Excited-state interactions in pyrrolidinofullerenes

Article abstract of DOI:10.1021/jp972756z

Two new pyrrolidinofullerenes, 1 and 2, have been synthesized, and their photophysical properties have been investigated. The pyrrolidinofullerene 1 has three phenyl groups attached to the 1,2, and 5 positions of the pyrrolidine ring. The pyrrolidinofulle

Full text of DOI:10.1021/jp972756z

J. Phys. Chem. A 1998, 102, 5341-5348
5341
Excited-State Interactions in Pyrrolidinofullerenes
K. George Thomas,* V. Biju, and M. V. George*,†
Photochemistry Research Unit, Regional Research Laboratory, TriVandrum 695 019, India
Dirk M. Guldi*,‡ and Prashant V. Kamat*,§
Radiation Laboratory, UniVersity of Notre Dame, Notre Dame, Indiana 46556
ReceiVed: August 25, 1997; In Final Form: December 2, 1997
Two new pyrrolidinofullerenes, 1 and 2, have been synthesized, and their photophysical properties have been
investigated. The pyrrolidinofullerene 1 has three phenyl groups attached to the 1, 2, and 5 positions of the
pyrrolidine ring. The pyrrolidinofullerene 2, on the other hand, has a flexible attachment with an
N-methylaniline end group and is the prototype of a fullerene-aniline dyad. Singlet and triplet excited-state
properties of these two functionalized fullerene derivatives have been examined with picosecond and nanosecond
laser flash photolysis. The singlet and triplet excited states of these fullerenes exhibit characteristic absorption
bands in the vis-IR region of the spectrum. The functionalization of C60 with pyrrolidine groups shift the
excited-state absorption maxima to the blue. Three different quenchers, O2, TEMPO, and ferrocene, are
employed to investigate the reactivity of triplet excited states. The bimolecular quenching rate constants
determined for these quenchers were in the range of 5.3 × 108 to 7.7 × 109 M-1 s-1. The excited-state
interactions of these functionalized fullerenes are compared to that of C60.
Introduction
CHART 1
Design of covalently linked donor-acceptor systems, which
can mimic photosynthetic reaction center and generate long-
lived radical pairs, employing photoinduced electron-transfer
reactions are of considerable interest (see for example, ref 1-3).
The efficiency of charge separation and recombination in such
systems can be conveniently tuned by varying the redox
properties of the donor as well as the acceptor and also the
distance between them. The remarkable photophysical proper-
ties of fullerenes,4-8 particularly their ability to undergo
multistep9-11 reduction, make them ideal acceptors for designing
such molecular dyads.
Fullerenes form charge-transfer complexes with electron
donors such as aromatic amines12-15 but undergo nucleophilic
addition reactions with aliphatic amines.4,16,17 The free energy
change, calculated for electron transfer from N,N-dimethylaniline
to 1C60 and 3C60, indicates that this is an energetically favorable
process (∆G ) -18.2 and -9.8 kcal mol-1, respectively).4 The
electrode materials.23 Photoinduced electron-transfer processes
in C60-based dyads containing donors such as ferrocene,24
porphyrin,25,26 and carotenoid,27 etc. have also been reported
in the literature. In this paper, we report the synthesis and the
detailed investigation on the photoinduced electron transfer of
a novel C60 derivative 2 (Chart 1), covalently linked to aniline
through a semiflexible chain. These results are compared with
the excited-state properties of a similar compound, 1, which
has three phenyl substituents on the pyrrolidine ring. The
intermolecular electron-transfer studies between 1 and N,N-di-
methylaniline (DMA) are also presented.
intermolecular photoinduced electron transfer in this system has
13-15,18,19
been investigated by various groups
using transient
absorption studies. More recently, Williams et al.20,21 have
studied the intramolecular electron-transfer processes in C60-
aniline dyads possessing rigid spacers. Preliminary results on
the photophysical and electrochemical studies of a few N-
phenylpyrrolidinofullerenes, prepared through the 1,3-dipolar
cycloaddition of azomethine ylides to C60, have been com-
municated in a recent communication.22 Similar pyrrolidine
derivatives of C60 have been used to modify SnO2 electrodes
for the purpose of developing photoelectrochemically active
Experimental Section
Methods. All melting points are uncorrected and were
determined on a Aldrich melting point apparatus. IR spectra
were recorded on a Perkin-Elmer model 882 IR spectrometer
and the UV-visible spectra on a Shimadzu 2100 or GBC 918
spectrophotometer. 1H NMR and 13C NMR spectra were
recorded on either a JEOL EX-90 or Varian VXR 500S
spectrometer. Mass spectra were recorded on a JEOL JMS AX
505HA mass spectrometer. The emission spectra were recorded
† Also at the Radiation Laboratory, University of Notre Dame, and the
Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore 560
064, India (E-mail: mvg@csrrltrd.ren.nic.in).
‡ E-mail: Guldi@macroni.rad.nd.edu.
§ E-mail: KAMAT.1@ND.EDU or http://www.nd.edu/∼pkamat.
S1089-5639(97)02756-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/30/1998

5342 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Thomas et al.
SCHEME 1
NMR (CDCl3) δ 6.62-6.82, 6.9-7.82, and 7.9-8.02 (34 H,
m, methine and aromatic). Anal. Calcd for C100H34N2: C,
95.02; H, 2.71; N, 2.21. Found: C, 95.80. H, 2.51; N, 1.93.
Mol wt calcd for C100H35N2 (MH+) 1264; found, 1264 (FAB).
Further characterization of this bis-adduct was not done, and it
was not used in the present studies.
Synthesis of the Fullerene-Aniline Dyad (2). A mixture
of C60 (90 mg, 0.12 mmol), the aldehyde 7 (32 mg, 0.12 mmol),
and N-methylglycine (11 mg, 0.12 mmol) in toluene (90 mL)
was stirred under reflux for 10 h. The reaction was cooled,
and removal of the solvent under reduced pressure gave a solid
residue, which was chromatographed over silica gel (100-200
mesh). Elution with petroleum ether gave 30 mg (33%) of the
unchanged C60. Further elution with a mixture (2:3) of toluene
and petroleum ether gave 48 mg (58%) of the fullerene-aniline
adduct, 2, mp > 400 °C. UV λmax (CHCl3) 256 nm (ꢀ 115 000),
306 (35 700); IR νmax (KBr) 2932, 2861, 1739, 1600, 1507,
SCHEME 2
1
1463, 1427, 1249, 1172, 1035, 825, and 744 cm-1; H NMR
(CDCl3) δ 2.85 (3 H, s, NCH3), 3.0 (3 H, s, NCH3), 3.6-3.8 (2
H, t, NCH2) 3.95-4.25 (3 H, m, OCH2 and CH pyrrolidine
ring), 4.8 (1 H, s, CH), 4.95 (1 H, d, CH pyrrolidine ring), and
6.5-7.8 (9H, m, aromatic); 13C NMR (CDCl3), 29.71, 39.16,
39.97, 51.92, 65.15, 68.96, 70.98, 83.13, 112.04, 116.46, 125.29,
128.2, 129.03, 129.14, 129.26, 130.49, 135.74, 135.78, 136.53,
136.77, 139.59, 139.90, 140.11, 140.15, 141.53, 141.67, 141.83,
141.95, 141.99, 142.03, 142.07, 142.12, 142.24, 142.54, 142.67,
142.97, 143.13, 144.38, 144.61, 144.69, 145.14, 145.22, 145.27,
145.33, 145.47, 145.53, 145.77, 146.09, 146.124, 146.20,
146.26, 146.30, 146.33, 146.51, 146.78, 147.29, 148.84, 153.58,
153.62, 154.09, 156.35, 158.70; exact mol wt calcd for
C78H22N2O (M+•) 1002.1732l; found, 1002.1725 (FAB high-
resolution mass spectrometry).
on SLM-8100 and Spex-Fluorolog, F112-X spectrofluorometers
equipped with a 450 W Xe lamp and a Hamamatsu R928
photomultiplier tube. The excitation and emission slits were 4
and 1, respectively. A 570-nm long-pass filter was placed
before the emission monochromator in order to eliminate the
interference from the solvent, and no corrections were applied
for the fluorescence spectra. Quantum yields of fluorescence
were measured by a relative method using optically dilute
solutions (absorbance was adjusted to 0.1 at 470 nm, and the
emission intensity was measured at 713 nm). C60 dissolved in
toluene (φf ) 2.2 × 10-4) was used as reference.28 Spectroscopy-
grade solvents were used for all measurements, and the solutions
were purged with argon before use.
Starting Materials. The fullero-1,2,5-triphenylpyrrolidine
1 was synthesized through the 1,3-dipolar cycloaddition of the
azomethine ylide 4 with C60 as shown in Scheme 1. The
azomethine ylide 4 itself was generated through the thermal
ring opening of 1,2,3-triphenylaziridine 3. The starting tri-
phenylazirdine 3,29 mp 99 °C, was prepared by a reported
procedure. C60 was purchased from SES Corporation. Syn-
thesis of the fullerene-aniline dyad 2 was achieved through
the sequence of reactions shown in Scheme 2.
Preparation of N-Methyl-(2-bromoethyl)aniline (6). To
an ice cold solution of N-methyl-N-(2-hydroxyethyl)aniline (5)
(2.25 g, 15 mmol) in dichloromethane (20 mL) was added PBr3
(4.05 g, 15 mmol), dropwise over a period of 1 h. The reaction
mixture was further stirred at room temperature for an additional
period of 3 h. The solvent was removed, and the crude product
was chromatographed on silica gel (100-200 mesh) using a
mixture (1:10) of ethyl acetate and petroleum ether to give 1.9
g (60%) of 6. IR νmax (neat) 2929, 1605, 1510, 1373, 1351,
1278, 1214, 1175, 1097, 1035, and 996 cm-1; 1H NMR (CDCl3),
δ 2.95 (3 H, s, NCH3) 3.2-3.6 (2 H, t, NCH2), 3.6-3.9 (2 H,
t, CH2), and 6.6-7.5 (5 H, m, aromatic); 13C NMR (CDCl3) δ
146.45, 127.74, 115.42, 110.41, 52.83, 36.90, and 26.87; exact
mol wt calcd for C9H12NBr (M+•) 213.0153; found, 213.0142
(FAB high-resolution mass spectrometry).
Synthesis of Fullero-1,2,5-triphenylpyrrylodine (1).
A
mixture of C60 (72 mg, 0.1 mmol) and 3 (27 mg, 0.1 mmol) in
toluene (40 mL) was refluxed for 6 h. The reaction mixture
was cooled, and removal of the solvent under reduced pressure
gave a solid residue, which was chromatographed over silica
gel. Elution with petroleum ether gave 20 mg (28%) of the
unchanged C60. Further elution with a mixture (1:4) of toluene
and petroleum ether gave 30 mg (42%) of the monoadduct 1,
mp > ,400 °C. IR νmax (KBr) 1601, 1500, 1543, 1450, 1265,
696 cm-1; UV λmax (CHCl3) 257 nm (ꢀ 123 000), 312 (38 000);
1H NMR (CDCl3) δ 6.68-6.82, 7.1-7.4, and 7.72-7.84 (17
H, m, methine and aromatic). Anal. Calcd for C80H17N: C,
96.94; H, 1.73; N, 1.41. Found: C, 96.63; H, 1.73; N, 1.21.
Mol wt calcd for C80H18N (MH+) 992.1439; found, 992.1460
(FAB, high-resolution mass spectrometry).
Preparation of N-Methyl-N-{(p-formylphenoxy)-2-ethyl}-
aniline (7). A suspension of the bromide, 6 (0.27 g, 1 mmol),
p-hydroxybenzaldehyde (0.24 g, 2 mmol), and potassium
carbonate (0.28, 2 mmol) was refluxed in acetone (15 mL) for
12 h. The reaction mixture was cooled, filtered, and concen-
trated under reduced pressure. The residue was chromato-
graphed on silica gel (100-200 mesh) using a mixture (1:10)
of ethyl acetate and petroleum ether to give 0.2 g (80%) of 7.
IR νmax (neat) 2934, 2759, 1696, 1602, 1476, 1262, 1161, 1028,
and 798 cm-1; 1H NMR (CDCl3) δ 2.95 (3 H, s, NCH3), 3.6-
3.8 (2 H, t, NCH2), 4.0-4.3 (2 H, t, OCH2), 6.4-7.9 (9H, m,
aromatic), and 9.8 (1 H, s, CHO); 13C NMR (CDCl3) δ 190.37,
163.67, 148.69, 131.90, 130.17, 129.30, 116.45, 114.74, 112.24,
96.15, 65.69, 51.79, and 39.17; exact mol wt calcd for C16H17-
NO2 (M+•) 255.1259; found, 255.1266 (FAB high-resolution
mass spectrometry).
Further elution of the silica gel column with a mixture (3:7)
of toluene and petroleum ether gave 20 mg (28%) of a bis-
adduct, mp > 400 °C. IR νmax (KBr) 1602, 1569, 1487, 695
1
cm-1; UV λmax (CHCl3) 252 nm (ꢀ 92 000), 312 (36 000); H

Excited-State Interactions in Pyrrolidinofullerenes
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5343
Figure 2. Fluorescence emission spectrum of (a) (s) 1 and (b) (- - -)
2 in methylcyclohexane at 77 K. (The excitation wavelength was at
400 nm.)
Figure 1. Absorption spectra of dyad 2 (trace a) and model system 1
(trace b), in dichloromethane. Inset shows the changes in the long-
wavelength absorption band of dyad 2 in dichloromethane on addition
of TFA. The concentrations of TFA were (a) 0 mM and (c) 25 mM.
nm, with those of various other [6-6] closed fullerene deriva-
tives is indicative of the partial perturbation of the fullerene
core. This can be rationalized in view of the fact that
monofunctionalization of the fullerene core impacts the elec-
tronic structure and leads to a change of the Ih-symmetry of
pristine C60, which converts into an effective C2ν′ symmetry.
Picosecond Laser Flash Photolysis. Picosecond laser flash
photolysis experiments were performed with 355-nm laser pulses
from a mode-locked, Q-switched Quantel YG-501 DP Nd:YAG
laser system (output 1.5 mJ/pulse, pulse width ∼18 ps).30 The
white continuum picosecond probe pulse was generated by
passing the fundamental output through a D2O/H2O mixture.
The output was fed to a spectrograph (HR-320, ISDA Instru-
ments, Inc.) with fiber optic cables and was analyzed with a
dual diode array detector (Princeton Instruments, Inc.), interfaced
with an IBM-AT computer. Time zero in these experiments
corresponds to the end of the excitation pulse. All the lifetimes
and rate constants reported in this study are at room temperature
(297 K) and have an experimental error of (10%. The
deaerated dye solution was continuously flowed through the
sample cell during the measurements.
Nanosecond Laser Flash Photolysis. Nanosecond laser
flash photolysis experiments were performed with a Laser
Photonics PRA/model UV-24 nitrogen laser system (337 nm,
2-ns pulse width, 2-4 mJ/pulse) with front face excitation
geometry. A typical experiment consisted of a series of 2-3
replicate shots per single measurement. The average signal was
processed with an LSI-11 microprocessor interfaced with a VAX
computer. Details of the experimental setup can be found
elsewhere.31
Absorption spectra of both the compounds in the visible
region are significantly red-shifted and distinctly different from
that of pristine C60. To investigate the possibility of a ground-
state charge-transfer formation at longer wavelengths, trifluo-
roacetic acid (TFA) was added to dichloromethane solutions
of 2. Dyad 2 possesses a long-wavelength absorption band
around 702 nm in dichloromethane, and a hypsochromic shift
of 14 nm was observed upon addition of 25 mM of TFA (Figure
1, trace c). These changes can be completely reversed by adding
30 mM of pyridine to the above solution. These results indicate
the possibility of ground-state electronic interaction between
C60 and aniline in dyad 2 as the protonation of nitrogen inhibits
such a charge-transfer interaction. In the case of 1, the
absorption spectrum remains unaffected even after adding 250
mM of TFA. Similar effects were observed for other N-
phenylpyrrolidinofullerenes.22
Steady-State Fluorescence. The fluorescence emission
spectrum of 1 in methylcyclohexane at 77 K exhibits a
maximum at 702 nm (Figure 2). Vibronic fine structure was
observed at 715, 725 (sh), 741, 754, 783, 806 (sh), and 824
nm. The structural pattern is in good agreement with other
monofunctionalized fullerene derivatives and with the mirror-
imaged UV-vis absorption features (see above). Since the
distribution of levels in the first-excited state, governing the
position of the absorption bands, resembles the distribution of
levels in the ground state, which is responsible for the position
of the fluorescence bands, fluorescence and absorption features
should be very similar. The experimental results show a good
mirror image, e.g., the absorption band (0 f *0, 701 nm) of
the longest wavelength and the corresponding emission band
(*0 f 0, 702 nm) of the shortest wavelength. This further
supports an unambiguous assignment of the 0 f 0 transition
band.
Pulse Radiolysis. Pulse radiolysis experiments were per-
formed by utilizing 50-ns pulses of 8-MeV electrons from a
model TB-8/16-1S Electron Linear Accelerator. Dosimetry was
based on the oxidation of SCN- to (SCN)2•-, which in N2O-
saturated aqueous solutions takes place with G ∼ 6 (G denotes
the number of species per 100 eV, or the approximate micro-
molar concentration per 10 J of absorbed energy). The radical
concentration generated per pulse amounts to (1-3) × 10-6
M
for all the systems investigated in this study.
Results and Discussion
The emission spectrum of 2 in methylcyclohexane (Figure
2), for example, displays a similar *0 f 0 band at 702 nm,
followed by some vibronic states emitting at 715, 725 (sh), 741,
754, 783, 806 (sh), and 824 nm. Although the emission pattern
is identical with 1 and various monofunctionalized fullerene
derivatives, the relative emission yield of derivative 2 is lower
compared to a model fullerene compound. Increasing the
Absorption Spectra. The absorption spectra of 1 and 2
recorded in dichloromethane at room temperature are shown in
Figure 1. Both show similar absorption properties in the UV
region with strong molar absorptivity at 211, 257, 310, and 430
nm. The close resemblance of the absorption pattern of 1 and
2, with absorption maxima at 701, 690, 680, 667, 655, and 636

5344 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Thomas et al.
TABLE 1: Fluorescence Quantum Yields (Absolute and
Relative) of 1 and 2 in Solvents of Varying Polarity at 77 K
φF × 104 (absolute)
solvents
1
2
methylcyclohexane
toluene
dichloromethane
toluene-acetonitrile (1:1)
6.0 (100)
4.4 (100)
4.2 (95)
3.8 (86)
1.9 (43)
Figure 3. Effect of addition of acetonitrile on the emission spectra of
dyad 2 in toluene solution. Acetonitrile v/v (a) 0%, (b) 10%, (c) 30%,
(d) 50%. (Corrections for the changes in the absorbance at the excitation
wavelength (470 nm) have been made for determining φf.)
dielectric constant of the solvents resulted in a moderate
amplification of this decrease. This observation suggests, in
line with the earlier assumption, that intramolecular interaction
between the photoexcited fullerene moiety and the aniline group
contributes to the overall deactivation process of photoexcited
2. The emission properties of the N-phenyl-substituted pyrrol-
idinofullerene, 1, and the dyad, 2, have been investigated in
solvents of a wide range of polarity, and these results are
summarized in Table 1. It is interesting to note that 2 exhibits
a solvent-dependent fluorescence whereas the emission proper-
ties of 1 are independent of the dielectric constant of the solvent.
To investigate the effect of solvent polarity, the emission
properties of 2 were studied in toluene-acetonitrile mixtures
of varying compositions (Figure 3). A marked decrease in the
quantum yield of fluorescence (φF) was observed upon increas-
ing the polarity of the solvent. The reductive quenching of the
singlet excited state of fullerene by aniline moiety is responsible
for the decrease in the fluorescence yield in polar solvents. The
fluorescence of 2 is restored upon addition of TFA to the above
solution. The protonation of nitrogen atom of the aniline group
inhibits the electron-transfer process in the donor-acceptor
system. Similar effects were earlier observed in fullerene-
aniline dyads connected through rigid spacers20,21 and fullerene-
ferrocene dyads.24 Earlier studies on the photophysical prop-
erties of N-methylpyrrolidinofullerene have indicated that the
pyrrolidine ring nitrogen is not involved in the quenching
process.24 In an earlier study we had found that φF of 1 (∼6 ×
10-4) is independent of solvent polarity and that the addition
of TFA does not affect its emission properties.22 However, the
aniline group is a good quencher for the singlet excited fullerene.
Independent fluorescence-quenching studies carried out with
dimethylaniline indicate that singlet excited 1 can be quenched
with DMA with a bimolecular rate constant of 3.8 × 1010 M-1
Figure 4. Transient absorption spectra recorded following 355-nm laser
pulse (pulse width 18 ps) excitation of toluene solution containing (A,
top) 1 at delay times (a) 60, (b) 1500, and (c) 5000 ps and (B, bottom)
2 at delay times (a) 60, (b) 1500, and (c) 4000 ps.
of 1 and 2 were thus completely abolished in solutions
containing ethyl iodide (methylcyclohexane, 2-methyltetrahy-
drofuran, ethyl iodide at 2:1:1 v/v). Instead, a new band
appeared at 824 nm that is assigned to the phosphorescence
emission. This finding implies a significant red shift of the
phosphorescence-related emission relative to pristine C60 (796
nm) but is well in line with various monofunctionalized
derivatives.
Singlet Excited-State Properties. The excited singlet of
pristine C60 has an absorption maximum at 920 nm with a
lifetime of 1.2 ns, and the triplet excited state has an absorption
maximum at 740 nm with a lifetime greater than 100 µs.33-35
The time-resolved transient absorption spectra recorded follow-
ing 355-nm laser pulse excitation of the two functionalized
fullerenes, 1 and 2, are shown in Figure 4 A,B. The spectra
recorded immediately after 355-nm laser pulse (pulse width 18
ps) excitation show the formation of singlet excited state with
a broad absorption maxima in the region of 880 and 890 nm
for 1 and 2, respectively. As the singlet excited state decays,
a new absorption band corresponding to the triplet excited state
appears with an absorption maximum in the region of 690-
700 nm. The intersystem crossing observed in the excited state
is quite efficient similar to the one observed for pristine C60. It
should be noted that the functionalization of C60 has a significant
impact on the electronic and spectral properties of the excited
state.
s-1
.
Steady-State Phosphorescence. External heavy atom effect
is known to accelerate the transformation rate from the excited
singlet to the excited triplet state. Fullerene related emissions

Excited-State Interactions in Pyrrolidinofullerenes
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5345
The kinetics of excited-state transients was probed by
recording absorption-time profiles at different wavelengths.
Few representative spectra and decay traces for the dyad 2 are
shown in Figure 5A,B. The excited singlet lifetimes of 1 and
2 were obtained from the transient decay recorded at 880 and
890 nm, respectively. The corresponding triplets monitored at
700 nm showed a growth with a rate constant similar to that
obtained from the corresponding decay of the singlet excited
state. The lifetime of the singlet excited states as monitored
from these kinetic traces were 1.25 and 1.05 ns for 1 and 2,
respectively. While the singlet excited lifetime of 1 is not
different from that of pristine C60, the singlet excited 2 had a
shorter lifetime. The decrease in lifetime observed in these
experiments is parallel to the decrease in fluorescence yield
observed in higher dielectric media. Presence of a covalently
attached aniline functionalizing group, which is a good electron
donor, prompts the possibility of an intramolecular quenching
with electron transfer.
It should be noted that the photoexcitation of 2 in benzonitrile
leads to decreased intersystem-crossing yield. The high dielec-
tric constant of the medium (e.g., benzonitrile relative to toluene,
CH2Cl2, etc.) is thermodynamically supportive of intermolecular
electron transfer. Furthermore, the flexible nature of the linking
spacer enables conformational rearrangements of the fullerene
and aniline centers. Thus it is conceivable that a combination
of these factors increase the thermodynamic driving force for
an intramolecular charge-transfer process and, in turn, allow
this route to compete with the fast intersystem crossing. The
observed shorter lifetime of singlet excited 2 supports this view.
Suggestions have been made in the past that C60 undergoes
charge-transfer interactions with electron-donating species such
as amines in both the ground and excited states.18,19,32 Similarly
charge-transfer complex formation has also been observed in
fullerene-donor type dyad moieties. Earlier studies have shown
that an excited fullerene-donor complex exhibits a broad
absorption in the 700-900-nm region.19 However, in the
present case it was not possible to obtain clean evidence for
such an excited charge-transfer complex formation because of
the dominance of singlet and triplet absorption in this spectral
region. Also, the observation of an isosbestic point between
the spectra of singlet and triplet excited states highlights the
major contribution from these two species to the transient spectra
in Figure 4.
Figure 5. (A) Time-resolved transient absorption spectra showing the
decay of singlet and formation of triplet excited states following the
355-nm laser pulse excitation of 2 in toluene. (B) Absorption-time
profiles showing the decay of singlet dyad, 2, in different solvents at
monitoring wavelength of 900 nm.
do not indicate the presence of any charge-separated state. This
makes us conclude that either the contribution of intramolecular
quenching to the overall decay is small or the charge-separated
pair recombines to yield triplet excited state as indicated in
reaction 3. The possibility of formation of triplet excited state
via recombination of charge-separated pair has been proposed
earlier for C60-amine complexes.33 By comparing the de-
creased fluorescence yield of 2 in high dielectric media, we
estimate the contribution from reactions 3 and 4 to the overall
decay to be around 30%. This sets the upper limit for the
observed interaction between the two moieties in the excited
state of the dyad molecule.
Triplet Excited-State Properties. Nanosecond laser flash
photolysis experiments were performed to investigate the triplet
excited-state behavior of 1 and 2 in various solvents. The time-
resolved transient spectra recorded following 337 (or 355)-nm
laser pulse excitation are shown in Figure 6. The spectrum
recorded immediately after the laser pulse excitation shows
spectral characteristics similar to the long-term spectra in Figure
4. This suggests that the triplet formation be completed within
the pulse duration of the excitation pulse. These experiments
also enabled us to characterize the absorption bands of triplet
excited states in the 350-400-nm region. The spectral char-
acteristics of the triplet excited states are summarized in Table
2.
In the case of excited singlet fullerene dyad such as 2, we
expect several possibilities for deactivation including those
illustrated in reactions 1-4.
C60-X + hν f 1(C60-X)* f C60-X + hν′
1(C60-X)* f 3(C60-X)*
(1)
(2)
1(C60-X)* f (C60•-‚‚‚X•+) f 3(C60-X)*
1(C60-X)* f (C60•-‚‚‚X•+) f (C60-X)
(3)
(4)
The time-resolved spectra recorded in Figure 6 did not
indicate the formation of any other transients during the
deactivation of the triplet excited state. The lifetimes of the
triplet excited states of the two title compounds were obtained
by fitting the decay traces to first-order kinetics. The lifetimes
of the triplet excited states of 1 and 2 were in the range of 7-17
µs. These lifetimes are shorter than the one reported for the
triplet excited C60. Although self-quenching processes such as
excited-state annihilation and ground-state quenching processes
significantly shorten the lifetime of the excited state of
The intersystem crossing is the major deactivation process in
the case of pristine C60 and monofunctionalized derivatives.
Presence of an electron donor such as the aniline group in 2
adds the possibility of a rapid intramolecular electron-transfer
quenching of excited singlet, which competes with intersystem
crossing and other deactivation processes. The fluorescence-
quenching studies indicated in the previous section strongly
support this view. However, the transient absorption studies

5346 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Thomas et al.
Figure 7. Dependence of pseudo-first-order rate constants for the triplet
decay of 2 on the concentration of different quenchers. The triplet decay
was monitored at 700 nm.
TABLE 3: Bimolecular Quenching Rate Constants
(M-1 s-1
)
O2
tempo
ferrocene
C60
1
2
1.6 × 109 (ref 42)
2.1× 109
3.3 × 109 (ref 39)
5.3× 108
3.3× 109(ref 43)
7.4 × 109
2.14 × 109
6.6 × 108
7.7 × 109
two transient species. A complex exponential decay kinetics
in the UV prevented, however, a meaningful analysis of the
associated processes.
Figure 6. Time-resolved difference absorption spectra recorded
following 337-nm laser pulse excitation of (A, top) 1 and (B, bottom)
2 in toluene. The spectra of 1 were recorded at delay times (a) 0, (b)
21.6, and (c) 53.6 µs and those of 2 at 0.25 and 5 µs, respectively, in
a nanosecond laser flash photolysis apparatus.
Triplet-State Quenching Reactions. To check the reactivity
of the triplet excited states we studied the quenching reactions
with three different quenchers, viz., O2, TEMPO, and ferrocene.
The triplet decay at the corresponding absorption maximum was
recorded at different quencher concentrations. The bimolecular
quenching rate constants were determined from the slope of
the plots on basis of the the relationship kd ) 1/τT + kq(Q),
where (Q) is the quencher concentration and τT is the lifetime
of the triplet excited state in the absence of the quencher. The
quenching plots for various quenchers are shown in Figure 7.
The rate constants determined from these plots are summarized
in Table 3.
TABLE 2: Excited-State Characteristics
S1 (λmax
)
τ (ps)
1200
1250 ( 100 690, 370
1050 ( 100 700
T1 (λmax
)
τ (µs)
C60 (from refs 34, 35, 37)
1
2
920
880
890
400, 750 >100
17 ( 1
7 ( 1
fullerenes,34-36 these processes alone cannot explain the short
triplet lifetimes of pyrrolidinofullerenes. A significant desta-
bilization of the fullerene triplet excited state was also evident
in corresponding measurements involving dyad 2. Since the
measurements of photoexcited 1 and 2 were conducted under
similar experimental conditions, the reduced lifetime, with decay
rates typically in the order of 1.25 × 105 s-1, suggests quenching
of the photoexcited fullerene moiety by the covalently attached
amine group. While the triplet quantum yield of 1 is similar to
that of pristine C60 (Φ ) 0.95),34,35,37 the triplet quantum yield
for 2 was 30% smaller. (We independently measured triplet
quantum yield of C60, 1 and 2 by triplet-triplet energy-transfer
method using the squaraine dye as an acceptor.38) Thus, the
lower triplet quantum yield observed for 2 further substantiates
the intramolecular quenching of the singlet excited state as a
competing process to the intersystem crossing.
In accordance with the accelerated picosecond dynamics, the
triplet yield of 2 in benzonitrile was appreciably smaller than
in toluene. As discussed in the radiolytic section, π-radical
anion’s of monofunctionalized fullerenes have a fairly strong
absorption around 400 nm, parallel to the coinciding absorption
of the triplet excited state (360 nm). The molar absorptivity of
the latter is equal or lesser than the dominant (T1* f Tn*)
transition around 700 nm. The differential absorption changes
recorded for 2 in benzonitrile indicate, however, a stronger
transition around 370-380 nm. Furthermore, the noticeable
red-shift can be interpreted as the concomitant generation of
The oxygen quenching rate constants for 1 and 2 are very
similar (∼2.0 × 109 M-1 s-1). As indicated in previous studies,
fullerenes quite efficiently generate singlet oxygen by the
energy-transfer method (reaction 5).
(3C60)-X* + O2 f C60-X + 1O2
(5)
Similarly, in the case of ferrocene as a quencher, the rate
constants are nearly diffusion-controlled ((3.6-7.7) × 109 M-1
s-1). The deactivation of the excited triplet state proceeds via
electron transfer to ferrocene. However, lack of transient
absorbance from long-lived intermediates other than the triplet
excited states makes it difficult to elucidate the mechanism of
electron-transfer quenching in the present experiments. It should
be noted that previous study of the interaction between excited
triplet and ferrocene has shown that photoinduced electron
transfer can occur in high dielectric media.24
The rate constants for the quenching by the stable nitroxyl
radical, TEMPO, were very similar for the two compounds, 1
and 2 (6.6 × 108 M-1 s-1 and 5.3 × 108 M-1 s-1, respectively).
The excited-state energy transfer to nitroxyl radical is unlikely
in the present case since doublet energy for TEMPO (∼47 kcal/
mol) is significantly higher than the triplet excited states of 1
and 2, which we expect to be in the range similar to that of
1C60* (37.5 kcal/mol). Therefore charge transfer between triplet

Excited-State Interactions in Pyrrolidinofullerenes
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5347
Figure 8. Square-wave voltammogram of 1 in toluene:acetonitrile (3:1
v/v). (Scan rate ) 100 mV/s; RE; SCE; WE, glassy carbon; CE, Pt;
supporting electrolyte, 0.1 M tetrabutylammonium perchlorate.)
TABLE 4: Reduction Potentials (ERed (mV) vs SCE in
Toluene:Acetonitrile (3:1))
E1
E2
E3
E4
E5
a
C60
1
2
-360
-648
-585
-830
-1060
-1010
-1420
-1676
-1580
-2010
-2100
-1980
-2600
-2250
-2200
Figure 9. Transient absorption spectrum (near-IR) of (A) 1 and (B) 2
obtained upon pulse radiolysis of 2.0 × 10-5 M of the respective
functionalized fullerene derivative in a nitrogen-saturated solvent
mixture containing 80 vol % toluene, 10 vol % 2-propanol, and 10 vol
% acetone.
a In benzene, from ref 44.
excited fullerene and TEMPO seems to be the likely possibility.
(3C60)-X* + TEMPO• f {(C60•-)-X‚‚‚TEMPO+}
f C60-X + TEMPO• (6)
This radical is known to react quite rapidly with C60, C70, and
C
C84
84 yielding the respective radical anions, e.g. C60•-, C70•-, and
•- 40,41
.
The same consideration is applied to study one-
electron reduction of functionalized fullerenes.
The lack of observation of radical anions of fullerenes suggests
that the back electron transfer of the products is quite efficient
and results in regeneration of original reactants. Low dielectric
constant of the solvent is likely to favor such a back reaction.
Similar observations have also been made in the previous studies
involving the quenching of 3C60* by various nitroxyl radicals.39
Electrochemical Reduction. Fullerenes undergo multiple
reductions when subjected to electrochemical and chemical
methods. Up to five reversible reductions have been observed
for C60. The square-wave voltammetry was employed to obtain
the reduction characteristics of the two functionalized fullerenes.
An example of successive reduction peaks observed in the
square-wave voltammogram of 1 is illustrated in Figure 8. We
also recorded cyclic voltammograms to cross-check the assign-
ment of our peak positions and the reversibility of the reduction
steps. The reduction potentials of 1 and 2 are compared with
that of pristine C60 in Table 4. The first reduction occurs in
the range of -0.585 to -0.65 V vs SCE, while the second
reduction occurred in the region of -1.01 to -1.06 V vs SCE.
Other reductions occurred at more negative potentials. It is
evident from Table 4 that the functionalization of fullerenes
with pyrrolidine moiety has rendered them slightly more difficult
to reduce than the pristine C60.
Under these reductive conditions, pulse irradiation of 1 and
2 (2 × 10-5 M) in a deaerated toluene/2-propanol/acetone
mixture (8:1:1 v/v) resulted in the formation of distinct and
characteristic absorption throughout the UV, visible and near-
IR region. The spectra of the reduced form of 1 and 2 are shown
in parts A and B of Figure 9, respectively. While the UV part
showed sharp bleaching around 320 nm, a range that corre-
sponds to a strong ground-state transition of the fullerene core,
the visible region is dominated mainly by maxima around 420
nm. More importantly, the differential absorption spectrum of
(C60•-) obtained upon pulse radiolysis of 1, for example, exhibits
a distinct maximum at 1005 nm. This characteristic fingerprint
•-
for reduced fullerene species is blue-shifted to pristine C60
(1080 nm), but in close resemblance to the spectral features
recorded for the π-radical anion of N-methylpyrrolidino-
fullerene,24 a structural analogue of 1. It is a general observation
that monofunctionalization leads to a significant blue-shift of
the corresponding radical anion bands, and thus, these features
are ascribed to the π-radical anion formed in the general
reaction.
(CH3)2 COH + C60-Xf (CH3)2CO + (C60•-)-X + H+
•
(9)
Pulse Radiolytic Formation of π-Radical Anions. Comple-
mentary to reductive triplet quenching, radical anions may be
produced via radical-induced reduction of the fullerenes. For
example, in an irradiated solvent mixture containing 80 vol %
toluene, 10 vol % 2-propanol, and 10 vol % acetone, the reduc-
ing radical is formed by hydrogen abstraction from 2-propanol.
Similarly, the π-radical anion absorption of 2 exhibits a sharp
maximum at 1010 nm, nearly identical with 1.
Conclusions
Pyrrolidinofullerenes exhibit singlet and triplet excited-state
properties that differ from those of pristine C60. The time-
resolved radiation and photochemical studies have revealed
informative details on the influence of functionalized groups
on the photophysical properties of functionalized C60. Com-
pared to N-phenyl-substituted pyrrolidinofullerene, the
N-methylanilinesfunctionalized fullerene dyad shows charge-
radiolysis
•
(CH3)2CHOH
8 (CH3)2 COH
(7)
The same species derives from acetone upon electron capture
and subsequent protonation.
(CH3)2CdO + esol- + H+ f (CH3)2 COH
(8)
•

5348 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Thomas et al.
(16) Hirsch, A. Synthesis 1995, 895.
transfer interactions in the ground and excited states, which were
dependent on the solvent polarity. The triplet quenchers such
as O2, TEMPO, and ferrocene interact with the triplet excited
states with bimolecular quenching rate constants of 5.3 × 108
to 7.7 × 109 M-1 s-1. The functionalization of fullerenes with
pyrrolidine moiety has rendered these pyrrolidinofullerenes
slightly more difficult to reduce than the pristine C60.
(17) Sun, Y. P.; Ma, B. Chem. Phys. Lett. 1995, 236, 285.
(18) Sension, R. J.; Szarka, A. Z.; Smith, G. R.; Hochstrasser, R. M.
Chem. Phys. Lett. 1991, 185, 179.
(19) Gevaert, M.; Kamat, P. V. J. Phys. Chem. 1992, 96, 9883.
(20) Williams, R. M.; Zwier, J. M.; Verhoeven, J. W. J. Am. Chem.
Soc. 1995, 117, 4093.
(21) Williams, R. M.; Koeberg, M.; Lawson, J. M.; An, Y.-Z.; Rubin,
Y.; Paddon-Row: M. N.; Verhoeven, J. W. J. Org. Chem. 1996, 61, 5055.
(22) Kamat, P. V.; Guldi, D. M.; Liu, D.; George Thomas, K.; Biju,
V.; S., D.; George, M. V. In Fullerenes; Kadish, K., Ruoff, R., Eds.; The
Electrochemical Socitey: Pennington, NJ, 1997; Vol. 4, p 122.
(23) Luo, C.; Huang, C.; Gan, L.; Zhou, D.; Xia, W.; Zhuang, Q.; Zhao,
Y.; Huang, Y. J. Phys. Chem. 1996, 100, 16685.
Acknowledgment. We thank Dr. Di Liu and Mr. Mark R.
Flumiani for their assistance in experimentation. P.V.K., D.M.G.,
and M.V.G. (in part) acknowledge the support from the Office
of Basic Energy Science, the U.S. Department of Energy and
K.G.T., V.B., and M.V.G. acknowldge the support from the
Council of Scientific and Industrial Research, Government of
India. One of the authors (M.V.G.) also thanks the Jawaharlal
Nehru Centre for Advanced Scientific Research, Bangalore, for
financial support. This is contribution No. NDRL 4018 from
the Radiation Laboratory and No. RRLT-PRU-86 from RRL,
Trivandrum.
(24) Guldi, D. M.; Maggini, M.; Scorrano, G.; Prato, M. J. Am. Chem.
Soc. 1997, 119, 974.
(25) Kuciavskas, D.; Lin, S.; Seely, G. R.; Moore, A. L.; Moore, T. A.;
Gust, D.; Drovetskaya, T.; Reed, C. A.; Boyd, P. D. W. J. Phys. Chem.
1996, 100, 1592.
(26) Imahori, H.; Hagiwara, K.; Aoki, M.; Akiyama, T.; Taniguchi, S.;
Okada, T.; Shirakawa, M.; Sakata, Y. J. Am. Chem. Soc. 1996, 118, 11771.
(27) Imahori, H.; Cardoso, S.; Tatman, D.; Lin, S.; Noss, L.; Seely, G.
R.; Sereno, L.; Chessa de Silber, J.; Moore, T. A.; Morre, A. L.; Gust, D.
Photochem. Photobiol. 1995, 62, 1009.
(28) Catalan, J.; Elguero, J. J. Am. Chem. Soc. 1993, 115, 9249.
(29) Taylor, T. W. J.; Owen, J. S.; Whitaker, D. J. Chem. Soc. 1938,
206.
(30) Ebbesen, T. W. ReV. Sci. Instrum. 1988, 59, 1307.
(31) Nagarajan, V.; Fessenden, R. W. J. Phys. Chem. 1985, 89, 2330.
(32) Sun, Y.-P.; Ma, B.; Lawson, G. E. Chem. Phys. Lett. 1995, 233,
57.
References and Notes
(1) Connolly, J. S.; Bolton, J. R. In Photoinduced Electron Transfer;
Fox, M. A., Channon, M., Eds.; Elsevier: Amsterdam, 1988; p 303.
(2) Kurreck, H.; Huber, M. Angew. Chem. Res. 1995, 34, 849.
(3) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res 1993, 26,
198.
(4) Arbogast, J. W.; Foote, C. S.; Kao, M. J. Am. Chem. Soc. 1992,
114, 2277.
(5) Foote, C. S. Top. Curr. Chem. 1994, 169, 347.
(6) Kamat, P. V.; Asmus, K.-D. Interface 1996, 5, 22.
(7) Kamat, P. V.; Guldi, D. In Fullerenes: Chemistry, Physics, and
New Directions; Ruoff, R., Kadish, K., Eds.; Electrochemical Society:
Pennington, NJ, 1996; Vol. VIII, p 254.
(8) Sun, Y.-P. In Molecular and Supramolecular Photochemistry;
Ramamurthy, V., Ed.; Marcel Dekker: New York:, 1997; Vol. 1.
(9) Zhou, F.; Jehoulelt, C.; Bard, A. J. J. Am. Chem. Soc. 1992, 114,
11004.
(10) Dubois, D. K.; Kadish, K. M.; Flanagan, S.; Haufler, R. E.;
Chibante, L. P. E.; Wilsson, L. F. J. Am. Chem. Soc. 1992, 114, 3978.
(11) Xie, P.; Perez-Cordero, E.; Echegoyen, L. J. J. Am. Chem. Soc.
1992, 114, 3978.
(12) Wang, Y. J. Phys. Chem. 1992, 96, 764.
(13) Park, J.; Kim, D.; Suh, Y. D.; Kim, S. K. J. Phys. Chem. 1994, 98,
12715.
(33) Ghosh, H. N.; Pal, H.; Sapre, A. V.; Mittal, J. P. J. Am. Chem.
Soc. 1993, 115, 11722.
(34) Ebbesen, T. W.; Tanigaki, K.; Kuroshima, S. Chem. Phys. Lett.
1991, 181, 501.
(35) Dimitrijevic, N. M.; Kamat, P. V. J. Phys. Chem. 1992, 96, 4811.
(36) Fraelich, M. R.; Weisman, R. B. J. Phys. Chem. 1993, 97, 11145.
(37) Biczok, L.; Linschitz, H.; Walter, R. I. Chem. Phys. Lett. 1992,
195, 339.
(38) Sauve, G.; Kamat, P. V.; Thomas, K. G.; Thomas, J.; Das, S.;
George, M. V. J. Phys. Chem. 1996, 100, 2117.
(39) Samanta, A.; Kamat, P. V. Chem. Phys. Lett. 1992, 199, 635.
(40) Guldi, D. M.; Hungerbuehler, H.; Janata, E.; Asmus, K. D. J. Phys.
Chem. 1993, 97, 11258.
(41) Kamat, P. V.; Sauve, G.; Guldi, D. M.; Asmus, K.-D. Res. Chem.
Intermed. 1997, 23, 575.
(42) Arbogast, J. W.; Darmanyan, A. P.; Foote, C. S.; Rubin, Y.;
Diederich, F. N.; Alvarez, M. M.; Anz, S. J.; Whetten, R. L. J. Phys. Chem.
1991, 95, 11.
(43) Guldi, D. M.; Maggini, M.; Scorrano, G.; Prato, M. Res. Chem.
Intermed. 1997, 23, 561.
(14) Sun, Y.-P.; Bunker, C. E.; Ma, B. J. Am. Chem. Soc. 1994, 116,
9692.
(15) Seshadri, R.; Rao, C. N. R.; Pal, H.; Mukherjee, T.; Mittal, J. P.
Chem. Phys. Lett. 1993, 205, 395.
(44) Dubois, D.; Moninot, G.; Kutner, W.; Jones, M. T.; Kadish, K. M.
J. Phys. Chem. 1992, 96, 7137.