Effect of cyclodextrin complexation on the photochemistry of the lignin model α-guaiacoxyacetoveratrone

Article abstract of DOI:10.1139/v99-137

The photodecomposition of α-guaiacoxyacetoveratrone (GAV) in homogeneous solvents and in aqueous cyclodextrin solutions was studied by following the fluorescence emission of the photoproducts formed. The same qualitative behavior as previously observed for the photodegradation quantum yield measurements was reproduced in the fluorescence studies in dry or wet methanol and acetonitrile. The yield for GAV photodegradation in water is smaller than in the organic solvents. Complexation of GAV to β- and γ-cyclodextrins leads to an increase in the photodecomposition yield, especially during the early part of the photodecomposition kinetics. The transients formed in the photolysis of GAV in water are similar to those observed in acetonitrile, and the lifetime of triplet GAV in water is 130 ns. In addition, solvated electrons were formed when GAV was photolyzed in water. Complexation of GAV to cyclodextrins leads to an increase in the triplet yield compared to the radical yield and a mode rate shortening of the triplet lifetime.

Full text of DOI:10.1139/v99-137

1356
Effect of cyclodextrin complexation on
the photochemistry of the lignin model
␣-guaiacoxyacetoveratrone
L.T. Okano, R. Ovans, V. Zunic, J.N. Moorthy, and C. Bohne
Abstract: The photodecomposition of α-guaiacoxyacetoveratrone (GAV) in homogeneous solvents and in aqueous
cyclodextrin solutions was studied by following the fluorescence emission of the photoproducts formed. The same
qualitative behavior as previously observed for the photodegradation quantum yield measurements was reproduced in
the fluorescence studies in dry or wet methanol and acetonitrile. The yield for GAV photodegradation in water is
smaller than in the organic solvents. Complexation of GAV to β- and γ-cyclodextrins leads to an increase in the
photodecomposition yield, especially during the early part of the photodecomposition kinetics. The transients formed in
the photolysis of GAV in water are similar to those observed in acetonitrile, and the lifetime of triplet GAV in water is
130 ns. In addition, solvated electrons were formed when GAV was photolyzed in water. Complexation of GAV to
cyclodextrins leads to an increase in the triplet yield compared to the radical yield and a moderate shortening of the
triplet lifetime.
Key words: α-guaiacoxyacetoveratrone, cyclodextrin, excited singlet states, excited triplet states, radicals.
Résumé : Utilisant l’émission de fluorescence des photoproduits formés, on a étudié la photodécomposition de l’α-
guaïacoxyacétovératrone (GAV) dans des solvants homogènes et dans des solutions aqueuses de cyclodextrine. Dans les
études de fluorescence dans le méthanol sec ou aqueux ainsi que dans l’acétonitrile, on a reproduit le même
comportement qualitatif que celui observé antérieurement lors de mesures de rendements quantiques pour la
photodégradation. Le rendement pour la photodégradation de la GAV dans l’eau est inférieur à celui observé dans les
solvants organiques. La complexation de la GAV par les β- et γ-cyclodextrines provoque une augmentation du
rendement de photodécomposition, particulièrement au début de la cinétique de photodécomposition. Les espèces
transitoires qui se forment lors de la photolyse de la GAV dans l’eau sont semblables à celles observées dans
l’acétonitrile et le temps de vie de l’état triplet de la GAV dans l’eau est de 130 ns. De plus, il se forme des électrons
solvatés lors de la photolyse de la GAV dans l’eau. La complexation de la GAV par les cyclodextrines provoque une
augmentation du rendement de l’état triplet par rapport au rendement en radical ainsi qu’à une diminution modérée du
temps de vie du triplet.
Mots clés : α-guaïacoxyacétovératrone, cyclodextrine, états singulets excités, états triplets excités, radicaux.
[Traduit par la Rédaction] Okano et al. 1365
Introduction
ways, the multiplicity of the excited states involved, and to
characterize spectroscopically the reactive intermediates in-
volved (3–13). α-Guaiacoxyacetoveratrone (GAV) has fre-
quently been used as a model compound (8, 9, 13, 14),
because it contains key features of lignin. In contrast to
other aryl ketones, it was shown that α-phenoxyacetophen-
ones, such as GAV, react from their excited singlet state (8,
11–14) to form the corresponding phenacyl and phenoxy
radicals (Scheme 1). This reactivity suggests that the β-
cleavage process from the excited singlet state readily com-
petes with intersystem crossing, leading to quantum yields
for the latter process that are significantly smaller than unity
(13). Compared to α-phenoxyacetophenone (15), the triplet
reactivity of GAV is diminished because its lowest triplet
state has a π,π* configuration (9). The triplet lifetime is de-
termined by the efficiency for intramolecular deactivation
through β-phenyl quenching, as was observed for β-arylpro-
piophenone and α-(aryloxy)acetophenones (15–21), and the
rate constant for β-cleavage (14). The photodecomposition
quantum yields and triplet lifetimes of GAV vary signifi-
cantly in different solvents (8, 14). This complex behavior is
The photodecomposition of lignin is responsible for the
yellowing observed in paper manufactured from mechanical
pulp. In recent years, a significant effort has been made to
find efficient inhibition strategies to eliminate or consider-
ably slow down the yellowing of lignin-containing paper.
The mechanisms for lignin photodegradation are complex
involving several reactive intermediates (1, 2). Most mecha-
nistic studies have been performed in solution employing
model compounds because lignin is a high molecular weight
polymer that is difficult to handle in solution. Solution stud-
ies were important to determine the decomposition path-
Received March 9, 1999.
L.T. Okano, R. Ovans, V. Zunic, J.N. Moorthy, and C.
Bohne.1 Department of Chemistry, University of Victoria,
P.O. Box 3065, Victoria, BC V8W 3V6, Canada.
1Author to whom correspondence may be addressed.
Telephone: (250) 721-7151. Fax: (250) 721-7147.
e-mail: bohne@uvic.ca
Can. J. Chem. 77: 1356–1365 (1999)
© 1999 NRC Canada

Okano et al.
1357
Scheme 1.
related to the reactivity from the GAV singlet excited state
and a solvent effect on the intramolecular quenching in the
triplet state.
the photochemistry of β-phenylpropiophenone (25, 26) and
α-phenoxyacetophenone (27) have been previously reported.
In addition, we chose to work in water, since paper has some
moisture (5–10% water), and lignin is probably exposed to a
hydroxylic environment.
Lignin is located in a constrained environment in the cell
wall of wood fibers that influences to a different extent the
deactivation pathways of the excited singlet and triplet states
when compared to these processes in solution. There are a
limited number of studies on the photoreactivity of lignin
model compounds in constrained environments or on solid
supports. The photodegradation of GAV has been shown to
decrease when this compound is adsorbed on paper fibers
(9). In addition, GAV is stable in the crystalline state, a fact
that was attributed to the formation of excimer-like arrange-
ments in the solid state (6). The photochemistry of α-
(aryloxy)acetoveratrones was studied on solid silica gel, Na-
X zeolites, and cellulose (4). In the case of cellulose, an in-
efficient photodecomposition and an enhanced triplet life-
time were observed. These experiments show that the
confinement of lignin in the fibers of paper can alter the
lignin photoreactivity. Furthermore, the mobility of oxygen
through the paper fiber may also be important, since yellow
products are only formed in the presence of oxygen (2). In
this respect, it has been shown that the quenching of reactive
intermediates on cellulose or paper by oxygen is very ineffi-
cient (3, 4, 22–24), suggesting that the reaction of reactive
intermediates with oxygen may be the rate-limiting step in
the photoyellowing process.
Experimental
The β- (lot F6080–191) and γ-CDs (lot E 8056) were a
kind gift from Cerestar and were used as received. Guaiacol
(98%, Aldrich) and 3,4-dimethoxyacetophenone (acetoveratrone)
(98%, Aldrich) were distilled under reduced pressure, meth-
anol (ACP, spectrograde) was distilled from CaH2, and
acetonitrile (ACP, HPLC grade) was distilled under N2 from
LiAlH4. Deionized water (SYBRON, Barnstead deionizing
system) was used for all aqueous samples. Whatman filter
papers (number 1001 125) were used as received.
The α-guaiacoxyacetoveratrone (3′,4′-dimethoxy-α-(2-
methoxyphenoxy)acetophenone, GAV) was synthesized by
the method previously reported (28). The required synthesis
of 3,4-dimethoxy-α-bromoacetophenone was modified from
that previously reported in that dimethoxyacetophenone was
reacted with pyridinium bromide perbromide in chloroform
and under nitrogen (29–33). GAV was purified by repeated
recrystallization from ethanol (95%), and its purity was
1
checked by mp (77–80°C) and H NMR measurements.
UV-vis absorption spectra were measured on Cary 1 or
The objective of this work is to establish how constrained
environments can affect the photochemistry of GAV and the
mobility of the radicals formed in its photodecomposition.
Since the work on solids has proven to be difficult, we de-
cided to investigate the effect of a constrained environment
on the photochemistry of GAV in solution. Cyclodextrins
(CDs) were chosen, since they provide a chemical environ-
ment similar to cellulose, and the effect of CD inclusion on
Cary
5
spectrometers from Varian. Induced circular
Jasco J-720
dichroism spectra were measured on
a
spectropolarimeter. Fluorescence was measured with a PTI
QM-2 fluorimeter at 20.0 ± 0.5°C. Most samples were ex-
cited at 280 nm, however, when necessary the excitation
wavelength was shortened to avoid the Raman emission
from the solvents. Emission spectra were collected before
and after the photodecomposition kinetics to monitor prod-
© 1999 NRC Canada

1358
Can. J. Chem. Vol. 77, 1999
Fig. 1. Induced circular dichroism of 15.8 µM GAV in the
presence of 5 mM β-CD (a, 1) and 10 mM γ-CD (b, 1). The
spectra for CDs (2) are also shown.
uct formation. Narrow slits (bandwidth 1–2 nm) were em-
ployed to measure the emission spectra to minimize the
photodecomposition of GAV. The photodecomposition ki-
netics of GAV was monitored simultaneously by the product
emission at 310 and 350 nm. The rate of decomposition was
dependent on the excitation intensity and bandwidth (5 or
7 nm) employed for the excitation slits. Solutions were con-
tained in 10 × 10 mm quartz cells and were purged with N2
or O2 for 30 min before the fluorescence measurements were
performed. Solid samples were either mounted on a triangu-
lar metal support or were suspended within the sample
holder. In both cases, the excitation beam and detector were
in a 90o arrangement, and for the solid samples, the surface
of the paper was placed at a 45° angle with respect to the ex-
citation beam.
1.5
(a)
1.0
1
0.5
2
0
200
300
400
The laser flash photolysis system used for the transient
studies has been previously described (34). Samples at 20 ±
2°C were excited either with a Lumonics excimer laser at
308 nm (<30 mJ/pulse, 5 ns pulse width) or a Spectra Phys-
ics Nd:YAG laser at 266 nm (<20 mJ/pulse, 6 ns pulse
width). A program written with Labview 4.0™ (National In-
struments) was used to control the laser flash photolysis sys-
tems, acquire the data, and for initial analysis of the decay
curves. A flow cell (7 × 7 mm Suprasil) was employed for
solutions purged with N2 and O2 (30–40 min) so that a fresh
portion of the solution was irradiated at each laser pulse.
The solution in the reservoir before entering the flow cell
was covered with aluminum foil to avoid photodecom-
position by ambient light. A 7 × 7 mm Suprasil cell
stoppered with a rubber septum was used for N2O purged
(20 min) solutions. The photodecomposition in the static
samples was followed by UV-vis absorption, and the sample
was subjected to a small number of laser shots so that only
limited decomposition of GAV occurred.
Solutions of GAV in all solvents except water were pre-
pared by dissolving the solid in the respective solvent.
Aqueous solutions were prepared by the injection of a small
amount (<1%) of the methanolic GAV solution into water. In
all cases, the concentration of GAV was 15.8 µM. The CD
solutions were prepared by dissolving, in the aqueous GAV
solution, the appropriate amounts of β- and γ-CD to reach fi-
nal concentrations of 10 and 30 mM, respectively. These so-
lutions were warmed in hot water for 5 min and were left
stirring for at least 12 h. Some insoluble material was re-
moved by filtering the solutions through 40 µm Millipore
filters. For this reason, it is likely that the GAV and CD con-
centrations quoted are somewhat lower than stated. During
preparation and storage, all solutions were shielded as much
as possible from ambient light.
(b)
1.5
1
1.0
2
0.5
0
200
300
400
Wavelength (nm)
described above. The immersion method leads to a more ho-
mogeneous distribution of additives in paper (35). The
amount of GAV added was established by comparing the
weight of each paper sample before and after the application
of GAV. Filter papers without any treatment were employed
as a reference. The samples were irradiated for 1 h in front
of a 75W Xe lamp (PTI).
Results
Several spectroscopic techniques can be employed to es-
tablish the complexation of guest molecules with CDs. The
primary consideration in choosing a technique is that a
change of a spectroscopic property has to be observed for
the free and complexed guest molecule. The formation of
the GAV–CD complexes was observed by circular dichroism
spectroscopy. GAV in water does not show a circular
dichroism signal because it is achiral. In the presence of β-
or γ-CD, induced circular dichroism signals were observed
in the spectral region where GAV absorbs (Fig. 1). These ex-
periments clearly show that GAV is incorporated to some
extent into both CD cavities. Unfortunately, circular
dichroism could not be employed to determine the equilib-
rium constants values and the complexation stoichiometries
due to the poor signal-to-noise ratio observed and the fact
that some photodestruction (<10%) occurs upon irradiation
Filter paper samples (5 × 1.1 cm) containing GAV were
prepared by using either the drop or immersion methods
(35). For the drop method, 0.15–0.5 mL of a methanolic or
methanol–water solution of GAV (≤0.01 M) was slowly
dropped with a microsyringe on samples mounted in Delrin
frames. Two samples were always prepared from the same
GAV solutions. After application of GAV, the samples were
protected from light and were dried for at least 12 h. In the
case of the immersion method, two samples were simulta-
neously placed for 20 min in Petri dishes containing GAV
solutions (≤0.01 M). After the immersion period, the paper
samples were mounted on Delrin frames and were dried as
1
in the spectrometer. Likewise, H NMR could not be em-
ployed to determine the complexation efficiency because of
the low solubility of GAV in water (<20 µM).
Fluorescence is, in principle, an ideal technique to estab-
lish the complexation of guest molecules with CDs, because
fluorescence properties, such as the emission quantum yields
© 1999 NRC Canada

Okano et al.
1359
Fig. 3. Fluorescence spectra of 15.8 µM acetoveratrone (a, λex
280 nm) and guaiacol (b, λex = 240 nm) in water (1), methanol
(2), and acetonitrile (3). The sharp peak at 260 nm
Fig. 2. Emission spectra before (2) and after (1) irradiation of
15.8 µM GAV for 1 h at 280 nm in (a) water, (b) methanol, and
(c) acetonitrile.
=
(b) corresponds to the Raman emission of the solvent.
(a)
(a)
0.5
1
1
2
0.3
2
0.1
3
300
400
500
350
400
450
500
(b)
3.0
(b)
2.0
1.0
0
1
1
2
2
3
300
350
400
450
(c)
250
300
350
400
450
Wavelength (nm)
1.0
0.6
0.2
fluorescence spectra for guaiacol were the same for all three
solvents, and its emission efficiency in acetonitrile and
methanol was somewhat higher (ca. 2) than in water. In con-
trast, the acetoveratrone emission is very solvent dependent
with emission maxima at 420, 390, and ca. 330 nm, respec-
tively, in water, methanol, and acetonitrile, and the emission
quantum yield is higher in water than in the organic sol-
vents.
1
2
300
350
400
450
The decomposition quantum yields of GAV in homoge-
neous solution are very dependent on the nature of the sol-
vent and the presence of water (8). For this reason, we used
the appearance of the photoproduct emission to measure rel-
ative photodecomposition yields of GAV in different sol-
vents. The kinetic traces for the appearance of photoproducts
were measured at 310 and 350 nm (Fig. 4). The rate of ap-
pearance was faster in methanol than in acetonitrile. At
310 nm, the photoproduct formation rate follows a fairly lin-
ear relationship, whereas at 350 nm, a considerable slow-
down is observed as the photodecomposition of GAV
proceeds. Most kinetic traces showed initially a fast increase
in the fluorescence intensity. This burst kinetics appears as
an initial jump in the traces shown in Fig. 4. The initial
jump also incorporates any product formed when spectra
were collected with narrow slits before the kinetic run and
any product formed due to exposure to ambient light during
sample preparation and handling.
Wavelength (nm)
and lifetimes, can be very sensitive to the environment
around the fluorophore. The first approach tried to establish
the complexation of GAV to β and γ-CD was fluorimetry,
since an emission assigned to the fluorescence of GAV in
methanol had been previously reported (6). Unfortunately,
we established that GAV does not fluoresce when the sam-
ples are shielded from the exposure to ambient light (Fig. 2).
Appreciable fluorescence is only observed after the samples
are exposed to the excitation beam in the fluorimeter
(Fig. 2). This emission was assigned to the fluorescence of
photoproducts. The spectral distribution of the emission is
solvent dependent. The emission observed in water and
methanol between 310 and 320 nm, after exposure of the
sample to the excitation beam, is not due to Raman emission
of the solvent because the spectra are the same at different
excitation wavelengths.
Relative yields for the formation of products were esti-
mated from the initial slopes of the photodecomposition ki-
netics after the burst phase (Table 1). In the absence of
oxygen, the rate for appearance of photoproduct emission at
310 nm is highest in methanol and comparable in
acetonitrile and water. At 350 nm, the photodecomposition
Guaiacol and acetoveratrone are likely products in the
photolysis of GAV (Scheme 1). For this reason, their fluo-
rescence spectra in various solvents was measured (Fig. 3).
Guaiacol has a higher fluorescence quantum yield by about
one order of magnitude than acetoveratrone. In addition, the
© 1999 NRC Canada

1360
Can. J. Chem. Vol. 77, 1999
Table 1. Relative rates for the appearance of the photoproduct emission at 310 or 350 nm, estimated from the
initial slopes and normalized to the rate at 310 nm in water in the presence of nitrogen.
Relative rates
Solvent
310 nm/N2
310 nm/O2
350 nm/N2
350 nm/O2
H2O
CH3OH
1
8
9
2
1
3
4
1
1
3
2
0.2
2
4
0.3
22
2
12
3
0.5
3
1
6
1
CH3OH/H2O (60/40)
CH3OH/H2O (40/60)
CH3CN
CH3CN/H2O (60/40)
CH3CN/H2O (40/60)
5
11
3
2
Fig. 4. Kinetics for the appearance of photoproduct emission for
the photolysis (λex = 280 nm) of 15.8 µM GAV monitored at
310 (a) and 350 (b) nm in methanol (1), acetonitrile (2), and
water (3). The opening of the shutter leads to the initial jump in
the emission intensities.
Fig. 5. Kinetics for the appearance of photoproduct emission for
the photolysis (λex = 280 nm) of 15.8 µM GAV monitored at 310
(a) and 345 (b) nm in the presence of 10 mM β-CD (1), 30 mM
γ-CD (2), 10 mM γ-CD (3), 5 mM β-CD (4), and in water (5).
The traces in (a) show the kinetics after the initial jump, while
the experiments in (b) show the burst kinetics at short times.
3.0
(a)
1
1
(a)
2
3
2.0
1.0
2
3
4
5
0
0
1.0
2.0
3.0
(b)
1
3.0
2.0
(b)
1
3
5
2
1.0
0
3
3.0
0
1.0
2.0
3
Time (10 s)
0
0.1
0.2
3
Time (10 s)
rate is highest in methanol and lowest in water (Fig. 4). The
presence of oxygen has only a small effect on the rate of
product appearance in water, due to the low solubility of ox-
ygen in this solvent (1.4 mM) (36). In methanol and
acetonitrile, the presence of oxygen led to a marked decrease
in the rate of emission of the photoproducts, suggesting that
oxygen efficiently quenched the excited state responsible for
product formation.
The presence of water in methanol was shown to decrease
the degradation quantum yield of GAV, whereas the opposite
effect was reported for the photolysis in acetonitrile (8). The
relative rates for photoproduct emission showed a similar
trend. The emission rate at 310 nm in methanol was only de-
creased at a high water content (60%), whereas a more sig-
nificant decrease of the rate was measured at 350 nm
(Table 1). In the case of GAV photolysis in acetonitrile, an
increase of the photoproduct formation was observed in the
presence of water (Table 1).
The emission spectrum of GAV in the presence of β- and
γ-CD is similar to that observed for GAV in water. However,
the contribution of the emission at 450 nm is higher in the
presence of CDs. The complexation of GAV to β- and γ-CD
led to a slight increase in the rate of appearance of the
photoproduct emission when monitored at 310 and 350 nm
(Fig. 5, Table 2). Some of this enhancement is quenched in
the presence of oxygen. A more significant effect of the CD
complexation is observed on the initial jump corresponding
to the burst phase (Table 3). The enhancement of the initial
jump increases when the β-CD concentration is raised, and
this enhancement is not sensitive to the presence of oxygen.
In contrast, the enhancement in the presence of γ-CD is sen-
sitive to the presence of oxygen (Table 3).
© 1999 NRC Canada

Okano et al.
1361
Fig. 6. Transient absorption spectra of 15.8 µM GAV/N2 after
excitation at 308 nm in water (a, (᭿) 90 and (᭺) 570 ns
delays), in the presence of 10 mM β-CD (b, (᭿) 70 and (᭺) 360
ns delays), and in the presence of 10 mM γ-CD (c, (᭿) 80 and
(᭺) 640 ns delays).
Table 2. Relative rate ratio for the estimated appearance of the
GAV photoproduct emission in the presence of β- and γ-CD with
respect to the appearance rate in water.
Relative rate ratio (CD/H2O)
CDs
310 nm/N2
350 nm/N2
0.006
5 mM β-
3
4
1
3
2
6
2
4
a
10 mM β-
10 mM γ-
30 mM γ-
0.004
Table 3. Ratios for the intensities of the initial jump in the
presence of β- and γ-CD with respect to the intensity in water.
0.002
0
Relative rate ratio (CD/H2O)
CDs
5 mM β-
310 nm/N2 310 nm/O2 350 nm/N2 350 nm/O2
0.008
b
2
2
3
2
10 mM β- 29
10 mM γ- 22
30 mM γ- 28
24
5
6
15
13
20
13
5
14
0.006
0.004
0.002
0
The emission of filter paper with and without GAV was
measured to determine whether fluorescence spectroscopy
could be employed to monitor the photodecomposition of
GAV absorbed into the paper matrix. The filter paper with-
out GAV showed some background emission. In the pres-
ence of GAV (≤3% w/w), the same emission spectra were
observed when the sample was excited at 270, 280, or
310 nm. The emission spectra were also measured after the
reference and paper sample containing GAV were irradiated
for 1 h. The emission intensities measured for the reference
and GAV samples were nonreproducible. In some cases, an
increase of the emission intensity was observed, whereas for
duplicate samples prepared from the same GAV solution a
decrease of the intensity was measured. The same pattern
was observed for samples prepared either by the drop or the
immersion methods. Although the contradictory results on
the emission intensity are difficult to explain at this point, it
was observed that the magnitude of the enhancement or the
decrease is dependent on the amount of GAV adsorbed into
the paper and the duration of the irradiation of the sample.
Laser flash photolysis was employed to study the transient
species formed in the photolysis of GAV. All transient stud-
ies, with the exception of those in the presence of N2O, were
performed using flow cells. When static cells were em-
ployed, a higher yield of radicals compared to the triplet
yield was observed. This is particularly noticeable when
GAV is excited in the presence of CDs. The higher yield of
radicals is probably due to the accumulation of long-lived
radicals after subsequent laser shots.
0.006
c
0.004
0.002
0
350
450
550
650
Wavelength (nm)
The transient phenomena of GAV in water have not been
previously described. The same transients as observed in
acetonitrile were also present in water (Fig. 6a). In addition,
in water we observed the formation of solvated electrons
that absorb above 600 nm. The solvated electrons are proba-
bly formed in a two-photon process leading to the
photoionization of GAV. The presence of solvated electrons
was established by the disappearance, in the presence of
N2O, of the short-lived absorption measured at 640 nm. The
triplet lifetime of GAV in water was (130 ± 10) ns, and it is
shortened to (76 ± 3) ns in oxygen saturated solutions. This
result is consistent with a diffusional quenching rate con-
stant by oxygen, taking into account the oxygen solubility in
water (1.4 mM).
Complexation of GAV to 10 mM β- or γ-CD leads to a
larger yield of triplets when compared to the radical yield
measured for the photolysis of GAV in water (Fig. 6). The
triplet GAV lifetime in the presence of β-CD is (110 ± 3) ns
and (104 ± 4) ns in the absence and presence of oxygen,
whereas in the presence of γ-CD the respective triplet life-
times are (97 ± 5) and (88 ± 7) ns. The residual absorption
The transient phenomena observed after excitation at
308 nm of GAV in acetonitrile were similar to those previ-
ously reported (5, 7, 8, 13). The short-lived triplet (τ = 254
ns) was observed around 400 nm, and the weak transient ab-
sorption in the 400–500 nm region of the phenacyl radical
was also present. Both these transients were efficiently
quenched by oxygen. The remaining transient absorption in
the 300–400 nm region in the presence of oxygen is due to
the guaiacoxy radical.
© 1999 NRC Canada

1362
Can. J. Chem. Vol. 77, 1999
Table 4. Average ratios for the solvated electron yield formed in
the photolysis at 308 nm of GAV in the presence of CDs and in
water. The number in parentheses corresponds to the number of
experiments averaged.
tum yield is about one order of magnitude higher than that
for acetoveratrone. The emission observed at 420 nm in wa-
ter and between 320 and 340 nm in methanol and
acetonitrile is assigned to products with the acetoveratrone
moiety. This blue shift of the emission maxima was also ob-
served when the fluorescence of acetoveratrone was mea-
sured in these organic solvents. The larger blue shift
observed for the photoproduct emission of GAV could be
due to a more significant solvent effect on the emission
spectra of the coupling products when compared to the shift
for acetoveratrone.
CD
Solvated electron ratio (CD/H2O)
10 mM β-
10 mM γ-
30 mM γ-
1.4 ± 0.2 (5)
1.2 ± 0.1 (5)
1.3 ± 0.2 (2)
observed after the decay of the triplet corresponds to the
guaiacoxy radical, which is very long lived (>100 µs).
In the presence of CDs, an increase was observed for the
yield of solvated electrons. The amount of solvated electrons
formed was estimated by subtracting the ∆A values at
640 nm in the presence of N2O, where all solvated electrons
are quenched, from the value determined in the presence of
N2. Since the photoionization of GAV is likely to be a two-
photon process, the absolute yield for solvated electron for-
mation is very dependent on the laser energy that excites the
sample. For this reason, the yields for solvated electron for-
mation have to be compared for measurements performed
for the same experimental conditions. Although the absolute
values for these yields changed between experiments per-
formed on different days, it was always observed that the
yields in the presence of β- and γ-CD were higher than those
measured in water (Table 4). Experiments performed by ex-
citing GAV at 266 nm in the presence of 10 mM of both
CDs led to the same ratios for solvated electron formation as
observed for the excitation at 308 nm.
A brief comment on the fact that acetoveratrone fluo-
resces is justified because it has implications on why GAV is
reactive from the excited singlet state. The fluorescence of
some methoxy substituted aryl ketones in water has been
previously reported (37), but in most cases, ketones do not
fluoresce because their intersystem crossing quantum yield
is close to unity (38). The intersystem crossing efficiency
depends on the configurations and the energy difference be-
tween the singlet and triplet states involved in the transition.
The configuration of the excited states of ketones can be in-
fluenced by the presence of electron-donating substituents
that stabilize π,π* states compared to n,π* states (39). In the
case of acetoveratrone, the presence of two methoxy sub-
stituents will stabilize the π,π* states in the singlet and triplet
manifold, leading to a decrease of the intersystem crossing
rate constant by either switching the configuration of S1 to a
π,π* one or by influencing the S1–Tn energy gap involved in
the intersystem crossing process. This decrease of the
intersystem crossing quantum yield leads to lengthening of
the excited singlet lifetime and the observation of fluores-
cence. This stabilization in the case of GAV leads to a less
reactive triplet state when compared to ketones with n,π*
triplets (13). Furthermore, the GAV photophysics will pri-
marily be determined by the acetoveratrone moiety. This en-
sures that the rate constant for intersystem crossing of GAV
is lower than for other ketones (13) and that the GAV ex-
cited singlet state can undergo a β-cleavage reaction. In con-
trast to acetoveratrone, GAV does not fluoresce in solution,
suggesting that the rate constant for β-cleavage is larger than
the radiative rate constant responsible for fluorescence. The
emission of the photoproducts of GAV in methanol was pre-
viously assigned incorrectly to the fluorescence of GAV (6),
and based on this assignment, the authors explained the
crystalline GAV emission with a maximum at 450 nm to be
due to an excimer formed from two adjacent acetoveratrone
moieties. The emission observed in the solid state may be
due to the emission of GAV, since this emission occurs in
the spectral region where acetoveratrone fluoresces in water.
However, other evidence previously reported (6) supports
this emission to be due to excimer formation. A definite as-
signment of this emission is problematic, since the emission
spectra of acetoveratrone was shown to depend on the sol-
vent polarity, and a similar effect is expected to be observed
for GAV.
Discussion
The photodecomposition of GAV in homogeneous solu-
tion leads to the formation several products (Scheme 1) (8,
13, 14). Some of these photoproducts fluoresce, and since
fluorescence spectroscopy is a very sensitive technique, a
small amount of product formation can be detected before
any decomposition is visible by other spectroscopic tech-
niques, such as UV-vis absorption and NMR. The spectral
distribution for the photoproduct emission following the
photolysis of GAV is very dependent on the nature of the
solvent. This difference could be due to changes in the prod-
uct distribution or different emission quantum yields for
each product in the different solvents. Since the emission
quantum yields of guaiacol or acetoveratrone were shown
not to vary by more than a factor of 2, we explain the differ-
ent spectra observed to be due to different product distribu-
tions.
The products from the photolysis of GAV will have differ-
ent fluorescence spectra as exemplified for guaiacol and
acetoveratrone. The coupling products shown in Scheme 1
are expected to have a fluorescence emission that is similar
to that of acetoveratrone, since the chromophore of the latter
is present in these compounds. The emission between 300
and 320 nm observed for the photoproduct formation of
GAV in methanol and water is assigned to the formation of
guaiacol. This spectral feature is absent for the photolysis of
GAV in acetonitrile, suggesting that guaiacol is not formed
in this solvent. The lack of detection of guaiacol is not due
to a sensitivity problem, since the guaiacol emission quan-
The objective of the kinetic studies for photoproduct ap-
pearance was to investigate if the relative rates observed in
the fluorescence measurements can be correlated to the re-
ported photodegradation quantum yields (8). It is important
to take into account that the fluorescence intensity is directly
proportional to the emission quantum yield of the excited
© 1999 NRC Canada

Okano et al.
1363
fluorophores. For this reason, the presence in solution of any
substance that quenches the emission of the fluorophore will
lead to a decrease of the measured fluorescence intensity.
For example, if during the photolysis of GAV one of the
products is an efficient quencher of the fluorescence from a
second product, the emission intensity of the latter will be
decreased. Due to this possible quenching effect, any analy-
sis of the shape of the kinetic trace (i.e., linear at 310 nm
and curved at 350 nm) was not attempted. Furthermore, the
photolysis of GAV leads to a mixture of products that can
fluoresce. At each wavelength, the intensity measured is a
function of the concentration of each product, its absorption
coefficient at the excitation wavelength, and its emission
quantum yield at the wavelength where the kinetics is being
monitored; we did not expect to observe the same relative
rates when the photoproduct formation was monitored at
310 and 350 nm. Clearly, no quantitative information is
available from the kinetic studies performed. However, the
relative ordering of the photodegradation yields in methanol
and acetonitrile as well as the effect of water addition to
these solvents were reproduced. This adequate qualitative
correspondence of the photodegradation quantum yield with
the relative fluorescence rates shows that fluorescence can
be used as a first screening technique to determine the
photodecomposition yields of lignin model compounds. The
advantage of using fluorescence for preliminary studies is
that the experiments are much quicker to perform than the
determination of quantum yields, and fluorescence can be
employed for the screening of different experimental condi-
tions before performing quantum yield measurements.
The fluorescence experiments were employed to estimate
the GAV photodecomposition quantum yield in water, since
this value has not been previously reported. The quantum
yield in water is much smaller than in methanol and is prob-
ably also smaller than in acetonitrile. The emission spectra
for the photoproducts indicate that the product distribution
observed in water is closer to that in methanol than the prod-
uct distribution observed in acetonitrile. These experiments
show that the photodecomposition pattern of GAV in water,
and probably other lignin model molecules, cannot be ex-
trapolated from studies in organic solvents. This is an impor-
tant point when trying to understand the photodegradation of
lignin because the paper matrix contains some water.
circular dichroism signals of GAV in β- and γ-CD suggests
that the guest is incorporated in a similar environment in
both cavities. No detailed information on the complexation
environment could be obtained because the signal-to-noise
ratio of the spectra is too poor to warrant any further analy-
sis. The complexation of α-phenoxyacetophenone with β-CD
has been previously reported (27), and a 1:1 complex in
which the benzoyl moiety is encapsulated was proposed.
GAV is too large to completely fit within the β-CD cavity,
and for 1:1 complexes, part of the molecule will be exposed
to the aqueous phase. However, the possibility exists that at
higher β-CD concentrations, a 1:2 (GAV:CD) complex is
formed that protects the guest from the aqueous environ-
ment. In the case of the larger γ-CD, the benzoyl and
phenoxy moieties could be included in the same cavity as
was previously proposed for the inclusion of β-phenyl-p-
methoxypropiophenone (25).
In principle, the complexation of GAV to CDs could lead
to either a decrease or increase of the photodegradation rate.
In the case of a 1:2 (GAV:CD) complex, we would expect a
decrease of the photodegradation rate, since separation of
the radicals would be inhibited, leading to an increase in the
probability for GAV regeneration. Since such an increase
was not observed, it is likely that 1:2 complexes are not
present to an appreciable amount. In contrast, the entrap-
ment of one of the radical moieties within the CD cavity and
efficient release of the second radical could lead to an en-
hancement of the GAV photodecomposition quantum yield.
Surprisingly, most of the enhancement effect is observed on
the initial jump and not on the relative rates at longer times.
The relative rates are increased moderately (2–6) by the
presence of CDs, indicating that complexation of the radi-
cals to the CD cavity does not affect significantly the rate of
product formation after the burst phase (initial jump). In the
laser flash photolysis studies, we established that at least the
phenoxy radicals are very long lived (>100 µs). Since the
exit rate constants of guests from CD cavities are of the or-
der of 105–106 s–1 (43, 44), the long radical lifetimes ensure
that radicals formed from the GAV photoreaction can exit
and enter CD cavities for many times before they are in-
volved in recombination reactions. Consequently, the entrap-
ment of the radicals does not lead to a significant alteration
of the rate for product formation.
We tried preliminary experiments to explore the fluores-
cence of product formation for the decomposition of GAV
when incorporated into paper. These experiments were in-
conclusive, suggesting that the decomposition rate and (or)
the emission properties of the photodecomposition products
are affected by the incorporation in the paper matrix. More-
over, these experiments show the complexity of the behavior
of fluorophores in paper and that degradation of lignin may
have a quenching effect when fluorescent brightners are em-
ployed to counteract the yellowing of paper made from me-
chanical pulp.
The complexation of GAV to β- and γ-CD was established
by the appearance of an induced circular dichroism signal
when GAV was solubilized in the presence of CD. The di-
rection and strengths of the induced circular dichroism sig-
nals of an achiral guest in the CD cavity can be related to
the position of the guest within the CD cavity (40–42). The
fact that the same direction was observed for the induced
The most pronounced effect of CD complexation on the
GAV photodecomposition was observed on the enhancement
of the amount of product formed during the burst phase. The
magnitude of this effect depends on the CD concentration,
indicating that this enhancement is due to the complexation
of GAV to CDs. This increase is not due to an inhibition of
the recombination probability of the phenacyl and guaiacoxy
radicals to regenerate GAV, since this inhibition would mean
that the exit of one of the radicals from the CD cavity is
faster than the separation of the radicals in the solvent cage
formed in water. This is an unlikely scenario because it was
shown that polar molecules such as triplet xanthone, which
are not completely included in the CD cavity (45), have an
exit constant of 8.4 × 106 s–1 (46), which is much slower
than the separation of encounter complexes of neutral mole-
cules in aqueous solution. For this reason, it is likely that the
phenacyl and guaiacoxy radicals remain in close contact
within the CD cavity for a longer period of time than an en-
© 1999 NRC Canada

1364
Can. J. Chem. Vol. 77, 1999
counter complex in the aqueous phase. In this case, the en-
hanced emission intensity is due to a larger yield of coupling
product formation in the CD cavities when compared to the
reaction in the bulk phase.
pure alcohols leading to
a
decrease of the
photodecomposition quantum yield of GAV (8). The triplet
lifetime of GAV in water is also shorter than in pure
acetonitrile. In contrast to the results in alcohols, a slight in-
crease (ca. 1.25) (8) was observed for the triplet lifetime in
acetonitrile containing 40% water, which is accompanied
with a significant increase (2.6) of the photodecomposition
quantum yield (14). Our qualitative fluorescence results in-
dicate that the photodegradation rate is smaller in water than
in acetonitrile, which is in line with the shorter triplet life-
times. The fluorescence results also show the same trend as
previously reported (8) for the GAV degradation in
acetonitrile–water mixtures. These results show that it is not
easy to predict the photochemistry of GAV in mixed sol-
vents.
In addition to the formation of the excited triplet states
and radicals, we also observed that GAV in water is
photoionized, leading to the observation of solvated elec-
trons in the laser flash photolysis experiments. This
photoionization reaction is common for ketones in aqueous
solution and is augmented when the ketone is complexed to
supramolecular structures, such as SDS micelles and bile
salt aggregates (47). The same effect was observed for GAV
when complexed to both CDs, suggesting that incorporation
of GAV into the CD cavity leads to a more efficient separa-
tion of the solvated electron and the GAV radical cation. No
transient signals were observed for the GAV radical cation
probably because its decomposition is very fast. A similar
effect has been reported for other ketones (47). Since the
photoionization process involves the sequential absorption
of two photons, this process can only contribute to the yel-
lowing of paper made from mechanical pulp when very high
irradiation intensities are employed. In most applications
this reaction will not be relevant.
In summary, complexation of GAV to β- and γ-CD leads to
an increase of the initial product formation but has only a
small effect on the rate following the burst kinetics. This re-
sult indicates that the constrained environment around lignin
within the paper matrix could determine the ratio between
coupling of the radicals at different positions of the phenol
ring and recombination to reform GAV, which will influence
the degree of photoyellowing of the paper. The small effect
observed for the relative rates after the initial burst are a
consequence of the fast entry and exit dynamics of the radi-
cals compared to the time domain for the radical–radical re-
actions. A more viscous environment will have to be
employed to study how the decrease of the mobility of radi-
cals by encapsulation into constrained environments affects
the photochemistry of GAV. Finally, we report that GAV
does not fluoresce in homogeneous solution, but the fluores-
cence of the products of GAV photolysis can be employed to
gain qualitative information on the photodegradation kinet-
ics.
An increase of the intersystem crossing quantum yield of
GAV when this molecule is complexed to the CDs could
also be responsible for the increase of products formed dur-
ing the burst phase. The laser flash photolysis data suggest
that the intersystem crossing quantum yield increased, since
a higher triplet to radical concentration is observed for GAV
in the presence of CDs than in water. The triplet lifetimes of
GAV in β- (110 ns) and γ-CD (97 ns) are short enough that
some β-cleavage will occur within the CD cavity. The radi-
cal pair formed from the cleavage of a triplet excited state
has triplet multiplicity, and an intersystem crossing process
has to occur before these radicals can react. For this reason,
it is likely that one of the radicals will have to escape the
CD cavity before any of the recombination or coupling reac-
tions can occur. In this respect, the increased intersystem
crossing for GAV leads to a decrease of the probability for
the recombination of GAV and an enhancement of the prod-
uct formation.
A puzzling effect was observed when oxygen was added
to the solutions containing GAV and CDs. The initial jump
for the fluorescence experiments decreased when GAV was
complexed to γ-CD, but no effect was observed in the pres-
ence of β-CD or for GAV solubilized in the aqueous phase.
Since the excited triplet state of GAV is quenched in the
aqueous phase, the lack of oxygen quenching in the homo-
geneous solution suggests that the products formed in the
initial burst are mainly due to the reactivity from the excited
singlet GAV. The same explanation can be employed for the
results with β-CD. In the case of γ-CD, the quenching by ox-
ygen indicates that some of the products formed during the
burst phase involve a long-lived transient that can react with
oxygen. The triplet lifetime of GAV when complexed to γ-
CD is decreased by a factor of 1.1 when oxygen is present in
solution. This decrease is much smaller than observed for
the quenching of triplet GAV in water, and it suggests that
the triplet GAV complexed to γ-CD is mainly protected from
the interaction with oxygen. This eliminates the triplet as the
candidate for the transient being intercepted by oxygen. The
phenacyl radical is the other transient in the GAV photolysis
that reacts readily with oxygen (7). Thus, the oxygen effect
in the presence of γ-CD could be explained by the trapping
of this radical when it exits the CD cavity. Although we can-
not fully explain the oxygen result, it is worth noting that the
GAV complexes with β- and γ-CD must have sufficiently dif-
ferent geometries to lead to the photochemical behavior ob-
served.
The transients formed in the photolysis of GAV in water
were studied by laser flash photolysis. As already empha-
sized above, the comparison of the photochemistry of GAV
in water to the previously reported reactions in organic sol-
vents is relevant, since lignin will be exposed to an aqueous
like environment in paper. The same transients, i.e., excited
triplet state, and phenacyl and guaiacoxy radicals as ob-
served in acetonitrile were also detected in water. The triplet
lifetime of GAV in water is shorter than observed in the or-
ganic solvents (8). This result explains why the GAV triplet
lifetimes in alcohols containing water are shorter than in the
Acknowledgments
The Natural Sciences and Engineering Research Council
of Canada (NSERC) is thanked for the financial support un-
der the Industrially Oriented Research Grant with Ciba Ad-
ditives Canada. In addition, the authors thanks Dr. J. Ausio
(Dep. Biochem. Microbiol.) for the use of the circular
© 1999 NRC Canada

Okano et al.
1365
22. F. Wilkinson, P.A. Leicester, L.F.V. Ferreira, and V.M.M.R.
Freire. Photochem. Photobiol. 54, 599 (1991).
23. L.F. Vieira Ferreira, A.R. Garcia, M.R. Freixo, and S.M.B.
Costa. J. Chem. Soc. Faraday Trans. 89, 1937 (1993).
24. P.P. Levin, L.F. Vieira Ferreira, and S.M.B. Costa. Chem.
Phys. Lett. 173, 277 (1990).
dichroism spectrometer, Dr. W.S. Patterson for the synthesis
of GAV, Cerestar for the kind gift of the cyclodextrins sam-
ples, and Dr. L. Netter for software support.
References
25. J.C. Netto-Ferreira and J.C. Scaiano. J. Photochem. Photobiol.
A: Chem. 45, 109 (1988).
1. C. Heitner and J.C. Scaiano. Photochemistry of lignocellulosic
materials. ACS Symp. Ser. 531 (1993).
26. C. Bohne, M. Barra, R. Boch, E.B. Abuin, and J.C. Scaiano. J.
Photochem. Photobiol. A: Chem. 65, 249 (1992).
27. N.C. de Lucas and J.C. Netto-Ferreira. J. Photochem.
Photobiol. A: Chem. 103, 137 (1997).
28. E. Adler, B.O. Lindgren, and U. Saedén. Svensk
Papperstidning. 55, 245 (1952).
29. A. Arrieta, I. Ganboa, and C. Palomo. Synth. Commun. 14,
939 (1984).
30. R.M. Cowper and L.H. Davidson. Org. Synth. Coll. vol. II,
480 (1943).
31. C. Djerassi and C.R. Scholz. J. Am. Chem. Soc. 70, 417 (1948).
32. L.L. Landucci, S.A. Geddes, and T.K. Kirk. Holzforschung,
35, 66 (1981).
33. K. Itoh, S. Tachibana, and M. Sumimoto. Mok. Gakk. 38, 579
(1992).
34. Y. Liao and C. Bohne. J. Phys. Chem. 100, 734 (1996).
35. M.P. Fernandez, T.C. Barros, and C. Bohne. J. Photochem.
Photobiol. A: Chem. 116, 171 (1998).
2. G.J. Leary. J. Pulp Paper Sci. 20, J514 (1994).
3. D. Shukla, N.P. Schepp, N. Mathivanan, and L.J. Johnston.
Can. J. Chem. 75, 1820 (1997).
4. L. Hurrell, L.J. Johnston, N. Mathivanan, and D. Vong. Can. J.
Chem. 71, 1340 (1993).
5. S.V. Jovanovic, J. Renaud, A.B. Berinstain, and J.C. Scaiano.
Can. J. Chem. 73, 223 (1995).
6. J.C. Scaiano, A.B. Berinstain, M.K. Whittlesey, P.R.L.
Malenfant, and C. Bensimon. Chem. Mater. 5, 700 (1993).
7. J.C. Scaiano, M.K. Whittlesey, A.B. Berinstain, P.R.L.
Malenfant, and R.H. Schuler. Chem. Mater. 6, 836 (1994).
8. J.A. Schmidt, A.B. Berinstain, F. de Rege, C. Heitner, L.J.
Johnston, and J.C. Scaiano. Can. J. Chem. 69, 104 (1991).
9. C. Vanucci, P. Fornier de Violet, H. Bouas-Laurent and A.
Castellan. J. Photochem. Photobiol. A: Chem. 41, 251 (1988).
10. A. Castellan, N. Colombo, C. Vanucci, P.F. de Violet, and H.
Bouas-Laurent. J. Photochem. Photobiol. A: Chem. 51, 451
(1990).
36. S.L. Murov, I. Carmichael, and G. L. Hug. Handbook of
photochemistry. Marcel Dekker, Inc., New York. 1993.
37. K. Lundquist, B. Josefsson, and G. Nyquist. Holzforschung,
32, 27 (1978).
38. J. Michl and V. Bonacic-Koutecky. Electronic aspects of or-
ganic photochemistry. John Wiley & Sons Inc., New York.
1990.
39. N.J. Turro. Modern molecular photochemistry. Benjam-
in/Cummings Publishing Co., Menlo Park. 1978.
40. M. Kodaka. J. Am. Chem. Soc. 115, 3702 (1993).
41. G. Marconi and B. Mayer. Pure Appl. Chem. 69, 779 (1997).
42. B. Mayer, G. Marconi, C. Klein, G. Köhler, and P. Wolschann.
J. Incl. Phenom. Molec. Recog. Chem. 29, 79 (1997).
43. M.H. Kleinman and C. Bohne. In Molecular and supramolecu-
lar photochemistry. Edited by V. Ramamurthy and K.S.
Schanze. Marcel Dekker Inc., New York. 1997. p. 391.
44. T.C. Barros, K. Stefaniak, J.F. Holzwarth, and C. Bohne. J.
Phys. Chem. A, 102, 5639 (1998).
11. W.U. Palm and H. Dreeskamp. J. Photochem. Photobiol. A:
Chem. 52, 439 (1990).
12. W.-U. Palm, H. Dreeskamp, H. Bouas-Laurent, and A.
Castellan. Ber. Bunsenges. Phys. Chem. 96, 50 (1992).
13. A.B. Berinstain, M.K. Whittlesey, and J.C. Scaiano. In
Photochemistry of lignocellulosic materials. Edited by C.
Heitner and J.C. Scaiano. American Chemical Society, Wash-
ington. 1993. p. 111.
14. J.A. Schmidt, E. Goldsmidt, C. Heitner, J.C. Scaiano, A.B.
Berinstain, and L.J. Johnston. In Photochemistry of ligno-
cellulosic materials. Edited by C. Heitner and J.C. Scaiano.
American Chemical Society, Washington. 1993. p. 122.
15. J.C. Netto-Ferreira, I.G.J. Avellar, and J.C. Scaiano. J. Org.
Chem. 55, 89 (1990).
16. F.R. Stermitz, D.E. Nicodem, V.P. Muralidhadran, and C.M.
O’Donnell. Mol. Photochem. 2, 87 (1970).
17. P.J. Wagner, P.A. Kelso, A.E. Kemppainen, A. Haug, and D.R.
Graber. Mol. Photochem. 2, 81 (1970).
45. R.S. Murphy, T.C. Barros, J. Barnes, B. Mayer, G. Marconi,
and C. Bohne. J. Phys. Chem. A, 103, 137 (1999).
46. Y. Liao, J. Frank, J.F. Holzwarth, and C. Bohne. J. Chem. Soc.
Chem. Commun. 199 (1995).
47. R. Boch, M.K. Whittlesey, and J.C. Scaiano. J. Phys. Chem.
98, 7854 (1994).
18. D.G. Whitten and W.E. Punch. Mol. Photochem. 2, 77 (1970).
19. T. Wismontski-Knittel and T. Kilp. J. Phys. Chem. 88, 110
(1984).
20. J.C. Netto-Ferreira, W.J. Leigh, and J.C. Scaiano. J. Am.
Chem. Soc. 107, 2617 (1985).
21. W.J. Leigh, J.H. Banisch, and M.S. Workentin. J. Chem. Soc.
Chem. Commun. 988 (1993).
© 1999 NRC Canada