Hydrogen-bonding effects on the properties of phenoxyl radicals. An EPR, kinetic, and computational study

Article abstract of DOI:10.1021/ja034963k

The effect of 1,1,1,3,3,3-hexafluoropropan-2-ol (HFP) on the properties of phenoxyl radicals has been investigated. HFP produces large variations of the phenoxyl hyperfine splitting constants indicative of a large redistribution of electron spin density, which can be accounted for by the increased importance of the mesomeric structures with electric charge separation. The conformational rigidity of phenoxyl radicals with electron-releasing substituents is also greatly enhanced in the presence of HFP, as demonstrated by the 2 kcal/mol increase in the activation energy for the internal rotation of the p-OMe group in the p-methoxyphenoxyl radical. By using the EPR equilibration technique, we have found that in phenols the O-H bond dissociation enthalpy (BDE) is lowered in the presence of HFP because it preferentially stabilizes the phenoxyl radical. In phenols containing groups such as OR that are acceptors of H-bonds, the interaction between HFP and the substituent is stronger in the phenol than in the corresponding phenoxyl radical because the radical oxygen behaves as an electron-withdrawing group, which decreases the complexating ability of the substituent. In phenols containing OH or NH₂ groups, EPR experiments performed in H-bond accepting solvents showed that the interaction between the solvent and the substituent is much stronger in the phenoxyl radical than in the parent phenol because of the electron-withdrawing effect of the radical oxygen, which makes more acidic, and therefore more available to give H-bonds, the OH or NH₂ groups. These experimental results have been confirmed by DFT calculations. The effect of HFP solvent on the reactivity of phenols toward alkyl radicals has also been investigated. The results indicated that the decrease of BDE observed in the presence of HFP is not accompanied by a larger reactivity. The origin of this unexpected behavior has been shown by DFT computations. Finally, a remarkable increase in the persistency of the α-tocopheroxyl radical has been observed in the presence of HFP.

Full text of DOI:10.1021/ja034963k

Published on Web 06/13/2003
Hydrogen-Bonding Effects on the Properties of Phenoxyl
Radicals. An EPR, Kinetic, and Computational Study
Marco Lucarini,*,† Veronica Mugnaini,† Gian Franco Pedulli,*,† and
Maurizio Guerra*,‡
Contribution from the Dipartimento di Chimica Organica “A. Mangini”, UniVersita` di Bologna,
Via S. Donato 15, I-40127 Bologna, Italy, and I.S.O.F., Consiglio Nazionale delle Ricerche,
Via P. Gobetti 101, I-40129 Bologna, Italy
Received March 3, 2003; E-mail: lucarini@alma.unibo.it; guerra@isof.cnr.it
Abstract: The effect of 1,1,1,3,3,3-hexafluoropropan-2-ol (HFP) on the properties of phenoxyl radicals has
been investigated. HFP produces large variations of the phenoxyl hyperfine splitting constants indicative
of a large redistribution of electron spin density, which can be accounted for by the increased importance
of the mesomeric structures with electric charge separation. The conformational rigidity of phenoxyl radicals
with electron-releasing substituents is also greatly enhanced in the presence of HFP, as demonstrated by
the 2 kcal/mol increase in the activation energy for the internal rotation of the p-OMe group in the
p-methoxyphenoxyl radical. By using the EPR equilibration technique, we have found that in phenols the
O-H bond dissociation enthalpy (BDE) is lowered in the presence of HFP because it preferentially stabilizes
the phenoxyl radical. In phenols containing groups such as OR that are acceptors of H-bonds, the interaction
between HFP and the substituent is stronger in the phenol than in the corresponding phenoxyl radical
because the radical oxygen behaves as an electron-withdrawing group, which decreases the complexating
ability of the substituent. In phenols containing OH or NH2 groups, EPR experiments performed in H-bond
accepting solvents showed that the interaction between the solvent and the substituent is much stronger
in the phenoxyl radical than in the parent phenol because of the electron-withdrawing effect of the radical
oxygen, which makes more acidic, and therefore more available to give H-bonds, the OH or NH2 groups.
These experimental results have been confirmed by DFT calculations. The effect of HFP solvent on the
reactivity of phenols toward alkyl radicals has also been investigated. The results indicated that the decrease
of BDE observed in the presence of HFP is not accompanied by a larger reactivity. The origin of this
unexpected behavior has been shown by DFT computations. Finally, a remarkable increase in the
persistency of the R-tocopheroxyl radical has been observed in the presence of HFP.
in catalyzing cytosine-tyrosine cross-linking in DNA damage.2
Introduction
Hydrogen bonding between tyrosyl radicals and a properly
The conversion of a phenol to a phenoxyl radical is of
considerable interest in a variety of biological and technological
processes including reduction of ribonucleotides,1 tyrosyl radi-
cal-mediated damage of proteins,2 production and utilization of
oxygen in photosystem II,3 and antioxidant activity of vitamin
E and related phenols.4
oriented protonated histidine residue has also been found to
regulate the redox behavior of the tyrosine/tyrosyl radical
couple.5,6 A model compound that allows the reversible electron
transfer between phenol and a phenoxyl radical with the
assistance of hydrogen bonding has recently been reported.7
Computational evidence has been obtained by DFT calcula-
tions that the phenoxyl/phenol hydrogen atom self-exchange
reaction can be regarded as a proton-coupled electron transfer.
The ability of phenoxyl radicals to accept hydrogen bonds
from hydrogen bond donors (HBD) is believed to be important
in changing their properties. EPR and computational studies have
shown that hydrogen bonding may have a stabilizing effect for
tyrosyl radicals in several enzymes.2,5 Hydrogen bond formation
by tyrosyl radicals has also been suggested to play a key role
(3) (a) Campbell, K. A.; Peloquin, J. M.; Diner, B. A.; Tang, X.-S.; Chisolm,
D. A.; Britt, D. A. J. Am. Chem. Soc. 1997, 119, 4787-4788. (b) O’Malley,
P. J. J. Am. Chem. Soc. 1998, 120, 11732-11737. (c) Tommos, C.;
Babcock, G. T. Acc. Chem. Res. 1998, 31, 18-25. (d) Proshlyakov, D. A.;
Pressler, M. A.; DeMaso, C.; Leykam, J. F.; DeWitt, D. L.; Babcock, G.
T. Science 2000, 290, 1588-1591. (e) Maki, T.; Araki, Y.; Ishida, Y.;
Onomura, O.; Matsumura, Y. J. Am. Chem. Soc. 2001, 123, 3371-3372.
(4) (a) Burton, G. W.; Joyce, A.; Ingold, K. U. Arch. Biochem. Biophys. 1982,
221, 281-290. (b) Burton, G. W.; Doba, T.; Gabe, E. J.; Hughes, I.; Lee,
F. L.; Prasad, L.; Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053-7065.
(c) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19, 194-201.
(5) (a) van Dam, P. J.; Willems, J.-P.; Schmidt, P. P.; Po¨tsch, S.; Barra, A.-L.;
Hagen, W. R.; Hoffman, B. M.; Andersson, K. K.; Gra¨slund, A. J. Am.
Chem. Soc. 1998, 120, 5080-5085. (b) Engstro¨m, M.; Himo, F.; Gra¨slund,
A.; Minaev., B.; Vahtras, O.; Agren, H. J. Phys. Chem. A 2000, 104, 5149-
5153.
† Universita` di Bologna.
‡ I.S.O.F.
(1) (a) Sjo¨din, M.; Styring, S.; Åkermark, B.; Sun, L.; Hammarstro¨m, L. J.
Am. Chem. Soc. 2000, 122, 3932-3936. (b) Stubbe, J. A.; van der Donk,
W. A. Chem. ReV. 1998, 98, 705-762. (c) Siegbahn, P. E. M.; Eriksson,
L.; Himo, F.; Pavlov, M. J. Phys. Chem. B 1998, 102, 10622-10629. (d)
Siegbahn, P. E. M.; Blomberg, M. R. A.; Crabtree, R. H. Theor. Chem.
Acc. 1997, 97, 289-300.
(2) Ban, F.; Lundqvist, M. J.; Boyd, R. J.; Eriksson, L. A. J. Am. Chem. Soc.
2002, 124, 2753-2761.
9
8318
J. AM. CHEM. SOC. 2003, 125, 8318-8329
10.1021/ja034963k CCC: $25.00 © 2003 American Chemical Society

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
Scheme 1
phenoxyl radicals experiencing small or negligible interactions
at the radical oxygen with HB donors. Actually, in the case of
hydrogen bond acceptor (HBA) solvents, the absence of
hydrogen bonding to the OH group of 2,6-di-tert-butyl phenols
was demonstrated by the observation in the 3200-3800 cm-1
region of their FT-IR spectrum of only the peak due to the free
OH group both in isooctane and in γ-valerolactone, that is, in
a very poor and in a strong HBA solvent.11 Moreover, the
absolute rate constants for the reaction of alkyl radicals with
2,6-di-tert-butylphenol in toluene and γ-valerolactone showed
a negligible kinetic solvent effect, this indicating that hydrogen
bond formation is unimportant.11,12 Experimental evidence
supporting the absence of hydrogen-bonding interactions of the
phenoxyl radicals from the 2,6-di-tert-butyl substituted phenols
of Scheme 1 with HBD solvents will be discussed later in the
paper.13
The effect of a HBD solvent was investigated using a
fluorinated alcohol such as 1,1,1,3,3,3-hexafluoropropan-2-ol
(HFP), characterized by a large R2H value (0.771) indicative of
an extremely strong hydrogen bond donor character, and by a
This mechanism requires the formation of a hydrogen bond and
therefore cannot operate in self-exchange reactions similar to
those in the benzyl/toluene hydrogen exchange.8
â2 value very close to zero (0.03),14,15 implying that no
H
interaction should occur with the hydrogen of OH or NH2
groups.
The stabilization of the semiquinone radicals derived from
catechols, due to intramolecular hydrogen bonding, is considered
to be the main effect controlling the excellent behavior of natural
catechols as antioxidants.9
Results and Discussion
HB-Effects on the hfs Constants of Phenoxyl Radicals.
The phenoxyl radicals were generated within the cavity of an
EPR spectrometer, by continuous UV irradiation of deoxygen-
ated di-tert-butyl peroxide/benzene solutions of one of the
phenols shown in Scheme 1, either in the absence or in the
presence of HFP. With a few representative phenols, spectra
were also recorded by using acetonitrile (ACN) or ethyl acetate
as solvents to show the effects of HB acceptors on the hyperfine
splitting (hfs) constants (see Table 1). An examination of this
table shows that the solvent dependence of the proton splittings
is moderate in the phenoxyl radicals from the three 2,6-di-tert-
butyl substituted phenols BHT, BHA, and BHQ,16 suggesting
that hydrogen bonding at the radical oxygen is unimportant.
On the other hand, the spectral parameters of the radicals from
the five unhindered phenols of Scheme 1 experience a very big
Finally, several DFT computational studies have shown that
the capability of phenoxyl radicals to accept hydrogen bonds
from hydrogen bond donors (HBD) is important in changing
their EPR spectroscopic properties.10
Despite the large interest concerning the effects of hydrogen
bonding on the properties of phenoxyl radicals, no quantitative
studies of such effects have been reported so far. Here we
present an EPR investigation carried out to provide a quantitative
description of the factors governing the formation of hydrogen
bonding in phenoxyl radicals and of the effects of this interaction
on their physical and chemical properties. A density functional
theory (DFT) study is also reported, which allows a better
rationalization of the experimental thermodynamic and kinetic
results.
We have investigated the series of phenols shown in Scheme
1. Some of them are sterically unhindered in the proximity of
the OH group and therefore should give rise to phenoxyl radicals
easily complexed by HBD solvents. Some others, containing
tert-butyl groups in both ortho positions, are expected to give
(11) Franchi, P.; Lucarini, M.; Pedulli, G. F.; Valgimigli, L.; Lunelli, B. J. Am.
Chem. Soc. 1999, 121, 507-514.
(12) (a) Ingold, K. U. Can. J. Chem. 1960, 38, 1092-8. (b) Ingold, K. U.; Taylor,
D. R. Can. J. Chem. 1961, 39, 481-7. (c) Ingold, K. U.; Taylor, D. R.
Can. J. Chem. 1961, 39, 471-80.
(13) Very recently, it has been reported that in sterically crowded 2,6-di-tert-
butylated phenols a small interaction between the phenolic hydroxyl group
and HBA molecule is possible only when the orientation of the OH group
is perpendicular to the aromatic plane. Litwinienko, G.; Megiel, E.; Wojnicz,
M. Org. Lett. 2002, 4, 2425-2428.
(6) (a) Uhlin, U.; Eklund, H. Nature 1994, 370, 533-539. (b) Nordlund, P.;
Eklund, H. J. Mol. Biol. 1993, 232, 123-164. (c) Sjoberg, B.-M. Structure
1994, 2, 793-796. (d) Siegbahn, P. E. M.; Blomberg, M. R. A.; Crabtree,
R. H. Theor. Chem. Acc. 1997, 97, 289-300. (e) Rova, U.; Goodtzova,
K.; Ingemarson, R.; Behravan, G.; Gra¨slund, A.; Thelander, L. Biochemistry
1995, 34, 4267-4275. (f) Schmidt, P. P.; Rova, U.; Thelander, L.; Gra¨slund,
A. J. Biol. Chem. 1998, 273, 21463-21472. (g) Un, S.; Atta, M.; Fontecave,
M.; Rutheford, A. W. J. Am. Chem. Soc. 1995, 117, 10713-10719. (h)
Fang, Y.; Liu, L.; Feng, Y.; Li, X.-S.; Guo, Q.-X. J. Phys. Chem. A 2002,
106, 4669-4678.
(14) â2H and R2H represent a general, thermodynamically related scale of solute
H
hydrogen bond basicities and acidities in CCl4, respectively. â2 values
range in magnitude from 0.00 for a non HBA solvent such as an alkane to
1.00 for hexamethylphosphoric triamide (HMPA). Values of R2H range from
0.00 (e.g., alkanes) to nearly 1.0 for strong acids (CF3COOH ) 0.951). (a)
Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J. J.;
Taylor, P. J. J. Chem. Soc., Perkin Trans. 2 1989, 699-710. (b) Abraham,
M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J. J. Chem.
Soc., Perkin Trans. 2 1990, 521-529.
(7) Maki, T.; Araki, Y.; Ishida, Y.; Onomura, O.; Matsumura, Y. J. Am. Chem.
Soc. 2001, 123, 3371-3372.
(8) (a) Mayer, J. M.; Hrovat, D. A.; Thomas, J. L.; Borden, W. T. J. Am.
Chem. Soc. 2002, 124, 11142-11147. (b) Kirby, J. P.; Roberts, J. A.;
Nocera, D. G. J. Am. Chem. Soc. 1997, 119, 9230-9236.
(15) Eberson, L.; Hartshorn, M. P.; Persson, O.; Radner, F. Chem. Commun.
1996, 2105-2111 and references therein.
(16) In the case of BHQ, the EPR spectral parameters of the observed radicals
were consistent with the phenoxyl radical centered on the hindered oxygen
atom. Actually, the spectrum of the phenoxyl radical showed coupling of
the unpaired electron with the two meta-protons (0.88 G) and with the
remaining hydroxyl proton (1.40 G). The formation of the other possible
species, with the radical centered on the oxygen in position 4, can be
discarded because larger splittings (ca. 6.5 G) from the protons in 3 and 5
would be expected in this case.
(9) (a) Barclay, L. R. C.; Edwards, C. E.; Vinqvist, M. R. J. Am. Chem. Soc.
1999, 121, 6226-6231. (b) Lucarini, M.; Pedulli, G. F.; Mugnaini, V. J.
Org. Chem. 2002, 67, 928-931.
(10) (a) O’Malley, P. J. J. Phys. Chem. A 1998, 102, 248-253. (b) Chipman,
D. M. J. Phys. Chem. A 1999, 103, 11181-11187. (c) Chipman, D. M. J.
Phys. Chem. A 2000, 104, 11816-11821. (d) Langella, E.; Improta, R.;
Barone, V. J. Am. Chem. Soc. 2002, 124, 11531-11540.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8319

A R T I C L E S
Lucarini et al.
Table 1. Room-Temperature (293 K) Hyperfine Splitting
Constants (Gauss ) 0.1 mT) for the Investigated Phenoxyl
Radicals in Different Solvents and Equilibrium Constants and Free
Energy Changes for Reaction 1
benzene/
HFP
∆G °/
kcal
mol-1
1
benzene ACN CH COOEt (70:30)
K /M-1
3
1
1a
2a
3a
4a
a(3H)
a(6H)
a(2H)
10.89
6.33
1.71
11.38 21.0 ( 1.8 -1.77
5.90
1.38
a(2H)o
a(2H)m 0.80
a(3H)
5.75
4.90 43.6 ( 4.1 -2.20
0.13
2.10
1.78
a(H)o
a(H)m
a(H)OH
5.22
0.93
1.40
1.00a
3.64a
4.68a
4.68a
5.07
0.83
1.51
4.48 75.2 ( 8.1 -2.51
0.54
1.73
1.30b
3.23b
5.00b
4.97b
Figure 1. Experimental dependence on the HFP concentration of a(H)ortho
for the phenoxyl radical from 2a (b) and of the smaller of the two a(CH3)ortho
(9) for the radical from 5a, at 293 K in benzene. The lines represent the
theoretical dependence of aH on the alcohol concentration calculated by
taking into account only the formation of hydrogen bonding and were
obtained by numerical fitting of the experimental data by using eq 2. K1
and abound were determined as adjustable parameters.
a(2H)
a(2H)
a(2H)
a(N)
5a
a(3H)
a(3H)
a(3H)
a(2H)
5.92
4.56
0.86
1.44
5.87
4.42
0.77
1.34
5.00 64.2 ( 5.2 -2.42
3.47
∼0
0.30
positions including the meta carbons, consistent with the
canonical structures A-D (Scheme 2). In the higher energy
polar structures E-H, on the other hand, the expected spin
density pattern is complementary to that predicted for structures
A-D, that is, positive for the meta and ipso positions and
negative for the positions ortho and para to the oxygen radical
center. The increased solvent polarity and the formation of a
hydrogen bond to the oxygen radical center induce a stabilization
of the polar structures E-H and opposite spin densities at the
ortho and meta carbons with respect to A-D with a consequent
reduction of the hyperfine splittings at the nuclei in these
positions.17 Figure 1 shows, as an example, the experimental
dependence on the HFP concentration of the ortho hyperfine
splittings in the phenoxyl radical derived from 2a and 5a in
benzene at 293 K.
BHT a(3H)
a(2H)
11.22 11.47
11.33
1.66
11.56
1.51
1.67
1.63
BHA a(3H)
a(2H)
1.55
0.95
1.62
0.85
1.57
0.92
1.70
0.68
BHQ a(H)
a(2H)
1.40
0.88
1.44
0.80
1.43
0.83
1.54
0.70
a Benzene/tert-butyl alcohol 60/40 (v/v). b Benzene/tert-butyl alcohol/
HFP 42/28/30 (v/v).
Scheme 2
This dependence is a clear indication of the formation of a
hydrogen-bonded complex between phenoxyl radical and HFP.
Because the formation and breaking of the hydrogen bonds are
fast on the EPR time scale, the experimental spectrum represents
the concentration-weighted average of the spectra due to the
free and hydrogen-bonded species, the molar fraction of the
latter one increasing with increasing HFP concentration.
It has been recently shown in the case of nitroxides18 that
the variation of the experimental hyperfine splittings is deter-
mined mainly by the hydrogen-bonding interaction at low HBD
concentration. Thus, by neglecting the macroscopic electrostatic
contribution on aH, we could determine the equilibrium constants
K1 (see eq 1) for the formation of the hydrogen-bonded complex
quantitatively from the EPR data by using eq 2.
change in the mixture benzene/HFP and only a small variation
in ACN or ethyl acetate. The behavior of the radicals from 1a-
5a can be explained in both cases in terms of the larger
importance in polar solvents of the higher energy mesomeric
structures (E-H) involving charge separation (see Scheme 2).
The different magnitude of the solvent effect for HBA and
HBD solvents is due to the fact that in the former case the small
preference for polar structures is only determined by the increase
of the dielectric constant with respect to benzene, while in HFP
the specific hydrogen-bonding interaction, strongly stabilizing
all structures bringing a negative charge on the oxygen atom,
increases significantly the contribution of the structures (E-H)
to the mesomeric system.
The decrease of the splittings at the aromatic protons or at
the methyl protons in the ortho and meta positions, observed
in HFP, can be explained as follows. These radicals are known
to have an odd alternant spin population pattern with large
positive spin densities at the oxygen and at the ortho and para
positions and small negative spin densities at the other ring
K1
A• + HFP y
\
z A•‚‚‚‚‚‚‚HFP
(1)
(2)
aH ) afreeXfree + aboundXbound
Here afree and abound represent the values of the hyperfine
splitting for the free and hydrogen-bonded radical species,
(17) The increase of the hfs constants at the nuclei of the para substituent
observed in HFP can be explained on the basis of the polar mesomeric
structure H involving spin delocalization on the substituent.
(18) Franchi, P.; Lucarini, M.; Pedrielli, P.; Pedulli, G. F. ChemPhysChem 2002,
3, 789-793.
9
8320 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
Figure 3. Plots of log(k) versus 1/T for the rotation of the p-methoxy group
in the radical from 2a. Rate constants in toluene (9) and in a HFP-toluene
mixture 30:70 v/v (b) are reported.
Figure 2. EPR spectra of the phenoxyl radical from 2a recorded at 293 K
in toluene (a) and in a mixture toluene/HFP (70:30) (b). In the lower
spectrum, selective line broadening effects are observed.
rate constant for the rotation about this bond was determined
by simulating the exchange-broadened EPR spectrum recorded
in toluene in the presence (a(H) ≈ 0 G, a(H) ) 0.23 G, a(3H)
) 2.29 G, a(H) ) 3.87 G, a(H) ) 5.33 G; T ) 232 K) and in
the absence (a(2H) ) 0.46 G, a(3H) ) 1.93 G, a(H) ) 4.56,
a(H) ) 6.02 G; T ) 218 K) of 30% (v/v) HFP using well-
established procedures21 based on the density matrix theory.22
By repeating these measurements at different temperatures
(see Figure 3), we obtained the activation parameters for the
internal rotation of the methoxy group about the C-O bond as
Ea ) 8.32 ( 0.40 kcal/mol and log A ) 14.5 ( 0.5 in toluene
solution and Ea ) 10.37 ( 0.45 kcal/mol and log A ) 15.2 (
0.5 in the mixture containing 30% HFP. Thus, the activation
energy undergoes an increase of about 2 kcal/mol in the presence
of the fluorinated alcohol.
This behavior can be explained as follows: when the O-Me
bond of the methoxy group lies on the same plane of the benzene
ring, maximum overlap between the lone pair on oxygen and
the π-aromatic electrons occurs with a strong conjugative
stabilization of the ground state of the phenoxyl radical. On
the other hand, when the methoxy group is rotated by 90°, its
electron-donating effect is reduced, leading to a weakening of
the hydrogen bond interaction. Although solvent effects on the
rotational barriers of diamagnetic molecules such as amides23
have been previously reported, to the best of our knowledge,
this is the first example of solvent effects on the rotational
energy barrier in an organic free radical.
respectively, and X is the corresponding molar fraction. Figure
1, reporting as a solid line the dependence of aH on the HFP
concentration calculated by using eq 2, shows that the experi-
mental data for 2a and 5a can be fitted very accurately at low
concentration of the fluorinated alcohol. The values of the
equilibrium constants for hydrogen-bonding formation, K1,
obtained from these fittings are reported in Table 1 together
with the corresponding ∆G° variation calculated at 293 K. By
assuming an entropy change for the formation of hydrogen
bonding18,19 of ca. 11.0 cal mol-1 K-1, we could estimate the
corresponding enthalpy change as 5.0, 5.4, 5.7, and 5.6 kcal/
mol for the phenoxyl radicals derived from 1a, 2a, 3a, and 5a,
respectively.20
Thermochemical data for the complexation by HFP of the
phenoxyl radical from 4a have not been determined because of
its low solubility in benzene.
HB-Effects on the Rotational Barrier of the p-Methoxy-
phenoxyl Radical. The formation of a hydrogen bond between
HFP and a phenoxyl radical leads to an increase of the relative
weight of the polar mesomeric structures (E-H), three of which
contain a formal double bond between the para carbon and the
substituent. It is therefore conceivable that complexation of
phenoxyl radicals by HBD solvents should induce a larger
conformational rigidity about this bond, providing additional
evidence for the formation of intermolecular hydrogen bonds.
We therefore examined the room-temperature EPR spectrum
of the p-methoxyphenoxyl radical in the presence of increasing
amounts of HFP. Actually, besides the change in the values of
the hyperfine splittings described above, a selective line width
broadening not visible in the absence of the fluorinated alcohol
started to be observed (see Figure 2). Such a line width effect
is indicative of restricted rotation about the Ar-OMe bond that
modulates the ortho-proton hyperfine splittings. The absolute
It is also remarkable that no line broadening effects have been
detected in the phenoxyl radical from 2,6-di-tert-butyl-4-
methoxyphenol (BHA) down to -70 °C in toluene (see
Supporting Information). This provides additional evidence that
phenoxyl radicals sterically crowded in the proximity of the
OH group are not capable of being complexed by HFP through
an intermolecular hydrogen bond.
Solvent Effects on the O-H BDE Values. The EPR radical
equilibration technique has been very successfully applied by
(19) (a) de Heer, M. I.; Korth, H. G.; Mulder, P. J. Org. Chem. 1999, 64, 6969-
6975. (b) Khafizov, F. T.; Breus, V. A.; Solomonov, B. N.; Kiselev, O. E.;
Konovalov, A. I. Dokl. Akud. Nauk SSRS 1988, 303, 916-919.
(20) The reported strengths (∆H, kcal/mol) of the hydrogen bonds between HFP
and some acceptors follow: 5.9 for acetone, 6.3 for tetrahydrofuran, 5.9
for acetonitrile, and 8.4 for pyridine.15 The strength for the H-bond between
HFP and the radical 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) is
instead 5.45.18
(21) Guerra, M.; Pedulli, G. F.; Tiecco, M.; Martelli, G. J. Chem. Soc., Perkin
Trans. 2 1974, 562-567.
(22) (a) Freed, J. H.; Fraenkel, G. K. J. Chem. Phys. 1963, 39, 326-348. (b)
Hudson, A.; Luckhurst, G. R. Chem. ReV. 1969, 69, 191-225.
(23) (a) Stewart, W. E.; Siddall, T. H., III. Chem. ReV. 1970, 70, 517-51. (b)
Wiberg, K B.; Rablen, P. R.; Rush, D. J.; Keith, T. A. J. Am. Chem. Soc.
1995, 117, 4261-4270.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8321

A R T I C L E S
Lucarini et al.
us to measure BDE (bond dissociation enthalpies) values of the
O-H bond in substituted phenols and of the N-H bond in
aromatic amines.24 This method involves the simultaneous
generation of two radicals (eq 3) by continuous UV irradiation,
within the cavity of an EPR spectrometer, of di-tert-butyl
peroxide/benzene solutions of two substrates, AH and BH,
present at known concentrations.
hV
AH + BH Me COOCMe 8 A• + B•
(3)
3
3
The resulting EPR spectra provide the concentration ratio of
the two substrate-derived radicals [A•]/[B•], which is solely
determined by the equilibrium of eq 4, if the H-atom transfer
between radicals and substrates is sufficiently fast. Thus, by
knowing the initial concentrations of the two phenols and the
molar ratio [A•]/[B•], determined by integration or by theoretical
simulation of the EPR spectrum, we could obtain the equilibrium
constant K4 and the ∆BDE for the couple AH and BH.
K4
AH + B• y
\
z A• + BH
(4)
(5)
Figure 4. Central part of the room-temperature (293 K) EPR spectra
observed under continuous irradiation of a benzene solution of di-tert-butyl
peroxide (0.2 M), BHT (0.15 M), and 1a (1.57 M) in the absence (a) and
in the presence (b) of 30% HFP. The six inner lines of the EPR spectrum
of the radical from BHT are marked with an asterisk. Spectrum c is the
theoretical simulation of (b).
∆G° ) -RT ln(K4)
The accuracy of the EPR method relies on the fact that even
relatively large errors in the measurement of radical concentra-
tions and therefore of the equilibrium constant, K4, give rise to
small errors in the ∆BDEs because of the logarithmic relation
connecting these two quantities (eq 5).
Scheme 3
The dependence of the BDE value of a given phenol on the
nature of the solvent can, in principle, be investigated with the
EPR radical equilibration technique, provided an appropriate
reference compound (BH) experiencing a negligible interaction
with the solvent is available. In this case, any interaction of the
investigated phenol/phenoxyl couple (AH/A•) with the solvent
should determine a change in the equilibrium constant easily
detectable in the corresponding EPR spectra.
Because we were interested in studying the properties of
phenoxyl radicals complexed with hydrogen bond donors
(HBDs), an appropriate reference compound not interacting with
HBD solvents was required. The most suitable candidates were
represented by 2,6-di-tert-butyl substituted phenols because the
steric protection around the oxygen atom by the bulky tert-
butyl groups was expected to minimize the formation of
hydrogen bonds with HBDs in the phenoxyl radicals.25 Experi-
mental evidence, based on the solvent dependence of the hfs
constants and of the rotational barriers of para methoxy groups,
that has been reported in the previous sections shows that this
is actually the case.
unhindered phenol (1a-5a, see Scheme 1) and BH is a 2,6-
di-tert-butyl substituted phenol (BHT, BHA, and BHQ). We
first examined the equilibria involving phenols with the same
substituent at the para position, that is, the couples trimethyl-
phenol (1a)/BHT, 2,5-di-tert-butylhydroquinone (3a)/BHQ, and
R-tocopherol (5a)/BHA in the presence of variable amounts of
HFP. These couples of phenols were chosen to rule out any
interaction between the fluorinated alcohol and the ring para
substituent so that any change in the equilibrium constant is
due only to the interaction between HFP and the phenol/
phenoxyl oxygen in the unhindered species (see Scheme 3 for
the equilibrium between R-tocopherol (5a) and BHA).
Figure 4 reports the central part of EPR spectra obtained
equilibrating 1a (1.57 M) with BHT (0.15 M) in benzene and
in the presence of 30% HFP. These spectra indicate that when
adding HFP the ratio between the two phenoxyl radicals
changes, with the trimethylphenoxyl radical becoming the
predominant species. Figure 5 shows that, in all three cases, by
increasing the alcohol concentration the equilibrium gradually
becomes less endothermic until reaching a constant ∆G° value
at a concentration of ca. 15% (v/v) HFP.
The first step of our investigation was focused on the
influence of HFP on the equilibrium of eq 4, where AH is an
(24) (a) Lucarini, M.; Pedulli, G. F.; Cipollone, M. J. Org. Chem. 1994, 59,
5063-5070. (b) Lucarini, M.; Pedrielli, P.; Pedulli, G. F.; Cabiddu, S.;
Fattuoni, C. J. Org. Chem. 1996, 61, 9259-9263. (c) Lucarini, M.; Pedrielli,
P.; Pedulli, G. F.; Valgimigli, L.; Gigmes, D.; Tordo, P. J. Am. Chem.
Soc. 1999, 121, 11546-11553. (d) Brigati, G.; Lucarini, M.; Mugnaini,
V.; Pedulli, G. F. J. Org. Chem. 2002, 67, 4828-4832.
(25) The effect of tert-butyl alcohol26 and ethyl acetate27 on the equilibrium
between a 2,6-di-tert-butyl substituted phenol and an unhindered phenol
has been previously investigated. On changing the solvent from benzene
to HBA solvent, the BDEs were found to increase by ca. 2 kcal/mol for
phenols without ortho substituents, while in 2,6-di-tert-butyl substituted
phenols they seem to be substantially unaffected. This behavior has been
interpreted by admitting that the BDE increase observed in HBA solvents
is essentially due to the solvation of the hydroxylic hydrogen which
stabilizes the phenol, leaving the energy of the phenoxyl radical unaltered.
Table 2 reports the equilibrium constants and the ∆G° values
for reaction 4 when HFP is present at 30%. The decrease of
endothermicity when the alcohol concentration is increased is
an indication that the equilibrium is shifted toward the unhin-
dered phenoxyl radical. Because, as discussed previously, no
9
8322 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
Scheme 4
Scheme 5
Figure 5. ∆∆G° variations for equilibrium 4 as a function of HFP content.
Table 2. Room-Temperature (293 K) Equilibrium Constants and
Free Energy Changes of Reaction 4
that arising from the interaction between the alcoholic OH and
the lone pairs on the ethereal oxygen of BHA. The experimental
data indicate (see Table 2) that in the presence of 30% HFP the
equilibrium is shifted toward the BHT phenoxyl radical by 1.19
kcal/mol. With a weaker hydrogen bond donor such as tert-
butyl alcohol, the equilibrium is shifted toward the BHT
phenoxyl radical by only 0.20 kcal/mol. The ∆∆G° values
indicate that hydrogen bond formation between alcohols and
the ethereal oxygen stabilizes BHA more than the corresponding
phenoxyl radical, leading to a strengthening of the associated
OH bond. The reason is that, on passing from phenol to
phenoxyl radical, the hydroxyl group changes its effect from
donor to acceptor when becoming a radical oxygen center. Thus,
the lone pair electrons of the conjugated OMe group are much
less available to accept hydrogen bonding in the phenoxyl radical
than in the phenol. It is worth pointing out that these results
are consistent with the larger solvent effect found with BHA
with respect to BHT when studying the hydrogen transfer
reaction between phenols and DPPH in tert-butyl alcohol.9a
When 2a and BHT are equilibrated in the presence of 30%
HFP, a smaller than expected ∆∆G° value (0.87 kcal/mol) was
found. This can be interpreted as the result of two opposite
effects (see Scheme 5): one being the preferential complexation
by the fluorinated alcohol of the radical oxygen in the phenoxyl
radical from 2a with respect to that of the hydroxyl oxygen of
2a, and the other is the preferential complexation of the ethereal
oxygen in 2a over that in the corresponding radical. The former
interaction induces a decrease of the O-H BDE value, while
the latter one has the opposite effect. The positive ∆∆G° is,
however, indicative that the dominant effect is the formation
of a stronger hydrogen bond with the oxygen of the phenoxyl
radical from 2a.
∆G °/
∆∆G °/a
4
4
AH
1a
BH
solvent
K
kcal mol-1 kcal mol-1
4
BHT Bzb
0.0578 ( 0.006
0.165 ( 0.018
1.66
1.06
0.60
0.87
1.68
Bz + 30% HFP
BHT Bz
Bz + 30% HFP
BHQ Bz
Bz + 30% HFP
2a
3a
3a
0.0462 ( 0.004
0.206 ( 0.017
1.79
0.92
0.0228 ( 0.003
0.409 ( 0.046
2.20
0.52
BHT Bz
1.49 ( 0.13
22.8 ( 2.6
8.70 ( 0.74
-0.23
-1.82
-1.26
1.03
Buc
1.59e
Bz + 30% HFP
4a
5a
BHT BuBzd
Bu/Bz + 30% HFP
55.0 ( 4.1
94.8 ( 9.7
-2.16
-2.65
0.49
1.54
BHA Bz
Bz + 30% HFP
1.18 ( 0.10
16.72 ( 1.8
-0.10
-1.64
BHA BHT Bz
Bu
105.3 ( 8.9
74.5 ( 8.3
13.61 ( 1.2
-2.71
-2.51
-1.52
-1.19
-0.20e
Bz + 30% HFP
BHQ BHT Bz
Bu
-2.41
-3.11
0.70e
a ∆∆G4° ) ∆GBz° - ∆GHFP 30%°. b Bz ) benzene. c Bu ) tert-butyl
alcohol. d Bu/Bz ) tert-butyl alcohol/benzene 60/40 (v/v). e ∆∆G4° )
∆GBz° - ∆GBu°.
interaction between HFP and the oxygen atom in the 2,6-di-
tert-substituted phenols should occur, the large solvent effect
on K4, that is, a 3-, 18-, and 14-fold increase for the couples
1a/BHT, 3a/BHQ, and 5a/BHA, respectively, can be explained
by assuming that the interaction with HFP is stronger in the
unhindered phenoxyl radical rather than in the corresponding
phenol and that the difference in the free energy of solvation is
equal to the ∆∆G° values measured by the EPR experiments.
A comparison of the three measured ∆∆G° values indicates
that the differences in the ability of the phenol and the
corresponding phenoxyl radical to accept a hydrogen bond from
HFP increase by increasing the electron-donating nature of the
para ring substituent (Me, OMe, OH). This behavior can be
related to the increased electron density at the phenoxyl oxygen
when the para substituent is an OH or an OMe group.
Equilibration studies have also been done on the couple
p-aminophenol/BHT. In this case, a quantitative comparison with
the previous experiments cannot be made because the equilib-
rium was investigated in the presence of a large amount of a
HBA solvent such as tert-butyl alcohol due to the very low
solubility of 4a in benzene. However, also in this case, in the
presence of 30% HFP the equilibrium is shifted toward the
p-aminophenoxyl radical by 0.49 kcal/mol.
To obtain from the equilibrium constants the BDEs of the
phenols in the presence of HFP, the entropy change ∆S° for
the hydrogen transfer reaction 4 should be negligible. Because
the validity of this assumption is not warranted, especially when
To study possible solvent effects on the para-substituent, we
have also investigated the effects of either HFP or tert-butyl
alcohol on the equilibrium between BHT and BHA (see Scheme
4). Because, in this case, no interaction between the alcohols
and the phenolic oxygen is expected due to the presence of the
ortho bulky substituents, the only appreciable effect should be
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8323

A R T I C L E S
Lucarini et al.
Table 3. Bond Dissociation Enthalpies (kcal/mol) in Different
Scheme 6
Solvents
phenol
benzenea
tert-butanol
Bz/HFP 30%
1a
2a
3a
82.7b
82.8
80.8
83.6d
85.0d
79.2
82.1
81.8
79.8
4a
5a
BHT
BHQ
BHA
78.3
78.2
81.0
78.6c
78.3
79.3d
81.0e
77.8
81.0e
79.3
79.5
78.5
a From ref 24b. b Uncertainty in the BDE values is (0.15 kcal/mol. c This
Table 4. Absolute Rate Constants (298 K) for the Abstraction of
the Phenolic Hydrogen Atom by Primary Alkyl Radicals Measured
in Different Solvents
work. d From ref 26. e Assumed.
the equilibrating couple forms and destroys hydrogen bonds in
different molecular environments during hydrogen transfer, we
studied the temperature dependence of the equilibrium constant
K4 in the case of the couple 1a/BHT in the temperature range
259-353 K in HFP/toluene 30/70 (v/v) (see Supporting
Information). The temperature dependence of the equilibrium
constants shows that K4 changes only slightly with temperature
(∆H° ) -1.28 ( 0.25 kcal mol-1 and ∆S° ) -0.66 ( 0.64
cal mol-1 K-1), thus indicating that the entropic contribution is
so small that it can be safely neglected. If the entropic term is
neglected, the OH bond dissociation enthalpy in the presence
of HFP can be calculated by using the ∆G° values reported in
Table 2. These values are collected in Table 3 together with
the values determined by EPR equilibration in tert-butyl alcohol
(for the discussion of these data, see Computational Details
section).
HB-Effects on the Reactivity of Phenoxyl Radicals and
of the Parent Phenols. Complexation by HFP of phenoxyl
radicals has important effects also on their persistency. Actually,
the intensity of EPR spectra is generally higher in the presence
of the fluorinated alcohol, this effect being particularly evident
in the case of the radical derived from R-tocopherol. In the latter
case, the EPR signal persists at room temperature for several
hours, demonstrating the enhanced lifetime of the hydrogen-
bonded R-tocopheroxyl radical (see Supporting Information).
Because the decay of R-tocopheroxyl radicals takes place by a
bimolecular disproportionation reaction, to occur in the presence
of HFP it requires the desolvation of at least one radical
molecule with a consequent reduced reactivity. This behavior
is similar to that found for the R-tocopheroxyl radical in the
presence of magnesium ions by Nakanishi et al.28
phenol
solvent
10-5 kH (M-1 s-1
)
R-tocopherol (5a)
benzene
6.0 ( 1.5
0.32 ( 0.04
0.20 ( 0.03
5.9 ( 1.2
tert-butanol
ethyl acetate
HFP/benzene (1/9 v/v)
p-methoxyphenol (2a)
benzene
2.3 ( 0.42
0.07 ( 0.008
2.0 ( 0.48
tert-butanol (4.8 M)
HFP/benzene (1/9 v/v)
by photolyzing oxygen-free solutions of the corresponding
bromide in the presence of hexabutyldistannane that was left
to react with the phenols (see Scheme 6).11
Experimental conditions were chosen to avoid significant
consumption of the phenols during the reaction. The reaction
products were analyzed by means of GC, and the rate constants
for hydrogen abstraction, kH, obtained by using the equation in
Scheme 6 are reported in Table 4.32 Kinetic experiments were
performed with R-tocopherol and p-methoxy phenol in benzene,
tert-butyl alcohol (HBA/HBD solvent), and HFP. As expected,
the rate constant for hydrogen abstraction decreased strongly
in the presence of the HBA tert-butyl alcohol11 due to the
formation of a hydrogen bond between solvent and the hydroxyl
proton with a consequent preferential stabilization of the starting
phenol with respect to the phenoxyl radical.33
This reduced reactivity toward alkyl radicals is consistent with
the observed increase of the O-H BDE for 2a and 5a in tert-
butyl alcohol (see Table 3).25-27 On the other hand, in the
presence of a strong hydrogen bond donor such as HFP the
reactivity is almost coincident to that in benzene, this being in
disagreement with the equilibration experiments that indicated
a reduced strength of the phenolic OH bond. To better
understand the reason for this surprising discrepancy between
the thermodynamic and kinetic experiments, we performed some
DFT calculations.
To verify if hydrogen bonding on phenoxyl radicals affects
also the transfer of a hydrogen atom from phenols, we decided
to measure the reactivity of a substituted phenol toward primary
alkyl radicals in the presence of HFP at 298 K by the radical
clock method. This is based on competition kinetics between a
known monomolecular process used as reference and the
bimolecular reaction to be timed.29 The competing mono-
molecular processes employed were the well-known 5-hexo-
DFT Results. The effect of HFP on the thermodynamic and
kinetic properties has been investigated by comparing the DFT
results for the phenol/phenoxyl-(HFP)n supermolecule com-
plexes with those obtained while employing in the calculations
a good HBA solvent (2-propanol) and an apolar solvent
(propane).
cyclization of the 1-hexenyl radical (kr ) 2.3 × 105 M-1 s-1 30
)
which has been previously demonstrated to be independent of
(31) (a) Maeda, Y.; Schmid, P.; Griller, D.; Ingold, K. U. J. Chem. Soc., Chem.
Commun. 1978, 525-526. (b) Beckwith, A. L. J.; Bowry, V. W.; Ingold,
K. U. J. Am. Chem. Soc. 1992, 114, 4983-4992.
(32) Under our conditions, products deriving from the reactions between
phenoxyl and alkyl radical to give radical-radical recombination products
were not observed.
(33) The first report providing clear quantitative evidence of the role that
hydrogen bonding plays on the H-atom donor activities of phenols has been
reported by Ingold and co-workers: Avila, D. V.; Ingold, K. U.; Lusztyk,
J.; Green, W. H.; Procopio, D. R. J. Am. Chem. Soc. 1995, 117, 2929-
2930.
the nature of the solvent.11,31 The alkyl radicals were generated
(26) Pedrielli, P.; Pedulli, G. F. Gazzetta 1997, 127, 509-512.
(27) Unpublished results.
(28) Nakanishi, I.; Fukuhara, K.; Shimada, T.; Ohkubo, K.; Iizuka, Y.; Inami,
K.; Mochizuki, M.; Urano, S.; Itoh, S.; Miyata, N.; Fukuzumi, S. J. Chem.
Soc., Perkin Trans. 2 2002, 1520-1524.
(29) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317-323.
(30) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1981,
103, 7739-7742.
9
8324 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
Scheme 7
Table 5. O-H BDE (kcal mol-1) for para-Substituted Phenols
Computed with the B3LYP Method
isodesmic approach
absolute BDE
a
6-31G*
+ ZPVE
6-31G*
(∆H
6-311++G**
//6-31G*
LDBS
(∆H
6-31G*
)
)
298
exptb
298
H
Me
OMe
OH
88.3
86.5
82.9
82.9
79.9
0.27
88.3
86.6
83.3
83.4
80.2
0.53
88.3
86.5
83.2
83.3
80.1
88.3
86.3
82.5
82.9
79.5
0.13
87.1
84.6
80.9
88.3
86.2
82.8
NH2
77.7
1.77
79.5d
MADc
0.43
a From ref 34. b From ref 24b. c Mean absolute deviation. d Estimated
from the substituent additivity rules (ref 24d).
Table 6. BDE Values for PhOH and Activation Energy Ea for
Hydrogen Atom Abstraction from PhOH by Methyl Radical
Computed with the B3LYP Method in the Presence of Two
Molecules of Solvent
a
BDE (kcal mol-1
)
E (kcal mol-1
)
a
6-31G*
+ ZPVE
6-311++G**
//6-31G* + ZPVE 6-31G* + ZPVE
6-31G*
b
solvent
6-31G*
Chipman10c using DFT calculations showed that water
hydrogen-bonding effects on the structure, vibrational frequen-
cies, and hfs constants of phenoxyl and p-aminophenoxyl
radicals are well accounted for by considering the interaction
of the radical species with two and four hydrogen-bonded water
molecules, respectively, that is, two water molecules for each
substituent. An analogous approach has been used in the present
study. First, we examined the interaction of two solvent
molecules with the unsubstituted phenol (PhOH) and with the
corresponding phenoxyl radical (PhO•) by placing one molecule
of solvent near the O-atom (HBD interaction: the solvent
behaves as hydrogen bond donor) and one near the hydroxyl
H-atom (HBA interaction: the solvent behaves as hydrogen
bond acceptor) of PhOH (structure I in Scheme 7) and two
molecules of solvent interacting with the O-atom (two HBD
interactions) in PhO• (structure I•).
We then examined the interaction of the solvent with para-
substituted phenols p-X-PhOH (X ) NH2, OH, OMe) and with
the corresponding radical species placing two additional mol-
ecules of solvent near the para-X group. Two HBA interactions
were considered with the amine group (structures II and II•),
and one HBA plus one HBD interaction with the methoxy and
hydroxyl group (structures III and III•). In the discussion, the
strength of these interactions was qualitatively related to the
atomic distance of the interacting atoms.
propane
2-propanol 91.3 (3.0)
HFP
87.8 (-0.5) 87.9 (-0.4)
91.4 (3.1)
85.9 (-2.4) 86.1 (-2.2)
88.2 (-0.1)
91.7 (3.4)
85.6 (-2.7)
3.4
7.4
3.0
2.2
6.3
2.0
a Values relative to gas-phase calculations are reported in parentheses.
b ZPVE computed at the B3LYP/6-31G* level.
correction to the enthalpy, as well as enlargement of the basis
set (6-311++G**) have negligible effects.
The mean absolute deviation (MAD) indicates that agreement
with experiment is slightly better than that shown by the absolute
values computed recently with the B3LYP method34 in conjunc-
tion with the locally dense basis set (LDBS) approach.35
Table 6 shows that an apolar solvent such as propane has
little effect on the OH BDE value of phenol while a strong HBA
solvent such as 2-propanol increases considerably the BDE value
in agreement with experiments.27-29 DFT calculations on the
phenol/phenoxyl-(HFP)2 supermolecule complexes confirm that
the fluorinated solvent has an effect opposite to 2-propanol
because it decreases the BDE of phenol.
The measure of the atomic distances reported in Table 7
suggests that 2-propanol interacts more efficiently with the
starting phenol as hydrogen bond acceptor rather than with the
phenoxyl radical as hydrogen bond donor. On the other hand,
with HFP, the stabilization of the phenoxyl radical by formation
of a strong three-center hydrogen bond overcompensates the
interaction with the starting phenol as suggested by the ESR
experiment. Interestingly, the variation of the O-H BDE
produced by different solvents computed at the B3LYP/6-31G*
level is similar to that computed with the large 6-311++G**
basis set and estimating the ZPVE correction at the 6-31G*
level. Hence, the effect of solvents on the BDEs can be studied
explicitly using the isodesmic approach even when phenols and
phenoxyl radicals interact with four molecules of solvent as in
para-substituted phenols.
The O-H BDE values for phenols (ArOH) have been
computed using the isodesmic reaction
ArOH + PhO• h ArO• + PhOH
(6)
in conjunction with the experimental O-H BDE value for
phenol as
BDE(ArOH) ) BDEexp(PhOH) + [Ecalc(ArO•) -
Ecalc(ArOH)] - [Ecalc(PhO•) - Ecalc(PhOH)] (7)
The effect of propane on the O-H BDE of para-substituted
phenols was not studied because this apolar solvent has little
effect on the BDE of phenol. First, the effect of 2-propanol on
the O-H BDE of para-substituted phenols p-X-PhOH has been
to cancel out errors due mainly to electron correlation.
Table 5 shows that the BDEs for para-substituted phenols
computed with the B3LYP-DFT method using the isodesmic
approach are in excellent agreement with the experimental
values even employing a small double ú basis set (6-31G*).
An estimate of the zero point vibration energy (ZPVE), thermal
(34) Wright, J. S.; Johnson, E. R.; DiLabio, G. A. J. Am. Chem. Soc. 2001,
123, 1173-1183.
(35) DiLabio, G. A. J. Phys. Chem. A 1999, 103, 11414-11424.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8325

A R T I C L E S
Lucarini et al.
Table 7. Relevant Atomic Distancesa d (Å) for the Interaction of Two Molecules of Solvent with PhOH, PhO• (in Parentheses), and
[PhOH-Me•]q Computed at the B3LYP/6-31G* Level
PhOH (PhO•)
d(H −O )
[PhOH−Me•]q
solvent
d(O −H )
d(O −H )
d(O −H )
d(H −O )
d(O −H )
d(H −C)b
P
S
P
S
P
P
P
S
P
S
P
P
P
propane
2.663
(2.666, 2.708)
1.911
(1.949, 1.965)
1.780
0.969
0.989
0.979
2.687, 2.702
1.968, 1.998
1.796, 1.968
1.114
1.137
1.117
1.457
1.429
1.453
2-propanol
HFP
1.746
1.796
2.953
2.738
(1.817, 1.858)
a P ) phenol/phenoxyl; S ) solvent (see Scheme 7). b C ) carbon of attacking methyl radical.
Table 8. O-H BDE (kcal mol-1) for para-Substituted Phenols Computed in the Presence of n Molecules of Solvent at the B3LYP/6-31G*
Level (Values Relative to the Corresponding Gas-Phase BDEs Are Reported in Parentheses)
2-propanol
HFP
n ) 4
n ) 2
n ) 2
n ) 4
n ) 2
n ) 2
X
near OH/O•
near X
near OH/O•
near X
H
Me
OMe
OH
NH2
91.3 (3.0)
85.9 (-2.4)
83.6 (-2.9)
78.5 (-4.4)
78.5 (-4.4)
74.0 (-5.9)
85.3 (2.4)
82.6 (-0.3)
73.7 (-6.2)
84.6 (1.7)
84.4 (1.5)
80.1 (0.2)
83.2 (0.3)
81.7 (-1.2)
75.8 (-4.1)
79.8 (-3.1)a
78.8 (-4.1)
71.4 (-8.5)
84.7 (1.8)a
83.6 (0.7)
78.4 (-1.5)
a Only one molecule of HFP near X (see text).
Table 9. Relevant Atomic Distancesa d (Å) for the HBD and HBA
Interactions Involving the X Group in para-Substituted Phenols
p-X-C6H4OH Interacting with Four Molecules of Solvent at the
B3LYP/6-31G* Level (Values for Phenoxyl Radicals Are Reported
in Parentheses)
stabilization of the phenoxyl radical produced by this HBA
interaction at the substituent counterbalances the stabilization
of the phenol produced by the HBA interaction of 2-propanol
with the phenolic OH group previously described, so that
2-propanol slightly reduces the O-H BDE. Indeed, a decrease
of BDE for 3a on going from an apolar solvent to alcohols (see
Table 3) was found experimentally.
2-propanol
HFP
X
d(O −H )
d(H −O )
d(C−X) d(O −H )
d(H −O )
d(C−X)
X
S
X
S
X
S
X
S
OMe 1.937
(2.017) (2.768)
1.900 1.761
(2.005) (1.690)
2.835
1.376
1.348 (2.018)
1.370
1.947
b
1.374
(1.352)
As expected, for X ) OMe, the HBA interaction of
2-propanol with the para-methoxy group is negligible (see Table
9 and Figure 6), and the BDE value in the presence of such a
strong HBA solvent increases similarly to the unsubstituted
phenol, the variation being slightly lower. Table 8 shows that
the effect of 2-propanol on the BDE due to interactions with
the OH/O• and X groups is not additive. The electronic effect
of the para substituent X on the HBA interaction of 2-propanol
with the phenoxylic group should be influenced by the change
of π electron density at X produced by interactions of 2-propanol
with the X group. This indirect effect should decrease largely
the computed BDE for X ) NH2 and moderately that for X )
OH and increase the BDE for X ) OMe in accord with
calculations.
OH
(1.344)
NH2
2.080, 2.104 1.399
(1.914, 1.948) (1.350)
2.208, 2.220 1.405
(1.967, 2.044) (1.353)
a S ) solvent (see Scheme 7). b One molecule of HFP near X.
studied at the simple B3LYP/6-31G* level. Further calculations
were carried out by placing only two molecules of 2-propanol
either at the OH/O• site or at the X site to study separately the
interactions of the solvent with the OH/O• group and the para
substituent. Table 8 shows that the presence of four molecules
of 2-propanol strongly decreases the BDE value for p-amino-
phenol as suggested by the experimental results (see Tables 3
and 5).
The effects of HFP have then been investigated. For X )
OMe, only one molecule of HFP was placed near the methoxy
group because the HBA interaction of 2-propanol with the
methyl of the methoxy group was found to be negligible.
Interestingly, in accord with the BHT/BHA radical-equilibrium
EPR experiment, the HBD interaction of HFP with the methoxy
group stabilizes the starting phenol with respect to the phenoxyl
radical because the polar structures in the radical species (E-H
of Scheme 2) bring a formal positive charge at the ethereal
oxygen. As expected, the increase of BDE is much larger in
HFP (1.8 kcal mol-1) than in 2-propanol (0.3 kcal mol-1), in
excellent agreement with the BDE determined experimentally
for BHA in different solvents (see Table 3). For X ) OH, the
weak HBA interaction of HFP with the hydrogen of the para-
OH group produces an opposite effect and reduces the de-
stabilization of the radical species in excellent agreement with
The measure of the atomic distances reported in Table 9
indicates that this lowering is due to the strong increase of the
HBA interaction between 2-propanol and the amine group on
going from p-aminophenol to the p-aminophenoxyl radical. In
the radical species, this interaction is stronger because of the
larger weight of a polar resonance structure having a positive
charge on the N-atom and a formal C-N double bond (see
Scheme 2). Actually, the C-N bond length decreases consider-
ably in the radical species.
For X ) OH, the measure of the atomic distance indicates
that 2-propanol interacts with the para-OH group both as a
strong H-bond acceptor and as a weak H-bond donor. These
interactions produce an opposite effect on the BDE. The
decrease of BDE on going from the gas phase to the super-
molecule complex in which 2-propanol interacts only with the
para-OH group confirms that the HBA interaction prevails. The
9
8326 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
Figure 6. B3LYP/6-31G* calculated structures of p-methoxyphenol (a) and p-methoxyphenoxyl radical (b) with H-bonded 2-propanol molecules. Aliphatic
and aromatic hydrogens have been removed for clarity.
the trend observed experimentally for the BDE value of BHQ
with the addition of HFP to benzene.
The solvent effect on the BDE of p-aminophenol is signifi-
cantly stronger in 2-propanol than in HFP, the decrease of O-H
BDE with respect to the unsubstituted phenol being 9.2 and
6.1 kcal mol-1, respectively. This is in accord with the weak
HBA character of HFP as confirmed by the measure of the
atomic distances reported in Table 9.
Calculations were also carried out on the p-methoxyphenoxyl
radical to determine the effect of HFP both on magnetic (hfs
constants) and on energetic (barrier Erot to the internal rotation
of the para-methoxy group) properties. At the UB3LYP/6-31G*
level, the sum of the ortho hfs constants decreases in absolute
value from -13.39 to -11.84 G in accord with the ESR results
displayed in Figure 1. The UB3LYP/6-31G* rotation barrier
increases from 6.8 kcal mol-1 computed in the gas phase to
8.4 kcal mol-1 in HFP in good accord with the variation of the
rotational barrier produced by adding HFP to a toluene solution
as determined by the line width broadening of the ESR spectra.
The ∆Erot (1.6 kcal mol-1) is slightly smaller than that observed
Figure 7. B3LYP/6-31G* calculated transition state structures for hydrogen
abstraction from phenol by methyl radical in the presence of two HFP
molecules. Numbers refer to atomic HS-OP distances (Å) and OPHSOS
bond angles.
experimentally (2.06 kcal mol-1). The computed Erot (8.3 kcal
mol-1) does not change significantly in calculations performed
placing only two molecules of HFP near the phenoxylic group.
Thus, the observed increase of Erot is due to a reduction in the
transition state of the strong HBD-stabilizing interaction of HFP
with the phenoxylic oxygen produced by the decrease of the
electron-donating effect of the para substituent that depends
on the π conjugation. Interestingly, the O-H BDE increases
by 3.5 kcal mol-1 (i.e., BDE ) 83.3 kcal mol-1) upon rotation
of the methoxy group by 90°. This value is close to the BDE
for p-methylphenol (83.6 kcal mol-1) computed in the same
solvent. This finding is in accord with the experimental evidence
that p-methoxy substituents constrained in the orthogonal
conformation by meta bulky groups have the same effect of a
para methyl group on the BDE of phenols.34
nature of the solvent on the activation energy Ea does not follow
the same trend as found for the O-H BDEs. The Ea value in
the presence of two molecules of 2-propanol is much higher
than in the other two cases as expected for a strong HBA solvent
that stabilizes to a larger extent the phenol rather than the
transition state. On the other hand, a HBD solvent might be
expected to decrease considerably Ea with respect to an apolar
solvent stabilizing preferentially the transition state. The decrease
of Ea is, however, small on going from propane to HFP.
Actually, the atomic OP-HS distances (1.796 and 1.968 Å)
reported in Table 7 and the corresponding OPHSOS bond
angles (177.8° and 134.7°) indicate that, in the transition state
computed in the presence of HFP, the second HBD interaction
is much weaker than that found in phenoxyl radicals (see Figure
7).
Finally, we have investigated the effect of solvent on the
activation energy Ea for the hydrogen atom abstraction from
phenol by methyl radical. Table 6 shows that the effect of the
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8327

A R T I C L E S
Lucarini et al.
Thus, the phenoxyl radical is not yet strongly stabilized by
HFP as expected for an early transition state. Actually, the O-H
bond (see d(OP-HP) values in Table 7) is stretched to a length
much shorter than the distance of the transferring hydrogen from
the abstracting carbon, d(HP-C). It is worth noting that these
atomic distances are similar in HFP and propane, while the
former is longer and the latter is shorter in 2-propanol as
expected for a transition state having higher activation energy.
As expected, the most evident change in the structural param-
eters of the phenol in the transition state is the rotation of the
O-H bond from the planar to a nearly orthogonal conformation,
the out-of-plane dihedral angle being 80.4° in HFP. This is a
consequence of the fact that in the hydrogen atom transfer the
π conjugated lone pair on oxygen in phenol must transform
into an in-plane σ lone pair in the phenoxyl radical.8a
the solvent with the phenolic OH group, so that the global effect
is a decrease of the BDE value.
The effect of HBD solvents on the reactivity is more puzzling.
In general, it has been found that the reactivity of phenols is
strictly correlated with their BDE values;24b,36 accordingly, the
logarithm of the rate constant for the reactions with peroxyl37
or alkyl radicals11 in the presence of HBA solvents decreases
parallel to the increase of the strength of the O-H bond, while,
in HBD solvents, the decrease of BDE is not accompanied by
a larger reactivity. The origin of this unexpected behavior has
been shown by DFT calculations, for the reaction with methyl
radical, to depend on the early character of the transition state
that resembles more the reagents than the products, so that the
stabilization of the phenoxyl radical by H-bond complexation
does not have any effect on the reactivity of the parent phenol.
A remarkable effect of H-bond complexation is instead repre-
sented by the much greater persistency of phenoxyl radicals,
for example, the R-tocopheroxyl radical, observed in the
presence of HFP. More detailed kinetic studies are however
required to clarify the reason for this behavior.
Conclusions
The present experimental and computational study demon-
strates that hydrogen bond donors can significantly change the
spectroscopic and thermochemical properties of phenoxyl
radicals. Thus, a strong H-bond donor such as hexafluoro-
propanol produces large variations of the hyperfine splitting
constants indicative of a large redistribution of electron spin
density which can be accounted for by the increased importance
of the mesomeric structures (E-H) with electric charge separa-
tion. The conformational rigidity of phenoxyl radicals with
electron-releasing substituents is also greatly enhanced in the
presence of HFP, this being due to the fact that three of the
polar structures (E-H) contain a formal double bond between
the aromatic ring and the substituent.
Experimental Section
Materials. Solvents were of the highest purity grade commercially
available and were used as received. Phenols 1a-4a, BHT, and BHA
were commercially available and were used as received. R-Tocopherol
was purified by column chromatography on silica gel according to a
previously described method.11 Hexa-n-butylditin and 6-bromo-1-hexene
were stored under nitrogen at 5 °C. Di-tert-butylperoxide was percolated
through activated basic alumina and stored at 5 °C. BHQ was prepared
according to literature by reducing the corresponding 2,6-di-tert-butyl
quinone.38
As far as the effect on the thermochemistry is concerned, it
has been found that in phenol and alkyl substituted phenols the
O-H bond dissociation enthalpy, which in HBA solvents is
larger than in hydrocarbons due to the preferential stabilization
of the starting phenol with respect to the phenoxyl radical, is
lowered in the presence of HBD solvents because they
preferentially stabilize the phenoxyl radical.
EPR Experiments. Deoxygenated benzene solutions containing the
phenol under investigation (0.1-0.5 M), an appropriate reference phenol
(0.1-0.5 M), and di-tert-butyl peroxide (5% v/v) were sealed under
nitrogen in a suprasil quartz EPR tube. The sample was inserted in the
thermostated cavity of an EPR spectrometer and photolyzed with the
unfiltered light from a 500 W high-pressure mercury lamp. The
temperature was controlled with a standard variable temperature
accessory and was monitored before and after each run with a copper-
constantan thermocouple.
The EPR spectra were recorded on a Bruker ESP 300 spectrometer
equipped with a Hewlett-Packard 5350B microwave frequency counter
for the determination of the g-factors, which were corrected with respect
to that of perylene radical cation in concentrated H2SO4 (g ) 2.00258).
The molar ratio of the two equilibrating radicals was obtained from
the EPR spectra and used to determine the equilibrium constant, K4,
by introducing the initial concentrations of the two reactants. Initial
concentrations were high enough to avoid significant consumption
during the course of the experiment. Relative radical concentrations
were determined by comparison of the digitized experimental spectra
with computer simulated spectra. In these cases, an iterative least-
squares fitting procedure based on the systematic application of the
Monte Carlo method was performed to obtain the experimental spectral
parameters of the two species including their relative intensities.24b
Kinetic Measurements. In a typical experiment, 200 mL of a
solution of the phenol (0.1-1 M) containing 6-bromo-1-hexene (0.005-
0.01 M) and hexa-n-butylditin (0.01 M) was sealed in a quartz tube,
after being deoxygenated by bubbling nitrogen. The reaction mixtures
were then irradiated for 30-120 min at 298 K in a thermostated
In phenols containing groups such as OR that are acceptors
of H-bonds, the interaction between the HBD solvent and the
substituent is stronger in the phenol than in the corresponding
phenoxyl radical because the radical oxygen behaves as an
electron-withdrawing group which decreases the complexating
ability of the substituent. This interaction partially compensates
the stabilizing effect on the phenoxyl radical due to the H-bond
formation at the radical oxygen, so that the overall effect of
HBD solvents is a decrease of the BDE value smaller that in
the case of alkyl substituted phenols.
A similar behavior is also experienced, in the presence of
HBD solvents, by phenols containing OH and NH2 groups.
Actually, phenols such as 3a, 4a, and BHQ display a previously
unreported effect when dissolved in the HBA solvent tert-butyl
alcohol because, instead of showing an increase of the O-H
BDE value as usual, they undergo a weakening of the phenolic
bond. The reason is that the stabilizing interaction between
solvent and the substituent is much stronger in the phenoxyl
radical than in the parent phenol because of the electron-
withdrawing effect of the radical oxygen, which makes more
acidic, and therefore more available to give H-bonds, the OH
or NH groups. This interaction between HBA solvent and
substituent in the radicals exceeds the stabilizing interaction of
(36) Burton, G. W.; Doba, T.; Gabe, E. J.; Hughes, L.; Lee, F. L.; Prasad, L.;
Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053-7065.
(37) Valgimigli, L.; Banks, J. T.; Lusztyk, J.; Ingold, K. U. J. Org. Chem. 1999,
64, 3381-3383.
(38) Worrel, C. J.; McLean, R. L. U.S. Patent 3415850, 1968.
9
8328 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Hydrogen-Bonding Effects on Phenoxyl Radicals
A R T I C L E S
photoreactor, built in our laboratories, equipped with a 125 W high-
pressure mercury lamp, and the products were analyzed by gas
chromatography. For each phenol, in each solvent or solvent mixture,
four measurements were made with a different phenol concentration,
and the reaction products ratio [UH]/[RH] was plotted versus the phenol
concentration to obtain the kH/kr ratio by linear regression of the
experimental data.
Computational Details. DFT calculations with the B3LYP func-
tional39,40 were carried out on para-substituted phenols p-X-PhOH (X
) H, Me, NH2, OH, OMe) and on the corresponding phenoxyl radicals
using the Gaussian 98 system of programs.41 The unrestricted wave
function was used for radical species. Geometries were obtained
employing a valence double-ú basis set supplemented with polarization
d-functions on heavy atoms (6-31G*).42 Single point calculations were
performed using a valence triple ú basis set augmented with diffuse
functions43 (6-311++G**). Zero point vibrational energies (ZPVE)
were computed at the B3LYP/6-31G* level from frequency calculations
using a scaling factor of 0.9806 to account for anharmonicity.44 O-H
BDEs were computed using the isodesmic approach. Activation energy
Ea for hydrogen abstraction from phenol by methyl radical in the
presence of two molecules of solvent has been computed at the B3LYP/
6-31G* level. The nature of the transition states was verified by
frequency calculations (one imaginary frequency), and the ZPVE
correction was also estimated.
Acknowledgment. Financial support from MURST (research
project “Free Radical Processes in Chemistry and Biology:
Fundamental Aspects and Applications in Environment and
Material Sciences”) is gratefully acknowledged.
(39) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(40) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
(41) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A., Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.7;
Gaussian, Inc.: Pittsburgh, PA, 1998.
Supporting Information Available: EPR spectra and experi-
mental plots (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA034963K
(42) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213-222.
(43) Clark, T.; Chandrasekhar, J.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4,
294-301.
(44) Scott, P. A.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8329