Laser flash photolysis study of the reaction mechanism in the photochromism of 1-(acyloxy)-2-methoxyanthraquinones

Article abstract of DOI:10.1021/jp962658y

Nanosecond laser flash photolysis has been used to study the primary processes in the photochromic reaction of five O-acylic derivatives of 1-hydroxy-2-methoxyanthraquinone with methyl (I), tolyl (II), phenyl (III), ethoxyl (IV), and diethylamino (V) groups in the migrating acyl. The triplet-triplet absorption spectra of the reactive triplet states of quinones IV and V were detected, and the rate constants of the primary photochemical step were measured. The temperature dependence of the rate constants of acyl migration in the triplet states of quinones IV and V was studied, and the Arrhenius parameters were determined. The rate constants of thermal acyl migration and their Arrhenius parameters were measured for compounds I-V. It was found that migrant nature significantly influences the activation energy of both thermal and photochemical reactions of acyl migration. The activation energy of thermal migration increases from 37.9 ± 0.2 kJ/mol for compound I up to 66.5 ± 0.7 kJ/mol for compound V. The photochemical process is characterized by considerably lower values of the activation energy (15.9 ± 0.8 and 26.0 ± 1.7 kJ/mol for compounds IV and V, respectively). It was confirmed in the present work that photochemical migration of acyl groups is an adiabatic process occurring on the triplet potential energy surface.

Full text of DOI:10.1021/jp962658y

794
J. Phys. Chem. A 1997, 101, 794-801
Laser Flash Photolysis Study of the Reaction Mechanism in the Photochromism of
1-(Acyloxy)-2-methoxyanthraquinones
N. P. Gritsan*
Institute of Chemical Kinetics and Combustion, Siberian Branch of the Russian Academy of Sciences and
NoVosibirsk State UniVersity, 630090 NoVosibirsk, Russia
A. Kellmann and F. Tfibel
Laboratoire de Photophysique Moleculaire du Centre National de la Recherche Scientifique, Batiment 213,
UniVersite Paris-Sud, 91405 Orsay, France
L. S. Klimenko
NoVosibirsk Institute of Organic Chemistry, Siberian Branch of the Russian Academy of Sciences,
630090 NoVosibirsk, Russia
ReceiVed: August 30, 1996; In Final Form: NoVember 1, 1996X
Nanosecond laser flash photolysis has been used to study the primary processes in the photochromic reaction
of five O-acylic derivatives of 1-hydroxy-2-methoxyanthraquinone with methyl (I), tolyl (II), phenyl (III),
ethoxyl (IV), and diethylamino (V) groups in the migrating acyl. The triplet-triplet absorption spectra of
the reactive triplet states of quinones IV and V were detected, and the rate constants of the primary
photochemical step were measured. The temperature dependence of the rate constants of acyl migration in
the triplet states of quinones IV and V was studied, and the Arrhenius parameters were determined. The rate
constants of thermal acyl migration and their Arrhenius parameters were measured for compounds I-V. It
was found that migrant nature significantly influences the activation energy of both thermal and photochemical
reactions of acyl migration. The activation energy of thermal migration increases from 37.9 ( 0.2 kJ/mol
for compound I up to 66.5 ( 0.7 kJ/mol for compound V. The photochemical process is characterized by
considerably lower values of the activation energy (15.9 ( 0.8 and 26.0 ( 1.7 kJ/mol for compounds IV and
V, respectively). It was confirmed in the present work that photochemical migration of acyl groups is an
adiabatic process occurring on the triplet potential energy surface.
O
C
O
C
1. Introduction
O
O
O
CH3
R1
CH3
O
O
O
Three types of photochromic processes are currently available
to quinone derivatives. They are characterized by light-induced
isomerization of the p-quinoid to the anaquinoid structure
by photochemical migration of aryl, hydrogen, or acyl
groups1-7
R1
hν
∆
R2
R2
R
1 = OCH3, NHCOCH3, N(CH3)2, R2 = H; R1 = H, R2 = OCH3
The photochemical products have a structure of substituted
9-acetoxy-1,10-anthraquinones (ana-quinones, a-Q).
O
XY
YO
X
hν1
To account for the relationship between the reactivity of these
compounds and their chemical structure, it was assumed that
photochemical migration of the acetyl group occurs adiabati-
cally.5 Later,6 the triplet state of the product (ana-quinone) was
detected as a precursor of the ana-quinone ground state in the
case of 1-acetoxy-2-methoxyanthraquinone (I). According to
this result, a triplet adiabatic mechanism was proposed for the
photochemical reaction of acyl group migration, but a detailed
study of the primary processes required a better time
resolution.6
hν2 or ∆
O
O
1, X = O, Y = Ph; 2, X = CR1R2, Y = H; 3, X = O, Y = C(O)R
The thermally reversible photochemical migration of acyl
groups was discovered previously for a series of 1-acetoxyan-
thraquinones with donor substituents in the anthraquinone
nucleus5-7
The influence of temperature and solvent and migrant nature
on the rate constant of thermal migration of the acyl group was
previously studied.6,7 It was found that donor substituents in
the migrating acyl group significantly reduce the rate constant
* Author to whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, January 1, 1997.
S1089-5639(96)02658-8 CCC: $14.00 © 1997 American Chemical Society

1-(Acyloxy)-2-methoxyanthraquinone Photochromism
J. Phys. Chem. A, Vol. 101, No. 5, 1997 795
of the thermal migration. It is reasonable to expect, therefore,
that donor substituents also reduce the rate constant of the
photochemical migration. Therefore, the purpose of the present
study was to investigate the primary photochemical step,
migration of the acyl group, by nanosecond laser photolysis and
to study the influence of the acyl group nature on the rate
constant of this process. For this purpose, we used a series of
O-acyl derivatives of 1-hydroxy-2-methoxyanthraquinone
(I-V).
SCHEME 1
2.4. Quantum Yields. The quantum yields of 2-methoxy-
9-(acyloxy)-1,10-anthraquinone (a-Qs) formation were deter-
mined by comparing the a-Q concentrations obtained after laser
excitation with that obtained by excitation at the same laser
energy of a solution of acridine in benzene chosen as the
standard. The triplet quantum yield of acridine is reported to
be 0.73 ( 0.07.8 In these experiments, less then 10% of the
anthraquinones and acridine molecules were converted respec-
tively into a-Qs and acridine triplet. The concentration of
acridine triplet was monitored at its absorption maximum (442
nm) using a triplet extinction coefficient8 ꢀ ) 2.7 × 104 M-1
cm-1. The a-Q concentration was monitored at the absorption
maximum at 510 nm. The value of ꢀ for the a-Q of compound
I was estimated at 77 K from the data of low-temperature
photolysis.6 This value was found to be 1.1 × 104 M-1 cm-1
at the absorption maximum 526 nm. We used this value of ꢀ
for the a-Qs of all compounds under study (I-V) because
substitution in the acyl group has only a minor effect on the
spectra. We also estimated ꢀ of the a-Q of compound IV from
the analysis of the intensity dependence of a-Q absorption.
The quantum yield of compound V irreversible decomposition
was determined from the decrease in optical density at the long-
wavelength absorption maximum. The photoisomerization of
2-(dimethylamino)-3-chloro-1,4-naphthoquinone in benzene (æ
) 0.10 ( 0.01)9 was used as an actinometric reaction.
O
O
O
OCR
RCO
O
OCH3
OCH3
hν
∆
O
O
I, R = CH3; II, R = CH2Ph; III, R = Ph; IV, R = OC2H5; V, R = N(C2H5)2
It was found, indeed, that donor substituents in the acyl
significantly reduce the rate constant of the photochemical
migration of the acyl group. The triplet-triplet absorption
spectra of quinones IV and V were detected, and the rate
constants of the primary photochemical step were measured.
The triplet- and ground-state absorption spectra of the photo-
product (ana-quinone) were observed; the quantum yields of
ana-quinone formation and the rate constants of elementary
reactions were measured in toluene at room temperature for
compounds I-V. It was confirmed in the present work that
photochemical migration of acyl groups is an adiabatic process
occurring on the triplet potential energy surface.
2. Experimental Details
2.5. Quantum Chemical Calculations. Experimental data
were interpreted by quantum chemical calculations using the
AM110 method based on the modified MNDO-85 program.11
The conventional BFGS procedure12 was used to optimize the
geometry. The geometry and electronic structure of excited
states were determined by the restricted Hartree-Fock technique
in a “half-electron” approximation.13
2.1. Materials and Solutions. O-Acyl derivatives (I-V)
of 1-hydroxy-2-methoxyanthraquinone were obtained by acy-
lation of 1-hydroxy-2-methoxyanthraquinone according to ref
6. The toluene used as the solvent was Merck Uvasol grade.
The concentrations of the quinones used were (2-5) × 10-5
M. Sample solutions were contained in 1-cm × 1-cm silica
cells and were deoxygenated by bubbling argon or oxygen
saturated by bubbling oxygen.
3. Results and Discussion
3.1. Reaction Scheme. Solutions of (2-5) × 10-5
M
2.2. Laser Flash Photolysis. The third harmonic (355 nm)
of a pulsed YAG laser (Quantel YG 441; pulse width at half-
maximum about 2 ns) was used as the excitation light source.
Relative measurements of the laser energy were performed by
focusing a small fraction of laser light on a pyroelectric
joulemeter. Transient transmission changes were monitored at
right angles to the laser beam using a xenon flash lamp (VQX
65N) as the probing light source. The detection system
consisted of a monochromator (Jarrell-Ash Type 82-410;
bandwidth 2 nm), a photomultiplier (HTV R928), and a digital
oscilloscope (Tektronix 2440). The electric signal from the
photomultiplier was fed into the oscilloscope through a field
effect transistor probe (Tektronix P6021). The time resolution
of the detection system was about 3 ns. The digital signals
were analyzed using a microcomputer.
The sample solutions of compound V were changed after
every laser pulse in order to avoid the effects due to the product.
The photochemical transformations of quinones I-IV are highly
reversible. Therefore, we could use the sample solutions of
quinones I-IV for 10-15 laser pulses without any changes in
the absorption spectra of the samples or in transient kinetics.
2.3. Absorption and Fluorescence. Conventional absorp-
tion and fluorescence measurements were performed using a
Cary 210 spectrophotometer (Varian) and MPF-3 spectrofluo-
rimeter (Perkin-Elmer), respectively.
1-(acyloxy)anthraquinones (Q) I-V in toluene were studied by
nanosecond laser photolysis in the temperature range 273-330
K in the presence and absence of oxygen. The results for
compounds I-IV are consistent with the occurrence of the
1
processes described in Scheme 1, where Q* represents the
excited singlet state, 3Q* the lowest triplet state of the quinones,
a-Q the ground state of the product with a structure of
9-(acyloxy)-1,10-anthraquinone (ana-quinone), and 3a-Q the
lowest triplet state of the product. The arguments which have
led to the proposal of this scheme are developed in the following
sections. The scheme describing the peculiarity of compound
V photolysis will be discussed in section 3.6.
3.2. Spectra and Kinetics. Transient optical density (OD)
changes obtained by laser excitation at 355 nm of deoxygenated
toluene solutions of compounds I-V at 297 K were monitored
in the spectral region 300-800 nm, over a time range extending
from a few nanoseconds to a few hundred microseconds after
the laser pulse. The OD changes were found to be linearly
dependent on the laser energy at low pulse energies but saturated
at higher energies due to ground-state depletion (Figure 1). In
our opinion, the linear relationship reflects the monophotonic
character of the processes investigated.
The transient spectra obtained after laser excitation of
compound IV (immediately after the end of the laser pulse, at

796 J. Phys. Chem. A, Vol. 101, No. 5, 1997
Gritsan et al.
Similar results were obtain for the [4/4] Pade approximation.
Using the limiting values of optical density, the extinction
coefficient of the a-Q of compound IV was estimated at the
long-wavelength maximum (ꢀ ) 1.16 × 104 M-1 cm-1 at 510
nm). This value is practically coincident with the value of ꢀ
estimated for the a-Q of compound I at 77 K6 (ꢀ ) 1.1 × 104
M-1 cm-1). This means that the [3/3] Pade approximation
(Figure 1) gives a reasonable result. The absolute spectra of
the intermediates (Figure 4a,b) were calculated from the
difference spectra (Figure 2) combined with the kinetic results
and using an estimated degree of depletion of the ground state.
It should be noted that the extinction coefficients of compound
IV triplet-triplet absorption are slightly underestimated (about
10-20%) due to its short lifetime (τ1 ) 12 ns), comparable
with the time resolution of the setup (about 3 ns).
3.3. Oxygen Effect. The concentration of molecular oxygen
in the solution has no effect on the yield and lifetime of the
long-lived intermediate a-Q. Saturation of the solution with
oxygen (9 × 10-3 M)15 efficiently accelerated the decay of its
precursor and did not reduce the amount of a-Q formed.
According to this result, the precursor of a-Q is its own excited
state. This state is a triplet since its lifetime is rather long (1.1
µs in the absence of oxygen). Therefore, the spectrum with
maxima at 390 and 500 nm can be attributed to the triplet-
triplet spectrum of a-Q.
Figure 1. Dependence of a-Q (compound IV) absorption at the long-
wavelength maximum (510 nm) on the laser intensity. Solid line: [3/3]
Pade approximation of intensity dependence.
3
The kinetics of the disappearance of a-Q is exponential as
noted before. The pseudo-first-order rate constant k2 is de-
pendent on the oxygen concentration in solution and can be
described by (Figure 5).
k2 ) k20 + k2Q[O2]
(1)
The quenching rate constant in the case of the 3a-Q of compound
IV is equal to (2.5 ( 0.1) × 109 M-1 cm-1. The similar values
3
of the quenching rate constant were obtained for the a-Qs of
compounds I, II, and V (Table 1). These values are indeed
characteristic of triplet quenching by oxygen.16,17
It is reasonable to assume that the end-of-pulse absorption
spectrum is due mainly to the triplet state (3Q) of anthraquinone
IV. A very similar spectrum was detected immediately after
the end of laser excitation of compound V (Figure 6). As in
the case of compound IV, it was attributed to the triplet-triplet
absorption spectrum of anthraquinone V. The lifetime of this
triplet state (τ1) is equal to about 1 µs at room temperature in
the absence of oxygen. Saturation of the solution with oxygen
efficiently accelerated the decay of this intermediate (Figure 7)
and reduced the yield of the following intermediates formed
(3a-Q and a-Q of compound V). The pseudo-first-order rate
constant k1 is linearly dependent on the oxygen concentration
in solution (Figure 8). The quenching rate constant k1Q is equal
to (6.6 ( 0.3) × 108 M-1 cm-1, which also lies within the range
of typical values for triplet quenching by oxygen.16,17 Since
Figure 2. Absorption spectrum of quinone IV (1) and transient
absorption spectra obtained upon excitation of compound IV (2.9 ×
10-5 M) in degassed toluene at 295 K by a laser pulse at 355 nm: 2,
spectrum detected immediately after the end of the laser pulse; 3, at
50 ns; 4, at 5 µs after the laser pulse.
50 ns and 5 µs after the pulse) are shown in Figure 2. Typical
kinetic traces at selected wavelengths are shown in Figure 3.
All kinetics are strictly first-order. Since the kinetic traces
observed can be fitted at all wavelengths by three exponentials
with the same time constants, these constants are assigned to
the lifetimes of three transient species; their values are respec-
tively 12 ns (τ1), 1.1 µs (τ2), and 150 µs (τ3).
The longer-lived transient spectrum with a maximum at 510
nm is assigned to a photoproduct with a structure of a-Q. This
absorption spectrum is identical to that of the low-temperature
(77 K) photoisomerization product.6 Note that a similar
spectrum had been detected for the a-Q of compound I in laser
flash photolysis experiments at 195 K.6 The precursor of a-Q
with a lifetime of 1.1 µs has two absorption bands in the
spectrum with maxima at 390 and 500 nm. The first short-
lived intermediate has two absorption maxima at 460 and 600
nm.
Under our experimental conditions, it was impossible to
achieve complete depletion of the ground state (Figure 1).
Nevertheless, approximation of the OD dependence by the ratio
of two polynomials (Pade approximation14) led to the limiting
value of a-Q optical density in the case of complete depletion
of the ground state. Figure 1 represents the [3/3] Pade
approximation (the ratio of two third-order polynomials).
3
the lifetime of the Q of compound IV is very short (12 ns),
saturation of the solution with oxygen has no effect on it.
3.4. Quantum Yields. The quantum yields of ana-quinone
formation were determined in toluene at 297 K by comparing
the a-Q concentrations obtained upon excitation of solutions of
compounds I and III-V with the concentration of the triplet of
acridine used as a standard (Table 2). It was found that for
compounds I, III, and IV, the quantum yields are close to 1.
The quantum yield of ana-quinone in the case of V is
significantly smaller, ∼0.1. The possible reasons for the lower
quantum yield in the case of V will be analyzed below (section
3.6).
3.5. Temperature Effect. The effect of temperature on the
kinetics was studied from 272 to 330 K. Temperature signifi-
cantly influenced the rate constant of the thermal migration of

1-(Acyloxy)-2-methoxyanthraquinone Photochromism
J. Phys. Chem. A, Vol. 101, No. 5, 1997 797
Figure 3. Kinetic traces showing the variation in optical density at different wavelengths (λ) on the excitation of anthraquinone IV (2.9 × 10-5
M) in degassed toluene at 295 K: (a) λ ) 600 nm; (b, c) λ ) 390 nm; (d) λ ) 510 nm.
TABLE 1: Values of the Lifetime (τ2) of the Triplet State of 9-(Acyloxy)-2-methoxy-1,10-anthraquinones in Toluene at Room
Temperature in the Absence of Oxygen and the Rate Constants (k2Q) of Their Quenching by Oxygen
compd
I
II
III
IV
V
R
CH3
1.13 ( 0.08
3.11 ( 0.11
CH2Ph
1.18 ( 0.12
1.66 ( 0.10
Ph
1.08 ( 0.05
OC2H5
1.15 ( 0.10
2.5 ( 0.10
N(C2H5)2
7.1 ( 0.9
0.32 ( 0.02
τ2, µs
10-9k2Q, M-1 cm-1
the acyl group (k3). Arrhenius plots were linear for all
compounds under study. Arrhenius parameters are presented
in Table 3. It is seen that the nature of the acyl group has only
a minor effect on the preactivation factor. The large effect of
the migrant nature on the rate constant of thermal migration is
due to the variation of activation energy. Substitution of a
methyl group by a diethylamino group results in an increase of
activation energy from 38 to 66.5 kJ/mol and, therefore, in a
decrease in the rate constant at room temperature of 6 orders
of magnitude.
the temperature dependence of k1 for compound IV is shown
in Figure 9. The activation parameters are listed in Table 3.
An Arrhenius treatment of the observed rate constant (kobs
)
of the decay of the triplet state of quinone V is presented in
Figure 10. The plot is clearly nonlinear. In this case, the rate
constant of acyl migration is decreased significantly and the
reactive triplet can also undergo intersystem crossing (ISC). The
temperature dependence of kobs can be described as a sum of
temperature-independent (kisc) and temperature-dependent (k1)
terms (Figure 10, insert)
The triplet-triplet absorption of initial quinone (3Q) and its
decay were detected only in the case of compounds IV and V
with donor substituents in the migrating acyl. In the case of
compounds I-III, we were unable to detect their triplets due to
the very high rate constant of acyl migration. The lifetime of
the 3Q of compounds I-III should be substantially shorter than
the time resolution of the detection system (about 3 ns). The
lifetime of the Q of compound IV is short enough (12 ns at
room temperature); nevertheless, its dependence on temperature
was studied from 272 to 315 K. An Arrhenius treatment of
kobs ) kisc + k1
(2)
kisc ) (4.3 ( 0.4) × 105 s-1, k1 ) (2.1 ( 1.1) × 1010
×
exp(-(3130 ( 190)/T) s-1
The rate constants of the decay of the product triplet excited
states (k2) are described for the a-Qs of compounds I-V by eq
1. The k20 values are very slightly dependent on temperature
for compounds I-IV. For instance, in the case of compound
III, k20 changes from (8.0 ( 0.2) × 105 to (9.2 ( 0.5) × 105 in
3

798 J. Phys. Chem. A, Vol. 101, No. 5, 1997
Gritsan et al.
Figure 6. Absorption spectrum of quinone V (1) and transient
absorption spectrum detected immediately after the end of the laser
pulse (2) in toluene solution (2.0 × 10-5 M) at 295 K.
A similar dependence
k20 ) 106.44(0.07 exp{-(280 ( 30)/T)
was measured previously6 for 3a-Q decay of compound I in the
temperature range 160-300 K. The preexponential factor is
low (about 106) and typical of a spin-forbidden process. The
activation energy of the interconversion is low (1.7-2.3 kJ/
mol). It reflects the participation in the deactivation process
of the vibrational mode with frequency 140-190 cm-1. An
analogous temperature dependence with a low activation energy
(about 500 cm-1) has been observed previously for the inter-
conversion of the excited triplet states of the hydrocarbons.18
3.6. Peculiarity of Anthraquinone V Photolysis. At room
temperature, kobs for the triplet-state decay of compound V is
equal to (9.5 ( 0.8) × 105 s-1. Using this result and the
estimated value of kisc, the quantum yield of formation of the
product triplet state (ΦT) can be calculated to be ΦT ) 0.5 (
0.1. This value is in good agreement with the estimation from
Figure 4. (a, top) Extinction coefficients (ꢀ) of anthraquinone IV (1)
and its lowest triplet state (2). (b, bottom) Extinction coefficients (ꢀ)
of the lowest triplet state of ana-quinone (3a-Q) (1) and the ground
state of ana-quinone (a-Q) (2) of compound IV.
3
the absorption of a-Q at 390 nm. If it is assumed that the
extinction coefficients of triplet-triplet absorption do not depend
on the substituent in the acyl group and ΦT can be estimated as
0.5 ( 0.1.
As described in section 3.4, the presence of oxygen signifi-
cantly influenced the lifetime of the triplet state of compound
V (Figure 8). Figure 11 (plot 1) shows the Stern-Volmer plot
for the quantum yield of the a-Q triplet state as a function of
oxygen concentration. The slope of the curve (KQ ) 680 ( 30
M-1) is in good agreement with the value of τ1k1Q ) 690 ( 70
M-1, extracted from the kinetic data (Figure 8).
It is seen from Table 2 that the quantum yields of a-Q
formation are close to unity except for compound V. In the
case of V, the quantum yield is significantly smaller (∼0.1).
But as we estimated before, the quantum yield of formation of
the product triplet state is about 0.5. A possible explanation of
3
this result is the occurrence of a side reaction in the a-Q of
compound V. In agreement with this assumption, we observed
a smaller effect of oxygen content on the yield of a-Q than on
the yield of its precursor a-Q (Figure 11). This is due to the
Figure 5. Dependence of the rate constant (k2) of ana-quinone triplet
state (3a-Q) deactivation on the oxygen concentration: a-Q of com-
pounds I (1), IV (2), and V (3).
3
quenching by oxygen of the side reaction.
We have estimated the quantum yield of irreversible decom-
position of compound V. In the absence of oxygen, it is high
and equal to 0.25 ( 0.05. Saturation of the solution with oxygen
results in a significant (68 ( 13 times) decrease of the quantum
yield of irreversible decomposition. This is consistent with the
assumption that a side reaction occurs mainly in 3a-Q. Indeed,
the temperature range 275-330 K. These changes can be
described in terms of an Arrhenius relationship in the following
form
k20 ) 106.2(0.1 exp{-(200 ( 50)/T)

1-(Acyloxy)-2-methoxyanthraquinone Photochromism
J. Phys. Chem. A, Vol. 101, No. 5, 1997 799
Figure 7. Kinetic traces showing the variation in optical density at 600 nm on excitation of anthraquinone V (2.0 × 10-5 M) at 295 K in degassed
(a) and oxygen saturated (b) toluene.
Figure 9. Arrhenius plot of the decay rate constant (k1) of the quinone
Figure 8. Dependence of the rate constant (k1) of the quinone V triplet-
IV triplet state (3Q).
state (3Q) decay on the oxygen concentration.
TABLE 2: Values of the Quantum Yields (Φ) of
9-(Acyloxy)-2-methoxy-1,10-anthraquinone Formation under
Irradiation of 1-(Acyloxy)-2-methoxy-9,10-anthraquinones in
Toluene at Room Temperature
The activation energy (15.1 ( 1.3 kJ/mol) is about 1 order of
magnitude higher than in the case of compounds I-IV (1.7-2
kJ/mol). It is reasonable to assume that this activation energy
is due to the side reaction.
compd
I
III
IV
V
One of the possible side reactions is the cleavage of the
carbon-oxygen bond and formation of two radicals, one of them
being an acyl radical. It is the well-known product of the photo-
Fries rearrangement.19 This photoprocess can be temperature
activated.19,20 But we did not isolate any product of rearrange-
ment except 1-hydroxy-2-methoxy-9,10-anthraquinone. The
chemical yield of 1-hydroxy-2-methoxy-9,10-anthraquinone
upon irradiation of compound V in dry toluene in the presence
of oxygen is equal to 88%. Further investigations are carried
out in order to understand the nature of the side reaction.
Nevertheless, according to the data available, we propose
Scheme 2 for the phototransformation of compound V.
3.7. Quantum Chemical Calculations. To interpret the
experimental results and the influence of chemical structure on
the rate constants of the elementary reactions (Scheme 1), we
performed the quantum chemical calculation of the electronic
Φ
0.85 ( 0.17 0.92 ( 0.18 0.91 ( 0.18 0.10 ( 0.02
in this case, quenching by oxygen of both triplets (3Q and
3a-Q) should lead to a very effective decrease of the quantum
yield of the side reaction (60 ( 13 times).
3
As noted before, the rate constants of a-Q decay (k2) are
very slightly dependent on temperature in the case of compounds
I-IV. It is in agreement with the fact that the main channel of
decay is intersystem crossing to the ground state. This is not
the case for compound V. An Arrhenius treatment of k2 for
compound V is shown in Figure 12. The plot is linear, and the
temperature dependence can be fitted by
k2 ) 108.3(0.3 exp(-1820 ( 160)/T)
TABLE 3: Values of the Rate Constants of Acyl Group Migration in the Excited Triplet State of p-Anthraquinones (k1) and in
the Ground State of Ana-anthraquinones (k3, Measurement Accuracy (10%) and Parameters of Their Arrhenius Plots
compd
R
k1, s-1
log A1, s-1
E1act, kJ/mol
k3, s-1
log A3, s-1
E3act, kJ/mol
I
II
III
IV
V
CH3
CH2Ph
Ph
OC2H5
N(C2H5)2
1.2 × 106
1.6 × 106
1.1 × 105
6.5 × 103
2.9
11.8 ( 0.1
37.9 ( 0.2
12.16 ( 0.13
12.33 ( 0.13
12.20 ( 0.13
40.5 ( 0.6
48.2 ( 0.8
66.5 ( 0.7
(8.2 ( 1.0) × 107
(5.2 ( 0.5) × 105
10.7 ( 0.2
10.3 ( 0.2
15.9 ( 0.8
26.0 ( 1.7

800 J. Phys. Chem. A, Vol. 101, No. 5, 1997
Gritsan et al.
Figure 10. Arrhenius plot of the decay rate constant (kobs) of the
quinone V triplet state (3Q) and fitting of kobs by exponential
dependence, kobs ) kisc + A1 exp(-E/RT) (insert).
Figure 13. Energy correlation diagram for the process of thermally
reversible photochemical acyl group migration. The formation enthalpies
of quinone IV, ana-quinone of IV, and their lowest triplet states were
calculated using the AM1 method. The energies of excited singlet states
were estimated from experimental absorption spectra of quinone and
ana-quinone. The position of the second triplet state (T2, nπ* type) of
quinone IV was estimated from the experimental data for 9,10-
anthraquinone.23 The energies of the transition states for the thermal
and photochemical reactions were calculated from experimental activa-
tion energies.
Figure 11. Stern-Volmer plots for the quantum yields of the a-Q
triplet state (1) and the a-Q ground state (2) vs oxygen concentration
under laser irradiation of anthraquinone V in toluene at 295 K.
SCHEME 2
quinones I-V and all the transients are practically independent
of the nature of the acyl group. According to this fact, calculated
positions of the triplet states of Q and a-Q are also very slightly
dependent on the substituent in the acyl group. Therefore, the
reaction enthalpy for acyl group migration in the excited triplet
state is practically independent of substituent in the migrant
(-66.6, -69.6, and 63.3 kJ/mol for compounds I, IV, and V,
respectively). Consequently, in this case, the influence of the
acyl substituent on the reaction rate constant does not correlate
with the thermodynamics of the process.
In a previous work,6 it was assumed that the influence of
substituents on the rate constant of thermal acyl migration is
due to the change of the charge on the carbon atom of the
migrating COR group: a higher rate of migration is observed
for a larger positive charge. Our calculation showed that a
significant positive charge is localized on a carbon atom of the
carbonyl group (about 0.4), and accordingly, a large negative
charge is localized on the quinone oxygen (about -0.4).
However, for compounds with donor substituent (a-Qs of IV
and V), our calculations gave an even bigger positive charge
(0.41) then for the a-Q of I (0.34). Therefore, the results of the
AM1 calculations do not confirm the assumption of our earlier
study.6
Figure 12. Arrhenius plot of the decay rate constant (k2) of ana-quinone
V triplet state (3a-Q).
structure and geometry of anthraquinones I-V, their a-Q forms,
and the lowest triplet states of Q and a-Q of I-V. Figure 13
presents the predicted relative energies of the ground and lowest
triplet states of quinone IV and its ana-quinone form. The
patterns for the other compounds under study are very similar.
The energy difference between the ground states of a-Q and Q
(reaction enthalpy for thermal acyl group migration) depends
very slightly on the nature of the substituent in the acyl group
(-62.0 and -69.6 kJ/mol for compounds I and V, respectively).
It was noted before that the experimental absorption spectra of

1-(Acyloxy)-2-methoxyanthraquinone Photochromism
J. Phys. Chem. A, Vol. 101, No. 5, 1997 801
migration of acyl groups are significantly higher. It can be due
to a bigger difference in nuclear configurations for the migration
in the ground state than in the excited triplet state.
The migrant nature significantly influences the activation
energy of both thermal and photochemical reactions of acyl
migration. The activation energy of thermal migration increases
from 37.9 ( 0.2 kJ/mol for compound I up to 66.5 ( 0.7 kJ/
mol for compound V with donor diethylamino substituent in
the migrating acyl. The activation energy of the photochemical
process changes from 15.9 ( 0.8 (compounds IV) to 26.0 (
1.7 kJ/mol for compound V. The influence of the nature of
the migratory group on the activation energy can also be due
to the magnitude of the change in the nuclear configurations of
the initial and final states of the reaction of acyl migration. The
preexponential factors are practically independent of the migrant
nature and are equal to 1010-1011 s-1 for the photochemical
process and about 1012 s-1 for the thermal migration.
The quantum yields of photochemical migration of the acyl
group in quinones I-IV are close to unity. Such values are
consistent with the adiabatic nature of the process. The very
low quantum yield (∼0.1) of ana-quinone formation and the
high quantum yield of irreversible photodecomposition in the
case of compound V are connected with the appearance of an
effective side reaction in the excited triplet state of ana-quinone.
The origin of this side reaction will be the subject of further
study.
Figure 14. Computer-generated drawing of 1-acetoxy-2-methoxy-
9,10-antraquinone (a) and 9-acetoxy-2-methoxy-1,10-antraquinone (b).
Geometry was optimized by the AM1 method.
It is seen from Figure 14 that the acetoxy group of quinone
I lies in the plane practically perpendicular to the plane of the
anthraquinone rings. The angle between these planes is even
less than 90° and equals 74° for the a-Q of I and 79° for the
a-Q of V. This angle is also smaller in the triplet states and
equals 70° and 72° for compounds I and V, respectively. It is
clear that the reaction coordinate has to be mainly a combination
of stretching and bending OC modes. According to the results
of the calculations, the difference in nuclear configuration of
the initial and final states will be smaller in the case of
compound I than compounds IV and V. Moreover, this
difference is smaller in the excited triplet states than in the
ground states. Let us assume that the potential energy depen-
dence on the reaction coordinate is harmonic with a frequency
slightly dependent on the substituent in the acyl group. In this
case, the barrier height (or activation energy) will depend both
on the enthalpy of reaction and on the difference in nuclear
configuration of the initial and final states. It is seen that in
our case (thermal and photochemical acyl migration), a higher
rate of migration is observed for the smaller difference in nuclear
configuration.
Acknowledgment. We are very grateful to Dr. L. Lindqvist
for helpful discussions and to Ministere de la Recherche et de
lÅEnseignement for financial support for one of us (Dr. Nina
Gritsan). This work was supported in part by the Russian
Foundation for Fundamental Research (Project 95-03-08920).
References and Notes
As we can see from the experimental data, the photochemical
acyl migration reaction occurs completely on the triplet potential
energy surface. We think that the absence of photoreaction in
the excited singlet state of compounds I-V is due mainly to
the short lifetime of this state. We failed to detect the
luminescence of compounds I-V, which means that the
quantum yield of fluorescence is smaller than 10-3. Using these
data and a rough estimate of the radiative lifetime (∼25 ns),
we can deduce that the lifetime of the singlet excited state is
significantly smaller than 25 ps. It is known that the rate
constant of intersystem crossing between excited states of
different symmetry (ππ* and nπ* states) can be very high
(∼1010-1011 s-1).21 This situation is realized in our case (Figure
13), and it is the reason for the very short lifetime of the singlet
excited ππ* state. It should be noticed that the growth lifetime
of the reactive triplet state of quinone IV from a precursor has
been estimated to be about 70 ps.22 This is longer than our
estimation for the lifetime of the singlet excited state and may
be equal to the lifetime of the intermediate triplet state T2 of
nπ* type.
(1) Gerasimenko, Yu. E.; Poteleschenko, N. T. Zh. Vses. Khim. O-Va.
im. D. I. MendeeleeVa, 1971, 16, 105.
(2) Malkin, J.; Zelenchenok, A.; Krongauz, V.; Dvornikov, A. S.;
Rentzepis, P. M. J. Am. Chem. Soc. 1994, 116, 1101.
(3) Gritsan, N. P.; Rogov, V. A.; Bazhin, N. M.; Russkikh, V. V.; Fokin,
E. P. Teor. Eksp. Khim. 1979, 15, 290.
(4) Gritsan, N. P.; Khmelinski, I. V.; Usov, O. M. J. Am. Chem. Soc.
1991, 113, 9615.
(5) Gritsan, N. P.; Russkikh, S. A.; Klimenko, L. S.; Plusnin, V. F.
Teor. Eksp. Khim. 1983, 19, 455.
(6) Gritsan, N. P.; Klimenko, L. S.; Shvartsberg, E. M.; Khmelinski,
I. V.; Fokin, E. P. J. Photochem. Photobiol. A: Chem. 1990, 52, 137.
(7) Gritsan, N. P.; Klimenko, L. S. J. Photochem. Photobiol. A: Chem.
1993, 70, 103.
(8) Kellmann, A. J Phys. Chem. 1977, 81, 1195.
(9) Gritsan, N. P.; Bazhin, N. M. IzV. AN SSSR, Ser. Khim. 1981, 280.
(10) Dewar, M. J. S.; Zoebich, E. G.; Healy, E. F.; Stewart, J. J. P. J.
Am. Chem. Soc. 1985, 107, 3902.
(11) Bliznyuk, A. A.; Voityuk, A. A. Zh. Struct. Khim. 1986, 27, N4,
190.
(12) Dennis, J. E.; Schnabel, R. B. Numerical Methods for Unconstrained
Optimization and Nonlinear Equations; Prentice-Hall: Englewood Cliffs,
NJ, 1983.
(13) Voityuk, A. A. Zh. Struct. Khim. 1983, 24, N3, 18.
(14) Baker, G. A. Ess. Pade Approx. N. Y. 1975.
(15) McKeown, A. B.; Hibbard, R. R. Anal. Chem. 1956, 28, 1490.
(16) Gijzeman, O. L. J.; Kaufman, F.; Porter, G. L. J. Chem. Soc.,
Faraday Trans. 2 1973, 69, 708.
(17) Ermolaev, V. L.; Bodunov, E. N.; Sveshnikova, E. B.; Shakhverdov,
T. A. Radiationless Transfer of Electron Excitation Energy; Nauka:
Lenigrad, 1977; p 190.
(18) Jones, P. F.; Siegel, S. J. Chem. Phys. 1969, 50, 1134.
(19) Bellus, D. AdV. Photochem. 1971, 8, 109.
(20) Gritsan, N. P.; Tsentalovich, Yu. P.; Yurkovskaya, A. V.; Sagdeev,
R. Z. J. Phys. Chem. 1996, 100, 4448.
(21) Plotnikov, V. G.; Ovchinnikov, A. A. Usp. Khimii (Russ. Chem.
ReV.) 1978, 47, 444.
(22) Malkin, Ya. N. Private communication.
(23) Nurmukhametov, R. N. Absorption and Luminescence of Organic
Compounds; Khimiya: Moscow, 1971.
4. Conclusions
In the O-acylic derivatives of 1-hydroxy-2-methoxyan-
thraquinone (I-V), photochemical migration of the acyl group
takes place via the triplet excited state. It was confirmed in
the present work that photochemical migration of acyl groups
is an adiabatic process occurring on the triplet potential energy
surface. The absence of photoreaction in the excited singlet
state of compounds I-V is due mainly to the short lifetime of
this state. The process of acyl migration in the triplet excited
state of quinones is termally activated, but the activation energy
is low (15.9 ( 0.8 and 26.0 ( 1.7 kJ/mol for compounds IV
and V, respectively). Activation energies for the thermal