Interaction and reaction of the phenyl radical with water: A source of OH radicals

Full text of DOI:10.1002/anie.200806268

Communications
DOI: 10.1002/anie.200806268
Radicals
Interaction and Reaction of the Phenyl Radical with Water:
A Source of OH Radicals**
Artur Mardyukov, Elsa Sanchez-Garcia, Rachel Crespo-Otero, and Wolfram Sander*
The phenyl radical (1) and the hydroxyl radical OHC are
highly reactive intermediates that play key roles in many
reaction mechanisms. Gas-phase reactions of these radicals
are of particular importance for our understanding of
processes such as the combustion of hydrocarbons at high
temperatures or the degradation of aromatic compounds in
the troposphere at ambient temperature. Reaction with the
hydroxyl radical is the only known pathway for the removal of
benzene (2) from the troposphere. Owing to this fundamental
importance, a large number of experimental and theoretical
Scheme 1. Pathways for the reaction of the hydroxyl radical with
benzene (2).
studies in this field have been described.[1–8]
An interesting, little-studied aspect of radical chemistry is
the role of weakly bound noncovalent complexes.[9] Com-
plexes between radicals and water and other highly polar
compounds might stabilize the radicals and thus modulate
their reactivity. The method of choice for investigating these
complexes is matrix-isolation spectroscopy.[10,11] Although the
inert-gas matrix introduces some degree of distortion into the
trapped complexes, it could be shown in many instances that
the spectra of even very weakly bound van der Waals
complexes can be reproduced well by gas-phase calculations.
Herein, we present evidence that both the phenyl radical (1)
and the hydroxyl radical form noncovalent complexes that are
of importance for understanding the reactivity of these
species.
There are two pathways for the reaction of benzene (2)
with OH: I) hydrogen abstraction to give radical 1 and water,
and II) addition to yield the 2-hydroxycyclohexadienyl radical
3 (Scheme 1). At higher temperatures, phenol (4) can be
formed from 3 through the loss of a hydrogen atom. The
thermochemistry of these reactions was evaluated by using
G3 theory.[7] According to these calculations, the hydrogen
abstraction, I, is exothermic by 4.2 kcalmolÀ1 with an
activation energy of 5.3 kcalmolÀ1. The calculated reaction
energy is in accordance with a simple estimation from the
experimental bond enthalpies (DH298) of the CH bond in
benzene (112.9 kcalmolÀ1) and the OH bond in water
(118.8 kcalmolÀ1): an enthalpy difference of 5.9 kcalmolÀ1.[12]
The addition reaction, II, is more exothermic (À16.2 kcal
molÀ1) according to the G3 calculation,[7] and the activation
barrier (2.8 kcalmolÀ1) is lower than that of reaction I.
To gain insight into the systems benzene/OHC and phenyl
radical/H2O, we investigated the photochemical reaction of
the phenyl radical with water in low-temperature argon
matrices. The phenyl radical (1) can be produced in high
yields by flash vacuum pyrolysis of azobenzene (5) at 5008C
and trapping of the products in argon at 10 K (see the
Supporting Information for experimental details).[13] The only
by-products observed by IR spectroscopy in these matrices
are benzene (2), formed by hydrogen abstraction from surface
contaminants of the pyrolysis oven, and traces of acetylenes
and other alkynes formed by ring opening of the phenyl
radical at higher temperatures.
When the matrix is doped with 1% water (Scheme 2),
several additional bands, which were absent for matrices
containing only the pyrolysis products of 5 or only 1% water,
were observed in the IR spectrum. The intensity of these
bands increased when the matrix was annealed at temper-
atures between 25 and 35 K. At these temperatures, the
diffusion of matrix-isolated water becomes rapid, and aggre-
[*] A. Mardyukov, Dr. E. Sanchez-Garcia, 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
Homepage: http://www.rub.de/oc2
[**] This research was supported financially by the Deutsche For-
schungsgemeinschaft (Forschergruppe 618) and the Fonds der
Chemischen Industrie.
Supporting information for this article (including experimental and
computational details) is available on the WWW under http://dx.
doi.org/10.1002/anie.200806268.
Scheme 2. Synthesis of the phenyl radical (1) and its reaction with
water.
4804
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 4804 –4807

Angewandte
Chemie
Table 1: IR spectroscopic data for the complex between the phenyl
gates of water molecules as well as aggregates between water
and other trapped molecules are formed.
radical (1) and water. Calculated frequencies (UM05-2X/6-311++G-
(2d,2p)) are given in italics underneath the experimental frequencies
(argon, 10 K); the frequency shift upon the formation of the complex
from the corresponding monomers is given in parenthesis.
One set of the new bands was assigned readily to the well-
known complex between benzene (2) and water.[14] In the
2···H2O complex, the strong out-of-plane (oop) deformation
mode of 2 at 675.0 cmÀ1 is blue-shifted to 682.0 cmÀ1
(+7.0 cmÀ1; Figure 1). This shift was taken as evidence for
OH stretch
CH oop deform.
CH oop deform.
H2O[14,15,20]
D2O[14,15,20]
H218O[21]
3639.4
3894.2
2658.8
2806.5
3630.7
3886.1
[16]
C6H5
705.8
734.0
518.0
657.4
681.8
–
[16]
C6D5
535.3
C6H5···H2O
C6D5···H2O
C6H5···D2O
C6D5···D2O
C6H5···H218O
3618.0 (À21.4)
3863.5 (À30.7)
3618.0 (À21.4)
3863.5 (À30.7)
2645.1 (À13.7)
2787.1 (À19.4)
2645.1 (À13.7)
2787.1 (À19.4)
3610.5 (À20.2)
3854.7 (À31.4)
711.4 (+5.6)
739.3 (+5.3)
522.3 (+4.3)
540.0 (+4.7)
711.8 (+6.0)
739.3 (+5.3)
522.3 (+4.3)
539.9 (+4.6)
711.5 (+5.7)
739.3 (+5.3)
659.8 (+2.4)
682.6 (+0.8)
–
659.8 (+2.4)
682.6 (+0.8)
–
659.5 (+2.1)
682.6 (+0.8)
The structure, energy of complexation, and IR spectra of
several weakly bound complexes between 1 and H2O were
calculated by using the UM05-2X functional[17] with a large
6-311 + + G(2d,2p) basis set (see the Supporting Information
for computational details). Since standard functionals, such as
B3LYP, fail to estimate dispersion energies (and thus are not
reliable for the calculation of weakly bound complexes, such
as 1···H2O), and CCSD(T) with a large basis set is prohib-
itively expensive, we used the UM05-2X functional, which
was developed for the calculation of van der Waals com-
plexes. The most stable complex, A (Scheme 2), is stabilized
by an OH···p hydrogen bond by À3.77 kcalmolÀ1 (Figure 2).
This value is in excellent agreement with the binding energy
of À3.69 kcalmolÀ1 obtained from a single-point coupled-
cluster calculation (RHF-UCCSD(T)/6-311 + + G(2d,2p)//
UM05-2X/6-311 + + G(2d,2p)). After corrections for the
zero-point vibrational energy (ZPE) and basis set super-
position errors (BSSE), the binding energy of complex A is
still À2.39 kcalmolÀ1, which demonstrates that radical 1 is
stabilized considerably by the interaction with water. Other
complexes between 1 and water have also been identified, but
these complexes are less stable. A comparison of the matrix
IR spectrum of the 1···H2O complex with that calculated for A
reveals an excellent agreement and thus confirms the assign-
ment (Table 1).
Figure 1. IR spectra showing the formation of a complex between the
phenyl radical (1) and water in solid argon. a) IR spectrum obtained
after trapping the products of the flash vacuum pyrolysis (FVP) of
azobenzene (5) in argon doped with 1% water at 10 K. b) IR spectrum
of the matrix in (a) after warming to 40 K and cooling back down to
10 K. c) IR spectrum after FVP of the perdeuterated [D10]5 as a
precursor of [D5]1 under conditions similar to those for (a). d) IR
spectrum of the matrix in (c) after warming to 40 K and cooling back
down to 10 K.
an interaction of the water molecule with the p system of 2.
A second set of new IR bands was observed at 3618.0, 711.4,
and 659.8 cmÀ1. These bands are close to the water OH
stretching vibration at 3639.4 cmÀ1[15] and to two b1-sym-
metrical oop deformation vibrations of radical 1 at 705.8 and
657.4 cmÀ1.[16] The band at 711.4 cmÀ1 is blue-shifted from the
band of 1 at 705.8 cmÀ1 by 5.6 cmÀ1, in the same way as the
band of the 2···H2O complex at 682.0 cmÀ1 is blue-shifted from
the corresponding band of 1. It was thus tempting to conclude
that this new compound was a complex between 1 and water.
To confirm this assignment, five isotopomers of the complex,
C6H5···H2O, C6D5···H2O, C6H5···D2O, C6D5···D2O, and
C6H5···H218O, were matrix-isolated (Table 1). For all iso-
topomers, the shifts of the IR bands were very similar to
those of the corresponding isotopomers of the benzene–water
complex.
Complex A proved to be photolabile upon irradiation
with near-UV light. Irradiation of the matrix for several
minutes with light of wavelength l > 350 nm resulted in the
bleaching of A (all three observed IR bands disappeared
simultaneously) and the formation of a new compound B with
IR bands at 3502.2, 1482.8, 1040.3, and 684.2 cmÀ1 (Figure 3,
Table 2). Other bands in the spectrum were not affected (for
example, the benzene–water complex does not show any
Angew. Chem. Int. Ed. 2009, 48, 4804 –4807
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4805

Communications
Figure 2. Structures and some geometrical data of the complexes
between the phenyl radical and water (top) and benzene and the
hydroxyl radical (bottom).
photochemistry). The vibrations of B at 1482.8, 1040.3, and
684.2 cmÀ1 closely resemble the vibrations of matrix-isolated
2 at 1483.3, 1040.8, and 675.0 cmÀ1.[14] In compound B, the oop
deformation mode of benzene is blue-shifted to 684.2 cmÀ1, in
analogy with the blue shift observed for the benzene–water
complex.[14] We therefore concluded that B was a p complex
of benzene. The other two benzene vibrations in B showed
very small red shifts of less than 1 cmÀ1 with respect to “free”
matrix-isolated 2 but could be assigned clearly to B and not to
free benzene by monitoring the intensity changes during
irradiation.
Figure 3. IR difference spectra showing the photochemical transforma-
tion of 1···H2O into 2···OH. A=absorbance. a) Bands pointing down-
wards assigned to 1···H2O disappear and bands pointing upwards
assigned to 2···OH appear after irradiation for 5 min with light of
wavelength l>350 nm. b) The equivalent experiment with H218O.
When D2O was used in the experiments, the n4 vibration
of benzene (2) in B showed a very large isotope shift to
614.2 cmÀ1, as expected for [D1]benzene (Table 2; see also the
Supporting Information). The band is blue-shifted by 6.0 cmÀ1
from that observed for matrix-isolated [D1]2 and thus
indicates the formation of a p complex with [D1]2 as one
component. This result demon-
cates the formation of a strong hydrogen bond, as expected
for B. With D2O instead of H2O, the vibration at 3502.2 cmÀ1
was red-shifted by 918.8 cmÀ1, and with H218O by 11 cmÀ1
(Table 2). These results prove that this band is indeed due to
an OH stretching vibration. The presence of other OH-
strates clearly that radical
1
abstracts a deuterium atom from
D2O to give [D1]2 during irradia-
tion. What remains from the water
molecule is an OD radical (or an
OH radical if unlabeled water is
used). It was thus tempting to con-
clude that compound B is a weakly
Table 2: IR spectroscopic data for the complex between benzene (2) and the hydroxyl radical. Calculated
frequencies (UM05-2X/6-311++G(2d,2p)) are given in italics underneath the experimental frequencies
(argon, 10 K); the frequency shift upon the formation of the complex from the corresponding monomers
is given in parenthesis.
OH stretch
ring deform.
ring deform.
CH oop deform.
OH[18,22]
3548.2
3803.6
2616.1
3537.1
bound complex between 2 and the OD[18,22]
18OH[18,22]
OH radical.
This assignment was corrobo-
rated by an analysis of the OH-
stretching region of the IR spec-
trum of B. The fourth observed IR
absorption of B at 3502.2 cmÀ1 is
red-shifted by 46 cmÀ1 with respect
to the OH stretching vibration of
the OH radical at 3548.2 cmÀ1 in
solid argon.[18,19] This red shift indi-
[16]
C6H6
1483.3
1040.8
675.0
1546.3
1334.7
1480.2
1081.1
816.3
1036.9
698.0
497.2
608.2
[16]
C6D6
C6H5D[16]
C6H6···OH
3502.2 (À46)
3759.5 (À44.1)
3502.2 (À46)
2583.4 (À32.7)
2583.4 (À32.7)
3491.2 (À45.9)
1482.8 (À0.5)
1040.3 (À0.5)
684.2 (+9.2)
711.7 (+13.7)
–
614.2 (+6.0)
504.6 (+7.4)
684.2 (+9.2)
1545.0 (À1.3)
1080.1 (À1.0)
C6HD5···OH
C6H5D···OD
C6D6···OD
–
–
–
–
–
–
C6H6···18OH
1482.7 (À0.6)
1040.9 (+0.1)
4806 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 4804 –4807

Angewandte
Chemie
containing compounds, such as water or phenol, can be
excluded by comparison of their known matrix IR spectra
with the matrix IR spectrum of B.
expect that this phenomenon is also of importance for the
understanding of radical reactions in biological media.
The observation that the 18O isotopic shift of the OH
radical in B of À11 cmÀ1 is identical to that of the matrix-
isolated OH radical and markedly different from that of H2O
serves as additional evidence that B is a complex of the OH
radical. Finally, the calculated IR spectrum of the 2···OH
p complex is in excellent agreement with the matrix IR
spectrum of B (Table 2). In particular, the predicted shifts of
the IR bands of the OH and benzene fragments match the
experimental values almost perfectly.
Received: December 22, 2008
Revised: March 22, 2009
Published online: May 19, 2009
Keywords: hydrogen bonds · IR spectroscopy · matrix isolation ·
.
phenyl radical · water
[1] T. Seta, M. Nakajima, A. Miyoshi, J. Phys. Chem. A 2006, 110,
5081 – 5090.
The crucial question remains as to why a complex between
benzene 2 and the OH radical is observed at all: Why does
this complex not immediately react further to give radical 3
(pathway II, Scheme 1)? The activation barriers for this
reaction have been predicted at the B3LYP level of theory
by Tokmakov and Lin[7] using a 6-311 + + G(d,p) basis set to
be absent (to À0.8 kcalmolÀ1) and by Hadad and Platz et al.[8]
with a smaller basis set (6-311 + G(d,p)) to 1.22 kcalmolÀ1. G3
[2] B. Bohn, C. Zetzsch, Phys. Chem. Chem. Phys. 1999, 1, 5097 –
5107.
[3] 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.
[4] C.-C. Chen, J. W. Bozzelli, J. T. Farrell, J. Phys. Chem. A 2004,
108, 4632 – 4652.
[5] S. Raoult, M.-T. Rayez, J.-C. Rayez, R. Lesclaux, Phys. Chem.
Chem. Phys. 2004, 6, 2245 – 2253.
[6] J. S. Poole, X. Shi, C. M. Hadad, M. S. Platz, J. Phys. Chem. A
2005, 109, 2547 – 2551.
[7] I. V. Tokmakov, M. C. Lin, J. Phys. Chem. A 2002, 106, 11309 –
11326.
theory gave a barrier of 2.8 kcalmolÀ1 [7]
, which is still very low,
even under the conditions of matrix isolation in cryogenic
matrices.
However, the barrier is considerably higher if OH and
benzene form a stabilized prereactive complex. The stabiliza-
tion energy of such a prereactive complex had been calculated
by Tokmakov and Lin[7] to be À1.8 and À2.6 kcalmolÀ1 at the
B3LYP/6-311 + + G(d,p) and G3 level of theory, respectively.
The complex was found to be very flexible with respect to
movements of the OH molecule across the benzene p system.
At the UM05-2X/6-311 + + G(2d,2p) level of theory we
found a stabilization energy of À4.71 kcalmolÀ1 for complex
B. As expected, this value is considerably larger than the
stabilization calculated with the B3LYP functional. This
stabilization increases the barrier to the addition reaction
and thus enables the isolation of the OH complex in the
matrix. It is also interesting to note that complex B is slightly
more stable than complex A. As a result, the endothermicity
of hydrogen-atom transfer from water to the phenyl radical is
decreased.
[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] R. Crespo-Otero, E. Sanchez-Garcia, R. Suardiaz, L. A. Mon-
tero, W. Sander, Chem. Phys. 2008, 353, 193 – 201.
[10] A. Mardyukov, E. Sanchez-Garcia, P. Rodziewicz, N. L. Dolt-
sinis, W. Sander, J. Phys. Chem. A 2007, 111, 10552 – 10561.
[11] E. Sꢀnchez-Garcꢁa, A. Mardyukov, A. Tekin, R. Crespo-Otero,
L. A. Montero, W. Sander, G. Jansen, Chem. Phys. 2008, 343,
168 – 185.
[12] S. J. Blanksby, G. B. Ellison, Acc. Chem. Res. 2003, 36, 255 – 263.
[13] A. Mardyukov, W. Sander, Chem. Eur. J. 2009, 15, 1462 – 1467.
[14] A. Engdahl, B. Nelander, J. Phys. Chem. 1985, 89, 2860 – 2864.
[15] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1629 – 1639.
[16] 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.
[17] Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157 – 167.
[18] B. M. Cheng, Y. P. Lee, J. F. Ogilvie, Chem. Phys. Lett. 1988, 151,
109 – 115.
[19] A. Engdahl, G. Karlstrom, B. Nelander, J. Chem. Phys. 2003, 118,
7797 – 7802.
[20] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1641 – 1650.
[21] G. P. Ayers, A. D. E. Pullin, Spectrochim. Acta Part A 1976, 32,
1689 – 1693.
In summary, we have isolated and characterized com-
plexes between the phenyl radical and water and between the
hydroxyl radical and benzene for the first time. We observed
light-driven hydrogen-atom transfer from water to the phenyl
radical within these complexes. Our results demonstrate the
importance of radical complexes for the radical reactivity. We
[22] A. Engdahl, B. Nelander, J. Chem. Phys. 2005, 122, 126101.
Angew. Chem. Int. Ed. 2009, 48, 4804 –4807
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4807