determined for any given dendrimer.9 As can been seen from
Table 1, the energy transfer efficiencies clearly show a bimodal
distribution between the two geometrical types of dendritic systems
mentioned above. This significant difference in WET can be
explained on the basis of intramolecular energy migration in the
clustering domains of anthracene groups. Indeed, dense packing of
anthracene groups in dendrimer systems has been demonstrated to
reduce fluorescence emission efficiency due to energy migration
processes as observed with benzene-cored anthracene dendrimers.6
Such an interpretation can also be applied to account for the
slightly decreased WET values for 3 and 6 in comparison to those of
the other members that share the common n1 numbers.
On the other hand, it is well accepted that quantum efficiency of
fluorescence resonant energy transfer (FRET) depends strongly on
the separation distances between the donor–acceptor pairs. In fact,
molecular modeling studies of fully extended structures of the
dendrimers indicate that averaged center-to-center distances from
the perylene core to the anthracene groups located at close and
remote sites of the dendritic shells are estimated to be
approximately 1.4 and 2.3 nm, respectively.10 From the viewpoint
of possibilities of long-range processes as suggested by these values,
Coulombic dipole–dipole interactions (Fo¨rster mechanism) would
provide a plausible rationale for the observed results.8 To consider
the mechanistic model, one must recall that this type of FRET
shows an inverse sixth power dependence of probability on the
donor–acceptor separations.9 However, the similarity in the net
WET values given for the related dendrimer series leads to an
apparently contradictory result because fractional contributions
for all possible FRET pathways in each single molecule are
suggested to be closely equivalent regardless of the number of
anthracene groups present at the most remote sites (n2 in Table 1).
In other words, an obvious conclusion that can be drawn from
these data is that every component of the energy transfer exhibits
distance independent behaviour in the dendrimer systems. This
remarkable phenomenon can be rationalized in terms of the
Fo¨rster critical radius (R0), which is a determinant of the extent of
dipole–dipole interaction effects during the FRET processes. From
the theoretical treatments, the FRET efficiency should be near
unity when the donor–acceptor distance is shorter than R0/2.9 In
comparing the R0 value of 3.6 nm reported for an analogous
anthracene–perylene pair,11 it is evident that plausible distances
between the donor–acceptor pairs are much less than R0 and even
closer to R0/2. These geometrical parameters meet the desired
criteria for achieving quantitative processes which provide an
effective solution to the observed lack of distance dependence in
the FRET efficiencies, although there exist significant differences
in the contacts of the core with the two discrete anthracene sites.
In conclusion, we have demonstrated that perylene-cored
anthracene dendrimers represent an excellent light-harvesting
antenna model. The quantitative analysis of the FRET efficiencies
led to a remarkable conclusion that confinement of antenna
elements within the nanoscopic dimensions of dendritic architec-
tures, especially inside the reach of the critical radii, would satisfy a
primary requirement for creating effective light-harvesting func-
tionalities via the Fo¨rster mechanism.12 Extensive research
ascertaining whether multistep energy hopping13–15 is feasible in
the dendrimer systems will be addressed elsewhere.16
Fig. 2 (A) Absorption and (B) normalized fluorescence spectra (lex
=
378 nm) of 1–6 in chloroform solutions (cabs 1026–1025 M, cfl 1028
–
1027 M). Insets: plots of (A) extinction coefficients and (B) excitation
intensities at 359 nm vs. total number of anthracene units (n) in the
dendrimer systems. Errors in fitting the experimental data: extinction
coefficients of 1–6, ¡4%; excitation intensities of 1–3, ¡4%; excitation
intensities of 4–6, ¡7%.
signals assigned to originate from the perylene core with complete
disappearance of the anthracene group emission in the shorter-
wavelength region (400–460 nm). These observations are consistent
with efficient intramolecular energy transfer which may arise from
through-space interactions between the excited anthracene units
and the perylene core.8 Moreover, the excitation spectra of 1–6 at
lem of the core at 500 nm matched well the corresponding
absorption spectra, exhibiting three distinct maxima attributed to
the anthracene groups. In the inset of Fig. 2B, the relative
intensities of these peaks, estimated by measuring the height in the
spectra normalized with respect to the perylene core excitation at
450 nm, were plotted as a function of the total number of
anthracene groups. In contrast to the linear dependence of the
absorption feature as mentioned above, the distribution of these
data can be well fitted with two comparable linear relationships in
accordance with two geometrical types of dendritic architectures
([1, 2, 3] and [4, 5, 6]), which are divided according to the number
of anthracene chromophores at the closest sites of the dendritic
shells (n1 in Table 1).
By comparing the absorption and excitation spectra normalized
at the acceptor peaks, the energy transfer efficiency (WET) was
Table 1 Number of anthracenes and energy transfer efficiencies
1
2
3
4
5
6
a
b
n1
n2
nc
4
0
4
0.69
4
4
8
0.69
4
8
12
0.65
8
0
8
0.54
8
8
16
0.54
8
16
24
0.48
d
WET
a
Number of anthracenes located at close sites of the dendritic shells.
Number of anthracenes located at remote sites of the dendritic
b
c
d
shells.
Total number of anthracene units.
Energy transfer
efficiencies determined by spectroscopic measurements in
chloroform.
This research was supported by a Grant-in-Aid for Scientific
Research from the Ministry of Education, Culture, Sports,
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 3084–3086 | 3085
Takahashi, Masaki
