Switchable open-cage fullerene for water encapsulation

Article abstract of DOI:10.1002/anie.201004879

A molecular container for a single water molecule was obtained by chemical transformation of a [60]fullerene cage. A phosphate moiety acts as an effective "stopper" in the orifice (see picture). Copyright

Full text of DOI:10.1002/anie.201004879

Angewandte
Chemie
DOI: 10.1002/anie.201004879
Molecular Flasks
Switchable Open-Cage Fullerene for Water Encapsulation**
Qianyan Zhang, Tobias Pankewitz, Shuming Liu, Wim Klopper,* and Liangbing Gan*
Fullerenes have a cavity large enough to encapsulate atoms
and molecules.[1] Making a suitable hole in the fullerene cage
followed by attaching an easily removable stopper into the
orifice can produce a single-molecule-based vial. There are
many possible applications for such a vial; for example, it may
be used as a carrier for radioactive atoms or small molecules
such as tritiated water (T2O) in radiopharmaceuticals for
diagnostic or therapeutic studies. Over the past decade, much
progress has been made in the area of fullerene cage
opening.[2] A number of open-cage fullerenes have been
reported, some of which can encapsulate small organic
molecules.[3] However, there have been no successful attempts
to cover the orifice with a removable blocker. The known
open-cage fullerenes are analogous to a round-bottom flask
without a stopper. Herein, we report the preparation of the
first open-cage fullerene with a stopper that can be attached
and removed through a single-step reaction. Water was shown
to be effectively trapped by the present compound.
Scheme 1. Addition of a phosphate blocker to the orifice.
We recently reported the preparation of compound 1a
and its water-encapsulated complex H2O@1a through a
peroxide-mediated cage-opening strategy.[4] Carbonyl groups
on the rim of the orifice are quite reactive towards various
nucleophiles. Treatment of 1 with triethylphosphite resulted
in compound 2 with a phosphate group attached above the
orifice (Scheme 1). Under basic conditions, the phosphate can
be hydrolyzed to retrieve compound 1. The phosphate
formation reaction is more efficient than the hydrolysis
process, as the reaction time at room temperature is 10 min
for the forward reaction but 2 h for the backward hydrolysis
reaction.
The formation of the phosphate moiety in compound 2
was unexpected. Analogous reactions of classical organic
carbonyl compounds with triethylphosphite usually yield a-
hydroxy phosphonates (Abramov reaction)[5] with the phos-
phorus atom connected to the carbonyl carbon atom, whereas
the hydrogen atom is connected to the carbonyl oxygen atom.
The opposite is observed in the present reaction with
compound 1.
A possible mechanism is proposed in
Scheme 1. Owing to the electron-deficient nature of the
fullerene cage and the presence of carbonyl groups in
compound 1, single electron transfer is favored in the first
step instead of phosphite attack at the carbonyl carbon atom.
The so-formed radical ion pair A then forms the zwitterionic
[*] Q. Y. Zhang, S. M. Liu, Prof. Dr. L. B. Gan
Beijing National Laboratory for Molecular Sciences, CAS Key
Laboratory for Organic Solids, Institute of Chemistry, Chinese
Academy of Science, Beijing 100080 (China)
ꢀ
intermediate B through formation of an O P bond. Steric
hindrance is probably responsible for the regioselectivity.
Finally, protonation of the carbon anion and hydrolysis of the
phosphonium yields compound 2. Alternatively, intermediate
B may be formed through phosphorus attack at the carbonyl
carbon atom to form a zwitterion, followed by a Phospha–
Brook rearrangement.[6]
Dr. T. Pankewitz, Prof. Dr. W. Klopper
Abteilung fꢀr Theoretische Chemie, Institut fꢀr Physikalische
Chemie, Karlsruher Institut fꢀr Technologie, Postfach 6980, 76049
Karlsruhe (Germany)
Prof. Dr. W. Klopper
Center for Functional Nanostructures (CFN), Karlsruher Institut fꢀr
Technologie, Postfach 6980, 76049 Karlsruhe (Germany)
E-mail: klopper@kit.edu
The structure of compound 2 was determined from
spectroscopic data. The number of carbon signals in the
13C NMR spectrum confirms the Cs symmetry of the structure.
There is just one signal for the two hydroxy groups at d = 7.63
and 8.11 ppm for 2a and 2b, respectively, further supporting
the Cs symmetry. The fullerenyl proton appears as a doublet at
d = 7.36 (2a) and 7.6 ppm (2b) owing to coupling to the
phosphorus atom. In the 1H NMR spectrum of compound 2a,
there is a minor signal at d = 7.64 ppm and a minor doublet at
d = 7.37 ppm, corresponding to the OH and fullerenyl proton
groups for the water encapsulated compound H2O@2a. Thus,
Prof. Dr. L. B. Gan
Key Laboratory of Bioorganic Chemistry and Molecular Engineering
of the Ministry of Education, College of Chemistry and Molecular
Engineering, Peking University, Beijing 100871 (China)
E-mail: gan@pku.edu.cn
[**] Financial support was provided by the NNSFC, MOST, and
Deutsche Forschungsgemeinschaft (CFN subproject C3.3).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201004879.
Angew. Chem. Int. Ed. 2010, 49, 9935 –9938
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9935

Communications
just like in the case observed by Komatsu and co-workers,[7]
proton chemical shifts on the rim of the orifice can be used to
tell the difference in structure inside the cavity.
A single-crystal X-ray structure was obtained for H2O@2a
with near Cs symmetry (Figure 1). In agreement with the
relatively low-field OH proton signal mentioned above, there
Figure 2. Water encapsulation for a mixture of 1a and 2a in C6D6/H2O
monitored by 1H NMR spectroscopy.
Figure 1. Single-crystal X-ray structure of H2O@2a. Hydrogen atoms,
except those of the OH groups, are omitted for clarity. Red O, yellow
P, blue N, gray C, white H.
tems, when benchmarking it to high-level wavefunction
calculations.[8] In both equilibrium (EQ) structures H2O@1a
and H2O@2a, the binding energy of the water molecule is
ꢀ46 kJmolꢀ1. The water molecule is located off-center, closer
to the unaltered hemisphere of the cage and about 3.2 ꢀ
above the bottom pentagon in both open cages. In accordance
with the X-ray structure, we found a strong intramolecular
hydrogen bond bridging between the OH groups in the
backbone of the cages and the imino groups (H2O@2a (BP86-
is strong intramolecular hydrogen bonding between the OH
group and the adjacent imino nitrogen atom with O-H…N
geometry of 2.16 ꢀ (1.99 ꢀ) and 1348 (1418) for the two sets
of almost equivalent hydrogen bonds. The phosphate moiety
is clearly shown to be positioned above the orifice, whereas
the fullerenyl hydrogen atom points away from the orifice.
The two ethoxy groups are almost symmetrically located on
the two sides of the mirror plane. A partial space-filling model
(Figure 1 right) indicates that the orifice is significantly
reduced and that it is quite difficult for the trapped water
molecule to escape from the cage.
Experimental data support the above “stopper” effect of
the phosphate moiety. A 1:1 mixture of empty compounds 1a
and 2a was heated in C6D6 with excess water added at various
temperatures (Figure 2). The process was monitored by
1H NMR spectroscopy. The signal for the trapped water
molecule in H2O@1a was evident after 36 h at 608C.
Continued heating of the same sample at 808C for 22 h
increased the encapsulation ratio only of H2O@1a, but the
signal for H2O@2a remained absent. A weak signal appeared
for H2O@2a only after heating the sample at 858C for another
22 h. In a related experiment, H2O@1a (at about 80%
encapsulation ratio) was heated in C6D6 with excess D2O
added. There was almost no change in the water signals for
H2O@1a at 458C for 24 h. When the temperature was
increased to 608C, the signal for H2O@1a began to decrease
gradually. However, the same exchange experiment with
H2O@2a did not show any detectable signal decrease. These
results indicate that the phosphate moiety in compound 2
serves as an efficient “stopper”.
ꢀ
D/def2-TZVP): N···H 1.85 ꢀ, O H···N 1438). Apart from
optimizing the equilibrium position of the water molecule
within the cages, transition state (TS) searches for the water
molecule leaving the cages were conducted (Figure 3).
Because of the numerous degrees of freedom within the
orifices, locating the TS is very demanding (for computational
details refer to the Experimental Section). Within a small
energy range of less than 5 kJmolꢀ1, we found several
stationary points only differing by the absolute position of
the water molecule with respect to rotation around its axis.
Thus, the displayed structures are expected to represent the
TS structures rather closely.
The encapsulation of a water molecule within the com-
pounds 1a and 2a was studied by density functional theory
including an empirical correction term for dispersion (DFT-
D). In previous computational studies, this approach has
provided reliable accuracy for chemically very similar sys-
Figure 3. Adopted TS structures for the water molecule leaving the
cage of 1a and 2a, respectively. The calculated barrier heights
neglecting solvent effects are 106 kJmolꢀ1 for H2O@1a and
129 kJmolꢀ1 for H2O@2a (BP86-D/def2-TZVP//BP86-D/def2-SV(P)).
9936 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 9935 –9938

Angewandte
Chemie
From Figure 3 it can be seen that in cage 2a the maximum
energy structure (the TS) for the water molecule leaving the
cage is reached later than in 1a with respect to the distance of
the water molecule to the bottom pentagon of the cage
(ca. 7.15 ꢀ compared with ca. 6.5 ꢀ, respectively). The calcu-
lated barrier heights (gas phase) are 106 kJmolꢀ1 in H2O@1a
and 129 kJmolꢀ1 in H2O@2a (Table 1). Furthermore,
molecule are drastically reduced relative to the unblocked
open cage. Thus, the reported system represents a molecular
vial that can be opened and closed in an easy single step
reaction, suggesting possible applications as a transporter or
carrier.
Experimental Section
Table 1: Calculated energy barriers [kJmolꢀ1] for the water molecule
leaving (DEout) and entering (DEin) the cages (BP86-D/def2-TZVP//BP86-
D/def2-SV(P)). Solvent effects on the barrier heights are estimated from
the perpendicular scans with and without the COSMO model.
Compound 2a: P(OEt)3 (50 mg, 0.27 mmol) was slowly added to a
solution of compound 1a (33 mg, 0.033 mmol) in toluene (10 mL) at
208C. The progress of the reaction was monitored by TLC. The
reaction was stopped when 2a reached its maximum yield
(ca. 10 min). The solution was directly chromatographed on a silica
gel column eluting with toluene/acetic ether (100:1). The first red
band was a trace amount of unreacted starting material 1a. The
second red band was collected and evaporated to give 2a as an orange
solid (24 mg, 0.021 mmol, 64%). Characterization data: 1H NMR
([D4]-ortho-dichlorobenzene (ODCB), 600 MHz): d = 7.87 (d, 4H),
7.62 (s, 2H), 7.37 (d, 1H), 7.20 (t, 4H), 6.90 (d, 2H), 4.58–4.45 (4H),
1.33 ppm (t, 6H). 13C NMR ([D4]-ODCB, 150 MHz): All signals
represent 2C except where otherwise noted. d = 185.33, 174.09 (1C),
154.77, 149.76, 149.03, 148.89, 148.82, 148.46, 148.21 148.03, 147.79,
147.43, 147.33, 147.22, 146.62, 146.06, 145.29, 144.73, 143.69, 143.17,
142.97, 142.86, 142.77, 142.40, 141.64, 140.61, 139.58, 133.85, 132.91,
132.57, 130.67, 128.85(4C), 127.95 (4C), 123.79, 80.86 77.70, 77.31,
64.35, 16.33 ppm. FTIR (microscope): 3251, 2966, 2920, 1738, 1463,
H2O@1a
H2O@2a
DEin
60
106
92
83
129
108
122
DEout
DEin(cosmo)
DEout(cosmo)
106
H2O@2a reveals a second, much lower barrier for the water
molecule moving further out of the cage (the approximate
TS2 structure lies about 53 kJmolꢀ1 below TS1), whereas
there is only a single barrier present in H2O@1a. For
H2O@2a, only the transition state associated with the
higher of the two barriers (TS1) is shown in Figure 3 and is
listed in Table 1, where also the barrier heights for the water
molecule entering the cages from the outside are given. To
1378, 1283, 1097, 1031, 964, 769, 695 cmꢀ1
. ESI-HRMS (for
C76H24N2O9P [M+H+]: calculated 1139.1214, found 1139.1269.
Crystal data for 2a: crystal size, 0.20 ꢁ 0.20 ꢁ 0.15 mm3, triclinic,
ꢀ
space group P1, a = 12.372(3), b = 13.972(3), c = 14.496(3) ꢀ, a =
74.66(3), b = 83.10(3), g = 76.24(3)8, V= 2342.8(8) ꢀ3, Z = 2, 1calcd
=
1
interpret the dynamic H NMR measurements performed in
1.619 Mgmꢀ3; T= 173(2) K; 28960 reflections collected, 10661 inde-
pendent (Rint = 0.0360) included in the refinement; min/max trans-
mission = 0.9793 and 0.9726; refinement method full-matrix least-
squares on F2. Final R indices [I > 2s(I)] R1 = 0.0525, wR2 = 0.1392, R
indices (all data) R1 = 0.0564, wR2 = 0.1428. One of the methyl
groups is disordered over two orientations.
solution, the computed gas phase barriers were corrected for
solvent effects employing a continuum solvation model. The
corrections were estimated by comparing the potential curves
of a perpendicular trajectory for the water moving towards
the orifice in H2O@1a and H2O@2a calculated with and
without the COSMO model (for details refer to the Exper-
imental Section).
The solvent corrected energy barriers for the water
molecule entering the cages of 1a and 2a (92 kJmolꢀ1 and
108 kJmolꢀ1, respectively) differ by 16 kJmolꢀ1. Assuming
first-order kinetics with an excess of empty cages 1a and 2a in
the 1H NMR experiments, this energy difference translates to
a factor of about 230 between the two reaction rate constants
at 808C. Thus, cage 1a incorporates water 230 times faster
than cage 2a, whose orifice is blocked by the stopper group.
This result concurs with the observations in the 1H NMR
experiments. From the encapsulation ratio of approximately
100:1 found in the experiments for cage 1a and 2a, a
difference in barrier heights of about 13 kJmolꢀ1 can be
estimated that compares well to the computed value of
16 kJmolꢀ1.
In summary, by chemical transformation of an intact
[60]fullerene cage, an open-cage fullerene was synthesized
that works perfectly as a lockable molecular container for a
single water molecule. A reversibly bound phosphate moiety
located above the orifice acts as an effective stopper to block
the opening of the cage and thus trap the encapsulated water
molecule. 1H NMR experiments and theoretical investiga-
tions show that with the stopper attached the reaction rates
for the encapsulation and release process of the water
CCDC 783509 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
Hydrolysis of 2a: KOH (5 mL, 0.1m) was added to a solution of
2a (12 mg, 0.011 mmol) in CH2Cl2 (15 mL) at room temperature.
After stirring for several minutes, [18]crown-6 (6 mg) was added to
the solution. Progress of the reaction was monitored by TLC. The
reaction was stopped when 1a reached its maximum yield (ca. 2 h).
The solution was washed with water (20 mL ꢁ 3). The organic layer
was separated and directly chromatographed on silica gel eluting with
toluene/acetic ether (100:1). The first red band was collected and
evaporated to give 1a as an orange solid (3 mg, 0.003 mmol, 27%).
Density functional theory: DFT calculations including an empiri-
cal correction term for dispersion (DFT-D) were conducted for the
empty cages 1a, 2a, and their water-encapsulation complexes
H2O@1a and H2O@2a with the program package turbomole using
the resolution of the identity (RI) approximation.[9] Based on our
previous successful benchmark and computational studies for carbon
nanotubes and fullerenone cages, the DFT-D parameterisation of
Grimme from 2004 along with the BP86 functional was employed.[8]
The equilibrium structures were optimized in the def2-TZVP basis
set, but for the determination of the reaction barriers, both the
equilibrium and the transition state structures were optimized using
the smaller def2-SV(P) basis set. The final binding energies and
barriers were subsequently calculated as single points with the large
def2-TZVP basis set.[10] All energies are corrected for the basis set
superposition error (BSSE).[11] To create best-guess starting structures
for the subsequent transition state search calculations, perpendicular
Angew. Chem. Int. Ed. 2010, 49, 9935 –9938
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9937

Communications
scans for water leaving the cage were performed for H2O@1a and
J. Y. Yao, D. Z. Yang, F. D. Wang, S. H. Huang, L. B. Gan, Z. S.
Jia, Z. P. Jiang, X. B. Yang, B. Zheng, G. Yuan, S. W. Zhang,
Z. M. Wang, J. Am. Chem. Soc. 2007, 129, 16149; g) H. Hachiya,
Y. Kabe, Chem. Lett. 2009, 38, 372; h) Q. Y. Zhang, Z. S. Jia,
S. M. Liu, G. Zhang, Z. Xiao, D. Z. Yang, L. B. Gan, Z. M. Wang,
Y. L. Li, Org. Lett. 2009, 11, 2772.
H2O@2a. Starting from the equilibrium position within the cage, the
water molecule was moved towards the orifice along the OX direction
in steps of 0.5 ꢀ (X is the center of the bottom pentagon in the cage,
opposite to the opening). For each structure along these pathways, the
perpendicular position of the water molecule was fixed by freezing
the distance, angle, and dihedral angle of the oxygen atom with
respect to the bottom pentagon while optimizing all other degrees of
freedom (BP86-D/def2-TZVP). The scans were refined by reducing
the step size to 0.1–0.2 ꢀ within the maximum energy region. The
maximum energy structures within these scans were then used for the
final saddle point searches employing the surface walking algorithm
of the module statpt. Several restarts with a recomputed exact
hessian were needed to finally locate the saddle points.[12] A true TS
(one imaginary frequency) was found for H2O@2a, whereas the
adopted TS structure for H2O@1a revealed two imaginary frequen-
cies with respect to not only the reaction coordinate but also an
energetically almost irrelevant rotation of the water molecule. To
estimate the influence of the solvent on the binding energies and
barrier heights, the perpendicular scans for H2O@1a and H2O@2a
were repeated with a continuum solvation model, the conductor-like
screening model (COSMO, e = 1).
[3] a) Y. Rubin, T. Jarrosson, G. W. Wang, M. D. Bartberger, K. N.
Houk, G. Schick, M. Saunders, R. J. Cross, Angew. Chem. 2001,
113, 1591; Angew. Chem. Int. Ed. 2001, 40, 1543; b) C. M.
Stanisky, R. J. Cross, M. Saunders, M. Murata, Y. Murata, K.
Komatsu, J. Am. Chem. Soc. 2005, 127, 299; c) Ref. [2e]; d) K. E.
Whitener, Jr., R. J. Cross, M. Saunders, S.-i. Iwamatsu, S.
Murata, N. Mizorogi, S. Nagase, J. Am. Chem. Soc. 2009, 131,
6338; e) Y. Morinaka, F. Tanabe, M. Murata, Y. Murata, K.
Komatsu, Chem. Commun. 2010, 46, 4532.
[4] Ref. [2h].
[5] R. Engel, Org. React. 1988, 36, 175; for a recent example; see:
P. G. Mandhane, R. S. Joshi, D. R. Nagargoje, C. H. Gill,
Tetrahedron Lett. 2010, 51, 1490.
[6] L. El Kaꢂm, L. Gaultier, L. Grimaud, A. D. Santos, Synlett 2005,
2335.
[7] S.-C. Chuang, Y. Murata, M. Murata, K. Komatsu, Chem.
Commun. 2007, 1751.
[8] a) T. Pankewitz, W. Klopper, Chem. Phys. Lett. 2008, 465, 48;
b) T. Pankewitz, W. Klopper, J. Phys. Chem. C 2007, 111, 18917;
c) S. Grimme, J. Comput. Chem. 2004, 25, 1463.
Received: August 5, 2010
Published online: November 16, 2010
[9] a) turbomole V6.1 V6.2 2009–2010, a development of Univer-
sity of Karlsruhe (TH) and Forschungszentrum Karlsruhe
GmbH, 1989 – 2007, TURBOMOLE GmbH since 2007; http://
www.turbomole.com; b) R. Ahlrichs, M. Bꢃr, M. Hꢃser, H.
Horn, C. Kꢄlmel, Chem. Phys. Lett. 1989, 162, 165; c) O.
Treutler, R. Ahlrichs, J. Chem. Phys. 1995, 102, 346; d) R.
Ahlrichs, Phys. Chem. Chem. Phys. 2004, 6, 5119.
[10] a) A. Schꢃfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97,
2571; b) F. Weigend, M. Hꢃser, H. Patzelt, R. Ahlrichs, Chem.
Phys. Lett. 1998, 294, 143; c) F. Weigend, R. Ahlrichs, Phys.
Chem. Chem. Phys. 2005, 7, 3297; d) F. Weigend, Phys. Chem.
Chem. Phys. 2006, 8, 1057.
Keywords: density functional calculations · fullerenes ·
molecular devices · supramolecular chemistry
.
[1] a) Y. Rubin, Top. Curr. Chem. 1999, 199, 67; b) J.-F. Nieren-
garten, Angew. Chem. 2001, 113, 3061; Angew. Chem. Int. Ed.
2001, 40, 2973; c) M. Murata, Y. Murata, K. Komatsu, Chem.
Commun. 2008, 6083; d) G. C. Vougioukalakis, M. M. Roubela-
kis, M. Orfanopoulos, Chem. Soc. Rev. 2010, 39, 817; e) L. B.
Gan, D. Z. Yang, Q. Y. Zhang, H. Huang, Adv. Mater. 2010, 22,
1498.
[2] a) J. C. Hummelen, M. Prato, F. Wudl, J. Am. Chem. Soc. 1995,
117, 7003; b) G. Schick, T. Jarrosson, Y. Rubin, Angew. Chem.
1999, 111, 2508; Angew. Chem. Int. Ed. 1999, 38, 2360; c) K.
Komatsu, M. Murata, Y. Murata, Science 2005, 307, 238; d) M.
Murata, Y. Murata, K. Komatsu, J. Am. Chem. Soc. 2006, 128,
8024; e) S.-i. Iwamatsu, T. Uozaki, K. Kobayashi, S. Y. Re, S.
Nagase, S. Murata, J. Am. Chem. Soc. 2004, 126, 2668; f) Z. Xiao,
[11] S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553.
[12] a) P. Deglmann, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2002,
362, 511; b) P. Deglmann, F. Furche, J. Chem. Phys. 2002, 117,
9535; c) P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem.
Phys. Lett. 2004, 384, 103.
9938 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 9935 –9938