Release of the Water Molecule Encapsulated Inside an Open-Cage Fullerene through Hydrogen Bonding Mediated by Hydrogen Fluoride

Article abstract of DOI:10.1002/chem.201502306

A reversible wetting/dewetting procedure is reported for an open-cage fullerene with an 18-membered orifice. In a homogeneous mixture of H₂O/EtOH/CHCl₃, water was encapsulated into the cavity of the open-cage compound quantitatively at 80 °C. Addition of aqueous hydrogen fluoride into the water-encapsulated complex removed the encapsulated water completely at room temperature. H-bonding between the trapped water and fluoride is shown to play a key role for the water release process. Please release me: Quantitative encapsulation and removal of a single water molecule in an open-cage fullerene has been observed under mild conditions through processes assisted by ethanol and hydrogen fluoride, respectively.

Full text of DOI:10.1002/chem.201502306

DOI: 10.1002/chem.201502306
Communication
&
Small-Molecule Encapsulation
Release of the Water Molecule Encapsulated Inside an Open-Cage
Fullerene through Hydrogen Bonding Mediated by Hydrogen
Fluoride
Liang Xu,[a] Hongjiang Ren,[b] Sisi Liang,[a] Jiahao Sun,[a] Yajun Liu,*[b] and Liangbing Gan*[a, c]
erties of open-cage fullerenes.[16] Herein we report the quanti-
Abstract: A reversible wetting/dewetting procedure is re-
ported for an open-cage fullerene with an 18-membered
orifice. In a homogeneous mixture of H2O/EtOH/CHCl3,
water was encapsulated into the cavity of the open-cage
compound quantitatively at 808C. Addition of aqueous
hydrogen fluoride into the water-encapsulated complex
removed the encapsulated water completely at room tem-
perature. H-bonding between the trapped water and
fluoride is shown to play a key role for the water release
process.
tative encapsulation and removal of a single water molecule in
an open-cage fullerene with ethanol and hydrogen fluoride, re-
spectively.
The cavity volume of open-cage fullerene is suited to fill just
one water molecule as shown by theoretical calculation[17] and
experimental results.[18] The size and shape of the opening and
functional group(s) on the rim of the opening are the key de-
termining factors for the water encapsulation process.[15] Water
encapsulation easily took place at room temperature with
a trace amount of water in the solvent during purification if
the opening is large enough and with suitable hydrophilic
groups such as carbonyl groups on the rim of the orifice. Heat-
ing the open-cage fullerene in a mixture of aromatic solvent
and water is usually required if the opening is relatively small.
In most cases the encapsulation ratio can only reach around
80%.[19–21] Murata et al. have achieved quantitative water en-
capsulation for an open-cage [60]fullerene by heating the com-
pound in toluene/water at 1208C for 36 h under high pressure
(9000 atm).[19]
Water in nonpolar cavities has been extensively studied to
understand the hydrophobic effect and to explore possible ap-
plications.[1–3] Water in the nonpolar interior channel of pro-
teins has been suggested as the mediator for proton transfer.[4]
Aquaporins transport water across biological membranes se-
lectively through hydrogen-bond-guided water movements.[5,6]
Protons have been shown to move with high mobility along
the linear water hydrogen bonds inside a carbon nanotube.[7]
Carbon nanotubes can also act as low-friction conduits for the
flow of water.[8–10] Hydrogen bonding between the trapped
water molecules with the interior was observed to stabilize the
carbon anion on the cage skeleton.[11] Open-cage [60]ful-
lerenes[12–14] have been shown to act as molecular containers
for a single water molecule.[15] The molecular nature of open-
cage fullerenes allows their study in detail with conventional
methods such as NMR, MS, and single crystal X-ray diffraction
analysis. We have been interested in the preparation and prop-
Compound 1 was previously characterized by spectroscopic
data.[22] Its single-crystal X-ray diffraction structure (Figure 1)
has now been obtained, and it confirmed the previous struc-
ture assignment.[23] The water encapsulation ratio was only
around 30% after routine purification procedure owing to its
relatively small orifice size. To increase the water content, we
[a] L. Xu, S. Liang, J. Sun, Prof. Dr. L. Gan
Beijing National Laboratory for Molecular Sciences, Key Laboratory of
Bioorganic Chemistry and Molecular Engineering of the Ministry of
Education, College of Chemistry and Molecular Engineering
Peking University, Beijing 100871 (P. R. China)
E-mail: gan@pku.edu.cn
[b] H. Ren, Prof. Dr. Y. Liu
Key Laboratory of Theoretical and Computational Photochemistry
Ministry of Education, College of Chemistry
Beijing Normal University, Beijing 100875 (P. R. China)
E-mail: yajun.liu@bnu.edu.cn
[c] Prof. Dr. L. Gan
State Key Laboratory of Organometallic Chemistry
Shanghai Institute of Organic Chemistry
Shanghai 200032 (P. R. China)
Figure 1. Single-crystal X-ray structure of compound H2O@1. The water en-
capsulation ratio is 31% as determined by H NMR spectroscopy integration.
Ellipsoids are set at 50% probability; hydrogen atoms are omitted for clarity.
13539 ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201502306.
Chem. Eur. J. 2015, 21, 13539 – 13543

Communication
heated its CDCl3 solution with water at 808C. The water encap-
sulation ratio increased to 80% after 18 h. Heating 1 in a mix-
ture of toluene and water gave similar encapsulation ratio. To
further increase the water content, we added some ethanol to
the CDCl3/H2O mixture to make a homogeneous solution in-
stead of separated CDCl3 and water phases. The heating pro-
1
cess was followed by H NMR, which indicated a quantitative
encapsulation after 18 h at 808C. The relatively small opening
also allowed the detection of the molecular ion signal of
H2O@1. Unlike the presence of both empty and water filled
molecular ion signals in partially filled samples, the mass spec-
trum of the ethanol assisted sample showed only the molecu-
lar ion signal of H2O@1 (Figure 2), in good agreement with the
1H NMR-derived quantitative water encapsulation.
Scheme 1. Reversible water encapsulation and removal processes.
The mass spectrum of 2 (bottom spectrum in Figure 2)
showed the monohydrofluoride adduct ion as a weak signal
together with empty 1 as the major signals, indicating that the
H-bond bound hydrogen fluorides can be eliminated under
1
the ESI-mass spectrum conditions. The H-decoupled 19F fluor-
ine NMR spectrum (top spectrum in Figure 3) of 2 showed two
Figure 2. ESI-HRMS spectra of 1, H2O@1, a mixture of 1+H2O@1 and 2.
A partial removal of the encapsulated water in H2O@1 can
be readily achieved by heating it in a dry solvent or evapora-
tion of its solution under vacuum. Complete removal of the
encapsulated water was achieved through an unexpected
hydrogen fluoride mediated procedure. In an effort to encap-
sulate hydrogen fluoride in compound 1,[24] we treated a mix-
ture of
1 and H2O@1 with aqueous hydrogen fluoride
(Scheme 1). The addition product 2 was produced slowly in-
stead of the expected hydrogen fluoride encapsulation prod-
1
Figure 3. 19F NMR spectra of 2 in CDCl3 (470 MHz). Above: H decoupled;
below: H coupled.
1
uct HF@1. H NMR spectra of 2 showed clearly that there is
1
hardly any detectable signal for trapped water, which should
appear in the high-field range above 0 ppm owing to the
shielding effect of the fullerene cage. The conversion of 1 into
2 was slow and the reaction usually stopped after around 50%
conversion. To make sure that both the empty 1 and water
filled H2O@1 can form compound 2, we used three different
samples of 1 with 25%, 50%, and 80% water encapsulation
ratio, respectively, all of which gave the same yield of 2 with-
out water trapped in the cavity. Compound 2 is not stable and
decomposes to empty compound 1 in a few days. Treating 2
with B(C6F5)3 eliminates hydrogen fluoride to give empty 1 in
10 min. Trifluoroacetic acid could also convert 2 into 1.
doublets. The high-field signal at À205.10 ppm is most likely
due to the fluorine atom that is well-shielded by the fullerene
1
cage. The H-coupled 19F fluorine NMR spectrum (bottom spec-
trum in Figure 3) showed a quartet at À119.25 ppm and an ap-
parent triplet at À205.10 ppm, which are due to coupling to
the bridging proton (coupling constants 75 and 157 Hz, re-
spectively) and coupling between the two different fluorine
atoms (coupling constant is 157 Hz). The substantial difference
1
of coupling constants between H and the two fluorine atoms
is in agreement with the structure assignment for 2, in which
Chem. Eur. J. 2015, 21, 13539 – 13543
www.chemeurj.org
13540
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
the hydrogen is closer to the terminal fluorine atom in the
F···HÀF moiety.
indicates that the protonation process does not result in re-
lease of the trapped water molecule. Comparison of UV/Vis
spectra of 1+H2O@1 and 3+H2O@3 (protonated by HCl gas)
showed no significant change except slight increase or de-
crease in the absorption coefficients.
In the reaction of 1 and H2O@1 with aqueous HF to form 2,
two unstable species 3 and H2O@3 could be observed by car-
rying out the reaction in CDCl3 and monitoring the progress
by 1H NMR spectrum. As in the 1H NMR spectrum of 1 and
H2O@1, there is no detectable difference in the range from 0
to 10 ppm between the empty 3 and water-filled H2O@3
except the trapped water signal. Attempted isolation resulted
in its decomposition back to compound 1 and H2O@1. To con-
firm the speculation that 3 and H2O@3 are due to protonation
of 1 and H2O@1, we treated CDCl3 solutions of 1 and H2O@1
with various acids, including CF3COOH, HCl (aqueous), and HCl
(gas) (Figure 4). All the acids resulted in a pair of tert-butyl sig-
Whitby et al. have recently successfully inserted hydrogen
fluoride into the cavity of an open-cage [60]fullerene by treat-
ing the open-cage compound with hydrogen fluoride in pyri-
dine (HF·Py, 70% w/w).[24] The hydrogen fluoride encapsulation
1
ratio can reach 50%. The H chemical shift of a HF encapsulat-
ed open-cage [60]fullerene appears at À6.55 ppm, with a 19F
coupling constant of 508 Hz. Inspired by this result, we treated
1 (30% water encapsulation ratio) in CDCl3 with HF·Py and
monitored the reaction by 1H NMR spectroscopy. As in the
above reaction with aqueous hydrogen fluoride, protonation
product 3 and H2O@3 were observed. However, there was no
detectable formation of 2 or any other HF addition product,
and the water encapsulation ratio of 1 remained unchanged.
Therefore, the presence of the basic pyridine inhibits the add-
ition of HF to compound 1. However, when we treated com-
pletely empty 1 with HF·Py (70% w/w) at room temperature,
we observed four peaks at À7.50, À7.53, À8.52, and
1
À8.74 ppm in the H NMR spectrum. The relative intensities of
these four peaks indicate that they are due to two doublets
centred at À8.01 ppm (À7.50/À8.52 ppm) and À8.14 ppm
(À7.53/À8.74 ppm) with a coupling constant of 506 Hz and
505 Hz, respectively. In light of Whitby’s work, we assign these
two doublets to the cation HF@3 and HF@1 (Scheme 2), re-
Scheme 2. Encapsulation of hydrogen fluoride in 1.
spectively. The HF encapsulation ratio is around 5% as deter-
mined from the integrals. Attempts failed to increase the en-
capsulation ratio by changing the solvent and the tempera-
ture. Density functional theory (DFT) calculations at the M06-
2X/6-31G** level of theory indicate that the binding energy of
H2O and HF is À42.39 and À21.31 kJmolÀ1 respectively in the
equilibrium structures of H2O@1 and HF@1. The much smaller
binding energy of HF is probably responsible for the difficulty
in its encapsulation.
Figure 4. 1H NMR spectra of reactions between 1+H2O@1 and different
acids in CDCl3 (for full spectra, see the Supporting Information).
nals at 1.35 and 1.25 ppm, which are the same as that of 3 and
H2O@3. The extent of protonation is dependent on the acidity
of the acid. Weak acids can only result in partial protonation.
Quantitative protonation was observed by bubbling HCl gas
into the solution. The water encapsulation ratio of 3 remains
the same as that of the starting compound 1 in both cases
when we treated two samples of 1 with 54% and 13% water
encapsulation ratios with different acids separately. This result
Hydrogen bonds with a fluorine atom in organic molecules
as the acceptor have been well studied, and controversial evi-
dence has been reported concerning their presence.[25,26]
Recent results strongly support that such organic fluorine
Chem. Eur. J. 2015, 21, 13539 – 13543
www.chemeurj.org
13541
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
hydrogen bonds CÀF···HX are possible and play an important
role in the functional properties of fluoro organic compounds.
Paquin et al. have reported that water accelerates alkyl fluoride
substitution through hydrogen bonding.[27] Hydrogen bond
like weak interactions were observed between the trapped
water molecules and fluorine atoms bound on the inner wall
of single wall carbon nanotubes through ab initio molecular
dynamics simulations.[28]
concerning this possibility, we heated 1 in toluene and H2O18
to obtain a mixture of 1 and H218O@1 in about 20:80 ratio, as
shown by an ESI-MS spectrum. Treatment of this mixture with
HF and subsequent decomposition with B(C6F5)3 gave com-
pound 1 without any 18O remaining in the molecule. This
result indicates that the carbonyl groups in 1 are not involved
in the escape of the trapped water molecule. The imino group
plays the key role in the hydrogen fluoride assisted removal of
trapped water.
The present hydrogen fluoride assisted removal of trapped
water from H2O@1 probably involves fluorine hydrogen bond-
ing as the key step as shown in Scheme 3. Addition of hydro-
In summary, a reversible wetting/dewetting procedure is re-
ported for an open-cage fullerene with an 18-membered ori-
fice. Both the encapsulation and removal processes are quanti-
tative (as determined by NMR spectroscopy) under relatively
mild conditions. Addition of ethanol to the organic solvent
containing the open-cage compound is shown to facilitate the
encapsulation process. The imino group on the rim of the ori-
fice is crucial for the fluorine hydrogen bond based removal of
the trapped water molecule. Hydrogen fluoride adds to the
imino group, and subsequent fluorine hydrogen bonding with
the trapped water activates the water removal process.
Experimental Section
Preparation of H2O@1: Three drops of water was added to a solu-
tion of compound 1 (8.2 mg, 0.0077 mmol) in 5 mL CDCl3. EtOH
was added carefully drop wise into the solution to make a homoge-
neous solution without the precipitation of compound 1. The re-
sulting solution was stirred at 808C in a sealed tube for 14 h. Water
(5 mL) was added to the solution and the organic phase was re-
moved under reduced pressure at 308C. CDCl3 (3 mL) was added
to the residue, about 1 mL of which was used for 1H NMR measure-
ment. After the NMR sample was recovered, the solvent was
evaporated in vacuo, and the residue was washed by 20 mL
hexane 3 times to afford H2O@1 (8.0 mg, 0.074 mmol, 96%).
Scheme 3. Possible pathway for the hydrogen fluoride assisted release of
trapped water in H2O@1.
gen fluoride to H2O@1 forms the intermediate A. The trapped
water moves up from the centre of the cavity to form the
hydrogen bond with the added fluorine atom analogous to
the hydrogen bond in compound 2. Hydrogen fluoride then
replaces the “activated water” to form compound 2. The driv-
ing force for water to escape the cavity is probably more
favourable hydrogen bonding with other water and/or hydro-
gen fluoride molecules in the solvent.[29,30] In hydrogen fluoride
solution, the concentration of (HF)2 is relatively high because
2: HF (3 mL, 30% aqueous solution) was added to a solution of
compound 1 and H2O@1 (28.3 mg, 0.026 mmol) in 8 mL CDCl3. The
resulting solution was stirred at 208C for 30 h. The reaction was
monitored by both TLC and 1H NMR spectroscopy. The 1H NMR
sample was poured back into the original reaction system together
with 0.5 mL 30% HF. The reaction was stopped with excess aque-
ous NaHCO3 and washed with water (5”5 mL). The organic layer
was dried over anhydrous Na2SO4 and directly chromatographed
on silica gel eluting with toluene/ethyl acetate (20:1) to recover
the starting material (12.5 mg), then CH2Cl2/MeOH (20:1) eluted
the product as a red band which was collected and evaporated
under 308C to give compound 2 (11.4 mg, 0.034, 35%, 72% based
on recovered starting material). Yield of 2 was quantitative based
on 1H NMR monitoring.
À
the hydrogen bond in HF2 is the strongest known hydrogen
bond.[31] In the formation of compound 2, addition of (HF)2 to
H2O@1 to form intermediate B is also possible. Just like A, B
loses the trapped water molecule through the H-bonding as-
sisted process. The DFT M06-2X calculations predict that the
energy change for the reaction from H2O@1 to A and B is
33.57 and 108.05 kJmolÀ1, and À14.91 and À105.54 kJmolÀ1
,
For details of other procedures and spectroscopic data, see the
Supporting Information.
in enthalpy and Gibbs free energy, respectively. So, the path-
way through intermediate B is possible, through A is infeas-
ible.
Another possible route for the escape of trapped water from
H2O@1 involves acid-catalysed formation of a hydrated ketone
moiety between the trapped water molecule with one of the
three carbonyl groups on the rim of the orifice. Dehydration of
the hydrated ketone moiety then releases the trapped water
molecule. This pathway appears unlikely from the HCl proton-
ation experiment shown in Figure 4, in which the encapsula-
tion ratio remained unchanged. To obtain further information
Acknowledgements
This work is supported by NNSFC (Grants 21272013 and
21132007) and MOST (2011CB808401 and 2015CB856600).
Keywords: fullerenes · hydrogen bonds · hydrogen fluoride ·
open-cage fullerenes · water encapsulation
Chem. Eur. J. 2015, 21, 13539 – 13543
www.chemeurj.org
13542
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[1] F. Biedermann, W. M. Nau, H.-J. Schneider, Angew. Chem. Int. Ed. 2014,
53, 11158; Angew. Chem. 2014, 126, 11338.
[20] Q. Y. Zhang, T. Pankewitz, S. M. Liu, W. Klopper, L. B. Gan, Angew. Chem.
2010, 122, 10131; Angew. Chem. Int. Ed. 2010, 49, 9935.
[21] Y. M. Yu, L. J. Shi, D. Z. Yang, L. B. Gan, Chem. Sci. 2013, 4, 814.
[22] S. M. Liu, Q. Y. Zhang, Y. M. Yu, L. B. Gan, Org. Lett. 2012, 14, 4002.
[23] A single crystal of 1 was obtained from diffusion at room temperature
[2] B. W. Matthews, L. J. Liu, Protein Sci. 2009, 18, 1135.
[3] J. C. Rasaiah, S. Garde, G. Hummer, Annu. Rev. Phys. Chem. 2008, 59,
713.
in CH2Cl2/pentane. C79H26.75N2O4.25
10.4729(13) Š, b=33.203(3) Š, c=15.3381(19) Š, b=107.731(13)8, V=
5080.2(10) Š3, T=180.01(10), space group P21/c (no. 14), Z=4,
,
Mr =1071.52, monoclinic, a=
[4] I. Schlichting, J. Berendzen, K. Chu, A. M. Stock, S. A. Maves, D. E.
Benson, R. M. Sweet, D. Ringe, G. A. Petsko, S. G. Sligar, Science 2000,
287, 1615.
m(Mo Ka)=0.087, 17255 reflections measured, 7414 unique (Rint
=
[5] M. Borgnia, S. Nielsen, A. Engel, P. Agre, Annu. Rev. Biochem. 1999, 68,
425.
0.0735), which were used in all calculations. The final wR2 was 0.2314
(all data) and R1 was 0.0795 (>2 sigma(I)). CCDC 1063410 contains the
supplementary crystallographic data for this paper. These data are pro-
vided free of charge by The Cambridge Crystallographic Data Centre.
[24] A. Krachmalnicoff, R. Bounds, S. Mamone, M. H. Levitt, M. Carravettaa,
R. J. Whitby, Chem. Commun. 2015, 51, 4993.
[6] U. Kosinska Eriksson, G. Fischer, R. Friemann, G. Enkavi, E. Tajkhorshid, R.
Neutze, Science 2013, 340, 1346.
[7] C. Dellago, M. M. Naor, G, Hummer, Phys. Rev. Lett. 2003, 90, 105902.
[8] G. Hummer, J. C. Rasaiah, J. P. Noworyta, Nature 2001, 414, 188.
[9] M. Majumder, N. Chopra, R. Andrews, B. J. Hinds, Nature 2005, 438, 44.
[10] J. K. Holt, H. G. Park, Y. M. Wang, M. Stadermann, A. B. Artyukhin, C. P.
Grigoropoulos, A. Noy, O. Bakajin, Science 2006, 312, 1034.
[11] E. E. Maroto, J. Mateos, M. Garcia-Borràs, S. Osuna, S. Filippone, M. .
Herranz, Y. Murata, M. Solà, N. Martín, J. Am. Chem. Soc. 2015, 137, 1190.
[12] M. Murata, Y. Murata, K. Komatsu, Chem. Commun. 2008, 6083.
[13] G. C. Vougioukalakis, M. M. Roubelakis, M. Orfanopoulos, Chem. Soc. Rev.
2010, 39, 817.
[25] H.-J. Schneider, Chem. Sci. 2012, 3, 1381.
[26] P. A. Champagne, J. Desroches, J.-F. Paquin, Synthesis 2015, 47, 306.
[27] P. A. Champagne, J. Pomarole, M.-E. Therien, Y. Benhassine, S. Beaulieu,
C. Y. Legault, J.-F. Paquin, Org. Lett. 2013, 15, 2210.
[28] J. K. Clark II, S. J. Paddison, Phys. Chem. Chem. Phys. 2014, 16, 17756.
[29] For solubility of water in benzene, R. Karlsson, J. Chem. Eng. Data 1973,
18, 290.
[30] For solubility of hydrogen fluoride in benzene, J. H. Simons, J. Am.
Chem. Soc. 1931, 53, 83.
[31] J. Emsley, Chem. Soc. Rev. 1980, 9, 91.
[14] L. B. Gan, D. Z. Yang, Q. Y. Zhang, H. Huang, Adv. Mater. 2010, 22, 1498.
[15] L. J. Shi, L. B. Gan, J. Phys. Org. Chem. 2013, 26, 766.
[16] L. B. Gan, Chem. Rec. 2015, 15, 189.
[17] A. Varadwaj, P. R. Varadwaj, Chem. Eur. J. 2012, 18, 15345.
[18] K. Kurotobi, Y. Murata, Science 2011, 333, 613.
Received: June 13, 2015
[19] S.-I. Iwamatsu, T. Uozaki, K. Kobayashi, S. Y. Re, S. Nagase, S. Murata, J.
Am. Chem. Soc. 2004, 126, 2668.
Revised: July 12, 2015
Published online on August 6, 2015
Chem. Eur. J. 2015, 21, 13539 – 13543
www.chemeurj.org
13543
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim