Cu(I)-mediated regioselective tri-addition of Grignard reagent to [70]fullerene. Synthesis of indenyl-type metal ligand embedded into graphitic structure

Article abstract of DOI:10.1039/b202130g

A variety of [70]fullerene derivatives bearing three organic groups C₇₀R₃H (1, R = aryl, methyl) have been synthesized in 90-99% isolated yield by the reaction of [70]fullerene with a Grignard reagent in the presence of CuBr·S(CH

Full text of DOI:10.1039/b202130g

View Article Online / Journal Homepage / Table of Contents for this issue
Cu(I)-mediated regioselective tri-addition of Grignard reagent to
[70]fullerene. Synthesis of indenyl-type metal ligand embedded into
graphitic structure
Masaya Sawamura,{ Motoki Toganoh, Hitoshi Iikura, Yutaka Matsuo, Atsushi Hirai{
and Eiichi Nakamura*
Department of Chemistry, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033,
Japan. Tel: 181-3-5800-6889; Fax: 181-3-5800-6889;
E-mail: nakamura@chem.s.u-tokyo.ac.jp
Received 28th February 2002, Accepted 24th April 2002
First published as an Advance Article on the web 10th May 2002
A variety of [70]fullerene derivatives bearing three organic groups C70R3H (1, R ~ aryl, methyl) have been
synthesized in 90–99% isolated yield by the reaction of [70]fullerene with a Grignard reagent in the presence
of CuBr?S(CH3)2. Deprotonation of C70R3H with a metal alkoxide gave the corresponding metal complex
M(g5-C70R3) [M ~ K (2), Tl (3)]. X-ray crystallographic analysis as well as theoretical study revealed the
2
nature of the 68p-electron ligand, C70R3
.
Introduction
Covalent functionalization of carbon clusters continues to
attract the interest of synthetic and materials chemists.1,2 With
all the intense research activities in the area, the major problem
still remains the poor yield of the synthetic transformations. In
order for carbon cluster science to become truly useful core
technology in the 21st century, the synthetic transformations
must be quantitative based on the starting carbon cluster used
for the reaction, and must be achieved in a simple and efficient
procedure without production of noxious waste. Unfortu-
nately, the current state of art is far from satisfactory. With
virtually no exceptions, the reported synthetic transformations
of carbon clusters produce the desired product together with
over-reacted products, recovered starting material, regio- and
stereoisomers, and therefore the product yield never
approaches 100%.
Some time ago, we reported that [60]fullerene reacts with a
phenyl Grignard reagent in the presence of a Cu(I) salt to form
a penta-addition product C60(C6H5)5H in 99% isolated yield
based on the analytically pure product [(eqn. (1)].3 The reaction
can be easily performed on a 10 g scale in an ordinary chemical
laboratory. This transformation represented, to our knowl-
edge, the only chemical transformation at that time that
enabled covalent functionalization of carbon clusters in
quantitative yield. The reaction was proven to be applicable
to the synthesis of variety of C60R5H compounds (R ~ aryl,
1-alkenyl and methyl) often in quantitative yield.3–5 The penta-
adduct C60R5H serves as a precursor to the pentahaptoful-
lerene metalcomplexes M(g5-C60R5), which are useful for
catalysis6 and for construction of nanostructures in water and
on the surface.7
ð1Þ
With the intention of obtaining higher fullerene analogues of
the penta-adduct, we examined the reaction of [70]fullerene
with a Grignard reagent in the presence of a Cu(I) salt. The
product was, however, not the expected penta-adduct (4 or 5)
but a tri-adduct C70R3H (1) (Scheme 1). Thus, the addition
reaction took place 100% selectively on the side of the football-
shaped fullerene rather than at the pointed end of the molecule,
and stopped precisely after tri-addition. After considerable
optimization of the reaction conditions, the tri-addition
reaction was made to proceed in 90–99% isolated yield for
methyl and aryl Grignard reagents. The tri-arylation reactions
were essentially quantitative, while the tri-methylation reaction
gave someside products. The reaction provided the second
example of quantitative covalent functionalization reaction of
carbon clusters,8 which was followed later by the third
example, quantitative tetra-amination reaction of [60]full-
erene.9
The tri-adduct 1 was deprotonated with a metal alkoxide to
obtain the corresponding metal complex M(g5-C70R3) [M ~
K (2), Tl (3)]. Single crystal X-ray crystallographic analysis of
the Tl complex 3a revealed that the compound possesses a
metal g5-indenyl structure flanked by three sp3 carbons. The
C70R3 anion ligand represents an unprecedented 68p-electron
ligand to a metal cation. The synthetic route to this class of
organofullerene is complementary to the stepwise synthesis of
C60R1 R2H from [60]fullerene,10 and may acquire significant
utility once the price of [70]fullerene goes downsignificantly.
Brief mechanistic studies indicated that the premature
{Present address: Department of Chemistry, Hokkaido University,
Sapporo, Hokkaido 060-0810, Japan. Fax: 181-11-706-4924; Tel: 181-
11-706-4924; E-mail: sawamura@sci.hokudai.ac.jp
{Present address: Department of Chemistry, Hokkaido University,
Sapporo, Hokkaido 060-0810, Japan. Fax: 181-11-706-4920; Tel: 181-
11-706-2716; E-mail: ahirai@sci.hokudai.ac.jp
2
DOI: 10.1039/b202130g
J. Mater. Chem., 2002, 12, 2109–2115
This journal is # The Royal Society of Chemistry 2002
2109

View Article Online
ð2Þ
Initial optimization was carried out for the copper reagent
(16 equiv.) derived from 4-CF3C6H4MgBr and CuBr?(CH3)2.
When 1,2-Cl2C6H4 was used instead of toluene, the reaction
was faster but low-yielding (7–32%) to produce numerous side
products as detected by HPLC. However, by lowering the
reaction temperature from 25 uC to 10 uC, the desired product
1a was obtained in 94% isolated yield without being accom-
panied by the mono-adduct 6a. Thus, a 1,2-Cl2C6H4 solution of
[70]fullerene was added to an aryl copper reagent (16 equiv.)
and the reaction mixture was stirred for an hour at 278 uC then
for 10 hours at 10 uC (Method A). After aqueous work-up, the
desired product was extractedwith toluene and purified by
precipitation.
For yet unknown reasons, some organocopper reagents
afforded the tri-adduct in better yield (e.g., less side products)
if the fullerene and the copper reagent were mixed at room
temperature rather than at –78 uC, and the reaction mixture
was let stand further at that temperature (Method B). Because
of a competing thermal homocoupling reaction of the copper
reagent, further excess of the copper reagent (20–30 equiv.) is
needed. As in the reaction of [60]fullerene,5 the reaction of the
methyl copper reagent necessitates the use of 1,3-dimethyl-2-
imidazolidinone in order to obtain a maximum product yield
(90% yield).
Table 1 summarizes the results of the tri-addition of various
Grignard reagents to [70]fullerene in the presence of
CuBr?S(CH3)2. Method A was applied for the cases of R ~
4-CF3C6H4, C6H5, 4-ClC6H4 and 1-naphthyl (entries 1–4). The
products were obtained in 93–95% isolated yield. When
applicable, Method B afforded the product of better purity
because of less side products derived from the fullerene (entries
5–7). For instance, the biphenyl adduct was obtained in 99%
yield. Tri-addition of CH3MgBr afforded, in 90% yield,
C70(CH3)3H, which is much less sterically hindered than
C70Ar3H and will serve as a useful ligand for transition metal
organometallics.
The structural identification of the unexpected tri-adduct
initially rested on spectroscopic analyses. The mass spectrum
of the tri-adduct 1a exhibits a parent peak at m/z ~ 1276
(FAB-MS, M1) that corresponds to the molecular formula
C70(4-CF3C6H4)3H. In the 1H NMR analysis, six doublet peaks
in the aromatic proton region appeared (d ~ 7.92, 7.84, 7.71,
7.66, 7.58–7.52: two doublets overlap on each other and
Scheme 1
quenching of the reaction gives a mono-adduct 6, which could
be converted further to the tri-adduct upon treatment with the
same copper reagent. In this article, we report a full account of
the synthesis of the C70R3H compounds.
Results and discussion
Expecting the formation of the penta-adduct 4 or 5 in ana-
logy with the formation of C60R5H from [60]fullerene, we
treated [70]fullerene with an organocopper reagent prepared
in situ from a one-to-one mixture of 4-CF3C6H4MgBr and
CuBr?S(CH3)2 under the conditions optimized for the [60]full-
erene reaction [eqn. (2)].3 A pre-cooled toluene solution of
[70]fullerene at 278 uC was transferred through a cannula to
a magnetically stirred mixture of the organocopper reagent
(16 equiv.) in THF at 278 uC. The resulting black slurry was
warmed to 25 uC. The reaction was monitored with HPLC,
which indicated complete consumption of [70]fullerene after
24 hours, and wasstopped by addition of aq. NH4Cl. After
purification of the organic products with preparative HPLC,
the tri-adduct C70(4-CF3C6H4)3H (1a) and the mono-adduct
C70(4-CF3C6H4)H (6a) were isolated in 43% yield and 47%
yield, respectively. None of the expected penta-adducts 4 and 5
were detected in the mixture. Unlike [70]fullerene itself, the tri-
addition product readily dissolves in toluene, carbon disulfide,
and chloroform.
Table 1 Three-fold organocopper addition to [70]fullerene with a
variety of Grignard reagents
Entry
R
Method
Yielda
1
2
3
4
4-CF3C6H4
C6H5
4-ClC6H4
1-naphthyl
C6H5
A
A
A
A
B
B
B
94% (1a)
93% (1b)
94% (1c)
95% (1d)
95% (1b)
99% (1e)
90% (1f)
5
6
4-(C6H5)C6H4
CH3
7b
aThe present yields were calculated on the basis of the purity of the
product and that of the starting [70]fullerene. b1,3-Dimethyl-2-imida-
zolidinone was used as additive.
2110
J. Mater. Chem., 2002, 12, 2109–2115

View Article Online
integrated twice as much.) indicating the existence of three non-
equivalent aryl groups. The proton directly attached to the
fullerene core appears at d ~ 4.40 ppm as a singlet. The 13C
NMR and 2D HMBC NMR spectra of 1a are also consistent
with the assigned C1 structure.
Deprotonation of 1a with KOC(CH3)3 afforded a potassium
complex K[C70(4-CF3C6H4)3] (2a). The singlet peak at d ~
1
4.40 ppm disappeared. The 13C NMR and H NMR spectra
became dramatically simplified, indicating much higher
symmetry of the potassium compound. The structure of 1a
was unequivocally determined by X-ray crystallographic
analysis of the corresponding thallium complex Tl[C70(4-
CF3C6H4)3] (3a) as described later.
Fig. 2 (a) A part of the optimized structure of K(C70H3) at the HF/
˚
3-21G(*) level. Bond lengths are in A. K–C bond lengths: K–C(1), 3.06;
K–C(2), 2.99; K–C(3), 2.92. (b) HOMO surface of K(C70H3).
Once the structure of 1a was solved, the structure of the
mono-adduct 6 followed, since 1a and 6 can be chemically
correlated. When the isolated C70(4-CF3C6H4)H (6a) was
treated again with the copper reagent, it was further converted
to the tri-adduct 1a whose structure was known from the
crystal structure of the corresponding thallium complex 3a
(vide infra).
analyzed in some detail, since the X-ray analyses are not good
enough to discuss the C–C bond lengths. Molecular orbitals of
the optimized structure were also examined.
As the potassium complex 2 did not give X-ray quality
crystals, we examined the structure of the model compound
K(C70H3) as optimized at the HF/3-21G(*) level12 under Cs
symmetry (total energy ~ 23233.615841 hartree,8 Fig. 2). The
potassium atom is bonded to the [70]fullerene sphere in an g5
coordination mode. The g5-moiety of the complex shows
relatively long C–C bond length of the [5,6] junction [C(3)–
The tri-adduct 1a was converted to the thallium complex
Tl[C70(4-CF3C6H4)3] (3a) and recrystallized from 1,2-Cl2C6H4–
hexane to obtain crystals suitable for X-ray crystallographic
analysis. While the data from a sample used in the preliminary
report8 did not successfully converge, we recently were able to
obtain reasonable data for a deep red single crystal of 3a?(1,2-
Cl2C6H4)2. The crystal structure of 3a with a space group P21/n
shows that the thallium atom is located equidistant from the
three aryl groups and bonded to the pentagon flanked by the
three sp3 carbons (Fig. 1). This structure was fully consistent
with the result of ab initio study described below. One aryl ring
of the three 4-CF3C6H4 groups is oriented perpendicular to
another aryl ring. In the light of the short distance of C(ortho)–
˚
C(3’); 1.44 A] and adjacent six-membered ring C–C bonds
˚
[C(3)–C(4) and C(3’)–C(4’); 1.43 A], which are the features
commonly observed for metal indenyl complexes (videinfra).
The molecular orbitals of the complex are essentially a hybrid
of an indenyl anion and [70]fullerene. As shown in Fig. 2b, the
HOMO (A@, 26.66 eV) of K(C70H3) is localized mainly at the
indenyl moiety but extends also over the whole fullerene cage.
The structure of the model compound Li(C70H3) was also
optimized at the HF/3-21G(*) level under Cs symmetry (total
energy ~ 22644.860708 hartree,8 Fig. 3). The lithium atom is
bonded to the [70]fullerene sphere in an g5 coordination mode.
The indenyl moiety of Li(C70H3) shows striking resemblance to
the X-ray crystal structure of Li(g5-indenyl)(temda).13
In view of the very nature of the reaction, the mechanism
likely involves a radical species or a radical-like species (e.g.,
organocopper(II) species). In the light of the mechanistic
correlation between the mono- and tri-adduct described above,
and the parallel studies carried out on the formation of C60R5H
from C60,10 we can draw a possible reaction pathway from
[70]fullerene to C70R3H under assumption of an organome-
tallic mechanism (Scheme 2). The initial addition of an anionic
R group in a putative diorganocuprate species takes place at
the most strained sp2 carbon center indicated as A, which
generates a 1,2-adduct 7. The 1,2-adduct 7 may equilibrate to 8
via1,3-metal shift. Removal of one electron from the copper
atom in 8 (possibly by a Cu(I) specied, fullerene or a fullerene
…
˚
C(ipso) [2.77 A], these two aryl group may have weak
H
…
C–H p interaction with each other. The third aryl ring is free
from intramolecular interactions. The lengths of the five
thallium–carbon bonds are nearly equal to each other [2.88–
˚
˚
2.93 A, av. 2.91 A] and the complex is judged to be an
g5-bonded Tl complex. These bond lengths are comparable
5
˚
˚
with those of Tl[g -C60(C6H5)5] [2.83–2.90 A, av. 2.87 A] that
we previously reported.3The structural data therefore indicate
that the [70]fullerene tri-adduct is a member of new indenyl-
type ligand incorporated into a graphitic structure.
Theoretical study was performed on model compounds
K(C70H3) and Li(C70H3).11 The optimized structures were
Fig. 3 (a) A part of the optimized structure of Li(C70H3) at the HF/
˚
˚
Fig. 1 Molecular structure of 3a?(1,2-Cl2C6H4)2. Solvent molecules are
omitted for clarity.
3-21G(*) level. Bond lengths are in A. Li–C bond lengths (A): Li–C(1),
2.24; Li–C(2), 2.20; Li–C(3), 2.18. (b) HOMO surface of Li(C70H3).
J. Mater. Chem., 2002, 12, 2109–2115
2111

View Article Online
(column: Buckyprep, 4.6 6 250 mm, Nacalai tesque; flow
rate: 1.0 to 3.0 ml min21; eluent: toluene/2-propanol ~ 7 : 3;
detector: SPD-M10Avp, Shimadzu). Purity of products was
determined by integrated area ratios at 350 nm absorption on
HPLC analysis. The present yields were calculated on the basis
of this purity and the purity of the starting [70]fullerene.
Common organic solvents as well as aqueous solutions were
deoxygenated by freeze-thaw cycles (over three times) or by
bubbling nitrogen (over 30 min).
All 1H NMR spectra were taken at 400 MHz (JEOL EX-400)
or 500 MHz (JEOL a-500), and 13C NMR spectra at 100 MHz
(JEOL EX-400) or 125 MHz (JEOL a-500). Spectra are
reported in parts per million from internal tetramethylsilane or
the residual proton of the deuterated solvent. Mass spectra
were measured with Shimadzu LCMS-QP8000 (APCI mode)
equipped with Buckyprep column or JEOL JMS SX102 (FAB
mode). Preparative HPLC was performed on a Buckyprep
column (20 6 250 mm) using toluene/2-propanol ~ 7 : 3 as
eluent (flow rate 12 to 20 ml min21, detected at 350 nm with
a UV spectrophotometric detector, Shimadzu SPD-6A).
Solvents
Anhydrous tetrahydrofuran (THF) was purchased from Kanto
Chemical Co., Inc. (free from stabilizer). Toluene and 1,2-
Cl2C6H4 were distilled under reduced pressure from CaH2 and
dried over molecular sieves 4A. The water content of the
solvent was determined with a Karl-Fisher Moisture Titrator
(MK-210, Kyoto Electronics Company) to be less than 20 ppm.
Materials
All commercially available reagents unless noted below were
distilled or recrystallized before use. KOC(CH3)3 in THF and
TlOC2H5 were purchased from Aldrich Inc. and used as
received. [70]Fullerene (98% purity) was received from Hoechst
Japan and used as received. CuBr?S(CH3)2 was freshly
prepared from CuBr (washed with methanol prior to use)
and (CH3)2S and reprecipitated twice from (CH3)2S and
pentane.14
Scheme 2
derivative) generates a transient Cu(II) species, which decom-
poses to give 9 and Cu(0). Experimentally, the formation of
metallic copper (sometimes as copper mirror) was observed.
The addition of the third R anion generates a stable
cyclopentadienide 10, which gives 1 upon protonation. The
cyclopentadienide anion in 10 is located at the flat, strain-
free belt region of [70]fullerene, and hence is stable so that
further decomposition does not take place. In the case of
[60]fullerene, the equivalent cyclopentadienide copper complex
is still strained and undergoesfurther reactions. Though this
organometallic mechanism appears to be reasonable, involve-
ment of an organic radical species cannot be discounted at this
time.
In summary, we have discovered that various Grignard
reagents undergo completely regioselective tri-addition to
[70]fullerene in the presence of CuBr?S(CH3)2 under mild
reaction conditions. The experimental procedure is simple and
requires only the skill of undergraduate organic experiments.
Most importantly, the product yields are excellent and often
quantitative, which is unique among the methods known for
covalent functionalization of carbon clusters. Together with
the results of penta-addition to [60]fullerene, the present results
suggest that the multi-addition of organocopper reagents will
be applicable to higher fullerenes and possibly also to carbon
nanotubes. The structural and theoretical data indicate that the
cyclopentadienyl anion of the [70]fullerene tri-adduct involves
68p-electrons and represents a member of new indenyl-type
ligand incorporated into a graphitic structure.
Synthesis and characterization of 2,5,10-tris(4-
trifluoromethylphenyl)-2,5,6,10-tetrahydro[70]fullerene (1a) and
2-(4-trifluoromethylphenyl)-1,2-dihydro[70]fullerene (6a)
4-CF3C6H4MgBr was freshly prepared by addition of a THF
solution (6.00 ml) of 4-CF3C6H4Br (1.13 g, 5.00 mmol) to Mg
turnings (130 mg, 5.30 mmol). A solution of 65.0 mg (77 mmol)
of C70 in 50.0 ml of toluene was cooled to 278 uC and
transferred through a cannula over 15 minutes to a magneti-
cally stirred solution of an organocopper reagent prepared
from 4-CF3C6H4MgBr (1.87 ml) and CuBr?S(CH3)2 (274 mg,
1.33 mmol) at 278 uC. The resulting dark green suspension was
gradually warmed to 220 uC over 3 hours. The mixture was
then allowed to warm to 25 uC. While stirring was continued at
this temperature, the reaction was monitored by HPLC. After
stirring for 3 hour at this temperature, the color of the solid
became dark brown, and the color of the supernatantbecame
dark brown-red. The reaction completed in 1 day and was
quenched with saturated NH4Cl solution. The mixture was
separated into two phases, and the aqueous phase was extracted
with toluene three times. The combined organic extracts were
washed with brine, dried by passing through a pad of Na2SO4,
and evaporated. The residual powder was washed with hexane
four times to obtain the dark-brown solid. Further purification
was achieved by HPLC to give C70(4-CF3C6H4)3H (1a, 40.2 mg,
31.5 mmol, 43%) and C70(4-CF3C6H4)H (6a, 33.6 mg, 34.1 mmol,
47%). 1a: 1H NMR (400 MHz, CDCl3/CS2) d 7.92 (d, J ~
6.50 Hz, 2H), 7.84 (d, J ~ 6.50 Hz, 2H), 7.71(d, J ~ 6.50 Hz,
2H), 7.66 (d, J ~ 6.50 Hz, 2H), 7.58–7.52 (m, 4H), 4.40 (s, 1H);
13C NMR (100 MHz, CDCl3/CS2) d 159.26, 155.03, 154.07,
Experimental
General
All reactions were performed in an oven-dried reaction vessel
under argon or nitrogen using standard Schlenk technique or in
a glove box. The reactions were monitored with HPLC
2112
J. Mater. Chem., 2002, 12, 2109–2115

View Article Online
153.45, 152.19, 152.10, 151.94, 150.26, 149.91, 149.66, 149.16,
148.98, 148.77, 148.59, 148.36, 148.29, 148.25, 147.96, 147.88,
147.86, 147.81, 147.73, 147.54, 147.27, 147.04, 147.02, 146.69,
146.66, 146.49, 146.20, 145.90, 145.75, 145.44, 145.38, 144.98,
144.94, 144.91, 144.89, 144.73, 144.71, 144.53, 144.12, 143.74,
143.40, 143.24, 143.06, 142.57, 142.43, 142.35, 141.94, 140.43,
140.07, 139.44, 137.19, 133.05, 132.85, 131.86, 131.83, 131.66,
131.53, 130.61, 127.58, 127.31, 127.22, 126.86, 126.20, 126.08–
126.00 (m), 125.84 (q, JC–F ~ 3.70 Hz), 59.54, 55.87, 55.72,
55.17 (Six quartetsignals corresponding to CF3 and CF3–C
carbons were not detected because of their low intensities.); FAB-
MS m/z 1276 (M1). 6a: 1H NMR (400 MHz, CDCl3–CS2) d 8.23
(d, J ~ 8.19 Hz, 2H), 7.91 (d, J ~ 8.19 Hz, 2H), 4.62 (s, 1H);
13C NMR (100 MHz, CDCl3–CS2) d 156.86, 154.36, 151.67,
150.92, 150.47, 150.38, 149.76, 149.38, 149.21, 148.60, 148.53,
146.97, 146.89, 146.76, 146.65, 146.47, 146.03, 145.69, 143.40,
142.93, 142.81, 142.65, 142.44, 140.42, 140.04, 137.52, 133.63,
98% purity) of C70(4-ClC6H4)3H. 1H NMR (500 MHz, CDCl3–
CS2) d 7.73 (d, J ~ 6.50 Hz, 2H), 7.63 (d, J ~ 6.50 Hz, 2H),
7.50 (d, J ~ 6.50 Hz, 2H), 7.38 (d, J ~ 6.00 Hz, 2H), 7.27–7.21
(m, 4H), 4.34 (s, 1H); 13C NMR (125 MHz, CDCl3–CS2) d
160.22, 155.60, 154.74, 153.97, 152.73, 152.59, 152.52, 150.53,
150.22, 149.96, 149.50, 149.47, 149.34, 149.12, 148.94, 148.90,
148.63,148.57, 148.34, 148.21, 148.13, 148.11, 148.04, 147.96,
147.55, 147.36, 147.01, 146.97, 146.79, 146.53, 146.21, 146.19,
145.76, 145.73, 145.67, 145.31, 145.22, 145.12, 145.05, 144.85,
144.38, 144.27, 143.98, 143.66, 143.27, 143.01, 124.65, 124.61,
140.81, 140.05, 139.81, 140.00, 137.38, 136.92, 134.74, 134.60,
134.19, 133.40, 133.25, 132.22, 132.02, 131.88, 130.85, 130.55,
129.57, 129.53, 129.3, 128.96, 128.68, 128.25, 127.68, 127.60,
126.63, 59.60, 55.93, 55.81, 55.67; Anal. calcd for C88H13Cl3: C
89.85; H 1.11%. Found: C 89.55; H 1.02%.
2,5,10-Tris(1-naphthyl)-2,5,6,10-tetrahydro[70]fullerene (1d).
This compound was synthesized by Method A (Reaction
time: 10 uC for 16 hours, then 15 uC for 10 hours). Starting with
50 mg (60 mmol) of C70, we obtained 69 mg (95% yield, 97%
purity) of C70(1-naphthyl)3H. 1H NMR (500 MHz, CDCl3/
CS2) d 9.65 (d, J ~ 8.54 Hz, 1H), 9.54 (d, J ~ 9.77 Hz, 1H),
8.92 (d, J ~ 8.85 Hz, 1H), 8.26 (d, J ~ 7.02Hz, 1H), 8.09 (d,
J ~ 7.63 Hz, 1H), 7.89–7.45 (m, 5H), 7.40 (m, 1H), 7.36 (m,
1H), 7.30–7.23 (m, 7H), 6.96 (m, 1H), 6.86 (m, 1H) 4.90 (s,1H);
13C NMR (125 MHz, CDCl3/CS2) d 161.34, 158.16, 157.43,
155.83, 154.51, 153.22, 152.61, 150.79, 150.26, 150.20, 150.13,
149.95, 149.60, 149.26, 149.19, 149.16, 148.55, 148.39, 148.35,
148.28, 147.94, 147.86, 147.60, 147.49, 147.26, 147.18, 147.02,
146.96, 146.86, 146.71, 146.65, 146.28, 146.09, 145.90, 145.73,
145.36, 145.31, 145.17, 145.10, 144.93, 144.81, 144.52, 144.20,
143.59, 143.45, 142.93, 142.01, 141.86, 141.26, 141.09, 139.86,
139.54, 136.71, 135.40, 134.93, 134.81, 134.75, 133.63, 133.58,
132.56, 132.47, 132.19, 131.98, 131.58, 130.84, 129.83, 129.70,
129.54, 129.48, 129.42, 129.21, 128.94, 128.43, 128.30, 128.15,
127.03, 126.56, 126.36, 126.33, 126.19, 125.76, 125.73, 125.47,
125.40, 125.19, 125.00, 124.85, 124.61, 121.48, 61.24, 57.58,
57.31, 54.22; APCI-MS m/z 1221 [(M 2 H)2].
131.16, 130.94, 130.75, 130.19, 130.14, 127.37, 126.52 (q, JC–F
~
4.15 Hz), 59.40, 56.00, two quartet signals corresponding
to CF3 andCF3–C carbons were not detected because of their
low intensities.
Optimized procedures of the three-fold organocopper addition to
[70]fullerene
Method
A. 2,5,10-Tris(4-trifluoromethylphenyl)-2,5,6,10-
tetrahydro[70]fullerene (1a). 4-CF3C6H4MgBr was freshly
prepared by addition of a THF solution (8.00 ml) of
4-CF3C6H4Br (0.930 ml, 1.90 mmol) to Mg turnings (173 mg).
A solution of 100 mg (0.119 mmol) of C70 in 100 ml of 1,2-
Cl2C6H4 was cooled to 278 uC and cannulated over 15 minutes
to a magnetically stirred solution of an organocopper reagent
prepared from 4-CF3C6H4MgBr (2.68 ml) and CuBr?S(CH3)2
(392 mg, 1.90 mmol) at 278 uC. The resulting dark green
suspension was stirred at this temperature over an hour. The
mixture was then allowed to warm to 10 uC.After an hour
stirring at this temperature, the color of the solid became dark
brown, and the color of the supernatant became dark brown-
red. While stirring was continued at this temperature, progress
of the reaction was monitored by HPLC. The reaction
completed in 10 hours and was quenched with a saturated
NH4Cl solution. The mixture was separated into two phases,
and the aqueous phase was extracted with toluene three times.
The combined organic extracts were washed with brine, dried
by passing through a pad of Na2SO4, and evaporated. The
residual powder was washed with hexane four times to remove
biaryl (91%, 158 mg) and dried in vacuo to obtain 140 mg (94%
yield, 96% purity) of C70(4-CF3C6H4)3H.
Method B. 2,5,10-Triphenyl-2,5,6,10-tetrahydro[70]fuller-
a suspension of CuBr?S(CH3)2 (374 mg,
ene (1b). To
1.82 mmol, 30 equiv.) in THF (23.0 ml) was added a THF
solution of C6H5MgBr (0.98M, 1.86 ml, 1.82 mmol, 30 equiv.)
at 28 uC and stirring was continued for 20 minutes at this
temperature. To the resulting yellow suspension was added a
degassed solution of C70 (49.6 mg, 59.0 mmol) in 1,2-Cl2C6H4
(25.0 ml) and stirring was continued for 24 hours at 28 uC. The
reaction mixture was quenched with a 5% HCl aqueous
solution. The crude mixture was washed with water and brine
continuously, dried over anhydrous MgSO4, filtered, evapo-
2,5,10-Triphenyl-2,5,6,10-tetrahydro[70]fullerene (1b). This
compound was synthesized by Method A. Starting with 100 mg
(119 mmol) of C70, we obtained 127 mg (93% yield, 92% purity)
of C70(C6H5)3H. 1H NMR (400 MHz, CDCl3) d 7.82–7.78 (m,
2H), 7.74–7.70 (m, 2H), 7.61–7.57 (m, 3H), 7.39–7.23 (m, 8H),
4.43 (s, 1H); 13C NMR (100 MHz, CDCl3–CS2) d 160.70,
155.67, 155.07, 153.80, 152.62, 153.14, 152.62, 152.19, 150.16,
149.87, 149.57, 149.42, 149.39, 149.17, 149.08, 148.95, 148.82,
148.68, 148.59, 148.30, 148.21, 148.09, 147.92, 147.88, 147.77,
147.68, 147.18, 147.06, 147.00, 146.65, 146.61, 146.44, 146.23,
146.18, 145.84, 145.64,145.51, 145.32, 145.06, 145.03, 144.92,
144.88, 144.74, 144.48, 144.15, 143.97, 143.43, 142.99, 142.81,
142.21, 142.18, 140.50, 140.40, 139.54, 139.20, 138.33, 136.87,
133.08, 131.96, 131.93, 131.82, 131.72, 131.59, 131.43, 131.41,
130.65, 128.96, 128.88, 128.78, 128.16, 128.04, 127.71, 127.60,
127.47, 127.27, 127.15, 126.70, 126.42, 56.23, 56.19, 55.5;
APCI-MS m/z 1072 [(M 2 H)2].
rated to
a small volume and precipitated by adding
CH3OH.The precipitated dark brown solid was washed
thoroughly with CH3OH, ether and water, then dried under
reduced pressure to obtain C70(C6H5)3H (61.2 mg, 95% yield,
96% purity).
2,5,10-Tris(biphenyl-4-yl)-2,5,6,10-tetrahydro[70]fullerene (1e).
To a suspension of CuBr?S(CH3)2 (374 mg, 1.82 mmol, 30 equiv.)
in THF (23.0 ml) was added a THF solution of 4-(C6H5)C6H4MgBr
(1.04M, 1.72 ml, 1.82 mmol, 30 equiv.) at 24 uC and stirring was
continued for 20 min at this temperature. To the resulting yellow
suspension was added a degassed solution of C70 (50.2mg, 59.7 mmol)
in 1,2-Cl2C6H4 (25.0 ml) and this was stirred for 14 hours at 24 uC. The
reaction mixture was quenched with a 15% HCl aqueous solution. The
crude mixture was dried over anhydrous MgSO4, filtered through a
pad of SiO2, evaporated to a small volume and precipitated by adding
CH3OH.The dark brown solid was washed thoroughly with CH3OH,
ether and water, then dried under reduced pressure to obtain C70[4-
2,5,10-Tris(4-chlorophenyl)-2,5,6,10-tetrahydro[70]fullerene
(1c). This compound was synthesized by Method A. Starting
with 50 mg (60 mmol) of C70, we obtained 67 mg (94% yield,
1
(C6H5)C6H4]3H (76.7 mg, 99% yield, 98% purity). H NMR (CS2,
400 MHz) d 8.21 (d, J ~ 8.4 Hz, 2H), 8.09 (d, J ~ 8.4 Hz, 2H), 7.98
J. Mater. Chem., 2002, 12, 2109–2115
2113

View Article Online
(d, J ~ 8.0 Hz, 2H), 7.51–7.95 (m, 21H), 4.82 (s, 1H); 13C NMR
spectra could not be obtained because of low solubility in common
organic solvents; APCI-MS m/z 1299 [(M 2 H)2].
Table 2 Crystal data and data collection parameters of 3a
Complex
Formula
3a?(1,2-Cl2C6H4)2
C103H20Cl4F9Tl
1774.47
monoclinic
P21/n (No. 14)
18.419(5)
17.565(5)
20.638(5)
90
Formula weight
Crystal system
Space group
2,5,10-Trimethyl-2,5,6,10-tetrahydro[70]fullerene (1f). To
a suspension of CuBr?S(CH3)2 (245 mg, 1.19 mmol, 20
equiv.) and 1,3-dimethyl-2-imidazolidinone (128 ml, 1.19 mmol,
20 equiv.) in THF (24.0 ml) was added a THF solution of
MeMgBr (1.0 M, 1.19 ml, 1.19 mmol, 20 equiv.) at 28 uC and
stirring was continued for 15 minutes at this temperature. To
the resulting yellow suspension was added a degassed solution
of C70 (49.8 mg, 59.2 mmol) in 1,2-Cl2C6H4 (25.0 ml) and this
was stirred for 24 hours at 28 uC. The reaction mixture was
quenched with a 10% HCl aqueous solution. The crude mixture
was washed with water and brine, dried over anhydrous
MgSO4, filtered, evaporated to a small volume and precipitated
by adding CH3OH.The precipitated dark brown solid was
washed thoroughly with CH3OH, ether and hexane, then dried
under reduced pressure to obtain C70(CH3)3H (51.0 mg, 90%
yield, 91% purity). 1H NMR (400 MHz, CDCl3–CS2) d 3.90 (s,
1H), 2.52, (s, 3H), 2.27 (s, 3H), 2.14 (s, 3H); 13C NMR (CS2,
100 MHz) d 33.91, 29.12, 26.12 (Other signals are too weak to
be identified.); FAB-MS m/z 887 (M1).
˚
a, A
˚
b, A
˚
c, A
a, deg
b, deg
c, deg
109.861(5)
90
6280(3)
3
˚
V, A
Z
4
1.877
3472
293(2)
Dcalcd, g cm23
F(000)
T, K
Crystal size, mm
2hmin, 2hmax, deg.
No. of refl. measured (Total)
No. of refl. measured (I w 2.0s(I))
No. of variables
R1, wR2 (all data)
R, Rw (I w 2.0s(I))
GOF on F2
0.32 6 0.22 6 0.10
5.0, 51.8
5791
2323
785
0.223, 0.319
0.097, 0.240
1.01
23
˚
D, eA
1.32, 20.63
THF-d8) d 8.09–8.07 (m, 6H),7.59–7.53 (m, 6H); 13C NMR
spectra could not be obtained because of low solubility in
common organic solvents.
Preparation of K and Tl complexes. K[g5-C70(4-CF3C6H4)3]
(2a). To a 0.50 ml KOC(CH3)3 solution (18 mM, 9.0 mmol) in
THF-d8 was added C70(4-CF3C6H4)3H (11 mg, 9.0 mmol) at
25 uC. The resulting dark red solution was used for 1H and 13
C
Crystallographic data collections and structure determina-
tion of 3a. A crystal of 3a suitable for the X-ray diffraction
study was mounted on a MacScience DIP2030 Imaging Plate
diffractometer for data collection using MoKa (graphite
monochromated, l ~ 0.71069) radiation. Crystal data [CCDC
180836. See http://www.rsc.org/suppdata/jm/b2/b202130g/ for
crystallographic data in .cif format.] and data statistics are
summarized in Table 2. The camera distance was 120 mm. All
data was corrected for Lorentz and polarization effects. The
structure of the complex 3a was solved by the direct method
(SHELXS-97)15 and expanded using Fourier techniques
(DIRDIF-94).16The non-hydrogen atoms except for several
carbon atoms of 3a were refined anisotropically by the full-
matrix least squares method. Hydrogen atoms were placed at
1
NMR measurements. H NMR (400 MHz, THF-d8) d 8.07–
8,05 (m, 6 H), 7.57–7.54 (m, 6H); 13C NMR (100 MHz, THF-
d8) d 166.18, 160.50, 153.30, 151.99, 151.04, 150.32, 149.79,
149.77, 149.17, 148.88, 148.82, 148.48, 148.44, 148.26, 147.14,
146.90, 146.81, 146.37, 146.19,146.07, 145.72, 145.61, 145.43,
144.71, 142.53, 142.28, 137.87, 136.81, 134.54, 134.46, 133.24,
132.92, 132.60, 129.36–129.23 (m), 128.93, 125.80–125.77 (m),
125.6 (q, J ~ 271.19 Hz), 121.07, 61.06, 59.11.
K[g5-C70(C6H5)3] (2b). 1H NMR (400 MHz, THF-d8) d
7.88–7.82 (m, 5H), 7.24–7.08 (m, 10H). 13C NMR spectra could
not be obtained because of low solubility in common organic
solvents.
˚
calculated positions (C–H ~ 0.95 A) and kept fixed. Measured
non-equivalent reflections were used for the structure determi-
K[g5-C70(4-ClC6H4)3] (2c). 1H NMR (400 MHz, THF-d8)
d 7.92–7.81 (m, 6 H), 7.27–7.24 (m, 6H); 13C NMR (100 MHz,
THF-d8) d 166.82, 160.75, 153.48, 152.11, 151.19, 149.90,
149.80, 149.29, 148.90, 148.86, 148.64, 148.47, 147.32, 147.02,
146.95, 146.82, 146.77, 146.42, 146.28, 146.18, 145.79, 145.60,
145.58, 145.26, 144.65, 142.84, 142.52, 142.48, 138.07, 136.78,
134.93, 134.50, 133.35, 133.02, 132.97, 132.76, 132.65, 130.41,
130.09, 129.50, 128.98, 121.44, 60.09, 58.79.
nation. In the subsequent refinement, the function gv(Fo22
Fc ) was minimized, where |Fo| and |Fc| are the observed and
calculated structure factor amplitudes, respectively. The
2 2
agreement indices are defined as R1 ~ S(||Fo| 2 |Fc||)/S|Fo|
and wR2 ~ [Sv(Fo 2 Fc ) /S(vFo4)]1/2.The quality of the
crystals of 3a was not good presumably due to the presence of
solvent molecule.
2
2 2
K(g5-C70(CH3)3) (2f). 1H NMR (400 MHz, THF-d8) d
2.86, (s, 3H), 1.97, (s, 6H); 13C NMR (100 MHz, THF-d8) d
168.84, 162.04, 153.62, 151.90, 151.20, 149.77, 149.09, 148.75,
148.63, 148.34, 148.29, 147.46, 147.28, 146.49, 146.36, 146.26,
146.17, 145.98, 145.84, 145.70, 145.47, 145.32, 143.80, 142.36,
142.27, 138.54, 135.84, 135.02, 133.86, 133.35, 133.09, 132.32,
130.17, 124.27, 52.96, 52.67, 33.07, 31.77.
Acknowledgement
This work was supported by Grant-in-Aid for Scientific
Research (Specially Promoted Research) from the Ministry
of Education, Science, Culture and Sports. Generous allotment
of computational time from the Institute for Molecular
Science, Okazaki, Japan, and the Intelligent Modeling
Laboratory, The University of Tokyo, is gratefully acknowl-
edged. M. T. thanks JSPS for a predoctral fellowship.
Tl[g5-C70(4-CF3C6H4)3] (3a). To a THF solution (10 ml)
of C70(4-CF3C6H4)3H (46 mg, 35 mmol) was added 10 mL
solution of TlOC2H5 (25 mL, 0.35 mmol) in THF (75 mL) via
syringe at 25 uC without stirring under protection from light.
After 2 hours, the resulting dark red suspension was freed of
the solvent in vacuo at 25 uC to give the Tl complex
quantitatively. Recrystallization by solvent diffusion technique
with 1,2-Cl2C6H4–hexane under darkness gave dark red
crystals suitable for X-ray analysis. 1H NMR (400 MHz,
References
1
W. Andreoni, The Physics of Fullerene-Based and Fullerene-
Related Materials, Kluwer Academic Publishers, The Netherlands,
2000.
2
A. Hirsh, Synthesis, 1995, 895; F. Diederich and C. Thilgen,
Science, 1996, 271, 317; M. Prato, J. Mater. Chem., 1997, 7, 1097;
2114
J. Mater. Chem., 2002, 12, 2109–2115

View Article Online
C. Thilgen, A. Herrmann and F. Diederich, Angew. Chem., Int. Ed.
Engl., 1997, 36, 2269; F. Diederich and R. Kessinger, Acc. Chem.
Res., 1999, 32, 537.
M. Sawamura, H. Iikura and E. Nakamura, J. Am. Chem. Soc.,
1996, 118, 12850; H. Iikura, S. Mori, M. Sawamura and
E. Nakamura, J. Org. Chem., 1997, 62, 7912; M. Sawamura,
H. Iikura, T. Ohama, U. E. Hackler and E. Nakamura,
J. Organomet. Chem., 2000, 32, 599.
M. Sawamura, N. Nagahama, M. Toganoh and E. Nakamura,
J. Organomet. Chem., in press.
M. Sawamura, M. Toganoh, Y. Kuninobu, S. Kato and
E. Nakamura, Chem. Lett., 2000, 270.
10 M. Sawamura, M. Toganoh, K. Suzuki, A. Hirai, H. Iikura and
E. Nakamura, Org. Lett., 2000, 2, 1919.
11 The calculations were performed with a GAUSSIAN 94 program
package. M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W.
Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith,
G. A. Petersson, J. A. Montogomery, K. Raghavachari, M. A.
Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman,
J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M. Challacombe,
C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong, J. L. Andres,
E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley,
D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon,
C. Gonzalez and J. A. Pople, Gaussian 94, revision D. 3, 1995,
Gaussian. Inc., Pittsburg, PA.
12 W. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, NewYork, 1986.
13 W. E. Rhine and G. D. Stucky, J. Am. Chem. Soc., 1975, 97, 737.
14 R. J. K. Taylor, Organocopper Reagents: A Practical Approach,
Oxford University Press, Oxford, 1994, p 38.
3
4
5
6
7
M. Sawamura, Y. Kuninobu and E. Nakamura, J. Am. Chem.
Soc., 2000, 122, 12407.
M. Sawamura, N. Nagahama, M. Toganoh, U. E. Hackler,
H. Isobe, E. Nakamura, S. Zhou and B. Chu, Chem. Lett., 2000,
1098; S. Zhou, C. Burger, B. Chu, M. Sawamura, N. Nagahama,
M. Toganoh, U. E. Hackler, H. Isobe and E. Nakamura, Science,
2001, 291, 1944; T. Shimada, H. Nakatani, K. Ueno, A. Koma,
Y. Kuninobu, M. Sawamura and E. Nakamura, J. Appl. Phys.,
2001, 90, 209.
Preliminary communication; M. Sawamura, H. Iikura, A. Hirai
and E. Nakamura, J. Am. Chem. Soc., 1998, 120, 8285, . Reported
total energy of the optimized structure of Li(C70H3) was incorrect.
H. Isobe, N. Tomita and E. Nakamura, Org. Lett., 2000, 2, 3663.
15 G. M. Sheldrick, Program for the Solution of Crystal Structures,
University of Goettingen, Germany, 1997.
16 P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman, R. de
Gelder, R. Israel and J. M. M. Smits, The DIRDIF-94 program
system, Technical Report of the Crystallography Laboratory,
University of Nijmegan, The Netherlands, 1994.
8
9
J. Mater. Chem., 2002, 12, 2109–2115
2115