Equilibrium isotope effects on noncovalent interactions in a supramolecular host-guest system

Article abstract of DOI:10.1021/ja2067324

The self-assembled supramolecular complex [Ga₄L ₆]^12- (1; L = 1,5-bis[2,3-dihydroxybenzamido]naphthalene) can act as a molecular host in aqueous solution and bind cationic guest molecules to its highly charged exterior surface or within its hydrophobic interior cavity. The distinct internal cavity of host 1 modifies the physical properties and reactivity of bound guest molecules and can be used to catalyze a variety of chemical transformations. Noncovalent host-guest interactions in large part control guest binding, molecular recognition and the chemical reactivity of bound guests. Herein we examine equilibrium isotope effects (EIEs) on both exterior and interior guest binding to host 1 and use these effects to probe the details of noncovalent host-guest interactions. For both interior and exterior binding of a benzylphosphonium guest in aqueous solution, protiated guests are found to bind more strongly to host 1 (K_H/K_D > 1) and the preferred association of protiated guests is driven by enthalpy and opposed by entropy. Deuteration of guest methyl and benzyl C-H bonds results in a larger EIE than deuteration of guest aromatic C-H bonds. The observed EIEs can be well explained by considering changes in guest vibrational force constants and zero-point energies. DFT calculations further confirm the origins of these EIEs and suggest that changes in low-frequency guest C-H/D vibrational motions (bends, wags, etc.) are primarily responsible for the observed EIEs.

Full text of DOI:10.1021/ja2067324

Article
pubs.acs.org/JACS
Equilibrium Isotope Effects on Noncovalent Interactions in a
Supramolecular Host−Guest System
Jeffrey S. Mugridge,†,‡ Robert G. Bergman,*,†,‡ and Kenneth N. Raymond*,†,‡
†Department of Chemistry, University of California, Berkeley, Berkeley, California 94720-1460, United States
‡Chemical Science Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720-1460, United States
S
* Supporting Information
ABSTRACT: The self-assembled supramolecular complex
[Ga4L6]12‑ (1; L = 1,5-bis[2,3-dihydroxybenzamido]-
naphthalene) can act as a molecular host in aqueous solution
and bind cationic guest molecules to its highly charged exterior
surface or within its hydrophobic interior cavity. The distinct
internal cavity of host 1 modifies the physical properties and
reactivity of bound guest molecules and can be used to catalyze
a variety of chemical transformations. Noncovalent host−guest
interactions in large part control guest binding, molecular
recognition and the chemical reactivity of bound guests.
Herein we examine equilibrium isotope effects (EIEs) on both exterior and interior guest binding to host 1 and use these effects
to probe the details of noncovalent host−guest interactions. For both interior and exterior binding of a benzylphosphonium
guest in aqueous solution, protiated guests are found to bind more strongly to host 1 (KH/KD > 1) and the preferred association
of protiated guests is driven by enthalpy and opposed by entropy. Deuteration of guest methyl and benzyl C−H bonds results in
a larger EIE than deuteration of guest aromatic C−H bonds. The observed EIEs can be well explained by considering changes in
guest vibrational force constants and zero-point energies. DFT calculations further confirm the origins of these EIEs and suggest
that changes in low-frequency guest C−H/D vibrational motions (bends, wags, etc.) are primarily responsible for the observed
EIEs.
examine host−guest exchange9,10 and binding11−16 processes
INTRODUCTION
Isotope effects (IEs) arise purely from differences in atomic
■
in a variety of supramolecular systems.
The self-assembled supramolecular complex [Ga4L6]12‑ (1;
Figure 1; L = 1,5-bis[2,3-dihydroxybenzamido]naphthalene)
mass.1,2 Molecules differing only in their isotopic substitution,
or isotopologues, have identical electronic structure, but often
differ in their chemical reactivity due to changes in the
vibrational frequencies and zero-point energies (ZPEs) of
isotopically substituted bonds. Since isotopic substitution can
perturb the thermodynamics and kinetics of chemical processes
without altering the electronic energy surface of the involved
molecular species, IEs provide a powerful tool to investigate the
fundamental intermolecular interactions operating during
chemical reactions. For this reason, numerous studies have
used IEs to elucidate the mechanism of chemical reactions in
both biological and chemical systems.3
Figure 1. (Left) Schematic of host 1 with only one ligand shown for
clarity. (Right) Solid-state structure of host 1 showing interior and
exterior binding of NMe3Bn+ guest molecules (orange).
Recently, many researchers have used IEs to probe
noncovalent intermolecular interactions in supramolecular
host−guest systems. These interactions, such as hydrogen
bonding, CH−π, π−π, and cation−π, are often largely
responsible for directing supramolecular self-assembly and
host−guest recognition.4−8 However, the weak, reversible and
dynamic nature of noncovalent interactions make their relative
importance and contribution to the overall free energy of
molecular recognition events difficult to dissect. Isotopic
substitution and the resulting kinetic and equilibrium IEs on
noncovalent interactions have been successfully used to
can act as a host for suitably sized cationic and neutral guest
molecules.17−19 The host ligand framework generates a large,
hydrophobic interior cavity (250−450 Å3) that can encapsulate
guest molecules with binding affinities of up to 105 M−1.20,21
Received: July 19, 2011
Published: December 6, 2011
© 2011 American Chemical Society
2057
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
Cationic molecules are also bound to the negatively charged
exterior of 1 with binding affinities of up to 102 M−1.22 We have
previously used the interior space of host 1, which differs
dramatically from the bulk solvent environment, to modify the
physical properties and reactivity of encapsulated guest
molecules and to catalyze chemical transformations with rate
accelerations of up to 106.23−27 This magnitude implies that,
like enzymes, the transition state of the reaction is stabilized.
The specific noncovalent interactions responsible for this host−
guest chemistry are difficult to study directly, however, since
there are many contributions to the overall free energy of
binding,28 including both enthalpic (heats of solvation/
desolvation, Coulombic attractions, and cation−π and CH−π
interactions) and entropic (translational entropy, entropies of
solvation/desolvation, and changes in the conformational
flexibility of both host and guest upon guest binding) terms.
Furthermore, small changes to guest size and shape have been
shown to dramatically alter guest binding21,29−31 and
reactivity,32−34 which precludes the use of structure−activity-
type studies (e.g., replacing hydrogen with fluorine atoms to
alter guest electronics) to systematically investigate host−guest
interactions. Thus, IEs provide a convenient and powerful
method to interrogate host−guest interactions in 1 because
isotopic substitution minimally changes guest structure, but
gives information about how specific C−H/D bond vibrational
frequencies change upon guest binding.
Recently we have examined kinetic isotope effects (KIEs) on
guest exchange35 and equilibrium isotope effects (EIEs) on
exterior guest binding in supramolecular host 1.36 The kinetic
studies revealed that deuterated [CpRu(η6-C6H6)]+-dn (Cp =
η5-cyclopentadienyl) guest molecules are displaced faster from
the host cavity than their protiated isotopologues. While these
KIEs are consistent with the smaller size of a C−D versus C−H
bond (and might thus be called a steric isotope effect), we
offered a more general explanation for the KIE based only on
changes in C−H/D vibrational frequencies and ZPEs, which
allows for contributions to the IE over all modes of C−H/D
vibration. The observed KIEs highlight the remarkable
sensitivity of the noncovalent interactions involved at the
transition state to guest exchange. In a separate work, we
adapted an NMR titration method originally developed by the
Perrin group37,38 for measuring IEs on acidity constants to
measure very small EIEs on the exterior binding of cationic
guests to host 1. This preliminary study showed that deuterated
benzylphosphonium guests bind more weakly to the exterior of
1 than their protiated isotopologues, however, this study was
focused primarily on the NMR titration method and its
application to our host−guest system, and the origins of the
observed EIEs were not explored. Here we report a
comprehensive study and analysis that well explains the
thermodynamic differences in isotopologue guest binding.
The EIEs on both interior and exterior binding of a number
of benzylphosphonium isotopologues to host 1 were measured;
protiated guest molecules are found to bind more strongly than
deuterated isotopologues. These EIEs are explained by changes
in guest vibrational force constants and ZPEs upon binding to
host 1. This model is further supported by DFT-level
computational studies.
Chart 1) was prepared by condensing trimethylphosphine or
trimethyl-d9-phosphine with the appropriate isotopologue of
Chart 1. Benzyltrimethyl Phosphonium Isotopologues (2-dn)
benzyl-dn bromide. This benzylphosphonium guest was chosen
for these EIE studies for several reasons: (a) it is strongly
bound to the host interior (log Kint ≈ 4.7 in D2O); (b) exterior-
interior equilibrium is quickly established (guest self-exchange
at 300 K in D2O is ∼16 s−1); (c) the guest has different
functionalities (phosphonium methyl/benzyl and aromatic
groups), which should each exhibit different noncovalent
interactions with the host (cation−π and CH−π/π−π,
respectively); and (d) a variety of isotopologues are syntheti-
cally accessible.
NMR competition titrations were carried out to measure the
relative binding affinities (Kd0/Kdn) of isotopologues 2-dn to the
interior of host 1. Rebek and co-workers have used a similar
approach to measure EIEs on guest binding in their cavitand
host systems.12 Phosphorus NMR can cleanly resolve both
interior and exterior isotopologues from one another and
titrations were carried out competing either 2-d0, 2-d7, and 2-
d9, or 2-d2, 2-d5, and 2-d9 (Figure 2); 2-d0 could not be
Figure 2. 31P{1H} NMR spectra from competition titrations used to
determine the EIE on guest encapsulation. Spectra from NMR
titrations competing 2-d0, 2-d7, and 2-d9 (top), and 2-d2, 2-d5, and 2-
d9 (bottom) are shown. Both exterior (red) and interior (blue) 2-dn
resonances are well resolved.
competed directly with 2-d5 due to peak overlap in the 31P{1H}
NMR spectrum. Using the relative concentrations of interior
and exterior 2-dn isotopologues obtained from the 31P{1H}
NMR spectra, linear plots were constructed to determine the
EIEs on interior guest binding (Figure 3); plotting [int 2-
d0][ext 2-dn] versus [int 2-dn][ext 2-d0] (where int and ext are
RESULTS AND DISCUSSION
■
Equilibrium Isotope Effects on Interior Guest Binding.
To investigate the EIE on interior guest binding, a series of
isotopologues of the guest benzyltrimethylphosphonium (2-dn,
2058
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
taking (Kd0/Kd2)*(Kd0/Kd5) = 1.07(2), which is very close to
the measured value of 1.103(7) for Kd0/Kd7.
All of the measured EIEs (Kd0/Kdn) on guest encapsulation
for 2-dn are greater than or equal to 1. This means that the
protiated isotopologues of benzyltrimethylphosphonium are
more strongly bound to the interior of host 1 than their
deuterated counterparts, with the exception of guest 2-d5,
which showed no measurable EIE.40 The EIE per deuterium
(EIE/D) values, which are the EIEs normalized for the number
of deuterium atoms in each isotopologue, show that the EIEs
are larger for the methyl/benzyl C−H/D positions than for the
aromatic C−H/D positions.
IEs are generally attributed to changes in bond vibrational
frequencies1,11,41−43 and the EIEs described above can be
explained in terms of changes in the guests’ bond vibrational
frequencies, force constants and ZPEs (Figure 4). Upon
Figure 3. Linear plots used to determine EIEs (Kd0/Kd7, top; Kd0/Kd9,
bottom) on interior guest binding of 2-dn. The plots show combined
data from multiple 31P{1H} NMR titrations competing the guests 2-d0,
2-d7 and 2-d9 ([1] = 5−15 mM; [2-dn]/[1] = 1.4−3.2) in D2O at 298
K. All linear fits are fixed through zero.
Figure 4. Diagram showing the qualitative changes in guest 2-dn C−H
and C−D ZPEs upon binding to the interior of host 1. Encapsulation
weakens the guest vibrational force constants, illustrated above by
shallower vibrational potential energy wells; this results in closer
spacing of the ZPE levels for the encapsulated 2-dn and a larger
association constant for protiated guests (Kd0/Kdn > 1).
interiorly and exteriorly bound species, respectively) gives a
straight line with slope of the EIE (Kd0/Kdn) and intercept
zero.39
Titrations competing guests 2-d0, 2-d7, and 2-d9, or 2-d2, 2-
d5, and 2-d9 were carried out over a range of guest/host ratios
([2-dn]/[1] = 1.4−3.2) and concentrations ([1] = 5−15 mM),
and the measured EIEs were found to be insensitive to both of
these parameters, as demonstrated by the excellent linearity of
the plots shown in Figure 3 (see Supporting Information for
individual titration data). Table 1 lists the EIEs on interior
encapsulation of guest 2-dn within 1, attractive interactions
between the guest and the aromatic walls of the host (cation−π,
CH−π) lower the vibrational force constants for guest C−H/D
motions. This results in lower vibrational frequencies for each
of these motions and more closely spaced C−H and C−D ZPE
levels when the guest is bound to the interior of 1, than when
the guest is free in aqueous solution. Smaller energy separations
between the C−H and C−D ZPEs in the thermodynamically
downhill encapsulated state mean that protiated guests will
have a larger association constant than deuterated guests (i.e.,
ΔEH > ΔED in Figure 4). Thus, despite the steric consequences
of encapsulation, which restrict the motional space available to
bound guest molecules,44 attractive interactions between host
and guest lead to lower C−H/D vibrational force constants, or
less restricted C−H/D vibrational motions, for encapsulated
guest molecules.
Considering the EIE/D values for interior binding of 2-dn,
the ZPE model proposed above predicts that when 2-dn is
encapsulated in 1, the methyl/benzyl C−H/D vibrational force
constants are weakened due to attractive cation−π interactions
between the positively charged phosphonium moiety and the
aromatic host walls, while the vibrational force constants for the
aromatic C−H/D bonds exhibit no measurable change. The
larger change in vibrational frequencies for the guest’s methyl/
benzyl positions suggests the importance of cation−π attractive
interactions between 2-dn and the interior of host 1. Likewise,
the lack of any EIE on guest encapsulation at the aromatic C−
Table 1. EIEs on Interior Guest Binding of 2-dn to host 1 in
D2O at 298 K
a
ratio
EIE
EIE/D
Kd0/Kd2
Kd0/Kd5
Kd0/Kd7
Kd0/Kd9
ratio
1.07(1)
1.00(2)
1.103(7)
1.14(1)
EIE
1.034(6)
1.000(3)
1.014(1)
1.015(1)
a
EIE/D
Kd0/Kd2
Kd0/Kd5
Kd0/Kd7
Kd0/Kd9
1.07(1)
1.00(2)
1.103(7)
1.14(1)
1.034(6)
1.000(3)
1.014(1)
1.015(1)
a
EIE/D = (Kd0/Kdn)1/n
.
guest binding for the isotopologues of 2-dn; because 2-d0 could
not be competed directly with 2-d2 and 2-d5, Kd0/Kd2 and Kd0/
Kd5 values were obtained by dividing Kd0/Kd9 by Kd2/Kd9 or
Kd5/Kd9 values, respectively. The validity of dividing EIE values
and the general accuracy of the data are supported by self-
consistency among independently measured EIEs: for example,
2059
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
H/D positions suggests that CH−π and π−π interactions
between the aromatic rings of 1 and 2-dn are a relatively small
contributor to attractive host−guest interactions. However,
some caution must be used here since the relative changes in
vibrational frequencies and force constants depend also on the
solvation environment and solvent−solute interactions between
2-dn and aqueous solution, and solvation of the cationic
methyl/benzyl groups is undoubtedly very different than
solvation of the hydrophobic aromatic group of the guest
molecule. This topic is further addressed by the DFT
calculations discussed near the end of this article.
guest binding make measurement of EIEs on that process
challenging, since: (a) the weak binding of guest molecules to
the host exterior suggests that the EIEs for exterior binding are
likely very small, and (b) rapid exchange between free and
exteriorly bound guest precludes the use of NMR techniques
that rely on integration, as were used to measure EIEs on
interior binding.
We have previously reported on an NMR titration method
that overcomes these difficulties and allows for the precise and
accurate measurement of very small EIEs on exterior guest
binding.36 This NMR titration methodology, originally
employed by the Perrin group in order to measure very small
IEs on acidity constants,37,38 relies only on changes in the
chemical shifts of directly competing isotopologues to measure
the relative equilibrium constants for an equilibrium that is fast
on the NMR chemical shift time scale. In our version of the
experiment, isotopologues of guest 2-dn are simultaneously
titrated into an aqueous solution of host 1 in which the interior
is blocked by the strongly binding and slowly exchanging guest
The observed EIE values of KH/KD > 1 for encapsulation of
guest 2-dn within 1 are consistent with other EIEs on
noncovalent interactions that have previously been measured
in aqueous solution. Aoyama et al. found EIEs of KH/KD
=
1.08−1.12 on the binding of small, neutral molecules to the
interior of their calixarene host molecules in water.45 Thornton
et al. reported IEs on the retention times of a series of nonpolar
small molecules for reverse-phase HPLC experiments in
methanol/water solution; they found that protiated solutes
had longer retention times than deuterated solutes and
interpreted these IEs as evidence of stronger attractive
interactions (manifested as weaker solute vibrational frequen-
cies) between the solute and the hydrophobic stationary
phase.43 The interior cavity of host 1 is very hydrophobic and
distinct from aqueous solution, and as such, guest encapsulation
might reasonably be thought of as a phase transfer, so although
the RP-HPLC and supramolecular host−guest systems are very
different, they do share some similarities and the IEs observed
in each can be explained in fundamentally the same way.
Our results and those described above for guest binding in
polar, aqueous solution stand in contrast to many examples
reported for EIEs on guest encapsulation in nonpolar, organic
solvent. For example, the Rebek group observed EIEs of KH/KD
= 0.83−0.75 on the encapsulation of p-xylene isotopologues
within their cylindrical cavitand hosts in mesitylene solution.12
These EIEs were explained by a computer model in which an
increase in C−H stretching frequencies of the xylene guest
occurred upon encapsulation (the guest and host were
simplified to a molecule of methane and benzene, respec-
tively).11 Similarly, Haino et al. reported EIEs of KH/KD < 1 for
encapsulation of neutral, aliphatic and aromatic guests in a
calixarene-based host in chloroform solution.14 The authors
here argued that an increase in the barrier to guest methyl
group internal rotations was responsible for the observed EIEs.
While it is difficult to draw very detailed conclusions from the
limited number of reported EIEs on guest encapsulation in
aqueous and organic solvent, it is clear that both specific, host−
guest noncovalent interactions as well as guest solvation play a
crucial role in determining how guest vibrational frequencies
change upon encapsulation.
+
NEt4 . As 2-dn is added to the host solution, the cationic guest
binds to the exterior of 1 and the guests’ 31P{1H} NMR
resonances shift upfield. The changes in these 31P{1H} NMR
signals are used to extract the relative exterior binding affinities
for two competing isotopologues, or the EIEs on exterior
association.
Very small EIEs are observed on the binding affinity of 2-dn
to the exterior of 1 upon deuteration of the aromatic, benzyl
and/or methyl positions of the phosphonium cation (Table 2).
Table 2. EIEs on Exterior Binding of 2-dn to Host 1 in D2O
a
at 296 K
c
ratio
EIE
EIE/D
Kd0/Kd2
Kd0/Kd5
Kd0/Kd7
Kd0/Kd9
Kd7/Kd9
1.012(3)
1.021(2)
1.0302(4)
1.005(1)
1.0038(5)
1.00426(6)
1.0051(2)
b
b
1.047(2)
1.017(3)
−
a
EIE and EIE/D values for Kd0/Kd7, Kd0/Kd9, and Kd7/Kd9 are repeated
b
from ref 36. EIEs reported as a weighted average of several titrations,
c
see Supporting Information for individual titration data. EIE/D =
(Kd0/Kdn)1/n
.
As with the EIEs on interior binding, isotopologues 2-d0, 2-d2,
and 2-d5 were not competed directly due to peak overlap in the
31P{1H} NMR spectra, and the values for Kd0/Kd2 and Kd0/Kd5
were obtained by competing 2-d2, 2-d5 with 2-d9 and then
taking the appropriate ratio with the independently measured
Kd0/Kd9. The validity of dividing these EIEs by one another
and, in general, the high precision and accuracy of these
measurements, are confirmed by the outstanding agreement
between the EIEs predicted from independent NMR titrations
and those measured directly. For example, (Kd0/Kd9)/(Kd0/Kd7)
= Kd7/Kd9 = 1.016(2), which perfectly matches the
experimentally determined EIE of Kd7/Kd9 = 1.017(3);
similarly, (Kd0/Kd2)(Kd0/Kd5) = Kd0/Kd7 = 1.033(4), which
Equilibrium Isotope Effects on Exterior Guest Bind-
ing. Cationic guest molecules can also bind to the exterior of
host 1 (Figure 1), and like interior binding, exterior association
involves a variety of noncovalent host−guest interactions.20
Exterior guest binding is generally much weaker than interior
binding, with association constants of Kext = 102 − 103 M−1 for
monocationic guest molecules.22 Exchange between free and
exteriorly bound guest molecules is rapid on the NMR time
scale. Therefore, only a single NMR resonance is observed for
unencapsulated guests, which is a weighted average of the
guest’s chemical shift in bulk solution and its chemical shift
bound to the host exterior. These characteristics of exterior
again nicely agrees with the experimental value of Kd0/Kd7
=
1.0302(4). Therefore, although many of the EIEs differ by less
than 0.01, this NMR titration method can precisely and
accurately discriminate these values from one another.
Deuteration of phosphonium cation 2-dn results in weaker
binding to the exterior of host 1 (KH/KD > 1) and the EIE/D
values show that the EIEs are largest when the benzyl/methyl
2060
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
phosphonium CH2/CH3 groups are deuterated. The EIEs on
exterior binding of 2-dn are smaller in magnitude than those
observed for interior binding, but the trends observed for both
are identical: protiated isotopologues of 2-dn are bound more
strongly to both the interior and exterior of host 1 and the EIEs
are larger for the phosphonium C−H/D bonds than for the
aromatic C−H/D bonds. This suggests that the local
noncovalent interactions operating between host and guest
on the exterior and interior of 1 are fundamentally the same,
despite evidence that the overall driving force for guest
parameters for the EIE on exterior guest binding were
determined by examining the temperature dependence of the
EIE. The exterior binding of 2-dn to 1 was chosen for this study,
rather than interior binding, since the EIE values for the former
are more precisely measured by NMR and experimentally easier
to obtain. Since the EIEs on interior and exterior host−guest
interactions follow very similar trends, as discussed above, it is
likely that the thermodynamic parameters for each of these
noncovalent interactions are also similar.
Aliquots of a D2O solution of [NEt4 ⊂ 1]11‑ were titrated into
an NMR tube containing an aqueous solution of isotopologues
2-d0, 2-d7, and 2-d9. After each addition of host, the 31P{1H}
NMR spectrum of the sample was measured at different
temperatures (294−320 K) and the phosphorus chemical shifts
of each isotopologue were recorded. Linearized plots were
constructed from the chemical shift data to determine the EIE
on exterior binding at each temperature (Table 3, see
+
association in each case is different (for NEt4 , entropy drives
encapsulation and enthalpy drives exterior ion-association).22
Although the host exterior is very hydrophilic due to its 12-
charge, it also has large aromatic surfaces that give rise to
noncovalent interactions such as cation−π, CH−π, and π−π,
between the host and exteriorly bound guest. Therefore, it is
not entirely surprising that the specific interactions between 2-
dn and the exterior or interior of host 1 would be very similar.
Given the alikeness of interior and exterior guest binding
discussed above, the EIEs on exterior binding of 2-dn to host 1
can be explained analogously to those observed for interior
binding: upon association of 2-dn to the exterior of 1, attractive
interactions (cation−π, CH−π) lower the C−H/D vibrational
force constants of the exteriorly bound cation, shrinking the
spacing between the guest’s C−H and C−D ZPE levels,
resulting in more favorable association of protiated isotopo-
logues to the host exterior (Figure 5). As with interior binding,
Table 3. EIEs on Exterior Binding of 2-dn to 1 at Different
Temperatures in D2O
T (K)
Kd0/Kd7
Kd0/Kd9
294
300
307
314
320
1.0302(4)
1.028(1)
1.025(1)
1.026(1)
1.0217(9)
1.047(1)
1.045(5)
1.041(4)
1.033(4)
1.030(4)
Supporting Information for raw NMR titration data). Van’t
Hoff analysis (Figure 6) revealed the thermodynamic
parameters (ΔΔH, ΔΔS) for the EIE on exterior binding of
2-dn to 1 (Table 4).
Figure 5. Diagram showing the qualitative changes in guest 2-dn C−H
and C−D ZPEs upon binding to the exterior of host 1. Exterior ion-
association weakens the guest vibrational force constants, illustrated
above by shallower vibrational potential energy wells, this results in
closer spacing of the ZPE levels for the exterior bound 2-dn and a
larger association constant for protiated guests (Kd0/Kdn > 1).
Figure 6. Van’t Hoff plots for EIEs on exterior association of 2-d0
■
●
versus 2-d7 (blue ) and 2-d0 versus 2-d9 (green ) to host 1 in D2O.
The errors on each EIE measurement are shown; the errors on the
Kd0/Kd9 measurements are larger due to the much broader 31P{1H}
resonance of 2-d9.
this resembles the IEs on RP-HPLC retention times discussed
above.43
Temperature Dependence of the EIE on Exterior
Binding. The above discussion for both interior and exterior
binding of 2-dn to host 1 rationalizes the observed EIEs on the
basis of changes in vibrational force constants and ZPE levels.
However, this explanation considers only enthalpic contribu-
tions to the IE and ignores the possibility that entropy could
play a role as well. The weak, reversible noncovalent
interactions between host and guest are not necessarily
dominated by enthalpy (as would typically be the case for
primary IEs, where bonds are formed or broken) and a
significant entropic contribution to the EIEs on guest binding is
not unlikely.14 To address this possibility, the thermodynamic
Table 4. Thermodynamic Parameters for the EIEs on
Association of Guests 2-dn to the Exterior of host 1 in D2O,
Obtained from a van’t Hoff Analysis
ΔΔH
ΔΔS
-TΔΔS (at 298 K,
ratio
(kcal/mol)
(cal/(mol K))
kcal/mol)
Kd0/Kd7
Kd0/Kd9
−0.05(1)
−0.13(2)
−0.12(3)
−0.33(5)
0.04(1)
0.10(2)
2061
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
The changes in the EIEs on exterior association with
temperature are small (1.030−1.022 for Kd0/Kd7; 1.047−1.030
Kd0/Kd9), but the measured EIEs are precise enough that the
trends with temperature are clear and reasonable van’t Hoff
plots can be constructed. The ΔΔH and ΔΔS values for both
Kd0/Kd7 and Kd0/Kd9 are negative, indicating that the
preferential association of the fully protiated isotopologue, 2-
d0, is favored by enthalpy, but opposed by entropy. Put another
way, association of deuterated isotopologues to the exterior of
host 1 is entropically favorable but enthalpically unfavorable.
Near room temperature, the magnitudes of the opposing
enthalpic (ΔΔH) and entropic (−TΔΔS) terms are nearly
equal, resulting in very small free energy differences (ΔΔG =
−0.01 − −0.03 kcal/mol at 298 K) for the exterior association
of 2-d0 versus 2-d7 or 2-d9.
The observation that the preferred binding of protiated
isotopologues to the exterior of host 1 (i.e., that Kd0/Kdn > 1) is
driven by enthalpic changes is consistent with the vibrational
force constant and ZPE model proposed above, for both
exterior and interior guest association. The entropic contribu-
tion to the differential binding of isotopologues, which favors
association of deuterated guests, is consistent with the changes
in vibrational energy level populations expected for association
of a protiated versus deuterated guest molecule. Assuming that
guest association results in a decrease in guest C−H/D
vibrational force constants, as proposed in the above models, a
simple Boltzmann analysis reveals that there is a larger increase
in the population of higher vibrational energy levels for low-
frequency (wags, bends, etc.) C−D versus C−H motions in the
associated guest (see Supporting Information for the details of
this analysis). This equates to a greater gain in entropy for the
deuterated isotopologue, consistent with the negative ΔΔS
value observed in experiment. There are other likely
contributors to the ΔΔS term, such as changes in the
frequencies of internal bond rotations, but a detailed and
quantitative deconstruction of all possible entropic contribu-
tions to the EIE is beyond the scope of the current work.
Instead, we believe the qualitative analysis presented above
provides a reasonable explanation for the observed thermody-
namic parameters and demonstrates that the origins of these
EIEs (the enthalpic and entropic contributions) can be
explained by changes in vibrational force constants and ZPEs.
DFT Computational Studies. To further demonstrate that
the EIEs on external or internal guest association to 1 are
caused by changes in vibrational force constants and ZPEs,
DFT-level calculations were carried out on model [2-dn−
solvent] complexes (A−D, Figure 7). Model complexes were
used here since calculations at a level of theory high enough to
give accurate vibrational frequencies are not feasible with the
full host−guest system, due to the large number of atoms
involved. Guest 2-d0 was placed near either a molecule of
naphthalene, to approximate the guest associated with the host,
or a cluster of 7 water molecules,46 to approximate the guest in
free solution. The solvent groups were arbitrarily positioned
either near the phosphonium methyl groups (geometries A and
B, for the naphthalene and water complexes, respectively), or
near the aromatic ring (geometries C and D, for the
naphthalene and water complexes, respectively) of 2-d0. The
geometries of these complexes were minimized, and the
vibrational frequencies for different isotopologues of [2-dn−
solvent] were calculated, at the B3LYP/6-311G++(d,p) level of
theory.47 Table 5 lists the calculated ZPEs and differences in
Figure 7. Optimized [2-d0−solvent] geometries, B3LYP/6-311G+
+(d,p).
Table 5. Calculated ZPEs for Isotopologues of [2-dn−
solvent] Geometries A−D
a
b
ZPE
ΔZPE
ΔΔZPE
geometry
isotopologue
(kcal/mol)
(kcal/mol)
17.896
(kcal/mol)
0.090
A
[2-d0−
238.99
naphthalene]
[2-d9−
221.09
naphthalene]
B
[2-d0−7H2O]
[2-d9−7H2O]
255.66
237.67
238.76
17.986
10.306
C
[2-d0−
0.034
naphthalene]
[2-d5−
228.45
naphthalene]
D
[2-d0−7H2O]
[2-d5−7H2O]
253.75
243.41
10.339
a
ΔZPE = ZPE(2-d0−X) − ZPE(2-dn−X); where X = naphthalene or
b
7H2O and n = 5 or 9. ΔΔZPE = ΔZPE(2-dn−7H2O) − ΔZPE(2-dn−
naphthalene).
ZPEs (ΔZPE and ΔΔZPE, respectively) for isotopologues of
the minimized [2-dn−solvent] geometries A − D.
Comparing geometries A versus B and C versus D, in both
cases the difference in ZPEs (ΔZPE) between [2-d0−solvent]
and [2-dn−solvent] isotopologues is largest for the [2-dn−
7H2O] interaction. Thinking about these ZPE changes in the
context of guest exterior or interior association, the [2-dn−
7H2O] geometries (B and D) represent the guest isotopologues
in free solution and the [2-dn−naphthalene] geometries (A and
C) represent the guest isotopologues associated with host 1.
Because the [2-dn−naphthalene] geometries are lower in
energy than the [2-dn−7H2O] geometries (both interior and
exterior guest association is thermodynamically favorable), the
calculated ΔZPE values correspond to EIEs of KH/KD > 1, or
more favorable association of protiated guest molecules. The
predicted EIEs on moving from geometry B to A (Kd0/Kd9) or
D to C (Kd0/Kd5) can be calculated from the ΔΔZPE values
(EIE = exp[-ΔΔZPE/RT]), and these values are Kd0/Kd9 = 1.16
and Kd0/Kd5 = 1.06. The magnitude and direction (Kd0/Kdn
>
1) of the computationally derived EIEs, and the larger EIE
2062
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
predicted for deuteration at the phosphonium methyl positions,
are very much consistent with the trends in EIEs on both
exterior and interior guest binding observed experimentally.
This suggests that upon association of guest 2-dn to the exterior
or interior of 1 (as approximated by interaction of the guest
with naphthalene), the guest’s vibrational force constants are
indeed lowered from those in aqueous solution (as
approximated by interaction of the guest with a cluster of
water molecules), and that the resulting ZPE differences
adequately explain the experimentally determined EIEs on
guest binding.
These calculations also provide information about which C−
H/D motions (stretching, wagging, bending, etc.) are primarily
responsible for the EIEs. To determine the contributions of
different vibrational modes to the EIE, the double difference of
the sum of vibrational frequencies (ΔΔΣυ, eq 1) for specific
C−H/D motions was calculated. To clarify the meaning of
ΔΔΣυ, note that it is directly related to ΔΔZPE by eq 2, and
constructing ΔΔΣυ for all vibrational normal modes, leads to
the ΔΔZPE values listed in Table 5. Considering only those
vibrational modes that involve motion of the methyl group H/
D atoms in [2-dn−solvent] geometries A and B,48 the ΔΔΣυ
values were calculated for both stretching and lower frequency
vibrations. The stretching vibrations are easily separated from
the other normal modes because they have much higher
frequencies, but the lower frequency vibrations (bends, wags,
scissors, etc.) are more difficult to separate from one another
because of similar frequencies and coupling between these
motions, and so are considered together. The calculated double
difference vibrational sums for stretching (ΔΔΣυ(stretch)) and
lower frequency (ΔΔΣυ(low)) C(H/D)3 motions in A and B
are: ΔΔΣυ(stretch) = −17.15 cm−1 and ΔΔΣυ(low) = 210.36
cm−1. These values indicate that the overall EIE is opposed by
changes in the phosphonium methyl group stretching
vibrations, and is due instead to lower frequency vibrations
such as wags, bends, rocks, etc. This is consistent with the
hypothesis that changes in low frequency vibrations are at least
in part responsible for the entropic component of the EIE.
is consistent with the larger EIE values observed at those
positions both experimentally (for interior and exterior guest
binding) and computationally.
Conclusion. For both interior and exterior guest binding,
EIEs were observed with supramolecular host 1. For the
benzyltrimethylphosphonium guest 2-dn, EIEs on guest interior
binding range from Kd0/Kdn = 1.00−1.14 and EIEs on guest
exterior binding range from Kd0/Kdn = 1.012−1.047 in aqueous
solution. Protiated guests are more strongly bound than their
deuterated isotopologues and the observed EIEs are largest for
the phosphonium methyl/benzyl positions. The stronger
binding of protiated guests is consistent with EIEs reported
for similar systems in aqueous solution,43,45 but the opposite of
what has been measured for a number of different host−guest
systems in organic solvent.11,12,14 Van’t Hoff analysis of the EIE
on exterior binding reveals that the preferential association of
protiated isotopologues is driven by enthalpy and opposed by
entropy.
The EIEs on guest encapsulation and exterior association are
explained by the changes in guest C−H/D vibrational force
constants and ZPEs upon binding to host 1. Association of the
cationic guest to the exterior or interior of the host introduces
attractive, noncovalent interactions, such as cation−π, CH−π,
and π−π, which weaken guest C−H/D vibrational motions.
The smaller vibrational force constants in the associated states
result in closer spacing of the C−H and C−D ZPE levels,
relative to aqueous solution, and this energy difference gives
rise to the enthalpically favored association of protiated
isotopologues. The entropic contribution to the EIE, which
favors association of deuterated isotopologues, can be
qualitatively explained using this same model, by considering
changes in vibrational energy level populations for low
frequency C−H/D motions.
DFT-level computational studies of model guest−solvent
complexes mirror the trends and magnitudes of the
experimentally determined EIEs. These calculations reveal
that the EIEs arise from closer spacing of the ZPEs in the
guest−naphthalene interaction, than in the guest-water
interaction, consistent with the vibrational force constant
model outlined above. Furthermore, the DFT studies indicate
that the EIEs are primarily the result of changes in the low
frequency vibrational motions (bends, wags, scissors, rocks,
etc.) of methyl/benzyl C−H/D bonds and that the EIEs are
opposed by changes in the stretching frequencies of those
bonds. The importance of changes in low frequency molecular
vibrations in determining the overall EIE is also consistent with
the proposed explanation for the entropic contribution to the
IE.
ΔΔ
− [
− {[
− [
∑
υ = {[
∑
9
υ(2‐d −7H O)]
0 2
∑
υ(2‐d −7H O)]}
2
∑
υ(2‐d −naphthalene)]
0
∑
υ(2‐d −naphthalene)]}
9
(1)
(2)
1
2
ΔΔZPE = hΔΔ
υ
∑
The remarkable guest stabilization, reactivity and cataly-
sis23−27 previously observed in supramolecular host 1 are
governed in large part by noncovalent host−guest interactions.
The EIEs described in this work provide a very sensitive probe
with which such nonbonding associations can be dissected and
explored in high detail. The measured EIEs reveal subtle
differences in the interactions of different guest C−H/D bonds
with the exterior and interior of host 1 and suggest the
importance of attractive cation−π interactions in guest binding.
These studies demonstrate that guest binding to the interior
and exterior of host 1 is exquisitely sensitive to even the
smallest perturbation of guest architecture and that isotopic
substitution has significant effects on the noncovalent
interactions that influence molecular recognition.
Carrying out a similar analysis, but now considering only
those vibrational modes that involve motion of the aryl H/D
atoms in [2-dn−solvent] geometries C and D, the calculated
double difference vibrational sums are: ΔΔΣυ(stretch) = −1.66
cm−1 and ΔΔΣυ(low) = −674.37 cm−1. This suggests that both
the stretching and lower frequency vibrational motions
involving the aromatic C−H/D bonds in geometries C and
D oppose the overall EIE. Therefore, the calculated EIEs
apparently arise from changes in vibrational frequencies
involving the methyl/benzyl C(H/D)3/C(H/D)2 groups,
even though these bonds are not in direct contact with the
solvent molecules in geometries C and D. The fact that the
phosphonium methyl/benzyl groups appear to be primarily
responsible for the overall EIEs, even for the 2-d5 isotopologue,
2063
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
of the host−guest complexes are reported at a guest/host ratio of ∼2/
EXPERIMENTAL SECTION
■
1.
General Information. Reagents were obtained from commercial
suppliers and used without further purification unless otherwise noted.
All solvents were sparged with nitrogen prior to use. The K12[1] host
assembly was prepared from ligand H4L (N,N′-(naphthalene-1,5-
diyl)bis(2,3-dihydroxybenzamide)) as previously described in the
literature17 and stored under nitrogen. Compounds 2-d7 and 2-d9 were
prepared as previously described.36 All deuterated precursors were
obtained from commercial suppliers with deuteration >99%.
K11[2-d0 ⊂ 1]. 1H NMR (600 MHz, D2O): δ 7.99 (d, JHH = 7.8 Hz,
12H, host ArH), 7.68 (d, JHH = 8.4 Hz, 12H, host ArH), 7.31 (d, JHH
=
7.8 Hz, 12H, host ArH), 6.99 (t, JHH = 7.8 Hz, 12H, host ArH), 6.74 (d,
JHH = 7.2 Hz, 12H, host ArH), 6.59 (t, JHH = 7.8 Hz, 12H, host ArH),
5.75 (br, 1H, encaps ArH), 5.05 (br, 2H, encaps ArH), 3.92 (br, 2H,
encaps ArH), −0.34 (m, 2H, encaps CH2), −1.26 (d, JPH = 13.2 Hz,
9H, encaps P(CH3)3). 31P{1H} NMR (243 MHz, D2O) δ 20.7 (s,
encaps P). HRMS (ESI-QTOF): calcd (found) m/z: [1 + 2-d0 + 8K+]3‑
1105.6757 (1105.6591), [1 + 2-d0 + H+ + 7K+]3‑ 1093.0238
(1092.9905), [1 + 2-d0 + 7K+]4‑ 819.5160 (819.4909), [1 + 2-d0 +
H+ + 6K+]4‑ 810.0271 (810.0029), [1 + 2-d0 + 2H+ + 5K+]4‑ 800.5381
(800.5182).
NMR Characterization. All NMR spectra were recorded using
either Bruker AV-500, AV-600 or DRX-500 spectrometers at the
1
indicated frequencies. All H NMR chemical shifts are reported in
parts per million (δ) relative to residual protic solvent resonances.
1
K11[2-d2 ⊂ 1]. 1H NMR (600 MHz, D2O): δ 7.96 (d, JHH = 7.8 Hz,
Multiplicities of H NMR resonances are reported as s = singlet, d =
doublet, t = triplet, m = multiplet and br = broad. For the NMR
chemical shift data of host−guest complexes, host denotes signals
corresponding to assembly 1 and encaps denotes signals corresponding
to encapsulated guest; only encapsulated guest signals are tabulated.
All 13C{1H} NMR spectra of host−guest assemblies were recorded
using an HSQC experiment; 1D 13C{1H} NMR lacks the sensitivity
necessary to obtain adequate spectra of these host−guest complexes.
Only 13C signals for carbon atoms attached directly to a hydrogen
atom are observed in the HSQC experiment. As such, 13C signals for
deuterated guest carbon atoms are not reported for any host−guest
complexes. All 31P{1H} NMR chemical shifts are referenced to an
internal standard of triethylphosphate.
Mass Spectrometry Characterization. All mass spectra were
recorded at the UC Berkeley Mass Spectrometry facility. Mass spectra
of all host−guest assemblies were acquired on a Waters QTOF API
mass spectrometer in methanol and all other mass spectra were
acquired on a Thermo Scientific LTQ−Orbitrap XL mass
spectrometer.
Computational Methods. All DFT calculations were carried out
in the UC Berkeley Molecular Graphics and Computation Facility
using Gaussian 09 software with GaussView graphical user interface.49
Benzyl-d2 Trimethyl Phosphonium Bromide (2-d2[Br]).
Benzyl-d2 bromide (0.30 mL, 1.74 mmol) was dissolved in 80 mL
diethyl ether in a warm, oven-dried 250 mL of Schlenk flask. The
solution was sparged with N2 for 10 min and trimethylphosphine (0.54
mL, 5.2 mmol) was added via syringe. The solution was stirred
overnight under nitrogen atmosphere and the resulting white
precipitate was collected by vacuum filtration and washed with diethyl
ether (3 × 30 mL). After removing residual solvent overnight under
high vacuum, the product was obtained as a white solid with yield 581
mg (92%). 1H NMR (600 MHz, D2O): δ 7.47 (br, m, 3H, 3 × ArH),
7.34 (br, d, 2H, 2 × ArH), 1.83 (d, J = 14.1 Hz, 9H, PMe3). 2H NMR
(92 MHz, D2O): δ 3.7 (br, CD2). 13C{1H} NMR (151 MHz, D2O): δ
130.0 (d, J = 4.8 Hz, ArC), 129.6 (d, J = 3.0 Hz, ArC), 128.5 (d, J = 3.6
Hz, ArC), 128.3 (d, J = 8.9 Hz, ArC), 29.8 (m, CD2), 7.1 (d, J = 55.6
Hz, PMe3). 31P{1H} NMR (243 MHz, D2O): δ 25.6 (s). MS (ESIHR)
for C10H14D2P, calcd (found) m/z: 169.1110 (169.1114).
Benzyl-d5 Trimethyl Phosphonium Bromide (2-d5[Br]). The
title compound was prepared analogously to 2-d2[Br] from benzyl-d5
bromide (0.35 mL, 2.97 mmol) and trimethylphosphine (0.92 mL, 8.9
mmol). The product was obtained as a white solid with yield 743 mg
(99%). 1H NMR (600 MHz, D2O): δ 3.69 (d, J = 15.9 Hz, 2H, CH2),
1.82 (d, J = 14.3 Hz, 9H, PMe3). 2H NMR (92 MHz, D2O): δ 7.5 (br,
ArD), 7.4 (br, ArD). 13C{1H} NMR (151 MHz, D2O): δ 129.6 (t, J =
24 Hz, ArCD), 129.0 (t, J = 25 Hz, ArCD), 128.1 (d, J = 8.9 Hz,
ArCD), 127.9 (d, J = 25 Hz, ArC), 30.1 (d, J = 50.5 Hz, CH2), 6.9 (d, J
= 55.3 Hz, PMe3). 31P{1H} NMR (243 MHz, D2O): δ 25.8 (s). MS
(ESIHR) for C10H11D5P, calcd (found) m/z: 172.1298 (172.1302).
General Preparation of Host−Guest Complexes K11[2-dn ⊂
1]. All host−guest complexes with 2-dn were prepared in situ, in a
nitrogen-filled glovebox, using degassed solvent. D2O stock solutions
of 2-dn[Br] (∼50 mM) and host K12[1] (∼20 mM) were combined in
the appropriate ratios (see below for titration procedure and
Supporting Information for individual titration data); guest
encapsulation is immediate and quantitative. Characterization details
12H, host ArH), 7.61 (d, JHH = 8.4 Hz, 12H, host ArH), 7.30 (d, JHH
=
8.4 Hz, 12H, host ArH), 6.96 (m, overlapping with exterior 2-d2, host
ArH), 6.73 (d, JHH = 7.2 Hz, 12H, host ArH), 6.58 (t, JHH = 7.8 Hz,
12H, host ArH), 5.67 (br, 1H, encaps ArH), 4.96 (br, 2H, encaps ArH),
3.87 (br, 2H, encaps ArH), −1.36 (d, JPH = 12.6 Hz, 9H, encaps
P(CH3)3). 31P{1H} NMR (243 MHz, D2O) δ 20.2 (s, encaps P).
HRMS (ESI-QTOF): calcd (found) m/z: [1 + 2-d2 + 8K+]3‑
1106.3466 (1106.3597), [1 + 2-d2 + H+ + 7K+]3‑ 1093.6946
(1093.6876), [1 + 2-d2 + 2H+ + 6K+]3‑ 1080.7094 (1080.7013), [1
+ 2-d2 + 7K+]4‑ 820.0192 (820.0231), [1 + 2-d2 + H+ + 6K+]4‑
810.5302 (810.5326), [1 + 2-d2 + H+ + 5K+]5‑ 640.4316 (640.4299),
[1 + 2-d2 + 2H+ + 4K+]5‑ 632.8404 (632.8329).
K11[2-d5 ⊂ 1]. 1H NMR (600 MHz, D2O): δ 7.96 (d, JHH = 7.8 Hz,
12H, host ArH), 7.61 (d, JHH = 9.0 Hz, 12H, host ArH), 7.30 (d, JHH
=
8.4 Hz, 12H, host ArH), 6.95 (t, JHH = 6.7 Hz, 12H, host ArH), 6.58 (t,
JHH = 7.8 Hz, 12H, host ArH), −0.45 (m, 2H, encaps CH2), −1.36 (d,
JPH = 13.2 Hz, 9H, encaps P(CH3)3). 31P{1H} NMR (243 MHz, D2O)
δ 20.3 (s, encaps P). HRMS (ESI-QTOF): calcd (found) m/z: [1 + 2-
d5 + 8K+]3‑ 1107.3528 (1107.3517), [1 + 2-d5 + H+ + 7K+]3‑
1094.7009 (1094.6860), [1 + 2-d5 + H+ + 6K+]4‑ 811.2849
(811.2872), [1 + 2-d5 + 2H+ + 5K+]4‑ 801.7960 (801.7955), [1 + 2-
d5 + H+ + 5K+]5‑ 641.2354 (641.2313), [1 + 2-d5 + 2H+ + 4K+]5‑
633.4442 (633.4349).
K11[2-d7 ⊂ 1]. 1H NMR (600 MHz, D2O): δ 7.97 (d, JHH = 7.8 Hz,
12H, host ArH), 7.64 (d, JHH = 8.4 Hz, 12H, host ArH), 7.30 (d, JHH
=
8.4 Hz, 12H, host ArH), 6.96 (t, JHH = 8.4 Hz, 12H, host ArH), 6.74 (d,
JHH = 7.8 Hz, 12H, host ArH), 6.58 (t, JHH = 7.8 Hz, 12H, host ArH),
−1.33 (d, JPH = 13.2 Hz, 9H, encaps P(CH3)3). 31P{1H} NMR (243
MHz, D2O) δ 20.3 (s, encaps P). HRMS (ESI-QTOF): calcd (found)
m/z: [1 + 2-d7 + 8K+]3‑ 1108.0237 (1108.0255), [1 + 2-d7 + H+ +
7K+]3‑ 1095.3717 (1095.3804), [1 + 2-d7 + 7K+]4‑ 821.2770
(821.2778), [1 + 2-d7 + H+ + 6K+]4‑ 811.5381 (811.5389), [1 + 2-
d7 + H+ + 5K+]5‑ 641.4379 (641.4271), [1 + 2-d7 + 2H+ + 4K+]5‑
633.8467 (633.8426).
K11[2-d9 ⊂ 1]. 1H NMR (600 MHz, D2O): δ 7.95 (d, JHH = 7.8 Hz,
12H, host ArH), 7.61 (d, JHH = 8.4 Hz, 12H, host ArH), 7.30 (d, JHH
=
7.2 Hz, 12H, host ArH), 6.98 (m, overlapping with exterior 2-d9, encaps
ArH), 6.73 (d, JHH = 7.2 Hz, 12H, host ArH), 6.58 (t, JHH = 7.8 Hz,
12H, host ArH), 5.66 (br, 1H, encaps ArH), 4.95 (br, 2H, encaps ArH),
3.87 (br, 2H, encaps ArH), −0.46 (m, 2H, encaps CH2). 31P{1H} NMR
(243 MHz, D2O) δ 19.4 (s, encaps P). HRMS (ESI-QTOF): calcd
(found) m/z: [1 + 2-d9 + 8K+]3‑ 1108.6945 (1108.6721), [1 + 2-d9 +
H+ + 7K+]3‑ 1096.0426 (1096.0227), [1 + 2-d9 + 2H+ + 6K+]3‑
1083.3907 (1083.3489), [1 + 2-d9 + 7K+]4‑ 821.7802 (821.7579), [1 +
2-d9 + H+ + 6K+]4‑ 812.2912 (812.2674), [1 + 2-d9 + 2H+ + 5K+]4‑
802.8023 (802.7799).
General Procedure for NMR Titrations to Determine EIEs on
Interior Guest Binding. In a typical experiment, D2O stock solutions
of 1 (∼20 mM), 2-d0, 2-d7, and 2-d9 (∼50 mM each) were each
prepared from a D2O solution containing 1,4-dioxane (4.7 mM) and
KPF6 (100 mM) as internal standards. The stock solutions were
measured by 1H NMR to obtain accurate concentrations of each
species and were combined in the desired ratios (see Supporting
Information for individual titration conditions and data) to give a 1−2
mL solution containing 1 (∼16 mM), 2-d0, 2-d7, and 2-d9 and the
2064
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
host−guest stock solution was filtered through a 0.2 μm syringe filter
to remove any undissolved particulates. Different volumes (500, 400,
300, and 200 μL) of the host−guest stock solution were then added to
each of four NMR tubes and diluted with D2O (4.7 mM dioxane, 100
mM KPF6) to a total volume of 500 μL in each tube. The set of NMR
tubes (now each with different absolute concentrations, ranging from
16−6 mM) were allowed to equilibrate overnight. The concentrations
of external and internal 2-dn isotopologues were measured by 31P{1H}
NMR on a Bruker AV-600 spectrometer, relative to the KPF6 internal
standard, at 298 K using an inverse-gated decoupling sequence
(zgig30) and a delay time of 7 s (the T1 relaxation times for interior
and exterior 31P guest signals was ∼2.5 s; the 7 s delay should allow
sufficient time for relaxation to obtain accurate integrals). Generally,
∼2 h of data collection were required to obtain adequate signal-to-
noise for each 31P{1H} NMR spectrum. This procedure was repeated
for different 1:2-dn ratios and the all data over a range of
concentrations and host:guest ratios was combined to determine the
EIEs on interior binding.
ACKNOWLEDGMENTS
■
The authors would like to thank Dr. Xinzheng Yang, Dr. Jamin
Krinsky and Dr. Kathleen Durkin for assistance with DFT
computational studies and acknowledge NSF Grants CHE-
0233882 and CHE-0840505, which fund the UC Berkeley
Molecular Graphics and Computational Facility. We also thank
Dr. Ulla Andersen for help with mass spectrometry experiments
and Prof. Charles Perrin for correspondence about his NMR
titration methods. This work has been supported by the
Director, Office of Science, Office of Basic Energy Sciences, and
the Division of Chemical Sciences, Geosciences, and
Biosciences of the U.S. Department of Energy at LBNL
under Contract No. DE-AC02-05CH11231 and an NSF
predoctoral fellowship to J.S.M.
REFERENCES
■
General Procedure for NMR Titrations to Determine EIEs on
Exterior Guest Binding. In a typical experiment, D2O solutions of 2-
d0 and 2-d7 (each ∼50 mM) were combined in an NMR tube with a
small amount (∼1 μL) of trimethyl phosphate as an internal standard
and the volume of the tube was adjusted as necessary to ∼500 μL with
D2O. The starting concentration of 2-dn was generally ∼20 mM. The
ratio of isotopologues (generally [d0]/[d7] ≈ 0.3) was empirically
chosen so that 31P{1H} peak heights remained similar to one another
throughout the titration. The 31P{1H} NMR spectrum of this mixture
was then measured to provide a starting chemical shift for each
isotopologue. Aliquots (ranging from 5 μL − 200 μL in volume) of a
D2O solution of 1 (∼40 mM) with excess (2−5 equiv, relative to 1)
NEt4Cl were then added to the NMR tube containing the
isotopologue via syringe. After each addition the NMR tube was
inverted 5−10 times to ensure adequate mixing and the 31P{1H} NMR
chemical shifts of each isotopologue were measured. All chemical shifts
were measured relative to that of the trimethyl phosphate internal
standard. The titration is finished when addition of the 1/NEt4Cl
solution causes no further upfield shifts in the 2-dn 31P{1H}
resonances. All 31P{1H} NMR spectra were acquired on a Bruker
AV-600 spectrometer with a delay time of 1.5 s, an acquisition time of
1.3 s and a digital resolution of 0.38 Hz/pt. For variable temperature
measurements, 31P{1H} NMR spectra were collected at several
different temperatures (294−320 K) after each addition of 1; the
sample was allowed to equilibrate for ∼10 min at each temperature in
the NMR probe. Four scans were generally enough to obtain adequate
signal-to-noise, but if more scans were needed a delay of at least 1 min
was incorporated between each set of four scans to avoid any sample
(1) Kohen, A.; Limbach, H. Isotope Effects in Chemistry and Biology;
Taylor & Francis Group: London, 2006.
(2) Anslyn, E. V.; Dougherty, D., A. Modern Physical Organic
Chemistry; University Science Books: Mill Valley, CA, 2006.
(3) For recent examples see: (a) Giagou, T.; Meyer, M. Chem.Eur.
J. 2010, 16, 10616−10628. (b) Zhang, H.; Wang, S.; Sun, Q.; Smith, S.
C. Phys. Chem. Chem. Phys. 2009, 11, 8422−8424. (c) Stojkovic,
Kohen, A. Isr. J. Chem. 2009, 49, 163−173. (d) de La Harpe, K.;
Crespo-Hernandez, C. E.; Kohler, B. J. Am. Chem. Soc. 2009, 131,
́
V.;
́
17557−17559. (e) Abe, T.; Miyazawa, A.; Konno, H.; Kawanishi, Y.
Chem. Phys. Lett. 2010, 491, 199−202. (f) Gonzalez-James, O. M.;
Singleton, D. A. J. Am. Chem. Soc. 2010, 132, 6896−6897. (g) Liu, M.;
Girma, E.; Glicksman, M. A.; Stein, R. L. Biochemistry 2010, 49, 4921−
4929. (h) Wiley, K. L.; Tormos, J. R.; Quinn, D. M. Chem. Biol.
Interact. 2010, 187, 124−127. For a review of IEs on noncovalent
interactions, see: Wade, D. Chem. Biol. Interact. 1999, 117, 191−217.
(4) Yoshizawa, M.; Klosterman, J. K.; Fujita, M. Angew. Chem., Int.
Ed. 2009, 48, 3418−3438.
(5) Saalfrank, R. W.; Maid, H.; Scheurer, A. Angew. Chem., Int. Ed.
2008, 47, 8794−8824.
(6) Oshovsky, G. V.; Reinhoudt, D. N.; Verboom, W. ChemInform
2007, 38, 2366−2393.
(7) Lehn, J.-M. Supramolecular Chemistry: Concepts and Perspectives;
Wiley-VCH: New York, 1995.
(8) Stang, P. J.; Olenyuk, B. Acc. Chem. Res. 1997, 30, 502−518.
(9) Felder, T.; Schalley, C. A. Angew. Chem., Int. Ed. 2003, 42, 2258−
2260.
1
heating from the H decoupling. Spectra were phased and calibrated
(10) Liu, Y.; Warmuth, R. Org. Lett. 2007, 9, 2883−2886.
(11) Zhao, Y.-L.; Houk, K. N.; Rechavi, D.; Scarso, A.; Rebek, J. J.
Am. Chem. Soc. 2004, 126, 11428−11429.
(phosphate standard set to 0 ppm) manually, and the isotopologue
chemical shifts were extracted by manually measuring peak centers.
(12) Rechavi, D.; Scarso, A.; Rebek, J. J. Am. Chem. Soc. 2004, 126,
7738−7739.
ASSOCIATED CONTENT
■
(13) Tran, C. D.; Mejac, I.; Rebek, J.; Hooley, R. J. Anal. Chem. 2009,
81, 1244−1254.
S
* Supporting Information
Data for NMR titrations used to determine EIE on interior
guest binding, representative 1H NMR spectrum from
unsuccessful interior EIE titrations, representative 31P{1H}
NMR spectra from exterior EIE titrations, data for NMR
titrations used to determine EIE on exterior guest binding,
error analysis used for exterior and interior EIE determination,
vibrational energy level population calculation, atomic coor-
dinates for DFT calculations, details of M06-2X DFT
calculations, full citation for ref 47, and a list of references for
this material. This material is available free of charge via the
Internet at http://pubs.acs.org.
(14) Haino, T.; Fukuta, K.; Iwamoto, H.; Iwata, S. Chem.Eur. J.
2009, 15, 13286−13290.
(15) Laughrey, Z. R.; Upton, T. G.; Gibb, B. C. Chem. Commun.
2006, 970−972.
(16) Liu, Y.; Warmuth, R. Angew. Chem., Int. Ed. 2005, 44, 7107−
7110.
(17) Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N.
Angew. Chem., Int. Ed. 1998, 37, 1840−1843.
(18) Parac, T. N.; Caulder, D. L.; Raymond, K. N. J. Am. Chem. Soc.
1998, 120, 8003−8004.
(19) Biros, S. M.; Bergman, R. G.; Raymond, K. N. J. Am. Chem. Soc.
2007, 129, 12094−12095.
(20) Pluth, M. D.; Johnson, D. W.; Szigethy, G.; Davis, A. V.; Teat, S.
J.; Oliver, A. G.; Bergman, R. G.; Raymond, K. N. Inorg. Chem. 2009,
48, 111−120.
AUTHOR INFORMATION
■
Corresponding Author
(21) Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.;
Raymond, K. N. J. Am. Chem. Soc. 2006, 128, 1324−1333.
rbergman@berkeley.edu; raymond@socrates.berkeley.edu
2065
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066

Journal of the American Chemical Society
Article
(22) Sgarlata, C.; Mugridge, J. S.; Pluth, M. D.; Tiedemann, B. E. F.;
Zito, V.; Arena, G.; Raymond, K. N. J. Am. Chem. Soc. 2010, 132,
1005−1009.
(44) Mugridge, J. S.; Szigethy, G.; Bergman, R. G.; Raymond, K. N. J.
Am. Chem. Soc. 2010, 132, 16256−16264.
(45) Fujimoto, T.; Yanagihara, R.; Kobayashi, K.; Aoyama, Y. Bull.
Chem. Soc. Jpn. 1995, 68, 2113−2124.
(23) Wang, Z. J.; Brown, C. J.; Bergman, R. G.; Raymond, K. N.;
Toste, F. D. J. Am. Chem. Soc. 2011, 133, 7358−7360.
(24) Hastings, C. J.; Pluth, M. D.; Bergman, R. G.; Raymond, K. N. J.
Am. Chem. Soc. 2010, 132, 6938−6940.
(46) The choice of 7 water molecules for these calculations was based
on the qualitative observation that this seemed to be the minimum but
sufficient number to surround either the phosphonium or aromatic
moieties and provide interactions between those groups’ C-H/D
bonds with nearby water molecules. Calculations with a shell of water
surrounding the entire phosphonium guest were also explored, but
with so many water molecules (∼50) and hence degrees of freedom
the minimizations could not be made to converge. Similar calculations
using only one molecule of water or one molecule benzene were also
explored, and these reproduce the trends for the calculations reported
in the manuscript.
(25) Brown, C. J.; Bergman, R. G.; Raymond, K. N. J. Am. Chem. Soc.
2009, 131, 17530−17531.
(26) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Acc. Chem. Res.
2009, 42, 1650−1659.
(27) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. Acc.
Chem. Res. 2005, 38, 349−358.
(28) Leung, D. H.; Bergman, R. G.; Raymond, K. N. J. Am. Chem.
Soc. 2008, 130, 2798−2805.
(47) Geometries A and B were also minimized and vibrational
frequencies calculated using M06−2X/6-311G++(d,p); the M06−2X
functional has been shown to more accurately calculate energies
involving significant dispersion forces: J. Comput. Chem.2009, 30, 51
. The M06−2X calculations are qualitatively consistent with the
B3LYP calculations reported above, with the M06−2X functional
giving an EIE of Kd0/Kd9 = 1.30 and ΔΔΣυ(stretch) = 1.19 cm−1 and
ΔΔΣυ(low) = 1552 cm−1 for model systems A versus B. Like the
B3LYP calculations, these closely reproduce the trends in EIEs
observed experimentally and suggest that the EIEs result from changes
in low frequency vibrational motions. Details of the M06−2X
calculations are shown in the Supporting Information.
(48) The normal modes primarily involving vibrational motion with
specific atoms were automatically selected by Gaussian09 (Gaussview)
software; see ref 49.
(49) Frisch, M. J.; et al.et al. Gaussian 09, Revision A.02, Gaussian,
Inc.: Wallingford CT, 2009.
(29) Hastings, C. J.; Pluth, M. D.; Biros, S. M.; Bergman, R. G.;
Raymond, K. N. Tetrahedron 2008, 64, 8362−8367.
(30) Parac, T. N.; Scherer, M.; Raymond, K. N. Angew. Chem., Int. Ed.
2000, 39, 1239−1242.
(31) Pluth, M. D.; Fiedler, D.; Mugridge, J. S.; Bergman, R. G.;
Raymond, K. N. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10438−10443.
(32) Leung, D. H.; Fiedler, D.; Bergman, R. G.; Raymond, K. N.
Angew. Chem., Int. Ed. 2004, 43, 963−966.
(33) Leung, D. H.; Bergman, R. G.; Raymond, K. N. J. Am. Chem.
Soc. 2006, 128, 9781−9797.
(34) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Angew. Chem.,
Int. Ed. 2007, 46, 8587−8589.
(35) Mugridge, J. S.; Bergman, R. G.; Raymond, K. N. Angew. Chem.,
Int. Ed. 2010, 49, 3635−3637.
(36) Mugridge, J. S.; Bergman, R. G.; Raymond, K. N. J. Am. Chem.
Soc. 2010, 132, 1182−1183.
(37) Perrin, C. L.; Fabian, M. A. J. Am. Chem. Soc. 1996, 68, 2127−
2134.
(38) Perrin, C. L.; Karri, P. J. Am. Chem. Soc. 2010, 132, 12145−
12149.
(39) The EIEs on interior binding of 2-dn to 1 were also measured
1
using analogous H NMR titrations. The EIEs determined this way
are: Kd0/Kd7 = 1.26, Kd0/Kd9 = 1.03 and Kd7/Kd9 = 0.97. These values
are qualitatively consistent with those determined via 31P NMR
titrations (i.e. Kd0/Kdn > 1), but are not internally consistent: (Kd0/
Kd9)/(Kd0/Kd7) = 0.82, significantly different than the measured value
1
of 0.97. This apparent inaccuracy in the H NMR titration method
1
may be due to small impurities in the H NMR, which could not be
removed (see Supporting Information, Figure S25). The 31P{1H}
NMR resonances for the isotopologues are cleanly resolved and
provide a more accurate measure of the concentrations of
isotopologues on the exterior and interior of 1. As such, only data
from the 31P{1H} NMR titrations are considered.
(40) In a previous study on kinetic isotope effects in host 1,35 we
measured an EIE on interior guest binding of Kd0/Kd6 = 0.96(1) for
the guest [CpRu(C6H6)]+-dn, suggesting that deuteration of the
aromatic C-H bonds on the benzene ring resulted in stronger binding
to the interior of host 1. The measured value of Kd0/Kd5 = 1.00(2) for
interior binding of the guest 2-dn reported in the current study does
not conclusively agree or disagree with the previous observation that
deuteration of guest aromatic C−H bonds results in stronger binding
to the interior of host 1. We can only say that the EIE observed upon
deuteration of the aromatic ring is small, consistent with the results
reported in ref 35. Comparison of the EIEs reported here for guest 2-
d9 versus 2-d5 also demonstrates that it is not surprising that such
variation in the EIE on interior binding is observed for different guests
or different guest moieties.
(41) Saunders, M.; Wolfsberg, M.; Anet, F. A. L.; Kronja, O. J. Am.
Chem. Soc. 2007, 129, 10276−10281.
(42) Hayama, T.; Baldridge, K. K.; Wu, Y.-T.; Linden, A.; Siegel, J. S.
J. Am. Chem. Soc. 2008, 130, 1583−1591.
(43) Turowski, M.; Yamakawa, N.; Meller, J.; Kimata, K.; Ikegami, T.;
Hosoya, K.; Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 2003, 125,
13836−13849.
2066
dx.doi.org/10.1021/ja2067324 | J. Am. Chem.Soc. 2012, 134, 2057−2066