C60 as a Photocatalyst of Electron-Transfer Processes: Reactions of Triplet C60 with Chloranil, Perylene, and Tritolylamine Studied by Flash Photolysis and FT-EPR

Article abstract of DOI:10.1021/jp960640h

Photoprocesses in benzonitrile solutions of C60 and chloranil (CA) have been studied by complementary techniques of nanosecond laser photolysis and Fourier transform EPR.Direct oxidation of ³C60 by CA is slow (k = (2.0 +/- 0.3)E7 M^-1 s^-1), consistent with the high oxidation potential of ³C60.However, the formation rate and yield of CA(1-) are much increased by addition of perylene (Pe) or tritolylamine (TTA) via the fast reactions ³C60 + Pe -> C60 + ³Pe, followed by ³Pe + CA -> Pe(1+) + Ca(1-), or ³C60 +TTA -> C60(1-) + TTA(1+), followed by C60(1-) + CA -> C60 + CA(1-).These reactions utilize the broad absorption and initial high triplet yield of C60, as well as the low oxidation potential of ³Pe or high reduction potential of ³C60, to catalyze efficient formation of CA(1-) and enhance separation of radicals.Triplet C60 also reacts with Pe by electron transfer, forming Pe(1+) and C60(1-) with rate one-third that of energy transfer.However, the Ca(1-) formed in the Pe-catalyzed reaction is strongly spin-polarized, indicating that it is formed primarily via the ³Pe pathway.The extinction coefficient of C60(1-) at 1080 nm is measured (ε = 18300 +/- 1100 M^-1 cm^-1) using the TTA reaction.

Full text of DOI:10.1021/jp960640h

8920
J. Phys. Chem. 1996, 100, 8920-8926
C60 as a Photocatalyst of Electron-Transfer Processes: Reactions of Triplet C60 with
Chloranil, Perylene, and Tritolylamine Studied by Flash Photolysis and FT-EPR
Carlos A. Steren,†,‡ Hans van Willigen,*,† Laszlo Biczo´k,§, Neeraj Gupta,§ and
Henry Linschitz*,§
Departments of Chemistry, UniVersity of Massachusetts at Boston, Boston, Massachusetts 02125;
Brandeis UniVersity, Waltham, Massachusetts 02254; and Central Research Institute for Chemistry,
Hungarian Academy of Sciences, P.O. Box 17, 1525, Budapest, Hungary
ReceiVed: February 29, 1996X
Photoprocesses in benzonitrile solutions of C60 and chloranil (CA) have been studied by complementary
3
techniques of nanosecond laser photolysis and Fourier transform EPR. Direct oxidation of C60 by CA is
3
slow (k ) (2.0 ( 0.3) × 107 M-1 s-1), consistent with the high oxidation potential of C60. However, the
formation rate and yield of CA- are much increased by addition of perylene (Pe) or tritolylamine (TTA) via
the fast reactions 3C60 + Pe f C60 + 3Pe, followed by 3Pe + CA f Pe+ + CA-, or 3C60 + TTA f C60
+
-
TTA+, followed by C60- + CA f C60 + CA-. These reactions utilize the broad absorption and initial high
triplet yield of C60, as well as the low oxidation potential of 3Pe or high reduction potential of 3C60, to catalyze
efficient formation of CA- and enhance separation of radicals. Triplet C60 also reacts with Pe by electron
transfer, forming Pe+ and C60- with rate one-third that of energy transfer. However, the CA- formed in the
3
Pe-catalyzed reaction is strongly spin-polarized, indicating that it is formed primarily via the Pe pathway.
The extinction coefficient of C60 at 1080 nm is measured (ꢀ ) 18 300 ( 1100 M-1 cm-1) using the TTA
-
reaction.
Introduction
pathway is illustrated by the C60/TTA/CA system. Reductive
quenching of the triplet by TTA, 3C60 + TTA f C60- + TTA+,
Triplet state C60 may react by reversible electron transfer with
either donors (arylamines,1,2 phenols2) or acceptors such as
tetracyanoquinodimethane (TCNQ),3 tetracyanoethylene
(TCNE),3-5 and chloranil (CA).6 We have previously studied
reductive reactions, including formation and decay of cage-
escape radical products, by complementary flash photolysis and
Fourier transform electron paramagnetic resonance (FT-EPR).7
These combined techniques are well-suited for such study since
the optical absorption spectrum of 3C608 is clearly separated from
that of C60 and its anion9,10 or cation10,11 radicals, and the unique
narrow-line EPR spectrum of the triplet7,12 is easily distinguished
from the highly resolved spectra of radical products. In addition
to information on chemical kinetics, the FT-EPR data provide
detailed information on spin dynamics.
yields C60 anion radicals, which then transfer the electron to
CA. A similar catalytic sequence involving C60- as intermediate
also plays a role, but to a lesser extent, in the C60/Pe/CA system,
leading to Pe+ and CA- radicals. The facts that light absorption
by C60 covers a broad range of the visible spectrum and radicals
formed from triplet precursors avoid immediate back-reaction
suggest its possible application in solar-driven photochemical
reactions.
In this work, we measure also the extinction coefficient of
-
the C60 anion radical and give rates and quantum yields of
reactions involved in these catalytic sequences.
Experimental Section
C60 (99.9%, SES Research), Pe (Gold label, Aldrich), and
benzonitrile (99.9% HPLC grade, Aldrich) were used as
received. CA (Aldrich) was purified by vacuum sublimation.
TTA was kindly donated by Professor R. I. Walter of the
University of IllinoissChicago. Studies were carried out on
freshly prepared solutions freed of oxygen by purging with
nitrogen or argon. Photochemical degradation over the course
of an experiment was found to be minimal and did not affect
the results.
Here we extend an earlier FT-EPR study of the C60/chloranil
(CA) system6 and demonstrate the function of C60 as a
photocatalyst for other electron-transfer reactions. It is shown
that in solutions of perylene (Pe) or tritolylamine (TTA) with
CA and C60, a fast and efficient electron-transfer reaction
between Pe or TTA and CA can be initiated by photoexcitation
of C60. In these reactions, C60 functions as a harvester of visible
light and an efficient channel to triplet states (φT > 0.9).8 In
3
Flash photolysis studies utilized a flash lamp driven dye laser
(Candela ED-200) giving 0.2 µs pulses at 590 nm (rhodamine
the case of C60/Pe/CA, triplet-triplet energy transfer, C60
+
Pe f C60 + 3Pe, leads to efficient production of perylene triplets
which are oxidized by CA to yield Pe+ and CA-. That this is
the primary route of CA- formation is directly demonstrated
by flash photolysis and is also deduced from the spin polariza-
tion (CIDEP) of the anion radicals. An alternative catalytic
6G). Initial triplet absorbance on flashing [C60] ∼ 7 × 10-5
M
in benzonitrile (∆D750 ∼ 0.6) corresponded to [3C60]0 ∼ 4 ×
10-5 M. Transient absorption spectra were followed by
conventional means,13 using a photomultiplier to 800 nm and
an InGaAs photodiode and fast amplifier (New Focus, 1811)
for the near-IR. Quenching rate constants were determined from
the linear dependence of pseudo-first-order decay constants on
quencher concentration. In averaging transient absorbances, ∆D
values measured on each flash were corrected for flash energy
variations by plotting ∆D’s vs energy and reading off the
absorbance at a given flash energy.
† University of Massachusetts at Boston.
‡ Present address: FAMAF, Universidad Nacional de Co´rdoba, 5000
Co´rdoba, Argentina.
§ Brandeis University.
CRIC, Hungarian Academy of Sciences.
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, May 1, 1996.
S0022-3654(96)00640-5 CCC: $12.00 © 1996 American Chemical Society

C60 as a Photocatalyst of Electron-Transfer Processes
J. Phys. Chem., Vol. 100, No. 21, 1996 8921
Reversible polarographic potentials of perylene (oxidation)
and C60 (reduction) were measured in the same benzonitrile
solution vs SCE, using an E.G.G. “Versastat” potentiostat, glassy
carbon working electrode, and 0.1 M tetrabutylammonium
hexafluorophosphate (TBAHP) electrolyte.
FT-EPR measurements were performed with a home-built
spectrometer.14 The nominal microwave pulse width used in
the measurements was 15 ns, and the free induction decay (FID)
was sampled, with quadrature detection, at 200 Msamples/s. A
CYCLOPS phase-cycling routine was used in the measurements.
Final spectra typically were the average of 2400 FIDs. C60 was
excited with the second harmonic (532 nm) of a Quanta Ray
GCR12 Nd:YAG laser (∼20 mJ/pulse, 10 Hz repetition rate).
Measurements in which 3Pe was generated by direct excitation
(440 nm, ∼ 2 mJ/pulse, 30 Hz) were performed using a dye
laser (Lambda Physik FL3001) pumped by an excimer laser
(Lambda Physik 103MSC). The time evolution of spectra was
monitored by recording FIDs for a series of delay times (τd),
ranging from 10 ns to 200 µs, between laser and microwave
pulses. Signal intensities were obtained from the FIDs with
the linear prediction-singular value decomposition technique15
or by integration of the peaks in the frequency-domain spectra.
Figure 1. Second-order decay of radical ions produced by excitation
of C60 (6 × 10-5 M) in the presence of tritolylamine (1.5 × 10-4 M)
in benzonitrile at room temperature. ∆D-1 versus time at 680 nm
(TTA+) and 1080 nm (C60-).
are made at different wavelengths. However, we feel that the
direct measurement of ∆D’s is a more reliable procedure and
take 18 300 ( 1100 M-1 cm-1 to be the preferred value. This
may be compared with previously given values (M-1 cm-1) from
electrochemical preparation of C60- in the presence of electro-
lyte: 1.2 × 104 in benzonitrile,9a 2.0 × 104 in tetrahydrofuran,9b
and 1.55 × 104 in methylene chloride.9c Values estimated from
γ-radiolysis are in the range (3 ( 1) × 104.10
Results and Discussion
Extinction Coefficient of C60-. To obtain quantum yields
of 3C60 reductions, we have previously measured the extinction
3
Reactions of C60 with Chloranil. Flash Photolysis. In
-
coefficients of C60 at its relatively weak absorptions in the
3
benzonitrile solution CA quenches C60 with rate constant kCA
visible.2,16,17 In connection with current and ongoing studies,
it is desirable to determine the extinction also for the strong
peak at 1080 nm where there is less overlap with other transient
spectra but for which only approximate and conflicting values
are presently available.9,10 We use, as before, the reaction 3C60
) (2.0 ( 0.3) × 107 M-1 s-1. However, in this solvent, and
+
on the 50 ns time scale, the expected C60 cation radical
absorption at 980 nm10,11 is not seen. Moreover, any optical
absorption by the presumed CA anion radical around 450 nm18
is obscured by the strongly overlapping ground state absorption
of CA itself. Thus, the detailed nature of this reaction in
benzonitrile cannot be determined simply by flash photolysis.
The absence of C60+ absorbance here may be related to the case
+ Am f C60 + Am+, where Am represents the efficient
-
quenchers tritolylamine (TTA) or tri-p-fluorophenylamine (TFPA),
whose cation radicals have well-established extinctions,
ꢀ(TTA+)680 ) 26 200 M-1 cm-1 and ꢀ(TFPA+)645 ) 26 600
3
of C60 quenching by TCNQ, in benzonitrile, in which again
M-1 cm-1 in benzonitrile.2 The extinction of C60 is then
-
neither C60+ nor TCNQ- absorptions are seen.3 This has been
attributed to formation only of a weakly polar triplet exciplex
absorbing at shorter wavelength.3
obtained by comparing radical absorptions at 1080 and 680 (or
645) nm, measured on the same solutions at equal flash energies
and corrected to take account of any overlapping radical
formation and decay.13 Absorbances measured at the amine
Time-ResolVed Magnetic Resonance (FT-EPR). More de-
tailed information is given by FT-EPR observation of both 3C60
and CA- radicals. Figure 2a shows FT-EPR spectra from a
solution of C60 (0.2 mM) with CA (9.9 mM) in benzonitrile for
a series of delay times (10 ns-200 µs) between photoexcitation
of C60 and a π/2 microwave pulse. In addition to the resonance
-
cation peaks were corrected for small contributions from C60
at these wavelengths (ꢀ645 ∼ 1700 M-1 cm-1; ꢀ680 ∼ 900 M-1
cm-1). There is no ground state absorption (i.e., bleaching)
-
beyond 630 nm. The values of ꢀ(C60
)
1080 thus found are 18 300
and 18 200 M-1 cm-1 ((5%), based on TTA and TFPA,
respectively.
3
peak from C60 (g ) 2.0012)7,12 observed at early times, the
spectra show a signal with g ) 2.0057 assigned to the chloranil
anion (CA-). We confirmed the assignment by comparison with
the spectrum from CA- prepared by in situ electrolysis.
The time dependence of the intensities of the resonance peaks
The C60- extinction coefficient was obtained also from kinetic
data. Assuming a second-order back-reaction with rate constant
k2 for radical decay,13 the slope, Sλ, of a plot of ∆Dλ-1 vs time
following triplet disappearance is k2/∑∆ꢀλ, where ∑∆ꢀλ is the
total differential extinction coefficient at wavelength λ. Thus,
in the C60/TTA system,
3
is displayed in Figure 2b. The signal from C60 increases in
intensity for 0 e τd e 1 µs as the triplet spin system relaxes to
thermal equilibrium.7,19 The signal from CA- initially is in
emission and becomes absorptive for τd > 5 µs. The emissive
signal can be attributed to spin polarization generated in the
transient radical pair [C60+‚‚‚CA-] by the radical pair chemically
induced dynamic electron polarization mechanism (RPM CI-
DEP20). Spin-lattice relaxation turns the signal from emission
into absorption at longer delay times. As in the optical transient
ꢀ1080 ) [S680/S1080
]
ꢀ
680
∑
Measured ∆D-1 vs time slopes of TTA+ and C60 peaks are
plotted in Figure 1. Using the extinction values of TTA+ and
C60- at 680 nm given above, the second-order plot in Figure 1
gives k2 ) (1.0 ( 0.1) × 1010 M-1 s-1, in close agreement
-
+
absorption measurements, no signal corresponding to C60 is
with values obtained previously using data at other wavelengths.2
detected, nor is C60- detected with FT-EPR in reductive triplet
quenching reactions.7 In both cases this is perhaps due to a
short T2.
-
From this we then obtain ꢀ(C60
)
) 19 500 M-1 cm-1, in
1080
fair agreement with the result given above. This kinetic method
is not subject to errors involving mismatch of initial triplet
concentration on successive flashes where ∆D measurements
The time development of the intensity of the resonances is
similar to that found in the study of reductive quenching of C60

8922 J. Phys. Chem., Vol. 100, No. 21, 1996
Steren et al.
CA- can be derived from the magnetic resonance data since
signal intensities are a measure of spin concentrations. The
3
analysis of the time profiles of the C60 and CA- resonances
makes it possible to calculate intensities that would obtain at
thermal equilibrium in the absence of decay processes. If the
CA- concentration equaled that of the triplet precursor, the ratio
of CA- signal intensity over 3C60 signal intensity at Boltzmann
equilibrium would be 3/8, the ratio of magnetic moments.
Experimentally, we find a ratio of ∼0.076 so that only 20 (
10% of the quenching events yield cage escape products. The
uncertainty in this result is relatively large because of propaga-
tion of errors in fitting parameters. The rather low yield of
cage escape products may be due to spin-orbit coupling in the
chlorinated quinone anion radical.23 It may be partly responsible
for the failure to detect the C60+ cation with flash photolysis or
FT-EPR.
∆G0 of the photoinduced electron-transfer reaction in ben-
zonitrile (dielectric constant 26) can be estimated using the
Rehm-Weller equation24 given that the triplet energy (ET) of
C60 is 1.57 eV,25-27 Eox of C60 is 1.76 V (SCE, in benzoni-
0
0
trile),28 Ered of CA is 0.07 V (SCE, in benzonitrile),29 and
assuming an acceptor-donor distance of ∼7 Å.1,2 With these
values ∆G0 ∼ 0.04 V. The low value of the electron-transfer
rate constant (2 × 107 M-1 s-1) derived from both flash and
FT-EPR data is consistent with a very small thermodynamic
driving force. The result may be compared with published data
on reductive quenching. For the reaction of 3C60 with tritolyl-
amine in benzonitrile (∆G0 ∼ -0.36 V) the rate constant is 3.5
× 109 M-1 s-1 2,7
,
and with diethylamine (∆G0 ∼ -0.17 V)
Figure 2. (a) Time evolution of the FT-EPR spectrum given by a
solution of C60 (2 × 10-4 M) with CA (9.9 × 10-3 M) in benzonitrile
following pulsed laser excitation of C60. (b) Time profiles of the
Arbogast et al.1 report 4.2 × 107 M-1 s-1
.
3
The rate constant of C60 quenching by CA in toluene,
determined by measuring the rate of decay of the FT-EPR signal
from the triplet, is 0.6 × 107 M-1 s-1. In this nonpolar solvent
FT-EPR spectra only show the resonance peak from 3C60. There
is no evidence of cage escape product.
3
intensities of the C60 and CA- resonance peaks. The solid lines
represent results of the data analysis described in the text.
triplets by hydroquinone and tritolylamine.7 As discussed
previously,7,19 the time profile of the 3C60 signal is determined
by the spin polarization (P) with which the triplets are born,
Flash Photolysis Study of Reactions of 3C60 with Perylene.
3
Kinetics. In benzonitrile solution Pe quenches C60 with rate
T
constant (1.4 ( 0.1) × 109 M-1 s-1, obtained from the decay
of the 750 nm triplet absorption. The transient spectrum given
in Figure 3 shows peaks at 1080, 555, and 485 nm which are
the spin-lattice relaxation rate (k1 ), and the pseudo-first-order
rate of triplet decay (kd). A small second-order contribution to
the triplet decay19,21 is ignored. The solid line in Figure 2b
3
-
gives the last-squares fit of the C60 data based on this model.
unambiguously assigned respectively to the ion radicals, C60
The analysis7,19 gives the following data: k1T ) 2.7 × 106 s-1
and Pe+,30 and to the triplet, 3Pe.31,32 Thus, both electron- and
energy-transfer reactions contribute to the quenching process:
(T1 ) 0.37 µs), kd ) (2.0 ( 0.3) × 105 s-1, and P (given as
T
T
fraction of Boltzmann polarization) is 0.21. The values of k1
and P are in good agreement with previous results.7,19 The
decay rate is given by kd ) k0 + kq [CA], where k0 (0.4 × 105
s-1)7 is the rate in the absence of quencher and kq is the rate
constant of electron-transfer quenching. Since [CA] ) 9.9 mM,
3C60 + Pe f C60- + Pe+
3C60 + Pe f C60 + 3Pe
(1)
(2)
the triplet signal decay gives kq ) (1.6 ( 0.3) × 107 M-1 s-1
,
Reaction 1 can be followed most clearly at the absorption peak
of C60-. Assuming pseudo-first-order radical formation and
second-order radical decay (kd),13 we fit the 1080 nm transient
in good agreement with the flash photolysis result.
The time evolution of the CA- signal is described by a
model22 that takes into account rates of radical formation and
decay as well as generation and decay of RPM CIDEP. The
least-squares fit of the CA- data, shown by the solid line in
Figure 2b, gives a pseudo-first-order rate constant of radical
formation of kf ) (2.0 ( 0.3) × 105 s-1 and a spin-lattice
relaxation rate of 0.19 × 106 s-1 (T1D ) 5.3 µs). This value of
profile with kq ()k1 + k2) ) (1.1 ( 0.3) × 109 M-1 s-1, in fair
-
agreement with the measurement at 750 nm, and kd/ꢀ(C60
)
1080
) (4.2 ( 0.5) × 105 cm s-1. With ꢀ(C60-) ) 18 300 ( 1100
M-1 cm-1, determined above, we obtain kd ) (7.7 ( 1.2) ×
109 M-1 s-1 for the bulk radical recombination rate, very close
to the diffusion-controlled value and comparable to that found2
D
T1 is in agreement with that given by a direct relaxation
for reaction of C60 with TTA+.
-
measurement using the standard inversion-recovery method. The
measured rates of 3C60 quenching and CA- formation are closely
matched. It can be concluded that the FT-EPR data establish
that C60 triplets are quenched oxidatively by CA giving CA-
free radicals.
Quantum Yields. Nonlinear least-squares analysis of the
transient profiles gives also the values of product absorbances
(∆D∞λ ) corrected for decay during the formation period and
corresponding to total product yields.13 Quantum yields referred
to the triplet are then given by
Triplet quenching is expected to favor cage escape of products
because of spin restriction on back electron transfer during the
φ ) [∆D∞λ (prod)/∆D7050][ꢀ(3C60)750/ꢀ(prod)λ]
3
lifetime of the [C60+‚‚‚CA-] radical pair. The yield of bulk

C60 as a Photocatalyst of Electron-Transfer Processes
J. Phys. Chem., Vol. 100, No. 21, 1996 8923
TABLE 2
rate constants (M-1 s-1 a
)
reactions
flash data
FT-EPR data ∆G0 (V)
3C60 + CA f C60+ + CA-
3C60 + Pe f C60- + Pe+
3C60 + Pe f C60 + 3Pe
2.0(0.3) × 107
3.6(0.4) × 108
1.1(0.2) × 109
1.6(0.2) × 107
0.04
-0.13
0
1.3(0.2) × 109 b
3C60 + TTA f C60- + TTA+ 3.5(0.3) × 109 c 3.1(0.3) × 109 d -0.36
3Pe + CA f Pe+ + CA-
C60- + CA f C60 + CA-
C60- + Pe+ f C60 + Pe
CA- + Pe+ f CA + Pe
5.6(0.3) × 109 e -0.5
2.5(1.0) × 109 f 4.4(0.3) × 109 g -0.4
1.2(0.2) × 1010 h
1.3(0.2) × 1010 j
-1.52i
-1.14i
a Solvent benzonitrile, at room temperature. b Overall (energy and
3
electron transfer) C60 quenching rate constant derived from [Pe]
3
dependence of C60 resonance intensity. c From ref 2. d From ref 7.
e From analysis of time profile of CA- signal intensity, Pe directly
Figure 3. Time-resolved absorption spectra from a solution of C60
(∼5 × 10-5 M) and perylene (5 × 10-4 M) in benzonitrile immediately
(b) and 3 µs (2) after flash.
excited with 440 nm laser light. f From effect of CA on decay of C60
-
3
under conditions where reductive quenching of C60 by TTA is fast
compared to electron transfer from C60- to CA. g From kinetics of CA-
generation upon excitation of C60 in the presence of TTA and CA,
conditions as in f. h From second-order decay of C60- monitored at 1080
nm. i In benzonitrile with 0.1 M TBAHP. j From second-order decay
of Pe+ (cf. Figure 4C).
TABLE 1: Data for Determination of Electron and Energy
3
Transfer Yields of C60 + Pe
wavelength (nm)
transient absorbance
485
555
750
1080
0.302a
0.303a
0.394b
0.122a
extinction coefficients, M-1 cm-1
assignm
3C60
16000c
2500d
3800d
1900g
18300e
-
C60
3Pe
13400f
5000h
Pe+
50000h
a ∆D∞λ , total product absorbance.13 b ∆D0, initial triplet absorbance.
c Reference 2. d Reference 2. e This work. f Reference 31. g Reference
32. h Reference 30.
0
3
where ∆D750 is the initial absorbance of C60. Since φT(3C60)
is close to unity,8 these quantum yields may be taken to be
essentially overall values.
Table 1 gives relevant data, including extinction coefficients
at triplet and radical peaks, and average values of the ∆D’s
normalized to constant flash energy. The quantum yield of the
electron transfer (LT) process, reaction 1, is determined directly
Figure 4. Flash transient time profiles of benzonitrile solutions of C60
(3 × 10-4 M) with Pe (6 × 10-4 M): A, At 750 nm (3C60); B, at 555
nm (mainly Pe+); C, at 555 nm after addition of CA (6 × 10-3 M).
Insert: A, 485 nm (mainly 3Pe), solution as in A, B above; B, 485 nm
solution as in C above.
from the initial 3C60 absorption (∆D0 ) and total C60- absorp-
750
tion (∆D01080),13 giving φLT ) 0.26 ( 0.03. The quantum yield
for the energy transfer (NT) process, reaction 2, was obtained
∞
3
from ∆D485, corresponding mainly to Pe absorption. Correc-
tions were made for absorption contributions at this wavelength
(see Table 1) by C60-, derived from the 1080 nm measurement,
and Pe+, which was taken to be stoichiometrically equivalent
below the diffusion-controlled value, matches that found for
similar cases35 and is consistent with the relation kNT ) kdif/[1
+ exp(-∆ET/RT)].35 The oxidation potential of Pe in benzo-
nitrile measured relative to the reduction potential of C60 in the
same solution is 1.52 V, directly giving ∆G0 for the ground
state electron-transfer reaction. Taking the interradical distance
as ∼7 Å,1,2 we then obtain ∆GL0T ∼ -0.13 eV for the triplet
quenching process.24 Since a larger reorganization energy
barrier is expected for electron transfer compared with energy
transfer, the slower electron-transfer rate despite the more
favorable ∆G0 is reasonable.
-
to C60-. Corrections for C60 bleaching due to C60 formation
were negligible. This gave φNT ) 0.76 ( 0.15, in which the
large uncertainty includes conservatively estimated errors in ꢀ’s
as well as ∆D’s. Finally, φLT was obtained also from the Pe+
peak at 555 nm, corrected again for overlapping absorbances
-
3
of C60 and Pe. This method gave φLT ) 0.20 ( 0.07, in
reasonable agreement with the 1080 nm based value. Within
the considerable uncertainty of these results, it thus appears that
the total yield of energy transfer and bulk radical formation in
the 3C60 + Pe reaction is not far from unity. The result suggests
that cage escape yield of radicals ions formed by reductive
quenching of 3C60 in this case is higher than that found (∼0.3-
3
Electron Transfer Mediated by C60. Flash Photolysis.
Addition of chloranil (CA) to the C60/Pe system at relative
3
concentrations such that primary quenching of C60 occurs
almost exclusively by reaction with Pe (see rate constants, Table
2) leads to the transient time profiles shown in Figure 4. We
see (Figure 4B,C) a marked increase in the peak at 555 nm
(Pe+) and a decrease in the absorbance at 485 nm due to
replacement of 3Pe by weakly absorbing Pe+ and CA- (Figure
3
0.4) in other C60 charge-transfer quenching reactions.2,33
3
From the rate constant of overall quenching of C60 by Pe,
(1.4 ( 0.1) × 109 M-1 s-1, and the quantum yields, we obtain
kNT ) (1.1 ( 0.2) × 109 and kLT ) (3.6 ( 0.4) × 108 M-1 s-1
for the rate constants of energy and electron transfer, respec-
tively. These rates may be compared with the ∆G0 values of
-
4, insert). CA also causes more rapid disappearance of the C60
absorption at 1080 nm; in fact, with the conditions and time
3
3
-
reactions 1 and 2. The triplet energies of C60 and Pe are
36.326,27 and 36.1 kcal/mol,34 respectively, so that the energy
transfer process is essentially thermoneutral. Its rate, slightly
scale that apply for Figure 4, no C60 is seen at all. An
associated formation of CA- is demonstrated by EPR measure-
ments described below. These results demonstrate the following

8924 J. Phys. Chem., Vol. 100, No. 21, 1996
Steren et al.
3
Figure 6. Rate of C60 resonance peak decay as function of perylene
concentration.
(weak) resonances due to Pe+ are observed. (c) The intensity
3
of the CA- resonance has increased relative to the C60
resonance.
In interpreting these results, we note first that the flash
3
photolysis measurement of C60 quenching by Pe is quantita-
tively confirmed by the dependence of the decay rate of the
3C60 resonance on [Pe] (Figure 6) which gives an overall
quenching rate constant of (1.3 ( 0.2) × 109 M-1 s-1, in good
agreement with the flash photolysis value. Furthermore, the
observation of resonance peaks due to Pe+ in the FT-EPR
spectra is in accord with the photolysis finding that C60
quenching by Pe occurs, at least in part, via the electron-transfer
route (reaction 1).
Figure 5. (a) Time evolution of the FT-EPR spectrum given by a
solution of C60 (2.3 × 10-4 M), perylene (6.3 × 10-4 M), and CA (5.3
× 10-3 M) in benzonitrile following pulsed laser excitation of C60. (b)
3
3
Time profiles of the intensities of the C60 and CA- resonance peaks.
The solid lines represent results of the data analysis described in the
text.
Given the two mechanisms of 3C60 quenching by Pe identified
by flash photolysis, CA- can be formed by (I) subsequent
electron transfer from C60- (reaction 3, ∆G0 ∼ -0.4 V) and/or
secondary processes,
3
(II) oxidative quenching of Pe (reaction 4, ∆G0 ∼ -0.5 V).
From the fact that CA- gives an emissive resonance at early
times, we conclude that the dominant signal contribution is made
k3
9
C60- + CA
8 C60 + CA-
(3)
(4)
3
by the Pe quenching route (II). This conclusion is based on
k4
the following considerations. The spin polarization with which
CA- is born if formed via (II) can contain a triplet mechanism
(TM) and a radical pair mechanism (RPM) component.20 Since
3Pe + CA
9
8 Pe+ + CA-
in which C60- and 3Pe are formed in the primary 3C60 quenching
reactions 1 and 2, respectively. The rate constant k3, derived
from the decay of 1080 nm absorbance as function of [CA], is
(2.5 ( 1) × 109 M-1 s-1. The final decay of 555 nm absorbance
3
3Pe is formed by energy transfer from C60, which initially
carries little or no polarization,7,19 the CA- TM signal contribu-
tion must be absorptive but small since it cannot exceed the
(triplet) thermal equilibrium value. On the other hand, spin state
evolution in the radical pair [Pe+‚‚‚CA-] gives rise to an
emissive signal contribution because g(CA-)(2.0057) > g(Pe+)-
(2.002 57).20 The magnitude of the RPM polarization is
expected22 to far exceed that given by the TM so that CIDEP
will give rise to an emissive resonance peak. For the alternative
reaction path, reductive quenching of 3C60 followed by reduction
of CA by the C60 anion, the spin polarization of CA- is that
carried over from C60-, and the TM contribution is absorptive
for the reason given above. The RPM contribution is absorptive
as well in this case since g(C60-) < g(Pe+).20 Therefore, this
route cannot account for the observed CIDEP effect.
is second order with rate constant (1.3 ( 0.2) × 1010 M-1 s-1
.
Under the conditions of Figure 4C, with almost total trapping
of 3C60 by Pe and of 3Pe by CA, ∆D5∞55 is due to Pe+ alone, and
we obtain the overall yield of Pe+, φrad ) 0.62 ( 0.04. The
3
contribution from the electron-transfer quenching of C60 by
perylene to φrad is ∼0.26 (vide supra). The remainder is formed
via the alternative route, energy transfer (2) followed by
3
oxidative quenching of Pe by CA (4). The value obtained
earlier for the quantum yield of (2) φNT ) 0.76 ( 0.15 so that
the yield of (4) becomes 0.47 ( 0.10. Assuming reasonably
that the unidirectional electron transfer in (3) occurs with unit
efficiency, the flash photolysis data show that the overall yield
of CA- also must be ∼0.62. FT-EPR measurements, which
permit following the formation and decay of CA- directly,
support and extend these findings.
FT-EPR Results. Figure 5 shows a series of FT-EPR spectra
given by a benzonitrile solution containing C60 (0.23 mM), Pe
(0.63 mM), and CA (5.3 mM) excited with 532 nm laser pulses.
A number of features distinguish these spectra from those given
by C60/CA (cf. Figure 2). (a) Both the rates of 3C60 decay and
CA- formation have increased by the introduction of Pe even
though the CA concentration has been reduced. (b) Additional
The TM contribution is expected to be small for both reaction
paths so that initial signal intensities will depend predominantly
on RPM CIDEP. On the basis of the g values of the free
radicals forming the radical pairs that are involved in CA-
formation, [C60-‚‚‚Pe+] and [CA-‚‚‚Pe+], it can be concluded
that RPM spin polarization generated by the pairs must be of
similar magnitude. However, the contribution to the CA-
3
resonance by reductive quenching of C60 by Pe followed by
-
electron transfer from C60 to CA is expected to be strongly
attenuated by the short spin-lattice relaxation time of the C60

C60 as a Photocatalyst of Electron-Transfer Processes
J. Phys. Chem., Vol. 100, No. 21, 1996 8925
anion.9b This combined with the flash photolysis finding that
3
energy transfer much exceeds electron transfer in the C60
+
Pe reaction leads to the conclusion that (I) makes only a minor
contribution to the CA- resonance for short delay times (i.e.,
before thermal equilibrium has been established).
The time profiles of the intensities of the 3C60 and CA- signals
were analyzed quantitatively in order to derive values of the
rate constants of the processes involved in the formation of the
3
anion radicals. The analysis of the time evolution of the C60
peak, as described earlier,7,19 gives the value of the overall triplet
quenching rate constant. The generation of CA- involves
contributions from routes I and II. An analysis of the time
profile of the CA- signal intensity in terms of a model that
takes both these contributions into account is not feasible
because of the large number of parameters involved. It is
known, however, that (II) is the dominant path and in addition
gives enhanced signal intensity because of spin polarization (vide
supra). Therefore, a treatment based on the assumption that
(I) makes a negligible contribution is expected to give reasonable
results that can serve to establish that CA- formation is a two-
step process. The following conditions were applied in the
3
numerical analysis of the data: (a) C60 quenching by Pe
(reaction 2) is a pseudo-first-order process ([Pe] . [3C60]) and
mainly determines the rate (kNT) of decay of C60 triplets, (b)
similarly, since [CA] . [3Pe], CA- formation by electron
transfer from 3Pe (reaction 4) is pseudo-first-order (rate constant
kLT), (c) first-order chemical decay of CA- (rate constant kd).
Actually, decay involves back electron transfer and is a second-
order process. The error introduced by this simplification is
not expected to affect values of rate constants that control time
evolution during the first 10-20 µs, which are of primary
interest here.
Figure 7. Time profiles of the intensities of the 3C60 and CA- resonance
peaks given by solutions of C60 (2 × 10-4 M), Pe, and CA in
benzonitrile: (a) Pe limiting, [Pe] ) 4 × 10-4 M, [CA] ) 4.1 × 10-3
M; (b) CA limiting, [Pe] ) 2 × 10-3 M, [CA] ) 3.3 × 10-4 M. The
solid lines represent results of the data analysis described in the text.
EPR signal intensity is proportional to the difference in
population (n) of â and R spin states in CA-. Given the
approximations outlined above, the time evolution of n is given
by
d
dt
D
[n] ) RkNT[3Pe] + k1 (nB - n) - kdn
(5)
In this equation, R represents CIDEP (TM and RPM) generating
processes, nB ) 0.75 × 10-3[CA-] is the thermal equilibrium
population difference, and k1D is the CA- spin-lattice relaxation
rate. A closed-form expression for the time evolution of n is
readily derived given the proposed reaction scheme and specified
conditions. Least-squares fitting of the experimental data to
the theoretical expression is aided by the fact that values of
rates kNT and k1D are given by the time development of the 3C60
resonance and relaxation time measurements, respectively.
According to the model, by adjusting the relative concentrations
of Pe and CA, triplet-triplet energy transfer (2) or electron
transfer (4) can be made the rate-limiting step in CA- formation.
Time profiles of the resonances for these limiting conditions
are shown in Figure 7. That 3C60 quenching can be decoupled
from CA- formation by adjusting the concentrations so that
Figure 8. Time profile of the CA- resonance peak given by a
benzonitrile solution of C60 (2 × 10-4 M), tritolylamine (1.9 × 10-3
M), and CA (3.6 × 10-4 M) following laser excitation of C60.
The C60/TTA/CA System. The kinetics and spin dynamics
of reaction 3 are conveniently studied with the C60/TTA/CA
system. Flash photolysis2 and FT-EPR7 measurements show
3
that C60 is reductively quenched by TTA at near diffusion-
3
quenching of Pe is the rate-limiting step is evident in Figure
controlled rate. Therefore, in the C60/TTA/CA (solvent ben-
7b. The solid lines in Figures 5 and 7 give the calculated CA-
signal intensity profiles. Calculated and experimental data are
in close agreement, providing additional support for the proposed
mechanism. From the relative intensities of the 3C60 and CA-
signals we calculate that 51 ( 15% of the triplets give rise to
anion radicals. This result, within the uncertainty of both
determinations, is comparable to the overall yield, Φrad ) 0.62
( 0.04, determined by flash photolysis (cf. Figure 4C).
Therefore, compared to the direct oxidative quenching of C60
triplets by CA, the yield of free radicals has more than doubled.
zonitrile) system, photoexcitation of C60 generates C60 and
-
TTA+ in the initial step. In a followup reaction CA- is formed
by electron transfer from C60-. Figure 8 shows that in this case
the signal from CA- is in absorption for all delay times in
agreement with the CIDEP pattern predicted for this reaction
sequence. In the initial triplet quenching step the C60 anion,
which is not observed in the FT-EPR spectrum probably because
of a short T2, is formed with positive spin polarization stemming
from TM and RPM CIDEP (g(C60-) < g(TTA+)36). In the
subsequent electron-transfer step, C60 + CA f C60 + CA-,
-

8926 J. Phys. Chem., Vol. 100, No. 21, 1996
Steren et al.
Materials; Kadish, K. M., Ruoff, R. S., Eds.; Electrochemical Society:
Pennington, NJ, 1994; p 818.
(7) Steren, C. A.; Levstein, P. R.; van Willigen, H.; Linschitz, H.;
Biczo´k, L. Chem. Phys. Lett. 1993, 204, 23.
any spin polarization of the precursor is transferred to the
quinone. Therefore, the CA- resonance is expected to be in
absorption.
The time profile of the intensity of the CA- resonance was
analyzed in terms of formation and decay steps taking spin
polarization into account.7. The least-squares fit to the experi-
mental data is given by the solid line in Figure 8.
Table 2 summarizes rate constants and yields measured in
this work.
(8) Reference 2 and references therein.
(9) (a) Lawson, D. R.; Feldheim, D. L.; Foss, C. A.; Dorhout, P. K.;
Elliott, C. M.; Martin, C. R.; Parkinson, B. J. Electrochem. Soc. 1992, 139,
268. (b) Schell-Sorokin, A. J.; Mehran, F.; Eaton, G. R.; Eaton, S. S.;
Viebeck, A.; O’Toole, T. R.; Brown, C. A. Chem. Phys. Lett. 1991, 195,
225. (c) Kato, T.; Kodama, T.; Shida, T. Chem. Phys. Lett. 1993, 205,
405.
(10) Kato, T.; Kodama, T.; Shida, T.; Nakagawa, T.; Matsui, Y.; Suzuki,
S.; Shiromaru, H.; Yamauchi, K.; Achiba, Y. Chem. Phys. Lett. 1991, 180,
446.
Conclusions
(11) Nonell, S.; Arbogast, J. W.; Foote, C. S. J. Phys. Chem. 1992, 96,
4169.
Flash photolysis and FT-EPR studies show that C60 triplets
are quenched by chloranil. FT-EPR spectra establish that
encounters between C60 and CA in benzonitrile leads to
(12) Closs, G. L.; Gautam, P.; Zhang, D.; Krusic, P. J.; Hill, S. A.;
Wasserman, E. J. Phys. Chem. 1992, 96, 5228. Regev, A.; Gamliel, D.;
Meiklyar, V.; Michaeli, S.; Levanon, H. J. Phys. Chem. 1993, 97, 3671.
(13) Andrews, L. J.; Levy, J. M.; Linschitz, H. J. Photochem. 1976-
77, 6, 335.
(14) Levstein, P. R.; van Willigen, H. J. Chem. Phys. 1991, 95, 900.
(15) de Beer, R.; van Ormondt, D. In AdVanced EPR: Applications in
Biology and Biochemistry; Hoff, A. J., Ed.; Elsevier: Amsterdam, 1989;
Chapter 4.
3
formation of CA- free radicals. The rate constant of the
electron-transfer reaction (cf. Table 2) is some 2 orders below
the diffusion-controlled limit which can be attributed to the small
thermodynamic driving force. A comparison of 3C60 and CA-
resonance peaks shows that only about 20% of the quenching
events give cage escape products.
(16) Dubois, D.; Kadish, K. M.; Flanagan, S.; Haufler, R. E.; Chibante,
L. P. S.; Wilson, L. J. J. Am. Chem. Soc. 1991, 113, 4364.
(17) Kamat, P. V. J. Am. Chem. Soc. 1991, 113, 9705.
(18) Andre, J. J.; Weil, G. Mol. Phys. 1968, 15, 97.
(19) Steren, C. A.; van Willigen, H.; Dinse, K.-P. J. Phys. Chem. 1994,
98, 7464.
(20) McLauchlan, K. A. In Modern Pulsed and Continuous-WaVe
Electron Spin Resonance; Kevan, L., Bowman, M. K., Eds.; Wiley: New
York, 1990.
(21) Dimitrijevic´, N. M.; Kamat, P. V. J. Phys. Chem. 1992, 96, 4811.
(22) Ebersole, M. H.; Levstein, P. R.; van Willigen, H. J. Phys. Chem.
1992, 96, 9311.
(23) Winter, G.; Steiner, U. Ber. Bunsen-Ges. Phys. Chem. 1980, 84,
1203 and references therein.
In benzonitrile solutions of CA with Pe (or TTA) and C60,
excitation with visible light initiates a fast and efficient reduction
of CA. In the presence of Pe, the dominant reaction path is
3
3
formation of Pe by energy transfer from C60 followed by
oxidative quenching of 3Pe by CA. In this process the yield of
CA- compared to that of the direct excited state electron transfer
from C60 to CA is more than doubled. With the amine, reductive
3
quenching of C60 is followed by fast electron transfer from
C60- to CA. It is noteworthy that C60 catalyses electron-transfer
reactions that cannot be driven by visible light (TTA/CA) or
would be very inefficient because of low triplet quantum yield32
(Pe/CA). Since C60 absorbs light over a broad range of the
visible spectrum and has a triplet quantum yield close to one,
it may find application in solar-driven chemical reactions.
(24) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259.
(25) Hung, R. R.; Grabowski, J. J. J. Phys. Chem. 1991, 95, 6073.
(26) Zeng, Y.; Biczo´k, L.; Linschitz, H. J. Phys. Chem. 1992, 96, 5237.
(27) v. d. Heuvel, D. J.; Chan, I. Y.; Gronen, E. J. J.; Schmidt, J.; Meijer,
G. Chem. Phys. Lett. 1994, 231, 111.
Acknowledgment. We much appreciate financial support
for this work provided by the Division of Chemical Sciences,
Office of Basic Energy Sciences, of the U.S. Department of
Energy to the University of Massachusetts (DE-FG02-84ER-
13242) and to Brandeis University (DE-FG02-89ER14072).
(28) Dubois, D.; Kadish, K. M.; Flanagan, S.; Wilson, L. J. J. Am. Chem.
Soc. 1991, 113, 7773.
(29) Handbook of Photochemistry; Scaiano, J. C., Ed.; CRC Press: Boca
Raton, FL, 1989; a correction of 0.09 V has been applied to account for a
change from acetonitrile to benzonitrile as solvent.2
(30) Kuroda, H.; Sakurai, T.; Akamatu, H. Bull. Chem. Soc. Jpn. 1966,
39, 1893.
References and Notes
(31) Carmichael, I.; Helman, W. P.; Hug, G. L. J. Phys. Chem. Ref.
Data 1987, 16, 239.
(1) Arbogast, J. W.; Foote, C. S.; Kao, M. J. Am. Chem. Soc. 1992,
114, 2273.
(2) Biczo´k, L.; Linschitz, H.; Walter, R. I. Chem. Phys. Lett. 1992,
195, 339; 1994, 221, 188. Biczo´k, L.; Linschitz, H.; Walter, R. I. Res.
Chem. Intermed. 1994, 20, 939.
(3) Nadtochenko, V. A.; Denisov, N. N.; Rubtsov, I. V.; Lobach, A.
S.; Moravskii, A. P. Chem. Phys. Lett. 1993, 208, 431.
(4) Michaeli, S.; Meiklyar, V.; Schulz, M.; Mo¨bius, K.; Levanon, H.
J. Phys. Chem. 1994, 98, 7444.
(5) Steren, C. A.; van Willigen, H. Proc. Indian. Acad. Sci. (Chem.
Sci) 1994, 106, 1671.
(32) Nakabayashi, K.; Toki, S.; Takamuku, S. Chem. Lett. 1986, 1890.
(33) Biczo´k, L.; Linschitz, H.; Treinin, A. In Proceedings of the
Symposium on Recent AdVances in Chemistry and Physics of Fullerenes
and Related Materials; Kadish, K. M., Ruoff, R. S., Eds.; Electrochemical
Society: Pennington, NJ, 1994; p 909.
(34) Handbook of Photochemistry; Murov, S. L.; Carmichael, I., Hug,
G. L., Eds.; M. Dekker: New York, 1993.
(35) Sandros, K. Acta Chem. Scand. 1964, 18, 2355.
(36) Neugebauer, F. A.; Bamberger, S.; Groh, W. R. Chem. Ber. 1975,
108, 2406.
(6) Steren, C. A.; van Willigen, H. In Proceedings of the Symposium
on Recent AdVances in Chemical Physics of Fullerenes and Related
JP960640H