Scheme 2. Thiol and thioether impurity formed in the
To circumvent this problem, the thioamidation reaction was
investigated using catalytic levels of Et3N, and this was found
to be successful using 5-20 mol % of Et3N based on starting
material 3. With this approach, Et3N remained in the reaction
mixture and was deactivated in the next step by methylation to
the quaternary ammonium salt, now present at lower catalytic
levels. This deactivation of Et3N is critical, as it prevents
reversion of 4 back to 3 upon application of heat and removal
of residual H2S. Also, excess residual H2S was found not to be
detrimental to the methylation reaction and could be removed
from the reaction mixture after that step.
thioamidation step
process were compatible with 1,4-dioxane, and this solvent was
particularly convenient for isolation of the final products. By
simply adding an excess of water to the reactor, a separation
of the technical-grade product as an oil occurred. Therefore,
we focused our efforts to develop a one-pot process for 1 using
1,4-dixoane as the common solvent for all reaction steps.
The key starting material, 2-chloro-6-fluorobenzonitrile 3,
was obtained from a 2,6-difluorobenzonitrile production plant
as a 92 GC area % pure compound. Initial preparation of
benzthioamide 4 from 3 was carried by the addition of H2S
gas to 3 in pyridine in the presence of Et3N,5,6 with the exception
that the reaction temperature was lowered to <0 °C to minimize
the formation of an undesired thiol byproduct 10 (Scheme 2).
This thiol byproduct arises from the displacement of fluoride
from 3 by H2S, rather than attack on the cyano functionality,
giving thiol 10. The thiol can then undergo further coupling
reactions with 3 in the presence of excess base to form higher
thioethers 11. Early laboratory experiments demonstrated that,
by carrying out the thioamidation reaction at -10 to -20 °C,
the coupling side reactions to form thioethers were minimized,
and cleaner thioamidate 4 was isolated as a white microcrys-
talline solid by quenching the reaction mixture with water in
which the thiol byproduct 10 was soluble.
Success in developing a one-pot process for a series of
reaction steps hinges on the ability to remove certain reagents
from the reaction mixture without isolating the key reaction
intermediates. This can be accomplished in various ways by
using aqueous wash steps, distillation operations, filtrations, and
such. The first goal in developing a one-pot process for 1 or 2
was to eliminate the isolation of benzthioamide intermediate
4, which would avoid the production of a large aqueous waste
stream and a solids isolation step.
Fortunately, initial tests showed that using 1,4-dioxane
solvent in place of pyridine in the thioamidation step led to
high yields of 4 in solution, with minimal formation of the
undesired byproduct 11. However, in order to take advantage
of this ability to use a common solvent across multiple reaction
steps and eliminate the isolation of 4, Et3N base used during
this thioamidation step must be removed from the reaction
mixture to prevent its methylation to MeEt3N+Br- by MeBr in
the subsequent reaction step. This quaternization reaction, while
not directly interfering with the desired product chemistry,
would consume an excess of methylating agent and produce a
relatively large amount of quaternary ammonium salt in the
waste stream. Initial attempts to remove excess Et3N from the
post thioamidation reaction mixture by distillation showed that
this reaction is easily reversible. Thus, as excess H2S was
removed and moderate heat applied, 4 reverted back to the
starting material 3 in the presence of Et3N.
In practice, it was found that the most convenient approach
to carrying out the thioamidation reaction was with a moderate
H2S pressure of 30 psig in a pressure vessel and heating at 60
°C in 1,4-dioxane solvent and 20 mol % Et3N. This approach
can also be readily applied to a commercial manufacturing scale.
However, for laboratory development efforts, we used the
following process under atmospheric conditions that could
readily be adapted to operating under mild pressure ranges in
a pilot-plant or commercial-scale process. The thioamidation
reaction could be effectively carried out at atmospheric pressure
in a standard jacketed glass reaction vessel by sparging H2S
gas into a 60 °C solution of 3 and catalytic Et3N in 1,4-dioxane.
A dry ice trap was applied to the reactor to condense liquid
H2S, and 1.5-1.7 mol equiv of H2S gas sparged into the 60 °C
solution as fast as the dry ice condenser allowed. In this manner
the thioamidation reaction was carried out to high conversion
within a 24 h period with >95 GC area % of 4 attainable.
Laboratory reactions had shown that methylation of 4 to give
the thioamidate salt 5 was most conveniently accomplished with
dimethylsulfate as the methylating agent. Using 1.3 mol equiv
of dimethylsulfate, a quantitative conversion of 4 to 5 was
achieved within 3 h at 75 °C. However dimethylsulfate is
considered too hazardous for scale-up to a commercial process.
Thus, the use of methyl bromide (MeBr) gas as the methylating
agent was investigated. A higher molar excess of MeBr of
1.5-2.0 equiv and slightly higher temperature (80 °C) was
required to attain high conversion of 4 to 5. The 20 mol % of
Et3N remaining from the previous step reacted with MeBr to
form the quaternary ammonium salt, and the excess H2S
prevented reversion of 4 back to 3. During the course of MeBr
addition, 5 (as the Br- salt) precipitated from solution. After
methylation, excess dissolved H2S gas was removed from the
reactor by a subsurface nitrogen sparge of the reaction mixture
to sweep residual dissolved H2S into a bleach scrubbing
solution.
The coupling step to attach thiophene acid chloride 6 or 8
to the imine nitrogen of 5 (Scheme 1), to give either the
thioimidate 7 or 9 required a pyridine base to promote the
reaction. As both H2S and MeBr will react with pyridine
necessitating additional quantities of the base, a thorough
sparging of the reaction mixture was required prior to the
acylation coupling step to remove these two materials. Triethy-
lamine (TEA), pyridine, and 3-picoline were investigated as
bases in this step. TEA led to lower yields in the subsequent
cyclization step with methylhydrazine. The other bases gave
good results for both the coupling and subsequent cyclization
steps, but 3-picoline was chosen on the basis of cost. At least
(5) Fairfull, A. E. S.; Lowe, J. L.; Peak, D. A. J. Chem. Soc. 1952, 742.
(6) Abbas, K. A.; Edward, J. T. Can. J. Chem. 1985, 63, 3075.
1126
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development