摘要:

Chemical Society Reviews Editorial Board Professor H. W. Kroto FRS (Chairman) Professor M. J. Blandamer Dr. A. R. Butler Professor E. C. Constable Dr. T. C. Gallagher Professor D. M. P. Mingos FRS Professor J. F. Stoddart FRS Consulting Editors Dr. G. G. Balint-Kurti Professor S. A. Benner Dr. J. M. Brown Dr. J. Burgess Dr. N. Cape Professor B. T. Golding Professor M. Green Professor A. Hamnett Dr. T. M. Herrington Professor R. Hillman Professor R. Keese Dr. T. H. Lilley Dr. H. Maskill Professor A. de Meijere Professor J. N. Miller Professor S. M. Roberts Professor B.H. Robinson Professor M. R. Smyth Dr. A. J. Stace Staff Editor Mr. K. J. Wilkinson University of Sussex University of Leicester University of St.Andrews University of Basel, Switzerland University of Bristol Imperial College London University of Birmingham University of Bristol Swiss Federal Institute of Technology, Zurich, Switzerland University of Oxford University of Leicester Institute of Terrestrial Ecology, Lothian University of Newcastle upon Tyne University of Bath University of Newcastle upon Tyne University of Reading University of Leicester University of Bern, Switzerland University of Sheffield University of Newcastle upon Tyne University of Gottingen, Germany Loughborough University of Technology University of Exeter University of East Anglia Dublin City University, Republic of Ireland University of Sussex Royal Society of Chemistry, Cambridge It is intended that Chemical Society Reviews will have the broad appeal necessary for researchers to benefit from an awareness of advances in areas outside their own specialities.Deliberate efforts will be made to solicit authors and articles from Europe which present a truly international outlook on the major advances in a wide range of chemical areas. It is hoped that it will be particularly stimulating and instructive for students planning a career in research. The articles will be succinct and authoritative overviews of timely topics in modern chemistry. In line with the above, review articles will not be overly comprehensive, detailed, or heavily referenced (ca. 30 references), but should act as a springboard to further reading.In general, authors, who will be recognized experts in their fields, will be asked to place any of their own work in the wider context. Review articles must be short, around 8-1 0 journal pages in extent. In consequence, manuscripts should not exceed 20-30 A4/American quarto sheets, this length to include text (in double line spacing), tables, references, and artwork. An Information to Authors leaflet is available from the Senior Editor (Reviews). Although the majority of articles are intended to be specially commissioned, the Society always considers offers of articles for publication. In such cases a short synopsis (including a selection of the literature references that will be cited in the review and a brief academic CV of the author), rather than the completed article, should be submitted to the Senior Editor (Reviews), Books and Reviews Department, The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF. @ The Royal Society of Chemistry, 1994 All Rights Reserved ’ No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic or mechanical, photographic, recording, or otherwise, without the prior permission of the publishers. Typeset by Servis Filmsetting Ltd. Printed in Great Britain by Blackbear Press Ltd.

ISSN:0306-0012    

DOI:10.1039/CS99423FX013

出版商:RSC    

年代:1994

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99423BX015

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

ISSN 0306-001 2 CSRVBR 23(4) 227-298 (1994) Chemical Society Reviews Volume 23 Issue 4 Pages 227-298 August 1994 /-\ I Linear Free Energy Relationships and Pairwise Interactions in Supramolecular Chemistry By Hans- Jorg Schneider (pp. 227-234) Association free energies G, of supramolecular complexes can be described by simple additive increments. These are taken either directly from plots of experimental Gt vs. the number of interactions occurring in the complexes, or -if many different ligand sites are participating -by using numbers for their electron donating or accepting properties, e.g. from single hydrogen bond measurements in carbon tetrachloride, in analogy to substituent constants in LFER. Thus empirically secured increments for Coulomb-, hydrogen bond-, van der Waals-, and ionophore interactions can be used for understanding and designing complexes of biological and industrial importance.Solution Chemistry of Lanthanide Macrocyclic Complexes By Franqoise Arnaud-Neu (pp. 235-241 ) In this article, the solution chemistry of the interactions between lanthanide cations and macrocyclic receptors is presented through selected examples referring to neutral and ionizable coronands and cryptands as well as phenolic calixarenes and their chemically modified derivatives. The different factors governing the thermodynamics of complexation and the selectivity, e.g. solvent effects, the respective sizes of the cations and the ligands, the number and nature of the binding sites, are discussed. Syntheses, Structures, and Properties of Methanofullerenes By Franqois Diederich, Lyle Isaacs, and Douglas Philp (pp.243-255) The chemistry of the fullerenes has been extensively explored -this review focuses specifically on the methanofullerenes. A comprehensive coverage of the known methanofullerene syntheses is presented. This is followed by a discussion of the electronic structure of the methanofullerenes which is based on experimental and computational studies. Lastly, we discuss the further functionalization of the methanofullerenes and their possible applications in materials and biomedical sciences. Non-ideality in Isotopic Mixtures By Gabor Jancso, Luis P. N. Rebelo, and W,Alexander Van Hook (pp. 257-264) Although widely considered to be ideal solutions, mixtures of isotopically substituted molecules (e.g.CH, and CD,) show small but significant deviations from ideality. This review discusses the experimental determination of non-ideality in solutions of isotopomers and their theoretical interpretation, including the molecular origins of the effect. Protein Structure from Linear Dichroism Spectroscopy and Transient Electric Birefringence By Michael Bloemendal (pp. 265-273) This paper discusses two less well-known techniques for studying protein structures in solution, linear dichroism (LD) spectroscopy and electric field induced transient birefringence (ETB). From the former, information on the orientation of chromophoric groups in molecules, on molecular characteristics such as shape, size, and electronic properties, and on binding parameters in molecular complexes, can be determined.From ETB hydrodynamic and electronic parameters, aggregational state, and intramolecular flexibility can be monitored. Both techniques are comparatively fast, and use relatively small quanties (0.2-1 mg) of protein at low concentration. After a general description of the principles of the techniques, their application for the study of a specific lens-protein, a-crystallin, is discussed in detail. Propagation of Interfacial Waves in Microgravity By Franqois Quirion, Marie-Claude Asselin, and Guy G. Ross (pp. 275-281 ) The propagation of waves at the interface of liquids is a topic relevant to fluid management and materials processing in space.The results obtained on earth and during parabolic flights emphasize the importance of wetting phenomena on the overall configuration of liquids and their response to perturbations in a reduced gravity environment. Crystal Engineering of Diamondoid Networks By Michael J. Zawororko (pp. 283-288) The review focuses upon recent research which has demonstrated that modular or multi-component strategies for designing and constructing diamondoid solids can be successful if they are based upon exploiting symmetry and functionality at the molecular level. Diamondoid networks can be sustained by a wide range of moieties and attractive forces can range from weak (n-hydrogen bonds) to strong (coordinate covalent bonds) as long as they are directional.The implications of the work in the general context of crystal engineering and future directions for research are also discussed. Microelectrodes: New Dimensions in Electrochemistry By Robert J. Furster (pp. 289-297) The development of electrodes with dimensions in the micrometre range has revolutionized electrochemistry by greatly extending the accessible sample environments and timescales. This article reviews the electrochemical properties of these small electrodes, as well as their application in electroanalysis, biological systems, studies of fast electron-transfer and chemical kinetics, and as probes in scanning electrochemical microscopy. Likely future advances including the direct investigation of single molecular events, the development of electron-transfer theory, and new electroanalytical techniques are also considered.Articles that will appear in forthcoming issues include Photooxidation Reactions of Transition Metal Carbonyls in Low-temperature Matrices M. J. Almond Some Aspects of the Metal-Insulator Transition J. K. Burdett The Insertion of Alkynes into Metal-Metal Bonds and Organic Chemistry of the Dimetallated Olefin Complexes R. D. Adams I, 10-Phenanthroline: A Versatile Ligand P. G. Sammes and G. Yahioglu Electrochemical Solid State Analysis: State of the Art F. Scholz and B. Meyer Aqueous Aluminates, Silicates, and Aluminosilicates T. W. Swaddle, J. Salerno, and P. A. Tregloan The Thermodynamics of Micellar Solubilization of Neutral Solutes in Aqueous Binary Surfactant Systems C.Treiner Oxidation of Some Organic Compounds by Aqueous Bromine Solutions J. Palou Biological Activity, Reactivity, and Use of Chromotropic Acid and its Derivatives J. Duda The Dynamics of Photodissociation R.N. Dixon Pericyclic Key Reactions in Biological Systems and Biomimetic Syntheses U. Pindur and G. H. Schneider Chemical Society Reviews (ISSN 030M012) is published bi-monthly by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge, CB4 4WF, England. All orders accompanied with payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd., Blackhorse Road, Letchworth, Herts., SG6 IHN, U.K. NB Turpin Distribution Services Ltd., distributors, is wholly owned by The Royal Society of Chemistry.1994 annual subscription rate E.C. 299.00, U.S.A. $186.00, Canada &I11.00+ GST, Rest of World 2106.00. Customers should make payments by cheque in sterling payable on a U.K. clearing bank or in U.S. dollars payable on a US. clearing bank. Second class postage is paid at Jamaica, N.Y. 11431. Air freight and mailing in the U.S.A. by Publications Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. U.S.A. Postmaster: Send address changes to Chemical Society Reviews, Publications Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. All other despatches outside the U.K. by Bulk Airmail within Europe and Accelerated Surface Post outside Europe. PRINTED IN THE U.K. Members of the Royal Society of Chemistry may subscribe to Chemical Society Reviews at 230.00 per annum; they should place their orders on the Annual Subscription renewal forms in the usual way.

ISSN:0306-0012    

DOI:10.1039/CS99423FP023

出版商:RSC    

年代:1994

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99423BP025

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Linear Free Energy Relationships and Pairwise Interactions in Supramolecular Chemistry Hans-Jorg Schneider FR Organische Chemie, Universitat des Saarlandes, D 6604I Saarbriicken, Germany 1 Introduction In recent years the chemistry of host-guest complexes has become one of the most rapidly expanding fields of science. lA3 Until now most of the activity has been directed towards the (frequently demanding) synthesis of host molecules, and recently also to functions of the supramolecular systems, e.g. as elements for sensors, molecular switches, etc. The essential basis of a multitude of possible applications is molecular recognition by non-covalent forces. Therefore in contrast to traditional organic chemistry, which most often concentrates on the deve- lopment of suitable methods for the synthesis of new com- pounds, almost all studies in the supramolecular field involve physical measurements, in the first place of complexation con- stants.The literature already contains thousands of such data, which until now in most cases have only been used for qualita- tive ad hoc interpretations. Systematic analyses by physical organic methods should allow the development of an experi- mental basis for a comprehensive description of the essential non-covalent forces and for a rational approach to the design of new host-guest systems. Linear free energy relationships have in the past provided the most useful way for the quantification of rates and equilibria in bond-breaking and -making reactions, for their dependence on microenvironment, for structure-activity correlations, and for obtaining insight into the reaction mechanism^.^ The analysis of binding energies and conformations with host-guest complexes in solution has exactly the same aims for the chemistry of the non-covalent bond.At the same time one can test and improve calculational models by properly designed experiments and systems, which -in contrast to biopolymers -allow focusing on any particular interaction, usually moreover in better defined conformational space. The strategy of the approach discussed in this review is based on the summation of pairwise interactions in the form of empirical complexation free energy increments; at the present time under the condition of sufficient geometrical matching Hans-Jorg Schneider was born in Stuttgart, Germany in 1935.He studied chemistry in Tiibingen, Munich, and Berlin and received his Ph.D. under Michael Hanack at the Tiibingen Institut Walter Hiickels in 1967. After postdoctoral research at the University oj California, San Diego (1967-1969) he worked on his Habili- tation and other research until 1971 in the laboratory of Walter Hiickel. In I971 he completed his Habilitation in Saarbriicken and since 1973 he has been Pro- fessor of Organic Chemistry at the Universitat des Saarlandes. His areas of interest include: conformational analyses, quantitative structure-reacti-vity relationships for aliphatic compounds, including steroids, NMR spectroscopic methods with an emphasis on I3C and 'H NMR shielding para-meters, selective functionaliza- tion of parafins, and in the last few years, supramolecular re- ceptor and enzyme models.between complementary binding sites, and/or negligible strain changes during complexation. For the control of both con- ditions, molecular mechanics calculations often play an essential role. NMR spectroscopy is of vital significance to check the geometric conditions in solution. Microenvironment effects can be quantified by, for example, empirical solvent hydrophobicity scales, or by Debye-Huckel correlations to ionic strength, etc. 2 Pairwise Interactions in Seemingly Simple 1:I Associations The description of molecular complexes by pairwise interactions has a long tradition in theoretical as well as in empirical approaches.Molecular mechanics calculations, for example, inherently involve summation of interactions between the parts of supramolecular entities. For relatively simple systems they have already led to encouraging results, particularly since the advent of the free energy perturbation methodY6 although the latter is very time consuming. The limitations here are particu- larly in the applied potential functions and their parametriza- tions. Thus, almost all force fields neglect electron lone-pair directionality as well as polarizations, although these can play a major role (see Section 7). The alternative approach of extracting parameters quantify- ing interactions of particular functionalities in organic mole- cules from appropriate series of measurements was first intro- duced to non-covalent systems by Gutmann7 and by Drago8 et al.They found that for many systems a single electron donor- acceptor parameter (e.g. CD, CA) for a particular function involved for example in a Lewis acid-base complex is sufficient to describe the equilibrium using multiplicative combination of CA and Cg, or by an additive combination of the corresponding free energy or enthalpies (equation 1). Sherry and Purcell, Ioghansen, as well as Abraham, Raevsky, Taft, and others have shown that simple 1:1 hydrogen-bonded associations in aprotic solvents obey similar linear combinations of parameters which essentially reflect basicities and acidities of the participating function^.^ In a number of cases additivity rules were also established on the basis of enthalpies AH instead of Gibbs values AG,as well as with mechanistic refinement of single parameters, reflecting for example charge-transfer, polarizibilities etc.839 Raevsky et al.were able to describe over 900 associations, measured in carbon tetrachloride by linear correlations (equa- tion l), with coefficients above r = 0.98 on the basis of single donor and acceptor parameters ED and EA;9a later these proved to be ideal for the description of ionophores (see Section 8). AG = const*CACe (1) All the associations discussed until now are based on mole- cules with single functions, generally assuming simple 1:1 com-plexes.The problems arising, however, from the occurrence of a multitude of other possible complexes (often including self- association) in these sterically mostly unbiased systems are evident from the conflicting results obtained for example with N-methylacetamide association in chloroform, which were expected to deliver cornerstone values for peptides;1° they vary from 4to 7 kJ/mol. Another serious limitation lies in ill-defined conformations, illustrated by the unresolved problem even of seemingly simple and intensively studied systems such as ammo- nia' la or benzene' lb dimers. This makes it difficult to apply such values to larger molecular associations with more selective 227 orientations. It is for these reasons that more straightforward answers are in fact obtained with more complicated, yet better defined structures as we see them in synthetic host-guest complexes.3 Additivity in Supramolecular Complexes A further problem in the application of additive interactions arises from the restriction of translatory and rotatory freedom accompanying association. The entropy disadvantage has been estimated by Jencks, Page, and others12 to amount to up to T*AS = 60-1 50 kJ/mol, part of which may be compensated by desolvation process. Supramolecular systems have the advan- tage of minimizing differential entropy effects since complex formation with the accompanying loss of mobility is already 'paid' to a large degree by one or few interactions so that additional pairwise interactions will cost less entropy sacrifice. Page and Jencks made use of this 'anchor' principle in 1971, applying it to ligand-biopolymer binding; however, only the result of exchanging one function was evaluated.The simultaneous pairwise action of many non-covalent forces is an important feature not only of synthetic, but also of natural receptor-effector complexes. Increasingly, the latter have also been used for deriving energetic factors of associ- ations. Williams et al. have estimated that adverse TAS contri-butions can be anywhere between 9 and 45 kJ/mol, being stronger as expected for more exothermic complex formation. l4 Site-directed mutagenisis has been used by Fersht et al. to estimate binding contributions of different amino-acid resi- dues.Unfortunately, the numbers for an amide-amide hydro- gen bond reported from the studies of some peptides and of proteins vary between 24 (Ref. 14) and 2 (Ref. 15) kJ/mol. The analyses of biological systems is hampered by their flexibility and complexity, leading to problems of identifying which groups are interacting in only approximately known geometries. Another reason for the observed large discrepancies must be seen in the again largely unknown involvement of water, ions, or other groups interfering at the ligand sites of biopolymers. With synthetic host-guest complexes these problems can be largely overcome. 5,1 6a Equation 2 summarizes the total com- plexation energy AG, as the result of the thermodynamic circle containing, besides the energy AGHGwhich results from direct interactions between host H and guest G, the desolvation energies AGdsof H and G -which might imply only desolvation at the contact sites -as well as re-solvation AG, of the complex.Again, the advantage of multifunctional host-guest complexes is that, for a given host, AGds of course remains constant, and that the desolvation A Gdsfor guests varies essentially with the number of functions to be desolvated before complexation, and thus will become part of the increment A A G per function which contributes to complexation. The latter assumption -like others -will be tested by success or failure of linear correlations between the experimentally observed AG, and the sum CAG from all participating functions with the increment AG for each (equation 3).A further test is whether one obtains the same or different increments AG using different host structures with, of course, the same type of interactions and the same environment with respect to solvents and salts. A complex between multifunctional host and guest structures (Figure 1) may contain different numbers 1, m,n etc. of different types of interaction characterized each by an increment A GL, AGM, dGNetc. In some complexes like aromatic ion pairs (see Section 7) one function may exert several types of interactions, such as, for example, an ion acting on a benzene ring as well as on a counterion. In the most simple case -such as in ion pairs within aliphatic frameworks (see Section 4) -the summation according to equation 3 results in a linear correlation of AG, against the number n of (for example) possible salt bridges.This indicates CHEMICAL SOCIETY REVIEWS, 1994 \ nt CI / Figure 1 Additive interactions between host and guest. In this example there are 4 equivalent interactions between complementary B sites, 2 different ones (2 of kind I, 2 of kind m) between the A sites, and 1 between the C sites, hence AG, = 4AGB+ 2dGA,,+ 2AGA,, i-1AGGC that one single increment is sufficient, which then can be obtained from the slope of the regression line. If more than one interaction type is involved than a multilinear regression will be required, or, better, increments for one type which have been determined independently will be used (see Sections 7, 8).A third, related approach is based on the use of relative para- meters, characterizing, for example, electron donor-acceptor capacities of functions occurring in a complex. These may stem from entirely different measurements or calculations (Section 8). 4 Salt Bridges: From Simple Ion Pairs to DNA Groove Binding The literature contains thermodynamic values for many inor- ganic ion pairs or, for example, tetramethylammonium salts, mostly in water as solvent, which often show large variations in entropy and enthalpy contributions.' Many of the organic host compounds which early on were proven to be very efficient binders for charged substrates, such as nucleotides,' * choline etc.rest essentially on Coulomb interactions. The latter can be identified if one compares the AG, values of uncharged eqiva- lents, i.e. the macrocyclic tetraphenolate (1) complex with either tetramethylammonmium ions or with the isosteric electro- neutral t-butyl alcohol (Scheme 1). This example19 also demon- strates how, to a first approximation, the number of salt bridges NeEt3 NeMe, CMe,OH I I 31.3 25.5 4.6 Scheme 1 The macrocyclic tetraphenolate, structure 1, and its complex with NMe,. Selected complexation free energies ([kJ/mol], water, + 25 "C,extrapolated to ionic strength = 0) (Ref. 19). SUPRAMOLECULAR LINEAR FREE ENERGY RELATIONSHIPS-H J SCHNEIDER is counted independently of the charge delocalization, which in this particular tetra-anion is very strong The plot of experimental free complexation energies A G, in water vs the number n of possible salt bridges shows a linear correlation (Figure 2) with a regression value of 5 f1 kJ per mole and per salt bridge 6aThe correlation, which within the error passes through the origin, today comprises over 70 ion pairs These range from smaller to larger ions with low and high polarizibility, such as metal, sulfonium or ammonium ions, halides, sulfonates, phosphates, carboxylates, phenolates, etc 2o The 10 kJ/mol observed for zinc sulfate in water1 6a for example is close to the expected value for divalent ions from Bjerrum theory,16 l7 21 although the latter neglects entropy contributions and should hold only for small and spherical ions The surprisingly small variability of the observed ddG increment reflects compensation effects well known from LFER studies In particular, a gain in AH, for example in hard-hard combinations of ions, will lead to less favourable AS values owing to the then stronger electrostriction of solvent molecules, the opposite will be true for interactions between soft ions Similarly, large and polarizable or soft ions will suffer from smaller Coulomb attractions but gain from charge polarization, vice versa arguments hold then for small and hard ions G * AMP^-G *ATP3-G ADP~-/pDN*+p* / O C,C' OA,B 3 6 9 12 "C Figure 2 Experimental complexation free energies dG, [kJ/mol] vs number n of salt bridges in ion pairs, in water For identification of the points see Ref 16a CD-N2 amino-cyclodextrin complexes (A V Eliseev and H -J Schneider, AngeM Chem Int Ed Engl 1993,32, 133l), DNA + PA selection from Figure 3 The affinities of polyamines with DNA also are predictable with the 5 kJ increment per charge 22 In line with this, there is a fairly linear correlation of the affinities, as measured by a fluorescence assay, and the number of nitrogen atoms in the arninesZ3 (Figure 3) The linearity observed is remarkable for an inhomogeneous biopolymer, larger deviations occur only if the amine used bears additional functionalities It should be noted not only that natural amines such as spermine are on the line but that permethylation does not alter the affinities 22 This means that hydrogen bonds play a minor role in these polyamine associations, in agreement with measurements of complexes with either protonated or again permethylated azoniacyclo- phanes which, within + 1 kJ/mol, showed the same AG, values with organic ions 24 What are the problems involved in this rather simple analysis? The first step requires the assignment of the number of interac- tions responsible for the complex formation For simple struc- lo-' t ; 2 d 5 6 ' Cationic charge Figure 3 Affinities of polyamines (PA) to ds-DNA (measured as C,, values with ethidiumbromide, see text) against the number of positive charges in the amines Values for (1) aliphatic protonated PA (Ref 23), (2) permethylated non-cyclic PA, (3) permethylated azoniacyclo- phanes (Ref 22), (4), (5) different new PA (D Ruf, B Palm, unpublished results), (6)PA with additional affinities by naphthalene and/or amide residues (B Palm, unpublished) Linear correlation coefficient with (1),(2),(4) r = 0 97 tures such as (1) + choline (Scheme 1) or tor small inorganic ion pairs one can assume contact, or secure it with the help of CPK or ball-and-stick models The latter are also helpful with more extended structures such as the azacrown ether derivatives complexing, for example, triphosphates The uncertainties introduced here are simply, for example, whether one accounts for 10 or 12 salt bridges, and thus do not alter the correlation (Figure 2) substantially However, for complicated structures such as DNA complexes computer-aided molecular modelling is almost indispensible for localizing the possible interactions Such simulations reported in the literature2 provided the basis of assigning an increment of 4-6 kJ mol and salt bridge to published DNA affinities of polyamines as well as of polyhisti- dine 22 Another complication arises if interactions other than salt bridges contribute in addition A possible way to handle this situation, which occurs in aromatic ion pairs, will be discussed in Section 7 5 Medium and Salt Effects on Ion Pairs Ion pair stabilities usually increase with decreasing solvent polarities l7 21 With some aromatic ion pairs we observe an increase, by an order of magnitude, in 80% dioxane compared with pure water 16a However, the simple dependencies (eg on the dielectric constant) assumed earlier do not generally hold For large polarizable host ions such as (1) (Scheme 1) an opposite stability increase with increasing water content in binary solvent mixtures has even been found 26 For the possible differentiation of tight and solvent-separated ion pairs which might blur the picture here the reader is referred to special texts l7 21 In contrast to solvent effects, salt effects on supramolecular ion complexes show more consistent trends as well as the expected increase in AGt with decreasing ionic strength 26 27 Even with large and quite anisotropic organic ions bearing multiple charges, surprisingly linear correlations with Debye- Huckel coefficients were observed26 27 (Figure 4) However, larger polarizable ions, which are often used in organic buffers, may within a cavity of suitable charge and size lead to strong deviations owing to specific complexation 26 6 Hydrogen Bonds in Peptide and Nucleobase Analogues Hydrogen bonds of the amide-amide type dominate in the most efficient artificial hosts of high selectivity, developed in particu- CHEMICAL SOCIETY REVIEWS, 1994 4.5 k ul -0 3.5 0.0 0.1 0.3 dZ/(l + dZ)l 0.5N [NaX] Figure 4 Correlation of ion pair lg K values with Debye-Huckel coefficients of ionic strength26 for cyclophane structure (1) + Et,NBr.Hamilton Still * B’ Scheme 2 Some examples for amide-type hydrogen bonds (see Refs.5 and 28 for details). lar by Hamilton, Rebek, Still, Zimmerman, and others (Scheme 2).5,28 They contain a number n of well-tailored hydrogen bond donors and acceptors which again show a linear correlation with the measured complexation free energies2* (Figure 5). Devi-ations from the correlation line are seen only for complexes which either suffer from particular entropy disadvantage as result of their acylic parts, or contain additional stacking interactions. The strong influence of the medium on the complexation can be quantified by the hydrogen bonding capacity of the solvent which competes in high molarity for the solute. This leads to an increase in the constants K by factor of 10-20 if CC1, is used instead of CHC13.2* How does the structure and arrangement of donor and acceptor groups affect the result^?^^,^^ Although the measure- ments with corresponding model compounds are far from being ac~urate~~,~we note that the Watson-Crick nucleobase pairs roughly follow the additivity scheme for amides (see above).For A-T with 2 hydrogen bonds one observes in chloroform K = lo2 M-l, for G-C with 3 bonds (Scheme 3) K = lo4 M- or higher,29*3 which is already more than expected. Other triply 11 ’./ 30 9 *’ / 012 /’ +10 0 Q 2o // 4+ /Pa 3+ / 10 2 4 6 n Figure 5 Free complexation energies of amide-type complexes measured in chloroform as a function of the number n of hydrogen bonds; for identification of points and explanation for the systems deviating from the regression see Ref.28. bound associations such as U-DAP (Scheme 3) show surpris- ingly low constants of around 102.29 Jorgensen et al., on the basis of molecular mechanics/MC calculations, have convincingly demonstrated that the reasons for these discrepancies are secondary electrostatic interactions between the adjacent donor and acceptor groups.29 Besides the direct hydrogen bond, to which we assign an increment P in RI A-T 2 P -2s EC&. = -44.4 Eexp.(HCC13) = -8.4 Eexp.(ccI,)= -1 4.3 Scheme 3 Hydrogen bonds in nucleobase analogues. Partial charges designated by + / -following Jorgensen et a1 (Ref. 29). Primary interactions (hydrogen bonds), P; secondary interactions, S; Ecalc,AG, calc~lated~~; AG, in CDC1, (all in [kJ/mol], for data see Refs.29, Eexpr 3 1). SUPRAMOLECULAR LINEAR FREE ENERGY RELATIONSHIPS-H -J SCHNEIDER 23 1 Scheme 3, the negative partial charges at the acceptor atoms 0 and N lead to additional attraction with a positively charged neighbouring proton, but to repulsion with neighbouring 0and N atoms The combination G-C has two such attractive second- ary minus/plus combinations, and two repulsive secondary combinations (assigned S in Scheme 3), which therefore cancel each other However, U-DAP has four repulsive plus/plus or minus/minus interactions, and A-T again two repulsive S increments, which explains the weak binding As Jorgensen et a1 29 have pointed out, their calculational results could be represented by an additive scheme with 30 kJ/mol for each primary hydrogen bond, and 11 kJ/mol for each either repulsive or attractive secondary electrostatic interaction Such additional interactions between adjacent groups are reminiscent of related problems with the Hammett equation, which for quite similar reasons cannot usually be applied to arenes with groups in the ortho-position (See note added in proof, p 234 ) 7 Van der Waals Interactions/Hydrophobic Effects, Aromatic Ion Pairs, Associations with Porphyrins Lipophilic and hydrophobic effects are known to play a major role in biological systems sb Earlier investigations of the aggre- gation behaviour of lipophilic model compounds32 has been hampered by the ill-defined stoichiometries and conformations from stacked aromatic compounds as well as by measurement problems Application of the additivity pnnciple to supramole- cular complexes allows, for the first time, the analysis of such complexes by reliable methods The strategy is to provide host and if necessary also guest compounds with ionic groups (Scheme 4), besides securing water solubility these can produce salt bridges for which the corresponding increments are known from the independent measurements discussed in Section 4 After correction of the observed dGt values with the 5 kJ/mol increments for each salt bridge one indeed observes linear correlations of, for example, the number m of phenyl rings occurring in aromatic ion pairs and A G,(Scheme 4, Figure 6) 6a (4 coo-Ocoo-po-)Q SO3-Scheme 4 Some aromatic ion pairs with additional van der Waals interactions (Ref 16) 012345678 m Figure 6 Correlation of non-ionic interactions energies with the number m of phenyl rings in aromatic ion pairs, after correlation for salt bridge contributions [see Ref 16a,also for explanation for deviating points (HI and (I)] This corresponds to equation 2, with AGvdW = AG, -n*5, where n = number of salt bridges Deviations are only observed for systems in which for steric reasons the ion cannot approach the n-moiety of an opposing ring The increment derived from Figure 6 for this van der Waals interactions is around 2 kJ per mol and per phenyl moiety That this increment reflects largely the dipole induced in the arene by the ion is supported by measurements with corresponding cleft systems (Scheme 4)for which we observe similar values with anion -7~ instead of cation -n interaction 166 A recent study of porphyrin associations allowed the normali- zation of such lipophilic interactions with respect to the number of T electrons involved, as well as separating them from solvo- phobic effects 33 The porphyrins used provide a n-surface of constant and large size interacting with a variety of smaller organic ligands, the observed AG, values are remarkably inde- pendent of the number or position of nitrogen atoms within the ligands A regular difference of 5 kJ per mol salt bridge is observed again for the difference between electroneutral and charged ligands The A G values, after being corrected as before (equation 2) for ionic contributions, if applicable, correlate surprisingly well with the number of n-electrons involved in the associations (Figure 7) Copper ions in the porphyrin have negligible influence, whereas zinc leads to deviations from the correlation due to the known demand of axial coordination Can one distinguish lipophilic from solvophobic contribu- tions in such systems? That saturated ligands show no binding contribution beyond their salt bridges (Scheme 5) indicates that, even in pure water, hydrophobic effects on ligands of this size are negligible, in rough agreement with published3" lower-end values of 0 1 kJ per mol and A2 In line with this, the observed increase of binding with the porphyrin associations with increas- ing water content in binary mixtures is much much smaller than observed, for example, for cyclodextrin complexes We have shown earlierz6 that the sensitivities of ligand complexation energies against water content can serve as a measure of hydrophobic binding contribution 8 Crown Ether and Cryptand Complexes -A Case for Multiple Regression Analysis? Ionophore complexes differ from Figure 1 in that the guest atom has only one binding site, which, however, can use quite different donor functions of a host Consequently, the rich chemistry which has been developed for this oldest and, in terms of application, most important class of host-guest complexes has provided a large range of different donor functions such as R-0-R, R-NH-R, where R = alkyl, phenyl, heterocycle, etc CHEMICAL SOCIETY REVIEWS, 1994 60 -0 o0 Scheme 5 Porphyrin complex, with typical AG, values ([kJ/mol], water, 25 "C) for aromatic and saturated ligands (Ref.33). 25 fi 20 15hE3 25 105Q 5 -5 0'< -5 0 5 10 15 x-electrons Figure 7 Correlation of non-ionic contributions to the binding with porphyrins with the number n of T electrons in the ligands; the point at n = 6 is the average value from 6 complexes, the one at n = 10 from 10 complexes (Ref. 33). These functions must be characterized by a whole range of separate parameters. In principle one could try to identify such parameters directly from the correlations, as outlined in the previous sections.Here one would need multilinear regression as it is to a lesser degree the number, but more the kind of interactions which determines the observed large variation of complex stability. However, although the complexation energies reported, e.g. for the potassium complexes of crown ether and cryptand type, range over 40 kJ/mol, even if only hosts of appropriate size are taken into consideration the topological variation in view of the obvious geometric restriction for iono- phores seems rather small. Quite in analogy to classical LFER, where substituent constants are taken from other reactions than the one to be analysed, we therefore decided to use electron donor factors ED9awhich were obtained from measurements of simple 1:1 hydrogen bond equilibria in carbon tetrachloride (cJ Section 1).For the analy~is,~ based on equation 3, all available stability constants were used except from those crown ethers or cryp- tands which are either too small or too large for a given cation, and therefore cannot materialize all possible interactions simul- taneously. If the latter are included one observes poor correla- tions (Figure S), very much in analogy to LFER in the case of steric distortions. Otherwise, the use of additive increments leads for the first time to a comprehensive and rather accurate Figure 8 Correlation of AG, for crown ether and cryptand complexes with K+ (in [kJ/mol], methanol, 25 "C) with the sum of ED increments (Ref. 35); including ionophores which are too large (mismatch).,,.,.,,,70 I 60 50 h-0E 40 325-8 30 20 10 0 0 2 4 6 8 10 12 14 16 18 20 cED Figure 9 Correlation of AG, for crown ether and cryptand complexes with K+ (in [kJ/mol], methanol, 25 "C) with the sum of ED increments (Ref. 35). Different 18C6 and [2.2.2]cryptand complexes. description of crown ether and cryptand complexation energy on a common scale (Figure 9). On the basis of only 11 ED group (substituent) parameters which were not treated as adjustible but stem from independent hydrogen bond measurements, we could in this way calculate over 120 complex ~tabilities.~~ This also included complexes with R +NH3 cations, without taking into account special factors for hydrogen bonding. This suggests that the binding here is of similar, largely electrostatic nature to that of metal cations.As with related LFER it is necessary to keep the solvent constant, as well as the anion. The sensitivities (slopes) for different cations again seem to be a simple linear function of their hydration energie~.~ 9 Additive Interactions in Protein-Ligand Complexes The identification of functional group contributions in ligand- protein complexes was started over 20 years ago by Page and Jencks.I3 Through the availability of many data, in particular on equilibrium constants of enzyme inhibitors, the fast progress of computer-aided structure simulation, and owing to the SUPRAMOLECULAR LINEAR FREE ENERGY RELATIONSHIPS-H J SCHNEIDER medicinal and industrial significance of rational drug design, such studies have gained considerable momentum In particular Andrew~,~~ andG~odford,~~ have extended the approach to the analysis of the many interactions acting simulta- neously in a protein-ligand complex More recently, Bohm has combined computer-aided automated searches for the identifi- cation of possible binding sites in drug receptors with scoring functions for obtaining free energy increments 38 The results of such a multilinear regression38c (Figure 10) demonstrate the promise of the additivity principle in this field It should be noted that the regression in Figure is obtained with adjustable increments The scatter demonstrates the problems involved in the extraction of single increments from biopolymer studies, for the reasons discussed in Section 3 15 12 -0hU Q,c0 9--0 2 Q v k-w 6-0 T 0 3 6 9 12 15 -log 4 (experimental) Figure 10 Estimated vs experimental binding constants K, for 45 protein-ligand complexes, after H -J Bohm (Ref 38c) 10 Conclusions Numerical descriptions of chemical systems will show their practical usefulness by systematical comparison between calcu- lated and experimental data, which is the reason why we have confronted the reader with so many linear correlations It is desirable to arrive at descriptions which will predict properties with few, simple, and chemically meaningful rules and para- meters, rather than to calculate them afterwards The supramo- lecular complex is not a black box, but a challenge for physical organic chemistry We hope to have shown that quantitative predictions are already possible within certain limits for com- plexes comprising many non-covalent bonds This should allow the experimental chemist to design new supramolecular systems, as well as to learn from experiment -in our case especially from stabilities and conformations of host-guest complexes -and thus to improve theoretical models Why, and for which cases, does one see such simple linear descriptions of supramolecular complexes7 First, the reader is reminded that there is no rigorous theoretical foundation for any LFER, the possible compensational factors, for example with respect to entropy/enthalpy, hard/soft behaviour etc have been mentioned in Section 4 A significant factor in the obvious tolerance of many supramolecular complexes to smaller steric distortions lies in the dominance of electrostatic forces, which fall off with distance very slowly, another one is the geometric tolerance in solvophobic interactions Future efforts will be directed towards refinement of the additivity approach by broadening the experimental basis as well as by taking into account (1) secondary interactions, such as in nucleobase pairs discussed in Section 6, (ii) unsufficient geometric matching, and (111) steric distortions with concommitant strain-energy changes 11 References 1 D J Cram, Angew Chem Int Ed Engl ,1986,25,1039, D J Crdm Science 1988,240, 760 2 J -M Lehn, Angew Chem Int Ed Engl , 1988,27,89, J M Lehn, in ‘Frontiers in Supramolecular Organic Chemistry and Photo- chemistry’, ed H -J Schneider and H Durr, VCH, Weinheim 199 I, p 1, J -M Lehn, Angew Chem ,1990,102,1347, Angew Chem Int Ed Engl , 1990,29, 1304 3 J F Stoddart, Annu Rep Progr Chem Sect B, 1988, 85, 353, ‘Host-Guest Molecular Interactions From Chemistry to Biology’, Wiley, Chichester, 1991 4 For a recent review highlighting the role of LFERiQSAR for the understanding of chemistry life interactions with physical organic chemistry “on the verge of a golden age” see C Hansch, Acc Chem Res, 1993,26, 147 5 See also H -J Schneider, Angew Chem Int Ed Engl , 1991, 30, 1417 6 For references on the computational work of Jorgensen, Kollman, Wipff, and others see (a)‘Computational Approaches in Supramole- cular Chemistry’, ed G Wipff, Kluwer Academic Publishers, 1994, (b)H -J Schneider, Rec Trav Chim Pays-Bas, 1993, 112, 412, (c) Ref 5 7 V Gutmann, ‘The Donor Acceptor Approach to Molecular Inter- actions’, Plenum Press, New York, 1978 8 R S Drago, in ‘Structure and Bonding’, Springer, Heidelberg, 1973, p 73ff 9 See recent reviews (a)0 A Raevsky, Russ Chem Rev, 1990, 59, 219, (b) M H Abraham, Chem SOC Rev, 1993, 22, 73, and references cited therein 10 G E Schulz and R H Schirmer, ‘Principles of Protein Structure, Springer, New York, 1979, J N Spencer, R C Garrtett, F J Mayer, J E Merkle, C R Powell, M T Tran, and S K Berger, Can J Chem , 1980,58, 1372, and references cited therein 1I See (a) R J Saykally and J G Loeser, J Chem Phys , 1992, 97, 4727, (b)S L Price and A J Stone, J Phys Chem , 1987,86,2859, for recent references see G Klebe and F Diederich, Phil Trans R SOC Lond A, 1993,345,37 12 See M I Page, Angew Chem Int Ed Engl, 1977, 16, 449, W P Jencks, Proc Natl Acad Sci USA, 1981,78,4046 13 M I Page and W P Jencks, Proc Nat Acad Sci USA, 1971,68, 1678 14 (a)M S Searle and D H Williams, J Am Chem SOC , 1992, 114, 10690, (b) M S Searle, D H Williams, and U Gerhard, J Am Chem Soc ,1992,114, 10697 15 (a)A R Fersht et a1 ,Nature, 1985,314,235, see also (b)S K Burley and G A Petsko, Adv Protein Chem ,1988,39, 125 16 (a)H -J Schneider, T Schiestel, and P Zimmermann, J Am Chem Soc , 1992, 114, 7698, see also (b) H -J Schneider, F Werner, and T Blatter J Phys Org Chem , 1993, 6, 590, and references cited therein 17 K Burger, ‘Solvation, Ionic and Complex Formation Reactions’, Elsevier, Amsterdam, 1983, U Meyer, Coord Chem Rev ,1976,21, 159 18 E Kimura, Top Curr Chem, 1985, 128, 131, 141, E Kimura, in ‘Crown Ethers and Analogous Compounds’, ed M Hiraoka, ‘Studies in Organic Chemistry’ Vol 45, Tokyo, 1992, p 381, M W Hosseini, J -M Lehn, and M P Mertes, Helv Chim Acta, 1983,66, 2454, M P Mertes, K Bowman, and K B Mertes, Acc Chem Res , 1990,23,413 19 H -J Schneider, D Guttes, and U Schneider, J Am Chem SOC , 1988,110,6449 20 H -J Schneider and I Theis, Angeu Chem Int Ed Engl , 1989,28, 753 21 A Marcus, ‘Ion Solvation’, Wiley, Chichester, 1986, Y Marcus, Chem Rev ,1988,88,1475, J A Gordon, in ‘The Organic Chemistry of Electrolyte Solutions’, Wiley, New York, 1975 22 H -J Schneider and T Blatter, Angew Chem Int Ed Engl , 1992, 31, 1207 23 K D Stewart and T A Gray, J Phys Org Chem , 1992,5,461 24 H -J Schneider, T Blatter, B Palm, U Pfingstag, V Rudiger, and I Theis, J Am Chem SOC , 1992, 114, 7704, and references cited therein 25 See e g L Stekowski, D B Harden, R L Wydra, K D Stewart, and W D Wilson, J Mol Recog, 1989,2, 158 26 H -J Schneider, R Kramer, S Simova, and U Schneider, J Am Chem SOC , 1988,110,6442 27 H -J Schneider and I Theis, J Org Chem , 1992,57, 3066 28 H -J Schneider, R K Juneja, and S Simova, Chem Ber , 1989,112, 121I 29 For arguments, data, and references see: J.Pranata, S. G. Wierschke, and W. L. Jorgensen, J. Am. Chem. SOC.,1991,113,2810. 30 See also T. J. Murray and S. C. Zimmermann, J. Am. Chem. SOC., 1992, 114,4010, and references cited therein. 3 1 H.-J. Schneider, H.-D. Junker, and J. Sartorius, unpublished results. 32 See K. A. Connors, A. Paulson, and D. Toledo-Velasquez, J. Org. Chem., 1988,53,2033,and earlier references. 33 (a) H.-J. Schneider and M. Wang, J. Chem. SOC.,Chem. Commun., 1994, 413; (b) unpublished results; see also (c) Ref. 16b. 34 F. M. Richards, Annu. Rev. Biophys. Bioeng., 1977,6, 151; a 100% higher value has also been estimated: K. A. Sharp, A. Nicholls, R. Friedman, and B. Honig, Biochemistry, 1991, 30,9686. Note Added in Proof A recent analysis shows that association free energies (in chloro- form) of about 40 nucleobase analogues can be described by single increments of 7.9 kJ/mol for primary hydrogen bond interactions, and of 2.9 kJ/mol for secondary interactions (H.-J. Schneider, J. Sartorius). CHEMICAL SOCIETY REVIEWS. 1994 35 H.-J. Schneider, V. Rudiger, and0. A. Raevsky J. Org. Chem., 1993, 58, 3648. 36 P. R. Andrews, D. J. Craik, and J. L. Martin, J. Med. Chem., 1984, 27, 1648. 37 P. J. Goodford, J. Med. Chem., 1985,28,849;D. N. A. Boobyer, P. J. Goodford, P. M. McWhinnie, and R. C. Wade, J. Med. Chem., 1989, 32, 1083. S. H. Rotstein and M. A. Murcko, J. Med. Chem., 1993,36, 1700;E. C. Meng, B. K. Shoichet, and I. D. Kuntz, J. Comp. Chem., 1992, 13, 505. For related factorizations see e.g. B. Tidar and M. Karplus Biochemistry, 1991,30, 321 7 and references cited therein. 38 (a) H.-J. Bohm, J. Comp. Chem., 1992, 6, 61; (b) 1992, 6, 593; (c) 1994, in press (I thank Dr. Bohm for a preprint).

ISSN:0306-0012    

DOI:10.1039/CS9942300227

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

SoIution Chemistry of Lant ha n ide MacrocycIic Corn plexes F. Arnaud-Neu Laboratoire de Chimie- Physique URA no405 au CNRS E. H.I. C.S. I rue Blaise Pascal 67000 Strasbourg France 1 Introduction Macrocyclic receptors define a hydrophilic cavity in which an ionic substrate like a metal ion can nest and be shielded from the environment by its lipophilic envelope.' Accordingly they can mediate cation transfer from an aqueous medium to a lipophilic phase and were first designed as synthetic models to mimic the ionophoric properties of natural antibiotics towards alkali and alkaline-earth cations. They have been extensively studied with these cations and it has been shown that the main factor affecting complexation selectivity is the size match between the cavity and the metal ion. The variety of macrocycles prepared which has increased greatly over the past two decades includes the coronands with the well-known and studied crown-ethers,* cryptands,' ~pherands,~ and amongst the last synthesized the calixarenes and their chemically modified derivative^.^ The success of numerous studies on the complexation of alkali and alkaline-earth cations by macrocycles led to similar studies with a related series of cations the lanthanides. These cations have a common +3 oxidation state (Ln3 +),although samarium europium and ytterbium can also exist as Ln2 ,and cerium as + Ce4+ . They behave as hard acids with a strong affinity for hard bases like oxygen and negatively charged groups and they form essentially non-directional electrostatic bonds. The size of Ln3+ cations decreases with increasing atomic numbers (lanthanidic contraction) and the achievement of high coordination numbers is an important requirement which can be met with macrocyclic ligands. Initial interest in these studies was academic viz. the investigation of the coordination properties of lanthanides and the understanding of the factors governing the complexation ability and selectivity the solvent effects and the stabilization of unusual oxidation states -but the increasing industrial develop- ment of lanthanides in the fields of catalysts glasses ceramics magnets optics and electronics now requires continuous improvement in separation processes. More recently owing to their paramagnetic and luminescence properties the possible applications of lanthanide macrocyclic complexes in chemistry biology and medicine as relaxation agents for imaging tech- niques structural and analytical probes as well as labels Francoise Arnaud-Neu is presently Research Director at the Centre National de la Recherche Scientijique -in the laboratory of Physical Chemistry of the Higher Institute of Chemistry of Strasbourg France. She graduated in Chemistry in 1968 from the University of Strasbourg where she also received her Ph.D. in 1973 studying the photo-chromic behaviour of amine molybda tes. After her thesis she turned her research interest towards the thermodynamics of complexation of metal ions in solution by macrocyclic recep- tors such as coronands crypt- ands and catenands. Since 1988 her research activity has been mainly focused on the study of the binding properties of calixarenes. She has some 60 publications to her credit. (fluoroimmunoassays) have provided additional reasons for their extensive study.5 In the following sections a non-exhaustive report of the solution thermodynamic properties of lanthanide macrocyclic complexes will be given associated with the different fundamen- tal or more applied interests cited above. First the basic question of the factors governing the complexation selectivity within the lanthanide series will be addressed where results concerning lanthanide coronates and cryptates will be pre- sented. Secondly the properties of lanthanide complexes with anionic macrocycles will be reviewed. Finally the last section will describe recent results on the complexing properties of calixarenes towards lanthanides. The binding of lanthanides by various macrocycles which is presented in this survey has been established mostly through stability constant determinations in homogenous media per- formed mainly by absorption spectrophotometry potentio- metry and calorimetry. In some cases other thermodynamic parameters such as the enthalpy (AH,) and entropy (AS,) of complexation have also been determined as well as biphasic extraction data of lanthanide cations from water to an organic phase. These results will be reported and illustrated graphically. However the latest results i.e. log /3 values pertaining to the complexation of lanthanide cations by calixarenic receptors and which are so far unpublished are given in Table 1. They correspond to the concentration ratios p = [LnL]/[Ln][L] where the charges are omitted and L is a neutral calixarene or the fully deprotonated form of the ligand in the case of calixarene-acids. Until 1986 references will be made throughout the text only to the exhaustive review by J. C. G. Biinzli on lanthanide complex- ation by synthetic ionophores.6 2 Lanthanide Complexes with Neutral Macrocycles 2.1 Complexes with Coronands. Influence of the Nature of the Heteroatoms A large variety of coronands -some of which are represented in Figure 1 -are able to complex lanthanide cations as demon- strated by the great number of complexes isolated in the solid state. For instance complexes of crown-ethers diazapolyoxa- monocycles and even polyazamonocycles are readily obtained whereas the synthesis of lanthanide complexes with ligands containing the softer sulfur atoms is less straightforward because of their lower affinity for hard cations. The stoichiome- tries of the solid complexes formed with the crown-ethers depend on different factors such as the nature of the medium the anion present and the respective size of the cation and the ligand cavity. Some of these stoichiometries have been con- firmed in solution. The mononuclear 1:l complexes the only ones which form with diazapolyoxamacrocycles are more stable than their homologues with crown ethers. For instance in propylene carbonate (PC) the difference in stability between 2 1 and 15C5 complexes of lanthanum is around 8 log units (Figure 2). Although both ligands are monocycles with the same size and number of donor atoms they differ only through the substitution of two ether oxygen atoms by two secondary amines. The same situation holds for 22 and 18C6. The presence of the nitrogen atoms in 21 and 22 should be partly responsible for the increase of stability owing to the higher polarizability and the higher flexibility of these ligands. This explanation is further supported by the examples of A 18C6 in which all the 235 CHEMICAL SOCIETY REVIEWS 1994 ' 1 log R (MeOH)I-0-0 X=O,m =1:15C5 m =1:21 X = 0,m = 2 18C6 m =2:22 X = NH m = 2 A618C6 m =n =0:211 m =O;n =1.221 pyo2[l8]dieneN6 m =n =1:222 Figure 1 Chemical structures of some coronands and cryptands. y-2-18C6 ~ 15C5 La Pr Sm Gd Er Yb OI' . I I I 1 56 62 60 74 Figure 2 Variation of log 6 with 2for some lanthanide cryptates and coronates in propylene carbonate. oxygens of the 18C6 are replaced by -NH groups,' and of pyo,[l 8]dieneN,.s Both ligands complex the trivalent lantha- nides strongly even in water with values of log ,!Iranging from 9.1 to 1 1.8 along the series for the former and values of 7.4 and 8.I for the latter with La3 + and Gd3 +.These results show that lanthanides have a great affinity not only for oxygen but also for nitrogen atoms. No sulfur-containing ligand has been investi- gated in solution with lanthanides so far. 2.2 Effect of the Lanthanidic Contraction Figure 3 shows the variation of log ,!I of 21 221 and 18C6 complexes in methanol (MeOH) with the atomic number Z. This parameter is chosen rather than the ionic radii because of their strong variation with the coordination number which is unknown. Complexation by 221 and 21 occurred only after 12 and 7 days of equilibration of the solutions respectively. A constant cryptate effect of ca. 1.5 log units is observed between 221 and 21 complexes. An increase in stability is found with 21 and 221 which contrasts with the behaviour of 18C6.(j In PC as in MeOH the effect of the lanthanide contraction is different with cryptands and crown-ethers (Figure 2).9 The stability of cryptates varies irregularly but the overall trend is a slight increase in the series. Generally the increasing charge density is invoked to explain the increase in stability of lanthanide chelates along the series. The peculiar behaviour of coronates could be predicted from their inclusive nature. When going from La3 + to Yb3+,the cations become less and less adapted to the cavity and this effect is predominent. The fact that cryptands do not exhibit the same trend should be related to the presence of nitrogen atoms which masks the expected size effect. 56 62 68 74 2 Figure 3 Variation of log /3 with Z for some lanthanide cryptates and coronates in methanol. 2.3 Solvent Effect -Transfer Properties The thermodynamic properties of trivalent lanthanide cryptates have also been investigated in solvents other than MeOH and PC :water,6 dimethylsulfoxide (DMSO) N,N-dimethylforma- mide (DMF),'O and acetonitrile (AN).' Their stability increases in the following order DMF < DMSO < H,O < MeOH < AN << PC As shown for europium cryptates (Figure 4),the most stable complexes are generally formed with 221 in poor solvating solvents and with 21 1 in strongly solvating ones. It is clear that the results cannot be interpreted in terms of size effects only. They emphasize the role of the solvent which is likely to interact with the highly charged cations. The role of the anion may also be important as it has been shown that Ln3 + can be complexed by C10 in poor solvating media.' As the log ,!Ivalues reported are mostly determined in the presence of Et,NC10 as support- ing electrolyte the results ought to be corrected for this effect. However these corrections should remain small in comparison with the variation of stability between the different solvents and should definitely not alter the trends observed. 20 0 22 A 21 + 211 A 221 222 10 0 18C6 LL A F2 0 + 0 6 0 0 VI I I 12 18 24 DN Figure 4 Variation of log /3 with the Gutmann's donor numbers of the solvents (DN) for trivalent europium cryptates. In PC and in AN lanthanide cryptates form two classes according to their degree of stabili~ation:~~~~' 1,1 those stabi- lized by both favourable enthalpy and entropy and those stabilized only by the enthalpic terms the entropy contribution being negative and thus unfavourable. As the enthalpy terms are rather similar in PC and AN (Figure 9 the extra-stability of lanthanide cryptates in PC is due to favourable entropy changes originating from a more important release of solvent molecules. The variation of the entropy with the atomic number suggests an incomplete desolvation of the cation. On the basis of full SOLUTION CHEMISTRY OF LANTHANIDE MACROCYCLIC COMPLEXES-F ARNAUD-NEU AG AH TAS La Pr Nd Er 30 kcal mol 1 AN 20 nrn AG 0 AH TAS 10 0 10 La Pr Nd Figure 5 Energetic diagram for some lanthanide cryptates of 221 in propylene carbonate and acetonitrile desolvation an entropy gain should be expected on going from La3+ to Nd3+ since the smaller the cation the greater its solvation The opposite trend is observed and shows that the cation retains part of its solvation shell in the complexes suggesting interactions between complexed lanthanide and the medium The single ion enthalpies of transfer from PC to AN of 221 and 222 cryptates determined using a thermodynamic cycle between the two solvents or the direct measurements of the enthalpies of solution of the lanthanide salts are independent of the ligands but vary with the cation l3 In fact they are equal to the enthalpies of transfer of the cations indicating a cation recognition by the solvent despite the presence of the ligand Therefore the 'cryptate' extrathermodynamic convention which implies the complete shielding of the cation by the cryptand in the cryptate is not valid for trivalent lanthanides 2.4 Stabilization of Unusual Oxidation States of Lanthanides Stabilization of divalent samarium europium and ytterbium can be achieved by complexation with cryptands or even cor- onands In H,0,6 DMF,l0 and MeOH,(j the reduction of trivalent lanthanide cryptates and coronates always occurs at higher potentials than the reduction of the uncomplexed cations indicating a stabilization of the divalent complexes but in PCi4 and AN,' no systematic stabilization is observed When these reductions are reversible and without any charge in the stoichio- metries of the complexes the stability constants of the divalent complexes can be calculated from the difference in the reduction potential of the complexed and uncomplexed cations and the stability of their trivalent homologues The variation of log p thus obtained with the Gutmann's donor numbers (DN) of the solvents (illustrated in Figure 6 in the case of the europium complexes) shows that (1) the stability of a given complex is quite independent of the solvent (11) 222 forms the most stable complexes in agreement with size considerations in all solvents and (111) a very strong macrobicyclic effect is observed the 222 VI I I 12 18 24 DN Figure 6 Variation of log /3 with the donor numbers of the solvent for divalent europium cryptates difference in stability between 222 and 22 complexes being 7 5 and 10 log units in AN and DMF respectively These results show that in contrast to trivalent lanthanides the stability of divalent lanthanide cryptates is controlled by size rather than by solvent effects 3 Lanthanide Complexes with Ionizable Macrocycles Another class of macrocyclic ligands with pendant ionizable functions has been and still is being widely explored with lanthanide cations The presence of carboxylic acids the ioniz- able groups most frequently met with in these macrocycles imparts two characteristics on these ligands (1) the formation of strong complexes owing to the establishment of electrostatic interactions and consequently the selective complexation of lanthanides over alkali and alkaline-earth cations (11) the solubi- lity in water of the ligands and their complexes These properties are essential for the application of these ligands to the analysis of biological material zn vzvo The stability of such complexes should also depend on different factors such as the ligand flexibility the number and the nature of the donor sites and the steric requirements of both the macrocyclic framework and the carboxylate groups Therefore in order to discover the best lanthanide selective agents a great variety of compounds of different sizes (from 9-to 18-membered rings) with different types of donor atoms have been synthesized and tested for complexation of lanthanides In some compounds the carboxy- hc acids have been replaced by phosphorous-containing acid functions Some of the most representative structures are shown in Figure 7 The generally high stability constants of the fully deproto- nated complexes of polyamino polycarboxylate ligands have often been determined by carrying out competitive experiments using a precipitation agent15 or an auxiliary ligand either Arsenazo-111' or EDTA (ethylenediaminetetraacetic acid) The ligand DACDA,6 derived from the diazatetraoxamono- cycle 22 by substitution of the two secondary amine functions with methylene carboxylic acids forms less stable complexes than EDTA as 10 8510g p< 12 2 instead of 15 %log /3< 19 8 The log /3 values are approximately constant for the lighter cations and decrease for the heavier ones This behaviour is in contrast with that of EDTA complexes the stability of which increases regularly along the series However there is a good selectivity for Ln3 + with respect to the other cations contrary to EDTA which is not able to discriminate between alkalis alkaline-earths and lanthanides The related 15-membered DAPDA6 forms complexes of similar stability (10 1 slog pi 1 1 9) and exhibits a small selectivity for Eu3 The octaden- + tate ligand DOTA' complexes lanthanide cations very strongly (28 2 Ilog /3 I29 2) approximately 10 times more than EDTA However it is less selective TETAl complexes have compar- able stabilities to those of EDTA (14 5 slog /3< 16 5 at 80 "C) CHEMICAL SOCIETY REVIEWS 1994 LNJ)< R=COZH NOTA R=COzH m =1 DAPDA R = HPO NOTP R=C02H m =2 DACDA R = PO(0H)OEt NOTPME R"'1/7CR(1 11 "I-(" R R=C02H DOTA R = COZH R = R = H D03A R = H2P03 DOTEP R=COZH R =H R -CH D03MA R = COzH TETA Figure 7 Chemical structures of some ionizable macrocycles Both DOTA and TETA behave as classical non-cyclic ligands as the stability of their complexes increases along the series with the increase of the charge density The predominance of the electro- static factor is further supported by the observation that log p values for the Gd complexes of DOTA and its 12-membered triazatriacetate homologue the hexadentate NOTA lie on the linear plot of the stability constants vs XpK for a variety of polyaminopolycarboxylate open ligands all forming 5-membered 0-Ln-N rings l6 Most of the more recent papers deal with the complexation of the highly paramagnetic Gd3 + cations which is the best suited for contrast agents by tricarboxylic acids able to form non-ionic complexes l7 2o The carboxylic arms can be substituted in various ways in order to increase the rigidity and/or the number of donor sites For instance D03MA in which the three propionic acid arms are a-methylated forms the most stable heptadentate gadolinium complex with log ,8 = 25 3 l8 It is nearly as stable as its DOTA homologue (log p = 25 8) despite the lower donor site number of the ligand but in agreement with the relative basicity of the two ligands Recently Martell et a1 studied a series a triazamacrocyclic ligands with three pendant acetate donor groups and increasing number of ether oxygen atoms l7 The general trend is a decrease of the stability of the complexes with the increase of the cavity size However the Gd complex of the large flexible 18-membered ring which enables a good adjustment of the cation to all the donor atoms in agreement with the structure of the solid complex is slightly more stable than its 15-membered homologue The more rigid 12-membered ligand is the best complexing agent as it favours a cage structure A series of macrocyclic bis-amide derivatives of EDTA and DTPA (diethylenetriamine pentaacetic acid) have also been tested with Gd3 + ,one of them displaying the highest Gd3+ /Zn2+ selectivity ever found with polyaminopolycarboxy- late ligands l9 All these ligands especially DOTA and D03MA are amongst the best sequestering agents for lanthanides However the accurate determination of large stability constants is difficult and several values have been reported for the GdDOTA complex in the range 28 0 to 23 60 log units according to the method and the background electrolyte Recently Desreux et a1 found a lower value of log /3 = 22 1 from kinetics measure- ments 21 Thus the GdDOTA complex would not be more stable than its DTPA homologue However the correlation between log /3 of polyaminopolycarboxylate complexes and XpK of the ligands still includes this new value and the stability of the DOTA complexes and related complexes depends on the electrostatic nature of the bonds with the metal ion rather than on steric factors However the presence of the macrocyclic structure leads to very rigid complexes in which the cation is well encapsulated They are thus prevented from dissociation in aqueous solution even in acidic conditions as compared with their open homologues In that respect GdDOTA -dissociates much more slowly than GdDTPA even in acidic medium (half- life of 85 days at pH 2 and > 200 days at pH 5) Moreover the high stability and kinetic inertness of macrocyclic polycarboxy- lic complexes improve their tolerance and prevent any compe- tition with endogenous cations such as Zn2 + ,Ca2+ ,or Cu2 + Thus they are very good candidates for water-soluble shift reagents and some of them are already used as contrast agents for NMR imaging (MRI) Compounds featuring phosphorus-containing ionizable groups like phosphonic acids (NOTP) and phosphinic acid (DOTEP),23 have been designed to monitor intracellular cation concentrations by lP-NMR spectroscopy (Figure 7) As the phosphinate RPO group is more electron-withdrawing than CO and PO ,the oxygen donor basicity falling in the same order all the donor atoms of DOTEP are less basic than those of DOTA or DOTP This results in a decrease of the metal-ligand interactions and in lower stability constants The fact that the stability of these complexes is still controlled by electrostatic interactions rather than by steric factors is further illustrated by the agreement between the experimental log ,f3 value for the Gd complex and the value predicted from the linear relationship between log ,f3 and the sum of the Iigand's pK (see above) However at physiological pH (7 4) the difference in stability between DOTEP and DOTA complexes is reduced because of the decreased competition between protons and Ln3 + for bind- ing DOTEP at this pH The LnDOTEP complexes are slow to form but dissociate slowly giving these systems similar kinetic advantages for in vzvo uses as LnDOTA On the other hand the methylenephosphonate monoethylester NOTPME22 forms considerably less stable complexes than either NOTP or NOTA 4 Lanthanide Complexes with Calixarenes Calixarenes are a new class of macrocyclic receptors formed by para-substituted phenolic units linked ortho by methylene groups The binding abilities of calixarenes towards metal ions were thought to be due to the existence of a template effect during some syntheses but were first demonstrated by transport experiments by Izatt et a1 24 These molecules transport alkali and alkaline-earth picrates from an aqueous source phase to an aqueous receiving phase through a bulk haloformic membrane provided that the source phase is basic enough to ensure partial deprotonation of the ligands and formation of neutral com- plexes Because of their great insolubility in most usual solvents the study of the complexing power of calixarenes in solution has been limited However further substitution of the phenolic hydrogens of calixarenes by various types of ligating groups led to more soluble compounds the structures of which seemed to be better pre-organized for hosting a metallic cation For instance the cone conformation preferentially adopted by the tetrameric oligomers defines a hydrophilic cavity as an exten- sion of the lipophilic one formed by the aromatic rings where a cation can be encaged In the past few years the complexing power towards alkalis and alkaline-earths of a great variety of calixaryl derivatives featuring ester amide ketone and carbo- xylic acid functions has been established on the basis of extrac- tion and transport experiments as well as stability constant determinations High complexation and extraction selectivi- ties have been demonstrated which were based mainly on the fit between the cation and the cavity sizes Parent phenolic calixar- enes and their chemically modified derivatives are also expected to bind lanthanide cations This was confirmed for the first time by Harrowfield et a1 through the isolation from DMF of a binuclear complex of europium with the p-t-butylcalix[8] arene 26 Its X-ray structure shows that the complexed ligand is SOLUTION CHEMISTRY OF LANTHANIDE MACROCYCLIC COMPLEXES-F ARNAUD-NEU -t + Ow0 Ow0OHOH NR (1 ) Phenols (2) amides (3) acids (4) dioxa-tetraester (5) trioxa-triacid (6) diacid-diester (7)monoacid-triester (8)diphenol-monoethylester-monoacid OH (OON hOO)uu (9) Schiff-base bridges calixarenes a R = (CH2)3 b R = (CH2)4 R=R (10) aza and oxa bridged calixarenes a R=CH2CH2 b R = CH2CH2NHCH2CH2 c R = CH2CH2(NHCH2CH2)2 d R = CH2CH2CH2NHCH2CH2CH2 e R = CH2CH2CH2NHCH2CH2NHCH2CH2CH2 f R = CH~CHZ(OCH~CH~)~ Figure 8 Chemical structures of some calixarenes and their chemically modified derivatives hexadeprotonated and can be considered as a ditopic receptor Each europium is coordinated to two solvent molecules and three phenolic groups in addition both cations are bridged by two phenoxide oxygen donor atoms of the macrocycle and by one molecule of DMF Complexes with other lanthanides and other oligomers were further obtained and characterized by X-ray crystallography 27 As the study of the complexation of lanthanides by calixarenes in solution has only just started we have tried in the following sections to report exhaustively on this topic The basic structures of calixarenes which have been studied in this frame are represented in Figure 8 4.1 Phenolic Calixarenes The complexing abilities of the p-t-butylcalix[8]arene (1 n = 8) in solution were also established by Harrowfield 26 This study was carried out in N,N-dimethylformamide by UV-visible absorption spectrophotometry after deprotonation of the ligands using Et,N in excess and it showed the simultaneous formation of a binuclear and a mononuclear complex of euro- pium for which the stability constants were 4 10SM-' and z 103M-l respectively Two other oligomers (R = iso-propyl n = 4,8)were later tested towards Pr3 +,Eu3+,and Yb3 + in the same conditions 28 The results confirmed the presence of the dinuclear and mononuclear complexes with the octamer whereas the only 1 1 species was found with the tetramer It must be noted that the degree of deprotonation of the ligands is unknown in either study and may be different in the free and complexed forms The complexes with the octamer are slightly more stable than their homologues with the tetramer With a given ligand there is no significant variation of stability on going from Pr3 + to Eu3+ and Yb3 +,and hence no selectivity of either the octamer or the tetramer within the lanthanides as far as the cations studied are representative of the whole series Sulfonation in the para position of the phenolic calixarenes (1 R = SOi) leads to water-soluble compounds The binding properties of such a tetramer were studied with all the lantha- nides in water by pHmetric measurements 29 This ligand behaves as a di-acid (LH if the electric charges of the sulfonate groups are omitted) as the dissociation of the third and the fourth phenols could not be achieved The overall stability constants of the LnL,0H2- species formed are very high -around 20 log units -and increase slightly along the series The results also indicate that this ligand is a useful colorimetric reagent for cerium(II1) in the presence of other rare-earth cations 4.2 Chemically Modified Calixarenes The introduction of functions containing 'hard' heteroatoms like esters or amides should lead to better complexing agents for lanthanides A 1 1 complex of the p-t-butylcalix[4]arene tetra- diethylamide (2 n = 4 R = Et) has been prepared with EuCl 30 This complex is water-soluble in contrast to the ligand which is sparingly soluble in methanol By analysing its luminescence spectrum it was established that only one molecule of water was coordinated to the complexed cation This indicates an almost complete shielding of the cation from the solvent This contrasts with the europium cryptate of 222 in which three water mole- cules still interact with the complexed cation * The high level of complexation of lanthanides by two p-t-butylcalix[4]arene tetraamides (n= 4 R = Et and R = pyrrolidinyl) has been confirmed by stability constant determinations (Table 1) 32 These compounds are looser and less selective binders than cryptand 221 Although they should provide a greater number of donor sites the corresponding hexaamides (n= 6 R = Et and R = pyrrolidinyl) form less stable complexes with log fl z 4-5 However the results of extraction of lanthanide pic- rates from water to dichloromethane by these calixarene amides do not exactly reflect the trends observed in complexation because (1) Eu3 + is better extracted than Pr3 + and Yb3 + by both tetramers and (11) Eu3 + is better extracted by the hexamers than by the tetramers 32 It can also be seen that substituent variation in the amide podands of calix[4]- calix[5]- and calix[6]arenes has significant effects on the extraction oi europium picrate as seen in Figure 9 which gives %E the percentage cation extracted They can influence the biphasic transfer of the cation according to their electronic steric and lipophilic characters For instance alkyl groups contribute to rather high extraction levels but the effect of an increasing number of carbons in the substituent (from ethyl to pentyl) is different for the pentamers and the hexamers Whereas a regular increase is observed with the pentamers the opposite variation takes place with the hexamers The presence of substituents with multiple bonds like ally1 or propargyl or with a rather bulky and rigid cyclic Table 1 Lanthanide cation binding with chemically modified calixarenes Log /3 (MeOH) Structure No n R R' Pr3+ Eu3+ Yb3+ 2 4 Et 85 87 81 4 6 pyrrol Et 81 80 44 82 6 pyrrol 52 3h 4 207 250 248 6l CH,CO,Bu' CH,CO,H 135 134 144 7 CH,CO,HCH,CO,H CH,CO,EtCH,CO,But 83 88 83 85 93 94 5h 170 159 180 8h 229 23 2 248 9 (CH2)3 48 9 9 (CH,) Ph 41 37 1% B(AN) 4 51 46 34 R = (CH,) (pyrrolidinyl) * simultaneous formation of I I 1 and 1 1 2 complexes simultaneous formation of 1 1 1 complexes tetraamides 0 pentaamides [3 hexaamides 40 20 0 Et Pyrr Pr Bu Pent ally1 Prgyl Bn R Figure 9 Influence of the substituent variation on the percentageextraction of europium picrate by calix[4]- calix[5]- and calix[6]arene amides structure like benzyl reduces drastically the %E values The steric constraints of the latter and the electron withdrawing of the former could be responsible for these low extraction levels The highest extraction by far is obtained with the hexa-pyrrolidi- nylamide derivative With a dioxa tetraester (R' = CH,CO,Et) bearing two 0-CH units next to two opposite bridging methylene (4) the stability of the complexes in acetonitrile decreases from Pr3 + to Eu3+ and Yb3 + (Table 1) 32 It is known from its X-ray structure that this ligand adopts the 1,2-alternate conformation and defines a larger cavity than would a simple tetramer in the cone conformation 25 The size of the cavity becomes less and less likely to be compatible with the cations along the series thus resulting in a drop in log /3 from Eu3 + to Yb3 + The p-t-butylcalix[4]arene tetracarboxylic acid (3 n = 4) has also been investigated as well as some related tetramers with mixed functionalities [(6) (7),and (S)] 32 The trioxa-triacid (5),a trimer possessing three additional oxygen atoms in the macro- cyclic framework was also tested for its ability to form neutral complexes with trivalent cations With all ligands very stable mononuclear unprotected complexes are formed (Table 1) However they are predominent or alone at higher pH only as the corresponding protonated species are present at lower pH The best complexing agents are in decreasing order the tetra- acid the mixed diphenol-monoethylester-monoacid,and the trioxa-triacid In all cases log /3 values increase from Pr3+ to CHEMICAL SOCIETY REVIEWS 1994 Yb3+ However these variations do not exceed 1 log unit except with the dip henol-monoet hy les ter-monoacid and the te traacid for which dlog /3 = I 9 and 4 1 respectively Apart from the trioxa-triacid and the triethylester-monoacid the Eu3 + com-plexes are at least as stable as their Pr3 + homologues Lantha- nide complexes with calixarene-acids are much stronger than those of alkali and alkaline-earth cations for instance dlog ,8 = 15 between the unprotonated complexes of europium and sodium,2s with the tetraacid a result consistent with the electro- static stabilization of the complexes which prevails with anionic ligands In most cases log values are higher than those found with the amide derivatives However although the tetraacid is for instance formally a stronger binder than the p-t-butylcalix [4]arene tetradiethylamide it would be less effective if used in acidic conditions as its complexation power is pH dependent The ability of p-t-butylcalix[4]arene tetracarboxylic acid to bind lanthanides has also been established by extraction experi- ments from water to toluene or chloroform 33 The extraction constants of the 1 2 metal ligand complexes for the water- toluene system follow the order Eu > Nd > Yb > Er > La which is consistent with the trends found with this ligand in complexation and in the same order found for stability con- stants of 1 1 complexes with simple carboxylic acids Upon addition of excess Na+ both extractibility and selectivity increase Lanthanides are better extracted from water to chloro- form by the p-t-butylcalix[6]arene hexaacid according to the same cation exchange mechanism as 12 metal ligand com-plexes The order of extractibility is Nd Eu > La > Er > Yb In the presence of excess Na+ in the aqueous phase this order did not change although the extractibility decreased and the lantha- nides were extracted as 1 1 complexes Recent results concerning extraction and transport experiments through bulk liquid mem- branes of a dimethoxy-diacid calix[4]arene show a great influence of the cation of the background electrolyte 34 In the presence of a quaternary ammonium cation no cation transfer is observed whereas in the presence of M = Na or K +,Ln3++ + are co-extracted with the alkali ions as LnML complexes This stoichiometry has been confirmed by fast-atom bombardment mass spectrometry (FAB-MS) The weak selectivity observed follows the order La < Y 5 Er Some bridged calixarenes in the cone conformation in which two opposite phenolic oxygen atoms are connected via bridges containing different kinds of donor sites should contribute to the high coordination numbers required by lanthanides better than the simple p-t-butylcalix[4]arene Existence of 1 1 com-plexes of europium in methanol with three Schiff-base bridged calixarenes (9a+) was detected Their stability decreases for the longer bridged (9b) and also for the more rigid aromatic ligand (9c) 35 Other bridged calixarenes containing poly(oxyethy1ene) and poly(azaethy1ene) links (lo) form weak complexes with Gd3 + and the rare-earths Y3 + and Sc3 +,as demonstrated by a FAB- MS study 36 The aza-derivative (IOa) is the best complexing agent for Sc3 + ,whereas the oxygen bridged calix[4]arene (100 is +the best for Gd3 and Y3 + The effect of methylation of the free OH groups leads to weaker complexes For trivalent cations no clear relationships between the ion radius and the size of the cavity is observed as the achievement of high coordination numbers is also an additional important factor 5 Conclusion In this survey we have tried to point out the more important factors of the complexation of lanthanide cations by macro- cyclic receptors by selecting some examples from amongst the numerous studies involving cryptands and neutral or ionizable coronands These macrocycles display many features (large number of donor sites presence of oxygen and nitrogen atoms or a combination of both types of heteroatoms ) favourable to a strong binding of lanthanide cations which may find numer- ous applications However in contrast to alkali and alkaline- earth cations high complexation selectivity amongst lanthanide SOLUTION CHEMISTRY OF LANTHANIDE MACROCYCLIC COMPLEXES-F ARNAUD-NEU 24 1 cations is difficult to achieve as the variation of the stability constants along the series results from an interplay of often antagonistic factors (size solvation and charge effects ligand flexibility) Data on the complexation of lanthanides by calixar- enes are still scarce but a great number of derivatives have not yet been tested with these cations The few results obtained so far indicate a generally high stability for lanthanide complexes of calixarenes which associated with their recently reported strong fl~orescence,~’ should make them useful for instance as luminescent probes and tags for the analysis of biological materials Acknowledgements The author thanks the CEC for supporting part of the work done within the framework of the Treatment of the Nuclear Waste Programme and of a CEC Twining Project The contributions of co-authors named on the cited references are gratefully acknowledged in particular Professors M J Schwing-Weill (E H I C S Strasbourg) M A McKervey (Queen’s University Belfast) M Gross (ULP Strasbourg) and Dr A F Danil de Namor (Surrey University) 6 References 1 (a)J M Lehn Structure and Bonding 1973 16 1 (b) J M Lehn J Incl Phenom 1988,6,351 2 (a) C J Pedersen J Am Chem Soc 1967 89 2495 (b) C J Pedersen J Incl Phenom 1988,6 341 3 (a)D J Cram Angew Chem Znt Ed Engl 1986,25,1039 (b)D J Cram J Incl Phenom 1988,6 397 4 ‘Calixarenes A Versatile Class of Macrocyclic Compounds’ J Vicens and V Bohmer ed Kluwer Academic Press Dordrecht 1990 5 ‘Lanthanide Probes in Life Chemical and Earth Sciences Theory and Practice’ ed J C G Bunzli and G R Choppin Elsevier Science Publishers New York 1989 6 J C G Bunzli ‘Handbook on the Physics and Chemistry of the Rare-Earths’ Chapter 60 ed K A Gschneider and L Eyring Elsevier Science Publishers Amsterdam 1987 7 W Szczepaniak B Juskowiak and W Ciszewska Znorg Chim Acta 1988 147 261 8 G L Rothermel Jr L Miao A L Hill and S C Jackels Inorg Chem 1992,31,4854 9 L E Loufouilou F Arnaud-Neu and M J Schwing-Weill J Chem SOC Dalton Trans 1986,2629 10 I Marolleau J P Gisselbrecht M Gross F Arnaud-Neu and M J Schwing-Weill J Chem Soc Dalton Trans 1989,367 11 I Marolleau J P Gisselbrecht M Gross F Arnaud-Neu and M J Schwing-Weill J Chem SOC Dalton Trans 1990 1285 12 J C G Bunzli A E Merbach and R M Nielson Znorg Chim Acta 1987 139 151 13 A F Danil de Namor M C Ritt M J Schwing-Weill and F Arnaud-Neu J Chem Soc Faraday Trans 1990,86,89 14 L E Loufouilou and J P Gisselbrecht Can J Chem 1988 66 2172 15 M F Loncin J F Desreux and E Merciny Znorg Chem ,1986,25 2646 16 W P Cacheris S K Nickle and A D Sherry Znorg Chem ,1987 26,958 17 (a) R Delgado Y Sun R J Ramunas J Motekaitis and A E Martell Znorg Chem ,1993,32,3320 (b)D Chen P J Squattrito A E Martell and A Clearfield Znorg Chem 1990,29,4366 18 S I Kang R S Ranganathan J E Emswiler K Kumar J Z Gougoutas M F Malley and M F Tweedle Znorg Chem 1993 32,2912 19 J F Carvalho S H Kim and C A Chang Znorg Chem 1992,31 4065 20 S Aime M Botta G Ermondi F Fedelli and F Uggeri Inorg Chem ,1992,31 1100 21 X Wang T Jin V Comblin A Lopez-Mut E Loncin and J F Desreux Znorg Chem 1992,31 1095 22 I Lazar,R Ramassamy,E Brucher,C F G C Geraldes,andA D Sherry Inorg Chim Acta 1992 195,89 23 I Lazar A D Sherry R Ramassamy E Brucher and R Kiraly Znorg Chem 1991,30 5016 24 (a)R M Izatt J D Lamb R THawkins P D Brown S R Izatt and J J Christensen J Am Chem Soc 1983,105 1782 (b) S R Izatt R 7 Hawkins J J Christensen and R M Izatt J Am Chem SOC,1985,107,63 25 (a) F Arnaud-Neu E M Collins M Deasy G Ferguson S J Harris,B Kaitner,A J Lough,M A McKervey,E Marques,B L Ruhl M J Schwing-Weill and E M Seward J Am Chem SOC 1989,111,8681 (b)F Arnaud-Neu M J Schwing-Weill K Ziat S Cremin S J Harris and M A McKervey New J Chem 1991,15 33 (c) F Arnaud-Neu S Cremin D Cunningham S J Harris P McArdle M A McKervey M McManus M J Schwing-Weill and K Ziat J Incl Phenom ,1991,10,329 (d)M J Schwing-Weill F Arnaud-Neu and M A McKervey J Phys Org Chem ,1992,5 496 (e) G Barrett M A McKervey J F Malone A Walker F Arnaud-Neu L Guerra M J Schwing-Weill C D Gutsche and D R Stewart J Chem Soc Perkin Trans 2 1993 1475 (f) F Arnaud-Neu G Barrett S J Harris M Owens M A McKervey M J Schwing-Weill and P Schwinte Inorg Chem 1993 32 26 (a)B M Furphy,J M Harrowfield,D L Kepert,B W Skelton,A H White and F R Wilner Znorg Chem 1987,26,4231 (b)J M Harrowfield 13th International Symposium on Macrocyclic Chemistry Hamburg (Germany) September 4-8 1988 27 (a) L M Engelhardt B M Furphy J M Harrowfield D L Kepert A H White and F R Wilner Aust J Chem 1988 41 1465 (6) B M Furphy J M Harrowfield M I Ogden B W Skelton A H White and F R Wilner J Chem Soc Dalton Trans 1989 2217 (c) J M Harrowfield M I Ogden and A H White Aust J Chem 1991,44 1237 (d)J M Harrowfield M I Ogden W R Richmond and A H White J Chem SOC Dalton Trans 1991,2153 (e)Z Asfari J M Harrowfield M I Ogden J Vicens and A H White Angew Chem Znt Ed Engl ,1991,30,854 28 F Arnaud R Abidi M J Schwing M Bourakhouadar R Perrin and J Vicens Societe Suisse de Chimie Assemblee d’Automne Berne 1989 29 I Yoshida N Yamamoto F Sagara K Ueno D Ishii and S Shinkai Chem Lett 1991 2105 30 N Sabbatini M Guardigli A Mecati V Balzani R Ungaro E Ghidini A Casnati A Pochini J Chem Soc Chem Commun 1990,878 31 N Sabbatini S Perathoner V Balzani B Alpha and J M Lehn in ‘Supramolecular Photochemistry’ ed V Balzani D Reidel Pub- lisher Dordrecht 1987 p 187 32 F Arnaud-Neu S Fanni M A McKervey D Marrs M J Schwing-Weill P Schwinte and K Ziat unpublished results 33 R Ludwig K Inoue and TYamato Solv Extr Zon Exch ,11,31 1 34 M Burgard and J Soedarsono unpublished results 35 R Seangprasertkij Z Asfari F Arnaud and J Vicens J Org Chem 1994,59 1741 36 R Ostaszewski T W Stevens W Verboom and D Reinhoudt Recl Trav Chim Pays-Bas 1991 110,294 31 (a) J C G Bunzli P Froidevaux and J M Harrowfield Inorg Chem ,1993,32,3306 (6) N Sato I Yoshida and S Shinkai Chem Lett 1993 1261

ISSN:0306-0012    

DOI:10.1039/CS9942300235

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Syntheses, Structures, and Properties of Methanofullerenes FranGois Diederich”, Lyle Isaacs, and Douglas Philp Laboratorium fur Organische Chemie, ETH Zentrum, Universitatstrasse 16, CH-8092Zurich, Switzerland 1 Historical Background The recent surge of interest1 in the chemistry and properties of carbon-rich molecules has been driven largely by the discovery2 and the isolation of macroscropic quantities3 of the fullerenes -all-carbon molecules which have the form of hollow, closed nets composed of 12 pentagons and n hexagons and the composition C,, + 2n. The most readily available member of the fullerene family, c60 (n = 20), possesses icosahedral symmetry, and, as a consequence, all sixty carbon atoms in the sphere are equivalent. However as a result of the presence of both five- and six- membered rings within the structure of c60, there are two types of bonds -namely bonds at the junction between two six- membered rings ([6,6]-bonds, Figure la), and bonds at the junction between a five- and a six-membered ring ([6,5]-bonds, Figure 1b).The electronic structure4 of the fullerenes is such that bonds at [6,6]-ring junctions (Figure la) have much double- bond character, while bonds at [6,5]-ring junctions (Figure 1b) are essentially single bonds. This arrangement results in c60 having a strongly bond-alternated structure which can best be described as a spherical tessellation of [5]radialene and 1,3,5- cyclohexatriene sub-units (Figure lc). The bond alternation observed in c60 governs its reactivity.Far from being ‘super- aromatic’, c60 behaves as an electron-deficient alkene in its solution-phase reactivity. This fact was exploited by several groups to prepare5 well-characterized organometallic deriva- tives of c60. The X-ray crystal structures of [y2-C60Pt(PPh3)2]5b and [y2-C601r(CO)(CI)(PPh3)2]5c(Figure 2) which incorporate metallacyclopropane sub-units were particularly intriguing to organic chemists. Formally, replacement of the metal atom in the metallacyclopropane by a carbon atom would give an organic derivative of c60, a methanofullerene, whose synthesis could be envisaged as the addition of carbenes or their equiva- Franqois Diederich was born in Ettelbruck (Luxemburg) in 1952. A.fter diploma and doctoral studies in chemistry with Heinz A.Staab at the University of Heidelberg, he moved in 1979 to the University of California at Los Angeles (UCLA) for postdoctoral work with Orville L. Chapman. He returned to Heidelberg in 1981, where he pursued research in host-guest chemistry at the Max- Planck-Institute for medical research and received his Habili- tation in 1985. He then moved back to UCLA as an acting associate professor and was promoted to full professor in 1989. In 1992 he came back to Europe to accept a chair in the Laboratory Franqois Diederich Lyle Isaacs lents to the c60 sphere. This proposal was further strengthened by the observation made by Diederich et al. in early 1991 that fullerene C7 0 reacts during the Kratschmer-Huffman produc- tion process with traces of oxygen to form an 0-bridged fullerene oxide C700,6 a finding which was soon after followed by the preparation of C6,0 by different method^.^ The presence of two different kinds of bonds in c60 suggests that, in principle, two isomers of the methanofullerenes may be formed by such addition reactions -namely the [6,5]-bridged and [6,6]-bridged isomers (Figure 3).Further, by analogy with the methanoannu- lenes,* the possibility of valence isomerism leads to a total of four possible isomers (Figure 3) depending on whether the transannular bond of the methanoannulene sub-unit in the fullerence sphere is ‘open’ or ‘closed’. Since the first reported synthesisg of a methanofullerene by Wudl and co-workers in late 1991, research in this field has expanded rapidly in many different directions.In this Review we will attempt not only to describe the syntheses and electronic structures of the methanofullerenes but also to highlight their potential applications in materials science and as potential therapeutic agents. 2 The Synthesis of Methanofullerenes The synthetic methods currently utilized to produce methano- fullerenes may be conveniently divided into three categories. The most popular route for methanofullerene synthesis is the thermal addition of diazo compounds to c60.The addition of free carbenes to c60 has also been demonstrated to give rise to methanofullerenes, as well as reactions which proceed by an addition/elimination mechanism.These methods will be dis- cussed in the following sections. for Organic Chemistry at the Eidgenossische Technische Hochs- chule (ETH) in Zurich. Lyle Isaacs was born in New York City in 1969. After receiving a B.S. degree from the University of Chicago in 1991, he joined the group of Franqois Diederich at the University of California at Los Angeles on a Department of Defense Graduate Fellowship. He received an M.S. degree in 1992 from UCLA before moving to Zurich to complete his doctoral studies with Franqois Diederich at the ETH. Douglas Philp was born in Glasgow, Scotland in 1967. He received his B.Sc. degree from the University of Aberdeen in 1989 and his Ph.D. degree from the University of Birmingham in 1993. He spent one year at the ETH, Zurich as a Royal Society European Science Ex- change Programme Postdoc-toral Fellow before taking up his current post as Lecturer in Organic Chemistry at the Uni- versity of Birmingham in April Douglas Philp 1994.243 CHEMICAL SOCIETY REVIEWS, 1994 (a) (b) (a Figure 1 Illustration of (a) a [6,6]-bond; (b) a [6,5]-bond; and (c) the [5]radialene and 1,3,5-cyclohexatnene substructures of C6,,. Figure 2 Stereoview of the X-ray crystal structure of +C,,,Ir(CO)(Cl)(PPh,), 5c [6,6Wpen [6,6]-Closed Figure 3 The four possible isomeric methanofullerenes. The methanoannulene-type subunits are highlighted in red. 2.1 The Addition of Diazo Compounds to Cs0 In late 1991, Wudl and co-workers reportedg that c60 reacts with diphenyldiazomethane in toluene solution at room temper- ature to give a [6,6]-ring bridged adduct.Subsequently, in an Accounts of Chemical Research article, it was shownlO that this reaction tolerates the presence of substituents on the phenyl rings, implying its general applicability to the production of functionalised fullerene derivatives. This account ignited an interesting debate on the structure of methanofullerenes, and the nature and resolution of the fundamental questions raised will be a recurring theme throughout this review. 2.1.1 Scope of the Thermal Addition of Diazo Compounds to C6 0 The scope of the thermal addition reaction of diazo compounds to c6,is quite broad. A complete compilation of the reactions of c60 with diazo compounds is presented in Table 1.Mono- and diphenyldiazomethanes bearing a variety of substituents on the phenyl rings have been successfully It has also proven possible to add diazomethane it~elf.~~J~ In three reports, our group has demonstrated the addition of electron-deficient diazoacetates and diazomalonates to c60.’leZ3 The addition of the diazoacetates is sluggish and requires refluxing in SYNTHESES, STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F. DIEDERICH, L. ISAACS, AND D. PHILP 245 Mixture ofIsomers Table 1 Compilation of the reactions of c60 with diazo compounds R' R2 R=H R=Me R = R = R = R = R = R = Br NMe, Ph0,C OMe OPr' (CH2),NHCOCH, R=H R e H R = C0,Bu' n=O n=l x = (CO) x = (CH,),X = (CH), x= K O?? R=H R R R = NO, n I- R = Et 0 ROk H R = But R = CH,CO,Et 0 R = Et ROX >OR R = But RX = N((CH,),),ORX = NHBn R J H RX = NHCH(Bn)CO,Bn RX = NH(CH,),CH(NH,)CO,Et Not reported Addition to two C,, molecules occurs under formation of dumbbell-type compounds.Yield (YO) Ref. nP 9 nr 10 nr 10 nr 10 nr 10 73 I1 nr 15 38 13 5Ih 14 39h 14 nr 10 nr 12 43 16 42 17 30 16 nr I5 nr 15 nr 10 nr 18 nr 18 nr 18 nr 10 nr 10 44 19 30 21 25 21 32 22 10 23 7 23 24 2&30 24 24 24 toluene for 7 hours -by comparison, the diphenyl diazometh- anes add rapidly at room temperature.The addition of diazoma- lonates can only be accomplished by the use of high tempera- tures and long reaction times. More recently, Skiebe and Hirsch described the reaction of c60 with diazo amide~.~~ In this case, refluxing toluene and a reaction time of 48 hours was required to produce the desired methanofullerenes. 2.1.2 The Initial [I ,3]-Dipolar Cycloaddition ofDiazo Compounds to c6, Wudl and co-workers proposed9 that diphenyldiazomethane adds initially as a [1,3]-dipole to c60, and that the pyrazoline intermediate, which was not isolated, subsequently extrudes nitrogen. When the addition of diazomethane to c60 was conducted in benzene at room temperature,19 a brown, ther- mally unstable, compound was isolated (Scheme 1).This com- pound was assigned the pyrazoline structure (1) on the basis of (i) a singlet in the 'H NMR spectrum at 6.51 ppm, (ii) the presence of 30 carbon resonances in the fullerene region of the 13C NMR spectrum (consistent with the assigned C, symmetry), and (iii) a N=N stretching vibration at 1560 cm- in the infrared spectrum. Thermolysis of (1) in refluxing toluene afforded (2) in quantitative yield. The lH NMR spectrum of (2) showed two doublets centred at 2.87 and 6.35 ppm, and the 13C NMR spectrum displayed 32 resonances between 135 and 150 ppm, in addition to a signal at 38.85 ppm, in agreement with the assigned [6,5]-open structure (2). Taken together, these experiments suggest that the initial step in the addition of diazo compounds to c60 is indeed a [1,3]-dipolar cycloaddition followed by thermal loss of nitrogen.However, all of the phenyl- and dip henyl-su bsti tuted methanofullerenes isolated previously O were bridged at the [6,6]-ring junction, and not at the [6,5]-ring junction. Thus, it was unclear whether the results obtained in the diazomethane addition, which yielded only a very small amount of the [6,6]-closed isomer, applied to the reactions of C,, with CHEMICAL SOCIETY REVIEWS, 1994 A --N2 I-hv -N2 Scheme 1 K""N2 66.79 PPm \r H'j+O,Et N2 + + Thermal Equilibration Scheme 2 di zo compound in general. In all of the previous cases, the solution was produced whose lH NMRspectrum showed only a initial product of the reaction was a mixture of at least two single set of resonances.This pure compound corresponded to isomers. Furthermore, it was possible to convert this isomeric one of the two minor components in the isomeric mixture and mixture thermally into a single [6,6]-ring bridged compound was obviously thermodynamically more stable than the other which had high symmetry (C2J -readily identified by the two isomers. This compound was assigned the [6,6]-closed simplicity of its IH and 13CNMRspectra. These two generaliza- structure (6) based on its 3CNMRspectrum and the lJCHvalue tions apparently did not apply to the diazomethane adduct (2). A for the methine carbon which clearly supported the presence of a connection between these seemingly unrelated facts was needed.cyclopropane ring (see Section 3 below). However, the 13C spectrum and JCH values of the isomeric mixture established that the two other isomers, which were produced under kinetic 2.I .3 Kinetic versus Thermodynamic Products control, had the [6,5]-open structures (4)and (5). The mechan- In the middle of 1992, our group observedz1 that the addition of ism of diazo additions to c60 (Scheme 2) was now reasonably alkyl diazoacetates to c60in refluxing toluene for 7 h (Scheme 2) secure. The initial reaction affords, as was shown by Wudl, a gave rise to three isomeric products (as determined by 'H NMR pyrazoline which is thermally unstable under the experimen- spectroscopic analysis) under kinetic control. Interestingly, one tal conditions and loses nitrogen rapidly to give [6,5]-open of the isomers predominated by a factor of three over the other methanofullerenes, together with some [6,6]-closed isomer.two. When this isomeric mixture, which like c60 solutions was Thermal equilibration of the isomeric mixture leads to the purple, was heated in refluxing toluene for 24 hours, a wine-red thermodynamically most stable [6,6]-closed methanofullerene. SYNTHESES. STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F. DIEDERICH, L. ISAACS, AND D. PHILP 247 Nearly identical results were subsequently reported' by Wudl and co-workers for the addition of bis(4-iso-propyloxypheny1)-diazomethane, (p-methoxyphenyl)diazomethane, and (p-meth- oxypheny1)diazoethane to c60. Similar to the results in the addition of alkyl diazoacetates, they found that the [6,5]-open diastereoisomer with the bulkier substituent located above the five-membered ring predominated in the product mixture formed under kinetic control.Related studies by Vogtle and co- workers,16 and those of Skiebe and HirschZ4 demonstrated the same phenomenon, suggesting that this diastereoselectivity in the formation of the kinetic [6,5]-open isomers may be a general phenomenon. Structures of the two [6,5]-open diastereoisomers, in all cases, were assigned based on the 'H NMR spectra which showed a strong downfield shift of the resonance for the methano bridge proton located above a pentagon as compared to the resonance of the proton above a hexagon (Scheme 2). In the case of the parent methanofullerene C61H2, it has not proved possible thermally to rearrange the [6,5]-open isomer (2) to the [6,6]-closed compounds (3).The latter could, however, be synthesized by Smith et al. by photolysis of pyrazoline (1).20 This reaction afforded both (2) and (3)in a 4:3 ratio, and separation of the mixture by HPLC on a C,, reverse-phase column gave (3),whose structure was convincingly elucidated by 'H and 13C NMR spectroscopy. The fact that (3)cannot be isomerized to (2) under either thermal or photolytic conditions implies that there is no easily accessible, low-energy pathway for their interconversion. Thermal isomerization of the diazoamide adducts of c60, reported by Skiebe and Hirs~h,~~ is also impossible, leading to decomposition of the [6,5]-open isomers, rather than equili- bration to the corresponding [6,6]-closed isomers. 2.2 The Addition of Carbenes to Cs0 The reaction of c60 with carbenes is less complicated than the addition of diazo compounds, since singlet carbenes seem to add exclusively to the [6,6]-ringjunction of C,, in one step, as would be expected if C,, behaves as an electron-deficient alkene.It appears that any method which generates carbenes is suitable for application in the synthesis of methanofullerenes. Table 2 Synthesis of methanofullerenes by carbene addition Precursor R' (7) R = BnU Rt0&RtO&,-.. (8) R = Piv" OR CI,CCOONa (1 1) c1 Me0 OMe NX0 (1 3) OMe 0(15)(16)R=Ph R = H 2.2.1 Diazirines The two reported examples of the addition of diazirines to c60 both come from the work of Vasella et al.25who, as a result of their interest in the potential biological activity of amphiphilic water-soluble fullerene derivatives, prepared fullerene-sugar conjugates.Thus, the reaction of the protected sugar derivatives (7) and (8) with c60 in toluene (Table 2) afforded the enantio- merically pure methanofullerenes (9) and (10) in 55 and 54% yield, respectively. The isolated compounds consisted of single structural isomers each showing two fullerene resonances between 77 to 80 ppm in the sp3-C-atom region of their respective 3CNMR spectra. These compounds were therefore assigned a [6,6]-closed structure. This finding is consistent with the intermediacy of the nucleophilic glycosylidene carbene, since a mechanism involving the isomeric diazo compound would presumably give rise to both [6,5]-open and [6,6]-closed methanofullerenes.2.2.2 Pyrolysis of a-Halo Carboxylates The addition of dichlorocarbene to c60 has been described by Nogami and co-workers.26 The pyrolysis (Table 2) of sodium trichloroacetate (11) in a mixture of benzene and diglyme generates dichlorocarbene which then adds, as expected on mechanistic grounds, directly to the [6,6]-ring junction giving the [6,6]-closed methanofullerene (12). The structure of the product was confirmed by the observation of a molecular ion at m/z 803amu in the negative ion fast atom bombardment (FAB) mass spectrum, and the presence of the 18 resonances in the 3C NMR spectrum expected for a C,,-symmetrical product.Six- teen of these resonances were observed between 138 and 146 ppm. A further two resonances are located at 64.I and 80.1 ppm and are ascribed to the methano and bridgehead C-atoms, respectively. 2.2.3 Thermolysis of Oxadiazoles The reaction of c60 with oxadiazole (13) in refluxing toluene, reportedz2 by Isaacs and Diederich, afforded [6,6]-closed (14) as the only one detectable monoaddition product, arising from the R2 Yield (YO) Ref. (9) R = Bn (10) R = Piv 55 54 25 25 c1 25 26 OMe 32 22 (17) R = H (18) R = Ph 44 10 29 29 H (19) CzCTIPS C=CTIPS 32 30N,N. Ts (20) C=CTMS CECTMS 24 30 R'% (21) C=CTIPS CECTMS 28 31 0 Bn = PhCH,, Piv = (CH,),CCO.addition of dimethoxycarbene to a [6,6]-ring junction. This compound was synthesized in order to assess the effect of methano bridge substituents on the position of the theoretically possible [6,6]-closed e [6,6]-open valence isomeric equilibrium. Such effects are well established in 1,6-methano[ lO]annulenes, where an electron-withdrawing substituent like the cyano group at the bridging C-atom favours a closed transannular bond whereas electron-donating substituents like the methyl group favour an open transannular bond or an electronic structure somewhere between the two extreme valence isomer^.^ 7328 The 3C NMR spectrum of (14) showed no unusual chemical shifts, thereby strongly supporting exclusive formation of a [6,6]-closed structure.This finding indicates that the fullerene itself and not the methano bridge substituents determine the preferred valence isomer. A simple rationale for this observation will be presented in Section 3. 2.2.4 Thermolysis of Cyclopropenone Acetals Nakamura and co-workers have described29 the reaction of C6, with the cyclopropenone acetals (1 5) and (16). In this case, the production of the carbene intermediate occurs by isomerization, generating vinyl carbenes which then undergo addition to c60. Hydrolysis of the intermediate ketene acetals gives the observed products (17) and (18) in 44 and 10% yield, respectively. Interestingly, when R = Et, a [3 + 21-cycloaddition product is formed, indicating that this method of generating carbenes may not be the one of choice if methanofullerenes are the desired products.2.2.5 Thermolysis of Tosylhydrazone Lithium Salts Two reports by Rubin and co-w~rkers~~ and Diederich and co- worker~~~demonstrated the use of diethynyl carbenes in the production of diethynylmethanofullerenes. Thermolysis of the lithium salts of tosylhydrazones (19), (20), and (21), derived from the corresponding 1,4-pentadiyne-3-0nes (Table 2), produces diethynylmethylene carbenes which add readily to C,, to give the symmetrically and unsymmetrically silyl-protected (TMS = trimethylsilyl, TIPS = triisopropylsilyl) [6,6]-closed diethynyl derivatives (22)--(24). 2.3 Reactions Occurring by Addition/Elimination Mechanisms c60 is capable of reacting with nucleophiles of all types, and the intermediate anions can be trapped with suitable electro- philes.lb~c~lo that c60 reacts with several Bingel demon~trated~~ stabilized a-halocarbanions (Table 3) to give the methanofuller- enes (25)--(28). This reaction is formulated as an addition of the R$Rz X t CHEMICAL SOCIETY REVIEWS, 1994 stabilized a-halocarbanion to c60, followed by intramolecular displacement of halide by the anionic centre generated on the fullerene sphere. The reaction is fast, clean, and proceeds in fair to good yield. Furthermore, it gives only [6,6]-closed methano- fullerenes, which obviates the need for tedious and time consum- ing thermal equilibration. By this method, Bingel prepared the first well-characterized C,,-derived methanofullerene.Addition was expected to occur to one of the two [6,6]-bonds in the regions of highest curvature and therefore highest reactivity on the C70 sphere -not unexpectedly only the one closest to the polar cap reacted. Conversion of C, ,with 1,8-diazabicyclo[5.4.O]-undec-7-ene (DBU) and bromodiethylmalonate in toluene at room temperature afforded (29) in 60% yield, whereas (30) was not detected in the reaction mixture. EtO*C, L’C02Et This methodology was extended (Scheme 3) by Hirsch et al. to produce well-characterized bis- and tris-adducts of C6,.33 They found that, upon addition of bromodiethylmalonate to (25)in the presence of NaH as a base, a mixture of bisadducts with [6,6]- closed structures was produced.Of the eight theoretically poss- ible regioisomers, seven were isolated by HPLC from the reaction mixture -the formation of the remaining isomer, the so-called ‘cis-1’isomer, is impossible on steric grounds. Further- more, repeating the same reaction (Scheme 3) on purified (31) and (32) gave the tris-adducts (33) and (34) in close to 40% yield. These compounds were isolated by HPLC, and their structures were assigned on the basis of their 13C NMR spectra. Lastly, in a recent report3 from this laboratory, it was shown that TMS-protected 3-bromopenta- 1,4-diyne reacts (Table 3) -xo c Table 3 Methanofullerenes by addition/elimination mechanisms R’ RZ X Yield (%) Ref. CO, Et COCH, COPh COPh CSTMS C0,Et C0,Me H Ph CSTMS Br C1 Br C1 Br (25) (26) (27) (28) (22) 45 27 21 25 55 32 32 32 32 31 SYNTHESES, STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F.DIEDERICH, L. ISAACS, AND D. PHILP 249 Five other Bis Adducts I Seven Bis Adducts Separated by HPLC I Eto2c’ ’CO2Et Tris Adducts Separated by HPLC Et02C/ Scheme 3 with c60 in the presence of DBU to give diethynyl methanoful- lerene (22). The addition could proceed either via direct deproto- nation of the TMS-protected 3-bromopenta-l,4-diyne by DBU or, alternatively, by nucleophilic displacement of the bromide with DBU to give a more acidic salt, which could be deproto- nated to form an ylid, which could attack C60. 3 The Electronic Structure of the Methanofullerenes The electronic structure of c60 is best described4 as a fusion of [5]radialene and 1,3,5-~yclohexatriene substructures (Figure l), and this strong, yet poorly understood propensity of c60 to avoid placing double bonds in five-membered rings results in poor electronic delocalization and gives c60 its alkene-like properties.Bridging can occur at both the [6,5]-or the [6,6]-ring junctions, and the possibility of valence isomerization poten- tially gives rise to four isomeric methanofullerenes (Figure 3). The chemistry of methanoannuIenes8 has shown that the substi- tuents at the bridging methylene carb~n~’.~~ play an important role in the determination of the valence isomeric equilibrium between the r-homoaromatic bond-open form and the a-homoaromatic bond-closed form; thus the electronic structure and valence isomeric preferences of the methanofullerenes were of great interest.In their initial report9 on their synthesis of methanofullerenes, Wudl and co-workers assigned a [6,6]-open structure to (35) on the basis of its ‘H NMR spectrum and the similarity of its UV/ VIS spectrum and cyclic voltammogram to the corresponding data of C,, itself; however, the compound was too insoluble to record a 13CNMR spectrum. Subsequently, a crystal structure of (36) was reportedlo in which the transannular bond length at the [6,6]-bridge was found to be 1.84 A, which, although shorter than an open transannular bond in methanoannulenes8, is clearly much longer than the normal sp3-sp3 bond length of 1.54 A.This result must however, be treated with caution since the crystals were of poor quality and probably contained more than one isomer. In addition, Wudl also determined the JCH coup-ling constant at the bridging carbon for (37).1° The measured value of 140 Hz is almost exactly that found in 11-phenyl-1,6- methano[ 101annulene with an open transannular bond, and thus was considered clear evidence for a [6,6]-open structure. Caution should be exercised, however, since, without lH and 13C NMR spectra of all three isomers, it is impossible to distinguish whether bridging had occurred at the [6,5]-or the [6,6]-ring junctions of the fullerene sphere. Despite these reservations the matter seemed settled in favour of the [6,6]-open structure, at this point.However, the synthesis25 of fullerene sugars by (35)R’ = R2 = Ph (36)R’ = R2 = p-BrC6H4 (37)R’ = H, R2 = Ph (38) R’ =R2=C021S~ Vasella et al. raised new questions. Compounds (9) and (10) (R = Bn or Piv) were soluble enough to permit the recording of high quality 13C NMR spectra, which clearly showed peaks corresponding to fullerene resonances in the 70 to 80 ppm region providing evidence for a [6,6]-closed structure. The synthesis of (6), (14), and (38) by Diederich and ~o-workers~~-~~ helped to clarify the situation further. Despite the widely different proper- ties of the substituents at the bridging carbon atom in these three compounds, 13C NMR data for the bridgehead fullerene reso- nances and the JcH coupling constant at the bridging C-atom clearly demonstrated that all three compounds had [6,6]-closed structures.All three compounds displayed fullerene resonances in the 70 to 85 ppm region, corresponding to the sp3-hybridized fullerene C-atoms in cyclopropane rings, and, in addition, the 'JcH coupling constant at the bridging C-atom in (6) was found to be around 165 Hz, typical of a cyclopropane ring. Close examination of the isomeric mixture of (4), (9,and (6) produced under kinetic control during the synthesis of (6) revealed that in both cases, the [6,5]-bridged isomers had an open transannular bond -the JCHcoupling constants at the bridging C-atom in (4) and (5) were found to be around 140-145 Hz, characteristic ofa 1,6-methano[1Olannulene-type structure, and all 32 resonances (C,-symmetry) of the carbon sphere appeared in the fullerene region above 130 ppm.Further support for the assignment of the closed structure to the [6,6]-isomer of methanofullerenes came from the photochemical preparation of the [6,6]-bridged isomer of the parent methanofullerene C61H2 (3) by Smith et ~1The. 3CNMR spectrum of (3) contained 18 signals, 16 between 136 and 150ppm, and more importantly, one at 30 ppm and one at 71 ppm which was assigned to the cyclopropane C-atoms in the fullerene sphere. Further evidence for the [6,6]-closed structure was obtained from the JCH coupling constant at the bridging C- atom, which, at 167 Hz, is clearly characteristic of a cyclopro- pane ring. In contrast, Wudl and co-workers did not observe any resonances in the 70-80 ppm region in the 3CNMRspectrum of the [6,5]-bridged isomer of Ce1HZ (2).19 They measured a JcH coupling constant at the bridging C-atom of around 146 Hz which, together with the 13C NMR data, provided clear evi- dence that the [6,5]-bridged isomer of C61H2 had an open transannular bond.All these results called into question the original assignment of a [6,6]-open structure to (35) and indeed, at this point, (35) and its derivatives were reformulated' as the corresponding [6,6]-closed structures. Thus, all the then-known [6,6]-bridged methanofullerenes had a closed transannular bond. Subsequent work has shown that this statement is almost certainly true for all [6,6]-bridged methanofullerenes, and that the substituent effects at the bridging C-atom which are observed in methanoannulenes are overwhelmed in the methanofullerenes by electronic and structural preferences of the fullerene sphere. In retrospect, it was not until [6,5]- and [6,6]-bridged compounds bearing the same substituents at the bridging C-atom became available, and were fully characterized, that all of the pieces of the puzzle could be fitted together. Today, it is clear that all [6,5]-bridged compounds possess an open transannular bond, and all [6,6]-bridged compounds have a closed transannular bond.Once the valence isomeric preferences of the [6,5]- and [6,6]- isomers of the methanofullerenes had been established experi- mentally, the attention turned to the interrelationship between the two isomers.Several groups had reported' 5,16,19,21,24 that the [6,5]-open isomer was the kinetic product of the addition of diazo compounds to c60 and that, in some cases, it could be isomerized thermally to the [6,6]-closed isomer. Clearly a simple rationale was desirable to explain the relative stabilities of the two experimentally isolated isomers and to link them via a feasible interconversion pathway. Progress in this area has come primarily from computational studies. Raghavachari and Sosa the parent methanofuller- ene C, H, using high-level ab initio methods. These calculations showed that the [6,6]-bridged isomer had a closed transannular bond (the [6,6]-open structure did not exist as a local minimum on the potential energy surface) and was between 3 and 9 kcal CHEMICAL SOCIETY REVIEWS, 1994 mol-more stable than the corresponding [6,5]-bridged isomer, which was found to have an open transannular bond.We carried out a comprehensive computational of five experimentally-known methanoannulenes (14), (25), (35), (38), and the parent C61H2 using the semi-empirical PM3 SCF- MOmethod. In all five cases studied, the [6,6]-bridged isomer was found to have a closed transannular bond, while the [6,6]- open isomer could not be located as a local minimum on the potential energy surface. In constrast, the [6,5]-bridged isomer was found strongly to prefer an open transannular bond. In four of the five cases studied, namely (14), (25), (35), and (38) the [6,6]-closed isomer was found to be significantly more stable thermodynamically than the corresponding [6,5]-open struc- ture, the exception being the parent compound C,,H2.These calculations also indicated the absence of classical substituent effects -the calculated geometries varied very little between compounds-suggesting that it is the electronic structure of the fullerene sphere which is dictating the valence isomeric prefer- ence of the [6,5]- and [6,6]-bridged isomers. Interestingly, we were able to locate shallow local minima corresponding to a [6,5]-closed isomer in all cases except that of C61H2. These structures were around 16 kcal mol-l less stable than the corresponding [6,5]-open isomers.Examination of the calcu- lated geometries and bond orders within the three isomers revealed the electronic basis for the observed trend in the relative stabilities. The [6,5]-closed isomer is forced (Figure 4a) to locate ~ two double bonds within five-membered rings of the fullerene ~ sphere, thus disrupting the strongly preferred [5]radialene type bonding seen in c60. The [6,5]-open isomer (Figure 4b) can avoid this fate, but only at the expense of placing double bonds at the bridgehead atoms, in violation of Bredt's Rule. Evidence for the strain induced in the structure by this compromise is found in the geometry of the formally @-hybridized bridgehead C-atoms-these are pushed out of the plane of the three adjacent C-atoms by almost 0.25 A.The [6,6]-closed isomer (Figure 4c) can both avoid placing double bonds in five-membered rings and violating Bredt's rule by having sp3-hybridized bridgehead atoms, which, although located in a cyclopropane ring, cannot be viewed as particularly unfavourable. This analysis also explains why the [6,6]-open isomer cannot be detected. Not only is it forced to locate three double bonds in five-membered rings (Figure 4d), but also two of these double bonds occur at the bridgeheads, making this isomer the least stable of the four. The location of a [6,5]-closed structure for (14), (25), (39, and (38) but not for C61H2 intrigued us and led us to propose the two-step pathway -valence isomerization followed by 1,5-shift -shown in Figure 5 for the interconversion of the [6,5]-open isomer into the thermodynamically more stable [6,6]-closed isomer.Thermal rearrangement takes place only in the case of substituted derivatives, for which a [6,5]-closed structure is located, whereas the parent methanofullerene c61H,, for which no [6,5]-closed structure was located, does not rearrange. This proposal awaits further experimental investigation. The experimental and computational results detailed above illustrate that the structural chemistry of the methanofullerenes is determined by the propensity of the fullerene sphere to retain the [5]radialene-type bonding found in c60. The preference for this bonding arrangement overwhelms any substituent effects expected on the basis of the presence of a methanoannulene sub- unit within the fullerene sphere.The control of the reactivity and the determination of the relative stability of adducts by the electronic characteristics of the carbon sphere may be a general feature of the chemistry of f~llerenes,~ and the studies described above developed some of the rules involved. 4 Potential Applications of Fullerene Derivatives -Further Functionalization of Met hanof u I lerenes Much speculation has appeared in the literature regarding the possible uses of the fullerenes and their derivatives, and, although no commercial applications of fullerene derivatives SYNTHESES, STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F. DIEDERICH, L. ISAACS, AND D. PHILP 25 1 [6,q-Cl-d [6,51-0pen [6,6]-Closed [6,6I-Open Three Double Bonds in Two Double Bonds in Five-membered Rings Five-membered Rings Bredt's Rule Violation Cyclopropane Ring and Bredt's Rule Violation + 21 kcaVmol + 6 kcaUmol 0 kcaUmol Not a Minimum Energy Structure Figure 4 Unfavourable structural elements (highlighted in red) of the four isomeric methanofullerenes and their relative energies from PM3 calculations.The quoted relative energies are the average over compounds (14), (25), (39, and (38). [6,514mn [6,5I=Closed [6,6]lClosed + 21 Figure 5 The proposed reaction coordinate for the thermal isomerization of [6,5]-open to [6,6]-closed methanofullerenes proceeding via the [6,5]- closed valence isomer. Relative energies are quoted in kcal mol -have emerged, the synthetic elaboration of a methanofullerene core has allowed several groups to explore the utility of fullerene derivatives in both biological and materials applications.4.1 Targeted Applications in Biomedical Science For potential biological applications of fullerene derivatives, it might prove desirable to combine the properties of biomolecules with particular physical properties of the c60 sphere such as efficient sensitization of singlet oxygen for~nation.~~>~~ Thus synthetic approaches to methanofullerene-based amino acids and peptides have received significant attention. Our approachZZ to the synthesis of methanofullerene amino acids focused on the production of the carboxylic acid (39) as a versatile intermediate. The synthesis of (39) (Scheme 4a) was achieved by the addition of (40) to C60,thermal isomerization of the isomeric mixture produced to the single [6,6]-closed isomer, and deprotection using BBr, in benzene.Amide bond forming reactions (Scheme 4a) with amino acids esters under peptide coupling conditions [dicyclohexyl carbodiimide (DCC), 1-hydroxybenzotriazole (HOBT)] gave the methanofullerene amino acids (41) and (42) in 71 and 80% yields, respectively. In the case of the phenylalanine derivative (42), the CD spectrum indicated that the coupling reaction had proceeded without racemization, allowing access to optically pure methanoful- lerene amino acids. Prato et al. utilized a similar approach' in their preparation of a fullerene pentapeptide (Scheme 4b).Reaction of c60 with [4-(t-butoxycarbonyl)phenyl]diazo-methane, after isomer equilibration, gave (43), which was con- verted with excess trifluoromethanesulfonic acid in dioxane into the corresponding acid. After activation with oxalyl chloride in benzene, reaction with an N-deprotected pentapeptide gave (44) in 31% yield. An exciting first step towards fullerene-based agents with possible therapeutic potential was taken by Wudl, Kenyon, and ~o-workers.~~One of the major targets in AIDS therapy is the Human Immunodeficiency Virus (HIV) protease, an aspartyl protease. This interdisciplinary group of scientists recognized that the active site of the HIV protease, which may be described as an open-ended cylinder, has an interior diameter approxi- mately the same size as C60.Since both the fullerene surface and the walls of the enzyme active-site are of pronounced hydro- phobic character, it was reasoned that the protease could be inhibited by suitably designed water-soluble c60 derivatives which would be bound through strong hydrophobic interac- tions.To test this hypothesis, the water-soluble methanoful- lerene (48) was prepared (Scheme 5)starting with the addition of (45) to C6,.13 Thermal equilibration to the [6,6]-closed isomer, followed by hydrolysis of the acetamide groups afforded the diamine (46) as its bis(hydroch1oride) salt. Homologation of (46) CHEMICAL SOCIETY REVIEWS, 1994 0 OH DCC, HOBT, Et3N, PhBr * HC'mH2NTco2Me R (39) (42) R = CH2Ph Hacoomu 1.CF3S03H, Dioxane CO(L-Ala-Aib)2-L-Ala-OMe= 2. Oxalyl Chloride, Benzene 3. H-(L-Ala-Aib),-L-Ala-OMe (43) (44) Scheme 4 1. qpR @N2 (45) R = (CH2),NHCOCH3 \/ 2. CH,COOH / aq. HCI 'I Scheme 5 OMe \ BBr3, edichlorobenzene -0 -25°C CICO(CH2)8COCI PhN02 Scheme 6 with succinic anhydride in pyridine afforded diacid (47)whose and polyurethanes which incorporate methanofullerenes in the bis(sodium) salt (48)was water-soluble to an extent of 1 mg polymer backbone. The reaction of bis(4-methoxypheny1)diazo-ml-'. Biological testing vindicated the prediction of anti-HIV methane (Scheme 6) with C60, followed by isomer equilibration, protease activity-it was found that (48)inhibits the enzyme with gave the [6,6]-closed isomer (49).After cleavage of the methyl a Ki of 5.3 pM.ethers using BBr,, condensation of the diphenol (50) with sebacoyl chloride or hexamethylene diisocyanate/diazabicyclo- octane (DABCO) (Scheme 6)gave the soluble polyester (51) and 4.2 Targeted Applications in Materials Science an insoluble polyurethane (52). The UV/VIS spectrum and The unusual physical and chemical properties of C,, prompted cyclic voltammogram of the polymer were almost identical to several groups to investigate the possibility of preparing those of the monomeric methanofullerene, suggesting that the polymers containing C,, spheres, either as pendant groups or properties of methanofullerene monomers can be transferred incorporated within the polymer backbone itself.unaltered to the polymers. Wudl and co-workers reported' the formation of polyesters 61,61-Diethynyl-l,2-methano[60]fullerene(C65H2), (53), SYNTHESES, STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F DIEDERICH, L ISAACS, AND D PHILP 253 Figure 6 (a) Glaser-Hay coupling of deprotected (56) should allow access to new molecular carbon allotropes consisting of a cyclic acetylenic core peripherally protected by Ce0spheres (b) The same reaction with the diethynylmethanofullerene (53) could afford all-carbon polymers possessing pendant methanofullerene units .R2HvH IR' =R~=TMSI ~~4(53) K2C03 MeOH / THF K2C03, MeOH / THF 20 eq TMSCSH, O2 CuCI, TMEDA CuCI, TMEDA, PhCl PhCI, 02 ITMS, TIPSVTMS Scheme 7 represents a key component In a variety of novel molecular and polymeric carbon allotropes (Figure 6) The recent synthesis of this compound and its silyl-protected analogues (22), (23), and (24) by Rubin and co-w~rkers~~ and our own group3 opens the way for a thorough investigation of its potential as a monomeric unit for all-carbon objects This initial work showed that the basic framework of (53) can be extended through oxidative acetylenic coupling, a critical methodology in the construction of all-carbon molecules and networksla Thus, the removal of the TMS protecting groups in (22) could be accomplished (Scheme 7) with potassium carbonate/methanol in THF Glaser-Hay heterocoupling [CuCl -TMEDA (N,N,N',N'-tetra- methylenediamine), O,] of (53) with excess trimethylsilylacety- lene afforded the butadiyne derivative (54) Similarly, selective deprotection of (24) permitted the homocoupling of (55) under Glaser-Hay conditions, producing the soluble dumbbell (56) Cyclization of deprotected (56) under similar conditions should give rise to new molecular carbon allotropes (Figure 6a), whereas the oxidative polymerization of (53) could provide a route to synthetic, all-carbon polymers (Figure 6b) Another area of interest is the encapsulation of a fullerene sphere within a dendntic shell The groups of Frkchet and W~d1~~synthesized the dendritic methanofullerene (58) using established dendrimer growth procedures (Scheme 8) The re- action of diphenol (50) with a four-fold excess of the fourth generation dendritic branch (57) in the presence of K2C03 afforded (58) as a light brown glass in 79% yield The dendritic arms dramatically increase the solubility of the fullerene and provide a more compact insulating layer than linear polymers Dendritic encapsulation techniques should in the future allow the preparation and study of fullerene molecules which, as a result of their isolation in a designed environment, may have altered physical properties compared to free fullerenes in bulk solid or liquid phase 5 Perspectives The fullerenes in general and the methanofullerenes in particular have already stimulated a tremendous amount of expenmental and theoretical work Experimentally, there is still much chemistry to be explored The facile access to regio- and stereo- specific multiple adducts is still an unsolved problem Are there CHEMICAL SOCIETY REVIEWS, 1994 T-4 r K2C03 Scheme 8 hitherto unrecognized solutions to fullerene regiochemistry similar to those developed for benzene denvatization? Well- defined multiple adducts of C6, would be desirable for several purposes -as cores for starburst dendrimers or for high-activity HIV protease inhibitors The field of fullerene chemistry in general has much to offer to the domain of synthetic method- ology The successful elaboration of methanofullerene side chains is a challenging task, because it is necessary to find the mildest conditions possible in order to avoid competitive reac- tions with the methanofullerene core This will almost certainly lead to the development of new synthetic methodology in the years to come On the other hand, the potential for methanoful- lerene-based matenals and industrial applications looks very promising Pure c60 itself has shown many intriguing proper- ties 40 Among these are the superconductivity of its alkali metal salts M&o as well as its non-linear optical properties 41It has already proven possible to prepare methanofullerene-based polymers which retain the interesting optical and electrochemi- cal properties of the parent methanofullerenes Methanofuller- enes may become components in new molecular and polymenc carbon allotropes with fascinating matenals properties la Lastly, methano-bridged derivatives of C,, and the higher f~llerenes,~~whose chemistry has only begun to be explored, might hold additional surprises in store The absorbance of the higher fullerenes in the near IR, when coupled to the efficient sensitization of singlet oxygen formation, promises potential applications in photodynamic therapy 37 For these reasons it is highly probable that the advances already made only represent the beginning of a long stream of methanofullerene-based developments in the realm of synthetic methodology as well as matenals and biomedical applications Note Added m Proof Two accurate low-temperature X-ray crystal structures of (54) as CS, and toluene solvates definitely confirm that methanofullerenes bridged at the [6,6]-ring junc- tion have a closed transannular bond [l 574(3) A], H L Anderson, C Boudon, F Diederich, J -P Gisselbrecht, M Gross, and P Seller, Angew Chem ,Int Ed Engl ,1994,33, in press Acknowledgments We thank the Royal Society for the Award of a European Science Exchange Programme Fellowship (to D P ) and the Swiss National Science Foundation for financial sup- port We are grateful to Drs Harry Anderson and Rudiger Faust for reading and improving the manuscript 6 References 1 (a)F Diederich, Nature, 1994,369, 199 (b)R Taylor and D R M Walton, Nature, 1993,363,685 (c)A Hirsch, Angeu Chem Int Ed Engl, 1993,32, 1138 2 H W Kroto, J R Heath, S C O’Brien, R F Curl, and R E Smalley, Nature, 1985,318, 162 3 W Kratschmer, L D Lamb, K Fostiropoulos, and D R Huffman, Nature, 1990,347, 354 4 (a)H W Kroto, Nature, 1987,329, 529 (b) R Taylor, Tetrahedron Lett , 1991,32, 3731 (c) R Taylor, J Chem Soc Perkin Trans 2, 1992,3 5 (a)J M Hawkins, T A Lewis, S D Loren, A Meyer, J R Heath, Y Shibato, and R J Saykally, J Org Chem ,1990,55,6250 (b)P J Fagan, J C Calabrese, and B Malone, Science f Washington DC) 1991,252,1160 (c) A L Balch, V J Catalano, and J W Lee, Inorg Chem ,1991,30,3980 6 F Diederich, R Ett1,Y Rubin,R L Whetten,R Beck,M Alvarez, S Anz, D Sensharma, F Wudl, K C Khemani, and A Koch Science (Washington DC), 199 1,252, 548 7 (a)K M Creegan, J L Robbins, W K Robbins, J M Millar, R D Sherwood,P J Tindal1,D M Cox,A B Smith,III,J P McCauley, Jr ,D R Jones, and R T Gallagher, J Am Chem Soc ,1992,114, 1103 (b)Y Elemes, S K Silverman, C Sheu, M Kao, C S Foote, M M Alvarez, and R L Whetten, AngeM Chem Int Ed Engl 1992,31, 351 8 (a) E Vogel, ‘Aromaticity’, Spec Pub1 No 21, The Chemical Society, London, 1967,113 (b)E Vogel, Pure Appl Chem ,1969 20 237 (c)E Vogel, Isr J Chem ,1980,20,215 (d)E Vogel, Puie Appl Chem , 1993,65, 143 9 T Suzuki, Q Li, K C Khemani, F Wudl, and 0 Almarsson, Science ( Washington DC), 199 1,254, 1 186 10 F Wudl, Acc Chem Res ,1992,25, 157 11 S Shi, K C Khemani, Q Li, and F Wudl, J Am Chem Soc , 1992, 114, 10656 12 M Prato, A Bianco, M Maggini, G Scorrano, C Toniolo, and F Wudl, J Org Chem , 1993,58,5578 13 R Sijbesma, G Srdanov, F Wudl, J A Castoro, C Wilkins, S H Friedman, D L Decamp, and G L Kenyon, J Am Chem Snc 1993,115,6510 14 T Suzuki, Q Li, K C Khemani, F Wudl, and 0 Almarsson, J Am Chem SOC, 1992,114,7300 15 M Prato, V Lucchini, M Maggini, E Stimpfl, G Scorrano, M Eiermann, T Suzuki, and F Wudl, J Am Chem Soc , 1993, 115, 8479 16 J Osterodt, M Nieger, P -M Windscheif, and F Vogtle, Chem Ber , 1993, 126,2331 17 S R Wilson and Y Wu ,J Chem Soc Cheni Coniniun ,1993,784 18 M Prato, T Suzuki, F Wudl, V Lucchini, and M Maggini, J Am Chem Soc , 1993,115,7876 19 T Suzuk1,Q Li,K C Khemani,and F Wudl, J Am Chem SOC, 1992, 114, 7301, and supplementary material SYNTHESES, STRUCTURES, AND PROPERTIES OF METHANOFULLERENES-F DIEDERICH, L ISAACS, AND D PHILP 255 20 A B Smith, 111, R M Strongin, L Brard, G T Furst, W J Romanow, K G Owens, and R C King, J Am Chem Soc , 1993, 115, 5829 21 L Isaacs, A Wehrsig, and F Diederich, Helv Chim Actu, 1993 76 22 L Isaacs, and F Diederich, Helv Chim Acta, 1993,76, 2454 23 F Diederich, L Isaacs, and D Philp, J Chem Soc Perktn Trans 2, 1994,391 24 A Skiebe and A Hirsch, J Chem SOC Chem Commun ,1994,335 25 A Vasella, P Uhlmann, C A A Waldraff, F Diederich, and C Thilgen, AngeN Chem Int Ed Engl , 1992,31, 1388 26 M Tsuda, T Ishida, T Nogami, S Kurono, and M Ohashi, Tetrahedron Lett , 1993, 34, 691 1 27 (a)H Gunther, H Schmickler, W Bremser, F A Straube, and E Vogel, Angeu Chem Inf Ed Engl, 1973, 12, 570 (b)H Gunther and H Schmickler, Pure Appl Chem ,1975,44,807 (c)E Vogel, T Scholl, J Lex, and G Hohlneicher, AngeH Chem Suppf, 1982, 1882 (6)L Frydman, B Frydman, I Kustanovich, S Vega, E Vogel, and C Yannoni, J Am Chem Soc , 1990, 112,6472 (e) R Arnz, J W d M Carneiro, W Klug.H Schmickler, E Vogel, R Breuckmann, and F -G Klarner, Angen Chem Int Ed Engl ,1991, 30,683 28 (a) R Hoffmann, Tetrahedron Lett, 1970, 2907 (b) H Gunther, Tetruhedron Lett, 1970, 5173 29 H Tokuyama, M Nakamura, and E Nakamura, Tetrahedron Lett , 1993,347429 30 Y -Z An, Y Rubin, C Schaller, and S W McElvany, J Org Chem , 1994,59,2927 31 H L Anderson, R Faust, Y Rubin, and F. Diederich, Angeu Chem Int Ed Engl , 1994,33,in press 32 C Bingel, Chem Ber , 1993,126, 1957 33 A Hirsch, I Lamparth, and H R Karfunkel, Angew Chew Znt Ed Engl , 1994,33,437 34 K Raghavachari and C Sosa, Chem Phys Lett , 1993,209,223 35 (a) D A Dixon, N Matsuzawa, T Fukunaga, and F Tebbe, J Phjs Chem ,1992,96,6107 (b)N Matsuzawa, D A Dixon, and T Fukunaga, J Phys Chem , 1992, 96, 7594 (c) R Taylor, Philos Trans R Soc London Ser A, 1993, 343, 87 (d)R D Haddon, Science (Washington DC), 1993,261, 1545 36 J W Arbogast, A P Darmanyan, C S Foote, Y Rubin, F Diederich, M M Alvarez, S J Anz, and R L Whetten, J Phys Chem , 1991,95, I1 37 H Tokuyama, S Yamago, E Nakamura, T Shinkai, and Y Sugiura, J Am Chem Soc , 1993,115,7918 38 S H Friedman, D L Decamp, R P Sijbesma, G Srdanov, F Wudl, and G L Kenyon, J Am Chem Soc , 1993,115,6506 39 K L Wooley,C J Hawker, J M J Frechet,F Wudl,G Srdanov, S Shi, C LI, and M Kao, J Am Chem Soc , 1993,115,9836 40 (a)‘Buckminsterfullerenes’, ed W E Billups and M A Ciufolini, VCH, Weinheim, 1993 (b)‘The Fullerenes’, ed H W Kroto, J E Fischer, and D E Cox, Pergamon Press, Oxford, 1993 41 H S Nalwa, Adv Muter, 1993,5, 341 42 F Diederich and R L Whetten, Acc Chem Res , 1992,25, 119

ISSN:0306-0012    

DOI:10.1039/CS9942300243

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Non-ideality in Isotopic Mixtures Gabor Jancso Central Research Institute for Physics Atomic Energy Research Institute PO5 49 I525 Budapest Hungary Luis P. N. Rebelo Department of Chemistry New University of Lisbon 2825 Monte de Caparica Lisbon Portugal W. Alexander Van Hook Chemistry Department University of Tennessee Knoxville Tennessee 37996 U.S.A. 1 Introduction mixtures ideal solutions will occur rather rarely’. Other exam- Ideal mixtures deserve special attention because (i) their behav- ples abound. iour is the simplest conceivable from either a mathematical or a physical point of view (ii) it is found experimentally that almost- ideal mixtures do exist so the ideal model is a useful one. Although any real mixture cannot be completely ideal the resemblances between many real mixtures and the ideal abstrac- tion are more striking than the differences. Isotopic mixtures i.e. mixtures of isotopomers (isotopomers are species that differ solely in isotopic content e.g. CH and CD,) have long been considered as textbook examples of ideal solutions for example Guggenheim states ‘statistical theory predicts that mixtures of very similar species in particular isotopes will be ideal’,’ Levine writes ‘the only truly ideal solutions would thus involve isotopic species’,* earlier Lewis and Randall had written ‘molecules which differ only by isotopic substitution . .. form ideal solu- tion~’,~and more recently Miinster claims ‘except for isotope Gabor Jancso MWS born in 1941 in Budapest Hungary. He graduated with a degree in chemistry from Eotvos Lorand Univer- sity in 1964 and immediately joined the Central Research Institute for Physics of the Hungarian Academy of Sciences where he is currently a Senior Research Scientist. In 1969-1970 and 1976- 1977 he spent one-year study leaves with W. A. Van Hook at the University of Tennessee. From 1982 to 1984 he was a visiting scientist at the Max Planck Institut fur Chemie in Mainz where he tt-orked btpith K. Heinzinger. His research interests include the investigation of the effects of isotopic substitution on the properties of liquids and mixtures molecular dynamics simulation of liquid ltuter and aqueous solutions and studies of the effects of intermo- lecular interaction on the vibrationalproperties of molecules in the condensed phase. More recentlev he has become involved with studies of the structural and dynamic properties of aqueous solutions using neutron diffraction. Luis Paulo Rebelo was born in Lisbon Portugal in 1960. He gruduated in 1982 in Chemical Engineering with honours from the Technical University of Lisbon. In 1989 he received his Ph.D. in PhqTsical Chemistry from the New University of Lisbon under the supervision of Professor Jorge Calado and Professor Manuel Gabor Jancso Luis Paulo Rebelo The ideal behaviour of isotopic mixtures is expected if one assumes that the intermolecular forces between pairs of like molecules of each type (e.g. CH,-CH CD,-CD,) and between unlike molecules (e.g. CH,-CD,) are all the same and further assumes the isotopomers to be the same size. Both assumptions are reasonable in first order. It follows that the detection of non- ideality in isotopic mixtures will yield informaton on the validity of the above assumptions and it will test the principles of the underlying mixture theories as well as the intermolecular poten- tial models. This review is concerned with the experimental determination of non-ideality in solutions of isotopomers and with the inter- pretation of such non-ideality in terms of the intermolecular interactions which occur within the solutions. A more compre- hensive treatment of the subject can be found in a recent review. Nunes da Ponte. He joined the New University of Lisbon in 1983 and has been an Assistant Professor since 1989. In 1989 he received the Student Government Association awardfor pedagogi- cal and scienti$c abilities. Rebelo has served as Visiting Scientist andlor Postdoctoral Researcher at Cornell University (1985) the Central Research Institute for Physics Budapest (1987) and the University of Tennessee (1990-1991). He has served as Member-Elect on the Board of Directors of the Portuguese Chemical Society (1987-1989). He is currently Principal Investigator of several research projects in the areas of cryogenic thermodynamics at moderate and high pressures molecular ther- modynamics of isotopically substituted species and phase eguili- bria in polymer solutions. W. Alexander Van Hook was born in I936 and received his education at the College of the Holy Cross (B.S.-mcl 1957) and Johns Hopkins University (M.A. 1960 Ph.D. 1961 mentor Paul Emmett). After a postdoctoral stay at Brookhaven National Laboratory spent with Jacob Bigeleisen and Max Wolfsberg he joined the University of Tennessee where he is presently Professor of Chemistry. Van Hook spent 1967-1968 on a Fulbright fellow- ship at the Universite Libre de Bruxelles studying with I. Prigogine and G. Thomaes and a good deal of 1972 at the Boris Kidric‘ Institute of Nuclear Sciences Belgrade on a National Acad- emy of Sciences Exchange Fellowship. Other shorter study visits have been spent at Lanzhou and Peking Uni-versities PRC and in Buda- pest and Lisbon. Van Hook’s research interests are in isotope eflects on the properties of con-densed phases solution thermodynamics and related Alexander Van Hook areas. 257 A mixture is said to be ideal if the chemical potential of component i,pl is a linear function of the logarithm of its mole fraction (x,) where pP(p,T) is the chemical potential of the pure liquid component at the same pressure and temperature as the mixture. This 'thermodynamic definition' of ideal mixing is equivalent to the one based on Raoult's law for a liquid mixture provided the vapours behave as perfect gases and provided the effect of pressure on the chemical potential of the liquid phase can be neglected. In the case of an ideal solution there is no heat evolved in the mixing process and neither is there any volume of mixing. The properties of real mixtures may be expressed in terms of excess molar properties. An excess property is defined in terms of the deviations of the real mixture from those of the ideal mixture of the same composition (in other words 'excess' means 'differ- ence' between real and ideal). The magnitudes of the excess Gibbs energy (GE) excess volume (VE) and excess enthalpy (HE)for mixtures of isotopomers indicate the extent to which these mixtures are non-ideal. For the excess Gibbs energy the measurement of the vapour pressures of isotopic mixtures of known composition is the most convenient way to determine non-ideal behaviour and that is the focal point of this review. Excess enthalpies can be determined by calorimetric measure- ments or from the temperature dependence of GE using the Gibbs-Helmholtz relation. The determination of excess volume in isotopic mixtures requires very precise density measurements. We will proceed first by describing the experimental technique used to measure vapour pressure differences between separated isotopomers and their mixtures. Using the theory of condensed phase isotope effects6 we will then show how the vapour pressure differences between isotopomers can be related to the intermole- cular forces acting in the liquid phase. Following this general description we will describe studies on some selected isotopic mixtures in detail. An important result will be the clear demon- stration that the proper consideration of the vibrational proper- ties of the component molecules is absolutely essential to the understanding of isotopomer mixtures. 2 Vapour Pressure Measurements Small differences in vapour pressure like those between separ- ated isotopomers or those between an isotopomer mixture and one or the other pure compound are usually expressed in terms of logarithm of the pressure ratio lnp'/p (the prime denotes the lighter molecule). This so-called vapour pressure isotope effect (VPIE) is said to be normal if lnp'/p > 0 i.e. if the lighter isotopomer has a higher vapour pressure than the heavier one and inverse if lnp'/p < 0. The temperature at which the vapour pressures of the isotopomers are the same (lnp'lp = 0) is called the crossover temperature. For small diff'erences lnp'/p M Ap/ p = (p' -p)/p.It follows that the best possible precision can be obtained by measuring Ap andp simultaneously using a differen- tial technique as shown in Figure la. Here the sample cells (A,A') containing the pure isotopomers are thermostatted the temperature is measured with a platinum resistance ther- mometer and the pressure difference between isotopomer sam- ples and the absolute pressure of the reference sample are determined with pressure transducers (P AP). For experiments on isotopic mixtures an additional differential vapour pressure must be measured i.e. that between the mixture and one of the pure isotopomers. If this is done simultaneously with the differential vapour pressure measurement between the separated pure isotopomers by using a third cell one obtains a truly double-differential method and the non-ideality can be assessed with the best possible accuracy.' The non-ideality of the binary isotopic mixture can be conveniently expressed in terms of the excess pressure dpE dpE = Ap,, -[x'p'C' + (1 -x')pC]+ p' CHEMICAL SOCIETY REVIEWS. 1994 a To b Figure 1 Schematic diagrams of common experimental techniques used for the determination of vapour pressure differences between isotopic isotopomers (a) and for the determination of the isotope separation factor (a)in a one-plate fractionation experiment (b). A and A' are sample containers for pure isotopomers P and LIPare pressure and differential pressure transducers; y y' and x x' are the concentrations of the two isotopomers in the vapour and liquid phases respectively; Tis a resistance thermometer. where dp,, is the experimentally observed vapour pressure difference between mixture and reference (usually the lighter isotopomer) x is the mole fraction and Cis the factor correcting for vapour non-ideality and liquid phase molar volume.8 (Note that ApE reduces to exactly the deviation from Raoult's law provided one neglects the correction factors C and C'.) The difference between the ApE/p ratio and zero gives one some feeling about the magnitude of the deviation of isotopic mixtures from ideal behaviour. For example the experimental data for an equimolar mixture of C6H6 and C6D6 shown in Figure 2 exhibit a non-ideality of about 0.02-0.03% which corresponds to a ApEvalue of 3.2 Pa (0.024mm Hg). For comparison the vapour pressure difference between C6D6 and C6H6 is 360 Pa at room temperature. The small scatter of the data around the correla- tion line demonstrates the high precision which can be achieved by present-day differential vapour-pressure measuring techniques. Another and quite different approach to determine the VPIE involves a measurement of the isotope separation factor (a) defined by the equilibrium between the liquid phase and its saturated vapour in a one-plate fractionation experiment. We define a = (y'/y)/(x'/x),where y and x are the concentrations in 3lo 1 1 12 10 8 106C T2 Figure 2 Excess pressures of equimolar mixtures of C6H6 and C6D6. For definition of dpEsee equation 2 pHis the vapour pressure of pure CISHI9 NON-IDEALITY IN ISOTOPIC MIXTURES-G JANCSO L P N the vapour and liquid phase respectively (Figure lb) The concentration ratios are straightforwardly related to the corres- ponding pressure ratios and free energy differences using standard thermodynamic relations The accuracy of this method is limited by the accuracy of the isotopic analysis of the two coexisting phases an improvement can be achieved by multipli- cation of the elementary separation process eg by using distillation columns The relationship between a and the vapour pressures of the pure isotopomers is given in the limit of infinite dilution (x' E 1) to a good approximation by the equation5 Here B is the second virial coefficient of the gas /3 is the isothermal compressibility of the liquid and Vo is the molar volume of the liquid In fractionation experiments the measure- ment is always made on an isotopomer mixture never on pure separated isotopes and this must be kept in mind throughout the interpretation We will return to this important consider- ation later 3 Vapour Pressure Differences between Pure lsotopomers Before we turn to a discussion of the excess properties of isotopic mixtures we first address the question of why there is any difference at all in the vapour pressures of isotopomers Our interpretation of the VPIEs6 is carried out within the framework of the Born-Oppenheimer approximation which states that the electronic structure of atoms and molecules is essentially inde- pendent of the isotopic distribution of nuclear mass It follows that the potential energy is isotope independent In other words isotope effects are nuclear mass effects resulting from the motion of nuclei of different mass on the same (or identical) potential energy surface(s) It may be noted in passing that this idea has been widely exploited by spectroscopists when they determine force constants from vibrational frequencies of isotopic molecules The isotope separation factor can be expressed in terms of the equilibrium constant for the liquid-vapour phase equilibrium X and X' denote different isotopomers also (4) Thus K is identical with the distillation separation factor Consequently a and VPIE may be expressed in terms of isotopic ratios of gas-phase and liquid-phase partition functions (Q)6 (The precise relationship should include the correction terms given in equation 3 ) A theory of VPIEs based on the above concepts was worked out by Bigeleisen9 in 1961 In model calculations of Q'/Q it is often assumed that all the degrees of freedom except the rotational and translational ones of the gas can be treated as harmonic oscillators In the vapour phase each freely rotating and translating gas molecule has 3n -6 vibrational modes (n is the number of atoms in the molecule 3n -6 is the number of internal degrees of freedom) In the condensed phase the 3nN(N is the number of molecules) degrees of freedom are very often treated within the framework of a simplified cell model of the condensed phase In this model O an 'average' condensed phase molecule is assumed to have 3n degrees of freedom of these 3n -6 are similar to the vibrational modes of the gas-phase molecule The other six external degrees of freedom correspond REBELO AND W A VAN HOOK to the gas-phase translations and rotations which have become bound in the condensed phase The assignment of positive force constants (binding) to the external degrees of freedom will always work in the direction of normal VPIE (p' > p) This is a consequence of the fact that the heavier isotopomer lies deeper in the intermolecular potential well than does the lighter isoto- pomer (see the left side of Figure 3) It therefore requires a higher heat of vaporization (this reasoning is valid strictly only at low temperatures) AF Figure 3 Potential energy curves for an external motion (left-side) and for an internal vibration (right-side) compared in the vapour (a) and liquid (b)phase For the external motion the zero-point energy change on condensation [(E -E;) -(E -E,)] = AE' -AE = E -E; < 0 and this leads in the direction of a normal VPIE (p' > p) For the internal vibration in the ordinary case (which is shown) AE' -AE = [(EL -E,) -(E' -El)]> 0 and this leads in the direc- tion of an inverse VPIE (p > p') The diagrams are schematic the shifts very much exaggerated in order to better illustrate the argu- ments in the text For monatomic substances the only degrees of freedom are translational and a normal VPIE is always observed but for polyatomic molecules the effect of condensed-phase intermole- cular forces on the internal vibrations must be also considered In most cases we know from IR and Raman investigations that the frequencies of the internal vibrations shift toward lower frequencies (red shift) when the molecule passes from the vapour phase to the condensed phase (Such frequency shifts are rather small e g the CH stretching frequency which lies around 3000 cm-' in the vapour phase decreases by about 10 cm-' on condensation ) When substituted into the partition function ratios which express the VPIE (equation 5) such a red shift (equivalent to a lowering of the internal force constant on condensation) leads to an inverse VPIE (p> p') The discussion in the last paragraph is nicely illustrated in Figure 3 where we consider a typical vibration at lower tempera- ture [I e in the zero-point energy (ZPE) approximation where the population in excited vibrational levels is negligible] For an internal degree of freedom the reduced mass of that particular vibrational motion results in the more lightly substituted isoto- pomer lying higher in the vibrational potential well (right side of Figure 3) In the harmonic approximation the isotopic ZPE difference is proportional to the square root of the force constant describing that particular vibration If on condensation the frequency of interest decreases (I e shifts to the red) its force constant decreases and the corresponding ZPE difference in the condensed phase will be smaller than that in the vapour phase and it is this negative difference of zero-point energy differences (liquid-vapour) which accounts for a negative (inverse) contri- bution to the VPIE For an external motion (hindered transla- tion or rotation left side of Figure 3) the gas-phase force constant is zero since the corresponding gas-phase frequency is zero On condensation then the frequency and its associated force constant increase (i e shift to the blue) This accounts for a positive (normal) contribution to VPIE The observed VPIE is the result of the interplay between the normal isotope effect from hindered translations and rotations and the inverse isotope effect arising from the internal degrees of freedom The temperature dependence of the effects is different the normal isotope effect usually falls off more steeply with temperature (proportionally to TP2)then does the inverse isotope effect (T-l) (The reason for the difference lies in the fact that the ZPE approximation is often inadequate for the low lying external frequencies and vibrational excitation must be con- sidered This leads to the T-dependence for these frequencies ) With internal and external effects of opposite sign each pro- portional to different powers of temperature it is not surprising that the VPIE can and often does cross from positive to negative (the crossover effect at low enough temperature the T-2term must dominate and the VPIE is necessarily positive) At higher temperature yet the inverse VPIE (when it occurs) must display a maximum before it decays to zero at very high temperature A good example which illustrates the crossover is the case of heavy water/light water H20has a higher vapour pressure than D20 up to 493 K (crossover temperature) but above this temperature D20is more volatile than H20 At higher temperature yet the inverse effect would decay to zero but before that decay is observed the critical point is reached the distinction between condensed and vapour phase vanishes and the point becomes moot The theoretical analysis described above couples thermody- namic observation with spectroscopically observed frequency shifts on condensation for isotope-sensitive vibrations By com- bining the available spectroscopic data on vapour-liquid vibra- tional frequency shifts with experimental VPIE data it follows that one can obtain information about the details of the intermo- lecular interactions in the liquid phase For example VPIE studies have furnished information on the vibrational coupling between internal vibrations and molecular translations and/or rotations which occurs in the condensed phase on the density dependence of the force constants which govern the external molecular motions and internal vibrations in the liquid on changes in vibrational anharmonicity which occur on condensa- tion and on the magnitude of the ‘dielectric correction’ to IR absorption peaks in condensed phases 4 Theory of the Excess Functions of Isotopic Mixtures The basic theory describing the excess functions of isotopic mixtures was developed by Prigogine Bingen and Bellemans’ (PBB) forty years ago in order to interpret the properties of hydrogen (H2 D T HD HT DT) and helium (3He 4He) isotopomer solutions The central idea of the PBB model is the assumption that since the intermolecular forces between differ- ent isotopomers are the same (Born-Oppenheimer approxima- tion vide supra) the deviations from ideality in isotopic mixtures can be rationalized in terms of the differences in the molar volume of isotopomers (molar volume isotope effect MVIE usually Yo‘2 Vo) The physical origin of the MVIE is well understood in terms of vibrational amplitude isotope effects which are a consequence of vibrational motion on an isotope- independent intramolecular potential surface PBB suggested that the excess Gibbs energy of mixing (GE) can be calculated in a two-step process (1) compression or expansion of the molar volumes of the pure components to the molar volume of the mixture (V,) (11) mixing the two compo- nents now at the same volume The essential point is that in the second step the mixing process can be considered ideal (the interactions between the components as well as their volumes are the same) therefore G is equal to the amount of work ( W) done in compressing the lighter isotope and in expanding the heavier one to the molar volume of the mixture l2 l3 XI Jr pdV -In step one it is necessary to apply positive pressure to compress CHEMICAL SOCIETY REVIEWS 1994 the lighter isotopomer from Vo’ to V and negative pressure (tension) to expand the heavier isotopomer from V O to V The relationship between p and V is defined by the equation of state and the application of negative pressures involves the extrapola- tion of the V vs p curve Note that both terms in equation 6 are positive I e GE calculated from PBB theory is always positive The statement IS equivalent to the prediction of positive devi- ations (only) from Raoult’s law In an infinitely dilute solution the process described above can be easily visualized first the molar volume of the minor compo- nent or ‘solute’ isotopomer is compressed or expanded to that of the ‘solvent’ isotopomer and then the components are mixed In this specific case the excess chemical potential (pEm),written here for the heavier isotopic species can be given as5 where /3 is the isothermal compressibility and ycc is the activity coefficient associated with the free energy transfer of the heavier isotopomer from its pure liquid state to infinite dilution in the lighter isotopomer as a solvent One expects that yc0 = p’,I e symmetrical behaviour around x = x‘ = 0 5 in which case GE can be represented with a one-term expansion of the Redlich- Kister type GE zxx’pECoBy using this expression infinite dilution values of GEcan be obtained from experimental data at finite concentrations (or vice versa) Jancso and Van Hook14 have suggested that the expression ( Vo’-V0)2/2pVomay significantly underestimate the value of pEccfor molecules with internal vibrations The reason is that the compressibility is principally determined by the overall (exter- nal) motions of the molecules in the intermolecular potential Therefore the expression containing /3 does not take proper account of the contributions of internal degrees of freedom to GE (or pEm) The internal vibrations are volume dependent Consequently during the compression/expansion step (I) their frequencies change and the work involved represents the inter- nal contribution to GE By using a harmonic oscillator model and writing the partition function Q of an ‘average’ molecule in the liquid phase in terms of 3n -6 internal and 6 external degrees of freedom and recalling A = -RTlnQ(v1,v2 ) one obtains14 pE7(V0’)=-RTS b’ C (i?LQ) -3n -(8”) dV cu 1 r avr where A is the Helmholtz energy and ul is the i-th vibrational frequency The volume dependence of vibrational frequencies can be deduced from their more readily available pressure dependence using (&/a V)= -(&/dp)( V/p) Another approach is to use the Gruneisen parameters T,= -dInv,/O?lnV which can be determined from the change in vibrational frequency and molar volume on melting or from the temperature dependence of the vibrational frequencies in the liquid phase l5 The use of rl parameters is equivalent to the assumption of the quasiharmo- nic approximation widely employed by solid-state scientists In this approximation the vibrations are assumed to remain har- monic about the new mean positions as the solid (liquid) phase expands In the following section examples of measurements on some selected isotopic mixtures will be considered and the experimen- tal excess thermodynamic data will be compared with the results of model calculations 5 Excess Properties of Some Selected Isotopic Mixt u res 5.1 36Ar/40ArMixtures Theoretical analysis of the experimental data for mixtures of monatomic isotopes is considerably facilitated by the fact that in this case there are neither rotational nor internal degrees of freedom The comparison of the isotope separation factors (a) obtained from one-stage liquid-vapour equilibrium measure- tt ments with VPIE values obtained from differential vapour pressure measurements on separated isotopes yields values for Iny"' (see equations 3 and 7) Bigleisen et af have investigated the 36Ar/40Ar system and from their experimental data one obtains Iny"' = -3 x lop4 at 83 K The negative sign is surprising and it cannot be rationalized in terms of the PBB theory (equation 7) or equation 8 In an analysis of the PBB theory Singh and Van Hook17 argued that the assumption of zero excess free energy of mixing for step ii (step ii refers to the process where the two isoto- pomers at identical molar volumes are mixed) can be expected only under congruent conditions By congruent they refer to a hypothetical situation where the radial distribution functions describing the liquid are identical for each of the separated isotopes If one assumes that the separated isotopes are congru- ent in their pure liquid states they can no longer be congruent after one of the isotopomers had been compressed (expanded) to the molar volume of the order This is a consequence of the fact that compression (expansion) along a normal thermodynamic equilibrium path not only shifts the radial distribution function along the Y axis (z e shifts the scale) but also changes its shape (I e changes the arrangement and/or number of neighbours around a central particle) The corrected calculation1 based on congruent expansion gives Iny"' = -2 x in reasonably good agreeement with the observed value Although the congru- ent/non-congruent correction for this mixture of monatomic isotopomers is important and even determines the sign of the effect we do not expect it to make a significant contribution in the general case (I e for mixtures of polyatomic isotopomers) In isotopic mixtures of polyatomic molecules a considerable part (sometimes most) of GE arises from the internal degrees of freedom and in most cases the 'congruent correction 'can be safely neglected 5.2 HJD Mixtures The excess properties (GE,HE,p)of H,/D mixtures have been thoroughly studied The PBB theory satisfactorily describes the significant deviations from Raoult's law observed in these solutions for example at 20 K G E(expt) =7 2 J mol- for the equimolar mixture while equation 6 gives GE(calc) =8 9 Jmol-' At infinite dilution equation 7 gives pEx(calc) =22 J mol- which compares favourably with the experimental value pEz(expt) =29 J mol- The contribution of the H-H stretch-ing vibration can be estimated from its frequency change during the process when H initially at its equilibrium molar volume is compressed to the equilibrium molar volume of D From the known values for the pressure-dependence of the vibrational frequency (dv/dp =2 5 cm-' kbar-l) the compressibility and the MVIE one finds that the intramolecular vibrational fre- quency of H blue-shifts 0 22 cm-' during the compression from Vo(H2) to Vo(DZ) That shift corresponds to a contribu- tion of 1 3 Jmol-' to GE Thus in this case the internal contribution is only a small fraction of the effect calculated from equation 7 We conclude that in HJD mixtures GE is domi- nated by the compressibility (equation 7) which in turn is principally determined by the intermolecular part of the potential 5.3 CHJCD Mixtures This system consists of simple non-polar quasi-spherical mole- cules and represents one of the simplest possible mixtures of polyatomic molecules Therefore the prediction of its excess properties by any liquid-state theory should be as simple and direct as possible and can be used to test the performance of that theory The vapour pressure differences between CH,/CD mixtures and CH have been determined as a function of concentration and temperature The excess free energies (G E and enthalpies (HE)are shown at 100 K and 100-120 K respectively in Figure 4 The experimental data are nicely represented by GE = xx'pEz O8 I 06 04 02 0 0 02 04 06 08 1 *D4 16 12 08 04 n-0 02 04 06 08 1 XCD4 Figure 4 Excess free energy (GE)and excess enthalpy (HE)for the CH,/ CD mixture as a function of the mole fraction of CD at I00 K and 100-120 K respectively It is interesting to compare interpretations of GE for isoto- pomer solutions using conventional and widely employed liquid-state structureless particle models with the 3nN dimen-sional harmonic cell model described earlier in this review Two popular and successful theoretical models for simple liquid mixtures were selected for comparison the one-fluid van der Waals theory developed by Leland et a1 (VDW-1 model) and the WCA-based perturbation theory of Kohler et af 2o (IcLJ model) In the application of both theories effective isotope- dependent Lennard-Jones potential parameters (a,E) and an adjustable parameter (ICHD) describing the deviation from the Berthelot geometric mean rule [EHD =(1-kHD)(EHHEDD)+] were used The results found for the equimolar mixture at 100 K are presented and compared with the experimental data in Table 1 The calculations based on equation 8 used the Gruneisen parameters rrot=0 and Ttr= 1 8 and density dependent inter- nal vibrational frequencies which are consistent with previous VPIE calculations The results of the cell model calculationls agree much better with experiment than do those of VDW- I and lcLJ models (see Table 1) Detailed analysis of the cell model Table 1 Values of the exce,s properties obtained for an equimolar mixture of CH and CD at 100 K (J mol ->" kHDh GE HE VDW-1 model 0 -0 02 -001 -2 9 x 10 5c 1 04 lcLJ model 0 -0 012 -000 -2 6 x 1 02 Theory of isotope 0 58 19 effects equation 8 Experimental 0 60 15 Ref 18 Value describing the deviation from Berthelot rule (see text) Obtained by fitting to reproduce HE results shows that both external and internal vibrational contri- butions to GE are important (at 100 K they amount to 67 and 33% of the total GE respectively). This explains the failure of models that do not consider internal vibration explicitly. Also in dealing with structureless particle models it is difficult to ensure consistency with the Born-Oppenheimer principle. An attract- ive feature of the cell model is that Born-Oppenheimer consist- ency is an integral part of the model. There is another interesting point of experimental detail which confirms the cell model interpretation. The theory pre- dicts GE = 0 when MVIE = 0 (see equations 7 and 8). For the CH,/CD system MVIE is normal (YO' > Vo)for T/K < 170 but becomes inverse (Vo> YO')at 170 f2 K.,l An extrapola- tion of the isotopomer mixture data (GEvs.1/T) gives GE = 0 at 171 f4 K. The agreement is excellent. 5.4 C,jH,j/C,jD,Mixtures The vapour pressure of C6H6 C6D6 and their equimolar mixture has been determined between 279 and 353 K.8 The vapour pressure of C6D6 is 2.8% higher than that of C& at the melting point and the relative vapour pressure difference slowly decreases with increasing temperature. The inverse VPIE is mainly due to the vapour-liquid red-shift in the six CH stretch- ing vibrational frequencies the contribution from the hindered translational and rotational motions accounts for less than 10% of the total VPIE.' A careful analysis of the temperature- dependence of the VPIE yielded values for the volume depen- dence of the liquid phase carbon-hydrogen stretching frequen- cies which are in reasonable agreement with those obtained from spectroscopic studies at elevated pressure and/or from analysis of the excess free energy of the isotopomer mixture (vide infra). A value for pEm=4GEequlmolar has been calculated from the experimental excess pressures displayed in Figure 2 and is compared with theoretical values in Table 2. To calculate the contribution of carbon-hydrogen stretching vibrations to G we used the available experimental data for dvcH/dp 1.5 cm-kbar -l isothermal compressibility and MVIE (the molar volume of benzene is 0.27% larger than that of deuterated benzene). The PBB approach predicted very nearly the same GE as did the contribution from the external degrees of freedom in the cell model formalism (0.33 vs.0.21 J mol-l). Once again this demonstrates that the PBB model accounts only for the external contribution which roughly is only about 10% of the total effect. It neglects the much more important contribution of internal vibrations. A second calculation which carried out the compression along a congruent path slightly improved the agreement with experiment. The agreement between the calcu- lated and observed excess enthalpies is less satisfactory. According to the calculations described above the free energy change accompanying the transfer of a C6H molecule from the neat liquid to infinitely dilute solution in C,D6 corresponds to a change of 0.04 cm-in the CH stretching frequency (-3000 cm-l). The spectroscopic observation of such a small frequency Table 2 Calculated and observed partial molar excess free energies (pEm)and enthalpies (HE") of benzene at infinite dilution in deuterobenzene at 298 K (Jmol-ly PE" HE% Contributions to pEx (calc) (obs) (calc) (obs) Internal 1.53 External 0.21 Total 1.74 2.3 f0.2 7.0 4.4 f2 Congruent pathb PBB 1.9 0.33 fl Ref. 5. See text. CHEMICAL SOCIETY REVIEWS 1994 shift is not possible by ordinary IR or Raman techniques because of the peak broadening in the liquid phase. However the Raman difference technique reported by Laane and co- workers2 enables the detection of shifts of even this order of magnitude. The frequency shifts Laane observed spectroscopi- cally in CGH~/C~D~ mixtures were interpretedz2 in terms of additive resonant intermolecular coupling and volume effects. It is the latter effect that is relevant to G since it takes into account the change in the molar volume when one isotopomer is dis- solved in the other. When C6H is dissolved in C6D6 it is slightly compressed consequently the repulsive forces between mole- cules become dominant and the CH stretching frequency increases. Although the magnitude of this volume effect (N 0.15 cm-l) was found to be three or four times larger than the value deduced from the G calculations the direction of the shift is the same in both cases. If one considers the very small magnitude of the shifts together with the uncertainties in the calculations and experiments both thermodynamic and spectroscopic one can conclude that the results of Raman difference spectroscopy of isotopic mixtures lend strong support to the present model of excess thermodynamic functions of isotopic mixtures. 5.5 H20/D20Mixtures Experimental data on various excess properties of H20/D,0 mixtures have accumulated for more than sixty years. The interpretation of the thermodynamic properties of this mixture is complicated by the existence of the disproportionation equili- brium H20+ D,O = 2 HOD i.e. the H20/D,0 mixture should be treated as a three-component system. The concen- tration of the different species in the liquid phase can be calculated from the equilibrium constant taken as 3.8 at 298 K. However there is another complication in the understanding of this system namely that the bulk properties of HDO are not directly measurable because this molecule cannot be isolated. The enthalpy of the disproportionation reaction (AH)is large and the ratio AHE/AHis difficult to measure because AHE is quite small. Calorimetric measurements (e.g. K~ga~~)are apparently not sensitive enough to permit an unambiguous conclusion regarding ideality in H,O/HDO/D,O mixtures. Vapour pressure measurements on H,0/D20 mixtures24 show that the mixtures do not deviate from the ideal behaviour within the limits of the experimental data. (The vapour pressure of HDO can be calculated from the expression h(PHOH/PDOD)/ h(PHOH/PHOD) = 1.92 f0.026; if the so-called 'law of the geo- metric mean' were valid for this system then instead of 1.92 one would use 2.00 for the ratio). The most reliable measure of excess properties in water isotopomer mixtures come from analysis of distillation data at high dilution which give lnym values of 8 f2 x lo- and -4 f2 x lop4for isotopic substitution at hydrogen and oxy- gen respecti~ely.~~ Note the opposite signs of the excess free energies which nicely correlate with the observed molar volume isotope effects (the molar volume of HDO is larger by 0.17% and that of H *Ois smaller by 0.15% than is the molar volume of H,160).6 The observation is important. The Iny" values calculated from equation 8 are in good agreement with experi- ment." The PBB approach or any of its subsequent modifica- tions taking account of contributions from external degrees of freedom alone predict positive values for both isotopic mixtures and thus are necessarily in error. 5.6 Polymer/Polymer Isotopomer Solutions In the sections above selected examples of excess free energies in isotopomer solutions have been given. In every case the effects are small difficult to measure not at all dramatic. It is only natural to look for ways to enhance the effects in order more convincingly to demonstrate their existence and more easily to measure their consequences. Since in first approximation the excess free energy parameter scales in proportion to the number of isotopic substitutions and the fractionation vapour pressure NON-IDEALITY IN ISOTOPIC MIXTURES-G JANCSO L P N REBELO AND W A VAN HOOK enhancement or other parameter of experimental interest scales in proportion to GE/T,we conclude that studies are best carried out with Tas low as possible or on molecules with many isotopically substituted bonds We have already described low temperature experiments in the sections on hydrogen argon and methane solutions For hydrogen especially the excess effects are large not the least because the temperature is so very low (x20K) For most liquids however it is impossible to lower the temperature very far -freezing intervenes Fortunately the alternative path is open By increasing the number of substituted bonds i e making the molecule larger and larger it is possible to increase GE/Tto the point where the solution becomes so non-ideal that phase separation occurs Since the excess free energy per bond is so small one is well into polymer/ polymer solutions before this happens The observation of precipitation in any isotopomer mixture is exciting It offers incontrovertible proof that the rather subtle effects we have been describing in this review do in fact exist It demonstrates that they have important physical consequences In the next few paragraphs we will review some of that evidence Within the past few years Bates and co-workers26 have reported phase separation in isotopomer mixtures of H/D polymers including polystyrenes polybutadienes and polyethy- lene/polypropylene Light and neutron scattering techniques have been employed The authors extracted a Flory-Huggins excess free energy parameter (x)from the scattering data As this interaction parameter x = (GE/ V)/#‘ increases to its critical limit x = 2RT the solution demixes at an upper critical solu- tion temperature TThe 4’s are volume fractions but for isotopomer solutions mole fractions would serve equally well The thermodynamics of demixing of isotopomer solutions of polymers was originally considered by Buckingham and Hents- che12’ but their theory was developed at the PBB level and has proved to be inadequate Following Singh and Van Hook2* we write the excess free energy of the solution of polymer isoto- pomers as AE/RT= 44’(NTr/2)(dV/V)(U‘-U) (9) Here N is the number of monomer units per molecule r the number of H/D substituted bonds per monomer rthe Grunei- sen parameter for the effective frequency (taken as the CH(CD) stretch) d V/V is the MVIE and u = hcv/kT The zero-point energy approximation is employed For polymers the relative contribution of external lattice modes is negligible and their effect is not included in equation 9 The conditions for phase separation are obtained by differentiating the total free energy with respect to concentration and setting second and third derivatives to zero This gives 4’ = 4 = 0 5 and N is the critical polymerization number p‘ and p are the reduced masses for the CH and CD oscillators respectively The excess free energy per H/D substituted bond is small just as small for example as the tiny effects discussed in the sections on small molecules above However the effects are cumulative and as the total number of substituted bonds Nr increases the excess free energy becomes large enough to cause phase separation For polybutadiene the critical polymerization number N is found to be 1 3 x lo3monomer units It is convenient to compare theory and experiment using pEa/RT= lnym ym is the activity coeffi- cient per monomer unit and for the one-term Flory-Huggins model describes the free energy of transfer of a given isotopomer from its Raoult’s law reference state to an infinitely dilute solution with the other isotopomer as solvent in this case lny“ = 2/Nc For polybutadiene experiment gives 1041nyx = 9 f2 the calculated value is 11 f4 The PBB approximation yields 4 f1 and an alternative analysis by Bates and Wignal15 26 gives 1041ny“ = 14 The authors26 chose an interpretation somewhat different than the one we described above and in the earlier sections of this review They related the contribution of internal vibrational modes to polarizability isotope effects (PIE’s) and employed typical values transferred from small molecule PIE’s This is a perfectly appropriate method just like MVIE the PIE is understood in terms of an isotope-independent potential energy surface describing the intramolecular vibrations The two effects share a common origin The observation of demixing in polymer/polymer isotopomer solutions comprises a powerful verification of the general ideas presented in this review The light scattering patterns which occur during the spinodal quench of isotopomer solutions offer incontrovertible evidence that a phase transition IS occurring The demixing is dramatic confirmation of non-ideality in the isotopomer solutions Its direct observation involves the recog- nition that x itself scales proportionally to the number of substituted bonds in the isotopomers being compared Although each CH/CD bond contributes only a very small amount to the excess free energy (as witness the minuscule effects in small molecule systems) they are additive and eventually result in precipitation 6 Conclusions High-precision measurements of the excess thermodynamic properties of mixtures of isotopomers have shown that even these very simple mixtures exhibit small but still significant deviations from ideal behaviour The magnitude of the excess Gibbs energy in isotopomer solutions is two to three orders of magnitude smaller than excess free energies typically observed in binary mixtures of non-isotopic simple molecules In the case of mixtures of H/D-substituted polymers the excess free energies are large enough to cause demixing Theoretical analysis demonstrates that the origin of non- ideality in isotopomer mixtures is closely connected with differ- ence between the molar volumes of isotopomers (the MVIE) Excess thermodynamic properties can be evaluated by integrat- ing the free energy which is volume-dependent across MVIE This is most conveniently done by expressing the free energies in terms of vibrational properties of the component molecules Thus the integration involves the volume-dependence of the vibrational properties of the component molecules It is import- ant to keep in mind that the MVIE is itself well understood in terms of isotope effects on vibrational amplitude and amplitude isotope effects in turn can be quantitatively described using a set of isotope independent inter- and intramolecular force constants Thus the development is self-consistent and is also consistent with the Born-Oppenheimer principle Those mixture theories which consider molecules as structure- less particles (I e do not include internal vibrations) fail to rationalize the excess properties of isotopomer mixtures properly Although vibrational effects can be and often are neglected in theoretical treatments of mixtures of non-isotopomers (since the non-ideality there is mainly caused by differences in intermolecular forces) that cannot be so for isotopomer mixtures where the force fields are identical (or nearly so Born-Oppenheimer approximation) Thus an import- ant conclusion from the work reviewed above is that vibrational effects although sometimes outweighed by other factors do contribute importantly in determining the properties of mix- tures and solutions In isotopomer solutions it is essential to consider the density dependence of vibrations explicitly In some cases the very small frequency shifts predicted from the analysis of isotopomer thermodynamics have been corroborated by Raman difference investigations AcknowIedgments Financial support from the Hungarian Research Fund under grant no OTKA-1846 and from JNICT/ BASE under contract no CEN/ 1 120/92 is gratefully acknowl- edged as is support from the U S Department of Energy Division of Materials Sciences under grant 91 -ER45374 and The National Science Foundation under grant CHE-9 1 I3636 264 7 References 1 E A Guggenheim ‘Thermodynamics’ North-Holland Publ Co Amsterdam 1986 2 I N Levine ‘Physical Chemistry’ McGraw-HI11 Kogakusha Ltd Tokyo 1978 3 G N Lewis and M Randall ‘Thermodynamics’ 2nd Edition revised by K S Pitzer and L Brewer McGraw-Hi11 Book Co Inc London 1961 4 A Munster ‘Statistical Thermodynamics’ Vol 11 Springer-Verlag Berlin 1974 5 G Jancso L P N Rebelo and W A Van Hook Chem Rev 1993 93,2645 6 G Jancso and W A Van Hook Chem Rev 1974,74,689 7 J C G Calado M Nunes da Ponte L P N Rebelo and L A K Staveley J Phys Chem 1989,93 3355 8 G Jakli P Tzias and W A Van Hook J Chem Phys 1978,68 3177 9 J Bigeleisen J Chem Phys 1961,34 1485 10 M J Stern W A Van Hook and M Wolfsberg J Chem Phys 1963,39,3179 11 J Bigeleisen M W Lee and F Mandel Acc Chem Res 1975,8 179 12 I Prigogine R Bingen and A Bellemans Physrca 1954 20 633 I Prigogine ‘The Molecular Theory of Solutions’ North-Holland Publ Co ,Amsterdam 1957 13 H F P Knaap R J J Van Heijningen J Korving and J J M Beenakker Physrca 1962,28,343 14 G Jancso and W A Van Hook Physica 1978,91A 619 G Jancso and W A Van Hook J Chem Phys 1978,68,3191 CHEMICAL SOCIETY REVIEWS 1994 15 J N C Lopes L P N Rebelo and G Jancso J Mol Liq ,1992,54 115 16 M W Lee S Fuks and J Bigeleisen J Chem Phis ,1970,53,4066 J TPhillips C U Linderstrom-Lang and J Bigeleisen J Chem Phys 1972,56 5053 17 R R Singh and W A Van Hook J Chem Phys 1987,86,2969 18 J C G Calado G Jancso J N C Lopes L Marko M Nunes da Ponte L P N Rebelo and L A K Staveley J Chem Phys 1994 100,4582 19 T W Leland J S Rowlinson and G A Sather Trans Furadaj Soc 1968,64 1447 20 F Kohler M Bohn J Fischer and R Zimmermann Monatsh Chem 1987,118 169 21 A F Grigor and W A Steele J Chem Phv~ 1968,48 1032 22 N Meinander M M Strube A N Johnson and J Lame J Chem Phys 1987,86,4762 and references cited therein 23 Y Koga J Chem Thermodyn ,1987,19,571 24 R C Phutela and D V Fenby Aust J Chem 1979 32 197 G Jancso and G Jakli Aust J Chem 1980,33,2357 25 G Jakli and W A Van Hook Geochem J 1981,15,47 26 F S Bates G D Wignall and W C Koehler Phjs Re1 Lett 1985,55,2425,F S Bates and G D Wignall Macromolecules 1986 19,932 F S Bates and P Wilthuis J Chem Phis 1989,91 3258 C Kedrowski F S Bates and P Wilthuis Macromolecules 1993 26 3448 27 A D Buckingham and H G E Hentschel J Poljm Sci Poljm Phys 1980 18 853 28 R R Singh and W A Van Hook Macromolecules 1987,20 1855

ISSN:0306-0012    

DOI:10.1039/CS9942300257

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Protein Structure from Linear Dichroism Spectroscopy and Transient Electric Birefringence Michael Bloemendal Department of Protein and Molecular Biology, Royal Free Hospital School of Medicine, Rowland Hill Street, London N W3 2PF, U.K. 1 Perspective Structural characterization of proteins plays an important role in fundamental science, medical research, and industry. X-Ray diffraction on protein crystals yields structures with a resolution better than 0.2 nm. However, many proteins, especially mem- brane-bound ones, resist crystallization. Moreover, the crystal structure may differ from the structure of the protein in its natural environment. Therefore, spectroscopic techniques to study biomolecules in solution are of great importance. Of these, nuclear magnetic resonance (NMR)* spectroscopy gives the most detailed results, but even for small molecules the spectra are complicated, and this problem becomes rapidly more severe with increasing size of the protein.Consequently, the applica- bility ofNMR is restricted to smaller proteins (MW <20 kDa).' In addition, both crystallographic X-ray diffraction and NMR experiments on proteins are time-consuming. Hence, these techniques are less appropriate for the comparison of a large set of mutants or closely related proteins, or for a single protein under different conditions. Moreover, the protein content in protein crystals is 50% or more and in NMR samples l-lOo/~, which is significantly higher than the physiological concen-tration for most proteins.Finally, these techniques have the disadvantage that relatively large quantities of proteins (5-10 mg at least) are required for accurate measurements. In this paper, two less well-known techniques to study protein structures in solution, linear-dichroism (LD) spectroscopy and electric-field induced transient birefringence (ETB), will be discussed. From the former can be obtained information on the orientation of chromophoric groups in molecules, on molecular characteristics such as shape, size, and electronic properties, and on binding parameters in molecular complexes. From ETB hydrodynamic and electronic parameters, aggregational state, and intramolecular flexibility can be deduced. Both techniques are comparatively fast, and use relatively small quantities (0.2-2 mg) of protein at low concentration. After a general Michael Bloemendal was born in Amsterdam, the Netherlands, in 1955.He obtained his B.Sc. in Chemistry, Biological Orientation (1977), and his M.Sc. (1980) and Ph.D. (1985) in Physical Chemistry (1985) at the Free University of Amsterdam. After postdoctoral research at the Hebrew University in Jerusalem, he was a recipient in 1987 of a C. & C. Huygens Fellowship from the Netherlands Organi- zation for Scientific Research to study the influence of the medium on the structure ofpro- teins with biophysical and ana- lytical-chemical methods. Cur- rently he is employed as senior research fellow at the Royal Free Hospital School of Medi-cine in London, where he is involved in the study of protein precipitation on biomedical materials.description of the principles of the techniques, their application for the study of a specific lens-protein, a-crystallin, will be discussed in detail. 2 Linear-Dic hroism Spectroscopy 2.1 The Concept In optical spectroscopy, light interacts with parts of molecules, the so-called chromophores. The dominant interaction is that of the electric field of the light with the local charges on the chromophore. Effectively, light is absorbed when its frequency matches the difference in energy levels between two electronic states (E =hv), and -less generally realized -when the inter- action yields an oscillating charge displacement on the chromophore.This displacement is represented by the so-called transition dipole moment (TDM). Due to its oscillating character, it has an orientation but no direction! Different electronic transitions each induce a distinct TDM with characteristic orientation in a chromophore. Figure 1 shows those for UV absorption by indole, toluene, and phenol, the functional groups of trypto- phan, phenylalanine, and tyrosine, respectively. As Figure 2 illustrates, charge displacement will be most effective, and hence absorption strongest, when the orientation of the electric field of the light coincides with that of the TDM. When the mutual angle P +45OT\TA Figure 1 (a) The major transition dipole moments (tf),L, and Lb,of indole according to Albinsson et a1.;2(b) TDM of phenol (R =OH) or toluene (R = CH,) according to Campbell and D~ek.~ Figure 2 Interaction between the electric field of light (<--->)and a transition dipole moment (<-> ):(a) maximum interaction; (b) no interaction; (c) intermediate interaction.* ETB = electric field induced transient birefringence; FEA = fluorescence emission anisotropy; IF =intermediate filament, IgG = immunoglobulin-G; LD = linear dichroism; NMR = nuclear magnetic resonance; TDM = transition dipole moment 265 CHEMICAL SOCIETY REVIEWS, 1994 is 90°,no charge displacement and no absorption will take place More generally, the absorption depends on both the size of the dipole moment, and the angle between the TDM and the electric field of the light This means that when the orientation of the TDM is kept constant, the absorption of linearly polarized light by the chromophore depends on the direction of the polariza- tion The difference in absorption of two perpendicular polar- ized light beams, dLDA,is called the linear dzchroism (LD) Hence, Detailed reviews on LD-spectroscopy are given by N~rden,~ tial direction of the molecules is obtained, where the extent of orientation depends on molecular properties like size and/or shape and, when electromagnetic fields are used, on dipole moments and polarizabilities or magnetic susceptibilities Since the intensity of the LD depends on the degree of orientation, this can be used to extract information on these properties from the LD measurements Although in principle LD can be measured on almost any commercially available CD spectrophotometer, generally a spe- cial device has to be mounted into the spectrophotometer to produce and maintain a net order of the molecules Charney5 (electric-field induced LD), Norden et a1 (especially LD on nucleic acids), and Bloemendal and van Grondelle7 (LD on proteins) The simplest way to measure LD is by inserting a polarizer in front of the sample chamber of a conventional absorption spectrophotometer However, most frequently a dedicated spectrophotometer is used that can generally be applied for both LD and circular-dichroism spectroscopy Details have been described 2.2 Orientation of the Sample The necessity of constant orientation of the TDM implies that the molecules must have a net order within the time frame of the measurements In most crystals and some biological systems, such as membranes, this is generally the case, but in a randomly oriented sample, as for example a protein in solution, special measures have to be taken in order to orient the molecules macroscopically Orientation of biological molecules has been obtained in electric and magnetic fields, by electrophoretic orientation, in stretched polymer-films, liquid flow, liquid crystalline phases, natural and artificial membranes, lamellae and lipids, by shearing concentrated solutions on optical flats, by drying them in salt solution, by so-called wet-spinning, and in squeezed polyacrylamide gels (for details and references, see Bloemendal and Van Grondelle7) While the first three approaches orient the molecules according to their electric and/ or magnetic properties, the others orient the particles with respect to their spacial symmetry axes A special orientation mechanism, the so-called photoselection, is possible for photo- reactive compounds, where illumination with linearly polarized light leads to oriented photoproducts The most commonly used orientation methods with some applications are collected in Table 1 It should be emphasized that none of the techniques mentioned produces fully ordered samples Rather, a preferen- Table 1 Some techniques for sample orientation in LD- spectroscopy and their biological applicationsa Technique Electric field Flow Stretched film Squeezed polymer gel Photoselection Shearing Drying in salt solution Wet-spinning Electrophoresis For details and references Brol Rep 1993 18 49 Some applications Chloroplasts, reaction centres, chlorophyll- protein interaction Nucleic acids, nucleic acid-protein complexes Nucleic acids DNA-protein interaction, coenzyme reonentation, proteins, pigment- complexes Photoreactive compounds, retinal, bacteriorhodopsin Nucleic acids Nucleic acids, reaction centres, pigment- protein complexes Nucleic acids Proteins see M Bloemendal and R van Grondelle Mol 2.3 Information from LD 2 3 1 General Consideration As explained in Section 2 1, LD is caused by the dependence of the absorption of linearly polarized light by a chromophore on the angle between the TDM of that chromophore and the direction of polarization This angle can be decomposed into the angle 5of the TDM with a certain axis in the molecule, and the angle between that axis and the direction of polarization The latter will depend on the mechanism and degree of orientation Mathematically this means that dLDA is a function (0of 4 and an orientation factor 4, For several orientation techniques and differently shaped mole- cules, expressions for f(4,+) have been derived 4-7 Different types of information obtained for biomolecules are given in Table 2, and will be discussed in the subsequent sections 2 3 2 Orientation of Chromophores Equation 2 implies that, when the orientations of the TDMs in the chromophore and those of the molecules in the sample are known, LD yields data on the orientation of the chromophores in the molecules This is illustrated in Figure 3 for the trypto- phans of bovine yII-crystallin, a 21 kDa monomeric lens protein containing 4Trp and 15 Tyr residues As Figure 1 shows, Trp has two dominating TDMs (called La and Lb) for UV-absorp- tion The (sub)spectra belonging to these TDMs are different and slightly shifted * By measuring LD as a function of wave- length information on the orientation of these two TDMS in y-crystallin, and thus on the average position of the tryptophans with respect to the molecular axes, can be extractedg By using light of different wavelengths, information can be obtained on the orientation of various chromophores An overview is given in Table 3 Table 2 Some applications of LD-spectroscopy on biological moleculesa Application Orientation of chromophores Global structure/electric parameters Identification and quantification Titration and binding Example y Crystallin structure (see Section 2 3 21,Reorientation of chromophores during reaction y-Crystallin structure (see Section 2 3 3) Retinal isomers in complex mixtures (see Section 2 3 4) Protein binding to nucleic acids Ligand binding to macromolecules (see Section 2 3 5) 0 For details and references see M Bloemendal and R van Grondelle MoI Biol Rep 1993 18 49 Figure 3 The orientation of the tryptophans (black) and tyrosines (dashed)with respect to the long axis (black) of bovine yII-crystallin (see Bloemendal et a1 9, Coordinates according to Wistow et al lo from Brookhaven Databank Table 3 Chromophores used in LD spectroscopy of proteins and their band positiona Chromophoric group Band position Aromatic residues 23CL300 nm Ligands/coenzymes/pigments Amide groups Carbonyls Ethylene and aromatic proton bonds 25CL900 nm 18OO-1600 cm-' 350CL3000 cm-', 1700-1500 1020-720 cm-' cm-' Nucleic acid interaction 240-350 nm For examples and references, see M Bloemendal and R van Grondelle, Mol Biol Rep 1993 18, 49 2 3 3 Global Shape, Electric Properties Alternatively, when the position of the TDMs in the chromo- phore and the conformation of the chromophores in the mole- cules are available, LD provides + (see equation 2) and hence information on the parameters that govern the orientation of the sample (like shape, size, dipole moments, and so on) Figure 4 shows the absorption and LD spectra of bovine yII- and yIVa- crystallin The X-ray structures of these closely related proteins have been determined at high resolution OP1 The primary structures are 80% homologous, and the tertiary structures show only minor differences l1 The protein molecules were oriented in a one-dimensionally squeezed polyacrylamide gel according to van Amerongen et a1 13---15 In this orientation technique the order of the molecules depends on their size (larger ones are better oriented) and shape (more spherical ones are less oriented) The absorption spectra look very similar, but the LD spectra are clearly different both in profile (the LD at 288 nm is positive for yII and negative for yIVa) and in magnitude From a combination of the LD spectra and the X-ray structures the 5 Ang I orientation factors (4 in equation 2) for 711 and yIVa were found to be 0 044 f 0 006and 0 018 f0 003, respectively This means that yII is much better oriented than yIVa, which is a surprising result for two so-closely resembling proteins A possible expla- nation is that, contrary to the situation in the crystal, the global shape of yIVa in solution is more spherical than that of 711 This interpretation is supportedg by size exclusion chromatography results,' which yield a significantly smaller apparent molecu-lar weight (and thus smaller excluded volume) for yIVa- cry stallin 2 3 4 Identification and QuantlJication The high sensitivity and specificity of the LD spectra of closely related compounds (see Figure 4)has been used to identify and quantify rare compounds in mixtures of closely related components * 2 3 5 Binding and Titration Unordered molecules do not yield LD (4 = 0 in equation 2) This has been used in binding studies by means of LD with liquid flow orientation Small molecules (even most proteins) are not oriented in a flow field Therefore, the appearance of their absorption bands in the LD spectrum of a macromolecule like a nucleic acid is an extremely sensitive probe for their binding This has been utilized for the calculation of equilibrium con- stants, the determination of binding sites, and the study of complex formation in titration studies (for references, see Nor- den et a1 ,6 and Bloemendal and van Grondelle') 3 Electric Field Induced Transient Birefringence 3.1 The Concept Particles with a permanent and/or inducible electric dipole moment can be oriented in solution by means of an electric field For optically anisotropic particles this orientation leads to a difference in the refractive indices of the solution parallel (n,) and perpendicular (nl) to the orientation direction, ' the so-called birefringence When the electric field is switched on, the mole- CHEMICAL SOCIETY REVIEWS.1994 -0.0005 260 280 300 320 Wavelength (nm) (a) \ 4 I 260 280 300 320 Wavelength (nm) (b) Figure 4 (a) LD and (b) absorption spectra of bovine yII (-) and yIVa (---) crystallin at identical concentration of protein (from Bloemendal et d9).cules become ordered and the birefringence (An = nlI-n,) is built up. When the field is switched off, the orientation and hence An decays due to the Brownian motion. This is illustrated in Figure 5 for a square pulsed electric field. The rate of orientation, and thus of the rise of An, depends on the hydrodynamic and electric properties of the particles, the steady state on the optical and electrical properties, and the (field-free) decay on hydrody- namic properties exclusively.The primary quantity obtained from the decay is the rota- tional diffusion coefficient, which can be used to get information on molecular dimensions (shape and size), aggregational state and intramolecular Molecular dimensions can, in principle, also be obtained from translational diffusion coeffi- cients as reflected in centrifugation, viscosity, electrophoresis, and so on. However, rotational diffusion coefficients are much more sensitive and hence are a better probe for these properties. 24 Rotational diffusion coefficients can also be deduced from fluorescence emission anisotropy (FEA) measurements.2 This technique is compared with ETB in Table 4. Although in principle FEA can be done on inherent fluorescent groups, like tyrosine or trytophan residues, often an exogenous fluorescent group has to be attached to the molecules to be probed.ETB does not use external labels, and the time domain in which the An i b Time Figure 5 Electric field-induced transient birefringence: the molecules (dashes), the electric field (E), and the birefringence (An) as function of time. The dashed areas S, and S, are the areas enclosed by the rise and decay curves, respectively. dynamics can be followed is not restricted by the life time of the fluorescent probe (usually < 100ns). It enables the simultaneous study of rotational processes of molecular groups, monomers, and aggregate~.~~J~ Moreover, in ETB there are no restrictions to the pH range usable, whereas most fluorescent labels are non- fluorescent at low pH.Self-evidently, binding of a probe group always bears the risk of affecting the natural conformation of the molecules under study. A disadvantage of ETB is that relatively high fields are used to orient the molecules. This may influence the conformation or integrity of the molecules (this can be checked, see e.g. van Haeringen et a1.20-22), and restricts the range of ionic strengths that can be used for solutions in ETB. FEA has also the advantage that its response is restricted to the molecules (or molecular groups) where the fluorescent label is bound. Hence, this technique is more selective than ETB. Consequently ETB results are sometimes less straightforward to interpret.Detailed reviews on (electric field induced transient) bire- fringence are given by Fredericq and Houssier19 and Charne~.~ Recent literature on ETB is summarized by Stellwagen26 and Curry and Krau~e.~’ Information on equipment can be found in these papers and that of van Haeringen et a1.20 Recently, an apparatus has been described, that may be used at physiological salt concentrations. 3.2 Information from ETB 3.2.1 Molecular Dimensions, Aggregation, and Flexibility The decay of the birefringence after the electric field is switched off reflects the rotational relaxation of the particles that cause the birefringence (see Figure 5). Analysis of the decay as function of time yields rotational relaxation times (T),which are related to the rotational diffusion coefficients.Asymmetric particles as well as mu1 ti-component samples yield mu1 ti- exponential decays. Computer programs like DISCRETE29 and CONTIN30 resolve the decays to provide the optimum number of relaxation times to describe the data. The use of these programs for ETB is discussed by van Haeringen et a1.20 Multiple relaxation is also found for particles that show internal flexibility. In that case some of the deduced relaxation times may reflect the orientation and disorientation of segments rather than molecules. Primarily, T depends on the ratio between the viscosity (7)and PROTEIN STRUCTURE FROM LD SPECTROSCOPY AND ETB-M Table 4 Electric-field induced transient birefringence (ETB) versus fluorescence emission anisotropy (FEA) for determination of rotational diffusion coefficientsa ETB FEA No external label Fluorescent label (often exogenous) Time domain upto microseconds Time domain < 100 ns No restriction on pH Usually not possible at low pH (High) electric field applied No external field applied Restricted ionic strength of No restriction to ionic strength of solution solution For discussion and explanation see text the absolute temperature (T),and on the cube of the longest axis, L, of the rotating group Expressions for the relation between 7 and the dimensions of the molecules have been derived for differently shaped particles Several of these are quoted in recent papers 21-24 26 27 When the relaxation is too fast to follow, the process can be retarded by increasing the viscosity of the solution, e g through addition of glycerol or sucrose, or by lowering the temperature Equation 3 allows correction of the 7 so-obtained to that at standard conditions In a studyZo of acid-induced structural changes of a mouse immunoglobulin G (IgG), several of the already-mentioned features of ETB are illustrated IgG is a Y-shaped molecule as shown in Figure 6 Rotational relaxation times at two pH values in water and glycerol-water (40% w/v) are given in Table 5 In water at pH 6 6 a single relaxation is found, that corresponds to the rotation of the IgG monomer The observed 72 of 151 ns agrees well with values obtained by other techniques 31 After correction to the viscosity of pure water, the glycerol-water measurement yields the same rotation time (this is an essential test to see whether the viscosity change affects the global structure), but also indicates a fast relaxation process (71 = 39 ns) This time describes the segmental motion of the two IgG Figure 6 Two models for IgG dimerization at pH 2 7 based on ETB measurements from van Haeringen et af 2o IgG monomers are represented by Y-shaped symbols BLOEMENDAL arms At low pH this flexibility decreases (7 increases), and a third relaxation time is apparent This slow time (440ns) is due to aggregation at low pH The ratio between 73 and 72 is about 3 Hence, equation 3 shows that the difference in length between the corresponding particles is 33 = 1 4 Also taking into account that the segmental flexibility is more or less retained, this result points to dimerization in one of the two forms given in Figure 6 The top arrangement would possibly involve considerable loss of segmental flexibility 3 2 3 Electric Properties The orientation of molecules in an electric field is caused by the torque exerted on their permanent and induced electric dipole moments It has been shown that in the absence of a permanent dipole moment (ppem)the rate of birefringence build-up, due to the induced dipole moment (wind) governed orientation, is equal to the rate of decay after switching off the electric field The presence of a permanent dipole moment slows down the rise of the birefringence l9 This can be used to estimate the relative contributions of ppermand plnd (called P/Q) to the orientation mechanism by taking the ratio between the areas excluded by the build-up curve and included by the decay curve, respectively (S, and S2in Figure 5) At low field strength where the birefringence is proportional to E2, S1/S2equals one for pure plnd caused orientation (P/Q = 0) For pure ppermruled orientation (P/ Q = co) it is four l9 The assumption is that the orientation directions favoured by pWrmand plnd are the same Equations for the dependene of the orientation on the strength of pWrm,plnd (reflected in the polarizability), and the electric field can be found in Fredericq and Houssier,lg and the references quoted there Most biomacromolecules have an ionic atmosphere of counter-ions These are polarized in an electric field The induced dipole moment due to this ionic cloud is generally much stronger than the intrinsic plnd In this case the induced dipole moment has a symmetry corresponding to the overall shape of the protein KOOijman et a1 23 have used S1/S2calculations in their study of the building units of the intermediate filament (IF) protein, vimentin The IF proteins are the major constituents of the cytoskeleton of most eukaryotic cells 32 Although widely stud- ied, there is no agreement on the size and structure of the primary units of these filaments23 32 Kooijman et a1 23 found a strong, although not exclusive, permanent dipole moment (S1/S2 = 3), which in combination with the observed multi- exponential decay indicated the presence of dimers, antiparallel tertramers, and hexamers (see Figure 7) The high value of S,/S2 is caused by the dipole moment of the dimers and hexamers 3 2 4 Optical Anisotropy The steady-state birefringence (An,) is less straightforward to interpret, since its value is determined by both the optical and electric properties of the molecules At low electric field, E, An, is proportional to the square of the external electric field strength (An, = &er,E2), &err being called the Kerr constant l9 Kerr constants of IgG at different pH are included in Table 5 The value at pH 2 7 is clearly higher than that at 6 6 Moreover, there Table 5 Rotational relaxation times (7,) and Kerr constants (KKerr)from ETB decays of IgG in water and water/glycerol (40% w/v) at two pHsa cmz/Vz)Solution PH TI(nS) 72(ns) 73@) KKerr(10-16 Water 66 151 f7 -45k01 G1 ycerol/water 66 39 2 158& 1 -NC Water (after I h) 27 NC NC NC 62k02 Water (after 24h) 27 -168f 11 411 f41 70f02 G1 ycerol/water 27 51&3 166 f8 438f 12 NC ~ Results from B van Haeringen et af 2o obtained by DISCRETE analysis of ETB decays Indicated uncertainties are standard deviations of the mean values in glycerol/water are corrected to standard conditions (25 "C viscosity of pure water) NC means not calculated 270 CHEMICAL SOCIETY REVIEWS, 1994 Figure 7 Building units of vimentin IFs indicated by SJS, calculation from ETB measurement^:^^ (a) dimer with dipole moment; (b) antiparallel staggered tetramer without dipole moment; (c) hexamer with dipole moment.Each arrow represents a vimentin monomer with electric dipole moment. is a further increase when the sample is kept at low pH for an extended period. Although separation of electric and optical effects is complicated, it is obvious that an instantaneous structural change followed by a slow process takes place. The latter might be the formation of aggregates as described in Section 3.2.1, whereas the fast process might be related to secondary structure changes that have been found immediately after decreasing the pH.33 4 Application of LD and ETB to Study the Structure of a-CrystalIin 4.1 Introduction a-Crystallin, one of the predominant lens proteins, has attracted renewed interest since the detection of its occurence in other tissues, and the discovery of its relation to heat-shock pro- tein~.~~?~~In lens a-crystallin contains 3&50 A- and B-type subunits (approximate ratio in bovine calf 3: 1) with a molecular mass of 20 kDa/subunit and 60% sequence homology.36 Despite extensive work by various groups, there is no consensus yet about its global shape and subunit-order.Some proposed a model in which the subunits are distributed in quasi-spherical or tetrahedral aggregates (see e.g. Tardieu et ~1.~~).Others also indicate more elongated structure^.^^^^^ Pure a-crystallin is readily isolated by size-exclusion chromatography of dissolved lenses (see Figure 8)39.We have studied the structure of a-crystallin by ETB and LD.' 5321,22 For LD measurements the protein molecules were oriented in one-dimensionally squeezed polymer gels, as des- cribed by Van Amerongen et al. 3-1 The orientation mechan- ism in these gels is illustrated in Figure 9. In the undeformed meshwork, all orientations of the embedded molecules are equally probable. When the network is deformed, the molecules show an average preference for the long deformation axis of the gel. It is obvious from Figure 9 that spherical particles will not be oriented at all in the squeezed gel, and that a sample of tetrahedral particles will show no net orientation. For orientation of a single chromophore in particles of rotational symmetry (disks, rods, ellipsoids) by this system, equation 2 become^:^^' 3-1 where A is the isotropic (normal) absorption by the chromo- phore.For a system with more than one chromophore, equation 4 has to be summed over all chromophores: 40 80 120 160 200 40 60 80 100 120 Figure 8 Isolation of a-crystallin: (a) Size-exclusion chromatography of dissolved lenses according to Vlaanderen et al.39The shaded area is used for analysis of total a-crystallin and for rechromatography. (b) Rechromatography of the shaded area of 8a to give subpools I-IV (from Van Haeringen et a1.22). Figure 9 The orientation of particles represented by arrows in (a) an undeformed and (b) a deformed polymer meshwork. where A,(A)/A(A)can be considered as the weight with which chromophore j contributes to the absorption at wavelength A, and the brackets (...) denote the average over all chromophores of typej in the sample. 6,is the angle between the TDM and the symmetry axis of the particles. The orientation function has been given by Van Amerongen et al.13,and depends on the shape of the molecules and the degree of squeezing. For positive 4, 4= 54.7" yields no LD; 6> 54.7" gives negative LD; and 5 < 54.7" positive LD. In fact equations 4 and 5 hold for any particle with an axis (the orientation axis) around which the distribution is at least statistically rotationally ~ymrnetric.~,~ PROTEIN STRUCTURE FROM LD SPECTROSCOPY AND ETB-M 4.2 Analysis of Total a-Crystallin Peak 4 2 1 Linear Dichroism The LD and absorption spectra from 250 to 320 nm of bovine a-crystallin are shown in Figure 10 The fact that any LD spectrum is found, immediately refutes the idea that a-crystallin is spheri- cally or tetrahedrally symmetric, since in that case no LD whatsoever would have been found (see Section 4 1) Appar-ently, at least part of the aggregates has a significant deviation from such symmetry Between 250 and 350 nm, protein absorp- tion spectra are essentially composed of the contributions of the aromatic residues tryptophan (Trp), tyrosine (Tyr), and phenyl- alanine (Phe) All these residues have their main TDMs in the plane of the aromatic ring The absorption spectrum of Trp has maxima around 270, 280, and 290 nm, that of Tyr a single maximum at about 275 nm, and that of Phe a series of maxima between 250 and 265 nm At its maximum the Trp absorption is about three and 25 times the maximal absorption for Tyr and Phe, respectively 40 The occurrence of these residues in a-crystallin is 1, 6, and 14 per aA-, and 2, 2, and 13 per aB- subunit 36 For an aggregate of 30 uA- and 10 aB-subunits this yields 50 Trp, 200 Tyr, and 650 Phe residues One could expect such a number of chromophores in so many subunits to have all possible orientations This would imply that absorption of parallel and perpendicular polarized light is almost equal From equation 1 it appears that for such a case no LD is observed Figure 10 shows not only that there is LD, but also that the LD spectrum resembles the absorption spectrum closely Even the fine structure of the Phe absorption spectrum (250-265 nm, 650 residues in the aggregate) shows up in the LD spectrum Apparently, there is a high structural order of the subunits within the aggregates and of the residues within the subunits, with a majority of the aromatic residues oriented in the same way This contradicts models that assume a random orientation of the subunits in the a-crystallin aggregate The fact that the LD is positive over the whole spectrum shows that the average angle 5 in equation 5 is smaller than 54 7" 0 0028 [ n, I I I h -10I 250 286 322 Figure 10 The absorption and LD spectra of bovine a-crystallin after orientation in a squeezed gel (from Bloemendal et afIs) The magnitude of the LD varies from batch to batch Values between 0 002 and 0 006 were found for the reduced dichroism, dLDA/(3dA), at 280 nm This value is 2 5 to 7 5% of the theoretical value of a rigid rod-like protein with all the planes of the aromatic residues parallel to the long axis For an aggre- gate of 30-50 subunits with about 900 aromatic residues this is a rather high value, again indicating a strong structural order of the subunits in the aggregates, and a significant deviation of spherical or tetrahedral symmetry for at least part of the molecules 4 2 2 Electric Field-induced Transient Birefringence The decay of the ETB of a-crystallin appeared to be bi- exponential with relaxation times (T)of 4 and lps, respectively BLOEMENDAL 27 1 Until a steady-state was reached the relative contribution of the slower time to the total decay of the birefringence increased, when longer electric pulses were used to orient the molecules In general there are three possible explanations for a bi-exponential decay (a) the rotation of a single type of molecules with asymmetnc ellipsoid shape,21 (b) the rotation of a single type of molecules that contain internal flexibility (see Section 3 2 l), and (c) a mixture of two or more species An asymmetric ellipsoid with T~/T~= 4would have a ratio of semi-axes of about 19 5 1 (see van Haeringen et a1 l) Such an extreme elongation is completely in disagreement with all other evidence Moreover, a single type of molecule is not likely to cause a change in relative contribution of the two times with duration of the electric pulse Internal flexibility of the a-crystallin aggregate cannot be ruled out completely However, electron microscopy pictures of U-crystallin give no indication whatsoever for a multi-lobal struc- ture or extensions that could cause internal flexibility Therefore the two times were interpreted as being caused by two different a-crystallin species If both types were spherical, the ratio of relaxation times would directly reflect their volume ratio, since in that case Tcorresponds to the cube of the radius (equation 3) Species with a volume ratio of four are separable by means of size-exclusion chromatography No such separation has been reported yet Moreover, the a-crystallin used for this study was taken from a single chromatographic peak (see Figure 8) Therefore it was concluded that a-crystallin as obtained from chromatography of dissolved lenses is composed of (at least) two differently shaped species This conclusion is supported by the observed dependence on the pulse duration of the relative contribution of the relaxation times Hydrodynamically smaller particles will orient faster Hence, at pulse lengths that do not orient all the particles, their contribution to the signal will be relatively large Increasing the pulse length will enhance the percentage of hydrodynamically larger particles that are oriented, and thus their contribution to the birefringence Evi- dence for two differently shaped a-crystallins was also found from electron microscopy 4.3 Analysis of a-Crystallin Subpools In order to study the structure of c cry stall in in more detail, the LD and ETB measurements were repeated on fractions of the total a-crystallin population This was achieved by taking subpools of the original chromatographic peak, as illustrated in Figure 8 Average apparent molar masses determined from comparison with reference proteins are given in Table 6 Since separation on a gel filtration column is in fact not based on differences in mass, but on variation in Stokes' radius,41 the latter are also included in this table Re-chromatography of the subpools did not lead to redistribution of molecular masses, which shows that the different a-crystallin species are not in reversible equilibrium with each other Analysis of the ETB decay curves yielded two relaxation times for all pools, again indicating at least two different species for each pool The faster of these times showed virtually no field strength and pulse duration dependence, suggesting that this time represents a well- defined particle The slower relaxation time, on the contrary, showed variation with pulse strength and/or duration Appar- Table 6 Average apparent molecular masses (Mapp)and Stokes' radii (Rs)of a-crystallin subpools determined from size exclusion chromatography" Pool Ma,,(kDa) R,(nm) I 1170 f33 98f07 I1 848 f25 87f06 I11 801 f23 85f06 IV 745 f22 83f06 0 For discussion and explanation see text ently this time is related to a mixture of particles, and the subpools are still too large for complete homogeneity.For pool I the contribution of the slower time (about 6 ps) to the total decay was 55%. For the other pools (5.0-5.5 ps) it was only 20%. Combining these relaxation times with the Stokes' radii obtained from chromatography (Table 6) yields molecular dimensions, provided that a certain shape is assumed for the m01ecules.~~~~~It appeared that the slow relaxation time can be related to cylindrically shaped particles with axial ratio between 3 and 6, and a diameter between 9 and 16 nm. In electron microscopy pictures of a-crystallin, asymmetric particles with axial ratios between 2 and 4 and diameters between 10 and 14 nm can be discerned among spherkal ones.2 The fast time could be related with more or less spherical molecules (axial ratio z 1) and diameters of 13, 14, 15, and 23 nm for pools IV, 111, 11, and I, respectively (those of pool I show some deviation from sphericality).Reduced LD at 280 nm, the ratio between the permanent and induced dipole contribution to the orientation (P/Q, see Section 3.2.3), and Kerr constants of the four pools as a function of their Stokes' radius are depicted in Figure 11. All quantities decrease from pool I to pool IV. The relatively high value of the LD of pool I can be explained by the larger contribution of asymmetric particles to this pool in comparison with the others (55% vs. 20%, see above).However, since no signifcant difference in the percentage of asymmetric particles is found for pools 11,111,and IV, the further decrease cannot be explained in this way. Possibly the spherical a-crystallin molecules are not completely symmetric and are still oriented in the squeezed gel. In that case the size difference will cause the variation in LD. Alternatively there is a steady decrease in order of the subunits throughout the chromatographic peak of a-crystallin. Interpretation of P/Q values and Kerr constants is less straightforward, since they are composed of contributions from both the slow and fast decaying particles. The P/Q value of pool IV is zero, which shows the absence of a permanent dipole moment for the particles of this pool.For pools I11 and I1 some permanent dipole contribution is found, whereas it is profound for pool I. Recalling that even pools I1 to IV contain 20% asymmetric particles, this shows that there is a decrease of permanent dipole moment throughout the chromatographic peak even within the asymmetric particles. This might be caused by a decrease in subunit order, which was also inferred from the decrease in LD. The observed Kerr constants also support this conclusion. The steady decrease from pool I to pool IV while the concentration of fast relaxating particles is similar for pools I1to Figure 11 Reduced linear dichroism (dL,A/3dA) at 280nm, P/Q values, and Kerr constants verms the Stokes' radius as determined from size- exclusion chromatography for a-crystallin subpools denoted by Roman numerals.10- 0.5 9- 8- 0.4 7- T T2 4 6-5-4- 0.3 0.2 3- 0.1 0.0 CHEMICAL SOCIETY REVIEWS, 1994 IV points to different electric properties and/or optic anisotropy for both the symmetric and asymmetric particles. 4.4 Conclusion: A New Model for a-Crystallin The combined results lead to the model for a-crystallin that is illustrated in Figure 12. a-Crystallin consists of two types of particles, a more or less symmetric one with a radius ranging from 13 to 23 nm, and an elongated one with axial ratio between 3 and 6 and diameter between 9 and 16nm. The concentration of asymmetric particles is larger at the beginning of the chromato- graphic a-crystallin peak.There exists a significant order of the subunits within the molecules and of the aromatic residues within the subunits, the average angle between these residues and the long axis of the molecule being less than 54.7'. This order decreases throughout the chromatographic peak. Part of the molecules contain a strong permanent dipole moment. As this is especially found for the fraction that contains more elongated particles, these asymmetric structures may be particularly polar. However, even for the asymmetric molecules a decrease in polarity is found throughout the chromatographic peak. Acknowledgement Dr. B. Van Haeringen is kindly acknowl- edged for his helpful remarks and critical reading of the manus- cript, and the IRC for their financial support.h \\ As----. Figure 12 A model for a-crystallin. I PROTEIN STRUCTURE FROM LD SPECTROSCOPY AND ETB-M 5 References 1 G Wagner, S G Hyberts, and T F Havel, Ann Rev Biophys Biomol Struct , 1992,21, 167 2 B Albinsson, M Kubista, B Norden, and E W Thulstrup, J Phys Chem , 1989,93,6646 3 I D Campbell and R A Dwek, ‘Biological Spectroscopy’, Benja- min/Cummings Pub1 ,Menlo Park, CA, 1980, p 29 4 B Norden, Appl Spectr Rev, 1978,14, 157 5 E Charney, Quart Rev Biophys , 1988,21, 1 6 B Norden, M Kubista, and T Kurucsev, Quart Rev Biophys, 1992,25, 51 7 M Bloemendal and R van Grondelle, Mol Biol Rep , 1993,18,49 8 K Fahmy, F Siebert, and P Tavan, Biophys J , 1991,60,989 9 M Bloemendal, J A M Leunissen, H van Amerongen, and R van Grondelle, J Mol Biol , 1990,216, 181 10 G Wistow, B Turnell, L Summers, C Slingsby, D Moss, L Miller, P Lindley, and T Blundell, J Mol Biol , 1983, 170, 175 11 H E White, H P C Driessen, C Slingsby, D S Moss, and P F Lindley, J Mol Biol , 1989,207,217 12 T Blundell, P Lindley, L Miller, D Moss, C Slingsby, I Tickle, B Turnell, and G Wistow, Nature, 1981, 289, 771 13 H van Amerongen, H Vasmel, and R van Grondelle, Biophys J , 1988,54,65 14 H van Amerongen, M E Kuil, F van Mourik, and R van Grondelle, J Mol Biol , 1988,204, 397 15 M Bloemendal, H van Amerongen, H Bloemendal, and R van Grondelle, Eur J Brochem , 1989,184,427 16 J G Bindels, J Bours, and H J Hoenders, Mech Ageing Dev ,1983, 21, 1 17 R J Siezen, E Wu, E D Kaplan, J A Thomson, and G B Benedek, J Mol Biol , 1988,199,475 18 J Horwitz and J Heller, J Biol Chem , 1973,248, 1051 19 E Fredericq and C Houssier, ‘Electric Drchroism and Electric Birefringence’, Clarendon Press, Oxford, 1973 20 B van Haeringen, W Jiskoot, R van Grondelle, and M Bloemen-dal, J Biomol Struct Dvn , 1992,9,991 21 B van Haenngen, D Eden, M R van den Bogaerde, R van BLOEMENDAL Grondelle, and M Bloemendal, Eur J Biochem , 1992,210,211 22 B van Haeringen, M R van den Bogaerde, D Eden, R van Grondelle, and M Bloemendal, Eur J Biochem , 1993,217, 143 23 M KOOiJman, M Bloemendal, H van Amerongen, P Traub, and R van Grondelle, J Mol Biof , 1994,236, 1241 24 J G Garcia de La Torre and V A Bloomfield, Quart Rev Biophys , 1981,14,81 25 S B Brown, ‘An Introduction to Spectroscopy for Biochemists’, Academic Press, London, 1980, chapter 3 26 N C Stellwagen, Biopolymers, 1991,31, 1651 27 J F Curry and S Krause,J Phys Chem, 1992,96,4643 28 D Porschke and A Obst, Rev Sci Znstrum , 1991,62, 818 29 S W Provencher, Biophys J, 1976, 16,27 30 S W Provencher, Comp Phys Commun , 1982,27,229 31 J Yguerabide, H F Epstein, and L Stryer, J Mof Biof , 1970,51, 573 32 P M Steinert and D R Roop, Ann Rev Biochem, 1988,57, 593 33 W Jiskoot, M Bloemendal, B van Haeringen, R van Grondelle E C Beuvery, J N Herron, and D J A Crommelin, Eur J Biochem , 1991,201,223 34 D Jones, R H Russnak, R J Kay, and E P M Candido, J Biol Chem , 1986,261, 12006 35 S P Bhat, J Horwitz,A Srinivasan,andL Ding,Eur J Biochem, 1991, 102, 775 36 H Bloemendal, CRC Crit Rev Biochem , 1982, 12, 1 37 A Tardieu, D Laporte, P Licinio, B Krop, and M Delaye, J Mol Biol , 1986, 192, 71 1 38 R C Augusteyn, E M Parkhill, and A Stevens, Exp Eye Res, 1992,54,219 39 I Vlaanderen, R van Grondelle, and M Bloemendal, J Liq Chromatogr , 1993, 16, 367 40 ‘Handbook of Biochemistry and Molecular Biology, Proteins’, ed G Fasman, 3rd Edn , CRC Press Cleveland, 1976, Vol 1, pp185-195 41 L C Davis, J Chromatogr Sci, 1983,21,214 42 C Sadron, Progr Biophys Biophys Chem , 1953,3, 237

ISSN:0306-0012    

DOI:10.1039/CS9942300265

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Propagation of Interfacial Waves in Microgravity FranGois Quirion", Marie-Claude Asselin, and Guy G. Ross" INRS-Energie et Materiaux, I650 Montee Ste- Julie, Varennes QC, Canada J3X IS2 1 Introduction It is quite difficult to imagine how the world, as we know it, would react to the disappearance of gravity but one thing is sure, the most spectacular events would arise from the redistribution of liquid in its environment. For example, water wets rock and gravity keeps it embedded into the ocean. In the absence of gravity, the ocean would climb up the cliffs and cover every surface that it can wet. Ships, which are often coated with water repellent, would be trapped inside gigantic air bubbles wander- ing within the liquid volume. These events are unlikely to occur on earth, but similar phenomena, based on the same physico- chemical principles, do occur in a reduced gravity environment such as the one provided by the space shuttle.It is now generally accepted that the investigation of inter- facial properties is essential to the development of materials science either on earth or in the absence of gravity. As a result, the Microgravity Sciences and Application Bibliography1 reports many more papers in the section covering Fluids, Interfaces, and Transport than in its section on Metals, Alloys, and Composites. In 1992, Dr. Rath, editor of the book Microgravity Fluid Mechanics,2 stressed the importance '. . . to get a wider basis and a precise comprehension of the different and interesting micro- gravity fluid mechanics phenomena.' The same year, an article in Science reported3 that several committees of microgravity researchers have agreed that the field of microgravity materials science needs '...to turn away from the current thrust of experimentation in orbit -into materials and processes that could quickly be commercialized -and dive unashamedly into good, old-fashioned basic research.' The propagation of a wave at the liquid-air or liquid-liquid interface is one of these old-fashioned topics that lacks infor- mation in the absence of gravity. It is relevant to the manage- ment of fl~ids~,~ as well as for the preparation and processing6 of Franqois Quirion Marie-Claude Asselin Guy G. Ross metals, glasses, polymers and their alloys, composites, and foams which are processed through the liquid state. Perturbing interfacial waves often occur during aircraft and shuttle experiments and they are also expected aboard the space station.' Let's just mention that they are a consequence of the residual accelerations, also known as g-jitter, arising from aerodynamic forces, routine crew activity, and equipment operation.The perturbing and wetting forces both determine the confi- guration of liquids in partly filled containers and it becomes important to understand how they relate to each other in order to predict the behaviour of liquid samples subjected to g-jitter in microgravity. To do so, the propagation of sinusoidal waves was investi- gated for liquid-air systems both on earth and during parabolic flights.Reduced gravity was also simulated using Plateau's neutral buoyancy technique.8 The dispersion relation9 was used to describe the movement of the liquids inside the container. A previous investigationlo showed that wetting governs the confi- guration of liquids in low gravity and one of the objectives of this work is to evaluate its effect on the propagation of interfacial waves. 2 Theoretical Background This section summarizes the theoretical background necessary to understand the phenomena described in this article. 2.1 Wetting In the absence of gravity, the mass distribution of a liquid within a container will depend on its ability to wet the container. * To whom correspondence should be addressed. Franqois Quirion received his Ph.D.in Physical Chemistry at the Universitk de Sherbrooke in 1986. He also spent one year at the University of Tennessee, working on light and neutron scatterivg of micellar solutions. He is now professor at the INRS-Energie et Matiriaux. His research interests are related to the physical chemistry of interfaces, particularly interfacial tension and wetting phenomena applied to liquids in micro- gravity, polymer mixtures, and colloidal suspensions. Marie-Claude Asselin is a third year Physics student at the Universiti de Sherbrooke. She participated in the ground and KC-13.5 experiments. Guy G. Ross receivedhis Ph.D. in Physics at the INRS-Energie et MatPriaux in 1985. He spent two years at the KFA-Juelich in Germany and three months at the Academia Sinica in China working on thermonuclear fusion.He is now professor at the INRS-Energie et Matkriaux. His research interests are related to the ion-matter interaction, ion beam modification of mater-ials, and thermonuclear fusion. 275 Wetting is characterized by the contact angle, 0, that originates from the balance of the interfacial tension at the wall-liquid, liquid-air, uL/A, and wall-air, uWIA,interfaces The correla- tion between the contact angle and the interfacial tensions is often referred to as the Young equation A well-known manifestation of wetting is the rise of a liquid on the surface of a material This forces the liquid interface to curve and the distance over which the curvature extends can be approximated by the capillary length, I, In our case, the liquid is between two plates and where dp is the density difference between the fluid phases and g is the Earth gravitational acceleration For water in a plexiglass container with a contact angle of 55", I, is about 2 mm This implies that the curvature due to the capillary rise between the plates of the box, should not affect the interface at distances greater than about 2 mm from the walls However, a hundred- fold decrease of the gravitational acceleration will produce a tenfold increase of the capillary length to around 20 mm 2.2 Capillary Force At low gdp, the capillary length becomes large so that the radius of curvature of the meniscus, t/2cos (0), is constant over the thickness of the box, t For a box of length L, this creates a capillary force of magnitude where 2(L + t) is the perimeter of the solid-liquid-air interface Hence, a perturbation will displace the liquid only if it generates a force, ma, greater than the capillary force In other words, a perturbation will change the configuration of the contained liquid only if it generates an acceleration, where m is the mass of liquid 2.3 Interfacial Waves The equation generally used to describe a sinusoidal wave at thc interface of two fluids is the dispersion relation Although this model does not account for the effect of wetting, it has been applied with success to the investigation of contained fluids k(Pl -P2) + &I (5)4= k,,[p,cotanh(k,h,) + p,cotanh(k,h,)l' where k,= 2r/A, and a,,= 2x8, with A, and 8, being the wavelength and frequency of the resonance wave corresponding to mode n For a box of length L, A, is 2L/n, and u and p are respectively the interfacial tension and the density of the immis- cible fluids in contact For all the experiments described here, the containers were always filled with either coloured water and air or coloured water and a solution of heptane and carbon tetrachloride The interface was set at half height so that h, = h, = h and thus equation 5 reduces to where AG is the apparent gravity also referred to as the Atwood number, CHEMICAL SOCIETY REVIEWS, 1994 with subscript 1 standing for the lower and subscript 2 for the upper phase Equation 6 contains a gravity and an interfacial tension contribution which will be referred to as the gravitational and interfacial waves, respectively At high frequencies the inter- facial waves are independent of gravity, and their investigation -for instance by light scattering12 -has proven an effective method for the determination of the interfacial tension of liquid-liquid and liquid-air systems The investigation of low frequency interfacial waves is poss- ible only when the apparent gravity or the gravity itself becomes small, for instance with immiscible liquids of similar density or during a curved trajectory where the centrifugal force balances gravity as in aircraft, rocket, and shuttle experiments Figure 1 compares the patterns expected for the low frequency waves corresponding to mode 1 to 4 with those observed on earth for the gravitational waves in a half-filled container Notice that for all these modes, there is a node in the middle of the box The main difference between the real and expected patterns is due to the wetting of the liquid which causes distortion of the interface near the walls The acceleration caused by a sine wave of amplitude A is f Aa2, and the variation of the amplitude of the wave at the interface, a, with the amplitude of the imposed sine wave, (dai dA), is defined here as the propagation efficiency At the resonance modes, a remains finite because of the viscosity of the liquid, 71 Kamotani and Ostrachs suggested a reciprocal square root dependence for the propagation efficiency with respect to the viscosity of Newtonian liquids 3 Experimental Part 3.1 Propagation of Perturbations The container was a thin plexiglass box having a length of 95 mm and a height of 44 mm The liquid-air and liquid-liquid inter- faces were positioned at half-height The width was 10 mm for the ground experiments and 15 mm for the parabolic flight experiments The perturbation imposed was a sine wave fed to a x-y plotter on which the box was fixed It was parallel to the liquid-air or liquid-liquid interface with an amplitude between 2 and 60 mm The resonance modes were identified at the frequency for which the amplitude of the extrema (see Figure 1) was maxi- mum The amplitude of the wave at the interface, a, was determined for at least ten values of the amplitude of the imposed wave, A, and the relation was linear with a slope, (da/ dA), corresponding to the propagation efficiency 3.2 Physicochemical Properties of the Liquids Used The interfacial tension was determined by the Du Nouy ring method The aqueous mixtures of Polyethylene Glycol 400 PEG 400 used for the effect of viscosity all behaved as Newtonian liquids, I e the shear stress is a linear function of the shear rate in the range of shear rate studied The density was measured with a vibrating tube densimeter l3 The contact angle was determined by direct visualization of a drop of liquid standing on plexiglass Table 1 summarizes the physicochemical properties of the liquids used to calculate the resonance fre- quencies from equation 6 Heptane, C,, and carbon tetrachloride, CCl,, are completely miscible in each other but immiscible in water They were used to simulate a reduced gravity environment The density of the mixtures of C, and CCl, is compared with the density of water in Figure 2 The arrows indicate the systems that were investigated here The apparent gravity of these systems, see equation 7, is calculated from the density of the organic mixture and the density of water When equilibrated with water, the mixtures of heptane and carbon tetrachloride wet the plexiglass walls with an angle smaller than 20" Thus, when water is on top, the system simulates wetting conditions, while it simulates non-wetting PROPAGATION OF INTERFACIAL WAVES IN MICROGRAVITY-F.QUIRION, M.-C. ASSELIN, AND G. G. ROSS Table 1 Physicochemical properties of the liquids used in the course of the present investigations. All data for water imply the presence of a colorant. v PU U e Yo (mPa.s) (kg/m3)"1("1 Figure 1 Comparison between the expected configuration of the liquidfor mode one to four with the patterns observed at the liquid-air interface on the ground. The right and the left pictures represent the interfaceat time 1and + 1/28, where 8is the frequency of the mode. Water 100 0.90 997 0.065 55 PEG 400 25.7 3.28 1044 0.040 49 48.3 13.3 1085 0.033 31 64.9 33.1 1105 0.041 38 c,cc1, 100 I00 679 1584 0.018h 0.018" 0' 0 * ce Densities were obtained at 25 "C.Interfacial tension between coloured water and mixtures of heptane and carbon tetrachloride. ' The liquids spread on the plexiglass surface. Relative concentration of heptane 1.0 0.8 Q.6 0.4 0.2 1.4 E 'jj 1.0 C2 0.8 I,IIIIII0.6 0.2 0.4 0.6 0.8 1.o Relative concentration of CC14 Figure 2 Density of the organic mixtures of heptane, C,, and carbon tetrachloride, CC1,. The horizontal line is the density of water in equilibrium with the organic mixtures and the arrows indicate the systems investigated during this work. The dotted line represents the isodensity condition where the organic phase (shaded) becomes heavier than water.conditions when water sits at the bottom with respect to the earth acceleration (see Figure 2). 3.3 Parabolic Flight Experiments It is important to realize that in aircraft and shuttle experiments, the gravitational force does act on the systems. For instance, an object orbiting at an altitude of 1600 km still senses 64% of the gravitational attraction of the Earth,14 i.e. 0.64 g. This is much higher than the 0.001 g encountered aboard spaceships. The reduced gravity arises from a balance of the centrifugal and gravitational forces that act on the vehicles as they orbit around the earth. The experiments presented here were performed aboard NASA's KC-135 aircraft, a modified Boeing 707, based in Houston, USA. During a typical parabola the aircraft under- goes about 25 seconds of reduced gravity.Because of pilot's manoeuvres and climatic disturbances, the level of gravity usually fluctuates in the range f0.01 g. The acceleration perpendicular to the liquid-air interface, z and parallel to the interface, y,are shown in Figure 3 along with a typical trajectory of the aircraft during a parabola. In the absence of the imposed sine wave, the interface remained almost flat so that the residual accelerations originat- ing from the aircraft could not be responsible for the large interfacial deformations associated to the resonance modes. 4 Results and Discussion Three types of experiments were conducted during this investi- gation. Ground experiments with liquid-air systems where the CHEMICAL SOCIETY REVIEWS, 1994 (c) -0.02I A -0.02I I 0 40 Time (s) Figure 3 Accelerations obtained during a typical parabola at the centre of gravity of the aircraft: (a) trajectory of the aircraft; (b) acceleration level along the z-axis (topbottom) including a magnification of the reduced gravity region; (c) acceleration along the y-axis (right-left).capillary length, 2 mm, was always smaller than the thickness of the box, 10 mm. Ground simulations of reduced gravity with liquid-liquid systems where the capillary length was between 2 and 4 mm -still smaller than the thickness. Finally, parabolic flight experiments with liquid-air systems subjected to residual accelerations of about 0.01 g where the capillary length increases to about 20 mm while the thickness of the box was 15 mm.4.1 Propagation at the Liquid-Air Interface on Earth In these systems, the capillary length is smaller than the thick- ness so that wetting should not much affect the propagation of the waves. Figure 4 shows that the experimental resonance modes for water and aqueous solutions of PEG 400 in equili- brium with air compare very well with the frequencies calculated from equation 6. For a liquid-air interface on Earth, these resonances are mainly due to the gravity contribution and they are not expected to depend much on the surface properties. The results also show that the resonance modes do not depend on the viscosity, at least for the Newtonian liquids studied. I I Ilm 1 I I I 10 20 30 40 Viscosity (mPa-s) Figure 4 Experimental resonance frequency for modes 1,2, and 3 at the air-liquid interface as a function of the viscosity of the liquid.Half- filled thin box (L= 0.095, H = 0.044, W= 0.010 m). 2.5 3 12.0 rn '0 1.5 t '"(!I ' 0!2 ' 0!4 ' 0:s ' 0:s ' IlO ' 112 [viscosity (rn~aos)]-1'2 Figure 5 Reciprocal square root dependence of the propagation efficiency with the viscosity of the liquid for the propagation of mode 2 at the liquid-air interface. Half-filled thin box (L= 0.095, H = 0.044, W= 0.010 m). The propagation efficiency was determined experimentally for mode 2 as a function of viscosity and the results are plotted in Figure 5. The reciprocal square root dependence reported by Kamotani and Ostrachs also applies to our data over the viscosity range investigated.4.2 Reduced Gravity Simulation with Immiscible Liquid Mixtures The dispersion relation also applies to liquid mixtures having different densities and it can be used to calculate their resonance modes. The major advantage of these systems is the possibility to reduce, at will, the gravity contribution to the resonance modes. In such immiscible liquid mixtures, the gravitational force may be partially or totally balanced with the hydrostatic force. Mathematically, a decrease of g or AG, which is related to the density difference between the immiscible liquids, should affect the gravitational contribution in the same way (see equation 6). As the apparent gravity decreases, the resonance modes become dominated by the interfacial tension contribution.This is shown in Figure 6 where a decrease of the gravitational contribution also results in a decrease of the frequency of the resonance modes. This explains the impression that liquids in space are in slow motion; they respond to lower frequencies. At the lower values of AG, the experimental frequencies seem systematically higher than the calculated ones. These systems correspond to capillary lengths around 6 mm so that wetting forces arising from the 10mm thick boxes may begin to interfere with the dispersion of the waves. As expected, the propagation efficiency decreases with the apparent gravity of the systems. A straightforward prediction is that the propagation of a mode will vanish when the density Apparent gravity (9) Figure 6 Experimental resonance frequency of modes I to 4 at the liquid-liquid interface with respect to the apparent gravity.Half-and- half mixtures of water and organic liquid consisting of heptane and carbon tetrachloride in a thin box (L= 0.095, H = 0.044, W= 0.010 m). PROPAGATION OF INTERFACIAL WAVES IN MICROGRAVITY-F. QUIRION, M.-C. ASSELIN, AND G. G. ROSS difference becomes zero. In such isopycnic (or isodensity) con- dition, there is no density gradient upon which the acceleration can act so that the system behaves as if its mass was homo- geneously distributed inside the container. The propagation efficiency does in fact extrapolate to zero when the apparent gravity is zero as seen in Figure 7.0.4 I I I1 wetting / 3 Apparent gravity (9) Figure 7 Linear dependence of the propagation efficiency of mode 2 at the liquid-liquid interface with respect to the apparent gravity. Half- and-half mixtures of water and organic liquid consisting of heptane and carbon tetrachloride in a thin box (L= 0.095, H = 0.044, W = 0.010 m). At high C, content, the organic phase stands on top of the water phase so that the interface is curved downwards, as opposed to high CCI, content where the interface is curved upwards. Two systems with downward curvature were investi- gated and their frequencies and propagation efficiency are the same as their parent systems with upward curvature.In the systems studied, the gravitational contribution is still quite high (> 0.05 g) and it may overwhelm the wetting contribution arising from the thickness. Much lower levels, 0.01 g, are attained in the course of parabolic flights experiments, where the liquid-air interface may be investigated directly. 4.3 Perturbation in Microgravity Parabolic flights experiments allowed us to investigate the propagation of interfacial waves at the liquid-air interface. However, the fluctuation in the level of acceleration during a parabola also causes the resonance frequency to oscillate around the constant frequency of the sine wave imposed on the system. Fortunately, the effect of the residual acceleration decreases rapidly for higher modes.For instance, at 0.01 g, the resonance frequency of the first mode may increase by 100% while it will cause only a 10% increase for the third mode. The patterns of modes 2 and 4as observed in reduced gravity are compared with their corresponding ground patterns in Figure 8. The first observation is the distortion of the interface which is affected by wetting forces. The extension of the curva- ture to larger distances is a consequence of the increase of the capillary length as discussed earlier. The curvature also extends in the direction of the thickness of the boxes so that it generates a capillary force that stabilizes the interface. Figure 9 compares the response of four liquid samples to a given perturbation in the course of a parabola.The thickness of the boxes increases from 5 to 20 mm and they all contain a 50/50 mixture of coloured water and PEG 400. The curvature of the liquid-air interface in the thinner box, top left, extends to about 20 mm, in accordance with the increase of the capillary length due to the reduced gravity. The picture was taken just after a perturbation parallel to the interface and one can see that the liquid in the thinner box did not respond much while that in the thicker box, bottom right, did. This is in accordance with the occurrence of a stabilizing capillary force in thinner containers. Once the liquid starts moving, it does so with either an advancing or a receding contact angle with the solid surface. v v v Figure 8 Comparison between the ground configuration of the liquid for mode two and four with the patterns observed in the course of KC-135 parabolic flight experiments.The right and the left pictures represent the interface at time f and t + 1/29 where 9 is the frequency of the mode. The box dimensions are L = 0.095, H = 0.044, W = 0.015 m. Figure 9 This picture shows the effect of thickness on the response of the liquid-air interface to a perturbation. The thicknesses are 0.005, 0.010, 0.015, and 0.020 m for the top left, top right, bottom left, and bottom right boxes, resDectivelv. The other dimensions of the boxes are L = 0.22 and H = 0:12 m. * These are different from each other, the advancing always being greater (or equal for ideal systems) than the receding, as illustrated in Figure 10.This is also true for the liquid that flows up and down in our flat boxes due to the imposed sine wave parallel to the interface. Hence, the change of contact angle, 8, upon rising or falling will affect the stabilizing capillary force (see equation 3). Figure 10 Picture of the liquid-air interface following a sudden right to left perturbation. The resulting advancing, OA, and receding, OR, contact angles of the liquid between the plates are also shown. In our case, the liquid wets the wall and the curvature is downward with respect to gravity. Thus, the rise of the liquid occurs with an advancing contact ang!e while its fall occurs with a receding contact angle. According to equation 3 and knowing that the advancing contact angle is always greater than the receding, the stabilizing capillary force will be smaller when the liquid rises than when it falls.The consequence is that the liquid will rise on both sides of the box as shown in Figure 11. This results in a stable configuration where the interfaces are almost perpendicular to the imposed perturbation. This stresses the importance of the contact angle on the configuration of con- tained liquids and the propagation of perturbations within such systems. Another phenomenon often observed in the course of a parabolic flight is the complete wetting of the container by the liquid which results in the encapsulation of air. For this to occur, the force caused by the acceleration due to the perturbation must be greater than the capillary force acting against the rise of the liquid on the side of the box. Figure 12 shows an encapsulation Figure 11 Sequence of events during a parabola which shows the rise of the liquid on both sides of the container leading to a U-shaped configuration.The box dimensions are L = 0.095, H = 0.044, W=0.015 m. CHEMICAL SOCIETY REVIEWS, 1994 Figure 12 Sequence of events during a parabola which shows the encapsulation of air due to a y-perturbation. The box dimensions are L = 0.095, H = 0.044, W = 0.015 m. as it progresses from a flat to a circular interface through the sloshing of the liquid on the side of box. For our experiment, placed at the centre of gravity, the typical y-perturbations were too small to cause encapsulation (f0.005 g). However, when a sine wave parallel to the interface was imposed, encapsulation was observed for accelerations (h2) between 0.01 7 and 0.026 g.Assuming that these accelerations act on the total mass of liquid in the container, equation 4leads to a contact angle for coloured water oc plexiglass in the range 66" to 52", respectively, in good accordance with the observed contact angle of 55" reported in Table 1. The encapsulation of air into the liquid phase generates a circular interface that may resonate, or not, with the pertur- bation. The two situations are shown in Figure 13 where, at the lowest frequency, one can clearly observe a wave rolling around the bubble. This is not the case at the higher frequency where the bubble is broken into smaller bubbles that remain evenly distributed within the box.5 Conclusions It was shown that the dispersion relation applies to the propaga- tion of perturbations at the liquid-air and liquid-liquid interface both on the ground and in the reduced gravity environment provided by parabolic flight experiments. The increase of the capillary length in microgravity enhances the wetting forces resulting in the deformation of the interface that may interfere with the usual propagation of a perturbation. For instance, it is suggested that the force generated by a perturbation must be greater than the stabilizing capillary force in order to displace the contained liquid. Under such conditions, the liquid may encapsulate the air and thus form a circular interface that may or may not resonate with the perturbation frequency.Acknowledgments. The authors wish to thank M. Jacques Pelle- tier for technical support and they express their sincere gratitude to Lawrence Vezina, CSA, and Glenn Campbell, CSA, whose advice and help were greatly appreciated. The research des- cribed in this paper is part of a project financed by the Canadian Space Agency. PROPAGATION OF INTERFACIAL WAVES IN MICROGRAVITY-F. QUIRION, M.-C. ASSELIN, AND G. G. ROSS 28 1 Figure 13 Comparison between a circular interface that (a) resonates with the imposed perturbation and (b) does not resonate with the imposed perturbation. The box dimensions are L = 0.095, H = 0.044, W= 0.015 m.6 References 1 ‘Microgravity Science and Application Bibliography’. 1991 Revision, NASA Technical Memorandum # 4348. 2 ‘Microgravity Fluid Mechanics’, ed. H. J. Rath, Springer-Verlag, Berlin-Heidelberg, 1992. 3 Ivan Ammato, ‘Microgravity Materials Science Strives to Stay in Orbit’, Science, 1992, 257, 882. 4 For instance see: R. J. Hung, C. C. Lee, F. W. Leslie, Advances in Space Research, 1991, 11, 201-208 & 209-216; D. Langbein, Microgravity Science and Technology, 1992,2,73-85; M. Weislogel, ‘Forum on Microgravity Flows’, American Society of Mechanical Engineers, New York, 1991, pp. 11-13. 5 Y. Kamotani and S. Ostrach, J. Thermophysics and Heat Transfer, 1989, 1, 83. 6 For a review see: ‘Fluid Sciences and Materials Science in Space’, ed.H. U. Walter, Springer-Verlag, Berlin-Heidelberg, 1987. 7 E. S. Nelson, ‘An Examination of Anticipated g-Jitter on Space Station and its Effects on Materials Processes’, NASA Technical Memorandum # 103775, April 1991. 8 J. A. F. Plateau, ‘Statique expkrimentale et theorique des liquides soumis aux seules forces molkculaires’, Gauthier-Villars, Paris, 1873. 9 K. 0.Mikaelian, Phys. Rev., 1990, A42,721 1. I0 F. Quirion, in proceedings of Spacebound 92, Canadian Space Agency, Ottawa, 1992, pp. 83-87. 11 T. Young, Philos. Trans. R. SOC. London, 1805,95,65. The topic of wetting has recently gained in popularity. For instance, see: ‘Wetta- bility’, ed. John C. Berg, Marcel Dekker Inc., New York, Surfactant Science Series, Vol. 49, 1993; ‘Contact Angle, Wettability, and Adhesion’, ed. K. L. Mittal, VSP, Utrecht, 1993. 12 For an exhaustive review of these techniques see ‘Light Scattering by Liquid Surfaces and Complementary Techniques’, ed. D. Langevin, Marcel Dekker Inc., New York, Surfactant Science Series, Vol. 41, 1992. 3 P. Picker, E. Tremblay, and C. Jolicoeur, J. Soh. Chem., 1974, 3, 377. 4 These numbers were taken from a recent book, ‘Fundamentals of Low Gravity Fluid Dynamics and Heat Transfer’, ed. B. N. Antar and V. S. Nuotio-Antar, CRC Press, 1993.

ISSN:0306-0012    

DOI:10.1039/CS9942300275

出版商:RSC    

年代:1994

数据来源: RSC