Quantum yield switching of fluorescence by selectively bridging single and double bonds in chalcones: Involvement of two different types of conical intersections

Article abstract of DOI:10.1021/jp992878m

From the fluorescence properties of chalcones as a function of solvent polarity, and by the comparison to derivatives with donors and acceptors and with various selectively bridged bonds, it can be concluded that two emissive and two nonemissive states ar

Full text of DOI:10.1021/jp992878m

9626
J. Phys. Chem. A 1999, 103, 9626-9635
Quantum Yield Switching of Fluorescence by Selectively Bridging Single and Double Bonds
in Chalcones: Involvement of Two Different Types of Conical Intersections
Knut Rurack,†,‡ Marina L. Dekhtyar,§ Julia L. Bricks,§ Ute Resch-Genger,‡ and
Wolfgang Rettig*,†
Institut fu¨r Physikalische und Theoretische Chemie, Humboldt UniVersita¨t zu Berlin, Bunsenstr. 1,
D-10117 Berlin, Germany, Bundesanstalt fu¨r Materialforschung und -pru¨fung (BAM),
Richard-Willsta¨tter-Str. 11, D-12489 Berlin, Germany, and Institute of Organic Chemistry,
National Academy of Sciences of the Ukraine, Murmanskaya 5, 253660, KieV-94, Ukraine
ReceiVed: August 13, 1999; In Final Form: October 1, 1999
From the fluorescence properties of chalcones as a function of solvent polarity, and by the comparison to
derivatives with donors and acceptors and with various selectively bridged bonds, it can be concluded that
two emissive and two nonemissive states are needed to describe the fluorescence behavior. Three of these
states are connected with bond twisting and lead to species with high or low dipole moment, two of them
situated in the proximity of a conical intersection.
1. Introduction
SCHEME 1: Chemical Structures of the Chalcone
Derivatives Studied; 1-4, 1-Cl, 1-OMe, 2-Cl, and 2-OMe
Have Been Experimentally Investigated in This Work,
5-8 Were Described in refs 3 and 7
The photophysical properties of chalcones have been studied
by numerous researchers1-7 since this class of dyes has been
widely used for various optical applications including nonlinear
optical materials for second harmonic generation,8 photorefrac-
tive polymers,9 holographic recording technology,10 and fluo-
rescent probes for the sensing of metal ions11 or the microen-
vironment in micelles.6 Whereas most 4,4′-substituted chalcones
absorb in the UV spectral region and are weakly fluorescent or
nonfluorescent in solvents of small or large polarity,5 4′-
(dimethylamino)chalcone 1 (Scheme 1) shows a strongly
solvatochromic emission behavior and a reasonable fluorescence
quantum yield in polar aprotic solvents.6,7
The results obtained during our studies on related fluoroiono-
phores, i.e., monoaza crown ether substituted heteroaromatic
analog of 1,11,12 have led us to a more detailed investigation of
the excited state processes governing the emission behavior of
chalcones and including the derivatives 1-4 (Scheme 1).
Furthermore, the findings reported by Wang3 and Wang and
Wu7 on the spectroscopic properties of the (partly or fully)
bridged compounds 5-8 (Scheme 1) suggested that a simple
reaction model involving only one photochemical funnel does
not sufficiently describe the deactivation of photoexcited 4′-
(dimethylamino)chalcones. But especially for applicational
purposes such as metal ion or environmental sensing, it would
be most helpful to be able to answer the question, how can
molecular bridging achieve fluorescence enhancement in a given
chromophore?
The original idea of bridging was connected with the loose
bolt theory13 which states that flexible groups or substituents
behave like a loose bolt causing the preference of nonradiative
decay pathways to the ground state and therefore causing
fluorescence quenching. Some well-known examples of suc-
cessful fluorescence enhancement are the bridged rhodamine
10114 and the bridged coumarin 102.15 However, when other
compounds with similar large flexible substituents are compared,
differences in fluorescence efficiency of several orders of
magnitude can be encountered which cannot be explained by
the loose bolt theory. For example, a few cases have been
described where bridging of flexible groups results in the
opposite effect, namely lowering of the fluorescence quantum
yields.16-19 This initiated the search for an adiabatic photo-
chemical mechanism which can explain the fluorescence losses.
† Institut fu¨r Physikalische und Theoretische Chemie.
‡ Bundesanstalt fu¨r Materialforschung und -pru¨fung.
§ Institute of Organic Chemistry.
10.1021/jp992878m CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/12/1999

Quantum Yield Switching of Fluorescence in Chalcones
J. Phys. Chem. A, Vol. 103, No. 48, 1999 9627
In a comparison of differently bridged triphenylmethane,20
rhodamine,21 and other xanthene dyes,22,23 arguments were given
for a simultaneous involvement of a twisting relaxation and a
charge-transfer interaction. This mechanism could explain the
500-fold weaker fluorescence quantum yield for a rhodamine
derivative23 where the usual acceptor carboxylic ester was
replaced by an amino donor group, as well as the restoring of
the large fluorescence quantum yield through protonation which
causes an increase of the loose bolt character but eliminates
the strong donor group. Such an “anti-loose bolt behavior” has
also been observed in connection with molecular bridging, e.g.,
for a series of donor-acceptor-substituted stilbenes (D-A-
stilbenes),16,17 indicating that not the flexibility or rigidity alone
is the decisive factor but that further adiabatically formed
excited-state product species may be involved and can have
intrinsic fluorescence properties depending on the specific case
considered. This photochemical mechanism of fluorescence
quenching or enhancement is closely related to the so-called
twisted intramolecular charge transfer (TICT) states (for reviews
see ref 24). With this in mind, we extended the series of
selectively bridged chalcones of varying donor and acceptor
strength by the synthesis of 1-Cl, 1-OMe, 2, 2-Cl, 2-OMe, 3,
and 4 (Scheme 1).
The experimental part of this work comprises steady state
absorption and fluorescence measurements of 1, 1-Cl, 1-OMe,
2, 2-Cl, 2-OMe, 3, and 4, where the study of the first compound
was done for a better comparison with the data published so
far. The quantum chemical calculations were performed in order
to elucidate theoretically the role of excited state bond rotation
on the relative energies of ground and the different types of
excited states. The aim of this paper is to analyze the mechanism
of bond rotation and its influence on the photophysics of
chalcones.
sphere; Gigahertz-Optik) and for the spectral irradiance of the
excitation channel (calibrated silicon diode mounted at a sphere
port; Gigahertz-Optik). The fluorescence quantum yields were
calculated from six independent measurements according to eq
1 employing the fluorescence standards coumarin 102 (φf )
0.6 in ethanol) and coumarin 153 (φf ) 0.4 in ethanol)25 and
the uncertainties of the measurement were determined to (10%
(for φf > 0.02), (20% (for 0.02 > φf > 5 × 10-3), and (30%
(for 5 × 10-3 > φf), respectively.
As Iif(λ) ni2
φif ) φsf
(1)
s
2
A
i If(λ) ns
Here, the subscripts i and s denote the quantities of unknown
compound and standard, Ax is the absorption at the excitation
wavelength, Ixf (λ) is the integrated fluorescence intensity, and
nx is the refractive index of the solvent.
Trans-cis isomerization experiments were performed em-
ploying a unique laser impulse fluorometer described else-
where.26 The sample was excited with the second harmonic
output (LBO crystal) of a regenerative mode-locked argon ion
laser-pumped Ti:sapphire laser at a repetition rate of 82 MHz.
The fluorescence was collected at right angles (polarizer set at
54.7°; monochromator with spectral bandwidths of 8 nm) and
the fluorescence spectra were recorded with a modified time-
correlated single photon counting setup. The laser beam was
attenuated using a double prism attenuator from LTB and
excitation energies were in the nanowatt to milliwatt range
(average laser power). The excitation energies were checked
and adjusted with a calibrated Si diode (model 221 with 100:1
attenuator 2550, Graseby) and an optometer (model S370,
Graseby).
Syntheses. 1, 3-(4-(dimethylamino)phenyl)-1-phenylprop-2-
en-1-one,27 1-OMe,28 and 1-Cl29 were synthesized by reacting
the (substituted) acetophenone precursor with the corresponding
4-substituted benzaldehyde derivative following a procedure
described previously.30 For the synthesis of 2, 2-OMe, and 2-Cl,
a mixture of 4-diethylaminosalicylaldehyde (1 mmol), the
corresponding bromoacetophenone (2.6 mmol), potassium car-
bonate (0.137 g), and tetrabutylammonium hydrogen sulfate
(0.137 g) in 14 mL dimethylacetamide was heated for 4 h at
100-105 °C. After cooling, the resulting mixture was poured
into 75 mL water and extracted with toluene (3 × 75 mL). The
extract was dried over MgSO4 and evaporated. The final product
was purified twice by column chromatography on silica gel
(toluene/ethyl acetate 1:1 and cyclohexane/ethyl acetate 1:1) and
crystallized from hexane.
(6-Diethylaminobenzofuran-2-yl)phenylmethanone (2). Yield:
68%. Dark-yellow needles, mp 76-77 °C. Calculated for
C19H19NO2: C 77.77, H 6.53, N 4.78. Found: C 77.56, 77.48;
H 6.45, 6.47; N 4.64, 4.68. 1H NMR (CDCl3) δ (ppm): 1.200-
1.246 (t, J ) 7 Hz, 6H), 3.405-3.475 (q, J ) 7.1 Hz, 4H),
6.740-6.771 (dd, J1 ) 2.4 Hz, J2 ) 8.8 Hz, 1H), 6.769-6.777
(d, J ) 2.4 Hz, 1H), 7.381-7.384 (d, J ) 0.8 Hz, 1H), 7.456-
7.608 (m, 4H), 7.950-7.982 (m, 2H). 13C NMR (CDCl3) δ
(ppm): 12.479; 45.021; 92.805; 111.077; 116.263; 118.481;
123.659; 128.313; 129.093; 132.035; 138.201; 149.391; 150.244;
159.312; 183.397.
2. Experimental and Calculations
Materials. All solvents were of UV spectroscopic grade and
were purchased from Aldrich. For the syntheses, starting
materials from Lancaster and Aldrich were used.
Apparatus. The chemical structures of the synthesized
compounds were confirmed by elemental analysis, H NMR,
1
and 13C NMR. Their purity was checked by reversed phase
HPLC (HPLC set up from Merck-Hitachi; RP18 column;
acetonitrile/water ) 75:25 as eluent) employing UV detection
(UV detector from Knaur; fixed wavelength at 310 nm). Melting
points (uncorrected) were measured with a digital melting point
analyzer IA 9100 (Kleinfeld GmbH) and NMR spectra were
obtained with a 500 MHz NMR spectrometer Varian Unityplus
500.
Spectroscopy. The steady-state measurements were per-
formed on a SPECORD M400/M500 spectrophotometer from
Carl Zeiss Jena and a Spectronics Instruments 8100 spectro-
fluorometer. For the fluorescence experiments, only dilute
solutions with an optical density (OD) below 0.01 at the
excitation wavelength (OD < 0.04 at the absorption maximum)
to avoid reabsorption in the case of small Stokes shifts were
used. The relative fluorescence quantum yields (φf) were
determined by adjusting the optical densities of the solutions at
the excitation wavelengths to 0.1 ( 0.001 in a 100 mm
absorption cell. These solutions were then transferred to a 10
mm quartz cell, and the fluorescence measurements were
performed with a 90° standard geometry and an emission
polarizer set at 54.7°. All fluorescence spectra presented here
are corrected for the spectral response of the detection system
(calibrated quartz halogen lamp placed inside an integrating
(6-Diethylaminobenzofuran-2-yl)(4-chlorophenyl)metha-
none (2-Cl). Yield: 52%. Dark-yellow crystals, mp 98-100
°C. Calculated for C19H18ClNO2: C 69.61, H 5.53, N 4.27.
1
Found: C 69.54, 69.59; H 5.63, 5.70; N 4.33, 4.34. H NMR
(CDCl3) δ (ppm): 1.213-1.241 (t, J ) 7.0 Hz, 6H), 3.423-
3.465 (q, J ) 7.0 Hz, 4H), 6.759-6.773 (broad, 2H), 7.394 (s,

9628 J. Phys. Chem. A, Vol. 103, No. 48, 1999
Rurack et al.
1H), 7.470-7.487 and 7.935-7.952 (dd, J ) 8.4 Hz, 4H),
7.473-7.491 (d, J ) 9.0 Hz, 1H). 13C NMR (CDCl3) δ (ppm):
12.450; 45.038; 92.686; 111.249; 116.182; 118.404; 123.752;
128.647; 129.106; 129.769; 130.593; 136.392; 138.425; 149.535;
150.089; 159.312; 181.854.
(6-Diethylaminobenzofuran-2-yl)(4-methoxyphenyl)metha-
none (2-OMe). Yield: 51%. Yellow needles, mp 97-99 °C.
Calculated for C20H21NO3: C 74.28, H 6.55, N 4.33. Found:
1
C 74.11, 74.24; H 6.60, 6.67; N 4.35, 4.44. H NMR (CDCl3)
δ (ppm): 1.198-1.245 (t, J ) 7.0 Hz, 6H), 3.402-3.473 (q, J
) 7.1 Hz, 4H), 3.895 (s, 3H), 6.744 (broad, 1H), 6.778 (broad,
1H), 6.972-7.020 (m, 2H), 7.389 (s, 1H), 7.459-7.488 (d, J
) 8.7 Hz, 1H), 8.020-8.049 (d, J ) 8.8 Hz, 2H). 13C NMR
(CDCl3) δ (ppm): 12.479; 45.005; 55.447; 92.917; 110.949;
113.618; 116.295; 117.348; 123.474; 130.733; 131.456; 149.142;
150.613; 159.046; 162.937; 182.094.
3 was synthesized according to a method described in ref
31. For 4, a similar strategy as for 1 was employed and the
synthesis of 7-formyljulolidine was described in ref 32. A
solution of 0.18 g (1.5 mmol) acetophenone and 0.301 g (1.5
mmol) 7-formyljulolidine in 0.5 mL 2 N aqueous NaOH was
refluxed for 3 min and stirred at room temperature for 20 h.
Water (50 mL) was added and the mixture was extracted with
toluene (3 × 50 mL). The organic layer was dried (MgSO4)
and concentrated under reduced pressure to give a yellow oil
(0.38 g). Purification by column chromatography (silica gel,
toluene) yielded crystalline 4.
Figure 1. Normalized steady-state absorption and emission spectra
of 1, 2, 3, and 4 in n-hexane (solid squares), diethyl ether (open circles),
and acetonitrile (crosses) at room temperature.
1-Phenyl-3-(2,3,6,7-tetrahydro-1H,5H-benzo[i,j]quinolizin-7-
yl)-2-propen-1-one (4). Yield: 61.5%. Red needles, mp 123-
125 °C (from n-hexane). Calculated for C21H21NO: C 83.13,
H 6.98, N 4.62. Found: C 82.98, 83.02; H 7.05, 7.12; N 4.75,
4.78. 1H NMR (CDCl3) δ (ppm): 1.944-1.992 (m, 4H), 2.746-
2.772 (t, J ) 6.4 Hz, 4H), 3.236-3.259 (t, J ) 5.8 Hz, 4H),
7.110 (s, 2H), 7.244-7.274 and 7.689-7.719 (2 × d, J ) 15.3
Hz, 2H), 7.453-7.482 (m, 2H), 7.515-7.544 (t, J ) 7.3 Hz,
1H), 7.979-7.998 (d, J ) 7.9 Hz, 2H). 13C NMR (CDCl3) δ
(ppm): 21.380, 27.592, 50.156, 111.730, 116.852. 122.555,
128.214, 128.338, 130.310, 132.023, 138.992, 145.741, 151.940,
190.590.
Quantum Chemical Calculations. AM1 calculations33 using
the AMPAC package34 were performed to yield the ground state
and excited state energies for various conformations of 1-8.
In the course of the ground-state energy calculations for the
90°-twisted conformations of 1, the structures were fully
optimized except for a single dihedral angle fixed at 90° and
associated with the corresponding twisted bond. The energies
of the excited states were calculated for the rigidized ground
state geometry, with eight molecular orbitals involved in the
configuration interaction including 200 single and multiple
excitations. The effect of solvent stabilization for the ground
and the excited states was calculated similar as in ref 35, on
the basis of the Onsager model.36,37 In the determination of the
Onsager cavity parameter by the mass-density formula,38 the
density of 1 (with the molar mass of 257 g) was taken to be
equal to that of dimethylaniline (0.95 g cm-3). The dielectric
constant of acetonitrile was taken as 37.5.
excitation spectra match and no different behavior has been
reported for 5-8. The intramolecular charge transfer (ICT)
character of all 12 donor-acceptor-substituted chalcones (D-
A-chalcones) is characterized by an increase in Stokes shift
with increasing solvent polarity pointing to a stronger stabiliza-
tion of the excited state in polar solvents. Consequently, the
excitation of the ICT state involves a charge shift from the
dimethylanilino (DMA) donor to the carbonyl-phenyl (COPh)
acceptor fragment, and the dipole moment of the resultant
excited state (µes) is larger than that of the ground state (µes
>
µgs). Analyzing the solvatochromic data according to the Bilot-
Kawski formalism by a plot of the Stokes shift vs BK (eq 239),
Wang and Wu determined (µes - µgs) to be 5.9, 5.6, 8.2, and
5.8 D for 1, 5, 7, and 8, respectively.7 In another paper, Wang
found a value of (µes - µgs) ) 10 D for 5 on the basis of
correlating the absorption maxima in different solvents according
to a method proposed by Varma and Groenen.3,40 In the present
work, again the BK formalism taking into account the polar-
izability of the solute was employed, and the change in dipole
moment was calculated from the slope of a plot according to
eq 2.
2(µes - µgs)2
∆ν˜(abs-em) ) ∆ν˜vac(abs-em) +
BK (2)
3
hc0aO
with
BK )
f(ꢀr) - f(n)
ꢀr - 1
2ꢀr + 1
,
f(ꢀr) )
,
2
(
)
(1 - âf(n)) (1 - âf(ꢀr))
3. Results
n2 - 1
2n2 + 1
Absorption and Fluorescence Spectroscopy. Steady-State
Spectra and SolVatochromism. The absorption and fluorescence
spectra of 1-4 in solvents of different polarity are shown in
Figure 1, and selected spectroscopic data of 1-8, 1-Cl, 1-OMe,
2-Cl, and 2-OMe are included in Table 1. For 1-4, 1-Cl,
1-OMe, 2-Cl, and 2-OMe, the absorption and fluorescence
and f(n) )
In eq 2, h is Planck’s constant, c0 is the speed of light in a
vacuum, ꢀr and n are the dielectric constant and refractive index
of the solvent, and â ) 2R/aO3 is a polarizability term including

Quantum Yield Switching of Fluorescence in Chalcones
J. Phys. Chem. A, Vol. 103, No. 48, 1999 9629
TABLE 1: Spectroscopic Data of 1-8, 1-Cl, 1-OMe, 2-Cl, and 2-OMe in Selected Solvents (Second Maxima Are Given in
Bracketsa)
n-hexane
diethyl ether
acetonitrile
ν˜ (abs)
ν˜ (em)
ν˜ (abs)
ν˜ (em)
ν˜ (abs)
ν˜ (em)
103 cm-1
103 cm-1
φf
103 cm-1
103 cm-1
φf
0.02
103 cm-1
103 cm-1
φf
0.15
1
2
3
26.4 (25.1)
26.1 (24.9)
26.7 (25.2)
24.9 (23.6)
25.6c
26.0
28.6
22.8 (23.9)
23.3 (21.9)
21.9 (23.4)
22.6 (21.3)
24.1
22.8
n.r.
1 × 10-4
0.046
25.3
25.1
25.6
23.8
24.7d
n.r.f
n.r.
n.r.
24.8
25.7
24.6
25.3
20.6
20.3
20.9
19.3
21.8
20.5
23.9
22.4
20.0
20.9
19.7
20.5
24.4
24.5
24.7
22.7
24.0
24.5
26.8
25.2
24.0
24.7
24.1
24.6
18.4
17.6
17.4
17.3
18.7e
18.2
21.1
20.4
17.8
19.0
16.9
18.3
0.051
2 × 10-3
0.14
1 × 10-3
0.027
0.015
0.045e
5 × 10-3
1
7 × 10-5
3 × 10-3
1 × 10-4
0.011
4
5b
7 × 10-3
0.024
0.87
6b
7e
7 × 10-3
0.50
8e
n.r.
n.r.
1
0.86
1-Cl
1-OMe
2-Cl
2-OMe
26.0 (24.6)
26.7 (25.4)
25.3 (24.1)
26.1 (25.0)
23.3 (21.8)
22.2 (23.9)
22.8 (21.5)
23.8 (22.5)
4 × 10-4
2 × 10-4
0.11
0.059
0.013
0.036
0.18
0.094
0.18
2 × 10-3
0.02
0.013
a Low-energy absorption band fitted to a progression of j ) 4 gaussian bands. b Taken from ref 3. c Mean value of two maxima at 26.2 and 25.1
(spectra not shown), taken from ref 3. d Mean value of two maxima at 25.1 and 24.4 (spectra not shown), taken from ref 3. e Taken from ref 7. f Not
reported.
TABLE 2: Dipole Moments of 1-4, 1-Cl, 1-OMe, 2-Cl, and
2-OMe Obtained from the Slope of Solvatochromic Plots
According to Eq 2
aOa Å
µgsb D
slope
(µes-µgs) D
1
2
3
4
5.7
5.7
5.7
5.7
5.9
6.6
5.9
6.6
3.9
4.7
3.9
4.6
4.1
2.7
4.8
3.7
3160
4770
3840
3810
3740
2750
5180
4450
7.6
9.4
8.4
8.4
8.7
8.9
10.3
11.3
1-Cl
1-OMe
2-Cl
2-OMe
a Determined on the basis of the AM1 optimized ground state
geometry according to a method proposed by Lippert68 for elongated
molecules. b Calculated for optimized ground state geometry by AM1.
the solute polarizability R and the Onsager cavity radius aO (R/
aO3 = 0.5 for the isotropic polarizability). The values of (µes
-
µgs) obtained for the chalcones studied in this work are collected
in Table 2. Apart from the crucial choice of the correct Onsager
cavity radius, the different values agree well for the molecules
1, 2, and 3 and give further support to the fact that selective
bridging of different bonds in the D-A-chalcones influences
the polar CT character of the emissive state(s) only to a minor
extent. The increase in acceptor strength from 1-OMe via 1 to
1-Cl and 2-OMe via 2 to 2-Cl is well documented by an
increase in the slope of the corresponding solvatochromic plots
(Table 2). The dependence of the spectral band positions on
the strength of the ICT interaction is further exemplified by a
good correlation of the absorption band maximum (cf. Table
1) and the Hammett constant σp (Cl, 0.23; H, 0.00; OCH3,
-0.2741) in a polar solvent such as, for instance, in acetonitrile.
The constant σp is a measure for the electronic effect (of both
inductive and resonance components) of the substituent in the
para position.41
Fluorescence Quantum Yields and SolVatokinetics. In contrast,
for all the compounds, the fluorescence quantum yield (φf)
strongly depends on all three parameters, i.e., the bridging
pattern, donor and/or acceptor strength, and the solvent polarity.
Plots of φf vs ET(N), Dimroth’s empirical solvent polarity scale,42
are depicted in Figures 2 and 3. The lines in these figures,
although they are only guides to the eye, indicate an opposing
behavior concerning the solvent polarity dependence of the
fluorescence quantum yields for the different compounds. A
reduction in solvent polarity strongly enhances the fluorescence
quantum yield of 2 and 2-Cl but leads to a loss in fluorescence
intensity for 1-OMe, 3, and 5. A maximum in such a plot is
Figure 2. Plots of φf vs ET(N) for the D-A-chalcones investigated.
From top to bottom: 1 (solid squares), 5 (open diamonds), and 6 (solid
triangles); 1 (solid squares) and 3 (crosses); 1 (solid squares), 7 (solid
circles), and 8 (stars). 1 is included in all the plots for a better
comparison. The solid lines are only guides to the eye, and data in
protic solvents are omitted. Data for 5-8 were taken from refs 3 and
7.
observed for 1-Cl, 2-OMe, 4, 6, and, to a lesser extent, 1. For
the highly bridged compounds 7 and 8, no pronounced polarity
dependence is observed and fluorescence quenching occurs only
in n-hexane. Interestingly, the solvent polarity dependence of
1-6, 1-Cl, 1-OMe, 2-Cl, and 2-OMe can be categorized into
a so-called “positive solvatokinetic behavior” with a decrease
in emission yield with increasing solvent polarity resulting from
enhanced excited state population of a highly polar charge
transfer species with strongly nonradiative properties43,44 or a
“negative solvatokinetic behavior”, i.e., an increase in fluores-
cence quantum yield with increasing solvent polarity indicative
of population of an emissive CT state. 1, 3, and 5 show a
strongly resembling negative solvatokinetic behavior (Figure 2)
with relative fluorescence quantum yields decreasing in the order
1 > 5 > 3 (Table 1, Figure 2). Accordingly, the rigid bridging
of both single bonds in the acceptor part as in 5 does not yield
any fluorescence enhancement but a fluorescence loss. On the
other hand, tuning of the donor or acceptor strength of a

9630 J. Phys. Chem. A, Vol. 103, No. 48, 1999
Rurack et al.
SCHEME 2: Chemical Structures of Possible Chalcone
Conformersa
a The bonds affected by a 180° twist are labeled according to formula
1 in Scheme 1. In the top row, the two stable conformers of the double
bond b67 E isomer and in the bottom row, the sterically unfavorable Z
conformers are shown. The conformation of bonds b67 and b78 is
indicated and the labels on the arrows denote the bonds which have to
rotate for the respective transition.
Figure 3. Plots of φf vs ET(N) for the D-A-chalcones investigated.
(Top) 1 (solid squares), 1-Cl (crosses), 1-OMe (open triangles), and 4
(open circles); (bottom) 1 (solid squares), 2 (open triangles), 2-Cl (solid
triangles), and 2-OMe (open squares). 1 is included in both plots for
a better comparison. The solid lines are only guides to the eye and
data in protic solvents are omitted.
N(tE)
) e-[E(tE) - E(cE)]/kT
(3)
molecule with a given bridging pattern by further substituents
(1 or 2) can affect the solvent polarity-dependent fluorescence
yield as well and enhance (2-Cl as compared to 2 in medium
polar solvents) or reduce (2-OMe as compared to 2 in medium
polar solvents) fluorescence losses.
N(cE)
However, the question if only one conformation is populated
in the ground state or if both conformers are already populated
in S0 and show identical spectra remains unanswered at present.
The corresponding ground state dipole moments (∆µ(tE-cE)
) +0.7 D for 1 and -0.6 D for 2) and transition energies
(∆ES0-S1(tE-cE) ) 320 cm-1 for 1 and 80 cm-1 for 2) are of
comparable magnitude as well. Accordingly, the same is true
for the dipole moments and excitation energies of the fixed cE
(3, 5, 6) and tE (7, 8) conformers and is supported by the
experimental data collected in Table 1. Geometry optimizations
for the tZ and cZ conformers (possible in 1, 3, 5, and 6) did not
converge, and these conformers are found to be highly energeti-
cally unfavorable. With fixed Z geometries at bond b67, high
energy differences of >35 kJ mol-1 with respect to the most
stable geometry are found. For instance, these differences
calculated for the cZ and the tZ conformers of 1 amount to >300
kJ mol-1 and >900 kJ mol-1, explaining the lack of an
experimental detection of Z conformers. Thus, besides the fixed
E conformers, for both 1 and 2 cE is the slightly more stable
ground state conformer, but for tE of 1 the higher ground state
dipole moment is calculated. If bond twisting is restricted by
bridging, the torsional angles (and bond lengths) obtained for
optimized ground state structures of cE and tE are very similar
for all the D-A-chalcone derivatives in most cases (for some
examples, see Table 3).
Energy barriers for twisting in the ground state were
calculated by optimizing the 90° twisted structure of the bond
concerned (“perp-structure”). The barrier thus calculated for
twisting of the aryl acceptor moiety is generally low for 1 and
2 (ca. 15 kJ mol-1). Rotation around b67 in the ground state is
connected with the usual high activation barrier for CdC double
bonds. These data should be viewed in conjunction with
experimental evidence: From X-ray studies of numerous
chalcones, it is known that either cE or tE conformers are
preferred in certain molecules but most of the chalcones with a
simple COPh acceptor were found to crystallize in cE.46
Trans-Cis Isomerization. The significant occurrence of
permanent trans-cis isomerization around the double bond was
excluded by recording the emission spectrum of 1 as a function
of laser excitation energy up to 10 mW (cw). Furthermore,
HPLC analysis with UV/vis diode array detection of an
irradiated sample of 1 showed only one peak, its absorption
spectrum being identical to that of a reference sample kept in
the dark. The same was found for all the freshly synthesized
compounds 1-4. Chalcone cis isomers described in the literature
so far show strongly hypsochromically shifted absorption spectra
and population of the short-lived species was only achieved
photochemically.1
Quantum Chemical Calculations. Single and Double Bond
Isomerism and the Ground State Surface. When dealing with
chalcone photophysics or photochemistry, one has to be aware
of a complex equilibrium involving different conformers for
both isomers possibly populated already in the ground state.
As depicted in Scheme 2, besides trans-cis isomerism at the
central double bond b67, s-cis (“c”) and s-trans (“t”) single bond
isomerism (bond b78) can occur (to avoid “cis”, “trans”, as well
as “double bond isomer” and “s-isomer” misunderstandings, the
double bond isomers are labeled E and Z and the single bond
isomers c and t, arriving at cE, tE, cZ, and tZ conformers, as
depicted in Scheme 2. Ground-state geometry optimizations for
both E conformers converge in the case of the b78 unbridged
compounds (for labeling see Scheme 1) yielding comparably
small energy differences for both pairs of tE and cE conformers,
∆E(tE-cE) ) 6 kJ mol-1 for 1 and 2 kJ mol-1 for 2,
respectively. Following the Boltzmann description for the ratio
of the population of two states (eq 3), values of N(tE)/N(cE) )
0.09 (for 1) and 0.45 (for 3) are obtained for T ) 298 K.

Quantum Yield Switching of Fluorescence in Chalcones
J. Phys. Chem. A, Vol. 103, No. 48, 1999 9631
TABLE 3: Molecular Geometrical Parameters (Bond
Length, l, and Dihedral Angle, æ) of the Stable
Ground-State Structures of 1, 2, 3, and 5 as Calculated by
AM1-Optimization (Bond Lengths of the tE Conformers Are
Very Similar to Those of cE and Are Omitted for Clarity)
cE
tE
l/Å
æ/deg
æ/deg
b12 b56 b67 b78 b89 b12 b56 b78 b89 b12 b56 b78 b89
1a 1.41 1.45 1.34 1.47 1.48 20 -10 -10 -33 20 -173 149 -33
2a 1.40 1.45 1.38 1.46 1.48 17
0
-9 -31 17
0 161 -35
3
5
1.40 1.45 1.35 1.48 1.48 16 53 -19 -4
1.41 1.45 1.34 1.50 1.48 20 28 -2
6
a The bond lengths and dihedral angles of the corresponding
conformers of 1-Cl 1-OMe, 2-Cl, and 2-OMe are very similar.
Figure 4. Ground-state and excited-state energies for 1 for the planar
conformation and for various conformations with the bond indicated
twisted to 90°. For numbering, see Scheme 1. Ground-state energy (solid
line and diamonds) corresponds to full optimization. The lowest
biradicaloid (BR) excited states are calculated for the same geometry
(Franck-Condon geometries) and are labeled BR-singlet (solid line,
open squares) and BR-triplet (dashed line, crosses). For comparison,
the lowest locally excited state (dotted line, solid triangles) is also
included.
Moreover, employing IR spectroscopy, Hayes and Timmons
showed that the unsubstituted chalcone (1,3-diphenyl-prop-2-
en-1-one) exists in a mixture of ca. 15% tE and 85% cE
conformers in chloroform solution.47 In 1H NMR investigations
of 1 in CDCl3, a coupling of o-phenyl protons was only observed
for H-R (CH adjacent to CO group), indicating a more stable
cE conformation also for the donor-substituted derivative.48
Thus, for a closer inspection of the photochemical behavior of
the chalcones, only the cE conformer of 1 was treated theoreti-
cally and will be described in more detail below.
The Effect of Bond Twisting on the Photophysics: The
“Photochemical Spectrum” of 1. Bond rotations in the excited
state are of major importance to understand the fluorescence
properties of these compounds.16,23,43 The ground state cE
structure of unbridged 1 is not planar but is distinguished by
some twist around all single bonds (see Table 3). For a given
perp-structure, the twist angles of the bonds not directly adjacent
to the 90°-twisted bond are not much different from those of
the planar structure, except for the b89 perp-structure which
allows the rest of the molecular skeleton to planarize in the
vicinity of bonds b56, b67, and b78. Likewise, bond lengths are
retained much the same in all perp-forms with the single bonds
twisted. In contrast, the twisting around the double bond (b67)
causes a substantial lengthening of this bond (up to 1.41 Å)
and a shortening of the adjacent single bonds down to 1.43 Å
for b56 and 1.45 Å for b78.
The energetics of the ground and excited states can be
summarized in a “photochemical spectrum” (Figure 4) showing
the raising and lowering of energies upon twisting successive
flexible bonds in the molecule. In every case, the nature of the
corresponding excited state is analyzed and the lowest states of
locally excited (LE) and of biradicaloid (BR) character are
reported. LE states involve transitions between orbitals located
on the same fragment, while BR states involve orbitals localized
on different sides of the twisted bond. BR states are often viewed
as possible candidates for enhanced nonradiative deactivation
because they can be located very close to a conical intersection
between S1 and S0.49 Here, we take a significant reduction in
the S1-S0 energy for the Franck-Condon (FC) conformation
as an indication that a conical intersection (photochemical
funnel) can be reached easily through relaxation of further
coordinates.50-52
As seen from the photochemical spectrum of 1 (Figure 4),
the calculations predict an efficient “funnel” for the nonradiative
degradation of the excited-state energy only for twisted double
bond b67, with a strong reduction of the energy gap between
the ground and the lowest BR excited state down to 2.47 eV in
the gas phase and expected to reduce further to 2.35 eV in a
polar solvent like acetonitrile. The BR states of the other perp-
structures, associated with single bond twistings, are rather high
lying in energy, especially that arising from the b89 twist, and
are not likely candidates for nonradiative deactivation (Table
4). Noteworthy is also that the BR state involving the twisted
double bond (b67) perp-structure possesses a considerable
singlet-triplet splitting which is attributable to some delocal-
ization of π-orbitals over both moieties despite the twisted
structure.
Table 4 and Figure 5 demonstrate that solvation effects should
cause no substantial changes in the photochemical behavior of
1, because the bond which might be most strongly responsible
for a possible twist-induced funneling situation, bond b67, is
characterized by the smallest difference in the dipole moments
between the ground and the lowest BR state (∆µBR-S0 ) 2.6
D).
4. Discussion
In an interpretation of the emission behavior of the D-A-
chalcones, possible deactivation routes of the initially excited
singlet state including rotations around certain bonds, the
electronic nature of the states populated and their dipole
moment, and isomerization reactions need to be considered.
Thus, to gain mechanistic access, we will leave aside the
possible occurrence of different E conformers at first and ask
the question, which of the various possible twisting coordinates
will most strongly lead to a nonradiative coupling of ground
and excited state. This can, for example, occur if thermally
available excited state conformations are populated which
possess a very narrow S0-S1 energy gap or are otherwise located
in the neighborhood of a conical intersection. A further
possibility for a strong fluorescence quenching is the coupling
of S1 and T1 through an enhancement of intersystem crossing
(ISC) for these active conformations. A well-known case for
such an enhancement of nonradiative effects is connected with
the proximity of nπ* and ππ* states, especially when the nπ*
state is the lowest singlet excited state.53
Solvatokinetic effects can help in learning something about
the adiabatic photochemical reaction mechanism involved. These
effects indicate how the fluorescence quenching reaction
depends on solvent polarity and can therefore give an indication
as to the polar properties of the species which quenches the
fluorescence. In the case of the D-A-stilbenes and related

9632 J. Phys. Chem. A, Vol. 103, No. 48, 1999
Rurack et al.
TABLE 4: Energies of the Ground and Excited States of 1 in Vacuum and the Polar Solvent Acetonitrile
vacuum
solventa
ES0
eV
ECT
eV
ELE
eV
∆ELE-CT
eV
µS0
D
µCT
D
µLE
D
∆µCT-S0
ES0
eV
ECT
eV
ELE
eV
structure
D
planar
b12
b56
b67
b78
0
3.86 (S1)
5.04 (S4)
5.84 (S5)
4.32 (S1)
6.24 (S1)
4.42 (S2)
4.64 (S1)
4.45 (S1)
5.59 (S2)
4.34 (S1)
4.16 (S1)
-0.56
0.4
1.39
-1.27
1.9
4.42
2.99
4.34
5.40
3.85
4.81
11.77
13.76
14.75
8.00
7.10
5.02
8.42
11.18
8.42
(13.54)
2.68
10.77
10.41
2.60
-0.07
0.25
0.03
1.74
0.04
3.36
4.35
5.05
4.09
4.99
0
4.24
4.55
4.19
5.14
4.08
3.49
0.28
0.10
1.85
0.09
0.05
18.56
14.71
b89
-0.03
a Acetonitrile; ꢀ ) 37.5; r ) 4.71 Å; F ) 0.95 g cm-3; ꢀ0 ≈ 0.1 (ESU cm-1).
pounds 1 and 4 as well as 5 and 6, at least for low polarities.
Accordingly, the differences in the maximum of the φf vs ET-
(N) plot for 1 (5) and 4 (6) (Figures 2 and 3) point to a stronger
charge transfer in the E* state in the latter compound with the
julolidyl group as a stronger donor. A similar explanation has
been given by Wang, his data being partly included in Table 1.
Results published recently by Fery-Forgues et al. on benzox-
azinone derivatives56 (Scheme 3) and by Ko¨hler et al. on D-A-
stilbenes32 (Scheme 3) point into the same direction of a stronger
donor character associated with the julolidyl group. The effect
of varying acceptor strength on the energetics of the excited
states involved will be discussed in more detail below in
combination with the applicability of the multiple state model
to 1-Cl 1-OMe, 2-Cl, and 2-OMe.
Considering the anilino bond (b56), the formation of a highly
polar excited state (A*) can occur via twisting around b56. This
is possible for all compounds except for 2 and the fully bridged
8. The highly fluorescent 7 demonstrates that any nonradiative
deactivation via a possible b56 twisting is weak to negligible.
The weak fluorescence property of 2 may therefore be related
to the impossibility of populating such a fluorescent A* state.
The positive solvatokinetic behavior of fluorescence quenching
in 2 indicates the involvement of a quenching state of higher
polarity than the precursor state E*, but it must be different
from the emissive A* state (around b56). This nonradiative state
may be connected with twisting of the bond b78 connecting the
heterocyclic system with the keto group. Because of its
connection with CO twisting, we will call this highly polar
quenching state K*, to distinguish it from the highly polar but
luminescent A* state reachable by twisting the anilino bond
b56.
Comparing the solvent-dependent fluorescence quantum yield
data of 1-8, a negative solvatokinetic behavior is only observed
for 1, 3, and 5 as well as (partly) 4 and 6 with an unbridged
anilino bond, suggesting the involvement of such an A* state.
Bridging of the C-C single bonds b78 and b89 as in 3 and 5
does not strongly alter the shape of the φf vs ET(N) plot (Figure
2). The curve has a similar shape but is of lower amplitude for
5 and especially for 3 as compared to 1.57 Accordingly, for these
single bonds, two contradictory observations are made: (i)
bridging the bonds adjacent to the CO group leads to increased
fluorescence quenching (3 and 5 as compared to 1) but (ii) the
lack of this bridge and the introduction of a different bridging
pattern can open an efficient nonradiative funnel as well (in
weakly fluorescent 2). Thus, C-C single bond-twisted states
(K*) in keto derivatives such as 2 seem to be especially weakly
emissive and the larger fluorescence quantum yield of 1 can be
explained as being due to the fact that an emissive A* state
competes successfully with the transition to the quenching K*
state. Very efficient K* state formation is known to be
responsible for the weak emission of compounds related to 2
such as Michler’s ketone59 and DMABK60 (Scheme 3). Since
related D-A-stilbenes containing a central double bond fixed
Figure 5. Energy stabilization of the ground (diamonds), TICT excited
(squares), and locally excited (triangles) states. Solid and dotted lines
represent state energies in the gas phase and in acetonitrile solution.
dyes,17 e.g., a negative solvatokinetic effect was observed with
quenching rates decreasing as solvent polarity increases, indicat-
ing a weakly polar biradicaloid state P* as the quenching state.54
In the case of the dual fluorescing TICT compounds (see refs
24,43), both precursor and product state of the reaction can be
observed in fluorescence, and the relative changes in fluores-
cence quantum yield can directly be related to a positive
solvatokinetic behavior with the photochemical reaction from
a less polar precursor to the highly polar TICT product being
accelerated with solvent polarity.24,55 In the chalcone case,
considering at first only the compounds with similar donor-
acceptor substitution pattern, we observe a positive solvatoki-
netic effect for the quenching reaction in 2 (reduction in solvent
polarity enhances the fluorescence), but a negative solvatokinetic
effect (reduction in solvent polarity quenches fluorescence) is
observed for 1, 3, and 5 (with the exception of the region of
high solvent polarity, i.e., acetonitrile for the former derivative).
This directly leads to the conclusion that different quenching
mechanisms are active in these two cases.
Multiple State Model. In analogy to the D-A-stilbenes,17
we can gain access to an understanding of the photophysical
behavior by using a mechanistic model which involves the all-
planar excited conformation E*, the CC double bond-twisted
species, with biradicaloid properties and a reduced dipole
moment (P* state), and one or several C-C single bond-twisted
species of highly polar nature, designated as A* state and related
to the TICT mechanism.
The opposite tendencies in solvatokinetic behavior of 1, 3,
and 5 (negative) on one hand and of 2 (positive) on the other
hand directly suggest that a weakly polar, nonemissive P* state
enhances the nonradiative losses in the former three compounds
upon decreasing solvent polarity. Moreover, a major influence
of a dimethylamino bond (b12) twisting on the excited state
deactivation of these chalcones can be directly excluded on the
basis that qualitatively similar effects are observed for com-

Quantum Yield Switching of Fluorescence in Chalcones
J. Phys. Chem. A, Vol. 103, No. 48, 1999 9633
SCHEME 3: Chemical Structures of the Related
Benzoxazinone,56 D-A-Stilbene,32 and Benzofuran61
Derivatives, Michler’s Ketone,59 DMABK,60 and
DMABI64
Figure 6. Normalized absorption and relative emission spectra
(normalized to the same optical density at the excitation wavelength)
of D-A-chalcones carrying different acceptors. Top: spectra of
1-OMe (solid squares), 1 (open circles), and 1-Cl (crosses) in
acetonitrile; the spectra of 4 (solid triangles) in the same solvent are
included for comparison. Bottom: spectra of 2-OMe (solid squares),
2 (open circles), and 2-Cl (crosses) in n-hexane.
behavior is reversed for 1-OMe, i.e., P* and A* are closer lying
in energy and the maximum of the φf vs ET(N) plot (minimum
of quenching interaction by P*) is only reached in acetonitrile,
the solvent of highest polarity employed (Figure 3). The
competetive interaction and highly polar nature of the other
quenching state(s) K* is also obvious from a comparison of
the plots in Figure 3. Upon going from dichloromethane (ET-
(N) ) 0.309) to acetonitrile (ET(N) ) 0.460), the fluorescence
losses are most pronounced for the compound with the strongest
acceptor (1-Cl, 3.5-fold decrease in φf), less pronounced for
intermediate 1 (1.6-fold) and for 1-OMe, the opposite, i.e., an
increase in fluorescence quantum yield is observed (from 0.13
to 0.18). For a better illustration of this influence of electron-
donating or -withdrawing substituents, Figure 6 contains a set
of spectra of 1, 1-Cl, and 1-OMe in acetonitrile. As already
mentioned above, the differences between 1 and 4 can be
understood on this basis as well. 4, showing bathochromically
shifted spectra and a maxmium in the φf vs ET(N) plot at lower
polarities than 1-Cl (ET(N) ) 0.28 for 4, Figure 3), is also the
weakest emitting chalcone of the series 1, 1-Cl, 1-OMe, and 4
in acetonitrile (Figure 6). The fluorescence quenching toward
higher polarities producing the maximum in the φf vs ET(N)
plot are due to the polar quenching state K* as discussed below.
Similar effects seem to account for the observations made
for the substitution of the COPh acceptor with a p-Cl or p-OCH3
group in bridged 2. Increasing the strength of the CT process
as in 2-Cl leads to a higher fluorescence yield in apolar solvents.
Here, initially excited, polar, and fluorescent E* is less quenched
by population of nonemissive weakly polar P* as, e.g., in 2
(Figure 3). Consequently, a nonemissive highly polar K* state
is already accessible in weakly or medium polar solvents such
as diethyl ether or THF and the curve of 2-Cl is already of
lowest amplitude in these solvents (Figure 3). Introducing a
donor group (methoxy group) in the acceptor part of 2, a
compound (2-OMe) with a similar behavior as observed for 4
by a similar bridge as in 2 (oxygen-bridged benzofuran
derivatives, Scheme 3) exhibit moderate fluorescence in polar
solvents (quantum yield of 0.13 in aqueous solution61), no
intrinsic quenching channel due to this oxygen bridge is
probable. Furthermore, both bridged derivatives 7 and 8 are
highly fluorescent despite an intramolecular oxygen bridge as
in 2.
The effect of donor-acceptor substitution pattern on the
relative energetic position of the states P*, A*, and K* is obvious
from a comparison of 1, 1-Cl, and 1-OMe as well as 2, 2-Cl,
and 2-OMe. Increasing the acceptor strength in the unbridged
derivative (1 f 1-Cl) shifts the maximum of the φf vs ET(N)
plot to the region of lower polarity (maxima at ET(N) ) 0.37
for 1 and 0.33 for 1-Cl in Figure 362). In 1-Cl, the possibility
of populating the fluorescent A* state is enhanced as compared
to the reaction toward the quenching state P*. Accordingly, this

9634 J. Phys. Chem. A, Vol. 103, No. 48, 1999
Rurack et al.
or 6 is obtained. For 2-OMe, in apolar solvents the interaction
of quenching mechanisms (P* state formation, proximity effect)
is dominant but with increasing solvent polarity, fluorescing
E* is more strongly stabilized and fluorescence is regained
(maximum in the φf vs ET(N) plot in THF in Figure 3). Upon
further increasing solvent polarity, the highly polar but non-
emissive K* state is responsible for enhanced nonradiative
deactivation. Accordingly, the influence of donor acceptor
strength on the photophysical properties of the bridged derivative
2 is again well-illustrated in the lower part of Figure 6,
emphasizing the apolar side of the solvent polarity range.
SCHEME 4: Generalized Excited State Reaction Model
for D-A-Chalcones (Triplet States, Reverse Reactions,
and Consecutive Population of Two Excited Species are
Omitted for Clarity)a
Returning to the comparison of 1 with 3 and 5 (where
quenching via a K* state is not possible), different reactivities
for nonradiative P* and luminescent A* state formation should
occur. The torsion angles around b56 given in Table 2 are much
larger in the case of the two bridged derivatives and thus not
only population of a fluorescent A* state but also reaction
toward a nonemissive P* state might be accelerated.63 Even in
the case of related DMABI (Scheme 3), where donor and
acceptor moiety are nearly planar in the ground state, Gulbinas
et al. recently attributed efficient P* state formation (bCdC, see
Scheme 3) to be responsible for the generally weak fluorescence
of this dye molecule (φf < 0.01 throughout the solvent polarity
region from n-hexane to DMSO) despite possible population
of an A* state (bC-C, see Scheme 3).64
a Dashed arrows indicate the second branch, coming into play when
both ground state conformers tE and cE are excited. Solid arrows denote
reactions and radiative deactivations, and wiggled arrows mark non-
radiative decays. Any triplet formation would occur most probably from
P*.
will be more strongly stabilized. If both ground state conformers
are present, the excited-state reaction scheme extends to a dual
scheme with individual species tA*, cA*, tP*, and cP* (Scheme
4). However, the close spectroscopic relation of both E
conformers is exemplified by the spectrally similar absorption
and emission characteristics of, e.g., 5 (fixed cE) and 8 (fixed
tE) and is supported by the quantum chemical calculations.
The large difference in fluorescence quantum yields between
1-6 and 7 and 8 in any solvent or medium of high polarity
demonstrates the efficient population of nonemissive K* and
P* states. But whereas the conical intersection responsible for
the P* type funnel is well approximated by the twist of the
double bond in the quantum chemical calculations, the behavior
connected with the formation of K* is not well reproduced.
Here, the one-dimensional approach employed in the calcula-
tions apparently fails to model the actual molecular changes
sufficiently well and additional coordinates are required to
describe this conical intersection.51,52,66 Future work using the
phase change rule51,52 may help to identify the missing
coordinates. This paper is meant to pave the way and establish
experimental facts.
Triplet Formation. With the data given in the results section,
the role of triplet state population cannot be elucidated any
further in this paper. However, intersystem crossing should occur
most probably from P*. In the case of the related bis(p-
(dimethylamino)benzylidene)acetone, DeVoe et al. found direct
experimental evidence for the formation of a “dark” P* state
by employing transient absorption spectroscopy.4 The lack of
the detection of any photoproducts in these studies,4 neither by
addition reactions nor by trans-cis isomerization,67 and the
energetically unstable Z conformers (any species formed by Cd
C double bond twisting should undergo a very rapid backward
reaction leading to the ground state E conformer) do not point
to a major contribution of intersystem crossing for the deactiva-
tion of excited 4′-(dimethylamino)chalcones.
Nonradiative Behavior in Alkanes. In apolar solvents such
as n-hexane, formation of a weakly polar P* state alone cannot
explain the decrease in fluorescence quantum observed for most
of the compounds regardless of the bridging pattern in the φf
vs ET(N) plot (Figures 2 and 3). Most probably, in these solvents,
the emitting polar state is sufficiently destabilized to experience
perturbation by the proximity effect, i.e., vibronic coupling to
energetically close lying nπ* states.53
Conformational Isomerism. In the following subsection, the
implications of the multiple state model will be examined more
closely for their consistency with the stereoisomerism of the
D-A-chalcones. The only two derivatives fixed in the tE
conformation are the highly fluorescent and largely bridged 7
and 8. Here, in apolar solvents, mainly the proximity effect
should lead to fluorescence quenching, but in all the more polar
solvents, only a polar fluorescent tE* state emits. Additionally,
in the case of 7, a highly polar, fluorescent tA* emitting species
is assumed to play a role. The higher fluorescence quantum
yield of 7 as compared to 8 in acetonitrile (Figure 2),7 the
increased solvatochromic behavior found by Wang and Wu for
7,7 and the assignment of the fluorescence of a derivative of 7
to CT fluorescence from an emissive TICT state65 given in the
literature so far support these assumptions.
None of the compounds 3, 5, and 6 with a fixed cE
conformation possesses a bridged double bond, and thus the
rate of intrinsic cE fluorescence remains unknown. But emission
in these compounds with a flexible double bond occurs most
probably via A* population (aniline twisting), counterbalanced
by (parallel or consecutive) reaction toward nonemissive P*
(double bond twisting). This is in accord with the three-state
model developed for the related D-A-stilbenes which predicts
an enhancement of the nonradiative losses due to the more
efficient population of the P* state in such a case of some
pretwisting of the double bond.17
5. Conclusion
The fluorescence efficiencies of chalcones with various
bridged bonds and different donors and acceptors can be
understood by considering (at least) four photophysically active
states: the highly emissive all-planar structure E*, a further
emissive and highly polar state of twisted structure (A*,
probably involving the twist of the anilino group), and two
nonemissive states of low (P*, twisting of the double bond)
and high dipole moment (K*). The quantum chemical calcula-
tions support the identification of P* with a twist of bond b67
Concerning 1, 2, and their substituted analogues, all able to
adopt the cE and/or the tE conformation, the ground state
equilibrium between both conformers will be largely dependent
on their stabilization by the respective solvent. Upon increasing
solvent polarity, the conformer with the higher dipole moment

Quantum Yield Switching of Fluorescence in Chalcones
J. Phys. Chem. A, Vol. 103, No. 48, 1999 9635
(30) Zubarovskii, V. M.; Bricks, J. L. Ukr. Khim. Zh. 1982, 48, 761.
Zubarovskii, V. M.; Bricks, J. L. Khim. Heterocycl. Soed. 1982, N5, 644.
(31) Shriner, R. L.; Teeters, W. O. J. Am. Chem. Soc. 1938, 60, 936.
(32) Rechthaler, K.; Ko¨hler, G. Chem. Phys. Lett. 1996, 250, 152.
(33) Dewar, M. J. S.; Zoeblich, E. G.; Healy, E. F.; Stewart, J. J. P. J.
Am. Chem. Soc. 1985, 107, 3202.
(34) Dewar, M. J. S.; Stewart, J. J. P.; Ruiz, J. M.; Liotard, D.; Healy,
E. F.; Dennington, R. D., II AMPAC 4.5 and AMPAC 5.0; Semichem:
Shawnee, 1993 and 1994.
(35) Le´tard, J.-F.; Lapouyade, R.; Rettig, W. Chem. Phys. 1994, 186,
119.
(36) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486.
(37) Bo¨ttcher, C. J. F. Theory of Electric Polarization, 2nd ed.;
Elsevier: Amsterdam, 1973; Vol. 1.
and suggest that it corresponds to a structure not far from a
conical intersection. From the experimental bridging studies,
K* is related with the twisting of the ketone function (b78), but
the calculations do not indicate a significant S0/S1 narrowing
as expected for the approach to a conical intersection. This might
be due to the involvement of a second coordinate to reach the
conical intersection66 which has not been considered in these
one-dimensional calculations. The nature of this second coor-
dinate remains to be determined. Recent qualitative improvement
in the understanding of conical intersections by the phase change
rule51,52 should be a promising approach to solving this question.
(38) Karelson, M.; Zerner, M. C. J. Am. Chem. Soc. 1990, 112, 9405.
(39) Bilot, L.; Kawski, A. Z. Naturforsch. 1962, 17a, 621.
(40) Varma, C. A. G. O.; Groenen, E. J. J. Recl. TraV. Chim. Pays-Bas
1972, 91, 296.
(41) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165.
(42) Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Justus
Liebigs Ann. Chem. 1963, 661, 1.
Acknowledgment. Financial support by the Deutsche For-
schungsgemeinschaft (DFG; Re 387/8-2 and Re 387/13-1) and
the Bundesministerium fu¨r Bildung und Forschung (BMBF,
13N7120) is gratefully acknowledged.
(43) Rettig, W. Top. Curr. Chem. 1994, 169, 253.
References and Notes
(44) In polar protic solvents (alcohols), the fluorescence of all D-A-
chalcones (note that no data are available for 7 and 8) is drastically quenched
as a consequence of hydrogen bond formation at the carbonyl group in the
acceptor part.3,45
(45) Inoue, H.; Hida, M.; Nakashima, N.; Yoshihara, K. J. Phys. Chem.
1982, 86, 3184.
(46) Rabinovich, D. J. Chem. Soc. B 1970, 11. Rabinovich, D.; Schmidt,
G. M. J.; Shaked, Z. J. Chem. Soc., Perkin Trans. 2 1973, 33.
(47) Hayes, W. P.; Timmons, C. J. Spectrochim. Acta 1968, 24A, 323.
(48) Rurack, K.; Radeglia, R. Unpublished results.
(49) Michl, J. J. Mol. Photochem. 1972, 243 and 257. Michl, J. Pure
Appl. Chem. 1975, 41, 507. Michl, J. Top. Curr. Chem. 1974, 46, 1.
Zimmerman, H. E. J. Am. Chem. Soc. 1966, 88, 1566. Bonacic-Koutecky´,
V.; Michl, J. J. Am. Chem. Soc. 1985, 107, 1765. Michl, J.; Bonacic-
Koutecky´, V. Electronic Aspects of Organic Photochemistry; J. Wiley &
Sons: New York, 1990. Bernardi, F.; Olivucci, M.; Robb, M. A. J.
Photochem. Photobiol. A: Chem. 1997, 105, 365. Celani, P.; Bernardi, F.;
Olivucci, M.; Robb, M. A. J. Chem. Phys. 1995, 102, 5733.
(50) Fuss, W.; Lochbrunner, S.; Mu¨ller, A. M.; Schikarski, T.; Schmid,
W. E.; Trushin, S. A. Chem. Phys. 1998, 232, 161. Garavelli, M.; Vreven,
T.; Celani, P.; Bernardi, F.; Robb, M. A.; Olivucci, M. J. Am. Chem. Soc.
1998, 120, 1285.
(51) Zilberg, S.; Haas, Y. J. Phys. Chem. A 1999, 103, 2364.
(52) Zilberg, S.; Haas, Y. Chem.sEur. J. 1999, 5, 1755.
(53) Wassam, W. A., Jr.; Lim, E. C. J. Chem. Phys. 1978, 68, 433.
(54) Rettig, W.; Majenz, W. Chem. Phys. Lett. 1989, 154, 335.
(55) Hicks, J.; Vandersall, M.; Babarogic, Z.; Eisenthal, K. B. Chem.
Phys. Lett. 1985, 116, 18. Rettig, W. In Dynamics and Mechanisms of
Photoinduced Electron Transfer and Related Phenomena; Mataga, N.,
Okada, T., Masuhara, H., Eds.; Elsevier: Amsterdam, 1992; p 57.
(56) Fery-Forgues, S.; Le Bris, M. T.; Pouget, J.; Rettig, W.; Valeur,
B. J. Phys. Chem. 1992, 96, 701.
(57) The generally lower fluorescence quantum yield in the case of 3
can be understood on the basis of the higher flexibility of the cyclohexene
ring undergoing a nearly barrierless ring inversion (boat-chair intercon-
version).58
(58) Jensen, F. R.; Bushweller, C. H. J. Am. Chem. Soc. 1969, 91, 5774.
(59) Lippert, E.; Rettig, W.; Bonacic-Koutecky´, V.; Heisel, F.; Miehe´,
J. A. AdV. Chem. Phys. 1987, 68, 1.
(60) Kirkor-Kaminska, E.; Rotkiewicz, K.; Grabowska, A. Chem. Phys.
Lett. 1978, 58, 379.
(61) Grynkiewicz, G.; Poenie, M.; Tsien, R. Y. J. Biol. Chem. 1985,
260, 3440.
(62) Maximum of a third-order polynomial fit of the φf vs ET(N) plot.
(63) The excited-state reaction mechanism might as well involve
reactions between states (e.g., P* and A*) which are not discussed here
any further.
(64) Gulbinas, V.; Kodis, G.; Jursenas, S.; Valkunas, L.; Gruodis, A.;
Mialocq, J.-C.; Pommeret, S.; Gustavsson, T. J. Phys. Chem. A 1999, 103,
3969.
(65) Revill, J. A. T.; Brown, R. G. J. Fluoresc. 1992, 2, 107. Revill, J.
A. T.; Brown, R. G. Chem. Phys. Lett. 1992, 188, 433.
(66) Longuet-Higgins, H. C. Proc. R. Soc. London, Ser. A 1975, 344,
147. Herzberg, G.; Longuet-Higgins, H. C. Trans. Faraday Soc. 1963, 35,
77.
(1) Nicodem, D. E.; de M. G. Matos, J. A. J. Photochem. 1981, 15,
193.
(2) Caldwell, R. A.; Singh, M. J. Am. Chem. Soc. 1983, 105, 5139.
(3) Wang, Y. J. Phys. Chem. 1985, 89, 3799.
(4) DeVoe, R. J.; Sahyun, M. R. V.; Schmidt, E.; Sadrai, M.; Serpone,
N.; Sharma, D. K. Can. J. Chem. 1989, 67, 1565.
(5) Gustav, K.; Bartsch, U.; Karnizschky, K. Z. Chem. 1989, 29, 213.
(6) Jiang, Y.-B.; Wang, X.-J.; Lin, L. J. Phys. Chem. 1994, 98, 12367.
(7) Wang, P.; Wu, S. J. Photochem. Photobiol. A: Chem. 1995, 86,
109.
(8) Su, G.-B.; He, Y.-P.; Li, Z.-D. Jiegou Huaxue (Chinese J. Struct.
Chem., Engl. Transl.) 1997, 16, 263.
(9) Mager, L.; Melzer, C.; Barzoukas, M.; Fort, A.; Mery, S.; Nicoud,
J.-F. Appl. Phys. Lett. 1997, 71, 2248.
(10) Monroe, B.; Smothers, W. K.; Keys, D. E.; Krebs, R. R.; Mickish,
D. J.; Harrington, A. F.; Schicker, S. R.; Armstrong, M. K.; Chan, D. M.
T.; Weathers, C. I. J. Imaging Sci. 1991, 35, 19.
(11) Rurack, K.; Bricks, J. L.; Kachkovskii, A. D.; Resch, U. J. Fluoresc.
1997, 7, 63S. Rurack, K.; Bricks, J. L.; Slominskii, J. L.; Resch-Genger,
U. In Near-Infrared Dyes for High Technology Applications; Da¨hne, S.,
Resch-Genger, U., Wolfbeis, O. S., Eds.; Kluwer Academic: Dordrecht,
1998; p 191.
(12) Rurack, K.; Bricks, J. L.; Radeglia, R.; Resch-Genger, U. J. Phys.
Chem. A, submitted.
(13) Lewis, G. N.; Calvin, M. Chem. ReV. 1939, 25, 273. Hofer, L. J.
E.; Grabenstetter, R. J.; Wiig, E. O. J. Am. Chem. Soc. 1950, 72, 203.
(14) Carstens, T.; Kobs, K. J. Phys. Chem. 1980, 84, 1871.
(15) Jones, II, G.; Jackson, W. R.; Choi, C.-Y.; Bergmark, W. R. J.
Phys. Chem. 1985, 89, 294.
(16) Le´tard, J.-F.; Lapouyade, R.; Rettig, W. J. Am. Chem. Soc. 1993,
115, 2441.
(17) Rettig, W.; Majenz, W.; Herter, R.; Le´tard, J.-F.; Lapouyade, R.
Pure Appl. Chem. 1993, 65, 1699.
(18) Lapouyade, R.; Czeschka, C.; Majenz, W.; Rettig, W.; Gilalbert,
E.; Rullie`re, C. J. Phys. Chem. 1992, 96, 9643.
(19) Sczepan, M.; Rettig, W.; Bricks, Y. L.; Slominski, Y. L.; Tolma-
chev, A. I. J. Photochem. Photobiol. A: Chem. 1999, 124, 75.
(20) Vogel, M.; Rettig, W. Ber. Bunsen-Ges. Phys. Chem. 1985, 89,
962.
(21) Vogel, M.; Rettig, W.; Sens, R.; Drexhage, K. H. Chem. Phys. Lett.
1988, 147, 452.
(22) Vogel, M.; Rettig, W.; Sens, R.; Drexhage, K. H. Chem. Phys. Lett.
1988, 147, 461. Vogel, M.; Rettig, W.; Fiedeldei, U.; Baumga¨rtel, H. Chem.
Phys. Lett. 1988, 148, 347.
(23) Plaza, P.; Dai Hung, N.; Martin, M. M.; Meyer, Y. H.; Vogel, M.;
Rettig, W. Chem. Phys. 1992, 168, 365.
(24) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D.
J.; Baumann, W. NouV. J. Chim. 1979, 3, 443. Rettig, W. Angew. Chem.,
Int. Ed. Engl. 1986, 25, 971. Rettig, W. In Modern Models of Bonding and
Delocalization; Liebman, J. F., Greenberg, A., Eds.; VCH: New York, 1988;
p 229.
(25) Drexhage, K. H. J. Res. Natl. Bur. Stand. (U.S.) 1976, 80A, 421.
(26) Resch, U.; Rurack, K. Proc. SPIE-Int. Soc. Opt. Eng. 1997, 3105,
96.
(27) Sachs, F.; Lewin, W. Ber. 1902, 35, 3569.
(67) Photochemical products have only been reported for the unsubsti-
tuted parent compound 1,3-diphenylprop-2-en-1-one1 and for bis(p-(di-
methylamino)benzylidene)acetone.4
(28) Grosscurt, A. C.; Van Hes, R.; Wellinga, K. J. Agric. Food Chem.
1979, 27, 406.
(29) Tewari, R. S.; Dubey, A. K.; Misra, N. K. J. Chem. Eng. Data
1981, 26, 106.
(68) Lippert, E. Z. Elektrochem. 1957, 61, 962.