Kinetic and thermodynamic consequences of the substitution of SMe for OMe substituents of cryptophane hosts on the binding of neutral and cationic guests

Article abstract of DOI:10.1039/b211363e

To investigate the origin of the high selectivity of cryptophane-E (1) towards Me₃NH⁺, Me₄N⁺, and CHCl₃, and particularly to discriminate the different contributions that stabilize the supramolecular complexes, we have synthesized the new cryptophane 2 bearing six MeS groups instead of MeO groups in 1. This led to a decrease of the negative charge density in the equatorial region of 2 without affecting notably the size of the molecular cavity. The binding properties of 1 and 2 towards the three guests were examined in solution and showed a slight decrease of the ΔG_a favoring the complexes of 1, accompanied by a significant modification of the δH_a vs. ΔS_a balance. The binding of the ammonium guests to 1 and 2 was strongly entropy driven, while that of CHCl₃ was purely enthalpy driven. A combination of spectroscopic and computational techniques was used to assign the main intermolecular interactions that occurred during the inclusion process. The neutral CHCl₃ molecule is more stabilized in the less negatively charged CTV cap of 1. The different behavior towards the ammonium cations can be explained in term of interactions with the electronegative heteroatoms and cation-π interactions. Moreover, this study revealed a considerable slowing down of the guest exchange kinetics with host 2, for which the association and dissociation rates are reduced by a factor 10³ to 1-⁴ with respect to 1. For example, at room temperature, the Me₄N⁺@2 complex exhibits a half-life of ca. 2 years, instead of a few hours for the corresponding complex of 1.

Full text of DOI:10.1039/b211363e

View Article Online / Journal Homepage / Table of Contents for this issue
Kinetic and thermodynamic consequences of the substitution of
SMe for OMe substituents of cryptophane hosts on the binding of
neutral and cationic guests†‡
Chantal Garcia, Delphine Humilière, Nathalie Riva, André Collet and Jean-Pierre Dutasta*
Stéréochimie et Interactions Moléculaires, École Normale Supérieure de Lyon,
UMR CNRS No5532, 46 Allée d’Italie, F-69364 Lyon 07, France. E-mail: dutasta@ens-lyon.fr
Received 19th November 2002, Accepted 1st May 2003
First published as an Advance Article on the web 12th May 2003
To investigate the origin of the high selectivity of cryptophane-E (1) towards Me3NHϩ, Me4Nϩ, and CHCl3, and
particularly to discriminate the different contributions that stabilize the supramolecular complexes, we have
synthesized the new cryptophane 2 bearing six MeS groups instead of MeO groups in 1. This led to a decrease of
the negative charge density in the equatorial region of 2 without affecting notably the size of the molecular cavity.
The binding properties of 1 and 2 towards the three guests were examined in solution and showed a slight decrease
of the ∆Ga favoring the complexes of 1, accompanied by a significant modification of the ∆Ha vs. ∆Sa balance. The
binding of the ammonium guests to 1 and 2 was strongly entropy driven, while that of CHCl3 was purely enthalpy
driven. A combination of spectroscopic and computational techniques was used to assign the main intermolecular
interactions that occurred during the inclusion process. The neutral CHCl3 molecule is more stabilized in the less
negatively charged CTV cap of 1. The different behavior towards the ammonium cations can be explained in term
of interactions with the electronegative heteroatoms and cation–π interactions. Moreover, this study revealed a
considerable slowing down of the guest exchange kinetics with host 2, for which the association and dissociation
rates are reduced by a factor 103 to 104 with respect to 1. For example, at room temperature, the Me4Nϩ@2 complex
exhibits a half-life of ca. 2 years, instead of a few hours for the corresponding complex of 1.
Introduction
The host–guest concept is by far the best approach to generate
new molecular devices and so far recognition of guests by arti-
ficial hosts has received much interest. The reversible encapsu-
lation of various substrates in molecular capsules, cavitands or
carcerands has allowed the formation of novel supramolecular
assemblies, used in the design of new materials.1–16 Associated
with this synthetic work, progress has been undertaken to
understand the recognition phenomena and to measure the
kinetic and thermodynamic parameters of the supramolecular
associations. The cryptophanes17,18 (see Chart 1) represent a
family of cyclophane hosts19 known for their ability to form
stable inclusion complexes with a variety of neutral and
cationic molecules.20–28 These studies have recently been
extended to monoatomic xenon, which proved to bind to the
smallest cryptophanes with an exceptionally large affinity in
organic solvents.29–32 These artificial host–guest (H–G) systems
represent useful tools to study the mechanisms of molecular
recognition, because they display association or dissociation
equilibria (H ϩ G
HG) that are easily observable and well
characterized by NMR techniques. In addition, representative
X-ray structures of the complexes have been solved, and sophis-
ticated simulations have been performed to reveal some of the
dynamic aspects of the molecular recognition that are not easy
to observe experimentally.33–36 The present work was under-
taken to gain some insight into the factors responsible for the
selectivity of cryptophane-E (1) towards Me3NHϩ, Me4Nϩ, and
CHCl3, for which values of association constants Ka = 1500,
225000, and 470 MϪ1 [(CDCl2)2, 300 K] had been measured.37
At the time these studies were first reported, the amazing
stability of the Me4Nϩ complex (∆Ga Ϫ30.9 kJ molϪ1) was
Chart 1
ascribed to cation–π interactions,38 the magnitude of which had
been evaluated by Schneider and coworkers at 4.2–5.4 kJ molϪ1
per interacting aromatic ring.39 This figure has recently been
challenged by Roelens and Torriti who have proposed a value
of 2 kJ molϪ1 per aromatic ring, with a saturation limit around
8 kJ mol .
Ϫ1 40 If these views are correct, then the earlier explan-
† Dedicated to the memory of Professor André Collet deceased on
October 1999.
‡ Electronic supplementary information (ESI) available: kinetics of
complexation. See http://www.rsc.org/suppdata/ob/b2/b211363e/
ation of the affinity of 1 towards Me4Nϩ should be revised.
In order to shed light on the nature of the interactions
between the cryptophane and its guests, we have made a small
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2207

View Article Online
Chart 2
structural modification of 1, and we have examined the effect
this perturbation has on the various thermodynamic and
kinetic parameters, which describe the behavior of the system.
This approach has the advantage that solvent contributions can
be minimized as solvation effects are very close for the parent
cryptophane and the modified form. In this paper, we describe
the synthesis and structural characterization of the new cryp-
tophane 2, in which six less electronegative SMe groups have
replaced the six OMe substituents of 1. Although the bulkier
SMe groups encumber the windows of the cryptophane host,
on going from OMe to SMe substituents does not modify the
size of the inner cavity of the cryptophane. This modification is
in fact designed to probe the effect of a decrease of the negative
charge density in the equatorial region of the cryptophane on
the properties of its inclusion complexes. This synthetic work is
followed by a detailed investigation of the interaction of 1 and 2
with Me3NHϩ, Me4Nϩ and CHCl3, a neutral guest isosteric
with Me3NHϩ, which shows a reversed surface polarity.
used for the preparation of the parent compound 1.41 This
method normally produces the anti D3 isomer (e.g., 1), with an
excellent selectivity with respect to the corresponding syn C3h
isomer (1Ј; see Chart 1). The key step in this synthesis is the
double trimerization of a bis-benzylalcohol such as 8, in which
the CH2OH groups precursor of the benzyl cations involved in
the reaction are situated para to the O(CH2)3O spacer. Reaction
of isothiovanillin 442 with 1,3-dibromopropane followed by
reduction of the intermediate dialdehyde 6 furnished the
desired diol 8 in 60% overall yield. The conversion of 8 to 2 was
carried out by reaction in formic acid containing a small
amount of CHCl3 to ensure dissolution of the diol, and gave 2
in 1.4% yield. This yield is much lower than that obtained in the
similar synthesis of 1 (ca. 15%).
In an attempt to improve on this, we examined the effect of
moving the CH2OH groups of the precursor to a position
para to the MeS groups. The diol 7 was thus prepared in 81%
yield from thiovanillin 3,41 using the same sequence as for the
preparation of 8. Reaction of 7 with formic acid afforded cryp-
tophane 2 in 2.1% yield. This detrimental influence of the MeS
groups on the formation of the cryptophane was not expected.
Earlier works have shown that MeS substituents do not prevent
the conversion of benzyl alcohols to cyclotriveratrylenes. Thus,
in the presence of formic acid, both the (3-MeO, 4-MeS) benzyl
alcohol 9 and its (3-MeS, 4-MeO) regioisomer 10 have been
Results
Synthesis of cryptophane 2
The cryptophane bearing MeS peripheral substituents was pre-
pared by the sequence shown in Chart 2, which is similar to that
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2208

View Article Online
converted to the corresponding C3-cyclotriveratrylene 11 in
70% and 60% yield, respectively.42 The larger steric requirement
of MeS vs. MeO is possibly responsible for the effect of making
more difficult the formation of the postulated ribbon-like
intermediate that eventually cyclizes to the cryptophane
structure.18
can be deduced from lineshape analysis of the free and bound
guest signals,46 or NOESY experiments. When the same
experiments were attempted on the CHCl3@2 system, the first
spectrum we could record ca. 7 min after dissolution of the
crystalline complex in (CDCl2)2 at 20 ЊC, showed 85% of the
host molecules containing a CHCl3 guest in their cavity (Fig. 2).
The amount of complex then slowly decreased, to reach a con-
stant value of 38% after ca. 100 min.
1
The anti structure was evidenced by H-NMR, which dis-
played a broad singlet for the inner CH2 protons of the
O(CH2)3O bridges (Fig. 1). In the D3 anti isomer 2 this CH2 lies
on a C2 axis and its two hydrogens are homotopic; the observed
broadening is due to the coupling with the adjacent CH2 groups
and probably to the conformational mobility of the spacer
bridges. In the C3h syn isomer these hydrogens are diastereo-
topic and would give two separate signals, as observed in
cryptophane-F 1Ј. A further confirmation of this assignment
was provided by HPLC analysis of the cryptophane using a
chiral stationary phase (Chiralpak-OTϩ), which showed a reso-
lution into two peaks of equal intensity, corresponding to the
enantiomers of 2 (the syn isomer is achiral).43
1
Fig. 2 500 MHz H-NMR spectra of 2 in (CDCl2)2, 20 ЊC; the slow
dissociation of the CHCl3 complex is evidenced by the spectra recorded
7 min (85% of complex) and 100 min (38%) after dissolution of a
crystalline complex of composition [2, 1.6 CHCl3].
Upon addition of Me3NHϩ picrate to a solution of empty 1
at 298 K, the complex forms immediately, and is evidenced by
the presence of a doublet at Ϫ0.36 ppm and a broad signal at
0.73 ppm representing the N(CH3)3 and NH resonances of the
incarcerated guest, respectively. The corresponding resonances
of the free guest are observed as two singlets at 2.94 ppm and
11.2 ppm. The coupling between the N(CH3)3 and NH hydro-
gens (3J = 5 Hz), which is observable in the complex, is not seen
in the free guest due to a too fast intermolecular exchange of
the NH proton. Most of the cryptophane resonances are
shifted downfield upon complexation, and the free and com-
plexed hosts are clearly visible in the spectrum. The largest shift
(ϩ0.08 ppm) is observed for the inner CH2 of the spacer
bridges; the OMe groups, the aromatic hydrogens, and the He of
the methylene bridges are also significantly shifted. The
Me3NHϩ complex of 2 formed at a much lower rate. The
1H-NMR chemical shifts of the Me3NHϩ@2 complex indicate
that its structure must be very similar to that of 1. The incarcer-
ated Me3NHϩ shows a doublet (3J = 5 Hz) at Ϫ0.33 ppm
(Me3N) and a broad signal at 0.83 ppm (NH), and most of the
cryptophane resonances are downfield shifted. The largest shift
(ca. ϩ0.1 ppm) is observed for the inner CH2 of the spacer
bridges. The binding of Me4Nϩ picrate to host 1 was still
relatively fast at room temperature. The free and bound guest
signals appear as sharp singlets at 3.35 and Ϫ0.26 ppm, respect-
ively. After ca. 30–40 min, the signal of the free guest was no
longer detectable and only the peak corresponding to the com-
plexed guest remained. This means that the association con-
stant was too large to be measured in this experiment. Thus,
competition experiments with Me3NHϩ and Me4Nϩ as guests,
were performed. The association of Me4Nϩ to host 2 was con-
siderably slower than its binding to host 1. At the beginning of
the experiment, the signals of the free guest (3.37 ppm) and of
the bound guest (Ϫ0.24 ppm) were both present. After ca. 55 h,
only the signal of the bound guest was visible, which means that
in this case also the value of Ka is too large to be measured
directly, and was determined from competition experiments.
It is thus obvious that the guest exchange rate is considerably
slower with 2 than it is with 1, a situation that has never been
Fig. 1 Part of the 500 MHz 1H-NMR spectra of 1, 2 and 1Ј in CDCl3,
showing the differences between anti and syn isomers.
Preliminary complexation studies
Crystalline complexes of 1 and 2 with CHCl3 were obtained by
evaporating to dryness solutions of the hosts in this solvent.
The G : H ratio of these solid samples was close to 2 : 1, one of
the CHCl3 molecules being incarcerated within the host cavity,
the other being located in voids of the crystal lattice.44 When
such complexes were dissolved in (CDCl2)2, the 1H-NMR
spectra recorded at room temperature exhibited two separate
signals for the free (δ 7.29 ppm) and bound CHCl3 (δ 2.86 for 1,
2.90 for 2). With host 1, the free–bound guest ratio did not
change with time and showed an immediate response to tem-
perature changes. This means that the guest exchange between
the solvent and the host cavity, which is slow on the spectro-
meter time scale,45 is nevertheless sufficiently fast to allow the
equilibrium to be reached during the experiment time. In this
case, the association constant Ka and its variations with tem-
perature can be measured in a direct way, and the exchange rate
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2209

View Article Online
Table 1 Kinetic parameters for association and dissociation of CHCl3 and ammonium picrates in the presence of hosts 1 and 2 at 293 K in (CDCl2)2
ka/MϪ1 sϪ1
kd/sϪ1
∆Ga /kJ molϪ1
∆Gd /kJ molϪ1
(t1/2)
∞
‡
‡
a
Guest
Host
CHCl3
1
2
1
2
1
2
760
1.27 (b)
55.6
77.0
61.9
86.6
67.8
85.4
71.2
90.4
79.2
103.8
98.7
116.3
0.55 s
24 min
14 s
97 h
13 h
1.16 × 10Ϫ1
76.6
4.88 × 10Ϫ4
4.81 × 10Ϫ2
1.99 × 10Ϫ6
1.49 × 10Ϫ5
1.0 × 10Ϫ8
Me3NHϩ
Me4Nϩ
2.32 × 10Ϫ3
7.2
2.88 × 10Ϫ3
792 days
a Half-life of the complex at infinite dilution. b Measured from lineshape simulation of the free and complexed guest signals.37
encountered in cryptophane complexes so far. In the following
experiments, the rate of dissociation can be measured by
recording NMR spectra as a function of time, using the kinetic
model defined below. The rate of association can similarly be
measured either by adding an excess of guest to a system
already in equilibrium, or by adding the guest to a solution of
empty host in (CDCl2)2.
Kinetic model
We assume that the behaviour of these cryptophane–guest
systems can be analysed in terms of H ϩ G
HG equilibria,
where the forward reaction (association) is second order (rate
constant ka in MϪ1 sϪ1) and the reverse reaction (dissociation) is
first order (kd in sϪ1). The ka/kd ratio represents the equilibrium
(association) constant Ka (in MϪ1). Accordingly, the rate of
formation or of dissociation of the complex is expressed by
eqn. (1).
(1)
Integration of this differential equation in the general case
led to kinetic equations for both association and dissociation
processes (see Supplementary Information).‡ A way to quantify
the kinetic stability of the complex is to determine its half-life
time (t1/2)∞ = ln2/kd, which only depends on kd, and would be
observed for a 1st order decay at infinite dilution, where the
reverse reaction is negligible. Relevant kinetic parameters for all
investigated complexes are assembled in Table 1.
Kinetic data for CHCl3, Me3NH؉ and Me4N؉ interacting with
hosts 1 and 2
Fig. 3 Association (a) and dissociation (b) kinetics for the CHCl3
complex of 2 in (CDCl2)2, 293 K. In (a) the host concentration was 2.18
× 10Ϫ3 M and 1.8 eq. of CHCl3 was added; in (b), the initial complex
concentration was 2.1 × 10Ϫ3 M, and 0.6 eq. of free CHCl3 was present.
Measurements performed in the present work on the CHCl3@1
system yielded thermodynamic and kinetic values in good
agreement with the earlier ones derived from line-shape simula-
tion of the 1H-NMR guest signals.23 At 293 K, the equilibrium
is reached within a few seconds. As said above, this is no longer
the case for the CHCl3 complex of 2. Fig. 3 shows the rates of
association and dissociation for this system at 293 K, derived
from 1H-NMR experiments. In both cases the equilibrium was
attained after ca. 100 min, allowing the determination of the
association constant Ka = 240 MϪ1 or 241 MϪ1 ( 10%), from the
constant, Ka = 1700 MϪ1, and the rate constants, ka = 9.16 MϪ1
s
Ϫ1 and kd = 5.39 × 10Ϫ3 sϪ1. From these data the association and
#
dissociation barrier were calculated to be ∆Ga# = 61.9 and ∆Gd
= 79.2 kJ molϪ1, respectively. The (t1/2)∞ at 275 K is thus 130 s.
Assuming that the barriers do not depend much on temper-
ature, the rate constants and (t1/2)∞ at 293 K can be estimated
from the Eyring equations, yielding a lifetime of 14 s for this
complex at room temperature (Table 1).
dissociation or association experiments, respectively (Ka
=
600 MϪ1 for the same complex of 1). The fit of the experimental
data to kinetic equations yielded the rate constants ka = 0.116
The corresponding Me3NHϩ complex of cryptophane 2
formed at a much slower rate (Fig. 4(b)). At 313 K, the
equilibrium is reached after ca. 15 h (60.6% of complex, Ka =
740 MϪ1). The data shown in Fig. 4(b) were fitted to give the
rate constants ka = 2.39 × 10Ϫ2 MϪ1 sϪ1 and kd = 3.25 × 10Ϫ5 sϪ1
M
Ϫ1 sϪ1 and kd = 4.88 × 10Ϫ4 sϪ1, respectively. The lifetime (t1/2
)
∞
of the CHCl3 complex of 2 at 293 K is thus of the order of
24 min, instead of 0.55 s for the same complex of 1. In terms of
∆G#, the barriers for association and for dissociation of the
CHCl3 complex are increased by approximately 20 kJ molϪ1 on
going from 1 to 2.
At 298 K, the equilibrium association constant for the
Me3NHϩ@1 complex was Ka = 1535 MϪ1 (∆Ga = Ϫ18.2 kJ
molϪ1). At 275 K, the rate of association became slow enough
to be measured (Fig. 4(a)). The equilibrium is attained after
approximately 30 min, allowing to calculate the association
(313 K). The dissociation barrier is thus ∆Gd# = 103.8 kJ molϪ1
,
i.e., 24 kJ molϪ1 higher than for the complex of 1. The estimated
(t1/2)∞ at 293 K is 97 h (Table 1).
At 280 K, the rate of association of Me4Nϩ picrate salt
to host 1 was slow enough to be measured by recording the
intensities of suitable NMR peaks as a function of time
(Fig. 5(a)). Since the association constant of Me3NHϩ to 1 and
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2210

View Article Online
Fig. 5 Kinetics of host–guest association upon addition of Me4Nϩ
Fig. 4 Kinetic of host–guest association upon addition of Me3NHϩ
(picrateϪ) to a solution of empty cryptophane 1 or 2 in (CDCl2)2, and fit
of the experimental points to eqn. (2); (a) host 1 (2.61 × 10Ϫ3 M), 0.45
eq. of guest added, 275 K; (b) host 2 (2.19 × 10Ϫ3 M), 1.5 eq. of guest
added, 313 K.
(picrateϪ) to a solution of empty cryptophane 1 or 2 in (CDCl2)2, and fit
of the experimental points to eqn. (2); (a) host 1 (2.17 × 10Ϫ3 M), 0.45
eq. of guest added, 280 K; (b) host 2 (2.83 × 10Ϫ3 M), 0.35 eq. of guest
added, 313 K.
systems exhibit negative enthalpies of association (exothermic
binding). A striking difference is observed for the corre-
sponding entropies, which are negative with CHCl3 and positive
for the ammonium guests. This means that a discussion of the
relative stabilities of the complexes in terms of their ∆Ga values
at room temperature cannot provide information on the origin
of the molecular recognition.
its variation with temperature can be easily measured, competi-
tion experiments with Me3NHϩ and Me4Nϩ as guests, were per-
formed in the range 293–323 K. The Ka value at 300 K was thus
found to be around 430000 MϪ1, a figure twice as large as that
previously reported.27 The new value is more reliable than the
earlier one, because the conditions used in the present work
allow a better precision in this determination. Using the same
kinetic treatment as above for the data of Fig. 5(a), the value of
∆Gd# for the dissociation of the Me4Nϩ complex of 1 is found to
be 98.7 kJ molϪ1 and its lifetime (t1/2)∞ at 293 K is estimated at
13 h (Table 1).
Discussion
The slowing down of the exchange rates on going from 1 to 2
has almost no impact on the equilibrium H–G affinities. This
conclusion can be drawn by observing that the association
Competition experiments were then performed between
Me3NHϩ@2 and Me4Nϩ@2 complexes to determine the value
of Ka for the Me4Nϩ@2 complex, which was found to decrease
from ca. 283000 MϪ1 at 293 K to ca. 136000 MϪ1 at 313 K. The
association of Me4Nϩ to host 2 at 313 K is shown in Fig. 5(b).
Treatment of the data then furnished a dissociation barrier of
#
#
(∆Ga ) and dissociation (∆Gd ) barriers for each guest (Table 1)
are in all cases increased by almost the same value on going
from 1 to 2 (ca. 20, 25, and 18 kJ molϪ1 for CHCl3, Me3NHϩ
and Me4Nϩ, respectively). The rationale for this effect is prob-
ably the fact that the replacement of OMe by SMe substituents
restricts the cross section of the host windows, without appre-
ciable consequences on the cavity dimensions. This was verified
by building a model of 2 from the X-ray structure of 1,23 and by
allowing the two cryptophane models to relax to their nearest
energy minimum (Tripos force field47) in the presence of the
same incarcerated ammonium guest. The only significant struc-
tural differences between 1 and 2 are the longer C–S bonds in
the latter (1.79 Å instead of 1.37 Å for C–O) and the fact that
Ar–SMe rotates more freely than does Ar–OMe. The bulkier
size of the SMe groups makes the cryptophane windows more
narrow, increasing in a similar way the association and dissoci-
ation barriers. This means that the thermodynamic parameters
assembled in Table 2 are meaningful to discuss the origin of
the host–guest affinities in these systems. The absolute values
116.3 kJ molϪ1, which corresponds to a (t1/2
) slightly greater
∞
than two years at 293 K (Table 1). This system is particularly
interesting due to its reversibility at elevated temperature, where
the complex can easily be formed, and its quasi-irreversibility at
room temperature. This circumstance may be exploited for
investigating the dynamic of the guest inside the molecular
cavity in the absence of guest exchange.
Thermodynamic data for CHCl3, Me3NH؉ and Me4N؉
interacting with hosts 1 and 2
For all systems the variation of Ka was determined in a range of
temperature sufficient to obtain significant RlnKa vs. (1/T ) plots
(Fig. 6). The values of ∆Ha and ∆Sa, together with other rele-
vant thermodynamic parameters are assembled in Table 2. All
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2211

View Article Online
a
Table 2 Thermodynamic parameters for complexation of CHCl3 and ammonium picrates by hosts 1 and 2 at 293 K in (CDCl2)2
Guest
Host
Ka/MϪ1
∆Ga/kJ molϪ1
∆Ha/kJ molϪ1
∆Sa/J molϪ1 KϪ1
CHCl3
1
2
1
2
1
2
600
270
1600
1120
475000
283000
Ϫ15.6
Ϫ13.7
Ϫ18.0
Ϫ17.1
Ϫ31.8
Ϫ30.5
Ϫ27.0
Ϫ15.1
Ϫ2.4
Ϫ12.2
Ϫ10.7
Ϫ20.2
Ϫ39
Ϫ5
ϩ53
ϩ17
ϩ72
ϩ35
Me3NHϩ
Me4Nϩ
a Errors on Ka estimated as 10%.
Chart 3 (a) Models for cryptophanes 1 and 2 used for the charge
calculations (X-ray conformation on the left); (b) CHELPG charges.
and the methylene bridges of the cyclotriveratrylene unit are
distinctly positive. In model 2, the heteroatoms are less neg-
atively charged than in model 1, Ϫ0.38 and Ϫ0.41 on the sulfur
and oxygen atoms, respectively. In contrast, the aromatic carb-
ons in model 2 are more negative than in model 1: Ϫ0.26 vs.
Ϫ0.21. These calculations suggest that the response of the H–G
complex to the structural change must depend on the type of
interactions the guest can develop with the host.
The CHCl3 complexes
We know from the crystal structure of the CHCl3@1 complex
that the guest tends to align its C3 axis with the North–South
axis of the host. This exposes the guest hydrogen to the shield-
ing effect of the three aromatic rings of one of the CTV caps,
and explains the large upfield shift of 4.43 ppm observed for
this hydrogen in the complex. The three chlorine atoms lie in the
equatorial region, and make very short contacts to the aromatic
carbons of the same CTV cap in which the C–H bond of the
guest is embedded. The closest Cl–C contacts (3.3–3.5 Å)
involve the aromatic carbons bearing the ether oxygens (the
sum of the Van der Waals radii for C–Cl is 3.45 Å). Fig. 7 is a
view of the X-ray structure of the complex, showing the way
in which the guest is stuck to one of the host CTV caps,
and evidencing the closest contacts. Thus, the less negatively
charged CTV cap in 1 favors the CHCl3@1 complex as com-
pared to the CHCl3@2 complex (Table 2). Consequently, we
observed a reduced entropy loss in the CHCl3@2 that could be
associated to a less organized assembly with 2 than with 1.
Fig. 6 Van t’Hoff plots, with black circles for host 1, open circle for
host 2; (a) CHCl3; (b) Me3NHϩ picrate; (c) Me4Nϩ picrate.
of these parameters comprise the enthalpy and entropy contri-
butions for the transfer of the guest from the solvent to the host
cavity. The magnitude of these solvophobic (or solvophilic)
contributions is difficult to appreciate, particularly for the
charged guests. By contrast, the changes of these parameters on
going from 1 to 2 are totally free from the guest–solvent inter-
actions. Under this condition it becomes possible to discuss the
origin of the host–guest interactions by examining the effects
that the structural changes of the host (i.e., the replacement of
OMe by SMe groups) have on the variation of the thermo-
dynamic parameters of the complexes.
In order to evaluate the consequences the substitution of the
SMe for the OMe groups have on the charge distribution
around the cavity, ab initio calculations were performed on the
simplified cyclotriveratrylene systems shown in Chart 3. Real-
istic models for 1 and 2 were built from the X-ray crystal struc-
ture of 1.23 In model 1, the ether oxygens are negatively charged
(Ϫ0.43 to Ϫ0.47), whereas the six aromatic carbons bear an
overall negative charge of Ϫ0.21. The two aromatic hydrogens
The Me3NH؉ and Me4N؉ complexes
Due to the lack of X-ray data for the ammonium complexes, we
performed molecular mechanic calculations to get structural
information on the ammonium@cryptophane assemblies.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2212

View Article Online
Fig. 7 Stereoview of the X-ray structure of the CHCl3@cryptophane 1 complex (ref. 23).
Fig. 8 Molecular models of the (a) Me3NHϩ@2 and (b) Me4Nϩ@2 complexes.
Starting with the known structure of the chloroform complex
should increase the rigidity and organization of the assemblies.
of 1, energy minima were determined (Tripos force field47)
for the Me3NHϩ@1, Me3NHϩ@2, Me4Nϩ@1 and Me4Nϩ@2
complexes. The minimized structures of the four complexes
showed the guest located inside the molecular cavity with one
N–H (Me3NHϩ) or one N–Me (Me4Nϩ) bond aligned with the
C3 axis of the host (Fig. 8).
This is also in agreement with the larger size of the sulfur
atoms, which congests the inner space of the host as compared
to the oxygenated one. Moreover, the highest solvation of the
charged guest compared to CHCl3, and the need of the com-
plete dissociation of the picrate salt for the encapsulation of the
positively charged guest into the cavity of the cryptophanes,
result in a more favorable entropy factor for the formation of
the ammonium complexes.48 This situation is totally different
from the process described with open-shell cyclophane recep-
tors, where the whole ion pair can be associated to the host, and
consequently is essentially enthalpic in origin.49
In the Me4Nϩ complexes the nitrogen atom is located in the
centre of the cavity, whereas in the Me3NHϩ complex it is
shifted towards a CTV unit by about 0.3 Å. In the four com-
plexes, the “axial” group aligned with the C3 axis is in close
contact with the aromatic rings of one CTV unit with distances
of the Hax or Cax atom to the centroids of the benzene rings in
the range 3.6–3.7 Å.
Conclusion
The distances to the heteroatoms are definitively larger. The
Hax atom lies 4.4–4.7 Å from the oxygen atoms in both com-
plexes and 4.95 Å from the sulfur atoms in Me3NHϩ@2. In
the Me4Nϩ complexes the average distances of the Cax atom to
the oxygen and sulfur atoms are 4.4 and 4.6 Å, respectively. For
the three groups located in the “equatorial” region, the dis-
tances of the Ceq carbons to the oxygen atoms are in the range
3.4–3.8 Å, and 3.6–4.2 Å for the distances to the sulfur atoms.
The distances to the benzene rings are in the range 3.9–4.4 Å.
Therefore, the major interactions of the three equatorial substi-
tuents mainly occur with the O or S atoms rather than with the
aromatic cavity. The axial group contribution in the stability of
the complex relies upon cation–π interactions, and can explain
the stabilizing ∆Ha values observed with the ammonium com-
plexes of cryptophane 2, which contains a more negatively
charged CTV unit, as compared to 1. However, the remarkable
stability of the ammonium-cryptophane complexes should
result from both binding modes and a simple interpretation in
terms of predominant cation–π interaction is not possible. With
the ammonium@2 complexes, the enthalpy–entropy compen-
sation seems more relevant as the stabilizing of the complexes
This work presents a detailed investigation about chloroform
and ammonium recognition by cryptophanes. Introducing
thiomethyl substituents on the cyclotriveratrylene units in the
molecule, instead of the more common methoxy group, pro-
vided the new spherical host 2 with a comparable cavity to that
of cryptophane-E 1. The structural modifications came only
from the larger MeS groups, which obstruct the portals of the
host, and we have shown that this small modification in the host
structure had significant changes in the complexation proper-
ties. This is evidenced by the higher values of the association
and dissociation energy barriers (average ∆∆G‡ = 21 kJ molϪ1
)
measured for the different complexes of 1 and 2. The half-life
time t1/2 are considerably increased for the guest@2 complexes.
For example, t1/2 for Me4Nϩ@2 is 19 × 103 h instead of 13 h for
Me4Nϩ@1 at 293 K in (CDCl2)2, acting so as a carcerand.1 The
binding of the ammonium guests to 1 and 2 was strongly
entropy driven, while that of CHCl3 was purely enthalpy
driven. As the MeO (1) or MeS (2) cryptophanes are built upon
the same aromatic architecture, the stability of the ammonium
cations complexes were attributed to interactions with the
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2213

View Article Online
heteroatoms and cation–π effects, showing that the latter is
not the unique stabilizing factor. These observations and their
interpretations clearly illustrate the many challenges of host–
guest chemistry, and justify the great interest of the cryp-
tophanes as model systems allowing performing thorough
experimental and theoretical studies for a better understanding
of the molecular recognition processes.
[HG]t = A × coth[Bt ϩ cothϪ1 (([HG]0 Ϫ D)/A)] ϩ D (2)
[HG]t = A × tanh[Bt ϩ tanhϪ1 (([HG]0 Ϫ D)/A)] ϩ D (3)
Parameters A, B and D in (2) and (3) are themselves defined
by relationships (4)–(6).
Experimental
(4)
Melting points were measured on a Perkin-Elmer DSC7 micro-
calorimeter. The 1H NMR spectra were recorded at 200 or
500 MHz on a Bruker AC200 and on Varian Unityϩ 500
spectrometers respectively. The 13C NMR spectra were recorded
at 50 MHz on the Bruker AC200 spectrometer. Elemental
analyses were performed by the Service Central d’Analyse du
C.N.R.S. Column chromatographic separations were carried
out over Merck silica gel 60 (0.040–0.063 mm); analytical and
preparative thin-layer chromatography (TLC) were performed
on Merck silica gel TLC plates F254.
B = Ϫka × A (in sϪ1
)
(5)
D = (1/2)[[H]0 ϩ [G]0 ϩ 2[HG]0 ϩ
(1/Ka)] (in mol LϪ1 or M) (6)
In practice, when Ka can be determined by direct measure-
ment of the appropriate equilibrium concentrations, then
parameters A and D can be calculated, knowing [H]0, [G]0 and
[HG]0. The variation of [HG]t as a function of t is then fitted to
eqns. (2) or (3) to give the unknown parameter B, which in turn
gives the rate constants ka and kd (since Ka is known).51 Details
for the resolution of the kinetic equation are reported in the
supplementary information.‡
Complexation studies
The rate of association or dissociation of the various guests
interacting with cryptophanes 1 and 2 as well as the corre-
sponding equilibria were investigated by variable temperature
1H-NMR at 500 MHz in 1,1,2,2-tetrachloroethane-d2. The
peak of residual C2DHCl4 set to δ = 6.00 ppm was used as
an internal standard. For CHCl3 and Me3NHϩ guests, the
appropriate solutions were kept in the spectrometer throughout
the experiment. In the case of Me4Nϩ, the solution was kept in
the spectrometer to record the beginning of the association or
dissociation process, then it was transferred into a thermostated
bath and the NMR spectra were recorded at regular intervals.
Samples of empty hosts 1 or 2 can be prepared by evaporating
to dryness, under vacuum, solutions of their CHCl3 or CH2Cl2
complexes in (CHCl2)2. The 1H-NMR spectra of the empty
cryptophanes 1 and 2 showed somewhat broadened signals,
which were due to the existence of slowly exchanging conform-
ations and to the presence in the host cavity of “invisible”
guests such as atmospheric N2 and (paramagnetic) O2
molecules.
Atomic charge distribution in cryptophanes 1 and 2
Only half of the cryptophane was used in ab initio calculations
(Gaussian 94, HF, 6.31G*, CHELPG). The inner CH2 of the
O(CH2)3O bridges being replaced by a terminal CH3 groups to
reduce the computation time while keeping the conformational
features of the molecules unaffected. In model 2, the OMe
group was replaced by a SMe one, using standard bond lengths
and angles for Ar–S–Me and a conformation similar to that of
a Ar–O–Me group (i.e., coplanar to the aromatic ring). Since
in the crystal structure the geometry of the molecule slightly
deviates from ideal C3 symmetry, the calculated charges were
averaged over the equivalent atoms or groups, to give the figures
shown in Chart 3.
Syntheses
The starting materials 3-hydroxy-4-methylthiobenzaldehyde 3
(mp 115 ЊC) and 4-hydroxy-3-methylthiobenzaldehyde 4 (mp
100 ЊC) were prepared in five steps from vanillin and isovanillin
respectively as described previously.42
Competition experiments
When the association constants were too large to be determined
directly, competition experiments were performed. Host 1
(2.18 × 10Ϫ3 M) was allowed to interact with both Me3NHϩ
(1.25 eq.) and Me4Nϩ (1.05 eq.), and NMR spectra were
recorded every 5 K in the range 293–323 K. After the equi-
librium was reached (>1 day at 20 ЊC), the four peaks corre-
sponding to the free and bound guests were visible, allowing to
calculate the ratio of the binding constants and in turn the
values of Ka for Me4Nϩ in this temperature range. Similarly,
NMR spectra of a solution of a mixture of host 2 (2.83 × 10Ϫ3
M), 1.03 eq. of Me4Nϩ and 3.46 eq. of Me3NHϩ, were recorded
in the range 293–313 K. Several weeks were necessary to
approach the equilibrium at 293K. The values of Ka were then
determined as for 1.
1,3-Bis(5-formyl-2-methylthiophenoxy)propane 5
Dibromopropane (2.4 mL; 23.6 mmol) was slowly added to
phenol 3 (8 g; 47 mmol) in acetonitrile (80 mL) in the presence
of K2CO3 (6.6 g; 47 mmol). The mixture was refluxed for 12 h.
The solvent was stripped off and the residue was taken into
water. The resulting solid was washed with aqueous KOH (10%
in weight), with water then with diethyl ether. Filtration of the
crude material over silica gel (dichloromethane) yielded 7.8 g
(88%) of 5, mp 148 ЊC (Found: C, 60.42; H, 5.23; S, 17.02.
C8H8O2S requires C, 60.61; H, 5.35; S, 17.03%); δH (CDCl3,
residual CHCl3 set to 7.24) 9.86 (s, CHO), 7.42 (dd, J 1.5 and
7.9, Ar), 7.32 (d, J 1.5, Ar), 7.16 (d, J 7.9, Ar), 4.35 (t, J 5.9,
OCH2), 2.44 (s, SCH3), 2.37 (quintet, J 5.9, CH2). δC (CDCl3,
residual CHCl3 set to 77) 191.3 (CHO), 154.9 (ArC–O), 137.9
(ArC–C), 133.6 (ArC–S), 125.2, 123.5, 108.3 (ArC–H), 65.0
(CH2O), 29.0 (CH2), 13.8 (SCH3).
Kinetic model
Integration of the differential equation (1) where [H]0, [G]0 and
[HG]0 represent the concentration of H and G and HG at time
t = 0, led to eqns. (2) and (3), which express the increase or
decrease of [HG] as a function of t, respectively.50 Eqn. (2),
which represents the association process, is valid at low initial
complex concentration, when [HG]0 is smaller than its equi-
librium concentration [HG]eq. This condition is fulfilled when
[HG]0 < Ka[H]0[G]0; conversely, eqn. (3), which represents the
dissociation, holds when [HG]0 > Ka[H]0[G]0.
1,3-Bis(4-formyl-2-methylthiophenoxy)propane 6
This compound was similarly prepared from phenol 4 (2.58 g;
15.3 mmol) and dibromopropane (0.77 mL; 7.6 mmol) in 30 mL
of acetonitrile in the presence of K2CO3 (2.12 g; 15.3 mmol).
Purification was done by digestion of the solid in hot dichloro-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2214

View Article Online
methane, giving 2.25 g (78%) of pure 6, mp 195 ЊC (Found: C,
59.26; H, 5.29; S, 16.47. C8H8O2S ϩ 0.5 H2O requires C. 59.19;
H, 5.49; S, 16.63%); δH (CDCl3, residual CHCl3 set to 7.24) 9.84
(s, CHO), 7.62 (d, J 1.9, Ar), 7.58 (dd, J 1.9 and J 8.3, Ar), 6.96
(d, J 8.3, Ar), 4.37 (t, J 5.9, OCH2), 2.45 (s, SCH3), 2.41 (quin-
tet, J 5.9, CH2); δC (CDCl3, residual CHCl3 set to 77) 190.6
(CHO), 159.7 (ArC–O), 130.4 (ArC–C), 129.9 (ArC–S), 129.7,
124.6, 110.3 (ArC–H), 65.1 (CH2O), 28.9 (CH2), 14.1 (SCH3).
References
1 D. J. Cram and J. M. Cram, Container Molecules and Their Guests,
The Royal Society of Chemistry, Cambridge, ed. J. F. Stoddart,
1994.
2 E. Maverick and D. J. Cram, in Comprehensive Supramolecular
Chemistry, ed. F. Vögtle, Pergamon, 1996, vol. 2, pp. 367–418.
3 J. Rebek Jr., Chem. Soc. Rev., 1996, 255–264.
4 D. M. Rudkevich, G. Hilmersson and J. Rebek Jr., J. Am. Chem.
Soc., 1998, 120, 12216–12225.
1,3-Bis(5-hydroxymethyl-2-methylthiophenoxy)propane 7
5 D. M. Rudkevich and J. Rebek Jr., Eur. J. Org. Chem., 1999, 1991–
2005.
Reduction of dialdehyde 5 (7.8 g; 20.7 mmol) was carried out in
200 mL of methanol by reaction with 3.9 g of NaBH4 at room
temperature overnight. The methanol was stripped off, the solid
was taken into water and collected by suction filtration. Yield
6.9 g (87%) of pure 7, mp 133 ЊC (Found: C, 60.05; H, 6.33;
S, 16.50. C19H24O4S2 requires C. 59.97; H, 6.35; S, 16.85%);
δH (CDCl3, residual CHCl3 set to 7.24) 7.08 (d, J 8.1, Ar), 6.96
(s, Ar), 6.90 (d, J 8.1, Ar), 4.58 (d, J 6, CH2OH), 4.58 (t,
J 6, OCH2), 2.39 (s, SCH3), 2.33 (quintet, J 6, CH2), 1.67 (t,
J 6, OH); δC (DMSO d6, residual DMSO set to 39.6) 153.6
(ArC–O), 135.5 (ArC–C), 126.7 (ArC–S), 123.8, 123.6, 111.2
(ArC–H), 64.9 (CH2OH), 62.7 (CH2O), 30.8 (CH2), 13.5
(SCH3).
6 A. Jasat and J. C. Sherman, Chem. Rev., 1999, 99, 931–967.
7 N. Chopra and J. C. Sherman, Angew. Chem., Int. Ed. Engl., 1999,
38, 1955–1957.
8 N. Chopra, C. Naumann and J. C. Sherman, Angew. Chem., Int. Ed.
Engl., 2000, 39, 194–196.
9 S. D. Starnes, D. M. Rudkevich and J. Rebek Jr., J. Am. Chem. Soc.,
2001, 123, 4659–4669.
10 A. M. A. van Wageningen, P. Timmerman, J. P. M. van Duynhoven,
W. Verboom, F. C. J. M. van Weggel and D. N. Reinhoudt,
Chem. Eur. J., 1997, 3, 639–654.
11 C. L. D. Gibbs, E. D. Stevens and B. C. Gibb, Chem. Commun.,
2000, 363–364.
12 C. L. D. Gibbs, H. Xi, P. A. Politzer, M. Concha and B. C. Gibb,
Tetrahedron, 2002, 673–681.
13 L. J. Prins, D. N. Reinhoudt and P. Timmerman, Angew. Chem.,
Int. Ed., 2001, 40, 2382–2426.
14 M. O. Vysotsky, I. Thondorf and V. Böhmer, Chem. Commun., 2001,
1890–1891.
1,3-Bis(4-hydroxymethyl-2-methylthiophenoxy)propane 8
The reduction of 6 (2.18 g; 5.8 mmol) was carried out in 50 mL
of methanol by reaction with 3.3 g (87 mmol) of NaBH4 at
reflux for 48 h. The methanol was stripped off, the solid was
taken into water and collected by suction filtration. Yield 1.7 g
(77%) of pure 8, mp 157 ЊC (Found: C, 59.53; H, 6.32; S, 16.75.
C19H24O4S2 requires C, 59.97; H, 6.35; S, 16.85%); δH (DMSO
d6, residual DMSO set to 2.49) 7.07 (s, Ar), 7.03 (d, J 8.1, Ar),
6.91 (d, J 8.1, Ar), 5.07 (t, J 5.6, OH), 4,41 (d, J 5.6, CH2OH),
4.19 (t, J 6, OCH2), 2.34 (s, SCH3), 2.14 (quintet, J 6, CH2);
δC (DMSO d6, residual DMSO set to 39.6) 155.5 (ArC–O),
138.8 (ArC–C), 126.6 (ArC–S), 125.7, 119.7, 110.1 (ArC–H),
65 (CH2OH), 65 (CH2O), 29.3 (CH2), 14.5 (SCH3).
15 L. J. Prins, F. De Jong, P. Timmerman and D. N. Reinhoudt, Nature,
2000, 408, 181–184.
16 M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita, T. Kusukawa and
K. Biradha, Chem. Commun., 2001, 509–518.
17 A. Collet, in Comprehensive Supramolecular Chemistry, ed. F. Toda,
Pergamon, 1996, ch. 9, Vol. 6, pp. 281–303.
18 A. Collet, in Comprehensive Supramolecular Chemistry,
ed. F. Vögtle, Pergamon, 1996, ch. 11, vol. 2, pp. 325–365.
19 F. Diederich, Cyclophanes, ed. J. F. Stoddart, Royal Society of
Chemistry, Cambridge, 1991.
20 L. Garel, J.-P. Dutasta and A. Collet, Angew. Chem., Int. Ed., 1993,
32, 1169–1171.
21 J. Canceill, L. Lacombe and A. Collet, J. Am. Chem. Soc., 1985, 107,
6993–6996.
22 J. Canceill, L. Lacombe and A. Collet, J. Am. Chem. Soc., 1986, 108,
Cryptophane 2 from 7
4230–4232.
23 J. Canceill, M. Cesario, A. Collet, J. Guilhem, L. Lacombe,
B. Lozach and C. Pascard, Angew. Chem., Int. Ed., 1989, 28, 1246–
1248.
24 L. Garel, H. Vezin, J.-P. Dutasta and A. Collet, J. Chem. Soc., Chem.
Commun., 1996, 719–720.
A mixture of diol 7 (2 g; 5.25 mmol) in CHCl3 (4 mL) and
formic acid (240 mL) was placed in a 500 mL rotatory evapor-
ator flask and heated in the water bath at 60 ЊC for 6 h, with
slow rotation (a precipitate began to form after 3 h). Evapor-
ation to dryness under vacuum gave a residue from which the
desired cryptophane 2 was isolated by column chromatography
(CH2Cl2). Yield 39 mg (2.1%) of 2 (solid, no mp) (Found: C,
57.45; H, 5.18. C57H60O6S6 ϩ 2.5 CH2Cl2 requires C, 57.36; H,
5.26%); δH (CDCl3, residual CHCl3 set to 7.24) 6.84 (s, ArH
ortho to SCH3), 6.49 (s, ArH ortho to OCH3), 4.64 (d, J 14,
CHa), 4.08 and 3.85 (two m, OCH2), 3.47 (d, J 14, CHe), 2.37 (s,
SCH3), 2.22 (m, CH2); δC (CDCl3, residual CHCl3 set to 77)
153.8 (ArC–O), 135.8, 131.8 (ArC–C), 125.4 (ArC–S), 125.2,
110.8 (ArC–H), 63.8 (CH2O), 36.4 (Ar–CH2–Ar), 30.2 (CH2),
17.8 (SCH3); m/z (FAB) 1033.3 (Mϩ. C57H60O6S6 requires
1033.49).
25 L. Garel, J.-P. Dutasta and A. Collet, New J. Chem., 1996, 20, 1265–
1271.
26 A. Collet, J.-P. Dutasta and B. Lozach, Bull. Soc. Chim. Belg., 1990,
99, 617–633.
27 L. Garel, B. Lozach, J.-P. Dutasta and A. Collet, J. Am. Chem. Soc.,
1993, 115, 11652–11653.
28 J. Lang, J. J. Dechter, M. Effemey and J. Kowalewski, J. Am. Chem.
Soc., 2001, 123, 7852–7858.
29 K. Bartik, M. Luhmer, J.-P. Dutasta, A. Collet and J. Reisse,
J. Am. Chem. Soc., 1998, 120, 784–791.
30 M. Luhmer, B. M. Goodson, Y.-Q. Song, D. D. Laws, L. Kaiser,
M. C. Cyrier and A. Pines, J. Am. Chem. Soc., 1999, 121, 3502–3512.
31 T. Brotin, A. Lesage, L. Emsley and A. Collet, J. Am. Chem. Soc.,
2000, 122, 1171–1174.
32 T. Brotin, T. Devic, A. Lesage, L. Emsley and A. Collet, Chem. Eur.
J., 2001, 7, 1561–1573.
33 P. D. Kirchhoff, M. B. Bass, B. A. Hanks, J. M. Brigss, A. Collet and
J. A. McCammon, J. Am. Chem. Soc., 1996, 118, 3237–3246.
34 P. D. Kirchhoff, J.-P. Dutasta, A. Collet and J. A. McCammon,
J. Am. Chem. Soc., 1997, 119, 8015–8022.
35 A. Varnek, G. Wipff and A. Collet, J. Comput. Chem., 1998, 19,
820–832.
Cryptophane 2 from 8
Similarly, a mixture of diol 8 (1.6 g; 4.2 mmol) in CHCl3 (8 mL)
and formic acid (200 mL) was stirred at 60 ЊC for 6 h. The same
chromatographic purification procedure as above gave 20 mg
(1.4%) of cryptophane 2 showing the same physical and
spectroscopic properties as the sample obtained from 7.
36 P. D. Kirchhoff, J.-P. Dutasta, A. Collet and J. A. McCammon,
J. Am. Chem. Soc., 1999, 121, 381–390.
37 B. Lozach, Thèse de Doctorat, Université Lyon 1, 1990. See also
references 26,27.
38 D. A. Dougherty, Science, 1996, 271, 163–168; J. C. Ma and
D. A. Dougherty, Chem. Rev., 1997, 97, 1303–1324.
39 H.-J. Schneider, T. Blatter, S. Simova and I. Theis, J. Chem. Soc.,
Chem. Commun., 1989, 580–581.
Acknowledgements
We are grateful to Professor J. Vidal and Dr D. Sakellariou for
their assistance in the development of the kinetic model used in
this work.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2215

View Article Online
40 S. Roelens and R. Torriti, J. Am. Chem. Soc., 1998, 120, 12443–
12452.
41 J. Canceill and A. Collet, J. Chem. Soc., Chem. Commun., 1988,
163–166.
48 The NMR chemical shift of the picrate anion in the complex
indicates that the anion is free in solution and does not interact
strongly with the host molecule (a high field shift of 0.07 ppm was
observed independently of the concentrations of salt and host).
Thus, an important counterion effect was excluded and only the
entropic contribution essentially due to solvation and ion pair
dissociation prevail.
49 S. Bartoli and S. Roelens, J. Am. Chem. Soc., 2002, 124, 8307–8315.
50 This differential equation is generally integrated under conditions
where it can be simplified, e.g., [H]0 = [G]0 and [HG]0 = 0. We were
unable to find any earlier example of its integration in a general case
such as this discussed here.
51 When Ka, is unknown or not measurable, a multivariable fitting of
the variation of [HG]t to eqns. (2) or (3) may in principle provide all
the unknown parameters, A, B and D, from which the rate constants
and equilibrium constant can be derived. This procedure, which
requires data of exceptional quality, was not employed in the present
work.
42 C. Garcia, C. Andraud and A. Collet, Supramol. Chem., 1992, 1,
31–45.
43 A. Tambuté, J. Canceill and A. Collet, Bull. Chem. Soc. Jpn., 1989,
62, 1390–1392.
44 This feature was observed in the X-ray crystal structure of
CHCl3@1 (reference 23).
45 The apparent slowness is in fact due to the large frequency difference
of the exchanging species resonances, ∆δ = 4.4 ppm i.e. ∆ν = 2200 Hz
at 500 MHz, which makes exchange processes having rate constants
smaller than π∆ν/√2 = 4.9 103 sϪ1 slow on the spectrometer time scale.
46 M. L. Martin, G. Martin, J.-J. Delpuech, in Practical NMR
Spectroscopy, Heyden & Son Ltd., London, 1980.
47 SYBYL 6.3 from Tripos, 1699 S. Hanley Road, St Louis,
MO 63144–2913.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 2 0 7 – 2 2 1 6
2216