Photochemistry and reactivity of the phenyl radical-water system: A matrix isolation and computational study

Article abstract of DOI:10.1002/chem.200903362

The reaction of the phenyl radical 1 with water has been investigated by using matrix isolation spectroscopy and quantum chemical calculations. The primary thermal product of the reaction between 1 and water is a weakly bound complex stabilized by an OH...π interaction. This complex is photolabile, and visible-light irradiation (λ > 420 nm) results in hydrogen atom transfer from water to radical 1 and the formation of a highly labile complex between benzene and the OH radical. This complex is stable under the conditions of matrix isolation, however, continuous irradiation with λ>420nm light results in the complete destruction of the aromatic system and formation of an acylic unsaturated ketene. The mechanisms of all reaction steps are discussed in the light of ab initio and DFT calculations.

Full text of DOI:10.1002/chem.200903362

FULL PAPER
DOI: 10.1002/chem.200903362
Photochemistry and Reactivity of the Phenyl Radical–Water System:
A Matrix Isolation and Computational Study
Artur Mardyukov,[a] Rachel Crespo-Otero,[a] Elsa Sanchez-Garcia,[b] and
Wolfram Sander*[a]
Dedicated to Professor Wolfgang Kirmse on the occasion of his 80th birthday
Abstract: The reaction of the phenyl
radical 1 with water has been investi-
gated by using matrix isolation spec-
troscopy and quantum chemical calcu-
lations. The primary thermal product
of the reaction between 1 and water is
a weakly bound complex stabilized by
an OH···p interaction. This complex is
photolabile, and visible-light irradiation
(l>420 nm) results in hydrogen atom
transfer from water to radical 1 and
the formation of a highly labile com-
plex between benzene and the OH rad-
ical. This complex is stable under the
conditions of matrix isolation, however,
continuous irradiation with l>420 nm
light results in the complete destruction
of the aromatic system and formation
of an acylic unsaturated ketene. The
mechanisms of all reaction steps are
discussed in the light of ab initio and
DFT calculations.
Keywords: density functional calcu-
lations · infrared spectroscopy · ma-
trix isolation · photochemistry · rad-
ical reactions
Introduction
The phenyl radical (1) is a reactive intermediate of funda-
mental importance to organic chemistry. It has been shown
to play a key role in the combustion of hydrocarbon fuels to
produce polycyclic aromatic hydrocarbons and soot.[1] It has
also been discussed as an intermediate in the chemistry of
the interstellar medium, especially for the formation of poly-
cyclic aromatic hydrocarbons,[2] and in tropospheric chemis-
try.[3,4] The most important reaction for the removal of ben-
zene (2) from the troposphere is the reaction with the OH
radical. For this reaction two channels have been observed
in the gas phase: 1) hydrogen abstraction to give 1 and
water and 2) addition to give 2-hydroxy-3,5-cyclohexadienyl
(3) (Scheme 1). In the troposphere the sequence of reactions
following the formation of radical 3 is fairly complex and
not well understood.
Scheme 1. Reaction channels for the reaction of 2 with OH in the gas
phase.
Due to its great importance, the reaction between 2 and
the OH radical has been the subject of a large number of
experimental[5–11] and theoretical studies.[12–22] At higher tem-
peratures (above room temperature) reaction channel a,
producing 1 and water dominates, whereas at room temper-
ature and below reaction channel b, the addition of OH to 2
to give 3, is the major pathway. One of the products of this
reaction channel is phenol.
In a computational study at the G3 level of theory Tok-
makov and Lin investigated the mechanism of the 2+OH
reaction and found that channel a proceeded by a single-
step reaction with an activation barrier of 5.3 kcalmolꢀ1 and
a reaction energy of ꢀ4.2 kcalmolꢀ1.[14] In contrast, the OH
addition, channel b, proceeds via a prereaction complex
2···OH. This complex shows a nonbonding interaction be-
tween the O atom and one of the C atoms of 2 and the OH
hydrogen atom pointing in the direction of the p system of
2. The complex is stabilized by 2.6 kcalmolꢀ1, which increas-
[a] A. Mardyukov, Dr. R. Crespo-Otero, Prof. Dr. W. Sander
Lehrstuhl fꢀr Organische Chemie II, Ruhr-Universitꢁt Bochum
44780 Bochum (Germany)
Fax : (+49)234-321-4353
E-mail: wolfram.sander@rub.de
[b] Dr. E. Sanchez-Garcia
Max-Planck Institut fꢀr Kohlenforschung, Mꢀlheim an der Ruhr
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim an der Ruhr (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200903362.
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8679

es the activation barrier for the addition of OH to 2 from
2.8 kcalmolꢀ1 (relative to the separated reactants) to
5.4 kcalmolꢀ1 (relative to the 2···OH complex). The forma-
tion of prereactive complexes in radical reactions has re-
cently been investigated by several authors.[23–26] It has been
shown that carbon-centered radicals indeed can function as
hydrogen-bond acceptors. However, due to the low electro-
negativity of carbon and since radicals provide only one
electron and not two like atoms with lone pairs, the hydro-
gen bonds to carbon-centered radicals are quite weak.
Results and Discussion
The water–phenyl radical complex: The phenyl radical (1)
was generated by flash vacuum pyrolysis (FVP) of azoben-
zene 4 at 5008C with subsequent trapping of the products in
solid argon at 10 K (Scheme 2). A second major product
In a combined experimental and theoretical study Platz
et al. investigated solvent effects for the addition of OH to
2.[8] By using laser flash photolysis it was shown that the ad-
dition in water is faster by a factor of 65 than in acetonitrile,
which could not be explained by the differences in the die-
lectric constants of these two solvents. DFT calculations
using the B3LYP functional revealed that the free energy of
activation at room temperature for the hydrogen abstraction
is about 1 kcalmolꢀ1 higher than for the addition, and thus
addition should predominate. In this study, a weakly bound
complex (similar to that described by Tokmakov and Lin
obtained with a slightly different basis set)[14] was used as a
starting point for the reactions between 2 and OH for both
reaction channels. At the B3LYP/6-311+G**//B3LYP/6-
31+G** level of theory, a binding energy of the complex of
ꢀ3.65 kcalmolꢀ1 was calculated and DG was calculated to
+3.96 kcalmolꢀ1. The solvent effect of the addition reaction
was explained by the formation of specific complexes be-
tween solvent molecules and the OH radical. The reaction
of OH complexed with water has a significantly lower acti-
vation barrier than that of OH complexed with acetonitrile.
In the temperature range between 325 and 380 K the gas-
phase reaction between OH and 2 shows a nonlinear Arrhe-
nius plot and the decay of OH is nonexponential. In a com-
putational study by Alvarez-Idaboy et al., this complex ki-
netic behavior was explained by the competition between
the addition reaction and hydrogen abstraction and by the
reversibility of the addition reaction.[22] Thus, the decompo-
sition of radical 3 back to 2 and OH is of importance to un-
derstand the experimental kinetic data.
Recently, we described the formation of a weakly bound
complex between the phenyl radical 1 and water under the
conditions of matrix isolation.[5] This complex proved to be
photolabile, and with visible-light irradiation reacted to a
complex between the hydroxyl radical and 2. We now found
that this second complex is also photolabile and irradiation
surprisingly results in ring opening and formation of a
highly unsaturated ketene. Herein we describe a detailed
mechanistic and computational study on the reaction be-
tween 1 and water. The sequence of reactions starts with 1
interacting with water and ends with the complete destruc-
tion of the aromatic ring system under fairly mild irradiation
conditions. A detailed knowledge of the elemental steps of
this reaction sequence can contribute to the understanding
of the highly complex processes involved in the combustion
and degradation of aromatic compounds.
Scheme 2.
formed under these conditions is 2, obtained from 1 either
by hydrogen abstraction from surface contaminations of the
pyrolysis oven or by bimolecular reactions or rearrange-
ments in the gas phase. Since pyrolysis of the perdeuterated
[D10]4 yields mainly [D6]2, the main source of hydrogen
comes from 4. The IR spectrum of matrix-isolated 1 with
the strongest absorptions at 705.8 and 657.4 cmꢀ1 is in agree-
ment with literature data.[27]
The advantage of generating 1 in the gas phase followed
by trapping with excess argon is that, under these conditions,
statistically most matrix cages contain only one molecule of
1, whereas the photolysis of a matrix-isolated precursor pro-
duces radical pairs in the matrix cages. Annealing of matri-
ces containing radical pairs results almost exclusively in rad-
ical recombination, whereas isolated radicals are able to un-
dergo bimolecular reactions with added dopands. This has
been shown previously in the case of O2-doped argon matri-
ces in which, at temperatures below 15 K, the diffusion of
trapped molecules is very slow and no reaction between 1
and O2 is observed. Annealing at 30 K, however, allows
rapid diffusion of O2 and the phenylperoxy radical is formed
in a clean bimolecular reaction.[28]
Doping argon matrices containing the products of the
FVP of 2 with 0.1–1% of H2O results in some significant
changes in the IR spectra. Most notably are these changes
in the area of the out-of-plane (o.o.p) CH deformation
modes between 650 and 720 cmꢀ1 (Figure 1). The strong fun-
ꢀ
damental n4 deformation of 2 (a2u symmetrical o.o.p. C H
deformation mode, in-phase movement of all hydrogen
ꢀ
Figure 1. Displacement vectors showing some o.o.p. C H deformation vi-
brations of 2 and 1.
8680
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8679 – 8689

Photochemistry and Reactivity of the Phenyl Radical–Water System
FULL PAPER
atoms) at 675.2 cmꢀ1 is blueshifted by 6.8 cmꢀ1 in the
2···H2O complex, in very good agreement with the observa-
tions of Engdahl and Nelander.[29] If the argon is doped with
only 0.1% H2O, this band is very weak after the deposition
of the matrix at 10 K, but increases in intensity during an-
nealing at temperatures above 25 K. In [D6]2 n4 shows a
very large isotopic shift to 497.4 cmꢀ1, and the [D6]2···H2O
complex shows a blueshift of 3.5 cmꢀ1. As expected, isotopic
labeling of water (deuterium or 18O) does not influence this
vibration.
same shifts as with H2O (Figure 2d and f and Table 2).
Other vibrations of 1 are not affected by the formation of
the 1···H2O complex.
ꢀ
The o.o.p. C D deformation mode n6 of [D5]1 is found at
518.0 cmꢀ1 (Figure 3).[27] In the presence of H2O, the forma-
tion of the [D5]1···H2O complex is indicated by a new band
at 522.3 cmꢀ1 that corresponds to a blueshift of 4.3 cmꢀ1 with
respect to the monomer.
Matrix-isolated 1 shows strong IR bands at 705.8 (n6) and
657.4 cmꢀ1 (n5), in good agreement with literature data.[27]
After annealing a matrix containing 1 and 0.1% of H2O for
several minutes at 30 K, these bands decrease in intensity
and new bands appear at 711.4 and 659.8 cmꢀ1. The blue-
shifts of 5.6 and 2.4 cmꢀ1, respectively, are similar to that of
the corresponding bands in the 2···H2O complex (Figure 2).
Figure 3. a) IR spectrum of a mixture of [D5]1 and [D6]2 in argon doped
with 1% H2O at 10 K, and b) after warming from 10 to 40 K with ap-
proximately 1 Kminꢀ1. c) IR spectrum of a mixture of [D5]1 and [D6]2 in
argon doped with 1% D2O at 10 K, and d) after warming from 10 to
40 K with approximately 1 Kminꢀ1
.
Water is quite mobile in argon, and even at low tempera-
ture and at low concentrations some molecules are rotating
and diffusion is rapid enough to easily form aggregates.[30–32]
Figure 2. a) IR spectrum of a mixture of 1 and 2 in argon doped with
0.1% H2O at 10 K, and b) after warming from 10 to 40 K with approxi-
mately 1 Kminꢀ1. c) IR spectrum of a mixture of 1 and 2 in argon doped
with 1% H2O at 10 K, and d) after warming from 10 to 40 K with ap-
proximately 1 Kminꢀ1. e) IR spectrum of a mixture of 1 and 2 in argon
doped with 1% D2O at 10 K, and f) after warming from 10 to 40 K with
approximately 1 Kminꢀ1. Bands of remaining azobenzene 4 are marked
*.
ꢀ
This results in complex IR patterns in the O H stretching
region, and therefore the assignment of bands of the 1···H2O
complex in this spectral region is difficult. The n2 mode of
the water monomer is observed at 3639.7 cmꢀ1 [30]
The for-
.
mation of 1···H2O results in a new band at 3618.0 cmꢀ1,
which is redshifted by 21.7 cmꢀ1 with respect to the mono-
mer (Figure 1AS in the Supporting Information). As expect-
ed, deuteration of the phenyl radical in [D5]1···H2O does not
ꢀ
The new bands appear only when both 1 and water are pres-
ent in the matrix, and the band intensities increase if the
concentration of one of the two
influence this vibration (Table 2). The O D stretching mode
(n2) of 1···D2O is observed at 2645.1 cmꢀ1, redshifted by
monomers is increased (Fig-
ure 2d). We therefore assign
Table 1. Calculated stabilization energies of C6H5···H2O complexes [kcalmolꢀ1].
UB3LYP[a]
UM05-2X[b]
RHF-UCCSD(T)[c]
ꢀ
these bands to the o.o.p. C H
Complex
DE
DECP
DECP+ZPE
DE
DECP
DECP+ZPE
DE
DECP
deformation modes n6 and n5 of
a 1···H2O complex. The bands
of this complex already appear
at low concentrations of water,
which indicates that a 1:1 com-
plex containing only one mole-
cule of water is formed. If D2O
is used in these experiments,
the o.o.p. vibrations of 1 in the
A
B
BI
C
CI
D
E
ꢀ1.71
ꢀ1.44
–
ꢀ0.67
–
ꢀ3.77
ꢀ3.01
ꢀ2.88[e]
ꢀ2.23
–
ꢀ3.26
ꢀ2.76
–
ꢀ2.38
ꢀ1.68
–
ꢀ3.69
ꢀ2.71
–
ꢀ2.45
ꢀ1.77
–
[d]
–
ꢀ2.00
ꢀ1.74
ꢀ0.79
–
–
–
ꢀ2.04
ꢀ1.09
ꢀ2.38
ꢀ1.69
ꢀ1.18
ꢀ1.09
ꢀ1.11
ꢀ1.02
ꢀ0.92
ꢀ0.96
ꢀ0.42
ꢀ0.27
ꢀ0.37
–
–
–
–
ꢀ2.10
ꢀ1.91
–
ꢀ1.03
–
ꢀ2.27
–
ꢀ1.59
–
ꢀ1.65[e]
[a] UB3LYP/6-311++G
UM05-2X/6-311++G(2d,2p). [d] Optimization of B at the B3LYP level leads to BI. [e] At the M05-2X level
BI is a transition state (TS) leading to B with one imaginary vibration at 25 cmꢀ1. [f] At the M05-2X level, E is
ACHTUNGTREN(GNUN 2d,2p). [b] UM05-2X/6-311++GACHTUNRTGEG(NNUN 2d,2p). [c] RHF-UCCSD(T)/6-311++GACHTUNGTRENNUNG(2d,2p)//
AHCTUNGTRENNUNG
complex show basically the a transition state (TS) leading to D with one imaginary vibration at 18 cmꢀ1
.
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8681

W. Sander et al.
Table 2. Experimental[a] and calculated (in italics)[b] IR spectroscopic
data of the C6H5···H2O complex.
13.4 cmꢀ1 from n2 of D2O at 2658.5 cmꢀ1 (Figure 1BS in the
Supporting Information).
OH str.
CH o.o.p. def.
CH o.o.p. def.
The frequencies and the frequency shifts in the 1···H2O
complex are similar to those of 2···H2O,[29] which suggests
that both complexes have similar structures with the water
molecule located on top of the aryl p system forming a
weak OH···p hydrogen bond. On the other hand, the radical
center in 1 could also act as a hydrogen-bond acceptor,
which should lead to a second complex with an in-plane
H2O
3639.4[c]
3894.2
D2O
2658.8[c]
2806.6
H218O
C6H5
C6D5
3630.7[e]
3886.1
705.8[d]
657.4[d]
681.8
–
734.0
water molecule and a OH···CACTHNUTRGENUGN(radical) hydrogen bond. We
518.0[d.]
therefore systematically investigated the interactions be-
tween 1 and water using DFT methods. Since dispersion is
important to describe these weakly bound complexes, we
used the M05-2X functional developed by Truhlar and
538.1
C6H5···H2O
A
B
C
3618.0 (ꢀ21.4)
3863.5 (ꢀ30.7)
3806.2 (ꢀ88.0)
3887.7 (ꢀ6.5)
3887.8 (ꢀ6.4)
3618.0 (ꢀ21.4)
3864.5 (ꢀ29.7)
711.4 (+5.6)
739.3 (+5.3)
740.9 (+6.9)
744.8 (+10.8)
746.6 (+12.6)
522.3 (+4.3)
540.2 (+2.1)
539.1 (+1.0)
541.3 (+3.2)
544.5 (+6.4)
711.8 (+6.0)
739.3 (+5.3)
740.9 (+6.9)
659.8 (+2.4)
682.6 (+0.8)
680.5 (ꢀ1.3)
684.1 (+2.3)
Zhao[33,34] with a large 6-311++G
ACTHNUTRGNE(UNG 2d,2p) basis set. At this
D
684.6 (+2.8)
–
–
–
–
–
level of theory four 1···H2O complexes A–D were found
(Figure 4) that are stabilized by the following intermolecular
interactions:
C6D5···H2O
A
B
C
3806.2
G
3887.7 (ꢀ6.5)
3887.8 (ꢀ6.4)
2645.1 (ꢀ13.7)
2787.1 (ꢀ19.5)
2862.0 (+56.0)
D
1) OH···p interaction between one hydrogen atom of the
water molecule and the p system of 1; this can be regard-
ed as a weak OH···p hydrogen bond.
C6H5···D2O
A
B
C
659.8 (+2.4)
682.6 (+0.8)
680.4 (ꢀ1.4)
2801.7
N
744.7
N
684.1 (+2.3)
2) OH···C
ACHTUNGTRENNUNG(radical) interaction between one hydrogen atom
D
2801.8 (ꢀ4.8)
2645.1 (ꢀ13.7)
2787.1 (ꢀ19.5)
2862.0 (+55.4)
2801.7 (ꢀ4.9)
2801.7 (ꢀ4.9)
3610.5 (ꢀ20.2)
3854.7 (ꢀ31.4)
3796.6 (ꢀ89.5)
3879.6 (ꢀ6.5)
3879.7 (ꢀ6.4)
746.5 (+12.5)
522.3 (+4.3)
539.9 (+1.8)
539.1 (+1.0)
684.5 (+2.7)
–
–
–
–
–
659.5 (+2.1)
682.6 (+0.8)
680.5 (+21.0)
of the water molecule and the radical center of 1. This
C6D5···D2O
A
B
C
interaction can be regarded as a weak OH···C
hydrogen bond.
ACHTUNGTRENN(UNG radical)
541.2
544.3
U
D
C6H5···H218O
711.5 (+5.7)
739.3 (+5.3)
740.9 (+6.9)
744.8 (+10.8)
746.6 (+12.6)
A
B
C
D
684.1ACHTUNGRTEN(NUNG +2.3)
684.6(+2.8)
AHCTUNGTRENNUNG
[a] Argon
matrix.
[b] M05-2X/6-311++GACHTUNGTRENNUNG
ences [29,35,36]. [d] Reference [27]. [e] Reference [37].
3) CH···O interaction between one hydrogen atom of the
phenyl radical and the oxygen atom of water. These are
weak van der Waals interactions.
In both complexes A and B, radical 1 acts as a Lewis acid.
While in A the p system acts as a hydrogen-bond acceptor,
in B the unpaired electron is the acceptor. Complex A is the
most stable complex (ꢀ2.39 kcalmolꢀ1 CP and ZPE correct-
ed) and is stabilized by interaction 1 between the OH group
of water and the p system of 1 (Table 1). The UB3LYP and
the UM05-2X calculated geometries of A are virtually iden-
tical.
The nonsymmetrical structure
B follows in energy
(ꢀ1.68 kcalmolꢀ1) and shows interaction 2 in which the OH
group of the water molecule interacts with the radical center
of 1. The OHC angle is 151.98 and the hydrogen-bonding
distance is 2.252 ꢃ. When the B3LYP functional is used, the
optimization of complex B leads to complex BI, with Cs
symmetry, a hydrogen-bond distance of 2.247 ꢃ, and an
OHC hydrogen-bond angle 173.68 (Figure 4). With the M05-
Figure 4. 1···H2O complexes A, B, C, and D calculated at the UM05-2X/
6-311++G(2d,2p) level of theory. The complexes B1, C1, and E are no
minima at this level and therefore the UB3LYP/6-311++G(2d,2p) geo-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
metries are presented herein. Hydrogen-bond lengths are shown in ꢃ
and angles in 8.
2X functional, the BI geometry is a transition state
8682
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8679 – 8689

Photochemistry and Reactivity of the Phenyl Radical–Water System
FULL PAPER
(25i cmꢀ1) leading to B. This behavior is not surprising, since
the potential energy surface of this kind of system is very
flat. The B3LYP functional predicts BI as the most stable
complex (ꢀ0.79 kcalmolꢀ1).
Much less stable than A and B are the Cs symmetrical
complexes C and D (ꢀ1.09 and ꢀ1.03 kcalmolꢀ1, respective-
ly) (Table 1). In these complexes, radical 1 acts as a Lewis
ꢀ
acid through its C H antibonding orbitals. They are both
stabilized by interaction 2 and differ mostly in the position
of the water molecule. In C the water molecule is located
between the ortho- and meta-hydrogen atoms of 1. In D, the
water is found between the meta- and para-hydrogen atoms
of 1.
At the B3LYP level of theory, complexes CI and E are
minima, whereas they are transition states with the M05-2X
functional. The geometry optimization of CI and E using
M05-2X leads to C and D, respectively (Figure 4). With
B3LYP the interaction energies of all complexes after CP
and ZPE corrections are below 1 kcalmolꢀ1 and thus much
smaller than at the M05-2X level. The differences in both
geometries and energies indicate that the inclusion of dis-
persion energy is crucial for the correct description of the
complexes.
Figure 5. Spin densities in the 1···H2O complexes A–D calculated at the
UM05-2X/6-311++GACTHNUTRGNE(NUG 2d,2p) level of theory.
The UM05-2X/6-311++GACTHUNGTRENNG(U 2d,2p) vibrational frequencies
of complex A are in good agreement with the experimental
values (Table 2). Since the interactions in the complexes are
very weak, the frequency shifts induced by complexation
(compared with the monomers) are small. The largest shifts
are found for the OH stretching mode of B (ꢀ88 cmꢀ1). In
A the shift of the OH stretching mode is only ꢀ30.7 cmꢀ1
(ꢀ26.4 cmꢀ1 at the B3LYP/6-311++G
ACHTUNGTREN(NGNU 2d,2p) level of
theory), which demonstrates that the stability of a hydro-
gen-bonded complex and the corresponding redshift of the
ꢀ
X H···Y vibration are not directly correlated.
In complex B the water molecule interacts directly with
the unpaired electron of the radical, which results in a con-
siderable spin transfer from 1 to the water molecule
(Figure 5). Natural bond orbital (NBO) calculations indicate
that the spin density on the water molecule is 0.02 e, mainly
localized at the oxygen atom. This spin transfer in complex
Figure 6. IR difference spectra showing the photochemical transforma-
tion of C6H5···H2O into C6H6···HO. Bands pointing downwards assigned
to C6H5···H2O disappear and bands pointing upwards assigned to
C6H6···HO appear after a) 1, b) 2, and c) 5 min irradiation at l>420 nm.
s
*
B results from the nC ! sOꢀH donor–acceptor interaction,
which contributes most to the stabilization of the complex.
For the other complexes, the distribution of the spin densi-
ties does not differ significantly from the monomers. Their
This selective photochemistry of 1···H2O allows for a de-
tailed analysis of its photochemistry. To identify the photo-
products, five isotopomers of the complex have been investi-
gated: 1···H2O, 1···D2O, 1···H218O, [D5]1···H2O, and
[D5]1···D2O.
The bands at 1482.8 and 1040.3 are very close to the in-
tense IR absorptions of matrix-isolated 2 at 1483 and
1041 cmꢀ1,[29] whereas the band at 684.2 cmꢀ1 is blueshifted
from the very strong fundamental n4 of 2 at 675.2 cmꢀ1 by
9 cmꢀ1. This indicates the formation of a complex of 2 simi-
lar to the 2···H2O complex. However, the 684.2 cmꢀ1 band is
clearly separated from the 682.0 cmꢀ1 band of 2···H2O, which
is also present in the matrix but not affected by the 420 nm
irradiation.
*
stabilization can be attributed to the pCꢀC ! sOꢀH (A) and
nsO ! sCꢀH (C and D) interactions. These donor–acceptor
*
interactions have similar contributions from the a and b or-
bitals, a behavior that has also been found in other radical–
water complexes in which the radical acts as a Lewis acid.[23]
Hydroxyl radical–2 complex: Visible-light irradiation (l>
420 nm) of an argon matrix containing the 1···H2O complex
and the other products described above (mainly matrix iso-
lated 1 and 2 and the 2···H2O complex) results in a rapid de-
crease of all IR absorptions assigned to 1···H2O, whereas
other bands are not affected. During photolysis, new bands
appear at 3502.2, 1482.8, 1040.3, and 684.2 cmꢀ1 (Figure 6).
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8683

W. Sander et al.
Table 3. Stabilization energies for the C6H6···OH complex.
If indeed 2 is formed during the photolysis of 1···H2O by
hydrogen atom transfer from water to 1, the other compo-
nent of the complex of 2 has to be the hydroxyl radical
(Scheme 3). The OH stretching vibration of the hydroxyl
6-311++G
(2d,2p)
DE
DECP
DECP+ZPE
UM05-2X
UB3LYP
UMP2
ꢀ4.71
ꢀ3.41
ꢀ4.50
ꢀ4.31
ꢀ2.97
ꢀ3.29
ꢀ3.08
ꢀ1.71
ꢀ2.80
Table 4. Experimental[a] and calculated (in italics)[b] IR spectroscopic
data of the C6H6···OH complex.
OH str.
Ring def.
Ring def.
CH o.o.p. def.
OH[c]
OD[c]
18OH[c]
3548.2
3803.6
2616.1
2769.1
3537.1
3791.1
[d]
C6H6
1483.3
1546.9
1334.7
1395.4
1480.2
1539.3
1392.6
1344.1
1041.2
1081.4
816.3
675.0
698.0
497.2
512.5
608.2
628.5
922.3
513.2
[d]
C6D6
839.6
C6H5D[d]
1036.9
1076.8
818.3
[e]
C6HD5
Scheme 3. Reaction of 1 with water.
C6H6···OH
3502.2 (ꢀ46) 1482.8 (ꢀ0.5) 1040.3 (ꢀ0.9) 684.2 (+9.2)
radical in argon has been reported at 3548.2 cmꢀ1, the new
band formed in this region during the photolysis of 1···H2O
at 3502.2 cmꢀ1 is redshifted from OH in argon by 46 cmꢀ1,
which indicates the formation of a complex between the hy-
droxyl radical and 2. Similar redshifts of the OH stretching
mode of the hydroxyl radical have been reported in the lit-
erature.[38,39]
The formation of the 2···HO complex is confirmed by iso-
topic labeling studies (Table 4) (Figure 2S in the Supporting
Information):
3759.5 (ꢀ44.1) 1545.5 (ꢀ1.4) 1080.8 (ꢀ0.6) 711.7 (+13.7)
C6HD5···OH 3502.2 (ꢀ46) 1392.2 (ꢀ0.4) 818.8 (+0.5) 927.4 (+4.1)
1343.3 (ꢀ0.8)
520.5 (+7.3)
614.2 (+6.0)
816.6 (+0.3) 504.6 (+7.4)
521.6 (+9.1)
C6H5D···OD 2583.4 (ꢀ32.7)
–
–
–
–
C6D6···OD
2583.4 (ꢀ32.7)
2737.1 (ꢀ32.0)
–
C6H6···18OH 3491.2 (ꢀ45.9) 1482.7 (ꢀ0.4) 1040.9 (ꢀ0.3) 684.2 (+9.2)
3748.6 (ꢀ42.5) 1545.1 (ꢀ1.8) 1080.6 (ꢀ0.8) 708.9 (+10.9)
[c] Argon matrix. [d] UM05-2X/6-311++GACHTNUTGRNEUNG(2d,2p). [e] Reference [40].
and UB3LYP methods with the 6-311++GACHTNUTRGNEUNG(2d,2p) basis set
1) If H218O is used in the experiments, the OH stretching vi-
bration of the 2···HO complex shows an isotopic shift of
ꢀ11 cmꢀ1, in excellent agreement with that of OH in
argon and clearly different from that of other OH-con-
taining compounds (e.g., water or phenol). Other vibra-
tions are not affected by 18O isotopic substitution.
2) If D2O is used, a deuterium atom is transferred to 1 and
thus [D1]2 is formed. This results in a large deuterium
isotopic shift of n4 of 2 from 684.2 in 2···HO to 614.2 in
[D1]2···HO, in excellent agreement with the reported iso-
topic shifts in [D1]2.[29] The OD vibration of the hydroxyl
radical is now found at 2616.1 cmꢀ1, again in perfect
agreement with the literature (Figure 2S in the Support-
ing Information).[39]
(Figure 7, Table 3). Very small spin contamination was
found for the MP2 structure. Several starting geometries
were used, however, no minima stabilized by CH···OH inter-
actions were found, and the only minima that could be lo-
cated are complexes with the OH···p interaction. With the
UB3LYP method (very tight optimizations), the O atom of
the hydroxyl radical is closer to the benzene ring than the H
atom and interacting with one of the C atoms (Figure 7
a).[14,16] This structure is an intermediate towards the addi-
tion of OH to 2. With UMP2 and UM05-2X, an OH···p
complex is predicted with the H atom of the OH radical
pointing towards the center of the p system (this complex
was found with UB3LYP only when loose optimization cri-
3) In a similar way, the photolysis of [D5]1···H2O results in
the formation of [D5]2···HO. These experiments clearly
show the transfer of hydrogen or deuterium atoms from
water to the phenyl radical.
The IR data of all isotopomers are summarized in
Table 4.
The spectroscopic data were further confirmed by calcula-
tions of the 2···HO complex using the UMP2, UM05-2X,
Figure 7. Geometry of the C6H6···HO complex calculated at several levels
of theory. a) UB3LYP/6-311++G
(2d,2p). c) UMP2/6-311++G(2d,2p).
ACHTUNGTNER(NUNG 2d,2p). b) UM05-2X/6-311++G-
A
ACHTUNGTRENNUNG
8684
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8679 – 8689

Photochemistry and Reactivity of the Phenyl Radical–Water System
FULL PAPER
teria are used, which is not advisable for these types of sys-
tems) The UMP2 and UM05-2X OH···p complexes are cal-
culated to be stabilized by ꢀ3.08 and ꢀ2.80 kcalmolꢀ1 (in-
cluding CP and ZPE corrections), respectively, whereas with
UB3LYP the binding energy of the HO···C complex is only
ꢀ1.71 kcalmolꢀ1 (Table 3).
Photochemistry of the 2···HO complex: Prolonged visible-
light irradiation (l>420 nm) of a matrix containing the
2···HO complex results in bleaching of all bands of the com-
plex and the appearance of
a very strong band at
2130.8 cmꢀ1 and several other new IR absorptions (Figure 9,
The frequencies calculated at the UM05-2X/6-311++G-
ACHTUNGTRENNUNG(2d,2p) level of theory agree very well with the experimental
IR data. In particular, the frequency shifts induced by the
formation of the complex are nicely reproduced by the cal-
culations. The largest frequency shift is observed for the OH
stretching vibration with ꢀ46 cmꢀ1. With UM05-2X, UMP2
and UB3LYP the predicted shift is ꢀ44.1, ꢀ49.1 cmꢀ1, and
+34.9 cmꢀ1, respectively, which clearly shows that B3LYP is
not reliable to predict the structure of the complex.[41–43]
The transition state a shown in Figure 8 was located for
the abstraction of a hydrogen atom from water by 1. Intrin-
sic reaction coordinate (IRC) calculations using the B3LYP
and M05-2X functionals confirm that this transition state di-
rectly connects the prereactive 1···H2O complex B with the
the 2···HO complex (Figure 8). The activation barrier for
the hydrogen abstraction starting from complex B was calcu-
lated to 13.63 (9.84) kcalmolꢀ1 with M05-2X (B3LYP). This
barrier and the fact that the calculated DE of the reaction is
2.87 (3.53) kcalmolꢀ1 with M05-2X (B3LYP) prevent a ther-
mal reaction under the conditions of matrix isolation. Our
experiments do not allow us to determine whether the hy-
drogen abstraction induced by visible-light irradiation pro-
ceeds on an excited-state surface and is thus a photochemi-
cal reaction or if the absorption of light just leads to vibra-
tional excitation (“hot-ground-state” chemistry). The re-
verse reaction, the abstraction of a hydrogen atom from 2
by a hydroxyl radical has an activation barrier of 6.05 kcal
molꢀ1 with respect to the isolated monomers and 10.76 kcal
molꢀ1 (M05-2X functional) with respect to the 2···HO com-
plex.
Figure 9. IR difference spectra showing the photochemistry of 1···H2O
and 2···HO during irradiation with l>420 nm in argon at 10 K. Bands
pointing downwards assigned to 1···H2O disappear and bands pointing
upwards assigned to ketene 5 appear after a) 1, b) 2, c) 5, d) 10, and
e) 20 min irradiation time.
Table 5). The absorption at 2130.8 cmꢀ1 is in the typical
region of ketene C=C=O stretching vibrations. The use of
H218O in the experiments results in a strong isotopic shift to
Table 5. IR spectroscopic data of the ring-opening product (Ar, 10 K).
C6H5···H2O C6H5···H218O C6H5···D2O C6D5···H2O C6D5···D2O Assignment
2130.8
1375.3
1373.5
904.7
903.0
592.4
588.9
2079.8
1375.2
1373.5
904.6
903.0
592.4
588.8
2130.7
1396.0
2130.6
2130.7
n
ACHTUNGTRENNUNG(C=C=O)
ꢀ
d
ACHTUNGTRENNUNG
747.2
736.2
727.7
wACHTUNGTRENNUNG
2079.8 cmꢀ1, thus confirming the presence of a ketene group.
Based on extensive isotopic labeling and comparison with
DFT calculations, the newly formed compound is assigned
to butadienylketene 5 (Scheme 3). Several configurations
and conformations are possible for ketene 5. However, due
to the similarity of their calculated spectra a clear assign-
ment of these geometrical isomers is not possible (Figure 3S
in the Supporting Information).
Mechanism of the ketene formation: The mechanism for the
formation of ketene 5 and a hydrogen atom by photolysis of
the 2···HO complex is not clear and requires a cascade of re-
action steps. In the low-temperature experiments, no evi-
dence for any intermediate was found for this reaction.
However, since the formation of 5, starting from azobenzene
4 and water, requires a long sequence of thermal and photo-
chemical steps, the overall yield of 5 is low and we expect to
observe intermediates only if they are formed in high sta-
tionary concentrations and if they contain strong IR chro-
Figure 8. Intrinsic reaction coordinate (IRC) for the hydrogen transfer
from water to 1.
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8685

W. Sander et al.
mophores. A band that could not be assigned at 3409.4 cmꢀ1
(shifted to 2509.8 cmꢀ1 if D2O was used) was formed as a
by-product during the photolysis of 2···HO. Products such as
phenol could be ruled out by comparison with literature
data and computed spectra.
Therefore, we undertook a detailed computational analy-
sis to shed some light into the mechanism of the formation
of 5 starting from radical 1. Although it is not clear if this is
Table 6. Relative energies of intermediates and transition states in the
3!5 reaction [kcalmolꢀ1].
UB3LYP[a]
UCCSD(T)[b]
E
EZPE
E
3
0
0
0
6+H
7
11
5-ccc+H
10
12
23.97
ꢀ5.39
42.87
70.34
49.73
28.28
15.86
50.13
29.74
25.60
44.57
78.62
51.73
72.11
49.53
39.72
81.84
50.78
57.61
51.18
94.73
18.23
ꢀ5.74
36.24
61.70
47.73
28.14
13.84
47.93
25.20
20.86
38.79
69.95
49.45
64.47
46.65
37.56
73.83
47.14
55.04
46.91
84.75
23.99
ꢀ3.83
41.24
71.19
52.84
28.32
20.40
52.14
31.27
27.57
46.64
81.06
54.74
75.23
51.21
41.46
89.04
53.55
58.00
57.54
97.37
a
photochemical or a hot-ground-state reaction, only
ground-state reactions were investigated. Four reaction
channels were calculated at the B3LYP level of theory
(Figure 10, Table 6 and Figures 9S and 10S in the Supporting
Information). Only the more reasonable channels a and b
are discussed herein, for the other two channels see the Sup-
porting Information.
13
14
TS3-6
TS6-7
TS7-10
TS11-5
TS7-10
TS10-5
TS3-12
TS12-13
TS13-5
TS3-7
TS3-14
TS12-13
TS6-11
[a] UB3LYP/6-311++G
N
[b] RHF-UCCSD(T)/6-311++G-
A
ACHTUNGTRENNUNG
Figure 10. Energy profile [kcalmolꢀ1] of reaction channels a and b from
radical 3 calculated at B3LYP/6-311++GACTHUNRGTNE(NUG 2d,2p) level of theory consid-
ering ZPVE.
Reaction channel a: The first and second reaction channels
involve the formation of phenol 6 and a hydrogen atom as
primary intermediates (Figure 10). The formation of 6 from
radical 3 has been observed in the gas phase and is consid-
ered to be of importance to tropospheric chemistry
(Scheme 4).[12,14] The barrier for the hydrogen loss is calcu-
lated to be 29.74 kcalmolꢀ1 (25.20 kcalmolꢀ1 considering
Scheme 4.
ZPE corrections) using B3LYP/6-311++G
G
(2.63 kcalmolꢀ1 with ZPE) and 3.58 kcalmolꢀ1 at the RHF-
31.27 kcalmolꢀ1 at the RHF-UCCSD(T)/6-311++G
UCCSD(T)/6-311++GACTHUGNTERN(UNG 2d,2p) level of theory. Radical 7 is
level of theory.
more stable than radical 3 by 5.74 kcalmolꢀ1.
A complex between 6 and a hydrogen atom stabilized by
weak interactions was not found either with the B3LYP or
with the M05-2X functional. Under the conditions of matrix
isolation, we expect that the hydrogen atom is trapped in
the host lattice in proximity to 6 and thus should be avail-
able for further reactions with 6.
The products of the addition of hydrogen to the meta and
para positions of 6 are radicals 8 and 9, respectively, which
are 4.34 and 3.99 kcalmolꢀ1, respectively, more stable than
3. The corresponding barriers are slightly larger than that
obtained for the ortho addition: TS6-8 2.66 kcalmolꢀ1
(3.65 kcalmolꢀ1 including ZPE) for the meta addition, and
TS6-9 2.23 kcalmolꢀ1 (3.29 kcalmolꢀ1 including ZPE) for
the para addition. Therefore, the next steps are starting
from the ortho radical 7, although for 8 and 9 also pathways
to 5 can be found.
The calculated DE for the addition of a hydrogen atom to
the ortho position of phenol 6 to produce radical 7 is
25.76 kcalmolꢀ1. The barrier for this reaction (TS6-7) is
1.63 kcalmolꢀ1 at the B3LYP/6-311++G
ACTHNUGRTENUNG(2d,2p) level
8686
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8679 – 8689

Photochemistry and Reactivity of the Phenyl Radical–Water System
FULL PAPER
Table 7. Calculated relative energies (Erel) and IR frequencies for the
ring opening product at B3LYP/6-311++GACTHUNGTRENNUNG
Isomers
Erel
]
[kcalmolꢀ1
]
n
(C=C=O)
d
N
wACTHNGURETNNU(G CH2)
w(CH)
ꢀ
5-ccc
5-cct
5-tcc
5-tct
5-ctc
5-ctt
5-ttc
5-ttt
7.56
4.89
3.43
1.27
5.41
3.63
1.75
0.00
2173.6
2182.1
2170.5
2183.4
2176.8
2181.7
2176.6
2182.5
1443.7
1456.0
1432.8
1418.4
1424.3
1429.4
1425.9
1399.1
955.5
940.1
929.2
926.3
927.1
925.5
916.6
916.7
562.2
610.8
567.9
609.1
655.8
606.4
579.2
615.2
Radical 7 can also be formed in a single step from 3 via
TS3-7. However, with
(44.53 kcalmolꢀ1 including ZPE, 50.53 kcalmolꢀ1 at the
RHF-CCSD(T)/6-311++G(2d,2p) level) the concerted re-
a
barrier of 49.96 kcalmolꢀ1
AHCTUNGTRENNUNG
action is much less favorable than the pathway via phenol 6
as an intermediate.
The tautomerism between phenol 6 and 2,4-cyclohexadie-
none (11) and the subsequent ring opening to form 5 is of
importance for the mechanism of the reaction between 2
and OH in the gas phase and therefore has been much in-
vestigated.[44–47] The calculated barrier for the 1,3-H migra-
tion in 6 to give 11 via TS6-11 is 70.76 kcalmolꢀ1 (B3LYP,
66.52 kcalmolꢀ1 including ZPE correction; 73.47 kcalmolꢀ1
with RHF-UCCSD(T), in agreement with the value of
68.7 kcalmolꢀ1 obtained by Xu and Lin at the G2//B3LYP/6-
Conclusion
Radical 1 and water easily form a weakly bound p complex
1···H2O, with a structure A similar to that of the complex
between 2 and water. A second slightly less stable, according
to our calculations, complex B with a hydrogen bond to the
radical center is not observed in the experiments. However,
complex B is the prereactive complex leading to the abstrac-
tion of a hydrogen atom from water. The activation barriers
for the rearrangements A!B and B!A have to be smaller
than the binding energy of A and B, respectively, which ex-
plains that even under the conditions of matrix isolation
only the most stable complex is observed. On the other
hand, complex B should be easily available upon thermal or
photochemical excitation of A.
According to intrinsic reaction coordinate (IRC) calcula-
tions, complex B is directly connected to the 2···HO com-
plex. The activation barrier of 13.63 (9.84) kcalmolꢀ1 at the
M05-2X (B3LYP) level of theory is too high for a thermal
reaction under the conditions of matrix isolation, however,
visible-light irradiation completely drives the reaction to the
2···HO complex. This reaction is slightly endothermic, but at
room temperature calculated to be exergonic.[8] Our experi-
ments do not allow us to differentiate between a photo-
chemical reaction and a hot-ground-state reaction. It is most
remarkable that the elusive complex between OH and 2
could be isolated and spectroscopically characterized, al-
though it exists in a very shallow minimum only and is both
thermo- and photolabile.
The structure calculated for the 2···HO complex depends
much on the theoretical method used for the calculation. At
the B3LYP level of theory, the structure is that of a shown
in Figure 7, in accordance with previous calculations.[8,14]
This structure corresponds to an intermediate to the addi-
tion of OH to 2 with the O atom interacting with one of the
ring carbon atoms. The hydrogen atom of OH only weakly
interacts with the p system of 2. In contrast, the structures
obtained at the UM05-2X or UMP2 levels of theory show a
strong OH···p interaction as in the 2···H2O complex. The sta-
bilization energy calculated with these methods is considera-
bly higher than that calculated with B3LYP. It is therefore
plausible to assume that the B3LYP calculated structure is
an artifact. This is confirmed by the comparison of the ex-
perimental IR spectrum of 2···OH with the calculated spec-
311GACHTUNGTRENNUNG
(d,p) level of theory.[47] The ring opening of 11 to 5
proceeds via TS11-5 with a barrier of 35.75 kcalmolꢀ1 (ZPE
corrected: 33.69 kcalmolꢀ1
molꢀ1).
; RHF-UCCSD(T): 40.82 kcal
Reaction channel b: This reaction path implies the ring
opening of 7 to produce radical 10 via TS7-10 (Figure 10).
The barrier is 57.12 kcalmolꢀ1 (ZPE corrected: 49.45 kcal
ꢀ1
ꢀ1
ꢀ
mol , RHF-UCCSD(T): 54.71 kcalmol ). The C C dis-
tance (2.451 ꢃ) is larger in TS7-10 than in TS11-5. The bar-
rier from 10 to 5 is 22.38 kcalmolꢀ1 with the B3LYP func-
tional (ZPE corrected: 16.74 kcalmolꢀ1) and 22.39 kcalmolꢀ1
with RHF-UCCSD(T). The structure of TS10-5 is similar to
those of the final product 5 (Figures 4S and 6S in the Sup-
porting Information).
In all reactions discussed above, ketene 5 is formed in the
cis-configuration, s-cis conformation 5-ccc (Figure 4S in the
Supporting Information). The rotations around all bonds in
5 were considered and 8 isomers were found (Figure 7S in
the Supporting Information). The most stable isomer is the
trans-configuration, anti- conformation for all bonds (5-ttt).
The isomer next in energy is 5-tct (1.27 kcalmolꢀ1), which
differs from 5-ttt in the configuration of the central bond.
The barriers for the rotation around the bonds with the
larger single bond character in the 5-ccc isomer to produce
5-cct and 5-tcc are 11.9 and 8.9 kcalmolꢀ1, respectively. The
barriers from 5-cct and 5-tcc to produce the 5-tct isomer are
4.0 and 3.7 kcalmolꢀ1. The rotation around the bond with
the larger double character in 5-tct has a barrier of 65.5 kcal
molꢀ1 to produce the most stable isomer 5-ttt. The spectra
of all isomers are very similar (Table 7), therefore, an unam-
biguous assignment is not possible.
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8687

W. Sander et al.
each case, and the PM3 and AM1[54,55] semiempirical Hamiltonians were
used for the geometry optimizations. The most relevant minima were
taken for further optimizations at higher levels of theory.
tra. The B3LYP calculations predict a blueshift of the OH
stretching vibration, which is in accordance with a C···O in-
teraction and the OH hydrogen atom not involved in hydro-
gen bonding. In contrast, the UM05-2X and UMP2 calcula-
tions predict a redshift, as expected for the OH···p interac-
tion, in excellent agreement with the experiment.
Irradiation of the 2···HO complex leads directly to ketene
5 as the major product, whereas radical 3 or other inter-
mediates are not observed. Radical 3 is the only plausible
product formed from the 2···HO complex, and all reasonable
mechanisms for the formation of 5 require the formation of
3 as an intermediate. We therefore conclude that 3 is indeed
formed, however, either with a large excess energy that
leads to further reactions or it is photochemically unstable
under the conditions of its formation. The formation of 5
and a hydrogen atom from radical 3 is calculated to be exo-
thermic by 61.7 kcalmolꢀ1. If we assume that the hydrogen
atom dimerizes to give molecular hydrogen, this value is re-
duced by 52 kcalmolꢀ1.
The DFT computations were performed by using the Gaussian 03 pro-
gram.[56] The equilibrium geometries and vibrational frequencies were ini-
tially calculated with the B3LYP functional.[57,58] Since the complexes in-
vestigated herein are only weakly interacting and the OH···p interactions
might play an important role, the hybrid meta-exchange correlation func-
tional M05-2X,[59] which performs very well for noncovalent and weak in-
teractions,[43,60–62] was also used. The Popleꢄs triple z basis set augmented
with diffuse and polarization functions 6-311++GACTHNUTRGNE(NUG 2d,2p) was em-
ployed.[63] All interaction energies were corrected for the BSSE by using
the counterpoise (CP) method developed by Boys and Bernardi.[64] The
stabilization energies were calculated by subtracting the energies of the
monomers from those of the complexes including ZPE corrections.
All complexes were calculated at the unrestricted level of theory. Since
the potential energy surface in the vicinity of the minima of the com-
plexes is very flat, the very tight convergence criterion was used for the
geometry optimizations.
Single-point RHF-UCCSD(T) calculations[65] for all complexes in the
UM05-2X/6-311++GACTHNUTRGNE(NUG 2d,2p) geometries were performed by using the
MOLPRO program.[66] For studying the reaction mechanisms, the geome-
tries of reactants, products, intermediates, and transition states were opti-
The most plausible reaction mechanism for the formation
of ketene 5 is the sequence 3!6+H!7!11+H!5+H. In
the first step, phenol (6) and a hydrogen atom is formed
from radical 3. The hydrogen atom adds to the ortho posi-
tion of 6 to give radical 7. This loses the OH hydrogen atom
to give 11, which finally ring opens to 5. Since we do not see
any of the proposed intermediates experimentally, and we
do not even know if the reaction proceeds in the ground
state or an excited state, the mechanism remains specula-
tive.
mized at the UB3LYP/6-311++GACTHNURTGNEU(GN 2d,2p) level of theory. Harmonic fre-
quencies were calculated in all cases. The energies were ZPE corrected,
too. The transition states were identified as saddle points on the potential
energies surfaces. A tight convergence criterion was used for all calcula-
tions. IRCs were calculated to confirm the connection between stationary
points.[67] Single-point RHF-UCCSD(T) calculations of all structures
were performed by using the UB3LYP/6-311++GACTHNUTRGNE(NUG 2d,2p) geometries.
The NBO analysis was performed by using NBO3.[68]
Acknowledgements
The reaction between 1 and water is highly efficient and
allows us to observe several reactive intermediates, for the
first time, that had been previously proposed in the degrada-
tion of 2. The reaction between 1 and water presented
herein allows us to gain insight into both reaction channels
of the reaction between 2 and the OH radical: the hydrogen
abstraction, coming from the product side, and the addition,
resulting in the complete destruction of the aromatic ring
system.
This work was financially supported by the Deutsche Forschungsgemein-
schaft (Forschergruppe 618) and the Fonds der Chemischen Industrie.
[1] Y. Li, L. Zhang, Z. Tian, T. Yuan, J. Wang, B. Yang, F. Qi, Energy
Fuels 2009, 23, 1473–1485.
[2] R. J. McMahon, M. C. McCarthy, C. A. Gottlieb, J. B. Dudek, J. F.
Stanton, P. Thaddeus, Astrophys. J. 2003, 590, L61L64.
[3] G. Ghigo, G. Tonachini, J. Am. Chem. Soc. 1998, 120, 6753–6757.
[4] R. Volkamer, B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N. Washida,
K. H. Becker, U. Platt, Phys. Chem. Chem. Phys. 2002, 4, 1598–
1610.
[5] A. Mardyukov, E. Sanchez-Garcia, R. Crespo-Otero, W. Sander,
Angew. Chem. 2009, 121, 4898–4901; Angew. Chem. Int. Ed. 2009,
48, 4804–4807.
[6] T. Seta, M. Nakajima, A. Miyoshi, J. Phys. Chem. A 2006, 110,
5081–5090.
[7] T. Berndt, O. Boege, Phys. Chem. Chem. Phys. 2006, 8, 1205–1214.
[8] M. P. DeMatteo, J. S. Poole, X. Shi, R. Sachdeva, P. G. Hatcher,
C. M. Hadad, M. S. Platz, J. Am. Chem. Soc. 2005, 127, 7094–7109.
[9] S. Raoult, M.-T. Rayez, J.-C. Rayez, R. Lesclaux, Phys. Chem.
Chem. Phys. 2004, 6, 2245–2253.
[10] R. H. Schuler, G. Albarran, Radiat. Phys. Chem. 2002, 64, 189–195.
[11] D. Johnson, S. Raoult, M.-T. Rayez, J.-C. Rayez, R. Lesclaux, Phys.
Chem. Chem. Phys. 2002, 4, 4678–4686.
[12] C.-C. Chen, J. W. Bozzelli, J. T. Farrell, J. Phys. Chem. A 2004, 108,
4632–4652.
[13] C.-C. Chen, J. W. Bozzelli, Chem. Phys. Proc. Combust. 2003, 13–16.
[14] I. V. Tokmakov, M. C. Lin, J. Chem. Phys. 2002, 116–117, 11309–
11326.
[15] D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank, T. Just, J. A.
Kerr, M. J. Pilling, J. Troe, R. W. Walker, J. Warnatz, J. Phys. Chem.
Ref. Data 1992, 21, 411–734.
Experimental Section
Matrix isolation: The complexes between 1 and water were generated by
codeposition of 1 and water with a large excess of argon at 10 K. In
highly diluted matrices mainly the monomers were observed by IR spec-
troscopy. The monomers were easily identified by comparison with the
data from the literature and with pure samples of matrix-isolated 1[27,48]
and water.[30,49,50] Water shows a high tendency for aggregation even in
low-temperature matrices. These aggregates have been extensively stud-
ied by IR spectroscopy.[31,32,51] Matrices with higher concentrations of
both 1 and water show IR absorptions that are not found in matrices con-
taining only one of these components. These absorptions are thus as-
signed to mixed complexes between 1 and water. By annealing the argon
matrices containing 1 and water at temperatures above 25 K, the small
trapped molecules were allowed to diffuse, which results in an increase
of the amount of aggregates. The dilution experiments in addition al-
lowed differentiating the dimers from higher aggregates.
Computations: The multiple minima hypersurface (MMH) approach[52,53]
was used to localize the minima for the 1–water system. One thousand
randomly arranged 1–water dimers were generated as starting points in
8688
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8679 – 8689

Photochemistry and Reactivity of the Phenyl Radical–Water System
FULL PAPER
[16] C. C. Chen, J. W. Bozzelli, J. T. Farrell, J. Chem. Phys. 2004, 120–
121, 4632–4652.
[17] F. P. Tully, A. R. Ravishankara, R. L. Thompson, J. M. Nicovich,
R. C. Shah, N. M. Kreutter, P. H. Wine, J. Phys. Chem. 1981, 85,
2262–2269.
[18] T. J. Wallington, D. M. Neuman, M. J. Kurylo, Int. J. Chem. Kinet.
1987, 19, 725–739.
[19] S. C. Lin, T. C. Kuo, Y. P. Lee, J. Chem. Phys. 1994, 101, 2098–2105.
[20] J. A. Manion, R. Louw, J. Phys. Chem. 1990, 94, 4127–4134.
[21] T. H. Lay, J. W. Bozzelli, J. H. Seinfeld, J. Phys. Chem. 1996, 100,
6543–6554.
[22] V. H. Uc, J. R. Alvarez-Idaboy, A. Galano, A. Vivier-Bunge, J. Phys.
Chem. A 2008, 112, 7608–7615.
[23] R. Crespo-Otero, E. Sanchez-Garcia, R. Suardiaz, L. A. Montero,
W. Sander, Chem. Phys. 2008, 353, 193–201.
[24] X. An, H. Liu, Q. Li, B. Gong, J. Cheng, J. Phys. Chem. A 2008, 112,
5258–5263.
[25] H. Hernandez-Soto, F. Weinhold, J. S. Francisco, J. Chem. Phys.
2007, 127, 164102.
[26] J. M. Merritt, S. Rudic, R. E. Miller, J. Chem. Phys. 2006, 124,
084301.
[27] A. V. Friderichsen, J. G. Radziszewski, M. R. Nimlos, P. R. Winter,
D. C. Dayton, D. E. David, G. B. Ellison, J. Am. Chem. Soc. 2001,
123, 1977–1988.
[28] A. Mardyukov, W. Sander, Chem. Eur. J. 2009, 15, 1462–1467.
[29] A. Engdahl, B. Nelander, J. Phys. Chem. 1985, 89, 2860–2864.
[30] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1641–1650.
[31] J. Ceponkus, G. Karlstroem, B. Nelander, J. Chem. Phys. 2005, 122–
123, 7859–7864.
[32] A. Engdahl, B. Nelander, J. Chem. Phys. 1987, 86, 4831–4837.
[33] Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157–167.
[34] Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2008, 112, 1095–1099.
[35] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1641–1650.
[49] G. P. Ayers, A. D. E. Pullin, Chem. Phys. Lett. 1974, 29, 609–615.
[50] J. P. Perchard, Chem. Phys. 2001, 273, 217–233.
[51] Y. Bouteiller, J. P. Perchard, Chem. Phys. 2004, 305, 1–12.
[52] L. A. Montero, A. M. Esteva, J. Molina, A. Zapardiel, L. Hernan-
dez, H. Marquez, A. Acosta, J. Am. Chem. Soc. 1998, 120, 12023–
12033.
[53] L. A. Montero, J. Molina, J. Fabian, Int. J. Quantum Chem. 2000, 79,
8–16.
[54] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart, J. Am.
Chem. Soc. 1985, 107, 3902–3909.
[55] J. J. P. Stewart, J. Comput. Chem. 1990, 11, 543–544.
[56] Gaussian 03, Revision E.01, M. J. Frisch, H. B. Schlegel, G. E. Scuse-
ria, J. R. C. M. A. Robb, J. A. Montgomery, Jr., K. N. Kudin, J. M.
Millam, S. S. Iyengar, J. Tomasi, B. M. V. Barone, M. Cossi, G. Scal-
mani, N. Rega, H. N. G. A. Petersson, M. Hada, M. Ehara, K.
Toyota, J. H. R. Fukuda, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, M. K. H. Nakai, X. Li, J. E. Knox, H. P. Hratchian, J. B.
Cross, J. J. C. Adamo, R. Gomperts, R. E. Stratmann, O. Yazyev,
R. C. A. J. Austin, C. Pomelli, J. W. Ochterski, P. Y. Ayala,
G. A. V. K. Morokuma, P. Salvador, J. J. Dannenberg, S. D. V. G.
Zakrzewski, A. D. Daniels, M. C. Strain, D. K. M. O. Farkas, A. D.
Rabuck, K. Raghavachari, J. V. O. J. B. Foresman, Q. Cui, A. G.
Baboul, S. Clifford, B. B. S. J. Cioslowski, G. Liu, A. Liashenko, P.
Piskorz, R. L. M. I. Komaromi, D. J. Fox, T. Keith, M. A. Al-Laham,
A. N. C. Y. Peng, M. Challacombe, P. M. W. Gill, W. C. B. Johnson,
M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian Inc., Pittsburgh PA,
2003.
[57] C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[58] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[59] Y. Zhao, N. E. Schultz, D. G. Truhlar, J. Chem. Theory Comput.
2006, 2, 364–382.
[60] Y. Zhao, D. G. Truhlar, J. Chem. Phys. 2006, 124-125, 5121–5129.
[61] Y. Zhao, D. G. Truhlar, J. Chem. Theory Comput. 2006, 2, 1009–
1018.
[62] Y. Zhao, D. G. Truhlar, J. Chem. Phys. 2006, 124–125, 10478–10486.
[63] R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem. Phys.
1980, 72, 650–654.
[36] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1629–1639.
[37] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1689–1693.
[38] A. Engdahl, G. Karlstrom, B. Nelander, J. Chem. Phys. 2003, 118,
7797–7802.
[39] A. Engdahl, B. Nelander, J. Chem. Phys. 2005, 122, 126101.
[40] L. H. P. Weldon, C. L. Wilson, J. Chem. Soc. 1946, 235–239.
[41] L. F. Holroyd, T. van Mourik, Chem. Phys. Lett. 2007, 442, 42–46.
[42] R. Crespo-Otero, L. A. Montero, W. D. Stohrer, J. M. G. de La Vega,
J. Chem. Phys. 2005, 123, 134107.
[43] Y. Zhao, D. G. Truhlar, J. Chem. Theory Comput. 2007, 3, 289–300.
[44] R. Cypres, B. Bettens, Tetrahedron 1974, 30, 1253–1260.
[45] A. B. Lovell, K. Brezinsky, I. Glassman, Int. J. Chem. Kinet. 1989,
21, 547–560.
[64] S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553.
[65] G. D. Purvis, R. J. Bartlett, J. Chem. Phys. 1982, 76, 1910–1918.
[66] MOLPRO, Version 2006.1, H.-J. Werner, P. J. Knowles, R. Lindh,
F. R. Manby, M. Schꢀtz, P. Celani, T. Korona, A. Mitrushenkov, G.
Rauhut, T. B. Adler, R. D. Amos, A. Bernhardsson, A. Berning,
D. L. Cooper, M. J. O. Deegan, A. J. Dobbyn, F. Eckert, E. Goll, C.
Hampel, G. Hetzer, T. Hrenar, G. Knizia, C. Kçppl, Y. Liu, A. W.
Lloyd, R. A. Mata, A. J. May, S. J. McNicholas, W. T. Meyer, M. E.
Mura, A. Nicklaß, P. Palmieri, K. Pflꢀger, R. Pitzer, M. Reiher, U.
Schumann, H. Stoll, A. J. Stone, R. Tarroni, T. Thorsteinsson, M. C.
Wang, A. Wolf.
[67] C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154–2161.
[68] NBO Version 3.1, E. D. Glendening, A. E. Reed, J. E. Carpenter, F.
Weinhold.
[46] L. Zhu, J. W. Bozzelli, J. Chem. Phys. 2003, 118-119, 3696–3703.
[47] Z. F. Xu, M. C. Lin, J. Chem. Phys. 2006, 124-125, 1672–1677.
[48] J. G. Radziszewski, M. R. Nimlos, P. R. Winter, G. B. Ellison, J. Am.
Chem. Soc. 1996, 118, 7400–7401.
Received: December 8, 2009
Revised: June 2, 2010
Published online: July 14, 2010
Chem. Eur. J. 2010, 16, 8679 – 8689
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8689