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/CS99423FX017

出版商:RSC    

年代:1994

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99423BX019

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

ISSN 0306-001 2 CSRVBR 23(5) 299-362 (1 994) Chemical Society Reviews Volume 23 Issue 5 Pages 299-362 October 1994 Some Aspects of the Metal-Insulator Transition By Jeremy K. Burdett (pp. 299-308) Several aspects of metals and insulators are discussed in this review. After an introduction which shows the relationship between electronic descriptions and transport phenomena (and in which it is suggested that the term ‘metallic bond’ should be dropped), four interesting examples are examined. (i) The metallic behaviour of supported monolayers, (ii) the behaviour of the Group 2 metals under pressure, (iii) the band gap in d6 perovskites, and (iv) high-temperature superconductors. Photooxidation Reactions of Transition Metal Carbonyls in Low-temperature Matrices By Matthew J.Almond (pp. 309-31 7) This article considers the photochemical reactions in low-temperature matrices in which transition metal carbonyls are oxidized. Typically the oxidant is dioxygen and the reactions proceed via oxocarbonylintermediates (in which both CO and oxygen, in one form or another, are co-ordinated to the metal centre) to yield ultimately molecular metal oxides. Techniques for characterizing such species by infrared spectroscopy are discussed. The use of oxidants other than dioxygen and the oxidation of species other than binary carbonyls are mentioned. Aqueous Aluminates, Silicates, and Aluminosilicates By Thomas W. Swaddle, Julian Salerno, and Peter A. Tregloan (pp. 31 9-325) Although aluminosilicates are commonly regarded as insoluble, the aqueous solution chemistry of aluminosilicate anions is important in various technological and biomedical contexts.NMR methods show that these anions resemble aqueous silicates, which form a remarkable range of small oligomeric structures in alkaline aqueous media, more closely than aqueous aluminium species, which show little structural diversity. The kinetic lability of many silicates and, especially, aluminates and aluminosilicates strongly influences their aqueous chemistry but limits the structural information obtainable from NMR spectra. 1,lo-Phenanthroline: A Versatile Ligand By Peter G. Sammes and Gokhan Yahioglu (pp. 327-334) Over the past decade the classical chelator, 1,lO-phenanthroline, has been catapulted into a starring role in the field of supramolecular chemistry and molecular recognition. This review aims to illustrate the use of 1,lO-phenantholine and its derivatives in areas such as the chiral recognition of DNA, as a probe for left- handed and right-handed forms; in DNA nicking reagents; as enzyme mimetics, promoting redox reactions; and as a sensitizer of europium in commercial heterogeneous immuno- and DNA assays, utilizing time- resolved luminescence.We report on our approach to the development of a homogeneous DNA assay using derivatives of this ligand. The Insertion of Alkynes into Metal-Metal Bonds and Organic Chemistry of the Dimetalled Olefin Complexes By Richard D. Adams (pp. 335-339) Recent studies on the insertion of activated alkynes into the metal-metal bond of dinuclear manganese and rhenium carbonyl complexes to yield 2-and E-dimetalled olefin complexes are reviewed.The dimetalled olefins can be readily functionalized by the insertion of additional small molecules into the metalkarbon bonds. Removal of the metal atoms has resulted in the formation of some novel organic molecules. Electrochemical Solid State Analysis-State of the Art By Fritz Scholz and Birgit Meyer (pp. 341 -347) The direct electrochemical analysis of insoluble solid substances has always been a great challenge to electrochemists. Digestion of samples, for example, leads to loss of information concerning the structures of the solids. If the solid sample is electron-conducting itself, as metals and alloys are, it can be used as a solid electrode.Solid compounds which are insulators, or which do not have sufficient conductivity, can be introduced into a carbon paste or mechanically attached to solid electrodes. With the electrochemistry of solid compounds their qualitative and quantitative identification is possible and information on thermodynamics and on different modifications is accessible. The Thermodynamics of Micellar Solubilization of Neutral Solutes in Aqueous Binary Surfactant Solutions By Claude Treiner (pp. 349-356) Unlike surfactants may form mixed micelles in water, introducing new surfactant solution properties. This topic is reviewed for one of the most characteristic surfactant properties -the micellar solubilization of scarcely soluble compounds.It is shown how regular solution theory predicts and experiments confirm that, in general, mixed micelles are less favourable to micellar solubilization than single surfactant micelles. Exceptions are thoroughly discussed. The relevance of parameters such as surfactant partial demixing and polydispersity of non-ionic surfactants is considered. Oxidation of Some Organic Compounds by Aqueous Bromine Solutions By Josefina Palou (pp. 357-362) This review deals with oxidation of some common functional groups by bromine in aqueous media, and covers the literature from 1967 to 1992. Some generalities about the chemistry of bromine are presented in this article besides the kinetics and mechanism of oxidation of several organic substrates with aqueous bromine.Articles that will appear in forthcoming issues include Biological Activity, Reactivity, and Use of Chromotropic Acid and its Derivatives J. Duda LIVERSIDGE LECTURE. The Dynamics of Photodissociation R. N. Dixon Pericyclic Key Reactions in Biological Systems and Biomimetic Syntheses U. Pindur and G. H. Schneider Mechanistic and Structural Investigations Based on the Isokinetic Relationship W. Linert Benzotriazole Mediated Arylalkylation and Heteroarylalkylation A. R. Katritzky and Xiangfu Lan HAWORTH MEMORIAL LECTURE. Experiments Directed Towards Glycoconjugate Synthesis T. Ogawa Surfactant Systems: Their Use in Drug Delivery M. J. Lawrence Molecular Mechanics Force Fields in Cyclopentadienyl Complexes B.Bosnich Chemical Society Reviews (ISSN 0306-001 2) 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. 699.00, U.S.A. $186.00, Canada El 11.OO+ GST, Rest of World E106.00. Customers should make payments by cheque in sterling payable on a U.K. clearing bank or in U.S. dollars payable on a U.S. 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 RevieM3sat E30.00 per annum; they should place their orders on the Annual Subscription renewal forms in the usual way.

ISSN:0306-0012    

DOI:10.1039/CS99423FP031

出版商:RSC    

年代:1994

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99423BP033

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Some Aspects of the Metal-Insulator Transition Jeremy K. Burdett Department of Chemistry and The James Franck Institute The University of Chicago Chicago IL 60637 U.S.A. 1 Introduction Studies of metals and semiconductors per se have been import- ant endeavours for the physics community for many years but one of the most interesting areas from both chemical and physical points of view is identification of the factors determin- ing whether a particular solid is a conductor of electricity how well it does it. and how external events such as pressure and temperature may move a system from one regime to the other. From the viewpoint of the electronic structure of the solid a metal is simply a system with a partially-filled energy band. But what are the rules associated with the generation of this state of affairs and what are the factors which compete with them and which lead to insulators? Certainly the predictive capabilities of existing ideas are not particularly good. Witness for example the surprise associated with the unravelling of the properties of the high-temperature superconductors and the difficulty in finding good models with which to describe them.Indeed one of the consequences of the discovery of this series of superconduct- ing copper and bismuth oxides has been the resurfacing of many old unsolved problems in this area. This review will try to set out some of the central electronic considerations in this field and highlight some of them with examples in more detail.2 The 'Metallic' Bond 2.1 Two Types of Electronic Description The question of metals and insulators (the 'dynamic' electron problem) makes immediate contact with the way that the electronic structure of solids (the 'static' electron problem) is described. Like many chemical problems the nub of some of the theoretical considerations may be phrased in terms of a balance between the one-electron and two-electron terms in the Hamil- tonian and so are frequently difficult to quantify. The simplest approach starts with the bonding situation in the H2 molecule using the molecular orbital or Mulliken-Hund model. * In-phase and out-of-phase combinations of the two Is orbitals lead to bonding and antibonding combinations (#,.,&,). The simplest wave function for the (&,)* configuration is Jeremjv Burdett was born in London and iras educated at Magda- lene College Cambridge and The University of Michigan. He obtained his Ph.D.,from Cambridge in 1972 btith J. J. Turner. He has been on the jacultjy of The University of Chicago since 1978 and is presently Professor and Chair- man. His interests are in plzysical aspects of inorganic chemistrjt especially structural -electronic relationships in molecules and solids. He is the author ofsome 200 articles and 4 books. with a total electronic energy of Here a = (+llXefll+t)and p the interaction integral = (~,lX"ffl4,). using the Huckel approach. Xeffis some effective Hamilito- nian for the problem. U = 1)4,(2) r;; 1)4/(2)) a two- electron term is the Coulombic repulsion between two electrons located in the same orbital on the same atom.Since there is a finite probability from the wavefunction of equation 1 that the two electrons can reside simultaneously on the same atom its contribution to the energy of equation 2 is non-zero. The energy of the ground state is thus the sum of the two types of terms those which depend on just one electron (a,P) and are negative and that which depends on two electrons (U)and is positive. The wavefunction of equation 1 describes an arrangement where the two electrons are delocalized via the MO description over the two hydrogen atoms. A different approach is used when the two hydrogen atoms are far apart. The chance that both electrons can reside on the same atom together is small and thus inclusion of terms such as 4 (1)4,(2) inappropriate.Instead a Heitler-London wavefunc- tion is written as (3) The electrons are now localized one on each of the hydrogen atoms (H.H.). The energy of this state is just ET = 2a. Two higher energy states where both electrons are forced to lie on the same atom as in H'H- and H-H+. have an energy of ET = 2a + U. The energy difference between delocalized and localized descriptions* is determined by the critical ratio K = U/4p. If K>> 1 the localized description is appropriate but if K>> 1 the localized picture is the one to use. K measures the importance of one-electron (p) and two-electron (U) contributions to the energy.Many observations in chemistry are of the same type. The value of P/A where P is the pairing energy and A the tzn/fn splitting (respectively two- and one-electron terms) discrimi- nates between high-spin and low-spin complexes of the transi- tion metals. These simple ideas may be extended to solids. If the interac- tions (p)between orbitals on adjacent atoms are small compared to the on-site repulsion (U) then the localized Heitler-London description is appropriate. This is the case for example for the copper x2 -y2 electrons in La,CuO,. The result is a semicon- ductor with an activation energy for electron movement the Coulombic penalty associated with placing two electrons on the same atom at the same time. If /3 is large compared to the on-site repulsion U then the Mulliken-Hund description is the one to use.Determination of the magnitude of the parameter K is not a simple task however. 2.2 Localized and Delocalized Bonding The pair of words 'delocalized' and its antonym 'localized' have a variety of meanings within the chemical literature and so one should be careful with their use. In the previous section we have used the terms to describe Mulliken-Hund and Heitler-London functions respectively in accord with the types of terms appear- ing in the wavefunctions of equations 1 and 3. In a separate but related usage the bonding in metals is frequently described as 299 delocalized since these materials are electrical conductors and therefore contain itinerant electrons By way of contrast the bonding in insulators is often described as localized since the electrons are not itinerant Although one should be much stricter in the choice of language since current usage is so ingrained it is difficult to be Included in the very broad spectrum of systems for which the Mulliken-Hund or molecular orbital picture IS appropriate are both the 0 and x manifolds of organic systems including the classic Huckel ‘delocalized’ picture for the 7~systems of conjugated molecules The photoelectron spectra of such compounds are well-matched by simple molecular orbital constructs However the 0 manifold of these molecules is usually described as being made up of ‘localized’ two-centre two-electron bonds Well-described too by the Mulliken-Hund approach are not only materials such ds diamond dnd Zintl phases but also those of metalloids such as bismuth The link between the two apparently contradictory theoretical pictures is ~ell-known,~ but definitely worth repeating here If in a molecule there is a set of N bonding orbitals each occupied by an electron pair and there are N close contacts in the structure then by taking judiciously chosen linear combinations of the delocalized orbitals a set of N localized two-electron two-centre bonds may frequently be generated -the bond orbitals As long as one is discussing ‘collective’ properties3 of the electrons such as total energy or electron density then either the molecular orbital or bond orbital approach is valid Non-collective proper- ties such as photoelectron spectra require the use of the molecu- lar orbital delocalized model In methane four equivalent ‘localized’ C-H ‘bonds’ may be generated from the set of four filled (a + f2) orbitals but the photoelectron spectrum shows two peaks If there are insufficient pairs of electrons for such bonds as in the x manifold of benzene (three bonding x pairs but six close contacts) then such a transformation is not possible A circle drawn inside the hexagon describes this state of dffairs Thus both s andp manifolds may be described by the molecular orbital method (this has to be used in the interpretation of photoelectron spectra) but the delocalized orbitals may be ‘localized’ in one case (a) but not in the other (x)to give two- centre two-electron bonds An analogous picture holds for solids For diamond the energy bands created vzu the delocalized picture may be localized in the same way as for methane by the construction of Wannier functions However here there is a very important restriction In order to do this the energy band must be filled with electrons Thus for quartz SiO the result is a localized picture with the two-centre two-electron bonds between silicon and oxygen as found in the Lewis structure However for the band structure which describes the x manifold of graphite just as in benzene such a transformation is not possible since the band is half-filled The delocalized picture has to be used The most important point of this discussion is that the option of an electronic description in terms of locdlized orbitals is available only for a subset of Mulliken-Hund systems namely those with a filled band or with the right number offilled orbitals in a molecule The delocalized picture is however always appropriate for all Mulliken-Hund systems Recognition of this fact has led in the molecular realm to dramatic progress in recent years in the electronic descriptions5 of a wide range of com-pounds including cage cluster and organometdllic examples Thus the view of the ‘metallic bond’ in solids we would like to stress is that it corresponds to a Mulliken-Hund description of the solid but without the option of the construction of locdlized Wannier functions Since the construction of such functions require the presence of a filled band of electrons a metal corresponds to the presence of a partially filled energy band Thus there is no special type of chemical bonding associated with the ‘metallic’ bond The band itself is generated by orbital overlap between adjacent atoms just as for the molecular orbitals of methane Probably the term ‘metallic bond’ should be dropped from the literature The challenge for the chemist is to identify the factors which lead to partially filled bands of this type for some systems and not for others CHEMICAL SOCIETY REVIEWS 1994 (a)ct-2J3,--161 /I-/ ‘-7 a+2p\lt-It \ DO -Figure 1 The relationship between (a) the Jahn Teller distortion in cyclobutddiene and (b)the Peierls distortion of a polyacetylene chain It It It tE S simple arsenic cubic Figure 2 (a) The three-dimensional Peierls distortion of the simple cubic structure for the Group 15elements to give (b) the structures of arsenic and black phosphorus (The simplest derivative structures are also shown ) NaCl SnSe P SnS As ? *“F Tw3-2.3 CDW and Peierls Distortions By analogy with Jahn-Teller ideas in molecules partially-filled energy bands are in principle Subject to geometrical distortions associated with a lowering of the total energy and variously described as Fermi surface instabilities Charge Density Waves (CDW) or Peierls distortions As the structure changes d gdp may open at the Fermi level to create an insulator or semicon- ductor Figure I shows a classic picture relating the Peierls distortion of the infinite one-dimensional chain (which would be metallic if half-filled)of polydcetylene to the Jdhn-Teller distor- tion of cyclobutadiene The lower energy structure in both systems is one where the bond lengths alternate Many CDW instabilities are triggered by lowering the temperature and occur in a range of systems which cover a wide range of chemical types SOME ASPECTS OF THE METAL INSULATOR TRANSITION-J K BURDETT 30 1 IONIC CsF R CsCI/ \ METALLIC COVALENT Cs Na Mg Al SI P S CI F Figure 3 A vdn Arkel Ketelddr didgrdm used to separate metallic ionic and covalent bonding in solids The indices used here are the sum dnd difference of the atomic electronegativities metal oxides and sulfides molecular metals and the elements themselve3 The three-dimensional analogue of Figure 1a shown in Figure ?a is the description’ of the structures (Figure 2b) of the Group 15 elements in terms of three-dimensional Peierls distortions of the simple cubic structure The latter with its half- filled collection of p orbitals would be a metal As a result of the distortion from cubic phosphorus the coordination number is reduced from six to three no energy bands are partially filled.and the three ‘sp3’bonds and lone pair of the localized model are dbk to be generated In this case the distortion may be reversed by pressure Under these conditions black phosphorus becomes inetal In dn exactly analogous way hydrogen is predicted to become d metal under (4 Mbar or higher) pressure Thus in addition to the criterion for the generation of a metal of a pdrtially filled energy band is the importance of CDW distor- tions which will open a gap and enable d localized picture to be drdwn Peierls drgued thdt in one-dimension d gap would always be opened on distortion (as in Figure la) but that this was generally not the case in two or three dimensions The details are beyond the scope of this review and we refer the reader else- where6 for d fuller discussion The vdridtion in the magnitude of the driving force from system to system is thus a crucial parameter to understand Figure 3 shows an interesting result namely a van Arkel-Ketelaar diagrdm used traditionally to separate compounds into ‘covalent’ ‘ionic’ and ‘metallic’ regions In view of the discus- sion above horizontal excursions across the diagram from ‘metallic’ to ‘covalent’ dre concerned with the variation in the magnitude of this driving force These diagrams may be con- structed qudntitdtively8 using ds d horizontal coordinate the sum ot the electronegativities (from conhgurationally averaged ionization energies) of the constituents.and ds d vertical axis their difference Obviously the electronegdtivity difference is d good measure of ionic character but why does the electronegati- vity sum lead to d good separation3 First it should be recognized thdt the Peierls distortion is driven by the energetic preferences (Figure la) of the occupied energy levels which lie at highest energy Resisting the distortion driven by these electrons are both the ‘elastic’ forces of the underlying electronic structure dnd the repulsive pdrt of the interatomic potential There are some simple considerations which show why elemental lithium has a close-packed structure dnd is d metdl but elemental fluorine is an insulator and is composed of F2dimers Calculations show9 a dramatic increase in the driving force for the distortion on moving from left to right dcross the didgram Although there are several ingredients the part of the electronic picture which is easiest to see is that dssocidted with the orbital interaction integrals which link adjdccnt dtoms pin the Huckel language Using the Wolfsberg- Helmholz dpproximdtion these are written as being directly proportiondl to the sum of the corresponding CI values Since these incredse smoothly in magnitude on moving from left to right across the periodic tdbk dn incredse In the magnitude of the distortion energy and thus the distortion itself should be expected from this source The a values used in calculations of this type come from atomic spectral data and so it is particularly interesting to see that the horizontal index for the van Arkel- Ketelaar diagram of Figure 3 is just the sum of the atomic electronegatives evaluated from the same atomic datd 2.4 The ‘Metallic’ Bond The discussion above emphasizes the need to remove from the vernacular the traditional idea that the metdllic bond represents a different bonding type Metals of the type described above are just those systems usually well-described by the orbital model but where the driving force is not large enough to open d gap dt the Fermi level An area where the traditional view of the metallic bond is especially inappropriate is that of the molecular metals exemplified most recently by the doped fullerenes These are simply systems held together by van der Waals forces where there is d smdll but importdnt overlap between the orbitals of edch unit to form energy bands On doping by either intrinsic or extrinsic means these bands may become partially filled and metallic conduction possible The electronic description of these indteridls is readily dccessible by tight-binding calculations the solid-state analogue of the molecular orbital approach Although outside the scope of this review for d wide range of solids the correlation between theory using this orbital model and experiment (see for example reference 6) in terms of the identification of the k vectors which nest the Fermi surface with those observed from diffuse X-ray scattering experiments is impressive 3 Metals and Insulators 3.1 Two Broad Classes of Metal-Insulator Transitions The metal-insulator transition is a many-faceted phenomenon but let us try to paint a picture which may be an oversimplificd- tion but will enable a broad overview In the previous section two routes to the generation of an insulator were outlined The first was one where the on-site Coulomb repulsion is so large that the electronic description of the system required the use of a Heitler-London wavefunction of the localized type Electrical conduction is activated the activation energy being associated with the accommodation of two electrons lying simultaneously in the sdme atomic orbital and thus subject to a strong Coulom- bic repulsion between them The second route led to an insulator by the filling of the highest occupied energy band Conduction may only now occur by excitation of electrons to the valence band which may not occur at ambient temperatures Because the bnnd is filled localized Wannier functions may be constructed for the system Sometimes as in diamond these localized functions look just like the ones expected from simple chemical considerations In MoS however the ‘localized’ functionlo is ‘delocalized’ over three atomic centres just like the bonding orbital in the H molecule Thus there is using the term ‘localized’ in somewhat different wdys a correspondence between insulating behaviour and ‘localized’ bonding Figure 4 shows these two broad classes of metal-insulator trdnsitions and how for the half-filled band.an insulator may be generated either by localization or via a CDW distortion CDW localization Figure4 How dn insulator may be generated either by locdlizdtion or viu d CDW distortion for the hdlf-filled bdnd Elemental hydrogen provides example of all three types Under high pressure hydrogen becomes a metal this corresponds to the picture in the centre of the figure Under ordinary pressures hydrogen exists as dimers regarded as being generated from the metallic structure via a CDW distortion at left The situation at the right describes the electronic picture for hydrogen atonis at large internuclear distances Mott suggested that systems where transitions from insulator to metal occur as the particle density increases may be regarded in terms of a screened Coulomb potential between the conduc- tion electrons and the nuclei There is a critical screening constant,ll determined by the electron density at which the electrons condense around the nuclei to give an insulator An example here might be the metal-insulator transition associated with the increase in density of electrons of solutions of sodium in liquid dmmonia Dilute solutions are blue ('localized' solvdted electrons) but more concentrated ones bronze and metallic The transition between the two is set by the concentration of sodium Similar considerations are probably behind the generation of metallic (and superconducting) Laz ,Sr,Cu04 for Y >0 05 by doping the insulator (I = 0) with strontium Here the dopant is 'holes' in contrast to the electrons of the sodium/liquid ammonia case The first holes which are introduced by doping are trapped at local sites by a local distortion the small polaron The idea suggested by Hubbard,' of competition between interatomic overlap leading to band formation and the localization of electrons when the on-site Coulomb repulsion (U) is high has been described earlier On this model the critical ratio determin- ing the metal-antiferromagnetic insulator transition for the half- filled band is U/W,a parameter related to K,where Wis the band width The two bands in the antiferromagnetic insulator are separated by a Hubbard or correlation gap The lower level at the far right of Figure 4 is best described by equation 3 and the upper level by equation 4 Thus the energy gap between them the Hubbard U,is simply the sum of the ionization potential and electron affinity of the separated atoms Evaluation of U for solids is not quite as simple Insulators of this type are frequently called Mott-Hubbard insulators Anderson' deve-loped a model where disorder leads to a mobility edge within the band of the ordered metal which separates localized and deloca- lized states Since disorder is invariably introduced by doping (e g ,in La -,Sr,CuO,) separation of the Mott Hubbard and Anderson aspects of the transition are difficult Of particular concern at present is how in detail doping an antiferromagnetic insulator leads to a metal 3.2 The R81e of Band Overlap If filled bands give rise to insulators and partially-filled bands to metals then one route to the generation of a metal is to allow filled and empty bands to overlap as in Figure 5 Elemental potassium is a metal because of a partially filled 4s band a result which might suggest that elemental calcium with a filled 4s band might be an insulator However under ambient conditions calcium is a metal because filled 4s and empty 4p bands overlap Elemental nickel is analogously a metal because filled 3d and empty 4s bands overlap In both of these cases band overlap is also associated with mixing or hybridization between the levels concerned Elemental grey arsenic is a semi-metal due to overlap of valence and conduction bands via interactions between adjacent sheets in the solid Band overlap may be achieved by increasing the pressure on the solid Iodine for example forms a molecular solid but the levels have a non-zero width via the formation of energy bands since there are significant interac- tions between the orbitals on adjacent molecules As pressure is applied the band gap smoothly drops to zero ( -17Okbar) and at the same time the solid becomes metallic and increasingly so as the band overlap increases (Figure 6) Sometimes the bands which overlap interact strongly with each other For example s-p mixing is so strong in diamond that a gap is generated4 and an insulator results In graphite an analogous process occurs but only with the in-plane p orbitals There can be no such interac- CHEMICAL SOCIETY REVIEWS 1994 or (n+ nsor Figure 5 The generdtion of d metdl by overldp of filled dnd empty bdnds Filled and empty bands are the ns and np respectively for the Group 2 elements dnd nddnd (n + 1)s for the Group 10 elements -i-l1 t I Pressure Figure 6 Broddening dnd eventud! overldp of moleculdr levels ds the intermolecular separation is decreased in iodine for example tion between 2s and 2p-rr orbitals by symmetry and the result is metallic graphite The converse of the above is the generation of an insulator or semiconductor by removal of band overlap Although grey arsenic is metallic because of band overlap the isoelectronic Zintl phases Sr[Sn,As,] K[SnSb] and K[SnAs] are semiconduc- tors They contain arsenic-like sheets' and the Group 14and 15 atoms alternate in the structure All three are semiconductors in contrast to the behaviour of grey arsenic For Sr[Sn,As,] the gap is smaller than in the other two compounds The reason is easy to see just by looking at the structure In the potassium-containing compounds these ions are arranged in layers which alternate with the SnAs or SnSb sheets and in the strontium-containing compound these metal ions are only found between every other pair of sheets The potassium ion spacer in K[SnAs] and K[SnSb] increases the separation between the sheets decreases the inter- action between them and removes the overlap of valence and conductor bands (the reverse of Figure 6a) The materials become semiconductors Because of the stoichiometry there is one set of non-bonded close contacts between adjacent SnAs sheets in Sr[Sn,As,] which are absent in the potassium-contain- ing compounds The semiconducting gap is thus smaller than in K[SnAs] and K[SnSb] In grey arsenic each sheet has close contacts on both sides The two examples that follow are much more subtle in an electronic sense 3.3 Metallic Behaviour of Supported Monolayers A novel metal-insulator transition occurs in metal mono-layers l3 Experiments in recent years have characterized the properties of monolayers of the late transition metals (M = Ni SOME ASPECTS OF THE METAL INSULATOR TRANSITION-J K BURDETT 04 02 0 -0 2 -0 4 I -0 6 hz-o a F Q)c ’ 10 ’ 20 30 40 0-081 -0 .10. . 20 30 40 w -0 34 -0 -0 39 -0 -0 44 -0 -0 49 -0 -O 593-0 -06410 .10 20 30 40 I -09J0 ’ 10. . 20’ 30 40 I DOS(states/Ry-Ni) Figure 7 Computed electronic density of states for (d) bulk nickel (b) d nickel monoldyer (c) d nickel monolayer on dn oxide support dnd (d) a two-atom thick multilayer of nickel on an oxide support Pd Ag Hg or Cu) supported on a ‘cushion’ of either X = H 0 or CO adsorbed on a W(110) surface W(l lO)/X/M Interest-ingly W( 1 10)/O/Ni does not behave metallically at all but W( 1 IO)/H/Ni does Thus the metallic properties of the overlayer metal M appear to be determined by the nature of the ‘cushion’ X This behaviour shows up clearly in electronic structure calculations carried out using the first principles LMTO approach Figure 7 compares the computed electronic densit- ies of states for bulk nickel and for a nickel monolayer Both are predicted to be metallic The band widths are smaller for the monolayer than for the bulk a result of the smaller coordination number in the monolayer The d-s band overlap and hybridiza- tion ChdrdCteriStiC of the transition metals leads to approxima- tely one (s + p) electron in the metallic structure of the sheet Thus the Fermi level for nickel lies below the top of the (largely) d band The situation is just as represented in Figure 5 However the presence of the underlying support changes the picture completely for the Group 10 case We know from studiess on complexes of transition metal atoms with ligands of a variety of types that the metal (n + 1)s orbitals are destabilized much more strongly than the corresponding nd orbitals on ligand coordination An exactly analogous process is calculated to occur when the monolayer is ‘coordinated’ to the cushion material The (n + 1)s dnd nd bands are separated in energy (Figure 5a)and non-metallic behaviour results for the Ni sheet Now the Fermi level for nickel lies above the top of the (largely) d band Since the monolayer in proximity to the cushion leads to d gap but the bulk metal is an electronic conductor the gap between d and s bands will disappear as the number of layers of the metal M increase n in W( 1 lO)/X/M Figure 7 also shows the calculated density of states for the case of n = 2 where the gap has indeed closed Analogous considerations for copper show it is metallic under all of these conditions 3.4 The Metal-Almost Insulator Transition of Calcium under Pressure We noted earlier that since the energy bands of solids broaden with a decrease in interatomic separation d generdl prediction is that all materials should become metallic under pressure A different result at intermediate pressures is found for the Group 2 elements Ca-Ba and some of the lanthanide elements with the s2 configuration Their conductivity decredses’ with dn increase in pressure and in the case of Yb a semiconductor IS generated Such behaviour is not found for the Group I2 metdls zinc and cadmium which have d d’O core However there is d structural distortion away from the hcp structure found for these elements which is connected to the rather unusudl observation for the Group 2 elements The experimental observations for the Group 2 elements are matched by band structure calculations which show thdt overlap of s and p bands of Figure 5 IS almost completely removed as the 304 CHEMICAL SOCIETY REVIEWS 1994 Cd Zn I-\ 'p' band lo'@' Figure 8 Thes andp bands of a linear chain (a) where the s p separdtion is small compared to the inter orbital interactions and (b)where the converse is true volume of the unit cell is reduced The result is the opposite to the 'usual' one of Figure 6 Although the three-dimensional band structure of thefcc solid is a non-trivial matter to derive study of the orbital properties of a simple one-dimensional chain will be quite sufficient to understand what is behind this interesting result Generation of the band structure of a linear atomic chain containing s and PO orbitals requires solution of the relevant secular determinant which includes three different values of p,appropriate for PO-~U s-s and pa-s interactions It may be readily written down as a + 2bsscoska -E 21/3 sinku -21p sinha a! -2& cosha -E (5) I I The algebraic solution will not be derived explicitly but there are important symmetry restrictions which control the form of the wavefunctions At the points (0 (27~/a)) theand X (3(27~ja)) group of k contains the elements (E,i)but for any other k the group contains {E only The group of order two which contains {E,I) contains two different symmetry species one symmetric (f and one antisymmetric (r,)with respect to inversion The group of G the general point which contains (El alone only contains a single irreducible representation (GI) Thus at the symmetry points f and X the s and pa orbitals transform asrl + r2(or X + X,) but as 2G at general points in between This important result seen algebraically in equation 5 too is that s-p mixing only occurs at general points in the Brillouin zone (and is a maximum fork = 7~/2a)At both zone centre and zone edge the crystal orbitals are either pure s or pure p in character Two different types of behaviour are thus forced by this last statement and are shown in Figure 8 If the s-p separation is large compared with the values of ,8 for the J and p bands then two bands each bounded by the energy of the relevant in-phase and out-of-phase orbital combinations are found (Figure 8a) Mixing between s and p orbitals at general k points occur and is shown by the dashed lines Figure 8b shows the case where the unmixed s andp bands (dashed lines) cross in energy in other words the in-phasepa combination lies deeper in energy than the out-of-phase s orbital combination Impor-Ca Sr -P-P -Figure 9 Expected behaviour of the electronic situations of Figure 8 under pressure tantly note that the lower energy band the 's-band while purely s-s bonding at the zone centre is purely p p bonding at the zone edge Extension of this discussion to three dimensions leads to an understanding (Figure 9) of the behaviour under pressure of the s2 and dl0s2metdls Whether case (d) or (b) is found depends upon the values of a ap and PI/ Larger values of PI/ usually correspond to shorter interatomic separations In fact elemental zinc corresponds to case (a) but elemental calcium to case (b) For the former the electronic situation is a little like that for He Bringing together two filled s bands should lead overall to a repulsion between the atoms An increase in metal-metal As-tance along c is the result found for zinc and cadmium Both elements crystallize in the hcp structure but with cja ratios (1 856.1 886 respectively) quite different from the ideal value of 1 633 Thus the structure is stretched along the c direction An increase in pressure leads to broadening of both bands just as in Figure 6 By way of contrast a reduction in the internuclear separation occurring with an increase in applied pressure should have very different effects on the relative energies of the top of the lower band and bottom of the upper band of Figure 9a Since bonding levels should drop in energy and antibonding ones should rise in energy the s-p gap should increase This is the case for calcium and as the metal-metal distance becomes smaller the gap between the two bands increases as shown by the full band structure of the material * 3.5 Band Gap in d6Perovskites Molecular transition metal complexes have a gap between the d orbitals of eKand t2K type which is determined in textbook fashion by the u and 7~ strength of the coordinated ligands It is interesting to see how the size of the band gap between the d bands of ee and tZR type in some solids is determined by somewhat more complex orbital considerations l6 The size of the gap determines properties of the solid of course The perovskite LaRhO is non-metallic with a gap of -1 6 eV but although the isoelectronic LaCoO with the same structure is also non-metallic it is onlyjust so The gap is small so that at higher temperatures a whole series of transformations take place triggered by thermal excitation of electrons from valence to conduction band Thus the chemical identity of the system is vitdl here Since the BO part of the ABO perovskite structure is composed of three identical mutually perpendicular 0-M-0-M-0 chains the essence of the electronic argu- ments will be contained in the band structure ofjust one of these SOME ASPECTS OF THE METAL-INSULATOR TRANSITION- J K BURDETT -10-1units As in the molecule by symmetry there is no mixing between the u (eg)or x (t2J bands It will be useful to take advantage of some further symmetry arguments As argued above for the simple sp chain at the points r (0 (2~/a))and X (+ (2~ia))the group of k contains the elements {E,z} but for any other k the group contains {Ejonly Also at the zone centre the wavefunction must be symmetric from cell to cell and at the zone edge antisymmetric This means that on a p-orbital-only model (Figure IOa) there can be no pa or p~ contribution at the bottom of either the pa (e,) band which lies at rorpX (t2J band which lies at X The bottom of both bands thus lie at the unperturbed d orbital energy The energies depicted come from the angular overlap model On this model there is no gap at all between the dbands of egand t2gtype and thus all d"perovskites (0 < n < 10) including those with the d6configuration should be metals The same results applies to the two-dimensional net (Figure lob) Inclusion of oxygen s orbitals changes the picture (Figure 10c) Whereas interaction of oxygen p with metal dis zero at r and non-zero at X the converse is true for interaction of oxygen s with metal d because of the difference in orbital parity Thus it is inclusion of oxygen s/metal d interactions which open the eg/t,ggap in the perovskites The magnitude of the gap is thus not solely determined by the balance between u and 7~ donor effects as in molecules but by the differential effects of metal-oxygen su and metal-oxygen pa interactions Thus prediction of the magnitude of the gap without a calculation is extremely difficult Unlike the case of the Zintl compounds where electronic saturation usually leads to semiconducting behaviour the eg/t,g gap in transition metal systems is fre- quently non-existent and the state of affairs system-dependent 4 High-temperature Superconductors 4.1 The Importance of Two-electron Terms Perhaps the classic cases of control of metallic or insulating properties by the balance of one- and two-electron terms in the energy come from the area' of doped transition metal oxides This is especially important for oxides containing first-row metals The recently-discovered group of high-temperature superconducting oxides based on doped cuprates are particu- larly interesting examples Identification of the nature of the superconducting state has proved elusive (however see refer- ence 17) but the geometrical control of the electronic structure and ultimately therefore the superconducting properties of these materials are striking -9I Three examples from the series of the superconducting cuprates will be discussed in this section As in all materials of this type they contain18 sheets of stoichiometry CuO (Figure -1 0 -1 1 -1 2 -1 3 (b) M r X M Figure 10 (d) Dlspersion behaviour of the bands of a one-dimensional perovskite chain The u and n bands overlap (Energies shown are from the angular overlap model ) (b) The d bands viametal d oxygen pa orbital interactions in d squdre MO net (Figure 1 Ib) (c) The same ds (b) but including metal d oxygen s orbital interactions 11) La ,Sr,CuO one of the very first 'high-temperature' superconductors with a T of 35 K for x -0 15 contains distorted CuO octahedra vertex-fused in two dimensions (Three-dimensional vertex fusion leads to the perovskite struc- ture and thus to the frequently used description of this family of compounds as 'perovskite superconductors' ) Jahn-Teller dis- tortions are typical of Cu" and here two long axial and four short equatorial Cu-0 distances are found The La (and Sr) atoms lie between these sheets Since these axial Cu-0 distances Figure 11 (a) The perovskite structure (b) A sheet of stoichiometry CuO found in all cuprate superconductors Often the sheet is not planar being slightly puckered or rumpled are long ones a good way of describing the solid is as alternating layers of rocksalt (La,Sr)O and perovskite CuO layers In superconductors containing bismuth and thallium Bi20 and T1,0 rocksalt layers are found here too The first ‘high- temperature’ superconductor to be discovered with a T above liquid nitrogen temperature was YBa,Cu30,,l8 with a T of-95 K Figure 12 shows the structure of the compound There is a range of oxygen stoichiometry possible in the inter-planar region namely YBa,Cu,O,- 0 < 6 < 1 and an interesting dependence of T on 6 There is a precipitous drop at around 6 = 0 65 where the compound becomes an antiferromagnetic insulator The parent compound YBa,Cu,O (Figure 12) con- tains two types of copper atom square pyramidal Cu(1) and square planar Cu(2) Since the fifth Cu( 1)-0 distance is quite long a good description of the structure is of chains of copper Cu(2) in square planar coordination sandwiched between planes of copper Cu(1) As oxygen atoms are lost from the chains with an increase in 6 linear two-coordinate copper atoms are generated so that in YBa,Cu,O all of the inter-planar copper atoms are two coordinate Many superconducting cuprates are now known and all may be described geometrically as containing sheets of CuO stoichiometry (where supercon- duction occurs) alternating with sheets or slabs of insulating material The latter described as ‘reservoir’ material for reasons which will become apparent below constitute a spatially distinct region separated from the CuO sheets by the long axial Cu-0 distances in the octahedra The formulae of the systems des- cribed above may then be rewritten as (La -,Sr,02)(Cu0,) and (Y)(Ba,CuO,)(CuO,) to highlight this geometrical (and elec- tronic) simplification In the former the reservoir material is a slab of rocksalt-like La -,Sr,O and in the latter there are two types of reservoir material (Figure 12) namely layers of Y atoms and the CuO chains of square planar Cu(2) linked by Ba atoms In superconductors containing bismuth and thallium (BiO) and (TIO) rocksalt layers are found TI,Ba,Ca,- ~CU,O~,+ written as (T1202)(Ba202)(Ca,l -l)((Cu02),J contains distorted T1,0 double rocksalt layers between the CuO sheets whereas TlBa,Ca,-,Cu,O,,,+ contains single TI0 layers (TIO)(Ba,O,) (Cam- 1)((CuO2)rJ Although some of these systems are formally electron-doped compounds (e g ,Nd -,Ce,CuO,) the majority are hole-doped (e g La -,Sr,CuO,) The compositional dependence of the properties of this second group appear to be reasonably similar As shown in the generic diagram18 of Figure 13 the materials may be converted from an antiferromagnetic insulator into a superconducting metal and then into a normal metal via a change in copper oxidation state Specifically for La -,Sr,-CuO T reaches a maximum at x -0 15 where the change in copper oxidation state is controlled by substitution of a two- valent ion for a three-valent one (e g ,Sr for La) The presence of defects (non-stoichiometric amounts of oxygen or bismuth or thallium for example) are effective in this regard too in many of the bismuth-or thallium-containing superconductors The reservoir material composition and structure determines the oxidation state of the copper atoms in the CuO planes There are considerable experimental problems in determining the exact stoichiometry in many of these systems The construction of a band picture for the CuO sheet for the region where the molecular orbital Mulliken-Hund approach is valid is 5traightforward The d orbital level splitting pattern expected for a planar CuO fragment and the band generated for the extended CuO array is shown in Figure 14 Since the axial Cu-0 distances are long the z2 orbital lies at low energy stabilized in this geometry too by d-s mixing The y2 -1.orbital of the metal the highest occupied orbital of the fragment is involved in 0 interactions with the ligands Since the atomic energies of the copper 3d and oxygen 2p orbitals are similar strong mixing between them is to be expected This level occupied by the single unpaired electron of Cull is antibonding between copper and oxygen,* a result which shows up experi- mentally as a strong dependence of Cu-0 distance on doping The band is metal-oxygen antibonding and thus addition of CHEMICAL SOCIETY REVIEWS 1994 C a Figure 12 The structure of YBa2Cu30 -Doping(Copper oxidation state) Figure 13 A generic picture showing how the antiferromagnetic insulat- ing state for d cuprate may be doped to give a metal and superconductor electrons (as in Nd -,Ce,CuO,) should lengthen and removal of electrons (as in La,-,Sr,CuO,) should shorten the Cu-0 distance By and large this is true but the steric demands of the reservoir material also influence these distances The result for La,CuO itself is a half-filled band situation Since the on-site Coulomb repulsion U for this first row metal K(Cu) is expected to be large will be large too and an antiferromagnetic insulator should be favoured This is what is found experimentally for the undoped material Addition of holes probably leads initially to the generation of small polarons (vrdesupra) which eventually collapse to give a metal at a critical doping level Addition of these holes suppresses the CDW instability expected for the half-filled metallic band and so on doping a metal is eventually generated It happens to be a superconductor too It is interesting to ask how important is the r6le played by copper in these systems For this we turn to some SOME ASPECTS OF THE METAL INSULATOR TRANSITION-J K BURDETT 307 row congener and the magnitude of the orbitdl interactions are f t Figure 14 Simple derivation of the band structureof a sheet of stoichio metry CuO b d b 6 Figure 15 (d) Symmetricdl dnd (b) distorted structures for platinum dnd nickel mixed-valence chains Expected Found Predicted if all Cu-0 distances were equal Figure 16 Band overlap in YBa,Cu,O (a)None leading to no plane chain charge transfer (b) the actual state of affairs (c) overlap expected if the plane and chain 0-0 distances are equal interesting results concerning the one-dimensional analogues of these CuO sheets namely the platinum mixed-valence com- pounds and their nickel analogues (Figure 15) Here the highest occupied band is the '2'band of the chain The oxidation state of the platinum in the symmetrical structure in Figure 15a is PtlIi and thus this 'z2' band is half-full of electrons A Peierls distortion is expected and indeed the symmetric system Figure 15a distorts to Figure 1% The initially metallic state has become insulating although as in the case of elemental iodine and phosphorus the conductivity of these salts increases mark- edly with application of pressure Some of the nickel analogues are found as antiferromagnetic insulators with an undistorted structure This behdviour is very similar to that found for molecular complexes of nickel and platinum and readily under- stood from Jahn-Teller considerations The singlet-triplet split- tings for the first-row nickel atom are larger than for its third- larger for platinum than for nickel The result is thus always diamagnetic square-planar complexes for PtIi (e g ,Pt(CN)i ) but the frequent observation of high-spin octahedral complexes for Nil1 (eg Ni(H,O)i+) The important parameter in the molecular case is again the critical ratio of these one-electron (Jahn-Teller stabilization energy) to two-electron (singlet-trip- let splitting) energy terms From this viewpoint one would suspect that silver analogues of the cuprate superconductors would be susceptible to strong Peierls distortions which would destroy the metallic Ago sheet Indeed no silver compounds structurally analogous to these cuprates are known 4.2 Band Overlap and the Generation of a Metallic State It was a straightforward matter to determine the copper oxi- dation state in La ,Sr,CuO but how the oxidation state of the planar copper atoms depend on oxygen stoichiometry in Y Ba2 Cu307 -b,is a more complex question Square planar coordina- tion is found for both Cull and CuiIi and square pyramidal geometries always with long apical bonds only for Cu" Such observations indicate that the formula YB~,(CU~~O,),(CU~~~O,) should be a correct description of the compound with 6 = 0 Similarly YBa,(CullO,),(CuiO,) should be a good way to describe the compound with S = 1 Linear two-coordination ISd common geometry for Cul Both pictures lead to a formdl oxidation state of Cull for the copper atoms in the superconduct- ing planes and do not answer the question of why YBa,Cu,07 is a metal but YBa,Cu,O an antiferromagnetic insulator Both compounds on the simplest model with a half-filled yZ -1 band should behave like La,CuO and be insulators It is vital to examine the form of the band structure of the materid12 to see how the CuO planes for 6 = 0 are doped away from the half- filled band and thus described by the generic picture of Figure 13 The details of the band structure of the material are both simple and interesting Since there are now two different types of square-planar copper there will be two types of xz -1 bands Both may be constructed in the same fashion as shown in Figure 14 (The z2 orbital may be ignored since it lies deep in energy ) Since the x2 -j2 band associated with the planes is of 6 symmetry with respect to the inter-planar material it is un-coupled from the energy bands of the chains since there dre no oxygen orbitals of that symmetry The xz -band for the chains needs to be relabelled as 3 -j since the chains run in the 1.2 plane The electronic description of YBa,Cu,O depends crucially on the relative locations of these two sets of bdnds the y2 -j band of the CuO planes and the z2 -1 band of the CuO chains Writing the S = I system as YBa,(CullO,) (Cd1'O3) forces the description of Figure 16a where the chain -2 -band is empty and thus Cu(2) described formally as Cull' and the plane y2 -band is half full so that Cu(1) is described as Cuil (The different numbers of copper atoms of the two types Cu(1,2) are indicated by drawing the boxes of different width ) The true situation is somewhat different since the two bands overlap (Figure 16b) so that the electron transfer occurs from planes to chains The half-filled band is avoided for the Cu( 1) atoms of the CuO planes and a metal may result In electronic terms the interplanar CuO unit plays therefore the same r6le as substitution of Sr for La in the Ld,CuO system The details of the local copper coordination geometry control the band overlap and hence charge transfer If the square-plandi bond lengths and the in-plane square-pyramidal bond lengths were equal the x2 -L~ and z2 -i2bands.while being of different widths would be similarly energeticdlly locdted as in Figure 16c Their overlap would lead to around 0 33 holes on each planar cop er atom Since the bond lengths in the pldnes (average 1 944 x) are substantially longer than those in the chains (average 1 888 A) the result is destabilization of the z2 -i2 bdnd reldtive to x2 -12 but maintdining overldp between them (Figure 16b) such that chargc transfer may occur between the two The same model allows access to the behaviour CHEMICAL SOCIETY REVIEWS 1994 of YBa,Cu,O Here there is no chain z2 -v2 band which may overlap with x2 -y2 and the planar copper oxidation state is that given by the formula YB~,(CU~~O~)~(CU~O~) Now the electronic state of affairs with a half-filled band is similar to that of undoped La2Cu04 and an antiferromagnetic insulator results In these cuprate superconductors there are thus two distinct ways to remove electron density from the planar y2 -1 band by doping and by band overlap As oxygen is removed from between the sheets the extent of plane-chain doping gradually decreases (Figure 17) and somewhere corresponding to a stoichiometry between YBa,Cu,O ,and YBa2Cu30 a metal-insulator transition is to be expected This shows up in the T dependence plot on 8 where the superconducting temperature drops precipitously to zero as the insulator is generated There are also structural changes in the material which show up most in changes in the c-axis of the material YBa2Cu307 YBa2Cu306 Y Ba2Cu306 n n u ki planes chains planes chains planes &y2 &y2 x2-y2 &y2 x 2-y Figure 17 Chain plane electron transfer in YBa,Cu,O and YBa Cu,O and its absence in YBa,Cu,O Band overlap controls the properties of other superconduc- tors too Overlap is found to occur by calc~lation~~ in Tl,Ba2 Ca,,- 1C~,,O2,~+4 which as noted above contains distorted T1,0 double rocksalt layers between the CuO sheets There is however no computed overlap in the material TIBa,Ca,,- CU,~O,,~+~which contains single T10 layers There is a simple electronic explanation for the difference between the two Just as the ‘band width’ of a simple polyene increases as the number of dtoms increases (asymptotically approaching the value found for the infinite system) so the contribution to the band width of the Tl,)Cl,* unit increases with the number of atoms or units (n)in the direction perpendicular to the slab Thus the Tl,O band-width is ldrger than that for the single T10 layer (Figure 18a b) Another pair of systems where metallic (and superconducting) behaviour is simply controlled by hand overlap is TlMRECuO Superconductors are found for RE = La Nd and M = Sr but not for its M = Ba analogues even though the two systems are isostructural There is though a small difference between the cell parameters of the Sr and Ba analogues resulting in a small difference in the Cu-0 bond lengths in the two materials [The crystallographic a parameters (roughly twice the Cu-0 dis-tance) are 3 849 8 (Ba) and 3 761 8 (Sr)] This difference of dbout 005 A in the Cu-0 distance is sufficient to raise the energy of the y2 -J band of the CuO sheet so that band overlap with the TI0 layer may occur in the Sr case but not for Ba as shown qualitatively in Figure 18c d These doping models are applicable2 to all of the presently known high-T superconductors The band model predicts a half-filled y2 -band for Cu2 +,an electronic configuration susceptible to two types of instability which lead to the creation of an insulator Both localization via a large on-site Coulomb iepulsion or a Peierls distortion are routes away from the metallic state Electrons need to be added or removed from this band to generate a metal Some of the high-temperature super- conductors appear to be electron doped but the majority are hole-doped where electrons are removed from the half-full bdnd There are two routes to this state of dffairs either by overlap with an empty band or a change in stoichiometry Although the second mechanism is easily understandable by consideration of the chemical formula via cation substitution n TIBa2Ca -1Cu,O2~3 T12Ba2Ca -1C%02,+4 (TI0 single layers) (TI202 double layers) or or TIBaRECu05 TISrRECuO Figure 18 Band overlap in Tl,Ba,Ca ICu,102,,+4which contains single T1,0 double rocksalt layers between the CuO sheets but no band overlap in TlBa,Ca,,- ,Cu,,O,,,+ ,which contains single T10 layers Similar results apply to TIMRECuO where M = Sr and Ba respectively cation vacancies or the presence of extra oxygen the first requires d detailed understanding of the elecronic structure In some materials certainly both mechanisms are at work 5 Epilogue It should be apparent from the discussion in this review that a prim prediction of metallic or insulating properties is not straightforward Some of the stoichiometry or structural changes involved are very small indeed For example substitu- tion of only 4% of the lanthanum by strontium in La2Cu04 is sufficient to convert an insulator into a metal and a supercon- ductor at that Similarly the difference in cell constant between TlBaRECuO (an insulator) and TISrRECuO (a superconduc- tor) is small The difference in the orbitdl interactions between metal and oxygen s andp orbitals which controls the band gap in tP perovskites is a parameter difficult to easily assess in qualita- tive terms The r6le of the structure itself is an important one and vdries from the obvious (eg diamond and graphite) to the subtle (L‘ g ,the doping mechanism in YBa,Cu,O,) 6 References I (u)J B Goodenough ‘Magnetism and The Chemical Bond’ Wiley 1963.(h)J B Goodenough Piog SoIidState Chem 1971,5,143,(c) P A Cox ‘The Electronic Structure dnd Chemistry of Solids’ Oxford 1987 2 J K Burdett in ref I~(N) 3 ((I) M J S Dewdr ‘Theory of Moleculdr Orbitdk’ McGrdw-Hill 1969,(h)A Haug.‘Theoretical Solid State Physics’ Pergdmon Press I972 4 J K Burdett ‘Chemicdl Bonding in Solids’ Oxford 1994 5 T A Albright J K Burdett and M -H Whangbo ‘Orbital Interactions in Chemistry’. Wiley 1985 6 E Cdnddell dnd M -H Whangbo Chem Re) 1991,91,965 7 J K Burdett and S Lee J Am Chrm Soc 1983 105 1079 8 L C Allen J Anr Chem Suc 1992 114 1510 9 W P Anderson J K Burdett and P T Czech J Am Chem Soc 1994,116,8808 0 K A Y Yee and T Hughbanks Znorg Chem 1991,30,2321 1 N F Mott ‘Metal-Insulator Transitions’ Taylor and Francis 1974 2 P C Schmidt D Stahl B Eisenmann R Kneip V Eyert and J Kubler J Solid State Chem 1992,97 93 3 N Shamir J C Lin and R Gomer J Chem Phys ,1989,90,5135 J E Whitten and R Gomer Surface Science 1994,316 36 14 J K Burdett and S Sevov. J Chem Phqs 1994 101,840 15 T L Brennan and J K Burdett Znorg Chem 1993,32 750 16 J K Burdett and S A Gramsch inorg Chem 1994,33,4309 17 J K Burdett inorg Chem 1993,32 3915 18 (a) ‘Chemistry of High-Temperature Superconductors’ ed T A Vanderah Noyes 1991 (b)J K Burdett Ad) Chem Phjs 1993 83,207 19 D Jung M -H Whangbo N Herron and C C Torardi Physica 1989 C160,381

ISSN:0306-0012    

DOI:10.1039/CS9942300299

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Photooxidation Reactions of Transition Metal Carbonyls in Low-temperature Matrices Matthew J. Almond Department of Chemistry University of Reading Whiteknights Reading RG6 2AD U.K. 1 Introduction Matrix isolation an invention of George Pimentel in 1953 is nowadays an established method for prolonging the lifetimes of transient chemical species. A full description of the technique is beyond the scope of this short review and in any case such experimental and theoretical details are given The principle behind the method is as follows when a chemical species is embedded at high dilution in a solid host at low temperature (a typical example being solid argon at 20 K) bimolecular reactions are suppressed and unimolecular decomposition is stopped for any process with an activation barrier larger than a few kJmol- * .The species is thus preserved. There are two main routes to generate matrix-isolated species. First the reactive entity may be generated in the gas phase by for example heating or the action of a microwave discharge and is then frozen rapidly with an excess of the matrix gas e.g. argon. Alternatively a stable precursor is trapped within the matrix and the reactive species is generated in situ typically by photolysis. Thus a wide range of atoms molecules ions and radicals may be isolated in low-temperature matrices then studied by a number of spectroscopic techniques.’ -3 One advantage of the matrix isolation method is that at least to some extent the method of detection may be tailored to the problem under investigation. In this review the photo-oxidation reactions of transition metal carbonyls in low-temperature matrices will be discussed. Many simple binary carbonyls and a number of ternary com- pounds have thus been studied by trapping the carbonyl at high dilution in a matrix containing molecular dioxygen. A typical mixture is M(CO),,:O,:Ar = 1:100:2000 at 20 K. The matrix is deposited upon a window transparent to IR radiation e.g. CsI held within a high vacuum shroud and is then subjected to UV- visible photolysis. Depending on the type of experiment being performed the photolysis may be ‘broad-band’ i.e. covering the whole of the UV and visible regions or at selected wavelengths. Broad-band photolysis is used typically to push the reaction through to its final products which are generally isolated molecu-lar binary oxides. Narrow band photolysis is used to form intermediate oxocarbonyl molecules in which both CO and Matthew J. Almond was born at Blackburn Lancashire in 1960. He obtained his BSc. from the University of Reading in 1981 then moved to the Inorganic Chemistry Laboratory of the University of Oxford. There he obtained his D.Phi1. for spectroscopic studies of un-stable species in lou2-tempera- ture matrices. Two years of postdoctoral work were shared between the University of 0.xjoi-d and the University of Munster (Germany) before in 1986 he returned to Reading to take up a post as lecturer. He has published tM.0 books and almost 50 research papers. He is a Fellow and currently also a member of Council of The Royal Societj) of Chemistry. oxygen in one form or another are co-ordinated to the metal centre and to explore the reaction pathway. In most of the examples mentioned here IR spectroscopy has been used to monitor the progress of the reaction and to characterize interme- diate and product compounds although Raman and UV-visible spectroscopy have been employed in a number of cases. Oxidation reactions of organometallic compounds is cur- rently an area of considerable research interest. Thus in transi- tion metal chemistry the synthesis of organometallic oxides such as (CH3)Re0,,4 and cp*ReO (cp* = $5Mes),S and the properties of such molecules as catalysts6 attract considerable attention. Under appropriate conditions main group organome- tallics will also yield isolable products with 0,. The 1992 Meldola lecture by Barron was concerned with reactions of the group 13 alkyls with dioxygen forming products which are sufficiently stable to be structurally characterized by single crystal X-ray diffra~tion.~ One aim of this review is to show that matrix isolation is an excellent means to study reactions of this type since intermediates may be stabilized and mechanisms explored. This information is of value to chemists working in diverse areas such as (i) selective oxidation of organic substrates at transition metal centres; (ii) transport of oxygen by living organisms; (iii) metal oxide chemistry; (iv) corrosion; (v) chemi- cal vapour deposition of oxide thin layers. 2 Identification of Products and Intermediates It is instructive to consider briefly how the structure of molecules such as binary oxides and oxocarbonyls may be studied in a matrix when vibrational spectroscopy is the only readily avail- able technique. The first question to answer is how does oxygen hind to the metal centre? Four different modes of bonding are commonly encountered. These are superoxo (end on) (1) or peroxo (side on) (2) binding of 0 or monoxo (3) or dioxo (4) co-ordination where the 0-0 bond has been ruptured. Superoxo derivatives typically show IR absorption arising from ~(0-0) vibrations in the region 1000-1200 cm-I; peroxo derivatives show IR absorption from the ~(0-0)vibration in the region 700-1000 cm-l the same region in which absorption from u(M=O) vibrations of (3) or (4)occurs. Thus a definitive characterization of the mode of binding of oxygen relies upon experiments utilising the isotope l80. ONO 0-0 0I \/ II o+ 50 M M M M Using a mixture of I6O2 *60180,and 1802 causes the absorption associated with the 0-0 stretching mode of the 7,-0 peroxo group to take the form of a single symmetrical triplet. A monoxo species gives rise to a simple doublet whilst under similar conditions (4)forms two asymmetric triplets derived from the antisymmetric and symmetric stretching fundamentals. The two unsymmetrical triplets are produced by the strong coupling of the two I60=M=l stretching modes caused by the reduction of local symmetry induced by the unsymmetrical isotope substitution. This situation is illustrated schematically in Figure 1. The superoxo species (2) would be expected to give a quartet of bands representing the four possibilities of isotopic 309 CHEMICAL SOCIETY REVIEWS 1994 I I I Figure 1 Predicted lS0isotopic splitting patterns in the infrared spectraof (a) the ~(0-0)band of a dioxygen complex and (b) the symmetric and antisymmetric v(M=O) bands of a bent dioxometal complex assuming a scrambled I2 1 I6O21601s0lSO2mixture (0= l60 = lSO) The horizontal arrows in (a) and (b) indicate that the dv(1602-160180)wavenumber spacing is equal to the dv(160180-1802)spacing in (a) but not in (b)(Reproduced with permission from reference 15 ) combinations of the 7-0-0 unit I e M-60160,M-601 M-I8Oi60 and M-180180 although it must be said that the splitting of the two central features resulting from the mixed oxygen isotopomers may be quite small It is possible also to distinguish v(M=O) from ~(0-0)vib-rations by the magnitude of the shift in frequency when l60is replaced by I8O In the case of MO (or indeed MO,) units the wavenumber shift will depend upon the LOM0 bond angle O The shift in position of vdsym(O=M=O) upon ls0substitution gives an upper limit 8 for this angle by the equation 160) -(m('so))R,]sin 5 = {m(X)[m( 12 2(m('60))(m('sO))(R1 -1) where m(X)is the mass of the appropriate metal atom and This calculation gives an upper limit R = (~(~~0)/v(l~O))~ rather than a true value for the bond angle since the vibration is necessarily anharmonic Similarly isotopic substitution of the metal atom (or observation of bands arising from isotopes of the metal atom in their natural abundance) gives a lower limit by where R = (V(~M)/V('M))~Unfortunately the calculated values of 8 become more sensitive to the precise value of the isotopic shift as the molecule approaches linearity It is also possible to approximate 8 by measurement of the intensity ratio of infrared bands arising from vdsym and vsym(O=M=O) (m(0)+ 2m(M)sin2(8/2))&=tan (e'2) (m(0)+ 2rn(M)cos2(8/2))Lyrn Likewise (C-M-C) angles of metal dicarbonyl moieties may be estimated by the simplified equation Zdsym/Zsym = tan2(O/2) To characterize fully more complex metal carbonyl fragments requires experiments utilising isotopic substitution with 3C0 or Cl80 and careful fitting of the positions and intensities of the pattern of bands observed by means of force constant calcula- tions generally involving the use of computer programs i 1 1 A 00 3 Matrix Reactions of the Group 6 Carbonyls with 0 Photooxidation of the Group 6 binary carbonyls M(CO) (M = Cr Mo or W) proceeds via oxocarbonyl intermediates ultimately to yield binary oxide products In each of these systems it has proved possible to generate selectively particular intermediates by specific wavelength photolysis Thus more-or- less complete mechanisms for the photooxidation processes have been proposed It is found that these pathways vary somewhat from metal to metal When M = Cr the principal oxocarbonyl intermediate observed is the molecule 'chromyl carbonyl' O,Cr(CO) (5) lo Intensity measurements of the bands arising from vsym and vdsym (Cr(CO),) and isotopic shift calculations on vdsym(OCrO) have allowed the C-Cr-C bond angle to be estimated as 118" and an upper limit for the O=Cr=O angle has been set at 129" This molecule is of interest as it was the first example of a chromium(1v) carbonyl compound yet it uppears to be one of the most stable intermediates in the photooxidation of Cr(CO) Structure (5) shows vdsym and ~,,,(Cr(CO>,) vibrations at 2125 and 2065 cm-l the high frequency of these features reflecting the limited ability of the metal centre to engage in back-donation to the T* orbitals of CO when in a high oxidation state The vibration vdsym(OCrO) gives rise to an infrared band at 981 cm l which as expected yields an asymmetrical triplet when the molecule is generated from a scrambled mixture of the oxygen isotopomers Continued pho- tolysis causes loss of all coordinated CO groups (n b by-products of all of the photooxidation reactions discussed in this article are invariably 'free' CO and CO,) and the formation of molecular CrO l1 CrO has previously been made in low- temperature matrices by the reaction of Cr atoms (formed by sputtering) and 0 Thus definitive characterization of our compound may be made by comparison with these spectra Our photooxidation experiments allow facile isotopic substitu- tion hence the upper limit for the 0-Cr-0 bond angle may be estimated as 1 17" When M = Mo or W analogues of 'chromyl carbonyl' are not seen as major intermediates Rather the principal oxocarbonyl species seen here are the previously unknown complexes trans-di 0x0 tetracarbon ylmol y bden um(rv) and trans-dioxo te tracar- bonyltungsten(1v) (6) l3 The square planar array of four CO groups is confirmed by isotopic substitution with I3CO The linear O=M=O moiety is demonstrated both by isotopic substi- tution with l80and by observation of splitting of the band arising from v~~~~(O=MO=O) due to the various isotopes of molybdenum in their natural isotopic abundance In Figure 2 is illustrated the observed spectrum Thus it can be seen that the observed molybdenum isotope splitting also allows confirma- tion of the dioxo rather than a monoxo formulation The high symmetry of these molecules is reflected in the simplicity of their infrared spectra and in the non-coincidence of infrared and PHOTOOXIDATION REACTIONS OF TRANSITION METAL CARBONYLS-M J ALMOND 311 co CO Visible light It 314 nrn VlsiMe light + 0,I co oc.. I ..-0 -%I-I '0Ioc' Figure 2 Observed and calculated Mo isotope structure in the IR dbsorption spectra of monooxo dnd dioxo species isolated in an 02-doped CH matrix at 20 K (a) calculated isotope frequency pattern for a monooxo species assuming the Mo=O group to be a simpleharmonic oscillator (b) observed molybdenum isotope structure for the 757 cm band of (6),(c) isotope pattern for a linear dioxo speciescalculated by using equation 2 with 0,/2 = 90" The calculated spectra nrn are depicted as stick plots with heights given by the relative natural abundances of the molybdenum isotopes (Reproduced with permission from reference 13 ) Raman bands Thus when M = Mo (6) shows an infrared- active v(C0) vibration at 21 10cm-' and a Raman-active v(C0) vibration at 2175 cm-l the symmetric and antisymmetric stretches of the linear O=Mo=O unit give rise to a Raman band at 820 and an infrared band at 758 cm- respectively Spectra of the tungsten analogue are similar Prolonged broad band photolysis of Mo(CO) in 0,-doped Ar matrices leads to the formation of a mixture of MOO and MOO l3 By contrast when W(CO) is treated similarly the final product is WO but this is not produced from WO Rather the precursor to WO appears to be WO unit coordi- nated to 0 (7) which forms WO by loss of an oxygen atom (equation 1) Evidence for the formation of oxygen atoms comes from experiments performed in (1) pure 0 (11) 0,-doped N and (111) 0,-doped CH matrices Under these conditions concomitant with the production of WO are seen (1) 0 (11) N,O and (111) methanol and formaldehyde (7) All of these reactions of 0 atoms are known from gas-phase studies the last three are typical of 0 atoms in the excited 'D state In order to establish the route by which these oxocarbonyl intermediates and oxide products are formed and to gain an understanding of the overall reaction mechanism experiments utilising narrow band photolysis and isotopic substitution were employed Thus the reaction mechanism given in Scheme 1 may be proposed The reaction is initiated by loss of CO from M(CO) to yield the 16-electron species M(CO) A common intermediate in these reactions is (O,)M(CO) (8) which con- tains a peroxo unit with an intact 0-0 bond Each of these molecules shows an infrared-active ~(0-0)vibration and the frequency of this mode decreases as the mass of the metal increases thus when M = Cr ~(0-0)= 995 2 cm-' M = Mo ~(0-0)= 954 0 cm-' M = W,v(O-0) = 914 0 cm-' This (7) (5)I Prolonged UV M = Cr Mo or W M' = Mo or W only M" = Cr or Mo only Scheme I sequence reflects presumably not only the effect of the mdss of the metal upon the ~(0-0)vibration but also the increased capacity of the heavier metals to donate electronic charge to the 7~*orbitals on the 0 group These peroxo species each yield upon near-UV photolysis (O),M(CO) (9) where the 0-0 bond has been cleaved Compound (9) may in turn act as d precursor to (O),M(CO) (5) or trans-(O),M(CO),(6) by loss or uptake of CO 4 Photoxidation of Fe(CO) and Ni(CO) in Low-tern peratu re Matrices Photooxidation of Fe(CO) in low-temperature matrices IS initiated in the same way as photooxidation of the Group 6 hexacarbonyls I e by loss of CO from the parent carbonyl to yield the corresponding unsaturated 16-electron species in this case Fe(CO) Fe(CO) differs from M(CO) (M = Cr Mo or W) in having a triplet as opposed to a singlet ground-state It has been noted that unsaturated metal carbonyl fragments are related to the divalent carbon compound carbenes by the so-called 'isolobal relationship" ' (n b two molecular fragments are said to be isolobal if the number symmetry properties shapes and approximate energies of their frontier orbitals are the same) Like Fe(CO) carbenes have triplet ground-states Thus it was hoped that some similarities might be found between the chemistry of Fe(CO) and of carbenes in low-temperature matrices containing 0 The carbene cyclopentadienylidene (lo) under these conditions is known to yield an end-bonded CHEMICAL SOCIETY REVIEWS 1994 adduct with 0 (1 1) which upon visible photolysis is isomerized to a side-bonded form (1 2) (Scheme 2). 8*1 Fe(CO) forms a side-bonded peroxo adduct ( 13).,O This complex shows an intriguing example of Fermi resonance between the v(0-0) fundamental and the first overtone of the ~,,,(FeO,) fundamental. Thus two bands rather than the expected single feature are seen in the region of the IR spectrum associated with the ~(0-0)vibration vide infra. These features are seen at 914 and 885 cm-l. It can be shown by experiments involving lSO that these bands do indeed arise from a ~(0-0) vibration rather than from vsym and vasym vibrations of an O=Fe=O unit. Thus the magnitude of the l60-ls0shift is too large for a dioxo unit and the intensity ratio of the two bands reverses upon 80-substitution. This latter phenomenon (see Figure 3) is only explicable in terms of Fermi resonance. A similar effect has been observed for the 0 adduct oxyhaemoglo- bin (almost certainly end-bonded 0,)where Fermi resonance is Wavenumbers (cm-' ) 900 800 Figure 3 IR spectrum of (13) in the region 800-1000 cm-' (bands marked by black infilling are due wholly or mainly to (13) (a) for a sample generated from 1602; (b) for a sample generated from l8O,; (c) for a sample generated from a mixture of I6O2 and 1802 (1602:180,= 3:2);(d) for a sample generated from a mixture of l60 and l80,160180 (1602:160180:180 = 1:2:1) Thesubscripts 1 and 2 are used to denote the two distinct bands due to (13) and the asterisks to denote bands due to Fe' species. (Reproduced with permission from reference 20.) c02+ co ? U* 0 2co o:Fe-CO 0I \-k2uy;/Fe u7$ (16) I ('*) 0 0' II 0 (19) The symbol A indicates annealing d the matrix Scheme 3 seen between the v(0-0) fundamental and the first overtone of the 4Fe-0,) fundamental. Continued near-UV irradiation causes the decay of (1 3) and the growth and decay of a number of oxocarbonyl intermediates [(14)-(17)]. Photolysis proceeds by rupture of the 0-0 bond and loss of CO. Unfortunately this system displays little photo- selectivity thus it is difficult to build up large concentrations of specific intermediates. However these molecules have been characterized more-or-less certainly and the pathway given in Scheme 3 has been proposed for the photo-oxidation of matrix- isolated Fe(CO) . The final products of this reaction are the binary oxides (q2-0,)Fe (18) and FeO (19).,l Compound (18) had previously been made by reaction of Fe atoms and 0,. FeO had not previously been observed although its infrared band at 945 cm- l which is assigned to vasYm(FeO3) is in an almost identical position to a band previously ascribed to matrix-isolated FeO,. -23 There has been some uncertainty regarding the probable structure of this 'FeO,' molecule with either the peroxo (~~-0,Fe)or dioxo (FeO,) forms being proposed. More recent reports favour the dioxo form~lation.~~ The spectra shown in Figure 4 in which the isotopes 1602,l60l8O,and 180 have PHOTOOXIDATION REACTIONS OF TRANSITION METAL CARBONYLS-M J ALMOND Wavenumbers (cm ’ ) 1000 950 900 1000 950 900 t 1 1 1I A‘ (b) (d) 1 1 Figure 4 IR spectrum of matrix-isolated (19) in the region 850-1000 cm (a)spectrum observed for the product generated from Fe(CO) and an equimolar mixture of I6O2and 1802 (b) spectrum predicted for v3 of a planar FeO molecule with D3/,symmetry (c) spectrumobserved for the product generatedfrom Fe(CO) and 16021601s0 and 1802 In the statistical proportions 1 2 5 1 2 (d) spectrum pre-dicted for U of d planar FeO molecule with D, symmetry In edch case the matrix hdd the composition Ar 0 Fe(CO) = ca 1000 50 1 and was maintained at ca 20 K photolysiswas at h = ca 213 nm for 300 min dnd then dt h = 290-370 nm for 360 min Cdlculdtions were based on an energy-fdctored force field for the v(Fe=O) modes (Reproducedwith permission from reference 2 I ) been utilised to generate the various 6Oand isotopomers of (19) leave little doubt however that (19) is not any form of FeO but rather FeO with planar D3,?symmetry 5 WCO) V(CO) is unique amongst simple binary metal carbonyls in being a paramagnetic 17-electron compound 0x0-vanadium species exhibit an extremely complex chemistry including a wide iange of peroxides such as the orthovanadate ion [0,V(q2- O,)l3 and the peroxovdnadium cation [V(q2-0,)]3 + As such there is an interest in studying the photooxidation of matrix- isolated V(CO) in order to search for novel 0x0 or peroxo compounds of vanadium and to compare the results with those obtained for the diamagnetic metal carbonyls The first step in the photooxidation of V(CO) in a low- temperature matrix is entirely analogous to that in the photooxi- dation of Fe(CO) or M(CO) (M = Cr Mo or W) z e loss of CO to generate V(CO) The reaction however is extremely rapid when broad-band UV-visible photolysis is employed all of the V(CO) is consumed after cu 20-60 minutes’ photolysis (the exact time depending upon the concentration of 0 in the matrix) It is perhaps for this reason that it is remarkably difficult to identify intermediate oxocarbonyl species in this system Moreover. the reaction scheme shows little photosensiti- vity thus even by utilising narrow-band filters it proved impossible to characterize unambiguously any of the reaction intermediates The speed of reaction does however facilitate the build up of a large yield of the final oxide product in a i elatively short time The product shows four infrared-active vibrations dt 1129 0 969 5 960 0 and 563 5 cm 24 Experi-ments utilising mixtures of the dioxygen isotopes including ‘“0’ and (Figure 5) show that this product is either the dnionic species (20) or the neutral compound (21) Although the symmetrical ti zplet of bands seen in place of the single band at 1 129 0 cm ’ when the three l60/l isotopes are utilised might point towards (20) ds the most likely structure on d balance of A I Q I; 1100 900 G/cm-’ Figure 5 The region 850-1 150cm * of the infrdred dbsorption spectrd of matrices (20 K) after broad-band UVjVIS photolysis for 20 min ([) initial mdtrix composition [V(CO),) I6O2Ar = LU 1 100 1000 ([I) initial matrix composition [V(CO),] lSOzAr = ((I 1 100 1000 ([u) initidl matrix composition [V(CO),]I6O2’*02Ar = (ii 1 50 50 1000 (it) initial matrix composition [V(CO),] l6OZ1601s0 1802 Ar = ca 1 25 50 25 1000 (Reproducedwith permission from reference 24 ) all the experimental evidence (21) is favoured The O=V=O bond angle is calculated by isotopic shift and intensity measure- ments to be in the range 108-1 15“ 6 Mn,(CO),.. Re,(CO),. and Co,(CO) The photochemistry of binuclear metal carbonyls is always likely to be more complex than that of the analogous mononuc- lear derivatives since more potential primary photochemical steps and hence more photochemical pathways exist The binuclear carbonyls Mn,(CO) and Re,(CO) both have M-M bonded structures with no bridging CO groups As such the two most likely primary photochemical steps are (I) loss of a CO group to yield M,(CO) (M = Mn or Re) or (11) cleavage of the M-M bond to form two M(CO) radicals Photolysis of Mn,(CO) ,in matrices containing 0 results in the rapid formation of the binary oxide Mn,O 25 This is a complex photochemical reaction involving at least four mole- cules of O the loss of all ten carbonyl groups the breaking of the Mn-Mn bond of the starting material and the formation of dn Mn-0-Mn bridge Unfortunately little can be concluded regarding the mechanrsm of the reaction since no intermediate CHEMICAL SOCIETY REVIEWS 1994 (OC)$Ie-0-0-Re(C0)5 (22) .\/& 2(CO),Re/,\;0 (22) 2 O=Re(CO) (23) h -YCOI +xo2-zcop t O3Re-O-Re03 Scheme 4 dn extremely complex mechanism As well as competition between Co-Co and Co-C bond cleavage there have been identified three separate isomers of Co,(CO) in matrices and these are known to interconvert photochemically In the event the photooxidation of Co,(CO) provides a rather curious result which is not in keeping with the findings obtained for the other binary carbonyls reported here Two oxocarbonyls are formed these are (y2-0,)C0(C0) (u = 1 or 2) (24) 28 Identification of these products is possible because the same compounds are formed when Co atoms are reacted with CO/O mixtures at 10-12 K 29 The only other molecule to be seen is the non-CO- bridged species Co,(CO) There is no sign of any mononuclear hmarj cobalt carbonyl nor of the species (y’-O,)Co(CO) which like (yi-02)Mn(CO) has been detected by ESR spectro- scopy 26 Nor intriguingly was there any sign of a binary cobalt oxide product (24) (x = 1 or2) I I00 I 000 900 800 cm-I Figure 6 IR absorption spectra in the region 750-1100 cm of d matrix initially containing Mn,(CO), 60,,and Ar in the approxi mate proportions 1 100 2000 A after deposition dt 12 K B dfter 20 min broad-band UVjVIS photolysis C after 150 min broad-band UViVIS photolysis. D after 220 min broad-band UViVIS photolysis E dfter 400 min broad-band UVjVIS photolysis (The spectrd show the build-up of absorptions due to Mn,O ) (Reproducedwith permission from reference 25 ) can clearly be identified As with V(CO)6 the photolysis is rapid and non-specific Although a low yield of Mn,(CO) is formed (then consumed) at an early stage of the reaction it is not clear whether or not this represents the primary photoproduct in the oxidation pathway There is no sign of the species (7,-0,)Mn(CO)5 which has been detected by ESR spectroscopy This result 26K77vapour at upon condensation of Mn,CO) is perhaps not too surprising when it is borne in mind that the 7 Mn(CO),CI and Re(CO),CI ESR technique is much more sensitive than is infrared spectro- scopy thus a radical formed at low yield may well escape detection by infrared measurements whilst giving rise to intense signals in the ESR spectrum The infrared spectra illustrated in Figure 6 show the formation of Mn,O from Mn,(CO) in an 0,-doped (5%) Ar matrix Although the photooxidation of Re,(CO) is also non- specific2’ the reaction does proceed much more slowly than that of the congener Mn,(CO) Thus more definite conclusions regarding reaction mechanism may be drawn The reaction product is Re,O It is formed vzu two detectable oxocarbonyl intermediates Both appear to contain Re(CO) units and they are believed to adopt the structures (22) (n b two structures have been proposed for this species -see Scheme 4 -either of which on the basis of the experimental evidence would appear to be equally plausible) and (23) Experiments using filtered photoly- sis suggest that the reaction is initiated by Re-Re rather than Re-C bond rupture Although this is a well-known primary photoreaction in solution it is less common in solid matrices since the ‘matrix cage effect’ suppresses the formation of the M(CO) radical In an 0,-containing matrix however the cage effect may be overcome if the radicals once formed react rapidly with 0 A plausible mechanism (albeit only showing some of the reaction steps) for the formation of Re,O from Re,(CO) is given in Scheme 4 The photochemical oxidation of Co,(CO) is likely to follow It is clear from the above discussion that photooxidation of matrix-isolated binary metal carbonyls is an excellent method to generate isolated molecular metal oxides The logical progres- sion of the work was to investigate whether or not it is possible to form ternart species such as oxohalides in the same way When the carbonyl halide Re(CO),Cl is subjected to photoly- sis in 0,-doped argon matrices a product (25) is formed which shows an infrared absorption of medium intensity at 971 cm-together with weak bands at 1004 and 435 cm 30 After ca 4 hours’ broad-band irradiation all of the starting material has been consumed and alongside (25) the only other matrix ingredients are free CO and CO The infrared spectrum of (25) shows a marked similarity to that of the molecule Re0,Cl which has previously been studied in low-temperature matrices The behaviour of the band 971 cm (which is assigned to the degenerate asymmetric stretch of the ReO unit) upon isotopic substitution with ‘*Osupports this assertion The weak band at 1004 cm is assigned to vsym(Re03) and that at 435 cm * to v(Re-CI) Unfortunately no isotopic shift data (160-180or 35C1-37C1) could be obtained for these very weak features CI I Reo// ,\*o0 PHOTOOXIDATION REACTIONS OF TRANSITION METAL CARBONYLS-M J ALMOND 0 2400 2200 Wave number (cm ') Figure 7 IR absorption spectra of a matrix (20 K) initially composed of Mn(CO),CI O and Ar in the approximate proportions 1 250 1000 (A) after spray-on (B) after 10 min broad-band UV/VIS photolysis (C) after 110 min broad-band UV/VIS photolysis Bands marked A are assigned to MnO,,CI-..CO and those marked * are due to Mn(C0) ,C1 (Reproduced with permission from reference 30 ) The molecule Mn(CO),CI is also readily photooxidized under these conditions generating free CO and CO and a product (26) which shows infrared absorptions at 21 71 (weak intensity) 976 (medium) 938 (weak) and 416/408 cm-I (weak) From these infrared data it is clear that (26) is not MnO,Cl since this molecule when isolated in an argon matrix shows vasym and vsym (MnO,) at 951 9 and 889 9 cm-' and v(Mn-Cl) at 459 (,,Cl) and 453 (37Cl) cm-' 31 Isotopic substitution experiments involving lS0 show that the features at 976 and 938 cm-I arise from vdsym and vsym(MnO,) vibrations while the separation of the low-frequency band into two components at 416 and 408 cm- suggest that it arises from a v(Mn-Cl) vibration where the splitting is caused by the presence of the isotopes and 37C1 in their natural abundance The absorption at 2171 cm-I shifts to low frequency upon substitution with I3CO it is thus presumed to belong to a co-ordinated CO group The high frequency is typical for a group where the CO moiety is behaving essentially as a 0-donor Compound (26) is thus likely to be the species Mn0,CI ...CO infrared spectra of (26) are illustrated in Figure 7 8 CpMn(CO) CpRe(CO) and CpV(CO) There has been some interest of late in the so-called organome- tallic oxides of rhenium especially MeReO CpReO and Cp*ReO (where Cp = +,H and Cp* = q-C,Me,) as such species have been implicated as useful catalysts for a range of important chemical processes including olefin oxidation metathesis and polymerization One aim of our current research is to extend our matrix photooxidation experiments to A r-2000 1000 900500 400 generate organometallic oxides which are not stable under normal conditions Our starting point was the oxidation reac- tion of cyclopentadienylmanganese tricarbonyl CpMn(CO) It is found that broad-band irradiation of this compound when isolated in an 0,-doped Ar matrix causes the formation of a product (27) which shows infrared bands which may be assigned to motion of the cyclopentadienyl ring at 1421 1020 1006,820 5 and 793 5 cm-' In addition two bands are seen at 938 and 893 cm- which isotopic substitution experiments with l8O show to arise from vasym and vsym(MnO,) vibrations (27) is thus believed to be CpMnO 32 The spectra illustrated in Figure 8 show the appearance and growth of features associated with (27) There is no sign of any further oxidation of the cyclopenta- dienyl group under the conditions of this experiment CpRe(CO) behaves under similar conditions in a rather different way It is oxidized very slowly the bands associated with the starting compound are seen to disappear only after 25 hours' broad-band photolysis of the material in a pure-0 matrix Here there is no sign of any oxidized product which contains a coordinated Cp group Rather the two most promi- nent features in the infrared spectrum of the photolysed matrix are at 966 and 891 cm-' together with a much weaker band at 815 cm- When CpRe(CO) is replaced by Cp*Re(CO) as the starting material photolysis proceeds even more slowly bands due to the starting material are still visible (though weak) after 66 hours' broad-band photolysis in a pure-0 matrix In this case the photolysed matrix shows in the Re=O stretching region bands of medium intensity at 965,925 and 898 cm- and a weak band at 8 15 cm- There is also some evidence for bands CHEMICAL SOCIETY REVIEWS 1994 I I 1 750 1050 Wavenu mber/cm-' Figure 8 The region 750-1050 cm of the IR absorption spectrum of a matrix initially composed of [Mn(~S-C,H,)(CO),] 0 and Ar in the approximate proportions 1 250 1000 (a)after deposition at 20 K (h) after 30 min of broad-band UVjVIS irradiation,(c) after 120 min of such irradiation and (6)after 260 min of such irradiation Bands marked + arise from vibrations of the cyclopentadienyl ring of the starting material while those marked * arise from vibrations of the cyclopentadienyl ring of (28)(Reproduced with permission from reference 32 ) due to a coordinated Cp* group in one of the products Interestingly the bands at 965 and 8 15cm-appear only at high 0 concentrations they are not seen in a 20% 0,-doped argon matrix One possible explanation is that these last two absorp- tions arise from the molecule Re207 The feature at 898 cm- is close to the reported position of v,,~,(R~O,) of the molecule Cp*ReO (28) in solution Compound (29) is a likely product of this photooxidation reaction it is known that (28) is formed from Cp*Re(CO) upon photolysis in the presence of 0 in solutions containing tetrahydrofuran The origin of the feature at 925 cm- remains unclear Thus a possible photooxidation route for the organometallic compounds CpRe(CO) and Cp*Re(CO) emerges in which the initial process yields the organometallic oxide (28) then further reaction results in oxi- dation of the organic group to produce Re,O CpV(CO) is oxidized rapidly under similar conditions and although it is clear that a number of 0x0-metal species are generated it has not so far proved possible to characterize any of these unequlvocally There is however evidence for the formation of the peroxo-compound Cp*V(CO),(O,) when Cp*V(CO) is photolysed in 0,-doped argon matrices 33 9 Use of Alternative Oxidants So far few experiments have been carried out in which oxidants other than 0 itself are utilised The only report concerns the use of N,O and CO as oxidants for the Group 6 carbonyls M(C0)6 (M = Cr or W) 34 Here photolysls proceeds more slowly than when 0 is the oxidant Ultimately binary metal oxides are formed and it appears that photolysis proceeds first viu loosely-bound complexes X*-M(CO) (X = OCO or N,O) and then through monoxo-metal species of the type O=M(CO) It is clear that this area is ripe for further experimentation In particular points to be addressed are the possible roles of 0 atoms alongside 0 molecules in the oxidation process and whether-or-not the oxidation reaction is affected by the electro-nic state of the 0 atoms 10 Conclusions Matrix isolation is an excellent method for monitoring photo- oxidation reactions This approach has allowed a number of binary metal oxide species to be synthesized and characterized and has recently been extended to some ternary compounds It is clear that there are many more systems which could be investi- gated in this way e g the range of organometallic oxides (of the type CpMnO,) could be enlarged and the work could be extended to the main group organometallics such as Me,Ga and Me,Zn It is instructive to consider briefly the results described above One equation to answer is to M hat evtent does the thermodj nu- mic stabiiitr of the oxides dictate Mhat products ate .formed? When the Group 6 hexacarbonyls are considered a clear and entirely expected trend emerges Thus Cr(CO) is oxidized to CrO (Cr in the + IV oxidation state) Mo(CO) gives a mixture of MOO (MolV) and MOO (MoIV) while W(CO) gives WO (Wvl) via (O,)WO (also Wv') This sequence would appear to reflect the greater ease of oxidation to high oxidation states of the heavier transition metals Likewise the highest oxidation state product observed upon photooxidation of Fe(CO) con- tains FeV1 Thus it is perhaps surprising to find that Mn,(CO) is readily oxidized to give the manganese(v1r) compound Mn207 When oxidation of the ternary compounds ClM(CO) and CpM(CO) (M = Mn or Re) is considered it is found that manganese is oxidized only to the (unusual) + V state whereas rhenium gives the +VII state Naturally a consideration of thermodynamic stability tells us little or nothing about the rate of oxidation Thus as a general observation we find that matrix- isolated rhenium compounds are invariably oxidized more slowly than their manganese analogues However it is import- ant to bear in mind that one of the overriding factors in many of these reactions may simply be the match between output of the photolysis source and absorption bands of reactants and intermediates It is presumably a similar point which dictates whether or not a range of intermediates are seen in these reactions The Group 6 hexacarbonyls show a high degree of photoselectivity thus several different intermediates may be built up selectively by photolysis at different wavelengths The Fe(CO) system shows little wavelength selectivity but it is possible to build up intermediates by carefully controlling the time of photolysis By contrast V(CO) and Mn,(CO) which are oxidized very rapidly give only fleeting glimpses of any possible intermediate species Presumably those systems where absorption bands of reactants and intermediates overlap will give only low concent- rations of intermediates Certainly this would appear to be the case for the Fe(CO) reaction where quite extensive UV-visible spectroscopic measurements have been made and have shown that absorption bands of reactant and intermediates do indeed overlap Lastly we may comment that the matrix isolation technique allows us to trap unusual molecules which would not be stable under normal conditions and we may thus compare the structure dnd bonding of these species with known molecules To take selected examples first the molecule O,M(CO) (M = Mo or W) adopts the trans configuration For a d2 system of this type simple molecular orbital considerations suggest that the tians isomer should be more stable than the czs and it is found that other similar but stable d2 species also adopt a trans configuration e g [O,Re(CN),I3-Secondly it is shown that the same oxide products are often obtained by oxidation of the PHOTOOXIDATION REACTIONS OF TRANSITION METAL CARBONYLS-M J ALMOND metal cdrbonyl ds by reaction of metal atoms with O thus it might be appropriate to consider the metal carbonyls under photolysis conditions simply as sources of isolated reactive metal atoms Upon photooxidation of Co,(CO) the same intermediate oxocarbonyl species are observed as when Co atoms are reacted with CO/O mixtures Thirdly it is of interest to consider some of the metal peroxides which have been generated in the work discussed above The peroxo-iron com- plex (O,)Fe(CO) shows spectral similarities to known peroxo- iron porphyrin complexes and even to oxyhaemoglobin itself Moreover the similarity between this peroxo-iron product and the organic peroxides species obtained upon oxidation of the cyclopentadienylidene radical wznj be explained in terms of the isolobal theory which suggests that there should be similarities in the chemistry of Fe(CO) and C,H, When the Group 6 compounds (O,)M(CO) (M = Cr Mo or W) are considered it is seen that the position of the infrared absorption arising from ~(0-0)follows d sequence which suggests increased back- donation to the 0 7~*orbital from the metal as the group is descended This is entirely in line with infrared spectroscopic dnd single-crystal diffraction measurements for various sequences of transition metal peroxides where the 0-0 bond is seen to weaken and lengthen as the group is descended 36 One last point is that this approach of matrix photooxidation allows some materials which are hazardous under normal conditions to be generated in small quantities and studied spectroscopically entirely in safety Thus Mn,O which is a highly explosive oil dnd quite difficult to handle ~rdinarily,~’ may conveniently be produced by photooxidation of Mn,(CO) in a low-tempera- ture matrix 11 References 1 M J Almond and A J Downs ‘Spectroscopy of Matrix-isolated Species’ ed R J H Clark and R E Hester J Wiley Chichester 1989 2 ‘Cryochemistry’ ed M Moskovits and G A Ozin Wiley-Inter- science New York 1976 3 R N Perutz Ann Rep Prog Chem Sect C Phvs Chem 1986,83 157 M J Almond and R H Orrin Ann Rep Prog Chem Sect C Phjs Chem 1991,89,3 4 I R Beattie and P J Jones Inorg Chem ,1979 18,2318 5 W A Herrmann M Taillefer C de M de Bellefon and J Behm Inorg Chem 199 1.30 3247 6 W A Herrmann,R W Fisher,andD W Marz,AngeM Chem Int Ed Engl 1991,30 1636 7 W A Herrmann and M Wang Angefi Chem Int Ed Engl ,1991 30 1641 8 T Kawai M Goto T Ishikawa and Y Yamasaki J Mol Cutal 1987,39,369 9 A R Barron Chem Soc Re\ 1993,22,93 10 M Poliakoff K P Smith J J Turner and A J Wilkinson J Chem Soc Dalton Trans 1982 65 1 11 M J Almond and M Hahne J Chem Soc Dulton Trans 1988 2255 12 L V SerebrennikovandA A Mal’tsev Vestn Mosk Unit Ser 2 1975 16,251 13 J A Crayston M J Almond A J Downs M Poliakoff and J J Turner Inorg Chem 1984,23 305 1 14 M J Almond and A J Downs J Chem Soc Dalton Trans 1988 809 15 M J Almond J A Crdyston. A J Downs M Polidkoff dnd J J Turner Inorg Chem 1986,25 19 16 M Poliakoff Chem Soc Rev 1978,7,527 17 R Hoffman Angel! Chem Int Ed Engl 1982,21 71 1 18 I R Dunkin and C J Shields,J Chem Soc Chem Commun ,1986 154 G A Bell I R Dunkin and C J Shields Spectrochim Actu Part A 1985,41 1221 19 0 L Chapman andT C Hess J Org Chem 1979.44.962,O L Chapman and T C Hess. J Am Clzem Soc 1984 106 1842 20 M Fanfarillo H E Cribb A J Downs T M Greene dnd M J Almond Inorg Chem 1992,31,2961 21 M Fanfarillo A J Downs T M Greene and M J Almond Inorg Chem 1992,31,2973 22 S Abramowitz N Acquista. and I W Levin Chem Phtc. Lett 1977,50,423 23 S Chang G Blyholder and J Fernandez Inorg Chern . 1981 20 2813. L V Serebrennikov Vestn Mosk Unrv Ser 2 Khrm 1988 29,451 24 M J Almond and R W Atkins J Chem Soc Dulton Trmc ,1994 835 25 M J Almond J Mol Struct 1988 172 157 26 S A Fieldhouse B W Fulham G W Nielson and M C R Symons J Chem Soc Dalton Tranr 1974 567 27 M J Almond and R H Orrin J Chem Soc Dulton Truns 1992 1229 28 M J Almond and R H Orrin J Organomet Chem 1993,444,199 29 G A Om A J L Hanlan and W J Power hoig Chem . 1979,18 2390 30 M J Almond and R H Orrin Polvhedron 1992 11,2157 31 E L Varetti and A Muller Z Anorg Allg Chem 1978 442 230 32 M J Almond R W Atkins and R H Orrin J Chem Soc Dalton Trans 1994 3 1 1 33 A J Rest M Herbernauld and M Schrepferman Organometallic c. 1992 11 3646 34 M J Almond A J Downs and R N Perutz Inorg Chem 1985 24 275 35 W S Caughey M G Choc and R A Houtchens in ‘Biochemical and Clinical Aspects of Oxygen’ ed W S Caughey Academic Press New York 1979 36 J S Valentine Chem Re\ 1973 73 235 37 W Levason J S Ogden and J W Turff J Chem Soc Dalton Trans 1983 2699

ISSN:0306-0012    

DOI:10.1039/CS9942300309

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Aqueous Aluminates SiIica tes and AluminosiIicates Thomas W. Swaddle Department of Chemistry University of Calgary Calgary Alberta Canada T2N I N4 Julian Salerno and Peter A. Tregloan School of Chemistry University of Melbourne Parkville Victoria Australia 3052 1 Introduction By far the most abundant elements in the Earth’s crust are oxygen silicon and aluminium (62 6 21 2 and 6 5 atom YO respectively) In this article we present an overview of the aqueous solution chemistry of these elements in the context of some applications of current interest with emphasis on the generally neglected topic of ciluminosilicates as aqueous solutes This relative neglect is due partly to experimental difficulties noted below but chiefly to the perception thdt the solubility of aluminosilicate rocks soils etc in water is very low under ordinary conditions of temperature and pH Dent Glasser and Hnrvey,’ however have shown that it is possible to create metdstdble solutions containing dissolved aluminosilicate ions at concentrations up to several tenths molar by mixing solutions of alumindtes and silicates such that the onset of deposition of gels may be delayed for hours days or even weeks depending on the concentrations of the reactants the identity of the counter-ion(s). and the pH The gels eventually crystallize to form zeolites aluminosilicates of the type M, [(A1O2),,]-~ (SiO,),,]*pH,O with open structures consisting of cages and channels (e g Figure 1 YM In. from Loewenstein’s rule -see below) and exchangeable cations M to balance the charge- + deficit of the A13+ An understanding of these processes is required in connection with the synthesis of zeolites for use as heterogeneous catalysts molecular sieves cation exchangers etc Furthermore aqueous aluminosilicates are important even at low concentrations in fields as diverse as environmental geochemistry stedm injection oil recovery bauxite refining for aluminium production kraft pulp mill operation and the toxicology of alumini~m,~ ’and we give first a brief outline of come of these applications to place the chemistry in context Soluble silicates (‘wdterglass’) dnd/or dlumindtes dre used in Peter Tregloan is a graduate of the University of Adelaide (Ph D 1969) A.fter postdoctoral ~orkat the Universitj of Kent at Canterbur? and a lectureship in Physical Chemistry at University College Dublin he joined the Inorganic Chemistr-y Department ut the Universitj of Melbourne in 1974 He has had a long standing interest in reactions of labile metal ions and new techniques .for their studj He has held visiting positions at the Universiti of Lausanne Comalco Research and the Universitj of Calgar 1 Peter Tregloan Julian Salerno 319 water clarification textile treatment adhesives cements and as detergent builders Figure 1 Schematic representation of the structure of (a) the sodalite cage showing how this may be assembled in principle by corner- linking four-Si/Al rings or fusion of six-Si/Al rings and (b) zeolite-A formed by corner-sharing of sodalite cages Each corner represents dn A1 or Si atom tetrahedrally surrounded by four oxygens dnd each strdight line includes d bridging oxygen (Reproduced by permission from TW Swaddle ‘Applied Inorganic Chemistry University of Calgary Press 1990 p I 13 ) Julian Salerno was born in Melbourne Australia in 1965 He obtained his 3 Sc (Hons ) from the Universitj of Melbourne where he is currently undertaking a multinuclear NMR investi- gation of speciation and dynamics in alkaline solutions of alumi-nates silicates and aluminosilicates The study is ajoint Industry1 University collaboration wth the support of the Australian Research Council and Comalco Research and Technologj Tom Swaddle a native of Newcastle upon Tyne read chemistrv at University College London (B Sc ,1958) and did research under Colin Eaborn at the Un- iversity of Leicester (Ph D 1961) In 1964,,follow~ing post- doctoral studies with John P Hunt and Edcr ard L King in the U S A ,he joined the academic staft”of the Universitj of Calg-ary where he is nowprqfessor of chemistry He has held visiting appointments at the Universrt- ies of Adelaide Lausanne and Melbourne and at the Tokjjo Institute of Technologv Tom Swaddle 2 Aluminosilicates in Aluminium Production Aluminium metal is generally produced by electrolytic reduction of alumina (a-A1203) in molten cryolite (Na3AIF,) with gra- phite electrodes -the Hall-Heroult process Alumina of the requisite purity is usually obtained by the Bayer process I e ,by leaching bauxite ore with aqueous NaOH (initially some 10-1 5 moll I) typically at about 165"C and 600 kPa whereupon the gibbsite y-Al(OH) in the ore goes into solution as sodium dluminate while most of the silica or silicdtes present precipitate as insoluble aluminosilicates and are removed along with Ti0 and iron(m) oxide/hydroxides as 'red mud' Pure gibbsite then crystallizes from the filtered solution as it cools below about 70 "C following seeding and is calcined to give alumina Not all of the silica however is precipitated as aluminosilicates at the 'red mud' stage some continues to deposit slowly usually as the zeolite-like solids sodalite or cancrinite (M = Na rn = n = 6 p = 8) or below ca 70°C zeolite-A (Figure I) so fouling the equipment and possibly contaminating the alumina product Similar problems arise from the use of alkaline pulping liquors in kraft paper mills It would appear then either that some long- lived soluble aluminosilicate species exist in the Bayer solutions or that the formation of sodalite or cancrinite precursors from aluminate and silicate species can be slow even at ca 165"C 3 Formation of Aluminosilicates by Water-Rock Interactions Many naturally-occurring aluminosilicate minerals form by geological hydrothermal processe~,~ and not surprisingly simi- lar solids may form as a consequence of industrial activity such as (for example) steam injection oil recovery Major reserves of petroleum exist in the Athabasca Oil (or Tar) Sands of northern Alberta in the form of very viscous bitumen in a sandy matrix The most promising method of recovering the more deeply buried reserves involves injection of steam or superheated water at temperatures up to 300°C and pressures up to 5 MPa and collecting the produced petroleum The sandy matrix however is significantly soluble in hot water especially if the pH is high and hydrothermal reactions occur between the small amounts of clay minerals dolomite etc ,and the predominant quartz Thus. in addition to silica reprecipitated in the cooler regions of the water flow path deposits of voluminous aluminosilicates such as montmorillonite and zeolites (notably analcime formation of which is associated with hot water-silica systems low in Al"') may form blocking the communication path and impeding production This process is not necessarily disadvantageous -it may close off a depleted section of the geological formation -but clearly if it is to be controlled knowledge of the solution chemistry of silicates and aluminosilicates is required 4 Synthesis of Zeolites Synthetic zeolites are generally made by mixing solutions of aluminates and silicates often with formation of a gel and maintaining the mixture at temperatures of 100"Cor more for selected periods The identity of the crystalline product depends upon the reaction time and temperature the solid surfaces present and the specific solution conditions -particu-larly the nature of the cation(s) and any organic solutes Thus zeolite A sodalite and cancrinite tend to form in aqueous Na+ media (as in the Bayer process) while zeolites with low A1 content are favoured by large cations such as tetraalkylammo- nium (R,N+) In his seminal study of the kinetics of formation of zeolites A B and X Kerr* found that nucleation in the gel was slow but both spontaneous nucleation and the subsequent OH -accelerated growth of the zeolite crystal were controlled by some solute species derived from the amorphous gel Indeed if the findings of Bell et al can be generalized such gels contain little Al and form through destabilization of a silica sol by the added aluminate so that the formation of aluminosilicates is solution-mediated Furthermore analcime (for example) can be CHEMICAL SOCIETY REVIEWS 1994 VI VII VIII IX XI1 x1n XIV xv XXI XXII XXIn (Q34) XXIV (Q310) xxv Figure 2 Silicate structures that have been detected by 29S~-NMRin dlkdline dqueous medid Filled circles represent Si dtoms tetrdhed- rally coordinated by 0atoms lines represent links through 0atoms Structures I-XIX and XXIII are numbered according to Harris and Knight (reference 20) From references 9 12 and 20 grown from clear non-colloidal silicate solutions very low in A1 Thus the consensus2 is that the formation of zeolites in general involves redissolution of any precipitated gel (or meta- stable zeolite intermediates) rather than internal reordering of the gel itself Water is important as a guest molecule in zeolite structures with relatively high A1 contents and consequently aqueous media favour their formation while salts have a parallel role in the stabilization of sodalite and cancrinite Highly siliceous zeolites such as silicalite and ZSM-5 which have zeolitic struc- tures but contain very little AllI1 and M'"+ are less hydrophilic than aluminous zeolites and their structures can be stabilized by certain organic guest molecules -notably amines alcohols and amino-alcohols There is however no obvious general correla- tion between the geometries of the organic molecules and the structures they promote Similarly although cations such as R4N+ are commonly said to direct the structure of the zeolitic product by acting as a template this should not be construed to mean that the ions act simply as objects of the right geometry around which Al- and Si-containing units assemble themselves Thus we need to understand not only the speciation and properties of aqueous Al- and Si-containing anions but also how these are affected by cations and organic molecules lo Direct correlations between the structures of precursor solutes and zeolite products however may be difficult to establish For example although the structure of ZSM-5 is based on pairs of rings containing five silicon atoms (linked by oxygens) there is no definite proof that the 'double-five-ring' aqueous silicate anion (Q:o in which Q' represents an SiO centre of connectivity x species XXIV of Figure 2) is a ~ precursor of ZSM-5 even though ZSM-5 certainly forms in solutions in which XXIV is known to be the predominant silicate species * Furthermore the smaller silicate1* and especially the aluminosilicate' species that have been identified in alkaline aqueous solutions are very labile in the exchange of silicate and/ or aluminate units although some larger silicate oligomers such as the cubic octamer (Qi) are relatively inert l4 Consequently the structure of aluminosilicate solids grown from aqueous solutions or gels need not necessarily reflect the structures of the predominant solute species as these with the exception of the 32 1AQUEOUS ALUMINATES SILICATES AND ALUMINOSILICATES-T W SWADDLE ET AL larger oligomers might rearrange 'on demand' at the growing crystal surface For example Figure 1 shows that the zeolite A structure can in principle be assembled by the corner-linking of Qi(cyclic tetramer species X) units but these are in labile equilibrium with other small species such as I-V and VII l2 Conversely if the larger oligomers are building-blocks for growth of the solid phase the rate of crystallization could be controlled by the rate of formation of these in solution One source of difficulty is that the mechanisms of nucleation and of subsequent crystal growth might well be different -small labile oligomers may add rapidly to a nuclear structure that formed only slowly In any event it is clear that knowledge of the kinetits of aluminate silicate and aluminosilicate exchange reactions at various pH and temperatures is a prerequisite for understanding hydrothermal zeolite formation 5 Toxicology of Aluminium and the Biological Role of Silicon Several excellent articles on the role of aluminium in medicine and biology have recently appeared l5 l7 Debate has tended to centre upon the highly controversial question of the apparent association of Allii with Alzheimer's disease (AD) -a form of senile dementia characterized by the development of neuro- fibrillary tangles and P-amyloid protein plaques and loss of neurons in the patient's brain Since 1973 several research groups have reported (though some have denied) that post nioi tern examinations of brain tissue from victims of AD showed aluminosilicate material within the P-amyloid plaques by con- ventional microscopy This seemed to vindicate inferences of the involvement of A1 in AD made from epidemiology {there are correlations between the incidence of AD and the A1 content of water supplies) observations of Al-induced dementia amongst renal dialysis patients retardation of the progress of the disease by administration of the A13 +-chelating agent desferrioxamine and so forth but in late 1992 Landsberg et al used nuclear microscopy to study neuritic plaques and found no evidence for the presence of increased levels of A1 in them They pointed out that aluminium is so widespread in the environment (e g in dust) that it is extremely difficult to eliminate it from experimen- tal materials -in this case from the staining solutions necessary in conventional microscopy This echoes a familiar problem in the analytical chemistry of low levels of Al in which samples often appear embarrassingly to have negative A1 contents relative to controls Medical interest in A1 from the standpoint of AD research has therefore waned in recent months (perhaps wrongly) but the broader issue of the toxicology of A13+ remains -for toxic it most certainly is being unequivocally implicated in bone deter- ioration (0steomalacia) and dementia (dialysis encephalopa t h y) in dialysis patients Because A13 + is strongly bound by transfer- rin in the blood a few tens ofpmol 1 of A]"' in the dialysis fluid (as when the fluid is made up with tapwater in some locations) will prevent A1 excretion and in fact cause Aliil to pass in reverse into the bloodstream leading to toxic effects The symptoms of A1 intoxication generally disappear when 'A1 free' dialysis fluid is used Fish are very susceptible to gill damage by A13 (aq) and massive fish kills through acidification of lakes to + pH < 5 by acid precipitation reflect the dissolution of the normally poorly-soluble Al(OH) Root growth in dcid-sensi- +tive plants is suppressed by The toxic action of AI3 at the cellular level is incompletely understood,' l9 but probably it displaces Mg2+ in several processes and binds strongly to phosphate functions in ADP ATP and phosphorylated proteins Our daily intake of A1 in food and water is typically 2-25 mg but can be much more [e g ,if AI(OH),-based antacids are taken] Why then is A1 poisoning rare? First the toxic form of AI"' is AI3 +(as) 7h whereas Martin7 has shown (Figure 3) that at physiological pH the only important form of aqueous Al"' is AI(OH) and the saturated concentration of AI3+(aq) becomes su b-nanomolar Second ingested A13 becomes firmly + 10 c 0 cU2 05 -a r" 0 5 Uc 8 M 10 20 I 15 202 3 4 5 6 7 8 9 10 Figure 3 Distribution of monomeric species ds d function of pH in hydrolysis of Al"' Upper frame mol fraction of soluble species Lower frdme molar concentrdtions of A13 (ddshed line) Al(OH) + (dotted line) dnd total dissolved Al (TAI solid curve) (Reprinted by permission from reference 19 Copyright 1991 bv Elsevier Science Inc ) bound to the intestinal mucosal cells and is excreted through the lumen as these die so that only some 2 pg A1 normally gets past the gastrointestinal barrier per day while up to 20 pg per day can be excreted in the urine Third the concentration of free A13+(aq) can be still further depressed by complexation with chelating agents such as citrates or maltol which are common constituents of foods (there is however a risk of enhanced A1 absorption with citrate since the chargeless citratoaluminium complex can pass relatively easily through biological bdrriers) or with dissolvedszlrca Silica is beneficial to many living things -e g it is essential for normal development of bones in rats and chickens -and Birchal17 l7 has argued convincingly that in many cases this property derives from complexation of A]"' by aqueous Si(OH) to give soluble alurninosilicate species such as (HO),SiOAl(OH) so depressing the free AI3 +(aq) level still further Thus the apparent correlation of the incidence of dementia with the A1 content of drinking water may actually reflect an inverse correlation with dissolved silica (and so expldin the absence of a dose-effect relationship in the A1 intake most of which comes from food not water) Similarly dissolved silica can protect fish from acute A1 poisoning in acidic ground waters 7d The characterization of these postulated soluble aluminosilicates is therefore important in biomedical contexts 6 Experimental Methods Until relatively recently it was difficult to learn much about aluminium and silicate solutes because of the lack of characteris-tic A1 or Si features in UV/VIS spectra (the weak UV bands sometimes reported for aluminates at high pH are probably artefacts due to traces of transition metal ions) redox electro- chemistry etc while pH measurements conductiometry and vibrational spectra give ambiguous results Trimethylsilylation of silicate solute species gives volatile products that can be separated and identified by gas chromatographic methods4 -for example the cubic octamer Si,O:; gives rise to Si8012 (OSiMe,) -but since many of the smaller silicate oligomers are very labile,12 l4 trimethylsilylation is likely to perturb the 322 equilibria among them and the yields of the trimethylsilylated product may not represent the original distribution of silicate solute species Nowadays however. with the i outine availability of Foul lei transform multinuclear magnetic resonance spectroscopy at magnetic fields of 4 7 Tand higher structural. thermodynamic. and kinetic information specific to 29Si 27Al and 170can be obtained 'non-invasively ' I . without perturbing the solution chemically Furthermore. NM R spectrd (chemicdl shifts) of species in solution can be compared with those of relevant solids of known structure Silicon-29 is the most useful nucleus,l0 l4 2o 23 as it gives sharp-line spectra (spin quantum number I = +) and hence detailed structural information. in the absence ofchemical exchange Its natural abundance (4 7%) and recepti- vity (about twice that of ") however are not high Additional structural information (from 29S1-29S~ coupling) as well as increased sensitivity is obtainable with highly enriched 29Si but this currently costs dround $(Us)8000for 100mg of 94% 29sias SiO Kinetic information on Si site exchange can be extracted from 29S1 line broadening,I2 21 2D-NMR,14 or selective inver- sion recovery methods l2 22 The longitudinal relaxation time T of 29S1 in aqueous silicates is fortunately much shorter than in solids or organosilicon compounds but is dependent in a perplexing way on the nature of the cation ion pairing is presumably involved 23 Aluminium-27 has good receptivity and 100% abundance but is quadrupolar (I = 3) and so gives rather broad lines 24 Oxygen-17 also has I = 4but its receptivity is low [0 061 that of 13C,natural abundance 0 037% 20% enriched water costs about $(US)320 per g] and it has been little used in this field 25 7 Silicate Solutions The solubility of silica in water is not simply stated as it depends markedly on the particle size and form of the solid (quartz cristobdhte tridymite vitreous etc ) and the degree of polymeri- zation of the solute and increases with increasing alkalinity temperature and pressure Z6 As working numbers however the solubility of quartz in pure water at 25 "C and 0 1 MPa may be taken to be 1 1 mg kg l. and that of amorphous silica to be 60-200 mg kg Sjoberg and co-workersz7 found the pK,'s for acid ionization of monomeric silicic acid Si(OH) and (HO),SiO to be 9 47 and I2 65 respectively at 25 "C and ionic strength 0 6 mol 1 I with various values for silicate oligomers [eg for (HO),SiOSi(OH),O .10 251. so that the average chargepet Siis essentially -1 at pH z 10-12 with correspondingly high solubilities The definitive work of Harris and KnightZo used high-field (1 1 75 T) 29Si NMR of enriched samples with homonuclear 29S~-:29S~;decoupling to identify unambiguously 18 silicate oligomers in KOH/SiO solutions with K+ Si = 1 1 to 2 I [Si] = 0 6 mol 1 I these are the species denoted by I-XVIII in Figure 2 Silicate centres with four-way connectivity (Q") are generally not observed in solut~on,~ evidently because their ' formdtion would nucleate gelation Knight et ul 28 have subse- quently used multiple quantum filtered NMR to confirm the presence of six (but onli six) single-Si-site silicate species in solutions with K+ Si = 1 1 [Si]z 1 5 mol 1-I these are species I 11. V X. XI. and with less certainty the P,O analogue XXIII * In particular the cubic (Qi,species XVII) and penta- gonal prismatic (Q:o species XXIV) structures identified in aqueous Me,N and Pr,N +/aqueous dimethylsulfoxide media + iespectively." l4 and the hexagonal prism characteristic of many zeolite structures were not detected ( < 3% of [monomer]) This illustrates the strong influence of cations on the distribution of silicate species in solution Similarly Kinrade and Pole" found that complexdtion of Nd + in sodium silicate solutions by CHEMICAL SOCIETY REVIEWS 1994 expected to pair more strongly than the larger On this basis the observations suggest that pairing of the smaller more labile silicate units with cations deuctivates them towards condensa- tion to the cage structures Lowered Si concentrations or increases in temperature or pH favour the monomer and small oligomers,12 so that the 29S1 spectrd become simpler as the alkali to Si ratio is increased Increased temperature and [Si] however also cause line broad- ening attributable largely if not entirely to the increased rate of Si site exchange between silicate units mediated mainly by the attack of Si(OH) or possibly (HO),SiO on 0-Si-in the cdse of intermolecular exchange or by intramolecular ring opening and closure (e g interconversion of species 111and V or of VII and the less labile X) the intramolecular processes being the more rapid ' * The dependence of the line-broaden- ing on [Si] reflects the intermolecular site exchange pathways but no simple rate equation can be derived because of the plethora of known and hypothetical12 silicate species that are likely to be simultaneously involved The doubly deprotonated monomer (H0)2SiOI and presumably other Si centres with more than one local negative charge are relatively inert towards exchange,' with the result that Si site exchange is slower and 29Si spectra sharper with fewer lines at high pH Caution must be exercised when interpreting NMR line broadening in terms of chemical exchange since field inhomoge- nities ion pairing relaxation by paramagnetic impurities. and other factors might mimic chemical exchange in causing line broadening that increases with rising temperature but the selec tile intersion recowj technique introduced by Creswell et a1 22 shows clearly that magnetization is transferred between Si sites at measurable rates that can only be due to chemical exchange It should be noted that the 29Si selective inversion recovery experiments described to date' 22 have been con-ducted at higher pH than most line-broadening studies for the sake of simplified spectra and so show slower siteexchange rates (eg site exchange time constants decrease by about 100-fold when the KOH SiO ratio is increased from 1 1 to 45 1) Because of the complexity of the exchanging systems it is difficult to extract reliable rate constants for specific reactions but for hydrolysis of O(HO),S~OSI(OH)~O the first order rate constant is 14 s-l at 25°C and AH* is 51 f2 kJ mol-' Thus substitution in this small oligomer is rapid on the prepara- tive timescale In contrast in Me,N silicate solutions that have + been quenched from boiling (where the monomer and small oligomers are the most abunddnt species) the prismatic hex- amer XI 'grows in' over a period of several hours at room temperature while the ultimately dominant cubic octamer XVII continues to grow at the expense of all other silicate species over several days l4 Rigid Q3cage species such as XI and XVII will be sterically very resistant to the SN2-type reaction mechanism that is typical of SiIV centres so once formed they may become 'black holes' into which smaller oligomers disappear From the standpoint of zeolite formation mechanisms however it would seem that the predominant aqueous silicate species at synthesis temperatures (usually 2 100°C) are the smaller labile oli- gomers so that rapid reorganization of these rather than the presence of specific preformed cage structures in solution at equilibrium may provide the primary route to crystal growth although nucleation may still require the formation of a cage 8 Aluminate Solutions Figure 3 summarizes the speciation of dilute All'' solutions ds a function of pH A striking feature is the very narrow pH range (compared with. say Fell1) between the predomindnce of A1(H20)3,+ and that of Al(OH) with only minor roles for A10H2+,AI(OH); and dissolved Al(OH)$ this reflects the cryptdnd 2 2 2 led to large increases in the (albeit still small) facility with which All'' can chdnge from 6- to 5-to 4-coordind- fractions of the single-Si-site cage structures XXIII (Qi),tion l6 I9 By contrast SiO centres form only at pressures of XVII (Qi)and less markedly XI (Qi)Undoubtedly anion-cation several GPd pairing is one of several influences acting on the species distribu- The 27Al-NMR-based work of Akitt's has done ti0r-1,~lo 23 with the smaller cations (hydrated alkali metal ions) much to clarify the chemistry of Al"' hydrolysis In acidic AQUEOUS ALUMINATES SILICATES AND ALUMINOSILICATES-T W SWADDLE ET AL solutions partially hydrolysed AI3 '(aq) can dimerize to (H20),A1(~-OH),Al(OH,)::+ but this is not normally found in significant concentrations l9 Indeed solid salts of this ion disproportionate in water to the monomer and polymers show- ing once again that solid phases need not show any direct structural relation to their progenitors in solution If however the temperature basicity and [All are kept low during hydroly- sis to avoid the formation of higher polymers,29 the remarkably stable tridecamer (AlO,)Al ,(OH),,(OH2)7; can form 24 This ion contains a central A13 + 111 a highly symmetrical tetrahedral0 environment surrounded by four blocks of three octahedralli coordinated edge-sharing A10 units (an isomer of the familiar Keggin structure) The quadrupolar 27Al nucleus gives relati- vely narrow NMR lines only when the electric field gradient at the nucleus is small or zero so that the single central A1 gives a well-defined NMR line but the 12 peripheral A1 nuclei are hard to detect because the comparatively low symmetry of their environments causes excessive line broadening 24 This problem of 'missing Al' is widespread and plagues efforts to use 27Al NMR quantitatively particularly at low magnetic field strengths and temperatures Early work on All1' speciation at pH 2 7 -that is the pH range relevant to the Bayer and kraft processes has been -summarized by Eremin et a130 From infrared/Raman and NMR spectra and other information it was recognized that the solutions were not colloidal and that the only important A1 species at 7 IpH I 13 is tetrahedral AI(OH) whether or not the solutions are allowed to 'age' (which results in some drift in pH if this is initially in the 8-10 range) Linear AIO; square planar Al(OH) AI(OH),(OH,); and significant concentra- tions of polymeric aluminates were specifically ruled out At pH > 13 however it might be anticipated that All1' might expand its coordination number to six to form AI(0H) ,which is known as the solid calcium salt while at high [All and low water activities AI(OH) does appear to condense to (HO) AlOAI(0H) ,the potassium salt of which is of known struc- ture and has infrared and Raman spectral features that corres- pond to those that appear in aluminate liquors as [All is raised above 1 5 rnol 1-Measurements of pH at high alkalinities are impractical because the electrodes are attacked Even liquid- phase *'AI NMR is not very helpful in concentrated solutions,24 as anion-cation pairing the higher viscosity and rapid AI- A1 site exchange may broaden the lines while the reduced site symmetry in any oligomeric species present may lead to a 'missing Al' problem Indeed the presence of polymeric alumi- nates may be revealed by the reduction in intensity of the Al(OH) line Nevertheless Akitt et a1 24 found 27A1-NMR evidence for Al(0H);- in addition to AI(OH) in extremely alkaline solutions (23 moll * NaOH 0 95 moll All'') while a recent solzd state 27Al NMR of Na aluminates precipi- tated by acid hydrolysis from solutions of high pH (OH A1 = 3 9-5 3 [All = 1 mol 1 ') suggests that some polyoxoa- nion(s) containing six-coordinate All1' are present in the solu- tion phase A critical piece of missing information in aqueous aluminate chemistry is the rate of substitution at A1 centres If chemical exchange in aluminate species were rapid on the NMR time- frame little or no structural information could be gleaned from 27A1 NMR spectra Kinetic information on aqueous alumino- silicates (see below) suggests that this may indeed be the case at room temperature and above and preliminary 170 NMR line- broadening studies in our Melbourne laboratory on the aqueous AI(0H)JOH exchange rate seem to confirm this rapid exchange process 9 Alumi nosi Iicate Solutions Engelhardt and Michel reviewed the NMR literature on aqueous aluminosilicates up to 1987,at which time the existence of aluminosilicate species in alkali metal hydroxide solutions had not been proven because of low solubility in typical prep- arations two Raman studies for example had led to opposite 4 0 -4 -8 -12 -16 -20 -24 ppm Figure 4 Vertically expanded 29S1NMR spectra (79 49 MHz) of solu- tions containing 1 1 mol kg SiO and 2 4 mol kg aqueous NaOH at 5 "C the lower spectrum is of a solution that also contained 0 07 mol kg Al"' The spectra are normalized to give equal areas of the Qo resonance and include 2 4 Hz drtificidl line-broddening (Reprinted with permission from reference 13 Copyright 1988 Ameri- can Chemical Society ) conclusions Experience with the 29S~ NMR of solid aluminosili- cates however enabled them4 to predict the likely chemical shift ranges of Si atoms in a Q\(J Al) environment (where Y is the total connectivity through oxygens to Si and A1 atoms and 1. is the number of those atoms that are Al) Thus it was expected that a Q'(1Al) resonance would generally lie about 5 ppm downfield from an analogous Q' line It was recognized that such features would be most likely to be observable in metastable solutions such as were described by Dent Glasser and Harvey,' but no lineshifts or new 29S~NMR lines were found only some broadening of the Qo and Qi lines We suggest that this broadening was due to rapid formation and hydrolysis of AI-0-Si links this would imply that the already weak aluminosilicate Si lines would be broadened too much to be observed and also that Al-0-Si links are more labile than their Si-0-Si analogues since the latter gave sharp 29S~ lines under the same conditions Thus Kinrade and Swaddle,13 again using Dent Glasser- Harvey supersaturated solutions were able to demonstrate the existence of Al-0-Si species with 29S1 NMR on a 94 Tspectrometer by cooling the solutions to 5 "C to slow down this inferred aluminate-silicate exchange Figure 4 shows that the addition of a small quanity of sodium aluminate solution to an excess of aqueous silicate produced new 29S1 resonances due to aluminosilicate solute species 4-5 ppm downfield from silicate lines of known origin and also broadened the silicate line to an extent consistent with Si site lifetimes in the aluminosilicate environments of some 60-120 ms implying exchange rates some lo2-lo3 times faster than for comparable silicate sites l3 Because of some overlapping of the new aluminosilicate bands A-E in Figure 4,the identification of the individual aluminosi- licate solute species (Figure 5) remains to some degree tentative M IIa IIIa IIIb IVa Va Xa XIa XVIIa Figure 5 Structures of aqueous aluminosilicate species assigned tentati- vely from 29S1and 27Al NMR spectra Code and numbering follow Figure 2 open circles represent Al atoms From references 4 13 23 and 34 324 CHEMICAL SOCIETY REVIEWS 1994 137 110 80 50 20 PPm Figure 6 27AlNMR spectra (52 12 MHz) for the sodium aluminate/ silicate solution of Figure4 Spectra are not normalized.and include 2 Hz artifical line broadening (Reprintedwith permission from reference I3 Copyright 1988 Ameri- can Chemical Society ) Figure 6 shows the 27Al NMR spectrum of the aluminate/ silicate solution of Figure 4 three distinct 7Al resonances emerge as the tempercture is raised from 4"C in accordance with the expected sharpening of quadrupolar resonance lines with rising temperature but above 41 "C this structure is lost because of chemical exchange line broadening Figure 6 con-firms that the aluminosilicates present are 2-3 orders of magni- tude more labile than the analogous silicate species Again the limitations placed on NMR-based structural studies by chemi- cal exchange kinetics are emphasized Bell and co-w~rkers,~~ using high-field (1 1 7 T) 29Si,27A1 23Na and 13,Cs NMR techniques including DANTE 'hole- burning' found that the extent of aluminosilicate formation in Na or Cs aluminate/silicate (A1 Si = 1 0) solutions increased with the ratio of silica to alkali I e ,as the solutions became less alkaline and that aluminate ions reacted more extensively with the more acidic silicate species such as the cyclic trimer (QZ) -in accordance with quantum mechanical calculation^^^ that pre- dict that the substitution of Si by A1 in an aluminosilicate frame- work should be favoured by the presence of a proton The ions solutions with Si A1 = 4 three new 29S1 NMR peaks show up clearly in addition to the singlet due to the Qi of silicate XI (Si,Oy ) the new peaks arise from the three distinct Si sites relative to the single A1 in the monosubstituted trigonal prisma- tic cage Si,AlO~ Ziemens et a133 found 29S1 NMR and trimethylsilylation evidence for the cubic Si,AlOg coexisting with Si,O! (silicate species XVII) in triethyl(hydroxyethy1)am-monium a1uminate:silicate solutions while a more detailed 29S1 NMR study of aqueous and methanolic Me4N+ aluminatel silicate solutions by Bell's group34 found these same species in high concentrations Aluminium-27 NMR of Me,N alumi-+ nate/silicate solutions has similarly provided evidence for A1 centres Al(OSi),,(O with Si connectivity n 24 Fahlke et u/ 35 noted that 27A1 NMR features attributed to aluminosili- cate solute species in the synthesis mixtures of zeolites X and Y were broad and interpreted this breadth as reflecting the presence of many such species but we suggest that it could also indicate rapid chemical exchange of aluminate units in these species as noted above Finally Bell et a1 34 have used both 29Si and "A1 NMR spectra ofdilute but highly alkaline tetrapropyl- ammonium aluminate/silicate solutions to identify Si centres with Q1(lAl) Q2d( IAl) {Q"(lAl) + Q"(lAl); and Q3(lAl) connectivity (Asignifies a 3-A1/Si ring) and A1 with Si connecti- vity 0 I 2 and 3 (Figure 5) Interestingly the trigonal prismatic hexameric silicate (Qi,species XI) appeared not to react with Al(OH) in these solutions which is consistent with Engelhardt and Michel's observation4 on Et,N solutions that the hexamer + is unreactive unless broken up at high temperatures Additions of 20% or more dimethylsulfoxide to the solutions increased the dbundances of silicate species XI and XVII whereupon new features in the 29Si and 27Al NMR spectra corresponding to A1 substitution into these were observed The readiness with which All" becomes incorporated into tetrahedral sites in aluminosilicates contrasts with the reluctance of Al(OH) units in solution to join together to form aluminate oligomers In aluminate solutions in aqueous alkali even the dimer (HO),AIOAl(OH)i forms (if at all) only under forcing conditions of low water activity Such polymers of as are known to exist in water (including the tridecamer which has one tetrahedral A1 surrounded by linked octahedral AlO units) are formed at pH < 7from octahedral A10 units as are also found in the gibbsite that precipitates from (alkaline) Bayer process liquors In aluminosilicate mineralogy this avoidance of direct links between A10 tetrahedra is known as Loewenstein's rule and may be thought of in terms of minimizing electrostatic repulsions between the excess negative chdrges associated with the A10 centres Recent theoretical studies3 show that extended aluminosilicate structures with Al-O-A1 links would be some 120 kJ mol-1 higher in energy per Al pair than equivalent structures with Al-O-Si-O-A1 links The experi- mental cvidence available to date suggests that All1' in aqueous aluminosilicates avoids A1-0- A1 links and prefers coordina- were also detected but the formation constants of these (par- ticularly the latter) seemed to be small under the experimental conditions Increased cation size (reduced ion pairing) also favoured aluminosilicate formation and this is precisely as would be expected if as proposed above pairing of the smaller oligomers with cations deactivates them toward reaction with further silicate or aluminate units The solubility limitations of alkali-metal aluminosilicate solu- tions can be avoided by going to R,N+ solutions but in these media specific silicate cage structures such as XI and XVII of Figure 3 tend to predominate Engelhardt and Miche14 summar- ize trimethylsilylation and 29Si NMR evidence for the presence of Q3(3A1) centres (presumably substituted into structure XVII) and less certainly some Q2(2A1) species in Me,N+ aluminate/ silicate solutions that have first been heated to 90°C and then cooled to room temperature Interestingly little or no substitu- tion of Si in cages by A1 seems to occur in unheated solutions showing once again that silicate cage anions are much less labile than the smaller oligomers In tetraethylammonium (Et4N +) and (HO),AlOSiO,H$~~~~)- tion by two Si tetrahedra over single connectivity l3(HO),AIOSi(OH),0Si03H$~~~)-The stability of aqueous aluminosilicates is difficult to express quantitatively At pH up to 12 solutions contain a large number of silicate oligomers into which one or more aluminate units could be incorporated At higher pH the stabilities of alumino- silicates are reduced and there are uncertainties over the possible occurrence of oligomeric aluminates or six-coordinate A1 centres as in Al(0H); ,as well as over the degree of secondary deprotonation of the (less numerous) silicate species Thus in a study of the solubility of sodalite at 95 "C (ionic strength 4 0 mol kg-I NaOH/NaCl) with [OH-] = 0 1 to 4 0 mol kg I Gas-teiger et a/ found that the apparent [Al][Si] product rose strongly with increasing [OH 1 but this reflects the deprotona- tion of (HO),SiO and (HO),SiO rather than any increase in the stability of soluble aluminosilicates Yokayama et a1 ,,'using 27Al NMR found a foimation constant of 22 8 1 rnol ' for H ,O,SiOAl(OH)'$+ from monosilicate and AI(OH) in 0 1 moll NaOH at 25 "C but no detectable reaction in 1 0 mol 1-I NaOH At physiological pH (74) and 25 "C Mar-tin3 estimates K = [AlOSi(OH)i +]/[A13 +][Si(OH),] = 3 x lo4 AQUEOUS ALUMINATES SILICATES AND ALUMINOSILICATES-T W SWADDLE ET AL 1 mol (the actual species present would be (HO),SiOAl(OH) dnd Al(OH),) which given that [Si]in blood plasma = 20 pmol l. means that some 60% of the Al"' burden of the bloodstream IS probably bound to silicic acid Exley and Bir~hall~~ have confirmed the importdnce of soluble aluminosilicates under physiological conditions and show that complexation of Al(OH) by silicic dcid inhibits the nucleation of Al(OH) precipitates At the still lower pH range (40-5 5) of acidic natural wdters Allll-Si(OH) binding IS complicated by the dcid ionization of A13+(aq) and its silicic acid complexes but Browne dnd Driscol15 estimate from fluorimetric measurements using morin complexdtion of free All1' that up to 95% of the total inorganic mononuclear Al"' is present as soluble alumino- silicates so that these may be controlling factors in the weather- ing of rocks or soils leading to new mineral deposition 10 Summary Aqueous All1' and Si'" readily form aluminosilicate complexes that can have significant solubilities particularly if gelation of aluminate/silicdte mixtures IS not prompt Their thermodyna- mic stabilities are lower at high pH The structures of these aluminosilicate species are much like those of the numerous silicate oligomers that have been characterized in alkaline aqueous solution by NMR methods and stand in contrast to the very limited rangc of structures known for aqueous aluminate species The effects of temperature pH and cations on the speciation nnd thermodynamic stability of Al Si and alumino- silicate oligomers are profound and need further clarification The larger cage-like silicate anions are kinetically rather inert at room temperature but the small silicate and especially alumino- silicate species are very labile These dynamic and thermodyna- mic aspects have far-reaching industrial biomedical environ- mental and scientific implications and further data are urgently needed particularly for aluminates Acknobz fedgernent We thank the Natural Sciences and Engi- neering Research Council of Cdnddd (TWS) and Comalco Aluminium Ltd (JS PAT) for financial support TWS thanks the University of Melbourne for a visiting fellowship 11 References L S Dent Glasser and G Hdrvey J Chem Soc Chem Commun 1984,664 I250 'Zeolite Synthesis' ed M L Occelli dnd H E Robson ACS Symposium Series 398 American Chemical Society Washington DC 1989 R M Bdrrer 'Hydrothermal Chemistry of Zeolites' Academic Press London 1982 G Engelhdrdt and D Michel 'High Resolution Solid-state NMR of Silicdtes dnd Zeolites ,John Wiley Chichester 1987 B A Browne dnd C TDriscoll Science 1992 256 1667 H A Gdsteiger W J Frederick dnd R C Streisel /nd Eng Chem Re\ 1992 31 I183 'Aluminium in Biology dnd Medicine' ed D J Chddwick dnd J Wheldn Cibd Founddtion Symposium 169 John Wiley and Sons Chichester 1992 G T Kerr. J Phis Chem . 1966 70 1047 1968 72 1385 9 D M Ginter C J Radke and A T Bell in 'Zeolites Facts Figures dnd Future' ed P A Jacobs dnd R A vdn Sdnten Elsevier Amsterdam 1989 pp 161-168 A V McCormick and A T Bell Cuter1 Rev Scz Eng 1989,31 97 10 W M Hendricks A T Bell and C J Radke J Phjs Chem 1991. 95.95 13,95 19 11 S D Kinrade and D L Pole Inorg Chem 1992. 31,4558 12 S D Kinrade and TW Swaddle Inorg Chem 1988 27 4253 4260 13 S D Kinrade and T W Swaddle Inorg Chem 1989,28 1952 14 C T G Knight R J Kirkpatrick and E Oldfield J Chem Soc Chem Commun 1986 66 J Mngn Reron 1988,78 31 15 Toxicologicdl Profile for Aluminum' U S Dept of Health dnd Humdn Services Public Health Service Agency for Toxic Sub- stances and Disease Registry TP-91/01. Washington DC 1991 16 'Aluminum in Chemistry Biology and Medicine' ed M Nicolini P F Zatta and B Corain Vol 1 Raven Press New York 1991 B Corain A Tapparo A A Sheikh-Osman and G G Bombi Coold Chem Re\ 1992 112 19 B Corain M Nicolini dnd P Zdttd Coord Cheni Re\ . 1992 112. 33 17 (a)J D Birchall in 'Food Nutrition dnd Chemical Toxicity' ed D V Parke C Ioannies and R Walker. Smith-Gordon London 1993 pp 215-226 (6) C Exley and J D Birchall J Thcw Bid 1992 159 83 (c) J D Birchall Chem Br ,1990 141 ((0J P Bellid J D Birchdll dnd N B Roberts Luncet 1994 343 235 18 J P Landsberg B McDonald and F Watt Nature 1992.360 65 19 R B Martin J Inorg Biochem 1991,44. 141 20 R K Hdrris and C T G Knight J Mu1 Struct 1982,78 273 J Chem Soc Furudcii Trans 2 1983 79 1525 1539 21 R K Hdrris J J0nes.C T G Knight,dnd R H Newmdn J Mol Liq 1984,29,63 22 C J Creswell R K Harris,andP T Jageland,J Chem Soc Chem Comniuri 1984 126 1 23 A V McCormick,A T Bell,andC J Radke J Phrr Chem 1989 93 1733 1737 1741 24 J W Akitt Prog NMRSpectr 1989-21,I J W Akitt W Gessner and M Weinberger. Magn Reyon Chem 1988,26 1047 25 C TG Knight A R Thompson A C Kunwar H S Gutowsky E Oldfield dnd R J Kirkpatrick J Chem Soc Dulton Truns 1989 275 26 R K Iler 'The Chemistry of Silica' Wiley-Interscience New York 1979 27 I L Svensson S SJOberg and L -0 Ohman J Chem Soc Faradui Trans 1. 1986. 82 3635 28 C T G Knight R J Kirkpdtrick and E Oldfield J Chem Soc Chem Commun 1989,919 29 J J Fitzgerdld L E Johnson dnd J S Frye J Mugri Rtwn ,1989 84 121 30 N I Eremin Y A Volokov dndV E Mironov Rush Chem Re 1974,43,92 31 S M Bradley and J V Hanna J Chem Soc Chem Conimun. 1993 1249 32 E G Derouane J G Fripiat and R von Ballmoos J Phi s Chem . 1990 94 1687 33 0 Ziemens 0 Rademacher dnd H Scheler Z Chem 1989 29 34 1 34 R F Mortlock A T Bell dnd C J Rddke J Phi5 Chem l991,95. 372 7847 1993 97 775 R F Mortlock A T Bell A K Chakraborty and C J Radke J Phir Chem 1991,95,4501 35 B Fahlke D Muller and W Wieker Z Anorg Allg Chem 1988 562 141 36 K -P Schroder dnd J Sduer J Phi.\ Chem 1993,97,6579 37 T Yokayama S Kinoshita H Wakita and T Tdrutdni Bull Chem Soc Jpn 1988,61 1002 38 R B Martin Polrhedron 1990,9 193 39 C Exley and J D Birchall Poliheclron. 1992 11 1901 1993 12 1007

ISSN:0306-0012    

DOI:10.1039/CS9942300319

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

I,lo-Phenanthroline: A Versatile Ligand Peter G. Sammes and Gokhan Yahioglu Molecular Probes Unit, Department of Chemistry, Brunel University, Uxbridge, Middlesex UB8 3PH, U.K. 1 Introduction I, 10-Phenanthroline (1) is the parent of an important class of chelating agents. Compared to the more common 2.2'-dipyridyl system (2), 1,lO-phenanthroline has several distinct properties: the rigid structure imposed by the central ring B means that the two nitrogen atoms are always held in juxtaposition, whereas, in (2). free rotation about the linking bond allows the two nitrogens to separate (2a *2b), in particular under basic or strongly acidic conditions. This entropic advantage for I, 10-phenanthroline means that complexes with metal ions can form more rapidly.a property of importance. for example. in the formation of cooperative complexes with lanthanide ions. Another conse- quence of the planar nature of 1,lO-phenanthroline is its ability to participate as either an intercalating or groove-binding species with DNA and RNA. One other important property of the phenanthroline nucleus is its ability to act as a triplet-state photosensitizer. especially in complexes with lanthanides such as europium. The metal chelating properties of I, 10-phenanthroline have been utilized in a range of analytical reagents and probes as well as herbicides. Since the seminal review of Summers2 in 1978, several natural products incorporating this heterocyclic nucleus have now been isolated, several of which possess interesting anti- cancer proper tie^.^ In the last decade the phenanthroline group has also been exploited by workers interested in molecular recognition and self-assembling systems4 This review covers some of the recent chemistry of 1,lO- phenanthroline and its derivatives and their use as chelating agents for the development of bioorganic reagents and probes.2 Synthetic Studies The enormous interest in molecular recognition processes and the increased use of 1,lO-phenanthrolines in this area has renewed interest in the synthetic manipulation of these systems. 2,9-Dimethyl- 1,IO-phenanthroline (3) ('neocuproine') has been converted into a range of oxidized derivatives (Scheme I). Selenium dioxide affords the dialdehyde (4), that can either be oxidized to the diacid (5)or reduced to the diol(6).Formation of the bisoxime from the dialdehyde and dehydration affords the dinitrile and, by reduction, the diamine (7). The diol (6) can be Peter Scrmmes HYIS trained at Imperial College. He held chairs in Orgunic Chemistrji at Citjg Universitj>, London und the Universitjq of Leeds before joining Brunel Universitjl in 1989. He spent several jaws in industry, ,first Li-ith Glaxo us a Junior Labora t ory Ass is tan t before he entered academia and, niore recentlj>, as a research Vice President at SmithKline und French. His current research trwk is jbcused on a Iwrietj. of mol-ecular probes. He has published over 200 research papers. 65 converted into its dibromide with HBr and this has been aminated to give a range of chelating agents.A useful alternative route to the diacid (5) is available by initial perchlorination of neocuproine with N-chlorosuccini- mide, followed by acid hydr~lysis.~ Methyl groups at positions 2,4,7. and 9 on 1,lo-phenanthro-lines can be lithiated with lithium diisopropylamide and then alkylated;6 the 3.8-dimethyl groups are resistant to such li thiation. 1,lO-Phenanthroline (I j can be oxidized to the mono-N-oxide (8), steric constraints precluding formation of the bis-N-oxide. The N-oxide reacts with benzoyl chloride and potassium cyanide to give the nitrile and, by hydrolysis, the monocarboxylic acid (9).3The nitrile hydrolysis is strongly assisted by metal ions such as Cu2 and Ni2 +,possibly by initial chelation and intramole- + cular delivery of hydroxide ions from the metal.' In I, 10-phenanthroline, the 5,6-double bond is most suscept- ible to electrophilic attack and even epoxides can be formed from it.* Nitration of either the parent (I) or neocuproine (3) leads to the 5-nitro-derivatives, e.g.(10) generally accompanied by some of the 5,6-dione (1 l), and the ratio of the two products can be partly controlled by careful choice of conditions.2 Reduction of the nitro-group to the amine is best achieved by a catalytic transfer hydrogenation. The recent reaction to convert &unsaturated nitro-com-pounds into pyrroles can be applied to the nitro-phenanthro- lines to give products of the type (I 2).9 The dione is also of importance since condensation with 1,2- diamines produce the pyrazines. e.g.1,2-diaminobenzene produces the chelating compound (1 3) (DPPZ), which acts as a solvent-sensitive ligand when chelated to ruthenium (see below).lo The introduction of aryl groups into positions 2 and 9 may be achieved by adding lithium aryls. Thus Sauvage treated phe- Gokhan Yahioglu NUS born in Cj'prus,greitq up in Turkej,,and moved to London in 1972. He obtained his B.Sc. from Brunel Urziversity, and after a period in industry?, he returned to Bru-nel University to study for a Ph.D. on neM' DNA ussuys with Prqfessor P.G. Sammes. He is currently a post-doctoral ,fel-low at Brunel Miorking on disco-tic liquid crystals ttyith Dr L.R. Milgrom. 327 328 CHEMICAL SOCIETY REVIEWS 1994gL 'R (1, R = H) \ 111 VI 02.@ \ + O@0 \ \ (13) Scheme 1 nanthroline with lithiated 4-bromoanisole, followed by MnO, oxidation of the intermediate to give the diarylated phenanthro- line Removal of the methyl protecting groups, by heating the product with pyridinium hydrochloride at 200 "C gave the phenol (14) in high yield Placement of functional groups at other positions on the phenanthroline skeleton is best achieved by ring synthesis, most approaches using the Skraup reaction starting with an 8-aminoquinoline Thus a Conrad-Limpach reaction of ethyl acetoacetate with 8-aminoquinoline leads to the 4-hydroxyphe- nanthroline, existing mainly as its pyridone tautomer (15), some of our studies, the diester (1 7)was produced by heating 8-amino-2-methoxycarbonylquinolinewith dimethyl acetylenedi- carboxylate followed by thermolysis in diphenyl ether 0 MeO2C (17) 3 Intercalating Properties of 1,lo-Phe na nt hroIine Derivatives Intercalation of planar heteroaromatics between stacked duplex bases in DNA and RNA was first recognized by Lerman in 1961 The phenomenon of intercalation depends on the type of double stranded DNA (ds-DNA) under examination ds-DNA com- monly adopts several motifs, the three mdin forms being the right-handed A- and B-forms, which differ in the degree of 0-@\N C02H H02C (16) although 0-alkylation may be readily effected Chandler and colleagues have used this method to generate the potentiometric sensors such as (16),12 oxidation of the methyl groups was effected by use of selenium dioxide followed by sodium chlorite oxidation to the acid, using isobutene as a chlorine scavenger In hydration, and the left-handed Z-form (see Figure 1) Of the right-handed forms the B-motif appears to be most common under normal saline conditions Intercalation can occur with either A- or B-DNA but is not observed with the Z-form, as the base pairs are not aligned in a sufficiently ordered stacked manner to allow the insertion of intercalators without a distor- tion of the chain, intercalators often act on Z-DNA to isomerize it into the B-form Binding of agents into one of the grooves running down the ds-DNA helices can often compete with intercalation, many heteroaromatic compounds can form tight complexes, particu- larly in the minor grooves, without intercalating, whilst others can either intercalate or groove bind as two competing processes 3291.10-PHENANTHROLINES: A VERSATILE LIGAND-P.G. SAMMES AND G. YAHIOGLU Figure I (a) A-DNA; (b) B-DNA: (c) Z-DNA (left-handed).Notc thc relative order and vertical alignment of the base pairs in the right- handed A and B forms as compared to that in Z-DNA. Intercalation leads to unwinding of the DNA helix and tests have been developed to distinguish between intercalation and groove-binding. Substituents on the heteroaromatic framework can either enhance or hinder intercalation.Thus proflavine (1 8) is a strong intercalator whereas its tetramethyl derivative (19) is not, acting as only a groove binder. Positively charged species also bind more strongly by electrostatic interactions with the bridging phosphate groups. (18, R = H) (19, R = Me) A detailed study has been made on a series of substituted N- methyl-1,l O-phenanthrolinium salts (20), which showed that although methyl substituents did not totally prevent intercala- tion into the right-handed B-form of DNA, some modifications in the selectivity for the ten identified intercalating sites was observed (Figure 2). Steric interference between A-T pairs favours binding between G-C sequences. The tetramethyl derivative (20; R, = R, = R, = R, = Me) was the most strongly bound and gave the greatest increase in viscosity (unwinding) of the helix, attributed to intercalating.In contrast the diphenyl derivative (20; R, = R, = Ph) showed only a relatively weak binding, a result in contrast to that of the corresponding metal chelates (see below). T---A A---T f 1 T--A C-----G t 1 C----G t 1 T-----A C---G G--C Figure 2 Possible intercalation sites. Replacement of the N-methyl group by a coordinated metal ion can also produce intercalators and a range of metals have been utilized including ruthenium, rhodium, cobalt, and copper. The trisphenanthroline complexes of Ru", for example (21), are coordinatively saturated, stable complexes inert to substitution but, because of their positive charge, they are generally soluble in water.They possess a very rigid structure, existing as racemates which may be resolved into the right handed A-and left-handed A-isomers (see Figure 3), which allow enantioselective studies on DNA. The complexes are luminescent, showing a strong metal- to-ligand charge-transfer excited state, which is perturbed by changes in the local environment. Intercalation usually leads to an enhancement in luminescence and an increased lifetime (from CHEMICAL SOCIETY REVIEWS. 1994 Figure 3 Representation of the twisted forms of tns-phenanthroline metal complexes. (a) the right-handed A form; (b) the left-handed A form R7 R2L R1 R2 (21) R, = R2 = H [R~(phe)~] (22) R, = R2= Me (23) Rf = R2 = Ph [Ru(~IP)~] (24)As 23 but Co Ill complex (25) As 23 but Rh Ill complex 0.6ps to > Sps).At least two modes of non-covalent binding to DNA have been shown to existI5 -external binding (such as groove binding) and intercalation, modes that have been con- firmed by using nitroxide labelled probes of the type (26) employing ESR spectroscopic studies. Binding is enantio- selective; intercalation into B-DNA (right-handed duplex) is favoured by the A-isomer from the major groove. Models indicate that this binding is consistent with the complementary shapes of the helix and the complex. External binding occurs with the A-isomer and NMR studies indicate that this occurs in the minor groove. l4 The discrimination is high enough to enable the use of B-DNA as a resolving agent for the enantiomers! By using variants on (2 l), these 'shape-selective' binding studies have been extended to other forms of DNA.Thus the A-isomer of the methylated derivative (22) binds more tightly than the d-enantiomer to A-form nucleic acid duplexes; in this case binding is in the major groove with the left-handed form being against the right-handed DNA helix and does not involve intercalation. The diphenyl derivative (23), [RuDip,], behaves in a similar fashion to (21) towards B-DNA, showing an even greater discrimination ('shape selectivity') between isomers, the d-isomer showing some intercalation involving the pendant phe- nyl groups whereas steric buttressing of the phenyl groups of the A-enantiomer prevents any binding.In contrast to the Ru (phe), isomers, the Ru (Dip), isomers both bind equally to the left- handed Z-DNA but the A-isomer shows hypochromicity on binding, suggesting a test for the Z-conformation. * Model studies on poly-dGC duplexes were informative since this duplex can be made to exist either in the B-conformation or the Z-form by changing the buffer conditions. Whereas ethidium (27) acts mainly as an intercalator, preferably from the minor groove, and, upon interaction with the DNA, changes the conformation of the Z-to the B-form, none of the isomers of (21) and (23) cause this interconversion; for the B-form the d-enantiomers bind more strongly whereas for the Z-form both isomers bind.The difference in behaviour with these complexes compared to ethidium was explained by assuming shape-selec- tive binding into the major groove of these isomeric helices is the strongest interaction. A detailed study using the analogous A-isomer of the Co"' complex (24), which can act as a DNA nicking agent on binding (see below), was used as a study on the Z-binding regions of two 0' plasmids. An analysis of the regions that were specifically nicked 1,lO-PHENANTHROLINES A VERSATILE LIGAND-P G SAMMES AND G YAHIOGLU 33 1 4p-0-o'p-O" Qp-0-0 0 b-0' 10-b-Figure 4 OH insertion mechanism for DNA nicking process by this reagent led to the suggestion that the Z-regions served as a genetic punctuation mark, used to describe the ends of transcription regions of geneslZo Shape-selective interactions with these phenanthroline complexes have also been used to study the nature of the folds in t-RNAphe 21 The phenanthroline derivative (28) shows interesting photo- physical properties 22 In aqueous solution the ruthenium com- plex shows no emission upon irradiation, whereas in non-aqueous systems it shows intense luminescence properties This change is explained by the formation of a metal-to-ligand charge-transfer system in the excited state, the increased electron density in the dipyridophenazine system has the effect of turning the excited-state pyrazine nitrogens into strongly basic groups that, in water, abstract a proton from the solvent to form the protonated, ground-state species, relaxing to the starting mater- ial in the process In aprotic solvents this protonation pathway is precluded and the charge-transfer state is in equilibrium with the excited metal species that then collapses to the ground state with emission of a photon This enviromental sensitivity has been elegantly exploited in studying intercalation Under intercalation conditions the dipyridophenazine system is effectively in a local aprotic region, no protonation can occur and any excited states exhibit lumines- cence A careful examination of the decay of luminescence against time indicates a biexponential process, suggesting the presence of two species, these were assigned as two different intercalated forms, rather than one intercalated form and one groove-bound form, since experiments to try to quench the luminescence with anions such as ferricyanide ion failed although this is known to quench the luminescence of groove- bound forms of such reagents 23 The luminescent enhancement on binding to DNA is 2 lo4, by comparison the enhancement observed when ethidium intercalates to DNA is 2 20 Thus intercalation can literally be seen as a 'switching on' of the luminescence This phenomenon has been exploited in specific probes for DNA The ruthenium ligand is first attached to a probe piece of DNA of known sequence, using a fairly flexible link When this probe meets a complementary piece of target DNA under hybridizing conditions, a local segment of duplex DNA is formed The ligand can then fold back and intercalate with this local duplex DNA, an event marked by the appearance of luminescence A minor problem with this approach is the inability to distinguish between a hit with the desired duplex piece of DNA and any adventitious binding of the dipyrido- phenazine chelate with other duplex DNA material that may be present in the assay mixture 24 4 DNA Nicking Reagents involving Phenant hrol i nes The herbicidal activity of many phenanthroline derivatives has been explained by the incorporation of copper, the ligand helping to transport this into the plant cells where the copper exerts its toxic behaviour The generation of free-radical species, such as hydroxyl groups, in the vicinity of DNA chains can lead to nicking of the links and cleavage of the chain Metal complexes of reducible species, such as the EDTA complexes of iron and copper, are particularly active By building these chelates into DNA probes the technique of DNA footprinting has been developed, where- by the point of attachment of the probe can be determined Since 1 ,lo-phenanthrolines act as good ligands for such metal ions it is not surprising to find that several recent studies have used these complexes to help probe the structure of DNA Sequence-specific scission of DNA has been achieved using the 2 1 complex of 1,IO-phenanthroline with cuprous ions [see (29)]with hydrogen peroxide as co-oxidant Use was made of 5-substituted DNA derivatives (30) to prove that a similar degradation occurred with both the complementary DNA or RNA sequences, after hybridization, to cleave sites on the substrate at base positions up to f3 nucleotides from the point of attachment of the ligand Related nicking agents, such as bleomycin, are found to be less active on RNA than DNA A mechanism involving attack at the site of base attachment to the sugar was indicated (see Figure 4)26 A similar approach has been adopted by Helene and his team 27 As mentioned above, the cobalt and rhodium complexes (24) and (25) can also catalyse the formation of free-radical species Cleavage of RNA occurs by irradiating the ruthenium com- plexes in the presence of oxygen which forms local concent- rations of singlet oxygen leading to nicking of the RNA chains CHEMICAL SOCIETY REVIEWS, 1994 After examining several of the reagents described, which show a variety of different nicking patterns, it was found that the rhodium(II1) complexes (25) and (31) show very selective nicking patterns at sites adjacent to and at the triple helix region at cruciform sites.It was argued that intercalation is possible at these centres since, at these, the helical structure is more open than in normal duplex regions. The observed breaking of the sugar-base bond, rather than at sugar-phosphate bonds, is evidence for the intimate association of the reagent with the RNA.Zs -\ \gfh:q 2 5 Phenanthrolines and Europium Europium(I1r) ions and related lanthanide species have attracted much rapidly increasing attention over the last decade as robust luminescent species in a variety of diagnostic probes and assays.The main features of Eu3 + photochemistry are: (i) the ground-state ion has only a very weak absorption coefficient, since the main excitation transitions are formally forbidden. (ii) Triplet sensitizers may be used to help populate the excited state, but these sensitizers need to be in close proximity with the ion, as in metal chelates, to observe efficient energy transfer. (iii) The excited state involves excitation of an inner f-shell electron and the transitions are shielded by the residual valence-shell electrons. Observed emissions occur as a series of sharp emission bands. (iv) The energy emitted is generally at a much longer wave- length than the energy absorbed by the sensitizer, leading to a large Stoke’s shift.Hence concentration-dependent self-quenching of the luminescent state is not observed. (v) Since the emission process is also formally spin-forbidden it is a relatively slow process (lifetimes in the microsecond to millisecond range), allowing the use of time-resolved measurements. This feature allows one to remove interfer- ences arising from background fluorescence, autofluores- cence, and scattering phenomena. (vi) Solvated water molecules can quench the luminescence by a vibronic-coupled deactivation of the excited state. The degree of quenching is dependent on the number of water molecules in the solvent shell; the lanthanide ions can form solvates with up to nine molecules of water.In order to observe efficient luminescence the majority of these water molecules have to be removed by using ligands that shield the ions from water. Two main ways for utilizing europium ions in probes have been developed. In the DELFIA system, developed by Soini, Lovgren, and colleague^,^^ europium is chelated to species like ethylenediamine tetraacetic acid conjugates of biological sub- strates, in a straight replacement of a corresponding radioactive tag. These compounds are themselves non-luminescent. After separation of the labelled reagent-substrate complex (such as an antigen-antibody complex), the europium is removed by sequestration at low pH with a large excess of an aromatic p-diketone reagent, which can also act as a sensitizer.In order to increase the europium luminescence a mixture of further rea- gents, such as surfactants and trioctylphosphine oxide, is added (an ‘enhancer’ solution) in order to form hydrophobic micelles and thus eliminate water quenching. The DELFIA system can only be used in heterogeneous assays (those in which a sepa- ration of the excess of reagent from the substrate-reagent complex has to be carried out). Rather than use an inert chelating agent, Diamandis et al.30 utilized derivatives of phenanthroline-2,9-dicarboxylicacid (5). This is a powerful sensitizing ligand for europium ions and avoids the need for a separate sensitizing ligand such as the 8-diketones used in the DELFIA approach.The main reagent used in these assays (e.g. the CyberFluor assay) is the bathophe- nanthroline derivative (32). A detailed study of the binding of Eu3+ with the diacid (5) showed that at neutral and acid pH a 1 :1 complex forms (Kiss 2 x lo8 M-l) but that at higher ligand concentrations a 2:1 complex can also form (Kiss2 x lo6 M -I). RS02 r5-l %co2H\N WN The 2:l complex produces high luminescence and shows only one molecule of water in its coordination shell (mean lifetime 0.72 ms). Luminescence of the 1 :I complex can be enhanced by drying. At pH values > 7, the complex collapses with formation of europium hydroxides and the 2:1 comple~.~ Enzyme-linked assays have been reported which involve oxidation of phenanthroline-2,9-dicarbohydrazide(33), which is oxidized to the dicarboxylic acid (5) and then assayed with europium./cYoNHNH2 (33) The DELFIA and CyberFluor systems are not readily appli- cable to homogeneous assays. In an approach to overcome this we have been investigating the use of a cooperative signalling system, one that is only turned on when two components meet. We have shown that, under defined conditions, europium can form discrete, mixed 1:l:l chelates. One ligand is used as a shielding agent, to protect the ion from water molecules, the other as a sensitizer. Reagents such as the diazatrioxa-[15]- crown (34) can be used as the shielding agent. This has a high binding constant to Eu3 + ions (Kiss > 1012 M-l) but does not coordinately saturate the metal.A molecule of the diacid (5) can also approach the ion in the pH range 6.5 to 8.0 to form a 1: 1: 1 complex (Kissca. lo6). Because this is now highly shielded from water the system shows efficient luminescence (T 0.72 ms), the phenanthroline acting as the sensitizer. A variety of shielding ligands may be used in place of the crown (34), including EDTA and its derivatives. The cooperative effect is concentration- dependent and, at concentrations below lop6 M, the phenanth- roline dicarboxylic acid starts to dissociate from the 1:l:l complex and luminescence disappears. (34) The cooperative approach has been developed for use as a homogeneous assay for DNA. This is outlined in Figure 5. In this, use is made of the organization created by formation of a segment of duplex DNA when a probe DNA strand meets its 1.10-PHENANTHROLINES: A VERSATILE LIGAND-P.G. SAMMES AND G. YAHIOGLU Figure 5 A homogeneous DNA assay. Only when the target and the probe DNA strands meet and hybridize can intercalation occur and the sensitization of the Eu3 ions be observed. T, target; P,probe; S,+ sensitizer; Eu, europium ion. Br H02 C (35) 5’ 0.A.A.G.A.T.G.A.T.A.T.T.T.T.C.T.T.T.A.A.T.G.G.TG I P-O(CH2)6NHCO.CH2rCH2CH2N(CH2CO2H)20” ‘0- CH2C02H (36) 3’ intensity t 550 600 650 700nm-Wavelength Figure 6 Curve a, reagents (35) and (36) with target (37); curve b; without target (37). Concentrations: (35) and (36), 2 x lops M; (37), 5 x 10-9 M.5’ C.C.T.T.T.G.T.G.GT.T.T.C.T.A.C.T.A.T.A.A.A.A.G.A.A.A.T.T.A.C.C.A.C. G.G.T.C.C.G.T.A.T.T.A (37) complementary sequence at a target and when the pair hybri- dize. In our assay system the probe DNA is linked at one end to a short handle bearing a molecule of EDTA to which is chelated europium. The EDTA complex of Eu3+ is very tightly bound (Kiss> 10’ M -l). At this stage the labelled duplex DNA shows no luminescence. However, duplex DNA can accommodate either groove-binding agents or intercalators, whereas the single-stranded target DNA alone cannot. This property is utilized to help increase the local concentration of the sensitizer molecule. We use the phenanthroline dicarboxylic acid deriva- tive (39, in which the linked phenanthridinium group can act as an intercalator.Since intercalation has a binding constant of cu. lo5,the effective binding constant of the sensitizer in the region of duplex is increased to ca. loll, i.e. an enhancement of luminescence is observed. Figure 6 shows the output from a typical test with the probe (36) against the target DNA strand (37) as against the background signal observed (due to adventi- tious approach of the europium-labelled probe DNA to sensit- izer molecules at the lo-* M concentrations used) when using a non-matching strand of DNA. The advantage of this approach is that it allows a direct test for a specific DNA sequence under homogeneous conditions. In situ assays of DNA from biologi- cal specimens are currently being developed.6 Su pramolecu lar Reagents ut iIizi ng Phenanthrolines A large number of studies on supram~lecularity~ involving 1,lO-phenanthroline derivatives have been made over the last decade and space limitations allow for the mention ofjust a few of these. Chandler and colleagues have studied a range of aza-crown derivatives of the diacid (9,such as the cyclic lactones (38)12 whilst Sauvage et ul. have studied related macrocyclic systems by making extensive use of the diphenol(l4). The rigid structure of this molecule, in conjunction with the large separation of the phenolic groups from the chelating nitrogens, have made it particularly useful in macrocyclic and topological studies. Inclu- sion of other metal chelating heterocyclic systems into the ring, such as in compound (38), provides a means for using metal chelation to control the formation of various new topological systems such as the catenanes.Recent successes in this area include the description of the molecular knot (39).32 In an alternative approach to that used by Barton’s group, Bannwarth has used 4,7-diphenylphenanthrolinecomplexes of ruthenium(I1) as general labels for nucleic acid strands to which they are attached by a covalent linker.33 The ruthenium lumi- nescence may be stimulated by a through-pace, fluorescence CHEMICAL SOCIETY REVIEWS, 1994 0 1 energy transfer (FRET) from an excited energy donor, such as the lumazine (40) Since the efficiency of energy transfer is related to its distance from the acceptor the system can be used as a molecular ruler Phenanthroline derivatives have also been used as enzyme mimics The redox properties of the tris( 1,lO-phenanthroline- 5,6-dione) ruthenium(I1) complex (41) and related compounds have been used as efficient mediators for the NAD+-promoted oxidation of alcohols The mediators oxidize NADH to NAD+ and are themselves reoxidized by either aerobic or anodic oxidation 34 Phenanthroline analogues of flavins have also been reported 35Chiral phenanthroline complexes, such as the pyrro- lidinemethanol derivative (42), in conjunction with metal ions such as zinc or cobalt, act as catalysts exhibiting enantioselecti- vity towards various peptide ester substrates 36 The above examples attest to the wide range of chemical applications made of the 1, 10-phenanthroline system other than its historical use as a chelating indicator We would expect further exciting revelations in the nedr future 7 References 1 P G Sammes, G Yahioglu, and G D Yearwood, J Chem Soc.Chem Commun , 1992, 1282 2 L A Summers, Adv Heierockcl Chem , 1978,22, 1 3 Cf J P Michael and G Pattenden, Angeit Chem Int Ed Engl, 1993,32. 1 4 J -M Lehn, Angeu Chenz Int Ed Engl , 1988, 27, 89, 1990, 29, I304 5 G R Newkome, G E Kiefer, W E Pickett, dnd T Vreeldnd, J Org Cheni , 1983,48, 5 1 12 6 A R Katritzky. Q -h Long, N Malhotra, T A Ramanarayanan. and H Vedage, Sjnthem, 1992, 91 1 7 R Breslow, R Fairweather, and J Keana, J Am Chem Soc , 1967, 89,2 135 8 S Krishnan, D J Kuhn, and G A Hamilton, J Am Chem Soc , 1977,99,8 12 1 9 T D Lash, B H Novak, and Y Lin, Tetrahedron Lett, 1994, 35, 2493, N Ono, H Hironaga, K Simizu, K Ono.K Kuwano, and T Ogawa, J Client Soc Client Conintun , 1994, 1019 10 M N Ackermann and L V Iterrante, Inorg Chem ,1984,23,3904 11 C 0 Dietrich-Buchecker and J -P Sauvage, Tetruhedrort Lett , 1983, 24, 5091, J -C Chambron and J -P Sauvage, Tetrahedron Lett, 1986,27,865, C 0 Dietrich-Buchecker, J -P Sduvdge, dnd J Weiss, Tetrahedron Lett , 1986, 27, 2257 12 C J Chandler, L W Deady, J A Reiss, and V Tzimos. J Heteiocjcl Cheni ,1982,19, 1017, C J Chandler, L W Deady, and J A Reiss, J Heterocj cl Chem , 1986, 23, 1327 13 L P G Walshe and M J Waring, 'DNA Intercalating Agents' in 'Comprehensive Medicinal Chemistry', ed C Hansch, P G Sammes, and J B Taylor, Pergamon Press, 1990, Vol 2, pp 703- 724, H W Zimmerman, Angel$ Chent Int Ed Engl , 1986,25, I I5 14 E J Gdbbdy, F Destefdno, dnd K Sdnford, Biocheni BiophtJ Re5 Commun , 1972, 46, 155, E J Gabbay, R E Scofield, and C S Baxter, J Am Chem Soc , 1973,957850 15 J P Rehmann and J K Barton, Biochemisrr), 1990,29, 1701, J K Barton, Acc Chem Res , 1990,23,271 16 M F Ottaviani, N D Ghatlia, S H Bossmann, J K Barton, H Durr, and N J Turro, J Am Chem Sue, 1992, 114,8946 17 H Y Mei and J K Barton, J Am Chem Soc, 1986, 108, 7414, idem,Proc Nail Acad Sci USA, 1988,85, 1339 18 J K Barton, L A Basile, A Danishefsky, and A Alexandrescu, Proc Natl Acad Sci USA, 1984,81. 1961 19 A E Friedman,C V Kumar,N J Turro,andJ K Barton, Nuclei( Acidc.Res, 1991, 19, 2595 20 J K Barton dnd A L Rdphdel, Proc Nut1 Acud Sci USA, 1985, 82,6460 21 M R Kirshenbaum, R Tribolet, dnd J K Barton, NucIeic Acid7 Res ,1988,16,7943, C S Chow and J K Barton, J Am Chem Soc , 1990,112,2839 22 Y Jenkins, A E Friedman, N J Turro, and J K Barton, Biochemistrh, 1992,31, 10809, E Amouyal, A Homsi, J -C Cham- bron, and J -P Sauvage, J Chem Soc Dalton Trans , 1990, 184 1 23 R M Hartshorn and J K Barton, J Am Chem Soc , 1992.114, 5919 24 Y Jenkins and J K Barton, J Am Chem Soc , 1992, 114, 8736 25 C -h B Chen dnd D S Seligmdn, Ace Cheni ReJ , 1986,19, 180, R Tamilarasan and D R McMillin, Inorg Chent , 1990,29,2798 26 C -h B Chen dnd D S Seligman, J Am Chem Soc, 1988, 110, 6570 27 J -C Francois, T Saison-Behmoaras, M Chassignol, N T Thuong, and C Helene, C R Acud Sci PutiJ, 1988, 307 111, 849 28 M R Kirshenbaum, A Tribolet, and J K Barton, Nucleic Acids Res . 1988,16,7943.C S Chowand J K Barton. J Am Client Soc . 1990, 112,2839 29 E Soini and T Lovgren, Crit Rev Anal Chem , 1987, 18, 105, I Hemmild, S Ddkubd, V -M Mukkdld, H Siitdri, dnd T Lovgren, Anal Biochem, 1984, 137, 335 30 E P Diamandis, Clin Biochem ,1988.21. 139. R A Evangelista. A Pollock, B Allore, E F Templeton, R C Morton and E P Diamandis, Clin Biochem, 1988,21, 173 31 E F G Templeton dnd A Polldk, J Luniiri , 1989, 43, 195 32 D M Walba, Q Y Zheng, and K Schilling, J Ant Chem Soc , 1992, 114,6259 33 W Bannwarth, W Pfleiderer, and F Muller, Hell Chim Acta, I99 I, 74, 1991 34 G Hilt and E Steckhan, J Clieni Soc Cheni Copliniun ,1993, 1706 35 K J Black, H Huang, S High, L Starks, M Olson, and M E McGuire, Inorg Chem , 1993,32, 5591 36 J G J Weijnen, A Koudijs, and J F J Engbersen, J Org Chem , 1992,57, 7258

ISSN:0306-0012    

DOI:10.1039/CS9942300327

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

The Insertion of Alkynes into Metal-Metal Bonds and Organic Chemistry of the Dimetallated Olefin Complexes Richard D. Adams Department of Chemistry and Biochemistry University of South Carolina Columbia SC 29208 U.S.A. 1 Introduction The activation of alkynes through complexation to metal atoms has played an important role in the development of the organic chemistry of these molecules. 1~2Recently there has been much interest in the coordination and reactivity of alkynes in polynuc- lear metal complexe~.~ Alkynes have been found to coordinate to the metal atoms in binuclear metal complexes in several different ways. By far the most common mode is the p-1 or di-x in which the ligand donates electrons from both of its x-bonds to the two metal atoms A.4 The direction of the C-C bond is perpendicular to the M-M bond.In these complexes the alkyne ligand serves as a four-electron donor. There are a few examples where the alkyne is coordinated in a skewed ‘asymmetric’ bridging mode B.,”. There are a number of examples where the alkyne is coordi- nated to the metal atoms through a-like bonds with one carbon bonded to each metal atom. In these cases the alkyne serves as a two-electron donor. Two basic geometries are possible. In one both metal atoms lie cis-or Z-positioned relative to the C=C R I 4 -Electron x -Donors M-M-M Dimet alIated Olef ins Di-oDonors 2 -Electron M-1 R ‘c=c/ M\ R M’ ‘M R R \c=c/ / R /\c=c M ‘R M p-I I cis or Z trans or E (C) (D) (E) Richard D. Adams is Professor of Chemistry at the University of South Carolina.He received a B.S. degree from the Pennsylvania State University in 1969 anda Ph.D. degree in chemistry in 1973 from the Massachusetts Institute of Technology. He was an Assistant Professor of Chemistry at the State University of New York at Buflalo 1973-75; Assistant Professor of Chemistry at Yale University 1975-80 and Associate Professor of Chemistry at Yale University 1980-84. He was a Fellow of the A. P. SIoan Foundation 197941. Since 1992 he has been the managing editor of the Journal of Cluster Science. His research interests lie in the area of metal cluster complexes particularly structure analy-sis polynuclear ligand coordi- nations and cluster catalysis.He is the author of over 300 scientijic publications. double bond C and D.The metal atoms may be bonded or not. This coordination mode is also referred to as p-11. In the other case the metal atoms lie trans-or E-to the C=C double bond E. The two structural types D and E in which the metal atoms are not mutually bonded are often referred to as dimetallated olefins. The first example of a dimetallated olefin complex [(NC>,CO(~-HC=CH)[CO(CN),]~was made by Wilkinson -from the reaction of HC-CH with K,[CO,(CN),,].~ A structure having Z-stereochemistry E was proposed7 and established many years later through an X-ray crystallographic analysis of the dicarboxylate-substituted derivative.* Since [Co,(CN) 0]6 -is well-known to split into monomeric [Co(CN),I3 -fragments in solution it seems most likely that the E-stereochemistry of these products is a consequence of the approach of the two [Co(CN),I3 -groups from opposite sides of the alkyne.Recently Beck and co-workers have prepared the E-dimetal- lated olefin complex [CpRu(PMe,),(p-Z-MeO,CCO,Me) [Ru(CO),Cp] (2) by the addition of the organometallic anion [CpRu(CO),]-to the alkyne ligand in the cationic complex [CpRu(PMe,),(MeO,CC=CCO,Me)] + .9 F H ‘C0,Me Cb N (1) (2) Some 2-dimetallated olefins have been prepared by the addition alkynes to dinuclear metal complexes by a formal ‘insertion’ of the alkyne into a metal-metal bond. Pridr to our studies all examples of these reactions have involved dinuclear metal complexes that are bridged by two chelating phosphine ligands e.g.equation I. O -PPh -PPh R-CEC-R ,pTTPh2r I 1CI-Pd-Pd-CI ____t Pd\ /II Pd (1) .p=CjR Ph,P-PPh h 2 P v pph2 The dimetallated olefin complexes formally contain a double bond between the two carbon atoms. Accordingly it is expected that the DeE or ZeE isomerization should have a high energy barrier. Theoretical treatments have indicated that this isomerization barrier should exceed 60 kcal mol- .6h In this review we will summarize the results of our recent studies on synthesis and reactivity of dimetallated olefin com- plexes of the manganese subgroup. The parent carbonyl com- plexes Mn,(CO),o and Re,(CO),o serve as the starting points for these studies.Ditechnetium complexes have not yet been investigated. The decacarbonyl complexes are relatively unreac- tive to alkynes under mild conditions. Mays has reported that Re,(CO) will react with internal alkyl and aryl alkynes at 190°C by the addition and coupling of the alkynes to form alkyne oligomers that bridge the two rhenium atoms.’ 335 336 2 The Insertion of Alkynes into Re-Re and Mn-Mn Bonds To enhance the reactivity of the decarbonyl compounds they are routinely converted into their acetonitrile derivatives Re,(CO),(NCMe) (3) and Mn,(CO),(NCMe) (4)by the decar- bonylation with Me,NO in the presence MeCN.12.13 The MeCN ligand in these complexes is readily displaced and bonding interactions between the alkyne and the metal atoms can be formed under mild conditions. All of the alkynes that we have studied contain at least one electron-withdrawing carboxy- late group or an OEt group.The first reaction of an alkyne with (3) that we studied involved the terminal alkyne HC=CCO,Me at 68°C. The product obtained in 8 1 %yield was shown to be the E-dimetal- lated olefin complex Re(CO),[cL-E-HC=C(COzMe)]Re(CO) (5) (see equation 2) and the structure was established by a single crystal X-ray diffraction analysis. \/Me -Re-C H /\ Interestingly the carbonyl oxygen atom of the carboxylate group was coordinated to one of the metal atoms to form a five- membered ring. The complex (5) was formed formally by the insertion of the alkyne into the Re-Re bond in (3).The question of mechanism was addressed immediately. In particular did the insertion occur by a non-dissociative 'intramolecular' process or did the complex somehow dissociatively split into mononuclear fragments which subsequently recombined? The latter process would easily explain the formation of the observed E-stereo- chemistry. Radical scavengers light and polar solvents had no detectable effect on the reaction which suggested that the process was instead intramolecular; but the most convincing evidence was that provided by an intermolecular crossover test. In this experiment a mixture of unlabelled Re,(CO),(NCMe) and fully 3C-labelled (990/,) Re,( ,CO),(NCMe) was allowed to react with the HC=CCO,Me to form the product (5). If the reaction was occurring via the intermediacy of fully dissociated mono- nuclear fragments then a statistical distribution of the four products Re(CO),[p-E-HC=C(CO,Me)]Re(CO) (5a) Re( 3CO),[cL-E-HC=C(C0 Me)]Re(CO) ,(5b) Re(CO),[p-E- HC=C(CO,Me)]Re( ,CO) (5c) and Re( WO),b-E- HC=C(C02Me)]Re(1 ,CO) (5d) should be formed.The non- dissociative mechanism should lead to the products (5a) and (5d) only. A mass spectrum of the products from this reaction showed the formation of (5a) and (5d) only. Accordingly it is believed that the insertion of the alkyne into the metal-metal bond occurs without a splitting of the molecule into mononuc- lear fragments. Further evidence was obtained from the reaction of (3) with EtO,CC=CCO,Et see below. It remained to explain how a product with E-stereochemistry was formed under such mild conditions if the Z* E transfor-mation has a high energy barrier as predicted theoretically.6h The mechanism shown in Scheme 1 was proposed. It is expected that the reaction proceeds initially by the displacement of the MeCN ligand by the alkyne to produce an intermediate F in which the alkyne is r-bonded to one of the rhenium atoms in the usual fashion.An interaction subsequently develops between the second rhenium atom and the carboxylate-substituted end to the coordinated alkyne G.In the process the Re-Re bond begins to develop a heteropolar character. In the limit H the metal-metal bond has a full heteropolar character. The next step is probably the slowest step and involves a spontaneous cleavage of the metal-metal bond with both electrons in the bond shifting to the Re(CO) grouping.This would reduce the number of CHEMICAL SOCIETY REVIEWS 1994 Scheme 1 valence electrons at the Re(CO) group to sixteen. This un-favourable condition could be alleviated by a conversion of the HC=C(CO,Me)Re(CO) grouping into an q2-coordination to the Re(CO) group through a 90" twist of the C(C0,Me) Re(CO) grouping to form the intermediate I. This 7,-HC=C(CO,Me)Re(CO) grouping would serve as a three-electron donor to the Re(CO) group. Recently a number of complexes containing related $-alkenyl ligands have been isolated and structurally characterized. l4 The intermediate I may be a transition state that is subsequently converted into the product (5) by a further 90" rotation of the C(CO,Me)Re(CO) grouping and a coordination of the carbonyl oxygen atom to the Re(CO) group.Evidence to support this mechanism was obtained from the reaction of (3) with the internal alkyne Et02CC=CC02Et. The product Re(CO)&-Z-(EtO,C)C=C(CO,Et)]Re(CO) (6) was structurally characterized and found to have the a Z-stereochemistry of the two metal atoms Figure 1. This complex- could be formed directly from a species analogous to H by cleavage of the heteropolar metal-metal bond and a stabiliza- tion of the intermediate by the coordination of the oxygen atom of the proximate carbonyl group which results in the formation of the four-membered metallacyclic ring. Compound (6) can be converted into an E-species Re(CO),~,-€-(EtO,C)=C(CO,Et)] Re(CO) (7) analogous to (5) but more forcing conditions are required.This is reasonable because the complex (6) is stabilized Figure 1 An ORTEP diagram of the molecular structure of the dimetal- lated olefin complex Re(CO),b-Z-( EtO,C)C=C(CO Et)]Re(CO) (2). Selected intramolecular bond distances (A) are as follows Re( l)-C(3) =2.193(8) Re(2)-C(2) =2.228(8) C( 1)-C(2) =1.49( 1). C(2)-C(3) =1.34(1) C(3)-C(4) =1.46(1); Re( 1)-0(3) =2.214(6). (Reproduced with permission from Urganonterullics 1994. 13 1264.) REACTIONS OF ALKYNES WITH DIMANGANESE AND DIRHENIUM COMPLEXES-R D ADAMS relative to the species traversed en route to (5) Compound (7) can be decarbonylated when heated to yield the E-complex Re(CO),[p-E-( MeO,C)C=C(CO Me)]Re(CO) (8) that con-tains two five-membered rings formed by the coordination of the carbonyl oxygen atoms from both carboxyldte groups equation 3 \-Interestingly complex (6) adds CO reversibly to yield the Z-dimetallated olefin complex Re(C0) ,[p-Z-(MeO,C)C=C(CO Me)]Re(CO) (9) having two Re(CO) groups equation 4 l3 OEt EtOPC ,CO,Et The irradiation of Mn,(CO) in the presence of EtO,CC= CC0,Et yields the compound Mn(CO),[p-Z-(EtO,C)C= C(CO,Et)]Mn(CO) (lo) which is structurally analogous to (6) in 36% yield in a matter of minutes A small amount of the two- ring compo und M n( CO) [p-E-(M eO C)C =C( CO ,Me)] Mn(CO) (1 I) analogous to (8) is also formed Interestingly compound (10) slowly isomerizes to the compound Mn,(CO),[p-(EtO,C>C=C(CO,Et)C=O] (12) by an insertion of a CO ligand into the metal-carbon bond of the Mn(CO) group The oxygen atoms of the inserted CO grouping and one of the carboxylate groupings are coordinated to the manganese atoms to form five-membered rings Figure 2 The reaction of Mn,(CO),(NCMe) with EtO,CC=CCO,Et proceeds slowly (hours) to yield (12) directly presumably via the intermediacy of (10) CO coupling to the alkynes is an important difference between the chemistry of the rhenium and manganese dimetal- lated olefins It is the norm in the manganese chemistry and quite rare in the rhenium chemistry Figure 2 An ORTEP diagram of the molecular structure of the complex Mn2(CO),[p-(Et02C)C=C(C02Et)C=O](12) (Reproduced with permission from J Am Chem Soc 1994 116 4467 ) The reaction of the terminal alkyne HC-CC0,Me with Mn,(CO),(NCMe) yields the compound Mn,(CO),{p-O=C [C(H)=C(CO,Me)],) (13) as the major product formed by the addition of two alkynes and their coupling to d single CO ligand This same product can be obtained by the UV irradia-tion of Mn,(CO) in the presence of HC=CCO,Me in a slightly lower yield l6 The reaction of Mn,(CO),(NCMe) with HCKCOEt yields the analogous product Mn,(CO),{p-O=C [C(H)=C(OEt)],J In these compounds the oxygen atom of the alkyne-coupled carbonyl group is coordinated to two metal atoms The metal groupings can be sequentially cleaved from the organic ligand by reaction with a mixture of HCI and CO Scheme 2 ' HH UH (13) R = C02Me or OEt Scheme 2 The coupling of a CO ligand to one' * and two' molecules of alkyne has been observed on many previous occasions. but the accompanying coordination of the oxygen atom of the cdrbonyl group as found in the compounds (1 2) and (1 3) appears to be unique to the manganese complexes 3 The Organic Chemistry of Dimetallated Olefin Complexes Most of our investigations of the organic chemistry of dimetal- lated olefins have been derived from the compound (5) In its original form compound (5) is fairly unreactive to organic reagents however it can be activated by conversion into its MeCN derivative Re(CO),[p-E-HC=C(CO,Me)JRe(CO) (NCMe) (14) by treatment with Me,NO in MeCN In this form it is quite reactive and number of interesting reactions hdve been studied The reaction of (14) with HC-CC0,Me results in dn oligo- merization of the alkyne through the addition and head-to-tail coupling of two equivalents of the alkyne to the existing alkyne ligand in (14) 2o The product (OC),Re[C(H)=C(CO,Me) C(H)=C(CO,Me>C(H)=C(CO,Me)]Re(CO) (1 5) exists in solution as a mixture of isomers (1 5a) and (1 5b) that differ by their stereochemistry at the C=C double bond see Scheme 3 Isomer ( 15a) was characterized crystallographically The coupling steps probably occur by displacement of the MeCN ligand in (14) followed by a series of two alkyne addition and insertion sequences at the adjacent metal-carbon bond When the reaction is performed under an atmosphere of CO the CO- stabilized complex (CO),Re[p-C( H)=C(CO,Me)C(H)=C(CO Me)]Re(CO) (16) containing only two coupled alkyne group- ings was isolated When heated to 98 "C the mixture of (1 5a) and (15b) was converted into the new compound Re(CO),[C,H (CO,Me)(CO,Me),],(17) by the elimination of one of the rhenium-containing fragments MeO.H FoZMe 25OC-c-Me0 Scheme 3 The chain of alkynes was cyclized and a 2,4,6-tricarboxylate- substituted phenyl ring was formed in which one of the carboxy- late groups was coordinated to the remaining metal atom The cyclization process must have involved the cleavage of the C-H bond on the five-membered ring and elimination of the hydro- gen atom with a Re(CO) grouping Compound (14) also reacts with the heterocumulenes ArN=C=S21 and CS2, by displacement of the MeCN ligand and insertion of the heterocumulene into the metal-carbon bond The compounds [Re(CO),(E-HC=C(CO,Me)C=N (C,H,-p-Me)S)Re(CO) (18 Ar = Ph p-tolyl and p-C,H,Cl) and [Re(CO),{E-HC=C(C0,Me)CS2)Re(CO)4] (1 9) respect-ively were formed in good yields Compound (1 8) undergoes a remarkable photo-induced cyclization to yield products con- taining quinoline-2-thiolate ligands The major product is the mononuclear metal complex Re(C0),(2-S,3-C02Me,6-R,NC H4)(20 R = H Me Cl) in which the quinolinethiolate ligand is chelated to the metal atom pMe Ar OMe R Me0,C-:c=N (20,R = H Me or CI)s< i The cyclization to form (20) involves an activation of a C-H bond on the aryl ring at a position ortlzo to the nitrogen atom a coupling of the ring carbon atom to the hydrogen-substituted carbon atom of the olefinic group and elimination of a 'HRe(CO),' grouping Compound (I 9) undergoes an unusual reaction with pyridine oxide or ethylene sulfide in which the oxygen or sulfur atom is transferred and inserted into the remaining metal-carbon bond to yield the compounds (OC),Re[EC(H)C(CO,Me)C(S)S] Re(CO) (21 E = 0) and (22 E = S) Compound (22) was Characterized crystallographically 22 The result of the sequence of two reactions CS plus pyridine oxide or ethylene sulfide with (14) is that the alkyne has been derivatized at both ends Curiously the compounds (18) do not engage in such a reaction with pyridine oxide or ethylene sulfide H 'OMe Ethyldiazoacetate reacts with compound (1 9) by transfer of a carbene grouping to the olefinic site Two compounds the metallated cyclopropane complex (OC),Re[C,H,(CO,Me) (CO Et)C(S)S]Re(CO) (23) and Re,(CO),[SC(S)C(CHCH CO,Et)C(OMe)O] (24) were formed The cyclopropane ring is opened when the compound (23) is heated and compound (24) is formed equation 5 EtO C;c=c;H Et0,CAC *xH H ,OMe CHEMICAL SOCIETY REVIEWS 1994 U 4 Figure 3 An ORTEP diagram of the molecular structure of the metal lated pyrdn complex Mn2(CO),~-~4-OCC(COzEt)C(C02Et)C(H)C (COzMe)1(25a)(Reproduced with permission from J An? Chem Soc 1994 116 4467 ) Recently we have found that the CO-coupled alkyne ligand in the complex (12) engages in a novel coupling reaction with addi- tional alkyne in the presence of UV/VIS irradiation The prod- ucts Mn2(CO),[p-y4-OCC(CO2Et)C(CO Et)C(H)C(CO Me)] (25a) and Mn,(CO),[p-~4-OCC(C0,Et)C(C0,Et)C(C0,Et)C (CO,Et)] (25b) and Mn2(CO),[p-~4-OCC(C02Et)C(C0,Et) CHCH] (25c) were formed by that addition of the alkynes HC=CCO,Me EtO,CC=CCO,Et and HC=CH to (12) 23 Compound (25a) was characterized crystallographically and a drawing of its structure is shown in Figure 3 This compound contains a six-membered pyran ring that is metallated at the carbon C(9) by the metal atom Mn(2) Carbon C(9) is formally a carbene centre The second metal atom Mn(1) is n-bonded to four of the carbon atoms of the pyran ring One carboxylate group is coordinated to the metal Mn(2) by its carbonyl oxygen atom The formation of the pyran ring is equivalent to a hetero Diels-Alder reaction between the enone grouping of (12) with the incoming alkyne molecule see Scheme 4 24u The role of the H dI OEtI CO,R:\l,CO,Me I *H (12) + alkyne (254 Scheme 4 irradiation has not been established but may be required to induce the loss of CO from (12) to clear a pathway for the addition of the alkyne to the enone grouping It is also possible that the alkyne may add to the decarbonylated metal atom prior to its coupling to the enone grouping Treatment of the com-pounds (25) with a mixture of CO and HCI gases results in removal of the metal atoms and formation of the free pyran molecules which possess the a-structure as established for the pyran (26c) obtained from (25c) equation 6 Et02C& + Mn(CO)&I (6)(25~)-COlHCl H H (2W REACTIONS OF ALKYNES WITH DIMANGANESE AND DIRHENIUM COMPLEXES-R D ADAMS The coupling of alkynes to CO generally leads to the forma- tion of cyclopentadieneone rings l The formation of pyrans in this manganese system appears to be a novel result and may be related to the fact that the oxygen atom of the enone grouping is coordinated to one of the metal atoms which enforces a cis-geometry at the C(2)-C(9) bond (Figure 3) that is required for the hetero Diels-Alder coupling Pyran rings are important functional groupings that are found in a wide variety of natural products 24 4 Conclusions These studies demonstrate the ability of alkynes to insert into metal-metal bonds in dinuclear metal complexes These inser- tions may have similarities to the insertions of alkynes into metal-hydrogen and metal-carbon bonds In reactions des- cribed in this review all of the alkynes have contained carboxy- late or ethoxy groupings It is clear that these substituents stabilize the products Coordination of the carboxylate group- ing is one important way Compounds (3)and (4) do react with dlkyl- and aryl-substituted alkynes but complex mixtures of products are obtained probably due to the lack of stabilization provided by these substituents Analysis of these reactions has not proved feasible to date Nevertheless the insertion of dlkynes into metal-metal bonds appears to be an important reaction pathway for dinuclear metal complexes of the Group VII Dimetallated olefin products are formed in general in the first step and these complexes can engage in reactions with a variety of reagents to functionalize the alkyne These products can lead to the formation of new organic compounds by the subsequent removal of the metal atoms Ackno~ledgments I must acknowledge the valuable contribu- tions of Dr Linfeng Chen who performed most of the research described in this review for his Ph D thesis at the University of South Carolina Financial support for this research was provided by the Office of Basic Energy Science of U S Depart-ment of Energy References ((1) N E Shore Chem Rev 1988,88,108 1 (b)H M Colquhoun D J Thompson and M V Twigg ‘Carbonylation Direct Synthesis of Cdrbonyl Compounds’ Plenum Press New York 1991 (L) G W Pdrshdll dnd S D Ittel ‘Homogeneous Catalysis’ Wiley-Inter- science NY 1992 Chapter 8 (a)R S Dickson Polyhedron 199 1.10 1995 (b)R D Adams J C Ddrdn and Y Jeannin J Cluster Sci 1992 3 1 (a)G Palyi G Vardi and L Marko ‘Stereochemistry of Organo- metallic and Inorganic Compounds’ ed I Bernal Elsevier Amster- dam 1986 Vol I p 358 (6) P R Raithby and M J Rosales Adv Inorg Radiochem 1985,29 169 (c)E Sappa A Tinpicchio and P Braunstein Chem Rev 1983,83 203 (6)E Sappa J Cluster Sci 1994,5211 4 D M Hoffman R Hoffmann,andC R Fisel J Am Chem Soc 1982 104 3858 and references therein 5 (a)F A Cotton and M Shang Inorg Chem 1990,29,508 (b)M J Calhorda and R Hoffmann Organometallm 1986,5,2 187 (c)F A Cotton and X Feng Znorg Chem 1990,29 3187 6 (a)D M Hoffman and R Hoffmann J Chem SOL Dalton Trans 1982 1471 (6) D M Hoffman and R Hoffmann Organometallics 1982,1 1299 7 (a)W P GriffithandG Wilkinson J Chem Soc 1959,1629 (b)M E Kimball J P Martella and W C Kaskd Znorg Chern 1967,6 414 8 K D Grande A J Kunin L S Stuhl and B M Foxman Inorg Chem 1983,22 1791 9 J Breimair M Steimann B Wagner and W Beck Chem Ber 1990 123,7 10 (a) J T Mague Poljhedron 1992 11 677 (h) M Cowie G Vasapollo B R Sutherland and J P Ennett Znorg Chern 1986 25 2648 (c)C -L Lee C T Hunt and A L Balch Inorg Chem 1981,20,2498 (6)B L ShawandS J Higgins,J Chem Soc Dalton Trans 1988,457 11 M J Mays D W Prest and P R Raithby J Chem Soc Dalton Trans 1981 771 12 R D Adams L Chen and W Wu Organornetallics 1993,12,1257 13 R D Adams and L Chen Organometallics 1994 13 1264 14 J L Templeton Adv Organomet Chem 1989,29 1 15 R D Adams L Chen and W Wu Organometallics,1993,12,4112 16 V V Derunov,O S Shilova,A S Batsanov,A I Yannovskii,Yu T Struchkov and N E Kolobova Metalloorg Khzm 199 1,4,1166 17 R D Adams L Chen and W Wu Organometallicy 1993,12,343 1 18 (a)G Hogarth F Kayser S A R Knox D A V Morton A G Orpen and M L Turner J Chem Soc Chem Commun ,1988,358 (h)B P Gracey S A R Knox K A Macpherson A G Orpen and S R Stobart J Chem Soc Dalton Trans 1985 1935 (c) J Takats J Cluster Sci 1992 3 479 19 W P Fehlhammer and H Stolzenberg in ‘Comprehensive Organo- metallic Chemistry’ ed G Wilkinson F G A Stone dnd E Abel Pergamon Oxford 1982 Chapter 3 1 4 p 548 20 R D Adams L Chen and W Wu OrganometallrcJ 1993,12,1623 21 R D Adams L Chen and W Wu Organometallrcs 1993,12,3812 22 (a)R D Adams L Chen and W Wu Organometallics 1994 13 1257 (b)R D Adams L Chen and W Wu Angel.c Chem Int Ed Engl 1993,33 568 23 R D Adams and L Chen J Am Chem Soc 1994 116,4467 24 (a)D L Boger and S M Weinreb ‘Hetero Diels-Alder Methodo- logy in Organic Synthesis’ Academic Press New York 1987 Chapter 7 (h)T Kametani and S Hibino in ‘Advances in Hetero- cyclic Chemistry ed A R Katritzky Vol 42 Academic Press Orlando 1987 pp 245-333 25 (a)P M Maitlis Acc Chem Res 1976,9 93 (6)A A H van der Zeljden H W Bosche and H Berke Organometallics 1992 11 563 (c)C Bianchini A Meli M Peruzzini and F Vizza Organo-metallics 1990 9 1146

ISSN:0306-0012    

DOI:10.1039/CS9942300335

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Electrochemical Solid State Analysis This Review isdedicated to Prof. Dr. Rolf Neeb on the State of the Art occasion of his 65th birthday Fritz Scholz and Birgit Meyer lnstitut fur Chemie Humboldt- Universitat Hessische StraBe 1-2 I01 15 Berlin Federal Republic of Germany 1 Introduction research field of pure electrochemistry of solids. It seems that the Solid-state analysis relies largely on various spectroscopic and latest developments in the field of voltammetry of solid sub- stances have opened the way into a new world of electro- diffraction methods. There are no fundamental reasons for the lack of electrochemical methods for solid-state analysis -but there are practical reasons for this situation. Most importantly until recently there were no handy methods available for the study of the electrochemistry of any solid sample independent of its electrical conductivity solubility etc. Since electro- chemistry always mirrors the thermodynamics and kinetics of interfacial reactions it follows that electrochemical measure- ments on solid compounds may provide valuable information concerning qualitative and quantitative composition chemical equilibria and kinetics of reactions of a solid compound. Of course it is clear that we cannot get direct information concern- ing crystal structure bond length bond angles etc. Voltammetry i.e. the recording of current versus electrode potential has become the most important electrochemical measuring technique in the field of pure and applied electro- chemistry. Jaroslav Heyrovsky's direct current polarography the first highlight of voltammetric analysis was followed by a great variety of sophisticated techniques,' which have won an undisputed place in the arsenal of modern analytical chemistry. The historic roots of electrochemistry lay in studies of pheno- mena occurring at interfaces of solids with solutions. However electrounulysis was almost exclusively focused on solution stu- dies with liquid mercury electrodes and the use of solid elec- trodes remained fairly limited.2 The first studies of the electro- chemistry of solids which were not metals alloys or semiconductors but insoluble metal oxides were performed as late as 1925.3 To allow electroanalytical studies of all solid substances. new experimental strategies had to be developed and new theoretical concepts for understanding this electro-chemistry were necessary. Potentiometric methods gave access to solubility products of sparingly soluble compounds in cases where electrodes of the second kind were amenable. Studies on galvanic cells with solid clectrolytes provided access to thermodynamic data such as free energy of formation etc. of various solid corn pound^.^ Obviously these methods are not suitable for the analysis of solid compounds but they are part of fundamental electrochemistry. This review is focused on the analj~ticaluse of voltammetry for the study of solid compounds. It does not intend to cover the Fritz Scholz Birgit Meyer chemistry. Nowadays it is reasonable to speak of electrochemical solid-state analysis as a field of research in its own right and many interesting discoveries can be expected in the future from studies on the electrochemistry of solid substances. 2 Electrography. The Early Beginnings of Electrochemical Sol id -state Analysis In the late twenties of our century Glazunov and Fritz indepen- dently developed electrography or as Fritz called it Elektro-Tupjelmethode (electric spot test method). The principle of electrography is that a sample of a metal an alloy or any other conductor is anodically oxidized while being pressed against a sheet of paper which has been impregnated with a reagent. The sample is connected to the positive pole of a current supply and the negative pole is connected to a counter electrode. This counter electrode is a sheet of metal beneath the impregnated paper (see Figure 1). Through anodic dissolution of the sample Figure 1 Equipment for electrography. A anode; Al cathode; N mineral; p paper with reagent. (Reproduced with permission from reference 5.) Fritz Scholz was born in Zeuthen Germany andgraduated with a Diplom in Chemistry,from Humboldt University Berlin where he was awarded a Ph.D. (Dr. rer nat.) in 1982 and Dr. sc. (habili-tation) in 1987. In 1981 he worked at the University of Warsaw and in 1987 and 1989 with A. M. Bond at Deakin Universitj! Australia. In 1993he became Professor ofAnalytica1 and Environ- mental Chemistry at Humboldt University. His research interests cover electrochemical trace analysis of dissolved species and electrochemistry of solid compounds. Birgit Meyer nhe Lange was born in Neustrelitz Germany and graduated with a Diplom in Chemistr-v from Humboldt University in 1990. Since 1991 she has been a Ph.D. student with Fritz Scholz. Her interest is focused on the electrochemistry of minerals kith applications to mineral analysis. 34 1 342 metal ions are released from the sample and form coloured compounds with the reagent in the paper In this way an image of the sample is produced which shows the distribution of for example nickel in a nickel ore when dimethyl glyoxime is the reagent Thus electrography was an early kind of spatial distri- bution analysis 3 Compact Electrodes Made of a Solid Substance This technique can be used only for solid materials which possess sufficient electrical conductivity and which can be manufactured into electrodes Hence only metals alloys and semiconductors can be investigated by this technique For analytical purposes the practical obstacles of this technique prevent a broader application It is indeed laborious to make compact electrodes of solids I e to shape them into a disc to polish this disc to mount the disc in a holder etc ,to make a voltammetric measurement on only one sample Therefore pressed cells have been developed These are electrochemical cells with a counter electrode and possibly a reference electrode in a cell without a bottom Around the ‘missing’ bottom there is a rubber ring so that the entire cell can be pressed against the metal surface which is to be studied (for a review of applications of pressed cells see reference 6) After filling the cell with electrolyte measurements can commence A way of manufacturing compact electrodes from powders of conducting material is to press tablets with or without a filling polymer In addition to the tedious procedures needed to make com- pact electrodes there are other more fundamental reasons for not applying such electrodes to solid-state analysis One of these reasons is that compact electrodes give rise to high currents High currents can lead to iR-distorted voltammograms which hide information on thermodynamics or kinetics In cases where there is more than one current signal the resolution is poor and a multicomponent analysis can become impossible High currents may also lead to unwanted chemical reactions on the surface of the electrode e g the precipitation of compounds which cover the electrode surface and inhibit further electrochemical reac- tions Published voltammograms obtained with compact elec- trodes often illustrate their unfavourable behaviour (Figure 2) For a review of applications of compact solid electrodes see references 7 and 8 EN 08v -0 0 0 Figure 2 Anodic voltammograms of Cu-Sn alloys obtained with d pressed cell (1) 92 13% Cu (2),53 80% Cu (3) 19 92% Cu Note the potential scale’ (Reproduced with permission from reference 6 ) CHEMICAL SOCIETY REVIEWS 1994 4 Carbon Paste Electrodes with Organic Binder and Addition of a Solid Substance One way of partly overcoming the disadvantages of solid electrodes that have been mentioned is to use carbon paste electrodes Here the solid compound is ‘diluted’ with graphite and formed to a paste with an organic binder This has the advantage that there is a reduced amount of substance and therefore the currents are smaller and surface coverage can be minimized Kuwana and French9 used this technique for vol- tammetric measurements on water-insoluble organic com-pounds such as ferrocene or anthraquinone The technique of modified paste electrodes has been extensively developed in the former Soviet Union by W G Barikov 0 A Songina N F Zakharchuk and Kh Z Brainina Several methods have been described for the quantitative analysis of powder mixtures e g Ag,O-Ago cubic and hexagonal In203 iron-magnetite-wus- tite mixtures etc ’A highlight of the application of modified paste electrodes has been the determination of the composition of thin oxide films (2-1 00 nm) which form on GaAs by anodic oxidation or by thermal oxidation lo Figure 3 shows the pub- lished derivative d c -voltammograms This figure illustrates how an identification of the following phases is possible a-,p- y- 6- c-Ga203 amorphous Ga(OH) claudetite and arsenolite (both As,O,) a-and P-Sb2O3 As,O As205 Sb205 GaAsO As and Sb (amorphous and crystalline) and Ga When these results were published there were no other analytical spectro- scopic techniques including ESCA and Auger electron spectro- scopy which would have allowed the simultaneous identifica- tion of all these compounds The thin film was mechanically removed with the help of diamond powder and the resulting mixture of film material and diamond powder was attached to the surface of a graphite paste electrode before commencing the voltammetric measurements A recent review of the fundamen- tals and limitations of modified graphite paste electrodes dis- cusses the influence of particle size on reproducibility of results and the entire procedure of paste preparation The influence of the organic binder on the electrochemistry of the solid com- pounds has not yet been fully elucidated but in all probability the organic compound obscures at least partly the faradayic reactions of the solid compound (ef the well-documented influence of adsorbed layers of surface-active compounds on the electrochemistry of dissolved species) 5 Carbon Paste Electrodes with Electrolytic Binder and Addition of a Solid Substance In this technique the carbon paste electrode is formed using an electrolytic binder instead of an organic binder So the exchange of electrons can in principle be achieved in the whole paste not only at the paste4ectrolyte interface In practice the iR-drop within the paste will certainly restrict the electrochemical reac- tion to a layer of paste adjacent to the electrolyte solution Such paste has to be housed in small cups with the upper surface exposed to the electrolyte solution which has to be the same composition as the electrolytic paste binder Bauer and Gaillo- chetI2 studied how the electrochemical behaviour of these electrodes depended on different parameters such as sweep rate paste volume and concentration of sample in the paste They concluded that the solid itself is involved in the electrochemical processes Lamache et ul investigated the oxidation of Cu,S in 1M H,SO at very low scan rates and found that the oxidation proceeds in steps forming copper sulfides of different stoichio- metry (Figure 4) Some of these sulfides are minerals occurring in nature Chouaib et ul l4 looked at the behaviour of different manganese oxides and they were able to distinguish different modifications through the different electrochemical reactions Eguren eta/ have described a method for the determination of the P-SnO contents in commercial tin dioxide A review on carbon paste electrodes with an electrolytic binder has been published by Batanero et a1 l6 Although paste electrodes do not suffer from the undesirable influence of organic compounds that ELECTROCHEMICAL SOLID STATE ANALYSIS STATE OF THE ART-F SCHOLZ AND B MEYER Figure3 Spectrum of the standard substances for the Ga-As-0 systemobtained by cyclic voltammetry Measurements were performed in 3 M HCl for As GaAsO As,O and As,O in I M HCl for As,O and in 0 5 M HCl for Ga,O Ga(OH) and Ga The reference electrode was a SCE (Reproduced with permission from reference 10 ) electrodes using organic binders experience the electrolytic binder does need to be carefully chosen Very aggressive binders like mineral acids or bases for example may chemically dissolve the solid compound before the commencement of the electro- chemistry thus affording erratic results 6 Voltammetry of Suspended Solid Particles Kolthoff and Stock” were the first to publish voltammograms of suspended silver bromide By chance Micka observed that suspended charcoal gives specific signals in polarography Later he and Kalvoda systematically investigated the polarography of vdrious solid substances suspended in electrolyte solution * To keep the particles in suspension and to ensure that they can come into contact with the electrode surface the suspensions have to be stirred Especially when a dropping mercury electrode is used this stirring leads to very noisy voltammograms (see Figure 5) This problem can be circumvented by using a rotating disc electrode the rotation of the electrode guarantees stable hydrodynamic conditions The results obtained by Micka indi- cdted that the voltammetric response of suspensions is con- nectcd with the point of zero charge (p z c ) because the suspen- sions gave peak-shaped current signals which were situated at or near the p z c Micka interpretated the peak-shaped curves as A1 A2 A3 Aq A5 CUPS t)CU~& t)CU~nS t)CUl6&3 + Cul31S t) CUS ,r c4 1 c5 Aq 05-A3 0 EN I1 I,,I Figure 5 Reduction of manganese dioxide of the rutile type Twelve millilitres of 0 2 M H,SO + 6 mg MnO (I) Electrically stirred suspension (2) Quit suspension Sensitivity ljl000 beginning from 0 V 200 mV/absc (Reproduced with permission from reference 18 ) p z c Applying the theory of Frumkin on the potential depen- dence of the stability of surface films they came to the following understanding At the p z c the electrolyte film between the electrode surface and the suspended particles has such instability that the particles can come into closest contact with the elec- trode where they are deposited on the surface by electrocapill- ary forces They called this phenomenon elect)ocupzllu~J deposi-tion This deposition ensures that the particles come into such close contact that electrons can be transferred between the particles and the electrode At potentials fdr away from the p z c the stability of the electrolyte film is so high that even forced transport of the particles to the electrode surface by stirring is unable to bring the particles into closest contact with the electrode In this connection it should be remembered that some compounds become electro-inactive when they are in colloidal dispersion It is then probably impossible to disrupt the electro- lyte film between the very small particles of the colloid and the electrode surface 7 Voltammetry of Solid Compounds Immobilized within a Polymeric Film on the Electrode Surface Voltammetry of polymer films on electrodes has been extensi- vely studied in the context of surface-modified electrodes dnd the term solid-state volturnrnetl-j has been used for those systems in which redox-active sites are incorporated in the polymer film 21 The idea of holding particles of a solid compound on the surface of a solid electrode by using a polymeric binder has been applied to the study of minerals * Franklin et a1 22 have des- cribed the voltammetry of suspended solid compounds in catio- nic surfactant-styrene-aqueous sodium hydroxide emulsions using platinum electrodes Under these conditions a hydropho- bic polymeric film is formed on the electrode surface which results in an enlarged potential window for measurements The solid particles are held on the electrode surface through the adsorption of surfactants embedding these particles From the literature it is not possible to answer the question of how the polymer film influences the electrochemistry of the solid parti- cles The polymer film may have a negative effect on the voltammograms but evidence for or against has not been given yet One thing however is certain -every measurement needs the preparation of a new electrode which makes the technique rather inconvenient to use 8 Voltammetry of Solid Compounds Sandwiched between Two Solid Electrodes Kulesza et a1 23 have shown that solid compounds which possess ion conductivity and mixed valence sites give voltammetric signals when they are sandwiched between two solid electrodes Although no deliberately added electrolyte solution is present in CHEMICAL SOCIETY REVIEWS 1994 these systems the degree of hydration strongly influences the voltammetric response since a certain water content is essential for the ion mobility This kind of voltammetry has been des- cribed for metal hexacyanoferrates and single crystals of silico- tungstic acid For hexdcyanoferrdtes the voltammetric response is due to the following reactions Thus it is electron hopping between the mixed valence sites and counter-ion flux which provides the current flow through the solid compound The electrochemistry is similar to that which is observed when d film of solid compounds e g prussian blue is deposited on a metal electrode and voltammograms are recorded in an electrolyte solution 24 Since the described technique is limited to a small number of compounds it cannot find a broader application for analysis but the results of these studies are very valuable for understand- ing the electrochemistry of solid compounds and will be of benefit for other techniques as well 9 Abrasive Stripping Voltammetry Abrasive stripping voltammetry (AbrSV) is a new approach for the direct study of solid samples which has recently been introduced by Scholz et ul 25 It makes use of the fact that (from Faraday's law) extremely small amounts of a sample are suffi- cient to give easily measurable currents The use of paraffin- impregnated graphite electrodes to fix such small amounts of solid particles on an electrode surface has proved to be very successful although in principle any solid electrode is suitable The transfer of the solid compound can be achieved simply by abrasion after which small solid particles stick to the electrode surface For the fabrication of paraffin-impregnated graphite electrodes (PIGE) soft graphite rods are put into molten paraffin under vacuum until air bubbles cease to evolve from the rods After re-establishing atmospheric pressure the rods are removed before the paraffin solidifies For transfer of the solid sample to the electrode surface the solid compound is powdered dnd placed on a glazed porcelain tile and spread out with an dgdte mortar to form d spot of finely distributed material Then the lower circular end of the PIGE is gently rubbed over that spot of sample Trace amounts down to about lop6 to lo-" mol of sample are mechanically immobilized on the electrode surface Effective abrasion is often also possible by rubbing the electrode on d smooth surfdce of d solid even when this is harder than the electrode surface In the case of an extremely hard material corundum powder may be added to support the abrasion After the transfer of the solid sample to the electrode surface the electrode is dipped into a conventional electrochemical cell with auxiliary and reference electrodes so that only the circular surface of the electrode is in contact with the electrolyte solution (see Figure 6) In this way it is possible to achieve good Workingelectrode Auxiliary electrode I I I Reference electrode \Electrolyte solution / Figure 6 Electrochemical cell for d brasive stripping voltammetry ELECTROCHEMICAL SOLID STATE ANALYSIS STATE OF THE ART-F. rcproducibilit) of the background current and an almost con- stant electrode area. After the measurement. the electrode is cleaned bq rubbing the surface on filter paper. The effectiveness of the cleaning procedure can be checked by recording a blank \ oltammogram. All the usual electrochemical techniques can be carried out by .4brSV. Thus. for example. the oxidation or reduction of solid compounds can be studied and used for analytical purposes. In addition it is possible to record so-called inverse voltammo- grams. This technique involves reducing the cations of a solid compound to metals which become deposited on the electrode surface.26 and in the following anodic voltammogram anodi- cally dissolving the metals in the electrolyte solution. The oxidation peaks obtained allow qualitative identification of the elements and quantification of composition of the solid (cf. Figure 7).By coprecipitation of mercury from the electrolyte solution during the reduction of a solid metal compound the metals formed dissolve in the in sifu plated mercury droplets on the electrode surface and the resolution of the following anodic \ oltammograms is remarkably improved. This technique results in relative standard deviations in quantitative analysis as low as 0,5O/0,~-Such voltammograms with in situ plated mercury are identical to the anodic stripping voltammograms which are well documented for mercury electrodes. -1 .o -0.2 -1 .o -0.2-1 .o -0.2 Potential Vvs.Ag/AgCI Figure 7 Incerse abrasiLe stripping coltammograms of different TI -Sn sulfo-salts in 1 M HCI. Scan rateO.O1 V s. deposition potential -1 .O\'. deposition time 60 s. differential pulse modus. AbrSV can be applied to all solid compounds which contain at least one electrochemically active element. Experiments were initially conducted on pure metals and simple alloys. Anodic dissolution peaks occur at potentials which arecharacteristic for the metals in specific electrolytes and allow their identification. In the case of alloys one can obtain qualitative information on the elemental constituents provided that there are no strong interactions between the constituents in the solid phase. When intermetallic phases are present they can be identified by their specific signals. Quantitative determination of the alloy consti- tuents is possible because the peak height ratios of the alloy constituents depend on the composition.* For this a calibration plot is necessary. AbrSV also provides information on the electrochemical corrosion behaviour of alloys. Detailed studies have been undertaken on dental amalgams.28 The results showed unambiguously that the corrosion of dental amalgams containing the often blamed rz-phase (Sn,,,Hg) is only worse than that of y2-phase-free amalgams when the electrolyte does not contain strongly complexing ions. In solutions with a high concentration of citric acid (e.g. beverages) y2-phase-free and y2-phase-containing amalgams have the same stability. A fascinating field of application is the voltammetric identifi- cation and analysis of minerals.26 With AbrSV. qualitative SCHOLZ AND B. MEYER Proustite 2--F 0 160 Timeis Figure 8 Abrasive stripping coulogram of mechanically transferred proustite in 1 M KCI. Potentials -0.6 V (20 s). -1.1 V (60 s) (reduction of proustite). -0.6 V (20 s). 0.2 V (60 s) (oxidation of silxer). information on the metallic constituents is accessible from an inverse voltammogram provided that no intermetallic phases are formed during the reduction. Every voltammogram is a voltammetric fingerprint for a particular mineral in a specific electrolyte; unambiguous identification of minerals can thus be achieved. Where there are different modifications of a mineral it is impossible to distinguish between them by conventional electrochemistry after digestion. because the same solution results. In AbrSV however the solid substance is not destroyed before analysis and therefore the influence of the structure of the solid on the electrochemistry is not lost. Thus different modifica- tions give different voltammograms. This is not true for inverse AbrSV which therefore cannot be used to distinguish between different modifications. In some cases it is possible and useful to perform coulometric measurements. Figure 8 depicts a coulogram of mechanically transferred proustite from which the ratio of charges for reduc- tion (Ag,AsS + 9e-+ 3H-+3Ag + ASH + 3S2-) to the charge for subsequent oxidation ofthe Ag (3Ag +3Ag' + 3e-) formed can be easily derived.z9 In some special cases if the electrochemistry is reversible. it is possible to calculate thermo- dynamic data for minerals.30 AbrSV is a valuable technique for the study of insoluble substances and provides hitherto inaccess- ible information on the electrochemistry and chemistry of such compounds. As an example. it was possible to compare the electrochemistry of mercury and lead dithiocarbamates in di- chloromethane solution with the electrochemistry of the solid compounds at the aqueous electrolyte electrode interface. It turned out that during cyclic reduction of the dithiocarbamates and oxidation of the deposited metals to the dithiocarbamates the formal potentials of these redox systems were determined by the conditional brutto stability constants of the complexes. These thermodynamic data could be easily determined and comparison with some known values of complexes showed excellent agreement.31.3z Bond et used the technique of abrasive stripping voltammetry in a detailed study of the electro- chemistry of solid microcrystalline cis-and rrcms-Cr(CO),(dpe) and tr~ns-[Cr(CO),(dpe),]~ complexes (dpe = Ph,PCH,CH PPh,). Those compounds are entirely insoluble and they are electrical insulators. Nevertheless they give well-defined voltam- mograms when mechanically attached to a pyrolytic graphite electrode (see Figure 9). With the help of X-ray electron probe analysis it was shown that the oxidation of these complexes is accompanied by the incorporation of perchlorate ions into the lattice of the solid compounds. In a similar way to the behaviour of the sandwiched hexacyanoferrates electron hopping between the metal centres occurs and electroneutrality is provided by a flux of counter ions. The chromium complexes differ from the hexacyanoferrates in that the former are not ion conductors and do not contain water. Dueber et have used AbrSV to study T 5cIA Figure 9 Cyclic voltammograms obtained in aqueous (0 1 M NaCIO,) media at 20 "C for solid trans-[Cr(CO),(dpe),] mechanically+ attached to a polished basal plane pyrolytic graphite electrode (scan rdte 50 mV/s) (a) Initial and final potential = 0 2 V switching potential = 1 2 V (b) initial and final potential = 0 2 V switching potentidl = 1 0 V (c) first cycle initial and final potential = -1 0 V switching potential = 1 2 V vs Ag/AgCl (d) as for (c) but six cycles (Reproduced with permission from reference 33 ) the intercalation of magnesium ions into uranium oxides The investigation of electrochemically-induced intercalation of ions into a host lattice is of utmost importance for the design of rechargeable batteries AbrSV has also proved to be applicable to the quantitative analysis of powder mixtures 35 Another group of compounds which has been studied by AbrSV is that of the high-temperature superconductors It was found that in the case of YBaCu high-T superconductors superconductivity was observed only when Cu3 +/Cu2 + and Cu2+/Cu+ couples were present 36 The theory of AbrSV has been addressed by Lovric et a1 37 38 for the case of square-wave voltammetry of immobilized reac- tants and by Jaworski et a1 39 for Iinear-sweep voltammetry of reversible reduction of metal salts For a review on abrasive stripping voltammetry see reference 35 10 Mechanistic Considerations Concerning the Electrochemistry of Solid Compounds Faradayic reactions of solid compounds at the interface with an electrolyte solution z e reactions involving an electron transfer between an inert electron conductor and a solid compound may proceed along different pathways They are always accompa- nied by an ion transfer between the solid phase of the reacting compound and the adjacent solution phase to preserve charge equilibrium in each phase The reaction pathway of a solid compound depends first on its electron conductivity In the case of a sufficient electron conductivity faradayic reactions can proceed directly on the surface of the solid compound Examples are the anodic dissolution of a metal likecopper the reduction of lead ions in solid lead sulfide PbS and also the oxidation of sulfide ions in the latter compound When copper is oxidized the copper ions are transferred into the electrolyte phase and the solid copper phase dissolves When the lead ions of PbS are reduced to metallic lead this metal forms a new solid phase and the sulfide ions are transferred into the electrolyte phase The oxidation of sulfide ions of PbS leads to the formation of a new CHEMICAL SOCIETY REVIEWS 1994 solid elemental sulfur and the transfer of lead ions into the solution phase In the case of solid compounds with an insufficient electron conductivity the faradayic reactions are more difficult to under- stand The transfer of electrons can only proceed at the interface of the inert electrode and the solid compound when there IS also the possibility for an exchange of ions between the solid com- pound and the solution phase This is easily possible dt those places where the three phases -inert electrode solid compound and electrolyte solution -are in contact with each other A similar mechanism is known for electrochemical reactions in emulsions 4n Some solid compounds possess the ability to conduct electrons by intra- or intermolecular redox redctions within the solid compound This electron hopping has to be accompanied by a flux of ions to preserve the charge neutrality Compounds which follow such a mechanism include the Prus- sian Blue analogue compounds and the organic chromium compounds discussed above In case of solid compounds which are sparingly soluble (or not fully insoluble) it can be observed that dissolved species undergo electrochemical reactions at the surrounding electrode surface When the product of this electrode reaction is again insoluble as is the case when dissolved lead dithiocarbamate is reduced to lead metal a restructuring of the electrode surface takes place in the course of cyclic polarization Besides the direct electrochemical reactions of solid com- pounds a mediated reduction or oxidation of solid compounds on the electrode surface is possible Hydrogen in particular has been suggested to be the mediator for the reduction of metal salts 11 Conclusions Electrochemistry now possesses many experimental tools for the study of phenomena at the interface between solid compounds and electrolyte solutions This experimental basis makes poss- ible an almost infinite number of new experimental investi- gations and challenges our theoretical understanding of electro- chemical phenomena It is to be expected that many branches of science will benefit from these developmcnts In analytical chemistry both organic and inorganic direct solid-stdte dndly-sis using electrochemical techniques will be attractive for many practical applications Battery and fuel cell development can benefit from solid-state electrochemical measurements because they enable the clear elucidation of electrode reactions Mdter- ials science and corrosion science get new tools for the character- ization of materials and for the study of their electrochemical corrosion Mineralogy gets microanalytical techniques for the identification and analysis of mineral phases even for the most tiny amounts Moreover mineralogy can use these electroche- mical techniques to get access to the chemical and electrochemi- cal reactions of minerals which are otherwise difficult to study Acknowledgement Since I 99 1 Deutsche Forschungsgemeinschajt has substantially supported the abrasive stripping voltammetry project at Humboldt University Without this help it would have been impossible to develop this kind of electrochemical solid- state analysis The authors also acknowledge support by Fonds der Chemischen Industrie We are especially indebted to all colleagues who are cited as co-authors of previous communica- tions concerning abrasive strippcng voltammetrq 12 References 1 'Analytical Voltammetry' ed M R Smyth and J G Vos Vol 27 of 'Wilson and Wilson's Comprehensive Analytical Chemistry ,ed G Svehld Elsevier Amsterddm London New York Tokyo 1992 2 A M Bond and F Scholz Z Cizeni 1990,30 117 3 K Fischbeck Z Anorg Allg Cheni 1925 148 97 4 H Rickert 'Electrochemistry of Solids An Introduction Springer Verlag Berlin I982 5 R Jirkovsky Mikrochemre 1934 15 (N F 9) 331 6 V V Slepushkin Zh Anal Kizrm 1987 42 606 ELECTROCHEMICAL SOLID STATE ANALYSIS STATE OF THE ART-F SCHOLZ AND B MEYER 7 Kh Z Brainina E Ya Neyman and V V Slepushkin ‘Inversion- nye elektro analiticheskie rnetody’ Khmiya Moskva 1988 8 V A Chanturiya and V E Vigdergauz ‘Elektrokhimiya sulfidov teoriya i praktika flotatsii’ Nauka Moskva 1993 9 T Kuwana and W G French Anal Chem 1964,36,241 10 V I Belyi T P Smirnova and N F Zakharchuk. Appl Surf Scl . 1989,39 161 1 1 W Gruner J Kunath L N Kalnishevskaja J V Posokin and Kh Z Brainina Electroanaljsrs 1993 5 243 12 D Bauer and M Ph Gaillochet Efectrochim Acta 1974 19 597 13 M Ldmdche and D Bauer Anal Chem 1979,51 1320 14 F Chouaib 0 Cauquil and M Lamache Electrocltirn Acta 198 1 26,325 I5 M Eguren M L Tascon M D Vdzques and P Sanchez Batanero Electrockmi Acta 1988 33 1009 16 P Sanchez Batanero M L Tascon Garcia and M D Vazques Bdrbddo Quimica Analiticu 1989 8 393 17 I M Kolthoff and J T Stock Anallst 1955,80,860 18 K Micka. In ‘Advances in Polarography’ Proceedings of the 2nd Int Congress on POldrOgrdphy Cambridge 1959 ed I S Longmuir Pergdmon Press New York Oxford London Paris 1960 Vol 3 (1959) pp 1182-1 190 19 R Kdlvodd In ref 18 pp 1172-1 181 20 M R Daushevd and 0 A Songina Usp Khrm 1973,42,323 21 TT Wooster M L Longmire H Zhang M Watanabe and R W Murray Anal Chem 1992 64 1 132 and references cited therein 22 T C Franklin J Darlington R Nnodimele and R C Duty in ‘Electrochemistry in Colloids and Dispersions’ ed R A Mackay. J Texter VCH Publishers Inc ,New York 1992 pp 319-329 23 P Kulesza and Z Galus J Electroanal Chem 1992,269,261 24 D N Upadhyay and D M Kolb J Electroanal Chem 1993,358 317 25 F Scholz. L Nitschke and G Henrion Naturwrssenschaften 1989 76,71 26 F Scholz L Nitschke G Henrion and F Darnaschun Fresenius Z Anal Chem 1989,335 189 27 F Scholz and B Lange Fresenius J Anal Chem 1990,338,293 28 F Scholz F Rabi and W -D Muller Electroanalysu 1992,4 339 29 S Zhang. B Meyer G Moh and F Scholz,Eleciroanalysu,in press 30 B Lange F Scholz H -J Bautsch F Damaschun and G Wappler Phrs Chem Minerals 1993 19,486 31 A M Bond and F Scholz J Phjs Chem 1991,95,7460 32 A M Bond and F Scholz Langmurr 1991,7 3 197 33 A M Bond R Colton F Daniels,D R Fernando F Marken,Y Nagdosa R F M Van Steveninck and J N Walter J Am Chem SOC 1993 115,9556 34 R E Dueber,A M Bond,and P G Dickens,J Elecrrochem SOC 1992,139,2363 35 F Scholz and B Lange Trendy Anal Chem 1992,11,359 36 S Scheurell F Scholz T Olesch and E Kemnitz Supercond Sci Technof 1992,5,303 37 M Lovric S Komorsky-Lovric and A M Bond J Electroanal Chem 1991 319 1 38 S Komorsky-Lovric M Lovric and A M Bond Anal Chrm Acra 1992,258,299 39 A Jaworski Z Stojek and F Scholz J Electroanal Chem 1993 354 1 40 H Fees and H Wendt Ber Bunsenges Phjs Chem 198I 85,914

ISSN:0306-0012    

DOI:10.1039/CS9942300341

出版商:RSC    

年代:1994

数据来源: RSC