Paired Electrosynthesis of Cyanoacetic Acid

Article abstract of DOI:10.1021/jo0358473

Cyanoacetic acid is formed by cathodic reduction of CO₂ and anodic oxidation of the tetraalkylammonium salt anion; the process is conduced in acetonitrile using a divided cell with a medium porosity glass-frit diaphragm. A mechanism for this paired electrochemical reaction is proposed.

Full text of DOI:10.1021/jo0358473

P a ir ed Electr osyn th esis of Cya n oa cetic Acid
Belen Batanero and Fructuoso Barba*
Department of Organic Chemistry, University of Alcal a´ , 28871 Alcal a´ de Henares, Madrid, Spain
Carlos M. S a´ nchez-S a´ nchez and Antonio Aldaz
Department of Physical Chemistry, University of Alicante, 03080 Alicante, Spain
fructuoso.barba@uah.es
Received December 19, 2003
2
Cyanoacetic acid is formed by cathodic reduction of CO and anodic oxidation of the tetraalkylam-
monium salt anion; the process is conduced in acetonitrile using a divided cell with a medium
porosity glass-frit diaphragm. A mechanism for this paired electrochemical reaction is proposed.
In tr od u ction
methods. Kawamata et al.13 describe its formation by
reaction between acetonitrile and alkaline phenoxides
using DMF as solvent. More recently, Gennaro et al.
The synthesis of cyanoacetic acid (CA) and its deriva-
tives is of much interest due to their industrial impor-
tance as starting materials. CA is mainly used for the
15
have prepared it by cathodic reduction of chloroacetoni-
trile in an undivided cell with a sacrificed aluminum
anode. Several syntheses of CA have already been
performed by different electrochemically (e.g., the anodic
1
2
production of the fungicide Cymoxanil, the cough remedy
3
4
Dextromethorphan, and antivirals such as Acyclovir,
5
6
Ganciclovir, or Famciclovir. Cyanoacetates have impor-
tant applications in the field of adhesives and are also
used for the synthesis of pharmaceuticals (such as the
16
17
oxidation of cyanoethanol or â-cyanoethyl ethers). In
these instances, large amounts of HCN were obtained,
thereby rending these syntheses inconvenient and dan-
7
antiepileptic valproic acid ), agrochemicals (preparation
of the rice herbicide Pyrazosulfuron-ethyl ), dyes (C.I.
Disperse Yellow 31 (Celliton Fast Yellow 7G )), and UV
absorbers (Etocrylene ).
gerous.
8
Tyssee published a patent14 in 1976 describing a
9
method to introduce a carboxylate group onto electro-
1
0
-
generated CH
2
CN by cathodic reduction of acetonitrile.
The conventional synthesis of CA is carried out by the
reaction of chloroacetic acid and alkaline cyanides. It is
a dangerous and environmentally unfriendly process that
should be replaced by a “green process”. For these
reasons, a large number of patents have described its
preparation using less dangerous reactants.11-14 One of
The current yield was only 5.8 g of CA/faraday. In the
patent, the nature of the electrodes is described: mer-
cury, cadmium, carbon, and lead were used as cathodes,
while the nature of the anode was irrelevant to the
electrochemical process. The electrolyses were performed
in a two-compartment cell with a medium-porosity glass-
frit diaphragm and tetraalkylammonium salts as sup-
porting electrolyte. However, as we demonstrate in the
present work, the electrosynthesis of CA does not take
place in the cathodic compartment, but takes place in
the anodic compartment through a paired reaction that
is described here for the first time. We show that the
presence of CA in the catholyte is due to its diffusion from
the anolyte.
the more interesting pathways involves the generation
-
of CH
2
CN, which is added to carbon dioxide. This anion
can be obtained by either chemical or electrochemical
*
To whom correspondence should be addressed. Tel: 34 91 8854617.
Fax: 34 91 8854686.
1) Hildbrand, S.; Pollak, P. Malonic acid and derivatives. In
(
Ullmann’s Encyclopedia of Industrial Chemistry, 6th ed. [Online]
Wiley-VCH: New York, 2002.
(2) Davidson, S. H. E.I. du Pont de Nemours and Co. US Pat. 3,-
9
54,992, 1973.
(3) Schnider, O.; Gruessner, A. Helv. Chim. Acta 1951, 34, 2211.
Gen er a l Electr och em ica l P r oced u r e
(4) Schaefer, H. J . Wellcome Foundation Ltd. DE 2539963, 1976.
Matsumoto et al. Chem. Pharm. Bull. 1988, 36, 1153.
5) Verheyden, J . P.; Martin, J . C. Syntex, Inc. US Pat. 4,355,032,
982; US Pat. 4,423,050, 1983.
Volta m m etr ic Exp er im en ts. Voltammetric curves
with IR drop compensation were obtained with an Amel
2053 potentiostat, an EGC 175 Universal Programmer,
and a Kipp-Zonen XY recorder. A three-electrode cell was
employed with a Pt bead (12.9 mm ) as working electrode
and a coiled wire Pt as counter electrode. As reference, a
(
1
(
(
(
(
(
(
6) Kincey, P. M. EP 355986, 1990.
7) Chignac, M. et al. Labaz US Pat. 4,155,929, 1979.
8) Yamamoto, S. et al. Nipon Noyakv Gakkaishi. 1996, 21, 259.
9) Hafenrichter, S. Textil-Praxis. 1961, 16, 273.
10) Holderbaum et al. BASF AG. DE 19519894, 1996.
11) Bulls, A. R.; Fellmann, J . D.; Periana, R. A. PCT Pat WO
2
9
1
212962, 1992.
(
12) Kaneko, H.; Masuda, T. J pn Pat. 07233135, 1995.
(15) Isse, A. A.; Gennaro, A. J . Electrochem. Soc. 2002, 149, D113-
117.
(16) Wermeckes, B.; Beck, F. Chem. Ber. 1985, 118, 3771.
(17) Wermeckes, B.; Beck, F. Electrochim. Acta 1985, 30, 1491.
(13) Kawamata, M.; Takahashi, T.; Fujikake, S. J pn Pat. 7571627,
1
0.1021/jo0358473 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/03/2004
J . Org. Chem. 2004, 69, 2423-2426
2423

Batanero et al.
Ag/AgCl electrode connected to the cell solution through
a Luggin capillary was employed. No correction for the
junction potential was made. The solution was deaerated
2
by N bubbling for 20 min. To avoid any concentration
change in the cell solution, the gas was previously
saturated by bubbling it through a similar solution. The
sweep rate was 50mV/s.
Electr och em ica l Syn th esis. The electrolyses were
carried out using both potentiostatic and constant current
conditions in a concentric cell with two compartments
separated by a porous (D3) glass-frit diaphragm and
equipped with a magnetic stirrer. As cathode, platinum
or lead plates were used, and as anode, a platinum net
or carbon plates were employed. As reference, a Ag/AgCl-
(satd) electrode was used. The solvent-supporting elec-
trolyte system (SSE) was a nominally anhydrous 0.05 M
solution of the different electrolytes: tetraalkylammo-
nium tetrafluorborate, perchlorate, bisulfate, chloride, or
bromide and lithium perchlorate in anhydrous acetoni-
trile. The volumes of catholyte and anolyte were 15 and
F IGURE 1. Voltammetric curves of the different SSE solu-
tions. Only the positive potential going scan is shown. Sol-
vent: acetonitrile. Supporting electrolyte concentration: 0.15
M. Key: (a) Bu
e) Et NClO , (f) Bu
electrode: Pt.
4
NBr, (b) Et
4 4 4 4
NCl, (c) Bu NHSO , (d) LiClO ,
NBF . Scan rate: 50 mV/s. Working
(
4
4
4
4
4
5 mL, respectively.
For controlled potential electrolyses the catholyte was
an SSE solution saturated with CO
bubbling of this gas, and the anolyte was a CO
2
by continuous
-free SSE
2
solution. The applied potential depended on the nature
of the employed SSE. The initial current was 200 mA
and remained almost constant during the electrolysis.
The reaction was finished after a given charge was
circulated through the cell. After that, the solvent in the
anolyte chamber was removed under reduced pressure.
The residue was extracted with ether/water (when water-
soluble electrolytes were used), and the organic phase
was dried over Na
When tetrabutylammonium tetrafluorborate was used as
electrolyte the residue was extracted with CHCl /NaOH
5%), and the aqueous phase was neutralized, further
concentrated, and extracted with ether. Finally, the ether
solution was dried over Na SO and then removed by
evaporation. A portion of this residue was treated with
methanol and some drops of sulfuric acid in order to
identify by gas chromatography (GC) the acid formed, as
the corresponding methyl ester.
2 4
SO and concentrated by evaporation.
3
(
2
4
F IGURE 2. Voltammetric curves of SSE: acetonitrile/Bu
NBF (0.15 M). Solid trace: CO -saturated SSE solution, dash
trace: CO -free SSE solution. Scan rate: 50 mV/s. Working
4
-
4
2
2
electrode: Pt.
For electrolyses carried out under constant current, the
same experimental conditions were employed; a divided
cell with platinum net as anode was used but a constant
current of 200 mA (current density of 15 mA/cm ) was
applied.
TABLE 1. Yield s (g/F of CA) a n d Cu r r en t Efficien cies
Obta in ed u n d er Differ en t Exp er im en ta l Con d ition s
no. of Faradays
current
2
anodic potential passed through
yield efficiency
+
electrolyte
(V, vs Ag/Ag )
the cell
(g/F)
(%)
Bu
Bu
4
4
NBF
NBF
4
4
+3.22
+3.2
+2.92
a
+1.67
+0.82
0.021
0.07
0.021
0.021
0.021
0.021
21
24.7
24.3
22.7
24.1
4.6
20.7
19.6
20.7
4
Resu lts
Et4
NClO
NHSO
Et NCl
4
Bu
4
4
Volta m m etr ic Exp er im en ts. Figure 1 shows the
voltammetric curves obtained for the different SSE
4
Bu
4
NBr
traces
systems employed: Bu
Et NClO , and Bu NBF
The increase of the current corresponds to the oxida-
4
NBr, Et
4
.
4
NCl, Bu
4
NHSO
4
, LiClO
4
,
a
Electrolysis under constant current conditions, I ) 200 mA.
4
4
4
tion of the anion of the supporting electrolyte. Figure 2
shows the voltammetric curves, corresponding to a nega-
tive potential direction scan, in the presence or absence
(current efficiency 24.7%) was isolated from the anodic
compartment after 0.021 F was circulated through the
cell. An increase of the circulated charge to 0.07 F did
not change the current efficiency of the process (see Table
1).
of CO
2 2
. Notice that the foot potential for reduction of CO
occurs before reduction of acetonitrile starts.
Electr olysis Exp er im en ts. For electrolyses carried
out at a controlled anodic potential of +3.22 V with a
When Et
trolled potential of +2.92 V was applied to get 200 mA
of current), similar results to those with Bu NBF were
obtained. Using Et NCl as the supporting electrolyte and
4 4
NClO was used as the electrolyte (a con-
platinum net anode and using Bu
4
NBF
4
as electrolyte,
4
4
4
35 mg of CA, which corresponds to 21 g of CA/faraday
4
2
424 J . Org. Chem., Vol. 69, No. 7, 2004

Paired Electrosynthesis of Cyanoacetic Acid
carrying out the electrolysis at +1.67 V led to a signifi-
cant decrease of the current efficiency and only 4 g of
CA/faraday was obtained. If the electrolysis was carried
It is known that in the first electron transfer for the
2
0-22
reduction of CO
ever, it has also been published that the CO
2
anion, formed in the CO cathodic reduction in acetoni-
2
a radical anion is obtained.
How-
2
-radical
out at +0.82 V using Bu
4
NBr as the supporting electro-
2
3
lyte, only traces of CA were found. In this case, dibro-
moacetonitrile was obtained as the major product (MS
trile/tetraalkylammonium salts, evolves to formate. For
this reason an electrolysis was carried out in absence of
+
+
m/e (relative intensity) EI: 201 (M + 4, 24), 199 (M +
2
8
2
CO , but in the presence of tetraethylammonium formate
+
, 48), 197 (M , 26), 120 (100), 118 (94), 93 (33), 91 (31),
1 (69), 79 (69)); a small amount of bromoacetonitrile was
dropped into the anodic compartment. In this case, no
CA was detected.
Another possibility described in the literature24 in-
volves the formation of oxalate in the catholyte through
also found.
Due to the possible industrial interest in this paired
reaction, the electrolysis was also carried out using one
dimerization of the electrogenerated CO -radical anion.
2
of the cheapest tetraalkylammonium salts, Bu
as the electrolyte. The results were similar to those
obtained with Bu NBF (see Table 1).
4
NHSO
4
,
In our case, small amounts of oxalic acid were found in
the catholyte. To eliminate the oxalate anion as the
species that migrates to the anolyte, electrolysis in
4
4
The type of separator employed in the divided cell has
a strong influence upon the CA yield. Thus, when a
cation-exchange membrane (Nafion) was used in place
of a glass-frit diaphragm, CA was not formed.
If the electrolysis is carried out by bubbling CO into
2
the anolyte, not the catholyte chamber, then CA is not
obtained.
absence of CO , but in the presence of added tetraethyl-
ammonium oxalate, dropped into the anodic compart-
2
ment, was carried out. However, no traces of CA were
detected.
It has been postulated24,25 that the CO -radical anion
2
reacts with a CO2 molecule to afford the -OCOOCO
•
radical anion dimer. Recently, it has also been postu-
lated26 that two CO -radical anions can react by a head-
The electrolysis has also been performed in an undi-
2
-
-
vided cell. Thus, using Bu
4
NBF
4
as electrolyte under the
to-tail reaction affording the dianion OCOOCO . How-
ever, the existence of a negative charge on a carbonyl
carbon seems highly unlikely, and this species should
same experimental conditions as described above, the
yield on CA decreased to 2.4 g/faraday.
Results similar to those obtained at constant potential
were obtained for electrolysis carried out at constant
current, even if the nature of the cathode was changed.
immediately evolve to CO and CO32 , as described.
-
24
In conclusion, the nature of the migrating molecule to
the anodic compartment should be either a CO -radical
2
•
For example, a constant current electrolysis (current
-
anion or the higher stabilized dimer OCOOCO radical
anion.
2
density: 15 mA/cm ) using Bu
4
NBF
4
as supporting
electrolyte in a divided cell and with Pb as cathode
afforded similar results than those obtained at controlled
potential. It was also demonstrated that when the
concentration of the supporting electrolyte drops near
during the electrolysis, the current also drops to zero.
(
3) What is the nature of the anodically generated
species that reacts with the migrating “CO -compound”?
2
It is well-known that acetonitrile is a very stable
molecule with a wide potential window, in both the
positive and the negative potential regions, and it is also
very well-known that the positive potential limit for the
SSE is fixed by the anodic oxidation of the anion of the
supporting electrolyte.27
As we have noted, the applied potential for the
synthesis of CA depends on the supporting electrolyte
employed and it corresponds to the oxidation potential
of the anion of each electrolyte.
Discu ssion
Before suggesting a mechanism it is necessary to
clarify the following questions:
(1) In which compartment does the formation of CA
take place?
2
The fact that CA was not obtained in a CO -saturated
catholyte when a cation exchange membrane was used
clearly points out that CA is not generated by cathodic
Cauquis et al. had demonstrated28 that the anodic
oxidation of supporting electrolyte anions leads to radical
intermediates. For instance, oxidation of tetrafluorbo-
rates at a platinum anode in acetonitrile yield the
reduction of acetonitrile and further attack of the anion
-
2 2
CH CN upon CO . Moreover, under our experimental
adduct29 CH
CN.BF . Its formation can be explained by
3 3
conditions acetonitrile is not reduced at the cathode,
because 3-aminocrotonitrile tetraalkylammonium salts
were not detected in the catholyte. The literature18,19
indicates that these compounds should be formed if
reduction of acetonitrile occurred. Thus, it is highlighted
that CA is formed only in the anolyte and that the CA
detected in the catholyte is only the result of its diffusion
from the anolyte.
(20) Desilvestro, J .; Pons, S. J . Electroanal. Chem. 1989, 267, 207.
(
(
21) Amatore, C.; Saveant, J .-M. J . Am. Chem. Soc. 1981, 103, 5021.
22) Gressin, J .-C.; Michelet, D.; Nadjo, L.; Saveant, J .-M. Nouv. J .
Chim. 1979, 3, 545.
23) Kushi, Y.; Nagao, H.; Nishioka, T.; Isobe, K.; Tanaka, K. Chem.
Lett. 1994, 2175.
24) Keene, R. R. In Electrochemical and Electrocatalytic Reactions
(
(
of Carbon Dioxide; Sullivan, R. B. P., Krist, K., Guard, H. E., Eds.;
Elsevier: New York, 1993; Chapter 1, p 9.
(
2
2) What is the nature of this migrating “CO -
(25) Nielsen, M. F.; Utley, J . H. P. In Organic Electrochemistry;
compound”? It is obvious that the formation of 1 equiv
of CA in the anolyte requires the transport of 1 equiv of
CO -reduction product from the catholyte to the anolyte.
2
Lund, H., Hammerich, O., Eds.; Marcel Dekker, Inc.: New York, 2001;
Chapter 21, p 831.
(
26) Flyunt, R.; Schuchmann, M. N.; von Sonntag, C. Chem. Eur.
J . 2001, 7, 796.
27) Lund, H. In Organic Electrochemistry; Lund, H., Hammerich,
(
(
18) Foley, J . K.; Korzeniewski, C.; Pons, S. Can J . Chem. 1988, 66,
O., Eds.; Marcel Dekker, Inc.: New York, 2001; Chapter 5, p 263.
(28) Cauquis, G.; Serve, D. J . Electroanal. Chem. 1970, 27, 3-6.
(29) Lund, H. In Organic Electrochemistry; Lund, H., Hammerich,
O., Eds.; Marcel Dekker, Inc.: New York, 2001; Chapter 5, p 264.
2
2
01.
(19) Batanero, B.; Barba, F.; Mart ´ı n, A. J . Org. Chem. 2002, 67,
369.
J . Org. Chem, Vol. 69, No. 7, 2004 2425

Batanero et al.
SCHEME 1
SCHEME 4
The reason CA is obtained, although in a small
amount, using Et NCl, but only traces of this acid are
obtained when Bu NBr is used can be explained consid-
ering that the abstraction of H by Cl radical is more
SCHEME 2
4
4
•
•
•
32
exothermic than that of Br , affording in this way larger
•
•
2 2
amount of CH CN. These CH CN radicals can react
either with the “reduced” form of CO
CN or with electrogenerated Cl to form ClCH
2
to give HOOCCH
2
-
•
2
CN. It is
also known that the polyhalogen substitution almost
invariably occurs, as side process, even if there is a large
molar ratio of substrate to halogen.33 However, when R
NBr is used the formation of dibromoacetonitrile is the
main process. A plausible explanation is that the Br
-
4
SCHEME 3
•
radical reacts with acetonitrile, as we have previously
pointed out, to give BrCH
on the concentration of CH
2
•
CN. This fact causes a decrease
2
CN, to the level of traces,
decreasing the rate of the reaction between this radical
and the migrating “reduced CO ”. Then, the “reduced
CO
” acts as an electrogenerated base23 versus the acidic
2
2
protons of BrCH
2
CN, affording a very stabilized anion
(formed by coupling of
-
(
BrCHCN) that reacts with Br
2
•
+
two Br ) or with Br (formed in the anodic oxidation of
decomposition of BF •
into BF
and the radical F . When
•
•
2
Br ) leading to Br
CHCN.
4
3
tetraalkylammonium chloride, bromide, or hydrogen
•
•
•
sulfate are used, obviously Cl , Br , or HSO
tively generated.
4
are respec-
Con clu d in g Rem a r k s
(
4) What is the final reaction between these electro-
Herein we describe an interesting alternative to the
synthesis of cyanoacetic acid through the generation of
generated species?
The anodically electrogenerated radical from the elec-
•
CH CN radical, obtained by anodic oxidation of electro-
2
trolyte anion abstracts a hydrogen atom of the solvent,
lyte anions and subsequent reaction with solvent mol-
•
•
acetonitrile, in an homogeneous reaction yielding CH
2
-
ecules (acetonitrile). These radicals CH CN are coupled
2
CN radicals.30 It should be noticed that the C-H bond
with cathodic CO -reduced species, coming through the
2
-
1 31
energy of H-CH
2
CN is 360 kJ mol
.
Reaction of the
medium-porosity glass-fritt diaphragm, to afford the
desired product.
later radical with a CO
2
-radical anion, or with the
corresponding dimer OCOOCO , and further protonation
-
•
We also show that the process could be of high
industrial interest and for this reason the reaction has
been carried out with the cheapest tetraalkylammonium
•
2
(see proposed mechanism) affords CA. The radical CH -
CN is formed through an homogeneous reaction in the
bulk of the solution.
Taking into account each of these points, we propose
4 4
salt Bu NHSO as the electrolyte system and under
constant current conditions, obtaining similar results
than those under controlled potential or employing
expensive electrolytes.
the mechanism shown in Schemes 1-3.
To account far the relatively moderate yield of CA
obtained in our paired reaction, a preparative-scale
anodic oxidation of CA was performed at constant
Ack n ow led gm en t. This work was supported by the
Spanish Ministry of Science and Technology through
projects BQU2001-1083 and BQU2000-0460. B.B. thanks
this Ministry for the “Ram o´ n y Cajal” contract, and C.S.
also thanks the University of Alicante for his FPU
scholarship.
potential (+3.22 vs Ag/AgCl), Pt anode, and CH
3 4
CN/Bu -
NBF as SSE. Oxidation of CA took place together with
4
the oxidation of the anion of the supporting electrolyte.
However, the regeneration of CA in the paired reaction
2
(under cathodically bubbled CO ) is plausible by reaction
between the decarboxylated product and the reduced-CO
molecule (see Scheme 4).
2
J O0358473
(
30) Ohmori, H.; Ueda, C.; Yamagata, K.; Masui, M.; Sayo, H. J .
Chem. Soc., Perkin Trans. 2 1987, 1065.
31) Clayden, J .; Greeves, N.; Warren, S.; Wothers, P. Organic
Chemistry; Oxford University Press: New York, 2001; p 1041.
(32) Clayden, J .; Greeves, N.; Warren, S.; Wothers, P. Organic
Chemistry; Oxford University Press: New York, 2001; p 1038.
(33) Advanced Organic Chemistry, 5th ed.; Smith, M. B., March, J .,
Eds.; J ohn Wiley and Sons, Inc.: New York, 2001; p 907.
(
2
426 J . Org. Chem., Vol. 69, No. 7, 2004