Electrogenerated base-promoted synthesis of organic carbonates from alcohols and carbon dioxide

Article abstract of DOI:10.1002/1099-0690(200007)2000:13<2445::AID-EJOC2445>3.0.CO;2-2

Electrogenerated bases promote the reaction between primary alcohols and carbon dioxide to give organic carbonates in excellent yields. Secondary alcohols are converted in moderate yields, whereas tertiary alcohols and phenols are unreactive. 1,2-Diols give a mixture of both cyclic and linear diand monocarbonates. These latter are intermediates in the reaction pathway leading to the cyclic derivatives.

Full text of DOI:10.1002/1099-0690(200007)2000:13<2445::AID-EJOC2445>3.0.CO;2-2

FULL PAPER
Electrogenerated Base-Promoted Synthesis of Organic Carbonates from
Alcohols and Carbon Dioxide
[a]
[a]
[b]
Maria Antonietta Casadei,* Stefania Cesa, and Leucio Rossi
Keywords: Electrogenerated bases / Carbon dioxide / Alcohols / Organic carbonates / Dimethyl carbonate
Electrogenerated bases promote the reaction between prim-
ary alcohols and carbon dioxide to give organic carbonates
active. 1,2-Diols give a mixture of both cyclic and linear di-
and monocarbonates. These latter are intermediates in the
in excellent yields. Secondary alcohols are converted in mod- reaction pathway leading to the cyclic derivatives.
erate yields, whereas tertiary alcohols and phenols are unre-
Introduction
ized. Finally, under optimized conditions, representative al-
cohols and 1,2-diols (Scheme 1) have been subjected to the
Organic carbonates are an important class of compounds carboxylation process, to establish the generality and limits
whose versatility allows their application in several fields of this new procedure.
of the chemical and pharmaceutical industry. Recently, the
utilization of organic carbonates and the methods for their
synthesis have been extensively reviewed.[ The most im-
1]
portant routes to these esters involve the direct or indirect
use of phosgene, a very toxic and corrosive reagent. Other
available methods use carbon monoxide, drastic conditions
and/or metal catalysts, whose potential environmental im-
pact should not be underestimated. We have already de-
scribed a new approach to organic carbonates based on the
electrocarboxylation of alcohols and 1,2-diols following ac-
tivation of carbon dioxide by either cathodic reduc-
[
2,3]
[3,4]
tion or electrogenerated superoxide anion.
Both
methods allow the synthesis of linear and cyclic carbonates
under very mild and safe conditions, but not in very high
yields. Therefore, we have reconsidered the synthesis of or-
ganic carbonates from alcohols and carbon dioxide with the
aim of increasing the yield of the process. Starting from the
hypothesis that the first step of the reaction involves sub-
strate deprotonation, we decided to use electrogenerated ba-
ses (EGBs), which can be obtained by reducing suitable
probases (PBs). The usefulness of EGBs in organic elec-
trosynthesis is well documented[ and the possibility of gen-
erating bases of different stability and reactivity by simply
controlling the nature of the counterion[ is of utmost util-
ity.
5]
6]
Scheme 1.
The efficiency of some representative EGBs in the carb-
oxylation of benzyl alcohol 5a, which was selected as a
model compound, has been tested. Under normal condi-
tions, the 2-pyrrolidone anion arising from PB 4 gives the Results and Discussion
highest yield of carbonate 6a. The experimental parameters
of the reaction involving 4 and 5a were subsequently optim-
Probases 1–4 have been tested in the carboxylation reac-
tion of benzyl alcohol 5a. Tetraethyl ethylenetetracarb-
[
[
a]
b]
Dipartimento di Studi di Chimica e Tecnologia delle Sostanze oxylate (1) has been reduced on a mercury pool cathode
Biologicamente Attive, Universit a` degli Studi ‘‘La Sapienza’’,
under potentiostatic control in the presence of 5a and CO2,
P. le Aldo Moro 5, I-00185-Roma, Italy
Fax: (internat.) ϩ 39-06/49913133
E-mail: macasadei@uniroma1.it
employing different PB concentrations and the different
current amounts necessary for its complete reduction to
EGB. The results of the HPLC analyses carried out on the
reaction mixture obtained after addition of an excess of EtI
Dipartimento di Chimica, Ingegneria Chimica e Materiali,
Universit a` degli Studi
I-67040-Monteluco di Roio, L’Aquila, Italy
Eur. J. Org. Chem. 2000, 2445Ϫ2448

WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ193X/00/0713Ϫ2445 $ 17.50ϩ.50/0
2445

M. A. Casadei, S. Cesa, L. Rossi
FULL PAPER
Table 1. Carboxylation of benzyl alcohol 5a (2 mmol): choice of tained using 5 mmol of PB and a current value of 6.0 mF.
3
the experimental conditions (Pt anode; catholyte ϭ CH CN/0.1 
A further increase of the PB concentration and of the total
current amount produces a decrease in the carbonate yield
and a greater number of unidentified products, some of
which arise from interactions between the EGB and the
solvent, as ascertained on the basis of their IR spectra. To
eliminate these undesired reactions, N,N-dimethylformam-
TEAP)
[
a]
mF[b]
Entry PB
mmol)
T
[°C]
E [V] or –2
Products
[c]
(
I [mA cm ] (yield% )
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
1 (1.0)
1 (1.5)
1 (2.0)
1 (3.0)
1 (4.0)
1 (5.0)
2 (4.0)
3 (4.0)
4 (5.0)
4 (10.0)
4 (5.0)
4 (5.0)
4 (5.0)
4 (5.0)
4 (10.0)
2.0
3.0
25 –1.0
5a (68), 6a (31)
25 –1.0
25 –1.0
25 –1.0
25 –1.0
25 –1.0
25 –0.5
25 –1.0
25 15
5a (52), 6a (35) ide was used instead of acetonitrile under otherwise ident-
4.0
5a (46), 6a (40)
ical experimental conditions, but the drastic decrease in the
carbonate yield (Table 1, entry 16) excluded the possibility
6.0
5a (33), 6a (46)
8.0
5a (30), 6a (49)
10.0
5a (28), 6a (50) of employing this solvent. Therefore, the optimized condi-
[
d]
6.1
8.0
4.0
8.0
4.0
4.0
4.0
6.0
8.0
6.0
5a (36), 6a (32)
tions for the carboxylation of 5a (2 mmol) to 6a require the
5a (65), 6a (24)
5a (20), 6a (63) reduction of pyrrolidone 4 (5 mmol) in acetonitrile (50 mL)
[
e]
1
1
1
1
1
1
1
25 15
6a (65)
at 0° C with a consumption of 6.0 mF of current.
Under these conditions, several representative alcohols
0
15
5a (18), 6a (78)
5a (27), 6a (68)
–25 15
–50 15
5a (24), 6a (58) and 1,2-diols (5 and 7 in Scheme 1) were subjected to the
0
0
0
15
15
15
6a (88)
EGB-promoted carboxylation process. The results of the
6a (78)
[
f]
analyses carried out on the corresponding reaction mixtures
are reported in Table 2 and allow us to draw some conclu-
sions. The primary alcohols 5a–c are converted into the cor-
responding carbonates in excellent yields (entries 1–3). Un-
4 (5.0)
5a (67), 6a (31)
[
a]
A Hg pool (probases 1–3) or a Pt gauze (probase 4) was used as
[
b]
[c]
cathode. –
Total current amount in mF. – HPLC analysis. –
[e]
[
d]
Compound 2 undergoes competitive dimerization. –
Uniden-
[f]
tified products did not allow to quantify 5a. – DMF was used der the same experimental conditions, the secondary alco-
as solvent.
hol 5d gives a lower yield of 6d (entry 4). A higher conver-
sion can be obtained by increasing the current value to 8.0
mF (entry 5), although a further increase of the PB concen-
are reported in Table 1 (entries 1–6). As expected, higher tration and of the total current amount (see Experimental
yields of benzyl ethyl carbonate (6a) are observed by in- Section) does not produce better results (entry 6). The ter-
creasing the PB concentration; they achieve a maximum tiary alcohol 5e and phenol 5f do not undergo the carb-
value when the PB concentration is twice that of the sub- oxylation reaction. It is likely that steric hindrance and/or
strate. However, we considered an equimolar concentration the unfavorable deprotonation equilibrium between EGB
[7]
[9]
of PB (and thus, theoretically, two equivalents of EGB ) and 5e do not allow the carbonate formation. On the
as a good compromise between chemical yield and current other hand the poor nucleophilicity of the 5f anion is prob-
consumption and chose such conditions to test the other ably responsible for the unsuccessful formation of 6f. When
probases. When using diethyl bromomalonate (2) and ethyl diols 7a–c are subjected to the above-described procedure,
2
(
-bromo-2-methylpropionate (3) the yield of 6a decreases complex mixtures are obtained in which the cyclic carbon-
Table 1, entries 7,8).[ On the other hand, the EGB arising ates 8a–c and linear monocarbonates 9a–c are present, to-
7]
from 2-pyrrolidone (4) promotes the highest yield in the gether with the dicarbonates 10a,b starting from 7a,b
[8]
conversion of 5a into carbonate 6a (Table 1, entry 9). In (Table 2, entries 9–16). Once again, the value of the PB con-
these conditions further increases of the PB concentration centration and the current amount affect the course of the
and of the current amount do not increase the carbonate reaction. In the case of 7a, the yield of cyclic carbonate
yield (Table 1, entry 10).
reaches a maximum for a current value of 6.0 mF and a
The reduction of 4 was carried out on a platinum gauze further increase and/or a higher PB concentration only
cathode under galvanostatic control. The reduction poten- bring about higher yields of dicarbonate 10a (Table 2, ent-
tial value of 4 is very negative, but its EGB is stable enough ries 9–12). A similar behavior is observed for 7b, although
to be generated and then utilized after the current is higher yields of 8b and 10b are obtained by increasing PB
switched off, thus allowing its use even with substrates elec- and total current amounts (Table 2, entries 13,14). Finally,
[5d]
troactive at the working potential.
higher yield of 6a (cf. entries 3,7–9 in Table 1), compound (Table 2, entries 15,16). It has already been proved that the
was selected to carry out all the other carboxylation reac- linear carbonates 9 are intermediates in the reaction path-
Since it gives the the yield of 8c is essentially unaffected by these parameters
4
tions. The next step of the research was the optimization of way yielding to the formation of 8. Actually, samples of
the experimental conditions. We first investigated the influ- compounds 9, independently synthesized by chemical
ence of the temperature on the formation of 6a, in the range methods, were converted into 8 when treated under carb-
from ϩ25 °C to –50 °C. HPLC analyses of these solutions oxylation conditions.[ Further evidence has been collected
are reported in Table 1 (entries 9,11–13) and show 0 °C to during the present study: both the treatment of isolated 9
be the optimum electrolysis temperature. Another set of ex- under workup conditions and the simple use of longer reac-
periments was carried out in order to establish the influence tion times promote the conversion of 9 into 8. The dicar-
of the PB concentration and of the employed electricity bonates 10 have never been detected in the reaction mix-
3]
(Table 1, entries 11,14,15). The best yield of 6a was ob- tures obtained employing the two previously reported
2446
Eur. J. Org. Chem. 2000, 2445Ϫ2448

Electrogenerated Base-Promoted Synthesis of Organic Carbonates
FULL PAPER
methods of electrocarboxylation.[3] It is likely that the con-
jugate base of 4 is strong enough to deprotonate both OH
groups, which are subsequently carboxylated. The forma-
tion of 10c from 7c is probably prevented for the reasons
suggested before in the case of 5e. It is interesting to note
that the formation of the cyclic carbonates takes place with
Experimental Section
General Remarks: The electrochemical apparatus, the cells, and the
reference electrode have already been described. The value of the
working potentials are reported relative to SCE. Acetonitrile (Rie-
del-de Ha e¨ n), N,N-dimethylformamide (Riedel-de Ha e¨ n) and tetra-
[3]
total retention of the absolute configuration. Actually, ethylammonium perchlorate (TEAP, Fluka) were purified as previ-
starting from meso-7a, only cis-8a is obtained. This is in ously described.[ All the starting materials and dimethyl carbon-
agreement with a reaction mechanism which does not in- ate 6c are commercially available. – Column chromatography (c.c.)
10]
was performed on Merck silica gel (70–230 mesh; 100 g per 1 g of
volve the cleavage of the C–O bond at the chiral carbon
crude reaction mixture). – IR, NMR, HPLC, GC and melting point
atom.
apparatus were as previously described. [ – H NMR spectra were
3]
1
recorded as solutions in CDCl
HPLC analyses were carried out using a Merck Hibar LiChrocart
250–4, 5 µm) RP-18 column with a CH CN/H O mixture in a
linear gradient from 35:65 to absolute CH CN over 20 min as elu-
3 4
, with Me Si as internal standard. –
Table 2. Carboxylation of 5a–f and 7a–c (PB ϭ 4; Pt anode and
cathode; catholyte ϭ CH CN/0.1  TEAP; T ϭ 0 °C; I ϭ 15
(
3
2
3
–
2
mAcm )
3
ent in the case of the solutions from 5a,b,d؊f. The same mixture,
in a linear gradient from 30:70 to 60:40 over 12 min, followed by
an isocratic step at this composition during 10 min, was used when
Entry Substrate mF[a] Products
yield%[
)
b]
(
2
starting from 7a,b. Finally, an MeOH/H O mixture in a linear gra-
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
5a
5b
5c
5d
5d
5d
5e
5f
7a
7a
7a
7a
7b
7b
7c
7c
6.0 6a (88) [74]
6.0 5b (20), 6b (80) [66]
6.0 6c (90)
dient from 30:70 to 60:40 over 10 min and a further isocratic step at
this composition during 10 min was employed when starting from
–1
6.0 6d (34) [31]
8.0 6d (50)
7c and to quantify alcohol 7b. The flow rate was always 1 mL min
.
–
GC analyses of the solution from 5c were carried out using a
12.0 6d (52)
Supelco Porapack PS 100 packed column (6 feet ϫ 1/8 inch) in the
range 90–190 °C. – Quantitative HPLC analyses were performed
with the internal standard method, whereas a calibration curve was
used in the case of GC analyses.
6.0 5e (98)
[c]
6.0 5f (51)
4.0 7a (71), 8a (20), 9a (7), 10a (2)
6.0 7a (63), 8a (29), 9a (4), 10a (4)
8.0 7a (62), 8a (30), 9a (2), 10a (5)
12.0 7a (50), 8a (29) [30], 9a (5) [2], 10a (12) [7]
1
1
1
1
1
1
1
6.0 7b (45), 8b (18) [24], 9b (11) [4], 10b (9) [9] Chemistry: Ethyl carbonates 6, cyclic carbonates 8, and monocar-
12.0 7b (38), 8b (28), 9b (10), 10b (16)
6.0 7c (62), 8c (19) [27], 9c (18) [9]
12.0 7c (55), 8c (21), 9c (15)
bonates 9b,c were prepared and characterized following a known
[3,4]
procedure.
The monocarbonate 9a and dicarbonates 10a,b were
obtained by reacting equimolar amounts of the corresponding diol
[
11]
[
a]
[b]
and ethyl chloroformate according to standard procedures, and
purified by c.c. with a mixture of light petroleum/AcOEt (3:2) in
Total current amount in mF. –
HPLC (GC in entry 3) yields
are given in parentheses; isolated yields in square brackets. –
[
c]
PhOEt was detected by HPLC analysis.
3 2
the case of 9a and 10a and CHCl /Me CO (95:5) in the case of 10b
as eluents.
Conclusion
Ethyl 2-Hydroxy-1,2-diphenylethyl Carbonate (9a): IR (film): ν ϭ
–
1
1
3
500, 1745 cm . – H NMR: δ ϭ 1.18 (t, J ϭ 7.2 Hz, 3 H, CH
We have developed a new method of carboxylation of 2.28 (s, 1 H, OH), 4.04 (q, J ϭ 7.2 Hz, 2 H, OCH ), 4.98 (d, J ϭ
alcohols with CO based on the employment of electrogen- 5.8 Hz, 1 H, CHOH), 5.70 (d, J ϭ 5.8 Hz, 1 H, CHO), 7.25 (s, 10
3
),
2
2
1
3
H, arom). – C NMR: δϭ 14.10, 64.16, 76.01, 82.40, 126.96,
erated bases. It allows the formation of linear carbonates
from primary and secondary alcohols in excellent to good
yields, respectively. Tertiary alcohols and phenols do not
react at all, so that selective carboxylation of suitable po-
lyhydroxy compounds can be anticipated. 1,2-Diols are
converted into a mixture of cyclic and linear carbonates, the
latter being precursors of the former. Dicarbonates are also
1
1
27.64, 127.69, 128.04, 128.11, 128.18, 128.52, 135.82, 139.25,
54.22. – C17 (286.31): calcd. C 71.31, H 6.34; found C 71.15,
18 4
H O
H 6.24.
Ethyl 2-Ethoxycarbonyloxy-1,2-diphenylethyl Carbonate (10a): m.p.
–1
1
1
10–112 °C. – IR (nujol) ν ϭ 1748 cm . – H NMR: δ ϭ 1.12 (t,
J ϭ 7.2 Hz, 3 H, CH ), 1.23 (t, J ϭ 7.2 Hz, 3 H, CH ), 4.10 (q,
3
3
formed, but not when a tertiary hydroxy group is present J ϭ 7.2 Hz, 2 H, OCH ), 4.11 (q, J ϭ 7.2 Hz, 2 H, OCH ), 5.90
2
2
1
3
in the molecule. If compared to the previously described (s, 2 H, 2 ϫ CH), 7.24–7.30 (m, 10 H, arom). – C NMR: δϭ
electrochemical methods, the use of EGBs sharply increases 14.08, 64.23, 79.92, 127.61, 128.06, 128.51, 135.19, 154.11. –
20 22 6
C H O (358.38): calcd. C 67.02, H 6.19; found C 66.85, H 6.08.
the yields of organic carbonates from alcohols and CO2.
Of relevant interest, owing to its wide and multi-purpose
applications in chemical industry, is the achievement of a
Ethyl 2-Ethoxycarbonyloxy-1-phenylethyl Carbonate (10b): IR
–
1
1
(
film) ν ϭ 1750 cm . – H NMR: δ ϭ 1.28 (t, J ϭ 7.2 Hz, 3 H,
CH ), 1.30 (t, J ϭ 7.2 Hz, 3 H, CH ), 4.18 (q, J ϭ 7.2 Hz, 2 H,
OCH ), 4.20 (q, J ϭ 7.2 Hz, 2 H, OCH ), 4.36 (d, J ϭ 5.5 Hz, 2
H, CHCH ), 5.87 (t, J ϭ 5.5 Hz, 1 H, CHCH ), 7.38 (s, 5 H,
arom). – C NMR: δϭ 14.09, 14.14, 64.23, 68.98, 76.88, 126.62,
9
0% yield of dimethyl carbonate. The mild and safe condi-
3
3
tions, which avoid the use of dangerous and polluting re-
agents, together with the high yields attained, make this
procedure advantageous with respect to the already avail-
2
2
2
13
2
able methods especially for the carboxylation of primary al- 128.69, 128.86, 135.68, 154.28, 154.78. – C H O (282.28): calcd.
1
4
18
6
cohols.
C 59.56, H 6.43; found C 59.49, H 6.31.
Eur. J. Org. Chem. 2000, 2445Ϫ2448
2447

M. A. Casadei, S. Cesa, L. Rossi
FULL PAPER
[
[
1]
2]
A. G. Shaikh, S. Sivaram, Chem Rev. 1996, 96, 951–976.
M. A. Casadei, A. Inesi, L. Rossi, Tetrahedron Lett. 1997, 38,
Electrochemistry
3
565–3568.
Reduction of Probases 1–3: The controlled-potential electrolyses
were carried out at a mercury pool cathode by stepwise addition
of the probase to a solution of 5a (2 mmol) in CH
[
3]
M. A. Casadei, S. Cesa, M. Feroci, A. Inesi, New J. Chem.
1
999, 23, 433–436.
3
CN/0.1  TEAP
[4]
M. A. Casadei, S. Cesa, M. Feroci, A. Inesi, L. Rossi, F. Mich-
(
50 mL) where CO was bubbling. The working potential value al-
2
eletti Moracci, Tetrahedron 1997, 53, 167–176.
[
5] [5a]
[5b]
lows the selective reduction of the PB. At the end of the reduction,
EtI (10 mmol) was added, and the mixture was maintained at room
temperature under stirring overnight. It was then analyzed.
J. H. P. Utley, Top. Curr. Chem. 1987, 142, 131–165. –
J.
H. P. Utley, S. K. Ling-Chung, C. Z. Smith, in Electroorganic
Synthesis (Eds.: R. D. Little, N. L. Weinberg ), Marcel Dekker
[
5c]
Inc:, New York, 1991, p. 377. –
T. Fuchigami, A. Konno,
in Electroorganic Synthesis (Eds.: R. D. Little, N. L. Weinberg),
Reduction of Probase 4: The electrolyses were carried out under
galvanostatic control (I ϭ 15 mA cm ) at a platinum gauze cath-
ode on solutions of 4 (5 mmol for current amounts up to 8.0 mF,
[5d]
Marcel Dekker Inc:, New York, 1991, p. 387. –
D. Bon-
–2
afoux, M. Bordeau, C. Biran, J. Dunogu e` s, J. Organomet.
Chem. 1995, 493, 27–32, and references cited therein.
[
[
6]
7]
10 mmol for 12.0 mF) in CH
3
CN/0.1  TEAP (50 mL). After the
T. Shono, M. Ishifune, T. Okada, S. Kashimura, J. Org. Chem.
1
991, 56, 2–4.
consumption of the predetermined number of mF, the current was
switched off, the substrate (2 mmol) was added to the mixture and,
It is well established that the reduction of 1 needs 2F/mol to
give the divalent EGB, whereas halo compounds such as 2 and
10 min later, the bubbling of CO
2
was started and maintained for
3
undergo a two-electron reductive cleavage to halide and car-
1
h. EtI (MeI in the reaction of 5c) (10 mmol) was added, the solu-
banion. It follows that, in theory, in order to use the same
number of equivalents of EGB, a twofold molar amount of 2
and 3 with respect to 1 must be employed.
8]
tions were stirred overnight at room temperature, and then ana-
lyzed. When the components of the mixtures were separated, a
[
The reduction of 4 needs 1F/mol to give an equivalent of EGB.
In general, under galvanostatic control conditions, an excess of
electroactive substrate is necessary to avoid the discharge of the
1
mL sample was withdrawn for HPLC analysis. The solvent was
removed from the remaining solution under reduced pressure and
the residue was extracted with Et O (5 ϫ 30 mL). The solvent was
[
5d,6]
supporting electrolyte. Nevertheless in the literature (refs.
)
2
an excess of current (1.5–2.0 F/mol) is employed to ensure the
complete reduction of the PB. We have chosen to use an excess
of PB and 4.0 mF to compare the efficiency of this EGB with
the others and, later on, to investigate the influence of PB con-
centration and current amount.
9]
further removed from the combined organic extracts under reduced
pressure. Column chromatography of the residues [with eluent mix-
tures of light petroleum/AcOEt (4:1) for substrates 5a,d and (3:2)
for 5b and 7a,b] allowed the separation of the products, whose iden-
tity was established by comparison with known standards.
[
F. G. Bordwell, Acc. Chem. Res. 1988, 21, 456–463. The au-
thors wish to thank one of the referees for this useful sugges-
tion.
[
[
10]
11]
M. A. Casadei, A. Gessner, A. Inesi, W. Jugelt, F. Micheletti
Moracci, J. Chem. Soc., Perkin Trans. 1 1992, 2001–2004.
T. W. Greene, P. G. M. Wuts, Protective Groups in Organic Syn-
thesis, 2nd ed., J. Wiley and Sons, Inc., NewYork 1991.
Received July 27, 1999
Acknowledgments
The authors wish to thank Mrs. Sonia Renzetti for typing the ma-
nuscript. Financial support from CNR and Universit a` La
Sapienza., Roma, Italy are gratefully acknowledged.
[O99464]
2448
Eur. J. Org. Chem. 2000, 2445Ϫ2448