On the mechanism of the Elbs peroxydisulfate oxidation and a new peroxide rearrangement

Article abstract of DOI:10.1016/S0040-4039(02)00495-1

Semiempirical calculations show that the intermediate formed by reaction between peroxydisulfate anion and the phenolate ion (the Elbs oxidation) is the species resulting from reaction of the tautomeric carbanion rather than the one resulting from attack by the oxyanion. This is confirmed by synthesis of the latter intermediate by reaction between Caro's acid dianion (peroxymonosulfate) and some nitro-substituted fluorobenzenes. This intermediate rearranges preferentially to give the phenol ortho-sulfate rather than the para-sulfate characteristic of the Elbs oxidation. This reaction of an arylperoxysulfate to give a phenol ortho-sulfate is a new rearrangement.

Full text of DOI:10.1016/S0040-4039(02)00495-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 3221–3224
On the mechanism of the Elbs peroxydisulfate oxidation and a
new peroxide rearrangement
a
b
b,
Elizabeth C. Behrman, Ssuhen Chen and Edward J. Behrman *
a
Department of Physics, Wichita State University, Wichita, KS 67260, USA
b
Department of Biochemistry, The Ohio State University, 484 West 12th Avenue, Columbus, OH 43210, USA
Received 16 January 2002; accepted 6 March 2002
Abstract—Semiempirical calculations show that the intermediate formed by reaction between peroxydisulfate anion and the
phenolate ion (the Elbs oxidation) is the species resulting from reaction of the tautomeric carbanion rather than the one resulting
from attack by the oxyanion. This is confirmed by synthesis of the latter intermediate by reaction between Caro’s acid dianion
(
peroxymonosulfate) and some nitro-substituted fluorobenzenes. This intermediate rearranges preferentially to give the phenol
ortho-sulfate rather than the para-sulfate characteristic of the Elbs oxidation. This reaction of an arylperoxysulfate to give a
phenol ortho-sulfate is a new rearrangement. © 2002 Elsevier Science Ltd. All rights reserved.
The Elbs peroxydisulfate oxidation is the reaction of
phenolate anions with peroxydisulfate ions to form a
mixture of the ortho- and para-sulfates of the parent
We have approached this problem in two ways: the first
is by calculation of the energies of both possible species
and the second by synthesis of one of these
intermediates.
1
phenol. The para product predominates in contrast to
the situation in the related Boyland–Sims oxidation of
aromatic amines in which the ortho product is favored.
A mechanistic point which has been in doubt for some
time is whether initial attack is at carbon or at oxygen
followed by rearrangement (Fig. 1). In the case of the
Boyland–Sims oxidation, the evidence favors attack at
We carried out calculations at the semi-empirical level
(MOPAC, AM1) of the heats of formation of the two
possible intermediates, I and II, resulting from attack at
oxygen or carbon. We find a difference of about 65 kcal
2
−
1
mol in favor of II. These are gas phase calculations,
−1
nitrogen followed by rearrangement to the ortho
but the estimated solvation energies (5–10 kcal mol )
are not likely to alter this preference. This appears to
3
product.
O
O-
S O -
A
2
8 2
B
H
O
O
OOSO3-
H
OSO3-
+
H
OSO3-
II
O-
I
III
OSO3-
Figure 1.
Keywords: phenol; peroxide; rearrangement; Elbs oxidation; Caro’s acid.
*
Corresponding author. Tel.: 614-292-9485; fax: 614-292-6773; e-mail: behrman.1@osu.edu
0
040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00495-1

3
222
E. C. Behrman et al. / Tetrahedron Letters 43 (2002) 3221–3224
establish the route as direct attack at carbon rather
than at oxygen followed by rearrangement, i.e. pathway
B. Further calculations give a comparison between the
para and ortho intermediates (i.e. between II and III).
Following a bulk correction for solvation, we find that
to find rearrangement products ortho and para to the
fluoro substituent. These isomers offer, respectively, an
ortho, both an ortho and a para, and a para position for
substitution. 2,6-Dinitrofluorobenzene reacted with
Caro’s acid at pH 7 to give 2,6-dinitrohydroquinone-4-
sulfate in about 25% yield (Fig. 2). Its identity was
−
1
the para-isomer is favored by about 3 kcal mol in
reasonable agreement with the observed para–ortho
1
established using H NMR by comparison with known
−
1 1
ratio of about 10 (1.5 kcal mol ).
material prepared by Elbs oxidation of 2,6-dinitrophe-
nol at pH 7 and confirmed by hydrolysis to 2,6-dinitro-
hydroquinone. However, when given a choice between
the ortho and para positions, the ortho product is
formed exclusively as shown by the case of 2,5-dini-
trofluorobenzene. Reaction of 2,5-dinitrofluorobenzene
with Caro’s acid at pH 8 gave the ortho sulfate (3,6-
dinitropyrocatechol monosulfate) as the major product
in about 20% yield with no trace of the para sulfate
4
The synthetic route relies on previous work which
suggested that Caro’s acid dianion, an alpha nucleo-
phile, ought to react with an appropriately substituted
fluorobenzene to form an intermediate of type I, path-
way A (Fig. 1). A direct synthesis of an intermediate of
type I was carried out according to Scheme 1. The
expectation was that IV would rapidly expel a fluoride
ion leading to a type I intermediate. Analysis would
then reveal if I gave rise to the normal products of the
Elbs oxidation or not.
†
(Fig. 2).
Since about 70% of the starting material was recovered,
the conversion was about 65%. The ortho sulfate was
identified by its characteristic pair of doublets, l=7.60
and 6.45, J=9 Hz, in the proton NMR spectrum of
We examined the reaction of Caro’s acid with three
dinitrofluorobenzenes (2,4; 2,5; 2,6) since we expected
F
O
S
-O
O
+
O-
O
R
F
OOSO3-
IV
R
OOSO3-
F
O-
F-
I
V
+
SO4
R
R
O-
F
O-
OSO3-
H
+
SO4
R
R
F-
O-
OSO3-
+
H+
R
Scheme 1.
†
A referee raised the question of why any product is formed in the 2,6-dinitro case. One possibility is that intermediate IV can undergo
intramolecular rearrangement to either the ortho- or para-position but that the ortho-rearrangement is much faster.

E. C. Behrman et al. / Tetrahedron Letters 43 (2002) 3221–3224
3223
F
OOSO3-
O-
O2N
NO2
O2N
NO2
O2N
NO2
SO5-2
OSO3-
F
OOSO3-
NO2
O-
NO2
-O3SO
O2N
NO2
SO5-2
O2N
O2N
F
OOSO3-
NO2
O-
NO2
-O3SO
NO2
SO5-2
NO2
NO2
NO2
Figure 2.
reaction mixtures. These were identical to those seen in
controlled reactions of chlorosulfonic acid with 3,6-
dinitropyrocatechol and also identical to those found in
small amounts in reaction mixtures of peroxydisulfate
with 2,5-dinitrophenol. Hydrolysis of the ortho sulfate
gave 3,6-dinitropyrocatechol. This material showed a
singlet within 0.1 ppm of the resonance of 2,5-dinitro-
hydroquinone (the para isomer), but was easily distin-
guished by TLC on silica. The pyrocatechol gives a
intermediates of type I could be challenged since we
realized that intermediates of type IV have at least two
pathways open to them for formation of the observed
products. Scheme 1 shows these possibilities. It is a
question of whether fluoride ion expulsion is faster than
cleavage of the peroxide bond followed by rearrange-
ment or vice versa. We therefore calculated the heats of
formation for the two intermediates, I and V. Interme-
−1
diate I is about 158 kcal mol (R=H) and 74 kcal
reddish spot with R =0.1 using dichloromethane:
−1
f
mol (R=2,5-dinitro) more stable than intermediate V.
methanol, 10:1; the hydroquinone gives a yellow spot
Application of the Hammond principle to this situation
suggests that decomposition of IV proceeds first by
expulsion of the fluoride ion and then rearrangement of
the peroxide.
with R =0.95. The reaction of peroxydisulfate in an
f
Elbs oxidation of 2,5-dinitrophenol at pH 7 gives, by
contrast, a mixture in which the hydroquinone sulfate
predominates; the ratio of the hydroquinone to the
pyrocatechol sulfate is about 90 (NMR integration).
We tested our calculational approach using an
6
analogous case from the literature in which t-butylhy-
2
,4-Dinitrofluorobenzene (Sanger’s reagent) reacted
droperoxide anion is known to react with 4-methyl-6-
chloropyrimidine to yield a stable pyrimidinyl peroxide
with Caro’s acid to give, following hydrolysis, 3,5-dini-
tropyrocatechol with melting point and IR spectrum
identical to material prepared via nitration of pyrocate-
that undergoes
a
slow rearrangement to
a
t-
5
chol diacetate.
butoxypyrimidone (Fig. 3). Calculation of the energies
of the two possible intermediates in this process involv-
ing either expulsion of the halide first (this is what
occurs) or peroxide cleavage first gives a result in
Our assumption that intermediates of type IV necessar-
ily decompose by expulsion of the fluoride ion to form
CH3
N
CH3
CH3
OO-
Fast
O
N
N
N
Slow
Cl
N
O
O
N
H
O
Figure 3.

3
224
E. C. Behrman et al. / Tetrahedron Letters 43 (2002) 3221–3224
agreement with experiment: halide cleavage is favored
by about 16 kcal mol . This result lends confidence to
the calculational conclusions.
the solution was boiled until the volume was about 100
−
1
ml. Glistening yellow crystals formed at 5°C after a few
5
days, 0.96–1.04 g, 53–56%, mp 168–170°C (lit. 166–
1
66.5°C). IR (Nujol): 3585, 3513, 1600, 1552, 1514,
There are some additional analogous cases in the litera-
ture involving other peroxides, but why a particular
1342, 1259, 1212, 1151, 1080, 994, 938, 888, 823, 811,
792, 77 737, 680, 656 cm . H NMR: (DMSO-d ) l
8.22, 7.77 (d, J=2.8 Hz).
−
1
1
6
6
reaction leads to rearrangement or simply to cleavage
7
of the peroxide is unclear at this time. It is particularly
notable that cumyl hydroperoxide sometimes shows
The intermediate sulfate ester was isolated in about
50% yield by allowing the reaction mixture of Caro’s
acid and Sanger’s reagent to stand at 5°C for 2 days
following the rt reaction (see above). Orange crystals of
the product were filtered, dried, washed with
dichloromethane, and recrystallized from isopropanol/
water (2:1). Fine yellow needles, mp 209–211°C, calcd
for C H N Na O S·H O: C, 21.05; H, 1.17; N, 8.19.
6
rearrangement, but gives rise only to 2,4-dinitrophenol
8
in the reaction with Sanger’s reagent.
Experimental
NMR was at 600 MHz. 2,6-Dinitrofluorobenzene was
prepared by a modification of the method of Parker
6
2
2
2
9
2
9
and Read. 2,5-Dinitrofluorobenzene, a new com-
Found: C, 20.8; H, 0.79; N, 7.87. IR (Nujol): 3625,
3441, 3105, 3056, 1668, 1618, 1598, 1573, 1548, 1482,
1433, 1363, 1339, 1321, 1283, 1260, 1237, 1200, 1169,
1080, 1052, 979, 933, 917, 834, 811, 799, 750, 728, 711,
pound, was prepared by dimethyldioxirane oxidation of
1
0
2
7
1
-fluoro-4-nitroaniline. Crystals from hexane, mp 74–
5°C, calcd for C H FN O : C, 38.72; H, 1.62; N,
6
3
2
4
−
1 1
5.05. Found: C, 39.07; H, 1.70; N, 14.69. 3,6-Dini-
685, 642, 588, 542 cm . H NMR (D O): l 8.80, 8.18
2
tropyrocatechol was prepared from the dinitro-
(d, J=3.1 Hz).
1
1
12
guaiacol by treatment with HBr. Elbs oxidation of
,5-dinitrophenol was carried out under standard con-
2
ditions using the ammonium salts of the reactants. Acid
hydrolysis and cooling gave crystals of 2,5-dinitrohy-
droquinone, mp 203–205°C (lit. 201–203°C).
References
13
1
. Behrman, E. J. Org. React. (N.Y.) 1988, 35, 421–511.
Reaction of Caro’s acid (12 mg Oxone, 8 mg NaDCO3)
with 2,6-dinitrofluorobenzene (3 mg) was carried out in
2. Behrman, E. J. J. Am. Chem. Soc. 1963, 85, 3478–3482.
3. Behrman, E. J. J. Org. Chem. 1992, 57, 2266–2270.
4. (a) McIsaac, J. E.; Subbaraman, L. R.; Subbaraman, J.;
Mulhausen, H. A.; Behrman, E. J. J. Org. Chem. 1972,
37, 1037–1041; (b) See also: Ritchie, C. D.; Sawada, M. J.
Am. Chem. Soc. 1977, 99, 3754–3761.
5. Heertjes, P. M.; Knape, A. A.; Talsma, H. J. Chem. Soc.
1954, 1868–1870.
6. Kropf, H.; Ball, M. Liebig’s Ann. Chem. 1976, 2331–2338.
D O, 0.2 M phosphate pD7 (1 ml), with stirring for 3 h
2
at rt. The reaction of Caro’s acid (12 mg Oxone) with
2
,5-dinitrofluorobenzene (6 mg) was carried out in a 7:3
mixture of 0.2 M phosphate, D O, pD 8: CD OD (5
2
3
ml). A trace of EDTA·2D O was added to both reac-
2
tions to decrease the autodecomposition of Caro’s acid.
Solid K CO was added as needed to maintain the pH.
2
3
7
. (a) Kropf, H. Houben-Weyl Meth. Org. Chem. 1988, 13,
762–763; (b) Heller, R. A.; Weiler, R. Can. J. Chem.
1987, 65, 251–255.
3
,5-Dinitropyrocatechol and 3,5-dinitropyrocatechol-1-
O-sulfonate, sodium salt
8. Kropf, H.; Ball, M.; Siegfriedt, K.-H.; Wagner, S. J.
2
,4-Dinitrofluorobenzene (1.86 g, 0.01 mol), Caro’s acid
Chem. Res. (M) 1981, 4001–4015 see p. 4007.
9. Parker, R. E.; Read, T. O. J. Chem. Soc. 1962, 3149–
3153.
10. Webster, B. M.; Verkade, P. E. Recl. Trav. Chim. Pays-
Bas 1949, 68, 77–87.
11. Oxford, A. E. J. Chem. Soc. 1926, 2004–2011.
12. Heertjes, P. M.; Nijman-Knape, A. A.; Talsma, H.;
Faasen, N. J. J. Chem. Soc. 1955, 1313–1316.
13. Kampouris, E. M. J. Chem. Soc. (C) 1967, 1235–1238.
(
6 g, 0.02 mol, Oxone), and disodium EDTA (10 mg)
were added all at once to 150 ml 0.5 M sodium
phosphate buffer, pH 7 and 50 ml methanol. The
mixture was stirred for 2–3 h at rt and then allowed to
stand for 3 days. It was then cooled on ice and acidified
with 10 ml acetic acid. Byproducts and unreacted start-
ing material were extracted with dichloromethane.
Conc. HCl was then added to the aqueous phase and