Encapsulation of ferrocene and peripheral electrostatic attachment of viologens to dimeric molecular capsules formed by an octaacid, deep-cavity cavitand

Article abstract of DOI:10.1002/chem.200701961

In aqueous media the deepcavity cavitand octaacid 1 forms stable dimeric molecular capsules 1₂, which are stabilized by hydrophobic effects. In this work we investigate the binding interactions in aqueous solution between these capsules and the redox active guests, ferrocene (Fc) and three 4,4×-bipyridinium (viologen) dications: methyl viologen (MV²⁺), ethyl viologen (EV²⁺), and butyl viologen (BV²⁺). Using NMR spectroscopic and electrochemical techniques we clearly show that the hydrophobic Fc guest is encapsulated inside 1₂. An interesting effect of this encapsulation is that the reversible voltammetric response of Fc is completely eliminated when it resides inside the 1₂ capsular assembly, a finding that is attributed to very slow electrochemical kinetics for the oxidation of Fc@1₂. Diffusion coefficient measurements (PGSE NMR spectroscopy) reveal that all three viologen guests are strongly bound to the dimeric capsules. However, the ¹H NMR spectroscopic data are not consistent with encapsulation and the measured diffusion coefficients indicate that two viologen guests can strongly associate with a single dimeric capsule. Furthermore, the (V²⁺)^2. 1₂ complex is capable of encapsulating ferrocene, clearly suggesting that the viologen guests are bound externally, via coulombic interactions, to the anionic polar ends of the capsule. The electrochemical kinetic rate constants for the reduction of the viologen residue in the V²⁺·1₂ complexes were measured and found to be substantially lower than those for the free viologen guests.

Full text of DOI:10.1002/chem.200701961

DOI: 10.1002/chem.200701961
Encapsulation of Ferrocene and Peripheral Electrostatic Attachment of
Viologens to Dimeric Molecular Capsules Formed by an Octaacid, Deep-
Cavity Cavitand
Dagmara Podkoscielny,[a] Ivy Philip,[a] Corinne L. D. Gibb,[b] Bruce C. Gibb,[b] and
Angel E. Kaifer*[a]
Abstract: In aqueous media the deep-
cavity cavitand octaacid 1 forms stable
dimeric molecular capsules 12, which
are stabilized by hydrophobic effects.
In this work we investigate the binding
interactions in aqueous solution be-
tween these capsules and the redox
active guests, ferrocene (Fc) and three
4,4’-bipyridinium (viologen) dications:
methyl viologen (MV2+), ethyl violo-
gen (EV2+), and butyl viologen (BV2+
ble voltammetric response of Fc is
completely eliminated when it resides
inside the 12 capsular assembly, a find-
ing that is attributed to very slow elec-
trochemical kinetics for the oxidation
of Fc@12. Diffusion coefficient meas-
urements (PGSE NMR spectroscopy)
reveal that all three viologen guests are
strongly bound to the dimeric capsules.
However, the 1H NMR spectroscopic
data are not consistent with encapsula-
tion and the measured diffusion coeffi-
cients indicate that two viologen guests
can strongly associate with a single di-
meric capsule. Furthermore, the (V2+)2·
12 complex is capable of encapsulating
ferrocene, clearly suggesting that the
viologen guests are bound externally,
via coulombic interactions, to the
anionic polar ends of the capsule. The
electrochemical kinetic rate constants
for the reduction of the viologen resi-
due in the V2+·12 complexes were mea-
sured and found to be substantially
lower than those for the free viologen
guests.
). Using NMR spectroscopic and elec-
G
trochemical techniques we clearly show
that the hydrophobic Fc guest is encap-
sulated inside 12. An interesting effect
of this encapsulation is that the reversi-
Keywords: electrochemistry · ferro-
cene · molecular capsules · supra-
molecular chemistry · viologen
Introduction
several receptors. A good deal of interest in these systems
revolves around the possibility of investigating the reactivity
and other properties of the encapsulated guests so that they
can be compared to those observed in bulk phases.[4,5] We
have previously reported on the encapsulation of redox-
active organometallic compounds[6,7] inside resorcinarene
molecular capsules and used ferrocenyl substituents to exert
redox control on the assembly of dimeric tetraurea calix[4]-
arene capsules.[8] In this work, we turn our attention to a
capsular system recently reported by two of us,[9] in which a
water-soluble, deep-cavity cavitand, octaacid 1, forms dimer-
ic molecular capsules that can include a variety of hydro-
phobic guests, such as steroids (Figure 1). The internal
cavity of 12 is estimated to be about 2 nm long and 1 nm in
diameter at the equator. The photochemistry of guests
inside 12 has proven to be very different from that observed
in homogeneous solution.[10–12] Here, we specifically report
on the binding interactions between several electroactive
guests, such as ferrocene and simple N,N’-dialkyl-4,4’-bipyri-
dinium (viologen) derivatives (see Figure 1 for structures),
Molecular encapsulation is one of the most intriguing as-
pects of supramolecular chemistry.[1] From Cramꢀs carcer-
ands and hemicarcerands[2] to the variety of molecular cap-
sule systems developed to date,[3] many receptors can be de-
signed and prepared to confine or trap guest molecules
within their cavities. Guest confinement or encapsulation
may take place in the cavity within a single molecular recep-
tor or in a cavity formed inside a well-defined assembly of
[a] D. Podkoscielny, Dr. I. Philip, Prof. A. E. Kaifer
Center for Supramolecular Science and Department of Chemistry
University of Miami, Coral Gables, FL 33124–0431 (USA)
Fax : (+1)305-284-5771
E-mail: akaifer@miami.edu
[b] Dr. C. L. D. Gibb, Prof. B. C. Gibb
Department of Chemistry, University of New Orleans
New Orleans, LA 70418 (USA)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
4704
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4704 – 4710

FULL PAPER
Figure 1. Structures of host 1 and the redox-active guests surveyed in this
work
and host 1. The host–guest interactions were found to be
very sensitive to the hydrophobic character of the guests
and provide a strong contrast between ferrocene and the vi-
ologen guests. While the former is fully encapsulated inside
the dimeric capsule, the viologen guests are strongly at-
tached to the capsuleꢀs surface via electrostatic interactions.
Results and Discussion
Encapsulation of ferrocene: The octaacid host 1 is known to
form dimeric capsules that can include hydrophobic com-
pounds in their inner cavities.[9,13,14] The formation of these
capsules is basically driven by hydrophobic forces, since mo-
lecular capsule formation and guest encapsulation minimize
the exposure of the guest surface and the inner surface of
the host cavity to water molecules. We have investigated in
detail the interactions between ferrocene (Fc), a hydropho-
bic, redox-active guest, and host 1. The low aqueous solubili-
ty of Fc limits the host/guest concentration ratios that can
be used in these experiments. However, Figure 2 shows the
proton NMR spectra of 1 as increasing concentrations of Fc
are added to the solution. The first experimental observa-
tion is that the presence of host 1 in the solution significant-
ly increases the solubility of guest Fc over the levels that
could be reached in the absence of 1. This finding suggests
the presence of binding interactions between Fc and 1.
Addition of small amounts of Fc to the D2O solution con-
taining 1 leads to the appearance of a new set of peaks for
the host protons. The only proton signal that is not affected
by the addition of Fc is that at d=7.55 ppm, which corre-
sponds to the aromatic protons located at the bottom of the
cavity (labeled “J” in Figure 1 and Figure 2), next to the
“feet” of the cavitand host. All other proton signals are sub-
stantially affected by the guest. The Fc-induced signals in-
crease, at the expense of the original host proton signals,
until the added amount of Fc reaches about 0.5 equivalents.
After this point, only the proton signals corresponding to
Figure 2. Partial 1H NMR spectra (400 MHz, 50 mm NaCl
borate buffer pH 8.9 in D2O) of host 1 (1.0 mm) in the presence of in-
creasing concentrations of Fc. Proton resonances labeled with a star cor-
respond to complexed 1.
+ 10 mm
the Fc complex are visible. The symmetry of the host is not
broken by the inclusion of the guest and could correspond
to the monomeric host (C4v) or the dimeric capsule (D4h).
However, the fact that the protons most affected by guest
inclusion are those at the cavity portal points to the head-
to-head dimerization of 1 to form a capsular assembly. Fur-
thermore, the stoichiometry of the binding interactions
(2 hosts/1 guest), as revealed by the NMR spectra, suggests
the encapsulation of the Fc guest inside the 12 assembly.
This is also consistent with the observed chemical shift for
the Fc protons (d=2.16 ppm), which are considerably shift-
ed upfield from their resonance frequency prior to encapsu-
lation. This pronounced encapsulation-induced upfield shift
is a result of the ring currents that the Fc protons experience
while surrounded by the aromatic walls of the capsule.
The pulse gradient stimulated echo (PGSE) NMR tech-
nique has become extremely popular for the determination
Chem. Eur. J. 2008, 14, 4704 – 4710
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4705

A. E. Kaifer et al.
of diffusion coefficients (Do) and as a general tool to investi-
gate molecular association phenomena in the solution
phase.[15] Therefore, we measured the diffusion coefficients
of the host and the guests used in this work. The Do value of
the octaacid in D2O solution also containing 10 mm sodium
tetraborate was measured as 2.2”10ꢀ6 cm2 sꢀ1, a value that
corresponds to the monomeric form of the host. In the pres-
ence of 0.5 equivalents of Fc, the diffusion coefficient of the
host decreases to 1.6”10ꢀ6 cm2 sꢀ1. The same value was ob-
tained by using the signal corresponding to the Fc protons,
which clearly reveals that the guest and the host diffuse at
the same rate. We must point out that the error margins in
the Do measurements are consistently less than 5%, and the
differences between the values assigned to dimeric and mo-
Figure 3. SWV response on a glassy carbon electrode (0.071 cm2) of
1.0 mm host 1 in 50 mm NaCl + 10 mm pH 8.9 borate buffer in the pres-
ence of 0.5 equiv (continuous line), 1.0 equiv (discontinuous line) and
A
tions of each value. Coupled to the rest of the experimental
data on this host-guest system, the diffusion coefficients pro-
vide strong support for the idea that ferrocene is encapsulat-
ed inside the capsular assembly formed by two molecules of
the host.
2.0 equiv (dotted line) guest Fc. Scan rate: 0.1 Vsꢀ1
.
Ferrocene undergoes reversible (fast) one-electron oxida-
tion to its positively charged form, ferrocenium. In the past,
we have investigated the electrochemistry of ferrocene trap-
ped inside a hemicarcerand[16] and inside hexameric molecu-
lar capsules formed by resorcinarenes.[6] We have also inves-
tigated the electrochemistry of dendrimers with a ferrocenyl
core, in which dendrimer growth results in the attenuation
of the corresponding electrochemical kinetics.[17,18] We were
thus quite interested in recording the electrochemical behav-
ior of Fc inside 12 molecular capsules using cyclic voltamme-
try (CV). The cyclic voltammograms recorded with solutions
containing 0.5–1.0 mm Fc in the presence of two equivalents
of host 1 are basically flat, with very small levels of faradaic
currents that can be associated with the oxidation of ferro-
cene. Unfortunately these Fc concentration levels cannot be
reached in the absence of the host, so a direct comparison
of the effect of the host is not possible in aqueous solution.
However, when excess amounts of Fc (in other words, when
[Fc]>0.5·[1]) are added to the solution, faradaic currents
corresponding to the Fc+/Fc redox couple are clearly detect-
ed and grow quickly with increasing concentrations of Fc.
This is consistent with the excess Fc remaining unassociated
with the molecular capsules and giving rise to larger levels
of current as anticipated for freely diffusing Fc. Solutions
with excess Fc become turbid, which is another indication
that the excess Fc stays away from the capsular assembly
and partially precipitates due to the low aqueous solubility
of this rather hydrophobic compound.
results, regardless of their intrinsic sensitivity differences.
The detection of faradaic currents in the presence of excess
Fc (when [Fc]>0.5·[1]) argues against any passivation of the
electrode surface. We must conclude that Fc inside the di-
meric molecular capsule 12 does not give rise to a measura-
ble current–potential response. In agreement with similar
observations in other cases in which Fc is encapsulated by a
large organic structure,[6] we interpret this experimental fact
as the result of slow heterogeneous electron transfer kinetics
between the encapsulated ferrocene center and the elec-
trode surface. Inclusion inside a sizable molecular capsule
leads to an increase in the distance of maximum approach
between the redox center and the electrode, which is expect-
ed to decrease the rate of electron transfer.
Interactions of simple viologen derivatives with the capsule
12: The binding interactions between methyl viologen
1
(MV2+) and 1 were initially investigated by H NMR spec-
troscopy in D2O solutions also containing 10 mm sodium
borate (pD=8.9) and 40–50 mm NaCl. Addition of one or
two equivalents of host 1 leads to minute changes in the
chemical shifts of the aromatic protons of the viologen
guest, but the resonance corresponding to the methyl pro-
tons shifts upfield by 0.15 ppm in the presence of two equiv-
alents of 1. Similar spectral changes were observed when
other viologen guests were used instead of MV2+. In the
case of butyl viologen (BV2+), the presence of two equiva-
lents of 1 results in upfield shifts of about 0.33 ppm for the
proton resonance of the terminal methyl and 0.25 ppm for
the adjacent methylene on each side arm (see the Support-
ing Information). These spectral changes are consistent with
interactions between the viologen guests and host 1, but on
their own, they do not indicate encapsulation. It is extreme-
ly important to note that the proton signals of host 1 were
not affected by the addition of any of the viologen guests, in
strong contrast with our experimental observations upon ad-
dition of Fc as a guest.
We also recorded the electrochemical behavior of Fc
using square-wave voltammetry (SWV), a technique more
sensitive than CV. In the presence of two equivalents of
host 1, no faradaic current response was detected for Fc
(Figure 3), in excellent agreement with the lack of faradaic
current response observed in similar CV experiments. When
the Fc concentration exceeds 0.5·[1] the faradaic response is
clearly observed and the solution turns turbid immediately.
Therefore, both voltammetric techniques provide the same
4706
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4704 – 4710

Encapsulated Redox Centers
FULL PAPER
We also used PGSE NMR measurements to determine
the Do values of the viologen guests and the host. The re-
sults are given in Table 1. In the absence of 1, all three viol-
Table 1. Diffusion coefficients (cm2 sꢀ1) of viologens guests measured at
258C in 40 mm NaCl/D2O buffered at pD 8.9 with 10 mm sodium borate.
The standard deviations of all values were found to be <0.1”
10ꢀ6 cm2 sꢀ1
.
Guest
N
1
1.0
0[a]
MV2+
EV2+
BV2+
7.6”10ꢀ6
6.7”10ꢀ6
5.4”10ꢀ6
1.7”10ꢀ6
1.7”10ꢀ6
1.7”10ꢀ6
Figure 4. The diffusion coefficient of the BV2+ guest as a function of the
guest/host concentration ratio. All values were measured using PGSE
NMR techniques at 258C in 40 mm NaCl/D2O buffered at pD 8.9 with
10 mm sodium borate and 1.0 mm host 1. The measured Do value for
[BV2+]/[1]=0 corresponds to the host and was measured using host pro-
tons.
[a] Values in this column correspond to the diffusion coefficient of the
host in the absence of viologens guests.
ogens exhibit Do values larger than 5”10ꢀ6 cm2 sꢀ1. Howev-
er, in the presence of one to two equivalents of 1, all the
measured values are below 2”10ꢀ6 cm2 sꢀ1. The pronounced
decrease in the diffusivity of the viologens clearly reveals a
strong interaction with a much larger molecule. In fact, the
Do values measured in the presence of two equivalents 1
were identical for all viologen guests (1.6”10ꢀ6 cm2 sꢀ1) and
very close to the Do value measured for the host (1.6”
10ꢀ6 cm2 sꢀ1) in the absence of any viologen guests. In the
absence of NaCl, the diffusion coefficient of octaacid 1 is
about 2.2”10ꢀ6 cm2 sꢀ1, which corresponds to its monomeric,
unassociated form. Moderate concentrations of NaCl (<
0.1m) tend to foster host dimerization and Do values in the
range 1.5–1.7”10ꢀ6 cm2 sꢀ1, and higher NaCl concentrations
give rise to larger aggregates with lower diffusion coeffi-
cients. Therefore, the data in Table 1 provide evidence for
the strong association of all three viologen guests with mo-
lecular capsules formed by two molecules of the octaacid
host (12).
Are the viologens incorporated inside the 12 molecular
capsules? The lack of effect of any of the viologen guests on
the proton NMR signals of the host affords a strong argu-
ment against encapsulation in this case. Encapsulation
would also be inconsistent with the minimal effects on the
aromatic proton signals of the viologens upon addition of
host 1. However, the host-induced shifts on the aliphatic
protons of the viologenꢀs N-substituents indicate that the
ends of the guests do interact with the host. To gain a better
understanding of this interaction, we measured the diffusion
coefficients of BV2+ at variable concentrations while keep-
ing constant the concentration of host 1 (Figure 4). When
[BV2+]<[1] the guestꢀs Do values are similar to those of the
capsule (1.6”10ꢀ6 cm2 sꢀ1). When both concentrations are
equal, the diffusion coefficient reaches a minimum value,
consistent with the formation of a tight external complex be-
tween the dimeric host capsule (12) and the two viologen
guests. Once the guestꢀs concentration exceeds the host con-
centration, the Do values start to increase. This behavior
strongly suggests that each 12 molecular capsule has two ex-
ternal sites for strong association of the viologen guests.
Once those sites are occupied, the excess guests cannot asso-
ciate with the capsule so effectively and tend to diffuse
freely in the solution, leading to an overall increase of the
average Do value recorded for the viologen guest. It seems
reasonable to postulate that the two external binding sites
for the dicationic viologen guests are the poles of the dimer-
ic capsules, each one containing four carboxylates. A violo-
gen can thus latch onto this polar anionic site via coulombic
attractive interactions. From our experimental data we
cannot be more specific about the type of interaction be-
tween the viologen guest and the anionic end of the capsule.
It is not clear whether the viologen inserts one of its posi-
tively charged ends in between the four anionic “feet” of
the capsule (with the other end still exposed to the solution)
or the electrostatic interaction involves both ends simultane-
ously. The first possibility would require rapid exchange be-
tween the two ends, since no difference is observed between
them in the proton NMR spectra. It is also possible to pos-
tulate that the viologen is undergoing fast exchange among
several docking conformations, as long as both ends of the
guest are responsible for most of the interaction. Although
the spectroscopic data support a polar location for the violo-
gen guests, this is still the first example of an empty host
that dimerizes to form a well-defined capsule. In this regard,
viologen interactions with the equatorial carboxylates might
also facilitate capsule formation.
The identification of two external binding sites for cation-
ic guests on the structure of dimeric 12 capsules opens the
possibility of filling the capsule up with a ferrocene guest
while maintaining the interaction with the two viologen
guests on the “polar caps”. We have indeed verified this
possibility using NMR spectroscopy. Figure 5 shows the
spectrum of 1 in the presence of one equivalent of EV2+ (a
12 capsule interacting with two ethyl viologen guests). Addi-
tion of ferrocene results in its encapsulation inside 12, as evi-
denced by the signal at d=2.16 ppm, with no significant
changes for the viologen protons. The Fc-induced shifts on
the host protons are all consistent with its incorporation
inside the capsule, leading to an unusual self-assembled
Chem. Eur. J. 2008, 14, 4704 – 4710
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4707

A. E. Kaifer et al.
1 eliminates the distortion observed on the anodic peak of
the second reduction process, indicating that 1 solubilizes
the two-electron reduced form MV. The presence of 1 also
increases the potential difference between the two reduction
processes, suggesting the differential stabilization of the
cation radical form.
Similar voltammetric results were recorded with EV2+
and BV2+ (see the Supporting Information), although exten-
sive precipitation effects were observed with the latter in
the absence of host 1. The voltammetric data are consistent
with the strong association of every viologen guest with the
dimeric 12 molecular capsules. The one-electron reduced
form of the viologen is differentially stabilized by the capsu-
les, which is reminiscent of similar stabilization observed in
solutions containing anionic micelles.[19,20] The combined bal-
ance of electrostatic/hydrophobic forces seems to favor the
interaction between the viologen cation radical and the
group of four carboxylate-terminated feet on the capsule
polar cap. However, the solubilization of the two-electron
reduced viologen form in the aqueous medium is not so
easy to explain, because any electrostatic interactions with
the anionic molecular capsule are lost due to the neutral
character of this viologen oxidation state. While one can
argue that the reduced viologen (MV) may efficiently move
towards the interior of the capsule, this is not consistent
with the relatively fast oxidation observed for the MV to
Figure 5. 1H NMR spectra (400 MHz, 50 mm NaCl
+ 10 mm borate
buffer pD=8.9 in D2O) of a) 2 mm EV2+ and 2 mm host 1 and b) 1 mm
Fc, 2 mm EV2+ and 2 mm host 1. The EV2+ protons are denoted by a
circle and the encapsulated Fc protons by an asterisk.
structure composed of five individual components: two host
molecules, an encapsulated hydrophobic guest (Fc) and two
dicationic guests externally bound to the polar ends of the
structure.
Viologens undergo two sequential, reversible, one-elec-
tron reductions. In aqueous solution, the first reduction pro-
cess (V2+/V+ ) is usually very fast, and the second reduction
C
(V+ /V), at more negative potentials, is often complicated
MV+ process in the reverse scan. It seems more reasonable
C
C
by precipitation of the neutral, uncharged form on the elec-
trode surface. These precipitation problems become more
pronounced and extend to the cation radical form as the hy-
drophobic character of the viologen increases. Figure 6
shows the contrast between the cyclic voltammetric respons-
es for MV2+ in the absence and in the presence of two
equivalents of host 1. In the absence of host, both reduction
processes are clearly observed, although the anodic peak for
the second redox couple is distorted due to the precipitation
of the fully reduced, uncharged MV form. In the presence
of two equivalents of 1, both reduction processes are also
observed, but the current levels for all the waves are sub-
stantially reduced, in agreement with the Do values obtained
in the PGSE NMR experiments. Notice that the presence of
to postulate that hydrophobic interactions are responsible
for maintaining some degree of interaction between MV and
the polar site.
Are the rates of heterogeneous electron transfer for the
viologen compounds affected by the association with the 12
molecular capsule? To answer this question we recorded the
scan rate dependence of the current–potential curves in the
range ꢀ0.3 to ꢀ0.9 V versus Ag/AgCl, thus focusing on the
first redox couple (V2+/V+ ). These experiments were done
C
with solutions containing 1 mm viologen guest and 2 mm
host 1, in which, according to our Do measurements, the spe-
cies present in solution is the V2+·12 complex. The corre-
sponding cyclic voltammograms are shown in the Supporting
Information and the standard rate constants (ko) for hetero-
geneous electron transfer were determined by the Nicholson
method,[21] using the diffusion coefficients obtained in the
PGSE NMR experiments. We obtained ko values of (7.2ꢁ
0.4)”10ꢀ3, (3.5ꢁ0.5)”10ꢀ3 and (1.7ꢁ0.2)”10ꢀ3 cmsꢀ1, for
MV2+·12, EV2+·12, and BV2+·12, respectively. For compari-
son purposes we also ran the same experiments for MV2+ in
the absence of host 1 and found ko to be too fast to be deter-
mined by this method, as expected (ko >0.8 cmsꢀ1). Our
data demonstrate that association of the viologen guest with
the dimeric molecular capsule results in a pronounced de-
crease of the ko value for all three viologens surveyed here.
We have encountered similar attenuations of electrochemi-
cal kinetic rates in other systems in which a redox-active
center is covalently or noncovalently encapsulated by a par-
tially or fully surrounding organic sheath.[22] The strong asso-
ciation of the viologen guest with a large organic structure
Figure 6. Cyclic voltammetric response on glassy carbon (0.071 cm2) of
0.5 mm MV2+ in the absence (continuous line) and in the presence (dis-
continuous line) of 1.0 mm 1. The aqueous solution also contains 40 mm
NaCl and 10 mm sodium borate (pH 8.9). Scan rate=0.1 Vsꢀ1
.
4708
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4704 – 4710

Encapsulated Redox Centers
FULL PAPER
leads to a substantially increased, average distance of maxi-
mum approach between the redox center and the electrode
surface. This is equivalent to saying that the outer Helm-
holtz plane (OHP) moves away from the electrode surface,
leading to slower electrochemical kinetics.
identical within experimental error margins (Table 1). We
note that the measured ko values decrease in the same order
as the increasing hydrophobic character of the viologen
guest. However, we must continue our investigation of this
system and obtain additional data before we can draw
meaningful relationships between the hydrophobic character
of the viologen guest and the electrochemical kinetic rates
for its reduction in these assemblies.
As part of our work with dendrimers containing a single
viologen unit, we have recently shown that their electro-
chemical kinetics in voltammetric experiments remains re-
versible as the size of the dendrimer increases from first to
third generation of growth.[23,24] Dendrimers containing a
single viologen residue are unique in this regard, as most
dendrimers of similar structure containing a single redox
unit show measurably slower electrochemical kinetics
(quasi-reversible to irreversible) by the second or third gen-
eration of growth.[25] We have attributed the unique behav-
ior of viologen-containing dendrimers to orientation effects
near the electrode-solution interface as the positive charge
of the viologen unit is likely to be attracted by the negative
charge density present on the electrode surface at the poten-
tials required for viologen reduction.[24] This is particularly
important in dendrimer structures in which the only charge
resides on the viologen unit. In contrast to these viologen-
containing dendrimers, the external complex formed be-
tween a viologen guest and a 12 dimeric capsule exhibits a
total of 16 negative charges (from the two octaacid mole-
cules) and two positive charges (from the viologen guest).
Under these conditions the predominant charge on the com-
plex is anionic and orientation effects associated with cou-
lombic interactions between the viologen and the negatively
charged electrode surface are not expected to play a signifi-
cant role. Therefore, the peripheral attachment of the violo-
gen to one of the anionic poles of the 12 molecular capsule
has the anticipated effect of increasing the average distance
of maximum approach between the viologen center and the
electrode surface, leading to a measurable attenuation of
the electrochemical kinetics rates.
Conclusion
The experimental results presented in this work clearly indi-
cate that 1) ferrocene undergoes encapsulation inside dimer-
ic 12 capsules, and 2) simple viologen guests bind strongly to
the tetra-anionic polar ends of the capsule. Ferrocene has a
pronounced hydrophobic character and undergoes encapsu-
lation as a result of interactions with the hydrophobic cavity
of the capsule, in excellent agreement with previous reports
on the encapsulation of other hydrophobic compounds by
12.[9,13,14] MV2+, EV2+, and BV2+ are organic dications with
variable degrees of hydrophobic character. Compared to Fc,
their greater hydrophilicity and +2 charge clearly factor
against encapsulation and lead to their strong interaction
with the polar ends of the capsule, where favorable electro-
static forces develop between each viologen dication and
the four carboxylates positioned on the capsuleꢀs end. It is
interesting that the two types of binding interactions with
the capsule are mutually independent, which makes possible
the formation of unique supramolecular aggregates
(Scheme 1) in which the hydrophobic guest (Fc) and one or
two cationic guests (V2+) are distributed among different,
albeit spatially close, locations. Among other issues, we are
currently starting an investigation on the possible advanta-
geous use of these supramolecular assemblies to carry out
photo-induced electron transfer reactions.
The case reported here allows us for the first time to com-
pare the ko values for three similar viologens (MV2+, EV2+
,
and BV2+) strongly associated to the surface of the same
molecular capsule system (Scheme 1). At this point we do
not fully understand the reasons behind the observed gradu-
al attenuation in the ko value as we move from MV2+·12 to
EV2+·12 to BV2+·12. The relative mass differences between
these species are minimal and their Do values are essentially
Experimental Section
Octaacid 1 was prepared as previously reported.[9] Methyl viologen chlo-
ride and ferrocene were commercially available. Ethyl viologen and butyl
viologen were prepared by exhaustive alkylation of 4,4’-bipyridine with
the corresponding bromoalkane. The voltammetric experiments were re-
corded on a BAS 100B/W electrochemical workstation, using a single-
compartment cell fitted with a glassy carbon working electrode, a Pt aux-
iliary electrode and a Ag/AgCl reference electrode. The working elec-
trode was polished with 0.05 mm alumina/water slurry on a felt surface.
The solutions were purged with purified nitrogen gas before the experi-
ments and kept under an inert nitrogen atmosphere throughout. Diffu-
sion coefficients were measured using PGSE NMR techniques as report-
ed before.[26]
Acknowledgements
Scheme 1. Pictorial representation of the structure of the supramolecular
assembly formed by the trapped Fc guest, the dimeric molecular capsule
12 and the two externally bound viologen guests. The negative charges on
the capsuleꢀs equator are not shown for simplicity.
The authors are grateful to the National Science Foundation (to A.E.K.,
CHE-0600795, to B.C.G., CHE-0414413) for the generous support of this
work.
Chem. Eur. J. 2008, 14, 4704 – 4710
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4709

A. E. Kaifer et al.
[13] C. L. D. Gibb, B. C. Gibb, Chem. Commun. 2007, 1635–1637.
[14] C. L. D. Gibb, B. C. Gibb, J. Am. Chem. Soc. 2006, 128, 16498–
16499.
[1] A. Lützen, Angew. Chem. 2005, 117, 1022–1025; Angew. Chem. Int.
Ed. 2005, 44, 1000–1002.
[2] D. J. Cram, J. M. Cram Container Molecules and Their Guests: The
Royal Society of Chemistry, 1994.
[15] Y. Cohen, L. Avram, L. Frish, Angew. Chem. 2005, 117, 524–560;
Angew. Chem. Int. Ed. 2005, 44, 520–554.
[3] For recent reviews, see: a) M. R. Johnston, M. J. Latter, Supramol.
Chem. 2005, 17, 595–607; b) J. L. Rebek, Jr., Angew. Chem. 2005,
117, 2104–2115; Angew. Chem. Int. Ed. 2005, 44, 2068–2078;
c) D. M. Vriezema, M. C. Aragones, J. A. A. W. Elemans, J. J. L. M.
Cornelissen, A. E. Rowan, R. J. M. Nolte, Chem. Rev. 2005, 105,
1445–1489; d) J. L. Atwood, L. J. Barbour, A. Jerga, Perspect. Su-
pramol. Chem. 2003, 7, 153–175.
[4] A. J. Parola, F. Pina, E. Ferreira, M. Maestri, V. Balzani, J. Am.
Chem. Soc. 1996, 118, 11610–11616.
[5] Z. S. Romanova, K. Deshayes, P. Piotrowiak, J. Am. Chem. Soc.
2001, 123, 11029–11036.
[16] S. Mendoza, P. D. Davidov, A. E. Kaifer, Chem. Eur. J. 1998, 4, 864–
870.
[17] C. M. Cardona, A. E. Kaifer, J. Am. Chem. Soc. 1998, 120, 4023–
4024.
[18] C. M. Cardona, T. D. McCarley, A. E. Kaifer J. Org. Chem. 2000, 65,
1857–1864.
[19] A. E. Kaifer, A. J. Bard. J. Phys. Chem. 1985, 89, 4876–4880.
[20] P. A. Quintela, A. E. Kaifer, Langmuir 1987, 3, 769–773.
[21] R. S. Nicholson, Anal. Chem. 1965, 37, 1351–1355.
[22] C. M. Cardona, S. Mendoza, A. E. Kaifer, Chem. Soc. Rev. 2000, 29,
37–42.
[6] I. Philip, A. E. Kaifer, J. Am. Chem. Soc. 2002, 124, 12678–12679.
[7] I. Philip, A. E. Kaifer, J. Org. Chem. 2005, 70, 1558–1564.
[8] K. Moon, A. E. Kaifer, J. Am. Chem. Soc. 2004, 126, 15016–15017.
[9] C. L. D. Gibb, B. C. Gibb, J. Am. Chem. Soc. 2004, 126, 11408–
11409.
[23] R. Toba, J. M. Quintela, C. Peinador, E. Roman, A. E. Kaifer,
Chem. Commun. 2001, 857–858.
[24] W. Ong, J. Grindstaff, D. Sobransingh, R. Toba, J. M. Quintela, C.
Peinador, A. E. Kaifer, J. Am. Chem. Soc. 2005, 127, 3353–3361.
[25] C. S. Cameron, C. B. Gorman, Adv. Funct. Mater. 2002, 12, 17–20.
[26] H. Sun, W. Chen, A. E. Kaifer, Organometallics 2006, 25, 1828–
1830.
[10] L. S. Kaanumalle, C. L. D. Gibb, B. C. Gibb, V. Ramamurthy, J. Am.
Chem. Soc. 2004, 126, 14366–14367.
[11] L. S. Kaanumalle, C. L. D. Gibb, B. C. Gibb, V. Ramamurthy, J. Am.
Chem. Soc. 2005, 127, 3674–3675.
[12] A. Natarajan, L. S. Kaanumalle, S. Jockush, C. L. D. Gibb, B. C.
Gibb, N. J. Turro, V. Ramamurthy, J. Am. Chem. Soc. 2007, 129,
4132–4133.
Received: December 11, 2007
Revised: February 26, 2008
Published online: March 28, 2008
4710
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4704 – 4710