Real time observation of the Photo-fries rearrangement

Article abstract of DOI:10.1063/1.1752885

Two-color femtosecond pump probe spectroscopy was used to investigate the photo-Fries rearrangement of 4-tert-butylphenyl acetate (BPA) dissolved in cyclohexane. Broadband measurements were performed in the spectral range between 450 and 700 nm to confirm the wavelength dependence of the various signal contributions. The recombination of the radicals was found to occur within 13 ps. The time resolved absorption measurements on BPA provide strong evidence that after photoexcitation the S₁ state is depopulated with a rate of 0.5 ps^-1 by a mechanism which leads to the cleavage of the O-C bond and the generation of a radical pair.

Full text of DOI:10.1063/1.1752885

Real time observation of the photo-Fries rearrangement
S. Lochbrunner, M. Zissler, J. Piel, E. Riedle, A. Spiegel, and T. Bach
Citation: The Journal of Chemical Physics 120, 11634 (2004); doi: 10.1063/1.1752885
View online: http://dx.doi.org/10.1063/1.1752885
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/120/24?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photoisomerization among ring-open merocyanines. I. Reaction dynamics and wave-packet oscillations induced
by tunable femtosecond pulses
J. Chem. Phys. 140, 224310 (2014); 10.1063/1.4881258
Photodissociation of thioglycolic acid studied by femtosecond time-resolved transient absorption spectroscopy
J. Chem. Phys. 134, 024514 (2011); 10.1063/1.3526746
Communications: Photoinitiated bond dissociation of bromoiodomethane in solution: Comparison of one-photon
and two-photon excitations and the formation of iso- CH 2 Br – I and iso- CH 2 I – Br
J. Chem. Phys. 132, 141102 (2010); 10.1063/1.3374680
Experimental and theoretical study of the photodissociation reaction of thiophenol at 243 nm : Intramolecular
orbital alignment of the phenylthiyl radical
J. Chem. Phys. 126, 034306 (2007); 10.1063/1.2424939
Photoisomerization of azobenzene from first-principles constrained density-functional calculations
J. Chem. Phys. 122, 094311 (2005); 10.1063/1.1861873
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.207.120.173 On: Tue, 30 Sep 2014 11:09:50

JOURNAL OF CHEMICAL PHYSICS
VOLUME 120, NUMBER 24
22 JUNE 2004
Real time observation of the photo-Fries rearrangement
a)
S. Lochbrunner, M. Zissler, J. Piel, and E. Riedle
Lehrstuhl f u¨ r BioMolekulare Optik, Sektion Physik, Ludwig-Maximilians-Universit a¨ t, Oettingenstrasse 67,
8
0538 M u¨ nchen, Germany
A. Spiegel and T. Bach
Lehrstuhl f u¨ r Organische Chemie I, Technische Universit a¨ t M u¨ nchen, Lichtenbergstrasse 4,
85747 Garching, Germany
͑
Received 18 November 2003; accepted 29 March 2004͒
The photo-Fries rearrangement of 4-tert-butylphenyl acetate dissolved in cyclohexane is
investigated by two-color femtosecond pump probe spectroscopy. The spectral transmission changes
are characterized in the visible and ultraviolet spectral region and allow for the first time to
temporally resolve the primary reaction steps. We find that the photoinduced homolytic cleavage of
the CO bond occurs within 2 ps and that the geminate recombination of the generated radical pair
to the intermediate substituted cyclohexadienone takes 13 ps. The experimental results support a
model in which the initial reaction proceeds from the originally excited ␲␲* state via a barrier to
a dissociative ␲␴* state. © 2004 American Institute of Physics. ͓DOI: 10.1063/1.1752885͔
I. INTRODUCTION
clohexadienone after irradiation of phenyl acetate have been
observed by time resolved flash photolysis and occur on a
The photochemistry of aryl esters and aryl amides is
time scale of some microseconds and of milliseconds,
1
,2
10,12,13
governed by photo-Fries rearrangements. Since the first
observation of this reaction in the case of catechol
respectively.
Up to now no direct observation of the
primary events has been reported. Time resolved Raman
studies gave only an upper limit of a few nanoseconds for the
3
monoacetate it has attracted much scientific interest. It has
4
11
been used for the synthesis of hydroxyphenones and mac-
appearance of the cyclohexadienone intermediate. Accord-
5
rocyclic diazonamides and it plays a significant role in the
ingly, the knowledge about the first steps after photoexcita-
tion is vague and the exact nature of the involved electronic
states is under debate.
6
degradation of polycarbonates, polyesters, and polyamides.
Its investigation gives insight into the mechanism of photo-
reactions involving radicals.7
Here we report our femtosecond pump probe studies on
4-tert-butylphenyl acetate ͑BPA͒ dissolved in cyclohexane.
They represent the first time resolved observation of the ap-
pearance of a radical pair and its geminate recombination in
the case of a photo-Fries rearrangement. Since the photo-
chemistry of BPA dissolved in cyclohexane had not yet been
characterized we carried out additional continuous wave
͑cw͒ experiments to determine the quantum yield and the
spectrum of the final ortho-hydroxyacetphenone photoprod-
uct 4-tert-butyl-2-hydroxyacetophenone ͑BOHA͒. BPA is
used because the para position is occupied by the 4-tert-
butyl moiety and the acyl group can only be transferred to
the ortho position. This simplifies the reaction scheme and
the data interpretation. The reaction product BOHA has also
been of synthetic interest for the preparation of potential ul-
traviolet ͑UV͒ absorbers.19 In the discussion we show that
the results support a model in which the bond cleavage is due
to the interplay of three electronic states.
Insight into the mechanism has been gained by system-
atic investigations of the photochemistry of phenyl and naph-
8
–15
thyl esters.
In Fig. 1 the photo-Fries rearrangements of
phenyl acetate and of 4-tert-butylphenyl acetate are shown.
After photoexcitation the acyl group is transferred to the
ortho or para position but not to the meta position and the
corresponding hydroxyacetophenones are formed.3 As a
side product phenol is found in many cases.
,4
The appearance of phenol and the observation of chemi-
cally induced spin polarization led to the hypotheses that
after photoexcitation a homolytic cleavage of the O–C bond
occurs and an acetyl and an aryloxy radical are
4
,9,10,16
generated.
Their recombination results in a dienone
which is transformed into the final hydroxyphenone by eno-
lization. The enolization has been shown to be mainly due to
intermolecular interaction under the conditions used in this
work with a small contribution from a sigmatropic hydrogen
1
3
shift. The selective recombination at the ortho or para po-
sition is explained by the density distribution of the unpaired
II. EXPERIMENT
electron that seems to be negligible at the meta position of
the aryloxy radical.1
7,18
If the aryloxy radical escapes from
BPA was prepared in 98% yield from 4-tert-butylphenol
͑9.45 g, 62.9 mmol͒ by treatment of a solution in acetic acid
anhydride ͑18.1 g, 16.8 mL, 151 mmol͒ with 1.5 mL of py-
ridine at 0 °C. Upon work-up the known compound20 was
the solvent cage phenol can be formed. The transformation
of the phenoxy radical to phenol and the enolization of cy-
a͒Author to whom correspondence should be addressed; electronic mail:
stefan.lochbrunner@physik.uni-muenchen.de
obtained as a white solid. Analytical data: R ϭ0.61 ͑pentane/
f
1
tert-butylmethyletherϭ80/20͒; H–NMR ͑200 MHz, CDCl )
3
0
021-9606/2004/120(24)/11634/6/$22.00
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.207.120.173 On: Tue, 30 Sep 2014 11:09:50
11634
© 2004 American Institute of Physics
1

J. Chem. Phys., Vol. 120, No. 24, 22 June 2004
The photo-Fries rearrangement
11635
which delivers 150 fs long laser pulses at 775 nm with a
repetition rate of 1 kHz. To generate the pump pulses a frac-
tion of 90 ␮J from the CPA output was frequency doubled in
a 500-␮m-thick BBO crystal cut for type I phase matching.
The second harmonic light and the copropagating remainder
of the fundamental were sum frequency mixed in a subse-
quent 500-␮m-thick BBO crystal cut for type II phase
matching yielding pulses at 258 nm with an energy of 0.75
␮J at the sample. Probe pulses at various wavelengths be-
tween 260 and 900 nm were generated with a noncollinearly
phase-matched optical parametric amplifier ͑NOPA͒ pumped
with a portion of the CPA output in the order of 200 ␮J per
pulse.2
1,22
For probe wavelengths below 450 nm the pulses
from the NOPA were frequency doubled in a 100-␮m-thick
BBO crystal.23 The energy of the probe pulses was reduced
to below 50 nJ by reflecting the beam from a fused silica
wedge. Pump and probe pulses were focused at the sample to
a spot size of about 200 and 120 ␮m in diameter, respec-
tively. The transmitted probe light and a reference reflection
of the probe beam were measured with UV sensitive photo-
diodes. The crosscorrelation had a width of typically 300 fs
full width at half maximum and was measured at the sample
position by difference frequency generation in a 100-␮m-
thick BBO crystal. For measurements with higher time reso-
lution pump pulses at 260 nm were generated by frequency
doubling pulses from a second NOPA operated at 520 nm.
By compressing the UV pump pulses and the probe pulses
with fused silica prism sequences a crosscorrelation width of
FIG. 1. ͑a͒ Photo-Fries rearrangement of phenyl acetate leading to the prod-
ucts ortho- and para-hydroxyacetophenone and phenol. ͑b͒ The investigated
photo-Fries rearrangement of BPA via the intermediate cyclohexadienone to
the final BOHA.
␦
͑
1.30 ͑s, 9 H͒, 2.17 ͑s, 3 H͒, 7.00 ͑d, Jϭ8.8 Hz, 2 H͒, 7.37
d, Jϭ8.8 Hz, 2 H͒. 13C NMR ͑50 MHz͒ ␦ 21.0 ͑q͒, 31.3 ͑q͒,
34.3 ͑s͒, 120.8 ͑d͒, 126.2 ͑d͒, 148.2 ͑s͒, 148.4 ͑s͒, 169.5 ͑s͒.
Cyclohexane ͑Uvasol, Merck͒ of spectroscopic purity was
used as the solvent.
The absorption spectrum of the final photoproduct
BOHA was measured after irradiating a 1.9 mM cyclohexane
solution of BPA in a 1-cm-thick fused silica cell for 90 h in
a fluorimeter ͑Fluorolog, SPEX͒ with spectrally filtered light
from a xenon lamp at 264 nm and a bandwidth of 5 nm.
During the irradiation the sample absorption was measured
in regular intervals with a steady state spectrometer ͑Lambda
1
00 fs was achieved. In addition to the two color experi-
ments the transient spectra were measured in the visible with
a white light continuum. The continuum was generated by
focusing a small portion of the CPA output in the order of 1
19, Perkin–Elmer͒. Towards the end of the irradiation time
␮J into a 3-mm-thick sapphire disc. After the sample the
no changes in the absorption were observed indicating that
all BPA molecules had undergone the photoreaction.
white light passed a monochromator and was detected with a
photodiode. The spectra were recorded by tuning the mono-
chromator. In this case BPA was excited at 270 nm with the
frequency doubled output of a NOPA.
To determine the quantum yield of the photo-Fries rear-
rangement under the conditions of the time resolved experi-
ment ͑see later͒ a BPA solution in a 1-mm-thick fused silica
cell was irradiated with femtosecond pump pulses at 258 nm
at a repetition rate of 1 kHz. The laser beam was vertically
enlarged by a cylindrical lens to irradiate most of the sample
simultaneously. The average excitation power in the sample
was 2.2 mW and the pulse energy 2.2 ␮J. Every minute the
cell was shaken to mix irradiated with nonirradiated fractions
and cw absorption spectra were measured. The concentration
in this experiment was only 0.5 mM because the photoprod-
uct absorbs stronger than BPA and at higher concentrations
the increasing product fraction would lead to a significant
reduction of the light intensity during the course of the ex-
periment. The energy of the pump pulses was measured be-
fore and after the sample cell. The arithmetic averaging in-
troduces an error of less than 1% to the measurement of the
energy per pulse within the sample since the BPA concentra-
tion was small and the absorption of the solution very low.
More seriously, the calibration error of the power meter lim-
its the accuracy to 5%.
The sample was pumped through a home made cell with
2
00-␮m-thick fused silica windows and an optical path
length of 120 ␮m. The excited volume was exchanged after
each laser pulse. The BPA solution had a concentration of
0
.14 M resulting in 70% absorption at the pump wavelength.
Additional measurements with pure cyclohexane were car-
ried out to check for solvent contributions to the pump probe
signals.
The results presented later were obtained with parallel
polarizations of pump and probe pulses. Experiments with
perpendicular polarizations yielded identical results within
the experimental accuracy. This can be understood by con-
sidering that the chromophore of BPA is the benzene moiety.
1
1
The benzene B2u S � A S0 transition in turn is vibroni-
1
1g
cally induced by coupling of the S state to the degenerate
1
1
24
E1u S3 state with a resulting lack of excitation anisotropy.
III. RESULTS AND DISCUSSION
A. Stationary spectroscopy
The time resolved absorption changes between 260 and
900 nm after photoexcitation of BPA were measured with a
two color pump probe spectrometer pumped by a regenera-
tive Ti:sapphire amplifier system ͑CPA 2001, Clark MXR͒
First we consider the photochemical properties of BPA
characterized by stationary spectroscopy. The absorption
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.207.120.173 On: Tue, 30 Sep 2014 11:09:50

11636
J. Chem. Phys., Vol. 120, No. 24, 22 June 2004
Lochbrunner et al.
FIG. 2. ͑a͒ Absorption spectra of BPA ͑solid line͒ and the photoproduct
BOHA ͑broken line͒. For better comparison the first band of the BPA spec-
trum is multiplied by 10 and shown as thick line. ͑b͒ Rise of the BOHA
absorption at 340 nm ͑circles͒ due to irradiation of BPA in dependence on
the excitation fluence. The solid line is a fitted exponential rise.
FIG. 3. Transmission change after excitation of BPA at 258 nm probed at
260, 330, and 700 nm ͑open circles͒. The solid lines are double exponential
functions fitted to the data.
solid trace is an exponential increase fitted to the data. From
the exponential rise and the extinction coefficient of BPA a
quantum yield for the photo-Fries reaction to BOHA of
spectrum of BPA and the final photoproduct BOHA is de-
picted in Fig. 2͑a͒. The first absorption band of BPA has the
characteristic vibronic structure of the benzene chromophore.
In comparison to benzene dissolved in cyclohexane the band
is redshifted by 9 nm and the structure is somewhat less
pronounced. The maximum extinction coefficient is 420
l/͑mol cm͒ and about four times larger than the one of ben-
zene due to the symmetry breaking substituents. Irradiating a
BPA sample for sufficient long time with continous light at
5
1%Ϯ8% was calculated. The given error reflects both the
uncertainty of the energy measurement and the slowly rising
absorption of the 258 nm light by the photoproduct. Since no
significant concentration of additional photoproducts were
detected the rest of the excited BPA molecules either under-
goes an internal conversion process back to the ground state
or recombines to the reactant after dissociation.
We have also attempted to measure the fluorescence
spectrum of BPA but no significant fluorescence signal was
found. Since the electronic transition dipole moment as cal-
culated from the absorption spectrum is quite small this ob-
264 nm or with femtosecond excitation pulses at 258 nm
leads to the same changes in the absorption and to the same
final spectrum. It is attributed to the photo-Fries product
BOHA and exhibits its first band at about 340 nm with a
maximum extinction coefficient of 2400 l/͑mol cm͒ ͓see Fig.
servation gives an upper limit for the S lifetime of 1 ns. Our
1
2
͑a͔͒. As expected, the band position and strength are very
time resolved experiments reveal that the lifetime is indeed
only 2 ps ͑see later͒.
2
5,26
similar to hydroxyacetophenone.
Both compounds have
an efficient internal conversion pathway from the electroni-
cally excited state to the ground state and do not react under
the continuing light exposure. Phenol has its first absorption
band at 271 nm with an extinction coefficient of 1450
B. Femtosecond spectroscopy
We characterized the dynamics of the photo-Fries rear-
rangement with femtosecond pump-probe spectroscopy. Af-
2
7
l/͑mol cm͒. In none of our measurements we have found
any indication for the appearance of this band. We conclude,
that the generation of substituted phenol is at most a very
minor channel in the photochemistry of BPA dissolved in
cyclohexane. This is in excellent agreement with the high
reaction yield of 91% reported for the synthetic application
of the investigated reaction.19 Therefore, we do not include
the phenol channel in the analysis of our time resolved data.
At the beginning of the sample irradiation the BOHA
spectrum rises linearly with the applied energy and ap-
proaches later asymptotically the final strength. This is dem-
onstrated in Fig. 2͑b͒ which shows the strength of the band
maximum at 340 nm in dependence on the fluence ͑total
energy per unit area͒ of the applied excitation pulses. The
ter exciting BPA in its S state with UV pulses at 258 nm the
1
induced transmission changes were measured between 260
and 900 nm. Three typical time resolved traces at the probe
wavelengths 260, 330, and 700 nm are depicted in Fig. 3.
The sharp feature at time zero is due to nonresonant signal
contributions from the solvent and the solute. At positive
delay times a reduced transmission is observed over the en-
tire probed spectral range. The transmission changes are larg-
est in the UV and significantly smaller in the visible and
near-infrared. During the first picoseconds a further transmis-
sion decrease occurs at wavelengths below 300 nm whereas
at longer wavelengths the transmission increases on the same
timescale. Subsequently the transmission increases at all
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.207.120.173 On: Tue, 30 Sep 2014 11:09:50

J. Chem. Phys., Vol. 120, No. 24, 22 June 2004
The photo-Fries rearrangement
11637
FIG. 5. Transmission changes in the visible at a delay of 1, 5, 20, and 50 ps
after excitation of BPA at 270 nm measured with a white light continuum.
amplitudes of the two exponential decays and the remaining
absorption in dependence on the probe wavelength. The am-
plitude of the short time constant is positive below 300 nm
and negative above resembling the absorption increase dur-
ing the first picoseconds for wavelengths below 300 nm. The
amplitude of the 13 ps decay is fairly strong in the UV and
its wing extends in the visible almost to 800 nm. The residual
absorption is only significant for wavelengths below 600 nm.
To check for contributions to the dynamics at shorter
time scales we performed pump probe experiments with a
crosscorrelation width of 100 fs at the probe wavelengths
260 and 500 nm ͑not shown͒. However, the identical dynam-
ics as with the longer crosscorrelation is observed and no
indication for an additional contribution is found.
FIG. 4. The amplitudes of the 2 ps ͑a͒ and the 13 ps ͑b͒ decay and the
residual absorption at long delay times ͑c͒. A negative amplitude reflects an
absorption decaying with time and a positive one an absorption increase or
the disappearance of bleaching or stimulated emission.
probe wavelengths corresponding to an exponential absorp-
tion decay with a time constant of 13 ps ͑see later͒. Above
600 nm the signal disappears completely whereas below 450
nm a significant absorption remains.
Measurements of the pure solvent show that after the
temporal overlap of pump and probe pulses the signal from
the solvent is negligible for wavelengths below 450 nm. At
longer wavelengths the fast decaying component and the re-
sidual absorption measured in the BPA solution are contami-
nated by solvent contributions. They are most probably due
to solvated electrons generated by two photon ionization.2
Since the absorption of the BPA solution at the pump wave-
length differs from the pure solvent the excitation intensities
in the samples are not comparable. Therefore, it is not pos-
sible to correct for the solvent signal by subtracting measure-
ments of pure cyclohexane. The absorption decay with the
longer time constant of 13 ps is completely absent in the pure
solvent and the corresponding signal contributions are in the
whole wavelength range solely due to the photodynamics of
BPA.
We fitted a model function to the experimental traces
that consists of two exponentially decaying components, an
absorption term remaining at long delay times and a Gauss-
ian dip for the artifact at time zero ͑see Fig. 3͒. The first
exponential time constant adopts values between 2 and 3 ps
at probe wavelengths below 450 nm and between 1 and 1.5
ps for longer wavelengths. The reduction of the decay con-
stant at longer wavelengths is probably due to solvent and
solvation contributions. The second time constant is found to
be 13 psϮ1 ps at all probe wavelengths. Figure 4 shows the
We performed in addition broadband measurements in
the spectral range between 450 and 700 nm to confirm the
wavelength dependence of the various signal contributions.
Transient spectra at various delay times are shown in Fig. 5.
Immediately after the excitation transient absorption is ob-
served in the whole visible range with increasing strength
towards shorter wavelengths. The transmission recovers after
the excitation with a time constant of about 13 ps and addi-
tional small contributions on the picosecond timescale like in
the single wavelength measurements discussed earlier. The
residual absorption in the visible is almost negligible and is
attributed to solvent contributions. The evolution of the tran-
sient spectra in Fig. 5 reflects the fitted decay amplitudes in
the visible range where the 13 ps component is the dominant
contribution ͑see Fig. 4͒.
The residual absorption at long delay times does not
show the pronounced absorption band of the final photoprod-
uct at 340 nm ͑compare Figs. 4 and 2͒. The final product is
obviously not yet formed on the subnanosecond time scale
accessible to our experiments. We tentatively attribute the
spectrum at long delay times to the intermediate cyclohexa-
dienone. The first absorption band of this compound is ex-
pected at wavelengths around 300 nm or below.13 The as-
signment agrees with the observation that the final hydrogen
8,29
shift occurs in the case of phenyl acetate on a time scale of
some tenth of a second.1
0,13
The absorption decaying with a
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.207.120.173 On: Tue, 30 Sep 2014 11:09:50

11638
J. Chem. Phys., Vol. 120, No. 24, 22 June 2004
Lochbrunner et al.
pare the process with the photochemistry of phenol. Phenol
can dissociate its O–H bond to form a phenoxy radical when
it is photoexcited. According to ab initio calculations this
process is mediated by a higher electronically excited state
with ␲␴* character.3
5,36
When the hydrogen is shifted along
its detachment coordinate this state decreases strongly in en-
ergy and forms an avoided crossing with the S state. The
1
␲
␴* state correlates directly with the electronic configuration
of the separated hydrogen atom and the phenoxy radical. If
phenol is excited to its S state the population is transferred
1
via the avoided crossing to the ␲␴* state and reaches along
the ␲␴* potential energy surface the ground state of the pho-
toproducts. The S state of phenol increases in energy along
the hydrogen detachment coordinate and the ab initio calcu-
FIG. 6. Model for the first steps of the photo-Fries rearrangement. The
excited state potential energy surface results from the interaction between
the optically excited ␲␲* state and a ␲␴* state. Within 2 ps the S1 state is
depopulated via the ␲␴* state which correlates with the ground state of the
two radicals. At the conical intersection with the electronic ground state of
the reactant the relaxing population splits. One part returns to the original
ground state and the other parts proceeds on the ␲␴* state and forms a
radical pair. The subsequent recombination of these radicals to the interme-
diate product occurs within 13 ps.
0
lations indicate that a conical intersection exists between the
35,36
S0 state and the ␲␴* state.
At this intersection a portion
of the excited molecules changes to the S state and returns
0
to the ground state of the reactant. It is the dynamics at the
intersection which determines the quantum yield of the
photoreaction.37
The hydrogen atom has an unpaired electron and repre-
sents a radical. We propose that the cleavage of the O–C
bond in BPA can be compared with the detachment of the
hydrogen in phenol and the acetyl radical plays a similar role
as the hydrogen atom. Ab initio calculations on phenyl ac-
etate indeed indicate that a correlation exists between the
electronic ground state of the radical pair and the electroni-
cally excited ␲␴* state of the reactant.38 Not all electronic
states of the dissociating BPA correlate with the phenol sys-
tem because the phenoxy acetyl biradical is tetratopic and the
phenoxy H-atom biradical tritopic. However, the correlations
breaks down only for singlet states at high energies where
they do not influence the reaction dynamics. The electronic
ground states of phenol and phenyl acetate correlate with the
lowest electronically excited singlet states in the biradicals
time constant of 13 ps is of similar strength in the UV as the
residual absorption but has a broad wing extending into the
visible.
To the best of our knowledge, absorption spectra of the
tert-butylphenoxy radical are not available in the literature.
We found spectral information, but no spectra for related
compounds. Cook reports a blue color for the 2,4,6-tri-tert-
butylphenoxy radical30 which would indicate a red absorp-
tion band around 630 nm.31 For the phenoxy radical itself
and several derivates a band around 400 nm is reported32
indicating that the butyl substitution at the para position is
critical to observe an absorption band in the visible. The first
absorption band of the acetyl radical is found around 220
3
3,34
nm
and cannot be used for the identification of this spe-
which are in both cases n␲* excitations of the phenoxy
cies in our measurements.
radical.3
6,38
From the visible transient absorption that decays
Due to this situation we attribute the broad and feature-
less absorption in the near UV and blue part of the visible to
the 4-tert-butylphenoxy radical. Consequently the absorption
decay should reflect the disappearance of the radicals. It in-
dicates that the recombination of the radicals occurs within
with 13 ps we conclude that this state is roughly 2 eV above
the biradical ground state in the case of BPA. We suppose
that the ␲␴* state of BPA decreases in energy if the O–C
bond is elongated and leads directly to the separated radicals
͑
see Fig. 6͒. It forms an avoided crossing with the ␲␲* state,
13 ps and leads to the cyclohexadienone generation ͑see Fig.
the S state in the Franck–Condon region, and subsequently
6
͒. The time for the recombination of the radicals agrees with
1
1
6
a conical intersection with the electronic ground state of the
reactant. The optically excited ␲␲* state correlates with a
very energetic doubly excited state of the biradical38 and
does not influence the dynamics beyond the crossing with the
the estimates deduced from CINDP studies.
The short time constant of 2 ps reflects the formation of
radicals due to the cleavage of the O–C bond ͑see Fig. 6͒.
The bond rupture seems to be associated with the return from
the electronically excited state to the ground state. The broad
initial absorption extending all over the visible indicates an
␲
␴* state.
The avoided crossing of the electronically excited states
results in an energy barrier between the S1 minimum and the
excited state absorption from the S state to higher electronic
1
Ϫ1
product region of the ␲␴ state. The finite rate of 0.5 ps
*
states. The transmission decrease associated with the 2 ps
time at wavelengths below 300 nm points to emission from
indicates that the process is not instantaneous and needs
some time to overcome the barrier. The excitation in our
experiments is on the blue side of the first absorption band
and leads to a fairly large amount of vibrational excess en-
ergy in the optically active modes. Possibly the redistribution
of the energy from the optically excited to the reactive vibra-
tional modes contributes to the 2 ps time. During this time
emission in the UV and excited state absorption contribute to
the S state of the excited BPA and its decay is due to the S1
1
depopulation.
C. Reaction mechanism
To understand the radical generation, i.e., the primary
step of the photo-Fries rearrangement it is instructive to com-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.207.120.173 On: Tue, 30 Sep 2014 11:09:50

J. Chem. Phys., Vol. 120, No. 24, 22 June 2004
The photo-Fries rearrangement
11639
the signal. Our attempts to see a change of the dissociation
rate with varying excitation wavelength did not yet provide
any significant evidence.
implementing the continuum setup. They gratefully acknowl-
edge funding by the Deutsche Forschungsgemeinschaft.
After the barrier crossing the molecular system evolves
on the potential energy surface of the ␲␴* state to the coni-
cal intersection ͑see Fig. 6͒. There about half of the popula-
1
tion is transferred to the original S state and returns to the
D. Bellu sˇ , Adv. Photochem. 8, 109 ͑1971͒.
0
2
M. A. Miranda and F. Galindo, in CRC Handbook of Organic Photochem-
istry and Photobiologiy, 2nd ed., edited by W. Horspool and F. Lenci
reactant minimum. The rest of the population reaches the
potential well of the separated radical pair. The cyclohexadi-
one intermediate is formed subsequently by the recombina-
tion of the radicals in ortho geometry. CIDNP studies on
͑
CRC Press, Boca Raton, 2004͒, p. 41–1.
3
4
5
6
J. C. Anderson and C. B. Reese, Proc. Chem. Soc. 1960, 217 ͑1960͒.
H. Kobsa, J. Org. Chem. 27, 2293 ͑1962͒.
P. Mangus and C. Lescop, Tetrahedron Lett. 42, 7193 ͑2001͒.
Lignins, edited by K. V. Sarkanen and C. H. Ludwig ͑Wiley, New York,
1
-naphtyl acetate indicate that recombination of the radicals
1
4
to the original compound is of minor importance and
thereby support the idea that the quantum yield is indeed
given by the dynamics at the conical intersection. The effi-
cient product formation and short recombination time of 13
ps indicate that the radicals cannot escape from the solvent
cage and propably stay in contact almost all the time.
1971͒.
7
8
9
0
D. Bellu sˇ and P. Hrdlovi cˇ , Chem. Rev. 67, 599 ͑1967͒.
M. R. Sandner and D. J. Trecker, J. Am. Chem. Soc. 89, 5725 ͑1967͒.
J. W. Meyer and G. S. Hammond, J. Am. Chem. Soc. 92, 2187 ͑1970͒.
C. E. Kalmus and D. M. Hercules, J. Am. Chem. Soc. 96, 449 ͑1974͒.
S. M. Beck and L. E. Brus, J. Am. Chem. Soc. 104, 1805 ͑1982͒.
R. Nakagaki, M. Hiramatsu, T. Watanabe, Y. Tanimoto, and S. Nagakura,
J. Phys. Chem. 89, 3222 ͑1985͒.
T. Arai, S. Tobita, and H. Shizuka, Chem. Phys. Lett. 223, 521 ͑1994͒.
N. P. Gritsan, Y. P. Tsentalovich, A. V. Yurkovskaya, and R. Z. Sagdeev, J.
Phys. Chem. 100, 4448 ͑1996͒.
1
11
12
IV. CONCLUSIONS
13
14
Our time resolved absorption measurements on BPA pro-
vide strong evidence that after photoexcitation the S state is
15
1
I. F. Molokov, Y. P. Tsentalovich, A. V. Yurkovskaya, and R. Z. Sagdeev,
J. Photochem. Photobiol., A 110, 159 ͑1997͒.
W. Adam, J. A. de Sanabia, and H. Fischer, J. Org. Chem. 38, 2571
Ϫ1
depopulated with a rate of 0.5 ps by a mechanism which
1
1
6
7
leads to the cleavage of the O–C bond and the generation of
a radical pair. The recombination of the radicals is found to
occur within 13 ps. It leads to the intermediate cyclohexadi-
enone product which transforms to the final product BOHA
on a time scale longer by far. This is the first direct observa-
tion of the primary steps of the photo-Fries rearrangement,
i.e., the formation and recombination of the radical pair. The
bond cleavage leading to the radicals seems to occur via a
͑
1973͒.
T. J. Stone and W. A. Waters, J. Chem. Soc. 1964, 213 ͑1964͒.
C. Cui, X. Wang, and R. G. Weiss, J. Org. Chem. 61, 1962 ͑1996͒.
A. Spiegel, T. Bach, T. Habeck, and F. Prechtl, J. Inf. Recording 25, 355
18
19
͑
2000͒.
B. S. Ross, M. Finkelstein, and R. C. Petersen, J. Am. Chem. Soc. 84,
139 ͑1964͒.
20
4
21
T. Wilhelm, J. Piel, and E. Riedle, Opt. Lett. 22, 1494 ͑1997͒.
22
E. Riedle, M. Beutter, S. Lochbrunner, J. Piel, S. Schenkl, S. Sporlein, and
W. Zinth, Appl. Phys. B: Lasers Opt. 71, 457 ͑2000͒.
¨
␲
␴* state which forms an avoided crossing with the optically
23
A. J. Wurzer, S. Lochbrunner, and E. Riedle, Appl. Phys. B: Lasers Opt.
excited ␲␲* state along a coordinate associated with the
lengthening of the O–C bond.
71, 405 ͑2000͒.
24
For a review, see J. M. Hollas, High Resolution Spectroscopy ͑Wiley,
Chichester, 1998͒, p. 385.
In molecules exhibiting excited state intramolecular pro-
ton transfer we found indications for a similar mechanism for
the subsequent internal conversion.3 A deformation of the
reaction center leads to an avoided crossing between the op-
tically accessible ␲␲* state and a ␲␴* state which forms
then a conical intersection with the electronic ground state.
The interplay of the three electronic states causes a very
efficient electronic relaxation route to the ground state. We
think that analogous to this situation and the photo-Fries re-
action the transition between two different electronic states is
often governed by the coupling to a third state like for ex-
ample a ␲␴* state which mediates the process. In this case
the energy gap between the originally excited state and the
ground state is of secondary importance and does not directly
determine the relaxation rate. Rather it is determined by the
shape of the potential energy surfaces, the resulting avoided
crossing and finally the accessibility of the conical
intersection.
25
J. Catal a´ n, F. Toribio, and A. U. Acu
˜
na, J. Phys. Chem. 86, 303 ͑1982͒.
S. Nagaoka, N. Hirota, M. Sumitani, and K. Yoshihara, J. Am. Chem. Soc.
05, 4220 ͑1983͒.
9,40
26
1
2
2
2
7
8
9
H. J. Jaffe and M. Orchin, Theory and Applications of Ultraviolett Spec-
troscopy ͑Wiley, New York, 1962͒, Chap. 12.
A. Migus, Y. Gauduel, J. L. Martin, and A. Antonetti, Phys. Rev. Lett. 58,
1559 ͑1987͒.
J. A. Kloepfer, V. H. Vilchiz, V. A. Lenchenkov, A. C. Germaine, and S. E.
Bradforth, J. Chem. Phys. 113, 6288 ͑2000͒.
C. D. Cook, J. Org. Chem. 18, 261 ͑1953͒.
C. D. Cook and B. E. Norcross, J. Am. Chem. Soc. 81, 1176 ͑1959͒.
R. Hermann, S. Naumov, G. R. Mahalaxmi, and O. Brede, Chem. Phys.
Lett. 324, 265 ͑2000͒.
30
31
32
33
H. Adachi, N. Basco, and D. G. L. James, Chem. Phys. Lett. 59, 502
͑
1978͒.
34
M. M. Maricq and J. J. Szente, Chem. Phys. Lett. 253, 333 ͑1996͒.
A. L. Sobolewski and W. Domcke, J. Phys. Chem. A 105, 9275 ͑2001͒.
A. L. Sobolewski, W. Domcke, C. Dedonder-Lardeux, and C. Jouvet,
Phys. Chem. Chem. Phys. 4, 1093 ͑2002͒.
W. Fuß, K.-L. Kompa, S. Lochbrunner, A. M. M u¨ ller, T. Schikarski, W. E.
Schmid, and S. A. Trushin, Chem. Phys. 232, 161 ͑1998͒.
S. Grimme, Chem. Phys. 163, 313 ͑1992͒.
3
3
5
6
37
3
3
8
9
ACKNOWLEDGMENTS
K. Stock, T. Bizjak, and S. Lochbrunner, Chem. Phys. Lett. 354, 409
͑
2002͒.
The authors thank Tanja Bizjak for performing the con-
tinuum probe measurements and Vincent De Waele for
40
S. Lochbrunner and E. Riedle, Recent Res. Devel. Chem. Physics 4, 31
͑2003͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.207.120.173 On: Tue, 30 Sep 2014 11:09:50
1