Influence of chain length on exciton migration to low-energy sites in single fluorene copolymers

Abstract

Fluorescence spectroscopy was performed on single molecules of two 9,9-dialkylfluorene–benzothiadazole (FxBT) copolymers with a 10-fold difference in average molecular weight. Molecules of both polymers exhibit red-shifted emission indicative of energy migration to low-energy sites (LES) on the polymer chains; however, “red” spectra are much more common for the longer polymer. Since singlet-exciton migration is found to occur on the molecular length scale in both cases, the increased number of red-shifted spectra observed for the longer polymer is evidence that the likelihood of LES formation increases with chain length. This relationship is discussed in terms of three possible causes of low-energy sites: local polymer conformations, chromophores with extended conjugation lengths, and random chemical defects.

---

1 Introduction

Conjugated polymers have garnered intense research interest as materials for light emitting devices.^1–3 Homo- and copolymers of polyfluorene have been studied extensively of late, due to improved stability for device applications and facile tunability of emission wavelength throughout the visible region. In particular, considerable experimental and theoretical work has been devoted to examining the cause(s) of red-shifted emission in polyfluorene-based molecules. Traditionally, these features have been ascribed to the formation of aggregates or excimers in the solid state; however, recent evidence suggests that they may instead be attributed to the presence of chemical keto defects along the polymer chains.

It is well known that low-energy exciton trap sites significantly influence conjugated polymer fluorescence. Both polymer conformations^4–9 and chemical defects^4,10,11 have been shown to produce low-energy sites (LES), resulting in “downhill” exciton migration and red-shifted emission. Investigation of LES in bulk samples is complicated by extreme heterogeneity; however, advances in single-molecule spectroscopy^12–15 allow for evaluation of LES concentrations and their effects on polymer photophysics through examination of isolated chains.^16–19 We have previously reported red-shifted emission in single molecules of the conjugated polymer MEH–PPV, resulting from efficient exciton migration to LES formed by chain–chain contacts.^20–22 The incorporation of increased numbers of synthetic saturation defects was found to result in decreased conformational order, shorter exciton migration lengths and fewer red-shifted spectra.²³

In this paper, we report on our observation of red-shifted emission in single molecules of two 9,9-di-n-alkylfluorene–benzothiadazole copolymers, F6BT and F8BT, with a ∼10-fold difference in average molecular weight (see Fig. 1). Further, we show that the percentage of red-shifted spectra observed for a particular population of single polymers, and the number of low-energy sites present per molecule, tend to increase with increasing chain length. The F6BT and F8BT molecules studied have different alkyl side chains (n-hexyl vs. n-octyl groups) and average molecular weights of 10 000 and 100 000 g mol⁻¹, respectively. These values correspond to ∼20 and ∼200 repeat units, likely comprising ∼5 and ∼50 weakly coupled chromophores.²⁴ Single molecules of both polymers exhibit narrow, red-shifted emission features, resulting from intramolecular excited-state energy transfer to LES;^19,22 however, these features are much more common for the longer F8BT molecules. Examination of fluorescence intensity fluctuations revealed that exciton migration is highly efficient in both polymers. Therefore, the increased probability of red-shifted emission observed for the longer F8BT molecules is attributed to an increased number of low-energy sites. Results of polarization anisotropy experiments suggest that local conformations may cause LES formation; the possible effects of conjugation lengths and chemical defects are also discussed.

Fig. 1 Chemical structures of poly(9,9-di-n-alkylfluorene-co-benzothiadazole)s F6BT and F8BT and DiO (3,3′-dioctadecyloxacarbocyanine perchlorate) dye.

2 Experimental

Sample preparation

F6BT was purchased from American Dye Source (Quebec, Quebec); F8BT was synthesized by CDT Ltd., and obtained as a gift, and DiO (3,3′-dioctadecyloxacarbocyanine perchlorate) was purchased from Molecular Probes (Eugene, OR).

Pristine FxBT films were spin-cast on glass cover slips from solutions ∼14 g L⁻¹ in toluene. Single-molecule samples were spin-cast from toluene solutions ∼10⁻⁹–10⁻¹⁰ M in FxBT (or DiO) and ∼1% (w/w) in polystyrene host. The resulting films had single-molecule concentrations of 0.2–0.7 molecules µm⁻². Single-molecule samples for fluorescence spectroscopy were exposed to vacuum (P ≤ 10⁻⁶ Torr) and coated with aluminium (≥100 nm) to minimize exposure to oxygen and thereby reduce oxygen-assisted quencher formation.²² The majority of samples for quenching studies were used as spin-cast; however, measurements were also made on oxygen-depleted (Al-coated) samples for comparison.

Fluorescence spectra

Fluorescence measurements on single molecules and pristine films were performed at room temperature with a sample-scanning confocal apparatus, described previously.¹⁹ Molecules and films were excited with cw irradiation using the 457-nm line of an argon-ion laser. A 100× oil-immersion objective (NA 1.25) focused the excitation beam to a diffraction-limited spot ∼300 nm in diameter (power ≤1 kW cm⁻²). Emission was collected with the same objective and filtered to remove scattered excitation light. An avalanche photodiode (APD) was used to detect emission for intensity measurements; fluorescence spectra were acquired with a polychromator equipped with a back-illuminated, liquid-nitrogen-cooled CCD camera. Spectra of dilute (∼10⁻⁷ M) solutions were acquired on a Spex Fluorolog fluorimeter (using 457-nm excitation) and corrected for PMT response.

Fluorescence quenching

Efficiencies of quenching events (F_q) in single polymers were estimated by comparing the partially quenched fluorescence intensities with the average of the unquenched intensities before and after the events, i.e.F_q = (I_uq − I_q)/I_q, where I_uq and I_q correspond to the fluorescence intensity in the unquenched and quenched state, respectively. Because signal-to-background ratios for F6BT molecules were often low, efficiencies are quoted only as greater or less than 50%. The number of quenching events per million excitations was estimated for each molecule, using the number of events per second and the average unquenched intensity for the molecule. In these calculations, the detection efficiency was assumed to be 10%, and near-unity quantum yields were used for both polymers.

Polarization anisotropy

For anisotropy measurements, an electro-optic modulator with a quarter-wave-plate analyzer was used to continuously rotate the linear excitation polarization from 0 to 180° (using a sawtooth waveform) at a frequency of 1 Hz. Total fluorescence intensity was monitored and summed for tens of cycles. Single-polymer modulation depths were determined by fitting the intensity data to a cosine dependence, as described below. Fluorescent beads (spheres) and single DiO molecules (simple dipoles) were used in control experiments, to verify instrumental capabilities.

3 Results

Single-molecule fluorescence spectra of F6BT and F8BT are shown in Fig. 2(A, B). A majority of molecules in both polymers emit near 520 nm, close to the fluorescence maxima in dilute toluene solutions (discussed below). F6BT and F8BT molecules also exhibit emission that is red-shifted 15–40 nm from the solution maxima, as shown for F8BT in Fig. 2(B). Individual spectra of each polymer vary widely in intensity, width, and extent of vibronic structure, but “blue” and “red” spectra of a given polymer show no systematic differences in properties. On average, F8BT spectra are slightly broader than F6BT spectra, with more increased structure; however, these differences are small compared to the diverse spectral behavior exhibited by single molecules of each polymer.

Fig. 2 (A) and (B) Fluorescence spectra of single F6BT and F8BT molecules in oxygen-depleted polystyrene matrices, elicited by excitation at 457 nm. (C) and (D) Histograms of λ_max values from single-molecule spectra (201 and 197 molecules, respectively).

The most dramatic contrast in the single-molecule spectra of the two polymers lies in the percentage of molecules displaying red-shifted emission, as shown in the histograms in Fig. 2(C, D). Only 9 of 201 F6BT molecules (4%) emit with peak wavelengths of at least 535 nm, while 83 of 197 F8BT spectra (42%) have λ_max ≥ 535 nm. As a result, the distribution of peak positions for F8BT is shifted 14–19 nm to the red of the F6BT distribution (comparing mean or most-probable λ_max values, respectively), and is somewhat broader, with a standard deviation of 10 vs 8 nm.

These differences in peak-wavelength distributions are reflected in “ensemble” spectra of the two polymers, generated by addition of the single-molecule spectra (Fig. 3A). The F6BT “ensemble” (dashed line) is peaked at 520 nm and is 81 nm wide at half-maximum. The F8BT “ensemble” (solid line) is red-shifted and broadened, with a maximum at 539 nm and a FWHM value of 92 nm. This significant contrast in the fluorescence of the two polymers is unique to the single-molecule level. Normalized spectra of F6BT and F8BT are nearly identical in dilute (∼10⁻⁷ M) toluene solutions (λ_max = 521 nm, FWHM = 89 nm), as shown in Fig. 3(B). Emission from pristine films is also very similar (λ_max = 555 nm, FWHM = 104 nm), though the F8BT spectrum is slightly broader (Fig. 3B). Comparison of single-molecule and bulk spectra reveals that the F6BT single-molecule “ensemble” spectrum is peaked near the toluene emission, with a smaller FWHM value. The F8BT “ensemble” spectrum falls approximately halfway between the solution and film spectra in wavelength, 18 nm to the red of the maximum in toluene. In addition, it is slightly (3 nm) broader than the inhomogeneously broadened solution spectrum.

Fig. 3 Fluorescence spectra of F6BT (dashed) and F8BT (solid), normalized at their maxima. (A) “Ensemble” spectra generated by addition of single-molecule spectra. (B) Spectra in dilute toluene solution (λ_max = 521 nm) and in pure films (λ_max = 555 nm). The relative intensities of the solution and film spectra as plotted are arbitrary.

Centered “ensembles” (not shown) were constructed for both polymers as a means of investigating the source(s) of broadening in the F8BT “ensemble.” Single-molecule spectra were shifted in wavelength so that their maxima overlapped, and then summed. For F6BT, the centered sum is 2 nm narrower than the original (79 vs. 81 nm FWHM), consistent with the relatively narrow distribution of λ_max values (the scarcity of “red” spectra). Centering the F8BT spectra resulted in a 6-nm decrease in the FWHM of the sum, from 92 to 86 nm, bringing the width below that of the inhomogeneously broadened toluene spectrum. This indicates that a significant amount (6 nm) of the broadening observed in the F8BT “ensemble” arises from the broader peak-wavelength distribution (the increased contribution of red-shifted spectra). The remainder of the width difference between “ensembles” of the two polymers is due to the greater intrinsic FWHM values of F8BT vs. F6BT single-molecule spectra, as evidenced by the 7-nm difference in widths of the centered spectra.

Fluorescence-quenching efficiency (F_q, see Experimental section) measurements were undertaken to estimate exciton-migration lengths in the two polymers and determine whether the increased likelihood of red-shifted emission in F8BT is the result of more efficient exciton migration to LES. Photo-generated quenchers are formed in F6BT and F8BT molecules after thousands or millions of excitations (in these studies taking milliseconds to seconds). Singlet excitons are quenched reversibly on the millisecond time scale, with milliseconds or seconds between events (Fig. 4(A)). Calculation of F_q indicates that quenching is highly efficient in F6BT and F8BT: 81% of quenching events in F6BT and 69% of events in F8BT had F_q greater than 50%. Exciton-migration lengths for both polymers are on the molecular scale.

Fig. 4 (A) A fluorescence intensity transient for a single F6BT molecule in a polystyrene host matrix, taken with cw excitation at 457 nm. The film was exposed to ambient oxygen. (B) and (C) Histograms of quenching frequencies in F6BT and F8BT molecules (55 and 60 molecules, respectively). (D) and (E) Histograms of the number of characteristic quenching depths exhibited by F6BT and F8BT molecules.

Quenching studies also afford some insight into the relative number of quenchers formed in molecules of F6BT vs. F8BT. Quenching frequencies were determined for single molecules of each polymer. The average number of quenching events per million excitations is greater in F8BT vs. F6BT molecules (2.8 vs. 1.1). In addition, histograms show that 21 of 60 F8BT molecules studied had 3 or more quenching events per million excitations, while only 2 of 55 F6BT molecules underwent quenching as frequently (Fig. 4(B, C)). Compared to samples exposed to ambient conditions, polymer molecules in oxygen-depleted films were found to undergo quenching at least ten times less frequently. This observation is consistent with the model of reversible, oxygen-assisted quenching in conjugated polymers, possibly involving polymer⁺/O₂⁻ adducts, which we proposed previously for MEH–PPV.^20,22,25

Both polymers exhibit reproducible quenching to one or more characteristic depths. Histograms of the number of quenching levels in F6BT and F8BT fluorescence-intensity transients are shown in Fig. 4(D, E). F8BT molecules were observed to visit more quenching levels, on average (2.1 per molecule vs. 1.1 for F6BT). Also, 22 of 60 F8BT molecules exhibited three or more characteristic quenching levels, while no F6BT molecules had more than two (Fig. 4(D, E)).

Polarization anisotropy measurements were performed on single molecules of the longer F8BT polymer, in order to gain insight into polymer conformations. The linear excitation polarization was continuously rotated from 0 to 180° in 1-s cycles, as described above, resulting in angle- (and time-) dependent modulation of single-molecule fluorescence intensity (see Fig. 5, insets). For each molecule studied, the fluorescence signal was collected for tens of cycles, summed, and fit to the cosine dependence shown in Fig. 5. Here, M is the modulation depth, a measure of anisotropy; M is expected to be zero for spheres and unity for dipoles. These expectations were confirmed in control experiments. A histogram for 32 DiO molecules (simple dipoles) is shown in Fig. 5(B). Here, M is peaked near 1, with an average modulation depth of 0.94, while measurements of fluorescent spheres (not shown) yielded an average modulation depth of 0.08. Together, these results confirm the instrument's capability to measure high and low anisotropies and establish an error for modulation depths of ≤ 0.1. The histogram of single-polymer modulation depths for 64 F8BT molecules is shown in Fig. 5(A). For this “ensemble,” M is peaked near 0.15, with an average value of 0.26, a significant tail to longer values, and a maximum modulation depth of 0.95.

Fig. 5 Histograms of single-molecule modulation depths, M, for (A) 64 F8BT molecules (〈M〉 = 0.26) and (B) 32 DiO molecules (〈M〉 = 0.94). Insets: Plots of total fluorescence intensity vs. rotation angle, θ, for representative single molecules.

4 Discussion

We have observed single-molecule emission spectra in F6BT and F8BT that are red-shifted 15–40 nm from the dilute-solution maxima, evidence of intramolecular excited-state energy transfer to low-energy sites (LES) on the polymer chains.^19,22 Histograms of peak positions (Fig. 2(C, D)) show that the longer F8BT molecules are more likely to exhibit red-shifted emission. These results raise several key issues for discussion, including: (1) the cause of the increased number of “red” spectra observed for F8BT, (2) the relationship between the length of a polymer chain and the number of low-energy sites it possesses, and (3) the nature of these low-energy sites. The implications of the present studies for intra- and interchain contributions to red-shifted emission will also be discussed.

The increased probability of red-shifted emission noted for F8BT molecules can be attributed to an increased number of LES. Increased exciton-migration lengths could also account for the higher percentage of “red” spectra, as observed previously for MEH–PPV samples with varying concentrations of synthetic tetrahedral defects.²³ In this case, however, the highly efficient quenching of singlet excitons observed for F6BT and F8BT molecules indicates that exciton migration occurs on the molecular length scale in both polymers. Since energy migration is highly efficient in F6BT, the lower percentage of “red” spectra observed for these molecules is evidence for a decreased number of LES. The analysis of quenching levels in fluorescence-intensity transients of the two polymers (Fig. 4(C, D)) also supports this conclusion. In previous studies of MEH–PPV, fluorescence spectra for unquenched and partially quenched forms of MEH–PPV molecules indicated that quenchers are formed at the same sites (LES) responsible for red-shifted emission.²² The presence of several characteristic F_q value in an intensity transient was attributed to the existence of several energy “funnels,” and several LES, in the molecule.^20,22 As such, the observation that F6BT molecules visit fewer quenching levels suggests that there are fewer LES in these molecules.

There are several known causes of low-energy sites in isolated polymer chains, including conformationally induced chain–chain contacts (intrachain excimer-like interactions), increased chromophore conjugation lengths, and random chemical defects, each of which are discussed below. Given these possibilities, we may first comment on the reason for the increased probability of red-shifted emission in F8BT. The only differences between the two polymers studied are the ten-fold difference in average molecular weight (and average chain length) and the presence of n-hexyl vs. n-octyl side chains. Side chains are known to alter emission properties (as well as packing arrangements and morphologies) in bulk polymer samples;^4,26–28 however, it is unlikely that these two very similar, straight-chain alkyl substituents would have significantly different effects on the formation of chain–chain (backbone–backbone) contacts, the generation of random chemical defects, or the conjugation lengths of polymer segments in isolated chains. The increased number of LES in F8BT molecules must instead be attributed to the increase in polymer chain length.

The polarization anisotropy results support the model of conformationally induced LES formed by intramolecular chain–chain contacts found previously for MEH–PPV.^21,22 Individual F8BT molecules exhibit significant anisotropies, indicative of a certain degree of conformational order, and consistent with local collapsed conformations and chain–chain contacts. Using previous Monte Carlo simulations of modulation-depth histograms for MEH–PPV as a rough guide, we speculate that F8BT molecules may adopt a defect-coil conformation, which was simulated to have an average modulation depth of 0.28, and is expected to contain chain–chain contacts.²¹

If LES are, in fact, conformationally induced, the noted dependence on polymer chain length means that chain–chain contacts must be more easily achieved with long chains. It seems reasonable that the folding of a longer chain may be more likely to produce chain–chain contacts. Also, given the broad molecular-weight distributions characteristic of conjugated polymers, it is reasonable to propose that the nine F6BT molecules with red-shifted emission spectra possessed conformationally induced LES due to increased chain lengths. Excimer broadening is not evident in the “red” single-molecule spectra; however, as noted previously for MEH–PPV,²² this may simply indicate that the chromophores lack the precise orbital orientations required for excimer formation. Thus, given our results, local collapsed conformations involving intramolecular chain–chain contacts are a likely cause of low-energy sites.

LES could also be caused by chromophores with increased conjugation lengths. Previous studies on PPV-type model compounds suggest that variations in conjugation length are unlikely to account for large shifts in emission spectra;^29,30 however, chromophores in those polymers likely consist of ∼10 repeat units.³¹ Conjugation lengths in FxBT molecules are expected to be significantly shorter. A study of 9,9-di-n-hexylfluorene oligomers suggests that dialkylfluorene homopolymers have conjugation lengths of ∼10 repeat units, though emission properties converge by n = 6.³² The copolymers studied here could have as few as 3–5 repeat units in each chromophore; therefore, it is conceivable that a small increase in conjugation length could result in a sizable red shift. Polymer chains with more repeat units may be more likely to contain chromophores with extended conjugation lengths.

A third possibility is that LES are formed by random chemical defects along the polymer chains. In this case, the number of LES would be expected to increase with chain length, as observed. Recent work on poly(9,9-dialkylfluorene)s suggests that keto defects are formed at the fluorene 9-position upon photo-oxidation.^4,6,10,11,33 These fluorenones emit near 2.2–2.3 eV (539–564 nm) and could conceivably give rise to the red-shifted emission we observe. However, in FxBT, the presence of the benzothiadazole comonomer alone is expected to give rise to emission in the same approximate region of the spectrum. In addition, our spectra were collected under extremely low oxygen concentrations, under which permanent photo-oxidation may not be possible. Based on our results, there is little to support keto-defect formation as the cause of low-energy sites.

Finally, we speculate on the implications of these single-molecule studies for the contributions of intra- and interchain interactions to solid-state fluorescence. Our results clearly indicate that intrachain interactions can give rise to red-shifted emission features in fluorene copolymers. However, the majority of these features occur to the blue of the pristine film spectra and lack the breadth observed in the solid-state fluorescence. Thus, intrachain interactions alone can not account for the fluorescence properties of pristine films. Recent work on films of 9,9-dialkylfluorene homopolymers suggests that no new spectroscopic features (such as exciton bands) arise from interchain interactions;^4,10,11 however, as these authors suggest, increased intermolecular order in the solid state could result in accelerated diffusion to LES ¹⁰ and/or conformations with increased conjugation lengths.⁴

5 Conclusions

In summary, red-shifted emission was observed from isolated chains of FxBT copolymers, arising from intramolecular excited-state energy migration to low-energy sites on the polymer chains. The percentage of red-shifted spectra and the preponderance of LES were found to increase with polymer chain length. These results emphasize the important contributions of intrachain interactions to conjugated-polymer fluorescence. The exact nature of the low-energy sites formed in FxBT is uncertain. Single molecules of the longer F8BT polymer adopt conformations with significant anisotropies, and are expected to experience chain–chain contacts which could be the cause of LES. However, we cannot completely exclude the possibility that LES are formed either by chromophores with extended conjugation lengths or by chemical keto defects on the polymer chains; further experiment and/or simulation would be required to assess these possibilities.

Acknowledgements

We thank Prof. Richard Friend (Cambridge University) and CDT Ltd. for providing the F8BT sample, Dr Carlos Silva (Cambridge University) for helpful discussions, and the National Science Foundation and Robert A. Welch Foundation for financial support of this work.

References

1. R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, R. N. Marks, C. Taliani, D. D. C. Bradley, D. A. Dos Santos, J. L. Brédas, M. Lögdlund and W. R. Salaneck, Nature, 1999, 397, 121.
2. A. W. Grice, D. D. C. Bradley, M. T. Bernius, M. Inbasekaran, W. W. Wu and E. P. Woo, Appl. Phys. Lett., 1998, 73, 629.
3. M. Gross, D. C. Müller, H.-G. Nothofer, U. Scherf, D. Neher, C. Bräuchle and K. Meerholz, Nature, 2000, 405, 661.
4. U. Scherf and E. J. W. List, Adv. Mater., 2002, 14, 477.
5. A. J. Cadby, P. A. Lane, H. Mellor, S. J. Martin, M. Grell, C. Giebeler, D. D. C. Bradley, M. Wohlgenannt, C. An and Z. V. Vardeny, Phys. Rev. B, 2000, 62, 15604.
6. V. N. Bliznyuk, S. A. Carter, J. C. Scott, G. Klärner, R. D. Miller and D. C. Miller, Macromolecules, 1999, 32, 361.
7. D. Beljonne, G. Pourtois, C. Silva, E. Hennebicq, L. M. Herz, R. H. Friend, G. D. Scholes, S. Setayesh, K. Müllen and J. L. Brédas, Proc. Natl. Acad. Sci. USA, 2002, 99, 10982.
8. L. M. Herz and R. T. Phillips, Phys. Rev. B, 2000, 61, 13691.
9. K.-H. Weinfurtner, H. Fujikawa, S. Tokito and Y. Taga, Appl. Phys. Lett., 2000, 76, 2502.
10. J. M. Lupton, M. R. Craig and E. W. Meijer, Appl. Phys. Lett., 2002, 80, 4489.
11. E. J. W. List, R. Guentner, P. S. de Freitas and U. Scherf, Adv. Mater., 2002, 14, 374.
12. W. E. Moerner, Single Molecule Spectroscopy, Springer Ser. Chem. Phys., 2001, 67, 32.
13. S. Weiss, Science, 1999, 283, 1676.
14. X. S. Xie and J. K. Trautman, Annu. Rev. Phys. Chem., 1998, 49, 441.
15. T. Basche, W. E. Moerner, M. Orrit and U. P. Wild, Single Molecule Optical Detection, Imaging, and SpectroscopyVerlag Chemie, Munich, 1996.
16. J. Yu, D. Hu and P. F. Barbara, Single Molecule Spectroscopy, Springer Ser. Chem. Phys., 2001, 67, 114.
17. T. Huser, M. Yan and L. J. Rothberg, Proc. Natl. Acad. Sci. USA, 2000, 97, 11187.
18. T.-Q. Nguyen, J. Wu, V. Doan, B. J. Schwartz and S. H. Tolbert, Science, 2000, 288, 652.
19. D. A. Vanden Bout, W.-T. Yip, D. Hu, D.-K. Fu, T. M. Swager and P. F. Barbara, Science, 1997, 277, 1074.
20. D. Hu, J. Yu and P. F. Barbara, J. Am. Chem. Soc., 1999, 121, 6936.
21. D. Hu, J. Yu, K. Wong, B. Bagchi, P. J. Rossky and P. F. Barbara, Nature, 2000, 405, 1030.
22. J. Yu, D. Hu and P. F. Barbara, Science, 2000, 289, 1327.
23. D. Hu, J. Yu, G. Padmanaban, S. Ramakrishanan and P. F. Barbara, Nano Lett.,, 2002, 2, 681.
24. H. Bassler and B. Schweitzer, Acc. Chem. Res., 1999, 32, 173.
25. S. J. Park, A. J. Gesquiere, J. Yu and P. F. Barbara, J. Am. Chem. Soc., 2004, 126, 4116.
26. S. Setayesh, A. C. Grimsdale, T. Weil, V. Enkelmann, K. Müllen, F. Meghdadi, E. J. W. List and G. Leising, J. Am. Chem. Soc., 2001, 123, 946.
27. R. Schroeder, W. Graupner, U. Scherf and B. Ullrich, J. Chem. Phys., 2002, 116, 3449.
28. J. Teetsov and D. A. Vanden Bout, Macromol. Symp., 2001, 167, 153.
29. J.-I. Lee, G. Klaerner and R. D. Miller, Chem. Mater., 1999, 11, 1083.
30. G. Padmanaban and S. Ramakrishnan, J. Am. Chem. Soc., 2000, 122, 2244.
31. S. Mukamel, S. Tretiak, T. Wagersreiter and V. Chernyak, Science, 1997, 277, 781.
32. H. Meier, U. Stalmach and H. Kolshorn, Acta Polym., 1997, 48, 379.
33. G. Klaerner and R. D. Miller, Macromolecules, 1998, 31, 2007.

---

Footnote

† Dedicated to Professor Hiroshi Masuhara on the occasion of his 60th birthday.

---