Electronic absorption and fluorescence properties of 2,5-diarylidene-cyclopentanones

Article abstract of DOI:10.1021/jp0219597

Spectroscopic properties for a series of 2,5-diarylidene-cyclopentanones are reported. Electronic absorption and fluorescence spectra have been measured for the all-E configurations of 2,5-dibenzylidene-cyclopentanone (1), 2,5-bis-(3-phenyl-allylidene)-cyclopentanone (2), and 2,5-bis-(5-phenyl-penta-2,4-dienylidene)-cyclopentanone (3). The absorption spectra have been assigned with the aid of INDO/S calculations. Molecular structures used for the INDO/S calculations were computed with the PM3 Hamiltonian. Agreement between absorption spectra obtained in cyclohexane at room temperature and the theoretical predictions is good. For 1, 2, and 3 the general features of the spectra are similar. The transition to S₁ (weak) is assigned as n → π* (A₂ ← A₁), to S₂ (strong) as π → π* (B₂ ← A₁), and to S₃ (moderate) as π → π* (A₁ ← A₁). The energy gap between S₁ and S₂ is seen to decrease as the length of the polyene chain increases in going from 1 to 3. Fluorescence is not observed for 1 in any of the solvents studied (protic and aprotic). Fluorescence is observed for 2 in protic solvents only. For 3, fluorescence is observed in a number of protic and aprotic solvents. Solvents which are able to induce fluorescence are believed to do so by inverting the order of ¹(nπ*) and ¹(ππ*) states. The influence of hydrogen bonding on the excitation spectra of 2 and 3 is discussed. Solvent-induced shifts in the absorption and fluorescence spectra of 3 in combination with the PM3 calculated ground-state dipole moment (2.8 D) are used to determine the excited-state dipole moment of 3 (6.4 D/protic solvents; 6.6 D/aprotic solvents). Fluorescence quantum yields in different solvents for 3 vary as the fluorescence maxima shift in these solvents, going through a maximum in the mid-frequency range. The variation in quantum yields with different solvents is primarily attributed to changes in the nonradiative rate of decay from S₁. Excitation, polarized excitation, and fluorescence spectra have been measured for 2 and 3 at 77 K in ethanol/ methanol glass. Vibronic features not observed in the broad spectra obtained in alcohols at room temperature become clearly resolved at 77 K. Evidence is provided that indicates that 2 and 3 undergo excited-state proton-transfer reactions in acetic acid at room temperature.

Full text of DOI:10.1021/jp0219597

7684
J. Phys. Chem. A 2003, 107, 7684-7691
Electronic Absorption and Fluorescence Properties of 2,5-Diarylidene-Cyclopentanones
Robert E. Connors* and Mine G. Ucak-Astarlioglu
Department of Chemistry and Biochemistry, Worcester Polytechnic Institute, Worcester, Massachusetts 01609
ReceiVed: August 27, 2002; In Final Form: May 21, 2003
Spectroscopic properties for a series of 2,5-diarylidene-cyclopentanones are reported. Electronic absorption
and fluorescence spectra have been measured for the all-E configurations of 2,5-dibenzylidene-cyclopentanone
(1), 2,5-bis-(3-phenyl-allylidene)-cyclopentanone (2), and 2,5-bis-(5-phenyl-penta-2,4-dienylidene)-cyclopen-
tanone (3). The absorption spectra have been assigned with the aid of INDO/S calculations. Molecular structures
used for the INDO/S calculations were computed with the PM3 Hamiltonian. Agreement between absorption
spectra obtained in cyclohexane at room temperature and the theoretical predictions is good. For 1, 2, and 3
the general features of the spectra are similar. The transition to S1 (weak) is assigned as n f π* (A2 r A1),
to S2 (strong) as π f π* (B2 r A1), and to S3 (moderate) as π f π* (A1 r A1). The energy gap between
S1 and S2 is seen to decrease as the length of the polyene chain increases in going from 1 to 3. Fluorescence
is not observed for 1 in any of the solvents studied (protic and aprotic). Fluorescence is observed for 2 in
protic solvents only. For 3, fluorescence is observed in a number of protic and aprotic solvents. Solvents
1
1
which are able to induce fluorescence are believed to do so by inverting the order of (nπ*) and (ππ*)
states. The influence of hydrogen bonding on the excitation spectra of 2 and 3 is discussed. Solvent-induced
shifts in the absorption and fluorescence spectra of 3 in combination with the PM3 calculated ground-state
dipole moment (2.8 D) are used to determine the excited-state dipole moment of 3 (6.4 D/protic solvents; 6.6
D/aprotic solvents). Fluorescence quantum yields in different solvents for 3 vary as the fluorescence maxima
shift in these solvents, going through a maximum in the mid-frequency range. The variation in quantum
yields with different solvents is primarily attributed to changes in the nonradiative rate of decay from S1.
Excitation, polarized excitation, and fluorescence spectra have been measured for 2 and 3 at 77 K in ethanol/
methanol glass. Vibronic features not observed in the broad spectra obtained in alcohols at room temperature
become clearly resolved at 77 K. Evidence is provided that indicates that 2 and 3 undergo excited-state
proton-transfer reactions in acetic acid at room temperature.
Introduction
Substituted and unsubstituted 2,5-diarylidene-cyclopentanones
(Figure 1), a class of diarylpolyene ketones, have received recent
attention for their potential applications in a variety of areas.
These include use as photosensitizers and photopolymer imag-
ers,1 solvent polarity probes,2 fluoroionophores,3 and nonlinear
optical materials.4 In addition, interesting photodimerization
reactions that have implications for crystal engineering have
been reported for crystalline 2,5-dibenzylidene-cyclopentanone
(R ) H, n ) 1, Figure 1).5 Despite the considerable interest in
applications related to their photoexcited-state properties, little
has been reported describing the electronic structure and
spectroscopy of these molecules.6 We have begun an investiga-
tion of the electronic structure, as well as the absorption and
luminescence properties, of molecules represented in Figure 1.
The focus of this initial report is on the all-E configurations of
the unsubstituted parent compounds (R ) H) with n ) 1, 2,
and 3: 2,5-dibenzylidene-cyclopentanone (1), 2,5-bis-(3-phenyl-
allylidene)-cyclopentanone (2), and 2,5-bis-(5-phenyl-penta-2,4-
dienylidene)-cyclopentanone (3) (see Figure 2). The combination
of aryl, carbonyl, and polyene chromophores presents an
interesting challenge for molecular orbital methods to provide
insight as to the nature of the spectral transitions. Any discussion
of the electronic structure and spectroscopy of molecules
Figure 1. Structure of 2,5-diarylidene-cyclopentanones.
Figure 2. Molecular structures of 1, 2, and 3.
containing polyene chains brings to mind the considerable body
of work regarding the relative location of the symmetry
forbidden 21Ag state (under C2h symmetry) and the strongly
* Corresponding author. Phone: (508) 831-5394. Fax: (508) 831-5933.
E-mail rconnors@wpi.edu.
10.1021/jp0219597 CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/30/2003

Properties of 2,5-Diarylidene-Cyclopentanones
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7685
TABLE 1: Calculated and Experimental Geometry of 1
bond lengths (Å)
bond order
PM3
PM3
expt.a
O1-C2
C2-C3
C3-C4
C3-C5
C5-C6
C6-C7
C7-C8
C8-C9
C9-C10
C10-C11
C11-C6
1.216
1.491
1.493
1.340
1.460
1.400
1.389
1.391
1.390
1.390
1.397
1.210
1.474
1.493
1.341
1.461
1.389
1.387
1.356
1.373
1.385
1.377
1.88
0.95
0.99
1.84
1.03
1.38
1.44
1.41
1.42
1.43
1.39
Figure 3. Reaction scheme for the synthesis of 5-phenyl-penta-2,4-
dienal.
allowed 11Bu state.7 Comparison with previous work on polyene
aldehydes and ketones will be given where pertinent throughout
the paper.
bond angles (°)
Experimental Section
PM3
expt.a
Compounds 1 and 2 were synthesized by the RuCl3 catalyzed
aldol cross condensation of cyclopentanone with benzaldehyde
and cinnamaldehyde, respectively.8 Compound 3 was prepared
by the base-catalyzed condensation of cyclopentanone with
5-phenyl-penta-2,4-dienal. The dienal was synthesized by two
carbon homologation of cinnamaldehyde using AgClO4 cata-
lyzed addition of the zirconocene complex derived from the
hydrozirconation of 1-ethoxyethyne (Aldrich) with bis(cyclo-
pentadienyl)zirconium chloride hydride (Strem) followed by
acidic hydrolysis, as shown in Figure 3.9 All final products were
purified by recrystallization or column chromatography, and
their purity was confirmed by HPLC.
Absorption spectra were measured with a Schimadzu UV2100U
spectrometer (2-nm band-pass). Fluorescence spectra were
obtained with a Perkin-Elmer LS 50B luminescence spectrom-
eter equipped with a R928 phototube. Fluorescence quantum
yields (φf), corrected for differences in refractive indices, were
determined by comparing corrected integrated fluorescence
spectra with that of coumarin 481 (Exciton), also known as
coumarin 35, in acetonitrile, with φf taken as 0.11.10 Absorbance
at the wavelength of excitation was kept low (<0.08). Degassing
of solutions had no measurable affect on the φf values.
Fluorescence spectra used to determine φf were corrected using
correction factors that were obtained by measuring the spectra
of compounds with known emission spectra.11 Other fluores-
cence spectra were uncorrected. All excitation spectra were
corrected for instrumental response. The band-pass was 5 nm
for both the excitation and emission monochromators. Low-
temperature fluorescence measurements were made with either
a quartz optical dewar (77 K) or an Air Products Displex 202
closed-cycle refrigeration system (10 K). Attempts to observe
phosphorescence out to 850 nm for 1, 2, and 3 in various
solvents at 77 K were unsuccessful.
Because there is limited X-ray crystallographic data available
for the structures discussed here, computed geometries were
used as input for our spectral calculations. Fully optimized
structures, as confirmed by finding all positive force constants
for the computed normal modes of vibration, were obtained with
the PM3 Hamiltonian12 using the MOPAC 2000 program.
Spectral properties were calculated for the optimized structures
using the INDO/S Hamiltonian13 with configuration interaction
as implemented in the MOS-F (V.4.2) program. Complete sets
of singly excited configurations (SCI) were employed in the
calculations for 1 and 2; 25 000 SCI configurations were
employed for 3. The calculations were also carried out using
49 SCI and 96 doubly excited configurations (DCI). It was found
that slightly better agreement between the calculated and
experimental band positions was obtained with the SCI calcula-
tions.
C1-C2-C3
126.1
109.0
127.8
126.6
118.3
120.2
120.2
119.9
120.2
120.2
119.2
125.6
108.8
130.6
131.1
118.4
121.4
120.1
119.8
120.1
121.4
117.2
C2-C3-C4
C4-C3-C5
C3-C5-C6
C5-C6-C7
C6-C7-C8
C7-C8-C9
C8-C9-C10
C9-C10-C11
C10-C11-C6
C11-C6-C7
dihedral angle (°)
PM3
35.5
expt.a
C3-C5-C6-C11
a From ref 5c.
3.0
Results and Discussion
PM3 Optimized Geometries. Initial structures for geometry
optimization were drawn with all carbon and oxygen atoms in,
or nearly in, the same plane. After minimization it was found
that 1 had adopted a structure with Cs symmetry and 2 and 3
had found minima of C2V symmetry. The major difference
between the Cs and C2V forms is the degree to which the phenyl
rings are rotated out-of-plane. PM3 predicts that the phenyl
groups of 1 are rotated 35° out-of-plane relative to the
cyclopentanone ring whereas 2 and 3 have nearly planar
structures. A crystallographic structure determination has been
reported for 1.5c The results of this study along with the PM3
calculated results are presented in Table 1. In general the
agreement is good except for the out-of-plane rotation angle of
the phenyl group which is 3.0° in the crystal phase. However,
crystal packing forces may be responsible for the deviation of
this experimental structural parameter from the calculated
value.14 Most of the computed structural parameters for 2 and
3 are similar to those of 1. All three compounds are predicted
to have a carbonyl group bond length of 1.216 Å. The single
bond connected to the phenyl group is predicted to have a bond
length of 1.460 Å for 1 and 1.458 Å for both 2 and 3. The
other exocyclic C-C single bond lengths fall in the range
1.448-1.449 Å for 2 and 3. The exocyclic double bond lengths
fall in the relatively narrow range of 1.340-1.344 Å for 1, 2,
and 3. Rotation of the phenyl rings out of the plane of the
cyclopentanone ring is predicted by PM3 to be less than 0.5°
for 2 and 3.
Absorption Spectra and INDO/S Calculations. The room-
temperature absorption spectra of 1, 2, and 3 in cyclohexane
are shown in Figures 4-6 along with the results of INDO/S-
CIS calculations. The spectra are similar in their gross appear-

7686 J. Phys. Chem. A, Vol. 107, No. 39, 2003
Connors and Ucak-Astarlioglu
solvent index of refraction, have been determined by measuring
the area under the strong transition. On the long wavelength
side of the strong transition for 1 a series of weak bands with
an average spacing of 1280 cm-1 is observed between 370 and
460 nm. This band system merges into the long wavelength
shoulder of the strong structured transition. For 2, less resolved
weak shoulders are observed between 430 and 490 nm on the
long wavelength side of the strong band system. These weak
long wavelength bands disappear for 1 and 2 when methanol is
used as the solvent. Although no clearly resolved bands are
observed for 3 on the long wavelength side of the strong
transition. It will be argued later that there is a separate
unresolved transition in this region that is masked by the tail of
the intense absorption. Turning to the short wavelength side of
the strong transition, moderate absorption is observed to fall
between 215 and 260 nm for 1, 230 and 290 nm for 2, and 260
and 320 nm for 3. Absorption is seen to rise again near the
200-nm cutoff for all three compounds.
Figure 4. Room-temperature absorption spectra of 1 in cyclohexane
and INDO/S-CIS calculated results. The diamond on the wavelength
axis indicates the computed location of the forbidden 1A2
transition.
r
1A1
INDO/S calculations provide insight as to the nature of the
long wavelength electronic transitions. Specific comments will
be restricted to excitations terminating at the lowest three excited
singlet states, S1-S3. For 1, 2, and 3, S1 is computed to be a
symmetry forbidden A2 r A1 transition (under C2V point group)
of the n f π* type, arising primarily from the orbital excitation
b2(n) f b1(π*), where b1(π*) is the LUMO and b2(n) is
HOMO-6 for 3 and HOMO-4 for 1 and 2. The weak bands
observed between 370 and 460 nm for 1 and between 430 and
490 nm for 2 are assigned to this transition. There, position,
intensity, and disappearance in alcohol solvent are entirely
consistent with n f π* behavior.15 Although not separately
resolved from the neighboring strong transition, it is believed
that S1 is also n f π* (A2 r A1) for 3 in cyclohexane.
Fluorescence data presented below support this assignment.
According to the INDO/S calculations, S2 is a strong, y-polarized
B2 r A1 transition corresponding primarily to the HOMO f
LUMO, a2(π) f b1(π*) configuration. Clearly, the strong
structured transition that dominates the spectra of the three
compounds can be assigned with confidence to this excitation.
This transition correlates with the 11Bu r 11Ag transition of an
idealized polyene under the C2h point group. The transition to
S3 is computed by INDO/S-SCI to be a weak, z-polarized A1
r A1 excitation with primary configurations HOMO-1 f
LUMO, a2(π) f a2(π*) and HOMO f LUMO + 1, b1(π) f
b1(π*). When the calculation is repeated to include doubly
excited configurations additional important configurations are
found to be HOMO-1, HOMO f LUMO, LUMO + 1, and
HOMO, HOMO f LUMO, LUMO. It appears that this state
correlates with the forbidden idealized polyene 21Ag r 11Ag
transition which is known to be S1 for longer polyene molecules.
It would be difficult to observe this relatively weak transition
directly in absorption because it is predicted to overlap the short
wavelength tail of the much stronger B2 r A1 transition for 1,
2, and 3. However, results from polarized fluorescence excitation
spectra (see below) are consistent with the INDO/S prediction
that there is a weak transition underlying the tail of the strong
band system. Table 2 summarizes experimental and computed
results for electronic transitions between the ground state and
the excited states S1-S3 of 1, 2, and 3.
Figure 5. Room-temperature absorption spectra of 2 in cyclohexane
and INDO/S-CIS calculated results. The diamond on the wavelength
axis indicates the computed location of the forbidden 1A2
transition.
r
1A1
Figure 6. Room-temperature absorption spectra of 3 in cyclohexane
and INDO/S-CIS calculated results. The diamond on the wavelength
axis indicates the computed location of the forbidden 1A2
transition.
r
1A1
ance. The dominant feature is a strong structured band found
between 270 and 370 nm for 1 that moves to longer wavelength
and becomes more intense as the polyene chain length in-
creases: 320-420 nm for 2, and 360-470 nm for 3. The low-
resolution spacing between the apparent 0-0 and 0-1 vibronic
bands of the strong transition falls in the range 1300-1440 cm-1
for these compounds. Oscillator strengths (f), uncorrected for
Fluorescence Properties. In examining the fluorescence
properties of these compounds it is found that each behaves
differently with respect to changes in solvent polarity and
hydrogen bonding strength. No fluorescence has been observed
for 1 in any of the solvents studied (nonpolar, polar aprotic, or
protic solvents). Fluorescence is observed for 2 in alcohols but

Properties of 2,5-Diarylidene-Cyclopentanones
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7687
TABLE 2: Absorption Spectral Data for 1, 2, and 3 in
Cyclohexane at Room Temperature and INDO/S Calculated
Results
1
2
3
S1 (A2)
λ (nm), exp
370-460
27.0-21.7
∼102
430-490
23.3-20.4
∼102
-
E (103 cm-1), exp
ꢀ (M-1 cm-1
)
-
λ
max (nm), calc
432
439
439
22.8
0.00
E (103 cm-1), calc
f, calc
23.1
22.8
0.00
0.00
S2 (B2)
λ
max (nm), exp
337
29.7
382
26.2
415
24.1
E (103cm-1), exp
ꢀmax (M-1 cm-1
f, exp
)
41.2 × 103
57.2 × 103
84.9 × 103
0.76
1.05
1.64
λ
max (nm), calc
317
372
383
E (103 cm-1), calc
f, calc
31.5
26.9
26.1
1.26
2.14
2.98
S3 (A1)
λ
max (nm), calc
284
35.2
0.12
321
31.2
0.12
331
30.2
0.09
E (103 cm-1), calc
f, calc
not in nonpolar or aprotic polar solvents. For 3, fluorescence is
observed in a variety of solvents including polar protic and
aprotic solvents and some aprotic nonpolar solvents. However,
no fluorescence has been observed for 3 in alkane solvents.
In situations where fluorescence has not been observed for
1, 2, or 3, it is believed that S1 is nπ* and S2 is ππ*. The absence
of measurable fluorescence when S1 is nπ* is attributed to
Figure 7. Normalized absorption (dashed line), excitation (left), and
fluorescence (right) spectra of 2 in alcohols: (a) 1-octanol; (b)
1-butanol; (c) 1-propanol; (d) ethanol; (e) methanol; (f) TFE; and (g)
HFIP. Excitation spectra were obtained by monitoring λmax of the
fluorescence.
1
efficient (nπ*)-3(ππ*) intersystem crossing which, as noted
by El-Sayed, is a direct consequence of the strong spin-orbit
coupling that exits between singlet and triplet states of different
orbital configurations.16 A solvent-induced inversion of state
order such that S1 becomes ππ* is believed to occur in those
situations where fluorescence is observed. The inversion of nπ*
and ππ* states when nπ* lies lower in energy in nonpolar
solvents is due to the well-known shift to higher energy
experienced by nπ* states and the shift to lower energy
experienced by ππ* states on going from nonpolar to polar
solvents. Hydrogen bonding solvents generally cause the largest
blue shift for nπ* transitions and the magnitude of the shift has
been correlated with the hydrogen bonding strength of the
solvent.17-19 As the INDO/S calculations indicate, and has been
found for short polyene ketones and aldehydes where nπ* is
lower in energy than ππ*, the energy of the nπ* state decreases
less rapidly than that of the ππ* state as the number of
conjugated double bonds increases.20,21 Thus, in environments
where S1 is nπ* for 1, 2, and 3, the S1(nπ*)-S2(ππ*) energy
gap is expected to decrease in going from 1 to 3. The gas-
phase INDO/S results show calculated S1(nπ*)-S2(ππ*) energy
gaps of 8400, 4100, and 3300 cm-1 for 1, 2, and 3, respectively.
Figures 7 and 8 present the normalized room-temperature
fluorescence, fluorescence excitation, and absorption spectra of
2 and 3 in the alcohol series 1-octanol, 1-butanol, 1-propanol,
ethanol, methanol, 2,2,2,-trifluoroethanol (TFE), and hexafluoro-
2-propanol (HFIP). Features to note are that the fluorescence
spectra red shift to a greater extent than do the absorption spectra
as the solvent pKa decreases. The excitation spectra and
absorption spectra for 2 do not overlap completely; whereas
there is reasonably good overlap of excitation and absorption
spectra for 3 in the series of alcohols. Changing the monitoring
wavelength throughout the fluorescence spectra of 2 and 3 does
not alter the appearance of the excitation spectra. Specifically,
it is found that the excitation spectra of 2 in alcohols are red
shifted relative to the absorption spectra, resulting in an apparent
dependence of φf on the S1 excitation wavelength (for example,
Figure 8. Normalized absorption (dashed line), excitation (left), and
fluorescence (right) spectra of 3 in alcohols: (a) 1-octanol; (b)
1-butanol; (c) 1-propanol; (d) ethanol; (e) methanol; (f) TFE; and (g)
HFIP. Excitation spectra were obtained by monitoring λmax of the
fluorescence.
see Figure 9). The effect is far less pronounced for 2 in the
strongly hydrogen bonding solvents TFE and HFIP. The results
for 2 in alcohols are strongly reminiscent of the 77 K results
reported by Becker22 for certain polyene ketones and aldehydes,
including retinals, in ether/isopentane/ethanol (EPA) and by
Christensen and co-workers for 2,4,6,8-decatetraenal in methanol/
ethanol at 10 K.19 The explanation provided in these previous
studies is applicable to the results presented here. Within the
alcohol solution a mixture of hydrogen bonded and non-
hydrogen bonded molecules of 2 exists. For the non-hydrogen
bonded members, nπ* remains below ππ* and does not

7688 J. Phys. Chem. A, Vol. 107, No. 39, 2003
Connors and Ucak-Astarlioglu
Figure 9. Absorption (dashed line), excitation (left), and fluorescence
(right) spectra and fluorescence quantum yield as a function of excitation
wavelength for 2 in methanol.
fluoresce. Hydrogen bonding inverts the order of these states
and induces fluorescence from S1 (ππ*). The energy of the ππ*
state of the hydrogen bonded (fluorescing) fraction of molecules
undergoes a characteristic red shift relative to that of the non-
hydrogen bonded (nonfluorescing) fraction; hence the lack of
agreement between the absorption spectrum and the excitation
spectrum. Presumably the failure of the nonfluoronated alcohols
to completely hydrogen bond with 2 is because much of the
solvent is tied up in self-association.19,23 As with 2,4,6,8-
decatetraenal, the extent of complex formation for 2 with the
stronger hydrogen-bonding solvents TFE and HFIP appears to
be complete as witnessed by the good agreement between
excitation and absorption spectra.19 Unlike decatetraenal, how-
ever, a strong enhancement in fluorescence intensity is not
observed for 2 in TFE when the temperature is lowered from
100 to 10 K.
Figure 10. Plot of (a) absorption frequency; (b) fluorescence frequency;
(c) Stokes shift against solvent orientation polarization. Open symbols
represent protic solvents; solid symbols are for aprotic solvents.
Contrary to the situation for 2, it is found that there is
generally good agreement between the excitation and absorption
spectra of 3 in alcohols. The conclusion here is that although
there is a distribution of hydrogen bonded and nonhydrogen
bonded molecules of 3, the nonhydrogen bonded molecules find
themselves in a polar solvent environment that is sufficient to
invert nπ* and ππ* and induce fluorescence. The observation
of fluorescence from 3 in aprotic polar solvents such as
acetonitrile shows that it is possible to achieve state inversion
in 3 without forming a hydrogen bonded complex with the
solvent.
Because 3 was found to fluoresce in a number of aprotic
solvents as well as protic solvents, a more complete study of
solvent effects on absorption and fluorescence spectra was
carried out. Results for 3 in protic and aprotic solvents are
plotted separately in Figure 10. Although not in a perfectly linear
manner, it is found that the absorption and fluorescence band
maxima shift to lower energy; whereas their difference, the
Stokes shift, moves to higher energy when plotted against ∆f,
the solvent’s orientation polarization. The orientation polariza-
tion is defined as ∆f ) (ꢀ - 1)/(2ꢀ + 1) - (n2 - 1)/(2n2 + 1)
where ꢀ is the dielectric constant and n is the index of refraction
for the solvent. The positive slope for the Stokes shift vs ∆f
plot is evidence that the dipole moment is larger for 3 in the
excited state than in the ground state.11 It is possible to calculate
the dipole moment of 3 in S1 from the solvent-induced shifts in
the absorption and fluorescence spectra by means of the ratio
method given in eq 124
Figure 11. Fluorescence quantum yield of 3 plotted against the
frequency of fluorescence in different solvents.
where µe and µg are the excited state and ground-state dipole
moments and ∆Ef and ∆Eg are the solvatochromic shifts for
fluorescence and absorption. The major advantage of using this
method over the Lippert-Mataga method25 is that it is not
necessary to assume a cavity radius for the solute. In employing
eq 1 to determine µe the underlying assumptions are that µe
and µg are collinear and that the cavity radius is about the same
in both states. Using the PM3 computed value for µg (2.79 D),
µe is calculated to be 6.4 D from the aprotic solvent data and
6.6 D from the protic solvent data. The gas phase INDO/S
calculation gives a slightly higher dipole moment of 7.2 D for
3 in the B2 state.
Fluorescence quantum yields have also been measured for 3
in different solvents and are plotted against the fluorescence
frequency maximum for each solvent in Figure 11. It is observed
that the quantum yield is approximately 10-3 when the
fluorescence maximum occurs above 19 000 cm-1 (nonpolar
solvents), increases sharply to 0.04 ( 0.01 between 19 000 and
17 000 cm-1 (polar solvents), and decreases gradually (0.03-
0.01) below 17 000 cm-1 (alcohols). The radiative rate constants
µe ) µg∆Ef/∆Ea
(1)

Properties of 2,5-Diarylidene-Cyclopentanones
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7689
TABLE 3: Absorption, Excitation, and Fluorescence Spectral Data and Fluorescence Quantum Yields for 2 in Alcohols
abs. λmax
(nm)
exc. λmax
(nm)
flu. λmax
(nm)
abs. νmax
(cm-1
exc. νmax
(cm-1
flu. νmax
(cm-1
λexcitation
for φf (nm)
solvent
)
)
)
pKa
φf
1-octanol
1-butanol
1-propanol
ethanol
methanol
TFE
412
412
409
410
408
420
432
431
428
430
420
416
430
434
505
510
512
515
522
542
558
24 272
24 272
24 450
24 390
24 510
23 810
23 148
23 202
23 364
23 256
23 810
24 038
23 256
23 041
19 802
19 608
19 531
19 417
19 517
18 450
17 921
0.020
0.018
0.020
0.027
0.026
0.035
0.013
409
409
409
405
400
409
409
16.2
15.9
15.5
12.4
9.3
HFIP
TABLE 4: Absorption and Fluorescence Spectral Data and Fluorescence Quantum Yields for 3
abs. λmax
flu. λmax
abs. νmax
(cm-1
flu. νmax
(cm-1
Stokes shift
λexcitation
solvent
(nm)
(nm)
)
)
(cm-1
3577
)
∆f
φf
for φf (nm)
carbon disulfide
cyclohexane
carbon tetrachloride
toluene
chloroform
ethyl ether
ethyl acetate
pyridine
dichloromethane
1-octanol
1-butanol
1-propanol
acetone
ethanol
442
415
425
430
450
415
425
448
447
450
450
445
435
445
434
477
451
464
525
22 624
24 067
23 529
23 256
22 222
24 096
23 529
22 321
22 371
22 222
22 222
22 472
22 989
22 472
23 041
20 964
22 173
21 552
19 048
0.0007
0.0002
0.0119
0.0131
0.1491
0.1669
0.1996
0.2124
0.2171
0.2263
0.2642
0.2746
0.2843
0.2887
0.3054
0.3092
0.3093
0.3159
0.0031
400
520
520
544
520
540
550
549
563
572
580
550
592
560
635
605
625
19 231
19 231
18 382
19 231
18 519
18 182
18 215
17 762
17 483
17 241
18 182
16 892
17 857
15 748
16 529
16 000
4299
4025
3840
4866
5011
4140
4156
4460
4740
5231
4807
5580
5184
5216
5644
5552
0.0008
0.0027
0.053
0.001
0.042
0.037
0.041
0.050
0.044
0.044
0.016
0.042
0.040
0.014
0.031
0.018
400
400
400
400
400
400
410
400
400
410
415
400
400
410
400
405
acetonitrile
HFIP
methanol
TFE
kf can be determined from a knowledge of φf and the measured
fluorescence lifetime τ, according to kf ) φf/τ. The rate of
nonradiative decay from S1 can be calculated from
where no fluorescence has been observed and where S1 is nπ*.
A blue shift of the nπ* state in polar or hydrogen bonding
solvents could shift it above S1(ππ*) and cut off this path for
3
intersystem crossing. An alternative mechanism for solvent-
dependent intersystem crossing that does not require nπ* to
3
knr ) (1/φf - 1)kf
(2)
be located below S1(ππ*) is second-order vibronic spin-orbit
coupling.29 Here, increasing the nπ*-ππ* energy gap with
solvent diminishes the degree of state mixing which in turn
reduces the rate of intersystem crossing.
In the absence of experimental values for τ, it is necessary to
rely on the method of Strickler and Berg26 to estimate kf from
the integrated absorption spectra. The radiative rates of decay
for 3 in CCl4 and methanol are calculated by the Strickler and
Berg method to be 5 × 108 and 4 × 108 s-1, respectively. From
these values of kf and the measured φf, the knr values are
calculated from eq 2 to be 6 × 1011 (CCl4) and 1 × 1010 s-1
(methanol) for 3. Limited as they may be,27 the data obtained
in these two solvents suggest that the order of magnitude
decrease in φf for 3 in solvents where emission occurs above
19 000 cm-1 compared to its value in those solvents where the
fluorescence occurs at lower frequencies is due primarily to a
larger value for knr. This is contrary to the trend expected from
the energy gap law for internal conversion which predicts that
the rate of internal conversion should increase as the energy
gap between S1 and S0 decreases.28 Explanations for the
enhanced rate of nonradiative decay for 3 in solvents where
emission occurs at frequencies above 19 000 cm-1 can be found
by considering that it is in these nonpolar solvents (CS2, CCl4,
toluene) that 3 is expected to have the smallest S1(ππ*)-S2-
(nπ*) energy gap. There are two pathways by which this small
energy gap could result in efficient nonradiative decay of S1.
One is through an enhancement in the rate of intersystem
crossing and the other is through an enhancement of the rate of
internal conversion. If 3nπ* lies between S1(ππ*) and T1(ππ*),
efficient intersystem crossing could account for the weak
fluorescence. Because of the dramatic difference in oscillator
strengths, kf for S1(ππ*) is expected to be much larger than
that for S1(nπ*). Hence, a small amount of fluorescence can be
observed for 3 in these solvents compared to alkane solvents
The enhanced internal conversion pathway involves the
“proximity effect” where vibronic interaction between closely
spaced nπ* and ππ* states is expected to lead to the distortion
and displacement of the potential energy surfaces of the coupled
states. Lim has argued through theory and experiment that such
distortion/displacement of S1 can dramatically increase the rate
of internal conversion.30 Any influence (solvent, substituents)
that increases the nπ*-ππ* energy gap between strongly
coupled states would reduce the vibronic coupling between them
and the associated distortion/displacement of potential energy
surfaces, thus decreasing the rate of internal conversion. For
many nitrogen heterocylic and aromatic carbonyl compounds
that do not fluoresce or fluoresce weakly in nonpolar aprotic
solvents, it has been observed that a switch to protic solvents
greatly increases φf. Studies indicate that for some of these
molecules the change in fluorescence behavior cannot be
attributed to a change in singlet-triplet intersystem crossing.31
The extent to which each of these pathways contribute to knr
when S1 of 3 is above 19 000 cm-1 awaits further investigation.
The rate that φf drops off with decreasing energy when S1 of
3 is below 17 000 cm-1 appears to be consistent with the energy
gap law for internal conversion which predicts an exponential
dependence of kic on ∆E, the S1-S0 energy gap20
kic ) Ce-R∆E
(3)

7690 J. Phys. Chem. A, Vol. 107, No. 39, 2003
Connors and Ucak-Astarlioglu
Figure 12. (a) Fluorescence, fluorescence excitation, and polarized
excitation spectra of 2 in ethanol/methanol (4:1) glass at 77 K. (b)
Fluorescence, fluorescence excitation, and polarized excitation spectra
of 3 in ethanol/methanol (4:1) glass at 77 K.
Figure 13. (a) Absorption (dashed line), excitation (left), and
fluorescence (right) spectra of 2 in acetic acid. (b) Absorption (dashed
line), excitation (left), and fluorescence (right) spectra of 3 in acetic
acid.
The room-temperature fluorescence properties of 2 and 3 are
summarized in Tables 3 and 4.
from the hydrogen bonded (unprotonated) ketone. A transfer
of electron charge density to the carbonyl oxygen atom upon
excitation from S0 to S1(ππ*) makes 2 and 3 stronger bases in
the excited state, thus promoting excited-state proton transfer
from acetic acid. A reduction in electron charge density on the
oxygen atom occurs when S1 is nπ* as appears to be the case
for 1 in acetic acid, making the molecule a weaker base in S1.
It is possible to protonate S0 of 1, 2, and 3 using strongly acidic
media.34
Fluorescence, fluorescence excitation, and polarized excitation
spectra were measured for 2 and 3 at 77 K in an ethanol/
methanol (4:1) glass and are shown in Figure 12. In comparing
the room-temperature spectra of 2 and 3 in ethanol with the 77
K spectra in ethanol/methanol, it is observed that the spectra
change from broad featureless bands to spectra exhibiting well-
resolved vibronic structure. It is also seen that the low
temperature absorption spectra are red shifted and the fluores-
cence spectra blue shifted relative to the room temperature
spectra, resulting in a much smaller Stokes shift. Small Stokes
shifts are commonly observed for molecules in low temperature
glasses and are attributed to the inability of the solvent matrix
to reorient and stabilize the molecule during the excited state
lifetime. The red shift of the absorption spectrum in going from
a room temperature liquid to a low temperature glass for a polar
molecule in a polar solvent is also frequently observed and may
be related to the thermochromic shift discussed by Suppan.32
The decrease in the degree of positive polarization observed
toward the short wavelength tail of the strong B2(y) r A1
transition for 2 and 3 may be due to the presence of a weak
underlying transition of perpendicular polarization. As noted
earlier, the INDO/S calculations do predict a weak A1(z) r A1
transition (S3) to occur in this region.
Excited-State Proton Transfer. The room-temperature
absorption, fluorescence, and fluorescence excitation spectra of
2 and 3 in acetic acid are presented in Figure 13. Fluorescence
is not observed for 1 in acetic acid. It is seen that the
fluorescence spectra of both 2 and 3 consist of two overlapping
bands (535, 615 nm for 2; 604, 685 nm for 3). The excitation
spectrum of each compound does not change when the monitor-
ing wavelength is moved from the shorter wavelength fluores-
cence band to the longer wavelength band. However, it is
observed that the longer wavelength fluorescence band decreases
in relative intensity and eventually disappears when the acid
solution is diluted gradually with water (pH increased). This,
along with additional evidence that will be presented in a
subsequent publication that deals with ground and excited-state
properties of the protonated cations of 1, 2, and 3, leads us to
assign the longer wavelength fluorescence band to emission from
protonated 2 and 3. The protonated cations are formed on the
excited-state surface and emit before returning to the ground
state.33 The shorter wavelength band corresponds to emission
References and Notes
(1) (a) Barnabus, M. V.; Liu, A.; Trifunac, A. D.; Krongauz, V. V.;
Chang, C. T. J. Phys. Chem. 1992, 96, 212. (b) Chambers, W. J.; Eaton, D.
F. J. Imaging Sci. 1986, 30, 230. (c) Baum, M. D.; Henry, C. P. U.S. Patent
3,652,275.
(2) (a) Pivovarenko, V. G.; Klueva, A. V.; Doroshenko, A. O.;
Demchenko, A. P. Chem. Phys. Lett. 2000, 325, 389. (b) Das, P. K.;
Pramanik, R.; Banerjee D.; Bagchi, S. Spectrochim. Acta A 2000, 56, 2763.
(3) Doroshenko, A. O.; Grigorovich, A. V.; Posokhov, E. A.; Pivo-
varenko, V. G.; Demchenko, A. P. Mol. Eng. 1999, 8, 199.
(4) (a) Kawamata, J.; Inoue, K.; Inabe, T. Bull. Chem. Soc. Jpn. 1998,
71, 2777. (b) Kawamata, J.; Inoue, K.; Inabe, T.; Kiguchi, M.; Kato, M.;
Taniguchi, Y. Chem. Phys. Lett. 1996, 249, 29. (b) Kaatz, P.; Shelton, D.
P. J. Chem. Phys. 1996, 105, 3918. (c) Kawamata, J.; Inoue, K.; Inabe, T.
Appl. Phys. Lett. 1995, 66, 3102.
(5) (a) Theocharis, C. R.; Alison, M. C.; Hopkin, S. E.; Jones, P.;
Perryman, A. C.; Usanga, F. Mol. Cryst. Liq. Cryst. 1998, 156, 85. (b)
Frey, H.; Behmann, G.; Kaupp, G. Chem. Ber. 1987, 120, 387. (c)
Theocharis, C. R.; Jones, W.; Thomas, J. M.; Montevalli, M.; Hursthouse,
B. J. Chem. Soc., Perkin Trans. 2 1984, 71. (d) Theocharis, C. R.; Thomas,
J. M.; Jones, W. Mol. Cryst. Liq. Cryst. 1983, 93, 53.
(6) (a) Issa, R. M.; Etaiw, S. H.; Issa, I. M.; El-Shafie, A. K. Acta
Chim. Acad. Sci. Hung., Tomus 1976, 89, 381. (b) Farrel, P. G.; Read, B.
A. Can. J. Chem. 1968, 46, 3685.
(7) (a) Hudson, B. S.; Kohler, B. E.; Shulten, K. In Excited States;
Lim, E. C., Ed.; Academic Press: New York, 1982; Vol. 6, pp 1-95. (b)
Christensen, R. L. In AdVances in Phtosynthesis: The Photochemistry of
Carotenoids; Frank, H. A., Young, A. J., Britton, G., Cogdell, R. J., Eds.;
Kluwer Academic Publishers: Dordrecht, The Netherlands, 1999; Vol. 8,
pp 137-159.
(8) Iranpoor, N.; Kazemi, F. Tetrahedron 1998, 54, 9475.
(9) Maeta, H.; Suzuki, K. Tetrahedron Lett. 1993, 34, 341.
(10) Rechthaler, K.; Ko¨hler Chem. Phys. Lett. 1994, 189, 99.
(11) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.;
Kluwer Academic/Plenum Publishers: New York, 1999.
(12) Stewart, J. J. P. J. Comput. Chem. 1989, 1, 209.
(13) Ridley, J.; Zerner, M. Theor. Chim. Acta 1973, 32, 111.
(14) Sanford, E. M.; Paulisse, K. W.; Reeves, J. T. J. Appl. Polym. Sci.
1999, 74, 2255.
(15) Baltrop, J.; Coyle, J. D. Principles of Photochemistry; Wiley &
Sons: New York, 1978; p 62.
(16) El-Sayed, M. A. J. Chem. Phys. 1963, 38, 2834.

Properties of 2,5-Diarylidene-Cyclopentanones
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7691
(17) Jaffe´, H. H.; Orchin, M. Theory and Applications of UltraViolet
Spectroscopy; John Wiley & Sons: New York, 1962; p 186.
(18) Brealey, G. J.; Kasha, M. J. Am. Chem. Soc. 1955, 77, 4462.
(19) Papanikolas, J.; Walker, G. C.; Shamamian, V. A.; Christensen,
R. L.; Baum, J. C. J. Am. Chem. Soc. 1990, 112, 1912.
(25) (a) Lippert, E. Z. Electrochem. 1957, 61, 962. (b) Mataga, N.; Kaifu,
Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1956, 29, 465.
(26) Strickler, S. J.; Berg, R. A. J. Chem. Phys. 1962, 37, 814.
(27) A limited supply of 3 prevented us from obtaining the extinction
coefficients in a wide variety of solvents.
(20) Klessinger, M.; Michl, J. Excited States and Photochemistry of
Organic Molecules; VCH Publishers: New York, 1995.
(21) Das, P. K.; Becker, R. S. J. Phys. Chem. 1978, 82, 2081.
(22) (a) Das, P. K.; Becker, R. S. J. Phys. Chem. 1978, 82, 2093. (b)
Takemura, T.; Das, P. K.; Hug, G.; Becker, R. S. J. Am. Chem. Soc. 1978,
100, 2626. (c) Takemura, T.; Das, P. K.; Hug, G.; Becker, R. S. J. Am.
Chem. Soc. 1976, 98, 7099.
(23) (a) Fletcher, A. N. J. Phys. Chem. 1972, 76, 2562. (b) Kuhn, L. P.
J. Am. Chem. Soc. 1952, 74, 2492.
(24) Suppan, P. Chem. Phys. Lett. 1983, 94, 272.
(28) Siebrand, W. J. Chem. Phys. 1967, 46, 440/2411.
(29) (a) Lim, E. C.; Yu, J. M. H. J. Chem. Phys. 1967, 47, 3270. (b)
Lim, E. C.; Yu, J. M. H. J. Chem. Phys. 1966, 45, 4742.
(30) Lim, E. C. J. Phys. Chem. 1986, 90, 6770.
(31) Madej, S. L.; Okajima, S.; Lim, E. C. J. Chem. Phys. 1976, 65,
1219.
(32) Suppan, P. Photochem. Photobiol., A 1990, 50, 293.
(33) Moomaw, W. R.; Anton, M. F. J. Phys. Chem. 1976, 80, 2243.
(34) Connors, R. E.; Ucak-Astarlioglu, M. G. Unpublished results.