J. Phys. Chem.

1996, 100, 12477-12482 12477

C60 and C70 Fullerene Ions in Nonpolar Liquids: Mobility and Radiation-Chemical Changes
G. Bakale*
Case Western ReserVe UniVersity, DiVision of Radiation Biology, 10900 Euclid AVenue,
CleVeland, Ohio 44106-4942

K. Lacmann and W. F. Schmidt

Hahn-Meitner Institut, Abteilung Strahlenchemie, Glienicker Strasse 100, 14109 Berlin, Germany
ReceiVed: December 1, 1995; In Final Form: May 1, 1996X

The drift mobilities of fullerene C60 and C70 ions were measured at several temperatures in n-pentane,
neopentane, cyclohexane, tridecane, and carbon disulfide to determine the degree of solute-solvent interaction
in these model systems of spherical C60 ions and ellipsoidal C70 ions in solvents of different shapes. One-
component ion current decays were generally observed in the fullerene solutions using a time-of-flight pulse-
conductivity technique, which indicates that ions of only one size were present in the solutions. These ions
Downloaded via UNIV NACIONAL AUTONOMA MEXICO on August 26, 2019 at 16:33:42 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

are assumed to be fullerene anions and cations formed by charge transfer from the solvents due to the higher
electron affinity and lower ionization potential of the fullerenes relative to the solvents. Marked decreases
of the ion mobilities were observed following γ-irradiation of the fullerene solutions, which indicate that
ions larger than C60 and C70 are produced by radiolysis. These lower mobility ions may be products of
radical addition to the fullerenes and/or dimerization products formed from radical or ionic precursors. Analysis
of the data with the Stokes-Einstein equation leads to the conclusion that the degree of solvation of the ions
is cyclohexane > neopentane = n-pentane > tridecane > carbon disulfide.

Introduction Dion ) kBTµion/eo (2)

The facile synthesis of C60 and C70 in macroscopic quantities1 where eo is unit charge. Combining eqs 1 and 2 yields
has triggered an avalanche of studies2 in which these unique
molecules have been characterized sufficiently well1-3 to make µion ) eo/(χπηRion) (3)
them attractive candidates to serve as tracer molecules in a
variety of physicochemical studies. Since C60 is characterized
From eq 3 it is evident that the product of mobility and viscosity
by a relatively low ionization potential of ∼7.6 eV and a high
is a constant, i.e.,
electron affinity of ∼2.5 eV, fullerene anions and cations can
be prepared in various matrices from primary charge-carriers µionη ) constant (4)
by electron transfer. Albrecht’s group used a time-of-flight
(TOF) method to measure the mobility of fullerene anions in which is also known as Walden’s rule. Numerous papers have
n-hexane4 in studies analogous to our preliminary TOF studies been published on the validity of the SE equation for diffusing
of fullerene ion mobility in liquid hydrocarbons before and after neutral species10 and of eq 3 for the mobility of ions,11 but
γ-irradiation.5 In the latter radiolysis studies, product ions were deviations of measured data from eq 1 or 3 are frequent and
observed having mobilities almost 2-fold less than those of C60 the applicability of the macroscopic SE model to the molecular
ions, which is a manifestation of the rich chemistry of highly level has been questioned.12 Despite such criticism, the SE
electrophilic fullerenes6 that are known as “radical sponges”.7 equation was recently referred to as “...a pillar of insight into
This reactivity and pulse8 and steady-state radiolysis studies of the molecular dynamics of liquids”13 and is used to interpret
fullerene solutions9 indicate an extremely complex radiation the mobilities measured in this work.
chemistry of fullerene. The results on µion of fullerenes in liquid hydrocarbons and
The well-defined geometries of C60 and C70 ions and their carbon disulfide reported herein extend our studies5,14 of the
large masses relative to other solute ions make the fullerenes applicability of the SE equation to quasi-spherical C60 ions and
ideal probes for studying the relationship between ion mobility ellipsoidal C70 ions in solvents having different shapes. Our
and solvent viscosity η. The Stokes-Einstein (SE) equation finding that preirradiation of fullerene/hydrocarbon solutions by
relates the diffusion coefficient, D, of a particle of radius R to 60Co γ-radiation changes the radius of the tracer fullerene ions
η by prompted a study to identify the radiolytic product by small-
angle X-ray scattering (SAXS) and mass spectral analysis that
D ) kBT/(χπηR) (1) was recently reported.15

where T is the temperature, kB is the Boltzmann constant, and Experimental Section

χ denotes a factor that depends on the degree of solute-solvent Cyclohexane, neopentane, and n-pentane (Phillips Research
interaction. A value of χ ) 6 denotes the “sticking” condition Grade, 99+ mol %) were purified using the same protocol as
of maximum interaction, whereas χ ) 4 applies to minimal described earlier16 except the low vapor pressure of tridecane
interaction in the “slipping” condition. For ions the diffusion (also Phillips, 99.8 mol %) precluded vacuum distillation of
coefficient, Dion, is proportional to the mobility, µion, this solvent. Therefore, tridecane was degassed by bubbling
with high-purity helium (99.99%) and stored under helium in a
X Abstract published in AdVance ACS Abstracts, June 15, 1996. glass cylinder that contained silica gel and a molecular sieve.
S0022-3654(95)03569-6 CCC: $12.00 © 1996 American Chemical Society
12478 J. Phys. Chem., Vol. 100, No. 30, 1996 Bakale et al.

Figure 1. Block diagram of the TOF pulse-conductivity setup used to

measure the drift time of ions. Electric field across electrodes is provided
by voltage supply U. Ions drifting in the field induce current across a
resistance R that is amplified and monitored by a digitizer/scope.
Figure 2. Oscilloscope traces of the decay of the ion current with E
The tridecane was transferred from this cylinder to an evacuated ) 40 kV/cm across electrodes separated by 0.075 cm: pure neopentane
glass conductivity cell in which the ion mobility was measured. (9); same neopentane + [C70] ) 7.0 µM (O); same C70/neopentane
solution γ-irradiated to 32 krads with cobalt-60 (+).
Carbon disulfide (Merck pro analysi, 99.7%) was degassed by
refluxing at -78 °C under vacuum and distilled directly into defined positive and negative ions give rise to linear decays
the conductivity cell. from which the drift time td can be determined, but if several
In the initial phases of study,5 fullerene solutions were ions having a range of mobilities are present, nonlinear decays
prepared from Strem Chemicals buckminsterfullerene, a mixture result, which makes determination of td difficult. The primary
of 10-12% C70 in C60. However, for the µion measurements cations and electrons produced via
reported herein, C60 and C70 each of Fluka purum grade (>97%)
were used. In addition, Hoechst C60 having a stated purity of M Df M+ + e- (5)
>98% based on HPLC analysis was also used in µion measure-
ments and was found to yield the same results as those obtained can react with C60/70 or other charge-accepting solutes present
using Fluka C60. Stock solutions of C60 and C70 were prepared in the liquid and produce well-defined secondary cations and
by dissolving the fullerenes in toluene (Phillips Research Grade, anions, respectively, via
99+ mol%) to yield ∼1-3 mM solutions from which appropri-
ate aliquots were used to prepare hydrocarbon or CS2 solutions M+ + C60/70f C60/70+ + M (6)
ranging from 2.5 µM for n-pentane to 12.5 µM for tridecane
with care taken not to exceed the solubility limits reported by e- + C60/70f C60/70- (7)
others.17 The solubility of C60 in tridecane was not measured
in these studies, but Ruoff et al. reported that the solubility of Since C60/70 cations and anions have approximately the same
C60 in n-decane is ∼100 µM and stated that the solubility of radius, their drift should give rise to a single linear decay as
C60 in the alkanes increases with increasing number of carbon illustrated in Figure 2 for fullerene ions in liquid neopentane,
atoms in the solvent.17b Also, Sivaraman et al. reported that which is discussed further in the next section. The point where
the solubility of C60 in dodecane increases from 91 to 126 µg/ the decay markedly deviates from linearity defines the end of
mL in tetradecane,17a which corresponds to 130-175 µM. Thus, the drift td, which for a parallel-plate cell with electrode
C60 should be soluble in tridecane at the 12.5 µM concentration separation d and an applied voltage U is related to the drift
studied in this work. The maximum concentration of the velocity Vd and µion via
fullerenes in carbon disulfide was <10 µM, which is well below
the reported solubility limit.17 Aliquots of the fullerene/toluene µion) Vd/E ) d2/(tdU) (8)
stock solutions were injected into the conductivity cell prior to
the addition of the solvent, and toluene was removed from the Application of eq 8 requires Vd to be proportional to E, which
cell via evaporation using a vacuum line to pump intermittently was observed to the highest field applied in this study, 55 kV/
on the 15-150 µL of fullerene solution in the cell until the cm. Within the error limits of our measurements of (15%,
line pressure was ∼10-4 Torr. Approximately 10-13 mL of we conclude that fullerene cations and anions have the same
the alkane solvent was then added to the cell either by distillation mobility.
for the more volatile solvents or by pouring under vacuum for In addition to the above TOF measurements of ionic drift
tridecane; the quantity of solvent added to the cell was times, electron half-lifes were measured in “pure” n-pentane
determined to (1% by weighing. Exposure of the fullerene and cyclohexane to obtain an indication of solvent purity. The
solutions to light was minimized throughout the experiment to electron half-life in pure cyclohexane generally exceeded 1 µs,
minimize possible photoinduced reactions that have been which indicates that the level of impurities having a diffusion-
reported.18 controlled rate constant was less than 250 nM, and the level of
Measurements of charge-carrier mobilities were made using impurities in the pentanes was estimated on a similar basis to
the pulse-conductivity technique illustrated schematically in be <150 nM. These electron half-lifes were measured using a
Figure 1 in which the solution is contained in a parallel-plate fast preamplifier, whereas ion drift times were measured with
conductivity cell irradiated by single pulses of radiation of a Keithley preamplifier (Model 427). In both cases the
negligible width (<1 µs) relative to the ion drift times (>1 ms). amplified signals were recorded with a Hewlett-Packard digitiz-
Glass cells having circular electrodes (1.5 cm diameter) ing oscilloscope (Model 54510B) and Labview software was
separated 0.55-0.75 mm were used across which an electric used to collect, process, and store data. Pulse widths ranged
field of 5-55 kV/cm was applied during irradiation by pulses from 5-200 ns and four to eight ion decay traces were
of 14 MeV bremsstrahlung from the HMI linear accelerator. sometimes averaged with the digitizer to obtain td. However,
Positive and negative charge-carriers produced in the irradiated a single trace was generally used to reduce smearing of the
volume drift in the electric field and induce ionization currents “knick” point that yields td. TOF measurements were made at
that decay as the ions are neutralized at the electrodes. Well- both polarities and averaged to preclude possible spurious
C60 and C70 Fullerene Ions J. Phys. Chem., Vol. 100, No. 30, 1996 12479

TABLE 1: Summary of Stokes Radii r (Å) of Fullerene Ions Derived from Measured Values of Mobility µ (cm2/(V‚s)) in
Neopentane, n-Pentane, Cyclohexane, Tridecane, and Carbon Disulfide before and after γ-Irradiation
C60 C70
unirradiated irradiated unirradiated irradiated
solvent T, °C η, cP µ r µ r µ r µ r
neopentane 0 0.33 3.6 7.2 2.3 11.3 4.2 6.2 2.6 10.0
26 0.20 5.0 8.5 3.3 12.9 6.3 6.7 3.8 11.2
r (average) 7.8 12.1 6.5 10.6
n-pentane -78 0.72 1.8 6.6 0.84 14.0 1.4 8.4 1.0 11.8
-32 0.37 3.4 6.8 1.9 12.4 3.1 7.4 2.0 11.5
24 0.21 5.5 7.4 3.4 11.9 5.2 7.8 3.8 10.6
r (average) 6.9 12.8 7.9 11.3
c-hexane 7 1.3 0.52 12.6 0.76 8.6 0.51 12.8
22 0.93 1.1 8.3 0.67 13.6 1.3 7.0 0.79 11.6
58 0.54 1.8 8.7 1.8 8.7 1.4 11.2
63 0.51 1.22 13.6
r (average) 8.5 13.3 8.1 11.9
tridecane 0 2.9 0.39 7.5 0.19 15.4 0.50 5.9 0.21 13.9
23 1.8 0.76 6.2 0.32 14.7 0.84 5.6 0.39 12.1
58 0.98 1.5 5.8 0.60 14.4 1.3 6.7 0.76 11.4
r (average) 6.5 14.8 6.1 12.5
carbon disulfide -78 0.85 2.2 4.5 1.6 6.2
0 0.44 4.5 4.3 3.8 5.1
26 0.35 5.5 4.4 5.1 4.7 4.8 5.0 4.2 5.7
r (average) 4.4 4.7 5.4 5.7

electrode effects. After td measurements the fullerene solution

was irradiated with γ-radiation at dose rates ranging from 0.3
to ∼4 krad/minute and td measurements were repeated soon after
irradiation and again ∼16-100 h later; no additional td changes
were observed after the latter times. The UV-visible spectra
of most fullerene solutions were checked after each experiment,
and the characteristic absorption of C60 at 327-329 nm19 and
the multiple absorption maxima of C70 between 310-380 nm19,20
were absent after irradiation of the solutions to >20 krads.

Results
The quality and signal/noise ratio of the of the ion current
decay data are illustrated in Figure 2 for “pure” neopentane and
a 7.0 µM solution of C70/neopentane before and after γ-irradia-
tion to 32 krad. The ion decays at 40 kV/cm illustrate the
marked change of td effected by fullerene and the further change
by γ-irradiation of the fullerene solution, which indicates that
a charge-carrier larger than the ion observed in pure neopentane
is present, or if more than one ion are present, the ions are of
similar mobilities. This is consistent with C60/70 cations and Figure 3. Ion drift velocity Vs electric field for 7.0 µM solutions of
anions formed via reactions 6 and 7, respectively, being the C70/cyclohexane before (O) and after (0) γ-irradiation to 27 krad.
dominant ions in the pulse-irradiated solutions, and upon
TABLE 2: Solvent Aggregation Numbers Derived from
γ-irradiation larger ions are produced and become the dominant Measured Stokes Radii Using Equation 10 for C60 and C70
charge-carriers. Ion mobilities were also measured in pure Ions before and after Irradiation to ∼25 krad
solvents irradiated to doses of 25-30 krads to ensure that none
solvent C-60 C-60 + γ C-70 C-70 + γ
of the observed radiation-induced mobility changes in the
fullerene solutions arose from radiation-induced changes of the cyclohexane 13.5 52 8.8 36
n-pentane 4.4 43 7.4 28
solvents (data not shown). neopentane 7.5 35 2.6 22
Values of td obtained from ion decays analogous to those in tridecane 1.5 32 0.8 19
Figure 2 were converted to ion drift velocities that were plotted carbon disulfide a a a 1.3
Vs electric field to obtain µion, an example of which is shown a Denotes indeterminate value of ni; see text.
in Figure 3 for solutions of C70 in cyclohexane before and after
irradiation. Values of µion thus derived for ions in the five
Discussion
solvents at various temperatures are listed in Table 1, which
also includes solvent viscosities and Stokes radii that were The Stokes radii and mobilities of C60 and C70 ions that are
obtained with eq 3. These radii were used to evaluate summarized in Table 1 constitute the crux of this study.
aggregation numbers of C60 and C70 ions in the solvents listed Differences between C60 and C70 ion radii are assumed to be
in Table 2. The dependences of C60/70 ion mobilities on negligible, which appears justified in view of Adams et al.
viscosity in the hydrocarbons plotted in Figure 4 illustrate that reporting a difference in the van der Waals volumes of these
γ-irradiation to doses of ∼20-30 krads decreases C60/70 ion fullerenes of 18%,21 which corresponds to <6% difference in
mobilities to almost half the values in unirradiated solutions as radii. These negligible differences are evident in Figure 4,
underscored by the lines with a slope of -1 that correspond to which also illustrates the inverse dependence of mobility on
radii of 7.0 and 12 Å before and after irradiation, respectively. viscosity in agreement with the SE equation and Walden’s law
12480 J. Phys. Chem., Vol. 100, No. 30, 1996 Bakale et al.

cautioned that agreement with Stokes’ law is often “an accident

of the rather narrow density ranges over which most experiments
have been performed”12 of which the reported agreement of Rion
and rvdW may be an example. We have no explanation,
however, for the 1.9 and 1.5 Å greater values of Stokes radii of
C60 and C70 ions, respectively, that we derived in n-pentane,
the solvent most similar to the n-hexane studied by Albrecht’s
group.
Fullerene Ion Interactions with Solvents. The underlying
interaction between fullerene ions and the surrounding solvent
is an ion-induced dipole that varies with the fourth power of r,
the distance from the ion that is assumed to be a point charge,
and is approximated by

r4 ) Reo2/(2kBT) (9)

where R and  are, respectively, the polarizability and dielectric

constant of the solvent (all in absolute cgs units). Negligible
differences should occur among fullerene ions interacting with
the various solvents via ion-induced dipoles owing to the fourth
root dependence of r on R for which a 2-fold difference in R
would effect <20% change in r. A rough approximation of
the solvent associated with ions via this interaction yields for
n-pentane with R ) 9.9 × 10-24 cm3 and  ) 2 a value of r )
5.2 Å as the distance where the energy of electrostatic
Figure 4. Dependence of mobility of C60 and C70 ions in hydrocarbon polarization interaction is equal to kBT at 295 K. This
solutions on viscosity before and after γ-irradiation: circles, cyclo-
approximation is consistent with a monolayer or less of
hexane; diamonds, neopentane; squares, n-pentane; triangles, tridecane.
Smaller symbols denote C60 and larger C70 ions. Open symbols denote n-pentane or the other solvents studied being bound to fullerene
before γ-irradiation to 25 ( 5 krads and closed symbols after ions by induced-polarization interaction.
γ-irradiation to 25 ( 5 krads. Among the unirradiated fullerene solutions, the system
cyclohexane/C60/70 exhibits the greatest solvation based on
over the viscosity range that was studied. The most salient measured Stokes radii of 8.5 and 8.1 Å for C60 and C70,
feature of Figure 4 is the nearly 2-fold decrease of µion effected respectively. This may be related to the special packing of
by γ-irradiation of the fullerene/hydrocarbon solutions, which cyclohexane molecules around neutral C60 molecules reported
corresponds to an increase of the Stokes radii from 7.0 and 12 by Nagano et al.24 who stated C60 is “massively solvated” by
Å derived using χ ) 6 in eq 3. 13.7 cyclohexane molecules and estimated C60(C6H12)13 to have
Differences in the Stokes radii of fullerene ions are evident a molar volume of 1830 cm3 mol-1, which yields an effective
not only before and after irradiation but also among the different radius of 8.8 Å. This value is in excellent agreement with the
solvents. For example, the average value of the Stokes radius Stokes radius reported in Table 1 and suggests that neutral C60
of C60 ions in the rod-like alkanes n-pentane and tridecane is is dressed by a shell of cyclohexane molecules that is not
6.7 Å, which is significantly different from the Stokes radii of significantly extended by ion-induced dipole interaction with
C60 ions in either cyclohexane (8.5 Å) or carbon disulfide (4.3 C60 ions. At the other extreme of solvation, the Stokes radius
Å), and analogous differences are apparent for the radii of C70 obtained for C60 ions in CS2 when evaluated with χ ) 6 in eq
ions. These radii can be compared with that of nonsolvated 3 is within our error limits equal to the van der Waals radius of
C60 for which the radius of the envelope through the carbon bare C60. This indicates that the slip condition of the SE
nuclei is 3.55 Å3 while the van der Waals radius is 5.0 Å.1,3,21 equation is more applicable to C60 ions in CS2, but setting χ )
Comparison of rvdW ) 5.0 of C60 molecules with the Stokes 4 in eq 3 increases the Stokes radius of C60 to ∼7.5 Å. These
radii of C60 ions listed in Table 1 indicates that hydrocarbon results imply that some degree of C60-CS2 solvation occurs
molecules are associated with C60 ions to varying degrees, but to a lesser extent than in hydrocarbon solutions, and an
whereas association with CS2 is negligible. A similar conclusion analogous conclusion is drawn for C70-CS2.
is drawn for ellipsoidal C70, which has major and minor axes The Stokes radii of C60/70 ions in Table 1 were used to
of 7.8 and 6.9 Å 22 that yield a van der Waals volume of 650 approximate the number of solvent molecules that is associated
Å3 from which an effective rvdW ) 5.4 Å is obtained23 that is with fullerene ions by applying the same method that Dubois
significantly less than the Stokes radii of C70 ions in hydrocar- et al. used in their voltammetry studies of C60 in more polar
bons reported in this work. aprotic solvents.25 In those studies Dubois et al. assumed that
These conclusions regarding solvation of fullerene ions by electrostriction was negligible in the solvation shell and
hydrocarbons are significantly different from those reported by calculated ni, the number of solvent molecules that fit in the
Albrecht’s group who reported a Stokes radius for C60 anions shell, using the equation
in n-hexane of 5 Å, which is equal to rvdW of the bare C60
molecule.4a The same group more recently reported a Stokes ni ) (4π/3)(rs3 - rvdW3)/Vi (10)
radius of 5.4 Å for C70 anions in n-hexane, which also agrees
well with the effective radius of C70 that they estimated to be in which Vi is the molar volume of the solvent molecules. The
5.6 Å.4c For both cases the agreement between the Stokes and values of ni thus obtained for fullerene and radiolytic product
van der Waals radii appears to be fortuitous, since χ ) 6 was ions are listed in Table 2 and are generally comparable to those
used in the SE equation to obtain the Stokes radii, but this is reported by Dubois et al. for C60 in the least polar solvents that
appropriate only if significant solute-solvent interaction occurs, they studied for which ni ranged from 4 to 11.25b In those
which, if occurring, would yield Rion > rvdW. Evans et al. studies, a mean Stokes radius of 7.4 Å was derived from which
C60 and C70 Fullerene Ions J. Phys. Chem., Vol. 100, No. 30, 1996 12481

Dubois et al. concluded “...a monomolecular layer of the solvent

is being dragged by the C60 molecule”, which coincides with
our conclusion for C60 in cyclohexane. Agreement is also noted
between the value of ni ) 13.5 in C60/cyclohexane solutions
and the solvation number of 13.7 reported by Nagano et al.24 It
should also be noted, however, that the cubic dependence of ni
on the difference in the two radii in eq 10 magnifies inaccuracies
of either radius and suggests that quantitative conclusions based
on ni should be viewed cautiously as indicated by three
indeterminate values of ni. These indeterminate values of ni
for CS2 also reflect less ion-CS2 interaction than the hydro-
carbons, which is consistent with the C60 radius of gyration of
3.8 Å measured in CS2 using small-angle neutron scattering.23
Between these solvation extremes are n-pentane and tridecane
in which the Stokes radii of C60 ions are, within experimental
error, the same as the effective hydrodynamic radius of C60 in
benzene of 6.4 Å measured using dynamic light scattering by
Ying et al.,26 who noted that this radius was the same as that of
C60(C6H6)4.27 Since the effective molecular diameter of benzene
is within 6% of that of n-pentane,28a ni ) 4.4 for C60 in n-pentane Figure 5. Sketch of proposed configuration of [C60(C6H12)10]2 dimer
seems reasonable, and the ∼40% greater diameter and 3-fold that is consistent with Stokes’ radius of ∼13 Å as well as results
greater volume of tridecane28b relative to n-pentane accounts reported in ref 15.
for significantly less aggregation of tridecane.
Radiation-Induced Changes of Fullerene Ions. The nearly are bonded by two carbon atoms of the C60 molecules analogous
2-fold increases of Stokes radii of C60/70 ions observed in to the structures reported by Morton et al.7d
γ-irradiation of fullerene/hydrocarbon solutions suggest that Mass spectral and SAXS analyses of the C60/cyclohexane
radiolytic dimerization of both fullerenes occurs, which was radiolysis product indicate that approximately 10 cyclohexyl
proposed in our preliminary studies.5 Supporting evidence for radicals add to C60,15 which forms a dimer similar to that shown
this proposal came from reports of dimerization of radical in Figure 5. This multiple addition of radicals to the same C60
adducts of C60 in liquids,7d photoinduced dimerization of C60,29 molecule may occur analogously to the addition of 34 or more
and laser desorption studies of C60 coalescence.30 These reports methyl radicals to the same C60 as was reported by Krusic et
suggested that the low-mobility ion formed in the radiolysis of al.,7b and multiple addition of H• atoms to C60 similarly occurs.32
C60/hydrocarbon solutions was a C120 dimer, but dimerization The ubiquitous presence of hydrogen atoms in the radiolysis of
is not consistent with the absence of an analogous radiolytic hydrocarbons33 as indicated in the proposed mechanism com-
product in the γ-irradiation of fullerene/CS2 solutions. Subse- bined with their known facile addition to C607,32 implies that
quent characterization of the product formed in γ-irradiated C60/ hydrogen-C60 products have a mobility that is not significantly
cyclohexane solutions revealed that it is formed via radical different from that of C60 or that such hydrogen-substituted
addition to C60, but an analogous product was not found in the products are inefficient charge scavengers and thus were not
radiolysis of fullerene/CS2 solutions.15 Thus, it became clear observed as dominant ions in these TOF studies.
that the hydrocarbon solvents play a key role in the radiolytic
In contrast to the radiolysis of fullerene/hydrocarbon solutions,
formation of low-mobility fullerene product ions. In order to
γ-irradiation of CS2 yields neither alkyl radicals nor H atoms
elucidate this role, the mechanism of hydrocarbon radiolysis
as indicated in the following mechanism proposed by Janssen
must be considered, critical elements of which are radicals, the
et al.:34
ubiquitous reactants in most radiolysis mechanisms,31 and the
radical-sponge properties of C60.7 Both elements are incorpo-
CS2Df CS• + S• (17)
rated in the following mechanism for the γ-irradiation of
solutions of C60 in alkane solvents, RH:
nCS• f (CS)n (18)
RH Df R• + H• (11)
R• + C60f RC60• (12) 8S• fS8 (19)

H• + C60f HC60• (13)

CS• + S• f CS2 (20)
2RC60• f RC60C60R (14)
CS• + S8 f CS2 + S7• (21)
2HC60• f HC60C60H (15)
The greater radiation stability of liquid CS2 at 25 °C relative
RC60• + HC60• f RC60C60H (16) to that of cyclohexane is evident from the G values (molecules/
100 eV absorbed) of radiolytic decomposition of both solvents,
Molecular hydrogen and cyclohexene are also produced in the G(-CS2) ) 0.8 34 and G(-c-C6H12) g 8,33 which undoubtedly
radiolysis of cyclohexane31 but are ignored, since neither is contributes to the greater radiation stability of the fullerenes in
pertinent to the ions observed in our TOF studies. The products CS2 relative to their stability in hydrocarbons. This difference
of radical addition to C60 in reactions 12 and 13 are also radicals combined with the apparent dominance of reactions 18-21 for
that can react with other radical-C60 adducts to form dimers CS• and S• relative to the addition of these radicals to C60 must
via reactions 14-16, which have been extensively studied by account for the absence of any S-C60 addition product that was
others.7 It should also be noted that the dimerization products detectable by mass spectral analysis.15
12482 J. Phys. Chem., Vol. 100, No. 30, 1996 Bakale et al.

In summary, the most salient conclusions drawn from this (9) Maity, D. K.; Palit, D. K.; Mohan, H.; Mittal, J. P. J. Chem. Soc.,
study are (i) fullerene ions serve as model charge-carriers that Faraday Trans. 1993, 89, 95.
(10) For a review, see the following: Edward, J. T. J. Chem. Educ.
can be used to probe ion-solvent interactions via the SE 1970, 47, 261.
equation, (ii) ions of C60 and C70 solvate hydrocarbons to varying (11) (a) Allen, A. O.; de Haas, M. P.; Hummel, A. J. Chem. Phys. 1976,
degrees but more than these ions solvate CS2, (iii) in the 64, 2587. (b) Gee, N.; Freeman, G. R. Can. J. Chem. 1989, 67, 27.
(12) Evans, D. F.; Tominaga, T.; Davis, H. T. J. Chem. Phys. 1981, 74,
γ-radiolysis of fullerene/hydrocarbon solutions ion formation 1298.
is strongly mediated by the solvent, and (iv) ions formed by (13) Tarjus, G.; Kivelson, D. J. Chem. Phys. 1995, 103, 3071
γ-radiolysis are a radical-adduct product with a radius that is (14) Katoh, R.; Lacmann, K.; Schmidt, W. F. Z. Naturforsch. 1994, 49A,
approximately 2-fold greater than that of the parent fullerene 793.
(15) Henderson, S. J.; Hettich, R. L.; Compton, R. N.; Bakale, G. J.
ion. These conclusions suggest new pathways for synthesizing Phys. Chem. 1996, 100, 5426.
novel fullerene derivatives that exploit radical-substituted ions (16) (a) Bakale, G.; Schmidt, W. F. Chem. Phys. Lett. 1990, 175, 319.
formed by radiolysis of C60 and C70 solutions. (b) Bakale, G.; Lacmann, K.; Schmidt, W. F. Phys .Lett. 1992, A168, 209.
(17) (a) Sivaraman, N.; Dhamodaran, R.; Kaliappan, I.; Srinavasan, T.
G.; Rao, P. R. V.; Mathews, C. K. J. Org. Chem. 1992, 57, 6077. (b) Ruoff,
Acknowledgment. This work was partially supported by the R. S.; Tse, D. S.; Malhotra, R.; Lorents, D. J. Phys. Chem. 1993, 97, 3379.
U.S. Department of Energy, Grant DE-FG02-88ER60617. G.B. (18) (a) Taylor, R.; Parsons, J. P.; Avent, A. G.; Rannard, S. P.; Dennis,
also thanks colleagues at HMI for their generous hospitality. T. J.; Hare, J. P.; Kroto, H. W.; Walton, D. M. Nature 1991, 351, 277. (b)
Rao, A. M.; Zhou, P.; Wang, K.-A.; Hager, G. T.; Holden, J. M.; Wang,
Y.; Lee, W.-T.; Xiang-Xin, B.; Ecklund, P. C.; Cornett, D. S.; Duncan, M.
References and Notes A.; Amster, I. J. Science 1993, 259, 955.
(1) Kratschmer, W.; Lamb, L. D.; Fostiropoulos, K.; Huffman, D. R. (19) (a) Allemand, P. M.; Koch, A.; Wudl, F.; Rubin, Y.; Diederich,
Nature 1990, 347, 354. F.; Alvarez, M. N.; Anz, S. J.; Whetten, R. L. J. Am. Chem. Soc. 1991,
(2) (a) For bibliography, see Braun, T.; Schubert, A.; Maczelka, H.; 113, 1050. (b) Bensasson, R. V.; Bienvenue, E.; Dellinger, M.; Leach, S.;
Vasvari, L. Fullerene Research, 1985-1993. A Computer-Generated Cross- Seta, P. J. Phys. Chem. 1994, 98, 3492. (c) Catalan, J. Chem. Phys. Lett.
Indexed Bibliography of the Journal Literature; World Scientific: River 1994, 223, 159.
Edge, NJ, 1995. For reviews and monographs, see the following. (b) Kroto, (20) (a) Lawson, D. R.; Feldheim, D. L.; Foss, C. A.; Dorhout, P. K.;
H. W.; Alaf, A. W.; Balm, S. P. Chem. ReV. 1991, 91, 1213. (c) Elliot, C. M.; Martin, C. R.; Parkinson, B. J. Phys. Chem. 1992, 96, 7175.
Fullerenes: Synthesis, Properties, and Chemistry of Large Carbon Clusters; (b) Bensasson, R. V.; Hill, T.; Lambert, C.; Land, E. J.; Leach, S.; Truscott,
Hammond, G. S., Kluck, V. J., Eds.; ACS Symposium Series 481; American T. G. Chem. Phys. Lett. 1993, 206, 197.
Chemical Society: Washington, DC, 1992. (d) Acc. Chem. Res. 1992, 25. (21) Adams, G. B.; O’Keeffe, M.; Ruoff, R. S. J. Phys.Chem. 1994,
(3) See references 1 and 2 and Hawkins, J. M.; Meyer, A.; Lewis, T. 97, 9465.
A.; Loren, S.; Hollander, F. J. Science 1991, 252, 312. (22) . McKenzie, D. R.; Davis, C. A.; Cockayne, D. J. H.; Muller, D.
(4) (a) Lim, S. K.; Burba, M. E.; Albrecht, A. C. Chem. Phys. Lett., A.; Vassallo, A. M. Nature 1992, 355, 622.
1993, 216, 405. (b) Lim, S. K.; Burba, M. E.; Albrecht, A. C. J. Phys. (23) Affholter, K. A.; Henderson, S. J.; Wignall, G. D.; Bunick, G. J.;
Chem. 1994, 98, 9665. (c) Burba, M. E.; Lim, S. K.; Albrecht, A. C. J. Haufler, R. E.; Compton, R. N. J. Phys. Chem. 1994, 99, 9224.
Phys. Chem. 1995, 99, 11839. (24) Nagano, Y.; Tamura, T.; Kiyobayashi, T. Chem. Phys. Lett. 1994,
(5) Bakale, G.; Lacmann, K.; Schmidt, W. F. ICDL Proceedings; Fuhr; 228, 125.
J., Biller, P., Eds.; IEEE Publishers: Piscataway, NJ 1993; p 131. (25) (a) Dubois, D.; Moninot, G.; Kutner, W.; Jones, M. T.; Kadish, K.
(6) (a) Wudl, F.; Hirsch, A.; Khemani, K. C.; Suzuki, T.; Allemand. M. J. Phys. Chem. 1992, 96, 7137. (b) Solvents studied in ref 25a with ni
P.-M.; Koch, A.; Eckert, H.; Srdanov, G.; Webb, H. M. Survey of Chemical ) 4-11 were chlorobenzene, 1,2-dichlorobenzene, and pyridine.
Reactivity of C60, Electrophile and Dieno-polarophile Par Excellence. (26) Ying, Q.; Marecek, J.; Chu, B. J. Chem. Phys. 1994, 101, 2665.
Fullerenes: Synthesis, Properties, and Chemistry of Large Carbon Clusters; (27) Meidine, M. F.; Hitchcock, P. B.; Kroto, H. W.; Taylor, R.; Walton,
Hammond, G. S., Kluck, V. J., Eds.; ACS Symposium Series 481; American D. R. M. J. Chem. Soc., Chem. Commun. 1992, 1534.
Chemical Society: Washington, DC, 1992; pp 161-175. (b) Taylor, R.; (28) (a) Ben-Amotz, D.; Herschbach, D. R. J. Phys. Chem. 1990, 94,
Walton, D. R. M., Nature 1993, 363, 685. (c) Haddon, R. C. Science 1993, 1038. (b) Ben-Amotz, D.; Willis, K. G. J. Phys. Chem. 1993, 97, 7736.
261, 1545. (29) Wang, Y.; Holden, J. M.; Dong, Z.-H.; Bi, X.; Ecklund, P. C. Chem.
(7) (a) Krusic, P. J.; Wasserman, E.; Parkinson, B. A.; Malone, B.; Phys. Lett. 1993, 211, 341.
Holler, E. R.; Keizer, P. N.; Morton, J. R.; Preston, K. F. J. Am.Chem.Soc. (30) Yeretzian, C.; Hansen, K.; Diedrich, F.; Whetten, R. L. Nature 1992,
1991, 113, 6274. (b) Krusic, P. J.; Wasserman, E.; Keizer, P. N.; Morton, 359, 44.
J. R.; Preston, K. F. Science 1991, 254, 1183. (c) McEwen, C. N.; McKay, (31) Spinks, J. W. T.; Woods, R. J. An Introduction to Radiation
R. G.; Larsen, B. S. J. Am.Chem.Soc. 1992, 114, 4412. (d) Morton, J. R.; Chemistry; J. Wiley & Sons: New York, 1964; Chapter 6.
Preston, K. F.; Krusic, P. J.; Hill, S. A.; Wasserman, E. J. Am.Chem. Soc (32) Howard, J. A. Chem. Phys. Lett. 1993, 203, 540.
1992, 114, 5454. (33) Spinks, J. W. T.; Woods, R. J. An Introduction to Radiation
(8) (a) Dimitrijevic, N. M.; Kamat, P. V. J. Phys.Chem. 1992, 96, 4811. Chemistry; J. Wiley & Sons: New York, 1964; Chapter 9.
(b) Guildi, D. M.; Hungerbuhler, H.; Janata, E.; Asmus, K.-D. J. Phys. (34) Janssen, O.; Henglein, A.; Perner, D. Z. Naturforsch. 1964 19B,
Chem. 1993, 97, 11258. (c) Bensasson, R. V.; Hill, T.; Lambert, C.; Land, 1005.
E. J.; Leach, S.; Truscott, T. G. Chem. Phys. Lett. 1993, 201, 326. (d) Guildi,
D. M.; Neta, P.; Asmus, K.-D. J. Phys. Chem. 1994, 98, 4617. JP9535699