Chain-reaction anion exchange between metal-organic cages

Article abstract of DOI:10.1021/ja311882h

Differential binding affinities for a set of anions were observed between larger (1) and smaller (2) tetrahedral metal-organic capsules in solution. A chemical network could thus be designed wherein the addition of hexafluorophosphate could cause perchlorate to shift from capsule 2 to capsule 1 and triflimide to be ejected from capsule 1 into solution.

Full text of DOI:10.1021/ja311882h

Article
pubs.acs.org/JACS
Chain-Reaction Anion Exchange between Metal−Organic Cages
Shucong Ma, Maarten M. J. Smulders,† Yana R. Hristova, Jack K. Clegg,‡ Tanya K. Ronson,
Salvatore Zarra, and Jonathan R. Nitschke*
Department of Chemistry, The University of Cambridge, Lensfield Road, Cambridge CB2 1EW, United Kingdom
S
* Supporting Information
ABSTRACT: Differential binding affinities for a set of anions were observed between larger
(1) and smaller (2) tetrahedral metal−organic capsules in solution. A chemical network
could thus be designed wherein the addition of hexafluorophosphate could cause perchlorate
to shift from capsule 2 to capsule 1 and triflimide to be ejected from capsule 1 into solution.
salt (4 equiv) led to the formation of the dark-green [Fe4L6]8+
INTRODUCTION
■
cage complex 1 (Scheme 1, top). The C2-symmetric bis-
In recent years, metal−organic capsules have attracted
bidentate pyridylimine ligands form the edges of the
continued attention due to their wide-ranging applications as
tetrahedron, bridging between the four 6-coordinate iron(II)
photoreactors,1,2 in the stabilization of reactive species,3,4 and in
ions at the vertices. The cage exhibits simple 1H and 13C NMR
molecular recognition.5−7 These abiological structures are also
spectra in solution (Figures S1 and S2), consistent with T point
of relevance as functional mimics of biological molecules such
symmetry. Cage 1 was prepared from iron(II) tetrafluoroborate
as protein receptors8−10 and enzymes.11,12
(1·(BF4)8), perchlorate (1·(ClO4)8), hexafluorophosphate
In a similar fashion to enzymes, container molecules can
(1·(PF6)8), or bis(trifluoromethane)sulfonimide (triflimide,
catalytically accelerate specific chemical transformations of their
1·(NTf2)8).
guests.13−20 In a first step toward mimicking the elegant
Other anilines could also be incorporated into the cage
enzymatic “assembly lines” that construct the most complex
corners; anisidine was chosen for subsequent studies because
and functional natural products,21 systems of catalytic container
the corresponding cage was easier to purify through
molecules might be used to carry out sequences of reaction
recrystallization, enabling the isolation of crystals of sufficient
steps wherein the product of one capsule’s transformation
quality for single-crystal X-ray diffraction analysis. It was
becomes the substrate for the next. The use of protective
determined that incorporation of different anilines does not
capsules in such sequences,4,22−30 capable of storing and
affect the binding affinity of anions for cage 1 (Figure S30).
preventing other substrates’ reactions, could allow more
Single crystals of the triflimide salt of 1 of sufficient quality
complex behavior within systems of transformative hosts.
for X-ray analysis could not be obtained despite numerous
As a precondition for designing systems of functional
attempts. However, crystal structures of both the tetrafluor-
container molecules, it is necessary to explore and understand
oborate (Figure 1) and perchlorate salts of 1 (Figure S17) were
how multiple guests interact with different hosts in ways that
obtained using synchrotron radiation at the Diamond Light
allow for complex behavior to emerge.31,32 In this study we
Source.45 The very small crystals were obtained by vapor
explore a system of two metal−organic hosts33−39 and multiple
guests, in which differential binding affinities between hosts and
guests allow transfer of a guest between hosts upon the addition
of a competing guest, laying the groundwork for networks of
diffusion of diisopropyl ether into acetonitrile solutions of the
cage. Each Fe4L6 complex has approximate (although not
crystallographic) T symmetry, with either a ΔΔΔΔ or ΛΛΛΛ
configuration at the metal centers. The metal−metal separa-
capsules to accomplish sequential multistep transformations.
tions are ∼13.4 Å, and the calculated cavity volume is
approximately 245 Å3, placing this cage among the larger
M4L6 species synthesized.40,46−48 The interior of 1 contains
three acetonitrile molecules in the structure of 1·(BF4)8, and a
mixture of diisopropyl ether and acetonitrile in 1·(ClO4)8.
RESULTS AND DISCUSSION
■
The present system comprises metal−organic container
molecule 1 (Scheme 1, top)40 and previously reported 2
(Scheme 1, bottom),41 both of which were prepared using
subcomponent self-assembly,42−44 and both of which were
observed to bind anions with different affinities.
−
−
Cage 1 was observed to bind the anions BF4 , PF6 , or
NTf2 , as confirmed by 19F NMR and a H−19F HOESY
−
1
The reaction of 5,5′-(1,4-phenylene)bis-2-pyridinecarboxal-
dehyde (6 equiv), anisidine (12 equiv), and a suitable iron(II)
Received: December 5, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Formation of Cages 1 and 2 from Iron(II) and
Either 5,5′-(1,4-Phenylene)bis-2-pyridinecarboxaldehyde
and Anisidine (1) or 6,6′-Diformyl-3,3′-bipyridine and
Aniline (2)
Figure 1. X-ray crystal structure of [(MeCN)3⊂1]·8BF4·0.5MeCN.
Anions and solvents outside of the central cavity are omitted for
clarity. Only one location of the disordered encapsulated acetonitrile
molecules is shown.
acetonitrile solution of 1·(NTf2)8. Although it was not possible
to obtain the entire binding isotherm due to the necessity of
having to estimate the section between 0 and 8 equiv, the
binding of NTf2− was moderate enough to allow observation of
−
the isotherm above 8 equiv. The affinity of NTf2 was thus
estimated to be (1.4 0.5) × 103 M−1. The data fit well to a
1:1 binding model, and we infer the Coulombic and steric
penalties incurred by forcing two triflimides into the same
cavity to be prohibitive. We thus conclude that host 1 exhibits
−
1:1 binding with NTf2 .
In order to quantify the hierarchy of anion binding in host 1,
competitive titration experiments were performed. The titration
of BF4 , OTf−, PF6 , and ClO4 against cage 1·(NTf2)8 in
acetonitrile solution allowed the competitive binding constants
of these four anions to be determined (Figures S23, S25, S27,
and S29). Again 1:1 binding was found for all the anions, as
concluded from the good fit of all titration curves to a 1:1
binding model. Cage 1 did not display strongly differential
binding between these anions,49−51 with the exception of
perchlorate, which was observed to bind to cage 1 an order of
−
−
−
experiment (Figure S12) for cage 1·(PF6)8, in which nuclear
−
Overhauser correlations between PF6 resonances and peaks
−
corresponding to the central phenylene and inward-facing
pyridine protons of the ligand were observed. In each case, the
19F NMR resonances for the guest molecules were broadened
and shifted compared to the resonances of the anions in the
absence of cage 1 suggesting fast exchange on the NMR time
scale (Figures S3, S7, and S11).
In order to perform anion binding studies from an “empty”
cage, we attempted to prepare cage 1 with a large counterion
whose volume exceeds the cavity volume. However, when the
synthesis of cage 1 was carried out with the larger counterions
tetraphenylborate or tetrakis(3,5-bis(trifluoromethyl)phenyl)-
borate, cage 1 was not observed to form, and only materials
showing negligible solubility in common laboratory solvents
were obtained.
magnitude more strongly than NTf2 (Table 1). The other
−
anions bound 2−3 times more strongly than NTf2 ; the quality
of the data fitting allowed us to conclude that the differences
between the best and worst of these binders are small but real.
Anion Exchange Studies. The different anion binding
affinities of cage 1 allowed an anion displacement experiment to
be carried out, which illustrated the exchange of bound anions
−
within the system. When BF4 (8 equiv) was added to a
solution of 1·(NTf2)8, 19F NMR (Figure S35) indicated that a
−
−
large proportion of bound NTf2 was displaced by BF4 , as
predicted from their binding constants (Table 1).
Having demonstrated the ability to follow anion displace-
ment events with 19F NMR, the different anion affinities of
cages 1 and 2 allowed for the design of more complicated
systems in which multiple guests’ interactions with the two
hosts could be studied. We first validated that cages 1 and 2
An estimate of the binding constant of triflimide for 1 was
−
obtained by titrating excess NTf2 (Figure S19) into an
B
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
anion exchange at 298 K (a temperature much higher than the
coalescence temperature, which must be below 233 K, as
demonstrated by low temperature NMR experiments) is
considerably greater than 88 s−1 and, therefore, more than 6
orders of magnitude faster than the most rapid ligand
dissociation rate of 10−5−10−6 s−1.32 We infer, thus, the guest
exchange of these anions to take place through the faces of the
host, on a significantly faster time scale than dynamic ligand
exchange.
Table 1. Summary of the Binding Constants of Anions for
Cages 1 and 2
binding affinity for cage 1
(M−1
binding affinity for cage 2
b
anion
)
(M−1
)
a
1.4(5) × 104
4.0(1) × 103
3.0(1) × 103
2.8(1) × 103
1.4(5) × 103
5.7(2) × 105
−
ClO4
ac
,
−
BF4
2.3(4) × 104
a
OTf−
5.2(8) × 104
1.3(5) × 106
a
−
PF6
−
NTf2
no binding observed
NMR spectra of cage 2 in the presence of PF6 and OTf−
−
a
Binding affinities for these anions in cage 1 were obtained from
showed slow guest exchange processes on the NMR time scale.
To probe the guest exchange mechanism for these systems we
determined the rates of guest exchange for OTf−, the largest
competitive titration experiments (see Supporting Information).
Standard deviations in the least significant digit are given in
b
−
parentheses. The value for ClO4 was determined as part of this
1
guest bound within 2, by H−1H EXSY NMR experiments. At
study (Figure S32), whereas those for BF4 , OTf−, and PF6− have been
−
298 K the first-order uptake rate constant k′in of OTf− into cage
2 was measured to be (1.8 0.4) × 10−2 s−1, and the release
rate constant was measured to be (5.2 0.6) × 10−2 s−1 (see
Supporting Information). Both exchange rates are ca. three
orders of magnitude higher than the rates of ligand
dissociation,32 leading us to conclude that the encapsulation
reported previously.41 A direct titration of excess BF4 into 1·(BF4)8
c
−
−
yielded a similar value for the binding affinity of BF4 for cage 1
(Figure S21).
could be studied in solution together. No scrambling of aniline
and anisidine residues was observed between the two cages
during the titration experiments discussed below, and further
mechanism for the “slowly” exchanging PF6 and OTf− into
−
cage 2 is also by diffusion of the guest through the faces of the
cage.
1
experiments showed no evidence of scrambling by H NMR
and ESI-MS under the experimental conditions employed
(Figures S39−S43). We infer that the imine exchange52
observed in similar structures41 is dependent on either the
concentration of free aniline or water,53 which have been
minimized in the present experimental conditions.
Next, the interactions of a single anion within a 1:1 system of
both cages 1 and 2 were studied in detail. Anion titration
experiments were carried out with BF4 , PF6 , OTf−, and
−
−
−
ClO4 against an equimolar solution of both cages 1·(NTf2)8
and 2·(NTf2)8 in acetonitrile (Table 2 shows the results of
anion titration experiments with 1 equiv of each anion). As
expected, the first equivalent of each anion was encapsulated
within empty cage 2, and only after more than one equivalent
had been introduced did the added anion begin to compete
with triflimide to bind within cage 1. We were able to combine
the equilibrium expressions for anion binding within cages 1
and 2 with the binding constants tabulated in Table 1 to derive
a series of analytical expressions for the composition of a
The mechanism of anion encapsulation for cages 1 and 2
may take place via one of two possible processes, either by the
diffusion of anions through the open faces of the cages assisted
by twisting of the ligands, or by the dissociation of the ligands
involving breakage of Fe−N bonds. Based upon investigations
of similar systems,32 we anticipate the rate of ligand dissociation
for cages 1 and 2 to be in the range of 10−5−10−6 s−1.32
Only one set of cage signals was ever observed in NMR
spectra of cage 1, indicating that all guests investigated
underwent fast exchange on the NMR time scale between the
free and cage-bound states. Fast exchange of triflimide, the
largest guest, was observed even at temperatures as low as 233
K (Figure S33). Similarly, NMR spectra of cage 2 in the
−
system containing 1, 2, NTf2 , and each of the other anions.
The equilibrium that is set when mixing cage 1·(NTf2)8 and
−
cage 2·(NTf2)8 with counteranionic guest G (where G = BF4 ,
−
−
OTf−, PF6 , or ClO4 ) starts with G encapsulated within cage
2, because triflimide cannot enter cage 2, then G will be
transferred to cage 1, leaving cage 2 empty:
−
−
presence of either BF4 or ClO4 displayed only one set of
signals. The observation of fast exchange is consistent with an
exchange rate of guests between the cage cavity and the bulk
solution that is higher than the difference in frequency between
fully bound host signal and free host signal in the NMR spectra.
(1)
G ⊂ 2 + X ⊂ 1 ⇌ 2 + G ⊂ 1 + X
The equilibrium constant for this equilibrium is given by eq 2:
1
The difference between the H NMR chemical shift values of
[2][G ⊂ 1][X]
[G ⊂ 2][X ⊂ 1]
[2] [G ⊂ 1]
[G ⊂ 2] [X ⊂ 1]
K* =
=
[X]
the host’s imine signal in the free and bound states was
estimated to be ca. 40 Hz (0.1 ppm at 400 MHz, see Figures
(2)
S19 and S32). This value would give an exchange rate kex
≈
where [2] is the concentration of empty cage 2, [G⊂1] the
concentration of cage 1 with anion G inside, [G⊂2] the
2.2Δν = 88 s−1 at the coalescence temperature.54 The rate of
Table 2. Proportions of 1 equiv of Each Anionic Guest Encapsulated in Cages 1 and 2, and Free in Solution, As Determined
from Titration Experiments Carried out on an Equimolar Solution of Cages 1·(NTf2)8 and 2·(NTf2)8
a
b
c
d
e
guest
y
[1]0 (M)
K1
K2 (M−1
)
q
(1 − y) − q
ClO4
0.93
0.87
0.88
0.98
5.0 × 10−4
1.0 × 10−3
5.0 × 10−4
5.0 × 10−4
10.1(1)
2.9(2)
2.2(1)
2.1(4)
5.7(2) × 105
2.3(4) × 104
5.2(8) × 104
1.3(5) × 106
0.03
0.05
0.04
0.01
0.04
0.08
0.08
0.01
−
−
BF4
OTf−
−
PF6
a
b
c
Fraction of cage 2 that incorporates guest. Competitive binding constant of the guest against NTf2− for cage 1. Binding constant of the guest for
d
empty cage 2. For K1 and K2, standard deviations in the least significant digits are given in parentheses. Fraction of cage 1 that incorporates guest.
e
Proportion of guest free in solution.
C
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
−
Scheme 2. Anion Exchange Sequence: Binding of ClO4− within Cage 2, Followed by Its Displacement by PF6 , and Subsequent
−
−
Displacement of NTf2 from Cage 1 by the Released ClO4
−
−
concentration of cage 2 with anion G inside, [X⊂1] the
exchanging BF4 and ClO4 , we related the chemical shifts of
the imine proton to the chemical shift of the same proton for
−
concentration of cage 1 with NTf2 inside, [X] the
−
concentration of free NTf2 , K1 the competitive binding
an empty cage and for a cage completely bound with BF4 or
−
constant of anion G against NTf2 for cage 1 and K2 the
−
ClO4 , and for the slow-exchanging OTf− and PF6 , we used
the comparison of the integration of the 1H signal of the imine
proton for both the empty and anion bound cages. The value q
was determined by taking the previously determined binding
−
−
binding constant of anion G to cage 2.
With the assumption that cage 2 can only encapsulate
−
anionic guest G (and not the NTf2 anion) and that no
constants, K1 and K2, the value of y as determined from H
1
“empty” cage 1 is available, and by considering the mass
balances for the total amount of host and guest species, the
equilibrium of eq 1 can be rearranged into eq 3 (full derivations
of all equations are reported in the Supporting Information):
NMR and the total concentration of cage 1. Since the total
concentration of the two hosts 1 and 2 and the guest G are all
identical, the difference between (1 − y) and q represents the
amount of guest G that is free in solution, as expressed in
equivalents with respect to cage 1 (final column in Table 2).
From the comparison of the y and q values of all four anions,
it can be seen that PF6− binds almost exclusively into cage 2. It
can also be observed that ClO4− competes most effectively with
and thus is best at displacing NTf2− for cage 1, and it is not the
strongest binder for cage 2, giving the possibility of it being
K1
K* =
K2
[2][G ⊂ 1][X]
=
[G ⊂ 2][X ⊂ 1]
q
1 − y
=
[16 − y][1]0
−
displaced from 2 upon addition of PF6 . From this analysis a
1 − q
y
(3)
sequence which capitalizes upon this binding behavior can be
deduced wherein an anion is displaced from cage 2 by the
addition of another higher affinity anion, and the supplanted
anion is then subsequently able to displace the anion already
present in cage 1, resulting in a chain-reaction exchange of
different anions within the 1:1 two cage system (Scheme 2).
State A (Scheme 2) shows the initial equilibrium of the
equimolar solution of cages 1·(NTf2)8 and 2·(NTf2)8, with cage
2 empty, and the 16 equiv of NTf2 present within the system
saturating cage 1. State B was thus entered into when 2 equiv of
ClO4 were added to the system in state A, with the ClO4
anion saturating cage 2, and a minority of the NTf2 being
where y is the fraction of cage 2 that incorporates guest G and q
is the fraction of cage 1 that incorporates guest G. The [X] =
(16 − y)[1]0 term means that at the start of the experiment, of
the 16 equiv of NTf2− per cage 1, 15 equiv are free in solution
and 1 equiv is incorporated in cage 1. At equilibrium, as a result
−
of displacement of the NTf2 anion by guest G in cage 1 the
total number of equivalents of free NTf2− will have become 15
+ (1 − y).
−
The degree of encapsulation of anions by cage 2 shown in
Table 2 (term y) was calculated from the NMR titration
experiments with a 1:1 system of cages 1 and 2. For the fast-
−
−
−
D
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
−
Preparation of Cage 1·(NTf2)8. 5,5′-(1,4-Phenylene)bis-2-pyr-
idinecarboxaldehyde (24 mg, 0.083 mmol), p-methoxyaniline (20 mg,
0.167 mmol), and iron(II) bis(trifluoromethane)sulfonimide dihydrate
(34 mg, 0.055 mmol) were added to a 25 mL Schlenk flask containing
degassed acetonitrile (6 mL). The flask was sealed and subjected to
three evacuation/nitrogen fill cycles. The reaction was stirred for 24 h
at 323 K. 1·(NTf2)8 was precipitated as a green powder by the addition
displaced from cage 1 by the remaining ClO4 , as indicated by
19F NMR (Figure S37). The addition of 1 equiv of PF6 then
−
−
brought the system into state C, where PF6 had completely
−
displaced the bound ClO4 from within cage 2 (Figure S36).
The system’s 2 equiv of ClO4 then acted to displace the
majority of the NTf2 from cage 1 (Figure S37), resulting in a
−
−
1
chain of anion exchange reactions.
of diethyl ether (yield 48 mg, 66%). H NMR (500 MHz, 298 K,
For each of the states shown in Scheme 2, experiments and
calculations thus reflect that the majority of each host binds the
guest shown in each of the system’s three states. The system’s
selectivity is imperfect, with cage 1’s lack of differentiation
between guests leaving room for future improvement. The large
CD3CN): δ = 8.82 (s, 2H, imine), 8.66 (d, J = 8.5 Hz, 2H, pyridine),
8.57 (d, J = 8.0 Hz, 2H, pyridine), 7.47 (s, 4H, phenyl), 7.44 (s, 2H,
pyridine), 6.79 (d, J = 8.5 Hz, 4H, p-methoxyaniline), 5.45 (s, DCM),
3.81 (s, 6H, methoxy); 13C NMR (125 MHz, 298K, CD3CN): δ =
173.90, 160.64, 158.33, 153.31, 144.17, 139.64, 137.56, 136.01, 131.04,
128.45, 123.78, 115.11, 56.22. 19F NMR (376 MHz, 298 K, CD3CN):
δ = −81.41 (−81.33 for LiNTf2). Elemental analysis (%) calcd for
chemical formula: C208H156F48Fe4N32O44S16·11H2O: C, 44.18; H,
3.17; N, 7.93. Found: C, 44.13; H, 2.88; N, 7.84. m/z: 1538.5
{[1·(NTf2)5]3+}, 1083.7 {[1·(NTf2)4]4+}, 810.9 {[1·(NTf2)3]5+}.
Crystallography. Data for 1·(BF4)8 and 1·(ClO4)8 were collected
at Beamline I19 of Diamond Light Source employing silicon double
crystal monochromated synchrotron radiation (0.6889 Å) with ω
scans at 100(2) K.58 Pictures of the crystal structure of
[(MeCN)1.5(iPr2O)0.5⊂1]·8ClO4·4MeCN·iPr2O and the crystal
packing are shown in the Supporting Information (Figure S17).
Data integration and reduction for [(MeCN)3⊂1]·8BF4·0.5MeCN
were undertaken with CrystalClear58 and data for [(MeCN)1.5(iPr2O)-
0.5⊂1]·8ClO4·4MeCN·iPr2O were first treated with ECLIPSE,59 the
cell orientations determined with CELL_NOW60 and then integration
and reduction performed with SAINT and XPREP.61 Subsequent
computations were carried out using the WinGX-32 graphical user
interface.62 Multiscan empirical absorption corrections were applied to
the data using the programs CrystalClear58 or TWINABS.63 Structures
were solved by charge flipping using SUPERFLIP64 then refined and
extended with SHELXH-97.65 In general, non-hydrogen atoms with
occupancies greater than 0.5 were refined anisotropically. Carbon-
bound hydrogen atoms were included in idealized positions and
refined using a riding model. The crystals were extremely unstable,
rapidly decaying once removed from the mother liquor and required
rapid handling at dry ice temperatures (<5 s) to facilitate data
collection. The crystals were also weakly diffracting with very few
reflections recorded at higher than 1.0 Å resolution for
−
amount of NTf2 present within the starting system also
reduces the system’s selectivity; the use of a larger counterion,
which does not bind to either 1 or 2, proved infeasible due to
the insolubility of the cages with larger anions. However, it is
−
significant that displacement of a majority of the bound NTf2
could be achieved through the addition of only 2 equiv of
−
−
ClO4 , despite the 16 equiv of NTf2 present in the system.
The exchange sequence did not include OTf− or BF4 due to
−
the observation that their binding within cage 2 was not strong
enough to displace bound perchlorate, and their binding with
cage 1 was also too weak to displace a significant portion of
−
NTf2 . Higher differentials between the affinities of 1 for
different anions (Table 1) would be necessary to improve the
selectivity of the system.
CONCLUSIONS
■
A cationic tetrahedral cage possessing a large enough cavity for
the binding of the triflimide anion was synthesized via
subcomponent self-assembly, and its host−guest chemistry
with a selection of anions was studied. The interactions of the
anions with an equimolar solution of this larger cage and a
smaller cage were also studied in detail. The resultant anion-
exchange sequence within the system of cages 1 and 2
demonstrates how different binding affinities may be harnessed
in order to design sequential reactions in which guests are
passed from one host to another. Although the selectivity in
this chain-reaction is not optimal, it provides the foundation
upon which future systems will be built. The coupling of
improved chains to the functions of catalytic cages14−19 could
allow for the design of new chemical systems in which multiple
reaction steps could be organized to progressively build up
molecular complexity. Such linked networks of capsules could
also serve to design indicator displacement assays55,56 that
respond differentially to combinations of analytes.
[(MeCN)3⊂1]·8BF4·0.5MeCN and 0.9
Å resolution for
[(MeCN)1.5(iPr2O)0.5⊂1]·8ClO4·4MeCN·iPr2O. Accordingly there
are a number of larger than ideal Ueq max/min ratios. Both structures
show disorder in the solvent and anions and a number of constraints
and restraints were required to facilitate realistic modeling. The
crystallographic data are summarized below.
[(MeCN)3⊂1]·8BF4·0.5MeCN. Formula C199H166.5B8F32Fe4N27.5O12,
M 4052.98, monoclinic, space group P21/c (No. 14), a = 27.191(11), b
= 39.825(15), and c = 20.852(8) Å, β = 96.369(3)°, V = 22442(15)
Å3, Dc 1.200 g cm−3, Z = 4, crystal size 0.1 × 0.01 × 0.001 mm, dark
green blade, temperature = 100(2) K, λ(synchrotron) = 0.68890 Å,
μ(synchrotron) = 0.338 mm−1, T(CrystalClear)min,max = 0.409, 1.000,
EXPERIMENTAL SECTION
2θmax = 40.30, hkl range −27 27, −37 39, −20 20, N = 118110, Nind
=
■
23420 (Rmerge = 0.1108), Nobs = 16737 (I > 2σ(I)), Nvar = 2045,
residuals R1(F) = 0.1310, wR2(F2) = 0.3031, GoF(all) = 1.487,
Δρmin,max = −0.847, 1.318 e− Å−3.
All reagents and solvents were purchased from commercial sources
and used as supplied. NMR spectra were recorded on a Bruker Avance
DPX400 or DPX500 spectrometer; δH values are reported relative to
acetonitrile-d3 at 1.94 ppm. δC values are reported relative to
acetonitrile-d3 at 118.10 ppm. 19F values are reported relative to the
internal standard hexafluorobenzene at −166.70 ppm. Mass spectra
provided by the EPSRC National MS Service Centre at Swansea were
acquired on a Thermofisher LTQ Orbitrap XL. Low-resolution
electrospray ionization mass spectra (ESI-MS) were obtained on a
Micromass Quattro LC, infused from a Harvard Syringe Pump at a
rate of 10 μL per minute. Cage 2 was synthesized as previously
reported.41 Iron(II) bis(trifluoromethane)sulfonamide dihydrate41,57
and hexakis(acetonitrile)iron(II) hexafluorophosphate32 were pre-
pared as previously reported. The syntheses of 5,5′-(1,4-phenylene)-
bis-2-pyridinecarboxaldehyde, 1·(BF4)8, 1·(PF6)8, and 1·(ClO4)8 are
reported in the Supporting Information.
The positions of only four of the eight anions could be located with
the remaining anions located in a region of smeared electron density.
Despite numerous attempts at modeling this region of disorder as a
combination of both disordered anions and solvent (including the use
of rigid bodies) no successful model could be obtained and the data
was treated with the SQUEEZE66 function of PLATON,67 which
resulted in significantly better residuals.
[(MeCN)1 . 5(iPr2 O)0. 5⊂1]·8ClO4 ·4MeCN·iPr2 O. Formula
C212H193.5Cl8Fe4N29.5O45.5, M 4389.46, triclinic, space group P1 (No.
̅
2), a = 20.495(9), b = 25.670(12), and c = 25.757(15) Å, α =
118.873(17), β = 109.64(4), and γ = 93.55(3)°, V = 10755(9) Å3, Dc =
1.355 g cm−3, Z = 2, crystal size 0.15 × 0.02 × 0.001 mm, dark green
blade, temperature = 100(2) K, λ(synchrotron) = 0.68890 Å,
E
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
μ(synchrotron) = 0.446 mm−1, T(TWINABS)min,max = 0.505585,
0.745532, 2θmax = 45.00, hkl range −22 21, −28 24, 0 28, N = 30953,
Nind = 30972 (Rmerge = 0.1827), Nobs = 18574 (I > 2σ(I)), Nvar = 2555,
residuals R1(F) = 0.0915, wR2(F2) = 0.2288, GoF(all) = 1.029,
Δρmin,max = −1.255, 1.692 e− Å−3.
To study the rate of exchange for the largest guests into cages 1 and
2, a solution of 1·(NTf2)8 (2.0 mg, 3.6 × 10−7 mol, 8 equiv) in
1
CD3CN (0.5 mL) was sealed in a J-Young tube and the H NMR
spectrum was acquired at 233 K. To the same tube, tetrabutylammo-
nium triflate (0.14 mg, 3.6 × 10−7 mol, 8 equiv) was added from a
premade CD3CN stock solution, and the solution was sealed in a J-
Despite appearing (at least visually) to be a single crystal, the
sample employed in this study proved to be a twinned by a 2-fold
rotation approximately around c. The BASF parameter refined to
1
Young tube and sonicated for 10 min. The H NMR spectrum was
again acquired at 233 K. 2D EXSY NMR experiments were run onto a
solution of OTf−⊂2·(NTf2)8 (7.0 mg, 3.8 × 10−3 mol, 1 equiv, OTf−
0.5 equiv) at 298 and 308 K. Rate constants are reported in the
Supporting Information.
−
0.55274. It was possible to locate the positions of all the ClO4
counterions (disordered over nine lattice sites) and the majority of the
solvent molecules (with a remaining solvent accessible void of 100 Å3
per unit cell) and SQUEEZE was not employed in this case. One of
the central phenyl rings shows positional disorder and was modeled in
two parts with isotropic thermal parameters. One of the periphery
OMe groups shows some unresolved disorder and was also modeled
isotropically. The occupancies of the two positions of the disordered
solvent molecules inside the tetrahedral cavity were refined and then
fixed at the obtained values.
Ka Determination. The derivation of the equations for competitive
binding studies and of the expression of the fraction of occupied host
as a function of total guest concentration was published elsewhere.41
The binding isotherms for all competitive binding titration experi-
ments show 1:1 binding for all anions. The concentration of cage
1·(NTf2)8 was kept constant throughout all titration experiments and
small aliquots of the second guest solution were titrated into a J-Young
NMR tube containing 1·(NTf2)8 before each spectrum was acquired.
All NMR data and binding constant fittings are reported in the
Supporting Information. The titration of fast exchanging [OTf−] =
0.102 M guest into a solution of [1·(NTf2)8] = 3.4 × 10−3 M resulted
A 1:1 mixture of 1·(NTf2)8 (1.0 mg, 1.8 × 10−7 mol, 1 equiv) and
2·(NTf2)8 (1.1 mg, 2.1 × 10−7 mol, 1 equiv) dissolved in CD3CN (0.4
mL) was sealed in a J-Young tube. The dark blue mixture was
sonicated for 10 min. The concentration of the mixture of cages
1·(NTf2)8 and 2·(NTf2)8 was kept constant throughout the titration
−
experiment and small aliquots of up to one equivalent of OTf−, BF4 ,
−
−
PF6 , or ClO4 solution were titrated into the mixture before each
spectrum was acquired. [cage] = 5.3 × 10−4 M. [anion] = 5.3 × 10−4
M. All NMR data are reported in the Supporting Information.
For anion exchange studies, a solution of 1·(NTf2)8 (1.0 mg, 1.8 ×
10−7 mol, 1 equiv) in CD3CN (0.5 mL) was sealed in a J-Young tube
and the dark green solution was sonicated for 10 min and the 19F
NMR spectrum was acquired. To the same tube, tetraethylammonium
tetrafluoroborate (0.4 mg, 1.8 × 10−6 mol, 10 equiv) was added from a
premade CD3CN stock solution and the 19F NMR spectrum was
acquired after sonication for 20 min. A solution of 1·(NTf2)8 (2.2 mg,
4.2 × 10−7 mol, 1 equiv) and 2·(NTf2)8 (2.0 mg, 4.2 × 10−7 mol, 1
equiv) dissolved in CD3CN (0.5 mL) were sealed in a J-Young tube.
The dark blue mixture was sonicated for 10 min to give state A. To the
same tube, tetrabutylammonium perchlorate (0.28 mg, 8.4 × 10−7 mol,
2 equiv) was added from a premade CD3CN stock solution giving
state B, and the 19F NMR spectrum was acquired after sonication for
20 min. To the same tube tetrabutylammonium hexafluorophosphate
(0.16 mg, 4.2 × 10−7 mol, 1 equiv) was added from a premade
CD3CN stock solution to give state C, and the 19F NMR spectrum was
again acquired after sonication for 20 min.
in a relative binding constant of 2.21
0.11. The titration of fast
−
exchanging [PF6 ] = 0.102 M guest into a solution of [1·(NTf2)8] =
3.4 × 10−3 M resulted in a relative binding constant of 2.06 0.35.
−
The titration of fast exchanging [BF4 ] = 0.054 M guest into a solution
of [1·(NTf2)8] = 1.8 × 10−3 M resulted in a relative binding constant
−
of 2.93 0.18. The titration of fast exchanging [ClO4 ] = 0.022 M
guest into a solution of [1·(NTf2)8] = 7.3 × 10−4 M resulted in a
relative binding constant of 10.06 1.12. Titration experiments with
cage 1 formed with aniline and NMR titrations for the binding affinity
−
ASSOCIATED CONTENT
* Supporting Information
of ClO4 for cage 2 are reported in the Supporting Information.
■
Estimates of the binding constants of triflimide for cage 1 were
S
−
obtained by titrating excess NTf2 into solutions 1·(NTf2)8. The
Experimental procedures; ligand and cage characterization data;
NMR titration studies; Ka determination; equilibrium equation
derivation. This material is available free of charge via the
Internet at http://pubs.acs.org. The crystallographic data have
been deposited with the Cambridge Crystallographic Data
Centre as entries CCDC 877966 and 877967.
derivation of the equations for these binding studies was published
elsewhere.41 The minimum of the total guest concentration (the X
parameter) was set to 0 in the fitting model, and the initial chemical
shift Y parameter was optimized by the binding algorithm to give a
−
value of 7.40 ppm. The binding of NTf2 was therefore estimated to
be (1.4 0.5) × 103 M−1. The concentration of cage 1·(NTf2)8 was
kept constant throughout the titration experiment and small aliquots
of NTf2− solution were titrated into a J-Young NMR tube before each
AUTHOR INFORMATION
Corresponding Author
jrn34@cam.ac.uk
■
−
spectrum was acquired. [1·(NTf2)8] = 2.3 × 10−4 M. [NTf2 ] = 0.003
M.
Estimates of the binding constants of BF4− for cage 1 were obtained
by titrating excess BF4− into solutions 1·(BF4)8. The concentration of
cage 1·(BF4)8 was kept constant throughout the titration experiment
and small aliquots of BF4− solution were titrated into a J-Young NMR
tube before each spectrum was acquired. [1·(BF4)8] = 2.3 × 10−4 M,
Present Addresses
†Biomolecular NanoTechnology Group, MESA+ Institute for
Nanotechnology, University of Twente, PO Box 217, 7500AE
Enschede, The Netherlands
−
‡School of Chemistry and Molecular Biosciences, The
University of Queensland, Brisbane St Lucia, QLD, Australia
4072
[BF4 ] = 0.003 M. The minimum of the total guest concentration (the
X parameter) was set to 0 in the fitting model, and the initial
−
parameter for Y was set to 7.40 ppm, and the binding of BF4 was
estimated to be (4.9 3.6) × 103 M−1.
Notes
Anion Exchange Experiments. To investigate potential scram-
bling effects between the two cages, a 1:1 mixture of 1·(NTf2)8 (1.0
mg, 2.1 × 10−7 mol, 1 equiv) and 2·(NTf2)8 (1.1 mg, 2.1 × 10−7 mol, 1
equiv) dissolved in CD3CN (0.5 mL) was sealed in a J-Young tube.
The dark blue mixture was sonicated for 10 min before being subjected
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the NMR service team at the Cambridge Chemistry
Department for carrying out some of the NMR spectroscopy,
the EPSRC Mass Spectrometry Service at Swansea for carrying
out the high-resolution mass spectrometry, and the Diamond
Light Source (UK) for synchrotron beamtime on I19 (MT6791
1
to H NMR and ESI-MS analysis. Extra aniline (0.08 mg, 8.4 × 10−7
mol, 4 equiv) from a premade CD3CN stock solution was added to the
1:1 cage mixture and left for 24 h at 323 K. 1H NMR spectra and MS
data of scrambling experiments are reported in the Supporting
Information.
F
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(36) Cangelosi, V. M.; Zakharov, L. N.; Johnson, D. W. Angew.
Chem., Int. Ed. 2010, 49, 1248.
and MT7114). This research was funded by EPSRC. M.M.J.S.
acknowledges The Netherlands Organization for Scientific
Research (NWO Rubicon Fellowship), Y.R.H. acknowledges
the Cambridge European Trust, and J.K.C. acknowledges the
Marie Curie IIF scheme of the 7th EU Framework Program.
(37) Turega, S.; Whitehead, M.; Hall, B. R.; Haddow, M. F.; Hunter,
C. A.; Ward, M. D. Chem. Commun. 2012, 48, 2752.
(38) Albrecht, M.; Shang, Y. L.; Rhyssen, T.; Stubenrauch, J.;
Winkler, H. D. F.; Schalley, C. A. Eur. J. Org. Chem. 2012, 2422.
(39) Ronson, T. K.; Fisher, J.; Harding, L. P.; Rizkallah, P. J.; Warren,
J. E.; Hardie, M. J. Nat. Chem. 2009, 1, 212.
(40) Ousaka, N.; Grunder, S.; Castilla, A. M.; Whalley, A. C.;
Stoddart, J. F.; Nitschke, J. R. J. Am. Chem. Soc. 2012, 134, 15528.
(41) Hristova, Y. R.; Smulders, M. M. J.; Clegg, J. K.; Breiner, B.;
Nitschke, J. R. Chem. Sci. 2011, 2, 638.
(42) Campbell, V. E.; Nitschke, J. R. Synlett 2008, 20, 2077.
(43) Domer, J.; Slootweg, J. C.; Hupka, F.; Lammertsma, K.; Hahn,
F. E. Angew. Chem., Int. Ed. 2010, 49, 6430.
(44) Zhou, X.-P.; Liu, J.; Zhan, S.-Z.; Yang, J.-R.; Li, D.; Ng, K.-M.;
Sun, R. W.-Y.; Che, C.-M. J. Am. Chem. Soc. 2012, 134, 8042.
(45) Nowell, H.; Barnett, S. A.; Christensen, K. E.; Teat, S. J.; Allan,
D. R. J. Synchrotron Radiat. 2012, 19, 435.
(46) Meng, W.; Clegg, J. K.; Thoburn, J. D.; Nitschke, J. R. J. Am.
Chem. Soc. 2011, 133, 13652.
(47) Scherer, M.; Caulder, D. L.; Johnson, D. W.; Raymond, K. N.
Angew. Chem., Int. Ed. 1999, 38, 1588.
(48) Biros, S. M.; Yeh, R. M.; Raymond, K. N. Angew. Chem., Int. Ed.
2008, 47, 6062.
(49) Sessler, J. L.; Camiolo, S.; Gale, P. A. Coord. Chem. Rev. 2003,
240, 17.
(50) Hua, Y. R.; Flood, A. H. Chem. Soc. Rev. 2010, 39, 1262.
(51) Chung, M. K.; Severin, K.; Lee, S. J.; Waters, M. L.; Gagne, M.
R. Chem. Sci. 2011, 2, 744.
(52) Belowich, M. E.; Stoddart, J. F. Chem. Soc. Rev. 2012, 41, 2003.
(53) Dirksen, A.; Dirksen, S.; Hackeng, T. M.; Dawson, P. E. J. Am.
Chem. Soc. 2006, 128, 15602.
REFERENCES
■
(1) Furutani, Y.; Kandori, H.; Kawano, M.; Nakabayashi, K.;
Yoshizawa, M.; Fujita, M. J. Am. Chem. Soc. 2009, 131, 4764.
(2) Gupta, S.; Choudhury, R.; Krois, D.; Brinker, U. H.;
Ramamurthy, V. J. Org. Chem. 2012, 77, 5155.
(3) Mal, P.; Breiner, B.; Rissanen, K.; Nitschke, J. R. Science 2009,
324, 1697.
(4) Furusawa, T.; Kawano, M.; Fujita, M. Angew. Chem., Int. Ed.
2007, 46, 5717.
(5) Nakabayashi, K.; Ozaki, Y.; Kawano, M.; Fujita, M. Angew. Chem.,
Int. Ed. 2008, 47, 2046.
(6) Clever, G. H.; Tashiro, S.; Shionoya, M. Angew. Chem., Int. Ed.
2009, 48, 7010.
(7) Ariga, K.; Ito, H.; Hillab, J. P.; Tsukube, H. Chem. Soc. Rev. 2012,
41, 5800.
(8) Anslyn, E. V. J. Org. Chem. 2007, 72, 687.
(9) Rakow, N. A.; Suslick, K. S. Nature 2000, 406, 710.
(10) Buryak, A.; Severin, K. J. Am. Chem. Soc. 2005, 127, 3700.
(11) Carlier, P. R. Angew. Chem., Int. Ed. 2004, 43, 2602.
(12) Kirkorian, K.; Ellis, A.; Twyman, L. J. Chem. Soc. Rev. 2012, 41,
6138.
(13) Fabbrizzi, L. Top. Curr. Chem. 2012, 323, 127.
(14) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Science 2007,
316, 85.
(15) Yoshizawa, M.; Tamura, M.; Fujita, M. Science 2006, 312, 251.
(16) Murase, T.; Horiuchi, S.; Fujita, M. J. Am. Chem. Soc. 2010, 132,
2866.
(17) Yoshizawa, M.; Takeyama, Y.; Okano, T.; Fujita, M. J. Am.
Chem. Soc. 2003, 125, 3243.
(18) Schmidtchen, F. P. Angew. Chem., Int. Ed. 1981, 20, 466.
(19) Leung, D. H.; Fiedler, D.; Bergman, R. G.; Raymond, K. N.
Angew. Chem., Int. Ed. 2004, 43, 963.
(20) Koblenz, T. S.; Wassenaar, J.; Reek, J. N. H. Chem. Soc. Rev.
2008, 37, 247.
(21) Fischbach, M. A.; Walsh, C. T. Chem. Rev. 2006, 106, 3468.
(22) Cram, D. J.; Tanner, M. E.; Thomas, R. Angew. Chem., Int. Ed.
1991, 30, 1024.
(23) Horiuchi, S.; Murase, T.; Fujita, M. J. Am. Chem. Soc. 2011, 133,
12445.
(54) Sanders, J. K. M.; Hunter, B. K. Modern NMR Spectroscopy: A
Guide for Chemists; Oxford University Press: Oxford, UK, 1993.
(55) Nguyen, B. T.; Anslyn, E. V. Coord. Chem. Rev. 2006, 250, 3118.
(56) Severin, K. Chem. Commun. 2006, 42, 3859.
(57) Sibi, M. P.; Petrovic, G. Tetrahedron: Asymmetry 2003, 14, 2879.
(58) Rigaku CrystalClear; Rigaku Americas and Rigaku Corporation:
Texas, 1997−2009.
(59) Parsons, S. ECLIPSE; The University of Edinburgh: Edinburgh,
UK, 2004.
(60) Bruker-Nonius Cell_NOW; Bruker AXS Inc.: Madison, WI,
2003.
(61) Bruker-Nonius APEX v2.1, SAINT v.7, and XPREP v.6.14;
Bruker AXS Inc.: Madison, WI, 2003.
(62) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837.
(24) Yoshizawa, M.; Kusukawa, T.; Fujita, M.; Yamaguchi, K. J. Am.
Chem. Soc. 2000, 122, 6311.
(63) Sheldrick, G. SADABS: Empirical Absorption and Correction
Software: University of Gottingen: Germany, 1996−2008.
̈
(25) Smulders, M. M. J.; Nitschke, J. R. Chem. Sci. 2012, 3, 785.
(26) Ma, D.; Hettiarachchi, G.; Nguyen, D.; Zhang, B.; Wittenberg, J.
B.; Zavalij, P. Y.; Briken, V.; Isaacs, L. Nat. Chem. 2012, 4, 503.
(27) Custelcean, R.; Bonnesen, P. V.; Duncan, N. C.; Zhang, X. H.;
Watson, L. A.; Van Berkel, G.; Parson, W. B.; Hay, B. P. J. Am. Chem.
Soc. 2012, 134, 8525.
(64) Palatinus, L.; Chapuis, G. J. Appl. Crystallogr. 2007, 40, 786.
(65) Sheldrick, G. M. SHELX-97: Programs for Crystal Structure
Analysis; University of Gottingen: Gottingen, 1997.
̈
̈
(66) van der Sluis, P.; Spek, A. L. Acta Crystallogr., Sect. A: Found.
Crystallogr. 1990, 46, 194.
(67) Spek, A. L. PLATON: A Multipurpose Crystallographic Tool;
Utrecht University: Utrecht, The Netherlands, 2008.
(28) Chakrabarty, R.; Mukherjee, P. S.; Stang, P. J. Chem. Rev. 2011,
111, 6810.
(29) Cangelosi, M. V.; Zakharov, L. N.; Fontenot, S. A.; Pitt, M. A.;
Johnson, D. W. Dalton Trans. 2008, 3447.
(30) Breiner, B.; Clegg, J. K.; Nitschke, J. R. Chem. Sci. 2011, 2, 51.
(31) Liu, Y. Z.; Hu, C. H.; Comotti, A.; Ward, M. D. Science 2011,
333, 436.
(32) Clegg, J. K.; Cremers, J.; Hogben, A. J.; Breiner, B.; Smulders,
M. M. J.; Thoburn, J. D.; Nitschke, J. R. Chem. Sci. 2013, 4, 68.
(33) Qiu, W.; Perman, J. A.; Wojtas, L.; Eddaoudi, M.; Zaworotko,
M. J. Chem. Commun. 2010, 46, 8734.
(34) Olive, A. G. L.; Parkan, K.; Givelet, C.; Michl, J. J. Am. Chem.
Soc. 2011, 133, 20108.
(35) Zebret, S.; Dupont, N.; Bernardinelli, G.; Hamacek, J. Chem.
Eur. J. 2009, 15, 3355 .
G
dx.doi.org/10.1021/ja311882h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX