Angel Kaifer Marielle Gomez-Kaifer
Supramolecular Electrochemistry
Angel Kaifer Marielle G6mez-Kaifer
Supramolecula...
114 downloads
1022 Views
13MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Angel Kaifer Marielle Gomez-Kaifer
Supramolecular Electrochemistry
Angel Kaifer Marielle G6mez-Kaifer
Supramolecular Electrochemistry
@WILEY-VCH Weinheim . New York . Chichester . Brisbane . Singapore . Toronto
Professor Angel E. Kaifer Dr. Marielle Gomez-Kaifer Chemistry Department University of Miami Coral Gables, FL 33124-0413 USA
This hook was carefully produced. Nevertheless, authors and publisher do not warrant the information contained therein to be free of errors. Readers are advised t o keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Library of Congress Card No. applied for A catalogue record for this hook is available from the British Library Die Deutsche Bibliothek - CIP-Einheitsaufnahme Kaifer, Angel E.: Supramolecular clcctrochemistry / Angel E. Kaifer ; Marielle Gomez-Kaifer. Weinheim ; New York ; Chichester ; Brishane : Singapore : Toronto : Wiley-VCH, 1999 ISBN 3-527-29591-6
0 WILEY-VCH Vcrlag GmbH. D-69469 Weinheim (Federal Republic of Germany), 1999
Printed on acid-free and chlorine-free paper All rights reserved (including those of translation into other languages). N o part of this hook may he reproduced in any form - by photoprinting, microfilm, or any other means nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this hook, even when not specifically marked as such are not to he considered unprotected by law. Printing: hetz-druck GmbH. D-64291 Darmstadt Bookbinding: Wilh. Osswald + C o . , D-67433 Neustadt Printed in the Federal Republic of Germany ~
To our parents
Angel, Barbara, Edward, Ellen and Emilia in thanks for their support of our curiosity and pursuit of knowledge
Preface
Supramolecular chemistry has different meanings for different people, and, perhaps, because of this ambiguity it is best to follow Professor Lehn's definition. In his words, supramolecular chemistry concerns the "chemistry beyond the covalent bond". This definition rightfully places emphasis on the importance of intermolecular forces present in supramolecular systems. From simple host-guest complexes to infinitely more complicated supramolecular assemblies described in the recent literature, intermolecular forces are at the core of all relevant supramolecular systems. Research in supramolecular chemistry already has a long and productive history and many reviews and several books have been devoted to this field of chemistry. In spite of the influence and importance that electrochemical techniques and concepts have had in the development of the field, when we started this work there were no monographs available on supramolecular electrochemistry. This book represents a modest attempt to correct this state of affairs. In launching a project such as this one, it is important to set clear goals. Our primary and foremost purpose was to provide the research community in supramolecular chemistry with an accessible and readable summary on the use of electrochemical techniques and the applications of electrochemical concepts to this new research area. A second purpose was to increase the level of interest in supramolecular systems from the electrochemical community. The book is thus intended as a tool to build bridges between these two rather separate communities and to foster some degree of cross-fertilization between the two research areas. In order to meet these goals, and due to the wide diversity of topics that we wanted to address, we could not, therefore, provide a comprehensive or thorough description of the subject matter. As is usually the case, we were forced to make many compromises concerning the selection of topics and the depth of coverage. The first seven chapters of the book are intended as an introduction to electrochemical techniques. Readers with a reasonable background in electrochemistry can probably skip these chapters. The remaining chapters address the electrochemistry of the most important types of supramolecular systems. Overall, the book should be useful to graduate students and postdocs, as well as more experienced researchers who are interested in expanding their research horizons at the frontier of electrochemistry and supramolecular chemistry. As stated above, this book does not even attempt a comprehensive coverage of the research topics presented. Therefore, literature citations were selected by the authors using very personal and, perhaps, seemingly arbitrary
VIII
Preface
criteria. We wish to apologize in advance to all those who feel that their work has not been appropriately represented here: this book is merely our personal view of the research landscape. Miami, June1999
Angel Kaifer and Marielle Gdmez-Kaifer
Acknowledgments
The authors owe their involvement in this research field to many people. First, they wish to express their gratitude to their common doctoral advisor, Professor Luis Echegoyen, who inspired them with his great enthusiasm and love for science. Both authors have been associated with the group of Professor Allen J. Bard, to whom they are indebted for his lucid teachings on electrochemistry, and his insights into the general importance of electrochemistry and the diverse ways in which it can be applied to almost any field of chemical research. We are fortunate to have been influenced by Professor Bards approach to research, which embodies all the best of collegiality and the true spirit of scientific endeavor. Over the years the authors have worked, discussed, and in many cases published, with a number of researchers in this field. Their contributions are important to this book and are reflected at many points throughout the manuscript. At the risk of missing someone, we wish to thank Jerry Atwood, Carmen Maria Casado, Alessandro Casnati, Cecil Criss, Isabel Cuadrado, Jeff Evanseck, George W. Gokel, David Gutsche, Moisks Morln, David Reinhoudt, Neil Spencer, J. Fraser Stoddart, Rocco Ungaro, and Frank van Veggel. The Kaifer group’s contribution to research in the field stems from the hard work of graduate students and postdoctoral associates. Their work cannot be overestimated. Julio Alvarez, Anna Bernardo, Richard Bissell, Claudia Cardona, Rene Castro, Emilio Cbrdova, Luis Godinez, Tim Goodnow, Mei Han, Rahimah Isnin, Jing Li, Jian Liu, Sandra Mendoza, Armen Mirzoian, Carlos Peinador, Maria Rojas, Esteban Romln, Yun Wang and Litao Zhang deserve our special thanks. The authors gratefully acknowledge the continued financial support from the U.S. National Science Foundation and NATO. Finally, the authors wish to express their sincere gratitude to their editor, Jorn Ritterbusch, for his encouragement, help, and above all else, his mfinite patience!
Contents
1Fundamentals of Electrochemical Theory
1
1.1Cell potentials and Electrochemical Reactions 1.2 Mass Transport 4 1.3Kinetics of Electrode Reactions 6 1.4 References 10
1
2 An Overview of Electrochemical Techniques
11
2.1 Faradaic and Nonfaradaic Currents 11 2.2 Classification of Electrochemical Techniques 13 2.3 Two-Electrode and Three-Electrode Cells 14 2.4 An Overview of Voltammetric Techniques 15 2.5 The Nernst Equation in Potential Controlled Experiments 2.6 Common Reversible Redox Couples 18 2.7 References 21 3 Potential Step Experiments
22
3.1The Cottrell Experiment 22 3.2 Chronoamperometry 26 3.3 Chronocoulometry 28 3.4 Bulk Electrolysis 29 3.5 References 31 4 Potential Sweep Methods
32
4.1 Linear Sweep Voltammetry 32 4.2 Cyclic Voltammetry 34 4.3 Pulsed Voltammetric Techniques 4.4 References 44
37
5 Ultramicroelectrodes and Their Applications 5.1 Characteristics of Ultramicroelectrodes 45 5.2 Scanning Electrochemical Microscopy 49 5.3 Electrochemistry of Single Molecules 51 5.4 Conclusions and Outlook 53 5.5 References 53
45
17
XI1
Contents
6 Practical Experimental Methods
55
6.1 Electrodes and Working Electrode Surfaces 6.2 Solvents and Supporting Electrolytes 64 6.3 Basic Cell Design 68 6.4 Vacuum Methods 72 6.5 References 75
7 Digital Simulation
55
77
7.1 Principles of Digital Simulation 77 7.2 Simulations of the CV Behavior of a Simple Redox Couple 7.3 Simulation of Electron Transfer Reactions Coupled to Homogeneous Chemical Processes 84 7.4 References 87
79
8 Electrochemical Considerations for Supramolecular Systems
89
8.1 Intramolecular Forces under Electrochemical Conditions 89 8.2 Self-Assembly and Fixed Association in Supramolecular Structures: Implications for Reversible Redox-Switching 93 8.3 Systems Involving Multiple Identical or Non-Identical RedoxActive Moieties 94 8.4 References 102
9 Electrochemical Switching
103
9.1 The Concept of Electrochemical Switching 103 9.2 Switchable Binding in a Redox-Active Cation Host 105 9.3 Electrochemical Switching as a Means of Controlling Molecular Devices and Other Structures 109 9.4 References 113 10 Electrochemically Switchable Cation and Anion Binding
10.1 Electrochemically-SwitchedCation-Binding Systems 10.2 Electrochemically-SwitchedAnion Binding 122 10.3 References 125
114
114
11 Redox-Switchable Cyclophanes and Other Molecular Receptors 127 11.1Early Cyclophane Studies and Metallocyclophanes 127 11.2 Redox-Active Cyclophanes as Molecular Receptors 130 11.3Viologen Based Cyclophanes- the Ideal n-Acceptor Host 132 11.4 Electroinactive Cyclophane Hosts and Their Binding of RedoxSwitchable Guests 135
Contents
XI11
11.5 Other Molecular Receptors 11.6 Conclusions 139 11.7 References 139
137
12 Electroactive Intertwined Structures
142
12.1 Electroactive Cyclodextrin-Based Rotaxanes and Pseudorotaxanes 143 12.2 Templated Metallocatenates and Metallorotaxanes 145 12.3 Catenanes Based on x-Donor and mAcceptor Interactions 150 12.4 Rotaxanes and Shuttles Based on x-Donor/ Acceptor Chemistry 155 12.5 Perspectives on the Future of Molecular Devices 160 12.6 References 161 13 Helicates, Racks Grids and Coordination Arrays 13.1Helicates 164 13.2 Molecular Racks, Grids and Coordination Arrays 13.3Conclusions 177 13.4 References 178 14 Electroactive Langmuir-Blodgett Films
164
175
180
14.1 Langmuir-Blodgett Films 180 14.2 Electron Transfer Studies in Langmuir and Langmuir-Blodgett Films 180 14.3 Other Electroactive LB Film Studies 183 14.4 References 190
15 Self-Assembled Monolayers
191
15.1 SAMs as Barriers for Electron Transfer Reactions 15.2 Electroactive Monolayers 195 15.3 Molecular Recogrution in SAMs 198 15.4 Photoswitchable SAMs 203 15.5 References 206 16 Electroactive Dendrimers
193
207
16.1 Dendrimers with Peripheral Electroactive Groups 207 16.2 Dendrimers with internal Electroactive Groups 213 16.3 References 220 17 Molecular Wires
222
17.1 The Concept of a Molecular Wire and its Electron Transfer Kinetics 17.2 Electrochemical Studies of Molecular Wires 223 17.3 References 227
222
XIV
Contents
18 Conclusions and Outlook 18.1 References
Index
233
231
228
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
1 Fundamentals of Electrochemical Theory
Electrochemistry is a branch of science with a long and prestigious history. The theoretical foundations of electrochemistry were laid out by Faraday, Volta, Galvani and many other prominent scientists; their names are now routinely used to designate constants, units, processes, or types of cells. Electrochemistry can be defined in a very general way as the study of chemical reactions to produce electric power or, alternatively, the use of electricity to affect chemical processes or systems. The first perspective concerns the so-called galvanic cells, while the second relates to electrolytic processes. Both have tremendous practical importance, industrially as well as in everyday life. From the electrolytic preparation of chlorine to the widespread use of batteries, electrochemistry is a branch of science that has a clear and marked impact in everyone's life. While the user of a cellular phone whose battery dies in the middle of an important conversation might all too clearly perceive the limitations of electrochemical technology, it is equally true that deveIopments and advances in electrochemical science hold the key to some important technological breakthroughs. Electric cars afford the primary example for this situation because attractive operational characteristics --that will make them competitive with vehicles based on the internal combustion engine-- require batteries with higher power densities and peak power outputs. As these better batteries become available, the feasibility and popularity of electric vehicles should improve.
1.1 Cell Potentials and Electrochemical Reactions As the simplest type of chemical reactidn, electron transfer processes are at the core of electrochemistry. Electrons, the key players in these phenomena, are also the carriers of electricity in metallic and semiconductor circuits. Therefore, the connection between chemistry and electricity is obvious. The science of electrochemistry has its origins in the fact that oxidation-reduction reactions can be performed in ways that allow the direct harvesting of the free energy released in these processes. Consider, for instance, the following spontaneous reaction
While it is possible to immerse Zn metal in a solution of Cu(I1) ions and observe the oxidation (dissolution) of the Zn metal along with the simultaneous reduction.of the Cu(I1) ions (to form metallic Cu deposits), the same overall reaction can be carried out by immersing a Zn strip in a solution of Zn(I1) ions
2
1
Fundamentals of Electrochemical
and a Cu strip in a solution of Cu(I1) ions (see Fig. 1.1).To start the reaction, one only needs to establish pathways for the charges (electrons and ions) to circulate between the sites at which the Zn oxidation and Cu(I1) reduction processes take place. This is accomplished by setting up a salt bridge to establish electrical contact between the two solutions. The salt bridge allows the circulation of ions between the two solutions while preventing their mixing. Under these conditions a potential difference between the Zn and Cu strips develops. If the circuit is closed externally, that is, if a so-called electrical "load" is connected to the metal electrodes, the existing potential difference will give rise to a current, a flow of electrons moving from the Zn electrode (negative pole) to the Cu electrode (positive pole). The free energy AGO of the overall chemical reaction taking place in the cell can be readily calculated as
where n is the total number of electrons transferred in the reaction, F is Faraday's constant and Eoceu is the standard cell potential of the cell.
t
I
Cathode
Anode Salt bridge Figure 1.1: Components of a Galvanic Cell.
Electrochemical reactions are heterogeneous in nature as they take place at interfaces, usually metal-solution boundaries. These active interfaces are usually referred to as electrodes. By definition, an electrode where a reduction
2 .I
Cell Potentials and Electrodlemical Reactions
3
(uptake of electrons by a solution species) takes place is called a cathode. Conversely, an anode is an electrode where an oxidation (loss of electrons by a solution species) occurs. Applying these definitions to the electrodes of the galvanic cell in Fig. 1.1,it is straightforward to conclude that the Zn electrode is the anode and the Cu electrode serves as the cathode. A net electrochemical reaction implies transfer of charge across the corresponding metal solution boundary and the flow of current across the electrode. The current i, a basic electrical quantity, affords an instantaneous measurement of the rate of the electrochemical reaction according to equation (3)
i = nFAr
(3)
where n is the number of electrons transferred in the interfacial reduction or oxidation process, F is Faraday's constant, A is the surface area of the metal solution boundary, and r is the instantaneous reaction rate. Since current measurements are easily done with modern instrumentation, a peculiar feature of electrochemical techniques is that they provide continuous monitoring of the reaction rate. Integration of the current over a period of time affords the electrical charge, Q, which can be transformed into the amount of material in moles, N, converted in the electrochemicalreaction using Faraday's law: Q=nFN
(4)
A third quantity of fundamental importance in electrochemistry is the electrode potential, which can be considered as an adjustable driving force for the electrochemical reactions. In general terms, as the potential of an electrode is made more negative, the average energy of the electrons in the metal, which is approximately equal to its Fermi level, becomes higher, giving the electrode more reducing power. Similarly, the oxidizing power of an electrode can be increased by making its potential more positive. While these qualitative arguments are perfectly straightforward, the definition of electrode potentials is complicated by the fact that the potential of a single electrode is not an experimentally measurable quantity. This experimental inaccessibility has given rise to many theoretical attempts to obtain absolute electrode potentials. However, to the authors' knowIedge, none of these attempts has gained universal acceptance and, therefore, relative values continue to be the only way in which electrode potentials can be quoted. Simply put, this means that electrode potentials are always measured versus a second, reference electrode, whose value is arbitrarily taken as zero. The potential of the normal hydrogen electrode (NHE) is generally assigned a standard value of zero and serves thus as the primary reference for any other electrodes. For a generalized process involving the transfer of n electrons, Ox -+ ne eRed
(5)
1 Fundamentals of Electrochemical
4
where Ox and Red represent the oxidized an reduced partners of the redox couple, the thermodynamic potential, E, of the corresponding electrode is given by the well known Nernst equation, which is unquestionably one the most important equations in electrochemistry,
RT a,, E = E" +-InnF aRed where Eo is the potential under standard conditions, uox and aRed are the activities of the oxidized and reduced species, respectively, and the remaining terms have their usual meaning. Extensive tabulations of standard potential values are available. To avoid the complications associated with the use of thermodynamic activities and activity coefficients, very often activities are replaced by concentrations. In this case, the standard potential is replaced by the formal potential, Eo', which is usually dependent on medium conditions since it includes the activity coefficients. Therefore, a more practical version of the Nernst equation is as follows
E = E"' +
2.303RT nF
log
[Ox] ~
[Redl
(7)
The factor 2.303 reflects the replacement of natural by decimal logarithms. At 25OC, 2.303RT/F is equal to the familiar O.O5916V, which every freshman chemistry student ends up committing to memory. The Nernst equation is a thermodynamic equation and, thus, can only be rigorously applied to equilibrium situations (i=O). In spite of this apparent limitation, eq. 7 is successfully applied when current flows across the electrode in question, as long as the heterogeneous electron transfer process is fast (reversible in electrochemical jargon). Under these conditions, the equation is useful to calculate the concentrations at the electrode surface of Ox and Red that are generated when specific potential values are imposed to the electrode. Fast electron transfer kinetics allows the electrochemical reaction to adapt quickly to the changing potential values on the electrode surface, maintaining a pseudoequilibrium situation as well as the validity of the Nernst equation. Therefore, the term nernstian is also used when describing kinetically fast or reversible electron transfer processes. Finally, we must point out the potential of a galvanic cell, such as that represented in Fig. 1.1, can always be calculated with the following equation,
in which the cathode and anode potentials are obtained individually using eq. 7.
1.2 Mass Transport Current is simply the movement of ions and/or electrons across conducting media. In electrochemical cells, the movement of charged and neutral species is
1.2 Mass Transport
5
fundamentally important. Quite often it is the rate of these movements that determine the potentials and currents measured in the cell. No treatment of electrochemistry can thus overlook mass transport mechanisms. The three relevant mechanisms that may arise in electrochemical cells are migration, convection, and diffusion. In most electrochemical techniques, conditions are chosen so that transport of the electroactive species is affected by a single mechanism, typically diffusion. A brief discussion of each of these modes of mass transport follows. Migration is the movement of ions under the mfluence of an electric field. Therefore, uncharged species are not affected by migration. Although migrational movements can be described mathematically, in most voltammetric techniques it is desirable to remove migration contributions to the mass transport of the primary electroactive species, that is, the molecule or ion under study or analysis. This is accomplished by adding a large excess of an easily ionizable salt, which will dissociate to produce a large amount of inert anions and cations. These ions become the migration current carriers, thus releasing the electroactive species (if charged) from migration effects. The ionizable salt used for this purpose is called the supporting electrolyte. To be effective, its concentration must be about 100 times higher than that of the electroactive species. A second beneficial effect of the supporting electrolyte is to increase the conductivity of the solution, thus decreasing cell resistance effects that are very detrimental for recording accurate current responses. Convection is mass transport resulting from movements of the solution as a whole. Convection can be driven by stirring, solution flow, or by movements of the electrodes. In quiet, thermostatted electrochemical cells, convection may arise from density gradients only after rather long experiments. In fact, it is usually the onset of convection that limits the maximum duration of voltammetric or chronoamperometric experiments. In shorter experiments convection is not a factor in mass transport as long as the solution is quiescent and the electrodes are stationary. D z ~ s i o nis mass transport driven by a gradient of chemical potential. Anytime that the concentration of a molecule or ion (charge is of no concern here) is uneven throughout a solution, mass transport will take place to restore the homogeneity of the solution. In other words, transport will proceed from regions of high concentration to regions of low concentration. Diffusional phenomena are very important across many scientific and engineering disciplines. Fortunately, diffusion can be described mathematically, which facilitates the quantitative treatment of many electrochemical phenomena. The rate of diffusion of any chemical species is described by its diffusion coefficient, D, that is usually expressed in units of cm2/s. Most small organic or inorganic molecules or ions have D values in the vicinity of 10-5 cm2/s. This value decreases with molecular size. For instance, for spherical molecules the StokesEinstein equation establishes that
1 Fundamentals of Electrochemical
6
where k is the Boltzmann constant, q is the solution viscosity and a stands for the effective hydrodynamic radius of the diffusing species. This equation also reveals explicitly that D values depend on the temperature and the composition of the solution. To quantitate one-dimensional diffusion rates the concept of material fzux is very useful. The diffusional flux, J, is defined as the number of particles crossing a unit surface area perpendicular to the direction of mass transport per unit time. Fick's first law establishes that the flux is directly proportional to the concentration gradient. The proportionality constant is precisely the diffusion coefficient, that is, J = -D .
(g)
and the negative sign denotes the fact that the material flux moves against the gradient. This equation is extremely useful to calculate currents under conditions of complete conversion, i.e., whenever all the molecules or ions reaching the electrode surface undergo instantaneous electrochemical reaction. In such cases, the flux at the electrode surface is directly proportional to the resulting current. Fick's second law permits the calculation of concentration changes as a function of time. Its mathematical expression is given below
"=.($) at
Fick's laws provide a complete and detailed description of diffusional mass transport for any species subject to concentration gradients. To find the analytical solutions of the resulting differential equations, appropriate boundary conditions must be provided detailing initial and limiting concentrations and extent of electrochemical conversion at the electrode surface. Some examples will be given in later chapters.
1.3 Kinetics of Electrode Reactions In most electrochemical experiments we are interested in recording a currentpotential curve. For instance, let us assume that we apply an increasingly positive potential to an electrode (or that we make its potential increasingly positive against a reference electrode). The more positive the potential becomes, the more oxidizing power is conferred to the electrode and, at some point, one of the cell components will start to undergo an oxidation reaction. This reaction will translate into current flowing across the cell, a situation that is represented in Fig. 1.2. Notice that this curve is composed of three distinctive regions. In the first region (low potentials), there is no significant current flow, because the potential is not sufficiently positive to drive the oxidation process. In the second region (intermediate potentials), the current increases with the potential, as one
1.3
Kinetics of Electrode Reactions
7
would generally anticipate from simple kinetic arguments. A third region (high potentials) is characterized by the leveling of the current, which reaches a constant or limiting value independent from the potential. This is due to limitations imposed by the finite rate of mass transport that can be achieved in the solution.
Region I
Current
D
Potential Figure 1.2 Typical current-potential curve.
An electrode reaction is a heterogeneous process that takes place at the interface between the electrode and the solution. Therefore, the overall rate or current depends on the rates of two distinct processes: the actual heterogeneous electron transfer process and the transport of the reactant species from the solution to the electrode surface. The slowest one of the two processes determines the overall current. Fig. 1.2 illustrates this situation clearly. In the intermediate potential region the kinetics of the electrode reaction controls the current level. In this region mass transport is still sufficiently fast to be "transparent", that is, it shows no effect on the overall current. However, at higher potentials, the electrochemical reaction is driven to very fast rates, increasing the demand for electroactive species to an extent that it becomes impossible for mass transport to keep pace. Therefore, a current plateau
1 Fundamentals of Electrocllemical
8
develops as the current reaches the maximum limit that mass transport processes can provide. These ideas are mathematically expressed by the simple equation: 1
1
i
i,
- =-+,
1 1,
in which i stands for the overall current, ik is the current that can be obtained at that potential and zr is the limiting current that can be reached through mass transport. In this section we will describe the potential dependence of the current assuming no limitations from mass transport. Any theoretical formulation of electrochemical kinetics must reduce to the thermodynamic limit (Nernst equation) when equilibrium is reached. Furthermore, the empirica1 Tafel equation establishes a mathematical relation between the current and the overpotential q (difference between the applied potential and the corresponding equilibrium potential for the electrode system in question) q = a + b .log i
(13)
where a and b are constant values characteristic of the system. Let us consider a generalized heterogeneous electron transfer process between species Ox and Red (see eq. 5). Using eq. 3, we can write for the forward reaction (Ox+Red) ic rf = k f.Cox(O,t)=-nFA
in which kf is the rate constant for the forward reaction, i, is the cathodic current. Notice that the heterogeneous character of the process is manifested by the fact that the reaction rate is directly proportional to the reactant concentration at the electrode surface Cox(O,t). We can write a similar equation for the reverse or backward process (Red+Ox)
The total current i flowing through the electrode is simply the difference between the cathodic and the anodic currents,
i = i, - i a = nFA(k, .cox(0, t) - kb .CRed(0,t)]
(16)
The way this equation is written implies that we have chosen to describe cathodic currents as positive and anodic currents as negative. This is a common, albeit completely arbitrary, choice that we will maintain throughout the book. Notice also that the rate constants have units of cm/s, a reflection of their heterogeneous character, provided that they operate on concentrations expressed in mol/cm3.
I .3 Kinefics ofElectro& Reactions
9
A key distinguishing feature of electrochemistry is that the reaction rates depend on the applied electrode potential. In fact, to further develop eq. 16 we must provide mathematical expressions to describe this dependence. The Butler-Volmer formulation is the most commonly used for this purpose. The corresponding equations are as follows k - kO .e-anF(E-Eo')/RT f -
and
k - kO -e(l-a)nF(E-Eo')/RT
(18)
b -
where ko is the standard rate constant and a is the so-called transfer coefficient.[*] It is possible to derive these equations using several physical models, but we will constrain ourselves here to explore some of the implications of the ButlerVolmer formulation. At equilibrium (E=Eeq) the net current is zero. By combining eqs. 16,17 and 18 we have nFAkOCo x (0, t) . e-anF(Eeq-Eo')/RT
= nFAk°CRed(0,t) . e
(1-a)F(E,,
-ED') / RT
Under equilibrium conditions, the concentrations of Ox and Red at the electrode surface are identical to those in the bulk solution and, thus, we can write
which is identical to the Nernst equation (eq. 7). The electrochemical equilibrium, as any other type of chemical equilibrium, is not static. In fact, the forward and backward processes take place at equal rates yielding no net current. However, the electrochemical activity at equilibrium can be expressed in terms of the exchange current, io, which is identical to the level of cathodic or anodic currrent. For instance, i o = i f = i b =nFAk 0 e -unF(E,,-E'')/RT which, after some manipulation, yields
io = nFAko[Oxll-a)[Red]"
(22)
The exchange current is directly proportional to the standard rate constant for the heterogeneous electron transfer process. Both parameters are used to express quantitatively the inherent rates of heterogeneous electron transfer reactions. Outside equilibrium conditions (q#O) the Butler-Volmer formulation leads to an important equation which is generally valid to describe the kinetics
1
10
Fundunientuls of Electrocllpmicul
of electrochemical reactions in the absence of mass transport limitations. Not surprisingly, this equation is commonly referred to as the Butler-Volmer equation and is given as
The right term in the equation describes the cathodic component (forward reaction) of the current while the left term describes the anodic component (reverse reaction). Of course, the sign of the overpotential determines which one of the two terms will predominate and control the overall current. At negative overpotentials (E<Eeq) the cathodic term predominates and at positive overpotentials (E>Eeq) the anodic term controls the total current. The transfer coefficient a is related to the degree of asymmetry in the electron transfer process. Many simple, one-step electrochemical reactions exhibit values of a close to 0.5. Kmetically sluggish processes or multi-step reactions may present transfer coefficients substantially different from 0.5. The Butler-Volmer equation reduces to Tafel conditions at extreme overpotentials. For instance, if q<
i = i , e -anFq / RT as the anodic term becomes negligibly small. Taking logarithms on both sides of the equation and rearranging yields
RT anF
q =-hi,
RT -----hi anF
which is identical in form to the empirical Tafel equation (eq. 13). It is also of interest to explore the Butler-Volmer equation under conditions of small overpotentials. Using the fact that for very small values of the exponent (x) the exponential (ex) can be approximated by l+x, it is easy to show that . i,nF 1 = --
RT
'
which demonstrates that the current is linearly related to the overpotential in a narrow range of potentials around Eeq. The kinetics of electrode processes has tremendous sigruficance in several technologically important fields. Unfortunately, many electrochemical reductions or oxidations of simple molecules occur via complicated, multi-step mechanisms. This makes their study far more difficult than this short treatment may reflect.
1.4 References 1. Like kO, a is temperature dependent. See M. T. M. Koper, J.Phys. Chem B, 1997, 202,3168-3173.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
2 An Overview of Electrochemical Techniques
In this chapter we will introduce several important ideas and modes of operation related to electrochemical techniques. We will also attempt to provide a classification of electrochemical techniques. The reader should keep in mind that the focus of the book is supramolecular chemistry. Therefore, some techniques that have not found many applications in this field of chemistry may not be appropriately represented here.
2.1 Faradaic and Nonfaradaic Currents Faradaic currents result from heterogeneous electron transfer processes. In faradaic electrochemical processes a component of the electrode-solution interface must be oxidized or reduced. Normally, this component is a solute in the solution side of the interface, but in some cases the electrode may also supply the corresponding reactant, i.e., oxidation of the metal surface to form insoluble oxides or soluble salts. Nonfaradaic currents originate from processes that do not involve electron transfer reactions across the electrode-solution interface, that is, no oxidations or reductions take place. Most often, nonfaradaic currents originate from the electrical capacitance present at the interface. It has been long known that the disruption of the bulk solution structure by the electrode and the density of electric charges at its surface exert a strong organizing effect on the adjacent solution layer. The classical term electrical double layer is still used to describe this interfacial region, which not surprisingly behaves like a capacitor. Numerous electrochemists have devoted intense research efforts to structural and physical studies of the double layer. Although this is a mature field of electrochemical research, it is also out of the scope of this book. Interested readers are referred to other -more authoritative- sources.[11 Like any capacitor, the electrode-solution interface gives rise to charging or discharging currents when the electrical potential across the interface changes. These currents flow transiently to insure that the two sides of the interface (equivalent to the two plates of a capacitor) adopt charges (9) commensurate with the effective potential according to the equation C=q/E
(1)
where C stands for the capacitance, and E is the effective potential across the electrode-solution interface. It can be readily shown[zI that, when the potential
12
2 An Ovewieru of Efectrocllemicul Techniques
across an electrode is changed to a fixed value E, a level of nonfaradaic current i will flow as described by
in which R, is the resistance of the cell and C is the capacitance of the electrode, which is assumed here to dominate the total capacitance of the electrochemical cell. This equation reveals that the current decays with time at a rate that depends on the product RsC, which is the time constant of the electrochemical cell. Ideally, the experimenter must always try to minimize this product, as shorter time constants guarantee the fast decay of nonfaradaic or chargzng currents. However, in experiments in which the potential varies continuously, such as in the linear potential scans often used in voltammetric techniques, the nonfaradaic current is approximated by[V
where Ei is the initial potential value and v is the scan rate, i.e., the rate of change of the potential vs time. In deriving eq. 3, it is assumed that the electrode capacitance C is constant over the potential range scanned. Notice that in this case the nonfaradaic or charging current is composed of two terms. The first one is independent from time and the second one exhibits the same time dependence expressed in eq. 2. This means that, regardless of how small the time constant might be, a continuous level of charging current (given by the product v.C) will flow throughout the experiment. From all these considerations it should now be clear why electrochemists have vigorously searched for methods to minimize nonfaradaic currents or to separate faradaic from nonfaradaic currents. The goal of minimizing nonfaradaic currents has led to the development of pulse voltammetric techniques. Using sophisticated electronic circuitry these methods can effectively reject a substantial fraction of nonfaradaic currents, thus increasing the analytical sensitivity, as small faradaic currents become readily measurable. The idea of fully separating faradaic from nonfaradaic currents has also attracted a lot of attention. However, most electrochemists have come to the conclusion that it is not possible to accomplish this separation rigorously. Let us consider any electrochemical experiment. One may perform the experiment in the absence of the oxidizable or reducible species and use the results as a blank to be subtracted from the data obtained in the presence of the target analyte. This scheme may appear to offer a good chance to fully remove nonfaradaic currents and produce clean faradaic currents. Unfortunately, there is a pitfall in this reasoning; nonfaradaic currents may change in the presence of the electroactive species, because its presence affects the structure of the double layer and, thus, the electrode capacitance. The proposed subtraction or blank removal scheme may work sometimes in an approximate fashion, but faradaic
2.2
Classification of Electrodreniical Techniques
13
and nonfaradaic current contributions are not truly separable. In any instance, nonfaradaic processes affect every electrochemical technique and their sensitivities and detection limits depend to a large extent on their ability to reject nonfaradaic currents.
2.2 Classification of Electrochemical Techniques One of the ways in which the breadth of electrochemistry makes itself evident is by the proliferation of electroanalytical techniques. It seems pertinent then to start the discussion of electrochemical techniques by providing a classification.L31 In principle, all electrochemical techniques can be divided in two classes, which we refer to as bulk and interfacial. Bulk techniques are based on phenomena that take place in the core of the solution. Conductometry constitutes an excellent example since the measured property, conductance, depends on the mobility of ions migrating through the solution phase. By contrast, interfacial techniques focus on phenomena taking place at the electrode-solution interface. Most electrochemical techniques of relevance to supramolecular chemistry fall in the latter class. Interfacial electrochemical methods can be further classified into equilibrium and dynamic methods. In equilibrium methods, measurements are taken under equilibrium conditions, that is, when there is no net flow of current through the cell. The best example of this class of techniques is afforded by potentiometric methods. In dynurnic methods, measurements are made in a cell that has been displaced away from equilibrium conditions. The measurements are done under conditions of potential or current control, which are termed also potentiostatic or galvanostatic conditions, respectively. While supramolecular chemistry, and particularly host-guest chemistry, has substantially influenced the development of potentiometric sensors, this book is clearly biased by the authors' expertise toward dynamic electrochemical methods. We must note that dynamic techniques not only make possible the investigation (thermodynamic, kinetic, etc..) of supramolecular systems, but they can be used as well to stimulate the system through the application of potential or current signals. This important idea will be further developed in later chapters. As mentioned before, dynamic methods can be divided into potential controlled (potentiostatic) and current controlled (galvanostatic) methods. Pofentiostutic methods are unquestionably more popular and include a large array of techniques, such as voltammetry, chronoamperometry, and bulk electrolysis. Indeed, all of these techniques involve current measurements while the cell potential is controlled. On the other hand, galvanostatic techniques monitor the cell potential under controlled current flow across the cell. Examples include coulometric titrations and the old-fashioned electrogravimetry.
14
2 An Overview of Electrochemical Tecliniques
2.3 Two-Electrode and Three-Electrode Cells Many electrochemical experiments can be conducted with a simple cell composed of two electrodes. Normally, the interest of the experimenter focuses in one of these electrodes, the so-called indicator or working electrode, but a second electrode is necessary simply because we cannot perform any useful electrochemical measurements with a single electrode. Therefore, the second electrode is usually a reference electrode with a fixed, known potential. For potentiometric methods this arrangement is ideal since the electrodes are not polarized away from equilibrium conditions and measurements are recorded in the absence of current. However, in dynamic methods the passage of current, which is caused by the prevalent nonequilibrium conditions, polarizes the electrodes and displaces them from their equilibrium potentials. This creates a particularly acute problem with the reference electrode. If the potential of the reference electrode shifts from its equilibrium value, how can we measure meaningful potentials in such a cell? The first solution to this problem was to use reference electrodes with very large surface area in order to maintain low current densities and, thus, minimize the polarization of these electrodes. Although this scheme often works fairly well in practice, the complete elimination of reference electrode polarization is only feasible with modern potentiostatic instrumentation based on operational amplifiers. These instruments require three electrodes for ideal operation. In addition to the working and reference electrodes, an auxiliary or counter electrode is added to complete the current loop with the working electrode, as shown in Fig. 2.1. The
Potentiostat
Cell
Figure 2.1: A three-electrodecell. W stands for working electrode, R stands for reference electrode and C stands for counter electrode.
2.4
An Overview of Voltammettic Techniques
15
key to this arrangement is that the connection of the reference electrode to the potentiostat is made at a point with very high input impedance. As a result, the current is diverted to the third electrode, the counter electrode, which will adopt any potential necessary to develop a current equal in absolute value and opposite in sign to that originating at the working electrode. Since the reference electrode is protected against current flow, its composition and potential remain constant and, thus, the potential difference measured or applied between the working and the reference electrode maintains its significance throughout the experiment.
2.4 An Overview of Voltammetric Techniques The group of voltammetric techniques started its development about three decades ago as an outgrowth from polarography. The toxicity of mercury has little by little taken a toll in the popularity of polarography while voltammetry in its many forms has flourished to become the most important and often used class of electrochemical methods. In any voltammetric technique, a potential function is applied to the working electrode. The response of the electrochemical cell is measured by recording the current flow. Usually, the electrochemist is interested in the corresponding current-potential curve or voltammogram. The different types of voltammetric techniques result from the diversity of potential excitation functions that can be applied to the working electrode as well as from the various ways to record and plot currents. A summary of the most common voltammetric techniques is given in Fig. 2.2. The two first techniques given in the figure (LSV and CV) were developed before the advent of digital electronic circuitry. Thus, the current is continuously measured throughout the whole potential scan. Modern implementations of these techniques using digital potentiostatic circuits make use of staircase potential excitation functions to approximate the required linear potential ramps. The current level is measured at each step in the potential staircase, normally often enough to approach closely the continuous current recorded with older, analog instrumentation. By contrast, the other three techniques shown constitute examples of pulse techniques. These more sophisticated methods rely on fast potentiostatic circuits to impose pulsed potential excitation functions with accurate timing. Current measurements (marked by arrows in the figure) are done also at carefully timed positions during the excitation function. Notice that all current measurements are done at the end of potential pulses, after the potential has been poised at a fixed value for a time long enough so that the charging current has decayed to small residual values. This scheme for current measurements gives pulse voltammetric techniques a great advantage over their linear sweep precursors, as it allows a substantial degree of rejection of nonfaradaic current contributions and affords much lower detection limits. Nowadays most manufacturers of electrochemical instrumentation offer computer controlled instrumentation capable of running all of these techniques
16
2
Technique
An Overview of Ekctrochernicd Techniques
Potential excitation function
Linear sweep voltammetry (LSV)
Cyclic voltammetry (CV) I
b
t
4 E
Normal pulse voltammetry (NPV)
t
E Square wave voltammetry (SWV)
Differential pulse voltammetry (DPV)
E
t
Figure 2.2: Names and potential excitation functions for the most common voltammetric techniques (the arrows indicate current measuring points).
2.5
TIE Nernst Equation in Potential-Controlled Experiments
17
in user-friendly ways. The operator must only make the connections from the potentiostatic instrument to the electrodes in the cell, select the technique to be run, and spec+ the parameters to fully describe the excitation function. The instrument takes care of the rest and stores the system response for further use.
2.5 The Nernst Equation in Potential Controlled Experiments As was mentioned in Chapter 1, the Nernst equation has a thermodynamic origin and, rigorously speaking, can thus only be used under equilibrium conditions, that is, in the absence of net current flow. However, in potential controlled experiments involving redox couples (Ox/Red) capable of fast heterogeneous electron transfer, the application of potentials to the working electrode results in the fast adjustment of the surface concentrations of Ox and Red to the ratio predicted by the Nernst equation. In other words, when the overall current flowing across the cell is not controlled by electrochemical kinetics, the Nernst equation can be used to determine the concentrations of electroactive species at the electrode surface as a function of the applied potential, that is,
E
RT, In [oxlo ~ = EO' ~ +~nF [Red], ~
(4)
where E,,p is the applied potential, n is the number of electrons exchanged by the Ox/Red couple, Eo' is the formal potential of the couple and [Oxlo and [Redlo are the respective concentrations at the electrode surface. As these electrochemical processes are kinetically very fast, the current flow is only limited by the rate of the mass transport processes in the solution. Due to the prevalence of eq. 4, these electrochemical processes (or the corresponding redox couples) are termed nernstian or reversible. A little more reflection seems appropriate here on the significance of eq. 4. This equation is the fundamental reason for the observation of voltammetric peaks or waves. Assume that we start an experiment with a moderate concentration of the oxidized form (Ox) of a reversible redox couple and no added Red. Normally, electrochemical experiments are started at a potential at which there is no faradaic current. Therefore, we will start the experiment at a potential Ei much larger than EO', so that [Oxlo is much higher than [Redlo, in agreement with the stated initial conditions. As the potential is scanned in the negative direction (cathodic scan) towards the formal value, the ratio of concentrations [Ox]o/[Red]ostarts to decrease according to eq. 4, which results in current flow as Ox is converted to Red. Eventually, as the applied potential becomes more negative than the formal value, Red will start to predominate and [Oxlo/ [Redlo becomes smaller and smaller. Whether the current peaks and decays or reaches a limiting value will depend on the details of the mass transport mechanisms. However, the observation of the voltammetric peak or wave is a manifestation of crossing from potentials that
2 An Overuiao of Elecfrocllemical Techniques
18
favor Ox to potentials that favor Red at the electrode surface and, therefore, reflects the electrochemical equilibrium between the two partners of the redox couple. In this sense, voltammetric peaks are inherently different from spectroscopic peaks, because they represent the conversions between two different, equilibrating redox partners. By contrast, spectroscopic peaks originate from the absorption of energy by a species to undergo a temporary transition to a short-lived excited state from which it quickly relaxes back to the ground state. The formal potential, EO',is by itself an expression of the relative stability of the two redox partners in the medium of interest. Any relative stabilization of Ox (or destabilization of Red) will cause a shift of the formal potential to more negative values. Conversely, a EV shift to more positive values reveals the relative stabilization of Red (or destabilization of Ox). These simple ideas are extremely useful in supramolecular electrochemistry as we will see in the following chapters.
2.6 Common Reversible Redox Couples The purpose of this section is to provide an overview of the electrochemical behavior of some fast, reversible redox couples usually encountered in supramolecular systems. The idea is to familiarize the reader with some of the electrochemical processes that will be found often in the remaining chapters of the book. Ferrocene or bis(cyclopentadienyl)iron(II) is probably the most used electroactive residue in supramolecular chemistry because of its synthetic accessibility. Ferrocene undergoes a fast one-electron oxidation to yield the cationic ferrocenium species at very accessible potentials. Ferrocene is a very stable 18-electron system. Its oxidized form is stable in most organic solvents, but can decompose quickly in aqueous media, especially in basic conditions. Ferrocene can be used as an internal reference (also known as a reference redox couple) for electrochemical experiments performed in low dielectric solvents that do not permit use of a standard reference electrode. See Chapter 6 for further information on this issue.
-e
&
(5)
Cobaltocenium is also a stable 18-electron bis(cyclopentadieny1) metal complex. It undergoes reduction to form the 19-electron cobaltocene at fairly negative
2.6
Common Reversible Redox Couples
19
potentials, accessible in nonaqueous solvents and under some conditions in aqueous media.
Tetrathiafulvalene, like ferrocene, is easily oxidizable. However, the resulting cation radical also can be reversibly oxidized at more anodic, although still accessible, potentials. These electrochemical processes are typically studied in nonaqueous solvent since the solubility of tetrathiafulvalene (TTF) derivatives is usually poor in aqueous media. As it is usually the case for two consecutive one-electron processes, the second transfer is kinetically slower than the first.
t
ps]p - 4 3
2+
l+
I-
S
S
S
S
Violonen is the common name used to designate N,N'-disubstituted 4,4'bipyridinium derivatives. These diquatemized compounds undergo two fast one-electron reductions as shown below. These processes are normally uncomplicated in nonaqueous solvents. In aqueous media, the electrochemical reactions are complicated by the low solubility of the reduced forms or by dimerization/oligomeriation reactions of the cation radicals. The simplest viologen, methylviologen (R=C&), is also an effective herbicide, that is widely known by its common name, paraquat.
20
2
An Overview of Electrocllemicaf Techniques
Ouinones are also commonly used in supramolecular systems as a reflection of its importance in natural electron transfer systems. In aqueous media the electrochemistry of most quinones is strongly pH dependent. In nonaqueous, aprotic solvents, quinones undergo two consecutive one-electron processes according to the following equations for naphthoquinone,
Nitrobenzene is another electroactive functional group relevant to a number of supramolecular systems. In aprotic solvents nitrobenzene undergoes two consecutive one-electron reductions. However, the second one is very sensitive to the presence of traces of water or nucleophiles and very often the focus is on the first reduction process.
Metal complexes. A wide array of metal complexes also play important roles in supramolecular chemistry and many of these are electroactive. Perhaps the two simplest complexes with redox activity that find use in supramolecular systems are hexacyanoferrato(I1) and hexaammineruthenium(II1). The first complex has
2.7
References
21
a very accessible one-electron oxidation, while the second one is easily reduced in an extremely fast one-electron process.
Finally, we cannot close this section without mentioning the porphyrin metal complexes and the large and important family of transition metal complexes with 2,2'-bipyridine and related ligands. All these metal complexes constitute important building blocks for the construction of numerous supramolecular systems and their redox activity affords important methods for the investigation of their properties.
2.7 References 1.For a recent authoritative review, see: A. J. Bard, H. D. Abrufia, C. E. Chidsey, L. R. Faulkner, S. W. Feldberg, K. Itaya, M. Majda, 0. Melroy, R. W. Murray, M. D. Porter, M. P. Soriaga, H. S. White J. Phys. Chem. 1993,97,7147-7173. 2. A.J. Bard and L. R. Faulkner Electrochemical Methods: Fundamentals and Appficafions,Wiley: New York, 1980; Chapter 1. 3. The classification given here is inspired by that in D.A. Skoog, F. J. Holler, and T. A. Nieman Principles of lnstrumental Analysis, 5 t h Ed.; Harcourt Brace: Philadelphia, 1998:Chapter 22.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
3 Potential Step Methods
In potential controlled electrochemical methods the working electrode is subjected to a so-called excitation function, that is, a potential that varies as a function of time. The simplest type of excitation function is the potential step, which implies an abrupt change of the potential of the working electrode from an initial rest value, at which no current flows, to a final value, at which the faradaic process of interest takes place at an appreciable rate. The sudden potential change is done at a certain time which, for convenience, is usually referred to as t=O. Often, the potential of the working electrode is set to a final value at which the electroactive species becomes unstable in contact with the electrode surface. For a reducible substance (Ox), this condition is simply met when the applied potential is at least 150 mV more negative than the formal potential (Eo’) of the corresponding redox couple (Ox/Red). Of course, this assumes that the electrochemical kinetics is fast in the timescale of the experiment. These types of experimental conditions are referred to as Cottrell conditions and give rise to faradaic currents which are determined by the rate of mass transfer of the electroactive species towards the electrode surface.
3.1 The Cottrell Experiment Let us consider a planar electrode immersed in an unstirred electrolyte solution and review the conditions of the Cottrell experiment from a more mathematical standpoint. Initially, we start with a perfectly homogeneous solution, as expressed by the following equations at t=O,
and the applied electrode potential is chosen so as not to disrupt these concentrations, that is, Eappp>Eo’, which favors the oxidized species. Here, we have selected to work with a solution containing a finite bulk concentration of Ox, Cb,, and no Red initially. This simplifies the mathematical equations somewhat, but other cases can be found in practice. As soon as the experiment starts (f >O) the applied potential is suddenly taken to a much more negative
3.7
The Cottrell Experiment
23
value, Eapp$<Eo’, at which the fast electrolysis of Ox fully depletes its concentration at the electrode surface (x=O), Cox(O,t) = 0 , for f
>O
(3)
At finite times after the beginning of the experiment, the surface concentration of Ox is reduced to zero, but its concentration remains unaltered and equal to its bulk value if we move sufficiently far away from the electrode surface. Mathematically, this condition can be expressed as:
Also, the principle of mass conservation requires that
Under these conditions, since the kinetics of heterogeneous electron transfer is very fast, the current flowing across the electrode is limited only by the rate of mass transfer of Ox from the solution to the electrode surface. The purpose of this section is to provide a mathematical expression to describe the resulting current as a function of time. As it turns out the solution to the problem depends on the size of the electrode because its size determines the type of diffusional mass transport prevalent in the experiment. This is clearly illustrated in Fig. 3.1. For a large planar electrode, edge effects are unimportant and semimfinite planar diffusion describes perfectly the mass transport situation encountered. In contrast, for small disk electrodes the mass transport is better described as hemispherical diffusion. The electrochemical literature contains mass transport treatments for many other electrode geometries,[l] but we will constrain our discussion to disk electrodes as they are by far the most common ones for experimentation.
B
A Semi-infinite diffusion
Large Electrode Figure 3.1: Diffusion at large and small disk electrodes.
HemisDherical diffusion
Small Electrode
24
3
Potential Step Methods
What determines if an electrode is large or small with respect to the diffusional differences illustrated in Fig. 3.1? Large or small are indeed relative terms and, in this case, the electrode radius (ro) must be compared to the characteristic distance traveled (diffused) by the electroactive species in the timescale of the electrochemical experiment. For common electroactive species having diffusion coefficients (D) around 10-5 cm*/s and experiments lasting several seconds, these diffusional distances are typically in the range of 10-100 pm. Therefore, large (conventional) electrodes have diameters of -1 mm or larger and small (ultramicrolectrodes) have diameters of -10 pm or less. In general terms, as the electrode becomes smaller, edge effects become more important and mass transport components parallel to the electrode surface come into play. Effectively, the solid angle occupied by the diffusion layer next to the electrode increases and more electroactive molecules reach the electrode per unit time and area than in the case of pure planar diffusion. The parameter v describes the extent of nonplanar diffusion.
where DO, stands for the diffusion coefficient of species Ox. Values of 17 larger than 6 lead to stationary currents under Cottrell conditions. For a disk electrode, the stationary current is simply given by
i = 4nFC&Doxr,
(7)
This stationary or steady state current value is reached after a time that depends on the diffusion coefficient of the electroactive species and the electrode radius. For a lpm disk electrode and an electroactive species with a D value of 10-5 cm2/s, the steady state current is reached in about 0.01 s. Ppre planar diffusion requires much smaller values of the parameter 7 and yields a time dependent current, given by the so-called Cottrell equation
i
=n
FAcbxE
where A represents the projected (geometric) area of the electrode. Notice that with conventional electrodes (planar diffusion) the current is always proportional to the electrode surface area. In contrast, with ultramicroelectrodes under diffusion control, the current is proportional to the electrode radius, which means that the current density (i/A) increases as the radius of the electrode decreases. The rest of this chapter focuses on the behavior of conventional size electrodes; ultramicroelectrodes will be treated in more detail in Chapter 5. For
3.1
The Cotfrell Experiment
25
planar electrodes of conventional size the current flowing immediately after the start of the experiment is extremely high (see eq. 8) and this gives rise to experimental problems to record current values at very early times (0-50 p). After this initial period, currents are easily measured and found to follow the Cottrell equation very closely. The continuous decay of the current observed in these experiments results from the continuous expansion of the diffusion layer deeper into the solution. The current decays with f-112, a characteristic feature of a diffusion controlled process. The corresponding concentration profile for the elecfxolyzed species Ox changes as a function of time according to the equation r
1
where the error function, commonly found in the mathematical treatment of diffusion problems, is defined as
As is shown in Fig. 3.2 the error function plot approaches unity asymptotically for z >2.
1.oo 0.80
0.60 0.40 0.20
0.00
Figure 3.2: Values of the error function for argument values in the range O< z <3.
26
3
Potential Step Methods
In terms of the concentration profile (eq. 9), this means that the concentration of Ox essentially returns to its bulk value when the following condition is met: X
2 & y 2 This equation can be utilized to estimate the thickness of the diffusion layer (d ) created with a conventional disk electrode after a certain electrolysis time t, d=4= Distance d also represents the thickness of the solution layer that is affected by the working electrode. This is a reminder that many electrochemical techniques are nondestructive as the bulk solution is unaffected by the experiment.
3.2 Chronoamperometry In chronoamperometry, as its name indicates, current is recorded as a function of time upon excitation by a potential step. The general shape of the excitation function and the current-time response are given in Fig. 3.3
EXCITATION FUNCTION
E
1
I
I
I f=O t t I
t=O
t
I
Figure 3.3: Typical potential excitation function and chronoamperometry.
current response in
If the step potential is sufficiently negative from the EO' value for the Ox/Red couple, Cottrell conditions are reached and the current response is given by eq. 8. On the other hand, for less negative potential values the current can be expressed
3.2
27
Clivonoaniperoriiet~~
(13)
where 5 is given by
and O is the ratio of concentrations at the electrode surface as determined by the Nernst equation,
Under Cottrell conditions 0=0 and eq. 13 reverts to eq. 8. Once again we must emphasize that these equations apply only to reversible (nernstian) electrochemical reactions. It is also possible to perform double-step chronamperometry in which the potential of the working electrode is switched to a second fixed value after a certain length of time z . Often, the potential is returned to the initial value, which enables the experimenter to check for chemical reactions affecting the species electrogenerated in the first potential step. A typical excitation function and current response are given in Fig. 3.4.
1
EXCITATION FUNCTION
4
i
0
RESPONSE RESPONSE
1
I bm I
t=O I
t=O
t=O
z z
t t
Figure 3.4: Potential excitation function and current response for double-step chronoamperometry .
3 Potential Step Methods
28
In spite of its importance to understand many other potential controlled experiments chronoamperometry is by itself not so commonly used in practice. However, chronocoulometry, which relies completely on chronoamperometric data, offers several important applications.
3.3 Chronocoulometry As implied by its name, in chronocoulometry we are interested in recording electrical charge as a function of time. The excitation function and the current response of the cell are identical to those described in the previous section but chronocoulometry is almost always performed under Cottrell conditions. Integration of the current given in eq. 8 as a function of time yields
Q=InFACb,,/v where all the symbols have their usual meaning. Therefore, the charge passed from the diffusion controlled reduction of solution species Ox increases with the square root of time. What gives chronocoulometry its key advantage against chronoamperometry? While the latter is based on raw data that can be directly measured in simple experiments, the former requires current integration (a very simple operation with computer-controlled instrumentation). The key difference, however, is that chronoamperometric currents decay quickly as a function of time, while chronocoulometry features a signal that increases continuously with time. Eq. 16 predicts that Q vs. t 112 plots will be linear with a zero intercept. In practice, the intercept is usually small but different from zero because some charge (Qdl) is passed at the very early stages of the experiment to charge the electrode-solution interface. Furthermore, chronocoulometry is a technique suitable for the study of species adsorbed at the electrode surface. If we assume that we have a surface excess rox of reducible electroactive species adsorbed at the electrode surface, one would anticipate that this material will be reduced instantaneously at the beginning of the experiment (diffusionplays no role in the reduction of adsorbed species). Therefore, we can write a more complete equation describing all the charge contributions that will be encountered in such an experiment
The two first terms are independent of time and their accumulated value sets the intercept of the Q vs. t 112 plot (see Fig. 3.5). Anson and coworkers were the first
3.4 Bulk Electrolysis
29
ones to propose the use of chronocoulometric plots to study species adsorbed on the electrode surface.[31 The quality of chronocoulometric data is typically very high. This fact fosters the application of chronocoulometry for the determination of diffusion coefficients (D ) of electroactive species. Normally, this is done after first determining the geometric area of the electrode in use. The determination of the electrode are can be easily done by obtaining chronocoulometric slopes with a well known electroactive species whose diffusion coefficient can be obtained from the literature. Once the electrode area has been determined with accuracy, the same electrode is utilized to record chronocoulometric data with the electroactive species of interest. The diffusion coefficient can be readily determined from the slope of the Q vs. t I/* plot. This method is one of the best ones available for the determination of diffusion coefficients.
Q
Figure 3.5: Charge components in a chronocoulometric (Q DS t 1 P ) plot.
3.4 Bulk Electrolysis In potential controlled methods the ratio of the electrode surface area to the volume of the solution (A/V) is normally low. We say that we work under small A/V conditions and the most obvious result is that the experiment is not destructive, i.e., the composition of the solution remains essentially unchanged after the experiment. However, there are instances in which we may wish to convert 100% of the electroactive species in the solution to a different oxidation state. This may be necessary to examine the spectroscopic properties or the reactivity of the electrogenerr +ed material. Full conversion of the electroactive
30
3 Potential Step Methods
species to another oxidation state can be accomplished by application of an appropriate constant potential value to the surface of the working electrode. To speed up the electrochemical conversion, an electrode with large surface area is employed while mass transport is enhanced by vigorous stirring of the solution or any other suitable means. This technique is generally referred to as bulk electrolysis under potential control. During a bulk electrolysis experiment the current decays exponentially with time as the remaining concentration of original electroactive species is gradually lowered by the electrochemical reaction. Ideally, the experimenter sets conditions and adjusts cell design in order to maximize the A/V ratio. The larger this ratio is, the more quickly the electrochemical conversion will be completed. This can be taken to the extreme of thin layer cells in which the electrolysis is over in a few seconds. On the other hand, if the solution volume that needs to be converted is larger, the electrolysis time also becomes longer. In any instance, the electrochemical conversion is taken as complete when the current level decreases to 5% or less of its initial value. Integration of the current vs. time over the length of the experiment leads to the total electrical charge passed, which can be correlated to the number of moles converted (N) using
n=- Q FVC Faraday's law. More often the measured charge (Q) can be utilized to determine the number of electrons ( n ) involved in the electrochemical reaction where F is Faraday's constant, V is the volume of electroactive solution and C is the concentration of electroactive species in the solution. A word of caution is in order for those interested in the coulometric determination of n values. It is extremely important to separate the counter electrode compartment from the working electrode compartment. While this is never necessary in experiments done with small A/V ratios, it is imperative in full conversion experiments. Assume that we are reducing Ox to Red at the working electrode in a bulk electrolysis experiment. Necessarily, an oxidation must be taking place at the counter electrode. One possibility would be that Red might be reoxidized at the counter electrode, thus recycling the electroactive material. Even if another species Red' is oxidized to yield Ox', the homogenous electron transfer between Red and Ox' would be equally detrimental as it would regenerate the starting materials Ox and Red'. Therefore, unless the two electrode compartments are separated by a glass frit of adequate porosity (see Chapter 6), the electric charge measured in the experiment will be artificially high, yielding erroneous values for n. Another complication that may arise in these experiments has to do with their longer timescales compared to chronocoulometric or voltammetric experiments. Slow reactions which do not affect the stability of electrogenerated species in faster experiments, may often lead to the decomposition of the same
3.5
References
31
species in bulk electrolysis, simply because the experimental timescale is several orders of magnitude longer. A related technique that is extremely useful in the investigation of some supramolecular systems is electrocrystallization, which relies on an extremely slow electrolysis carried out to foster the crystallization of electrogenerated species. Although rigorously speaking electrocrystallization is usually performed under conditions of controlled current, this technique is conceptually related to bulk electrolysis. Electrocrystallization allows the investigation of the solid state properties of reduced or oxidized materials. X-ray diffraction, electrical conductivity and other physical properties can thus be investigated with materials that would be almost impossible to prepare by other means. In this sense, eIectrocrystakation is a unique, valuable and perhaps underutilized method.
3.5 References 1. [a] J, Heinze, Angew. Chem. Int. Ed. Engl., 1993,32,1268-1288; [b] C. Amatore in Physical Electrochemistry,(Ed. I. Rubinstein), Dekker, New York, 1995. 2. A. J. Bard, L. R. Faulkner Electrochemical Methods: Fundamentals and Applications, Wiley, New York, 1980, Chapter 5. 3.[a] F. C. Anson, Anal. Chem., 1966,38,54; [b] J. H. Christie, R. A. Osteryoung, F. C. Anson, J. Electroanal. Chem., 1967, 13,236.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
4 Potential Sweep Methods: Voltammetry
In potential sweep methods the potential of the working electrode (measured against the reference electrode of choice) is varied continuously according to a predetermined potential waveform (also called the excitation function), while the current (or some current function) is measured as a function of the potential. In general terms, potential sweep methods or voltammetric techniques are equivalent designations for this group of methods, which have gained considerable popularity and are in widespread use today as they provide a quick and rather straightforward assessment of the redox behavior of molecular systems. Voltammetric techniques thus constitute the group of electrochemical techniques most commonly used by supramolecular chemists. This chapter is devoted to the fundamentals of voltammetric methods, from their operation to the interpretation of experimental results.
4.1 Linear Sweep Voltammetry The simplest potential sweep method is linear sweep voltammetry (LSV). In this technique the potential of the working electrode is varied linearly with time between two values, the initial and final potentials (Eiand Ef, see Fig. 4.1). Since the electrode potential is always changing throughout the experiment, a level of capacitive or charging current flows continuously. Faradaic current will also flow when the potential reaches values at which the species in solution can undergo electrochemical conversions. For instance, let us assume that we start the experiment with a solution containing a reducible compound Ox as the only electroactive species. The voltammetric scan is started at a potential at which no electrochemical reactions may take place, that is, E>E1/2. Normally, the potential will be linearly scanned in the negative direction and faradaic currents will be detected near, around, and beyond the half-wave value, that is, in the potential region where the conversion Ox + Red is favored. If the solution is kept quiescent (so that diffusion is the only mass transport mechanism possible) and the Ox/ Red couple is electrochemically reversible, the electrochemical conversion gives rise to a characteristic cathodic wave (Fig. 4.2), with a maximum current value given by the Randles-SevCik equation:
i,
= (2.69~10~)
at250 c
(4.1)
4.1
Linear Sweep Voltamrnety
33
where the peak current ip is given in PA, A is the projected electrode area (in cm2), Dox is the diffusion coefficient of the electroactive species expressed in cm2/s, Cox is its concentration (mM), and v is the scan rate in V/s. It is important to use the specified units as the equation contains a numeric factor that results from the evaluation of several constants. The Randles-Sevtik equation is one of the most important equations in voltammetry. Of course, it applies only when the current is diffusion controlled and hemispherical diffusion is unimportant (we are assuming that a planar
time
-
Figure 4.1: Excitation waveform used in linear sweep voltammetry (cathodic scan).
2 80 2 30
180
b
1 u
130 080 030
-0 20 000
-0 20
-040
-060
-0SO
POTENTIAL Figure 4.2 A typical linear sweep voltammogram.
El/:=
-0.500 V.
-1
00
34
4
Potential Sweep Methods
electrode of conventional size is used.) Note that the current depends on the square root of the scan rate. The implicit time dependence ( W 2 ) is identical to that expressed by the Cottrell equation for a potential step experiment. It is important to point out here that the potential of the voltammetric peak does not equal the half-wave potential of the corresponding redox couple. For reversible electrochemical couples, the cathodic peak occurs 20-30 mV more negative than the E 1 p value and its position is independent of the scan rate. The position of the peak represents the onset of diffusion control on the current. That is, beyond the peak potential the current does not depend on the potential anymore and is fully controlled by the rate of diffusion, which decreases gradually as the thickness of the diffusion layer increases. Therefore, it is necessary to go past the half-wave potential to reach the necessary Cottrell-like conditions. For slower (irreversible)electrochemicalcouples, a peak may or may not be reached. If the voltammogram exhibits a peak, the corresponding peak potential will shift cathodically as the scan rate increases.
4.2 Cyclic Voltammetry Cyclic voltammetry (CV) is based on the same principles as linear sweep voltammetry. However, in CV the potential of the working electrode is scanned back after reaching a certain value E,, the so-called switching potential (see Fig. 4.3). This figure shows a typical excitation waveform for CV. It is also possible to utilize excitation waveforms with more than two h e a r segments. In Fig. 4.3, the reverse scan is set to end at the initial potential, but this does not have to be the case in every CV experiment. It is not unusual to extend the reverse scan
time Figure 4.3 A typical potential excitation waveform used in CV.
*
4.2
Cyclic Voltmimetry
35
past the initial potentia1 and have a third linear segment to take it back to the initial value. Scan rates can also be varied for each linear segment of the waveform. The key advantage of CV over simple LSV results from the reverse scan. Reversing the scan after the electrochemical generation of a species is a direct and straightforward way to probe its stability. A stable electrogenerated species will remain in the vicinity of the electrode surface and yield a current wave of opposite polarity to that observed in the forward scan. An unstable species will react as it is formed and no current wave will be detected in the reverse scan. A typical cyclic voltammogram for the reversible reduction of Ox to Red is shown in Fig. 4.4.The electrochemical process is fast in the time scale of the experiment and the electrogenerated species Red is perfectly stable in the electrolytic solution. Under those conditions, and assuming that the solution is kept unstirred during the experiment, the ratio of the cathodic and anodic peak currents (the peak currents measured in the forward and reverse scans, respectively) should be equal to one. Deviations from unity reveal the presence of chemical reactions involving either redox partner (Ox or Red) or both partners. The average of the two peak potentials affords the half-wave potential for the corresponding couple, that is,
-2
-3
000
-020
-040
-0GU
-080
-1 00
POTENTIAL
Figure 4.4: Cyclic voltammetric response for a reversible redox couple. E i p = -0.500 V.
36
4
Potential Sweep Methods
The differential equations describing the diffusional movements of the electroactive species cannot be solved exactly along with the boundary conditions for LSV or CV experiments. Therefore, the current-potential curve cannot be described analytically. The voltammetric response can be calculated using numerical techniques or digital simulation techniques (see Chapter 7). The current-potential curves shown in Figs. 4.2 and 4.4 were simulated by the authors using the Electrochemical Simulation Package (ESP) written by Professor C. Nervi and freely available at his internet site.Pl The lack of analytical equations for the voltammetric current-potential responses makes it advisable to describe the observed response in detail as we discuss the parameters that can be derived from CV experiments. Fig. 4.4 shows that the flow of faradaic current does not start until a potential value of about 0.4 V is reached. If we were to reverse the potential scan at -0.4 V and return to the initial values, we would record a flat voltammogram having approximateIy constant levels of cathodic (in the forward scan) and anodic (in the reverse scan) current. As we already know, these currents are due to the capacitive charging of the working electrode's double layer. At any potential, the difference between the cathodic and anodic current (Ai) is given by: Ai=2vC
(4.3)
where v is the scan rate and C is the capacitance of the electrode at the potential of choice. This equation provides a simple method to determine the capacitance of the working electrode. However, electrode capacitance values obtained this way should only be considered estimates. In the forward scan the peak current is gwen by the Randles-Sevcik equation (eq. 4.1) as it is in LSV experiments. This equation is often used to analyze the behavior of a redox couple by plotting peak currents as a function of the square root of the scan rate. A linear plot is taken as evidence for the reversible character of the couple and demonstrates that the currents are controlled by planar diffusion to the electrode surface. The slope of such a plot can also be used to determine the diffusion coefficient of the electroactive species (Ox in our discussion) if A and CO, are known beforehand. This is not, however, a recommended method to determine diffusion coefficient values, as the peak currents are usually obtained with sizable error margins and the slope of the plot depends only on the square root of the diffusion coefficient. Chronocoulometry or voltammetric experiments with ultramicroelectrodes are much preferred for the determination of diffusion coefficient values. Another method to assess the reversibility of a redox couple is the evaluation of the potential difference between the peak potentials (AEp) of the anodic and cathodic peaks associated with the couple. Based on numerical solutions of the current-potential response in CV experiments,[2]a value of 57/n mV (at 25oC, first cycle voltammogram) is expected for a reversible redox couple. It is extremely important to realize that this value will only be obtained if the switching potential is at least 200 mV beyond the peak potential observed in the forward scan. The proximity of the switching potential to the voltammetric peaks leads to increased AEp values. Furthermore, the presence of
4.2
Cyclic Volturnmetnj
37
uncompensated cell resistance also leads to increased A€, values. If the researcher can insure that the levels of uncompensated resistance in the electrochemical cell are small and the switching potential is at least 200 mV beyond the forward scan peak potential, the observed deviations from the theoretical A€, value can be used to estimate the standard rate constant (ko) for the heterogeneous electron transfer process.[31 We should note that this method yields only estimates of ko values. As mentioned above the half-wave potential ( E l p ) can be readily obtained from the midpoint between the two peak potentials (eq 4.2) for a reversible or quasi-reversible redox couple. This value is characteristic of a redox couple and is typically within a few mV of the formal potential for the couple ( E o ' ) according to the following equation: RT El,, = E"'- -In 2nF
Do, ~
(eq. 4.4)
DRed
where the ratio of the diffusion coefficients Dox and DRed is usually very close to unity. The easy determination of half-wave potentials and estimation of formal potentials is an extremely attractive feature of CV.
4.3 Pulsed Voltammetric Techniques Although CV is a powerful and extremely useful electrochemical technique, capacitive charging currents set its detection limit to about 10-4 M under optimal conditions. This is inadequate for many analytical problems. From the standpoint of supramolecular chemistry, solubility limitations and/or material availability concerns would be eased by electrochemical techniques exhibiting higher sensitivity. The most successful way to accomplish this goal relies on the use of pulsed waveforms as potential excitation functions. These techniques take advantage of the sophisticated capabilities for potential control, current measurement, and timing in the millisecond domain that are accessible with modern microcomputers. In this chapter we will review the three most popular and potentially useful pulse voltammetric techniques: normal pulse voltammetry (NPV), differential pulse voltammetry (DPV), and square wave voltammetry (SWV).
4.3.1 Normal Pulse Voltammetry The potential excitation function used in NPV is illustrated in Fig. 4.5. It essentially consists of a series of short duration pulses of gradually increasing magnitude. After each pulse the potential returns to the initial value, a feature
38
4 Potentiai Sweep Methods
that is unique to this technique and gives rise to special applications as we will see later in this section.
-4
Figure 4.5 Potential excitation function for NPV experiments. The dots indicate current measuring points.
As indicated in the figure, the current is measured at the end of each pulse. Measuring current at the end of a period through which the potential remains constant is a common feature of many pulse voltammetric techniques. This is done to minimize capacitive charging currents, taking advantage of the exponentially fast decay of the charging current once the potential reaches a fixed value. Using pulse widths (tp) in the millisecond regime, the current measured at the end of each pulse is essentially faradaic in nature. The scan rate can be readily calculated by dividing the potential step size (AE,) by the period of the waveform (T). The current potential response in this technique takes the form of a sigmoidal, steady state voltammogram (Fig. 4.6) from which one can easily determine the half-wave potential and other voltammetric parameters. The relative rejection of capacitive currents results in concomitant gains in sensitivity. However, NPV does not afford the sensitivity levels provided by DPV or SWV (vide infra). As mentioned before, the single feature that makes NPV a useful technique for supramolecular chemists is the periodic return of the potential to the initial value. This is particularly useful in cases in which the electrogenerated species is insoluble in the electrolytic solution, as the cyclic return to the initial potential periodically regenerates the initial conditions, cleaning the electrode surface from insoluble deposits and leading to currentpotential curves that are relatively unaffected by the precipitation of the electrogenerated species. The authors' group has recently reported an example on this application of NPV.[41
4.3 Pulsed Volturnmetric Techniques
-5.0 0.00
-0.20
39
-0.40
-0.60
-0.80
-1.00
POTENTIAL
Figure 4.6: A typical current-potential curve obtained using NPV.
El,?
= -0.500 V.
4.3.2 Differential Pulse Voltammetry The potential excitation function used in DPV is illustrated in Fig. 4.7. The waveform is composed of a series of potentiaI pulses. After each pulse the potential returns to a value which is slightly more negative (in a cathodic scan, or more positive in an anodic scan) than the value preceding the pulse. This potential difference (AEs in the figure) is the net potential change that takes place after a full waveform cycle. As in NPV, the scan rate is given by the ratio between AEs and the period of the cycle, T. Two current samples are taken during every cycle of the excitation function. The current measuring points are indicated by the numbers 1 and 2 in the figure. The quantity of interest in DPV is the dzference between the currents measured at the end of the pulse (point 2) and immediately before the pulse (point l), 6i = iz - i ~ . The differential pulse voltammogram is simply a plot of 6i against the potential value at the beginning of the corresponding waveform cycle. The differential nature of the current measurement results in a peaked output, a key difference in comparison to the wave-like current-potential curves obtained in most other voltammetric techniques.
40
4
5-
Potential Sweep Methods
-
-
Figure 4.7:A typical excitation function for DPV. See text for symbol definitions.
0.5
0
0.4
I= 0.3
L
E u
o.2 0.1
0.0 0.00
-0.20
-0.40
-0.60
-0.80
-1.00
POTENTIAL Figure 4.8: A typical differential pulse voltammogram.
E l l 2 = -0.500 V
and AEp = -50 mV.
The shape of the DPV response can be quantitatively treated. The events during each waveform cycle correspond to those in a double potential step experiment. At the beginning of the cycle, the base potential E is enforced until the application of the pulse. After the pulse a new fixed potential E + AEp
4.3 Piilsed Voltammetric Techniques
41
(AE, is the pulse amplitude) is applied during the pulse width t,. shown that
It can be
(eq. 4.5)
where ( t 2 - t l ) is the time difference between the two current readings, and the parameters P and 0 are defined as follows: (eq. 4.6)
.=exp( nF ' AEp
]
2RT
The bracketed factor of eq. 4.6 describes the potential dependence of the differential current 6i. Its shape, that is, the shape of a typical differential pulse voltammogram is given in Fig. 4.8. At € >> Eo', P is very large and 6i is essentially zero. At E << Eo', P is very small and 6i approaches zero. It can be readily shown that the maximum differential current value is reached when P = 1. Therefore, using eq. 4.6 one can demonstrate that (eq. 4.8)
which combined with eq. 4.4 yields (eq. 4.9) Note that AE, values are negative in cathodic scans and positive in anodic scans. Typical AEp values range from 10 to 100 mV. The maximum or peak current is given by the expression (eq. 4.10)
Since 0 depends on the value of A€, (see eq. 4.7) the magnitude of the pulse amplitude is crucial for determination of the peak current. It can be readily shown that the ratio (l-o)/(l+o)decreases with I AE, 1 . However, it would be detrimental to increase I AEp I too much as this would tend to broaden the peak,
42
4
Potential Sweep Methods
thus jeopardizing the resolution of peaks arising at close potential values. The best compromise is reached by using 1 AEpl values in the range 25 to 50 mV. Small I AE, I values are used to resolve close peaks. Effective rejection of capacitive currents and the differential nature of faradaic current measurements are the factors responsible for the substantially increased sensitivity of DPV compared to CV. Under optimal conditions DPV allows the detection of electroactive species at concentrations as low as 10-8 M. This technique is indeed very attractive for voltammetric studies with species available in very limited supply. In addition to this, its peak-shaped output favors the resolution of voltammetric features having close EO' values.
4.3.3 Square Wave Voltammetry This technique relies on excitation functions that combine the features of a largeamplitude square wave modulation with a simple staircase waveform (see Fig. 4.9). The current is sampled at the end of each potential plateau. In each potential cycle, the current function is the differential value 6i = il - iz. which, when plotted against the average potential of each waveform cycle, affords peak-shaped voltammograms with excellent sensitivity and charging current rejection. Fig. 4.10 shows the potential dependence of the normalized current responses for a reversible redox couple.
I
*fP I
P
-t
. 4
1 1
f k
Figure 4.9 A typical excitation function for square wave voltammetry.
1
4.3 Pulsed VolfammefricTechniques
43
Generally the current-potential response of a reversible redox couple Ox/Red to a cathodic scan patterned after the potential excitation function, such as that given in Fig. 4.9, is given by the following expression[51
.
nFAD ox1/2 C
bx . + p , A E , )
(eq. 4.11)
where f, stands for the waveform's pulse width (one half of the waveform's period in this case) and F'(AEp, AE,) is the dimensionless current function which depends on the waveform's potential parameters, as defined in Fig. 4.9. This dimensionless current function is useful because for a given excitation waveform all reversible redox couples will exhibit the same normalized current response. Unfortunately, the response is hard to calculate, requiring the use of numerical techniques.
0.5
JI
0.0
-0.5
0.2
I
I
I
0.0
-0.2
-0.4
Figure 4.10 Normalized square wave voltammetric responses for a reversible redox couple. yi, y2 and A y correspond to the currents measured at points 1,points 2, and their differences, respectively.[61 (See Fig. 4.9 for definitions of these points.) Reprinted with permission ofthe American Chemical Society.
Fig. 4.10 shows the current response as measured at points 1and points 2. Normally, the preferred current function is that calculated by subtracting these two current measurements for every cycle. This current function is
44
4
Potential Sweep Methods
symmetrical around the half-wave potential and, therefore, the peak potential occurs at the Ell2 value of the redox couple. The peak height depends on the pulse amplitude much as it does in DPV. Decreasing the pulse amplitude leads to decreasing peak currents, but large pulse amplitudes give rise to broadened waves. As in DPV, the best compromise calls for AEp values around 50 mV. According to eq. 4.11 the current response depends on (fp)-1/2 or f 1/2, where f is the frequency of the excitation waveform (l/~). In fact, a h e a r plot for the peak current againstf1/2 is taken as evidence for a diffusion controlled electrochemical process. SWV allows the use of faster scan rates than DPV. This is one of the main advantages of SWV over DPV. In SWV, the scan rate is given by the productf. AEs. Typical parameters such as f = 200 MHz and AES= 2 mV give rise to a relatively fast scan rate of 600 mV/s. By contrast, in DPV the working electrode is maintained at the base potential for times approaching a full second to establish a deep diffusion layer before every pulse. This affords very slow scan rates, normally in the range of a few mV/s. SWV is therefore an inherently faster technique. It does, however, lack DPV's ability to resolve close peaks. SWV is also slightly superior to DPV in its ability to reject capacitive charging currents and offers thus an edge in terms of sensitivity. Overall, SWV appears as a highly promising and powerful electrochemical technique. In principle, its is possible to calculate the current response for any system, regardless of the presence of kinetic complications or coupled chemical reactions. As software packages capable of digital simulations of the corresponding current-potential curves become more readily available, the popularity of SWV can only increase.
4.4 References 1. This DOS-based software package can be downloaded from the following Internet address: http:/lleni.cli.unito.it/chemist~/electrochemistrv.l~tml 2. A. J. Bard and L. R. Faulkner, Electrochentical Methods: Fiiizdarnenfals and Applications, Wiley, New York, 1980, chapter 6. 3. R. S. Nicholson, Anal. Chew. 1965,37,1351-1355. 4. Y. Wang, S. Mendoza, and A. E. Kaifer, lnorg. Cliem. 1998,37,317-320. 5. J. G. Osteryoung, Acc. Clzeiii. Res. 1993, 26,77-83. 6 . J. G. Osteryoung, R. A. Osteryoung, Aiinl. Cliem. 1985,57,10lA.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
5 Ultramicroelectrodes and Their Applications
The terms macroelectrodes, microelectrodes, and ultramicroelectrodes are used to classify electrodes according to their sizes. Historically, these terms have been given different meanings depending on the uses and manufacturing technologies of the time. Currently, macroelectrodes is the term used to designate electrodes with dimensions in the centimeter range or larger while ultramicroelectrodes have dimensions in the micrometer range or smaller. Therefore, microelectrodes is the term of choice to describe any electrodes with sizes intermediates between those specified above. While this terminology is rooted on the historical development of electrochemistry, we can advance a less casual definition of ultramicroelectrodes as those that have at least a characteristic dimension similar to or smaller than the thickness of the diffusion layer created in the electrochemical experiment. In planar electrode surfaces, for instance, this size relationship gives rise to diffusion components parallel to the electrode surface which increase the overall effectiveness of mass transport to or from the solution. This is in contrast to larger electrodes, in which diffusion is always perpendicular to the electrode surface. In this chapter, we will describe in some detail the advantages of ultramicroelectrodes and the electrochemical behavior observed with them, as well as some of their important applications, such as Scanning Electrochemical Microscopy (SECM).
5.1 Characteristics of Ultramicroelectrodes The use of ultramicroelectrodes results in several significant advantages for the electrochemist. The following are the most important ones: 1.
2.
3.
4.
Currents are very small and, thus, the associated ohmic drops, given by the product IR, are correspondingly small, leading to minimized errors from residual cell resistance values. The small electrode surface leads to small electrode capacitance values. Electrode time constants, RC, are thus shortened, allowing faradaic currents to be measured, essentially free from charging currents, at shorter times. Ultramicroelectrodes give rise to steady state currents more easily than larger electrodes due to the more efficient mass transport mechanisms operating in the former. The small size of ultramicroelectrodes allows their use to probe small spatial regions, such as nerve synapses.
5 Ultramicroelectrodesand their Applications
46
There are also some difficulties associated with the use of ultramicroelectrodes. For instance, they are more difficult to fabricate and maintain than conventional size electrodes. By the same token, commercially available ultramicroelectrodes are more expensive than conventional size electrodes. In addition, they give rise to smaller currents, although these can be measured with the electrochemical instrumentation currently offered by most manufacturers. While a few years ago only specialists would engage in research with ultramicroelectrodes, nowadays they can be considered routine tools for any electrochemist. We have already mentioned that ultramicroelectrodes must have at least one characteristic dimension similar to the thickness of the diffusion layer created in the experiment. Thk point is emphasized by the range of shapes that ultramicroelectrodes can adopt (see Fig. 5.1). Notice, for instance, that a band electrode may be several meters long and still retain ultramicroelectrode behavior as long as its width is just a few micrometers or smaller. Similarly, an array may behave as an ultramicroelectrode if the radius of the individual disks is in the micrometer range regardless of the total area of the array.
Disk
Sphere
Wire
e e e Band
e e e Array
@ Ring
m
Interdigitated Array
Figure 5.1: Common ultramicroelectrode shapes.
The diffusion of electroactive species to or from ultramicroelectrodes was discussed already in Chapter 3. As mentioned before, the effective modes of mass transport resulting from the similar size of the electrode dimensions and the diffusion layer thickness lead to the very fast development of steady state (time independent) currents at these electrodes. Normally, the smaller the size of the electrode the more quickly steady state currents are reached. The detailed treatment of diffusional currents at the different ultramicroelectrode configurations shown in Fig. 5.1is beyond the scope of this book. A number of good reviews have been published on this subject."] It is perhaps simpler to compare the mathematical expressions that describe the limiting currents at some of these
5.1
Cluracteristics of Ultramicroslechodes
47
electrode configurations. In this context, limiting currents are defined as those observed under Cottrell conditions at a stationary electrode immersed in a quiescent solution. Some of the pertinent equations are listed in Table 5.1
Table 5.1 Limiting currents for some ultramicroelectrodeconfimrations. -
Sphere
(rois the radius)
i = 4nronFDC*
Hemisphere (ro is the radius)
i = 2nronFDC'
Disk
i = 4ronFDC*
(yo
is the radius)
Band (approximate) (I is the band length)
1 i =Z~FDC? ln(4(Dt)/ I021
Let us consider the voltammetric response of a fast (nernstian or reversible) redox couple (Ox/Red) on a disk ultramicroelectrode. If the voltammetric experiment is performed at slow or moderate scan rates, the resulting currentpotential curve will exhibit the typical sigmoidal wave shape observed in any type of steady state voltammetric experiment (see Fig. 5.2). The equation describing this current potential curve has the well-known form,
RT E=EX+nFlni
i,-i
where E is the applied potential, E 1 p is the half-wave potential, i stands for the current and iL is the limiting current value, as given in Table 5.1. The remaining symbols have their usual meaning. For slower redox couples, deviations from this shape can be used to determine the standard rate constant for heterogeneous electron transfer (ko) of the corresponding redox couple. As the scan rate becomes faster, the redox couple has less time to reach steady state currents at each potential value (see eq. 3.6) and, therefore, the voltammogram gradually changes its shape towards a transient current-potential response, with well defined current peaks in both the forward and the reverse scans (see Fig. 5.2). At sufficiently fast scan rates, the voltammogram becomes identical to those recorded with conventional size electrodes. The specific scan rate around which the transformation from steady state to transient behavior takes place depends on the actual electrode radius as well as the diffusion coefficient of the electroactive species.
48
5
Ultramicroelectrodes and their Applications
It is important to note here that ultramicroelectrode currents are not proportional to the electrode area, as it is the case with larger electrodes. In fact, the Limiting current on a disk ultramicroelectrode is directly proportional to the electrode radius (Table 5.1). This means that the current density actually increases as the electrode radius decreases.
im
m D
I".
-m -a
~
250,
I
Figure 5.2: Simulated (DigiSimG3 2.1) cyclic voltammetric responses of a reversible redox couple (EI/z= -0.500 V) at ultramicroelectrodes. (A) 1 pm radius, 5 mV/s, (B) lpm radius, 0.5 V/s, (C) l p m radius, 5.0 V/s. (D) 10 p m radius, 5 mV/s, (E) 10 pm radius, 0.1 V/s, (F) 10 pm radius, 1V/s.
5.2
Scanning Electrochemical Microscopy
49
The development of ultramicroelectrodes has opened a number of areas to electrochemical research. For instance, the use of ultramicroelectrodes has led to electrochemical experiments in low conductivity media. They are also fundamentally important to researchers in neurochemistry as ultramicroelectrode tips can be positioned in the small solution volume contained in a synapse and used to monitor the flow of neurotransmitters. Ultramicroelectrodes are also particularly useful to investigate electrochemical kinetics, as their response is affected by resistance to a much lower extent than that of conventional size electrodes. In the remainder of this chapter we will describe a powerful set of electrochemical methods that has been made entirely possible by ultramicroelectrodes.
5.2 Scanning Electrochemical Microscopy The group of methods known as Scanning Electrochemical (SECM) was developed by Bard and coworkers, starting in the late 1980's. The theory of SECM is usually introduced to illustrate its applications for recording microscopic images of surfaces. The fundamental principle of any SECM experiment is that the limiting current measured (under Cottrell conditions) with a disk ultramicroelectrode immersed in a solution of an electroactive species (Ox) changes in a predictable fashion with the distance between the electrode tip and the substrate to be imaged. As we already know, the current when the electrode tip is far away from the substrate is given by the simple expression ZL = 4r,nFDCt. This current will be affected if the distance between the tip and the substrate is comparable to the diffusion layer. If the substrate is an electrical insulator, its effect will be to hamper the diffusional flow of Ox to the electrode surface. Therefore, for nonconducting substrates, the current decreases from its limiting i L value as the tip approaches the substrate. In SECM, this is referred to as the negative feedback mode (see Fig. 5.3). On the other hand, if the substrate is electrically conducting, the species electrogenerated at the tip, Red, will be converted back to Ox at the surface of the substrate, and the overall current increases as the tip approaches the substrate (Fig. 5.3). This is known as the positive feedback mode. The currentdistance relationship has been quantitatively established for conducting and nonconducting substrates.Pl It is thus clear that the faradaic current measured with the ultramicroelectrode tip can be used to determine the tip-substrate distance and the conducting or nonconducting character of the substrate. Therefore, if the tip is scanned over the substrate in a rastered pattern at a constant height z, the plot of current measured against tip position (in the x-y plane) affords an image of the substrate topography. Unlike scanning tunneling microscopy (STM), imaging is possible even if the substrate is not a conductor. SECM can provide information on the conducting nature of the substrate as well as on the chemical composition of the surface, features that are not normally accessible with atomic force microscopy (AFM). However, the imaging resolution that can be attained with SECM is limited by the tip radius. Micrometer resolution is now routine in SECM, but
50
5
m
Ultramicroelectrodes and their Applications
Nonconducting substrate
No feedback
Negative feedback
( i = 4r0nFDc*J
i < 4r,nFDC*
C on d uc tin g substrate
Figure 5.3: Main modes of current generation in SECM.
substantial resolution improvements would require considerable advances in the routine fabrication of electrodes with nanometer dimensions. In this regard, SECM is currently inferior to both STM and AFM, which, in favorable cases, can provide images of surfaces with atomic resolution. The applications of SECM, however, are not limited to surface imaging. SECM has developed into a group of electrochemical techniques that can be used
5.3 E1ectroc~~nzistt-1~ of Single Molecules
51
to study electrochemical and chemical kinetics. For instance, consider the situation represented in Fig. 5 . 3 ~ .Any process affecting the rate of regeneration of Ox will have a marked effect on the overall tip current observed at a given tip-substrate distance. Therefore, one can investigate the electrochemical kinetics of the Red+Ox conversion at the surface of the conducting substrate. Furthermore, any chemical reaction that may affect the concentration of Red will also determine the overall tip current. These ideas have been developed to investigate the kinetics of chemical and electrochemical reactions using SECM measurementd41 The literature also contains a number of examples in which the tip of an SECM apparatus can be used to impress a structural pattern on the surface of the substrate.[5] For instance, an electrode poised at a potential positive enough to oxidize Br- ion to BrZ can be used to etch a pattern on the surface of a semiconductor substrate by taking advantage of the oxidizing strength of Brz. The instrumentation for SECM is rather simple. Essentially, the working ultramicroelectrode is immersed in a solution of an appropriate electroactive species along with reference and auxiliary electrodes of conventional size. The position of the ultramicroelectrode tip is set by piezoeIectric controllers and its potential against the reference electrode is determined by a potentiostat (Fig. 5.4). Often a bipotentiostat is utilized because it also affords control of the substrate potential. This is particularly useful when dealing with conducting substrates.
Bipotentiostat
I
Piezo
ri Piezo
1
Computer
\
Figure 5.4: Schematic diagram of the components of a SECM instrument.
5.3 Electrochemistry of Single Molecules The detection of electrochemical events from single molecules is among the most compelling applications of SECM. Single-molecule electrochemistry (SME) is an impressive achievement because it represents the ultimate detection limit of
52
5
Ultramicroelectrodes and their Applications
electroanalytical techniques.[6] In addition to this, in recording the electrochemical behavior of a single molecule, one may hope to uncover effects that are lost or perhaps averaged while recording the collective behavior of a large number of molecules. Again, it was the group of Professor Bard, at the University of Texas, Austin, who was responsible for this breakthrough.[q The successful detection of SME depends on the use of electrodes of submicron sizes (normally, 10-20 nm) which are sealed in wax, in such a way that their tips are recessed inside the surrounding wax sheath. Such an electrode configuration gives rise to a pocket in which a small solution volume can be entrapped when the electrode tip approaches a planar substrate (see Fig. 5.5).
I
conductive substrate
I
Figure 5.5: Trapping of an electroactive molecule in the small solution volume contained between a wax-covered, recessed ultramicroelectrode and a planar conducting substrate.
Single molecule isolation is accomplished with a SECM in which the tip electrode and the planar substrate are used to define a small solution volume that contains one molecule on the average. For instance for a 1.0 mM solution of an electroactive species, this volume will be approximately equal to 1.7 x 10-18 cm3. Actually, the number of molecules contained in this volume will be given by the Poisson distribution and, in different experiments, we can encounter data produced by zero, one, two, or, much less likely, three or more molecules. In addition to this statistical complication, the instrumentation currently available does not allow the detection of currents from single electron transfer events. Therefore, detection of the electrochemica1 behavior of a single molecule requires amplification. This is naturally achieved in the SECM, by appropriately biasing the tip and the substrate in such a way that the species electrogenerated at the tip is recycled at the substrate. This recycling process can be performed at high frequency because the tip-substrate distance is very short. For instance, if the s for D = 5 x distance is 10 nm, the transit time is given by d 2 / 2 D , or 1 x cmZ/s. If each cycle gives rise to one electron transfer event, the resulting current
5.4 Conclusions and Outlook
53
will be approximately equal to 0.8 PA. Therefore, the positive feedback mode of the SECM provides the amplification needed for SEM detection. Bard and coworkers have performed positive feedback SECM measurements with a water-soluble ferrocene derivative trapped in a small solution volume as discussed above.[q The tip current versus time data showed fluctuations that were interpreted to indicate the presence of zero, one or two ferrocene molecules in the trapped solution volume. Careful statistical analysis of the experimental data was necessary as the recorded faradaic currents were only slightly larger than the level of noise present in these experiments (about 0.2 PA).
5.4 Conclusions and Outlook In the last two decades electrochemistry has clearly been one of several fields of science in which miniaturization has played a very visible role. Electrochemists have gone from using electrodes of millimeter size to routinely fabricating and using electrodes of submicrometer dimensions. These ultramicroelectrodes have distinct technical and operational advantages. However, their most exciting feature is that they approach molecular dimensions. On the other hand, supramolecular chemists are creating molecules and molecular assemblies of progressively larger sizes. As a particularly current example, dendrimers are molecules with nanometer dimensions. In fact, we have already reached a level of technical and scientific development that allows supramolecular chemists to design and synthesize molecules of comparable sizes to those of the smallest electrodes available. Can we take advantage of this state of affairs? Can electrochemists provide supramolecular chemists with tools to study single-molecule events? Can electrodes be used to wire a single molecule to the external word, allowing electrochemical readouts of molecular states? What other applications and developments can be extrapolated from these technical advances? These and many other questions are open at this point. The authors believe that the progress of supramolecular electrochemistry relies on the answers that the near future will provide.
5.5 References 1. [a] C. Amatore in Physical Electrochemistry, I. Rubinstein, Ed., Marcel Dekker, New York, 1995, Chapter 4. [2] J. Heinze, Angew. Chem. Int. Ed. Engl. 1993, 32, 1268-1288. 2. [a] A. J. Bard, G. Denuault, C. Lee, D. Mandler and D. 0. Wipf, Acc. Chem. Res. 1990, 23, 357-363. [b] A. J. Bard, F.-R. Fan and M. V. Mirkin in Physical Electrochemistry, I. Rubinstein, Ed., Marcel Dekker, New York, 1995, Chapter 5. 3. J. Kwak and A. J. Bard, Anal. Chem. 1989, 62,1221-1227. 4. [a] M. V. Mirkin, T. C. Richards and A. J. Bard, 1. Phys. Chem. 1993,97,7672-7677. [b] M. V. Mirkin, M. Arca and A. J. Bard, I. Phys. Chem. 1993, 97, 10790-
54
5
Ultramicroelectrodes and their Applications
10795. [c] M. Arca, M. V. Mirkin and A. J. Bard, J. Phys. Chern. 1995, 99, 50405050. 5. See, for irlstance: D. Mandler and A. J. Bard, J. Electrochem. SOC.1990, 137, 10791086. 6. A. J. Bard and F.-R. Fan, Ace. Chern. Res. 1996,29,572-578. 7. F.-R. Fan and A. J. Bard, Science 1995,267,871-874.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
6 Practical Experimental Methods
An important consideration for the supramolecular chemist lies in the mechanics of the electrochemical experiments themselves. While some who are not electrochemists by training may feel happily challenged by taking up a new technique, others new to the field may feel that they do not know the "tricks of the trade" that facilitate studying their systems. The goal of this chapter is to provide a very basic introduction and suggest resources for further study.
6.1 Electrodes and Electrode Surfaces 6.1.1 Working Electrodes At the present time most working electrodes for conventional voltammetry can be divided into two classes: solid electrodes, including disk electrodes, bead electrodes, ultramicroelectrodes, and optically transparent electrodes, and nonsolid electrodes, which include mercury electrodes, carbon paste electrodes, and amalgam electrodes, among others. Nonsolid electrodes are usually employed in experiments with special applications and are not discussed here. The reader is directed to reference [l]for a more detailed discussion of these types of electrodes. Practical aspects of working with microelectrodes, employed in high resistance solvents or for high-speed voltammetry, are also beyond the scope of this basic introduction. The reader is directed to references from Chapter 5 for further mformation.
6.1.1.A Disk Electrodes
The disk minielectrode is perhaps We most widely used electrode for conventional voltammetry. Typical surfaces are gold, platinum and glassy carbon. Each of these surfaces has its limits with respect to the useful potential range and the tendency for electroactive species to adsorb onto the surface. The solvent, and its condition, ultimately determines the limits of the potential range. The primary factor that limits the background potential window is solvent decomposition or the formation of oxides on a metal electrode surface due to traces of water. Decomposition of the supporting electrolyte can also be a factor in some cases. For work in organic solvents, the glassy carbon disk electrode is probably the most commonly used.
56
6
Practical Experimental Methods
Glassy carbon has long been a preferred electrode surface for organic solvents due to its wide potential range in comparison to gold and platinum surfaces. The origin of this large window lies in its relative inertness to the presence of water. The potential range of a glassy carbon electrode under vacuum conditions can extend to almost 6 V (+3 to - 3 V) in a few very dry organic solvents. In contrast, gold and platinum have narrower potential windows, due in large part to the fact that they can undergo oxide formation in the presence of traces of water and that H+ reduction is more facile on these surfaces. Minielectrodes of glassy carbon are readily available from several manufacturers, or can be assembled from glassy carbon rods by a deft hand. Commercially available glassy carbon (GC) disk electrodes are typically sealed in a "shroud" of glass or an inert plastic such as Kel-F or PEEK. Plastic shrouds are reasonably resistant to organic solvents, however the electrode should never remain immersed in the solvent for long periods of time unless it is being used. Long periods of immersion can result in swelling of the polymer, which can degrade the seal around the region of the disk, especially if the electrode is subject to vacuum conditions. Homemade glassy carbon disks can be prepared with glass shrouds (this does require a skilled hand). Glass does not seal perfectly with the carbon rod so a sealant must be used in order to prevent leakage. Here again, the sealing material must be completely resistant to the type of solvent to be employed. While GC disks work well for a wide variety of experiments, like any other tool they are subject to wear and tear. Typical problems that may be observed include leaks, increases in resistance and adsorption of organic species. Leakage around the carbon disk may be observed as erratic increases in currents, or peculiarly shaped voltammograms (since the characteristic diffusion profile for electroactive species no longer reflects exclusively the diffusion to a planar disk surface). Often, inspection in strong light with a magnifier or dissecting microscope may reveal damage at the shroud/disk interface. Such electrodes can be discarded or the disk can be removed and recycled into a homemade electrode. The electrode's internal electrical contacts may also be subject to degradation over time, leading to increases in resistance. Resistance problems are of critical concern in accurate assessments of electrochemical reversibility. Electrodes that consistently show increases in resistance should be used with caution, even if IR compensation is applied. A final problem with these electrodes can be the appearance of small "bumps" in the background voltammograms. Such faradaic processes can be attributed to 'poisoning' of the electrode surface. Due to the high surface activity of carbon, bonds to hydrogen, hydroxyl and even carbonyl and quinones have been reported@] Activation and deactivation of the GC surface may be of interest to the reader.[zb] The behavior of the electrode under various pH conditions has been correlated with the presence of such poisoned sites on the electrode surface. Electroactive species may also adsorb strongly onto the electrode surface. Typically, both such defects can be removed by polishing with somewhat rougher grade diamond paste (see below), followed by the regular final polishing material. Electrochemical cleaning (applications of strong potentials in an appropriate solvent) may also improve the electrode's behavior. Rarely, an electroactive
6.1
Electrodes and Electrode Surfaces
57
species may become trapped in the gaps near a degraded seal where it cannot be removed by polishing. In this case if applications of moderate potentials in mildly acidic or basic solutions do not degrade the species (by EC mechanisms) the electrode may be a loss. Gold and platinum disk electrodes are also widely used and are desirable because of their superior conductivity. As with GC electrodes, these disks are typically sealed in plastic or glass shrouds. Platinum provides a long lasting surface and quite a reasonable potential window in organic solvents. In non-acidic aqueous solvents however, its cathodic potential window (with respect to SCE) is somewhat more limited than that of glassy carbon. Platinum can also have catalytic effects, which may be of import in examining some structures. Gold exhibitsk more negative potential limit than platinum under identical conditions. In the anodic range, both metals are readily oxidized in aqueous solution. Thus, in organic solvents, traces of water can also bode a limited anodic potential range. For this reason, carbon electrodes, which are less sensitive to the presence of water, are preferred. Additionally, both gold and platinum can be very sensitive to the adsorption of electroactive halides, an important consideration if the oxidative electrochemistry of species with halide counter ions is to be examined. While the oxidation of halides can be observed as a diffusion controlled process on glassy carbon surfaces, iodide adsorbs strongly on gold and platinum, resulting in stripping waves. All of the problems of electrode wear noted for glassy carbon disk electrodes can exist for gold and platinum disk electrodes as well. Disk electrodes may be polished manually or with the aid of a polishing wheel. A polishing compound such as alumina or diamond paste (synthetic or natural) is mixed with water or a polishing extender and the electrode is polished, with moderate pressure, on a polishing felt. (Some prefer to polish with a circular motion, while others advocate patterns such as a figure eight to avoid uneven wear on the surface.) The goal of polishing is to obtain a mirrorlike surface that is uniform to the eye and does not bear any scratches or other marring of the surface. Various particle sizes are available for the polishing compounds. In diamond paste, these sizes usually range from 15pm to 0.25 pm pastes. Large sizes can be used to remove scratches and other defects while the very fine pastes (0.25pm to 1 pm) are used for the final polish. Alumina is most typically used in the 0.05pm to 0.5pm range. The preference for polishing with alumina versus diamond paste varies from electrochemist to electrochemist. Immediately after polishing, the electrode should be rinsed with deionized water and then with more nonpolar solvents, as appropriate. Many electrochemists prefer to sonicate the polished electrode in deionized water instead of rinsing, in order to remove any adhering polishing media. We should also note that many electrochemists are advocates of activation procedures for “activating the surface” of glassy carbon electrodes. These procedures may be warranted in instances where the kinetics of the electron transfer reactions studied are very slow. There are ample references for these procedures in the literature.
58
6
Practical Experimental Methods
6.1.1.B Smooth Bead Electrodes
Gold and platinum bead electrodes are frequently used in the study of monolayers and supramolecular assemblies based on monolayers. A characteristic example is the use of gold bead electrodes as a substrate for the organization of thiol-based assemblies. Bead electrodes are usually made by heating a wire of high purity (99.999+%) in a gas/oxygen flame in order to melt the tip of the wire into a small bead. This bead is then flame polished to a smooth surface. The wire is usually fastened to a gold connector that permits connection to a simple copper or alloy wire in order to make the electrical contact with the potentiostat. The gold or platinum wire is then wrapped in teflon tape or otherwise insulated from the solution, permitting only the bead to contact the solution. While standard glass blowing safety glasses can be used to prepare gold bead electrodes, special precautions must be taken when preparing platinum bead electrodes because of the high temperatures necessary to melt the platinum wire. Considerable radiation is emitted, warranting additional protection of the eyes. One of the authors has used quartz glass blowing safety glasses for this purpose. Once an electrode has been prepared, reasonable care is required to prevent adsorption of volatile organics or of halides onto the electrode’s smooth surface. These electrodes should ideally be prepared in a laboratory free of organic solvent vapors. Bead electrodes should be rinsed carefully in deionized water after preparation and can be immersed in deionized water until they are ready to be used. When removed from the storing solution, any excess water can be dislodged by evaporation in a stream of argon or high quality nitrogen. In many instances, it will be useful to gauge the area of the electrode. One well established real surface area determination involves depositing an adsorbed layer of iodide and integrating the charge in order to determine the corresponding areaP1 The geometric surface area should closely approximate the real surface area on these electrodes and can be determined in the usual fashion, by employing a probe with well known diffusion coefficients, such as Ru(NH3)bCb or Fe(CN)&. After surface area determinations have been made, adequate rinsing of the electrode should remove any electrolytes and permit preparation of monolayers, etc. 6.1.1.C Other Working Electrodes
Other working electrodes include optically transparent electrodes (OTEs) which can be employed in spectroelectrochemical studies. These semiconductor electrodes are most frequently films of tin oxide doped with antimony or indium and deposited on glass. Resistance is typically higher on these electrodes and requires careful compensation. Even brief amounts of excessive IR compensation will permanently damage the electrode surface. Very thin gold, platinum and carbon films have also been employed as OTEs.
6.1 Electrodes and Electrode Surfaces
59
Working electrodes for bulk electrolysis experiments are typically platinum mesh flags. Due to the longer timescale of electrolysis experiments, the large surface area provided by the mesh increases the available surface for redox conversion. Carbon fiber bundles are used in industry for large scale electrolysis. Carbon fiber products may be useful for performing electrolyses in which any of the reduced or oxidized species are prone to adsorbing onto a platinum working electrode surface.
6.1.2 Counter Electrodes
The counter electrode is typically a platinum wire or mesh, although occasionally tungsten wire is also used. Several configurations for this electrode can be used, including a simple straight wire, coiled wire, an L-shaped wire that sweeps under the working and reference electrodes, as well as the aforementioned mesh types. This electrode plays a sacrificial role and the material must be robust in this respect. Metals that react readily with organic radicals or with halide supporting electrolytes, e.g. silver, are thus unsuitable as counter electrodes as they may become inactivated during the course of the electrochemical experiment or introduce impurities into the analyte solution. It is important to remember that the total surface area of the counter electrode must always exceed that of the working electrode. This assures that the currents required to achieve the potential at the working electrode do not become limited by those that can occur at the counter electrode. This electrode does not have to be polished, although often it is cleaned in acidic solution and flame polished to remove any adsorbed material.
6.1.3 Reference Electrodes
The most commonly employed true reference electrodes are AglAgCl, the saturated calomel (SCE) and sodium chIoride saturated calomeI (SSCE). These reference electrodes are commercially available or can be prepared in the laboratory. The basic design of any Ag/AgCl reference electrode is a simple silver wire coated with a thin layer of AgC1, immersed in a saturated solution of potassium chloride. Contact with the analyte solutions is made via a salt bridge, which can be composed of a fine frit of glass, porous vycor glass or even ceramic materials. The Ag/ AgCl reference is preferred for many experiments because the electrode itself is easily prepared and maintained, and is easily accommodated in most cells. Fig 6.1 shows three common Ag/AgCl electrodes. Although the basic design is the same, the scale of the voltammetric cell and sample volume determine the optimum size of the reference electrode. All of these designs can be prepared in the laboratory, if necessary, as the components are commercially available. For low volume cells (1 - 10 mL
60
6
Practical Experimental Methods
capacity), a small diameter (> 5mm) glass shroud can be used, while very low volume (< 1 mL) cells can utilize a flexible teflon spaghetti tubing type design. High purity silver wire is employed and the AgCI layer can be formed on the clean wire by passing a current of a few milliamps through the wire while immersed in a 1 M solution of HCl. Passing the current for a few tens of seconds produces a thin layer of AgCl on the surface of the wire. (Too thick a layer is problematic, however.) Once assembled, the electrode should be immersed in additional filling solution overnight, in order to saturate the salt bridge. Salt bridges with a low rate of filling solution leakage, such as porous vycor, are suitable for use in more polar organic solvents. Care must be taken however to prevent dehydration of vycor frits in particular, as the conductance across the membrane like frit may be altered, creating problems in accurately referencing the potential at the working electrode. Electrodes with vycor bridges should be maintained in their storing solution whenever the electrode is not in use and this solution adjusted for evaporation as necessary. Some vendors recommend that, if the vycor bridge is allowed to dry out, the entire bridge should be replaced. If lower dielectric solvents are to be used, the effects of the solvent/supporting electrolyte system should be gauged on the salt bridge by performing routine voltammetry in pure supporting electrolyte solution for time periods similar to those to be employed in the actual experiment. Increases in resistance, shifting in potentials or distortions in the normal background may indicate that the salt bridge is not compatible with the solvent system and may necessitate use of a pseudoreference, vide infra. Changes in the appearance of the vycor tip (loss of translucency, discoloration, etc. may also indicate problems due to solvent
Figure 6.1: A homemade Ag/AgCl reference electrode (left) and two commercially available designs, a small glass shroud type (center) and a flexible teflon tube type (right)M
6.1
Electrodes and Electrode Surfaces
61
incompatibility or contamination of the vycor with organic matter. If the normal appearance of the vycor cannot be restored after soaking in the electrode storing solution, the tip should be replaced. Excessive crystal formation inside the electrode shroud is a telltale sign of improper storing, resulting from loss of the internal filling solution and consequent precipitation of the excess concentration of KC1 ions. Most typically this occurs when the electrode is allowed to stand in the open air, or is left soaking in an organic solvent. Crystal formation that blocks the salt bridge can result in undesirable additional junction potentials. (Air bubbles that may become lodged in this area may also be problematic.) Electrodes can be immersed for several hours in deionized water to decrease the internal concentration, then reimmersed overnight in the original f i l h g solution. If this fails to redissolve the crystals of KC1, the electrode should be disassembled and the filling solution replaced. Care should be taken when handling glass, vycor or ceramic frits as procedures like sonication may introduce microfractures, which increase the rate of leakage to unacceptable levels. Calomel electrodes are somewhat less readily assembled and are perhaps falling out of favor due to their use of mercury. Fig. 6.2 shows a typical design. This H-type configuration is the most commonly seen design for calomel electrodes, although it can be rather awkward for some cell designs. Commercially available saturated calomel electrodes (SCE) are typically smaller
Filling solution (KC1 or NaCI)
Tungsten wire contact
-1
Figure 6.2: Schematic representation of the Saturated Calomel electrode.
62
6
Pructicd Experimentd Methods
(though still H-shaped) than those assembled in the lab. Their smaller size is advantageous for low volume (5 - 10 mL capacity) cells. Homemade calomels, while attractive because of their low cost, may be challenging for novice glassblowers to assemble. The electrical contact with the mercury interface must be provided by a wire (typically tungsten) sealed in the glass shroud.[51 Those with a moderate level of experience may find preparation of the glass body for the calomel electrode quite simple, however. In any calomel electrode, the electrical contact wire has an interface with a layer of liquid mercury. Resting above the layer of mercury is a layer of the material that gives the electrode its name, mercurous chloride, or calomel. In fact, several different types s f calomel electrodes can be prepared, the most common being the simple saturated calomel electrode (SCE), in which the filling solution of the electrode is saturated KCI. Nonsaturated KCI solutions result in different potentials versus the normal hydrogen electrode (NHE). The normal calomel electrode (NCE) contains a 1M solution of KCl. This electrode is less commonly used, because of the necessity to maintain the constant KCl concentration. Evaporation of water from the filling solution will shift the potential vs. NHE. Care must be taken in assuring that the concentration and degree of saturation of the filling solution are known and kept constant for reference electrodes. The sodium saturated calomel electrode (SSCE) is also used by some authors. The filling solution, as its name implies, contains saturated NaCl in place of KCI. In some instances, experiments may require the exclusion of chloride ions from the solvent. In this instance, both the Ag/AgCl and the SCE or SSCE reference may prove unacceptable, unless the rate of leakage from their frit-type junctions with the analyte solution is almost vanishingly small. A double bridge, i.e. a second salt bridge filled with pure supporting electrolyte solution, can isolate the reference electrode salt bridge from the analyte solution and the working electrode. Double bridges may be used in high dielectric solvents, since the additional junction potential is unlikely to pose a serious problem. Double bridges can be awkward to use however. In order to avoid double bridge requirements variants on the above electrodes can be employed. For instance, a Hg/HgS04 electrode, with a K2S04, Na~S04or HzS04 filling solution has been used by some authors in place of the SCE. The potential of this electrode is shifted by at least 0.40 volts vs. NHE from that of the SCE however. Other authors have used Ag/AgN03 type systems for some applications. Ag/AgN03 electrode kits are also commercially available for use in organic solvents. Other variations on the standard references may be found in the literature. Finally, in a number of situations, use of either SCE or Ag/AgCl type electrodes becomes impractical. Under vacuum conditions, these electrodes cannot be employed at all, while even under atmospheric conditions, highly nonpolar solvents may not be compatible with the use of salt bridges. In these instances it is not uncommon to use a silver wire as a pseudoreference during the experiment and then add a small amount of a well known redox couple, such as ferrocene, at the end of the experiment to correct all the potentials to this redox couple’s half-wave potential. This method is termed a reference redox couple or an
6.1 Electrodes and Electrode Surfaces
63
infernal reference. If the reference electroactive species interacts with the silver wire, a platinum wire can be used, or the silver reference can be insulated from the analyte solution by prefilling a glass fritted shroud with supporting electrolyte solution, as shown in Fig. 6.3.[61 A variation on this type of shroud employs a platinum wire sealed into the glass at the base of the shroud, in place of the frit. In either design, the supporting electrolyte solution is introduced prior to dissolution of the analyte. (Employing this type of pseudoreference under vacuum conditions necessitates a cell design that permits isolation of the dry analyte (see section 6.4). An alternative true reference can be prepared by mixing the supporting electrolyte with dry silica gel, and AgN03 with silica and placing the mixtures in two separate layers (supporting electrolyte, then Ag/AgN03) inside a shroud similar to that shown in Fig. 6.3, but lacking the side opening.[q The supporting electrolyte solution is allowed to slowly saturate the gel mixture, soaking up through the glass frit. The saturated gel forms a thick slurry of silver salt. This reference electrode is only compatible with relatively polar nonviscous organic solvents and is designed for use under vacuum conditions. It is attractive because a constant concentration of AgN03 will be maintained inside the shroud, even if the cell must be agitated slightly. The presence of the silver salt should permit more stable potential control over the course of long experiments. The matrix of silica gel makes it somewhat less likely that electroactive silver will be released into the analyte solution.
Removable TOP
Supporting Electrolyte Solution
Supporting Electrolyte Solution Is Introduced via Opening in Shroud
Silver Wire
Fine Glass Frit Figure 6.3: A design for a silver wire pseudoreference for use in experiments in which the analyte interacts with silver.[61
64
6
Practical Experimental Methods
In general, with respect to reference electrodes under vacuum conditions, the various problems of potential stability, particularly in the case of long experiments such as bulk electrolyses with multiple reductions, are notorious. The reader is advised to approach experiments with precious samples with great caution. As detailed below, vacuum methods should first be practiced with materials available in abundant supply.
6.2 Solvents and Supporting Electrolytes 6.2.1 Solvents for Electrochemical Experiments
The number of solvents ordinarily used by electrochemists is relatively few in comparison to the wide realm of solvents used in synthesis. The need for reasonably high conductivity represents a stringent requirement that eliminates many solvents from consideration. Beyond concerns of conductivity, there are further criteria to be met in order for a solvent to be useful in an electrochemical experiment. The solvent must prove suitably solubilizing for the analyte and must also exhibit both electrochemical and chemical inertness. Thus, the solvent must not be readily oxidized or reduced in the potential range of interest, nor should it interact with the analyte in its reduced or oxidized state (e.g. quinones should be studied in aprotic media). In short, the ideal solvent should have low resistance, a high degree of inertness, ease of handling and purification, and provide excellent solubility for all oxidation states of the analyte material. Solubility can become an important issue for systems undergoing multiple electron transfers. Oftentimes a solvent that seems ideal for the initial state of the analyte proves inadequate for solubilizing its reduced or oxidized products, which may be highly charged, generating precipitation in nonpolar solvents, or neutral, resulting in adsorption onto the electrode surface in polar solvents. In some instances, no ideal single solvent will be available. A compromise can sometimes be attempted. Binary solvent mixtures have been used by many authors to overcome solubility and resistance lirnitations.I81 Ease of purification of a solvent for electrochemistry is yet another practical concern. Organic solvents should be as dry as possible if a large potential window is desired. Parts per thousand traces of water in an organic solvent can result in sharply curtailed potential limits. Most solvents are purified with the goal of low parts per million range of water. Thus, dry to the electrochemist may be quite different from the degree of dryness sought for some synthetic routines. Much of the time involved in setting up an experiment will be devoted to preparing the solvent. In this respect, the ease of handling also becomes a consideration. Solvents with high boiling points may be difficult to distill or transfer under vacuum. They may also make recovery of the analyte difficult if limited quantities of the analyte warrant its recovery. On the other hand, solvents with low boiling points may be readily evaporated while purging with gas, making it difficult to maintain solution concentrations for longer electrochemical experiments.
6.2
Solvents and Supporting Electrolytes
65
Many organic solvents are obtained as HPLC grade, stored in their original bottles, over activated molecular sieves (preferably 3 A pore size), and with an atmosphere of argon or nitrogen. The solvent can be removed by needle and syringe or by double cannula techniques, under argon or nitrogen pressure to a dry vessel for further purification, if necessary. For a number of solvents, further purification may not be necessary if the solvent is well stored and the potential range to be examined is not too wide. If dryness is of concern, the solvent shouId be dried further, imrnediufely prior to use, by distillation from the appropriate drying agent. Distillation under vacuum provides the best results. Although impractical for some laboratory setups, vacuum distillation has its advantages. Moderate volumes (50 - 100 mL) of some purified solvents can be stored under vacuum for days at a time. In the presence of a strong drying agent such as sodium or sodium-potassium amalgam (NaK), transferring the solvent to the electrochemical cell must also be performed under vacuum. Shriver and Drezdzon provide an excelIent description of solvent purification under these conditions.Pl Table 6.1 presents a list of commonly used electrochemical solvents along with relevant physical constants and some recommended drying methods. The methods known to dry most efficiently are underlined. Methods followed by the exclamation symbol require caution. A qualitative description of the advantages and disadvantages of the more common solvents is useful. Among the polar solvents, dimethylformamide (DMF) and acetonitrile have long been popular electrochemical solvents. Both solvents show relatively low solution resistance, although IR compensation is usually still necessary. DMF has a sizable cathodic range and is an excellent solvent because of its solubilizing capability. HPLC grade DMF can be effectively dried in small quantities using P4010or a small amount of sodium. These methods can produce a cathodic range of almost 3 V vs. ferrocene, under vacuum conditions. A disadvantage of this solvent is its high boiling point, which makes vacuum transfer lengthy, and its toxicity. Due to the latter point, many electrochemists now prefer to use dimethylsulfoxide (DMSO) in place of DMF, although DMSO does not have the same solubilizing power as does DMF. Acetonitrile has both a good anodic and cathodic range, is easily dried by distillation at atmospheric pressure with calcium hydride, has a moderate boiling point and few drawbacks. Acetonitrile is a good first choice solvent for many electrochemical experiments. Among the nonpolar solvents, dichloromethane, tetrachloroethane (TCE) and tetrahydrofuran (THF) are the most frequently used. While their conductivity is still sufficiently high enough to permit routine voltammetric studies, resistance in these solvents is considerable, even with ampIe supporting electrolyte concentrations. Kinetics measurements performed in these solvents, with minielectrodes by conventional voltammetry, should be verified with microelectrode studies. Solvents with such low dielectric constants can prove problematic for studies involving titrations with alkali or alkaline earth salts. These titrants will not be soluble in such low dielectric solvents. Nonetheless, dichloromethane is commonly used for many nonpolar organic molecules. TCE has seen greater use in recent years and provides a good alternative solvent to dichloromethane, which is quite volatile. THF can be difficult to use in some conditions, has relatively narrow
66
6
Practical Experimental Methods
Table 6.1 A selection of common solvents for electrochemical purposes, along with their relevant physical constants and recommended drying methods. Solvent
Er
T,,/OC
TbP/Oc
Recommended
Water 78.30 0.0 100.0 N/A Propylene Carbonate 64.92 -54.5 241.7 1 Dimethylsulfoxide 46.45 18.5 189.0 1,z N,N-Dimethylformamide 36.71 -60.4 153.0 1,2, 3!b,4! Nitromethane 35.94 -28.55 101.2 1,5 Acetonitrile 35.94 -43.8 81.6 L2’,6 Methanol 32.66 -97.7 64.5 1 Benzonitrile 25.20 -12.75 191.1 1,2,6 Ethanol 24.55 -114.5 78.5 1 Dichloromethane 8.93 -94.9 39.60 1,2,6 Tetr achloroethaned 8.20 -43 147 L26 Tetrahydrofuran 7.58 -108.4 66.0 1,3!e Toluenef 2.38 -95.0 110.6 1,2 Benzene* 2.27 5.5 80.1 1,2 Many of these solvents will benefit from storage in the dark after purification. Drying methods: 1. Molecular Sieves. 3 A. Powdered material will increase the available surface area but may make isolation of solvent without sieve particulates difficult. 2. Calcium hydride. 3. Sodium metal. Explosion hazard in the presence of water or peroxides. 4. P4010. 5. Calcium chloride. 6. Filtration through a short column of activated alumina. .Adapted from Solvents and Solvent Effects in Organic Chemistry, by Christian Reichardt, 1988, VCH, Weinheim Germany, pp.408-410 and p.414. bDMF can be very effectively dried in small (< 15 mL) quantities using sodium and a very small amount of benzophenone as a color indicator. We caution the reader that this method should be applied only to small volumes of the solvent. CAcetonitrile should not be left stirring over CaH2 for long periods of time in the presence of water vapor. Formation of acetic acid can result. Walues drawn from Lunge’s Handbook of Chemistry, 13th Edition, McGraw Hill, New York, 1985. .Further drying on sodium/potassium amalgam yields extremely dry THF, possessing a pale blue color. NaK amalgam poses serious a explosion hazard in the presence of peroxides or water. Prepurification should always precede introduction of this reducing agent. This agent is considered to be as reactive as cesium metal. fDue to their very poor resistance, these solvents are commonly used as a component of a binary mixture with a higher dielectric solvent such as CH3CN.
6.2
Solvents and Supporting Electrolytes
67
potential windows and can readily form polymers in the presence of alkali perchlorates. For the latter reason THF may be unsuitable for use in bulk electrolysis experiments in the presence of alkali metal perchlorates. Extremely nonpolar solvents such as toluene and benzene are typically used as components of binary solvents systems. In tandem with a polar solvent such as acetonitrile, these solvents can work effectively to provide solubility for highly nonpolar materials such as the fullerenes. Two solvents, not mentioned here in detail, that have been used by electrochemists for special applications, are liquid sulfur dioxide and liquid ammonia. Further information on these specialized systems is available in the literature. Maintaining solvent purity during the course of the electrochemical experiment is another point to examine. Water can begin to saturate a dry solvent system in a cell that is not under vacuum, decreasing the potential range as the experiment is performed. Water can be introduced from a variety of sources. One of the most common problems involves water vapor introduced by purging the electrochemical with nitrogen or argon gas. Gas cylinders prepared in humid climates are notorious in this respect. Scrubbing the gas with a prefilter containing a drying agent such as drierite, NaOH, or calcium chloride can eliminate a considerable amount of water. Water can also be introduced by the need to repeatedly polish the working electrode or due to the addition of hygroscopic titrants, such as salts. In this instance, the addition of small amounts of activated alumina to the cell solution can eliminate additional water. Finally, introduction of oxygen from purging gases may also occur. Oxygen is a common contaminant in both nitrogen and argon gas. Electrochemical evidence for the presence of dissolved oxygen is observed as a diffusion controlled wave in the cathodic range, typically near -0.8 V vs. Ag/AgCl. The current intensity of the wave may actually increase the longer the cell is purged. A number of oxygen scavengers are reported in the literature.['01 Commercially available scavengers include supported copper products such as Ridox[llI or BTS catalyst[121 may be used in addition to a drying agent such as drierite. (A typical purification train alternates drierite, Ridox, and drierite columns) A Ridox column can be activated by passage of nitrogen/hydrogen (or argon) 95/5 v/v gas mixture at temperatures around 200 0C for several hours, or until the characteristic rusty brown color of the reduced copper pellets is achieved.[l21 Due to the scrubbing reaction process, a drying column should both precede and follow the oxygen scrubber column. Ridox is reported to drop the oxygen level to 1 ppm in a flowing stream of inert gas.[121
6.2.2 Supporting Electrolytes for Electrochemical Experiments
In aqueous solution a wide range of salts can be employed for supporting electrolytes. These include alkali chlorides, sulfates, nitrates, perchlorates, phosphates (for pH buffered solutions), and many others. Lithium or sodium fluorides may be of use in instances in which charge compensation is hindered by bulky structures near the electrode surface, as can arise in some modified
68
6
Practical Experimental Methods
electrodes. Since a supporting electrolyte is used in very high concentration (typically 0.10 to 0.50 M), it is important that the salts used be of the highest purity. Salts of chlorides should have negligible iodide content since this anion can adsorb onto the electrode surface. Bromide is also readily oxidized and can limit the usefuI oxidative potential range. Purity of supporting electrolytes is of special concern when working with gold or platinum bead electrodes. Electrolytes for organic solvents tend to be more limited due to the lower dielectric constant of the solvents. The most common electrolytes are tetraalkylammonium salts, especially tetrabutylammonium hexafluorophoshate (TBAPF6). Hexafluorophosphate is the most common anion, although perchlorate have also been employed. (The latter anion has fallen out of favor due to the fact that very dry perchlorates can be an explosion hazard.) These electrolytes are typically recrystallized from ethanol at least twice before use. Recently a procedure for the recycling of TBAPF6 has been reported.[l31 Triflates (trifluoromethanesulfofonates) may not be suitable for conventional cyclic voltammetry as they can increase the capacitive currents observed at the working electrode. They may, however, be successfully employed with pulsed voltammetric methods, which minimize capacitive current. A good resource of information on purification of organic and inorganic chemicals is that of Perrin and Armarego.P41 Their text contains information on purification of a wide variety of supporting electrolytes, as well as solvents.
6.3 Basic Cell Design The design of an electrochemical cell is dictated by the type of experiment desired. As the reader will note below, and in the following section, glassblowing skills bestow the freedom of designing and making cells that suit a particular experiment. To this end, glassblowing is a skill well worth attaining. Many resources are available"4 and with practice, most chemists can develop at least a modest level of skill. Nonetheless, simple cells for conventional voltammetry can be readily assembled without the aid of the torch. For voltammetric techniques the typical cell design requires three electrodes. The scale of the cell, i.e. its solution volume, is best determined by the quantity of material available and the nature of the experiment desired. Typical cells can range from moderate volume cells (a minimum of around 5 -10 mL) to microcells with less than 0.5 mL volume, or even thin layers cells in which the cell volume may be 100 pL. With the exception of the latter, the basic design is typically the same, as shown in Fig. 6.4. Each cell shown has three openings for the electrodes and one for a purging line, such as teflon tubing or a thin pipette. Homemade cells can be assembled from a small straight sided beaker or weighing jar (10 or 15 mL volume) and.a teflon cap which is trimmed to fit the beaker diameter. Cells for volumes of 1 mL or less can be assembled from conical microsample vials in which the flexible spacer pad on an open ring cap has been replaced by a wafer of teflon drilled with appropriate size openings for
6.3
Basic Cell Design
69
Figure 6.4 CommonIy seen cell setups for different solution volumes. The celIs a11 show the working electrode, counter electrode, purging line, and reference electrode, from left to right, along with a magnetic stir bar. A cell similar to the central cell can be assembled
from a small beaker and a drilled teflon cylinder cap.
the mini- or microelectrodes. In a variation on the doubIe bridge theme described in the previous section, a small volume bridge can also act as a cell, bridging the working electrode, rather than the reference electrode. The analyte/supporting electrolyte solution is placed in the cell and the working electrode is immersed in this solution. This minicell is then immersed in a larger cell's supporting electrolyte solution, near the reference and counter electrode. Communication between the reference and working electrodes is via the fritted bridge in the minicell. (Since this system does, in effect, still use a double bridge, solution resistance and additional junction potentials are still a concern.) Rotating electrode voltammetries will require a cell with a slightly larger opening for the working electrode. Very low volume cells are not compatible with the large size of the rotating electrode. A simple cell design for this method features a central tubular type of opening for the rotating disk and side openings for the counter and reference electrodes and a purging h e , as shown in Fig. 6.5. Such cells can easily be assembled by a glassblower. Thin layer cells can be fabricated using a variety of means. The working and counter electrodes are usually composed of an extremely fine, thin mesh of gold or platinum, although the working electrode surface can also be an optically transparent electrode. Small spacers separate the two sides of the cell, resulting in the very low solution volumes. Some researchers have designed
70
6
Practical Experimentd Methods
Figure 6.5:A cell design for rotating electrode voltammetry.
cells with flow capacity in order to facilitate washing away the cell’s solution after each cycle at the working electrode surface. Since thin layer cells are most often employed in spectroelectrochemical experiments, the reader is directed to reference 16 for detailed mformation. Bulk electrolysis experiments require a different cell design. Since bulk electrolyses are very long experiments, the reactions occurring at the counter electrode become of greater importance. As mentioned previously, the function of the counter electrode is to take any current necessary in order to achieve the desired potential at the working electrode versus the reference. Its presence assures that the reference remains unpolarized. However, in fulfilling this function, chemical reactions may take place in the region of the counter electrode. Indeed, even the analyte may participate in these reactions, as can the solvent and any supporting electrolyte species. In the short timespan of even a moderately long voltammetric scan, such chemical processes are of little concern. In the course of a bulk electrolysis experiment however, their occurrence becomes a prime concern. Species reacting at the counter electrode will easily diffuse throughout the cell solution over the course of the experiment and thus will possibly interact with the species created at the working electrode, and vice versa. Since bulk electrolyses are performed to determine the number of electrons transferred, or to generate reduced or oxidized species for spectroelectrochemical studies, even a small amount of contamination or loss of the solution created at the working electrode can be problematic. Rather than risk such contamination, the counter electrode is usually placed in a separate
6.3 Basic Cell Design
71
compartment, and its solution is isolated by means of a fine glass frit. A simple cell design is shown in Fig. 6.6. The two compartments are separated by a frit and the cell is sealed with simple rubber stoppers. It is important that adequate room for a magnetic stir bar be incorporated into the design, as efficient stirring is important in these experiments. Gas inlet and outlet ports permit initial purging of the cell solution, if required. This U-shaped cell has many variants, permitting a variety of experiments. In the scheme shown, there are two working electrodes: one for voltammetry and one for the bulk electrolysis. Such a design permits comparison of the voltammetric behavior of the analyte before and after the bulk electrolysis experiment. The sidearm port of the cell permits introduction of a purging line or may be used to deliver additional materials, e.g. for binding studies. Additional ports can be attached as needed. This type of cell can be used to quantify the number of electrons transferred in a redox process or to assess the stability of an electrolysed species for further work. Because the design is not completely "air-tight" however, it is not compatible with producing electrolysed solutions for spectroelectrochemical experiments. Unless the electrolysed species are extremely stable, such experiments are only possible under vacuum conditions. (The design is, however, readily adapted for such experiments by addition of a port for connection to a vacuum line, and ground glass or teflon stoppers, as described in the next section.)
CE
BE WE RE
t
Figure 6.6: A simple bulk electrolysis cell. CE is the counter electrode, BE, the bulk electrolysis working electrode, WE the normal working electrode and RE, the reference.
72
6
Pructicul Experimental Methods
Cells for electrocrystallization, that is for the growth of crystals of reduced or oxidized materials, are analogous to vacuum bulk electrolysis cells. The working and counter electrode compartments are separated by a very fine glass frit. Only working and counter electrodes are employed (typically flame polished platinum wires), e.g. no reference, since this is a controlled current process. The electrical contacts are threaded through ground glass stoppers, as described in the following section.
6.4 Vacuum Methods The final section of this chapter is devoted to vacuum methods, which comprise some of the most difficult experiments an electrochemist can conduct. There are several indicators which may point to the fact that a given analyte is better studied under vacuum conditions. In the majority of cases electrochemical processes occur at easily accessible potentials and generate reasonably stable species. Occasionally however, some species are more evanescent or intractable. The high degree of reactivity of such species may require that experiments be conducted under vacuum conditions in order to provide a carefully controlled environment."al Additionally, some experiments to be performed on radical species are simply better suited to their preparation under vacuum conditions. These might include ESR experiments or UV/VIS/NIR studies of a radical. In order to assure stability of the electrolyzed species, the entire experiment is performed under vacuum conditions. Finally, in some instances, such a small amount of material is available that any suspicion of its instability, or sensitivity to the presence of water, may warrant conducting the experiment under vacuum conditions. One of the greatest concerns about doing experiments under vacuum is their inherent difficulty. Vacuum experiments can take hours of preparation and require meticulous care to be successful. This said, if adequate preparation and careful design are employed, these experiments can often produce beaubful results that make them seem well worth their trouble. The potential window widens under these conditions, opening the possibility of observing species that could never be examined under normal atmospheric conditions in the same solvents. The key to success in electrochemical experiments under vacuum conditions is careful planning. The goal is to try to foresee any problem areas and circumvent them. By way of example, the simplest of problems under regular atmospheric conditions, adsorption onto or spoiling of the working electrode surface, may bring an unexpected end to an experiment conducted under vacuum. Breaking the vacuum in a cell in order to polish the electrode will introduce large amounts of oxygen and moisture that one has spent hours carefully removing. (For this reason, many electrochemists will choose to design cells that accommodate two working electrodes, providing a backup electrode in case of trouble.) A thorough evaluation of the planned experiment, its variables, and its potential for problems can assure at least a reasonable chance of success. The reader is therefore strongly encouraged to first practice experimental
6.4
Vacuum Methods
73
techniques for vacuum conditions on materials available in abundance. Once a successful routine is arrived at with confidence, experiments can be conducted with materials in limited supply. Cell design is the first consideration. A number of successful cell types have been published.[191 Some cell designs have been large and somewhat ungainly. Ease of handling should be an important consideration. A cell designed for this type of work should ideally permit facile movement of the supporting electrolyte solution into the reference electrode shroud, and any sidearms. For simple voltammetry the cell should permit the three (if not four) electrodes in the central cell volume. The number of openings should be kept to a bare minimum, as each additional opening or port presents a potential leak site. Teflon components have been used with success by some authors, however for high vacuum work, in the range of 10-6 torr, ground glass joints will yield more reliable results. Glassblown cells of the elegant design shown in Fig. 6.7 have proven very successful in the group of L. Echegoyen.[*O]The central cell top is the electrode attachment site. Tungsten electrical contacts can be threaded through a large diameter (30 mm) ground glass stopper15J71 and attached to the electrodes via nickel wires terminating in gold-plated brass connectors. A sidearm port branching off from the central volume provides the cell's
I Cell top
L
1 Sidearm port to vacuum line
\
Analyte reservoir
Figure 6.7 A cell for voltammetric experiments under vacuum
74
6 Practical Experimental Methods
attachment to the vacuum line, while a second sidearm serves as the solvent reservoir during distillation. Additionally, in order to obtain a voltammogram of the background supporting electrolyte solution, the analyte must be sequestered in a third sidearm. The cell body should be annealed in order to assure that the glass can take the stress of the heating and cooling required in vacuum work. Ground glass joints should be matched manually with their ground tops to assure the optimum resistance to developing leaks. The cell's central top, with tungsten electrical contacts sealed into the pyrex cannot be annealed, but can be hand-matched by grinding with corundum then fine silica. This design can be further modified to permit addition of a cell for ESR or other spectroscopic methods. In order to give the reader a feel for how a voltammetry experiment is conducted, a brief summary of the method follows. The cell, solvent vessels, and all stopcocks to be used should be extremeZy clean. Appropriate cleaning methods are available in the literature. The solvent is transferred to its flask and dried under low or high vacuum conditions, as required by the experiment. (See section above.) For high vacuum work a double manifold system, with vacuum line and inert gas line, is strongly recommended. The cell is assembled by placing the supporting electrolyte salt into the central compartment and the analyte into its sidearm. The electrodes are connected and their proper height confirmed in the assembled cell. Once all components of the setup are confirmed to be assembled correctly a small amount of perfluorinated vacuum grease is used to grease the ground glass stoppers. An inert, perfluorinated grease is recommended, since these are less subject to dissolution by most solvents. The adherence of any solubilized grease to the working electrode surface will prove problematic. Once assembled, the cell is pumped under vacuum conditions and can be heated with a heat gun to drive off any water adhering to the glass, or solvent trapped in the electrolyte. Due caution is assumed with heat sensitive analytes. The assembled and pumped cell can be checked for leaks with a Tesla probe. After sufficient pumping has taken place, the solvent is carefully transferred to the central cell volume using freeze-pump-thaw techniques.[211 The cell is closed and disconnected from the vacuum line. Stirring provides uruform dissolution of the supporting electrolyte, which is then transferred into the reference electrode shroud. After obtaining the background voltammogram, a small amount of the supporting electrolyte solution is carefully spilled onto the analyte in its sidearm receptacle. The analyte solution is washed back into the central volume. This process is repeated several times in order to assure complete dissolution of the analyte. The voltammetry may then be recorded. After obtaining the desired electrochemical results, an internal reference, such as ferrocene, must be added to the solution in order to determine the reference potentials. If a double manifold vacuum system is available this is easily accomplished by reattaching the cell to the vacuum line, pumping the stopcock, and briefly placing the cell under argon or nitrogen atmosphere. Here we assume that the standard drying and oxygen removal agents are used to purify the gas. Under argon atmosphere the cell is easily opened. (In fact, care
6.5
References
75
must be taken to assure that the gas pressure is not sufficient to displace the ground glass stoppers.) The ferrocene is instilled under a low gas flow in the sidearm compartment used to hold the analyte and as before the cell solution is washed onto the ferrocene. After dissolving the fewocene (the reader is cautioned that ferrocene easily sublimes) the cell is briefly reevacuated by rapid alternation between opening and closing it to the vacuum line. The reference voltammogram is obtained and the experiment is complete. Typical timescales for a single experiment (independent of time spent cleaning glassware, etc.) range from three to six hours. Much depends on the amount of time spent achieving the desired level of vacuum, the degree of purification required for the solvent, and the rate of solvent transfer to the cell. Slowly transferring solvents such as DMF, DMSO, or TCE can require a considerable effort on the part of the experimentalist. These solvents, because of their slow rate of transfer, are more likely to recondense into the vacuum line itself rather than in the desired reservoir of the cell and may require continuous heating of the vacuum line with a heat gun. Some solvents, such as THF, will readily dissolve nonperfluorinated grease, resulting in contamination of the cell solution and possible spoiling of the electrode surface. Bulk electrolyses can also be performed under vacuum conditions. The cell design illustrated in Fig. 6.6 can easily be modified to accommodate ground glass joints. The procedure for the solvent transfer is similar to that described above. The supporting electrolyte should be added to both compartments. In the case of a very fine frit separating the working and counter electrode compartments, an estimate of the volume of each compartment should be made to assure correct supporting electrolyte concentration in each. The analyte may be kept in a separate sidearm compartment (attached to the working compartment) until a background voltammogram has been obtained.
6.5 References 1. P. T. Kissinger, W. R. Heineman Laboratoy Techniques in Electroanalytical Chemistry, 2nd Edition, Marcel Dekker, New York: 1996. For practical experimental information this book, and its first edition, are excellent resources. 2. [a] C. M. A. Brett, A. M. Oliveira Brett, Electrochemisty, Oxford, New York, 1993, pp. 130-133. [b] Activation of glassy carbon electrodes has been studied in some detail by McCreery, Kuwana, Hu and others. Among more recent references are : M. T McDermott, C.AIIred McDermott, R. LMcCreery, Anal. Cliem. 1993, 65,937-44 and K. R. Kneten, R. L. McCreery, Anal. Chem. 1992, 64, 2518-24. 3. [a]. J. F. Rodriguez, T. Mebrahtu, M. P. Soriaga J. Electround. Chem., 1987, 233, 283-289; [b] B. G . Bravo, S. L. Michelhaugh, M. P.Soriaga, J. Phys. Chem., 1991, 95,5245-5249. 4. Electrodes of this sort are available from Bioanalytical Systems and Cypress Systems. See Appendix for information on these suppliers.
76
6
Practical Experimental Methods
5. The use of uranium doped glass (Corning Glass No. 3320) greatly enhances the success of the tungsten wire attachment. This glass bears a thermal expansion coefficient more compatible with the tungsten than pure Pyrex. The uranium glass is melted onto the tungsten wire at the desired attachment point and is then merged with the pyrex glass of the H-cell. 6. This electrode shroud, designed by E. Perez-Corder0 and Q. Xie was described in M. Gbmez-Kaifer, P. A. Reddy, C. D. Gutsche, L. Echegoyen, 1. Am. Chem. SOC.,1994,116,3580-3586. 7. A reference of this unpublished design (by E. Perez-Cordero) has been employed in the group of L. Echegoyen. The supporting electrolyte mixture should reside below the tip of the Ag wire, while the AgN03 mixture should contact the wire. 8. Binary solvent systems have been most notably successful in examinations of the fullerenes. See for instance Q. Xie, L. Echegoyen, J. A m . Chem., 1992, 114, 3978-80. 9. D. F. Shriver, M. A. Drezedzon, The Manipulation of Air-Sensitive Compounds, Wiley and Sons, New York, 1986, pp. 84- 96. 10. Ridox is available from Fisher Scientific Co. 11. BTS catalyst is available from Fluka Chemical Corp. 12. Ref. 9, pp. 74 - 80. 13. S. Diimmling, E. Eichhorn, S. Schnieder, B. Speiser and M. Wiirde, Cuwent Sep., 1997, 15(1). (Current Separations may be accessed online at www.bioanalytical.com/ cursep) 14. D. D. Perrin, W. L. F. Armarego, Purification of Laboratory Chemicals, 3rd Edition, Pergamon Press, New York, 1993. 15. [a] J. E. Hammesfahr, Clair L. Strong, Creative Glassblowing: Scientific and Ornamental, Freeman, New York, 1968, [b] E. Carberry, Glassblowing, 2nd Edition, MGLS Publishing, Marshall, Minnesota, 1994. The first of these books is out of print, but still available from the glassblowing supplier Wale Apparatus Co., 400 Front Street, Hellertown, Pennsylvania, 18055. 16. R. J. Gale, Ed. Specfroelectrochemistry: Theory and Practice, Plenum, New York, 1988. See also reference 1, pp. 280 - 283 for thin layer cell design. 17. Reference 15a provides information on this process (pp. 178 - 180). Further mformation is also available in Laboratory Glass Blowing with Corning’s Glasses, Pamplet B-72, from Corning Glass Works, Corning, New York. This pamphlet is available from Wale Apparatus. 18. An alternative to using vacuum line techniques is to perform the electrochemistry in a dry box. In many instances, this method is equally difficult. For further mformation the reader is directed to Ref. 1, pp. 569 - 581 and references therein. 19. See for example Ref. 1,pp. 557 - 565. 20. This elegant and easily handled cell was designed collaboratively by E. Perez-Cordero, Q. Xie and L. Echegoyen and has proven successful for a wide variety of vacuum experiments. As mentioned in the text, it is readily adapted for ESR and other spectroscopic experiments. 21. Ref. 9, p. 104.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
7. Digital Simulations
As discussed in Chapter 4 researchers must often deal with voltammetric responses (current-potential curves) that cannot be represented analytically. In order to determine the parameters that describe the thermodynamics and/or kinetics of the electron transfer processes surveyed, as well as those of any homogeneous chemical reactions coupled to them, it is necessary to utilize numerical approximation methods. The literature also contains numerous procedures that, making use of dimensionless parameters and working curves, allow the extraction of thermodynamic and/or kinetic data for certain commonly occurring mechanisms. Nowadays, the omnipresence of personal computers greatly facilitates their use to perform digital simulations of electrochemical experiments. This approach was pioneered by Feldberg[l] and, in principle, can be applied to any mechanism, regardless of its complexity. The central idea is to optimize the fitting of simulated and experimental data in order to determine the thermodynamic and kinetic parameters of any electrochemical or chemical reactions involved in the mechanism in question. This chapter will introduce the reader to the fundamental principles of digital simulations of electrochemical experiments. We will also discuss two software packages and some of the results that they produce when applied to some simple mechanisms.
7.1 Principles of Digital Simulation Nowadays digital simulations are often utilized to represent and understand a variety of complex physical phenomena. The general idea is to set up a mathematical model of the phenomenon in question in a digital computer and let it evolve, according to well-defined rules or equations, to extract conclusions and make predictions on its behavior. In simple terms, a digital simulation is a computerized, mathematical representation of a fraction of our physical reality. The simulation of electrochemical experiments requires the discretization of the electrolytic solution, which is divided in a series of small volume elements[2](see Fig. 7.1.). The first element ( i = 1)is in contact with the electrode surface and the nth represents the bulk solution. In principle, the quality of the simulation will improve as the number of volume elements increases. However, since the electrochemical conversions take place in the first volume element, the elements close to the electrode surface are more important than those removed from the surface. In a fixed grid, such as that of Fig. 7.1, all elements or cells have the same width Ax. Exponentially expanding grids consist of volume elements whose width increases exponentially with their distance to
7
78
Digital Simulations
the electrode surface. These have become very popular as they combine accuracy with computational speed since little computational effort is invested in the elements close to the bulk solution. Every volume element i is characterized by a concentration Ci for each species of interest.
Electrode
Figure 7.1: Space discretized model for the digital simulation of electrochemical experiments on an electrode with a projected area A.
The time evolution of the system is determined by equations that represent the physical phenomena that take place at the electrode-solution interface and in the solution. For instance, for a reversible system, the Nernst equation will determine the concentrations of electroactive species at the electrode surface (i = 1).Diffusional flow of species between volume elements is represented by approximate equations derived from Fick's laws. The simulation starts from a perfectly homogeneous solution in which the concentration of each species is constant across all volume elements. Perturbations are introduced, normally through potential changes at the electrode surface, and the concentrations in the first volume element change (yielding faradaic current) to adapt to the new potential values. These concentration changes start diffusional flows that are calculated and propagated through the grid in successive iterations. The variable time is introduced as the interval (At) between two consecutive iterations. Thermodynamic and kinetic equations to describe equilibrium or non-equilibrium chemical processes can also be applied to the concentrations of all species in every volume element. In simulations of electrochemical experiments, two fundamental types of finite difference calculations can be used. In explicit finite difference (EFD) methods the properties of the system at t+At are calculated from the properties
7.2
CV Simulations of the CV Bellavior of a Sirtiple Redax Couple
79
of the system in the previous iteration (at t). By contrast, implicit finite difference (IFD) methods rely on the properties of the system at t, t-At, and even values at t+At to determine the remaining system properties at f+At. Obviously, IFD methods are more powerful and suffer from less "propagation" problems than EFD methods. On the other hand, IFD methods are more difficult to program than EFD methods. Since the advent of modern voltammetric techniques in the 1960's, electrochemists that wished to analyze their experimental data with digital simulations had to write their own computer code adapted to the mechanisms that described their electrochemical problems. For many years the idea of a universal simulator, that is, a software package that could simulate electrochemical experiments for any combination of electron transfer steps and coupled homogeneous chemical reactions, was just wishful thinking. In this regard, the introduction a few years ago of DigiSim8[31 was an important breakthrough that made digital simulations accessible to nonspecialists. DigiSimB is a versatile and user friendly software package that allows the simulation of cyclic voltammetric data for any mechanism entered by the user. This package can also simulate experiments in which hemispherical diffusion prevails, that is, it can simulate ultramicroelectrode behavior. The program code uses a fast IFD algorithm and an exponentially expanding grid to optimize the robustness and computational efficiency of the simulations. Another interesting software package is the Electrochemical Simulation Package (ESP) that is freely available at Professor C. Nervi's internet site.@]This package can simulate cyclic voltammetry, square wave voltammetry, chronoamperometry, and normal pulse voltammetry. It is DOS-based and less user friendly than DigiSimB, which runs in the Windows environment. Nonetheless, it is quite powerful and allows simulations that are not available with DigiSimB. The availability of these software packages has made the tool of digital simulations accessible to any researchers regardless of their degree of sophistication with computer programming. In particular, supramolecular chemists can take advantage of these simulation programs to improve the interpretation of their electrochemical data. In supramolecular chemistry it is often found that electron transfer reactions are coupled to chemical processes, and digital simulations constitute a powerful tool to obtain the kinetic and thermodynamic parameters of these electrochemical and chemical reactions. This chapter will provide several examples of the applications of digital simulations to electrochemical and supramolecular problems.
7.2 Simulations of the CV behavior of a Simple Redox Couple Let us consider a reversible redox reaction such as A+ + e = A, with a formal potential ED = 0.4 V. The solution initially contains 1.0 mM A and no A+. The simulation of this trivial voltammogram is easily set in DigiSimB. A nice feature of this software package is that it provides concentration profiles for the partners of the electrochemical reaction at any point in the voltammogram. Fig.
80
7
Digital Simulations
7.2A shows the concentration profiles for both A and A+ at a potential (0.338 V) smaller than the formal value. Notice that the initial concentrations are essentially unchanged and they only depart slightly form the initial values in the vicinity of the electrode surface. Fig. 7.2B shows the concentration profiles at 0.412 V, at which point the changes near the electrode surface become larger and reach deeper into the solution. At this potential, the surface concentration of A+ is already larger than that of A, as the potential is slightly larger than the formal value. As the potential scan continues, these trends become more obvious, that is, the surface concentration of A+ becomes larger until it reaches the maximum possible value of 1.0 mM and the surface concentration of A is depleted to a zero value while the thickness of the diffusion layer increases.
Figure 7.2 Concentration profiles for A and A+ at (A)0.338 V and (B)0.412 v.
7.2
Simulations oftlie CV behavior of a Simple Redox Couple
81
When the scan is reversed, eventually we will reach potential values at which the surface concentrations will start to return to the initial values. Thus, at 0.376 V in the reverse scan, Fig. 7.3A shows that the surface concentration of A+ is again lower than that of A. Of course the diffusion layers created in the forward scan do not follow this trend immediately, and farther away from the eIectrode surface there are areas in which [A'] > [A]. Slowly, as the reverse scan follows its course (see Fig. 7.3B), the initial conditions are slowly reestablished.
Figure 7.3 Concentration profiles for A and A+ at (A) 0.376 V and (B) 0.000 V
In cyclic voltammetry the current levels observed for a redox couple clearly depend on the square root of the diffusion coefficient (DO) for the electroactive species initially present in the solution (see eq. 4.1). This dependence is illustrated by the simulated voltammograms shown in Fig. 7.4 corresponding to three different DO values.
7
82
Digital Simulations
Figure 7.4: Simulated (DigiSimB) CV responses for a reversible redox couple (Eo' = -0.275 V) with D,values of lxlO-5,5x10-6and 1x10-6cm2/s. Scan rate: 0.1 V/s.
Cyclic voltammetry is also sensitive to the electron transfer kinetics of the redox couple under investigation. As the electrochemical kinetics become slower (decreasing ko values) the anodic and cathodic voltammetric peaks broaden and shift away from each other (see Fig. 7.5)
0.2
0.1
2
3
0.0
E;
5?! -0.1 -0.2
-0.3 0.7
0.6
0.4 0.3 0.2 Potential in V vs. Ag/AgCl
0.5
0.1
Figure 7.5 Simulated (DigiSimB) CV responses for a redox couple (Eo' values of l.O,lO-3,10-' and 10-5cm/s. Scan rate 0.1 V/s.
0.0
=
0.400 V) with ko
7.2
Simulations of the CV bellavzor ofa Simple Redox Couple
83
Also the charge transfer coefficient c1 may affect the CV responses of kinetically slow redox couples. Fig. 7.6 illustrates this effect for a quasi-reversible redox couple (ko = 0.005 cm/s). Notice that when a=0.75, the current levels of the cathodic peak increase at the expense of the anodic peak current levels, while the reverse is true when ct=0.25. The magnitude of these effects would be even larger for kinetically slower (smaller ko values) redox couples.
Potential in V vs. Ag/AgC1
Figure 7.6: Simulated (DigiSimB) CV responses for a redox couple
(Eo' =
0.400 V, ko
=
0.005 cm/s) with a values of 0.75, 0.50 and 0.25. Scan rate: 0.1 V/s.
Square wave voltammetry (SWV) is also very sensitive to slow electrochemical kinetics. As the ko values decrease the SWV peak broadens and shifts to more negative potentials (in a cathodic scan, more positive in an anodic scan). Fig. 7.7 illustrates this behavior with a set of ESP simulations.
-
1 4 . 00 2 2 .. 0 0
0.0 8.0 6.0
4.0 2.0
1-
0.0 0.00
A
'
I -0.20
-0 .40
-0.60
-0 . 8 0
-1.00
Potential, V
Figure 7.7: Simulated (ESP) responses for a redox couple (EO' = -0.500 V) with ko values of 0.1, 0.01, 0.001 and 0.0001 cm/s. Scan rate: 0.18V/s (60 Hz x 3 mV step size).
7
84
Digital Simulations
The authors' group has used digital simulations to analyze the electrochemical behavior of ferrocene encapsulated inside Cram's hemicarcerands. These hosts have four equatorial portals through which a guest can reach their inner cavities. If the sizes of the guest and the portals are similar, the complexation reaction occurs very slowly. Heating the host in the presence of an excess of guest yields the inclusion complex (hemicarceplex), which becomes kinetically stable upon coohg. Cram termed this type of host-guest interactions .constrictive binding.[51 Experimentally we found that ferrocene hemicarceplexes exhibit quasi-reversible electrochemical behavior,[61 in contrast to the fast electrochemical kinetics observed with uncomplexed ferrocene. The cyclic voltammograms obtained were simulated with the DigiSimB package to determine the pertinent ko values. Our results revealed that hemicarcerand encapsulation lowers the ferrocene ko value by a factor of 10-50, depending on the host structure, due to the larger distance of maximum approach of ferrocene to the electrode surface in the hemicarceplex.
7.3 Simulation of Electron Transfer Reactions Coupled to Homogeneous Chemical Processes Either the oxidized or reduced species in electron transfer processes are often involved in homogeneous chemical reactions. Usually, the chemical reaction can be detected from the observed electrochemical behavior, which can then be used to gain information on the thermodynamic and/or kinetic parameters of the chemical reaction. A situation commonly found in electrochemical experiments involves the decomposition of the electrogenerated species. For instance, let us consider the following mechanism: O + e = R
R
(1)
+ Products
This is a simple example of the so-called EC mechanism, where EC means an heterogeneous electron transfer step followed by a chemical reaction. By the same token, we can speak of CE mechanisms (electron transfer preceded by a chemical reaction), EE mechanisms (electron transfer followed by a second electron transfer step on the products of the first one), ECE mechanisms (electron transfer followed by a chemical reaction, which in turns produces new electroactive species), etc.. These abbreviated designations can be made more specific by using subscripts to indicate the reversible (r) or irreversible (i) character of each step. Coming back to the E,C, mechanism represented by eqs. 1 and 2, we must ask the question: How does the "following"chemical reaction affect the electrochemical behavior? Clearly one would expect reversal techniques, such as CV, to be strongly affected as they sample the fate of the electrogenerated species (R in our mechanism). However, the extent of the perturbations depends on the relative time scales of the experimental technique
7.3
simulation of Electron Transfer Reactions Coupled to Homogeneous Chemical Processes
85
and the chemical reaction. This is clearly illustrated by the range of CV responses shown in Fig. 7.8. If the time scale of the experimental technique is much faster than the lifetime of the electrogenerated species no effect is observed on the voltammetric response. This is the case if we use a fast scan rate (5.0 V/s). At slower scan rates the reverse peak shows decreased current levels until it disappears completely at 0.1 V/s.
15.0
10.0
$
-5.0
-10.0
-15.0
-0.2
4.3
-0.4
-0.5
4.6
4.7
4.8
POTENTIAL,V
Figure 7 . 8 Simulated (ESP) CV responses for the EC mechanism of the previous page. EO' = -0.500 V for the redox process. k = 1.0 s-1 for the "following"chemical process. Scan rates: 0.1,0.5,1.0 and 5.0 V/s.
In supramolecular chemistry researchers often find situations in which one of the partners of a redox couple interacts with a host forming a stable noncovalent complex, while the other redox partner is not significantly complexed. For instance, the cobaltocenium/cobaltocene redox couple affords an interesting example regarding its interactions with the well-known host pcyclodextrin @-CD).171 The positively charged cobaltocenium is not bound by pCD, but the uncharged and hydrophobic reduced form, cobaltocene, is an excellent guest for this host. The cathodic voltammetric behavior of
86
7
Digital Simulations
cobaltocenium is complicated by the insolubility of its reduced form in aqueous media. However, similar p-CD binding properties are found in more watersoluble cobaltocenium derivatives, such as carboxycobaltocenium (Cob+-COOH) which is acidic enough to be present as its anion at pH>3. The cathodic voltammetric behavior for Cob+-COO- exhibits a reversible one-electron reduction wave centered at -0.99 V vs Ag/AgCl. In the presence of (3-CD this wave shifts anodically, while the AEp value tends to increase from its theoretical 57 mV for a reversible process. These findings are consistent with the complexation of the electrogenerated cobaltocene by the p-CD host. Using electrochemical and 1H-NMR spectroscopic data, we have proposed the following E,C, mechanism for the reduction of Cob+-COO-in the presence of 0CD.[q
coo+ e
&
coo-
+
Figure 7.9 Proposed mechanism for the reduction of carboxycobaltocenium in the presence of (3-CD.
Digital simulations were utilized to validate this mechanism and to determine the equilibrium ( K ) and kinetic rate (kf)constants for the association between p-CD and cobaltocene. Good fits between the simulated and experimental voltammograms were obtained,[71yielding optimum values of K = 1,800 M-l and kf = 3.6 x lo7 M-Is-'. A good test of the accuracy of these parameters is that they yield simulated voltammograms that fit the experimental ones very well through the entire range of 0-CD concentrations surveyed (see for instance Fig. 7.10 in the next page). Notice that the proposed mechanism does not take into account the direct oxidation of the inclusion complex. As the
7.4
References
87
dissociation of the guest from the complex is fast, the electron transfer reaction takes place on the free guest. Similar results have been reported with other cyclodextrin inclusion complexes of electroactive guests.[8-101 The lack of electrochemical activity of these complexes is consistent with the thermodynamic and kinetic hindrances that we have observed on the electron transfer reactions of fully encapsulated redox centers.
t -2.01
-0.4
'
-0.6
'
'
-0.8
'
-1.0
'
'
-1.2
'
-1.4
Potential, V vs Ag/AgCI
Figure 7.10 Experimental (continuous Iine) and simuIated (circles) voltammograms for the reduction of 1.0 mM Cob+-COO-+ 10 mM p-CD in 0.1 M phosphate buffer (pH = 7 ) . Scan rate: 0.1 V/s.
In principle, similar methods can be used to analyze electrochemical data obtained with any supramolecular system in which electron transfer reactions are coupled to chemical processes. In practice, however, when the number of parameters that must be fitted increases so does the uncertainty associated with their estimation. In those cases the task is greatly facilitated if one can determine some of the pertinent equilibrium and kinetic rate constants using independent methodology.
7.4 References 1.S. W. Feldberg and C. Auerbach, Anal. Chem. 1964,36,505-509. 2. For reviews, see: (a) M. Rudolph in Physical Electrochemistry, I. Rubinstein, Ed., Marcel Dekker, New York, 1995, Chapter 3. (b) S. Feldberg in Electroanalytical Chemistry, Vol. 3, A. J. Bard, Ed.; Marcel Dekker: New York, 1969, Chapter 4. 3. M. Rudolh, D. P. Reddy and S. Feldberg, Anal. Chem. 1994,66,589A-600A.
88
7
Digital Simulations
4. Copyright by Professor Carlo Nervi. This package can be downloaded at the Internet address: http://lem.ch.unito.itlchemistrv/electrochemistry.html 5. D. J. Cram and J.M Cram, Monographs in Supramolecular Chemistry, Vol. 4: Molecular Containers and Their Guests, J. F. Stoddart, Ed.; Royal Society of Chemistry, Cambridge, 1994. 6. S. Mendoza, P. D. Davidov and A. E. Kaifer, Chem. Eur. J. 1998,4,864-870. 7. Y. Wang, S. Mendoza and A. E. Kaifer, lnorg. Chem. 1998,37,317-320. 8. T. Matsue, D. H. Evans, T. Osa and N. Kobayashi, J. Am. Chem. SOC.1985,107, 3411-3417. 9. R. Isnin, C. Salam and A. E. Kaifer, J. Org. Chem. 1991,56,35-41. 10. A. Mirzoian and A. E. Kaifer, Chem. Eur. J. 1997,3,1052-1058.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
8 Electrochemical Considerations for Supramolecular Systems
In this chapter we shall review several basic concepts relevant to the study of electrochemistry of supramolecular systems. The first of these is a consideration of intermolecular interactions and their relevance to electrochemical processes. The role played by these forces in host-guest chemistry and self-assembly will be discussed briefly. Considerations of molecular design and its effect on electrochemical reversibility will also be addressed. Finally, we shall consider the electrochemical behavior of systems with multiple redox active sites.
8.1 Intermolecular Forces under Electrochemical Conditions The intermolecular interactions that comprise the basis of host-guest interactions or self-assembling systems have been discussed extensively in the literature. In supramolecular systems, as in biological systems, these interactions are largely electrostatic in character, encompassing ion-ion, ion-dipole and dipole-dipole, hydrogen bonding, n--71 and cation-7t interactions. In self-assembling systems, particularly those relying on monolayer formation, Van der Waals, hydrophobic or solvophobic interactions may also be of significance. While much has been written about the nature of these interactions, the complicated interplay between these forces and their nature under electrochemical conditions has been less explored. The requirements of simple electrochemical experiments, e.g., solvents with reasonably high dielectric constants and a large excess of supporting electrolyte, in and of themselves, can significantly affect the magnitude of electrostatic interactions, particularly in the case of host-guest complexes. Thus, a cogent topic of discussion is the consideration of these fundamental interactions and how electrochemical experimental conditions may mfluence or alter them. This can further lead us to several salient points for reflection on the design of experiments and their comparison to mformation obtained by spectroscopic methods. Electrostatic forces play the most sigruficant role in supramolecular electrochemistry, since a change in the oxidation state of an electroactive species may result in changes in the interaction energy (AE) between that species and its prospective hosts or guests. The dielectric constant of the solvent medium is of obvious import, as demonstrated by consideration of the well known equation for the Coulombic interaction energy:
90
8 Electrochemical Considerations for Supramolecular Systems
AE =
Z,Z,E2
Dr12 where r12 is the distance between the two charges, ZI and 2 2 are their unit charges (a positive or negative integer), E is the unit of electronic charge and D is the dielectric constant of the solvent. In most instances, the solvent chosen for experimental work will be a compromise between one that provides the best solubility for the electroactive species (in its initial and switched oxidation states, reflecting oxidation or reduction), and one providing the highest possible dielectric constant, in order to reduce the resistance of the medium. Given the choice between employing CHKL (D = 8.93)"l or DMSO (D = 46.5)[11 as a solvent, most electrochemists would certainly prefer the latterP1 (Water, with its high dielectric constant of 78.30,['1 would be even more desirable.) From the above equation we can see that the magnitude of the coulombic interaction is reduced when the solvent dielectric is large. Furthermore, a large amount of supporting electrolyte is required to reduce the overall solution resistance and maintain diffusion controlled, rather than migratory, conditions. The high supporting electrolyte concentration may further affect the dielectric of the medium, while the electrolyte's ions may interact extensively with any charged or polar host/guest species. Clearly therefore, under the foregoing conditions, some of the electrostatic forces between an electroactive species and its prospective host or guest may be minimized, or of diminished significance. Nonetheless cooperative interactions, i.e. the summation of several low energy electrostatic contacts, or nonelectrostatic forces, can still dominate the binding process. Indeed, this point is borne out by the-innumerable examples of successful redox-switchable binding studies in the literature. In this respect we can consider two points: a molecular design of the host and guest species that maximizes these interactions, and the selection of a supporting electrolyte that is unlikely to compete with them. In the first case we can see that molecular design is of paramount importance: a thoughtful approach to optimizing the strength of the intermolecular interactions between the host-guest pair. While it is obvious that this is crucial to any host-guest system, it is even more so for a system which is to be electrochemically switched. In at least one of its redox states, the binding of a guest to a host should permit the maximum possible favorable interaction, e.g., in hydrogen-bonded systems the target atoms should have contacts at the appropriate distance and angle. The supporting electrolyte can also be chosen to lessen the degree of interaction with a host-guest complex, e.g., ions too large or too small to be tightly bound in a charged host may be employed. Both the size and the relative hardness or softness of the electrolyte species can be considered. In organic solvents with relatively low dielectric constants (e.g. CHClz, THF, and toluene, with dielectrics of 8.93, 7.20 and 2.38, respectively, at 25 C"]) electrostatic forces become more significant. However, in these solvents incomplete dissociation of supporting electrolyte ion pairs may be observed and
8.2
Intermolecular Forces under Electrochemical Conditions
91
solution resistance is markedly higher. While these solvents may favor an increase in the role of electrostatic forces in a host-guest complex, the solubility of a charged reduced or oxidized electroactive species can be quite low in such media and may result in precipitation at the electrode surface, making electrochemical study difficult or impossible.PI High supporting electrolyte concentrations are a requirement in such solvents and a charged analyte species may also exhibit a stronger interaction with the electrolyte in these solvents. We should note that the advent of ultramicroelectrodes, which permit electrochemical study in higher resistance solvents, in some instances in the absence of supporting electrolyte, shows promise for analyzing the electrochemical systems under conditions closer to those employed in traditional spectroscopic studies. Ultramicroelectrodes are discussed in Chapter 5. Although they may be employed with some success in higher resistance solvents, their behavior under such conditions is often still far from ideal. Considering hydrogen bonding as a special class of electrostatic interaction, it is obvious that the choice of solvent is also important for systems heavily dependent upon molecular recognition via this class of interaction. The higher dielectric solvents that are desirable for electrochemical work (e.g. water, alcohols, or the aprotic DMF and DMSO) are also more likely to be capable of hydrogen bonding with the guests or hosts, thus competing as hydrogen bond acceptors and/or donors. Host-guest complexes that are designed to be stabilized by hydrogen bonding are obviously most appropriately studied in nonpolar solvents such as CH2C12 or, less ideally, DMF or DMSO. Supporting electrolytes containing non-hydrogen bond-accepting anions are also a necessity. Van der Waals, or dispersion forces, which are present between any two atoms, are not influenced by the special solvent media requirements of electrochemical experiments. However, the complex processes leading to hydrophobic interaction~,[~] or the so-called solvophobic interactions, can, of course, be profoundly affected by the solvent medium. These forces have been used to advantage in the self-assembly of a number of electroactive supramolecular assemblies at interfaces. The extent of the effect of hydrophobic or solvophobic interactions on complexation and molecular recognition has been a topic of recent interest. Diederich and coworkers have studied apolar complexation and have found that apolar arene binding occurs in solvents of all polaritie~,[~] although the stabilities of such complexes formed in organic solvents, rather than water, is greatly reduced. Aromatic systems have been noted to interact via n--71 arrangements, a process that should not be perturbed by the presence of most solvents used for electrochemistry. An exception is the fullerene-based systems, which are most frequently studied in aromatic solvents such as toluene, benzene and benzonitrile, or in binary systems containing these solvents. Such solvent systems might be anticipated to interfere with the n-n interactions that would contribute to the binding of fullerenes by aromatic hosts such as cyclophanes. Cation-n interactions have been a topic of interest in recent years. This type of interaction has been theorized to be largely electrostatic, attributable
92
8 Electrochemical Considerations for Supramolecular Systems
primarily to ion-quadrupole effects, but also relying on additional contributions from polarizablities, and dispersion forces of the components.[51 In experimental terms, these interactions have largely been studied in aqueous solution[5,61and so little is known about the effect of solvent polarity and dielectric constant or of other ions (due to a supporting electrolyte) on these interactions. Since evidence suggests that aromatic molecules such as benzene can compete effectively with water for solvation of larger cations (K+, Rb'), one can speculate that moderately nonpolar, low dielectric solvents (excluding, of course, x-donor aromatic solvents like benzene and toluene) may foster the electrostatic component of these interactions, strengthening their character. Thus, such solvents may enhance the binding of cations in aromatic systems such as cyclophanes. We should note that even organic cations, such as tetramethylammonium ions and acetylcholine, have been shown to exhibit cation-x interactions. Alkylammonium ions are, of course, among the most commonly used supporting electrolyte cations for organic solvent systems. The surprisingly high estimated strength of such interaction~[~] (-9 kcal/mol per cation-n-face interaction for tetramethylammonium ion in aqueous solution) implies that the sum interaction energy might be significant enough to exert an effect on the electrochemistry of, for example, redox active cyclophanes, which can often undergo reduction to more electron rich charge states. Should this point be considered when performing electrochemical binding studies in organic solvent systems using alkylammonium ions as supporting electrolyte? Clearly the trend suggested by these authors indicates less interaction should be anticipated between bulkier cations such as tetrabutylammonium ion and aromatic species. In this instance, the magnitude of the effect should be anticipated to be quite small. However the magnitude of the interaction that is reported for the tetramethylammonium ion makes this a point of interest when the smaller alkylammonium ions are employed. In this instance we might wish to consider the possibility that the redox potentials for an aromatic electroactive receptor or guest may already display a shiftfrom its ''true" redox potential, due to interactions with the supporting electrolyte. Thus the strength of the electroactive species' binding with a positively charged target species may be underestimated. This is a question that might bear further study. Ascertaining the magnitude of any such an effect is, however, a challenging undertaking. From the foregoing points, it is clear that in order to obtain consistent and meaningful assessments of binding in redox-switchable ligands we should study the binding of initial states of any host/guest pair under conditions that correspond as closely as possible to those employed for the electrochemical switching studies. By this we mean that binding studies of the initial states of the system should preferably be performed in the same solvent and supporting electrolyte system to be employed in electrochemical experiments. In some instances this may not be possible, e.g. in some spectroscopic studies the presence of a given supporting electrolyte may interfere with the desired spectral window. However, in such cases efforts can be made to try to mimic electrochemical conditions with alternative electrolytes, if possible.
8.2
SelfAssenibly and Fixed Associntion
111
Siipranioleciilar Structures
93
8.2 Self-Assembly and Fixed Association in Supramolecular Structures: Implications for Reversible Redox-Switching Over the past two decades the goal of self-assembly of electroactive supramolecular systems has been achieved exploiting combinations of the forces detailed above. Self-assembly has no doubt been a popular route to such structures because it affords a more facile means of preparation. One avenue of development has been interfacial assemblies, e.g., the preparation of selfassembled monolayers or SAMs. Amphiphilic aggregation and thiol/ disulfide or silane attachment of amphiphiles directly onto electrode surfaces is still a rapidly expanding research area. These assemblies rely primarily on apolar and van der Waals forces to drive aggregation at the interface. Multicomponent supramolecular systems have proven capable of electrochemical interfacial molecular recognition.[7] In most iistances-however, molecular recognition will still rely on diffusion of a guest species to the immobilized host. Alternatively, some supramolecular systems have employed a host and guest that are covalently interlocked (mechanically linked) or maintained in some other type of ”fixed association” that prevents the complete dissociation (into separate solution components) of the host and guest from one another. This would appear to permit rapid interchange between their complexed and “dissociated” states. Self-assembled structures in this class have included catenanes, rotaxanes, shuttles, helicates, stacks and grid-like structures, many of which are discussed in detail in the chapters ahead. Typically these systems rely on multiple electrostatic interactions, especially x-donor and n- or metal ion acceptor type interactions. A clear advantage of systems in a “fixed association” is that many of the types of interactions that may suffer under typical electrochemical conditions, high ion concentrations and polar solvents, are compensated by the guaranteed proximity of the host and guest. In these systems, albeit on a much smaller scale, some of the same effects- proximity and exclusion of infewening solvent and electrolytes- that drive weak interactions to become dominant in proteins, may begin to become more apparent. Thus, electrostatic interactions may be somewhat more significant in such structures than might be anticipated (vide supra). The host-guest pairing may work far more efficiently for electrochemical switching. However, it is possible to envision scenarios in which this may not be the case. While fixed association guarantees the prospect of complexation, important criteria to be considered are the resulting effects on the kinetics and thermodynamics of the redox processes of interest. Changes in the dynamic association-dissociation of the host-guest complex may be expected to have significant effects on kinetic barriers and thermodynamic stability. For example, the work of Evans has shown that oxidation of ferrocene to ferrocenium in the
94
8 Electrochemical Considerations for Supramolecular Systems
presence of P-cyclodextrin takes place only when the ferrocene dissociates from the cyclodextrin.[81What then could be anticipated in the instance of a ferrocenyl moiety trapped with a host molecule? Recently, we have explored such a case for ferrocene encapsulated within a hernicarcerand.Ig1 In this instance the permanent association between host and guest alters the heterogeneous kinetics to such an extent that electrochemical conversion (switching) is significantly hindered. Achieving higher charge states in such a non-polar host environment may be difficult in host-guest systems with a fixed association. In other words, it might be reasonable to anticipate that the thermodynamic stability of the system could change, resulting in a shift in redox potentials, while a loss of electrochemical reversibility, as indicated by a larger separation (AEp) between the peak potentials for a redox couple, would attest to heterogeneous kinetic complications. Such points have been of recent interest both for their bearing on redox processes in biological systems, where redox centers may be buried in a hydrophobic protein core, as well as for their relation to the development of molecular information and storage devices. Other examples of systems in which the electrochemistry of redox active moieties is affected by fixed association have been noted in the literature. One example, presented in Chapter 12, involves a well known bis-paraquat cyclophane host, bearing a 4+ net charge, acting as a bead in a rotaxane. The highly charged cyclophane exerts a dramatic effect on the redox behavior of an aromatic unit in the thread of the rotaxane. This effect, both thermodynamic and kinetic, reflects the substantial electrostatic repulsion created by oxidation of the thread moiety, which generates a more highly charged (from overall charge of +4 to the oxidized states of +5 and +6) system. Due to the dramatic shift in the oxidation potential of the thread moiety and its sluggish kinetics, the usefulness of this particular system as a redox-switch is lost. Thus, the potential for loss of facile electrochemical reversibility in such systems is an important consideration for the supramolecular chemist, because of the implications for switching control. Ideally, electrochemical switching control in supramolecular systems should be via processes that are fast and reversible, in order to assure complete conversion (switching)within reasonable potential limits.
8.3 Systems Involving Multiple Identical or Non-Identical Redox-Active Moieties Many redox-switchable supramolecular systems are designed with the capacity to undergo electron transfer at multiple sites. The voltammetric behavior of such systems can be strongly influenced by the extent of electronic coupling between these sites. In this section we consider the voltammetric behavior presented in three different scenarios- one involving non-identical redox sites and two cases, in which identical redox moieties may be uncoupled or strongly coupled.
8.3
Systems Involving Multiple Identical or Non-Identical Redox-Active Moieties
95
In systems with several non-identical redox active moieties, the electronic effect of one redox-active substituent on another may exert an d u e n c e on both the thermodynamics, e.g. El/& and the kinetics, e.g. the magnitude of AEp, of electron transfer of either or both moieties. Interpretation of the observed electrochemical behavior should consider whether the sites are covalently or mechanically linked. In contrast to the mechanically linked structures mentioned above, the comparison of monomeric to dimeric (or higher order) structures may be less straightforward because of changes in the electrondonating or -accepting character of the extended structural framework. Here we illustrate one such example by considering a simple system composed of two metal cluster sites. Ligand 1, prepared by Haga and coworkers, provides two benzimidazole sites that can coordinate to Ru(I1) and Os(I1) ionsP" The authors also examined the analogous complexes formed with monomeric benzimidazole 2. The metal complexes employ two bipyridine (bipy) ligands to complete the coordination sphere of each metal ion.
2
First we shall consider the redox behavior of the monomeric structures. Os(II)2(bipy)ndemonstrates reversible oxidation of Os(I1) at 0.59 V vs. Ag/AgCl, while the Ru(II)2(bipy)zcomplex exhibits its (reversible) oxidation at 0.96 V vs. Ag/ AgC1. The mixed ligand (bipy)zOs(II)lRu(II)(bipy)z exhibits small shifts in the oxidation potentials of the metal ions. The oxidation of Os(I1) occurs at 0.56 V vs. Ag/AgCl, while that of Ru(I1) occurs at 0.99 V. Contrary to what we might expect, that the oxidation potential of the Os(I1) site to its higher redox state might be negatively mfluenced electrostatically by the presence of the Ru(I1) site, the Os(I1) site in the dimer is actually somewhat more easily oxidized. Oxidation of the Ru(I1) site is rendered slightly more difficult, however. Several factors may determine the extent of interdependence on the electrochemical behavior of
96
8
Electrochemical Considerations for Supramolecular Systems
the neighboring non-identical sites. For covalently linked redox sites both the separation distance between the sites, and the extent to which the molecular framework permits electronic communication between sites, are obviously sigmficant determinants of the electronic coupling. The differences between the oxidation potential for the respective metal centers in the monomeric complexes and the asymmetric dimeric complex are relatively minor. In fact, similar shifts in the oxidation potentials are also exhibited with the symmetric bis-Os(I1) and bis-Ru(I1) complexes of 1. The bis-Os(I1) complex shows its oxidations at 0.53 and 0.58 V vs. Ag/AgCl, while the bis-Ru(I1) complex shows its oxidations at 0.94 and 0.98 V vs. Ag/AgCl. Thus, the more facile oxidation of the first site in complexes of 1 appears to reflect the greater electron-donating capacity of the dimer's extended ligand. The extended x-ligand can offer greater stabilization of the higher charge state on the first oxidized metal center. The small positive shift in the oxidation of the second metal center probably reflects electrostatic repulsion, although the reduced electron-donating capacity of the ligand (which is stabilizing the first oxidized site) may also play a role. In the case of mechanically linked redox sites no such molecular framework communication is present. Clearly the extent of electrostatic attraction or repulsion is affected by the proximity of the redox sites. Controlling the proximity of redox-active sites in mechanically linked structures has been successfully employed in molecular switching. For instance, redox-switched changes in the proximity of electroactive components of rotaxanes form the basis of redox-controlled molecular shuttles (see Chapter 12). When the proximity between the sites is not altered by redox-switching, the resulting electrostatic repulsion or attraction may yield sigmficant changes in the observed kinetics and half-wave potentials of one or both sites, as was mentioned in the previous section. The symmetric metal complexes with ligand 1 allow us to pose the question of what behavior we should expect for structures bearing multiple identical redox sites. Should we expect that both Os(I1) sites will be oxidized at the same potential in the symmetric dimer? How can we interpret the 50 mV separation between the first and second half-wave potentials for the Os(II/III) redox couples? The redox behavior of molecules bearing multiple identical redox sites has been examined in detail by Shain,lW and Bard[l'bI and has continued to be a topic of interest for a number of electrochemists. Sav6ant was the first to point out that a molecule bearing two identical noninteracting redox sites would yield a voltammetric wave with the shape of a single electron transfer reaction, although more than a single electron is transferredP1 The characteristic separation between the half-wave potentials of the two electron transfer processes, in this case equal to a value in volts of (RT/F) In 4, is determined by simple statistics. Bard and Anson further studied the voltammetric response of poly(vinylferrocene) of various weight distributions and extended the statistical treatment.[W Based on the work of these authors, consideration of the expected behavior for systems bearing multiple identical noninteracting electroactive centers is quite straightforward.
8.3
Systems lnvolving Multiple ldentical or Non-ldentical Redox-Active Moieties
97
For a polymeric molecule bearing n independent identical electroactive sites each site should have the same standard potential, E,o and a corresponding half reaction can be defined for the reduction of any one of these sites:
Emo
XXXOXXXX +
e-
+-
e-
XXXXXXXO
Emo
XXXRXXXX
XXXXXXXR
0 and R represent the oxidized and reduced states of the center while X represents a site in either state. At equilibrium the probability that any site i is in the reduced state is given by:
where 0 is given by: 8 = exp[ &(E
-
EO,
i]
and E is the equilibrium potential. The net oxidation state of the molecule is given by the sum of the difference between the total number of sites n and the number of sites in the reduced state, j , that is ( n - j ). Simple binomial distribution leads to the fraction of polymer molecules present with exactly j number of redwed centers: f. = I
[;)(
0 )(n-j)( 1 )i
1+0
(3)
1+0
In a molecule bearing noninteracting centers it is possible to calculate formal potentials for the successive oxidation states. By use of eq. 2 and 3 it can
98
8 Electrochemical Considerations for Supramolecular Systems
be shown that the formal potential for the molecule in oxidation state given byh is:
Based on eq. 5 we can see that for n = 2 the first formal potential (for XR + XO or RX + OX) occurs at -17.8 mV from the observable oxidation wave ( E I ~ ) , while the second formal potential (RO + RR or OR + RR) is symmetrically distributed about this point at +17.8 mV from the observable E 1 p . These small shifts are not discernable by most voltammetric methods. Thus, the voltammogram has the appearance of a normal, single electron process, albeit with two superimposed peaks with differences in their half-wave potentials of 35.6 mV, and a higher current intensity, reflecting the fact that a two electron process is occurring. The peak current intensity may be close to twice that expected for a single electron process. We should note that current intensity is nof an accurate method for assessing the number of electrons transferred. Bulk electrolysis is the appropriate method for this determination. In the case of a dimeric species, the diffusion coefficient is likely to be lower than that of the monomer and thus we would anticipate that the current intensity will be proportionately lower. (Recall from Chaps. 3 and 4 that the current intensity in potential sweep methods is proportional to the square root of the diffusion coefficient.) For molecules with three identical sites (n = 3) we can see that the three formal potentials are distributed about the observed E1/2 at-28.5, 0.0 and +28.5 mV. In a molecule with four identical sites (n = 4) the spacing of the formal potentials occurs at -35.6, -10.4, +10.4 and +35.6 mV. For n = 5 the spacing between the successive formal potentials is distributed about the half-wave potential in according to -41.4, -17.8, 0.0, +17.8, and 41.4 mV. As the number of identical sites increases, the large overlap of concentrations of the partially reduced species begins to affect the appearance of the voltammogram. The observed voltammetric wave begins to broaden. Fig. 8.1 shows the positions of the successive formal potentials for n = 2-6. The formal potentials for the first and last pair of oxidation states in the molecule are given by:
Thus for a molecule of ten identical noninteracting sites the first and last formal potential are separated by 118.3 mV. Again, the voltammetry of such a molecule would reflect both the slower diffusion of the large structure to the electrode surface and a broadened appearance, attributable tp the multiple species undergoing oxidation as the region of the half-wave potential is scanned.
8.3 Systems Involving Multiple Identical or Non-ldentical Redox-Active Moieties
ElF- EnF
E1,20bS
I
n=2
I I
I
n=4
c
n=5
n=6
35.6
I
n=3
I
57
I
71.2
I
I
I
99
I
82.8
90.9
Figure 8.1: Separations for the successive formal potentials of molecules with n identical, noninteracting redox sites. Note that for n 2 4 the separations become nonuniform.[W E i F - EJ is the difference between the formal potentials for the first and last oxidation states.
Based on the foregoing discussion we can see that the symmetric bisOs(I1) complex of 1 shows very slight evidence of interaction between the two redox sites. Rather than displaying the anticipated single wave with the expected but not observable peak to peak separation of 35.6 mV, the separation between the oxidation potentials, El - Ez, was estimated to be around 50 mV. Such small separations in half-wave potentials are often difficult to discern accurately and are best examined by differential pulse voltammetry (see Chap. 4). This value presents a deviation from the behavior expected for a system with identical noninteracting sites. The magnitude of the separation between halfwave potentials of each site is a measure of the extent of electronic coupling interaction between the redox sites. Isomeric covalently linked systems with multiple identical sites provide examples of the variation in the extent of intramolecular interaction according to changes in the molecular framework. Metal-metal interactions can provide particularly dramatic examples of electronic cn*ipling. Below we briefly mention two cases of electronically
100
8
Electrochemical Considerations f o r Supramolecular Systems
coupled redox sites, one on isomeric calixarene-based structures and one based on metal helicates. Diquinonecalix[4]arenes3 and 4 differ in their placement of the quinone rings. Not surprisingly, this leads to differing degrees of interaction between the two quinone sites. Diquinone 3, in which the quinones are proximal, shows two well-separated redox waves corresponding to the first reduction of each of the two quinone moieties, as shown in Fig. 8.2a. The difference between the two half-wave potentials (AE1/2), is 297 mV. In contrast, the voltammetric response for 4 reveals that when the quinones oppose one another, the extent of electronic communication is lessened: AE1/2 decreases to 141 mV.I*41
Figure 8.2: Cyclic voltammetric response on a GC electrode obtained for 1mM solutions of 3 (a) and 4 (b) in CH3CN/O.l M TBAPF6. Scan rate was 100 mV/s, potentials referenced to Ag/AgCl.[**l Used with permission of the author.
8.3
Systems Involving Multiple Identical or Non-Identical Redox-Active Moieties
101
Helicates are a special class of metal complexes covered in detail in Chap. 13. Ligand 5 (a terpyridine derivative) forms Cu(I/II) helicates composed of two metal ions and two ligands. The solution structure of the Cu(1) complexes of 5 has been reported to have an unusual diamond-like tetrahedral ge~rnetry"~] (see inset in Fig. 8.3). This compIex, which is discussed in greater detail in Chap. 13, is novel in that it shows an extremely dramatic separation of 860 mV between the first and second Cu(I/II) redox couples, as shown in Fig. 8.3. Both processes were quasireversible in CH3CN. n
= Pyridine
I
1
1
1
+1.W
I
I
I
I
I
0= Cu(1)
0 = Phenyl
I
M.50
I
I
1
1
I
I
I
0.0
E vs SSCE Figure 8.3: Cyclic voltammograms of a 1.1mM solution of [Cu(II)52] in CH3CN/TBAPF6. The scan rate was 100 mV/s and the electrode surface was platinum. The inset shows the space-filling representation of [Cu(I)5,] based on the x-ray structure of right-handed helicate.Ps1 Reprinted w i t h permission of the American Chemical Society.
102
8 Electrochemical Considerations for Supramolecular Systems
8.4 References 1. C. Reichardt, Solvent Efects in Organic Chemisfy, 2nd Ed., VCH, New York, 1988, p. 408 - 410. 2. Clearly here we take into consideration only the decrease in solution resistance provided by the higher dielectric solvent. Other considerations such as solvent volatility may make CH2C12 more ideal from the standpoint of sample recovery. 3. Under these conditions the reader is referred to Chapter 4 for the discussion of normal pulse voltammetry, which may aid in the study of analytes which tend to precipitate on the electrode surface in their switched oxidation states. 4. [a] D. B. Smithrud, T. B. Wyman, F. Diederich, J. Am. Chem. SOC., 1991, 113, 5420 - 5426; [b] D. B. Smithrud, F. Diederich, J. Am. Chem. SOC., 1990,112,339 - 343. 5. D. A. Dougherty, Science, 1996,271,163 - 168. 6. M. A. Petti, T. J. Shepodd, R. E. Barrans, Jr., D. A. Dougherty, J. Am. Chem. SOC., 1988,110,6825 - 6840. 7. [a] M. T. Rojas, R. Koniger, J. F. Stoddart, A. E. Kaifer, J. Am. Chem. SOC., 1995, 11 7,336 -343; [b]M. T. Rojas, A. E. Kaifer, 1. Am. Chem. SOC., 1995,117,5883 5884; [c] 0.Chailapakul, D. Crooks, Langrnuir, 1995,ZZ,1329 - 1340. 8. T. Matsue, D. H. Evans, T.Osa, N. Kobayashi, J. Am. Chem. SOC.,1985,107,3411 - 3417. 9. S. Mendoza, P. D. Davidov, A. E. Kaifer, Chem. Eur. J., 1998,864 - 870. 10. M.-a. Haga, T.-a. Ano, T. Ishizaki, K. Kano, K. Nozaki, I. Chern. SOC., Dalfon, 1994,263 - 272. 11. [a] R. L. Myers, 1. Shain, Anal. Chem. 1969, 41 980-990. [b] A. J. Bard, Pure App. Chem., 1971,25,379-393. 12. F. Ammar, J.-M. Saveant, Electroanal. Chem., 1973,47,215-221. 13. J. B. Flanagan, S. Margel, A. J. Bard, F. Anson, J. Am. Chem. SOC., 1978, ZOO, 4248 - 4253. 14. M. Gomez-Kaifer, Ph.D. Dissertation, University of Miami, 1997, 145-147. Used with with permission of the author. 15. K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abruiia, C. Arana, Inorg. Chem., 1993,32,4450-4456.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
9 Electrochemical Switching
Electrochemical switching is perhaps the most important application of electrochemistry in the field of supramolecular chemistry. This concept, whose use has been widespread in the past two decades, has been exploited to such an extent that a thorough examination of its use is well beyond the scope of this book.111 The early focus of this concept in supramolecular research was primarily cation recognition. During the 1980s much interest centered on redox-switchable cation binding systems as potential mimics of biological cation transport or the development of analytical methods for cation sensing. More recently this emphasis has broadened to include anion and molecular recognition. Without a doubt however, the predominant interest in this field has shifted toward the development of switchable molecular devices. Electrochemical switching provides an easy means for con trolling the molecular architecture of redox-active supramolecular systems. In the four chapters that follow, we will see the prominence of this concept in the field. In this chapter we will examine the basic concept of electrochemical switching, and by way of example, contemplate its application in two areas: cation binding, and its power to exert control over molecular architecture.
9.1 The Concept of Electrochemical Switching The concept of an electrochemically switchable molecule is a simple one. Such a molecule displays differing affinity with a second species based on its redox state. The oxidation state of the redox-switched component of the pair determines the thermodynamic stability of the complex formed between the two species. The basis of this differential affinity is purely electrostatic. Perturbation of the charge in a redox-active host or guest can result in increased or decreased binding affinity. When the magnitude of this change in interaction energy is strong, the electrochemistry may clearly reflect two different redox states, i.e. the interacting and noninteracting species may give rise to different half-wave potentials. In electrochemically-switchable systems either host, guest, or both, may be redox active. The requirements for such a host or guest are essentially the same. First, in order to provide meaningful switching control, the redox-active moiety to be switched should exhibit reversible heterogeneous electron transfer kinetics. Without reversible kinetics, the switch itself is in essence rendered too slow to be useful. A second requirement is that at least one of the redox states
104
9
Electroclleniical Switching
must interact strongly with its targeted species. These two prerequisites apply for any switchable system- from ion binding agents to molecular shuttles. Analytical work presents additional requirements. In general, for work involving sensors, the half-wave potentials of the free and bound states should ideally be well separated from one another in order to see true two wave behavior. This point is important because the intensity of a voltammetric wave is proportional to the concentration of the corresponding complexed or free molecule. Thus, for most quantitative analytical work there should be two clearly defined states that are readily discerned via voltammetric techniques. These states are defined as "on-off" or more accurately, "high-low" states of interaction. For the easiest analysis of a switchable binding scenario, both forms of the redox-active host or guest should have sufficiently high binding affinity if separate redox waves for the two states are to be observed by voltamrnetry.[21 This is an important consideration when trying to extract accurate binding constant information from voltammetric data. In most instances however, digital simulations provide the best analysis. A simple thermodynamic square scheme can be employed to elucidate the equilibria in a redox-switchable system. Such a scheme is shown in Fig. 9.1. In this system, the redox equilibrium is coupled with a reversible binding reaction. For this example, we assume that the host H is electroactive and is switched from its low to high binding state by reduction. In this square scheme, host H forms more stable complexes with guest G when it is reduced (H-)than when it is oxidized (H). Therefore, the association constant KH is much larger than KL. In this scenario, when the initial state shows a lower association constant than the electrochemically switched state the magnitude of the ratio of association constants KH/KL is defined as the binding enhancement. The value of this term may be estimated (vide inpa) by the difference in formal potentials for the free and complexed host as shown in Eq. (1):
KH/KL
= exp[-F(E$
-Eg)/RT]
In this expression EF is the formal potential of the free ligand, Ec the formal potential of the complexed ligand, (both values are usually approximated by the
Figure 9.1:A square scheme for binding equilibria with a redox switchable host.
9.2
Switcluzble Binding in a Redox-Active Cation Host
105
half-wave potentials), F is the Faraday constant and R the molar gas constant in J/mol.K. The larger the difference in half-wave potentials, the greater the magnitude of the binding enhancement. The value of KL determines whether H or H-G is reduced at the electrode surface. When KL is large, the complex H-G, already formed, is electrochemically reduced to a higher affinity state, H--G. In this situation, the diffusion of the guest species is not a relevant factor. In contrast, if KL is not large the species undergoing reduction will be H, which will subsequently bind the guest to yield H--G. In this instance, after reduction of the free host to the "high" binding state (H-), the complexation process is essentially diffusion controlled, i.e. determined by the diffusion of available guest to the reduced receptor. Generally, two separate voltammetric waves corresponding to the redox processes of the free (H) and complexed (H-G) species will not be observed for systems with a low KL. In this binding regime the rate constants for complexation and decomplexation may be an important consideration (vide infiu). The typical electrochemical response observed for a system with a very low KL will be a shift of the half-wave potential for the free host species as guest is added to the solution, i.e., separate waves will likely not be observed for the free and bound species.PI We shall consider these various aspects in greater depth in the foIIowing section.
9.2 Switchable Binding in a Redox-Active Cation Host Crown ether and cryptand structures are well known hosts for a variety of metaI ions. When redox-active substituents are added to a crown ether or cryptand structure, the binding of a metal ion may affect the redox behavior of the host. Ferrocenyldimethyl-[2.2]-cryptand1 is a sensitive redox probe for the presence of a variety of ions, including Na+, K+, Ca*+,and Ag+.[4]1exhibits the expected voltammetric behavior of a ferrocene derivative, with a reversible monoelectronic oxidation at +0.216V vs. SSCE. Based on the nature of ferrocene electrochemistry, we would expect that binding of a cation in the cryptand region of this host would shift the oxidation of ferrocene to more positive potentials, due largely to electrostatic repulsion. This expectation is confirmed when we examine the cyclic voltammetric behavior of 1in the presence of an
1 ' Oq\
I I
106
9
I
-0.2
Electrochemical Switching
I
0.7 -0:2 €/V vs. SSCE
0.7
Figure 9.2: Cyclic voltammetry of 1 in CHKN/O.lO M TBAPF6, scan rate of 50 mV/s. (a) No Na'; (b) 0.25 equivalents of Na'; (c) 0.50 equivalents; (d) 0.75 equivalents; (e) 1.0 equivalents, ( f ) 3.0 Reprinted with permission of the Royal Society of Chemistry.
increasing concentration of Na', as shown in Figure 9.2. In the presence of 0.25 equivalents of Na+, a second redox wave emerges at 0.402 V vs. SSCE (Fig. 9.2b) As the concentration of the salt is increased this second wave grows in intensity at the expense of the current observed for the original. The fact that 1-Na' is oxidized at more positive potentials confirms that the binding of Na' to the cryptand destabilizes the ferrocenium form of the complex due to the increased positive charge near the redox active site. Examining Fig. 9.2, we can easily see that both the free and complexed 1 yield reversible voltammetric behavior. What can we assume about the nature of 1-Na' complex from the fact that it too, yields a reversible redox couple? Here we can pause to analyze each of the components in the switching scheme. Fig 9.3 shows the square scheme for 1, along with the association constants for 1 and 1+ with Na+, as determined for 1 by potentiometric experiments and for 1' as calculated by the shift in the half-wave potentials (more on this point later).
9.2
Switchable Binding in a Redox-Active Cation Host
107
While it is evident that 1 has a high binding association constant for Na+ it would seem that 1' also has a moderate binding affinity for Na+. Based on the
Kal= 2 x lo6
1
+ Na'
1' + Na'
-
-
L
1-Na'
-
1'-Na'
Figure 9.3: A square scheme for the equilibria of 1and it's Na+complex.
voltammetric results, this is a reasonable assumption since the new redox wave that appears for the 1-Na+ complex is reversible. What behavior would we expect if the binding constant for 1' were substantially lower, for instance on the order of 1 M-I? We might expect the appearance of the voltammogram to be different, namely in that after oxidation of the complex 1-Na', the significant loss of binding avidity might result in decomplexation of Na' and a subsequent decrease in the currents attributable to the reduction of 1'-Na'. This argument makes sense, assuming that the rate of decomplexation of the Na' ion is very fast. If we begin to examine the kinetics regime for ion complexation and decomplexation however, we know that the rate of decomplexation, from cryptands in particular, can actually be quite slow. What effect would the kinetics of decomplexation have on the voltammetric appearance of 1-Na'? If the value of the association constant were very low and the rate of decomplexation very slow, the voltammetric behavior might approach reversibility. If Na' cannot decomplex rapidly, even with low affinity for its host's new redox state, on the voltammetric timescale we might still see the reduction of 1-Na+with currents approaching those observed for the oxidation process. In this instance scan rate studies might be useful, although there are lower limits on how slowly we can sweep the potential and still avoid convection. In such situations, the only way to truly comprehend the electrochemical binding scenario is via digital simulations. Digital simulations of 1reveal that the ratio of binding constants for 1is actually much higher than that estimated by the difference in half-wave potentials. Although K1/KJ51 was initially estimated at 1500, digital simulations of this system later showed that this value must be closer to 3 x lo4 in order to reproduce the voltammetric behavior.@a] This ratio of constants drops the association constant for 1+-Na+to 64 M-I! Yet, even this low value does not necessarily imply that the kinetics of decomplexation is the determining factor in the voltammetric appearance of the reversible redox wave for 1-Na+. In the
108
9
- 0.2
~
Electrochemical Switching
~1
+ 0.8
Potential, V us SSCE a solution containing 1.0 mM 1 and 0.50 mM NaC104. (a) Experimental voltammogram at 100 mV/s; (b) simulated voltammogram using KI= 2 x 106 M-1 and K2=50.[4al Reprinted with permission of the American Chemical Society.
Figure 9.4:Voltammetric response of
simulation described, no assumptions were made about the kinetics of decomplexation, i.e., it is implicit that the rate of decomplexation is fast on the timescale of the vokammetric experiment, in order to allow all species to be at equilibrium. In spite of this lack of kinetic considerations, the excellent fit of the simulations to the experimental data is evident in Fig. 9.4. The authors have theorized that while the rate of decomplexation of Na' may indeed be slow, the complexation equilibria appear to be frozen on the voltammetric timescale.[4] Even very slow scan rate studies were unable to differentiate a decrease in the reduction wave for 1-Na' complex that would suggest decomplexation. Evidence of decomplexation effects has been noted in related systems, however. Ferrocenyl macrocycle 2 is the synthetic precursor of 1.This cryptand also exhibits significant binding and distinct oxidation waves for both the free and bound forms of 2 in the presence of Be2'.[61 In this instance however,, differences in the cathodic and anodic current intensities suggested dissociation of the 2+-Be2' complex on the cyclic voltammetric timescale. These effects were evident even with moderately fast scan rates (400 mV/s). In the presencqof Mg2', Ca2+,Sr2+or Ba2+,2 exhibited voltammetric behavior typical of lower KL
9.3
Electroclleniical Switching
a5
a Means of Controlling Molecular Devices
109
systems: a single oxidation wave was observed for the free and bound ligand, and the E1/2 shifted from that of the free ligand.161 What factors can aid in the analysis of this class of redox-switchable binding systems? Digital simulation is clearly one of the most important tools. In titration experiments a wide concentration range of the bound species should be examined, i.e., ranging from substoichiometric amounts to as large an excess number of equivalents as solubility permits. Examining the redox behavior in the presence of excess guest permits the determination of more accurate halfwave potentials for bound states of the redox-active host. When KLis quite low, recaIl that shifts in the half-wave potential may occur, thus the true position of voltammetric peaks may not be evident without such a thorough titration study. NMR experiments may yield information about the rate of guest complexation or decomplexation, while both NMR and potentiometric studies (when possible) may provide a good experimental value for the binding constant of the initial state of the host. The foregoing analysis points to the complex factors that are at work in a redox-switchable host-guest system. These points are important when one is concerned with switchabIe systems for transport or sensing. However, in many instances the more recent applications of switching have a completely different emphasis. In contrast to sensing or transport systems, which may operate on a "high/low" principle, redox-switching in many of the other avenues of supramolecular research may truly operate on an "on/off" principle. In these cases, the idea of electrochemical switching is applied with a different focus.
9.3 Electrochemical Switching as a Means of Controlling Molecular Devices and Other Structures The goal of one avenue of supramolecular research is the development of switchable molecular devices. These electrochemically or photochemically active molecules can undergo a change in molecular structure that permits assignment of clearly defined on/off states. To be useful, a switch must show reversibility, and as we have mentioned in Section 9.1, a redox-switchable structure must have fast heterogeneous kinetics. This implies that the rate of switching itself will not be limited by the rate of electrochemical processes.
110
9
Electrocllernical Switching
What types of molecules display this device-like switching? And what types of changes are induced by chemical or electrochemical redox-switching? If we consider that a switch requires clearly defined on/off states we can relax the requirement that both states of the redox-switchable species interact with a second species. For instance, in a molecular shuttle the bead-like host is switched in between two topologcally linked guest sites on the rotaxane thread.l71 Electrochemical-switching of one of the guest sites results in translocation of the bead along the thread to the other site. In this instance, while the bead still interacts with the thread, in reality, the host is now interacting with a different moiety on the thread. The same type of reactions could be observeU if all three individual components (the host and two nonlinked guests) were in solution. However, the great limiting factor of such a switching system would be the slow rates of diffusion, as the host searches for a new guest that exists in a more compatible redox state. A design in which the host and two guests are linked in a rotaxane structure, greatly enhances the functional rate of the switch. Nevertheless, occasionally effective switchable systems can involve more than two solution species. Rotello and coworkers have developed just such a three component system, in which a change in the oxidation state of one molecule switches its affinity for a second species to a thirdP1 Broadening the definition of switching even further, we can eliminate the requirements for a second species and consider the effects of redox-switching on the molecule’s self-interaction. A process as familiar as electrochemicallyinduced dimerization could be considered an example of switching: a different species, with potentially very different structural and spectral properties, is generated. A prominent example of this concept is the helicates, a class of molecules in which redox-switching can induce the reversible self-assembly of new species.I91 These various systems are shown schematically in Fig. 9.5.
Figure 9.5: Concepts of several electrochemically switchable systems. From top: a three component switch, a molecular shuttle and a helicate.
9.3 Electroclzemical Switching as a Means of Controlling Molecular Devices
111
Drawing upon the above-mentioned concept, many of the most promising switchable systems are those in which all the switching components are topologically linked. In this sense, while the components of the system may not be covalently bound to one another, their proximity to one another is assured. In this class of molecules there are many examples of redox-switchable intertwined or helical structures. Although many of these systems have been explored eIectrochemically, there are s t i l l a sigruficant number whose redoxswitching has been examined by chemical means. One such intriguing molecule is 3, a helical iron binding ligand that can translocate iron ions to two distinct coordination sites on the ligand as their redox state is changed from Fe2+and Fe3+ states.[I01 Readily discerned spectroscopic changes in the visible range sigrufy the change in the redox state and coordination sphere. 3 is a three armed ligand with an internal tris(hydroxymate) binding Fe3+ site and an external tris(2,2'-bipyridine) Fe*+ coordination site. The Fe3+ coordinated system is readily reduced by ascorbic acid and rapidly generates the tris(bpy) complex of Fe2+.The reverse process, oxidation with ammonium persulfate, regenerates the hydroxymate Fe3+complex on a slightly slower timescale, with the half-life of the reduction process on the order of 15 s. Curiously, a chiral ligand structurally related to 3 (bearing an alanyl residue) displays an even longer half-life for the reduction process, -45 s. In an important experiment pointing to the efficacy of modular design, the authors were unable to accomplish the analogous complexation intermolecular processes by employing two tridentate ligands bearing the hydroxymate and bipyridine residues, respectively. These
s I
,1
0
Fez+ -binding site
P O <
Edin
"So
+eL
c
-0-
1
site
J4 N
Light-Brown
Purple-Red
Figure 9.6: Helical structure 2 is capable of translocating iron ions between binding sites based on their redox stateS'01 Reprinted with permission of Nature.
9
112
Electrochemical Switching
experiments also appear to confirm that the iron does not dissociate from the helix but in fact translocates from one binding site to the another as the redox state is changed. Although conventional DC voltammetry has not been reported for this system, related thiolated structures have been studied by AC impedance spectroscopy.[lll Finally, we mention one last example of the application of electrochemical switching to supramolecular chemistry. By employing redoxactive amphiphilic compounds, redox switching can be utilized to control aggregation of vesicles and micelles. Gokel and Saji[**lhave pioneered this area of research. The first successful redox-switchable amphiphile reported to form aggregate structures, prepared by Gokel and coworkers, was a ferrocene substituted cholestanyl.[13] This compound formed multilamellar vesicles when the ferrocene substituent was oxidized to ferrocenium. Upon chemical reduction the vesicles collapsed. Later, the same group reported preparation of two unrelated metalloamphiphiles that also afford redox-switched vesicles.[141 The amphiphilic phenanthroline derivative 4 was shown to form multilamellar vesicles with a hydrodynamic diameter of 1500 A. These vesicles collapsed upon introduction of a chemical reducing agent.
2 a04
4
More recently, Gokel has reported a series of diferrocenyl bolaamphiphiles.[~51Structure 5 is one of the bolaamphiphiles studied. The diferrocenyl compounds all pdssess a single oxidation wave for the oxidation of the two ferrocenes to ferroceniums, indicating that the two moieties are independent of one another, This suggests that no electronic communication occurs across the hydrocarbon bridge and that the molecule does not adopt a solution conformation that permits communication between the ferrocene units. Vesicle formation is observed upon oxidation of the bolaamphiphile and is disrupted upon reduction. Thus redox-switching can be employed to direct aggregation in these molecules.
Fe
Fe 5
9.4 References
113
9.4 References 1.A number of thorough reviews are available on the topic of redox-switching. These include [a] A. E. Kaifer in Comprehensive Supramolecular Chemisby, Vol. 1,
(Eds.: G. W. Gokel), Pergamon, Tarrytown, NY, 1996; [b] A. E. Kaifer, L. Echegoyen, in Cation Binding by Macrocycles, (Eds. Y. Inoue, G. W. Gokel) Dekker, New York, 1990; [c] P. D. Beer,, Chem. SOC.Rev.,1989,18,409-450; 2. [a] Ref. 2a, p.701; [b] S.R. Miller, D.A. Gustowski, Z.-H. Chen, G.W. Gokel, L.A. Echegoyen, A.E. Kaifer, Anal. Chem., 1988,60,2021-2024. 3. Ref. la, p. 709-711. 4. [a] ] J. C. Medina, T. T. Goodnow, M. T. Rojas, J. L. Atwood, B. C. Lynn, A. E. Kaifer, G. W. Gokel, J. Am. Chem. SOC.,1992, 114,10583-10595; [b] J. C. Medina, T. T. Goodnow, S. Bott, J. L. Atwood, A. E. Kaifer, G. W. Gokel, Chem. Commun., 1991,290-292. 5. We should note that for the cases in which the initial state of the switchable host has the higher binding constant, than it's alternative redox form, we can no longer call this ratio of Kl/K2 a binding enhancement. In fact, in this situation the alternative redox state offers the opposite of enhancement.. One could perhaps consider the term binding depression (or suppression) as an alternative. 6. C. D. Hall, N. W. Sharpe, I. P. Danks, Y. P. Sang, J. Chem. SOC., Chem. Commun., 1989,419-421. 7. R. A. Bissell, E. Cordova, J. F. Stoddart, A. E. Kaifer, Nature, 1994,369,133-137. 8. R. Deans, A. Niemz, E. C. Breinlinger, V. M. Rotello, J. Am. Chem. SOC., 1997, 119,10863-10864. 9. See for example: J.-P. Gisselbrecht, M. Gross, J.-M. Lehn, J.-P. Sauvage, R. Ziessel, C. Piccinni-Leopardi, J. M. Arrieta, G. Germain, M. V. Meerssche, Nouv. J. Chimie, 1984,8,661-667. 10. L. Zelkovich, J. Libman, A. Shanzer, Nature, 1995,374, 790-792. 11. Y. Gafni, H. Weizman, J. Libman, A. Shanzer, I. Rubinstein, Chem. Eurp. I., 1996,2, 759-766. 12. [a] K. Hoshino, T. Saji, Y. Ishii, M. Goto, 1. Am. Chem. Soc., 1991, 113,450-456; [b] K. Hoshino, T. Saji, 1. Am. Chem. SOC., 1987, 109, 5881-5883; [c] T. Saji, K. Hoshino, S. Aoyagui, J. Chem. SOC., Chem. Commun., 1985,865-866. 13. J. C. Medina, I. Gay, Z. Chen, L. Echegoyen, G. W. Gokel, J. Am. Chem. SOC., 1991, 113,365-366. 14. S . Muiioz, G. W. Gokel, J. Am. Chem. SOC.,1993,115,4899-4900. 15. K. Wang, S. Muiioz, L. Zhang, R. Castro, A. E. Kaifer, G. W. Gokel, J. Am. Chem. SOC.,1996,118,6707-6715.
Supramolecular Electrochemistry Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
10 Electrochemically Switchable Cation and Anion Binding
Much of the early interest in redox-switching has centered on ion binding. Cation-binding systems were the first switchable receptors .designed with the goal of attempting to mimic biological transport systems.[ll Among the electrochemically-switched systems designed were lariat ethers, ferrocenyl macrocycles and cryptands, quinone-based ligands, and more recently, calixarenes. Anion binding systems, recent in their development, have begun to enjoy some success. We shall examine several examples of these types of systems.
10.1 Electrochemically-Switched Cation-Binding Systems In the fifteen years since electrochemical switching was first successfully applied to the field of cation binding and transport this area has grown extensively. At the present date an abundance of literature is available on this topic. Thus, in this section we shall examine seIect examples from four classes of cation-binding systems: the lariat ethers, ferrocenyl-, and quinone- based crowns and cryptands, and caharenes. For readers interested in further information a number of books and thorough reviews are available.[21 10.1.1The Lariat Ethers
Lariat ethers were the first macrocyclic receptors to show significant electrochemically-switched binding capacity. A typical lariat ether is composed of a crown ether macrocycle to which an electroactive sidearm is appended. Structures 1-3 are three redox-active lariat ethers in which the nitroaromatic moiety provides the switching capability. Taken together, these three strucutres represent a progression in synthetic design in order to enhance binding. In these systems the goal is to enhance the binding of the cation by providing coordinating atoms from both the macrocyclic receptor and from the nitro group on the sidearm. The expected observation is that a bound Na' ion will shift the half-wave potential of the nitroaromatic group if this group interacts with the ion. Easier reduction of this group sigrufies the thermodynamic stabilization of the complex. Ether 1 bears a nitroaromatic sidearm with the nitro group in the para position. Due to the para placement of the nitro group, this ligands electrochemical response is relatively insensitive to the presence of added Na+.r31
115
10.1 Electrochemically-Switched Cution-Binding
1
2
3
In contrast, lariat ether 2, with its nitro group placed ortho to the pivoting sidearm, was found to be extremely sensitive to the presence of Na+.[31 Substoichiometric additions of the ion resulted in development of a new redox wave associated with the reduction of the 2-Na' complex, at more positive potentials than the reduction observed for the free ligand 2, as anticipated. These findings indicate how crucial the orientation of the redox moieties may be to attaining successful switchable binding enhancement. Lariat ether 3 possesses the pivoting sidearm attached to a nitrogen atom. This variation increases the flexibility of the sidearm and provides an even greater switchable binding sensitivity than possessed by 2. When Na' is bound by the macrocyclic crown ring, it can simultaneously interact strongly with the nitroaromatic group, as shown in Fig. 10.1. At substoichiometric amounts of Na+, two reduction waves are seen, one for the free and one for the Na+-bound species of 1,see Fig. 10.2ae.C41 The reduction potential of 3-Na' was anodically shifted by 260 mV compared to that of 3 because the nitroaromatic anion radical is stabilized by Na'. The cyclic voltammetric behavior of 3-Na' was readily simulated, as shown in Fig.10.2f-jJ41 The behavior demonstrated by this system is reminiscent of that seen in Chap. 9 for a ferrocenyl cryptand. In this instance, however, the initial (neutral) state of the ligand does not show the higher binding affinity. The estimated binding enhancement for 3-Na' was 2.5 x 104 in favor of the reduced species. ESR spectra of the anion radical of several nitrobenzene lariat ethers in
3
3-Na+
3--Na+
Figure 10.1:Structure of free 3, and schematic structures of %Na+and 3=Na+.
10
116
Electrocllemically Switchable Cation and Anion Binding
Experiment
0.0
-
Simulation
I
-1.0
E (V vs. SSCE)
-1.35
~
0.0
-
-1.0
-1.35
E (V vs. SSCE)
Figure 10.2: Cyclic voltammetry of a 1 mM solution of 3 in CHCN/TBAPF6, in the presence of increasing amounts of NaC104. (a) - (e) experimental voltammograms, (f) (j)simulated voltammograms, at a scan rate of 100 mV/s. From the top, the amounts of NaC104 are 0.0, 0.25, 0.50, 0.75, and 1.0 equivalents for each set of voltammograms.[41 Reprinled with permission of the American Chemical Society.
the presence and absence of alkali metal cations have been reported.[51 The electrochemical switching response of this ligand promoted further interest in this area of research. 10.1.2Quinone-Based Ligands
Quinones are important redox-active substituents in a variety of biological systems. Bock et al., were the first to report a quinone-based redox-switchable ligand, although their system employed chemical reduction.[b] In keeping with the goal of exploring electrochemically driven cation transport with synthetic ligands designed to mimic biological systems, lariat ethers incorporating
117
7 0.1 Electrochemically-Switched Cation-Binding
0
quinone switches were examined. Structure 4, shows a lariat with an anthraquinone moiety on the pivoting arm. The reduced anthraquinone 4demonstrated good transport capabilities when used in the source interphase of a transport cell, in combination with 4 (neutral) in the receiving phase. The presence of the neutral ligand in the receiving phase was found to speed up the transport rate. This method of employing both the reduced and neutral states of the carrier was termed "pumping".[7-1 The related bismacrocyclic ligand 5 shows a more substantial binding enhancement than 4PI Other switchable quinoidal ligands include 6 and 7, an anthraquinone crown and cryptand, respectively. The free ligand of crown compound 6 exhibits two reversible monoelectronic reduction waves.[91 In the presence of 0.5 equivalents of Na' two new waves appear, corresponding to the reduction processes of the free and complexed ligand. Cryptand 7 demonstrated a more dramatic response, with a 400 mV anodic shift for the reduction of the anthraquinone moiety upon binding two Li'
&
0
\
io 0
O
0 O3
C
O
J
6
7
10 Electrochemically Switchable Cation and Anion Binding
118
ions. Such a shift corresponds to a binding enhancement of 8 x 105, while reduction to the dianion leads to a total binding enhancement of 1011 relative to the neutral cryptandF" This magnitude of enhancement is significant. Cryptand 7 also binds Na+ and K+. In its reduced state, 7 binds K+ so strongly that it cannot be removed competitively by cryptand [2.2.2].[1°1 Other recent anthraquinone ligands with electrochemical-switching capacity include a series of anthraquinones bearing pendant diazacrown ethers, and bis(anthraquinone) systems with diazacrown ether spacers.[lll Na' binding enhancements obtained upon one electron reduction in several of these systems have exceeded lo5. 10.1.3 Ferrocenyl Macrocycles and Cryptand Switches
Designs of electrochemically-switchable receptors for cations made further strides throughout the late 1980's and early 1990's. Ferrocene based systems have become prevalent in the literature, due in large part to their ease of preparation and ferrocene's reversible heterogeneous et kinetics. Ferrocene shows success as a redox antenna in that its oxidation potentiaI is sensitive to the binding of other molecules. Two ferrocenyl cryptands were described in Chapter 9. Most ferrocenyl ligands follow the pattern seen with those two cryptands, i.e., the binding of a cation in the macrocycle makes the ferrocene moiety more difficult to oxidize. Thus, it is a surprising result that Fabrizzi and coworkers have prepared a novel ferrocene receptor that has an increased affinity for a cation upon oxidationP1 The ferrocenyl ligand 8 bears diamine ketone side arms. Reduction of the carbonyls yields 9, whose side arms are reminiscent of an open cyclam. In the presence of Ni2+ the cyclic voltammetry of 8 demonstrates an unexpected shift for the ferrocene oxidation to more negative potentials. The origins of this finding lie in the fact that 8, unlike 9, can undergo loss of two amido protons upon binding Ni2+ in aqueous solution, thus yielding a neutral square planar complex. In contrast, 9 retains the full charge of the Ni2+ ion upon binding it, shifting its ferrocene oxidation to more positive potentials.
0
8
9
Other sigruficant results involving ferrocenyl macrocycles include the cation transport, via an electrochemically switchable ferrocenyl crown ether, across a CHzCIz bulk liquid membrane, reported by Saji and Kinoshita in 1986.1'31
10.I Electroclzemically-SwitchedCation-Binding
119
Transport was found to be enhanced upon oxidation of the ferrocene, due to the destabilization of the Na+ complex in the oxidized state. This favors release of the Na' ion to the receiving phase.
10.1.4 Calixarenes Calixarenes provide an attractive framework for ion binding ligands. Synthesis of the parent caluc[4] and calix[b]arenes is facile, and the potential for preorganization of the binding site is high, especially in calix[4]arenes. While a large number of calixarenes have been employed as cation binding ligand~['~], only a fraction have possessed electrochemical switching capacity. In this area the two most commonly observed electroactive moieties are quinones and nitroaromatic residues. One limitation of the electroactive nitrocalixarenes has been their poor solubility in solvents compatible with ion binding studies.[151 Calixquinones have shown a greater degree of promise in the area of switching. Yet even these systems have shown that reduction of multiple quinone sites in the presence of alkali metal ions can result in film formation of the electrode surface.Il61Among the most promising of the calixquinones are those structures bearing one or two quinones in the calix[4] or calix[6]arene framework. An additional consideration, discussed further below, is that different conformations of a calixarene may have different binding affinities for a cation. In conformationally mobile structures, or even occasionally in those structures that appear locked at room temperature, the presence of cations may alter the conformational distribution.[17-1*1 Structures 10 - 12 are several calix[4]quinones that have demonstrated various binding affinities for Na' and Ag+.[W Fig. 10.3 shows an Osteryoung square wave voltammogram of the first reduction of monoquinone 10 in the presence and absence of Na'. In the absence of the salt, this reduction occurs at -0.828 V vs. Ag/AgCl, while in the presence of a two-fold excess of the ion, the reduction of the lO-Na+ occurs at -0.457 V. This potential shift equates
10
11
12
120
10 Electroclu?nzically Switcllable Cation and Anion Binding
Potential/V
Figure 10.3 Osteryoung square wave voltammogram of 0.5 m M 10 (CH3CN/TBAPF6) .in the absence (solid line) and presence (dashed line) of 1equlivalentof Na+.The wane rate was 375 mV/s.[151 Reprinted with permission of the author.
to a binding enhancement on the order of 106. The binding constant of the neutral calixarene 10 was determined by 23Na-NMRto be 186 M-1,P51and that of the reduced species 10-Na' was estimated by digital simulations to be on the order of 108 M-1. Calixarenes 11 and 12 are two conformations of the same diquinone structure, termed syn and anti. (The specific conformations relative to typical calixarene terminology are not appropriate since the quinone moieties rotate rapidly at room temperature.1151) The anti structure 11 was found to have moderate binding capacity. Separate waves are observed for the first reduction of each of the quinone units, however these waves are collapsed into a single wave in the presence of Na+.WI The modest binding enhancement (with respect to the first reduction) observed in this system is on the order of 34 . The binding constant for the neutral ligand was determined by ZNa-NMR to be less than 50 M-1.[151 In contrast to the behavior of 11, syn conformer 12 appears to have a sigruficantly higher binding affinity in both it's neutral and reduced states. The binding constant of neutral 12 was estimated to be 856 M-1 by 1H and Z3NaNMR. This no doubt reflects its better preorganization for ion binding.[*7-181The reduction potential of the free ligand 12 is -0.601 V vs. Ag/AgCl, while that of reduced 12- is -0.480 V yielding a binding enhancement of lo3, suggesting a binding constant of 105 M-1 according to digital simulations. Fig. 10.4 shows the differential pulse voltammogram for a mixed solution of 11 and 12 in the presence and absence of Na+. This solution was generated by heating 11 over several hours in order to generate 12. One caveat in examining this ethoxytailed diquinonecalix[4]arene is that initial studies of the neutral calixarene in the absence of salt suggest the calixarene is conformationally locked at room temperature.[1&191 In fact, in the presence of Na+, 11 undergoes a slow conversion to 12 a room temperature and even at temperatures as low as 5 "C, a point of concern for analysis of binding studies. This point has not been directly
121
10.1
Y.0
*
ii
+32-
1 1
I , I .
'* I I
: I
22.4-
I
,
I
1
*
I
I
'
1 I
, ,
! '' I
Figure 10.4: Differential pulse voltammogram of a mixed solution (CH3CN/TBAPFb) of 11 and 12, two conformational isomers, in the presence (dashed line) and absence (solid line) of Na'. The lower intensity peaks (-0.48 V and -0.60 V) correspond tol2-Na+ and 12, respectively. Peaks at -0.61V and -0.70 V correspond to 11 and 11-Na+. Scan rate was 100 mV/s. Potentials vs. Ag/AgC1."51 Reprinted with permission of the author.
noted as a sigdicant factor to consider when interpreting the electrochemistry of the free and bound calixarenes, but the finding that different conformations[15, 191 may well have different reduction potentials does make conformational change an important consideration when determining binding constants of calixarenes by electrochemical methods. Care must be taken to assure that any potential shift used to estimate binding or binding enhancements does not include a component due to a concomitant change i n the conformation of the calixarene. A nice study of a series of diquinonecalix[4]arenes and their voltammetry has been reported by Ungaro, Casnati and coworkers.[l91 These authors reported a slight difference in the first and a larger difference in the second reduction wave of a pair of conformers, the first observation of such conformational effects in calixquinones. Several other diquinonecalix[4]arenes have been studied for their cation binding ability. The alkaline earth and ammonium ions have been noted to have an effect on the voltammetric behavior of these systerns.[20-211 With binding constants that are on the order of those exhibited by the crown ethers, the first electron reduction potentials of these systems is found to exhibit positive potential shifts from 90 mV to as much as 600 mV in the presence of a bound ion.[20-211 The calix[b]quinones have also been studied. Nam and coworkers have recently reported binding results of a 1,4-diquinonecalix[6]arene with ester tails.IzJ Alkyl ester tails have been employed for cation binding by many
122
10
Elecfroc/tPrnicullySwifcliuble Cation and Anion Binding
researchers in nonelectroactive calixarenes. Structure 13 binds Cs' in the lower rim ester region, resulting in a 0.17 V shift in the reduction potential for 13-Cs'. Interestingly, when t-butyl groups occupy the upper rim position this shift increases to 0.23 V, suggesting that preorganization is better in the more sterically hindered host.
I
13
10.1.5 Conclusions about Cation Binding
To some extent, work in the field of cation binding has slowed in the past several years, no doubt in part due to the present interest in the recognition of molecular guests. This area is still of interest however, as the binding of heavy transition metal ions, for separations, recovery or detection purposes, is still an important environmental concern. Challenges still can be found in this field. True binding selectivity coupled with electrochemical switching enhancement is an issue that has yet to be widely addressed and remains a goal for future work.
10.2 Electrochemically-SwitchedAnion Binding The area of switchable anion-binding has been less explored than that of cation binding. The first report to examine anion binding was that of Wrighton and coworkers, who in 1982 reported the electrostatic interaction of anions with and electroactive polymerP1 The first macrocycle with anion recognition capability of a redox-active guest was reported in 1983 by Lehn and coworkers. They examined a polyaminemacrocycle, 14, that binds the comparatively small ferrocyanide ionP1 In this system the oxidation of ferrocyanide to ferricyanide is shifted to more positive potentials upon binding to 14. At less than one equivalent of the polyamine ligand, two redox waves are observed. Following addition of 1.2 equivalents of the macrocycle, a single wave is observed, see Figure 10.5. The wave shows noticeably lower current intensity than that
2 0.2
E/ectroclremica//?l-SwitclledAnion-Binding
123
14 corresponding to the uncomplexed ferrocyanide.[*4] Occasionally, lower peak currents are observed for a complexed species in comparison with those of the free species. This effect is typically noted when a small guest, rather than the (typically) larger host, is the redox active species. For redox couples with reversible heterogeneous kinetics, the lower currents observed for the complexed species are attributable to the lower diffusion coefficient of the complex. Differences between the diffusion coefficients of the complexed and free species have been observed in the electrochemistry of other supramolecular
0.o
i'
0.5
Figure 10.5:Electrochemistry of ferrocyanide (0.9 mM in 0.1M KCL, pH= 5.5) in presence of increasing amounts of 14. Scan rate= 50 mV/s; R= [2]/[Fe(CN)&].1241 Reprinted with permission from Elsevier Science, SA.
10 Electrocllemically Switchable Cation and Anion Binding
124
systems such as cyclodextrin inclusion complexes.[251For example, a decrease in the diffusion coefficient of ferrocyanide was detected upon complexation by an amino substituted P-cyclodextrin host.[W In these cases, simulations of the voltammetric behavior must take into account the difference in diffusion coefficients in order to accurately reproduce the observed behavior, which can include shifts in the observed redox potentials. Simulation may also be employed to extract binding constant information based on changes in the diffusion coefficient.[*5”1 Throughout the 1980’s and early 1990’s several other reports in the area of anion recognition were publishedP1 Since then, Beer and coworkers have dominated the field and pioneered the structural design of electrochemicallyswitchable anion receptors.[27l Their early work in the area employed multiple positively charged redox moieties, which were switched electrochemically to higher charge states. With this type of system an excess of the anion guest had to be added in order to see measurable changes. The effects of binding were rather modest, on the order of 50 mV or less, indicating very low binding enhancements (< 10). More recently however, the authors have had some success with dihydrogenphosphate binding exhibited by a calixarene-based receptor with a pendant Ru(bpy)s moiety, as shown in structure 1 4 P I This receptor shows selectivity for H2P04- anions. In the presence of an equimolar amount of this anion, the reduction potential for the amide-linked bpy unit (-1.40 V vs. Ag/Ag+) is shifted by 175 mV to more negative potentials. Similar results are obtained even in the presence of a 10-fold excess of HS04- and C1-,
I
H I
14
10.3
RefPrences
125
both of which had relatively low binding affinity for the receptor. This result suggests a clear preference for the target anion. The binding site of the anion is the lower rim of the calixarene, presumably due to interaction with the amide protons. X-ray diffraction studies of the complex appear to confirm this. The binding constant for 14 was estimated by 1H-NMR to be 2.8 x 104 M-1. Sessler and coworkers have recently published work on the anion binding of several calix[4]pyrroles with a pendant ferrocene.1291 Binding studies in the presence of F-, C1-, and H2P04- were reported. Work in the area of anion binding will no doubt continue to progress. Since the analytical determination of anions is generally more difficult than that of cations, this area of research shows promise.
10.3 References 1. For a review of the broad range of techniques employed for switching of cationic receptors see: S. Shinkai in Comprehensive Supramolecular Chemistry, Vol. 1, (Eds.: G. W. Gokel), Pergamon, Tarrytown, NY, 1996 2. [a] A. E. Kaifer in Comprehensive SupramolecuZar Chemistry, Vol. 1, (Eds.: G. W. Gokel), Pergamon, Tarrytown, NY, 1996; [b] A. E. Kaifer, L. Echegoyen, in Cation Binding by Macrocycles, (Eds. Y. Inoue, G. W. Gokel) Dekker, New York, 1990; [c] P. D. Beer,, Chem. SOC.Rev.,1989,18,409-450. 3. A. E. Kaifer, L. Echegoyen, D. A. Gustowski, D. M. Goli, G. W. Gokel, J. Am. Chem. SOC.,1983, 205,7168-7169. 4. [a] S. R. Miller, D. A. Gustowski, Z. H. Chen, G. W. Gokel, L. Echegoyen, A. E. Kaifer, Anal. Chem., 1988, 60, 2019-2022. [b] A. E. Kaifer, L. Echegoyen, D. A. Gustowski, V. J. Gatto, R. A. Schultz, T. P. Cleary, C. R. Morgan. A. M. Rios, G. W. Gokel, J. Am. Chem. SOC.,1985, 107,1958-1965. 5. M. Delgado, L. Echegoyen, V. J. Gatto, D. A. Gustowski, G W. Gokel, J. Am. Chem. SOC.,1986, 108,4135-4138, 6. H. Bock, B. Bierholz, F. Vogtle, G. Hollman, Angew. Chem. Inf. Ed. EngZ., 1984, 23,57-56 . 7. L. E. Echegoyen, H. K. Yoo, V. J. Gatto, L. Echegoyen, G. W. Gokel, J. Am. Chem. SOC.,1989, 111,2440-2443. 8. L. Echegoyen, Y . Hafez, R.C. Lawson, J. de Mendoza, T. Torres, J. Org. Chem. 1993,58,2009-2012. 9. [a] D.A. Gustowski, M. Delgado, V.J. Gatto, L. Echegoyen, G.W. Gokel, J. Am. Chem. SOC.1986,308,7553-7560; [b] M. Delgado, D.A. Gustowski, H.K. Yoo, V.J. Gatto, G.W. Gokel, L. Echegoyen, J. Am. Chem. SOC.1988,110,119-124. 10. Z. Chen, 0. F. Schall, M. Alcalh, Y. Li, G.W. Gokel, L. Echegoyen, J. Am. Chem. SOC.1992, 114,444-451. 11. L. Echegoyen, Y. Hafez, R.C. Lawson, J. de Mendoza, T. Torres, Tetrahedron Lett. 1994, 35, 6383-6386; b) L. Echegoyen, R.C. Lawson, C. Lopez, J. d e Mendoza, Y. Hafez, T. Torres, 1. Org. Chem. 1994,59,3814-3820; 12. G. De Santis, L. Fabrizzi, M. Licchelli, P. Pallavicini, A. Perotti, J. Chem. SOC., Dalton, 1992,3283-3284. 13. T. Saji, 1. Kinoshita, J. Chem. SOC., Chem. Commun., 1986,716-717.
126
10 Electrochemically Switchable Cation and Anion Binding
14. M. A. McKervey, M.-J. Schwing-Weill, F. Arnaud-Neu, in Comprehensive Supramolecular Chemistry, Vol. 1, (Eds.: G. W. Gokel), Pergamon, Tarrytown, NY, 1996. 15. M. G6mez-Kaifer, Ph. D. Dissertation, University of Miami, FL, 1997. 16. M. Gomez-Kaifer, P. A. Reddy, C. D. Gutsche, L. Echegoyen, 1994,116,35803587. 17. J. Blixt, C. Detellier, J. Am. Chem. SOC.1995,117,8536-8540. 18. M. Gomez-Kaifer, P. A. Redily, C. D. Gutsche, L. Echegoyen, 1997,119,52225229. 19. A.Casnati, E. Comelli, M. Fabbi, V. Bocchi, G. Mori, F. Ugozzoli, A.M. Manotti Lafredi, A. Pochini, R. Ungaro, Red. Trav. Chim. Pays-Bas 1993, 112, 384-392. 20. P. D. Beer, Z. Chen, P. A. Gale, Tetrahedron, 1994,50,931-940. 21. T. D. Chung, D. Choi, S. K. Kang, S. K. Lee, S.-K. Chang, H. Kim, J. Elecfroanal. Chem., 1995,396,431-439. 22. K. C. Nam, S. 0.Kang, H. Lee, S. Jeon, H. J. Cho, S.-K. Chang, Bull kor. Chem. SOC.,1998, 279-281. 23. James A. Bruce, Mark S. Wrighton, J. Am. Chem. Soc., 1982,104,74-82. 24. F. Peter, M. Gross, M. W. Hosseini, J.-M. Lehn, J. Elecfroanal. Chem. 1983,144, 279-292. 25. See for example: a) L.A. Godinez, J. Lin, M. MuAoz, A.W. Coleman, A.E. Kaifer, J. Chem. SOC., Faraday Trans. 1996, 92, 645-650; b) P.M. Bersier, J. Bersier, B. Klingert, Electroanalysis 1991,3,443-455 and references therein. 26. [a] Juan Arago, Andrea Bencini, Antonio Bianchi, Antonio Domenech, Enrique Garcia-Espana, J. Chem. Soc., Dalton Trans., 1992, 319-24; [b] Richard A. Simon, Thomas E. Mallouk, Karen A. Daube, Mark S. Wrighton, Inorg. Chem. 1985,24,3119-26; [c] K. M. Kadish, R. K. Rhodes, Inorg. Chem. 1983, 22, 1090-4. 27. [a] P. D. Beer, S. E. Stokes, Polyhedr. 1995, 14, 2631-2635; [b] P. D. Beer, M.G.B. Drew, D. Hesek, R. Jagessar, J. Chem. SOC.Chem. Comrnun. 1995, 11871189; [c] P. D. Beer, M. G. Drew, C. Hazelwood, D. Husek, J. Hodacova, S. E. Stokes, J. Chem. SOC.Chem. Commun. 1993,229-231. 28. F. Szemes, D. Hesek, Z. Chen, S. W. Dent, M. G. B. Drew, A. J. Goulden, A. R. Graydon, A. Grieve, R. J. Mortimer, T. Wear, J. S. Weightman, P. D. Beer, Inorg. Chem., 1996,35,5868-5879. 29. J. L. Sessler, A. Gebauer, P. A. Gale, Gazz. Chem. Ital., 1997,127, 723-726.
Supramolecular Electrochemistry Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
11 Redox-Switchable Cyclophanes and Other Molecular Receptors
In the past three decades a vast amount of research has been devoted to cyclophane systems. At present, cyclophanes are frequently seen as integral components of supramolecular systems. Due to their aromatic nature, many cyclophanes are electroactive, however until recently, electrochemical studies of redox-active cyclophanes had been comparatively few. Much of the early work in the field centered on small systems in which the aryl units were held within close proximity to one another. These systems provide an opportunity to examine proximity effects on redox behavior and to examine delocalization of electrons in the cyclophane framework. Metal complexes to the 7[: faces in these smaller cyclophane systems, the so-called metallocyclophanes, were also examined. More recent developments in the field have involved the redoxswitchable binding capabilities of the cyclophanes and various molecular guests. Interest in this area has led to studies of redox-switchable catenanes, rotaxanes and molecular shuttles. In this chapter we will examine three general areas relevant to cyclophane electrochemistry: (i) the early cyclophane studies including systems involving metal coordination to cyclophanes e.g. metallocyclophanes, in which the metal ions comprise the redox-active site of interest, (ii) cyclophanes in which the aromatic components are the electroactive site of interest, and whose redox states can influence molecular binding, and (iii) electroinactive cyclophane hosts capable of exhibiting molecular recognition of redox-active guests. Redox-switchable molecular binding by several noncyclophane systems will be examined at the end of the chapter. Metallocyclophanes will be discussed only briefly, as our primary focus will be molecular binding of redox-active cyclophanes. Calix[4]arenes, a special class of cyclophane with redox-switchable cation or anion binding properties, are discussed in Chapter 10. Cyclophanes as components of catenanes, rotaxanes and shuttles or other intertwined structures are discussed in Chapter 12.
11.1 Early Cyclophane Studies and the Metallocyclophanes One goal of early cyclophane research was the examination of systems that might provide insight into biologically relevant structures. To this end, the earliest published work on the electrochemistry of a cyclophane was that of Wasielewski and coworkers in 1978.1’1 These authors reported synthesis and electrochemistry of a bis(chlorophyl1) cyclophane system, 1, shown
128
71
Redox-Switchable Cyclophanes and Other Molecular Receptors
0
kofl Porphyrin face of 1
1
Figure 11.1: Schematic representation of Wasliewki's chlorophyll cyclophane.[~l (schematic structure* reprinted with permission of the American Chemical Society).
schematically in Fig. 31.1. Their intent was examination of a system analogous to chlorophyll special pairs, which are integral to the photosynthetic process. Along with a variety of spectroscopic studies, these authors examined the ac voltammetric behavior of the cyclophane, in comparison with that of the monomeric chlorophyll analogue and noted modest shifts in both the anodic and cathodic redox response. Oxidation of the dimeric cyclophane occurred approximately 70 mV earlier than the monomer, which was attributed to delocalization of unpaired electron density over both macrocycles. The reductions of cyclophane 1 were found to be more difficult than that of the monomeric chlorophyll by 150 mV, leading the authors to postulate that the electron density generated in radical monoanion species remained localized on one of the porphyrin macrocycles. Shortly thereafter, in another important article for the field, Sat0 and Torizuka reported the cyclic voltammetry of a series of simple benzyl and naphthyl para- and metacyclophanes and compared these results with the voltammetry of model compounds and analogue acyclic Such studies permitted insights into the effects of fixing redox centers in close proximity to one another. Additionally, their work contrasted the redox behavior of dimeric systems, both cyclized and linear, with that of equivalent monomers. Among the systems examined was compound 2, which possesses cis and trans conformational isomersP" The authors noted that the anodic peak potentials of the cis isomer (+1.65 V vs. SCE) differed slightly from that of its monomeric analogue. 3 (+1.74 V), in contrast to the voltammetric behavior observed for the trans isomer of cyclophane 2. The first oxidation wave of 2trans occurs 260 mV earlier than that of 2-cis, at +1.39 V. A linear acyclic model,
129
Early Cyclophane Studies and the Metulloajcloplianes
2-t~ans
2
I . 3.0
4
. . . , . 2.0
1.0 V versus 5.c.e.
0
Figure 11.2 Cyclic voltammetric behavior of (a) monomer 3, @) 2-cis, and (c) 2-trans in CH3CN/TBAC104 at 500 mV/s.[*I Reprinted with permission ofthe Royal Chemical Society.
4, showed behavior intermediate between the two isomers of 2, with its oxidation occurring at +1.54 V vs. SCE. Sat0 postulated that the more open conformations of 2-trans and 4 permit greater cation radical stabilization
130
11
Redox-Active Cyclophanes as Molecular Receptors
between the two naphthalene groups. The difference in the voltammetric behavior observed for the isomers of 2 is striking, as the trans isomer undergoes two oxidation processes in the range examined, see Fig. 11-2. At least minor shifts in the oxidation potentials, relative to the monomeric and acyclic systems, were observed in all cases: cyclophanes were more readily oxidized, in some instances by as much as 1 V. Oxidation in these systems was virtually always irreversible. Studies on metacyclophanes, with various substituents,[3] and of larger [2n]paracyclophane systems[4] followed. The effect of face-to-face 'ilbonded systems has continued to fascinate chemists. In one interesting report Paquette and coworkers have prepared a cyclooctatetraenophanesystem. Anion species ranging from the monoanion to the tetranion were studied by low temperature voltammetry.[51 Recently, Sonnenschein et al. have delved further into conformational effects of redox-active cyclophanes with their studies on redox-switchable 3,3'-biindolizine-based cyclophanes.~6lOne of these authors' systems is discussed briefly below. Interest in metallocyclophanes began with the iron and ruthenium complexes of [2n]cyclophanes and their electrochemistry in work of Boekelheide, et al.17 Complexes with other transition metals have also been The group of Boekelheide has also been active in the study of cyclophanes as components of oligomeric metallocene structures, such as structure 5,[91 an area that led to studies of the formation redox-active polymer films on electrode surfaces.l101 More recently, Okuno and coworkers have studied iron-sulfur clusters in a cyclophane cage,["] and Hassan, et al. have studied a diamagnetic diiron-[2.2]paracyclophane cornplex.[121 The majority of this work has been classically inorganic in its orientation. In the area conventionally thought of as supramolecular, the metal complexes of porphyrinbased cyclophanes have been a topic of interest, with work represented by Staab, and coworkers[~31and Diederich and coworkers.[14] More recently however, metallocyclophanes have been extended to the formation of catenanes and complex "knotted systems, a topic covered in Chapter 12.
5
11.2 Redox-Active Cyclophanes as Molecular Receptors Within the past decade much attention has centered on the molecular binding capabilities of redox-switchable cyclophanes. Diederich and coworkers
11.2
Redox-Active Cyclophanes as Molecular Receptors
External Association
I
I
131
Internal Association
. 2e
- 2 e-
.o,c
c 0,-
Figure 11.3:The isoalloxazine unit of flavinophane Gundergoes structural changes from a planar to a butterfly conformation, opening the cyclophane for internal binding."51
were the first to report a redox-switchable cyclophane system."5] A flavin cyclophane, or flavinophane (6) displayed switchable binding activity due to a change from planarity to a "butterfly," or bent shape, noted in its isoalloxazine unit after two electron reduction, as shown in Fig. 11.3. The reduced form of compound 6 was noted to bind napthalene derivatives by inclusion formation, while the oxidized form displays noncavity binding (external association) of the same order of magnitude. This system showed quasi-reversible heterogeneous electron transfer kinetics and possessed a single half-wave potential of -0.581 V vs. Ag/AgCl in an aqueous borate buffer at pH lO.Ll51 Kuroda and coworkers have recently reported cyclophanes composed of two and three NAD' analogues in their main framework.[161 Kreher and Sonnenschein have reported that cyclophanes 7a and 7b (prepared as an inseparable 1:6 diastereomeric mixture) interact with 2,4,7trinitrofluoren-9-one (TNF).@l This interaction, as detected by cyclic
0
TNF
132
11
Redox-Active Cycloplzanesas Molecular Receptors
voltammetric response, was not clearly attributable to internal (cavity) binding. The authors noted substantial changes in the appearance of the cyclic voltammograms, but no effect on the observed half-wave potentials for the host. X-ray diffraction and molecular modeling data suggest that TNF is too large to be included by the host.[bal Thus, the voltammetric results may be attributable to external association, in a manner similar to that observed for the unreduced form of flavinophane 6.
11.3 Viologen Based Cyclophanes- The Ideal .n-Acceptor Host Perhaps the most studied cyclophane of the last decade is the bis-viologen cyclophane The electrochemistry of this compound was first described as a component of a catenane system in 1989[181 and later studied as an effective redox switchable host for electron rich aromatic amino acids"9"I and neurotransmittersW" 8 has also been studied as a component of both rotaxane[*o]and molecular shuttle systems.[21] The electrochemistry of related viologen based cyclophanes with the xylylene bridges attached at the ortho or meta positions has also been describedP1 Cyclophane 8 shows reversible redox processes that are characteristic of the parent methyl viologen unit, i.e. two oneelectron redox waves are observed for each viologen unit. The +4 charge on the r
8
1
I
0 :0
-1.2
E (V vs SSCE)
Figure 11.4: Cyclic voltammetric response of a 0.5 mM solution of 8 in CH3CN/TBAl'F6 obtained at 25 oC, at a scan rate of 50 mV/s.Ilsl Reprinted with permission of the American Chemical Society.
11.3
Viologen Based Cyclophanes- The [deal n-Acceptor Host
133
cyclophane makes it an excellent host for electron rich aromatic guests, while it's reduced forms (the +2 or neutral states) show a substantially decreased nacceptor character. The cyclic voltammetric behavior of 8, shown in Fig. 11.4, displays two redox waves.[l*l The first two electron wave corresponds to the monoelectronic reduction of each viologen moiety to its cation radical state at 0.279 V vs. SSCE in CH3CN/TBAPF6, while the second two electron wave corresponds to the monoelectronic reduction to the neutral viologens at -0.703 V. As mentioned above, cyclophane 8 has been used as a redox switchable receptor for both aromatic amino acids[lga1 and catechol-type neurotransmitters.[*9bl The electron accepting character of 8, along with its ideal separation (approximately 3.5 A, or the thickness of a x-ring) between viologen units makes it an excellent host for aromatic guests. While the solubility of 8 in organic solvents is high in even its fully reduced, neutral state, its solubility in aqueous solution in the reduced state is quite poor resulting in precipitation onto the electrode surface. Thus, the study of its binding to the neurotransmitters was initially carried out using modified electrodes in order to permit detection in aqueous solution. The electrochemistry of cyclophane 8 was studied in Nafion films cast onto glassy carbon electrodes and immersed in aqueous solutions of dopamine, norepinephrine, epinephrine, catechol, serotonin or indole. Differential pulse voltammetry of 8 was useful in the detection of binding to catechol and indole and their related neurotransmitters in the micromolar concentration range. The half-wave potential shifts exhibited by the host-guest complex were modest (10- 70 mV) but well within analytical detection range. Only the first redox couple was useful for this analysis, due to the tendency of the neutral viologen to adsorb onto the electrode surface. The limited dynamic range and analytical sensitivity of this system lead to the design of a cyclophane, 9, which could be chemisorbed onto a gold electrode surface via disulfide/ thiol chemistry. This is shown schematically in Fig. 11.5. The surfaceattached cyclophane demonstrated surface coverages of the bound catechol or indole guests in the range of 3 x 10-11mol/cm2. Potential shifts upon binding the guest were similar to those obtained with host 8 in Nafion films.
S @N.
9
134
11
-
Monolayer
Redox-Active Cyclophnnes as Molecular Receptors
-
Solution
+
Figure 11.5: An idealized scheme for detection of catechol by surface attached host Reprinted with permission of the American Chemical Society.
9J31
The molecular recognition capacity of 8 has been used by several authors to create self-complexing systems. These molecules have been termed electrochemically controlled molecular switches.[W One such system, 10, threads a dioxynaphthalene through the cyclophane. In this system the first two-electron reduction occurs at -0.35 V vs. SCE, shifted some 60 mV from that of cyclophane 8, indicating that the n-electron donating dioxynaphthalene moiety stabilizes the unreduced cyclophane. The fact that the second twoelectron reduction occurs at the same potential as the free cyclophane suggests that no donor-acceptor interaction exists, i.e. expulsion of the electron donating thread has occurred.
4PF;
I I .4 Electroinactive Cycloplme Hosts and Their Binding of Redox-Switchable Guests
135
Becher and coworkers have designed similar systems with cyclophane 8 bearing TTF tethers of two different lengths.[261 Structure 11 shows one such molecule. A strong color difference was noted between the self-complexed and decomplexed molecule. The reversibility of this complexation/ decomplexation process was noted to be dependent on the tether length. One concern in selfcomplexing systems is the fact that both inter- and intramolecular complexation can occur. Voltammetric studies have not yet been reported for this system, however Becher has reported results for a sandwich type complex, 12, that is conceptually related to self-complexed 11. The two oxidation waves for the TTF moiety occur at 0.64 and 0.94 V vs. Ag/AgCl, anodically shifted about 100 mV from the free TTF precursorP1 The authors postulate that electrostatic repulsion contributes to the more difficult oxidation of this sandwich complex. Spectral evidence suggests that the TTF is able to interact strongly with only one of the two viologen acceptor units due to conformational constraints.[Zq
I
d
1
N
/”+ 11
+
4PF,-
+
rNL
+N
--N3 X
N +
12
11.4 Electroinactive Cyclophane Hosts and Their Binding of Redox-Switchable Guests Among electroinactive cyclophanes perhaps the most common hosts have been the calixarenes. While calix[4]arenes have generally been employed in ion binding work, calix[5], calix[6] and calix[8]arenes have all been employed in electrochemical studies of various redox active guests. One particularly common such host has been calix[6]arenesulfonate 13 or its 0-methylated analogue 14. Both calixarenes bear high negative charge (8- and 6-, respectively) and hence have proven to be good hosts for small positively charged guests. Studies of ferrocene and cobaltocenium complexes with the two calix[6]arenesulfonates have recently been reportedP1 The studies of the cobaltocenium complexes with 13 and 14 are interesting in that 2:l complexes are formed. The double partial cone conformation of calixarene 13 permits one
136
Redox-Active Cyclophanes as Molecular Receptors
11
138-
146-
P
16
-1
-
-3
-
-5
-
4 -7
.
I
I
.-
-0.5
2 2
-0.7
I
I
I
I
-0.9
-1.1
-1.5
-1.3
(4
1-
0-1
-2
-
'
I
4.3
I
-0.5
I
I
I
-0.7 -0.9 -1.1 E I V vs. Ag I AgCl
I
-1.3
I
-1.5
Figure 11.6: Cyclic voltammograms of a) 15 in the absence (solid line) and presence of 4mM 13 (- - - ) and both 4 mM 13 and 4 m k P-cyclodextrin. b) same conditions for 16. Voltammetry was performed in 0.1 M phosphate buffer (pH = 7.0) at a scan rate of 100 mV/s.[2Bal Reprinted with permission ofthe Royal Society of Chemistry.
11.5
Otlzer Molecular Receptors
137
cobaltocenium 15 to bind on each of the upper and lower rims, as evidenced by NMR studies.[Bal In the absence of 13, reduction of 15 to the neutral cobaltocene results in the adsorption of this more hydrophobic species onto the electrode surface, as shown in solid line voltammogram in Fig. 11.6a. In the absence of a host that can adequately solubilize neutral 15, this guest is better studied by normal pulse voltammetry. In the presence of 13, 15 becomes more difficult to reduce, indicative of its stabhation by the negatively charged 13 (dashed line in Figure 11.6a). Distortion of the voltammogram’s appearance indicates that the adsorption of the neutral cobaltocene persists, since 13 does not interact with the neutral form of 15. In a variation on competitive binding studies, 13 which can form the 1:2 complex with 15, and j3-cyclodextrin, which can form a 1:lcomplex with the more hydrophobic cobaltocene, were added to sohtions of 15. Under these conditions, both redox states of guest 15 are stabilized by a host. The resulting voltammetric behavior is greatly improved. The slower overall rate of diffusion for the complexes contributes to the lower currents observed. In order to gain better solubility for the reduced cobaltocene, cobaltoceniumcarboxylate (16)was employed. Due to its net negative charge at pH 7,16 does not adsorb onto the electrode surface upon reduction, as does 15. The results of the mixed host study with guest 16 are shown in Fig. 11.6b. Host-guest interaction is still evident with this guest. Voltammetric studies were also performed with calixarene 14, which does not adopt a conformation well suited to binding. Little if any potential shifts were observed for complexes of this host and 15 or 16.[28bI Complexes between calixarenes and fullerenes have been reported by a number of authors. While the fullerenes field has developed rapidly, the low solubility of the fullerenes has proved challenging for the electrochemist. Though a number of complexes have been examined by spectroscopic methods, few have been analyzed for their electrochemical behavior. Beer and coworkers reported electrochemical studies of C ~ and O a series of calix[4], calix[5] and calix[€i]arenes in a mixture of toluene/acetonitrile or toluene/DMSO and have noted adsorption on the electrode surface accompanied by negative potential shifts in what appear to be the redox waves for the Go-calixarene complexes.~331
11.5 Other Molecular Receptors Cyclophanes have not been the sole framework upon which redox-switchable molecular receptors have been based. Two recent examples from the current literature provide examples of noncyclophane based systems. Rotello and coworkers have reported a three component molecular switch.[351 Naphthalimide is the redox active guest in their system, while an anthracene and an acylated diaminopyridine are the hosts for the naphthalimide and its reduced state. Fig. 11.7 shows the structures and the scheme for this system. The anthracene host 18 interacts with the naphthalimide guest 17 via both hydrogen bonding and x-n stacking interactions, while the diaminopyridine host 19 interacts solely via hydrogen bonding interactions. While 18 interacts
11
138
Redox-Active Cyclophmes as Molecular Receptors
with both the oxidized and reduced forms of 17 with binding constants of the same order of magnitude, the anion radical 17- interacts far more strongly with host 19, which in turn shows only negligible interaction with the oxidized form, 17. The reduction potential of 17 is shifted almost 150 mV more positive in the presence of host 19, indicating that the host favors reduction of 17 to its radical anion. In contrast, reduction of 17 results in destabilization of the 7c-n stacking interaction between 17 and the anthracene moiety of 18. The authors also report ESR studies of this system.
Guest 17
Host 18
n
0-
(17-)
(19) D
-8
U
u
Figure 11.7 A three component molecular switch and a scheme illustrating the principles behind the electrochemical switching of the system. (Adapted from Rotello, e t al, Reference 35, used with permission of the American Chemical Society.)
Another recent and elegant work reported by Kochi, et al. deals with an electrochemically switchable system for detection of NOJ361 The so-called "Venus flytrap" ligand 20 is shown below. The pentamethylphenyl moieties are forced into close proximity of one another by the rigid bicyclooctene unit. Oxidation of 20 yields a relatively stable cation radical which turns bright blue upon exposure to gaseous nitric oxide. An impressive decrease in the oxidation potential for 20 is observed in the presence of NO. The half-wave potential
11.6
Conclusions
139
20 shifts from 1.49 V vs. SCE in the absence of NO to 0.58 V vs. SCE in the presence of NO. The cation radical of 20 is capable of complete uptake of nitric oxide and the estimated binding constant for the cation radical complex is greater than 3 x 106 M-1. The complex is stable enough to permit crystallization at -23 oC. Reduction of the [20-NO]+complex restores the free NO to solution. The authors suggest that 20 provides a molecular pumping system for N0.P61
11.6 Conclusions Redox-switchable non-cyclophane systems show promise in the area of molecular recognition and will no doubt continue to be developed. Clearly systems with selectivity require thoughtful design and will employ multiple non-covalent interactions which are influenced by the electron density of the switchable host or guest. The work of Rotello and Kochi point to the promise of further work in the field. In the area of cyclophane research, the emphasis in the last decade has clearly shifted beyond preparation of novel structures toward the goal of making cyclophanes as part of functional assemblies. Whether as catalytic sites1371 or as components of complex intertwined structures, cyclophanes are now among the most prevalent supramolecular building blocks. Research that has gone into understanding the basic redox behavior of these systems has provided the foundation for controllable supramolecular assemblies.
11.7 References 1. M. R. Wasielewski, W. A. Svec, B. T. Cope, 1. Am. Chem. SOC.1978,100, 19611962. 2. [a]T. Sato, K. Torizuka, J. Chem. SOC., Perkin Trnns. 11, 1978, 1199-1204; [b] T. Sato, H. Matsui, R. Komaki, J. Chem. SOC.Perkin 1,1976,2051-2053.
140
11 Redox-Active Cycloplianes as Molecular Receptors
3. T. Sato, K. Torizuka, R. Komaki, H. Atobe, 1. Chem. SOC., Perkin Trans. 11, 1980, 561-568. 4. B. Thulin, J. Chem. SOC., Perkin Trans. 11,1981,664-667. 5. L. A. Paquette, M. A. Kesselmayer, G. E. Underiner, S. D. House, R. D. Rogers, K. Meerholz, J. Heinze, 1. Am. Chem. SOC.,1992,114,2644-2652. 6. [a] T. Kreher, H. Sonnenschein, 8.Costisella, M. Schneider, 1. Chem. Soc., Perkin 1,1997,3451-3457; [b] M. B. Leitner, T. Kreher, H. Sonnenschein, B. Costisella, J. Springer, 1. Chem. SOC.,Perkin 2,1997,377-381; [c] H. Sonnenschein, T. Kreher, E. Grundemann, R.-P. Kriiger, A. Kunath, V. Zabel, J. Org. Chem., 1996,61,710714. 7. [a] E. V. Laganis, R. H. Voegeli, R. T. Sann, R. G . Finke, H. Hopf, V. Boekelheide, Organornetallics, 1981,1,1415-20; [b] W. J. Bowyer, W. E. Geiger, V. Boekelheide, Organomefallics,1984,3, 1079-1086 8. [a] P. Jutzi, U. Siemeling, A. Mueller, H. Boegge, Organometullics, 1989,8, 17441750; [b] C . Elschenbroich, J. Schneider, M. Wuensch, J. L. Pierre, P. Baret, P. Chautemps, Chem. Ber. 1988,121,177-183. 9. R. T. Swann, A. W. Hanson, V. Boekelheide, J. Am. Chem. SOC.,1986,108,33243334. 10.8. Gollas, B. Speiser, J. Sieglen, J. Strahle, Organornefallics,1986, 25,260-271. 11.T. Tomohiro, H. (Y.) Okuno, lnorg. Chim. Acfa, 1993,204,147-152. 12. R. Hassan, M. Lacoste, M.-H. Delville-Desboise, J. Ruiz, B. Gloaguen, N. Ardoin, D. Astruc, A. L. Beuze, J.-Y. Saillard, et al., Organometullics, 1995, 14, 5078-5092. 13. H. A. Staab, G. Voit, J. Weiser, M. Futscher, Chem. Ber., 1992,125,2303-2310. 14. D. R. Benson, R. Valentekovich, S.-W. Tam, F. Diederich, Helv. Chim. Acfa, 1993, 76,2034-2060. 15. E. M. Sward, R. B. Hopkins, W. Sauerer, S.-W. Tam, F. Diederich, J. Am. Chem. SOC., 1990,Z 12,1783-1790. 16. Y. Kuroda, H. Seshimo, T. Kondo, M. Shiba, H. Ogoshi, Tetrahedron Left., 1997,38,3939-3942. 17. [a] R. R. Lilienthal, D. K. Smith, Anal. Chem., 1995, 67, 3733-9; [b] E. A. Smith, R. R. Lilienthal, R. J. Fonseca, D. K. Smith, Anal. Chem., 1994, 66, 30133020; [c] P. R.Ashton, C. L.Brown, E. J. T.Chrysta1, T. T. Goodnow, A. E. Kaifer, K. P. Parry, D. Philp, A. M. Z. Slawin, N. Spencer, J. F. Stoddart, J. Chem. SOC., Chem. Commun., 1991, 634-9; [d] R. J. Fonseca, J. T. Colina, D. K. Smith, J. Elecfroanal. Chem., 1992,340,341-8, among many others. 18. P. R. Ashton, T. T. Goodnow, A. E. Kaifer, M. V. Reddington, A. M. Z. Slawin, N.Spencer, J. F. Stoddart, Angew. Chem. lnf. Ed. Engl., 1989, 34, 13961399. 19. [a] T. T. Goodnow, M. V. Reddington, J. F. Stoddart, A. E. Kaifer, J. Am. Chem. SOC.,1991,113,43354337. [b] A. R. Bernardo, J. F. Stoddart, A. E. Kaifer, J. Am. Chem. SOC., 1992,114,10624-10631. 20. R. A. Bissell, E. Cordova, A. E. Kaifer, J. Org. Chem. 1994,100,42484254. 21. [a] R. A. Bissell, E. Cordova, J. F. Stoddart, A. E. Kaifer, Nature 1994,369,133137; [b] A. C. Benniston, A. Harriman, V. M. Lynch, J. Am. Chem. SOC., 1995, 117, 5275-5291; [c] A. C. Benniston, A. Harriman, Angew. Chem. Int. Ed. Engl., 1993,32,1459-1461; [d] P. R. Ashton, R.A. Bissell, N. Spencer, J. F. Stoddart, M.
11.7 References
141
S. Tolley, Syn. Lett., 1992, 923-926; [el F. Vogtle, W. M. Miiller, U. Miiller, M. Bauer, K. Rissanen, Angew. Chem. Int. Ed. Engl., 1993,32, 1295. 22. W. Geuder, S. Hiinig, A. Suchy, Tetrahedron, 1986,42,1665-1677. 23. M. T. Rojas, A. E. Kaifer, 1.Am. Chem. SOC.,1995,117,5883-5884. 24. V. Balzani, M. G6mez-Lopez, J. F. Stoddart, Acc. Chem. Res., 1998,405-414. 25. P. R. Ashton, R. BalIardini, V. Balzani, S. E. Boyd, A. Credi, M. T. Gandolfi, M. Gomez-Lopez, S. Iqbal, D. Philp, J. A. Preece, L. Prodi, H. G. Ricketts, J. F. Stoddart, M. S. Tolley, A. P. J. White, D. J. Williams, Chem. Eur. J., 1997, 3,10521058. 26. M. Brerndsted Nielson, S. Brerndsted Nielson, J. Becher, Chem. Commun., 1998, 475-476. 27. K. B. Simonsen, N. Thorup, M. P. Cava, J. Becher, Chem. Commun., 1998,901902. 28. [a] Y. Wang, J. Alvarez, A. E. Kaifer, Chem. Commun., 1998, 1457-1458; [b] J. Alvarez, Y. Wang, M. Gomez-Kaifer, Chem. Commun., 1998,1455-1456. 29. S. Shinkai, J. Chem. SOC.,Chem. Comm. 1994,22,2587-2588. 30. R. M. Williams, J. W. Verhoeven, Recl. Trav. Chim. Pay-Bas 1992,111,531-532. 31. J. L. Atwood, G. A. Koustantonis, C. L. Raston, Nature 1994,368,229-231. 32. T. Suzuki, K. Nakashima, S. Shinkai, Chem. Lett. 1994,699-702. 33. Z . Chen, J. M. Fox, P. A. Gale, A. J. Pilgrim, P.D. Beer, M. J. Rosseinsky, J. Electroanal. Chem. 1995,392,101 - 105. 34. R. Castillo, S. Ramos, R. Cruz, M. Martinez, F. Lara, J. Ruiz-Garcia, J. Pkys. Chem. 1996,100,709-713. 35. R. Deans, A. Niemz, E. C. Breinlinger, V. M. Rotello, 1.Am. Ckem. SOC., 1997, 119,10863-10864. 36. R. Rathore, S. V. Lindeman, J. K. Kochi, Angew. Chem. Int. Ed. Engl., 1998,37, 1585-1587. 37. F. Diederich has a long running series of articles on cataIytic cyclophanes. See for example: P. Mattei, F. Diederich, Helv. Chim. Acta, 1997,80,1555-1588.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
12 Electroactive Intertwined Structures
In recent years one of the most active fields of supramolecular research has been that of intertwined structures such as catenanes, rotaxanes and molecular shuttles.[ll Although it is the pursuit of molecular devices that fuels this research, the aesthetic element of these systems has contributed to the immense interest these compounds have garnered. The complexity and artfulness of many of these structures provides one of the few instances in which the chemist can compete with nature in creating beautiful molecular architecture. These fascinating structures have been the subject of a number of excellent reviews within the past several years.[Zl In this chapter we shall concentrate on catenanes, rotaxanes and related structures. Helicates, a special class of intertwined structures, are discussed in Chapter 13. The basic structures of catenanes, rotaxanes, pseudorotaxanes and shuttles, well known to most readers, are shown schematically in Fig. 12.1, for the sake of clarity. Many of the reviews in reference 2 can give the reader an appreciation of the scope and rich history of this field of research. Although catenanes have been known for some time, the difficulty of their synthesis was a limiting factor in studies of this class of compounds. While early work in the field concentrated on the preparation of these fascinating intertwined structures, the incorporation of redox-active moieties into catenanes, rotaxanes and related structures rapidly led to the idea of creating redox-switchable functional assemblies. The goal of the work by leading research groups in this area has
Figure 12.1: Schematic representations (left to right) of a catenane, pseudorotaxane, rotaxane and a shuttle.
12.1
Electroactive Cyclodextn'n-Based Rotaxanes and Pseudorotmanes
143
been the development of molecular machines, and in particular molecular mformation storage devices.[31 Switchable catenanes, rotaxanes and related molecules undergo structural changes in response to either photochemical, electrochemical, or chemical triggers. Analogies to biological systems are also apparent. We shall begin our review of the field with a brief survey of redoxactive cyclodextrin-based systems, followed by sections devoted to redox switchable metallocatenanes and metallorotaxanes, and n-donor/acceptor systems.
12.1 Electroactive Cyclodextrin-Based Rotaxanes and Pseudorotaxanes In this section we will briefly examine several electroactive rotaxane structures that do not exhibit redox-switching in a manner similar to the n-donor/acceptor systems discussed below. Cyclodextrins (CDs) are quite well known as hosts with a preference for forming inclusion complexes with primarily hydrophobic or nonpolar guests.[41 This binding capacity of cyclodextrins has been exploited extensively in supramolecular chemistry.[5] In the area of catenanes and rotaxanes several thorough reviews are availableP, 61 We have selected several structures to examine. While a number of cyclodextrin rotaxanes with redox-active moieties have been prepared, only a few have been examined for their electrochemical behavior. Notable among the structures that have been prepared are a [Co(III)en]-stoppered a-CD rotaxane prepared by Ogino and coworkers,[q a related system prepared by Yamanari and Shimura,[81a [3]-pseudorotaxanewith a porphyrin core prepared by Lawrence, et a1,PI Macartney's [(NC),Fe(bpy)13+ terminated rotaxanes,[lO]and an a-CD rotaxane with a bis-benzylviologen core, prepared by Wenz and coworkers.[1*1 Isnin et al. prepared rotaxanes 1and 2 with an a-CD bead threaded by a 7 or 11 carbon chainP1 Due to the inherent asymmetry of the thread, two
l a (n = 7)
2a(n = 11)
lb(n=7) 2b (n = 11)
7SO^ K+ -
144
12
Electroactive Intertwined Structures
isomers (due to differing orientation of the cyclodextrin) were observed in differing proportions (a 4:6 ratio for 1a:lb and a 6:4 ratio for 2a:2b). Stability of the two isomers of either rotaxane were observed to differ. The a isomer forms, with the wide end of the CD toward the naphthalenesulfonate moiety, were subject to dethreading in polar solvents. These rotaxanes are also interesting in that the free ferrocene end of the molecules can bind yet another cyclodextrin, albeit 0-CD. Cyclic voltammetry of 11131 (isomeric *mix) yields a half wave potential of +0.40 V vs. SSCE for the oxidation of the ferrocene subunit. Addition of P-CD shifts this oxidation to +0.44 V vs. SSCE. This behavior is in agreement with previously observed values for other ferrocene derivatives in the presence of P-CDPI P-Cyclodextrin and its derivative heptakis-(2,6-O-dimethyl)Pcyclodextrin (DM-P-CD), were employed as bead components of [2]pseudorotaxanes based on viologen thread structures 3 - 5 P l In this system the viologen is a fully controllable recognition site- in a twist. on the'usual rotaxane formation scenario, only the reduced forms of the viologens are threaded through the p-CD cavity. Thus, as shown in Figure 12.1, the viologen thread is partially complexed by the host upon its reduction to the monocation radical and fully complexed when reduced to the neutral species. This scenario
[Weak Binding]
LStrong Binding)
3R=
/\/O\/\O/\/OH
4R= O -H
II
0
Figure 12.1: Scheme showing the formation of [2]pseudorotaxanes via threading of the viologen threads 3 - 5 by p-CD.
12.2
Teniplated Metallocatenates and Metallorotaxanes
145
is evident in the cyclic voltammetry. Fig. 12.2 shows the voltammetric response of 4 in the presence and absence of DM-P-CD. In this figure we can see that the presence of an excess of DM-P-CD affects the first reduction wave of 4 only slightly. The second reduction wave is strongly affected, with E1p shifted by approximately 160 mV to more positive potentials, indicative of strong stabilization of the fully reduced viologen in the presence of the cyclodextrin. Digital simulations were used to estimate the association constants of threads 3 5 with the two hosts. The association constant for the monoreduced state of 4 and DM-P-CD was estimated to be 100 M-1 while that of the fully reduced thread 4 and the host was 7000 M-1..[151Similar trends were observed for 3 and 5 and for all three threads with p-CD.
15.01
10.0 -
-
%
5.0-
V
0.0;
+"
-5.0 -
-
-la01
'
' ' -500
-..*' '
'
'
I
'
'
'
"
-700 -900 Potential vs SSCE, mV
'
'
"
-1100
'
-1300
Figure 12.2: Cyclic voltammograms (0.1 V/s) of a 0.5 mM solution of 4 in the absence (solid line) an presence (dashed line) of 0.125 M DM-p-CD in a pH 7, 0.1 M phosphate buffer.1151 Copyright VCH.
12.2 Templated Metallocatenates and Metallorotaxanes In this section we will address catenanes and rotaxanes whose structure is organized by the presence of a templating ion, typically Cu(1). This field has been pioneered by Sauvage and coworkers . Sauvage has used the term catenate to describe a metal templated catenane whose aromatic ligands provide coordination to the templating metal ion. Catenates are typically prepared stepwise, beginning with the synthesis of a macrocycle containing an
12
146
6
Electroactive Intertwined Structures
7
appropriate aromatic ligand, threading this macrocycle as a bead in a pseudorotaxane and then closing the pseudorotaxane to form the catenate. Catenated structures possess considerable stability: demetallation of the bound cation, or even the neutral metal atom, is more difficult in catenates than is seen in the analogous uncatenated (acyclic) complexes. By way of example, catenate 6 resists electrochemical demetallation, while the acyclic helical analogue 7 is demetallated at a mercury electrode at -1.7 V vs. SCE.[161 This unusual kinetic stability has been termed the cafenand effect. (Catenand is the term used by Sauvage for the free catenating ligand). This enhanced stability of the catenated systems makes them attractive as potential molecular information storage devices. To this end, Sauvage has sought to design molecules with well defined structures that undergo profound conformational changes upon changing the redox state of the bound ion. Thus, the differing states resulting from an external trigger or signal are readily observed. Among the catenates recently prepared by Sauvage and coworkers, perhaps the most intriguing structure is that of the "swinging catenate" 8, shown in Fig. 12.3, which displays linkage isomerism. Interlocking diphenylphenanthroline (dpp) and dpp/ terpyridine (terpy) cyclophanes provide coordination sites for a copper ion which may be switched between its Cu(1) and Cu(I1) redox states."7I This system provides an example of how redox-switching of coordinated ions can bring about broad structural changes in the catenates. Oxidation of copper(1) to copper(I1) results in a rearrangement of the asymmetric catenane from the four coordinate structure favored by copper(1) to a five coordinate structure favored by copper(I1). This rearrangement takes place via a sliding mechanism- the dpp/ terpy cyclophane rotates through the dpp cyclophane. Two redox couples can be observed for this system- one corresponding to CuI/IINs processes at 0.63 V vs. SCE and another corresponding to CuII/IN5. processes at -0.07 V vs. SCE. In keeping with the designations of Sauvage, N4 represents copper coordination to two dpp ligands and Ns coordination to one
12.2
Templafed Metallocatenates and Mefullorofnxnnes
8 (CN=4)
Red-Brown Color
147
8 (CN=5) Olive Green Color
Figure 12.3: "Swinging catenane" 8, in its two possible isomers, as determined by the redox state of the coordinated Cu ion.
dpp and one terpy ligand. The electrochemical behavior of this system is complex and bears closer examination. Oxidation of CufN4 to Cu"N4 occurs at +0.63 V vs. SCE. When this redox couple is examined at a scan rate of 200 mV/s in an electrolytically generated bulk solution of 8-Cu(II), this process is reversible. Restating this, after conversion from Cu"4 to CulIN4 by bulk electrolysis, the cyclic voltammogram obtained immediately following electrolysis shows a reversible redox couple of the expected current intensity. Examination of the cyclic voltammetric behavior of this electrolysed solution of 8-Cu(II) several minutes later does not yield this result however. Instead the redox couple at 0.63 V vs. SCE is diminished whilst a new peak, at -0.07 V vs. SCE begins to appear. This new peak has been assigned to the CuWN5 redox couple and indicates that the cyclophane with two ligating moieties has rotated in order to provide a five coordinate environment for Cu(I1). This process is marked by a color change in the analyte solution, which goes from a red-brown color (CulN4 ) to dark green (Cu"N4) and finally to an olive green color (Cu"N5). The course of the change from Cu11N4 to Cu11N5 (determined by disappearance of the redox couple at +0.63 V vs. SCE) takes around 20 h When the reverse electrolysis experiment is conducted- reduction of CuIIN5 to the Cu(1) state, the red color begins to reappear immediately upon electrolysing the solution. The cyclic voltammogram obtained following this electrolysis shows a redox wave at +0.63 V vs. SCE suggesting that electroIysis regenerates the Cu"4 species. Fig. 12.4 shows that the disparity in the rates for the conformational changeover processes are visible even in the cyclic voltammograms obtained in a solution of 8-Cu(II). Depending on the initial scanning direction, either the Cu1/11N4or Cu"/lN5 appear to be irreversible.
148
12 Electroactive Intertwined Shuctures
Figure 12.4: Cyclic voltammogram of 8 as CuIIN5 (in CHEN/TBABF4) illustrating the different rate constants for the two changeover processes (CN=4 vs. CN=5):(a) scanning from 1 V to -0.5 V first, (b) scanning from -0.5 V to 1 V first. The scan rate was 200 mV/s.[l7al Reprinted with permission of the American Chemical Society.
This curious finding, slower kinetics for the Cu"4 to CuW5 changeover process vs. those observed for the reverse process, has recently been attributed by these authors to the effects of solvent, traces of water and especially the presence of ligating anions such as chloride. All have been determined to drastically influence the kinetics of the conversion from CuIIN4 to C U ~ N ~ , [ ~ ~ ~ ] presumably because of the higher charge of Cu(I1). The presence of chloride ion was found to greatly accelerate the kinetics of the conformational change. In CHzCh the authors found that the rate for CuIIN4 to Cu"N5 conversion was dropped from on the order of 20 hours to around 2 min. As the authors point out, at some point in the changeover from CN=4 (dpp-dpp) to CN=5 (dppterpy), half of the coordination sphere around the Cu ion is lost. This perturbation of the coordination sphere is likely more difficult for the Cu(I1) ion than for Cu(1). Thus, chloride effectively lowers the activation barrier for the EPR changeover process by stabilizing this "half-naked" studies of this system have recently been published."81 Sauvage and coworkers have also studied related systems, including pseudorotaxanes, rotaxanes and catenated polymers.[*91 In the area of rotaxanes and catenanes the group of Sauvage has recently reported di- and trimetallic systems in which two different transition metal ions are present.1201 Structure 9 is a representative structure. The voltammetric behavior of 9 is quite simple in comparison with that of 8. The
12.2
Ternplated Metallocatenates and Metallorotaxanes
149
9
U V )vs SCE
Figure 12.5:Cyclic voltammogram of 7 showing (a) Ru and Cu redox couples and (b) the Ru waves for the rotaxane formed after Cu demetallation of 9. CHyZN/O.lOM TBABF4P" Reprinted with permission of the American Chemical Society.
12 Electroactive Intertwined Structures
150
voltammogram in Fig. 12.5 shows three reversible waves, with the half-wave potentials for Ru(III)/Ru(II) at +1.24 V, Cu(II)/Cu(I) at 0.58 V, and Ru(II)/Ru(I) at -1.32 V (all potentials vs. SCE)P" Cu(I1) can be cleanly demetallated from 9, yielding a rotaxane with Ru(II)(terpy)z stoppers. The demetallated species displays half-wave potentials for the ruthenium couples at virtually the same position as 9. The rotaxane can be remetallated with a variety of ions.
12.3 Catenanes Based on n-Donor and n-Acceptor Interactions One of the largest classes of electroactive intertwined structures includes those systems based on the mutual attraction of n-electron donors and acceptors for one another. Familiar from the concepts of host-guest chemistry, these systems employ the simple design concept of pairing n-electron deficient hosts with nelectron rich guests in order to construct what are sometimes very complex structures. Changing the redox state of the host or guest allows "tuning" of its acceptor or donor character. Electrochemical switching of the system can be translated into mechanical control over its conformation or topology. These structures have also been touted as possible molecular devices, As mentioned in Chapter 11, the bis-viologen cycIophane 10 is perhaps the most studied cyclophane of the last decade. The electrochemistry of 10 was first described as a component of a catenane system in 1989PI The electrochemistry of related viologen based cyclophanes with the xylylene bridges attached at the ortho or meta positions has also been describedP1 We can briefly review the electrochemistry of 10, which was discussed in Chapter 11. Cyclophane 10 shows reversible redox processes that are characteristic of the parent methylviologen unit, i.e., two single-electron redox waves are observed for each viologen unit. In the cyclophane structure, both viologens will undergo these reductions at the same potentials. This is the expected behavior for a molecule bearing two identical noninteracting redox The +4 charge on the cyclophane makes it an excellent n-deficient host for n-electron rich aromatic
+ *+
N
+N
10
12.3 Catenanes Based on Ir-Donor and rr-Acceptor Interactions
151
guests. Reduced forms of 10 (either the net +2 cation diradical or net neutral states) show a substantially decreased n-acceptor character, as shown in Fig. 12.6. The cyclic voltammetric behavior of 10 displays two redox waves, as shown in Fig. 12.7a. The first two electron wave corresponds to the monoelectronic reduction of each viologen moiety to its cation radical state at 0.279 V vs. SSCE in C&CN/TBAPFb, while the second two-electron wave corresponds to the monoelectronic reduction to the neutral viologens at -0.703 V.["]
v
- e-
Tetracationic 1 0 n-Electron Acceptor
gg$$
e-
A 7
- e-
NwN Dicationic, diradical 10: Weak n-Electron Acceptor
Neutral 10: n-Electron Donor
Figure 12.6:Effects of reduction on the n-acceptor character of 10.
The electrochemistry of the cyclophane 10 as a component of catenane 11 displays markedly different behavior from that of uncatenated 10, as shown in Fig. 12.7b. In contrast to the two reduction waves observed for 10, three waves are observed for 11, with the first two waves corresponding to separate monoelectronic reductions of the two viologen moieties. The reduction of the first viologen unit occurs at -0.307 V vs. SSCE, while the second viologen couple occurs at -0.438 V. The third and final reduction process occurs as a single two electron wave, corresponding to reduction of both viologens to their neutral state at -0.841 V. The first observation that can be made is that both viologen moieties in 11 are stabilized, to differing extents, by the presence of the nelectron donating hydroquinol units. This is evidenced by the fact that their reductions occur at more negative potentials than do the same processes in uncatenated 10. The splitting of the first reduction processes in catenane 11 has been attributed to the topological differences between the "inside" and "alongside" viologen units in the catenane. The inner viologen is stabilized to a greater extent by the two electron rich hydroquinol units, making this viologen
152
12
Electroactive lntertwined Structures
V -1.2
-0.6 E (V vs SSCE)
0.0
Figure 12.7: Cyclic voltammograms of 0.50 mM solutions of (a) 10 and (b) 11, in 0.1 M TBAPFb/CH&ZN at a scan rate of 50 mV/s obtained at 25 oC.[251 Reprinted with permission of the American Chemical Society.
12.3
Catenaries Based on Ir-Donor and Ir-Acceptor Interactions
0
0
0 'Nq
153
0 N'
---' W 11
unit substantially more difficult to reduce than the "alongside" viologen, which interacts with a single hydroquinol unit. The shifts in the half-wave potentials for reduction of 10 vs. 11 reflect the magnitude of this stabilization: 28 mV more negative for the alongside viologen cation radical, 159 mV further negative for the inside viologen cation radical, and finally another 138 mV more negative for the simultaneous reduction of both units to the neutral species. The tight interlocking fit of the macrocycles no doubt plays an important role in the observed voltammetric behavior of 11. Solvent interaction with the inside viologen is limited and this may influence the ease of reduction of this moiety. It is interesting to note, for instance, that a surface attached catenane based on cyclophane 10 and a bis(thio1)-terminated poly(ether) hydroquinol thread does not show a splitting of the first reduction processes, but rather exhibits behavior similar to Fee 10, displaying a single two-electron reduction for the first reduction of the viologens.[24] The extended polyether chains in this system, along with their mode of attachment to the electrode surface, may permit greater solvent and electrolyte interaction with both faces of the cyclophane. Catenane systems involving 10 and napthyl, rather than hydroquinol, macrocycles or cyclophanes with nextended viologens have recently been examined by Stoddart and Balzani.[251 The electrochemistry of 10 in systems with the napthyl cyclophanes exhibits splitting of the second electron transfer processes, as well as the first. Stoddart and Balzani have recently reported a desymmetrized catenane in which the crown ether macrocycle possesses two different n-donor moieties.[26]Catenane 12 has the TTF moiety inside the cavity of the tetracationic cyclophane 12. Electrochemical switching of 12 occurs when the TTF unit is oxidized. The crown ether ring undergoes circumrotation so that the 1,5dioxynaphthalene moiety resides inside the cavity of 10.1261 Cyclophane 10 has also been used as a component of a series of novel [3]pseudocatenanes by Becher et al., in a system employing TTF-based macrocycles.[271Structures 13 and 14 are &.and trans isomers (with respect to
154
72 Electroactive Intertwined Structures
' +
0
/ O
x.-; "
@oj29dN+ / O 0
N
J cyclophane 10 and its orientation about the TTF moiety), while 15 is the free TTF bismacrocycle. The electrochemistry of 10 in these systems was not reported, however the oxidative electrochemistry of the various isomers of these catenane systems was interesting in that cis and trans isomers exhibit different heterogeneous kinetics for the TTF redox processes. The voltammetry of 15 reveals three waves, two reversible monoelectronic waves at +0.50 V, and +0.79 V vs. Ag/AgCI, and an irreversible wave at +I38 V vs. Ag/AgCl. The presence of the electron accepting cyclophane 10 stabilizes the TTF units in the catenanes 13 and 14, as evidenced by the positive shifts in the TTF oxidation potentials. The half-wave potentials observed in the two isomeric catenanes were almost identical to one another. The first wave appeared at +0.63 V, the second at +0.94 V and the third at + I 3 4 V vs. Ag/AgCl for 13, while the oxidation potentials were +0.62, +0.92, and +1.34 V vs. Ag/AgCl for 14Pl We should note however that the first two oxidation waves are irreversible for 14, while those of 13 are reversible, a rather puzzling finding. Based on other cis and trans structures
13
14
15
12.4 Rotaxanes and Shuttles Based on n-Donor/Acceptor Clleniisty
155
reported, the effect on the heterogenous kinetics cannot be attributed solely to the orientation of 10 in these isomersP1 Other novel catenanes include a variant on catenane 11 reported by Benniston and Harriman. In their system a xylyl moiety in cyclophane 10 is replaced by a 22'-bipyridyl moiety. This substitution permits formation of a Ru(bpy)s modified catenane. Both the free Ru(bpy)s derivatized cyclophane and the nonmetalated catenane were also examined. The authors report electrochemical and photochemical studies on this system with the goal of elucidating redox asymmetry of cofactors in the bacterial photosynthetic reaction centerP1
12.4 Rotaxanes and Shuttles Based on n-Donor/Acceptor Chemistry Considering its popularity, it 1s hardly surprising that cyclophane 10 has been the dominant building block of many rotaxane and shuttle systems. In this section we shall trace the development of rotaxane thread components into components of an electrochemically controllable molecular shuttle. 10 has been used as the bead component in a rotaxane system[291 that is a precursor to a chemically and electrochemically controllable shuttle.[301 In structures 16 and 17, both the cyclophane bead and the thread (an
16
17
12
156
Electroactive Intertwined Structures
extended 1,4-phenylenediamine or benzidine) are electroactive and exhibit reversible electrochemistry. In rotaxane 16 two redox waves are observed for the first reduction of the viologen moieties, which occur at -0.272 V and -0.362 V vs. Ag/AgCl in CH3CN/TBAPF6. These half-wave potentials are shifted 21 and 111 mV negative from the single two electron wave observed for the free cyclophane 10, indicating the n-electron donor character of the phenylenediamine station, as well as electronic communication between the neighboring viologen units. (The degree of electronic communication has been noted to be dependent on the donor character of the guest.) The stabilization, as indicated by the negative shift in the half-wave potential for the first reduction processes of 16 is similar to that noted in the catenane system 11. The second electron transfer onto the viologen moieties occurs in a single step at a half-wave potential of -0.780 V vs. Ag/AgCl. The thread portion of this system exhibits far more dramatic shifts in the presence of cyclophane 10, as shown in Fig. 12.8. While the free thread exhibits two oxidation waves at 0.203 V and 0.732 V respectively, the rotaxanized thread undergoes oxidation at 0.463 V and 1.505 V (all potentials vs. Ag/AgCl) and exhibits dramatically slower ET kinetics in comparison to the free thread. The reason for this effect appears to lie in the limited mobility of the tetracationic cyclophane bead. 10 is unable to move far enough from the phenylenediamine station to avoid the high degree of electrostatic repulsion that results from creating the fully oxidized phenylenediamine, which raises the overall charge on the rotaxane to +6.
I
I
1
1
0.0
1.0
1.2
1.8
Potential in V vs. Ag/AgCl
Figure 12.8:Anodic voltammetry of 1 mM solutions of (a) the free thread from 16 and @) rotaxane 16 in CH3CN/O.10 M TBAPF6 at 25 oC. Scan rate was 200 mV/s).l291 Reprinted with permission of the American Chemical Society.
12.4 Rotaxanes and Shuttles Based on rr-Donor/Acceptor C l m z i s t y
157
A similar effect is noted, albeit to a lesser extent, in the benzidine rotaxane 17. As seen with 16, the viologen cyclophane in 17 undergoes its electron transfers at slightly more negative potentials than does free 10: the first of these are at -0.265 and -0.317 V and the second occur at -0.780 V vs. Ag/AgCl. The benzidine thread of 17 shows a more modest shift in the oxidation potentials when bound by the cyclophane, with oxidations occurring at 0.529 V and 0.714 V in the free thread, and 0.695 V and 0.998 V vs. Ag/AgCl in the rotaxane. The cyclic voltammetry of 17 is shown in Fig. 12.9.
r
I
1.0
0.0
I
1
1.2 1.3
Potential in V vs. Ag/AgCl
Figure 12.9: Anodic voltammetry of 1 mM solutions of (b) the free thread from 17 and (c) rotaxane 17 in CHSCN/O.10 M TBAPF6 at 25 oC. Scan rate was 200 mV/s)PI Reprinted with permission of the American Chemical Society.
Cyclophane 10 was used yet again as the bead component of a special type of rotaxane, molecular shuttle 18, whose thread contains both benzidine and biphenol stations.[301 In this system the cyclophane can move back and forth (shuttling) between the two donor stations on the thread. Fig. 12.10 shows the shuttling capability of this rotaxane. Since 10 will tend to prefer being localized on either of the two donor stations one can delineate two trundutionul isomers of 18. Spectroscopic evidence indicates that at -4.4 OCthe cyclophane spends more time near the benzidine station than the biphenol station (isomer ratio 86:14), i.e., benzidine acts as a better n-electron donor for 10. The electrochemistry of 10 in this system was found to be similar to that of the free cyclophane, with the two-electron reductions occurring in single waves at -0.235 V and -0.755 V
158
12
H
Electroactive Intertwined Structures
H
/\
Figure 12.10: The shuttling motion of the cyclophane bead in 18 generates two translational isomers. Oxidation of the benzidine moiety in 18 traps the cyclophane in the region of the biphenol station.
vs. Ag/AgCl in CHsCN. The electrochemistry of the biphenol station in the thread was not examined in any detail, as it is irreversible. The oxidation of the benzidine station in this system was found to have a first redox wave shifted to more positive potentials, 0.570 V vs. AglAgCl, while the second wave was comparable to that of thefvee, elongated benzidine thread used in rotaxane 17, 0.720 V vs. AglAgCl. Fig. 12.11 shows a comparison of the voltammetric behavior of the elongated benzidine thread from 17, rotaxane 17, and shuttle 18. The more difficult first oxidation of the benzidine station is attributable to the s t a b k i n g presence of the cyclophane 10 bead, while the ease of the second oxidation (comparable to that of the free elongated benzidine thread) is indicative of shuttling- the movement of 10, off the benzidine station and to other portions of the thread, presumably to the second, weaker donor station (biphenol). One explanation of the lack of splitting of the first electron transfers
159
12.4 Rotaxanes and Slzuttles Based on n-Donor/Acceptor Clwnistnj
\ I I
r
I
00
+1.2
Potential (V vs Ag/AgCI)
Figure 12.11: Cyclic voltammetric response of (a) 1 mM solution of the free thread from 17 (b) 0.50 mM solution of 18, and (c) 0.70 mM solution of rotaxane 17. The solvent system was CH3CN/O.10 M TBAPF6 and the scan rate was 100 rnV/s.[301 Reprinted with
permission of Nature.
on the viologen bead is that since occupation of the benzidine station is not 100% when the station is in the reduced state, the electrochemistry of the cyclophane bead is not affected to the same extent as it is in the case of rotaxane 10. Other shuttles have been reported (see for instance Ref. 2c) however most of these use cyclophane 10 as the bead component of the rotaxane. In a reversal of this common design, Balzani and Stoddart have reported a multistation rotaxane system in which 4,4'-bipyridinium units are incorporated into the thread component of a rotaxane. The bishydroquinol macrocycle that was a component of catenane 11 is used as a bead which can move from viologen station to stationP1 Structure 19 shows one such rotaxane, one of the more interesting of those examined. In this system the cathodic voltammetry reveals four monoelectronic reductions at -0.33 V, -0.44 V, -0.77 V, -0.86 V vs. SCE. The reductions at -0.33 V and -0.77 V can be assigned to the viologen reductions observed for the free thread of rotaxane 19. Those at -0.44 V and 0.86 V can be correlated with the potentials observed when the thread of 19 is
160
12
Electyoactive Intertwined Structures
0
1
19
‘tBu
bound by two crown ether beads, one bead localized on each of the viologen Thus, one set of reductions appears to correspond to a free viologen, stati0ns.1~~1 whilst the other appears to correspond to a viologen engaging in the CT interaction with the bishydroquinol bead. The shuttling of the bead between the bipyridinium stations was observed to be fast on the 1H-NMR timescale. Clearly the voltammetry indicates that given a single cyclophane bead, application of a potential to reduce the first viologen moiety results in the bead shuttling to interact with the other station, shifting the second station’s half-wave potential some 110 mV further negative for the first reduction and even affecting the second reduction by 90 mV.
12.5 Perspectives on the Future of Molecular Devices New reports of catenanes and rotaxanes continue to appear with frequency. These systems are sure to be the source of further exciting electrochemical
12.6 References
161
research in the future. But what of the future directions in the field? Undoubtedly, some of the molecules described in this chapter exhibit not only fascinating topologies, but also a very remarkable switchable character. Electron transfer events can be used to influence strongly their molecular dynamics, in such a way that internal motions and predominant molecular configurations may be determined through electrochemical or chemical redox reactions. In this sense, these compounds reinforce the belief that molecules can be utilized as minimal components for electronic circuits or information storage/retrieval devices. We must realize, however, that many difficulties lie ahead to transform such ideas into technological reality. Perhaps the most important problem that remains to be solved is that of addressing individual molecules. Switching of these molecules is now done in the solution phase and, of course, the switching stimulus acts simultaneously on a huge number of molecules. The key advantage to be gained from using molecular switches derives from their small size. To realize this advantage one must be able to address reproducibly one or just a small number of switchable molecules. Switching trillions of molecules at the same time is not technologically fruitful. For switchable molecules to become practical as circuit or memory devices, we should be able to fix the location in space of the target molecule (or small group of molecules) so that dormation can be stored and retrieved in a reproducible way. A possible future direction might be to immobilize the switchable molecules on solid surfaces. Just such an approach has recently been employed by Heath and coworkers.[321 We should also prevent communication transfer between adjacent molecules, which may lead to information 'scrambling' effects. The stability of these organic molecules, particularly under conditions of frequent switching, is also a serious concern. It has been pointed out that biological systems perform many switching functions successfully with organic materials. However, it is also true that living matter is renewed with some frequency. In any instance, living systems provide a very clear inspiration to continue this research work. ,It took billions of years of molecular evolution to develop efficient and complex living systems. It is difficult to anticipate how long it will take for mankind to develop synthetic systems capable of complex functions. However, the authors believe that it is not farfetched to anticipate systems such as molecular computers in a not too distant future.
12.6 References 1. D. N. Reinhoudt, J. F. Stoddart, R. Ungaro, Chem. Eur. I., 1998,4,1349-1351 2. [a] J.-P. Sauvage, Acc. Chem. Res. 1998, 31, 611-619; [b] V. Balzani, M. GomezLopez, J. F. Stoddart, Acc. Chem. Res., 1998,31,405-414; [c] J.-C. Chambron, J.-P. Sauvage, Chem. Eur. I., 1998,4,1362- 1366; [d] D. B. Amabilino, J. F. Stoddart, Chem. Rev., 1995, 95,2725 -2828; [el R. Jager, F. Vogtle, Angew. Chem. lntl. Ed. Engl., 1997,36,930 - 940. 3. Some photoswitchable systems can store more complex 1ogicaI states than just on/off states. See for example: [a] A. P. de Silva, H. Q. N. Gunaratne, C. P.
162
12 Electroactive lntertwined Structures
McCoy, Nature, 1993, 364, 42-44; [b] A. P. de Silva, C. P. McCoy, Chem. Indus., 1994, 992-996.4. There are exceptions to this binding preference, for instance the binding of some anions by cyclodextrin is well known. See for example: [a] L. A. Godinez, B. G. Schulz-Fiehn, S. Patel, C. M. Criss, J. D. Evanseck; A. E. Kaifer, Supramol. Chem., 1996, 8, 17-22; [b] R. I. Gelb, L. M. Schwartz, M. Radeos, D. A. Laufer, J. Phys. Chem., 1993,87,3349. 5. [a] G. Wenz, Angew. Chem. Int. Ed. Engl., 1994,33,803- 822; [b] Cyclodexfrins,J. Szetli, T. Osa, Eds., Comprehensive Supramolecular Chemistry, Volume 3, J. L Atwood, J. E. D. Davies, D. D. MacNicol, F. Vogtle, Series Eds., Elsevier, Oxford, U. K., 1996. The reader is also directed to issue number 5, volume 98 of Chemical Reviews, 1998, for a number of articles on the topic of cyclodextrins. 6. S. Nepogodiev, J. F. Stoddart,, Chem. Rev., 1998, 98,1959-1976. 7. H. Ogino, J. Am. Chem. SOC.,1981,103,1303-1304. 8. K. Yamanari, Y. Shimura, Bull. Chem. SOC.Jpn., 1983,56,2283-2289. 9. J. S. Manka, D. S. Lawrence, J. Am. Chem. SOC.,1990, 112,2440-2442. 10. R. S. Wylie, D. H. Macartney, J. Am. ChemSoc., 1992,114, 3136-3138. 11. G. Wenz, E. von der Bey, L. Schmidt, Angew. Chem. Intl. Ed. Engl., 1992,31, 783-785. 12. [a] R. Isnin, A. E. Kaifer, J. Am. Chem. SOC., 1991, 113, 8188-8190; [b] R. Isnin, A. E. Kaifer, Pure Appl. Chem., 1993,65,495-498. 13. R. Isnin, Ph.D. Dissertation, University of Miami, 1992, pp.148-149. 14. Ibid, pp. 41-61 and references therein. 15. A. Mirzoian, A. E. Kaifer, Chem. Eur. I. 1997,3,1052-1058. , 16. C. 0. Dietrich-Buchecker, J.-P. Sauvage, J.-M. Kern, J. Am. Chem. SOC.,1984, 3043. 17. [a] A. Livoreil, J.-P. Sauvage, N. Amaroli, V. Balzani, L. Flamigni, B. Ventura, I. Am. Chem. SOC., 1997, 219, 12114-12124; [b]J. P.Sauvage, J. Am. Chem. SOC., 1994,116, 4248-4254. 18. F. Baumann, A. Livoreil, W. Kaim, J.-P. Sauvage, Chem. Commun., 1997,35-36. 19. [a] J.-P. Collin, P. Gaviila, J.-P. Sauvage, New.1. Chem., 1997, 21, 525-528; [b] J.-P. Collin, P. Gaviiia, J.-P. Sauvage, Chem. Comrnun., 1997,2005-2006; [c] J.-M. Kern, J.-P. Sauvage, J.-L. Weidmann, N. Armaroli, L. Flamigni, P. Ceroni, V. Balzani; [d] J.-M. Kern, J.-P. Sauvage, G. Bidan, M. Billon, B. Divisia-Blohorn, Adv. Mafer., 1996, 8, 580-582; [el D. J. CBrdenas, A. Livoreil, J.-P. Sauvage, J. Am. Chem. Sot., 1996, 118, 21980-11981. 20. D. J. Cardenas, P. Gaviila, J.-P. Sauvage, J. Am. Chem. SOC., 1997, 119, 26562664. 21. [a] P. L. Anelli, P. R. Ashton, R. Ballardini, V. Balzani, M. Delgado, M. T. Gandolfi, T. T. Goodnow, A. E. Kaifer, D. Philp, M. Pietraszkiewicz, L. Prodi, M. V. Reddington, A. M. Z. Slawin, N. Spencer, J. F. Stoddart, C. Vicent, D. J. Williams, 1. Am. Chem. SOC., 1992, 114, 210-214; [b] P. R. Ashton, T. T. Goodnow, A. E. Kaifer, M. V. Reddington, A. M. Z. Slawin, NSpencer, J. F. Stoddart, Angew. Chem. Int. Ed. Engl., 1989,34,1396-1399. 22. W. Geuder, S. Hiinig, A. Suchy, Tetrahedron, 1986,42, 1665-1677.
12.6 References
163
23. [a] J. B. Flanagan, S. Margel, A. J. Bard, F. C. Anson, 1. Am. Chem. SOC.1978, 100, 4248-4254. [b] A. El-Kasmi, D. Lexa, P. Maillard, M. Momenteau, J.M. Saveant, 1.Phys. Chem. 1993,976090-6095and references cited therein. 24. T. Lu, L. Zhang, G. W. Gokel, A. E. Kaifer, J. Am. Chem. SOC.1993,115, 25422543. 25. P. A. Ashton, R. Ballardini, V. Balzani, A. Credi, M. T. Gandolfi, S. Menzer, L. Perez-Garcia, L. Prodi, J. F. Staoddart, M. Venturi, A. J. P. White, D. J. Williams, J. Am. Chem. SOC.1995,117,11171-11197. 26: M. Asakawa, P. R. Ashton, V. Balzani, A. Credi, C. Hamers, G. Mattersteig, M. Montalti, A. N. Shipway, N. Spencer, J. F. Stoddart, M. S. Tolley, M. Venturi, A. J. P. White, D. J. Williams, Angew. Chern. lnt. Ed. Engl., 1998, 37,333337. 27. [a] Z.-T. Li, P. C. Stein, J. Becher, D. Jensen, P. Mark, N. Svenstrup, Chem. Eur. J., 1997,2, 624-633; [b] Z.-T. Li, P. C. Stein, N. Svenstrup, K. H. Lund, J. Becher, Angew. Chem. l n f . Ed. Engl., 1995,34,2524-2528. 28. A. C. Benniston, P. R. Mackie, A. Harriman, Angew. Chem. Inf. Ed. Engl., 1998, 3 7,354-356. 29. R. A. Bissell, E. Cordova, A. E. Kaifer, 1. Org. Chem. 1995,60,1033-1038. 30. R. A. Bissell, E. Cordova, J. F. Stoddart, A. E. Kaifer, Nature, 1994, 369, 133137. 31. P. R. Ashton, R. Ballardini, V. Balzani, M. BZlohradsky, M. T. Gandolfi, D. Philp, L. Prodi, F. M. Raymo, M. V. Reddington, N. Spencer, J. F. Stoddart, M. Venturi, D. J. Williams, J. Am. Chem. SOC.,1996,118,4931-4951. 32. C. P. Collier, E. W. Wong, M. BZlohradsky, F. M. Raymo, J. F. Stoddart, P. J. Kuekes, R. S. Williams, J. R. Heath, Science, 1999,285,391-394.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
13 Helicates, Racks, Grids, and Coordination Arrays
In this group of beautiful supramolecular structures we encounter molecular architecture that is largely directed by the coordination preferences of metal cations in their various redox states. Self-assembly is typically employed to prepare these complexes and their initial structure is based on directing the ligands’ spatial orientation in order to satisfy a preferred coordination geometry. In some of these systems, e.g. the helicates, a subsequent change in the redox state of the metal ions may result in profound changes in the molecular architecture, yielding exchanges between mono-, di-, tri- and even higher order metallic complexes. Thus, interpretation of electrochemical studies may be complicated by the fact that the forward and reverse potential steps or scans may, in fact, involve diflerent species, with a differing number of redox active sites in each species. In some structures, interactions between neighboring redox sites will result in differing reduction and oxidation potentials for each redox active site. Spectroelectrochemicalstudies may prove invaluable in the examination of these complicated processes.
13.1 Helicates Helicates are a class of molecules that have been of considerable interest over the past decade. Early examples of these molecules involved short oligopyridine ligands which self-assemble in the presence of metal ions, forming helical structures. Several very fine reviews of these structures are available in this area.[ll More recently, these helical systems have involved use of extended oligoterpyridine ligands to form longer and intricate structures.[2] As the structures have become more elaborate however, fewer electrochemical studies have been performed, owing perhaps to limited amounts of material, as well as lower solubility of some of these materials. In this section we have chosen to discuss only copper helicates, since the clear preferences for differing coordination geometries exhibited by Cu(1) and Cu(I1) ions result in dramatic and often complicated voltammetric behavior. Examination of the copper helicates formed by Iigands ranging from quaterpyridines to septipyridines can provide examples of the challenges involving interpretation of electrochemical data in these systems. Lehn and coworkers, who were among the first to examine the electrochemical behavior of the helicates, studied a Cu(I1)tetramethylquaterpyridine (pQP) complex, 1, and observed unusual voltammetric behavi0r.1~1The voltammetry of this system, and that of a closely
13.1
Helicufes
165
1 related system described below, are illustrative of the complexity encountered even in this most basic of the helicate systems. pQP-Cu(I1) 1 undergoes conversion to a dimeric dimetallic species upon reduction of the Cu(I1) ion and serves as a fine example of the ability of redox switching to control molecular architecture. In the oxidized monomeric Cu(I1) form, the pQP complex 1exists as a tetragonal pyramid with an apical water molecule, providing the higher coordination favored by Cu(I1) ions. Upon reduction, the Cu(1) complex dimerizes rapidly to a distorted tetrahedrally coordinate dinuclear helical form, [CU(I)~(PQP)~], reflecting the preferred coordination for Cu(1) ions. Lehn and coworkers found that subsequent controlIed oxidation of this helical dimer can successfully generate the mixed valence [Cu(I)Cu(II)(pQP)z]complex. Further oxidation regenerates the monomeric Cu(I1) form, albeit on a slower timescale.
reduction, dimerization (fiust)
dimerization (slow)
P
O
= Pyridine
= Cu(I1)
0= Cu(1)
Figure 13.1: Structural changes observed in 1upon changing the redox state copper ion.
166
13 Helicates, Racks, Grids, and Coordination Arrays
This difference in timescale, for the electrochemical steps driving conversion between the monomeric and dimeric species, is clearly evident in the voltammetric response. These structural changes are shown schematically in Fig. 13-1. (Throughout the section on helicates we have chosen to use the schematic diagrams after the style of Abrufia and coworkers in order to illustrate the changes in molecular structure resulting from the redox chemistry.) The fact that the above scheme is clearly delineated in the cyclic voltammetric behavior of 1, shown in Fig. 13-2, is quite unusual.[31 The start point of the voltammogram shows the behavior observed for a solution of the [Cu(II)pQP] monomer as the potential is scanned in a negative direction, beginning at a potential of + 1.0 V vs. SCE. A single wave, corresponding to the reduction of Cu(I1) ions to Cu(I), is observed near 0.10 V, however, upon reversing the potential scan at -0.5 V vs. SCE and scanning in the positive direction, only a slight residual anodic wave, corresponding to the foregoing reduction process, is observed. The loss of the corresponding anodic process clearly signifies that a chemical process has occurred following the reduction of the Cu(I1) ions to their Cu(1) state, i.e. that an EC process occurs with respect to reduction of 1.
Figure 13.2 Cyclic voltammogram showing reduction of [Cu(II)pQP] monomer and the resulting voltammetric behavior of the redox generated [Cu(I)z pQPz] in CH3CN with 0.10 M TEAP at a scan rate of 0.30 V/s. Potentials are shown vs. a SCE reference, current is in pAS31 Reprinted with permission of Elsevier.
13.1
Helicates
167
As the scan continues in the positive direction two oxidation waves are observed. The first of these waves, at 0.56 V, has been assigned to oxidation of a single Cu(1) ion in a dimeric [Cu(I)2(pQP)z]complex to form [Cu(I)Cu(II)(pPQ)z]. The second wave, at 0.76 V vs. SCE, corresponds to oxidation of this mixed valent species to [Cu(II)2(pQP)2]. The dimeric Cu(I1) species is stable enough to permit observation of its corresponding reduction waves, at 0.50 and 0.70 V, respectively. The fact that reduction to Cu(1) does not occur at the same potential as was observed in the initial negative-going scan highlights the point that the species at the electrode surface now differs from that present in the original (bulk) solution. Further studies (including synthesis of the Cu(I1) dimer and examination of its electrochemical and spectroscopic behavior) permitted the authors to determine with certainty that the voltammetric behavior corresponds to the process shown in Fig. 13-1. The differing half-wave potentials for the oxidation of the Cu(1) ions in the [CU(I)Z-~QPZ] dimer indicates that the Cu ions are not independent of one another. This difference in potentials can be interpreted as arising from a change in the coordination sphere of the remaining Cu(1) ion, which could be sufficiently perturbed by the oxidation of the first Cu(1) ion as to make the remaining ion considerably more difficult to oxidize.Pl The second oxidation demands further reorganization of the molecular architecture (and indeed to a more unstable state, since the [Cu(II)2(pQP)2]dimer was found to be unstable in comparison to the monomeric [Cu(II)(pQP)]).A The cyclic voltammetric and bulk electrolysis behavior indicates that the [Cu(I)-pQP] dimerized species is quite stable. The Cu(1) ions undergo oxidation at much more positive potentials than would be predicted for the corresponding reversible redox process in the monomeric Cu(1) species. The stability of the [Cu(I)-pQP] dimer is also supported by the finding that reduction of the Cu(1) ions to the neutral copper species is shifted more than 1 V negative versus the reduction potential in an analogous Cu(I)(bipy)z complex. The oxidized dimer [Cu(II)2(pQP)2]was not found to be stable over time and always reverted to the monomeric form, if not stored in the presence of a reducing solvent or agent.W Following their work on 1, Lehn and coworkers continued to pursue studies of helicating ligands, preparing oligopyridines with as many as five bipyridine ligands, bridged by dimethylether linkages. 141 Their work has recently extended to preparation of Cu(I1) complexes based on oligoterpyridine structures.[21 Constable and coworkers[51, as well as AbruAa, et aW have also explored oligoterpyridine structures, as well as systems based on the original oligobipyridine ligands. These groups have examined a number of such helicate systems, complexed with a variety of transition metal ions and have reported electrochemical results. Abruiia, and coworkers have studied the electrochemical behavior of a series of alkylthio-substituted 2,6-oligopyridines, in great detail. Constable’s work, as a result of this group’s strong synthetic orientation, has focused on the preparation of a variety of different helicate systems. These include such novel structures as a quinquepyridine system bearing pendant “redox spectator” ferrocene moieties.[q The ligands employed by AbruAa and Pott’s systems are pyridine, methylthio- or propylthiopyridine (mt-py or pt-py), subunits, in which the
13 Helicates, Racks, Grids, and Coordination Arrays
168
appended thioalkyl groups enhance solubility. An example of their Cu(1I) methylthioquaterpyridine complex [Cu(II)(mt-qpy)] is shown in structure 2.R (The overall structure is quite similar to that of 1.) The authors have reported, in rich detail, the electrochemical behavior of an entire series of Cu helical complexes based on oligopyridines from quater-, quinque-, sexi-, and septipyridine ligands. Based on the results of Lehn et al. for their simple quaterpyridine system, it is no surprise that the structural variations observed for helicates have become considerably more complicated as the length of the ligands increases. OCH,
-S
I
It is surprising however, that the voltammetric results obtained by Abruiia and coworkers for 2 differ somewhat from the behavior noted by Lehn and coworkers for their analogous methylquaterpyridine system 1.181 Both the Cu(I1) complexes with ligands 1 and 2 clearly dimerize upon reduction to the corresponding Cu(1) complex. In the case of 2 only a single wave for the oxidation of the bimetallic Cu(1) system to the Cu(I1) state is observed (in DMSO or DMF/TBAP), see Fig. 13-3. This result is clearly in contrast with that observed by Lehn and coworkers for their system, i.e. in Abruiia and Potts system no mixed valenf state (appearing as two discrete waves for the reduction of the first, and then the remaining, Cu(1) ion) is apparent in the voltammetry. Furthermore, the regeneration of monomeric [Cu(II) mt-qpy] appears to be much more rapid in the thiosubstituted complex. There is also the notable difference in the potentials for formation of reduced Cu(1) species (+0.13 V vs. SCE for Lehn and -0.07 V vs. SSCE for Abruiia), that is not accounted for by the negligible difference between the reference electrodes. Abruiia points out that the solvent interactions may differ significantly in the case of Lehn’s complex, which was examined in CHKN (vide infra).rsl (This result clearly proves the cautionary note to electrochemists seeking to compare results in differing solvent systems!) The ligand electrochemistry yielded four reversible to quasireversible single electron waves at -1.38, -1.58, -1.83 and -2.02 V vs. SSCE. Spectroelectrochemistry of this system was also reported.181
13.7
169
Helicates
I I I I I I I1 I
4.85
I 1 I
I I I I I I II I
I
0.00
E(V) vs SSCE Figure 13.3: Cyclic voltammogram of [Cu(II) mtqpyIPF6 (2) in DMSO with 0.10 M TBAP as supporting electrolyte, at a scan rate of 0.10 V/s.[W Reprinted with permission of the
American Chemical Society.
Cu(I1) quinquepyridine complexes have been examined by both Abruiia and Constable. In Abrufia's work, ligand 3 (mt-qnpy), shown in Fig. 13-4, was used to generate the double stranded helicate depicted in Fig. 13-4PI The authors prepared the trimetallic Cu(1) complex, formed with two strands of ligand 3. This trimetallic species was found to be air sensitive when in solution, and is readily oxidized to the bimetallic mixed valent complex. A prepared bimetallic mixed valent complex was found to be stable. The bimetallic hornovalent complex [Cu(II)2 (mt-qnpy)~][PF6]4 possesses octahedral and tetrahedral geometries about the two Cu ions, as does the mixed valence complex. (The remaining sites for octahedral coordination of the second Cu(I1) ion are provided by an acetate ion.) The trimetallic complex should possess tetrahedral geometry and appeared to be highly symmetric according to NMR spectra.[81 The voltammetric behavior of the mixed valent ([Cu(I)/Cu(II) (mtqnpy)z] [PF6]3) and the homovalent ([cu(II) (mt-qnpy)]z [PF6]4) bimetalhc complexes was found to be identical. The voltammetry, performed in CHKN/TBAP, yielded two, one electron waves at -0.08 V and +0.48 V vs. SSCE. (Depending on the oxidation state of the Cu ion, the wave at +0.48 V was assigned as an oxidation or reduction by measuring the current response by RDE at several rotation rates.[81) A small additional wave at +0.35 V vs. SSCE confirms that CH3CN interacts with the Cu ions in these complexes (vide supra). The electrochemistry of the trimetallic complex [Cu(I)3(mt-qnpy)~][PF6]3 shows
170
13
Helicates, Racks, Grids, and Coordination Arrays
three one electron waves at -0.04, +0.50 and +1.03 V vs. SSCE, all of which may be assigned to oxidations to Cu(I1). (Note the similar potentials vs. the bimetallic complex- in the presence of molecular oxygen the oxidation of the trimetalllic species should result in a slow loss of the third wave.) The substantial separation between the redox waves suggests that considerable metal-metal interaction exists in the trimetallic complex. Ligand reductions for all three complexes were identical, yielding half-wave potentials of -1.44, -1.62 and -1$8 V vs. SSCE. Constable and coworkers have studied a similar, unsubstituted system.191
‘S
S’
3
pm = Pyridine
ox
ox
+ ‘red
%
4
+a-
Q
0
Q
= Cu(1I)
Figure 13.4 Structural changes observed in helicates composed of ligand 3 upon changing the redox state of the copper ion.
A quinquepyridine ligand 4, bearing two ferrocenyl moieties that the authors term “redox spectators,” was prepared by Constable and coworkers.[q This ligand was employed in the preparation of helicates with several different transition metal ions, however only the results on the copper based system are described here. The helicating ligand readily formed a bimetallic mixed valent
13.1
Helicates
171
complex. The cyclic voltammetry of [Cu(I) Cu(I1) 421 [PF6]3 yielded two oxidation waves, at +0.48 V and +0.69 V vs. SSCE (potentials converted from the authors' internal reference potentials vs. Fc/Fc+) corresponding to the oxidation of a single Cu(1) ion and the four ferrocenyl moieties, respectively (assignments made on the basis of current intensities). The reduction of the Cu(I1) ion in the complex occurred at -0.04 V vs. SSCE. Dissociation of the dimeric complex at very positive potentials, which might have been expected on the basis of the high positive charge generated by the presence of two Cu(I1) ions and the four pendant ferrocenium moieties in a single complex, was not observed. Constable, and coworkers reported ligand reductions at -0.81, -1.36, -1.58, and -1.84 V vs. SSCE for this system.[51
4 Abruiia and coworkers have also recently published work on a terpyridyl helicate that is conceptually related to the quinquepyridine system 3, discussed above. We mention this system because of the strong interaction observed between the copper ions in the complex and the large shifts in the anticipated redox potentials. Ligand 5, a methylthio-diphenyl-substituted terpyridine (mtdp-terpy), was used to probe stability of Cu(1) terpyridine (terpy) complexes, which are air sensitive and readily oxidize to form the Cu(II)(terpy)n complexes.[~01The solution structure of the complex appears to be an unusual diamond-llke tetrahedral geometry. The distance between the Cu(1) ions is only 2.6 suggesting that strong metal-metal interactions might be anticipated. While the closely related [Cu(II) (mt-terpy)z] undergoes reduction at -0.29 V vs. SSCE and Cu(0) plates onto the electrode surface at -0.62 V, the helicate structure [Cu(I)2 (mtdp-terpy)z] has its redox potential for conversion between the Cu(I)/Cu(II) couple at +0.12 V and +0.98 V vs. SSCE in the same solvent (CHKN). The 400 mV shift in the first Cu(I)/Cu(II) couple for the helicate suggests that the diphenyl substituted ligand stabilizes the Cu(1) oxidation state significantly more than the mt-terp ligand itself. The two half-wave potentials for the Cu(I)/Cu(II) couple were found to be solvent dependent but the very large AE value of 860 mV clearly suggests that a strong metal-metal interaction does exist in this complex. Constable[5] and LehnPI have also reported
A,
172
13 Helicates, Racks, Grids, and Coordination Arrays
\
S'
S
n
5
Q.Yx2 0
0 Cu(1) =
= Pyridine
0 = Phenyl
Figure 13.5:Solution structure observed for the Cu(1) helicate ligand 5.
work in the area of terpyridine ligands, preparing a variety of complexes. The electrochemistry of a dinuclear Cu(I1)-methylthiosexipyridine ([Cu(II)(mt-sexpy)]~[PF6]4) complex, 6 shown in Fig. 13-6, along with its t r h e t a k analogue [Cu(I)3(mt-sexpy)z][PF6]3, was reported by Abruiia, along with helicate complexes with a number of other transition metals.[llI The electrochemistry of the dinuclear copper complex (CH3CN/TBAPF6) demonstrates a simultaneous reduction of the two Cu(I1) ions in a quasireversible redox process (+0.02 V vs. SSCE) and is followed by incomplete reoxidation (+0.11 V vs. SSCE) to the initial dinuclear form. A second wave in the anodic range was observed at +0.45 V vs. SSCE. This wave appeared to be related to an EC process. Repeated voltammetric cycling demonstrated a gradual increase in the intensity of the wave at +0.45 V, at the expense of the anodic wave at 0.11 V vs. SSCE. The new peak was shown to correspond to the oxidation of the trimetallic species [Cu(I)s(mt-sexpy)z] [PF6]3, which was generated as a product of the reduction of the corresponding dimetallic species ([Cu(II)(mt-sexpy)]~[PF6]4). A prepared sample of the trimetallic complex demonstrated the ability to generate the corresponding dimetallic species by oxidation. Ligand based reductions were noted at -1.38, -1.61, -1.78, -1.91 V vs. SSCE, although the authors note that this complex begins to adsorb onto the electrode surface after the third ligand based reduction.
23.2 Helicafes
173
S'
w
= Cu(1I)
0 = Pyridine
Figure 13.6: Structural changes observed in helicates composed of ligand 6 upon changing the redox state of the copper ion.
The final example we mention among the helicates is the electrochemistry of an Cu(I)/Cu(II)-alkylthiolseptipyridine (at-septipy) complex formed with ligand 7,shown in Fig. 13-7. In the presence of both ions the ligand formed helical structures bearing two, three and four metal ions. The dinuclear complex contained two Cu(I1) ions, the trinuclear complex two Cu(1) and one Cu(I1) ions, and the tetranuclear complex four Cu(1) ions. The geometry of the trimetallic mixed valent system was ascertained to include one octahedral (for Cu(I1) ) and two tetrahedral (for Cu(1) ) sites. The geometry of the tetrametallic system was less easily determined, having some indications to be a variant of the diamond-like tetrahedral geometry determined for the terpy complex (ligand 6) above. The dimetallic complex is assumed to be octahedrally coordinate for both ions. The cyclic voltammetry of the tetrametallic complex [Cu(I)4 (at-septipy)Z ] [PF& was found to yield four distinct and reversible oxidations in CH3CN at + 0.02, + 0.44, + 0.84, + 1.08 V vs. SSCE. The large AE values are suggestive of strong metal-metal interactions. The mixed valent [Cu(I)2Cu(II) (at-septipy)z ] [PF6]4 possessed quasireversible (AE, ranging from 160 mV to 220 mV) oxidation waves at very close to the same values: + 0.01, + 0.43, + 0.82 V vs. SSCE. The dimetallic complex [Cu(II)2 (at-septipy)~] [PF6]4 had two quasireversible waves at + 0.05 and + 0.24 V vs. SSCE. Interestingly, the reported ligand reductions in the tetra- and trimetallic complexes appear to differ markedly, a finding not noted in the smaller oligopyridine helicates. Reduction of the ligands was noted at -1.11,-1.27, and -1.49 (irrev) V vs. SSCE in the tetrametallic complex, while the same processes occur at -1.42, -1.62, and -1.83 (irrev) V vs. SSCE in the trimetallic complex. The transformation processes resulting in the interconversion of the complex species, from the tetrametallic to the dimetallic complex, were found to be relatively slow. Spectroelectrochemical results suggest that multiple species exist in equilibrium at a given potential rather than a simple equilibrium between two species.
13
174
Helicates, Racks, Grids, and Coordination Arrays
-S
S-
7
= Pyridiae
!!if! 0 =Cu(I) 0
= Cu(I1)
Figure 13.7: Structural changes observed in helicates composed of ligand 7 upon changing the redox state of the copper ion.
The studies of the various helicates clearly demonstrate the complexity that can be involved in the interpretation of the voltammetric behavior of systems that can undergo changes in molecular architecture. In these systems the combined approach of spectroelectrochemical experiments are likely to provide the best keys for unraveling the intricacies observed. Spectroscopic approaches followed by Lehn and coworkers, and Abrufia and colleagues, typically relied on changes in the MLCT transitions in the visible range spectra upon appropriate potential steps. For mixed valent complexes AbruAa employed near-IR spectra. In situ EPR studies may also provide useful mformation.
13.2 Molecular Racks, Grids and Coordination Arrays
175
13.2 Molecular Racks, Grids and Coordination Arrays Another area of very recent interest has been the preparation of molecules possessing linear arrays of coordination sites. These systems also self-assemble in the presence of ions such as Ag(1) or Cu(1). The majority of the work in this area has been that of Lehn and coworkers. These are fascinating systems, which Lehn has reported can be extended into small grids of 3 X 3 coordination sites. The possibilities of such small grids, such as switching of the redox states of ions in order to create small potential gaps or "holes", are evident. While the obvious advantage of this approach for creating switchable devices is the easy self assembly of the systems, the solubility of the resulting complexes remains an important consideration from the electrochemical standpoint. Surprisingly, reports on the development of molecular arrays and grids have preceded those of the simpler racks and ladder-type systems. One of the early array-type systems was reported by Youinou and coworkers.[~31Ligand 8 spontaneously formed a simple 2 X 2 type array with coordination sites for four
Figure 13.8: Ligand 8 and two views of the 2 X 2 grid this ligand can for with four &(I) ions.
Cu(1) atoms, as shown in Fig. 13-8. The cyclic voltammetry reported by these authors yielded four reversible single electron ligand based reductions to at 0.88, -1.01, -1.16 and -1.32 V vs. SCE. Three additional waves at more negative potentials appeared to correspond to the second reductions of the ligands, while the final reduction appeared to be obscured by the solvent window. (The free ligand reductions begin at -2.07 V vs. SCE and are irreversible.) Curiously, the authors did not report the oxidative electrochemistry of the system, although the oxidation of the Cu(1) ions could clearly influence the architecture, as we have seen with the helicates. A 3 X 3 grid system was reported by Youinou and Lehn in 1994.[141 Ligand 9 and nine Ag(1) ions were used to compose the grid (see Fig. 13-9). The authors reported the crystal structure and various NMR spectra but electrochemistry of the grid-like array was not reported. Other authors have also explored electroactive multinuclear arrays and grid-like strucutres, albeit based on different schemes. These include Harriman and Ziessel,[*5]and Michl.1161
13 Helicates, Racks, Grids, and Coordination Arrays
1 76
9 Figure13.9 Ligand 9 and a schematic representation of its 3 X 3 grid formed with nine.
Molecular racks are in many respects simpler structures than molecular arrays. A linear ligand provides multiple binding sites for ions and the remaining coordination sites for the ions are fulfiLled, orthogonally, by a second, smaller set of ligands. The electrochemistry of three such systems, structures 10,
10 11 12
23.3
Conclusions
177
11, and 12 will be mentioned here.[17] Both the metals ions and the ligands are redox active. The Ru(I1) complexes display octahedral coordination geometry. The Ru-Ru distance in complex 11 is 6.40 A, suggesting that the metal-metal interaction will not be very strong. The cyclic voltammetric response of 10 demonstrates two separate reversible oxidations of the Ru(I1) centers at +1.41 V and +1.57 V vs. SCE, indicating that some interaction between the Ru(I1) ions still exists, in spite of their rather large separation, no doubt due to the nature of the bridging ligand. The ligand based reductions attributed to the bipyrimidine units occur at -0.43 and -1.03 V vs. SCE, while the terpyridine units reduce at 1.50, and -1.59 V vs. SCE. The latter two reductions generate the neutral material, resulting in adsorption and are irreversible in CH3CN but reversible in the more solubilizing DMFPJ Further reduction results in desorption of the negatively charged species. The electrochemical response of 11 is similar, indicating that the presence of the anthryl spacer does not influence the electrochemical properties of the complex. The metal-based voltammetry of 10 and ligand-based voltammetry of 11 is shown in Fig. 13-10.[*81 The voltammetry of the trimetallic rack 12 differs from that observed for 10 and 11, with the oxidations for the Ru(I1) ions occurring in two steps, a two electron wave at +1.47 V and a single electron wave at +1.88 V vs. SCE.[W The latter process is assigned to the central Ru(I1) ion, which is likely to be more difficult to oxidize when placed between the two higher charged centers. The ligand based reductions in this system occur at -0.33 and -1.28 V for the bipyrimidine centers, and, at -1.62 V (all V vs. SCE), a single irreversible adsorption wave is noted for the terpyridine units. Other rack systems include rotaxane-type racks reported by Lehn and Rissanen."91 These interesting hybrid structures, bearing a linear coordinating ligands similar to 9, possess phenanthroline cyclophane ledges to complete the rack coordination sites for Cu(1) ions. No voltammetry has been reported for these systems.
13.3 Conclusions Recent developments in the field of coordination arrays have clearly shown the promise of creating redox or photoswitchable nanodevices. A number of authors have suggested that lightweight organic switches could be used for memory storage, with the redox state of the ions or other redox active components coding for information. That such devices could be self assembled bodes well for the field. From the electrochemical perspective however, there are many questions that are raised. As we have seen, in many of the systems detailed above, the redox active sites interact with one another. This poses a fundamental design problem in that the neighboring sites affect the capacity to store information at discreet sites. Furthermore, changes in the redox states can convey profound changes in the molecular architecture and thus there is also the question of maintaining the integrity of the molecular design as potentials are
178
13 Helicates, Racks, Grids, and Coordination Arrays
0.0
-0.5
E (V) vs SCE
l I L I I I I I I I l #
+2.0
t1.5
+I .o
1
1
1
1
C0.5
1
1
1
, I
t0.0
E (V) vs SCE
Figure 13.10 Cyclic voltammetry of molecular racks 10 and 11 in DMF/TBAPFb: a) the ligand-based reductions for rack 10; b) the two Ru(I1)-based oxidation waves observed for l l . l l S 1 Copyright VCH.
changed. Nonetheless, these systems clearly show the promise of this area of research and are, in and of themselves, fascinating structures.
13.4 References 1. A good introduction to the use of helical structures in the field of supramolecular chemistry is provided by [a] A. E. Rowan, R. J. M. Nolte, Angm. Chem. Intl. Ed. Engl., 1998, 37, 63-68. Comprehensive reviews of the
topic of helicates are provided by [b] E. C. Constable, Progress in Inorganic
13.4
References
179
Chemistry, Vol. 42, (Ed. K. D. Karlin) 1994, 67-138; [c] E. C. Constable, Tetrahedron, 1992, 48, 10014-10059. The first report of synthesis and electrochemistry of oligopyridine helicate systems (with Ni(I1) and Co(I1) ions) was that of E. C. Constable, J. Lewis, M. Schroder, Polyhedron, 1982,2,311-312. 2. See for example, B. Hasenkopf, J.-M. Lehn, G. Baum, D. Fenske, Proc. Nafl. Sci. Acad., 1996,93,1397-1400 and references therein. 3. J.-P. Gisselbrecht, M. Gross, J.-M. Lehn, J.-P.Sauvage, R. Ziessel, C. PiccinniLeopardi, J. M. Arrieta, G. Germain, M. V. Meerssche, N o w . J. Chimie, 1984, 8,661-667. The reader should note that the figure captions for figures 3 & 4, which display the voltammetric behavior of the complex 1, and its dimeric analogue, were mistakenly reversed in the original article. 4. See for example, R. Kramer, J.-M. Lehn, A. Marquis-Rigault, Proc. Nafl. Acad. Sci. USA, 1993,90,5394-5398 and references therein. 5. E. C. Constable, P. Harveson, D. R. Smith, L. Whall, Polyhedron, 1997,26,36153623, and references therein. 6. K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abrufia, C. Arana, Inorg. Chem., 1993,32,4450-4456. 7. E. C. Constable, R. Martinez-Mafiez, A. M. W. CargillThompson, J. V. Walker, J. Chem. SOC., Dalton, 1994,1585-1594. 8. K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abruiia, C. R. Arana, Inorg. Chem., 1993,32,4422-4435. 9. [a] E. C. Constable, M. G. B. Drew, M. D. Ward, J. Chem. Soc., Chem. Commun., 1987, 1600; [b] M. Barely, E. C. Constable, S. A. Corr, C. S. McQueen, J. C. Nutkins, D. Ward, M. G. B. Drew, J. Chern. SOC., Dalton Trans., 1988,2655. 10. K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abrufia, C. Arana, Inorg. Chem., 1993,32,4450-4456. 11. K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abrufia, C. Arana, Inorg. Chem., 1993,32,4436-4449. 12. K. T. Potts, M. Keshavarz-K, F. S. Tham, K. A. GheysonRaiford, C. Arana, H. D. Abrufia, Inorg. Chem., 1993,32,5477-5484. 13. M.-T. Youinou, N. Rahmouni, J. Fischer, J. A. Osborne, Angew. Chemie, Intl. Ed. Engl., 1992,32,733-735. 14. P. N. W. Baxter, J.-M. Lehn, J. Fischer, M.-T. Youinou, Angew. Chem. Inf. Ed. Engl., 1994,33,2284-2287. 15. A. Harriman, M. Hissler, R. Ziessel, A. De Cian, J. Fischer, J. Chem. SOC., Dalton Trans., 1993,4067-4080 16. L. PospiSil, M. Heyrovsky, J. Pecka, J. Michl, Langmuir, 1997,23,6294-6301. 17. G. S. Hanan, C. Arana, J.-M. Lehn, D. Fenske, Angew. Chem. Intl. Ed. Engl., 1995,34,1122-1124. 18. G. S. Hanan, C. R. Arana, J.-M. Lehn, G. Baum, D. Fenske, Chem. Eur. I., 1996, 2,1292-1302. 19. H. Sleiman, P. Baxter, J.-M. Lehn, K. Rissanen, J. Chem. SOC., Chem. Commun., 1995,715-716.
SupramolecularElectrochemistry Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
14 Electroactive Langmuir-Blodgett Films
Over the past several decades, considerable interest in the field of electrochemistry has centered on the development of modified electrodes. Modification of the electrode surface can alter the reactivity of the surface by reducing the size and accessibility of reactive sites at the interface. Chief among the goals of modification of the electrode surface has been the potential for controlling heterogeneous kinetics and/or molecular recognition processes. Two of the most popular methods that have been exploited to create modified electrodes involve Langmuir-Blodgett and self-assembly techniques. LangmuirBlodgett (LB) methods, which rely on mechanical compression of the molecules into an organized film, can yield monolayer, bilayer and multilayer modified electrodes. Self-assembly methods have typically been used to prepare monolayer modified surfaces, the so-called SAMs, although bilayer and multilayer systems may also result from self-assembly. The latter method, and its applications to supramolecular chemistry, are covered in Chapter 15. Each method has its particular advantages. The great ease of preparation, along with their potential for greater stability on the electrode surface, have led to SAMs emerging as perhaps the more accessible technique employing modified electrodes. Nonetheless, some very elegant work has been reported in the area of LB films, which in some instances can yield more tightly packed films that are less subject to defects than are SAMs. In this chapter we shall briefly review LB film preparation and methods employed for film deposition on electrode surfaces. We shall then examine selected topics in the area. A recent review of the area provides a more comprehensive treatment of the field."]
14.1 Langmuir-Blodgett Films The basic methods of Langmuir-Blodgett (LB) film preparation have been described in detail by a number of authors. (The reader is directed to Reference 2, and references therein for a comprehensive description of LB techniques.) In a typical experiment an amphiphilic species, dissolved in a volatile organic solvent) is spread onto the aqueous surface. After evaporation of the organic solvent the surface area is compressed by means of a moving barrier, forming an organized film in which the nonpolar tails of the amphiphile rise above the aqueous subphase, while the polar head groups remain immersed, as shown in Fig. 14.1. Free-floating films on the trough surface are termed Langmuir films. Deposition of these films onto the surface of a substrate
14.1 Lnngrnuir-BlodgettFilms
181
Figure 14.1: Organization of amphiphiles before (a) and after (b) compression in the Langmuir trough.
results in Langrnuir-Blodgettfilms. Monolayer LB films may be deposited on the electrode surface in either of two possible orientations, wh,ich can be described simply as tails in (X type, transfer on the downstroke) or heads in (Y or Z type, transfer on the upstroke) Vertical depositions yielding these two orientations are shown in Fig. 14.2. Subsequent layers may be added by repeated vertical dipping (VD) in either mode. (Note that the hydrophilicity of the substrate may influence the ease of preparation of monolayers in a particular orientation, i.e. heads in Y , Z-type deposition is likely to be easier on a more hydrophilic substrate. Also, VD can only be applied on the rising portion of the pressure-area isotherm.) Electrodes prepared in this fashion are typically immersed in a cell containing supporting electrolyte solution for electrochemical study. Films deposited in this fashion, while initially organized, are subject to relaxation of the packed film, and herein lies one of the chief difficulties of this method for preparing modified electrodes. The organization of the film is degraded after transfer: without the mechanical pressure provided by the floating barrier, the tight packing of the film is quickly lost. Furthermore, the adhesion of the film’s monomeric units to the surface is dynamic. Even if a small amount of the amphiphile dissolves into the cell solution the number of potential pinhole sites may be significant.
182
14
Electroactiue Langmuir-Blodgett Films
t
c Figure 14.2 Different modes of vertical dipping (VD) yield monolayers in two possible orientations: on the left the substrate is dipped into the trough after compression (X mode), on the right the substrate is pulled from the trough after compression (Y or Z mode).
In order to overcome the problems associated with relaxation of the organized film structure one can alternatively employ the horizontal touching (HT) technique. In this method the substrate is gently lowered onto the surface of the trough and the amphiphiles adhere to the substrate in an X (tails in) orientation. The electrode is then raised ever so slightly and surface tension holds the film in contact with the trough solution, as shown in Fig. 14.3. This method permits in-trough electrochemical studies. Because the electrode substrate remains in contact with the trough solution, which is typically maintained at constant pressure, the film deposited on the surface of the electrode is less subject to relaxation. HT can also be used over a wider range of trough conditions than can VD. The majority of studies discussed below will employ HT techniques. For the purpose of in-trough electrochemical studies, the aqueous subphase contains supporting electrolyte. The trough is equipped with a reference and counter electrode. The working electrodes most commonly employed for these experiments are optically transparent electrodes such as tin oxide or sputtered gold on glass slides or mica. In situ spectroelectrochemical studies have been reported by some authors.
14.2 Electron Transfer Studies in Langmuir and Langmuir-Blodgett Filnis
183
Figure 14.3: The horizontal touching (HT) technique avoids some of the problems associated with relaxationof VD films.
14.2 Electron Transfer Studies in Langmuir and Langmuir-Blodgett Films A topic of interest in the area of LB films has been the effects of film environment on the heterogeneous kinetics of electron transfer within the film. A well packed film, free of pinhole defects, poses an almost ideal environment for studying electron transfer to the electroactive amphiphiles in the film. For example, if the film is coated onto the electrode's active and inactive areas does electron transport extend out beyond the active electrode area covered by the film? What role does such lateral electron transport between electroactive sites in the film play and under what conditions is it likely to be observed? What is the maximum distance, e.g. film thickness, at which electron transfer may still occur, without substantial hindrance? Both these topics were addressed in two classical studies, which we shall briefly examine here. Majda and coworkers have examined the issue of lateral electron transfer in some detail.[31 Two mechanisms of electron propagation were studied by these authors- lateral diffusion of the amphiphilic redox species and lateral electron hopping. One of their most elegant studies involved Langmuir monolayers of ferrocenyl amphiphile 1 and lateral electron hopping that is
1
observed when the electrochemical behavior of 1 is probed by means of a thin gold strip microelectrode.~41 At higher pressures lateral diffusion of the
184
24 Electroactive Langmuir-Bfodgett Films
amphiphiles becomes less likely to occur. The observation of currents larger than those expected for the calculated surface coverages under such pressures suggests that lateral electron hopping is the mode of propagation. We should note that this effect has been reported only in electrodes of very small dimensions. Bard and coworkers have reported the electrochemical behavior of films composed of 2 in 1mMNaC104/0.4 M NaS04 (pH 4.5)P Their studies reached several important conclusions about the nature of surfactants resting on the trough subphase and their compression into films. Rather than distributing isotropically over the subphase surface in an idealized “gaseous-like” state, the authors found that 2 tended to form monolayer aggregates of various sizes. This finding suggests that distribution of the surfactant is uneven across the trough surface until the film is considerably compressed, a determination that is significant when one considers preparation of modified substrates. The authors also compared the characteristics of electrodes prepared by HT vs. VD. Electrodes prepared by the HT method showed higher surfaces coverages (near maximum theoretical values) even at lower surface pressures. In contrast, substrates prepared by VD showed the aforementioned loss of surfactant due to reimmersion into the supporting electrolyte solution. The nature of this loss of surfactant at the electrode active area was explored and found to involve not just loss into the solution but expansion of the uncompressed film onto other (electroinactive) areas of the electrode itself. The greater stability of HT films is attributed to the substantial lateral surface pressure exerted on the film by its remaining in contact with the trough film. This lateral contact was further thought to prevent flipping of surfactant molecules, which can contribute to film disorder. Bard has also reported the electrogenerated chemiluminescent behavior of 2 in LB films.PI
Additional interesting results of this study revolved around the electron propagation across bi- and trilayers of 2, which yield some insights into the maximum distance dependence for electron transfer. Multilayers prepared by the HT and VD methods have differing orientation with respect to heads to tails. As shown in Fig. 14.4 this orientation leads to differing voltammetric responses for multilayer films prepared by the two methods. In Fig. 14.4a the current intensity was found to increase with the addition of each new layer (signified on
14.2 Electron Transfer Studies in Langmuir and Langnruir-Blodgett Films
185
the CVs by 1, 2,3 for mono-, bi-, and trilayers of 2 respectively). In contrast, Fig. 14.4b shows the opposite trend: the current intensities indicate that electron propagation cannot take place over the long intervening distance between the bilayer head sites. The slight decrease in current observed in these voltammograms can be attributed to a decrease in the surface concentration of surfactant in the initial layer (due to expansion) with each reimmersion for deposition of subsequent layers. Guo, Facci and McLendon have examined the distance dependence of electron transfer in LB films composed of ferrocenyl amphiphiles. [71
2
a
b
-123
1.6
~
o
1,1
V vs. SCE
0.6
~
1.6
1.1
A
0.6
V vs. SCE
Figure 14.4: CVs of 2 coated on indium tin-oxide electrodes at 30 dyn/cm by HT (a) and VD (B) methods obtained in 1 mM NaC104/0.4 M NazS04 aqueous solution (pH 4.5) at 200 mV/s. The numbers 1, 2, and 3 refer to electrodes coated with mono-, bi-, and trilayers, respective1y.W Reprinted w i t h permission of the American Chemical Society.
Zhang and Bard also examined the cyclic voltammetric behavior of dissolved reactants in the presence of HT prepared mono- and multilayers of
14
186
Electroactive Langmuir-Blodgett Films
2J61 They observed partial blocking of the voltammetric response for Os(bpy)32+ and Fe(CN)b4-. Decreases in current intensity were observed and attributed to hindered mass transfer of the dissolved reactants to the electrode surface. Surprisingly, the blocking effect did not appear to be dependent on the charge of the dissolved species as the suppression of the cyclic voltammetric response was similar for both Os(bpy)32+and Fe(CN)b4.
14.3 Other Electroactive LB Film Studies Facci and Leidner have examined quinone derivatized phospholipid monolayers at the air-water and gold-water interface in some detail. LB methods were employed to prepare monolayers of an anthraquinone derivatized dipalmitoylphospatidylcholine (AQ-DPPC, 3) and mixed monolayers of AQ-DPPC and DPPC.[sl The films were analysed by ellipsometry, which indicated that the molecules transferred at high pressures were in fully extended, heads-in conformation. The authors observed a well behaved cyclic voltammetric response for films containing mixed and pure 3, with the response sensitive to the transfer pressure, supporting electrolyte composition and pH. Expanded films gave the best voltammetric response with small supporting electrolyte anions. Microheterogeneity of more compressed films resulted in more complex voltammetric responses.
3
Fujishima and coworkers have examined the cis-trans isomerization of amphiphilic azobenzene derivatives (4) in LB monolayers.[9] Azobenzene derivatives have been widely studied photoactive molecules due to their profound structural changes exhibited as a result of thermally or photoinduced isomerization. Electrochemical switching of azobenzene (AB) and hydrazobenzene (HAB) films has yielded interesting results, namely that the potentials at which cis-AB films are reduced to HAB are substantially more positive than are those films composed of trans-ABs. This yields the rather interesting result that HAB films are thus reoxidized to the more Reconversion to the cis isomer is thermodynamically stable trans-AB. accomplished by irradiation. This has led the authors to propose that films composed of 4 show potential for mformation storage, due to the clean transformation between states. The isomerization was found to occur
14.3 Other Electroactive LB Film Studies
187
electrostatically, without passage of any apparent faradaic current and with no change in the capacitive current. The electrochemically induced isomerization was found to be indistinguishable from the thermally or photoinduced reaction.PI Liu and coworkers have also examined the voltammetric behavior of cis-azobenzene derivatives in mixed SAM and LB films.[l01 Recently, these authors have reported the voltammetry of LB films composed of a diaza-18crown-6 derivatized azobenzene amphiphile.[llI
n=7,11
4
Nakashima, et al. have reported the transmembrane (or mediated) electron transfer of LB films composed of x-conjugated vinylpyridinium derivatives 5 and 6, whose structure the authors liken to molecular wiresP1 5 and 6 demonstrated the ability to mediate electron transfer to [Fe(CN)6I3-as shown by voltammetry and double potential step chronocoulometric studies. Their results are reminiscent of results obtained for mediated electron transfer to R u ( N H ~ ) ~in+ ~amphiphilic viologen SAM assemblies.l13-141 In a typical experiment the reduction of the solution species (in this case [Fe(CN)6I3-)is prevented by the blocking monolayer. Upon reaching the reduction potential of the monolayer's electroactive moieties however, mediated electron transfer can occur from the monolayer to the solution species via several possible mechanisms.[151 The appearance of the voltammogram is altered from that anticipated for the monolayer in the absence of the solution species. An increase in the cathodic current attributable to reduction of the monolayer, typically proportional to the concentration of the solution species, is observed, while a
188
14 EIectroactiue Langniuir-BlodgettFilms
decrease in the corresponding anodic current (reflecting the decreased number of reduced monolayer monomeric units) is noted. Studies of LB films of rigid rod molecules and polymers have also been of interest in the last decade. The resulting anisotropic films obtained for some structures suggest possible materials applications. Wegner, et al. examined LB films of poly(siloxyphtha1ocyanines) 7, which formed highly ordered multilayers exhibiting anisotropy. The rod-like polymers were found to orient themselves parallel to the substrate surface and dipping direction (rather than approximately perpendicular, as is usually the case)."61 The oxidation of the phthalocyanine rings was characterized as facile although the extent and rate of oxidation was found to be anion dependent.Ll61 Kobayashi, Leznoff, and Shirai have also studied the spectroscopic and voltammetric properties of LB films composed of binuclear phthalocyanines.[151 Miller and coworkers have examined the spectroscopic and voltammetric beeavior of LB films formed from rigid oligoimides with extended lengths to 78 A P ] Based on the appended substituents and symmetry of the rod-like structures varying degrees of anisotropy were observed.
Murray and coworkers were the first to report polymer assemblies employing different redox polymers"91 however the most commonly employed methods in such assemblies have been electropolymerization, which affords little control over orientation and spacing of the redox components. In the area of heterogeneous films, Miyashita and Aoki have reported the electrochemical behavior of mixed LB films composed of ferrocenyl (Fc) and Ru(bpy)s copolymers of dodecylacrylamide.[*0]In their experiments trilayers of the Ru copolymer were deposited on ITOs, followed by a bilayer of the Fc copolymer, as shown in Fig. 14.5. In a good example of mediated electron transfer, the trilayer of Ru copolymer, which is oxidized at more positive potentials, mediates the electron transfer to the ferrocenyl groups. Upon cycling, only the wave attributed to the Ru copolymer is observed. This is attributed to the fact that the Ru copolymer cannot mediate the reduction of the Fc' units, as this process would occur at potentials less positive than that at which the Ru copolymer is
14.4 References
189
I-
Figure 14.5: Cyclic voltammograms for redox polymer LB films on I T 0 electrodes examined at 10 mV/s in 1.0 M NaC104. Voltammogram A corresponds to an I T 0 coated solely with the Fc copolymer, B the two polymers deposited according to the scheme in the upper right hand figure, and C the same electrode as in B, at the second scan of the potential range.[*olReprinted with the permission of the Chemical Society of Japan.
already reduced.[201 Whether the charged Fc' layer remains trapped at the electrode surface has not been clearly established. Finally, in a novel combination of LB and colloidal gold methods Jiang and coworkers have recently reported the stabilization of 9-cis-retinal/digitonin based films by incorporation of colloidal gold particles in the 10 - 50 nm size rangeP1 The authors characterized the films obtained by this method by employing cyclic voltammetry, quartz crystal microbalance and absorption spectroscopy on films deposited on ITOs. Increased photocurrents and enhanced stability of the film were observed.
190
14
Electroactive Langmuir-Blodgett Films
14.4 References 1. L. M. Goldenberg, Russ. Chem. Rev., 1997,66,1033-1052. 2. See for instance.[a] M. C. Petty, Langmuir Blodgeft Films, Cambridge University Press, Cambridge, 1996; [b] G. Roberts, Ed. Langmuir-Blodgett Films, Plenum, New York, 1990. 3. [a] C. J. C Mder, M. Majda, J. Am. Chem. SOC., 1986, 210, 3118-3020; [b] C. J. Miller, C. A. Widrig, D. H. Charych, M. Majda, J. Phys. Chem., 1988, 92,19281936; [c] C . A. Goss, C. J. Miller, M. Majda, 1. Phys. Chern., 1988,92,1937-1942. 4. C. A. Widrig, C. J. Miller, M. Majda, 1.Am. Chem. SOC., 1988,110,2009-2011. 5. X. Zhang, A. J. Bard, I. A m Chem. SOC., 1989,222,8098-8105. 6. C . J. Miller, P. McCord, A. J. Bard, Langmuir, 1991, 7,2781-2782. 7. L.-H. Guo, J. S. Facci, G. McLendon, J. Phys. Chem., 1995, 99,8458-8461. 8. M. D. Liu, C. R. Leidner, J. S. Facci, J. Phys. Chem., 1992,96,2804-2811. 9. T. Enomoto, H. Hagiwara, D. A. Tryk, Z.-F Liu, K. Hashimoto, A. Fujishima, J. Phys. Chem. B, 1997, 102, 7422-7427 and references therein. The authors also provide a useful figure detailing the cell design used in their elegant photoand electrochemical studies. 10. Z. Liu, C . Zhao, M. Tang, S. Cai, J. Phys. Chem., 1996,100,17337-17344. 11.Y. Q. Wang, H. Z. Yu, T. Mu, Y. Luo, C. X. Zhao, Z. F. Lu, J. Electround. Chem, 1997,438,127-131. 12. M Kunitake, K. Nasu, 0. Manabe, N. Nakashima, BulI. Chem. SOC.Jpn., 1994, 67,375-378. 13. S. E. Creager, D. M. Collard, M. A. Fox, Langmuir, 1990,6,1617-20. 14. M. Gomez, J. Li, A. E. Kaifer, Langmuir, 1991, 7,1797-1806. 15. These include the so-called “through space” and ”through bond” mechanisms. Discerning which mechanism is operative is not trivial. 16. A. Ferencz, N. R. Armstrong, G. Wegner, Macromolecules, 1994,27,1517-1528 and references therein. 17. N. Kobayashi, H. Lam, W. A Nevin, P. Janda, C. C. Leznoff, T. Koyama, A. Monden, H. Shirai, J. Am. Chem. SOC.,1994,116,879-890. 18. V. Cammerata, L. Atanasoska, L. L. Miller, C . J. Kolaskie, B. J. S t a h a n , Langmuir, 1992,8,876-886. 19. H. D. Abruiia, P. Denisevich, M. Umana, T. J. Meyer, R. W. Murray, J. Am. Chem. SOC., 1981,103,l-5. 20. A. Aoki, T. Miyashita, Chem. Lett., 1996,563-4. 21. Y. H. Sun, J. R. Lee, B. F. Lee, L. Jiang, Langmuir, 1997,13,5799-5801.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
15. Self-Assembled Monolayers
In the last 15 years one of the main topics of electrochemical research has centered on electrodes covered with organized monolayers. These supramolecular assemblies have allowed researchers unprecedented control on the molecular architecture at the electrode-solution interface and opened several new avenues of research work. A variety of methods have been developed to prepare organized monolayer assemblies on electrode surfaces. LangmuirBlodgett techniques (Chapter 14) rely on forced mechanical organization of the molecules at the air-water interface. By contrast, self-assembly takes advantage of a large number of weak and/or moderate interactions to spontaneously generate organized interfacial structures. Therefore, the term self-assembled monolayers (SAMs) is used to describe organized molecular assemblies (onemolecule thick) whose spontaneozis formation and stability depend on favorable, if weak, intermolecular forces, as well as on forces between each of the individual component molecules and the solid substrate or support. From the standpoint of electrochemistry, the most important class of SAM is formed by the chemisorption of thiolates on gold (although other metals, such as mercury or silver, may also serve as effective support materials). The prototypical thiolate-on-gold SAM is prepared by adsorption of long chain alkanethiol molecules on clean gold surfaces. The main driving forces for this self-assembly process are: (i) the formation of thiolate-gold bonds, and (ii) the favorable lateral interactions among the aliphatic chains of neighboring alkanethiol molecules. Energetically, the thiolate-gold bond has been estimated to release 40-50 kcal/mol.[11 The lateral interactions rely on much weaker van der Waals forces, but their cumulative magnitude is considerable for well-packed monolayers made from long chain alkanethiols. These SAMs are usually prepared by exposing a clean gold surface to a solution of the alkanethiol. For alkanethiols with chains longer than 10-12 carbons, it is generaIIy true that: (1) A stable, organized and densely packed monolayer is readily formed on the gold surface. (2) The thickness of the monolayer is directly proportional to the length of the alkanethiol's chain. (3) The main axes of individual molecules are tilted by an angle of approximately 300 from the normal to the surface. (4) On Au(ll1) crystalline faces (predominant in polycrystalline gold), the sulfur atoms adopt a (d3xd3)R3OUadlayer arrangement. (5) Every chemisorbed molecule loses the hydrogen atom from the thiol group. The surface attached molecules are thus better formulated as thiolates.
192
15
Self-Assembled Monolayers
In spite of the large number of papers published on these monolayers, the detailed mechanism of SAM formation is not so well understood. For instance, the exact nature of the bond between thiolate and gold surface is still the subject of much controversy.[~Jl
Figure 15.1:Schematic representation of a thiolate-on-gold SAM.
The figure illustrates another important feature of SAMs. Synthetic modification of the thiol permits the attachment of functional groups (R) to the monoIayer. Among other applications, this feature can be used to control the wetting properties of the surface, to attach redox active groups at defined distances to the gold (electrode) surface, and to incorporate binding sites that may impart some selectivity to the electrochemical reactions taking place at the electrode-solution interface. We will discuss some of these applications later in this chapter. The structure and properties of SAMs can be characterized by a variety of techniques.[l~~lContact angle measurements, FT-IR spectroscopy, surface enhanced Raman spectroscopy, X-ray photoelectron spectroscopy, ellipsometry and surface plasmon resonance measurements are among the most commonly used techniques. The presence of gold as the substrate for monolayer preparation makes it possible to add several types of electrochemical measurements to the arsenal of characterization techniques.[ZI Not surprisingly, the electrode properties of the gold surface are considerably altered by the deposition of SAMs. For instance, electrode capacitances decrease drastically
15.1 S A M s as Barriers for Electron 7'rans)r Reactions
193
after monolayer formation, and the electrode surface becomes passivated by the SAM (to an extent that depends on monolayer quality and thickness). The decreased capacitance means that the background or charging currents observed in linear scan voltammetric experiments are minimized. On the other hand, the SAM hinders the observation of faradaic currents from the oxidation or reduction of solution species. This is particularly true with hydrophilic electroactive species which are rejected by the monolayer and forced to exchange electrons with the electrode surface through much longer distances, leading to slower rates for the heterogeneous electron transfer reactions. Thiolate-on-gold SAMs have been the subject of several excellent reviews."l The electrochemistry of electrodes derivatized with thiolate SAMs has also been reviewed.L21 The reader is referred to these articles for more comprehensive treatments of the area. In this chapter we will try to summarize some of the research work in which thiolate SAMs have been used as a novel framework for supramolecular concepts.
15.1 SAMs as Barriers for Electron Transfer Reactions In general terms the derivatization of a gold surface with an alkanethiolate SAM results in the attachment of a monolayer of hydrophobic molecules, which acts as a barrier between the electrode surface and hydrophilic electroactive species (Ox) in the solution. We say that the electrode is passivated or 'blocked' by the monolayer. The shortest possible distance between Ox and the electrode surface increases, due to the presence of the SAM, by a length similar to the thickness of the monolayer. The heterogeneous electron transfer process is severely affected by this distance increase, imposing a tunneling mechanism. Actually, in high quality SAMs, the experimentally observed currents follow closely the predictions from tunneling theory.[31 Unfortunately, most SAMs contain a certain density of defects or pinholes, at which electron transfer reactions between Ox and the electrode surface may take place with considerable less hindrance. If the pinhole density is large enough, the blocking effect may even become undetectable. Currents resulting from faradaic reactions at pinholes are affected by mass transfer phenomena while kinetic currents are not. In general terms, the extent of electrode blocking also depends on the nature of the electroactive species, Ox. While hydrophilic metal ions or metal ion complexes are rejected by the monolayer, more hydrophobic organic molecules may perhaps partition with ease into the SAM structure, giving rise to higher current levels for their associated electrochemical reactions. From an experimental standpoint, electrode blocking is typically investigated using two probes: Ru(NH3)b3+and Fe(CN)b4-. Aside from their obvious difference in charges, these two redox active metal complexes also differ in their inherent kinetics of heterogeneous electron transfer reactions. The R u ( N H ~ ) ~ ~redox +/~+ couple exhibits much faster electrochemical kinetics than Fe(CN)63-/4,which, in practical terms, means that the former is a more demanding probe of the monolayer's blocking ability. Generally, the use of Ru(NH3)2+as the probe of
15 Self-Assembled Monolayers
194
monolayer quality is strongly recommended. In some specific cases, Fe(CN)& or Fe(CN)63- (stable salts of both ions are commercially available) might be preferred. For instance, the authors' group has used the later redox couple to estimate the pKa of thioctic acid SAMs.141 This simple disulfide forms rather stable, but poorly packed, SAMs on gold electrodes (we will designate these electrodes as Au/TOA). The cyclic voltammetric response of Au/TOA electrodes in 1.0 mM Fe(CN)63-solutions depends strongly on the pH of the solution. At low pH the thioctic acid molecules are protonated and the response is similar to that observed on bare gold eIectrodes (Fig. 15.2A), as this SAM is rather thin and not densely packed. At high pH values the thioctic acid molecules are ionized and their negative charges establish a coulombic barrier for the approach of F~(CN)C,~to the electrode surface, resulting in effective blocking. By plotting the current (at a fixed potential near the Eo' value) as a function of solution pH, one obtains a sigmoidal curve (Fig. 15.2B) from which the pKa of the chemisorbed thioctic acid molecules can be estimated. It is noteworthy that the pKa value in the SAM was found to be higher than the value
Potentisl, V vs AglAgCI
\-
Figure 15.2 (A) Cyclic voltammetric response of Au/TOA electrodes in 1.0 mM Fe(CN)$-, 0.1 M NaCl at pH 2 (solid line) and at pH 8 (dotted line). Scan rate = 0.1 V/s. (B) Cathodic current at 0.2 V as a function of solution pH. The pH values were adjusted by addition of HCl or NaOH.141 Reprinted with the permission of the American Chemical Society.
15.2
Electroactive Monolayers
195
in homogeneous aqueous solution, a clear effect of the forced proximity of the carboxylic acid groups in the monolayer. It is possible to prepare SAMs with controlled densities of pinholes. For instance, Majda and coworkers[51 have used a mixture of HS(CH2)17CH3 and HO(CH2)17CH3 (70:30) which they pre-organize by compression (in a Langmuir trough) at the air-water interface and subsequently transfer to a smooth gold substrate. The formation of thiolate-gold bonds results in an excellent quality gold-supported monolayer. Mixing of foreign molecules during spreading in the trough leads to the incorporation of 'defects' in the gold-supported monolayer. These defective sites or pinholes are comparable to molecular gates as they provide sites were electron transfer reactions are facile. Vitamin K1 and ubiquinone are examples of foreign molecules useful for the generation of molecular gates. The concentration of foreign molecules during the compression of the monolayer in the Langmuir trough determines the density of pmholes and, thus, the voltammetric behavior observed with the SAM-covered electrodes immersed in R u ( N H ~ ) solutions. ~~+ In related experiments, Chailapakul and Crooks[61 have shown that the co-adsorption of thiophenol derivatives and HS(CH2)15CH3 on gold surfaces leads to SAMs with controlled number of pinholes, as their surface density is determined by the [thiophenol]/ [alkanethiol] ratio in the deposition solution (see Fig. 15.3)
Figure 15.3: Schematic representation of pinholes (marked by arrows) generated by coadsorption of thiophenol in a HS(CH&sCH3 SAM.
15.2 Electroactive Monolayers SAMs have also been used as barriers for electron transfer reactions of redox active residues covalently attached to some of the thiolates in the monolayer
15 Self-Assembled Monolayers
196
assembly. This allows a more precise positioning of the redox active centers and essentially eliminates the currents associated with pinholes or monolayer defects. On the other hand, new problems arise due to interactions among the monolayer-anchored redox units. This potential problem may be addressed by diluting the redox active thiol with redox inactive thiols. However, it is not entirely clear that this approach always produces the desired result, that is, the isolation of the redox active residues from one another. In any instance, electroactive SAMs show reversible electrochemical behavior when the currents are recorded with a scan rate slow enough so that the response is controlled by thermodynamic, not by kinetic, factors. Since the electroactive centers are all close to the electrode surface, diffusion does not have any influence in the experiment. Cyclic voltammetric responses for surface confined redox centers are given by the following equationsP I
15.1
where ip-
n2F2AT V 4RT
nF RT
8 = -(E
-Eo')
15.2
15.3
Unlike in experiments with diffusion controlled currents, the peak current zp is directly proportional to the scan rate, v. The surface coverage, r, is usually expressed in mol/cm2 units and can be determined by integration of the voltammetric wave current. The remaining symbols have their standard meanings. In these systems, the formal potential can be obtained from the average of the anodic and cathodic peak potentials. The difference between the two peak potentials (AEP) is theoretically zero for surface confined electroactive species. In practice, small values are found due to residual uncompensated resistance and/ or interactions between the redox centers. The voltammetric waves in these systems should in theory be symmetrical with a width at halfheight ( A E h h ) of 90.6/n mV at 2 5 C In these SAM systems, the most extensively studied redox subunit is probably ferrocene, due to the synthetic accessibility of its derivatives. Starting with Chidsey's pioneering electron transfer studies,[q a good number of papers have reported on the electrochemical behavior of ferrocene centers covalently attached to SAM components. Chidsey reported on the electron transfer reactions between monolayer-anchored ferrocene residues and the underlying gold electrode surface. The ferrocene centers were attached via ester linkages to the ends of (CH2)lbS thiolate chains. The SAM system was prepared by the coadsorption of these ferrocene-terminated thiols with excess CHs(CH2)15SH as
15.2
Elecfroactive Monolayers
197
diluent. The electrochemistry of the ferrocene centers was found to be almost ideal as the AEp values were very close to zero and AEfwhm were -95 mV at slow and moderate scan rates. The kinetics of electron transfer was studied by monitoring the chronoamperometric responses of the SAM system to a range of potential step perturbations. The data could be quantitatively explained using Marcus theory.[q
Figure 15.4 Schematic representation of Chidsey's experiments.
This work was the first of many reports on electron transfer between monolayer-anchored redox subunits and the underlying electrode. SAMs provide a nice molecular framework for the controlled placement of electroactive centers at well-defined distances from the electrode surface. Therefore, many studies have focused on the distance dependence of heterogeneous electron transfer rates. Often in these SAMs, the electron transfer rates are determined by the electronic c o u p h g between the electroactive center and the electrode surface. Thus, the observed rate constant, k, is expected to decrease exponentially with the distance, d, according to the equation:
k = k" .exp(-o.d)
15.4
where p gives a quantitative measure of the distance dampening effect on the electron transfer process. Usually, p values determined experimentally hover around 1.0 A-1, but substantially higher or lower values have also been obtained. Other reports have focused on using SAM structures to control the environment of covalently attached redox units. Rowe and Creagerf81 investigated the behavior of ferrocene buried to different depths in alkanethiolate SAMs and found that the formal potential for ferrocene oxidation shifted to more positive values as the ferrocene residue was buried deeper into the monolayer (see Fig. 15.5). This reflects the poor solvation that the oxidized
15 Self-AssembledMonolayers
198
form of ferrocene experiences inside the SAM'S hydrophobic interior. These authors have also investigated the effects of ion pairing on the voltammetric behavior of monolayer-anchoredferrocene residues.[8]
E(V) vs. Ag/AgCl sat. KCl Figure 15.5: CV responses in 1.0 M HClOl of gold electrodes coated with the mixed SAMs shown at right.P] Scan rate 0.1 V/s. S = 1.0 PA. Reprinted with the permission ofthe American Chemical Society.
15.3 Molecular Recognition in SAMs Electrochemical techniques generally suffer from poor analytical selectivity. Although the electrode potential can be adjusted to control which
25.3 Molecular Recognition in SAMs
199
species will undergo electrochemical reactions at the electrode surface, only limited selectivity can be derived in this way. For more than two decades one of the main topics of electrochemical research has centered on the derivatization of electrodes surfaces with predominantly organic layers. Among the key ideas fueling this research work has been the possibility of using the modified organic layer as permselective membranes to increase electrode selectivity by allowing only certain species to approach the electrode surface and, thus, dictating which species will react electrochemically. SAMs are indeed very attractive as permselective membranes since they can be readily prepared as densely packed assemblies. Good-quality SAMs passivate the electrode surface effectively. Therefore, the key to permselective properties is the incorporation into the SAMs of molecular channels or binding sites that may provide the required selectivity to the electrode through molecular recognition of specific analytes. The pioneering work of Rubinstein and coworkers opened the utilization of SAMs for these purposes. In 1988 this group published an innovative report on the preparation and voltammetric behavior of gold electrodes modified with mixed monolayers of octadecanethiol and 2,2‘thiobisethyl acetate (TBEA).L91 The former thiol acts as an inert diluent and the latter provides recognition sites for specific metal ions (Cu”). These authors demonstrated that the monolayer-covered electrode exhibits selectivity properties similar to those of the active component (TBEA) and affords an electrochemical response to Cu2+ions in the presence of competing Fe3+ions.
Figure 15.6 Rubinstein’sSAMs for selective metal ion detection.
200
15
Self-Assembled Monolayers
These authors have continued their work in this area and found that selective ion binding is crucially dependent on the applied potential.[lO] More recently, Reinhoudt and coworkers have exploited the properties of crown ethers and prepared SAMs containing thiolates derivatized with pendant crown ether subunits such as those represented below.[lll
Figure 15.7:Reinhoudt'scation-responsive SAMs.
These monolayers can be utilized for the detection of electroinactive cations using impedance spectroscopic methods. The complexation of alkali metal ions by the monolayer crown subuoits can also be detected using Ru(NH~)~ as~ a+ probe. For instance, a SAM prepared from the 12-crown-4 derivatives shown in Fig. 15.7 does not block the voltammetric response of the underlying gold electrode to R u ( N H ~ ) ~ ions. ~ + However, in the presence of NaCl, effective blocking is observed, as the bound Na' ions improve the barrier properties of the monolayer towards the positively charged Ru(II1) complex ions. These SAMs exhibit selectivity properties that differ somewhat from those found with individual crown ethers in homogeneous solution. Monolayers of 12-crown-4 derivatives detect Na+ in the presence of a large excess of K', but SAMS of the 15-crown-5derivatives (at right in Fig. 15.7) are selective to K' ions. The authors interpreted these findings as a result of the accumulation of crown ethers on the monolayer-solution interface, which favors the formation of 2:l (crown-to-ion) complexes.[1lI Dilution of the 15-crown-5 ether thiolate derivatives with heptanethiol was found to decrease the K+/Na+selectivity of the SAM.
15.3 Molecular Recognition in S A M s
201
The Reinhoudt group has also reported extensively over the last six years on the properties of gold electrodes covered with SAMs prepared from sulfide-functionalized cavitands and other sulfur-containing hosts. [I*] These monolayers have been shown to exhibit some selectivity for guest binding from gas and solution phases. However, from the standpoint of binding electroactive guests, larger cavity hosts must be assembled at the electrode-solution interface. In this regard, the authors' group reported in 1995 the preparation and properties of gold electrodes derivatized with per-6-thio-~-cyclodextrin(t-p-CD), shown below. [I31
This cyclodextrin derivative was found to chemisorb on gold by forming the expected thiolate-gold bonds. The chemisorbed receptor faces the gold surface with its narrow opening while the wider opening (with the secondary OH groups) is exposed to the contacting solution. This arrangement should optimize the receptor's interfacial binding efficiency. However, the maximum surface coverage of t-p-CD molecules on gold is limited to about -75% of the gold surface due to the lack of favorable lateral interactions between the immobilized cyclodextrins. Subsequent exposure of the Au/ t-p-CD electrode to pentanethiol in the presence of ferrocene (to protect the CD cavities) improves the total surface coverage while maintaining most of the cyclodextrin cavities ready for binding. This was verified by reductive desorption, electrode capacitance and blocking measurements. After washing and removal of all the ferrocene used in the preparation of the monolayers, the resulting mixed SAMs formed from t-PCD and pentanethiol bind ferrocene when this electroactive guest is added to the contacting solution at concentration levels in the range 1-60 pM. Surfaceconfined waves were clearly observed for the reversible oxidation of ferrocene. Control experiments showed that the observation of ferrocene surface waves is a direct result from the presence of 0-CD cavities in the interfacial SAM structure. Competition studies between two guests were also performed. Both ferrocene and rn-toluic acid are excellent guests for binding by p-CD hosts, but the former is electroactive and the latter is electroinactive in the potential range surveyed in this work. Therefore, gradual addition of m-toluic acid to a solution containing 5 pM ferrocene leads to the progressive disappearance of the ferrocene voltammetric response, as the electroinactive rn-toluic acid molecules displace the ferrocene molecules from the available binding sites in the SAM (Fig. 15.8). The adsorption (or interfacial binding) of ferrocene was mathematically treated according to the Langmuir isotherm to determine the binding constant between the ferrocene molecules and the surface-immobilized t-0-CD receptors.[l31 The
15 Self-Assembled Monolayers
202
value obtained (4 x lo4 M-1) is about an order of magnitude larger than the values normally measured in homogeneous solution for binding of ferrocene derivatives by p-CD hosts. The reasons for this discrepancy are still unclear but may be related to the spatial accumulation of binding sites in the SAM structure.
1
-'-'
POTENTIAL. V
vs
SSCE
O.'
Figure 15.8: CV response of a gold electrode modified with a mixed SAM (t-O-CD + pentanethiol) at 0.5 V/s in 0.2 M Na2S04 also containing (A) 5 pM ferrocene, (B) 5 pM ferrocene + 5 pM rn-toluic acid, and (C) 5 pM ferrocene + 7.5 pM m-toluic acid.[ls] Reprinted with the permission of the American Chemical Society.
Although the immobilization of synthetic hosts in SAMs holds substantial promise for the development of new types of sensors, it is important to realize that these hosts do not afford selectivity properties as favorable as those exhibited by biological molecules, such as antibodies or enzymes. Can
15.4 Photoswitchable SAMs
203
thus we expect that this research work may eventually lead to the development of sensors with enough selectivity to be useful in practical analytical situations involving complicated samples? The answer is not easy but the work of Crooks and coworkers holds some important clues for the future of this fieldP1 This group has published extensively on the use of functionalized SAMs as sensor surfaces. Their approach is not based on the development of sufficient selectivity to bind a particular analyte in a SAM while rejecting all possible interfering species. They have rather proposed the use of multiple functionalized surfaces in what amounts to an array of sensors having different properties. In this approach, even if each SAM exhibits poor selectivity, the combination of all the responses from the various sensing surfaces in the array, coupled with the use of pattern recognition software, may yield the sought-after analytical data, i.e., the composition of the sample regardless of its complexity. Quite recently, this group has also developed methodology for the attachment of dendrimers to carboxylate-terminated SAMsPI Interestingly, the surfaceimmobilized dendrimers constitute a barrier in which only molecules with the right size to fit in the dendrimer's empty spaces are incorporated. This method provides an alternative and very promising way for the preparation of sensing surfaces that show reasonable selectivity.
15.4 Photoswitchable SAMs The idea of combining the high degree of molecular organization found in SAMs with the spatial confinement of their components to create switchable molecular assemblies has driven the research work of many groups over the last few years. Let us assume that an external stimulus, capable of altering the chemical composition of the SAM, can be translated into differing rates of interfacial electron transfer. In other words, this means that the SAM-covered gold electrode becomes a transducer in which the external stimulus is converted to electrons. Photoswitchable SAMs are the most prominent in this regard. Willner and coworkers have reported extensively on the general idea of using light to control the behavior of electrochemical interfaces modified with photoswitchable SAMs.rl51 In 1994, these authors reported on photoswitchable SAMs that offer light-control on the rate of electron-transfer between the protein cytochrome c and the underlying electrode surface. It is well known that cytochrome c undergoes relatively fast electron transfer on an electrode covered with 4-mercapto pyridine. These authors assembled a mixed monolayer of cystamine and 4-mercapto pyridine and attached covalently nitromerocyanine to the cystamine molecules. In this form, the nitromerocyanine is positively charged and repels the positively charged side of cyctochrome c that contains the redox active, iron-heme group of the protein. Therefore, no electron transfer between the protein and the electrode is observed. Irradiation of the SAM with visible light (A 2 475 nm) transforms the nitromerocyanine groups to nitrospiropyran groups, which are neutral, giving rise to faster and observable electron transfer between the protein and the electrode (see Fig. 15.9).
204
15
Self-Assembled Monolayers
Illumination at shorter wavelengths (360 < h < 400 nm) reverts the process, yielding again the positively charged, nitromerocyanin form of the SAM which shuts down the electron transfer.process.
360
I. < 400 nm
t
h > 475 nm
electron transfer
Figure 15.9 Willner's photoswitchable SAM. Photochemical control of communication between the electrode and cytochrome c.
the
Very recently the Willner group has also reported a method to photochemically imprint molecular recognition sites in SAMs.[l71Their approach relies on the preparation of mixed SAMs from tetradecanethiol and a thiolfunctionalized naphthacene quinone. Irradiation of this monolayer (320 < h < 380 nm) photoisomerizes the quinone units to their "anal'-quinone forms (Fig. 15.10), which in turn can be removed from the monolayer by reaction with free amines, leaving behind well-defined empty sites which can recognize diffusing
15.4
Pkotosruitcliable SAMs
205
naphthacene quinone molecules. These sites reveal high selectivity in their recognition ability as they do not show any significant binding affinity for structurally related molecules.
Figure 15.10: Willner's method for the photochemical imprinting of recognition sites in SAMs.
These examples illustrate some of the possibilities that are being explored with photoswitchable SAMs. Other groups have used light to fabricate two-dimensional patterns on monolayer assemblies. While basic research in SAMs appears to have reached maturity, this area is still very active and new developments continue to appear in the literature.
206
15
Self-Assembled Monolayers
15.5 References 1. (a) L. H. Dubois and R. G. Nuzzo Ann. Rev.Phys. Chem. 1992,43,437-63. (b) G. M. Whitesides and P. E. Laibinis Langmuir 1990, 6,87-96. 2. (a) H. 0. Finklea in Electroanalyfical Chemistry, A. J. Bard and I. Rubinstein, Eds.; Dekker: New York, 1996; Vol. 19, pp109-335. (b) D. Mandler and I. Turyan Electroanalysis, 1996, 8, 207-213. 3. C. Miller, P. Cuendet and M. Gratzel, J. Phys. Chem. 1991,95,877-886. 4. Y. Wang and A. E. Kaifer 1. Phys. Chem. B 1998,102,9922-9927. 5. R. Bilewicz, T. Sawaguchi, R. V. Chamberlain I1 and M. Majda Langmuir 1995, 11,2256-2266. 6 . 0 . Chailapakul and R. M. Crooks Langmuir 1993,9,884-888. 7. C. E. D. Chidsey Science 1991,251,919-922. 8. G. K. Rowe and S. E. Creager Langmuir 1991, 7,2307-2312. 9. I. Rubinstein, S. Steinberg, Y. Tor, A. Shanzer and J. Sagiv Nature 1988, 332,426-429. 10. S. Steinberg, Y. Tor, E. Sabatani and I. Rubinstein J. Am. Chem. SOC.1991, 113, 5176-5182. 11. S. Flink, B. A. Boukamp, A. van den Berg, F. C. J. M. van Veggel and D. N. Reinhoudt J. Am. Chem. SOC.1998,120,4652-4657. 12. See, for instance: E. U. T. van Velzen, J. F. J. Engbersen, P. J. de Lange, J. W. G. Mahy and D. N. Relnhoudt 1. Am. Chem. SOC.1995,117,6853-6862. 13. M. T. Rojas, R. Koniger, J. F. Stoddart and A. E. Kaifer J. Am. Chem. SOC.1995, 117,336-343. 14. For reviews, see: (a) R. M. Crooks and A. J. Ricco Acc. Chem. Res. 1998, 31, 219-227. (b) A. J. Ricco, R. M. Crooks and G. C. Osbourn Acc. Chem. Res. 1998, 31,289-296. 15. I. Willner Acc. Chem. Res. 1997, 30,347-356. 16. M. Lion-Dagan, E. Katz and I. Willner J. Chem. Soc., Chem. Commun. 1994, 2741-2742. 17. M. Lahav, E. Katz, A. Doron, F. Patolsky and I. Willner J.Am. Chem. SOC.1999, 121,862-863.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
16 Electroactive Dendrimers
During the past few years one of the most active areas of research in chemistry has focused on dendrimers. This term describes a special type of macromolecules that are built from a central core and expand in three dimensions forming globular structures with well-defined surfaces.il1 Quite often, dendrimers are prepared layer by layer using a repetitive synthetic sequence. Therefore, the term generation was coined to describe the extent of growth of these macromolecules as measured by the number of synthetic iterations needed for their preparation. Typically, a first generation dendrimer is a rather simple molecule with 3 or 4 identical functional groups on its periphery, while a fourth or fifth generation dendrimer may contain nearly 100 tightly packed functional groups and have a molecular mass in the range of 10-100K daltons. A distinguishing feature of dendrimers as compared to linear chain polymers is that they typically exhibit a much higher degree of monodispersity. As supramolecular chemists strive to prepare nanometer scale structures, dendrimers have become a natural framework for this type of work. In this chapter, we will limit ourselves to review some of the recent developments involving electroactive dendrimers, as well as dendrimer use in electrochemistry. We define electroactive dendrimers as those that contain functional groups capable of fast electron transfer reactions. We can class+ all electroactive dendrimers into two general classes: (i) Dendrimers with peripheral electroactive groups, and (ii) dendrimers with internal electroactive groups. In the second class, the most common situation arises when the central core of the dendrimer is the single electroactive group. However, the literature contains several examples of dendrimers with multiple internal electroactive groups.
16.1 Dendrimers with Peripheral Electroactive Groups These dendrimers are prepared by attaching a number of identical electroactive functional groups to the periphery of a dendritic macromolecular framework. Each dendrimer macromolecule is thus capable of exchanging mxn electrons, where m is the number of electroactive residues and n is the number of electrons exchanged per residue in the corresponding electrochemical reaction. The multielectron character of the electrochemical reactivity of these dendrimers might find use in catalysis or in other applications involving
208
I6
Electronctive Dendrimers
me
Polar, cationic dendrimer
Nonpolar, neutral dendrimer
Figure 16.1:Polarity changes experienced upon electron transfer reactions by dendrimers containing peripheral electroactivegroups.
chemical amplification, although these possibilities have not been explored yet. One of the key problems in this type of work is that the polarity and solubility properties of these dendrimers change drastically as a result of the multiple electron transfer reactions that they can experience, hindering their use in catalytic cycles. Bryce and coworkers reported one of the first electroactive dendrimers of this class in 1994.[21 Using a convergent synthetic approach relying on a repetitive coupling/ deprotection sequence, they prepared the symmetric dendrimer 1, which has 12 tetrathiafulvalene (TTF) units in the periphery, as well as several other related TTF-containing macromolecules. Dendrimer 1 exhibits the electrochemical behavior typical of TTF, with two quasi-reversible oxidations in acetonitrile solution at half-wave potentials of 0.43 and 0.81 V vs Ag/AgCl. The authors concluded that there is no significant interaction between the TTF units in any of their oxidation states. In a later paper,c31 they report thin layer cyclic voltammetric data confirming that all TTF subunits are oxidized. One of the most interesting properties of TTF is its ability to form charge transfer complexes with electron acceptors. The TTF cation radical (TTF") dimerizes, giving rise to a characteristic band at 830 nm. The association of TTF" to form dimers and ordered stacks is of considerable interest as they give rise to organic solids with electronic conductivity. These authors explored the behavior of oxidized 1 and similar TTF-containing dendrimers and found spectral evidence for inter-dendrimer interactions relying on TTF+'- TTF" contacts. The group of Cuadrado and Moran have also prepared several classes of dendrimers with organometallic subunits on their periphery. In 1995 these authors reported on two novel dendrimers with a silicon backbone and 4 or 8
16.1
Dendrimers with Peripheral Elecfroactive Groups
209
O Y 0
g '0 0
3 0
I 7TF
I TTF ~ T F
STF
equivalent ferrocene subunits on the periphery.[41 Voltammetric results in CHzC12 solution indicate that the oxidized forms of these dendrimers deposit on the electrode surface forming stable electroactive films. Platinum electrodes modified in this fashion could be transferred to fresh solutions containing no dendrimers, yielding persistent voltammetric responses. A similar class of dendrimers with peripheral ferrocene subunits has been utilized by Astruc and coworkers to demonstrate the so-called dendritic effect in molecular recognition interactions.151 Oxidation of the nine ferrocene groups of dendrimer 2 yields a species with a 9+ charge, which is capable of anion recognition. Cyclic voltammetry of 2 in CH2C12 shows a single anodic reversible wave at E1p = 0.69 V vs SCE. Titration with n-Bu4N+H~P04-,for instance, leads to the gradual development of a new wave at less positive potentials at the expense of the original wave. The replacement of the initial wave by the anion-induced one is completed after the addition of 1 equiv of nBu4N+H2P04-per ferrocene group. This reveals that the oxidized dendrimer binds the anions much more strongly than the reduced dendrimer. A more quantitative analysis of the data shows that the spatial accumulation of positive charges on the dendrimer surface (dendritic effect) leads to substantial increases of the apparent binding constants compared to those that would be measured with isolated ferrocene receptors. Even larger dendritic effects are observed with a higher generation dendrimer analog possessing 18 ferrocene groups in its surface. Since anion recognition results from the combination of hydrogen
210
16
Electroactiue Dendrimers
0
2
bonding to the amidic N-H and coulombic attraction between the anion and the positive charge of the oxidized ferrocenium residue, the binding ability and the dendritic effect are larger for H2P04- and HS04- than for C1- and NOs-.[5l These results open interesting research avenues for the use of dendrimers as active components in electrochemical sensors. The group of Cuadrado and Mordn has also reported on another series of dendrimers possessing 4,8,16,32 and 64 peripheral ferrocene groups.[61 This series of macromolecules was built around poly(propy1eneimine) frameworks. The electrochemistry of these dendrimers reveals the non-interacting character of all the ferrocene subunits. In nonpolar solvents such as CH2C12, the oxidized form of the dendrimers was found to deposit as robust films on the working electrode surface. These dendrimers were also shown to act as multisite guests for complexation by P-cyclodextrin hosts,[? yielding large supramolecular assemblies in aqueous solution that could be disrupted by electrochemical or chemical oxidation of the ferrocene residues. The same group has also reported a unique example of silicon-based dendrimers that exhibit a substantial extent of interaction among their peripheral ferrocene For instance, the cyclic voltammetry of 3 (see structure in the next page) in CH2Clz:CHKN (5:l v/v) reveals two distinct waves (separated by 190 mV) for the oxidation of the peripheral ferrocene
16.1
Dendrimers with Peripheral Electroactive Groups
4.0
f 1Vvr.SCE
211
Q.Q
Figure 16.2: Cyclic voltammogram of 3 in CH$&/CfiCN (5:l) v/v with 0.10 M TBAPF,j.[al Reprinted w i t h permission of the American Chemical Society.
212
16 Electroactive Dendrimers
groups. This finding indicates that the two ferrocene centers attached to the terminal silicon atom in each of the dendritic wedges undergo oxidation at different potentials, due to their strong mutual interactions. Dendrimer 5 and a higher generation analog (16 ferrocenes) deposit on the working electrode surface upon oxidation in CHzClz, yielding electroactive films which maintain the same kind of two-wave voltammetric response. A recent report from the same authors (working in collaboration with the authors’ group) has focused on the synthesis and electrochemistry of dendrimers based on the poly(propy1eneimine) backbone and containing 4, 8, 16, and 32 cobaltocenium subunits on their (the structure of the tetrameric dendrimer is shown in Fig. 16.3.). These cationic dendrimers were isolated as their hexafluorophosphate salts. The voltammetric behavior of these dendrimers in aqueous media resembles that of simple cobaltocenium, which undergoes a fast one-electron reduction to yield cobaltocene. This neutral and hydrophobic compound is not soluble in water and, therefore, precipitates on the working electrode surface. The shape of the anodic peak in the reverse voltammetric scan, a very sharp spike, reflects the stripping and re-dissolution of the deposited cobaltocene upon oxidation. In spite of this complication, the reduction of all the cobaltocenium subunits in these dendrimers takes place in a single, if distorted, wave. In other words, all the cobaltocenium subunits are independent or non-interacting. We have shown previously (Chapter 7) that
Figure 16.3: Electrochemically-driven binding of cobaltocene dendrimers by p-CD hosts.
7 6.2
Dendrimers with Internal Electroactive Groups
213
cobaltocene forms a stable inclusion complex with p-CD, while cobaltocenium is not bound by this host. Are these binding interactions maintained between the peripheral subunits of these dendrimers and the p-CD hosts in the solution? Voltammetric data obtained with the cobaltocenium-containing dendrimers (4, 8, and 16 residues) in the presence of p-CD strongly suggest that, while the oxidized dendrimers do not interact with the CD hosts, multiple inclusion complexes are formed upon dendrimer reduction to the cobaltocene form (see Fig. 16.3). This constitutes an excellent example of large molecular weight supramolecular assemblies that are formed only after appropriate electrochemical stimuli are applied to the system.
16.2 Dendrimers with Internal Electroactive Groups The previous section has shown that the surface functionalization of dendrimers with redox active residues yields macromolecules capable of multielectron transfer reactions. A general observation common to the majority of these systems is that the electrochemical properties of the dendrimers resemble those of the monomeric redox active species. Furthermore, the electroactive centers usually behave as non-interacting subunits. In contrast to these findings, one would anticipate that internal electroactive residues should exhibit electrochemical properties reflecting their environment and, thus, be more sensitive to the surrounding dendritic backbone. In fact, this is often the case. In some instances (vide infra), analogies can be established between dendrimers having internal electroactive groups and redox proteins. Furthermore, internal redox active groups provide interesting possibilities regarding electric charge and information storage in these macromolecules. Diederich and coworkers pioneered the exploration of these issues by preparing dendritic systems around a metal porphyrin core.[101 The two dendrimers shown below are representative structures of the two series of macromolecules synthesized by this group.1111 Dendrimer 4 has a zinc porphyrin core and a lipophilic surface, which renders it soluble in nonpolar solvents, such as THF and CH2C12. On the other hand, 5 has a Fe core (with an additional chloride in the fifth coordination site) and more hydrophilic peripheral groups that make it water soluble. The third generation analog of 4 has 108 methyl carboxylates on its surface and a molecular weight of 19,054 daltons. Computer modeling suggests that it has a dense structure and globular shape, with a diameter of ca. 4 nm, a size similar to that of the redox protein cytochrome c. These dendrimers exhibit half-wave potentials for the first porphyrin-centered oxidation and reduction processes that shift to more negative values with increasing dendrimer generation.[lOJlI This finding is counterintuitive, because it means that the generation of positive charge on the porphyrin residue is thermodynamically favored by the growth of the surrounding dendritic mass. The voltammetric behavior of these dendrimers also reveals that the kinetics of the electron transfer reactions slow down as the dendrimer generation increases.
214
16
Electroactive Dendrimers
Y
O M 0
e
The series of iron dendrimers was also studied in aqueous solution.[111 In going from dendrimer 5 to its next generation analog the potential for the Fe(III)/Fe(II)redox couple becomes more positive by 420 mV. This finding was rationalized by the increased hindrance experienced by water molecules as they solvate the more highly charged Fe(II1) oxidation state through the increased dendrimer mass. These same Fe dendrimers showed only small potential differences when their electrochemical behavior was recorded in CH2Ch. In collaboration with Collman's group, these authors have also reported on the 0 2 and CO binding affinity of the Fe(I1) forms of these dendrimersP21 (with imidazole replacing chloride at the fifth coordination spot). They found that the oxygen affinity is 1,500 times higher than that of hemoglobin, while the CO affinity is similar. Both diatomic gases were bound reversibly by the Fe(I1) porphyrin dendrimers. Aida and coworkers have also reported on the electron transfer reactions,[131 and 0 2 and CO binding of Frechet-type, porphyrincontaining dendrimers"41. Recently, Gorman and coworkers have described the synthesis and properties of a novel series of dendrimers"51 built around a redox active ironsulfur core of the form [Fe4S4(SR)4I2-.This core exhibits a quasi-reversible oneelectron reduction. The voltammetric behavior of these dendrimers demonstrates that the reduction processes become kinetically and thermodynamically hindered with increasing dendrimer generation (zeroth to fourth). This finding was interpreted to result from the increased steric bulk imposed by the dendritic ligands R.
16.2 Dendrimers with Internal Electroactive Groups
215
Figure 16.4 The basic concept of Newkome's dendrimer systems, which contain a single bis(terpyridyl)Ru(II) center.
Newkome and coworkers have developed an interesting approach to the assembly of dendrimers that relies on the coordination of metal ions.[161 The main idea is represented in Fig. 16.4. Two terpyridine-derivatized dendritic wedges can be assembled into a larger structure by using an appropriate metal ion to which the terpyridines will act as ligands. These authors have used Ru(I1) for the preparation of a series of dendrimers containing either a single[161 or several[171bis(terpyridyl)Ru(II) centers. For the systems with a single metallic center, the kinetics of the electron transfer reactions was found to slow down as the dendritic wedges became bulkier and surround the metal complex. The group of Balzani, at the University of Bologna, has also made extensive use of transition metal coordination in the design and preparation of many novel dendrimers.['s] Their systems are completely synthesized relying on metal coordination, using the directionality of ligand-to-metal bonds for the structural design of the dendrimers. The main building blocks used in this approach, along with Ru(I1) and Os(I1) nucleating metal ion centers, are illustrated in Fig. 16.5. The resulting dendrimers can be considered as organized assemblies of coordinated metal centers. In addition to very interesting spectroscopic properties, these dendrimers exhibit rich electrochemical behavior related to metal-centered one-electron oxidations or ligand-centered reductions. The magnitude of the interactions among the metal centers varies from small to considerable depending on their relative proximity. Therefore, the electrochemical behavior can be predicted if one considers these effects as well as others resulting from the macromolecular nature of these systems. As an illustrative example, let us look at the behavior of the decanuclear Ru(I1) dendrimer 6. This dendrimer contains three kinds of Ru(I1) centers: One in the exact center of the molecule, three internal ones, and six peripheral ones. Since
216
16
Electroactive Dendrimers
\
/
, /
Terminal ligands
Bridging ligands
bPY
Q-Me
Q / -\ ( = = \
2,3-Medpp+
Figure 16.5: Ligands used by Balzani and coworkers for dendrimer assembly.
120+
bpy is known to be a better electron donor than 2,3-dpp, the six peripheral. centers are expected to undergo oxidation at less positive potentials than the four more deeply buried redox centers. The differential pulse voltammogram shows only one anodic peak at 1.53 V vs SCE that involves six electrons. The
16.2
Dt-ndrirners with Internal Electroactive Groups
217
oxidation of the internal Ru(I1) centers was not observed probably due to the building u p of positive charge in the periphery of 6 which hinders the extraction of more electrons from the macromolecule, shifting the oxidation potentials beyond the accessible potential window. Synthetic replacement of the central Ru(I1) center by a more easily oxidizable Os(I1) results in a differential pulse voltammogram exhibiting the same six-electron oxidation of the peripheral Ru(I1) units at 1.53 V plus a poorly resolved shoulder at 1.35 V corresponding to the oxidation of the central Os(1I) metal center. The utilization of different ligands and metal ions makes it indeed possible to design and establish electrochemical potential gradients in these dendrimers. The generation of radial electrochemical potential gradients is crucial to afford directional electron transfer properties in dendritic macromolecules, an important feature for the development of materials with useful charge and/or mformation storage properties. Selby and Blackstock have recently reported on a new redox active polyarylamine dendrimer (7,see structure in the next page) that exhibits a radial redox gradientJ91 The cyclic voltammogram of 7, as shown in Fig. 16.6, exhibits three well-defined anodic waves. The first two (at halfwave potentials of 0.48 and 0.63 V vs SCE) involve one and two electrons and correspond to the oxidation of the three internal p-phenylenediamine residues. The last wave (at 0.88 V vs SCE) involves three electrons and is assigned to the oxidation of peripheral arylamino groups. Dendrimer 7 exhibits other irreversible oxidation waves at more positive potentials. The gradient of
Meo\O_NMoMe
OMe
I
I 7
OMe
OMe
218
16
Electroactive Dendrirners
-a, a0 -16.
an
- 1 p. OD
-B. aU
-4.130 0.00
4, DO
Figure 16.6: Cyclic voltammogram of 7 in CH2C12/0.10 M TBABF4 at 25 C; the scan rate l with permission of the American Chemical Society. was 200 m V / ~ . " ~Reprinted
electrochemical potentials present in this molecule leads the authors to hypothesize that oxidation of the core to yield ' 1 should be a facile process, while reduction of 1+ must be hindered. Measurements of the self-exchange electron transfer rate constant for the 1+/1 couple yield a value of 1.8 x lo5 M-1 s-1. Self-exchange electron transfer rate constants for simpler, unprotected p-phenylenediamine derivatives were found to show values consistently above 1x l o 9 M-1 s-1. This difference was attributed to the countergradient charge transfer barrier imposed by the peripheral arylamino groups in the dendrimer. One feature that is commonly observed in redox proteins is the unsymmetric positioning of the prosthetic group. Often, the redox active center is partially buried in the polypeptide backbone and unsymmetrically located "off center" in the protein structure. This is one of Nature's methods to control electron transfer reactivity in biological systems, favoring reactions with desirable partners and preventing reactions with unwanted partners. For instance, in cyctochrome c, the redox active iron-heme group is buried in the backbone, but remains close to several surface lysine residues that are positively charged at physiological pH values. This protein undergoes rather facile electron transfer with appropriate partners that interact with the lysinedz01 However, reactions with other partners, those that approach the protein from other directions, are more sluggish. Recently, the authors' group begun to use dendritic frameworks to express this type of directional redox reactivity. Using ferrocenecarboxylic acid as the redox active nucleus and Newkome's synthetic
16.2
Dendrimers with Internal Electroactive Groups
219
methodology, we synthesized a series of three novel dendrimers containing a single, unsymmetrically located ferrocene center.[211 Compound 8 is the third generation and largest member of the series. The electrochemical behavior of these molecules was investigated in CHzC12 solution. As expected the rate constant for heterogenous electron transfer decreased as the dendrimer generation increased. The half-wave potential for the one-electron oxidation of the ferrocene subunit was found to shift to less positive potentials in going from the first to the third generation compound. Again, this is not an intuitive finding, but it is in excellent agreement with the trend observed by Diederich's group on the oxidation of their Zn-porphyrin dendrimers in nonpolar solvents[101 (CH2C12 and THF). Hydrolysis of the peripheral esters in 8 and lower generation analogs produces a new series of three carboxylate-terminated, water-soluble dendrimers. In pH 7 buffered solution, these dendrimers are almost fully ionized and exhibit large negative charges. Their electrochemical behavior is still under investigation as this is written, but their anionic nature at neutral pH offers an obvious "handle" to restrict their free rotation at the electrode-solution interface and force them to adopt specific orientations that may affect their rates of electron transfer with the electrode surface.
220
16 Electroactive Dendrimers
There is no question that the chemistry of dendrimers offers extraordinary promise. The work summarized in this chapter is concerned only with dendrimers functionalized with electroactive groups. We have not addressed other interesting topics such as the modification of electrode surfaces with dendrimerdzl or their use as templates for the preparation of catalytically active metal colloidal particles.[~l The work described here suggests that dendrimers may find interesting applications as materials for molecule-based storage of charge and/or information. In addition to this, the analogies between some electroactive dendrimers and redox proteins are very interesting and open research avenues geared to mimicking these important biological systems and their functions.
16.3 References 1.See, for instance: (a) G. R. Newkome, C. N. Moorefield and F. Vogtle, Deizdritic Molecules: Concepts, Syntheses, Perspectives, VCH, Weinheim, 1996. (b) F. Zeng and S. C. Zimmerman, Chem. Rev. 1997,97,1681-1712. 2. M. R. Bryce, W. Devonport and A. J. Moore, Angew. Chem. Int. Ed. Engl. 1994, 33,1761-1763. 3. W. Devonport, M. R. Bryce, G. J. Marshallsay, A. J. Moore and L. M. Goldenberg, J. Muter. Chem. 1998,8,1361-1372. 4. B. Alonso, M. Moran, C. M. Casado, F. Lobete, J. Losada and I. Cuadrado, Chem. Muter. 1995, 7,1440-1442. 5. C. Valerio, J.-L. Fillaut, J. Ruiz, J. Guittard, J.-C. Blais and D. Astruc, J. Am. Chem. SOC.1997,119,2588-2589. 6. I. Cuadrado, M. Moran, C. M. Casado, B. Alonso, F. Lobete, B. Garcia, M. Ibisate, J. Losada, Orgunometullics 1996, 15,5278-5280. 7. R. Castro, I. Cuadrado, B. Alonso, C. M. Casado, M. Moran and A. E. Kaifer, J. Am. Chem. SOC.1997,119,5760-5761. 8. I. Cuadrado, C. M. Casado, B. Alonso, M. Moran, J. Losada and V. Belsky, ,i. Am. Chem. SOC.1997,119,7613-7614. 9. B. Gonzalez, C. M. Casado, B. Alonso, I. Cuadrado, M. Moran, Y. Wang and A. E. Kaifer, Chem. Commun. 1998,2569-2570. 10. P. J. Dandliker, F. Diederich, M. Gross, C. B. Knobler, A. Louati and E. M. Sanford, Angew. Chem. lnt. Ed. Engl. 1994,33,1739-1742. 11.P. J. Dandliker, F. Diederich, A. Zingg, J.-P. Gisselbrecht, M. Gross, A. Louati and E. Sanford, Helv. Chim. Actu 1997,80,1773-1801. 12. J. P. Collman, L. Fu, A. Zingg, F. Diederich, Chem. Commun. 1997,193-194. 13. R. Sadamoto, N. Tomioka and T. Aida, J. Am. Chem. SOC.1996,118,3978-3979. 14. D. L. Jiang and T. Aida, Chem. Commun. 1996,1523-1524. 15. C. B. Gorman, B. L. Parkhurst, W. Y. Su and K.-Y. Chen, J. Am. Chem. SOC. 1997,119,1141-1142. 16. G. R. Newkome, R. Guther, C. N. Moorefield, F. Cardullo, L. Echegoyen, E. Perez-Corder0 and H. Luftmann, . Angew. Chem. Int. Ed. Engl. 1995, 34, 20232026.
16.3
References
221
17. G. R. Newkome, F. Cardullo, E. C. Constable, C. N. Moorefield, A. M. W. C. Thompson, J. Chem. SOC., Chem. Commun. 1993,925-926. 18. V. Balzani, S. Campagna, G. Denti, A. Juris, S. Serroni and M. Venturi, Acc. Chem. Res. 1998,31,26-34. 19. T. D. Selby and S. C. Blackstock, I. Am. Chem. SOC. 1998,120,12155-12156. 20. E. M. Bowden, Interface 1997, 6(4), 40-44. 21. C. M. Cardona and A. E. Kaifer, 1. Am. Chem. SOC.1998,120,4023-4024. 22. M. Wells and R. M. Crooks, 1.Am. Chem. SOC.1996,118,3988-3989. 23. M. Zhao and R. M. Crooks, Angew. Chem. Int. Ed. Engl. 1999,38,364-366.
Supramolecular Electrochemistry
Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
17 Molecular Wires
In this chapter we shall briefly survey an area of supramolecular chemistry with great potential for development. It has been more than twenty five years since the theory of a "molecular wire" was first described."] The concept of molecular electronics has progressed to include molecular switches which will be interconnected via molecular wires to form arrays of sites that can perform a variety of functions. Unlike the plethora of successful molecular switches that have been reported in the literature, the preparation of successful molecular wires has proven a challenging undertaking. In this chapter we shall examine the design goals of molecular wires and review a few of the wires that have been the subject of published electrochemical studies. The majority of work in this area has involved photochemical rather than electrochemical methods. It is our hope that this brief chapter will incite further interest in the area. We should note at the outset that as conduction in the wires becomes more efficient, electron transfer rates exceed a level that can be probed by common electrochemical methods. Alternative voltammetric techniques such as AC voltammetry have been employed by some
17.1 The Concept of a Molecular Wire and its Electron Transfer Kinetics Perhaps the simplest description of a molecular wire is that of a conductive path of atomic orbitals permitting long-range electron or hole transfer.f31 Most proposed structures have involved bridged donor-acceptor pairs. Considerable insight into the behavior of these systems has been reaped from extensive research in donor-acceptor systems over the past forty yearsJ41 The first reported molecules purporting to exhibit wire-like behavior were naturally occurring chlorophyll and acceptor assemblies bearing polyisoprene chains.[51 While the possibility of efficient conduction of electrons along the molecular framework has, as mentioned above, been around for twenty five years, bringing this idea to reality has proven a challenge. In most systems prepared thus far, the rate of electron transfer has remained linked to the donoracceptor distance, with increasing separation between the sites adversely effecting the rate of electron transfer. This distance dependence of the observed electron transfer rate has been described by the expression:
k,, = k, expI-PR,,l
(17.1)
17.2
Electrocllenzical Studies of Molecular Wires
223
where ko is a kinetic prefactor, RDA the center to center distance between the donor and acceptor and p a scaling factor of units A-1, characteristic of the nature of the bridge moiety. Typical reported values of J3are 1.0-1.4 A-1 for proteins, 20.2- 1.4 A-1 for DNA, 0.8 -1.0 A-1 for saturated hydrocarbons, and 0.2 - 0.6 A-1 for unsaturated phenylene, polyene, polyyne bridges.[6] Recently however, Wasielewski and coworkers have reported truly impressive p values including an astonishing p of 0.04 A-1 for wire 1.161
1 Clearly from the standpoint of electrochemical conversion, as values of p approach the theoretical minimum, the electron transfer rate increases toward the limit of the heterogeneous electron transfer rate for the redox moiety in question. As mentioned above, when the rate (ko) becomes large it can no longer be measured using the typical electrochemical methods. Creager and coworkers have employed AC voltammetry,[2] while Chidsey has used a coulostatic indirect laser induced temperature jump method for assessing electron transfer in ferrocenyl wires, vide infua."I Assessment of the electron transfer kinetics in slower systems is still quite complicated. Both hole tunneling and electron tunneling contribute to the overall observed transfer rates. The contribution of each component may be resolved in order to give insight into the nature of conduction in the bridge. The results obtained must be examined carefully however, with interpretation of true conduction tempered by consideration of interchain percolation effects along with other factors. Allara, Tour, Weiss and coworkers have advocated use of STM to prove both conductivity and to image the molecule in question to verlfy that it is, in actuality, the conductive component in an assembly.[*] Bearing in mind these caveats, electrochemical study of wires yields important insights and is a direct step toward the development of electrically switched sensors.
17.2 Electrochemical Studies of Molecular Wires In this section we shall address four recent reports on molecular wires. This area of research is enjoying rapid growth with a number of recent successes in increased efficiency of electron transfer. Thus, this section can provide only a
17 Molecular Wires
224
cursory overview of the field. We have selected several studies that have included electrochemical work and that are illustrative of the salient factors for research in the area. Harriman and Ziessel have examined electron localization, exchange and transfer in alkyne bridged metal terpyridyl complexes in some detail.191 While the majority of their studies revolve around photophysical measurements, the authors report half-wave potentials for their trinuclear complexes 2 where M is variously Fe, Zn or Co. Based on a variety of laser excited lifetime measurements the p value in 2-Co(II) system was estimated to be in the range of 0.17 A-1 with the contribution of hole tunneling to p estimated to be more substantial than that of electron tunneling.
2 Tolbert and coworkers reported a prototype for a molecular wire based on bis(ferrocenyl)polymethines.l~~ Their system employs an odd-alternant polyenyl (polymethine ion) bridge, termed a soliton. In this system the soliton, which exhibits its own conduction of negative or positive charge, enhances communication between the terminal acceptor groups, in this instance the ferrocene moieties. Wires of general structure 3' were probed by absorption spectroscopy and voltammetry, with the latter technique used to assess the extent of electronic communication between the terminal ferrocene groups. Amazingly, separate voltammetric waves are observed for the oxidation of the ferrocenyl groups even when the soliton chain length is increased to 13 carbons between the metal centers. The separation between successive oxidations ranges from 0.33 V for n=l to 0.04 V for the n=13 case.131 (This finding is highly unusual in dimeric ferrocenes with bridging of more than two carbons.) The timescale of the electrochemical measurements did not permit resolution of the dynamic events leading to the peak separation, i.e. the effects of the soliton migration or mixed valence states could not be distinguished.
7 7.2 Electroclteniicul Studies ofMoleculur Wires
3
3+
Ferrocenyl wires have also been examined by Chidsey and coworkersJV Mixed SAMs composed of ferrocene terminated polyakynylthiols, general structure 4, and various alkyltl-dols on gold were employed for studies of electron transfer rates by a coulostatic indirect laser induced temperature jump method (ILIT). Surface coverages on the gold electrodes were measured by cyclic voltammetry. The estimated p value for these systems was 0.57 A-l, lower than that for the analogous systems with saturated bridges,[*O]although still high in comparison to those observed for polyene spacers.
0 4
Creager and coworkers have recently reported the electron transfer rates of conjugated ferrocenyl wires based on the general structure 5 where n=O - 2 and X is an alkyl extension bearing variously a pyridyl or silyl group.[lll Mixed monolayers of the ferrocenyl wires and 16-mercaptohexadecanol were prepared on gold bead electrodes and studied by AC voltammetry, a method that along with other impedance methods has emerged as a valuable tool in studying redox kinetics of mixed monolayers with dilute surface coverages of electroactive material.[2J21 By varying the frequency of application of the AC potential insight into the electron transfer rate can be deduced. Fig. 17.1 shows
5
17 Molecular Wires
226
the AC voltammogram for a mixed monolayer of 5 in which n=l. At a low frequency of 100 Hz for the applied AC potential a peak is observed at +0.274 V vs. Ag/AgCl, while no such peak is clearly distinguishable at a frequency of 104 Hz. This indicates that the alternation of potential exceeds the rate of the ferrocene redox reaction. The diminution of the ac peak current relative to the baseline in a series of AC voltammograms acquired at different frequencies can yield a value for the standard electron transfer rate constant ko of an immobilized species.[21 Although the structures employed by Creager and coworkers are similar to those of Chidsey, Creager’s group found a p value of 0.36 A-1 for their systems. While Creager’s monolayers are more dilute (by a 4.Oe-7 D
3.5e-7
{
3.0e-7
6 0
E
2.5t-7
2.08-7
+. E3
u
2
1%-7 1.0s-7
5.00-6 O.Oe+O
J 0.1
0.2
0.3
0.4
0.5
0.6
0.5
0.6
Potential (Volts vs. ref) 4.0e-5
g
-, 10,OOO Hr
3.s-5
E q 3.0e-5
E
1Se-5
P(
1.00-5
2
5.0e-6 O.Oe+O
J
0. I
0.2
0.3
0.4
Potential (Volts vs. ref)
Figure 17.1: AC voltammograms for mixed monolayers of 5 (n=l) and 16mercaptohexadecanol on a gold bead electrode in 1M NaC104.[111 Potentials are reported vs. a Ag/AgCI reference. Peak amplitude of applied voltage was 25 mV 950mV peak to peak), sampling period was Is and sampling interval 10 s. The top voltammogram was acquired at 100 Hz, the bottom at 10,000 Hz. Reprinted with permission of the American Chemical Society.
17.3
References
227
factor of ten in some instances) the reasons for the differing values observed are not clear. They may reflect an effect of differing diluent alkylthiols or perhaps variations of the two methods employed to estimate ko or, a combination of these two factors. One important difference to be noted between the two methods is that ILIT is suited for the measurement of very rapid rates of electron transfer (up to 106 s-1 reportedl71) , while the AC voltammetric method has been recommended by its author for probing slightly slower electron transfer kinetics regimes in the range of 105 s-1.[21
17.3 References 1.A. Aviram, M. A. Ratner, Chem. Physics Lett., 1974,29,277- & X J . 2. S. E. Creager, T. T. Wooster, Anal. Chem., 1998, 70,4257-4263. 3. L. Tolbert, X. Zhao, Y. Ding, L. A. Bottomley, 7. Am. Chem. Soc., 1995, 117, 12891-12892. 4. P. F: Barbara, T. J. Meyer, M. A. Ratner, I. Phys. Chem., 1996,100,13148. 5. A. F. Janzen, J. R.Bolton, (Eds. K. W. Boeer, B. H. Glenn,) Sun 2, Proc. lnt. Sol. Energy SOC. Silver Jubilee Congr., (Pergamon, Elmsford, N. Y.) 1979,1,117-120. 6. W. B. Davis, W. A. Svec, M. A. Retner, M. R. Wasielewski, Nahtre, 1998,396, 80-83. 7. S. B. Sachs, S. P. Dudek, R. P. Hsung, L. R. Sita, J. F. Smalley, M. D. Newton, S. W. Feldberg, C. E. D. Chidsey, J. Am. Chem. Soc., 1997,119,10563-10564. 8. L. A. Bumm, J. J. Arnold, M. T. Cygan, T. D. Dunbar, T. P. Burgin, L. Jones 11, D. L. Allara, J. M. Tour, P. S. Weiss, Science, 1996,271,1705-1707. 9. V. Grosshenny, A. Harriman, R. Ziessel, Angew. Chem. lntl. Ed. Engl., 1995,35, 2705-2708. 10. J. F. Smalley, S. W. Feldberg, C. E. D. Chidsey, M. R. Linford, M. D. Newton, Y.-P. Liu, J. Phys. Chem., 1995,99,13141-13149. 11. S. E. Creager, C. J. Yu, C. Bamdad, S. OConnor, T. MacLean, E. Lam, Y. Chong, G. T. Olsen, J. Luo, M. Gozin, J. Faiz Kayyem, 7. Am. Chem. Soc., 1999, 121, 1059-1064. This paper is strongly recommended as an introduction to methods for determining ko, and electronic coupling in wire systems.
SupramolecularElectrochemistry Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
18 Conclusions and Outlook
We hope that the readers will agree with us that the last two decades have witnessed fast progress in supramolecular chemistry and that electrochemical concepts and techniques have played a significant role in many of the advances in this active research field. In this book we have shown how changes in the oxidation states of appropriate subunits can be used to exert control on supramolecular structures and assemblies. We have also discussed many examples that illustrate how supramolecular structure may affect the kinetics and thermodynamics of electron transfer reactions. These ideas constitute the central core of supramolecular electrochemistry and have been established in a wide variety of systems. The key question that this chapter attempts to address is: Where can we go from here? This type of question is indeed extremely difficult to answer. Innovation has always been one of the driving engines of science and predicting the scientific directions and accomplishments of the future is obviously risky. There are some trends, however, that are clearly visible today. We can extrapolate them and make reasonable guesses on what the near future will bring us in this field of science. During the last two decades electrochemists have learned to manufacture and use electrodes of micrometer and even nanometer dimensions. Electrode miniaturization has brought about the miniaturization of the systems that can be addressed with electrochemical techniques. In chapter 6 we described some of the recent results on single molecule electrochemistry. As this is written, a report by Crooks and coworkers[11demonstrates the feasibility of using carbon nanotubes as ultramicroelectrodes. While these developments have changed the field of electrochemistry, supramolecular chemists have been at work preparing, characterizing and functionalizing larger and larger molecules and supramolecular assemblies. For instance, we can now easily prepare dendrimers having dimensions in the nanometer range. We must realize, thus, that electrodes and molecules (or assemblies) are reaching similar sizes simultaneously, opening new and fascinating possibilities for singlemolecule manipulation and study. Although considerable technical difficulties need to be overcome, these ideas should no doubt develop into a highly exciting area of research during the next few years. Electrochemically or, more generally, redox switchable molecules are one of the most promising types of multi-stable molecular systems. Electrochemicalreactions can be effectively used to reversibly control the state of these interesting molecules. However, in order to move this research work towards more practical applications it is necessary to find ways to address a single molecule or a small group of molecules at a time. Switching large
18
Conclusions and Outlook
229
numbers of molecules in the solution phase is an elegant and worthwhile research accomplishment, but unrealistic and unpractical if we want to build information storage or processing systems, which truly resemble our current electronic circuits. The key reason to replace silicon-based circuit elements by molecdes is miniaturization. Therefore, switching or addressing trillions of molecules at the same time serves no useful purpose. To accomplish the level of required addressability researchers must move away from the solution phase to work either in the solid state or at surfaces, media in which molecules can be anchored at specific locations and addressed individually or as small groups. In this regard, the equalization of sizes between electrode surfaces and molecular systems (vide supra) opens the possibility of using each molecule as an individual circuit component addressable through its corresponding electrode. Scanning probe microscopic techniques may also play an important role in this area. Molecular recognition in interfacial environments has extraordinary relevance in biochemistry as many important recognition and binding phenomena take place on the surface of cells and membranes. It is thus hardly surprising that investigations on interfacial molecular recognition have become more commonplace in the last decade. In this regard, electrochemical techniques occupy a privileged position because the electrode-solution interface offers an excellent framework for these studies. In the next few years, we should see increasing research activity on host-guest phenomena at electrochemical interfaces. We should note here that the field of supramolecular chemistry keeps expanding and that researchers in this field are continuously applying their basic concepts and ideas to new systems. As an illustrative and pertinent example, we can mention the metal-solution interface of colloidal particles as a novel platform for molecular recognition studies that has attracted some recent attention. Work by several groups[241is breaking ground in this novel area and expanding the horizons of supramolecular electrochemistry. In the past, electrochemical experiments were customarily carried out with a millimolar solution of the electroactive species in an appropriate supporting electrolyte/solvent medium. Nowadays, the sensitivity of electrochemical techniques is greatly improved. Pulse voltammetric techniques are gaining wider popularity, among other reasons, because they afford excellent sensitivity. Square wave voltammetry, for instance, can be used in solutions with concentrations of electroactive species as low as 10 pM. In addition to this, ultramicroelectrodes can be employed in fairly resistive solutions, eliminating the requirement for large concentrations of supporting electrolyte that may be incompatible with certain supramolecular systems. Therefore, the possibilities for applications of electrochemical techniques from an experimental standpoint are only increasing. A key problem in supramolecular chemistry is the characterization of high molecular weight assemblies. As the complexity and molecular weight of the supramolecular assemblies prepared increase so do the difficulties associated with their characterization. Chemists have at their disposal a rather limited number of techniques useful for these purposes and the use of electrochemical techniques may prove increasingly fruitful as long as electroactive subunits are
230
18
Conclusions and Outlook
suitably inserted as markers in the supramolecular structures to be characterized. Diffusion coefficients, electrochemical potentials, and electron transfer rate constants may provide different types of information that can be of assistance in the characterization of high molecular weight supramolecular structures. Selectivity is a perennial problem in modern chemical research. We measure enviously the high binding selectivities exhibited by biological hosts, such as antibodies or enzymes. As chemists, we strive to reach similar selectivity values when we design a molecular host for a particular guest. Unfortunately, the selectivity presented by synthetic receptors, with very few exceptions, is orders of magnitude below that exhibited by natural systems. Of course, biological systems are the result of extremely lengthy processes of molecular evolution, and may contain large fractions of material without any useful purpose, accumulated over millions of years. Chemists must design within a much shorter time scale and with material economy in mind. We must admit, however, that we cannot compete with Nature, at least, not yet. From the standpoint of the development of sensors, catalytic systems, and other practical systems we must be concerned with selectivity since it often determines the difference between a practical development and a system with mere academic interest. Supramolecular chemists must be at the forefront in searching for solutions to this problem. In particular, electrochemists can perhaps make use of the peculiar character of interfacial environments to enhance selectivity properties. Alternatively, sensor arrays, aided by pattern recognition software, may offer practical ways to circumvent problems associated with lack of selectivity. Combinatorial chemistry has changed the approach to chemical problem solving in many branches of chemistry. Not surprisingly, combinatorial approaches are also starting to emerge and gain importance in supramolecular chemistry. Electrochemical techniques and concepts may play a significant role in the future development of this field providing methods to analyze and sort the most effective compounds within a library. Also, molecular recognition and selectivity at electrochemical interfaces may also benefit from combinatorial approaches. In general, the future of supramolecular electrochemistry appears bright and exciting. We have only started to develop the possibilities of this field. Research in supramolecular electrochemistry ideally requires collaborative efforts from several groups because of its inherent multidisciplinary character. As more researchers travel across and along the rather arbitrary frontiers between the several branches of chemistry and other sciences, the frontiers themselves will tend to disappear. Multidisciplkary research areas, such as supramolecular electrochemistry, will thus enjoy their best days.
18.1
References
231
18.1 References 1.J. K. Campbell, L. Sun, R. M. Crooks, J. Am. Chem. SOC.1999,121,3779-3780. 2. J. J. Storhoff, R. Elghanian, R. C. Mucic, C. A. Mirkb, R. L. Letsinger, J. Am. Chem. SOC.1998,120,1959-1964. 3. J. Liu, S. Mendoza, E. Roman, M. J. Lynn, R. Xu, A. E. Kaifer, J. Am. Chem. SOC. 1999,121,4304-4305. 4. A. K. Boal, V. M. Rotello, J. Am. Chem. SOC.1999,121,49154915.
Supramolecular Electrochemistry Angel Kaifer, Marielle G6mez-Kaifer 0 WILEY-VCH Vcrlag GmbH. 1999
Index
AN conditions, bulk electrolysis 29 acceptors - catenanes 150 - supramolecular systems 91 acetonitrile 65 f acidic solutions, cleaning 59 Ag/AgCl electrodes 59 alkali chlorides, supporting electrolytes 67 alkenethiols, SAMs 191 alkylammonium, supramolecular systems 92 alkyne, molecular wires 224 alumina, polishing compounds 57 amalgam electrodes 55 amphiphiles, LB films 180 amphiphilic aggregation 93 anion binding 114 ff, 122 anodes 2 f anthracene hosts 137 anthraquinone moiety 117 approxima tion methods, numerical 77 f aproticDMF 91 arene bindings 91 argon atmosphere 65 aromatic amino acids 132 f aromatic systems 91 array ultramicroelectrodes 46 associa tion - fixed 93 f - flavinophane 131 azobenzene derivatives, LB films 186 backward processes, electron transfer 8 band ultramicroelectrodes 46
bead electrodes 55,58 benzene - solvents 66 - supramolecular systems 91 benzidine rotaxanes 157 benzimidazole 95 benzonitrile - solvents 66 - supramolecular systems 91 binary systems, solvents 91 binding - calixarenes 119 - switching 103 f, 114 ff bipyridine ligands 95 bipyridinum 159 bipyridinum derivatives 19 bis(chlorophyl1) cyclophane system 127 f bis(cyclopentadienyl)iron(II) see ferrocene boiling points, solvents 64 bolaamphiphiles 112 Boltzmann constant 6 bromides, supporting electrolytes 68 BTS catalysts, drying agents 67 bulk electrolysis 29 - celldesign 70 bulk solutions, digital simulations 77 bulk techniques 13 Butler-Volmer equation 9
calcium hydridekhloride 66 calixarene based structures 100 calixarenes - cyclophanes 135 - switching 114, 119
234
calomel electrodes 59 ff capacitance, ultramicroelectrodes 45 carbon paste electrodes 55 carbonyl bonds 56 carboxylcoba1tocenium reduction 86 catechol solutions, cyclophanes 133 catenand effect 146 ca tenanes - cyclophanes 132 - intertwined structures 142 - n-donor/acceptor interactions 150 ff - redox-switchable 127 - self assembly 93 cathodes 2 ff cation binding, switching 114 ff cation host, redox switchable 105 cation-K interactions, supramolecular systems 89 f cation recognition, switching 103 cationic dendrimers 208 cavity binding, cyclophanes 132 cell design 68 f cell potentials 1 ff cell types 14 CH2C13 supramolecular systems 90 charging currents 12,45 chlorides supporting electrolytes 67 chlorophyll, molecular wires 222 chronoamperometry 13,% f chronocoulometry 28 cleaning - counter electrodes 59 - vacuum methods 73 cobaltocene - dendrimers 212 - redox couples 85 cobaltocenium - cyclophanes 135 - redox couples 18,85 concentration profiles, redox groups 80 constrictive binding 84 contamination, counter electrodes 70 controlling, molecular devices 109
Index
convection 5 coordination arrays 164,175 f copper helicates 164 ff Cotrell conditions - ultramicroelectrodes 49 - potential step methods 22 f coulombic interactions - dendrimers 210 - supramolecular systems 89 coulometric titrations 13 counter electrodes 14,59,70 coupling, redox sites 94 f covalently interlocking, self assembly 93 crown ethers, switching 105,114,153 cryptands, switching 105,114 ff crystal formation, reference electrodes 61 current genera tion modes, SECM 50 current potential curve, electrode reactions 7 currents - faradaic 11 - ultramicroelectrodes 45 cyclic voltammetry (CV) 16 - dendrimers 211 - digital simulations 79 - ECmechanism 85 - helicates 166 - potential sweep methods 34 f - ultramicroelectrodes 48 cyclodextrin - inclusion complexes 124 - redox couples 85 - self assembly 94 cyclodextrin based rotaxanes, intertwined structures 143 cyclophanes - intertwined structures 157 - redox-active 130 - redox-switchable 127 ff - supramolecular systems 91 - viologens 150
235
Index
decomposition, supporting electrolytes 55 dendrimers, electroactive 207 ff dendritic effect, molecular recognition 209 diamine ketone, switching 118 diamond paste, polishing compounds 57 diazacrown ethers, switching 117 dichloromethane 65 f dielectric constants 89 f diferrocinyl bolaampiphiles 112 differential pulse voltammetry (DPV) 16,39 diffusion - digital simulations 78 - mass transport 5 - potential step methods 23 f - chronocoulometry 29 - redox groups 81 diffusion layer thickness 46 DigiSim 79,82 digital simulations 77 ff dihydrogenphosphate binding 124 dimerization, helicates 165 dimethylformamide (DMF) 65 f dimethylsulfoxide (DMSO) - solvents 65 f - supramolecular systems 90 dioxynaphthalene moiety, cyclophanes 134 dipole-dipole interactions 89 diquinone structures 120 diquinonecalixarenes 100 discretization, electrolpc solutions 77 disk electrodes 55 f disk ultramicroelectrodes 46 dispersion 91 dissolution 1 disulfide attachement, self assembly 93 DNA, molecular wires 223 donor acceptor systems - intertwined structures 143
molecular wires 222 donors - catenanes 150 - supramolecular systems 91 dopamine solutions, cyclophanes 133 double bridges, reference electrodes 62 double layer, electrical 11 drierite, drying agents 67 drying, solvents 64 ff dynamic methods 13
-
EC/ECE/EE mechanisms 84 electroactive dendrimers 207 ff electroactive guests, cyclodextrin 87 electroactive intertwined structures 142 ff electroactive Langmuir-Blodgett films 180 ff electroactive SAMs 195 electrochemical conditions, intermolecular forces 89 f electrochemical reactions 1 ff electrochemical simulation package (ESP) 79 electrochemical switching 103 ff electrochemical techniques 11 ff electrochemistry, single molecules 51 f electrocrystallization 31,72 electrode blocking, SAMs 193 electrode size, potential step methods 24 electrode surfaces 45 ff, 55 ff - precipitation 91 electrode types 14,32,45,55 f electrodes 2 ff electrogravimetry 13 electroinactive cyclophane hosts 135 electrolytic solutions, discretization 77 electron movements 4 electron transfer lff, 8 - homogeneous processes 84 f
Index
236
LBfilms 183 molecular wires 222 SAMs 193 epinephrine solutions 133 equilibrum methods 13 error function, Cottrell conditions 25 ester hydrolysis, dendrimers 219 ethanol 66 ethers 114 excitation functions 15 - chronoamperometry 26 - cv 34 - DPV 40 - LSV 33 - NPV 37 - potential step methods 22 - S W V 42 experimental methods 55 ff explicite finite difference (EFD) 78 explosion hazard, NaK amalgam 66
-
faradaic currents 11 - digital simulations 78 - ultramicroelectrodes 45 Faraday law 3 ff, 30 feedback, ultramicroelectrodes 49 Fermilevel 3 ferrocene - cyclophanes 135 - dendrimers 209 - hemicarcerand encapsulation 84 - pseudoreference 62 - reversible redox couples 18 - SAMs 197 - self assembly 93 - switching 105 - ultramicroelectrodes 53 - vacuum methods 75 ferrocenyl - LBfilms 183 - molecular wires 225 - switching 114,118 ferrocyanine, switching 122 Fick law, digital simulations 6,78 flame polishing 59
flavinophane isoalloxazine 131 forward reactions, electron transfer 8 forward scan, CV 34 fullerenes - cyclophanes 137 - supramolecular systems 91 future trends 228 ff galvanic cells 2 ff galvanostatic methods 13 glass shrouds 56 glassblown cells, vacuum methods 74 glassy carbon surfaces 55 gold disk electrodes 55 f grid-like structures, self assembly 93 grids 164,175f guests - intertwined structures 143 - self assembly 93 half wave potentials cv 37 - helicates 167 - switching 103 ff halide adsorption 57 helicates 164 ff - redox-active moieties 101 - self assembly 93,110 hemicarcerand - ferrocene 84 - self assembly 94 heterogeneous electron transfers 11 hexaamineruthenium(I1) 20 hexacyanoferra to(I1) 20 hexafluorophosphate 68 hole transfer, molecular wires 222 homogeneous processes, electron transfer 84 f horizontal touching technique, LB films 182 hosts - electroinactive cyclophanes 135 - ferrocene 84 - intertwined structures 143, 150 -
Index
redox switchable 104 self assembly 93 - sterically hindered 122 - supramolecular systems 89 HPLC grade, solvents 65 hydrogen bonding - dendrimers 209 - electrode surfaces 56 - molecuar receptors 137 - supramolecular systems 89 f hydrophobic interactions 89 f hydroxymate, iron binding 111
-
-
implicite finite difference (IFD) 79 inclusion complexes 87,124 indicator electrodes 14 indole solutions, cyclophanes 133 interdigitated array ultramicroelectrodes 46 interfacial techniques 11 f intermolecular interactions 89 internal electroactive groups, dendrimers 213 internal reference 63 intertwined structures, electroactive 142 ff ion binding, switching 114 ion-dipole interactions 89 ion-ion interactions 89 ion movements 4 ion-quadrupole effects 92 iron binding ligand, helical 111 iron complexes, metallocyclophanes 130 isomeric covalently linked systems 99 isomers - catenanes 153 - cyclophanes 129 iterations, digital simulations 78 Kel-F 56 kine tics - digital simulations 77 - electrode reactions 6 ff
231
Langmuir-Blodgett (LB) films 180 ff lariat ethers 114 leakage 56 ligands - dendrimers 216 - helicates 164 ff - redox-active moieties 95 - switching 114 limiting currents, ultramicroelectrodes 47 linear sweep voltammetry (LSV) 16,32 linking, self assembly 93 lithium fluorides 67 macrocycles ferrocenyl 108 intertwined structures 146 - switching 114 macroelectrodes 45 mass transport 4 ff - potential step methods 23 - ultramicroelectrodes 45 mechanical linking 93 mercapto pyridine 203 mercury, toxicity 15 mercury electrodes 55,61 metal complexes, reversible redox couples 20 metallic circuits 1 metalloca tenanes, intertwined structures 143,145 f metallocyclophanes 127 ff metallorotaxanes, templated 145 f methanol 66 methyl carboxylates 213 micelles, switching 112 microelectrodes 45 moieties, redox-active 94 molecular devices 109,160 molecular receptors, cyclophanes 130,137 molecular recognition - dendrimers 209 - SAMs 198
-
238
Index
molecular wires 222 ff monodispersivity, dendrimers 207 monolayers - LBfilms 181 - self-assembled 191 - supramolecular systems 89 monomers, cyclophanes 129 monoquinone, SQV 119 multilayers, LB films 184 Nafion films, cyclophanes 133 naphthalene groups, cyclophanes 130 f naphthalene quinone 205 naphthalimides 137 negative feedback, ultramicroelectrodes 49 Nernst equation 4 ff - digital simulations 78 - potential controlled systems 17 - ultramicroelectrodes 47 neurotransmitters, catechol-type 133 nitrates 67 nitroaromatic groups, switching 114,119 nitrobenzene, reversible redox couples 20 nitrobenzene lariat ethers 115 nitrogen atmosphere 65 nitromerocyanine 203 nitromethane 66 non-faradaic currents 11 nonpolar solvents 91 nonsolid electrodes 55 norepinephrine solutions 133 normal hydrogen electrode (NHE) 3,62 normal pulse voltammetry (NPV) 16,37 numerical approxima tion methods 77 f ohmic drops 45 oligopyridine ligands
164 f
optically transparent electrodes 55,58 organic solvents - LBfilms 180 - supramolecular systems 90 f overpotential 8 oxydation reactions 1 ff, 6 f oxygen scavenger 67
PEEK 56 pentamethylphenyl moiety, molecular receptors 138 perchlorates 67 perfluorinated vacuum 74 peripheral electroactive groups, dendrimers 207 f pertubations, digital simulations 78 phenanthroline derivatives 112 phenylenediamine derivatives, dendrimers 218 phophates 67 phospholipid monolayers 186 photoswitchable SAMs 203 physical constants, solvents 66 n-acceptor hosts, cyclophanes 132 f n-donor/acceptor interactions - catenanes 150 ff - rotaxanes 155 f n-donors, intertwined structures 143 n-n interactions 89 f pinholes, SAMs 195 planar diffusion, potential step methods 24 plastic shrouds 56 platinum disk electrodes 57 platinum electrodes 209 platinum surfaces 55 poisoning, electrode surfaces 56 polar dendrimers 208 polar solvents 65 polishing, electrode surfaces 57 f polyaminemacrocycles 122 polyisoprene, molecular wires 222 poly(siloxyphtha1ocyanines) 188 poly(viny1ferrocene) 96
I?dex
porph yrin cyclophanes 128 - dendrimers 213 - intertwined structures 143 positive feedback, ultramicroelectrodes 49 potential controlled systems, Nernst equation 17 potential excitation functions see excitation functions potential step methods 22 ff potential sweep methods 32 ff potential window, cathodic 57 potentiostatic methods 13 precipitation, electrode surfaces 91 propylene carbonate 66 pseudoreference, silver wires 62 pseudorotaxanes 142 pulsed voltammetric techniques 12 f, 37 ff purification, solvents 64 purity, supporting electrolytes 68 purple red, helicates 111 pyridyl groups, molecular wires 225 pyridyne ligands, helicates 164 f
-
quinone based ligands, switching 114 ff quinones - electrode surfaces 56 - redox-active moieties 100 - reversible redox couples 20 - switching 119 racks 164,175 f Randles Sevcik equation 32 f receptors - molecular 127 ff - switching 114 redox active guests, anion binding 122 redox active moieties 94 - intertwined structures 142
239
redox active supramolecular systems 103 f redox groups - CV behavior 79f - surface confined 196 redox couples - pseudoreference 62 - reversible 18,35 - ultramicroelectrodes 47 redox switchable cation hosts 105 reducing agents, solvents 66 reduction 1 reference electrodes 14,32,59 resistances 45, 56 reverse scan, CV 34 Ridox, drying agents 67 ring ultramicroelectrodes 46 rotating electrode voltammetry 69 rotaxanes - cyclodextrin-based 143 - intertwined structures 142 - x-donor/acceptor chemistry 155 f - redox-switchable 127 - self assembly 93 ruthenium complexes 130 salt bridges 2,60 saturated calomel electrode (SCE) 59 f scanning electrochemically microscopy (SECM) 45,49 self assembled monolayers (SAM) 191 ff self assembly - helicates 164,175 - supramolecular systems 89, 93 semiconductor circiuts 1 semiconductor electrodes 58 serotonin solutions 133 shapes, ultramicroelectrodes 46 shroud types 56 shuttles - intertwined structures 142 - n-donor/acceptor chemistry 155 f - redox-switchable 127 - self assembly 93
240
silane attachement, self assembly 93 silyl groups, molecular wires 225 single molecule electrochemistry (SME) 51 f silver wires, pseudoreference 62 smooth bead electrodes 57 sodium chloride saturated calomel electrode (SSCE) 59 f sodium fluorides 67 sodium metal, explosion hazard 66 sodium potassium amalgam 65 solid electrodes 55 solitons, molecular wires 224 solubility, solvents 64 solution flow, mass transport 5 solution layer thickness 26 solvents 64 ff - supramolecular systems 91 solvophobic interactions 89 f spacers, switching 117 sphere ultramicroelectrodes 46 spontaneous formation, SAMs 191 square wave voltammetry (SWV) 16,42 - digital simulations 79 - redox groups 83 stacks, self assembly 93 staircase excitation functions 15 stirring, mass transport 5 Stokes-Einstein equation 5 f structural changes, helicates 165 f sulfates 67 supporting electrodes 5 supporting electrolytes 64 ff, 89 f - decomposition 55 - pseudoreference 63 supramolecular systems 13 ff, 89 ff - reversible redox couples 18 surface confined redox centers, SAMs 196 surfaces - electrodes 55 f - ultramicroelectrodes 45 ff swinging catenanes 147
Index
switching, electrochemical 103 ff switching potential, CV 34 Tafel equation 8 technique classifications 13 f teflon components, vacuum methods 73 teflon tubing, cell design 68 templa ted metallocatenanes/ rotaxanes 145 f terpyridyl, alkyne bridged 224 Tesla probe, vacuum methods 74 tetra butylammonium hexafluorophosphate (TBAPF6), supporting electrolytes 68 tetrachloroethane (TCE) 65 f tetrahydrofuran 65 f tetramethylammonium 92 tetrathiafulvalene (TTF) - dendrimers 208 - reversible redox couples 19 - supramolecular systems 90 thermodynamics, digital simulations 77 thin layer cells 69 thiobisethyl acetate (TBEA) 199 thioVdisulfide attachment, self assembly 93 thiolate gold SAMs 191 three electrode cells 14 time constant, electrochemical cell 12 time evolution, digital simulations 78 timescales, vacuum methods 75 tin oxide films, semiconductor electrodes 58 toluene - solvents 66 - supramolecular systems 91 toxicity - mercury 15 - solvents 65 trapping, electroactive molecules 52 triflates 68
24 1
Idex
triggers, intertwined structures 143 tunneling, molecular wires 223 two electrode cells 14 ultramicroelectrodes
45 ff
vacuum conditions 56 vacuum methods 72 ff van der Waals interactions, supramolecular systems 89 f Venus flytrap ligands 138 vertical dipping, LB films 181 vesicles, switching 112 vinylpyridinium derivatives, LB films 187 viologens - cyclophanes 132 f, 150
pseudorotaxanes 144 reversible redox couples voltammetry 13 ff, 32 f voltammograms see cyclic voltammograms 15 volume elements, digital simulations 77 vycor bridges 60
-
-
19
water 66,90 wave form potential parameters, swv 43 wire ultramicroelectrodes 46 wires, molecular 222 ff working electrodes 14,32,55 f xylylene bridges, cyclophanes
132