New process for the preparation of methyl carbonates

Article abstract of DOI:10.1021/ol034274d

(Matrix presented) The methyl carbonate of HOBt was developed for the conversion of alcohols to carbonates. This method is superior to the use of methyl chloroformate or methyl pyrocarbonate, especially with more hindered alcohols. The reagent is a stable solid that is easily prepared on a multigram scale.

Full text of DOI:10.1021/ol034274d

ORGANIC
LETTERS
2003
Vol. 5, No. 9
1483-1485
New Process for the Preparation of
Methyl Carbonates
Peter G. M. Wuts,* Scott W. Ashford, Andrew M. Anderson, and
Joseph R. Atkins1
Chemical Process Research and DeVelopment, 1500-91-201, Pharmacia,
Kalamazoo, Michigan 49001
peter.g.wuts@pharmacia.com
Received February 17, 2003
ABSTRACT
The methyl carbonate of HOBt was developed for the conversion of alcohols to carbonates. This method is superior to the use of methyl
chloroformate or methyl pyrocarbonate, especially with more hindered alcohols. The reagent is a stable solid that is easily prepared on a
multigram scale.
Carbonates may serve as protecting groups2 or serve as
superior leaving groups in a variety of metal-catalyzed
transformations.3 In connection with another project, we
wished to take advantage of the carbonate as a protecting
group and of its superior leaving group ability in a transition
metal-catalyzed reaction on the tricarbonate 2c. It soon
became clear that the existing technology was inadequate
for conversion of the triol 1 to the tricarbonate 2c. Numerous
attempts using methyl chloroformate and dimethyl pyro-
carbonate failed to give good conversions to the desired
tricarbonate 2c. Table 1 summarizes some of these results
showing that this transformation was indeed problematic. The
cleanest reactions were obtained using TMEDA4 as the base,
and in this case the dicarbonate 2b was formed with some
selectivity; however, this reaction could not be scaled to make
more than a few grams of material. In all these reactions, it
appears that chloroformate decomposition is a problem and
that the released methanol competes with the substrate 1.
The method using MgBr2/(MeO2C)2O was a derivative of a
method used by Vedejs5 for making hindered esters. These
results led us to explore other possible carbonate transfer
agents.
Reagents 3-6 were prepared and tested as possible
carbonate transfer agents (Figure 1). Of these only the HOBt-
Table 1. Carbonate Formations with MeOCOCl and
(MeOCO)2O
entry
reagent
base
solvent
results
1
2
(MeOCO)2O
(MeOCO)2O MgBr2 DIPEA
DMAP
CH2Cl2 40% di, 44% tri
THF
1:1:1 mixture of
mono/di/tri
Figure 1. Potential carbonate-transfer agents.
3
4
5
MeOCOCl
MeOCOCl
MeOCOCl
Pyr
Pyr
CH2Cl2 29-48% di,
30-17% tri
derived reagent 66 was suitable for the preparation of the
tricarbonate 2c.7 The thiadiazole reagent 5 was unstable, and
Pyr
mixture of
mono/di/tri
TMEDA CH2Cl2 <78% di
(1) Summer intern from University of Illinois.
10.1021/ol034274d CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/05/2003

agents 3 and 4 failed to transfer the carbonate using pyridine/
TEA/DMAP. With 2 equiv/OH of 6a, we were routinely able
to obtain a 91% yield of the tricarbonate 2c by simply adding
water to the reaction mixture when TLC showed complete
reaction and then filtering and drying. The solvent for this
reaction proved to be important. Pyridine was the best
solvent, but DMF could also be used although at the expense
of rate and efficiency. The reaction proceeds only very slowly
in solvents such as EtOAc and CH3CN.
DMAP was beneficial as a catalyst as determined by
measuring the heat flow using the Mettler-Toledo Multimax
system. In this experiment, the heat of reaction for the
preparation of carbonate 19b was monitored by measuring
Tr - Tj. Without DMAP, there is only minimal conversion
on the basis of TLC. The curve in Figure 2 shows that when
Figure 2. Heat flow for carbonate formation.
DMAP was added at about 3500 s to the reaction mixture,
a rapid reaction ensues. The reaction was complete at 7400
s on the basis of the heat flow and the TLC.8
Figure 3. Results for the formation of carbonates.
The stability of 6a was also examined using DSC, which
shows melting at 152 °C followed by two exothermic events,
one with an onset at 159 °C (111 J/g, heating rate ) 10
°C/min) and one with an onset at 191 °C (1286 J/g). A DSC
measured in pyridine (10:1 ratio) shows a small exotherm
at 61.5 °C (14.4 J/g) and a larger one at 211.7 °C (118 J/g).
During the course of this work, we found that for the more
hindered alcohols, 2 equiv of 6a were required to drive the
reaction to completion as a result of a competitive side
reaction to form BtOMe 7. It was determined that when 6a
and HOBt in pyridine/TEA were combined that 7 slowly
forms with the liberation of CO2. This was determined using
a ReactIR and a mass spectrometer to monitor the CO2
formation. The benzyl derivative 6c is more prone to this
(6) These reagents are known to be N-acyl derivatives: Fruchier, A.;
Elguero, J.; Hegarty, A. F.; McCarthy, D. G. Org. Magn. Reson. 1980, 13,
339. Singh, J.; Fox, R.; Wong, M.; Kissick, T. P.; Moniot, J. l.; Gougoutas,
J. Z.; Maley, M. F.; Koey, O. J. Org. Chem. 1988, 53, 205.
(7) Use of the allyl reagent for selective protection of unhindered
carbohydrate alcohols: Harada, T.; Yamada, H.; Tsukamoto, H.; Takahashi,
T. J. Carbohydr. Res. 1995, 14, 165.
(2) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic
Synthesis, 3rd ed.; Wiley: New York, 1999.
(3) Tsuji, J.; Minami, I.; Shimizu, I. Tetrahedron Lett. 1983, 24, 1793.
(4) Sano, T.; Ohashi, K.; Oriyama, T. Synthesis 1999, 1141.
(5) Vedejs, E.; Daugulis, O. J. Org. Chem. 1996, 61, 5702.
(8) Dip in the curve at 7400 s is a result of removing heat by taking the
TLC sample.
1484
Org. Lett., Vol. 5, No. 9, 2003

steric hindrance, which is in contrast to the 11R-hydroxy
steroids that react efficiently. Only reagent decomposition
was observed. All other alcohols tested gave clean conver-
sions to their respective carbonates.9
Scheme 1
Acknowledgment. We thank Sean Lapekas for perform-
ing the ReactIR experiment and Steve L. Snyder for running
the DSCs.
Supporting Information Available: Experimental pro-
cedures, NMR, and HRMS data for the preparation of 2b,
2c, 8-20, and 22. This material is available free of charge
via the Internet at http://pubs.acs.org.
decomposition pathway, but the ethyl derivative 6b is more
stable than 6a. It is for this reason that very hindered alcohols
fail to give carbonates with these reagents.
OL034274D
With this new method in hand, we explored the generality
of the process. Figure 3 shows that the overall efficiency
and generality of the reaction is excellent. Two of the
alcohols examined, 7-TES-baccatin III 21 and the 11â-
hydroxyl of steroid 9, could not be derivatized because of
(9) In general, these reactions were performed by dissolving the alcohol
in 10 mL/g of pyridine and 2.5 mL/g of TEA and adding reagent 6 (2
equiv/OH group) and a catalytic amount of DMAP. When the reaction was
complete as determined by TLC or heat flow, the product was isolated by
extraction with EtOAc. In some of the steroid examples, isolation was
performed simply by slowly adding water to crystallize the product.
Org. Lett., Vol. 5, No. 9, 2003
1485