Designing efficient light harvesting materials is animportant goal, considering the need for superior re-newable energy sources. An important feature ofmany multichromophoric artificial and natural lightharvesting systems is the presence of excitations de-localized over several chromophores and the possibil-ity of a wavelike energy transfer mechanism [1–4].Enhancement of the nonlinear optical properties inmultichromophoric systems beyond what is expectedfor isolated chromophores is also associated withstrong electronic coupling [5,6]. Two-photon excited(TPE) fluorescence measurements have demon-strated great potential for many applications such asmicroscopy and bioimaging as well as for TPA cross-section measurements [7–10]. It was also discoveredthat steady-state fluorescence anisotropy measure-ments under two-photon excitation are intrinsicallymore sensitive than one-photon excitation (OPE) dueto its higher anisotropy values that, in some cases,can lead to significant advantages when comparedwith traditional one-photon methods [7–9,11,12].Time-resolved TPE fluorescence measurements werereported on the picosecond time scale, and theyshowed essentially the same dynamics as comparedwith OPE. This may be due to the fast internal con-version of the TPE state to the relaxed fluorescencestate (Vavilov–Kasha rule), which is missed with pi-cosecond experiments [7,8,13]. However, TPE fluores-cence anisotropy measurements with femtosecondtime resolution have not been previously reported.
In this contribution we investigate the propertiesof ultrafast, TPE fluorescence anisotropy of potentialartificial-light-harvesting macrocyclic thiophene sys-tems. We show that femtosecond time-resolved TPEfluorescence anisotropy has strong potential in eluci-dating the properties of the collective excited molecu-lar states and can be utilized to detect the presence ofdelocalized excitations in such systems as the
The structure of the molecules used in this work isshown in Fig. 1. Conjugated oligothiophenes are ar-ranged in a circular geometry comprising ter-thiophene and acetylene units. The synthesis of thismacrocycle system has been reported previously [14].An �-quaterthiophene molecule was used as the ref-erence oligomer representing the linear buildingblock of the ring [15]. Toluene solutions were used inthe experiment.
To probe the excited state dynamics of the chro-mophore, polarized time-resolved fluorescence upcon-version measurements were carried out using eitherthe fundamental (�ex=800 nm, TPE) or the secondharmonic of a Ti-sapphire femtosecond laser (�ex=400 nm, OPE) for excitation. The thiophene macro-cycle was excited to the blue side of its fundamentalabsorption peak at �430 nm [16]. The femtosecondupconversion spectroscopy setup has been describedin our previous reports [16,17]. Briefly the samplewas excited with fundamental or frequency doubledlight from a mode-locked Ti-sapphire laser (Tsunami,Spectra Physics). The luminescence emitted from thesample was upconverted in a nonlinear crystal of�-barium borate using the pump beam at �800 nmthat was first passed through a variable delay line.The fluorescence upconversion unit has an instru-
Fig. 1. Structures of oligothiophene systems investigatedin this work. a, Macrocyclic thiophene oligimer C�3T �DA�5;
ment response function of �200 fs (FWHM) for bothvisible and near-infrared (NIR) excitations. This sys-tem also has the capability to carry out ultrafastemission anisotropy decay measurements using a Be-rek compensator to control the polarization of the ex-citation beam. The energy per excitation pulse didnot exceed 50 pJ for one-photon (visible) excitationand 7 nJ for two-photon (NIR) excitation experi-ments. For the fluorescence anisotropy measure-ments the setup has been calibrated using severalstandard dyes (Perylene, Coumarin 30, Rhodamine6G). The fluorescence of macrocycle has been de-tected near the peak of the fluorescence spectrum at540 nm [16].
Fluorescence anisotropy dynamics for a macrocy-clic thiophene is shown in Fig. 2 in comparison withthat for near linear system of �-quaterthiophene. Itis clearly seen that initial anisotropy for TPE�-quaterthiophene amounts to �0.54, which is muchhigher than that for OPE (�0.38, not shown) in ac-cordance with other observations for near linear mol-ecules using steady-state or time-correlated photoncounting techniques [8,9,12,18]. In the inset to Fig. 2we have also showed TPE time-resolved anisotropyfor reference dye Rhodamine 6G in propylene glycolusing the same setup. Both the initial value of 0.495and flat anisotropy profile for Rhodamine 6G are inexcellent agreement with those obtained with muchlower time resolution reported previously [9]. The ini-tial anisotropy for the symmetrical thiophene ringsystem demonstrated rather different behavior start-ing with a smaller value than that for linear analogand followed by a risetime feature. In general, theTPA cross section for a transition from the groundelectronic state �g� to a final excited state �f� (degen-erate TPA, the sum-of-states approximation) is pro-
Fig. 2. Time-resolved fluorescence anisotropy decay undertwo-photon excitation for (1) macrocyclic thiophene as com-pared with that for (2) �-quaterthiophene. The fluorescencedetection wavelength is 540 nm. The instrument responsefunction for fluorescence intensity is also shown (dashed–dotted curve). 3, Fluorescence anisotropy dynamics forRhodamine 6G at 575 nm.
portional to the squared two-photon tensor [19]
�2 ����i
N 2�p��� ig���� fip� �
�ig − � + i�ig�2� , �1�
where p� is the unit polarization vector for the laserbeam, and �ig and �ig are the frequency and the ho-mogeneous linewidth associated with the i state, re-spectively, �� nm is the transition dipole moment be-tween states n and m �m�n� or the permanentdipole moment for the state n �m=n�. � denotes theaverage over molecular orientation. It is seen fromthis expression that the TPA possesses tensorialcharacter and depends on values, mutual orienta-tions, and resonance conditions of at least two tran-sitions (Fig. 3, inset). The two-photon fluorescenceanisotropy rTPA in turn depends on the mutual orien-tations of at least three dipole moments participatingin TPA and fluorescence, and corresponding to tran-sitions S0–S1��01�, S1–Sf ��1f�, S1–S0��fl� [18–20](inset Fig. 3). For a planar molecule it was shown us-ing a three-state model that TPE fluorescence aniso-tropy rTPE can be expressed as [18]
rTPE =18 cos� − �em�cos��em�cos − 7 cos2 + 1
7�2 cos2 + 1�,
�2�
where is the angle between �01 and �1f and �em isthe angle between �01, �fl. When all three transitiondipoles �01, �1f, and �fl are parallel to each other (lin-ear chromophore) the TPE anisotropy is 0.57—a well-established result supported by many experiments[8,9,18]. On the other hand if the transitions �01 and�1f are perpendicular to each other, rTPA=0.14 inde-pendently of the orientation of emission transitionmoment �fl thus making the anisotropy insensitive toany dynamics associated with reorientation (in the
Fig. 3. Fluorescence anisotropy decay of the macrocyclicthiophene after two- and one-photon excitation. Best fitsfor the anisotropy after convolution procedure are shownby solid curves. Instrument response function for fluores-cence intensity is represented by the dashed-dotted curve.Inset: Best fit fluorescence anisotropy (before convolutionprocedure) for OPE (dash curve) and TPE (solid curve). En-
ergy level structure for three-state model is also shown.
molecular plane) of the fluorescence dipole (e.g., en-ergy transfer).
For the planar molecules possessing axial symme-try C3 or higher with lowest degenerate state the sec-ond transition moment between the excited state de-generate levels �1f oriented perpendicularly withrespect to the fundamental transition from theground state to the excited state �01 [21,22]. As wementioned above, Eq. (2) predicts anisotropy of 0.14with a flat time profile as it is independent of any dy-namics (reorientation) of the fluorescence transitionmoment �fl. However if the excitation is not delocal-ized over the entire molecule the initial anisotropycan be quite high, reaching the 0.57 level for the ex-citation localized on a single linear building block(weak coupling limit). This high initial anisotropywill decay due to the energy transfer between differ-ently oriented chromophores to the residual value0.14 (for the planar system) before rotational diffu-sion starts. Thus the TPA-anisotropy time profile forthe strongly coupled system is expected to be verydifferent from that for the weakly coupled system. Onthe contrary, for OPE anisotropy the fluorescence an-isotropy always starts from high initial values [7,23],and fluorescence anisotropy decay time rather thaninitial anisotropy is used as a key experimental pa-rameter to detect the presence of strong coupling[5,6]. Experimental comparison of the anisotropy dy-namics for TPE with that for OPE is shown in Fig. 3.The OPE anisotropy starts from a high value of 0.4(Fig. 3, inset) typical for isolated chromophores andquickly decays to the residual anisotropy of 0.08. Inthe case of TPE the anisotropy starts from a low ini-tial value, �0.1, which is very different from that forisolated (weakly interacting) chromophores (0.54,Fig. 2). While qualitatively in accordance with theabove theoretical predictions for the strongly coupledring system, the TPE anisotropy showed an unex-pected fast risetime feature followed by the residualanisotropy higher than the theoretically predictedlevel of 0.14. We suggest that this deviation is asso-ciated with the imperfect planarity of the system andpossible inclination of the transition moment(s) fromthe molecular plane, which is not taken into accountin the simple theory given above. The risetime canreflect dephasing and population flow between thedegenerate levels of the symmetrical molecule simi-lar to the decay feature for OPE [23]. The internalconversion process can also contribute to this initialdynamics. However the quantitative aspects of TPEanisotropy in planar macromolecules with distortionswill require additional experimental study and ex-tended modeling.
In conclusion we provide what is to the best of ourknowledge the first systematic investigation of ul-trafast two-photon excited fluorescence anisotropyand have experimentally demonstrated that two-photon excited time-resolved fluorescence anisotropyis a powerful method to directly detect the presenceof delocalized excitations in macromolecules of cyclic
and other axial symmetry topologies that are impor-
tant for both light harvesting applications and non-linear optics. We believe that this new experimentalapproach combined with extended modeling offersnew opportunities for the development of new or-ganic multichromophoric systems possessing delocal-ized electronic excitations with strongly enhancedlight harvesting and nonlinear optical capabilities.
T. Goodson III acknowledges the Army ResearchOffice and the National Science Foundation for sup-port. The work at Ulm University has been supportedby the German Research Foundation (DFG) in theframe of SFB 569.
References
1. G. S. Engel, T. R. Calhoun, E. L. Read, T.-K. Ahn, T.Mančal, Y.-C. Cheng, R. E. Blankenship, and G. R.Fleming, Nature 446, 782 (2007).
2. G. R. Fleming and G. Scholes, Nature 431, 256 (2004).3. M. Dahlbom, T. Pullerits, S. Mukamel, and V.
Sundström, J. Phys. Chem. B 105, 5515 (2001).4. S.-H. Lim, T. G. Bjorklund, F. C. Spano, and C. J.
Bardeen, Phys. Rev. Lett. 92, 107402 (2004).5. T. Goodson III, Annu. Rev. Phys. Chem. 56, 581 (2005).6. T. Goodson III, Acc. Chem. Res. 38, 99 (2005).7. J. R. Lakowicz, Principles of Fluorescence Spectroscopy
(Kluwer, 1999).8. J. R. Lakowicz, I. Gryczynski, Z. Gryczynski, E.
Danielsen, and M. J. Wirth, J. Phys. Chem. 96, 3000(1992).
9. A. Volkmer, D. A. Hatrick, and D. J. S. Birch, Meas.Sci. Technol. 8, 1337 (1997).
10. C. Xu and W. W. Webb, J. Opt. Soc. Am. B 13, 481(1996).
11. W. M. McClain, J. Chem. Phys. 58, 324 (1973).12. W. M. McClain, Acc. Chem. Res. 7, 129 (1974).13. S. V. Frolov, Z. Bao, M. Wohlgenannt, and Z. V.
Vardeny, Phys. Rev. Lett. 85, 2196 (2000).14. J. Krömer, I. R. Carreras, G. Fuhrmannn, C. Munsch,
M. Wunderlin, T. Debaerdemaeker, E. Mena-Osteriritz,and P. Bäuerle, Angew. Chem., Int. Ed. 39, 3481(2000).
15. X. Zhang and A. Matzger, J. Org. Chem. 68, 9813(2003).
16. A. Bhaskar, G. Ramakrishna, K. Hagedorn, O.Varnavski, E. Mena-Osteritz, P. Bäuerle, and T.Goodson III, J. Phys. Chem. B 111, 946 (2006).
17. O. Varnavski, X. Yan, O. Mongin, M. Blanchard-Desce,and T. Goodson III, J. Phys. Chem. C. 111, 149 (2007).
18. J. Fu, O. V. Przhonska, L. A. Padilha, D. J. Hagan, E.W. Van Stryland, K. D. Belfield, M. V. Bondar, Y. L.Slominsky, and A. D. Kachkovski, Chem. Phys. 321,257 (2006).
19. C. Wan and C. K. Johnson, Chem. Phys. 179, 513(1994).
20. P. Cronstrand, Y. Luo, and H. Angren, J. Chem. Phys.117, 11102 (2002).
21. A. M. Kelley, L. C. T. Shoute, M. Blanchard-Desce, G.P. Bartolomew, and G. C. Bazan, Mol. Phys. 104, 1239(2006).
22. D. Beljonne, W. Wenseleers, E. Zojer, Z. Shuai, H.Vogel, S. J. K. Pond, J. W. Perry, S. R. Marder, andJ.-L. Bredas, Adv. Funct. Mater. 12, 631 (2002).
23. K. Wynne and R. M. Hochstrasser, Chem. Phys. 171,
