Energy and electron transfer in bifunctional non-conjugated dendrimers

Article abstract of DOI:10.1021/ja044778m

Nonconjugated dendrimers, which are capable of tunneling energy from the periphery to the core followed by a charge-transfer process from the core to the periphery, have been synthesized. The energy and electron donors involve a diarylaminopyrene unit and

Full text of DOI:10.1021/ja044778m

Published on Web 12/16/2004
Energy and Electron Transfer in Bifunctional Non-Conjugated
Dendrimers
K. R. Justin Thomas,† Alexis L. Thompson,‡ Aathimanikandan V. Sivakumar,†
Christopher J. Bardeen,*,‡ and S. Thayumanavan*,†
Contribution from the Department of Chemistry, UniVersity of Massachusetts,
Amherst, Massachusetts 01003 and Department of Chemistry, UniVersity of Illinois,
Urbana, Illinois 61801
Received August 30, 2004; E-mail: bardeen@scs.uiuc.edu; thai@chem.umass.edu
Abstract: Nonconjugated dendrimers, which are capable of funneling energy from the periphery to the
core followed by a charge-transfer process from the core to the periphery, have been synthesized. The
energy and electron donors involve a diarylaminopyrene unit and are incorporated at the periphery of these
dendrimers. The energy and electron acceptor is at the core of the dendrimer, which involves a chromophore
based on a benzthiadiazole moiety. The backbone of the dendrimers is benzyl ether based. A direct electron-
transfer quenching of the excited state of the periphery or a sequential energy transfer-electron-transfer
pathway are the two limiting mechanisms of the observed photophysical properties. We find that the latter
mechanism is prevalent in these dendrimers. The energy transfer occurs on a picosecond time scale,
while the charge-transfer process occurs on a nanosecond time scale. The lifetime of the charge separated
species was found to be in the range of microseconds. Energy transfer efficiencies ranging from 80% to
90% were determined using both steady-state and time-resolved measurements, while charge-transfer
efficiencies ranging from 70% to 80% were deduced from fluorescence quenching of the core chromophore.
The dependence of the energy and charge-transfer processes on dendrimer generation is analyzed in
terms of the backfolding of the flexible benzyl ether backbone, which leads to a weaker dependence of the
energy and charge-transfer efficiencies on dendrimer size than would be expected for a rigid system.
Introduction
systems are extremely efficient in their native conditions, these
systems are neither cheap nor robust enough to be widely useful
Designing efficient light harvesting materials is an important
goal, considering the need for effectively using renewable energy
sources.1 Nature has already evolved an effective pathway for
harvesting sunlight in to useful chemical energy that is stored
in the form of ATP. Nature’s light harvesting assembly, as
exemplified in purple photosynthetic bacteria, consists of a
multichromophore array that is capable of absorbing photons
of a broad spectral range from the sunlight.2 The placement of
the chromophores relative to the reaction center plays a crucial
role in the high efficiency of this light harvesting process.3 The
photosynthetic apparatus involves two key steps: (i) electronic
energy transfer (EET) where the solar energy is funneled to a
reaction center and (ii) a sequence of charge transfer (CT) events
from the excited state of the reaction center that ultimately
results in chemical energy. While the nature’s light harvesting
as components of solar cells. Considering this, several groups
have been interested in developing light harvesting systems with
custom-designed molecules.4 These molecules include small
covalent arrays containing photoactive units,5 or supramolecu-
lar,6 and polymeric7 systems containing multichromophore
arrays.
Dendrimers,8 which consist of peripheral units, a core, and
intervening branching units, are interesting scaffolds for light
harvesting applications. The decreasing number of functionalities
from the periphery to the core of the dendrimer makes these
molecules excellent candidates for light harvesting antenna
(Figure 1). The periphery of the dendrimers functionalized with
light absorbing chromophores could funnel the energy to a lower
energy acceptor at the core. Since dendrimers are assembled
layer-by-layer and can be synthesized with excellent control in
molecular weight, control over the relative placement of
photoactive units can be easily achieved.8,9 Considering this
possibility, several groups have utilized dendrimers as light
† Department of Chemistry, University of Massachusetts.
‡ Department of Chemistry, University of Illinois.
(1) Hoffert, M. I.; Caldeira, K.; Jain, A. K.; Haites, E. K.; Harvey, L. D. D.;
Potter, S. D.; Schlessinger, M. E.; Schnedier, S. H.; Watts, R. G.; Wigley,
T. M.; Wuebbles, D. J. Nature 1998, 395, 881.
(2) (a) McDermott, G.; Prince, S. M.; Freer, A. A.; Hawthornthwaite-Lawless,
A. M.; Papiz, M. Z.; Cogdell, R. J.; Isaacs, N. W. Nature 1995, 374, 517.
(b) Pullerits, T.; Sundstro¨m, V. Acc. Chem. Res. 1996, 29, 381. (c)
Ku¨hlbrandt, W.; Wang, D. N.; Fujiyoshi, Y. Nature 1994, 367, 614.
(3) (a) Hu, Y. Z.; Takashima, H.; Tsukiji, S.; Shinkai, S.; Nagamune, T.; Oishi,
S.; Hamachi, I. Chem. Eur. J. 2000, 6, 1907. (b) Hu, X.; Damjanovic, A.;
Ritz, T.; Schulten, K. Proc. Nat. Acad. Sci. U.S.A. 1998, 95, 5935. (c)
Oijen, A. M. V.; Ketelaars, M.; Ko¨hler, J.; Aartsma, T. J.; Schmidt, J.
Science 1999, 285, 400.
(4) (a) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2001, 34, 40.
(b) Wasielewski, M. R. Chem. ReV. 1992, 92, 435. (c) Fox, M. A. Acc.
Chem. Res. 1999, 32, 201. (d) Balzani, V. Tetrahedron 1992, 48, 10443.
(e) Wagner, R. W.; Johnson, T. E.; Lindsey, J. S. J. Am. Chem. Soc. 1996,
118, 11166. (f) Kurreck, H.; Huber, M. Angew. Chem., Int. Ed. Engl. 1995,
34, 849. (g) Hagfeldt, A.; Gra¨tzel, M. Acc. Chem. Res. 2000, 33, 269. (h)
Choi, M.-S.; Yamazaki, T.; Yamazaki, I.; Aida, T. Angew. Chem., Int. Ed.
2004, 43, 150.
9
10.1021/ja044778m CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 373-383
373

A R T I C L E S
Justin Thomas et al.
form a charge separated state. In such a system, the donor fulfills
two roles: energy harvester and electron donor. It is easy to
imagine that dendrimers are also interesting scaffolds for
photoinduced charge transfer from the core to the periphery.
As one moves from the core to the periphery of the dendrimers,
the number of units at least doubles with each layer. Therefore,
dendrimers could provide an entropic driving force for charge
separation from the core to the periphery. Guldi et al. reported
conjugated phenylene vinylene based dendrimers, in which an
energy level dependent energy transfer and/or electron transfer
was achieved.14a Qu et al. reported polypheneylene based
conjugated, rigid dendrimers in which intramolecular energy
transfer and electron transfer were observed.14b In this paper,
we report on a nonconjugated dendrimer, in which energy
transfer from the periphery to the core of the dendrimer is
followed by an electron-transfer quenching of the core excited
state. In conjugated dendrimers, the backbone serves as a vehicle
for the through-bond communication both for electron and
energy transfer. However, in nonconjugated dendrimers, the role
Figure 1. Cartoon of dendrimers for light harvesting.
harvesting antennae, in which efficient energy transfer from the
periphery to the core of the dendrimer has been observed.10
Studies have been carried out both with conjugated11 and
nonconjugated12 dendrimers.
There has also been considerable effort to design dendrimers
with favorable electron-transfer properties.13 Efforts to combine
these two steps, energy transfer and charge transfer, to mimic
both preliminary events of the photosynthetic apparatus, have
been more limited.14 This sequence of events is outlined in
Figure 1, where a donor molecule first transfers its energy to
the core acceptor, which then is able to oxidize the donor and
(8) (a) Newkome, G. R.; Moorefield, C. N.; Vo¨gtle, F. Dendrimers and
Dendrons: Concepts, Syntheses, Applications, 2nd ed.; Wiley-VCH: New
York, 2001. (b) Dendrimers and Other Dendritic Polymers; Fre´chet, J. M.
J., Tomlia, D. A., Eds.; John Wiley & Sons: New York, 2002. (c) Bosman,
A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665. (d)
Fischer, M.; Vo¨gtle, F. Angew. Chem., Int. Ed. Engl. 1999, 38, 884. (e)
Fre´chet, J. M. J. Science 1994, 263, 1710. (f) Matthews, O. A.; Shipway,
A. N.; Stoddart, J. F. Prog. Polym. Sci. 1998, 23, 1. (g) Dendrimers II:
Architecture, Nanostructure and Supramolecular Chemistry; Vo¨gtle, F.,
Ed.; Springer: Berlin, 2000; Vol. 210. (h) Zeng, F.; Zimmerman, S. C.
Chem. ReV. 1997, 97, 1681.
(9) For methodologies that allow variations among every monomer unit of a
dendrimer, see: (a) Sivanandan, K.; Britto, S. S.; Thayumanavan, S. J.
Org. Chem. 2004, 69, 2937. (b) Vutukuri, D.; Sivnandan, K.; Thayumana-
van, S. Chem. Commun. 2003, 796. (c) Sivanandan, K.; Vutukuri, D.;
Thayumanavan, S. Org. Lett. 2002, 4, 3751.
(10) (a) Balzani, V.; Ceroni, P.; Maestri, M.; Vicinelli, V. Curr. Opin. Chem.
Biol. 2003, 7, 657. (b) Adronov, A.; Fre´chet, J. M. J. Chem. Commun.
2000, 1701. (c) Balzani, V.; Campagna, S.; Denti, G.; Juris, A.; Serroni,
S.; Venturi, M. Acc. Chem. Res. 1998, 31, 26. (d) Dirksen, A.; De Cola, L.
C. R. Chimie 2003, 6, 873. (e) Serroni, S.; Campagna, S.; Puntoriero, F.;
Loiseau, F.; Ricevuto, V.; Passalacqua, R.; Galletta, M. C. R. Chimie 2003,
6, 883. (f) Reference 4h.
(11) (a) Devadoss, C.; Bharathi, P.; Moore, J. S. J. Am. Chem. Soc. 1996, 118,
9635. (b) Shortreed, M. R.; Swallen, S. F.; Shi, Z.-Y.; Tan, W.; Xu, Z.;
Devadoss, C.; Moore, J. S.; Kopelman, R. J. Phys. Chem. B. 1997, 101,
6318. (c) Mellinger, J. S.; Pan, Y.; Kleiman, V. D.; Peng, Z.; Davis, B. L.;
McMorrow, D.; Lu, M. J. Am. Chem. Soc. 2002, 124, 12002. (d) Wang.
Y.; Ranasinghe, M. I.; Goodson, T., III. J. Am. Chem. Soc. 2003, 125,
9562. (e) Ranasinghe, M. I.; Varnavski, O. P.; Pawlas, J.; Hauck, S. I.;
Louie, J.; Hartwig, J.; Goodson, T., III. J. Am. Chem. Soc. 2002, 124, 6520.
(f) Gronheid, R.; Hofkens, J.; Ko¨hn, F.; Weil, T.; Reuther, E.; Mu¨llen, K.;
DeSchryver, F. C. J. Am. Chem. Soc. 2002, 124, 2418. (g) Weil, T.; Reuther,
E.; Beer, C.; Mu¨llen, K. Chem. Eur. J. 2004, 10, 1398. (h) Liu, D. J.; De
Feyter, S.; Cotlet, M.; Stefan, A.; Wiesler, U. M.; Herrmann, A.; Grebel-
Koehler, D.; Qu, J. Q.; Mullen, K.; De Schryver, F. C. Macromolecules
2003, 36, 5918. (i) Benites, M.d. R.; Johnson, T. E.; Weghorn, S.; Yu, L.;
Rao, P. D.; Diers, J. R.; Yang, S. I.; Kirmaier, C.; Bocian, D. F.; Holten,
D.; Lindsey, J. S. J. Mater. Chem. 2002, 12, 65.
(12) (a) Stewart, G. M.; Fox, M. A. J. Am. Chem. Soc. 1996, 118, 4354. (b)
Jiang, D. L.; Aida, T. J. Am. Chem. Soc. 1998, 120, 10895. (c) Adronov,
A.; Gilat, S. L.; Fre´chet, J. M. J.; Ohta, K.; Neuwahl, F. V. R.; Fleming,
G. R. J. Am. Chem. Soc. 2000, 122, 1175. (d) Neuwahl, F. V. R.; Righini,
R.; Adronov, A.; Malenfant, P. R. L.; Fre´chet, J. M. J. J. Phys. Chem. B
2001, 105, 1307. (e) Hahn, U.; Gorka, M.; Vo¨gtle, F.; Vicinelli, V.; Ceroni,
P.; Maestri, M.; Balzani, V. Angew. Chem., Int. Ed. 2002, 41, 3595. (f)
Choi, M.-S.; Aida, T.; Yamazaki, T.; Yamazki, I. Chem. Eur. J. 2002, 8,
2668.
(13) (a) Lor, M.; Thielemans, J.; Viaene, L.; Cotlet, M.; Hofkens, J.; Weil, T.;
Hampel, C. Mullen, K.; Verhoeven, J. W.; Van der Auweraer, M.; De
Schryver, F. C. J. Am. Chem. Soc. 2002, 124, 9918. (b) Capitosti, G. J.;
Cramer, S. J.; Rajesh, C. S.; Modarelli, D. A. Org. Lett. 2001, 3, 1645. (c)
Sadamoto, R.; Tomioka, N.; Aida, T. J. Am. Chem. Soc. 1996, 118, 3978.
(14) (a) Guldi, D. M.; Swartz, A.; Luo, C.; Go´mez, R.; Segura, J.; Mart´ın, N.
J. Am. Chem. Soc.2002, 124, 10875. (b) Qu, J.; Pschirer, N. G.; Liu, D.;
Stefan, A.; De Schryver, F. C.; Mu¨llen, K. Chem. Eur. J. 2004, 10, 528.
(c) For a study in which the branching units are partially oxidized to generate
a mixed valence compound, which was involved in an electron-transfer
quenching of the excited state of light absorber, see: Belser, P.; von
Zelewsky, A.; Frank, M.; Seel, M. Vo¨gtle, F.; De Cola, L.; Bargelletti, F.;
Balzani, V. J. Am. Chem. Soc. 1993, 115, 4076. (d) Choi, M.-S.; Aida, T.;
Luo, H.; Araki, Y.; Ito, O. Angew. Chem., Int. Ed. 2003, 42, 4060.
(5) (a) Sazanovich, I. V.; Kiramaier, C.; Hindin, E.; Yu, L.; Bocian, D. F.;
Lindsey, J. S.; Holten, D. J. Am. Chem. Soc. 2004, 126, 2664. (b) Tomizaki,
K.-y.; Yu, L.; Wei, L.; Bocian, D. F.; Lindsey, J. S. J. Org. Chem. 2003,
68, 8199. (c) Campagna, S.; Serroni, S.; Puntoriero, F.; Loiseau, F.; De
Cola, L.; Kleverlaan, C. J.; Becher, J.; Sorensen, A. P.; Hascoat, P.; Thorup,
N. Chem. Eur. J. 2002, 8, 4461. (d) Jullien, L.; Canceill, J.; Valeur, B.;
Bardez, E.; Lefe`vre, J.-P.; Lehn, J.-M.; Marchi-Artzner, V.; Pansu, R. J.
Am. Chem. Soc. 1996, 118, 5432. (e) Liddell, P. A.; Kodis, G.; Andreasson,
J.; De la Garza, L.; Bandyopadhyay, S.; Mitchell, R. H.; Moore, T. A.;
Moore, A. L.; Gust, D. J. Am. Chem. Soc. 2004, 126, 4803. (f) Kodis, G.;
Herrero, C.; Palacios, R.; Marno-Ochoa, E.; Gould, S.; De la Garza, L.;
Van Grondelle, R.; Gust, D.; Moore, T. A. Moore, A. L. J. Phys. Chem. B.
2004, 108, 414. (g) Bennett, I. M.; Farfano, H. M. V.; Bogan, F.; Primak,
A.; Liddell, P. A.; Otero, L.; Sereno, L.; Juana, J.; Moore, A. L.; Moore,
T. A.; Gust, D. Nature 2002, 420, 398. (h) Weiss, E. A.; Chemick, E. T.;
Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 2326. (i) Gaimo, J. M.;
Gusev, A. V.; Wasielewski, M. R. J. Am. Chem. Soc. 2002, 124, 8530. (j)
Bell, T. D. M.; Ghiggino, K. P.; Jolliffe, K. A.; Ranasinghe, M. G.;
Langford, S. J.; Shephard, M. J.; Paddon-Row: M. N. J. Phys. Chem. A.
2002, 106, 10079. (k) Biemans, H. A. M.; Rowan, A. E.; Verhoeven, A.;
Vanoppen, P.; Latterini, L.; Foekema, J.; Schenning, A. P. H. J.; Meijer,
E. W.; de Schryver, F. C.; Nolte, R. J. M. J. Am. Chem. Soc. 1998, 120,
11054.
(6) (a) Belser, P.; von Zelewsky, A.; Frank, M.; Seel, C.; Vo¨gtle, F.; De Cola,
L.; Barigelletti, F.; Balzani, V. J. Am. Chem. Soc. 1993, 115, 4076. (b)
Balaban, T. S.; Goddard, R.; Linke-Schaetzel, M.; Lehn, J.-M. J. Am. Chem.
Soc. 2003, 125, 4233. (c) Balaban, T. S.; Eichhofer, A.; Lehn, J.-M. Chem.
Eur. J. 2000, 6, 4047. (d) Kercher, M.; Ko¨nig, B.; Zieg, H.; De Cola, L.
J. Am. Chem. Soc. 2002, 124, 11541. (e) Takahashi, R.; Kobuke, Y.; J.
Am. Chem. Soc. 2003, 125, 2372. (f) Anderson, S.; Anderson, H. L.; Bashall,
A.; McPartlin, M.; Sanders, J. K. M. Angew. Chem., Int. Ed. Engl. 1995,
34, 1096. (g) Drain, C. M.; Nifiatis, F.; Vasenko, A.; Batteas, J. D. Angew.
Chem., Int. Ed. 1998, 37, 2344. (h) Prodi, A.; Indelli, M. T.; Kleverlaan,
C. J.; Scandola, F.; Alessio, E.; Gianferrata, T.; Marzilli, L. G. Chem. Eur.
J. 1999, 5, 2668. (i) Sugou, K.; Sasaki, K.; Iwaki, T.; Kuroda, Y. J. Am.
Chem. Soc. 2002, 124, 1182.
(7) (a) Webber, S. E. Chem. ReV. 1990, 90, 1469. (b) Watkins, D. M.; Fox,
M. A. J. Am. Chem. Soc. 1994, 116, 6441. (c) Fox, H. H.; Fox, M. A.
Macromolecules 1995, 28, 4570. (d) Ng, D.; Guillet, J. E. Macromolecules
1982, 15, 724. (e) Ng, D.; Guillet, J. E. Macromolecules 1982, 15, 728. (f)
Jones, G., II.; Rahman, M. A. Chem. Phys. Lett. 1992, 200, 241. (g) Sein,
J.; Schultze, X.; Adronov, A.; Fre´chet, J. M. J. Macromolecules 2002, 35,
5396. (h) Peng, K. Y.; Chen, S. A.; Fann, W. S. J. Am. Chem. Soc. 2001,
123, 11388.
9
374 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Bifunctional Non-Conjugated Dendrimers
A R T I C L E S
Chart 1. Structures of Dendrimers and Model Compounds
of the backbone is only structural and not functional. Therefore,
studying these dendrimers should be useful in dissecting the
architectural advantages of dendrimers as light harvesting
materials. In addition, a nonconjugated backbone also provides
the ability to independently tune the energy levels of the donors
and acceptors. The flexibility of the dendrimer arms may
enhance solubility, and also provide a way for the molecule to
compact itself via folding. Our molecular system consists of
benzyl ether dendrons functionalized with diarylaminopyrene
units in the periphery attached to benzthiadiazole core. We find
that this series of dendrimers exhibits picosecond energy transfer
from the periphery to the core followed by nanosecond charge
transfer from the core to the periphery. The lifetime of the charge
separated state was found to be in the microsecond time scale.
Furthermore, we find that while the EET rate decreases with
dendrimer branch length, the CT rate (as deduced from
fluorescence quenching of the core) remains essentially un-
changed.
periphery, 3,5-dihydroxybenzyl alcohol as the nonconjugated
repeat unit, and the benzthiadiazole-based chromophore unit as
the core. Syntheses of these dendrimers were approached in a
modular fashion using the convergent approach. Thus, the
benzthiadiazole based core chromophore containing two phe-
nolic functionalities and the diarylaminopyrene containing
dendrons were synthesized separately and then assembled in
the last step (vide infra). Synthesis of the chromophore core 9
was achieved using a two-step one-pot procedure starting form
4,7-bis(5-bromothiophen-2-yl)benz[c][1,2,5]thiadiazole (7) and
the aryl stannane 8. These two compounds were coupled using
the Stille methodology,15 with Pd(PPh3)2Cl2 as the catalyst and
DMF as the solvent. This reaction was followed by a desilylation
procedure with aqueous KF in the same pot (Scheme 1). Note
that aqueous KF is commonly used as a reagent to wash-off
the tin-based byproducts in the Stille coupling reactions. In our
case, this solution also served to cleave the tert-butyldimeth-
ylsilyl protecting groups.
Synthesis of the peripheral unit 6 was achieved using the
palladium catalyzed C-N coupling reaction as the key step
(Scheme 2). To synthesize the aldehyde intermediate 13, we
attempted the reaction between N-phenyl-1-pyrenylamine (10)
Results and Discussion
Synthesis and Characterization. Structures of the dendrim-
ers, control core chromophore 5, and the control peripheral
diarylaminopyrene 6 are shown in Chart 1. The dendrimers
contain diarylaminopyrene moieties as the energy donor at the
(15) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 375

A R T I C L E S
Scheme 1
Justin Thomas et al.
and 18-Crown-6 in THF to obtain the dendrimers 1, 2, 3, and
4 respectively.
1
All new compounds were characterized by H, 13C NMR,
and mass spectrometry. As successive generations were added,
multiple NMR signals attributable to the benzylic protons were
observed. For instance, only one benzylic signal is observed
for compounds 1 and 5. However, four different signals in the
ratio of 8:4:2:1 are observed in the case of 4. All the dendrimers
exhibited the parent ion peak with calculated isotopic distribution
pattern in the MALDI-ToF spectra. Additionally, the purity of
the samples was also elucidated using GPC. All dendrimers
exhibited a single sharp peak in the size exclusion chromatogram
indicating the presence of a single large species. We also
characterized the dendrimers 1-4 using absorption spectra. The
relative number of the diarylaminopyrene periphery units and
the benzthiadiazole core increases with increasing generation.
It is expected then that there will be a linear increase in the
relative absorbance of these two species in dendrimers 1-4.
Our observations are consistent with these expectations (vide
infra).
Scheme 2
Steady-State Absorption and Emission Spectroscopy. We
have used both linear absorption and emission spectroscopy for
the photophysical characterization of the dendrimers. Absorption
spectra of compounds 1, 5, and 6 are shown in Figure 2. The
spectrum of compound 5 represents the absorption characteristics
of the core chromophore, since this molecule does not contain
any diarylaminopyrene donor. Similarly, the spectrum of
compound 6 represents the absorption features of the diary-
laminopyrene unit. The spectrum of compound 1 can be very
well approximated by a weighted sum of the absorption spectra
of compounds 5 and 6. Similar behavior is also observed for
the dendrimers 2-4. This provides evidence for the lack of
direct electronic communication between the diarylaminopyrene
periphery and the benzthiadiazole core in the ground electronic
state.
and 3-bromobenzaldehyde under Hartwig’s conditions (Pd(dba)2,
P(t-Bu)3, t-BuONa, Toluene, 80 °C).16 The reaction failed to
afford any desired compound. Using less active bases such as
K2CO3 or Cs2CO3 produced only trace amounts of product.
When the reaction was carried out with the protected derivative,
2-(3-bromophenyl)-1,3-dioxolane (11), the product 12 was
obtained in 82% yield. Deprotection of the acetal moiety by
treatment with acetic acid resulted in the aldehyde 13 in 95%
yield. The aldehyde moiety was then reduced to the correspond-
ing hydroxymethyl group to afford the peripheral triarylamine
6. This compound was used as the model for the energy donor
component of the dendrimers 1-4. The hydroxymethyl group
was then converted to the corresponding bromomethyl com-
pound 14 upon treatment with NBS/PPh3. Compound 14 was
used for further elaboration in to the dendrons.
In the convergent assembly of dendrons, compound 14 was
treated with 3,5-dihydroxybenzyl alcohol (15) under Williamson
alkylation conditions to afford the G-1 dendron 16 in 78% yield
(Scheme 3). Conversion of the hydroxymethyl compound 16
to the bromomethyl version 17 was achieved using NBS/PPh3.
Note that the success of this second step depends on the use of
concentrated mixture and short reaction time. Compound 17
was taken through the above two synthetic steps iteratively to
obtain the G-2 and G-3 monodendrons 19 and 21 with a
bromomethyl functionality at the focal point. In the final step
of the assembly, the bromomethyl functionalized compounds
14 (G0), 17(G1), 19 (G2), and 21 (G3) were treated with the
phenol functionalized core 9 in the similar fashion using K2CO3
Comparison of the absorption spectra of the dendrons G0
through G3 (1 through 4) shows that the absorbance of these
compounds increases steadily with increasing generation. This
is attributed to the increasing number of diarylaminopyrene units
with generation. We utilized this feature to further assert the
identity of the chromophore-cored dendrimers 1-4. Figure 3
shows that the absorbance at 395 nm increases steadily with
generation when normalized for the chromophore core absorp-
tion at 490 nm. A plot of the number of pyrene units with the
absorbance at 395 nm for compounds 1-5 exhibits a linear
relationship that passes through the origin (see inset in Figure
3). This further attests to the purity of the dendrimers studied.
The emission spectrum of the control molecule 5 has the
maxima centered at 602 nm, when excited at 490 nm. Similarly,
excitation of the diarylaminopyrene molecule 6 at 395 nm
exhibits an emission spectrum with the maxima centered at 464
nm. The emission maxima for the dendrimers 1-4 is fairly
insensitive to the excitation wavelength, with most of the
emission centered around 600 nm. Since this peak corresponds
to the chromophore core, this is taken as preliminary evidence
to suggest that there is an efficient energy transfer from the
diarylaminopyrene unit to the chromophore core in all these
dendrimers. The efficiency of the energy transfer process was
quantified using both steady-state and time-resolved spectro-
scopic data (vide infra).
(16) (a) Hartwig, J. F. Angew. Chem., Int. Ed. Engl. 1998, 37, 2046. (b) Wolfe,
J. P.; Wagaw, S.; Marcoux, J.-F.; Buchwald, S. L. Acc. Chem. Res. 1998,
31, 805. (c) For a C-N bond formation methodology that uses a cheaper
copper based catalyst, see: Gujadhur, R. K.; Bates, C. G.; Venkataraman,
D. Org. Lett. 2001, 3, 4315.
9
376 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Bifunctional Non-Conjugated Dendrimers
A R T I C L E S
Scheme 3
Energy Level Estimations. As mentioned above, we hy-
pothesized that these dendrimers should be capable of sequen-
tially funneling energy from the chromophores at the periphery
to the chromophore at the core followed by a charge-transfer
quenching of the excited core by the functionalities in the
periphery. In order for these processes to occur, the core and
peripheral functionalities should have appropriate relative ener-
gies. We estimated the HOMO and LUMO energy levels using
the combination of absorption and emission spectra along with
electrochemical studies. The intersection of the absorption and
emission spectra was assumed to be an estimate of ∆E0-0 gap
for the chromophores. Oxidation or reduction potential deter-
mined by electrochemistry were taken to be an estimate of the
HOMO or LUMO of the molecule, respectively.
Cyclic voltammograms of the compounds 1 and 6 are shown
in Figure 4. Redox potentials are reported with respect to
ferrocene/ferrocenium couple as the internal standard. The
Figure 3. Comparison of absorption spectrum of compounds 1-5. The
spectrum is normalized at the low-energy peak that corresponds to the
chromophore core. Inset: Plot of Number of pyrene in dendrimers vs
absorbance at 395 nm.
Figure 2. Absorption spectrum of compounds 1, 5, and 6.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 377

A R T I C L E S
Justin Thomas et al.
Figure 5. Absorption and emission spectra of compounds 5 and 6. Note
the overlap between the emission spectrum of 6 with the absorption spectrum
of 5.
Figure 4. Cyclic voltammogram of compounds 1 and 6.
tures of 5 and 6 showed that intermolecular EET was nonexistent
under our experimental conditions. To quantify the extent of
energy transfer in compounds 1-4, excitation spectra of these
compounds were obtained at λem of the core at 603 nm. The
excitation spectra were then compared with the respective
absorption spectra. The relative ratio of the peaks corresponding
to the donor, while normalizing for the peaks corresponding to
the acceptor, was used to estimate the energy transfer ef-
ficiency.17 The estimated quantum yields for energy transfer
are 87%, 85%, 81%, and 79% for compounds 1, 2, 3, and 4
respectively. These results show that energy transfer from the
periphery to the core is the major pathway. This also shows
that direct quenching of the excited state of the diarylaminopy-
rene by electron transfer is very small, if any. The fact that
these energy transfer efficiencies are relatively insensitive to
the polarity of solvents further supports the mechanism.
These high EET efficiencies are consistent with the picosec-
ond time-resolved data shown in Figure 6 as well. In Figure
6a, we show the picosecond fluorescence decay for G0-G3,
along with the results of a biexponential fit to the data. The fits
consist of a fast component whose lifetime increases from 21
ps in G0 to 162 ps in G3, and a small, long-lived component
with a lifetime which is obtained from the decay of 6. Note
that the long-lived component of the decay appears flat on the
time scale of the figure. The results of our fits to the donor
decays in DMF are summarized in Table 1. The short-time
component is relatively insensitive to solvent, although it does
decrease measurably in the two larger dendrimers, G2 and G3,
by ∼34% in toluene and ∼18% in CH2Cl2 relative to the times
obtained in DMF. In Figure 6b, we also show the short-time
dynamics of the acceptor emission at 600 nm. As expected in
a light-harvesting system, the acceptor exhibits a fast rise on a
time scale similar to the donor decay.
benzthiadiazole core exhibited an oxidation potential of 595 mV
and a reduction peak at -1674 mV. This corresponds to a
HOMO-LUMO gap of about 2.27 eV. This value is close to
the estimate of 2.18 eV that utilizes the ∆E0-0 calculated from
the absorption and emission spectra. The oxidation potential of
the diarylaminopyrene periphery 6 is centered at 444 mV. The
∆E0-0 for compound 6 was estimated to be 2.82 eV using the
intersection of absorption and emission spectra. These estimates
clearly suggest that the diarylaminopyrene functionality in the
periphery of compounds 1-4 is capable of transferring energy
to the benzthiadiazole core. It is also evident from these
estimates that the excited state of the chromophore core could
be reduced by the diarylaminopyrene moiety. As mentioned
above, our dendrimer is designed to funnel the energy from the
periphery to the core followed by a charge transfer process. Both
these processes seem to be energetically feasible. It is to be
noted however that it is also energetically feasible to obtain
the final charge transfer product by directly quenching the
excited state of the diarylaminopyrene periphery by electron
transfer. To distinguish these possibilities, it is necessary to
estimate the extent of energy transfer from the periphery to the
core. This is addressed both using steady-state spectroscopy and
time-resolved spectroscopy (see below). Time- and wavelength-
resolved fluorescence measurements are obtained using a streak
camera (Hamamatsu, streak scope C4334) attached to a
spectrometer. The solutions are excited using 150 fs, 400 nm
pulses, and the fluorescence is collected at 90° relative to the
excitation beam. The emission is monitored from 420 to 750
nm, and no time dependence of the spectrum is observed.
Additional information can be found in the Supporting Informa-
tion.
Energy and Charge-Transfer Studies. Linear absorption
and emission spectra of compounds 5 and 6 are shown in Figure
5. The overlap between the emission spectrum of 6 and the
absorption spectrum of 5 indicates that Fo¨rster energy transfer
is possible from diarylaminopyrene to the benzthiadiazole core.
A calculation of the Fo¨rster radius for this donor-acceptor pair
yields an R0 of 48 Angstroms, which is quite large. Although
excitation at 395 nm should excite mostly the diarylaminopy-
rene, the fluorescence of compounds 1-4 is dominated by
emission from the benzthiadiazole core. Experiments on mix-
The origin of the small long-lived component, whose lifetime
is similar to that of the isolated donor, is an open question. Such
a long-lived component of the donor population has been
observed previously in other light-harvesting benzyl-ether
dendrimers and ascribed to an energy transfer saturation
phenomenon.12d The suggestion in those femtosecond pump-
(17) Haugland, R. P.; Yguerabide, J.; Stryer, L. Proc. Natl. Acad. Sci. U.S.A.
1969, 63, 23.
9
378 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Bifunctional Non-Conjugated Dendrimers
A R T I C L E S
taneously on a single dendrimer should be roughly proportional
to the square of the light intensity. The relative amplitudes of
the fast and slow decay components were observed to be
independent of laser power over 2 orders of magnitude, however,
which effectively rules out this mechanism for the long decay.
Another possible explanation is that the long-lived component
results from residual donor impurities in our sample. Repeated
purifications of the final product failed to change its amplitude
appreciably, however. Aggregation of the chromophores might
also be expected to lead to a separate, long-lived emission due
to excimer formation. Generally, excimers have a broadened,
redshifted emission with a longer lifetime due to a reduced
radiative rate. This explanation is contradicted by two observa-
tions: that the spectrum of the long-lived species is identical
to that of the short-lived species, i.e., the monomeric pyrene
emission; and its lifetime is equal to or less than that of the
monomer, instead of being longer. A final explanation is that
conformational disorder in the dendrimers leads to a subpopu-
lation where at least one donor is oriented such that its Forster
transfer probability to the core is essentially zero. Such disorder
would be expected to increase for the larger dendrimers,
consistent with the increase in the amplitude of the long-lived
component seen in Table 1. Nevertheless, we cannot rule out
the possibility of impurity emission as well, and so we do not
attempt to analyze the amplitude of the long-lived decay in any
detail. Since in all cases it contributes 10% or less of the total
signal, its role in the overall photophysics of these molecules
is relatively minor.
To quantitatively describe the dendrimer photophysics, we
use a simple kinetic model to describe the EET dynamics. First,
to describe the donor dynamics, we must assume two separate
(1)
donor
(2)
donor
populations of excited donors, N
and N
, whose tem-
poral evolution is described by the equations below
dN (1)
donor ) -(kEET + kdonor)N (1)
(1)
donor
dt
Figure 6. a. Picosecond donor fluorescence decays for G0 (1) (shortest
decay), G1 (2), G2 (3), and G3 (4) (longest decay) in DMF. Biexponential
fit lines for each molecule are shown in red. One time constant is fixed to
the value fit from a longtime decay of the donor 6 fluorescence. Values are
in Table 1. Figure 6b. Acceptor fluorescence signal for G3 (4) in DMF
with triexponential fit (red) as discussed in the text. In the fitting, the decay
times are fixed to the values for the long time acceptor decay (Table 3)
and the short time donor decay Table 1, with the amplitudes allowed to
vary.
dN (2)
donor ) k
N
(2)
(3)
(2)
donor donor
dt
dN
dt
acc ) -kaccNacc + kEET
N
(1)
donor
(1)
donor
where N
represents the donor population that undergoes
Table 1. Fit Parameters for Curves Shown in Figure 6aa
EET to the core, while the population N(d2o)nor only undergoes its
intrinsic fluorescence decay given by kdonor. Since the donor
and acceptor emissions are well separated in these compounds,
the two populations can be monitored independently. The
observed fluorescence signals of both the donor and acceptor
are then given by the biexponential functions
A1
τ1 (ps)
A2
τ2 (ns)
G0
G1
G2
G3
0.889
0.927
0.911
0.902
21.2
61.3
104.8
162.2
0.025
0.045
0.077
0.097
9.88
9.88
9.88
9.88
a The τ2 times, which are obtained from the decay of 6, are not allowed
to vary in the fitting.
(1)
donor
Fdonor(t) ) N
(0) exp[-(kEET + kdonor)t] +
(2)
N
(0) exp[-kdonort] (4)
probe experiments was that high power excitation could excite
two donors within a single dendrimer, only one of which could
undergo energy transfer. The other donor would then remain
in its excited state for the duration of its normal excited state
lifetime. If this were the case, then one would expect a strong
dependence of the slow decay component on the laser pulse
energy, since the probability of having two excitations simul-
donor
(1)
kEET
N
(0)
donor
Facc(t) )
exp[-(kEET + kdonor)t] +
(
)
kacc - kEET - kdonor
(1)
k
N
(0)
EET donor
Nacc(0) -
exp[-kacct] (5)
(
)
kacc - kEET - kdonor
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 379

A R T I C L E S
Justin Thomas et al.
Table 2. Fit Parameters for Curve Shown in Figure 6ba
(1)
(2)
N
(0), N
(0) and Nacc(0) are the initial excited-state
donor
donor
populations created by the laser pulse, which are determined
by the relevant concentrations and absorption coefficients.
Equation 4 describes the observed biexponential decay of the
A
τ (ps)
component 1
component 2
component 3
-1.603
1.157
0.385
162.2
900
2490
donor, while eq 5 predicts that the acceptor signal also has two
(1)
donor
components: a rapid rise due to EET from the N
popula-
a The τ times, which are obtained from the decays of 4 and 5, are not
tion followed by a slower decay given by kacc. While the relative
sizes of N(d1o)nor(0) and Nd(2o)nor(0) can be obtained directly from
fitting the donor decay, the relative amplitudes of the prefactors
in eq 5 can be obtained from the measured fluorescence decay
rates (kEET, kdon, and kacc) and by assuming that the initial
populations (Nd(1o)nor(0) and Nacc(0)) are proportional to the
number of each type of chromophore present in the molecule
and their measured absorption coefficients at the excitation
wavelength, 400 nm. Note that if Nacc(0) , N(d1o)nor(0), then the
amplitudes of the two components are essentially equal. In the
case of the G3 dendrimer, for example, we obtain kEET in the
usual manner from τ1, the fast component of the donor decay
allowed to vary.
1
1
1
kEET
)
)
-
(6)
τEET τ1 τdonor
Since there are 16 donors and a single acceptor in G3, and the
relative ratios of the absorption coefficients are known, we have
Figure 7. Measured electronic energy transfer times (τEET) are shown as
squares. The triangles are the electronic energy transfer times calculated
from the sum of the energy transfer rates using the donor-acceptor
distances, R, from MM2 minimizations. The N6 (dashed) and N3 (solid)
lines are shown normalized to the G0 value, where N is the number of
bonds between the donor and acceptor for each molecule. Distance values
are shown in Table 3.
N (d1o)nor(0) 0.9 × 16 × ꢀdonor(400 nm)
Nacc(0) 1 × ꢀacc(400 nm)
where we have taken into account that in G3 10% of the donors
do not participate in the EET. Using the values ꢀdonor(400 nm)
) 11800 and ꢀacc(400 nm) ) 5500, we find that the relative
amplitudes of rise and decay terms should be equal to within a
few percent. If we fix the acceptor and donor decay rates to be
those in Tables 1 and 4, and then allow the amplitudes of the
multiexponential fit to vary, then we do find that the amplitude
of the rising component and the sum of the two decaying
components are the same to within 5%. Figure 6b shows the
signal calculated using the parameters discussed above, along
with the times and amplitudes listed in Table 2, overlaid with
the experimental signal. As can be seen from the figure, the
agreement between the two is reasonably good. Similar calcula-
tions can be done for G0-G2, but are complicated by the more
rapid EET rate, which means that the time-response of the
detection system must be taken explicitly into account. This
complicates the fitting, although preliminary data workup shows
that the expected trends are present in the other data sets as
well. In any case, the results shown in Figure 6 demonstrate
that our model reproduces not only the rapid EET seen in the
donor fluorescence, but also the evolution of the acceptor
fluorescence.
higher than those obtained from the steady-state data. The origin
of this discrepancy probably lies in our assumption that the only
contributions to the donor decay are energy transfer kEET and
its intrinsic lifetime kdonor which is measured using compound
6. In reality, there can be other nonradiative channels in the
dendrimers (donor aggregation, conformational distortion, and
rapid electron transfer) which quench the donor fluorescence
without leaving an excited acceptor. The steady-state calcula-
tions automatically take such processes into account, while our
simple kinetic model neglects them and lead to an overestima-
tion of the EET efficiencies.12c,18 In light of the high efficiencies
calculated using both the steady-state and time-resolved data,
these channels are relatively minor in the compounds studied
here, and our neglect of them in the modeling does not affect
the main conclusions of this paper.
Unlike the nanosecond decay time, the short-time decay of
the donor luminescence shows a marked dependence on
dendrimer size. The Forster radius R0 for the pyrene-benzthia-
diazole donor-acceptor pair is 48 Angstroms in toluene, and
all the dendrimers are smaller than this length, so the rapid EET
times are consistent with a Forster mechanism. In Figure 7 we
plot τEET, obtained using eq 6, as a function of dendrimer
generation. Although the fluorescence lifetimes of the donor
differ between solvents, we expect that the R0 values will also
change because of a solvent dependent fluorescence quantum
yield, and therefore the tEET curves will be similar for the
different solvents. Although we only show the values obtained
for the dendrimers in DMF, very similar curves are obtained
Using the parameters extracted from our time-resolved data,
we can make a separate estimation of the EET efficiencies using
the relation
kEET
k(Folbs) - kdonor
ηEET
)
)
(7)
(obs)
kEET + kdonor
k
Fl
We obtain efficiencies of 0.97, 0.95, 0.91, and 0.89 for G0-
G3, respectively. These efficiencies are consistently about 10%
(18) Mugnier, J.; Pouget, J.; Bourson, J.; Valeur, B. J. Lumin. 1985, 33, 273.
9
380 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Bifunctional Non-Conjugated Dendrimers
A R T I C L E S
for toluene and CH2Cl2. If the EET is due to a Fo¨rster process,
we expect τEET to scale as R6, where R is the center-to-center
distance between the donor and acceptor. The question then
arises as to how R changes as the size of the dendrimer increases.
One approach is to simply assume that R is directly proportional
to the number of intervening bonds. This basically assumes a
fully extended structure of the dendrimer, as shown in Chart 1,
where the overall molecular size increases linearly with N, the
number of bonds between donor and acceptor. In this case, τEET
R6 N,6 which allows us to calculate how τEET should scale
with dendrimer generation. The results of such a calculation
are shown in Figure 7, where the calculated rates have been
scaled to match the measured G0 value. This simple estimation
of the energy transfer rate clearly overestimates the sensitivity
of τEET to dendrimer size. This is not surprising, since the
benzyl-ether linkages are known to be quite flexible. A more
realistic model would take the flexibility of the dendrimer arms
x
Table 3. Distances for EET Calculations from N, N, and MM2
Calculations, along with the Experimental τEET Values and the
Distances Computed from them
average D−A
distance and
standard deviation
in Å from
experimental R
(Å) calculated
using eq 8
τ
EET (ps)
xN
from eq 6
MM2 calculations
N
G0
G1
G2
G3
21.2
61.7
106
19.3
23.1
25.3
27.2
18.8 ( 0.6
22.6 ( 2.6
24.7 ( 4.9
28.9 ( 4.8
15
20
25
30
3.87
4.47
5
167
5.48
into account. In the case of a flexible polymer, the actual extent
19
x
of the chain increases as N, rather than N. If we naively
assume that the same physics applies to the benzyl-ether arms
6
3
of the dendrimer, we have τEET R6 ( N) ) N . We see
from Figure 7 that such a calculation does a much better job of
reproducing the experimental results.
x
It is not immediately clear that a model developed for
polymers can also be applied to these relatively small benzyl-
ether dendrimers, so we have also used the MM2 force-field to
examine the low energy conformers of G0-G3. As has been
observed previously,20 we see considerable backfolding in these
molecules with multiple conformations separated by less than
kT, the available thermal energy at room temperature. In Figure
7, we plot the τEET dependence on generation as estimated from
the MM2 calculations, assuming isotropic transfer and R0 )
48 Angstroms. The total EET rate is taken to be the average of
the rates due to each donor in the low energy structure
Figure 8. Acceptor emission of G2 dendrimer 3 shown in toluene (blue),
methylene chloride (red), and DMF (black).
Our results in these dendrimers are consistent with flexible,
backfolded arms that enable very rapid EET even in the largest
dendrimer, G3. The flexibility of the benzyl-ether linkages
means that τEET only scales as the square root of dendrimer
size, rather than linearly as would be expected in a rigid
structure. The fact that each donor also includes a tri-arylamine
(TAA) group containing pyrene led us to investigate whether
charge transfer could also be observed in these molecules.
Electron transfer from the TAA group to the excited benzthia-
diazole is expected to quench the core fluorescence at 600 nm.
Furthermore, this quenching should become more pronounced
in polar solvents, which increase the CT rate. Indeed, this is
what is seen in Figure 8, where the acceptor decay for the G2
dendrimer becomes more rapid as we go from nonpolar solvents
(toluene) to very polar solvents (DMF). This trend of more rapid
fluorescence decays in more polar solvents is the same for the
other dendrimers as well. In general, the fluorescence decays
are slightly nonexponential, and in some cases were fit using
biexponentials. The results of our fits to the acceptor fluores-
cence decays in toluene, CH2Cl2, and DMF for dendrimers G0-
G3 are given in Table 4. From the data in Table 4, we estimate
the CT efficiency using the relation
6
M
R0
1
1
)
(8)
∑
6
τEET τdonor
i)1
Ri
where Ri is the distance of each donor in the structure, and M
ranges from 2 (G0) to 16 (G3). The average distances obtained
for the calculations, along with the standard deviations, are
shown in Table 3, along with the R distances obtained using
the standard Fo¨rster relation
R6
τEET ) τ
(9)
donor R0
6
with τdonor ) 4.95 ns, the lifetime of the amino-pyrene 6 in the
absence of the dendrimer. The agreement between the experi-
mentally measured τEET and that calculated from the MM2
distances is fairly good, considering that the MM2 calculations
were only a single run, which is not guaranteed to find the global
minimum. The reason multiple runs were not done is that these
runs are prohibitively slow for the larger dendrons.
0
kCT kacc - kacc
ηCT
)
)
(10)
kacc
kacc
where kacc0 is the fluorescence decay rate of the acceptor without
the attached dendrimer, and where kacc, in the case of a
biexponential decay is given by
(19) Tanford, C. Physical Chemistry of Macromolecules; Wiley & Sons: New
York, 1961.
(20) For example of studies suggesting backfolding, see: (a) Wooley, K. L.;
Klug, C. A.; Tasaki, K.; Schaefer, J. J. Am. Chem. Soc. 1997, 119, 53. (b)
Gorman, C. B.; Hager, M. W.; Parkhurst, B. L.; Smith, J. C. Macromol-
ecules 1998, 31, 815.
A + B
AτA + BτB
kacc
)
(11)
Using these relations and the values in Table 4, for example,
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 381

A R T I C L E S
Justin Thomas et al.
Table 4. Acceptor Fluorescence Decays (excited at 400 nm) and CT Efficiencies Calculated Using eq 10a
toluene
B
methylene chloride
B
DMF
B
A
τA
τB
ηCT
A
τA
τB
ηCT
A
τA
τB
ηCT
acceptor
G0
G1
G2
G3
6.62
6.59
6.51
6.56
6.98
7.66
4.97
5.36
4.76
5.49
7.49
1.72
2.57
2.24
2.49
0.0045
0.0166
0.0553
0.0042
0.351
0.461
0.566
0.518
0.770
0.702
0.789
0.805
0.64
0.62
0.60
0.36
0.38
0.40
1.95
0.98
1.00
0.76
0.51
0.37
0.24
0.49
0.63
1.17
0.89
0.90
0.95
0.93
0.05
0.07
0.44
1.43
a Decays with two time scales are fit with a biexponential fit of the form A exp[(-t/τA)] + B exp[(-t/τB)]. All times are in ns.
oxidation methods, its long lifetime and its behavior with solvent
polarity, we can assign the induced absorption to the radical
cation of the TAA donor group.
The photophysics indicate that after EET from the peripheral
donors to the core acceptor, an electron is then transferred from
one of the donors to the excited core. Due to experimental
limitations, we were unable to independently obtain a quantum
yield for the generation of the CT species from the transient
absorption experiments, and so we must rely on the fluorescence
quenching to estimate the CT quantum yield. We note that these
efficiencies represent an upper bound on the actual yield, since
they do not take into account the possible contributions of other
nonradiative processes (intersystem crossing, internal conver-
sion) which might also be solvent dependent. There is no
experimental evidence that such processes play a significant role
in the current experiments, e.g., no large triplet absorption was
observed in the transient absorption experiments. Also, we have
synthesized dendrimers with benzyl ether arms without the
diarylaminopyrene donors incorporated in them. The fluores-
cence lifetime of the core benzthiadiazole is unaffected by the
presence of various generations of benzyl ether dendrons. The
lifetimes range from 6 ns in toluene to 7 ns in CH2Cl2 to 8 ns
in DMF. In other words, there is no evidence that the addition
of benzyl ether backbone leads to an increase in other nonra-
diative processes. The yields obtained from the fluorescence
quenching are reasonably high, and if we multiply the yield for
energy transfer in G3 (0.81) by the yield for CT (0.70), we
obtain an overall efficiency, absorption to charge separated state,
on the order of 50%. This efficiency is probably limited by the
fact that CT must occur within the nanosecond singlet lifetime
of the core benzthiadiazole. If the core could rapidly cross to a
long-lived triplet with favorable redox properties, like C60, then
we would anticipate a significant increase in the CT yield.
Figure 9. Absorption spectrum of the amino-pyrene compound 6 after
oxidation using a stoichiometric amount of SbCl5 (red), after electrochemical
oxidation (blue) and transient absorption of G3 in DMF (squares).
we calculate that G0 in DMF has a 77% yield of charge
separated species, based on the fluorescence quenching alone.
The presence of a charge separated state can be confirmed
using transient absorption on the microsecond time scale, after
the initial singlet excited-state emissions and absorptions have
decayed. We have done nanosecond experiments on the den-
drimers in the same solvents in which we did the fluorescence
measurements. For the acceptor by itself in any solvent, and
for the dendrimers in nonpolar toluene, no long-lived (>10 ns)
transient absorption could be detected, as expected in the
absence of charge transfer. But in both CH2Cl2 and DMF, very
similar induced absorptions in the visible region were observed.
Figure 9 shows this absorption in G3, along with the absorption
measured when compound 6 is oxidized, either using SbCl5 or
electrochemically. A similar absorption is observed for den-
drimers G1 and G2 as well, although our signal-to-noise was
not sufficient to resolve this feature in G0. The absorption
spectra of radical cation of the TAA and of the long-lived
species produced by photoexcitation of the dendrimers are
similar in shape, but the dendrimer transient absorption is
redshifted by about 50 nm. Such a redshift can be due to either
solvatochromic effects within the polarizable dendrimer envi-
ronment or the radical anion of benzthiadiazole as the counte-
rion, which are not present in the model compound. In both
cases, one could expect the observed red shift. The lifetime of
this species depends on solvent polarity. In DMF, the lifetime
of the absorption decays varies from 2 to 3 µs and shows no
systematic trend with dendrimer size. In CH2Cl2, the lifetime
ranges from 0.9 to 1.6 µs for all dendrimers, as expected if the
stability of the charge separated state is decreased in this less
polar solvent. In light of the similarity of the transient absorption
spectrum with that obtained using chemical and electrochemical
Perhaps more interesting is the fact that the CT yield does
not appear to have any significant dependence on dendrimer
size, as can be seen from Table 4. This is especially surprising
in light of the fact that τEET exhibited a strong generation
dependence as seen in Figure 7. The most likely explanation
for this lack of size dependence in the CT rate is that the overall
CT rate only depends on the distance of the closest ground state
electron donor. The backfolding observed in these systems
provides many opportunities for the TAA groups to closely
approach the core. In the case of EET, the incoming photon
excites all donors with equal probability, so even if one is close,
the overall rate will reflect the contribution from the more distant
donors as well. But the CT event will only reflect the presence
of the nearest donor, and if that distance of closest approach is
roughly equal in all the dendrimers, then the net CT rate will
also be the same. As the dendrimer size increases, the average
distance to the peripheral donors increases, but the additional
9
382 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Bifunctional Non-Conjugated Dendrimers
A R T I C L E S
donors also provide more opportunities for conformational
disorder to bring one back into contact with the core. Such an
increase in disorder with increasing size can be seen from the
standard deviation data in Table 3. If this hypothesis is correct,
then the key to maintaining reasonable CT rates with larger
dendrimers is to maintain the flexibility of the backbone.
Other CT mechanisms such as the more classical through-
bond tunneling or solvent assisted mechanisms are possible. If
the CT event is based on a through-bond mechanism, then a
systematic dependence (or generation dependence) would be
expected. However, this trend is complicated by the fact that
the number of TAA units increase with generation, which will
have the opposing effect on the CT quenching efficiency. It is
also possible that the conformational changes by the solvent
has an effect on CT. If this were the case in these dendrimers,
then significant differences in EET efficiencies should have been
observed in different solvents. Therefore, we believe that this
mechanism is less likely. We believe that the observations could
be more simply explained by the mechanism proposed in the
previous paragraph. Experiments that could clearly distinguish
these mechanistic possibilities are also underway in our
laboratories.
transfer states. Estimates based on fluorescence quenching
indicate CT efficiencies as high as 70% in the most polar
solvent, DMF. Although these yields may be taken as upper
limits, the present work clearly demonstrates that these non-
conjugated, flexible dendrimers can simultaneously exhibit
reasonably high energy transfer and electron transfer efficiencies
which scale favorably with size.
The main goal of future work will be to investigate how to
increase energy and charge transfer efficiencies. The energy
transfer efficiency is currently limited by the size of the slowly
decaying donor component, and identifying its origin is the key
to raising the light-harvesting efficiency to close to 100%. To
increase the charge transfer efficiency, one possibility we are
currently pursuing is the incorporation of additional amino
groups in the dendrimer arms to increase the probability of
contact with the core.22 Another area of interest is increasing
the lifetime of the charge separated state. A final question which
we have not addressed is the role of dynamic conformational
fluctuations in enhancing energy or charge transfer processes.
The flexibility of the nonconjugated dendrimer arms not only
allows for more efficient packing, but may also provide transient
opportunities for the approach of the electron donating amino
groups to the core, facilitating charge transfer via a diffusional
encounter mechanism. All of these issues suggest that multi-
functional, nonconjugated dendrimers merit further study.
Experimental Section
All experimental details are provided as Supporting Information to
this paper.
Acknowledgment. This work is supported by U.S. Depart-
ment of Energy, Basic Energy Sciences program. DOE Grant
(DE-FG02-04ER15503 to S.T.; DEFG02-01ER15270 to C.J.B.).
We thank A. Nantalaksakul (UMass) and Dr. Gordon Hug
(Radiation Lab, University of Notre Dame) for help with the
transient absorption measurements. The Radiation Laboratory
of the University of Notre Dame is also supported by the U.S.
Department of Energy, Basic Energy Sciences program. S.T.
is a Cottrell Scholar, C.J.B. is a Sloan Fellow, and A.L.T. is a
NSF Predoctoral Fellow.
Conclusion
We have studied the photophysical properties of a family of
benzyl-ether based dendrimers, which exhibit both light-
harvesting and electron-transfer functionalities. The large (80-
90%) energy transfer efficiencies scale favorably with increasing
dendrimer size. The reason τEET is observed to scale only as
x
N, as opposed to the N scaling expected in a rigid structure,
is the ability of the flexible arms to backfold and decrease the
average distance between donor and acceptor. As observed
previously by Frechet and co-workers,21 this flexibility appar-
ently does not come at the cost of increasing donor aggregation
and decreased energy transfer efficiency. After energy transfer,
these molecules show fluorescence quenching and transient
absorption characteristic of the creation of long-lived charge-
Supporting Information Available: All experimental details
are provided in the Supporting Information. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA044778M
(21) Serin, J. M.; Brousmiche, D. W.; Frechet, J. M. J. Chem. Commun. 2002,
(22) (a) Bronk, K.; Thayumanavan, S. Org. Lett. 2001, 3, 2057. (b) Bronk, K.;
Thayumanavan, S. J. Org. Chem. 2003, 68 , 5559.
2605.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 383