Phosphorescence and Triplet State Energies of Oligothiophenes - The

Citation data is made available by participants in Crossref's Cited-by Linking ..... Juan T. López Navarrete , David L. Officer , Pawel Wagner , John...
0 downloads 0 Views 130KB Size
Download PDF
4410

J. Phys. Chem. B 2005, 109, 4410-4415

Phosphorescence and Triplet State Energies of Oligothiophenes Dorothee Wasserberg,† Philippe Marsal,‡ Stefan C. J. Meskers,† Rene´ A. J. Janssen,*,† and David Beljonne*,‡ Laboratory of Macromolecular and Organic Chemistry, EindhoVen UniVersity of Technology, P.O. Box 513, 5600 MB EindhoVen, The Netherlands, and Chemistry of NoVel Materials, UniVersity of Mons-Hainaut, Place du Parc 20, 7000 Mons, Belgium ReceiVed: September 21, 2004; In Final Form: NoVember 26, 2004

The phosphorescence spectra of a series of small oligothiophenes (nT, n ) 1-3) incorporating a variety of substituents, end cappers, and functional groups have been recorded for the first time using gated detection in combination with nanosecond excitation in frozen solution at 80 K. The vibrationally resolved emission spectra provide accurate estimates of the T1 and S1 levels, and the singlet-triplet energy gap. Theoretical quantum chemical calculations performed at the DFT (B3LYP/6-31G*) level reproduce all experimental trends accurately and provide quantitative description of the S0-T1 energy difference. The geometry relaxation in the excited state shows that the “natural” size of the triplet exciton is about 3-4 thiophene units.

Introduction The T1 triplet excited state of oligothiophenes has been a subject of continuous interest ever since terthiophene (3T), a constituent of marigolds, was identified for its nematicidal activity in the presence of light.1 More recently, oligothiophenes (nT, with n the number of thiophene rings) have received considerable attention as organic semiconductors with potential applications in field-effect transistors, lasing, nonlinear optics, light-emitting diodes, and solar cells.2 The triplet T1 state is formed in oligothiophenes with considerable quantum yield via intersystem crossing from the photoexcited singlet S1 state to the triplet manifold involving low-lying Ti triplet excited states.3 Its long lifetime enhances energy migration, trapping, and the possibility to interact with electron acceptors to form charge carriers.4 The triplet states of various short and lengthy oligothiophenes (nT, n ) 2-17) have been identified and studied by recording their transient Tn r T1 absorption spectra.5 Despite this continued interest, one of the most basic features of the oligothiophene triplet state, the triplet energy ET, has long been elusive. Early attempts by Reyftmann and Scaiano to observe phosphorescence of the triplet state of 3T at 77 K were unsuccessful.6-8 Later, the triplet energy ET of 3T was estimated to be 1.72 ( 0.05 eV by recording the T1 r S0 absorption, utilizing heavy-atom-induced absorption spectroscopy.9 The first truly quantitative information on the triplet state energy of a series of oligothiophenes was reported by Landwehr et al., who recorded high-resolution triplet photoexcitation spectra by monitoring the delayed fluorescence of crystalline, unsubstituted oligothiophenes (nT, n ) 2-5).10 More recently, Seixas de Melo et al. and Rentsch et al. also reported triplet energies of oligothiophenes, using different approaches.11,12 Seixas de Melo et al. employed time-resolved photoacoustic calorimetry (PAC), in combination with spectroscopic data on absorption and fluorescence, and measured quantum yields for fluorescence and triplet formation to determine ET for nT (n ) 2-5, and 7).11 * Corresponding authors. E-mail (Janssen): [email protected]. E-mail (Beljonne): [email protected]. † Eindhoven University of Technology. ‡ University of Mons-Hainaut.

Rentsch et al. used a more direct method, photodetachment photoelectron spectroscopy (PD-PES) of oligothiophene anions, to measure ET for nT (n ) 2-4) in the gas phase.12 Moreover, they were able to measure, for the first time, the phosphorescence of 2T by gated detection in solution at cryogenic temperatures, but this method failed for 3T.12 Gated detection has been used successfully in recent years for observing phosphorescence of various conjugated polymers.13 In combination with laser excitation and sensitive detection, we now utilize this technique to record the first phosphorescence spectra of a series of oligothiophenes with up to 3 thiophene units and a range of substituents and functional groups (Figure 1) at 80 K in frozen solution. The triplet energy ET and the S1-T1 gap ∆EST are directly obtained from the emission spectra. On the theory side, triplet excitations in oligothiophenes and derivatives have been characterized by means of quantumchemical calculations, yet at the semiempirical level.14 Here, the results of Density Functional Theory (DFT) calculations, including geometry optimization in S0 and T1, are discussed that provide a firm quantitative description of the lowest triplet excited state. The DFT results reproduce not only the experimental triplet energy with a small (∼0.1 eV) and most importantly systematic underestimation; in addition, these allow to investigate the extension of the triplet exciton and its spin distribution in more detail. Experimental Section Materials. The synthesis of the nTBP and EDOnT oligothiophenes has been reported previously.15 The other oligomers were obtained from Aldrich (3T, 3TBA) or were synthesized according to published methods.16 3TOH, 3TC12, 3TMABr, 3TBP, EDO2T, and EDO3T were carefully purified by preparative HPLC. The purity of all oligomers was checked to avoid fluorescence or phosphorescent signals of minor quantities of contaminants. GC-MS was used for compounds with a molecular mass below 500 g/mol, while HPLC with diode array detection was employed for higher-mass compounds. The final products exhibited only a single peak in the respective chromatograms.

10.1021/jp0457085 CCC: $30.25 © 2005 American Chemical Society Published on Web 02/12/2005

Triplet State Energies of Oligothiophenes

Figure 1. Structure of oligothiophenes studied.

Sample Preparation. 2-Methyltetrahydrofuran (MeTHF) was predried over KOH for 3 days, distilled from CaH2, and stored under an inert atmosphere. All sample solutions were prepared in a glovebox at concentrations around 10-5 M (OD around 1 at λmax) for phosphorescence measurements and at about 10-6 M (OD < 0.3 at λmax) for fluorescence measurements. All samples were kept under a protective atmosphere at all times. Low-temperature spectra were recorded using an Oxford Optistat CF continuous flow cryostat or an Oxford Optistat DN bath cryostat. In all experiments, the temperature was kept constant at 80 (0.1 K during measurements. Absorption and Photoluminescence. UV-visible-near-IR absorption spectra were recorded on a Perkin-Elmer Lambda 900 spectrophotometer. Fluorescence spectra were recorded on an Edinburgh Instruments FS920 double-monochromator spectrometer equipped with a Peltier-cooled red-sensitive photomultiplier. Phosphorescence. A pulsed Nd:YAG laser (Surelite II-10, Continuum, pulse width 4 ns, repetition rate 10 Hz) was used for excitation, combined with an optical oscillator (Panther OPO, Continuum) to continuously tune the excitation energy. The excitation energy was chosen close to the absorption spectrum. Emission spectra were recorded with an intensified CCD camera (PI-MAX:1024HQ, Princeton Instruments) after dispersion of the photoluminescence by a spectrograph (SP306, Acton Research). The emitted light was collected and focused into a collecting optical fiber using 3 lenses at an angle of about 30 degrees with respect to the incident light beam. Signal acquisition by the camera was electronically gated. Phosphorescence spectra were recoded at delay times of >100 ns to avoid any residual fluorescence and gate widths of 0.1-20 ms. All phosphorescence spectra were smoothed (FFT) to reduce noise levels. The CCD camera has a reduced sensitivity above 1.5 eV, resulting in loss of spectral intensity in this region. Results and Discussion Fluorescence and Phosphorescence. The UV-visible absorption, steady-state fluorescence, and gated phosphorescence spectra of terthiophene (3T) in frozen MeTHF at 80 K are shown in Figure 2. The fluorescence (S1 f S0) spectrum exhibits a well-defined 0-0 transition at 3.06 eV and a vibronic progression with 0.18 eV energy spacing, characteristic for the CdC stretching mode of the conjugated system. In addition, a second vibrational mode with 0.07 eV energy spacing can be discerned. Phosphorescence spectra were recorded using pulsed (4 ns) excitation in combination with gated detection at fixed time delays after the laser pulse. Although previous attempts to record

J. Phys. Chem. B, Vol. 109, No. 10, 2005 4411

Figure 2. UV-visible absorption, steady-state fluorescence, and gated phosphorescence spectra of 3T in MeTHF at 80 K. Phosphorescence spectra were recorded with gated detection during 2 ms at different delay times (0.2, 2, 20, and 200 µs) after excitation. Excitation energy was at 3.32 and 3.35 eV in fluorescence and phosphorescence experiments, respectively.

the phosphorescence of 3T were unsuccessful,6,7,11 the gated detection technique proved to be rather straightforward in measuring this emission. The 0-0 vibronic band in the phosphorescence (T1 f S0) spectrum at 1.82 eV gives a first spectroscopic estimate for the T1 energy under these conditions and gives a value of 1.24 eV for the relaxed S1-T1 singlettriplet energy gap when compared with fluorescence at 3.06 eV. The triplet energy of 3T determined from the phosphorescence spectrum (1.82 eV) agrees reasonably well with previous values obtained from gas-phase PD-PES (1.92 eV),12 direct excitation in crystals at 6 K (1.87 eV),10 and time-resolved PAC in dioxane at 293 K (1.92 eV).11 The phosphorescence spectrum exhibits a vibronic progression with two bands and a shoulder at the lowest energy. When shifted over the energy axis, the positions of the peaks in the fluorescence and phosphorescence spectra are essentially the same, confirming that they correspond to Franck-Condon transitions to the same state (S0). With increasing time delay, the intensity of the phosphorescence signal is reduced as a result of the decay of the triplet state in time. Similar UV-visible absorption, fluorescence, and phosphorescence spectra were successfully recorded for all oligothiophenes shown in Figure 1 in frozen MeTHF at 80 K. Vibronic progressions with a dominant mode in the 0.14-0.19 eV range could be identified in these spectra for most oligothiophenes. As an example, Figure 3 shows the spectra of the EDOnT oligomers (n ) 1-3). These spectra clearly reveal that the first excited singlet and triplet states of the EDOnT oligomers evolve to lower energies with increasing number of thiophene rings. Within the EDOnT series, the absorption and fluorescence spectra are best resolved for the longest oligomer (n ) 3). As for 3T, the fluorescence spectrum of EDO3T shows two vibronic progressions (0.18 and 0.05 eV). The S1 and T1 energy levels and the corresponding S1-T1 energy gap for all 13 compounds are given in Table 1. A plot of the T1 energy versus the S1 energy reveals an almost linear relation (Figure 4). The slope of 0.57 of the least-squares linear fit indicates that the energy of the S1 excited state is more affected by chemical functionalization and a change in the conjugation length than the T1 excited state. Comparing the nTBP and EDOnT oligomers reveals that the energies of nTBP are systematically somewhat higher than of EDOnT. Table 1 and Figure 4 also demonstrate that the length of the oligomer has a pronounced effect on the energies of both

4412 J. Phys. Chem. B, Vol. 109, No. 10, 2005

Wasserberg et al. TABLE 1: Experimental and Theoretical Singlet and Triplet Energies (eV) of Oligothiophenes in Figure 1 experiment

theory

compound S1 r S0a S1 f S0b T1 f S0b S1-T1b 3T EDO3T EDO2T EDO1T 3TMK 3TMA 3TBA 3TMABr 3TBP 2TBP 1TBP 3TC12 3TOH

3.17 2.58 2.88 3.38 2.89 2.84 2.74 2.82 2.71 2.97 3.45 3.46e 3.12

3.06 2.56 2.84 3.30 2.75 2.69 2.69 2.69 2.63 2.90 3.34 3.03 3.02

1.82 1.55 1.74 2.03 1.74 1.70 1.65 1.67 1.62 1.76 2.04 1.84 1.86

1.24 1.01 1.10 1.27 1.01 0.99 1.04 1.02 1.01 1.14 1.30 1.19 1.16

S1-S0 T1-S0 INDO/SCIc DFTd 3.13 2.55 2.89 3.52

1.72 1.45 1.64 2.02

2.81 2.70

1.60 1.53

2.77 3.07 3.67

1.52 1.70 2.09

a From UV-visible absorption (0-0 transition). b From emission spectra (0-0 transitions). c INDO/SCI calculation at DFT S0 optimized geometry. d DFT calculations at optimized geometries for S0 and T1. e Maximum absorption, no resolved vibrational progression.

Figure 4. Triplet energy as a function of the singlet energies for oligothiophenes. Solid line represents a least-squares fit to the data.

Figure 3. UV-visible absorption, steady-state fluorescence, and gated phosphorescence spectra of EDOnT in MeTHF at 80 K. Phosphorescence spectra were recorded 200 ns after excitation and with gated detection. (a) EDO1T: fluorescence excitation 3.62 eV, phosphorescence excitation 3.40 eV, gate 10 ms. (b) EDO2T: fluorescence excitation 3.06 eV, phosphorescence excitation 2.88 eV, gate 100 µs. (c) EDO3T: fluorescence excitation 2.76 eV, phosphorescence excitation 2.95 eV, gate 1 ms.

S1 and T1. The singlet and triplet state energies of the EDOnT (n ) 1-3) and nTBP (n ) 1-3) oligomers are inversely proportional to the number of double bonds (NDB) as shown in Figure 5. The weaker slope of the T1 energy as compared to the S1 energy indicates that the triplet exciton is less extended.13b,17 The linear fits indicate that the conjugation length dependence of the singlet and the triplet excited-state energies of the EDOnT oligomers are slightly stronger (slopes of 18 (S1) and 12 (T1)) than for the nTBP oligomers (slopes of 17 (S1) and 10 (T1)). As can be seen in Table 1 and Figure 5, the S1-T1 energy gap in the nTBP and EDOnT oligomers decreases from ∼1.3 to

Figure 5. Relation between the inverse number of double bonds (1/ NDB) and the singlet and the triplet excited-state energies of EDOnT and nTBP oligomers.

∼1.0 eV, going from n ) 1 to 3. This decrease of exchange energy with conjugation length is in accordance with the 0.7 (0.1 eV S1-T1 energy gap found experimentally for many conjugated polymers.17 Also, chemical functionalization of terthiophene affects the singlet and triplet energies. Substitutents such as -CHO and -C(O)CH3 (in 3TMA, 3TMK, 3TBA, and 3TMABr) extend the conjugation length of 3T and sas a consequenceslower

Triplet State Energies of Oligothiophenes

J. Phys. Chem. B, Vol. 109, No. 10, 2005 4413 TABLE 2: Theoretical Triplet State Energies (eV) of Oligothiophenes T1-S0 DFTa

a

Figure 6. Site-selective excitation of the phosphorescence of 3T. Excitation wavelengths are indicated in the spectra. The vertical lines indicate the emission maximum for the spectrum taken after excitation at 3.18 and 3.06 eV, respectively.

Figure 7. Quantum-chemical (filled symbols) and experimental (open symbols) of the T1 state (squares) and S1 (circles) excited-state energies for oligothiophenes. The experimental data are from the 0-0 transitions in the fluorescence (S1) and phosphorescence (T1) spectra.

the excited-state energies of both T1 and S1 more strongly than the alkyl substituents in 3TC12 and 3TOH. This increased conjugation also lowers the exchange energy to ∼1.0 eV compared to 1.3 eV in 3T. In a frozen solvent, all oligothiophene molecules will have slightly different energetic environments as a result of small variations in conformation and the orientation of solvent molecules. To further elucidate the correlation between singlet and triplet state energies, site selective excitation experiments were performed. In these experiments, the phosphorescence spectra of 3T were recorded for excitation energies between 3.18 eV (at the top of the 0-0 transition in absorption) and 3.06 eV (at the half-maximum of 0-0) (Figure 6). The latter is close to the region where only the energetically most relaxed 3T molecules can be excited. Figure 6 reveals that as a result of the lowering of the excitation energy the onset of the phosphorescence spectrum (at half-maximum) shifts by 0.023 eV from 1.871 to 1.848 eV, while the maximum intensity undergoes a similar shift from 1.664 to 1.640 eV. This demonstrates that we are beginning to see features of selective excitation and indicates that photoexcitation is preserved at specific sites. Quantum-Chemical Calculations. To elucidate the nature of the triplet state in more detail, quantum-chemical calculations were performed, initially focusing on accurately determining the triplet energies. We used density functional theory (DFT)

n

nTBP

EDOnT

6 5 4 3 2 1

1.31 1.35 1.41 1.52 1.70 2.09

1.22 1.27 1.34 1.45 1.64 2.02

DFT calculations at optimized geometries for S0 and T1.

to describe the geometric and electronic structures in the S0 singlet ground state and T1 triplet excited state of these molecules. Here, we have adopted a potential including both Becke and Hartree-Fock exchange and Lee Yang and Parr correlation (B3LYP), which has been shown to provide a reliable description of both the electronic singlet ground state and the lowest triplet excited states in other conjugated compounds.18 The use of a 6-31G* split-valence basis set leads to accurate excitation energies at a reasonable computational cost.19 Singlet vertical transition energies were computed with the semiempirical intermediate neglect of differential overlap (INDO) method coupled to a single configuration interaction (SCI), INDO/SCI, based on DFT geometries (the MatagaNishimoto potential20 was adopted to depict electron-electron interactions). The active space size was set according to the number of π electrons in the molecule. The ZINDO and Gaussian9821 packages were used for the semiempirical INDO/SCI and DFT calculations, respectively. The calculations were performed for a selection of oligothiophenes depicted in Figure 1, that is, 3T, 3TMA, 3TBA, and the series EDOnT (n ) 1-3, and 6) and nTBP (n ) 1-6) (Tables 1 and 2). The T1-S0 energy differences calculated from the DFT data agree by a systematic underestimation of ∼0.1 eV where experimental data are available; the 0-0 transition energies are determined from the phosphorescence spectra (Table 1, Figure 7). Moreover, all experimentally observed trends with introduction of functional groups and chain length elongation are reproduced satisfactorily. In using the optimized S0 and T1 geometries, the quantum-chemical T1 energy corresponds to the energy difference between the minima of the potential energy surfaces, while the experimental 0-0 transition corresponds to the difference between the zero-point energies (ZPEs) of the two states. Since the dominant vibrational modes and frequencies are similar for S0 and T1, the error introduced by neglecting the vibrational effects is likely to be smaller than the observed differences. For calculating the vertical S1 energies, the DFT-optimized S0 ground-state geometries were used for INDO/SCI singlepoint calculations. The INDO/SCI S1 energies reproduce the experimental values and the trends (Table 1, Figure 7). However, the deviations in S1 energy between theory and experiment are less systematic and generally somewhat larger than for the DFT T1 energies. The larger differences are most likely due to the lower theoretical level of the INDO/SCI method and the fact that the S1 geometry was not optimized and only a vertical transition is computed. Since the calculations refer to the gas phase, the strong oscillator strength of the S1 f S0 transition may induce a medium-induced spectral shift, which could contribute to the observed larger differences between experiment and theory for the S1 energy level. For the T1 level, such spectral shift would not occur because the T1 f S0 transition has a weak oscillator strength.

4414 J. Phys. Chem. B, Vol. 109, No. 10, 2005

Wasserberg et al.

Figure 9. Total and beta spin density per ring for 6TBP and EDO6T triplet molecules.

Figure 8. Changes in bond lengths upon transition from the singlet ground state to the first excited triplet state as calculated with DFT (B3LYP/6-31G*) for EDOnT (a) n ) 1-3 and (b) n ) 6. The horizontal arrow indicates the extension of the triplet exciton, taking a cutoff of ∼|0.02 Å|. The bonds are consecutively numbered along the chain, setting the central bond to 0.

The DFT calculations reveal that, in the S0 ground state, the carbon-carbon bonds alternate in length along the backbone of the oligothiophene. The thiophene rings have aromatic character, with longer bonds between the rings. Going from the S0 to the T1 state, the lengths of the carbon-carbon bonds are subject to change in an alternating fashion. In the T1 excited state the inter-ring bonds shorten, while the double bonds of the S0 ground state become longer. The extent of the changes in bond length are an indication for the localization or delocalization of the excited state, that is, bonds that are subject to significant changes are involved in accommodating the new electronic configuration. The DFT calculations demonstrate that these changes occur symmetrically around the central thiophene ring(s) and that the effects decrease toward the ends of the molecules. As a consequence, the changes in bond lengths are more pronounced for the thiophene moieties than for phenyl or carbonyl moieties. Only for EDO1T and 1TBP are the terminal phenyl rings affected to a considerable extent. The single carbonyl group of 3TMA induces asymmetry in the bond length alternation, with stronger changes at the side of the carbonyl moiety. Figure 8 shows the changes in bond lengths in the T1 state of the EDOnT oligomers compared to the S0 ground state. The changes are similar in magnitude for all oligomers. Considering a cutoff of |0.02 Å|, the geometric deformation taking place in the triplet state extends over 3-4 repeating units. A more detailed analysis of the dihedral angles around the carbon-carbon bonds reveals that all aromatic rings are essentially planar. Some torsion occurs around the bonds connecting two rings, or a ring and an end group. In the singlet states the oligothiophene segments are essentially planar for all compounds, except for 3T and 3TBP where all the thiophene rings are twisted by ∼16 degrees (yet the S0 ground-state potential is very flat, with an energy destabilization of

∼0.35 [2.86] kJ/ mol when going from the global minimum to the fully planar structure in 3T [3TBP]). In this respect, the carbonyl end groups must be responsible for the complete planarization in 3TMA and 3TBA. In contrast to the carbonyl end groups, the phenyl end groups are usually twisted in the singlet state by 20 degrees for EDOnT and 27 degrees for nTBP. The triplet states of the oligothiophenes are in general planar. Only for nTBP and EDOnT does some twisting (6-7 degree) occur in the triplet state around the thiophene-phenyl bonds of most nTBP and EDOnT (n ) 1-6) oligomers. To gain more information on the “natural” size of an exciton in oligothiophenes, we have analyzed the spin density distributions in the largest compounds investigated (6TBP and EDO6T), see Figure 9. As expected from the localized character of the triplet, the total spin density (difference between spin-up or alpha electron and spin-down or beta electron densities) per ring is maximized around the central part of the two molecules, with ∼85 and ∼80% of the total spin density contained within the inner four rings of 6TBP and EDO6T, respectively. Examination of the beta spin density distributions, reported in Figure 9, provides complementary insight into the triplet exciton size. Because of spin polarization phenomena, the increased alpha spin density around the center of the molecule is accompanied by a redistribution of the beta spin density that is depleted at the center and peaks close to the edges of the triplet excited species. In case the molecule size is larger than the exciton size, the total beta spin densities are thus expected to decrease at both ends of the molecule. This effect is clearly seen in Figure 9, showing a typical inverted W shape for the beta spin density distribution. From this figure, we infer an exciton size of about four units in both 6TBP and EDO6T that is consistent with the above analysis of the geometric deformations induced by going from the singlet ground state to the triplet excited state. These results also agree very well with the triplet size (four units) in unsubstituted oligothiophenes, as inferred from the analysis of the measured zero-field-splittings (ZFS) obtained from electron spin resonance (ESR) spectroscopy within a model based on an extended Hubbard Hamiltonian and going beyond the dipole model.22 As pointed out by Bennati et al.,22 phenomenological models based on the dipolar approximation underestimate the mean distance between the dipoles contributing to the ZFS, with values on the order of the length of one repeating unit in nT 23 or polythiophene.24 Figure 10 shows the triplet state energies (ET) of the nTBP (n ) 1-6) and EDOnT (n ) 1-6) oligomers versus the inverse number of double bonds (1/NDB). As expected, the dependence of ET on 1/NDB levels off for high NDB and seems to approach

Triplet State Energies of Oligothiophenes

J. Phys. Chem. B, Vol. 109, No. 10, 2005 4415 References and Notes

Figure 10. Relation between the inverse number of double bonds (1/NDB) and computed triplet excited-state energies of nTBP oligomers (n ) 1-6) and EDOnT (n ) 1-6).

an asymptotic value for a chain length corresponding to ∼4 thiophene units. Conclusions We have recorded the first phosphorescence spectra for a series of small, functionalized oligothiophenes (nT, n ) 1-3) in frozen solution at 80 K using nanosecond excitation in combination with gated detection to eliminate spurious fluorescence emission. The phosphorescence spectra exhibit a vibronically resolved progression with a dominant vibrational mode that corresponds to the CdC stretch vibration. The 0-0 transitions provide a direct spectroscopic probe of the triplet energy ET. Apart from a small, but systematic, underestimation of ∼0.1 eV, the experimental energies and their variation with the changes in chemical structure of the oligomer are accurately reproduced by DFT (B3LYP/6-31G*) calculations when using optimized geometries for S0 and T1. The combined experimental and theoretical results demonstrate that there is a significant dispersion of both triplet and singlet excited-state energies with chain length and a decreasing S1-T1 gap with increasing n. In the triplet state, there is a significant change in the bond length alternation compared to the singlet ground state. The triplet exciton localizes at the center of the oligomer, and an analysis of the spin density reveals that the “natural” size is approximately 3-4 thiophene rings. Acknowledgment. We thank Dr. L. Groenendaal (Agfa) for a gift of the EDOnT oligomers and R. Bovee and Dr. X. Lou for assistance with preparative HPLC. This research has been supported by The Netherlands Organization for Scientific Research (NWO) through a grant in the PIONIER program. The research of S. C. J. Meskers has been made possible by a fellowship of the Royal Netherlands Academy of Arts and Sciences. The work in Mons is partly supported by the Belgian Federal Services for Scientific, Technical, and Cultural Affairs (InterUniversity Attraction Pole 5/3) and the Belgian National Science Foundation (FNRS). David Beljonne is Research Associate from FNRS. The research of P. Marsal has been made possible by a fellowship of the European Commission, under the RTN network ‘LAMINATE’.

(1) Bakker, J.; Gommers, F. J.; Nieuwenhuis, I.; Wynberg, H. J. Biol. Chem. 1978, 254, 1841. (2) Handbook of Oligo- and Polythiophenes; Fichou, D., Ed.; WileyVCH: Weinheim, 1999. (3) Beljonne, D.; Shuai, Z.; Pourtois, G.; Bre´das, J.-L. J. Phys. Chem. A 2001, 105, 3899. (4) Janssen, R. A. J.; Moses, D.; Sariciftci, N. S. J. Chem. Phys. 1994, 101, 9519. (5) (a) Evans, C. H.; Scaiano, J. C. J. Am. Chem. Soc. 1990, 112, 2694. (b) Wintgens, V.; Valat, P.; Garnier, F. J. Chem. Phys. 1994, 98, 228. (c) Janssen, R. A. J.; Smilowitz, L.; Sariciftci, N. S.; Moses, D. J. Chem. Phys. 1994, 101, 1787. (d) Becker, R. S.; Seixas de Melo, J.; Mac¸ anita, A. L.; Elisei, F. Pure Appl. Chem. 1995, 67, 9. (e) Grebner, D.; Helbig, M.; Rentsch, S. J. Phys. Chem. 1995, 99, 16991. (f) Apperloo, J. J.; Janssen, R. A. J.; Malenfant, P. R. L.; Fre´chet, J. M. J. J. Am. Chem. Soc. 2001, 123, 6916. (6) Reyftmann, J. P.; Kagan, J.; Santus, R.; Morliere, P. Photochem. Photobiol. 1985, 41, 1. (7) Evans, C.; Weir, D.; Scaiano, J. C.; MacEachern, A.; Arnason, J. T.; Morand, P.; Hollebone, B.; Leitch, L. C.; Philoge`ne, B. J. R. Photochem. Photobiol. 1986, 44, 441. (8) Scaiano, J. C.; MacEachern, A.; Arnason, J. T.; Morand, P.; Weir, D. Photochem. Photobiol. 1987, 46, 193. (9) Scaiano, J. C.; Redmond, R. W.; Mehta, B.; Arnason, J. T. Photochem. Photobiol. 1990, 52, 655-659. (10) Landwehr, P.; Port, H.; Wolf, H. C. Chem. Phys. Lett. 1996, 260, 125. (11) Seixas de Melo, J.; Silva, L. M.; Arnaut, L. G.; Becker, R. S. J. Chem. Phys. 1999, 111, 5427. (12) Rentsch, S.; Yang, J. P.; Paa, W.; Brickner, E.; Schiedt, J.; Weinkauf, R. Phys. Chem. Chem. Phys. 1999, 1, 1707. (13) (a) Romanovskii, Y. V.; Gerhard, A.; Schweitzer, B.; Scherf, U.; Personov, R. I.; Ba¨ssler, H. Phys. ReV. Lett. 2000, 84, 1027. (b) Hertel, D.; Setayesh, S.; Nothofer, H.-H.; Scherf, U.; Mu¨llen, K.; Ba¨ssler. H. AdV. Mater. 2001, 13, 65. (c) Romanovskii, Y. V.; Ba¨ssler, H. Chem. Phys. Lett. 2000, 326, 51. (d) Hertel, D.; Ba¨ssler, H.; Guenthner, P.; Scherf, U. J. Chem. Phys. 2001, 115, 10007. (e) Ko¨hler, A.; Wislon, J. S.; Friend, R. H.; AlSuti, M. K.; Khan, M. S.; Gerhard, A.; Ba¨ssler, H. J. Chem. Phys. 2001, 116, 9457. (f) Rothe, C.; Hintschich, S.; Monkman, A. P.; Svensson, M.; Andersson, M. R. J. Chem. Phys. 2002, 116, 10503. (g) Rothe, C.; Monkman, A. Phys. ReV. B 2002, 65, 073201. (h) Hayer, A.; Ba¨ssler, H.; Falk, B.; Schrader, S. J. Phys. Chem. A 2002, 106, 11045. (i) Laquai, F.; Im, C.; Kadashchuk, A.; Ba¨ssler, H. Chem. Phys. Lett. 2003, 375, 286. (j) Reufer, M.; Schindler, F.; Patil, S.; Scherf, U.; Lupton, J. M. Chem. Phys. Lett. 2003, 381, 60. (k) Rothe, C.; Monkman, A. P. Phys. ReV. B 2003, 68, 075208. (14) Beljonne, D.; Cornil, J.; Friend, R. H.; Janssen, R. A. J.; Bre´das, J. L. J. Am. Chem. Soc. 1996, 118, 6453. (15) Apperloo, J. J.; Groenendaal, L.; Verheyen, H.; Jayakannan, M.; Janssen, R. A. J.; Dkhissi, A.; Beljonne, D.; Lazzaroni, R.; Bre´das, J.-L. Chem. Eur. J. 2002, 8, 2384. (16) (a) Nakayama, J.; Nakamura, Y.; Yoichi, T.; Tajiri, T.; Hoshino, M. Heterocycles 1986, 24, 637. (b) Wei, Y.; Wang, B.; Wang, W.; Tian, J. Tetrahedron Lett. 1995, 36, 665. (c) van Haare, J. A. E. H.; Havinga, E. E.; van Dongen, J. L. J.; Janssen, R. A. J.; Cornil, J.; Bre´das, J.-L. Chem. Eur. J. 1998, 4, 1509. (17) Ko¨hler, A.; Beljonne, D. AdV. Funct. Mater. 2004, 14, 11. (18) Pogantsh, A.; Heimel, G.; Zojer, E. J. Chem. Phys. 2002, 117, 5921. (19) Nguyen, K. A.; Kennel, J.; Patcher, R. J. Chem. Phys. 2002, 117, 7128. (20) Mataga, N.; Nishimoto, K. Z. Phys. Chem. 1957, 13, 140. (21) Frisch, M. J., et al. Gaussian98 (Revision A.10); Gaussian, Inc.: Pittsburgh, PA, 1998. (22) Bennati, M.; Nemeth, K.; Surjan, P. R.; Mehring, M. J. Chem. Phys. 1996, 105, 4441. (23) Bennati, M.; Grupp, P.; Ba¨uerle, P.; Mehring, M. Mol. Cryst. Liq. Cryst. 1994, 256, 751. (24) Swanson, L. S.; Shinar, J.; Yoshino, K. Phys. ReV. Lett. 1990, 65, 1140.