triethylamine was reduced to the minimum required to
complete the oxidation of compound 2 within 2 h. The
oxidation of compound 2 was successful at large scale (>300
Scheme 4. Corey-Kim oxidation of compound 2
The addition of compound 2 to the Corey-Kim inter-
mediate resulted in a slow but steady decrease in the intensity
of the 1011 cm-1 FT-IR band, as shown in Figure 5. HPLC
analysis of this reaction revealed that the desired product,
ketone 3, is formed in ∼20% prior to the addition of
triethylamine. In this instance, the tertiary amine in com-
pound 2 was acting as the base to cause the oxidation. The
partial conversion of compound 2 to 3 would consume a
fraction of the Corey-Kim intermediate and was consistent
with the slow decrease in the intensity of the 1011 cm-1
FT-IR band shown in Figure 5.
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
273

kg) maintaining the temperature below 0 °C with a slight
excess of the Corey-Kim intermediate.
intermediate was equal to the amount of benzyl alcohol
consumed.
Evaluation of the Effect of Triethylamine on the
Corey-Kim Intermediate. The Corey-Kim intermediate
was prepared in the 1-L RC-1 calorimeter at -8 °C as
described above. The mixture was then stirred for 40 min
while being periodically monitored by FT-IR. Triethylamine
(35 g, 0.35 mol, 48 mL) was added dropwise via addition
funnel while maintaining the internal temperature in the
vessel at -8 °C. The mixture was stirred for 1 h, and a
sample was analyzed by GC-MS.
Evaluation of the Corey-Kim Oxidation of Compound
2. The Corey-Kim intermediate was prepared in the 1-L
RC-1 calorimeter at -8 °C as described above using
N-chlorosuccinimide (31.6 g, 0.237 mol), dichloromethane
(525 g), and dimethyl sulfide (15.4 g, 0.248 mol). The
mixture was stirred for 1.5 h while being periodically
monitored by FT-IR. A solution of compound 2 (131.2 g,
0.15 mol) in dichloromethane (457 g) was added over 1 h.
The mixture was stirred for 1 h, and then triethylamine (28.7
g, 0.284 mol) was added dropwise via addition funnel over
20 min while maintaining the internal temperature in the
vessel at -8 °C. The mixture was stirred and periodically
monitored by FT-IR.
GC-MS analysis of major components: compound:
formula, formula weight; GC: retention time; MS: m/z
(intensity, rel %), molecular ion, M+. Triethylamine: C6H15N,
FW ) 101; GC RT ) 5.8 min; MS: 86 (100%), M+ 101
(18%). Methylthiomethyl chloride, 6: C2H5ClS, FW ) 96;
GC: RT ) 8.7 min; MS: 35 (17%), 45 (35%), 47 (8%),
49/51 (8%/3%), 61 (100%), 81/83 (3%/1%), M+ 9 6/98
(46%/18%). Isopropyl methylthiomethyl ether: C5H12OS,
FW ) 120; GC: RT ) 10.6 min; MS: 43 (100%), 61 (45%),
73 (45%), 105 (1%), M+ 120 (18%). Succinimide, 5: C4H5-
NO2, FW ) 99; GC: RT ) 15.2 min; MS: 42 (11%), 56
(74%), 70 (6%), M+ 99 (100%). Methylthiomethyl succin-
imide, 7a,b: C6H9NO2S, FW ) 159; GC: RT ) 18.2 min;
MS: 47 (8%), 55 (100%), 61 (11%), 84 (65%), 100 (12%),
112 (76%), 144 (26%), M+ 159 (71%).
Experimental Section
General Experimental. The reactions were carried out
in a Mettler-Toledo reaction calorimeter (RC1), and the FT-
IR spectra were recorded using a Mettler-Toledo AutoChem
ReactIR 1000. An air background spectrum was recorded
before starting each FT-IR experiment. The GC-MS
spectra were recorded on a Hewlett-Packard 5890/5972 with
a jet separator interface, fitted with a J&W, DB-624 30 m
× 0.53 mm, 3 µm column; helium carrier gas, flow rate 6.8
mL/min; injector 150 °C, jet separator 210 °C, detector 280
°C, oven program: 45 °C, hold 7 min, rate 20 °C/min, 210
°C, hold 10 min. The HPLC analyses were performed on a
Zorbax or Phenomenex C8 column, eluting with a gradient
of acetonitrile and water (0.1% phosphoric acid). The
chemicals and solvents were purchased form the Aldrich
Chemicals and EM Science companies and used as re-
ceived. Compound 2 was prepared according to the known
procedure.5b
Evaluation of Corey-Kim Intermediate at Different
Temperatures over Time. To the 1-L RC-1 calorimeter
were charged N-chlorosuccinimide (41 g, 0.31 mol) and
dichloromethane (681 g). The mixture was stirred and
adjusted to the desired temperature. Dimethyl sulfide (20 g,
0.32 mol, 24 mL) was added dropwise via addition funnel.
The mixture was then stirred for 20 h at the desired
temperature while being periodically monitored by FT-IR.
Following the 20 h mix, isopropanol (11 g, 0.18 mol, 14
mL) was added dropwise via addition funnel. The resulting
mixture was stirred at the desired temperature for ap-
proximately 1 h, and then triethylamine (35 g, 0.35 mol, 48
mL) was added dropwise via addition funnel. The mixture
was stirred for 1 h, and a sample was analyzed by GC-MS.
Determination of the Amount of Corey-Kim Inter-
mediate by Reaction with Benzyl Alcohol. The Corey-
Kim intermediate was prepared as described above at 20%
scale using a jacketed reactor at the desired temperature. The
mixture was stirred for the desired amount of time, and then
benzyl alcohol (6.641 g, 0.06141 mol, 1.0 equiv) was added.
The mixture was warmed to 20 °C over 45 min and then
quantitatively transferred to a volumetric flask, diluting to
volume with acetonitrile. After serial dilution, the amount
of unreacted benzyl alcohol was quantified by HPLC relative
to a benzyl alcohol standard. The amount of Corey-Kim
Acknowledgment
We thank Howard Morton and M. Robert Leanna for
review of the manuscript.
Received for review October 13, 2006.
OP0602122
274
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development