Large concentration-dependent nonlinear optical responses of starburst diphenylaminofluorenocarbonyl methano[60]fullerene pentads

Article abstract of DOI:10.1039/b615697e

We demonstrate an approach toward the design of starburst C ₆₀-keto-DPAF assembly by applying a starburst macromolecular configuration with C₆₀ as the core center, which is encapsulated by multiple bulky groups leading to the increas

Full text of DOI:10.1039/b615697e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Large concentration-dependent nonlinear optical responses of starburst
diphenylaminofluorenocarbonyl methano[60]fullerene pentads{
Hendry I. Elim,a Robinson Anandakathir,b Rachel Jakubiak,c Long Y. Chiang,*b Wei Ji*a and
Loon-Seng Tan*c
Received 30th October 2006, Accepted 12th February 2007
First published as an Advance Article on the web 1st March 2007
DOI: 10.1039/b615697e
We demonstrate an approach toward the design of starburst C60-keto-DPAF assembly by
applying a starburst macromolecular configuration with C60 as the core center, which is
encapsulated by multiple bulky groups leading to the increase of intermolecular separation and
aggregation barrier. Molecular compositions of the resulting C60(.DPAF-C9)2 triad and
C60(.DPAF-C9)4 pentads were clearly confirmed by MALDI-MS (positive ion) detection of
protonated molecular mass ions. Both C60(.DPAF-C9)2 2 and C60(.DPAF-C9)4 (structural
isomers, 3a and 3b) exhibited nonlinear optical transmittance reduction responses in the
femtosecond (fs) region with a lower transmittance value for the latter at the high laser power
above 80 GW cm22. This was attributed to the larger fs 2PA cross-section values of 3a and 3b
than that of 2 at the same concentration and, apparently, correlated to a higher number of
DPAF-C9 subunits in the structure of 3. As the concentration was decreased to 1024 M, a
clear monotonous increase of the s2 value change (Ds2) from 13.9, 33.2, to 48.1 and
68.2 6 10248 cm4 s photon21 molecule21 (or 6820 GM for the latter) for the structural variation
from the monoadduct 1, bisadduct 2, to tetraadducts 3b and 3a, respectively, was observed. We
interpreted the concentration-dependent phenomenon as being due to the high tendency of
fullerene-DPAF chromophores to form nanoscale aggregates at concentrations above 1023 M.
We also proposed that starburst structures, as exemplified by C60(.DPAF-C9)4, in a multipolar
arrangement resembling encapsulation of C60 by DPAF-C9 pendants, provide a useful means to
increase the degree of molecular dispersion and maintain high nonlinear optical efficiency.
given the formation of hexaarmed hydrophilic fullerene cages,
Introduction
such as those of hexa(n-sulfobutyl)[60]fullerenes (FC4S) mole-
Chemical functionalization of C60 with organic groups to
cules, were proven to allow the occurrence of fullerenyl photo-
produce the corresponding monoadducts and bisadducts by
activation processes in the long wavelength region. Examples
converting one or two fullerenyl double bonds into two or four
of these photoprocesses were demonstrated recently in singlet
sp3 carbons, respectively, in general does not alter much the
overall p-conjugation and photophysical properties of the
fullerene cage.1–3 In fact, functional modification of C60
disrupts symmetry-forbidden transitions of the parent cage
and allows more ground state optical absorption of the
derivatives in the visible region. Certain structural alterations
oxygen generation,4 photodynamic therapeutic fibrosarcoma
tumor treatment,5 and photodynamic antibacterial coatings6
with the use of either a red-light laser or a visible light source.
During the application, a significant quantity of singlet oxygen
produced by FC4S implied the existence of a photoinduced
triplet excited transient state (3C60*) upon irradiation. This
observation supports our hypothesis that the attachment of a
limited number (¡6) of addends onto C60 forming a starburst
molecular structure may not adversely affect its ability to form
an excited triplet fullerene state, which is considered to be one
of the crucial prerequisites for the reverse saturable absorption
(RSA) and optical limiting properties of C60-derived mate-
rials.7–9 The phenomenon of RSA is related to the well-
recognized fact that excited states of C60 are more polarizable
with larger absorption cross-sections than those of the ground
states. The near co-existence of fullerenyl singlet and triplet
excited states fits well with the required facile generation of
populated excited states for obtaining strong nonlinear optical
responses10 and the enhanced excited state absorption.7,11
These RSA properties have been applied as the fundamental
principle in the development of effective materials for
protection from high-intensity laser pulses.12–21
aDepartment of Physics, National University of Singapore, 2 Science
Drive 3, Singapore 117542, Singapore. E-mail: phyjiwei@nus.edu.sg;
Fax: 65-67776126; Tel: 65-68742664
bDepartment of Chemistry, University of Massachusetts Lowell, Lowell,
MA 01854, USA. E-mail: Long_Chiang@uml.edu;
Fax: +1-(978)-934-3013; Tel: +1-(978)-934-3663
cAir Force Research Laboratory, AFRL/MLBP, Wright-Patterson Air
Force Base, Dayton, OH 45433, USA.
E-mail: Loon-Seng.Tan@wpafb.af.mil; Fax: +1-(937)-255 -9157;
Tel: +1-(937)-255-9141
{ Electronic supplementary information (ESI) available: MALDI-
TOF mass spectrum of C60(.DPAF-C9) 1; comparison profiles of
MALDI-TOF mass spectra among C60(.DPAF-C9), C60(.DPAF-
C9)2, and C60(.DPAF-C9)4; MALDI-TOF mass spectrum of
C60(.DPAF-C9)4 3b; comparison of 13C NMR spectra of
C60(.DPAF-C9) and C60(.DPAF-C9)2; procedure of proton integra-
tion ratio measurement and the calculation of proton counts in 1H
NMR spectra of C60(.DPAF-C9), C60(.DPAF-C9)2, and
C60(.DPAF-C9)4. See DOI: 10.1039/b615697e
1826 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Nonlinear absorption behavior of C60 appeared to be
dominated by the reverse saturable absorption in the visible
region and the two-photon absorption (2PA) process in the
NIR–IR region.22–24 Recently, large enhancement of mole-
cular 2PA cross-sections in the NIR region has been achieved
by periconjugation of C60 with conjugated chromophores or
grafted polymers.25–29 One typical example of such conjugated
chromophores was given by 9,9-diethyldiphenylaminofluorene
(DPAF-C2) which produces fluorescence emission centered at
498 nm in CHCl3 upon excitation (y400 nm) near the ground
state optical absorption peak maximum.30 The emission is
efficiently quenchable by C60 intermolecularly in a solution
mixture of methano[60]fullerene and DPAF-Cn or intramole-
cularly in the case of the conjugated molecule C60(.DPAF-C2)
via energy transfer in nonpolar solvents, such as toluene,
benzene, and CS2. This energy transfer is plausible since the
lowest excited singlet energies of C60(.1DPAF*-C2) and
1C60*(.DPAF-C2) were estimated to be 2.74 (452 nm) and
1.74 (714 nm) eV, respectively, in toluene based on steady state
fluorescence measurements.31 The formation of the fullerene
triplet state was evident from the detection of the triplet–triplet
starburst hindered diphenylaminofluorenocarbonyl methano-
[60]fullerene pentads C60(.DPAF-C9)4 3 and the observed
significant decrease of relative 2PA cross-section values,
especially at high solution concentration (1022 M). In this
concentration region, a high tendency for intermolecular
clustering and nano-aggregative interactions of this class of
materials is believed to exist.
Materials and methods
Reagents and chemicals
Pure C60 (99.5%) was produced from NeoTech Product
Company, Russia and confirmed by thin-layer chromatography
(TLC, SiO2, toluene). Reagents of tris(dibenzylideneacetone)
dipalladium(0), rac-2,29-bis(diphenylphosphino)-1,19-binaphthyl
(BINAP), 2-bromofluorene, and all other chemicals were
purchased from Sigma-Aldrich Chemicals.
Physical measurements for material characterization
Infrared spectra were recorded from KBr pellets on a Thermo
Nicolet 370 series FT-IR spectrometer. UV-Vis spectra were
recorded on a Perkin Elmer Lambda-9 UV/VIS/NIR spectro-
meter. 1H NMR and 13C NMR spectra in solution were
taken on a Bruker Avance Spectrospin-500 spectrometer.
Fluorescence spectra were collected on a FLUOROLOG (ISA
Instruments) spectrofluorometer. Mass spectroscopic measure-
ments were performed by the use of the positive ion matrix-
assisted laser desorption ionization (MALDI-TOF) technique
on a micromass M@LDI-LR mass spectrometer. In a typical
measurement, two fractions (100 ml) each from the solution of
C60(.DPAF-C9)x (1.0 mg) in chloroform (1.0 ml) and the
matrix solution of a-cyano-4-hydroxycinnamic acid (10 mg) in
acetone (1.0 ml) were mixed together. From this sample–
matrix mixture, a small quantity (2.0 ml) was taken and
deposited on the stainless steel sample target plate, dried, and
subsequently inserted into the ionization source of the
instrument. The resulting sample blended or dissolved in the
matrix material was irradiated by a nitrogen UV laser at
337 nm with 10 Hz pulses under high vacuum. MALDI mass
spectra were acquired in reflection mode. Each spectrum was
produced by averaging 10 laser shots; at least 25 spectra were
acquired from different regions of the sample target. Mass ion
peaks were identified for the spectrum using the MassLynx
v4.0 software.
3
absorption band of C60*(.DPAF-C2) at about 730 nm in
nanosecond transient absorption measurements after 532 nm
pulse laser irradiation, in close resemblance to that of the
3
parent C60 occurring at 750 nm for C60* upon similar
excitation.32 Therefore, photoexcitation of C60(.DPAF-Cn)
by 780 nm irradiation with respect to nearly half of
DPAF-Cn’s HOMO–LUMO energy gap should facilitate both
two-photon absorption of the DPAF-Cn moiety and excited
state absorption of the fullerene moiety in a similar wavelength
3
range. Combination of two-photon absorption and C60*
absorption in a nearly concurrent event may largely increase
the overall nonlinear optical absorption capability of the
materials on the ns timescale, in line with the proposed concept
of 2PA–RSA combination at the excited states of complex
chromophores toward the reduction of optical transmittance
at high irradiance.33–36 Accordingly, transient absorption data
obtained from femtosecond pump–probe experiments at
800 nm on samples of sterically hindered [60]fullerenyl dyad
C60(.DPAF-C9) 1 and triad C60(.DPAF-C9)2 2, where C9 is
3,5,5-trimethylhexyl, unambiguously verified the occurrence of
two-photon excitation processes in air-saturated benzene and
subsequent efficient energy transfer from the two-photon
pumped DPAF-C9 moiety to the C60 cage moiety.28 A similar
study on C60(.DPAF-C9)2 in CS2 (1.0
6
1022 M)
also indicated good intrinsic fs 2PA cross-sections around
0.824 610248 cm4 s photon21 molecule21 (or 82.4 GM).
Synthetically, C60-DPAF conjugates can be modified to
further increase the number of chromophore arms on a single
C60 cage resulting in a starburst structure for enhancing 2PA
cross-sections. In this structural motif, the C60 cage serves as a
molecular core for attaching multiple 2PA-active chromophore
antenna components to harvest light in the visible–NIR wave-
length ranges and, subsequently, condense the excited-state
energy to the central fullerene core. The approach resembles
photoresponsive dendritic chromophore structures synthesized
recently that embodied the fluorescence resonance energy
transfer (FRET) phenomenon.37 Here, we report the first
concentration-dependent 2PA cross-section measurements on
Preparation of 7-a-bromoacetyl-9,9-di(3,5,5-trimethylhexyl)-2-
diphenylaminofluorene (7)
A modified aluminium chloride quantity compared to the
literature procedure was applied.30,38 A solution of 9,9-
di(3,5,5-trimethylhexyl)-2-diphenylaminofluorene
6 (1.4 g,
2.4 mmol) in 1,2-dichloroethane (30 ml). was added to a
suspension of aluminium chloride (1.12 g, 8.4 mmol, 3.5 equiv.)
in 1,2-dichloroethane (50 ml) at 0 uC. The mixture was then
added to a-bromoacetyl bromide (0.25 ml, 6.5 mmol) dropwise
over a period of 5.0 min while maintaining the reaction
temperature between 0 and 10 uC. The mixture was slowly
warmed to ambient temperature and stirred for an additional
8.0 h. At the end of the reaction, it was quenched by slow
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1827

View Article Online
addition of ice cold water (100 ml) while maintaining the
temperature below 45 uC. The organic layer was separated
and washed sequentially with dilute HCl (1.0 N, 50 ml) and
water (50 ml, twice). The liquid was dried over magnesium
sulfate and concentrated in vacuo to obtain the crude
product as a crystalline yellow solid. Further purification
was carried out by column chromatography (silica gel)
using a mixture of hexane–EtOAc (9 : 1) as the eluent.
A chromatographic fraction corresponding to Rf = 0.6
on TLC (SiO2, hexane–EtOAc 9 : 1 as the eluent) was
isolated to afford 7-a-bromoacetyl-9,9-di(3,5,5-trimethyl-
hexyl)-2-diphenylaminofluorene 7 as a yellow viscous oil in
68% yield (1.2 g). The spectroscopic characteristics of 7 were
identical to the reported data.29
mixture of hexane–toluene (2 : 3) as the eluent, gave the
bisadduct C60(methanocarbonyl-9,9-di(3,5,5-trimethylhexyl)-
2-diphenylaminofluorene)2, C60(.DPAF-C9)2 2, as a brownish
solid in 9% yield (130 mg). Spectroscopic data of 2: MALDI-
12
MS (TOF) calcd for
C
150
1H11014N216O2 m/z 1970.8; found,
m/z 1974, 1973, 1972, 1971 (M+), 1347, 1346, 870, 861, 710,
705, 672, 666, 630, 629, 612, and 583; FT-IR (KBr) nmax 3429
(br), 3163, 2951 (s), 2922, 2863, 1680, 1595 (vs), 1492 (vs),
1465, 1420, 1400, 1316(w), 1277, 1200, 1156 (w), 1093, 1030
(w), 816 (w), 753, 670 (s), and 527 cm21; 1H NMR (500 MHz,
CDCl3, ppm) d 8.5–8.1 (m, 4H), 7.9–7.4 (m, 4H), 7.24 (m,
12H), 7.12 (m, 8H), 7.05 (m, 4H), 5.8–5.3 (m, 2H), 2.1–1.8 (m,
8H), 1.35–1.15 (m, 8H), and 1.1–0.5 (m, 60H).
Synthesis of the [60]fullerenyl tetraadducts
Synthesis of the [60]fullerenyl monoadduct 7-(1,2-dihydro-1,2-
methanofullerene[60]-61-carbonyl)-9,9-di(3,5,5-trimethylhexyl)-
2-diphenylaminofluorene, C60(.DPAF-C9) 1, and the
C60[methanocarbonyl-7-(9,9-di(3,5,5-trimethylhexyl)-2-
diphenylaminofluorene)]4, C60(.DPAF-C9)4 3a and 3b
Finely divided C60 (550 mg, 0.76 mmol), 7-a-bromoacetyl-
(9,9-di(3,5,5-trimethylhexyl)-2-diphenylaminofluorene (3.24 g,
4.6 mmol, 6.0 equiv.), and anhydrous toluene (500 ml) were
placed in a reaction flask under nitrogen atmosphere. The
mixture was stirred at ambient temperature until a clear
solution was obtained. To this solution was added 1,8-
diazabicyclo[5.4.0]undec-7ene (DBU, 850 mg, 5.5 mmol,
7.2 equiv.). The reaction mixture was stirred at room
temperature for a period of 8.0 h to result in a fine-solid
suspended solution. Solid particles in the solution were filtered
off and washed with toluene. The combined filtrate was
concentrated to 50 ml. To this concentrated liquid was added
methanol (100 ml) to effect precipitation of the crude products,
which were isolated by centrifugation. The crude solids were
washed with methanol (3 6 20 ml) and subsequently purified
by column chromatography (silica gel) using a solvent mixture
of hexane–toluene (3 : 1) as the eluent. Prior to column
chromatography, the crude sample was evaluated on analytical
TLC plates to observe at least 4 product spots each in a
different relative quantity. However, only two major fractions
were collected by column chromatography separation and
subsequently repurified separately on preparative TLC plates
[SiO2, 2000 mm layer thickness, using a solvent mixture of
hexane–toluene (4 : 1) for the less polar fraction and hexane–
toluene (3 : 1) for the more polar fraction as the eluent].
During the TLC separation, only a center narrow dense
band was isolated to afford the [60]fullerenyl multiadducts.
These two product fractions were identified to be the
tetraadducts: C60[methanocarbonyl-7-(9,9-di(3,5,5-trimethyl-
hexyl)-2-diphenylaminofluorene)]4 3a, corresponding to the
chromatographic band at Rf = 0.5 on the analytical TLC plate
using hexane–toluene (2 : 3) as the eluent, as a brownish
solid in 18% yield (440 mg); and 3b, corresponding to the
chromatographic band at Rf = 0.25 on the analytical TLC
plate, as a brownish solid in 7% yield (170 mg). All other
unidentified minor products from column chromatography
were accounted for a total 470 mg.
corresponding bisadduct C60[methanocarbonyl-7-(9,9-di(3,5,5-
trimethylhexyl)-2-diphenylaminofluorene)]2, C60(.DPAF-C9)2 2
A modified reagent quantity compared to the literature
procedure was applied.30 C60 (700 mg, 0.97 mmol) dissolved
in toluene (600 ml) was added to 7-a-bromoacetyl-9,9-
di(3,5,5-trimethylhexyl)-2-diphenylaminofluorene 7 (686 mg,
0.97 mmol, 1.0 equiv.) under an atmospheric pressure of
nitrogen. To this solution was added 1,8-diazabicyclo[5.4.0]
undec-7-ene (DBU, 0.18 ml, 1.16 mmol) and the resulting
mixture was stirred at room temperature for a period of 5.0 h.
At the end of the stirring period, suspended solids of the
reaction mixture were filtered off, washed with toluene, and
the filtrate was concentrated to a 10% volume. Methanol
(100 ml) was then added to cause precipitation of the crude
product, which was isolated by centrifugation. Analytical
thin-layer chromatography (TLC, SiO2) of the solid sample
indicated a mixture of several products at different Rf values.
Only two major narrow bands were isolated by column
chromatographic separation on silica gel using a solvent
mixture of hexane–toluene (1 :2) as the eluent. The first
chromatographic band, corresponding to a spot at Rf = 0.85
on the analytical TLC plate using a mixture of hexane–toluene
(2 : 3) as the eluent, afforded the monoadduct 7-(1,2-dihydro-
1,2-methanofullerene[60]-61-carbonyl)-9,9-di(3,5,5-trimethyl-
hexyl)-2-diphenylamino-fluorene, C60(.DPAF-C9) 1, as a
brown solid (680 mg, 68% based on recovered C60 of
135 mg). Spectroscopic data of 1: MALDI-MS (TOF) calcd
12
for
C
1H5514N16O m/z 1345.4; found, m/z 1349, 1348, 1347,
105
1346 (MH+), 894, 877, 862, 855, 845, 840, 802, 704, 687, 683,
672, 665, 650, 631, 629, 612, and 585; FT-IR (KBr) nmax
3435 (br), 2947 (s), 2922, 2861 (m), 1678, 1593 (vs), 1491 (s),
1464, 1426, 1402, 1347 (w), 1274 (s), 1199 (s), 1185, 1155 (w),
1028 (w), 816 (w), 752 (s), 695 (s), 576, and 526 (vs) cm21
;
1H NMR (500 MHz, CDCl3, ppm) d 8.47 (d, J = 8.0 Hz,
1H), 8.33 (s, 1H), 7.82 (d, J = 8.0 Hz, 1H), 7.65 (d, J = 8.0 Hz,
1H), 7.26 (m, 6H), 7.13 (m, 4H), 7.05 (m, 2H), 5.67 (t, J =
4.0 Hz, 1H), 2.1–1.9 (m, 4H), 1.3–1.1 (m, 4H), and 0.9–0.6 (m,
30H).
Spectroscopic data of the nonpolar tetraadduct 3a:
12
MALDI-MS (TOF) calcd for
C
240
1H22014N416O4 m/z
The second chromatographic band, corresponding to a spot
at Rf = 0.73 on the analytical TLC plate using a solvent
3221.6; found, m/z 3228, 3227, 3226, 3225, 3224, 3223, 3222
(M+), 2639, 2619, 2560, 2599, 2597, 2540, 2127, 2104, 2102,
1828 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
1972, 1563, 1547, 1543, 1479, 1468, 1456, 1347, 878, 870, 857,
827, 629, 612, and 531; FT-IR (KBr) nmax 3421 (br), 2952 (vs),
2924 (vs), 2866, 1681, 1595 (vs), 1493 (vs), 1465, 1421, 1377
(w), 1347 (w), 1277, 1200, 1156 (w), 1029, 819, 752, 697 (s),
Results and discussion
The compounds 9,9-dialkyldiphenylaminofluorenes (DPAF-
Cn) are electron donors (D) and highly fluorescent materials
showing emission at y450 nm upon photoexcitation at
,400 nm. Conjugation with the C60 acceptor (A) cage
facilitates intramolecular electron- and energy-transfer events
from the DPAF moiety to the fullerene moiety in C60-DPAF
assemblies at an ultrafast kinetic rate of .10 ps after 150 fs
laser irradiation.31 In a nonpolar solvent, such as toluene, the
1
525, and 475 cm21; H NMR (500 MHz, CDCl3, ppm) d 8.5–
8.0 (m, 8H), 7.9–7.5 (m, 8H), 7.24 (m, 20H), 7.1–6.9 (m, 28H),
5.6–5.0 (m, 4H), 2.4–1.7 (m, 16H), 1.4–1.1 (m, 16H), and 1.1–
0.5 (m, 120H).
Spectroscopic data of the polar tetraadduct 3b: MALDI-MS
12
(TOF) calcd for
C
1H22014N416O4 m/z 3221.6; found, m/z
240
3
3228, 3227, 3226, 3225, 3224, 3223, 3222 (M+), 2785, 2770,
2757, 2755, 2754, 2753, 2731, 2730, 2680, 2658, 2622, 2599,
2145, 2130, 2128, 2106, 2105, 2104, 1974, 1972, 1958, 1067,
1003, 948, 947, 946, 827, 629, 612, and 531; FT-IR (KBr)
nmax 3434 (br), 2951 (s), 2924 (vs), 2864, 1683, 1594 (vs),
1492 (vs), 1466, 1420 (w), 1363 (w), 1316 (w), 1278, 1211,
triplet state of the fullerene cage in C60*(.DPAF-C2) has
a lifetime of 33 ms.31 As we extended this linear D–A
arrangement to include multiple DPAF-Cn attachments on
one C60 cage, corresponding starburst A-(D)n or C60-(DPAF-
Cn)x structures were formulated. In this structural motif, the
C60 cage becomes a molecular core and multiple DPAF-Cn
subunits are photoactive antenna components to harvest light
from the visible (for single photon excitation) to the near
infrared (for two-photon excitation) wavelength ranges.
Subsequent transfer of the excited-state energy to the central
fullerene core may enhance its reverse saturable absorption
capability. In a very dilute solution giving a nearly molecularly
separated state of each C60-(DPAF-Cn)x, interference of
intermolecular energy transfer and excited state quenching
may be minimized. As their concentration increases in the
materials application, molecular aggregation and coalescence
becomes a common phenomenon that may significantly
affect the photoresponsive efficiency of the compound. One
approach to extend and maintain the molecular photo-
efficiency in highly concentrated solution or the solid state is
to apply highly bulky groups in the structure as a physical
barrier to favor molecular separation. One recent example was
given by hindered C60(.DPAF-C9) and C60(.DPAF-C9)2,
where C9 is a 3,5,5-trimethylhexyl group, showing enhanced
2PA cross-section values and nonlinear optical responses.28
Reduction to some extent of the molecular aggregation in a
dense solution without much decrease of the chromophore
density is believed to be responsible for this enhancement.
Cluster formation of fullerenes in solution can be a frequent
occurrence during sample preparation owing to a strong
tendency for hydrophobic attraction among fullerene cages via
weak intermolecular van der Waals interactions in nonpolar
solvents, such as benzene, toluene, and CS2.40,41 Evidently, the
phenomenon exists even in fairly dilute toluene solution in a
concentration of 2 6 1024 M showing fractal clusters (C60)n
(n ¢ 3) with sizes .1.2 nm as well as single C60 molecules.42
The quantum efficiency of the triplet population can be
considerably depleted in the form of particles and clusters, via
triplet–triplet annihilation.
1177 (w), 1089, 819, 752, 697 (s), 525, and 474 (br) cm21
;
1H NMR (500 MHz, CDCl3, ppm) d 8.6–8.1 (m, 8H), 7.9–7.5
(m, 8H), 7.24 (m, 20H), 7.11(m, 20H), 7.03 (m, 8H), 5.6–4.9
(m, 4H), 2.2–1.8 (m, 16H), 1.4–1.1 (m, 16H), and 1.1–0.4 (m,
120H).
Optical absorption measurements
Optical absorption and transmission spectra, in the range of
180–1200 nm, of C60(.DPAF-C9)x (x = 1, 2, or 4) samples
dissolved in CS2 were determined in a 1 mm thick quartz
cuvette with a UV-Vis-NIR spectrophotometer (Model UV
3600, Shimadzu, Kyoto, Japan).
Two-photon absorption cross-section (s2) and nonlinear optical
transmittance measurements
Two-photon absorption cross-section (s2) and nonlinear
optical transmittance measurements of C60(.DPAF-C9)x
(where x = 1, 2, or 4) samples in a CS2 solution were carried
out with femtosecond Z-scans and energy-dependent trans-
mission counting at the wavelength of 780 nm. To reduce
accumulative thermal effects and to minimize the contribution
of triplet–triplet state absorption in C60(.DPAF-C9)x, we
employed 150 fs laser pulses at 1 kHz repetition rate. Laser
pulses were generated by a mode-locked Ti : sapphire laser
(Quantronix, IMRA), which seeded
a Ti : sapphire
regenerative amplifier (Quantronix, Titan), and focused onto
a 1 mm thick quartz cuvette containing a solution of
C60(.DPAF-C9)x with a minimum beam waist of y12 mm.
By adding CS2 to C60(.DPAF-C9)x, the concentration of
the samples was adjusted to either 1.0 6 1024, 1.86 6 1023
,
or 1.0 6 1022 M for the measurement. Incident and
transmitted laser powers were monitored as the cuvette was
moved (or Z-scanned) along the propagation direction of the
laser pulses.
In the presence of an extended aromatic DPAF ring in the
C60(.DPAF-C2) structure, strong fullerenyl p-interactions
were found to drive the molecular assembly over DPAF-C2
subunits into a highly ordered array of the fullerene cages in its
single crystal structure.30 The structural arrangement also
forces planar DPAF-C2 rings to orient side-by-side with no
direct p–p interaction between DPAF moieties, indicating
stronger interactions among C60 cages than among DPAFs
or between C60 and DPAF. Therefore, it may require at least
four DPAF-Cn subunits, such as in C60(.DPAF-C9)4, to
Two-photon absorption is described by the change in the
absorption coefficient Da = bI, where b and I are the 2PA
coefficient and the light intensity, respectively. The 2PA
coefficient can be extracted from the best fitting between the
Z-scan theory39 and the data. The 2PA cross-section value
was then calculated from the 2PA coefficient by the formula
s2 = bhv/N, where hv is the photon energy and N is the
number of molecules.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1829

View Article Online
encapsulate the C60 cage in addition to the attachment of
hindered 3,5,5-trimethylhexyl groups on the C9 position of the
fluorene ring to prevent direct intermolecular stacking contact
of both fullerene cages and DPAF-Cn rings. In addition, when
C60 cages are wrapped by four hindered dialkylated chromo-
phores, enhanced solubility of the resulting macromolecule
should lead to a larger amount of C60 cages being drawn into
the solvent for a highly concentrated solution.
another roughly 20% yield of minor, unidentified tetraads,
pentaads, or hexaads.
The chemical structures of 1 and its related monoadduct
C60(.DPAF-C2) were well characterized by spectroscopic
methods with the latter also by X-ray single crystal structural
analyses.30 Their spectroscopic data greatly aided the struc-
tural correlation, characterization, and identification of 2, 3a,
and 3b in this report. Monofunctionalization of C60 occurs
only on one-half of the sphere leaving the other half-sphere
intact. Accordingly, the infrared spectrum of 1 (Fig. 1b)
showed two sharp characteristic bands at 576 and 526 cm21 in
nearly identical peak positions to those of pristine C60 (Fig. 1a)
at 575 and 526 cm21, respectively. Therefore, these two sharp
bands are characteristically useful as structural indicators for
the bisadducts and tetraadducts if all functional groups are
attached on C60 in the same half-sphere. In comparison with
the molecular diameter of C60, the size of the 7-keto-9,9-
di(3,5,5-trimethylhexyl)-2-diphenylaminofluorene (DPAF-C9,
Scheme 1) moiety is relative bulky and can be considered as
a hindered arm. It is plausible to assume that the attachment of
two DPAF-C9 arms on one C60 should result in regioisomers
each with two DPAF-C9 subunits located at a separation
distance sufficient to accommodate the gyration radius of these
chromophore arms. A large reduction in intensity of two sharp
bands at 576 and 526 cm21 in the infrared spectrum (Fig. 1c)
of the triad C60(.DPAF-C9)2 revealed only a minor quantity
of a regioisomer with two DPAF-C9 subunits located on the
same half-sphere in the isolated major chromatographic band.
These two sharp bands disappeared fully in the infrared
spectra (Fig. 1d and e) of the pentad isomers C60(.DPAF-C9)4
3a and 3b, indicating a distribution of four bulky DPAF-C9
subunits around the full sphere of C60. Undoubtedly, the IR
spectra of all C60-DPAF conjugates 1, 2, 3a, and 3b displayed
nearly identical profiles of the bands corresponding to the
optical absorption of the DPAF-C9 moiety. It included the
characteristic band centered at 1683 cm21, corresponding to
absorption of a keto group bridging a methano[60]fullerene
(C60.) cage and a fluorene moiety, shifted by roughly 50 cm21
Synthesis and structural characterization of C60-DPAF
conjugates
Preparation of the dyad C60(.DPAF-C9) 1, 7-(1,2-dihydro-1,2-
methanofullerene[60]-61-carbonyl)-9,9-di(3,5,5-trimethylhexyl)-
2-diphenylaminofluorene, and the triad C60(.DPAF-C9)2 2,
C60[methanocarbonyl-7-(9,9-di(3,5,5-trimethylhexyl)-2-diphenyl-
aminofluorene)]2, followed a modified synthetic procedure
reported recently.30 The modification included the use of a
larger quantity of aluminium chloride (3.5 equiv.) in a
Friedel–Crafts acylation reaction with a-bromoacetyl bro-
mide than the reported amount to obtain a good product
yield of 7-a-bromoacetyl-9,9-di(3,5,5-trimethylhexyl)-2-diphe-
nylaminofluorene 7. Isolation and purification of 2 were
achieved by the collection of only a narrow, major chromato-
graphic band on preparative TLC plates corresponding to a
spot at Rf = 0.73 on the analytical TLC plate using a
solvent mixture of hexane–toluene (2 : 3) as the eluent.
Synthesis of starburst C60-DPAF conjugates was carried out
by using multiple equivalents of a-bromo-DPAF-C9 7 per
C60 under similar reaction conditions. However, a greater
quantity of 7 than 2.0 equiv. used in the reaction resulted in
a distribution of bisadducts, trisadducts, and tetraadducts
along with other minor higher multiadducts with little
control. Therefore, we applied a maximum quantity of 7
of 6.0 equiv. for this reaction by considering the bulky size
of DPAF-C9 with a sufficient reaction time to complete the
highest possible number of chromophore attachments on
each C60. The approach minimizes the products containing
lower numbers of DPAF-C9 arms and maximizes the pro-
duct yield of high multiadducts that simplifies the product
separation. Evidently, at least 4 chromatographic spots were
clearly visible on analytical TLC plates, each in a different
relative quantity. Among them, only two major fractions
were collected by column chromatographic separation and
subsequently repurified individually as a narrow band on
preparative TLC plates (SiO2) to afford the major multi-
adduct products identified later to be [60]fullerene pentad
isomers. No significant quantities of [60]fullerene hexaads or
higher adducts were isolable. The first fraction of these two
products, corresponding to the chromatographic band at
Rf = 0.5 on the analytical TLC plate using hexane–toluene
(2 : 3) as the eluent, is less polar. It was found to be the
tetraadduct C60(.DPAF-C9)4, C60[methanocarbonyl-7-(9,9-
di(3,5,5-trimethylhexyl)-2-diphenylaminofluorene)]4 3a, in the
form of a brownish solid in 18% yield. The next major polar
fraction, corresponding to the chromatographic band at Rf =
0.25 on the analytical TLC plate, was the second tetraadduct
3b isolated as a brownish solid in 7% yield. All other
product fractions in a combined quantity accounted for
Fig. 1 Infrared spectra of (a) pure C60, (b) C60(.DPAF-C9) 1, (c)
C
60(.DPAF-C9)2 2, (d) C60(.DPAF-C9)4 3a, and (e) C60(.DPAF-
C9)4 3b, all showing nearly identical characteristic carbonyl
absorption band at 1683 cm21 largely shifted from 1720 cm21
a
.
1830 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
(m/z 612) and CH(DPAF-C9) (m/z 625) were also observed
revealing facile fragmentation at the bonds immediately next
to the C60 and methano[60]fullerene cage. The relatively simple
spectrum in the high mass region revealed the good stability
of the aromatic keto-diphenylaminofluorene moiety under
MALDI-MS measurement conditions.
In the MALDI-MS investigation of the tetraadducts 3a and
3b, a significantly higher laser power than that used in the
measurement of 2 was necessary to overcome the low intensity
of the mass ion above m/z 3000. However, application of a
high laser power created additional mass ion peaks as spectral
contaminants with one or two carbon mass units (m/z 15, 24,
or higher, etc.) higher than the molecular ion or fragmented
ion masses. In the high mass region, a group of sharp mass ion
peaks corresponding to the molecular mass of C60(.DPAF-
C9)4 3a at m/z 3222 (M+) and its isotope peaks at m/z 3222–
3228 were detected, as shown in Fig. 3. Two additional groups
of higher mass ions at m/z 3247 (M + 24)+ and 3263 were,
presumably, the result of the high laser power conditions.
The next major group of mass fragmentation peaks occurring
at m/z 2599 (Fig. 3a) matched well with the mass of protonated
C60(.DPAF-C9)3 by the loss of one CH(DPAF-C9) group
(m/z 625) from 3a. Further fragmentation of the C60(.DPAF-
C9)3 mass peak to give the corresponding C60(.DPAF-C9)2
and C60(.DPAF-C9) mass ions at m/z 1972 and 1347,
respectively, was also observed at a relatively low intensity.
In the low mass ion region, two sharp peaks at m/z 612 and 629
corresponding to the mass of DPAF-C9 and CHx(DPAF-C9),
respectively, were detected at a high peak intensity, implying
fast loss of two CH(DPAF-C9) groups after the C60(.DPAF-
C9)3 fragment. These mass spectral data evidently sub-
stantiated the mass composition of 3a as C60(.DPAF-C9)4.
The mass ion peak at m/z 2104 was the result of decarboxyla-
tion fragmentation of a C60(.DPAF-C9)2-(a-cyano-4-hydro-
xycinnamic acid) adduct generated under the high voltage
MALDI-MS conditions.
1
A comparison of the H NMR spectra of 1, 2, 3a, and 3b is
shown in Fig. 4, where the chemical shift values of all the
protons of C60(.DPAF-C9) were used as the reference for the
proton assignment of the bisadducts and tetraadducts.
Interestingly, the a-proton (Ha, next to the carbonyl group)
peak of 1 appeared as a triplet-like signal, perhaps due to long-
range 1H–1H coupling, at d 5.67 (Fig. 4a) with a large
downfield shift of 1.15 ppm from that of 7 (d 4.52, singlet). It
was accompanied by three groups of phenyl proton peaks of
H5, H6, and H8 of the DPAF moiety showing corresponding
chemical shifts at 7.82 (d, J = 8 Hz), 8.46 (d, J = 8 Hz) and
8.33 (s), respectively. Benzyl proton (H9) peaks at d 1.9–2.1 are
apparently indicative of non-equal H9s showing two groups of
multiplets, instead of one triplet. Peak splitting of benzyl
protons is caused by steric hindrance of the adjacent bulky
diphenylamino group. In the case of C60(.DPAF-C9)2, a
comparable 1H NMR spectrum (Fig. 4b) to that of Fig. 4a
was obtained. Accordingly, we assigned multiplet peaks at d
8.1–8.5 (4H), 7.63–7.8 (2H), 5.3–5.8 (2H), and 1.8–2.1 (8H) to
the chemical shifts of H6–H8/H69–H89, H5/H59, Ha/Hb, and
H9protons, respectively, of 2. Considering two possible
sterically non-equivalent CH(DPAF-C9) subunits and the
peak splitting pattern of H6, H5, Ha, and H9 as the reference,
Scheme 1 Reagents and conditions: i. (CH3)3CCH2CH(CH3)CH2-
CH2-OMs, t-BuOK, THF, 10 uC to rt, 15 h; ii. diphenylamine,
tris(dibenzylideneacetone)dipalladium(0) (cat.), rac-BINAP (cat.),
t-BuONa, toluene, reflux, 8.0 h; iii. bromoacetyl bromide, AlCl3,
ClCH2CH2Cl, 0 uC to rt, 8.0 h; iv. C60, DBU, toluene, rt, 5.0 h.
from the normal carbonyl stretching band at 1720–1750 cm21
due to the influence of the fluorene ring structure.
The molecular weight of the triad 2 was clearly confirmed by
positive ion matrix-assisted laser desorption ionization mass
spectrum (MALDI-MS), using a-cyano-4-hydroxycinnamic
acid as the matrix, displaying a group of mass peaks with a
maximum peak intensity centered at m/z 1972, as shown in
Fig. 2. Analysis of this group in Fig. 2b indicated the
molecular ion mass (M+) of 2 at m/z 1971 with following
protonated MH+ and isotope peaks at m/z 1972–1974. It was
followed by a group of mass peaks centered at m/z 1346
matching well with that of the monoadduct 1 resulting from
the loss of exactly one CH2(DPAF-C9) (m/z 626) subunit from
the molecular mass. In the low mass ion region, several major
peaks at m/z 612–629 corresponding to the mass of DPAF-C9
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1831

View Article Online
Fig. 2 Positive ion MALDI-TOF mass spectra of the bisadduct C60(.DPAF-C9)2 2.
Fig. 3 Positive ion MALDI-TOF mass spectra of the tetraadduct C60(.DPAF-C9)4 3a.
the profile of the a-proton peak in the region of d 5.3–5.8
was separated into two groups at d 5.3–5.55 and 5.55–5.8
with a proton integration ratio of roughly 2.5 : 1. This revealed
the same molecular ratio of two possible regioisomers of
C60(.DPAF-C9)2 in a similar molecular polarity. By applying
the analyses made on its IR spectrum, the minor regioisomer
might contain two CH(DPAF-C9) arms on the same half-
sphere, whereas the major regioisomer might contain two
CH(DPAF-C9) arms located on opposite half-spheres. As the
spectrum extended to those of C60(.DPAF-C9)4 3a and 3b, as
shown in Fig. 4c and d, a slight upfield shift of the a-proton
multiplet to d 4.9–5.6 with increasing complexity was detected.
In principle, a large number of four DPAF-C9 addends per C60
as in 3a and 3b, giving a close resemblance to molecular
encapsulation of C60, should increase significantly the spatial
crowdedness surrounding the fullerene cage and decrease the
degree of freedom for each DPAF-C9 arm to rotate freely
around the C61–(CLO) bond. This reduces the peak resolution
of a-protons Ha, Hb, Hc, and Hd of 3 and, thus, further
broadens their peak profiles. Therefore, based on the peak
profiles of H6/H69/H60/H6-, H8/H89/H80/H8-, and H5/H59/H50/
H5- of 3a (Fig. 4c) and 3b (Fig. 4d) at d 7.65–8.5, both with
reasonably close similarity in peak resolution and shape to
those of 2 (Fig. 4b) in the same chemical shift region, we
suggest a small possible number of non-isolable regioisomers,
such as that of 2, in the narrow chromatographic fractions of
3a and 3b.
Relative integration ratios of aromatic proton peaks at d
7.0–8.6 in the 1H NMR spectra of C60(.DPAF-C9),
C60(.DPAF-C9)2, and C60(.DPAF-C9)4 can be used as
the value for quantitative counting of the total aromatic
protons that, in turn, correlate directly to the number
of DPAF-C9 units per molecule. The compound 1,4-diazabi-
cyclo[2.2.2]octane (DABCO) is a symmetrical molecule show-
ing only one singlet 1H peak at d 2.78 for 12 protons in a
chemical shift region containing no aliphatic proton peaks of
1832 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 4 1H NMR spectra of (a) C60(.DPAF-C9) 1, (b) C60(.DPAF-C9)2 2, (c) C60(.DPAF-C9)4 3a, and (d) C60(.DPAF-C9)4 3b in CDCl3 with
relative integration values, which were calibrated by the use of an internal DABCO standard at the same concentration for (a)–(d), indicated in
parentheses.
the DPAF-C9 moiety. Therefore, this proton peak taken in a
known applied concentration can be used as an internal
standard for proton counting of the peaks shown in the 1H
NMR spectra of Fig. 4a–d. The procedure of relative
integration ratio measurements and the calculation are given
as a part of the ESI.{ Consequently, the results showed total
numbers of aromatic protons of 1 (Fig. 4a), 2 (Fig. 4b), and 3
(Fig. 4c and d) in good agreement with the corresponding
chemical structures containing one, two, and four DPAF
addends, respectively, per C60 cage. The data provided
further confirmation of the structural composition of these
compounds.
fullerenyl chromophores by employing ultrafast 150 fs laser
pulses at 1 kHz repetition rate. Nevertheless, it cannot rule out
completely the contribution of thermally induced scattering
effects to result in somewhat larger measured 2PA cross-
sections. By assuming this thermal effect on the scattering is
similar for all C60(.DPAF-C9)x solutions, the 2PA data
should be qualitatively useful for comparison analyses.
Prior to 2PA measurements, the concentrations of all
C60(.DPAF-C9)x solutions in CS2 was adjusted to 1.86 6
1023 M which gives a linear (or low-fluence) transmittance (T)
of 88.6% at a nearly identical intensity for all samples at
780 nm, as shown in Fig. 5. As the concentration was increased
to 1.0 6 1022 M, the solution transmittance of 1, 2, 3a, and 3b
decreased to slightly different levels of 77, 82, 80, and 82%,
Two-photon absorption cross-section (s2) measurements of
C60-DPAF conjugates at a high laser power
In the present study, two-photon absorption (2PA) properties
of C60(.DPAF-C9)x (where x = 1, 2, and 4) dissolved in CS2
were investigated with femtosecond Z-scans measurements at
780 nm. To correlate the 2PA data with the results of energy-
dependent nonlinear transmission studies, a relatively high
energy power source at 163 GW cm22 (or 24.5 mJ cm22) was
applied. This power intensity range also facilitates the signal
detection on the same sample at a low solution concentration
of 1.0 6 1024 M. However, high laser-pulse irradiation
of fullerene solutions has been reported to cause thermally
induced scattering effects.11,43 Therefore, recently measured
2PA cross-section values of two C60-DPAF conjugates
C60(.DPAF-C9) and C60(.DPAF-C9)2 using a low laser
power of 10–30 GW cm22 in both femtosecond and nano-
second regions28 were used as the reference. To prevent much
deviation of measured 2PA cross-section values from intrinsic
two-photon absorptions, effort was taken to reduce accumu-
lative thermal effects to a possibly low level and to minimize
the contribution of triplet–triplet state absorption in these
Fig. 5 UV-Vis-NIR transmission spectra of C60(.DPAF-C9) 1,
C
60(.DPAF-C9)2 2, C60(.DPAF-C9)4 3a, and C60(.DPAF-C9)4 3b,
with the latter four compounds dissolved in CS2 in the concentrations
indicated.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1833

View Article Online
two-photon absorption cross-section value. To differentiate
the molecular contribution vs. the number-sum component
contribution to the measured 2PA cross-section values (s2)
and to gain an insight of the structure–property relationship
leading to the enhancement of s2, we carried out comparative
Z-scans (Fig. 6) to include measurements based on the
cumulative number-sum of individual C60 and CH3(DPAF-
C9) components in the same composition ratio as their
corresponding conjugated compounds C60(.DPAF-C9)x. For
comparison purposes, the Z-scans were performed in CS2 with
a concentration of 1.0 6 1024 M. As a result, a systematic
increase of s2 from 12.7 6 10248 cm4 s photon21 molecule21
(or 1270 GM) for the combination of one C60 and one
CH3(DPAF-C9) model component to 29.6 and 46.4 6
10248 cm4 s photon21 molecule21 (or 4640 GM for the latter)
for the combination of one C60 with two and four CH3(DPAF-
C9) components, respectively, was observed, as shown in
Table 1. At a relatively low concentration of 1.0 6 1024 M, a
unimolecular process of either C60 or CH3(DPAF-C9) model
component in response to laser excitation may be expected.
Based on this assumption and the simple cumulative sum of
2PA data measured, the average 2PA contribution of one
molecular fraction of CH3(DPAF-C9) component to the s2
Fig. 6 Open-aperture Z-scans of 1 mm thick solutions of (a)/(b)
60(.DPAF-C9)4, (c) C60(.DPAF-C9)2, and (d) C60(.DPAF-C9) in
C
CS2 in comparison with the sum of the individual model components
[C60 and CH3(DPAF-C9)], performed at the same concentration of
1.0 6 1024 M and irradiance of 163 GW cm22. The data are vertically
shifted for clear presentation. Solid lines are the theoretical fit for two-
photon absorption.
value can be estimated to be roughly 11–12 6 10248 cm4
s
photon21 molecule21. This implied that DPAF-C9, being the
major structural moiety in C60(.DPAF-C9)x, is responsible
for photoexcitation in 2PA processes at 780 nm. The
implication is reasonable since we believe the 2PA peak of
structurally modified C60 to be at wavelengths .950 nm.24
Similar open-aperture Z-scans carried out under an
irradiance of 163 GW cm22 were taken for 1, 2, 3a, and 3b
in CS2 in three concentrations from 1.0 6 1024 to 1.0 6
1022 M giving the plots shown in Fig. 7. These Z-scans
displayed positive signs for absorptive nonlinearities. All
calculated values of 2PA coefficient b and cross-sections s2
of these chromophores are listed in Table 1, where nonlinear
absorption in neat carbon disulfide was also measured
and included as the reference which gave the b value of
respectively, at 780 nm. A relatively small deviation of
transmittance intensity for the monoadduct C60(.DPAF-C9)
solution might be attributed to its higher tendency to aggregate
in at high concentration than the rest of the multiadducts.
Molecular aggregation of 1 and 2 was expected to result in
broadening and the shift of their optical absorption bands
from the maxima at 326 (for the fullerene cage) and 410 (for
the DPAF moiety) nm to a slightly longer wavelength along
with a shift of the transmission edge from roughly 570 to
680 nm.
Both C60 and diphenylaminofluorene components are
photoresponsive chromophores capable of undergoing simul-
taneous two-photon excitation, making contributions to the
0.03 cm GW21 in a similar range as the value (0.05 cm GW21
)
Table 1 Two-photon absorption cross-section values (s2 = bhv/N) of C60(.DPAF-C9)x (x = 1, 2, or 4) and the sum of their individual molecular
components, measured at 780 nm with 150 fs laser pulses and I = 163 GW cm22
s2/10248 cm4
s
s
2,adj/10248 cm4
s
Ds2,adj/10248 cm4
s
Chromophore/solvent
[C]/M
b/cm GW21
photon21 molecule21
photon21 molecule21 a
photon21 molecule21
CS2
C60 + 1 [CH3(DPAF-C9)]
C60(.DPAF-C9) 1/CS2
0.03
1.0 6 1024
1.0 6 1024
1.86 6 1023
1.0 6 1022
1.0 6 1024
1.0 6 1024
1.86 6 1023
1.0 6 1022
1.0 6 1024
1.0 6 1024
1.86 6 1023
1.0 6 1022
1.0 6 1024
1.86 6 1023
1.0 6 1022
0.033
0.036
0.080
0.100
0.037
0.041
0.088
0.130
0.041
0.051
0.120
0.280
0.045
0.097
0.160
12.7
25.3
11.4
2.9
29.6
46.4
13.2
4.2
46.4
88.6
20.4
10.5
63.3
15.2
5.5
9.3
21.9
8.0
13.9
7.7
0.3
C60 + 2 [CH3(DPAF-C9)]
C60(.DPAF-C9)2 2/CS2
26.2
43.0
9.8
33.2
9.0
0.8
C60 + 4 [CH3(DPAF-C9)]
C60(.DPAF-C9)4 3a/CS2
43.0
85.2
17.0
7.1
59.9
11.8
2.1
68.2
9.9
C60(.DPAF-C9)4 3b/CS2
48.1
9.7
a
Calibrated s2 values to the reported data collected at 1.0 6 1022 M using a lower laser energy source of 10–30 GW cm22 28
.
1834 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 8 (a) Open-aperture Z-scans at different excitation irradiances
for C60(.DPAF-C9)4 3a with experimental data presented as scatter
graphs and fitting curves, calculated by Z-scan theory,38 as solid lines
and (b) irradiance independence of the s2 profile of 3a in two
concentrations indicated.
insensitive or relatively constant over the irradiance range.
Thus, the observed change (Ds2) in the nonlinear absorption
could be purely due to the different 2PA processes.
In the comparison of s2 values among C60(.DPAF-C9)x
samples collected at 1.0 6 1024 M concentration, a consistent
reduction of the transmittance in Fig. 7a–d is indicative of a
larger 2PA coefficient of each C60(.DPAF-C9)x compound
than that for the sum of compositional model components.
Accordingly, an increase of 2PA cross-sections by 2.0, 1.6, 1.9,
and 1.4 times their corresponding model component s2 values
for 1, 2, 3a, and 3b, respectively, revealed clearly the con-
tribution of extending molecular branching and p-conjugative
polarization of donor–acceptor moieties to the 2PA enhance-
ment over the simple cumulative number-sum effect.
Fig. 7 Open-aperture Z-scans of
1
mm thick solution of
C
60(.DPAF-C9), C60(.DPAF-C9)2, and C60(.DPAF-C9)4 in CS2
performed at different concentrations of (a) 1.0 6 1024, (b) 1.86 6
1023, and (c) 1.0 6 1022 M with an irradiance of 163 GW cm22. Solid
lines are the theoretical fit for two-photon absorption.
reported recently.44 As thermally induced scattering effects
caused by high irradiances may influence the accuracy of
overall 2PA cross-sections measured, we undertook the
measurement of Z-scans of C60(.DPAF-C9)4 3a in CS2
(1.86 6 1023 M) at different irradiances, as shown in Fig. 8a,
with the resulting 2PA cross-section values plotted as a
function of the laser power intensity in Fig. 8b. Interestingly,
the s2 values apparently remain nearly constant in approxi-
mately a straight line across the irradiance range from
40 GW cm22 up to 600 GW cm22 without an obvious
increase, confirming the independence of the s2 values from
the laser power intensity. This can be interpreted as the ther-
mally induced scattering contribution being either fluctuation
Most importantly, we observed a significantly large increase
of 2PA cross-sections of all C60(.DPAF-C9)x compounds in a
consistent trend upon decreasing the solution concentration
from 1.0 6 1022 to 1.0 6 1024 M in CS2 (see Fig. 9).
Since 2PA cross-sections for 1 and 2 at 2.0 and 1.0 6 1022
M
in CS2 were reported with values of 0.306 and 0.824 6
10248 cm4 s photon21 molecule21 (or 30.6 and 82.4 GM),
respectively, using
30 GW cm22 28
a
lower laser energy source of 10–
,
these two values were used for the calibration
by reducing the measured s2 value of 2 at 1.0 6 1022 M to the
same amount as the reported one. Correspondingly, all other
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1835

View Article Online
Fig. 9 Concentration dependence of 2PA cross-section values s2 of
60(.DPAF-C9), C60(.DPAF-C9)2, and C60(.DPAF-C9)4 in CS2.
C
s2 values measured were then deducted by the same quantity
of 3.4 6 10248 cm4 s photon21 molecule21, which accounted
for the partial factor arising from the thermally induced
scattering,11,43 to the new s2,adj values, as shown in Table 1. As
an example for 1, the s2,adj values were adjusted to be 8.0
and 21.9 6 10248 cm4 s photon21 molecule21 (or 800
and 2190 GM) at concentrations of 1.86 6 1023 and 1.0 6
1024 M, respectively, showing a clear increase of 2PA cross-
sections upon decrease of the concentration. The systematic
increase of the s value, Ds2,adj, at incremental decreases of the
concentration was calculated as listed in Table 1. Interestingly,
regardless of the multiadduct structure, the Ds2,adj values in the
high concentration range (1023–1022 M) were relatively
close to each other in a narrow range of 7.7–9.9 6
10248 cm4 s photon21 molecule21, revealing domination of
the aggregation factor over the difference of the molecular
structures. As the degree of molecular separation increases at a
low concentration of 1024 M, a clear leap of the Ds2,adj values
from 13.9, 33.2, to 48.1 and 68.2 6 10248 cm4 s photon21
molecule21 for the modification of starburst structures going
from the monoadduct 1, bisadducts 2, to tetraadducts 3b and
3a, respectively, was found. This observation implied the
significance of branched fullerenyl multiadducts in enhancing
the 2PA cross-sections at the laser power range related to the
nonlinear power transmission application.
Fig. 10 (a)
Nonlinear
transmission
of
C60(.DPAF-C9),
C
60(.DPAF-C9)2, and C60(.DPAF-C9)4 in CS2 and (b) the measure-
ment of transient transmission change (2DT/T) in solutions of
60(.DPAF-C9) and 3a. All the measurements were conducted with
C
150 fs laser pulses at l = 780 nm. All four samples showed nearly
identical linear transmission T = 88.6% for an optical path of 1 mm.
energy-dependent transmission measurements at the wave-
length of 780 nm using the same setup as that in 2PA
measurements conducted by 150 fs laser pulses, except that the
sample was fixed at the focus area when the incident pulse
power was varied. Technical details of the measurements were
described elsewhere.45,46 Nonlinear transmittance results for
1 mm thick solutions of 1, 2, 3a, and 3b in CS2 at a medium
concentration of 1.86 6 1023 M are shown in Fig. 10a. All
the samples showed a linear transmission (T = 88.6%) with
an input intensity of up to 36 GW cm22, or a fluence of
5.4 mJ cm22. When the incident intensity was increased above
36 GW cm22, the transmittance began to deviate from the
linear transmission line and decrease indicating the initiation
of nonlinearity and the limiting effect. The transmitted fluence
further departed from linearity upon increasing the incident
fluence. A systematic trend showing higher efficiency in
reducing optical transmittance with the increase of irradiance
level up to ,1000 GW cm22 was observed for the starburst
tetraadducts C60(.DPAF-C9)4 than for the bisadduct
C60(.DPAF-C9)2 and the monoadduct C60(.DPAF-C9).
This improvement in lowering the transmittance can be
correlated to a higher number of DPAF-C9 subunits in the
structure of 3 than in 1 and 2, consistent with the positive
contribution of a starburst structure such as 3 to a larger
We interpret the concentration-dependent phenomena by
the high tendency of fullerene-DPAF chromophores to form
nano-aggregates, instead of a molecular dispersion, at con-
centrations above 1023 M. Nano-aggregation might play a role
in reducing 2PA quantum efficiency leading to the decrease
of the s2 value. The starburst structures of C60(.DPAF-C9)4
pentad regioisomers 3a and 3b gave the largest s2 enhancement
over the C60(.DPAF-C9) dyad and the C60(.DPAF-C9)2 triad
when molecular dispersion became possible in dilute solution.
Nonlinear optical transmittance measurements of C60-DPAF
conjugates
Nonlinear optical transmittance properties of C60-DPAF
conjugates C60(.DPAF-C9)x in CS2 were investigated by
1836 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007

View Article Online
two-photon absorption, concluded by Z-scans in Fig. 7. This
correlation was made on the assumption of a similar degree of
nonlinear scattering among the solutions.
character. The molecular compositions of the resulting
C60(.DPAF-C9)2 triad and C60(.DPAF-C9)4 pentads were
confirmed by the analyses of various spectroscopic data.
Both C60(.DPAF-C9)2 2 and C60(.DPAF-C9)4 (3a and 3b)
exhibited nonlinear optical transmittance reduction responses
in the femtosecond region with a lower transmittance value for
In the molecular system of C60(.DPAF-Cn)x, simultaneous
absorption of two photons energy at 780 nm leads to fast
release of the fluorescence emission of DPAF-Cn at roughly
450 nm, which can be utilized for subsequent photoexcitation
of the C60 cage via intramolecular energy transfer. The kinetic
rate of following intersystem crossing going from the excited
singlet state of the fullerene cage moiety (1C60*.) to the
corresponding triplet state (3C60*.) took place within a 1.4 ns
time scale.28 Therefore, triplet–triplet state related absorption
the latter in the high laser power region above 80 GW cm22
.
This was attributed to the larger fs 2PA cross-section value of
C60(.DPAF-C9)4 than for 2 at the same concentration and
correlated to a higher number of DPAF-C9 subunits in the
structure of 3.
The observed high 2PA cross-section values were explained
by an increasing degree of intramolecular interaction and
polarization. As the concentration of the solution was varied
systematically to induce molecular separation, a consistently
large increase of 2PA cross-sections of all C60(.DPAF-C9)x
(x = 1, 2, or 4) chromophores in the fs region was observed
upon concentration dilution from 1.0 6 1022 to 1.0 6 1024 M.
The difference of the 2PA cross-section value change (Ds2)
among all samples in response to variation of the concentra-
tion was not sensitive to the compound structure in the high
concentration range. However, at a low concentration of
1024 M, Ds2 values were evaluated to increase clearly for the
structural modification from the monoadduct 1, bisadducts 2,
to tetraadducts 3b and 3a. This large Ds2 value enhancement
for 3b and 3a was correlated to the hindered starburst
structures which may give a higher degree of molecular
separation than 1 and 2. Therefore, we interpreted the
concentration-dependent phenomenon by the high tendency
of fullerene-DPAF chromophores to form nano-aggregates at
concentrations above 1023 M. In fact, nano-aggregation might
play a role in reducing the 2PA quantum efficiency leading to
the decrease of the s2 value. The starburst structure of
C60(.DPAF-C9)4 in a multipolar arrangement resembling
encapsulation of C60 by multiple DPAF-C9 chromophores is
likely to increase the degree of molecular dispersion at high
solution concentration and hamper the aggregation tendency,
thus maintaining high nonlinear optical efficiency. These
results provided an insight into structural control toward
managing molecular and nanoscale aggregation phenomena of
C60-based starburst and dendritic suprastructures in the
materials design of 2PA-based photonic devices and photo-
dynamic therapeutics.
3
processes in the C60*. cage moiety can be neglected in this
study since the current measurements were performed by using
150 fs laser pulses. The observed transmittance reduction
behavior may be attributed predominantly to the high
absorption cross-section of 2PA processes, mainly on DPAF-
C9 moieties, and the excited state absorption of the lowest
1
1
excited singlet state C60*.. Population of the latter C60*.
state can be highly enhanced by ultrafast intramolecular
energy transfer from the two-photon pumped excited singlet
state of the DPAF-C9 moiety in C60(.1DPAF*-C9) to form
1C60*(.DPAF-C9), which occurs at rate as fast as 130 fs28
to 15 ps.31
Under the conditions of Z-scan experiments, C60(.DPAF-
C9)4 showed surprisingly good photostability, which was
evaluated by comparing the absorption spectra before and
after laser irradiation. The results showed no difference in the
spectral profiles for all the samples used. Even though the laser
power applied was relatively high, and possibly had reached
above the decomposition threshold of the C60-DPAF mole-
cule, the application of ultrafast irradiation on a femtosecond
time scale did not seem to have accumulated sufficient energy
to impact the chromophore sample with noticeable damage.
To evaluate the recovery time of observed nonlinearities and to
reveal the underlying mechanism, we conducted a degenerate
pump–probe experiment with 150 fs, 780 nm laser pulses
close to the 2PA peak position of C60(.DPAF-C9) and
C60(.DPAF-C9)4 3a in CS2 solution. As shown in Fig. 10b,
the maximum transmittance change (2DT/T) in 3a solution
was found to be greater than that in the solution of 1,
indicating larger 2PA coefficients and s2 values of 3a than
those of 1. Furthermore, the response time of 3a was evaluated
to be y15 ps in a time scale necessary for intramolecular
energy transfer from the DPAF-C9 moiety to the methano-
[60]fullerene (C60.) moiety. It should be pointed out that an
oscillation in the relaxation time was present. The origin of this
oscillation is unclear. It may be attributed to the interplay
between the 2PA and energy transfer processes.
Acknowledgements
We thank the Air Force Office of Scientific Research for
funding under the contract number FA9550-05-1-0154.
References
Conclusions
1 D. M. Guldi and M. Prato, Acc. Chem. Res., 2000, 33, 69–73.
2 D. M. Guldi and P. V. Kamat, Fullerenes, Chemistry, Physics and
Technology, ed. K. M. Kadish and R. S. Ruoff, Wiley-Interscience,
New York, 2000, pp. 225–281.
3 M. Maggini and D. M. Guldi, Molecular and Supramolecular
Photochemistry, ed. V. Ramamurthy and K. S. Schanze, Marcel
Dekker, New York, 2000, vol. 4, pp. 149–196.
4 C. Yu, T. Canteenwala, M. E. El-Khouly, Y. Araki, L. Y. Chiang,
B. C. Wilson, O. Ito and K. Pritzker, J. Mater. Chem., 2005, 15,
1857–1864.
We demonstrated an approach toward the design of starburst
C60-keto-DPAF assemblies to extend and maintain the
molecular photoefficiency in highly concentrated solution by
applying
a combination of a starburst macromolecular
configuration with C60 as an encapsulated core and multiple
bulky groups. The structure was made for the progressive
increase of the aggregation barrier and the unimolecular
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1826–1838 | 1837

View Article Online
5 C. Yu, T. Canteenwala, H. H. C. Chen, B. J. Chen,
M. Canteenwala and L. Y. Chiang, Proc. Electrochem. Soc.,
1999, 12, 234–249.
27 Y. Zhao, Y. Shirai, A. D. Slepkov, L. Cheng, L. B. Alemany,
T. Sasaki, F. A. Hegmann and J. M. Tour, Chem.–Eur. J., 2005,
11, 3643–3658.
6 C. Yu, T. Canteenwala, L. Y. Chiang, B. Wilson and K. Pritzker,
Synth. Met., 2005, 153, 37–40.
28 P. A. Padmawar, J. O. Rogers, G. S. He, L. Y. Chiang,
T. Canteenwala, L.-S. Tan, Q. Zheng, C. Lu, J. E. Slagle,
E. Danilov, D. G. McLean, P. A. Fleitz and P. N. Prasad,
Chem. Mater., 2006, 18, 4065–4074.
29 B. Z. Tang, H. Xu, J. W. Y. Lam, P. P. S. Lee, K. Xu, Q. Sun and
K. K. L. Cheuk, Chem. Mater., 2000, 12, 1446–1455.
30 P. A. Padmawar, T. Canteenwala, L.-S. Tan and L. Y. Chiang,
J. Mater. Chem., 2006, 16, 1366–1378.
31 H. Luo, M. Fujitsuka, Y. Araki, O. Ito, P. Padmawar and
L. Y. Chiang, J. Phys. Chem. B, 2003, 107, 9312–9318.
32 T. W. Ebbesen, K. Tanigaki and S. Kuroshima, Chem. Phys. Lett.,
1991, 181, 501–504.
33 For a review of 2PA-active materials: T. C. Lin, S. J. Chung,
K. S. Kim, X. Wang, G. S. He, J. Swiatkiewicz, H. E. Pudavan and
P. N. Prasad, Adv. Polym. Sci., 2003, 161, 157–193.
34 G. S. He, J. D. Bhawalkar, C. F. Zhao and P. N. Prasad, Appl.
Phys. Lett., 1995, 67, 2433–2435.
35 J. E. Ehrlich, X. L. Wu, L.-Y. Lee, Z.-Y. Hu, H. Roeckel,
S. R. Marder and J. W. Perry, Opt. Lett., 1997, 22, 1843–1845.
36 M. G. Silly, L. Porres, O. Mongin, P.-A. Chollet and M. Blanchard-
Desce, Chem. Phys. Lett., 2003, 379, 74–80.
37 W. R. Dichtel, J. M. Serin, C. Edder, J. M. J. Frechet,
M. Matuszewski, L. S. Tan, T. Y. Ohulchanskyy and
P. N. Prasad, J. Am. Chem. Soc., 2004, 126, 5380–5381.
38 P. A. Padmawar, T. Canteenwala, S. Verma, L.-S. Tan and
L. Y. Chiang, J. Macromol. Sci., Pure Appl. Chem., 2004, 41,
1387–1400.
39 M. Sheik-Bahae, A. A. Said, T. Wei, D. J. Hagan and E. W. Van
Stryland, IEEE J. Quantum Electron., 1990, 26, 760.
40 Q. Ying, J. Marecek and B. Chu, Chem. Phys. Lett., 1994, 219, 214.
41 R. V. Hongchuck, T. W. Cruger and J. Milliken, J. Am. Chem.
Soc., 1993, 115, 3034.
42 L. A. Bulavin, I. I. Adamenko, V. M. Yashchuk,
T. Yu. Ogul’chansky, Y. I. Prylutskyy, S. S. Durov and
P. Scharff, J. Mol. Liq., 2001, 93, 187–191.
7 L. W. Tutt and A. Kost, Nature, 1992, 356, 225–226.
8 D. G. McLean, R. L. Sutherland, M. C. Brant, D. M. Brandelik,
P. A. Fleitz and T. Pottenger, Opt. Lett., 1993, 18, 858–860.
9 T. H. Wei, T. H. Huang, T. T. Wu, P. C. Tsai and M. S. Lin, Chem.
Phys. Lett., 2000, 318, 53–57.
10 Q. L. Zhou, J. R. Heflin, K. Y. Zamani-Khamiri and A. Garito,
Phys. Rev. A, 1991, 43, 1673.
11 A. Kost, L. W. Tutt, M. B. Klein, T. K. Dougherty and W. E. Elias,
Opt. Lett., 1993, 18, 334–336.
12 F. Z. Henari, K. H. Cazzini, D. N. Weldon and W. J. Blau, Appl.
Phys. Lett., 1996, 68, 619.
13 Y.-P. Sun and J. E. Riggs, Int. Rev. Phys. Chem., 1999, 18, 43.
14 J. E. Riggs and Y.-P. Sun, J. Phys. Chem. A, 1999, 103, 485–495.
15 Y. L. Song, G. Y. Fang, Y. X. Wang, S. T. Liu, C. F. Li, L. C. Song,
Y. H. Zhu and Q. M. Hu, Appl. Phys. Lett., 1999, 74, 332–334.
16 M. Maggini, C. D. Faveri, G. Scorrano, M. Prato, G. Brusatin,
M. Guglielmi, M. Meneghetti, R. Signorini and R. Bozio, Chem.–
Eur. J., 1999, 5, 2501–2510.
17 K. Dou and E. T. Knobbe, J. Nonlinear Opt. Phys. Mater., 2000, 9,
269–287.
18 Z. B. Liu, J. G. Tian, W. P. Zang, W. Y. Zhou, C. P. Zhang,
J. Y. Zheng, Y. C. Zhou and H. Xu, Appl. Opt., 2003, 42,
7072–7076.
19 E. Koudoumas, M. Konstantaki, A. Mavromanolakis, S. Couris,
M. Fanti, F. Zerbetto, K. Kordatos and M. Prato, Chem.–Eur. J.,
2003, 9, 1529–1534.
20 G. Kopitkovas, A. Chugreev, J. F. Nierengarten, Y. Rio,
J. L. Rehspringer and B. Honerlage, Opt. Mater., 2004, 27,
285–291.
21 H. I. Elim, J. Quyang, S. H. Goh and W. Ji, Thin Solid Films, 2005,
477, 63–72.
22 K. A. Drenser, R. J. Larsen, F. P. Strohkendl and L. R. Dalton,
J. Phys. Chem. A, 1999, 103, 2290–2301.
23 S. V. Rao, D. N. Rao, J. A. Akkara, B. S. DeCristofano and
D. V. G. L. N. Rao, Chem. Phys. Lett., 1998, 297, 491–498.
24 G. Banfi, D. Fortusini, M. Bellini and P. Milani, Phys. Rev. B:
Condens. Matter, 1997, 56, R10075–R10078.
25 L. Y. Chiang, P. A. Padmawar, T. Canteenwala, L.-S. Tan,
G. S. He, R. Kannan, R. Vaia, T.-C. Lin, Q. Zheng and
P. N. Prasad, Chem. Commun., 2002, 1854–1855.
43 L. B. Justus, Z. H. Kafafi and A. L. Huston, Opt. Lett., 1993, 18,
1603.
44 R. A. Ganeev, A. I. Ryasnyansky, N. Ishizawa, M. Baba,
M. Suzuki, M. Turu, S. Sakakibara and H. Kuroda, Opt.
Commun., 2004, 231, 431–436.
45 H. I. Elim, J. Y. Ouyang, J. He, S. H. Goh, S. H. Tang and W. Ji,
Chem. Phys. Lett., 2003, 369, 281.
26 P. A. Padmawar, T. Canteenwala, S. Verma, L.-S. Tan, G. S. He,
P. N. Prasad and L. Y. Chiang, Synth. Met., 2005, 154, 185–188.
46 H. I. Elim, W. Ji, G. C. Meng, J. Y. Ouyang and S. H. Goh,
J. Nonlinear Opt. Phys., 2003, 12, 175.
1838 | J. Mater. Chem., 2007, 17, 1826–1838
This journal is ß The Royal Society of Chemistry 2007