134 109 9MB
English Pages 312 [313] Year 2006
121 Structure and Bonding Series Editor: D. M. P. Mingos
Editorial Board: P. Day · T. J. Meyer · H. W. Roesky · J.-P. Sauvage
Structure and Bonding Series Editor: D. M. P. Mingos Recently Published and Forthcoming Volumes
Single-Molecule Magnets and Related Phenomena Volume Editor: Winpenny, R. Vol. 122, 2006 Non-Covalent Multi-Porphyrin Assemblies Synthesis and Properties Volume Editor: Alessio, E. Vol. 121, 2006 Recent Developments in Mercury Sience Volume Editor: Atwood, David A. Vol. 120, 2006 Layered Double Hydroxides Volume Editors: Duan, X., Evans, D. G. Vol. 119, 2005 Semiconductor Nanocrystals and Silicate Nanoparticles Volume Editors: Peng, X., Mingos, D. M. P. Vol. 118, 2005 Magnetic Functions Beyond the Spin-Hamiltonian Volume Editor: Mingos, D. M. P. Vol. 117, 2005 Intermolecular Forces and Clusters II Volume Editor: Wales, D. J. Vol. 116, 2005 Intermolecular Forces and Clusters I Volume Editor: Wales, D. J. Vol. 115, 2005 Superconductivity in Complex Systems Volume Editor: Müller, K. A. Vol. 114, 2005
Principles and Applications of Density Functional Theory in Inorganic Chemistry II Volume Editors: Kaltsoyannis, N., McGrady, J. E. Vol. 113, 2004 Principles and Applications of Density Functional Theory in Inorganic Chemistry I Volume Editors: Kaltsoyannis, N., McGrady, J. E. Vol. 112, 2004 Supramolecular Assembly via Hydrogen Bonds II Volume Editor: Mingos, D. M. P. Vol. 111, 2004 Applications of Evolutionary Computation in Chemistry Volume Editors: Johnston, R. L. Vol. 110, 2004 Fullerene-Based Materials Structures and Properties Volume Editor: Prassides, K. Vol. 109, 2004 Supramolecular Assembly via Hydrogen Bonds I Volume Editor: Mingos, D. M. P. Vol. 108, 2004 Optical Spectra and Chemical Bonding in Transition Metal Complexes Special Volume II dedicated to Professor Jørgensen Volume Editor: Schönherr, T. Vol. 107, 2004
Non-Covalent Multi-Porphyrin Assemblies Synthesis and Properties Volume Editor: Enzo Alessio
With contributions by E. Alessio · I. Bouamaied · T. Coskun · G. Ercolani L. Flamigni · M. J. Gunter · V. Heitz · J. T. Hupp · E. Iengo Y. Kobuke · J.-P. Sauvage · F. Scandola · E. Stulz
13
The series Structure and Bonding publishes critical reviews on topics of research concerned with chemical structure and bonding. The scope of the series spans the entire Periodic Table. It focuses attention on new and developing areas of modern structural and theoretical chemistry such as nanostructures, molecular electronics, designed molecular solids, surfaces, metal clusters and supramolecular structures. Physical and spectroscopic techniques used to determine, examine and model structures fall within the purview of Structure and Bonding to the extent that the focus is on the scientific results obtained and not on specialist information concerning the techniques themselves. Issues associated with the development of bonding models and generalizations that illuminate the reactivity pathways and rates of chemical processes are also relevant. As a rule, contributions are specially commissioned. The editors and publishers will, however, always be pleased to receive suggestions and supplementary information. Papers are accepted for Structure and Bonding in English. In references Structure and Bonding is abbeviated Struct Bond and is cited as a journal. Springer WWW home page: springer.com Visit the Struct Bond content at springerlink.com
Library of Congress Control Number: 2006922800
ISSN 0081-5993 ISBN-10 3-540-32542-5 Springer Berlin Heidelberg New York ISBN-13 978-3-540-32542-0 Springer Berlin Heidelberg New York DOI 10.1007/11677529
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com c Springer-Verlag Berlin Heidelberg 2006 Printed in Germany The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: Design & Production GmbH, Heidelberg Typesetting and Production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig Printed on acid-free paper 02/3100 YL – 5 4 3 2 1 0
Series Editor Prof. D. Michael P. Mingos Principal St. Edmund Hall Oxford OX1 4AR, UK [email protected]
Volume Editor Prof. Enzo Alessio Dipartimento di Scienze Chimiche Università di Trieste Via L. Giorgieri 1 34127 Trieste, Italy [email protected]
Editorial Board Prof. Peter Day
Prof. Herbert W. Roesky
Director and Fullerian Professor of Chemistry The Royal Institution of Great Britain 21 Albermarle Street London W1X 4BS, UK [email protected]
Institut for Anorganic Chemistry University of Göttingen Tammannstr. 4 37077 Göttingen, Germany [email protected]
Prof. Jean-Pierre Sauvage Prof. Thomas J. Meyer Department of Chemistry Campus Box 3290 Venable and Kenan Laboratories The University of North Carolina and Chapel Hill Chapel Hill, NC 27599-3290, USA [email protected]
Faculté de Chimie Laboratoires de Chimie Organo-Minérale Université Louis Pasteur 4, rue Blaise Pascal 67070 Strasbourg Cedex, France [email protected]
Structure and Bonding Also Available Electronically
For all customers who have a standing order to Structure and Bonding, we offer the electronic version via SpringerLink free of charge. Please contact your librarian who can receive a password or free access to the full articles by registering at: springerlink.com If you do not have a subscription, you can still view the tables of contents of the volumes and the abstract of each article by going to the SpringerLink Homepage, clicking on “Browse by Online Libraries”, then “Chemical Sciences”, and finally choose Structure and Bonding. You will find information about the – – – –
Editorial Board Aims and Scope Instructions for Authors Sample Contribution
at springer.com using the search function.
Preface
Among many interesting light-induced functions, those inspired by natural photosynthesis have attracted special attention. In fact, as a result of extensive spectroscopic studies and detailed structural determinations, the photoinduced charge separation taking place in the reaction centers and the antenna effect carried out by the light-harvesting units are now understood in great detail. Artificial systems mimicking the natural photoinduced processes are the subject of continuing research activity, fostered by the problem of artificial solar energy conversion. In many of these artificial systems, porphyrin (and/or metalloporphyrin) units play a major role as active chromophores. In fact, aside from their chemical resemblance to the chlorophylls of the natural systems, porphyrins can be considered as ideal components for the construction of artificial systems because of several appealing features: rigid, planar geometries; high stability; inherent symmetry; facile tunability of their optical and redox properties by metalation/functionalization. Supramolecular synthetic strategies have emerged as viable alternatives to covalent synthesis in the construction of large multicomponent porphyrin architectures. Both coordination and multiple hydrogen bonds can be used as the source of the noncovalent bonding interactions, often in combination with other noncovalent interactions and/or with more conventional covalent procedures, for the efficient construction of robust and shape-persistent assemblies of nanoscopic dimensions in which the number, position, relative orientation and distance of the chromophores are well defined. Multifunctionalized porphyrins with a high degree of symmetry are easily accessible synthetically, thus topologically controlled multiporphyrin arrays can be assembled by judicious design of the substitution pattern and geometry of the building blocks. This book summarizes recent work on metal-mediated and hydrogenbonded multiporphyrin assemblies which are stable enough to be characterized in solution. Assemblies coordinated only in solids (e.g. crystals) are not included. Emphasis is given on the photophysical and molecular recognition properties of the supramolecular multiporphyrin systems. In the first chapter, by Imenne Bouamaied, Timur Coskun and Eugen Stulz, the use of axial coordination to metalloporphyrins for constructing multinuclear complexes is discussed. The porphyrin-as-ligand concept combined with orthogonal binding modes is presented, and selected examples show that in
X
Preface
this way a large diversity of multiporphyrin assemblies can be achieved. New emerging concepts, such as dynamic combinatorial chemistry, porphyrinfullerene complexes and porphyrins assembled around gold nanoclusters or on surfaces are also presented. These systems are expected to play a leading role in the design of new materials in the near future. A particularly intriguing kind of supramolecular self-assembly by axial coordination to metalloporphyrins, namely self-complementary coordination, is reviewed in the second chapter by Yoshiaki Kobuke. Self-complementarity affords large stability constants. The fascinating systems described in Kobuke’s contribution were prepared and investigated as models for components of photosynthetic natural systems: self-complementary dimers of porphyrins mimic the special pair of the photosynthetic reaction centers, while macrocyclic and three-dimensional porphyrin supramolecules were prepared as photosynthetic antenna models. The third chapter, by Elisabetta Iengo, Franco Scandola and Enzo Alessio, describes yet another approach to the metal-driven construction of multiporphyrin assemblies: the formation of coordination bonds between exocyclic donor site(s) on the porphyrins and metal centers belonging to coordination compounds. Several discrete multiporphyrin assemblies, prepared by the authors and by others, are reviewed. The structural aspects and the photophysical properties of some ruthenium-mediated assemblies of porphyrins are treated in detail. The following chapter, by Joseph T. Hupp, focuses on how rhenium ions, and especially tricarbonylrhenium(I)chloro fragments, have been used to organize and link pyridine-functionalized porphyrins into neutral molecular squares, rectangles, planar dimers, and more complex structures. Emphasis is given on the properties of the multiporphyrin assemblies. Films and membranes generated from porphyrin metallacycles and displaying molecular-scale porosity for molecular sieving are described. Additional applications of films, membranes, and free assemblies concern chemical sensing, oxidative catalysis, and light-to-electrical energy conversion. An overview of the thermodynamic principles governing self-assembly in solution, with particular reference to multiporphyrin architectures, is presented by Gianfranco Ercolani in the fifth chapter. The topic is discussed in order of increasing complexity, from simple acyclic assemblies to multicyclic assemblies. The principles are illustrated by selected examples of metalmediated assemblies of porphyrins, many of which have been described in the previous chapters. An intriguing series of electro- and photoactive porphyrin catenanes and rotaxanes, obtained by the authors through copper(I)-templated synthesis, is reviewed in the sixth chapter by Lucia Flamigni, Valérie Heitz, and Jean-Pierre Sauvage. The photo-induced processes – energy and electron transfer reactions – occurring in the interlocked structures upon light absorption are discussed in detail and critically compared to closely related systems reported by others.
Preface
XI
The last chapter, by Maxwell Gunter, gives a comprehensive overview of multiporphyrin arrays assembled through hydrogen bonding and shows how the geometric precision and strong directionality of H-bonds between relatively rigid donor and acceptor groups can be incorporated into the architecture of porphyrin supramolecular arrays to produce some remarkably complex assemblies through simple mixing of the component parts. This chapter demonstrates that composite H-bonds can be used to advantage for the assembly of relatively robust and well-defined multiporphyrin arrays whose integrity can be ensured in solution, provided that a nonhydrogen-bonding solvent is utilized. Maxwell’s review is constrained to those systems which principally result in multiporphyrin arrays, although often templated by nonporphyrinic component parts. I am grateful to the authors for their contributions on these important topics and hope that this volume will be of use to scientists working in this area and a source of references and inspiration. Finally, I wish to acknowledge the precious contribution of my coworker, Dr. Elisabetta Iengo, in the shaping of this book. Trieste, February 2006
Enzo Alessio
Contents
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies I. Bouamaied · T. Coskun · E. Stulz ....................................................... Porphyrin Supramolecules by Self-Complementary Coordination Y. Kobuke ......................................................................................................
1
49
Metal-Mediated Multi-Porphyrin Discrete Assemblies and Their Photoinduced Properties E. Iengo · F. Scandola · E. Alessio . . . . . . . . . . . . . . . . . . . . . 105 Rhenium-Linked Multiporphyrin Assemblies: Synthesis and Properties J. T. Hupp 145 Thermodynamics of Metal-Mediated Assemblies of Porphyrins G. Ercolani
167
Porphyrin Rotaxanes and Catenanes: Copper(I)-Templated Synthesis and Photoinduced Processes L. Flamigni · V. Heitz · J.-P. Sauvage . . . . . . . . . . . . . . . . . . . . 217 Multiporphyrin Arrays Assembled Through Hydrogen Bonding M. J. Gunter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Author Index Volumes 101–121 . . . . . . . . . . . . . . . . . . . . . . 297 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Contents of Volume 111 Supramolecular Assembly via Hydrogen Bonds II Volume Editor: D. Michael P. Mingos ISBN: 3-540-20086-X Hydrogen Bonding Interactions Between Ions: A Powerful Tool in Molecular Crystal Engineering D. Braga, L. Maini, M. Polito, F. Grepioni Hydrogen-Bonded Supramolecular Chain and Sheet Formation by Coordinated Guanidine Derivatives P. Hubberstey, U. Suksangpanya Hydrogen-Bonding Templated Assemblies R. Vilar Hydrogen Bonded Network Structures Constructed From Molecular Hosts M. J. Hardie
Struct Bond (2006) 121: 1–47 DOI 10.1007/430_021 © Springer-Verlag Berlin Heidelberg 2006 Published online: 15 February 2006
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies Imenne Bouamaied · Timur Coskun · Eugen Stulz (u) Department of Chemistry, University of Basel, St. Johanns-Ring 19, 4056 Basel, Switzerland [email protected] 1 1.1 1.2 1.3
General Aspects of Axial Coordination to Metalloporphyrins . . . . . . Thermodynamic and Kinetic Stability of Axial Coordination Compounds Characterization of Axial Coordination Compounds Using Spectroscopic and Mass Spectrometric Methods . . . . . . . . . . . Geometry of Axial Coordination Compounds . . . . . . . . . . . . . . . .
. .
2 3
. .
4 7
.
7
2.2 2.3
Nitrogen Ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mono-porphyrin Complexes with Functionalised Pyridyl and Imidazolyl Ligands . . . . . . . . . . . . Bis- and Multi-Porphyrin Complexes through Bis-Nitrogen Ligands . . . Multi-Porphyrin Complexes through Tri- and Polytopic Nitrogen Ligands
. . .
8 12 23
3
Miscellaneous Ligands: O, S, Se . . . . . . . . . . . . . . . . . . . . . . . .
25
4 4.1 4.2
Porphyrins as Ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pyridyl-Porphyrin Binding to Zinc, Ruthenium and Rhodium Porphyrins . Mixed Multi-Porphyrin Assemblies . . . . . . . . . . . . . . . . . . . . . .
28 29 32
5
Dynamic Combinatorial Libraries . . . . . . . . . . . . . . . . . . . . . . .
38
6
Porphyrin-Fullerene Complexes . . . . . . . . . . . . . . . . . . . . . . . .
40
7
Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
2 2.1
Abstract The use of axial coordination to metalloporphyrins is discussed on the basis of constructing multinuclear complexes. Starting with single metalloporphyrin-ligand complexes where the ligand is designed to bring a functional moiety close to the porphyrin, the discussion further expands to the design, synthesis, and detailed analysis of multiporphyrin assemblies. The porphyrin-as-ligand concept combined with orthogonal binding modes is presented, and selected examples show that in this way a large diversity in multiporphyrin assemblies can be achieved. New emerging concepts such as dynamic combinatorial chemistry, porphyrin-fullerene complexes and porphyrins assembled around gold nanoclusters or on surfaces are presented as well, because these systems are expected to play a leading role in the design of new materials in near future. Keywords Metalloporphyrins · Axial coordination · Non-covalent multiporphyrin assembly · Porphyrin-as-ligand · Dynamic combinatorial libraries · Self-assembled monolayers
2
I. Bouamaied et al.
1 General Aspects of Axial Coordination to Metalloporphyrins Nearly every metal of the periodic table has been inserted into porphyrins. A detailed discussion on the properties of metalloporphyrins and their axial ligands can be found in the Handbook of Porphyrins and Related Macrocycles [1, 2]. The square planar arrangement of the tetrapyrrole chelate ligand (Fig. 1) greatly reduces both the number of metals suitable for having an additional fifth or sixth ligand, and the type of ligand to be recognized by the metalloporphyrin. Firstly, the metal is required to be coordinatively unsaturated when located in the cavity of the porphyrin; otherwise it will not be useful for the construction of multinuclear complexes using the axial ligand approach. For this reason, central metals like nickel(II), platinum(II) or palladium(II) can already be ruled out as “useful metals”. Secondly, the central metal is normally located in the center of the porphyrin, thus monodentate ligands will be most convenient for axial coordination. However, in some rather exotic metalloporphyrins, such as zirconium(IV), the metal is located about one angstrom above the plane of the porphyrin, thus giving access to the complexation of chelate ligands. These have so far not been used in the creation of multinuclear assemblies and will thus not be covered here. Thirdly, synthetic accessibility and magnetic properties should also be taken into account. Some metalloporphyrins require harsh conditions during metallation, thus they may not be suitable in combination with delicate substituents. For ease of characterization using NMR spectroscopy (vide infra), diamagnetic metals are most convenient. Because of the ease of synthesis, nitrogen ligands contribute an overwhelming majority of the donor atoms, and the respective Sect. 2 is devoted to those ligands. Selected examples are discussed with increasing complexity in the systems. Some subsections are starting with systems that are intended to study the metalloporphyrin-ligand interaction in detail (if applicable), followed by examples with specific applications. It is not possible to provide full coverage of the field; therefore many potentially interesting complexes had to be left out of this review. Section 3 includes the remaining important donor atoms for axial coordination, and summarizes the recent systems com-
Fig. 1 General structure and cartoon representation of the metalloporphyrin and its ligands
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
3
posed of oxygen, sulfur and selenium complexes. Section 4 covers many of the multiporphyrin assemblies realized to date using either pyridyl-porphyrin or multiple orthogonal bindings. The concept of porphyrin-as-ligand in the construction of supramolecular assemblies will be discussed based on selected examples. Section 6 outlines how the new concept of dynamic combinatorial chemistry can be applied to selectively synthesize multiporphyrinic assemblies. In Sect. 7 porphyrin-fullerene complexes are briefly presented. Though the C60 is not directly coordinated to the central metal but the association is through π – π-interactions with the porphyrin core and displaying similar geometrical features, the importance of this new class of assemblies makes it worthwhile to be addressed. It should be noted that in some cases cartoon representations of the porphyrins will be used due to the increasing complexity of the systems as we go on in the discussion. These are subsequently introduced in the course of the sections. In many cases the side chains on the β-pyrrole positions or the substituents on the meso-phenyl groups have been omitted (which may not explicitly be mentioned) as their role is primarily related to solubility control, and the reader is referred to the original literature for full structural details. 1.1 Thermodynamic and Kinetic Stability of Axial Coordination Compounds It is important to have an understanding of both the thermodynamic and kinetic stability of axial coordination compounds before attempting to construct a supramolecular assembly using these features. The simple reason is that if the mode of connectivity is badly chosen, the systems will not be stable in solution, and dissociation to yield the individual building blocks will be inevitable. The basic concepts are discussed in more detail in other chapters in this book; therefore we will give only a brief overview. The association constants Ka span a wide range from about 102 M–1 to 9 10 M–1 . From a supramolecular point of view, the higher the better to construct stable arrays. For Ka -values at the lower end of the scale, spectroscopic analyses will reveal mixtures of associated and dissociated species, whereas systems connected by strong metal-ligand interactions will show mainly the coordinated assemblies. As a rule of thumb, the binding affinities to the same metal in a ligand series follow the pKa values of the ligand, which roughly reflects the σ -donor properties of the ligand atom. However, in some cases, and especially with phosphorus as donor atom, π-back bonding can contribute an important factor to the stability of the axially coordinated ligand. The following series can generally be applied to predict the relative stabilities of complexes, if steric effects do not interfere: H2 N – R > HN – RR > N – RR R > Imidazole > Pyridine – OH > – O – R > – O C – R 2
4
I. Bouamaied et al.
Ru(II)/Rh(III) – N > Zn(II) – N Zn(II) – N > Zn(II) – P > Zn(II) – O Ru(II)/Rh(III) – P > Ru(II)/Rh(III) – S > Ru(II)/Rh(III) – N The thermodynamic stabilities do not always correlate with the kinetic parameters of ligand exchange reactions, which can be very fast despite high Ka values. This is particularly the case for soft metal-ligand interactions like Ru(II)/Rh(III)-porphyrins or Zn(II)-porphyrins with amines and pyridines. On the other hand, hard Lewis acid-base interactions tend to be kinetically more inert, as can be seen in the examples of tin(IV) carboxylate or zirconium(IV) carboxylate complexes. Thermodynamic data for these complexes are rather scarce because of the slow ligand exchange. The kinetic lability of axially coordinated ligands has to be taken into account, when supramolecular multinuclear arrays are constructed which include several different binding modes, as we shall see later in the chapter. Multipoint complexes will, in this case, certainly add to the overall stability of the array. 1.2 Characterization of Axial Coordination Compounds Using Spectroscopic and Mass Spectrometric Methods Most conveniently, the porphyrin core is an 18 π-electron aromatic system that provides us with a variety of specific properties useful in spectroscopic analysis of axial ligand coordination. The aromatic system induces a ring current, and the ligands are thus located in an electronically shielding region that can easily be detected using 1 H NMR spectroscopy; of course other nuclei such as 13 C or 31 P serve as useful probes as well. Since NMR measurements of paramagnetic metal complexes is rather tedious (but not impossible), most researchers restrict the design of multinuclear complexes to diamagnetic metal complexes. A representative example of induced chemical shifts is shown in Fig. 2a (the resonances of the uncomplexed ligand appear at δ 7–7.5 ppm). Best results may be obtained by keeping the porphyrin concentration in the millimolar or hundred micromolar ranges. It should be noted that in the 1 H NMR spectra, an up field shift is observed upon ligand binding. Contrary to this, the nucleus which is directly bound to the central metal may experience a down field shift. This is most probably due a stronger deshielding effect resulting from the σ -donation of electron density upon bond formation as compared to the shielding effect of the porphyrin. For many of the larger oligomers which may not possess a high degree of symmetry, unambiguous structure determination is far from trivial. Multiple interconverting isomers can lead to NMR spectra with broad lines and many resonances, and only strong binding and restricted conformations leads to NMR spectra displaying sharp lines and characteristic shift and symmetry patterns.
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
5
Fig. 2 Spectra of a ruthenium(II) porphyrin phosphine complex: a 31 P and 1 H NMR spectra, DPAP denotes the free ligand, (DPAP)Ru(CO)(DPP) the mono-phosphine complex; b UV-vis spectra of the porphyrin and the complex; c MALDI-ToF MS of the complex
The extended π-electron system of the porphyrin has a very characteristic absorbance spectrum with a major maximum at λmax of about 400 to 450 nm. The exact value depends very much on the nature of the metal, its axially bound ligands, and on the substitution pattern of the porphyrin. This absorbance is denoted B-band (or Soret band), and stems from the second symmetry allowed transition. Weaker absorbances can be found in the range of 450 to 700 nm, denoted as Q-bands (first symmetry forbidden transitions). The number of peaks and their relative intensities depend on the same features as for the B-band. Tables with data are numerous [3], and there is no point in going into detail of the aspects influencing the overall spectrum of a given porphyrin here. However, generally a bathochromic shift is induced when an axial ligand is bound, which is sometimes accompanied by a hypochromicity of the absorbance. This feature allows for both the detection of ligand coordination and for the measurement of association constants through titration experiments using UV-vis spectroscopy. A typical example is given in Fig. 2b. The very high molar absorption coefficient of ∼ 105 allows the detection of interactions in the concentration range of 10–4 to 10–6 M. If for some reason the absorbance
6
I. Bouamaied et al.
spectrum should not show clear changes upon ligand coordination, the measurement of the emission spectra might provide a useful alternative. The fluorescence—if it is not quenched by internal relaxation processes as in iron or nickel porphyrins—has a maximum which can be found in the region of 500 to 700 nm. The emission intensities are very sensitive to the nature of the axially bound ligand(s), thus fluorescence spectroscopy can be applied to detect the metal coordination, usually at very low concentrations (10–7 to 10–10 M). The high absorption coefficient of the porphyrins renders these compounds suitable for analysis using MALDI-ToF mass spectrometry, mostly without the need of an additional matrix, which is then referred to as LDI [4]. In fact, simple porphyrins such as TPP can itself be used as a matrix in MS analysis. The response in the MS is usually very high, and hardly any fragmentation occurs. In case of problems to obtain a decent mass spectrum, addition of 4-Nitroanilin as a neutral matrix may help. Other ionization methods like ESI or FAB have also been used. When studying metal-ligand interactions, the medium (solvent, matrix) might strongly interfere through ligand exchange reactions. Therefore, the use of strongly coordinating solvents and acidic or basic matrices should be avoided. We found that LDI provides the most versatile method for ionization, and direct interactions between the central metal of the porphyrin and its potential axial ligand can be observed (Fig. 2c). To detect chirality in the supramolecular systems, CD spectroscopy provides another powerful tool. Again, it is the high absorption coefficient of the aromatic core that leads to the simple detection of excitonic interactions between the chromophores. So far, CD spectroscopy of porphyrins has mainly been used to determine their binding modes to DNA, or to assign the chirality of axially bound ligands. Examples of studies with multinuclear porphyrin arrays are rather limited; therefore this method has yet to be fully recognized for its multipurpose applicability. Infrared spectroscopy, on the other hand, used to be widely applied in studying simple complexes of metalloporphyrins. Nowadays, this analytical method is only rarely used though it can provide valuable information about the metal-to-ligand interactions such as π-back bonding effects. But this knowledge is usually less important in the design of multinuclear complexes and becomes only essential when a detailed understanding of new interactions is required. Since the most convenient methods of analysis are 1 H NMR and UVvis spectroscopy, together with mass spectrometry, the synthesis of multinuclear arrays using axial coordination has mainly focussed on the use of diamagnetic metals. Together with the ease of synthesis of both metalloporphyrins and the ligands, the vast majority of systems are therefore comprised of the Zn(II)-nitrogen, Ru(II)- or Rh(III)-nitrogen and Sn(IV)-oxygen interactions.
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
7
1.3 Geometry of Axial Coordination Compounds The term “axially bound” already indicates that the fifth and sixth ligands on the metalloporphyrin form the corners of a distorted octahedron comprised of the four pyrrole nitrogen atoms and the respective ligands (Fig. 3). The metalligand bonding usually is near to perpendicular with respect to the porphyrin, but deviations of up to 10 degrees are not unusual. As a first approximation, it can be assumed that a pyridine ligand will be orthogonal to the porphyrin, and the substituents on a phosphine ligand will be at an angle which is determined by the tetrahedral structure of the phosphorus atom. The same of course applies to imidazoles, amines, and alcohols etc. as ligands. Carboxylates are monodentate ligands if the metal lies within the plane of the porphyrin [e.g., Sn(IV)], but are bidentate chelate ligands if the metal is coordinated above the porphyrin plane [e.g., Zr(IV)]. To predict the geometry involves both the hybridization of the donor atom as well as its specific substitution pattern. To apply the simple structure model without any large deviation from the ideal coordination geometry normally gives a fairly good starting point to get an idea of the overall geometry which the array will adopt. The cartoon representation in Fig. 3 shows the T-shape (or cross-shape) pattern as one idealized model, and the directionality of the ligands at various angles.
Fig. 3 Cartoon representation of various idealized geometries in metalloporphyrin complexes
2 Nitrogen Ligands Attachment of pyridyl or amino groups to electronically active groups such as bipy or terpy transition metal complexes, fullerenes, TTF, or porphyrins
8
I. Bouamaied et al.
Fig. 4 Schematic representation of the mono-amino complexes
is easily achieved. Because of the potential interest in studying the electronic interactions between two different porphyrins, or between porphyrins and other electrochemically or photophysically active molecules, the axial coordination of metalloporphyrins with substituted nitrogen ligands has gained widespread applications. In fact, the vast majority of non-covalent supramolecular systems are built through this approach. This chapter deals with complexes that have been synthesized through axial coordination of nitrogen ligands except porphyrins as ligands; these multiporphyrin arrays will be discussed in Sect. 4. Simple complexes with pyridine or other amino ligands will not be covered here, but only those where additional functionalization to yield new assemblies is introduced, e.g., for the study of photoinduced electron transfer. These complexes can be generalized with cartoon representations as shown in Figs. 4, 9 and 22 in this section. 2.1 Mono-porphyrin Complexes with Functionalized Pyridyl and Imidazolyl Ligands A systematic variation in the redox potential of imide-substituted pyridine as ligands to ZnTPP by Otsuki shows how the rate of electron transfer in a supramolecular assembly can be varied (Fig. 5) [5]. The photo irradiated ZnTPP in 1 acts as an electron donor, and the electron transfer rate was measured through quenching of the porphyrin fluorescence. In a range of 0.8 V difference in redox potential of various ligands, the electron transfer yield varies from 0.11 to 0.91. The plot of electron transfer rate vs. the free energy gap ∆G◦ between the singlet excited state of ZnTPP and the acceptors shows that increasing the energy gap results in an increased electron transfer rate, which according to classical Marcus theory is an expected “normal” behavior. Whether the electron transfer occurs through space or through intermolecular bonds could not be differentiated. Further extension to a ternary complex 2 (Fig. 5) showed similar results [6]. This time additional hydrogen bonding (salt bridge) to the electron acceptor was introduced. The spacer, N-propyl-isonicotinamide, binds nitro-benzoic acids through amidinium-carboxylate salt formation on one side, and on the other side has a pyridine moiety for axial coordination to ZnTPP. The overall thermodynamic stability of the system was determined to be 1.9 × 108 M–1 in
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
9
Fig. 5 ZnTPP complexes, where the porphyrin acts as electron donor (1, 2) [5, 6] or electron acceptor (3) [7] from the photoexcited state
DCM. The fluorescence of the porphyrin was quenched with an intracomplex quenching rate of 1.9 × 109 s–1 . Photoinduced intracomplex electron transfer was suggested as a likely mechanism of the fluorescence quenching and triplet bleaching. In these two cases, the electron was transferred from the porphyrin to the axial ligand. When using a pyridyl-ferrocene as ligand instead in the complex 3, the electron transfer now occurs from the ferrocene moiety to the porphyrin (Fig. 5) [7]. The relative energies of the ground and excited states were determined by luminescence spectroscopy and cyclic voltammetry and help to understand the fluorescence quenching of the porphyrin, which is due to electron transfer from the ferrocene to the singlet excited state of the zinc(II) porphyrin. The life time of the charge-separated state Por– -Fc+ is too short to be detectable using transient absorption spectroscopy due to ultrafast back electron transfer. A supramolecular system capable of responding to light by undergoing reversible changes in structure was developed by Branda [8, 9]. The Lewis basicity of a pyridine-functionalized 1,2-dithienylcyclopentene (Fig. 6) was modulated by interconverting the compound between its insulated ring-open (4) and electronically connected ring-closed (5) form by irradiation with the appropriate wavelength, as demonstrated by the binding affinity to a ruthenium(II) porphyrin. The ring-closed form, which is conjugated to the electron deficient pyridinium cation, has a decreased nucleophilicity compared to the “isolated” pyridine in the ring-open form, which results in a lower binding constant to the metalloporphyrin in 7 as compared to the complex 6. This outcome is in accordance with the general observation when electron-withdrawing substituents are placed on nitrogen ligands. The decreased affinity can be monitored by 1 H NMR and UV-vis spectroscopy, and a cycling experiment showed no degradation of the system even after prolonged irradiation (365 nm and > 490 nm for > 2 h). This example shows that a supramolecular event can modulate the reactivity of electronically active ligands to metalloporphyrins.
10
I. Bouamaied et al.
Fig. 6 Modulation of the reactivity of an electronically active ligand [8, 9]
Various fullerenes (C60 and C70 ) were coordinated to zinc(II) and magnesium(II) porphyrins via functionalized pyridines or imidazoles (Fig. 7) by D’Souza and Ito [10–22]. Both “single-point” [10, 11] and “two-point” [12–14] binding strategies were employed, together with additional covalent functionalization of the porphyrins with ferrocene (Fc) [10] or boron dipyrrin (BDP) [16]. Similar systems were also studied by Guldi, Diederich, Nierengarten and Schuster, and the results on the intermolecular and supramolecular photoinduced electron transfer (PET) processes of fullerene-porphyrin and phthalocyanine systems were reviewed recently [23, 24]. Since PET is
Fig. 7 a Two-point binding of a pyridine-fullerene to a ZnTPP-carboxylate [12], and b ab initio B3LYP/3-21G(∗ )-optimized geometry of a supramolecular triad formed by an imidazole-C60 and a ZnTPP-Fc (right); reprinted with permission from [10], Copyright (2004) American Chemical Society
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
11
of great importance in chemistry and biology, and since fullerenes exhibit a number of characteristic electronic and photophysical properties that make them promising candidates for the investigation of PET processes, the design of supramolecular systems combining the porphyrin moiety with fullerenes has gained much attraction (see also Sect. 7). Supramolecular systems composed of functionalized fullerenes that are coordinated to the central metal of the porphyrin have mainly been studied to mimic the photosynthetic system. If the self-assembled donor-acceptor system is properly designed so that a sufficient amount (> 99%) of the complex is present in solution, intramolecular PET occurs from the short-lived singlet excited state of the porphyrin to produce charge separated species. The axial coordination allows controlling the orientation and the distance of the donor and acceptor moiety, which is an important factor in tuning electron transfer efficiency. The association constants for the N-fullerene ligands are in the range of 104 M–1 (pyridyl) to 105 M–1 (imidazolyl), and the binding events can be monitored by UV-vis titration. The estimated electron transfer rates to generate the charge separated complexes range from ∼ 5 × 107 s–1 to ∼ 2 × 1010 s–1 with quantum yields between 10 and 99% and lifetimes in the nanosecond range. The systems are very sensitive to the polarity of the solvent and the relative orientation of the ligand (i.e., the fullereneporphyrin distance). Coordinating solvents such as benzonitrile can compete with the ligand, and generate solvent-separated ion-pairs after dissociation of the ionic species. Noteworthy, metal-coordinated fullerene-ruthenium(II) porphyrin dyads can be deposited on surfaces via Langmuir–Blodgett film deposition [25]. Covalent attachment of electronically active groups such as N,N-dimethylaminophenyl (8) [12] or ferrocene (9) [10] as a second electron donor further improves the systems performance (Fig. 7). After photoirradiation of the Donor-Por-C60 system, a charge-separated state Donor+ -Por-C60 – is generated with high quantum efficiency and moderate charge stabilization. In the BDP-Por-C60 complex, the boron dipyrrin unit acts as an antenna unit in that it transfers its excited state energy to the porphyrin, which is followed by electron transfer to the fullerene to give the complex BDP-Por+C60 – , reminiscent of the combined antenna-reaction center events in natural photosynthesis. Photoinduced electron transfer in multicomponent arrays was also detected in the substituted porphyrin ZnTBTPP-heterofullerene (C59 N) dyads 10 by Guldi and Hirsch [26] [TBTPP = tetra-(tert-butyl-phenyl)porphyrin]. Mannich functionalization of the dimer of C59 N yielded a pyridine donor substituted heterofullerene (Fig. 8), which coordinates to the zinc(II) porphyrin to give a quasi-linear donor-acceptor ensemble. The heterofullerene system seems to be advantageous over the C60 system in that a nitrogen replacement creates an open shell system, and the C59 N derivatives are better electron acceptors. Detailed fluorescence titration and time-resolved flu-
12
I. Bouamaied et al.
Fig. 8 ZnTBTPP-heterofullerene dyads by Guldi and Hirsch [26]
orescence measurements, together with measurements in various solvents differing in the dielectric constant, revealed that either intramolecular energy or electron transfer from the porphyrin to the heterofullerene takes place, depending on the solvent used. In toluene, the large energy gap (– ∆G = 0.52 eV) for the energy transfer route evokes deactivation of the porphyrin singlet excited state via energy transfer, whereas in o-dichlorobenzene a metastable charge-separated radical pair (ZnTBTPP)+ -pyridine-(C59 N)– is generated. Thus, the combination of porphyrins with heterofullerenes adds a new system to the already intensively studied porphyrin-C60 complexes. 2.2 Bis- and Multi-Porphyrin Complexes through Bis-Nitrogen Ligands Two of the “classical” bis-nitrogen ligands to assemble porphyrins (predominantly zinc metallated) are the bidentate 1,4-diazabicyclo[2.2.2]octane (DABCO) and 4,4 -bipyridine (bipy). Other di-topic nitrogen-ligands studied include hydrazine, 4,4 -bipyrimidine, diaza-pyrene, 5,5 -dicyano-2,2 -bipyridine, perylene-bisimide, extended bipyridine units and non-covalently linked di-pyridyl ligands via bipy-metal interactions. The formation of various bisporphyrin and multiporphyrin assemblies using these versatile ligands were studied in detail by Sanders, Anderson, Hunter, Branda and many others. The structures of the supramolecular multi-porphyrin architectures are schematically represented in Fig. 9. The system is analytically complicated by the presence of two competing binding sites, which can undergo fast or slow ligand exchange depending on the solvent used [27]. For DABCO (Fig. 10), the single binding K11 to a zinc(II) porphyrin in toluene is ∼ 106 M–1 , whereas the value drops to
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
13
Fig. 9 Schematic representation of the porphyrin complexes obtained from ditopic nitrogen ligands
Fig. 10 Formation of DABCO-bridged porphyrin sandwich complexes to study the cooperativity in binding by Anderson [27]
∼ 105 M–1 in CHCl3 due to the higher dielectric constant and its ability to solvate the ligand by acting as a hydrogen bond donor. Under dilute conditions (∼ 10–6 M) which are used for UV-vis measurements only the monomeric complexes are observed. Using higher concentrated solutions for 1 H NMR studies (∼ 10–3 M), both 1 : 1 and 2 : 1 porphyrin : DABCO complexes 11 and 12, respectively, can be detected, depending on the ratio porphyrin to ligand used. The association constant K21 for the second binding drops to a value of ∼ 3 × 103 M–1 in toluene and ∼ 5 × 103 M–1 in CHCl3 . The interaction factor, α, which is defined as α = 4K21 /K11 , is dependent on the structure of the porphyrins used. In the case of free meso-positions, α is in the range of 0.7–1.0 indicating rather little cooperativity, whereas in the case of mesodiphenyl porphyrins α 1 indicates negative cooperativity in the system. This is attributed to steric factors in the bis-porphyrin complex. The value of α, however, appears to depend strongly on the system studied, and changes in substituents and geometry will have great impact [28]. The concept was extended to conjugated ladder type complexes consisting of two to six porphyrins, and a detailed analysis revealed all-or-nothing formation of the complexes (no partially formed ladders were observed) and self-sorting of mixed systems in analogy to poly-bipy metal complexes [27]. An increasing Hill coefficient reflects increasing cooperativity of ladder assembly with
14
I. Bouamaied et al.
increasing length of the porphyrin conjugate, which is also reflected in the increased stability of the complexes (determined by the Gibbs free energy). This is analogous to the DNA duplex formation. Changes in the UV-vis spectra show that both DABCO and bipy ladders are conformationally restricted and more ordered that in the “single-strand” form. The binding of various bi-dentate ligands to ruthenium(II) and rhodium(III) porphyrins have been studied thoroughly by Sanders et al. [29–33]. As already mentioned in Sect. 1, the binding of nitrogen ligands to ruthenium(II) or rhodium(III) porphyrins is several orders of magnitude stronger than to the corresponding zinc(II) porphyrins. Therefore, stable bisporphyrin complexes are readily formed in solution; however, the formation of monomeric rhodium(III) porphyrin-ligand adducts is reported upon addition of excess ligand [33]. In the case of the ruthenium(II)-DABCO system, a slow equilibrium between mono- and bis-porphyrin complexes is observed by NMR spectroscopy [29]. The complexes shown in Fig. 11 were analyzed thoroughly by 1 H NMR spectroscopy, and of most of the compounds single crystals could be grown which were suitable for X-ray analysis. This provides a very useful set of data for comparing the different metal-ligand interactions both in solution and in the solid state. The complexation induced chemical shifts ∆δ of the species are almost exactly the sum of the shifts induced by one porphyrin, implying additivity of the ring current effects. Since the signals are sharp and well resolved, slow exchange between bound and unbound species (if present) on the NMR-time scale can be assumed. Also, the occurrence of only one set of signals hint for rapid rotation around the metal-ligand bond, and an overall highly symmetrical structure results in solution. This is corroborated by the fact that ∆δ for the ligands is almost identical in all cases for protons with comparable distances to the porphyrin plane. The solid state structures, on the other hand, differ substantially when the bridging ligand is changed. Tilt angles between the aromatic ligand and porphyrin best-fit planes can be as small as 71◦ . Deviation from the “ideal” geometry is most dramatic in the complex with the (much weaker) ligand 5,5 -dicyano2,2 -bipyridine. Using sp3 -hybridized nitrogen ligands such as hydrazine and substituted hydrazines instead of aromatic ligands changes the relative orientation of the cofacially arranged porphyrins in the complexes (Fig. 11) [32]. Both rigid and flexible porphyrin sandwiches were studied by Hunter et al. [28, 34]. A rigid bis-porphyrin, connected via naphthalene diimide (Fig. 12), theoretically forms a number of complexes with DABCO, depending on the porphyrin to ligand ratio. To characterize the supramolecular system, and to find the optimum conditions for cyclization to form the sandwich complex, the effective molarity (EM) was determined by 1 H NMR and UV-vis titration experiments. The EM may represent a composite of the real effective molarity and cooperative effects (α), but it is nevertheless a useful parameter for predicting the stability of the sandwich complex. Here, the EM was determined to be 2 mM, which means that the macrocyclic complex 14 is
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
15
Fig. 11 a Schematic representation of the bis-porphyrin complexes investigated by Sanders et al. [29–33], b X-ray structure of a 5,5 -dicyano-2,2 -bipyridine-bis-Rh-porphyrin complex; reprinted with permission from [33], Copyright (2001) American Chemical Society, c X-ray structure of a hydrazine-bis-Rh-porphyrin complex; reprinted with permission from [32], Copyright (2001) American Chemical Society
Fig. 12 Rigid bis(naphthalene-diporphyrin-DABCO) complex by Hunter [28]
only present when the concentration is significantly lower than this value. At higher concentrations, insoluble polymers were formed. In a series of flexible zinc(II) bis-porphyrins sharing the same linker but differing in the substitution pattern (ortho, meta and para in the mesophenyl group), the self-assembled porphyrin arrays with DABCO show different structures as shown in Fig. 13 [34]. A detailed analysis using UV-vis
16
I. Bouamaied et al.
and 1 H NMR titrations, together with simulated speciation profiles, revealed a very different behavior for the three different bis-porphyrins. The ortho- and meta-substituted bisporphyrins form 1 : 1 intramolecular sandwich complexes. In both cases, upon addition of an excess of DABCO, the porphyrin and DABCO resonances assigned to the sandwich complex gradually shifted in the 1 H NMR spectra, indicating fast exchange equilibrium between the sandwich complex and a 2 : 1 DABCO-porphyrin complex. The parasubstituted bisporphyrin cannot adopt the cofacial conformation required for this type of complex and forms a higher-order 2 : 2 intermolecular assembly, which is stable over a wide range of DABCO concentrations. Molecular modelling revealed the energetic reasons for this behavior. The modelled intramolecular complex of DABCO with the para-isomer is 54 kJ mol–1 less stable than the meta-isomer, because a cofacial arrangement of the porphyrin units can only be achieved by distorting the planarity of the amide groups. On the other hand, molecular modelling shows that the para-isomer can form a strain-free 2 : 2 assembly with DABCO. This complex is 50 kJ mol–1 per bisporphyrin more stable than the 1 : 1 complex. By substituting the porphyrins with electron withdrawing groups (EWG), electron-deficient porphyrin building blocks are obtained and show significant differences in the binding of axial ligands. Most conveniently the EWG effect is achieved by introducing meso-C6 F5 groups. Though the perfluorophenyl groups have little effect on the geometry of the porphyrin, the electronic impacts are much larger and are best described by the four-orbital model for the porphyrin ground and excited states. The theory and influence on the electronic spectra and redox potentials shall not be discussed here (see the chapter by J. Hupp in this series). The main point is that the perfluorophenyl-substituted porphyrins bind axial ligands much stronger, i.e., nearly an order of magnitude higher in the case of pyridines to porphyrins such as shown in the sandwich complex of 15 with bipy (Fig. 14) [35]. The enhanced Lewis-acidity of the rhenium-linked bis-porphyrin relative to the analogous non-fluorinated porphyrin system allows for the formation of a somewhat more robust supramolecular architecture compared to an analogous non-fluorinated sandwich-complex [36].
Fig. 13 AM1 optimized structures of the DABCO-porphyrin complexes by Hunter: a ortho-isomer, b meta-isomer, c both 1 : 1 and 2 : 2 sandwich complexes of the paraisomer. Reprinted with permission from Wiley from [34]
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
17
Extended bis-nitrogen ligands such as open-chain and rigid crossconjugated bis-pyridines provide versatile alternatives to both DABCO and 4,4 -bipy in supramolecular chemistry, because they allow to vary both distance and geometry of the assembly. Heterodimeric porphyrin macrocycles were obtained from the binding of a covalently linked bis-porphyrin with the appropriate bis-pyridyl ligand (Fig. 15) [37]. Since both the bis-porphyrin and the bis-pyridyl ligand are conformationally flexible, a range of strained conformations in the macrocycle are possible. The stability constants Ka for a series of macrocycles are in the range of 1.1 × 105 M–1 to 1.3 × 107 M–1 with values for the EM ranging from 0.02 to 6 M, as determined by UV-vis titration. The differences in stability were attributed to the linker in the bisporphyrin system. A 2,6-carboxamido-pyridine linker favors the necessary conformation through internal hydrogen bonding, which is not the case for
Fig. 14 Sandwich complex of an electron-deficient bis-porphyrin [35]
Fig. 15 Bis-porphyrin macrocycle stabilized by a ditopic ligand [37]
18
I. Bouamaied et al.
the analogous arene linker. The linker in the bis-pyridyl ligand, on the other hand, has little or no effect on the stability of the macrocycle. The macrocycle 16 has amide groups directed into the cavity which are capable of forming hydrogen bonds with other amides such as terephthalimides or alkyl diamides. The binding of a diamide moiety in the cavity leads to the formation of pseudorotaxanes and rotaxanes, as evidenced by 1 H NMR spectroscopy. The formation of the rotaxanes would not be possible without the structure stabilization through the ditopic ligand. The extended bipyridine ligand, based on a 3,5-diethynyl pyridine building block, is capable of binding to two ruthenium(II) porphyrins, thus forming discrete macrocyclic complexes such as 17 (Fig. 16) [38, 39]. Unlike the short 4,4 -bipy unit, this extended conjugated macrocyclic ligand has the potential of forming porous solid-state structures, and exocyclic substituents allow the manipulation of the physical properties such as solubility and stability. The thus obtained complexes were characterized both in solution (1 H NMR, UV-vis and luminescence spectroscopy) and in the solid state (X-ray crystallography). It is noteworthy that some of the complexes are stable enough to be chromatographed on silica or alumina. The solid state analysis revealed clear differences in the relative conformation of the enyne core of the macrocycle. Both planar and chair-like conformations can be adopted depending on the substitution pattern on the ligand, thus very different crystal packing induced by these conformational changes is observed. Self-organization of multiporphyrin arrays into well-defined structures, where two or more porphyrins are covalently linked together, form another set of supramolecular structures arranged through axial coordination. A combination of porphyrins with C60 has been realized in several ways (Fig. 17). The C60 unit forms a covalent bridge between two porphyrins, which are then oriented in a cofacial arrangement via DABCO-binding as in
Fig. 16 General structure of the extended macrocyclic complex [38] and X-ray structure of a complex displaying the chair conformation of the bridging ligand; reprinted with permission from [39], Copyright (2002) American Chemical Society
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
19
18 [40]. Alternatively, a strapped porphyrin-fullerene is synthesized, which is then dimerised with DABCO to form a π – π stacked porphyrin-fullerene dyad (19) [41]. Or the fullerene itself is substituted with two pyridyl units, thus forming a porphyrin:fullerene:porphyrin array 20 via axial coordination (viz. Sect. 2.1) [15, 23]. In all cases, the occurrence of charge-separated species was reported. However, in the example of DABCO-driven structure stabilization, fundamentally different deactivation pathways were observed depending on the presence or absence of the axial ligand. Supramolecular donor-acceptor systems based on a zinc(II) porphyrin dimer and a bidentate acceptor ligand allows control of the relative orientation and thus fine-tuning of electron transfer between the units [42, 43]. The binding constant between the zinc(II) porphyrin and the dipyridine-quinone in 21 was estimated to be ∼ 107 M–1 , and the rate of charge separation was obtained to be 1.6 × 1010 s–1 , based on time-resolved single-photon counting fluorescence studies. Again,
Fig. 17 Donor-acceptor systems which show formation of charge-separated species upon photoirradiation [15, 23, 40–43]
20
I. Bouamaied et al.
the kinetic lability of the coordination could lead to breaking of the system immediately after electron transfer, leading to a charge-separated state with long lifetime. Multicomponent porphyrin assemblies were used as a platform to create catalysts by Reek [44–48]. A series of stable arrays, all incorporating a rhodium or palladium phosphine or phosphite complex, were synthesized and investigated in their ability to maintain catalytic activity. The supramolecular complex 22 (Fig. 18) shows one example. All systems proved to be highly active and regioselective in hydroformylation of alkenes or in allylic alkylation reactions. These examples demonstrate in an elegant way that multiporphyrin arrays do not only serve to model the photosynthetic reaction center, but highly active catalysts can be designed and synthesized. Here, the axial coordination chemistry to porphyrins is not used to arrange the porphyrins in a specific way, but rather the porphyrins serve to form an active catalyst via geometry stabilization of the ligated metal complex. A porphyrin calixarene conjugate 23 was also self-assembled using DABCO as a ligand to the zinc metallated porphyrins [49]. A calix-bisporphyrin forms a simple 2 : 2 complex, whereas a calix-tetraporphyrin showed remarkable flexibility (Fig. 19). Initially, an intramolecular 1 : 1 complex between two (probably adjacent) zinc(II)-porphyrins and DABCO is formed. Further stepwise addition of DABCO leads to formation of the supramolecular 2 : 4 complex, held together via four DABCO units. 1 H NMR spectroscopy reveals that the ligands in the intermediate and final complexes are in slow exchange, but after addition of excess ligand a fast exchange system is obtained consisting mainly of the (calix-porphyrin)-(DABCO)4 complex. The detailed spectroscopic analysis of this system reveals that using titration data over a wide concentration range (micromolar to millimolar) simple complexation and higher-order self-assembly can be distinguished. The conclusion is that the stability of higher-order assemblies increases significantly with concentration, whereas simple complexes are much less sensitive to concentration.
Fig. 18 Example of a supramolecular catalyst by Reek and PM3 optimized molecular structure of an active Pd-allyl complex of the catalyst. Reprinted from [45], Copyright 2005, with permission from Elsevier
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
21
Fig. 19 Self-assembly processes involved in the dimerization of a calix-tetraporphyrin 23 with DABCO; reprinted with permission from [49], Copyright (2003) American Chemical Society
The use of pyridyl-substituted terpy complexes, i.e., bis[4 -(4 -pyridyl)2,2 :6 ,2 -terpyridine]ruthenium(II) and osmium(II) complexes, allows to generate linear multicomponent arrays with ruthenium(II) porphyrins located at either end of the array (24) [50]. Both porphyrins and metal-terpy complexes have a rich photochemistry, therefore this system combines two potentially interacting chromophores via axial ligand coordination (Fig. 20). Indeed, the photophysical investigations showed intramolecular quenching of the terpy luminescence. Similarly, oxorhenium(V) compounds with monocoordinated pyrazines could be used as bridging ligands for ruthenium(II) porphyrins (25, Fig. 20) [51]. Addition of two equivalents of RuTPP to a CDCl3 solution of the preformed oxorhenium(V) complex lead to rapid formation of the assembly which showed sharp and well-resolved resonances in
Fig. 20 a Terpy-metal complex (24) [50] and oxorhenium(V) complex (26) [51] as bridging units for porphyrins; b zinc(II)-porphyrin tweezer for the stereochemical assignment of chiral ligands; the schematic picture on the right side shows the induced chirality from a face view [52–61]
22
I. Bouamaied et al.
the 1 H NMR spectrum. It can thus be assumed that no significant dissociation occurs in the millimolar range. Zinc(II) porphyrin tweezers, which can bind diamines in a host-guest complex of the general structure 26, were used to determine the absolute configuration of chiral ligands extensively by Nakanishi and Berova (Fig. 20) [50, 52–59, 61]. In this case, the axial coordination brings the two porphyrins in a chiral arrangement which is predetermined by the stereochemistry of the ligand and subsequently can be analyzed by CD spectroscopy. Depending on whether a positive or negative Cotton effect is measured, a clockwise or counter-clockwise arrangement of the two chromophores is present. From this, the geometry of the host-guest complex can be determined which directly leads to the assignment of the stereochemistry of the ligand. This method is not only applicable to diamines but can be extended to the analysis of amino alcohols or amino acids. A great challenge in the controlled organization of functional molecules into self-assembled arrays is to obtain a detailed understanding of the factors that play a role in the complex and dynamic self-assembly processes. The use of scanning probe techniques has advanced to the point where the self-assembled structures can be visualized down to the single molecule level. The porphyrin hexamer 27 reported by Nolte and Rowan [62] formed porphyrin arrays at a solid-liquid interface with DABCO as bridging ligand,
Fig. 21 a Structure of the propeller-hexaporphyrin (side-chains on the meso-aryl groups omitted). b STM topography and high-resolution image of a domain of edge-oriented 27-DABCO complexes on HOPG, c proposed consequences on the addition of DABCO resulting in a better defined columnar structure through intermolecular connection, d schematic side-view of the interconnected hexaporphyrins. Reprinted with permission from Wiley from [62]
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
23
and the process of self-assembly could be monitored using STM (Fig. 21). The propeller-like architecture forms stable Langmuir monolayers on water and micrometer sized rings on solid substrates. This is mainly due to π – π interactions between the porphyrin cores. Addition of DABCO led to the instantaneous formation of large domains (> 400 nm2 ) of lamellar assemblies, which remained stable for hours. The STM analysis revealed an edge-on arrangement, where the DABCO ligand increases the stability and definition of the lamellae. If 4,4 -bipy is used as a ligand instead, no lamellar structure was observed on the surface, but the bipy ligands act as inhibitor to columnar stacking and direct the molecules into a face-on arrangement on the surface. 2.3 Multi-Porphyrin Complexes through Tri- and Polytopic Nitrogen Ligands Polytopic nitrogen ligands such as tripyridyl-triazine, non-covalently linked multi-pyridyl ligands via bipy-metal interactions or built up through dendritic cores, were used for the formation of higher-ordered architectures. Several examples shall be presented here to demonstrate the utility of these polytopic ligands for the assembly of metalloporphyrins. The tritopic ligands tripyridyl-triazine (Py3 T) is not only versatile for the template directed synthesis of covalently linked oligo porphyrins [63, 64], but it has also been used to assemble three ruthenium(II) porphyrins around this core (Fig. 23a) [30]. Though the crystallization and X-ray analysis of the Py3 T-Por3 complex proved to be rather tedious, the crystal structure shows that the porphyrins are in a C3 -symmetrical arrangement, and the aromatic rings of the Py3 T-ligand are coplanar. A hexameric wheel, where six ruthenium(II) porphyrins spontaneously assemble into an octahedral array through a metal-directed synthesis, was achieved using a tetrapyridyl ligand [65]. This ligand, which is a 2,2 -connected bipy, can on one side form stable octahedral complexes with ruthenium or iron to build the core of the construct, while on the other side leave six pyridine groups for the assembly of either RuTPP or RuOEP to get complexes of the general structure 28 (Fig. 23b). Significant differences in the rates of formation of the arrays were observed for the two different porphyrins. The OEP-complexes were read-
Fig. 22 Schematic representation of the porphyrin complexes obtained from tritopic nitrogen ligands
24
I. Bouamaied et al.
Fig. 23 a X-ray structure of a triporphyrin assembled around Py3 T [30], reproduced by permission of The Royal Society of Chemistry (RSC) on behalf of the Centre National de la Recherche Scientifique (CNRS); b hexaporphyrin assembled around a hexatopic metalbipy complex [65]
ily formed at ambient temperature as judged from the 1 H NMR spectra, but using the sterically much more demanding TTP-derivatives the fully assembled array was only formed after heating. A combination of complementary hydrogen bonding and axial ligand coordination leads to either strapped porphyrin rosettes [66] or strapped porphyrin cleft receptors 29 (Fig. 24a) for rotaxanes [67]. In the rosettes, the porphyrins are substituted with a cyanuric acid moiety which can form a complementary hydrogen bonding pattern with melamines. The resulting hexagonal rosette incorporates three porphyrins. The porphyrins contain an additional aliphatic strap between two adjacent meso-phenyl groups, so that a potential axial ligand can only approach from one side. A self-assembled porphyrin
Fig. 24 a Strapped porphyrin cleft 29 as receptor for the synthesis of rotaxanes [67]; b gold nanoparticles (shaded sphere) coated with methyl-imidazole to bind ZnTPP or triporphyrins [70]; c self-assembled mixed multilayers of porphyrins on gold surfaces [72]
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
25
dimer (29) was shown to bind either dipyridyl molecules or dipyridyl substituted [2]rotaxanes between the aryl moieties of the ligated spacer. A dendrimer based approach was also used to assemble porphyrins [68]. Here, a covalently linked triporphyrin unit was complexed by a dendrimer which was functionalized with pyridine units. The dendrimer which can theoretically bind 16 porphyrins showed to be satisfied after binding 11 porphyrins. When imidazole units were immobilized onto gold nanoparticles via a thiol-alkyl linker, mono- and tris-porphyrin units could be arranged around this scaffold (Fig. 24b) [69]. Even though the cooperative binding of the methyl imidazoles on the nanoparticles to a tris-porphyrin is at least as good as that of a complementary tripodal ligand in solution [70], the relative gain in binding constant and effective molarity for intracomplex binding become lower with increasing amounts of porphyrins. Self-assembled monolayers (SAMs) on gold electrodes have also been serving as a platform to form metalloporphyrin monolayers (Fig. 24c) [71, 72]. The SAMs contained imidazole-terminated adsorbates and were shown to bind a series of metalloporphyrins such as ruthenium(II) and osmium(II) porphyrins. The ruthenium(II) porphyrins used contained two axially bound N2 -molecules, which are easily displayed by other nitrogen ligands. After assembly of the first monolayer, the surface is actually coated with a monolayer of ruthenium(II) porphyrins which have a labile N2 -ligand. Treatment of this monolayer with pyrazine leads to ligand exchange, and this “activated” monolayer can again be treated with a ruthenium(II) porphyrin. In this way, a system of the composition SAM-porphyrin-pyrazine-porphyrin was obtained. Applying a cycle of pyrazine and porphyrin binding using structurally different porphyrins, a mixed multilayer metalloporphyrin stack on gold electrodes could be assembled specifically.
3 Miscellaneous Ligands: O, S, Se In fact, hardly any donor atoms other than nitrogen have been used to create multiporphyrin assemblies. Using hard Lewis acids as central metal opens the possibility to bind two axial ligands on either side of the porphyrins. With ditopic ligands such as diols this can lead to polymerization. However, a few examples can be found in the literature, where the central metal of the porphyrin is axially coordinated by oxygen, sulfur or selenium to assemble porphyrins. The use of (non-porphyrinic) phosphine ligands apparently has so far not been successful in the construction of supramolecular multiporphyrin arrays. A detailed analysis of sulfur and selenium ligands has been reported by Sanders et al. [32]. Analogous to the hydrazine-diporphyrin complexes (see Sect. 2.2), both dimethyl disulfide (Me2 S2 ) and dimethyl diselenide (Me2 Se2 )
26
I. Bouamaied et al.
are capable of acting as ditopic ligands to rhodium(III) porphyrins. The 1 H NMR spectroscopic studies showed that bridging and non-bridging complexes were formed in solution, depending on the concentrations and stoichiometries used. In general, the proton resonances in the bridged systems are broadened indicating dynamic ligand exchange behavior, which was slower in the case of Me2 Se2 than with Me2 S2 . The affinity constants log(Ka ) for the monomeric binding were determined to be 5.7 for Me2 S2 and 6.6 for Me2 Se2 . Oxygen ligands form very stable complexes with hard Lewis-acid metalloporphyrins such as Sn(IV), Zr(IV), Mn(III), Mo(V), but also with P(V)-porphyrins. Mixed-metal dimers were synthesized from Al(Me)OEP and phosphorus, arsenic or antimony porphyrins in the form of µ-oxo dimers [73]. In most cases, polymeric structures are obtained because Sn(IV)-, Mn(III)- and Fe(III)-porphyrins can bind two axial ligands on either side of the porphyrin referred to as trans-binding. This is schematically represented in Fig. 25a. An example includes the linear trinuclear or polymers µ-trans-dioxo-MoTPP [74, 75]. A linear tri-nickel(II) complex as bridging ligand between zinc(II) or manganese(III) porphyrins was used to form the one-dimensional polymer 30 (Fig. 25b) [76]. The carboxylate group of carboxyl pyridine (X in Fig. 25b) binds to manganese porphyrins, whereas the pyridine nitrogen was attached to the tri-nickel(II) complex. According to UV-vis spectroscopy measurements, there is no noticeable electronic interaction between the nickel units and the porphyrins in the chains. Iron(III) chloride complexes of porphyrins are susceptible to hydrolysis, meaning that the chloride ligand is rapidly exchanged with hydroxide if treated with a hydroxide solution. The formation
Fig. 25 a Schematic representation of the polymeric porphyrin strings; b linear manganese porphyrin-polymer 30 connected through a tri-nickel-complex [76]; c optically active polymer 31 formed from chiral di-porphyrins [77]
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
27
of µ-oxo dimers brings the two porphyrin rings together in a cofacial arA. In chiral bis-porphyrin systems, rangement with a separation of about 3.8 ˚ where the two porphyrins are covalently linked through an optically active binaphthyl group, the formation of an intramolecular oxo-bridge is not possible [77]. Therefore, a polymer (31) with intermolecular µ-oxo-bridges is formed having a molecular weight and polydispersity of about 6 × 104 g mol–1 and 2.35, respectively, as estimated by gel permeation chromatography. The chiral twist of these self-assembled porphyrin dimers can be analyzed by CD spectroscopy and shows that the predetermined design of the building blocks is indeed retained in the polymer. Phosphorus(V) porphyrins are relatively simple to prepare from the freebase porphyrins and POCl3 . The central phosphorus can bind two oxygen ligands to form a trans-complex. This bonding, however, is rather covalent in nature than coordinative, but adds greatly to the diversity in creating polyporphyrin complexes. Both dimers and one-dimensional arrays connected with various lengths of either conducting (32, Fig. 26a) or insulating molecular wires were synthesized by Shimidzu [78–80]. Photoirradiation of the polymer with a conducting wire showed an enhanced conductivity, where the porphyrin acts as an electron acceptor (hole generator), and the wire tends to be an electron donor. A great variety of covalently or non-covalently bound porphyrin-terpyridine motifs are known to closely mimic the initial photoinduced electron transfer (PET) or excitation energy transfer (EET) events of natural photosynthetic reactions. A first example of covalent connection of a terpyridine group to the axial position of a phosphorus porphyrin as in 33 (Fig. 26b) was presented by Kumar and Maiya [81]. The terpy moiety itself is a strong chelator towards other transition metal ions (see also 24). Upon postirradiation, the axial terpy subunits act as a donor and P(V) porphyrin as an acceptor. In fluorescence titrations it was found that the PET and EET reactions are mod-
Fig. 26 a One-dimensional P(V)-porphyrin array connected with a conducting wire [78–80]; b axial covalent attachment of a terpy-moiety that modulates the electronic properties of the array through additional metal complexation [M = Zn(II) or Cd(II)] [81]
28
I. Bouamaied et al.
ulated by metal coordination with the terpy subunits. When excited at 566 nm in the presence of Zn2+ or Cd2+ , the PET process from the ground state of the terpy subunit to the excited state of the porphyrin was inhibited, resulting in an enhancement of porphyrin fluorescence. On the other hand, when excited at the isosbestic point at 300 nm, PET was suppressed and EET from the excited terpy subunit to the ground state of the porphyrin was enhanced.
4 Porphyrins as Ligands Substituents are in most cases relatively easy attached post-synthetic to porphyrins, or they are directly incorporated during porphyrin synthesis. The use of functionalized benzaldehydes is of course the most convenient way to do so, because they are commercially available with a very large diversity in substitution pattern. It is therefore not surprising that porphyrins themselves were designed to act as axial ligand(s) to another metalloporphyrin. As we have seen in the previous chapters, substituted pyridines provide the most extensively studied class of ligands to metalloporphyrins, hence why not having a porphyrin as a substituent to the pyridine ligand? The attachment of a meso-pyridyl group is synthetically very easy to achieve since p-formyl pyridine is cheap and reactive in porphyrin synthesis. Therefore, most of the multi-porphyrin arrays were realized by using meso-pyridine substituted porphyrins, and used in all possible combinations of stoichiometry (mono- to tetra-pyridyl porphyrin) and substitution pattern (5,10- or 5,15-di-pyridyl porphyrin). The schematic representation in Fig. 27 shows how the different substitution pattern of the pyridyl-porphyrin (arrowed square) affects the overall geometry of the array, and linear, angular, T-shape or cruciform complexes can be designed. The systems are of course not restricted to the 4-pyridyl group as a substituent of the porphyrin, which can then act as a ligand to another porphyrin. Depending on the nature of this substituent-as-ligand, additional geometrical variety can also be incorporated into the design, e.g., by using
Fig. 27 Schematic representation of the porphyrin-as-ligand complexation scheme
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
29
3-pyridyl, imidazolyl or amine substituents. Insertion of zinc into the thus modified porphyrin leads to self-assembly which shall not be covered here, but only the binding to other porphyrins leading to hetero-porphyrin arrays is discussed in this section. The Sect. 4.1 focuses on the use of pyridylporphyrins in supramolecular porphyrin chemistry, and in Sect. 4.2 more possibilities of porphyrin-as-ligand structures are presented, together with a selection of multiporphyrin assemblies constructed thereof. 4.1 Pyridyl-Porphyrin Binding to Zinc, Ruthenium and Rhodium Porphyrins Many groups have studied the binding of pyridyl porphyrins to both zinc(II) porphyrins [36, 82, 83], ruthenium(II) porphyrins [29, 82, 84–88] or osmium porphyrins [89]. Recent reviews also give a good overview on the field [90, 91]. Since the Zn-porphyrin-pyridine complex is more labile than the Ruporphyrin-pyridine complex, the latter interaction is much more interesting in supramolecular chemistry, because it readily leads to stable arrays in solution. Perpendicularly arranged porphyrin dimers [83, 84, 87], trimers [87], tetramers [84, 86] and pentamers [29, 83, 88, 92] are available by using the pyridyl-porphyrin-as-ligand approach. The pyridyl-porphyrin can be either free base or zinc metallated, and the coordination to either RuTPP or RuOEP leads to stable arrays such as 34 (Fig. 28). The use of a 3 -pyridyl group instead of the 4’-pyridyl group is less often used, but yields stable arrays as well and displaying a slightly different geometry [83, 85, 92]. The larger Ka value for the ruthenium(II)-pyridine complex compared to the zinc-pyridine complex allows the selective binding of a zinc metallated pyridyl-porphyrin to a ruthenium(II) porphyrin by shifting the equilibrium towards the desired array. Due to synthetic procedures of ruthenium-metallation of porphyrins, these are usually isolated in the form of the carbonyl (CO) complex. The additional ligand also helps to stabilize the +2 oxidation state of the ruthenium. The CO ligand can be displaced photolytically, which leaves the ruthenium
Fig. 28 Schematic representation of a TPyP-(RuP)4 complex
30
I. Bouamaied et al.
porphyrin coordinatively unsaturated, and an additional sixth axial ligand can thus be bound. In this way, perpendicular arranged trimers were synthesized [87]. Usually, a combined covalent-coordination approach is applied to improve the stability of the arrays, because in this way receptors employing a chelate effect can be created. A di-zinc(II) bis-porphyrin receptor that binds a free-base 5,10-dipyridyl-porphyrin (cis-DPyP) was reported by Alessio and Sauvage [93]. Two zinc(II)-porphyrins are connected by a phenanthroline linker in an oblique fashion and can act as a bifunctional receptor. 1 H NMR spectroscopy evidenced quantitative formation of the tri-porphyrin macrocyclic assembly 35 (Fig. 29b), and the two-point binding together with an almost perfect geometrical match leads to a remarkably high binding constant of 6 × 108 M–1 which was determined using UV-vis absorption and emission titration experiments. This tri-porphyrin complex compares very well will the systems 20 and 21 (Fig. 17) and provides an illustrative example how using a structurally similar receptor can bind either functionalized ditopic ligands or di-pyridyl-porphyrins. By using a ruthenium-complex connected bis-porphyrin metallacycle and a trans-DPyP as a bridge, relatively stable mixed zinc(II)-free base porphyrin sandwiches were obtained in solution by Alessio (36, Fig. 29c, see also the chapter by E. Alessio in this volume for a modified version of the X-ray structure) [36, 82]. As judged by the relative broadness of the 1 H NMR resonances of the porphyrin-ligand, the complex is in slow to moderate exchange with the dissociated species, hence leading to an equilibrium mixture at room temperature. At temperatures below – 20 ◦ C, the complexes become more robust and could be characterized using 2D NMR techniques. The discrete supra-
Fig. 29 a Di-zinc(II) bis-porphyrin receptor complex with 5,10-dipyridyl porphyrin (substituents on the phenyl groups omitted) [93]; b X-ray structure of the sandwich-complex 36; reprinted with permission from [36], Copyright (2002) American Chemical Society
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
31
molecular box-like structure was confirmed both in solution- and in the solid-state. This again demonstrates that a chelate-effect helps to stabilize the supramolecular structure, though the individual binding event would not be favorable to form the array (see Sect. 2.1). The meso-position of the porphyrins offers an anchor point for connecting porphyrins covalently together. By using alkynyl linkers instead of orthogonally oriented aryl linkers (Fig. 30a), the aromatic cores are in direct π-conjugation which opens the way to new π-extended systems that are potentially interesting for the design of materials with enhanced non-linear optical behavior. The conjugated porphyrins explore a range of torsion-angles in solution [94]. However, formation of the triple strand porphyrin array 37 leads to a stable system in which the porphyrins are held in a co-planar arrangement. This is shown by UV-vis spectroscopy because the spectrum of the array cannot be reproduced by a simple sum of the spectra of the individual diporphyrins. Increased splitting of the Soret band and increased red-shift of the Qx bands indicate increased conjugation between the porphyrin units due to a conformationally restricted coplanarity. Extended versions of ditopic pyridyl porphyrins [95], tritopic and hexatopic pyridyl-triporphyrins [96] bind into the cavities of hexameric porphyrin wheels. Very high binding constants in the range of 109 to 1010 M–1 , e.g., in 38 (Fig. 30b) are reported for these encapsulated host-guest systems by Gossauer as is expected for an additive effect of the individual zinc-pyridine
Fig. 30 a Triple strand porphyrin array by Anderson [94]. The bottom picture shows the electronic absorption spectra of (a) the free base dimer, (b) the Zn-dimer-(pyridine)2 complex and (c) 37 in DCM. Spectrum (b) was recorded in the presence of excess pyridine and is scaled ×2 to facilitate comparison with (c). Arrows highlight regions of increased or decreased absorption in the complex 37. Reproduced by permission of The Royal Society of Chemistry. b Triporphyrin bound to a hexameric wheel 38 [96]
32
I. Bouamaied et al.
binding event. The hexameric wheels can act as an antenna complex, and about 43% energy transfer from the zinc(II)-wheel to the free base-host was observed using fluorescence spectroscopy. 4.2 Mixed Multi-Porphyrin Assemblies Functionalization of C60 with a porphyrin on one side and with a pyridyl ligand on the other side leads to the supramolecular triad 39 formed by a covalent-coordination approach (Fig. 31a) [18, 20, 24]. The photophysical properties of this system were studied in detail and show that energy-transfer occurs from the singlet excited state of the free-base porphyrin to the zinc(II) porphyrin. A charge separated state with formation of the C60 radical anion and the zinc(II) porphyrin radical cation is produced at an increased efficiency upon complex formation of the array. The stability of assembled diporphyrin arrays can substantially be improved by combining the axial coordination with hydrogen bonding as shown in the dyads 40 by Weiss (Fig. 31b) [97–99]. The N-unsubstituted imidazole group, which is attached to a free-base porphyrin via an alkynyl linker, has a high affinity to phenanthroline-strapped zinc(II)-porphyrins through a unique two-point binding mode. The association constant was determined to be ∼ 106 M–1 , which is unprecedented for a zinc-imidazole complex. The dyad shows photoinduced energy transfer from the zinc(II) porphyrin to the free-base porphyrin, because the emission of the free-base at 717 nm was increased by 38% in the array compared to that of the free-base porphyrin alone [97]. If an additional boron dipyrrin moiety is covalently attached to the zinc(II) porphyrin, then a photonic wire can be created which showed highly efficient stepwise energy transfer from the BDP input to the free-base output unit [99]. From a structural point of view, the phenanthroline strapped porphyrins display induced fit related distortions to accommodate the rather bulky substrates on the hindered face of the zinc(II) porphyrin [98].
Fig. 31 Structures of the a a Zn-porphyrin-C60 -free base-porphyrin triad (39) [18, 20, 24], and b a BDP-Zn-porphyrin-free base porphyrin triad (40) [99]
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
33
A combined covalent-coordination approach was used to realize multiporphyrin arrays by Sanders et al. The approach employs either oligomeric porphyrins as donor ligands that coordinate to a central ruthenium(II) porphyrin, or the oligomeric porphyrin unit contains a ruthenium(II) porphyrin which can be assembled around a pyridyl porphyrin. In a first study, ruthenium(II) porphyrin centered porphyrin pentamers were used as building blocks for the construction of large porphyrin arrays, where nickel(II) and zinc(II) porphyrins as peripheral sites were attached to a central ruthenium(II) porphyrin [29, 88]. Both flexible (41) and rigid (42) pentamers were synthesized (Fig. 32a). The penta-porphyrin unit 41 was tried to be assembled around tetra-pyridyl porphyrin which would lead to a 21-porphyrin array in one step. However, as judged from the 1 H NMR spectrum the steric congestion associated with coordination of the rather flexible pentamer seems so great that the product distribution is dominated by the coordination of only three such pentamers about the tetrapyridyl porphyrin. Using an extended version of the tetrapyridyl porphyrin showed that now four molecules of pentaporphyrin can be assembled. Also, titration of either 0.5 equivalents of a 5,15-dipyridine porphyrin to 41 and 42, or 0.25 equivalents of tetrapyridyl a)
b) R
R
R
N
N But
R N
M N
O
N
O R
R
R
O
O
O
N
t Bu
N
N
R N
N M
O R
R
N
N
O
R
R
R
R
N
N
O
M N
O
O
N
O
2H
2H
N
t Bu
M
N
R
t Bu
But
O R
R But
t Bu
t Bu
Ru
N
R
R
R
N
2H
41 M = Ni, Zn But
N
t Bu
Ru
N N N Zn N
R
R
2H
R
R
R
O
R
R N
O
N
R
O O
N N Zn N N
R
R
R
R
But t Bu
42
t Bu
t Bu
R
R
R
N Zn N
t Bu
N
N N
2H
2H
But
Ru N
R
t Bu
N t Bu
R
R N N Zn N N
R
But
But
t Bu
But
R
43
t Bu
But
Fig. 32 Covalently linked tri- and pentatporphyrins as building blocks for large multiporphyrin arrays by Sanders et al. [29, 30, 88]
34
I. Bouamaied et al.
porphyrin to 42, results in the formation of the corresponding 11-porphyrin and 21-porphyrin arrays, respectively. Characteristic changes in the chemical shifts of the inner NH protons as well as of the ortho-protons of the pyridyl groups and of the meso-protons of the porphyrin ligand clearly indicate the formation of the large multiple porphyrin complexes. By exploiting the inherent flexibility of the molecules prepared and the binding affinity of pyridyl ligands to three metals in the order Ru > Zn > Ni, the potential of these porphyrin arrays as large building blocks was demonstrated. The steric effect due to the porphyrin arrays surrounding the core or bridging ligands is the crucial factor controlling the formation of the large porphyrin complexes. By careful design of building blocks and choosing suitable bridging ligands, porphyrin arrays can be fabricated with the size increasing dramatically with very few steps. The sterical factor is very nicely demonstrated in this example, because the flexible building block 41 does not readily lead to the formation of the 21-mer array with tetrapyridyl porphyrin, whereas the more rigid structure of the pentamer 42 favors the formation of the saturated 21-mer array. Photolytic removal and substitution of the CO by additional ligands expands the scope of this approach even further. In another system, the approach is slightly changed in a way that the porphyrin oligomers (dendrons) contain pyridine units for coordination to ruthenium(II) porphyrin monomer building blocks [30]. The dendrons readily form a mono-coordinated complex with the ruthenium(II) porphyrin. After irradiation, the CO ligand is lost, and a second unit can bind to form the dendritic multiporphyrin array 43 (Fig. 32b), where a total of seven porphyrins are assembled. Large heterometallic porphyrin arrays were also successfully assembled using a combination of Sn(IV) and Rh(III) porphyrin coordination chemistry [33]. This is a further extension of the concept to combine porphyrins via axial coordination and makes use of orthogonal binding modes. Here, a Sn(IV) porphyrin acted as a core around which were coordinated two carboxylic acid functionalized porphyrins or porphyrin trimer dendrons. Rh(III) porphyrins were coordinated to pyridyl groups at the periphery of these entities. In this way an eleven porphyrin array, with four different porphyrin metallation states, was assembled. The diamagnetic nature of both the Rh(III) and Sn(IV) porphyrins, the slow ligand exchange kinetics on the NMR time scale, and tight ligand binding permitted the porphyrin arrays to be analyzed by two-dimensional 1 H NMR techniques. The integrity of the eleven-porphyrin assembly 44 (Fig. 33) could thus be proven. This example elegantly demonstrates that the composition of such multi-porphyrin arrays can be tuned by choosing the appropriate combination of substituted porphyrins-as-ligands together with the metalloporphyrins that act as acceptor porphyrins. By employing the mutually non-interfering coordination properties of the zinc(II), ruthenium(II), and tin(IV) centers, further multi-metal arrays
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
35
Fig. 33 Heterometallic undecamer by Sanders using orthogonal binding modes [33]
are accessible [100, 101]. Zn(II) porphyrins prefer nitrogen donor ligands, adopt 5-coordinate square pyramidal geometry, and are kinetically very labile, while Ru(CO) porphyrins form stable and inert complexes with nitrogen donor ligands and adopt 6-coordinate geometry. Sn(IV) porphyrins, on the other hand, prefer oxygen donor ligands, adopt 6-coordinate octahedral geometry, and exchange carboxylate ligands rather slowly. When complementary binding functions are introduced into appropriately designed Zn(II), Ru(II), and Sn(IV) porphyrins, it is possible to assemble multiple components in a controlled manner (Fig. 34a). A dynamic heterometallic oligoporphyrin assembly 45 was thus designed and synthesized, based on the covalent attachment of carboxylate and pyridyl ligands to the porphyrins. The porphyrinligands not only had orthogonal binding modes incorporated, but the geometry also matched well with the anticipated formation of the supramolecular structure. The carboxylate bound tin porphyrin was shown to spin rapidly around the axle because the binding to the zinc(II) center is kinetically labile. Upon addition of the ruthenium(II) porphyrin, the system becomes inert, and the different pyridyl groups binding to two different metal centers can now be distinguished using 1 H NMR spectroscopy. In another system, the bridging ligand isonicotinic acid was used as a bifunctional ligand. The Ru-Sn-Ru trimer 46 (Fig. 34b) was thus constructed by both step-wise and one-pot methods. The UV-vis spectra provide evidence for the absence of any A apart. electronic interaction between the porphyrin planes, which are > 4 ˚
36
I. Bouamaied et al.
Fig. 34 Assembly of a dynamic heterometallic oligoporphyrin using cooperative coordination modes by Sanders. a Zn-Sn-Ru trimer 45 [100] and b Ru-Sn-Ru trimer 46 [101]; c the X-ray structure is of the rhodium(III) analogue of 46; reprinted with permission from [33], Copyright (2001) American Chemical Society
However, the luminescence measurements show quenching of fluorescence in the trimer, which is attributed to a photo-induced electron transfer from the axial ruthenium(II) porphyrin to the excited state of the basal tin porphyrin. Not only ruthenium(II) porphyrins, but also rhodium(III) porphyrins can easily be incorporated into the arrays with the same strategy [33]. Again, the isonicotinic acid is first reacted with the bis-hydroxy tin porphyrin to give the bis-isonicotinic acid complex. Addition of two equivalents of rhodium(III) porphyrin readily yields the trimeric array of the composition Rh-Sn-Rh. The X-ray structure of this complex, which is shown in Fig. 34c, shows that the ligands on the tin center (carboxylates) are in an off-direction which is close to orthogonal to the porphyrin plane, and the three porphyrins adopt a near coplanar arrangement. The tin porphyrin is tilted by about 8.6◦ with respect to the rhodium(III) porphyrins. Phosphorus(V) in its most stable oxidation state forms robust chelate complexes with the porphyrin, and retains the capability of binding two oxygen ligands in a trans-axial mode as outlined in Sect. 3. This strategy was used by Maiya [81, 102–104] and Tanaka and Segawa [105]. The coordination chemistry of phosphorous(V) porphyrins is similar to germanium(IV), tin(IV), gallium(III), and aluminium(III) porphyrin coordination chemistry [102, 106]. These metalloporphyrins were shown to form multiporphyrin complexes with phenol-substituted porphyrins. Because of the sp3 -configuration of the oxo-ligand, the arrays formed show a bent centreto-edge type geometry as shown in structure 47 (Fig. 35). Simply by heating the appropriate mixtures of porphyrins, a series of triporphyrin arrays were obtained incorporating either P(V), Sn(IV), or Ge(IV) in the central porphyrin, and either VO(IV), Co(II), Ni(II), Cu(II), or Zn(II) in the peripheral porphyrins. A detailed electrochemical and photophysical analysis of these arrays was undertaken and should allow the fine-tuning of the electronic in-
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
37
Fig. 35 General structure of the phosphorus-porphyrins with axially bound phenolateporphyrins by Maiya [81, 102–104] and Tanaka and Segawa [105]. M = VO(IV), Co(II), Ni(II), Cu(II), Zn(II), or P(V)
teractions in such multiporphyrin arrays based on these results [102]. It is also possible to first coordinate free-base phenol porphyrins to the central phosphorus porphyrin, and then to metallate the outer porphyrins, for example again with phosphorus. This repetitive cycle of axial ligand binding and subsequent metallation is a simple and versatile method for the construction of supramolecular systems comprised of phosphorus porphyrins. The phosphorus atom itself can also be used as a ligand to metalloporphyrins, especially to ruthenium(II) and rhodium(III) porphyrins [107–112]. Based on a detailed study of different phosphine ligands [108–110, 112], alkynyl-diphenyl phosphine porphyrins were chosen as most appropriate porphyrin-as-ligand building blocks. The phosphine can be attached to the porphyrin on either para- or meta-position of the meso-phenyl group, and both mono- and bis-phosphines can be synthesized. With appropriate amounts of the building blocks mixed, linear dimeric and trimeric arrays have been synthesized and analyzed by 1 H NMR and 31 P NMR spectroscopy. The complex 48 in Fig. 36 shows a representative example. The Ru(II)/Rh(III) acceptor porphyrins can be located either at the periphery or in the center of the array. Likewise, the monophosphine porphyrins can be positioned at the periphery, thus allowing a high degree of freedom in the overall composition of the arrays. This way, both donor and acceptor porphyrins can act as chain extenders or terminators. One of the trimeric complexes with two
Fig. 36 Phosphine-substituted porphyrins as building blocks for the construction of heterometallic porphyrin arrays. 48 displays the Rh-Zn-Rh trimer, the X-ray structure is representing a Ni-Ru-Ni trimer (sidechains on the Ni-porphyrins are omitted for clarity); reprinted with permission from [111], Copyright (2003) American Chemical Society
38
I. Bouamaied et al.
nickel(II) and one ruthenium(II) porphyrin has also been analyzed by X-ray crystallography (Fig. 36). Unlike with pyridyl ligands, no tilting of the ligand was observed, and the metal-phosphorus bonding is near to perpendicular to the porphyrin plane. Attempts have also been made to synthesize higherorder arrays by mixing appropriate amounts of the porphyrins; however, from the NMR data it cannot be concluded if monodisperse five, seven, or nine porphyrin arrays are present or if the solutions are composed of a statistical mixture of smaller and larger arrays. The kinetic lability of the complexes is in this case working against the rational design of long oligomers of well-defined structure and composition.
5 Dynamic Combinatorial Libraries Despite the fact that the concept of dynamic combinatorial chemistry has gained widespread attraction since its first report [113], the concept is hardly used in porphyrin chemistry. Figure 37 displays a schematic representation of the concept. So far only two systems have been realized and studied in more detail [114–117]. The key point is to have a system where the axial coordination is on one side kinetically labile to allow interconversion of the members of the library, and on the other side thermodynamically stable enough so that the complexes remain intact to a sufficiently high quantity. The templating effect can be achieved for example by π – π-interactions or metal-to-ligand coordination. The latter has lead to the selection and virtually complete amplification of cyclic heterometallic porphyrin tetramers, which are stabilized through orthogonal binding modes [114, 115]. The ruthenium and rhodium phosphorus bonding was used as the kinetically labile and thus reversible connection between the porphyrin units to create the library.
Fig. 37 Schematic representation of the formation of a dynamic combinatorial library in thermodynamic equilibrium, and the selection and amplification of a target complex through templating. Arrowheads represent the ligands (P, N) to the central metal of the porphyrins (bars)
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
39
When mixing stoichiometric amounts of a bis-phosphine porphyrin with either a ruthenium(II) or a rhodium(III) porphyrin, a complex mixture of multiporphyrin arrays are obtained which differ in composition and conformation. This can be seen by the complexity of the 1 H and 31 P NMR spectra. Addition of 0.5 equivalents of 4,4 -bipy per zinc(II) porphyrin as ditopic template to the solution resulted in amplification of the cyclic porphyrin tetramer up to virtually 100%. The amplification process could be monitored in situ using 1 H NMR spectroscopy upon titrating the guest into the host solution. Both the proton and the phosphorus NMR spectra showed only one significant species being present. All proton resonances were sharp and well resolved. Even though the building blocks contain a built-in predisposition to form cyclic arrays, this specific tetramer is not accessible by rational synthesis because of the lability of the Rh – P bonding and the cis/trans isomerism of the zinc(II)-phosphine porphyrin. In a further study, the structural diversity in cyclic tetraporphyrin cages was enlarged [114]. Both rhodium(III) and ruthenium(II) porphyrins have been incorporated, and the homoporphyrinic cages (with respect to the Ru(II) or Rh(III) porphyrin) can be selected and amplified from dynamic combinatorial libraries by using various ligands as templates. The cage 49 (Fig. 38) displays an example of an amplified host-guest complex. In general, the rhodium(III) cages are easier to handle, because of the inherent problems of removing the carbonyl ligand from ruthenium(II) porphyrins. In addition, the 1 H NMR spectra obtained with rhodium(III) porphyrins are usually better resolved than those of their ruthenium counterparts. The absolute amount of amplification from the DCLs is strongly dependent on the combination of the Ru(II)/Rh(III) porphyrin and the template: the more sterically demanding the porphyrin, the smaller the template should be to obtain virtually complete amplification. In the case of bulky tertiary butyl groups on the meso phenyl substituents, cages are not formed at all. The largest template diaza-pyrene forms cages quantitatively only with OEP, the least sterically demanding porphyrin. The X-ray crystal structure of one specific
Fig. 38 Structural drawing and X-ray structure of the amplified cyclic (4,4 -bipy)⊂[(Zn)2 / (Rh)2 ] host-guest complex 49 [114]
40
I. Bouamaied et al.
cage demonstrates that the cages can adopt severely distorted conformations to accommodate the relatively short template bipy. The solid-state geometry deviates considerably from the conceptual box-structure. An extension to mixed DCLs showed that only limited selectivity is displayed by the various templates. Formation of mixed cages that contain two different rhodium(III) porphyrins prevents effective selection, although the kinetic lability of the systems allows for some amplification. This lability, however, also prevents isolation of the cages. Effective amplification can only be achieved if the system is biased in such a way that some of the porphyrins used are inherently unable to form cages, i.e., having bulky substituents.
6 Porphyrin-Fullerene Complexes In Sects. 2.1 and 2.2, the reasons for constructing supramolecular assemblies with porphyrins and fullerenes were already outlined. The C60 unit can be bound to the porphyrin via axial coordination chemistry because the fullerene can be substituted with nitrogen ligands. However, also unmodified C60 and C70 readily forms assemblies with porphyrins through π – πstacking. Even though this is not axial coordination in its proper sense, these interactions were studied in detail and used for the construction of supramolecular porphyrin-fullerene complexes. The potential of the complexes in the areas of porous framework solids and photovoltaic devices makes it worthwhile to give an overview of the current status of the systems under investigation. Rather unusual is the short distance between the porphyrin core and A compared to other π-stacking distances [118]. The the C60 of ∼ 2.75 ˚ porphyrin-fullerene association is attractive and structure-defining. Almost any free-base or metallo-porphyrin can be co-crystallized with fullerenes from the corresponding mixtures, and in the solid state a zigzag motif seems to dominate [119, 120]. In the absence of steric effects, the hierarchy of interactions in the solid-state is porphyrin-porphyrin > porphyrin-fullerene > fullerene-fullerene; the metallation state of the porphyrin also may have an influence on these secondary interactions. This is seen in the crystal structures of either C60 and free-base TPP or C60 and CoOEP as representative examples (Fig. 39a) [118]. The π – π-interaction can in part be optimized by a deviation from planarity of the porphyrins, e.g., ruffling or doming. The porphyrin-fullerene attraction is not merely a feature of the solid-state, but it is also present in solution. Both 1 H and 13 C NMR spectroscopy, as well as UV-vis spectroscopy show indication of the attractive interactions. Cyclic diporphyrins with the appropriate spacer form highly stable inclusion complexes with C60 [121]. Using this interaction in solution allowed the selection
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
41
Fig. 39 a Zigzag structures of a C60 -H2 TPP (top) and a C60 -CoOEP (bottom) complex in the crystal; reprinted with permission from [118], Copyright (2004) American Chemical Society; b the X-ray structure of a C60 -bis-porphyrin inclusion complex which was obtained from a dynamic combinatorial library [117]; reproduced by permission of The Royal Society of Chemistry
and amplification of a bis-porphyrin host for the C60 guest from dynamic combinatorial libraries (Fig. 39b) [117]. Composite molecular nano-clusters of fullerene and porphyrins can be prepared in solution, usually using toluene or toluene-acetonitrile mixtures, and then be deposited on surfaces [122, 123]. The mixtures undergo closepacked stacking to produce various shaped micro-cocrystallites. The composite cluster films exhibit incident photon-to-photocurrent efficiencies of up to 17% which are significantly higher than those observed from individual porphyrin clusters (1.6%) or fullerene clusters (5%), which demonstrates the synergy of these systems. Multiporphyrin arrays, which are based on an oligopeptide [124] or a gold nanocluster [125] scaffold, can form clusters with C60 , which show enhanced activity in photoelectronic devices (Fig. 40). The porphyrin-C60 complex forms linear or cyclic arrays with an alternating porphyrin and fullerene composition. These arrays are preformed in solution and subsequently deposited on SnO2 -surfaces. The systems showed efficient photoresponse from the UV to the near-IR region and a high light-energy conversion efficiency. Both porphyrins and fullerenes can be deposited onto surfaces under ultra-high vacuum (UHV) conditions as studied by Diederich et al. [126, 127]. In a first step, a monolayer of porphyrins or extended porphyrin analogues can be formed which can be analyzed using AFM or STM imaging techniques. UHV deposition of a second layer of C60 on top of a full monolayer of porphyrins leads to the formation of supramolecular assemblies, and the precise structure of the arrangement of the C60 on the patterned layer can be controlled by the porphyrin structure. The observed modes of self-assembly result from a delicate balance between the fullereneporphyrin interaction and the conformational motion within the porphyrin monolayer.
42
I. Bouamaied et al.
Fig. 40 Porphyrin-fullerene assemblies in the fabrication of photovoltaic cells for enhanced performance: a a peptide-porphyrin-C60 cluster; reprinted with permission from [124], Copyright (2005) American Chemical Society; b a gold nanocluster approach; reprinted with permission from Wiley from [125]
7 Conclusions and Outlook The methodologies to design and synthesize multiporphyrin assemblies on a rational basis have advanced to the point, where almost any metalloporphyrin can be incorporated into oligomeric heterometallic arrays. The use of
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
43
the axial coordination is an alternative route to the direct covalent connection of porphyrins and allows using smaller building blocks. This greatly reduces the synthetic and purificational efforts that have to be made en route to the final construct. With this strategy a great number of multiporphyrin assemblies can be realized, as was outlined in this chapter. The complexes range from axially coordinated functional molecules to one porphyrin over multiporphyrin complexes with up to 21 porphyrins to multicomplexes that can be adsorbed on surfaces. It should be kept in mind that the axial coordination, also using the porphyrin-as-ligand approach, has its limitations in terms of metals and ligands used. The limitations arise mainly from the synthetic accessibility of both the metalloporphyrin and the substitutions that can be attached to the porphyrins or other functional groups. Analysis using most convenient techniques such as NMR spectroscopy, UV-vis spectroscopy, mass spectrometry and X-ray crystallography may also be restrictive in terms of size and specific composition of the analyte. However, in this chapter the recent advancements are summarized, and since this is by far not done on an exhaustive basis the amount of highly elaborate multiporphyrin assemblies that are presented demonstrate that this is still a highly active field of research. Not only pure coordination chemistry is applied, but a combination of covalent-coordination chemistry together with orthogonal binding modes opens the way to almost any desired composite. Photophysically highly active complexes for efficient photoinduced energy or electron transfer, enantioselective catalysts or new materials for structure determination and sensing were developed. New strategies such as dynamic combinatorial chemistry are still in its infancy with respect to porphyrin chemistry, but the concept may prove very useful in future for the selection and amplification of hitherto inaccessible multiporphyrin assemblies. A recently emerging trend concerns the absorbance of porphyrins and conjugates thereof with various additives on surfaces. This is a very promising concept for the production of a new generation in photovoltaic devices. Of great potential is the combination of porphyrins with fullerenes, as was demonstrated in representative examples throughout the entire chapter. It is very probable that we shall see the emergence of new materials based on the porphyrin-C60 combination in near future. Despite the fact that the first porphyrin synthesis was performed more than a century ago by Fischer, the continuous use of metalloporphyrins in all fields of research, from organic synthesis over materials science to biology and medicinal chemistry, proves that these molecules have by far not yet been fully explored, and new applications will certainly not cease to appear for a while.
44
I. Bouamaied et al.
References 1. Sanders JKM (2000) In: Kadish KM, Smith KM, Guilard R (eds) The Porphyrin Handbook. Academic, New York 2. Sanders JKM, Bampos N, Clyde-Watson Z, Darling SL, Hawley JC, Kim H-J, Mack CC, Webb SJ (2000) In: Kadish KM, Smith KM, Guilard R (eds) The Porphyrin Handbook. Academic, New York 3. Kalyanasundaram K (1992) Photochemistry of Polypyridine and Porphyrin Complexes. Academic, London 4. Stulz E, Mak CC, Sanders JKM (2001) J Chem Soc Dalton 5:604 5. Otsuki J, Harada K, Toyama K, Hirose Y, Araki K, Seno M, Takatera K, Watanabe T (1998) Chem Commun 15:1515 6. Otsuki J, Takatsuki M, Kaneko M, Miwa H, Takido T, Seno M, Okamoto K, Imahori H, Fujitsuka M, Araki Y, Ito O, Fukuzumi S (2003) J Phys Chem A 107:379 7. Kashiwagi Y, Imahori H, Araki Y, Ito O, Yamada K, Sakata Y, Fukuzumi S (2003) J Phys Chem A 107:5515 8. Norsten TB, Branda NR (2001) Adv Mater 13:347 9. Samachetty HD, Branda NR (2005) Chem Commun 22:2840 10. D’Souza F, Smith PM, Gadde S, McCarty AL, Kullman MJ, Zandler ME, Itou M, Araki Y, Ito O (2004) J Phys Chem B 108:11333 11. D’Souza F, El-Khouly ME, Gadde S, McCarty AL, Karr PA, Zandler ME, Araki Y, Ito O (2005) J Phys Chem B 109:10107 12. D’Souza F, Deviprasad GR, Zandler ME, El-Khouly ME, Fujitsuka M, Ito O (2003) J Phys Chem A 107:4801 13. D’Souza F, Chitta R, Gadde S, Zandler ME, McCarty AL, Sandanayaka ASD, Araki Y, Ito O (2005) Chem-Eur J 11:4416 14. D’Souza F, Chitta R, Gadde S, Zandler ME, Sandanayaka ASD, Araki Y, Ito O (2005) Chem Commun 10:1279 15. D’Souza F, Gadde S, Zandler ME, Itou M, Araki Y, Ito O (2004) Chem Commun 20:2276 16. D’Souza F, Smith PM, Zandler ME, McCarty AL, Itou M, Araki Y, Ito O (2004) J Am Chem Soc 126:7898 17. D’Souza F, Deviprasad GR, Zandler ME, Hoang VT, Klykov A, VanStipdonk M, Perera A, El-Khouly ME, Fujitsuka M, Ito O (2002) J Phys Chem A 106:3243 18. D’Souza F, Deviprasad GR, Zandler ME, El-Khouly ME, Fujitsuka M, Ito O (2002) J Phys Chem B 106:4952 19. D’Souza F, Gadde S, Zandler ME, Arkady K, El-Khouly ME, Fujitsuka M, Ito O (2002) J Phys Chem A 106:12393 20. D’Souza F, Deviprasad GR, El-Khouly ME, Fujitsuka M, Ito O (2001) J Am Chem Soc 123:5277 21. D’Souza F, Zandler ME, Smith PM, Deviprasad GR, Arkady K, Fujitsuka M, Ito O (2002) J Phys Chem A 106:649 22. Zandler ME, Smith PM, Fujitsuka M, Ito O, D’Souza F (2002) J Org Chem 67:9122 23. El-Khouly ME, Ito O, Smith PM, D’Souza F (2004) J Photochem Photobiol C 5:79 24. D’Souza F, Ito O (2005) Coord Chem Rev 249:1410 25. Da Ros T, Prato M, Carano M, Ceroni P, Paolucci F, Roffia S, Valli L, Guldi DM (2000) J Organomet Chem 599:62 26. Hauke F, Swartz A, Guldi DM, Hirsch A (2002) J Mater Chem 12:2088 27. Taylor PN, Anderson HL (1999) J Am Chem Soc 121:11538 28. Hunter CA, Tregonning R (2002) Tetrahedron 58:691 29. Mak CC, Bampos N, Sanders JKM (1999) Chem Commun 12:1085
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
45
30. Darling SL, Mak CC, Bampos N, Feeder N, Teat SJ, Sanders JKM (1999) New J Chem 23:359 31. Kim HJ, Redman JE, Nakash M, Feeder N, Teat SJ, Sanders JKM (1999) Inorg Chem 38:5178 32. Redman JE, Feeder N, Teat SJ, Sanders JKM (2001) Inorg Chem 40:3217 33. Redman JE, Feeder N, Teat SJ, Sanders JKM (2001) Inorg Chem 40:2486 34. Ballester P, Costa A, Castilla AM, Deya PM, Frontera A, Gomila RM, Hunter CA (2005) Chem Eur J 11:2196 35. Splan KE, Stern CL, Hupp JT (2004) Inorg Chim Acta 357:4005 36. Iengo E, Zangrando E, Minatel R, Alessio E (2002) J Am Chem Soc 124:1003 37. Ballester P, Costa A, Deya PM, Frontera A, Gomila RM, Oliva AI, Sanders JKM, Hunter CA (2005) J Org Chem 70:6616 38. Campbell K, McDonald R, Branda NR, Tykwinski RR (2001) Org Lett 3:1045 39. Campbell K, McDonald R, Tykwinski RR (2002) J Org Chem 67:1133 40. Guldi DM, Luo CP, Swartz A, Scheloske M, Hirsch A (2001) Chem Commun 12:1066 41. Guldi DM, Luo CP, Da Ros T, Prato M, Dietel E, Hirsch A (2000) Chem Commun 5:375 42. Imahori H, Yamada K, Yoshizawa E, Hagiwara K, Okada T, Sakata Y (1997) J Porphyr Phthalocyanines 1:55 43. Sakata Y, Imahori H, Tsue H, Higashida S, Akiyama T, Yoshizawa E, Aoki M, Yamada K, Hagiwara K, Taniguchi S, Okada T (1997) Pure Appl Chem 69:1951 44. Kleij AW, Kuil M, Tooke DM, Spek AL, Reek JNH (2005) Inorg Chem 44:7696 45. Reek JNH, Roder M, Goudriaan PE, Kamer PCJ, van Leeuwen PWNM, Slagt VF (2005) J Organomet Chem 690:4505 46. Slagt VF, van Leeuwen P, Reek JNH (2003) Angew Chem Int Ed 42:5619 47. Slagt VF, van Leeuwen P, Reek JNH (2003) Chem Commun 19:2474 48. Slagt VF, Roder M, Kamer PCJ, van Leeuwen P, Reek JNH (2004) J Am Chem Soc 126:4056 49. Baldini L, Ballester P, Casnati A, Gomila RM, Hunter CA, Sansone F, Ungaro R (2003) J Am Chem Soc 125:14181 50. Chichak K, Branda NR (1999) Chem Commun 6:523 51. Iengo E, Zangrando E, Mestroni S, Fronzoni G, Stener M, Alessio E (2001) J Chem Soc Dalton Trans 8:1338 52. Huang XF, Borhan B, Rickman BH, Nakanishi K, Berova N (2000) Chem Eur J 6:216 53. Huang XF, Rickman BH, Borhan B, Berova N, Nakanishi K (1998) J Am Chem Soc 120:6185 54. Huang XF, Fujioka N, Pescitelli G, Koehn FE, Williamson RT, Nakanishi K, Berova N (2002) J Am Chem Soc 124:10320 55. Kurtan T, Nesnas N, Koehn FE, Li YQ, Nakanishi K, Berova N (2001) J Am Chem Soc 123:5974 56. Kurtan T, Nesnas N, Li YQ, Huang XF, Nakanishi K, Berova N (2001) J Am Chem Soc 123:5962 57. Fujimoto Y, Ishihara J, Maki S, Fujioka N, Wang T, Furuta T, Fishkin N, Borhan B, Berova N, Nakanishi K (2001) Chem Eur J 7:4198 58. Huang X, Nakanishi K, Berova N (2000) Chirality 12:237 59. Huang XF, Nakanishi K, Berova N (2000) Chirality 12:237 60. Nehira T, Parish CA, Jockusch S, Turro NJ, Nakanishi K, Berova N (1999) J Am Chem Soc 121:8681 61. Berova N, Borhan B, Dong JG, Guo JS, Huang XF, Karnaukhova E, Kawamura A, Leu JH, Matile S, Nakanishi K, Rickman B, Su J, Tan Q, Zanze I (1998) Pure Appl Chem 70:377
46
I. Bouamaied et al.
62. Elemans J, Lensen MC, Gerritsen JW, van Kempen H, Speller S, Nolte RJM, Rowan AE (2003) Adv Mater 15:2070 63. Marvaud V, VidalFerran A, Webb SJ, Sanders JKM (1997) J Chem Soc Dalton Trans 6:985 64. Vidal-Ferran A, Bampos N, Sanders JKM (1997) Inorg Chem 36:6117 65. Chichak K, Branda NR (2000) Chem Commun 13:1211 66. Shao XB, Jiang XK, Zhu SZ, Li ZT (2004) Tetrahedron 60:9155 67. Shao XB, Jiang XK, Zhao X, Zhao CX, Chen Y, Li ZT (2004) J Org Chem 69:899 68. Ballester P, Gomila RM, Hunter CA, King ASH, Twyman LJ (2003) Chem Commun 1:38 69. Fantuzzi G, Pengo P, Gomila R, Ballester P, Hunter CA, Pasquato L, Scrimin P (2003) Chem Commun 8:1004 70. Felluga F, Tecilla P, Hillier L, Hunter CA, Licini G, Scrimin P (2000) Chem Commun 12:1087 71. Kanayama N, Kanbara T, Kitano H (2000) J Phys Chem B 104:271 72. Offord DA, Sachs SB, Ennis MS, Eberspacher TA, Griffin JH, Chidsey CED, Collman JP (1998) J Am Chem Soc 120:4478 73. Yamamoto G, Nadano R, Satoh W, Yamamoto Y, Akiba K (1997) Chem Commun 14:1325 74. Hamstra BJ, Cheng BS, Ellison MK, Scheidt WR (2000) Inorg Chem 39:1454 75. Hamstra BJ, Cheng BS, Ellison MK, Scheidt WR (1999) Inorg Chem 38:3554 76. Tsao TB, Lee GH, Yeh CY, Peng SM (2003) Dalton Trans :1465 77. Kimura M, Kitamura T, Sano M, Muto T, Hanabusa K, Shirai H, Kobayashi N (2000) New J Chem 24:113 78. Susumu K, Kunimoto K, Segawa H, Shimidzu T (1995) J Photochem Photobiol A 92:39 79. Susumu K, Kunimoto K, Segawa H, Shimidzu T (1995) J Phys Chem 99:29 80. Shimidzu T (1996) Synth Met 81:235 81. Kumar PP, Premaladha G, Maiya BG (2005) Chem Commun 30:3823 82. Prodi A, Chiorboli C, Scandola F, Iengo E, Alessio E, Dobrawa R, Wurthner F (2005) J Am Chem Soc 127:1454 83. Prodi A, Indelli MT, Kleverlaan CJ, Scandola F, Alessio E, Gianferrara T, Marzilli LG (1999) Chem-Eur J 5:2668 84. Alessio E, Ciani E, Iengo E, Kukushkin VY, Marzilli LG (2000) Inorg Chem 39:1434 85. Gupta I, Ravikanth M (2004) J Org Chem 69:6796 86. Gupta I, Agarwal N, Ravikanth M (2004) Eur J Org Chem 69:1693 87. Funatsu K, Kimura A, Imamura T, Ichimura A, Sasaki Y (1997) Inorg Chem 36:1625 88. Mak CC, Bampos N, Darling SL, Montalti M, Prodi L, Sanders JKM (2001) J Org Chem 66:4476 89. Kariya N, Imamura T, Sasaki Y (1997) Inorg Chem 36:833 90. Imamura T, Fukushima K (2000) Coord Chem Rev 198:133 91. Burrell AK, Officer DL, Plieger PG, Reid DCW (2001) Chem Rev 101:2751 92. Guldi DM, Da Ros T, Braiuca P, Prato M, Alessio E (2002) J Mater Chem 12:2001 93. Iengo E, Zangrando E, Alessio E, Chambron J-C, Heitz V, Flamigni L, Sauvage J-P (2003) Chem Eur J 9:5879 94. Wilson GS, Anderson HL (1999) Chem Commun 16:1539 95. Ambroise A, Li JZ, Yu LH, Lindsey JS (2000) Org Lett 2:2563 96. Rucareanu S, Mongin O, Schuwey A, Hoyler N, Gossauer A, Amrein W, Hediger HU (2001) J Org Chem 66:4973 97. Paul D, Wytko JA, Koepf M, Weiss J (2002) Inorg Chem 41:3699
Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies
47
98. Paul D, Melin F, Hirtz C, Wytko J, Ochsenbein P, Bonin M, Schenk K, Maltese P, Weiss J (2003) Inorg Chem 42:3779 99. Koepf M, Trabolsi A, Elhabiri M, Wytko JA, Paul D, Albrecht-Gary AM, Weiss J (2005) Org Lett 7:1279 100. Kim HJ, Bampos N, Sanders JKM (1999) J Am Chem Soc 121:8120 101. Maiya BG, Bampos N, Kumar AA, Feeder N, Sanders JKM (2001) New J Chem 25:797 102. Giribabu L, Rao TA, Maiya BG (1999) Inorg Chem 38:4971 103. Rao TA, Maiya BG (1995) J Chem Soc Chem Commun 9:939 104. Rao TA, Maiya BG (1996) Inorg Chem 35:4829 105. Susumu K, Tanaka K, Shimidzu T, Takeuchi Y, Segawa H (1999) J Chem Soc Perkin Trans 2:1521 106. Kumar PP, Maiya BG (2003) New J Chem 27:619 107. Darling SL, Stulz E, Feeder N, Bampos N, Sanders JKM (2000) New J Chem 24:261 108. Stulz E, Maue M, Feeder N, Teat SJ, Ng YF, Bond AD, Darling S, Sanders JKM (2002) Inorg Chem 41:5255 109. Stulz E, Sanders JKM, Montalti M, Prodi L, Zaccheroni N, de Biani F, Grigiotti E, Zanello P (2002) Inorg Chem 41:5269 110. Stulz E, Scott SM, Bond AD, Otto S, Sanders JKM (2003) Inorg Chem 42:3086 111. Stulz E, Scott SM, Ng YF, Bond AD, Teat SJ, Darling SL, Feeder N, Sanders JKM (2003) Inorg Chem 42:6564 112. Stulz E, Maue M, Scott SM, Mann BE, Sanders JKM (2004) New J Chem 28:1066 113. Brady PA, BonarLaw RP, Rowan SJ, Suckling CJ, Sanders JKM (1996) Chem Commun 3:319 114. Stulz E, Scott SM, Bond AD, Teat SJ, Sanders JKM (2003) Chem-Eur J 9:6039 115. Stulz E, Ng Y-F, Scott SM, Sanders JKM (2002) Chem Commun 5:524 116. Kieran AL, Pascu SI, Jarrosson T, Gunter MJ, Sanders JKM (2005) Chem Commun 14:1842 117. Kieran AL, Pascu SI, Jarrosson T, Sanders JKM (2005) Chem Commun 10:1276 118. Boyd PDW, Reed CA (2005) Acc Chem Res 38:235 119. Zhang J, Cai RF, Chen MQ, Weng LH, Zhou XG (2005) Eur J Inorg Chem :3298 120. Konarev DV, Neretin IS, Saito G, Slovokhotov YL, Otsuka A, Lyubovskaya RN (2004) Eur J Inorg Chem :1794 121. Tashiro K, Aida T, Zheng JY, Kinbara K, Saigo K, Sakamoto S, Yamaguchi K (1999) J Am Chem Soc 121:9477 122. Hasobe T, Imahori H, Fukuzumi S, Kamat PV (2003) J Phys Chem B 107:12105 123. Imahori H, Liu JC, Hotta H, Kira A, Umeyama T, Matano Y, Li GF, Ye S, Isosomppi M, Tkachenko NV, Lemmetyinen H (2005) J Phys Chem B 109:18465 124. Hasobe T, Kamat PV, Troiani V, Solladie N, Ahn TK, Kim SK, Kim D, Kongkanand A, Kuwabata S, Fukuzumi S (2005) J Phys Chem B 109:19 125. Imahori H, Fujimoto A, Kang S, Hotta H, Yoshida K, Umeyama T, Matano Y, Isoda S (2005) Adv Mater 17:1727 126. Bonifazi D, Spillmann H, Kiebele A, de Wild M, Seiler P, Cheng FY, Guntherodt HJ, Jung T, Diederich F (2004) Angew Chem Int Ed 43:4759 127. Armaroli N, Accorsi G, Song FY, Palkar A, Echegoyen L, Bonifazi D, Diederich F (2005) Chem Phys Chem 6:732
Struct Bond (2006) 121: 49–104 DOI 10.1007/430_023 © Springer-Verlag Berlin Heidelberg 2006 Published online: 24 February 2006
Porphyrin Supramolecules by Self-Complementary Coordination Yoshiaki Kobuke Graduate School of Materials Science, Nara Institute of Science and Technology, Takayama 8916-5, 630-0101 Nara, Japan [email protected] 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
2 2.1 2.2
Characteristics of Coordination Organization . . . . . . . . . . . . . . . . Stability Constants for Supramolecular Assemblies . . . . . . . . . . . . . . Chlorophylls in Natural Systems . . . . . . . . . . . . . . . . . . . . . . . .
51 51 53
3 3.1 3.2 3.2.1 3.2.2 3.2.3 3.3 3.4
Self-Complementary Coordination of Porphyrins . . . Dimer . . . . . . . . . . . . . . . . . . . . . . . . . . . Macrocycle and Box . . . . . . . . . . . . . . . . . . . . Antenna Mimics of B850 and Box Types . . . . . . . . . Porphyrin Macrocycle Capable of Guest Incorporation Size-Tunable Giant Macrocyclic Arrays . . . . . . . . . One-Dimensional Arrays . . . . . . . . . . . . . . . . . Two-Dimensional Arrays . . . . . . . . . . . . . . . . .
. . . . . . . .
59 59 64 64 74 75 76 82
4
Coordination Organization of Phthalocyanines . . . . . . . . . . . . . . .
84
5
Structural Characterization of Supramolecular Assemblies . . . . . . . . .
87
6 6.1 6.2 6.3 6.4
Properties of Porphyrin Assemblies . . . . . . . . . . Photoinduced Electron Transfer in Special Pair Mimic Energy Transfer in Photosynthetic Antenna Mimics . Nonlinear Optical Properties . . . . . . . . . . . . . . Photocurrent Generation . . . . . . . . . . . . . . . .
. . . . .
89 89 91 93 96
7
Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
100
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . . . . .
. . . . .
Abstract The supramolecular organization of porphyrins using self-complementary coordination is reviewed. Complementarity affords large stability constants for coordination organization, and as a result various structural variations of porphyrin supramolecules were obtained. Dimeric porphyrins are of special interest as they mimic the special pair of the photosynthetic reaction centers. A very stable slipped cofacial dimer was obtained and the dimer effect on photoinduced charge separation is discussed. Macrocyclic porphyrin supramolecules were prepared as photosynthetic antenna models and rapid energy transfer rates among the components were estimated. Three-dimensional organization represents another type of approach to light-harvesting antenna systems. Oneand two-dimensional arrays of porphyrins have also been prepared using this methodology. Energy and electron transfers along the array provide the basis for photonic and
50
Y. Kobuke
electronic applications. Introduction of molecular polarization in the array leads to the development of nonlinear optics materials. Photocurrent generation is enhanced by array formation on electrode surfaces. Keywords Antenna · Charge separation · Macrocycle · Nonlinear optics · Photosynthesis
1 Introduction Light is the ultimate energy source for all living systems. Photosynthetic systems utilize its energy to excite chlorophyll or bacteriochlorophyll within organisms [1, 2]. The electron in this photoexcited state is then transferred to the primary electron acceptors, initiating the oxidation–reduction reaction known as photosynthesis. After the photosynthetic organisms employed reducing agents of more negative reduction potentials, they found an evolutionary way to use water as a more abundant reducing agent, thus evolving oxygen as a by-product. The subsequent accumulation of oxygen allowed the appearance of animals, which depend totally on the products of photosynthetic plants by consuming their component carbohydrates for food and using the oxygen they produce for respiration. How do natural photosynthetic systems transfer the excitation energy generated by the incident light to the reaction center through multistep energytransfer chains with minimum energy loss? How do these natural systems avoid back-electron transfer, the most favorable side reaction resulting in the loss of the excitation energy as heat, instead converting the photoexcited state into a reduced acceptor and generating a hole in the original framework? How do they subsequently initiate cascade electron transfers and supply electrons to these holes, regenerating the original electronic state of the reaction center? It is intriguing to apply the answers found in natural systems to the problem of designing artificial photosynthetic systems [3–5]. Indeed, work on this secondary problem, which has a range of applications, may open the door to a more genuine understanding of natural photosynthesis. This chapter focuses on supramolecular porphyrin systems that are stable enough to be characterized in solution. Assemblies coordinated only in solids (e.g., crystals) will not be included. In order to maintain these supramolecular structures in solution, strong complementarity and multitopic coordinations are required, rather than a simple monotopic axial coordination. Of these conditions, this discussion will be concerned only with those necessary for the complementary coordinations, and with the presentation of various examples of these coordinations. The issue of multitopic coordination will be dealt with in another forum. Self-complementary coordination can be used to obtain dimers, rings, and one- and two-dimensional arrays. These supramolecular systems should not only influence structure formation, however, but also be closely related to their functions. It is therefore natural to discuss the
Porphyrin Supramolecules by Self-Complementary Coordination
51
photophysical properties associated with these unique supramolecules from this viewpoint [6–9]. This review will focus on Mg and Zn ions as useful candidates from among the other metal ions that might be inserted into the center of porphyrin macrocycles. Photosynthetic systems certainly employ Mg ions. However, extensive studies focusing on the Zn complexes of porphyrins have been undertaken, since Zn(II), having a d10 structure, can be a good substitute for Mg(II). Although the fluorescence quantum yields of Zn(II) porphyrins are inferior to those of Mg(II) porphyrins, the former have obvious advantages such as permitting easy introduction of the metal into the structure, stability in acidic conditions, and durability when exposed to oxidation conditions. In fact, organisms living under conditions of extreme heat and acidity near sea volcanoes have been found to employ Zn substitutes in place of the more common chlorophyll [10]. Most other transition metal ions quench the singlet excitation energy, and these metalloporphyrins may be regarded as functional dyestuffs rather than expressing photosynthetic models. There are many interesting examples of structure formation using these transition metal ions, and they can be important in the transfer of electrons and holes in the system because they are readily oxidized and reduced. With these points in mind, some transition metal porphyrins are included in this review as well.
2 Characteristics of Coordination Organization 2.1 Stability Constants for Supramolecular Assemblies Relatively strong coordination to zinc porphyrins is obtained when neutral nitrogen ligands are employed. The relation between the basicity of various nitrogen ligands (pKa of the conjugated acid) and their stability constants with zinc porphyrins is summarized in Fig. 1 [11]. It is reasonable for these typical, or d10 , metal ions that a correlation should be observed between these two parameters, considering the similarity between the coordination of N to metal and the bonding of N to H+ provided that steric requirements are insignificant. Among the important aromatic ligands, imidazole is preferred to pyridine because of its high basicity: the pKa is more favorable by a value of 2 for imidazole compared with that for pyridine. When steric requirements become significant, the linearity of the relationship between the stability constant and the basicity drops sharply, as is observed for 2-methylpyridine or 2,6-dimethylpyridine. Aliphatic amines follow the order of the stability constant as primary > secondary > tertiary, depending on their steric demands rather than their pKa value. Cyclic aliphatic amines are favored under these
52
Y. Kobuke
Fig. 1 Stability constants of neutral nitrogen ligands with zinc tetraphenylporphyrin (ZnTPP) as a function of pKa of the conjugated acid; • pyridines, × imidazole, aliphatic amines, o-substituted pyridines
conditions, since steric hindrance is properly avoided by ring closing, as is best demonstrated for diazabicyclooctane. Considering the general trend observed, the stability constant of imidazole to zinc porphyrin is expected to be of the order of 104 M–1 at most. Table 1 shows calculations of the stability constants required to ensure that more than 90 or 99% of the initial self-assembling monomer may exist as Table 1 Minimum stability constants for 90 and 99% initial monomers to exist as 2- to 4-mers at initial concentrations of 1.0 and 10 µM
Dimer Trimer Tetramer
at 1.0 × 10–6 M > 90%
> 99%
at 1.0 × 10–5 M > 90%
> 99%
108 M–1 1015 M–2 1022 M–3
1010 M–1 1018 M–2 1026 M–3
107 M–1 1013 M–2 1019 M–3
109 M–1 1016 M–2 1022 M–3
Porphyrin Supramolecules by Self-Complementary Coordination
53
Fig. 2 Schematic illustration of ring systems made of metalloporphyrins with selfcomplementary coordination
dimers, trimers, or tetramers in solution at initial concentrations of 1.0 and 10 µM (typical concentrations for photophysical measurements of porphyrin derivatives). It is clear that a significant enhancement in the stability constant is required to satisfy the necessary conditions. Di- or multitopic interactions are among the most efficient methods satisfying the above conditions, and will be treated in detail by another reviewer. An alternative method exploits the high level of self-complementarity between the monomeric units to induce efficient mutual coordination. As is schematically illustrated for the cases of ring systems in Fig. 2, matching the coordinating angle permits the formation of various macrocycles. Details of the molecular design required to adequately satisfy the self-complementarity condition will be discussed thoroughly with reference to various examples later in the discussion. 2.2 Chlorophylls in Natural Systems In the core of all photosynthetic reaction centers, two (bacterio)chlorophyll molecules are arranged by coordination of their imidazolyl side chains from the transmembrane α helices to the central magnesium(II) ions. This coordination brings the two chlorophylls into a slipped cofacial orientation, where two π-electronic planes are placed in an almost parallel orientation with a slippage of the central Mg centers. The interplanar distances depend on the photosynthetic species, and range from 3.2 to 5.0˚ A [12–15]. These coordination modes are illustrated in Fig. 3. In the bacterial special pair (Fig. 3b), A. two (bacterio)chlorophylls interact strongly at their closest distance of 3.2 ˚ One of the pyrrole rings in each chlorophyll is overlapping, with a Mg · · · Mg A [12]. Here, the close chromophore interaction is believed to distance of 7.4 ˚ drive efficient charge separation, leading to the pheophytin acting as the primary electron acceptor. In photosystem I (Fig. 3c), the interplanar distance A. In this case, two pyrrole rings is longer, with an intermediate value of 3.6 ˚ A [14]. in each chlorophyll overlap, shortening the Mg · · · Mg distance to 6.3 ˚
54
Y. Kobuke
Fig. 3 Coordination modes of the special pair in photosynthetic systems. a and b photosynthetic bacteria, c photosystem I, d photosystem II
Charge separation probably occurs mainly as a result of one of the chlorophyll pair, with some involvement of the accessory chlorophyll. Rather than pheophytin acting as an acceptor, in this case the electron is transferred to the chlorophyll. Since the oxidation/reduction potential of the acceptor chlorophyll is negatively shifted, it plays a role in the subsequent photosynthetic reduction. In the case of photosystem II (Fig. 3d), the interplanar and A [15], respectively, Mg · · · Mg distances are further increased to 5.0 and 10.0 ˚ and the two chlorophylls in the pair are now regarded as almost independent.
Porphyrin Supramolecules by Self-Complementary Coordination
55
These differences in the chlorophyll pair structures are interesting in view of the evolution of the photosynthetic reaction centers. In light-harvesting antenna systems, too, X-ray crystallographic studies have shown evidence of similar structural evolutionary effects [16–20]. The first clear antenna structure was reported in 1995 for the bacterial light-
Fig. 4 Arrangement of chlorophylls in a LHI and b B850 in LHII
56
Y. Kobuke
harvesting complex LHII [16]. In the same year, cryomicroscopy provided an image of LHI [17], which is similar to LHII, but with a larger ring structure containing the reaction center in its central pore. LHII’s B850 ring is arranged in a barrel form, and both it and the similar (quasi)ring of LHI [18] employ the same process for constructing antenna structures, as is shown in Fig. 4. Here, two (bacterio)chlorophylls are arranged almost in a slipped cofacial orientation (similar to the special pair discussed above) by the coordination of imidazolyl side chains from the transmembrane α helices. This dimer unit is further assembled into a barrel form, thus constituting a macrocyclic ring composed, in the case of B850, of 18 (bacterio)chlorophylls [19]. In the case
Fig. 5 High-resolution AFM image of photosynthetic units in bacterial membrane
Fig. 6 Energy transfer rates in photosynthetic membrane of purple bacteria
Porphyrin Supramolecules by Self-Complementary Coordination
57
of LHI, the originally reported ring is composed of 32 (bacterio)chlorophylls. In some reports, however, this structure is modified and the ring is open to afford a gate for quinone to pass through [18]. Recent developments in AFM techniques elucidate further the closely packed structures of these antennae in the membrane (Fig. 5) [20]. These ordered structures, existing as they do in close proximity, are thought to enable the rapid, low-loss transfer of excitation energy to the (bacterio)chlorophyll components in the rings and to the neighboring macrocycles. The rates estimated are much faster compared with the lifetime, and are of the order of femto- to picoseconds, as is shown in Fig. 6 [21–26]. The ring structures observed in these bacteria systems have some obvious advantages over other structures in terms of facilitating energy transfer. The energy level of every component in the ring is almost identical, and smooth energy migration is expected to occur, since there is no large intervening en-
Fig. 7 X-ray structure of the LHC-II trimer. a Top view of trimer: chlorophylls assemble to form rings 1 and 2. b Side view in grana membrane: chlorophylls are located near the membrane surfaces along the dotted line
58
Y. Kobuke
ergy barrier. As the cyclic structures are densely packed in a two-dimensional space, each one will have many surrounding rings, so that rapid inter-ring energy transfer becomes easy. Also, the reaction center (RC) complex trapped in the large LHI antenna can ultimately accept the excitation energy circulating in this ring, thus starting the photoinduced oxidation–reduction reaction sequence. Compared to the antenna systems in bacteria that we have considered so far, those found in chloroplasts (including cyanobacteria) have quite different properties. Figure 7 shows the structure deduced via X-ray crystallography of one of the antenna systems of chloroplasts [27]. This is an antenna of outer source working mainly as the antenna system for photosystem II (PSII). Besides this antenna, many chlorophylls are also incorporated as antennae inside both PSI and PSII. The outer antennae, containing 12 chlorophylls in each unit, assemble to make a trimer. Many trimer units further assemble around the PSII. The chlorophyll arrangement in this trimer unit is completely different from that of the bacteria previously discussed. The principle of the assembly seems to be that the chlorophylls assembled in the inner part (ring 1 in Fig. 7a) transfer the excitation energy to the outer assembly (ring 2 in Fig. 7a), through which the energy is further transferred to the adjacent trimer unit. Furthermore, it must be noted that all of these chlorophylls are located near the outer membrane surface (i.e., along the dotted lines in Fig. 7b), so that the energy may in fact be transferred across the grana membrane. In contrast to the structure of bacteria, this arrangement is clearly quite complex, making it difficult to extract the structural principles of the plant antenna system. When grown in iron-depleted conditions, even more antenna molecules attach themselves to the PSI surface. Figure 8 shows such a situation, in
Fig. 8 Electron micrograph of antenna rings assembled further around trimeric photosystem I in cyanobacteria grown under iron-deficient conditions
Porphyrin Supramolecules by Self-Complementary Coordination
59
which a further 18 outer antennae (containing 18 × 12 = 216 chlorophylls) have assembled at the outer surface of a system which already contained 270 chlorophylls as components of its inner antenna. Clearly, the creation of extra antennae assists the maximum functioning of the reaction center when it is grown under starving conditions.
3 Self-Complementary Coordination of Porphyrins 3.1 Dimer On receiving light energy from the surrounding antenna complexes, the first oxidation–reduction step in the photosynthetic process takes place via electron transfer from the photoexcited special pair to pheophytin in the bacterial reaction center. The charge separation is accomplished with a time constant of 3 ps [29], which is exceptionally fast considering the edge-to-edge distance A. The special pair is considered to play an important role, as it is of 10 ˚ known to affect the efficiency of the charge separation process [30, 31]. For this reason, this configuration has long been a target of intensive research, and several porphyrin dimers mimicking the structure of the special pair have been synthesized by both covalent [32–35] and noncovalent [36–38] approaches. In covalent approaches, two porphyrins are attached to each other through rigid aromatic or cyclophane-type bridges, as shown in structures 1–3. (Note that the bold numbers refer to illustrated structures throughout.) In these models, two porphyrins are placed either in a face-to-face arrangement, or with a dihedral angle of 60◦ . Electronic interactions between the porphyrin components are related to transitions in the Soret band, and have been explained using Kasha’s exciton coupling theory [39]. Face-to-face dimers shift the Soret band to a shorter wavelength, and the fluorescence in this case is quenched significantly due to the transfer of electronic excitation energy to the vibrational energy levels. This is an effect that has also been observed in
60
Y. Kobuke
other H-type aggregates of various chromophores [40]. The arrangement of two porphyrins with a finite dihedral angle causes, in principle, a splitting in the energy bands due to the existence of both face-to-face and head-to-tail orientations of the two in-plane transition moments of the porphyrin. This is also the case for slipped cofacial arrangements, where the existence of two similar transition moments produces split Soret bands. As an alternative, we have reported a noncovalent approach to the construction of slipped cofacial porphyrin dimers, with close reference to the X-ray crystallography of the photosynthetic reaction center [36]. If one looks into the formation of special pairs, it becomes apparent that the transmembrane helices from the L and M subunits in the reaction center provide the imidazolyl ligand to the chlorophylls. This coordination structure itself could be cut off from the surrounding peptide moieties, and the two imidazolyl-tochlorophyll pairs could be turned 180◦ in a disrotatory fashion along the axes, thus penetrating the chlorophylls. The imidazolyl group can be transferred by this symmetry operation to a position from which it can be connected directly to the counterpart chlorophyll, providing the required complementary structure (Fig. 9). According to this idea, porphyrin and Zn(II) ions were substituted for chlorophyll and Mg(II), respectively. The employment of Zn(II) instead of Mg(II) was motivated by a consideration of the properties of magnesium porphyrin, which is too readily oxidized, and from which heat and acid easily remove the metal. Also, Mg(II) insertion is considerably more difficult than Zn(II) insertion. The second important point for any special pair mimic designed to replicate the photosynthetic model is that the central metal ion should not quench the singlet excitation energy. Zn(II) fulfills this requirement. Furthermore, zinc porphyrin has an additional merit from the viewpoint of coordination chemistry. Since Zn(II) can accept only one axial coordination, the dimer is the only coordination product without any additional complexation events. As was discussed in Sect. 2.1, Zn(II) gives, in general, large stability constants with neutral nitrogen ligands. It should be noted here that selfcomplementarity results in an extraordinarily large stability constant (on
Fig. 9 C2 operation on the special pair to produce a self-complementary dimer of imidazolyl porphyrins
Porphyrin Supramolecules by Self-Complementary Coordination
61
the order of 1011 M–1 in toluene). Therefore, the dimer is practically the sole species existing in nonpolar solutions, even at concentrations as low as micromoles. If coordinating solvents are added, however, the dimer structure collapses and its degree and rate of formation are dependent on the nature and amount of solvent. This does not necessarily represent a drawback, however, but can be regarded as a property allowing more sophisticated supramolecular structures to be constructed. Combining the processes of formation, breakage, and reorganization of supramolecules does indeed lead to some very useful results, and various examples of the application of this methodology to obtain more elaborate supramolecules will be described in later sections. Pyridyl ligands are certainly candidates for forming stable self-complementary coordination assemblies and are conventionally employed for supramolecular structure formation. Zn and Mg complexes of meso- and β-pyrrole-substituted 2-pyridylporphyrins (4a–c) were therefore prepared [37, 38]. Crystallographic analysis was used to verify that the required slipped cofacial arrangement had been generated, and demonstrated the overlapping of one of the pyrrolic rings in each porphyrin. This confirmed the structural similarity between the sample and the special pair. The stability constant was reported to be 5 × 105 M–1 for the Mg dimer of the meso-pyridyl case 4b. The stability constants for the monotonic coordination to zinc porphyrin by N-methyl-2-imidazole, compared to the constant for coordination by pyridine, is larger by one order of magnitude (∼ 104 M–1 compared to ∼ 103 M–1 , respectively) [41]. Several factors contribute to this difference. One important factor is certainly the difference in their basicities, since the imidazolyl group has a larger pKa value than pyridine (see Sect. 2.1). The proton occupying the alpha position relative to the nitrogen hinders the nitrogen coordination more appreciably for the six-membered heterocycle than is the case for the five-membered ring. Another factor affecting the complementary coordination process may be the deviation of the coordination angle from the perpendicular orientation, since the carbon occupying the alpha position relative to the nitrogen is connected to the porphyrin ring. For selfcomplementary dimer formation, the 2-imidazolyl-substituted dimer 5 must
62
Y. Kobuke
experience considerable angle strain, since it deviates from the ideal 90◦ to 72◦ (18◦ deviation) in the molecule. This strain energy may lower the stability constant from the ideal value. However, the 2-pyridyl ligand must accommodate an even higher strain energy in the molecule, since the angle in that case deviates by 30◦ from the perpendicular coordination. Such an angular strain could be avoided by connecting the pyridyl group through an angle-adjusting connector, such as an arylamide spacer 6 [42]. In this case, however, the distance between the two porphyrin rings becomes longer, and little stabilizing contribution from electronic and π–π interactions between porphyrin chromophores is expected. In other words, the desired complementarity is much less likely in this case. The many rotational degrees of freedom of the connecting bonds decrease the stability constant, too. X-ray crystallography was used to obtain the structure of the selfcomplementary dimer of N-methylimidazolyl zinc porphyrin (Fig. 10; Inaba Y, Nomoto A, Kobuke Y, personal communication). The angle strain of the imidazolyl coordination is eased by rotation of the imidazolyl ring to widen the angles of Cmeso –imidazolyl C2 –imidazolyl N3 and of imidazolyl C2 –imidazolyl N2 –Zn to 127.9 and 130.5◦ , respectively, from the 126◦ expected for the putative strain-free model. The slight bending (3.9◦ ) of the Cmeso and imidazolyl–C2 bonds from the porphyrin plane may also help to decrease the strain. It is noteworthy that the interplanar distance between the A, implying strong electronic interaction betwo porphyrins is close to 3.2 ˚ A, making the slipped cofacial tween the two. The zinc-to-zinc distance is 6.13 ˚ structure very similar to the special pair in the bacterial reaction center [12]. A strong exciton interaction was observed for 5 in the Soret band, which split into a twin peak with a splitting energy of ∼ 1300 cm–1 . Two transition moments, Mx and My , interact in this case, resulting in two allowed transition dipoles (the head-to-tail and face-to-face orientations). These are observed as red- and blue-shifts of the Soret band, respectively. This split Soret band,
Porphyrin Supramolecules by Self-Complementary Coordination
63
Fig. 10 a X-ray crystal structure of the dimer of self-complementary imidazolyl zinc porphyrins. b Mode of strain release in self-complementary dimer
64
Y. Kobuke
along with upfield shifts in the resonances of the coordinating imidazolyl protons from the porphyrin in the 1 H NMR spectrum, are highly characteristic of self-complementary dimer formation. A similarly split Soret band, with a splitting energy of ∼ 1040 cm–1 , was observed in the 2-pyridyl-substituted dimer 4a at a concentration of 5.2 × 10–5 M in CH2 Cl2 [37]. 3.2 Macrocycle and Box The beautiful structures of bacterial antennae have been shown to possess, within natural photosynthetic systems [19, 20], extremely important biological functions. This alone provides more than adequate motivation for examining their photophysical properties, since there is no doubt that the inherent nature of energy capturing and its transfer must be associated with them. It is exciting to be able to use analyses of synthetic systems to extract the reasons why nature uses such rings for this process. By analyzing simple systems, we can both understand the complexity of the natural process, and find ways to improve the efficiency of energy transfer in general, which may lead to the use of excitation energy in various applications. Compared to the simpler bacterial cases, it is difficult in cyanobacteria or chloroplast systems to extract the principles of operation and explain why nature employs intrinsic antenna systems. In ways similar to those described for the photosynthetic reaction center, evolution may have deformed a simpler original structure, added extra factors, and deleted others. For this reason, research necessarily focuses on the bacterial system. The antennae of regular macrocyclic structures are sufficiently characterized in this system by photophysical measurements based on conservation of excitation energy and its hopping efficiency. This part of the review is primarily concerned with supramolecular organization. Covalent approaches are not described. It must be briefly noted here that there are many reports on the synthesis and characterization of macrocycles with artificial antennae in various forms of macrorings 7–14 [43–48] and wheels [49]. Selective macroring formation has been achieved in the presence of suitable templates for macrocyclization [50–52]. These have been the subject of various review articles [53–55]. 3.2.1 Antenna Mimics of B850 and Box Types As described in Sect. 2, imidazolyl zinc porphyrin is an excellent candidate for self-complementary coordination due to its extremely large stability constant. In order to arrange these complementary coordination units into a hexameric macrocycle, two imidazolyl zinc porphyrins were connected through an m-phenylene group to fix the angle at 120◦ , offering a gable porphyrin
Porphyrin Supramolecules by Self-Complementary Coordination
65
66
Y. Kobuke
arrangement. The initial Zn insertion, however, gave a complex mixture which was characterized by gel permeation chromatography (GPC) analysis (Fig. 11). This mixture was then dissolved in CHCl3 /MeOH (1 : 1, v/v) at a dilute concentration of 3.5 µM, and the whole solvent was simply evaporated. GPC analysis of the remaining solid showed only two peaks with a 1 : 1 ratio, excluding almost all of the oligomeric peaks of higher molecular weights. The two peaks were separated easily with a preparative GPC column. The products were stable in the absence of coordinating solvents and the first and second peaks were determined to be macrocyclic hexamer 16 and pentamer 17, respectively [56, 57]. The reorganization process can be explained as follows. The dissolution of the substance in a large amount of solvent containing MeOH resulted in significant cleaving of the coordination bonds. During the subsequent evaporation process, complementary coordination bonds regenerated. Provided the solution is highly dilute, it is unlikely that the molecules will encounter other molecular species. Any molecular terminals with an unsaturated coordination bond try to find their counterpart in the same molecule to close the ring structure. This process follows the principle of macrocycle preparation (Fig. 12). The pentameric ring can form by sharing the angle strain in the macrocycle components. Tetrameric and heptameric rings are difficult to form because of the large strain energy associated with these structures.
Fig. 11 GPC charts immediately after zinc insertion (dashed curve) and after reorganization (solid curve)
Porphyrin Supramolecules by Self-Complementary Coordination
67
Fig. 12 Reorganization into macrocyclic 6-mer 16 and 5-mer 17 from gable imidazolyl zinc porphyrin 15
As we have just outlined it, supramolecular structure formation sounds very simple. The elucidation of the structure after its formation is, however, a tougher problem. This is because the intensity of the peak associated with the monomeric gable porphyrin unit in all the mass spectra is overwhelmingly high compared to the tiny peak associated with the dimer. Initially, no sign was found of such large macrocycles as 5- or 6-mers, although all the data from the GPC and atomic force microscopy (AFM) analysis, and from the small-angle X-ray scattering experiments, agreed with the hypothesis that there was pentamer and hexamer formation. All of this latter information, however, was only suggestive, rather than being conclusive. Direct evidence was eventually obtained using an olefin metathesis reaction (Fig. 13) [58]. Two trans meso positions in each porphyrin were appended with allyloxycarbonylethyl groups (18) and quantitatively connected to each other via a Grubbs catalyst to give 19. Then all the complementary coordination pairs in the macrocycles of tetra(allyloxyester)-substituted gable porphyrins 20 and 21 were metathesized doubly at two meso positions to yield 22 and 23, respectively. The parent peaks in the mass spectra were now identical to those calculated for 5- and 6-mers 22 and 23, respectively, thus allowing their structures to be identified (Fig. 14) [59]. Hunter [60] extended the idea of self-complementary dimer formation to higher macrocycles by adjusting the coordinating angle of the appended amidopyridyl ligands to selectively obtain the trimeric 24 and tetrameric
68
Y. Kobuke
Fig. 13 Covalent linking of self-complementary coordination dimer 18
Fig. 14 Fixation of the cyclic structure through covalent linking by metathesis reaction
macrocycles 25 of zinc porphyrin. In this case, the stability constants are not large enough to ensure the success of photophysical measurements at micromole concentrations because of the presence of freely rotating bonds in the substituents. In order to obtain larger macrocycles, some improvement in the stability constants is required. Since Co(II) porphyrin can accept two axial ligands, the introduction of two meso substituents leads to a one-dimensional array or macrocycles by
Porphyrin Supramolecules by Self-Complementary Coordination
69
adjustment of the angle between the two porphyrins in each complementary coordination pair. When the angle was adjusted to 30◦ by the choice of two meso-amidopyridyl substituents of different lengths (26), it has been claimed that macrocyclic dodecamer 27 (360/30 = 12) is formed by successive complementary dimer formation at concentrations lower than 0.5 mM (Fig. 15) [61]. When the concentration exceeds this effective molarity for cyclization, the macrocycle opens up, resulting in mixtures of polymers with higher molecular weights, as is demonstrated by their GPC behaviors. In order to address the ring size and its specificity or homogeneity, more direct and precise
Fig. 15 Co(II) porphyrin with two different pyridyl substituents (26) and self-assembled macrocyclic oligomer 27
70
Y. Kobuke
information on the molecular weights is definitely required. The ability of this form of porphyrin to accept two axial coordinations certainly results in facile structure formation, but cobalt quenches the porphyrin fluorescence, hampering the investigation of energy transfer events between chromophores. Similar features are associated with most metal ions that accept hexadentation. Osuka et al. reported numerous porphyrin arrays of various types, mostly obtained through covalent approaches, but also reported several tetrameric supramolecular macrorings 28 composed of mono-p-pyridyl-substituted zinc porphyrins [62]. When two pyridyl zinc porphyrins are connected directly at their meso positions, the coordination organization allows for the formation of porphyrin boxes 29 by two-point self-coordination. Interestingly, macrocyclization proceeds through a homochiral self-sorting assembly, as is evidenced by the successful optical separation of two enantiomers to display strong Cotton effects. The stability constant of the substance in CHCl3 was increased significantly from 1.4 × 1015 M–3 for monocyclization (n = 0 series) to > 1025 M–3 for the box formation described. The presence of the intervening phenylene groupings between the porphyrin and pyridine decreased the stability constants of the macrotetramerization [63]. The rigidity of the product contributed to an increase in the local concentration of the threedimensional cyclization. Aida et al. reported box formation from bis- (30) and tris-pyridyl zinc porphyrins (32), both connected through butadiynyl linkages. In these cases, the porphyrins can rotate around the connecting butadiynyl bonds, and it is proposed that the linked porphyrins are in an equilibrium between coplanar and orthogonal conformations in the resulting boxes 31 and 33 (Fig. 16) [64, 65]. A ruthenium-connected dimer of bis-(4 -
Porphyrin Supramolecules by Self-Complementary Coordination
71
Fig. 16 Cyclotetramerization of alkynylene-bridged pyridyl zinc porphyrin dimer 30 and trimer 32
72
Y. Kobuke
pyridyl) zinc porphyrin also underwent self-complementary coordination to make a porphyrin tetramer 34. The structure was determined by X-ray crystallography in the solid state, confirming the analysis of the highly symmetric S4 structure observed in solution using NMR. Both NMR and UV/Vis spectra were concentration independent, implying high coordination stability in a concentration range from milli- to micromoles [66]. Balaban et al. suggested that there was preferential tetramer 35 formation from meso-2-aminopyrimidinyl zinc porphyrin in dry toluene at millimolar concentrations. The Soret band shows temperature-dependent equilibrium between the monomer and its associated species. The associated enthalpy
Porphyrin Supramolecules by Self-Complementary Coordination
73
change was estimated to be 220 kJ/mol, which corresponds to about five times the Zn-pyrimidine ligation energy of 35–40 kJ/mol. X-ray crystallography confirms that the tetrameric structure of 35 is given by head-to-tail Zn – N(aromatic) coordination bonds [67]. In this context, Aida reported the preferential formation of cyclotetrameric structures from meso-(3-pyridyl) zinc porphyrin 36. The supramolecular pyridyl-coordinated porphyrins displayed thermochromic change in the range 0 to 100 ◦ C. The color was sensitive to the number of meso-alkynyl groups. Monoalkynyl-pyridyl porphyrin showed the most distinct stepwise color change, changing from green to yellow to red as the temperature moved from 0 to 50 to 100 ◦ C [68]. We noted earlier that a Co(II) porphyrin appended with meso-5-(1-methyl)imidazolyl undergoes dynamic equilibrium between cyclic trimer 37 and tetramer 38 depending on the concentration and the solvent employed (Fig. 17). The trimer is the predominant species in MeOH, while the tetramer is preferentially obtainable by raising the concentration of the CHCl3 solution [69]. Although pyrazole is inferior to imidazole as a ligand coordinating to Zn porphyrins, it has an additional NH group that can assist the coordination by hydrogen-bond formation. When 5-(4-pyrazolyl) zinc porphyrin is appended with a 20-(o-benzoate) group, it is automatically cyclized to the
Fig. 17 Formation of cyclic trimer 37 and tetramer 38 from imidazolyl Co(II) porphyrin
74
Y. Kobuke
trimer 39 by angle restriction on coordination with a stability constant of 2.3 × 1015 M–2 [70]. Ru(CO) and RhCl porphyrins with a meso-(p-pyridyl) substituent give rise to the formation of macrocyclic tetramers 40 and 41, respectively. The Ru tetramer structures were converted by an excess amount (103-fold) of pyridine into monomers. The Soret band sharpened during the change, showing the effect of excitonic interactions between the cofacially arranged porphyrins [71, 72]. 3.2.2 Porphyrin Macrocycle Capable of Guest Incorporation In the above gable porphyrin approach, the angle between the two imidazolyl zinc porphyrins was adjusted to 120◦ to facilitate the formation of the macrocyclic hexamer. When two porphyrins are linked through a central porphyrin via m-phenylene connectors to adjust the dihedral angle of the two terminal imidazolyl zinc porphyrins to 60◦ , the cyclic trimer 42 is the only macrocycle formed after the dilution and reorganization process (Fig. 18) [73]. In this case, neither the cyclic dimer nor the tetramer are produced because of their high angle-strain energies. The characteristic feature of this macrocycle is that the central zinc porphyrin units do not participate in the structure formation by complementary coordination and can accept the axial coordination, especially from the central pore side. In this way tripodal ligands can, if properly designed for cooperative coordination, be specifically trapped within the central pore. UV/visible titration of the cyclic trimer 42 with the tetrapyridyl ligand 43 gave 44 with a stability constant of 8 × 108 M–1 in toluene. The large stability constant enables us to construct an antenna reaction center composite by introducing electron acceptors into the fourth arm of the tetrapodal ligand.
Fig. 18 Specific formation of macrocyclic trimer 42 and incorporation of a tripodal guest in the pore
Porphyrin Supramolecules by Self-Complementary Coordination
75
3.2.3 Size-Tunable Giant Macrocyclic Arrays An interesting way of adjusting the size of macrocycles is to use a spacer which, instead of having a fixed angle, allows hingelike motion. With this
Fig. 19 Formation of giant macrocycles by self-coordination of ferrocene-bridged zinc trisporphyrin units
76
Y. Kobuke
in mind, ferrocene spacers were employed to connect the porphyrin units in the trisporphyrin building block with two imidazolyl terminals (45). Bis(imidazolyl)-trisporphyrin with two ferrocenyl spacers dimerized spontaneously on Zn insertion by the complementary coordination of two terminal imidazolyl zinc porphyrins. Dimer 46 was dissolved in pyridine and simple evaporation of the pyridine turned the compound into polymer of high molecular weight. This phenomenon is accounted for by the fact that the concentration exceeds the critical monomer concentration (cmc), above which only linear polymer increases occur [74]. When left to stand in commercial CHCl3 , which contains 0.5% EtOH, the polymers gradually changed size. After 14 days, at least nine peaks (including the dimer 46) were observed in the GPC elution curve. All of the peaks were separated and subjected to ring-closing metathesis reactions, and were thus identified as a series of macrocycles composed of three to ten trisporphyrins 47 (Fig. 19). The largest ring was composed of 30 porphyrins and 20 ferrocenes, where 40 meso substituents were metathesized pairwise. The fluorescence from these macrocycles was quenched by rapid electron transfer from the ferrocenyl part of the ring. Single molecular images of the cross-linked macrocycles were obtained using scanning tunneling microscopy (STM) [75] under ultrahigh vacuum conditions with a pulse injection technique [76, 77]. Clear circular arrangements of ten and five bright spots, corresponding to the coordinated dimers in the respective supramolecules, were observed. These coordinating dimer units seem to be standing perpendicular to the substrate, with the central porphyrin units in the trisporphyrin lying horizontally. 3.3 One-Dimensional Arrays A one-dimensional porphyrin array is the most basic product to come from the connection of porphyrin units. The preparation of one-dimensional porphyrin arrays has been approached in various ways, using not only covalent but also noncovalent approaches. There are various ways to connect π-conjugated molecular systems ∼ 1 nm2 in size. Planar, conjugated, fused, orthogonal, and helical array connections are all available. Depending on the mode of connection, these arrays may be used as materials for molecular wires, photocurrent generation, optoelectronics, and so on. Therefore, many studies have been reported on the preparation of one-dimensional porphyrin arrays. Meso–meso coupling of porphyrins has opened the way to orthogonally linked arrays, where a meso-free Zn porphyrin is treated successively until the number of porphyrins in the array reaches 2n after n steps of Ag(I)promoted oxidative couplings (Fig. 20) [78, 79]. Ten such repetitions make available a 1024 (= 210 )-mer (48) as a monodispersed porphyrin array. Com-
Porphyrin Supramolecules by Self-Complementary Coordination
77
Fig. 20 Stepwise syntheses of meso–meso covalently linked linear multi-porphyrin arrays
bined interactions of coordination and hydrogen bonding convert the linear array into a helical one. The structure and photophysical properties of these one-dimensional arrays have been characterized by 1 H NMR, AFM, STM, and time-resolved spectroscopies. Porphyrin tapes highly conjugated by triply fused rings were prepared; their Q-bands were shifted to dramatically longer wavelengths, with those of the porphyrin 12-mer 49 appearing in the IR region [80, 81]. Anderson has reported a conjugated, butadiyne-connected polymeric zinc porphyrin array that transforms into a porphyrin ladder 50 upon titration with a bridging ligand such as 4,4 -bipyridyl [82, 83]. These long π-conjugated porphyrin systems are interesting in view of their applicability in the field of nonlinear optics [84–86] and as two-photon absorption materials [87, 88]. Self-complementary dimer formation of imidazolyl zinc porphyrin allowed two such units to be connected directly at their meso positions to give 51. Since the two coordination assemblies grow in opposite directions in this case, successive self-complementary coordination leads to a linear multi-
78
Y. Kobuke
Fig. 21 Formation of one-dimensional linear multi-porphyrin supramolecular arrays 52
porphyrin array (52) with a large stability constant (Fig. 21) [89]. Figure 22 shows a GPC elution curve of the product along with the curves for standard polystyrene mixtures. The molecular weights at the maximum and distribution peaks were estimated to be 5 × 105 and 1 × 105 , respectively, based on a calibration against the polystyrene mixtures. This corresponds to arrays of 800 and 160 porphyrin units. The AFM image of this sample (Fig. 23) shows a molecule that is a few hundreds of nanometers in length. The absorption spectrum for this array clearly demonstrates how the selfcomplementary coordination is exploited during the array formation. The Soret band at the longer wavelength is now shifted to 490 nm from the position of the dimer signal (450 nm), the red-shift occurring due to successive head-to-tail excitation interactions as described above. On the other hand, the Soret band at the shorter wavelength does not shift significantly from the dimer signal position because face-to-face interactions in the meso– meso dimer are orthogonal to each other and cannot be transferred through the linkage. Although the self-complementary coordination is very stable in nonpolar solvents, it is adversely affected by the addition of coordinating solvents. The addition of MeOH decreases the splitting width of the Soret bands. The face-to-face interaction is cleaved, inducing a small red-shift of the shorter wavelength Soret band. On the other hand, the collapse of the head-to-tail interaction results in a large blue-shift of the longer wavelength band, which passes through a clear isosbestic point. From this it can be concluded that the choice of solvents or other external ligands can be used to control the length of the linear multi-porphyrin array produced.
Porphyrin Supramolecules by Self-Complementary Coordination
79
Fig. 22 GPC charts of 52 (solid line) and a mixture of polystyrene standards (dotted line) eluted by CHCl3 . The exclusion limit of the column was 5 × 105 Da
Fig. 23 AFM image of linear multi-porphyrin arrays 52 dispersed on a mica plate
These formation and dissociation equilibria can be used to manipulate the multi-porphyrin assemblies, as was demonstrated in the following experiments. A mixture of coordination polymer 52 and dimer 53 was dissociated by the addition of MeOH followed by simple evaporation. The GPC elution curves before (Fig. 24a) and after (Fig. 24b) the evaporation were completely different. The new peaks (1, 2, and 3) in Fig. 24b were separated and identified as oligomers terminated by monomeric porphyrins 54 (n = 0, 1, 2, and 3). This result demonstrates the possibility of controlling the length of the porphyrin array and terminating it with other imidazolyl zinc porphyrins
80
Y. Kobuke
Fig. 24 GPC chart of a mixture of 52 and 53 in a 5 : 1 molar ratio (dotted line) and the same sample mixture when dissolved in CHCl3 /MeOH (1 : 1), then evaporated and redissolved in CHCl3 (solid line)
(Fig. 25). Alternatively, one can “reverse” the process. Porphyrin array formation can start with imidazolyl zinc porphyrin, which can be immobilized using, for example, thiolate attachment on the electrode surface [90, 91]. In this case the porphyrin attached to the electrode can serve as a molecular solder connecting the electrode to polymers made of hundreds of meso–mesoor bis(acetylene)-linked bis(imidazolyl zinc porphyrins) 55 [92]. Molecular wires or light-energy conversion systems could be built on the basis of such technologies. If we encourage hexa-coordination by using a central metal ion, even a single porphyrin unit can produce an array structure by having two coordinating ligands attached to it. The Mg(II) ion is typically chosen for this process. However, the sixth coordination is rather weak and gives a long array structure. Bis(imidazolyl) Mg porphyrin, for example, has been found to produce only one-dimensional oligomers (56) up to the heptamer level, as detected by
Fig. 25 Porphyrin oligomers 54 terminated by dimer porphyrin 53
Porphyrin Supramolecules by Self-Complementary Coordination
81
electrospray ionization mass spectrometry [93]. When imidazolyl rather than N-methylimidazolyl substituents are appended at two meso positions of the porphyrin, and the ion inserted is Ga(III), the resulting Ga(III) porphyrin can accept both an imidazolide anion and a neutral imidazolyl ligand. As a result, a porphyrin array with a staircase structure (57) can be obtained [94]. AFM measurements suggest that this results in the formation of rodlike assemblies a few hundreds of nanometers in length. This type of one-dimensional multi-porphyrin array is a possible candidate for not only a one-dimensional antenna, but also a molecular wire for molecular electronic applications. The photoirradiation of a film of such arrays has been shown to enhance the electric conduction across a coated electrode. Co(III) can form a long linear array (58) by accepting a much stronger coordination from imidazolyl, although the singlet excitation energy is quenched in this case [95]. As is shown in Fig. 26, the introduction of two N-methylimidazolyl substituents resulted in a widely split Soret band at 404 and 474 nm. This was in contrast to the Co(III) monoimidazolyl porphyrin dimer signal (dashed), which indicates the development of extensive exciton coupling. The results of a GPC analysis suggested a molecular weight (MW) of over 30 000, and AFM images showed many rodlike assemblies. These were mostly 50–100 nm in length, which corresponds to 90- to 170-mers. The maximum observed length was 1.1 µm (a 1900-mer). Hunter and Michelsen developed a methodology to obtain linear as well as macrocyclic arrangements of porphyrins by using the properties of Co(II) porphyrin, which can accept two axial ligands. Thus, a cobalt porphyrin with
Fig. 26 UV/visible absorption spectra of self-complementary Co(III) dimer (dotted line) and polymer 58 (solid line)
82
Y. Kobuke
two identical amidopyridyl substituents appended to it was shown to give rise to self-complementary dimers successively arranged in a linear fashion (59) [96]. (Note: These authors do not exclude the possibility of oxidation to the Co(III) state.) The molecular size, as deduced from GPC analysis, was found to grow depending on the concentration of the species. For this test, the concentration was varied from 55 to 7000 µM. 3.4 Two-Dimensional Arrays The introduction of supramolecular multi-porphyrin assemblies is expected to contribute significantly to the field of molecular electronic and photonic devices because of their excellent levels of light absorption and charge sep-
Porphyrin Supramolecules by Self-Complementary Coordination
83
aration, and their ability to facilitate the transfer of the resulting excitons, holes, and electrons. As has been demonstrated in the previous sections, selfcomplementary coordination is a powerful technique for connecting a large number of porphyrins into one-dimensional arrays. This methodology can be extended further to create two-dimensional networks by self-coordination of the cross-shaped zinc porphyrin pentamer 60, which has imidazolyl porphyrins at four meso positions. This array can propagate in four directions to form a planar porphyrin aggregate 61 (Fig. 27) [97]. When organized in a concentrated solution, only insoluble material was obtained. Well-organized species could be prepared, however, by performing a reorganization procedure. The compound was dissolved first in pyridine/nitrobenzene (1 : 10, v/v), after which gentle evaporation of the pyridine was allowed to take place. An absorption spectrum of a film of 61 on a glass plate shows split Soret bands with a peak separation of 930 cm–1 , indicating the presence of complementary self-coordination. AFM images of this substance on a mica substrate show a series of spots spread over 350 nm with a height distribution of 2.0–2.4 nm. Given a zinc-to-zinc diagonal distance of 3 nm, the spot length corresponds to an assembly of ∼ 130 porphyrin pentamer units. The thickness of the sample suggests a filmlike assembly formation of six to seven porphyrin layers, which accords well with the suggestion that there is a random approach from the top and bottom sides of the sample during coordination. Two-dimensional network formations of 62 using covalent approaches have also been achieved after multistep synthesis [98]. Two-dimensional propagation of self-complementary coordination provides a more robust network system for electrical conduction, since the network is not fully disrupted by defect formation in the structure. This property
Fig. 27 Formation of two-dimensional supramolecular porphyrin array 61
84
Y. Kobuke
is of the utmost importance if these substances are to be used to build conductive molecular devices.
4 Coordination Organization of Phthalocyanines Phthalocyanines are typically characterized by wide color variation along with high levels of stability in response to chemical reactions, heat, and illumination with light. They are often used as functional dyes of extreme durability [99]. Exploitation of their excellent electronic and photoconductive properties has led to their widespread use as materials for xerography, organic electroluminescence, solar energy conversion systems, and many others. Their expanded π-electronic plane is ideal for electron–hole delocalization due to the stabilization of anion or cation radicals in the large π-electronic system [100]. Phthalocyanines have large extinction coefficients in the Q-band region and high fluorescence quantum yields. Their possible application in the field of nonlinear optics [101], as optical data storage devices, and in the production of molecular-scale devices is also of extreme interest. In spite of extensive research, supramolecular methodologies for the organization of phthalocyanines through coordination processes have not been very successful. Various sandwich complexes have been prepared, and their
Porphyrin Supramolecules by Self-Complementary Coordination
85
electronic and electrochromic properties have been extensively studied. However, other examples of organization through coordination are rather limited. The pyrido-substitution of one of the benzo groupings in zinc phthalocyanine, where an edge-to-face monotopic complexation (63) was suggested by NMR and GPC analyses, has been reported [102, 103]. Kobayashi also reports the connection of titanium bis-phthalocyanine 64 to octahydroxyphthalocyanine through chelating or rather covalent bondings [104]. The incorporation of phthalocyanines into a multichromophore supramolecular system was attempted by attachment to imidazolyl zinc porphyrin 65 (Fig. 28). A combination of strong absorption of the Soret band of porphyrin and of the Q-band of phthalocyanine makes compound 66 an attractive dyestuff similar to chlorophylls and covering the whole visible re-
Fig. 28 Formation of self-complementary dimer 66 from phthalocyanine-substituted imidazolyl zinc porphyrin 65
86
Y. Kobuke
gion [105]. The porphyrin part exists as a self-complementary dimer and acts as an excellent energy donor for the phthalocyanine, because the fluorescence wavelength of porphyrin overlaps with the absorption band of phthalocyanine. Strong fluorescence, with a high quantum yield of 0.71, was observed from the free base phthalocyanine part of the system upon excitation of the whole absorption area, since fast and efficient energy transfer was observed when the porphyrin part was excited. Given these results, it seemed that the methodology developed in the field of porphyrin supramolecules could be extended to phthalocyanine if appropriate modifications were made. The introduction of an imidazolyl substituent at one of the benzo positions of metallo(Zn or Mg)phthalocyanine 67 spontaneously gave the self-complementary dimer 68 in a way similar to the behavior of the porphyrins discussed earlier (Fig. 29) [106]. The Q-bands were split into a twin peak and converged into a single band through a clear isosbestic point by the addition of a large excess of N-methylimidazole. From this competitive titration curve, the stability constants for the complementary coordination were estimated to be of the order of 1011 –1012 M–1 , depending on the metal ion species (Zn or Mg) and ring substituents ((t-Bu)4 or
Fig. 29 Self-assembled imidazolyl phthalocyanine dimer 68
Porphyrin Supramolecules by Self-Complementary Coordination
87
(OBu)8 ). These extremely large stability constants are certainly a consequence of the nature of complementary self-coordination itself, a fact which was established by investigation of the 1 H NMR spectra of the systems. These spectra demonstrated a characteristic shielding of the imidazolyl and half of the phthalocyanine ring protons. Since the substituent at one of the benzo positions of phthalocyanine does not lie along the molecular symmetry axis, imidazolyl coordination gives rise to the formation of two geometrical isomers. The NMR analysis established that there is a 1 : 1 mixture of species with parallel and oblique orientations. These isomers behave similarly and do not introduce any additional complexity into the spectroscopic characterization of the sample. Measurements of the oxidation potentials yielded results which reflected the close proximity of the two phthalocyanine planes. The first two oxidation potentials (one-electron oxidations from each phthalocyanine ring) were ∼ 100 mV apart, suggesting the delocalization of the cation radical over the two phthalocyanines’ π-electronic framework. This behavior shows some similarity to that of the porphyrin dimers, and is expected to favor energy- and electron-transfer reactions. Another characteristic feature of this phthalocyanine dimer is that its fluorescence quantum yields are almost the same as those of the corresponding monomers (0.45, 0.26, and 0.76 for Zn(OBu), Zn(t-Bu), and Mg(t-Bu) phthalocyanines, respectively). Such a highly fluorescent phthalocyanine dimer has never been reported before. In this way, self-complementary coordination developed in the field of porphyrin chemistry was successfully transferred to a related field, and has been shown to be a successful first step toward similar advances in phthalocyanine chemistry. Since phthalocyanine has several advantages over other materials in a range of important applications, this development should open up the field of supramolecular chemistry in phthalocyanines as an area of much more active research.
5 Structural Characterization of Supramolecular Assemblies As has now been demonstrated through various examples, supramolecular methodology is a useful technique for generating unique structures spontaneously, provided that the initial building blocks are properly designed. When successful, it certainly saves considerable time and effort compared to other, more tedious synthetic processes. In the approach using only covalent bonds, identification of the final product is, in principle, accomplished relatively easily. The identification is generally based on information about the specific molecular weight, which is found using mass spectroscopy measurements, and on other spectroscopic evidence, most notably on the results of
88
Y. Kobuke
NMR spectroscopy. In the case of the supramolecular sciences, synthesis of the target unit molecules which are to be identified is only the start of the chemistry. The supramolecular structure depends on the concentrations of the substances, the solvents employed, and many other factors, and in general is not readily verifiable. A detailed list of structure elucidation techniques has been described in a previous review [9]. Brief comments on these techniques are made below. GPC (gel permeation chromatography): GPC is very powerful not only for analytical but also for preparative purposes, if the stability constants of the supramolecules are large enough for them to be subjected to the process. Only nonpolar solvents are allowed for certain gels, but polystyrene-based gels allowing the use of polar solvents are also available. Difficulties may occur due to adsorption of the sample to the resin, in which case it may be worthwhile to try the addition of LiBr salt to the eluent. Mass Spectrometry: Modern organic chemistry depends on the development of technologies, such as soft ionization of molecules without bond breaking, which are now an integral part of mass spectrometry. ESI-TOF mass spectrometry via the cold ion-spray method is very powerful. In our limited experience, however, this was not a valid method for our compounds. Almost all of the species analyzed in this way gave only the dissociated assembly unit, even though the supramolecules belong to a class with amongst the largest stability constants. The best solution to this problem was to use metathesis to link the self-complementary coordination pair. For example, owing to the almost quantitative nature of the reaction, 40 allylic meso substituents were successfully connected pairwise in a decamer of ferrocene-connected trisporphyrin units (see Sect. 3.2.3). After the metathesis linking, the parent peak was observed at the calculated mass number of 21 783.72 (M + 1) with a sufficient peak intensity for it to be properly identified. AFM (atomic force microscopy): Recently, a number of AFM images have been published which disclose the assembled structure of both natural and reconstituted photosynthetic systems [20]. The scanning is accomplished by the application of smaller cantilever amplitudes and weaker forces than are normally used, the measurements being made with an oxide-sharpened Si3 N4 tip with a 2–5-nm edge. Just as elucidation of the photosynthetic reaction center and the antenna by X-ray crystallography in the 1980s–1990s stimulated the construction of artificial photosynthetic unit structures, similarly their total assembled structures (which can now be viewed for the first time) have become interesting targets for construction. STM (scanning tunneling microscopy): Pulse injection of the sample solution onto a metal surface under ultrahigh vacuum gives excellent images of single molecules within the sample [76, 77]. This contributes greatly to the observation of nonsublimable supramolecules of chemical significance. While it must be remembered that these images result from a strong interaction between the target molecule and the substrate, this technique permits the ac-
Porphyrin Supramolecules by Self-Complementary Coordination
89
Fig. 30 STM images of macrocyclic 10-mer (left) and 5-mer (right) from ferrocenebridged zinc trisporphyrin unit 45
quisition of unique information about single molecules that would otherwise be unobtainable (Fig. 30). VPO (vapor pressure osmometry): Unfortunately, this classical technique is becoming more and more difficult to employ due to the limited availability of ultrahigh-precision thermostats. Machines produced recently lack the high sensitivity required. Moreover, the molecular weights of recent interest have become in general too high to apply the concentrations required for detection with any meaningful level of precision.
6 Properties of Porphyrin Assemblies 6.1 Photoinduced Electron Transfer in Special Pair Mimic It is interesting to examine how the special pair mimic which has been created behaves in a photoinduced charge separation reaction. First, the redox properties of imidazolyl porphyrin dimer 5 were investigated. Differential pulse voltammetry of the dimer showed four one-electron oxidation peaks at 414, 632, 968, and 1120 mV vs Ag/Ag+ in chloroform. The first oxidation potential is close to or higher than the corresponding potential for the monomer (381 mV), which was obtained by the addition of an excess amount (220 equiv) of N-methylimidazole. It is important to note that dimer formation never favors the first oxidation of the porphyrin. However, the second oxidation potential, corresponding to one-electron oxidation of the second porphyrin unit in the pair, was shifted up by 228 mV. It is clear that the cation radical, once generated on one of the porphyrin units, must be subsequently be delocalized over the π-electronic system of the strongly interacting porphyrin dimer. Attachment of pyromellitdiimide as the electron acceptor to this dimer and to the dissociated monomer porphyrins with two different separation
90
Y. Kobuke
distances, 69, 70 and 71, 72, respectively, quenched the porphyrin fluorescence efficiently. In all cases, the fluorescence was quenched more efficiently in the dimer than in the monomer (as can be seen in the last column of Table 2), which suggests that electron transfer is occurring more efficiently in the dimer [107]. In addition, the time constants for both photoinduced charge separation (CS) and charge recombination (CR) were determined by measuring the fluorescence lifetime and the transient absorption of the porphyrin cation radicals and the pyromellitdiimide anion radicals. Unfortunately, in the case of the monomers, the CR rates were faster than the CS rates. In contrast, dimer 69 showed an accelerated CS rate (with the lifetime decreasing from 10 to 2 ps) and a decelerated CR rate (with the time taken increasing from 3 to 12 ps). The same is true for the 70 series, where the CS lifetime decreased from 1000 to 120 ps and the CR time increased from < 100 to 160 ps. As discussed above with reference to the oxidation behavior of the system, the cation radical is delocalized over two closely interacting porphyrins in the dimer, reducing the charge density of the dimer compared with that of
Table 2 Kinetic parameters of charge separation in dimers 69 and 70, and the corresponding monomers 71 and 72 in CHCl3 Compound
τCS (ps) a
τCR (ps) a
τf (ps) b
Qeff (%) c
71 69 72 70
10 2 1000 120
3 12 < 100 160
10 2 920 120
99.1 99.5 70.3 94.6
a
Time constants obtained by time-resolved transient absorption Fluorescence lifetime c Quenching efficiencies defined as Q = (1 – I eff compound /Ireference ) × 100, where I is fluorescence intensity. b
Porphyrin Supramolecules by Self-Complementary Coordination
91
Fig. 31 Schematic drawings of the Marcus parabolas for dimer and monomer
the monomer. This leads to a smaller solvent reorganization energy for the dimer, shifting the Marcus parabola [108–111] to the left. Since the CS rate is located in the normal region, this shift induces the observed increase in the rate, while the CR rate in the inverted region decreases (Fig. 31). It is clear that dimer formation accelerates the CS rate and decelerates the CR rate, facilitating the total photoinduced charge separation efficiency. This is a result of the dimer having a smaller reorganization energy than the monomer due to charge delocalization over the enlarged π-electronic framework. This simple model system thus clearly demonstrates the vital role which the special pair arrangement plays in the electron-transfer process during photosynthesis. 6.2 Energy Transfer in Photosynthetic Antenna Mimics Supramolecular porphyrin rings have been designed to mimic the antenna function of bacterial photosynthesis, taking into account the fact that the random assembly formation usually loses its singlet excitation energy due to the formation of energy trapping sites. The antenna rings investigated, macrocyclic pentamer 17 and hexamer 16, both behaved ideally. Almost no fluorescence quenching was observed, and they maintained a constant fluorescence lifetime of 2.2 ns both before and after the assembly formation. The hopping rate of the excitation energy was estimated via anisotropy depolarization and exciton–exciton annihilation dynamics (Fig. 32) [112]. Although the latter technique does not yield a unique time constant, the slowest of the laser power-dependent decay processes can be considered representational of the energy hopping process. This energy hopping is caused by exciton an-
92
Y. Kobuke
Fig. 32 Energy hopping rates in cyclic antenna complexes 16 and 17
nihilation due to the disproportionation of two excitons. As for anisotropy depolarization, one can extract the real hopping rate from the other decay processes by observing the rate of in-phase anisotropy decay and molecular rotation. Finally, this rate must be verified by considering the coincident time constant from two decay processes. According to the analysis, the energy hopping rates were estimated as 8.0 and 5.3 ps for the 5-mer and 6-mer systems, respectively. A larger exciton–exciton interaction level for the latter, as demonstrated by the larger splitting of its Soret band, may be responsible for the faster energy hopping rate. The combined evidence clearly demonstrates the fact that these self-assembled cyclic multi-porphyrin arrays are excellent mimics of the light-harvesting antenna of B850. The energy hopping rates analyzed for three-dimensional box-type assemblies are schematically shown in Fig. 33 [63]. In these cases, the rates were analyzed as if they were occurring in an xy plane without any anisotropy along the z direction. The rate was found to be strongly dependent on the center-to-center distance, indicating that the excitation energy migration processes are well described by a Förster-type incoherent energy hopping mode. Similar analyses have been made for covalently connected light-harvesting models [113]. A linear assembly formed through successive self-complementary coordination of meso–meso-coupled bis(imidazolyl zinc porphyrin) units can also transfer excitation energy along the one-dimensional array. In order to estimate the efficiency of this energy transfer, imidazolyl Mn porphyrin was added to the linear array of imidazolyl zinc porphyrins 52. The terminals of the array can accept complementary coordination with imidazolyl Mn porphyrin [114]. As a result, the fluorescence of Zn porphyrin (φ = 0.053) in the one-dimensional array is quenched. This quenching never happens
Porphyrin Supramolecules by Self-Complementary Coordination
93
Fig. 33 Energy hopping rates in antenna boxes 29
Fig. 34 Energy hopping along linear antenna complex and quenching at terminal Mn porphyrins
when Mn tetraphenylporphyrin is added without an imidazolyl substituent. It can therefore be concluded that efficient quenching occurs through a fast energy hopping process along the one-dimensional array to the terminal Mn porphyrin coordinated by complementary coordination. GPC analysis of the mixture of the bis(imidazolyl zinc porphyrin) array and imidazolyl Mn porphyrin indicates that the energy hopping quenching occurs through 130 porphyrin units in an array containing 200 porphyrin units in total (Fig. 34). 6.3 Nonlinear Optical Properties Nonlinear optical (NLO) materials are interesting in view of their possible photonic applications, such as ultrafast optical switching and modulation [115–117]. The general principle for obtaining large third-order susceptibility χ (3) values, which are required if nonlinear optical effects are to be observed, requires molecules composed of a π-conjugated core in association
94
Y. Kobuke
with polarizing donor/acceptor terminals. The coordination reorganization process introduced in Sect. 3.1 seems to afford an ideal tool for obtaining such an assembly for porphyrin [118]. A mixture of bis(imidazolyl zinc porphyrin) and monozinc-bis(imidazolyl porphyrin) was therefore reorganized from a CHCl3 /MeOH solution, and a series of oligomeric compounds with and without free base porphyrin terminals (73–75 and 76–78, respectively) were isolated. The optical Kerr effect (OKE), which is an NLO property, was assessed for these compounds. The measurement was made under off-resonant conditions at 800 nm using a femtosecond laser with a CS2 reference. The results are shown schematically in Fig. 35. The molecular second-order hyperpolarizabilities |γyyyy | of oligomers 73–75 were extremely large compared with those of the porphyrin polymers separated by a phenylene or an acetylene spacer. The values belong to the class of molecules showing the largest γ value, as is evident when one considers the value for a donor–acceptor pair of 2,7-diethynylfluorene and N,N,N ,N -tetrakis(4-phenyl)-4,4 -diamino-1,1 -biphenyl under similar conditions (4.5 × 10–30 esu) [119]. These large |γyyyy | values were observed only for those compounds having free base porphyrins at the molecular terminals. The combination of terminal free base porphyrins and a core of coordinationlinked zinc porphyrins can therefore be seen to be effective for enhancing the |γyyyy | value through efficient molecular polarization. The bisporphyrin discussed above is connected directly at meso–meso positions, and the two π-electronic systems are almost orthogonal to each other.
Fig. 35 Third-order nonlinearity of porphyrin self-assemblies from femtosecond OKE measurements
Porphyrin Supramolecules by Self-Complementary Coordination
95
In order to obtain a better conjugation, two porphyrins were then connected by a butadiyne grouping and organized in a similar way, so that a free baseterminated bisporphyrin dimer was generated. Two-photon absorption (2PA) cross section values (σ (2) ) were then measured using a femtosecond open aperture Z-scan method in the off-resonant region [120]. In contrast to the small σ (2) value (370 GM at 964 nm) measured for monozinc meso–meso-connected bisporphyrin 73, butadiyne-linked bisporphyrin 79 (Fig. 36) gave an extremely large value (7600 GM at 887 nm), suggesting that the increased π conjugation had a significant effect on the absorption rate. This value belongs to the largest class of such measurements for organic compounds in the femtosecond timescale [121–135]. In contrast, when the monomer had been dissociated by the addition of excess N-methylimidazole, the result was a value of 1800 GM. Monomeric biszinc complex 81 and free base 82 gave values of 1200 and 1000 GM, respectively. Comparison of these values indicates that the complementary coordination and polarization of the asymmetric zinc-free base structure also contributes to the enhancement of the 2PA cross section. Furthermore,
96
Y. Kobuke
Fig. 36 Two-photon absorption spectra of 73 (filled triangles) and 79 (filled circles) measured by the femtosecond Z-scan method
the self-assembled imidazolyl porphyrin polymer 52, with a mean molecular weight of Mn = 1.5 × 105 , had an extremely large σ (2) value of 4.4 × 105 GM at 873 nm [136]. The two-photon absorption cross section of any material depends on the square of the incident light intensity and occurs only at the focal point of the laser light. Therefore, much attention has been paid to this process as it applies to such developments as two-photon photodynamic therapy, threedimensional optical data storage, optical limiting, three-dimensional microfabrication, and fluorescence microscopy. Porphyrins and phthalocyanines have been candidates for 2PA materials because of their highly conjugated π systems. Although monomeric porphyrin derivatives have been shown to have relatively small σ (2) values (less than 102 GM), the examples shown here suggest possibilities of further improvements. 6.4 Photocurrent Generation Photocurrent generation is one of the most interesting direct applications of photosynthetic studies. The adsorption of sensitizers onto semiconductor surfaces has been found to be an efficient way to generate photocurrents and has been studied extensively. Ruthenium bipyridyl complexes, in particular, have been the focus of recent research [137–139]. In these cases, only the first layer of molecules, which is in direct contact with the surface, is active. A highly porous semiconductor material was therefore employed to compensate for the low level of absorption of the single molecular layer. Other varieties of chromophores, semiconductor materials, and electron carriers for totally solid systems have been the subjects of extensive studies. The present
Porphyrin Supramolecules by Self-Complementary Coordination
97
review addresses specifically the application of self-complementary coordination to this system. Photocurrent generation systems on flat electrode surfaces may become important in the development of applications involving flexible materials such as plastic electrodes. A simple self-assembled monolayer (SAM) of photosensitizers (including porphyrins) cannot generate a large enough photocurrent density to be practicable. This is because the absorbance of a single molecular layer is small even with chromophores having a strong Soret band, and its sharp absorption band without exciton interactions is too narrow to cover the whole visible range. In order to overcome these inherent problems, antenna layers can be induced to grow on the SAM from the electrode. Our approach [90, 91] involved the use of an imidazolyl porphyrin free base to which ω-mercaptoalkyl substituent 83 had been appended. This was
Fig. 37 Scheme of photocurrent generation. A porphyrin antenna was grown from a SAM on a gold surface by self-complementary coordination followed by covalent immobilization by a ring-closing metathesis reaction
98
Y. Kobuke
attached to a gold surface by a simple soaking process [140]. After the introduction of Zn(II), the SAM substrate was immersed in a solution of meso–meso-coupled imidazolyl zinc porphyrin dimer 84, which was dissociated partially in a CH2 Cl2 /MeOH solution. This was followed by a rinse with CH2 Cl2 to facilitate the imidazolyl-to-zinc coordination. After the bisporphyrins were coordinated complementarily, the coordination-organized pair was immobilized covalently through a metathesis reaction by simply dipping the plate into a solution of Grubbs catalyst. Alternating coordination deposition and metathesis reaction cycles were repeated until multiporphyrin layers were obtained (Fig. 37). The successive accumulation of multi-porphyrin arrays was clearly indicated by the gradual increase in the UV/Vis absorbance of the sample over a wide range (Fig. 38). Both the complementary imidazolyl-to-zinc coordination and the covalent linking of the allyl side chains are indispensable to the stable growth of the array structure. As a result of these processes, the covalently immobilized porphyrin arrays are resistant to dissociation when the next organizing solution is added, even though it contains a dissociation-inducing solvent. The optimum spacer length for the covalent linking was employed to allow a 95% yield in the solution, and the dimer connection proceeded efficiently on the solid substrate. Cathodic photocurrent generation was observed using an aqueous electrolyte solution containing viologen as the electron carrier. Figure 39 shows a plot of the light-harvesting effect by integrating the total absorbance in the whole range and the total current values at all the excitation wavelengths in the action spectrum. The sharp increase in both the total photocurrent and
Fig. 38 Photocurrent-action spectra normalized to a constant light intensity at 174.6 W cm–2 at – 200 mV (Ag/AgCl)
Porphyrin Supramolecules by Self-Complementary Coordination
99
Fig. 39 Plot of integrated total current values vs total absorbance for different array lengths. Numbers in the figure correspond to the number of the accumulation cycle
the current intensity emphasizes the efficient “light-harvesting” function of the surface-grafted multi-porphyrin arrays.
7 Concluding Remarks Self-complementary coordination is a powerful tool for supramolecular organization of various porphyrins. Molecules equipped with a set of complementary coordination (i.e., ligands and metalloporphyrins) spontaneously organize into their self-complementary structures, if they are appropriately designed. Due to this self-complementarity, the stability constants of these substances are far greater than those of the same ligand–metalloporphyrin combinations without complementarity. The structures must be rigid by the complementarity requirement. Therefore, this is the process of choice for designing specific arrangements of molecular systems with large stability constants. The smallest product is the coordination dimer, where the porphyrin π planes must be slip-stacked by the requirement of mutual selfcoordination. In other words, the dimer is essentially a J-type aggregate. This results in some of the key advantages of this coordination organization, since the fluorescence of the substance is not quenched and the absorption band is red-shifted. When the coordination angle is adjusted, larger rings or box-type organizations become possible. The employment of hexa-coordinating metal ions in the porphyrin center is a way to extend the supramolecular structure from the dimer to various
100
Y. Kobuke
macrocycles. In such cases, however, photonic application is limited due to the quenching of singlet excited states by these hexa-coordinating metal ions (mostly transition metal ions). An alternative way to obtain higher-level organized structures is through the connection of self-complementary sets. Various connection methods are associated with new organized structures. Direct connection of two zinc porphyrins at the meso–meso position gives rise to the formation of linear arrays, along which excitons, electrons, or holes can be transferred. Connection via acetylenic bonds allows the highly conjugated zinc porphyrin arrays to develop improved electronic conduction levels. The attachment of energy (or electron) donor or acceptor substituents provides materials for nonlinear optics by increasing the polarization levels of the material. metaPhenylene linkage of two porphyrins provides a method for obtaining artificial light-harvesting complexes. Not only two but also three or more porphyrins can be connected to yield other interesting organized structures. The methodology reviewed in this chapter will be further improved and developed to provide interesting products, not only in the field of scientific research, but also in the realm of technological applications.
References 1. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P (2001) Molecular biology of the cell. Garland, New York 2. Green BR, Parson WW (eds) (2003) Light-harvesting antennas in photosynthesis, vol 13. Kluwer, Dordrecht 3. Wasielewski MR (1992) Chem Rev 92:435 4. Gust D, Moore TA, Moore AL (2001) Acc Chem Res 34:40 5. Choi MS, Yamazaki T, Yamazaki I, Aida T (2004) Angew Chem Int Ed 43:150 6. Ogawa K, Kobuke Y (2003) Encyclopedia of nanoscience and nanotechnology X: 1. American Scientific, Los Angeles 7. Kobuke Y, Ogawa K (2003) Bull Chem Soc Jpn 76:689 8. Kobuke Y (2004) J Porphyrins Phthalocyanines 8:156 9. Satake A, Kobuke Y (2005) Tetrahedron 61:13 10. Wakao N, Yokoi N, Isoyama N, Hiraishi A, Shimada K, Kobayashi M, Kise H, Iwaki M, Ithoh S, Takaichi S, Sakurai Y (1996) Plant Cell Physiol 37:889 11. Kirksey CH, Hambright P, Storm CB (1969) Inorg Chem 8:2141 12. Deisenhofer J, Epp O, Miki K, Huber R, Michel H (1984) J Mol Biol 180:385 13. Yeates TO, Komiya H, Chirino A, Rees DC, Allen JP, Feher G (1988) Proc Natl Acad Sci USA 85:7993 14. Jordan P, Fromme P, Witt HT, Kukas O, Saenger W, Krauss N (2001) Nature 411:909 15. Zouni A, Witt HT, Kern J, Fromme P, Krauss N, Saenger W, Orth P (2001) Nature 409:739 16. McDermott G, Prince SM, Freer AA, Hawthornthwaite-Lawless AM, Papiz MZ, Cogdell RJ, Isaacs NW (1995) Nature 374:517 17. Karrasch S, Bullough PA, Ghosh R (1995) EMBO J 14:631 18. Roszak AW, Howard TD, Southall J, Gardiner AT, Law CJ, Isaacs NW, Cogdell RJ (2003) Science 302:1969
Porphyrin Supramolecules by Self-Complementary Coordination
101
19. Papiz MZ, Prince SM, Howard T, Cogdell RJ, Isaacs NW (2003) J Mol Biol 326:1523 20. Bahatyrova S, Frese RN, Siebert CA, Olsen JD, Van Der Werf KO, Van Grondelle R, Niederman RA, Bullough PA, Otto C, Hunter CN (2004) Nature 430:1058 21. Hess S, Akesson E, Cogdell RJ, Pullerits T, Sundström V (1995) Biophys J 69:2211 22. Shreve AP, Trautman JK, Frank HA, Owens TG, Albrecht AC (1991) Biochim Biophys Acta 1058:280 23. Jimenez R, Dikshit SN, Bradforth SE, Fleming GR (1996) J Phys Chem 100:6825 24. Hu X, Ritz T, Damjanovic A, Schulten K (1997) J Phys Chem B 101:3854 25. Hess S, Chachisvilis M, Timpmann K, Jones MR, Hunter CN, Sundström V (1995) Proc Natl Acad Sci USA 92:12333 26. Beekman LMP, van Mourik F, Jones MR, Visser HM, Hunter CN, van Grondelle R (1994) Biochemistry 33:3143 27. Liu Z, Yan H, Wang K, Kuang T, Zhang J, Gui L, An X, Chang W (2004) Nature 428:287 28. Boekema EJ, Hifney A, Yakushevska AE, Piotrowski M, Keegstra W, Berry S, Michel KP, Pistorius EK, Kruip J (2001) Nature 412:745 29. Fleming GR, Martin JL, Breton J (1988) Nature 333:191 30. Deisenhofer J, Norris JR (eds) (1993) The photosynthetic reaction center I and II. Academic, New York 31. Michel-Beyerle ME (eds) (1996) The reaction center of photosynthetic bacteria. Springer, Berlin Heidelberg New York 32. Chang CK, Abdalmuhdi I (1983) J Org Chem 48:5388 33. Chang CK, Liu HY, Abdalmuhdi I (1984) J Am Chem Soc 106:2725 34. Meier H, Kobuke Y, Kugimiya S (1989) J Chem Soc Chem Commun 923 35. Osuka A, Nakajima S, Nagata T, Murayama K, Toriumi K (1991) Angew Chem Int Ed Engl 30:582 36. Kobuke Y, Miyaji H (1994) J Am Chem Soc 116:4111 37. Stibrany RT, Vasudevan J, Knapp S, Potenza JA, Emge T, Schugar HJ (1996) J Am Chem Soc 118:3980 38. Gerasimchuk NN, Mokhir AA, Rodgers KR (1998) Inorg Chem 37:5641 39. Kasha M, Rawls HR, El-bayyoumi MA (1965) Pure Appl Chem 11:371 40. Kobayashi T (ed) (1996) J-Aggregates. World Scientific, Singapore 41. Tabata M, Nishimoto J (1999) In: Kadish K, Smith KM, Guilard R (eds) The porphyrin handbook, vol 9. Academic, New York, p 221 42. Hunter CA, Sarson LD (1994) Angew Chem Int Ed Engl 33:2313 43. Nakamura Y, Hwang IW, Aratani N, Ahn TK, Ko DM, Takagi A, Kawai T, Matsumoto T, Kim D, Osuka A (2005) J Am Chem Soc 127:236 44. Mackay LG, Wylie RS, Sanders JKM (1994) J Am Chem Soc 116:3141 45. Vidal-Ferran A, Clyde-Watson Z, Bampos N, Sanders JKM (1997) J Org Chem 62:240 46. Mongin O, Schuwey A, Vallot MA, Gossauer A (1999) Tetrahedron Lett 40:8347 47. Peng X, Aratani N, Takagi A, Matsumoto T, Kawai T, Hwang IW, Ahn TK, Kim D, Osuka A (2004) J Am Chem Soc 126:4468 48. Li J, Ambroise A, Yang SI, Diers JR, Seth J, Wack CR, Bocian DF, Holten D, Lindsey JS (1999) J Am Chem Soc 121:8927 49. Anderson HL, Sanders JKM (1990) Angew Chem Int Ed Engl 29:1400 50. Anderson S, Anderson HL, Sanders JKM (1995) J Chem Soc Perkin Trans 1 2247 51. Yu L, Lindsey JS (2001) J Org Chem 66:7402 52. Rucareanu S, Mongin O, Schuwey A, Hoyler N, Gossauer A (2001) J Org Chem 66:4973 53. Anderson S, Anderson HL, Sanders JKM (1993) Acc Chem Res 26:469
102
Y. Kobuke
54. 55. 56. 57. 58. 59. 60. 61.
Kim D, Osuka A (2004) Acc Chem Res 37:735 Holten D, Bocian DF, Lindsey JS (2002) Acc Chem Res 35:57 Takahashi R, Kobuke Y (2003) J Am Chem Soc 125:2372 Takahashi R, Kobuke Y (2005) J Org Chem 70:2745 Ohashi A, Satake A, Kobuke Y (2004) Bull Chem Soc Jpn 77:365 Ikeda C, Satake A, Kobuke Y (2003) Org Lett 5:4935 Chi X, Guerin AJ, Haycock RA, Hunter CA, Sarson LD (1995) Chem Commun 2567 Haycock RA, Hunter CA, James DA, Michelsen U, Sutton LR (2000) Org Lett 2: 2435 Tsuda A, Nakamura T, Sakamoto S, Yamaguchi K, Osuka A (2002) Angew Chem Int Ed 41:2817 Hwang IW, Kamada T, Ahn TK, Ko DM, Nakamura T, Tsuda A, Osuka A, Kim D (2004) J Am Chem Soc 126:16187 Tsuda A, Hu H, Watanabe R, Aida T (2003) J Porphyrins Phthalocyanines 7:388 Tsuda A, Hu H, Tanaka R, Aida T (2005) Angew Chem Int Ed 44:4884 Iengo E, Zangrando E, Geremia S, Graff R, Kieffer B, Alessio E (2002) Chem Eur J 8:4670 Balaban TS, Goddard R, Linke-Schaetzel M, Lehn JM (2003) J Am Chem Soc 125:4233 Tsuda A, Sakamoto S, Yamaguchi K, Aida T (2003) J Am Chem Soc 125:15722 Tanaka A, Ryuno A, Okada S, Satake A, Kobuke Y (2005) Isr J Chem 45:281 Ikeda C, Tanaka Y, Fujihara T, Ishii Y, Ushiyama T, Yamamoto K, Yoshioka N, Inoue H (2001) Inorg Chem 40:3395 Funatsu K, Imamura T, Ichimura A, Sasaki Y (1998) Inorg Chem 37:1798 Fukushima K, Funatsu K, Ichimura A, Sasaki Y, Suzuki M, Fujihara T, Tsuge K, Imamura T (2003) Inorg Chem 42:3187 Kuramochi Y, Satake A, Kobuke Y (2004) J Am Chem Soc 126:8668 Shoji O, Okada S, Satake A, Kobuke Y (2005) J Am Chem Soc 127:2201 Shoji O, Tanaka H, Kawai T, Kobuke Y (2005) J Am Chem Soc 127:8598 Tanaka H, Kawai T (1997) J Vac Sci Technol B 15:602 Tanaka H, Kawai T (2003) Surf Sci 539:L531 Aratani N, Osuka A, Kim YH, Jeong DH, Kim D (2000) Angew Chem Int Ed 39:1458 Aratani N, Takagi A, Yanagawa Y, Matsumoto T, Kawai T, Yoon ZS, Kim D, Osuka A (2005) Chem Eur J 11:3389 Tsuda A, Osuka A (2001) Science 293:79 Tsuda A, Nakamura Y, Osuka A (2003) Chem Commun 1096 Anderson HL (1994) Inorg Chem 33:972 Taylor PN, Anderson HL (1999) J Am Chem Soc 121:11538 Cho HS, Jeong DH, Cho S, Kim D, Matsuzaki Y, Tanaka K, Tsuda A, Osuka A (2002) J Am Chem Soc 124:14642 Yoon DH, Lee SB, Yoo K-H, Kim J, Lim JK, Aratani N, Tsuda A, Osuka A, Kim D (2003) J Am Chem Soc 125:11062 Kuebler SM, Denning RG, Anderson HL (2000) J Am Chem Soc 122:339 Screen TEO, Thorne JRG, Denning RG, Bucknall DG, Anderson HL (2002) J Am Chem Soc 124:9712 Ikeda C, Yoon ZS, Park M, Inoue H, Kim D, Osuka A (2005) J Am Chem Soc 127:534 Ogawa K, Kobuke Y (2000) Angew Chem Int Ed 39:4070 Nomoto A, Kobuke Y (2002) Chem Commun 1104 Nomoto A, Mitsuoka H, Ozeki H, Kobuke Y (2003) Chem Commun 1074 Ogawa K, Ohashi A, Kobuke Y, Kamada K, Ohta K (2003) J Am Chem Soc 125:13356
62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92.
Porphyrin Supramolecules by Self-Complementary Coordination 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112.
113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130.
103
Kobuke Y, Miyaji H (1996) Bull Chem Soc Jpn 69:3563 Nagata N, Kugimiya S, Fujiwara E, Kobuke Y (2003) New J Chem 743 Ikeda C, Satake A, Fujiwara E, Kobuke Y (2003) Chem Commun 616 Michelsen U, Hunter CA (2000) Angew Chem Int Ed 39:764 Morikawa S, Ikeda C, Ogawa K, Kobuke Y (2004) Lett Org Chem 1:6 Sugiura K, Tanaka H, Matsumoto T, Kawai T, Sakata Y (1999) Chem Lett 1193 McKeown NB (2002) Phthalocyanine materials: synthesis, structure and function. Cambridge University Press, Cambridge Marks TJ (1985) Science 227:881 Torre G, Vazquez P, Agullo-Lopez F, Torres T (2004) Chem Rev 104:3723 Ishii K, Abiko S, Fujitsuka M, Ito O, Kobayashi N (2002) J Chem Soc Dalton Trans 1735 Cook MJ, Jafari-Fini A (1997) J Mater Chem 7:2327 Kobayashi N, Muranaka A (2000) Chem Commun 1855 Kameyama K, Satake A, Kobuke Y (2004) Tetrahedron Lett 45:7617 Kameyama K, Morisue M, Satake A, Kobuke Y (2005) Angew Chem Int Ed 44:4763 Ozeki H, Nomoto A, Ogawa K, Kobuke Y, Murakami M, Hosoda K, Ohtani M, Nakashima S, Miyasaka H, Okada T (2004) Chem Eur J 10:6393 Marcus RA (1959) Can J Chem 37:155 Marcus RA (1965) J Chem Phys 43:679 Rehm D, Weller A (1969) Ber Bunsenges Phys Chem 73:834 Schmidt JA, Liu JY, Bolton JR, Archer MD, Gadzekpo VPY (1989) J Chem Soc Faraday Trans 1 85:1027 Hwang IW, Park M, Ahn TK, Yoon ZS, Ko DM, Kim D, Ito F, Ishibashi Y, Khan SR, Nagasawa Y, Miyasaka H, Ikeda C, Takahashi R, Ogawa K, Satake A, Kobuke Y (2005) Chem Eur J 11:3753 Morandeira A, Vauthey E, Schuwey A, Gossauer A (2004) J Phys Chem A 108:5741 Furutsu D, Satake A, Kobuke Y (2005) Inorg Chem 44:4460 Prasad PN, Williams DJ (eds) (1991) Introduction to nonlinear optical effects in molecules and polymers. Wiley, New York Zyss J (ed) (1994) Molecular nonlinear optics: materials, physics, and devices. Academic, Boston Bosshard C, Sutter K, Prêtre P, Hulliger J, Flörsheimer M, Kaatz P, Günter P (1995) Organic nonlinear optical materials. Gordon and Breach, Amsterdam Ogawa K, Kobuke Y (2002) J Am Chem Soc 124:22 Zhan X, Liu Y, Zhu D, Huang W, Gong Q (2001) Chem Mater 13:1540 Sheik-Bahae M, Said AA, Wei TH, Hagan DG, van Stryland EW (1990) IEEE J Quantum Electron 26:760 Drobizhev M, Karotki A, Kruk M, Rebane A (2002) Chem Phys Lett 355:175 Drobizhev M, Karotki A, Kruk M, Mamardashvili NZ, Rebane A (2002) Chem Phys Lett 361:504 Drobizhev M, Karotki A, Kruk M, Krivokapic A, Anderson HL, Rebane A (2003) Chem Phys Lett 370:690 Goyan RL, Cramb DT (2000) Photochem Photobiol 72:821 Kamada K, Ohta K, Iwase Y, Kondo K (2003) Chem Phys Lett 372:386 Iwase Y, Kondo K, Kamada K, Ohta K (2003) J Mater Chem 13:1573 Audebert P, Kamada K, Matsunaga K, Ohta K (2003) Chem Phys Lett 367:62 Antonov L, Kamada K, Ohta K, Kamounah S (2003) Phys Chem Chem Phys 5:1193 Kamada K, Matsunaga K, Yoshino A, Ohta K (2003) J Opt Soc Am B 20:529 Abbotto A, Beverina L, Bozio R, Facchetti A, Ferrante C, Pagani GA, Pedron D, Signorini R (2002) Org Lett 4:1495
104
Y. Kobuke
131. Kim OK, Lee KS, Woo HY, Kim KS, He GS, Swiatkiewicz J, Prasad PN (2001) Chem Mater 13:4071 132. Chung SJ, Kim KS, Lin TC, He GS, Swiatkiewicz J, Prasad PN (1999) J Phys Chem B 103:10741 133. Abbotto A, Beverina L, Bozio R, Facchetti A, Ferrante C, Pagani GA, Pedronb D, Signorini R (2003) Chem Commun 2144 134. Belfield KD, Liu Y, Negres RA, Fan M, Pan G, Hagan DJ, Hernandez FE (2002) Chem Mater 14:3663 135. Porres L, Mongin O, Katan C, Charlot M, Pons T, Mertz J, Blanchard-Desce M (2004) Org Lett 6:47 136. Ogawa K, Ohashi A, Kobuke Y, Kamada K, Ohta K (2005) J Phys Chem B 109:22003 137. Nazeeruddin MK, Kay A, Rodicio I, Humphry-Baker R, Müller E, Liska P, Vlachopoulos N, Grätzel M (1993) J Am Chem Soc 115:6382 138. Hagfeldt A, Grätzel M (1995) Chem Rev 95:49 139. Stuart L (2002) Semiconductor electrodes and photoelectrochemistry. In: Bard AJ, Stratmann M (eds) Encyclopedia of electrochemistry, vol 6. Wiley, Weinheim 140. Morisue M, Yamatsu S, Haruta N, Kobuke Y (2005) Chem Eur J 11:5563
Struct Bond (2006) 121: 105–143 DOI 10.1007/430_025 © Springer-Verlag Berlin Heidelberg 2006 Published online: 1 March 2006
Metal-Mediated Multi-Porphyrin Discrete Assemblies and Their Photoinduced Properties Elisabetta Iengo1 · Franco Scandola2 · Enzo Alessio1 (u) 1 Dipartimento
di Scienze Chimiche, Università di Trieste, Via L. Giorgieri 1, 34127 Trieste, Italy [email protected]
2 Dipartimento
di Chimica, Università di Ferrara, 44100 Ferrara, Italy
1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
Design Strategy: Multi-Porphyrin Assemblies Connected by Exocyclic Coordination to Metal Fragments . . . . . . . Multi-Porphyrin Adducts Assembled via one External Metal Fragment 2D and 3D Multi-Porphyrin Adducts Assembled via two or more External Metal Fragments . . . . . . . . . . Ruthenium-Mediated Assemblies . . . . . . . . . . . . . . . . . . . . . .
106
. . . .
107 110
. . . .
115 118
Higher Order Assemblies of Porphyrins through a Hierarchical Metal-Driven Synthetic Approach . . . . . . . . . .
122
4 4.1 4.2
Photophysical Properties of Metal-Connected Assemblies of Porphyrins . Metallacycles of Porphyrins and their Model Compounds . . . . . . . . . . Higher-Order Assemblies . . . . . . . . . . . . . . . . . . . . . . . . . . . .
126 129 135
5
Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . .
139
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
141
2.1 2.2 2.3 3
Abstract The metal-driven construction of multi-porphyrin assemblies, which exploits the formation of coordination bonds between exocyclic donor site(s) on the porphyrins and metal centers, has recently allowed the design and preparation of sophisticated supramolecular architectures whose complexity and function begin to approach the properties of naturally occurring systems. Within this framework, meso-pyridyl/phenyl porphyrins (PyPs), or strictly related chromophores, can provide geometrically well-defined connections to as many as four metal centers by coordination of the pyridyl peripheral groups. Several discrete assemblies of various nuclearities, in which the pyridylporphyrins are connected through external coordination compounds, have been constructed in recent years. In this review, we summarize recent work in this field from our and other laboratories. The photophysical properties of some ruthenium-mediated assemblies of porphyrins prepared by our group are also described. Keywords Multi-porphyrin assemblies · Metal-mediated assemblies · Ruthenium-dmso coordination compounds · Pyridylporphyrins · Photo-induced processes
106 Abbreviations 4 MPyP 4 -cisDPyP 4 -transDPyP 4 -transDPyP-npm 4 TPyP TPP dppp en OTf BINAP
E. Iengo et al. 5-(4 pyridyl)-10,15,20-triphenylporphyrin 5,10-bis(4 pyridyl)-15,20-diphenylporphyrin 5,15-bis(4 pyridyl)-10,20-diphenylporphyrin 5,15-bis(4 -pyridyl)-2,8,12,18-tetranormalpropyl-3,7,13,17-tetramethylporphyrin 5,10,15,20-tetra(4 pyridyl)porphyrin tetraphenylporphyrin bis(diphenylphosphino)propane ethylenediamine triflate, CF3 SO3 – 2,2 -bis(diphenylphosphino)-1,1 -binaphtyl
1 Introduction A major challenge in molecular photonics is to develop synthetic architectures of nanoscopic dimensions for the absorption of light and the subsequent manipulation of the captured excited state energy, i.e. lightharvesting antenna systems, artificial reaction centers, light-driven molecular machines, molecular switches and optical memories, fluorescent sensors, etc. Porphyrins and/or metalloporphyrins are ideal components for photoactive assemblies because they offer a variety of appealing features such as: a rigid, planar geometry; high stability; an intense electronic absorption band in the visible region; a relatively long fluorescence decay time; a flexible tunability of their optical and redox properties by appropriate metalation/functionalization. Synthetic multiporphyrin assemblies are investigated as models of naturally occurring multichromophore aggregates [1–11], such as those found in the photosynthetic reaction centers and in the light harvesting complexes of photosynthetic organisms [12–34] and also in materials science and nanotechnology (e.g. molecular-scale electronics, optoelectronic devices, sensors) [35–38]. Supramolecular synthetic strategies provide a “bottom-up” approach to this challenging field. In particular, the metal-driven assembly is an efficient, highly convergent synthetic strategy that is used, often in combination with other non-covalent interactions and/or with more conventional covalent procedures, for constructing robust and shape-persistent assemblies of nanoscopic dimensions in which the number, position, relative orientation and distance of the chromophores are well defined [39–41]. The use of metalloporphyrins for the construction of multichromophore assemblies through axial binding to polytopic ligands and/or to other porphyrins is treated in detail in other chapters of this book. Here we report on the exocyclic coordination of porphyrins, functionalized with appropriate ligands appended at the β-pyrrolic and/or meso
Metal-Mediated Multi-Porphyrin Discrete Assemblies
107
positions, to external metal centers as a convenient and extremely versatile synthetic strategy towards supramolecular architectures that conjugate increasing complexity with good thermodynamic and kinetic stability. Some examples of higher order assemblies, in which the coordination capabilities of such chromophores are exploited both towards metal complexes and metalloporphyrins, will also be described. The artificial multi-chromophore assemblies prepared in our group have been investigated for their photophysical properties, in particular with respect to mimicking the two important light-induced functions of natural photosynthesis: photoinduced energy transfer (i.e. light harvesting—antenna effect) and photoinduced electron transfer (i.e. charge separation).
2 Design Strategy: Multi-Porphyrin Assemblies Connected by Exocyclic Coordination to Metal Fragments A large and diverse number of transition-metal coordination numbers and geometries can be used in the construction of coordination-driven assemblies, giving access to different topologies rather difficult to obtain with the classical synthetic methods. The synthetic process is under thermodynamic control when relatively labile metal centers are used (true supramolecular self-assembly), while it is under kinetic control with inert metals (unless high temperatures are used). The metal centers can be naked ions or, more conveniently, coordination compounds with some binding sites occupied by labile, readily replaceable ligands. With naked ions the number and geometry of the coordination sites used for assembly purposes are dictated by the nature and oxidation state of the metal cation exclusively. Conversely, coordination complexes (or organometallic compounds) offer a variety of different binding geometries which depend on the nature both of the metal center and of the non-labile ligands. In addition, a judicious choice of the ancillary ligands allows fine-tuning of the charge and polarity (and thus solubility) of the final adduct and to introduce additional functionalities (e.g. chirality). Finally, also the metallic connectors, if suitable coordination complexes are appropriately chosen, may introduce useful photophysical properties into the assemblies. The metal-mediated synthetic approach has recently provided many examples of large discrete assemblies of porphyrins [39–41] and of solid-state arrays as well [39–48]. For the purpose of metal-driven synthesis, porphyrins and metal centers are to be considered essentially as rigid modules. Thus, in principle, the final shape, dimension, and topology of the self-assembled architecture will be defined by the number and relative geometries of the coordination sites on the building blocks. Within this framework, mesopyridylporphyrins (PyPs, Fig. 1) are largely exploited as they can provide
108
E. Iengo et al.
Fig. 1 Schematic drawing of 4 meso-pyridylporphyrins (4 PyPs) with numbering scheme. Grey circles symbolize exocyclic metal centers
geometrically well-defined connections to as many as four metal centers by coordination of the meso pyridyl groups [39–41, 49–73]. The remaining meso and β positions can be differently substituted with aryl groups or alkyl chains, depending on the type of condensation reaction employed for the preparation of the chromophore. The choice of these residues allows us to tune the solubility and the physical properties of the porphyrin units, and thus of the final adducts. The peripheral N atom(s) of PyPs can be either in the 4 or in the 3 position. With 4 PyPs the exocyclic coordination bonds are established in the plane of the porphyrin, along the meso bond axes; with 3 PyPs instead, since the meso pyridyl rings are tilted, the coordination bonds are directed out of the plane of the chromophore (Fig. 2)1 . The coordination of PyPs to metal fragments can be performed using them either as free-bases (i.e. with no metal inserted in the internal core) or as metallated units. Alternatively, inner metal centers can be introduced in a second step, after external coordination, depending on the design rationale. The presence of one or more metalloporphyrin units in the final assembly, in addition to introducing novel reactive functionalities (i.e. the inner metal ions), can vary dramatically its photophysical properties. The most widely used external metal connectors are cis or, more seldom, trans bis-acceptor complexes, which can be defined as 90◦ angular or linear fragments, respectively. The angular connectors employed 1
The 2 position cannot be easily exploited for coordination for steric reasons.
Metal-Mediated Multi-Porphyrin Discrete Assemblies
109
Fig. 2 Different orientations of the exocyclic coordination bonds in 4 - and 3 -PyPs
Fig. 3 Schematic drawing of the square planar and octahedral metal fragments (and of their precursors) most commonly used in the metal-driven construction of multiporphyrin assemblies
so far are relatively few (Fig. 3): dicationic fragments are obtained from square planar Pd(II) or Pt(II) complexes with two labile anionic ligands, such as [M(dppp)(OTf)2 ] [51–54, 63, 64] [M(BINAP)(OTf)2 ] [53, 64] or [M(en)(ONO2 )2 ] (M = Pd, Pt) [62], while neutral fragments are generated by precursors with neutral leaving ligands, such as cis-[PtCl2 (dmso)2 ] [51],
110
E. Iengo et al.
cis-[PtCl2 (PhCN)2 ] [50], [ReX(CO)5 ] (X = Cl, Br) [55–57], trans-[RuCl2 (dmso-S)4 ] [65–73], and trans,cis,cis-[RuCl2 (dmso-O)2 (CO)2 ] [65–73]. The number of linear precursors employed is even more limited, being restricted to trans-[PdCl2 (dmso)2 ] [51, 60, 61], trans-[PdCl2 (PhCN)2 ] [50, 58, 59, 67], and 1,4-bis-[trans-Pt(PEt3 )2 (OTf)]benzene [53]. [M(CH3 CN)4 ][Y]2 (M = Pd, Pt; Y = OTf, BF4 ) and cis-[RuX2 (dmso)4 ] (X = Cl, Br) have been used as cruciform fragments since they can provide four coplanar coordination sites [51, 59]. The assemblies mediated by the neutral 90◦ angular Re(I) fragment fac[ReX(CO)3 ] are treated in detail in the contribution by J. Hupp (in this volume) and will be mentioned only briefly here. 2.1 Multi-Porphyrin Adducts Assembled via one External Metal Fragment Linear neutral bisporphyrin adducts were obtained by reaction of a two-fold excess of 4 mono-pyridyl porphyrins with trans-[PdCl2 (dmso)2 ] or trans[PdCl2 (PhCN)2 ]. The corresponding angular compounds were similarly prepared using cis-[PtCl2 (dmso)2 ], cis-[PtCl2 (PhCN)2 ], or cis-[M(dppp)(OTf)2 ] (M = Pd, Pt) [49–51, 58, 59]. The adducts in which the two mono-pyridyl porphyrins bear long 4-hexadecyloxyl chains at the para positions of the meso phenyl groups were also assembled into stable monolayers on a water surface [59]. A kinked bisporphyrin palladium complex was obtained by Reed, Boyd, and co-workers by coordination of two 3 MPyP units to the trans-PdCl2 fragment [60]. This jaw-like cleft was proved to be an efficient host for fullerene (Fig. 4) and the host-guest complex was also structurally characterized in the solid state by X-ray crystallography [61]. Tetraporphyrin systems of the type [M(4 MPyP)4 ][Y]2 (M = Pd, Pt; Y = OTf, BF4 ) [51], and trans-[RuCl2 (Pd · 4 MPyP)4 ] [59] were isolated by reaction of 4 monopyridyl porphyrins (either free-base or metallated) with the charged square planar complexes [M(CH3 CN)4 ][Y]2 or the neutral octahedral complexes cis-[RuX2 (dmso-S)4 ], respectively [59, 74]2 (Fig. 5). In the first case the reaction occurred quantitatively upon four-fold addition of the chromophore at room temperature, whereas for the preparation of the neutral tetraporphyrin adducts a large excess of chromophore, heating, and column purification were needed. We mainly exploited the coordination ability of the two Ru(II) complexes, trans-[RuCl2 (dmso-S)4 ] (1) and trans,cis,cis-[RuCl2 (dmso-O)2 (CO)2 ] (2), that behave as neutral cis bis-acceptor fragments upon selective replacement, 2
cis-[RuX2 (dmso-S)4 ] (X = Cl, Br) complexes are known to isomerize to the trans geometry when treated with excess monodentate nitrogen ligands (e.g. pyridine, py), thus leading to trans[RuX2 (py)4 ] [74].
Metal-Mediated Multi-Porphyrin Discrete Assemblies
111
Fig. 4 Schematic drawing of the inclusion adduct of fullerene inside the jaw-like bisporphyrin Pd complex (see [60])
Fig. 5 Schematic representations of the cruciform tetraporphyrin adducts [M(4 MPyP)4 ] [Y]2 (M = Pd, Pt; Y = OTf, BF4 ), and trans-[RuCl2 (Pd·4 MPyP)4 ] (see [51] and [59])
112
E. Iengo et al.
under mild conditions, of two adjacent dmso ligands [75, 76]. Thus, treatment of 1 or 2 with a slight excess of 4 MPyP (chloroform, room temperature), yielded the corresponding disubstituted complexes trans,cis,cis[RuCl2 (dmso-S)2 (4 MPyP)2 ] (3) and trans,cis,cis-[RuCl2 (CO)2 (4 MPyP)2 ] (4), respectively (Fig. 6), which were unambiguously characterized in solution by NMR spectroscopy [65, 72, 73]. Similarly, treatment of 1 or 2 with excess 4 -cisDPyP under mild conditions led, after chromatographic purification, to the isolation of the corresponding bisporphyrin mono-nuclear compounds trans,cis,cis-[RuCl2 (X)2 (4 -cisDPyP)2 ] (Scheme 1, X = dmso-S, 5; X = CO, 6) [66, 72, 73]. Compounds 5 and 6 have one residual unbound 4 N(py) ring on each of the two cis coordinated 4 -cisDPyP units and are thus examples of metal-containing ligands, capable of chelating suitable cis bis-acceptor metal fragments (see below). More recently we prepared the cis disubstituted ruthenium-nitrosyl complex mer-[RuCl3 (NO)(4 MPyP)2 ] (7) by treatment of the anionic precursor [nBu4 N]trans-[RuCl4 (NO)(dmso-O)] with a slight excess of monopyridylporphyrin in refluxing chloroform [70]. Compound 7 represents the first example of an octahedral coordination compound with two adjacent porphyrins that has been structurally characterized in the solid state by X-ray crystallography (Fig. 7). The geometrical parameters of this structure show significant differences as compared to those reported by Woo and co-workers for the square-planar bisporphyrin complex cis-[Pd(dppp)(4 MPyP)2 ](OTf)2 (see above) (Fig. 7) [51]. In the neutral ruthenium complex (7) the porphyrin
Fig. 6 Schematic drawing of trans,cis,cis-[RuCl2 (dmso-S)2 (4 MPyP)2 ] (3) and trans,cis,cis[RuCl2 (CO)2 (4 MPyP)2 ] (4)
Metal-Mediated Multi-Porphyrin Discrete Assemblies
113
Scheme 1
A, the angle defined by the metal cencentroid-to-centroid distance is 14.9 ˚ ter and the two porphyrin centroids is 98.7◦ , and that between the porphyrin mean planes is 83.78◦ . The corresponding parameters for the charged diphosA, 56.1◦ , and 20.4◦ , respectively. These phine palladium complex are 9.1 ˚ marked differences might be ascribed to the different nature of the ancillary ligands, modestly hindering in the mer-RuCl3 (NO) fragment, and quite sterically demanding in the cis-Pd(dppp)2+ one. Also, specific interactions between the phenyl rings of the dppp bridge and the porphyrin pyridyl groups could play a role in determining the solid state arrangement of the porphyrins in the Pd complex (see discussion below).
114
E. Iengo et al.
Fig. 7 Comparison of the molecular structures of the two bisporphyrin complexes fac-[RuCl3 (NO)(4 MPyP)2 ] (7, adapted from [70]) and cis-[Pd(dppp)(4 MPyP)2 ](OTf)2 (adapted from [51], triflate anions omitted). The orientations are such that in both cases one porphyrin is perpendicular to the viewer
It should be noted here that information derived from solid state data of discrete metal-mediated multi-porphyrins arrays is crucial for a better understanding of the preferred geometry of the system, the flexibility of the building blocks, the inter- and intra-molecular interactions, and thus for assessing their potentialities for further exploitation (e.g., host-guest interactions, higher order assemblies). A general common feature for all the ruthenium-mediated multi-porphyrin systems described above is that the newly formed Ru – N(pyridyl) bonds are both stable and inert and, as a consequence, the corresponding adducts are very robust. In the 1 H NMR spectra of the cis disubstituted 4 PyP species 4–7 we identified a signature that is quite characteristic of two mutually cis-coordinated pyridylporphyrins in free rotation about the metal–pyridyl axis [65, 72, 73]. Coordination to the ruthenium fragment induces downfield shifts for the H2,6 and H3,5 resonances of the 4 N(py) ring involved in the new bond (∆δ H2,6 from 0.3 to 0.9 ppm, ∆δ H3,5 from 0.03 to 0.18 ppm, depending on the compound. See Fig. 1 for numbering scheme). Conversely, the resonances of the adjacent pyridyl and/or phenyl rings experience an upfield shift of ca. 0.2 ppm. This is caused by the rotation of the porphyrin about the Ru–pyridyl axis, which brings the six-membered rings at the 10 and 20 positions into the shielding cone of the adjacent porphyrin. The relatively small upfield shift
Metal-Mediated Multi-Porphyrin Discrete Assemblies
115
suggests that the two porphyrins rotate freely and that the average time spent in the orientations inducing mutual shielding is rather short. 2.2 2D and 3D Multi-Porphyrin Adducts Assembled via two or more External Metal Fragments Bidimensional metallacycles of porphyrins, also called molecular squares, were first obtained by treatment of meso-bis(4 pyridyl)porphyrins (i.e. 4 cisDPyP and 4 -transDPyP) with the appropriate linear or 90◦ -angular metal fragments. Reaction of 4 -cisDPyP with a linear connector such as trans[PdCl2 (PhCN)2 ] led to the 4 + 4 metallacycle in which the trans-[PdCl2 ] units are the sides and the porphyrins are the corners of the square (Fig. 8) [49, 59]. Conversely, reaction of linear 4 -transDPyP units with 90◦ -angular connectors, such as cis-[PtCl2 (PhCN)2 ] [49], [ReX(CO)5 ] [55, 57], and [M(dppp) (OTf)2 ] (M = Pd, Pt) [52, 53], yielded the corresponding 4 + 4 metallacycles in which the metal fragments define the corners and the porphyrins the sides of the molecular square, and are actually oriented (as an average) perpendicularly to its plane (Fig. 9). Smaller 2 + 2 metallacycles were obtained by reaction of 4 -cisDPyP with ◦ 90 -angular metal fragments, such as [Pd(dppp)]2+ (Fig. 10) [49, 53–57]. Interestingly, the use of optically pure metal connectors of the type [Pd(S(–)-BINAP)]2+ or [Pd(R(+)-BINAP)]2+ was found to promote the selective formation of chiral porphyrin metallacycles [53]. With a similar synthetic approach, Drain and co-workers prepared also more elaborate architectures, such as linear trimers, tapes and tessellated
Fig. 8 Example of a 4 + 4 metallacycle (molecular square) of porphyrins obtained by reaction of the angular 4 -cisDPyP unit with the linear connector trans-[PdCl2 ] (see [49] and [59])
116
E. Iengo et al.
Fig. 9 Example of a 4 + 4 metallacycle (molecular square) of porphyrins obtained by reaction of the linear 4 -transDPyP unit with the 90◦ angular connector cis-[PtCl2 ] (see [49])
Fig. 10 Example of a 2 + 2 metallacycle of porphyrins obtained by reaction of the angular 4 -cisDPyP unit with the 90◦ angular connector [Pd(dppp)]2+ (see [53] and [54])
arrays of porphyrins by one-pot self-assembly of different metal precursors and/or porphyrin building blocks [50]. For example, the molecular tape shown in Fig. 11 was quantitatively obtained by mixing a cis-PtCl2 precursors with 4 TPyP and 4 -cisDPyP units in the appropriate 3 : 1 : 1 ratio [50]. By treatment of the tetratopic 3 -tetrapyridyl porphyrin (in place of the 4 analogue) with a two-fold excess of the angular precursor [Pd(en)(ONO2 )] in refluxing CH3 CN/H2 O mixtures, the group of Fujita obtained in quantitative yield the charged tridimensional prism of porphyrins [{Pd(en)}6 (Zn· 3 TPyP)3 ](NO3 )12 (Fig. 12), that was also characterized in the solid state [62].
Metal-Mediated Multi-Porphyrin Discrete Assemblies
117
Fig. 11 Metal-mediated molecular tape of porphyrins obtained by one-pot reaction of cis-PtCl2 precursors with 4 TPyP and 4 -cisDPyP units in 3 : 1 : 1 ratio (see [50])
Fig. 12 Schematic drawing of the charged tridimensional prism of porphyrins [{Pd(en)}6 (Zn·3 TPyP)3 ]12+ prepared by Fujita and co-workers (see [62])
Shinkai and co-workers reported on the formation of a charged bisporphyrin tridimensional capsule by coordination of two equal porphyrins, bearing four extended 4 -pyridyl arms each, to four 90◦ -angular Pd(dppp)2+ connectors (Fig. 13) [63]. Such assembly, after metalation of the porphyrin cores with Zn(II) atoms, proved capable of encapsulating bypiridyl guests, via nitrogen-zinc axial coordination. Furthermore, the corresponding chiral capsule was obtained when the optically pure S(–)- or R(+)-BINAP bisphosphine was introduced as ancillary ligand on the Pd fragments. The chirality of the final adduct was demonstrated to result from the helical twisting of the porphyrin units induced by the chiral non-participating ligands [64].
118
E. Iengo et al.
Fig. 13 Schematic drawing of the charged Pd-mediated bisporphyrin capsule described by Shinkai and co-workers (see [63])
2.3 Ruthenium-Mediated Assemblies We found that the reaction of the cis bis-coordinating ruthenium precursors trans-[RuCl2 (dmso-S)4 ] (1) and trans,cis,cis-[RuCl2 (dmso-O)2 (CO)2 ] (2) with an equimolar amount of the angular porphyrin building block 4 cisDPyP leads to the formation of the neutral 2 + 2 metallacycles of formula [trans,cis,cis-RuCl2 (X)2 (4 -cisDPyP)]2 (X = dmso-S, 8; X = CO, 9, Scheme 2) [68, 72, 73]3 . The NMR spectra of 8 and 9 unambiguously established the metallacyclic nature and high symmetry of these species. A recent X-ray structural determination performed on the zincated derivative of 9, 9Zn, showed that in the solid state the metallacycle (C2 symmetry, with the two-fold axis passing through the Ru atoms) is flat, with an almost perfect co-planar arrangement of the two porphyrins (Fig. 14). The Ru · · · Ru and the Zn · · · Zn distances are As a minor product of the reaction between 2 and 4 -cisDPyP we also isolated in pure form, after chromatographic workup, the trinuclear metallacycle [RuCl2 (CO)2 (4 -cisDPyP)]3 , in which the two porphyrins coordinated to the same trans,cis-RuCl2 (CO)2 fragment are not coplanar [73]. 3
Metal-Mediated Multi-Porphyrin Discrete Assemblies
119
Scheme 2
Fig. 14 Solid state structure of the neutral 2 + 2 metallacycle [trans,cis,cis-RuCl2 (CO)2 (Zn· 4 -cisDPyP)(EtOH)]2 (9Zn) (adapted from [71]). Top: front view; bottom: side view along the Ru· · ·Ru axis, evidentiating the almost perfect coplanar geometry
120
E. Iengo et al.
A, respectively; metal-to-porphyrin center distances 14.009(3) and 14.028(3) ˚ A [71]. (sides of the metallacycle) are about 9.9 ˚ Similarly, the reaction of precursor 2 with an equimolar amount of 3 -cisDPyP, rather than 4 -cisDPyP, yielded the corresponding neutral 2 + 2 metallacycle [trans,cis,cis-RuCl2 (CO)2 (3 -cisDPyP)]2 (10) [71]. NMR spectroscopy provided unambiguous evidence that only one highly symmetrical metallacycle, in which the two chromophores are held in a slipped cofacial arrangement by the external Ru(II) metal fragments, exists in solution (Fig. 15). We had no evidence of the formation of other possible conformers of 10, which might derive from different orientations of the 3 N(py) rings. The unprecedented staggered geometry of 10, with an interplanar distance A and a lateral offset (center-to-center between the two porphyrins of 4.18 ˚ A, was confirmed in the solid state by X-ray structural indistance) of 9.819 ˚ vestigation (Fig. 16). The geometry of 10 is reminiscent of those of the special pair of bacteriophylls in the reaction centers and of adjacent B850 units in the LH2 light-harvesting antenna systems of photosynthetic bacteria. Thus, as anticipated (Fig. 2), the change in the position of the N atom in the peripheral pyridyl rings of cisDPyP from 4 to 3 led indeed, upon coordination to the same cis bifunctional ruthenium fragment trans,cis-[RuCl2 (CO)2 ], to porphyrin cyclic assemblies with the same nuclearity but with very different geometries: from a flat two-dimensional assembly (9) to a staggered three-dimensional structure with a rigid spatial arrangement of the two chromophores (10) [71].
Fig. 15 Schematic representation of the slipped-cofacial porphyrin metallacycle [trans,cis, cis-RuCl2 (CO)2 (3 -cisDPyP)]2 (10) and of the corresponding fully- and semi-zincated derivatives
Metal-Mediated Multi-Porphyrin Discrete Assemblies
121
Fig. 16 Molecular structure of [trans,cis,cis-RuCl2 (CO)2 (3 -cisDPyP)]2 (10); side-view along the Ru· · ·Ru axis, showing the staggered geometry (adapted from [71])
Compared to the other neutral 2 + 2 metallacycles of porphyrins that feature octahedral fac-[ReX(CO)3 ] corner units, compounds 8–10 have the advantage of being more symmetrical: the two axial ligands of the Re(I) corners are different (one halide and one CO), thus generating mixtures of isomers when more units are present in the same assembly. In addition, reactions leading to metallacycles are performed at room temperature with the Ru-dmso precursors 1 and 2, whereas prolonged heating (typically refluxing THF/toluene) is required for the [ReX(CO)5 ] precursors. On the other hand, while the rheniumporphyrin metallacycles are thermodynamic products and are usually obtained with high yields and purities, the ruthenium cyclic assemblies are kinetic products, formed in lower yields, and often needing column purification. Nevertheless, under kinetic conditions, cyclic adducts with higher nuclearity3 , as well as products with different stoichiometry may become accessible, for example the reactive bisporphyrin complexes trans,cis,cis-[RuCl2 (dmso-S)2 (4 cisDPyP)2 ] (5) and trans,cis,cis-[RuCl2 (CO)2 (4 -cisDPyP)2 ] (6). Indeed, the hetero-bimetallic 2 + 2 metallacycles of porphyrins [Pd(dppp){trans,cis,cisRuCl2 (X)2 (4 -cisDPyP)2 }](OTf)2 (X = dmso-S, 11; X = CO, 12), featuring one neutral octahedral Ru(II) and one dicationic square-planar Pd(II) fragment at opposite corners, were quantitatively assembled by titration of the cis bisacceptor metal fragment [Pd(dppp)(OTf)2 ] into chloroform solutions of either 5 or 6 (Scheme 1) [66, 72, 73]. According to the X-ray structure (Fig. 17), the metric parameters of 12 are very similar to those found in the corresponding homometallic ruthenium metallacycle 9Zn (see above) [71]. However, while in the solid state 9 is almost perfectly flat, 12 exhibits a butterfly conformation, similar to that found by Stang and co-workers in the corresponding Pd(II) macrocycle [Pd(dppp)(4 -cisDPyP)]2 (OTf)4 [54] (see Fig. 10), with porphyrin least-squares planes that form a dihedral angle of ca. 138◦ , approximating a Cs symmetry. Optimization of the stacking interactions between the phenyl rings of the diphosphine bridge and the pyridyl rings of the porphyrins is very likely the driving force leading to this distortion in the solid state.
122
E. Iengo et al.
Fig. 17 Solid state structure of the cationic hetero-bimetallic 2 + 2 metallacycle of porphyrins [Pd(dppp){trans,cis,cis-RuCl2 (CO)2 (4 -cisDPyP)2 }]2+ (12, adapted from [72, 73])
3 Higher Order Assemblies of Porphyrins through a Hierarchical Metal-Driven Synthetic Approach As anticipated already, metalation of the ruthenium-mediated multiporphyrin adducts described above introduces new sites for further axial coordination and basically makes them suitable for being used as building blocks in the hierarchical assembly of higher order adducts. Thus, insertion of Zn(II) into the two porphyrins of 6 yielded trans,cis,cis[RuCl2 (CO)2 (Zn·4 -cisDPyP)2 ] (6Zn), which is a novel type of metal-containing building block featuring two donor (the uncoordinated 4 N(py) atoms) and two acceptor (the Zn atoms) sites, whose relative orientation depends on the torsion angles about the Ru – N bonds. Owing to the many degrees of rotational freedom of this compound, it is difficult to predict the nature and geometry of the self-coordination product that, in principle, might be either discrete or polymeric. 1 H NMR spectroscopy and single crystal X-ray analysis established that 6Zn self-assembles in chloroform solution to yield selectively the stable dinuclear species ([6Zn]2 ), in which the four 4 N(py) sites and the four Zn ions of two 6Zn units are mutually saturated through inter-molecular axial coordination (Fig. 18) [69, 72, 73]. The adduct [6Zn]2 , that features four porphyrins and six metal atoms (two Ru and four Zn), is an unprecedented example of a discrete and highly ordered array of porphyrins obtained by two-point metal-mediated self-assembly.
Metal-Mediated Multi-Porphyrin Discrete Assemblies
123
Fig. 18 Perspective view of the solid-state molecular structure of [trans,cis,cisRuCl2 (CO)2 (Zn·4 -cisDPyP)2 ]2 ([6Zn]2 ); for clarity reasons the two 6Zn units have different shades (adapted from [69])
Treatment of the metallacycle 9 with zinc acetate in chloroform/methanol mixtures led, depending on the Zn/Ru ratio, to the isolation in pure form of the corresponding fully zincated adduct 9Zn or of the semi-zincated metallacycle 9Zn1/2 , in which only one of the two chromophores has an inner zinc ion (see below) [71]. Owing to the thermodynamic and kinetic stability of the Ru-pyridyl bonds, solutions of 9Zn can be treated with excess of other N-ligands without the occurrence of decomposition or scrambling processes. Thus, 9Zn is a rigid 2D module with two central junctions that can be exploited for the construction of more elaborate supramolecular adducts upon treatment with appropriate polytopic ligands. As unambiguously evidenced by NMR spectroscopy, titration of 9Zn in CDCl3 solution with one equivalent of a linear ditopic N-ligand leads rapidly to the quantitative assembling of sandwich-like 2 : 2 supramolecular adducts of formula [(9Zn)2 (µ-L)2 ] (L = 4,4 -bipy, 13; L = 4 -transDPyP, 14; L = 4 -transDPyP-npm, 15), formed by two parallel metallacycles connected by two bridging ligands which are axially bound to the zinc-porphyrins (Scheme 3) [68, 72, 73]. Formation constants higher than 1018 M–3 were estimated in chloroform for compounds 13–15 and the NMR spectra of CDCl3 solutions, diluted to the limit of detection, showed the resonances of the intact assemblies
124
E. Iengo et al.
Scheme 3
exclusively4 . The high stability of the adducts was attributed to cooperativity in the stepwise coordination of the two bridging ligands and to entropy considerations: owing to the rigidity of the fragments, formation of the sandwich molecules requires limited conformational changes. Compounds 14 and 15, which feature six porphyrins each, might be better defined as multiporphyrin molecular boxes. Compound 14 was also characterized in the solid state by single-crystal X-ray analysis (Fig. 19). A. The The distance between the two opposite metallacycles is ca. 19.5 ˚ two bridging 4 -transDPyP ligands are cofacial and slightly bowed inward, 4
Hupp and co-workers reported a formation constant of ca. 1018 M–3 for a similar sandwich system formed by two rhenium-mediated metallacycles with electron deficient zinc-porphyrins and two 4,4 -bipy axial ligands (see [57]).
Metal-Mediated Multi-Porphyrin Discrete Assemblies
125
Fig. 19 X-ray structure of the multiporphyrin molecular box [(9Zn)2 (µ-4 -transDPyP)2 ] (14) (adapted from [68])
A. The apical 4 N(py) rings bound to the Zn ions at a distance of ca. 11.4 ˚ are approximately perpendicular to the porphyrin basal planes. The CDCl3 1 H NMR spectrum of 14 is perfectly consistent with the solid state structure, suggesting that the same geometry, including the cofacial orientation of the two bridging pyridylporphyrins, is maintained in solution. Conversely, despite the unambiguous NMR evidence that in solution compound 13 exists as a discrete sandwich molecule, X-ray analysis showed that in the solid state it has a different structure, which consists of an infinite wire of porphyrin metallacycles bridged by 4,4 -bipy ligands axially coordinated alternatively on the two opposite faces of each 9Zn unit, [(9Zn)(µ-4,4 -bipy)]∞ [68, 72, 73]. We are now interested in the preparation of new sandwich assemblies using chromophores other than porphyrins as linear bridging ligands. Indeed, multi-chromophore assemblies of nanoscopic dimensions, that incorporate different light-absorbing units such as porphyrins and perylenebisimide dyes (PBI), offer the greatest chances for light-induced applications. PBI’s represent a class of stable organic chromophores possessing excellent photophysical properties and whose absorption and emission bands can be tuned by appropriate substitutions in the bay-area,
126
E. Iengo et al.
Fig. 20 Schematic drawing of the bisporphyrin/perylene-bisimide assembly [{Ru(TPP) (CO)}2 (µ-PBI)] (16)
the aromatic core of the molecule [77–83]. They can be rather easily included in supramolecular systems as bis-pyridyl perylene-bisimides. As a first step in this direction we recently prepared a bis-porphyrin adduct [{Ru(TPP)(CO)]}2 (µ-PBI) (16) (Fig. 20), whose photophysical properties are described below [84].
4 Photophysical Properties of Metal-Connected Assemblies of Porphyrins In this section, the photophysical behaviors of some prototypal porphyrins are summarized, so as to facilitate the subsequent discussion of the supramolecular systems. We use here as convenient experimental cases 4 MPyP, ZnTPP, and [Ru(TPP)(CO)(py)] (Fig. 21) [85]. These systems can be taken as good examples of the general behavior of free-base porphyrins (Fb), zinc-porphyrins (Zn), and ruthenium-porphyrins (Ru). Minor quantitative changes in the spectroscopic and photophysical behaviors of these chromophores are only expected to occur as a consequence of chemical modifications such as, for example, replacement of peripheral phenyl with pyridyl groups or axial coordination of an additional ligand (e.g. pyridine) to the metal in the case of Zn.
Metal-Mediated Multi-Porphyrin Discrete Assemblies
127
Fig. 21 Prototype porphyrins 4 MPyP (left), ZnTPP (center), and [Ru(TPP)(CO)(py)] (right)
The absorption spectrum of a free-base porphyrin (Fig. 22) involves, besides the intense Soret band, a typical Q-band system made of four distinct vibronic bands. The metalloporphyrins, on the other hand, exhibit a characteristic two-band pattern in the Q-band region. The 0 – 0 transition is strongly blue shifted upon metallation. The reasons for these different types of behaviors can be discussed in terms of the photophysical mechanisms schematized in Fig. 23. It can be seen that the free-base and zinc-porphyrin have excited singlet states with nanosecond lifetimes, that deactivate via both intersystem crossing to the triplet state (90–95% efficiency) and fluorescent emission (5–10% efficiency). This is the typical behavior of many organic chromophores such as, for example, aromatic hydrocarbons. By contrast, the behavior of the ruthenium porphyrin is similar to that of typical inorganic chromophores, being dominated by the strong spin-orbit coupling provided by the heavy ruthenium center. Thus, ultrafast intersystem crossing prevents any measurable
Fig. 22 Absorption and emission spectra (dotted lines) of prototype free-base (Fb), zinc(Zn), and ruthenium-porphyrin (Ru)
128
E. Iengo et al.
Fig. 23 Energy level diagrams and photophysical mechanisms of prototype free-base (Fb), zinc- (Zn), and ruthenium-porphyrin (Ru)
singlet-state fluorescence. Furthermore, measurable phosphorescence from the triplet state is observed in solution. Experimentally, the singlet states of free-base and zinc-porphyrins can be conveniently monitored not only by fluorescence spectroscopy but also by picosecond transient absorption (Fig. 24). By comparison with the absorption/emission spectra of Fig. 22, it can be seen that the transient spectra consist of a broad featureless positive absorption throughout the visible region, with superimposed bleaching of the ground-state Q-bands and additional apparent bleaching corresponding to stimulated fluorescent emission. As such,
Fig. 24 Transient absorption spectra obtained for free-base and zinc-porphyrin in ultrafast spectroscopy (excitation at 560 nm). Time delay after excitation pulse: 1 ps. Appreciably constant in the 1–1000 ps time range
Metal-Mediated Multi-Porphyrin Discrete Assemblies
129
Fig. 25 Triplet difference absorption spectra obtained in the laser flash photolysis of freebase, zinc-, and ruthenium-porphyrin
the picosecond transient spectra provide good fingerprints for the excited states, particularly useful in supramolecular systems containing various types of chromophores. The triplet states of free-base, zinc-, and ruthenium-porphyrins can be conveniently monitored in nanosecond laser flash photolysis (Fig. 25). The triplet transient spectra exhibit significant differences between the free-base and the metalloporphyrins that can be used for diagnostic purposes in multichromophore supramolecular systems. 4.1 Metallacycles of Porphyrins and their Model Compounds The photophysics of adducts 5, 8, 17–20 (Fig. 26), involving different 4 PyPs and octahedral Ru(II) fragments with different ancillary ligands and stereochemistry (cis,cis-RuCl2 (CO)(dmso-S)2 and trans,cis-RuCl2 (dmso-S)2 ), has been studied [86, 87]. In these compounds the Ru(II) fragments are virtually non-absorbing in the visible, so that selective excitation of the pyridylporphyrin is easily achieved. In all the adducts the pyridylporphyrin emitting singlet excited state is shorter-lived compared to the parent molecule. The effect, not very large but significant, increases with the number of ruthenium centers attached to each chromophore (Fig. 27). The magnitude of the effect also depends on the nature of the Ru center, being higher for the Ru-dmso adducts
130
E. Iengo et al.
Fig. 26 Porphyrin-ruthenium compounds that have been investigated for establishing the effect of peripheral ruthenium complexes on the photophysical behavior of the chromophores
5, 8 than for the Ru – CO series 17–20. Similar lifetime shortening effects were observed by Hupp in metallacycles with Re(I) corners [56] and by us in side-to-face assemblies in which the pyridylporphyrins are axially bound to Ru(TPP)(CO) units [85]. The origin of this lifetime shortening is interesting, since the most obvious quenching mechanisms, singlet energy transfer and photoinduced electron transfer, are prohibited on energetic grounds [87]. A likely explanation of the pyridylporphyrin singlet quenching in the arrays can be identified in
Metal-Mediated Multi-Porphyrin Discrete Assemblies
131
Fig. 27 Porphyrin fluorescence lifetime quenching by coordinated Ru centers in compounds 5, 8, 17–20 (py stands for pyridylporphyrin)
the heavy-atom effect of the metal. The conventional notion of this effect is that heavy-atom substituents introduce spin-orbit coupling into a molecule, thereby relaxing spin selection rules. In simple molecular systems, the heavyatom effect quenches the lowest excited singlet state by enhancing intersystem crossing to the triplet. In supramolecular systems of the type we are dealing with here, besides intersystem crossing within the pyridylporphyrin chromophore (kISC ), an additional spin-forbidden channel is available for deactivation of the pyridylporphyrin singlet, i.e., singlet-triplet energy transfer to the attached metal fragment (kSTEn ) (Fig. 28). In principle, both channels are expected to be sensitive to the heavy-atom effect of ruthenium, although their
Fig. 28 Possible pathways for pyridylporphyrin fluorescence quenching by heavy-atom perturbation
132
E. Iengo et al.
relative importance is difficult to predict: for kISC the heavy atom is remote, but the process is an intra-component one; for kSTEn the heavy metal center is directly involved, but the process is an inter-component one. Comparative laser flash photolysis experiments suggest that both mechanisms are operative in the adducts studied, with kISC prevailing in the Ru – CO compounds 17–20 and kSTEn prevailing in the Ru-dmso adducts 5 and 8. The reasons for the switch in mechanism is not obvious, in the absence of direct information on the triplet energies of the two types of Ru(II) complexes. Ligand field arguments could suggest that the singlet-triplet energy transfer channel may become energetically unavailable for adducts with carbonylcontaining Ru(II) centers [87]. The six metallacycles 9, 9Zn, 9Zn1/2 and 10, 10Zn, 10Zn1/2 contain two cis dipyridylporphyrins and two trans,cis-[RuCl2 (CO)2 ] corners (see Scheme 2 and Fig. 14). They differ in the geometry of the pyridyl substituents on the porphyrin, and hence in the overall geometry: 9, 9Zn, 9Zn1/2 contain 4 pyridyl groups and are planar, 10, 10Zn, 10Zn1/2 have 3 -pyridyl groups and have a slipped cofacial geometry (see previous section for detailed structural information). For each structural type, zinc-porphyrin homo-dimers (9Zn, 10Zn), or free-base/zinc-porphyrin hetero-dimers (9Zn1/2 ), 10Zn1/2 are obtained by controlling the degree of zinc insertion. The photophysics of these six metallacycles has been studied in chloroform [71]. As expected for weakly interacting systems, the absorption spectra of the homonuclear species 9, 10 and 9Zn, 10Zn are very similar to those of the parent free-base and zinc-porphyrin chromophores in the Q-band region (Fig. 22), except for minor spectral shifts. A prominent difference between the planar and the slipped cofacial macrocycles is found in the Soret band region, in which a clear exciton splitting (of ca. 500 cm–1 ) is present only for the latter compounds (10 and 10Zn). This result is as expected on the basis of the relative center-to-center distance in the two types of metallacycles A in the slipped cofacial geometry as compared to 14.1 ˚ A in the pla(10.1 ˚ nar one). The photophysics of the homo-dimers is very similar to that of the corresponding monomeric species. In particular, 9 and 10 exhibit the typical fluorescence of the free-base or zinc-porphyrin units (9 λmax = 655, 716 nm, τ = 5.7 ns; 10 λmax = 656, 716 nm) and 9Zn and 10Zn that of Zn-porphyrins (9Zn λmax = 608, 651 nm, τ = 1.1 ns; 10Zn λmax = 600, 651 nm). The lifetimes (9 and 10, 5.5 ns; 9Zn and 10Zn, 1.04 ns) are somewhat shortened (by 30–40%) with respect to the porphyrin components, as a consequence of the heavy-atom effect of the external ruthenium centers (see above for a detailed account of this phenomenon). The behavior of the hetero-dimers 9Zn1/2 and 10Zn1/2 is more interesting, as in these systems a substantial energy gradient exists between the two chromophores (Fig. 23). The absorption spectra of these semi-zincated species reflect with reasonable approximation the superposition of those of the freebase and Zn-porphyrin components, as shown for 10Zn1/2 in Fig. 29.
Metal-Mediated Multi-Porphyrin Discrete Assemblies
133
Fig. 29 Comparison between the absorption spectrum of 10Zn1/2 and those of free-base (Fb) and Zn-porphyrin (Zn) monomeric models
An important consequence is that in the hetero-dimers, besides selective excitation of the free-base porphyrin at λ > 600 nm, substantial excitation of the zincated porphyrin can be obtained with light of ca. 550 nm. In the semi-zincated species, free-base fluorescence (9Zn1/2 λmax = 658, 719 nm, τ = 5.0 ns; 10Zn1/2 λmax = 654, 718 nm, τ = 5.3 ns) is always observed, regardless of the excitation wavelength. This demonstrates the occurrence of efficient intramolecular singlet energy transfer from the Zn-porphyrin to the free-base unit. The efficiency of this process can be estimated by an appropriate comparison of the fluorescence intensity observed upon 550 nm excitation (ca. 70% Zn-porphyrin absorption) with that obtained upon 646-nm excitation (100% free-base absorption). The emission intensities obtained for the two excitation wavelengths are identical within experimental error, indicating that for both 9Zn1/2 and 10Zn1/2 , the efficiency of singlet energy transfer is ≥ 0.95. The energy transfer process can be conveniently monitored by ultrafast (femtosecond) spectroscopy, using 555-nm excitation pulses so as to achieve substantial excitation of the Zn-porphyrin chromophore. The spectral changes obtained for the slipped cofacial hetero-dimer 10Zn1/2 are shown in Fig. 30. Consistent with the partitioning of the exciting light, the initial difference spectrum is similar to that of the isolated Zn-porphyrin chromophore (Fig. 24). Very rapidly, however, the transient spectrum undergoes relevant changes, with decreasing absorbance in the short-wavelength region and a characteristic bleaching developing at 515 nm. The final, constant spectrum reached after 40 ps matches very closely that of the free-base porphyrin (Fig. 24). These spectral changes provide clear direct evidence for the occurrence of intramolecular singlet energy transfer in 10Zn1/2 from the Znporphyrin to the free-base unit. The time constant of this process is 12 ps, corresponding to a rate constant k = 8.3 × 1010 s–1 . This value is close to (ac-
134
E. Iengo et al.
Fig. 30 Transient spectral changes obtained for 10Zn1/2 in CHCl3 . Inset: decay kinetics recorded at 515 nm. Excitation wavelength, 555 nm
Fig. 31 Center-to-center distances and singlet energy transfer time constants for planar and slipped cofacial hetero-dimers 9Zn1/2 (top) and 10Zn1/2 (bottom)
tually slightly higher than) what is expected on the basis of the Förster theory of dipole–dipole energy transfer. The results of the femtosecond experiments for the planar semi-zincated metallacycle 9Zn1/2 are very similar to those described above for the slipped
Metal-Mediated Multi-Porphyrin Discrete Assemblies
135
cofacial analogue. The kinetics give a time constant of 14 ps, again very similar to that obtained for 10Zn1/2 . This result is somewhat surprising, as the A in 9Zn1/2 versus 10.1 ˚ A in 10Zn1/2 , longer center-to-center distance (14.1 ˚ Fig. 31) would imply a smaller dipole–dipole interaction and substantially slower Förster energy transfer. A plausible explanation can be offered considering that singlet energy transfer in these systems may involve, in addition to through-space dipole– dipole interaction, also some through-bond exchange interaction. Since in the ruthenium-bridged porphyrin metallacycles, through-bond interactions are expected to be more effective with 4 - rather than for 3 -pyridyl units (para rather than meta conjugation), in the planar macrocycle a better exchange coupling could compensate for a weaker dipole–dipole interaction [71]. A similar photo-induced energy transfer was observed by Hupp and co-workers in a 4 -cisDPyP semi-zincated 2 + 2 metallacycle featuring ReCl(CO)3 , rather than RuCl2 (CO)2 , corners [56]. 4.2 Higher-Order Assemblies As discussed in Sect. 3, various multiporphyrin molecular boxes can be obtained that combine both external metal-mediated and side-to-face assembling motifs. The chemical structure of the molecular box 14, resulting from side-to-face self-assembling of two 9Zn metallacycles and two 4 -transDPyP units, is shown in Fig. 32 (for the X-ray structure, see Fig. 19). The photophysics of this interesting supramolecular system has been studied in chloroform (where the system is stable at concentrations > 7 × 10–5 M)5 . The molecular box 14 contains two types of chromophores, zinc-porphyrin and free-base porphyrin. The behaviors of the monomeric models of these units, Zn and Fb, have been summarized in Sect. 4.1. As expected for supramolecular species, the absorption spectrum of the box is a good superposition of those of the molecular components (Fig. 22). The energy level diagram for the molecular box, obtained as a combination of those of the Fb and Zn models (Fig. 23), shows a significant driving force for energy transfer from the Zn-porphyrin to the free base units. Selective excitation of the free-base chromophore in the molecular box can be carried out at 640 nm, yielding the typical fluorescence of this unit. When excitation is carried out at 569 nm, on the other hand, light is predominantly absorbed by the Zn-porphyrin chromophores. Very weak fluorescence is obtained from this unit, however, whereas intense free-base fluorescence is again observed, clearly indicating the occurrence of singlet energy transfer in the molecular box. The energy transfer process can be directly monitored by 5
Prodi A et al., 2006, unpublished results
136
E. Iengo et al.
Fig. 32 Schematic representation of the molecular box 14 (phenyl groups of the pillar porphyrin are omitted for clarity), and of its antenna effect
ultrafast spectroscopy (Fig. 33), where the process is demonstrated by the disappearance of the zinc-porphyrin absorption in the 450–550 nm range and the appearance of the characteristic free-base bleaching (see, for comparison, the spectra of the Zn and Fb models in Fig. 24). The singlet energy transfer is very fast. The complex kinetics (best fit obtained with time constants of 2.4 ps and 32 ps) can be attributed to an appreciable degree of conformational freedom in the molecular box. A quantitative estimate of the sensitization efficiency, performed by comparing the intensity of the free-base emission upon free-base or zinc-porphyrin excitation, yields a value of 0.5, indicating that some additional fast process competes with energy transfer for deactivation of the zinc-porphyrin singlet excited state. Oxidative electron transfer quenching can be considered as a plausible competing deactivation channel. In conclusion, the molecular box 14 provides a good example of the antenna effect, whereby the light energy absorbed by the four Zn-porphyrin units is funneled by means of singlet–singlet energy transfer to the pillar free-base units (Fig. 33). Of particular interest in this regard is the fact that the molecular box has a central cavity that could, in principle, be used to host a variety of species capable of establishing π stacking in-
Metal-Mediated Multi-Porphyrin Discrete Assemblies
137
Fig. 33 Transient spectral changes observed in the ultrafast spectroscopy of 14 (excitation at 569 nm)
teractions with the free-base units (e.g., aromatic hydrocarbons, aromatic bisimides, fullerenes). This could open the way towards more complex functional systems. As discussed in Sect. 3, interesting assemblies can be produced using bispyridyl perylene-bisimide (PBI) units as linear bridging ligands. In terms of structural role, PBI can be considered analogous to a ditopic porphyrin ligand such as, e.g., 4 -transDPyP. Unlike the porphyrins, however, perylene bisimides undergo facile electrochemical reduction and are thus more likely to be involved in photoinduced electron transfer processes. Moreover, the high fluorescence yield and the clear spectroscopic signatures of the radical anion form make these chromophores ideally suited for photophysical studies [83, 88]. In the bisporphyrin sandwich compound [{Ru(TPP)(CO)}2 (µ-PBI)] (16) (Fig. 20), two ruthenium porphyrins are axially bound in a side-toface fashion to a central PBI chromophore. The photophysics has been studied in dichloromethane (where the assembly is stable at concentrations > 2 × 10–5 M) [84]. As expected, the absorption spectrum of 16 corresponds very closely to the sum of those of its molecular components (Fig. 34). Experimentally, this implies that selective excitation of the two types of chromophores is feasible (e.g., 100% PBI at 585 nm, 62% Ru(TPP)(CO) at 530 nm). The energy level diagram of 16 (Fig. 35) can be considered as a simple superposition of those of the separated constituents, with the addition of a charge transfer state in which PBI is reduced and the Ru porphyrin is oxidized (energy estimated from electrochemistry). Ultrafast spectroscopy experiments performed at the two excitation wavelengths afforded evidence that the side-to-face assembly 16 incorporating a perylene bisimide chromophore exhibits complex, interesting photophysics schematically summarized in Figs. 35 and 36.
138
E. Iengo et al.
Fig. 34 Absorption spectra of 16 (continuous line), PBI (dash-dot line), and [Ru(TPP)(CO) (py)] (broken line, ×2)
Fig. 35 Photophysical pathways in [{Ru(TPP)(CO)}2 (µ-PBI)] (16). Red, excitation at 530 nm; blue, excitation at 585 nm
Compound 16 provides a rather striking example of wavelength-dependent behavior, in that a relatively small change in excitation wavelength (from 585 to 530 nm) causes a sharp change in photophysical response (from intramolecular electron transfer to triplet energy transfer). The triplet state of the perylene bisimide, inaccessible in the free chromophore due to high fluorescence quantum yields and negligible intersystem crossing efficiency, is efficiently accessed in 16 by means of intramolecular sensitization [84].
Metal-Mediated Multi-Porphyrin Discrete Assemblies
139
Fig. 36 Schematic representation of the complex, wavelength-dependent electron/energy transfer behaviors of 16. Left: photoinduced electron transfer upon excitation of the PBI unit at 585 nm; right: photoinduced energy transfer upon excitation of the Ru(TPP)(CO) units at 530 nm
5 Conclusions and Outlook The formation of coordination bonds between peripheral donor sites on the porphyrins and external metal fragments has proved to be a very versatile and powerful synthetic approach for the construction of discrete, ordered multiporphyrin assemblies. Several robust and shape-persistent metal-mediated assemblies of increasing structural complexity and beauty, that gradually approach those of the multichromophore systems found in Nature, have been prepared in recent years. In our contributions we have mainly exploited the coordination capabilities of two neutral Ru(II)-dmso precursors, trans-[RuCl2 (dmso-S)4 ] (1) and trans,cis,cis-[RuCl2 (dmso-O)2 (CO)2 ] (2), that selectively replace two adjacent dmso ligands. Nevertheless, the number of multitopic metal fragments used so far in this synthetic approach is still limited (most of them are 90◦ cis bis-acceptor fragments). Thus, we believe that in future the exploitation of a larger variety of metal connectors, in terms of geometries and coordination numbers, will provide even more structurally sophisticated assemblies. Also the charge of the metal fragments is an important parameter to be considered, as it will determine that of the assembly. As a general consideration, neutral multi-porphyrin assemblies are more suitable
140
E. Iengo et al.
compared to the corresponding charged species for being used as units for the construction of useful framework solids. In fact, their crystal lattices are likely to define cavities and semi-infinite channels which, owing to the absence of counter-ions, might be used (after solvent removal) for microporous materials applications (e.g. selective separation, chemical sensing, or catalysis). When the coordination bonds that hold these assemblies together are both stable and inert, their formation from the components occurs under kinetic rather than thermodynamic control and it is thus disputable if they can be truly defined as supramolecular systems. On the other hand, owing to their kinetic stability, some of these species can be further exploited as building blocks in the construction of higher order architectures through a hierarchical self-assembly approach (see for example the molecular sandwiches 13–15). Through this modular approach, multichromophore systems become easily accessible on demand, with a limited synthetic effort. Even though solution spectroscopic measurements provide a wealth of information about the nature of the metal-mediated products, very often solid state X-ray structural determinations have proved to be essential for establishing the real composition and geometry of the multiporphyrin assemblies. This is particularly true for highly symmetrical adducts. Among the remarkable number of X-ray structures that we contributed to the still limited database of metal-mediated multi-porphyrin assemblies, we found several examples of unexpected geometries and/or distortions. Thus, while, in general, thinking of the porphyrins and metal fragments as rigid building blocks may be an acceptable approximation for the design of new systems, their surprising (and often unexpected) flexibility should not be forgotten when it comes to understanding the geometry of new assemblies in the absence of an X-ray structural determination. In addition, when porphyrins such as 3 -PyPs are used as building blocks, in which the peripheral binding groups display relevant torsional freedom, conformational parameters may be expected to affect strongly the geometry of the final assembly. Some of the new assemblies of porphyrins, in particular when they define a cavity with specific shape and size, might behave as selective hosts in molecular recognition reactions. Sensing properties, that might eventually lead to practical applications, may be found for the multi-porphyrin adducts in the case that their photophysical properties (e.g. their fluorescence) change as a result of molecular recognition. For example, molecular boxes 14 and 15 are particularly stimulating for further investigation, as they feature two A (the bridging ligands) that cofacial porphyrins at a distance of about 11.4 ˚ might induce the inclusion of guest molecules of appropriate shape and size through π – π interactions. Interesting guests could be simple molecules to be used as energy or electron acceptors (e.g., aromatic hydrocarbons, aromatic bisimides, fullerenes). Finally, despite the increased structural complexity, the photo-induced functions of elaborated artificial assemblies of chromophores are most often
Metal-Mediated Multi-Porphyrin Discrete Assemblies
141
still far from those of natural systems. We believe that in future, when the construction principles will be better understood, a more appropriate choice of the metal connectors in terms of their energy levels and photophysical properties will allow us to improve the photo-induced response of the assemblies. In addition, towards practical utilization, it would be appropriate to further organize a large number of identical functional systems into ordered structures. For example functional multi-porphyrin arrays might be anchored on solid surfaces, such as metals or wide bandgap semiconductors, or on metal nanoparticles. Acknowledgements Most of the X-ray structures of ruthenium-mediated assemblies of porphyrins mentioned or showed in this contribution were solved by Prof. Ennio Zangrando (Dipartimento di Scienze Chimiche, Università di Trieste), who also very kindly provided the new drawings. Part of the work summarized here was financially supported by the donors of the Petroleum Research Fund, administered by the ACS, (grant ACS PRF# 38892-AC3) and by MIUR (PRIN 2003 no. 2003035553 and FIRB-RBNE019H9K projects).
References 1. Deisenhofer J, Epp O, Miki K, Huber R, Michel H (1984) J Mol Biol 180:385 2. Allen JP, Feher G, Yeates TO, Komiya H, Rees DC (1987) Proc Natl Acad Sci USA 84:5730 3. Deisenhofer J, Epp O, Sinning I, Michel H (1995) J Mol Biol 246:429 4. Zouni A, Witt H-T, Kern J, Fromme P, Krauß N, Saenger W, Orth P (2001) Nature 409:739 5. Jordan P, Fromme P, Witt HT, Klukas O, Saenger W, Krauß N (2001) Nature 411:909 6. Ferreira KN, Iverson TM, Maghlaoui K, Barber J, Iwata S (2004) Science 303:1831 7. McDermott G, Prince SM, Freer AA, Hawthornthwaite-Lawless AM, Papiz MZ, Cogdell RJ, Isaacs NW (1995) Nature 374:517 8. Kuhlbrandt W (1995) Nature 374:497 9. Pullerits T, Sundström V (1996) Acc Chem Res 29:381 10. Ritz T, Park S, Schulten KJ (2001) Phys Chem B 105:8259 11. Papiz MZ, Prince SM, Howard T, Cogdell RJ, Isaacs NW (2003) J Mol Biol 326:523 12. Wasielewski MR (1992) Chem Rev 92:435 13. Gust D, Moore TA, Moore AL (1993) Acc Chem Res 26:198 14. Harriman A, Sauvage J-P (1996) Chem Soc Rev 25:41 15. Bennett IM, Vanegas Farfano HM, Bogani F, Primak A, Liddell PA, Otero L, Sereno L, Silber JJ, Moore AL, Moore TA, Gust D (2002) Nature 420:401 16. Aratani N, Osuka A, Cho HS, Kim D (2002) J Photochem Photobiol C 3:25 17. Smirnov SN, Liddell PA, Vlassiouk IV, Teslja A, Kuciauskas D, Braun CL, Moore AL, Moore TA, Gust D (2003) J Phys Chem A 107:7567 18. Guldi DM (2003) Pure Appl Chem 75:1069 19. Gould SL, Kodis G, Palacios RE, de la Garza L, Brune A, Gust D, Moore TA, Moore AL (2004) J Phys Chem B 108:10566 20. Rybtchinski B, Sinks LE, Wasielewski MR (2004) J Am Chem Soc 126:12268 21. Anderson S, Anderson HL, Bashall A, McPartlin M, Sanders JKM (1995) Angew Chem Int Ed Engl 34:1096
142
E. Iengo et al.
22. Davila J, Harriman A, Milgrom LR (1987) Chem Phys Lett 136:427 23. Prathapan S, Johnson TE, Lindsey JS (1993) J Am Chem Soc 115:7519 24. Hsiao J-S, Krueger BP, Wagner RW, Johnson TE, Delaney JK, Mauzerall DC, Fleming GR, Lindsey JS, Bocian DF, Donohoe RJ (1996) J Am Chem Soc 118:11181 25. Nakano A, Osuka A, Yamazaki I, Yamazaki T, Nishimura Y (1998) Angew Chem Int Ed Engl 37:3023 26. Kuciauskas D, Liddell PA, Lin S, Johnson TE, Weghorn SJ, Lindsey JS, Moore A, Moore TA, Gust D (1999) J Am Chem Soc 121:8604 27. Anderson HL (1999) Chem Commun 2323 28. Kim YH, Jeong DH, Kim D, Jeoung SC, Cho HS, Kim SK, Aratani N, Osuka A (2001) J Am Chem Soc 123:76 29. Yang SI, Prathapan S, Miller MA, Seth J, Bocian DF, Lindsey JS, Holten D (2001) J Phys Chem B 105:8249 30. Taniguchi M, Ra D, Kirmaier C, Hindin EK, Schwartz JK, Diers JR, Bocian DF, Lindsey JS, Knox RS, Holten D (2003) J Am Chem Soc 125:13461 31. Takahashi R, Kobuke Y (2003) J Am Chem Soc 125:2372 32. Balaban TS, Bhise AD, Fischer M, Linke-Schaetzel M, Roussel C, Vanthuyne N (2003) Angew Chem Int Ed 42:2140 33. Pen X, Aratani T, Matsumoto T, Kawai T, Hwang I-W, Ahn TK, Kim D, Osuka A (2004) J Am Chem Soc 126:4468 34. Choi M-S, Yamazaki T, Yamazaki I, Aida T (2004) Angew Chem Int Ed 43:150 35. Gust D, Moore TA, Moore AL (2001) Acc Chem Res 34:40 36. Holten D, Bocian DF, Lindsey JS (2002) Acc Chem Res 35:57 37. Campbell WM, Burrell AK, Officer DL, Jolley KW (2004) Coord Chem Rev 248:817 38. Kim D, Osuka A (2004) Acc Chem Res 37:735 39. Chambron J-C, Heitz V, Sauvage J-P (2000) In: Kadish KM, Smith KM, Guillard R (eds) The Porphyrin Handbook, chap 40, vol 6. Academic Press, San Diego 40. Wojaczy´ nski J, Latos-Gra˙zy´ nsky L (2000) Coord Chem Rev 204:113 41. Imamura T, Fukushima K (2000) Coord Chem Rev 198:133 42. Chou J, Kosal ME, Nalwa HS, Rakow NA, Suslick KS (2000) In: Kadish KM, Smith KM, Guillard R (eds) The Porphyrin Handbook, chap 41, vol 6. Academic Press, San Diego 43. Kumar KR, Goldberg I (1998) Angew Chem Int Ed 37:3027 44. Pan L, Noll BC, Wang X (1999) Chem Commun 157 45. Sharma CVK, Broker GA, Huddleston JG, Baldwin JW, Metzger RM, Rogers RD (1999) J Am Chem Soc 121:1137 46. Pan L, Kelly S, Huang X, Li J (2002) Chem Commun 2334 47. Kosal ME, Chou J-H, Wilson SR, Suslick KS (2002) Nat Mater 1:118 48. Suslick KS, Bhyrappa P, Chou J-H, Kosal ME, Nakagaki S, Smithenry DW, Wilson SR (2005) Acc Chem Res 38:283 49. Drain CM, Lehn J-M (1994) J Chem Soc, Chem Commun 2313 50. Drain CM, Nifiatis F, Vasenko A, Batteas JD (1998) Angew Chem Int Ed 37:2344 51. Yuan H, Thomas L, Woo LK (1996) Inorg Chem 35:2808 52. Stang PJ, Fan J, Olenyuk B (1997) Chem Commun 1453 53. Fan J, Whiteford JA, Olenyuk B, Levin MD, Stang PJ, Fleischer EB (1999) J Am Chem Soc 121:2741 54. Schmitz MM, Leininger S, Fan T, Arif AM, Stang PJ (1999) Organometallics 18:4817 55. Slone RV, Hupp JT (1997) Inorg Chem 36:5422 56. Splan KE, Keefe MH, Massari AM, Walters KA, Hupp JT (2002) Inorg Chem 41:619 57. Splan KE, Stern CL, Hupp JT (2004) Inorg Chim Acta 357:4005
Metal-Mediated Multi-Porphyrin Discrete Assemblies
143
58. Schenning APHJ, Benneker FBG, Geurts HPM, Liu XY, Nolte RJM (1996) J Am Chem Soc 118:8549 59. Foekema J, Schenning APHJ, Vriezema DM, Benneker FBG, Nørgaard K, Kroon JKM, Bjørnholm T, Feiters MC, Rowan AE, Nolte RJM (2001) J Phys Org Chem 14:501 60. Sun D, Tham FS, Reed CA, Chaker L, Burgess M, Boyd PDW (2000) J Am Chem Soc 122:10704 61. Sun D, Tham FS, Reed CA, Chaker L, Boyd PDW (2002) J Am Chem Soc 124:6604 62. Fujita N, Biradha K, Fujita M, Sakamoto S, Yamaguchi K (2001) Angew Chem Int Ed 40:1718 63. Ikeda A, Ayabe M, Shinkai S, Sakamoto S, Yamaguchi KA (2000) Org Lett 3707 64. Ayabe M, Yamashita K, Sada K, Shinkai S, Ikeda A, Sakamoto S, Yamaguchi K (2003) J Org Chem 68:1059 65. Alessio E, Macchi M, Heat SL, Marzilli LG (1997) Inorg Chem 36:5614 66. Iengo E, Milani B, Zangrando E, Geremia E, Alessio E (2000) Angew Chem Int Ed 39:1096 67. Alessio E, Ciani E, Iengo E, Kukushkin VY, Marzilli LG (2000) Inorg Chem 39:1434 68. Iengo E, Zangrando E, Minatel R, Alessio E (2002) J Am Chem Soc 124:1003 69. Iengo E, Zangrando E, Geremia S, Graff R, Kieffer B, Alessio E (2002) Chem Eur J 8:4670 70. Gianferrara T, Serli B, Zangrando E, Iengo E, Alessio E (2005) New J Chem 29:895 71. Iengo E, Zangrando E, Bellini M, Alessio E, Prodi A, Chiorboli C, Scandola F (2005) Inorg Chem 44:9752 72. Alessio E, Iengo E, Marzilli LG (2002) Supramol Chem 14:103 73. Iengo E, Zangrando E, Alessio E (2003) Eur J Inorg Chem 237 74. Alessio E (2004) Chem Rev 104:4203 75. Alessio E, Mestroni G, Nardin G, Attia WM, Calligaris M, Sava G, Zorzet S (1988) Inorg Chem 27:4099 76. Alessio E, Milani B, Bolle M, Mestroni G, Faleschini P, Geremia S, Calligaris M (1995) Inorg Chem 34:4722 77. Würthner F, Thalacker C, Sautter A (1999) Adv Mater 11:754 78. Würthner F, Thalacker C, Sautter A, Schärtl W, Ibach W, Hollricher O (2000) Chem Eur J 6:3871 79. Schenning APHJ, v. Herrikhuyzen J, Jonkheijm P, Chen Z, Würthner F, Meijer EW (2002) J Am Chem Soc 124:10252 80. You C-C, Würthner F (2003) J Am Chem Soc 32:9716 81. Würthner F, Sautter A (2003) Org Biomol Chem 1:240 82. Miura A, Chen Z, Uji-i H, De Feyter S, Zdanowska M, Jonkheijm P, Schenning APHJ, Meijer EW, Würthner F, De Schryver FC (2003) J Am Chem Soc 125:14968 83. Würthner F (2004) Chem Commun 1564 84. Prodi A, Chiorboli C, Scandola F, Iengo E, Alessio E, Dobrawa R, Würthner F (2005) J Am Chem Soc 127:1454 85. Prodi A, Indelli MT, Kleverlaan CJ, Scandola F, Alessio E, Gianferrara T, Marzilli LG (1999) Chem Eur J 5:2668 86. Prodi A, Indelli MT, Kleverlaan CJ, Alessio E, Scandola F (2002) Coord Chem Rev 229:51 87. Prodi A, Kleverlaan CJ, Indelli MT, Scandola F, Alessio E, Iengo E (2001) Inorg Chem 40:3498 88. Salbeck J, Kunkely H, Langhals H, Saalfrank RW, Daub J (1989) Chimia 43:6
Struct Bond (2006) 121: 145–165 DOI 10.1007/430_027 © Springer-Verlag Berlin Heidelberg 2006 Published online: 17 February 2006
Rhenium-Linked Multiporphyrin Assemblies: Synthesis and Properties Joseph T. Hupp Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, USA [email protected] 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 2.1 2.2 2.3
Synthesis of Rhenium-Porphyrin Supramolecular Assemblies Squares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rectangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Planar Dimers . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
146 146 149 151
3
Porous Materials Based on Supramolecular Assemblies . . . . . . . . . . .
155
4 4.1 4.2 4.3 4.4 4.5
Applications and Function . . . . . . . . . . . . . . . Film- and Membrane-Based Molecular Sieving . . . . Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . Fundamental Studies of Energy and Electron Transfer Light-to-Electrical Energy Conversion . . . . . . . . .
. . . . . .
159 159 161 161 162 163
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
164
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
145
Abstract Re(CO)5 Cl readily reacts with pyridine-derivatized porphyrins to give molecular squares or planar dimers featuring fac-Re(CO)3 (Cl) corners. Chelating ligands, in sequential combination with pyridine-derivatized porphyrins, can be used to obtain related cofacial porphyrin assemblies. Films and membranes displaying molecular-scale porosity can be generated from squares by: (1) van der Waals aggregation, (2) layer-bylayer assembly chemistry based on zirconium-phosphonate links, or (3) polymerization at liquid–liquid interfaces. In addition to molecular sieving, films, membranes, and free assemblies have been used for chemical sensing, oxidative catalysis, and light-to-electrical energy conversion. The assemblies have also been used to investigate fundamental aspects of light-induced energy and electron transfer. Keywords Rhenium · Molecular square · Supramolecular · Porphyrin
1 Introduction Rhenium ions, and especially tricarbonylrhenium(I)chloro fragments, have been used to organize and link pyridine-functionalized porphyrins as molecular squares, rectangles, planar dimers, and more complex structures.
146
J.T. Hupp
The idea followed from successful syntheses of molecular squares featuring smaller ligand edges such as 4,4 -bipyridine and pyrazine. The attractions of Re(CO)3 Cl in the initial studies were three: (a) The majority of imine and azine complexes of Re(CO)3 Cl are photo luminescent [1, 2]. They typically display long excited-state lifetimes, with emission coming from radiative decay of nominally triplet metal-to-ligand charge transfer states. Early on, emission was viewed as a potential signal transduction scheme in host:guest type chemical sensing applications [3]. For ligand-centered emission, such as porphyrin fluorescence, however, the presence of rhenium centers is a detriment. The proximal metals enhance spin-orbit coupling within the photo-excited ligand (the “heavy atom” effect). This accelerates singlet excited state to triplet excited state intersystem crossing, thereby shortening the lifetime of the singlet state, and diminishing the fluorescence yield. (b) Neutral compounds are readily obtainable. The single positive charge on rhenium is balanced by a coordinated halide. If neutral ligand edges are used for square construction, the isolated square is charge neutral. Analogous assemblies based on Pt(II) or Pd(II) coordination typically are octa-cationic [4, 5]. Charge neutrality for the rhenium squares eliminates the possibility of channel blocking by counter ions in the solid state. (Often, though, for analogous cationic Pt(II) and Pd(II) assemblies, counter ions have been found to occupy sites between square layers rather than within square cavities, obviating the concern about channel blocking.) Charge neutrality also confers insolubility in water—a potentially useful property if the squares are used as thin-film molecular aggregate materials, as discussed below. Of course, insolubility is a disadvantage if solution-phase applications are envisioned. (c) Rhenium-imine bonds are inert at ambient temperature. In contrast, analogous Pt(II) and especially Pd(II) assemblies are somewhat labile, especially if competing ligands are introduced. At elevated temperatures, however, rhenium-imine bonds are labile enough to allow for conversion of kinetic structures (such as open oligomers) to thermodynamic structures during the assembly process. With one exception [6], attempts to use less expensive manganese centers as a substitute for rhenium have been unsuccessful because of the greater lability of Mn(I).
2 Synthesis of Rhenium-Porphyrin Supramolecular Assemblies 2.1 Squares In weakly coordinating solvents such as mixtures of toluene and tetrahydrofuran, the 1 : 1 combination of Re(CO)5 X (X = Cl, Br, or I) and a rigid or semi-rigid dipyridyl ligand generally produces molecular squares in high yield [7–9]. The strong trans labilizing effect of CO allows two of the carbonyl
Rhenium-Linked Multiporphyrin Assemblies
147
ligands of Re(CO)5 X to be replaced with solvent molecules. The octahedral geometry of the rhenium coordination sphere ensures that the reactive sites are cis to each other. It also ensures that fac rather than mer compounds are obtained; see Scheme 1. Subsequent coordination of pyridine-terminated ligands yields complexes defining a roughly 90◦ pyridine-Re-pyridine angle. With difunctional ligands, the number of Re-imine bonds is maximized by forming cyclic as opposed to open structures. Since the Re-imine bonds are stronger than Re-solvent bonds, cycle formation is enthalpically favored. Strain is minimized by forming structures having square geometries (although triangular assemblies are known [10]). Contributing to the typically high yields is the generally greater solubility of open structures than squares. Also important is the lability of the Re-imine bond under refluxing conditions in weakly coordinating solvents. Lability allows “mistakes” to be undone, and new bonds to be formed until thermodynamic structures are generated. Note that the halide ligand for each rhenium can be oriented either up or down with respect to the square framework. Four isomers are possible (Fig. 1). While single isomers have occasionally selectively crystallized, the syntheses are believed to produce statistical mixtures of isomers. The first rhenium porphyrin square to be reported was compound 1. The square has resisted crystallization, but the structure is supported by FAB mass spectral measurements. From molecular modelling the edge length (rhenium-to-rhenium) is ca. 20 ˚ A. As described below, closely related porphyrinic squares featuring slightly larger cavities have been reported. Characterization of these has proven particularly challenging because of insufficient volatility in high-molecular-weight mass spectrometry experiments. The unique CO infrared signature of fac-Re(CO)3 (X)L2 species and the relative simplicity of proton NMR spectra for high symmetry cyclic structures enable cycles to be distinguished from acyclic oligomers or polymers, but do
Scheme 1
148
J.T. Hupp
Fig. 1 Four isomers can be formed for molecules having the formula {Re(CO)3 (X) (µ-ligand)}4
not indicate the size of the cycle. In at least one case, vapor phase osmometry has been used to confirm porphyrin square formation [11]. Molecular squares define cavities that are potential hosts for appropriate molecular guests. In methylene chloride as solvent, pyridine binds to the available Zn(II) sites of 1Zn with an association constant of ∼ 103 M–1 . Whether binding occurs on the square interior or exterior (or a combination of both) has not been established. Recall, however, that Zn porphyrins generally axially bind only one ligand. The free base form of the dipyridyl porphyrin used to assemble the squares binds to 1Zn with an equilibrium constant of 3 × 106 M–1 [7]. The stronger association is a consequence of two-point binding—necessarily occurring within the cavity. Tetrapyridyl-
Structure 1
Rhenium-Linked Multiporphyrin Assemblies
149
Structure 2
porphyrin (TPyP) binds to all four zinc ions, yet is only marginally better associated: Kbinding = 4 × 107 M–1 . The benefit of additional binding sites is evidently largely offset by the cost of orienting all four porphyrins normal to the plane defined by the rhenium centers. A structural analysis by molecular mechanics and density-functional methods indicated that the porphyrin walls of the lowest energy conformers of the empty square are tilted away from a “vertical” or box-like geometry [12]. A version of the square lacking alkyl substituents but having a pair of pentafluorobenzyl substituents (2Zn) was found to bind a representative ligand, 4-phenylpyridine, about an order of magnitude more strongly than does 1Zn [13]. The stronger binding reflects the electron-withdrawing nature of the perfluorinated substituents and the resulting enhancement of the Lewis acidity of the Zn(II) centers. 2.2 Rectangles Although exceptions exist [14, 15], attempts to prepare tetra-rhenium rectangles simply by combining short and long difunctional ligands with Re(CO)5 (halide) have tended to yield mixtures of the corresponding small and large squares instead. An alternative approach is to strengthen the metal-ligand interactions along the short edges of the rectangle by first chelating the rhenium centers to yield either Re2 (CO)6 Cl2 (µ-bipyrimidine) or Re2 (CO)8 {µ-bis(benzimidazolate)} [16]. Subsequent stoichiometric reaction of the bis(benzimidazolate)-containing edge unit with [5,15-bis(4ethynlpyridyl)porphyrinato]zinc(II) results in CO displacement and formation of the neutral rectangle 3Zn. Extraction of chloride ligands with Ag+ allows reaction of the bipyrimidine-containing edge with the same porphyrin to give the cationic rectangle 4Zn. The syntheses were initially attempted using porphyrin 5. The failure to obtain isolable rectangles was attributed
150
J.T. Hupp
Fig. 2 ORTEP representations of 3Zn showing the side view (top) and top view (bottom) of the co-facial porphyrin systems; hydrogen atoms have been omitted for clarity; the thermal ellipsoids represent 50% occupancy. Adapted from [16]
Structure 3
to unfavorable steric interactions associated with making the pyridine substituents coplanar with the porphyrin. A crystal structure of 3Zn showed that cavity collapse occurs—the driving force being van der Waals interactions between the porphyrins; see Fig. 2.
Rhenium-Linked Multiporphyrin Assemblies
151
Structure 4
Structure 5
Consistent with strong van der Waals attraction, attempts to force open the cavity by binding the difunctional ligand 1,4-diazaobicyclo[2.2.2]octane (DABCO), a strong Lewis base, to the available pair of zinc ions were not successful. 2.3 Planar Dimers The strong preference for cis substitution of pentacarbonyl halo rhenium(I) synthons has been exploited to prepare dimers and hetero dimers of 5,10pyridyl derivatized porphyrins; see compounds 6–8 and Scheme 2 [13, 17]. Related chemistry has been demonstrated with platinum(II) and palladium(II) compounds where square planar coordination is used to an advantage [18]. In contrast to the square planar compounds, the octahedral
152
J.T. Hupp
Scheme 2
coordination of rhenium yields distinct syn and anti isomers with respect to halide ligand configuration. (For simplicity, only one of the compounds has been sketched in both syn and anti form. Higher symmetry octahedrally coordinated Ru(II) compounds, lacking the complications of isomer formation, have also been reported [19] and are discussed elsewhere in this volume.) The isomers display identical 1 H NMR spectra, absorption spectra and fluorescence spectra, but they are separable via chromatography on alumina. To assign the configurations of separated isomers an interesting photophysical approach, transient direct-current photoconductivity, was used. This technique reports on changes in dipole moment, µ, upon photoexcitation. More precisely, it measures the quantity: µ2 (ground state) – µ2 (excited state). The ground-state dipole moment for the anti isomer is zero by symmetry, but greater than zero for the cis isomer. Photo excitation causes identical changes in dipole moment (µ (ground state) – µ (excited state)) for the two isomers, but different changes in µ2 (ground state) – µ2 (excited state). A useful extension (not yet reported) would be to replace the halide ligands with larger functional ligands (other chromophores, energy or electron accepting units, receptors for molecular guests, etc.). Another potentially attractive extension (not yet reported) would be incorporation of tetrapyridyl
Rhenium-Linked Multiporphyrin Assemblies
153
Structure 6
porphyrin repeat units as suggested in structure 9. An extension that has been reported is dimer and hetero dimer formation from sulfur and oxygen substituted porphyrins; see compounds 10–12 [20]. In the presence of 4,4 -bipyridine, the zinc metallated dimers 6b and 8b form ligand-pillared cofacial pairs of dimers, 6d and 8c; see Scheme 3 [13]. These assemblies are highly reminiscent of the ruthenium-linked tetraporphyrinic assemblies of Iengo and coworkers [21]. These assemblies, like the ruthenium ones, are formed in an “all or nothing” fashion, as evidenced by the lack of intermediate species in titrations of the dimers with the pillars. In other words, the assembly process is a cooperative one: the binding constant for the second pillar is significantly larger than the first. The cooperative behavior is a consequence of the pre-organization achieved by the receptors
Structure 7
154
Structure 8
Structure 9
Scheme 3
J.T. Hupp
Rhenium-Linked Multiporphyrin Assemblies
155
once the first pillar is in place. An indication of the degree of cooperativity in binding was provided by comparative studies with the monofunctional ligand, 4-phenyl-pyridine. To account for differences in stoichiometry, values of c50 —the concentration where half the assemblies are dissociated—were determined. The values were about 20-fold lower for 8c than for the 4-phenylpyridine assembly. They were about 4-fold lower for 6d than for 8c, consistent with the greater Zn(II) Lewis acidity for the pentafluorophenyl-derivatized dimer.
3 Porous Materials Based on Supramolecular Assemblies Several studies have described the behavior of thin films of molecular square 1 (see Sect. 4). The films consist simply of molecular aggregates where advantage is taken of the insolubility of the squares in water, in particular, to retain the molecules as functional films. Careful evaporative casting from solvents such as chloroform was found to yield essentially pinhole-free films, in some cases with average film thicknesses as low as ∼ 25 nm [22]. Useful films have been prepared on glassy carbon, indium-tin oxide, glass, and mesoporous polyester membranes. Modification of the thin-film materials by Zn(II) ligation of various imines and other nitrogen donors was demonstrated [23]. Interestingly, binding proved to be much more persistent in the film environment than in solution (based on comparisons to soluble porphyrins). More robust films have been obtained by using the phosphonate-functionalized square, 13Zn, and employing a layer-by-layer assembly scheme [24]. This scheme, which builds on work by several research groups, exploits the high affinity of phosphonates for Zr(IV) [25, 26]. Film growth can be initiated either by direct attachment of the octa-phosphonated square to a metaloxide surface or by first derivatizing the supporting surface with a zirconium binding ligand. Note that two steps are entailed in addition of each layer (Scheme 4). This ensures that only one layer is added per assembly cycle, so allows for very good control over film thickness. This point is illustrated in Fig. 3 where film thicknesses, measured by atomic force microscopy, are plotted against the number of assembly cycles. From the figure, the average layer thickness is ca. 1.8 nm, compared with an expected thickness of 2.5 nm if the walls of the square are strictly vertical. That the observed value is smaller suggests that the squares are configured in a partially collapsed form. Film characterization by X-ray reflectivity, X-ray fluorescence, and long-period X-ray standing wave measurements corroborate the partial collapse, or tilting of the walls, of filmconfined squares [27]. The measurements also indicate that appreciably more zirconium is incorporated than implied by Scheme 4.
156
Scheme 4
Structure 10
J.T. Hupp
Rhenium-Linked Multiporphyrin Assemblies
157
Fig. 3 AFM heights for ZrP films of 13 assembled on glass. Adapted from [24]
Scheme 5 Synthesis and interfacial polymerization of molecular square 14. For the interfacial polymerization, the square is confined to the aqueous phase and the acid chloride linker to the organic phase. The polymer structure shown is idealized for illustrative purposes; the actual structure lacks net monomer orientation and is almost certainly cross-linked
158
J.T. Hupp
Fig. 4 Scanning electrochemical microscopy images of 200 µm × 200 µm areas on micropatterned conductive glass slides: the columns (from left to right) correspond to electrode areas featuring a thin-film overlayer of open or cavity-modified squares. The rows correspond to the following redox mediators: (A) 2 mM ferrocene-methanol (diameter ∼ 4.5 ˚ A); (B) 3 mM Ru(NH3 )6 3+ (diameter ∼ 5.5 ˚ A); (C) 5 mM Fe(1,10-phenanthroline)2+ 3 (diameter ∼ 13 ˚ A) Eplat = 1.0 V, i = 3.5 nA; and (D) 10 mM Fe(4,7-phenylsulfonate-1,0˚ phenanthroline)4– 3 (diameter ∼ 24 A). All solutions contained 0.1 M KNO3 electrolyte. Adapted from [22]
The layer-by-layer assembly approach also allows defects or pinholes to be overwritten. Electrochemical measurements indicated that detectable pinholes are typically eliminated once three layers are formed [24]. A limitation of the layer-by-layer assembly method is that the resulting films must be supported by a rigid platform. An alternative approach that enables free-standing films (membranes) to be fabricated is polymerization at liquid/liquid interfaces. Porphyrin squares featuring reactive phenolic substituents (14Zn) were dissolved in buffered water and allowed to react with small bis(acid chloride) cross-linkers present in a water-immiscible chloroform phase (see Scheme 5) [28]. Condensation yields ester linkages and a thin porous polymeric film. Closely related nonporous film formation chemistry has been described by Wamser and coworkers based on appropriately functionalized pairs of monomeric porphyrin reactants [29]. Because film formation at the liquid/liquid interface inhibits reactant transport, the polymerization process is self-limiting (Fig. 4). Additionally, because the reactant flux is highest at defect sites in the membrane, pinholes are automatically filled. The thicknesses of harvestable films ranged from about 300 to 2500 nm.
Rhenium-Linked Multiporphyrin Assemblies
159
4 Applications and Function 4.1 Film- and Membrane-Based Molecular Sieving Simple molecular-aggregate type films have proven surprisingly effective as molecular sieves. In one study involving films of 1 cast on mesoporous polyester membrane supports, molecular transport was examined using a U-tube configuration, with the membrane as the size-selective barrier between solutions [30]. Transport was observed via UV-vis absorption by the permeant in the receiving solution. Films were found to be permeable to phenol and to a 13-˚ A diameter probe, but blocking toward a 24-˚ A diameter probe. No dependence on molecular charge was seen. The findings are A diameter for the square cavity. In an adconsistent with an estimated 18-˚ ditional study, the square cavity was bisected by binding a fifth dipyridyl porphyrin to two of the four available Zn(II) sites of 1Zn. These cavity-modified A probe. Synfilms proved permeable to phenol but blocking toward the 13-˚ chrotron X-ray studies showed the films to be amorphous. Cross-sectional microscopy studies indicated film thickness of ca. 16 µm plus another 9 µm of porphyrinic material infiltrating the support pores. Vapor permeation of membrane-supported films of 1Zn has also been studied [31]. Examined as permeants were benzene, toluene, 4-picoline, 2-picoline, cyclohexane, and methylcyclohexane. In pair-wise comparisons modest permeation selectivities were observed (i.e., factors of 1.1 to 9, depending on the permeant pairs compared). The competitive transport measurements were made at equal vapor pressures for the component pairs. The selectivities decrease if comparisons are made at equal reduced vapor pressures, P/Po , where Po is the vapor pressure at saturation. Nonspecific sorption of volatile compounds at a given partial pressure generally inversely correlates with values for the saturated vapor pressure. Scanning electrochemical microscopy was used to examine molecular transport through the same material, but now on a conductive platform. Water-soluble redox probe molecules of various sizes were used, with electrochemical signals being observed only from those small enough to permeate the film. Figure 4 shows how the size cutoff changes as the cavity size is altered by incorporating either a dipyridyl- or tetra-pyridyl porphyrin guest. Similar measurements using either walljet electrochemistry or steady-state microelectrode voltammetry have been reported for layer-by-layer assembled and interfacially polymerized materials, respectively [24, 28]. Additional measurements were made spectrophotometrically with polymerized porphyrin squares by using a U-tube. Results summarized in Fig. 5 revealed the following: (a) After normalizing for differences in film thickness, transport through polymeric membranes is two to three orders of magnitude faster
160
J.T. Hupp
Fig. 5 PDf values for ZrP-assembled (right axis) and interfacially polymerized (left axis) films of porphyrinic molecular squares 13Zn () and 14Zn (•), respectively, as a function of the inverse radius of the permeant. (PDf = permeability = (solution-to-film partition coefficient) × (film-based diffusion coefficient). Adapted from [24]
than through layer-by-layer assembled films. (b) A sharp size cutoff exists for the polymeric materials, but layer-by-layer assembled films show a gradual cutoff. (c) For the polymeric material the size cutoff agrees well with the
Fig. 6 Probe molecule flux as a function of film thickness for an interfacially polymerized film derived from molecular square 14Zn and succinyl chloride. The line is drawn to show a first order inverse fit to the data. (Inset shows linearity of the reciprocal plot.) Adapted from [28]
Rhenium-Linked Multiporphyrin Assemblies
161
cavity size for an isolated square. For layer-by-layer assembled materials it appears to be smaller. These differences have been ascribed to swelling and flexing of the polymeric material versus rigidity for the zirconium phosphonate linked films. Shown in Fig. 6 is the dependence of molecular flux on polymer film thickness. The observed inverse correlation is characteristic of permeationcontrolled transport. In other words, fluxes are limited by rates of diffusion through the film rather than partitioning from the solution to the film. Similar behavior has been observed for molecular aggregate films. 4.2 Sensing A brief report showed that molecular square 1Zn could be used to detect binding of sodium and potassium cations by a crown ether featuring a pendant pyridine [32]. The crown ether ligand was observed to bind via the pendant pyridine to the available Zn(II) sites of the porphyrin square. Accompanying the binding was a decrease in fluorescence intensity that was reversed upon capture of an alkali metal cation by the crown. Micropatterned porous films of molecular square 1Zn were observed to diffract visible light. Uptake of guests, either in the vapor phase or from aqueous solution, was found to modulate the diffraction efficiency, with the extent of modulation scaling with amount of guest (analyte) taken up as measured independently by quartz crystal microgravimetry [33]. While many of the analytes were sorbed nonspecifically, a few examples of specific binding or molecular recognition were described. Aqueous 2,4,6-tris(4pyridyl),1,3,5-triazine was sensed with patterned films of 1Zn but not with films of 1Zn·TPyP (see Fig. 4 for structure), consistent with uptake based on axial ligation of Zn(II). Molecular square cavity functionalization with tris(aminoethyl)amine yielded films that responded to aqueous Zn2+ . Functionalization with 1,6-hexanedithiol yielded films responsive to molecular iodine, due to iodine/thiol charge-transfer complex formation. 4.3 Catalysis Molecular square 1Zn has been used as an encapsulant for pyridine functionalized manganese porphyrin guests used as catalysts for olefin epoxidation. The protection afforded by guest encapsulation was found to increase the catalyst lifetime by an order of magnitude or more, evidently by inhibiting formation of catalytically inactive oxo-bridged dimers of the manganese porphyrins. The host assembly also engendered substrate size selectivity by preventing access to the active site by large substrate molecules. Modifying the molecular square cavity by ligating pyridine derivatives at two of the
162
J.T. Hupp
Zn(II) sites, while anchoring the catalyst at the other two, resulted in partial tunability of the substrate size selectivity [34]. 4.4 Fundamental Studies of Energy and Electron Transfer Molecular square 1 weakly fluoresces in methylene chloride as solvent [7]. The singlet lifetime is reported to be 2.4 ns. The weak luminescence and relatively short lifetime are presumably consequences of accelerated intersystem crossing caused by proximal heavy atoms (rhenium). When bound as a guest, TPyP is reported to quench ∼ 90% of the luminescence of 1Zn. The mechanism of quenching was not established. Excitation of the metallated porphyrins in planar heterodimers 6c and 7c was observed to sensitize emission from the free-base porphyrins (Scheme 6) [17]. Energy transfer (EnT) rates of about 1 × 1010 s–1 were inferred from quantum yields. A factor contributing to efficient transfer may be the porphyrin–porphyrin coplanarity enforced by rhenium coordination. It was argued on energetic grounds that EnT was not mediated by the Re(I) centers. One-electron reduction of the cofacial assembly 4Zn yields a porphyrinbased mixed-valence compound [35]: The reduced compound displays an intense intervalence band in the near infrared region. The intervalence transition is nominally a symmetrical charge-transfer transition. Electroabsorbance measurements indicated, however, that only a small amount of charge-transfer character occurs, leading the ground-state form of the assembly to be assigned as a borderline valencedelocalized compound. Electronic structure calculations established that electronic coupling occurs almost entirely by direct donor–orbital/acceptor– orbital overlap. Contributions from superexchange coupling through the rhenium centers were estimated to be less than 10 cm–1 . Three-electron reduction also yields a mixed-valence species, but much less strongly electronically coupled as evidenced by the absence of a de-
Scheme 6
Rhenium-Linked Multiporphyrin Assemblies
163
Structure 11
tectable intervalence absorption band. The reduction in coupling was ascribed to a geometry change: coulombic repulsion between singly and doubly anionic redox centers (i.e., porphyrin ligands) was assumed to yield an opencavity geometry featuring minimal donor-orbital/acceptor-orbital overlap. 4.5 Light-to-Electrical Energy Conversion Phosphonated molecular squares have been examined as multilayer sensitizers of indium-tin oxide (conductive glass) surfaces in photoelectrochemical solar cells [36]. Generally, these cells convert energy by the following sequence: (a) dyes are photo-excited, (b) energy is transferred from outer layers to the innermost layer of dye molecules, (c) electrons are transferred from the excited inner layer to the transparent electrode, leaving oxidized dye molecules, (d) dyes are restored to their original form by reduction with a redox shuttle, and (e) the oxidized shuttle molecules diffuse to a dark electrode and are reduced, completing the circuit [37]. Multilayer sensitization typically is only marginally better than monolayer sensitization despite the collection of more photons [38]. One of the problems is believed to be difficulty in moving charge-compensating ions within or through chromophoric layers. In principle, this difficulty could be overcome by using porous molecular square films as light absorbers. A study done in water as solvent with iodide/triiodide as the redox shuttle exhibited the hoped-for systematic increase in photocurrent with increasing number of dye layers [36]. (Cells with 1 to 15 layers were examined.) The
164
J.T. Hupp
Fig. 7 Mechanism of cathodic current generation by 13Zn/ITO multilayer electrodes. Adapted from [36]
direction of current flow, however, was the opposite of what was expected. Further study revealed that the cell operated according to the mechanism shown in Fig. 7. The key feature is direct quenching of the photo-excited porphyrinic square by a pre-associated triiodide ion. Acknowledgements I gratefully acknowledge the contributions of many coworkers whose names are cited in the descriptions of the portions of the reviewed work that were done at Northwestern. I gratefully acknowledge the U.S. Department of Energy, Basic Energy Sciences Offices, for support of our own work on photophysics and energy conversion (grant No. DE-FG87ER13808) and on membranes and molecular sieving (grant No. DE-FG0201ER15244), and the Northwestern Institute for Environmental Catalysis for support of our work on catalysis.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
Giordano PJ, Wrighton MS (1979) J Am Chem Soc 101:2888 Caspar JV, Sullivan BP, Meyer TJ (1984) Inorg Chem 23:2104 Slone RV, Yoon DI, Calhoun RM, Hupp JT (1995) J Am Chem Soc 117:11813 Fujita M (1998) Chem Soc Rev 27:417 Leininger S, Olenyuk B, Stang PJ (2000) Chem Rev 100:853 Benkstein KD, Hupp JT, Stern CL (2000) Angew Chem Int Ed 72:3122 Slone RV, Hupp JT (1997) Inorg Chem 36:5422 Slone RV, Hupp JT, Stern CL, Albrecht-Schmitt TE (1996) Inorg Chem 35:4096 Lee SJ, Lin W (2002) J Am Chem Soc 124:4554 Sun SS, Lees AJ (2000) J Am Chem Soc 122:8956 Keefe MH, O’Donnell JL, Bailey RC, Nguyen ST, Hupp JT (2003) Adv Mater 15:1936 Miljacic L, Sarkisov L, Ellis DE, Snurr R (2004) J Chem Phys 121:7228 Splan KE, Stern CL, Hupp JT (2004) Inorg Chim Acta 357:4005 Rajendran T, Manimaran B, Liao RT, Lin RJ, Thanasekaran P, Lee GH, Peng SM, Liu YH, Chang IJ, Rajagopal S, Lu KL (2003) Inorg Chem 42:6388
Rhenium-Linked Multiporphyrin Assemblies
165
15. Thanasekaran P, Liao RT, Liu YH, Rajendran T, Rajagopal S, Lu KL (2005) Coord Chem Rev 249:1085 16. Benkstein KD, Stern CL, Splan KE, Johnson RC, Walters KA, Vanhelmont FWM, Hupp JT (2002) Eur J Inorg Chem 2002:2818 17. Splan KE, Keefe MH, Massari AM, Walters KA, Hupp JT (2002) Inorg Chem 41:619 18. Fan J, Whiteford JA, Olenyuk B, Levin MD, Stang PJ, Fleischer EB (1999) J Am Chem Soc 121:2741 19. Iengo E, Milani B, Zangrando E, Geremia S, Alessio E (2000) Angew Chem Int Ed 39:1096 20. Santosh G, Ravikanth M (2005) Inorg Chim Acta 358:2671 21. Iengo E, Zangrando E, Minatel R, Alessio E (2002) J Am Chem Soc 124:1003 22. Williams ME, Hupp JT (2001) J Phys Chem B 105:8944 23. Belanger S, Keefe MH, Welch JL, Hupp JT (1999) Coord Chem Rev 192:29 24. Massari AM, Gurney RW, Schwartz CP, Nguyen ST, Hupp JT (2004) Langmuir 20:4422 25. Kaschak DM, Lean JT, Waraksha CC, Saupe GB, Usami H, Mallouk TE (1999) J Am Chem Soc 121:3435 26. Katz HE (1994) Chem Mater 6:2227 27. Libera JA, Gurney RW, Nguyen ST, Hupp JT, Liu C, Conley R, Bedzyk MJ (2004) Langmuir 20:8022 28. Keefe MH, O’Donnell JL, Bailey RC, Nguyen ST, Hupp JT (2003) Adv Mater 15:1936 29. Wamser CC, Bard RR, Senthilathipan V, Anderson VC, Yates JA, Lonsdale HK, Rayfield GW, Friesen DT, Lorenz DA, Stangle GC, Eikren PV, Baer DR, Ransdell RA, Golbeck JH, Babcock WC, Sandberg JJ, Clarke SE (1989) J Am Chem Soc 111:8485 30. Czaplewski KF, Hupp JT, Snurr RQ (2001) Adv Mater 13:1895 31. Czaplewski KF, Li JL, Hupp JT, Snurr RQ (2003) J Memb Sci 221:103 32. Chang SH, Chung KB, Slone RV, Hupp JT (2001) Synth Met 117:215 33. Mines GA, Tzeng BC, Stevenson KJ, Li J, Hupp JT (2002) Angew Chem Int Ed 41:154 34. Merlau ML, Mejia MDP, Nguyen ST, Hupp JT (2001) Angew Chem Int Ed 40:4239 35. Dinolfo PH, Coropceanu V, Bredas JL, Hupp JT, Lee SJ (2005) Inorg Chem 44:5789 36. Splan KE, Massari AM, Hupp JT (2004) J Phys Chem B 108:4111 37. Hagfeldt A, Grätzel M (2000) Acc Chem Res 33:269 38. Taniguchi T, Fukasawa Y, Miyashita T (1999) J Phys Chem B 103:1920
Struct Bond (2006) 121: 167–215 DOI 10.1007/430_019 © Springer-Verlag Berlin Heidelberg 2006 Published online: 4 February 2006
Thermodynamics of Metal-Mediated Assemblies of Porphyrins Gianfranco Ercolani Dipartimento di Scienze e Tecnologie Chimiche, Università di Roma Tor Vergata, Via della Ricerca Scientifica, 00133 Roma, Italy [email protected] 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
168
2 2.1 2.2 2.3 2.4
Thermodynamics of Acyclic Assemblies . . . . . . . Generalities on Metalloporphyrin-Ligand Interactions Closed Acyclic Assemblies and Cooperativity . . . . . Acyclic Polymeric Assemblies . . . . . . . . . . . . . Selected Examples . . . . . . . . . . . . . . . . . . . .
3 3.1 3.2 3.3
Thermodynamics of Cyclic Assemblies Chelate Effect and Effective Molarity . Stability of Cyclic Assemblies . . . . . . Selected Examples . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
4 4.1 4.2 4.3 4.4
Thermodynamics of Multicyclic Assemblies . Stability of Multicyclic Assemblies . . . . . . . Cooperativity in Multicyclic Assemblies . . . . Topologically Reducible Multicyclic Assemblies Selected Examples . . . . . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
5
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
170 170 171 174 177
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
182 182 185 191
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
200 200 203 204 205
Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . .
211
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
212
. . . .
. . . .
. . . .
Abstract An overview of the thermodynamic principles governing self-assembly in solution, with particular reference to multiporphyrin architectures, is presented. The topic is discussed in order of increasing complexity, from simple acyclic assemblies to multicyclic assemblies, by emphasizing the role played by two fundamental physicochemical quantities, namely, the stability constant of the metalloporphyrin-ligand interaction and the effective molarity, the latter being a measure of the ease of formation of a given cyclic structure or substructure. Criteria to assess cooperativity in self-assembly are also discussed. The principles are illustrated by selected examples involving metal-mediated assemblies of porphyrins in which the metal is incorporated into the porphyrin ring. Keywords Cooperativity · Effective Molarity · Multivalency · Porphyrin · Self-assembly Abbreviations b C
number of bonds joining the assembly components total monomer concentration
168 C Copt DP EM EMm i Kinter Kintra Kintra(m) Ksa lsac m n nH OEP p p PyTPP r TPP trans-DPyDPP wsa wsa(max) wi y α σp σr
G. Ercolani total monomer concentration in the chain fraction optimal self-assembly concentration number average degree of polymerization effective molarity microscopic effective molarity running index intermolecular stability constant intramolecular stability constant microscopic intramolecular stability constant stability constant of the supramolecular assembly lower self-assembly concentration number of binding sites of a multivalent ligand number of assembly components (order of the assembly) Hill coefficient octaethylporphyrin extent of reaction extent of reaction in the chain fraction 5-(4-pyridyl)-10,15,20-triphenylporphyrin occupancy 5,10,15,20-tetraphenylporphyrin 5,15-bis(4-pyridyl)-10,20-diphenylporphyrin weight fraction of the supramolecular assembly maximum weight fraction of the supramolecular assembly weight fraction of the linear i-mer degree of saturation cooperativity factor statistical factor of products statistical factor of reactants
1 Introduction Self-assembly consists of the spontaneous generation of ordered supramolecular architectures from a given set of components under thermodynamic equilibration [1–9]. The process, driven by non-covalent interactions, plays a crucial role in determining the “intelligent” behavior observed in the biological world, where huge amounts of information depend ultimately on the shape, size, and binding properties of a limited number of building blocks. The porphyrin ring is one of the most prominent, featuring a wealth of photochemical and redox properties. Not surprisingly, the self-assembly of porphyrin derivatives has attracted considerable attention as a viable route to the synthesis of nanometer scale photonic devices mimicking photosynthetic functions such as light harvesting and charge separation [10–18]. Designed porphyrin assemblies have been formed mainly by hydrogen bonding and metal-ligand coordination; the exploitation of the defined directionality of these bonds has allowed the formation of several topologically complex struc-
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
169
tures which have been characterized in solution and/or in the solid state. Needless to say, structural characterization of self-assembled architectures is inherently difficult [19, 20]. Indeed, owing to the lability of non-covalent interactions, often supramolecular architectures are stable only at a well defined range of reactant concentrations, and, when a single crystal is available, the X-ray structure may be not representative of the structure(s) present in solution. NMR and optical spectroscopic methods, especially UV-vis, are essential to gain structural information in solution and to evaluate stability constants. Sometimes, however, they are insufficient to establish the stoichiometry of oligomeric assemblies. Other techniques, such as vapor phase osmometry, gel permeation chromatography, NMR diffusion experiments, and smooth ionization techniques in mass spectrometry, such as FAB, ESI, and MALDI-TOF, may be helpful to this end. In spite of the successful characterization of many sophisticated and esthetically appealing multiporphyrin assemblies, the theoretical and rational modeling of self-assembly processes is far less developed. In principle, to describe the behavior in solution of a mixture of molecules capable of reversible association with each other, one should know, beside their initial concentrations, the stability constants of all the species that could possibly form upon association. This knowledge, however, is generally lacking, so a number of simplifying assumptions, approximations, and reasonable guesses are necessary to treat such complicated systems. Over of the years, this process has singled out the fundamental features that make self-assembly possible, highlighting concepts, such as preorganization, cooperativity, chelate effect, effective molarity, etc., which are nowadays indispensable tools to understand the dynamic behavior of self-assembling systems, to envisage new building blocks, and to make predictions about the outcome of self-assembly experiments. The aim of this review is to provide a conceptual overview of the principles underlying the thermodynamics of self-assembly in solution and to illustrate them by selected examples involving metal-mediated multiporphyrin assemblies in which the metal is incorporated into the porphyrin ring. The term “multiporphyrin assembly” is used here in the restrictive sense of a supramolecular architecture that is formed upon association of at least two porphyrin units, thus excluding those supramolecular complexes formed by a multiporphyrin receptor bound to a non-porphyrin ligand. The reported examples will be necessarily limited to those cases where more or less detailed data about the stability of the assembly in solution are available. Since one of the requisites of genuine self-assembly is reversibility, this excludes those molecular architectures whose building blocks are held together by thermodynamically very stable metal-ligand bonds, as they tend to be kinetically inert. Among the many metal-ligand combinations, the most successful for the preparation of porphyrin assemblies is the labile Zn(II)-amine interaction, where the amine can be either aliphatic or aromatic (pyridine, imidazole, etc.), thus many of the discussed examples will involve this interaction.
170
G. Ercolani
2 Thermodynamics of Acyclic Assemblies 2.1 Generalities on Metalloporphyrin-Ligand Interactions Typically a metalloporphyrin reversibly adds either one or two ligands depending on the nature of the metal, as schematically shown in the cases (a) and (b) of Fig. 1. If a kinetically inert ligand, e.g. CO, is already axially bound to a metal capable of 6-coordination, only one binding site is available, and this case, omitting the fixed ligand, can be also represented by Fig. 1a. The geometry of the complex is determined by the orthogonal orientation of the electron pair of the ligand’s donor atom to the average plane of the porphyrin ring. Metal-ligand bonds cover a large range of energies depending on the nature of the metal and the ligand. For several combinations, the addition is rapid, simple, and reversible. Since the stability of any assembly is strongly correlated to the strength of the bonds holding its constituent building blocks together, knowledge of the relevant association constants is of fundamental importance for any physicochemical discussion. The reader is addressed to the contributions of Stulz, and Kobuke in this volume, and to a previous review of Sanders et al. [21], for adequate information about the axial coordination chemistry of metalloporphyrins. An extensive collection of stability constants by Tabata and Nishimoto is also available [22]. Owing to the importance of the Zn(II)-amine interaction, however, its principal characteristics will be briefly illustrated here. Zincporphyrins only add one nitrogen ligand, the addition being fast and reversible. The association constants are primarily determined by the basicity and steric hindrance of the amine. For example the order of affinity of typical amines toward zinc 5,10,15,20-tetraphenylporphyrin
Fig. 1 Coordination processes of a monodentate (a) and a bidentate (b) metalloporphyrin with a labile ligand
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
171
Table 1 Equilibrium constants, enthalpies, and entropies for the addition of pyridine to ZnTPP in various solvents at 298 K Solvent
K (mol–1 L)
∆H ◦ (kcal mol–1 )
∆S◦ (cal mol–1 K–1 )
Refs.
Cyclohexane Methylene chloride Toluene Benzene Chloroform
25 100 6900 6000 3900 610
– 10.0 ± 0.2
– 13.5 ± 0.7
– 8.8 ± 1.2 – 4.0 ± 0.4
– 13 ± 5 0±3
[26] [27] [25] [28] [28]
(ZnTPP), or analogous porphyrin derivatives, in toluene or benzene at 298 K, is (log K in parentheses): DABCO (5.2) [23] > piperidine (4.8) [24] ∼ imidazole (4.7) [25] > propylamine (4.3) [24] > pyridine (3.8) [25] > dipropylamine (3.1) [24] > triethylamine (1.1) [23]. The effect of basicity is evident in the comparison between imidazole and pyridine; the compounds have similar steric hindrance, but the more basic imidazole has an affinity about 10 times greater. In contrast, though propylamine is much more basic than imidazole, its affinity is slightly lower because of the greater steric hindrance. The depressing effect of steric hindrance increases on passing from primary to secondary, to tertiary amines, as exemplified by propylamine, dipropylamine, and triethylamine. However, the steric hindrance of secondary and tertiary amines is greatly reduced if the amine is cyclic, as dramatically illustrated by the comparison of DABCO and triethylamine. The stability constants are significantly affected by the nature of solvent and temperature as well; for the sake of illustration, in Table 1 are reported the association constants of ZnTPP with pyridine in various non-coordinating solvents at 298 K, together with the available thermodynamic parameters. The constants decrease in the order cyclohexane > methylene chloride ∼ toluene ∼ benzene > chloroform. The weaker binding in chloroform can be attributed to its ability to solvate pyridine by hydrogen bond. This would also be apparent by the higher entropy of complexation in chloroform indicating a larger desolvation of the ligand. In all cases the reaction is exothermic; then, according to the van’t Hoff equation, K increases with decreasing temperature, suggesting that Zn(II)-pyridine based assemblies will be more stable at low temperature. This behavior, in the absence of significant strain effects (Sect. 3.1), is expected to be general for metal-mediated porphyrin assemblies. 2.2 Closed Acyclic Assemblies and Cooperativity Porphyrins and metalloporphyrins can be exploited in two ways as modules for the construction of supramolecular assemblies. Porphyrins can act
172
G. Ercolani
as donor building blocks provided that the periphery has ligands that can suitably coordinate to metal centers, whereas metalloporphyrins can act as acceptor building blocks provided that the metal atom inside the porphyrin core has at least one axial site available for coordination. Fully saturated (closed) acyclic assemblies are formed when at least one of the building blocks is monodentate. Some examples are schematically shown in Fig. 2. The cases (a) and (d) do not differ substantially from the cases in Fig. 1 and thus will not be discussed further. More interesting are the cases in which a multitopic ligand interacts with a monodentate metalloporphyrin. If the donor groups in the ligand are identical and there is no interaction among them, the stepwise association constants are easily predictable by the microscopic equilibrium constant of the metal-donor interaction (Fig. 1a) and the appropriate statistical factor. For example, consider the assembly in Fig. 2e; the corresponding stepwise association equilibria are described by Eq. 1, where A3 is
Fig. 2 Cartoons showing closed acyclic assemblies in which at least one of the building blocks is monodentate
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
173
the tritopic ligand and B is the metalloporphyrin.
(1) If K is the microscopic equilibrium constant for the association of the monotopic ligand A with B, the stepwise constants are: K1 = 3K; K2 = K; K3 = K/3. The statistical factors are easily understood: in the first equilibrium, three equivalent binding sites A are available for binding; in the second, two equivalent binding sites A are available in the forward reaction but two equivalent bonds A—B can dissociate in the reverse reaction; in the third, there are three equivalent bonds A—B which can dissociate in the reverse reaction. In general, for the interaction of a symmetrical m-topic ligand with a monotopic metalloporphyrin, the stepwise constants Ki are given by Eq. 2, from which Eq. 3 is easily derived [20, 29–31]. Ki = K(m – i + 1)/i Ki+1 i(m – i) = Ki (i + 1)(m – i + 1)
(2) (3)
The ratio Ki+1 /Ki , given by Eq. 3, provides a criterion to assess cooperativity: if the equality in Eq. 3 is satisfied the binding is non-cooperative or statistical; if the binding of a B unit favors the binding of a subsequent B unit, the experimental ratio of Ki+1 /Ki is larger than the statistical ratio predicted by Eq. 3, then we speak of positive cooperativity; on the contrary, if the former ratio is smaller than the latter, indicating that the binding of a B unit impedes the binding of a subsequent B unit, the binding is negatively cooperative (or anticooperative). In the specific case m = 2, where in the absence of cooperativity K2 /K1 = 1/4, an interaction parameter (or cooperativity factor) α = 4K2 /K1 can be defined, which provides a quantitative measure of cooperativity; in the three cases above, α being = 1, > 1, < 1, respectively [29]. There are other equivalent tests to assess cooperativity, mainly graphical ones, based on the calculation of the occupancy r, that is to say the average number of occupied sites of the m-topic ligand [20, 29–31]. A plot of r/[B] as a function of r is known as the Scatchard plot; non-cooperative behavior is characterized by a straight line (Eq. 4), whereas a concave downward curve or a concave upward curve are diagnostic for positive or negative cooperativity, respectively. r = – Kr + mK (4) [B] Analogously, a plot of log[y/(1 – y)] vs. log[B], known as the Hill plot, where y is the degree of saturation (r/m), is linear with unit slope in the absence of cooperativity (Eq. 5), whereas the presence of cooperativity is evidenced by two lines of unit slope connected by an S-shaped curve. The value of the slope
174
G. Ercolani
in the central region of the curve is called the Hill coefficient (nH ). Values of nH > 1, and < 1 are diagnostic for positive, and negative cooperativity, respectively. Since nH can vary between 0 and m, it provides a quantitative measure of cooperativity. y = log[B] + log K (5) log 1–y It is worth remarking that the use of Eq. 3, as well as the Scatchard plot and the Hill plot, must be confined to the case in which a symmetric multitopic donor (or acceptor) interacts with a monotopic acceptor (or donor). Use of these methods to assess cooperativity outside this specific case is improper and has been criticized [31]. 2.3 Acyclic Polymeric Assemblies Polymeric assemblies are formed upon reaction of multidentate building blocks assembling in a linear or almost linear fashion. The most common cases involve either a single building block of the type A—B or an equimolar mixture of two building blocks of the type A—A and B—B (Fig. 3). In case (a), the building block A—B is constituted by a metalloporphyrin covalently bound to a suitable ligand yielding a module capable of acting simultaneously as a donor and acceptor so as to form a homo-polymer; in the case (b), a hetero-polymer is formed from a bidentate ligand (A—A) and a bismetalloporphyrin (B—B); in the case (c), a bidentate ligand (A— A) and a porphyrin with a 6-coordinate metal give rise to a hetero-polymer dubbed “shish-kebab”, because of its shape. In all of the cases a mixture of oligomers is formed whose distribution depends on the concentration of the monomeric building blocks and the relevant association constant(s). The principle that the thermodynamic reactivity of end groups is independent of the length of the chain to which they are attached is the foundation for theoretically derived molecular weight distribution relationships [32]. Accordingly, for monomers of the type A—B, the overall equilibrium between end groups is given by Eq. 6, where K is the microscopic association constant (Fig. 1a), C is the total monomer concentration (equal to the initial concentration of end groups) and p is the extent of reaction at equilibrium.
(6) The equilibrium constant K is then given by Eq. 7. This equation can be rearranged to give Eq. 8 which gives the extent of reaction as a function of K and
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
175
Fig. 3 Homo-polymers (a) and hetero-polymers (b and c) by metal-mediated porphyrin self-assembly
C. It simplifies when (KC)1/2 1. p K= C(1 – p)2 √ 1 – 1 + 4KC 1 p=1+ ≈1– √ 2KC KC
(7) (8)
The number average degree of polymerization, DP, and the weight fraction (the yield) of a given linear i-mer, wi , as a function of p, are those characteristic of the most probable distribution (Eqs. 9 and 10) [32]. 1 1–p wi = i(1 – p)2 pi–1
DP =
(9) (10)
In Fig. 4 is reported the average degree of polymerization as a function of the product KC in the range 1 ≤ KC ≤ 103 . For KC values greater than 103 , the approximate expression DP = (KC)1/2 is accurate enough. It is evident
176
G. Ercolani
Fig. 4 Average degree of polymerization vs. KC
from the plot that KC represents the driving force for the process of polymerization. From Fig. 4 it appears that for K < 105 mol–1 l, corresponding to the typical Zn(II)-amine interaction, only relatively short oligomers occur in solutions more dilute than 10–2 mol l–1 (DP < 32). To obtain high DP values, it is necessary to reach association constant higher than 108 –109 mol–1 l; this can be done by designing molecular architectures involving multicyclic assemblies (Sects. 4.3 and 4.4). In Fig. 5 is reported the weight fraction of the first five oligomers as a function of KC. Figure 5 shows that, although each oligomer presents a rather well defined range of monomer concentration in which its weight fraction is maximal, the selectivity for a specific oligomer is always rather low and tends to worsen when the degree of polymerization increases. It has been suggested that a strong bias in favor of a well defined linear oligomer in solution can be obtained by adding calculated amounts of monofunctional reactants to the oligomeric mixture [33]. However, the suggestion has been criticized on the basis of a rigorous thermodynamic analysis showing that the addition of monofunctional reactants, apart from providing a cap to the ends of oligomers, is analogous to dilution; it simply involves a shift to the left in the abscissa of Fig. 5, i.e. depolymerization [34]. The above molecular distribution also applies to the case A—A + B—B provided that the concentrations of the two monomers are exactly equal. Of course, while in the case A—B, C coincides with the initial monomer concentration, in the case A—A + B—B, C is the sum of the initial concentrations of the two monomers. It must also be remarked that in the case A—A + B—B,
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
177
Fig. 5 Weight fractions of linear oligomers, from monomer (a) to pentamer (e), vs KC
there are two types of oligomers with the same odd number of units, namely those having A groups at both ends and those having B groups at both ends; thus when i is odd, the weight fraction given by Eq. 10 represents the sum of the weight fractions of the two types of oligomers. In the case (c) of Fig. 3, the bidentate metalloporphyrin behaves as a monomer of the type A—A, with two distinct association constants (Fig. 1b). If the two constants, corrected for the statistical factors, are largely different, the above analysis cannot be applied since it is based on the principle of equal thermodynamic reactivity of end groups. In this case the process of polymerization will be essentially directed by the lower constant. However, if the two constants are of the same order of magnitude, the above theoretical analysis can still be applied, although with a greater approximation, by defining an average microscopic constant K = (K1 K2 )1/2 . 2.4 Selected Examples Among the simple acyclic assemblies, the most investigated one is the 2 : 1 zincporphyrin-DABCO system forming a sandwich motif that has been exploited to assemble a number of complex architectures [35–41]. The dynamics of the system can be described by taking into account the equilibria shown in Fig. 6. Consider the reaction of zinc 5,10,15,20-tetra(4-npentylphenyl)porphyrin with DABCO in CHCl3 , recently studied by Ballester,
178
G. Ercolani
Fig. 6 Assembly (a) and disassembly (b) of a zincporphyrin-DABCO system
Hunter et al. [40, 41]. At micromolar concentrations of porphyrin, only the 1 : 1 complex is formed, and so UV-vis titrations were used to determine the value of K1 = 1.8 × 105 mol–1 l, from which the microscopic constant K = K1 /2 = 8.9 × 104 mol–1 l is calculated. When one DABCO nitrogen is bound to a zincporphyrin, the affinity of the second nitrogen for another zincporphyrin may be affected, i.e. the two binding events may be not independent. The cooperativity factor α is used to quantify this effect, and its value can readily be determined from 1 H NMR titrations. At millimolar concentrations of porphyrin, until 0.5 equiv of DABCO is added, only the 2 : 1 sandwich complex is formed, but this opens up to form the 1 : 1 complex with excess DABCO. The association constant for formation of the 2 : 1 complex is too large to be measured at these concentrations, but the equilibrium constant for breakdown of the sandwich complex (K3 = K1 /K2 = 4/α) is weaker and can be measured by monitoring the changes in chemical shift. This experiment gave a value of α = 0.8 ± 0.2 in CDCl3 , indicating a very small, if any, anticooperative effect. An analogous α value (0.77) was obtained by 1 H NMR titrations of rutenium(II) carbonyl mesoporphyrin-II-dimethyl ester with DABCO in CH2 Cl2 [36]; in this case all the species were in slow exchange and their concentrations could be directly determined from the 1 H NMR spectrum. Lower α values, 0.20 in CHCl3 and 6.0 × 10–3 in toluene, were obtained in the presence of two bulky 3,5-di-tert-butylphenyl groups attached at the 5,15 meso positions of the porphyrin ring (see structure 32 with i = 1 in Sect. 3.3) [38]. Such low values were ascribed to steric interactions between the aryl groups
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
179
in the sandwich complex; however, the significant role played by the solvent is still not clear. Reek et al. reported binding studies in toluene of the trivalent ligands tris(3-pyridyl)phosphine [42] and tris(4-pyridyl)phosphine [43] with ZnTPP to yield adducts of the type schematized in Fig. 2e. In the first case the stepwise binding constants were K1 = 3.7 × 103 mol–1 l, K2 = 7.8 × 103 mol–1 l, K3 = 1.2 × 104 mol–1 l. Both the experimental ratios K2 /K1 (= 2.1) and K3 /K2 (= 1.5) are larger than the expected statistical ratio 1/3 evaluated by Eq. 3, thus indicating the action of positive cooperativity. This was ascribed, as suggested by molecular modeling, to favorable π – π interactions between the meso-phenyl rings of the porphyrins associated to the template. In the second case the stepwise binding constants were K1 = 2.1 × 103 mol–1 l, K2 = 2.3 × 103 mol–1 l, K3 = 1.9 × 103 mol–1 l. Also, in this case, the ratios K2 /K1 (= 1.1) and K3 /K2 (= 0.8) indicate, unlike the conclusion of the authors, a positively cooperative binding, although the effect is smaller than that of the first template. There is an obvious interest in building multiporphyrin assemblies of increasing order to mimic the function of the antenna complex in the photosynthetic light-harvesting system which is one of the largest porphyrin assemblies in nature. One of the exploited motifs is that of a multivalent ligand capable of coordinating a large number of porphyrin units to its periphery. Interesting examples are those reported by Kuroda et al. in which a core porphyrin is functionalized with 4 (1), and 8 (2 and 3) pyrazine subunits, as schematically illustrated in Fig. 7 [44, 45]. Each pyrazine is capable of coordinating a zincporphyrin dimer, 4, whose monomeric units, constituted by a zincporphyrin bearing at the meso position four 2-carboxy4-nonylphenyl groups, are held together by hydrogen bonding between the carboxylic groups. The association constant between pyrazine and the dimer is of the order of 107 mol–1 l. The process of assembly formation of the ligands 1–3 with the dimer 4 showed that each pyrazine subunit behaves independently from the others, with no sign of cooperativity. The experiments further indicated that, at the given experimental concentrations, by adding an equivalent amount of dimer to each ligand, a mixture of adducts is obtained in which the fully saturated assembly is the dominant species. Also successful was the use of N-methylimidazole-functionalized gold nanoparticles as multivalent ligands [46]. Two monolayer protected gold nanoclusters were prepared with mixed monolayers composed of a 1+1 (5) and 4+1 (6) mixture of dodecanethiolates and thiolates functionalized with N-methylimidazole (Fig. 8). The binding constant of the model methylimidazole derivative 7 to ZnTPP (K = 3.1 × 104 mol–1 l in CH2 Cl2 at 25 ◦ C) is slightly larger than the binding constants measured for the two clusters under the same conditions (K = 1.1 × 104 mol–1 l for 5, and K = 1.4 × 104 mol–1 l for 6), suggesting some steric hindrance when the porphyrins approach the nanoparticle’s surface. Preliminary experiments carried out by saturating the
180
G. Ercolani
Fig. 7 Schematic representation of the multivalent ligands 1–3 capable of forming nonameric and heptadecameric porphyrin assemblies upon binding of the dimer 4
Fig. 8 Functionalized gold nanoparticles and reference ligand
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
181
monolayer surface of 5 with ZnTPP indicate that ca. 70% of the imidazoles is available for binding. As to the thermodynamically controlled formation of acyclic polymeric assemblies in solution, there are only a few examples in the literature. The main problems are due to: a) competition with cyclization processes; b) low association constants; c) low solubility of the oligomers. In fact the first reported polymeric assembly in solution made of zinc 5-(4-pyridyl)-10,15,20triphenylporphyrin (ZnPyTPP) units [47], actually consists, as successively demonstrated, of a cyclic tetramer (Sect. 3.3). The linear polymer, however, was unambiguously detected in the solid state by X-ray analysis of a ZnPyTPP single crystal [47]. This is one of the first examples in the literature of metalloporphyrins, illustrating the fact that the stability of an assembly may be strongly dependent on the aggregation state. An interesting case in which the assembly properties can be regulated by light is that regarding the zincporphyrins 8 and 9 [48]. According to Burrell et al. the trans form 8 assembles into a linear polymer whereas the cis form 9 yields a cyclic dimer. Conversion of 8 into 9 occurred upon irradiation with UV light, thus switching the geometry of the assembly in CHCl3 solution from linear to cyclic. However, considering the strength of the association of pyridine with zincporphyrins in CHCl3 (Table 1), the DP value of 8 at millimolar concentrations, interpolated from Fig. 4, is very low (< 3). Hunter et al. reported the formation of a polymer in CHCl3 from a zincporphyrin bearing one para-aniline group at the meso position [49]. The formation of the polymer was deduced by broadening of the 1 H NMR signals at a concentration as high as 2.5 × 10–2 mol l–1 , that, considering an association constant K = 190 mol–1 l, translates into a DP value of ca. 3. A hetero-polymer formed in CH2 Cl2 solution at 25 ◦ C by the rigid biszincporphyrin 10 and the short flexible bispyridine 11 was reported by Twyman and King [50]. This is an example of an A—A + B—B polymerization in which the competition with cyclization is reduced by the use of a flexible monomer. A binding constant K = 2500 mol–1 l was measured with the reference ligand 12, therefore an equimolar solution of 10 and 11, each at a concentration of 1 × 10–2 mol l–1 , was estimated to produce a hetero-polymer
182
G. Ercolani
with a DP value of ca. 7. However the presence of rings in solution cannot be excluded, as predicted by the Jacobson–Stockmayer theory [51, 52]. To the best of our knowledge there are no reports of shish-kebab oligomers obtained in solution under thermodynamically controlled conditions. Typically the oligomers precipitate from the solution after stoichiometric amounts of the reagents are mixed. For example a series of shish-kebab polymers, [M(OEP)(L – L)]i , (M = Ru, Os; OEP = octaethylporphyrin; L – L = pyrazine, 4,4 -bipyridine, DABCO), were synthesized by dissolving the dimers [M(OEP)]2 in toluene, then adding 1 equiv of the bridging ligand, and recovering the precipitate by filtration [53]. Typical values of DP were 20–25.
3 Thermodynamics of Cyclic Assemblies 3.1 Chelate Effect and Effective Molarity Intramolecular reactions are often more favored than analogous intermolecular reactions. This advantage is known either as the proximity effect or as the chelate effect. To set it on a quantitative scale, the effective molarity (EM) parameter has been devised [54–56]. For thermodynamically controlled processes, the EM is defined by Eq. 11, where Kintra and Kinter are the equilibrium constants for the intramolecular process leading to a cyclic species and the corresponding intermolecular process leading to a linear adduct (Fig. 9). EM = Kintra /Kinter
(11)
The EM represents the concentration of one of the reactants, -A or -B, considered in large excess with respect to the other, needed for the intermolecular process to proceed with the same extent of reaction of the intramolecular process. More generally, it represents an intramolecular reactivity corrected for the inherent reactivity of end groups. It is rather common in the self-assembly literature to speak of the chelate effect in terms of cooperativity. We strongly
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
183
Fig. 9 An intramolecular process (a) and the analogous intermolecular process (b)
argued against this usage for the following reasons [31]: a) cooperativity is intended as a deviation from the statistical behavior of multiple recognition sites non-interacting with each other (Sect. 2.2); b) cooperativity is independent of concentration. Obviously before claiming the presence of cooperativity, one should define the expected behavior in the absence of cooperativity, but there is no rational way to predict the constant Kintra in the supposed absence of cooperativity from the knowledge of Kinter , for the fundamental reason that intramolecular and intermolecular processes are radically different from each other. This is also evident from the different dimensions in which the constants in Fig. 9 are expressed. Moreover an intramolecular process is favored over the intermolecular one, only if the concentration of end groups is lower than the EM; at higher concentrations the opposite occurs. According to the classical analysis of Page and Jencks [57–59], the advantage of an intramolecular process with respect to the analogous intermolecular one depends mainly on entropy. A free molecule has 3 degrees of translational freedom and three degrees of overall rotational freedom. When two molecules condense to form one, 3 degrees of each are converted into 6 degrees of vibrational freedom having lower entropy content; thus an intermolecular reaction proceeds with a significant entropy loss. In contrast, if two groups are already tightly bound to each other, formation of an additional bond proceeds without any appreciable entropy loss. Thus, in the absence of solvent effects, the maximum advantage of an intramolecular reaction is given by the entropy loss suffered by the analogous intermolecular reaction. Page and Jencks estimated that this advantage, for bond forming reaction commonly encountered in organic chemistry, is about 35 cal mol–1 K–1 on the basis of 1 mol l–1 standard state, corresponding to a maximum EM ≈ 108 mol l–1 . This value should be lower for metal-ligand bonds because they are generally looser than ordinary covalent bonds in organic molecules; indeed low-frequency vibrations of metal-ligand bonds and other motions make a significant contribution to the internal entropy of chelate complexes. For example, in the case of the interaction of pyridine with
184
G. Ercolani
a zincporphyrin, the lowest ∆S◦inter value recorded in Table 1 is that obtained in cyclohexane (∆S◦inter = – 13.5 cal mol–1 K–1 ); by assuming that in this case the effect of solvent is negligible, a maximum EM of about 103 mol l–1 can be estimated. Usually EMs are lower than these limit values because the intramolecular process of cyclization may involve a loss of entropy due to restriction of torsional motions [60], and a gain in enthalpy due to the possible build up of strain in the formed ring. Indeed applying thermodynamics to Eq. 11, Eq. 12 is obtained, showing that the EM has both an enthalpic and an entropic component. 0 0 0 ∆Sintra – ∆S0inter – ∆Hinter ∆Hintra exp (12) EM = exp – RT R In most of the cases the enthalpic component coincides with the strain energy of the ring; so the first exponential is lower than 1 unless a strainless ring is formed, in which case, it is equal to 1 and the EM is solely dependent on entropy. Besides the classical sources of strain, the enthalpic component can be affected by attractive or repulsive interactions taking place among the subunits of an oligomeric assembly as long as these are electrically charged or possess strong dipole moments. Although the EM is usually independent of the nature of solvent [55], the presence of these interactions can make the EM solvent dependent. Of course, the enthalpic component determines the behavior of the EM on changing temperature; the EM increases with increasing temperature if the ring is strained whereas it is independent of temperature for strainless rings. The entropic component decreases on increasing the number of rotatable bonds in the linear precursor. It also decreases because of the adverse statistical factor n involved in the formation of a cyclic n-mer. Indeed considering the equilibrium in which the ends of a linear n-mer, e.g. (A—B)n , react to give a cyclic n-mer, c – (A—B)n , it is evident that in the reverse reaction there are n equivalent bonds – AB – that can undergo fission, thus the equilibrium is disfavored by a factor n. It is sometimes convenient to treat separately this statistical factor by defining a statistically corrected (or microscopic) constant Kintra(m) = nKintra , and consequently a microscopic EMm = nEM = Kintra(m) /Kinter . The theoretical evaluation of EMs is a challenge that should be faced by theoretical and computational chemists. Evaluation of the strain of a ring actually presents no problems both with molecular mechanics and quantum mechanical methods. In contrast, the accurate estimation of entropy by molecular modeling is still problematic [61], mainly because low frequency vibrations and internal rotations, possessing the higher entropic content, have nonharmonic character. Theory is well developed for long flexible chains, say more than 25–30 skeletal bonds, yielding strainless rings; it predicts that the EMm of an n-meric chain is proportional to n–3/2 , the factor relating to the probability that a Gaussian chain of n repeating units has its ends coincident [51, 62]. No general theory is available for shorter chains, although an
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
185
approximate empirical relationship yielding the entropic component of EMm as a function of the number of rotatable bonds has been established by Mandolini [54, 55, 63]. This relationship, however, has been tested only in the case of organic rings mostly made of methylenic units; outside this case its validity is not warranted. Obviously the stability of a cyclic assembly will benefit from a large EM. As a general rule, the EM tends to dissipate as it is shared out among more and more components, because of increasing low frequency torsions and skeletal vibrations. Thus, it is easier to reach high EMs with lower order assemblies. Apart from this general consideration, the EM can be maximized by preorganization, consisting of the preparation of rigid, possibly small, building blocks, with no rotatable bonds, whose shapes exactly match each other so as to give a cyclic assembly devoid of strain. Preorganization is not without risks; if the match is not accurate enough the built-in rigidity of the building blocks may become a handicap giving rise to severe strain. In spite of accurate design, however, sometimes molecules are flexible enough to prefer assembling into a strained lower order assembly, because of the inherent entropic advantage, rather than assembling into an assembly of higher order devoid of strain [5, 64]. In the case in which accurate preorganization of the building blocks is difficult, predisposition may be preferable. This has been defined as a strong conformational or structural preference expressed by the building block once incorporated into a larger structure, which gives rise to a thermodynamic preference for a particular product [65, 66]. This statement implies that the building block has a certain degree of conformational mobility and that the final conformation in the cyclic assembly is not the same as, or dictated by, its ground-state conformation. An additional factor that can contribute to the value of the EM is the stabilization provided by a template. Indeed if an assembly can serve as a host for a guest molecule, it will be selectively stabilized by host-guest interactions. This is the physical principle at the basis of the adaptive behavior of dynamic combinatorial libraries [66–68]. 3.2 Stability of Cyclic Assemblies A cyclic assembly can be considered as a cyclic oligomer formed by monomers either of the type A—B or of the type A—A + B—B as schematically shown in Fig. 10. The formation of a cyclic n-mer requires n – 1 intermolecular bonds and one intramolecular bond. If all the n – 1 intermolecular processes occur with the same constant Kinter , the stability constant of the supramolecular assembly is given by Eq. 13, where σr and σp are statistical factors; the former accounting for the number of equivalent binding sites of the reactants [σr = 1 for the case (a), and σr = 2n/2 2n/2 = 2n for the case (b)], and the latter accounting for the number of equivalent bonds in the product
186
G. Ercolani
Fig. 10 Formation of cyclic n-meric assemblies from monomers of the type A—B (a), and A—A + B—B (b)
that can dissociate in the reverse reaction [σp = n for both the cases (a) and (b)]. σr n–1 n Kintra(m) = σr Kinter EM (13) Ksa = Kinter σp Since it appears that there is no general consensus about the need for the factor σr = 2n for the case A—A + B—B (see, for example, [39–41]), we feel necessary to provide an independent proof of its necessity, starting from the general consensus about the validity of Eq. 13 when the monomer is of the type A—B and σr = 1 [69]. Consider the virtual thermodynamic cycle reported in Fig. 11. The vertical equilibria are hypothetical processes in which the end groups of the monomers are exchanged. The stability of a cyclic n-mer does not depend whether its monomeric constituents are of the type A—B or A—A + B—B, thus the equilibrium constant at the right is equal to 1. As to the equilibrium at the left, if we consider a process in which the end groups are statistically scrambled, we obtain at equilibrium: 25% A—A, 25% B—B, and 50% A—B, corresponding to an equilibrium constant equal to 2n , thus unequivocally demonstrating the correctness of the factor above. The process of self-assembly of a specific cyclic oligomer with degree of polymerization n occurs in competition with the formation of other cyclic oligomers and with the process of linear polymerization. Self-assembly takes place when the monomer has a structure preorganized or predisposed in such a way that formation of all the cyclic oligomers except the n-mer is prevented by strain or, in the case of high order cyclic oligomers, by unfavorable entropic effects. This is the first condition to be met for selective self-assembly. However, to evaluate the range of monomer concentration in which a cyclic assembly is stable, the competition with the process of linear polymerization must be taken into account [70]. Focusing on the case A—B, when the
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
187
Fig. 11 A virtual thermodynamic cycle aimed at demonstrating the correctness of the statistical factor σr = 2n in Eq. 13 for the case A—A + B—B
monomer has attained the equilibrium in solution, its total concentration C is distributed between the cyclic n-mer, c – (A—B)n , and the linear polymeric fraction, C (Eq. 14). C = n [c – (A – B)n ] + C
(14)
Considering Eq. 7, Eq. 15 is obtained, where p is the extent of reaction in the linear fraction. C =
p Kinter (1 – p )2 1
(15)
The equilibrium concentration of the monomer A—B is given by the product w1 C , where w1 is its weight fraction at equilibrium given by Eq. 10, so Eq. 16 is easily obtained. [A – B] =
p Kinter
(16)
For the case A—B, Eq. 13 can be rewritten as Eq. 17, which on substitution of Eq. 16 becomes Eq. 18 n EM[A – B]n [c – (A – B)n ] = Kinter [c – (A – B)n ] = EMpn .
(17) (18)
188
G. Ercolani
Substituting Eqs. 15 and 18 into Eq. 14, the following mass balance equation as a function of p can be written p n C = EM np + . (19) EMKinter (1 – p )2 The weight fraction of the supramolecular assembly, given by n[c – (A—B)n ]/C, is expressed, taking into account Eqs. 18 and 19, by Eq. 20. –1 1 . (20) wsa = 1 + nEMKinter (1 – p )2 pn–1 Equations 19 and 20 can be used to build plots of wsa vs C/EM for a given n, that only depend on the product EMKinter . This is done by compiling a list of p values in the range 0 ≤ p < 1 and then calculating, for a given value of the product EMKinter , the couple of C/EM and wsa corresponding to each p , by Eqs. 19 and 20, respectively. Such plots are shown, by way of illustration, in Fig. 12 for the cases n = 4 and EMKinter = 10, 25, 100, and 1000, respectively. It is apparent from Fig. 12 that the driving force for self-assembly is given by the product EMKinter ; a large value of EMKinter is beneficial not
Fig. 12 Plots of the weight fraction of a self-assembled cyclic tetramer vs C/EM. The four curves refer to EMKinter = 10 (a), 25 (b), 100 (c), and 1000 (d)
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
189
only to the weight fraction of the self-assembling macrocycle but also to the amplitude of the concentration range over which self-assembly takes place. The plots in Fig. 12 illustrate the fact that a self-assembly curve goes through a maximum. It has been demonstrated that the height of the maximum is approximately given by Eq. 21, whose plot (Fig. 13) is useful to evaluate the maximum weight fraction of a cyclic assembly obtainable with a given driving force [70]. –1 n (21) wsa(max) ≈ 1 + 1.85 EMKinter Self-assembly is virtually complete when the weight fraction of the cyclic n-mer is greater than 0.99. This can occur in a certain concentration range only if the condition in Eq. 22 is satisfied [70]. EMKinter ≥ 185n
(22)
This is the second condition for self-assembly, namely the condition of stability of the self-assembling macrocycle over the acyclic oligomers. This is an important relation showing that the required driving force for self-assembly directly depends on the number of monomer units constituting the cyclic
Fig. 13 Plot of the maximum weight fraction of a cyclic n-meric assembly as a function of EMKinter /n
190
G. Ercolani
oligomer. The plots reported in Fig. 12 nicely illustrate this point; only for the plot relative to EMKinter = 1000 the condition in Eq. 22 is satisfied. Figure 12 also shows that the value of the monomer concentration C is crucial to determine the stability of a cyclic assembly in solution. If the concentration is too low, the assembly lacks the driving force to form and thus only low order linear oligomers will be present in solution; on the other end, if the concentration is too high, the cyclic assembly loses the competition with linear polymer. It has been demonstrated that the optimal monomer concentration, corresponding to the maximum of the weight fraction of the assembly, is approximately given by Eq. 23, wsa(max) being given by Eq. 21 [70]. Note that when the condition in Eq. 22 is satisfied, Copt ≈ 0.1nEM. Copt ≈ 0.1
nEM wsa(max)
(23)
Provided that Eq. 22 holds, the next question regards the monomer concentration at which self-assembly becomes significant. This is the so-called lower self-assembly concentration (lsac), that is the initial monomer concentration at which half of the monomer is assembled into the cyclic n-mer (Eq. 24). 2 (24) lsac = n/(n–1) n1/(n–1) EM1/(n–1) Kinter The expression in Eq. 24 is due to a slight revision [70] of a previous expression published by Hunter et al. as the csac (critical self-assembly concentration) [69], and is based on the approximation that the chain fraction is essentially constituted by the monomer only. Equations 19–24 also hold for the case A—A + B—B (Fig. 10b) provided that the concentrations of the two monomers are exactly equal and the total monomer concentration is taken as the sum of the initial concentrations of the two monomers. It must be noted that in the case A—A + B—B, only assemblies in which n is an even number are allowed. In conclusion, a large EM value is fundamental for the selective formation of a cyclic architecture. It allows the assembly to win the competition with both the other cyclic oligomers and the linear polymer. In this respect it must be noted that self-assembly of high order cyclic oligomers is hampered by inherent difficulties. Indeed preorganization of monomers to give selectively a specific cyclic oligomer is increasingly difficult, if not impossible, on increasing the order of the assembly. Moreover the condition in Eq. 22, to maintain its validity on increasing n, would require that the EM increases at least proportionally to n, implying that EMm should increases at least proportionally with the square of n; in fact, EMm decreases on increasing n. A large value of Kinter can provide part of the driving force; but, apart from the fact that the effect of Kinter is unselective, it cannot be increased above certain limits because large Kinter values tend to slow down the equilibration process.
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
191
3.3 Selected Examples Since the highest EMs are reached by assemblies of the lowest order, it is not surprising that there are many examples of cyclic dimers. These have attracted considerable attention as biomimetic models of the special pair of bacteriochlorophylls which plays a key role in the photosynthetic process. A notable example regards the zinc and magnesium complexes of a porphyrin bearing two N-methylimidazoles at the 5,15-meso positions; the two N-methyl groups being oriented either cis or trans with respect to the plane of the porphyrin ring. The trans-zincporphyrin forms a slipped cofacial cyclic dimer (13) in CHCl3 by coordination of the imidazolyl groups to the zinc centers [71]. The self-organization is maintained even at a concentration as low as 10–9 mol l–1. This observation and competitive titration with Nmethylimidazole indicated that Ksa > 1010 mol–1 l. Considering that the intermolecular association of imidazole to a zincporphyrin in CHCl3 is of the order 104 mol–1 l [72], an EM > 102 mol l–1 can be estimated by Eq. 13. Of course such a large EM value finds justification also in the attractive π – π interactions between the two porphyrin rings. The cis-zincporphyrin forms an analogous dimer (15) even more stable than the trans [73]. Analogous arrangements were found for the corresponding magnesium complexes 14 and 16 [73]. The cis dimer 16 is the dominant species at the concentration level of 10–7 mol l–1. The trans dimer 14 needs a higher concentration by a factor of 10–100 to become the dominant species; however, in contrast with the behavior of the magnesium cis-porphyrin, the magnesium trans-porphyrin tends to form a trimer on increasing the concentration more than 10–5 mol l–1; it was estimated that the trimer content is about 34% in 2 × 10–3 mol l–1 CHCl3 solution [73]. The trimerization, found for magnesium but not for zinc complexes, was explained by the higher preference of magnesium for 6-coordination. Similar slipped cofacial dimers, 17 and 18 (meso substituents omitted), were reported for zinc and magnesium porphyrins functionalized with 2-pyridyl groups at the meso position [74, 75]. The stepwise formation con-
192
G. Ercolani
stants of the magnesium porphyrin dimer 18 in CH2 Cl2 , coinciding with Kinter and Kintra , were about 1 × 106 mol–1 l and 0.5, respectively [75]. This would indicate a very low EM value (≈ 5 × 10–7 mol l–1). The rationale for this anomalous behavior is that the acyclic dimer 19 is strongly stabilized by π – π cofacial interactions, much more than its cyclic isomer that probably also suffers from some strain because of the less than optimal geometry. It was estimated that in the absence of stabilizing π – π interactions in the acyclic dimer, Kinter would be about 75 mol–1 l [75] and therefore, considering that the overall formation constant of the cyclic dimer is Ksa ≈ 5 × 105 mol–1 l, an EM of ca. 90 mol l–1 would be estimated; a value more in keeping with the expectations. Zincporphyrins provided with a more extended pyridine side arm have been exploited in the formation of other dimeric assemblies such as 20 and 21 (meso-aryl substituents omitted for clarity). Hunter reported the selfassembly of the dimer 20 in which the orientation of the coordinating pyridine is controlled by the pattern of hydrogen bonding between the pyridine nitrogen and amide protons [76]. The stability constant Ksa = 2 × 108 mol–1 l and the reference constant Kinter = 5.6 × 103 mol–1 l, obtained in CH2 Cl2 , allowed to calculate an EM value of ca. 6 mol l–1 [69]; a quite high value that is justified by the level of preorganization of the side arm due to extended
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
193
π conjugation and hydrogen bonding. Interestingly, the rather large cavity of the assembly is capable of hosting terephthalic acid derivatives giving rise to [2]rotaxane architectures [77]. Shinkai reported the formation of the cyclic dimer 21 upon mixing a 1 : 1 mixture of a meso-dicatechol zincporphyrin and pyridine-3boronic acid dimethyl ester in CH2 Cl2 [78]. The dimerization constant Ksa = 1.1 × 104 mol–1 l and the reference constant Kinter = 340 mol–1 l, yield an EM ≈ 0.1 mol l–1 . The authors claimed complete formation of the dimer 21, however, the product EMKinter /n ≈ 17 does not satisfy the condition for complete self-assembly expressed by Eq. 22. According to the plot in Fig. 13 the weight fraction of the dimer can be in no case higher than ca. 0.9. The zincporphyrin-aniline interaction, although is an order of magnitude weaker than the corresponding zincporphyrin-pyridine interaction, has also been exploited to form dimers. Hunter et al. reported the formation of dimers 22 and 23 (meso-aryl groups omitted) from the corresponding porphyrins bearing at one meso position ortho and meta aniline groups, respectively [49]. According to the authors the stability constant of the dimer 22 in CDCl3 is Ksa = 160 mol–1 l and the reference constant Kinter = 10 mol–1 l, corresponding to an EM ≈ 1.6 mol l–1 . Also in this case, the product EMKinter /n ≈ 8 does not satisfy the condition for complete self-assembly and interpolation from the plot in Fig. 13 yields a maximum weight fraction of the assembly wsa(max) ≈ 0.8. According to Eq. 23, wsa(max) would be reached at Copt ≈ 0.4 mol l–1 , thus at the concentrations used to investigate the process, ca. 2.5 × 10–2 mol l–1 , the weight fraction of the assembly could have been significantly lower than the maximum. The equilibrium data for the assembly 23 are Ksa = 1080 mol–1 l and Kinter = 130 mol–1 l, yielding EM ≈ 6.4 × 10–2 mol l–1 and EMKinter /n ≈ 4, corresponding to wsa(max) ≈ 0.7. Thus, both 22 and 23 are not very effective in the competition with linear chains. According to the definition of multiporphyrin assembly given in the Introduction (Sect. 1), a dimeric hetero-assembly can be formed by A—A and B—B monomers only if both contain porphyrin subunits. An example is given by the dimer 24 (meso-aryl groups omitted) whose stability constant, Ksa = 3 × 108 mol–1 l, was determined in CH2 Cl2 by a fluorescence titration
194
G. Ercolani
method [79]. The value is very close to that of the dimer 20 on which the design of 24 was based; thus the same reference constant Kinter = 5.6 × 103 mol–1 l can be assumed. From Eq. 13, taking into account that for the present case σr = 4, an EM ≈ 2 mol l–1 can be calculated, which is in agreement with the expectation of a slighly lower rigidity of 24 with respect to 20. Other analogous dimeric hetero-assemblies were reported by von Borczyskowski et al. [80], and by Warrener et al. [81]. There are few equilibrium data available for cyclic trimers. It was preliminarily reported that 25 (meso-aryl groups omitted) yields a cyclic trimer in CH2 Cl2 [69, 82]; however the EM was suspiciously high for a cyclic trimer (≈ 100 mol l–1 ), especially if compared with that of the entropically more favored cyclic dimer 20 whose monomer has a sidearm similar to that of 25. The findings were refuted by subsequent experiments suggesting that the assembly formed by 25 actually consists of a cyclic dimer [14]. In a detailed study, Ikeda et al. reported that a zincporphyrin bearing a pyrazol-4-yl group in the 5-position and an ortho-ethoxycarbonylphenyl group in the 15-position, self-assembles into the cyclic trimer 26 (meso-aryl groups omitted) in CH2 Cl2 with a stability constant Ksa = 6.0 × 1013 mol–2 l2
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
195
at 22 ◦ C [83]. The stability constant of the reference adduct 27 obtained under the same conditions, was Kinter = 1.2 × 104 mol–1 l, yielding an EM ≈ 35 mol l–1 for the cyclic trimer. No doubt such a high EM value is due to the important contribution of the hydrogen bonding illustrated in structure 26 that, restricting the torsional motion about the pyrazolyl nitrogen-zinc bond, makes cyclization occur with a lower entropy loss. This was confirmed by experiments carried out with an analogous zincporphyrin in which the group in 15-position, responsible for hydrogen bonding with the pyrazolyl group, was replaced by a para-tolyl group. In this case the stability constants were Ksa = 2.8 × 108 mol–2 l2 for the cyclic trimer and Kinter = 1.2 × 103 mol–1 l for the reference reaction, yielding an EM ≈ 0.16 mol l–1, i.e. about 200 times lower than that of 26. This example highlights the importance of multipoint binding as a strategy to increase the EM of cyclic assemblies. Kobuke et al. reported an interesting case regarding a cobalt(III) porphyrin chloride bearing at one meso position a 1-methylimidazol-5-yl group. The compound assembles into a slowly equilibrating trimer-tetramer mixture as depicted in Fig. 14 (meso-aryl groups omitted) [84]. Of course the molar ratio of the assemblies is concentration dependent, as lower order assemblies are favored by dilution. Indeed, in going from a CDCl3 solution of the monomer 7.4 × 10–2 mol l–1 to 1.48 × 10–2 mol l–1 , the ratio 28 : 29 changed from 1 : 8.2 to 1 : 2. More intriguing is the effect of the solvent; in 2.0 × 10–2 mol l–1 solutions the percentage of trimer 28 increased on increasing the polarity (or coordinating ability) of the solvent as follows: 28% in CDCl3 , 31% in CD2 Cl2 , 89% in CD3 COCD3 , 100% in both CD3 OD and DMF-d6 . By a simple thermodynamic cycle, it would be easy to show that the constant of the equilibrium shown in Fig. 14 is given by the ratio EM43 /EM34 , where EM3 and EM4 are the EMs of the trimer and tetramer, respectively. Usually EMs are independent of the nature of the solvent and thus one would expect, contrarily to the facts, that the ratio trimer to tetramer is not affected by a solvent change. The ob-
Fig. 14 Dynamic equilibrium between the trimeric and tetrameric assemblies, 28 and 29
196
G. Ercolani
served behavior was ascribed to the close association, in less polar solvents, of the positively charged subunits of the assemblies with chloride anions, which would minimize electrostatic interactions among the subunits. In more polar solvents the cationic charges are less shielded by the counteranions and thus the tetracationic assembly would be less favored than the tricationic assembly because of greater coulombic repulsions. The example illustrates the fact that in the presence of attractive or repulsive interactions among the assembly subunits, the nature of the solvent can significantly affect the EM value. Cyclic tetramers are rather common. One of the most studied cases regards the aggregation of ZnPyTPP. Fleischer and Shachter reported a study on the aggregation of ZnPyTPP both in solution and in the solid state [47]: X-ray crystallography of ZnPyTPP showed that the pyridyl nitrogen is bound to the metal center of an adjacent porphyrin to form a linear polymeric aggregate; since a spectrophotometric study of ZnPyTPP in CHCl3 solution showed a good fit to a model involving reversible linear polymerization, it was concluded that the linear polymeric structure is present not only in the solid state but also in solution. However, Hunter et al. pointed out that the stability constant for the formation of the polymeric assembly in solution obtained by Fleischer and Shachter (Kinter = 3.1 × 104 mol–1 l) is unusually large for pyridine coordination of a zincporphyrin and suggested that ZnPyTPP probably forms in solution the cyclic tetrameric structure 30 (meso-phenyl groups omitted) [82]. Imamura et al. reported a variable 1 H NMR study of ZnPyTPP in CDCl and found that at – 40 ◦ C the spec3 trum is the same as that of the cyclic tetramer c– [RuPyTPP(CO)]4 with sharp pyridyl and β-pyrrole signals that broaden at 25 ◦ C [85]. They concluded that at low temperature the cyclic tetramer is the major component and that a fast exchange, relative to the NMR time scale, occurs at 25 ◦ C between cyclic and linear structures. Ercolani et al. studied the aggregation of ZnPyTPP both in toluene and CHCl3 by UV-vis spectrophotometry examining both the models of linear polymerization and cyclotetramerization [86]. The fitting procedure gave unrealistic parameters for the first model, definitively ruling out the polymerization hypothesis, whereas the cyclotetramerization model was confirmed with the following stability constants: Ksa = 5.7 × 1015 mol–3 l3 , Kinter = 4.1 × 103 mol–1 l, EM = 20 mol l–1 , in toluene; Ksa = 4.3 × 1013 mol–3 l3 , Kinter = 1.3 × 103 mol–1 l, EM = 15 mol l–1, in CHCl3 . It is worth noting that, as expected, the EMs are practically unaffected by the nature of solvent. Schaafsma et al. also studied the aggregation of ZnPyTPP in toluene and pointed out that the Soret band shows a splitting which can be explained by applying simple excitonic theory to the tetramer, whose stability constant was determined as Ksa = 6.2 × 1013 mol–3 l3 [87]. Osuka et al. reported an X-ray crystal structure of the cyclic tetramer formed from an analogous zinc pyridylporphyrin, namely 5-(4-pyridyl)-15-(3,5-di-tert-butylphenyl)porphyrin, demonstrating that the
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
197
tetrameric structure can be obtained even in the solid state [88]. They also evaluated the stability constant of the tetramer in CHCl3 solution as Ksa = 1.4 × 1015 mol–3 l3 . In a successive study Osuka et al. also evaluated the stability constants of cyclic tetramers obtained by inserting one, and two 1,4-phenylene spacers between the meso-position and the 4-pyridyl group of their porphyrin [89]. The corresponding stability constants of the selfassembled squares were 1.8 × 1012 and 7 × 1011 mol–3 l3 , respectively, showing, as expected, a decrease on increasing the flexibility of the monomeric units. A cyclic tetramer with a large cavity, 31, was previously reported by Hunter et al. [69, 82]; the reported stability constant, Ksa , of the order 1012 –1013 mol–3 l3 in CH2 Cl2 allowed to calculate an EM value of ca. 0.6–0.9 mol l–1 [69], which is about 30 times lower than that of the more preorganized tetramer 30. Interestingly the dynamics of self-assembly of the cyclic tetramers of certain zincporphyrins were held responsible for temperature dependent UV-vis spectra [90] and thermochromism [91]. A number of cyclic tetrameric hetero-assemblies have been reported involving the interaction of a biszincporphyrin with DABCO. Anderson reported the formation of the cyclic tetramer 32 (i = 2) whose stability constant is Ksa = 9.7 × 1019 mol–3 l3 in toluene, and 1.8 × 1018 mol–3 l3 in CHCl3 [38]. The interaction of DABCO with the reference zincporphyrin to form the sandwich complex 32 (i = 1), shows some degree of anticooperativity (cf. Sect. 2.4) that complicates the choice of the reference constant Kinter ; the most practical solution is to assume Kinter equal to the geometrical average of the constants K1 and K2 defined in Fig. 6a. By this choice Kinter = (K1 K2 )1/2 = 7.0 × 104 mol–1 l in toluene, and 2.2 × 104 mol–1 l in CHCl3 . From
198
G. Ercolani
Eq. 13, taking into account that for a tetrameric hetero-assembly σr = 16, an EM ≈ 0.3 mol l–1 in toluene, and ≈ 0.5 mol l–1 in CHCl3 , can be calculated, showing, as expected, little dependency on the nature of solvent. Other studies regard the interaction of the biszincporphyrins 33–35 with DABCO to yield the corresponding cyclic tetramers [39–41]. UV-vis titration studies of the self-assembly of 33 and 34 with DABCO were not capable of discriminating between cyclic dimer and tetramer formation, however the association constants obtained for the two binding models could be used to predict the assembling behavior at NMR concentrations, where the two binding models look quite different. In this way it was possible to ascertain that 33 and 34 assemble with DABCO to form cyclic tetramers [40, 41]. Unfortunately, the self-assembly of 35 and DABCO was studied by UV-vis only because insoluble polymeric assemblies formed at NMR concentrations; however, owing to the high level of preorganization of the building blocks, the formation of the cyclic tetramer is warranted also in this case [39]. It must be noted that in these studies, the authors ignored the statistical factor σr = 2n that appears in Eq. 13, thus the reported EMs must be divided by 16. The corrected EM value for the cyclic tetramer made of 33 and DABCO in CH2 Cl2 is 2 × 10–3 mol l–1 indicating a less than optimal preorganization of 33 [40]; indeed from the reported Kinter = 3.6 × 104 mol–1 l, a factor EMKinter /n ≈ 21 can be calculated, corresponding to wsa(max) ≈ 0.9. The tetrameric assembly formed from 34 and DABCO has a corrected EM ≈ 6 × 10–2 mol l–1 in CHCl3 and a Kinter ≈ 9 × 104 mol–1 l [41], thus the degree of preorganization of 34 is better than that of 33 but still not as high as that found for the bisporphyrins forming the complex 32 (i = 2). A surprising low corrected EM ≈ 1 × 10–4 mol l–1 can be assigned to the cyclic tetramer formed from 35 and DABCO in CHCl3 [39]. Considering the rigidity of the assembly, the low EM could be ascribed to the electrostatic repulsions between the stacked dipoles associated to the imido groups of 35. From the reported
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
199
Kinter = 2.1 × 105 mol–1 l, a factor EMKinter /n ≈ 5 can be calculated, corresponding to wsa(max) ≈ 0.7, showing that also in this case the assembly process is far from being complete. An interesting biszincporphyrin 36 was obtained by Alessio et al. as an inert bis-ruthenium complex [92]. It forms cyclic tetrameric assemblies with ditopic amines such as 4,4 -bipyridine and 5,15-bis(4-pyridyl)10,20-diphenylporphyrin (trans-DPyDPP). Single-crystal X-ray investigations showed that in the solid state these sandwich structures are maintained in the case of the trans-DPyDPP ligand whereas are disrupted to originate polymeric chains in the case of 4,4 -bipyridine. The stability constants of the cyclic tetramers formed from the analogous biszincporphyrins, 37 and 38, and 4,4 -bipyridine were determined in CHCl3 as Ksa = 3.1 × 1016 mol–3 l3 and 2.8 × 1018 mol–3 l3 , respectively [93]. The reference constants, Kinter = 1.5 × 104 mol–1 l and 1.8 × 105 mol–1 l, evaluated with 4-phenylpyridine, were significantly different in the two cases owing to the remarkable electronwithdrawing effect of the perfluorophenyl groups in 38 that increases the Lewis acidity of the zinc centers. The calculated EMs, equal to 0.04 mol l–1 and 2 × 10–4 mol l–1 , respectively, are smaller than the values expected on the basis of the rigidity of 37 and 38, suggesting that the stacked biszincporphyrins experience mutual electrostatic repulsions due to the polar Re(CO)3 Cl groups and, in the case of 38, perfluorophenyl groups. It must be noted however that 4-phenylpyridine may be not appropriate as reference ligand because its basicity is larger than that of 4,4 -bipyridine.
200
G. Ercolani
Cyclic tetramers, 39 (porphyrin substituents omitted), consisting of two bis-phosphine substituted zincporphyrin units as ligand donors and two rhodium(III) or ruthenium(II) porphyrins as ligand acceptors, were selected and amplified from dynamic combinatorial libraries by using appropriate templates in CDCl3 [67, 68]. The templating was achieved by using ditopic amines, such as 4,4 -bipyridine, acting through zinc-nitrogen coordination. Removal of the template, by addition of a competitor for the amine, e.g. ZnTPP, led to re-equilibration, with subsequent collapse of the host. Thus the template behaves as a scaffold necessary to maintain the host structure intact. This is an example of a cyclic assembly whose apparent EM is significantly increased by the presence of a template. As the number of monomeric units of a cyclic assembly increases, its formation in solution becomes unfavorable. Although there is a report regarding the formation of a cyclic hexamer in the solid state [94], to the best of our knowledge, there are no evidences for the complete assembly in solution of simple multiporphyrin rings of order greater than four.
4 Thermodynamics of Multicyclic Assemblies 4.1 Stability of Multicyclic Assemblies Multicyclic assemblies, as well as cyclic assemblies, are generally made either from a unique monomeric building block (homo-assemblies) or two monomeric building blocks (hetero-assemblies). The two types of architectures can be exemplified by the assembly 40 prepared by Osuka et al. [88, 89]
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
201
and by the ladder complexes 32 (i = 3–6) prepared by Anderson and Taylor. It is useful to define n as the number of assembly components, and b as the number of bonds joining them. Since the number of intermolecular bonds joining the monomeric units is n – 1, the number of intramolecular bonds, coinciding with the degree of cyclicity of the assembly, is b – n + 1. If the assembly has a symmetrical structure so that all the constituent rings are identical, and if cooperative effects are absent both in intermolecular and intramolecular processes, then the self-assembly equilibrium constant Ksa can be expressed as a function of two single microscopic constants, Kinter and K(intra)m , as shown by Eq. 25, which represents the extension of Eq. 13 to multicyclic architectures. σr n–1 b–n+1 K(intra)m (25) Ksa = Kinter σp Owing to the complexity and variety of multicyclic structures, a systematic method is needed for the evaluation of the statistical factors σr and σp . A practical way to evaluate statistical factors in difficult cases stems from the consideration that they coincide with the product of the symmetry numbers of reactants or products each raised to the corresponding stoichiometric coefficient [95]. The symmetry number of a molecule is defined as the total number of independent permutations of identical atoms or groups in a molecule that can be arrived at by simple rotations of the entire molecule, or by rotations about freely rotating single bonds within the molecule [96]. In practice,
202
G. Ercolani
the symmetry number is found by multiplying the symmetries of the independent rotational axes. These axes may be of two types, external or internal. External axes generate an identical arrangement of atoms by rigid rotation of the molecule as a whole; internal axes do so by rotations around single bonds within the molecule. Moreover, if a species is chiral and is present as racemic mixture, a factor 1/2 raised to the corresponding stoichiometric coefficient must also be considered; this correction accounts for the entropy of mixing of the two enantiomers [96]. For example, consider the formation of the homo-tetramer 40 (n = 4): the monomer has a C2 axis (2n = 16), it is chiral and was used as aracemic mixture (1/2n = 1/16), thus σr = 1; the assembly 40 (symmetry group D4 ) has a principal C4 axis and one orthogonal independent C2 axis (4 × 2 = 8), it is chiral and formed as racemic mixture (1/2), thus σp = 4. Consider now the formation of the hetero-assembly 32 (i = 3–6): the array of porphyrins can be schematized as A—(A)i–2 —A, and DABCO as B—B; both the sketched monomers have a C2 axis and are achiral thus σr = 22 × 2i = 2i+2 ; the sketched ladder complexes have two orthogonal C2 axes and are achiral thus σp = 4. Note that the use of sketched monomers is often sufficient to establish the symmetry corrections. For example there is no need to take into account the C3 axis of DABCO, because this is compensated in the ladder complex, where each DABCO unit has a 3-fold symmetrical internal rotation. Taking into account the definition of EMm , Eq. 25 can be rewritten as Eq. 26. Note that in the case of multicyclic assemblies, owing to the variety of their symmetries and therefore to the many values that σp can assume, it is convenient to consider separately this factor and EMm . σr b b–n+1 EMm (26) Ksa = Kinter σp The self-assembly of a multicyclic structure occurs in competition with non-linear polymerization. When the monomeric units have attained the equilibrium in solution, they can be considered as partitioned in two fractions in equilibrium between them, one constituted by an infinite number of oligomers having one or more loops in their structure and including the assembly, and the other constituted by an infinite number of more or less branched oligomers devoid of loops. Self-assembly takes place when the monomers have a rigid structure preorganized or predisposed in such a way that formation of the self-assembling complex is strongly favored over other cyclic or polycyclic species. The latter are disfavored by one or more of the following: (i) the presence of unreacted end groups, (ii) strained loops with high enthalpy content, (iii) large loops involving a high entropy loss. It is assumed therefore that the assembly and branched acyclic oligomers are the only significant species in solution. This assumption is further justified by the fact that the most important theory of non-linear polymerization, due to the contributions of Flory and Stockmayer, completely neglects the formation of
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
203
looped structures [32]. The competition between the self-assembly of a multicyclic architecture and nonlinear random polymerization has been treated in a way analogous to that regarding the competition between the formation of a cyclic assembly and linear polymerization, illustrated in Sect. 3.2 [97]. Among the principal results, it has been demonstrated that complete selfassembly, requiring wsa(max) > 0.99, can occur in a certain concentration range only if the condition in Eq. 27 is satisfied. b–n+1 1 n2 EMm Kinter ≥ 185 σp b
(27)
Equation 27 is a generalization of Eq. 22. Indeed, in the case of monocyclic assemblies, since σp = b = n, and EMm /n = EM, Eq. 27 reduces to Eq. 22. On the understanding of the importance of large EM and Kinter values, Eq. 27 points out that the driving force for self-assembly increases exponentially with the degree of cyclicity and that self-assembly will become easier and easier the higher is b, the number of bonds joining the assembly components, with respect to n, the number of assembly components. Equation 27, therefore, provides the physical basis for the principle of maximum occupancy stating that closed assemblies are generally more favored than the corresponding open isomers [1, 98], for narcissistic self-sorting [38], and for the concept of multivalency which is becoming one of the dominant themes of supramolecular chemistry [99–101]. A high degree of cyclicity is also the key for the preparation of stable assemblies with a large number of subunits which could not be prepared by simple cyclization. 4.2 Cooperativity in Multicyclic Assemblies Cooperativity has been rigorously defined in the case of multiple intermolecular binding of a monovalent ligand to a polyvalent receptor (Sect. 2.2). However, until recently, cooperativity had not been adequately defined in the case of intramolecular binding of a multivalent ligand, and much confusion was present in the literature, especially in the field of supramolecular chemistry where multiple intramolecular interactions play a key role. The source of such confusion can be traced to the lack of recognition that intramolecular and intermolecular processes are radically different from each other, as also evidenced by the different dimensions in which the constants Kintra and Kinter are expressed. Ercolani has recently proposed a method to assess cooperativity in self-assembly processes which is based on a clear distinction between intermolecular and intramolecular processes [31]. According to the author only virtually identical processes described by equilibrium constants having the same dimensions should be compared. Therefore intermolecular and intramolecular processes should be considered as forming two distinct groups within which cooperativity is assessed independently. In most cases, however,
204
G. Ercolani
it is difficult, if not impossible, to extract the relevant stepwise equilibrium constants, whereas the overall self-assembly constant is generally more accessible experimentally. If both the microscopic constant for the intermolecular processes, Kinter , and the microscopic constant for the intramolecular processes, K(intra)m , are available from reference compounds, the statistical self-assembly constant can be calculated by Eq. 25 and compared with the experimental one; if the latter exceeds the former, there is positive cooperativity, whereas if the opposite occurs, negative cooperativity has taken place. Application of the method to two classical cases, namely the self-assembly of helicates and of porphyrin ladders, by using data previously published by the groups of Lehn [102] and Anderson [38] (the latter are briefly discussed in Sect. 4.4), has shown, contrarily to the conclusions of the authors, that in both cases there is no cooperativity [31]. The methods previously used to assess cooperativity, in particular, the Scatchard plot and/or the Hill plot, have been criticized as being generally inappropriate for self-assembly, because they are pertinent to a specific case only, namely, the intermolecular binding of a monovalent ligand to a multivalent receptor (Sect. 2.2), a case very different from self-assembly which involves both inter- and intramolecular interactions (see, however, Sect. 4.3 for possible exceptions). The conclusion is that positive cooperativity in artificial self-assembling systems is probably much rarer than it was previously thought. 4.3 Topologically Reducible Multicyclic Assemblies In Fig. 15 are illustrated two examples of multicyclic assemblies, one closed and the other open, that possess a peculiarity: if we consider the stability of the cyclic subunit constituting the assembly, given by Eq. 13, equal to a pseudo-intermolecular constant K, the assembly formally behaves as an acyclic assembly with a microscopic binding constant equal to K. In other words the assembly can undergo a virtual topological reduction as illustrated in Fig. 15. The advantage of this topological reduction is that all the theory regarding cooperativity in closed acyclic assemblies (Sect. 2.2), as well as the theory of supramolecular polymerization (Sect. 2.3), can be immediately applied to topologically reducible assemblies. In Sect. 2.3 it has been pointed out that to obtain high DP values, it is necessary to reach association constant higher than 108 –109 mol–1 l; while such high values are hardly attainable by labile metalloporphyrin-ligand interactions, they are within the reach of pseudo-intermolecular constants. Thus the use of topologically reducible assemblies can be an excellent strategy for the construction of robust and large assemblies that would be otherwise difficult or impossible to assemble. Of course the concept of topologically reduction can be also applied to supercyclic structures; these are reducible to simple cyclic structures, as illus-
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
205
Fig. 15 Cartoons illustrating the topological reduction of a bicyclic assembly to a closed acyclic assembly (a), and of a multicyclic polymeric assembly to an acyclic polymer (b)
Fig. 16 Cartoons illustrating the topological reduction of a supercyclic assembly to a cyclic architecture
trated in Fig. 16. Accordingly the stability of the assembly can be expressed in terms of a pseudo-intermolecular constant and of the EM of the supercyclic structure. It should be noted that this strategy, not only provides pseudo-intermolecular constants significantly greater than simple intermolecular constants, but also increases the rigidity of the monomeric units constituting the reduced ring, so that an EM significantly larger than that of a simple cyclic structure is expected. This paves the way for the formation of supercyclic structures with an oligomerization degree higher than that of ordinary cyclic assemblies. 4.4 Selected Examples The characterization of multicyclic assemblies of porphyrins and the evaluation of their stability constitute a field of research that is still in its infancy. In particular, there is a significant paucity of data about the stability of mul-
206
G. Ercolani
ticyclic assemblies in solution, probably because of the difficulties to evaluate equilibrium constants which can assume huge values. Among the homo-assemblies, one of the most topologically simple is the tricyclic dimer of the biszincporphyrin 41 prepared by Alessio et al. [103]. Single crystal X-ray analysis of the dimer showed that the two pyridyl nitrogens of one 41 unit are coordinated to the two zinc sites of the other unit forming an achiral complex in which the two convergent units have opposite helical chirality. Both NMR and UV-vis spectra were found to be concentration independent in the range from 10–3 to 10–6 mol l–1 in CDCl3 e CHCl3 , indicating that the disassembly of the dimer is negligible even at the lowest concentration examined, but unfortunately the stability constant of the assembly was not determined. The chiral pentacyclic tetramer 40, prepared by Osuka et al. [88, 89] by homochiral self-sorting, is remarkably stable; its stability constant was estimated to be > 1025 mol–3 l3 from the independence of fluorescence spectra up to 10–8 mol l–1 . Interestingly, in the GPC analyses with CHCl3 as an eluent, 40 exhibited a sharp elution band at a distinctly shorter retention time than that of the free base monomer, which indicated the preservation of the discrete porphyrin box in the HPLC moving phase under highly diluted conditions. By exploiting this behavior, optical resolution of racemic 40 was accomplished with a chiral HPLC setup. Tetrameric boxes larger than 40 were obtained by analogous monomers with longer side arms, obtained by inserting one, and two 1,4-phenylene spacers between the 4-pyridyl groups and the bisporphyrin meso positions. An interesting modification of the monomer giving rise to 40 was carried out by Aida et al. by inserting the ethyne-1,2-diyl bridge between the porphyrin rings [104]. The monomer so modified self-assembles in CHCl3
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
207
to form a rectangular shaped nano box, 42 (meso substituents omitted), in which all the monomeric units preferentially adopt a planar conformation with the pyridyls oriented syn-parallel to each other. While the perpendicular conformation of the monomer in the assembly 40 is forced by the steric hindrance to rotation around the meso-meso linkage, the monomer in 42, not suffering from such hindrance, prefers the planar conformation because of its more extended π-conjugation. Interestingly, by inserting a longer bridge between the porphyrin rings, namely the butadiyne-1,4-diyl spacer, an equilibrium mixture of two tetrameric boxes is obtained, one, analogous to 40, in which the monomeric units adopt the perpendicular conformation (⊥) and the other, analogous to 42, in which the monomeric units adopt the planar conformation (//), the ratio of the two boxes ⊥ : // being 3 : 1 at 20 ◦ C and 3 : 2 at 50 ◦ C [105]. The greater stability of the ⊥ box was attributed to more favorable interactions among the dipole moments associated with the directions of the eight pyridyl groups. This effect would be more important than π-conjugation with the butadiyne-1,4-diyl bridge, in contrast with the ethyne-1,2-diyl bridge where the opposite trend was observed. Unfortunately stability constants of the boxes with respect to the constituting monomer were not determined. One of the first examples of topologically reducible homo-polymers was reported by Hunter et al. [106]; they designed the cobalt(II) porphyrin 43 (meso substituents omitted) symmetrically functionalized with two pyridine ligands which self-assembles to yield the polymer 44 (Fig. 17). Sizeexclusion chromatography yielded quantitative information about the size of the polymer as a function of the initial monomer concentration in the range 5 × 10–5 –10–2 mol l–1 . At the higher concentrations, DP was about 100, corresponding to a pseudo-intermolecular constant of the order of 106 mol–1 l. However, there is a puzzling aspect of these results that needs further investi-
Fig. 17 Self-assembly of a multicyclic homo-polymer via coordination of a cobalt porphyrin symmetrically functionalized with two pyridine ligands
208
G. Ercolani
gation. The cobalt(II)-amine interaction is known to occur with a decreased stability for the 6-coordinate complex. For example, 3,4-lutidine binds to cobalt(II) 5,10,15,20-tetra(p-anysil)porphyrin in toluene at 25 ◦ C with constants K1 = 1100 mol–1 l and K2 ≈ 0.1 mol–1 l (Fig. 1b) [107, 108]. The EM of the cyclic subunit of the polymer, evaluated on a zincporphyrin model compound, is ≈ 0.4 mol l–1 [109]. Thus the process of polymerization can be thought as occurring in two steps: the first step would lead to dimers with a pseudo-intermolecular constant equal to K12 EM ≈ 4.8 × 105 mol–1 l; the second step should join the dimers to form the polymer with a pseudointermolecular constant equal to K22 EM ≈ 4 × 10–3 mol–1 l. However, this value is so low that the second process could not occur, so that only cyclic dimers with 5-coordinate cobalt should be present in solution, in disagreement with the reported results. By altering the length of one of the sidearm, as in 45, the system could be biased to form supercyclic structures (Fig. 18) [109]. The lengths of the sidearms were such that metal-ligand coordination gives a 30◦ angle between adjacent porphyrin planes suggesting the formation of the dodecamer 46. Size-exclusion chromatography showed a trace over the concentration range 5 × 10–6 –5 × 10–4 mol l–1 , whose retention time is consistent with either a dodecameric structure or another cyclic oligomer close in size or a mixture of two or three such species. Only when the concentration was increased further, high molecular weight polymers began to emerge. The EM of the supercyclic structure was evaluated as ca. 5 × 10–4 mol l–1 . By assuming a pseudo-intermolecular constant of the same order of magnitude of that experimentally evaluated for the polymer above (ca. 106 mol–1 l), a factor EMK/n ≈ 40 can be calculated corresponding to wsa(max) ∼ 0.96 (from the plot in Fig. 13) indicating that the assembly of a supercyclic dodecamer is justified on account of the available driving force.
Fig. 18 Self-assembly of a dodecameric supercyclic structure via coordination of a cobalt porphyrin unsymmetrically functionalized with two pyridine ligands
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
209
Other interesting topologically reducible homo-polymers and supercycles obtained by exploiting the self-coordination of meso-(N-methyl)imidazolyl metalloporphyrins, were reported by the Kobuke group [17, 110–116]. These are reviewed by Kobuke himself in this volume. As to multicyclic hetero-assemblies, Anderson and Taylor investigated the formation of a series of ladder complexes 32 (i = 3–6) and determined their stability constants in both toluene and CHCl3 as reported in Table 2 [38]. They claimed that ladder formation exhibit many indications of positive cooperativity. In contrast, Ercolani pointed out that the self-assembly of porphyrin ladders occurs without cooperative effects because the stability constants follow the expected statistical behavior [31]. The analysis of Ercolani was based on the application of Eq. 25 to the self-assembly equilibria of 32 (i = 3–6), yielding Eq. 28. i+1 i–1 Ksa = 2i Kinter Kintra(m)
(28)
From the stability constants of 32 with i = 1, the following intermolecular constants were obtained: Kinter = 7.0 × 104 mol–1 l in toluene and 2.2 × 104 mol–1 l in CHCl3 (Sect. 3.3). The case i = 2 was taken as reference for the intramolecular reaction and used to estimate log Kintra(m) (7.1 × 104 in toluene and 4.1 × 104 in CHCl3 ). Knowing Kinter and Kintra(m) , the stability constants of the ladder complexes 32 (i = 3–6) were calculated by Eq. 28; the results are reported in Table 2 together with the corresponding errors calculated by the theory of error propagation. The accordance between experimental and statistical log Ksa values is extraordinarily good, ruling out the presence of cooperative effects. Self-assembly of the interesting calixarene-porphyrin cage 48 in CH2 Cl2 was reported by Hunter et al. (Fig. 19) [40]. On mixing the calix-tetraporphyrin 47 and DABCO, four different complexes are mainly formed depending on the stoichiometry and concentration of the reactants. At a calixtetraporphyrin:DABCO ratio of 2 : 4, the major species is the tricyclic asTable 2 Experimental and calculated log Ksa values in toluene and chloroform for ladder complexes 32 (i = 3–6) i
(log Ksa )exp in PhMe a
(log Ksa )calc in PhMe b
(log Ksa )exp in CHCl3 a
(log Ksa )calc in CHCl3 b
3 4 5 6
29.7 ± 0.2 39.9 ± 0.2 50.3 ± 0.3 60.5 ± 0.3
30.0 ± 0.4 40.0 ± 0.6 50.0 ± 0.8 60.0 ± 0.9
27.2 ± 0.2 37.2 ± 0.2 46.5 ± 0.3 56.5 ± 0.3
27.5 ± 0.4 36.8 ± 0.6 46.1 ± 0.8 55.4 ± 0.9
a
Data from [38] Data from [31]
b
210
G. Ercolani
Fig. 19 Self-assembly of the calix [4]arene 47 equipped with four zincporphyrins, and DABCO (2 : 4) to form the cage 48
sembly 48 that features a large internal cavity. The stability constants were Ksa = 6.3 × 1033 mol–5 l–5 for the cage assembly and Kinter = 4.7 × 104 mol–1 l–1 for the reference intermolecular constant. According to Eq. 26, considering that σr /σp = 32 (note that the authors did not take into account this factor to evaluate the EM; see also the discussion about statistical factors in Sects. 3.2 and 4.1), a corrected EMm = 0.02 mol l–1 can be calculated that is in reasonable agreement with the EMm = 9.2 × 10–3 L–1 calculated for the formation of the cyclic tetramer made of 33 and DABCO 2 : 2. As to topologically reducible hetero-assemblies, two cases, both regarding bicyclic assemblies, are worth discussing. The first case regards the self-assembly of a free base porphyrin having four symmetrical pyridinyl sidearms that binds two biszincporphyrins to form a pentaporphyrin trimer [117]. The case can be schematically illustrated by the cartoons in Fig. 15a. The pseudo-intermolecular constant for each of the two binding events is 2 × 106 mol–1 l, indicating the absence of cooperativity effects. The second case regards the formation of the assembly 49 in which a tetrapyridyl-
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
211
bisporphyrin binds two biszincporphyrins [118]. Also, this case is reducible as shown in Fig. 15a, and thus, a reaction scheme analogous to that defined in Fig. 6 can be used. UV-vis titrations in CH2 Cl2 showed the formation of the 1 : 2 complex 49 with an overall stability constant too high to determine accurately (αK 2 ≈ 1016 –1019 mol–2 l2 , where K is the pseudo-intermolecular constant). However, 1 H NMR titrations allowed the determination of the disassembly constant that only depends on the cooperativity factor, α. The obtained value of α = 1.8 indicates the presence of positive cooperativity. This can be explained by considering that formation of the 1 : 1 complex tends to hold the tetrapyridyl ligand in a coplanar conformation, increasing its affinity for the second biszincporphyrin.
5 Conclusion and Perspectives Significant developments in the construction of multiporphyrin assemblies have been seen over recent years and many more will be seen in the years to come. No doubt the rational design of complex and intriguing supramolecu-
212
G. Ercolani
lar architectures will benefit from the knowledge of the theoretical principles governing self-assembly in solution. This contribution has been written with the aim of providing a unified view of such principles by emphasizing the role played by two fundamental physicochemical quantities, namely, the stability constant of the metalloporphyrin-ligand interaction and the effective molarity, the latter being a measure of the ease of formation of a given cyclic structure. Of course, the evolution of physical principles in this area has required and still requires a large body of experimental data about the stability of assemblies in solution. Unfortunately most of the multiporphyrin assemblies have been characterized without such associated experimental data and this state of affairs has certainly slowed down progress in the field. In this respect we are obliged to the English school of Jeremy Sanders, Harry Anderson, and Christopher Hunter for having provided many of the available quantitative studies on metal-mediated porphyrin assemblies and for having emphasized the important role played by the EM. In terms of future efforts in this area, we expect that multicyclic oligomeric assemblies will play a dominant role, because of their high stability, their rich stereochemical potentialities, and the possibility to be organized in cages and capsules capable of hosting a range of guest molecules. Needless to say, since such large multicyclic assemblies consist of cyclic subunits, the EM will continue to be of paramount importance for the correct appraisal of their stability. In this respect it is desirable that theoretical and computational chemists face the challenges posed by the theoretical evaluation of the EM. No doubt an affordable method for the a priori evaluation of the EM of a given supramolecular architecture would give a strong impulse to the field. We conclude by remarking that self-assembly ultimately rests on the chelate effect, possibly amplified by multivalency and cooperativity. No doubt, these will be the major themes in the field of supramolecular chemistry in the next years, influencing the approach of supramolecular chemists to nanotechnology.
References 1. Lehn J-M (1995) Supramolecular chemistry: concepts and perspectives. Wiley, Weinheim 2. Lawrence DS, Jiang T, Levett M (1995) Chem Rev 95:2229 3. Philp D, Stoddart JF (1996) Angew Chem Int Ed Engl 35:1154 4. Lehn J-M, Atwood JL, Davies JED, MacNicol DD, Vögtle F, Sauvage J-P, Hosseini MW (eds) (1996) Comprehensive supramolecular chemistry, vol 9. Pergamon, Oxford 5. Caulder DL, Raymond KN (1999) J Chem Soc Dalton Trans 1185 6. Lindoy LF, Atkinson I (2000) Self-assembly in supramolecular systems. In: Stoddart JF (ed) Monographs in supramolecular chemistry. The Royal Society of Chemistry, Cambridge, UK 7. Fujita M (ed) (2000) Struct Bond, vol 96
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
213
8. Holliday BJ, Mirkin CA (2001) Angew Chem Int Ed 40:2022 9. Seidel SR, Stang PJ (2002) Acc Chem Res 35:972 10. Chambron J-C, Heitz V, Sauvage J-P (2000) Noncovalent multiporphyrin assemblies. In: Kadish KM, Smith KM, Guilard R (eds) The porphyrin handbook, vol 6. Academic, New York, p 1 11. Sanders JKM (2000) Coordination chemistry of oligoporphyrins. In: Kadish KM, Smith KM, Guilard R (eds) The porphyrin handbook, vol 3. Academic, New York, p 347 12. Imamura T, Fukushima K (2000) Coord Chem Rev 198:133 13. Wojaczynski J, Latos-Graznski L (2000) Coord Chem Rev 204:113 14. Baldini L, Hunter CA (2002) Adv Inorg Chem 53:213 15. Iengo E, Zangrando E, Alessio E (2003) Eur J Inorg Chem 2371 16. Harvey PD (2003) Recent advances in free and metalated multiporphyrin assemblies and arrays; a photophysical behavior and energy transfer perspective. In: Kadish KM, Smith KM, Guilard R (eds) The porphyrin handbook, vol 18. Elsevier Science (USA), p 63 17. Satake A, Kobuke Y (2005) Tetrahedron 61:13 18. Balaban TS (2005) Acc Chem Res 38:612 19. Lehn J-M, Atwood JL, Davies JED, MacNicol DD, Vögtle F, Davies JED, Ripmeester JA (eds) (1996) Comprehensive supramolecular chemistry, vol 8. Pergamon, Oxford 20. Schneider H-J, Yatsimirsky AK (2000) Principles and methods in supramolecular chemistry. Wiley, Chichester 21. Sanders JKM, Bampos N, Clyde-Watson Z, Darling SL, Hawley JC, Kim H-J, Mak CC, Webb SJ (2000) Axial coordination chemistry of metalloporphyrins. In: Kadish KM, Smith KM, Guilard R (eds) The porphyrin handbook, vol 3. Academic, New York, p 1 22. Tabata M, Nishimoto J (2000) Equilibrium data of porphyrins and metalloporphyrins. In: Kadish KM, Smith KM, Guilard R (eds) The porphyrin handbook, vol 9. Academic, New York, p 221 23. Sakuragi M, Ichimura K, Sakuragi H (1992) Bull Chem Soc Jpn 65:1944 24. Imai H, Nakagawa S, Kyuno E (1992) J Am Chem Soc 114:6719 25. Bhyrappa P, Krishnan V, Nethaji M (1993) J Chem Soc, Dalton Trans 1901 26. Vogel GC, Stahlbush JR (1977) Inorg Chem 16:950 27. Nappa M, Valentine JS (1978) J Am Chem Soc 100:5075 28. Cole SJ, Curthoys GC, Magnusson EA, Phillips JN (1972) Inorg Chem 11:1024 29. Perlmutter-Hayman B (1986) Acc Chem Res 19:90 30. Connors KA (1987) Binding constants: the measurement of molecular complex stability, chap 2. Wiley, New York 31. Ercolani G (2003) J Am Chem Soc 125:16097 32. Flory PJ (1946) Chem Rev 39:137 33. Drain CM, Shi X, Milic T, Nifiatis F (2001) Chem Commun 287 34. Ercolani G (2001) Chem Commun 1416 35. Hunter CA, Meah MN, Sanders JKM (1990) J Am Chem Soc 112:5773 36. Anderson HL, Hunter CA, Meah MN, Sanders JKM (1990) J Am Chem Soc 112:5780 37. Anderson HL (1994) Inorg Chem 33:972 38. Taylor PN, Anderson HL (1999) J Am Chem Soc 121:11538 39. Hunter CA, Tregonning R (2002) Tetrahedron 58:691 40. Baldini L, Ballester P, Casnati A, Gomila RM, Hunter CA, Sansone F, Ungaro R (2003) J Am Chem Soc 125:14181 41. Ballester P, Costa A, Castilla AM, Deyà PM, Frontera A, Gomila RM, Hunter CA (2005) Chem Eur J 11:2196
214
G. Ercolani
42. Slagt VF, Kamer PCJ, van Leeuwen PWNM, Reek JNH (2004) J Am Chem Soc 126:1526 43. Kleij AW, Kuil M, Tooke DM, Spek AL, Reek JNH (2005) Inorg Chem 44:7696 44. Kuroda Y, Sugou K, Sasaki K (2000) J Am Chem Soc 122:7833 45. Sugou K, Sasaki K, Kitajima K, Iwaki T, Kuroda Y (2002) J Am Chem Soc 124:1182 46. Fantuzzi G, Pengo P, Gomila R, Ballester P, Hunter CA, Pasquato L, Scrimin P (2003) Chem Commun 1004 47. Fleischer EB, Shachter AM (1991) Inorg Chem 30:3763 48. Burrell AK, Officer DL, Reid DCW, Wild KY (1998) Angew Chem Int Ed 37:114 49. Gardner M, Guerin AJ, Hunter CA, Michelsen U, Rotger C (1999) New J Chem 309 50. Twyman LJ, King ASH (2002) Chem Commun 910 51. Jacobson H, Stockmayer WH (1950) J Chem Phys 18:1600 52. Ercolani G, Mandolini L, Mencarelli P, Roelens S (1993) J Am Chem Soc 115:3901 53. Collman JP, McDevitt JT, Leidner CR, Yee GT, Torrance JB, Little WA (1987) J Am Chem Soc 109:4606 54. Kirby AJ (1980) Adv Phys Org Chem 17:183 55. Mandolini L (1986) Adv Phys Org Chem 22:1 56. Galli C, Mandolini L (2000) Eur J Org Chem 3117 57. Page MI, Jencks WP (1971) Proc Nat Acad Sci USA 68:1678 58. Page MI (1973) Chem Soc Rev 2:295 59. Jencks WP (1975) Adv Enzym 43:219 60. Ercolani G (1999) J Org Chem 64:3350 61. Darian E, Hnizdo V, Fedorowiczf A, Singh H, Demchuk E (2005) J Comp Chem 26:651 62. Flory PJ, Suter UW, Mutter M (1976) J Am Chem Soc 98:5733 63. Cacciapaglia R, Di Stefano S, Mandolini L (2004) Acc Chem Res 37:113 64. Fujita M, Sasaki O, Mitsuhashi T, Fujita T, Yakazi J, Yamaguchi K, Ogura K (1996) Chem Commun 1535 65. Rowan SJ, Hamilton DG, Brady PA, Sanders JKM (1997) J Am Chem Soc 119:2578 66. Rowan SJ, Cantrill SJ, Cousins GRL, Sanders JKM, Stoddart JF (2002) Angew Chem Int Ed 41:898 67. Stulz E, Ng Y-F, Scott SM, Sanders JKM (2002) Chem Commun 524 68. Stulz E, Scott SM, Bond AD, Teat SJ, Sanders JKM (2003) Chem Eur J 9:6039 69. Chi X, Guerin AJ, Haycock RA, Hunter CA, Sarson LD (1995) J Chem Soc, Chem Commun 2563 70. Ercolani G (1998) J Phys Chem B 102:5699 71. Kobuke Y, Miyaji H (1994) J Am Chem Soc 116:4111 72. Rudkevich DM, Verboom W, Reinhoudt DN (1995) J Org Chem 60:6585 73. Kobuke Y, Miyaji H (1996) Bull Chem Soc Jpn 69:3563 74. Stibrany RT, Vasudevan J, Knapp S, Potenza JA, Emge T, Schugar HJ (1996) J Am Chem Soc 118:3980 75. Gerasimchuk NN, Mokhir AA, Rodgers KR (1998) Inorg Chem 37:5641 76. Hunter CA, Sarson LD (1994) Angew Chem Int Ed Engl 33:2313 77. Hunter CA, Low CMR, Packer MJ, Spey SE, Vinter JG, Vysotsky MO, Zonta C (2001) Angew Chem Int Ed 40:2678 78. Sarson LD, Ueda K, Takeuchi M, Shinkai S (1996) Chem Commun 619 79. Hunter CA, Hyde RK (1996) Angew Chem Int Ed Engl 35:1936 80. Chernook AV, Shulga AM, Zenkevich EI, Rempel U, von Borczykowski C (1996) J Phys Chem 100:1918 81. Johnston MR, Gunter MJ, Warrener RN (1998) Chem Commun 2739
Thermodynamics of Metal-Mediated Assemblies of Porphyrins
215
82. Chi X, Guerin AJ, Haycock RA, Hunter CA, Sarson LD (1995) J Chem Soc, Chem Commun 2567 83. Ikeda C, Tanaka Y, Fujihara T, Ishii Y, Ushiyama T, Yamamoto K, Yoshioka N, Inoue H (2001) Inorg Chem 40:3395 84. Tanaka A, Ryuno A, Okada S, Satake A, Kobuke Y (2005) Isr J Chem 45:281 85. Funatsu K, Imamura T, Ichimura A, Sasaki Y (1998) Inorg Chem 37:1798 86. Ercolani G, Ioele M, Monti D (2001) New J Chem 25:783 87. Yatskou MM, Koehorst RBM, Donker H, Schaafsma TJ (2001) J Phys Chem A 105:11425 88. Tsuda A, Nakamura T, Sakamoto S, Yamaguchi K, Osuka A (2002) Angew Chem Int Ed 41:2817 89. Hwang I-W, Kamada T, Ahn TK, Ko DM, Nakamura T, Tsuda A, Osuka A, Kim D (2004) J Am Chem Soc 126:16187 90. Balaban TS, Goddard R, Linke-Schaetzel M, Lehn J-M (2003) J Am Chem Soc 125:4233 91. Tsuda A, Sakamoto S, Yamaguchi K, Aida T (2003) J Am Chem Soc 125:15722 92. Iengo E, Zangrando E, Alessio E (2002) J Am Chem Soc 124:1003 93. Splan KE, Stern CL, Hupp JT (2004) Inorg Chim Acta 357:4005 94. Knapp S, Vasudevan J, Emge TJ, Arison BH, Potenza JA, Schugar HJ (1988) Angew Chem Int Ed 37:2368 95. Benson SW (1958) J Am Chem Soc 80:5151 96. Lowry TH, Richardson KS (1987) Mechanism and theory in organic chemistry, 3rd edn. Harper & Row, New York, p 175 97. Ercolani G (2003) J Phys Chem B 107:5052 98. Piguet C, Borkovec M, Hamacek J, Zeckert K (2005) Coord Chem Rev 249:705 99. Mammen M, Choi S-K, Whitesides GM (1998) Angew Chem Int Ed 37:2754 100. Mulder A, Huskens J, Reinhoudt DN (2004) Org Biomol Chem 2:3409 101. Badjicä JD, Nelson A, Cantrill SJ, Turnbull WB, Stoddart JF (2005) Acc Chem Res 38:723 102. Pfeil A, Lehn J-M (1992) J Chem Soc, Chem Commun 838 103. Iengo E, Zangrando E, Geremia S, Graff R, Kieffer B, Alessio E (2002) Chem Eur J 8:4670 104. Tsuda A, Hu H, Watanabe R, Aida T (2003) J Porphyrins Phthalocyanines 7:388 105. Tsuda A, Hu H, Tanaka R, Aida T (2005) Angew Chem Int Ed 44:4884 106. Michelsen U, Hunter CA (2000) Angew Chem Int Ed 39:764 107. Walker FA (1973) J Am Chem Soc 95:1150 108. Walker FA (1973) J Am Chem Soc 95:1154 109. Haycock R, Hunter CA, James DA, Michelsen U, Sutton LR (2000) Org Lett 2:2435 110. Ogawa K, Kobuke Y (2000) Angew Chem Int Ed 39:4070 111. Ikeda C, Fujiwara E, Satake A, Kobuke Y (2003) Chem Commun 616 112. Takahashi R, Kobuke Y (2003) J Am Chem Soc 125:2372 113. Takahashi R, Kobuke Y (2005) J Org Chem 70:2745 114. Furutsu D, Satake A, Kobuke Y (2005) 44:4460 115. Shoji O, Okada S, Satake A, Kobuke Y (2005) J Am Chem Soc 127:2201 116. Shoji O, Tanaka H, Kawai T, Kobuke Y (2005) J Am Chem Soc 127:8598 117. Haycock RA, Yartsev A, Michelsen U, Sundström V, Hunter CA (2000) Angew Chem Int Ed 39:3616 118. Wilson GS, Anderson HL (1999) Chem Commun 1539
Struct Bond (2006) 121: 217–261 DOI 10.1007/430_018 © Springer-Verlag Berlin Heidelberg 2006 Published online: 10 February 2006
Porphyrin Rotaxanes and Catenanes: Copper(I)-Templated Synthesis and Photoinduced Processes Lucia Flamigni1 (u) · Valérie Heitz2 (u) · Jean-Pierre Sauvage2 (u) 1 Istituto
per la Sintesi Organica e la Fotoreattività (ISOF)-CNR, Via P. Gobetti 101, 40129 Bologna, Italy fl[email protected] 2 Laboratoire de Chimie Organo-Minérale, UMR 7513 du CNRS, Institut Le Bel, Université Louis Pasteur, 4 rue Blaise Pascal, 67000 Strasbourg Cedex, France [email protected], [email protected] 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
218
2 2.1
Porphyrin Stoppered-Rotaxanes . . . . . . . . . . . . . . . . . . . . . . . The First Systems, Consisting of a Bis-Porphyrin (PZn, PAu+ ) Dumbbell Threaded Through a 1,10-Phenanthroline-Containing Ring . . . . . . . . [3]- and [5]Rotaxanes with Porphyrin Stoppers . . . . . . . . . . . . . . . [2]Rotaxanes Consisting of a Dumbbell with Two PZn Stoppers and a PAu+ -Appended Ring . . . . . . . . . . . . . . . . . . . . . . . . . Porphyrin [2]Rotaxanes for Which a PAu+ Fragment is an Integral Part of the Ring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
221
. .
221 227
.
234
.
241
Porphyrin-Containing Catenanes . . . . . . . . . . . . . . . . . . . . . . . A [2]Catenane Whose Two Rings Incorporate a PZn and a PAu+ Unit. . . . Copper(I)-Templated Synthesis of [2]Catenanes Bearing Pendant Porphyrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
246 246 253
4
Related Catenanes and Rotaxanes Recently Reported in the Literature . .
254
5
Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . .
258
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
259
2.2 2.3 2.4 3 3.1 3.2
Abstract The approach based on the copper(I)-templated synthesis of porphyrin catenanes and rotaxanes developed by the authors’ group is here reviewed. Zn(II) porphyrins and gold(III) porphyrins were chosen as electron donors and electron acceptors, respectively, to constitute the electro- and photoactive parts of the present systems. The processes—energy and electron transfer reactions—occurring in the interlocked structures upon light absorption in the presence or absence of Cu(I) are presented, their rates and efficiencies critically compared and discussed with respect to properties of the components and of the ensemble. A detailed examination of differences and analogies in photoreactivity between the present and closely related systems reported by others is presented. Keywords Catenanes · Charge separation · Energy transfer · Porphyrins · Rotaxanes
218
L. Flamigni et al.
1 Introduction Photosynthesis is a fascinating process, which has attracted formidable interest from various fields of science over several decades. The publication of the X-ray structure of the reaction center (RC) of Rhodopseudomonas viridis in 1984 by German researchers was without doubt a major event [1, 2]. It not only confirmed what the biologists knew about the spatial arrangement of the various components of the RC, but it also inspired many research groups and thus initiated the design and synthesis of numerous artificial donor-acceptor systems expected both to display structural analogy with the RC and to fulfill some of its photochemical and electron-transfer functions. Among the most relevant synthetic molecular systems proposed by various researchers, those containing two or more porphyrins are particularly promising [3–8]. It is now well established that in the RC, electron transfer (ET) between tetrapyrrolic chromophores is indeed the primary charge separation step following generation of the special-pair (SP) singlet excited state. The arrangement of the cofactors of the core of the RC is represented in Fig. 1a. This ET process occurs in ∼ 3 ps with the formation of a bacteriopheophytinreduced state (BPh–. radical anion) [9, 10]. Subsequent electron transfer to a quinone is much slower (∼ 200 ps). The first system that we designed some time ago to mimic photoinduced electron transfer in the RC was based on the following requirements: (i) An oblique bis-porphyrin looked appealing since in the RC the porphyrinlike nuclei are mostly organized in an oblique fashion to one another. The cytochrome part consists of 4 hemes and the transmembrane part contains 4 BCh’s and 2 BPh’s, the only pair of relatively closely lying and parallel rings being that of the SP [1]. (ii) If intramolecular electron transfer between two porphyrin units is to be the main reaction occurring after light excitation, very precisely defined electrochemical properties for both components are required. In addition, the various excited state energy levels will also have to be positioned in a well-controlled manner. The electron donor part will have to be a singlet excited state (∗1 D). The acceptor porphyrin (A) will have to display a very high-lying singlet excited state so as to avoid undesired energy transfer from ∗ D. In order to favor electron transfer, the acceptor porphyrin should, of course, be relatively easy to reduce and the singlet excited state of the donor (∗1 D) must be a strong reductant. These electronic properties can be governed by the nature of the metal centers introduced in the central coordination sites of each porphyrin. We selected zinc(II) and gold(III) as metals for the donor component and the acceptor subunit, respectively. The [Zn(II), Au(III)] couple also displays an interesting property related to photoinduced electron transfer stud-
Porphyrin Rotaxanes and Catenanes
219
ies. If the porphyrin substituents are carefully chosen, a very convenient spectral window around 600 nm in the absorption spectrum allows one to selectively excite the zinc-containing tetrapyrrolic ring without significantly affecting the gold(III) porphyrin [11]. It is worth noting that the [Zn(II), Au(III)] strategy is reminiscent of the photosynthetic RC. In Nature, energy transfer has been avoided by decreasing the energy level of the donor (SP: a porphyrin dimer instead of a porphyrin monomer) [12], whereas in our approach the same effect is obtained by raising the excited state level of the electron acceptor [gold(III) porphyrin]. The system which fulfilled these requirements is the 2,9-diphenyl-1,10phenanthroline (dpp) bridged Zn(II)/Au(III) bis-porphyrin conjugate 1+ (represented in Fig. 1b) [13]. Time-resolved fluorescence and transient absorption spectroscopy showed that forward electron transfer took place in 55 ps and charge recombination was delayed by a factor of ten [11]. Presumably, electron transfer takes place via a through-bond pathway, the dpp bridge playing the role of a superexchange relay, like the accessory bacteriochlorophyll (BCh, Fig. 1a) in the photosynthetic reaction center, according to several authors [14, 15]. More recently, porphyrin units have been incorporated in complex molecules such as catenanes [16–18] and rotaxanes [19–22]. Whereas catenanes are composed of interlocking rings only, the simplest rotaxanes
Fig. 1 a A fragment of the photosynthetic RC with the three important components: SP, BCh and BPh [1]. b Chemical structure of the gold(III)/zinc(II) bis-porphyrin 1+
220
L. Flamigni et al.
are molecules made from a macrocycle threaded onto dumbbell-shaped molecules consisting of axle-bearing large, bulky end-groups (or stoppers) to prevent the macrocycle from slipping out. Both contain independent portions that are not bonded to each other by any covalent bond but nevertheless remain linked. This particular mode of bonding is often called a mechanical link [23, 24]. These interlocked structures offer therefore the opportunity to study photoinduced electron or energy transfer processes between non-covalently linked porphyrins. In this way they mimic more closely the tetrapyrrolic arrangement of natural pigments found in light harvesting antennae or in reaction centers. Catenanes and rotaxanes are also ideally suited for undergoing large amplitude motions under the action of an external signal which makes them particularly attractive for the building of molecular machines and motors [25–28]. Among the different synthetic methods developed to make catenanes and rotaxanes, the template synthesis, which uses either organic templates (van der Waals interactions, hydrogen-bonds, π – π stacking) or transition-metal templates (metal-ligand interactions) proved to be the most efficient [29–36]. A unique feature of the use of a transition metal is that the template can easily be removed at the end of the synthesis whereas, using strategies based on organic fragments, the interactions between the template and the components of the final rotaxane structure will, most of the time, be maintained. The first successful application of such an approach was in the field of catenane chemistry and came through previous work from our laboratories [37]. It was based on the fact that Cu(I) forms extremely stable complexes with bidentate 2,9-diphenyl-1,10-phenanthroline (dpp) ligands [38]. Two such ligands are entwined around the Cu(I) cation forming a tetrahedral complex. By using the preferential tetrahedral coordinating properties of Cu(I), a transition-metal templated synthesis of rotaxanes or catenanes was devised. The synthetic strategy is shown in Fig. 2. The threading step is a complexation reaction, which involves a dpp chelate-based macrocycle (A), and a linear fragment incorporating the same dpp chelate (B) end-functionalized by the appropriate reacting groups. The Cu(I) cation templates the assembly of the threaded complex (C). This strategy is suitable for the construction of a large variety of rotaxane- or catenane-bearing porphyrins since (i) the threading reaction shown in Fig. 2 is quantitative, as
Fig. 2 The Cu(I)-directed threading reaction: A is a macrocyclic compound incorporating a dpp fragment (indicated with a thick line), B is a dpp-containing molecular thread and the black disk is Cu+ . C is the threaded complex formed
Porphyrin Rotaxanes and Catenanes
221
proven by work done in our group; and (ii) because it is modular, which enables modules A and B to be functionalized with various porphyrins as will be described in the following sections.
2 Porphyrin Stoppered-Rotaxanes 2.1 The First Systems, Consisting of a Bis-Porphyrin (PZn, PAu+ ) Dumbbell Threaded Through a 1,10-Phenanthroline-Containing Ring The approach used for construction of porphyrin stoppered-rotaxanes is based on the template synthesis described in Fig. 2. On the threaded complex (C), porphyrins playing the role of big stoppers can be constructed, to prevent dethreading of the dumbbell from the ring after removal of the template metal. In Fig. 3 two possible rotaxanes are indicated that can be formed (a [2]- or a [3]rotaxane) by following this construction principle. This figure also shows that this strategy permits us to introduce two different metalated porphyrins in a stepwise manner. By following the strategy described in Fig. 3, the ideal precursor corresponding to B is a non-symmetrical dpp derivative attached to a porphyrin at one end and to an aromatic aldehyde at the other. The macrocycle A of Fig. 3 can obviously be the dpp-incorporating 30-membered ring used previously for making catenates and catenanes [39–41]. As shown by the handling of CPK molecular models, it is sufficiently small to prevent release of the phenanthroline-bridged bis-porphyrin dumbbell. The starting materials are represented in Fig. 4. In compound 2+ , the porphyrin incorporates a trivalent gold which was selected for two reasons: (i) it forms very stable porphyrin complexes and will thus not be lost during the synthesis of the second porphyrin nucleus [13]; and (ii) because of its strong electropositive character, it confers to the aromatic porphyrin ring to which it is complexed a remarkable electronaccepting ability with a resulting very accessible reduction potential [41–44]. The reactions leading to the transition metal-complexed [2]- and [3]rotaxanes 82+ and 104+ , respectively, are indicated in Fig. 5 [45]. Prerotaxane 52+ was first formed. Macrocycle 3 was complexed with Cu(I) by reaction with Cu(CH3 CN)4 BF4 . Then the gold porphyrin 2+ (BF4 – ) was added. Noteworthy, although Cu(dpp)2 + -type complexes are notoriously highly colored [46], the presence of the gold(III) porphyrin made any color change in the course of the formation reaction of 52+ virtually undetectable. Thin-layer chromatography and NMR spectroscopy showed that complex 52+ had formed quantitatively. It was used in the next step, without further purification. Importantly, the synthesis of the second porphyrin had to be
222
L. Flamigni et al.
Fig. 3 The macrocycle A incorporating a coordinating fragment (thick line) interacts with a metal center (black circle) and an asymmetrical open chain chelate B bearing one porphyrin and a precursor function X which is small enough to pass through the ring; after the threaded intermediate C is formed, the additional porphyrin ring is constructed, affording the transition metal-complexed rotaxanes D and E. Demetalation leads to the free-ligand rotaxanes F and G from D and E, respectively
compatible with the existence of the threaded pre-rotaxane intermediate C of Fig. 3. A mild reaction had thus to be selected, for which no demetalation of the copper(I) complex 52+ was expected to take place. Since Rothemund– Adler [47–49] synthesis provides too stringent a set of conditions for copper(I) bis-chelate complexes to survive the reaction quantitatively, we investigated a milder procedure developed by Lindsey and co-workers [50]. The condensation reaction was performed by mixing 52+ (BF4 – )2 , di-tert-butyl3,5-benzaldehyde [51] and dipyrrylmethane 4 in the molar ratio 1 : 4 : 40. Subsequently, a large excess of chloranil was added in order to oxidize the intermediate porphyrinogen. After workup and chromatographic separations, three porphyrins were isolated: etioporphyrin 6, the desired copper(I) [2]rotaxane 82+ , and the compartmental bis-copper(I) [3]rotaxane 104+ . The rotaxanes 82+ and 104+ were isolated as their PF–6 salts in 25% and 32% yields, respectively. All three compounds contained a free-base etioporphyrin and
Porphyrin Rotaxanes and Catenanes
223
Fig. 4 Precursors to the synthesis of [2]rotaxane 12+
Fig. 5 Synthesis of the various [2]- and [3]rotaxanes: (i) CH2 Cl2/CH3 CN, room temperature; (ii) di-tert-butyl-3,5-benzaldehyde, (diethyl-3,3 -dimethyl-4,4 -dipyrryl2,2 )methane 4, CF3 COOH, CH2 Cl2 , room temperature, then chloranil, CH2 Cl2 , reflux
could be readily metalated with Zn(OAc)2 ·2H2 O to afford 7, 92+ , and 114+ respectively. The rotaxanes, which were synthesized in this way, are complexed
224
L. Flamigni et al.
Fig. 6 Decomplexation reaction of the Cu(1)-complexed [2]-rotaxane 92+ (PF6 – )2 leading to the free-ligand [2]-rotaxane 12+ (PF6 – ): (i) KCN, CH3 CN/CH2 Cl2 /H2 0
copper(I) rotaxanes. The two constituent parts, a ring (macrocycle 3) and the dumbbell (the phenanthroline-bridged bisporphyrin), are held together by coordination to copper(I). Copper(I)- free [2]rotaxane 12+ (PF6 – ) was obtained by treatment of 2+ 9 (PF6 – )2 with excess KCN (Fig. 6). The fact that demetalation of Cu+ complexed rotaxane 92+ afforded a single product 12+ and the absence of release of the bis-porphyrin by macrocycle 3 is proof that the compound synthesized is a true rotaxane. Linking is provided only by steric crowding of the porphyrin moieties, preventing unthreading of the bis-porphyrin from macrocycle 3. Photoinduced Processes 1+ represents an important building block of several rotaxanes discussed in the present review article. In addition, the couple (PZn, PAu+ ) turned out to be an excellent choice for studying electron or energy transfer between porphyrins so that it is necessary to describe the properties of 1+ in detail, even if this molecule does not belong to the family of porphyrin catenanes and rotaxanes. The oblique bisporphyrin 1+ has been extensively investigated [11] and a schematic energy level diagram for the photoinduced processes occurring in 1+ in DMF is reported in Scheme 11 . Following excitation of zinc or 1
A generic nomenclature is used for the arrays: i.e. PZn represents the tetra-aryl or the etio zinc(II) porphyrin, PAu represents the positively charged corresponding gold(III) porphyrin, and Cu represents the copper(I) bis phenylphenantroline complex and so on. Therefore, a generic PZn-Cu-PAu represents any triad made of the above components. The processes between states are represented with arrows identified, in most cases, by numbers.
Porphyrin Rotaxanes and Catenanes
225
Scheme 1 Generic energy level diagram for compounds 1+ and 12+
gold porphyrin units (dotted lines), 1 PZn – PAu or PZn – 3 PAu are formed, respectively, and lead to the charge-separated state PZn+ – PAu– , reactions (1) and (2). Reaction (1) consists of an electron transfer from the LUMO localized on zinc porphyrin to the LUMO localized on gold porphyrin; reaction (2) is an electron transfer from the HOMO localized on PZn to the HOMO localized on PAu+ . The yield of these reactions is not unity: whereas 1 PZn – PAu intercrosses in part—according to its intrinsic deactivation—to the triplet 3 PZn – PAu (wavy line), PZn – 3 PAu transfers energy to the zinc porphyrin moiety in competition with electron transfer and yields 3 PZn – PAu with reaction (3). The latter excited state has enough energy to yield the chargeseparated state upon transfer of an electron from the zinc porphyrin triplet excited state to the gold porphyrin unit, reaction (4). The charge-separated state has a lifetime of ca. 600 ps and undergoes recombination to the ground state. Whereas reactions (2) and (4) are inactive in a rigid ethanol glass at 77 K, reaction (1) occurs under these conditions and an equilibrium is established between the singlet excited state 1 PZn – PAu and the charge-separated state PZn+ – PAu– , whose energy level is destabilized by the absence of solvent polarization in the rigid solvent by ca. 0.75 eV with respect to room temperature [52]. The recombination of the charge-separated state to the ground state takes place at 77 K with a rate of 7 × 107 s–1 . Detailed studies on the photoinduced electron transfer mechanism between PZn and PAu+ in 1+ point to the involvement of the orbitals of the bridging phenantroline, in a superexchange mechanism, whereas the charge recombination step (5) is likely to occur through the space separating the two units, very likely with the intermediation of solvent molecules [53].
226
L. Flamigni et al.
Scheme 2 Energy level diagram for compounds 92+ Table 1 Lifetimes of the main reactions reported in Scheme 1 and Scheme 2 a
1+ 9+ 12+ a
1 (ps)
2 (ps)
5 (ps)
6 (ps)
55 1.7 36
120 17 71
580 22 540
2500
Butyronitrile, room temperature
Following excitation of the porphyrin unit of the copper(I) [2]rotaxane 92+ (Scheme 2), electron transfer reactions (1) and (2) are faster than in 1+ , Table 12 . The CS state PZn+ – Cu – PAu– , in competition with charge recombination to the ground state, displays a charge shift reaction, (5), where the zinc porphyrin cation oxidizes the central copper(I) complex. Finally, the charge-separated state PZn – Cu+ – PAu– recombines to the ground state with a lifetime of 2.5 ns (6). The results obtained for 92+ , where the bridging ligand is incorporated into a copper [2]rotaxane that modulates the energy of HOMO and LUMO localized on the phenanthroline with respect to system The lifetimes τ(1/k) are reported on the tables. The reported lifetime represents in most cases the lifetime of the state from which the reaction is generated, i.e. the reverse of all the reaction rates depleting the state (1/Σk), therefore when more than one reaction compete to deplete the state it is not strictly correct to assign the lifetime of the state to the process. However, generally, the discussed processes are indeed the fastest and Σk can be approximated to k of the major process with very good approximation. When this is not the case, i.e. for slow reactions, the determination of the rate of the reaction is performed according to the corrected formula k = 1/τ – 1/τ0 , where τ and τ0 stand for the lifetime of the state in the presence of the depleting reaction and in the model, respectively. 2
Porphyrin Rotaxanes and Catenanes
227
1+ , were interpreted as confirmation of the superexchange mechanism for reaction (1) [54]. This was further supported by the results obtained with 12+ , the free rotaxane, where photophysical properties very similar to 1+ were evidenced, Scheme 1. The experimental kinetic data collected for these structures (Table 1) allowed us to derive a good correlation between the rates of electron transfer and the reciprocal of the energy gap between the relevant orbitals on the porphyrin excited state and on the phenanthroline ligand, confirming a superexchange mechanism for electron transfer, for reactions (1), (2) and probably also (4) [55]. 2.2 [3]- and [5]Rotaxanes with Porphyrin Stoppers The synthetic strategy is based on the threading of two rings onto the same two-chelate string (Fig. 7). As in the other cases, the driving force behind this reaction is the coordination of all the dpp chelates to Cu(I) centers. Since appropriate end-functions were previously introduced in the string, porphyrin stoppers can be constructed. As previously shown with similar compounds [56, 57], the threading steps (1) can be quantitative. The second step (2), can involve formation of terminal porphyrins only (upper route) or entail construction of both terminal and bridging porphyrin units (lower route). The chemical structures of the real precursors are given in Fig. 8. 1,10phenanthroline was reacted with the 4-lithio derivative of protected ben-
Fig. 7 Synthetic strategy for making porphyrin-stoppered multirotaxanes. The starting dialdehyde is a two-coordination site molecular string, the ring is also able to bind the metal (dot) while being threaded by the string. The threading (step 1) is assumed to be spontaneous and will thus not require any additional reagent. The porphyrin-forming reaction (step 2) requires the proper reagents and can, in principle, be directed mostly towards the construction of porphyrin stoppers or of both terminal and bridging porphyrins
228
L. Flamigni et al.
Fig. 8 Chemical structures of the precursors to the [3]- and [5]rotaxanes
zaldehyde [58] at 0 ◦ C in THF under argon. Following a classical procedure [59, 60], after hydrolysis and MnO2 re-oxidation, 13 was isolated pure in 71% yield. 14 was obtained by treating 13 with 1,4-dilithiobutane [61, 62] under similar conditions, eventually leading to the dialdehyde 15 after deprotection, the overall yield from 13 being 73%. Clearly, the dialdehydic string 15 can readily be made at the multi-gram scale. 3 is one of the most universally used cyclic building blocks in the Strasbourg group [63, 64]. The threading step [(1) of Fig. 7] was carried out by first mixing equimolar amounts of Cu(CH3 CN)4 + PF6 – and 3 and then adding 0.5 equivalents of 15. 172+ was indeed obtained quantitatively, as indicated by 1 H – NMR spectroscopy. It was treated with the dipyrrole derivative 16 [65] and 3,5di-tert-butylbenzaldehyde [51], under experimental conditions analogous to those originally described by Lindsey et al. [50]. The three components 172+ , 16 and 3,5-di-tert-butylbenzaldehyde (molar ratio 1 : 10 : 8) were
Porphyrin Rotaxanes and Catenanes
229
treated with CF3 COOH at room temperature in CH2 Cl2 . The porphyrinogen species formed were oxidized to the corresponding porphyrins with chloranil (31 equiv). After purification by repeated chromatography, three porphyrin-containing compounds could be isolated from the reaction mixture (Fig. 9): 5,15-bis(3,5-di-tert-butylphenyl)-2,8,12,18-tetrahexyl-3,7,13,17tetramethylporphyrin (18) in 32% yield, (Cu2 )-[3]rotaxane 192+ in 34% yield, and (Cu4 )-[5]rotaxane 204+ in 8% yield, the copper complexes being isolated as PF6 – salts [66, 67]. Noteworthy is the high yield of (Cu2 ) – [3]rotaxane 192+ . Since two porphyrins are created simultaneously, the yield of individual porphyrin formation is ca. 60%. (Cu4 )-[5]rotaxane 204+ deserves some comment. First of all, it is a molecule with four threaded macrocycles, that is, it has a [5]rotaxane structure; second, it results from the simultaneous formation of three porphyrins, the central one resulting from the condensation of two prerotaxane-like units. The yield of individual porphyrin formation is ca. 45% in this case. This copper(I)-complexed [5]rotaxane may be considered a “compartmental” [5]rotaxane, since two groups of two macrocycles each are separated by an inner porphyrin blocking group. The compounds of Fig. 9 (prerotaxane and rotaxane), as well as all the other [3]rotaxanes of this study, were characterized by fast atom bombardment (FAB) or electrospray (ES) mass spectrometry and 1 H NMR spectroscopy. Reaction of (Cu2 )-[3]rotaxane 192+ with zinc acetate afforded, after chromatography, (Cu2 )-[3]rotaxane 222+ in 83% yield. Porphyrin 18 and
Fig. 9 One-pot preparation of copper(I)-complexed [3]- and [5]rotaxanes: (i) di-tertbutyl-3,5-benzaldehyde (8 equiv), 16 (10 equiv) cat. CF3 CO2 H, CH2 Cl2 , room temperature, then chloranil (35 equiv), CH2 Cl2 , reflux
230
L. Flamigni et al.
(Cu4 )-[5]rotaxane 204+ were treated similarly, affording, respectively, Zn(II)porphyrin 21 and (Cu4 )-[5]rotaxane 234+ (Fig. 9) in 79% and 85% isolated yields, respectively. Demetalation of copper(I)-complexed catenates and rotaxanes is routinely performed by competitive complexation with cyanide, using a large excess of KCN [typically 20–25 mol/mol of Cu(I)] [68]. This reaction performed in similar conditions starting from 222+ [68 mol of KCN/mol of 222+ i.e. 34 mol/mol of Cu(I)] afforded cleanly and almost quantitatively the monodemetalated species 25+ , to our great surprise (Fig. 10). Prolonged reaction times or heating to 64 ◦ C did not improve the extent of demetalation. The simplest explanation is that, after removal of the first copper(I), which proceeds at a normal rate (i.e., similar to other analogous complexes), the molecule undergoes some profound rearrangement. The new conformation obtained is such that the second coordination site has now become fully protected from attack by CN– and is thus totally inert toward decomplexation. The 2,9-diphenyl-1,10-phenanthroline chelates embedded respectively in the macrocyclic and the dumbbell-shaped components of (Cu)-[3]rotaxane
Fig. 10 Chemical structure of metalated [3]rotaxanes 242+ and 25+ , as well as free [3]rotaxanes 26+ . Some models used for photophysical studies: 27+ , 28+ , 29+
Porphyrin Rotaxanes and Catenanes
231
25+ are in fact imprints of the metal template which has been partially removed. The tetradentate cavity initially formed around Cu(I) could be restored, by reacting (Cu)-[3]rotaxane 25+ , with AgBF4 in a CH2 Cl2 /CH3 CN mixture. (Ag,Cu)-[3]rotaxane 242+ , was isolated in 51% yield after chromatographic purification (Fig. 10). To test the influence of the metal cation incorporated in the porphyrin stoppers on the outcome of the demetalation reaction, (Cu2 )-[3]rotaxane containing a Zn(II)- and a Au(III)-porphyrin as stoppers was reacted with a slight excess of KCN in a CH2 Cl2 /H2 O mixture. Unexpectedly, total removal of the Cu(I) templating ions took place, leaving the free [3]rotaxane 26+ as product. It was isolated in 59% yield after chromatography. Photoinduced Processes The photophysical properties of [3]rotaxanes 222+ , 242+ , 25+ and 26+ were investigated in butyronitrile [69]. The discussion of processes occurring in 222+ , 242+ , 25+ the [3]rotaxanes containing zinc porphyrins as stoppers will be made with reference to a common energy level scheme, Scheme 3, where a generic PZn – Cu – M – PZn (with M = Cu, or Ag or nothing) will represent the systems. Excitation of the PZn unit in 222+ resulted in energy transfer from the singlet excited state localized on one of the zinc porphyrins, 1 PZn – Cu – Cu – PZn, to the excited state localized on the nearby Cu(dpp) + 2 unit, PZn – ∗ Cu – Cu – PZn, reaction (1). ∗ Cu is the metal-to-ligand charge transfer (MLCT) excited state of the Cu(dpp)2 + unit, at room temperature an equilibrium mixture of singlet and triplet excited state [70]. In 25+ —the compound missing one of the coordinating ions—deactivation of the zinc porphyrin singlet excited state, 1 PZn – Cu – PZn, displays a bi-exponential behavior with a 50% component unquenched with respect to the model PZn.
Scheme 3 Generic energy level scheme for compounds 222+ , 242+ , 25+ . M stands for Cu(I), Ag(I), or nothing
232
L. Flamigni et al.
We interpret the data as an energy transfer occurring only to the Cu(dpp)2 + unit in proximal position to the excited donor, whereas the excited donor which has an empty position in close proximity is unable, during its lifetime, to transfer energy to the distal Cu units. In practice, for this system, reaction (2) is inactive, Table 2. In 242+ , coordination of Ag+ introduces further levels in the energy scheme, i.e. the excited state localized on Ag(dpp)2 + , a ligand-centered (LC) excited state with an energy of ca. 2.5 eV (not shown in the diagram since its formation is endoergonic), and a charge-separated state reported in Scheme 3. The latter involves electron transfer (3) from the excited state localized on zinc porphyrin to the close Ag(dpp)2 + unit, PZn – Cu – Ag– – PZn+ and has an energy of ca. 1.3 eV. Quenching of the excited state 1 PZn – Cu – Ag – PZn is bi-exponential, reactions (1) and (2), and this is assigned to energy transfer from the zinc porphyrin localized excited state to either the proximal or distal Cu(dpp)2 + units. With respect to the case of 25+ , which has an empty coordination position, coordination of Ag+ increases the rigidity of the system and the electronic coupling between the different components, allowing in this case energy transfer (2) to the distal donor-acceptor couple. Formation of a charge-separated state by reaction (3) seems ineffective, since no transient absorbance due to CS state could be detected by a picosecond resolution apparatus, but we cannot exclude that there is some contribution of electron transfer to the decay of the excited state localized on zinc porphyrin in system 242+ . A common feature of the above discussed compounds, 222+ , 242+ , 25+ , is that the primary deactivation step is energy transfer leading from 1 PZn – Cu – M – PZn to the formation of PZn – ∗ Cu – M – PZn. This state, which is formed in different yields for the different systems, decays faster than the resolution (ca. 30 ps) by energy transfer to the triplet state localized on zinc porphyrin 3 PZn – Cu – M – PZn reaction (4). The latter reaction in 222+ , 242+ , 25+ can be evidenced by Table 2 Lifetime of the main reactions reported in Scheme 3 and Scheme 4 a
222+ 242+ 25+ 26+ a
1 (ps)
2 (ps)
4 (ps)
5 (ns) b
180 300 490 4600 d
180 4300 c — —
≤ 30 ≤ 30 ≤ 30 n.d.
5800 6300 6500
Butyronitrile, room temperature The lifetime of the model 3 PZn is 8000 ns c Calculated from a lifetime of PZn – Cu – M – 1 PZn of 1.3 ns compared to a lifetime of the model of 1.85 ns, according to: τ2 = 1/[(1/1300 ps) – (1/1850 ps)] d Calculated from a lifetime of 1PZn – PAu of 1.32 ns compared to a lifetime of the model of 1.85 ns, see note c b
Porphyrin Rotaxanes and Catenanes
233
the increase in the yield of 3 PZn compared to the model and its subsequent decay (5) occurs unperturbed. Quite remarkably, both quenching of 1 PZn – Cu – M – PZn luminescence and sensitization of 3 PZn – Cu – M – PZn phosphorescence takes place in 222+ , 242+ and 25+ also at 77 K in a rigid glass, confirming the energy transfer nature of the processes [69]. Lifetime data for the reactions in Scheme 3 are collected in Table 2. [3]rotaxane 26+ differs from the rotaxanes discussed above because of the presence in the structure of the electron-accepting unit gold (III) porphyrin and the absence of possible energy-acceptor units, Cu in the previous cases. Furthermore, because of the absence of coordinating metals, a more flexible structural frame is expected for this rotaxane which could exist in different conformations. Steady state and time-resolved luminescence data of 26+ indicate a different degree of quenching for the zinc porphyrin luminescence with respect to the model: 45% from steady state and 28% from the lifetime quenching, which could be fitted by a single exponential decay of 1.3 ns. This fact could be explained by the presence of two extreme conformations in 26+ ; a bent conformation with the porphyrins very close resulting in an immediate (“static”) through-space quenching which can only be detected by steady state methods and an extended conformation where the two partners are kept apart and the electron transfer occurs slowly through the linking chain. The latter is the conformation for which we can detect the exponential decay and which will be discussed with reference to the energy level diagram of Scheme 4. The quenching is assigned to electron transfer from the zinc porphyrin unit to the gold porphyrin unit, (1), to yield the charge-separated state PZn+ – PAu– which is at ca. 1.25 eV above the ground state. Electron transfer from the triplet states localized on both gold porphyrin and zinc porphyrin, reactions (2) and (3), are thermodynamically allowed but do not take place,
Scheme 4 Energy level diagram for 26+
234
L. Flamigni et al.
as testified by their decay which is the same as in the pertinent models, ca. 400 ps for 3 PAu+ and several microseconds for 3 PZn. The charge-separated state escaped our detection by transient absorption spectroscopy, because of the low formation yield (ca. 30%) and because of the close similarity of the CS spectrum with those of others transients, not allowing the detection of the charge recombination rate, reaction (4). At 77 K in a rigid glass, only a very modest quenching of the zinc porphyrin luminescence is detected. This is in agreement with an electron transfer process, since an important destabilization of the CS state upon freezing is expected and this would render the electron-transfer reaction iso- or endo-ergonic. The most important kinetic parameters of reactions in Scheme 3 are collected in Table 2. 2.3 [2]Rotaxanes Consisting of a Dumbbell with Two PZn Stoppers and a PAu+ -Appended Ring Rotaxanes made of a pendant gold-porphyrin macrocycle threaded inside a bis zinc-porphyrin stoppered dumbbell were built using a template approach, in order to assemble the two parts of the system, the dumbbell and the macrocycle. The template principle of construction is depicted in Fig. 11.
Fig. 11 Principle of transition metal-templated synthesis of a [2]rotaxane. A thick line represents a dpp chelate, a black dot represents a metal cation, a hatched diamond represents a Au(III) porphyrin and an empty diamond represents a Zn(II) porphyrin. The transition metal controls the threading of Au(III) porphyrin-pendant macrocycle (A) onto chelate (B), to form prerotaxane (C). Construction of the porphyrin stoppers at the X functions leads to the metal complex [2]rotaxane (D). Removal of the template cation forms the free rotaxane (E)
Porphyrin Rotaxanes and Catenanes
235
The precursor C to the rotaxane E, called prerotaxane, is obtained in one step from a gold porphyrin macrocycle A and a difunctionalized thread B, thanks to the gathering properties of the transition metal (black dot). Construction of the porphyrin blocking groups is the kinetically templated key step leading to the metal-complexed rotaxane structure D. The desired rotaxane E is obtained after removal of the metal template from D. The Cu(I)-complexed [2]rotaxane 332+ consists of a macrocycle incorporating a dpp chelate and bearing a pendant Au(III) porphyrin and two Zn(II) porphyrins as stoppering groups. The precursors and the synthetic route leading to this rotaxane are shown in Fig. 12 [71, 72]. Threading of the macrocycle 30+ [73, 74] onto 31 afforded the prerotaxane 322+ quantitatively. The two porphyrin stoppers were constructed in a one-pot procedure by reacting the above prerotaxane, 3,5-di-tert-butylbenzaldehyde [51], and the dipyrrole derivative 16 (see Fig. 8) [65]. After oxidation of the porphyrinogen with chloranil, and metalation with Zn(II), Cu(I)-complexed [2]rotaxane 332+ was obtained in 17% yield. Removal of Cu(I) was achieved quantitatively by treatment with KCN, giving 34+ (Fig. 13). As evidenced by 1 H NMR spectroscopy, demetalation is followed by a translation motion of the dumbbell component towards the Au(III) porphyrin of the macrocycle, showing that in both compounds (332+ and 34+ ) the two porphyrin stoppers of the dumbbell sandwich the phenanthroline subunit of the macrocycle. A rich coordination chemistry starting from 34+ was developed at the bis-dpp tetrahedral site left by the metal template. Thus, rotaxane 34+ was metalated with monocations Ag+ (352+ ) and Li+ (362+ ) by treatment with AgBF4 or LiBF4 , respectively (Fig. 13) [73–75]. In the metalated rotaxanes 332+ , 352+ and 362+ the components incorporating the chromophores and electrophores are connected by metal–ligand bonds, being thus appropriate models for the study of electron transfer through metal–ligand bonds. In the demetalated rotaxane 34+ , there is no classical chemical connection between the electroactive moieties, so this molecule is a suitable model for studying through-space electron transfer processes. In order to have a suitable reference compound for the photophysical properties of 332+ , compound 37+ represented in Fig. 14 was also synthesized using a procedure similar to the preparation of 332+ . Photoinduced Processes The photophysical properties of 332+ , 34+ and 352+ were investigated in DMF and the results are here summarized [72, 75]. In these systems the redox active porphyrins are on different parts of the structure which are held together by the coordinated metal in 332+ and 352+ , whereas in 34+ they are held together only by mechanical bonds and no covalent electron-transfer pathways exist. These systems are characterized by some flexibility of the polyether gold porphyrin appended macrocycle which could cause, upon
236
Fig. 12 Synthesis of the Cu(I)-complexed [2]rotaxanes 332+
L. Flamigni et al.
Porphyrin Rotaxanes and Catenanes
237
Fig. 13 Illustration of the demetalation/remetalation reactions carried out on the Cu(I)complexed [2]rotaxane 332+ to afford free [2]rotaxane 34+ and metallo-[2]rotaxanes 352+ and 362+ . In the free [2]rotaxane 34+ , the macrocycle is more deeply buried inside the cavity formed by the bis-porphyrin dumbbell compared to the complexed [2]rotaxanes
bending, a close approach of the terminal donor and acceptor. The photophysical data in DMF presented here are interpreted by the presence of two independent conformations, presumably a bent and an extended one, non-
238
L. Flamigni et al.
Fig. 14 Cu(I)-complexed [2]rotaxanes 37+
equilibrating on the timescale of the determinations. This results in a double exponential behavior for the kinetics processes of the intermediates, since the two conformations display a quite different reactivity. The energy level scheme for 34+ is reported in Scheme 5. Excitation of the zinc porphyrin yields the excited state 1 PZn2 – PAu, localized on one of the zinc porphyrin stoppers, which decays with a biphasic rapid electron transfer to the gold porphyrin unit to yield the CS state PZn2 + – PAu– , (1). The same CS state is formed also upon excitation into the gold porphyrin, reactions (2). The triplet state localized on zinc porphyrin, 3 PZn2 – PAu, is capable of forming the charge-separated state, reaction (3), but its rate(s) can not be resolved. The CS state finally decays with two lifetimes, assigned to the recombination to the ground state of the two different conformations, reaction (4) [72]. The lifetimes of the reactions are summarized in Table 3. In 352+ the zinc- and gold porphyrin subunits are connected by an Ag(dpp)2 + fragment; the energy level scheme can be considered similar to 34+ except for a few differences detailed below (Scheme 6). Whereas the excited state localized on Ag(dpp)2 + , a ligand-centered triplet excited state (3 LC) is at an energy (ca. 2.5 eV) well above the energies involved here and can be ignored, a charge-separated state with the oxidized zinc porphyrin and reduced Ag(dpp)2 + , PZn2 + – Ag– – PAu, could be formed by 1 PZn – Ag – PAu, with a ∆G0 = – 0.7 eV [72, 76]. However, 352+ displayed 2 the same reactivity as 34+ upon excitation in the zinc porphyrin manifold, Table 3, allowing us to exclude any involvement of Ag(dpp)2 + in the processes originated from 1 PZn2 – Ag – PAu. On the contrary, excitation of the gold por-
Porphyrin Rotaxanes and Catenanes
239
Scheme 5 Energy level diagram for 34+ Table 3 Lifetime of the main reactions reported in Schemes 5–7 a . The reactions are biexponentials for the presence of two different conformations (see text) 1 (ps)
2 (ps)
4 (ns)
34+
80, 570
90, 1000
352+ 332+
80, 700 90, 300
50, 400 25, 470
10, 10, 10, 12,
a b c
35 b 40 c 40 40
5 (ps)
6 (ps)
—
—
180
650
Dimethylformamide, room temperature Excitation of the zinc porphyrin unit Excitation of the gold porphyrin unit
phyrin unit gave different results: PZn2 – Ag – 3 PAu reacts faster to form the charge-separated state PZn2 + – Ag – PAu– than PZn2 – 3 PAu in 34+ to form the charge-separated state PZn2 + – PAu– , in spite of identical energetic and structural parameters (Table 3). Coordination of Ag+ increases the electronic coupling between the phenanthrolines on the macrocycle and on the dumbbell and this results in an enhancement of the reaction rate of (2) operated by a superexchange mechanism. The effect of Ag+ on reaction (2) and not on reaction (1) is interpreted as a major effect of Ag+ coordination on the HOMO of the phenanthroline, involved in electron-transfer reaction (2), rather than on the LUMO orbitals, involved in electron-transfer reaction (1) [75]. Coordination of the Cu+ ion in 332+ yields a different picture; the energy level scheme is reported in Scheme 7. The MLCT excited state of the Cu(dpp)2 + unit, ∗ Cu, is lower in energy than the singlet excited state of
240
L. Flamigni et al.
Scheme 6 Energy level diagram for 352+
Scheme 7 Energy level diagram for 332+ . The numeration of the reactions is consistent with Schemes 5 and 6
the zinc porphyrin component and has some role in the photoinduced processes. Following excitation of the zinc porphyrin to 1 PZn2 – Cu – PAu, an electron transfer to the gold porphyrin reaction (1), competes with an energy transfer to the Cu(dpp)2 + unit, reaction (5), leading to the excited state PZn2 – ∗ Cu – PAu. This in turn transfers energy to the zinc porphyrin unit, (6), yielding 3 PZn2 – Cu – PAu. The further evolution is similar to that detected for 34+ and no involvement of the charge-separated state PZn2 – Cu+ – PAu– , where the copper unit has transferred an electron to the gold porphyrin moiety, was evidenced. The final recombination of the chargeseparated state PZn2 + – Cu – PAu– , reaction (4), is essentially identical to that
Porphyrin Rotaxanes and Catenanes
241
detected for the copper-free rotaxane. Excitation of the gold porphyrin unit results in enhancement of the rate of reaction 2 with respect to the free rotaxane 34+ , in analogy to the Ag+ rotaxane 352+ . In this case, the effect of Cu+ could be that of introducing a real state, PZn2 – Cu+ – PAu– , in the electrontransfer process, rather than of introducing a virtual state as in the case of the enhanced superexchange mechanism with Ag+ ion. However, the steps could not be resolved and only the overall rate of reaction (2) could be determined (Table 3). 2.4 Porphyrin [2]Rotaxanes for Which a PAu+ Fragment is an Integral Part of the Ring Other systems of interest in order to study through-space electron transfer processes or through-bond electron transfer with different metals connecting the macrocycle and the dumbbell were designed. In these [2]rotaxanes, the Zn(II) bis-porphyrin dumbbell component remains the same as in Sect. 2.3, whereas the gold porphyrin is incorporated in the macrocycle. This affects the mutual arrangement of the chromophores and therefore their electronic communication. It also changes remarkably the dynamic of the system after removal of the metal template as will be described. Cu(I)-complexed [2]rotaxane 402+ was prepared in a similar way as 332+ (see Fig. 12) using the transition-metal technique to assemble the gold(III) porphyrin-containing macrocycle 38+ and a precursor to the dumbbell, the open chelate 2,9-bis[p-(formylphenyl)]-1,10-phenanthroline 31 (Fig. 15) [77, 78]. Prerotaxane 392+ was obtained quantitatively and used directly in the next step. Porphyrin stoppers were constructed at the protruding aldehyde functions as follows: a mixture of prerotaxane 392+ (1 equiv.), 3,5-di-tertbutylbenzaldehyde [51] (8 equiv.), 16 (10 equiv.) [65], and a few drops of trifluoroacetic acid in CH2 Cl2 was stirred at RT overnight. Tetrapyrrole assembly was fixed by controlled oxidation of the porphyrinogens with chloranil. Cu(I)-complexed [2]rotaxane 402+ was isolated in 13% yield, after chromatographic purification and Zn(II) insertion into the porphyrin stoppers. The metal template (Cu(I)) was selectively extruded by reacting the [2]rotaxane complex 402+ with KCN (50 equiv.) [68]. This decomplexation reaction liberated the free [2]rotaxane 41+ quantitatively. As studied by 1 H NMR spectroscopy, and depicted in Fig. 16, the template imprint (a bis-dpp, tetrahedral coordination sphere) has completely vanished through rearrangement of the threaded macrocycle around its axle [78]. Recomplexation of 41+ with Ag+ or Li+ by reaction with AgBF4 or LiBF4 restored the template imprint and afforded the Ag+ - and Li+ -[2]rotaxane complexes 422+ and 432+ quantitatively. The rotaxanes described in this paragraph constitute representative examples showing that complexing or decomplexing the appropriate metal in a coordination site can bring to close proximity, or spread a long distance
242
L. Flamigni et al.
Fig. 15 Cu(I)-directed template synthesis of [2]rotaxane 402+
apart, the porphyrin components of the system, in relation to molecular machines [25–28]. In the complex, the Au-porphyrin is remote from the two Zn-porphyrins. After removal of the central metal, weak forces may favor
Porphyrin Rotaxanes and Catenanes
243
Fig. 16 Illustration of the demetalation/remetalation reactions carried out on Cu(I)complexed [2]rotaxane 402+ to afford free [2]rotaxane 41+ and metallo-[2]rotaxanes 422+ and 432+ and showing the pirouetting motion of the Au(III)-incorporating macrocycle upon removal of the central metal. Indicated with arrows are the rOe correlations as observed by proton NMR spectroscopy
244
L. Flamigni et al.
an attractive interaction between the Au(III) porphyrin and the Zn(II) nuclei leading to a situation in which the Au(III) porphyrin is pinched between the two Zn(II) porphyrin units. The interconversion between both situations implies a half-turn rotation of the threaded fragment within the ring. It also leads to dramatic electron transfer property differences between both situations. Photoinduced Processes The photophysical properties of the copper rotaxanes 402+ and the free rotaxane 41+ have been investigated in acetonitrile and the results are summarized below [78]. These structures differ from 332+ and 34+ previously described because of a reduced flexibility; the macrocycle is now more rigid due to the insertion in the ring of the rigid gold (III) porphyrin structure. As a consequence in 402+ no evidence was found of the existence of different conformations. This is not the case for the free rotaxane 41+ which, upon removal of the copper ion, can exist in two extreme conformations: a compact one with the gold porphyrin between the zinc porphyrin stoppers, referred to as “short” and a fully extended conformation referred to as “long” with the gold and zinc porphyrins at opposite positions resembling the structure of the metal rotaxanes 402+ , 422+ and 432+ . The short to long conformer distribution depends on the solvent and in acetonitrile the population consists of ca. 70% short to 30% long, as determined by comparative time resolved and steady state luminescence studies [78]. The data reported below for 41+ are related to the long conformer—here the processes could be resolved—whereas in the short conformer the processes occur through space at close contact and are faster or of the same order of time resolution (ca. 30 ps). Excitation of 41+ into the zinc porphyrin local-
Scheme 8 Energy level diagram for 41+
Porphyrin Rotaxanes and Catenanes
245
ized singlet, 1 PZn2 – PAu, Scheme 8, results in quenching of the luminescence lifetime to 170 ps. This is ascribed to an electron transfer, reaction (1), to the gold porphyrin unit to yield the charge-separated state PZn2 + – PAu– . Excitation in the gold porphyrin manifold which yields rapidly the triplet localized on gold porphyrin unit PZn2 – 3 PAu, does not result in any change in the lifetime of the gold porphyrin triplet with respect to the model 3 PAu+ , indicating that the gold porphyrin excited state is inactive in promoting intramolecular processes in the long conformation. Its short lifetime, 1.5 ns, and the distance from the zinc porphyrin partner hampers its participation in energy- or electron-transfer reactions, both possible on energetic grounds, see Scheme 8. Recombination to the ground state of PZn2 + – PAu– , reaction (2), takes place with a lifetime of ca. 6 ns and it could be dictated by the inter-conversion lifetime of the long to short conformation. In the latter in fact, recombination is instantaneous. Photoinduced processes in 402+ are more complex due to the presence of the Cu(dpp)2 + unit, which introduces further levels in the energy diagram, Scheme 9. Upon excitation, the luminescence of the zinc porphyrin excited state, 1 PZn2 – Cu – PAu, is quenched with a lifetime of 180 ps, identical to the lifetime of the energy transfer process detected in the strictly correlated [3]rotaxane 222+ and was therefore ascribed to energy transfer to the copper complex unit to form PZn2 – ∗ Cu – PAu, reaction (1) (Table 4). Transient absorption spectra show, with the same lifetime of 180 ps, the formation of the charge-separated state PZn2 + – Cu – PAu– . This is assumed to be formed by the consecutive fast reactions (3) and (4). The former is an electron transfer from the excited state localized on the copper complex PZn2 – ∗ Cu – PAu
Scheme 9 Energy level diagram for 402+ . The numeration of the reactions is consistent with Scheme 8
246
L. Flamigni et al.
Table 4 Lifetime of the main reactions reported in Schemes 8 and 9 a
41+ 402+ a b c
1 (ps)
2 (ns)
170 b 160 c 180 b 170 c
5.5 5
3 (ps)
4 (ps)
≤ 30
≤ 30
Acetonitrile, room temperature From luminescence quenching of the zinc porphyrin unit From transient absorbance
to the gold porphyrin to yield PZn2 – Cu+ – PAu– , the latter is a charge shift from one of the zinc porphyrin stoppers to the copper unit, easier to reduce in this array than the zinc porphyrin. Energy transfer from PZn2 – ∗ Cu – PAu to the zinc porphyrin unit to yield 3 PZn2 – Cu – PAu was shown to be unimportant or null on the basis of the measured triplet yield. The gold porphyrin localized triplet excited state PZn2 – Cu – 3 PAu does not show any intramolecular reactivity and decays with the same lifetime detected in the corresponding model, 1.5 ns. The charge-separated state PZn2 + – Cu – PAu– , which is essentially the only product of the photoreaction, recombines to the ground state with a lifetime of 5 ns, reaction (2).
3 Porphyrin-Containing Catenanes 3.1 A [2]Catenane Whose Two Rings Incorporate a PZn and a PAu+ Unit. A natural extension of the work described in Sect. 2.4 was to prepare a Cu(I)complexed [2]catenane with macrocycles incorporating the Zn and Au porphyrins. [2]catenanes are topologically non-trivial molecules (non-planar molecular graph) in which two rings are interlocked but not linked [23, 24]. Therefore, differentiating the rings with Zn- and Au-porphyrins enables us to study photoinduced electron transfer in mechanical bond systems and also to have information on the conformation of the system. The target Cu(I)-complexed [2]catenane contains two different macrocycles. Therefore, two routes can be envisaged for its construction by the transition metal templated strategy [64]. They are shown in Fig. 17. Both involve the preparation of an intermediate precatenane species, (C) or (F), in which either Zn or Au porphyrin-containing macrocycle (A) or (E) is threaded onto chelate (B), thanks to copper(I) coordination. Formation of the second, inter-
Porphyrin Rotaxanes and Catenanes
247
Fig. 17 Two strategies for the transition metal-templated synthesis of a [2]catenane made with Zn and Au porphyrin-incorporating interlocked macrocycles. The thick lines represent chelating fragments, the black disk symbolises copper(I), the empty diamonds are Zn(II) porphyrins, and the hatched diamonds are Au(III) porphyrins
locking macrocycle is achieved in the next step by reaction of the precatenane with the appropriate porphyrin, (D) in the case of (C), (G) in the case of (F), to produce the desired Cu(I)-complexed [2]catenate (H). Finally, removal of the metal template affords the free catenane species (I). The precursors of the target Cu(I) complexed [2]catenate 492+ are shown in Fig. 18. The two steps of the most efficient route leading to the Cu(I) complexed [2]catenate are shown in Fig. 19 [79]. Mixing of equimolar solutions of [Cu(CH3 CN)4 ]PF6 in acetonitrile and Zn porphyrin-containing macrocycle 44 [80] in dichloromethane, followed by addition of phenanthroline derivative 45 in dichloromethane afforded in quantitative yield precatenate 48+ . Subsequently, this complex was combined with a stoichiometric amount of gold(III) 5,10-di(p-hydroxyphenyl)-15,20di(3,5-di-tert-butylphenyl)porphyrinate [47] PF6 in DMF. The resulting so-
248
L. Flamigni et al.
Fig. 18 Precursors to [2]catenate 492+ and a model compound 51+
lution was treated portionwise with a suspension of Cs2 CO3 in DMF. This procedure allows us to overcome the relative instability of the present precatenate in basic medium [81, 82]. In these conditions, Cu(I) complexed [2]catenate [49] (PF6 )2 was isolated in 11.5% yield after chromatography. The alternative route, which involves Au porphyrin-containing macrocycle 38+ [77] and Zn porphyrin 46 as reactants afforded the same copper catenate in 5% yield. Demetalation leading to the free [2]catenane species 50+ was carried out by treating the Cu(I) complex with 100 mol % KCN [64]. [2]catenane 50+ was obtained in 83% yield after purification by column chromatography (Fig. 20). The 1 H NMR spectrum of the free [2]catenane 50+ is dramatically different from that of its parent Cu(I) complex 492+ . Dipolar correlations show that upon demetalation the catenane goes from an extended conformation to a more compact one in which the two porphyrins are closer together on average but have no significant interactions, as represented in Fig. 20. Photoinduced Processes The photophysical properties of [2]catenane 50+ , of the copper(I) catenate 492+ and of the models 51+ , 44 and 38+ have been investigated in acetonitrile solutions [83]. Compared to the [2]rotaxanes addressed in the previous section, 402+ and 41+ , the center-to-center distance between donor and acceptor has increased from ca. 1.9 nm to ca. 2.6 nm and the zinc porphyrin is differ-
Porphyrin Rotaxanes and Catenanes
249
Fig. 19 Two-step preparation of Cu(I)-complexed [2]catenate 492+
ent. The two etio-type zinc-porphyrin stoppers of the rotaxanes have in fact been replaced here by a tetra-aryl zinc porphyrin which is included in one of the interlocked macrocycles. The redox and spectroscopic properties of the etio porphyrins are slightly different from the tetra-aryl derivative. The ex-
250
L. Flamigni et al.
Fig. 20 Demetalation of [2]catenate 492+ to afford [2]catenane 50+
cited states of the latter are slightly lower in energy, the singlet by ca. 0.08 eV and the triplet by ca. 0.15 eV, and the oxidation is more difficult by ca. 0.1 V. This introduces some changes in the energy levels, in the thermodynamic parameters and in the reaction rates of the various reactions with respect to the corresponding rotaxanes 402+ and 41+ . The copper free [2]catenane 50+ can assume different conformations: a short one as represented in Fig. 20, and a long one, similar to that of the copper [2]catenane 492+ . In acetonitrile the short to long conformer distribution is of ca. 60% short to 40% long, determined by comparative steady-state and time-resolved luminescence. In this case some of the processes occurring in the short conformation could be resolved, being slower than in the corresponding short conformation of [2]rotaxane 41+ , but in this chapter we will only discuss the processes occurring in the long conformation [83].
Porphyrin Rotaxanes and Catenanes
251
Scheme 10 Energy level diagram for compounds 50+
Excitation of 50+ into the zinc porphyrin localized singlet, 1 PZn – PAu, Scheme 10, results in quenching of the luminescence lifetime to 570 ps. This is ascribed to an electron transfer to the gold porphyrin unit, reaction (1), to yield the charge-separated state PZn+ – PAu– . Excitation in the gold porphyrin manifold, which yields rapidly the triplet localized on the gold porphyrin unit PZn – 3 PAu, does not result in any change of the lifetime of this state with respect to the model 3 PAu+ , indicating that the gold porphyrin excited state does not participate in intramolecular processes in the long conformation, as already observed for [2]rotaxane 41+ . Recombination to the ground state of PZn+ – PAu– , reaction (2), takes place with a lifetime of 12 ns. Very likely this recombination lifetime is governed by the inter-conversion rate from the long to the short conformation, where recombination takes place immediately. The reaction lifetimes are collected in Table 5.
Table 5 Lifetime of the main reactions reported in Schemes 10 and 11 a
50+ 492+ a b c
1 (ps)
2 (ns)
570 b 570 c 540 b 550 c
12 10
Acetonitrile, room temperature From luminescence quenching of the zinc porphyrin unit From transient absorbance
3 (ps)
≤ 30
252
L. Flamigni et al.
Scheme 11 Energy level diagram for compounds 492+
The energy level of the copper(I) catenate 492+ (Scheme 11) is remarkably different from that of the corresponding [2]rotaxane 402+ : the chargeseparated states with the hole on zinc porphyrin, PZn+ – Cu – PAu– , and the one with the hole on the copper unit, PZn – Cu+ – PAu– , for 492+ are at the same energy (compare Scheme 9 with Scheme 11). In fact, at variance with the properties of 402+ , in 492+ the copper complex and the porphyrin donor oxidize at the same potential and this causes isoergonicity of the two chargeseparated states. Excitation in this array of the zinc porphyrin, leading to 1 PZn – Cu – PAu, results in quenching of the luminescence to a lifetime of 540 ps, compared to a lifetime of the model 44 of 2.1 ns. In the model catenate 51+ without the gold porphyrin, where the only possible deactivation path is energy transfer to the copper complex, the lifetime of the excited zinc porphyrin is 320 ps. As a consequence, quenching of 1 PZn – Cu – PAu was identified as a similar (with some perturbation for the presence of the gold porphyrin) energy transfer process to PZn – ∗ Cu – PAu, process (1). The latter state deactivates faster than our resolution by electron transfer to the gold porphyrin unit yielding PZn – Cu+ – PAu– , process (3). No increase in the yields of the triplets localized on both zinc- and gold porphyrin, formed by the possible alternative deactivation paths, were detected, confirming the proposed mechanism. For the ensuing PZn – Cu+ – PAu– charge-separated state there is no driving force to promote an electron from the terminal zinc porphyrin unit to the oxidized copper complex unit (Scheme 11). Therefore, the recombination to the ground state occurs from PZn – Cu+ – PAu– , as testified by the absence of the typical zinc porphyrin cation band (λmax at 670 nm) in the spectrum of the charge-separated state. The lifetime of charge recombination (2) is 10 ns (Table 5). Excitation in the gold porphyrin unit leads to PZn – Cu – 3 PAu which decays unperturbed to the ground state with the same lifetime as the model.
Porphyrin Rotaxanes and Catenanes
253
3.2 Copper(I)-Templated Synthesis of [2]Catenanes Bearing Pendant Porphyrins The main objective of this work was to synthesize a [2]catenane, with donor and acceptor porphyrins arranged linearly on each side of the catenane core [73]. The chemical realization of this aim, the Cu(I)-complexed catenate
Fig. 21 Synthetic steps to prepare Cu(I)-complexed [2]catenate 552+
254
L. Flamigni et al.
552+ , consisting of macrocyclic components, one attached to a gold(III) porphyrin and the other to a zinc(II) porphyrin, was achieved using the synthesis depicted in Fig. 21, whereas other synthetic routes where also tested [74]. The precatenane 532+ incorporating the macrocycle bearing a gold(III) porphyrin 30+ was first generated quantitatively by formation of the copper(I) complex of 30+ and subsequent threading of the diiodo-derivative 52. This complex was then reacted with the diphenolic porphyrin 54 in the presence of cesium carbonate to afford the desired Cu(I)-complexed [2]catenate in 16% yield. In addition, the component macrocycles and the bis-zinc(II) porphyrin catenate were isolated. As in the previous catenation, these are formed as a result of decomposition and “scrambling” of the initial precatenane 532+ .
4 Related Catenanes and Rotaxanes Recently Reported in the Literature Several examples of photo-active porphyrin-containing catenanes and rotaxanes have been reported recently [84–86], but we will here examine only the few cases based on copper(I) template synthesis of the interlocked structure and as such, strictly related to the structures here discussed [87–90]. These reports, by Schuster and Guldi, present both the synthetic procedure and an account of the photoinduced processes taking place in the array. The latter will be briefly summarized and discussed with reference to the main results on the systems presented in the previous chapters. It should be noticed that the choice of the photoactive components for the arrays reported by Schuster and Guldi is in part different from the choices discussed above, whereas a similar zinc(II) porphyrin was used as photosensitizer and electron donor, [60] fullerene was used as the electron acceptor in the array. Replacing the gold(III) porphyrin acceptor with C60 has several immediate effects on the properties of the systems. The molar absorption coefficients in the visible range decrease by a factor of ca. 2, because of the poor absorption properties of the C60 component above 400 nm, and the energy which can be stored in the excited state of the fullerene is lower than the one stored in the corresponding excited states of the gold (III) porphyrin. However, C60 excited states are much longer lived, a few nanoseconds for 1C 3 60 and a few tens of microseconds for C60 [90] to be compared with 1 + 240 fs [91] for PAu and 1.5 ns [78] or ca. 0.4 ns [69] for 3 PAu+ depending on the substitution pattern. The longer lifetime would allow the involvement of C60 excited states in processes with a poor driving force, i.e. rather slow. In this case, in fact, even slow intra-molecular processes could compete with the intrinsic deactivation processes. The reduction potentials for the two electron acceptors are similar, ca. – 1000 mV vs. Fc/Fc + [78, 87] but the reorganization energy is lower for 1 C60 (ca. 0.6 eV [92]) than for gold(III) porphyrin (1.2 eV [93]), because of the three-dimensional structure of this
Porphyrin Rotaxanes and Catenanes
255
electron acceptor which can easily accommodate one or more electrons without large distortions. The latter parameter is very critical for the kinetic stability of the charge-separated state formed upon photoinduced electron transfer since it can shift the Marcus inverted region to less negative ∆G0 and, in conditions of similar driving force for charge recombination with respect to the gold(III) porphyrin acceptor, slow down the recombination of charges [94]. To summarize, the class of compounds where C60 has replaced gold(III) porphyrin is expected to display less positive performances with respect to light absorption, but to have a better yield of charge separation and longer lived charge-separated states, i.e. better performance in the conversion of light- into chemical-energy. The systems studied by Schuster and Guldi are reported in Fig. 22; 56+ and + 57 are analogous to rotaxanes 332+ and 402+ with two zinc(II) porphyrins as stoppers, a Cu(I)(dpp)2 + as the spacer and a C60 as the electron acceptor are appended to the macrocycle or included in it. 58+ is a catenane derived from 56+ upon axial binding by a bidentate ligand to the zinc ions of the porphyrin stoppers. Rotaxanes 59+ and 60+ differ from 56+ and 57+ in having two C60 electron acceptors as stoppers and a zinc porphyrin electron donor appended to the macrocycle and differ from each other in the distance between the electron donor and the copper complex, which has been increased by insertion of a phenylamido group.
Fig. 22 Chemical structures of compounds 56+ –60+ [87–89]
256
L. Flamigni et al.
Intramolecular association between C60 and zinc porphyrins are rather common events in multi component systems containing these chromophores [95], however, none of these interlocked structures are reported to display such a tendency, in spite of the rather flexible polyether chain connecting the donor and the acceptor to the central copper phenanthroline, which could easily allow a close approach of the terminal units. A general, schematic energy level diagram for structures 56+ –60+ , where the main photodeactivation paths are reported, is shown in Scheme 12. We have made use of the same type of nomenclature adopted in this presentation; the generic array containing zinc(II) porphyrin(s), copper(I) bisphenyl phenanthroline and a C60 moiety is named PZn – Cu – C60 , which allows us to easily identify each component and the transfer of energy or electrons between the units; the reaction paths are identified with numbers. The first process occurring in all structures upon excitation of the zinc porphyrin component, is identified as an energy-transfer process (1) from the singlet excited state localized on the porphyrin(s), 1 PZn – Cu – C60 , to the MLCT excited state localized on the copper complex, PZn – ∗ Cu – C60 . This is in full agreement with what occurs in systems 222+ , 402+ , 48+ , and 492+ discussed above, where energy transfer takes place with lifetimes ranging from 0.17 ns to 0.54 ns, depending on the thermodynamic and distance parameters. Energy transfer occurs in 1 ns in 56+ , 57+ and 58+ , reflecting similar conformation, distance, and thermodynamics for the three structures, but it is quite faster in 59+ and 60+ , respectively 0.22 ns and 0.44 ns. This has been explained both by the changes in through-bond distance between the porphyrin donor and the copper complex acceptor—shorter for 59+ with respect to 60+ and both shorter compared to 56+ , 57+ and 58+ —and to different
Scheme 12 Generic energy level scheme for compounds 56+ , 57+ , 58+ , 59+ , 60+
Porphyrin Rotaxanes and Catenanes
257
conformations. In fact, whereas in 59+ the zinc porphyrin and the closest phenanthroline planes are coplanar, they are almost perpendicular in 60+ . The subsequent electron transfer reaction (2) from the MLCT excited state localized on the complex, PZn – ∗ Cu – C60 , to the fullerene moiety which yields the charge-separated state with an oxidized copper complex and a reduced C60 unit, PZn – Cu+ – C60 – , is independent of the structure, being of the order of 0.5 ns for all samples. The authors could determine the rate of reaction (2) by detecting the luminescence quenching of the MLCT excited state of the copper phenanthroline complex (λmax = 710 nm). This luminescence could not be observed for any of the systems previously reported because of the intense emission of the zinc porphyrin donor in the spectral region of interest—Φfl for porphyrin is of the order of 0.5 and for Cu(phen)2 + it is of the order of 10–4 —and because of the poor concentration developed for this state in a consecutive reaction with a formation rate (1) of the same order or lower than the decay rate (2). For this reaction the authors can identify a transient absorption spectrum assigned to the fully charge-separated state PZn+ – Cu – C60 – , where the hole is localized on the zinc porphyrin and the extra electron localized on the C60 ; the band of C60 – can be detected around 1100 nm and the band of PZn+ around 650 nm. No estimate of yield of charge separation or of long-lived states, e.g. the triplet state localized on zinc porphyrin, was attempted. The formation of the fully charge-separated state, PZn+ – Cu – C60 – , was assigned to a very rapid and complete charge shift from the intermediate charge-separated state PZn – Cu+ – C60 – , reaction (3). The recombination to the ground state from PZn+ – Cu – C60 – exhibits a dependence on the driving force, slower for larger driving forces, typical of reactions occurring in the Marcus inverted region. It is in fact slower for the bis-substituted fullerene derivative 57+ , more difficult to reduce with respect to the mono-substituted fullerene in 56+ and it is slower for the less polar solvent tetrahydrofurane (ε = 7.58) compared to dichloromethane (ε = 8.93) for compounds 59+ and 60+ (Table 6). Catenane 58+ , formed with an association constant Ka of ca. 2 × 105 M–1 by coordination of 56+ with ditopic ligands 4,4 -bipyridine and DABCO, does not display any difference from the parent rotaxane 56+ . It is evident from this Table 6 Lifetimes of the reactions reported in Scheme 12
56+ 57+ 58+ 59+ 60+
1 (ns)
2 (ns)
5 CH2 Cl2 (µs)
5 THF (µs)
1 1 1 0.22 0.44
0.50 0.56 — 0.58 0.59
0.49 1.17 0.5 0.73 29
— — — 0.89 32
258
L. Flamigni et al.
that the stabilization in the energy level of the excited state localized on the zinc porphyrin upon complexation, of the order of ca. 0.05 eV [96] which alters the ∆G0 of the energy transfer reaction (1), does not lead to any sizeable change in the rate of reaction (1). The most remarkable feature of the systems discussed in this chapter is the fast and complete occurrence of reaction (3). The driving force for this reaction is null, since both the copper(I) complex and the zinc(II) porphyrin oxidize at the same potential, nonetheless the formation of the fully chargeseparated state is fast and complete and, in spite of its extremely long lifetime, the reverse reaction (4) does not occur. This is quite different from the results reported above for 492+ where, in the presence of a null driving force for a charge shift to the fully charge-separated state PZn+ – Cu – PAu– , a charge shift did not take place and recombination occurred from the intermediate charge-separated state PZn – Cu+ – PAu– , with the extra charges localized on the proximal copper complex and gold porphyrin. In conclusion, these systems appear quite unique, since they combine a large efficiency of charge separation, which can be calculated from the lifetimes reported by the authors in the range of 76–93%, and display remarkable charge-separated state lifetimes, from 0.5 µs to 32 µs. Whether this is due to the intrinsic properties of the C60 chromophore or is a combination of several favorable conditions met in the present systems is difficult to assess on the basis of the present data.
5 Conclusions and Perspectives Several families of porphyrin rotaxanes and catenanes have been synthesized using the classic 3D-template effect of copper(I). The molecules contain up to 5 constitutive organic fragments ([5]rotaxane). By removing the metal template, totally different geometries are obtained. Detailed electron- and energy-transfer studies have been carried out, on the copper(I) complexes, on the demetalated species and, in some cases, on complexes obtained by exchanging the copper center for another cationic metal (Li+ , Ag+ ). The photophysical properties have been studied in different solvents, however, being solvents of similar polarity, the results for different systems can be considered comparable. Time-resolved measurements allowed us to analyze in detail several electron and energy transfer processes occurring in cascade in some of the highly multicomponent catenanes and rotaxanes of the present article. In addition, it has been possible to identify very distinct conformers in solution, some of them containing two chromophores located close to one another and others with remote chromophores. It is difficult to derive simple and absolute rules to improve photoinduced charge separation efficiency in interlocked assemblies, but we have gained
Porphyrin Rotaxanes and Catenanes
259
some insights: (a) rotaxanes and catenanes behave similarly provided the reaction parameters are kept constant; (b) the photochemical properties of the systems depend dramatically on the presence or absence of a metal center in the dpp-based coordination site, providing both a structural and a functional role; (c) the nature of the central metal is important in determining the deactivation paths of the primary excited states, whereas Cu(phen)2 + acts as both energy acceptor and electron donor, Ag(phen)2 + acts as a real or virtual electron acceptor; (d) including the chromophore in the rings, rather than appending it, increases the rigidity thus improving the performances. The present compounds could be regarded as models of the highly complex multicomponent assemblies found in natural photosynthesis. In addition, they could be used in the future as molecular elements of complex logic gates within the futuristic field of molecular electronics. Acknowledgements Financial support from Italian CNR (Project PM-P04-ISTN-C1-ISOFM5), MIUR (FIRB, RBNE019H9K), Cost Action D31/0003/04, and the French CNRS are gratefully acknowledged.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Deisenhofer J, Epp O, Miki K, Huber R, Michel H (1984) J Mol Biol 180:385 Deisenhofer J, Epp O, Miki K, Huber R, Michel H (1985) Nature 318:618 Boxer SG, Bucks RR (1979) J Am Chem Soc 101:1883 Nagata T, Osuka A, Maruyama K (1990) J Am Chem Soc 112:3054 Harriman A, Magda DJ, Sessler JL (1991) J Chem Soc Chem Commun 345 Gust D, Moore TA, Moore AL, Mkings LR, Sely GR, Ma X, Trier TT, Gao F (1988) J Am Chem Soc 110:7567 Johson DG, Niemczyk MP, Minsek DW, Wiedereecht GP, Svec WA, Gaines III GL, Wasielewski MR (1993) J Am Chem Soc 115:5692 Helms A, Heiler D, McLendon G (1992) J Am Chem Soc 114:6227 Martin JL, Breton J, Hoff AJ, Migus A, Anonetti A (1986) Proc Natl Acad Sci 83:957 Dressler K, Umlauf E, Schmidt S, Hamm P, Zinth W, Buchanan S, Michel H (1991) Chem Phys Lett 183:270 Brun AM, Harriman A, Heitz V, Sauvage JP (1991) J Am Chem Soc 113:8657 Boxer SG, Goldstein RA, Lokhart DJ, Middendor TR, Tkiff L (1989) J Phys Chem 93:8280 Heitz V, Chardon-Noblat S, Sauvage JP (1991) Tetrahedron Lett 32:197 Arlt T, Schmidt S, Kaiser W, Lauterwasser C, Meyer M, Scheer H, Zinth W (1993) Proc Natl Acad Sci 90:11757 Franzen S, Goldstein RF, Boxer SG (1993) J Phys Chem 97:3040 Gunther MJ, Johnston MR (1992) J Chem Soc Chem Commun 1163 Le Bras F, Loock B, Momenteau M (1994) Tetrahedron Lett 35:3289 Chambron JC, Heitz V, Sauvage JP (1995) Bull Soc Chim Fr 132:340 Ashton PR, Johnston MR, Stoddart JF, Tolley MS, Wheeler JW (1992) J Chem Soc Chem Commun 1128 Chambron JC, Heitz V, Sauvage JP (1992) J Chem Soc Chem Commun 1131 Vögtle F, Ahuis S, Baumann S, Sessler JL (1996) Liebigs Ann 921 Chichak K, Walsh MC, Branda NR (2000) Chem Commun 847
260
L. Flamigni et al.
23. Schill G (1971) Catenanes, Rotaxanes and Knots. Academic Press, New York 24. Sauvage JP, Dietrich-Buchecker CO (eds) (1999) Molecular Catenanes, Rotaxanes and Knots. Wiley, Weinheim 25. Balzani V, Venturi M, Credi A (2003) Molecular Devices and Machines: a Journey into the Nanoworld. Wiley, Weinheim 26. Sauvage J-P (2001) Molecular Machines and Motors, Structure & Bonding. Springer, Berlin Heidelberg New York 27. Balzani V, Credi A, Raymo FM, Stoddart FJ (2000) Angew Chem Int Ed 39:3348 28. Special issue on Molecular Machines (2001) Acc Chem Res 34:341 29. Dietrich-Buchecker CO, Sauvage JP (1987) Chem Rev 87:795 30. Dietrich-Buchecker CO, Sauvage JP (1991) Bioorganic Chemistry Frontiers, vol 2. Springer, Berlin Heidelberg New York 31. Sauvage JP (ed) (1993) Special Issue New J Chem 17 32. Sauvage JP, Dietrich-Buchecker CO, Chambron JC (1996) In: Lehn JM (ed) Comprehensive Supramolecular Chemistry. Pergamon Press, Oxford 33. Belohradsky M, Raymo FM, Stoddart JF (1996) Collect Czech Chem Commun 61:1 34. Amabilino DB, Stoddart JF (1995) Chem Rev 95:2725 35. Amabilino DB, Stoddart JF (1993) Pure Appl Chem 65:2351 36. Philp D, Stoddart JF (1996) Angew Chem Int Ed Engl 35:1154 37. Sauvage JP (1990) Acc Chem Rev 23:219 38. Arnaud-Neu F, Marques E, Schwing-Weill MJ, Dietrich-Buchecker C, Sauvage JP, Weiss J (1988) New J Chem 12:15 39. Dietrich-Buchecker CO, Sauvage JP (1983) Tetrahedron Lett 24:5091 40. Dietrich-Buchecker CO, Sauvage JP, Kintzinger JP (1983) Tetrahedron Lett 24:5095 41. Dietrich-Buchecker CO, Sauvage JP (1990) Tetrahedron Lett 46:503 42. Antipas A, Dolphin D, Gouterman M, Johnson EC (1978) J Am Chem Soc 100:7705 43. Jamin ME, Iwamoto RT (1978) Inorg Chim Acta 27:135 44. Abou-Gamra Z, Harriman A, Neta P (1986) J Chem Soc Faraday Trans 2 82:2337 45. Chambron JC, Heitz V, Sauvage JP (1993) J Am Chem Soc 115:12378 46. Ichinaga AK, Kirchhoff JR, McMillin Dr, Dietrich-Buchecker CO, Marnot PA, Sauvage JP (1987) Inorg Chem 26:4290 47. Rothemund PJ (1936) J Am Chem Soc 58:625 48. Rothemund PJ (1939) J Am Chem Soc 61:2912 49. Adler A, Longo FR, Finarelli JD, Goldmacher J, Assour J, Korsakoff L (1967) J Org Chem 32:476 50. Lindsey JS, Hsu HC, Schreiman IC (1986) Tetrahedron Lett 27:4969 51. Newman MS, Lee LF (1972) J Org Chem 37:4468 52. Harriman A, Heitz V, Ebersole M, van Willigen H (1994) J Phys Chem 98:4982 53. Harriman A, Heitz V, Sauvage JP (1993) J Phys Chem 97:5940 54. Chambron JC, Harriman A, Heitz V, Sauvage JP (1993) J Am Chem Soc 115:6109 55. Chambron JC, Harriman A, Heitz V, Sauvage JP (1993) J Am Chem Soc 115:7419 56. Chambron JC, Dietrich-Buchecker CO, Nierengarten JF, Sauvage JP (1993) J Chem Soc Chem Commun 801 57. Chambron JC, Dietrich-Buchecker CO, Nierengarten JF, Sauvage JP, Solladié N, Albrecht-Gary AM, Meyer M (1995) New J Chem 19:409 58. Marshall JA, Flynn GA (1979) Synth Commun 9:123 59. Dietrich-Buchecker CO, Marnot PA, Sauvage JP (1982) Tetrahedron Lett 23:5291 60. Dietrich-Buchecker CO, Sauvage JP, Kintzinger JP, Maltèse P, Pascard C, Guilhem J (1992) New J Chem 16:931 61. West R, Rochow EG (1953) J Org Chem 18:1739
Porphyrin Rotaxanes and Catenanes 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96.
261
McDermott JX, White JF, Whitesides G (1976) J Am Chem Soc 98:6529 Dietrich-Buchecker CO, Sauvage JP (1983) Tetrahedron Lett 24:5091 Dietrich-Buchecker CO, Sauvage JP (1990) Tetrahedron Lett 46:503 Young R, Chang CK (1985) J Am Chem Soc 107:898 Solladié N, Chambron JC, Sauvage JP (1999) J Am Chem Soc 121:3684 Solladié N, Chambron JC, Sauvage JP (1996) Angew Chem Int Ed Engl 35:906 Dietrich-Buchecker CO, Sauvage JP, Kern JM (1984) J Am Chem Soc 106:3043 Flamigni L, Armaroli N, Barigelletti F, Chambron J-C, Sauvage J-P, Solladié N (1999) New J Chem 23:1151 Kirchhoff J, Gamache RE, Blaskie MW, del Paggio AA, Lengel RK, McMillin DR (1983) Inorg Chem 22:2380 Linke M, Chambron JC, Heitz V, Sauvage JP (1997) J Am Chem Soc 119:11329 Anderson M, Linke M, Chambron JC, Davidsson J, Heitz V, Hammarström L, Sauvage JP (2002) J Am Chem Soc 124:4347 Amabilino DB, Sauvage JP (1996) Chem Commun 2441 Amabilino DB, Sauvage JP (1998) New J Chem 22:395 Anderson M, Linke M, Chambron JC, Davidsson J, Heitz V, Hammarström L (2000) J Am Chem Soc 122:3526 Dietrich-Buchecker C, Sauvage J-P, Kern J-M (1989) J Am Chem Soc 111:7791 Linke M, Chambron JC, Heitz V, Sauvage JP, Semetey V (1998) Chem Commun 2469 Linke M, Chambron JC, Heitz V, Sauvage JP, Ensinas S, Barigelletti F, Flamigni L (2000) J Am Chem Soc 122:11834 Linke M, Fujita N, Chambron JC, Heitz V, Sauvage JP (2001) New J Chem 25:790 Linke M, Chambron JC, Heitz V, Sauvage JP (1999) Chem Commun 2419 Jörgensen T, Becher J, Chambron JC, Sauvage JP (1994) Tetrahedron Lett 35:4339 Raehm L, Kern JM, Sauvage JP (1999) Chem Eur J 5:3310 Flamigni L, Talarico AM, Chambron J-C, Heitz V, Linke M, Fujita N, Sauvage J-P (2004) Chem Eur J 10:2689 Watanabe N, Kihara N, Furusho Y, Takata T, Ataki Y, Ito O (2003) Ang Chem Int Ed 42:681 Sandanayaka ASD, Watanabe N, Ikeshita K-I, Yasuyuki A, Kihara N, Furusho Y, Ito O, Takata T (2005) J Phys Chem B 109:2516 Flamigni L, Talarico AM, Serroni P; Puntoriero F, Gunter JM, Johnston MR, Jeynes TP (2003) Chem Eur J 9:2649 Li K, Schuster DI, Guldi DM, Herranz MA, Echegoyen L (2004) J Am Chem Soc 126:3388 Li K, Bracher PJ, Guldi DM, Herranz MA, Echegoyen L, Schuster DI (2004) J Am Chem Soc 126:9156 Schuster DI, Li K, Guldi DM, Ramey J (2004) Org Lett 6:1919 Guldi DM, Prato M (2000) Acc Chem Res 33:695 Andreasson J, Kodis G, Lin S, Moore AL, Moore TA, Gust D, Martensson J, Albinsson B (2002) Photochem Photobiol 76:47 Guldi DM (2002) Chem Soc Rev 31:22 Ou ZP, Kadish KM, Wenbo E, Shao JG, Sintic PJ, Ohkubo K, Fukuzumi S, Crossley MJ (2004) Inorg Chem 43:2078 Marcus RA (1993) Ang Chem Int Ed Engl 32:1111 Armaroli N, Marconi G, Echegoyen L, Bourgeois J-P, Diederich F (2000) Chem Eur J 6:1629 Flamigni L, Talarico AM, Ventura B, Rein R, Solladié N (2006) Chem Eur J 12:701
Struct Bond (2006) 121: 263–295 DOI 10.1007/430_020 © Springer-Verlag Berlin Heidelberg 2005 Published online: 22 December 2005
Multiporphyrin Arrays Assembled Through Hydrogen Bonding Maxwell J Gunter Chemistry, School of Biological, Biomedical and Molecular Sciences, University of New England, NSW 2351 Armidale, Australia [email protected] 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
264
2 2.1 2.2
Carboxylic Acid Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . Homocomposite Carboxylic Acid Systems . . . . . . . . . . . . . . . . . . . Carboxylate-Amide Base Systems . . . . . . . . . . . . . . . . . . . . . . .
266 266 271
3 3.1 3.2
Heterocyclic Base Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . Homocomposite Heterocyclic Base Pairing . . . . . . . . . . . . . . . . . . Heterocomposite Nucleobase and Heterocyclic Base Pairing . . . . . . . . .
274 274 276
4
DNA-Porphyrin Conjugates . . . . . . . . . . . . . . . . . . . . . . . . . .
278
5
Cyanuric Acid-Melamine/Barbiturate Systems . . . . . . . . . . . . . . . .
282
6
Self-Organisation Through Selection . . . . . . . . . . . . . . . . . . . . .
285
7
Catenanes and Rotaxanes . . . . . . . . . . . . . . . . . . . . . . . . . . . .
287
8
Polymeric Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
287
9
Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . .
293
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
293
Abstract Although relatively weak in isolation, composite H-bonds can be used as an advantage for the assembly of relatively robust and well-defined arrays of molecular components. Porphyrins, with their inherent symmetry, synthetic accessibility and functionality offer ideal base units for the assembly of multicomponent systems by the reversible yet strong intermolecular forces of H-bonding. The geometric precision and strong directionality of H-bonds between relatively rigid donor and acceptor groups can be incorporated into the architecture of porphyrin supramolecular arrays to produce some remarkably complex high-definition assemblies through simple mixing of the component parts. Nevertheless, the very reversibility which allows for ease of construction becomes problematic in ensuring integrity of structure in solution; temperature, solvent and concentration become increasingly important. Measurement techniques are limited to those which can discriminate between the various possible combinations of the component parts to produce discrete oligomeric or polymeric entities. In many cases the measured properties of the assemblies underlying the reason for their construction in the first place also corroborate their structural integrity. Although H-bonding has been
264
M.J. Gunter
used to construct heterotopic porphyrinic arrays comprising porphyrins integrated with other molecular entities, this review is constrained to those systems which principally result in multiporphyrin arrays, although often templated by non-porphyrinic component parts. Keywords Multiporphyrin arrays · Hydrogen bonding · Self-assembly · Supramolecular · Porphyrin
1 Introduction If one were to plan design principles for new and unnatural self-assembled systems, then it would be prudent to select from nature those motifs which have had the advantage of millions of years of evolution for the construction of biological systems with exquisite control of shape and function. It is remarkable how predominantly hydrogen bonding is featured in the organisation for the specific functional and structural roles that typify such complex systems as cells, viruses and other higher organisms. It is not only the ultimate shape and function that is determined by these interactions, but of equal importance is their role in molecular self-assembly that ensures, through thermodynamic control, that the information required for the selforganisation and assembly is built-in. Reversibility allows re-iterative error correction during the construction phase, leading exclusively to the desired outcome with absolute efficiency and control. Even the use of enzymes or templates for the construction of higher order systems from fundamental components relies on intermolecular interactions that must necessarily be directional, defined and specific. The topologies which are critical for the structure and function of proteins, or the base pairing in DNA and RNA are obvious examples of the critical role of hydrogen bonding in natural systems. Hydrogen bonding is ideally suited as an inter- and intramolecular organisational tool for assembling complex structures. It is relatively strong, it is directional, and it can exploit multipoint interactions in multicomponent arrays and systems. By the use of multiple H-bonding interactions, which individually are relatively weak, the total stabilisation energies associated with any particular system can be amplified considerably by cooperative interactions of many complementary hydrogen bonds. Thus, it is not surprising that hydrogen bonding has featured prominently in many of the design principles used in the construction of new nanoscale materials and nanotechnological devices through supramolecular assembly processes; the principles of hydrogen bonding in supramolecular chemistry have been the subject of several recent reviews [1–10]. The strongly directional motif of hydrogen bonding combined with relatively rigid building blocks can lead to well-defined and predicted structures
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
265
whose integrity can be ensured in solution, with a caveat that a non H-bonding solvent is utilised. Being under active thermodynamic control is an advantage for the assembly of the arrays, yet this can be a disadvantage in that the structures are often temperature, solvent and concentration dependent. The integrity of the structure can be compromised during temperature-dependent measurements (e.g. dynamic NMR), and careful choice of the solvent is needed to avoid complete or partial collapse of the assembly, or formation of a different structure incorporating the solvent. Polymeric structures are favoured at higher concentrations, while monomeric or partially assembled species will predominate at lower concentrations. Any design of supramolecular systems incorporating H-bonding needs to cater to the fact that each H-bond has a donor (D) and acceptor (A) unit. This can be used as an advantage for strengthening intermolecular interactions through careful design. For example, the classical Watson–Crick base pairing of guanine–cytosine (G –– C) involves an AAD sequence of cytosine, which is complementary to the DDA sequence of guanine mutually oriented at 120◦ ; triaminotriazine has a DAD sequence, and this is complementary to the ADA sequence of barbituric acid. If subunits such as these are incorporated into separate monomer units, six-membered H-bonded rings with complementary sequences can be formed under the correct circumstances, leading to particularly stable heterotopic complexes. Often the structures are maintained from solution to the solid state, and many of the H-bonded systems characterised to date are from solid-state structures obtained from X-ray crystallography [1, 11, 12]. It is more problematic to determine whether the solid state structure is maintained in solution, and a raft of less direct but nevertheless reliable methods can be utilised to define the solution-phase structures [12]. In the solid state, there is often a wide variety of other reversible and weaker interactions (e.g. dipolar, van der Waals, π – π interactions) that supplement the main H-bonding forces that are the major factors in determining the structure [2, 13–18]. Porphyrins, with their rigid core structure, are ideal scaffolds for assembling supramolecular and multicomponent linked systems. There have been several more recent reviews of multiporphyrinic supramolecular systems assembled through non-covalent forces including coordination, H-bonding, electrostatics, surface attachment and aggregation [7, 19–21]. Generally, these arrays can be categorised as being closed, discrete, open, or polymeric. Multifunctionalised porphyrins with a high degree of symmetry are easily accessible synthetically, and when armed with complementary H-bonding moieties, they can be used to construct multicomponent H-bonded systems through self-assembly. For ditopic monomer units with more than one complementary H-bonding group attached to the same porphyrin, topologically-controlled multiporphyrin arrays can then be assembled by judicious design of the substitution pattern and geometry of the building block.
266
M.J. Gunter
2 Carboxylic Acid Systems 2.1 Homocomposite Carboxylic Acid Systems Porphyrin carboxylic acids will form classical carboxylate dimers in lesspolar solvents. For example the monocarboxylic acid porphyrin 1 dimerizes with an association constant of 1.1 × 103 M–1 in CH2 Cl2 [22]. For multicarboxylic acid porphyrins, the dimerization can be enhanced if the carboxylate groups are appropriately aligned. Kuroda and co-workers [23, 24] have utilized H-bonding to assemble a system which has a central single freebase porphyrin surrounded by eight zinc porphyrins as a model for core and antenna units in a light harvesting photosynthetic mimic. The tetrapyrazinyl porphyrin 2a was synthesized by condensation of mesotetrakis(4-aminophenyl)porphyrin with 5-(2-pyrazinyl)pentanoic acid. The complementary tetracarboxylic acid porphyrin [meso-tetrakis(2-carboxy-
Structure 1
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
267
4-nonylphenyl)porphyrinato]zinc(II) 3 is dimeric in solution. Thus, selfassembly through both hydrogen bonding and Zn-pyrazine coordination results in a nonameric array 4, with the pyrazinyl porphyrin as a central energy acceptor and the four pairs of carboxylate zinc porphyrin dimers as antennas. Titration data can be analysed for a 4 : 1 complex through four independent binding processes with identical binding constants K = (4 ± 1) × 107 M–1 , the large value reflecting the complementarity of the H-bonding re-inforced by the bidentate axial ligation to the Zn porphyrins. The results suggest that over 92% of 3 in the solution of 1 : 4 mixture of these porphyrins is bound to 2a under the experimental conditions. The center-to-center distance between the freebase and zinc porphyrins in extended conformations of the alkyl chains supporting the pyrazinyl moieties is estimated to lie in the
Structure 2
268
M.J. Gunter
A with only weak electronic interaction between porphyrins in range 16–21 ˚ the resulting assembly. The fluorescence properties of the 1 : 4 mixed system under excitation at 564 nm were compared to the emission intensities and quantum yields of the individual components when irradiated at the same wavelength. The resulting emission spectrum of the 1 : 4 mixture can be fitted to 18 times the emission from the freebase porphyrin 2a plus 0.18 times the emission of the Zn-dimer/pyrazine complex 3·pyrazine. Since only 18% of the emission of 3·pyrazine complex contributes to the final spectrum, this implies an 82% efficiency of the energy transfer from the 3·pyrazine complex to the freebase porphyrin 2a in the assembly. Although the observed efficiency of the energy transfer is low compared to other covalently linked multiporphyrin systems, the 18 times enhanced fluorescence of 2a in the nonamer is demonstration of the antenna effect of the array, which enhances the light absorption efficiency of the system rather than energy transfer. Furthermore, since the energy transfer efficiency is approximately independent of the type of linkers for the pyrazine chains (ether 2b or amide 2a), it appears that the energy transfer is true Förster type. Subsequently, this same group [25] have extended the size of these types of systems to a heptadecameric array 5. The same zinc carboxylate dimer
Structure 3
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
Structure 4
Structure 5
269
270
M.J. Gunter
is used as the antenna components, but the central acceptor porphyrin unit 2c, 2d or 2e is now armed with eight pyrazines; the result is the extended 17-component light harvesting antenna and acceptor systems 5 and 6. The pyrazinyl porphyrins side chains are designed to give different topological arrangements of the antennas, so that the pyrazines are either in “parallel” as in 2c and 2d, or in “series” as in 2e within each chain, leading to the heptadecameric arrays 5 and 6, respectively. This leads to a 77 times enhancement of the fluorescence of the freebase core porphyrin in the largest “parallel” heptadecamer 5, and clearly vindicates the design principles of these systems as true antenna molecules: the more antennas, the greater the energy transfer. Another interesting outcome from these systems is the topological dependence of the energy transfer: 6, which has a serial arrangement of the antenna
Structure 6
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
271
pigments, has an unexpectedly low energy transfer efficiency, even lower than that for the smaller nonameric systems 4. This may indicate (a) the linker chain rigidified by attachment of two highly bulky antenna pigments in the single strand of 6 fixes the position of the antenna pigments further from the central porphyrin, or (b) the energy transfer between inner and outer antenna pigments, as is observed in the natural photosynthetic system, operates in 6 and this interferes with the energy transfer from the antenna to the central porphyrin. An acyclic dimer of a dendritic zinc porphyrin bearing six carboxylic acid functionalities 7 interacts with fullerenes, such as C60 and C70, to form “supramolecular peapods”, composed of a hydrogen-bonded zinc porphyrin nanotube and fullerenes 8 [26]. H-bonding between the R3 carboxylates produces a bis-fullerene “hamburger” type unit 9, which then forms extended H-bonding interactions in the second dimension through R1 and R2 to form elongated peapod-like assemblies 8, with the fullerenes arranged pairwise along its interior. The outside of the peapod is encased in the dendritic units, which comprise R4 of the monomer. TEM images show that the peapods are very long (> 1 µm) and have a uniform diameter of 15 nm. Without fullerenes, the zinc porphyrin dimer forms only a heavily entangled, irregular assembly. In contrast to 7, an ester version of the acyclic dimer without hydrogen-bonding capability shows little interaction with fullerenes, indicating the crucial role of H-bonding in stabilising the assembly. 2.2 Carboxylate-Amide Base Systems Otsuki et al. [27] have demonstrated that amidinium-carboxylate salt bridges, which have been used earlier to construct electron donor-acceptor dyads or a donor-spacer-acceptor triad, can also be used to assemble energy donoracceptor dyad 13 and pentad 14. The salt bridge consists of complementary double hydrogen bonds and electrostatic interactions and, therefore, offers
Structure 7
272
M.J. Gunter
Structure 8
the advantages of well-defined directionality and strength. The Zn-porphyrin complex with an amidine group 10 as the energy donor was combined with the freebase porphyrins bearing either one 11, or four 12 carboxy groups to assemble the 1 : 1 donor-acceptor pair 13, and the antenna type 4 : 1 assembly 14, respectively. The steady state and time-resolved fluorescence measurements unequivocally showed that efficient singlet–singlet excited state energy transfer from the Zn-porphyrin complex to the freebase porphyrin takes place in these assemblies. Indeed, the observed energy transfer rates in both types of assemblies are much faster than those which the Förster mechanism would suggest. This infers that a through-bond mechanism is in operation in the excited state energy transfer in these Zn/freebase porphyrin assemblies through
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
273
Structure 9
the amidinium-carboxylate salt bridges. Thus, it is suggested that hydrogen bonds, augmented by electrostatic attraction, can mediate through-bond energy transfer. These seem to be the first reports of through-bond energy transfer involving intermolecular bonds which invoke faster rates than estimated from a Förster mechanism, but the true mechanism of the fast singlet–singlet intra-ensemble excited state energy transfer remains to be deciphered. The twisted conformation inevitably adopted by meso–meso linked porphyrins has provided the basis for defined helical structures from dimeric and trimeric bis-porphyrins 15 appropriately substituted with carboxylic acid functionalities, in the presence of an ancillary cyclic urea 16 [28]. The helical porphyrin arrays 17 that result in solution from complementary H-bonding exhibit chirality amplification and enhanced two-photon absorption properties.
274
M.J. Gunter
3 Heterocyclic Base Systems 3.1 Homocomposite Heterocyclic Base Pairing Pyrazolyl and substituted pyrazolyl porphyrins have been shown to form linear dimers 18 and cyclic tetramers 19 in CDCl3 solution [29]. The aggregation is destroyed in protic solvents such as methanol, and dilution experiments were utilized to estimate the possible structure of the association. The data were fitted to a mixture of dimer and tetramer, and the presence of other cyclic n-mers was discounted on the basis of the curve fitting. The dimerization constant, K2 = 39 M–1 obtained from the curve fitting of 18, was lower than the tetramerization constant K4 = 9.3 × 103 M–3 of 19. Although it has been demonstrated unequivocally that the discrete structures discussed above are maintained in solution under the conditions described, thermodynamic reversibility will inevitably result in mixtures of condition-dependent n-mers. This is particularly evident as the concentration is varied within a given series, and the same set of porphyrins can be monomeric, open oligo- or polymeric, or cyclic oligomeric. Under a given set of conditions, it is expected that at least several entities will be present at
Structure 10
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
275
any time, although a major species can be selected for by adjustment of the conditions. Drain [13] has utilized the self-association of 3,5-diacetamido-4-pyridyl substituted porphyrins as a design stratagem for the formation of a variety of H-bonded oligomeric tapes and squares. A monosubstituted derivative 20 can form the simplest homodimer 21 through a quadruple H-bond between diacetamidopyridine units in the meso (5-) position of each porphyrin; a linear trimeric tape 22 predominates in a 2 : 1 stoichiometric mixture of
Structure 11
276
M.J. Gunter
Structure 12
the 5- and 5,15-disubstituted derivatives 20 and 23; and a linear tetramer 24 from a 2 : 2 stoichiometric ratio. NMR evidence confirms the predominance of these species in solution, but the measured association constants indicate relatively weak binding in each case. Nevertheless, it should be stressed again that the NMR results do not imply 100% of each of these species in all cases, but a predominance of the tape structures in each case as a result of the stoichiometries and ∆G. In fact fits of the binding isotherms indicate that although the simple dimer 21 contains ca. 2 units (K = 160 dm3 mol–1 ), the trimer 22 indicates ca. 3.2 units (K = 110 dm6 mol–2 ) in the 2 : 1 mixture, and the tetramer 24 ca. 4.1 units (K = 70 dm9 mol–3 ) in the 2 : 2 mixture (note that direct comparisons of K are difficult because of the different powers of the concentration units). However the 5,10-disubstituted derivative 25, where the diacetamidopyridine units are located on adjacent meso-positions of the porphyrin and hence 90◦ disposed from each other, forms a tetrameric closed square 26. Here higher values of K (2400 dm9 mol–3 ) indicate an enhanced cooperativity in the formation of the closed square with a more favourable ∆G; at the mM NMR concentrations used, the square tetramer is favoured over an open chain tetrameric tape with “unbonded” ends by ca. 8 kJ mol–1 . 3.2 Heterocomposite Nucleobase and Heterocyclic Base Pairing The use of the classical guanine–cytosine (G–C) Watson–Crick base pairing to form heterocomposite porphyrin systems has been established for some time, through the early work of the Sessler [30–32], Hamilton [33, 34] and
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
277
Structure 13
Lehn [35, 36] groups, amongst others. From the initial systems which formed well-characterized heterodimers, there have evolved increasingly complex and sophisticated systems with well-defined architectures and topologies. These have departed from the classical nucleobase pairing (Watson–Crick and Hoogsteen) to arrays based on pairing between a variety of different nucleobases [37], urea/carboxylic acids, barbiturate/diaminopyridine, uracil/diacetamidopyridine, and triaminotriazines/barbiturates. Since 2,6-diacetamidopyridines also form complementary H-bonding interactions with uracil groups, a combination of porphyrin monomer units 25 described above with 90◦ disposed diacetamido units, and a second porphyrin unit with uracil groups at equivalent 90◦ 5,10-positions 27 formed the square heterotetramer 28 [13]. Fits of NMR chemical shift data relative to concentration indicate complementary interactions between the uracyl and diacetamidopyridyl moieties in tetrahydrofuran-d8 which are much stronger than either self-complementary interaction between the monomer units in CDCl3 even in a solvent known to be less favorable to H-bonding than chloroform. These results were also compared to the thermodynamics of forming the open array, with the complementary groups on opposite (5,15-) meso positions of each of two different porphyrins (180◦ disposed). This allowed an assessment of the cooperativity advantage for forming the square array vs an open polymer.
278
M.J. Gunter
These systems have no doubt evolved conceptually from the earlier Lehn bis-porphyrin triaminotriazine motif [36], which spontaneously formed the beautiful 3-D cyclic hexaporphyrin structure 31 from three units of the bisporphyrin 29 and three of dialkylbarbituric acid 30. The structural integrity of these systems incorporating 18 N – H – N and N – H – O hydrogen bonds was implied from solution (UV-vis and NMR, ESI-MS, VPO and DOSY 2-D NMR) studies.
4 DNA-Porphyrin Conjugates DNA provides an ideal supramolecular scaffold for assembling porphyrins and other functional units in well-defined structures which are capable
Structure 14
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
279
of self-assembly given an appropriate connectivity to a specific nucleobase. Such a strategy also allows for chirality to be built into the array through the ribose moiety. Thus Endo and Majima [38, 39] have synthesised a tetraphenylporphyrin-based DNA-porphyrin conjugate 32 with linkages between each meso-phenyl group and the central phosphorus atom of a 10-mer strand of DNA (the spacing of the 10 bases in a 10-mer corresponds to one helical turn, so that the linkers end up on the same side of a subsequent duplex). By the use of a thiolate tether through a phosphoramidite linkage 33, and reaction with a maleimide-substituted porphyrin 34, two synthetic oligonucleotides with sequences CGGCTpACTCC and GTGCTpAGCGG 32 were assembled, where p in the sequence denotes the position of the phosphoramidite at which the porphyrin is linked. The DNA-porphyrin strands 32 were then mixed with 4 equivalents of a complementary 20-mer non-porphyrinic strand, resulting in a doublehelical complex with two DNA-porphyrin strands linked to four complementary strands 35. The stoichiometry was confirmed by absorption spectroscopy and non-denaturing polyacrylamide gel electrophoresis.
Structure 15
280
M.J. Gunter
To probe the photophysical properties of the double-stranded complexes, the corresponding zinc derivatives were also made. The assembly technique allows the production of freebase and zinc derivatives within the one com-
Structure 16
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
281
plex. Fluorescent lifetimes of the systems indicated a singlet energy transfer between the freebase and zinc porphyrins within the one complex. These results prove that it is possible to incorporate multiple and different porphyrin chromophores into the DNA structures by appropriate sequence programming of the DNA strands. Accordingly, Stulz has outlined the synthesis of porphyrin-substituted uridine and deoxyuridine derivatives 36a and 36b [40], and the dimerized analogues 37a and 37b [41], which contain either different or the same porphyrin subunits. The substitution patterns on the porphyrins were selected to infer different solubility properties of the conjugates, which can either
Structure 17
282
M.J. Gunter
be organic-solvent soluble (ester substituents) or water soluble (carboxylate substituents). For the dimers, there is electronic interaction between the porphyrin subunits as indicated by electronic spectroscopy, but these are strongly dependent on the nature of the porphyrin; meso-tetraphenyl derivatives show little interaction while the 5,15-diphenyl substituted derivatives are strongly coupled. These dinucleotides could be prepared either by solution or solid-phase synthetic routes, and other nucleotide bases (e.g. diadenosine) could also be incorporated. More recently, this group1 has incorporated these and related porphyrin-substituted nucleosides into tetranucleotides using standard solid support techniques to assemble chiral homo- and heteroporphyrinic arrays 38a–c. In this way, the composition and the physical properties of the array can easily be made to order simply by programming the DNA synthesis on an automated DNA synthesizer. The tetranucleotides were then able to form duplexes with the complementary strands (e.g. tetra-adenosine dA4 39). It was shown that the porphyrin electronic ground state environment is not particularly influenced by the presence or absence of the complementary strand, but the fluorescence intensity is decreased. CD measurements confirm the helical nature of the duplexes.
5 Cyanuric Acid-Melamine/Barbiturate Systems The self-assembly behaviour and complexing properties of several strapped porphyrin-incorporated melamine-cyanuric or melamine-barbiturate-based rosette supramolecules in chloroform-d has produced some intriguing structures. The almost perfect H-bonding complementarity of the constituent units, amplified by their multiplicity, results in particularly stable structures whose integrity is maintained in solution [42]. Strapped porphyrin cyanuric acid 40a and its Zn (II) complex 40b were designed to combine with melamine derivatives 41 to afford stable porphyrin rosettes 42. The new porphyrin rosettes 42 could efficiently complex tripyridyl derivative 43 through intermolecular, cooperative coordination between Zn (II) and pyridine to form the stable rosette 44. New pyridine-bearing barbiturates 45a, which form the stable rosettes 45 were also synthesized. Mixing equimolar amounts of 45a with 41 in chloroform-d led to the formation of new isomeric rosettes 45 resulting from different orientation of the pyridine unit in the rosettes. It was also established that porphyrin-bearing rosette 42 could complex pyridine-bearing rosette 45, leading to the formation of new double-layer type supramolecular architectures 46, in which one rosette assembles the three zinc porphyrin 1
Bouamaied I, Fendt L-A, Stulz E (2005) private communication
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
Structure 18
Structure 19
283
284
Structure 20
Structure 21
M.J. Gunter
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
285
units by covalent bonds, while the second rosette is coordinated to each of the zinc atoms.
6 Self-Organisation Through Selection Although careful design and synthesis can give rise to stable and well-defined arrays that self-assemble by H-bonding and other reversible processes, ultimate flexibility and efficiency can ensue through the use of dynamic combinatorial libraries (DCLs). Although the general applicability of DCLs is in relative infancy, there have been significant advances in recent years using several different approaches to achieve thermodynamic reversibility of the active library components. The high thermodynamic stability of multiple H-bonded assemblies together with their usually high kinetic lability makes them entirely suitable for the generation of DCLs in which different combinations of receptors equilibrate through reversible formation of multiple H-bonds. In the presence of a suitable template, the most effective receptor is selected and amplified from an equilibrating mixture of building blocks. This has been illustrated in a variety of systems, but the use of H-bonding as the reversible assembly process can only be effective where the final array is particularly stable [7]. Reinhoudt and co-workers [43] have used this concept effectively for the construction of stable porphyrinic arrays based on the melamine-cyanuric acid motif. For example they reported the non-covalent synthesis of a family of H-bonded (non-porphyrinic) assemblies consisting of nine different components held together by 36 cooperative H-bonds [44]. For the construction of porphyrin-based arrays, a dimelamine-appended calixarene with two covalently attached zinc porphyrin units 47 is used as the basic unit. When this is mixed with 2 equivalents of the cyanuric acid 48 in the presence of the tridentate ligand 49, it leads to clean self-assembly of the symmetrical homomeric array 50 containing six identical porphyrin units, three on the top and three on the bottom of the assembly. Structural diversity can be built into the system by simply mixing a variety of components. When the non-porphyrinic component 51 is also introduced into the mixture at room temperature, a statistical (1 : 3 : 3 : 1) mixture of the double rosette assemblies is formed immediately. In the presence of a tripyridine 49 of suitable dimensions as a template, clean amplification from all the possible combinations resulted in a 1 : 1 mixture of the two homomeric species, the most stable receptor 50 and the symmetrical non-porphyrinic analogue 52. The structures of the homomeric species was established from solution NMR studies, utilising the inherent symmetry of the final product in comparison to the other possible structures. Subsequently [43, 45] the limits of the concept were tested by a remarkable assembly of nanostructures (3.0 × 5.5 nm dimensions and molecular weight
286
M.J. Gunter
Structure 22
20 kDa), comprising 8 different rosette layers that are held together by 144 cooperative H-bonds!
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
287
7 Catenanes and Rotaxanes Although there have been several approaches to porphyrinic catenanes and rotaxanes [46–48], these have been based on assembly through charged [49] or neutral [50, 51] π-donor/acceptor concepts, or by using coordination chemistry [47]. Nevertheless, there is a series of non-porphyrinic amidebased catenanes and rotaxanes that have received considerable attention [52, 53]. The primary recognition factor in these systems is H-bonding. Gunter [54] has extended these systems to include porphyrinic supramolecules that use similar H-bonded amide-based motifs for the self-assembly of multiporphyrin catenanes and rotaxanes. Thus the synthesis of a porphyrin strapped by an isophthalamide moeity under conditions deliberately chosen to favour H-bonding between two of the strap units resulted in good yield of the corresponding bis-porphyrin [2]catenane 53 as well as the expected singly-strapped derivative 54. Unlike many of the other H-bonded assemblies discussed in this review, these catenanes are chemically robust, and cannot disassemble under conditions that disfavour the H-bonding that was the templating factor responsible for their assembly in the first place. The dynamic aspects of these catenanes were studied by NMR methods, and were found to be affected by solvent and temperature, factors which clearly affect the internal H-bonding. The singly-strapped porphyrin side-product 54 from the catenane synthesis still retains the isophthalamide unit, and so can potentially form relatively stable H-bonded supramolecular complexes with different openchain isophthalamide-containing species. These range from inclusion complexes with simple isophthalamides, to psuedo-rotaxanes with longer-chain derivatives. Indeed, when the pseudo-rotaxane thread unit is terminated with a pyridyl ligand as in 55, then a (pseudo)rotaxane with large metalloporphyrinic stopper units can be assembled under thermodynamic (reversible) conditions [51]. For example, the rotaxane 56 forms on simply mixing the components of the assembly: the isophthalamide-strapped porphyrin 54, the pyridine-terminated isophthalamide thread unit 55, and two equivalents of a bulky ruthenium carbonyl porphyrin 57. Again, the dynamics of this system could be controlled by variations in temperature and solvent, and even chloride ion which acts as a competitive binder in the isophthalamide moiety of the strapped porphyrin.
8 Polymeric Systems Without specific geometrically constrained donor or acceptor moieties, porphyrin monomer units with less-defined H-bonding substituents tend to-
288
Structure 23
M.J. Gunter
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
289
wards polymeric arrays, but designed arrangement of the substitution pattern can give rise to ordered polymers. Thus a porphyrin with imidazoles at opposite meso positions has been shown to assemble through H-bonding into an offset-stacked co-facial polymeric structure 58 [55]. This motif is maintained in non-protic solvents such as CDCl3 , but the polymeric structure is progressively destroyed by the addition of CD3 OD, which competes for the internal H-bonding in the polymer. Solvent-dependent NMR and UV-vis spectroscopic evidence is consistent with a face-to-face H-bonded structure, but the fluorescence behaviour appears to be independent of solvent, and hence indicates no significant self-quenching in the supramolecular array. Nevertheless, there is more efficient quenching of external acceptors by energy-electron transfer for the assembled polymer than the monomeric unit. This is attributable to a delocalised excitation energy in the polymer so that energy transfer can occur at any of the individual components of the polymer chain; in effect the polymer behaves as a crude antenna complex. If the long alkyl chains in the monomer unit are replaced by carboxylateterminated chains 59, the amphiphilic nature of the resultant molecule leads to well-defined homotopic liposomal dispersions 60 in water, without any added lipid components such as lecithin [56]. Gel filtration was used to separate the fraction corresponding to unilamellar vesicles, and dynamic light scattering measurements indicated a liposome diameter of 27 ± 8 nm, which is consistent with AFM images which showed a diameter of about 26 nm. A TEM image of the dispersion also showed a predominance of particles in the 20–30 nm range. The cavity formed by the liposome 60 is large enough to accommodate polar solutes such as pyranine (a fluorescence probe) inside, which is slowly released to the exterior aqueous phase. The formation of these liposomes in an aqueous phase is due to a network of H-bonds involving the imadazolyl moieties assembled around the hydrophobic porphyrin cores, which are themselves assembled in a central belt by π – π interactions. The low permeability of the entrapped pyranine guest across the liposomal membrane suggests a strong H-bonding network in the interior. These liposomes obtained from a single amphiphilic unit are regarded as large light-
Structure 24
290
M.J. Gunter
Structure 25
harvesting antenna units, and future experiments will be directed towards this end. Systems incorporating sugar units are an inevitable choice for the design of H-bonded systems. Aggregates and polymers based on sugar-appended porphyrins have been studied for their ability to form gels and fibres from solution [57]. Thus amphiphilic porphyrins 61 bearing β-D-galactopyranose groups at the periphery, attached to the p-positions of meso-tetraaryl porphyrins were shown to aggregate in a unidimensional columnar stack 62, resulting in robust gels in DMF/alcohol mixed solvents. In this case, there is a synergism between the H-bonding provided by the sugar units and π – π stacking of the porphyrins. The resultant thermotropic and lyotropic gels were characterized by SEM and TEM techniques. In a different strategy for designing porphyrin-based organogelators [58], hydrogen bond-donating (carboxylic acid)/accepting (pyridine) substituents or electron-donating (dialkylamino ester)/withdrawing (pyridine) substituents were introduced into peripheral positions of a porphyrin (63a or 63b, respectively), and the gelation properties were compared with those of symmetrical reference compounds bearing two pyridyl substituents or two ester groups (65 or 64b, respectively). The solubility properties of the
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
291
Structure 26
series of molecules is critically dependent on whether or not they have a dipole moment; the symmetrical molecules 64a and 64b are much less soluble in organic solvents. Looking at 63a, 63b, and 64b, they all formed gels with cyclohexane, methylcyclohexane, and several alcoholic solvents, but scanning electron microscopy (SEM) and transmission electron microscopy (TEM) observations revealed quite different superstructures in the organogels. In cyclohexane, 63a resulted in a sheet-like structure, whereas 63b and 64b resulted in a fiber-like structure. The difference is attributed to the two-dimensional interactive forces in 63a consisting of porphyrin– porphyrin π – π stacking and carboxylic acid-pyridine hydrogen bonding. When the H-bonding interaction was weakened by alcoholic solvents or by adding pyridine or N,N-(dimethylamino)-pyridine, the sheet-like structure was transformed to the fiber-like structure. Further detailed analyses of their aggregation modes were conducted by spectroscopic methods such as ultraviolet-visible (UV-vis) absorption, Fourier transform infrared (FT-IR), and X-ray diffraction (XRD). On the basis of these findings, the influence of the peripheral substituents on gel formation and the aggregation mode was rationalized in terms of different 1-D, 2-D and 3-D structural motifs, in-
292
Structure 27
Structure 28
M.J. Gunter
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding
293
volving hydrophobic interactions between the hexyloxyphenyl chains in one dimension, offset pyridine-carboxylate H-bonding interactions in a second dimension, and π-stacking of the porphyrin units in the third as indicated in the diagram (Structure 28).
9 Conclusions and Perspectives The next era of supramolecular chemistry will undoubtedly be one of adaptability and evolution. As the complexity and the sophistication of these systems increases the need for viable alternatives to covalent synthesis is essential. One such is offered by the exploitation of weak intermolecular forces for the construction of arrays and aggregates with defined composition, shape and functionality. Hydrogen bonding provides and ideal motif for such systems, and together with the well-established reactivity patterns and shape adaptability of porphyrins, the way is clear to the development of new systems with increasing chemical diversity and controlled dynamics. It is clear that the most rapid advances will be made through the use of dynamic combinatorial libraries, which allows for the chemical evolution of mixtures and amplification of specific components, with built-in error checking and control. Nevertheless, the creation of new forms of complex matter will need to wait for a full understanding of the driving principles at a fundamental level, but with ever-increasing sophistication. From the selected examples given in this short review, it is apparent that the self-assembly of multiporphyrin homotopic and heterotopic arrays through H-bonding is clearly a reality. However, as their complexity increases, so too does the degree of difficulty in their characterisation, particularly so for non-solid state assemblies. Being dynamic is a two-edged sword: essential for self-assembly, but problematic for definition. To maintain their credibility, the increasing sophistication and sensitivity of measurement techniques must keep pace with the evolution of these new materials.
References 1. 2. 3. 4. 5. 6. 7. 8.
Desiraju GR (2002) Acc Chem Research 35:565 Steiner T (2002) Angew Chem Int Ed Engl 41:48 Schmuck C, Wienand W (2001) Angew Chem Int Ed Engl 40:4363 Archer EA, Gong HG, Krische MJ (2001) Tetrahedron 57:1139 Subramanian S, Zaworotko MJ (1994) Coord Chem Rev 137:357 Aakeroy CB, Seddon KR (1993) Chem Soc Rev 22:397 Prins LJ, Reinhoudt DN, Timmerman P (2001) Angew Chem Int Ed Engl 40:2382 Turner DR, Pastor A, Alajarin M, Steed JW (2004) In: Michael D, Mingos P (eds) Structure and Bonding, vol 108. Springer, Berlin Heidelberg New York, p 97
294
M.J. Gunter
9. Vilar R (2004) In: Michael D, Mingos P (eds) Structure and Bonding, vol 111. Springer, Berlin Heidelberg New York, p 85 10. Hardie MJ (2004) In: Michael D, Mingos P (eds) Structure and Bonding, vol 111. Springer, Berlin Heidelberg New York, p 139 11. Burrows AD (2004) In: Michael D, Mingos P (eds) Structure and Bonding, vol 108. Springer, Berlin Heidelberg New York, p 55 12. Aliev AE, Harris KDM (2004) In: Michael D, Mingos P (eds) Structure and Bonding, vol 108. Springer, Berlin Heidelberg New York, p 1 13. Shi X, Barkigia KM, Fajer J, Drain CM (2001) J Org Chem 66:6513 14. Balaban Teodor S, Eichhöfer A, Lehn J-M (2000) Eur J Org Chem 24:4047 15. Bhyrappa P, Wilson SR, Suslick KS (1997) J Am Chem Soc 119:8492 16. Suslick KS, Bhyrappa P, Chou JH, Kosal ME, Nakagaki S, Smithenry DW, Wilson SR (2005) Acc Chem Research 38:283 17. Kosal ME, Suslick KS (2000) J Solid State Chem 152:87 18. Suslick KS, Kosal ME, Bhyrappa P (1998) Abstr Pap Am Chem Soc 216:290 19. Satake A, Kobuke Y (2005) Tetrahedron 61:13 20. Ogawa K, Kobuke Y (2004) In: Nalwa HS (ed) Encyclopedia of Nanoscience and Nanotechnology, vol 9. American Scientific Publishers, Stevenson Ranch, California, p 561 21. Chen X, Drain CM (2004) In: Nalwa HS (ed) Encyclopedia of Nanoscience and Nanotechnology, vol 9. American Scientific Publishers, Stevenson Ranch, California, p 593 22. Turró C, Chang CK, Leroi GE, Cukier RI, Nocera DG (1992) J Am Chem Soc 114:4013 23. Kuroda Y, Sugou K, Sasaki K (2000) J Am Chem Soc 122:7833 24. Kuroda Y, Kawashima A, Hayashi Y, Ogoshi H (1997) J Am Chem Soc 119:4929 25. Sugou K, Sasaki K, Kitajima K, Iwaki T, Kuroda Y (2002) J Am Chem Soc 124:1182 26. Yamaguchi T, Ishii N, Tashiro K, Aida T (2003) J Am Chem Soc 125:13934 27. Otsuki J, Iwasaki K, Nakano Y, Itou M, Araki Y, Ito O (2004) Chem Eur J 10:3461 28. Ikeda C, Yoon ZS, Park M, Inoue H, Kim D, Osuka A (2005) J Am Chem Soc 127:534 29. Ikeda C, Nagahara N, Motegi E, Yoshioka N, Inoue H (1999) Chem Commun 17:1759 30. Harriman A, Magda DJ, Sessler JL (1991) J Phys Chem 95:1530 31. Harriman A, Magda DJ, Sessler JL (1991) Chem Commun :345 32. Sessler JL, Wang B, Harriman A (1995) J Am Chem Soc 117:704 33. Slobodkin G, Fan E, Hamilton AD (1992) New J Chem 16:643 34. Hamilton AD (1995) Tetrahedron 51(2):343 35. Drain CM, Fischer R, Nolen EG, Lehn J-M (1993) J Chem Soc, Chem Commun :243 36. Drain CM, Russell KC, Lehn J-M (1996) Chem Commun 3:337 37. Sessler JL, Jayawickramarajah J (2005) Chem Commun 15:1939 38. Endo M, Seeman NC, Majima T (2005) Angew Chem Int Ed 44:6074 39. Endo M, Shiroyama T, Fujitsuka M, Majima T (2005) J Org Chem 70:7468 40. Bouamaied I, Stulz E (2005) Synlett 9:1579 41. Bouamaied I, Stulz E (2005) Chimia 59:101 42. Shao X-B, Jiang X-K, Zhu S-Z, Li Z-T (2004) Tetrahedron 60:9155 43. Calama MC, Timmerman P, Reinhoudt DN (2000) Angew Chem Int Ed 39:755 44. Calama MC, Timmerman P, Reinhoudt DN, Hulst R, Timmerman P, Fokkens R, Nibbering NMM (1998) Chem Commun 9:1021 45. Paraschiv V, Crego-Calama M, Fokkens RH, Padberg CJ, Timmermann P, Reinhoudt DN (2001) J Org Chem 66:8297 46. Ogoshi H, Mizutani T, Hayashi T, Kuroda Y (2000) In: Kadish KM, Smith KM, Guilard R (eds) The Porphyrin Handbook, vol 6. Academic Press, San Diego, p 279 47. Chambron J-C, Heitz V, Sauvage J-P (2000) In: Kadish KM, Smith KM, Guilard R (eds) The Porphyrin Handbook, vol 6. Academic Press, San Diego, p 1
Multiporphyrin Arrays Assembled Through Hydrogen-Bonding 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.
295
Gunter MJ (2004) Eur J Org Chem :1655 Gunter MJ, Jeynes TP, Turner P (2004) Eur J Org Chem :193 Gunter MJ, Farquhar SM (2003) Org Biomol Chem 1:3450 Gunter MJ, Bampos N, Johnstone KD, Sanders JKM (2001) New J Chem 25:166 Schalley CA, Beizai K, Vogtle F (2001) Acc Chem Research 34:465 Breault GA, Hunter CA, Mayers PC (1999) Tetrahedron 55:5265 Gunter MJ, Farquhar SM, Mullen KM (2004) New J Chem 28:1443 Nagata N, Kugimiya S-I, Kobuke Y (2000) Chem Commun 15:1389 Nagata N, Kugimiya S-I, Kobuke Y (2001) Chem Commun 8:689 Tamaru S-I, Nakamura M, Takeuchi M, Shinkai S (2001) Org Lett 3:3631 Tanaka S, Shirakawa M, Kaneko K, Takeuchi M, Shinkai S (2005) Langmuir 21:2163
Author Index Volumes 101–121 Author Index Vols. 1–100 see Vol. 100
The volume numbers are printed in italics Alajarin M, see Turner DR (2004) 108: 97–168 Aldinger F, see Seifert HJ (2002) 101: 1–58 Alessio E, see Iengo E (2006) 121: 105–143 Alfredsson M, see Corà F (2004) 113: 171–232 Aliev AE, Harris KDM (2004) Probing Hydrogen Bonding in Solids Using State NMR Spectroscopy 108: 1–54 Alloul H, see Brouet V (2004) 109: 165–199 Amstutz N, see Hauser A (2003) 106: 81–96 Anitha S, Rao KSJ (2003) The Complexity of Aluminium-DNA Interactions: Relevance to Alzheimer’s and Other Neurological Diseases 104: 79–98 Anthon C, Bendix J, Schäffer CE (2004) Elucidation of Ligand-Field Theory. Reformulation and Revival by Density Functional Theory 107: 207–302 Aramburu JA, see Moreno M (2003) 106: 127–152 Arˇcon D, Blinc R (2004) The Jahn-Teller Effect and Fullerene Ferromagnets 109: 231–276 Atanasov M, Daul CA, Rauzy C (2003) A DFT Based Ligand Field Theory 106: 97–125 Atanasov M, see Reinen D (2004) 107: 159–178 Atwood DA, see Conley B (2003) 104: 181–193 Atwood DA, Hutchison AR, Zhang Y (2003) Compounds Containing Five-Coordinate Group 13 Elements 105: 167–201 Atwood DA, Zaman MK (2006) Mercury Removal from Water 120: 163–182 Autschbach J (2004) The Calculation of NMR Parameters in Transition Metal Complexes 112: 1–48 Baerends EJ, see Rosa A (2004) 112: 49–116 Bard AJ, Ding Z, Myung N (2005) Electrochemistry and Electrogenerated Chemiluminescence of Semiconductor Nanocrystals in Solutions and in Films 118: 1–57 Barriuso MT, see Moreno M (2003) 106: 127–152 Beaulac R, see Nolet MC (2004) 107: 145–158 Bebout DC, Berry SM (2006) Probing Mercury Complex Speciation with Multinuclear NMR 120: 81–105 Bellandi F, see Contreras RR (2003) 106: 71–79 Bendix J, see Anthon C (2004) 107: 207–302 Berend K, van der Voet GB, de Wolff FA (2003) Acute Aluminium Intoxication 104: 1–58 Berry SM, see Bebout DC (2006) 120: 81–105 Bianconi A, Saini NL (2005) Nanoscale Lattice Fluctuations in Cuprates and Manganites 114: 287–330 Blinc R, see Arcˇcon D (2004) 109: 231–276
298
Author Index Volumes 101–121
Boˇca R (2005) Magnetic Parameters and Magnetic Functions in Mononuclear Complexes Beyond the Spin-Hamiltonian Formalism 117: 1–268 Bohrer D, see Schetinger MRC (2003) 104: 99–138 Bouamaied I, Coskun T, Stulz E (2006) Axial Coordination to Metalloporphyrins Leading to Multinuclear Assemblies 121: 1–47 Boulanger AM, see Nolet MC (2004) 107: 145–158 Boulon G (2004) Optical Transitions of Trivalent Neodymium and Chromium Centres in LiNbO3 Crystal Host Material 107: 1–25 Bowlby BE, Di Bartolo B (2003) Spectroscopy of Trivalent Praseodymium in Barium Yttrium Fluoride 106: 193–208 Braga D, Maini L, Polito M, Grepioni F (2004) Hydrogen Bonding Interactions Between Ions: A Powerful Tool in Molecular Crystal Engineering 111: 1–32 Brouet V, Alloul H, Gàràj S, Forrò L (2004) NMR Studies of Insulating, Metallic, and Superconducting Fullerides: Importance of Correlations and Jahn-Teller Distortions 109: 165– 199 Buddhudu S, see Morita M (2004) 107: 115–144 Budzelaar PHM, Talarico G (2003) Insertion and β-Hydrogen Transfer at Aluminium 105: 141–165 Burrows AD (2004) Crystal Engineering Using Multiple Hydrogen Bonds 108: 55–96 Bussmann-Holder A, Keller H, Müller KA (2005) Evidences for Polaron Formation in Cuprates 114: 367–386 Bussmann-Holder A, see Micnas R (2005) 114: 13–69 Canadell E, see Sánchez-Portal D (2004) 113: 103–170 Cancines P, see Contreras RR (2003) 106: 71–79 Cartwright HM (2004) An Introduction to Evolutionary Computation and Evolutionary Algorithms 110: 1–32 Christie RA, Jordan KD (2005) n-Body Decomposition Approach to the Calculation of Interaction Energies of Water Clusters 116: 27–41 Clot E, Eisenstein O (2004) Agostic Interactions from a Computational Perspective: One Name, Many Interpretations 113: 1–36 Conley B, Atwood DA (2003) Fluoroaluminate Chemistry 104: 181–193 Contreras RR, Suárez T, Reyes M, Bellandi F, Cancines P, Moreno J, Shahgholi M, Di Bilio AJ, Gray HB, Fontal B (2003) Electronic Structures and Reduction Potentials of Cu(II) Complexes of [N,N -Alkyl-bis(ethyl-2-amino-1-cyclopentenecarbothioate)] (Alkyl = Ethyl, Propyl, and Butyl) 106: 71–79 Cooke Andrews J (2006) Mercury Speciation in the Environment Using X-ray Absorption Spectroscopy 120: 1–35 Corà F, Alfredsson M, Mallia G, Middlemiss DS, Mackrodt WC, Dovesi R, Orlando R (2004) The Performance of Hybrid Density Functionals in Solid State Chemistry 113: 171–232 Coskun T, see Bouamaied I (2006) 121: 1–47 Crespi VH, see Gunnarson O (2005) 114: 71–101 Daul CA, see Atanasov M (2003) 106: 97–125 Day P (2003) Whereof Man Cannot Speak: Some Scientific Vocabulary of Michael Faraday and Klixbüll Jørgensen 106: 7–18 Deeth RJ (2004) Computational Bioinorganic Chemistry 113: 37–69 Delahaye S, see Hauser A (2003) 106: 81–96 Deng S, Simon A, Köhler J (2005) Pairing Mechanisms Viewed from Physics and Chemistry 114: 103–141
Author Index Volumes 101–121
299
Di Bartolo B, see Bowlby BE (2003) 106: 191–208 Di Bilio AJ, see Contreras RR (2003) 106: 71–79 Ding Z, see Bard AJ (2005) 118: 1–57 Dovesi R, see Corà F (2004) 113: 171–232 Duan X, see He J (2005) 119: 89–119 Duan X, see Li F (2005) 119: 193–223 Egami T (2005) Electron-Phonon Coupling in High-Tc Superconductors 114: 267–286 Eisenstein O, see Clot E (2004) 113: 1–36 Ercolani G (2006) Thermodynamics of Metal-Mediated Assemblies of Porphyrins 121: 167– 215 Evans DG, see He J (2005) 119: 89–119 Evans DG, Slade RCT (2005) Structural Aspects of Layered Double Hydroxides 119: 1–87 Ewing GE (2005) H2 O on NaCl: From Single Molecule, to Clusters, to Monolayer, to Thin Film, to Deliquescence 116: 1–25 Flamigni L, Heitz V, Sauvage J-P (2006) Porphyrin Rotaxanes and Catenanes: Copper(I)Templated Synthesis and Photoinduced Processes 121: 217–261 Fontal B, see Contreras RR (2003) 106: 71–79 Forrò L, see Brouet V (2004) 109: 165–199 Fowler PW, see Soncini A (2005) 115: 57–79 Frenking G, see Lein M (2003) 106: 181–191 Frühauf S, see Roewer G (2002) 101: 59–136 Frunzke J, see Lein M (2003) 106: 181–191 Furrer A (2005) Neutron Scattering Investigations of Charge Inhomogeneities and the Pseudogap State in High-Temperature Superconductors 114: 171–204 Gàràj S, see Brouet V (2004) 109: 165–199 Gillet VJ (2004) Applications of Evolutionary Computation in Drug Design 110: 133–152 Golden MS, Pichler T, Rudolf P (2004) Charge Transfer and Bonding in Endohedral Fullerenes from High-Energy Spectroscopy 109: 201–229 Gorelesky SI, Lever ABP (2004) 107: 77–114 Grant GJ (2006) Mercury(II) Complexes with Thiacrowns and Related Macrocyclic Ligands 120: 107–141 Gray HB, see Contreras RR (2003) 106: 71–79 Grepioni F, see Braga D (2004) 111: 1–32 Gritsenko O, see Rosa A (2004) 112: 49–116 Güdel HU, see Wenger OS (2003) 106: 59–70 Gunnarsson O, Han JE, Koch E, Crespi VH (2005) Superconductivity in Alkali-Doped Fullerides 114: 71–101 Gunter MJ (2006) Multiporphyrin Arrays Assembled Through Hydrogen Bonding 121: 263– 295 Gütlich P, van Koningsbruggen PJ, Renz F (2004) Recent Advances in Spin Crossover Research 107: 27–76 Guyot-Sionnest P (2005) Intraband Spectroscopy and Semiconductor Nanocrystals 118: 59–77 Habershon S, see Harris KDM (2004) 110: 55–94 Han JE, see Gunnarson O (2005) 114: 71–101
300
Author Index Volumes 101–121
Hardie MJ (2004) Hydrogen Bonded Network Structures Constructed from Molecular Hosts 111: 139–174 Harris KDM, see Aliev (2004) 108: 1–54 Harris KDM, Johnston RL, Habershon S (2004) Application of Evolutionary Computation in Structure Determination from Diffraction Data 110: 55–94 Hartke B (2004) Application of Evolutionary Algorithms to Global Cluster Geometry Optimization 110: 33–53 Harvey JN (2004) DFT Computation of Relative Spin-State Energetics of Transition Metal Compounds 112: 151–183 Haubner R, Wilhelm M, Weissenbacher R, Lux B (2002) Boron Nitrides – Properties, Synthesis and Applications 102: 1–46 Hauser A, Amstutz N, Delahaye S, Sadki A, Schenker S, Sieber R, Zerara M (2003) Fine Tuning the Electronic Properties of [M(bpy)3 ]2+ Complexes by Chemical Pressure (M = Fe2+ , Ru2+ , Co2+ , bpy = 2,2 -Bipyridine) 106: 81–96 He J, Wei M, Li B, Kang Y, G Evans D, Duan X (2005) Preparation of Layered Double Hydroxides 119: 89–119 Heitz V, see Flamigni L (2006) 121: 217–261 Herrmann M, see Petzow G (2002) 102: 47–166 Herzog U, see Roewer G (2002) 101: 59–136 Hoggard PE (2003) Angular Overlap Model Parameters 106: 37–57 Höpfl H (2002) Structure and Bonding in Boron Containing Macrocycles and Cages 103: 1–56 Hubberstey P, Suksangpanya U (2004) Hydrogen-Bonded Supramolecular Chain and Sheet Formation by Coordinated Guranidine Derivatives 111: 33–83 Hupp JT (2006) Rhenium-Linked Multiporphyrin Assemblies: Synthesis and Properties 121: 145–165 Hutchison AR, see Atwood DA (2003) 105: 167–201 Iengo E, Scandola F, Alessio E (2006) Metal-Mediated Multi-Porphyrin Discrete Assemblies and Their Photoinduced Properties 121: 105–143 Iwasa Y, see Margadonna S (2004) 109: 127–164 Jansen M, Jäschke B, Jäschke T (2002) Amorphous Multinary Ceramics in the Si-B-N-C System 101: 137–192 Jäschke B, see Jansen M (2002) 101: 137–192 Jäschke T, see Jansen M (2002) 101: 137–192 Jaworska M, Macyk W, Stasicka Z (2003) Structure, Spectroscopy and Photochemistry of the [M(η5 -C5 H5 )(CO)2 ]2 Complexes (M = Fe, Ru) 106: 153–172 Jenneskens LW, see Soncini A (2005) 115: 57–79 Jeziorski B, see Szalewicz K (2005) 116: 43–117 Johnston RL, see Harris KDM (2004) 110: 55–94 Jordan KD, see Christie RA (2005) 116: 27–41 Kabanov VV, see Mihailovic D (2005) 114: 331–365 Kang Y, see He J (2005) 119: 89–119 Keller H (2005) Unconventional Isotope Effects in Cuprate Superconductors 114: 143–169 Keller H, see Bussmann-Holder A (2005) 114: 367–386 Khan AI, see Williams GR (2005) 119: 161–192 Kobuke Y (2006) Porphyrin Supramolecules by Self-Complementary Coordination 121: 49–104
Author Index Volumes 101–121
301
Koch E, see Gunnarson O (2005) 114: 71–101 Kochelaev BI, Teitel’baum GB (2005) Nanoscale Properties of Superconducting Cuprates Probed by the Electron Paramagnetic Resonance 114: 205–266 Köhler J, see Deng (2005) 114: 103–141 van Koningsbruggen, see Gütlich P (2004) 107: 27–76 Lein M, Frunzke J, Frenking G (2003) Christian Klixbüll Jørgensen and the Nature of the Chemical Bond in HArF 106: 181–191 Leroux F, see Taviot-Gueho C (2005) 119: 121–159 Lever ABP, Gorelesky SI (2004) Ruthenium Complexes of Non-Innocent Ligands; Aspects of Charge Transfer Spectroscopy 107: 77–114 Li B, see He J (2005) 119: 89–119 Li F, Duan X (2005) Applications of Layered Double Hydroxides 119: 193–223 Liebau F, see Santamaría-Pérez D (2005) 118: 79–135 Linton DJ, Wheatley AEH (2003) The Synthesis and Structural Properties of Aluminium Oxide, Hydroxide and Organooxide Compounds 105: 67–139 Lux B, see Haubner R (2002) 102: 1–46 Mackrodt WC, see Corà F (2004) 113: 171–232 Macyk W, see Jaworska M (2003) 106: 153–172 Mahalakshmi L, Stalke D (2002) The R2M+ Group 13 Organometallic Fragment Chelated by P-centered Ligands 103: 85–116 Maini L, see Braga D (2004) 111: 1–32 Mallia G, see Corà F (2004) 113: 171–232 Margadonna S, Iwasa Y, Takenobu T, Prassides K (2004) Structural and Electronic Properties of Selected Fulleride Salts 109: 127–164 Maseras F, see Ujaque G (2004) 112: 117–149 Micnas R, Robaszkiewicz S, Bussmann-Holder A (2005) Two-Component Scenarios for Non-Conventional (Exotic) Superconstructors 114: 13–69 Middlemiss DS, see Corà F (2004) 113: 171–232 Mihailovic D, Kabanov VV (2005) Dynamic Inhomogeneity, Pairing and Superconductivity in Cuprates 114: 331–365 Millot C (2005) Molecular Dynamics Simulations and Intermolecular Forces 115: 125–148 Miyake T, see Saito (2004) 109: 41–57 Moreno J, see Contreras RR (2003) 106: 71–79 Moreno M, Aramburu JA, Barriuso MT (2003) Electronic Properties and Bonding in Transition Metal Complexes: Influence of Pressure 106: 127–152 Morita M, Buddhudu S, Rau D, Murakami S (2004) Photoluminescence and Excitation Energy Transfer of Rare Earth Ions in Nanoporous Xerogel and Sol-Gel SiO2 Glasses 107: 115–143 Morsch VM, see Schetinger MRC (2003) 104: 99–138 Mossin S, Weihe H (2003) Average One-Center Two-Electron Exchange Integrals and Exchange Interactions 106: 173–180 Murakami S, see Morita M (2004) 107: 115–144 Müller E, see Roewer G (2002) 101: 59–136 Müller KA (2005) Essential Heterogeneities in Hole-Doped Cuprate Superconductors 114: 1–11 Müller KA, see Bussmann-Holder A (2005) 114: 367–386 Myung N, see Bard AJ (2005) 118: 1–57
302
Author Index Volumes 101–121
Nishibori E, see Takata M (2004) 109: 59–84 Nolet MC, Beaulac R, Boulanger AM, Reber C (2004) Allowed and Forbidden d-d Bands in Octa-hedral Coordination Compounds: Intensity Borrowing and Interference Dips in Absorption Spectra 107: 145–158 O’Hare D, see Williams GR (2005) 119: 161–192 Ordejón P, see Sánchez-Portal D (2004) 113: 103–170 Orlando R, see Corà F (2004) 113: 171–232 Oshiro S (2003) A New Effect of Aluminium on Iron Metabolism in Mammalian Cells 104: 59–78 Pastor A, see Turner DR (2004) 108: 97–168 Patkowski K, see Szalewicz K (2005) 116: 43–117 Patoˇcka J, see Strunecká A (2003) 104: 139–180 Peng X, Thessing J (2005) Controlled Synthesis of High Quality Semiconductor Nanocrystals 118: 137–177 Petzow G, Hermann M (2002) Silicon Nitride Ceramics 102: 47–166 Pichler T, see Golden MS (2004) 109: 201–229 Polito M, see Braga D (2004) 111: 1–32 Popelier PLA (2005) Quantum Chemical Topology: on Bonds and Potentials 115: 1–56 Power P (2002) Multiple Bonding Between Heavier Group 13 Elements 103: 57–84 Prassides K, see Margadonna S (2004) 109: 127–164 Prato M, see Tagmatarchis N (2004) 109: 1–39 Price LS, see Price SSL (2005) 115: 81–123 Price SSL, Price LS (2005) Modelling Intermolecular Forces for Organic Crystal Structure Prediction 115: 81–123 Rabinovich D (2006) Poly(mercaptoimidazolyl)borate Complexes of Cadmium and Mercury 120: 143–162 Rao KSJ, see Anitha S (2003) 104: 79–98 Rau D, see Morita M (2004) 107: 115–144 Rauzy C, see Atanasov (2003) 106: 97–125 Reber C, see Nolet MC (2004) 107: 145–158 Reinen D, Atanasov M (2004) The Angular Overlap Model and Vibronic Coupling in Treating s-p and d-s Mixing – a DFT Study 107: 159–178 Reisfeld R (2003) Rare Earth Ions: Their Spectroscopy of Cryptates and Related Complexes in Glasses 106: 209–237 Renz F, see Gütlich P (2004) 107: 27–76 Reyes M, see Contreras RR (2003) 106: 71–79 Ricciardi G, see Rosa A (2004) 112: 49–116 Riesen H (2004) Progress in Hole-Burning Spectroscopy of Coordination Compounds 107: 179–205 Robaszkiewicz S, see Micnas R (2005) 114: 13–69 Roewer G, Herzog U, Trommer K, Müller E, Frühauf S (2002) Silicon Carbide – A Survey of Synthetic Approaches, Properties and Applications 101: 59–136 Rosa A, Ricciardi G, Gritsenko O, Baerends EJ (2004) Excitation Energies of Metal Complexes with Time-dependent Density Functional Theory 112: 49–116 Rudolf P, see Golden MS (2004) 109: 201–229 Ruiz E (2004) Theoretical Study of the Exchange Coupling in Large Polynuclear Transition Metal Complexes Using DFT Methods 113: 71–102
Author Index Volumes 101–121
303
Sadki A, see Hauser A (2003) 106: 81–96 Saini NL, see Bianconi A (2005) 114: 287–330 Saito S, Umemoto K, Miyake T (2004) Electronic Structure and Energetics of Fullerites, Fullerides, and Fullerene Polymers 109: 41–57 Sakata M, see Takata M (2004) 109: 59–84 Sánchez-Portal D, Ordejón P, Canadell E (2004) Computing the Properties of Materials from First Principles with SIESTA 113: 103–170 Santamaría-Pérez D, Vegas A, Liebau F (2005) The Zintl–Klemm Concept Applied to Cations in Oxides II. The Structures of Silicates 118: 79–135 Sauvage J-P, see Flamigni L (2006) 121: 217–261 Scandola F, see Iengo E (2006) 121: 105–143 Schäffer CE (2003) Axel Christian Klixbüll Jørgensen (1931–2001) 106: 1–5 Schäffer CE, see Anthon C (2004) 107: 207–301 Schenker S, see Hauser A (2003) 106: 81–96 Schetinger MRC, Morsch VM, Bohrer D (2003) Aluminium: Interaction with Nucleotides and Nucleotidases and Analytical Aspects of Determination 104: 99–138 Schmidtke HH (2003) The Variation of Slater-Condon Parameters Fk and Racah Parameters B and C with Chemical Bonding in Transition Group Complexes 106: 19–35 Schubert DM (2003) Borates in Industrial Use 105: 1–40 Schulz S (2002) Synthesis, Structure and Reactivity of Group 13/15 Compounds Containing the Heavier Elements of Group 15, Sb and Bi 103: 117–166 Seifert HJ, Aldinger F (2002) Phase Equilibria in the Si-B-C-N System 101: 1–58 Shahgholi M, see Contreras RR (2003) 106: 71–79 Shinohara H, see Takata M (2004) 109: 59–84 Sieber R, see Hauser A (2003) 106: 81–96 Simon A, see Deng (2005) 114: 103–141 Slade RCT, see Evans DG (2005) 119: 1–87 Soncini A, Fowler PW, Jenneskens LW (2005) Angular Momentum and Spectral Decomposition of Ring Currents: Aromaticity and the Annulene Model 115: 57–79 Stalke D, see Mahalakshmi L (2002) 103: 85–116 Stasicka Z, see Jaworska M (2003) 106: 153–172 Steed JW, see Turner DR (2004) 108: 97–168 Strunecká A, Patoˇcka J (2003) Aluminofluoride Complexes in the Etiology of Alzheimer’s Disease 104: 139–180 Stulz E, see Bouamaied I (2006) 121: 1–47 Suárez T, see Contreras RR (2003) 106: 71–79 Suksangpanya U, see Hubberstey (2004) 111: 33–83 Sundqvist B (2004) Polymeric Fullerene Phases Formed Under Pressure 109: 85–126 Szalewicz K, Patkowski K, Jeziorski B (2005) Intermolecular Interactions via Perturbation Theory: From Diatoms to Biomolecules 116: 43–117 Tagmatarchis N, Prato M (2004) Organofullerene Materials 109: 1–39 Takata M, Nishibori E, Sakata M, Shinohara M (2004) Charge Density Level Structures of Endohedral Metallofullerenes by MEM/Rietveld Method 109: 59–84 Takenobu T, see Margadonna S (2004) 109: 127–164 Talarico G, see Budzelaar PHM (2003) 105: 141–165 Taviot-Gueho C, Leroux F (2005) In situ Polymerization and Intercalation of Polymers in Layered Double Hydroxides 119: 121–159 Teitel’baum GB, see Kochelaev BI (2005) 114: 205–266 Thessing J, see Peng X (2005) 118: 137–177
304
Author Index Volumes 101–121
Trommer K, see Roewer G (2002) 101: 59–136 Tsuzuki S (2005) Interactions with Aromatic Rings 115: 149–193 Turner DR, Pastor A, Alajarin M, Steed JW (2004) Molecular Containers: Design Approaches and Applications 108: 97–168 Uhl W (2003) Aluminium and Gallium Hydrazides 105: 41–66 Ujaque G, Maseras F (2004) Applications of Hybrid DFT/Molecular Mechanics to Homogeneous Catalysis 112: 117–149 Umemoto K, see Saito S (2004) 109: 41–57 Unger R (2004) The Genetic Algorithm Approach to Protein Structure Prediction 110: 153–175 van der Voet GB, see Berend K (2003) 104: 1–58 Vegas A, see Santamaría-Pérez D (2005) 118: 79–135 Vilar R (2004) Hydrogen-Bonding Templated Assemblies 111: 85–137 Wei M, see He J (2005) 119: 89–119 Weihe H, see Mossin S (2003) 106: 173–180 Weissenbacher R, see Haubner R (2002) 102: 1–46 Wenger OS, Güdel HU (2003) Influence of Crystal Field Parameters on Near-Infrared to Visible Photon Upconversion in Ti2+ and Ni2+ Doped Halide Lattices 106: 59–70 Wheatley AEH, see Linton DJ (2003) 105: 67–139 Wilhelm M, see Haubner R (2002) 102: 1–46 Williams GR, Khan AI, O’Hare D (2005) Mechanistic and Kinetic Studies of Guest Ion Intercalation into Layered Double Hydroxides Using Time-resolved, In-situ X-ray Powder Diffraction 119: 161–192 de Wolff FA, see Berend K (2003) 104: 1–58 Woodley SM (2004) Prediction of Crystal Structures Using Evolutionary Algorithms and Related Techniques 110: 95–132 Xantheas SS (2005) Interaction Potentials for Water from Accurate Cluster Calculations 116: 119–148 Zaman MK, see Atwood DA (2006) 120: 163–182 Zhang H (2006) Photochemical Redox Reactions of Mercury 120: 37–79 Zerara M, see Hauser A (2003) 106: 81–96 Zhang Y, see Atwood DA (2003) 105: 167–201
Subject Index
Acyclic assemblies, thermodynamics 170 Amidinium-carboxylate salt bridges 272 Antenna 49 Antenna mimics, B850 64 Arrays, one-dimensional 76 –, two-dimensional 82 Assemblies, acyclic polymeric 174 –, cyclic, thermodynamics 182 –, metal-mediated 105 –, multicyclic, thermodynamics 200 –, multi-porphyrin 32, 105 –, ruthenium-mediated 118 –, topologically reducible multicyclic 204 Axial coordination compounds 3 – – –, geometry 7 B850 64 Barbiturate/diaminopyridine 277 Base pairing, homocomposite heterocyclic 274 – –, nucleobase 276 Bismetalloporphyrin 174 Bis-porphyrin, (PZn/PAu+ ) dumbbell 221 – –, bis-nitrogen ligands 12 Box 64 Carboxylate-amide base systems Carboxylic acid systems 266 Catalysis 161 Catenanes 287 –, pendant porphyrins, copper(I)-templated synthesis –, porphyrin-containing 246 –, PZn/PAu+ units 246 Charge separation 49, 217 Chelate effect 182 Chlorophylls 53 Cooperativity 173
271
253
Coordination organization 51 Cyanuric acid-melamine/ barbiturate 282 Cyclic assemblies, thermodynamics
182
DABCO 151 2,6-Diacetamidopyridines 277 Di-tert-butyl-3,5-benzaldehyde 223 Dimers, planar 151 DNA/RNA, hydrogen bonding 264 DNA-porphyrin conjugates 278 Dynamic combinatorial libraries (DCLs) 1, 38, 285 Electron transfer 162 – –, photoinduced, special pair mimic 89 Energy conversion, light-to-electrical 163 Energy transfer 217 Etioporphyrin 222 Fullerenes 40, 271 Gold nanoparticles, functionalized 180 Gold(III) porphyrins 217 Guanine–cytosine 276 Heterocyclic base systems/base pairing 274, 276 Homocomposite heterocyclic base pairing 274 Hydrogen bonding 263 Isophthalamide 287 Light-to-electrical energy conversion Macrocycle 49, 64 Macrocyclic arrays, size-tunable 75 Mesopyridylporphyrins 107
163
306 Mesotetrakis(4-aminophenyl)porphyrin 266 Metallacycles 129 Metalloporphyrins 1 – –, ligand interactions 167, 170 Metal-mediated assemblies 105 Molecular sieving, film-/membrane-based 159 Molecular square 145 Monolayers, self-assembled 1 Monoporphyrin complexes, functionalised pyridyl/imidazolyl ligands 8 Multicyclic assemblies, thermodynamics 200 Multiporphyrin adducts, 2D/3D, external metal fragments 115 – –, external metal fragment 110 Multiporphyrin architectures 167 Multiporphyrin arrays 263 Multiporphyrin assemblies 32, 105 – –, bis-nitrogen ligands 12 – –, exocyclic coordination, metal fragments 107 – –, non-covalent 1 – –, rhenium-linked 145 – –, tri-/polytopic nitrogen ligands 23 Nickel(II) 2 Nitrogen ligands 7 – –, tri-/polytopic 23 Nonlinear optics 49, 93 Nucleobase, base pairing 276 O ligands 25 Palladium(II) 2 Peapods, supramolecular 271 1,10-Phenanthroline-containing ring, bis-porphyrin (PZn/PAu+ ) dumbbell 221 Photocurrent generation 96 Photo-induced processes 105 Photosynthesis 49 Photosynthetic antenna mimics, energy transfer 91 Phthalocyanines, coordination organization 84 Planar dimers 151 Platinum(II) 2
Subject Index Polymeric assemblies 287 – –, acyclic 174 Porous materials, supramolecular assemblies 155 Porphyrin assemblies 89 Porphyrin-fullerene complexes 40 Porphyrin macrocycle, guest incorporation 74 Porphyrin [2]rotaxanes, PAu+ 241 Porphyrins, higher order assemblies 122 –, ligands 1, 28 –, metal-connected assemblies, photophysical properties 126 –, metallacycles 129 –, self-complementary coordination 59 –, stoppered-rotaxanes 221 Pseudorotaxane 287 Pyrazines 270 5-(2-Pyrazinyl)pentanoic acid 266 Pyridyl-porphyrin binding 29 Pyridylporphyrins 105 Re(CO)5 (halide) 149 Reaction center (RC) 218 Rectangles 149 Rhenium 145 Rhenium porphyrin, supramolecular assemblies 146 Rhenium-imine bonds 146 Rhodium porphyrins 29 Rhodopseudomonas viridis, reaction center (RC) 218 Rotaxanes 287 –, PZn stoppers/PAu+ -appended ring 234 [3]/[5]Rotaxanes, porphyrin stoppers 227 Ruthenium porphyrins 29 Ruthenium-DMSO coordination compounds 105 Ruthenium-mediated assemblies 118 S ligands 25 Se ligands 25 Self-assembled monolayers 1 Self-assembly 167, 263 Sensing 161 Shish-kebab 174 Special-pair (SP) singlet excited state 218 Squares 146
Subject Index Supramolecular assemblies, stability constants 51 – –, structural characterization 87 Supramolecular peapods 271 Tetranucleotides 283 Tetraporphyrin adducts 111 Tetrapyrazinyl porphyrin 266 Triaminotriazines 265 –, /barbiturates 277
307 Uracil/diacetamidopyridine Urea/carboxylic acids 277
277
Zinc porphyrins 29, 170, 217 –, DABCO 177 Zirconium-phosphonate 145 Zn(II)-amine interaction 169