guest binding, redox, and molecular transport properties of...
TRANSCRIPT
Guest Binding, Redox, and Molecular Transport Properties of Supramolecular
Coordination Assemblies
by
Bryan Erik Tiedemann
B.S. (California Institute of Technology) 2002
A dissertation submitted in partial satisfaction of the
requirements for the degree of
Doctor of Philosophy
in
Chemistry
in the
GRADUATE DIVISION
of the
UNIVERSITY OF CALIFORNIA, BERKELEY
Committee in Charge:
Professor Kenneth N. Raymond, Chair Professor Jeffrey R. Long Professor Alexander Katz
Spring 2007
Guest Binding, Redox, and Molecular Transport Properties of
Supramolecular Coordination Assemblies
Copyright © 2007
Bryan Erik Tiedemann
- 1 -
Abstract
Guest Binding, Redox, and Molecular Transport Properties of Supramolecular
Coordination Assemblies
by
Bryan Erik Tiedemann
Doctor of Philosophy in Chemistry
University of California, Berkeley
Professor Kenneth N. Raymond, Chair
The rich host-guest chemistry of the M4L6 tetrahedral cluster developed by
Raymond and coworkers has been studied for the last ten years, yet this dissertation will
argue that its full capabilities have yet to be realized. Several different physical
techniques have been used to explore the properties of the M4L6 supramolecular cluster,
providing a different perspective. A brief summary of the known properties of M4L6 is
given in Chapter 1, as well a description of how electrochemical techniques can apply to
supramolecular systems.
Chapter 2 is concerned with molecular transport, using diffusion NMR methods to
measure diffusion coefficients. Exterior binding of alkylammonium cations was clearly
observed from changes in the diffusion coefficient, confirming that strong ion pairing
interactions persist in aqueous solution. The effect of different alkali cations on the
diffusion rate is also discussed.
- 2 -
The thermodynamics of host-guest interactions are investigated in Chapter 3 using
isothermal titration calorimetry to directly measure reaction heats. Large heats are
observed when [Ga4L6]12- is titrated with R4N
+ in aqueous solution (R = Me, Et, Pr), and
the resulting thermodynamic parameters for the overall processes are reported.
A novel partial guest encapsulation mode is described in Chapter 4. Cationic
ruthenium sandwich complexes with pendant linear chains are encapsulated by [Ga4L6]12-
with the result that part of the long chains protruding out of the host to the exterior.
Chains terminated by an anionic group are permanently extruded through the three-fold
symmetric aperture in a triangular face of the host, confirming the non-dissociative guest
exchange mechanism described by Davis et al. The methyl terminus of neutral chains
can reside either inside or outside the cluster, and such a chain with six carbons rapidly
extends and retracts at room temperature, moving the methyl group in and out of the host
cavity. This fluxional structure exemplifies the dynamic second-order Jahn-Teller effect.
Chapters 5 and 6 describe electrochemical experiments with M4L6. Chapter 5
explores whether redox-active cations can be reduced while encapsulated within a redox-
silent host; the answer is no for the systems examined. Chapter 6 investigates redox-
active vanadium complexes, including the [V4L6]8- cluster with a redox-silent Et4N
+
guest. The results demonstrate that, while the vertices are redox-active in the M4L6
cluster, there is no interaction between them.
- i -
For my grandfather
Albert E. Morjig
- ii -
ACKNOWLEDGEMENTS
My family has always been extremely supportive of my interests, whatever they
turned out to be. My fiancée Anishya Mathai has provided critical support and
inspiration during the most stressful times of my thesis preparation. I thank my parents,
Al and Nancy Tiedemann, and my grandparents, Florence and Al Morjig, for their
continued love and support. My brother Sanders and my sisters Whitney and Lori have
always been dear friends to me, always in the mood for a good time together.
Ken Raymond has been a wonderful teacher and a great friend to me throughout
my graduate career. I’ve always enjoyed going up to the white board to muddle through
some theoretical analysis explaining exactly why my hypothesis was wrong. He
provided me opportunities to do world-class research, and showed me the joys of sailing.
I will truly miss working with him.
My colleagues in the Raymond Group have been wonderful coworkers, all very
talented, and all willing to help one another. Didier Pomeranc taught me synthetic
chemistry, particularly air-sensitive manipulations. Anna Davis, Rob Yeh, Dorothea
Fiedler, and Dennis Leung were my mentors in the supramolecular group. Dorothea
deserves a lot of credit for help with the partial guest encapsulation project. Georg
Seeber was a good friend and was great to work with while writing our review article.
Mike Pluth is without a doubt the most talented chemistry gradate student I’ve ever met,
and our regular discussions were always stimulating. Shannon Biros is not only an
extremely talented supramolecular chemist, but a wonderful person who made 505 a
- iii -
much warmer place. Jeff Mugridge will be my successor for the ITC experiments, and it
was a pleasure to work with him for the short period we overlapped.
Rudi Nunlist deserves full credit for making UC Berkeley’s NMR facility a world
class laboratory. He was always willing to teach me all sorts of crazy things about NMR
and Linux, and without his help setting up the diffusion NMR experiments on the AV-
500, Chapter 2 would have been quite a bit slimmer. Herman van Halbeek was
invaluable for helping Mike and I set up the initial diffusion NMR experiments on the
AVB-400, and without his help Chapter 2 might not even exist.
Our collaborators in Sicily, Prof. Giuseppe Arena, Dr. Carmelo Sgarlata, and
Valeria Zito at the Università di Catania, were wonderful colleagues to work with.
During the two weeks while I was there, we not only achieved some fantastic scientific
feats, but they also showed me the true meaning of hospitality. They are all highly
skilled analytical chemists.
The redox-active guest electrochemical studies was also a collaborative effort
between our group and Burak Ulgut and Prof. Héctor D. Abruña at Cornell University. I
worked in their lab for a week, where I picked up all sorts of great tips and tricks for
electrochemical experiments. Much of the vanadium electrochemistry work was done
with a lot of advice from Prof. Marcin Majda here at UC Berkeley. The propanethiol
treatment of the Hg electrode was his idea, and it worked great.
My undergraduate advisor, Harry Gray, allowed my passion for chemistry to
flourish, and always let me visit him in his office to ask stupid questions. Randy
Villahermosa was Harry’s graduate student who mentored me for three years, I
appreciate all his patience while we worked together.
- iv -
TABLE OF CONTENTS
CHAPTER 1 – Introduction 1
Overview 1
Supramolecular Hosts 2
Supramolecular Electrochemistry 5
Summary 10
References 10
CHAPTER 2 – Diffusion of a Supramolecular Cluster: Ion Pairing Effects in
Aqueous Solution
14
Introduction 14
Results and Discussion 19
Summary 27
Experimental 28
References 33
CHAPTER 3 – Thermodynamics of Guest Binding: Calorimetry Experiments 36
Introduction 36
Results and Discussion 41
Summary 51
Experimental 52
References 55
- v -
CHAPTER 4 – Partial Guest Encapsulation Modes 58
Introduction 58
Zwitterionic Guests 59
Monocations with Pendant Alkanes 69
Summary 78
Experimental 78
References 91
CHAPTER 5 – Electrochemical Properties of Monocations Encapsulated by
an M4L6 Host
93
Introduction 93
Results and Discussion 94
Summary 111
Experimental 112
References 119
CHAPTER 6 – Redox-Active Vanadium Complexes 122
Introduction 122
Synthesis and Characterization of Vanadium(IV) Complexes 124
Electrochemical Studies of Vanadium Complexes 128
Summary 150
Experimental 151
References 160
- vi -
APPENDIX 1 – Supplementary Spectra for Chapter 4 163
APPENDIX 2 – Born Solvation Energies and Guest Encapsulation 184
APPENDIX 3 – Crystallographic Data for K4[V2LH
3]⋅⋅⋅⋅6.7 DMF⋅⋅⋅⋅Et2O⋅⋅⋅⋅0.3 H2O 188
- 1 -
CHAPTER 1
Introduction
Overview
This dissertation focuses on various aspects of the M4L6 supramolecular clusters
developed by Raymond and coworkers, where M = GaIII or VIV and H4L = 1,5-bis(2,3-
dihydroxybenzamido)naphthalene (Figure 1.1).1 These chiral, self-assembled tetrahedral
complexes have proven to be extremely versatile hosts, encapsulating a wide variety of
guests within their hydrophobic cavities.3, 4 While encapsulated, guests can undergo
reactions – both stoichiometric and catalytic – with significant rate enhancement and
improved product selectivity in some cases.5-7 This may lead to practical applications,
warranting a detailed study of the guest binding, redox, and molecular transport
Figure 1.1. (Left) Schematic structure of the M4L6 tetrahedral cluster illustrating the structure of L4- and its coordination to the metal ion vertices. (Right) Diagram of the host-guest complex [Et4N ⊂ Ga4L6]
11-, based on the X-ray structure coordinates, with Et4N
+ guest shown in blue.
- 2 -
properties of the M4L6 supramolecular coordination assemblies, which are the subjects of
this thesis.
Supramolecular Hosts
Supramolecular chemistry takes advantage of self-assembly to prepare large,
discrete structures from relatively simple subunits. Using metal-ligand interactions,
π-stacking, and/or hydrogen bonds to link the various subunits together, many elegant
examples of supramolecular assemblies have been designed and synthesized which have
been the subject of some excellent review articles.8-10 Some of these assemblies feature a
binding site for guest encapsulation through noncovalent interactions. Examples of the
utilization of host-guest properties of such clusters to modify guest reactivity include
work done by Fujita,11, 12 Reek,13 and Rebek.14
Raymond and coworkers used symmetry and geometrical constraints to rationally
design the components of the M4L6 cluster shown in Figure 1.1.1, 9, 15 When the rigid
two-fold symmetric H4L ligands are combined with a hard octahedral metal ion
(M = AlIII, FeIII, GaIII, InIII, TiIV, VIV, GeIV, SnIV) in the presence of a suitable guest, the
only product which forms is the M4L6 tetrahedral cluster.3, 10 Four pseudo-octahedral
metal ions are located at the vertices of the tetrahedron, linked by six bis-bidentate
catecholamide ligands spanning the edges. Although the cluster is constructed from
achiral subunits, each tris-bidentate metal center exhibits helical chirality, and coupling
by the rigid ligands forces all vertices to be homochiral; thus, the assembly has pure
rotation point group T symmetry. When prepared with trivalent metal ions such as
gallium(III), the anionic cluster has an overall -12 charge, making K12[Ga4L6] very water
- 3 -
soluble. The interior of the tetrahedron is surrounded by six naphthalene rings from the
ligands, creating a hydrophobic pocket that sharply contrasts the polar exterior
environment surrounding the solvated host. This 250 – 500 Å3 expandable cavity binds
lipophilic guests, preferably monocations such as tetraethylammonium (Et4N+) and
cobaltocenium (CoCp2+).
The encapsulation of a solvated molecule G by the M4L6 host (abbreviated as H)
to form the host-guest complex [G ⊂ H] (abbreviated as HG), with the guest located
inside the host cavity, can be expressed by the following equilibrium:
H + G = HG Kb =[HG]
[H][G]Kb =
[HG]
[H][G]
When comparing guest binding affinities, saying that A is a “stronger” or “better” guest
than B is commonly used to mean Kb(A) > Kb(B). This jargon can be convenient, and
will be used periodically in this dissertation.
Two different types of binding interactions are possible for lipophilic cations with
the M4L6 cluster: encapsulation (discussed above), and exterior binding interactions to
form an ion pair. These two processes are interrelated, and guest exchange is thought to
proceed through an ion pair intermediate.7, 16 The X-ray crystal structure of
K5[Et4N]7[Fe4L6] shows one Et4N+ cation inside the host cavity, and the other six Et4N
+
cations closely associated with aromatic rings in the ligands.1, 3 This exterior binding
persists in aqueous solution, as evident from upfield shifts of exterior Et4N+ 1H NMR
resonances in the presence of [Ga4L6]12-. Exterior ion pairing is driven not only by
coulombic attraction with the anionic cluster, but also favorable cation-π interactions with
the aromatic rings, as well as van der Waals forces. Guests with π systems can also
participate in favorable internal π-π interactions with the naphthalene rings.11 Exterior
- 4 -
ion pairing is explored in Chapter 2 using diffusion NMR measurements, and guest
binding thermodynamics are discussed in Chapter 3, using data from isothermal titration
calorimetry (ITC) experiments.
Guest Exchange Mechanism
Interior and exterior guest species can dynamically exchange in solution, and the
kinetics of both self- and cross-exchange reactions have been measured for a variety of
guests.6, 16, 17 During a cross-exchange reaction, a strongly bound guest will displace a
weaker guest to form a more stable host-guest complex. Davis and Raymond observed
that [Ga4L6]12- and [Ti4L6]
8- hosts facilitate guest exchange at essentially the same rate,
thus demonstrating the guest exchange mechanism does not require a ligand dissociation
step.2 [Titanium(IV) catechol complexes are very inert to ligand exchange, while
gallium(III) catechol complexes are labile, so much slower guest exchange rates would
be observed for [Ti4L6]8- if ligand dissociation was required.2, 18 ]
Small openings exist in the triangular faces of the host, and concerted cluster
distortion acts to enlarge these gaps for guests to pass through (Figure 1.2). The
Figure 1.2. The egress of Et4N+ (green molecule) from the [Ga4L6]
12- cavity proceeds through the three-fold symmetric apertures of an intact cluster, without breaking any metal-ligand bonds. The reaction energy diagram shown at left reveals the transition state features the guest partially extruded through a dilated aperture (point C). The increase in energy at the far right of the reaction coordinate is an artifact of the calculation being performed in the gas phase. Reprinted with permission from Davis and Raymond.2
- 5 -
activation barrier height for the guest exchange process is determined by how wide the
aperture must be to accommodate the steric bulk of the guest in the transition state. Since
the distortion resembles a vibrational breathing mode, one can approximate the potential
energy profile with Hooke's law, proportional to the square of the radial distortion from
its equilibrium size. Small guests such as Me4N+ require small distortions, and fast self-
exchange rates are observed from line broadening analysis.19 Larger guests such as Pr4N+
require larger distortions to pass through the aperture, and slower self-exchange rates are
observed (for Pr4N+, k298 = 1.4 s-1).16 However, for asymmetric guests, where one
dimension is longer than the others, the aperture only has to open enough to allow the
smallest cross section to pass through. This is exemplified by Me2Pr2N+, whose self-
exchange rate of k298 = 4.4 s-1 is much faster than Pr4N+ because it can enter methyl-
group first, with the long propyl chains trailing behind.16 In contrast, ground state effects
of the smaller yet strong binding Et4N+ guest make its self-exchange rate very slow
(k298 = 0.009 s-1).16
Chapter 4 explores the consequences of the nondissociative guest extrusion
mechanism by synthesizing two sets of RuII sandwich complexes that have pendant alkyl
chains with different terminal groups. The mechanism is supported by the observation
of a stable cluster in D2O with part of the chain protruding through one of the three-fold
symmetric openings in the triangular faces of the tetrahedron.
Supramolecular Electrochemistry
The two primary topics of supramolecular electrochemistry are systems with
multiple redox sites and electrochemical switching. Chapter 5 explores the use of
- 6 -
[Ga4L6]12- as a redox-silent host for switchable binding of redox-active metallocene
guests, and Chapter 6 describes the electrochemical behavior of M4L6 with electroactive
vanadium ions at the vertices, as well as the analogous dinuclear vanadium helicate and a
mononuclear vanadium tris(catecholamide) model complex. The primary technique used
to study the electrochemical behavior of these systems is cyclic voltammetry. A brief
introduction to some important electrochemical concepts related to supramolecular
systems is given here.
Electrochemical Switching
Electrochemical switching is the most important application of electrochemistry
in the field of supramolecular chemistry.20 In a redox-switchable system, guest binding
events can be precisely controlled simply by applying voltage to a solution, changing the
oxidation state of a redox-active site on either the host or the guest. The two oxidation
states exhibit different binding equilibrium constants, preferably one much larger than the
other. For practical purposes, the redox-active site must exhibit reversible electron
transfer kinetics at the electrode; otherwise, kinetic limitations would render the switch
too slow to be useful.
The coupling of the electrochemical and guest binding equilibria can be described
by a simple square scheme. Following the description by Kaifer et al., such a scheme is
shown in Figure 1.3.20 For this example, the guest G is electroactive and is switched
from its low to high binding state by reduction, forming more stable complexes with
redox-silent host H when it is reduced (G) than where it is oxidized (G+). Therefore, the
binding equilibrium constant Kred is larger than Kox. The reduction potential of the host-
- 7 -
guest complex E±
b0E±
b0 is shifted from that for the free guest E±
f0E±
f0 according to the following
expression:
E±
b0 = E±
f0 +
RT
FlnKred
Kox
E±
b0 = E±
f0 +
RT
FlnKred
Kox (1.1)
where F is the Faraday constant, R is the molar gas constant, and T is the temperature in
Kelvin. A similar expression can be derived for a redox-active host binding a redox-
silent guest, with the reduction potentials E±
b0 and E±
f0 corresponding to the host with and
without bound guest, respectively.
With Kred > Kox, the redox potential of the host-guest complex will be
anodically shifted, i.e. oxidation will require a more positive potential due to the
stabilization imparted by guest binding. If both Kox and Kred are large (¸ 104M¡1),
diffusion of the empty host species is not a relevant factor, and if Kred=Kox ¸ 103, two
separate voltammetric waves can be observed.21 In this scenario, the magnitude of the
ratio Kred=Kox is defined as the binding enhancement, and can be estimated by the
difference in the half-wave potentials:
Kred
Kox
= e¡F (E±
f0¡E±
b0)=RT
(1.2)
Figure 1.3. Coupled electrochemical and binding equilibria for an electroactive guest G binding with a redox-silent host H.
- 8 -
In contrast, if Kox is small, the free guest is the species being reduced, not the
host-guest complex, and after reduction to the strongly bound G the complexation
process is limited by the diffusion of available empty host. In this case, a single wave
will be observed, with a shift in the observed half wave potential due to guest binding.20
Numerous electrochemical studies of redox-active guests encapsulated by a
redox-silent host have been reported, with ferrocene (Fc) derivatives being the most
popular choices for electroactive guest studies. Osella et al. studied the influence of Fc
encapsulation on its redox properties by β-cyclodextrin (β-CD), a well-characterized
neutral host.22 They observed a single FeIII/II wave in the cyclic voltammogram for the
host-guest complex, anodically shifted from the free Fc+/0 potential due to the high
binding constant, and hence stability, of the neutral host-guest complex. Diminished
currents were also observed upon encapsulation, due to the lower diffusion coefficient of
the encapsulated Fc relative to that for the free guest. The redox behavior of Fc in the
presence of β-CD served as the basis for models describing how host-guest interactions
affect electrochemical behavior, including an insightful analytical discussion by Mendoza
et al.23
In another study, Fujita and coworkers reported cyclic voltammetry measurements
with a PdII-linked supramolecular nanocage capable of encapsulating up to four
molecules of Fc.24 A single wave was observed with the host-guest complex pre-
synthesized with four encapsulated Fc molecules, and the observed 73 mV anodic shift
was attributed to a destabilization of the Fc+ oxidation product by the cationic
environment of the cage. When one equivalent of a water-soluble, neutral ferrocene
- 9 -
derivative was combined with the host, a single anodically shifted wave was observed,
with diminished currents due to the slower diffusion of the host-guest complex.
Multiple identical redox sites
If two or more redox-active sites are present in a molecule, the voltammetric
behavior can be strongly influenced by electronic coupling between the individual sites.
Extended π systems allow efficient long range communication, and π-d overlap for redox
active metal centers makes this interaction even stronger. The ultimate example of metal-
metal communication is the Creutz-Taube ion (Figure 1.4), where a single electron is
completely delocalized between both metals across the pyrazine bridge.25 The formal
oxidation states of both ruthenium atoms is 2.5, and this mixed valence state is denoted
[2,3] (for RuII and RuIII). Two reversible one-electron waves are observed: one for the
[2,2]/[2,3] couple at +0.37 V vs. NHE, and the other for the [2,3]/[3,3] couple at +0.76 V
vs. NHE.26 This separation of 0.39 V is a consequence of the strong communication,
since reduction of one site leads to an increase in electron density at the other site, and the
next oxidation becomes more difficult.
If the individual sites are not linked by a continuous extended π system, and are
separated far enough to make electrostatic interactions negligible, then little to no
communication between redox sites will exist. In such a case, the free energy of
successive reductions differ purely for entropic reasons. Statistical analysis dictates that
Figure 1.4. Creutz-Taube ion, where both ruthenium atoms share the same formal oxidation state of 2.5.
- 10 -
the kth successive formal reduction potential differs from the first according to27
E±
k ¡ E±
1 = ¡2RT
Fln kE±
k ¡ E±
1 = ¡2RT
Fln k
(1.3)
For two noninteracting sites, E±
2 ¡ E±
1 = ¡0:036E±
2 ¡ E±
1 = ¡0:036 V, which is extremely difficult to
measure experimentally. This is discussed in further detail in Chapter 6.
Summary
This dissertation uses several different approaches in order to understand the
guest binding, redox, and molecular transport properties of supramolecular coordination
assemblies. Both synthetic and analytical methods are used to investigate this common
theme, which is of direct interest for future industrial applications.
References
1. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
2. Davis, A. V.; Raymond, K. N., “The Big Squeeze: Guest Exchange in an M4L6 Supramolecular Host.” J. Am. Chem. Soc. 2005, 127, 7912-7919.
3. Caulder, D. L.; Brückner, C.; Powers, R. E.; König, S.; Parac, T. N.; Leary, J. A.; Raymond, K. N., “Design, Formation, and Properties of Tetrahedral M4L4 and M4L6 Supramolecular Clusters.” J. Am. Chem. Soc. 2001, 123, 8923-8938.
4. a) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Enantioselective Guest Binding and Dynamic Resolution of Cationic Ruthenium Complexes by a Chiral Metal-Ligand Assembly.” J. Am. Chem. Soc. 2004, 126, 3674-3675; b) Fiedler, D.; Pagliero, D.; Brumaghim, J. L.; Bergman, R. G.; Raymond, K. N., “Encapsulation of Cationic Ruthenium Complexes into a Chiral Self-Assembled Cage.” Inorg. Chem. 2004, 43, 846-848; c) Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am. Chem. Soc. 1998, 120, 8003-8004; d) Parac, T. N.; Scherer, M.; Raymond, K. N., “Host within a Host: Encapsulation of Crown
- 11 -
Ethers into Ga4L6 Supramolecular Host.” Angew. Chem. Int. Ed. 2000, 39, 1239-1242; e) Tiedemann, B. E. F.; Raymond, K. N., “Dangling Arms: A Tetrahedral Supramolecular Host with Partially Encapsulated Guests.” Angew. Chem. Int. Ed.
2006, 45, 83-86.
5. a) Fiedler, D.; Bergman, R. G.; Raymond, K. N., “Supramolecular catalysis of a unimolecular transformation: Aza-Cope rearrangement within a self-assembled host.” Angew. Chem. Int. Ed. 2004, 43, 6748-6751; b) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Selective Molecular Recognition, C-H Bond Activation, and Catalysis in Nanoscale Reaction Vessels.” Acc. Chem. Res. 2005, 38, 351-360; c) Leung, D. H.; Fiedler, D.; Bergman, R. G.; Raymond, K. N., “Selective C-H Bond Activation by a Supramolecular Host-Guest Assembly.” Angew. Chem. Int. Ed. 2004, 43, 963-966.
6. Fiedler, D.; van Halbeek, H.; Bergman, R. G.; Raymond, K. N., “Supramolecular Catalysis of Unimolecular Rearrangements: Substrate Scope and Mechanistic Insights.” J. Am. Chem. Soc. 2006, 128, 10240-10252.
7. Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Scope and Mechanism of the C-H Bond Activation Reactivity within a Supramolecular Host by an Iridium Guest: A Stepwise Ion Pair Guest Dissociation Mechanism.” J. Am. Chem. Soc. 2006, 126, 9781-9797.
8. a) Lawrence, D. S.; Jiang, T.; Levett, M., “Self-Assembling Supramolecular Complexes.” Chem. Rev. 1995, 95, 2229-2260; b) Leininger, S.; Olenyuk, B.; Stang, P. J., “Self-Assembly of Discrete Cyclic Nanostructures Mediated by Transition Metals.” Chem. Rev. 2000, 100, 853-908; c) Conn, M. M.; Rebek, J., Jr., “Self-assembling capsules.” Chem. Rev. 1997, 97, 1647-1668; d) Hamilton, T. D.; MacGillvray, L. R., “Enclosed Chiral Environments from Self-Assembled Metal-Organic Polyhedra.” Cryst. Growth Des. 2004, 4, 419-430; e) Manteos-Timoneda, M. A.; Crego-Calama, M.; Reinhoudt, D. N., “Supramolecular chirality of self-assembled systems in solution.” Chem. Soc. Rev. 2004, 33, 363-372; f) Johnson, D. W.; Raymond, K. N., “The Role of Guest Molecules in the Self-Assembly of Metal-Ligand Clusters.” Supramolecular Chem. 2001, 13, 639-659; g) Davis, A. V.; Yeh, R. M.; Raymond, K. N., “Supramolecular Assembly Dynamics.” Proc. Nat. Acad. Sci.
USA 2002, 99, 4793-4796.
9. a) Caulder, D. L.; Raymond, K. N., “Supermolecules by Design.” Acc. Chem. Res.
1999, 32, 975-982; b) Caulder, D. L.; Raymond, K. N., “The Rational Design of High Symmetry Coordination Clusters.” J. Chem. Soc., Dalton Trans. 1999, 8, 1185-2000.
10. a) Yeh, R. M.; Davis, A. V.; Raymond, K. N., “Supramolecular Systems: Self-Assembly.” In Comprehensive Coordination Chemistry II, Fujita, M.; Powell, A.; Creutz, A., Eds. Amsterdam, 2003; Vol. 7, pp 327-355; b) Seeber, G.; Tiedemann, B. E. F.; Raymond, K. N., “Supramolecular Chirality in Coordination Chemistry.” In Top. Curr. Chem., Reinhoudt, D. N.; Crego-Calama, M., Eds. Springer: Berlin, 2006; Vol. 265, pp 147-183.
- 12 -
11. Fujita, M.; Tominaga, M.; Hori, A.; Therrien, B., “Coordination Assemblies from a Pd(II)-Cornered Square Complex.” Acc. Chem. Res. 2005, 38, 369-378.
12. Yoshizawa, M.; Takeyama, Y.; Kusukawa, T.; Fujita, M., “Cavity-directed, highly stereoselective [2+2] photodimerization of olefins within self-assembled coordination cages.” Angew. Chem. Int. Ed. 2002, 41, 1347-1349.
13. a) Slagt, V. F.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Reek, J. N. H., “Encapsulation of Transition Metal Catalysts by Ligand-Template Directed Assembly.” J. Am. Chem. Soc. 2004, 126, 1526-1536; b) Slagt, V. F.; Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen, P. W. N. M., “Assembly of Encapsulated Transition Metal Catalysts.” Angew. Chem. Int. Ed. 2001, 40, 4271-4274; c) Slagt, V. F.; van Leeuwen, P. W. N. M.; Reek, J. N. H., “Multicomponent Porphyrin Assemblies as Functional Bidentate Phosphite Ligands for Regioselective Rhodium-Catalyzed Hydroformylation.” Angew. Chem. Int. Ed. 2003, 42, 5619-5623.
14. a) Hof, F.; Craig, S. L.; Nuckolls, C.; Rebek, J., Jr., “Molecular Encapsulation.” Angew. Chem. Int. Ed. 2002, 41, 1488-1508; b) Kang, J. M.; Rebek, J., Jr., “Acceleration of a Diels−Alder reaction by a self-assembled molecular capsule.” Nature 1997, 385, 50-52; c) Kang, J. M.; Santamaria, J.; Hilmersson, G.; Rebek, J., Jr., “Diels-Alder Reactions through Reversible Encapsulation.” J. Am. Chem. Soc.
1998, 120, 3650-3656; d) Kang, J. M.; Santamaria, J.; Hilmersson, G.; Rebek, J., Jr., “Self-Assembled Molecular Capsule Catalyzes a Diels-Alder Reaction.” J. Am.
Chem. Soc. 1998, 120, 7389-7390.
15. Beissel, T.; Powers, R. E.; Raymond, K. N., “Symmetry-Based Metal Complex Cluster Formation.” Angew. Chem. Int. Ed. Engl 1996, 35, 1084-1086.
16. Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.; Raymond, K. N., “Guest Exchange Dynamics in an M4L6 Tetrahedral Host.” J. Am. Chem. Soc. 2006, 128, 1324-1333.
17. Pluth, M. D.; Bergman, R. G.; Raymond, K. N., “Acid Catalysis in Basic Solution: A Supramolecular Host Promotes Orthoformate Hydrolysis.” Science 2007, 316, 85-88.
18. Lincoln, S. A.; Merbach, A. E., Adv. Inorg. Chem. 1995, 42, 1-88.
19. Caulder, D. L. The Rational Design of High Symmetry Coordination Clusters. University of California, Berkeley, CA, 1998.
20. Kaifer, A. E.; Gómez-Kaifer, M., Supramolecular Electrochemistry. Wiley-VCH: Weinheim, 1999.
21. Kaifer, A. E.; Mendoza, S., “Redox-switchable Receptors.” In Molecular
Recognition: Receptors for Cationic Guests, Gokel, G. W., Ed. Pergamon: Tarrytown, NY, 1996; Vol. 1, pp 701-732.
- 13 -
22. Osella, D.; Carretta, A.; Nervi, C.; Ravera, M.; Gobetto, R., “Inclusion Complexes of Ferrocenes and β-Cyclodextrins. Critical Appraisal of the Electrochemical Evaluation of Formation Constants.” Organometallics 2000, 19, 2791-2797.
23. Mendoza, S.; Castaño, E.; Meas, Y.; Godínez, L. A.; Kaifer, A. E., “Analysis of the Voltammetric Response of Electroactive Guests in the Presence of Non-Electroactive Hosts at Moderate Concentrations.” Electroanalysis 2004, 16, 1469-1477.
24. Sun, W.-Y.; Kusukawa, T.; Fujita, M., “Electrochemically Driven Clathration/Declathration of Ferrocene and Its Derivatives by a Nanometer-Sized Coordination Cage.” J. Am. Chem. Soc. 2002, 124, 11570-11571.
25. Creutz, C.; Taube, H., “A Direct Approach to Measuring the Franck-Condon Barrier to Electron Transfer between Metal Ions.” J. Am. Chem. Soc. 1969, 91, 3988-3989.
26. Ward, M. D., “Metal-Metal Interactions in Binuclear Complexes Exhibiting Mixed Valency, Molecular Wires and Switches.” Chem. Soc. Rev. 1995, 121-134.
27. Bard, A. J.; Faulkner, L. R., Electrochemical Methods: Fundamentals and
Applications. 2nd ed.; John Wiley & Sons: Hoboken, 2001.
- 14 -
CHAPTER 2
Diffusion of a supramolecular cluster: Ion pairing effects in aqueous solution
Introduction
In studies of guest exchange in the anionic [Ga4L6]12- host supramolecular cluster
it was found that ion pair formation between the cationic guest and the anionic [Ga4L6]12-
host is an important part of the guest exchange process. This was surmised from indirect
kinetic observations, sometimes involving a large number of experiments fit to
mechanistic models.1, 2 Measurements of complex diffusion offer means to explore ion
pairing more directly, since ion pair formation increases the size of the diffusing species,
which in turn decreases the rate of diffusion. Up to the late 1980’s, an entire day was
usually required to obtain a single diffusion coefficient measurement, with the sample
isolated from vibrations and under precise temperature control throughout the course of
the experiment. Thanks to major advancements in NMR methods and instrumentation
over the past two decades, particularly with the inclusion of z-gradient capabilities into
standard probes, diffusion coefficients can be measured in a single 40 minute experiment
with good accuracy (± 2%).
This chapter explores how diffusion NMR measurements can be used to
investigate exterior cation interactions with the [Ga4L6]12- anion by addressing two
important questions. First, will exterior binding of alkylammonium cations (Et4N+ =
tetraethylammonium, Pr4N = tetrapropylammonium) lead to observable changes in rates
of diffusion of the cluster or the exterior R4N+ ion? Second, how do different alkali
- 15 -
cations influence the diffusion of [Et4N⊂Ga4L6]11-? The first question assesses whether
or not diffusion NMR can actually detect ion pair formation, while the second question is
concerned with the basic mass transport characteristics for the cluster. Both questions are
addressed using quantitative measurements of diffusion coefficients in D2O solutions.
Background
Mass transport in solution occurs by diffusion, migration, and convection. In a
quiescent solution with no external electric field gradient, diffusion is the only means for
solutes to move from a region with high concentration to a region with low concentration.
Fick’s laws are differential equations describing the molar flux of a substance and its
concentration as a function of time and position. Following the description of Bard and
Faulkner,3 consider the case of linear (one-dimensional) diffusion. The flux of species i
at a position x and time t is Ji(x; t)Ji(x; t), with units of mol s-1 cm-2, and represents the number
of moles of i that pass point x per second per cm2 of area normal to the axis of diffusion
(Figure 2.1). According to Fick’s first law, the flux is proportional to the concentration
gradient:
Ji(x; t) = ¡Di
µ
@ci(x; t)
@x
¶
Ji(x; t) = ¡Di
µ
@ci(x; t)
@x
¶
(2.1)
x x0
Ji(x0 , t )
ci ci(x, t )
Figure 2.1. Schematic of one dimensional diffusion, where the decreasing concentration profile ci(x,t) results in a net flux of species i in the positive x direction.
- 16 -
where ci(x; t)ci(x; t) is the concentration of species i at position x and time t, and Di is the
diffusion coefficient of species i, with units of cm2 s-1. The negative sign arises because
the direction of flux is towards decreasing concentration. Fick’s second law describes the
change in concentration of i with time:
@ci(x; t)
@t= Di
µ
@2ci(x; t)
@x2
¶
@ci(x; t)
@t= Di
µ
@2ci(x; t)
@x2
¶
(2.2)
For a given initial concentration profile, the diffusion coefficient is the only factor
which can lead to different rates of diffusion for different chemical species. A higher
diffusion coefficient leads to larger flux at any given time, as well as a faster response to
changes in the concentration profile; this is experimentally observed as a faster rate of
diffusion.
The diffusion coefficient is an empirical quantity which depends on the solution
conditions. Due to the complex nature of solvent-solute interactions, there is no single
analytical model which reliably and accurately predicts diffusion coefficients, particularly
for ions in aqueous solutions. With that said, there are relatively simple models that are
very good at qualitatively predicting how the diffusion coefficient will change when a
specific parameter is varied, such as the Stokes-Einstein relation. By modeling diffusion
as a hydrodynamic process where a sphere moves through a viscous medium, Stokes and
Einstein obtained the following expression:4
D =kBT
6¼´rh
D =kBT
6¼´rh (2.3)
where kB is Boltzmann’s constant, T is the absolute temperature, η is the dynamic
viscosity of the solution, and rh is the “hydrodynamic radius” of the diffusing species.
- 17 -
The hydrodynamic radius is in fact defined by the Stokes-Einstein relationship,
and is an empirical quantity. For a large, neutral molecule, the hydrodynamic radius is
very similar to its crystallographic radius; but this is not so for ionic species. Solvent
molecules organize themselves around ions, and these solvation spheres tend to diffuse
with the ion, increasing the hydrodynamic radius. This is especially true for small ions in
aqueous solution: the crystallographic radius of Li+ in LiCl is 0.76 Å,5 but its
hydrodynamic radius in aqueous solution is 3.0 Å.6 The discrepancy between
crystallographic and hydrodynamic radii becomes much less apparent for larger ions,
allowing experimental trends in the diffusion coefficient to be correlated to trends in the
actual size of the diffusing species, measured from crystallographic data.
Diffusion NMR
A number of NMR techniques are available to monitor diffusion and have been
described in a recent review article.7 Pulsed gradient spin-echo (PGSE) NMR methods
have recently attracted increasing interest, since this technique allows diffusion
coefficients to be measured with high signal to noise ratios and have been successfully
used with 7Li and 31P as well as 1H NMR.8 In a PGSE experiment, gradient pulses are
used to spatially label nuclei along the z axis of the NMR tube, and after a delay ∆ to
allow for diffusion, a second gradient pulse is applied before data acquisition. This pulse
sequence is repeated while systematically changing the strength of the gradient. The
attenuation of resonance intensity I with applied gradient strength G is related to the
diffusion coefficient according to:
- 18 -
I = I0 exp
·
¡D°2 G2 ±2
µ
¢¡ ±
3
¶¸
I = I0 exp
·
¡D°2 G2 ±2
µ
¢¡ ±
3
¶¸
(2.4)
where I0 is the signal intensity measured with G = 0, D is the diffusion coefficient, γ is
the gyromagnetic ratio of the resonating nucleus, δ is the gradient pulse width in seconds,
and ∆ is the “diffusion time,” or time between the first and second gradient pulses. The
gradient strength G is typically expressed as G = f ¢GmaxG = f ¢Gmax, where Gmax is the maximum
gradient strength in gauss cm-1, and the fraction f serves as the dependent variable. In
practice, f is expressed as a percentage, and is limited to a range from 2% to 95% to
ensure linear behavior of the gradient amplifier in the probe.9 The observed data can be
fit to Equation 2.4 using nonlinear regression to obtain the value for D (Figure 2.2).
All values of D obtained from this fit are computed using the empirical value
Gmax, which should be accurately calibrated for the particular probe used for
measurements. This was done indirectly by measuring the diffusion decay curves for
dextrose and β-cyclodextrin, whose diffusion coefficients are known to a high degree of
Figure 2.2. Example of a diffusion decay curve with proper choices for δ and ∆. This particular curve corresponds to Li11[Et4N⊂Ga4L6] + 0.1 M LiCl in D2O at 300 K, with δ = 7 ms and ∆ = 90 ms.
0 20 40 60 80 1000
0.2
0.4
0.6
0.8
1
% Gradient Strength
No
rma
lize
d In
teg
ral
Observed Simulated Fit
- 19 -
accuracy. These reported values are Ddex = 6.728 x 10-6 cm2 s-1 and
Dβ-CD = 3.224 cm2 s-1, measured in aqueous solution by monitoring Rayleigh optical
interference fringes.10 The measured value of Gmax = 30.12 ± 0.1 gauss cm-1 was the
same for both probes used for this work, despite their use at different magnetic field
strengths (400 MHz vs. 500 MHz).
Results and Discussion
Ion pairing of alkylammonium cations
In collaboration with Michael D. Pluth, the ion pairing of Et4N+ and Pr4N
+ with
the [Ga4L6]12- host in D2O was investigated using PGSE 1H NMR to measure diffusion
coefficients. When a solution of K12[Ga4L6] is titrated with R4N+ (R = Et, Pr), the
observed diffusion coefficient of the host (DH) decreases with increasing [R4N+] (Figure
2.3) . This indicates that after the first equivalent is encapsulated, the excess R4N+ binds
to the exterior of the host, increasing its overall size and decreasing its diffusion
coefficient. Saturation occurs because the host has a limited number of exterior binding
0 5 10 152.0
2.1
2.2
2.3
DH ,
1
0-6 c
m2 s
-1
# equiv. Et4N
+
0 5 10 15
1.9
2.0
2.1
2.2
DH ,
10
-6 c
m2 s
-1
# equiv. Pr4N
+
a) Et4N+ b) Pr4N
+
Figure 2.3. Ion pairing interactions cause the diffusion coefficient of [Ga4L6]12- in D2O to decrease with
addition of a) Et4NCl with 0.1 M K2CO3 buffer (pD 11) in the host solution, and b) Pr4NBr.
- 20 -
sites available – as many as six sites are suggested from the crystal structure of
K5(Et4N)6[Et4N⊂Fe4L6], with each ligand providing a π-basic naphthalene ring surface
for binding lipophilic cations.11, 12 It is important to note that different solution
conditions were used for the Et4N+ and Pr4N
+ titration experiments in both Figure 2.3 and
Figure 2.4: samples with Et4N+ were prepared with 0.1 M K2CO3 buffer (pD 11.0), while
no buffer was used to prepare Pr4N+ samples.1 Therefore, it is difficult to assess the
relative binding affinities by comparing the data in Figure 2.3.
The weak exterior binding interactions occur in parallel with the much stronger
guest encapsulation equilibrium, causing most ion pair interactions to involve hosts with
an encapsulated guest. Identical diffusion coefficients are obtained from separate fits of
the diffusion decay curves for the signals from the host and encapsulated guest. This
confirms the host and guest diffuse together as a single host-guest complex.
Furthermore, the diffusion coefficient of [Ga4L6]12- measured in the absence of guest is
approximately the same as that measured with one equivalent of R4N+ present. This
implies that [Ga4L6]12- and [R4N⊂Ga4L6]
11- have similar hydrodynamic radii, and similar
solvation shell sizes. We see two possible explanations for this: either size differences
between solvation shells for z = -11 and z = -12 ions are fairly small due to the large size
of the clusters themselves, or the “empty” host contains a deuteron plus solvent (e.g.
D3O+, D5O2
+, etc.) leading to an actual charge of z = -11. These two situations cannot be
distinguished from diffusion data. Kinetic evidence suggests that neutral species are
encapsulated and later protonated inside of the [Ga4L6]12- assembly, and the effective
1 The experiments were carried out separately by two different people, explaining the different conditions.
- 21 -
shift in guest protonation constants has been seen to be no bigger than 4 log units, thus
favoring the first of the two hypotheses.13
Ion pair formation also decreases the observed exterior R4N+ diffusion coefficient,
since the bound alkylammonium cations must diffuse with the much larger [Ga4L6]12-
cluster. Recall that interior and exterior protons are readily distinguished by the large
upfield shifts observed upon encapsulation.11, 14 When a solution of K12[Ga4L6] with two
equivalents of R4N+ (R = Et, Pr) is titrated with KCl the observed diffusion coefficient of
the exterior alkylammonium cation rapidly increases with increasing [KCl] (Figure 2.4).
One R4N+ cation is encapsulated by the cluster, leaving the other R4N
+ cation on the
exterior to form an ion pair. Since the exterior R4N+ cation rapidly equilibrates between
the ion paired and free states on the NMR timescale, its observed diffusion coefficient is
the population average of the two states. In the absence of salt, the exterior R4N+
diffusion is only slightly faster than the host-guest complex, particularly for Et4N+,
consistent with tight ion pairing. The added salt disrupts the ion pairing of R4N+ to the
0.0 0.5 1.0 1.52.5
3.0
3.5
4.0
4.5
D,
10
-6 c
m2 s
-1
[KCl], M
0.0 0.5 1.0 1.5
3.0
3.5
4.0
4.5
D,
10
-6 c
m2 s
-1
[KCl], M
2 eq. Et4N+ 2 eq. Pr4N
+
Figure 2.4. Diffusion coefficient of exterior R4N+ as a function of KCl concentration. Addition of KCl to
K12[Ga4L6] in D2O with a) 2 equivalents of Et4NCl + 0.1 M K2CO3 buffer (pD 11) or b) 2 equivalents of Pr4NBr results in higher diffusion coefficients observed for the exterior alkylammonium cation, because the added salt disrupts ion pairing to the host exterior.
a) b)
- 22 -
host exterior, causing the diffusion coefficient of the exterior R4N+ cation to increase
dramatically.
Much higher diffusion coefficients are observed for the free alkylammonium
cations in the absence of host (Table 2.1). For Et4N+ with 1 M KCl, interaction with
[Et4N⊂Ga4L6]11- reduced the observed diffusion coefficient of the exterior cation to less
than half the value observed in the absence of host, despite the 100-fold excess of KCl;
similar effects were observed with Pr4N+, but to a somewhat lesser extent. Thus the
favorable exterior interactions between R4N+ and [Ga4L6]
12- cannot be solely attributed to
simple coulombic attractions, since K+ will exhibit similar, if not higher, coulombic
attractive forces to the anionic host. If R4N+ binding were caused by coulombic
attraction alone, a large excess of KCl would eliminate any interactions with the anionic
host, and the observed R4N+ diffusion rate would be equal to that observed in the absence
of host. This is clearly not the case, and additional attractive forces must be involved,
such as cation-π binding and/or van der Waals interactions.
Table 2.1. Diffusion coefficients of R4N+ in D2O with 1 M KCl measured in the absence and presence of
[Ga4L6]12- at 27 °C.
D – no [Ga4L6]12- (cm2 s-1) D – with [Ga4L6]
12- (cm2 s-1)
Et4NCl 9.9 x 10-6 4.3 x 10-6
Pr4NBr 7.6 x 10-6 4.1 x 10-6
Alkali cation interactions
In a study by Leung et al. involving reactive monocationic half-sandwich iridium
guests, a stepwise mechanism of guest dissociation through an ion paired intermediate
was described in detail based on kinetic data.2 In an elegant set of experiments, it was
- 23 -
demonstrated that K+ is much more effective than Na+ in disrupting the exterior binding
interaction, suggesting K+ interacts much more strongly with the ion pair adduct than
Na+. Whether or not K+ exhibits a particularly favorable interaction with the host itself,
however, remained unclear. Diffusion measurements may shed some light on whether
the anionic cluster shows different affinities for the various alkali cations.
Five different alkali salts of A11[Et4N⊂Ga4L6] (A+ = Li+, Na+, K+, Rb+, Cs+) were
prepared to investigate how different alkali cations interact with the cluster. The
crystallographic radii of the alkali cations increase with atomic number, i.e. Li+ < Na+ <
K+ < Rb+ < Cs+. However, their ionic mobilities observed in highly dilute aqueous
solutions indicate their hydrodynamic radii decrease with increasing atomic number, i.e.
for rh, Rb+ ≤ Cs+ < K+ < Na+ < Li+.6 The slower diffusion for smaller cations, particularly
for Li+, reflects their large solvation spheres. The solvation shells surrounding small ions
such as Li+ are tightly held to the central ion due to its high charge to surface area ratio.
Since the single positive charge for large ions is spread over a much larger area, Rb+ and
Cs+ exhibit relatively thin shells of loosely associated solvent molecules. Also, different
binding characteristics are expected for the different alkali cations based on hard-soft
acid-base theory, since Li+ is small and hard while Cs+ is large and soft due to
polarizability differences.15
The diffusion coefficients of [Et4N⊂Ga4L6]11- were measured for all five alkali
salts in D2O solutions, and the results are summarized in Figure 2.5. Changing the alkali
cation has a subtle but significant effect on the cluster’s diffusion coefficient. Two
different sets of samples were prepared: one set containing A11[Et4N⊂Ga4L6] (A+ = Li+,
Na+, K+, Rb+, Cs) in pure D2O (“non-salt” samples) and the other set containing the same
- 24 -
compounds but with 0.1 M ACl present (ACl is the chloride salt of the corresponding A+
cation). The data in Figure 2.5 were measured using two different spectrometers, with
much higher 1H sensitivity for the AV-500 compared to the AVB-400 due to different
probe configurations. All measurements with the “non-salt” samples were made with the
AVB-400 spectrometer, whereas the AV-500 instrument was primarily used for
measurements on samples containing 0.1 M ACl.2 The data for the “non-salt” samples –
measured on the AVB-400 spectrometer – suffered low poor signal to noise ratios at high
gradient strengths, suggesting their diffusion coefficients in Figure 2.5 may not be as
accurate as those shown for the samples with 0.1 M ACl.
2 Only one of the three values for the Li+ and Na+ systems was measured on the AV-400 out of all the samples with added salt, and the resulting diffusion coefficients were identical to one of the other two diffusion coefficients calculated from AV-500 to two decimal places (i.e. within 0.01 x 10-6 cm2 s-1).
Figure 2.5. Diffusion coefficients at 300 K measured for A11[Et4N⊂Ga4L6] (A+ = Li+, Na+, K+, Rb+, Cs+)
in pure D2O (blue bars) and in D2O with 0.1 M ACl present (purple bars). When average values of repeated experiments are reported, the number of experiments averaged together is listed above the corresponding bar, with the measured standard deviations used for the error bar. Otherwise, error bars for single measurements are calculated from the standard deviation of the measured gradient strength.
2.00
2.20
2.40
2.60
2.80
Li Na K Rb Cs
DH
G, 1
0-6
cm
2 s
-1
No added salt
With 0.1 M ACl
3
3
3
2
- 25 -
Comparing the measurements for the samples with 0.1 M ACl, the diffusion
coefficients for the Na+, K+ and Cs+ systems are all equal within experimental error, with
DK = 2.37(3) x 10-6 cm2 s-1. The Li+ system shows the slowest diffusion, with
DLi = 2.24(3) x 10-6 cm2 s-1, which is about 5% lower than the Na+, K+ and Cs+ values.
The Rb+ system exhibits the fastest diffusion, with DRb about 8% greater than the Na+, K+
and Cs+ diffusion coefficients, and about 14% higher than that for Li+. The non-salt data
also show the host-guest complex diffusion is fastest with Rb+ counterions, and the
slowest with Li+ counter ions.
Clearly, the alkali counterions must play an active role during the cluster diffusion
process. If A+ formed ion pairs with the [Et4N⊂Ga4L6]11- anion, and if different alkali
cations bind with different affinities to cause the diffusion coefficients to differ, then a
significant cation concentration dependence would be observed. However, the presence
of 0.1 M ACl – twenty times higher than the 5 mM cluster concentration – has very little
effect on the diffusion coefficient for Li+, Na+, and Cs+, implying their ion pairing can be
neglected (or the binding sites are already saturated). For Rb+, addition of salt leads to a
4% increase in the diffusion coefficient, and for K+ a 5% increase is observed with KCl.
These small changes may be artifacts from the poor signal to noise of the no-salt data
measured on the AVB-400.
The slow diffusion observed with Li+ counterions is consistent with its much
lower ionic mobility.6 To maintain charge neutrality, counterions must co-diffuse with
the solvated host-guest anion, and the observed diffusion coefficient of [Et4N⊂Ga4L6]11-
will depend on the alkali cation’s mobility. For infinitely dilute solutions, the diffusion
coefficient can be estimated from the individual ionic mobilities at infinite dilution:
- 26 -
D± =2u±+u
±
¡RT
zF (u±+ + u±¡)
D± =2u±+u
±
¡RT
zF (u±+ + u±¡)
(2.5)
where u±+u±
+ and u±¡u±¡
are the respective mobilities (cm2 V-1 s-1) of the cation and anion at
infinite dilution, z is the charge number (z = 11 for [Et4N⊂Ga4L6]11-), and F is the
Faraday constant.6, 16 At infinite dilution, the cluster’s mobility u±HGu±HG is unaffected by the
nature of the cation, and we can calculate the ratio of diffusion coefficients for two
different cations with mobilites u±1u±
1 and u±2u±
2:
D±
2
D±
1
=
µ
u±2u±1
¶
u±1 + u±HG
u±2 + u±HG
D±
2
D±
1
=
µ
u±2u±1
¶
u±1 + u±HG
u±2 + u±HG (2.6)
Due to its large size and charge, the cluster’s mobility u±HGu±HG should be less than
that for any alkali cation. If u±2 > u±1 > u±HGu±2 > u±1 > u±HG, then D±
2 > D±
1D±
2 > D±
1 according to Equation 2.6.
Since Li+ has the lowest mobility of any alkali cation, Equation 2.6 predicts the diffusion
of [Et4N⊂Ga4L6]11- will be slowest for the lithium salt, consistent with the experimental
observations in Figure 2.5.
Rb+ has the highest ionic mobility in aqueous solution, about 6% faster than K+,
and much higher than Na+ or Li+, so this may explain why the host diffuses fastest with
Rb+ cations. It remains unclear why the cluster’s diffusion coefficient observed with Rb+
is relatively high compared to Cs. The mobilities of Rb+ and Cs+ are very similar, so
Equation 3.6 cannot explain why the diffusion coefficient with Rb+ is about 8% larger.
Ion pairing effects are unlikely – addition of KCl or RbCl to leads to a 5% and 4%
increase in the observed diffusion coefficients, respectively, with essentially no change
for Cs+. For Et4N+ and Pr4N
+, exterior binding led to slower diffusion, since the
relatively large cations increased the hydrodynamic radius of the ion pair. Although the
radii of K+ and Rb+ are much smaller than the alkylammonium ions, ion pairing still
- 27 -
cannot account for the faster diffusion observed in the presence of salt. The
crystallographic radii for the two cations are r(K+) = 1.38 Å and r(Rb+) = 1.52 Å,5 so if
the two cations did bind to the anionic cluster, the size differences predict faster diffusion
for K+, but the larger Rb+ shows faster diffusion, inconsistent with ion pair formation.
The solvated radii of the two ions are approximately equal, so binding of the solvated
cations cannot account for the difference either.3
The diffusion coefficients observed with Na+ and K+ are about the same,
suggesting the anionic cluster does not show any particular preference for binding one
cation over the other. In contrast, Leung et al. observed K+ was much more effective
than Na+ at disrupting the binding of an iridium half-sandwich cation to the exterior of
[Ga4L6]12-. While the reason for this cation dependence remains unclear, the diffusion
data suggest that ion pairing to the host itself is not significant for alkali cations. Perhaps
the stronger solvation of Na+ compared to K+ due to its smaller radius is an important
factor. The crystallographic radii for the chloride salts are rNa = 1.02 Å for Na+ and
rK = 1.38 Å for K+, and from the Born model K+ desolvation requires about 35% less
energy than Na+ in the same solution. Future studies are needed to test this hypothesis.
Summary
Diffusion NMR can be a very powerful tool to observe relatively weak
interactions with ease. Ion pairing interactions can be observed by monitoring changes in
the diffusion coefficients for the host and the guest, rather than relying on detailed kinetic
studies involving complex reaction mechanisms to indirectly measure exterior binding
3 Ternary species with bridging ion-paired cations such as HG11-⋅⋅⋅A+⋅⋅⋅HG11- will be strongly disfavored by the large coulombic repulsive forces between the two -11 anions.
- 28 -
equilibria. The exterior binding of Et4N+ and Pr4N
+, originally deduced from the Et4N+
concentration dependence observed for the NMR chemical shifts, has been firmly
established by diffusion NMR methods as an important secondary binding interaction.
Furthermore, diffusion measurements provided valuable insight on how different alkali
cations interact with the anionic host. Since diffusion coefficients are related to
molecular size, diffusion NMR can be a valuable analytical method to study
supramolecular systems.
Experimental
General considerations
Unless noted otherwise, reagents were obtained from commercial suppliers and
used without further purification. Standard Schlenk techniques were used for reactions
carried out under argon, and a glove box continuously purged with nitrogen was used to
manipulate and store air-sensitive solids. When necessary, solvents were degassed by at
least six pump/fill cycles while vigorously stirring, using argon for the fill step.
Methanolic alkali hydroxide stock solutions were standardized by Michael D. Pluth by
performing a titration with aqueous HCl using a colored pH indicator, taking the average
value from three titrations for each solution. Tetraethylammonium chloride (Et4NCl) was
recrystallized from absolute ethanol/ether and dried in vacuo over P2O5 at room
temperature for 12 hours, then dried in vacuo over molecular sieves at 60 °C for 18 hours
and stored under nitrogen. H4L (H4L = 1,5-bis(2,3-dihydroxybenzamido)naphthalene)
and K12[Ga4L6] were synthesized according to literature procedures.11, 17 Routine mass
spectrometry and elemental analysis was performed by the Mass Spectrometry
- 29 -
Laboratory and Microanalysis Facility in the College of Chemistry at the University of
California, Berkeley.
Synthetic procedures
Cs11[Et4N ⊂⊂⊂⊂ Ga4L6]·(Me2CO).12 In a 250 mL round-bottom Schlenk flask, 200 mg
(0.465 mmol) of H4L and 113.8 mg (0.310 mmol) of Ga(acac)3 (Aldrich) were combined,
to which 100 mL of methanol was added. To this opaque white mixture was added
0.79 mL (0.077 mmol) of methanolic Et4NCl (97 mM) via syringe, the reaction mixture
was degassed via several pump/fill cycles, and 4.2 mL (0.95 mmol) of methanolic CsOH
(0.226 M) was added via syringe, causing the color to change from white to yellow. The
reaction mixture was degassed again, and stirred under argon at room temperature for
18 hours. The pale yellow solid suspended in the methanolic reaction mixture was
filtered under a stream of nitrogen, washed with acetone (3 x 10 mL) and petroleum ether
(2 x 15 mL), and dried in vacuo overnight for 12 hours to yield 300 mg (86%) of yellow
powder. The 1H NMR spectrum was similar to that published for K11[Et4N⊂Ga4L6]. The
number of co-precipitated acetone molecules was determined from 1H NMR integration.
Rb11[Et4N⊂⊂⊂⊂Ga4L6]·(Me2CO)(Et2O).12 A procedure similar to the Cs+ salt synthesis was
used, with 200 mg (0.465 mmol) of H4L, 113.8 mg (0.310 mmol) of Ga(acac)3, 0.75 mL
(0.073 mmol) of methanolic Et4NCl (97 mM), and 2.4 mL (0.95 mmol) of methanolic
RbOH (0.397 M). After stirring at room temperature for 4 days, the reaction mixture was
cloudy, indicating some product was suspended. After reducing the volume to
approximately 50 mL using a vacuum pump to remove solvent, the reaction mixture was
- 30 -
filtered, leaving a small amount of yellow solid on the frit. Additional solvent was
removed from the clear yellow filtrate to reduce its volume to about 30 mL, generating a
small amount of yellow precipitate. Addition of 30 mL of degassed acetone caused this
solid to re-dissolve, affording a transparent yellow solution. Addition of 30 mL of Et2O
to this methanol/acetone solution led to precipitate formation after 10 min., and 30 L of
additional Et2O was added and allowed to equilibrate for an additional 15 min. The solid
was collected on the same frit used for the first filtration, washed with acetone/ether
(1 x 10 mL; some product dissolved in this washing), then pure Et2O (2 x 15 mL). The
solid was dried on the frit under a stream of nitrogen, yielding 220 mg (70%) of pale
yellow powder. The 1H NMR spectrum was similar to that published for
K11[Et4N⊂Ga4L6]. The number of co-precipitated solvent molecules was determined
from 1H NMR integration.
Li11[Et4N⊂⊂⊂⊂Ga4L6]·(Me2CO)5.12 A procedure similar to the Cs+ salt synthesis was used,
with 200 mg (0.465 mmol) of H4L, 113.8 mg (0.310 mmol) of Ga(acac)3, 0.75 mL
(0.073 mmol) of methanolic Et4NCl (97 mM), and 2.8 mL (0.95 mmol) of methanolic
LiOH (0.343 M). After stirring at room temperature for 7 days, solvent was removed
with a vacuum pump to reduce the reaction mixture volume to 60 mL. The slightly
turbid mixture was passed through 0.2 µm nylon syringe filter disks, and the volume of
the clear yellow filtrate was reduced to approximately 10 mL using a vacuum pump to
remove solvent. Degassed acetone (150 mL total) was gradually added via cannula while
slowly stirring to form a precipitate, which was collected on a frit under a stream of
nitrogen, washed with acetone (2 x 10 mL) and petroleum ether (3 x 20 mL), yielding
- 31 -
160 mg (62%) of yellow-green powder. This solid is hygroscopic, steadily gaining mass
while exposed to air. The 1H NMR spectrum was similar to that published for
K11[Et4N⊂Ga4L6]. The number of co-precipitated acetone molecules was determined
from 1H NMR integration.
Na11[Et4N⊂⊂⊂⊂Ga4L6]·(Me2CO)3.12 A procedure similar to the Li+ salt synthesis was used,
with 1.45 mL (0.93 mmol) of methanolic NaOH (0.640 M) as the base, stirring at room
temperature overnight. Yield: 180 mg (68%) of yellow powder. The 1H NMR spectrum
was similar to that published for K11[Et4N⊂Ga4L6]. The number of co-precipitated
acetones was determined from 1H NMR integration.
K11[Et4N⊂⊂⊂⊂Ga4L6]·(Me2CO)2.12 A procedure similar to the Li+ salt synthesis was used,
with 1.9 mL (0.95 mmol) of methanolic KOH (Aldrich, 0.5 M) as the base, stirring at
room temperature overnight. Yield: 100 mg (37%) of yellow powder. The 1H NMR
spectrum was similar to that published for K11[Et4N⊂Ga4L6]. The number of co-
precipitated acetone molecules was determined from 1H NMR integration.
Diffusion NMR experiments
PGSE diffusion 1H NMR measurements were performed on either a Bruker
AVB-400 spectrometer with a z-gradient broadband coil or a Bruker AV-500
spectrometer with a TBI-P probe with a z-gradient coil, using the ledbpgp2s pulse
program with diffusion time ∆ = 90-100 ms, bipolar gradient pulse duration δ = 7 ms
(2 x 3.5 ms), 8 scans per experiment, pre-pulse delay of 6 sec., and a linear gradient
- 32 -
strength ramp of 32 increments from 2% to 95%.9 The 90° RF pulse width was
calibrated for each sample, particularly with different ionic strengths. A constant
temperature of 300 K was maintained using an automated temperature controller,
allowing the sample temperature to equilibrate in the probe for at least 10 minutes before
starting each measurement. The integrated areas (normalized by the low-G integral
value) were averaged for resonances on the same molecule, and fit to the expected
exponential decay equation18 using nonlinear regression to evaluate the diffusion
coefficient. The regression weighting scheme was based on the observed standard
deviation for the averaged resonances. The probe gradient power was calibrated from a
fit of the diffusion decay curve of dextrose and β-cyclodextrin in D2O using literature
values for the diffusion coefficients.10
Sample preparation: Et4N+ titration. In a thin-walled NMR tube, 25 mg (7 µmol) of
K12[Ga4L6] was dissolved in 500 µL of D2O buffered to pD = 11.0 with 0.1 M K2CO3,
and a sealed capillary with ferrocene in CDCl3 was inserted in the tube for an internal
diffusion standard. For each titration point, a stock solution of 1.0 M Et4NCl in D2O was
added to the NMR tube in 7 µL increments, and mixed with the host solution by inverting
the tube several times before replacing the tube in the probe for the next measurement.
Actual number of equivalents of Et4N+ was determined from 1H NMR integrals.
Sample preparation: Pr4N+ titration. In a glass vial, 14 mg (4 µmol) of K12[Ga4L6]
was dissolved in 0.7 mL of D2O, filtered through a glass wool plug, and the filtrate was
collected in a thin-walled NMR tube. For each titration point, a stock solution of 0.2 M
- 33 -
Pr4NBr in D2O was added to the NMR tube in 20 µL increments in a similar manner
described for the Et4N+ titration.
Sample preparation: KCl titration with Et4N+. Using six thin-wall NMR tubes, 10 mg
(2.8 µmol) of K12[Ga4L6] was combined with different amounts of KCl in each tube
(0 mg, 2.16 mg, 3.44 mg, 9.15 mg, 18.06 mg, 33.75 mg) and each dissolved in 400 µL of
D2O buffered to pD = 11.0 with 0.1 M K2CO3. To each sample was added 100 µL of a
stock solution with 60 mM Et4NCl in the same D2O buffer for a final volume of 500 µL
in each tube.
Sample preparation: KCl titration with Pr4N+. Two separate stock solutions were
prepared in D2O: one with 62.1 mg (16.9 µmol) of K12[Ga4L6] and 9.0 mg (34 µmol) of
Pr4NBr dissolved in D2O to 2.00 mL for a stock solution with 8.5 mM host and 19 mM
guest; the other with 3.0 M KCl. Using seven medium-walled NMR tubes, 200 µL of
host/guest stock solution was added to each tube, and different amounts of KCl stock
solution was added (0 µL, 13.3 µL, 33.3 µL, 66.7 µL, 100 µL, 133 µL, 200 µL) and the
appropriate amount of pure D2O was added for a final volume of 400 µL per tube. Each
sample was degassed in its tube via three freeze-pump-thaw cycles, flame sealed under
vacuum, and allowed to equilibrate overnight before measurements.
References
1. a) Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.; Raymond, K. N., “Guest Exchange Dynamics in an M4L6 Tetrahedral Host.” J. Am. Chem. Soc. 2006, 128, 1324-1333; b) Fiedler, D.; Bergman, R. G.; Raymond, K. N., “Supramolecular catalysis of a unimolecular transformation: Aza-Cope rearrangement within a self-
- 34 -
assembled host.” Angew. Chem. Int. Ed. 2004, 43, 6748-6751; c) Fiedler, D.; van Halbeek, H.; Bergman, R. G.; Raymond, K. N., “Supramolecular Catalysis of Unimolecular Rearrangements: Substrate Scope and Mechanistic Insights.” J. Am.
Chem. Soc. 2006, 128, 10240-10252.
2. Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Scope and Mechanism of the C-H Bond Activation Reactivity within a Supramolecular Host by an Iridium Guest: A Stepwise Ion Pair Guest Dissociation Mechanism.” J. Am. Chem. Soc. 2006, 126, 9781-9797.
3. Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and
Applications, 2nd ed.; John Wiley & Sons: Hoboken, 2001; p 148-150.
4. Welty, J. R.; Wicks, C. E.; Wilson, R. E.; Rorrer, G. Fundamentals of Momentum,
Heat, and Mass Transfer, 4th ed.; John Wiley & Sons: New York, 2001.
5. Lide, D. R., Handbook of Chemistry and Physics. 81st ed.; CRC Press: Boca Raton, 2000; p 12.14-15.
6. Stern, K. H.; Amis, E. S., “Ionic Size.” Chem. Rev. 1959, 59, 1-64.
7. Johnson, C. S., Jr., “Diffusion ordered nuclear magnetic resonance spectroscopy: principles and applications.” Prog. NMR. Spectrosc. 1999, 34, 203-256.
8. a) Fernández, I.; Martínez-Viviente, E.; Breher, F.; Pregosin, P. S., “7Li, 31P, and 1H Pulsed Gradient Spin-Echo (PGSE) Diffusion NMR Spectroscopy and Ion Pairing: On the Temperature Dependence of the Ion Pairing in Li(CPh3), Fluorenyllithium, and Li[N(SiMe3)2] amongst Other Salts.” Chem. Eur. J. 2005, 11, 1495-1506; b) Pregosin, P. S.; Martínez-Viviente, E.; Anil Kumar, P. G., “Diffusion and NOE spectroscopy. Applications to problems related to coordination chemistry and homogeneous catalysis.” Dalton Trans. 2003, 4007-4014.
9. Kerssebaum, R., DOSY and Diffusion by NMR. In User Guide for XWinNMR 3.1/3.5
Version 1.03, Bruker BioSpin GmbH: Rheinstetten, Germany, 2002.
10. Longsworth, L. G., “Temperature Dependence of Diffusion in Aqueous Solutions.” J.
Phys. Chem 1954, 58, 770-773.
11. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
12. Caulder, D. L.; Brückner, C.; Powers, R. E.; König, S.; Parac, T. N.; Leary, J. A.; Raymond, K. N., “Design, Formation, and Properties of Tetrahedral M4L4 and M4L6 Supramolecular Clusters.” J. Am. Chem. Soc. 2001, 123, 8923-8938.
- 35 -
13. Pluth, M. D.; Bergman, R. G.; Raymond, K. N., “Acid Catalysis in Basic Solution: A Supramolecular Host Promotes Orthoformate Hydrolysis.” Science 2007, 316, 85-88.
14. Chatterjee, A. K.; Choi, T.-L.; Sanders, D. P.; Grubbs, R. H., “Catalyzed Alkene Metathesis.” J. Am. Chem. Soc. 2003, 125, 11360-11370.
15. Ahrland, S., Factors Contributing to (b)-Behaviour in Acceptors. In Structure and
Bonding, Jørgensen, C. K.; Neilands, J. B.; Nyhoum, R. S.; Reinen, D.; Williams, R. J. P., Eds. Springer-Verlag: Berlin, 1966; Vol. 1, pp 207-220.
16. Euken, A. Lehrbuch der chemischen Physik, Akademische Verlagsgesellschaft: Leipzig, 1949; Vol. 2, p 798.
17. Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am.
Chem. Soc. 1998, 120, 8003-8004.
18. Stejskal, E. O.; Tanner, J. E., “Spin Diffusion Measurements: Spin Echoes in the Presence of a Time-Dependent Field Gradient.” J. Chem. Phys. 1965, 42, 288-292.
- 36 -
CHAPTER 3
Thermodynamics of Guest Binding: Calorimetry Experiments
Introduction
The fact that guest encapsulation by the [Ga4L6]12- host can be highly favorable is
readily apparent from 1H NMR spectroscopy. When K12[Ga4L6] is combined with one
equivalent of Et4N+ in D2O, almost quantitative encapsulation is observed from the
integrated areas of the upfield-shifted CH2 and CH3 interior resonances, with very little
signal remaining for the exterior Et4N+ protons. The binding equilibrium constant for this
reaction, measured by 1H NMR integral ratios, is 1.96 x 104 M-1 in D2O at 25 °C.1 What
is the driving force behind this strong binding affinity? Previous van't Hoff studies using
NMR data indicated that guest encapsulation was an endothermic process, driven by a
large increase in entropy due to solvent release.2 However, the van't Hoff plot used for
that study assumed the enthalpy and entropy of binding were both temperature
independent, which in general is not the case. By neglecting the ∆Cp contribution –
which can be quite relevant for host-guest binding interactions – the values of ∆H° from
a van’t Hoff plot can be incorrect.3 Isothermal titration calorimetry (ITC) offers a direct
method to measure the enthalpy of a reaction, ∆H°, without relying on any assumptions
about temperature dependence.4, 5 ITC experiments were carried out in collaboration
with Prof. Giuseppe Arena and Dr. Carmelo Sgarlata at the Università di Catania in
Catania, Italy. This chapter represents the initiation of a long-term collaboration between
the Raymond and Arena groups for thermodynamic studies.
- 37 -
Background: Guest binding interactions
The host-guest chemistry of the M4L6 tetrahedral cluster has been studied in great
detail since its preparation was first reported in 1998 by Caulder et al.1 When prepared
with trivalent metal ions such as gallium(III), the anionic M4L6 cluster has an overall -12
charge, making K12[Ga4L6] very water soluble. The interior of the tetrahedron is
surrounded by six naphthalene rings from the ligands, forming a 300–500 Å3
hydrophobic cavity that preferentially binds lipophilic monocations as guests.2 The X-
ray crystal structure of K5(Et4N)7[Fe4L6] shows one Et4N+ cation occupying the interior
cavity, and six Et4N+ cations in close contact with the aromatic rings of the ligands.6 As
discussed in Chapter 2, these exterior Et4N+ cations remain closely associated with the
cluster in aqueous solution, demonstrating that the M4L6 cluster has a second set of guest
binding sites on the exterior.
Cross peaks in the 1H NOESY spectrum have been observed between the exterior
guest resonances and the catechol and naphthalene signals from the host, in contrast to
the interior guest signals which only show cross peaks with naphthalene proton
resonances.6-8 For Et4N+, the intensities of the catechol and naphthalene cross peaks are
similar, consistent with the X-ray crystal structure showing Et4N+ associated with both
the catechol and naphthalene rings.6 For the half-sandwich iridium complex 1, cross
peaks between the Cp* proton signals and the
host catechol protons are much more intense
than those with naphthalene.7 The anionic
charge of the [Ga4L6]12- cluster is
concentrated at the metal catecholate vertices
- 38 -
of the assembly, suggesting electrostatic attraction to monocationic 1 is the predominant
interaction, which causes strong ion pairing to the anionic host. In contrast, ammonium
cations such as Me4N+ and Et4N
+ can also bind to the aromatic naphthalene faces of the
ligand scaffold via cation-π interactions. Encapsulated guests can only interact with the
naphthalene ring walls surrounding the cavity, so cation-π and π-π interactions are very
favorable for interior binding. Clearly, interactions with the ligand are extremely
important for favorable guest binding interactions. To illustrate the charge distribution of
a coordinated ligand in the [Ga4L6]12- host, the potential density surface of the
hypothetical model complex [K2L]2- was generated from a quantum mechanical
computational model (Figure 3.1).
Exterior binding and guest encapsulation are not independent processes. In fact,
recent kinetic studies have demonstrated that guest encapsulation and dissociation
Figure 3.1. Potential density surface of the [K2L]2- complex generated from the Hartree-Fock quantum mechanical calculation, as a hypothetical model to approximate the electronics of the coordinated ligands in the [Ga4L6]
12- cluster. Negatively charged regions are shown in red, and positive charges are colored blue.
- 39 -
proceed in a stepwise process via an exterior ion-pair intermediate.7, 9, 10 What are the
factors which determine how favorable interior and exterior guest binding will be?
Noncovalent interactions such as coulombic attraction, cation-π interactions, π-π
interactions, and van der Waals forces are all enthalpically favorable in the gas phase, but
entropy is lost when the guest binds to the host. However, due to the large charge of the
[Ga4L6]12- host, solvation effects can be extremely powerful, particularly in aqueous
solution. When a monocationic guest is encapsulated, the overall charge of the host
decreases from -12 to -11, and the solvated cation must be completely stripped of solvent
to enter the interior cavity. Both desolvation processes involve an enormous enthalpic
cost, but the liberation of water molecules from the highly organized solvation shells
leads to a large increase in entropy. Thus, guest encapsulation may be entropically
driven, even if it is an endothermic process.
Solvation of soft, polarizable donors and acceptors is usually less powerful, so
binding is typically enthalpically favorable but entropically disfavored.11 Cation-π
interactions involve soft, polarizable acceptors, and exothermic binding interactions are
often observed even in aqueous solution.12 Thus, exterior binding of ammonium cations
such as Me4N+ and Et4N
+ to the naphthalene ring faces of the host may be enthalpically
driven, but entropically disfavored.
Born solvation theory
According to the Born solvation model, for a spherical ion of charge zi with a
radius ri the Gibbs free energy of solvation relative to the gas phase is given by:13
- 40 -
¢Gsolv = ¡µ
Nae2
8¼²0
¶
z2i
ri
µ
1¡ 1
²
¶
= ¡¯µ
z2i
ri
¶ µ
1¡ 1
²
¶
where ¯ ´ Nae2
8¼²0
¢Gsolv = ¡µ
Nae2
8¼²0
¶
z2i
ri
µ
1¡ 1
²
¶
= ¡¯µ
z2i
ri
¶ µ
1¡ 1
²
¶
where ¯ ´ Nae2
8¼²0
(3.1)
where ² is the dielectric constant of the solvent (modeled as a continuum), Na is
Avogadro’s number, e is the charge of an electron, and ²0 is the permittivity of vacuum.
Since dG = V dP ¡ SdTdG = V dP ¡ SdT , S = ¡µ
@G
@T
¶
P
S = ¡µ
@G
@T
¶
P
, and thus the entropy of solvation is:
¢Ssolv = ¯z2
i
ri
µ
1
²2d²
dT
¶
¢Ssolv = ¯z2
i
ri
µ
1
²2d²
dT
¶
(3.2)
Furthermore, since ¢G = ¢H ¡ T¢S¢G = ¢H ¡ T¢S, the enthalpy of solvation is:
¢Hsolv = ¡¯z2
i
ri
µ
1¡ ²¡1 ¡ T²¡2 d²
dT
¶
¢Hsolv = ¡¯z2
i
ri
µ
1¡ ²¡1 ¡ T²¡2 d²
dT
¶
(3.3)
In H2O, = 78.36 and at 25 °C,14 and the constant
β = 6.95 x 10-5 J·m mol-1, so Equations 3.1 – 3.3 can be expressed numerically for
convenience:
where 1 e.u. ≡ 1 cal mol-1 K-1, and ri has units of Å in the above formulae.
The Born solvation free energies, entropies, and enthalpies for the ions considered
in this chapter are listed in Table 3.1 for comparison purposes. Note that the host
solvation enthalpy cost for guest encapsulation is about 400 kcal mol-1 due to the
reduction in overall charge from -12 to -11. Furthermore, note the enthalpy cost for
desolvation of a dicationic species is much larger than for the monocation due to the z2
- 41 -
dependence of Equation 3.3, explaining why a dication has never been observed inside
the [Ga4L6]12- host cavity.2
Table 3.1. Born thermodynamic parameters for solvation of various ions in H2O at 25 °C calculated using Equations 3.1 – 3.3. In addition to the alkylammonium monocations studied in this chapter, a hypothetical dication with the same radius as Et4N
+ is listed for comparison.
Ion ri (Å) zi ¢Gsolv¢Gsolv
(kcal mol-1) ¢Hsolv¢Hsolv
(kcal mol-1) ¢Ssolv¢Ssolv (e.u.)c
Me4N+ 2.93a 1 -55.9 -56.9 -3.3
Et4N+ 3.48a 1 -47.1 -47.9 -2.8
Pr4N+ 3.90a 1 -42.0 -42.7 -2.5
Bu4N+ 4.25a 1 -38.6 -39.2 -2.3
Hypothetical dication 3.48 2 -188 -192 -11.1
[Ga4L6]12- 9.5b -12 -2483 -2527 -146
[R4N⊂Ga4L6]11- 9.5b -11 -2087 -2123 -122
a) Radius calculated from volume of molecular model (semi-empirical minimization). b) Radius estimated from X-ray crystal structure of K5(Et4N)7[Fe4L6]. c) 1 e.u. ≡ 1 cal mol-1 K-1.
Results and Discussion
When K12[Ga4L6] is dissolved in H2O with 0.1 M KCl and titrated with R4NCl
(R = Me, Et, Pr, Bu), the observed heats are very different depending on the size of the
guest being added (Figure 3.2). Exothermic reactions are observed for R = Me, Et, and
Pr, but Bu4N+ switches from exothermic to endothermic as the titrant concentration is
increased. (The addition range for Bu4N+ was limited to 7 equivalents due to
precipitation problems at higher ratios). Values for ∆H° and log β were obtained for
Me4N+, Et4N
+, and Pr4N+ by fitting the data with Hyp∆H,15 and ∆S° was calculated from
the fitted quantities (Table 3.2). Simple 1:1 (guest:host) binding models were used for
Me4N+ and Et4N
+, but an additional 2:1 interaction could be elucidated with the Pr4N+
- 42 -
system. Unfortunately, the Bu4N+ titration curve could not be fit at all to any binding
model, probably because the host does not encapsulate the large cation. It is thought that
Bu4N+ may bind to the exterior of the host,10 but the observed ITC data cannot confirm
nor deny this hypothesis. Since this remains unresolved, Bu4N+ will not be discussed
further in this chapter.
Table 3.2. Thermodynamic parameters of complex formation of Me4N+, Et4N
+, and Pr4N+ with the
[Ga4L6]12- host calculated from isothermal titration calorimetry data measured at 25 °C in 0.1 M KCl.
Titrant Product Species log β ∆H° (kcal mol-1) ∆S° (e.u.)a
Me4NCl (Me4N)1Host 1.1(3) -6.02(5) -15 ± 1
Et4NCl (Et4N)1Host 1.8(1) -8.82(1) -21.3(5)
(Pr4N)1Host 2.0(3) -5.97(5) -11 ± 1 Pr4NCl
(Pr4N)2Host 4.5(4) -0.7(1) 18 ± 2
a) 1 e.u. = 1 cal mol-1 K-1.
The data in Table 3.2 clearly demonstrate that the interactions of R4N+ cations
with the [Ga4L6]12- cluster are enthalpically favored (R = Me, Et, Pr). The values of log β
for Me4N+ and Pr4N
+ are similar to the equilibrium constants for guest encapsulation by
0 5 10 15 20 25
-5
0
5
10
15
20
To
tal H
ea
t, m
J
[G]/[H]
Me4N
+
Et4N
+
Pr4N
+
Bu4N
+
0 5 10 15 20 25
-2
-1
0
1
2
Incre
me
nta
l H
ea
t, k
J m
ol-1
titra
nt
[G]/[H]
Me4N
+
Et4N
+
Pr4N
+
Bu4N
+
Figure 3.2. ITC reaction heats observed when 1 mM K12[Ga4L6] is titrated with R4NCl (R = Me, Et, Pr, Bu) in aqueous 0.1 M KCl at 25 °C. a) Incremental heat released per mole of titrant injected. b) Cumulative heat curves observed for the four different cations interacting with the host. All data were corrected for dilution effects. Note that the vertical axes correspond to the heat released by the system, and therefore an exothermic reaction is observed as a positive heat.
- 43 -
[Ga4L6]12- observed via 1H NMR in D2O
4 (for Pr4N+, Kb = 1.1 x 102 M-1, and for Me4N
+
Kb << 102 M-1),1 but the value of log β for Et4N+ listed in Table 3.2 is two orders of
magnitude lower than its value of log Kb = 4.3 measured via NMR!1 The high affinity of
Et4N+ for the [Ga4L6]
12- interior cavity has been verified by literally hundreds of NMR
experiments independently performed by several different researchers over the last ten
years, so its validity is not in question. However, ten separate ITC experiments have
been performed that all indicate log β = 1.8 for the interaction of Et4N+ with the host in
0.1 M KCl, ruling out artifacts from experimental errors. If additional equilibria with
different stoichiometries are included in the model, such as the 2:1 (or even 7:1) species
expected for ion pairing to the host-guest complex, the fit does not converge.5 Thus, the
equilibrium constant for Et4N+ listed in Table 3.2 may be primarily describing exterior
binding interactions, or perhaps something much more complicated; one hypothesis
proposed involves the guest encapsulation and host assembly equilibria coupled in a
square scheme. Work towards understanding this system is currently in progress, and we
hope to resolve this issue as the collaborative project continues.
The low but nonzero value of log β for Me4N+ is driven entirely by enthalpy, and
is strongly entropically disfavored. Whether Me4N+ interacts primarily with the exterior
or interior of the host is unclear. The favorable exothermic interaction is almost
4 For the remainder of this chapter, Kb refers to the equilibrium constant for encapsulation of a guest into the interior cavity of [Ga4L6]
12- in aqueous solution. 5 ITC experiments with [Ga4L6]
12- had been performed previously by Dr. Martin Michels in collaboration with Dr. Linfeng Rao at the Lawrence Berkeley National Laboratory. Titrations of the host with many different guests were performed, including Et4N
+, and the solution conditions he used were the same as those used in this chapter. Dr. Michels tabulated both exterior and interior binding equilibrium constants, stoichiometries, and reaction enthalpies for Et4N
+ from his ITC measurements in an unpublished report (which has been cited several times in various publications). The data were fit using the BindWorks software, with a multiple binding sites model. The fits obtained with this software are questionable, since its computational algorithms are prone to false minima. Furthermore, five variables were refined simultaneously in a nonlinear fit, with a large degree of correlation between the parameters. These data have not yet been analyzed using the methods described in this chapter.
- 44 -
completely neutralized by the free energy cost associated with the large decrease in
entropy. Cation-π interactions with the aromatic naphthalene rings of the ligands are
likely responsible for the favorable enthalpy, and numerous examples of exothermic
cation-π interactions with Me4N+ in aqueous solution have been described.12, 16 The
highly negative ∆S° is contrary to what one would expect for interior binding based on
solvation effects – the entropy gained due to desolvation of host (z = -12 to -11) and guest
from the Born equation is about 25 e.u. It is possible that the observed heat is due to
exterior binding interactions, but encapsulation may be involved as well. Similar
unexpected entropic costs have been observed by Arena et al., where encapsulation of a
pendant Me3N+ group into the pocket of a calixarene host is entropically disfavored in
aqueous solution as well.16 The authors attributed this entropy loss to both confinement
of the guest to the cavity and a stiffening of the surrounding host system.
A “stiffening” of the [Ga4L6]12- cluster upon encapsulation of Me4N
+ may also be
involved here. In the absence of guest, the cluster is ill-formed in aqueous solution, with
ligand dissociation readily observed via 1H NMR as extra peaks gradually appear in the
aromatic region of the spectrum of K12[Ga4L6] in degassed unbuffered D2O (pD ~ 8).
Addition of one equivalent of guest causes the cluster to snap together, and six sharp
aromatic signals are observed (characteristic of T-symmetric host), with no extra peaks.
This guest-templated self-assembly results in a large negative ∆S due to many partially
assembled microstates collapsing into one single species. Furthermore, ligand
coordination may contribute to the relatively high value of -∆H°, since the GaIII-
catecholamide coordination is most likely exothermic. While no experimental enthalpy
information with GaIII catecholates was found in the literature, this prediction is justified
- 45 -
since the coordination of enterobactin to FeIII at pH 9 is exothermic for the following
equilibrium reaction17
H3ent3- + Fe3+ = [Fe(ent)]3- + 3H+ ∆H = -6.5(3) kcal/mol
Since both enterobactin and H4L feature catecholamide chelating moieties, and the
chemistry of GaIII and FeIII are similar,18 the coordination of catecholamides to GaIII are
likely exothermic reactions as well.
The Pr4N+ system is very interesting, since the thermodynamic parameters for
both the 1:1 and 2:1 complexes could be quantified. For the 1:1 complex, does the Pr4N+
cation bind to the interior cavity to form [Pr4N+⊂Ga4L6]
11-, or does it bind to one of the
naphthalene rings on the exterior to form the [(Pr4N)···[Ga4L6]]11- ion pair? This may be
explored by comparing the stepwise binding equilibria (Table 3.3).
Table 3.3. Thermodynamic parameters for the stepwise binding equilibria for Pr4N+ with the [Ga4L6]
12- host (H) in 0.1 M aqueous KCl at 25 °C, computed from the calorimetry data listed in Table 3.2.
Reaction logKilogKi ¢H±
i¢H±
i , kcal/mol ¢S±
i¢S±
i , e.u.
H + Pr4N+ ÐÐ [(Pr4N
+)1·H] 2.0(3) -5.97(5) -11 ± 1
[(Pr4N+)1·H] + Pr4N
+ ÐÐ [(Pr4N+)2·H] 2.5(7) 5.2(1) 29 ± 4
The first Pr4N+ cation binding step with the free [Ga4L6]
12- cluster is an
exothermic process, but the second Pr4N+ binding reaction with the [(Pr4N
+)1H] adduct is
endothermic. Conversely, binding the first Pr4N+ cation is entropically disfavored, but
the second cation binding event is driven by a large increase in entropy. The two
stepwise binding equilibrium constants are similar, with their difference less than the
standard error of the values. However, K1 is driven entirely by enthalpy, and K2 is driven
entirely by entropy.
- 46 -
The value of ¢H±
1¢H±
1 is equal to ¢H±¢H± observed for Me4N+ within error, but less
entropy is lost during the first Pr4N+ binding step than the Me4N
+ reaction, causing the
1:1 binding constant for Pr4N+ to be an order of magnitude higher than that for Me4N
+.
The Born solvation entropy for Pr4N+ is only about 1 e.u. less negative than that for
Me4N+, so solvation cannot account for difference in reaction entropy observed for
forming 1:1 complexes with Me4N+ and Pr4N
+. However, if the first Pr4N+ cation binds
to the exterior and the second is encapsulated inside the host cavity, solvation effects can
account for the large difference between ¢S±
1¢S±
1 and ¢S±
2¢S±
2. This endothermic, entropically
driven encapsulation equilibrium for Pr4N+ was observed from a van’t Hoff plot using
variable temperature NMR to measure the encapsulation equilibrium constant.2
As discussed by Parac et al.,2 the host charge changes from z = -12 to z = -11
upon guest encapsulation, and the Born model predicts an entropy gain of about 23 e.u.
due to the change in solvation. In addition, the aqueous Pr4N+ ion must be completely
stripped of solvent to fit inside the host cavity, leading to an additional entropy gain of
about 2.5 e.u. Finally, about 8-10 water molecules are expected to occupy the interior
cavity of aqueous [Ga4L6]12-, which are liberated upon Pr4N
+ encapsulation for an
additional entropy gain. (The transfer of water molecules “frozen” in hydrated salts into
bulk water typically show an entropy change of about 6.7 e.u.).19 The positive value
observed for ¢H±
2¢H±
2 can also be attributed to solvation changes, since a large enthalpic cost
must be paid for the desolvation of the host and guest. Furthermore, the larger
Pr4N+ cation does not fit inside the host cavity as well as Et4N
+, and expansion of the host
cavity to accommodate the larger guest may also contribute to the positive ∆H2 value.
- 47 -
When weaker guests such as Me4N+ or Pr4N
+ are present in the working cell with
the host, the heat curves observed during titration of Et4N+ are significantly altered
(Figure 3.3). When two equivalents of the weakly encapsulated Me4N+ guest are
combined with the host solution in 0.1 M KCl and allowed to equilibrate for over an
hour, the heat released during the first addition of Et4N+ (~0.8 equiv) is five times larger
than that observed with the host alone. However, the two cumulative heat curves are
almost perfectly parallel, demonstrating the equilibrium constants characterizing these
two systems are essentially the same. Thus, the presence of two equivalents of Me4N+ in
solution with [Ga4L6]12- causes the reaction of the first equivalent of Et4N
+ with the host
to be much more exothermic. The difference in the Born solvation enthalpies for the two
cations is consistent with the observed increase in heat released. Consider the guest
exchange reaction where Me4N+ is displaced by Et4N
+:
[Me4N⊂Ga4L6]11- + Et4N
+(aq) ÐÐ [Et4N⊂Ga4L6]
11- + Me4N+
(aq)
0 5 10 15 200
5
10
15
20
25
30
To
tal H
ea
t, m
J
[Et4N
+]/[Host]
0.1 M KCl
KCl + 2 eq. Me4NCl
0.1 M Me4NCl
0 5 10 15 200
5
10
15
20
25
30
To
tal H
ea
t, m
J
[Et4N
+]/[Host]
0.1 M KCl
KCl + 2 eq. Pr4NCl
0.1 M Pr4NCl
a) b)
Figure 3.3. The thermograms observed during the titration of K12[Ga4L6] with Et4NCl are strongly affected by the presence of a) Me4N
+ and b) Pr4N+ as secondary guests. The black curves in both plots are identical,
measured in the absence of secondary guest with 0.1 M KCl as the ionic medium. The red curves were measured with 2 equivalents of secondary guest added to the host solution, also with 0.1 M KCl as the ionic medium. The green curves were measured when the KCl ionic buffer was replaced with the specified alkylammonium salt in both the analyte and titrant solutions. Heats are corrected for dilution effects.
- 48 -
The enthalpic cost of Et4N+ desolvation is offset by the much more negative resolvation
enthalpy of the smaller Me4N+ cation upon ejection from the host cavity, with the Born
model predicting a net ∆Hsolv ≈ -9 kcal mol-1.
For the larger Pr4N+ cation, the solvation enthalpy is less favorable than that for
Et4N+ due to the 1/r dependence in Equation 3.2. The simple solvation model described
above predicts that when two equivalents of Pr4NCl are combined with the host in a
similar fashion, the heat released after the first injection of Et4N+ should be lower than
that observed in the absence of secondary guest. This is not the case – the heat is higher
in the presence of two equivalents of Pr4N+, although to a lesser extent than with Me4N
+.
The reason for this is not known at this time, and is currently under investigation.
However, one possible explanation is that the interior binding affinity of Et4N+ is truly
enthalpically favored.
With a large excess of Me4N+ present, very different results are observed. The
green curve in Figure 3.3a was obtained by replacing the 0.1 M KCl ionic strength buffer
with 0.1 M Me4NCl in both the titrant (Et4NCl) and analyte (K12[Ga4L6]) solutions.
Thus, approximately 100 equivalents of Me4N+ per [Ga4L6]
12- host was present
throughout the experiment. A sharp break in the total heat curve, followed by a relatively
flat plateau, indicates that the primary exothermic reaction is essentially complete after
addition of two equivalents of Et4N+ per unit host. Sharp break points are characteristic
of systems with high binding constants (Kb ≥ 104 M-1).4 It is possible that the exterior
binding sites of the host are saturated by the large excess of Me4N+, allowing the Et4N
+
cation to bypass the exterior binding intermediate and directly bind to the host interior.
Despite its much weaker binding affinity, the very large excess of Me4N+ competes to a
- 49 -
small extent with the much stronger Et4N+ guest for binding to the host cavity. For this
reason, a satisfactory fit of the data can only be obtained if the model includes the 1:1
binding equilibria of both Et4N+ and Me4N
+. Furthermore, such a model accurately
accounts for the displacement reaction of encapsulated Me4N+ by the superior Et4N
+
guest. This analysis revealed that for Et4N+ encapsulation, ∆H° = -4.8 kcal mol-1,
log Kb = 4.6, and ∆S° = 5 e.u. Note that by including the Me4N+ binding equilibrium in
the model, these values are for the true guest binding equilibrium between Et4N+ and
[Ga4L6]12-, as opposed to the displacement reaction. These preliminary results must be
confirmed by titrations over a narrower concentration range by injecting a more dilute
Et4N+ stock solution; such experiments were being performed by our collaborators in
Catania at the same time this dissertation was written.
A similar effect was observed with 0.1 M Pr4NCl, although the break point in
Figure 3.3b is much less pronounced compared to that observed with Me4N+. The
shallower curve is consistent with the higher binding constant for Pr4N+, requiring higher
amounts of Et4N+ to fully displace the encapsulated Pr4N
+. After the break point is
attained early in the titration, corresponding to quantitative Et4N+ encapsulation, heat
continues to be released as more Et4N+ is injected. This may be due to displacement of
exterior bound Pr4N+ by the smaller Et4N
+, since cation-π interactions are typically more
favorable for smaller cations.12
The binding of Et4N+ with [Ga4L6]
12- in 0.1 M KCl is clearly an exothermic
process in all three cases, as shown in Table 3.2 and Figure 3.3a. However, the
enthalpies inferred from van’t Hoff plots of alkylammonium binding to the [Ga4L6]12-
host in D2O indicated guest encapsulation was endothermic.2 Dr. Martin Michels, a
- 50 -
previous postdoctoral scholar working with Professor Raymond through May of 2001,
also noted this discrepancy during previous calorimetry experiments with the [Ga4L6]12-
host. Using variable temperature NMR spectroscopy to generate van’t Hoff plots, Dr.
Michels found that in the absence of ionic strength buffer, the observed slopes indicate
encapsulation of both Et4N+ and Pr4N
+ are endothermic. In the presence of KCl,
however, the slopes change sign, indicating that encapsulation becomes exothermic for
both guests at higher ionic strength.20
To test whether added salt has an appreciable impact on the guest binding
enthalpy, two similar ITC experiments were performed: one with 0.1 M aqueous KCl,
and the other using only water (“non-salt” system). Both analyte solutions contained
1 mM K12[Ga4L6], and both titrant solutions contained 10 mM Et4NCl. Although
numerical values could not be fit to the non-salt data due to spurious dilution heats
caused by large ionic strength changes, the data in Figure 3.4 clearly show that the
reaction is exothermic whether or not KCl is present. The difference in slope may be due
to uncompensated endothermic host dilution that becomes significant in the absence of
ionic strength buffer.
Since metal-catechol binding equilibria are strongly pH dependent, with
formation of tris-bidentate metal catechol complexes favored in basic solution, should a
buffer system be used to regulate pH during the encapsulation process? For equilibrium
studies, buffers can be very useful if a chemical reaction leads to a pH change. A 1 mM
solution of [Ga4L6]12- in 0.1 M KCl was prepared, and the pH was measured to be 7.56 in
the absence of guest. To this was added Et4NCl titrant solution in several aliquots,
allowing the system to equilibrate for 20 minutes between each addition. No significant
- 51 -
pH change was observed (i.e. ±0.03 pH units) even up to 16 equivalents of Et4N+ per
host. This clearly demonstrates that protons are neither consumed nor released during the
guest binding interaction.
Summary
Isothermal titration calorimetry is the most accurate method for measuring
reaction enthalpies, since heat is measured directly. The calorimetry data described in
this chapter unambiguously demonstrate that the encapsulation of Et4N+ by the [Ga4L6]
12-
host in aqueous solution is exothermic by nearly 5 kcal mol-1, in contrast to the
endothermic binding enthalpy inferred from the van’t Hoff plot. However, the complex
binding interactions are difficult to identify from this method, since the only observable
parameter involved is heat. This chapter represents the beginning of a long term research
Ethyl With or Without KCl
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0.00 0.20 0.40 0.60 0.80 1.00
# eq. ethyl
Cu
mu
lati
ve h
eat,
mJ
no KCl
with KCl
Figure 3.4. Comparison of the cumulative heats released during titration of 1 mM [Ga4L6]12- with
9.23 mM Et4NCl at 25 °C for solutions prepared with and without KCl ionic strength buffer.
- 52 -
collaboration, and the unresolved discrepancies between NMR and ITC observations will
be addressed as this project continues.
Experimental
General considerations
Unless noted otherwise, reagents were obtained from commercial suppliers and
used without further purification. Standard Schlenk techniques were used for reactions
carried out under argon, and a glove box continuously purged with nitrogen was used to
manipulate and store air-sensitive solids. (Due to less rigorous glovebox conditions in
Catania, vials containing host samples were flushed with argon, capped, wrapped in
Parafilm, and placed in a -20 °C freezer for later use). When necessary, solvents were
degassed by at least six pump/fill cycles while vigorously stirring, using argon for the fill
step. H4L (H4L = 1,5-bis(2,3-dihydroxybenzamido)naphthalene) and K12[Ga4L6] were
synthesized according to literature procedures.1, 2 Tetraethylammonium chloride (Et4NCl)
was recrystallized from absolute ethanol/ether and dried in vacuo over P2O5 at room
temperature for 12 hours, then dried in vacuo over molecular sieves at 60 °C for 18 hours
and stored under nitrogen. Tetramethylammonium chloride (Me4NCl),
tetrapropylammonium chloride (Pr4NCl), and tetrabutylammonium chloride (Bu4NCl)
were not recrystallized, but were dried over P2O5 in a manner similar to that for Et4NCl.
Aqueous silver nitrate, used as a titrant for chloride determination, was standardized by
Valeria Zito (Dipartimento di Scienze Chimiche at the Università di Catania, Italy) via
seven titrations into aqueous NaCl, using K2CrO4 as an indicator
([AgNO3] = 0.0488(2) M). This AgNO3 solution was used within one week after
- 53 -
standardization. Stock solutions of the R4NCl titrants (R = Me, Et, Pr, Bu) were
standardized using this AgNO3 titrant (with K2CrO4 indicator) to determine the chloride
concentration to three significant figures. Routine mass spectrometry and elemental
analysis was performed by the Mass Spectrometry Laboratory and Microanalysis Facility
in the College of Chemistry at the University of California, Berkeley.
Thermogravimetric analyses (TG) were performed by the analytical facilities in the
Dipartimento di Scienze Chimiche at the Università di Catania, Italy.
K12[Ga4L6]·(Me2CO)2(H2O)8. The following reaction was carried out under argon. A
suspension of 1.51 g (3.50 mmol) of H4L and 855 mg (2.33 mmol) of Ga(acac)3 in
150 mL of degassed methanol was prepared, and the slurry was degassed via seven
pump/fill cycles. Via syringe, 16 mL of 0.5 M KOH in methanol was added to the
reaction mixture while stirring, immediately followed by via seven pump/fill cycles to
remove any oxygen introduced with the base solution. Addition of base caused most
solid to dissolve, and the color changed from white to yellow. Nearly all solid dissolved
after stirring for 15 min. The dark yellow solution was stirred overnight at room
temperature, filtered through a fine frit to remove insoluble residual impurities, filtered
again through several 0.2 µm nylon syringe filter disks, and the volume was reduced to
30 mL with a vacuum pump. Any solid that formed during this evaporation step was re-
dissolved with degassed methanol. The product was precipitated by slowly adding
600 mL of degassed acetone via cannula, stirred slowly for 5 min to aid mixing of
acetone and methanol, and left alone for two hours. The yellow precipitate that formed
was collected on a frit under a stream of nitrogen, washed with acetone (3 x 50 mL), and
- 54 -
dried in vacuo 12 hours to yield 1.3 g (65%) of light yellow powder. Duplicate Anal.
Calc. (found) for C144H84Ga4K12N12O36·(Me2CO)2·(H2O)8: %C, 50.51 (50.86); H, 3.17
(3.12); N, 4.71 (4.75). Thermogravimetric Analysis: ~8% volatile components. The solid
is somewhat hygroscopic, gaining mass on an analytical balance after removing from
glove box. Samples for thermogravimetric analysis were stored in an air-filled vial for
several hours before the TG experiment, so the sample had become fully hydrated by the
time the experiment was performed.
Isothermal titration calorimetry (ITC) experiments
Measurements were performed using a CSC 5300 Nano-Isothermal Titration
Calorimeter III at the Università di Catania, Italy. The ambient temperature of the room
was controlled to 25 ± 1 ºC, with double doors for the entrance minimizing temperature
fluctuations. All measurements were performed with a cell temperature of 25.000(5) ºC,
with 0.1 M KCl in the dummy cell. Samples were degassed by stirring under vacuum for
20 minutes to ensure no air bubbles formed in the cell or syringe. Prior to beginning each
experiment, the instrument response was calibrated with 10 identical power pulses
(50-300 uJ) using the sample for the specific experiment in the working cell. The cell
volume was 1036 µL, and was operated on overfill mode.21 Titrant was automatically
injected in 8 µL increments using a special syringe with a stirrer tip, rotating at 200 rpm
to mix the solution. For experiments in the presence of host, a 600 s delay between
injections was used up to 4:1 guest/host ratio due to sluggish kinetics in this region. The
time intervals between injections were incrementally shortened as higher guest/host ratios
were reached, eventually reaching 300 s as the minimum delay for greater than 10 equiv.
- 55 -
guest. A dilution experiment in the absence of host was performed with the same titrant
solution and ionic medium for each set of conditions.
The raw data were corrected for baseline heat and integrated using BindWorks,22
and the integrated dilution heat was subtracted from the integrated gross heat for each
interval to obtain the net incremental heat for each injection. Values for K and ∆H° were
obtained with the computer program Hyp∆H,15 which was derived from the general
analysis suite HYPERQUAD.23
Molecular modeling calculations
Quantum mechanical calculations of the hypothetical [K2L]2- complex were
performed with Spartan ’02 for Linux.24 The equilibrium geometry of the complex was
first determined using a molecular mechanics calculation (MMFF method). The
electronic structure of this minimized structure was then computed with a Hartree-Fock
molecular orbital calculation, using the 6-31G* basis set. From the results of this single-
point energy calculation, the potential density surface map was calculated to produce
Figure 3.1.
References
1. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
2. Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am.
Chem. Soc. 1998, 120, 8003-8004.
3. a) Diederich, F.; Dick, K.; Griebel, D., “Complexation of arenes by macrocyclic hosts in aqueous and organic solutions.” J. Am. Chem. Soc. 1986, 108, 2273-2286; b) Meric, R.; Vigneron, J.-P.; Lehn, J.-M., “Efficient complexation of quaternary
- 56 -
ammonium compounds by a new water-soluble macrobicyclic receptor molecule.” J.
Chem. Soc. Chem. Comm. 1993, 129-131; c) Smithrud, D. B.; Wyman, T. B.; Diederich, F., “Enthalpically driven cyclophane-arene inclusion complexation: Solvent-dependent calorimetric studies.” J. Am. Chem. Soc. 1991, 113, 5420-5426; d) Ferguson, S. B.; Sanford, E. M.; Seward, E. M.; Diederich, F., “Cyclophane-arene inclusion complexation in protic solvents: solvent effects versus electron donor-acceptor interactions.” J. Am. Chem. Soc. 1991, 113, 5410-5419; e) Stauffer, D. A.; Barrans, R. E., Jr.; Dougherty, D. A., “Concerning the thermodynamics of molecular recognition in aqueous and organic media. Evidence for significant heat capacity effects.” J. Org. Chem. 1990, 55, 2762-2767; f) Diederich, F., “Complexation of Neutral Molecules by Cyclophane Hosts.” Angew. Chem. Int. Ed. Engl. 1988, 27, 362-386; g) Ferguson, S. B.; Seward, E. M.; Diederich, F.; Sanford, E. M.; Chou, A.; Inocencio-Szweda, P.; Knobler, C. B., “Strong enthalpically driven complexation of neutral benzene guests in aqueous solution.” J. Org. Chem. 1988, 53, 5593-5595; h) Petti, M. A.; Shepodd, T. J.; Barrans, R. E., Jr.; Dougherty, D. A., “"Hydrophobic" binding of water-soluble guests by high-symmetry, chiral hosts. An electron-rich receptor site with a general affinity for quaternary ammonium compounds and electron-deficient π systems.” J. Am. Chem. Soc. 1988, 110, 6825-6840.
4. Fisher, H. F.; Singh, N., “Calorimetric Methods for Interpreting Protein-Ligand Interactions.” Methods in Enzymology 1995, 259, 194-221.
5. a) Horn, J. R.; Brandts, J. F.; Murphy, K. P., “van't Hoff and Calorimetric Enthalpies II: Effects of Linked Equilibria.” Biochemistry 2002, 41, 7501-7507; b) Arena, G.; Calì, R.; Maccarrone, G.; Purrello, R., “Critical review of the calorimetric method for equilibrium constant determination.” Thermochim. Acta 1989, 155, 353-376.
6. Caulder, D. L.; Brückner, C.; Powers, R. E.; König, S.; Parac, T. N.; Leary, J. A.; Raymond, K. N., “Design, Formation, and Properties of Tetrahedral M4L4 and M4L6 Supramolecular Clusters.” J. Am. Chem. Soc. 2001, 123, 8923-8938.
7. Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Scope and Mechanism of the C-H Bond Activation Reactivity within a Supramolecular Host by an Iridium Guest: A Stepwise Ion Pair Guest Dissociation Mechanism.” J. Am. Chem. Soc. 2006, 126, 9781-9797.
8. Tiedemann, B. E. F.; Raymond, K. N., “Dangling Arms: A Tetrahedral Supramolecular Host with Partially Encapsulated Guests.” Angew. Chem. Int. Ed.
2006, 45, 83-86.
9. Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.; Raymond, K. N., “Guest Exchange Dynamics in an M4L6 Tetrahedral Host.” J. Am. Chem. Soc. 2006, 128, 1324-1333.
10. Fiedler, D.; van Halbeek, H.; Bergman, R. G.; Raymond, K. N., “Supramolecular Catalysis of Unimolecular Rearrangements: Substrate Scope and Mechanistic Insights.” J. Am. Chem. Soc. 2006, 128, 10240-10252.
- 57 -
11. Ahrland, S., “Thermodynamics of Complex Formation between Hard and Soft Acceptors and Donors.” In Structure and Bonding, Jørgensen, C. K.; Neilands, J. B.; Nyhoum, R. S.; Reinen, D.; Williams, R. J. P., Eds. Springer-Verlag: Berlin, 1968; Vol. 5, pp 118-149.
12. Dougherty, D. A.; Ma, J. C., “The Cation-π Interaction.” Chem. Rev. 1997, 97, 1303-1324.
13. Born, M., “Volumen und Hydratationswärme der Ionen.” Z. Physik 1920, 1, 45-48.
14. Lide, D. R., Handbook of Chemistry and Physics. 81st ed.; CRC Press: Boca Raton, 2000; p 6.149-6.150.
15. Gans, P.; Sabatini, A.; Vacca, A. Hyp∆H, 1.0.62; Protonic Software: Leeds, UK, 2007.
16. Arena, G.; Casnati, A.; Contino, A.; Lombardo, G. G.; Sciotto, D.; Ungaro, R., “Water-Soluble Calixarene Hosts that Specifically Recognize the Trimethylammonium Group or the Benzene Ring of Aromatic Ammonium Cations: A Combined 1H NMR, Calorimetric, and Molecular Mechanics Investigation.” Chem.
Eur. J. 1999, 5, 738-744.
17. Scarrow, R. C.; Ecker, D. J.; Ng, C.; Liu, S.; Raymond, K. N., “Iron(III) Coordination Chemistry of Linear Dihydroxyserine Compounds Derived from Enterobactin.” Inorg. Chem. 1991, 30, 900-906.
18. Loomis, L. D.; Raymond, K. N., “Kinetics of gallium removal from transferrin and thermodynamics of gallium-binding by sulfonated tricatechol ligands.” J. Coord.
Chem. 1991, 23, 361-387.
19. Phillips, C. S. G.; Williams, R. J. P., Inorganic Chemistry. 1st ed.; Oxford University Press: New York, 1965; Vol. 1, p 260.
20. Michels, M.; Raymond, K. N., unpublished results.
21. Spokane, R. B.; Gill, S. J., “Titration microcalorimeter using nanomolar quantities of reactants.” Rev. Sci. Instrum. 1981, 52, 1728-1733.
22. BindWorks, Version 3.1.8; Calorimetry Sciences Corp.: Lindon, Utah, 2007.
23. Gans, P.; Sabatini, A.; Vacca, A., “Investigation of equilibria in solution. Determination of equilibrium constants with the HYPERQUAD suite of programs.” Talanta 1996, 43, 1739-1753.
24. Spartan '02 for Linux, Build 119; Wavefunction, Inc.: Irvine, CA, 2001.
- 58 -
CHAPTER 4
Partial Guest Encapsulation Modes
Introduction
Guest exchange for the M4L6 tetrahedral host was demonstrated to proceed
through a nondissociative mechanism, where the cluster remains fully intact throughout
the entire process.1, 2 Guest ingress and egress into and out of the host cavity is thought
to occur through expandable apertures in each of the host’s four triangular faces, centered
along the three-fold rotation axis passing through the opposite vertex. This conjecture
raised the following question: could this aperture be used to extend part of a guest outside
the cavity while another part remains bound to the host interior? The answer is yes. This
unique encapsulation mode will be called “partial guest encapsulation.” Two different
categories of guests are discussed in this chapter where partial guest encapsulation has
been observed: zwitterions (RuCn = [CpRu(η6-C6H5(CH2)nSO3)], n = 4, 6, 8, 10) and
monocations with pendant linear alkanes (RuAn+ = [CpRu(η6-C6H5(CH2)nH)]+, n = 4, 6,
8, 10, 12) whose structures are shown in Scheme 4.1. Both guest systems feature linear
Scheme 4.1. Structural diagrams of the RuCn zwitterion and the RuAn+ cation, where n corresponds to the
total number of carbon atoms contained in the linear hydrocarbon chains for both structures.
- 59 -
hydrocarbon chains to pass through the small aperture in the host’s triangular faces
without much distortion, minimizing the energy cost. The work described in this chapter
has been previously described in two separate communications.3, 4
Zwitterionic Guests
With the [Ga4L6]12- ion as the host, suitable guest compounds had to be designed
with appendages capable of protruding through the opening of the host into the bulk
solvent. Such a guest should have three regions: a lipophilic monocationic head group
with a high affinity for the host interior, a linear chain of variable length to protrude
through the aperture, and a hydrophilic anionic end group that cannot be encapsulated
(Figure 4.1). The sandwich complex [CpRu(η6-C6H6)]+ is known to be encapsulated by
the [Ga4L6]12- ion,5 and was chosen as the cationic head group for RuCn. This series of
compounds features linear alkyl chains, with one end bound to the cationic sandwich
complex at the phenyl ligand, and the other bound to a sulfonate anion.
Although several other choices are feasible for the cationic head group, such as
Et4N+ derivatives, the RuII sandwich complex was chosen because its synthesis was
Figure 4.1. a) Cartoon of the M4L6 host encapsulating the cationic head group (red ball) of a zwitterion while the anionic end (blue ball) remains outside the cavity. For this to occur, the linear alkyl chain (wavy line) must pass through an aperture in one of the host’s triangular faces. b) Relationship between the cartoon guest and the actual structure of the RuCn zwitterion.
- 60 -
relatively straightforward. (Attempts to prepare Et3N+-(CH2)n-SO3
- were hindered by
purification difficulties). Commercially available n-phenylalcohols were treated with
PBr3, and the reaction of the resulting bromide with sodium sulfite afforded the
phenylalkanesulfonate product as the sodium salt. These soap-like compounds were then
combined with [CpRu(MeCN)3]PF6 in degassed, anhydrous chloroform to yield the
desired products as the NaPF6 adduct (Scheme 4.2).
Addition of a stoichiometric amount of RuCn to K12[Ga4L6] in D2O led to the
encapsulation of the ruthenium head group for all chain lengths (n = 4 – 10). The 1H
NMR spectra of the resulting host-guest complexes display signals for the Cp and phenyl
rings of the guest shifted to significantly higher fields relative to the values observed in
the absence of host (Figure 4.2). Such an upfield shift is a diagnostic feature observed for
encapsulated guest protons and is caused by the magnetic shielding from the naphthalene
groups surrounding the cavity of the [Ga4L6]12- host.6-8 Furthermore, the two sets of
mirror-related phenyl protons become diastereotopic upon encapsulation of the sandwich
complex by the host because of the chiral environment of the cavity.
Diastereotopic splitting is also observed for most, but not all, geminal methylene
resonances, accompanied by varying upfield shifts. With the aid of 2D COSY and/or
Scheme 4.2. Generalized synthesis for RuCn·NaPF6
- 61 -
TOSCY NMR spectroscopic analysis (Appendix), all chain protons were fully assigned
(Figure 4.2). Methylene carbon toms are numbered sequentially along the chain and
begin with the carbon atom bound to the sandwich complex. For RuC4, the geminal
methylene protons on C1–C3 become diastereotopic upon encapsulation by [Ga4L6]12-,
and diastereotopic geminal methylene protons are observed for C1–C5 for encapsulated
RuC6, RuC8, and RuC10. Upfield shifts tend to increase for protons closer to the head
Figure 4.2. Guest region of the 1H NMR spectra (D2O, 500 MHz) of the host-guest complexes [RuCn⊂Ga4L6]
12-. Guest resonance assignments follow the labeling scheme illustrated for RuC4 (top), with the same numbering pattern used for all chain lengths. Cp denotes the 5H singlet from the Ru-bound cyclopentadiene ring. Interior protons are highlighted with red labels, and all signals that integrate to two protons are identified by subscripted labels.
- 62 -
group. In comparison, several methylene protons remain enantiotopic and show little, if
any, upfield shift. Furthermore, the signal from the methylene group adjacent to the
sulfonate moiety is essentially unaffected by encapsulation of the cationic head group for
all chain lengths. These 1H NMR observations indicate that part of the alkyl chain bound
to the cationic sandwich complex resides within the chiral host, and the rest of that alkyl
chain lies outside the cavity with the terminal sulfonate group. At least one methylene
group is found outside the host in all four systems. Thus, the [Ga4L6]12- cluster
encapsulates only part of the RuCn zwitterion. This represents the first reported example
of partial guest encapsulation with a “closed” host.3
Symmetry of host-guest complexes
The M4L6 cluster has pure rotation point group T symmetry. Three-fold rotation
axes pass through each metal center at the vertices of the tetrahedron, and two-fold
Figure 4.3. Host resonances from the 1H NMR spectra observed for a) point group T typically observed for host-guest complexes, and b) [RuC10⊂Ga4L6]
12- characteristic of C3 symmetry.
- 63 -
rotation axes pass through opposite pairs of naphthalene rings surrounding the cavity.
Most fully encapsulated guests do not change the observed symmetry of the host, because
the guest can rapidly tumble inside the cavity. For these T-symmetric host-guest
complexes, the signals for the 72 host protons are observed as six sets of 12 protons
(Figure 4.3a) because the catechol and naphthalene ring edges are interrelated, each with
three adjacent nonequivalent protons.
In the presence of the protruding arm, however, the host aromatic resonances are
split into 24 sets of three protons (Figure 4.3b). This observation indicates that the
overall symmetry of the cluster is reduced from point group T for the host alone to C3
upon encapsulation of RuCn. Such a symmetry reduction is expected to occur when the
alkyl sulfonate arm of the guest is extruded through the opening in a triangular face of the
tetrahedron (Figure 4.4). This perturbation breaks the ligand two-fold symmetry, but
C3 (observed) D2 (not observed)
Figure 4.4. Breaking the T symmetry of the M4L6 cluster can result in two different lower symmetry subgroups. The protruding tail for RuCn is collinear with one three-fold axis, but breaks the two-fold ligand symmetry, resulting in 24 resonances from the two sets of unsymmetric ligands. The above C3 structure shows one set in red, and the other in blue, with different shades to highlight the lack of ligand symmetry. Alternatively, distortion along a two-fold axis results in the D2 subgroup, which exhibits 18 resonances from the three sets of chemically inequivalent C2 symmetric ligands.
- 64 -
retains the C3 axis running from a gallium(III) vertex through the aperture of interest.
The six ligands that span the edges of the tetrahedron separate into two chemically
nonequivalent groups: three “base” ligands that surround the protruding arm and three
“side” ligands connected to the gallium(III) vertex opposite the protrusion, with 12
different protons for each. According to 2D COSY NMR spectroscopy, the 24 host
signals originate from four sets of catechol protons and four sets of naphthalene protons
(Appendix).
Structural details from NOE interactions
The 2D NOESY NMR spectra of the [RuCn⊂Ga4L6]12- system provide additional
information about the structures of the host-guest complexes. Figure 4.4 shows the
spectrum of the [RuC6⊂Ga4L6]12- system with focus on the cross section between the host
and guest resonances. The phenyl and Cp signals of the guest show strong correlations
with signals from three out of the four sets of host naphthalene protons. No cross peaks
are observed between the phenyl or Cp resonances of the guest and catechol proton
signals of the host. This confirms that the cationic head group is buried deep within the
host cavity, near the naphthalene ring walls. Similar 2D NOESY NMR spectroscopic
observations have been reported for other guests within the [Ga4L6]12- cluster, including
Et4N+ and [CpRu(η6-C6H6)]
+ ions.5, 6
- 65 -
Cross peaks are also observed between several host resonances and most, if not
all, methylene groups on the alkyl chain. The first six to eight methylene groups all show
cross peaks with the same two host resonances – one doublet and one triplet, highlighted
in Figure 4.5. According to molecular modeling studies (Figure 4.5) and 2D COSY
NMR spectra (Appendix), these signals originate from two adjacent naphthalene protons,
located ortho and meta to the amide nitrogen atom. Because of the orientation of the
naphthalene ring, the para proton is directed away from the alkyl chain and strong cross
peaks are not observed with the guest.
Figure 4.5 The 2D NOESY NMR spectrum (D2O, 500 MHz) of the [RuC6⊂Ga4L6]12- system shows cross
peaks between host resonances (horizontal axis) and guest resonances (vertical axis). Resonances from naphthyl protons that border the aperture, ortho and meta to the amide nitrogen atom, are highlighted in red. The catechol resonance highlighted in blue shows cross peaks with signals from exterior methylene protons (nap = naphthalene, cat = catechol).
- 66 -
Relative distances between the host and guest protons in the [RuC6⊂Ga4L6]12-
system were determined from NOE interaction growth rates, showing these ortho and
meta protons are the closest to the protruding alkyl chain, so they must be located on the
C3-related edges of the three naphthalene rings that surround the aperture. These
hydrogen atoms are directed away from the cluster center and can be used as boundary
markers to distinguish the host interior from the exterior. Considering the guest’s alkyl
chain, methylene groups 4 and 5 are closest to the boundary, whereas methylene group 1
is the farthest. This difference suggests that the two naphthyl protons are situated
between the C4 and C5 carbons of the alkyl chain, with C4 on the interior side and C5 on
the exterior side of the cavity boundary. C1, immediately adjacent to the encapsulated
cationic head group, lies deep inside the cluster.
Cross peaks between protons at different positions along the alkyl chain provide
information about the specific conformation of the guest’s alkyl chain. The actual
conformation adopted by the alkyl chain can be described by two extremes: fully
extended for the longest chain, or helical coiling for the shortest chain. In a fully
Figure 4.6. MM3 minimized structural model of the [RuC10⊂Ga4L6]12- host-guest complex.
- 67 -
extended chain, protons on carbon C(i) will only show NOE interactions with C(i+2)
and/or C(i-2), since all dihedral angles are 180°. Helical coiling leads to additional NOE
interactions such as C(i+3) and C(i+4) due to the gauche orientations of the chain. Such
gauche interactions were observed by Rebek and co-workers upon encapsulation of n-
alkanes by a cylindrical supramolecular host, particularly when the length of the extended
alkane exceeded the dimensions of the dimer’s interior cavity.9 For the [RuCn⊂Ga4L6]12-
system, the only intra-chain cross peaks observed are from C(i ± 1) methylenes and
diastereotopic geminal protons bound to the same carbon atom, indicating the alkyl chain
is in the extended conformation. The chain is probably drawn into solution by the
sulfonate anion, extended away from the anionic host.
Guest binding affinities
Competitive binding experiments with Et4N+ were carried out to evaluate the
binding affinities of RuCn in D2O. Equimolar amounts of RuCn, Et4N+, and [Ga4L6]
12-
were combined in D2O and the system was allowed to equilibrate overnight. With Kref
known (for Et4N+ in D2O with no added KCl, Kref = 1.96 x 104 M-1),8 the unknown
binding constant Kn can be obtained from the relative concentrations of the two host-
guest complexes from the following equation:
Kn = Kref
µ
[RuCn ½ Ga4L6]
[Et4N ½ Ga4L6]
¶2
Kn = Kref
µ
[RuCn ½ Ga4L6]
[Et4N ½ Ga4L6]
¶2
(4.1)
where n refers to the length of the guest’s alkyl chain. The results are summarized in
Table 4.1.
- 68 -
Table 4.1. Host-guest binding equilibrium constants for [RuCn⊂Ga4L6]12- in D2O, from Et4N
+ competitive binding experiments.
Guest Kn, M-1 log10 Kn
RuC4 (≤ 103) ---
RuC6 1.7 x 103 3.2
RuC8 8.7 x 103 3.9
RuC10 6.9 x 103 3.8
No signal for [RuC4⊂Ga4L6]12- was observed by NMR when 1 equivalent of
Et4N+ was added, so a value for K4 could not be obtained from these experiments.
However, these data clearly establish the relative ordering for binding affinities:
K4 < K6 < K10 ~ K8
For formation of a host-guest complex with the M4L6 cluster, the optimal chain length for
[RuCn⊂Ga4L6]12- is eight carbons. As the chain length increases from n = 4 to n = 8, the
sulfonate anion can move farther away from the anionic host to minimize coulombic
repulsion, lowering the energy of the host-guest complex.
Mass spectrometry
Host-guest complex formation was also confirmed by high resolution negative ion
electrospray mass spectrometry (ESI-MS) (Figure 4.7). Spectra were obtained for
solutions of all four [RuCn⊂Ga4L6]12- systems, showing peaks for the z = -3 and -4 charge
states of the host-guest complexes with K+, Na+, and/or H+ counterions. The mass
spectra of the [RuC4⊂Ga4L6]12- and [RuC10⊂Ga4L6]
12- systems show additional peaks that
correspond tot the z = -5 charge state of the host-guest complex.
- 69 -
The above study shows that the [Ga4L6]12- tetrahedron remains intact upon
incorporation of the cationic head of a zwitterion, while the alkyl chain attached to the
anionic tail passes through one of the small openings at the centers of the triangular faces.
This process is fully consistent with the nondissociative guest egress mechanism
described by Davis and Raymond in a previous publication.1
Monocations with Pendant Alkanes
When the cationic head group of RuCn is encapsulated, coulombic repulsion with
the anionic host requires the sulfonate anion at the opposite end of the hydrocarbon chain
to remain in solution. This design feature forced the hydrocarbon to be permanently
extruded through one aperture in the cluster as long as the cation remained bound to the
Figure 4.7. Portion of the ESI MS of the [RuC10⊂Ga4L6]12- system that shows two adjacent peaks for the
z = -4 charge state, with predicted isotopic distribution patterns for two particular fragment ion formulae.
- 70 -
host interior.3 What happens if the sulfonate group were replaced by a simple hydrogen
atom, so the pendant chain terminates with a neutral methyl group? To answer this
question, the series of cations RuAn+ was prepared (n = 4, 6, 8, 10, 12) by combining the
commercially available n-phenylalkanes with [CpRu(CH3CN)3]PF6 in a manner similar
to the RuCn synthesis. By eliminating repulsion from the sulfonate anion, the neutral
chain is free to completely retract inside the host, as long as there is enough space to
accommodate it.
When equimolar amounts of solid [RuAn]PF6 and K12[Ga4L6] were combined in
D2O (n = 4, 6, 8, 10), homogeneous solutions were obtained despite the very low water
solubilities of [RuAn]PF6 , particularly for longer chain lengths. This enhanced solubility
is due to encapsulation of the cationic ruthenium center by the water soluble [Ga4L6]12-
host, with formation of [RuAn⊂Ga4L6]11- readily apparent from the large upfield shifts of
the guests’ Cp and phenyl resonances in the 1H NMR spectra (Figure 4.8).3, 6 The
diastereotopic splitting patterns for the five phenyl proton resonances observed upon
encapsulation of RuAn+ are very similar to those for [RuCn⊂Ga4L6]
12-. For RuA12+,
MeOD was used instead of D2O to combine the two solids, but similar upfield shifts
observed upon encapsulation of the head group confirmed quantitative formation of
[RuA12⊂Ga4L6]11- in MeOD. Host-guest complex formation was confirmed via ESI MS
for all chain lengths (Appendix).
In contrast to the RuCn system, the point symmetry of the [RuAn⊂Ga4L6]11- host-
guest complexes is chain length dependent. For long side chains (n ≥ 8), the entire alkane
cannot fit inside the cavity and remains extruded through one of the host’s faces, with C3
symmetry observed for the host-guest complexes. With a short side chain (n = 4, RuA4+),
- 71 -
the entire guest is completely encapsulated by the host, and rapid tumbling of the guest
within the cavity leads to overall T symmetry for the [RuA4⊂Ga4L6]11- host-guest
complex. With an intermediate chain length (n = 6, RuA6+), the side chain rapidly
Figure 4.8. Guest region of the 1H NMR spectra (500 MHz) of [RuAn⊂Ga4L6]11- for the different chain
lengths n. Samples were dissolved in D2O for all spectra except for n = 12 (CD3OD). Selected assignments for the guest protons are shown in blue, with the regions where different groups of protons resonate labeled as appropriate. A label printed in one spectrum applies to all those below it as well.
- 72 -
extends and retracts at room temperature such that it dynamically protrudes through all
four triangular faces of [Ga4L6]12- on the NMR timescale (Figure 4.9). During this
dynamic process, the side chain retracts such that RuA6+ is fully encapsulated long
enough so that its tumbling averages out to T symmetry for the host-guest complex.
Lowering the temperature reveals that the ground state configuration has C3 symmetry,
with the alkyl chain fully extended and protruding through one of the host’s four
triangular faces. Since all four apertures are identical in the retracted, T-symmetric state,
there are four degenerate C3 states. Rapid interconversion between C3 states on the NMR
timescale leads to time-averaged T symmetry at higher temperatures.
The dynamic nature of the [RuA6⊂Ga4L6]11- system is apparent by monitoring the
host region during variable temperature 1H NMR measurements. At room temperature,
Figure 4.9. While the cationic RuII sandwich complex (red ball at top) of RuA6+ remains within the
[Ga4L6]12- cluster, the pendant alkyl chain rapidly retracts and extends at room temperature, moving the
terminal methyl group (black ball) inside and outside the host cavity. The higher energy retracted state exhibits T symmetry, and the four degenerate C3-symmetric states are lower in energy.
- 73 -
six signals – two broad and four sharp – integrating to 12H each are observed for the 72
aromatic host protons, consistent with point group T (Figure 4.10). At low temperature
(T < -40 °C), the host resonances split into 24 overlapping signals integrating to 3H each,
with the 2D COSY spectrum at -60 °C (Appendix) consistent with a C3-symmetric host-
guest complex whose guest side chain protrudes from one of the face apertures. At high
temperature (T ≥ 50 °C), the host signals coalesce into six sharp peaks.
Coalescence of the host resonances for [RuA8⊂Ga4L6]11- was observed at a much
higher temperature, with six broad host resonances observed at 75 °C (Figure 4.11). This
demonstrates that the longer eight carbon chain can fully retract into the cavity, but the
Figure 4.10. Aromatic region of the 1H NMR spectra for [RuA6⊂Ga4L6]11- showing the signals for the
host’s 72 aromatic protons observed at a) -60 °C in MeOD, b) 20 °C in D2O, and c) 50 °C in D2O.
- 74 -
high coalescence temperature indicates that the T state is much higher in energy than the
C3 ground state with the eight carbon chain, most likely due to the limited space of the
host cavity.
The modest upfield shift observed for the CH3 triplet (Figure 4.8) depends on the
percentage of time the alkyl chain spends retracted (T state) vs. extended (C3 state), since
the magnetically shielded environment of the host cavity leads to an upfield shift for
interior protons. For [RuA6⊂Ga4L6]11-, conversion between C3 and T states at room
temperature is fast on the NMR timescale, and the observed chemical shift for the CH3
Figure 4.11. Aromatic region of the 1H NMR spectra for [RuA8⊂Ga4L6]11- (D2O, 500 MHz) showing the
signals for the host’s 72 aromatic protons observed at a) room temperature, b) 50 °C, and c) 75 °C.
- 75 -
signal δobs is a population average of the chemical shifts for the C3 and T states, denoted
δC3 and δT, respectively:10
±obs = ±C3xC3
+ ±TxT
= ±C3+ (±T ¡ ±C3
)xT
±obs = ±C3xC3
+ ±TxT
= ±C3+ (±T ¡ ±C3
)xT (4.2)
where xC3xC3
and xTxT denote the fractional populations for the two states, with xC3+ xT ´ 1xC3+ xT ´ 1.
Substitution and rearrangement leads to the following expression for xTxT :
xT =±obs ¡ ±C3
±T ¡ ±C3
xT =±obs ¡ ±C3
±T ¡ ±C3 (4.3)
With [RuA4⊂Ga4L6]11- and [RuA10⊂Ga4L6]
11- as systems in the T and C3 states, values
for δT = -0.116 ppm and δC3 = 1.01 ppm can be estimated from the methyl triplet in their
1H NMR spectra. From Equation 4.3, with δobs = 0.741 ppm, the [RuA6⊂Ga4L6]11- system
exists in the T state (retracted chain) for 24% of the time, and exists in the C3 state
(extended chain) the remaining 76% of the time. This corresponds to an equilibrium
constant K = [T]/[C3] = 1/3, which means the energy difference between the T and C3
states is about 0.6 kcal mol-1 (Figure 4.9).
The 2D NOESY spectrum of [RuA6⊂Ga4L6]11- at 27 °C shows a cross peak
between the guest’s terminal methyl group and one of the host’s catechol protons
(Appendix). Because the catechol protons point away from the interior of [Ga4L6]12-, this
confirms that, on average, the methyl group is located outside the host cavity. Similar
NOE interactions were observed with the RuCn zwitterions, where a host catechol
resonance showed cross peaks with signals from the unencapsulated portion of the
guest’s alkylsulfonate chain.3 In contrast, fully encapsulated guests such as RuA4+ and
Et4N+ only show NOE cross peaks with the three naphthalene signals.6
- 76 -
The Second-Order Jahn Teller Effect in [RuA6⊂Ga4L6]11-
The preceding observations demonstrated that the [RuA6⊂Ga4L6]11- system
rapidly oscillates between four degenerate C3 ground states through a T-symmetric
intermediate at room temperature (Figure 4.9). The reaction mechanism consequences of
such a dynamical system have been described by Pearson as based on the second-order
Jahn-Teller effect.11 The original Jahn-Teller theorem applies to specific types of
degenerate electronic systems.12 Chemists are generally familiar with the first-order
Jahn-Teller effect explaining, for example, the elongated axial bond distances in pseudo-
octahedral copper(II) complexes.13 An often cited example of a dynamic second-order
Jahn-Teller distortion is XeF6 in the gas phase.14 Its sterically active lone pair distorts
one face of the octahedron to yield a C3v ground state, yet is in dynamic exchange
through an essentially octahedral intermediate on its way to the other seven degenerate
C3v structures, with the lone pair occupying the other seven faces in the octahedron.15
Most studies concerning second-order Jahn-Teller effects involve bond distortions or
ruptures within a single covalent network.16 However, the fluxional structure for
[RuA6⊂Ga4L6]11- described above represents the first reported example of a second-order
Jahn-Teller distortion in a supramolecular host-guest system.4
Following the description of Pearson, after a displacement q = Q – Q0, the
Hamiltonian may be expanded as a Taylor-Maclaurin series as a function of the reaction
coordinate Q:
H = H0 +
µ
@U
@Q
¶
q +1
2
µ
@2U
@Q2
¶
q2 + ¢ ¢ ¢H = H0 +
µ
@U
@Q
¶
q +1
2
µ
@2U
@Q2
¶
q2 + ¢ ¢ ¢
(4.4)
with U equal to the potential energy of the system. From second-order perturbation
theory, the energy E becomes:
- 77 -
E = E0 +
¿
Ã0
¯
¯
¯
¯
@U
@Q
¯
¯
¯
¯
Ã0
À
q +
¿
Ã0
¯
¯
¯
¯
@2U
@Q2
¯
¯
¯
¯
Ã0
À
q2
2+
X
k>0
D
Ã0
¯
¯
¯
@U@Q
¯
¯
¯Ãk
E2
q2
E0 ¡ Ek
E = E0 +
¿
Ã0
¯
¯
¯
¯
@U
@Q
¯
¯
¯
¯
Ã0
À
q +
¿
Ã0
¯
¯
¯
¯
@2U
@Q2
¯
¯
¯
¯
Ã0
À
q2
2+
X
k>0
D
Ã0
¯
¯
¯
@U@Q
¯
¯
¯Ãk
E2
q2
E0 ¡ Ek
(4.5)
where E0 is the original energy at point Q0 (q = 0), Ã0Ã0 is the electronic wave function for
the ground state, and ÃkÃk is the wave function for the kth excited state (k = 0, 1, 2, …).11
Although Figure 4.9 is a cartoon diagram of the energy of the host-guest complex,
it serves to illustrate Equation 4.5. At maxima or minima, the slope (@U=@Q) = 0(@U=@Q) = 0, and
the term linear in q vanishes. For Q0 defined as the C3 ground state with a fully extended
chain, E = E0E = E0, and the force constant (@2U=@Q2) > 0(@2U=@Q2) > 0. When Q is at the transition state,
with the chain partially retracted, E > E0E > E0, and the force constant (@2U=@Q2) < 0(@2U=@Q2) < 0. When
Q is at T, with the chain fully retracted inside the cavity, ET > E0ET > E0 by 0.6 kcal mol-1, but
now (@2U=@Q2) > 0(@2U=@Q2) > 0 because the system is at a local minimum. The last term in
Equation 4.5 is negative, and mixing excited states with Ã0Ã0 lowers the energy when Q
approaches T from the C3 transition state. These are the descriptors of a second-order
Jahn-Teller distortion.
At low temperatures, the system does not have enough thermal energy to
sufficiently populate the appropriate excited states to overcome the larger activation
barrier ¢E1¢E1 (defined in Figure 4.9), and the guest’s alkyl chain remains extruded through
a single face. At higher temperatures, there is a greater statistical population of excited
electronic states ÃkÃk, permitting distortion to the transition state towards the T-symmetric
intermediate. The system can easily overcome the smaller activation barrier ¢E2¢E2 to
return to one of the C3 ground states. At 50 °C, this distortion is very fast, allowing the
system to rapidly interconvert between the four degenerate C3 states through the T
intermediate state, leading to the T time-averaged symmetry on the NMR timescale.
- 78 -
Summary
The development of the novel partial guest encapsulation mode illustrates the
power of rational design. Once the nondissociative guest egress mechanism for the M4L6
host was established, a set of simple design principles could be developed to take
advantage of its predictions. By deliberately creating the partial guest encapsulation
mode, complex interactions could be studied and carefully controlled simply by changing
the length and terminus of a linear hydrocarbon. The flexible apertures of the M4L6
tetrahedron might be used for future applications such as controlling catalytic polymer
growth to completely prevent branching, or linking two M4L6 clusters to form a
supramolecular “dumbbell.”
Experimental
General considerations
All reagents were obtained from commercial suppliers and used without further
purification unless noted otherwise. Standard Schlenk techniques were used for reactions
carried out under argon, and a glove box continuously purged with nitrogen was used to
store the moderately air sensitive cluster. When necessary, solvents for reactions were
degassed by at least six pump/fill cycles while vigorously stirring, backfilling with argon.
H4L (H4L = 1,5-bis(2,3-dihydroxybenzamido)naphthalene) and K12[Ga4L6] were
synthesized according to literature procedure.7, 8 NMR spectra of host-guest complexes
were measured on a Bruker AV-500 spectrometer with a TBI-P probe featuring a high-
sensitivity inner 1H coil. Chemical shifts δ vs. SiMe4 are listed for 1H NMR spectra, and
vs. CFCl3 for 19F NMR spectra. Routine mass spectrometry and elemental analysis was
- 79 -
performed by the Mass Spectrometry Laboratory and Microanalysis Facility in the
College of Chemistry at the University of California, Berkeley. High resolution TOF
electrospray mass spectra (ESI-MS) were recorded on a Waters QTOF API mass
spectrometer equipped with a Z-spray source, located either at the UC Berkeley Mass
Spectrometry facility (RuAn+ complexes) or the Waters Corporation, Dublin, CA (RuCn
complexes).
Guest synthetic procedures
1-bromo-4-phenylbutane (C4Br). A stirred solution of 1.0 g (6.7 mmol) of 4-phenyl-1-
butanol and 3 mL of pyridine in 75 mL of CH2Cl2 was cooled to -10 °C in an ice/salt
bath. An addition funnel containing 6 g (20 mmol) of phosphorous tribromide dissolved
in 25 mL of CH2Cl2 was affixed to the reaction vessel, and this solution was added
dropwise to the stirred solution over a 30 min period. The reaction mixture was stirred
for 15 h at room temperature, filtered to remove an orange solid, then washed with dilute
brine (2 x 300 mL), dilute sulfuric acid (250 mL), 1 M hydrochloric acid (2 x 250 mL),
and concentrated brine (250 mL). The organic fraction was collected, dried over MgSO4,
and the solvent removed via rotary evaporation, yielding a yellowish oil. Purification via
chromatography (basic alumina, CH2Cl2) and removal of solvent yielded 0.50 g (36%) of
a clear oil. 1H NMR (CDCl3, 400 MHz): δ 1.77 (m, 2H), 1.9 (m, 2), 2.65 (t, 2, J=7.6 Hz),
3.42 (t, 2, J=6.8 Hz), 7.2 (m, 5).
Sodium(4-phenylbutane-1-sulfonate) (C4SO3Na).17 A solution of 1.0 g (8.0 mmol) of
sodium sulfite in 20 mL of H2O was added to 0.50 g (2.3 mmol) of C4Br. The reaction
- 80 -
mixture was heated at reflux for 24 h, cooled to room temperature, and filtered to collect
the white crystals, washing once with cold H2O. The volume of the filtrate was reduced,
cooled to 4 °C, and the white crystalline solid was collected on a frit and washed with
cold H2O, then dried for 24 h in vacuo at 50 °C to yield 570 mg (96%) of white solid.
Anal. Calc. (found) for C10H13NaO3S: %C, 50.84 (50.45); H, 5.55 (5.31). MS(FAB+):
m/z 237 ([MNaH]+), 259 ([MNa2]+). 1H NMR (D2O, 300 MHz): δ 1.73 (m, 4H), 2.67 (t,
J=7.1 Hz, 2H), 2.93 (t, J=7.5 Hz, 2H), 7.3 (m, 5H).
1-bromo-6-phenylhexane (C6Br). A stirred solution of 2.52 g (14.1 mmol) of 6-phenyl-
1-hexanol in 80 mL of Et2O was cooled to 0 °C in an ice bath. An addition funnel
containing 4 g (20 mmol) of phosphorous tribromide in 25 mL of Et2O was affixed to the
reaction vessel, and this solution was added dropwise to the stirred solution over a 30 min
period. The reaction mixture was stirred for 15 h, allowing the ice bath to gradually
return to room temperature. The ether solution was treated with aqueous sodium
bicarbonate (2 x 250 mL) followed by brine (100 mL). The organic layer was collected,
dried over MgSO4, and the solvent removed via rotary evaporation to yield a cloudy,
colorless oil. Purification via chromatography (basic alumina, CH2Cl2) and removal of
solvent yielded 1.16 g (34%) of a clear oil. 1H NMR (CDCl3, 400 MHz): δ 1.39 (m, 2H),
1.49 (m, 2H), 1.66 (m, 2H), 1.88 (m, 2H), 2.64 (t, J = 7.6 Hz, 2H), 3.42 (t, J = 6.8 Hz,
2H), 7.21 (m, 3H), 7.31 (m, 2H).
Sodium(6-phenylhexane-1-sulfonate) (C6SO3Na).17 A solution of 2.1 g (15 mmol) of
sodium sulfite in 50 mL of H2O/EtOH (4:1) was added to 1.16 g (4.81 mmol) of C6Br.
- 81 -
The reaction mixture was heated at reflux for 24 h, cooled to 4 °C, and filtered to collect
the white crystals. The solid was dried for 24 h in vacuo at 50 °C to yield 1.2 g (96%) of
white solid. Anal. Calc. (found) for C12H17NaO3S·(H2O)0.5: %C, 52.73 (52.21); H, 6.64
(6.28). MS(ES-): m/z 241.1 (M-), 483.3 ([HM2]-), 505.2 ([NaM2]
-). 1H NMR (MeOD,
300 MHz): δ 1.39 (m, 4H), 1.62 (m, 2), 1.77 (m, 2), 2.60 (t, 2, J=7.8 Hz), 2.76 (t, 2,
J=8.1 Hz), 7.17 (m, 5).
1-bromo-8-phenyloctane (C8Br). A stirred solution of 2.87 g (13.9 mmol) of 8-phenyl-
1-octanol in 75 mL of anhydrous Et2O was cooled to 0 °C under air in an ice/salt bath. A
solution of 3 g (10 mmol) of phosphorous tribromide in 15 mL of Et2O was added
dropwise over 30 min, and the reaction mixture was stirred for 90 additional min at 0 °C.
The reaction mixture was treated with aqueous NaHCO3, the organic layer collected and
dried with MgSO4, and the solvent removed to yield a cloudy colorless oil. Purification
via chromatography (basic alumina, CH2Cl2) and removal of solvent yielded 1.0 g (27%)
of a clear, colorless oil. 1H NMR (CDCl3, 400 MHz): δ 1.38 (br s, 8H), 1.65 (m, 4H),
2.66 (t, 2H, J=7.6 Hz), 3.46 (t, 2H, J=6.8 Hz), 7.3 (m, 5H).
Sodium(8-phenyloctane-1-sulfonate) (C8SO3Na).17 To 1 g of C8Br was added a
solution of 1.5 g (12 mmol) of sodium sulfite dissolved in 40 mL of H2O with 10 mL of
ethanol. The reaction mixture was heated at reflux under air for 24 hours, then allowed to
cool to room temperature. A white polycrystalline solid formed, which was collected on
a frit and washed with cold H2O (2 x 5 mL and isopropyl alcohol (2 x 20 mL), then dried
overnight to yield 0.50 g (46%) of shiny white flakes. Anal. Calc. (found) for
- 82 -
C14H21NaO3S: %C, 57.51 (57.15); H, 7.24 (7.41); S, 10.97 (10.90). MS(ES-): m/z 269.1
(M-). 1H NMR (D2O, 300 MHz): δ 1.3 (m, 8H), 1.57 (m, 2H), 1.67 (m, 2H), 2.59 (t,
J=7.8 Hz, 2H), 2.84 (t, J=8.0 Hz, 2H), 7.3 (m, 5H).
1-bromo-10-phenyldecane (C10Br). A stirred solution of 2.56 g (10.9 mmol) of 10-
phenyl-1-decanol in 75 mL of anhydrous Et2O was cooled to 0 °C under air in an ice
bath. An addition funnel containing 3 g (10 mmol) of phosphorous tribromide in 25 mL
of Et2O was affixed to the reaction vessel, and this solution was added dropwise to the
stirred solution over a 30 min period. The reaction mixture was stirred for 90 min, then
treated with aqueous sodium bicarbonate (250 mL). The organic layer was collected,
dried over MgSO4, and its solvent removed to yield a cloudy, colorless oil. Purification
via chromatography (silica, CH2Cl2) yielded 0.85 g (25%) of a colorless oil. 1H NMR
(CDCl3, 400 MHz): δ 1.30 (br s, 10H), 1.43 (m, 2H), 1.62 (m, 2H), 1.86 (m, 2H), 2.615
(t, 2H, J=7.6 Hz), 3.421 (t, 2H, J=6.8 Hz), 7.24 (m, 5H).
Sodium(10-phenyldecane-1-sulfonate) (C10SO3Na).17 To 0.84 g (2.8 mmol) of C10Br
was added a solution of 1.2 g (8.4 mmol) of sodium sulfite dissolved in 35 mL of H2O
with 10 mL of ethanol. The reaction mixture was heated at reflux under air for 24 h, then
allowed to cool to room temperature. A white polycrystalline solid formed, which was
collected on a frit and washed with 10 mL of cold H2O, then for two days at 60 °C in a
vacuum oven to yield 0.76 g (84%) of shiny white flakes. Anal. Calc. (found) for
C16H25NaO3S: %C, 59.97 (60.16); H, 7.86 (7.73). MS(ES-): m/z 297.2 (M-). 1H NMR
- 83 -
(MeOD, 400 MHz): 1.35 (m, 12H), 1.60 (m, 2H), 1.78 (m, 2H), 2.59 (t, J=7.6 Hz, 2H),
2.78 (t, J=8.0 Hz, 2H), 7.2 (m, 5H).
[Ru(η5-C5H5)(η6-C6H5(CH2)10SO3)]·NaPF6 (RuC10·NaPF6). This reaction was
performed under argon using standard Schlenk techniques. To 160 mg (0.49 mmol) of
C10SO3Na and 200 mg (0.46 mmol) of [Ru(η5-C5H5)(CH3CN)3]PF6 was added 100 mL of
anhydrous CHCl3. The reaction mixture was heated at reflux for 24 h, during which time
a brown solid appeared on the walls of the flask. The solvent was removed, and the light
brown residue was dissolved in methanol, filtered, and precipitated with ether to yield
150 mg (52%) of beige solid. Anal. Calc. (found) for C21H30F6NaO3PRuS: %C, 39.94
(39.05); H, 4.79 (4.73); S, 5.08 (4.82). MS(ES+): m/z 465.1 ([MH]+), 487.1 ([MNa]+).
1H NMR (MeOD, 500 MHz): δ 1.34 (s, 6H), 1.40 (m, 6H), 1.63 (m, 2H), 1.78 (m, 2H),
2.53 (t, J=8.0 Hz, 2H), 2.78 (t, J=8.0 Hz, 2H), 5.41 (s, 5H), 6.13 (m, 1H), 6.17 (m, 2H),
6.23 (d, J=6.0 Hz, 2H). 1H NMR (D2O, 300 MHz): δ 1.32 (m, 12H), 1.58 (m, 2H), 1.71
(m, 2H), 2.48 (t, J=7.8 Hz, 2H), 2.87 (t, J=8.0 Hz, 2H), 5.33 (s, 5H), 6.07 (m, 3H), 6.14
(m, 2H). 19F NMR (MeOD, 376 MHz): δ -74.0 ppm vs. CFCl3 (d, J=709 Hz).
[Ru(η5-C5H5)(η6-C6H5(CH2)4SO3)]·NaPF6 (RuC4·NaPF6). A procedure similar to the
synthesis of RuC10·NaPF6 was used, with 127 mg (0.278 mmol) of [Ru(η5-
C5H5)(CH3CN)3]PF6 and 71 mg (0.28 mmol) of C4SO3Na in place of C10SO3Na.
Recrystallization from methanol/ether yielded 60 mg (40%) of beige powder. Anal.
Calc. (found) for C15H18F6NaO3PRuS: %C, 32.91 (32.35); H, 3.31 (3.55); S, 5.86 (5.50).
MS(ES+): m/z 381.1 ([MH]+), 403.1 ([MNa]+). 1H NMR (D2O, 500 MHz): δ 1.74 (m,
- 84 -
2H), 1.81 (m, 2H), 2.55 (t, J=7.7 Hz, 2H), 2.94 (t, J=7.5 Hz, 2H), 5.35 (s, 5H), 6.08 (m,
3H), 6.18 (d,J=5.5 Hz, 2H).
[Ru(η5-C5H5)(η6-C6H5(CH2)6SO3)]·NaPF6 (RuC6·NaPF6). A procedure similar to the
synthesis of RuC10·NaPF6 was used, with 215 mg (0.495 mmol) of [Ru(η5-
C5H5)(CH3CN)3]PF6 and 140 mg (0.50 mmol) of C6SO3Na in place of C10SO3Na.
Recrystallization from methanol/ether yielded 150 mg (53%) of beige powder. Anal.
Calc. (found) for C17H22F6NaO3PRuS: %C, 35.48 (35.30); H, 3.85 (4.15); S, 5.57 (5.29).
MS(ES+): m/z 409.1 ([MH]+), 431.1 ([MNa]+). 1H NMR (D2O, 400 MHz): δ 1.43 (m,
4H), 1.61 (m, 2H), 1.72 (m, 2H), 2.50 (t, J=7.8 Hz, 2H), 2.89 (t, J=8.0 Hz, 2H), 5.34 (s,
5H), 6.06 (m, 3H), 6.16 (d, J= 5.6 Hz, 2H).
[Ru(η5-C5H5)(η6-C6H5(CH2)8SO3)]·NaPF6 (RuC8·NaPF6). A procedure similar to the
synthesis of RuC10·NaPF6 was used, with 185 mg (0.425 mmol) of [Ru(η5-
C5H5)(CH3CN)3]PF6 and 125 mg (0.425 mmol) of C8SO3Na in place of C10SO3Na.
Recrystallization from methanol/ether yielded 170 mg (66%) of beige powder. Anal.
Calc. (found) for C19H26F6NaO3PRuS: %C, 37.81 (37.13); H, 4.34 (4.56); S, 5.31 (5.28).
MS(ES+): m/z 437.1 ([MH]+), 459.1 ([MNa]+). 1H NMR (D2O, 400 MHz): δ 1.38 (m,
8H), 1.60 (m, 2H), 1.71 (m, 2H), 2.49 (t, J=7.8 Hz, 2H), 2.88 (t, J=8.0 Hz, 2H), 5.34 (s,
5H), 6.07 (m, 3H), 6.16 (d, J=5.6 Hz, 2H). 19F NMR (MeOD, 376 MHz): δ -71.3 ppm
vs. CFCl3 (d, J=709 Hz).
- 85 -
[(cyclopentadienyl)(n-phenylbutane)ruthenium(II)]hexafluorophosphate
([RuA4]PF6). The following reaction was carried out under argon. To 95 mg
(0.22 mmol) of [CpRu(CH3CN)3]PF6 (Strem, 98%) was added 50 mL anhydrous
chloroform (Aldrich, 99%, stabilized with amylenes) via cannula, then 70 µL (60 mg,
0.5 mmol) of n-butylbenzene (TCI) was added to the orange solution via syringe. An
additional 5 mL of chloroform was added to wash any stray reactant from the walls of the
flask. After heating the reaction mixture at 75 °C for 24 h, the solvent was removed via
rotary evaporation. The residue was re-dissolved under air in 15 mL of dichloromethane,
filtered through Celite, and after washing the Celite plug with ~100 mL of
dichloromethane, the solvent was removed again via rotary evaporation. The residue was
re-dissolved in a minimum volume (~5 mL) of fresh dichloromethane, and 50 mL of
diethyl ether was added to this yellow-brown solution to form a precipitate. The solid
was collected on a frit, washed with diethyl ether (3 x 15 mL) and petroleum ether
(2 x 10 mL), and dried in vacuo for 6 h to yield 70 mg (72%) of off-white powder. Anal.
Calc. (found) for C15H19F6PRu: %C, 40.45 (40.35); H, 4.30 (4.44). MS (ESI+, MeOH):
m/z 301.1 (RuA4+). 1H NMR (CDCl3, 400 MHz): δ 0.98 (t, 3H, J=7.2 Hz), 1.44 (m, 2H),
1.6 (m, 2H), 2.54 (t, 2H, J=7.6 Hz), 5.40 (s, 5H), 6.15 (m, 5H).
[(cyclopentadienyl)(n-phenylhexane)ruthenium(II)]hexafluorophosphate
([RuA6]PF6). The following reaction was carried out under argon. To 98 mg
(0.22 mmol) of [CpRu(CH3CN)3]PF6 was added 40 mL of anhydrous chloroform and
150 µL (130 mg, 0.80 mmol) of n-hexylbenzene. After stirring at room temperature for
2 days, a mustard yellow solution with a ring of dark brown solid was present in the
- 86 -
reaction vessel. The reaction mixture was filtered through Celite under air, and after
thoroughly washing the Celite plug with 150 mL of dichloromethane, the solvent was
removed via rotary evaporation, leaving a dark, viscous, mustard colored oil in the flask.
This oil was dissolved in 3 mL of dichloromethane, and addition of 50 mL of diethyl
ether led to phase separation after 20 min, with droplets of mustard colored oil at the
bottom of the flask. After adding an additional 50 mL of diethyl ether, the biphasic
mixture was vigorously stirred for 45 min to afford a light-colored solid precipitate. The
solid was collected on a medium frit, washed with diethyl ether (3 x 15 mL), and dried in
vacuo for 4 h to yield 55 mg (52%) of pale tan solid. Anal. Calc. (found) for
C17H23F3PRu: %C, 43.13 (42.83); H, 4.90 (4.79). MS (ESI+, MeOH): m/z 329.2
(RuA6+). 1H NMR (CDCl3, 300 MHz): δ 0.87 (t, 3H, J=6.6 Hz), 1.3 (m, 6H), 1.6 (m,
2H), 2.47 (t, 2H, J=7.8 Hz), 5.35 (s, 5H), 6.1 (m, 5H).
[(cyclopentadienyl)(n-phenyloctane)ruthenium(II)]hexafluorophosphate
([RuA8]PF6). The following reaction was carried out under argon. To 98 mg
(0.23 mmol) of [CpRu(CH3CN)3]PF6 was added 50 mL of anhydrous chloroform via
cannula, then 1.0 mL (0.86 g, 4.5 mmol) of n-octylbenzene was added to the orange
solution via syringe. After heating the reaction mixture at reflux for 18 h, the reaction
mixture was allowed to cool, filtered through Celite, and the plug was washed with
CH2Cl2 to obtain a yellow-brown solution. Solvent was removed via rotary evaporation,
leaving a biphasic oily residue: small yellow-brown drops within a clear oily liquid. The
residue was dissolved in 5 mL of CH2Cl2, and 50 mL of ether was added and stirred for
one hour, forming a silky precipitate. The powder was collected on a frit, washed with
- 87 -
ether (2 x 20 mL), and dried in vacuo for three days to yield 65 mg (58%) of off-white
powder. Anal. Calc. (found) for C19H27F6PRu: %C, 45.51 (45.39); H, 5.43 (5.36). MS
(ESI+, MeOH): m/z 357.1 (RuA8+). 1H NMR (CDCl3, 300 MHz): δ 0.87 (t, 3H,
J=6.6 Hz), 1.3 (m, 10H), 1.57 (m, 2H), 2.48 (t, 2H, J=7.8 Hz), 5.36 (s, 5H), 6.1 (m, 5H).
[(cyclopentadienyl)(n-phenyldecane)ruthenium(II)]hexafluorophosphate
([RuA10]PF6). The title compound was prepared using the same procedure described for
[RuA8]PF6, with 100 mg (0.23 mmol) of [CpRu(CH3CN)3]PF6, 0.40 mL (0.34 g,
1.6 mmol) of n-phenyldecane, and 40 mL of anhydrous CHCl3. Recrystallization of the
product from CH2Cl2/ether yielded 50 mg (41%) of off-white powder. Anal. Calc.
(found) for C21H31RuPF6: %C, 47.63 (47.32); H, 5.90 (5.88). MS (ESI+, MeOH):
m/z 385.1 (RuA10+). 1H NMR (CDCl3, 400 MHz): δ 0.86 (t, 3H, J=6.2 Hz), 1.24 (m,
12H), 1.32 (m, 2H), 1.56 (m, 2H), 2.45 (t, 2H, J=7.8 Hz), 5.34 (s, 5H), 6.05 (m, 2H), 6.11
(m, 3H).
[(cyclopentadienyl)(n-phenyldodecane)ruthenium(II)]hexafluorophosphate
([RuA12]PF6). The following reaction was carried out under argon. To 99 mg
(0.23 mmol) of [CpRu(CH3CN)3]PF6 was added 40 mL of anhydrous 1,2-dichloroethane
via cannula, then 0.3 mL (0.3 g, 1 mmol) of n-dodecylbenzene was added to the orange
solution via syringe. The reaction mixture was heated at reflux for 5 h, then removed
from heat and stirred for an additional 12 h at room temperature. The brown reaction
mixture was filtered through Celite under air, and after thoroughly washing the Celite
plug with 100 mL of CH2Cl2, the solvent was removed via rotary evaporation. The
- 88 -
yellow-brown residue was dissolved in 3 mL of CH2Cl2, and addition of approximately
40 mL of diethyl ether afforded a light-colored precipitate. The solid was collected on a
fine frit, washed with diethyl ether (3 x 5 mL) and petroleum ether (2 x 15 mL), and dried
in vacuo overnight to yield 65 mg (50%) of off white powder. Anal. Calc. (found) for
C23H35RuPF6: %C, 49.55 (49.38); H, 6.33 (6.48). MS (FAB+): m/z 413 (RuA12+). 1H
NMR (CDCl3, 400 MHz): δ 0.86 (t, 3H, J=7.2 Hz), 1.23 (m, 16H), 1.32 (m, 2H), 1.56 (m,
2H), 2.46 (t, 2H, J=8.0 Hz), 5.34 (s, 5H), 6.06 (m, 2H), 6.12 (m, 3H).
Host-guest complex synthetic procedures
K12[RuC10⊂⊂⊂⊂Ga4L6]·(Me2CO)·7H2O. This reaction was performed under argon using
standard Schlenk techniques. A suspension of 102 mg (0.237 mmol) of H4L, 58 mg
(0.16 mmol) of Ga(acac)3, and 25 mg (0.040 mmol) of RuC10·NaPF6 in 75 mL of MeOH
was heated at reflux for 12 h. The solvent was removed to leave a light tan residue.
After drying this residue in vacuo for three hours, 50 mL of MeOH was added. Addition
of 0.95 mL (0.47 mmol) of methanolic KOH (0.5 M) caused the off-white suspension to
become a yellow solution. The reaction mixture was re-degassed with three pump/fill
cycles immediately after the addition of base, then stirred at room temperature for 2 h.
Undissolved solid was removed by cannula filtration. The volume was reduced to 5 mL,
and 150 mL of acetone was added to precipitate a pale yellow/brown solid. This solid
was collected on a frit under a stream of nitrogen, washed with acetone (4 x 10 mL), and
dried in vacuo overnight to yield 120 mg (80%) of yellow-beige powder. Anal. Calc.
(found) for C165H114Ga4K12N12O39RuS·Me2CO·7H2O: %C, 51.03 (51.01); H, 3.42 (3.29);
N, 4.25 (4.13). MS(ES-): (see text). 1H NMR (D2O, 500 MHz): δ -1.20 (m, 1H), -1.13
- 89 -
(br m, 1H), -1.02 (br m, 1H), -0.91 (br m, 1H), -0.25 (m, 2H), 0.29 (m, 1H), 0.52 (m, 1H),
0.66 (m, 1H), 0.80 (m, 1H), 1.18 (m, 2H), 1.36 (m, 2H), 1.49 (m, 2H), 1.87 (m, 2H), 2.29
(s, 5H), 3.05 (m, 3H), 3.14 (d, J=6.0 Hz, 1H), 3.20 (d, J=5.9 Hz, 1H), 3.28 (t, J=5.8 Hz,
1H), 4.35 (t, J=5.6 Hz, 1H), 6.49 (t, J=7.8 Hz, 3H), 6.58 (t, J=7.8 Hz, 3H), 6.64 (m, 9H),
6.74 (d, J=7.3 Hz, 3H), 6.79 (m, 6H), 6.82 (t, J=8.2 Hz, 3H), 6.90 (t, J=8.0 Hz, 3H), 7.17
(m, 6H), 7.23 (d, J=8.3 Hz, 3H), 7.27 (t, J=8.3 Hz, 3H), 7.36 (d, J=8.3 Hz, 3H), 7.39 (d,
J=8.3 Hz, 3H), 7.47 (d, J=7.6 Hz, 3H), 7.76 (d, J=8.6 Hz, 3H), 7.79 (d, J=8.6 Hz, 3H),
7.80 (d, J=8.6 Hz, 3H), 7.87 (d, J=7.8 Hz, 3H), 8.08 (d, J=8.7 Hz, 3H), 8.11 (d, J=7.8 Hz,
3H), 8.37 (d, J=7.8 Hz, 3H). 19F NMR (D2O, 376 MHz): No signal observed.
NMR Experiments of Host-Guest Complexes
General considerations. NMR spectra of host-guest complexes were acquired using a
Bruker AV-500 spectrometer with a TBI-P probe featuring a high-sensitivity inner 1H
coil. All 2D NMR spectra were measured at a constant, controlled temperature. For the
2D gCOSY spectrum of [RuA6⊂Ga4L6]11- measured at -60 °C in MeOD, a 12 second
delay was used between FID acquisition and the next pulse due to the very long 1H
relaxation times at low temperature.
Sample preparation: RuCn guests. K12[Ga4L6] (10.0 mg, 2.97 µmol) and RuCn·NaPF6
(3 µmol) were combined in a vial and dissolved in 0.6 mL of D2O at room temperature.
The solution was filtered through a glass fiber plug and transferred to an NMR tube, and
the spectrum recorded 10 min after dissolution. Solutions were discarded within 24 hours
due to slow oxidation of the ligands in the cluster.
- 90 -
Sample preparation: RuAn+ guests. Aqueous samples of host-guest complexes for
NMR spectroscopy were prepared using the same method, so the procedure used for
[RuA6⊂Ga4L6]11- is given as a representative example. In a threaded glass vial, 30.0 mg
(8.62 µmol) of K12[Ga4L6]·2(Me2CO) and 4.08 mg (8.62 µmol) of [RuA6]PF6 were
combined, corresponding to equimolar proportions. To this was added 2.0 mL of D2O,
and the mixture was sonicated for approximately 2 min to aid solvation. The clear yellow
solution was passed through a 0.2 µm nylon disk filter, and 0.5 mL of the filtrate was
transferred to a medium wall glass NMR tube. The solution was degassed in the tube via
three freeze-pump-thaw cycles, and flame sealed under vacuum. The sample was
allowed to equilibrate at room temperature for at least 12 h before measurements.
The NMR sample of [RuA6⊂Ga4L6]11- in MeOD used for low temperature studies
was prepared in a similar fashion, with 15.0 mg (4.31 µmol) of K12[Ga4L6]·2(Me2CO)
and 2.04 mg (4.31 µmol) of [RuA6]PF6 dissolved in 0.7 mL of MeOD after brief
sonication (< 1 min). The clear yellow solution was filtered through a 0.2 µm nylon disk
filter into a thin wall glass NMR tube. The solution was degassed and sealed under
vacuum as above, and allowed to equilibrate overnight at room temperature before use.
Spectra of [RuA6⊂Ga4L6]11- in D2O and MeOD were nearly identical, showing the same
host signal broadening behavior between 1 °C and 40 °C.
- 91 -
Mass spectrometry: Sample preparation
Samples for ESI-MS were prepared by combining equimolar amounts of RuCn·NaPF6 or
[RuAn]PF6 and K12[Ga4L6] in methanol or H2O/methanol, except for [RuC10⊂Ga4L6]12-
where the pre-synthesized host-guest complex was used.
References
1. Davis, A. V.; Raymond, K. N., “The Big Squeeze: Guest Exchange in an M4L6 Supramolecular Host.” J. Am. Chem. Soc. 2005, 127, 7912-7919.
2. Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.; Raymond, K. N., “Guest Exchange Dynamics in an M4L6 Tetrahedral Host.” J. Am. Chem. Soc. 2006, 128, 1324-1333.
3. Tiedemann, B. E. F.; Raymond, K. N., “Dangling Arms: A Tetrahedral Supramolecular Host with Partially Encapsulated Guests.” Angew. Chem. Int. Ed.
2006, 45, 83-86.
4. Tiedemann, B. E. F.; Raymond, K. N., “Second-Order Jahn-Teller Effect in a Host-Guest Complex.” Angew. Chem. Int. Ed. 2007, in press.
5. Fiedler, D.; Pagliero, D.; Brumaghim, J. L.; Bergman, R. G.; Raymond, K. N., “Encapsulation of Cationic Ruthenium Complexes into a Chiral Self-Assembled Cage.” Inorg. Chem. 2004, 43, 846-848.
6. Caulder, D. L.; Brückner, C.; Powers, R. E.; König, S.; Parac, T. N.; Leary, J. A.; Raymond, K. N., “Design, Formation, and Properties of Tetrahedral M4L4 and M4L6 Supramolecular Clusters.” J. Am. Chem. Soc. 2001, 123, 8923-8938.
7. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
8. Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am.
Chem. Soc. 1998, 120, 8003-8004.
9. Scarso, A.; Trembleau, L.; Rebek, J., Jr., “Helical Folding of Alkanes in a Self-Assembled, Cylindrical Capsule.” J. Am. Chem. Soc. 2004, 126, 13512-13518.
10. Macomber, R. S., A Complete Introduction to Modern NMR Spectroscopy. John Wiley & Sons: New York, 1998; pp 164-165.
- 92 -
11. a) Pearson, R. G., Symmetry Rules for Chemical Reactions. John Wiley and Sons: New York, 1976; pp 12-25; b) Pearson, R. G., “The Second-Order Jahn-Teller Effect.” Journal of Molecular Structure - Theochem 1983, 103, 25-34.
12. Jahn, H. A.; Teller, E., “Stability of Polyatomic Molecules in Degenerate Electronic States. I. Orbital Degeneracy.” Proc. R. Soc. London, A 1937, 161, 220-235.
13. Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M., Advanced Inorganic
Chemistry. Sixth ed.; John Wiley and Sons: New York, 1999; p 865.
14. Bartell, L. S., “Evidence for Pseudo-Jahn-Teller Effect in XeF6.” J. Chem. Phys.
1967, 46, 4530-4531.
15. Pitzer, K. N.; Bernstein, L. N., “Molecular Structure of XeF6.” J. Chem. Phys. 1975, 63, 3849-3856.
16. Pearson, R. G., “Concerning Jahn-Teller Effects.” Proc. Nat. Acad. Sci. USA 1975, 72, 2104-2106.
17. Truce, W. E.; Hoerger, F. D., “The Chemistry of Sultones. II. Alkylation of Organometallic and Related Compounds by Sultones.” J. Am. Chem. Soc. 1955, 77, 2496-2500.
- 93 -
CHAPTER 5
Electrochemical Properties of Cations Encapsulated by an M4L6 Host
Introduction
Supramolecular chemistry relies on labile interactions to assemble intricate
structures from relatively simple components.1 Large, discrete molecular systems may be
designed and synthesized, sometimes exhibiting rich functionality such as guest binding
equilibria.2 While X-ray crystallography and NMR are often the preferred analytical
tools used to characterize supramolecular assemblies, electrochemical methods offer
unique advantages for investigating supramolecular systems with redox-active
components.3 Supramolecular electrochemistry goes beyond the typical redox activity
encountered with most unimolecular covalent species, involving complex behavior such
as electrochemical switching.4, 5 Direct modulation of host-guest binding properties
through electron transfer events can be an elegant means for applications ranging from
chemical sensing to electrochemically driven self-assembly reactions.3, 6, 7 The work
described in this chapter explores the electrochemical behavior of redox-active guests
encapsulated by an M4L6 cluster assembled with redox-silent M.
Ferrocene (Fc) and cobaltocenium (CoCp2+) are popular redox-active guests for
electrochemical binding studies with redox-silent hosts, since they typically show
reversible redox behavior. Cationic sandwich complexes such as CoCp2+ can be
encapsulated by the [Ga4L6]12- host, with binding constants exceeding 104 M-1 in some
cases.8, 9 Reduction of GaIII and naphthalene require extremely negative potentials,
- 94 -
making [Ga4L6]12- electrochemically inert at potentials below the onset of ligand
oxidation. If a guest is electrochemically active within this window, is it possible to
reduce a monocationic guest while encapsulated within the host? To address this
question, cyclic voltammetry experiments were carried out in collaboration with Burak
Ulgut and Prof. Héctor D. Abruña at Cornell University, with CoCp2+,
decamethylcobaltocenium (CoCp*2+), and decamethylferrocenium (FeCp*2
+) chosen as
electroactive species encapsulated by [Ga4L6]12-. To support the electrochemical data, the
binding constant of CoCp2+ with [Ga4L6]
12-, the diffusion coefficient of its host-guest
complex and the rate of encapsulation of CoCp*2+ by [Ga4L6]
12- were measured in DMF-
d7 solutions using 1H NMR methods.
At negative potentials associated with cobalt(III) reduction, the cationic guest
could not be reduced while inside the cavity, suggesting encapsulation by [Ga4L6]12-
rendered the guest electrochemically inert. Adsorption of the host-guest complex onto
the platinum working electrode occurs at more positive potentials associated with
iron(III) reduction, causing the observed response to be much different than the
analogous cobalt systems. Furthermore, exterior ion pairing of the unencapsulated
metallocenium cations with the [Ga4L6]12- anionic host affects both the current and
potential of the observed waves.
Results and Discussion
Synthesis and characterization of host-guest complexes
Solid samples of the 1:1 host-guest complexes were prepared by combining
equimolar quantities of K12[Ga4L6]·(Me2CO)3 and the hexafluorophosphate salt of the
- 95 -
appropriate cationic sandwich complex in methanol under argon, stirring overnight, and
evaporating the solvent in vacuo. The resulting 1H NMR spectra showed the
encapsulated CoCp2+, CoCp*2
+, and FeCp*2+ resonances shifted to significantly higher
fields relative to their unencapsulated values. This diagnostic upfield shift is a
consequence of the host cavity’s magnetically shielded environment, due to ring current
effects in the six surrounding naphthyl moieties.10 In contrast, the neutral FeII
counterparts do not show an appreciable affinity for the host interior. When K12[Ga4L6]
was combined with excess neutral Fc or FeCp*2 in DMF-d7, no peaks were observed for
an encapsulated guest, even after one week.
Figure 5.1. (Top) 1H NMR spectrum (300 MHz) of [FeCp*2]PF6 in acetone-d6 at 20 °C. (Bottom) 1H NMR spectrum (500 MHz) of [FeCp*2⊂Ga4L6]
11- in CD3OD at -20 °C, with an expansion of the central region shown as an inset. The broad resonance for the Cp* methyl protons is marked with a red asterisk in each spectrum.
- 96 -
The 1H NMR spectra for [CoCp2⊂Ga4L6]11- and [CoCp*2⊂Ga4L6]
11- match the
spectra from previous publications, and the signals from the host’s 72 aromatic protons
convey information about the overall symmetry of the host-guest complex.8, 11 Six
signals integrating to 12H each are observed for complexes with T symmetry, including
[CoCp2⊂Ga4L6]11-, but the very large size of CoCp*2
+ causes the host to bulge along one
of its 2-fold axes to accommodate the bulky guest, reducing the overall symmetry to D2
due to loss of the C3 axes. The resulting 1H NMR spectrum of a complex with D2
symmetry will display 18 aromatic resonances integrating to 4H each, observed as
overlapping signals for [CoCp*2⊂Ga4L6]12-. Despite its paramagnetism, the
encapsulation of FeCp*2+ by [Ga4L6]
12- is confirmed by the resulting 1H NMR spectrum
(Figure 5.1). The broad 30H guest resonance, typically observed at very high fields in
the absence of host due to paramagnetism, is shifted nearly 10 ppm further upfield upon
encapsulation. The aromatic host signals, which normally appear between 6 – 9 ppm, are
observed as 18 resonances spread from -3 to +28ppm, each integrating to 4H. (Two of
these resonances overlap to yield a single 8H peak). Interactions with the paramagnetic
FeIII complex are responsible for the wide chemical shift distribution of the aromatic
signals, but the observed splitting from 6 to 18 resonances is caused by a reduction of
symmetry from T to D2. This is exactly the same behavior observed for the isostructural
CoIII analogue, but the paramagnetic FeIII ion makes the splitting due to loss of host
symmetry much more pronounced.
- 97 -
UV-visible spectroscopy has a limited role for characterization of host-guest
complexes. The electronic spectrum of the [Ga4L6]12- host consists of very strong ligand
π-π transitions in the near UV region, with a tail extending through 420 nm responsible
for its yellow color. The much weaker CoIII d-d transitions occur in the same region as
the tail, and therefore cannot be readily observed. For the 17-electron ferrocenium
derivatives, there is a ligand to metal charge transfer (LMCT) band in the visible region
overlapping with two weaker d-d transitions at shorter wavelengths.12 UV-visible
spectroscopy was used to show whether or not the FeIII sandwich complex remained
intact when combined with [Ga4L6]12-. For [FeCp*2]PF6, λmax = 779 nm for the LMCT
band in methanol with no host. Upon encapsulation, this LMCT peak is slightly red
shifted, with λmax = 786 in the presence of host (Figure 5.2). The persistence of that
transition after encapsulation indicates the FeCp*2+ sandwich complex remains intact,
while the small red shift of the LMCT peak is due to a change in solvation environment
(from methanol to the hydrophobic host interior).
Figure 5.2. Visible absorption spectra (methanol, room temperature) showing d-d bands and a ligand to metal charge transfer transition for FeCp*2
+ (blue) and [FeCp*2⊂Ga4L6]11- (red).
500 600 700 800 900
0
100
200
300
400
500
600
786 nm
779 nm
ε ,
M-1cm
-1
Wavelength, nm
Free
Encapsulated
- 98 -
In contrast to the fully methylated derivative, when either Fc+ or Me2Fc+
(Me2Fc+ = 1,1’-dimethylferrocenium) is combined with [Ga4L6]12-, decomposition
occurs rather than encapsulation of the cationic sandwich complex, associated with a
color change and formation of black insoluble solid. This decomposition was monitored
for Me2Fc+ via UV-visible spectroscopy in degassed methanol, with the intensity of the
absorption at λmax = 649 nm diminishing in the presence of [Ga4L6]12-. The stronger
oxidizing abilities of Fc+ and Me2Fc+ relative to FeCp*2+ account for their different
reactivities observed with the host; oxidation of the catechol chelating groups in the host
ligands, accompanied by reduction of FeIII to FeII is the most likely decomposition
pathway.
Guest binding and diffusion measurements
The binding constant of CoCp2+ with [Ga4L6]
12- was measured in DMF-d7 from
1H NMR integration ratios. Four different samples containing 1:1 host-guest
stoichiometries (0.4 mM) were used for redundant data points: three prepared using stock
solutions of [CoCp2]PF6 and K12[Ga4L6]·(Me2CO)4 and one from the solid host-guest
sample used for electrochemical measurements. The binding constant
Kb = 3.3 ± 0.2 x 105 M-1 in DMF-d7 is over ten times higher than that measured in D2O
(1.6 x 104 M-1)13 for the same host-guest complex. Although binding constants have not
been quantified for other guests in DMF, competitive binding between CoCp2+ and Et4N
+
has been observed via NMR. When 1 equivalent of Et4NCl is added to a solution of
[CoCp2⊂Ga4L6]11- in DMF-d7 and allowed to equilibrate at room temperature, over 60%
- 99 -
of the CoCp2+ guest is displaced by Et4N
+.6 This indicates Et4N+ has a slightly higher
binding constant than CoCp2+ in DMF, which is also the case for aqueous solutions (for
Et4N+ binding in D2O, Kb = 1.96 x 104 M-1).14 Thus, changing the solvent from H2O to
DMF leads to very similar increases in the binding affinities for Et4N+ and CoCp2
+,
despite their very different structures, suggesting a general solvent dependence for Kb.
The observed solvent dependence of the binding constant can be partially
explained by the differences in the Born solvation energies. The dielectric constant of
DMF is 38, compared to 78 for H2O, so there is a lower guest desolvation energy cost in
DMF compared to H2O. By invoking a simple thermodynamic cycle to describe the
encapsulation process in the gas phase, solvation effects on the free energy of the
encapsulation reaction ∆Genc can be isolated. Using the Born theory of solvation, the free
energy difference ∆Genc(DMF) – ∆Genc(H2O) ≈ -5.6 kcal mol-1 (Appendix 2). This
estimate qualitatively agrees with experimental observations, predicting a higher binding
constant in DMF compared to H2O, although the energy difference calculated from the
Born model is three times larger than the energy difference calculated from the observed
binding constants.
In cyclic voltammetry measurements with solvated redox couples, the magnitude
of the observed current is proportional to the square root of the diffusion coefficient for
the electroactive species. To properly interpret changes in current, accurate values for
diffusion coefficients are needed. The diffusion coefficient DHG of [CoCp2 ⊂ Ga4L6]11- in
DMF-d7 with 11 mM Bu4NPF6 was measured to be DHG = 3.1 ± 0.1 x 10-6 cm2 s-1 at
6 The exterior Et4N
+ resonances were near the large solvent residual peaks in DMF-d7, preventing accurate binding constant measurements.
- 100 -
25 °C by pulsed gradient spin-echo (PGSE) diffusion NMR (Figure 5.3).7 This value was
measured with a host-guest concentration of 0.40 mM, with potassium counterions.
The binding affinity of the sterically demanding CoCp*2+ guest in DMF is similar
to that for the smaller CoCp2+, as indicated from the small amount of dissociation
observed in the 1H NMR spectrum of 0.4 mM [CoCp*2⊂Ga4L6]11- two weeks after
preparation. Similar diffusion coefficients are also expected for the two host-guest
complexes, since guest encapsulation has a very small effect on the effective radius.
However, the two cations will be encapsulated by the host at very different rates due to
their size differences, with slower kinetics expected for the bulky CoCp*2+. To assess the
upper limits of the time scale where significant guest encapsulation occurs, the rate of
encapsulation of CoCp*2+ by [Ga4L6]
12- in DMF-d7 was measured via 1H NMR
(Figure 5.4). The linear relationship of xf-1 vs. time, where xf is the mole fraction of free
7 For a detailed discussion of diffusion NMR techniques, see Chapter 2 in this dissertation.
Figure 5.3. PGSE 1H NMR diffusion decay curve for [CoCp2⊂Ga4L6]11- in DMF-d7 with 11 mM Bu4NPF6,
using data obtained from host and guest signals. The solid line was fit to the mean of the three decay curves shown.
0 20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
% Gradient Strength
Norm
aliz
ed Inte
gra
l
Host amide Host aromaticsGuest Cp Simulated Fit
- 101 -
(unencapsulated) CoCp*2+, indicates the overall rate law is second order, with the
reaction proceeding via a bimolecular process:
−−+ ⊂→+ 11642
12642 ]LGa*[CoCp]L[Ga*CoCp 2k
The second order reaction rate constant k2 characterizing this reaction is 2.6 M-1 s-1. At
0.4 mM, this corresponds to a reaction half life of approximately 7 min.
Cyclic voltammetry
Cyclic voltammetry of solutions of the free metallocenes in DMF led to reversible
or quasireversible waves for the MIII/II redox couple (Table 5.1). Both cobalt complexes
showed reversible waves for the CoIII/II couples, with E1/2(CoCp*2) occurring at 0.55 V
more negative potentials than the unsubstituted CoCp2+/0 wave. A second quasireversible
wave appeared at very negative potentials with CoCp2+, associated with the CoII/I redox
couple. The larger ∆Ep values associated with the ferrocene derivatives seem to suggest
the FeIII/II redox processes are quasireversible, particularly for FeCp*2. Values reported
0 10 20 30 400.1
0.2
0.3
0.4
0.5
0.6
0.7
Mo
le f
ractio
n F
ree
Co
Cp
* 2
+
Time, minutes
Figure 5.4. Kinetic trace obtained by monitoring the integrated 1H NMR signal from free CoCp*2+
diminishing due to encapsulation by [Ga4L6]12- in DMF-d7 at room temperature.
- 102 -
for the heterogeneous rate constant k0 in the literature reveal ferrocene and its derivatives
demonstrate mediocre electron transfer kinetics with a Pt electrode in DMF solutions,
leading to the observed quasireversible behavior.15
Table 5.1. Half wave potentials (E1/2) and peak separations (∆Ep) measured from the cyclic voltammograms observed for various metallocenes in DMF with 0.1 M Bu4NPF6 (v = 100 mV/s).
Couple E1/2, V vs. Ag/AgCl ∆Ep, mV
CoIII/II –0.76 60 [CoCp2]PF6
CoII/I –1.73 107
[CoCp*2]PF6 CoIII/II –1.31 59
FeCp*2 FeIII/II 0.09 88
Me2Fc FeIII/II 0.47 66
Cyclic voltammograms of [CoCp2⊂Ga4L6]11-, [CoCp*2⊂Ga4L6]
11- and
[FeCp*2⊂Ga4L6]11- in DMF measured with a platinum working electrode are shown in
Figure 5.5. A single wave is observed for the MIII/II couple in all three cases. Comparing
the three systems, it becomes clear that the currents measured for the two cobalt systems
are about an order of magnitude less than that for the iron system at the same scan rate,
despite the fact that the concentrations are approximately the same for all three systems.
Furthermore, the iron system features large anodic currents and small cathodic currents,
but the cathodic and anodic peak currents are approximately equal for the two cobalt
systems. Clearly, there is a different process occurring in the iron system that is not
present in the cobalt systems.
Consider first the cobalt systems. The diffusion coefficient of the host-guest
complex is about a third of that for the guest alone, which would reduce the peak current
by only 32% in a reversible system – not nearly enough to account for the observed order
- 103 -
of magnitude change. Although interference with background current makes
measurement of peak parameters somewhat ambiguous, the peak potential differences
∆Ep measured for the encapsulated CoCp2+/0 couple do not increase with scan rate,
suggesting electrode kinetics do not limit the observed current. (Peak parameters could
not be accurately measured for the encapsulated CoCp*2+/0 couple due to the wave’s
proximity to the residual H2O reduction barrier). The half wave potentials are shifted to
slightly more negative potentials than their values measured in the absence of host. The
observed cathodic shift is somewhat larger for CoCp2+/0 (40-80 mV) than for CoCp*2
+/0
Figure 5.5. Cyclic voltammograms for a) [CoCp2⊂Ga4L6]
11-, b) [CoCp*2⊂Ga4L6]11-, and
c) [FeCp*2⊂Ga4L6]11- measured at 50 mV/s in DMF
using a polished Pt electrode. Data for the host-guest complexes are shown in blue, whereas data measured for the corresponding guests in the absence of host are shown in red. For the cobalt systems shown in a) and b), the currents displayed for the free guests are reduced by 20% from their actual values to aid visual comparison with the much weaker signal for the host-guest complex.
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6[CoCp
2 < Ga
4L
6]11-
: 50 mV/s
x 0.2
Curr
ent,
µA
Potential, V vs. Ag/AgCl
Host-Guest
Free Guest
a)
-1.8 -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
[CoCp*2 < Ga
4L
6]11-
: 50 mV/s
x 0.2
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
Host-Guest
Free Guest
b)
-0.2 0.0 0.2 0.4
-2
-1
0
1
2
3
[FeCp*2 < Ga
4L
6]11-
: 50 mV/s
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Host-Guest
Free Guestc)
- 104 -
(10-20 mV). A cathodic shift indicates the cationic state is stabilized in the presence of
host, as discussed later in this chapter.
The reason the currents are so low for the cobalt host-guest systems is that the
guest is rendered electroinactive while encapsulated within the [Ga4L6]12- host. The
observed signal is from the small amount of unencapsulated CoCp2+ present at
equilibrium. This was confirmed by adding fractional equivalents of CoCp2+ to a
solution of K6(Me4N)5[Me4N⊂Ga4L6] in
DMF (Figure 5.6). The very weak Me4N+
guest is readily displaced by the much
more strongly bound CoCp2+, and served
only as a filler to ensure the cluster
remained intact. As expected, no signal
was observed for [Me4N⊂Ga4L6]11- at
potentials below 0 V, with the solvent
reduction barrier appearing below -1.5 V
due to water introduced with the solid
host-guest sample (a dihydrate). When
0.5 equivalents of CoCp2+ was added and
the solution was allowed to equilibrate for
35 minutes, essentially no signal was
observed in the cyclic voltammogram
other than a small feature in the anodic
segment appearing at -0.76 V, with a
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2
1 eq.
0.5 eq.
0 eq.
Potential, V vs. Ag/AgCl
1 uA
Figure 5.6. Following the addition of 0.5 equivalents of CoCp2
+ to a solution of [Me4N⊂Ga4L6]11- in DMF,
no signal for the CoIII/II redox process is observed in the cyclic voltammogram. When 1 equivalent of CoCp2
+ is present, a small wave appears. Scan rate: 100 mV/s
- 105 -
current of approximately 30 nA above the baseline at 100 mV/s. A second addition of
CoCp2+ increased the total CoCp2
+:[Ga4L6]12- ratio to 1:1, and the solution was allowed to
equilibrate. The resulting cyclic voltammogram was very similar to that for the
[CoCp2⊂Ga4L6]11- host-guest system, featuring a small reversible wave centered at
-0.79 V vs. Ag/AgCl, with peak currents of about 0.15 µA at 100 mV/s. Unfortunately,
addition of CoCp2+ beyond 1 equivalent led to precipitation of the host.
The absence of a signal with 0.5 equivalents of CoCp2+ but the presence of a
signal with 1 equivalent of CoCp2+ per host is due to the much lower amount of free
CoCp2+ available at equilibrium with 0.5 equivalents added compared to 1 equivalent.
The observed binding constant of CoCp2+ encapsulation by [Ga4L6]
12- is 3.3 x 105 M-1 in
DMF, and assuming the competitive binding of Me4N+ is negligible, the concentration of
free CoCp2+ is about 30 µM for a 1:1 host-guest stoichiometry at 0.4 mM. In contrast,
with only 0.5 equivalents of CoCp2+ present under the same conditions, the free CoCp2
+
concentration is only about 3 µM. This large concentration difference explains the
observed CoCp2+/[Me4N⊂Ga4L6]
11- titration results if [CoCp2⊂Ga4L6]11- is redox silent.
Furthermore, the magnitudes of the peak currents observed in the [CoCp2⊂Ga4L6]11- and
[CoCp*2⊂Ga4L6]11- cyclic voltammograms are consistent with the predicted current for a
reversible system with a bulk concentration of about 30 µM.
The reason why the observed current is much larger for [FeCp*2⊂Ga4L6]11- is that
the anionic host has a tendency to adsorb onto the Pt electrode at more positive potentials.
The current response for this system is extremely sensitive to the condition and history of
the electrode: after polishing the electrode, the response will vary from one experiment to
the next when the same experimental parameters are used (Figure 5.7). When the
- 106 -
electrode is polished prior to each run, the anodic current is much larger than that
observed for the initial cathodic sweep, presumably due to some irreversible host
oxidation process. After many experiments without polishing, the current response
stabilized, exhibiting quasireversible behavior with the half wave potential cathodically
shifted approximately 20 mV versus free FeCp*2+/0. When the Pt working electrode was
removed from the cell after the response stabilized, the metal disk was yellow!
Obviously, some bulk layer formed on the electrode over the course of the experiment,
probably containing the yellow [Ga4L6]12- cluster.
To test whether or not this yellow adsorbate was electroactive, the working
electrode was rinsed with clean electrolyte and immersed in fresh degassed DMF with
0.1 M Bu4NPF6, and then the cyclic voltammetric response was observed (Figure 5.8).
Two different redox processes are observed for this surface-confined material: a
quasireversible wave centered at about 0.06 V, and an irreversible oxidation with a peak
at about 0.7 V vs. Ag/AgCl. The quasireversible process persisted upon further cycling,
but the irreversible oxidation did not. The quasireversible wave is probably from
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
[FeCp*2 < Ga
4L
6]11-
: Polished Pt
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
[FeCp*2 < Ga
4L
6]11-
: Nonpolished Pt
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Several experiments
No polishing
Figure 5.7. Cyclic voltammograms for [FeCp*2⊂Ga4L6]11- measured at 50 mV/s with a freshly polished Pt
electrode (left) and with the same electrode after 20 consecutive experiments (right).
- 107 -
FeCp*2+ trapped in the film, and oxidation of the catecholamide groups of L4- to the
quinone form leads to the irreversible anodic wave.
If a solution of K12[Ga4L6] is titrated with neutral FeCp*2, another type of
behavior is observed. Cyclic voltammetry experiments started with an anodic sweep
from -0.3 V to 0.3 V to generate the cation, which could then go on to interact with
[Ga4L6]12- before the cathodic sweep regenerated the neutral species. By adding aliquots
of FeCp*2 stock solution to K12[Ga4L6] in DMF electrolyte, measurements were made at
several guest:host concentration ratios ranging from 0.25 to 4 molar equivalents of
FeCp*2 per host (Figure 5.9). The resulting cyclic voltammograms exhibit peak current
ratios which depend not only on scan rate but on the guest:host ratio as well. The FeCp*2
concentration dependency is most pronounced at slow scan rates, with the ipc/ipa values
measured at 10 mV/s increasing linearly with [FeII] up to 1 equivalent, then increasing
much more gradually as excess FeCp*2 is added. The half wave potentials observed for
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
-2
0
2
4
6
8
Adsorbate Film in Clean Electrolyte
Curr
en
t, µ
A
Potential, V vs. Ag/AgCl
Figure 5.8. After many cyclic voltammetry experiments with [FeCp*2⊂Ga4L6]11- in DMF, the yellow film
deposited on the Pt electrode surface remained adsorbed when transferred to clean electrolyte, and its redox activity was demonstrated by measuring the above cyclic voltammogram for the film itself.
- 108 -
the FeIII/II redox couple are shifted 10-20 mV cathodically in the presence of [Ga4L6]12-,
similar to the shifts observed for the cobalt host-guest systems.
The reason for the lower cathodic currents at slow scan rates is likely due to a
reduction in the effective FeCp*2+ diffusion coefficient due to association with [Ga4L6]
12-.
However, the rate of encapsulation is too slow to account for this association: at 1:1 host-
guest stoichiometry, the second order reaction half life is about 7 minutes for
encapsulation of CoCp*2+, but at v = 10 mV/s FeCp*2 is oxidized and then re-reduced in
about 30 seconds for the potential range used. The host-guest interaction involved in this
situation must then be exterior ion-pair formation. Ion pairing with [Ga4L6]12- has been
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3-0.2
0.0
0.2
0.4
0.6
Potential, V vs. Ag/AgCl
Host OnlyC
urr
ent,
µA
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.25 eq.
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-0.5
0.0
0.5
1.0
1.5
2.0
0.5 eq.
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
0.75 eq.
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
1 eq.
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-2
-1
0
1
2
2 eq.
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Figure 5.9. Cyclic voltammograms of FeCp*2 with K12[Ga4L6] in DMF at 10 mV/s. The host solution was titrated with a FeCp*2 stock solution for 0.25 – 2 equivalents of Fe per host.
- 109 -
shown to play an important role in other systems, proposed as an intermediate step
between guest ejection and complete guest dissociation.8, 16
Effect of ion pairing on redox potentials
The small but real cathodic shifts observed for MIII/II redox waves in the presence
of [Ga4L6]12- indicate the cluster stabilizes the metallocenium cation, making its reduction
slightly more difficult. Metallocenium complexes can interact with the cluster in two
different ways: encapsulation or exterior binding. As discussed in Chapter 2, cations are
known to bind to the exterior of the cluster to generate an ion pair, even in aqueous
media, and ion pair formation has been shown to be an important intermediate step
during the guest exchange process.8, 16, 17
Since encapsulated cobalt guests are redox-silent, but the residual signal from
exterior CoIII exhibits a cathodic shift, the stabilizing interaction must be exterior binding.
The electrochemical and exterior binding equilibria are coupled to one another in the
square scheme shown in Figure 5.10.7 Let E±
f0E±
f0 and E±
b0E±
b0 denote the reduction potentials for
the unencapsulated guest when free (solvated) and bound to the exterior of the cluster,
respectively, and let Kox and Kred denote the equilibrium constants for exterior binding of
Figure 5.10. Coupled electrochemical and binding equilibria for a redox-active guest G+/0 forming an exterior adduct with a redox-silent host H.
- 110 -
the cationic and neutral sandwich complexes, respectively. The reduction potential of the
host-guest adduct (guest bound to exterior, not encapsulated) is determined by the other
three equilibria:
E±
b0 = E±
f0 +
RT
FlnKred
Kox
E±
b0 = E±
f0 +
RT
FlnKred
Kox (5.1)
Separate waves for the solvated and bound species will not be observed unless the
exterior binding constant is very high (≥ 105 M-1); rather, a single wave for the free guest
will appear, shifted cathodically in the presence of host.4 A kinetic study by Leung et al.
exploring exterior ion pairing interactions with [Ga4L6]12- indicates Kox ~ 102 M-1 – far
too small to expect two-wave behavior to be observed.16 Both coulombic attraction and
cation-π interactions operate in favor of exterior binding interactions for the cationic
species, but not the neutral species, suggesting Kred < Kox, and a cathodic shift is predicted
from Equation 5.1.
For CoCp*2+ and FeCp*2, the redox waves shift cathodically 10-20 mV when
[Ga4L6]12- is present, while a larger shift of 40-80 mV is observed with CoCp2
+. The
magnitudes of these shifts are consistent with an exterior binding constant Kox ~ 101 -
102 M-1 for the cationic species. The unsubstituted CoCp2+ has a higher affinity for the
exterior of [Ga4L6]12- compared to its decamethyl analogues. One possible explanation
for this difference is that the π-acidic Cp rings of the sandwich complex favorably
interact with the π-basic naphthalene rings in the ligands of the cluster, but the steric bulk
from the methyl groups prevents the Cp* rings from forming as close a contact with the
naphthalene π system, reducing the strength of the favorable π-π interaction.
Although structurally FeCp*2+ and CoCp*2
+ are nearly identical, they display very
different electrochemical responses when combined with [Ga4L6]12-. The redox potential
- 111 -
of the cobalt complex is much more negative than the iron complex, indicating different
electrode interactions for the host-guest complexes are involved at negative and positive
potentials. Even with these problems, platinum working electrodes led to much simpler
voltammograms than glassy carbon or mercury electrodes. With mercury, multiple
adsorption peaks dominated the observed response for [CoCp2⊂Ga4L6]11-, while data
from glassy carbon suffered from a large background signal and poor reproducibility.
Summary
Encapsulation of an electroactive metallocenium cation within [Ga4L6]12- renders
it electrochemically inert. The host prevents electron transfer from the electrode to the
encapsulated guest, demonstrating a neutral guest cannot be electrochemically generated
from the monocation while it resides in the [Ga4L6]12- host cavity. This is likely due to
coulombic repulsion between the negatively charged cathode and the highly anionic host-
guest complex. The residual unencapsulated guest can bind to the host exterior, and this
ion-paired complex is redox-active. Unsubstituted CoCp2+ binds more strongly to the
exterior of [Ga4L6]12- than CoCp*2
+ and FeCp*2+, as inferred from the relative
magnitudes of the cathodic shifts of the half wave potentials. At more positive potentials,
the cluster strongly adsorbs to the Pt electrode surface, forming an electroactive bulk film
with the guest trapped inside of it. These results demonstrate that [Ga4L6]12- does not
exhibit redox-switchable guest binding – encapsulation simply shuts down a guest’s
redox activity.
- 112 -
Experimental
General Considerations
All reagents were obtained from commercial suppliers and used without further
purification unless noted otherwise. A glove box continuously purged with nitrogen was
used to manipulate and store all air-sensitive solids. Standard Schlenk techniques were
used for reactions carried out under argon. When necessary, solvents were degassed by
at least six pump/fill cycles, using argon for the fill step. DMF was dried and stored over
molecular sieves. Tetraethylammonium chloride (Et4NCl) was recrystallized from
absolute ethanol/ether and dried in vacuo over P2O5 at room temperature for 12 hours,
then dried in vacuo over molecular sieves at 60 °C for 18 hours and stored under
nitrogen. Ferrocene was resublimed in air prior to use, and 1,1’-dimethylferrocene
(Me2Fc) was recrystallized under argon from ethanol/water. H4L (H4L = 1,5-bis(2,3-
dihydroxybenzamido)naphthalene) was synthesized according to literature procedure,10
and purified by washing the solid with methanol. Routine mass spectrometry and
elemental analysis was performed by the Mass Spectrometry Laboratory and
Microanalysis Facility in the College of Chemistry at the University of California,
Berkeley. Elemental analysis of host-guest complexes was performed by Desert
Analytics, Tucson, AZ.
Synthetic procedures
Decamethylferrocenium hexafluorophosphate ([FeCp*2]PF6).18 In air, 3 mL of
concentrated H2SO4 was added to 0.40 g (1.2 mmol) of decamethylferrocene (FeCp*2).
The orange-red color of FeCp*2 immediately changed to dark green upon addition of
- 113 -
acid. After stirring overnight (~12 h) at room temperature, the very dark green viscous
reaction mixture was diluted with water to 50 mL, and the resulting green mixture was
filtered through paper to remove the yellow solid byproduct. A solution of 1.03 g
(6.13 mmol) of NaPF6 dissolved in 5 mL of H2O was gradually added to the filtrate while
stirring, immediately forming a green precipitate. The suspension was stirred for 30 min,
and the green solid was collected on a medium frit, washed with cold H2O (4 x 10 mL),
transferred to a vial, and dried overnight in vacuo over Drierite to yield 0.49 g (87%) of
green powder. Anal. Calc. (found) for C20H30FePF6: %C, 50.97 (51.08); H, 6.42 (6.16).
MS(ESI+, CH3CN): m/z 326.2 ([FeCp*2]+). 1H NMR (acetone-d6, 300 MHz):
δ -37.3 ppm (br s, 30H, -CH3). Visible spectrum (MeOH, λ > 400 nm): λmax(εmax) 779 nm
(543 M-1 cm-1); Gaussian Fit: Peak 1: 646 nm (168 M-1 cm-1); Peak 2: 740 nm
(299 M-1 cm-1); Peak 3: 786 nm (363 M-1 cm-1).
1, 1’-dimethylferrocenium hexafluorophosphate ([Me2Fc]PF6).19 The title compound
was prepared using a modified literature procedure reported for the synthesis of the
unsubstituted FcPF6. In air, 10 mL of concentrated H2SO4 was added to 1.0 g (4.7 mmol)
of recrystallized Me2Fc. The orange color of Me2Fc immediately changed to dark blue
upon addition of acid. After stirring for 4 h at room temperature, the very dark blue
viscous reaction mixture was diluted to 150 mL with water, and the resulting blue
mixture was filtered through paper to remove a dark solid by-product. A solution of 3.6 g
(21 mmol) of NaPF6 dissolved in 10 mL of H2O was gradually added to the filtrate while
stirring, forming a dark precipitate suspended in a dark mother liquor. The reaction
vessel was sealed with a glass stopper and placed in a refrigerator to cool overnight. A
- 114 -
dark blue amorphous solid was collected on a course frit, washed with cold H2O
(3 x 15 mL), and partially dried on frit by pulling vacuum through solid cake under a
stream of nitrogen. Note that the filtrate was dark blue, due to the moderate solubility of
the product in H2O. The gooey solid was transferred to a vial and dried overnight in
vacuo over Drierite (periodically pausing to break up clumps of solid with spatula) to
yield 0.94 g (56%) of dark cerulean amorphous powder. While the product does have a
faint sulfur odor, elemental analysis indicates its purity is sufficient for use without
further purification. Anal. Calc. (found) for C12H14FePF6: %C, 40.14 (40.34); H, 3.93
(3.84). MS(ESI+, CH3CN): m/z 214.0 (Me2Fc+). 1H NMR (D2O, 500 MHz): δ 34.5 (br s,
4H, Cp-Ha), 31.5 (br s, 4H, Cp-Hb), 10.39 ppm (br s, 6H, -CH3). Visible spectrum
(MeOH, λ > 400 nm): λmax(εmax): 649 nm (304 M-1 cm-1), 569 nm (sh, 202 M-1 cm-1),
466 nm (184 M-1 cm-1).
K12[Ga4L6]·4Me2CO.14 The following reaction was carried out under argon using a
procedure adapted from the literature. A suspension of 2.00 g (4.65 mmol) of H4L in
125 mL of degassed methanol was prepared, to which 19.5 mL of 0.5 M KOH in
methanol was added via syringe while stirring, followed by via three pump/fill cycles to
remove any oxygen introduced with the base solution. Addition of base caused most
solid to dissolve, and the color changed from white to yellow. Addition of 25 mL
degassed methanol to the turbid reaction mixture caused the remaining solid to dissolve
after stirring for 10 min at room temperature. To this yellow solution, 1.14 g
(3.11 mmol) of Ga(acac)3 was added, and the reaction mixture was immediately degassed
via six pump/fill cycles. The dark yellow solution was stirred overnight at room
- 115 -
temperature, filtered through a fine frit to remove insoluble residual impurities, and
solvent was removed using a vacuum pump. The dark yellow residue was dried in vacuo
for 3 h to remove 2,4-pentanedione, re-dissolved in a minimal amount of methanol
(100 mL), and the product was precipitated with 600 mL of degassed acetone, then stirred
for 2 h to allow the heterogeneous system to equilibrate. The yellow precipitate was
collected on a frit under a stream of nitrogen, washed with acetone (3 x 50 mL), and dried
in vacuo over molecular sieves and Drierite for 12 h to yield 2.5 g (93%) of light yellow
powder. 1H NMR integrals confirmed the number of acetone molecules present in the
sample.
K11[CoCp2⊂⊂⊂⊂Ga4L6]·KPF6·2H2O. The following reaction was carried out under argon.
To 300 mg (0.0848 mmol) of K12[Ga4L6]·4Me2CO and 28.3 mg (0.0848 mmol) of
[CoCp2]PF6 (Aldrich) was added 50 mL of degassed methanol, leading to a homogeneous
solution. After stirring at room temperature overnight, solvent was removed with a
vacuum pump, and the residue was dried in vacuo for 3 h. The yellow solid was scraped
from flask walls, transferred to an Erlenmeyer flask, swirled in pentane, collected on a
frit under a stream of nitrogen, washed with pentane (3 x 20 mL), and dried overnight in
vacuo over molecular sieves to yield 285 mg (92%) of orange-yellow powder. Anal.
Calc. (found) for C154H98CoF6Ga4K12N12O38P: %C, 50.31 (50.32); H, 2.68 (2.76); N, 4.57
(4.48); Co, 1.60 (1.46); Ga 7.59 (8.15). The number of co-crystallized H2O molecules
per host was confirmed by 1H NMR in DMSO-d6, using the DMSO/H2O peak integrals
from pure DMSO-d6 for comparison. 1H NMR (D2O, 300 MHz): δ 7.98 (br, 12H, ArH),
- 116 -
7.84 (br, 12H, ArH), 7.30 (d, J=8.1 Hz, 12H, ArH), 7.11 (t, J=8.2 Hz, 12H, ArH), 6.75
(d, J=7.0 Hz, 12H, ArH), 6.59 (t, J=7.8 Hz, 12H, ArH), 2.21 (s, 10H, Cp-H, encaps.).
K11[CoCp*2⊂⊂⊂⊂Ga4L6]·KPF6·2H2O. The title compound was synthesized from 301 mg
(0.0850 mmol) of K12[Ga4L6]·4Me2CO and 40.2 mg (0.0848 mmol) of [CoCp*2]PF6
(Aldrich) in a manner analogous to that described for K11[CoCp2⊂Ga4L6]·KPF6·2H2O.
Yield: 290 mg of dark yellow grainy solid. 1H NMR (D2O, 400 MHz): δ 8.26 (d,
J=7.9 Hz, 4H, ArH), 8.22 (d, J=7.8 Hz, 4H, ArH), 7.87 (d, J=7.7 Hz, 4H, ArH), 7.68 (d,
J=8.5 Hz, 4H, ArH), 7.62 (d, J=8.5 Hz, 4H, ArH), 7.47 (m, 8H, ArH), 7.37 (m, 8H, ArH),
6.85 (t, J=8.2 Hz, 4H, ArH), 6.81 – 6.62 (m, 28H, ArH), 6.57 (t, J=7.8 Hz, 4H, ArH),
-0.58 (s, 30H, CH3, encaps.).
K11[FeCp*2⊂⊂⊂⊂Ga4L6]·KPF6·2H2O. The title compound was synthesized from 150 mg
(0.0425 mmol) of K12[Ga4L6]·4Me2CO and 20.1 mg (0.0426 mmol) of [FeCp*2]PF6 in a
manner analogous to that described for K11[CoCp2⊂Ga4L6]·KPF6·2H2O. Yield: 140 mg
(88%) of light green powder. Anal. Calc. (found) for C164H118FeF6Ga4K12N12O38P:
%C, 51.65 (51.80); H, 3.12 (3.45); N, 4.41 (4.28); Fe, 1.46 (1.38). 1H NMR (MeOD,
500 MHz, -20 °C): δ 27.86 (s, 4H, ArH), 21.66 (s, 4H, ArH), 20.32 (s, 4H, ArH), 9.37 (d,
J=8.9 Hz, 4H, ArH), 9.01 (s, 4H, ArH), 8.91 (d, J=8.1 Hz, 4H, ArH), 8.43 (s, 4H, ArH),
6.77 (s, 4H, ArH), 6.42 (d, J=7.9 Hz, 4H, ArH), 4.70 (s, 8H, ArH), 4.06 (m, 4H, ArH),
3.85 (d, J=7.6 Hz, 4H, ArH), 2.79 (br s, 4H, ArH), 2.09 (br s, 4H, ArH), 0.20 (br s, 4H,
ArH), -2.01 (br s, 4H, ArH), -3.08 (br s, 4H, ArH), -46.78 (br s, 30H, -CH3, encaps.).
- 117 -
Binding constant measurements
The binding equilibrium constant for encapsulation of CoCp2+ by [Ga4L6]
12- in
DMF-d7 was calculated from the 1H NMR spectra measured for four different samples
using a Bruker AV-500 spectrometer with a TBI-P probe at ambient temperature. All
four samples were prepared with [host]=[guest]=0.4 mM. One sample was prepared
using pre-synthesized K11[CoCp2⊂Ga4L6]·KPF6, and the other three samples were
prepared by mixing stock solutions of K12[Ga4L6] and [CoCp2]PF6 in a nitrogen-filled
glovebox. The binding constants were computed from the integrated areas of the exterior
and interior guest resonances for each spectrum, and the resulting values were averaged
to obtain the reported binding constant.
Encapsulation kinetic measurements
In a screw-cap vial, 0.83 mg (0.24 µmol) of K12[Ga4L6] was dissolved in 0.6 mL
of DMF-d7 and transferred to an NMR tube. A Bruker AV-500 spectrometer with a
TBI-P probe was used to monitor the 1H NMR signal. After shimming and tuning, the
sample was ejected from the probe, and 10 µL (0.2 µmol) of [CoCp*2]PF6 stock solution
(20 mM in DMF-d7) was added to the NMR tube at t = 0. The tube was repeatedly
inverted to mix the solution, and spectra were acquired at approximately 1 min intervals
using one scan each with a 90° pulse. The first point corresponded to t ≈ 2 min, and the
last to t ≈ 40 min.
- 118 -
Diffusion coefficient measurements
In a nitrogen-filled glovebox, a solution containing 0.4 mM K12[Ga4L6], 0.4 mM
[CoCp2]PF6, and 11 mM Bu4NPF6 in DMF-d7 was prepared, transferred to a standard
NMR tube, and capped with a rubber septum. PGSE diffusion 1H NMR measurements
were performed on a Bruker AVB-400 spectrometer with a z-gradient coil, using the
ledbpgp2s pulse program with diffusion time ∆ = 100 ms, bipolar gradient pulse duration
δ = 7 ms (2 x 3.5 ms), 108 scans per experiment, and a linear gradient strength ramp of
32 increments from 2% to 95%.20 A constant temperature of 298 K was maintained
during the eight hour experiment. The integrated areas were fit to the exponential decay
equation21 using nonlinear regression to evaluate the diffusion coefficient. The probe
gradient power was calibrated from a fit of the diffusion decay curve of dextrose and β-
cyclodextrin in D2O using literature values for the diffusion coefficients.22
Electrochemical measurements
All electrochemical experiments were carried out at ambient temperature (~22 °C)
using either a Solartron 1280B (at Cornell University) or a BAS 100A potentiostat (at UC
Berkeley). Data from the Solartron 1280B were collected without noise filtration in the
post processing. Dry DMF was used for the solvent, and 0.1 M Bu4NPF6, recrystallized
from hot ethyl acetate, served as the supporting electrolyte. Free guest solutions (no
[Ga4L6]12- present) were prepared with analyte concentrations of approximately 1 mM,
and host-guest solutions were prepared with 0.4 mM [Ga4L6]12- due to the host’s limited
solubility in the electrolyte solution. All samples were sparged with dry nitrogen or
- 119 -
argon for at least 10 minutes before the first measurement, and then held under a
continuous stream of inert gas until measurements were finished.
A scintillation vial served as the vessel for the standard three-electrode
electrochemical cell. A platinum disc working electrode and a coiled platinum wire
counter electrode were used, and a homemade Ag/AgCl reference electrode was used
with saturated NaCl filling solution. All potentials are reported relative to this electrode
potential. The working electrode was mechanically polished with 1 µm diamond paste
and electrochemically polished in 0.1 M aqueous H2SO4 at least once before starting
experiments with a different sample. The electrochemical polishing procedure consisted
of four steps (potentials vs. Ag/AgCl electrode described above): 1) hold potential at
1.6 V for 30 sec, 2) hold potential at -0.3 V for 30 sec, 3) rapid cyclic voltammetry
(v = 3 – 5 V/s) from -0.3 V to 1.6 V for 10 cycles, returning to -0.3 V and 4) slow cyclic
voltammogram (v = 0.1 V/s) to test surface characteristics (usual range of -0.2 V to 1.5 V,
depending on H2 reduction barrier). For data collection, potentials were held at the
starting potential for several seconds before beginning the cyclic voltammetry runs.
References
1. a) Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; Vogtle, F.; Lehn, J.-M., Comprehensive Supramolecular Chemistry. Pergamon: Oxford, 1996; b) Lehn, J.-M., Supramolecular Chemistry: Concepts and Perspectives. VCH: Weinheim, 1995; c) Conn, M. M.; Rebek, J., Jr., “Self-assembling capsules.” Chem. Rev. 1997, 97, 1647-1668; d) Lawrence, D. S.; Jiang, T.; Levett, M., “Self-Assembling Supramolecular Complexes.” Chem. Rev. 1995, 95, 2229-2260; e) Leininger, S.; Olenyuk, B.; Stang, P. J., “Self-Assembly of Discrete Cyclic Nanostructures Mediated by Transition Metals.” Chem. Rev. 2000, 100, 853-908.
2. a) Davis, A. V.; Yeh, R. M.; Raymond, K. N., “Supramolecular Assembly Dynamics.” Proc. Nat. Acad. Sci. USA 2002, 99, 4793-4796; b) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Selective Molecular Recognition, C-H Bond Activation, and Catalysis in Nanoscale Reaction Vessels.” Acc. Chem. Res.
- 120 -
2005, 38, 351-360; c) Johnson, D. W.; Raymond, K. N., “The Role of Guest Molecules in the Self-Assembly of Metal-Ligand Clusters.” Supramolecular Chem.
2001, 13, 639-659; d) Pluth, M. D.; Raymond, K. N., “Reversible Guest Exchange Mechanisms in Supramolecular Host-Guest Assemblies.” Chem. Soc. Rev. 2007, 36, 161-171; e) Rebek, J., Jr., “Reversible Encapsulation and Its Consequences in Solution.” Acc. Chem. Res. 1999, 32, 278-286.
3. Kaifer, A. E.; Gómez-Kaifer, M., Supramolecular Electrochemistry. Wiley-VCH: Weinheim, 1999.
4. Kaifer, A. E.; Mendoza, S., “Redox-switchable Receptors.” In Molecular
Recognition: Receptors for Cationic Guests, Gokel, G. W., Ed. Pergamon: Tarrytown, NY, 1996; Vol. 1, pp 701-732.
5. Boulas, P. L.; Gómez-Kaifer, M.; Echegoyen, L., “Electrochemistry of Supramolecular Systems.” Angew. Chem. Int. Ed. 1998, 37, 216-247.
6. a) Venturi, M.; Credi, A.; Balzani, V., “Electrochemistry of coordination compounds: an extended view.” Coord. Chem. Rev. 1999, 185-186, 233-256; b) Gisselbrecht, J.-P.; Gross, M.; Lehn, J.-M.; Sauvage, J.-P.; Ziessel, R.; Piccinni-Leopardi, C.; Arrieta, J. M.; Germain, G.; Van Meerssche, M., “p-Quaterpyridine complexes: crystal structure of the mononuclear copper(II) complex. Electrochemical studies of the monomeric copper(II) and dimeric copper(I) complexes, of their interconversion, and of the bis[RuII(bipy)2] complex.” Nouv. J. Chimie 1984, 8, 661-667; c) Moon, K.; Kaifer, A. E., “Dimeric Molecular Capsules under Redox Control.” J. Am. Chem.
Soc. 2004, 126, 15016-15017; d) Fabbrizzi, L.; Poggi, A., “Sensors and Switches from Supramolecular Chemistry.” Chem. Soc. Rev. 1995, 24, 197-202.
7. Miller, S. R.; Gustowski, D. A.; Chen, Z. H.; Gokel, G. W.; Echegoyen, L.; Kaifer, A. E., “Rationalization of the Unusual Electrochemical Behavior Observed in Lariat Ethers and Other Reducible Macrocyclic Systems.” Anal. Chem. 1988, 60, 2021-2024.
8. Davis, A. V.; Fiedler, D.; Seeber, G.; Zahl, A.; van Eldik, R.; Raymond, K. N., “Guest Exchange Dynamics in an M4L6 Tetrahedral Host.” J. Am. Chem. Soc. 2006, 128, 1324-1333.
9. Fiedler, D.; Pagliero, D.; Brumaghim, J. L.; Bergman, R. G.; Raymond, K. N., “Encapsulation of Cationic Ruthenium Complexes into a Chiral Self-Assembled Cage.” Inorg. Chem. 2004, 43, 846-848.
10. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
11. a) Davis, A. V.; Raymond, K. N., “The Big Squeeze: Guest Exchange in an M4L6 Supramolecular Host.” J. Am. Chem. Soc. 2005, 127, 7912-7919; b) Tiedemann, B. E.
- 121 -
F.; Raymond, K. N., “Dangling Arms: A Tetrahedral Supramolecular Host with Partially Encapsulated Guests.” Angew. Chem. Int. Ed. 2006, 45, 83-86.
12. Sohn, Y. S.; Hendrickson, D. N.; Gray, H. B., “Electronic Structure of Metallocenes.” J. Am. Chem. Soc. 1971, 93, 3603-3612.
13. Parac, T. N.; Raymond, K. N. Unpublished results.
14. Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am.
Chem. Soc. 1998, 120, 8003-8004.
15. Matsumoto, M.; Swaddle, T. W., “The Decamethylferrocene(+/0) Electrode Reaction in Organic Solvents at Variable Pressure and Temperature.” Inorg. Chem. 2004, 43, 2724-2735.
16. Leung, D. H.; Bergman, R. G.; Raymond, K. N., “Scope and Mechanism of the C-H Bond Activation Reactivity within a Supramolecular Host by an Iridium Guest: A Stepwise Ion Pair Guest Dissociation Mechanism.” J. Am. Chem. Soc. 2006, 126, 9781-9797.
17. Fiedler, D.; Bergman, R. G.; Raymond, K. N., “Supramolecular catalysis of a unimolecular transformation: Aza-Cope rearrangement within a self-assembled host.” Angew. Chem. Int. Ed. 2004, 43, 6748-6751.
18. Nesmeyanov, A. N.; Materikova, R. B.; Lyatifov, I. R.; Kurbanov, T. K.; Kochetkova, N. S., “sym-Polymethylferricinium Hexafluorophosphates.” J.
Organomet. Chem. 1978, 145, 241-243.
19. Yang, E. S.; Chan, M.-S.; Wahl, A. C., “Rate of Electron Exchange between Ferrocene and Ferricenium Ion from Nuclear Magnetic Resonance Studies.” J. Phys.
Chem 1975, 79, 2049-2052.
20. Kerssebaum, R., “DOSY and Diffusion by NMR.” In User Guide for XWinNMR
3.1/3.5 Version 1.03, Bruker BioSpin GmbH: Rheinstetten, Germany, 2002.
21. Stejskal, E. O.; Tanner, J. E., “Spin Diffusion Measurements: Spin Echoes in the Presence of a Time-Dependent Field Gradient.” J. Chem. Phys. 1965, 42, 288-292.
22. Longsworth, L. G., “Temperature Dependence of Diffusion in Aqueous Solutions.” J.
Phys. Chem 1954, 58, 770-773.
- 122 -
CHAPTER 6
Redox-Active Vanadium Complexes
Introduction
The work in Chapters 2 through 5 used only M4L6 clusters assembled with GaIII
metal ions, since gallium(III) has several properties which are very convenient for
different applications. For example, the d10 GaIII ion is diamagnetic and thus NMR silent,
and the very high reduction potential of GaIII makes it electrochemically inert. However,
other metal ions can be used to prepare M4L6 tetrahedra, since the catechol binding sites
on L4- are very powerful chelators for hard metal ions with pseudo-octahedral
coordination geometries. Previous studies have shown M4L6 clusters can be assembled
with trivalent MIII = AlIII, FeIII, GaIII, InIII, as well as tetravalent MIV = TiIV, GeIV, and
SnIV.1 In this chapter, redox-active vanadium(IV) is used to prepare the M4L6 tetrahedral
complex, as well as the related M2L3 dinuclear helicate and a mononuclear ML3 model
complex. Where the preceding chapter investigated redox-active guests with redox-silent
hosts, this chapter concerns redox-active hosts with redox-silent guests.
Vanadium tris(catechol) complexes have been characterized with trivalent,
tetravalent, and pentavalent vanadium ions.2 The three oxidation states for [V(cat)3]z-
(z = 1, 2, 3) can be electrochemically interconverted, with E±0(VV=IV) = ¡0:035 VE±0(VV=IV) = ¡0:035 V and
E±0(VIV=III) = ¡0:86 VE±0(VIV=III) = ¡0:86 V vs. SCE for [V(cat)3]z- in acetonitrile (cat2- = ortho-catechol
dianion). This means that for V4L6, eight electrons are released when oxidizing all four
vanadium(III) ions in [V4L6]12- to vanadium(V) in [V4L6]
4-. Due to the large difference
- 123 -
between the two reduction potentials, the waves for [V4L6]4-/8- and [V4L6]
8-/12- are well
separated, but how the reduction of individual vanadium ions will proceed must be
determined. Will reduction of one vanadium in a complex influence the reduction of
adjacent vanadium ions?
When multiple redox sites are involved, three scenarios can occur, depending on
how the individual redox sites interact. Consider the reduction of four identical sites. At
one extreme, if the reduction of one site leads to cooperative interactions that make the
reduction of the remaining three sites much more favorable than the first, all reductions
occur simultaneously and a single four electron wave is observed in the cyclic
voltammogram (Figure 6.1a). At the other extreme, if reduction of one site makes the
next more difficult, each consecutive reduction requires more energy than the last,
requiring more negative electrode potentials, and four separate one electron waves are
observed (Figure 6.1b). Finally, if individual redox sites do not interact at all with one
Figure 6.1. With four identical redox sites, the observed redox wave depends on how the sites interact. a) If reduction of one site makes the next reduction more favorable, all four electrons are transferred at once. b) If consecutive reductions become more difficult, four separate one electron waves are observed. c) If there is no interaction between sites, the four one electron waves overlap to generate one broad wave.
60
40
20
0
-20
-40
-60
-0.25 -0.5 -0.75 -1 -1.25E(V):
I/µA
a)
15
10
5
0
-5
-10
-0.25 -0.5 -0.75 -1 -1.25E(V):
I/µA
b) 37.5
25
12.5
0
-12.5
-25
-0.25 -0.5 -0.75 -1 -1.25E(V):
I/µA
c)
- 124 -
another, the free energy of successive reductions are affected only by statistical (entropic)
factors, and the small differences in potentials result in four overlapping one-electron
waves, observed as a single broad wave in the cyclic voltammogram (Figure 6.1c).3
Synthesis and characterization of vanadium(IV) complexes
In addition to the tetranuclear [V4L6]8- cluster, the mononuclear complex
[V(cam)3]2- and the dinuclear helicate [V2L
H3]
4- were synthesized as the potassium salts,
with the structures of H2cam and H4LH shown in Figure 6.2 (H2cam = N-methyl-2,3-
dihydroxybenzamide; H4LH = 1,4-bis(2,3-dihydroxybenzamido)benzene). A similar
400 500 600 700 800 9000
5
10
15
20
25
ε ,
10
3 M
-1 c
m-1
Wavelength, nm
[V2LH
3]4-
300 400 500 600 700 8000.0
0.2
0.4
0.6
0.8
1.0
Absorb
an
ce
Wavelength, nm
[V(cam)3]2-
400 500 600 700 800 9000
10
20
30
40
50
60
70
ε,
10
3 M
-1 c
m-1
Wavelength, nm
[V4L6]8-
H2cam =
H4LH
a)
b) c)
Figure 6.2. Structural formulae of ligands H2cam and H4LH (upper right), and the UV-visible spectra for
aqueous solutions of a) K2[V(cam)3], b) K4[V2LH
3], and c) K7[Et4N⊂V4L6].
- 125 -
procedure was used to prepare the three vanadium(IV) complexes, where a suspension of
ligand in degassed methanol was combined with the appropriate stoichiometric amount of
VO(acac)2 under argon (H-acac = 2,4-pentanedione) and 2 equivalents of KOH per
VO(acac)2, and stirred at room temperature for several hours. The amount of KOH must
be carefully measured, since the VO2+ ion is favored in basic conditions, and the oxo
ligand must be displaced by the bidentate catechol unit to form the pseudo-octahedral
VL3 coordination geometry in each complex. The resulting products are extremely dark
blue, appearing black in the solid state. This intense color is from broad overlapping
ligand to metal charge transfer bands spanning the entire visible region (Figure 6.2), and
is characteristic of vanadium(IV) tris(catechol) derivatives.
Negative ion electrospray mass spectrometry confirmed the formation of
[V(cam)3]2-, [V2L
H3]
4-, and [V4L6]8-, and elemental analysis confirmed the purity of the
powders. The infrared spectra showed no evidence for a V=O stretch expected near
2500 3000 3500 4000
H (gauss)
[V2LH
3]4-
[V(cam)3]4-
Figure 6.3. EPR spectra of glassy (Et3NH)4[V2LH
3] and K2[V(cam)3] in MeOH/EtOH (9:1) at 8 K.
- 126 -
1000 cm-1 for the vanadyl ion, consistent with pseudo-octahedral vanadium coordination
spheres in all three complexes. EPR spectra of K2[V(cam)3] and (Et3NH)4[V2LH
3] were
measured at 8 K as MeOH/EtOH glasses (Figure 6.3). Both displayed the eight-line
splitting pattern characteristic of the 51V nucleus (I = 7/2), and are similar to the EPR
spectrum of glassy (Et3NH)2[V(cat)3] in aqueous catechol.4 More informative is that the
two spectra are very different from that observed for VO(cat)2, since the lower symmetry
of the square pyramidal vanadyl complex leads to a high degree of anisotropy observed in
the EPR spectra. This further supports the tris(bidentate) coordination mode for
vanadium(IV) expected with formation of [V(cam)3]2- and [V2L
H3]
4-. Due to the
paramagnetic nature of the d1 vanadium(IV) ion, NMR methods could not be used to
characterize the complexes described here.
The X-ray crystal structure of K4[V2LH
3] was determined, confirming that
[V2LH
3]4- is a discrete, dinuclear triple helicate (Figure 6.4). This represents the first
structural characterization of a non-oxo vanadium(IV) supramolecular assembly. The
general structural features of the [V2LH
3]4- helicate are very similar to other dinuclear
Figure 6.4. Illustrations of the anionic fragment [V2LH
3]4- based on the X-ray crystal structure coordinates,
as viewed from the side of the helicate (left), and along the metal-metal axis (right).
- 127 -
helicate structures reported by Raymond and coworkers: two pseudo-octahedral metal
ions, each coordinated by three catecholamide groups joined by a rigid para-phenylene
linker. The compound crystallizes in space group P21/n with Z = 4, with a large amount
of disordered solvent (one Et2O, about seven DMF molecules, and about 0.3 H2O
molecules per helicate). Each helicate is homochiral, with both ∆,∆ and Λ,Λ enantiomers
present in the centrosymmetric unit cell. The V1 – V2 distance is 12.04 Å, with the
metal-metal axis distorted from three-fold symmetry due to crystal packing effects. The
average trigonal twist angle is 33°, but this varies up to three degrees between the six
chelating groups due to packing effects. A summary of bond lengths and angles for the
vanadium coordination spheres is given in Appendix 3.
To gauge the influence of the VIV coordination geometry on the helicate structure,
the structure of K4[V2LH
3] will be compared with two previously published dinuclear
helicates: (s-nic)6[Ga2LH
3] and K4[Ti2LA
3] (H4LA = 2,6-bis(2,3-
dihydroxybenzamido)anthracene).5 The average trigonal twist angle for the GaIII
structure is 46.5°, and only varies up to one degree from this mean value. For the TiIV
structure, the average twist angle is 34.5°, just slightly greater than the average for
vanadium. However, due to the variation between individual dihedral angles, the TiIV
and VIV twist angles are equal within the standard deviation. Thus, the additional d
electron in VIV does not appreciably influence the twist angle. Furthermore, the Ga1 –
Ga2 distance in [Ga2LH
3]6- is 11.71 Å, 0.33 Å shorter than the V1 – V2 distance in
[V2LH
3]4-. The shorter distance in the GaIII structure is due to the larger twist, coiling the
ligands inward. Differences between V-O bond distances observed here and M-O bond
- 128 -
distances (M = TiIV, GaIII) in analogous structures are consistent with differences in ionic
radii.
Electrochemical Studies of Vanadium Complexes
The electrochemical behavior of tris(bidentate) catecholamide vanadium(IV)
complexes is dependent on experimental conditions. In particular, the choice of solvent,
pH of aqueous solutions, and type and conditioning of the working electrode surface are
all factors which strongly influence the observed response during a cyclic voltammetry
experiment. For this reason, this section will be divided into two parts: DMF solutions
will be described first, followed by aqueous solutions. Note that the potential of the
Ag/AgCl (sat’d NaCl) reference electrode used here was -0.080 V vs. SCE.
DMF Solutions – [V2L3]4-
and [V4L6]8-
Cyclic voltammograms of [V2LH
3]4- and [V4L6]
8- in DMF measured using a
hanging mercury drop electrode (HMDE) are shown in Figure 6.5, with potentials
Figure 6.5. Cyclic voltammograms for a) [V2LH
3]4- and b) [V4L6]
8- in DMF. The scan rates range from 0.4 V/s (red) to 2.0 V/s. Inset: peak current for the cathodic (blue) and anodic (red) waves plotted against the square root of the scan rate.
-0.6 -0.8 -1.0 -1.2 -1.4
-5
0
5
10
0.0 0.5 1.0 1.50
2
4
6
8
10
Pe
ak C
urr
en
t, µ
A
sqrt(v), (V/s)1/2
Cathodic
Anodic
Curr
en
t, µ
A
Potential, V vs. Ag/AgCl
-0.6 -0.8 -1.0 -1.2 -1.4
-5
0
5
10
0.0 0.5 1.0 1.50
2
4
6
Pe
ak C
urr
en
t, µ
A
sqrt(v), (V/s)1/2
Cathodic
Anodic
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
a) V2LH
3 b) V4L6
- 129 -
reported vs. Ag/AgCl (saturated NaCl filling solution). Note that all DMF solutions
prepared for electrochemical experiments contained 0.1 M Bu4NPF6 as the supporting
electrolyte. Quasireversible waves for the VIV/III redox couple are observed for both
supramolecular systems, with E1/2 = -0.87 V vs. Ag/AgCl for [V2LH
3]4-/6- and
E1/2 = -0.88 V vs. Ag/AgCl for [V4L6]8-/12-. Plots for the peak cathodic and anodic
currents vs. pv (where vv is the scan rate in V/s) yield straight lines for both systems,
characteristic of a diffusion-controlled process (insets in Figure 6.5).
Assuming reversible behavior, the slopes of the regression lines fit to the peak
current plots in Figure 6.5 can give information about the number of electrons involved.
For the reversible reaction O + ne = R, the peak cathodic current (in µA) for a linear
potential sweep is3
ip;c = (2:69£ 105)n3=2AD1=2O C¤
O
pvip;c = (2:69£ 105)n3=2AD
1=2O C¤
O
pv (6.1)
where n is the number of electrons transferred in a single step, A is the electrode area
in cm2, DO is the diffusion coefficient of species O in cm2 s-1, CO* is the concentration of
O in the bulk solution in mM, and vv is the scan rate in V/s. In theory, the anodic peak
current ip,a measured during the reverse sweep can also be described using the same
equation.
From the slope of ip,c vs. pv , n can be determined if A, DO, and CO* are known.
The value of CO* corresponds to the analyte concentration determined from the sample
preparation. The electrode surface area A = 0.01 cm2 (“SMALL” drop size setting) was
determined from the average mass of the Hg drop dispensed by the apparatus while the
capillary tip is immersed in electrolyte solution. The diffusion coefficients of [V2LH
3]4-,
- 130 -
[V2LH
3]6-, [V4L6]
8-, and [V4L6]12- have not been measured, but they can be estimated from
measured values for similar complexes by using the Stokes-Einstein equation6
D =kBT
6¼¹rh
D =kBT
6¼¹rh (6.2)
where kB is Boltzmann’s constant, T is the temperature in Kelvin, µ is the dynamic
viscosity of the solvent, and rh is the hydrodynamic radius of the diffusing species.
Although rh is unknown, for large molecules we can assume it varies proportionally with
changes in the ion’s size as measured from crystallographic data.8
The diffusion coefficient of [CoCp2 ⊂ Ga4L6]11- with 11 mM Bu4NPF6 in DMF-d7
is DGa = 3.1(1) x 10-6 cm2 s-1 at 25 °C (Chapter 5), measured by diffusion NMR. To use
this observed value for the vanadium system, we can make the following assumptions:
1) the identity of the encapsulated guest does not affect the exterior dimensions of the
cluster, and 2) size changes due to substitution of VIII for GaIII can be neglected.
Therefore, the hydrodynamic radii of [CoCp2⊂Ga4L6]11- and [Et4N⊂V4L6]
11- are equal,
and D = 3.1 x 10-6 cm2 s-1 for [Et4N⊂V4L6]11-. For large ions in nonaqueous solvents,
DO ≈ DR (where R denotes the reduced species), so the diffusion coefficients of the VIII
and VIV clusters are approximately equal.
Although the dinuclear helicate is certainly not a sphere, its effective radius can
be estimated to be about half the center-to-center distance between adjacent complexes in
the crystal lattice. This approximation seems reasonable because a large amount of
solvent co-crystallized with each [V2LH
3]4- ion. From the crystal structure described
earlier in this chapter, the above estimate gives reff ≈ 6.8 Å. This means the effective
radius of the M2L3 helicate is about 2/3 of that for the M4L6 cluster, estimated to be 10 Å
8 For further information about diffusion, see Chapter 2 in this dissertation.
- 131 -
from the center-to-center distances between clusters in the K5(Et4N)6[Et4N⊂Fe4L6]
crystal lattice.7 According to the Stokes-Einstein equation, the diffusion coefficient
varies with 1/r, and thus DM2L3 ≈ (3/2)·DM4L6 = 4.5 x 10-6 cm2 s-1 in DMF electrolyte.
Using the above values for the diffusion coefficients, the slopes of ip,c vs. pv for
the data in Figure 6.5 correspond to n ≈ 1.0 for [V2L3]4-/6- and n ≈ 2.3 for [V4L6]
8-/12-. It
should be emphasized that in general the value of n computed using Equation 6.1 does
not necessarily correspond to the actual number of electrons transferred for multielectron
processes. Only if all electrons are transferred simultaneously in a single, perfectly
reversible step will this value of n actually equal the number of electrons involved. One
familiar example of a true multielectron process is the reduction of aqueous CuII to Cu0.
The peak to peak separation ∆Ep in the cyclic voltammogram would be 59/n mV for a
true multielectron process. For the cyclic voltammograms in Figure 6.5, ∆Ep ≈ 100 mV,
independent of scan rate.
The observed behavior does correspond to that expected if a species contains
multiple identical non-interacting sites, each capable of a one electron reversible redox
reaction. For [V2LH
3]4- there are two sites, and for [V4L6]
8- there are four. In other
words, the reduction of one vanadium(IV) vertex to vanadium(III) has little to no
influence on the reduction potential for the remaining vanadium(IV) ions in the same
molecule. With no communication between metal centers, the free energy of successive
reductions differ purely for entropic reasons. Statistical analysis dictates that the kth
successive formal reduction potential differs from the first according to:3
E±
k ¡ E±
1 = ¡2RT
Fln k
(6.3)
- 132 -
where R is the universal gas constant and F is the Faraday constant. For this sort of
system the wave observed in the cyclic voltammogram will resemble a broadened one-
electron reversible wave centered about the average of the individual reduction potentials.
The value of n calculated from the slope of ip,c vs. pv
pv using Equation 6.1 will be less
than the number of individual one-electron redox steps contributing to the observed
wave.
The oxidation wave for the VV/IV couple can be observed using a glassy carbon
electrode instead of a mercury electrode. A wide potential window allows both waves for
the VV/IV and the VIV/III redox couples to be observed for [V4L6]8- in a single cyclic
voltammogram (Figure 6.6). The VV/IV quasireversible wave appears at 0.47 V vs.
Ag/AgCl, and the two waves are separated by a very large potential difference.
0.5 0.0 -0.5 -1.0 -1.5
-20
-10
0
10
V(4+/3+)
V(5+/4+)
Cu
rren
t, µ
A
Potential, V vs. Ag/AgCl
100 mV/s
Figure 6.6. Cyclic voltammogram of K7[Et4N⊂V4L6] in DMF at a glassy carbon electrode showing the two separate waves for the VV/IV oxidation and the VIV/III reduction at 100 mV/s.
- 133 -
Aqueous Solutions – [V(cam)3]2-
The VIV/III redox couple of this mononuclear model compound was studied with
cyclic voltammetry at a mercury drop electrode. The observed electrochemical behavior,
while highly reproducible, depended strongly on the composition and pH of the
electrolyte solution. The common feature between all systems was a diffusion-controlled
redox wave, corresponding to the reaction
[V(cam)3]2-
(aq) + e- [V(cam)3]3-
(aq).
In addition to the diffusion-controlled wave, various adsorption phenomena were
observed. Treatment of the mercury electrode surface with propanethiol (PrSH) in situ
had a noticeable effect on adsorption. To facilitate discussion, the conditions used are
divided into five categories:
(1) “Naked Electrode” – Low pH: Buffered with phosphate to pH < 7.
(2) “Naked Electrode” – High pH: Buffered with carbonate to pH > 8.
(3) TRIS Solutions – Buffered with TRIS to pH ≥ 8.
(4) Propanethiol Treatment
(5) Excess Ligand –Saturated H2cam in pH 9.5 carbonate buffer solution
A discussion of the behavior observed under these five different conditions will follow
the description of the results, including the influence of complex formation equilibria.
Naked Electrode – Low pH
Cyclic voltammograms in pH 6.7 KCl and pH 6.3 NaClO4 appear very similar.
Representative voltammograms for each solution at low (25 mV/s) and high (500 mV/s)
scan rates are shown in Figure 6.7. The dominant feature at all scan rates is a large
- 134 -
cathodic postpeak appearing as the potential is swept in the negative direction (cathodic
sweep); no corresponding anodic postpeak is observed in the reverse scan. This postpeak
shifts cathodically (to more negative potentials) as the scan rate is increased. At pH 6.7,
an adsorption prepeak is also observed during the initial cathodic sweep, overlapping
with the diffusion wave. The prepeak is not observed after the first cycle. Although an
overlapping prepeak is not observed at pH 6.3, a small peak is present during the first
cycle around -0.2 V that is absent from additional cycles. Additionally, at pH 6.3 there is
an additional feature in the anodic wave near -0.25 V that becomes more prominent at
Figure 6.7. Cyclic voltammograms for aqueous [V(cam)3]2- (pH < 7) measured with a mercury electrode.
0.0 -0.2 -0.4 -0.6 -0.8
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
pH 6.34
(1 M NaClO4)
v = 25 mV/s
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
0.0 -0.2 -0.4 -0.6 -0.8
-1
0
1
2
3
4
5
6
pH 6.34
(1 M NaClO4)
v = 500 mV/s
Curr
ent,
µA
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
Cycle 3
Cycle 4
Cycle 5
-0.2 -0.4 -0.6 -0.8 -1.0
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
pH 6.73
(0.5 M KCl)
v = 25 mV/s
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
Cycle 3
Cycle 4
-0.2 -0.4 -0.6 -0.8 -1.0
-1
0
1
2
3
4
5
6
pH 6.73
(0.5 M KCl)
v = 500 mV/s
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
Cycle 3
Cycle 4
Cycle 5
- 135 -
higher scan rates (vv ≥ 200 mV/s). A similar oxidation process may account for the
“flattened tail” of the anodic wave at pH 6.7.
Repeated cycling of the electrode potential affects both the diffusion wave and the
adsorption postpeak (Figure 6.8). For the diffusion wave, the peak current ratio ip,c/ip,a
increased with additional cycles until it reaches a plateau. In particular, at 500 mV/s in
pH 6.7 KCl, the ratio holds constant at 0.95 by the eighth scan. The variation of the
cathodic to anodic peak current ratio ip,c/ip,a with repeated cycling primarily reflects the
increasing cathodic peak current for the diffusion wave, since the peak anodic current
remained relatively constant. For the cathodic adsorption postpeak, the peak current
decreased with additional cycles. It seems likely that the two processes are
interdependent. This behavior is consistent with reduction of one species that is
electroactive toward reduction both while solvated and while adsorbed on the mercury
surface, with rapid desorption of the reduced species once formed.
Figure 6.8. Effect of repeated cycling on: (a) the diffusion wave peak current ratio and (b) the cathodic postpeak current observed in the cyclic voltammogram of [V(cam)3]
2- at pH 6.7 (scan rate: 500 mV/s).
0 2 4 6 8 10 12 14 16
0.6
0.7
0.8
0.9
1.0
i p,c /
ip
,a
Cycle Number
0 2 4 6 8 10 12 14 160.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Pe
ak C
urr
en
t, µ
A
Cycle Number
a)
b)
- 136 -
Naked Electrode – High pH
Solutions buffered with carbonate to pH 8.7 (KCl) and pH 9.6 (NaClO4) exhibited
sharp adsorption postpeaks in both the cathodic and anodic sweep of their cyclic
voltammograms (Figure 6.9). These peaks are sharper and much less intense than the
enormous postpeak observed for solutions at lower pH. The postpeak intensity remained
essentially unchanged upon repeated cycling, but additional cycles resulted in peaks
slightly shifted (~8 mV) to more negative potentials from the peak observed in the first
cycle. For the pH 9.6 solution, the cathodic postpeak current increased linearly with the
scan rate. The small anodic postpeak is fairly well separated from the main anodic
diffusion wave at pH 9.6, but at pH 8.7 the anodic adsorption peak and diffusion wave
overlap. Two prepeaks were also observed for both solutions during the initial cathodic
sweep: one located around -0.3 V, and another overlapping with the main diffusion-
controlled wave at about -0.5 V. These prepeaks were not observed after the first cycle.
Figure 6.9. Cyclic voltammograms for K2[V(cam)3] measured with a mercury electrode in pH > 8 aqueous solutions buffered with carbonate.
0.0 -0.2 -0.4 -0.6 -0.8 -1.0
pH 9.6 (NaClO4)
Potential, V vs. Ag/AgCl
200 mV/s
500 mV/s
1003 mV/s
2 uA
0.0 -0.2 -0.4 -0.6 -0.8
pH 8.7 (KCl)
Potential, V vs. Ag/AgCl
1 uA
- 137 -
The [V(cam)3]2-/3- diffusion wave is quasireversible in both cases. In the first
cycle, the cathodic wave is shifted anodically and has a larger peak current compared to
subsequent cycles. In the first cycle, ip,c/ip,a = 1.1 – 1.2, whereas for every subsequent
cycle ip,c/ip,a = 0.9 – 1.0. Since additional cycles produce nearly identical waves, the
second cycle will be used for analysis. At pH 9.6, E1/2 = -0.48 V vs. Ag/AgCl (sat’d
NaCl). The peak separation was scan rate dependent, ranging from ∆Ep = 168 mV at
50 mV/s to ∆Ep = 336 mV at 1000 mV/s.
TRIS solutions
The cathodic postpeak decreased at pH 8.1, but a shoulder in the positive sweep
suggests an overlapping anodic postpeak is present (Figure 6.10). At pH 9.2, with
[TRIS] = 0.76 M, the postpeak disappears, but its shape suggests weak adsorption of
[V(cam)3]2- may occur.8 This buffer system should not be used with mercury for these
compounds, since its specific adsorption complicates the system unnecessarily.
Figure 6.10. Cyclic voltammograms for K2[V(cam)3] in aqueous TRIS/TRIS·HCl buffered at a) pH 8.1 and b) pH 9.2, with 0.5 M KCl supporting electrolyte (Hg electrode, scan rate: 200 mV/s).
-0.2 -0.4 -0.6 -0.8 -1.0
-2
-1
0
1
2
3
pH 8.1
[TRIS] = 0.17 M
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
-0.2 -0.4 -0.6 -0.8 -1.0
-1
0
1
2
pH 9.2
[TRIS] = 0.76 M
Cu
rrent,
µA
Potential, V vs. Ag/AgCl
a) b)
- 138 -
Propanethiol treatment
Alkanethiols (CnH2n+1SH) are known to undergo oxidative adsorption onto the
surface of mercury.9 Short-chain thiols (n ≤ 5) form bulk low-density films at open
circuit potential or at potentials more positive than -0.7 V vs. Ag/AgCl.9, 10 In contrast to
their longer chain counterparts that form self-assembled monolayers,11 these low-density
films are permeable to solvated species, and do not hinder electron transfer appreciably.
However, adsorption of the short-chain thiols will decrease the mercury surface tension,
reducing the free energy released upon adsorption of other species (and thus their binding
affinities), including [V(cam)3]2-.
Addition of PrSH directly to the analyte solution allowed modification of the
mercury electrode surface at thiol concentrations of 10 µM. At these PrSH
concentrations, the prepeaks were eliminated, even with a mercury drop less than a few
seconds old (Figure 6.11a). The adsorption postpeak intensity may be reduced by
allowing the mercury drop to “age” in the presence of PrSH at open circuit potential for
several minutes. At high pH, the sharp adsorption postpeaks broaden as the drop is
Figure 6.11. Cyclic voltammograms of aqueous K2[V(cam)3] (pH 9.8, 1 M NaClO4) with 10 µM PrSH measured at a mercury electrode (a) immediately after dispensing a fresh drop, and (b) after aging another mercury drop for 10 min. at open circuit (scan rate: 200 mV/s).
0.0 -0.2 -0.4 -0.6 -0.8 -1.0
-1.0
-0.5
0.0
0.5
1.0
1.5
Cu
rre
nt,
µA
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
0.0 -0.2 -0.4 -0.6 -0.8 -1.0
-1.0
-0.5
0.0
0.5
1.0
1.5
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
Cycle 1
Cycle 2
a) b)
Fresh drop Aged 10 min.
- 139 -
“aged,” with a corresponding decrease in the peak current relative to baseline. After
10 min, the postpeaks were gone, leaving only the reversible VIV/III diffusion wave
(Figure 6.11b). The center of this wave is located at E1/2 = -0.48 V vs. Ag/AgCl at
pH 9.8, which corresponds to the VIV/III reduction potential for [V(cam)3]2-.
The surface coverage of PrSH on the mercury surface was estimated by
chronocoulometry to be about 6 x 10-11 mol cm-2. This measurement was made with
50 µM PrSH in aqueous carbonate buffer (pH 9.4) with 1 M NaClO4 by aging the
mercury drop for 10 min. at open circuit, followed by a double potential step from 0 V to
-1.0 V for 0.5 sec. and back to 0 V vs. Ag/AgCl, taking the difference between the two to
correct for double layer charging effects. The charge per unit area is 11 µC/cm2 from the
average of three measurements, and knowing n = 2 electrons for the reduction of PrSH on
Hg leads to the surface concentration given above. This number is about ten times lower
than the alkanethiol surface coverages reported by Majda and coworkers,12 possibly due
to the lower PrSH concentrations used here or the fact that PrSH accumulation was
performed at open circuit rather than at a fixed potential with a completed circuit.
In contrast, at pH 6.3 the large adsorption postpeak is not completely quenched by
a long soak in PrSH, although its intensity is reduced considerably. Even after aging the
Hg drop for two hours, a significant adsorption postpeak remained for the acidic systems.
This is consistent with the pH dependence of mercuric thiolate formation, with adsorption
strongly preferred in basic solution.13 At low pH, a longer drop age resulted in lower
peak currents in the second cycle compared to those observed with a fresh drop; the
original peak currents could be restored by forming a fresh mercury drop.
- 140 -
Eliminating the adsorption postpeaks at high pH leads to approximately reversible
waves, and the peak current is described by Equation 6.1. This equation predicts ip;c=pvip;c=
pv
should remain constant as long as the bulk analyte concentration CO* does not change.
With PrSH-treated electrodes at pH 9.8, a plot of ip;c=pvip;c=
pv vs. time shows this ratio is
certainly not constant (Figure 6.12). Instead, it decays over time as the solution
equilibrates. (Time is measured using the time stamp written in the data file for the
experiment; t = 0 is defined as the time of the first data file). The decay is significant –
after about 80 minutes, ip;c=pvip;c=
pv is half its original value. Since the electrode area remains
essentially constant over time, the observed decay must be due to a steady decrease in
analyte concentration. In the low pH PrSH system, however, the peak current of the
diffusion wave measured at a newly formed drop remained essentially constant over time.
Note that the solution remained under argon at all times. Thus, some homogeneous
decomposition process occurs in basic solution.
Figure 6.12. Time dependence of ip,c/v1/2 ratio for K2[V(cam)3] in aqueous solution (pH 9.8) with 10 µM
PrSH. For each measurement, the mercury drop electrode was aged for at least five minutes.
10 20 30 40 50 60 70 801.5
2.0
2.5
3.0
3.5
i p,c /
v1
/2,
µA
*(V
/s)-1
/2
Elapsed Time, minutes
- 141 -
Excess ligand
Solid H2cam was added to a 1 mM solution of K2[V(cam)3] in 1 M NaClO4,
buffered with carbonate to pH 9.5 (pH measured before adding H2cam and K2[V(cam)3]).
Some undissolved H2cam remained in the cell, indicating the solution was saturated with
camH2. With judicious choice of a potential range, a reversible [V(cam)3]2-/3- wave was
observed (Figure 6.13). No adsorption peaks were observed with a potential window
[Emax, Emin] = [-0.25, -0.725] V vs. Ag/AgCl. When the potential window was expanded
to [0, -1.0] V, additional broad features appeared. However, these are well separated
from the main diffusion wave, and could be avoided with a narrower potential range.
The wave is centered at E1/2 = -0.48 V vs. Ag/AgCl The observed peak currents
exhibited a linear dependence on pv
pv . After sitting under argon for about 3 hours, a 20%
decrease was observed in the peak cathodic current. This amount of decay was consistent
between three different scan rates.
Figure 6.13. Cyclic voltammograms of aqueous K2[V(cam)3] with excess H2cam measured with a mercury electrode using a (a) wide and (b) narrow potential window (scan rate: 200 mV/s). Prior to adding H2cam, the pH was 9.5 (carbonate buffer with 1 M NaClO4 supporting electrolyte).
0.0 -0.2 -0.4 -0.6 -0.8 -1.0-2
-1
0
1
2
Curr
ent, µ
A
Potential, V vs. Ag/AgCl
-0.2 -0.3 -0.4 -0.5 -0.6 -0.7-2
-1
0
1
2
Cu
rre
nt, µ
A
Potential, V vs. Ag/AgCl
a) b)
- 142 -
Discussion – Mononuclear complex
The behavior observed for aqueous [V(cam)3]2- at pH < 7 is due to
the large difference in affinities for the mercury electrode surface between
the VIV and VIII complexes, and is very similar to that for aqueous [V(cat)3]2-
with mercury electrodes reported in the literature.14 In that study, the authors reported
that the amount of adsorbed [V(cat)3]2- is limited only by the rate of diffusion to the
mercury surface, and not on accumulation potential. The favorable binding interaction
stabilizes the adsorbed VIV species, and thus more negative potentials are required to
reduce the adsorbed species. Following reduction of the adsorbed VIV complex, which
leads to the observed postpeak current, the VIII complex immediately desorbs from the
Hg surface, diffusing into solution.
With both oxidation states electroactive in the solvated state, this difference in
surface binding accounts for the behavior of both the diffusion wave and the adsorption
postpeak observed upon repeated cycling in Figure 6.8. To facilitate discussion, VIV
refers to [V(cam)3]2-, and VIII refers to [V(cam)3]
3-. The VIII formed at the end of the first
cathodic sweep rapidly desorbs, and only solvated VIII is present during the anodic return
sweep, which is oxidized to generate solvated VIV. Before the surface concentration of
VIV is able to return to its original value, the next cycle begins. With more solvated VIV
present at the start of the next cycle, a higher cathodic current is observed for the
diffusion-controlled reduction process, and a lower current is observed for the postpeak.
Each additional cycle begins with more VIV in solution and less VIV adsorbed on the
electrode compared to the previous cycle, eventually reaching a steady state after a large
number of cycles.
- 143 -
In basic solutions, changes in the electrode-solution interface lead to very
different features in the observed cyclic voltammograms. When a non-adsorbing
carbonate buffer is used, adsorption features no longer dominate the observed current
response, but large peak separations in the diffusion wave suggest sluggish electrode
kinetics are involved. Changing the buffer system to TRIS while maintaining a similar
pH dramatically changes the behavior, since TRIS specifically adsorbs onto mercury.
This highlights just how sensitive the electrochemical response can be for complexes
such as [V(cam)3]2- due to changes in solution conditions that affect the composition of
the electrochemical double layer.
The fact that PrSH treatment is able to suppress the stray peaks confirms they
originate from adsorbed electroactive species. Adsorption is suppressed not only by
PrSH treatment, but by excess H2cam as well. Catechol (H2cat) is known to adsorb onto
mercury,15 so its derivative H2cam should exhibit similar adsorption properties. Previous
aqueous electrochemical studies of [V(cat)3]2- derivatives published by Raymond and
coworkers were performed in the presence of excess ligand. In particular, Cooper et al.
merely note that for aqueous [V(cat)3]2-, reduction at a mercury drop “is not reversible in
solutions that lack excess catechol.”16 They attribute this “irreversible cyclic
voltammogram” to dissociation of catechol from the VIII species, which they expected to
exhibit a smaller stability constant than the higher oxidation states. No mention of
adsorption is made. However, the presence of excess ligand presumably prevented
electrochemical observation of the decomposition reaction in that study.
With interference from adsorption peaks successfully suppressed by PrSH, the
decomposition of solvated V(cam)32- may be studied without adding additional ligand. In
- 144 -
aqueous solution, several homogeneous chemical equilibria occur in parallel to
heterogeneous electron transfer and adsorption reactions at the electrode. These
homogeneous reactions are:
[V(cam)3]2¡ + H2O Ð [VO(cam)2]
2¡ + H2cam KVO =[VO(cam)2][H2cam]
[V(cam)2¡3 ]
[V(cam)3]3¡ + 2H+
Ð [V(cam)2]¡ + H2cam K3 =
[V(cam)¡2 ][H2cam]
[V(cam)3¡3 ][H+]2
H2cam Ð Hcam¡ + H+ Ka1 =[Hcam¡][H+]
[H2cam]
Hcam¡Ð cam2¡ + H+ Ka2 =
[cam2¡][H+]
[Hcam¡]
[V(cam)3]2¡ + H2O Ð [VO(cam)2]
2¡ + H2cam KVO =[VO(cam)2][H2cam]
[V(cam)2¡3 ]
[V(cam)3]3¡ + 2H+
Ð [V(cam)2]¡ + H2cam K3 =
[V(cam)¡2 ][H2cam]
[V(cam)3¡3 ][H+]2
H2cam Ð Hcam¡ + H+ Ka1 =[Hcam¡][H+]
[H2cam]
Hcam¡Ð cam2¡ + H+ Ka2 =
[cam2¡][H+]
[Hcam¡]
To assess the significance of the four equilibria under various conditions, one may
approximate the four homogeneous equilibrium constants by using literature data
reported for similar compounds. The vanadium(IV)-Tiron system reported by Buglyó
and Kiss17 was chosen to approximate KVO, while the iron(III)-DMB system reported by
Harris et al.18 was chosen to approximate K3 (DMB = N,N-dimethyl-2,3-
dihydroxybenzamide). For vanadium(IV)-Tiron at 0.2 M ionic strength, log KVO = -2.0,
pKa1 = 7.47, and pKa2 = 12.2. For iron(III)-DMB at 0.1 M ionic strength, log K3 = 12.0,
pKa1 = 8.42, and pKa2 = 12.1. Using these values, the following equilibrium relations are
obtained for the VIV and VIII concentration ratios:
log
µ
[V(cam)2¡3 ]
[VO(cam)2¡2 ]
¶
= 2 + log[H2cam]log
µ
[V(cam)2¡3 ]
[VO(cam)2¡2 ]
¶
= 2 + log[H2cam]
(6.4)
log
µ
[V(cam)3¡3 ]
[V(cam)¡2 ]
¶
= 2pH + log[H2cam]¡ 12log
µ
[V(cam)3¡3 ]
[V(cam)¡2 ]
¶
= 2pH + log[H2cam]¡ 12
(6.5)
Consider first the VIV equilibrium. If the concentration of ligand in solution is
small, most VIV exists as the vanadyl complex at equilibrium. In most experiments
- 145 -
described above, no H2cam was added when preparing the solution for analysis. In these
systems, ligand dissociation from the tris complex accounts for all free H2cam present in
solution. Most free ligand exists as H2cam at pH 7 and below. By setting [H2cam] =
[VO(cam)22-] and CM° = [V(cam)3
2-] + [VO(cam)22-], one can simplify Equation 6.4 and
solve the resulting quadratic equation for the equilibrium concentration of the vanadyl
product if the initial V(cam)32- concentration CM° is known. Using the above
assumptions, and with CM° = 10-3 M, 90% of the vanadium will be found as the vanadyl
complex at equilibrium, with only 10% of the original tris-bidentate complex remaining.
In a more basic solution, such as the pH ~9-10 solutions described above, most of
the free ligand will exist as Hcam-. Under these conditions, the VIV equilibrium can be
re-expressed as:16
[V(cam)3]2¡ +OH¡
Ð [VO(cam)2]2¡ + Hcam¡ K¤
VO = KVOKa1
Iw
[V(cam)3]2¡ +OH¡
Ð [VO(cam)2]2¡ + Hcam¡ K¤
VO = KVOKa1
Iw
where the ionization constant of water Iw = 10-14 at 25 °C. For DMB, pKa1 = 8.4, and
using logKVO = ¡2logKVO = ¡2 gives logK¤
VO = 3:6logK¤
VO = 3:6. Performing a similar calculation with
[V(cam)32-]0 = 10-3 M reveals that at pH 10.4, VO(cam)2
2- is the only vanadium species
present at equilibrium. (This estimate considers only one equilibrium, with Hcam- as the
only free ligand species present in solution). The decomposition reaction is slow
compared to the timescales used in cyclic voltammetry.17
The decomposition of [V(cam)3]2- into [VO(cam)2]
2- and Hcam- in basic solution
accounts for the observed decay of ip;c=pvip;c=
pv over time shown in Figure 6.12. At
equilibrium, nearly all VIV will be found as the vanadyl species according to the above
- 146 -
calculations. Nearly two hours after the solution was first prepared,9 over half of the
vanadium in solution was converted to the vanadyl complex. Despite this, the slope in
Figure 6.12 remained essentially constant, indicating the system was far from
equilibrium. Addition of H2cam results in a higher concentration of [V(cam)3]2- at
equilibrium. This is consistent with the slower decay of ip;c=pvip;c=
pv observed for solutions
saturated with H2cam.
Next consider the VIII equilibrium, using data reported for [Fe(DMB)3]3- to
estimate its formation constants. At pH ≤ 7, the ligand exists primarily as the fully
protonated species H2cam. If no free H2cam was added to the system, [H2cam] =
[V(cam)2-]. If the initial V(cam)3
3- concentration is 1 mM, then at equilibrium 90% of the
complex dissociates into V(cam)2- at pH 7. Therefore, following the reduction of
[V(cam)3]2- to [V(cam)3]
3- at pH ≤ 7, the ligand dissociation process should also be
considered when interpreting the results. Furthermore, if [V(cam)2]- is electroactive
towards oxidation, its formation at pH < 7 could account for the anodic feature observed
at higher scan rates in Figure 6.7.
At pH ≥ 9.5, the ligand primarily exists as the monobasic species Hcam-, and
dissociation primarily proceeds according to:
[V(cam)3]3¡ + H+
Ð [V(cam)2]¡ +Hcam¡ K¤
3 = K3Ka1[V(cam)3]3¡ + H+
Ð [V(cam)2]¡ +Hcam¡ K¤
3 = K3Ka1
with logK¤
3 = logK3 ¡ pKa1 = 3:6logK¤
3 = logK3 ¡ pKa1 = 3:6. Using this equilibrium, the vanadium(III) complex
ratio is given by the following expression:
log
µ
[V(cam)3¡3 ]
[V(cam)¡2 ]
¶
= pH+ log [Hcam¡]¡ 3:6log
µ
[V(cam)3¡3 ]
[V(cam)¡2 ]
¶
= pH+ log [Hcam¡]¡ 3:6
(6.6)
9 While the last data point in Figure 6.12 is at t = 80 min, the solution was actually prepared approximately 30 min prior to t = 0, since the degassing process had to be completed before the first measurement.
- 147 -
If no free H2cam was added to the system, [H2cam] = [V(cam)2-]. If the initial
[V(cam)3]3- concentration is 1 mM, then at equilibrium only 3% of the complex
dissociates into [V(cam)2]- at pH 9.5, and only 2% at pH 10. Therefore, during cyclic
voltammetry experiments with basic solutions the homogeneous dissociation process of
the VIII product can be neglected.
The reversible waves observed with PrSH treatment and with excess ligand in
basic aqueous buffer exhibit the same half-wave potentials, indicating their half-wave
potentials correspond to the standard reduction potential for the VIV/III
couple. This
reduction potential is -0.48 V vs. Ag/AgCl (sat’d NaCl), and calibration of the reference
electrode potential gives E°’ = -0.56 V vs. SCE. This is 0.16 V more positive than the
reduction potential for the aqueous tris(catechol) complex reported in the literature,2
which means reduction to the vanadium(III) complex is easier for the catecholamide
complex than with unsubstituted catechol ligands. The electron-withdrawing amide
group reduces the electron density surrounding the metal, leading to the observed anodic
shift. In contrast, the unsubstituted catechol dianion is highly electron rich, making its
VIII complex a very strong reductant.
Aqueous Solutions – [V2LH3]4-
and [V4L6]8-
.
Complex redox behavior is exhibited by both [V2LH
3]4- and [V4L6]
8- in aqueous
solution. In addition to adsorption processes similar to those observed for [V(cam)3]2-,
additional reduction waves appear at more negative potentials which do not exhibit
typical quasireversible behavior.
- 148 -
The dinuclear complex [V2LH
3]4- exhibits two well separated sets of redox waves
during cyclic voltammetry (Figure 6.14a). The first is a quasireversible diffusion wave
centered at E1/2 = -0.57 V vs. SCE, with an overlapping adsorption-controlled wave
present in the first cycle only. The adsorption-controlled wave may be isolated by
subtracting the current-voltage data for cycle 2 from cycle 1, generating a peak such as
that illustrated in Figure 6.14b. The difference peak current is linearly dependent on the
scan rate, consistent with an adsorption-controlled process present in cycle 1 but absent in
cycle 2 (and additional cycles). Treatment with PrSH reduces the interference from this
adsorption process to a certain extent, although a small peak remains in the difference
wave between the first and second cycles, indicating reactant adsorption still occurs.
Plotting the peak cathodic current of the first redox wave in cycle 2 versus the
square root of the scan rate yields a straight line, characteristic of a diffusion-controlled
process. Assuming D2 = 3 x 10-6, the slope gives a value of n = 1, similar to the DMF
results. A second, smaller redox wave appears at higher scan rates (vv > 0.2 V/s), with a
peak current around -1.23 V vs. SCE. The nature of this wave is not clear – plotting the
Figure 6.14. (a) Cyclic voltammograms for [V2LH
3]4- in aqueous solution (pH 9.8). The first and second
cycles are shown in different colors. (b) Plot of (cycle 1) – (cycle 2) difference, illustrating the absorption peak. The red line is the absorption peak baseline used for analysis.
0.0 -0.5 -1.0 -1.5
1.0 V/s
0.5 V/s
0.2 V/s
Potential, V vs. SCE
Cycle 1
Cycle 2
1 uAa)
-0.2 -0.4 -0.6 -0.8 -1.0
0.0
0.5
Curr
ent
Diffe
rence
, µ
A
Potential, V vs. SCE
b)
- 149 -
peak current against either pv
pv or vv yields a curve which may be considered a straight line
in either case.
For the tetranuclear complex [V4L6]8-, at least three separate reduction waves
appear in the cathodic sweep, with a possible fourth wave appearing in the solvent
reduction tail (Figure 6.15). The first three waves seem to be diffusion controlled, as
suggested by linear plots of peak current versus pv
pv . The first wave is quasireversible,
centered at E1/2 = -0.57 V vs. SCE. The next three waves are irreversible, with the last
exhibiting erratic behavior. Since no additional waves were observed for [V4L6]8- in
DMF, these additional waves are most likely due to various side reactions specific to
aqueous solutions, as opposed to communication between redox sites. The fact that four
peaks can be observed for the cluster with four vanadium redox sites may simply be a
coincidence. Since simple cyclic voltammograms can be observed in DMF, the
complicated aqueous behavior will not be explored further.
-0.5 -1.0 -1.5
-1
0
1
2
Cu
rre
nt, µ
A
Potential, V vs. SCE
Cycle 1
Cycle 2
Figure 6.15. Cyclic voltammogram of [V4L6]8- in H2O (pH 9). Scan rate: 0.2 V/s.
- 150 -
Discussion – Supramolecular complexes
The half-wave potentials of the quasireversible diffusion waves for [V2LH
3]4- and
[V4L6]8- are both equal to E1/2 = -0.57 V vs. SCE = -0.49 V vs. Ag/AgCl. This is very
close to the VIV/III reduction potential of -0.48 V vs. Ag/AgCl for the mononuclear
[V(cam)3]2- complex in basic solution, confirming that the additional vanadium metal
centers have no influence on the individual VIV/III reduction potentials in aqueous
solution. This is also true for DMF solutions, where the reduction potentials for the
dinuclear and tetranuclear complexes differ by only 0.01 V. However, the potentials of
the supramolecular complexes in DMF are 0.38-0.39 V more negative than in aqueous
solutions. This is due to stabilization of the VIII complex in aqueous solution from
hydrogen bonding, which acts to delocalize the excess negative charge into the
surrounding solvent. DMF is aprotic, and no hydrogen bonding interactions exist to
stabilize the anion. Similar solvent effects were observed for reduction of [V(cat)3]2-,
where the VIV/III redox wave was 0.14 V more negative in aprotic CH3CN compared to its
aqueous reduction potential; the smaller cathodic shift in that particular case was due to
stabilization from the Et3NH+ counterions that form strong hydrogen bonds with the
anionic complex in acetonitrile.2
Summary
When redox-active metal ions are used to assemble large supramolecular
complexes such as the M2LH
3 helicate or the M4L6 tetrahedron, there is no interaction
between the individual redox sites. The individual reduction potentials differ only due to
statistical (entropic) factors, with the observed half-wave potentials nearly identical for
- 151 -
the mononuclear, dinuclear, and tetranuclear vanadium catecholamide complexes.
Hydrogen bonding stabilizes the anionic vanadium(III) complexes in aqueous solution,
but the observed electrochemical response in water is complicated by adsorption to the
mercury electrode. Adsorption can be minimized by treating the electrode with
propanethiol, or by adding excess free ligand if sufficiently soluble; basic conditions
should be maintained by buffering with non-adsorbing carbonate. Both VV/IV and VIV/III
redox couples lead to quasireversible waves in DMF for the multiple non-interacting
redox sites, with the two waves separated by nearly 1.5 V. For [V4L6]12-, oxidation of the
VIII ions to VV releases eight electrons in two sets of four one electron reactions. This
suggests potential applications as an electron transport mediator.
Experimental
General considerations
Reagents were obtained from commercial suppliers and used without further
purification unless noted otherwise. Reactions were carried out under argon using
standard Schlenk techniques, and a glove box continuously purged with nitrogen was
used to manipulate and store all air-sensitive solids. Water was purified by passing
house-distilled water through a Millipore Milli-Q system until the effluent stream
conductivity reached 18 MΩ·cm. Dry CH2Cl2 and THF were obtained by flowing
through a bed of activated alumina under nitrogen. When noted, methanol (MeOH),
triethylamine (Et3N), and CH3CN were distilled from CaH2 under N2. H4L and H4LH
were synthesized according to literature procedure.19 Solvents were removed from air-
sensitive systems with a vacuum pump connected to a Schlenk line, and from non-air-
- 152 -
sensitive systems with a rotary evaporator. Electrospray ionization (ESI) mass
spectrometry was performed on K4[V2LH
3] by Robert M. Yeh on a Finnigan LCQ
quadrupole ion trap mass spectrometer equipped with a microspray ionization source, and
on K7[Et4N⊂V4L6] by Michael D. Pluth on a Waters Q-TOF API mass spectrometer
equipped with a Z-spray source. All other mass spectrometry and elemental analyses
were performed by the Mass Spectrometry Laboratory and Microanalysis Facility,
respectively, in the College of Chemistry at the University of California, Berkeley.
Synthetic procedures
N-methyl-2,3-dimethoxybenzamide (Me2cam). To 6.1 g (33 mmol) of 2,3-
dimethoxybenzoic acid was added 40 mL (55 mmol) of thionyl chloride (SOCl2) and
1 mL of DMF. The reaction mixture was allowed to stir for 12 hours under a moderate
flow of N2, and the exhaust stream containing the gaseous HCl and SO2 products passed
through a bubbler filled with aqueous NaOH and vented into a fume hood. The resulting
solid was dried under vacuum for one hour, yielding the acid chloride as a yellow-white
residue. The flask was filled with argon, and 150 mL of dry THF was added to dissolve
the residue. The resulting solution was transferred via cannula to a dry addition funnel
temporarily affixed to an empty flask to minimize exposure to air. To a separate 500 mL
Schlenk flask was added 8.4 mL (97 mmol) of methylamine (40% solution in H2O),
50 mL of THF, and a magnetic stirbar. This flask was affixed to the addition funnel, and
its contents were allowed to cool to -10 °C in an ice/salt bath while stirring vigorously.
The acid chloride solution was added dropwise over the course of two hours, and the
reaction mixture was vigorously stirred for two additional hours at -10 °C. Solvent was
- 153 -
removed, and excess water was removed by azeotropic evaporation with ethanol to form
an oily yellowish residue. The residue was dissolved in 100 mL of CH2Cl2, washed with
1 M HCl (1 x 350 mL), 1 M NaOH (1 x 350 mL), and brine (1 x 200 mL). The organic
layer was collected and its solvent removed to form an oil. Column chromatography on
silica gel with 95%/5% CHCl3/MeOH (500 mL) was performed, with the desired product
collected in the first of two bands. The volume was reduced to 100 mL before drying
with MgSO4 and filtering through a fine frit. The remaining solvent was removed with a
rotary evaporator to form an oil, which slowly solidified overnight at 4 °C.
Recrystallization from CH2Cl2/hexane yielded 3.6 g (55%) of large colorless
orthorhombic crystals, mp 79-80 °C. IR (Nujol): 3329, 1638 cm-1. 1H NMR (400 MHz,
CDCl3): δ 3.02 (d, 3H, J=4.8 Hz), 3.89 (br s, 6H), 7.03 (dd, 1H, J1=1.6 Hz, J2=8.0 Hz),
7.14 (t, 1H, J=8.0 Hz), 7.70 (dd, 1H, J1=1.6 Hz, J2=8.0 Hz), 7.93 (br s, 1H). 13C NMR
(100 MHz, CDCl3): δ 26.59, 56.04, 61.30, 115.19, 122.75, 124.37, 126.76, 147.41,
152.53, 165.87. MS (FAB+): m/z 196.1 (M+), 165.1 ([M–CH3NH·]+). Anal. Calcd.
(found) for C10H13N1O3: %C, 61.53 (61.36); H, 6.71 (6.83); N, 7.18 (7.16). The structure
was confirmed by X-ray crystallography.
N-methyl-2,3-dihydroxybenzamide monohydrate (H2cam·H2O). To a solution of
1.70 g (8.71 mmol) of Me2cam in 150 mL of CH2Cl2 cooled to -78 °C was added 8.7 g
(35 mmol) of BBr3 via syringe. (CAUTION: BBr3 is very corrosive, and violently reacts
with water and alcohols). The reaction mixture was allowed to reach room temperature
and stirred overnight. Solvent was removed under vacuum, leaving a pale yellow
residue. The flask was uncapped, immersed in liquid nitrogen, and 200 mL of methanol
- 154 -
was added in small portions. The open flask was heated at reflux for 8 hours,
maintaining a minimum volume of 75 mL with periodic additions of fresh methanol. The
solvent was then allowed to evaporate, and 150 mL of H2O was added to the grey solid.
The reaction mixture was boiled on a hot plate for three hours, the volume was reduced to
25 mL, and the mixture was cooled in an ice bath to 0 °C. The white precipitate was
collected on a frit, washed with water (4 x 15 mL), and dried overnight under vacuum to
yield 1.36 g (84%) of off-white needles. 1H NMR (d6-acetone, 400 MHz): δ 2.93 (d, 1H,
J=4.4 Hz), 6.72 (t, 1H, J=8.0 Hz), 6.97 (dd, 1H, J1=1.2 Hz, J2=8.0 Hz), 7.23 (dd, 1H,
J1=1.2 Hz, J2=8.0 Hz), 7.7 (br s, 1H), 8.1 (br s, 1H), 13.2 (br s, 1H). 13C NMR (d6-
acetone, 100 MHz): δ 25.46, 114.47, 116.46, 118.08, 118.23, 146.45, 150.00, 170.91.
MS (FAB+): m/z 168.1 (MH+). Anal. Calcd. (found) for C8H9N1O3·H2O: %C, 51.89
(51.95); H, 5.99 (6.03); N, 7.56 (7.51). Note: If further purification is desired, camH2
may be slowly recrystallized from hot water to yield off-white flakes (needles merged
together). The structure was confirmed by X-ray crystallography.20
K2[V(cam)3]·(CH3OH). A slurry of 285 mg (1.54 mmol) of H2cam in 45 mL of MeOH
was prepared, and addition of 1.8 mL (0.90 mmol) of 0.5 M KOH in methanol caused the
solid to dissolve, forming a pale pinkish-brown solution. A solution of 120 mg
(0.45 mmol) of vanadyl bis-acetylacetonate (VO(acac)2) in 10 mL MeOH was added via
cannula, immediately causing the color to change to intense blue-violet. The reaction
mixture was stirred for 90 min at room temperature. Solvent was removed, and the black
residue was dried under vacuum for one hour. The solid was redissolved in 5 mL of
methanol, and addition of 60 mL of EtOAc/Et2O (1:1) formed a powdery precipitate.
- 155 -
The solid was collected on a fine frit, washed with cold EtOAc/Et2O (3 x 20 mL) and
Et2O (2 x 20 mL) and dried overnight under vacuum to yield 225 mg (76%) of fine black
powder. MS(ESI-): 585.0 (KM-), 569.1 (NaM-), 547.1 (HM-), 273.1 (M2-). UV-visible
(H2O): λmax, nm (εmax, M-1s-1) 320 (11700), 425 (sh, 3200), 590 (5020), 680 (sh, 4600).
Anal. Calcd. (found) for C24H21K2N3O9V·(CH3OH): %C, 45.73 (45.62); H, 3.84 (3.91);
N, 6.40 (6.28). IR (Nujol) showed no V=O stretch near 1000 cm-1. EPR (9:1
MeOH/EtOH, 8 K): see Figure 6.3.
K4[V2LH
3]·3(CH3OH)·2H2O. The title compound was synthesized from 350 mg
(0.92 mmol) of H4LH, 160 mg (0.61 mmol) of VO(acac)2, and 2.5 mL (1.2 mmol) of
0.5 M KOH in methanol following a procedure similar to that used to prepare
K2[V(cam)3]·CH3OH. After stirring for three hours at room temperature, the volume was
reduced to 5 mL. Ethyl acetate was steadily added via cannula until solid began to form
(100 mL total), and 25 mL of Et2O was added to form a precipitate suspended in solution.
The flask was immersed in an ice bath for 30 min, and the black solid was collected on a
frit, washed with ethyl acetate (3 x 20 mL) and Et2O (2 x 15 mL), and dried overnight
under vacuum to yield 420 mg (91%) of coarse black solid. Anal. Calcd. (found) for
C60H36K4N6O18V2·3CH3OH·2H2O: %C, 49.80 (49.84); H, 3.45 (3.45); N, 5.53 (5.36).
UV-visible (H2O): λmax, nm (εmax, M-1s-1) 432 (sh, 9420), 583 (11900), 666 (sh, 11100).
IR (Nujol) showed no V=O stretch near 1000 cm-1. The structure was confirmed by
X-ray crystallography (see Figure 6.4, experimental details listed separately below).
- 156 -
(Et3NH)4[V213]·3H2O. To a suspension of 100 mg (0.26 mmol) of H4LH in 40 mL of
distilled methanol was added 0.3 mL (2 mmol) of distilled Et3N. A solution of 46 mg
(0.18 mmol) of VO(acac)2 in 5 mL of distilled methanol was added to the reaction
mixture via cannula, and the color immediately changed to an intense blue. The solution
was stirred for 5 hours at 50 °C, followed by stirring at room temperature for an
additional 12 hours. Removal of solvent afforded a black residue, which was dried for
two hours under vacuum. Recrystallization from CH2Cl2/hexane overnight afforded
extremely thin black plates. The solid was collected on a frit, washed with hexane
(2 x 30 mL) and dried overnight under vacuum to yield 120 mg (81%) of black solid.
Anal. Calcd. (found) for C84H100N10O18V2·3H2O: C, 59.57 (59.74); H, 6.31 (6.09);
N, 8.27 (8.19). UV-visible: λmax, nm (εmax, M-1s-1) 438 (12,900), 578 (17,100), 660 (sh,
16,000). IR (Nujol) showed no V=O stretch near 1000 cm-1. EPR (9:1 MeOH/EtOH,
8 K): see Figure 6.3.
K7[Et4N⊂⊂⊂⊂V4L6]. In a 250 mL Schlenk flask, 1.00 g (2.32 mmol) of H4L and 411 mg
(1.55 mmol) of VO(acac)2 were combined as solids and placed under argon. To this was
added 125 mL of degassed methanol via cannula while stirring, and the resulting mixture
consisted of light colored solid suspended in a very dark blue solution. A stock solution
with 100 mM of Et4NCl in methanol was separately prepared, and 3.8 mL (0.38 mmol) of
this solution was added to the stirred suspension, and the reaction mixture was degassed
via four pump/fill cycles. After stirring at room temperature for 25 min, 6.2 mL
(3.1 mmol) of 0.5 M KOH in methanol was added via syringe, and the reaction mixture
was degassed again immediately after addition of base via several pump/fill cycles. The
- 157 -
nearly black reaction mixture was heated at reflux overnight under argon. After allowing
the mixture to cool to room temperature, the black reaction mixture was filtered using a
Buchner funnel to remove undissolved solids, degassed again, and 150 mL of 1:1
acetone/Et2O was added via cannula to precipitate a black solid. This solid was collected
on a frit, washed with 1:1 acetone/Et2O (3 x 50 mL) and petroleum ether (2 x 50 mL),
and dried overnight in vacuo to yield 1.05 g (83%) of grainy black solid. Anal. Calc.
(found) for C152H104N13O36K7V4·(H2O)6: %C, 55.76 (55.89); H, 3.57 (3.61); N, 5.56
(5.54). UV-visible (H2O): λmax, nm (εmax, M-1s-1) 435 (18150), 593 (24680), 683 (sh,
22420). IR (Nujol) showed no V=O stretch near 1000 cm-1.
X-Ray crystallography
Black crystals of K4[V2LH
3]·6.7DMF·Et2O·0.3H2O were grown by vapor diffusion
of ether into a slightly wet DMF solution of K4[V2LH]3 at 4 °C. Crystal data were
collected using a Bruker SMART diffractometer21 equipped with a CCD area detector
with graphite monochromated Mo Kα radiation (λ = 0.71073 Å). Frames corresponding
to an arbitrary hemisphere of data were collected using ω scans of 0.3° counted for a total
of 30 seconds per frame at T = -151 °C. Peak integrations, cell refinement, and data
reduction were performed by using the Bruker SAINT software package.22 Data were
corrected for Lorentz and polarization effects. The compound crystallizes in monoclinic
space group P21/n, with a = 26.129(5), b = 12.842(2), c = 29.694(5) Å, β = 99.738(2)°,
V = 9820(3) Å3, Z = 4. An empirical absorption correction was applied using
SADABS.23 The structure was solved by direct methods (SIR92) and expanded using
Fourier techniques using the teXsan crystallographic software package.24 Further least
- 158 -
squares refinement to model solvent disorder was done in SHELXL-97.25 All non-
hydrogen atoms, excluding solvent molecules, were refined anisotropically; solvent
atoms were refined with isotropic thermal parameters. The positions of the hydrogen
atoms were included in the structure factor calculation but were not refined. Additional
experimental details for this structure are listed in Appendix 3.
Five of the seven DMF molecules in each asymmetric unit were disordered. Four
of these five were modeled by allowing each DMF to partially occupy two positions. For
three of the four pairs of DMF molecules, the thermal parameters were refined with the
occupancy fixed. The atoms in the two parts of the fourth DMF pair were constrained to
lie in a rigid group corresponding to an idealized DMF, and the occupancy was refined
with fixed isotropic thermal parameters. The fifth disordered DMF showed substitutional
disorder, and was modeled as a DMF with 70% occupancy and a water with 30%
occupancy. Full matrix least squares with 925 parameters yielded R1 = 0.088 for 5711
reflections with I > 2σ(I).
Electrochemical measurements
All electrochemical experiments were carried out at ambient temperature (~22 °C)
using a Bioanalytical Sciences BAS-100A Electrochemical Analyzer. No corrections
were made for uncompensated resistance or junction potential, but uncompensated
resistance was estimated using the “IR Test” function in the control software. Analyte
concentrations of 0.3-1.1 mM were used. Unless noted otherwise, a Princeton Applied
Research EG&G Model 303A Static Mercury Drop Electrode (SMDE) apparatus was
used as the working electrode. A platinum wire served as the counter electrode. A
- 159 -
silver/silver chloride reference electrode filled with saturated NaCl/AgCl was used. The
potential of the reference electrode was -0.080 V vs. SCE. This was measured in aqueous
electrolyte using a voltmeter and a commercial electrode of known potential, and
confirmed using [CoCp2]+/0 as a reference compound in DMF with 0.1 M
tetrabutylammonium hexafluorophosphate (Bu4NPF6). For measurements with glassy
carbon, the working electrode was a BAS glassy carbon disk electrode (3 mm diameter)
polished with 0.1 µm alumina before use, and a coiled platinum wire counter electrode
was used. A similar Ag/AgCl reference electrode filled with saturated NaCl was used
with glassy carbon, and exhibited the same potential as the electrode used with mercury.
Unless noted, a fresh drop of mercury was dispensed immediately before each
experiment. Samples were bubbled with inert gas while stirring for at least twenty
minutes before any measurements, and for at least 30 s between each individual
experiment. A steady stream of argon flowed over the solution surface when not
bubbling. DMF solutions contained 0.1 M Bu4NPF6 as the supporting electrolyte. For
aqueous solutions, electroanalytical grade KCl (0.5 M) or reagent grade NaClO4 (1 M)
was used as the supporting electrolyte. Solution pH was buffered using potassium or
sodium phosphate (pH 6.34, 6.73), Tris/HCl (pH 8.13, 9.16), or NaHCO3/Na2CO3
(pH 8.73, 9.53, 9.64, 9.79) at 0.1 M buffer ionic strength. The argon used for aqueous
measurements was first passed through a gas scrubbing tower filled with aqueous
VCl2/HCl and then a bubbler filled with 1 M aqueous NaCl.
Experiments with propanethiol-modified mercury electrodes were performed with
the following procedure. Propanethiol was added to a prepared NaClO4/carbonate buffer
solution, and this solution was either used to prepare the analyte solution, or was added to
- 160 -
an existing analyte solution via syringe, followed by bubbling with argon for 4 min while
stirring. In either case, [PrSH] ≈ 10 µM in the actual analyte solution. Mercury drops
were “aged” as follows: after the normal 30 s bubble/stir cycle between experiments, the
solution was allowed to settle for 1 min, and at least five drops were dispensed and
dislodged to remove any contaminants from the capillary tip. A new drop was dispensed,
and the system was left undisturbed without potential control. Time was measured using
a stopwatch. After the desired time elapsed, the experiment was started.
Digital cyclic voltammetry simulations were performed using BAS DigiSim.26
References
1. Seeber, G.; Tiedemann, B. E. F.; Raymond, K. N., “Supramolecular Chirality in Coordination Chemistry.” In Top. Curr. Chem., Reinhoudt, D. N.; Crego-Calama, M., Eds. Springer: Berlin, 2006; Vol. 265, pp 147-183.
2. Cooper, S. R.; Koh, Y. B.; Raymond, K. N., “Synthetic, Structural, and Physical Studies of Bis(triethylammonium) Tris(catecholato)vanadate(IV), Potassium Bis(catecholato)oxovanadate(IV), and Potassium Tris(catecholato)vanadate(III).” J.
Am. Chem. Soc. 1982, 104, 5092-5102.
3. Bard, A. J.; Faulkner, L. R., Electrochemical Methods: Fundamentals and
Applications. 2nd ed.; John Wiley & Sons: Hoboken, 2001.
4. Branca, M.; Micera, G.; Dessi, A.; Sanna, D.; Raymond, K. N., “Formation and Structure of the Tris(catecholato)vanadate(IV) Complex in Aqueous Solution.” Inorg.
Chem. 1990, 29, 1586-1589.
5. a) Yeh, R. M.; Ziegler, M.; Johnson, D. W.; Terpin, A. J.; Raymond, K. N., “Imposition of Chirality in a Dinuclear Triple-Stranded Helicate by Ion Pair Formation.” Inorg. Chem. 2001, 40, 3922-3935; b) Scherer, M.; Caulder, D. L.; Johnson, D. W.; Raymond, K. N., Angew. Chem. Int. Ed. 1999, 38, 1588-1592.
6. Welty, J. R.; Wicks, C. E.; Wilson, R. E.; Rorrer, G., Fundamentals of Momentum,
Heat, and Mass Transfer. 4th ed.; John Wiley & Sons: New York, 2001.
7. Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N., “The Self-Assembly of a Predesigned Tetrahedral M4L6 Supramolecular Cluster.” Angew. Chem. Int. Ed.
1998, 37, 1840-1842.
- 161 -
8. Wopschall, R. H.; Shain, I., “Effects of Adsorption of Electroactive Species in Stationary Electrode Polarography.” Anal. Chem. 1967, 39, 1514-26.
9. Stevenson, K. J.; Mitchell, M.; White, H. S., J. Phys. Chem. B 1998, 102, 1235-1240.
10. Muskal, N.; Mandler, D., “Thiol self-assembled monolayers on mercury surfaces: the adsorption and electrochemistry of omega-mercaptoalkanoic acids.” Electrochim.
Acta 1999, 45, 537-548.
11. Slowinski, K.; Chamberlain, R. V.; Miller, C. J.; Majda, M., J. Am. Chem. Soc. 1997, 119, 11910-11919.
12. Slowinski, K.; Chamberlain, R. V.; Miller, C. J.; Majda, M., “Through-Bond and Chain-to-Chain Coupling. Two Pathways in Electron Tunneling through Liquid Alkanethiol Monolayers on Mercury Electrodes.” J. Am. Chem. Soc. 1997, 119, 11910-11919.
13. Birke, R. L.; Mazorra, M., “A Study of the Electrochemical Characteristics of some Thiols by Differential Pulse Polarography and Other Electrochemical Techniques.” Anal. Chim. Acta 1980, 118, 257-269.
14. Ivanov, V. D.; Kaplun, M. M., “Studying the adsorption accumulation of vanadium-catechol complexes at a mercury electrode.” Zhurnal Anal. Khim. 1997, 52, 362-368.
15. Sarangapani, S.; Venkatesan, V. K., “Adsorption of Phenols at Mercury-Solution Interface.” Proc. Indian Natn. Sci. Acad. A 1983, 49, 124-142.
16. Cooper, S. R.; Koh, Y. B.; Raymond, K. N., J. Am. Chem. Soc. 1982, 104, 5092-5102.
17. Buglyo, P.; Kiss, T., “Formation of a Tris Complex in the Vanadium(IV)-Tiron System.” J. Coord. Chem 1991, 22, 259-268.
18. Harris, W. R.; Carrano, C. J.; Cooper, S. R.; Sofen, S. R.; Avdeef, A. E.; McArdle, J. V.; Raymond, K. N., J. Am. Chem. Soc. 1979, 101, 6097-6104.
19. a) Kersting, B.; Meyer, M.; Powers, R. E.; Raymond, K. N., J. Am. Chem. Soc. 1996, 118, 7221-7222; b) Caulder, D. L.; Powers, R. E.; Parak, T. N.; Raymond, K. N., Angew. Chem. Int. Ed. 1998, 37, 1840-1842.
20. Escalada, J.; Freedman, D.; Werner, E. J., “2,3-Dihydroxy-N-methylbenzamide monohydrate.” Acta Crystallogr. Sect. E 2004, E60, o1296-o1298.
21. SMART Area Detector Software Package, v5.052d; Bruker Analytical X-ray Systems, Inc.: Madison, WI, 1999.
22. SAINT: SAX Area-Detector Integration Program, v7.01A; Bruker Analytical X-ray Systems, Inc.: Madison, WI, 2002.
- 162 -
23. Blessing, R. H., “An Empirical Correction for Absorption Anisotropy.” Acta
Crystallogr. Sect. A 1995, 51, 33-38.
24. teXsan: Crystal Structure Analysis Package, Molecular Structure Corp.: The Woodlands, TX, 1992.
25. Sheldrick, G. M. SHELXL97, Universität Göttingen: 1997.
26. Rudolph, M.; Feldberg, S. W. DigiSim, v3.03b; Bioanalytical Systems, Inc.: West Lafayette, IN, 2004.
- 163 -
APPENDIX 1
Supplementary Spectra for Chapter 4
2D NMR spectra for [RuCn⊂⊂⊂⊂Ga4L6]12-
Figure A1.1. Portion of the 2D TOCSY 1H NMR spectrum (D2O, 400 MHz) of
[RuC4⊂Ga4L6]12- showing the cross peaks between encapsulated RuC4 resonances.
- 164 -
Figure A1.2. Portion of the 2D COSY 1H NMR spectrum (500 MHz, D2O) of
[RuC6⊂Ga4L6]12- showing the cross peaks between encapsulated RuC6 resonances.
- 165 -
Figure A1.3. Portion of the 2D COSY 1H NMR spectrum (500 MHz) of
[RuC10⊂Ga4L6]12- showing the cross peaks between encapsulated RuC10 resonances.
- 166 -
Figure A1.4. Portion of the 2D COSY 1H NMR spectrum (D2O, 500 MHz) for
[RuC10⊂Ga4L6]12-, showing the coupling between host protons.
- 167 -
Figure A1.5. Portion of the 2D NOESY 1H NMR spectrum (D2O, 400 MHz) of
[RuC4⊂Ga4L6]12-, showing the cross peaks between host and guest signals.
- 168 -
Figure A1.6. Portion of the 2D NOESY 1H NMR spectrum (D2O, 400 MHz) of
[RuC10⊂Ga4L6]12-, showing the cross peaks between host and guest signals.
- 169 -
ESI MS for [RuCn⊂⊂⊂⊂Ga4L6]12-
Figure A1.7. Portion of the ESI- mass spectrum of [RuC4⊂Ga4L6]12- in 75% H2O with
25% methanol. Peaks from fragments containing the host-guest complex in -3,-4, and -5
charge states are shown here.
- 170 -
Figure A1.8. Portion of the ESI- mass spectrum of [RuC6⊂Ga4L6]12- in methanol. Peaks
from fragments containing the host-guest complex in -3 and -4 charge states are shown
here.
- 171 -
Figure A1.9. Portion of the ESI- mass spectrum of [RuC8⊂Ga4L6]12- in methanol. Peaks
from fragments containing the host-guest complex in -3 and -4 charge states are shown
here.
- 172 -
Figure A1.10. Portion of the ESI- mass spectrum of [RuC10⊂Ga4L6]12- in methanol.
Peaks from fragments containing the host-guest complex in -3, -4, and -5 charge states
are shown here.
- 173 -
Full 1D NMR spectra for [RuAn⊂⊂⊂⊂Ga4L6]11- (n = 4, 6, 8)
Figure A1.11. 1H NMR spectrum of [RuA4⊂Ga4L6]11- in D2O at room temperature.
The guest’s four carbon side chain is entirely contained within the host cavity, resulting
in a T-symmetric host-guest complex.
- 174 -
Figure A1.12. 1H NMR spectrum of [RuA6⊂Ga4L6]11- in D2O at room temperature. The
guest’s six carbon side chain rapidly extends and retracts in this system.
- 175 -
Figure A1.13. 1H NMR spectrum of [RuA8⊂Ga4L6]11- in D2O at room temperature. The
guest’s eight carbon side chain is extruded through one of the host’s facial apertures,
resulting in a C3-symmetric host-guest complex.
- 176 -
2D NOESY NMR Spectra for [RuAn⊂⊂⊂⊂Ga4L6]11- (n = 4, 6, 8)
Figure A1.14. 2D NOESY 1H NMR spectrum of [RuA6⊂Ga4L6]11- at 27 °C in D2O, with
a mixing time τ = 0.4 sec.
- 177 -
Figure A1.15. Expanded region of the 2D NOESY spectrum of [RuA6⊂Ga4L6]11- at
27 °C in D2O shown in Figure S4 showing the cross peaks between host and guest
signals.
- 178 -
Figure A1.16. 2D NOESY 1H NMR spectrum of [RuA4⊂Ga4L6]11- (n = 4) at 27 °C in
D2O, with a mixing time τ = 0.4 sec.
- 179 -
2D COSY NMR Spectra: [RuAn⊂⊂⊂⊂Ga4L6]11- (n = 4, 6, 8)
Figure A1.17. 2D COSY 1H NMR spectrum of [RuA6⊂Ga4L6]11- at 27 °C in D2O
showing the aromatic host resonances for T symmetry (time-averaged). Assignments are
also included.
- 180 -
Figure A1.18. 2D COSY 1H NMR spectrum of [RuA6⊂Ga4L6]11- at -60 °C in MeOD
showing the aromatic host resonances for the C3 state. Although the peaks are broad and
overlap with each other, especially below 7 ppm, the cross peaks are consistent with eight
sets of three adjacent protons, expected for point group C3. This COSY spectrum is
qualitatively similar to that observed upon encapsulation of RuCn.
- 181 -
Electrospray Ionization Mass Spectra (ESI-MS) of [RuAn⊂⊂⊂⊂Ga4L6]11-
Pluth /R aym ond, B ryan4 in M eO H
815 820 825 830 835 840 845 850 855 860 865 870m /z0
100
%
Q T 0155 50 (0.855) Cm (1 :153) TO F M S E S - 7 .19e3833.542
833.287
833.043
832.788
824.046
823.803
823.549
823.306
823.053
822 .810
815.056814.068
822.546
815.560
832 .543
824 .796
825.050832.288
825.304
832 .044825.547
831 .789825.812
834.042
843.029
834 .286
842.783
834.542
842.526
834 .786
842.281
835.041842.024
835.286841.779
841 .533835.542
835.786
843 .274
843 .777
844.034
844 .280
844.537
852 .762844.783
852.526845.039
852.279
845.275
853.021
855.032
856.012
856.346
a)
Pluth /R aym ond, B ryan4 in M eO H
815 820 825 830 835 840 845 850 855 860 865 870m /z0
100
%
Q T 0155 50 (0.855) Cm (1 :153) TO F M S E S - 7 .19e3833.542
833.287
833.043
832.788
824.046
823.803
823.549
823.306
823.053
822 .810
815.056814.068
822.546
815.560
832 .543
824 .796
825.050832.288
825.304
832 .044825.547
831 .789825.812
834.042
843.029
834 .286
842.783
834.542
842.526
834 .786
842.281
835.041842.024
835.286841.779
841 .533835.542
835.786
843 .274
843 .777
844.034
844 .280
844.537
852 .762844.783
852.526845.039
852.279
845.275
853.021
855.032
856.012
856.346
a)
Pluth/Raymond, Bryan4 in MeOH
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848m/z0
100
%
0
100
%
0
100
%
QT0155 (0.017) Is (1.00,1.00) C159H102N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12833.767
833.267
833.017
832.767
832.517
832.267
832.017
834.267
834.517
834.768
835.018
835.268
QT0155 (0.017) Is (1.00,1.00) C159H101N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12843.256
842.756
842.506
842.256
842.006
841.756
841.506
843.756
844.006
844.256
844.507
844.757845.007
QT0155 50 (0.855) Cm (1:153) TOF MS ES- 7.19e3833.542
833.287
833.043
832.788
832.543
832.288
832.044829.540830.303
834.042
843.029834.286
842.783834.542
842.526834.786
842.281
835.041842.024
835.286 841.779835.542
843.531
843.777
844.034
844.280
844.537
845.039
845.789
K4Na2H1[1 ⊂ Ga4L6]4-
(simulated)
K5Na2[1 ⊂ Ga4L6]4-
(simulated)
b)
Observed Data
Pluth/Raymond, Bryan4 in MeOH
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848m/z0
100
%
0
100
%
0
100
%
QT0155 (0.017) Is (1.00,1.00) C159H102N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12833.767
833.267
833.017
832.767
832.517
832.267
832.017
834.267
834.517
834.768
835.018
835.268
QT0155 (0.017) Is (1.00,1.00) C159H101N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12843.256
842.756
842.506
842.256
842.006
841.756
841.506
843.756
844.006
844.256
844.507
844.757845.007
QT0155 50 (0.855) Cm (1:153) TOF MS ES- 7.19e3833.542
833.287
833.043
832.788
832.543
832.288
832.044829.540830.303
834.042
843.029834.286
842.783834.542
842.526834.786
842.281
835.041842.024
835.286 841.779835.542
843.531
843.777
844.034
844.280
844.537
845.039
845.789
K4Na2H1[1 ⊂ Ga4L6]4-
(simulated)
K5Na2[1 ⊂ Ga4L6]4-
(simulated)
b)
Observed Data
Figure A1.19. (a) Electrospray ionization mass spectrum (ESI-MS) for [RuA4⊂Ga4L6]11-
(n = 4) in MeOH showing peaks for the z = -4 charge state. (b) Simulated isotopic
distributions for two particular fragment ion formulae, with the observed data shown
below. Here, 1+ = RuA4+.
- 182 -
Pluth/Raym ond, Bryan6 in M eOH
828 830 832 834 836 838 840 842 844 846 848 850 852 854 856 858 860 862 864m /z0
100
%
Q T0156 41 (0.701) Cm (1:133) TO F M S ES- 1.33e4850.292
849.798
849.540
840.551
840.305
840.049
839.804
839.558
839.292
839.057831.067
830.558
830.060
838.801831.810 836.553
840.796
841.052 849.293
841.298
849.047841.554
841.800848.789
842.056848.543
842.302848.285
850.538
850.796
851.043
860.035
851.290
859.279
859.031851.548
858.783851.795
858.534852.042
852.289 858.286
860.283
860.531
860.780
861.039
861.287
861.547
861.795
862.303
862.801
a)
Pluth/Raym ond, Bryan6 in M eOH
828 830 832 834 836 838 840 842 844 846 848 850 852 854 856 858 860 862 864m /z0
100
%
Q T0156 41 (0.701) Cm (1:133) TO F M S ES- 1.33e4850.292
849.798
849.540
840.551
840.305
840.049
839.804
839.558
839.292
839.057831.067
830.558
830.060
838.801831.810 836.553
840.796
841.052 849.293
841.298
849.047841.554
841.800848.789
842.056848.543
842.302848.285
850.538
850.796
851.043
860.035
851.290
859.279
859.031851.548
858.783851.795
858.534852.042
852.289 858.286
860.283
860.531
860.780
861.039
861.287
861.547
861.795
862.303
862.801
a)
Pluth/Raymond, Bryan6 in MeOH
838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853m/z0
100
%
0
100
%
0
100
%
QT0156 (0.017) Is (1.00,1.00) C161H106N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12840.775840.525
840.275
840.025
839.775
839.525
839.275
839.025
841.025
841.275
841.525
841.775
842.026
842.276
842.526
QT0156 (0.017) Is (1.00,1.00) C161H105N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12850.264850.014
849.764
849.514
849.264
849.014
848.764
848.514
850.514
850.764
851.014
851.264
851.514
851.765
852.015
QT0156 41 (0.701) Cm (1:133) TOF MS ES- 1.33e4850.292850.045
849.798
849.540
840.551840.305
840.049
839.804
839.558
839.292
838.801
840.796 849.293
841.298849.047
841.554
841.800 848.789
842.056 848.543
850.538
850.796
851.043
851.290
851.548
851.795
852.042
852.547853.300
K5Na2[2 ⊂ Ga4L6]4-
(simulated)
K4Na2H1[2 ⊂ Ga4L6]4-
(simulated)
b)
Observed Data
Pluth/Raymond, Bryan6 in MeOH
838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853m/z0
100
%
0
100
%
0
100
%
QT0156 (0.017) Is (1.00,1.00) C161H106N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12840.775840.525
840.275
840.025
839.775
839.525
839.275
839.025
841.025
841.275
841.525
841.775
842.026
842.276
842.526
QT0156 (0.017) Is (1.00,1.00) C161H105N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12850.264850.014
849.764
849.514
849.264
849.014
848.764
848.514
850.514
850.764
851.014
851.264
851.514
851.765
852.015
QT0156 41 (0.701) Cm (1:133) TOF MS ES- 1.33e4850.292850.045
849.798
849.540
840.551840.305
840.049
839.804
839.558
839.292
838.801
840.796 849.293
841.298849.047
841.554
841.800 848.789
842.056 848.543
850.538
850.796
851.043
851.290
851.548
851.795
852.042
852.547853.300
K5Na2[2 ⊂ Ga4L6]4-
(simulated)
K4Na2H1[2 ⊂ Ga4L6]4-
(simulated)
b)
Observed Data
Figure A1.20. (a) ESI-MS for [RuA6⊂Ga4L6]11- (n = 6) in MeOH showing peaks for the
z = -4 charge state. (b) Simulated isotopic distributions for two particular fragment ion
formulae, with the observed data shown below. Here, 2+ = RuA6+.
- 183 -
Pluth /R aym ond, Bryan8 in M eO H
836 838 840 842 844 846 848 850 852 854 856 858 860 862 864 866 868 870 872m /z0
100
%
Q T0157 14 (0.239) Cm (1:153) TO F MS ES- 1.75e4857.305
856.810
856.562
847.567
847.320
847.063
846.817
846.560
846.313
846.056838.076
837.330
836.830
838.822 845.810
856.303848.060
848.317
856.055
848.564
848.811 855.808
849.057855.560
849.315855.312
857.553
857.801
858.060
867.045866.546
858.308
866.297
866.048858.556
865.799858.804
865.550859.063
865.301859.311
867.294
867.544
867.804
868.043
868.303
868.552
868.802
869.312
869.562
869.801
a)
Pluth /R aym ond, Bryan8 in M eO H
836 838 840 842 844 846 848 850 852 854 856 858 860 862 864 866 868 870 872m /z0
100
%
Q T0157 14 (0.239) Cm (1:153) TO F MS ES- 1.75e4857.305
856.810
856.562
847.567
847.320
847.063
846.817
846.560
846.313
846.056838.076
837.330
836.830
838.822 845.810
856.303848.060
848.317
856.055
848.564
848.811 855.808
849.057855.560
849.315855.312
857.553
857.801
858.060
867.045866.546
858.308
866.297
866.048858.556
865.799858.804
865.550859.063
865.301859.311
867.294
867.544
867.804
868.043
868.303
868.552
868.802
869.312
869.562
869.801
a)
Pluth/Raymond, Bryan8 in MeOH
845 846 847 848 849 850 851 852 853 854 855 856 857 858 859m/z0
100
%
0
100
%
0
100
%
QT0157 (0.017) Is (1.00,1.00) C163H110N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12847.783847.533
847.283
847.033
846.783
846.533
846.283
846.033
848.033
848.283
848.533
848.783
849.033
849.284
849.534
QT0157 (0.017) Is (1.00,1.00) C163H109N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12857.272857.022
856.772
856.522
856.272
856.022
855.772
855.522
857.522
857.772
858.022
858.272
858.522
858.772
859.023
QT0157 14 (0.239) Cm (1:153) TOF MS ES- 1.75e4857.305857.057
856.810
856.562
847.567847.320
846.817
846.560
846.313846.056
845.575
847.814 856.303
848.317856.055
848.564
848.811855.808
855.560849.315
857.553857.801
858.060
858.308
858.556
858.804859.063
859.560
b)
K5Na2[3 ⊂ Ga4L6]4-
(simulated)
K4Na2H1[3 ⊂ Ga4L6]4-
(simulated)
Observed Data
Pluth/Raymond, Bryan8 in MeOH
845 846 847 848 849 850 851 852 853 854 855 856 857 858 859m/z0
100
%
0
100
%
0
100
%
QT0157 (0.017) Is (1.00,1.00) C163H110N12O36Ga4Ru1K4Na2 TOF MS ES- 1.19e12847.783847.533
847.283
847.033
846.783
846.533
846.283
846.033
848.033
848.283
848.533
848.783
849.033
849.284
849.534
QT0157 (0.017) Is (1.00,1.00) C163H109N12O36Ga4Ru1K5Na2 TOF MS ES- 1.18e12857.272857.022
856.772
856.522
856.272
856.022
855.772
855.522
857.522
857.772
858.022
858.272
858.522
858.772
859.023
QT0157 14 (0.239) Cm (1:153) TOF MS ES- 1.75e4857.305857.057
856.810
856.562
847.567847.320
846.817
846.560
846.313846.056
845.575
847.814 856.303
848.317856.055
848.564
848.811855.808
855.560849.315
857.553857.801
858.060
858.308
858.556
858.804859.063
859.560
b)
K5Na2[3 ⊂ Ga4L6]4-
(simulated)
K4Na2H1[3 ⊂ Ga4L6]4-
(simulated)
Observed Data
Figure A1.21. (a) ESI-MS for [RuA8⊂Ga4L6]11- (n = 8) in MeOH showing peaks for the
z = -4 charge state. (b) Simulated isotopic distributions for two particular fragment ion
formulae, with the observed data shown below. Here, 3+ = RuA8+.
- 184 -
APPENDIX 2
Born Solvation Energies and Guest Encapsulation
According to the Born theory of solvation, a spherical ion of charge zi with a
radius ri, the Gibbs free energy of solvation relative to the gas phase is given by:1
¢Gsolv = ¡µ
Nae2
8¼²0
¶
z2i
ri
µ
1¡ 1
²
¶
= ¯
µ
z2i
ri
¶ µ
1¡ 1
²
¶
where ¯ ´ Nae2
8¼²0
¢Gsolv = ¡µ
Nae2
8¼²0
¶
z2i
ri
µ
1¡ 1
²
¶
= ¯
µ
z2i
ri
¶ µ
1¡ 1
²
¶
where ¯ ´ Nae2
8¼²0
(A2.1)
where ²² is the dielectric constant of the solvent, Na is Avogadro’s number, e is the charge
of an electron, and ²0²0 is the permittivity of vacuum. Suppose an anionic host H12-
encapsulates a monocationic guest G+ to form the host guest complex [G⊂H]11-,
abbreviated as HG11-. The solution-state binding equilibrium reaction is
H12¡(s) +G+
(s) Ð HG11¡(s) ¢G(s)
enc = ¡RT lnKbH12¡(s) +G+
(s) Ð HG11¡(s) ¢G(s)
enc = ¡RT lnKb
where subscript (s) denotes solvated species, ¢G(s)enc¢G(s)enc is the Gibbs free energy of the
binding reaction in the solution state, Kb is the binding equilibrium constant, R is the
universal gas constant, and T is the absolute temperature.
Solvation effects can be analyzed by considering a hypothetical pathway where
guest encapsulation occurs in the gas phase, denoted by subscript (g):
G+(s) ¡! G+
(g) ¡¢Gsolv(G+)
H12¡(s) ¡! H12¡
(g) ¡¢Gsolv(H12¡)
H12¡(g) +G
+(g) ¡! HG11¡
(g) ¢G(g)enc
HG11¡(g) ¡! HG11¡
(s) ¢Gsolv(HG11¡)
G+(s) ¡! G+
(g) ¡¢Gsolv(G+)
H12¡(s) ¡! H12¡
(g) ¡¢Gsolv(H12¡)
H12¡(g) +G
+(g) ¡! HG11¡
(g) ¢G(g)enc
HG11¡(g) ¡! HG11¡
(s) ¢Gsolv(HG11¡)
- 185 -
¢G(s)enc = ¢G(g)
enc +¢Gsolv(HG11¡)¡¢Gsolv(H
12¡)¡¢Gsolv(G+)¢G(s)
enc = ¢G(g)enc +¢Gsolv(HG
11¡)¡¢Gsolv(H12¡)¡¢Gsolv(G
+) (A2.2)
Assume that G+, H12-, and HG11- are spheres with effective radii rG for G+ and rH
for H12- and HG11-. Since guest encapsulation does not change the overall size of the
cluster very much, one can assume that the effective radii of H12- and HG11- are equal. Of
course, neither [Ga4L6]12- nor guests such as CoCp2
+ are actually spheres, but this
assumption is sufficient for a rough estimate of solvation effects. Using the Born
equation (Equation A2.1), the free energies of solvation are:
¢Gsolv(G+) = ¡ ¯
rG
µ
1¡ 1
²
¶
¢Gsolv(G+) = ¡ ¯
rG
µ
1¡ 1
²
¶
(A2.3)
¢Gsolv(HG11¡)¡¢Gsolv(H
12¡) = ¯
µ
1¡ 1
²
¶
(112 ¡ 122)
rH
= 23¯
rH
µ
1¡ 1
²
¶
¢Gsolv(HG11¡)¡¢Gsolv(H
12¡) = ¯
µ
1¡ 1
²
¶
(112 ¡ 122)
rH
= 23¯
rH
µ
1¡ 1
²
¶
(A2.4)
Upon substitution, the free energy of encapsulation in the solution state is
¢G(s)enc = ¢G(g)
enc + ¯
µ
1¡ 1
²
¶ µ
23
rH
+1
rG
¶
¢G(s)enc = ¢G(g)
enc + ¯
µ
1¡ 1
²
¶ µ
23
rH
+1
rG
¶
(A2.5)
Assuming ¢G(g)enc¢G(g)enc is independent of solvent properties because it corresponds to a gas
phase reaction, the above expression can be used to estimate the difference in guest
binding affinities between two different solvents.
Suppose there are two different solvents, solvent A and solvent B, with dielectric
constants ²A²A and ²B²B, respectively. If guest binding occurs in both solvent systems, and if
²A 6= ²B²A 6= ²B, the difference in the free energy of guest binding in the two solvents is
- 186 -
¢G(B)enc ¡¢G(A)
enc = ¯
µ
23
rH+
1
rG
¶ µ
1¡ 1
²B
¶
¡ ¯
µ
23
rH+
1
rG
¶ µ
1¡ 1
²A
¶
= ¯
µ
23
rH+
1
rG
¶ µ
1
²A¡ 1
²B
¶
¢G(B)enc ¡¢G(A)
enc = ¯
µ
23
rH+
1
rG
¶ µ
1¡ 1
²B
¶
¡ ¯
µ
23
rH+
1
rG
¶ µ
1¡ 1
²A
¶
= ¯
µ
23
rH+
1
rG
¶ µ
1
²A¡ 1
²B
¶
(A2.6)
Equation A2.6 allows the relative binding affinities for two different solvent
systems to be estimated from their dielectric constants and the effective radii of host and
guest. If ²B < ²A²B < ²A, then ¢GBenc ¡¢GA
enc < 0¢GBenc ¡¢GA
enc < 0, and therefore the binding constants
KBb > KA
bKBb > KA
b . In other words, the binding constant is higher in less polar solvents because
the desolvation energy cost is smaller, since the reactants are less strongly solvated. Note
that most of this energy cost is due to reducing the host charge from -12 to -11, whereas
much less energy is required to fully desolvate the monocationic guest; this observation
was discussed in a previous publication.2
We would like to compare the observed binding constant differences for CoCp2+
encapsulation by [Ga4L6]12- in water and DMF solutions with the predicted difference
from Equation A2.6. For ferrocenium (Fc+), rG ≈ 4 Å,3 and for [Et4N⊂Fe4L6]11-, the
center-to-center distance between adjacent clusters in the crystal lattice gives rH ≈ 9.5 Å.4
At 25 °C, the dielectric constant for water ²A = 78²A = 78 and for DMF ²B = 38²B = 38. The energy
difference ¢GBenc ¡¢GA
enc¢GBenc ¡¢GA
enc from Equation A2.6 is -5.6 kcal mol-1. This predicts a higher
binding constant for DMF than H2O based purely on solvation differences, in qualitative
agreement with the observed differences. However, the experimentally observed free
energy difference is -1.8 kcal mol-1, much smaller than the value from Equation A2.6.
References
1. Born, M., “Volumen und Hydratationswärme der Ionen.” Z. Physik 1920, 1, 45-48.
- 187 -
2. Parac, T. N.; Caulder, D. L.; Raymond, K. N., “Selective Encapsulation of Aqueous Cationic Guests into a Supramolecular Tetrahedral M4L6 Anionic Host.” J. Am.
Chem. Soc. 1998, 120, 8003-8004.
3. Matsumoto, M.; Swaddle, T. W., “The Decamethylferrocene(+/0) Electrode Reaction in Organic Solvents at Variable Pressure and Temperature.” Inorg. Chem. 2004, 43, 2724-2735.
4. Caulder, D. L.; Brückner, C.; Powers, R. E.; König, S.; Parac, T. N.; Leary, J. A.; Raymond, K. N., “Design, Formation, and Properties of Tetrahedral M4L4 and M4L6 Supramolecular Clusters.” J. Am. Chem. Soc. 2001, 123, 8923-8938.
- 188 -
APPENDIX 3
Crystallographic Data for K4[V2LH
3]·6.7DMF·Et2O·0.3 H2O
formula C87.1H92.2K4N13.7O27V2 formula weight 2021.22 temperature (K) 122 crystal system monoclinic space group P21/n (no. 14) a (Å) 26.129(5) b (Å) 12.842(2) c (Å) 29.694(5) β (degrees) 99.738(2) Z 4 V (Å3) 9820.15 µcalc (mm-1) 0.44 Tmin, Tmax (transmission) 0.8979, 0.9783 F(000) 4194.8 crystal size (mm) 0.25 x 0.15 x 0.05 radiation Mo Kα (λ = 0.71073 Å) h,k,l range collected -26 ≤ h ≤ 20, -12 ≤ k ≤ 10, -29 ≤ l ≤ 29 2θ range 3.54° – 41.75° scan type ω scans scan speed (°/min) 0.6 reflections collected 31142 unique reflections 10302 data: Fo
2 > 2σ(Fo2) 5711
number of parameters 925 data/parameter ratio 6.2 for Fo
2 > 2σ(Fo2)
number of restraints 44 R 0.088 with Fo
2 > 2σ(Fo2)
wR2 0.2130 with Fo2 > 2σ(Fo
2) goodness of fit 1.25
- 189 -
Vanadium Coordination Sphere Information
Twist Angle O – V – O Angle O1, O2 35.49° 80.38° O5, O6 36.48 79.66 O9, O10 32.16 80.37 Average for V1 34.71° 80.14° Std Deviation 2.26° 0.41°
Twist Angle O – V – O Angle O4 , O3 33.12° 79.89° O8, O7 30.56 79.94 O12, O11 35.95 79.65 Average for V2 31.41° 79.83° Std Deviation 1.48° 0.15°
Distance (Å) Distance (Å) V1-O1 1.9126 V1-O2 1.9596 V1-O5 1.9311 V1-O6 1.9618 V1-O9 1.9248 V1-O10 1.9370 V2-O4 1.9377 V2-O3 1.9572 V2-O8 1.9142 V2-O7 1.9562 V2-O12 1.9367 V2-O11 1.9374 Avg dout 1.926(11) Avg din 1.952(11)
V1 – V2 Distance = 12.041 Å
V1
O5
O9
O2
O6
O1
O10
V2
O4
O8
O11
O3
O12
O7