Electrochemical and H/D-labeling study of oxazolino[60]Fullerene rearrangement

Article abstract of DOI:10.1021/jo1023798

Isomerization of an oxazoline cycle from a [6,6]-to [5,6]-junction on the C₆₀ sphere of dianionic [60]fullero-oxazoline (1^2-) during a 1,4-addition is studied by electrochemistry and a stepwise addition of PhCH₂Br and PhCD₂Br. Cyclic voltammerty of the in situ generated 1^2- shows a very unusual positive shift for the anodic peak corresponding to the oxidation of 1^2-, indicating that the C ₆₀ cage of dianionic 1 bears only one unit negative charge due to the heterolytic cleavage of the C₆₀-O bond. Further study with a stepwise addition of PhCH₂Br and PhCD₂Br, which are used to differentiate the aryl groups added at each step onto dianionic 1, shows explicitly there is an exclusive selectivity of the C-O bond for the ring-opening and ring-closure during the isomerization of the heterocycle. A reaction mechanism is proposed on the basis of the experimental results and computational calculations.(Figure Presented)

Full text of DOI:10.1021/jo1023798

pubs.acs.org/joc
Electrochemical and H/D-Labeling Study of Oxazolino[60]Fullerene
Rearrangement
Wei-Wei Yang, Zong-Jun Li, Fang-Fang Li, and Xiang Gao*
State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry,
Graduate School of the Chinese Academy of Sciences, Chinese Academy of Sciences, 5625 Renmin Street,
Changchun 130022, Jilin, China
xgao@ciac.jl.cn
Received December 2, 2010
Isomerization of an oxazoline cycle from a [6,6]- to [5,6]-junction on the C60 sphere of dianionic
[60]fullero-oxazoline (12-) during a 1,4-addition is studied by electrochemistry and a stepwise
addition of PhCH2Br and PhCD2Br. Cyclic voltammerty of the in situ generated 12- shows a very
unusual positive shift for the anodic peak corresponding to the oxidation of 12-, indicating that the
C60 cage of dianionic 1 bears only one unit negative charge due to the heterolytic cleavage of the
C60-O bond. Further study with a stepwise addition of PhCH2Br and PhCD2Br, which are used to
differentiate the aryl groups added at each step onto dianionic 1, shows explicitly there is an exclusive
selectivity of the C-O bond for the ring-opening and ring-closure during the isomerization of the
heterocycle. A reaction mechanism is proposed on the basis of the experimental results and com-
putational calculations.
Introduction
recorded during the early days of fullerene study,1a,b while
the unusual migration from [6,6]- to [5,6]-junction has been
shown until recently for the cases of Y3N@N-ethylpyrrolidino-
Isomerization of cyclic functional groups on fullerene
spheres between the [5,6]- and [6,6]-junctions has attracted
intense interest,1 since it reveals the unique reactivity of
fullerenes. Echegoyen and co-workers have reported an
electrochemically induced “moonwalking” of cyclopropane
rings over the C60 sphere, which involves the consequential
migration of a singly bonded malonate between the [5,6]- and
[6,6]-junctions.1c Migration from the less stable [5,6]-junction
to the more stable [6,6]-junction is typical, and has been
1f
C80 and aziridinofullerenes.1g The rearrangement of the
cyclic functional groups on fullerene spheres is closely re-
lated to the important retrocycloaddition reactions,2 where
the remaining singly bonded linkage during the isomerization
process is further cleaved and the addends are completely
(2) (a) Knight, B.; Martın, N.; Ohno, T.; Ortı, E.; Rovira, C.; Veciana, J.;
´ ´
Vidal-Gancedo, J.; Viruela, P.; Viruela, R.; Wudl, F. J. Am. Chem. Soc. 1997,
€
119, 9871–9882. (b) Kessinger, R.; Crassous, J.; Herrmann, A.; Ruttimann, M.;
Echegoyen, L.; Diederich, F. Angew. Chem., Int. Ed. 1998, 37, 1919–1922.
´ ´
(c) Martın, N.; Altable, M.; Filippone, S.; Martın-Domenech, A.; Echegoyen,
L.; Cardona, C. M. Angew. Chem., Int. Ed. 2006, 45, 110–114. (d) Filippone, S.;
(1) (a) Eiermann, M.; Wudl, F.; Prato, M.; Maggini, M. J. Am. Chem.
Soc. 1994, 116, 8364–8365. (b) Janssen, R. A.; Hummelen, J. C.; Wudl, F.
J. Am. Chem. Soc. 1995, 117, 544–545. (c) Li, Z.; Shevlin, P. B. J. Am. Chem.
ꢀ
ꢀ
ꢁ
Soc. 1997, 119, 1149–1150. (d) Kessinger, R.; Gomez-Lopez, M.; Boudon,
C.; Gisselbrecht, J. P.; Gross, M.; Echegoyen, L.; Diederich., F. J. Am. Chem.
Soc. 1998, 120, 8545–8546. (e) Hall, M. H.; Lu, H.; Shevlin, P. B. J. Am.
Barroso, M. I.; Martı
´
n-Domenech, A.; Osuna, S.; Miquel Sola, M.; Martı
n, M.; Altable, M.; Filippone, S.;
nez-Alvarez, R.; Suarez, M.; Plonska-Brzezinska,
M. E.; Lukoyanova, O.; Echegoyen, L. J. Org. Chem. 2007, 72, 3840–3846.
(f) Delgado, J. L.; Oswald, F.; Cardinali, F.; Langa, F.; Martın, N. J. Org. Chem.
´
n, N.
Chem.;Eur. J. 2008, 14, 5198–5206. (e) Martı
´ ´
Martın-Domenech, A.; Martı
´
ꢀ
Chem. Soc. 2001, 123, 1349–1354. (f) Rodrıguez-Fortea, A.; Campanera,
´
J. M.; Cardona, C. M.; Echegoyen., L.; Poblet, J. M. Angew. Chem., Int. Ed.
2006, 45, 8176–8180. (g) Tsuruoka, R.; Nagamachi, T.; Murakami, Y.;
Komatsu, K.; Minakata, S. J. Org. Chem. 2009, 74, 1691–1697.
´
2008, 73, 3184–3188. (g) Wang, G.-W.; Saunders, M.; Cross, R. J. J. Am. Chem.
Soc. 2001, 123, 256–259.
1384 J. Org. Chem. 2011, 76, 1384–1389
Published on Web 01/20/2011
DOI: 10.1021/jo1023798
r
2011 American Chemical Society

Yang et al.
JOCArticle
SCHEME 1. Illustrated Structures of 1 and 2a
aThe hydrogen atoms are omitted for clarity.
removed from the C60 sphere. However, different from the
retrocycloaddition reactions, where cyclic structures includ-
ing cyclopropane,2a,b pentagonal pyrrolidino,2c,d isoxazolino,2e
pyrazolino,2f and cyclopentadiene2g have been reported, the
isomerizations of organofullerenes are mostly limited to the
cyclopropane rings,1a-e with only few examples of other
cyclic structures.1f,g In addition, details of the isomerization
process remain elusive due to the difficulty of following the
reaction.
We have recently reported the synthesis of a five-membered
heterocyclic C60 derivative, 1,2-cyclic phenylimidate C60
(1, Scheme 1) and a cis-1 1,4-dibenzyl-2,3-cyclic phenylimi-
date C60 (2, Scheme 1).3 Notably, the pentacyclic oxazoline
ring undergoes an unusual migration from the [6,6]- to [5,6]-
junction from 1 to 2 upon 1,4-addition of benzyls. Such a
migration is quite intriguing because it is the first example of
isomerization involving an oxazolino structure for organo-
fullerenes, and it makes one wonder if there is a selectivity for
the migration.
Fukuzumi and co-workers have shown that the addition
of organic halides to dianionic C60 is a stepwise process,
where the reaction is initiated by a single-electron transfer
(SET) process, followed by an SN2 reaction.4 More impor-
tantly, it has been demonstrated that the addition of addends
can be controlled by taking advantage of the stepwise nature
of the reaction, which leads to the formation of organoful-
lerenes having different addends by adding various organic
halides with a time interval.4,5 Inspired by such results, we
employed the stepwise reaction of dianionic 1 with benzyl
bromide (PhCH2Br) and benzyl bromide-R,R-d2 (PhCD2Br)
to study the isomerization mechanism of the oxazoline ring
during the transformation from 1 to 2.
FIGURE 1. Cyclic voltammograms of 1 in benzonitrile containing
0.1 M TBAP at a scan rate of 0.1 V s-1. (a), (b), and (c) are for
neutral 1, while (d) is for the in situ generated dianionic 1.
3
in benzonitrile containing 0.1 M TBAP (tetra-n-butyl-
amonium perchlorate), and the stability of dianionic 1 was
examined by in situ cyclic voltammetry. Figure 1 shows the
cyclic voltammogram of the in situ generated dianionic 1
along with the cyclic voltammograms of the neutral 1 for
comparison. As for the neutral compound, the first redox is a
reversible one-electron transfer process with an E1/2 of -0.44 V,
which is almost identical to the value of C60, but different
from the typical 100-150 mV negative potential shift ob-
served for C60 derivatives,6 suggesting that the addend is
rather electronegative. As for the second reduction, it shows
a very interesting and unusual redox process. The compound
is cleanly reduced with an Epred at -0.86 V, which is also very
similar to that of C60. However, as the potential is reversed,
the anodic peak corresponding to the oxidation of this
dianionic species, which is usually about 60-100 mV more
positive with respect to the Epred, is completely missing in the
expected potential region. Instead, a small anodic peak
appears around -0.54 V, followed by a reversible anodic
peak corresponding to the oxidation of monoanionic 1 at
-0.39 V. It indicates that the heterocycle is still on the C60
cage of dianionic 1, and no decomposition of dianionic 1 to
C60 occurs, otherwise the anodic peak due to the oxidation of
dianionic C60 would appear around -0.80 V; however, a
heterolytic cleavage of one C60-X (X = O or N) bond occurs
when compound 1 receives two electrons, consistent with
previous report on C60 heterocycles upon reduction.7 The
heterolytic cleavage results in the bearing of only one unit of
negative charge on the C60 cage in dianionic 1, with the other
charge located on the heteroatom. Consequently, dianionic 1
will not be oxidized at a potential suitable for removing one
electron from the C60 core having two electrons, and the
oxidation peak around -0.54 V should correspond to the
electron removal from the addend. Such a charge redistribu-
tion has been proposed for anionic malonated fullerenes to
Results and Discussions
Electrochemical Study of 1 and 12-. Dianionic 1 was gen-
erated electrochemically by reducing 1 at -1.10 V vs SCE
(3) Zheng, M.; Li, F. -F.; Ni, L.; Yang, W.-W.; Gao, X. J. Org. Chem.
2008, 73, 3159–3168.
(4) (a) Subramanian, R.; Kadish, K. M.; Vijayashree, M. N.; Gao, X.;
Jones, M. T.; Miller, D. M.; Krause, K.; Suenobu, T.; Fukuzumi, S. J. Phys.
Chem. 1996, 100, 16327–16335. (b) Fukuzumi, S.; Suenobu, T.; Hirasaka, T.;
Arakawa, R.; Kadish, K. M. J. Am. Chem. Soc. 1998, 120, 9220–9227.
(5) (a) Allard, E.; Cheng, F.; Chopin, S.; Delaunay, J.; Rondeaub, D.;
Cousseau, J. New J. Chem. 2003, 27, 188–192. (b) Yang, W.-W.; Li, Z.-J.;
Gao, X. J. Org. Chem. 2010, 75, 4086–4094.
(6) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31, 593–601.
(7) Wang, G.-W.; Li, F.-B.; Zhang, T.-H. Org. Lett. 2006, 8, 1355–1358.
J. Org. Chem. Vol. 76, No. 5, 2011 1385

Yang et al.
JOCArticle
account for the retroaddition mechanism,8 but no such
electrochemical behavior has been recorded. It is noteworthy
that such a charge redistribution between the C60 cage and
addend is unlikely to occur without a heterolytic cleavage of
the C60-X bond, since the C60 core of 1- would be more
electronegative than the heterocycle as evidenced by the
reduction potentials (-0.86 V vs SCE for 1-/12-; -2.49 V
vs SCE for an oxazoline reduction9). The result therefore
asserts that there is a ring-opening via heterolytic cleavage
for dianionic 1, and the appendage is singly bonded to the
C60 cage in this intermediate, however, it is not known at this
stage which bond, C-O or C-N, is cleaved.
SCHEME 2. Reaction of PhCH2Br and PhCD2Br with Dia-
nionic 1
As the potential is further scanned to the third reduction as
shown in Figure 1c, multiple reduction waves appear in the
region, indicating that the remaining C60-Y (Y = O or N)
bond in dianionic 1 is further cleaved, and the heterocycle is
removed completely to afford C60. The observed anodic
peaks are likely due to the decomposed product of C60,
especially the one around -0.80 V, which is absent in
Figure 1b, confirming that 1 decomposes to C60 completely
upon acquiring three electrons. The decomposition is further
evidenced by the broadening of the first oxidation peak
(around -0.39 V vs SCE), which should consist of the
oxidation of both 1- and C60-. In the meantime, it corro-
borates that the dianionic 1 is indeed stable without decom-
position to C60 under the experimental conditions. Figure 1d
shows the cyclic voltammogram of the in situ generated
dianionic 1, which is similar to the CV shown in Figure 1b,
with no oxidation wave appearing around -0.80 V, consis-
tent with the formation of a singly bonded appendage
intermediate. Compared with Figure 1b, the anodic peak
around -0.53 V is much more prominent, and the height of
current is almost identical to that of the second reduction,
indicating that it is indeed the oxidation process countering
the one reduced at -0.86 V. As for the reason for a smaller
current of this anodic peak exhibited in Figure 1b, it is likely
caused by dianionic 1 diffusing away, the generation of
which was only in tiny amounts by cyclic voltammetry, from
the electrode. Notably, this peak was used previously as an
indicator to monitor the formation of C60 oxazolines even
though the origin of it was not fully understood at that
time.3,11 Due to the unusual positive shift of this wave
potential, it indicates that the oxidation likely originates
from a process where the electron is taken away from the
hetero addend directly rather than from the C60 cage. In
addition, judging by the reversibility of the subsequent
oxidation wave (1-/1), it indicates that compound 1 is fully
recovered at the stage of 1-, i.e., the C60-X bond is formed
Stepwise Addition of PhCH2Br and PhCD2Br and HMBC
NMR. Scheme 2 illustrates the reaction procedures of the
stepwise addition of PhCH2Br and PhCD2Br to dianionic 1,
where X and Y are used to denote the heteroatoms, and X
stands for the atom that is heterolytically cleaved from C60,
while Y stands for the one that remains bonded to the C60
core. Two different routes in which PhCH2Br and PhCD2Br
were added in opposite order were employed with the pur-
pose to obtain a complete picture of the reaction. Equivalent
amounts of PhCH2Br and PhCD2Br (molar ratio to 1 = 5:1)
were used to eliminate the concentration effect, and the
chemicals were added separately with a time interval of 3
min. The products obtained from route A and route B are
respectively designated as 2A and 2B.
Compound 2A or 2B was purified by HPLC eluted with
toluene over a Buckyprep column with an isolated yield of
∼50%. Only one fraction for each product was obtained,
even though both compounds consist of actually an isotopic
mixture containing both hydrogen and deuterium benzyls as
indicated by NMR characterizations. The identification of
the compound has been confirmed with positive ESI FT-ICR
MS (Electrospray Ionization Fourier Transform Ion Cyclo-
tron resonance Mass Spectrometry) and NMR spectroscopy
(see the Supporting Information). In the mass spectrum
of 2A, two [M þ H]þ peaks at 1022.15520 and 1024.16514
are shown, which correspond to the C81H20NO and C81H18-
D2NO; and in the mass spectrum of 2B, the [M þ H]þ peak
appears at 1024.16457, corresponding to C81H18D2NO.
The 1H and 13C NMR of 2A are in good agreement with
the data reported previously.3 Two AB quartets centered at
4.10 ppm (ABq, ΔνAB=324 Hz, JAB=13.2 Hz) and 4.06 ppm
(ABq, ΔνAB=258 Hz, JAB=13.2 Hz) are shown in the 1H
NMR. As expected, the intensity of one of the two AB
quartets is decreased due to the presence of excessive
PhCD2-, which facilitates the assignment of the correlated
doublets within each quartet. The AB quartet in the lower
field has shown a higher intensity, implying that the corre-
sponding benzyls should bear more methylene hydrogen
atoms, indicating that this quartet is likely to originate from
the benzyls added in the first step, since PhCH2Br was used in
the first step. The AB quartet in the upper field shows a
reduced intensity, and it is therefore likely to originate from
the benzyls that were added to dianionic 1 during the second
step. Notably, the intensity reduction is only about 25%, less
than the ideal situation where one of the AB quartets should
completely disappear, indicating that the reaction was not
carried out in a completely clean stepwise manner, and part
upon giving up an electron from 12-
.
The stability of dianionic 1 in PhCN was further confirmed
by oxidizing the species back to neutral at 0 V vs SCE. HPLC
analysis of the obtained crude product showed the full recovery
of 1 with only trace amount of C60 formed (see the Supporting
Information). The result is different from that obtained in
DMF,10 where the dianionic 1 undergoes a significant decom-
position, probably due to the difference of solvent polarities.
(8) Ocafrain, M.; Herranz, M. A.; Marx, L.; Thilgen, C.; Diederich, F.;
Echegoyen, L. Chem.;Eur. J. 2003, 9, 4811–4819.
(9) Ryan, C. M.; Kariv-Miller, E. Tetrahedron 1988, 44, 6807–6810.
(10) Li, F.-F.; Gao, X.; Zheng, M. J. Org. Chem. 2009, 74, 82–87.
(11) Li, F.-F.; Yang, W.-W.; He, G.-B.; Gao, X. J. Org. Chem. 2009, 74,
8071–8077.
1386 J. Org. Chem. Vol. 76, No. 5, 2011

Yang et al.
JOCArticle
FIGURE 3. Expanded HMBC NMR spectrum of product 2B in
CS2 with DMSO-d6 as the external lock.
FIGURE 2. Expanded HMBC NMR spectrum of product 2A in
CS2 with DMSO-d6 as the external lock.
protons resonating in the lower field compared with those
located besides the C-O bond.
of the first-step added benzyls may contain PhCD2-, while
the benzyls added in the second-step may also contain
PhCH2-. As a result, 2A is actually an isotopic mixture,
A control experiment was carried out following route B as
outlined in Scheme 2, to confirm the assignment of the
benzyls. The 1H and 13C NMR spectra of 2B are essentially
the same as those of 2A, except the intensity of the two AB
quartets are reversed, which is in good agreement with the
reversed order of the addition. The results verify that the AB
quartet in the lower field is due to the methylene protons of
benzyls added during the first step, while the AB quartet in
the upper field corresponds to the methylene protons of
benzyls added during the second step. Figure 3 shows the
HMBC NMR of product 2B, which exhibits the same H-C
couplings as those for 2A, confirming that the addition
of benzyls to dianionic 1 is a regioselective and step-dependent
process.
1
and the integration area ratio in the H NMR is there-
fore deviated from the theoretical value. Fortunately, the
observed 25% peak reduction is significant enough to discern
the benzyls with specific addition step. In the 13C NMR of
2A, the attention is paid to the sp3 C60 region, since this is
the region that can be used to distinguish the structure of
3
the compound with HMBC NMR via the JCH couplings
between the methylene protons and the carbon atoms
bonded to heteroatoms.11 It shows two resonances at 97.1
and 90.8 ppm, corresponding to the sp3 C60 carbon atoms
bonded to the oxygen and the nitrogen, respectively.3,7,11,12
Resonances due to the C60 sp3 carbon atoms bonded to the
benzyls appear at 61.2 and 62.2 ppm, consistent with pre-
vious results.3
Figure 2 shows the expanded HMBC NMR for 2A. The
spectrum shows that the AB quartet corresponding to the
methylene protons of the benzyls involved in the first-step
addition (higher intensity) is correlated with the C60 sp3
carbon atoms at 62.2 and 90.8 ppm, while the AB quartet
corresponding to the methylene protons of the benzyls in-
volved in the second-step addition (lower intensity) is corre-
lated with the C60 sp3 carbon atoms at 61.2 and 97.1 ppm. The
result indicates that in compound 2A the first-added benzyl
group is selectively located adjacent to the C-N bond,
while the second-added benzyl is selectively positioned next
to the C-O bond. The benzyl assignment by HMBC
is consistent with previous calculations,10 which predict
Reaction Mechanism. Since the addend in dianionic 1 is
singly bonded to the C60 core, the addition site for the first
benzyl is only available at either the o- (a [6,6]-bond) or
p-position with respect to the existing bond4a in considering
that the appendage is not very bulky.13 If the addition of the
first benzyl took place at the p-position, it would lead to the
cleavage of the remaining C60-Y bond for the subsequent
addition of the second benzyl in order to form product 2,
which would result in the complete removal of the hetero-
cycle. The experiment starting from anionic 1,4-(PhCH2)2C60
has shown that the obtained oxazoline product has a
1
different structure from that of 2 (see the H NMR in the
Supporting Information), indicating that compound 2 will
not be formed via the retroaddition pathway. It is therefore
only possible that the first benzyl will be added to C60 at the
o-position with respect to the existing C-Y bond. In this
case, the heteroatom correlated with the first-added benzyl
in HMBC is the one that remains bonded to the C60 sphere
in dianionic 1. Accordingly, the N atom is the one that is
bonded to the C60 sphere during the reaction, while the
that the N atom has a greater C-H X bonding interac-
tion with the adjacent methylene protons than the O atom,
and would therefore result in the neighboring methylene
3 3 3
(12) (a) Li, F.-B.; Liu, T.-X.; Wang, G.-W. J. Org. Chem. 2008, 73, 6417–
6420. (b) Li, F.-B.; Liu, T.-X.; You, X.; Wang, G.-W. Org. Lett. 2010, 12,
3258–3261. (c) Li, F.-B.; You, X.; Wang, G.-W. Org. Lett. 2010, 12, 4896–
4899.
(13) Kokubo, K.; Arastoo, R. S.; Oshima, T.; Wang, C.-C.; Gao, Y.;
Wang, H.-L.; Geng, H.; Chiang, L. Y. J. Org. Chem. 2010, 75, 4574–4583.
J. Org. Chem. Vol. 76, No. 5, 2011 1387

Yang et al.
JOCArticle
FIGURE 4. Calculated NBO charge distributions of (a) dianionic 1, and (b) reaction intermediate with one benzyl bonded to C60 next to the
C-N bond. See the Supporting Information for the charge distributions on the other side of the C60 sphere. The orange ball stands for the
nitrogen atom, and the purple ball stands for the oxygen atom.
SCHEME 3. Proposed Mechanism of the Rearrangement Reaction of Oxazoline Ring on C60 Surface
C60-O bond is cleaved heterolytically with the oxygen
atom bearing a negative charge. The result is quite reason-
able in considering that the oxygen anion is more stable
than the nitrogen counterpart since the oxygen atom is
more electronegative.
To obtain a better understanding of the reaction, a com-
putational calculation at the HF/6-31G theoretical level with
the Gaussian 03 program was carried out. Previous work has
shown that the reaction of charged C60 species is usually a
charge-directed process, where the addition takes place at the
carbon atom with the largest charge density.4a,11,14
Figure 4a shows the natural bond orbital (NBO) charge
distributions of the optimized dianionic 1, where the oxygen
atom is positioned away from the C60 [6,6]-junction and stays
over a pentagon in order to leave the reaction site open for
the coming benzyl. The result shows that the carbon atom of
C60 at the o-position of the C-N bond indeed has a much
greater negative charge (-0.139) than other C60 carbon atoms.
In addition, the oxygen atom cleaved from the C60 sphere bears
a significant amount of negative charge, in agreement with the
experimental observations. It is noteworthy that, even though
the oxygen anion carries a greater amount of negative charge
as predicted by the calculations, no evidence has been found
experimentally to show that the oxygen anion is involved in the
SN2 reaction with the benzyl bromide, which can be rationa-
lized by the greater stability of the oxygen anion due to the
strong electronegativity of the oxygen atom.
After the addition of the first benzyl, the number of
negative charges on the intermediate is reduced to one. Since
the C60 cage is more electron affinitive than the singly bonded
hetero appendage as shown in the cyclic voltammogram
(Figure 1d, Epox = -0.38 V when an electron is removed
from the C60 core for monoanionic 1; Epox=-0.53 V when
an electron is removed from the singly bonded hetero
appendage), the remaining one unit charge in the intermedi-
ate will then be surrendered from the hetero appendage to the
C60 core via a ring-closure pathway. Consequently, the
C60-O bond is reconstructed but at the [5,6]-junction, since
the carbon atom for the [6,6]-bond has been occupied by the
first added benzyl. Figure 4b shows the NBO charge dis-
tributions of this reaction intermediate with one benzyl
bonded to C60 next to the C-N bond. It shows that the
carbon atom, which is at the p-position to the added benzyl
and is next to the C-O bond, has the greatest negative charge
(-0.235), predicting that this is the reaction site for the
benzyl addition in the second step, in agreement with
experimental results. It is noteworthy that the observed
ring-opening and ring-closure processes for C60 oxazoline
compound undergo the same pathway for the ring-opening
of 2-oxazolines.15 The proposed transformation mecha-
nism for the oxazoline cycle on C60 surface is illustrated
in Scheme 3.
(15) (a) Meyers, A. I.; Himmelsbach, R. J.; Reuman, M. J. Org. Chem.
1983, 48, 4053–4058. (b) Wilson, K. J.; Sabat, M.; McGarvey, G. J. J. Org.
Chem. 1993, 58, 6180–6181. (c) Lee, S.-H.; Bok, J.; Qi, X.; Kim, S. K.; Leed,
ꢀ
Y.-S.; Yoon, J. Tetrahedron Lett. 2007, 48, 7309–7312. (d) Sanchez, R. S.;
Zhuravlev, F. A. J. Am. Chem. Soc. 2007, 129, 5824–5825.
(14) Feng, L.; Nakahodo, T.; Wakahara, T.; Tsuchiya, T.; Maeda, Y.;
Akasaka, T.; Kato, T.; Horn, E.; Yoza, K.; Mizorogi, N.; Nagase, S. J. Am.
Chem. Soc. 2005, 127, 17136–17137.
1388 J. Org. Chem. Vol. 76, No. 5, 2011

Yang et al.
JOCArticle
Conclusions
(1C), 142.2 (1C), 142.0 (1C), 141.6 (1C), 141.55 (1C), 141.3 (1C),
141.2 (2C), 141.0 (1C), 139.3 (1C), 139.0 (1C), 137.6 (1C), 136.8
(1C), 136.4 (1C), 135.9 (1C), 135.8 (1C), 134.5 (1C, Ph), 134.2
(1C, Ph), 133.7 (1C, Ph), 131.5 (1C, Ph), 131.5 (2C, Ph), 131.3
(1C, Ph), 131.3 (1C, Ph), 131.3 (1C, Ph), 128.3 (2C, Ph), 128.2
(2C, Ph), 127.5 (2C, Ph), 127.4 (2C, Ph), 126.6 (1C, Ph), 126.3
(1C, Ph), 97.1 (1C, sp3, C-O), 90.8 (1C, sp3, C-N), 62.2 (1C,
sp3, C-CH2Ph), 61.2 (1C, sp3, C-CH2Ph), 46.2 (1C, CH2), 45.3
(1C, CH2).
Synthesis of 2B (Route B). The procedures for the synthesis
are similar to those of route A, except that 5 equiv of PhCD2Br
(35.4 μL) was added into the system first, and the reaction was
allowed to proceed for 3 min, followed by adding 5 equiv of
PhCH2Br (35.4 μL). Compound 2B was purified by HPLC
eluted with toluene over a Buckyprep column with an isolated
yield of ∼50%.
The rearrangement reaction of oxazoline ring from [6,6]-
to [5,6]-junction on C60 surface upon 1,4-addition of benzyl
bromide has been explored by using cyclic voltammetry and
stepwise addition PhCH2Br and PhCD2Br, where the ben-
1
zyls with specific addition step are discerned by H NMR.
The results show that the C-O bond is heterolytically
cleaved in dianionic 1, and there is an exclusive selectivity
of the C-O bond for the ring-opening and ring-closure
during the isomerization process of the heterocycle, which
is consistent with the pathway for the ring-opening of 2-
oxazolines. The results have extended the scope of cyclic iso-
merization reactions of organofullerenes and are helpful to gain
a better understanding on the chemistry of fullerenes.
Spectral characterization of 2B: positive ESI FT-ICR MS:
C81H17D2NO, m/z [M þ H]þ calcd 1024.1665, found 1024.1646;
1H NMR (600 MHz, in CS2, DMSO-d6 was used as the external
lock solvent) δ 8.13 (d), 7.40 (t), 7.35 (t), 7.10 (d), and from 6.95
to 6.80 (m), 4.10 (ABq, ΔνAB = 324 Hz, JAB = 13.2 Hz), 4.06
(ABq, ΔνAB = 258 Hz, JAB = 13.2 Hz); 13C NMR (150 MHz,
CS2/ DMSO-d6) δ 161.6 (1C, CdN), 152.6 (1C), 152.4 (1C),
148.7 (1C), 148.5 (1C), 148.4 (1C), 148.2 (1C), 147.5 (1C), 146.6
(1C), 146.6 (1C), 146.5 (2C), 146.3 (1C), 145.9 (1C), 145.6 (1C),
145.5 (2C), 144.9 (2C), 144.8 (1C), 144.7 (2C), 144.6 (1C), 144.2
(1C), 144.1 (1C), 144.0 (1C), 144.0 (2C), 143.8 (1C), 143.8 (1C),
143.7 (2C), 143.6 (1C), 143.4 (2C), 143.3 (1C), 143.0 (1C), 142.9
(1C), 142.8 (1C), 142.6 (1C), 142.4 (1C), 142.3 (1C), 142.2 (1C),
142.0 (1C), 141.6 (1C), 141.6 (1C), 141.3 (1C), 141.2 (2C), 141.0
(1C), 139.3 (1C), 139.0 (1C), 137.6 (1C), 136.8 (1C), 136.4 (1C),
135.9 (1C), 135.8 (1C), 134.5 (1C, Ph), 134.2 (1C, Ph), 133.7 (1C,
Ph), 131.5 (1C, Ph), 131.5 (1C, Ph), 131.4 (1C, Ph), 131.3 (1C,
Ph), 131.3 (1C, Ph), 128.3 (2C, Ph), 128.2 (2C, Ph), 127.5 (2C,
Ph), 127.4 (2C, Ph), 126.6 (1C, Ph), 126.3 (1C, Ph), 97.1 (1C, sp3,
C-O), 90.8 (1C, sp3, C-N), 62.2 (1C, sp3, C-CH2Ph), 61.2 (1C,
sp3, C-CH2Ph), 46.2 (1C, CH2), 45.3 (1C, CH2).
Experimental Section
Synthesis of 2A (Route A). Typically, 50 mg (69.4 μmol) of 1,
which was obtained via the method reported previously,3 was
electroreduced at -1.10 V vs SCE in 50 mL of freshly distilled
PhCN solution containing 0.1 M TBAP under a nitrogen atmo-
sphere. The potentiostat was switched off when the theoretical
number of coulombs required for a full conversion of 1 to 12-
was reached. Then 5 equiv of PhCH2Br (35.4 μL) was added into
the solution, and the reaction was allowed to proceed for 3 min,
followed by adding 5 equiv of PhCD2Br (35.4 μL). The reaction
was allowed to proceed for about 2 h with stirring. The solvent
was removed with a rotary evaporator under reduced pressure
after the reaction was finished, and the residue was washed
with methanol to remove TBAP before further purification. The
slurry was separated by filtration, and the residue was put into
toluene and sonicated. The soluble part was further purified by
HPLC eluted with toluene over a Buckyprep column. Com-
pound 2A was obtained with an isolated yield of ∼50%.
Spectral characterization of 2A: positive ESI FT-ICR MS:
C81H17D2NO, m/z for [M þ H]þ calcd 1024.1665, found
1
1024.1651; H NMR (600 MHz, in CS2, DMSO-d6 was used
Acknowledgment. The work was supported by the Na-
tional Natural Science Foundation of China (Grant No.
20972150), and the Solar Energy Initiative of the Chinese
Academy of Sciences (Grant No. KGCX2-YW-399þ9).
as the external lock solvent) δ 8.13 (d), 7.40 (t), 7.36 (t), 7.10 (d),
and from 6.95 to 6.80 (m), 4.11 (ABq, ΔνAB = 324 Hz, JAB
=
13.2 Hz), 4.06 (ABq, ΔνAB = 258 Hz, JAB = 13.2 Hz); 13C
NMR (150 MHz, CS2/DMSO-d6) δ 161.6 (1C, CdN), 152.6
(1C), 152.4 (1C), 148.7 (1C), 148.5 (1C), 148.4 (1C), 148.2 (1C),
147.5 (1C), 146.6 (1C), 146.5 (1C), 146.5 (2C), 146.3 (1C), 146.0
(1C), 145.6 (1C), 145.5 (2C), 144.9 (2C), 144.8 (1C), 144.7 (2C),
144.6 (1C), 144.2 (1C), 144.1 (1C), 144.0 (1C), 144.0 (2C), 143.8
(1C), 143.78 (1C), 143.7 (2C), 143.6 (1C), 143.4 (2C), 143.3 (1C),
143.0 (1C), 142.9 (1C), 142.8 (1C), 142.6 (1C), 142.4 (1C), 142.3
Supporting Information Available: General methods, char-
acterizations of 2A and 2B, 1H NMR of the oxazoline product
obtained from the reaction of anionic 1,4-(PhCH2)2C60, and
calculation details. This material is available free of charge via
the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 76, No. 5, 2011 1389