Cyclic mechanism in the trapping of carbonyl oxides with alcohols and carboxylic acids

Article abstract of DOI:10.1016/j.tetlet.2004.01.158

The reaction of diphenylcarbonyl oxide with alcohols and carboxylic acids, which has been classified as a nucleophilic trapping, is shown to be in the reactivity order: AcOH?MeOHCF₃CH₂OHEtOH?t-BuOH. A laser-flash spectroscopy indicated that the reaction of carboxylic acids is very fast, that is, one-tenth of the diffusion rate. These results suggest that the hydroxyl compounds react as an acid and a nucleophile at the same time and the major reaction is via the seven- and five-membered cyclic mechanism for RCO₂H and ROH, respectively.

Full text of DOI:10.1016/j.tetlet.2004.01.158

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 2559–2561
Cyclic mechanism in the trapping of carbonyl oxides
with alcohols and carboxylic acids
Hiroshi Hanaki, Yuta Fukatsu, Masaki Harada and Yasuhiko Sawaki*
Department of Applied Chemistry, Graduate School of Engineering, Nagoya University, Chikusa-ku, Nagoya 464-8603, Japan
Received 5 January 2004; revised 28 January 2004; accepted 30 January 2004
Abstract—The reaction of diphenylcarbonyl oxide with alcohols and carboxylic acids, which has been classified as a nucleophilic
trapping, is shown to be in the reactivity order: AcOH ꢀ MeOH > CF3CH2OH > EtOH ꢀ t-BuOH. A laser-flash spectroscopy
indicated that the reaction of carboxylic acids is very fast, that is, one-tenth of the diffusion rate. These results suggest that the
hydroxyl compounds react as an acid and a nucleophile at the same time and the major reaction is via the seven- and five-membered
cyclic mechanism for RCO2H and ROH, respectively.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Carbonyl oxides and related species are known as a key
intermediate in the ozonolysis of olefins1 and have
attracted much attention in the various fields such as
synthetic, biological, mechanistic, and atmospheric
chemistry.2 Chemical property of the peroxidic species is
significantly dependent on their substituents.2a;3 For
example, an unique reactivity has been known for a
carbonyl oxide (CO) carrying a potent electron-with-
drawing4 or -donating group.5 One of recent interests is
a generation of hydroxyl radical from COÕs in gas
phase.6 Typical reactions of COÕs are the cycloaddition
to aldehydes and the trapping with alcohols and carb-
oxylic acids.1;2 The latter reaction with hydroxyl com-
pounds has been regarded as the evidence of CO
intermediates in the ozonolysis of olefins1 and classified
as the nucleophilic trapping.2b We became interested in
the trapping reaction and obtained unexpected results
that the trapping proceeds by way of a cyclic mecha-
nism, carboxylic acids being the most reactive as
described in the following.
benzophenone quantitatively, and the reaction in the
presence of ROH resulted in the formation of a-al-
koxyhydroperoxide (3),7b as shown in Eq. 1, as the
trapped product from diphenylcarbonyl oxide (2), for
example, 51% yield of 3 when R ¼ Me.
OOH
ROH
Sens/hν/O2
Ph2CN2
Ph2COO
Ph2C
ð1Þ
OR
1
2
3
It is apparent that 2 was trapped efficiently since the
selectivity for 2 from 1 and 1O2 is known to be 50–60%.8
Comparable yields (i.e., 40–65%) of 3 were resulted with
ethyl alcohol (R ¼ Et) and acetic acid (R ¼ Ac), but no
product was obtained from t-BuOH. Relative reactiviy
of these hydroxyl compounds toward 2 were determined
from the relative yields of 3 in their presence of 1:1 ratio
in MeCN.9 The resulting relative reactivity of AcOH–
MeOH–CF3CH2OH–EtOH–t-BuOH are 50:5.0:1.4:
1.0:<0.01; unexpectedly, acetic acid was shown to be the
most reactive in trapping 2. A different trapping ability
has been reported to be AcOH < t-BuOH < MeOH,10
but the order was estimated indirectly during an olefin
ozonolysis.
The reaction of diazo compounds with singlet oxygen is
well known to produce the corresponding COÕs.7 Thus,
the irradiation of diphenyldiazomethane (1) and Rose
Bengal in acetonitrile at >400 nm under oxygen afforded
In order to confirm whether carboxylic acids react with
COÕs facilely or not, the trapping reaction was studied
by means of laser flash spectrophotometry.11 Carbonyl
oxide 2 is known to decay slowly according to second-
order kinetics, yielding benzophenone ultimately.12
Keywords: Carbonyl oxides; Cycloadditions; Alcohols; Carboxylic
acids and derivatives; Peroxides; Mechanisms.
* Corresponding author at present address: Department of Applied
Chemistry, Aichi Institute of Technology, Toyota 470-0392, Japan.
Tel./fax: +81-561-74-0266; e-mail: ysawaki@apchem.nagoya-u.ac.jp
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.01.158

2560
H. Hanaki et al. / Tetrahedron Letters 45 (2004) 2559–2561
philic trappingÕ,2b the order of RCO2H ꢀ ROH clearly
4.0
3.0
suggests that the trapping is not controlled by their
nucleophilicity. Then pathway (a) or (b) is conceivable
as the mechanism of trapping with hydroxyl compounds
(Scheme 1). The assumption of hydrogen bonding be-
tween the outer oxygen of CO and hydroxyl compounds
is reasonable since the oxygen is of minus charge and
behaves as a nucleophilic oxidant.2b;3;7b The observed
trapping reactivity of RCO2H ꢀ MeOH > CF3CH2OH
> EtOH is not compatible with pathway (a) where the
product selectivity would be dependent on the nucleo-
philicity of RCO2H or ROH. Especially, the observed
orders of RCO2H ꢀ ROH and of CF3CH2OH >
EtOH15 clearly deny pathway (a) since RCO2H and
CF3CH2OH is known to be less nucleophilic.
•
2.0
1.0
0.0
•
•
•
0.0
0.5
1.0
1.5
2.0
2.5
[ HA ] / mM
Figure 1. Effect of concentrations of HA on the decay rate of Ph2COO
(2) in acetonitrile at 20 ꢁC: d, AcOH; m, PhCOOH.
On the other hand, a cyclic mechanism such as path (b)
may explain the trapping orders since the hydroxyl
compounds could react as an acid and a nucleophile at
the same time. The observed higher trapping reactivity
of RCO2H in comparison to alcohols could be explained
only by cyclic mechanism involving a hydrogen bonded
transition state such as 5. According to the cyclic
mechanism the trapping reactivity is governed by the
hydrogen bonding ability and partially by their nucleo-
philicity. No reaction of t-BuOH is explicable by the
steric hindrance between the bulky tert-butyl and two
phenyl groups in 3 or 5.
When acetic acid was added, CO 2 decayed rapidly
according to the first-order dependence on [2] (e.g., the
lifetime of 0.37 s in the presence of 1.0 mM AcOH). As
shown in Figure 1, the decay rate is proportional to the
acid concentrations, and the resulting second-order rate
constant for AcOH is k2 ¼ 1:61 ꢁ 109 Mꢂ1s ꢂ1 (vs [2] and
[AcOH]; see Fig. 1). The k2 values of 2 for benzoic and
m-chlorobenzoic acids are 1.56 and 1.40 · 109 Mꢂ1 sꢂ1
,
respectively, and that of (p-ClC6H4)2COO with AcOH
was 1.69 · 109 Mꢂ1 sꢂ1. Apparently, the reaction of 2
with carboxylic acids is as fast as one-tenth of the dif-
fusion rate and is independent of their acidity. For
comparison, the k2 value for the reaction of 2 and
Suitability of the cyclic transition state (TS) 5 is con-
sidered according to the semi-empirical AM1 method by
which the Criegee mechanism was successfully ana-
lyzed.16 According to the method17 a five-membered
cyclic TS 6 could be calculated for the addition of ROH
to parent H2COO (Eq. 2). Resulting TS 6 (R ¼ H) is of
an envelope type and similar to that for the addition of
H2C@O.16a
PhCHO was determined to be 5.6 · 107 Mꢂ1 sꢂ1 13
, which
is only 1/30 of the value with RCO2H.
How about the rate of reaction between COÕs with
alcohols? For the case of Ph(MeO2C)COO, a reactive
CO, the k2 value for MeOH is reported to be
3.4 · 106 Mꢂ1 sꢂ1 14
.
On the other hand, the reaction rate
of 2 with alcohols has not been reported because
the reaction is too slow for the accurate determination.
The k2 value of 2, however, could be estimated from the
initial decay in the presence of MeOH, affording the
approximate one of 2.7 · 104 Mꢂ1 sꢂ1. The estimation
seems to be acceptable since 2 is trapped efficiently by
MeOH.
O
H
H
O
O
H
H
O
C
C
O
H
H2COO + ROH
R
O
H
R
6
7
ð2Þ
These kinetic study on the trapping of 2 confirmed the
interesting fact that the reactivity order is
RCO2H ꢀ RCHO ꢀ ROH. Although the reactions with
hydroxyl compounds have been classified as Ônucleo-
The heat of addition of H2O to yield adduct 7 is cal-
culated to be highly exothermic (i.e., )59 kcal/mol)18 by
OOH
OR'
R'OH
(a)
Ph
Ph
O
ROH
Ph C = O − O - - - -
2
HOR
Ph2C
C = O
4
2
O
O
H
OOH
OR
Ph2C
R
Ph2C
(b)
O
5
Scheme 1. Trapping pathways of CO 2 with hydroxyl compounds.

H. Hanaki et al. / Tetrahedron Letters 45 (2004) 2559–2561
2561
2. (a) Sander, W. Angew. Chem., Int. Ed. Engl. 1990, 29, 344;
(b) Bunnelle, W. H. Chem. Rev. 1991, 91, 335; (c)
McCullough, K. J.; Nojima, M. In Organic Peroxides;
Ando, W., Ed.; Wiley: New York, 1992; p 661.
3. (a) Ishiguro, K.; Nojima, T.; Sawaki, Y. J. Phys. Org.
Chem. 1997, 10, 789; (b) Ishiguro, K.; Sawaki, Y. Bull.
Chem. Soc. Jpn. 2000, 73, 535.
4. Nojima, T.; Hirano, Y.; Ishiguro, K.; Sawaki, Y. J. Org.
Chem. 1997, 62, 2387.
5. (a) Kopecky, K. R.; Xie, X.; Molina J. Can. J. Chem.
1993, 71, 272; (b) Ishiguro, K.; Hirabayashi, K.; Nojima,
T.; Sawaki, Y. Chem. Lett. 2002, 796.
6. Gutbrod, R.; Kraka, E.; Schindler, R. N.; Cremer, D.
J. Am. Chem. Soc. 1997, 119, 7330.
7. (a) Higley, D. P.; Murray, R. W. J. Am. Chem. Soc. 1974,
96, 3330; (b) Sawaki, Y.; Kato, H.; Ogata, Y. J. Am.
Chem. Soc. 1981, 103, 3832.
the AM1 method, but the activation energy to form TS 6
(R ¼ H) is as high as 30.3 kcal/mol, indicating that the
reactions of alcohols is not so facile.19 The calculated
activation energy for the addition of MeOH (R ¼ Me,
Eq. 2) is 32.4 kcal/mol, but that of formic acid
(R ¼ HCO) is as high as 43.5 kcal/mol. Thus, five-
membered TS 6 cannot explain the observed fast reac-
tion of carboxylic acids. For the case of acids, however,
seven-membered TS 8 could be pictured.
H
O
C
O
O
H
H
O
C
H
8
8. Nojima, T.; Ishiguro, K.; Sawaki, Y. J Org. Chem. 1997,
62, 6911.
9. Acetonitrile solution of
1 (7.5 mM), Rose Bengal
(0.15 mM), ROH (0.615 M each) was irradiated for 1.5 h
In fact, such TS 8 was suggested by the AM1 method,
the activation energy being )48.0 (AM1) or )30.2 kcal/
mol (a DFT).20 The activation energy is predicted to be
barrierless, which is consistent with the observed fast
reaction of carboxylic acids. In order to examine the
incorporation of hydrogen bonding, the deuterium iso-
tope effect was determined in the reaction of 2 with
AcOH and AcOD. The resulting small value of
kH=kD ¼ 1.44 (MeCN, 25 ꢁC) is in a line with the
hydrogen-bonded cyclic TS 8 in the fast reaction close to
the diffusion rate.
at >400 nm under oxygen. Product yields were determined
1
by H NMR measurement. Ph2C(OR)OOH: 3a (R ¼ Me)
dH (CDCl3): 3.33 (s, 3H), 7.2–7.7 (m, 10H); 3b (R ¼ Et):
2.02 (t, 3H), 3.52 (q, 2H), 7.2–7.7 (m, 10H); 3c
(R ¼ CH2CF3): 3.84 (s, 2H), 7.2–7.7 (m, 10H); 3d
(R ¼ COMe): 2.16 (s, 3H), 6.8–7.5 (m, 10H).
10. Yamamoto, Y.; Niki, E.; Kamiya, Y. Bull. Chem. Soc.
Jpn. 1982, 55, 2677.
11. Nano-second spectroscopy was carried out as described
previously. (a) Ref. 8; (b) Yokoi, H.; Nakano, T.; Fujita,
W.; Ishiguro, K.; Sawaki, Y. J. Am. Chem. Soc. 1998, 120,
12453.
12. (a) Casal, H. L.; Sugamori, S. E.; Scaiano, J. C. J. Am.
Chem. Soc. 1984, 106, 7623; (b) Scaiano, J. C.; McGimp-
sey, W. G.; Casal, H. L. J. Org. Chem. 1989, 54, 1612.
13. The k2 value for MeCHO is reported to be
Finally, some should be noted on the apparent dis-
crepancy between the kinetic rate ratio and the com-
petitive trapping reactivity. The rate ratio for the
carboxylic acid and MeOH is 109:104, that is, the 105-
fold difference, while that from competitive trapping
experiments was 10:1, that is, only 10-fold. The dis-
crepancy suggests clearly that in the co-presence of
RCO2H and R0OH (see Scheme 1) the major reaction is
pathway (b) but pathway (a) operates as a minor one
(up to 10%).
3.1 · 106 Mꢂ1 sꢂ1 12b
.
14. Fujiwara, Y.; Tanimoto, Y.; Itoh, M.; Hirai, K.; Tomioka,
H. J. Am. Chem. Soc. 1987, 109, 1942.
15. (a) McClelland, R. A. In Organic Reactivity: Physical and
Biological Aspects; Golding, B. T., Griffin, R. J., Maskill,
H., Eds.; Royal Soc. Chem.: Cambridge, 1995; p 301; (b)
Kawamura, S.; Takeuchi, R.; Masuyama, A.; Mojima,
M.; McCullough, K. J. J. Org. Chem. 1998, 68, 5617.
16. (a) Dewar, M. J. S.; Hwang, J. C.; Kuhn, D. R. J. Am.
Chem. Soc. 1991, 113, 735; (b) Ponec, R.; Yuzhakov, G.;
Haas, Y.; Samuni, U. J. Org. Chem. 1997, 62, 2757.
17. Spartan package: MACSPARTAN Plus 2.0, Wavefunction,
Inc., Irvine, CA 92612.
18. Calculated heat of reaction of Eq. 2 are )59 and )52 kcal/
mol for R ¼ H and HCO, respectively. The high heat of
reaction seems, when the reaction of COÕs with water or
carboxylic acid occurs in gas phase, to indicate the
spontaneous homolysis of weak O–O bond in 7 leading
to the formation of hydroxyl radical.
In conclusion, the present paper on the reaction of COÕs
revealed experimentally that the addition of carboxylic
acids proceeds very fast via seven-membered TS while
that of alcohols by way of five-membered TS. So-called
nucleophilic trapping of COÕs is to be classified as the
trapping via five- or seven-membered cyclic mechanism.
References and notes
19. In contrast, the cycloaddition of H2C@O to H2COO was
shown to have a low activation energy of below 2 kcal/
mol.16a
20. A DFT calculation: pBP/DN**(Spartan)/AM1.17
1. Bailey, P. S. In Ozonation in Organic Chemistry; Aca-
demic: New York, 1978; Vol. 1, 1982; Vol. 2.