Pulsed Gradient Spin Echo (PGSE) diffusion measurements as a tool for the elucidation of a new type of hydrogen-bonded bicapsular aggregate

Article abstract of DOI:10.1002/chem.200500021

Compounds formed by linking two tris(ureidobenzyl)amine modules with a hexamethylene tether are described. These compounds self-assemble to form bicapsular aggregates featuring two rings of six hydrogen-bonded ureas. ¹H and ¹H/¹

Full text of DOI:10.1002/chem.200500021

FULL PAPER
DOI: 10.1002/chem.200500021
Pulsed Gradient Spin Echo (PGSE) Diffusion Measurements as a Tool for
the Elucidation ofa New Type ofHydrogen-Bonded Bicapsular Aggregate
Mateo Alajarín,*[a] Aurelia Pastor,*[a] Raffll-Angel Orenes,[a]
Eloísa Martínez-Viviente,[b] and Paul S. Pregosin[b]
Abstract: Compounds formed by link-
ing two tris(ureidobenzyl)amine mod-
ules with a hexamethylene tether are
described. These compounds self-as-
semble to form bicapsular aggregates
featuring two rings of six hydrogen-
bonded ureas. 1H and 1H/1H ROESY
NMR spectroscopy, together with
pulsed gradient spin echo (PGSE)
NMR diffusion measurements, have
been used to characterize the dimers in
solution. The results have been com-
pared with energy-minimized struc-
tures. The new compounds are kineti-
cally stable on the NMR timescale, and
their thermodynamic stabilities are
comparable to other capsular aggre-
gates derived from tris(ureidobenzyl)-
amines.
Keywords: diffusion
· hydrogen
bonds · NMR spectroscopy · self-
assembly · structure elucidation
Introduction
Multiple hydrogen-bond formation has opened the door to a
useful approach for assembling individual molecular compo-
nents into functional organic nanostructures that are held
together in a reversible manner.[1] Molecules containing the
correct structural information spontaneously assemble into
helices,[2] rosettes,[3] cylinders,[4] capsules,[5] dendrimers,[6]
polymers,[7] or grids.[8] Urea functionalities have a great ten-
dency to form hydrogen bonds, not only in the solid state,
but also in solution.[9] Recently, some of us have proved that
the triureas 1, derived from the tribenzylamine skeleton, as-
semble to form dimeric aggregates (Scheme 1).[10] In these
dimers two molecules associate by forming hydrogen bonds
between the six ureas, forming a head-to-tail directional
array of 12 hydrogen bonds.[10] The result is a capsular ag-
Scheme 1. Tris(2-ureidobenzyl)amines 1 dimerize in the solid state and in
solution to form capsular aggregates.
[a] Prof. Dr. M. Alajarín, Dr. A. Pastor, R.-A. Orenes
Departamento de Química Orgµnica
Facultad de Química, Universidad de Murcia
Campus de Espinardo, Murcia-30.100 (Spain)
Fax : (+34)968-364-149
gregate with an internal cavity, that is too small for the en-
capsulation of organic molecules.[10a,c]
Unimolecular capsules could in principle be achieved
from a molecule containing two tripodal subunits of tris(ur-
eidobenzyl)amine, linked through their urea residues by
means of a flexible tether. This spacer should be long
enough to allow the folding of the molecule to form the ring
of hydrogen-bonded ureas, but short enough to minimize
the increase in entropy brought about by the freely rotating
single bonds. A hexamethylene spacer has previously been
E-mail: alajarin@um.es
aureliap@um.es
[b] Dr. E. Martínez-Viviente, Prof. Dr. P. S. Pregosin
Laboratory of Inorganic Chemistry
ETHZ, Hçnggerberg, 8093 Zürich (Schweiz)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author. It contains a
1H/1H ROESY spectrum of the triurea 1a (CDCl3), as well as energy-
minimized structures of compounds 2b·2b and 6.
Chem. Eur. J. 2006, 12, 877 – 8W86iꢀley2-0V0C6H Verlag GmbH & Co. KGaA, Weinheim
877

Results and Discussion
Synthesis: The synthetic procedure leading to the hexaureas
2a,b is shown in Scheme 2. The key intermediate 3 was pre-
pared in 93% yield by the reaction of 2-azidobenzylamine
with 2-nitrobenzyl iodide in the presence of Na2CO3. The
formation of the corresponding iminophosphorane by the
reaction of 3 with trimethylphosphine, followed by hydroly-
sis in THF/H2O, gave the amine 4 (67–90%). After reaction
with 1,6-diisocyanatohexane, the bis(tribenzylamine) deriva-
tive 5 was isolated in 69% yield. Finally, the hexaureas 2
were obtained from 5 by the following two-step sequence.
First, catalytic hydrogenation of 5 afforded the tetraamino
intermediate 6 in 91% yield. Second, this intermediate was
Figure 1. Schematic representation of the self-assembly of 2 to yield un-
imolecular, bimolecular, or oligomeric aggregates.
used to successfully link two tetraureidocalixarenes, leading
to unimolecular capsules.[11] Consequently, we decided to in-
troduce this structural fragment into our tris(ureidobenzyl)-
amine system. In principle, such a modified structure would
allow several types of assembly to occur (Figure 1), such as:
1) intramolecular self-assembly to give a unimolecular cap-
sule, 2) intermolecular assembly to give a bimolecular spe-
cies, and 3) oligomerization to higher order systems.
Whereas, examples of assembly by hydrogen bonding to
afford unimolecular[11,12] or polymeric capsules[7c–e] are well-
known in the literature, bicapsular aggregates assembled ex-
clusively by hydrogen bonding are rare. Rebek[7c,11] has re-
ported an example based on a heterotrimer. Cram,[13] Sher-
man,[14] Nolte,[15] and Liu[16] have reported several bicapsules
based on the highly restricted resorcarene skeleton and on
two metallobridged bis(b-CyDs) complexes, but these capsu-
les were formed by covalent synthesis, and not exclusively
by hydrogen-bond assembly.
The characterization of assembling systems in solution
often relies on a combination of NMR spectroscopy, mass
spectrometry, vapor pressure osmometry (VPO), and gel
permeation chromatography (GPC). However, as the mole-
cules increase in complexity, new analytical tools are re-
quired.[17] Recently, pulsed gradient spin echo (PGSE) NMR
diffusion measurements[18] have proved to be a useful tool
for probing encapsulation and studying the structure of hy-
drogen-bonded capsules in solution.[19] The diffusion coeffi-
cient (D) depends on the molecular shape and size (the
larger the molecule, the lower the D value). Thus, it should
be possible to distinguish between the unimolecular, bimo-
lecular, and oligomeric species in Figure 1 on the basis of
their diffusion coefficients. We present here our PGSE diffu-
sion measurements on the hexaureas 2a,b (Scheme 2) and
related species. These results, together with 1H NMR and
2D NOE spectroscopic data, provide conclusive evidence of
the solution structures of these systems.
Scheme 2. Reagents and reaction conditions for the synthesis of the hex-
aureas 2a,b: a) 2-nitrobenzyl iodide, Na2CO3, MeCN, reflux, 24 h.
b) i) PMe3, THF, 08C, 30 min. ii) THF/H2O, 20 8C, 20 h. c) OCN-
(CH2)6NCO, CHCl3, 1008C, 30 h. d) H2, PtO2, THF, 208C, 20 h.
e) ArNCO, CHCl3, reflux, 24 h.
878
www.chemeurj.org
ꢀ 2006Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 877 – 886

Bicapsular Aggregates
FULL PAPER
reacted with the corresponding aryl isocyanate to give 2a,b
in 58–60% yield.
matic protons of the pendant p-tolyl groups afford the usual
AA’XX’ pseudodoublets in the region of d=7.00–7.30 ppm.
These chemical shifts for 2a are similar to those found for
1a in [D6]DMSO (Figure 2a).[10a,c]
1
1H NMR spectroscopy: Figure 2 shows H NMR spectra for
solutions of 1a and 2a in [D6]DMSO and CDCl3 . In a
strongly hydrogen-bonding solvent, such as [D6]DMSO, 1a
and 2a exist as monomeric, nonassembled species. For 2a
(Figure 2b), the two types of benzylic methylene groups
appear as two singlets at d=3.56and 3.59 ppm. The p-tolyl
methyl groups give a singlet at d=2.21 ppm, and the aro-
The change to a noncompetitive solvent, such as CDCl3,
produces significant alterations in the solution structures of
1a and 2a, as indicated by the H NMR spectra (Figure 2c
1
and d). As mentioned in the introduction, the triurea 1a has
been previously investigated.[10a,c] A combination of techni-
ques (X-ray analysis, NMR and IR spectroscopy, and ESI-
MS) provided evidence for a dimeric, capsular structure
1a·1a, both in the solid-state and in CDCl3.[10a,c] Three as-
pects of that research are of relevance for the present dis-
cussion: 1) When 1a is dissolved in CDCl3, the 1H NMR
spectrum (Figure 2c) shows a loss of symmetry with respect
to the spectrum in [D6]DMSO (Figure 2a). 2) The resonan-
ces of the aromatic and methyl protons of the p-tolyl groups
in CDCl3, are shifted to lower frequencies with respect to
those in [D6]DMSO. This effect can be attributed to local
anisotropic effects within the capsule 1a·1a.[10a,c] 3) The
ROESY spectrum in CDCl3 (see Supporting Information,
Figure S.1[10c]) shows NOE contacts between the aromatic
and methylenic protons of the tribenzylamine moiety and
the aromatic protons of the p-tolyl groups.[10c] As shown in
Figure S.1,[10c] these contacts nicely support the formation of
a capsular dimeric aggregate 1a·1a, as they are difficult to
explain in a monomeric 1a.[10c] Indeed, these NOE contacts
are not present in the ROESY spectrum of 1a in
[D6]DMSO, in which the compound exists as a nonassem-
bled monomer.[10c]
Although compound 2a was purified by chromatography
(EtOAc/hexane 1:3), and was thus expected to be a single
1
species, the H NMR spectrum of 2a in CDCl3 is poorly re-
solved and shows a mixture of different species (Figure 2d).
By a comparison of this spectrum with that of 1a·1a in
CDCl3 (see Figure 2c and the previous paragraph), the
broad resonances at approximately d=1.90, 6.20, and
6.50 ppm can be assigned to the methyl and aromatic pro-
tons of the p-tolyl groups in a capsular derivative of 2a. As
shown in Figure 1, this derivative could, in principle, be a
self-assembled monomer, a dimer, or an oligomer. Other
signals in this spectrum might belong to a nonassembled
monomeric species, for example, the peak at d=2.26ppm is
in agreement with the methyl signal observed for 2a in
[D6]DMSO, in which this compound is a nonassembled
monomer. All the resonances in CDCl3 are broad and com-
plex, pointing to a distribution of isomers, both for the as-
sembled and nonassembled species. Figure 3 shows a com-
1
parison between the methyl region from the H NMR spec-
trum of 2a in CDCl3 at 2”10ꢀ3 m (Figure 3a, 700 MHz) and
at 0.020m (Figure 3b, 600 MHz, expansion from Figure 2d).
At the lower concentration, the nonassembled species (indi-
cated by the higher resonance frequencies) are predominant.
Figure 2. 1H NMR spectra (300 and 600 MHz) of 0.020m solutions of
a) 1a in [D6]DMSO, b) 2a in [D6]DMSO, c) 1a in CDCl3, and d) 2a in
CDCl3. A and N-A mark the resonances exclusively due to assembled
and nonassembled species, respectively. The asterisk labels the signal for
residual water.
1
We will return to the H NMR spectrum of 2a in a later dis-
cussion of the NOEs.
1
The H NMR spectra of 2b (which contains an n-butyl in-
stead of a methyl substituent) in [D6]DMSO and CDCl3 are
Chem. Eur. J. 2006, 12, 877 –W88il6eꢀy-V20C0H6 Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
879

M. Alajarín, A. Pastor et al.
Figure 4. 1H NMR spectrum of a) 2b in [D6]DMSO (0.020m, 600 MHz)
and b) 2b in CDCl3 (1”10ꢀ3 m, 500 MHz). The asterisk labels the signal
for residual water.
Table 1. Diffusion coefficients (D [m2 sꢀ1]) and hydrodynamic radii (rH
[Š]) of 1a, 2a, 2b, and 6 in CDCl3 (2”10ꢀ3 m) at 299 K.[a]
Figure 3. Methyl region of the 1H NMR spectrum of 2a in a) CDCl3 (2”
10ꢀ3 m), measured at 700 MHz and b) CDCl3 (0.020m), measured at
600 MHz. The complexity and line width of the signals suggest several
isomers of both species, that is assembled and nonassembled.
[c]
d [ppm]
D[b] ”10ꢀ10
rH
1a
2a
assembled (1a·1a)
assembled (2a·2a)
nonassembled (monomer)
mixture
assembled (2b·2b)
nonassembled (monomer)
precursor
1.89, 7.43, 7.965.24
1.93, 6.22, 6.46
2.264.28
7.0–7.3
6.5
7.8
3.31
9.6
12.4
4.11[d]
3.16
4.04
5.76
–
2b
6
13.0
10.1
7.1
shown in Figure 4. Based on the anisotropy criterion men-
tioned above, the multiplets between d=6.10 and 6.60 ppm
correspond to the aromatic protons of the p-BuC6H4 groups
in an assembled species (monomer, dimer, or oligomer). At
first sight, it is not clear whether nonassembled, monomeric
2b is also present in this solution. However, since the inten-
sity of the weak resonance at d=2.5 ppm increases with
added DMSO, we suggest that this weak signal might corre-
spond to the butyl CH2 group in nonassembled 2b (see
below for further discussion).
2.5
6.61, 6.73
[a] h=0.534”10ꢀ3 Kg sꢀ1 mꢀ1 (CDCl3, 299 K). [b] Experimental error is
approximately ꢁ2%. [c] Standard deviation is approximately ꢁ0.1 Š.
[d] Average of values for d=7.06(4.08”10 ꢀ10 m2 sꢀ1), 7.19 (4.19”
10ꢀ10 m2 sꢀ1), and 7.22 ppm (4.06”10ꢀ10 m2 sꢀ1).
molecules, this number corresponds to a mass ratio of ap-
proximately 2:1.[17b,21] As the nonassembled molecules must
be monomers, a dimeric structure is proposed for the aggre-
gates. We do not find polymers of 2a in CDCl3, at least not
at measurable concentrations. The PGSE measurements on
the aromatic region (d=7.0–7.3 ppm) of this mixture 2a/
2a·2a afford a D value (4.11”10ꢀ10 m2 sꢀ1) that is only slight-
ly smaller than the value assigned to the monomer (4.28”
10ꢀ10 m2 sꢀ1, Table 1). Indeed, in this aromatic region the
most intense signals are expected to stem from the p-tolyl
protons of the monomer, which at 2”10ꢀ3 m represents the
predominant species (ratio 2a:2a·2a 1:0.4, Figure 3a).
For the solution of 2b in CDCl3, the D value measured
using the resonances at d=6.5 ppm (3.16”10ꢀ10 m2 sꢀ1,
Table 1) is similar to that found for 2a·2a (3.31”
10ꢀ10 m2 sꢀ1), confirming the assignment of these signals to
the dimeric species 2b·2b. The substitution of Me in 2a·2a
with nBu in 2b·2b produces the expected modest decrease
Clearly, although the proton spectra for 2a and 2b in
CDCl3 are informative, the nature of the aggregates remains
uncertain.
PGSE measurements: To estimate the molecular volumes
(and thus the aggregation states) of the two different species
present in CDCl3, we have carried out PGSE diffusion
measurements on solutions of 2a and 2b in CDCl3. Solu-
tions of the triurea 1a and the precursor 6 in CDCl3 (shown
in Scheme 2) were also studied (Table 1 and Figure 5).
Table 1 includes the calculated hydrodynamic radii, obtained
from the D values by the Stokes–Einstein equation.[20] The
chemical shifts of the peaks used in the determination of the
D values are indicated in the Table.
The ratio between the D values for the assembled and
nonassembled species in 2a is 4.28/3.31=1.29. For spherical
880
www.chemeurj.org
ꢀ 2006Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 877 – 886

Bicapsular Aggregates
FULL PAPER
PGSE results for both solvents. For 6, the aggregation state
of which does not change with the solvent, the rH value in
[D6]DMSO is 1 Š larger than in CDCl3, and we attribute
this difference to solvation and/or hydrogen bonding by the
[D6]DMSO solvent. Since the solvation should be at least as
strong for 1a, 2a, and 2b as for 6, we have made a crude
empirical correction of ꢀ1 Š for all of the rH values ob-
tained in [D6]DMSO. The resulting radii (rc in Table 2) are
the suggested upper limits for the molecular radii of the spe-
cies present in [D6]DMSO. These rC values for 2a and 2b
are similar to the rH values obtained for the monomers of
these species in CDCl3 (also shown in Table 2).
Figure 5. Plot of ln(I/Io) versus arbitrary units proportional to the square
of the gradient amplitude for 1H PGSE diffusion measurements on solu-
tions of 2a (circles), 1a (triangles), and 6 (squares) in CDCl3 (2”10ꢀ3 m).
For 2a, the full circles correspond to the measurements on the signals of
the dimer (d=1.93, 6.22, and 6.46 ppm) and the open circles correspond
to the measurements on the signals of the monomer at d=2.26ppm. For
1a, all of the compound is present as a dimer at this concentration. The
NMR spectroscopic parameters are the same for all of the reproduced
measurements (D=93 ms, d=1.75 ms, 16scans, for an explanation see
Experimental Section).
We were curious to know if the average D value of the
solvent [D6]DMSO would be affected by the presence of 1a,
2a, 2b, or 6 in the solution, due to hydrogen bonding with
the solute. Table 3 shows the D values of a reference sample
Table 3. Diffusion coefficients (D [m2 sꢀ1]) for the residual protonated
solvent in solutions of 1a, 2a, and 6 in [D6]DMSO (2”10ꢀ3 m) at 299 K.
D[a] ”10ꢀ10
D[a] ”10ꢀ10
[b]
DMSO
1a
2a
6.55
6.58
6.54
2b
6
–
6.57
in D. For the weak resonance at d=2.5 ppm, D=4.04”
10ꢀ10 m2 sꢀ1 and rH =10.1 Š support the tentative assignment
of this resonance to the monomer. Other signals from the
spectrum of 2b in CDCl3 afforded intermediate D values,
presumably, due to an overlap of the peaks produced by the
monomer and dimer.
[a] Experimental error is approximately ꢁ2%. [b] The D value for the
solvent in a solution of 2b in [D6]DMSO could not be determined, be-
cause the signal from the residual protonated solvent was superimposed
with signals from 2b.
The triurea 1a exists only as the dimer 1a·1a in CDCl3 at
the concentration studied. The hydrodynamic radius of the
monomer of 2a (9.6Š) is larger than for 1a·1a (7.8 Š), al-
though the structures of these compounds seem at first sight
to be similar. The difference in size may be explained by the
presence of the tether linking group in 2a, as well as by a
more compact, assembled structure in 1a·1a. The precursor
6 is the smallest molecule shown in Table 1, although it is
almost as large as the compact assembly 1a·1a. As the
“smallest” compound, 6 reveals the larger slope in Figure 5.
In [D6]DMSO, 1a and 2a exist as monomeric, nonassem-
bled species (Table 2). As a result of the higher viscosity of
of DMSO, plus the D values of the solvent in the solutions
studied. Not surprisingly, in these relatively dilute solutions
(2”10ꢀ3 m), such an effect was not observed.
Molecular models: To compare the expected sizes of our
molecules with the results from the diffusion measurements,
we have used the program MacroModelꢁ 8.1 (AMBER*
force field) to build molecular models of the new com-
pounds.
For 2a·2a, the results from the molecular modeling con-
ducted in CHCl3 led to the energy-minimized structure de-
picted in Figure 6(the energy-minimized structure for 2b·2b
can be found in the Supporting Information). The calculated
structure for 2a·2a reveals a pair of classical capsular aggre-
gates of “almost S6 symmetry”, linked by two hexamethy-
lene tethers. However, this is only one of, eventually, many
local energy minima. This variety of conformations, together
with the possibility of different diastereomers arising from
the clockwise or counterclockwise orientation of each ring
of the hydrogen-bonded ureas in 2·2[22] might explain the
multiple resonances found in the 1H NMR spectrum of
these species in CDCl3.
Table 2. Diffusion coefficients (D [m2 sꢀ1]) and hydrodynamic radii (rH
[Š]) of 1a, 2a, 2b, and 6 in [D6]DMSO (2”10ꢀ3 m) at 299 K. The rH
values for the monomers in CDCl3 are given for comparison.[a]
[c]
[d]
[c]
D[b] ”10ꢀ10
rH
rC
rH (monomer, CDCl3)
[e]
1a
2a
2b
6
1.51
1.05
1.03
1.40
7.5
10.7
11.0
8.1
6.5
9.7
10.0
7.1
–
9.6
10.1
7.1
[a] h=1.940”10ꢀ3 Kg sꢀ1 mꢀ1 ([D6]DMSO, 299 K). [b] Experimental error
is approximately ꢁ2%. [c] Standard deviation is approximately ꢁ0.1 Š.
[d] Corrected for [D6]DMSO solvation. [e] For 1a in CDCl3 there was no
monomer present.
It has not been possible to obtain a convergent energy-
minimized structure for the self-assembled monomers from
1
2a and 2b, which have not been detected in the H NMR
this solvent, all the D values in [D6]DMSO are considerably
smaller than in CDCl3. However, as the calculated rH values
are viscosity-corrected,[20] they allow for a comparison of the
spectra taken in CDCl3. It seems that such a folded confor-
mation for these hexaureas, to afford unimolecular capsules,
is strongly disfavored.
Chem. Eur. J. 2006, 12, 877 –W88il6eꢀy-V20C0H6 Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
881

M. Alajarín, A. Pastor et al.
figure). The other encircled peaks in the figure mark the
NOE cross-peaks between the ortho- and meta-protons of
the p-tolyl groups in 2a·2a, as well as between the meta-pro-
tons and the adjacent NH groups, at approximately d=
8.00 ppm. This high NH chemical shift is typical for ureido
NH groups engaged in strong hydrogen bonding (in 1a·1a,
these NH protons appear at d=7.94 ppm).
The cross-peaks surrounded by rectangles in Figure 7a are
particularly important, as they support the capsular geome-
try proposed for 2a·2a in CDCl3. These cross-peaks corre-
late the protons associated with the tribenzylamine moiety
(methylene groups at d=3.10–3.60 ppm and aromatic pro-
tons at d=7.00–7.50 ppm) with the aromatic protons of the
pendant p-tolyl groups of 2a·2a (at approximately d=6.20
and 6.50 ppm). By analogy to 1a (see Figure S1 in the Sup-
porting Information),[10c] these interactions point to a dimer-
ic, bicapsular structure, 2a·2a, in which the p-tolyl groups of
one 2a molecule are close to the tribenzylamine moiety of
the other.
Figure 6. Energy-minimized structure (MacroModelꢁ 8.1, AMBER*
force field) of 2a·2a. All hydrogen atoms have been omitted for clarity.
The files generated in MacroModelꢁ were introduced into
the Chem 3Dꢁ program, in which the molecular radii can be
estimated by measuring distances between atoms. In addi-
tion to 2a·2a and 2b·2b, an energy-minimized model of
Figure 7b shows the ROESY spectrum for a solution of
2b in CDCl3. The black arrow in the figure marks, for
2b·2b, the intramolecular NOEs between the ortho and the
methylene protons of the p-BuC6H4 groups. The weak cross-
peak marked by the white arrow might correspond to the
analogous interaction within the monomer 2b, which is pres-
ent at a very low concentration in this solution. The rectan-
gles in Figure 7b surround the peaks assigned to intermolec-
Table 4. Estimated hydrodynamic radii for the dimeric assemblies.
rH (model)[a] [Š]
rH (PGSE)[b] [Š]
1a·1a
2a·2a
2b·2b
7.67.8
12.3
13.5
12.4
13.0
[a] Determined by Chem 3Dꢁ from the calculated structures in CDCl3.
[b] Determined from the D values in CDCl3.
1a·1a was also generated in
Chem 3Dꢁ. For these elongated
molecules, the hydrodynamic
radii in CDCl3 were considered
to be approximately 85% of
the rotational radii, according
to a suggestion made by Matti-
son et al.[23] In spite of the
crude approach,[23] the estimat-
ed
hydrodynamic
radii
(Table 4) are in fairly good
agreement with the results from
the PGSE measurements.
Exchange processes and NOE:
The ROESY spectrum of a so-
lution of 2a in CDCl3 (Fig-
ure 7a) shows NOE cross-peaks
between the methyl and the
ortho-protons of the p-tolyl
groups (marked by arrows in
the figure). These contacts con-
firm that these ortho-protons
resonate at approximately d=
7.00 ppm for the nonassembled
monomer 2a (M in figure 7a),
and at a lower frequency, ap-
proximately d=6.50 ppm, for
Figure 7. a) ROESY spectrum of a solution of 2a in CDCl3 (500 MHz, 2”10ꢀ3 m), showing the cross-peaks re-
sulting from the NOE contacts within the monomer 2a (M) and the dimer 2a·2a (D). The cross-peaks be-
tween the ortho-protons and the methyl protons of the p-tolyl groups are marked with arrows. The intramolec-
ular cross-peaks within each unit of 2a in 2a·2a are surrounded with circles and ellipses, while the intermolec-
ular cross-peaks between the two units of 2a in 2a·2a are marked with rectangles. b) ROESY spectrum of a
CDCl3 solution of 2b (500 MHz, 5”10ꢀ3 m), showing the cross-peaks resulting from the NOE contacts within
the monomer 2b (M) and the dimer 2b·2b (D). The contacts between the ortho and some of the methylene
protons of the BuC6H4 groups are marked with arrows. The intramolecular cross-peaks within each unit of 2b
in 2b·2b are surrounded with circles and ellipses, while the intermolecular cross-peaks between the two units
the dimer 2a·2a (D in the of 2b in 2b·2b are marked with rectangles.
882
www.chemeurj.org
ꢀ 2006Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 877 – 886

Bicapsular Aggregates
FULL PAPER
ular NOE interactions between the tribenzylamine moiety
and the p-BuC6H4 groups, which provides further evidence
for the capsular structure of 2b·2b.
m/z=1533. In this case a MS/MS spectrum of the signal at
m/z=3065 could not be measured due to its low intensity.
All the protons involved in intermolecular NOE contacts
within 2a·2a and 2b·2b are separated by less than 3.50 Š in
the energy-minimized structures. When no intermolecular
NOE cross-peaks are observed (e.g., between the methylen-
ic protons of the tribenzylamine moiety and the protons
ortho to the Me/Bu substituents), the interatomic distances
in the models are larger than 3.50 Š.
Equilibria: 1D NMR spectroscopic saturation experiments
on a solution of 2a in CDCl3 did not reveal saturation trans-
fer between the methyl signals of the monomer and dimer,
pointing to a rate constant for the exchange that is smaller
than 0.01 sꢀ1. Similarly, the ROESY spectra of 2a and 2b in
CDCl3 (Figure 7) did not show cross-peaks due to exchange
processes.
Nevertheless, we have observed that an increase in con-
centration favors the formation of the dimer (compare the
two spectra of 2a in Figure 3). The equilibrium is reached
within minutes, and the ratio of species remains unchanged
for more than one month. For 2b, the equilibrium is dis-
placed towards the dimer.
ESI mass spectra: The ESI mass spectrum of the hexaurea
2a measured in CHCl3 (1.00mm) produced a base peak cor-
responding to [2a·2a+H]+ at m/z=2730 and the signal for
the protonated monomer at m/z=1366 (Figure 8a).[24] The
To provide semiquantitative data on the relative stability
of the dimers, their solutions in CDCl3 were titrated with
1
[D6]DMSO, and the process followed by H NMR spectro-
scopic analysis. For 2a·2a, 2%vol [D6]DMSO was required
for 50% dissociation (i.e. [Dimer]/[Monomer]ffi1), while for
1a·1a and 1b·1b 8–9%vol [D6]DMSO was required to
reach the same level of dissociation. These results reflect the
lower stability of 2a·2a relative to 1a·1a and 1b·1b. For
2b·2b, the point of 50% dissociation cannot be distinguish-
1
ed by H NMR spectroscopy, as a result of the superposition
of the signals from the monomer and dimer. However, titra-
tion with [D6]DMSO until the point of complete dissociation
afforded a result of 19%vol [D6]DMSO[25] needed for
2a·2a, versus 29%vol [D6]DMSO for 2b·2b,[25] indicating
the higher relative stability of the later. For 1a·1a and
1b·1b 33–34%vol [D6]DMSO was required for total de-
struction of the dimer.[10a,c,25] The dimers of calix[4]arenes,
bearing four urea functions at the wider rim, are usually de-
stroyed by 2–10%vol of DMSO added to an apolar sol-
vent.[19b,26]
The dimerization constant (KD)[27] for 2a, was determined
at three different concentrations (0.040–0.020m) from the in-
tegrals of the 1H NMR spectrum measured at 298 K, and
gave values of 440, 540, and 610mꢀ1. However, these num-
bers can be taken only as a rough estimate, since the KD
values increased with the decreasing concentration of the
hexaurea. These results points to the presence of small
quantities of oligomeric species, which may lead to very
1
broad and undetectable signals in the H NMR spectra (see
above).[26c]
We had previously observed[10c] that the stability of the
dimers 1·1 is strongly dependent on the terminal substitu-
ents of the urea functionality (KD =91200mꢀ1 for 1a and
KD =4000mꢀ1 for 1c) and, to lesser extent, on the substitu-
ents located on the tribenzylamine skeleton (KD =1000mꢀ1
for 1d). The KD value found for the hexaurea 2a is slightly
lower than that of the triurea 1d.
Figure 8. ESI mass spectrum, in the range of 800–4000, of a CHCl3 solu-
tion (1.00 mm) of a) 2a and b) 2b.
MS/MS (magnetic sector mass spectrometry) spectrum for
the signal at m/z=2730 was also measured and produced a
peak at m/z=1365, which was assigned to the monomer.
For 2b, the ESI mass spectrum in CHCl3 (1.00mm) pro-
duced a signal for the protonated dimer [2b·2b+H]+ at
m/z=3065 (Figure 8b).[24] In addition, the signal for the pro-
tonated monomer was also observed as the base peak at
Chem. Eur. J. 2006, 12, 877 –W88il6eꢀy-V20C0H6 Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
883

M. Alajarín, A. Pastor et al.
higher viscosity of this solvent. The relaxation delay was always 5 s and
the number of scans 16(except for the measurements on the signal at d=
2.5 ppm, in the CDCl3 solution of 2b, for which 80 scans per increment
were required). Typically, 14–20 points were used for the regression anal-
ysis and the experimental time was approximately 40 min. All of the ob-
served data leading to the reported D values afforded lines whose corre-
lation coefficients were above 0.999.
Conclusion
The hexaureas 2a,b, formed by two tris(2-ureidobenzyl)a-
mine moieties linked by an hexamethylene tether, are prone
to assemble in CDCl3 to afford bicapsular dimeric aggre-
gates, 2a·2a and 2b·2b, which both contain two rings of six
hydrogen-bonded ureas. In solution, these dimers are in
equilibrium with the nonassembled monomers 2a and 2b,
but show kinetic stability on the NMR timescale. No poly-
mers or self-assembled monomers were detected in CDCl3,
at least not in measurable concentrations. The thermody-
namic stabilities of 2a·2a and 2b·2b are comparable to
those of other capsular aggregates derived from tribenzyla-
mines. The characterization of these new supramolecules in
solution was carried out by 1H NMR and 1H/1H-ROESY
spectroscopic studies, and PGSE diffusion studies. The
PGSE diffusion methodology has been shown once more to
be an extremely efficient tool for studying supramolecular
interactions in solution. As far as we know, the dimeric spe-
cies 2·2 constitutes as a unique example of a bicapsular ag-
gregate, which is associated strictly by hydrogen bonding.
This new type of assembly demonstrates the efficiency of
the tribenzylamine subunit in the design of new supramolec-
ular systems.
All molecular mechanics calculations were carried out by using the
AMBER* force field as implemented within Maestro/MacroModelꢁ 8.1.
Standard potentials and atomic charges, as provided by the AMBER*
force field, were employed without modifications. AMBER* and
OPLAA force fields produce essentially the same results in related struc-
tures. Calculations were initially performed under vacuum and then in
chloroform (GB/SA solvation model). Most complex structures were vir-
tually identical under both conditions. Energy minimizations were con-
ducted over 500 iterations on a silicon graphics computer. Minimized
structures were then subjected to conformational searches with 5000-step
Monte Carlo multiple minimum simulations. All conformations within
15 kJmolꢀ1 of the computed global minimum were stored, and the repre-
sentative lowest energy structure was analyzed.
The ROESY experiments were measured on solutions of 2a (2”10ꢀ3 m)
and 2b (5”10ꢀ3 m) in CDCl3, with a 500 MHz Bruker AVANCE spec-
trometer, equipped with a multinuclear inverse probe. A spin-lock pulse
of 400 ms was used. The number of scans was 128 for 2a and 280 for 2b.
Total experimental times were approximately 20 h.
CAUTION: Azido compounds may represent an explosion hazard when
being concentrated under vacuum or stored neat. A safety shield and ap-
propriate handling procedures are recommended.
Bis(2-nitrobenzyl)(2-azidobenzyl)amine
(3):
2-Azidobenzylamine[30]
(1.60 g, 10.8 mmol) and 2-nitrobenzyl iodide[10c] (5.68 g, 21.6 mmol) were
added, as solutions in dry acetonitrile (2 and 5 mL respectively), to a sus-
pension of Na2CO3 (6.58 g, 62.1 mmol) in acetonitrile (10 mL), and the
reaction mixture stirred under reflux for 24 h. After cooling, the inorgan-
ic salts were filtered and washed with cold acetonitrile (5”5 mL). The fil-
trate was collected, the solvent was removed, and the residue was puri-
fied by silica-gel chromatography, eluting with EtOAc/hexanes 1:6( Rf =
0.34), to afford 3 (93% yield) as pale yellow prisms (an analytical sample
was obtained by recrystallization from CH2Cl2/Et2O 1:3). M.p. 84–858C;
1H NMR (200 MHz, CDCl3, 258C, TMS): d=3.55 (s, 2H), 3.91 (s, 4H),
7.04–7.11 (m, 2H), 7.22–7.37 (m, 4H), 7.51 (td, 3J(H,H)=7.5 Hz,
4J(H,H)=1.3 Hz, 2H), 7.65 (dd, 3J(H,H)=7.7 Hz, 4J(H,H)=1.1 Hz, 2H),
7.78 ppm (dd, 3J(H,H)=8.0 Hz, 4J(H,H)=1.3 Hz, 2H); 13C NMR
(50 MHz, CDCl3, 258C): d=54.2 (t), 55.6(2t), 118.1 (d), 124.3 (2d),
124.6(d), 127.9 (2d), 128.6(s), 128.9 (d), 131.0 (2d), 131.4 (d), 132.6
(2d), 134.1 (2s), 138.9 (s), 149.6ppm (2s); IR (Nujol): n˜ =2131 (N3),
1531 (NO2), 1522, 1342 cmꢀ1 (NO2); MS (70 eV, EI): m/z (%): 419 (43)
[M+1]+, 418 (11) [M]+, 286 (66), 105 (100); elemental analysis calcd (%)
for C21H18N6O4 (418.4): C 60.28, H 4.34, N 20.09; found: C 60.01, H 4.50,
N 20.45.
Experimental Section
General: 1H and 13C NMR spectra were measured on Bruker AC 200
(1H: 200 MHz, 13C: 50 MHz), Varian Unity-300 (1H: 300 MHz, 13C:
75 MHz), Bruker AVANCE 600 (1H: 600 MHz, 13C: 151 MHz), and
Bruker AVANCE 700 (1H: 700 MHz) spectrometers with TMS (d=
0.00 ppm) or the solvent residual peak as internal standards. IR spectra
were recorded on a FT–IR Nicolet Impact 400 infrared spectrometer and
melting points were taken on a Reichert apparatus and are not corrected.
The PGSE NMR diffusion measurements were carried out by using the
stimulated echo pulse sequence,[18b] as has been explained elsewhere.[17c,28]
All the experiments were performed on a 500 MHz Bruker AVANCE
spectrometer, equipped with a microprocessor-controlled gradient unit
and a multinuclear inverse probe with an actively shielded Z-gradient
coil. The sample was not spun and the airflow was disconnected. The
shape of the gradient pulse was rectangular, and its strength varied auto-
matically during the course of the experiments. The D values were deter-
mined from the slope of the regression line ln(I/Io) versus G2, according
to Equation (1).
Bis(2-nitrobenzyl)(2-aminobenzyl)amine (4): PMe3 in toluene (1.0m,
11.7 mL, 11.7 mmol) was slowly added at 08C to a solution of 3 (4.10 g,
9.8 mmol) in freshly distilled THF (30 mL) under N2. The reaction mix-
ture was then stirred at this temperature for 30 min. After this time, H2O
(15 mL) was added and the reaction mixture was stirred at 208C for a
further 20 h. After removal of the organic solvent, H2O (40 mL) was
added and the aqueous phase was extracted with CH2Cl2 (3”20 mL).
The combined extracts were dried over MgSO4, the solvent was evaporat-
ed, and the residue was purified by silica-gel chromatography, eluting
with EtOAc/hexanes 1:4 (Rf =0.16), to afford 4 (67–90% yield) as yellow
prisms (an analytical sample was obtained by recrystallization from
CHCl3/Et2O 1:4). M.p. 96–998C; 1H NMR (200 MHz, CDCl3, 258C,
TMS): d=3.55 (s, 2H), 3.85 (s, 4H), 4.13 (s, 2H), 6.53–6.65 (m, 2H),
6.98–7.06 (m, 2H), 7.22–7.31 (m, 2H), 7.37–7.48 (m, 4H), 7.66 ppm (dd,
3J(H,H)=8.1 Hz, 4J(H,H)=1.0 Hz, 2H); 13C NMR (50 MHz, CDCl3,
258C): d=56.3 (2t), 59.4 (t), 115.6 (d), 117.8 (d), 121.4 (s), 124.2 (2d),
128.2 (2d), 128.9 (d), 131.2 (d), 132.0 (2d), 132.6(2d), 133.4 (2s), 146.5
(s), 149.7 ppm (2s); IR (Nujol): n˜ =3486(NH), 3390 (NH), 1623, 1522
(NO2), 1340 cmꢀ1 (NO2); MS (70 eV, EI): m/z (%): 393 (42) [M+1]+, 375
2
lnðI=IoÞ ¼ ꢀðgdÞ G2ðD-d=3ÞD
ð1Þ
In Equation (1) I/Io =observed spin echo intensity/intensity without gra-
dients, G=gradient strength, D=delay between the midpoints of the gra-
dients, D=diffusion coefficient, and d=gradient length.
The calibration of the gradients was carried out by a diffusion measure-
ment of HDO in D2O (DHDO =1.9”10ꢀ9 m2 sꢀ1).[29] The values reported in
Tables 1–3 are an average of three different measurements, which yielded
D values within a maximum of ꢁ1.5% of the reported value. All the
1
measurements were carried out by using the H NMR spectroscopic reso-
nances. For the measurements in CDCl3, the gradient length (d) was set
to 1.75 ms and the diffusion delay was set to approximately 68, 93, and
168 ms. For the experiments in [D6]DMSO the diffusion delays were
longer (approximately 168, 218, and 268 ms), to compensate for the
884
www.chemeurj.org
ꢀ 2006Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 877 – 886

Bicapsular Aggregates
FULL PAPER
(39), 286(47), 257 (82), 120 (100); HRMS (EI):
m/z: calcd for
137.2 (s), 137.9 (s), 152.9 (s), 155.9 ppm (s), three signals are overlapped;
IR (Nujol): n˜ =3324 (NH), 1655 cmꢀ1 (C=O); MS (FAB): m/z (%): 1364
(4) [M]+, 625 (11), 239 (89), 132 (100); elemental analysis calcd (%) for
C82H88N14O6 (1365.7): C 72.12, H 6.49, N 14.36; found: C 71.62, H 6.52, N
14.50.
C21H20N4O4: 392.148455; found 392.148647; elemental analysis calcd (%)
for C21H20N4O4 (392.4): C 64.28, H 5.14, N 14.28; found: C 64.05, H 5.15,
N 14.33.
1-{2-{[Bis(2-nitrobenzyl)amino]methyl}phenyl}-3-{6-{3-{2-{[bis(2-nitroben-
zyl)amino]methyl}phenyl}ureido}hexyl}urea (5): 1,6-Diisocyanatohexane
(0.42 g, 2.5 mmol) was added, under N2, to a solution of the amine 4
(1.94 g, 4.9 mmol) in dry CHCl3 (60 mL). After stirring at 1008C for 30 h
in a sealed tube the solvent was removed and the residue was purified by
silica-gel chromatography, eluting with EtOAc/hexanes 1:1!4:1 (Rf =
0.16in EtOAc/hexanes 1:1), to afford 5 (69% yield) as pale yellow
prisms (an analytical sample was obtained by recrystallization from
CH2Cl2/Et2O 1:1). M.p. 99–1048C; 1H NMR (300 MHz, CDCl3, 258C,
TMS): d=1.52–1.54 (m, 4H), 1.65–1.71 (m, 4H), 3.37 (q, 3J(H,H)=
6.5 Hz, 4H), 3.69 (s, 4H), 3.90 (s, 8H), 5.77 (t, 3J(H,H)=5.3 Hz, 2H),
6.85 (td, 3J(H,H)=7.5 Hz, 4J(H,H)=1.1 Hz, 2H), 7.07 (dd, 3J(H,H)=
7.4 Hz, 4J(H,H)=1.4 Hz, 2H), 7.13–7.25 (m, 10H), 7.38 (td, 3J(H,H)=
7.6Hz, 4J(H,H)=1.2 Hz, 4H), 7.44 (s, 2H), 7.59 (dd, 3J(H,H)=8.1 Hz,
4J(H,H)=1.2 Hz, 4H), 7.96ppm (dd, 3J(H,H)=8.1 Hz, 4J(H,H)=0.6Hz,
2H); 13C NMR (75 MHz, CDCl3, 258C): d=26.7 (2t), 30.0 (2t), 40.2 (2t),
57.3 (4t), 61.2 (2t), 121.1 (2d), 121.9 (2d), 124.3 (4d), 124.6 (2s), 128.2
(4d), 129.0 (2d), 130.9 (2d), 132.0 (4d), 133.1 (4d), 133.4 (4s), 139.2 (2s),
148.9 (4s), 155.4 ppm (2s); IR (Nujol): n˜ =3297 (NH), 1622 (C=O), 1522
(NO2), 1342 cmꢀ1 (NO2); MS (FAB): m/z (%): 953 (12) [M+1]+, 818
(10), 666 (12), 419 (49), 106 (100); elemental analysis calcd (%) for
C50H52N10O10 (953.0): C 63.01, H 5.50, N 14.70; found: C 63.05, H 5.56, N
14.71.
Hexaurea 2b: 4-Butylphenyl isocyanate (0.21 g, 1.20 mmol) was added,
under N2, to a solution of 6 (0.25 g, 0.30 mmol) in dry CHCl3 (25 mL).
After stirring under reflux for 24 h the solvent was removed and the resi-
due purified by recrystallization from EtOH to afford 2b (58% yield) as
1
colorless prisms. M.p. 164–1668C; H NMR (600 MHz, [D6]DMSO, 258C,
TMS): d=0.86(t, 3J(H,H)=7.4 Hz, 12H), 1.22–1.29 (m, 12H), 1.37 (m,
4H), 1.49 (quint, 3J(H,H)=7.6Hz, 8H), 2.47 (t, 3J(H,H)=7.6Hz, 8H),
2.99 (q, 3J(H,H)=6.0 Hz, 4H), 3.56 (s, 4H), 3.58 (s, 8H), 6.28 (brs, 2H),
6.95 (t, 3J(H,H)=7.4 Hz, 2H), 7.03–7.04 (m, 12H), 7.08 (t, 3J(H,H)=
7.6Hz, 2H), 7.17 (t, 3J(H,H)=7.5 Hz, 4H), 7.29 (d, 3J(H,H)=8.4 Hz,
8H), 7.46(d, 3J(H,H)=7.4 Hz, 2H), 7.50 (d, 3J(H,H)=7.5 Hz, 4H), 7.53
(d, 3J(H,H)=8.1 Hz, 2H), 7.56(d, 3J(H,H)=8.0 Hz, 4H), 7.70 (s, 2H),
7.86(s, 4H), 8.66ppm (s, 4H);
13C NMR (151 MHz, [D6]DMSO, 258C):
d=13.8 (q), 21.7 (t), 26.2 (t), 29.7 (t), 33.3 (t), 34.2 (t), 39.3 (t), 54.5 (t),
54.7 (t), 118.2 (d), 123.0 (d), 123.6(d), 127.2 (d), 128.4 (d), 128.8 (d),
129.0 (d), 129.1 (s), 129.6(s), 135.6(s), 137.1 (s), 137.3 (s), 137.9 (s), 153.0
(s), 155.8 ppm (s), three signals are overlapped; IR (Nujol): n˜ =3310
(NH), 1653 cmꢀ1 (C=O); MS (FAB): m/z (%): 1534 (1) [M+1]+, 709 (6),
281 (100); elemental analysis calcd (%) for C94H112N14O6 (1534.0): C
73.60, H 7.36, N 12.78; found: C 73.24, H 7.42, N 12.81.
1-{2-{[Bis(2-aminobenzyl)amino]methyl}phenyl}-3-{6-{3-{2-{[bis(2-amino-
Acknowledgements
benzyl)amino]methyl}phenyl}ureido}hexyl}urea
(6):
PtO2
(0.67 g,
2.9 mmol) was added to a solution of 5 (1.40 g, 1.5 mmol) in freshly distil-
led THF (100 mL), and the resulting reaction mixture was stirred at 208C
for 20 h under H2. After filtration over a pad of celite, the solvent was re-
moved to afford the tetraamine 6 (91% yield) as colorless prisms. This
compound was deemed pure enough for the following step (an analytical
sample was obtained by recrystallization from CH2Cl2/Et2O 1:2). Com-
pound 6 could not be completely dried, even by placing under a high
temperature (1008C) and vacuum (0.15 Torr). M.p. 126–1288C; 1H NMR
(300 MHz, CDCl3, 258C, TMS): d=1.36–1.44 (m, 4H), 1.52–1.56 (m,
4H), 3.19 (q, 3J(H,H)=6.5 Hz, 4H), 3.41 (s, 8H), 3.48 (s, 4H), 3.83 (brs,
This work was supported by MCYT-FEDER (Project BQU2001–0010),
Fundación SØneca-CARM (Project 00458/PI/04), the Swiss National Sci-
ence Foundation, and the ETH Zurich. A.P. is grateful to the Ministerio
de Educación y Ciencia of Spain and the University of Murcia (Spain)
for a Ramón y Cajal contract. R.-A.O. thanks the Fundación Cajamurcia
for a fellowship. E.M.-V. thanks the Fundación SØneca-CARM for a
grant. We thank Dr. Heinz Rüegger for the helpful discussions and
advice about the NMR measurements, as well as for the acquisition of
1
the 700 MHz H NMR spectrum of 2a.
3
3
4
8H), 6.01 (t, J(H,H)=5.6 Hz, 2H), 6.61 (dd, J(H,H)=8.4 Hz, J(H,H)=
0.9 Hz, 4H), 6.73 (td, 3J(H,H)=7.4 Hz, 4J(H,H)=1.0 Hz, 4H), 6.88 (td,
3J(H,H)=7.5 Hz, 4J(H,H)=1.0 Hz, 2H), 7.06–7.12 (m, 10H), 7.18 (td,
3J(H,H)=7.8 Hz, 4J(H,H)=1.5 Hz, 2H), 7.81 (s, 2H), 8.03 ppm (d,
3J(H,H)=8.1 Hz, 2H); 13C NMR (75 MHz, CDCl3, 258C): d=26.3 (2t),
30.0 (2t), 39.6 (2t), 56.9 (4t), 58.0 (2t), 116.3 (4d), 118.8 (4d), 120.9 (2d),
121.80 (2d), 121.84 (4s), 124.9 (2s), 128.8 (2d), 129.3 (4d), 131.0 (2d),
132.4 (4d), 138.7 (2s), 144.8 (4s), 155.5 ppm (2s); IR (Nujol): n˜ =3373
(NH), 3313 (NH), 1677 cmꢀ1 (C=O); MS (FAB): m/z (%): 833 (7)
[M+1]+, 359 (7), 106(100); elemental analysis calcd (%) for
C50H60N10O2·H2O (851.1): C 70.56, H 7.34, N 16.46; found: C 70.48, H
7.13, N 16.53.
[1] a) C. A. Schalley, Adv. Mater. 1999, 11, 1535–1537; b) R. P. Sijbes-
ma, E. W. Meijer, Curr. Opin. Colloid Interface Sci. 1999, 4, 24–32;
c) L. J. Prins, D. N. Reinhoudt, P. Timmerman, Angew. Chem. 2001,
113, 2446–2492; Angew. Chem. Int. Ed. 2001, 40, 2382–2426;
d) E. A. Archer, A. E. Sochia, M. J. Krische, Chem. Eur. J. 2001, 7,
2059–2065.
[2] S. J. Geib, C. Vicent, E. Fan, A. D. Hamilton, Angew. Chem. 1993,
105, 83–85; Angew. Chem. Int. Ed. Engl. 1993, 32, 119–121.
[3] G. M. Whitesides, E. E. Simanek, J. P. Mathias, C. T. Seto, D. N.
Chin, M. Mammen, D. M. Gordon, Acc. Chem. Res. 1995, 28, 37–
44.
Hexaurea 2a: 4-Methylphenyl isocyanate (0.22 g, 1.68 mmol) was added,
under N2, to a solution of 6 (0.35 g, 0.42 mmol) in dry CHCl3 (20 mL).
After stirring under reflux for 24 h the solvent was removed and the resi-
due was purified by silica-gel chromatography, eluting with EtOAc/hex-
anes 1:3 (to remove N,N’-bis(4-methylphenyl)urea) !EtOAc, to afford
2a (Rf =0.85 in EtOAc; 60% yield) as colorless prisms (an analytical
sample was obtained by recrystallization from CH2Cl2/n-hexane 1:1).
M.p. 176–1788C; 1H NMR (600 MHz, [D6]DMSO, 258C, TMS): d=1.20–
1.26(m, 4H), 1.34–1.40 (m, 4H), 2.21 (s, 12H), 3.00 (q, 3J(H,H)=6.0 Hz,
4H), 3.56 (s, 4H), 3.59 (s, 8H), 6.29 (brs, 2H), 6.96 (t, 3J(H,H)=7.4 Hz,
2H), 7.02–7.05 (m, 12H), 7.09 (t, 3J(H,H)=7.6Hz, 2H), 7.17 (t,
3J(H,H)=7.6Hz, 4H), 7.29 (d, 3J(H,H)=8.3 Hz, 8H), 7.46(d, 3J(H,H)=
7.4 Hz, 2H), 7.51 (d, 3J(H,H)=7.5 Hz, 4H), 7.53 (d, 3J(H,H)=8.2 Hz,
2H), 7.56(d, 3J(H,H)=8.0 Hz, 4H), 7.71 (s, 2H), 7.86(s, 4H), 8.67 ppm
(s, 4H); 13C NMR (151 MHz, [D6]DMSO, 258C): d=20.3 (q), 26.2 (t),
29.7 (t), 39.3 (t), 54.4 (t), 54.7 (t), 118.2 (d), 123.0 (d), 123.5 (d), 123.6(d),
127.1 (d), 128.8 (d), 129.0 (s), 129.1 (d), 129.6(s), 130.5 (s), 137.1 (s),
[4] M. R. Ghadiri, J. R. Granja, R. A. Milligan, D. E. McRee, N. Khaza-
novich, Nature 1993, 366, 324–327.
[5] a) M. M. Conn, J. Rebek, Jr., Chem. Rev. 1997, 97, 1647–1668; b) V.
Bçhmer, M. O. Vysotsky, Aust. J. Chem. 2001, 54, 671–677; c) F.
Hof, S. L. Craig, C. Nuckolls, J. Rebek, Jr., Angew. Chem. 2002, 114,
1556–1578; Angew. Chem. Int. Ed. 2002, 41, 1488–1508.
[6] S. C. Zimmerman, F. Zeng, D. E. C. Reichert, S. V. Kolotuchin, Sci-
ence 1996, 271, 1095–1098.
[7] a) R. P. Sijbesma, F. H. Beijer, L. Brunsveld, B. J. B. Folmer, J. H. K.
Ky Hirschberg, R. F. M. Lange, J. K. L. Lowe, E. W. Meijer, Science
1997, 278, 1601–1604; b) H.-A. Klok, K. A. Jolliffe, C. L. Schauer,
L. J. Prins, J. P. Spatz, M. Mçller, P. Timmerman, D. N. Reinhoudt, J.
Am. Chem. Soc. 1999, 121, 7154–7155; c) R. K. Castellano, J. Re-
bek, Jr., J. Am. Chem. Soc. 1998, 120, 3657–3663; d) R. K. Castella-
no, C. Nuckolls, S. H. Eichhorn, M. R. Wood, A. J. Lovinger, J. Re-
bek, Jr., Angew. Chem. 1999, 111, 2764–2768; Angew. Chem. Int.
Ed. 1999, 38, 2603–2606; e) R. K. Castellano, R. Clark, S. L. Craig,
Chem. Eur. J. 2006, 12, 877 –W88il6eꢀy-V20C0H6 Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
885

M. Alajarín, A. Pastor et al.
C. Nuckolls, J. Rebek, Jr., Proc. Natl. Acad. Sci. USA 2000, 97,
12418–12421.
Kelvin, h=viscosity of the solution (for diluted solutions, the viscos-
ity of the solvent can be used), and r=hydrodynamic radius (the
radius of a hypothetical hard sphere that diffuses with the same
speed as the particle under examination). See: J. T. Edward, J.
Chem. Educ. 1970, 47, 261–270.
[8] P. Lipkowski, A. Bielejewska, H. Kooijman, A. L. Spek, P. Timmer-
man, D. N. Reinhoudt, Chem. Commun. 1999, 1311–1312.
[9] a) M. C. Etter, Z. UrbaÇczyk-Lipkowska, M. Zia-Ebrahimi, T. W.
Panunto, J. Am. Chem. Soc. 1990, 112, 8415–8426; b) X. Zhao, Y.-L.
Chang, F. W. Fowler, J. W. Lauher, J. Am. Chem. Soc. 1990, 112,
6627–6634; c) J. J. Kane, R.-F. Liao, J. W. Lauher, F. W. Fowler, J.
Am. Chem. Soc. 1995, 117, 12003–12004; d) S. Boileau, L. Bouteil-
ler, F. LauprÞte, F. Lortie, New J. Chem. 2000, 24, 845–848; e) F.
Lortie, S. Boileau, L. Bouteiller, Chem. Eur. J. 2003, 9, 3008–3014.
[10] a) M. Alajarín, A. López-Lµzaro, A. Pastor, P. D. Prince, J. W.
Steed, R. Arakawa, Chem. Commun. 2001, 169–170; b) M. Alajarín,
A. Pastor, R.-A. Orenes, J. W. Steed, J. Org. Chem. 2002, 67, 7091–
7095; c) M. Alajarín, A. Pastor, R.-A. Orenes, J. W. Steed, R. Ara-
kawa, Chem. Eur. J. 2004, 10, 1383–1397.
[11] M. S. Brody, C. A. Schalley, D. M. Rudkevich, J. Rebek, Jr., Angew.
Chem. 1999, 111, 1738–1742; Angew. Chem. Int. Ed. 1999, 38, 1640–
1644.
[12] J. M. Rivera, J. Rebek, Jr., J. Am. Chem. Soc. 2000, 122, 7811–7812.
[13] J. Yoon, D. J. Cram, Chem. Commun. 1997, 2065–2066.
[14] a) N. Chopra, J. C. Sherman, Angew. Chem. 1997, 109, 1828–1830;
Angew. Chem. Int. Ed. Engl. 1997, 36, 1727–1729; b) N. Chopra, C.
Naumann, J. C. Sherman, Angew. Chem. 2000, 112, 200–202;
Angew. Chem. Int. Ed. 2000, 39, 194–196.
[21] M. Valentini, H. Rüegger, P. S. Pregosin, Helv. Chim. Acta 2001, 84,
2833–2853.
[22] The reorientation of the hydrogen-bonded urea groups would be
slow on the NMR timescale; this is analogous to what is observed
for the dimers 1·1 (otherwise, averaged C3v symmetries would be ob-
served).
[23] The rotational radius (rr) is the radius of the sphere determined by
rotating a molecule around its geometric center. The rr values are
usually too big when compared with rH. See: K. W. Mattison, U.
Nobbmann, D. Dolak, Am. Biotechnol. Lab. 2001, 19, 6 6 –6 7. We
have taken the average of the two longest diagonals of the mole-
cule.
[24] These masses are not isotopic but averaged due to the low resolu-
tion of the equipment in the range of 800–4000 Da. For 2a the cal-
culated values for the averaged masses of the protonated dimer and
monomer are 1366 and 2732, respectively. For 2b the calculated
values are 1535 and 3069, respectively. Although the spectrometer
was calibrated before the measurements, there is a slight difference
(2–4 Da) between the calculated and the experimental values, which
may due to technical problems encountered with the equipment in
this mass range.
[25] The percentage of [D6]DMSO is given with respect to the overall
volume ([D6]DMSO+CDCl3), while in references [10a,c] the per-
centage of [D6]DMSO relative to the volume of CDCl3 was given.
In these experiments, [D6]DMSO was added to a solution of CDCl3
until the monomer/dimer ratio=95:5 (300 MHz) in the 1H NMR
spectra. In the mixture CDCl3/[D6]DMSO the monomer/dimer equi-
librium remains slow on the NMR timescale.
[26] a) O. Mogck, M. Pons, V. Bçhmer, W. Vogt, J. Am. Chem. Soc. 1997,
119, 5706–5712; b) M. O. Vysotsky, I. Thondorf, V. Bçhmer, Chem.
Commun. 2001, 1890–1891; c) M. O. Vysotsky, O. Mogck, Y. Rudze-
vich, A. Shivanyuk, V. Bçhmer, M. S. Brody, Y. L. Cho, D. M. Rud-
kevich, J. Rebek, Jr., J. Org. Chem. 2004, 69, 6115–6120.
[27] Each value is an average of two measurements see: D. J. Cram, H.-J.
Choi, J. A. Bryant, C. B. Knobler, J. Am. Chem. Soc. 1992, 114,
7748–7765.
[15] F. Venema, A. E. Rowan, R. J. M. Nolte, J. Am. Chem. Soc. 1996,
118, 257–258.
[16] Y. Liu, Y. Song, Y. Chen, X.-Q. Li, F. Ding, R.-Q. Zhong, Chem.
Eur. J. 2004, 10, 3685–3696.
[17] a) M. Shaul, Y. Cohen, J. Org. Chem. 1999, 64, 9358–9364; b) P.
Timmerman, J.-L. Weidmann, K. A. Jolliffe, L. J. Prins, D. N. Rein-
houdt, S. Shinkai, L. Frish, Y. Cohen, J. Chem. Soc. Perkin Trans. 2
2000, 2077–2089; c) D. G. Regan, B. E. Chapman, P. W. Kuchel,
Magn. Reson. Chem. 2002, 40, S115-S121; d) X. Ribas, J. C. Dias, J.
Morgado, K. Wurst, M. Almeida, T. Parella, J. Veciana, C. Rovira,
Angew. Chem. 2004, 116, 4141–4144; Angew. Chem. Int. Ed. 2004,
43, 4049–4052.
[18] a) E. O. Stejskal, J. E. Tanner, J. Chem. Phys. 1965, 42, 288–292;
b) P. Stilbs, Prog. Nucl. Magn. Reson. Spectrosc. 1987, 19, 1–45;
c) W. S. Price, Annu. Rep. NMR Spectrosc. 1996, 32, 51–142; d) C. S.
Johnson, Jr., Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 203–256.
[19] a) L. Frish, S. E. Matthews, V. Bçhmer, Y. Cohen, J. Chem. Soc.
Perkin Trans. 2 1999, 669–671; b) L. Frish, M. O. Vysotsky, S. E.
Matthews, V. Bçhmer, Y. Cohen, J. Chem. Soc. Perkin Trans. 2 2002,
88–93; c) L. Avram, Y. Cohen, Org. Lett. 2002, 4, 4365–4368; d) L.
Avram, Y. Cohen, J. Am. Chem. Soc. 2002, 124, 15148–15149; e) L.
Avram, Y. Cohen, Org. Lett. 2003, 5, 1099–1102; f) L. Frish, M. O.
Vysotsky, V. Bçhmer, Y. Cohen, Org. Biomol. Chem. 2003, 1, 2011–
2014.
[28] P. S. Pregosin, E. Martínez-Viviente, P. G. A. Kumar, Dalton Trans.
2003, 4007–4014.
[29] H. J. V. Tyrrell, K. R. Harris, Diffusion in Liquids, Butterworths,
London, 1984.
[30] P. A. S. Smith, G. F. Budde, S.-S. P. Chou, J. Org. Chem. 1985, 50,
2062–2066.
Received: January 10, 2005
Revised: June 1, 2005
Published online: September 27, 2005
[20] The Stokes–Einstein: D=(kBT)/(6phr). In this equation, D=diffu-
sion coefficient, kB =Boltzman constant, T=temperature in degrees
886
www.chemeurj.org
ꢀ 2006Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 877 – 886