Kinetic and product studies on the side-chain fragmentation of 1-arylalkanol radical cations in aqueous solution: Oxygen versus carbon acidity

Article abstract of DOI:10.1002/(SICI)1521-3765(19990604)5:6<1785::AID-CHEM1785>3.0.CO;2-0

A kinetic and product study of the side-chain fragmentation reactions of a series of 1-arylalkanol radical cations (4-MeOC₆H₄CH(OH)R^?+) and some of their methyl ethers was carried out; the radical cations were generated by pulse radiolysis and γ radiolysis in aqueous solution. The radical cations undergo side-chain fragmentation involving the C_α-H and/or C_α-C_β bonds, and their reactivity was studied both in acidic (pH ≤ 4) and basic (pH 10 - 11) solution. At pH 4, the radical cations decay with first-order kinetics, and the exclusive reaction is C_α-H deprotonation for 1^?+, 2^?+, and 3^?+ (R = H, Me, and Et, respectively) but C_α-C_β bond cleavage for 5^?+, 6^?+, and 7^?+ (R = tBu, CH(OH)Me, and CH(OMe)Me, respectively). Both types of cleavage are observed for 4^?+ (R = iPr). The radical cations of the methyl ethers 8^?+, 9^?+, and 10^?+ (R = H, Et, and iPr, respectively) undergo exclusive deprotonation, whereas C-C fragmentation predominates for 11^?+ (R = tBu). Large Cα deuterium kinetic isotope effects (4.5 and 5.0, respectively) were found for 1^?+ and its methyl ether 8^?+. Replacement of an α-OH group by OMe has a very small effect on the decay rate when the radical cation undergoes deprotonation, but a very large, negative effect in the case of C-C bond cleavage. It is suggested that hydrogen bonding of the α-OH group with the solvent stabilizes the transition state of the C-C bond fragmentation reaction but not that of the deprotonation process; however, other factors could also contribute to this phenomenon. The decay of the radical cations is strongly accelerated by HO^-, and all the α-OH substituted radical cations react with HO^- at a rate (≈10¹⁰M^-1S^-1) very close to the limit of diffusion control and independent of the nature of the bond that is finally broken in the process (C-H or C-C). The methyl ether 8^?+, which exclusively undergoes C-H bond cleavage, reacts significantly slower (by a factor of ca. 50) than the corresponding alcohol 1^?+. These data indicate that 1-arylalkanol radical cations, which display the expected carbon acidity in water, become oxygen acids in the presence of a strong base such as HO^- and undergo deprotonation of the O-H group; diffusion-controlled formation of the encounter complex between HO^- and the radical cation is the rate-deter^- mining step of the reaction. It is sug^- gested that, within the complex, the proton is transferred to the base to give a benzyloxyl radical, either via a radical zwitterion (which undergoes intramolecular electron transfer) or directly (electron transfer coupled with deprotonation). The latter possibility seems more in line with the general base catalysis (β ≈ 0.4) observed in the reaction of 5^?+, which certainly involves O-H deprotonation. The benzyloxyl radical can then undergo a β C-C bond cleavage to form 4-methoxybenzalde-hyde and R^? or a formal 1,2-H shift to form an a-hydroxybenzyl-type radical. The factors of importance in this carbon/ oxygen acidity dichotomy are discussed.

Full text of DOI:10.1002/(SICI)1521-3765(19990604)5:6<1785::AID-CHEM1785>3.0.CO;2-0

FULL PAPER
Kinetic and Product Studies on the Side-Chain Fragmentation of 1-Aryl-
alkanol Radical Cations in Aqueous Solution: Oxygen versus Carbon Acidity
Enrico Baciocchi,*[a] Massimo Bietti,[b] and Steen Steenken*[c]
.
.
‡
‡
Abstract: A kinetic and product study
of the side-chain fragmentation reac-
tions of a series of 1-arylalkanol radical
its methyl ether 8 . Replacement of an
a-OH group by OMe has a very small
effect on the decay rate when the radical
cation undergoes deprotonation, but a
very large, negative effect in the case of
C C bond cleavage. It is suggested that
hydrogen bonding of the a-OH group
with the solvent stabilizes the transition
state of the C C bond fragmentation
reaction but not that of the deprotona-
tion process; however, other factors
could also contribute to this phenomen-
on. The decay of the radical cations is
strongly accelerated by HO , and all the
a-OH substituted radical cations react
ing alcohol 1 . These data indicate that
1-arylalkanol radical cations, which dis-
play the expected carbon acidity in
water, become oxygen acids in the
presence of a strong base such as HO
and undergo deprotonation of the O H
group; diffusion-controlled formation of
the encounter complex between HO
and the radical cation is the rate-deter-
mining step of the reaction. It is sug-
gested that, within the complex, the
proton is transferred to the base to give
a benzyloxyl radical, either via a radical
zwitterion (which undergoes intramo-
lecular electron transfer) or directly
(electron transfer coupled with depro-
tonation). The latter possibility seems
more in line with the general base
catalysis (b ꢁ 0.4) observed in the reac-
.
‡
cations (4-MeOC6H4CH(OH)R ) and
some of their methyl ethers was carried
out; the radical cations were generated
by pulse radiolysis and g radiolysis in
aqueous solution. The radical cations
undergo side-chain fragmentation in-
volving the Ca H and/or Ca Cb bonds,
and their reactivity was studied both in
acidic (pH ꢀ 4) and basic (pH 10 ± 11)
solution. At pH 4, the radical cations
decay with first-order kinetics, and the
exclusive reaction is Ca H deprotona-
.
.
.
‡
‡
‡
1
tion for 1 , 2 , and 3 (R ˆ H, Me, and
with HO at a rate (ꢁ1010 m 1 s ) very
Et, respectively) but Ca Cb bond cleav-
close to the limit of diffusion control and
independent of the nature of the bond
that is finally broken in the process
.
.
.
‡
‡
‡
age for 5 , 6 , and 7 (R ˆ tBu,
CH(OH)Me, and CH(OMe)Me, respec-
tively). Both types of cleavage are
.
.
‡
‡
(C H or C C). The methyl ether 8
,
tion of 5 , which certainly involves
O H deprotonation. The benzyloxyl
radical can then undergo a b C C bond
cleavage to form 4-methoxybenzalde-
.
‡
observed for 4 (R ˆ iPr). The radical
which exclusively undergoes C H bond
cleavage, reacts significantly slower (by
a factor of ca. 50) than the correspond-
.
.
‡
‡
cations of the methyl ethers 8 , 9 , and
.
‡
10 (R ˆ H, Et, and iPr, respectively)
undergo exclusive deprotonation,
whereas C C fragmentation predomi-
.
hyde and R or a formal 1,2-H shift to
form an a-hydroxybenzyl-type radical.
The factors of importance in this carbon/
oxygen acidity dichotomy are discussed.
Keywords: alcohols ´ cleavage reac-
´ pulse radiolysis ´ radical
cations ´ radical reactions
.
‡
nates for 11 (R ˆ tBu). Large Ca deu-
tions
terium kinetic isotope effects (4.5 and
.
‡
5.0, respectively) were found for 1 and
Introduction
[a] Prof. E. Baciocchi
The degradation of lignin, a three-dimensional polymer
composed of phenylpropane units (Figure 1), is a process of
Á
Dipartimento di Chimica, Universita La Sapienza
P. le A. Moro, 5, I-00185 Rome (Italy)
Fax: (‡39)6-490421
OH
E-mail: baciocchi@axcasp.caspur.it
[b] Dr. Massimo Bietti
(H,H3CO)
R
O
HO
H3CO
O
R
Dipartimento di Scienze e Tecnologie Chimiche
Á
Universita Tor Vergata, Via della Ricerca Scientifica
I-00133 Rome (Italy)
O
HO
H3CO
[c] Prof. S. Steenken
Max-Planck-Institut für Strahlenchemie, Stiftstrasse 34 ± 36
D-45470 Mülheim (Germany)
OCH3
OH
(H,OCH3)
Fax: ‡(49)208-3063951
Figure 1. Schematic representation of lignin showing the b-O-4 inter-unit
linkages and the presence of benzyl alcohol groups.
E-mail: steenken@mpi-muelheim.mpg.de
Chem. Eur. J. 1999, 5, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1785 $ 17.50+.50/0
1785

E. Baciocchi, M. Bietti, S. Steenken
FULL PAPER
enormous practical importance both for the
production of cellulose in the pulp and
paper industry and for preparing useful
aromatic compounds from a continuously
renewable source.[1] Recently, the use of
lignin-degrading fungi has generated much
interest, and most attention has been fo-
cused on the white-rot fungus Phanero-
chaete chrysosporium, which secretes lignin
peroxidase (LiP), a ferric hemoprotein that
depolymerizes both lignin and its model
compounds.[2] Convincing evidence exists
that this enzyme unergoes an electron-
transfer reaction to form radical cations
that structurally resemble the 1-arylalkanol
radical cations A where R can be a complex group, for
example, CH(CH2OH)Ar or CH(CH2OH)OAr. The key step
in the degradation process is then the cleavage of the C R
bond [Eq. (1)], a typical side-chain fragmentation reaction of
alkylaromatic radical cations.[3] Very likely, this reaction also
plays a major role when lignin is degraded by industrial
chemical[4] or electrochemical[5] oxidations.
R
R
OH
CH OH
CH OMe
CD2OX
H3C
C
CH3
OCH3
OCH3
OCH3
OCH3
1: R = H
2: R = Me
3: R = Et
4: R = iPr
5: R = tBu
6: R = CH(OH)Me
7: R = CH(OMe)Me
8: R = H
9: R = Et
10: R = iPr
11: R = tBu
12: R = H
13: R = Me
14
.
‡
water (the solvent in which LiP operates) with 4-methoxyl-
substituted substrates; the methoxyl group results in radical
cations that absorb in the UV/Vis region of the spectrum and
have a lifetime sufficient for detection by pulse radiolysis.
Results
R
Generation of the radical cations: Radical cations of subtrates
1 ± 14 were generated in aqueous solution either by pulse
radiolysis or steady-state g radiolysis with sulfate radical
CHOH
CHO
.
anion SO4 or Tl2‡ as the oxidant. In the former case
+
R
+
H+
(1)
(method 1) Equations (2) ± (5) apply. Radiolysis of water
OCH3
OCH3
OCH3
OCH3
A
The presence of a-OH groups is a structural feature of
fundamental importance in lignin and in the model compound
.
‡
A
. The OH group strongly favors cleavage of the side-chain
Ca Cb bond (simply referred to as C C bond in the following)
in aromatic radical cations [Eq. (1)].[6] Clear evidence in this
respect comes from our previous study on the fragmentation
reactions of 1- and 2-arylalkanol radical cations.[7] The
important role of the a-OH group was evident in the
significantly higher rate of potassium 12-tungstocobalt(iii)ate
(Co(iii)W) induced oxidation of 4-MeOC6H4CH(OH)tBu
relative to its methyl ether. On the basis of a pulse radiolysis
study of the reactivity of the 4-MeOC6H4CH(OH)CH2Ph
radical cation, it was suggested that the favorable effect of the
a-OH group is mainly due to the stabilization by hydrogen
bonding of the transition state that leads to C C bond
cleavage.
.
leads to formation of hydroxyl radicals ( OH) and hydrated
electrons (eaq) [Eq. (2)]. The former is scavenged by 2-meth-
yl-2-propanol [Eq. (3); k ˆ 6  108 m 1 s ],[9] while eaq reacts
1
.
with the peroxodisulfate anion to yield SO4 [Eq (4); k ˆ
.
1.2 Â 1010 m 1 s ].[9] Then SO4 reacts with the aromatic
1
compounds[10] to form the corresponding radical cations
[Eq (5); k ˆ 5  109 m 1 s for anisole and derivatives[7, 11]].
1
For reactions carried out at pH ꢀ 4, Tl2‡, produced by
‡
irradiating N2O-saturated aqueous solutions of Tl [Eqs. (2),
(6) ± (8)] was used as oxidant (method 2). The role of N2O is
Given our continuing interest in the reactivity of alkylar-
omatic radical cations and the importance of Equation (1)
with respect to the catalytic reactivity of LiP, we felt that
further investigation of the effects of structure and the basicity
of the medium on the reactivity of benzyl alcohol and a-
alkylbenzyl alcohol radical cations was necessary. We now
report on a pulse and steady-state radiolysis study of the
fragmentation reactions of several 1-(4-methoxyphenyl)alka-
nol radical cations and some of their methyl ethers (substrates
1 ± 14), in the pH range 3.5 ± 11.[8] This study was carried out in
1
to scavenge eaq (k ˆ 9.1  109 m 1 s ),[12] leading to formation
.
of hydroxyl radicals [Eq. (6)]. Oxidation of Tl by OH[13]
‡
produces Tl2‡ [Eq (7); k ˆ 1.2  1010 m 1 s ],[14] which is
1
known to react with aromatic compounds by one-electron
1786
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1786 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 6

Fragmentation Reactions
1785±1793
Table 1. Product distributions obtained from the decomposition of
4-MeOC6H4CH(X)R radical cations in aqueous solution.[a]
transfer to give the corresponding radical cations [Eq (8); k ꢁ
1
5 Â 108 m 1 s ].[11]
R
X
pH[b]
Aldehyde [%]
Ketone [%]
.
.
‡
‡
2
3
Me
OH
4
11
4
< 0.1
< 0.1
0.5
70.0
0.5
ꢂ 99.5
ꢂ 99.5
11.0
95.0
1.5
> 95
> 95
88
> 99.9
> 99.9
99.5
30.0
99.5
±
±
89.0
5.0
Product analysis: It is well known that two fragmentation
pathways are possible for 1-arylalkanol radical cations
(Scheme 1, R=H): heterolytic cleavage of the Ca H bond
(path a; from now on, Ca H is referred to as C H) and
homolytic cleavage of the C C bond (path b).[15±17]
Et
OH
11
.
.
‡
‡
9
6
7
4
Et
OMe
OH
OH
OH
4
[c]
CH(OH)Me
CH(OMe)Me
iPr
.
‡
[c]
.
‡
4
11
4
4
11
4
OH
O
-
- H+
- e , - H+
.
‡
10
iPr
tBu
OMe
OH
98.5
ArCR
ArCR
a
.
‡
[d]
H
C
C
C
5
OH
[d]
ArCHR
.
‡
11
tBu
OMe
12
- H+, - R
b
ArCHO
[a] The radical cations were generated at room temperature by method 1,
by steady-state 60Co g radiolysis of argon-saturated aqueous solutions
containing 0.5 ± 5.0mm substrate, 0.2 ± 1.0mm K2S2O8, and 0.2 ± 1.0m
Scheme 1. The two possible fragmentation pathways for 1-arylalkanol
radical cations.
2-methyl-2-propanol (substrate/oxidant ratio between
2 and 5). The
reported values are the average of two to four experiments. [b] The pH
of the solution was adjusted to 4 or 11 with HClO4 or NaOH. [c] The
oxidation reactions were carried out in 55/45 AcOH/H2O at 508C with
Co(iii)W as oxidant (ref. [7] and unpublished results). [d] A small amount
of ketone (<5%) was observed under all conditions. The fact that this
amount is not influenced significantly by the presence of the base is
probably due to the existence of a background hydrogen atom abstraction
For path a, an a-hydroxybenzyl-type radical is formed that,
under oxidizing conditions (vide infra), undergoes oxidation
followed by proton loss to give the corresponding ketone.
Path b leads directly to an aromatic aldehyde. Hence, the
study of the reaction products provides quantitative informa-
tion on the relative importance of the two fragmentation
pathways of 1-arylalkanol radical cations.
.
reaction by SO4 that eventually leads to formation of the ketone. In
agreement with this hypothesis is the observation that in the oxidation of 5
by Co(iii)W, a well-known outer sphere one-electron oxidant but not an H
atom abstractor,[18, 19] the only observed product was 4-methoxybenzalde-
hyde.[7, 20]
.
.
.
‡
‡
‡
The reaction products of radical cations 2 ± 5 and 9
±
.
‡
11 were analyzed after steady-state g radiolysis, using
method 1 to produce the radical cations. Argon-saturated
aqueous solutions containing 0.5 ± 5.0mm substrate, 0.2 ±
1.0mm K2S2O8, and 0.2 ± 1.0m 2-methyl-2-propanol were
irradiated at room temperature with a 60Co g source at dose
radical cations.[7, 11] The rates of decay of the radical cations
were determined spectrophotometrically by measuring the
decrease in optical density at 440 ± 450 nm or by monitoring
1
rates of 0.5 Gys until 40% conversion with respect to
‡
the production of H by the ac-conductance technique.
peroxodisulfate was attained. In order to minimize over-
oxidation of the initial products, all experiments were carried
out with a substrate/oxidant ratio ꢂ2. The product distribu-
tion was studied at pH ꢁ 4 and 11. Products were identified
and quantitatively determined by HPLC (comparison with
authentic samples). The results are collected in Table 1.
It is evident that at pH 4 deprotonation is the exclusive
In acid media (pH ꢀ 4), the latter technique was generally
preferred, and radical cations were more often generated by
method 2 than by method 1. The conductance technique has
higher sensitivity under our pulse-radiolysis conditions and
hence allows the use of significantly lower dose/pulse ratios
than the spectrophotometric technique. The advantage of
.
method 2 is that scavenging of OH by 2-methyl-2-propanol is
.
‡
reaction of the a-alkylbenzyl alcohol radical cations 2 and
not necessary, so that the high concentration of
.
.
‡
‡
3
(R ˆ Me and Et, respectively), whereas 5 (R ˆ tBu)
.
‡
.
CH2C(CH3)2OH radicals formed in this process [Eq. (3)] is
undergoes almost exclusively C C bond cleavage. For 4
(R ˆ iPr), both C C and C H bond cleavage are observed.
avoided. Use of low doses and thus low concentrations of
radicals in solution is particularly important when dealing
with long-lived radical cations (lifetime ꢂ 0.1 ms) for which
decay by second-order radical ± radical reactions begins to
compete with unimolecular decay.[21] This problem is much
less important in basic media, in which decomposition of the
radical cations is significantly faster. Therefore, under basic
conditions, the radical cations were always generated by meth-
od 1, and the rates were monitored by spectrophotometry.
As an example, the time-resolved spectra obtained for the
oxidation of 2 at pH 3.9 are shown in Figure 2, in which the
.
.
‡
‡
The methyl ethers 9 and 10 (R ˆ Et and iPr, respectively)
.
‡
undergo exclusively deprotonation, whereas with 11 (R ˆ
tBu) partitioning between the two pathways occurs. The C C
.
.
‡
‡
bond cleavage is the only pathway for 6 and 7 , in which an
OX group (X ˆ H, Me) is present in the b position of the side
chain.
When the reactions are carried out in the presence of 1mm
NaOH (pH 11), there is a significant increase in the extent of
.
‡
C C bond cleavage, which becomes the main path for 3
(R ˆ Et) and the almost exclusive path for 4 (R ˆ iPr).
.
‡
.
‡
characteristic absorptions due to the radical cation 2
,
centered at 290 and 440 nm,[7, 11] are clearly visible. These
absorptions reach a maximum 5 ms after the pulse [completion
of radical cation formation, Eq. (5)] and then decrease
according to first-order kinetics (see inset a for absorption
at 440 nm) that reflects the decay of the radical cations. This
Pulse radiolysis: The radical cations were produced by
methods 1 and 2 with a 300 ns, 3 MeV electron pulse. In all
cases they exhibited the characteristic UV/Vis absorption
bands, centered around 290 and 440 ± 450 nm, of anisole-type
Chem. Eur. J. 1999, 5, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1787 $ 17.50+.50/0
1787

E. Baciocchi, M. Bietti, S. Steenken
FULL PAPER
.
.
‡
‡
For 6 and 7 , the decay rate of the radical cation is very
high. It is therefore necessary to find conditions under which
formation of the radical cation [Eq. (5)] is not rate-determin-
ing. This requirement is satisfied at concentrations of the
.
‡
parent substrates of ꢂ7mm and 1mm for 6 (Figure 3) and
.
Figure 2. Time-resolved absorption spectra for the reaction of SO4 with 2
(0.2mm) recorded on pulse radiolysis of an Ar-saturated aqueous solution
*
(pH 3.9) containing 0.1m 2-methyl-2-propanol and 10mm K2S2O8, at 5 ( ),
~
&
200 ( ), and 700 ms ( ) after the 300 ns, 3 MeV electron pulse. Insets:
.
‡
a) First-order decay of 2 monitored at 440 nm; b) Buildup of conductance
.
‡
‡
(due to production of H ) in the decay of 2 ; c) Buildup of absorption at
280 nm due to formation of 4-MeOC6H4C (OH)CH3 (Ar, ), which is
.
*
Figure 3. Plot of kdec versus [6]. The radical cation was generated by
method 1 from an argon-saturated aqueous solution containing (0.1 ±
12mm) 6, 20mm K2S2O8, and 0.2m 2-methyl-2-propanol, and its decay
was monitored at 450 nm (Ar ˆ 4-MeOC6H4).
*
scavenged by O2
(
). For the determination of the extinction coefficient,
.
G(radical cation) ˆ G(SO4 ) ˆ 3.1  10 7 molJ 1 was used.[22]
decay is accompanied (inset c) by a corresponding buildup of
optical density around 280 nm, which indicates the formation
of a product that absorbs at this wavelength more strongly
than the radical cation itself (an isosbestic point is visible at ꢁ
310 nm). In this case, the decay of the radical cation was
.
.
.
‡
‡
‡
7 , respectively. Thus, for 6 and 7 , the decay rates
obtained after reaching the plateau in the plots of kdec versus
substrate concentration were taken as the unimolecular
decomposition rates. (In all other cases, the decay rates are
much smaller than the formation rates even at lower substrate
concentrations.)
‡
also followed by monitoring the production of H by
the conductance technique (inset b), and the first-order
rate constant obtained in this way was the same as that
obtained by measuring the decrease in absorption at 440 nm
(inset a).
The first-order rate constants for decay in water at pH 4 of
the investigated radical cations are listed in Table 2. These
Inset c shows that the absorption buildup at 280 nm is no
longer observed when O2 is present, and this indicates that the
species responsible for this absorption is rapidly scavenged by
oxygen. This suggests that this species is a carbon-centered
radical and that therefore the decay of the radical cation
involves C H deprotonation (Scheme 1, path a) to give the
Table 2. Rate constants kdec for the uncatalyzed decay of radical cations
.
.
‡
‡
1
± 14
, generated by pulse radiolysis of the parent substrate
4-MeOC6H4CH(X)R in aqueous solution (pH ꢀ 4) at 258C.
1
[a]
R
H
X
kdec [s
]
Bonds cleaved
C H
.
‡
1
OH
1.5 Â 104
1.8 Â 104[b]
3.3 Â 103
1.5 Â 104
3.0 Â 103
7.0 Â 103
9.0 Â 103[b]
5.4 Â 103
1.6 Â 103
3.5 Â 103
1.1 Â 103
1.5 Â 105[b]
.
‡
‡
4-methoxy-a-hydroxybenzyl-type
radical
4-MeOC6H4-
12
8
D
H
D
Me
OH
C D
C H
C D
C H
.
.
‡
.
CH (OH)CH3, as is also indicated by the products (Table 1).
Similar results were obtained with the other substrates, with
OMe
OMe
OH
13
.
‡
2
.
.
.
.
‡
‡
‡
‡
the exception of 5 , 6 , and 7 . In the case of 5 , decay of
the radical cation led to the formation of a species which has a
strong absorption at 285 nm and does not react with oxygen.
This species was identified as 4-methoxybenzaldehyde,
formed by homolytic C C bond fragmentation of the radical
cation (Scheme 1, path b), again in agreement with the study
of the reaction products.[7]
.
.
.
‡
‡
‡
.
3
9
4
Et
Et
iPr
iPr
OH
OMe
OH
OMe
OH
OMe
OH
C H
C H
C H, C C
C H
C C
C H, C C
C C
C C
‡
‡
10
.
‡
.
5
tBu
tBu
CH(OH)Me
CH(OMe)Me
Me,Me[d]
11
23
.
‡
6
7
1.0 Â 107[b, c]
1.0 Â 106[b]
2.9 Â 102
.
.
‡
‡
.
‡
The decay of 6 and 7 led to the formation of a species
that absorbs strongly around 280 nm and is quenchable by
oxygen. This species was identified as the 4-methoxy-a-
OH
OH
.
‡
14
C C
[a] The radical cations were generated by method 2, with doses such that
ꢀ1mm of radicals were produced. Rates of decay were measured by
monitoring the increase in conductance at pH ꢁ 3.5. [b] The radical cations
were generated by method 1 in argon-saturated aqueous solutions contain-
ing 0.1 ± 1.0mm substrate, 2.0mm K2S2O8, and 0.1m 2-methyl-2-propanol,
the pH was adjusted to about 4 with HClO4, and the rates of decay were
measured by monitoring the decrease in optical absorption at 440 ±
450 nm: dose ꢁ 1 Gy per pulse. [c] Obtained by using 7 ± 12mm of
substrate (see text). The pulse width was reduced to 25 ns. [d] 4-
Methoxycumyl alcohol (4-MeOC6H4C(OH)Me2), in which the second a-
hydrogen atom is also replaced by a methyl group.
.
hydroxy benzyl radical 4-MeOC6H4CH (OH), formed by
heterolytic C C bond fragmentation of the radical cations
(Scheme 2: Arˆ 4-MeOC6H4; X ˆ H, Me).
+
ArCH(OH)CH(OX)Me
ArCHOH
+
CH(OX)Me
Scheme 2. Heterolytic C C bond fragmentation of the radical cations
(Arˆ 4-MeOC6H4; X ˆ H, Me).
1788
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1788 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 6

Fragmentation Reactions
1785±1793
rate constants did not change in the pH range 3 ± 5, but
significantly increased on addition of HO . This effect was
is drastically different for C C bond cleavage. The rate of this
process is slowed down by almost four orders of magnitude on
.
.
.
.
‡
‡
‡
‡
studied quantitatively for the radical cations 1 ± 5 , 8 , and
going from the alcohol radical cation 5 to its methyl ether
.
.
.
‡
‡
‡
12 ± 14 , the decay rates of which were measured spectro-
photometrically at 440 ± 450 nm as a function of the HO
concentration. Clean first-order decays were observed, and
linear dependencies of the observed rates (kobs) on the
concentration of HO were found. From the slope of these
plots, the second-order rate constants (kOH ) for the reaction
of HO with the radical cations were determined (Table 3).
11 . This rate effect is particularly striking if one considers
.
.
‡
‡
that 5 and 11 have very similar C C bond dissociation
energies (BDE).[23]
This large difference between the effects of a-OH and a-
OMe on the cleavage of the C C bond was already noted by
us, although indirectly, in a study of the chemical oxidation of
5 and 11. It was suggested that the OH group permits a strong
stabilization of the transition state of the C C bond fragmen-
tation reaction by engaging in hydrogen bond formation with
solvent molecules.[7] This concept can be extended to explain
the intriguing fact that, in contrast to C C cleavage, there is
no significant difference between a-OH and a-OMe when the
cleavage of the radical cation involves the C H bond.
Scheme 3 (Arˆ 4-MeOPh) illustrates, with the mesomeric
structures a ± e, how the a-OH group may engage in hydro-
gen-bond stabilization of the transition state in the homolytic
C C bond cleavage reaction. Mesomer c, with a positively
Table 3. Rate constants (kOH ) for the OH-promoted decay of radical
.
‡
cations 4-MeOC6H4CH(X)R generated by pulse radiolysis of the parent
substrate in aqueous solution at 258C.[a]
1
[b]
R
X
kOH [m 1 s
]
Bonds cleaved
.
‡
.
1
H
D
H
D
Me
Et
iPr
OH
OH
OMe
OMe
OH
OH
OH
OH
OH
1.2 Â 1010[c]
1.1 Â 1010[c]
2.5 Â 108
C H
C D
C H
C D
‡
‡
12
.
‡
.
8
13
1.4 Â 108
.
.
.
.
‡
‡
‡
‡
.
2
3
4
5
1.4 Â 1010[c]
1.2 Â 1010
1.2 Â 1010
1.3 Â 1010[c]
1.2 Â 1010[c]
C H
C H, C C
C H, C C
C C
ˆ
charged oxygen atom and a fully formed C O double bond,
should be particularly important, and for this structure strong
stabilization by hydrogen bonding with the solvent water is
anticipated.
tBu
Me, Me[d]
‡
14
C C
[a] The radical cations were generated by method 1 from oxygen-saturated
aqueous solutions containing 1.0mm substrate, 10mm K2S2O8, and 0.1m
2-methyl-2-propanol with doses such that ꢀ3mm of radicals were
produced. 1mm Na2B4O7 was added to buffer the pH of the solution.
[b] The observed rates were measured by monitoring the decay of the
optical absorption at 440 ± 450 nm. The pH of the solution was varied
between 8.5 and 10.5, and the second-order rate constants for reaction of
the radical cations with HO were obtained from the slope of the plots of
the observed rates kobs versus the NaOH concentrations. [c] Lower rates
were observed in the presence of 0.5m Na2SO4 (negative salt effect).[8] See
also Table 4. [d] 4-Methoxycumyl alcohol (4-MeOC6H4C(OH)Me2), in
which the second a-hydrogen atom is also replaced by a methyl group.
In contrast, the transition state for the C H bond cleavage
(Scheme 4; the proton-accepting base H2O is also shown) is
envisaged as a resonance hybrid of structures f ± j. The most
important structure is probably j, which represents the
heterolytic cleavage of the C H bond, whereas structure h
(corresponding to c in Scheme 3) should be unimportant since
it involves unfavorable homolytic cleavage of the C H bond.
‡
An additional aspect is that OH (s ˆ 0.92)[25] is a
‡
significantly more efficient electron donor than OMe (s ˆ
0.78).[25, 26] This difference in the ability for electron
donation is likely to be felt more in the transition state of
the C C bond cleavage that in that for C H cleavage, since
there is evidence that a greater accumulation of positive
charge (by bond delocalization) is required in the former
transition state than in the latter.[27±29]
Discussion
Studies in water at pH 4: The results of product studies and
pulse-radiolysis experiments (Tables 1 and 2) show that the
.
‡
.
‡
decay of radical cations 1
4
±
.
.
.
.
‡
‡
‡
‡
, 8 ± 10 , 12 , and 13
H
O
H
O
H
O
H
O
H
O
proceeds by cleavage of the
C H bond. This is well sup-
ported by the large deuterium
kinetic isotope effect (kH/kD ˆ
H
H
H
H
H
H
H
H
H
H
+
O
C
H
O
C
H
O
C
H
O
O
C
H
+
+
Ar
R
Ar
R
Ar
R
Ar
R
Ar
R
C
H
.
.
‡
‡
4.5 for 1 /12 and kH/kD ˆ 5.0
.
.
‡
‡
for 8 /13 ). Probably, H2O is
the proton-abstracting base as
the rate constants did not
change between pH 3 and 5.
Replacement of an a-OH
group by an a-OMe group has
a very small effect upon the
deprotonation rate, as shown by
the fact that the reactivities of
a
b
c
d
e
Scheme 3. The transition state for the homolytic C C bond cleavage reaction (Arˆ 4-MeOPh).
H
O
H
O
H
O
H
O
H
O
H
H
H
H
H
H
H
H
H
H
+
O
O
C
H
O
C
H
O
C
H
O
+
+
Ar
H
Ar
H
O H
Ar
H
Ar
H
O H
Ar
H
O H
H
C
H
C
H
.
.
.
‡
‡
‡
O H
H
O H
1
, 3 , and 4 are similar to
H
H
H
those of the corresponding
.
.
‡
‡
j
f
g
h
i
methyl ethers
8
,
9
,
and
.
‡
10 . This situation, however,
Scheme 4. The transition state for the C H bond cleavage with the proton-accepting base H2O.
Chem. Eur. J. 1999, 5, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1789 $ 17.50+.50/0
1789

E. Baciocchi, M. Bietti, S. Steenken
FULL PAPER
Â
Saveant et al. recently pointed out that the (rate-determin-
There is a decrease in reaction rate (ca. 50-fold) on going
.
.
‡
‡
ing) C C bond cleavage step in a radical cation may be
reversible. If so, the overall rate will be high only if the
subsequent reactions of the cleavage fragments are sufficient-
ly fast.[30] This is certainly the case for 5 , which gives rise to
4-MeOC6H4CH OH, from which deprotonation to
4-MeOC6H4CHO should be extremely rapid, but not for
11 , which yields 4-MeOC6H4CH OMe, a stabilized carbo-
cation that cannot deprotonate. Therefore 5 should be more
reactive than 11 , as is observed (Table 2).
from 1 to its methyl ether 8 , which undergoes C H bond
.
‡
cleavage and accordingly has a kH/kD value of 1.8 (cf. 8 and
.
.
.
‡
‡
‡
13 ). The difference in reactivity of 1 and 8 cannot be
attributed to a difference in the electronic effects of OH and
OMe, since in acid medium the influence of these groups on
C H deprotonation is very similar (see previous section).
Hence, this observation together with the very similar
.
‡
‡
.
‡
‡
.
.
.
‡
‡
‡
reactivity of 1 and 5 , in spite of the different types of
side-chain fragmentation, point to a mechanism in which the
a-OH group plays a key and specific role.
.
‡
For the radical cations with an OX group (X ˆ H, Me) in
.
.
‡
‡
the b position, such as 6 and 7 , a different situation applies
since the C C bond cleavage is now heterolytic (Scheme 2).
At first sight, this observation is surprising, as the oxidation
The high (diffusion-controlled) rates are typical of thermo-
dynamically favored proton-transfer reactions between elec-
tronegative atoms.[41] The most reasonable conclusion is that
in the radical cations with a-OH substituents, the reaction
center is the a-OH group itself. In other words, these radical
cations, which in water display the expected carbon acidity,
become oxygen acids in the presence of a strong base such as
HO and undergo OH deprotonation as the first step of the
decay process.
The fact that the deprotonation rate is diffusion-controlled
indicates that the interaction between HO and the radical
cation to form the encounter complex is the rate-determining
step of the reaction.[41] Therefore, the radical cation must be a
stronger oxygen acid than H2O, that is, its pKa must be much
smaller than 15.7.[42] Once the complex is formed, several
pathways linking the complex to the observed C C and C H
fragmentation products can be envisaged (Scheme 5, Arˆ 4-
MeOC6H4).
.
potentials in MeCN of CH(OX)Me (X ˆ H, Me: 0.45 V vs.
.
SCE)[31] and 4-MeOC6H4CH OH ( 0.51 V vs. SCE)[32, 33]
suggest a slight preference for homolytic cleavage. However,
in H2O the difference in solvation energy between
4-MeOC6H4CH OH and the smaller and less delocalized
carbocation CH(OX)Me should be much larger than in
MeCN.[34] Hence, in H2O the oxidation potential of
‡
‡
.
CH(OX)Me (X ˆ H, Me) is likely to be significantly more
.
negative than that of 4-MeOC6H4CH OH, and heterolytic
.
‡
C C bond cleavage is therefore the favored pathway for 6
.
‡
and 7 .
The above-mentioned difference in the electronic effects of
.
‡
OH and OMe may also explain the higher reactivity of 6
.
compared with 7 .[35, 36] On the same basis, the observation
‡
.
.
‡
‡
that 7 fragments more rapidly than 5 suggests that the
OMe group is more effective than two methyl groups in
stabilizing the partial positive charge that develops on the
scissible C C bond in the transition state for C C bond
cleavage.
R
R
R
a
b
-
-
-
ArCHOH
+
OH
ArCHOH OH
ArCHO
- H2O
Finally, information on the effect of the a-alkyl group upon
the rate of C H cleavage is provided by comparison of the
f
- H2O
c
g
.
.
‡
‡
radical cations 1 ± 4 , which undergo C H bond cleavage
.
.
‡
‡
(Table 1) exclusively (1 -3 ) or as the main reaction path
R
R
.
‡
(4 ). The kinetic data in Table 2 show that by changing the a-
alkyl group the deprotonation rate decreases in the order H >
Me > Et > iPr, as expected on the basis of stereoelectronic
effects.[37, 38] The bulkier the alkyl group, the more energeti-
cally costly it is to reach the conformation most suitable for
bond cleavage, that is, one in which the C H bond is collinear
with the p-electron system. The differences in rate for the
three alkyl groups are rather small, but this is not surprising
since when dealing with steric effects in the alkyl-group series,
drastic changes are generally observed only on going from iPr
to tBu.[39] Probably, this holds also in the present case, but
unfortunately we were unable to determine a reliable reaction
e
d
ArCHO
ArCHO + R
ArCOH
Scheme 5. Possible pathways linking the complex to the observed C C and
C H fragmentation products (Arˆ 4-MeOC6H4).
Within the complex, the proton may be transferred to the
base to give a benzyloxyl radical directly (intramolecular
electron transfer concerted with deprotonation, path f), or via
a radical zwitterion which undergoes intramolecular electron
transfer (paths b, c). Two routes are then available for the
benzyloxyl radical: b C C bond cleavage to form 4-methoxy-
.
.
‡
rate for the deprotonation of 5 , since it undergoes almost
benzaldehyde and the radical R (path d), or a formal 1,2-H
exclusively C C bond cleavage (see footnote [d] in Table 1).
shift to give an a-hydroxybenzyl-type radical (path e).
Both C C b cleavage[44] and 1,2-H shift reactions[45] are
well-known processes for alkoxyl radicals. The latter seem to
be limited to aqueous solution[46] and require the participation
of solvent molecules (a direct shift of the H atom from carbon
to oxygen would have too high an energetic cost).[47] The
competition in the benzyloxyl radical between b cleavage and
Reactions in the presence of HO : Remarkably, all a-OH-
substituted radical cations react with HO at the same rate
(Table 3), that is, that of a diffusion-controlled reaction
1
(ꢁ1010 m 1 s ),[40] independent of the nature of the bond
eventually broken in the process (C H or C C). In line with
.
.
.
‡
‡
this, comparing 1 with its Ca-deuterated counterpart 12
1,2-H shift should depend on the stability of R . Hence, our
reveals no deuterium kinetic isotope effect.
experimental observations (Table 1) lead to the conclusion
1790
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1790 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 6

Fragmentation Reactions
1785±1793
that when R is H or Me, the 1,2 shift is the exclusive pathway,
whereas when R is tBu, only b cleavage occurs. An inter-
mediate situation apparently holds for R ˆ Et and iPr.
Satisfactory linear Brùnsted plots were obtained from these
data, and the reactions of the a-OH-substituted radical
.
.
.
‡
‡
‡
cations 1 , 2 , and 5 have practically the same b value
A further alternative is concerted electron transfer and
side-chain fragmentation in the zwitterion (path g). Thus, the
products might be formed without the intermediacy of the
benzyloxyl radical. Intramolecular electron transfer coupled
with side-chain fragmentation was suggested to occur in the
reactions of aromatic radical anions.[48, 49] However, we were
recently able to directly observe the 4-methoxycumyloxyl
radical in the HO -promoted C C bond fragmentation of the
(ca. 0.4), a value higher than that of 0.25 for the reaction of
.
‡
the methyl ether 8 , whose deprotonation involves the C H
bond. However, it has to be pointed out that with the buffer
.
‡
bases the rate constants for reaction with 8 and the a-OH-
substituted radical cations are very similar so that the b values
are practically determined by the rate constants for the HO -
induced reaction, which are in all cases similar. Thus, we
.
‡
cannot be sure whether, with weaker bases than HO , 1 and
.
.
‡
‡
4-MeOC6H4C(CH3)2OH radical cation 14
to form
2
undergo C H or O H deprotonation, and a meaningful
.
4-MeOC6H4COCH3 (Scheme 6).[50] Since 14 also reacts with
HO at a diffusion-controlled rate (see Table 3), it is very
likely that this reaction occurs with the mechanism shown in
Scheme 5 (paths a ± d or alternatively a, f, d).
discussion of their b values is therefore not possible. This,
‡
.
‡
however, is not the case for the reaction of 5 , which should
undergo O H deprotonation. The fact that with the buffer
bases the rate of this reaction is much lower than the
diffusion-controlled limit suggests that with the weaker bases
the proton transfer is thermodynamically uphill. This is
because the benzyl alcohol-type radical cations are weaker
oxygen acids than the conjugate acids of the buffers.
Interestingly, the Brùnsted value is smaller than unity, the
value expected for uphill proton transfer between oxygen
atoms.[41] This may mean that proton transfer is not the only
process involved in the rate-determining step,[51] but that
it is coupled with intramolecular electron transfer that leads
directly to the oxyl radical (Scheme 5, step f). How-
ever, deprotonation coupled to some extent with C C
bond cleavage may also be consistent with the above
results. At present no clear-cut choice between these alter-
natives is possible, and further studies with a wider range of
structures and redox properties of the substrates are neces-
sary.
OH
C
O
C
O
C
H3C
CH3
H3C
CH3
CH3
- H2O
-
+
OH
+
CH3
OCH3
OCH3
OCH3
Scheme 6. The 4-methoxycumyloxyl radical as intermediate in the HO -
promoted C C bond fragmentation of the 4-MeOC6H4C(CH3)2OH radical
cation.
These data support the hypothesis of the formation of an
intermediate benzyloxyl radical in the HO -promoted decay
of 1-arylalkanol radical cations. However, some caution has to
be exerted with respect to the generality of this conclusion.
The results for cumyl alcohol refer to a system in which
intramolecular electron transfer in the radical zwitterion leads
to a relatively stable benzyloxyl radical as its fragmentation
produces the methyl radical, that is, the least stable alkyl radical.
To obtain further information on the detailed mechanism of
Regardless of the detailed mechanism by which the
encounter complex is converted into the products, our results
clearly indicate that benzyl alcohol radical cations exhibit the
expected carbon acidity in water (pH ꢀ 4), but become
oxygen acids in the presence of HO . This shift from carbon
to oxygen acidity, unprecedented in the chemistry of aromatic
radical cations, is both interesting and surprising, since only
the C H but not the O H bond can overlap efficiently with
the p system of the aromatic radical cation and, from the
(thermodynamic) acidity point of view, the deprotonation of
the side-chain a carbon atom (carbon acidity pKa ˆ 7.5 in
MeCN)[24, 34] is strongly favored over deprotonation at oxy-
gen.[42] We feel that an explanation may be found in terms of
the concept of hard and soft acids and bases. The O H group
is a much harder acid center than C H, and its hardness is
probably also increased by the presence of the positive charge
in the ring. Thus, the interaction of the O H group with the
charged hard base HO might be particularly favorable owing
to an effective electrostatic interaction to form a relatively
stable hard ± hard complex.[52] With the uncharged base H2O,
no favorable electrostatic interaction can occur and reaction
takes place at the softer acid center C H.
.
‡
the O H deprotonation process, we studied the decay of 1
2
,
.
.
.
‡
‡
‡
, 5 , and 8 in buffer systems. At constant pH, the rates
depended upon the buffer concentration, indicating general
base catalysis. The second-order rate constants for the
2
reactions with AcO , HCO3 , and HPO4 are listed in
Table 4, together with those for the reaction with HO .
.
.
.
.
‡
‡
‡
‡
Table 4. Second-order rate constants for reaction of 1 , 2 , 5 , and 8
,
generated by pulse radiolysis in aqueous solution, with different bases at
258C.[a]
2
k(AcO )
k(HCO3
)
k(HPO4
)
k(HO )
1
1
1
1
[b, c]
[m 1s
]
[m 1s
]
[m 1s
]
[m 1s
]
.
‡
1
2
5
8
2.8 Â 105
7.8 Â 104
8.0 Â 104
4.2 Â 105
6.0 Â 105
7.3 Â 105
7.2 Â 105
7.0 Â 105
2.7 Â 106
9.9 Â 105
1.5 Â 106
1.9 Â 106
5.4 Â 109[d]
5.3 Â 109[d]
6.0 Â 109
.
.
.
‡
‡
‡
2.0 Â 108
[a] The radical cations were generated by method 1, with doses such that
ꢀ3mm of radicals were produced. The ionic strength of the solution was
buffered with 0.5m Na2SO4. In the experiments with NaOAc, NaHCO3,
and Na2HPO4, the pH of the solution was kept constant (5.5, 7.0, and 8.0
respectively). [b] The pH of the solution was varied between 8.5 and 10.5,
and 1mm Na2HPO4 was added to avoid undesired pH changes. [c] The
values are lower than those reported in Table 3 due to the presence of
0.5m Na2SO4, which exerts a negative salt effect. [d] Saturated with
oxygen.
The problem can also be dealt with in terms of the Marcus
theory of proton-transfer reactions on the basis of the fact that
carbon acids have a much larger intrinsic barrier for proton
transfer than oxygen acids.[53] In particular, Saveant has
Â
recently shown for the case of radical cations from NADH
Chem. Eur. J. 1999, 5, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1791 $ 17.50+.50/0
1791

E. Baciocchi, M. Bietti, S. Steenken
FULL PAPER
sodium hydroxide, disodium tetraborate decahydrate, potassium thiocya-
nate, thallium(i) sulfate, perchloric acid, and 2-methyl-2-propanol were of
the highest commercially available quality. Milli-Q-filtered (Millipore)
water was used for all solutions. 4-Methoxybenzyl alcohol (1; Aldrich) and
1-(4-methoxyphenyl)ethanol (2; Aldrich) were used as received. 1-(4-
Methoxyphenyl)-1-propanol (3), 1-(4-methoxyphenyl)-2-methyl-1-propa-
nol (4), 1-(4-methoxyphenyl)-2,2-dimethyl-1-propanol (5), threo-1-(4-
methoxyphenyl)-1,2-propanediol (6), 1-(4-methoxyphenyl)-2-methoxy-1-
propanol (erythro/threo mixture) (7), and the methyl ethers 9 ± 11 were
prepared according to literature procedures.[7, 55] a,a-[D2]-4-Methoxyben-
zyl alcohol (12) was synthesized by reduction of 4-methoxybenzoic acid
with LiAlD4.[56] The corresponding methyl ether (13) was prepared by
reaction of the alcohol 12 with methyl iodide and sodium hydride in
anhydrous THF. Both compounds had the expected 1H NMR and MS
spectra.
analogues and alkylaromatic compounds that the intrinsic
barrier for C H deprotonation correlates with the homolytic
bond-dissociation energy.[54] Thus, when a very weak base
(H2O) is involved, the effect of the much larger driving force
of C H deprotonation predominates, whereas the intrinsic
barrier that favors O H deprotonation comes into play when
the base is strong (HO ).
Further work, including theoretical calculations, aimed at a
better understanding of this mechanistic dichotomy is under
way.
Conclusions
Reaction products: 4-Methoxybenzaldehyde, 4-methoxyacetophenone,
and 4-methoxypropiophenone (Aldrich) were used as received. 4-Methoxy-
isobutyrophenone and 4-methoxyphenyl-tert-butyl ketone were prepared
according to literature procedures.[55]
The results presented above have unveiled the existence of a
hitherto unknown mechanistic dichotomy for the reactions of
a-R-substituted (R ˆ H, Me, Et) benzyl alcohol radical
cations. In water, in the absence of any other base, these
species display the expected carbon acidity of alkylaromatic
radical cations and undergo Ca H deprotonation with for-
mation of an a-hydroxyl-substituted carbon radical. However,
when HO is present, they behave as oxygen acids, and
deprotonation involves the alcoholic Ca OH bond. It is
possible that a benzyloxyl radical is formed via a radical
zwitterion that undergoes an intramolecular electron transfer
or directly (concerted proton transfer and intramolecular
electron transfer). The latter possibility seems more in line
Product analysis: g irradiation was carried out with a panorama 60Co g
source (Nuclear Engineering) at dose rates of 0.5 Gys 1. In a typical
experiment, 5 mL of an argon-saturated aqueous solution containing the
substrate (0.5 ± 5.0mm), potassium peroxodisulfate (0.2 ± 1.0mm) (sub-
strate/oxidant ratio: 2 ± 5) and 2-methyl-2-propanol (0.2 ± 1.0m) was irradi-
ated at room temperature (ꢁ258C) until 40% conversion of peroxodisul-
fate was attained. Reaction products were identified and quantitatively
determined by HPLC (comparison with authentic samples) on a Shimadzu
LC 6A instrument equipped with a Shimadzu SPD 6A UV/Vis detector
(wavelength of detection 285 nm) and a Nucleosil-5-C18 column (125 Â
4.6 mm; Macherey&Nagel). Solvent: methanol/water 1:1 (0.8 mLmin 1).
Blank experiments were performed under all conditions and showed the
presence of negligible amounts of products. Two to four experiments were
performed under all conditions with very good reproducibility (within 5%)
and mass balance.
.
‡
with the general base catalysis observed in the reaction of 5
(b ꢁ 0.4), a process that certainly involves deprotonation of
the O H group. Once formed, the benzyloxyl radical under-
goes a 1,2-H shift to give an a-hydroxyl-substituted carbon
radical. When R=H, the benzyloxyl radical can also undergo
b cleavage of the C C bond. This mechanism accounts well
for the products and is also supported by the direct
observation of the 4-methoxycumyloxyl radical in the HO -
promoted reaction of the radical cation of 4-methoxycumyl
alcohol. However, at present the possibility that side-chain
fragmentation is to some extent concerted with OH depro-
tonation cannot be excluded, at least as far as C C bond
cleavage is concerned. The observed shift from carbon acidity
in water to oxygen acidity in the presence of HO can be
interpreted in terms of the concept of hard and soft acids and
bases as well as by the Marcus theory.
Another interesting result is that when water is the only
base present, the a-OH group exerts a large favorable effect
upon the decay rate of a-alkylbenzyl alcohol radical cations if
the decay involves cleavage of the C C bond but not in the
case of C H deprotonation. By considering the structures that
contribute to the transition states of the two processes, we
have noticed that only in the former case does the presence of
the a-OH group allow significant stabilization of the tran-
sition state by hydrogen bonding. Other factors of importance
include the different electronic effects of OH and OMe as well
as the different reactivity of the fragments formed in the
cleavage.
Pulse radiolysis: The pulse-radiolysis experiments were performed with a
3 MeV van de Graaff accelerator which supplied 300 ns pulses with doses
such that 0.5 ± 3mm of radicals were produced. A temperature-controlled
continuous-flow cell was employed in all experiments. The pulse-radiolysis
setup and the methods of data processing are described elsewhere.[57]
Dosimetry was performed with N2O-saturated 10mm KSCN
.
.
1
aqueous solutions with G( OH) ˆ 6.0  10 7 molJ and e[(SCN)2 ] ˆ
7600m 1 cm
at 480 nm.[58] Experiments were performed in argon-
1
saturated aqueous solutions containing the substrate (0.1 ± 1.0mm),
peroxodisulfate (2 ± 10mm) and 2-methyl-2-propanol (0.1m). Alter-
natively, N2O-saturated aqueous solutions (pH ꢀ 3.5) containing the
substrate (0.1 ± 0.2mm) and thallium(i) sulfate (0.5 ± 2.0mm) were
employed. The pH of the solutions was adjusted with NaOH or
HClO4. The temperature of the solutions was kept constant at 25 Æ
0.28C. Rate constants were obtained by averaging 8 ± 14 values, each of
which consisted of the average of 10 ± 30 shots and was reproducibile to
within 3%.
The second-order rate constants for reaction of the radical cations with
HO (kOH ) were obtained from the slopes of the plots of the observed rates
(kobs) versus the concentration of NaOH. For these experiments the
solution containing 0.5 ± 1.0mm substrate, 10mm potassium peroxodisul-
fate, and 0.1m 2-methyl-2-propanol was saturated with argon or oxygen,
and 1mm sodium tetraborate was added to avoid undesired pH variations
upon irradiation.
The second-order rate constants for reaction of the radical cations with
different bases (kbase) were obtained from the slope of the plots of the
observed rate constants (kobs) versus the concentration of the added base.
For these experiments the ionic strength of the solution was buffered with
0.5m sodium sulfate and the pH of the solution was adjusted with NaOH or
HClO4 to 5.5, 7.0, or 8.0 for the experiments with NaOAc, NaHCO3, and
Na2HPO4, respectively. In these experiments the following base concen-
trations were employed: NaOAc, 0 ± 100mm; NaHCO3, 0 ± 70mm; Na2H-
PO4, 0 ± 50mm. In the experiments with NaOH, 1.0mm Na2HPO4 was
added to avoid undesired pH variations, and the pH of the solution was
varied between 8.5 and 10.5.
Experimental Section
Reagents: Potassium peroxodisulfate, potassium dihydrogenphosphate,
disodium hydrogenphosphate, sodium acetate, sodium hydrogencarbonate,
1792
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1792 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 6

Fragmentation Reactions
1785±1793
[34] In MeCN the smaller and less localized methoxymethyl cation has a
solvation energy that is about 15 kcalmol 1 more exergonic than that
of the benzyl cation: D. D. M. Wayner, V. D. Parker, Acc. Chem. Res.
1993, 26, 287 ± 294.
[35] It should also be considered that 6 bears a b-OH group, and it is well
known that such a structural situation can significantly increase the
rate of C C bond cleavage.[7]
[36] Alternatively, it is possible that, concerted with heterolytic C C
fragmentation (Scheme 2), the b-OH group is deprotonated, and that
this process profits from the large hydration energy of the (incipient)
proton.
Acknowledgements
E.B. and M.B. thank the Ministry of University and Scientific and
Technological Research (MURST), the National Research Council
(CNR), and the Progetto Vigoni for financial support. M.B. also thanks
the CNR for a short-term research fellowship. We are grateful to Prof. J.-M.
Saveant for discussion and suggestions and to Prof. R. A. Rossi for
providing us with a manuscript prior to publication.
.
‡
Â
[1] Biotechnology in the Pulp and Paper Industry (Ed.: K.-E. L. Eriks-
son), Springer, Berlin, 1997.
[37] E. Baciocchi, M. Mattioli, R. Romano, R. Ruzziconi, J. Org. Chem.
1991, 56, 7154 ± 7160.
[2] H. E. Schoemaker, K. Piontek, Pure Appl. Chem. 1996, 68, 2089 ±
2096, and references therein.
[3] E. Baciocchi, Acta Chem. Scand. 1990, 44, 645 ± 652.
[4] a) R. DiCosimo, H.-C. Szabo, J. Org. Chem. 1988, 53, 1673 ± 1679;
b) T. H. Fisher, S. M. Dershem, T. P. Schultz, J. Org. Chem. 1988, 53,
1504 ± 1507.
[38] L. M. Tolbert, R. K. Khanna, A. E. Popp, L. Gelbaum, L. A.
Bottomley, J. Am. Chem. Soc. 1990, 112, 2373 ± 2378; L. M. Tolbert,
Z. Li, S. R. Sirmanne, D. G. VanDerveer, J. Org. Chem. 1997, 62,
3927 ± 3930.
[39] For example, the conformational energies of monosubstituted cyclo-
1
hexanes are 1.7, 1.8, 2.2, and 4.9 kcalmol for Me, Et, iPr, and tBu,
[5] V. L. Pardini, R. R. Vargas, H. Viertler, J. H. P. Utley, Tetrahedron
1992, 48, 7221 ± 7228; V. L. Pardini, C. Z. Smith, J. H. P. Utley, R. R.
Vargas, H. Viertler, J. Org. Chem. 1991, 56, 7305 ± 7313.
[6] D. M. Camaioni, J. A. Franz, J. Org. Chem. 1984, 49, 1607 ± 1613.
[7] E. Baciocchi, M. Bietti, L. Putignani, S. Steenken, J. Am. Chem. Soc.
1996, 118, 5952 ± 5960.
[8] Preliminary communication: E. Baciocchi, M. Bietti, S. Steenken, J.
Am. Chem. Soc. 1997, 119, 4078 ± 4079.
[9] G. V. Buxton, C. L. Greenstock, W. P. Helman, A. B. Ross, J. Phys.
Chem. Ref. Data 1988, 17, 513 ± 886.
respectively: L. E. Eliel, S. H. Wilen, Stereochemistry of Organic
Compounds, Wiley, New York, 1994, p. 697.
[40] R. P. Bell, The Proton in Chemistry, 2nd ed., Cornell University Press,
Ithaca, 1973, pp. 124 ± 128.
[41] M. Eigen, Angew. Chem. Int. Ed. Eng. 1964, 3, 1 ± 19.
[42] The pKa of benzyl alcohol is around 16 ± 17.[43] The positive charge in
the ring in the radical cation could decrease the pKa to about 12 ± 13.
[43] Estimated on the basis of a pKa value of 17.4 for 2,4,6-trimethybenzyl
alcohol: N. C. Deno, J. J. Jaruzelski, A. Schriesheim, J. Am. Chem. Soc.
1955, 77, 3044 ± 3051.
[10] P. Neta, V. Madhavan, H. Zemel, R. W. Fessenden, J. Am. Chem. Soc.
1977, 99, 163 ± 164.
[44] K. U. Ingold, in Free Radicals, Vol. 1 (Ed.: J. K. Kochi), Wiley, New
York, 1973, pp. 99 ± 102.
[11] P. OꢁNeill, S. Steenken, D. Schulte-Frohlinde, J. Phys. Chem. 1975, 79,
2773 ± 2777.
[45] B. C. Gilbert, R. G. G. Holmes, H. A. H. Laue, R. O. C. Norman, J.
Chem. Soc. Perkin Trans. 2 1976, 1047 ± 1052.
[12] E. Janata, R. H. Schuler, J. Phys. Chem. 1982, 86, 2078 ± 2084.
[13] K.-D. Asmus, M. Bonifacic, P. Toffel, P. OꢁNeill, D. Schulte-Frohlinde,
S. Steenken, J. Chem. Soc. Faraday Trans. 1 1978, 74, 1820 ± 1826.
[14] H. A. Schwarz, R. W. Dodson, J. Phys. Chem. 1984, 88, 3643 ± 3647.
[15] C. Walling, G. M. El-Taliawi, C. Zhao, J. Org. Chem. 1983, 48, 4914 ±
4917.
[46] The 1,2-H shift also takes place in alcoholic solvents: P. E. Elford, B. P.
Roberts, J. Chem. Soc. Perkin Trans. 2 1996, 2247 ± 2256.
[47] S. Saebù, L. Radom, H. F. Schaefer, J. Chem. Phys. 1983, 78, 845 ± 853;
G. F. Adams, R. J. Bartlett, G. D. Purvis, Chem. Phys. Lett. 1982, 87,
311 ± 314; C. Sosa, H. B. Schlegel, J. Am. Chem. Soc. 1987, 109, 7007 ±
7015.
[16] M. E. Snook, G. A. Hamilton, J. Am. Chem. Soc. 1974, 96, 860 ± 869.
[17] W. S. Trahanovsky, J. Cramer, J. Org. Chem. 1971, 36, 1890 ± 1893;
W. S. Trahanovsky, N. S. Fox, J. Am. Chem. Soc. 1974, 96, 7968 ± 7974.
[18] L. Eberson, J. Am. Chem. Soc. 1983, 105, 3192 ± 3199.
[19] E. Baciocchi, M. Bietti, M. Mattioli, J. Org. Chem. 1993, 58, 7106 ±
7110.
[20] When 5 was oxidized under Fenton conditions the following product
distribution was observed: 83% aldehyde and 17% ketone.[16]
[21] M. Bietti, E. Baciocchi, S. Steenken, J. Phys. Chem. A 1998, 102,
7337 ± 7342.
[22] J. L. Faria, S. Steenken, J. Phys. Chem. 1992, 96, 10869 ± 10874.
[23] Based on the reasonable assumption of identical values of C C BDE
and oxidation potential[24] for the neutral parent compounds 5 and 11.
[24] E. Baciocchi, T. Del Giacco, F. Elisei, J. Am. Chem. Soc. 1993, 115,
12290 ± 12295.
Â
Â
[48] J.-M. Saveant, J. Phys. Chem. 1994, 98, 3716Ð3724; J.-M. Saveant,
Tetrahedron 1994, 50, 10117 ± 10165; R. A. Rossi, R. H. de Rossi,
Aromatic Substitution by the SRN1 Mechanism, ACS Monograph, 1983,
178, chapter 6; A. B. Pierini, J. S. Duca, Jr., J. Chem. Soc. Perkin Trans.
2 1995, 1821 ± 1828.
[49] Note that the present case refers to a radical zwitterion rather than to
a radical cation.
[50] E. Baciocchi, M. Bietti, O. Lanzalunga, S. Steenken, J. Am. Chem. Soc.
1998, 120, 11516 ± 11517.
[51] W. P. Jenks, Catalysis in Chemistry and Enzymology, McGraw-Hill,
New York, 1969, p. 234.
[52] In line with this hypothesis are the results of semi-empirical
calculations on 3,4-dimethoxybenzyl alcohol radical cation,
which show that the alcoholic H atom bears much more positive
charge than the benzylic H atoms: T. Elder, Holzforschung 1997, 51,
47 ± 56.
[25] N. B. Chapman, J. Shorter, Correlation Analysis in Chemistry, Plenum,
New York, 1979, Table 10.2.
[53] R. A. Marcus, J. Phys. Chem. 1968, 72, 891 ± 899; A. O. Cohen, R. A.
Marcus, J. Phys. Chem. 1968, 72, 4249 ± 4256.
‡
[26] The s values are considered, since it is likely that in the transition
state the resonance effect is more important than the inductive effect.
[27] P. Maslak, Top. Curr. Chem. 1993, 168, 1 ± 46.
Â
[54] A. Anne, P. Hapiot, J. Moiroux, P. Neta, J.-M. Saveant, J. Am. Chem.
Soc. 1992, 114, 4694 ± 4701; A. Anne, S. Fraoua, V. Grass, J. Moiroux,
[28] O. Takahashi, O. Kikuchi, Tetrahedron Lett. 1991, 32, 4933 ± 4936.
[29] D. M. Camaioni, J. Am. Chem. Soc. 1990, 112, 9475 ± 9483.
Â
J.-M. Saveant, J. Am. Chem. Soc. 1998, 120, 2951 ± 2958.
[55] E. Baciocchi, S. Belvedere, M. Bietti, O. Lanzalunga, Eur. J. Org.
Â
[30] A. Anne, S. Fraoua, J. Moiroux, J.-M. Saveant, J. Am. Chem. Soc. 1996,
Chem. 1998, 1, 299 ± 302.
118, 3938 ± 3945.
[56] A. D. N. Vaz, M. J. Coons, Biochemistry 1994, 33, 6442 ± 6449.
[31] D. D. M. Wayner, D. J. McPhee, D. Griller, J. Am. Chem. Soc. 1988,
[57] V. Jagannadham, S. Steenken, J. Am. Chem. Soc. 1984, 106, 6542 ±
111, 132 ± 137.
6551.
[32] D. D. M. Wayner, B. A. Sim, J. J. Dannenberg, J. Org. Chem. 1991, 56,
[58] R. H. Schuler, A. L. Hartzell, B. Behar, J. Phys. Chem. 1981, 85, 192 ±
4853 ± 4858.
199.
.
[33] The values reported in refs. [31] and [32] refer to CH(OEt)Me and
.
4-MeOC6H4CH OMe, respectively, which should be very close to
.
.
those of CH(OX)Me (X ˆ H, Me) and 4-MeOC6H4CH OH.
Received: December 1, 1998 [F 1465]
Chem. Eur. J. 1999, 5, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0506-1793 $ 17.50+.50/0
1793