Infrared spectrum of 2-hydroxyphenoxyl radical and photoisomerization between trans and cis 2-hydroxyphenyl nitrites

Article abstract of DOI:10.1016/j.cplett.2006.10.109

Photoreaction of 2-nitrophenol has been investigated by low-temperature matrix-isolation IR spectroscopy and density-functional-theory (DFT) calculation. A reaction intermediate was identified as syn-trans 2-hydroxyphenyl nitrite, and the reaction product caused by dissociation of NO from the intermediate as 2-hydroxyphenoxyl radical. Both syn-trans and syn-cis 2-hydroxyphenyl nitrite isomers were produced by recombination of 2-hydroxyphenoxyl radical with NO in annealing at 27 K after the UV irradiation. The conformational change from cis to trans around the O-N axis occurred upon visible-light irradiation. The rate constants of photoreaction were determined by a least-squares fitting of the changes in the IR band intensities.

Full text of DOI:10.1016/j.cplett.2006.10.109

Chemical Physics Letters 432 (2006) 446–451
www.elsevier.com/locate/cplett
Infrared spectrum of 2-hydroxyphenoxyl radical and
photoisomerization between trans and cis 2-hydroxyphenyl nitrites
*
Maki Nagaya, Satoshi Kudoh, Munetaka Nakata
Graduate School of BASE (Bio-Applications and Systems Engineering), Tokyo University of Agriculture and Technology, Naka-cho,
Koganei, Tokyo 184-8588, Japan
Received 15 September 2006; in final form 24 October 2006
Available online 10 November 2006
Abstract
Photoreaction of 2-nitrophenol has been investigated by low-temperature matrix-isolation IR spectroscopy and density-functional-
theory (DFT) calculation. A reaction intermediate was identified as syn-trans 2-hydroxyphenyl nitrite, and the reaction product caused
by dissociation of NO from the intermediate as 2-hydroxyphenoxyl radical. Both syn-trans and syn-cis 2-hydroxyphenyl nitrite isomers
were produced by recombination of 2-hydroxyphenoxyl radical with NO in annealing at 27 K after the UV irradiation. The conforma-
tional change from cis to trans around the O–N axis occurred upon visible-light irradiation. The rate constants of photoreaction were
determined by a least-squares fitting of the changes in the IR band intensities.
Ó 2006 Elsevier B.V. All rights reserved.
1. Introduction
NO, the energy barrier being calculated to be
262 kJ molÀ1 [5].
Nitro–nitrite isomerism has recently drawn much
attention in experimental and theoretical chemistry. For
example, the photolysis of nitrobenzene 1 is known to
produce phenoxyl radical 3, as shown in Scheme 1 [1–
9]. This indicates the existence of phenyl nitrite 2, an
isomer of nitrobenzene. Since this intermediate 2 is a
strong absorber of the second photon leading to rapid
photodissociation of NO from the nitrite, no infrared
(IR) spectrum was identified even by the method of
matrix-isolation [5].
The existence of phenyl nitrite 2 was confirmed in a
recent study [6], which identified a reverse reaction in a
low-temperature argon matrix by annealing of the co-
photoproducts, phenoxyl radical 3 and NO. This photo-
reaction mechanism has been supported by quantum
chemical calculations [5,7–10]; the nitro–nitrite rearrange-
ment is the most plausible pathway for production of
We have recently studied the photolysis of 2-nitrophe-
nol (2-NP) 4, which has a strong intramolecular hydro-
gen bond between the NO2 and OH groups [11]. We
have detected the aci-nitro isomer 5 (Scheme 2) pro-
duced by overcoming the lowest energy barrier on the
potential surface by irradiation from a 248-nm KrF exci-
mer laser, where the hydrogen atom of the OH group
moves to the nitro group through the hydrogen bond.
This reaction is found to be reversible; the aci-nitro iso-
mer returned to the original 2-NP upon visible-light irra-
diation (k > 450 nm).
The present Letter reports on our further study of the
photoreaction mechanism of 2-NP. Besides the reversible
reaction from 4 to 5 by UV irradiation and from 5 to 4
by visible-light [11], we have discovered a photoreaction
pathway of nitro–nitrite isomerism, which produces 2-
hydroxyphenoxyl (2-HPO) radical 7 through 2-hydroxy-
phenyl nitrite (2-HPN) 6 as an intermediate species
(Scheme 3). We have assigned the observed IR spectra of
these species and determined their conformations by com-
parison with the spectral patterns calculated by density-
*
Corresponding author. Fax: +81 42 388 7349.
E-mail address: necom@cc.tuat.ac.jp (M. Nakata).
0009-2614/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2006.10.109

M. Nagaya et al. / Chemical Physics Letters 432 (2006) 446–451
447
NO
O
2. Experimental and calculation methods
NO
O
O
N
λ> 250 nm
λ> 250 nm
O
The details of the experimental and calculation methods
are described in Ref. [11]. The deposition time was ꢀ1.5 h.
UV radiation from a superhigh-pressure mercury lamp
(USHIO Model UI-501 C) was used to induce photoreac-
tion. Cutoff optical filters, UV36 (k > 350 nm, HOYA),
L42 (k > 410 nm, Toshiba) and Y52 (k > 510 nm, HOYA),
were used to choose radiation energy, and a water filter to
remove extra thermal reaction. The DFT calculations were
performed using the GAUSSIAN 03 program [12] with the
basis set of 6-31++G** or 6-311++G** to optimize the
geometrical structures and to estimate the vibrational
wavenumbers.
nitrobenzene 1
phenyl nitrite 2
phenoxyl radical 3
Scheme 1.
O
O-
OH
O
N
λ = 248 nm
λ > 450 nm
N
OH
O
2-nitrophenol 4
aci-nitrophenol 5
Scheme 2.
3. Results and discussions
3.1. Identification of photoproducts
NO
OH
O
N
OH NO
O
An IR spectrum of 2-NP in an argon matrix is reported
in Ref. [11]. When the matrix sample was exposed to UV
and visible-light (k > 220 nm) from the mercury lamp with-
out short-cutoff optical filters, new product IR bands
appeared; they differed from those of the aci-nitro isomer
produced by 248-nm radiation. A ‘difference spectrum’
was obtained by subtracting the spectrum measured before
the 800-min irradiation from that after the irradiation, as
shown in Fig. 1a, where the decreasing bands were assigned
to the reactant, 2-NP, and the increasing bands to a mix-
ture of the photoproducts. From the dependence of the
band intensity on the irradiation time, as described in
OH
λ > 220 nm
λ> 220 nm
O
O
2-NP 4
2-HPN 6
2-HPO radical 7
Scheme 3.
functional-theory (DFT) calculations. The proposed
assignments are confirmed by a kinetic analysis of the
changes in their band intensities.
Fig. 1. Observed and calculated IR spectra of 2-HPO: (a) Observed difference spectrum between the spectra measured after minus before UV irradiation
(k > 220 nm) for 800 min, (b) and (c) calculated spectral patterns of syn and anti 2-HPO, respectively, obtained at the DFT/B3LYP/6-31++G** level,
where the scaling factors of 0.95 and 0.98 were used in the regions higher and lower than 3000 cmÀ1, respectively. The intensity in the high-wavenumber
region is expanded three times from that in the low-wavenumber region.

448
M. Nagaya et al. / Chemical Physics Letters 432 (2006) 446–451
each other in Table 1. The 2-HPO radical was investigated
previously in the photolysis of catechol with OH radical by
ESR spectroscopy [13] and IR spectroscopy with modu-
lated electric excitation [14], but its geometrical structure
and vibrational spectrum except for the OH stretching
region has never been reported.
Fig. 2. Optimized structures of syn and anti 2-HPO obtained at the DFT/
B3LYP/6-31++G** calculation level.
In the region of the O–H stretching mode in Fig. 1a, a
slightly broad band of syn 2-HPO radical appeared at
3336 cmÀ1, which is close to the value of 3318 cmÀ1
observed in the CCl4 solution [14]. This red shift from that
of the free O–H stretching mode suggests that this radical
has an intramolecular hydrogen bond between the O and
H group and the O atom with an unpaired electron,
although it is weaker than that of 2-NP, which has the cor-
detail below, the product bands were classified into four
groups.
To identify the photoproducts, the calculated spectral
patterns of some candidates were compared with the
observed spectra. As a result, it is found that the calculated
spectral pattern of 2-HPO radical can satisfactorily repro-
duce the observed bands marked (m), where the conforma-
tion around the C–O axis for the H atom of the OH group
is found to be syn against the OÅ atom but not anti (see
Fig. 2). The former is predicted to be more stable than
the latter by 36.7 kJ molÀ1. The observed and calculated
wavenumbers and relative intensities are compared with
responding broad band at 3230 cmÀ1
.
This interpretation is supported by the optimized
structures obtained by our DFT calculation, where the
˚
O–H bond length of syn isomer, 0.983 A, is longer than
˚
that of anti isomer, 0.968 A, as shown in Fig. 2. The C–
O–H angle of the syn isomer, 105°, is smaller than that
of anti isomer, 110°. A similar hydrogen bond between
the OH group and the C atom with an unpaired electron
in benzene ring was pointed out in the optimized struc-
tures of 2-hydroxyphenyl radical [15], but its stabilization
energy, 7.7 kJ molÀ1, is much smaller than that of 2-
HPO.
Table 1
Observed and calculated wavenumbers (in cmÀ1) of syn 2-hydroxyphen-
oxyl (2-HPO) radical
The band marked (n) appearing at 1868 cmÀ1 is assign-
able to NO. This band shows a red shift of 4 cmÀ1 from an
isolated NO in an argon matrix [16], implying that it inter-
acts with syn 2-HPO radical.
Observeda
Calculated
Wavenumberb
Intensity
3336 (m)
3377
3059
3055
3043
3031
38
1
3
4
1
When the matrix sample was annealed at 27 K, the
intensities of the 2-HPO (m) and NO (n) bands
decreased, while the product bands marked (ꢁ) increased,
as shown in Fig. 3a. In addition, new bands marked (ꢂ)
appeared near the bands marked (ꢁ). It is highly plausi-
ble that thermal recombination between 2-HPO and NO
produces 2-hydroxyphenyl nitrite (2-HPN), similarly to
the result of annealing for co-photoproducts of nitroben-
zene [6], as described in Section 1. Thus, we performed
DFT calculations for 2-HPN to confirm the identification
of these new product bands marked (ꢁ) or (ꢂ).
1579 (w)
1585
1550
28
15
1518 (m)
1457 (s)
1412 (w)
1503
1468
1411
1356
1314
26
100
6
20
38
c
–
1304 (m)
1240 (m)
1230 (m)
1240
44
The four isomers for 2-HPN, displayed in Fig. 4, were
optimized in our calculations at the DFT/B3LYP/6-
31++G**level. The first Letter S or A denotes syn or anti
around the C–O axis for the hydrogen atom of the OH
group against the ONO group similarly to the 2-HPO rad-
ical. The second Letter T or C denotes trans or cis around
the O–N bond similarly to HONO. The most stable isomer
among the four is ST, which is less stable than 2-NP by
20.8 kJ molÀ1. The second stable isomer is SC, the energy
1169 (m)
1151 (w)
1107 (w)
1170
1147
1109
993
953
925
22
<1
3
1
<1
2
864
833
2
<1
770 (m)
752
744
709
16
5
45
difference against ST being calculated to be 4.6 kJ molÀ1
.
Two anti isomers, AT and AC, are less stable than the
syn isomers, ST and SC, by ꢀ10 kJ molÀ1, probably
because of the repulsion between the lone pairs of the
OH and ONO groups. Another isomer for ST, with all
atoms planar, was optimized at the 6-311++G**, but gave
an imaginary wavenumber.
a
Letters in parentheses represent relative intensities: ‘s’, ‘m’, and ‘w’
means strong, medium, and weak, respectively.
b
Calculated at the DFT/B3LYP/6-31++G** level. Scaling factors of
0.95 and 0.98 are used in the regions higher and lower than 3000 cmÀ1
respectively.
,
c
Disturbed by strong bands of 2-NP.

M. Nagaya et al. / Chemical Physics Letters 432 (2006) 446–451
449
Fig. 3. Observed and calculated IR spectra of 2-HPN: (a) observed difference spectrum after minus before annealing at 27 K, (b) observed difference
spectrum between the spectra measured after 10-min visible-light (k > 410 nm) irradiation subsequent to annealing, and (c) calculated spectral patternsÀo1f
2-HPN, obtained at the DFT/B3LYP/6-31++G** level, where scaling factors of 0.95 and 0.98 were used in the regions higher and lower than 3000 cm
respectively. Up and down represent ST and SC, respectively. The scale of expansion for each region is marked on top.
,
1700 cmÀ1, and the CC ring stretching mode around
1500 cmÀ1 consist of two peaks marked (ꢂ) and (ꢁ). This
implies that two isomers of 2-HPN are produced. When
the matrix sample was exposed to the visible-light through
a short-cutoff optical filter (k > 410 nm), the (ꢂ) bands
decreased and the (ꢁ) bands increased. This means that
photoisomerization between the (ꢂ) and (ꢁ) species
occurred in the visible-light. By comparison of the
observed difference spectrum shown in Fig. 3b with the cal-
culated spectral patterns obtained by the DFT calculation,
Fig. 3c, the bands marked (ꢂ) are assigned to SC, and those
marked (ꢁ) to ST. Our assignments are confirmed by the
observation of split bands in the O–H and N = O stretch-
ing modes; the O–H stretching mode of ST is calculated to
be 13 cmÀ1 higher than that of SC, which is consistent with
the observed 3585 (ꢁ) and 3572 (ꢂ) cmÀ1 bands, while the
N = O stretching mode of ST is calculated to be 15 cmÀ1
lower than that of SC, which is consistent with the
observed 1735 cmÀ1 (ꢂ) and 1722 cmÀ1 (ꢁ) bands. This con-
formational isomerism was also supported by our TD-
DFT calculations, where the wavelength of absorption by
SC is longer than that by ST. The observed and calculated
wavenumbers and relative intensities of 2-HPN are listed in
Table 2.
H
H
O
O
O
O
O
N
O
N
O
ST
SC
H
O
O
N
(20.8)
(25.4)
O
H
H
N
O
N
O
O
O
2-NP
(0.0)
O
AT
AC
(36.0)
(33.0)
Fig. 4. Calculated relative energies (in kJ molÀ1
) of four possible
conformational isomers of 2-HPN against 2-NP. The former Letter S or
A represents syn or anti with O–H, while the latter Letter C or T represents
the conformation around the O–NO bond for the position of the N–O
bond against the benzene ring. Optimized structural parameters are
available upon request.
In the difference spectrum (Fig. 3), the observed bands
corresponding to the O–H stretching mode around
3600 cmÀ1
,
the N = O stretching mode around

450
M. Nagaya et al. / Chemical Physics Letters 432 (2006) 446–451
Table 2
3.2. Kinetic analysis
Observed and calculated wavenumbers (in cmÀ1) of 2-hydroxyphenyl
nitrite (2-HPN)
As described above, 2-NP changes to 2-HPN upon
UV irradiation (k > 220 nm). 2-HPN immediately
absorbs the second photon and releases NO to produce
2-HPO radical. Unlike phenyl nitrite, 2-HPN is stabilized
by intramolecular hydrogen bonding. Then 2-HPN was
detected as a photoreaction intermediate. To confirm this
reaction pathway, we examined the dependence of the
band intensities on the irradiation time. The depletion
of the IR band for 2-NP (1130 cmÀ1) and the growth
of those for 2-HPO radical (3336 cmÀ1), NO (1868
cmÀ1), and 2-HPN (1722 cmÀ1) are drawn in Fig. 5. 2-
HPN rapidly increases in the beginning of the photoreac-
tion and then slowly increases. The growth speed for NO
seems to be equal to that for 2-HPO, suggesting that
both species are produced simultaneously. Although
syn-cis 2-HPN may be produced by the photolysis of
2-NP and recombination of 2-HPO radical and NO, only
syn-trans 2-HPN was observed upon UV–visible irradia-
tion, suggesting that the isomerization from cis to trans
is very rapid.
Thus we assumed Scheme 4 for our least-squares fitting
to obtain the rate constants. The ratio of absorption
coefficients for 2-HPO and NO radical, e2-HPO/eNO, was
estimated from the observed bands, since they are co-prod-
ucts, and was fixed to 1.22. The e2-HPN/eNO was also fixed
to 8.98, estimated from the observed bands of syn-trans
2-HPN and NO measured under suitable conditions: by
410-nm irradiation to decompose most part of syn-cis 2-
HPN, and then by 350-nm irradiation to decompose the
remaining syn-trans 2-HPN to 2-HPO and NO. The
observed value, 8.98, is roughly consistent with 12.33 calcu-
lated by the DFT method. The e2-NP/eNO was treated as a
parameter, since no exact value can be estimated from
the observed spectra.
As shown in Fig. 5, the calculated values obtained by
a least-squares fitting are consistent with the correspond-
ing observed values within experimental errors. The rate
constants obtained are 0.0112 0.0002, 1.2 0.1, and
0.12 0.03 hÀ1 for k1, k2, and k3, respectively. The k2
value exceeds k1 by two orders of magnitude; this differ-
ence is probably caused by the performance of the used
mercury lamp that the photon numbers of k > 220 nm
for reaction of 2-HPN is ꢀ100 times higher than that
of k < 250 nm for reaction of 2-NP. The recombination
rate k3 of 2-HPO with NO is 10 times larger than the
k1 value. It is plausible that the absorption of UV–visi-
ble-light by 2-NP and/or 2-HPN causes local heating in
the argon matrix [19], like the annealing at 27 K. The
ratio of the absorption coefficients, e2-NP/eNO = 3.45, is
about four times larger than the calculated value, 0.89.
This difference may originate from the minor photoreac-
Observeda
Calculatedb
ST
SC
Wavenumberc
3589
Intensity
17
wavenumberc
intensity
3585 (w)
3572 (w)
3576
3057
3051
3041
3127
24
2
3
2
<1
3058
3052
3043
3032
1
2
2
<1
1735 (m)
1722 (s)
1791
100
1776
1618
1610
100
4
6
1617
1604
7
10
1497 (m)
1493 (m)
1494
25
1490
1470
1353
1304
35
1
4
1472
1355
1304
<1
2
4
5
1267 (w)
1260 (m)
1217 (w)
1211 (w)
1178 (w)
1264
1213
1168
18
20
11
1258
1207
31
29
1173 (w)
1166
1154
16
2
1155
1092
1
2
1100 (w)
1099 (w)
1091
1028
960
3
4
<1
1
1030
957
924
887
3
<1
2
931
6
880 (m)
882
848
14
844
825
2
9
<1
829 (m)
751 (m)
825
750
5
23
749 (w)
749
719
21
25
730
8
a
Letters in parentheses represent relative intensities: ‘s’, ‘m’, and ‘w’
means strong, medium, and weak, respectively.
b
Conformations are drawn in Fig. 4.
Calculated at the DFT/B3LYP/6-31++G** level. Scaling factors of
c
0.95 and 0.98 are used in the regions higher and lower than 3000 cmÀ1
respectively.
,
Similar conformational changes from cis to trans in
HONO and CH3ONO are known to be induced by IR
radiation, while visible irradiation induces release of
NO [17,18]. As for 2-HPN, the band intensities of 2-
HPO and NO also increased slightly in Fig. 3. This
means that visible irradiation (k > 410 nm) induces pho-
todissociation with photoisomerization. The dissociation
pathway is found to be negligible upon irradiation with
k > 510 nm.
tions of 2-NP to produce other species marked ( )
*
besides 2-HPN. The photoreaction pathways of 2-NP
including the previous result [11] is summarized in
Scheme 5.

M. Nagaya et al. / Chemical Physics Letters 432 (2006) 446–451
451
Fig. 5. Irradiation time dependence of IR band intensity. Symbols (n), (ꢁ), (D) and (m) represent the band intensities for 2-nitrophenol (1130 cmÀ1), syn-
trans 2-HPN (1722 cmÀ1), NO (1868 cmÀ1) and 2-HPO radical (3336 cmÀ1). Solid lines represent calculated values obtained by a least-squares fitting,
where Scheme 4 is assumed.
References
OH
OH
O
N
OH
k2
k3
O
k1
O
O
N
O
[1] T.B. Brill, K.J. James, Chem. Rev. 93 (1993) 2667.
[2] D.B. Galloway, T. Glenewinkel-Meyer, J.A. Bartz, L.G. Huey, F.F.
Crim, J. Chem. Phys. 100 (1994) 1.
NO
4
6a
7
[3] T. Yu, A.M. Mebel, M.C. Lin, J. Phys. Org. Chem. 8 (1995) 47.
[4] F. Berho, F. Caralp, M.-T. Rayez, R. Lesclaux, E. Ratajczak, J. Phys.
Chem. A 102 (1998) 1.
Scheme 4.
[5] J. Spanget-Larsen, M. Gil, A. Gorski, D.M. Blake, J. Waluk, J.G.
Radziszewski, J. Am. Chem. Soc. 123 (2001) 11253.
[6] R. Yang, X. Jin, W. Wang, K. Fan, M. Zhou, J. Phys. Chem. A 109
(2005) 4261.
[7] D.B. Galloway, J.A. Bartz, L.G. Huey, F.F. Crim, J. Chem. Phys. 98
(1993) 2107.
NO
OH
O
OH
OH
O
O
O
N
λ > 220 nm
λ > 220 nm
N
O
Annealing
[8] T. Glenewinkel-Meyer, F.F. Crim, J. Mol. Struct. (Theochem) 337
(1995) 209.
7
6a
4
[9] G.M. Khrapkovskii, A.G. Shamov, G.A. Shamov, V.A. Shlyapochni-
kov, Zh. Org. Khim. 35 (1999) 891.
[10] S. Xu, M.C. Lin, J. Phys. Chem. B 109 (2005) 8367.
[11] M. Nagaya, S. Kudoh, M. Nakata, Chem. Phys. Lett. 427 (2006) 67.
[12] M.J. Frisch et al., GAUSSIAN 03, Revision B.04, Gaussian Inc.,
Pittsburgh, PA, 2003.
OH
O
O
N
O
N
OH
O
[13] H. Yoshioka, Y. Ohashi, H. Fukuda, Y. Senda, H. Yoshioka, J. Phys.
Chem. A 107 (2003) 1127.
[14] M. Foster, K. Lith, M. Andrist, U.P. Fringeli, Hs.H. Gunthard,
Chem. Phys. 17 (1976) 59.
6b
Scheme 5.
5
[15] M. Nagata, Y. Futami, N. Akai, S. Kudoh, M. Nakata, Chem. Phys.
Lett. 392 (2004) 259.
[16] M. Nakata, H. Frei, J. Am. Chem. Soc. 111 (1989) 5240.
[17] J.D. Baldeschwieler, G.C. Pimentel, J. Chem. Phys. 33 (1960) 1008.
[18] J.F. Ogilvie, J. Chem. Soc. Chem. Comm. (1973) 450.
[19] E.I. Tarasova, A.M. Ratner, V.M. Stepanenko, I.Ya. Fugol, J. Chem.
Phys. 98 (1993) 7786.
Acknowledgement
The authors thank Professor Kozo Kuchitsu (BASE,
Tokyo University of A and T) for his helpful discussions.