Comprehensive evaluation of the absorption, photophysical, energy transfer, structural, and theoretical properties of ct-oligothiophenes with one to seven rings

Article abstract of DOI:10.1021/jp960852e

A large basis set of ct-oligothiophenes with two to seven rings (a.2-a.7), also including thiophene, al, have been investigated in five solvents regarding absorption, fluorescence and phosphorescence, quantum yields of fluorescence (Φ) and triplet formation (0r), lifetimes of fluorescence and the triplet state, quantum yields of singlet oxygen production (<￡A), all rate constants kF, k[C, Msc, and several of the foregoing as a function of temperature. Ten different theoretical calculations across several levels including three levels of ab initio have been carried out regarding which conformer is lowest in energy and the A//'s among all conformers of O.2, a.3 and o5, as well as calculations of transitions energies of the a-oligothiophenes. We have shown that the (1) 'Bu state is the lowest singlet state for all o2-a? in any solvent, in contradiction to previous predictions for the higher members. Based on absorption and fluorescence data and calculations of atomic charges in So and Si, the ground state is twisted while the excited state is planar (quinoidal-like). Significant charge transfer occurs between So and Si but not So and T|. For all a.2-a.7, c is small, ff F is approximately constant while Arise decreases significantly from a.2 to aJ. The decrease is A'isc is believed to arise from a decrease in matrix elements of the type (1cr|H'|3W|). The essential lack of phosphorescence is assigned as originating from inter-ring twisting mode coupling between TI and SQ. Triplet energy transfer to 3Oa to produce 'Oi is highly efficient for a.2-o5. Based on all data, the first an representative of a-polythiophene is a5.

Full text of DOI:10.1021/jp960852e

J. Phys. Chem. 1996, 100, 18683-18695
18683
Comprehensive Evaluation of the Absorption, Photophysical, Energy Transfer, Structural,
and Theoretical Properties of r-Oligothiophenes with One to Seven Rings
Ralph S. Becker,*,†,‡ J. Seixas de Melo,†,§ Anto´nio L. Mac¸anita,†,| and Fausto Elisei
Instituto de Tecnologia Qu´ımica e Biolo´gica, Rua da Quinta Grande n°6, 2780 Oeiras, Portugal, Department
of Chemistry, UniVersity of Arkansas, FayetteVille, AR, Departamento de Qu´ımica, UniVersidade de Coimbra,
3000 Coimbra, Portugal, Departamento de Qu´ımica do IST, Lisboa, Portugal, and Dipartimento di Chimica,
UniVersita´ di Perugia, Perugia, Italy
ReceiVed: March 20, 1996; In Final Form: August 22, 1996X
A large basis set of R-oligothiophenes with two to seven rings (R2-R7), also including thiophene, R1, have
been investigated in five solvents regarding absorption, fluorescence and phosphorescence, quantum yields
of fluorescence (φF) and triplet formation (φT), lifetimes of fluorescence and the triplet state, quantum yields
of singlet oxygen production (φ∆), all rate constants kF, kIC, kISC, and several of the foregoing as a function
of temperature. Ten different theoretical calculations across several levels including three levels of ab initio
have been carried out regarding which conformer is lowest in energy and the ∆H’s among all conformers of
R2, R3 and R5, as well as calculations of transitions energies of the R-oligothiophenes. We have shown that
the (l) 1Bu state is the lowest singlet state for all R2-R7 in any solvent, in contradiction to previous predictions
for the higher members. Based on absorption and fluorescence data and calculations of atomic charges in S0
and S1, the ground state is twisted while the excited state is planar (quinoidal-like). Significant charge transfer
occurs between S0 and S1 but not S0 and T1. For all R2-R7, φIC is small, k0F is approximately constant while
kISC decreases significantly from R2 to R7. The decrease is kISC is believed to arise from a decrease in matrix
1
3
elements of the type ΨCT|H′| Ψ1 . The essential lack of phosphorescence is assigned as originating from
inter-ring twisting mode coupling between T1 and S0. Triplet energy transfer to 3O2 to produce 1O2 is highly
efficient for R2-R5. Based on all data, the first Rn representative of R-polythiophene is R5.
Introduction
energy of states, but again, these have been limited essentially
to R2-R3, particularly R2. We shall again consider all of the
foregoing at the appropriate time.
The R-oligothiophenes (designated R1, R2, R3, ... where n
) 1, 2, 3, ..., respectively)
One of the dominant questions of importance is the state
order, particularly relating to the two lowest singlet π,π* states.
The symmetry of the trans oligothiophenes varies depending
upon the odd-even number of rings, C2V for odd, C2h for even.
1
1
The states of interest are the Bu and Ag ones within C2h
symmetry, and this notation, or 1B and 1A, is commonly carried
through for C2V symmetry. This order is considered important
in terms of the relationship/identity of R-oligothiophenes and
polyenes with the same number of double bonds. For R2, in
solution the (1)1Bu-like state is the lowest where the (2)1Ag-
like state has been located above it by two-photon spectroscopy,5
and this (2)1Ag-like state also has been located above the (1)1Bu-
like state in a crystalline film of R66 (all state symbols should
read “like”, since the exact molecular symmetry and degree of
state mixing is not known; however, we shall drop this addition
for simplicity). Based on some spectroscopic data on R,ω-
dithienylpolyenes, restricted CI (configuration interaction) cal-
culations, extrapolated curves of Ag and Bu states energies
of some polythiophene oligomers, and the difference in state
energies expected for R6 in the crystalline state vs alkane
solution, the (2)1Ag state was assigned below the (1)1Bu state
for R6 and all other polythiophene oligomers with more than
six rings (R6, R7, ....).7 We will show that this is not the case.
are currently of keen interest because (1) they are interesting
analogues of polyenes, (2) they are good singlet oxygen
sensitizers and biophotosensitizers, and (3) particularly those
with higher n are important for their use in nonlinear optics
applications, charge storage, and molecular electronics.
There has been only limited information available on the
fluorescence, fluorescence quantum yields (φF) and lifetimes
(τF), triplet yields (φT) and lifetimes (τT), singlet oxygen yields
(φ∆), and the rate constants connecting the excited states and
those connecting excited states to the ground state. Recently,1
all the above aspects except φ∆ were evaluated in one solvent.
Also, R2-R6 have been studied2,3 relative to φF and φF for R2-
R6 and φ∆ for R3-R6.4 Other scattered data exist for some
Rn’s regarding some of the photophysical parameters. Some
information exists for films, but again it is limited, even more
than for solutions, in terms of the number of oligothiophenes
and the photophysical parameters evaluated. One area receiving
considerable attention is theoretical, expecially regarding tor-
sional (inter-ring) barriers, ∆H of trans and cis conformers, and
1
1
Other calculations8 predict the (2)1Ag state already below the
(1)1Bu state at R2-R4, with crossing occurring between R4 and
R5. However, by restriction of the CI (which seems to have
no real rational justification as was also done before),7 the (1)1Bu
state was now predicted to be lowest for all Rn’s through R6,
R7. Still other calculations9-11 on R2 and some higher Rn’s in
† Instituto de Tecnologia Qu´ımica e Biolo´gica.
‡ University of Arkansas.
§ Universidade de Coimbra.
| Departamento de Qu´ımica do IST.
Universita´ di Perugia.
X Abstract published in AdVance ACS Abstracts, November 1, 1996.
S0022-3654(96)00852-0 CCC: $12.00 © 1996 American Chemical Society

18684 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
some cases give varying results regarding the relative ordering
of the states for R2 and place the (2)1Ag state lowest for R3
and R4.9
The molar extinction coefficients (ꢀ) were obtained with seVen
solutions of different concentrations. The slope of the plot of
the absorption values (at the maximum wavelength of absorp-
tion) vs the concentration values gave us the ꢀ values with
correlation values g0.999.
We will show that despite any theoretical and/or comparative
experimental data to the contrary, the (1)1Bu state is the lowest
excited singlet state in all R-oligothiophenes from R2 to and
including R7. We will also determine the following in fiVe
solvents for R2-R7: (1) fluorescence (φF) and triplet (φT)
quantum yields, (2) fluorescence (τF) and triplet (τT) lifetimes,
The radiative rate constant kR was calculated with the
Strickler-Berg14 equation
ꢀ(νj)
νj
-3
kSRB ) (2.88 × 10-9)n2 νjF
dνj
-1
∫
AV
(3) all determinable rate constants k0 , kISC (intersystem cross-
F
ing), kIC (internal conversion), and kNR (nonradiative), (4)
absorption (and ꢀ values) and fluorescence spectra (and phos-
phorescence for R1) at 298 and 77 K, (5) singlet oxygen
quantum (φ∆) yields of R2-R5 and R7, (6) τF and φF as a
function of temperature for R3 and R7, (7) 10 different
theoretical calculations of various levels for a broad variety of
Rn’s, including three levels of ab initio that evaluate which
conformer is the lowest in energy for all conformers of R2, R3,
R5, (8) calculations of atomic charges and their changes between
S0 and S1 as well as T1, (9) calculation of transition energies
and oscillator strengths for R1-R5, and (10) correlation of
absorption, emission, and photophysical parameters as a function
of n or 1/n where n is the number of rings. All these will be
integrated to elucidate the behavior of the absorption and
photophysical properties of R1-R7.
with
I(νj) dνj
∫
F
-3
νjF
)
-1
AV
νj -3I(νjF) dνjF
∫
F
using the integrated first absorption band and the fluorescence
band. For thiophene, since there is no fluorescence, the use of
the equation of Fo¨rster15 was needed:
ꢀ(νj)(2νj0 - νj)3
kF ) (2.88 × 10-9)n2
dνj
∫
R
νj
Fluorescence decays were obtained using a time-correlated
single-photon-timing technique with a home-built apparatus as
previously described16 except for the wavelength shift, which
is now 300 fs/nm. The obtained fluorescence decays were
deconvoluted in a Microvax 3100 employing the method of
modulating functions.17
Experimental Section
Thiophene was purchased from Riedel de Ha¨en and was used
as received. Bithiophene and terthiophene were purchased from
Aldrich and were respectively purified by sublimation and
recrystallization. Tetrathiophene and dibutylheptathiophene
were a kind gift from J. Kagan and H. Winberg, respectively.
These were purified by column chromatography. Hexathiophene
was provided by H. Naarmann (BASF).
The experimental setup used for triplet spectra and triplet
yields of R2-R5 was that described elsewhere.18,19 Absorption
spectra were recorded every 10 nm averaging at least 5 shots
per wavelength recorded. The triplet-triplet molar extinction
coefficients (ꢀT) were evaluated using the energy transfer method
from naphthalene (in benzene, dioxane, and ethanol) and from
benzophenone (in benzene and acetonitrile) to Rn’s. A solution
of the sensitizer, S, had an absorbance of 1.5-1.6 and the Rn
had an absorbance of 0.15-0.25 at the excitation wavelengths
of 308 and 347 nm for naphthalene and benzophenone,
respectively.
Pentathiophene was synthesized by bromination of ter-
thiophene in the R positions followed by reaction with 2-bro-
mothiophene Via Grignard coupling in the presence of a nickel
catalyst (Ni[C6H5)2PCH2CH2P(C6H5)2]Cl2) synthesized as else-
where reported (and dried with SOCl2).12 The final product,
R5, was recrystallized from dichloromethane.
The equation20
All the solvents were of spectroscopic or equivalent grade,
except for methylcyclohexane (which was purified over a
mixture of H2SO4/HNO3, then distilled, and finally chromato-
graphed on an Al2O3/SiO4 column) and ethanol (dried and
purified by distillation over CaO). All the other solvents were
used without further purification. The solutions used (1 × 10-5
to 10-6 M) were deoxygenated by either N2 or Ar bubbling.
Anthracene, naphthalene (J. K. Baker Chemical Co.), and
benzophenone (Aldrich Chemical Co.) were used without further
purification.
-1
∆AA - ∆Adir k2 - k
ln(k2/k3)
ꢀT(A) ) ꢀT(S)
1 exp -
{
}
[
]
∆ASfS
k2
(k2/k3) - 1
was used where the acceptor A ) Rn and S is the sensitizer.
This equation takes into account the following: ∆AS, the
absorbance change of sensitizer alone; ∆AA, the absorbance
change of the maximum; k2, the decay rate constant of the
sensitizer in the presence of the acceptor; k3, the decay rate
constant of the sensitized triplet of the acceptor; ∆Adir, the direct
excitation of the acceptor; fS, the fraction of light absorbed by
the sensitizer in the solution of S + A with respect to the
sensitizer alone. Triplet properties of the sensitizers in some
solvents were taken from the literature21,22 (ꢀT in M-1 cm-1):
for naphthalene (benzene), ꢀT ) 13 200 (420 nm), φT ) 0.75;
for naphthalene (ethanol), ꢀT ) 40 000 (420 nm); for benzophe-
none (benzene), ꢀT ) 7200 (530 nm), φT ) 1.0; for benzophe-
none (acetonitrile), ꢀT ) 6500 (530 nm), φT ) 1.0. The ꢀT
values of anthracene and naphthalene in dioxane were deter-
mined in this work. The φT of anthracene was determined by
the heavy atom effect of 4-bromo-N,N′-dimethylaniline on the
fluorescence quantum yield and triplet population.19,23 Energy
transfer between naphthalene (donor) and anthracene (acceptor)
were used to obtain ꢀT of naphathalene in dioxane: for
Absorption and fluorescence spectra were run with a Beckman
DU-70 and a SPEX Fluorolog spectrometer, respectively. All
the fluorescence spectra were corrected for the wavelength
response of the system.
The fluorescence quantum yields at 293 K were measured
using several standards, namely, methyl 1-pyrenoate (φF ) 0.83
in cyclohexane) and 3-chloro-7-methoxy-4-methylcoumarin (φF
) 0.12 in cyclohexane13). Terthiophene in ethanol (φF ) 0.054)
was also used for internal verification of the obtained values.
The fluorescence quantum yields at 77 K were obtained by
running under the same experimental conditions the solution
done at 293 K, avoiding by this way external interferences. The
φF value was than obtained by assuming a 20% “shrinkage” of
the ethanol solvent on going from 293 to 77 K.

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18685
TABLE 1: Photophysical and O∆ Data for r-Oligothiophenes in Benzenea,b
cpd
φF
τF (ns)
kF0
kNR
φT
φ∆
0.96
kIC
kISC
22
5.9
1.7
0.72
τT (µs)
R2
R3
R4
R5
R6
R7
0.026
0.07
0.18
0.34
0.44
0.36
0.046
0.16
0.44
0.82
0.97
0.82
0.55
0.44
0.41
0.41
0.45
0.43
21e
0.99
0.95
0.73
0.59
0.11c
0.03c
0.20
104
88
38
5.8e
1.9
0.81d
0.72
0.56
∼0.36f
h
0.81
0.58
0.79
0.098
24
g17g
21
e0.6
0.049
0.73
a Solutions degassed with nitrogen or argon by bubbling. k’s are 109 s-1, and R7 is the di-n-butyl-substituted R7. b Also see Table 2, acetonitrile.
c Assumes (1 - φT - φF) ) 0.005. d The value of φ∆ ranged from 0.65 to 0.85 depending on the technique used (see Experimental Section for
1
1
e
techniques). This value was that from measuring emission of O2. In acetonitrile a value of 0.74 was obtained by measuring O2 emission. kNR
obviously cannot be less than kISC + kIC, but recall that errors in φT for large φT particularly effect the accuracy of kIC; see text. f Reference 4. We
were unable to obtain sufficient concentration in benzene or dioxane to obtain a value using our techniques (see Experimental Section for techniques).
g Wintengs, V.; Valat, P.; Garnier, F. J. Phys. Chem. 1994, 98, 228 indicate 24 µs in dichloromethane and also R4 ) 35 µs. h We have determined
a tentative value of 0.25 in dioxane.
naphthalene (dioxane), ꢀT ) 15 000 (420 nm); for anthracene
(dioxane), ꢀT ) 50 000 (430 nm), φT ) 0.62.
spectra and extinction coefficients, fluorescence and T-T
maxima, as well as the experimental k0F and kF0 from the
Strickler-Berg equation (SB).14 Other tables will be considered
in the Results and Discussion sections.
The product ꢀTφT for each Rn was obtained by the laser
energy effect on the change of absorbance of an Rn measured
at the λmax of optically matched solutions (A ≈ 0.08) compared
with benzophenone in acetonitrile and naphthalene in benzene
as references/standards. Plots of ∆A vs laser dose were linear
and passed through zero, indicating that only one-photon
processes were occurring. Moreover, there was only a <5%
difference between ꢀTφT calculated with the two different
references/standards. The triplet yield of Rn’s was then
calculated from the ꢀTφT/ꢀT ratio.
Absorption and Emission. Dependence on Solvent and
Temperature. The solution absorption spectra at room tem-
perature in any solvent for R2-R7 are generally devoid of
structure (Figure 1). Table 3 gives representative absorption
spectral data in dioxane. There is a relatively small effect of
solvent on these maxima. However, the low-temperature
absorption, as represented by the fluorescence excitation spectra,
can have significant structure (Figure 2 and ref 1), particularly
for R2 and R3. Moreover, there is a considerable red shift,
∼1600 cm-1, in the maxima. There is an excellent linear
correlation between 1/n and the maximum of the first transition,
as well as the 0-0 energy particularly for R4-R7 (Figure 3).
If R2 and R3 are included, these points scatter somewhat
compared to the others and change the slope slightly. Nonethe-
less, the linearity is good. This observation regarding the
deviation of R2 in particular occurs for other kinds of relation-
ships involving 1/n (or 1/N where N is the length of the molecule
in Å) such as that for the maximum of the T-T absorption,
among others (Figure 3). These types of data plus photophysical
results indicate that R2, and even R3, are not yet legitimate
representatives of a polythiophene and that such a representation
seemingly begins with R4. Actually, additional consideration
of all the photophysical data indicates that the first true
representatiVe of a polythiophene is R5.
On the basis of the plot of 1/n vs absorption maxima of R4-
R7 (corrected for the presence of two n-butyl groups), we can
predict the expected maxima for R8-R13 and R∞: R8, 456
nm; R9, 466 nm; R10, 473 nm; R11, 480 nm; R12, 485 nm;
R13, 489 nm; R∞, 555 nm. The latter value for the absorption
maximum of R∞ is quite insensitive to a change in the number
of rings down to 50 (λmax ) 538 nm), 40 (λmax ) 535 nm), or
even 30 (λmax ) 526 nm), any one of which numbers may be
a realistic value for the number of thiophene units in a real
polythiophene
For R7, the spectra and triplet parameters were obtained using
the third harmonic of a Q-switched Nd:YAG laser where spectra
and ∆A changes were obtained. The ꢀT was obtained by the
partial depletion technique21 using benzophenone in benzene
as a reference (actinometer) to measure φT (ꢀ ) 7200 at 530
nm, φT ) 122,24). When the triplet yield was measured, special
care was taken to have optically matched dilute solutions (A ≈
0.2 in a 10 mm cell) and a low laser energy (e1.8 mJ) to avoid
multiphoton and T-T annihilation effects. The region moni-
tored for R7 was at the depletion minimum of 450 nm ((10
nm), which appeared to be outside the T-T absorption.
Nonetheless, the φT obtained should be an upper limit, since, if
the triplet absorbs here, the ꢀT obtained will be too small. In
our case a value of 87 000 M-1 cm-1 ((15%) was obtained.
The φT was 0.60 ( 15%.
The singlet oxygen yields (φ∆) of Rn’s were determined by
1
measuring the emission of O2 in air-equilibrated benzene and
acetonitrile solutions with a germanium diode detector. The
luminescence emerging from the cuvette was passed through a
filter combination (Glenn-Creston, cutoff 1050 nm, and Kodak
Wratten 87c gelatin filter, cutoff 870 nm) and collected by a
germanium diode (Judson J16 85p, 5 mm diameter). The
detection system was at right angles to the excitation beam.
After amplification with a homemade amplifier (100 MHz, 14
dB), the output was fed into a Tektronix DSA 602 digital
sampling analyzer. The amplified signal extrapolated to zero
time (in mV) was plotted as a function of laser dose and
compared with that of a reference, phenalenone, where φ∆ is
0.98.25 The φ∆ of the Rn’s was obtained from a ratio of the
slopes and the known φ∆ of phenalenone.
Figure 4 shows plots of ꢀ vs n and 1/n for R1-R7. If you
plot ꢀ values for R3-R5 vs 1/n and n, both plots are very linear.
If you then use these to predict the extinction coefficients of
R6 and R7 by extrapolation, the results are excellent from the
1/n plot (ꢀ of R6 predicted to be 47 400 compared to the
experimental 47 910; ꢀ of R7 predicted to be 50 800 compared
to the experimental 50 470). On the other hand, extrapolation
of the n (vs ꢀ) plot gives 63 400 for R7 vs 50 470 found. This
is a warning that even though there may be a linear relationship
over a restricted basis set, it may in fact not permit any accurate
predictive capability for other members of higher or lower Rn.
We will see later that this is true for other kinds of considerations
Results
Table 1 presents comprehensive photophysical data of R1-
R7 (di-n-butyl), including singlet oxygen quantum yields (φ∆)
in benzene. Table 2 contains much of the same photophysical
data, except for some φ∆ and φT values, with four additional
solvents. In most cases R6 could not be included because of
its limited solubility. Table 3 presents data regarding absorption

18686 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
TABLE 2: Photophysical Properties of r-Oligothiophenes for n ) 1-7 in Various Solventsa l
Methylcyclohexane
cpd
φF
τF (ns)
kF0
kNR
R2
R3
R4
R5
R7
0.014
0.057
0.18
0.33
0.32
e0.1
0.19
0.49
0.83
0.76
g0.14
0.30
0.37
0.39
0.42
g9.9
5.0
1.7
0.81
0.90
Ethanol
cpd
φF
φF (77 K)
τF (ns)
0.046b
kF0
kNR
∼21
5.0
φT
1.0
0.96
0.70
kIC
kISC
τT (µs)
R2
R3
R4
R5
R7
0.014
0.054
0.18
0.32
0.29
0.04
0.12
0.18
0.34
0.28
0.35
0.28
0.34
0.36
0.35
0.10
∼21
100
91
43
0.19
0.52
0.88
0.82
0.03c
0.23
5.1
1.3
1.6
0.77
0.87
20
e0.6b
0.13
0.73
Acetonitrile
cpd
φF
τF (ns)
kF0
kNR
φT
φ∆
kIC
g0.6
0.24
0.27
0.011
0.011
kISC
τT (µs)
R2
R3
R4
R5
R7
0.013
0.056
0.16
0.33
0.30
e0.1
0.18
0.48
0.90
0.90
g0.133
0.311
0.33
0.37
0.33
g10.1
5.1
0.93
0.90
0.71
0.63
0.98
0.74
0.72
0.50
g9.5
5.0
1.5
0.70
0.67
124
62
40
1.8
0.74
0.78
20
e0.6b
Dioxane
cpd
φF
τF (ns)
kF0
kNR
φT
0.94
0.93
0.67
0.6d
kIC
kISC
20
4.4
1.4
0.73
τT (µs)
R2
R3
R4
R5
R6e
R7
0.017
0.066
0.18
0.36
0.41
0.34
0.046b
0.21
0.49
0.82
1.0
0.37
0.31
0.36
0.44
0.41
0.40
21
4.5
1.7
0.78
0.55
0.78
0.93
0.019
0.31
0.05
146
108
48
29
0.85
e0.6b
0.07
0.71
g15
a Solutions degassed with nitrogen or argon by bubbling. k’s are 109 s-1 and R7 is the di-n-butyl-substituted R7. The lifetime of thiophene (R1)
is ∼1 µs (by sensitization from xanthone). b Assumes value in benzene. c Assumes (1 - φT - φF) ) 0.005. d Assumes average of values in acetonitrile
and benzene. e R6 is “not soluble” in methylcyclohexane, acetonitrile, and ethanol.
TABLE 3: Absorption, Fluorescence, Fluorescence Rate Constants and Triplet-Triplet Data in Dioxanea at Room Temperature
cpd
abs max λmax/nm
ꢀ/mol-1 dm3 cm-1
k0F(SB)/109 s-1
k0F/109 s-1
fluoresc maxb
T-T maxa
R1
231
8340c
0.33d
305
0.023d
R2
303
12440
0.22e
0.38e
0.31
0.36
0.44
0.41
0.40
362
385
R3
R4
R5
R6
R7h
354
392
417
436
441
22080f
31560g
42670
47910
50500
0.30
0.34
0.38
407, 426
437, 478
482, 514
502, 537
522, 560
460
560, 700
630
685
720
0.39
a T-T maxima in other solvents vary by only 10 nm at the maximum. The R1 datum is with acetonitrile. b The italicized wavelength is the band
maximum. The other one is another distinct band (see Figure 4). The R1 does not fluoresce but shows phosphorescence (see text). c Assumes same
value as for ethanol. d First value using Fo¨rsters equation instead of Strickler-Berg (SB) with integration from 38 000 cm-1 (263 nm) to 48 000
cm-1 (208 nm). Second (lower) value from integration over entire first absorption band (same wavelength/wavenumber) limits but different ν0
(absorption maximum value). e First (upper) value using a τF from cyclohexane (0.081 ns) and second one using τF value in benzene of 0.046 ns
with a different instrument. First value has more potential error because normal τ lower limit is g100 ps with the instrument used. f ꢀ(benzene) )
22 540 M-1 cm-1, ꢀ(acetonitrile) ) 24 215 M-1 cm-1 and ꢀ(ethanol) ) 23 590 M-1 cm-1 at their maxima of 354, 352, and 355 nm, respectively.
g ꢀ(acetonitrile) is 34 700 at its maximum. h For the di-n-butyl-substituted R7.
as wellsfor example, φF and φT of R3-R5 vs n if used to predict
R2, R6, and R7.
(100-200 cm-1 ) in the fluorescence maxima between room
temperature and 77 K. This is in marked contrast to that of
absorption (see above). Note also that there can be a switch
between the first two bands regarding which one is the
maximum at room temperature as the number of rings increases
(from the second band to first band from R4 on) and between
room- and low-temperature (77 K) bands for a given Rnsfor
example, see R4 in particular. We have also found an excellent
correlation between the energy of the first (and second) band
of fluorescence vs 1/n (or 1/N).
The room temperature solution fluorescence of R2-R7 is
shown in Figure 5. Note that there is definitely more resolution
than for absorption. Table 3 gives representative fluorescence
spectra maxima in dioxane. As in the case of absorption, there
is little solvent effect on fluorescence. Note in Figure 5 at 77
K (as well as in ref 1) that there is some considerable increase
in vibrational resolution compared to that at room temperature,
particularly noticeable for R2, less for R3, and continuing to be
less as the number of rings increases to R4 and on to R7. Of
notable importance is the fact that there is almost no red shift
We were unable to see phosphorescence in three different
laboratories for n ) 2 or n ) 3 at 77 K in various glasses

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18687
Figure 1. Room temperature absorption spectra of R1-R7 in dioxane.
Figure 4. Extinction coefficients (ꢀ) in dioxane vs 1/n (O) and n (9).
Figure 2. Absorption and fluorescence at room temperature (‚‚‚) and
fluorescence and fluorescence excitation spectra (-) at 77 K for R3
and R7 in ethanol.
Figure 5. Room- (- - -) and low-temperature (77 K) (-) fluorescence
spectra of R2-R7 in ethanol, except for R6 where the solvent is
2-methyltetrahydrofuran.
Figure 3. ET -T (max) b, ES -S (max) 9, and ES -S (0-0) 1 vs 1/n.
1
n
0
1
0
1
triplet energies could not be the same. Furthermore, there is
some published evidence that the lowest triplet of R3 is near
720 nm (0-0)30 and we have evidence that it may be at still
lower in energy at ∼788 nm (0-0) based on photoacoustical
experiments. Presumably, R3 could possibly have a maximum
at such a wavelength (826 nm), but then surely the polymer
must be at considerably longer wavelength.
including an aliphatic hydrocarbon, ethanol, and one containing
10% ethyl iodide, and others could not see phosphoresecence
for R3 and R4.1,26-28 Particularly, n ) 2 (as well as n ) 3)
emission(s) should be quite easy to observe even with a quantum
yield of 10-3 (or even lower), since they are expected to emit
in a spectral region (500-700 nm) easily detectable by sensitive
photomultiplier tubes. Also, no phosphorescence was observed
for dodecyl-substituted R6, R7, and longer ones.27 However,
it has been reported29 that R3 (in undefined environmental and
temperature conditions) showed a sharp, single band phospho-
rescence at 826 nm, which was stated to be identical with that
in shape and energy of poly(3-hexylthiophene) in a thin film at
18 K. This phosphorescence emission from R3 seems very
strange particularly from the point of view that it is at the same
wavelength as a multithiophene component polymerssurely the
Very interestingly, we did see a weak phosphorescence of
thiophene, R1, with a maximum at ∼430 nm and a 0-0 band
near 362 nm (27600 cm-1). Others31,32 found the lowest S0 f
T1 absorption by electron impact to have an onset near 370-
375 nm. Thus, there is excellent agreement for the energy origin
(0-0) of the lowest triplet state being very close to 27 300 cm-1
.
This would give a singlet-triplet (S1-T1) separation of close
to 12 300 cm-1. For R3 using 720 nm30 (∼13 900 cm-1) as
the origin of T1, there is a S1-T1 separation of 10 800 cm-1
,

18688 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
Figure 6. Triplet-triplet absorption spectrum of R1 at room temper-
ature in benzene.
Figure 8. φF (dark symbols) and φT (open symbols) vs n for R2-R7
in four solvents.
Figure 7. Triplet-triplet absorption spectra of R2-R7 at room
temperature in benzene.
-1
Figure 9. kF (3), kISC (0), and kIC (O) vs 1/n for R2-R7 in benzene.
whereas using 788 nm (∼12 700 cm
), the separation is
12 000 cm-1, close to that of thiophene.
number of rings except perhaps for R2. Also, for R3, k0F (in
Table 2) seems somewhat different (lower) than for R4-R7,
except for perhaps benzene and acetonitrile. The k0F values in
benzene generally appear larger than in the other solvents
(Tables 1 and 2). Despite this, it is generally valid that k0F is
quite constant as a function of both solvent and the number of
rings. The kNR Values show a clear decreasing trend with
increasing number of rings (Tables 1 and 2), but apparently,
kNR attains a nearly constant value from R5 to R7. Since we
have all the photophysical data required, we can further
determine both kISC ) φT/τF and kIC ) (1 - φF - φT)/τF, which
are shown in Tables 1 and 2. It is clear that kISC decreases
with increasing number of rings, reaching a constant value from
R5 to R7. The trend in kIC is much more difficult to discern,
but from R2 to R5-R7, there is a decrease in kIC with the data
for acetonitrile, while the other solvents produce a scatter on
the way from R2 to R5-R7. All the above rate constants are
illustrated in Figure 9. The reason for the scatter of kIC is that
kIC depends on the difference 1 - φT - φF and φT + φF is
close to 1 for R2 and R3. Moreover, the error in φT is (15%
so that the kIC value can be very sensitive to this (where φT )
0.90-0.99 and a 2% change in φT results in a minimum of a
300% change in kIC). This is not true for kISC where the change
expected in the kISC is essentially no more than about the same
as the error in φT (φT ) (15%, only ∼7% error in τF).
Triplet-Triplet Absorption. Our triplet-triplet absorption
data are shown in Figure 6 for R1 and in Figure 7 for R2-R7.
Table 3 gives representative spectral data in dioxane. As before,
there is little effect of solvent on these values. Note again the
shift of the absorption maximum to progressively longer
wavelengths as the number of rings increase. We also consider
this point in Figure 3 where the energy of the T-T maxima
are plotted as a function of 1/n. Thiophene has a weak shoulder
at a wavelength longer than that of the maximum, near 360
nm, and R3 and R4 clearly have band shoulders (bands for R4)
at wavelengths both shorter and longer than that of the main
maximum. Also, the onset of the 460 nm band maximum of
R3 is at an unusually long wavelength, essentially equal to that
of R5 (where the maximum is at 630 nm). It appears that for
R3 there must be more than one T f Tn transition in the broad
absorption region of ∼700-460 nm, and probably for R4 also.
The extinction coefficient of the principal maximum progres-
sively increases as n increases. Some other data exist for some
of the individual oligomers.1,4,26-28,30 For different alkyl-
substituted R9 and R11, T1-Tn maxima occur at 774 and 805
nm, respectively.27 The longer (dodecyl) alkyl chain substitu-
tions cause a blue shift in the maxima compared to shorter ones
(butyl).
Quantum Yields, Lifetimes, and Rate Constants. The φF
and φT values are remarkably solVent independent (Tables 1
and 2), but φF clearly increases and φT clearly decreases as the
number of rings increase and both become essentially constant
from R5 to R7 (also see Figure 8). The τF values are also quite
solVent independent, but clearly, the τF increases as the number
of rings increase, becoming essentially constant from R5 to R7.
The k0 ) 1/τ0 values, where τ0 ) τF/φF, derived from φF and
Our lifetimes of fluorescence data are given in Tables 1 and
2, and other fluorescence lifetime data also exist for some of
the oligomers. Data on R3-R62-4 are within a few percent to
10% of ours.2 Other data33 for R3 and R6 are ∼30% shorter
than ours (theirs in dichloromethane), and for others34 the R3
is some 30% longer than ours while R4 is close to ours. The
F
F
F
τF data in Tables 1 and 2 are quite constant as a function of the

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18689
R5 in rigid poly(methyl methacrylate) has τF ) 0.9 ns,4 similar
to ours in five solvents.
TABLE 4: Relative Stabilities of Cis and Trans Conformers
of r2 by Various Theoretical Approximationsa
method
result (kcal/mol)
t < c, 1.16 ( 0.13b; 0.2 41, 32, 39, 40
ref/footnote
Recently, others have measured some photophysical param-
eters but there is no complete set even on any one compound;
commonly, data on R2 are missing, and there is no data on R7.
Our φF and τF data are given in Tables 1 and 2. The φF values
of R3-R6 have been determined by Chosrovian et al.2 and
Colditz et al.3 (R2-R6), and generally they are within 10% of
our values. The former authors believed that kNR was largely
due to S1 f S0 internal conversion. However, this was based
on φT values of approximately 0.2 for R3 and R4, which are
certainly not correct (see later discussion). Others3,4 have
measured φF of R2-R6, and except for R2, these were also
within 10% of our values. Moreover, φF data33 are available
on R2-R4 and R6 and all agree with ours except for R6 (which
is lower than ours and that of others given above). In addition,
some data are available on the scattering of compounds: φF of
R3 is similar to ours and others26,30,34 but R4 is noticeably
lower34 than ours and that of others2-4,26 (and τF’s are reasonably
similar to ours).
exptl
MM2
MOPAC
t < c, 1.04
c < t, 0.46
t < c, 0.16
t < c, 0.25
t < c, 1.3
t < c, 1.20
t < c, 0.64
t < c, 0.8
this work
this work
this work
54
c
d
e
AMPAC/AM1
AMPAC/AM1
ab initio STO 3G
ab initio STO 3G
ab initio 3-21G*
ab initio RHF/DPZ
plus electron correlation t < c, 0.4
ab initio 3-21G
ab initio 3-21G, 631
Columbus ACPF gradient
f
f
f
t < c, 1.9
t < c, 0.6-0.7
t < c, 0.5
40
g
a See Table 5 for more details on our calculations regarding other
R-oligothiophenes and their conformers. b See text for a possible caveat
regarding whether this truly represents the ∆H of cis and trans
conformers (ref 41). c Bredas, J. L.; Street, G. B.; Thiemans, B.; Andre,
J. M. J. Chem. Phys. 1985, 83, 1323. d Jones, D.; Guerra, M.; Favaretto,
L.; Modelli, A.; Fabrizio, M.; Distefano, G. J. Phys. Chem. 1990, 94,
5761. e Distefano, G.; Colle, D.; Jones, D.; Zambianchi, M.; Favaretto,
L.; Modelli, A. J. Phys. Chem. 1993, 97, 3504. f Quattrocchi, C.;
Lazzaroni, R.; Bredas, J. L. Chem. Phys. Lett. 1993, 208, 120.
g Kofranck, M.; Kovar, T.; Lischka, H.; Karpfen, A. J. Mol. Struct.
1992, 259, 181.
Triplet yields for R3 of 0.2,26 g0.9,30 and 0.9535 exist vs our
values of 0.90-0.96 (Tables 1 and 2). Based on all available
data, a φT for R3 of 0.2 cannot be correct. For R4, a value of
∼0.226 exists for φT vs our values of 0.63-0.73 (Tables 1 and
2). On the basis of our φT data and φ∆ for R4 (0.69),4 we believe
the value of ∼0.2 for φT of R4 is not correct.
TABLE 5: Relative Stability of the Conformers All-Cis and
All-Trans r-Oligothiophenes (this Work)
In Tables 1 and 2 our triplet state lifetimes are shown. Some
other data also exist: for R3, a value of 30 µs in methanol at
zero laser dose and concentration has been given.28 Nonetheless,
this is much shorter than our values ranging from ∼62
(acetonitrile) to 108 µs (dioxane) (Table 2). For R3 and R4,
others26 report a value of 57 (R3) and 45 µs (R4) in ethanol vs
our values of 91 (R3) and 43 µs (R4) in the same solvent. The
lifetime of R5 in dioxane in nitrogen has been given4 to be 7.7
µs (0.24 µs in air), which is considerably shorter than our value
in dioxane and all other solvents (Tables 1 and 2). Based on
induced dichroism decay, the decay of the triplet of R3 in a
viscous medium has been given to be a few nanoseconds.36
Presumably, oxygen was present, but nonetheless, this is an
unusually short lifetime for a highly viscous medium even in
the presence of oxygen. Lifetimes of 1- to 3-dodecyl-substituted
R6, R7, R9, and R11 were determined in frozen 2-methyltet-
rahydrofuran at 80 K by monitoring the photoinduced absorption
assigned as T1-T2 absorption.27 Lifetimes were 380-470 µs
for R6, 350-380 µs for R7, 300 µs for R9, and 250 µs for R11.
The trend to shorter lifetimes with increasing number of rings
is the same as what we find for R2-R7 (Tables 1 and 2). An
interesting case is that of R2 in a seeded free-jet expansion where
the triplet lifetime has been given as 550 ns.37 This is extremely
short compared to our values of 100-146 µs (∼225-fold) at
room temperature in fluid solution. This also seems to be very
short for a triplet of an isolated, cold molecule (of R2) (see
later discussion).
cpd
methoda,b
rel stability
∆H cis-trans (kcal)
R2
MOPAC
MM2
AMPAC/AM1
exptl
c < t
t < c
t < c
t < c
0.46
1.04
0.16
1.16 ( 0.13;c 0.2
R3
MOPAC
AMPAC/AM1
MM2
MNDO 93, AM1
MNDO 93, PM3
DGauss-LDF
DGauss-NLSD
ab initio STO 3G
ab initio 6-311 G
ab initio 6-31 G(d)
cc < tt
tt < cc
tt < cc
tt < cc
cc < tt
tt < cc
tt < cc
tt < cc
tt < cc
tt < cc
0.89
0.35
2.12
0.42
0.83
1.39
1.18
2.41
2.22
1.47
R4
R5
AMPAC/AM1
ttt < ccc
2.73
MOPAC
AMPAC/AM1
MM2
MNDO 93, AM1
MNDO 93, PM3
DGauss·LDF
DGauss·NLSD
ab initio STO 3G
cccc < tttt
tttt < cccc
tttt < cccc
tttt < cccc
cccc < tttt
tttt < cccc
tttt < cccc
tttt < cccc
1.54
1.41
8.4
1.72
1.52
6.46
5.79
4.81
a Gaussian 92 code was used for all ab initio calculations. For one
of them, 6-31G(d), a polarization function (d-orbital) was added for C
and S atoms. For DGauss, double-ú potential basis sets were used for
orbital functions and a nonlocal spin density (NLSD) correction was
used in one of the calculations with DGauss. b For MNDO 93 (AM1
or PM3) and ab initio STO 3G calculations, initial input geometries
were from MM2. For DGauss, input geometries were planar. For ab
initio 6-311 G and 6-31G(d), input geometries were from STO3G
optimized geometries (planar). A second set of calculations using
MNDO 93, AM1 geometry as input gave ∼0.2 kcal difference from
planar input but did not change the relative stability. c See text for
possible caveat regarding whether this truly represents the ∆H of cis
and trans conformers.
Calculations, Geometry, Stability of Conformers, and
Transition Energies. We have done an extensive number (10)
of different types of calculations and at a number of levels
including three at the ab initio level regarding the relative
stability of the all-cis and all-trans conformers of R2 (Table 4),
as well as R3-R5 (Table 5). In Table 4, for R2 we compared
a number of calculations of the literature with those of ours
and proposed ∆H values between cis and trans. We did
significantly more and higher level calculations on R3 and R5,
since their was very little or no significant data on these latter
R-oligothiophenes. It can be seen in Table 4 that for R2, the
two experimental ∆H values are quite different from one
another. This potentially complicates the comparison with
theory. Nonetheless, it appears that the trans conformer of R2,
albeit twisted around the inter-ring bond, is lower in energy.32,38-40
Note that MOPAC and MNDO93,PM3 incorrectly predict the

18690 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
TABLE 6: Calculated and Experimental First Transition
cis conformer as the more stable one. Also, the presumably
less sophisticated MM2 approach predicts ∆H in the same
vicinity as several of the other methods including the ab initio
level calculations.
Energies of Some r-Oligothiophenesa
transition energy,b nm
theory (f)
exptlc
geometry
opt method
cpd
conformer
Some literature data propose that for R2 in the ground state,
both cis and trans conformers coexist in an R2-seeded free-jet
expansion and that the inter-ring twisting is about the same in
both conformers.41 The ∆H between cis and trans was given
to be 1.16 ( 0.13 kcal/mol (406 ( 46 cm-1) with the trans
conformer being the more stable. On the other hand, different
but parallel supersonic jet experiments42 involving R2 and using
the same emission techniques, plus hole burning, point to the
existence of a torsionally twisted equivalent pair (a double
minimum) around the trans conformer in the ground state in
place of identifying the coexistence of cis and trans conformers
as done earlier.41 The equilibrium structure (trans) existed with
a twist of ∼21° from planarity and with a barrier of ∼25 cm-1
between the two minima (at 21°). In the first excited state by
contrast, R2 was found to be trans and planar with a deep, steep
single minimum torsional potential well around the equilibrium
structure. As we will see later, our spectroscopic data strongly
indicate a similar situation in the first excited state. The
difference in the supersonic jet experiments41,42 makes it difficult
to be certain that the proposed41 ∆H (1.16 kcal) between ground
state cis and trans is actually appropriate.
di-
tri-
MM2 (S, sp2)
t
c
t
c
t
c
348 (0.70) 302 (303)
345 (0.72)
336 (0.65) 302 (303)
335 (0.66)
MOPAC
AMPAC/AM1
MOPAC
345
325
302 (303)
all-t
all-c
all-t
all-c
all-t
all-c
all-t
all-c
all-t
all-c
all-t
all-c
all-t
398 (1.07) 350 (354)
397 (1.09)
AMPAC/AM1
AMPAC/AM1
373
354
398
333
350 (354)
tetra-
386 (392)
quinque-d MM2 (S, sp2)
MOPAC
431 (1.65) 410 (417)
480 (1.68)
422 (1.59) 410 (417)
473 (1.60)
AMPAC/AM1
405
355
410 (417)
STO 3G
405 (1.56) 410 (417)
a See Table 7 for calculations on more Rn’s and by different methods.
Here, ZINDO is used for theory. All are this work. b See Table 7 for
other transitions for these and other R-oligothiophenes. c Values in
methylcyclohexane. Values in dioxane are in parentheses. d In all cases
except STO3G there were finite values of twist for the S-C-C-S
dihedral angle while for STO 3G there was no twist (totally planar).
For R3, the higher level calculations (DGauss and several
levels of ab initio) predict the tt conformer as the more stable
as does the MM2 approach and with a fairly comparable ∆H
(Table 5). If all three conformers are calculated (tt, ct, cc), then
tt is still the most stable except by MOPAC and MNDO93,-
PM3. The ∆H between tt and cc in general seems to be in the
2.0 kcal/mol range ((0.4) for ab initio methods and ∼1.2 kcal/
mol for the DGauss approach (lower for AMPAC and MNDO,
AM1) (Table 5). Again, MOPAC and MNDO93,PM3 very
likely predict stability in the wrong order, cc < tt.
For R4 and R5, all methods except MOPAC and MNDO93,-
AM1 predict the all-t conformer as the most stable. The ∆H
beween the all-trans and all-cis conformer varies according to
the calculation method (Table 5). If we consider all possible
conformers of R5 (10 of them), the all-trans conformer is still
the most stable where generally the greater the number of
adjacent cis pairs, the higher the relative energy.
geometry (no twisting) for the all-trans conformer, as well as
for the all-cis one (interestingly enough, essentially all the
calculation methods give near planar geometry for the all-cis
conformer). Some X-ray data on R4 and R645 give evidence
for an all-trans, nearly planar (0-10° twist) geometry. There
would be no reason to believe R5 should be different.
In Table 6 we present the lowest transition energy predicted
for the all-trans and all-cis conformers of R2, R3, and R5 and
a comparison with experimental data. It can be clearly seen
that for R5, all three methods of geometry optimization show
that the all-trans conformer spectral data are very much closer
to experimental results than the all-cis. The best agreement is
from STO3G, and recall that this was the one that gave no
twisting of the inter-ring bond. We believe this is perhaps the
strongest evidence that R5 is all-trans with very little inter-ring
twisting, and a similar situation should be true for R3.
Table 7 gives experimental data for the discernible transitions
in R1-R7, extinction coefficients, and theoretical predictions
for a number of R-oligothiophenes.
We have determined the dihedral angle of all the conformers
of R3 and for the all-cis and all-trans conformers for R5. All
ab initio STO3G, DGauss (with planar input), and MM2 predict
either planar (ab initio STO3G) or near planar geometry
(essentially no twisting) for the cc and tt conformers of R3.
This is also true for MOPAC, but remember, it predicts the cc
conformer as the more stable vs tt for the other calculations
considered here (and all the rest except MNDO93,PM3). X-ray
data of trithiophene (R3)43 give a clear indication that the tt is
the more stable, that is, this is the conformeric form in the crystal
with inter-ring angles of about 6°-9°. Also, a dibutyl-
substituted trithiophene has a tt configuration44 with a higher
angle of inter-ring twist than for R3 itself (not unexpected).
However, the barrier to rotation was given to be 19.7 kcal/mol
in the ground state and 4.2 kcal/mol in the lowest excited singlet
Discussion
State Order. One of the most important questions for the
R-oligothiophenes (hereafter denoted Rn) concerns the state
order: does the state order change with increasing n, and if so,
when?
Based on earlier spectroscopic date of R,ω-dithienylpolyenes,
calculations, and some extrapolated data, the (2)1Ag state was
assigned below the (1)1Bu for oligothiophenes of n ) 6 and
greater.7 For R2, two-photon spectroscopy5 has located the
1Ag above the Bu. Theoretically, it has been predicted8 that
1
1
1
1
state using H NMR. This seemed quite incompatible with a
the Ag state would be below the Bu state for R2-R4 with
crossing between R4 and R5. However, reduction in the degree
of CI (with no rational justification) now resulted in the Bu
“single” inter-ring bond in the ground state and a more double
inter-ring bond in the lowest excited singlet state. Very recently,
reinvestigation44 of the same molecule at higher resolution now
suggests a ground state barrier of 0.5-1 kcal/mol with the
excited state barrier similar to that reported earlier.44 These
new data are now compatible with our findings (see later
discussion). In the case of R5, only STO3G gives planar
1
1
state being lowest from R2 toR7 (others predicted Ag to be
lowest for R6 and longer chain). Also, more recent calculations9
for R1-R4 predicted that the 1Ag state (0 level) would be below
1
1
the Bu state (0 level). They believed the Ag state observed
was the (3)1Ag state. Multi-CI ab initio calculations on

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18691
TABLE 7: Experimental (with E) and Theoreticala Energies
(with f) for All-Trans Conformers of Some
r-Oligothiophenes
excited state for R2-R7 is of allowed character and is (1)1Bu
and not (2)1Ag. This was noted earlier1 for R2-R5 and R7 in
benzene, but we can now include R6 as well and, moreover,
safely conclude that this order is independent of the solVent for
R2-R7.
Relative to the question of whether the (1)1Bu and (2)1Ag
state cross and, if so, when, it is difficult to answer when, but
surely they will cross, since the two states are getting closer as
n increases (for Rn). Earlier, on the basis of only two data points
for extrapolation of the (2)1Ag energy, we believed crossing
might occur between n ) 9 and n ) 11. However, the (2)1Ag
data were from different phases of different Rn’s, so we no
longer believe the crossing will occur at these low n values.
Others3 also note that the states will cross, but based on some
CNDO/S and FEMO models, they predicted crossing would be
beyond n ) 50.
exptlb
theoretical
λ
(nm)
ꢀ
λ
(nm)
cpd
(M-1 cm-3
)
f
thiophene 232
8340
227
211
0.26 (PPP)d
0.27
170, 169 1.5
di-
tri-
303 (302)c
247
12440
6200
311
257
204
342
276
0.83 (PPP)d
0.39
205
0.19
354 (350)c
22100
8600
1.40 (PPP)d
0.02
252
250, 245 0.24
221
205
234
222
0.15
0.22
quater-
391 (386)c
308 (sh)
252
31560
4300
10700
Temperature Effect on Absorption and Ground-State
Geometry. Recall earlier in the Results section that we noted
the very significant red shift (∼1600 cm-1) of the absorption
maxima of R2-R7 upon going from 298 to 77 K (see Figure 3
and ref 1). If, however, we look at the 0-0 band (as determined
by the overlap of absorption and fluorescence spectra at 298
205
quinque-
417 (410)c
∼340 (sh)
290
42650
405
286
1.56 (ZINDO)
0.12
7300
256
207-214
12700
256, 260 0.06
and 77K), the shift is greatly reduced to only ∼200 cm-1 for
236
196
0.10
1.71
-1
R2 and R3, ∼300 cm
for R4, and ∼400 cm-1 for R5 and
sexi-
436
314
47910
R7. We interpret the absorption results to mean that as the
temperature is lowered, there is a greater average planarity
among existing conformers; said in another way, the virtual
molecule at 77 K is more planar than one at 298 K. That is,
the potential energy minimum in the ground state becomes more
vertically aligned to that of the excited state such that the vertical
transition now is to a lower vibronic level of the excited state,
yet the actual state energy difference (0-0’s) has changed
relatiVely little. Usually, the red shift seen in the absorption
upon going to low temperature is ascribed to increased planarity
of the ground state with subsequent greater conjugation,
lengthening of the π-electron path, and decreased energy gap
between S0 and S1. However, based on our spectroscopic data
as described above, this is not what principally occurs. It does
appear that there is some small effect on relative state energies
for R5-R7 and perhaps for R4.
259
septa-
441 (433)c
361 (sh)
325
50460
10900
18200
282, 298 (sh)
256
205
a PPP model with geometries from an MMX (force field) approxima-
tion (Serena Software, Bloomington IN) for thiophene, bithiophene,
and trithiophene. For quinquethiophene data are from ZINDO using
geometry obtained via STO 3G. b In ethanol for thiophene, all others
in dioxane. See Experimental Section for methodology for determining
ꢀ. c Values in parentheses are for the compound in methylcyclohexane.
d Parameters for PPP calculation were IP(C) ) 11.20, γ(C) ) 11.10,
IP(S) ) 17.20, γ(S) ) 10.00, âCS ) -1.8, âCC ) -2.4.
1
thiophene (R1) place the (2)1A1 state below the B2.46 More
There are clear indications that R2 is twisted in both the cis
and trans forms, that they coexist, and that the trans is more
stable than the cis. Furthermore, the absorption (and fluores-
cence) of R2 undergo the most dramatic increase in vibrational
resolution as a function of temperature compared with all other
Rn’s (see Figure 5 and ref l). However, the red shift of the
absorption maxima at low temperature is essentially the same
as for R3-R5 (∼1600 cm-1 ) and for the others, while the
fluorescence maximum shows essentially no shift (e200 cm-1).
We interpret all the absorption results to mean that for R2 the
ground state becomes more planar, as it is for all Rn’s, and a
greater population of a more singular conformer exists for
R2 compared to other Rn’s at low temperature.
Temperature Effect on Fluorescence and Excited-State
Geometry. A negligible shift of the maximum of fluorescence
(100-200 cm-1 ) as well as the 0-0 band occurs upon lowering
the temperature. We believe that based on the Franck-Condon
forbidden shape of the first absorption band, an obvious change
in geometry has occurred. This could be consistent with a
quinoidal-like contribution to the structure of the excited
stateswe will support this belief more fully shortly. At room
temperature, solvent shell relaxation around the now quinoidal-
like excited state can occur, so emission occurs from the
equilibrium excited state to a twisted ground state. However,
at 77 K in a rigid matrix environment solvent shell relaxation
1
recently, ab initio calculations11 (multi-CI) for R2 give a Bu
1
1
state below the Ag state and, in fact, predicts two Bu states
below the lowest 1Ag. Very recent calculations3 used CNDO/S
both with single and single plus double CI to calculate transition
energies and state orders. Good results compared with experi-
mental results were obtained with single CI but not with single
plus double CI. This problem with double CI and of rational-
izing the results when ignoring double CI has been a major
problem (see refs 3, 7, and 8).
From our experimental data the k0F values in benzene (Table
1) for six oligothiophenes, R2-R7, show that the k0F is
remarkable constant. Furthermore, k0F is essentially constant in
all four additional solVents in Table 2, especially for R4-R7.
Moreover, the k0F values for all the R-oligothiophenes in all the
solvents only vary from (0.28-0.45) × 109 s-1, most values
being in a narrower range still (others3 found a range of (0.25-
0.33) × 109 s-1 in dioxane). These correspond to τ0F values of
2.3-3.5 ns, which are clearly typical of emission lifetimes from
allowed π,π* states. Also, as given earlier1, kF0 values calcu-
lated based on the Strickler-Berg (SB) equation over the first
absorption band for R2-R5 and R7 agreed within 0-14% (only
one case was as large as 14%; others were less than 7%) of
those from experiment. The result of all of these considerations
provides experimental conVincing eVidence that the lowest

18692 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
TABLE 8: Atomic Charges and Dipole Moments in the Ground and Excited Statesa
cpd
atom
GS charge
µ (D)
ES (S1) charge
µ (D)
ES (T1) charge
R1
1.84
2.52
S5
+0.02
-0.14
-0.03
-0.22
+0.03
-0.01
C1, C2 (R-C)
C3, C4 (â-C)
R2
R3
0.007
1.84
0.009
1.77
S5, S13
+0.004 (+0.58)
-0.14 (-0.45)
-0.03 (-0.15)
-0.04 (-0.16)
-0.035 (-0.30)
+0.075
-0.15 (+0.38)
-0.01 (-0.35)
+0.015 (-0.09)
-0.025 (-0.11)
+0.05 (-0.29)
+0.04
(+0.53)
(-0.47)
(-0.10)
(-0.20)
(-0.24)
C4, C11 (R-C)
C1, C10 (â-C)
C2, C9 (â-C)
C3, C12 (R-C, bonded)
H7, H14
S5, S20
+0.004
-0.14
-0.14
-0.02
+0.01
-0.014
+0.04
-0.16
+0.04
-0.005
+0.04
+0.04
+0.04
+0.04
C4, C18 (R-C)
C1, C17 (â-C)
-0.03
C2, C8 (â-C)
-0.04
C3, C19 (R-C, bonded)
-0.035
+0.001
-0.04
S14
C12, C13 (R-C, bonded)
C10, C11 (â-C)
-0.04
H7
+0.074
+0.077
+0.076
+0.074
H15
H16
H21
R5
2.11
2.16
S26, S34
+0.02
-0.12
C25, C19, C18, C11,
C12, C4, C3, C32c
C16, C31 (â-C)
C23, C30 (â-C)
C24, C33 (R-C)
H27, H22, H21, H15,
H14, H7, H6, H36
S10, S13, S5
-0.05c
+0.03c
-0.04
-0.03
-0.15
+0.075b
+0.00
+0.01
-0.04
+0.04
+0.02
-0.035
-0.04
-0.14
+0.005
-0.003
C17, C1 (â-C)
C10, C9, C8, C2 (â-C)
a See Figure 10 for structures and numbering of atoms. Charges are formal charges calculated by the ZINDO approximation, and for R2, the
numbers in parentheses are those calculated using the AMPAC approximation using the AM1 Hamiltonian (full SCF and configuration interaction).
b This represents an average of values between 0.074 and 0.077. c R-C, bonded.
cannot occur and emission is from the Franck-Condon excited
state to a “planar” ground state. It is this combination of the
different circumstances that results in essentially no shifting of
fluorescence bands between 298 and 77 K. The above
interpretations regarding absorption (previous section) and
fluorescence behavior are supported in two major ways. One
source of support regards the spectroscopic studies of bithiophene
(R2) in a supersonic jet 42 where there is a clear indication that
in S1 the R2 is trans planar with a single minimum and that a
deep, steep potential well exists. However, in S0, a much
broader potential well exists (and in fact there is a double
minimum, and equivalent twisted forms exist). There is no
reason to believe a parallel situation is not true in general for
all of the Rn’s albeit the ground state may not have multiminima
but that the excited state has a higher bond order for the inter-
ring bonds and is planar quinoidal-like compared to the ground
state. A second source of support involves a consideration of
changes in atomic charges between the S0 and S1 state (as well
as the T1 state, which we will need to consider shortly). Table
8 gives the atomic charge and dipole moments in the ground
state and S1 state for all atoms in R1-R3 and R5 and atomic
charges in T1 for R2, where Figure 10 gives the appropriate
structures and numbering for these four molecules. Note that
in all cases, there is large change in charge on sulfur and the
R-carbon (non inter-ring bonded) between S0 and S1 (Table 8).
In the case of sulfur, the change involves a large addition of
negative charge [∼0.14-0.24e] while for R-carbons that are
Figure 10. Numbering of atoms of various oligothiophenes for
reference to atomic charges given in Table 8.
non-inter-ring bonded, there is a quite large subtraction of
negative charge (∼0.12-0.19e). The bonded R-carbons (to
another R-carbon) still show subtraction of negative charge but
smaller in magnitude (∼0.07e). One obvious important impli-
cation of this is that the lowest energy transition (longest
wavelength one) now clearly established to be
1A f (1)1Bu

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18693
3
1ψCT|H′| ψi
has some charge transfer character. Secondly, with such large
changes in atomic charges there must be an accompanying
change in the geometry of the excited singlet state (as previously
noted, based simply on the Franck-Condon forbidden shape
of the first transition). A quinoidal-like structure is compatible
with these charge changes and totally in harmony with the
change in the ground and excited singlet state potential energy
curves along the inter-ring RC-RC torsional coordinate found
in supersonic jet experiments42 where for R2, the inter-ring RC-
RC bond was clearly much more double bonded in S1 compared
to S0.
where H′ is the spin-orbit (coupling) operator and contains the
atomic spin-orbit coupling factor ú for sulfur, which is large,
363 cm-1. Remember that the decrease in φT (and increase in
φF) is not due to some significant increase in k0, which is
F
essentially constant, or kIC (kNR as the number of rings increases)
but is due to a real decrease in kISC (Tables 1 and 2). It is
clear that as the number of rings is increased, a like number of
heavy atom sulfur centers is also increased. Yet, kISC (and fT)
still decreases. We believe this occurs because of a decrease
in the magnitude of the matrix elements described above,
resulting from a decreased charge transfer mixing of 1ψCT and
3ψi as the number of rings increases. This may result from the
smaller overlap of the electron donor-electron acceptor mo-
lecular orbitals because of the more spread-out nature of the
molecular orbitals proceeding from R2 to R5, which becomes
essentially constant from R5 to R7.
In addition to studying the effect of temperature on absorption
and fluorescence spectral maxima and 0-0 origins, we examined
the dependence of φF on temperature. In Table 2 for ethanol
as the solvent, it is interesting to note that φF for R2 and R3
clearly increase from 298 to 77 K by 2.8-fold and 2-fold,
respectively. However, there is essentially no change in φF for
the other Rn’s (R4-R7). Others35 also found a very similar
change for R3 and, furthermore, by assuming that the only
temperature dependent process was nonradiative, calculated an
activation energy of ∼1.2 kcal/mol. We also examined the
effect of temperature on φT of R3 and found little or no change
over the range 290-140 K. Others35 believed that the activation
energy was for S1 f Tn and that about 55% of the φT arose
from this path.
Absence of Phosphorescence and Triplet State Lifetimes.
Significant inter-ring bond torsional (twisting) coupling to the
ground state could occur, resulting in a large radiationless rate
constant for T1 f S0 with a quenching of phosphorescence
emission. Recall that the lifetime of the triplet state undergoes
a substantial decrease from R2 to R7 in all solvents (see Tables
1 and 2) by 5- to 10-fold. If the lifetimes are considered to be
essentially the radiationless lifetimes, then the radiationless
intersystem crossing T1 Df S0 would strongly dominate and
phosphorescence would be “absent”. In such a case, the
lifetimes would be expected to decrease because of the decrease
in the T1-S0 energy gap (R2-R7) resulting in a significant
increase in the Franck-Condon overlap integrals between T1
and S0 upon which the radiationless rate constants depend.47
We make one additional observation regarding R2. Some
two-photon photoionization experiments37 assign the final state
as a triplet state (related to the dominant photoionization
channel). Earlier in the Results section, we compared some
triplet lifetimes and in the particular case of R2, we determined
τT to be 100-146 µs (solvent dependent) at room temperature
in a fluid solution (Tables 1 and 2). The τT given for the triplet
state of R2 in the supersonic jet experiment above37 was 550
ns. This triplet lifetime seems to be extraordinarily short for a
molecule in a nearly isolated condition presumably without (or
nearly without) external intermolecular interaction (at the usually
expected chamber background pressures) and where, although
it is not strictly possible to define the temperature, a temperature
on the order of 10 K seems reasonable. The lifetime is ∼225-
fold shorter than our triplet lifetimes in a variety of solvents
that were not vacuum degassed (but N2 bubbled) at room
temperature (vs temperature on the order of 10 K for the
supersonic jet experiment), where intermolecular interactions
must be more plentiful than in the supersonic jet experiment.
These considerations make us very doubtful of the assignment
of the T1 triplet state (of R2) as the state to be associated with
the 550 ns lifetime.
Variation of Rate Constants. Geometry and Charge
Transfer Mixing. First of all, recall that based on the k0F and
τ0F values in the fiVe solvents (Tables 1 and 2) and the
correlation with those determined by the Strickler-Berg equa-
tion, the evidence is totally convincing that the 1Bu state is the
lowest for all Rn’s from R2 to R7, and moreover, this order is
solvent independent.
In general, the k0F is quite constant (∼20%) as a function of
the number of rings (R2-R7) and solvent (except perhaps
benzene where it is sometimes greater) while kISC undergoes at
least a 20-fold (vs 20% for k0F) decrease between R2 and R5
and a leveling of kISC occurs from R5 to R7. The trend of kIC
is not so definitive because as noted earlier φT + φF ≈ 1 and
because of the potentially significant error in φT ((15%); the
difference term (1 - φT - φF) required to calculate kIC can
generate notable error in the kIC. However, it is clear that it is
significantly smaller than kISC (Tables 1 and 2). Furthermore,
there is a clear, progressive decrease in kNR (Tables 1 and 2)
from R2 to R5 and than a leveling from R5 to R7 (as for kISC).
A similar trend is apparent for kIC in dioxane (Table 2). On
the basis of the relationship of kISC, kF, and kIC as well as the
fact that φF + φT generally is >0.92, kIC is very small, and
therefore, internal conversion from S1 is small or negligible (φIC
≈ 0.1-0.0).
We believe that the preceding facts can be understood by a
combination of circumstances. There is evidence from two
sources that the first excited singlet state has a quinoidal-like
character that would account for the very small internal
conversion observed, since there would be very little torsional
mode coupling (around the inter-ring bond) to the ground state.
There is obviously significant spin-orbit coupling that has to
be due in part to the heavy atom effect47 of the sulfur.
Nonetheless, φT and kISC show a progressive decrease from R2
to R5 and then a leveling off (while φF shows a progressive
increase from R2 to R5 and then a leveling off from R2 to R7).
We believe that most of the spin-orbit coupling is not due to
the classical heavy atom effect47 but that this coupling is
mediated by charge transfer (CT) mixing involving matrix
elements of the type
Recall that thiophene itself, R1, did show a weak phospho-
rescence but “no” fluorescence (φF e 5 × 10-4). We do not
know whether phosphorescence is weak because φT is low (φIC
is high) or because the nonradiative process from T1 is high.
Nonetheless, phosphorescence does exist at a level greater than
any other Rn so far examined (R2-R4, R6). Given that it would
be reasonable to expect φT to be quite high because of matrix
elements of the type described in the previous section above
(thiophene also has a large amount of charge transfer going from
S0 f S1, Table 8), we would expect φP to be high. Thus, the
reason phosphorescence is weak is very likely because of a

18694 J. Phys. Chem., Vol. 100, No. 48, 1996
Becker et al.
highly efficient nonradiative process out of T1. This is supported
by the short τT (1 µs) compared to those of all other Rn’s.
lifetime was given as shorter than 600 ps. Recall that our work
in solutions clearly defines the lowest excited state of the
oligomers R2-R7 as (1)1Bu.
Singlet Oxygen Formation by Energy Transfer. In Table
1, it can be seen that the φ∆ decreases with an increase in the
number of rings. Note that for R3, the efficiency of triplet
Absorption data for a single crystal of R6 indicated that the
maximum was red-shifted compared to the solution data,45
whereas for oriented thin films, it was blue-shifted.51 Other
studies on thin films of R6 found the excited singlet state lifetime
to be ∼40 ps and a triplet-triplet (sharp) absorption near 790
nm with a lifetime of 4 ns for the triplet state.52 The singlet
lifetime is very short compared to that in solution (24-fold),
but in general φF values are also much lower, so some quenching
mode exists. The short 4 ns triplet lifetime is at least partially
explained by the presence of air.
Thin polycrystalline films of R4-R6 showed transient
absorptions that were assigned as T1 f Tn absorption.53
Lifetimes of 310 ns (R6), 550 ns (R5), and 1.3 µs (R4) were
reported (presumably in the presence of air). These are 40-
50 times shorter than ours in degassed solutions (see Tables 1
and 2). In a crystalline film (at 5-10 Torr) the τT of R6 (∼170
µs) is very much longer than that found above (4 ns)52 and even
much longer than in degassed solutions (∼17 µs, Table 1). The
lifetimes of a film of R7 (∼100 µs) is also much longer than
that in degassed solution (21 µs, Table 1).
1
energy transfer to produce O2, S∆ ) φ∆/φT, is 0.85 while for
R2, R4, and R5 it is near 1. These results clearly indicate energy
transfer is highly efficient from many oligothiophenes to ground
state oxygen. Although we are not totally confident about the
φ∆ for R7 (or φT), it appears to be in the 0.2-0.3 range, giving
a S∆ that is relatively low (∼0.5). This may be because of the
presence of two n-butyl groups preventing efficient encounter
transfer and the fact that the triplet is expected to be quite low
for R7 (9-12 000 cm-1). The situation for R6 is unclear,
although based on φT data in two solvents for R5 and one for
R7, it seems that φT for R6 should be in the 0.5 area. If true,
the S∆ is definitely lower than for the others except possibly
for R7, based on the literature value (Table 1) of 0.364 for φ∆.
There are several literature values for the φ∆ of R3: 0.15,26
∼0.8,1,4 ∼0.8,48 0.7330 (which was believed to be more accurate
than an earlier 0.86 value determined by some of the same
authors), and 0.6 or 0.86 depending on the technique used49
compared with our values of 0.81 in benzene and 0.74 in
acetonitrile. We believe the 0.1526 value is very probably
incorrect for R3. On the basis of all the data available, we
believe φ∆ for R3 is 0.75 ( 0.05.
Summary/Conclusions
1. Based on the k0F of R2-R7 in any and all five solvents,
The φ∆ values of other Rn’s have been given to be4 ∼0.7 for
R4, ∼0.5 for R5, and ∼0.36 for R6, where both R4 and R5 (as
well as R3) are close to our values (see Table 1). Elsewhere,26
a value of 0.24 has been determined for R4. We believe this
value of 0.2426 is quite probably incorrect for R4 based on our
value of φT of ∼0.7. Furthermore, we and others4 find a value
of ∼0.7 for φ∆ of R4 and, of course, φ∆ cannot be greater than
φT as it would be if φT ) 0.24.26 The φ∆ values of some
substituted bithiophenes (acetylenic, olefinic) are known,50 and
these all seem to be quite low, ∼0.05, compared to what we
obtain for bithiophene itself, ∼0.97 (Tables 1 and 2).
the lowest excited singlet state is (1)1Bu (π,π*).
2. There is a significant shift (∼1600 cm-1 ) of the absorption
maxima of R2-R7 upon going from 295 to 77 K, but much
smaller shifts occur for the 0-0 band (200-400 cm-1 ). We
interpret this to be largely due to an increase in the overall
planarity of the ground state among the existing conformers
(or virtual molecule) at 77 K. That is, the potential energy
minima of S0 and S1 are more vertically aligned such that the
vertical transition is to a lower vibronic level of S1 (but the
actual state energy difference is little changed).
3. Significant changes in atomic charge occur on sulfur and
the R-carbons of all Rn’s when comparing S1 to S0 (and T1)
but not when comparing T1 to S0 (R2). These are compatible
with geometry changes between S0 and S1 but not between S0
and T1 (also see items 4-6, below).
Comparison of Solution and Film Data. Recall that there
is great interest in oligothiophenes regarding their application
in nonlinear optical charge storage and molecular electronics
devices. In most of these applications, the solid phase is the
phase of most interest. However, because of film quality,
thickness, homogeniety, and crystallinity, as well as the ambient
conditions, the absorption, fluorescence, and other photophysical
properties truly intrinsic to a particular Rn are not well defined.
Therefore, a study as carried out here can provide intrinsic
properties at the molecular level, and these can then be used as
a reference to better understand and interpret solid state
absorption and photophysical results.
4. On the basis of absorption spectral shape, charge density
calculations of the ground and first excited states, and existing
literature,42 we believe the excited singlet state, S1, has a
quinoidal-like form and therefore is essentially planar while the
ground state is twisted for all Rn’s.
5. There is almost no shift (100-200 cm-1) of the fluores-
cence maxima of R2-R7 (in contrast to absorption). We
interpret this as the result of a combination of circumstances,
including a quinoidal-like planar S1 state, a twisted ground state,
and the presence of (at 295K) or absence of (at 77 K) of solvent
reorientation.
For films of the oligothiophenes, assignment of the lowest
excited state varies. Studies of films and solutions4 showed
that the φF of ultrathin and thin films were considerably less in
general than those in solutions (10-103 times) and that the
absorption maximum was strongly blue-shifted51 (although
absorption could continue to near the same onset). Also,
fluorescence decay curves of films are strongly nonexponen-
tial,51 in contrast to our work in solution. This indicates
considerable inhomogeniety in the surroundings of the oligo-
mers. The lowest excited state was assigned as (2)1Ag for all
oligothiophenes including R2-R83. However, although this
assignment for all was repeated again,51 there was some question
about how unambiguous it was. Others45 have assigned the
lowest excited state for films R6 as (1)1Bu, which was noted to
be some 900 cm-1 below the (2)1Ag (comparing lowest exciton
levels of each exciton band). The fluorescence (radiative)
6. The small magnitude of φIC (e0.1) is interpreted to be
the result of small coupling between modes of the planar
(quinoidal) S1 state and the essentially single-bonded (inter-
ring bond) twisted S0 state.
7. The decrease in kISC as one proceeds from R2 to R7 (with
plateauing at/near R5, is believed to be the result of the decrease
in magnitude of matrix elements of the type
3
1ΨCT′|H′| Ψi
which may be caused by a decrease in the overlap of the electron
donor-electron acceptor orbitals. In any event, the classical
heavy atom effect alone is not able to explain this observation.

R-Oligothiophenes
J. Phys. Chem., Vol. 100, No. 48, 1996 18695
(13) Seixas de Melo, J.; Becker, R. S.; Mac¸anita, A. L. J. Phys. Chem.
1994, 98, 6054.
(14) Strickler, S. J.; Berg, R. A. J. Chem. Phys. 1962, 37, 814.
(15) Fo¨rster, Th. Fluoreszenz Organischer Verbindungen; Vandenhoeck
und Ruprecht: Go¨ttingen, Germany, 1951.
8. The increase in φF with increasing n (R2-R7) arises
primarily because of a decrease in kisc (see item 7 above) and
not from an increase in k0F (kF0 is essentially constant over all
Rn’s).
9. “No” phosphorescence has been abserved by us or
others26-28 for R2-R4, although the φT is very high (0.70-
1.0). This is believed to be the result of a large radiationless
rate constant.
(16) Mac¸anita, A. L.; Costa, F. P.; Costa, S. M.; Melo, E. C.; Santos,
H. J. Phys. Chem. 1989, 93, 336.
(17) Stricker, G. In EffectiVe Implementation of Modulation Functions
in DeconVolution and ReconVolution of Analytical Signals; Bouchy, M.,
Ed.; University Press: Nancy, France, 1982.
10. There is a dramatic decrease in τT from R2 to R7 (g100
to 20 µs). This is interpreted based on the premise that the
lifetimes represent essentially the radiationless ones. The change
in τT would then be expected, since the lifetimes should decrease
as the T1-S0 energy gap decreases because of a signficant
increase in the Franck-Condon factors between T1 and S0 upon
which the radiationless rate depends.
(18) Elisei, F.; Aloisi, G. G.; Go¨rner, H. J. Phys. Chem. 1991, 95, 4225.
(19) Aloisi, G. G.; Elisei, F.; Latterini, L. J. Chem. Soc., Faraday Trans.
1992, 88, 2139.
(20) Dawe, E. A.; Long, D. A.; Land, E. J. J. Chem. Soc., Faraday
Trans. 1, 1977, 73, 1319.
(21) Charmichael, I.; Hug, G. L. J. Phys. Chem. Ref. Data 1986, 15, 1.
(22) Murov, S.; Charmichael, I.; Hug, G. L. Handbook of Photochem-
istry; M. Dekker Inc.: New York, 1993.
(23) Medinger, T.; Wilkinson, F. Trans. Faraday Soc. 1965, 61, 620.
Horrocks, A. R.; Wilkinson, F. Proc. R. Soc., Ser. A 1968, 306, 257.
(24) Hurley, J. K.; Sinai, N.; Linschitz, H. Photochem. Photobiol. 1983,
38, 9.
(25) Wilkinson, F.; Helman, W. P.; Ross, A. B. J. Phys. Chem. Ref.
Data 1993, 22, 113.
(26) Reyftman, J. P.; Kagan, J.; Santus, R.; Morliere, P. Photochem.
Photobiol. 1985, 41, 1.
(27) Janssen, R. A.; Smilowitz, L.; Sariciftci, N. S.; Moses, D. J. Chem.
Phys. 1994, 101, 1787.
(28) Evans, C.; Weir, D.; Scaiano, J. C.; MacEachern, A.; Arnason, J.
T.; Morand, P.; Hallebone, B.; Leitch, L. C.; Philogene, J. R. Photochem.
Photobiol. 1986, 44, 441.
11. All φF, φF, kisc, and τT reach a plateau at or near R5.
These results, coupled with some absorption and fluorescence
data, are interpreted as indicating that R5 acts as the first
representative of an “R-polythiophene”.
3
1
12. Triplet Rn’s can transfer energy to O2 to produce O2
with high efficiency where the φ∆ values are essentially the same
as φT for R2, R4, and R5 and only 15% lower than φT for R3.
It appears that the efficiency for R6 and R7 is less, which could
be caused by the low energy of the triplets (and possibly also
steric factors for the dibutyl-substituted R7).
13. Up to 10 different methods of calculations have been
used to explore the most stable geometric forms of R2-R5
(twisting S-C-C-S angles) and the ∆H between the all-trans
and all-cis forms. These include three levels of ab initio. In
general MOPAC and MNDO93,PM3 predicted a most stable
geometry of all-cis that disagreed with all other calculation
methods that predicted all-trans (the former of which is certainly
incorrect). For R2-R5, the all-trans conformer is predicted to
be the most stable from among all possible conformers (2 for
R2, 3 for R3, and 10 for R5). For R3, MOPAC,MM2, DGauss,
and ab initio STO3G indicate a near planar structure for the tt
conformer. For R5, ab initio STO3G alone indicates a near t
planar structure for the tttt conformer.
(29) Xu, B.; Holdcroft, S. J. Am. Chem. Soc. 1993, 115, 8447.
(30) Scaiano, J. C.; Redmond, R. W.; Mehta, B.; Arnason, J. T.
Photochem. Photobiol. 1990, 52, 655.
(31) Van Veen, E. H. Chem. Phys. Lett. 1976, 41, 535.
(32) Flicker, W. M.; Mosher, O. R.; Kuppermann, A. Chem. Phys. Lett.
1976, 38, 489.
(33) Garcia, P.; Pernaut, P.; Hopiat, P.; Wintgens, V.; Valat, P.; Garnier,
F.; Delabouglise, D. J. Phys. Chem. 1993, 97, 513.
(34) Rossi, R.; Carpita, A.; Ciofalo, M.; Houben, J. L. Gazz. Chim. Ital.
1990, 120, 793.
(35) Rossi, R.; Ciofalo, M.; Carpita, A.; Ponterini, G. J. Photochem.
Photobiol., A 1993, 70, 59.
(36) Charra, C.; Fichou, D.; Nunzi, J. M.; Pfeffer, N. Chem. Phys. Lett.
1992, 192, 566.
(37) Buma, W. J.; Kohler, B. E.; Shaler, T. A. J. Phys. Chem. 1994,
98, 4990.
Acknowledgment. Part of this work was supported by
JNICT-Portugal (STRD/C/AMB/49792) and by the Italian
Ministero per l’Universita´ e la Ricerca Scientifica e Tecnologica
(MURST) and Consiglio Nazionale delle Ricerche (CNR). Drs.
J. Kagan, H. Naarmann, and H. Winberg are again acknowl-
edged for providing samples of R4, R6, and R7, respectively.
J.S.M. thanks Dr. I. Conc¸alves for experimental support in the
synthesis of R5. We thank Dr. Paritosh Das, Phillips Petroleum
Co., for carrying out several of the higher level calculations.
(38) Visser, G. J.; Heeres, G. J.; Wolters, J.; Vos, A. Acta Crystallogr.
B. 1968, 24, 467.
(39) Samdal, S.; Samuelson, E. J.; Volden, H. Synth. Met. 1993, 59,
259.
(40) Bucci, P.; Longeri, M.; Veracini, C. A.; Lunazzi, L. J. Am. Chem.
Soc. 1974, 96, 1305.
(41) Chadwick, J. E.; Kohler, B. E. J. Phys. Chem. 1994, 98, 3631.
(42) Takayanagi, M.; Gejo, T.; Hanazaki, F. J. Phys. Chem. 1994, 98,
12893. Hanazaki, F. Personal communication.
(43) Van Bolhuis, F.; Wynberg, H.; Havinga, E. E.; Meijer, E. W.;
Staring, E. G. I. Synth. Met. 1989, 30, 381.
(44) De Witt, L.; Blanchard, G. J.; Le Goff, E.; Benz, M. E.; Liao, J.
H.; Kanatzidis, M. G. J. Am. Chem. Soc. 1993, 115, 12158. Blanchard, G.
J. Personal communication.
(45) Taliani, C.; Danieli, R.; Lazzaroni, R.; Periasamy, N.; Ruani, G.;
Zamboni, R. Synth. Met. 1993, 55-57, 4714.
References and Notes
(1) Becker, R. S.; Seixas de Melo, J.; Mac¸anita, A. L.; Elisei, F. Pure
Appl. Chem. 1995, 67, 9.
(2) Chosrovian, H.; Rentsch, S.; Grebner, D.; Dahn, D. U.; Birchner,
E. Synth. Met. 1993, 60, 23.
(3) Colditz, R.; Grebner, D.; Hebig, M.; Reutsch, S. Chem. Phys. 1995,
201, 309.
(4) Oelkrug, D.; Egelhaaf, H.-I.; Wilkinson, F. J. Fluoresc., in press.
(5) Birnbaum, D.; Fichou, D.; Kohler, B. E. J. Chem. Phys. 1992, 96,
165.
(46) Andres, L. S.; Merchan, M.; Fulscher, M.; Roos, B. O. Chem. Phys.
Lett. 1993, 211, 125.
(47) Becker, R. S. Theory and Interpretation of Fluorescence and
Phosphorescence; Wiley-Interscience: New York, 1969.
(48) Foote, C. S. Light ActiVated Pesticides; Heitz, J. R., Downum, K.
R., Eds.; ACS Symposium Series 339; American Chemical Society:
Washington, DC, 1987; p 36.
(6) Periasamy, N.; Daneli, R.; Ruani, G.; Zamboni, R.; Talani, C. Phys.
(49) Ciofalo, M.; Ponterni, G. J. Photochem. Photobiol., A 1994, 83, 1.
(50) D’Auria, M.; D’Onofrio, F. Gazz. Chim. Ital. 1988, 118, 633.
(51) Egelhaaf, H.-I.; Oelkrug, D. SPIE, in press.
(52) Nunzi, J. M.; Pfeffer, N.; Charra, F.; Fichou, D. Chem. Phys. Lett.,
in press.
ReV. Lett. 1992, 68, 919.
(7) Birnbaum, D.; Kohler, B. E. J. Chem. Phys. 1992, 96, 2492.
(8) Belijonne, D.; Shuai, Z.; Bredas, J. L. J. Chem. Phys. 1993, 98,
8819.
(9) Negri, F.; Zgierski, M. Z. J. Chem. Phys. 1994, 100, 2571.
(10) Soos, Z. B.; Galva˜o, D. S. J. Phys. Chem. 1994, 98, 1029.
(11) Rubio, M.; Mercha´n M.; Orti, E.; Roos, B. Personal communication.
(12) Van Hecke, G. R., Horrocks, De W., Jr. Inorg. Chem. 1966, 5,
1968.]
(53) Cheng, X.; Ichimura, K.; Fichou, D.; Kobayashi, T. Chem. Phys.
Lett. 1991, 185, 286.
JP960852E