Kinetic substituent and solvent effects in 1,3-dipolar cycloaddition of diphenyldiazomethanes with fullerenes C60 and C70: A comparison with the addition to TCNE, DDQ, and chloranil

Article abstract of DOI:10.1021/jo052580u

Kinetics of 1,3-dipolar cycloaddition of a series of meta- and para-substituted diphenyldiazomethanes (DDMs) with fullerenes C₆₀ and C₇₀ as dipolarophiles have been investigated in toluene at 30 0°C. Fullerene C₆₀ was ca. 1.5 times more reactive than C ₇₀. The rate constants (k) for the primary [3 + 2] additions increased with the increase of the electron-releasing ability of the meta and para substituent. The log k/k₀ values were well correlated by the Yukawa-Tsuno (Y-T) equations with the smaller negative p values (-1.6 and -1.7 for C₆₀ and C₇₀) and the reduced resonance reaction constants r (0.22 and 0.17) compared to similar reactions of common acceptors, TCNE, DDQ, and chloranil (CA). The plots of log k (acceptor) versus log k (C₆₀) as reference gave good regression equations and the slopes became larger in the order of TCNE > DDQ > CA > C₇₀ ≥ C₆₀. The rates were also found to decrease with the increase of solvent polarity due to the ground-state solvation of fullerenes. However, the relative reactivity of these acceptors toward the unsubstituted DDM increased in the order of DDQ > C₆₀ ≥ C₇₀ > TCNE > CA. The unexpected higher reactivity of fullerenes was interpreted in terms of the inherent steric strain by the pyramidalization of the sp² C-atoms as well as the shorter [6,6] bonds with larger π-electron densities.

Full text of DOI:10.1021/jo052580u

Kinetic Substituent and Solvent Effects in 1,3-Dipolar Cycloaddition
of Diphenyldiazomethanes with Fullerenes C and C : A
6
0
70
Comparison with the Addition to TCNE, DDQ, and Chloranil
Takumi Oshima,* Hiroshi Kitamura, Taijiro Higashi, Ken Kokubo, and Nozomu Seike
Department of Applied Chemistry, Graduate School of Engineering, Osaka UniVersity, Suita,
Osaka 565-0871, Japan
oshima@chem.eng.osaka-u.ac.jp
ReceiVed December 15, 2005
Kinetics of 1,3-dipolar cycloaddition of a series of meta- and para-substituted diphenyldiazomethanes
(
DDMs) with fullerenes C60 and C70 as dipolarophiles have been investigated in toluene at 30 °C. Fullerene
60 was ca. 1.5 times more reactive than C70. The rate constants (k) for the primary [3 + 2] additions
increased with the increase of the electron-releasing ability of the meta and para substituent. The log k/k
values were well correlated by the Yukawa-Tsuno (Y-T) equations with the smaller negative F values
-1.6 and -1.7 for C60 and C70) and the reduced resonance reaction constants r (0.22 and 0.17) compared
C
0
(
to similar reactions of common acceptors, TCNE, DDQ, and chloranil (CA). The plots of log k (acceptor)
versus log k (C60) as reference gave good regression equations and the slopes became larger in the order
of TCNE > DDQ > CA > C70 g C60. The rates were also found to decrease with the increase of solvent
polarity due to the ground-state solvation of fullerenes. However, the relative reactivity of these acceptors
toward the unsubstituted DDM increased in the order of DDQ > C60 g C70 > TCNE > CA. The
unexpected higher reactivity of fullerenes was interpreted in terms of the inherent steric strain by the
2
pyramidalization of the sp C-atoms as well as the shorter [6,6] bonds with larger π-electron densities.
Introduction
release leads to [6,5]-open fulleroids and then valence isomer-
ization provides more stable [6,6]-closed methanofullerenes.5
Besides the mechanistic study, the quantitative elucidation
of the chemical reactivity as well as the physicochemical
properties of fullerenes are very important issues in view of
the syntheses of modified derivatives, the control of regio-
Since the discovery of macroscopic synthetic methods of
fullerene C60, a large number of reports have been published
on the preparation and characterization of fullerene adducts and
1
derivatives. Due to the highly conjugated low-lying LUMO
orbital, C60 is demonstrated to easily undergo electrocyclic
2
(1) (a) Hirsch, A. Chemistry of the Fullerenes. Thieme, Stuttgart 1994.
(b) Diederich, F.; Thilgen, C. Science 1996, 271, 317-323. (c) Hirsch, A.
Top. Curr. Chem. 1999, 199, 1-65. (d) Wilson, S. R.; Schuster, D. I.; Nuber,
B.; Meier, M. S.; Maggini, M.; Prato, M.; Taylor, R. Fullerene, Chemistry,
Physics and Technology; John Wiley & Sons Inc.: New York, 2000; pp
91-176. (e) Nakamura, Y.; O-Kawa, K.; Nishimura, J. Bull. Chem. Soc.
Jpn. 2003, 76, 865-882. (f) Hirsch, A.; Brettreich, M. Fullerenes Chemistry
and Reactions; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim,
Germany, 2005.
reactions such as 1,3-dipolar cycloadditions and Diels-Alder
3
reactions. Among the established reactions, the dipolar cy-
cloaddition of diazoalkanes has attracted particular attention as
a result of the interesting reaction features depending on the
structures of diazoalkanes.4 The primary [3 + 2] addition
preferentially occurs at the [6,6] ring conjunct double bond to
yield [6,6]-closed pyrazoline adducts.4 Subsequent nitrogen-
a-c
1
0.1021/jo052580u CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/18/2006
J. Org. Chem. 2006, 71, 2995-3000
2995

Oshima et al.
chemical functionalization, and the further development of
fullerene chemistry. For example, Haddon et al. reported a
comparative study of electron-withdrawing substituents on
methanofullerene (C61), correlating the first reduction potential
SCHEME 1
6
with the Hammett σm. Scorrano et al. tried to quantitatively
evaluate the substituent effect of fullerene ring from the pK value
7
of a fullerene based benzoic acid. However, to the best of our
knowledge, a systematic kinetic study of the reaction of C60
and higher C70 has not yet been reported.3b In addition, the
question remains open as to whether the C60 is more reactive
than C70 and on how large the reactivity of C60 is as compared
with the common acceptor dipolarophiles.
To gain more insight into the chemical reactivity of fullerenes
C60 and C70, we have investigated the kinetics of 1,3-dipolar
cycloaddition of a series of meta- and para-substituted diphen-
yldiazomethanes (DDMs, 1a-m) with spherical C60 and C70
as a model reaction (Scheme 1). We have also compared them
with similar additions of DDMs to TCNE, DDQ, and chloranil
(CA) to explore the steric and electronic effects of fullerene
reactions.
Results and Discussion
Kinetic Measurements. The kinetic measurements were
carried out under pseudo-first-order conditions, using a large
(
2) (a) Suzuki, T.; Li, Q.; Khemani, K. C.; Wudl, F.; Almarsson, O¨ .
Science 1991, 254, 1186-1188. (b) Diederich, F. Nature 1994, 369, 199-
2
07. (c) Taylor, R.; Walton, D. R. M. Nature 1993, 363, 685-693. (d)
Hirsch, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1138-1141. (e)
Diederich, F.; Issacs, L.; Philp, D. Chem. Soc. ReV. 1994, 23, 243-255.
excess of DDMs (25-50 equiv) with respect to C60 and C70.
The progress of the reaction was followed by HPLC, using a
Buckeyprep column. The representative HPLC chart obtained
(
3) (a) Pang, L. S. K.; Wilson, M. A. J. Phys. Chem. 1993, 97, 6761-
6
7
763. (b) Mestres, J.; Duran, M.; Sol a` , M. J. Phys. Chem. 1996, 100, 7449-
454 and references therein. (c) Ohno, M.; Azuma, T.; Kojima, S.;
-
3
for the kinetic solution initially containing C60 (1.0 × 10 mM)
Shirakawa, Y.; Eguchi, S. Tetrahedron 1996, 52, 4983-4994. (d) Wang,
-
2
G.-W.; Saunders, M.; Cross, R. J. J. Am. Chem. Soc. 2001, 123, 256-259.
and 1c (4.0 × 10 mM) in toluene at 30 °C on 0.5 h standing
was shown along with the time course of the product distribu-
tions in Figure 1. There is the buildup of two monoadducts
(fulleroid (B) and methanofullerene (C)) and some bisadducts
(
3
e) Kr a¨ utler, B.; M u¨ ller, T.; Duarte-Ruiz, A. Chem. Eur. J. 2001, 7, 3223-
235. (f) Miller, G. P.; Mack, J. Org. Lett. 2000, 2, 3979-3982. (g) Miller,
G. P.; Mack, J.; Briggs, J. Org. Lett. 2000, 2, 3983-3986. (h) Paquette, L.
A.; Graham, R. J. J. Org. Chem. 1995, 60, 2958-2959. (i) Chronakis, N.;
Orfanopoulos, M. Org. Lett. 2001, 3, 545-548. (j) Chronakis, N.; Froadakis,
G.; Orfanopoulos, M. J. Org. Chem. 2002, 67, 3284-3289. (k) Sarova, G.
H.; Berberan-Santos, M. N. Chem. Phys. Lett. 2004, 397, 402.
(
D) between peaks of 1c and C60 (A). The identity of adduct B
as fulleroid is due to the complete thermal isomerization to the
methanofullerene C at 160 °C for 15 h in o-dichlorobenzene.
Monitoring the kinetic reaction by HPLC showed the gradual
increase of both B and C accompanied by the first-order decay
of C60 (A). Similarly, we also observed the formation of several
1:1 and 1:2 adducts in the reaction with C70 (for more details
(4) (a) Smith, A. B., III; Strongin, R. M.; Brard, L.; Furst, G. T.;
Romanow, W. J.; Qwens, K. G.; Goldschmidt, R. J.; King, R. C. J. Am.
Chem. Soc. 1995, 117, 5492-5502. (b) Suzuki, T.; Li, Q.; Khemani, K.
C.; Wudl, F. J. Am. Chem. Soc. 1992, 114, 7301-7302. (c) Smith, A. B.,
III; Strongin, R. M.; Brard, L.; Furst, G. T.; Romanow, W. J.; Qwens, K.
G.; King, R. C. J. Am. Chem. Soc. 1993, 115, 5829-5830. (d) Schick, G.;
Hirsch, A. Tetrahedron 1998, 54, 4283-4296. (e) Prato, M.; Suzuki, T.;
Wudl, F.; Lucchni, V.; Maggini, M. J. Am. Chem. Soc. 1993, 115, 7876-
8
see the Supporting Information). The natural logarithmic plots
of the signal intensities of the remaining C60 and C70 relative
7
7
1
877. (f) Wilson, S. R.; Wu, Y. J. Chem. Soc., Chem. Commun. 1993, 784-
86. (g) Suzuki, T.; Li, Q.; Khemani, K. C.; Wudl, F. J. Am. Chem. Soc.
992, 114, 7301-7302. (h) Shi, S.; Khemani, K. C.; Li, Q.; Wudl, F. J.
Am. Chem. Soc. 1992, 114, 10656-10657. (i) Prato, M.; Bianco, A.;
Maggini, M.; Scorrano, G.; Toniolo, C.; Wudl, F. J. Org. Chem. 1993, 58,
5
578-5580. (j) Wooley, K. L.; Hawker, C. J.; Fr e` het, J. M. J.; Wudl, F.;
Srdanov, G.; Shi, S.; Li, C.; Kao, M. J. Am. Chem. Soc. 1993, 115, 9836-
837.
5) (a) Wudl, F. Acc. Chem. Res. 1992, 25, 157-161. (b) Isaacs, L.;
9
(
Wehrsing, A.; Diederich, F. HelV. Chim. Acta 1993, 76, 1231-1250. (c)
Prato, M.; Lucchini, V.; Maggini, M.; Stimpfl, E.; Scorrano, G.; Eiermann,
M.; Suzuki, T.; Wudl, F. J. Am. Chem. Soc. 1993, 115, 8479-8480. (d)
Diederich, F.; Isaacs, L.; Philp, D. J. Chem. Soc., Perkin Trans. 2 1994,
3
91-394. (e) Hummelen, J. C.; Knight, B. W.; LePeq, F.; Wudl, F.; Yao,
J.; Wilrins, C. L. J. Org. Chem. 1995, 60, 532-538. (f) Li, Z.; Bouhadir,
K. H.; Shevlin, P. B. Tetrahedron Lett. 1996, 37, 4651-4654. (g) Ishida,
T.; Shinozuka, K.; Nogami, T.; Kubota, M.; Ohashi, M. Tetrahedron 1996,
5
1
1
2, 5103-5112. (h) Li, Z.; Shevlin, P. B. J. Am. Chem. Soc. 1997, 119,
149-1150. (i) Hall, M. H.; Lu, H.; Shevlin, P. B. J. Am. Chem. Soc. 2001,
23, 1349-1354.
FIGURE 1. (a) Typical HPLC trace of the sample aliquot (t ) 0.5 h)
of the reaction of 1c and C60 with a mixture of toluene/hexane (4/1,
v/v). (b) Time course of the reaction of 1c with C60 in toluene at 30
(
6) Keshavarz, M.-K.; Knight, B.; Haddon, R. C.; Wudl, F. Tetrahedron
°
C. Relative intensity of each component was obtained by dividing its
1
996, 52, 5149-5159.
7) Bagno, A.; Claeson, S.; Maggini, M.; Martini, M. L.; Prato, M.;
Scorrano, G. Chem. Eur. J. 2002, 8, 1016-1023.
(
signal intensity by that of the first aliquot C60: C60 (A), fulleroid (B),
and methanofullerene (C).
2996 J. Org. Chem., Vol. 71, No. 8, 2006

1
,3-Dipolar Cycloaddition of Diphenyldiazomethanes
TABLE 1. Second-Order Rate Constants (k/M- s-1) for Reaction
1
+
log(k/k ) ) -1.66(σ + 0.167∆ σj ) + 0.00
0
R
of DDMs with C60 and C70 at 30 °C in Toluene
2
0 k / M s-1
2
a
-1
10 k / M s-1
2
a
-1
(R ) 0.98, n ) 12) for C (2)
70
1
DDMs
C60 C70
DDMs
C60
C70
It is easily noted that C60 gives a slightly smaller negative F
value (-1.61) but somewhat larger resonance reaction constant
1
1
1
1
1
1
1
a
b
c
191
104
1h
1i
1j
1k
1l
1m
7.16
6.66
3.58
3.13
1.29
0.475
c
3.29
2.17
2.30
0.742
0.303
49.9
40.6
20.7
12.4
11.3
34.7
32.4
12.6
8.29
7.83
7.19
(
r ) 0.216) compared to C70 (-1.66 and 0.167, respectively).
d
e
These smaller negative F and positive r values imply that the
appearance of positive charge on the diazo-carbon as well as
its resonance stabilization is rather small in the rate-determining
transition state. This is in harmony with the concerted mecha-
b
f
g
9.76
a
b
Average of at least two measurements. Error limit is (2%. The rate
13
nism of symmetry allowed [π4s + π2s] cycloaddition of larger
2
-1 -1
constants (10 k/M
s ) at other temperatures (38, 44, and 50 °C) are 17.9,
HOMO-LUMO energy difference. The differential reactivity
between C60 and C70 (as to the above point 2) will be discussed
in the last section.
2
3.4, and 31.8 for C60 and at 40 and 50 °C they are 13.0 and 23.0 for C70,
respectively. Not observed.
c
A Comparison with TCNE, DDQ, and Chloranil (CA)
Reactions. (a) Dependency of Rates on the Meta and Para
Substituents. Fundamental knowledge of the chemical reactivity
of fullerenes as compared to the common acceptors is very
eagerly desired from synthetic and theoretical viewpoints.
Fortunately, we have previously investigated the kinetic sub-
stituent effects in the 1,3-dipolar cycloaddition of DDMs with
to those of the first aliquots (t ) 0) gave the linear correlation
against time more than 70% completion of the reaction. The
obtained slope, that is, the pseudo first-order rate constant kobs,
was divided by the corrected concentration of DDMs to provide
the second-order rate constants k2. The k2 values thus obtained
for various DDMs (1a-m) were collected in Table 1.
Kinetic Substituent Effects. A brief survey of Table 1
indicates the two noticeable points: (1) the rates were acceler-
ated by the electron-releasing (donor) substituent of DDMs and
9
1
4
15
some typical acceptors such as TCNE, DDQ, and chloranil
16
(CA). Before describing the kinetic substituent effects of these
representative acceptors, we must consider the reaction features
of these 1,3-dipolar cycloadditions. TCNE is known to react
with DDMs at the electron-deficient central CdC double bond
to afford tetracyanocyclopropane derivatives via a rapid nitrogen
release from the primary [3 + 2] adduct pyrazolines.14 On the
other hand, due to steric reasons, DDQ and CA react exclusively
at the less hindered outer CdO group rather than the inner CdC
bond to give very labile 3-oxapyrazolines which quickly lose
(2) C60 tended to react with DDMs 1.3-2 times faster than C70.
We first discuss point 1. This rate enhancement can be easily
rationalized in view of the typical 1,3-dipolar cycloaddition
between DDMs (type I class dipole as characterized by
10
Sustmann ) and fullerenes (acceptor dipolarophile). Namely,
the electron-releasing meta and para substituents lift the HOMO
energies of DDMs and hence diminish the energy gap between
HOMO (DDMs) and LUMO (fullerenes). This electronic effect
leads to the acceleration of rate by the increased energy gain in
the transition state according to the simple application of FMO
1
5,16
nitrogen to become the betaine intermediates.
Although the solvents were different from that of the present
fullerene reactions (toluene), these reactions exhibited rather
more enhanced kinetic substituent effects with larger negative
F values and raised resonance reaction constants (r), i.e., TCNE
10
theory to the comparison of reactivity sequences. Thus, the
p-MeO-substituted 1b provided ca. 110-fold larger rate constant
than p-NO2-substituted 1m for both C60 and C70 (Table 1).
(
F ) - 2.67 for the Brown-Okamoto equation (corresponding
1
4
to r ) 1.0 for the Y-T equation by definition), in benzene at
0 °C), DDQ (F ) - 2.33 for Y-T equation (r ) 0.47), in
To assess the degree to which the fullerene reactions respond
to the meta and para substituent changes of DDMs, we applied
the free energy relationships for log(k/k0), where k0 refers to
the rate constant for the unsubstituted DDM (1f). It was found
3
15
benzene at 30 °C), and CA (F ) -1.69 for Y-T equation
(
16
r ) 0.66), in THF at 30 °C), respectively. Assuming that the
substituent effects are not so significantly affected by changing
solvents in a narrow range of solvent polarity as represented
11
that the two-parameter Yukawa-Tsuno (Y-T) equation gave
excellent results for both C60 and C70 reactions (eqs 1 and 2)
17
by Reichardt’s ET values, i.e., toluene (ET ) 33.9), benzene
(
for the single parameter treatment with the Hammett σ and
(
34.3), and THF (37.4), it can be said that the electron
+
12
the Brown σ scale, see the Supporting Information).
2
withdrawal from the sp diazo-carbon of DDMs increases in
the order of TCNE > DDQ > CA > C70 g C60.
+
log(k/k ) ) -1.61(σ + 0.216∆ σj ) + 0.02
We can more explicitly recognize the relative magnitude of
the kinetic substituent effects for each acceptor when the log
k(acceptor) was plotted against the log k(C60) as reference
0
R
2
(
R ) 0.98, n ) 13) for C (1)
60
(Figure 2). First noticeable is that a beautiful linear free energy
(
8) In the case of C70, we observed the formation of several 1:1 and 1:2
relationship has been found to hold for all acceptors investigated.
1
2
adducts by HPLC. We have also confirmed by HPLC and H NMR that
the fulleroid thermally rearranges to the corresponding methanofullerene
as found for the reaction with bis(4-anisyl)diazomethane (1a).
The regression lines provided the slopes of 2.56 (R > 0.99)
2
2
for TCNE, 1.57 (R ) 0.99) for DDQ, 1.29 (R ) 0.99) for
(
9) The corrected concentration of DDMs was obtained by subtracting
the half amount of C60 or C70 from the initial one.
10) (a) Sustmann, R. Tetrahedron Lett. 1971, 12, 2717-2720. (b)
Sustman, R.; Trill, H. Angew. Chem., Int. Ed. Engl. 1972, 11, 838-840.
c) Sustmann, R. Pure Appl. Chem. 1974, 40, 569-593.
11) Tsuno, Y.; Fujio, M. AdVances in Physical Organic Chemistry;
Bethell, D., Ed.; Academic Press: Tokyo, Japan, l999; Vol. 32, pp 267-
85.
12) Williams, A. Free Energy Relationships in Organic and Bio-organic
Chemistry; Royal Society of Chemistry: Letchworth, UK, 2003.
(13) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl. 1969,
8, 781-853.
(
(14) Oshima, T.; Yoshioka, A.; Nagai, T. J. Chem. Soc., Perkin Trans.
2 1978, 1283-1288.
(
(
(15) Oshima, T.; Nagai, T. Bull. Chem. Soc. Jpn. 1981, 54, 2039-2044.
(16) Oshima, T.; Nagai, T. Bull. Chem. Soc. Jpn. 1980, 53, 3284-3288.
(17) (a) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358. (b) Reichardt,
C. SolVents and SolVent Effects in Organic Chemistry; Wiley-VCH Verlag
GmbH & Co. KGaA: Weinheim, Germany, 2003.
3
(
J. Org. Chem, Vol. 71, No. 8, 2006 2997

Oshima et al.
FIGURE 3. LUMO of TCNE, DDQ, and chloranil (CA) by MOPAC
PM3 calculation.
SCE) of the acceptors studied increase in the order of DDQ
2
0
21
(
+0.50 V, CH3CN) > TCNE (+0.24 V, CH3CN) > CA
FIGURE 2. Plots of log k (acceptor) versus log k (C60) at 30 °C: (1)
a, (2) 1b, (3) 1c, (4) 1d, (5) 1e, (6) 1f, (7) 1g, (8) 1h, (9) 1i, (10) 1j,
11) 1k, (12) 1l, and (13) 1m. The solvents are benzene (for DDQ and
2
0
22
1
(
(-0.01 V, CH3CN) > C70 (-0.41 V, CH2Cl2) > C60 (-0.44
2
2
V, CH2Cl2). Therefore, the smaller negative F values for C60
TCNE), toluene (for C60 and C70), and THF (for CA). The broken lines
a and b stand for C60 and C70, respectively. The values in parentheses
are the slopes of the regression lines (unit slope for C60). All the
and C70 can be explained by the weaker electron affinity as
23
well as the smaller LUMO coefficients.
By contrast, TCNE exhibited the highest substituent depen-
dency despite a lower E than DDQ. This discrepancy may
2
regressions provided good correlation coefficients (R g 0.99).
red
2
be rationalized by the considerably larger LUMO coefficients
of TCNE at the central CdC double bond as compared to the
DDQ carbonyl function (Figure 3), because the larger LUMO
orbital coefficients cause the more pronounced charge transfer
from the HOMO of DDMs.
CA, and 1.003 (R > 0.99) for C70 (broken line b), respectively.
By definition, the reference C60 reaction gives the unit slope
through the origin (broken line a). The magnitudes of these
slopes refer to the relative sensitivity of each acceptor on the
substituent changes of DDMs as compared with reference C60.
The steepest TCNE line crossed with the C70 and C60 lines.
To address these substituent effects, we must consider several
factors which affect the electronic state of the central diazo-
carbon of DDMs. It is evident that the strong acceptor will
significantly withdraw electron from the relevant diphenyl-
bearing carbon atom (i.e., charge transfer from DDMs to
acceptor) in the transition state. One of the most important
Strangely, however, CA gave the larger resonance reaction
constant (r ) 0.66) than DDQ (0.47), suggesting the occurrence
of more enhanced π-electron-donating delocalization by donor
substituent. The π-resonance effects are highly dependent on
2
4
the conformation of adjoining aromatic rings. As a conse-
quence, the delicate balance of these crucial factors, i.e., the
electron affinity (LUMO energy), the LUMO coefficients, as
well as the conformation of the π-resonating aromatic ring,
would determine the magnitudes of kinetic substituent effects
as depicted in Figure 2.
red
criteria for the electron affinity is the reduction potential (E )
or LUMO energy of the acceptor18 because more positive E
red
(
the more lower LUMO) brings about more enhanced electron
(b) Dependency of Rates on Solvents. In Figure 2, we can
withdrawal from the transition state DDMs, thus inducing more
development of positive charge on the central diazo-carbon. In
addition, the LUMO coefficients at the reaction site of the
acceptor play a crucial role in the magnitude of the orbital
interaction with the HOMO of DDMs. According to the FMO
model, reactions take place preferably in the maximal HOMO-
LUMO overlap with the product of (large) × (large) orbital
coefficients than the others.19 Moreover, the steric bulk of
acceptor exerts the some inhibition about performing the ideal
orbital interaction with DDMs.
easily grasp the unexpected higher reactivity of C60 and C70 as
(20) Sasaki, K.; Kashimura, T.; Ohura, M.; Ohsaki, Y.; Ohta, N. J.
Electrochem. Soc. 1990, 137, 2437-2443.
(
21) Briegleb, G. Angew. Chem. 1964, 76, 326-341.
(22) Dubois, D.; Kadish, K. M.; Flanagan, S.; Haufer, R. E.; Chibante,
L. P. F.; Wilson, L. J. J. Am. Chem. Soc. 1991, 113, 4364-4366.
(23) For instance, theoretical calculations of the molecular orbitals of
C60 show that the lowest LUMO (t1u-symmetry) and the LUMO+1 (t1g-
symmetry) are triply degenerated and the orbitals possessing the high
coefficients are distributed at the multisites of the spherical surface of C60;
see: (a) Reference 1a. (b) B u¨ hl, M.; Hirsch, A. Chem. ReV. 2001, 101,
Keeping in mind the above considerations, we will discuss
the kinetic substituent effects of fullerenes as compared to those
of DDQ, TCNE, and CA. The first reduction potentials (E vs
1
153-1183. (c) Hirsch, A.; Chen, Z.; Jiao, H. Angew. Chem., Int. Ed. 2000,
39, 3915-3917.
red
(24) (a) Fujio, M.; Nomura, H.; Nakata, K.; Saeki, Y.; Mishima, M.;
Kobayashi, S.; Tsuno, Y. Tetrahedron Lett. 1994, 35, 5005-5008. (b) Fujio,
M.; Morimoto, H.; Kim, H.-J.; Tsuno, Y. Bull. Chem. Soc. Jpn. 1997, 70,
1403-1411. (c) Nakata, K.; Fujio, M.; Saeki, Y.; Mishima, M.; Nishimoto,
K.; Tsuno, Y. J. Phys. Org. Chem. 1996, 9, 561-572, 573-582. (d) Nakata,
K.; Fujio, M.; Mishima, M.; Nomura, H.; Tsuno, Y.; Nishimoto, K. Bull.
Chem. Soc. Jpn. 1999, 72, 581-590.
(
18) Lobach, A. S.; Strelets, V. V. Russ. Chem. Bull. 2001, 50, 1593-
595.
19) (a) Fukui, K. Bull. Chem. Soc. Jpn. 1966, 39, 498-503. (b) Fukui,
K. Acc. Chem. Res. 1971, 4, 57-64.
1
(
2998 J. Org. Chem., Vol. 71, No. 8, 2006

1
,3-Dipolar Cycloaddition of Diphenyldiazomethanes
TABLE 2. Second-Order Rate Constants (k/M- s-1) for Reaction
1
of DDM (1f) with C60 in Various Solvents at 30 °C
10 k/M s-1
2
-1
Dπa
ETb
solvent
1
2
3
4
5
6
. benzonitrile
. anisole
. benzene
. toluene
. p-xylene
. mesitylene
17.8
13.5
12.0
11.5
8.23
4.60
-0.398
-0.043
0.000
0.394
0.846
1.36
41.5
37.1
34.3
33.9
33.1
32.9
a
Solvent soft basicity parameter, see ref 25a. The Dπ values for anisole,
p-xylene, and mesitylene were newly obtained according to ref 25a.
b
Solvent polarity parameter, see ref 17.
compared to that of TCNE and CA, although the reaction
solvents are different from each other, i.e., C60 and C70 (toluene),
CA (THF), TCNE, and DDQ (benzene). Strictly speaking, we
must compare the kinetic results which were obtained under
the same conditions since the solvents effects are often so
noticeable in the chemical reactions.1 Moreover, it is apparent
that the acceptor molecule suffers specific solvation, so-called
electron-pair donor/electron-pair acceptor (EPD/EPA) interac-
tion.25 When compared in the same reaction conditions at 30
FIGURE 4. Plots of log k(C60) at 30 °C versus solvent soft basicity
parameter Dπ.
7b
°
C in solvent “toluene”, the reaction of C60 with unsubstituted
DDM (1f) (point number 6 in Figure 2) was found to be ca. 3
and 70 times faster than those of TCNE2 and CA, respec-
tively, but only 9 times slower than the strongest acceptor
DDQ.15 Here, it is interesting to know the solvent dependency
of the rate of C60 because the π-donor solvents also seem to
play some role in the fullerene reactions. In fact, the rates for
the parent reaction of C60 with DDM (1f) were apt to moderately
decrease with the increase of π-electron donating ability of the
solvents (Table 2). The obtained k2 was decreased 4 times by
changing solvent from the less basic benzonitrile to the highly
basic mesitylene as represented by the solvent soft basicity
5a
26
FIGURE 5. C60: [6,6] bond (a) and [6,5] bond (b). C70: Four types
of [6,6] bonds (R-δ).
strain energy at the individual carbon atoms in fullerenes and
28
represented as the useful index of pyramidalization angle (θp).
The major driving force for addition reactions to fullerenes is
the resulting relief of strain energy accompanied by rehybrid-
2
5a
parameter Dπ. However, there is no appreciable relationship
between the rates and the solvent polarity (ET). The plots of
2
3 28a
ization (sp f sp ).
As seen in Figure 5, the C60 (Ih symmetry) has a single set
of equivalent atoms with θp ()11.64°) regardless of the bond
2
log k versus Dπ gave a fairly good regression equation (R )
0
.95) as shown in Figure 4. The negative slope (m) implies that
28a
type, [6,6] and [6,5] bond (types a and b). This strained angle
the π-donor aromatic solvents predominantly stabilize the
ground-state fullerenes, retarding the reaction. The C60 reaction
introduces a large strain energy into C60, which is estimated to
-1
be ca. 8 kcal mol per carbon. The [6,6] bond (1.401 Å) is
(
(
(
m ) -0.307) provided a more gentle slope than the DDQ
considerably shorter than the [6,5] bond (1.458) and thereby
the larger π-bond order is responsible for the exclusive additions
-1.14), TCNE (-1.00 by definition), and CA reactions
2
7
-0.598). Here it is noteworthy that these m values are
at the [6,6] bond as found for the various addition reactions of
red
correlated with the above-mentioned E
acceptors, though no corrections of E were made for the
solvent change; m ) -0.933E - 0.693 (R ) 0. 96) (for the
plots, see the Supporting Information). This equation proves
that the susceptibility of C60 to the solvent soft π-basicity
of the relevant
2-4
C60.
In recent years, several workers have performed the
red
theoretical investigation on the mechanism and regioselectivity
of these cycloadditions with typical dipoles, representing the
red
2
30
total agreement with the experimental observations.
On the other hand, C70 (D5h symmetry) has 8 types of
red
increases with the increase of E .
chemically different C-C bonds, half of which are [6,6] bonds
Factors Influencing the Reactivity of Fullerenes. As
described above, fullerenes exhibited the unexpected higher
reactivity despite of the poor electron affinity compared with
TCNE and CA. This may be attributed to two determining
3b
(
(
types R-δ). Among the [6,6] bonds, the most reactive bonds
type R) have slightly larger angles (θp ) 11.91° and 11.93°
3b
for the two conjunct carbons) than that of C60. They scatter at
the 10 apical positions in the pole sites (Figure 5). The second
reactive bonds (type â) possessing fairly reduced θp ()10.57°
for both conjunct carbons) and the shortest bond length are
2
factors: (1) the strain due to pyramidalization of the sp C-atoms
in the spherical structures of fullerenes28 and (2) the short [6,6]
2
9
bonds with larger π-bond orders and higher π-density. The
2
magnitude of deviation from the sp planarity is related to the
(28) (a) Haddon, R. C. Science 1993, 261, 1524-1550. (b) Haddon, R.
C. J. Am. Chem. Soc. 1997, 119, 1797-1798. The θp is obtained as θp )
(θσπ - 90)° from the π-orbital axis vector (POAV) analysis where the vector
makes equal angle (θσπ) to the three σ-bonds at the conjugated carbon atom.
(29) Baker, J.; Fowler, P. W.; Lazzeretti, P.; Malagoli, M.; Zanashi, R.
Chem. Phys. Lett. 1991, 184, 182-186.
(30) (a) Cases, M.; Duran, M.; Mestres, J.; Mart ´ı n, N.; Sol a` , M. J. Org.
Chem. 2001, 66, 433-442. (b) Kavitha, K.; Venuvanalingam, P. J. Org.
Chem. 2005, 70, 5426-5435.
(
25) (a) Oshima, T.; Arikata, S.; Nagai, T. J. Chem. Res. 1981, (S) 204-
05, (M) 2518-2536. (b) Fawcett, W. R. J. Phy. Chem. 1993, 97, 9540.
c) Isaacs, N. S. Physical Organic Chemistry; Longmann Science &
2
(
Technical: Essex, UK, 1995; Chapter 5, pp 220-222. (d) Chen, T.; Hefter,
G.; Marcus, Y. J. Solution Chem. 2000, 29, 201-216.
(
(
26) Oshima, T.; Nagai, T. Bull. Chem. Soc. Jpn. 1982, 55, 555-560.
27) Oshima, T.; Nagai, T. Tetrahedron Lett. 1985, 26, 4785-4788.
J. Org. Chem, Vol. 71, No. 8, 2006 2999

Oshima et al.
located at the 10 hexagon sides conjugated with the above type
R bonds. Other [6,6] bonds (γ and δ) which are located at the
equatorial belt regions are less reactive because of the more
reduced strain as well as the longer bond character. Indeed, the
addition reactions of C70 proceed via the preferential attack at
negative F values (-1.6 and -1.7) and diminished resonance
reaction constants (r ) 0.22 and 0.17), compared with the
corresponding reactions with stronger acceptors TCNE, DDQ,
and CA. The plots of log k(acceptor) versus log k(C60) as
reference also provided the excellent linear free energy relation-
ships and the magnitudes of slopes increased in the order TCNE
> DDQ > CA > C70 g C60. The poor substituent dependency
of C60 and C70 can be rationalized by the considerably reduced
electron affinity of fullerenes compared with the strong accep-
tors. On the other hand, the relative reactivity of these acceptors
toward the parent unsubstituted DDM increased in the order
DDQ > C60 g C70 > TCNE > CA. The unexpected higher
reactivity of fullerenes can be inferred by the inherently strained
3
1
32
the type R bond as found for the osmylation and 1,3-dipolar
cycloaddition of CH2CN2.4
a,33
For instance, in the addition of
diazomethane, the reactivity ratio of 13:2 for type R:type â bond
was deduced from the product ratio of the corresponding primary
adduct pyrazolines.4 Thus, it is likely that ca. 90% of the rate
may be attributed to the attack at the type R bond and the rest
to type â attack Assuming that DDMs also attack C70 with a
similar site-selectivity, we could imagine that the 1.5 times larger
reactivity of C60 relative to C70 is related to the excess total
numbers (30) of the reactive [6,6] bonds (type a) in C60 than
the less abundant type R bonds (10) in C70. Based on this
statistical reasons, it is interesting to note that the intrinsic
reactivity per reactive [6,6] bond is ca. 2 times higher for C70
than for C60. However, the possibility that the inherent higher
reactivity of C70 with the slightly lower LUMO is reversed by
the basic solvation by toluene could not be thoroughly ruled
out by considering the above-mentioned solvent effects.
More interestingly, Duran et al. theoretically analyzed the
Diels-Alder (DA) reaction of butadiene with C60 and C70 and
predicted slightly larger reactivity of C70 than C60 in the gas
a
2
pyramidalization of sp C-atoms, which strongly assists the
additions leading to the relief of strain. The kinetic solvent
effects elucidated the ground state stabilization of fullerenes
especially by the π-basic donor solvents as represented by the
good linear free energy relationship with the solvent basicity
parameter Dπ. Consequently, the useful information on the
chemical reactivity of fullerenes will provide beneficial insight
into the physicochemical understanding of reactivity as well as
further development of fullerene chemistry.
Experimental Section
Materials. All diphenyldiazomethanes (DDMs) were prepared
just before use by oxidation of the corresponding hydrazones with
yellow mercury oxide, according to the literature procedure. The
solid DDMs (1a-d, 1f, 1i, 1k, and 1m) were purified by
recrystallization from a hexanes-ether mixture. The melting points
were in accordance with the literature values.14 The oily DDMs
(1e, 1g, 1h, 1j, and 1l) were assessed by HPLC and NMR and
found not to be contaminated by the possible azines, carbene
dimmers, and benzophenones, as well as the unreacted hydrazones.
The C60 and a mixture with C70 were purchased from Tokyo Kasei
Kogyo Co., Ltd.. The C70 was separated from C60 by a preparative
HPLC with a Buckeyprep column (20 mm × 250 mm). Toluene
3
b
phase. By contrast, they obtained the reverse reactivity order
in toluene solution like the present case. This prediction is
consistent with an ca. 7 times enhanced rate for C60 than C70 in
the DA reaction with cyclopentadiene at 20 °C in toluene as
reported by Pang and Wilson.3a The higher reactivity of C60
was rationalized on the basis of the Arrenius data which were
very different between C60 and C70. The activation energy (Ea)
of C60 was by far smaller than half that of C70, whereas C70
recovered the handicap with a 2.3 times larger logarithmic
frequency factor A. However, as seen in Table 1, the present
kinetic data for the reaction of DDM (1f) with fullerenes in
various temperatures exhibited no essential difference in the
(predried for the synthetic procedure) was commercial and was used
without further purification. All other solvents were purified
according to the usual manner.34
q
-1
activation parameters, i.e., ∆H ) 39.1 and 40.1 kJ mol and
Kinetic Procedure. For the sake of simplicity, kinetic reaction
was carried out in first-order conditions, using a large excess of
q
-1
-1
∆S
) -134 and -134 J K
mol
for C60 and C70,
-
3
DDMs (25-50 equiv) with respect to C60 and C70 (1 × 10 mM).
Thus, a toluene solution (1 mL) of the excess DDMs was rapidly
introduced into a screw-capped test tube containing a 1 mL toluene
respectively. Our results suggest that the 1,3-dipolar cycload-
dition of DDM with C60 and C70 proceeds through a very similar
pathway on the potential energy surface. Current efforts are
directed toward the kinetics in the secondary addition of DDMs
to these primary products (fulleroids and methanofullerenes)
to elucidate the steric and electronic factors which control the
reactivity of fullerene derivatives. We also intend to clarify the
regioselective reactivity of the four types of [6,6] bonds of C70.
-
3
solution of C60 (1 × 10 mM), which is preheated at 30 °C in a
thermostated bath. The reaction solution was heated for a requisite
time at 30 °C and a constant aliquot (10 µL) was removed by a
microsyringe at regular time intervals for HPLC analysis as
described below. The progress of reactions was followed by
monitoring the decrease of the absorption peak intensity of C60 with
a HPLC instrument equipped with a UV operating at 280 nm, a
calculating integrator, and an analytical Buckey-prep column (4.6
mm × 250 mm). A hexane-toluene mixture (∼30 % v/v) was used
as a mobile phase.
Conclusions
Kinetics of model 1,3-dipolar cycloaddition of a series of
meta- and para-substituted diphenyldiazomethanes (DDMs) with
fullerens C60 and C70 have been investigated at 30 °C in toluene
in comparison with similar dipolar addition with TCNE, DDQ,
and chloranil (CA). It was found that C60 is ca. 1.5 times more
reactive than C70. The log k/k0 values were well correlated by
the Yukawa-Tsuno (Y-T) equation with relatively smaller
Supporting Information Available: Five figures for the typical
HPLC trace of reaction of 1a with C70 as well as the HPLC charts
showing the thermal rearrangement of the fulleroid to the metha-
1
nofullerene, H NMR charts exhibiting the thermal rearrangement
of fulleroid to the methanofullerene, plots of log k/k
0
vs Hammett
σ and Brown σ for C60 and C70, respectively, and then plots of
solvent dependency on the D parameter vs the reduction potential
E of acceptors. This material is available free of charge via the
+
(31) (a) Thilgen, C.; Herrmann, A.; Diederich, F. Angew. Chem., Int.
π
red
Ed. Engl. 1997, 36, 2268-2280. (b) Thilgen, C.; Diederich, A. Top. Curr.
Chem. 1999, 199, 135-171.
Internet at http://pubs.acs.org.
(
32) Hawkins, J. M.; Meyer, A.; Solow, M. A. J. Am. Chem. Soc. 1993,
JO052580U
1
15, 7499-7500.
33) Smith, A. B., III; Strongin, R. M.; Brard, L.; Furst, G. T.; Romanow,
W. J.; Owens, K. G.; Goldschmidt, R. C. J. Chem. Soc., Chem. Commun.
994, 2187-2188.
(
(34) Riddick, J. A.; Bunger, W. B. Techniques of Chemistry, Organic
SolVents; Weissberger, A., Ed.; John Wiley & Sons: New York, 1970.
1
3000 J. Org. Chem., Vol. 71, No. 8, 2006