Communications
DOI: 10.1002/anie.200704005
Anion–p Interactions
Fluoride-Selective Host Based on Anion–p Interactions, Ion Pairing,
and Hydrogen Bonding: Synthesis and Fluoride-Ion Sandwich
Complex**
Mark Mascal,* Ilya Yakovlev, Edward B. Nikitin, and James C. Fettinger
The theoretical description of the anion–p interaction by
Frontera, Deyà, and co-workers;[1] Alkorta and co-workers;[2]
and our group[3] has, in the five years since the publication of
these studies, given rise to a large body of work describing
physical observations of close anion–p contacts as well as
further theoretical investigations.[4] Such reports are often
accompanied by the suggestion that this phenomenon could
form the basis for anion recognition and binding, and indeed,
this is turning out to be the case.[5] For example, a synthetic
halide-ion channel composed of electron-deficient aromatic
rings has recently been reported,[6] as well as a chloride–p
complex formed in a calix-like cavity within a dendrimer.[7]
We ourselves described a hypothetical series of purpose-
designed fluoride receptors that took advantage of the
cylindrophane effect, that is, the high conformational stability
and limited vertical flexibility of [1,3,5]cyclophanes bridged
by chains with odd numbers of atoms, such as 1 (Scheme 1).
addition to acute steric issues associated with accommodating
the larger chloride anion in the cavity, led to a calculated
23.5 kcalmolÀ1 difference in binding energy between the two
halides in aqueous solution.[9] Herein, we present the proof of
principle behind the design of 2.
The synthesis of a cyanuric acid based cylindrophane is
shown in Scheme 2. Although 2 itself was useful for modeling
purposes, it was anticipated that substitution at the chain
nitrogen atoms would be required to eventually be able to
work with the tricationic receptor in common organic
solvents. Thus, commercially available diethanolamine (3)
was alkylated with n-hexyl bromide to give 4. Statistical
monoprotection of 4 as the triisopropylsilyl (TIPS) ether and
subsequent conversion of the remaining hydroxy group to
chloride gave 6. Installation of the first cyanuric acid ring
could be accomplished in good yield by simply heating the
components together in DMF in the presence of the base 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU). Desilylation of 7 and
subsequent chlorination gave the cage precursor 9. Heating
trichloride 9 with cyanuric acid in the presence of DBU at
high dilution gave the macrocyclic product 10 in good yield
for this type of reaction.
Cylindrophane 10 is freely soluble in dichloromethane but
only poorly so in methanol and crystallizes from a mixture of
these solvents. An X-ray crystal structure of the empty cage is
shown in Figure 1. Unexpectedly, the conformation of one of
the three bridging chains is gauche-anti-gauche-gauche, devi-
ating from the gauche-anti-anti-gauche conformation of the
other chains. There are two molecules in the asymmetric unit
of the crystal (Z’ = 2), and this distortion appears in both of
them, causing the cyanuric acid rings to tilt relative to each
other such that the height of the cylinder varies from 4.30 to
4.91 .[10] In one of the structures (not shown), one of the
three hexyl sidechains also has a gauche kink.
The cage was armed for the binding of fluoride by
protonation. Thus, 10 was suspended in methanol, and an
excess of concentrated aqueous hydrochloric acid was added
to give a clear solution, the evaporation of which gave
[H310]Cl3 as a white solid. The corresponding tetrafluorobo-
rate salt could similarly be prepared by the use of aqueous
HBF4. Redissolution of either [H310]3+ salt in methanol and
treatment with an excess of aqueous HF gave a white solid
after evaporation of the solvent. The first evidence of the
presence of a fluoride complex in this material came from the
electrospray mass spectrum, which showed a base peak at
m/z 370, corresponding to the [H310F]2+ ion.[12] Also present
in the spectrum were signals corresponding to [H210]2+
(m/z 360), [H10]+ (m/z 718), and [H210F]+ (m/z 738). No
Scheme 1. Structures of cation host 1 and the proposed anion host 2.
While 1 has been shown to be a selective silver(I) and
copper(I) host,[8] it was demonstrated that an inversion of its
electronic character from electron-rich to electron-poor, such
as in 2, would produce an anion host.[9] In the theoretical
treatment of this receptor, the complexation of fluoride ion
was shown to benefit from intrinsically stronger p···XÀ···p and
+NH···XÀ interactions than for chloride. This difference, in
[*] Prof. M. Mascal, I. Yakovlev, Dr. E. B. Nikitin, Dr. J. C. Fettinger
Department of Chemistry
University of California, Davis
Davis, CA 95616 (USA)
Fax: (+1)530-752-8995
E-mail: mascal@chem.ucdavis.edu
[**] The authors gratefully acknowledge Prof. Marilyn Olmstead and Dr.
Martyn Jevric for assistance with X-ray crystallography and mass
spectrometry, respectively.
Supporting information for this article is available on the WWW
8782
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8782 –8784
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Take advantage of blue reference links
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&