Photochemistry of substituted 4-halogenophenols: Effect of a CN substituent

Article abstract of DOI:10.1039/a900141g

The photochemistry of 5-chloro-2-hydroxybenzonitrile 1 was studied in aqueous solution using transient absorption spectroscopy and product analysis. The triplet carbene 3-cyano-4-oxocyclohexa-2,5-dienylidene 2 (λ_max/nm 385, 368) was successfully detected and identified on the basis of its characteristic reactivity. This transient species is converted into cyanobenzo-1,4-quinone-O-oxide (λ_max/nm 470) by reaction with oxygen and is reduced into 2-cyanophenoxyl radical (λ_max/nm 402, 387) by propan-2-ol. The product studies confirm the intermediary formation of carbene 2. 2,5-Dihydroxybenzonitrile 3 and the biphenyls 4 and 5 are primary photoproducts in deoxygenated solutions whereas 2-hydroxybenzonitrile 9 and 5-bromo-2-hydroxybenzonitrile 10 are cleanly produced upon addition of propan-2-ol (0.13 M) and bromide ions (10^-2 M), respectively. In oxygen-saturated solutions, cyanobenzo-1,4-quinone 8 is the main photoproduct. The quantum yield of carbene formation (0.062) is reduced by 40% in the presence of oxygen and is increased up to a value of 0.20 upon addition of bromide or iodide ions. These results can be interpreted in terms of triplet quenching and heavy-atom enhancement and support the assumption that the carbene 2 is formed from the triplet excited state of 1; this assumption is supported by a detailed study of the phototransformation of 1 in ethanol. Mono- and biphotonic formations of solvated electrons and 4-chloro-2-cyanophenoxyl radicals (λ_max/nm 427, 408) are also observed from neutral 1. The effects of CN substitution can be traced to deprotonation of the lowest excited singlet state on the one hand (pK* = 0.12 ± 0.04) and to an increase of the triplet lifetime on the other.

Full text of DOI:10.1039/a900141g

View Article Online / Journal Homepage / Table of Contents for this issue
Photochemistry of substituted 4-halogenophenols: effect of a CN
substituent
a
b
a
a
Florent Bonnichon, Gottfried Grabner, Ghislain Guyot and Claire Richard *
a
Laboratoire de Photochimie Moléculaire et Macromoléculaire, UMR 6505,
Université Blaise Pascal (Clermont-Ferrand II), F-63177 Aubière Cedex, France
Institut für Theoretische Chemie und Strahlenchemie, Universität Wien, Althanstrasse 14,
b
A-1090 Wien, Austria
Received (in Cambridge) 5th January 1999, Accepted 22nd April 1999
The photochemistry of 5-chloro-2-hydroxybenzonitrile 1 was studied in aqueous solution using transient absorption
spectroscopy and product analysis. The triplet carbene 3-cyano-4-oxocyclohexa-2,5-dienylidene 2 (λmax/nm 385, 368)
was successfully detected and identified on the basis of its characteristic reactivity. This transient species is converted
into cyanobenzo-1,4-quinone-O-oxide (λmax/nm 470) by reaction with oxygen and is reduced into 2-cyanophenoxyl
radical (λmax/nm 402, 387) by propan-2-ol. The product studies confirm the intermediary formation of carbene 2. 2,5-
Dihydroxybenzonitrile 3 and the biphenyls 4 and 5 are primary photoproducts in deoxygenated solutions whereas
2
-hydroxybenzonitrile 9 and 5-bromo-2-hydroxybenzonitrile 10 are cleanly produced upon addition of propan-2-ol
Ϫ2
(
0.13 M) and bromide ions (10 M), respectively. In oxygen-saturated solutions, cyanobenzo-1,4-quinone 8 is the
main photoproduct. The quantum yield of carbene formation (0.062) is reduced by 40% in the presence of oxygen
and is increased up to a value of 0.20 upon addition of bromide or iodide ions. These results can be interpreted in
terms of triplet quenching and heavy-atom enhancement and support the assumption that the carbene 2 is formed
from the triplet excited state of 1; this assumption is supported by a detailed study of the phototransformation of 1
in ethanol. Mono- and biphotonic formations of solvated electrons and 4-chloro-2-cyanophenoxyl radicals (λmax/nm
4
27, 408) are also observed from neutral 1. The effects of CN substitution can be traced to deprotonation of the
lowest excited singlet state on the one hand (pK* = 0.12 ± 0.04) and to an increase of the triplet lifetime on the other.
Introduction
bromo to iodo, in contrast to what would be expected on the
grounds of an internal heavy-atom effect.
We reasoned that the substitution of 4-halogenophenols by
an electroattracting group such as the CN group would change
the photoreactivity. First, the substitution of phenols by a CN
Halogenophenols exhibit an interesting photoreactivity in
aqueous solution (see refs. 1 and 2 for reviews). In the case of 4-
halogenophenols, heterolytic photoelimination of the halo-
gen in polar protic solvents leads to formation of the carbene 4-
oxocyclohexa-2,5-dienylidene (λmax/nm 370, 384) whose lifetime
in water is long enough for detection on the microsecond time-
substituent is known to lower the pK values. In particular,
a
2
-hydroxybenzonitrile is a strong acid in the first singlet excited
9
state (pK* = 0.66). Second, the presence of an electroattracting
group is expected to hinder the loss of the halide ion and
may lengthen the lifetime of the excited states. In the hope of
gaining further insights into the mechanism of the carbene
formation, we studied the photolysis of 5-chloro-2-hydroxy-
benzonitrile 1 using transient absorption spectroscopy and
product studies techniques.
3
scale. This carbene can be trapped by oxygen with formation
4,5
of benzo-1,4-quinone-O-oxide (λmax/nm 460), can add on the
starting 4-halogenophenol with production of adducts (λmax/
3
nm 405) and can abstract an H-atom from alcohols yielding
3
the phenoxyl radical (λmax/nm 385, 400). In the case of
4
-chlorophenol, the quantum yield of the carbene formation is
very high (0.75). All the photoproducts observed under con-
tinuous irradiation (hydroquinone, benzo-1,4-quinone, coupl-
ing products and phenol) can be understood from the reactions
of the carbene with water, oxygen, the starting halogenophenol
and H-donor molecules. The carbene exhibits triplet reactivity;
it was assumed to be formed directly from the triplet excited
Experimental
General
5-Chloro-2-hydroxybenzonitrile was purchased from May-
bridge (England) and was recrystallized from water prior to use.
Water was purified with a Milli-Q (Millipore) device. All other
3
state of the 4-halogenophenol, but no experimental evidence
could be presented in favour of this assumption. Alternatively,
it might be proposed that the carbene is initially formed in the
singlet state and is then converted into the triplet by rapid inter-
system crossing. However, how efficient this intersystem cross-
ing step would have to be is easily seen from the fact that singlet
carbenes very efficiently insert into the O–H bond of water
molecules.
This highly efficient photoreactivity of 4-halogenophenols
has subsequently been corroborated by other authors, using
Fourier-transform electron spin resonance spectroscopy or
1
reagents were used as received. H NMR spectra were recorded
on a Bruker AC400 spectrometer. Mass analyses were per-
formed by the Service d’analyse of the Université d’Orléans,
France. Ammonia was used for chemical ionisation mass
spectra. UV–visible spectra were recorded on a Cary 3 (Varian)
spectrophotometer. Analytical HPLC was carried out on a
Waters apparatus equipped with a photodiode array detector
and a conventional reverse phase 5 µm column. The following
eluting gradient was used: from 0 to 30 min mixture water with
6
7
nanosecond transient absorption spectroscopy. Recently, the
H PO4 (0.1%)–MeOH (55:45, v/v) then at 35 min mixture
3
assumption of carbene formation from the triplet excited state
of the phenol was questioned on the grounds that the photo-
transformation quantum yield of 4-halogenophenols decreases
when the substituent is changed from fluoro over chloro and
water with H PO4 (0.1%)–MeOH (30:70, v/v). Preparative
3
8
HPLC was performed on a Gilson apparatus with UV detec-
tion using a semi-preparative microsorb 3 µm column and
water–methanol mixtures (55:45, v/v) and (30:70, v/v) as elu-
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210
1203

View Article Online
ents. Chloride ions were titrated using a Waters Quanta 4000
capillary HPLC equipped with a capillary column (0.75 µm,
Ϫ
0
.06 m). The electrolyte was CrO4 and surfactant Waters
Ϫ
OFM-OH . The pK of the ground state 1 was evaluated by a
spectrophotometric method. Phosphate buffers (6 × 10 M)
a
Ϫ3
were used to adjust the pH within the range 5.67–6.96 and
HClO and NaOH to adjust the pH at 3 and 12. The pK was
4
a
computed using the relationship: pK = pH ϩ log(A Ϫ A)/
a
b
(
A Ϫ A ), where A , A and A are absorbances measured at 336
a b a
nm in solutions at pH = 12, 3 and at variable pH, respectively.
Fluorescence measurements
A Perkin-Elmer MPF-3L spectrofluorimeter equipped with an
IP 28 photomultiplier was used for the fluorescence measure-
ments. Fluorescence spectra were recorded using solutions of 1
Ϫ4
at a concentration of 10 M. The pH was adjusted with HClO4
or NaOH solutions. The Hammett acidity scale H was used
0
10
Fig. 1 Emission spectra of 1 1) in water pH = 4.4; 2) in water pH = 0.8;
3) in water pH = 12; 4) in ethanol. Inset: dependence of the fluorescence
instead of the pH scale when HClO concentrations exceeded
4
Ϫ3
0
3
.1 M. Phenol (10 M) was used as the standard (Φ = 0.13 at
f
11
intensity measured at 340 nm on pH or H . Excitation wavelength set at
0
00 K) for the evaluation of fluorescence quantum yields. All
3
15 nm.
measurements were made at room temperature. Fluorescence
lifetimes were measured by a single-photon-correlated tech-
5Љ-H), 7.69 (1H, d, J 2.6, 4-H), 7.75 (1H, d, J 2.6, 6-H), 7.78
12
nique as described earlier.
(1H, dd, J 2.9 and 8.8, 6Љ-H), 7.87 (2H, d, J 2.9, 2Љ-H and
ϩ
4
Ј-H), 7.93 (1H, d, J 2.9, 2Ј-H); CI-NH m/z 388 and 400 (MH ,
3
Nanosecond laser flash photolysis
25%).
Authentic samples of 2,5-dihydroxybenzonitrile 3 and cyano-
Transient absorption experiments were carried out using a
frequency-quadrupled Nd:YAG laser (Quanta-Ray DCR-1,
benzo-1,4-quinone 8 were synthesized using the following pro-
cedures. A solution of 2-hydroxybenzonitrile (4.8 g, 0.04 mol),
K S O (10.8 g, 0.04 mol) and NaOH (0.8 g, 0.02 mol) was
2
66 nm, pulse duration 10 ns). The procedures used for transi-
2
2
8
ent absorption spectroscopy measurements have been described
3,13
stirred at room temperature for 48 h. 5-Sulfo-2-hydroxybenzo-
nitrile potassium salt was recovered after cooling the solution
previously. For some experiments requiring low pulse ener-
gies (P/mJ < 0.1), the excitation wavelength was set at 280 nm
using the frequency-doubled output of a parametric oscillator
14
and was hydrolyzed into 3 upon stirring in hot acidic solution.
Finally, 3 was recovered by recrystallization. An acidic mixture
(
Lambda Physik Scanmate OPPO) pumped by the third
of 3 (10 mg, 0.074 mmol) and NaNO (14 mg, 0.2 mmol) was
2
harmonic of an Nd:YAG laser (Coherent Infinity).
stirred for 8 h at room temperature. NaNO was extracted with
2
ether and 8 finally recovered after evaporation.
Steady-state irradiations
3
2
,5-Dihydroxybenzonitrile 3: λ /nm (H O) 317 (ε/dm
max 2
Ϫ1 Ϫ1
For quantum yield measurements and analytical runs, photol-
yses were carried out at 313 nm using a Bausch and Lomb
monochromator equipped with a high pressure mercury lamp
mol cm 3500 ± 200); δ (400 MHz; CD OD) 7.00 (1H, dd,
H
3
J 2.9 and 8.4, 4-H), 7.10 (1H, d, J 2.9, 6-H), 7.24 (1H, d, J 8.4,
ϩ
ϩ
3-H); m/z 135 (M , 100%) and 107 (M Ϫ 18).
3
(
200 W). Potassium ferrioxalate was used as chemical actin-
Cyanobenzo-1,4-quinone 8: λ /nm (H O) 254, 446 (ε446/dm
max
2
Ϫ1
Ϫ1
ϩ
ometer. Measurements were made in triplicate. For preparative
separations, samples were irradiated in a device equipped with
six sunlamps emitting from 280 to 350 nm with a maximum of
emission located at 310 nm. Solutions were deoxygenated by
nitrogen or argon bubbling.
mol cm 1800 ± 200); m/z 133 (M , 100%).
5-Bromo-2-hydroxybenzonitrile 10 was produced by irradiat-
Ϫ3
ing a deoxygenated solution of 1 (10 M) containing NaBr
Ϫ2
(10 M) at 313 nm for 15 min. After irradiation, the solution
was evaporated to dryness. The residue was dissolved in meth-
anol and analyzed by mass spectrometry. m/z 197 and 199 (M ,
ϩ
Identification of products
100%).
Ϫ3
Two 100 ml portions of 1 (2 × 10 M) were saturated with N2
for 30 min prior to the irradiation, irradiated for 1 h and finally
Results
evaporated to 20 ml. Four photoproducts were isolated by pre-
Absorption and fluorescence measurements
1
parative HPLC. The assignments were based on H NMR and
mass spectra.
The maxima of the longest-wavelength (S ) absorption bands
1
3
2
-Hydroxy-5-chloro-6-(3Ј-cyano-4Ј-hydroxyphenyl)benzo-
of neutral and anionic 1 are located at λ /nm 306 (ε/dm
max
Ϫ1
Ϫ1
3
Ϫ1
Ϫ1
nitrile 4: λmax/nm (H O) 309; δ (400 MHz; CD OD) 7.13 (1H,
mol cm 3400 ± 100) and λmax/nm 335 (ε/dm mol cm
2
H
3
d, J 9.0, 3-H), 7.25 (1H, d, J 8.8, 5Ј-H), 7.63 (1H, dd, J 2.0 and
4800 ± 200) respectively. The pKa of the ground state was
9
2
.0, 6Ј-H), 7.71 (1H, d, J 2.0, 2Ј-H), 7.73 (1H, d, J 8.9, 4-H); m/z
70 and 272 (M , 50%).
evaluated as 6.44 ± 0.03. This value is close to that measured
for 2-hydroxybenzonitrile (6.97 ± 0.02).
ϩ
9
2
-Hydroxy-5-chloro-3-(3Ј-cyano-4Ј-hydroxyphenyl)benzo-
The fluorescence emission spectra of 1 are shown in Fig. 1.
The emission maxima are located at λmax/nm 348 for neutral
1 and λmax/nm 396 for anionic 1. The pH dependence of the
fluorescence intensity of neutral 1 is depicted in the inset of
Fig. 1. The fluorescence intensity measured at 340 nm decreases
3
Ϫ1
Ϫ1
nitrile 5: λmax/nm (H O) 313 (ε/dm mol cm 7500 ± 700); δH
400 MHz; CD OD) 7.20 (1H, d, J 8.8, 5Ј-H), 7.59 (1H, d, J 2.6,
-H), 7.61 (1H, d, J 2.6, 6-H), 7.80 (1H, dd, J 2.3 and 8.8, 6Ј-H),
.90 (1H, d, J 2.3, 2Ј-H); m/z 270 and 272 (M , 100%).
2
(
4
7
3
ϩ
2
-Hydroxy-5-(3Ј-cyano-4Ј-hydroxyphenyl)benzonitrile
6:
as H or the pH is increased from Ϫ1.0 to 2.0. A plateau value
0
λmax/nm (H O) 265, 306; δH (400 MHz; CD OD) 7.20 (2H, d,
corresponding to 35% of the intensity measured for H = Ϫ1.05
2
3
0
J 8.7, 3-H and 5Ј-H), 7.86 (2H, dd, J 2.5 and 8.9, 4-H and 6Ј-H),
is reached at pH = 2. The decrease of fluorescence intensity
results from the dissociation of the phenolic group in the lowest
excited singlet state after excitation of neutral molecules. From
the plateau value, it can be deduced that 65% of the neutral
ϩ
7
.90 (2H, d, J 2.4, 6-H and 2Ј-H); m/z 236 (M , 50%).
-Hydroxy-5-chloro-3-(5Ј,3Љ-dicyano-6Ј,4Љ-dihydroxybiphen-
yl)ylbenzonitrile 7: δH (400 MHz; CD OD) 7.24 (1H, d, J 8.8,
2
3
1
204
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210

View Article Online
Fig. 2 Hydrated electron formation from 1: plot of A715/P versus P
Fig. 3 Transient spectrum (᭺) measured at pulse end in a deoxygen-
(
(
P = laser pulse energy). All data normalized to the same absorbance
0.5) at the wavelength of excitation. ᮀ: pH = 4, ᭹: pH = 11.
Ϫ3
Ϫ1
ated solution of 1 (2 × 10 M) at pH = 4.0; P = 0.56 mJ pulse . ᭹:
spectrum obtained after correction for the contribution of the 4-chloro-
2
-cyanophenoxyl radical (dashed line).
excited state molecules deprotonate at pH values exceeding pH
. From this pH dependence, the pK * of the singlet excited
2
a
state of 1 was evaluated as 0.12 ± 0.04.
The fluorescence quantum yields were estimated by com-
11
parison with the emission of phenol (Φ = 0.13 at 300 K). The
f
values obtained for neutral and anionic 1, respectively, are
0
.007 and 0.03. The lifetime of the singlet excited states was
evaluated from fluorescence decay measurements as 0.32 ns
in ethanol, 0.34 ns in water at pH = 0.8, 0.42 ns in water at
pH = 4.4 and 0.52 ns in water at pH = 12.
Laser flash photolysis studies
Formation of hydrated electrons. The characteristic absorp-
tion of hydrated electrons was detected at pulse end on excit-
ation of both neutral and anionic 1 with 10 ns pulses (λexc/nm
2
66). Since phenolic compounds are susceptible to mono-
photonic as well as biphotonic ionization processes via their S
1
15
Ϫ
state, the absorbance, A , at the maximum of e absorption
7
15
aq
Fig. 4 Absorption spectra of the 4-chloro-2-cyanophenoxyl radical
(ᮀ) and of the 2-cyanophenoxyl radical (᭹) produced by photolysis of
Ϫ1
Ϫ1
16
(
λ/nm 715, ε715/M cm 19000) was monitored as a function
of the laser pulse energy, P. Plots of A715/P versus P measured
for neutral and anionic 1 are linear and parallel (Fig. 2). The
linearity of this function is predicted for a mixed one- and
two-photon mechanism according to eqn. (1).
acidic solutions containing K S O (0.1 M) and 1 or 2-hydroxybenzo-
2 2 8
Ϫ4
nitrile (5 × 10 M, absorbance multiplied by 2.5).
therefore be generated with the same yield. There was no indi-
cation of a corresponding transient at λ/nm > 350.
2
A715 = a × P ϩ b × P
(1)
Formation of 4-chloro-2-cyanophenoxyl radicals. The transi-
ent spectrum in the near-UV region obtained on laser flash
photolysis of 1 at pH values between 1 and 6 is dominated by
the strong absorption of anionic 1 arising from excited-state
deprotonation and extending up to 350 nm. A search for other
photoinduced transients had therefore to be focused on longer
wavelengths.
The coefficients a and b in this expression are related to the
efficiencies of the one- and two-photon ionization processes,
respectively. The relatively small value obtained for the coef-
ficient b (the slope of the lines in Fig. 2) is consistent with the
short S lifetime of 1. The one-photon ionization quantum yield
can be determined from the coefficient a (the intercepts in Fig.
1
2
) to be Φ = 0.023 ± 0.005 at pH = 4 and Φ = 0.034 ± 0.005 at
The photolysis of slightly acidic (pH around 4.0) and
deoxygenated solutions of 1 (2 × 10 M) produces at pulse end
the transient spectrum shown in Fig. 3. Four maxima appear
within the wavelength range 350–440 nm, suggesting the pres-
ence of more than one species. Since photoionization will pro-
Ϫ3
pH = 11. Φ is lower at pH = 4 than at pH = 11 by 32%; this is in
line with the fluorescence-based result that 65% of 1 deproton-
ate in the S state. Moreover, the observation that the slopes of
the regression lines shown in Fig. 2 are parallel indicates that
the same excited state absorbs the second photon in the two-
photon ionization process, irrespective of the pH value. It can
therefore be concluded that ionization of neutral 1 arises from
1
duce 4-chloro-2-cyanophenoxyl radicals concomitantly with
aq, this radical is expected to contribute to the overall spec-
Ϫ
e
trum. In order to study it independently, we generated it by
sulfate radical-induced one-electron oxidation in an acidic solu-
its S state after excited-state deprotonation.
1
Ϫ
Ϫ4
The e formed by photoionization of 1 decay by a pseudo-
tion of K S O (0.1 M) and 1 (5 × 10 M) irradiated at 266 nm.
aq
2
2
8
first-order reaction with the substrate; the rate constants of this
The resulting radical spectrum, shown in Fig. 4, exhibits two
bands at λmax/nm 427, 408 which conform to the two longer-
wavelength bands of the overall spectrum in Fig. 3. The
4-chloro-2-cyanophenoxyl radical spectrum, shown as a dashed
line in Fig. 3, was then subtracted from the overall spectrum
assuming that it is the only species contributing to it at λ/
nm > 420. The spectrum remaining after subtraction shows
again two maxima, λmax/nm 385, 368.
Ϫ1 Ϫ1
10
9
reaction are k/M
s
= 1.4 × 10 at pH 4 and 8.6 × 10 at pH
17 Ϫ
1
1. As in the case of halobenzonitriles the e are expected to
aq
add to the substrate molecules. In order to check a possible
contribution of radical anions of 1 to the transient absorption
4Ϫ
spectrum, a mixture of 1 and Fe(CN) Ϫ6 was subjected to 266
nm pulses at pH 4. In this system, e are produced with a
aq
18
quantum yield of the order of 0.5, and radical anions should
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210
1205

View Article Online
Fig. 5 Pulse energy dependence of absorbances measured in a
Ϫ3
deoxygenated solution of 1 (2 × 10 M) at pH = 4.0. ᭹: absorbance at
4
05 nm measured at the end of pulse, ᭿: absorbance at 405 nm, differ-
Fig. 6 Transient spectrum measured 200 ns after pulse end in an
Ϫ3
ence between absorbances measured 40 µs after the pulse and absorb-
ances measured at the end of pulse, ᭡: absorbance at 385 nm corrected
for the absorbance of the 4-chloro-2-cyanophenoxyl radical. Inset: time
course of absorbances at 385 and 405 nm.
oxygen-saturated solution of 1 (2 × 10 M) at pH = 4.0; P = 0.65 mJ
Ϫ1
pulse . Inset 1: dependence of the absorbance measured at 470 nm
with P. Inset 2: time course of absorbance measured at 470 nm.
Measurements of the dependence of the absorbances at 385
and 405 nm on pulse energy, P (Fig. 5) confirmed the hypoth-
esis of two different species absorbing at these wavelengths. At
3
85 nm, a linear relationship was found as expected for a
monophotonic formation process. On the other hand, the
absorbance at 405 nm was shown to increase in a nonlinear way
with P, indicating a biphotonic contribution as expected from
Ϫ
the e formation data (Fig. 2).
aq
Formation and characterization of the carbene 2. The assign-
ment of the residual species present after subtraction of the
4
-chloro-2-cyanophenoxyl radical spectrum is aided by the
observation that its two-band structure is reminiscent either of
a further phenoxyl radical, or of a carbene resulting from a
heterolyticdehalogenationstep;indeed, thecarbene, 4-oxocyclo-
hexa-2,5-dienylidene previously observed as the primary transi-
Fig. 7 Transient spectra measured in a deoxygenated solution of 1
Ϫ1
containing propan-2-ol (0.52 M); P = 0.73 mJ pulse . ᭡: absorbances
3
ent following photolysis of 4-chlorophenol spectrally differs
measured 500 ns after pulse end, ᭺: absorbances measured 500 ns after
pulse end and corrected for the absorption of the 4-chloro-2-cyano-
phenoxyl radical, ᭿: absorbances measured 80 µs after the pulse.
from the unsubstituted phenoxyl radical merely by a 16-nm
hypsochromic shift. This similarity between the radical and the
related carbene reflects the fact that the latter has a triplet
19
ground state giving it a radical-like character.
with the spectra of these species (Fig. 4) confirms this assign-
ment. The formation of the 2-cyanophenoxyl radical, which is
absent in solutions not containing an H-donor, is explained by
rapid reaction of 2 with propan-2-ol. It is interesting to note
that the absorption at 425 nm grows in with time to reach a
maximum at about 80 µs after the laser pulse; at this time, only
4-chloro-2-cyanophenoxyl radicals are present in the spectrum,
indicating conversion of 2-cyanophenoxyl radicals into the
former by reaction with 1.
The only further phenoxyl-type radical that could conceiv-
ably arise from photolysis of 1 is the 2-cyanophenoxyl radical.
This species was produced using one-electron oxidation of
2
-cyanophenol by sulfate radicals and has a two-band spectrum
(
λmax/nm 402, 387) shown in Fig. 4. The residual transient in
Fig. 3 therefore cannot be assigned to a phenoxyl radical;
instead, we propose its identification with the carbene 2 result-
ing from heterolytic dechlorination of 1.
Characteristic trapping reactions were used to firmly estab-
lish the intermediacy of the carbene 2 in the photolysis of 1.
Triplet 4-oxocyclohexa-2,5-dienylidene is known to react with
A further transient species appears in slightly acidic
deoxygenated solution of 1 when the time course of the absorb-
ance is monitored at 405 nm. This latter transient grows in
within 40 µs following the pulse (see the inset of Fig. 5) and is
produced by a monophotonic process as shown by the linear
dependence on pulse energy (Fig. 5). This species has a lifetime
in the millisecond range and probably represents a product
formed as a consequence of addition of the carbene 2 on 1,
although it is probably not the immediate product of this addi-
tion. A transient of this type has also been observed in the
4,5
oxygen yielding benzo-1,4-quinone-O-oxide.
This latter
species exhibits a broad absorption band with a maximum at
Ϫ1
4
1
60 nm and decays in a first order reaction with k /s = 5.5 ×
d
4 3
0 in neutral medium. As shown in Fig. 6, photolysis of
oxygen-saturated solutions of 1 produces a monophotonic
transient exhibiting an absorption spectrum (λmax/nm 470) and
Ϫ1
4
kinetic characteristics (k /s = 8.4 × 10 ) very similar to those
d
3
of benzo-1,4-quinone-O-oxide. This transient can therefore be
assigned to the cyanobenzo-1,4-quinone-O-oxide.
photolysis of 4-chlorophenol.
Another reaction characterizing a triplet carbene is its reduc-
tion by H-atom donors into a phenoxyl radical. The irradi-
ation of deoxygenated solutions of 1 containing propan-2-ol
0.52 M) yields the transient spectra given in Fig. 7. The spec-
Transients in ethanolic solution and triplet quenching studies.
Flash photolysis of 1 in deoxygenated ethanol leads to the
appearance of a short-lived transient with a broad absorption
spectrum shown in Fig. 8 (Spectrum A). This transient cannot
3,5
(
trum obtained in this system exhibits three maxima which can
be deconvoluted into the superposition of a 2-cyanophenoxyl
radical and a 4-chloro-2-cyanophenoxyl radical; comparison
be observed in O -saturated solution and reacts with acryl-
amide with a rate constant of 2 × 10 M s ; it is therefore
assigned to the triplet state of 1. In the absence of quenchers,
2
8
Ϫ1 Ϫ1
1
206
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210

View Article Online
Ϫ3
Fig. 8 Transient spectra measured in a solution of 1 (4.75 × 10 M)
Ϫ1
in ethanol (0.4 mJ pulse ). ᭿: absorbance measured 100 ns after pulse
end in deoxygenated solution, divided by factor 5, ᭹: absorbance
measured 5 µs after pulse end in deoxygenated solution, ᭡: absorbance
Fig. 9 Pulse energy dependence of transients measured in a deoxygen-
Ϫ3
ated solution of 1 (4.75 × 10 M) in ethanol. (A) absorbance measured
00 ns after pulse end at 402 nm (O) and 415 nm (X), (B) absorbance
1
measured 100 ns after pulse end in O -saturated solution.
2
measured 5 µs after pulse end at 402 nm, (C) absorbance measured 5 µs
after pulse end at 415 nm. Insert: Stern–Volmer plots for the quenching
of 2-cyanophenoxyl radical formation by acrylamide (A) in aqueous
solution (pH 4) containing 0.26 M propan-2-ol, (B) in aqueous solution
(pH 4) containing 0.52 M propan-2-ol, (C) in ethanolic solution.
6
Ϫ1
the triplet signal decays with a rate constant of 1.4 × 10 s ,
independently of the substrate concentration.
A triplet energy transfer experiment was conducted as a test
Ϫ3
of this interpretation. In an ethanolic solution of 1 (1.5 × 10
Ϫ4
M) containing 2.75 × 10 M anthracene, triplet energy transfer
lengths. The short-lived absorption measured at pulse end,
however, does not show this difference between the two measur-
ing wavelengths. It can be concluded from this observation that
the absorption of the 2-cyanophenoxyl radical is not present at
short times after the pulse, in agreement with the hypothesis
that it is formed concomitantly with the decay of the triplet
state of 1.
The influence of the triplet quencher acrylamide on the
formation of the 2-cyanophenoxyl radicals was examined in
order to subject this hypothesis to a further test. As seen in the
insert in Fig. 9, the radical yield diminishes with increasing
from 1 to anthracene was observed, yielding an intersystem
crossing yield of 0.60 for 1, assuming that the transfer is
quantitative.
After decay of the triplet signal, a long-lived absorption
remains which in deoxygenated solution is similar to that
obtained from photolysis of 1 in aqueous solution in the pres-
ence of an alcohol (Fig. 7). Again, a sharp maximum at 402 nm
witnesses the formation of the 2-cyanophenoxyl radical result-
ing from dehalogenation of 1. A smaller band at 425 nm can be
assigned to a contribution from the 4-chloro-2-cyanophenoxyl
radical. It can therefore be assumed that the photochemistry of
acrylamide concentration in accordance with a Stern–Volmer
Ϫ1
1
proceeds in the same way in aqueous and in ethanolic solu-
relationship, yielding k τ/M 80 for the product of the quench-
q
tions; the same observation has been made before in the study
of the photochemistry of 4-chlorophenol.
ing rate constant and the lifetime of the quenched state in the
3
Ϫ1 Ϫ1
8
absence of the quencher. Since k /M
s
1.9 × 10 could be
q
The possibility of observing the triplet state of 1 in ethanolic
solution opened the way to studying its role in the formation of
the photoinduced transients. As shown in Fig. 8, the transient
determined by measurement of triplet decay, τ/ns 420 results
from the Stern–Volmer treatment, in good agreement with the
lifetime of the triplet state in deoxygenated solution (490 ns) as
determined from the triplet decay. This result directly confirms
that the cyanophenoxyl radicals originate from the triplet state
of 1.
spectrum obtained in O -saturated solution consists only of the
2
4
2
-chloro-2-cyanophenoxyl radical; the contribution from the
-cyanophenoxyl radical is absent.
The lack of formation of the 2-cyanophenoxyl radical in O2-
saturated solution could be due to quantitative trapping of the
The same experiment, i.e. acrylamide quenching of 2-cyano-
phenoxyl radical formation, was also conducted in deoxygen-
ated aqueous solutions containing two concentrations of
propan-2-ol (0.26 and 0.52 M). As in the ethanolic solution, the
data show Stern–Volmer behaviour, but the slopes depend on
carbene 2 by O , thus preventing H transfer from ethanol to the
2
carbene. This does not however seem to be the case because
there is no sign of the absorption of the cyanobenzo-1,4-quin-
one-O-oxide which should result from this reaction, and which
is indeed easily observed in solutions of 4-chlorophenol in
the concentration of propan-2-ol (see the insert in Fig. 9); the
Ϫ1
values obtained are k τ/M 128 at 0.26 M propan-2-ol and
q
3
Ϫ1
alkanols. It must therefore be concluded that quenching of the
kqτ/M 70 at 0.52 M propan-2-ol. This indicates that in the
triplet state also quenches the formation of the photoproducts.
The relationship between triplet and photoproduct absorp-
tion was further investigated by studying the pulse energy
dependence of laser-induced absorptions in deoxygenated solu-
tions at two wavelengths corresponding to the maximum of the
case of the aqueous solution the acrylamide interacts with a
precursor of the 2-cyanophenoxyl radical that also reacts with
the alcohol; this is most likely the carbene. Triplet quenching by
acrylamide is of minor importance, probably due to the fact
that the triplet lifetime is considerably shorter in water than in
ethanol; this is supported by the fact that it was not possible to
detect a triplet–triplet absorption in aqueous solution, and by
the observation that photolysis is only partially suppressed by
2
-cyanophenoxyl absorption (402 nm) on the one hand, and to
the nearest wavelength to the red where this radical does not
significantly absorb any more (415 nm). It can be seen from
Fig. 8 that both the triplet–triplet and the 4-chloro-2-cyano-
phenoxyl absorption do not significantly change in this wave-
length interval. The results of this experiment are shown in Fig.
Ϫ1
O saturation (see below). Assuming k τ/M < 20 for triplet
2
q
Ϫ1 Ϫ1
8
quenching by acrylamide and k /M
s
1.9 × 10 as in eth-
q
anolic solution, a triplet lifetime τ/ns < 100 can be estimated
for the aqueous medium. In ethanolic solution, interaction of
acrylamide with the carbene is unlikely because the lifetime of
the latter can be estimated to be about 5 ns if the rate constant
for its reaction with ethanol is assumed to be similar to that
9
. The long-lived radical absorption is consistently higher at 402
nm than at 415 nm, demonstrating the contribution of the
-cyanophenoxyl radical. As seen before in aqueous solution
Fig. 5), a marked departure from linearity reflects the two-
photon character of the photoionization process at both wave-
2
(
Ϫ1 Ϫ1
6
3
measured in the 4-chlorophenol system (k/M
s
6.4 × 10 ).
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210
1207

View Article Online
Ϫ4
Steady-state irradiations
when the concentration of 1 is increased from 2.5 × 10 to
1
Ϫ3
.0 × 10 M. Acidification of the solutions favours the form-
Photolysis in deoxygenated medium. The typical HPLC
chromatogram of a deoxygenated and slightly acidic solution
of 1 irradiated at 313 nm at a conversion extent of 20% is shown
in Fig. 10. The main photoproducts are cyanohydroquinone, 3
and the biphenyls 4, 5 and 6. The terphenyl 7 is additionally
ation of 3 and suppresses that of 8 indicating that the oxidation
product 8 is produced by thermal oxidation of anionic 3 during
the sampling. The main coupling product is 5, accounting for
3
2 to 38% of converted 1.
Ϫ3
formed when the concentration of 1 exceeds 2 × 10 M. Small
Photolysis in the presence of oxygen. In oxygen-saturated and
amounts (between 2 and 4%) of cyanobenzo-1,4-quinone 8 are
found even in deoxygenated solutions in slightly acidic, neutral
or basic solution.
Quantum yields of the photodegradation of 1 and chemical
yields of 3, 5 and of chloride ions in deoxygenated medium are
given in Table 1. The quantum yield of the disappearance of 1,
acidic solutions, 8 is the main photoproduct, accounting for 42
to 60% of converted 1. This result is in line with the detection
of cyanobenzo-1,4-quinone-O-oxide in laser-flash photolysis
experiments. Quantum yields of the photodegradation of 1 and
chemical yields of 3, 5, 8 and of chloride ions in O -saturated
2
solution are given in Table 2.
Φ , is 0.08 ± 0.01 in slightly acidic medium and about half this
1
value in basic medium. Chloride ions are produced with a
smaller quantum yield (0.056 in acidic medium). This difference
can be explained by the formation of coupling products
between photoproduced transients such as 2 on the one hand
and 1 on the other. Of the yields of these coupling prod-
ucts, only that of 5 could be determined; the product balance
is therefore not quantitative. The chemical yield of 3 is
concentration-dependent: it shows a decrease from 16 to 10%
Photolysis in the presence of propan-2-ol. In the presence of
propan-2-ol (0.13 M) and in the absence of oxygen, 1 is trans-
formed into 2-hydroxybenzonitrile 9 with 95% yield. In solu-
tions containing both oxygen and propan-2-ol, 8 and 9 are
formed competitively (see Table 2): as the concentration of
propan-2-ol is increased from 0.13 to 1.3 M, the chemical yield
of 9 increases from 36 to 59% whereas that of 8 decreases from
3
6 to 8%.
In the presence of propan-2-ol and in the absence of oxygen,
Φ is equal to 0.063 ± 0.005 whereas it is only 0.035 ± 0.004 in
1
oxygenated solution. This decrease in photodegradation yield
in the presence of oxygen may be the result of a triplet quench-
ing and therefore the indication that a triplet-state reaction is
involved in the photoprocess.
Photolysis in ethanol. The irradiation of 1 in deoxygenated
ethanol yields exclusively 9 with a quantum equal to
0
.075 ± 0.008. In oxygen-saturated solution, the photolysis of
1
is negligible. The reaction is inhibited by one-half by acryl-
Ϫ3
amide (6.3 × 10 M).
Involvement of the triplet state of 1 in aqueous solution.
Experiments involving the heavy-atom intersystem crossing
enhancers bromide and iodide ions which do not absorb at 313
nm, were undertaken in order to confirm the involvement of
the triplet. The results are summarized in Table 3. Addition
of bromide or iodide ions significantly enhances the photolysis
Fig. 10 HPLC chromatogram of a deoxygenated solution of 1
Ϫ4
(
5 × 10 M) irradiated for 15 min at 313 nm. Detection set at 230 nm.
Table 1 Photolysis of 1 in deoxygenated medium. Quantum yield of
1
photolysis (Φ ), quantum yield of chloride ion formation (Φ
chemical yields of photoproducts 3 and 5
Ϫ
) and
quantum yield, Φ , the effect of iodide being more pronounced
1
Cl
1
than that of bromide. These results can be interpreted in terms
of a heavy-atom effect. Saturation with oxygen in the presence of
[
1]/M
pH
Φ1
ΦCl
Ϫ
3 (%)
5 (%)
iodide (0.48 M) again reduces Φ . In the presence of bromide or
1
iodide ions, the reaction is very clean as far as products are
concerned: bromo-2-hydroxybenzonitrile 10 was identified as
the sole photoproduct in solutions containing bromide ions.
Ϫ4
Ϫ4
Ϫ3
Ϫ4
Ϫ4
a
2
5
1
2
2
.5 × 10
.0 × 10
.0 × 10
.5 × 10
.5 × 10
S.a.a
S.a.a
S.a.
3.2
0.080 ± 0.008
0.073 ± 0.007
0.080 ± 0.008
0.080 ± 0.008
0.042 ± 0.004
0.058 ± 0.004
0.056 ± 0.004
16
15
10
24
12
38
36
32
16
46
0.057 ± 0.004
Ϫ3
b
Upon addition of acrylamide (1.5 × 10 M) and in the
N.t.b
absence of oxygen, the quantum yield of 1 disappearance is not
affected, but the formation of 3 and 5 is drastically inhibited.
These results confirm that acrylamide does not interact with the
10.7
N.t.
a
b
S.a.: slightly acidic. N.t.: not titrated.
Table 2 Photolysis of 1 in oxygen-saturated medium. Quantum yield of 1 photolysis (Φ ), quantum yield of chloride ion formation (ΦCl
Ϫ
) and
1
chemical yields of photoproducts 3, 5, 8 and 9
[
1]/M
Conditions
Φ1
ΦCl
Ϫ
3 (%)
5 (%)
8 (%)
9 (%)
Ϫ4
Ϫ4
Ϫ3
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
Ϫ4
a
2
5
1
2
2
2
2
2
2
.5 × 10
.0 × 10
.0 × 10
.5 × 10
.5 × 10
.5 × 10
.5 × 10
.5 × 10
.5 × 10
S.a.a
0.038 ± 0.004
0.037 ± 0.004
0.035 ± 0.004
0.031 ± 0.003
0.02 ± 0.005
0.030 ± 0.005
0.041 ± 0.005
0.055 ± 0.005
0.053 ± 0.005
0.038 ± 0.004
0.037 ± 0.004
0.031 ± 0.004
13
11
9
10
Traces
—
—
—
—
—
—
—
—
—
—
—
—
—
53
54
60
42
30
36
19
14
8
—
—
—
—
—
36
54
44
59
S.a.a
S.a.
b
pH = 3.2
N.t.b
pH = 10.7
N.t.b
N.t.b
N.t.b
N.t.
a
S.a., propan-2-ol 0.13 M
a
S.a., propan-2-ol 0.39 M
a
S.a., propan-2-ol 0.65 M
a
b
S.a., propan-2-ol 1.3 M
N.t.
a
b
S.a.: slightly acidic. N.t.: not titrated.
1
208
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210

View Article Online
Table 3 Photolysis of deoxygenated and slightly acidic solutions of
points to a more complex reaction mechanism in comparison to
Ϫ4
1
(5.0 × 10 M) in the presence of acrylamide, bromide and iodide
that of 4-chlorophenol. It is conceivable that CN substitution
destabilizes at least some of the possible products of addition
of the carbene on 1, thereby promoting their dissociation into
radicals. The presence of the non-halogen-containing cyano-
hydroxybiphenyl 6 may be taken as an indication of such a
process.
ions. Quantum yield of 1 photolysis (Φ ) and chemical yields of photo-
1
products 3, 5 and 10
3
(%)
5
(%)
10
(%)
Additive
[Additive]/M
Φ1
Secondary photolytic transformations of the coupling
products are observed in moderately concentrated solution of
1. The presence of the terphenyl 7 among the products could
be analytically demonstrated; its formation can be explained by
the secondary photolysis of 5 (Scheme 2).
No additive
Acrylamide
Bromide ions
0.073 ± 0.007
0.055 ± 0.005
0.062
0.051
0.064
0.11
0.19
0.060
0.076
15
2
36
5
—
—
95
95
95
95
95
—
—
—
—
Ϫ3
1.5 × 10
0.01
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
0
0
0
2
.1
.3
.82
.0
Iodide ions
0.01
NC
Cl
NC
.
0
0
.048
.48
.
hν
0.20
0.085
HO
HO
Iodide ions,
oxygen-
saturated solution
0.48
HO
CN
O
CN
5
1
triplet but with the carbene to yield new and non-identified
Cl
photoproducts.
CN
OH
NC
Discussion
HO
Transient absorption as well as product studies have demon-
strated that the carbene, 3-cyano-4-oxocyclohexa-2,5-dienyl-
idene 2, is the key intermediate in the photolysis of 1. This
hitherto unknown carbene could be characterized by its UV
HO
CN
7
Scheme 2 Mechanism of 7 formation.
absorption and by its characteristic reactions, addition of O to
2
From the point of view of the overall quantum yield, the CN-
substituted derivative is much less photoreactive than 4-chloro-
phenol. By taking Φ1 values measured in the presence of
propan-2-ol or bromide ions (10 M), i.e. in conditions where
carbene 2 is fully trapped, we can estimate the quantum yield of
give a cyanobenzoquinone-O-oxide and H-abstraction from
propan-2-ol to give a 2-cyanophenoxyl radical. Its properties
are thus similar to those of unsubstituted 4-oxocyclohexa-2,5-
Ϫ2
3
dienylidene resulting from photolysis of 4-chlorophenol; in
particular, its spectrum and reactivity indicate that it has a trip-
let ground state too. Unfortunately, the substrate solubility
and low photoreaction quantum yield precluded a quantitative
study of carbene reaction kinetics. However, the photoproduct
studies are in complete agreement with the carbene hypothesis.
All observed photoproducts can be traced to carbene reactions,
as shown in Scheme 1.
2
formation as 0.062. In the case of 4-chlorophenol, the quan-
tum yield of carbene formation was 0.75, i.e. more than ten
times higher. This difference points to limiting processes on the
3
time scale of the primary photoinitiated steps, since the sub-
strate can hardly be regenerated once dehalogenation has taken
place.
Transient absorption spectroscopy has demonstrated that
photoionization of 1, both monophotonic and biphotonic, con-
tributes to photolysis on the nanosecond–microsecond time
OH
OH
CN
CN
3
scale, in contrast to 4-chlorophenol, where it could be detected
in alkaline medium only. This is the consequence of the high
excited-state acidity of 1 brought about by CN substitution.
However, the products determined in conditions of continuous
irradiation bear no trace of this pathway. This observation may
be explained by lack of reactivity of the radicals produced in
the photoionization step, in line with the generally low reactiv-
Br
–
OH
3
Br
H2O
CN
1
0
OH
O.
O
CN
CN
ROH
ROH
.
.
2
NC
NC
OH
9
20
ity of phenoxyl radicals on the one hand and with the low
O2
1
HO
dehalogenation tendency of m-chlorobenzonitrile radical
17
O
O
anions on the other; if so, geminate radicals would have a
high tendency towards in-cage recombination. Photoionization
thus conceivably contributes little to the phototransformation
of 1, and the overall quantum yield is correspondingly reduced.
Other (and quantitatively speaking more relevant) processes
limiting the phototransformation yield may be of a photo-
physical nature. In this connection, it is important to consider
the fact that the carbene 2 (as well as the carbene resulting from
Cl
4
CN
CN
NC
Cl
O
O
HO
+
O
–
8
HO
5
CN
Scheme 1 Reactivity of carbene 2.
4
-chlorophenol dehalogenation) has a triplet ground state. An
Comparison of the product distribution obtained in the pres-
ent study with that from 4-chlorophenol yields some interest-
intersystem crossing step must therefore come into play at some
stage of the photoprocess. With the 4-halogenophenols, it was
until now not possible to decide whether this step occurs at
the substrate molecule level or at the carbene level. In this
respect, CN substitution has turned out to be a valuable tool by
enabling triplet quenching as well as heavy atom effects on the
photoproduct yields.
3
ing differences. The carbene 2 has a distinctly lower tendency to
Ϫ4
give the photohydrolytic product: at 2.5 × 10 M, 4-chloro-
3
phenol yielded about 45% of hydroquinone, whereas 1 pro-
duces only 16% of 3. Substitution of the carbene by the CN
group thus favors the coupling reactions. On the other hand,
competition between the yields of 3 and 5 is not clear-cut. This
The study of the photoinduced reactions of 1 in ethanolic
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210
1209

View Article Online
solution, both by transient spectroscopy and product analysis,
gave clear evidence in support of the assumption that inter-
system crossing on the molecular level precedes dehalogen-
ation. The study of transients and photoproducts confirmed
that the photoreaction mechanism is the same in aqueous and
in ethanolic solution. The triplet–triplet absorption of 1 was
easily measurable in this latter solvent. Triplet quenching by O2
dramatically reduced the formation of the transients generated
by carbene trapping on the one hand, and the degradation of
The substitution of 4-chlorophenol by the CN group in the
ortho position to the hydroxy group has the double consequence
of lowering the triplet yield of the neutral molecule, due to fast
deprotonation of the excited singlet in competition with inter-
system crossing, and of increasing the triplet lifetime. Both
effects could be exploited to show that the photochemistry of 1
involves the triplet state of the molecule, and thus provide evi-
dence in support of the assumption that the intersystem cross-
ing step on the way to the triplet ground state of the carbene
takes place on the molecular and not on the carbene level. Add-
itionally, excited-state deprotonation entails formation of
hydrated electrons from the singlet excited state of the anion,
but this process does not affect the photochemistry in a detect-
able way.
1
following continuous irradiation on the other. A detailed
examination of the laser-induced transients in deoxygenated
ethanolic solution showed that H atom abstraction from the
alcohol by the carbene is not a very fast process and may be
rate-determined by triplet decay. The study of the influence of
acrylamide on triplet decay and 2-cyanophenoxyl formation led
to the same conclusion. It has to be noted that the triplet quan-
tum yield measured by energy transfer to anthracene is much
higher than the quantum yield of phototransformation; the
latter contributes only about 12% to overall triplet decay.
Acknowledgements
The authors wish to acknowledge the Ministère des Affaires
Etrangères, France, for financial support under program
o
The second evidence for involvement of the molecular triplet
came from the experiments using bromide or iodide as heavy-
atom enhancers of intersystem crossing in aqueous solution.
The results clearly show that the phototransformation yield is
enhanced by addition of the halides, in a manner consistent
with the heavy-atom effect. This means that in aqueous solution
the triplet yield must be relatively low in the first place; this
again constitutes a difference with respect to the 4-chloro-
phenol case, where bromide addition changed the product
distribution (in the same way as with 1, i.e., towards a clean
Amadeus n 96007.
References
1
H. D. Burrows, L. S. Ernestova, T. J. Kemp, Y. I. Skurlatov, A. P.
Purmal and A. N. Yermakov, Prog. React. Kinet., 1998, 23, 145.
C. Richard and G. Grabner, in The Handbook of Environmental
Chemistry, Environmental Photochemistry, ed. P. Boule, Springer-
Verlag, Berlin, 1999, vol. 2, Part L, p. 182.
2
3 G. Grabner, C. Richard and G. Köhler, J. Am. Chem. Soc., 1994,
116, 11470.
3
4
5
W. Sander, J. Org. Chem., 1988, 53, 2091.
substitution of Cl by Br) but not the photoreaction yield. The
B. R. Arnold, J. C. Scaiano, G. F. Bucher and W. Sander, J. Org.
Chem., 1992, 57, 6469.
reason for this difference probably lies in the fact that in the
case of 1 intersystem crossing has to compete with excited-state
deprotonation, thereby lowering the triplet yield. In the anionic
form of the phenol, the triplet yield is inherently lower; this is
shown by the lower photoreaction yields obtained in alkaline
6
A. Ouardaoui, C. A. Steren and H. van Willigen, J. Am. Chem. Soc.,
1
995, 117, 6803; A. Ouardaoui, D. M. Martino, C. A. Steren and
H. van Willigen, Appl. Magn. Reson., 1997, 13, 275.
7 A.-P. Y. Durand, R. G. Brown, D. Worrall and F. Wilkinson,
3
J. Photochem. Photobiol. A, 1996, 96, 35.
A.-P. Y. Durand, R. G. Brown, D. Worrall and F. Wilkinson,
J. Chem. Soc., Perkin Trans. 2, 1998, 365.
S. G. Schulman, W. R. Vincent and W. J. M. Underberg, J. Phys.
Chem., 1981, 85, 4068.
solution from 1 as well as from 4-chlorophenol.
8
9
In contrast to the ethanolic solution, a time-resolved study of
the triplet state could not be conducted in aqueous solution
on account of its shorter lifetime (τ/ns < 100) and lower yield.
Indeed, triplet quenching by O in water is only partial, thus
10 M. A. Paul and F. A. Long, Chem. Rev., 1957, 57, 1.
11 G. Grabner, G. Köhler, J. Zechner and N. Getoff, J. Phys. Chem.,
2
allowing observation of cyanobenzo-1,4-quinone-O-oxide pro-
duction, which is not seen in ethanol; in the case of acrylamide
interaction with the triplet is even minor, while the majority of
product quenching probably involves an interaction with the
carbene. Nevertheless, compared to 4-chlorophenol the effect
of CN substitution is clearly that of a lengthening of the triplet
1
980, 84, 3000.
1
1
2 G. Köhler, J. Photochem., 1986, 35, 189.
3 G. Grabner, G. Köhler, G. Marconi, S. Monti and E. Venuti, J. Phys.
Chem., 1990, 94, 3609.
1
4 E. Seebeck, Chem. Acta, 1947, 30, 149.
15 G. Grabner, G. Köhler, J. Zechner and N. Getoff, Photochem.
Photobiol., 1977, 26, 449.
3
lifetime; in the former case there was no detectable influence of
1
1
1
6 F.-Y. Jou and G. R. Freeman, J. Phys. Chem., 1979, 83, 2383.
7 P. Neta and D. Behar, J. Am. Chem. Soc., 1981, 103, 103.
8 M. Shirom and G. Stein, J. Chem. Phys., 1971, 55, 3372.
O on the yield of phototransformation.
2
Conclusions
19 E. Wassermann and R. W. Murray, J. Am. Chem. Soc., 1964, 86,
4
203; W. Sander, W. Müller and R. Sustmann, Angew. Chem., Int.
This study shows that the carbene 3-cyano-4-oxocyclohexa-
Ed. Engl., 1988, 113, 5311.
20 D. R. Armstrong, C. Cameron, D. C. Nonhebel and P. G. Perkins,
J. Chem. Soc., Perkin Trans. 2, 1983, 563.
2
,5-dienylidene is the key intermediate in the photolysis of 1.
This transient is converted into substrate addition products
substituted biphenyls), cyanobenzo-1,4-quinone-O-oxide or 2-
(
hydroxybenzonitrile by reaction with 1, oxygen or propan-2-ol,
respectively.
Paper 9/00141G
1
210
J. Chem. Soc., Perkin Trans. 2, 1999, 1203–1210