ADVANCED INORGANIC CHEMISTRY A Comprehensive Text F. ALBERT COTTON ROBERT A. WELCH DISTINGUISHED PROFESSOR OF CHEMISTRY TEXAS A AND M UNIVERSITY COLLEGE STATION, TEXAS, USA’.* and GEOFFREY WILKINSON SIR EDWARD FRANKLAND PROFESSOR OF INORGANIC CHEMISTRY IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY UNIVERSITY OF LONDON, ENGLAND Fourth Edition, completely revised from the original literature A WILEY-INTERSCIENCE PUBLICATION JOHN WILEY & SONS NEW YORK + CHICHESTER + BRISBANE + TORONTO407th Copyright © 1980 by John Wiley & Sons, Inc. All rights reserved. Published simultaneously in Canada, Reproduction or translation of any pa¥ of this work beyond that permitted by Sections 107 and 108 of the 1976 United States Copyright Act without the permission of the copyright owner is unlawful. Requests for permission or further information should be addressed to the Permissions Department, John Wiley & Sons, Inc. Library of Congress Cataloging in Publication Data Cotton, Frank Albert, 1930- ‘Advanced inorganic chemistry. “A Wiley-Interscience publication.” Includes index. 1. Chemistry, Inorganic. 1. Wilkinson, Geoffrey, 1921- joint author. _II. Title. QDIS1.2.C68 1980 54679-22506 ISBN 0-471-02775-8 Printed in the United States of America 10987654321 — AS 26 /Preface to the Fourth Edition It is remarkable how the subject of inorganic chemistry has not only grown but changed in form and emphasis since the early 1970s. In this fourth edition we have made major alterations in the arrangement of our material in an effort to reflect these changes. The main purpose of our book has not changed. It is to provide the student, or other reader, with the knowledge necessary to read with comprehension the con- temporary research literature in inorganic chemistry and certain areas of or- ganometallic chemistry. We have, as usual, also updated our coverage of all topics to include developments published as late as the first half of 1979. This, too, has entailed a significant in- crease in the amount of factual material. To keep the size of the book from getting out of hand we have omitted still more of the relatively elementary material included in earlier editions. There are now a number of more elementary textbooks, including our own Basic Inorganic Chemistry, where elementary topics are fully covered. For this reason we believe that the absence of this material here is pedagogically acceptable. We again wish to thank all of those who have been kind enough to give us con- structive suggestions and insight into fields where they are experts and we are not. We continue to be receptive to sucl contributions. F. ALBERT COTTON GEOFFREY WILKINSON College Station, Texas, USA London, England January 1980Preface to the First Edition It is now a truism that, in recent years, inorganic chemistry has experienced an impressive renaissance. Academic and industrial research in inorganic chemistry is flourishing, and the output of research papers and reviews is growing exponen- ly. In spite of this interest, however, there has been no comprehensive textbook on inorganic chemistry at an advanced level incorporating the many new chemical developments, particularly the more recent theoretical advances in the interpretation of bonding and reactivity in inorganic compounds. It is the aim of this book, which is based on courses given by the authors over the past five to ten years, to fill this need. It is our hope that it will provide a sound basis in contemporary inorganic chemistry for the new generation of students and will stimulate their interest in a field in which trained personnel are still exceedingly scarce in both academic and industrial laboratories. The content of this book, which encompasses the chemistry of all of the chemical elements and their compounds, including interpretative discussion in the light of the latest advances in structural chemistry, general valence theory, and, particularly, ligand field theory, provides a reasonable achievement for students at the B.Sc. honors level in British universities and at the senior year or first year graduate level in American universities. Our experience is that a course of about eighty lectures is desirable as a guide to the study of this material. We are indebted to several of our colleagues, who have read sections of the manuscript, for their suggestions and criticism. It is, of course, the authors alone who are responsible for any errors or omissions in the final draft. We also thank the various authors and editors who have so kindly given us permission to reproduce diagrams from their papers: specific acknowledgements are made in the text. We sincerely appreciate the secretarial assistance of Miss C. M. Ross and Mrs. A. B. Blake in the preparation of the manuscript. F, A. COTTON G. WILKINSON Cambridge, Massachusetts London, England viiContents PART ONE Introductory Topics 1. Nonmolecular Solids 2, Symmetry and Structure 3. Introduction to Ligands and Complexes 4. Classification of Ligands by Donor Atoms 5. Stereochemistry and Bonding in Main Group Compounds PART TWO Chemistry of the Main Group Elements 6. Hydrogen 7, The Group I Elements: Li, Na, K, Rb, Cs 8. Beryllium and the Group II Elements: Mg, Ca, Sr, Ba, Ra 9. Boron 10. The Group II Elements: Al, Ga, In, TI 11. Carbon 12. The Group IV Elements: Si, Ge, Sn, Pb 13. Nitrogen 14. The Group V Elements: P, As, Sb, Bi 15, Oxygen 16. The Group VI Elements: S, Se, Te, Po 17. The Halogens: F, Cl, Br, I, At 18. The Noble Gases 19. Zine, Cadmium, and Mercury 28 61 107 195 215 253 an 289 326 352 374 407 438 502 542 577 589x CONTENTS PART THREE Chemistry of the Transition Elements 20. The Transition Elements and the Electronic Structures of Their Compounds 21. The Elements of the First Transition Series . Titanium, 692 . Vanadium, 708 . Chromium, 719 . Manganese, 736 Iron, 749 Cobalt, 766 Nickel, 783 Copper, 798 e Elements of the Second and Third Transition Series . Zirconium and Hafnium, 824 Niobium and Tantalum, 831 Molybdenum and Tungsten, 844 . Technetium and Rhenium, 883 The Platinum Metals, 901 Ruthenium and Osmium, 912 Rhodium and Iridium, 934 . Palladium and Platinum, 950 Silver and Gold, 966 23. The Lanthanides; also Scandium and Yttrium 24, The Actinide Elements 22. FmMO™MOOP> STO ™MIOw> 5 s gms PART FOUR Special Topics 25, Metal Carbonyls and Other Complexes with -Acceptor Ligands 26. Metal-to-Metal Bonds and Metal Atom Clusters 27. Transition Metal Compounds with Bonds to Hydrogen and Carbon 28, Reaction Mechanisms and Molecular Rearrangements in Complexes 29, Transition Metal to Carbon Bonds in Synthesis 30. Transition Metal to Carbon Bonds in Catalysis 31. Bioinorganic Chemistry Appendices 1. Units, Fundamental Constants, and Conversion Factors 619 689 822 981 1005 1049 1080 1113 1183 1234 1265 1310 1347CONTENTS 2. Ionization Enthalpies of the Atoms 3. Enthalpies of Electron Attachment for Atoms 4. Atomic Orbitals 5. The Quantum States Derived from Electronic Configurations 6. Magnetic Properties of Chemical Compounds Index xi 1349 1351 1351 1354 1359 1367Abbreviations in Common Use 1. Chemicals, Ligands, Radicals, etc. Ac acetyl, CH;CO acac acetylacetonate anion acacH acetylacetone AIBN azoisobutyronitrile am ammonia (or occasionally an amine) Ar aryl or arene (ArH) aq aquated, H,O ATP adenosine triphosphate 9-BBN 9-borabicyclo[3,3,1]nonane bipy 2,2’-dipyridine, or bipyridine Bu butyl (superscript n, i, s or t, normal, iso, secondary or tertiary butyl) Bz benzyl COD orcod —_cycloocta-1,5-diene COT or cot cyclooctatetraene Cp cyclopentadienyl, CsHs cy cyclohexyl depe 1,2-bis(diethylphosphino)ethane depm 1,2-bis(diethylphosphino) methane diars o-phenylenebisdimethylarsine, o-CeHa(AsMe2)2 dien diethylenetriamine, HXN(CH2CH,NH)2H diglyme diethyleneglycoldimethylether, CH30(CH2CH20).CH3 diop {[2,2-dimethyl-1 ,3-dioxolan-4,5-diyl)bis(methylene) ]bis(diphen- ylphosphine)} diphos any chelating diphosphine, but usually 1,2- bis(diphenylphosphino)ethane, dppe DME dimethoxyethane DMF ordmf —_N,N’-dimethylformamide, HCONMe dmg dimethylglyoximate anion dmgH> dimethylglyoxime dmpe 1,2-bis(dimethylphosphino)ethane DMSO or dimethylsulfoxide, MezSO dmso xiiixiv dppe DPPH dppm EDTAHs EDTAHS, en Et Fe Fp glyme hfa HMPA L M Me Mes Megtren NBD or nbd NBS np? np} NTAH3 OAc Ox Pe Ph phen pn PNP (= np?) pp? PPNt+ Pr py pz QAS QP R Re Ss sal salzen or salen ABBREVIATIONS IN COMMON USE 1,2-bis(diphenylphosphino)ethane diphenylpicrylhydrazyl bis(diphenylphosphino)methane electrophile or element ethylenediaminetetraacetic acid anions of EDTAH4 ethylenediamine, H2NCH2CH2NH2 ethyl ferrocenyl Fe(CO)2Cp ethyleneglycoldimethylether, CH;0CH2CH,OCH3 hexafluoroacetylacetonate anion hexamethylphosphoric triamide, OP(NMe)3 ligand central (usually metal) atom in compound methyl mesityl tris-(2-dimethylaminoethyl)amine, N(CH)CH,NMe)3 norbornadiene N-bromosuccinimide Bis-(2-diphenylphosphinoethyl)amine, HN(CH2CH2PPh»)2 Tris-(2-diphenylphosphinoethyl)amine, N(CHxCH>PPh2)3 nitrilotriacetic acid, N(CH2COOH)3 acetate anion oxalate ion, C,0j- phthalocyanine phenyl, CeHs 1,10-phenanthroline propylenediamine (1,2-diaminopropane) Bis-(2-diphenylphosphinoethyl)amine, HN(CH2CH2PPh»)2 Tris-(2-diphenylphosphinoethyl)phosphine, P(CH2CHPPh2)3 [(Ph3P)2N]* propyl (superscripts, m or i) pyridine pyrazolyl Tris-(2-diphenylarsinophrnyl)arsine, As(o-CsH4AsPhy)3 Tris-(2-diphenylphosphinopheny!) phosphine, P(o-CgH4PPha)3 alkyl (preferably) or aryl group Perfluoro alkyl group solvent salicylaldehyde bis-salicylaldehydeethylenediimineTAN TAP TAS TCNE TCNQ terpy TFA THF or thf TMED tn TPN (= np3) TPP tren trien triphos TSN TSP TSeP TTA x ABBREVIATIONS IN COMMON USE xv Tris-(2-diphenylarsinoethyl)amine, N(CH2CH2AsPh))3 Tris-(3)dimethylarsinopropyl)phosphine, P(CH2CH2CH>AsMez)3 Bis-(3-dimethylarsinopropyl)methylarsine, * MeAs(CH2CH2CH2AsMe2)2 tetracyanoethylene 7,7,8,8-tetracyanoquinodimethane terpyridine trifluoroacetic acid tetrahydrofuran N,N,N’,N’-tetramethylethylenediamine 1,3-diaminopropane(trimethylenediamine) Tris-(2-diphenylphosphinoethyl)amine, N(CH2CH2PPh2)3 meso-tetraphenylporphyrin Tris-(2-aminoethyl)amine, N(CH2CH2NH2)3 Triethylenetetraamine, (CH2NHCHCH,NH?)> 1,1,1-tris(diphenylphosphinomethyl)ethane Tris-(2-methylthiomethy!)amine, N(CH»CH2SMe)3 Tris-(2-methylthiophenyl)phosphine, P(o-CsH4SMe)3 Tris-(2-methylselenophenyl)phosphine, P(o-CsH4SeMe)3 thenoyltrifluoroacetone, Cs H3SCPCH2CPCF3 halogen or pseudohalogen 2. Miscellaneous A AOM asym bee BM b.p. cep CFSE CFT CIDNP cm! CT d d- ESCA esr or epr eV Ft & Angstrom unit, 107! m angular overlap model asymmetric or antisymmetric body centered cubic Bohr magneton boiling point cubic close packed crystal field stabilization energy crystal field theory chemically induced dynamic nuclear polarization wave number charge transfer decomposes dextorotatory electron spectroscopy for chemical analysis electron spin (or paramagnetic) resonance electron volt Fourier transform (for nmr) g-valuesxvi (g) A hep HOMO Hz ICCC ir TUPAC Ie (O} LCAO. LFSE LFT LUMO MO MOSE mp. nmr P.E. R (s) SCE SCF SCF-Xa-SW Sp or spy str sub sym tbp SREZT™N SEC a ABBREVIATIONS IN COMMON USE gaseous state Planck’s constant hexagonal close packed highest occupied molecular orbital hertz, sec“! International Coordination Chemistry Conference infrared International Union of Pure and Applied Chemistry levorotatory liquid state linear combination of atomic orbitals ligand field stabilization energy ligand field theory lowest unoccupied molecular orbital molecular orbital molecular orbital stabilization energy melting point nuclear magnetic resonance photoelectron (spectroscopy) gas constant solid state saturated calomel electrode self consistent field self-consistent field, Xa, scattered wave (form of MO theory) square pyramid(al) vibrational stretching mode sublimes symmetrical trigonal bipyramid(al) lattice energy ultraviolet valence bond atomic number molar extinction coefficient frequency (cm! or Hz) magnetic moment in Bohr magnetons magnetic susceptibility Weiss constantADVANCED INORGANIC CHEMISTRY A Comprehensive Text FOURTH EDITION1 INTRODUCTORY TOPICSCHAPTER ON Nonmolecular Solids 1-1. Introductory Remarks Inorganic chemistry deals with substances having virtually every known type of physical and structural characteristic. It is not, therefore, easy to decide where to begin a book that will, in the end, have had something to say about many of them. There is no uniquely logical or convenient approach. No matter what order of topics is chosen, there are bound to be some carts before horses, since many facets of the subject interrelate in such a way that neither can be considered as simply prereq- uisite to the other. We begin this edition with a discussion of substances that exist in the solid state as extended arrays rather than molecular units. Although there are doubtless some arguments for not starting a book this way, we have several reasons for doing so. There are a great many important substances that are nonmolecular. Most of the elements themselves are nonmolecular. Thus more than half the elements are metals in which close-packed arrays of atoms are held together by delocalized electrons, while others, such as carbon, silicon, germanium, red and black phos- phorus, and boron involve infinite networks of more localized bonds. There are also many compounds, such as SiOz and SiC, in which the array is held together by localized heteropolar bonds. The degree of polarity varies, of course, and this class of substances grades off toward the limiting case of the ionic arrays in which there are well-defined ions held together principally by the Coulombic forces between those of opposite charge. There are also solids that consist neither of small, well-defined molecules nor of well-ordered infinite arrays of atoms; examples are the glasses and polymers, which, for reasons of space, are not explicitly discussed here. It is, of course, true that most molecular substances form a crystalline solid phase; but because of the relatively weak intermolecular interactions, crystallinity is usually of little chemical importance, though, of course, of enormous practical significance in that it facili- 34 INTRODUCTORY TOPICS tates the investigation of molecular structures, namely, by X-ray crystallog- raphy. Finally, we note that the area of materials science, which is a marriage of applied chemistry and applied physics, deals to a considerable extent with inorganic ma- terials: semiconductors, superconductors, ceramics, refractory compounds, alloys, and so on. Virtually all these are substances with nonmolecular structures. 1-2. Close Packing of Spheres The packing of spherical atoms or ions in such a way that the greatest number oc- cupy each unit of volume is one of the most fundamental structural patterns of Nature. It is seen in its simplest form in the solid noble gases, where spherical atoms are concerned, in a variety of ionic oxides and halides, where small cations can be considered to occupy interstices in a close-packed array of the larger spherical anions, and in metals where close-packed arrays of metal ions are permeated by a cloud of delocalized electrons binding them together. All close-packed arrangements are built by stacking of close-packed layers of the type shown in Fig. 1-1a; it should be evident that this is the densest packing arrangement in two dimensions. Two such layers may be brought together as shown Fig. 1-1. (a) One close-packed layer of spheres. (b) Two close-packed layers showing how tetrahedral (A, A’) and octahedral (B) interstices are formed,NONMOLECULAR SOLIDS 5 fo) (v) Fig. 1-2. (a) The stacking of close-packed layers in the ABC pat closest packing cep. (b) Another view of the cep pattern, emphasi rn that on repetition, gives cubic its cubic symmetry. in Fig. 1-1b, spheres of one layer resting in the declivities of the other; this is the densest packing arrangement of the two layers. It will be noted that between the two layers there exist interstices of two types: tetrahedral and octahedral. When we come to add a third layer to the two already stacked, two possibilities arise. The third layer can be placed so that its atoms lie directly over those of the first layer, or with a displacement relative to the first layer, as in Fig. 1-2a. These two stacking arrangements may be denoted ABA and ABC, respectively. Each may be continued in an ordered fashion so as to obtain Hexagonal close packing (hcp): ABABAB . Cubic close packing (cep): ABCABCABC. .. It is immediately obvious that the hcp arrangement does indeed have hexagonal symmetry, but the cubic symmetry of what has been designated cep may be less evident. Figure 1-2b provides another perspective, which emphasizes the cubic symmetry; it shows that the close-packed layers lie perpendicular to body diagonals of the cube. Moreover, the cubic unit cell is not primitive but face centered. There are, of course, an infinite number of stacking sequences possible within the definition of close packing, all of them, naturally, having the same packing density. The hep and cep sequences are those of maximum simplicity and symmetry. Some of the more complex sequences are actually encountered in Nature, though far less often than the two just described. In any close-packed arrangement, each atom has twelve nearest neighbors, six surrounding it in its own close-packed layer, three above and three below this layer. In the Acp structure each layer is a plane of symmetry and the set of nearest neighbors of each atom has D3, symmetry. In ccp the set of nearest neighbors has D3q symmetry. (See Section 2-4 for definition of these symbols.) 1-3. Metals It is evident, even on the most casual inspection, that metals have many physical Properties quite different from those of other solid substances. Although there are6 INTRODUCTORY TOPICS Fig. 1-3. A body-centered cubic (cc) structure. individual exceptions to each, the following may be cited as the characteristic properties of metals as a class: (1) high reflectivity, (2) high electrical conductance, decreasing with increasing temperature, (3) high thermal conductance, and (4) mechanical properties such as strength and ductility. An explanation for these Properties, and for their variations from one metal to another, must be derived from the structural and electronic nature of the metal. Metal Structures. Almost all metal phases have one of three basic structures, or some slight variation thereof, although there are a few exceptional structures that need not concern us here. The three basic structures are cubic and hexagonal close-packed, which have already been presented in Section 1-2, and body-centered cubic (bcc), illustrated in Fig. 1-3. In the bcc type of packing each atom has only eight instead of twelve nearest neighbors, although there are six next nearest neighbors that are only about 15% farther away. It is only 92% as dense an ar- rangement as the hcp and ccp structures. The distribution of these three structure types, hep, ccp, and bec, in the Periodic Table, is shown in Fig. 1-4. The majority of the metals deviate slightly from the ideal structures, especially those with hep structures. For the Acp structure the ideal value of c/a, where c and a are the hex- agonal unit-cell edges, is 1.633, whereas all metals having this structure have a smaller c/a ratio (usually 1.57~1.62); zinc and cadmium, for which c/a values are 1.86 and 1.89, respectively, are exceptions. Although such deviations cannot in general be predicted, their occurrence is not particularly surprising, since for a given atom its six in-plane neighbors are not symmetrically equivalent to the set of six lying above and below it, and there is consequently no reason for its bonding to those in the two nonequivalent sets to be precisely the same. Metallic Bonding. The characteristic physical properties of metals as well as the high coordination numbers (either twelve or eight nearest neighbors plus six more that are not too remote) suggest that the bonding in metals is different from that in other substances. Clearly there is no ionic contribution, and it is also ob- viously impossible to have a fixed set of ordinary covalent bonds between all adjacent pairs of atoms, since there are neither sufficient electrons nor sufficient orbitals. Attempts have been made to treat the problem by invoking an elaborate resonance of electron-pair bonds among all the pairs of nearest neighbor atoms, and this ap- proach has had a certain degree of success. However the main thrust of theoreticalNONMOLECULAR SOLIDS 7 J BOO pel ss! «| | Lo H@ a00 @ es ©) rs | r Fig. 1-4. The occurrence of hexagonal close-packed (ficp), cubic close-packed (ccp), and body-centered (bcc) structures among the elements. Where two or more symbols are used, the largest represents the stable form at 25°C. The symbol labeled Acp/ccp signifies a mixed... ABCABABCAB . .. type of close packing, with overall hexagonal symmetry. [Adapted, with permission, from H. Krebs, Funda- mentals of Inorganic Crystal Chemistry, McGraw-Hill Book Co., 1968.] work on metals is in terms of the band theory, which gives in a very natural way an explanation for the electrical conductance, luster, and other characteristically metallic properties. A detailed explanation of band theory would necessitate a level of mathematical sophistication beyond that appropriate for this book; however the qualitative features are sufficient. Let us imagine a block of metal expanded, without change in the geometric re- lationships between the atoms, by a factor of, say, 10°. The interatomic distances would then all be 10? greater, that is, about 300 to 500 A. Each atom could then be described as a discrete atom with its own set of well-defined atomic orbitals. Now let us suppose the array contracts, so that the orbitals of neighboring atoms begin to overlap, hence to interact with each other. Since so many atoms are involved, this gives rise, at the actual internuclear distances in metals, to sets of states so close together as to form essentially continuous energy bands, as ilfustrated in Fig. 1-5. Spatially these bands are spread through the metal, and the electrons that occupy them are completely delocalized. In the case of sodium, shown in Fig. 1-5, the 3s and 3p bands overlap, Another way to depict energy bands is that shown in Fig. 1-6. Here energy is plotted horizontally and the envelope indicates on the vertical the number of elec-8 INTRODUCTORY TOPICS Energy Internuclear separation —3> Fig. 1-5. Energy bands of sodium as a function of internuclear distance; ro represents the actual equilibrium distance. [Reproduced by permission from J. C. Slater, Introduction 10 Chemical Physics, McGraw-Hill Book Co., 1939.] trons that can be accommodated at each value of the energy. Shading is used to indicate filling of the bands. Completely filled or completely empty bands (Fig. 1-6a) do not permit net clectron flow, and the substance is an insulator. Covalent solids can be discussed from this point of view (though it is unnecessary to do so) by saying that all electrons occupy low-lying bands (equivalent to the bonding orbitals), while the high-lying bands (equivalent to antibonding orbitals) are entirely empty. Metallic conductance occurs when there is a partially filled band, as in Fig. 1-6b; the transition metals, with their incomplete sets of d electrons, have partially filled d bands; and this accounts for their high conductances. The alkali metals have half-filled s bands formed from their s orbitals, as shown in Fig. 1-5; actually these 5 bands overlap the p bands, and it is because this overlap occurs also for the Ca group metals, where the atoms have filled valence shell s orbitals, that they nevertheless form metallic solids, as indicated in Fig. 1-6c. Cohesive Energies of Metals. The strength of binding among the atoms in metals can conveniently be measured by the enthalpies of atomization. Figure 1-7 plots the energies of atomization of the metallic elements, lithium to bismuth, from their standard states. It is first notable that cohesive energy tends to maximize with el-NONMOLECULAR SOLIDS 9 @ No. of electrons Energy —> > ® No. of electrons Energy —> > © No, of electrons Energy —p> Fig. 1-6. Envelopes of energy bands; shading indicates filling, ements having partially filled d shells (i.e., with the transition metals). However it is particularly with the elements near the middle of the second and third transition series, especially Nb-Ru and Hf-Ir, that the cohesive energies are largest, reaching 837 kJ mol! for tungsten. It is noteworthy that these large cohesive energies are principally duc to the structural nature of the metals whereby high coordination numbers are achieved. For a hcp or ccp structure, there are six bonds per metal atom (since each of the twelve nearest neighbors has a half-share in each of the twelve bonds). Therefore each bond, even when cohesive energy is 800 kJ mol~!, has an energy of only 133 kJ mol-!, roughly half the C—C bond energy in diamond where each carbon atom has only four near neighbors. 1-4. Intersti | Compounds! The term “interstitial compounds” refers primarily (though usage is sometimes more flexible) to combinations of the relatively large transition metal atoms with the small metalloid or nonmetal atoms such as hydrogen, boron, carbon, and ni- trogen. The substances may be thought of as consisting of a metal host lattice with \ L.E. Toth, Transition Metal Carbides and Nitrides, Academic Press, 1971; Refractory Carbides, G. V. Samsonov, Ed., Consultants Bureau, 1974; H. J. Goldschmidt, Interstitial Alloys, Plenum Press, 1967; E. K. Storms, The Refractory Carbides, Academic Press, 1967; B. Aronsson, T. Lundstrom, and S, Rundqvist, Borides, Silicides, and Phosphides, Wiley, 1965; Transition Metal Hydrides, E. L. Muetterties, Ed., Dekker, 1971: H. A. Johansen, Surv. Progr. Chem., 1977, 8, 57.INTRODUCTORY TOPICS fatal ie i i Ta TCT SE 100 Wl ena i i eA i CVT AS a i Me a nent hae a en AT ty 2 HEEL any Hillis Eigen i SH 0 010 11 VET GEL Ata WWESEaUIVSTVVSunI SUN GITUUNTIISUBITUGLUDNNTVTERIDUGUHI SIN STIATTNTH Aitinit By finan is a TT TE sanUai ‘Een Gav uO Ue eT TENT TT A itn 100 ee jaii tis vi vtiy i il it ust VRID BTV eivUR Ve VV edivapuatve nn ur aL i TH in ch ee EUIU TELAT ATV LUAU GATT Tata it 's of atomization of metals AH®29s for M(s) > M(g). [Reproduced by permission from W. E. Dasent, Jnorganic Energetics, Penguin Books, Ltd., 1970.] the small nonmetal atoms occupying the interstices (octahedral or tetrahedral in, c.g., the Aep and cep structures). However the array of metal atoms need not (and usually does not) correspond to any known packing arrangement of the pure metal. One common type of interstitial structure is that obtained by filling every octahedral in a cep structure with a small atom. This arrangement (Fig. 1-8a) is intersti in geometric form to the well-known sodium chloride structure for ionic identicalNONMOLECULAR SOLIDS 11 ‘Tungsten carbide structure CrgCp structure Fig. 1-8. Three structures of interstital compounds; open circles are metal atoms; small filled circles are nonmetal atoms. (a) The NaCI-like structure obtained by filling all octahedral holes in a cep metal array. (b) The WC structure. WN has the same structure, thus showing the unimportance of conven- tional valences. (c) The Cr3C3 structure. materials that we shall discuss later. Figure 1-8b shows a structure in which the metal atoms themselves do not define a close-packed structure. In general the complexity of the structures of interstitial compounds depends on the ratio, R = r,/rm, of the guest atom atomic radius r, to that of the transition metal atom ry. The atomic radii themselves are in most cases equal to one-half the internuclear distances for crystals of the pure elements, though there are ex- ceptions, and small corrections are often made. It has been observed that when R is less than 0.59 the metal atoms form simple structures such as Acp, ccp, bec, or the arrangement shown in Fig. 1-8b. Interstitial compounds with such structures are called Hagg compounds after the Swedish chemist who first noted the radius- ratio rule. When R exceeds 0.59, much more complicated structures, such as that of Cr3C3 (Fig. 1-8c), are adopted.12 INTRODUCTORY TOPICS Not only are these substances rarely interstitial in the sense that the native metal structure is usually not retained, but they are generally not “compounds” either if that term is considered to imply exact stoichiometry. They are usually phases whose structure remains intact over a range of composition. For example, the “compound,” VC actually varies in composition from C/V = 37 to 47%, and VC of exact 1:1 composition does not exist at all. The interstitial compounds are characterized, typically, by great hardness (though also great brittleness) and high melting points, as well as retention of typically metallic characteristics such as luster and good conductivity for heat and electricity. At high temperatures many of them even acquire metallic-type me- chanical properties such as malleability. 1-5. Tonic Crystal Structures Tonic Radii. One of the major factors in determining the structures of the sub- stances that can be thought of, at least approximately, as made up of cations and anions packed together, is ionic size, especially the ratio of radii of the two or more ions present. It is obvious from the nature of wave functions that no ion or atom has a precisely defined radius. The only way radii can be assigned is to determine how closely the centers of two atoms or ions actually approach each other in solid substances and then to assume that such a distance is equal or closely related to the sum of the radii of the two atoms or ions, Even this procedure is potentially full of ambiguities, and further provisions and assumptions are required to get an empirically useful set of radii. The most ambitious attempt to handle this problem is that of Paulin 1g, and some of his arguments and results are briefly summarized here. We begin with the four salts NaF, KCl, RbBr, and CsI, in each of which the cation and anion are isoelectronic and the radius ratios (rcation/Tanion = r+/r-) should be similar in all four cases. Two assumptions are then made: 1. The cation and anion are assumed to be in contact, so that the internuclear distance can be set equal to the sum of the radii. 2. For a given noble gas electron configuration, the radius is assumed to be in- versely proportional to the effective nuclear charge felt by the outer electrons. The implementation of these rules may be illustrated by using NaF, in which the internuclear distance is 2.31 A. Hence rvat trp = 231A Next, using rules developed by Slater to estimate how much the various electrons in the 1s?25?2p® configuration shield the outer electrons from the nuclear charge, we obtain 4.15 for the shielding parameter. The effective nuclear charges, Z, felt by the outer electrons are then, for Nat with Z = 11, 114.15 = 6.85 and for F~, with Z = 9: 9-415 = 4.85NONMOLECULAR SOLIDS 13 According to rule 2, the radius ratio Nat/rp~ must be inversely proportional to these numbers; hence n trat(re = Nel * 685/485 Solving this and the previous equation for the sum of the radii simultaneously we obtain This method, with certain refinements, was used by Pauling to estimate individual ionic radii. Earlier, V. M. Goldschmidt, using a somewhat more empirical method, also estimated ionic radii. The radii for a number of important ions, obtained by ‘0 procedures, are given in Table I-1. A more recent set of most probable ionic ii is also given in Table 1-1. These are based on a combination of shortest in- teratomic distances and experimental electron density maps. Important Ionic Crystal Structures. Figure 1-9 shows six of the most important structures found among essentially ionic substances, In an ionic structure each ion is surrounded by a certain number of ions of the opposite sign; this number is called the coordination number of the ion. In the first three structures shown, namely, the NaCl, CsCl, and CaF> types, the cations have the coordination numbers 6, 8, and 8, respectively. We now ask why a particular compound crystallizes with one or another of these structures. To answer this, we first recognize that ignoring the possibility of me- tastability, which seldom arises, the compound will adopt the arrangement providing the greatest stability, that is, the lowest energy. The factors that contribute to the energy are the attractive force between oppositely charged ions, which will increase with increasing coordination number, and the forces of repulsion, which will increase very rapidly if ions of the same charge are “squeezed” together. Thus the optimum. arrangement in any crystal should be the one allowing the greatest number of op- positely charged ions to “touch” without requiring any squeezing together of ions with the same charge. The ability of a given structure to meet these requirements will depend on the relative sizes of the ions. Let us analyze the situation for the CsCl structure. We place eight negative ions of radius r~ around a positive ion with radius r* so that the M+ to X- distance is r+ +r~ and the adjacent X~ ions are just touching. Then the X~ to X~ distance, a, is given by ee a= Bet trys or Now, if the ratio r—/r* is greater than 1.37, the only way we can have all eight X7 ions touching the M* ion is to squeeze the X~ ions together. Alternatively, if r~/r* is greater than 1.37, and we do not squeeze the X~ ions, they cannot touch the M* ion and a certain amount of electrostatic stabilization energy will be un-14 INTRODUCTORY TOPICS TABLE 1-1 Goldschmidt (G),# Pauling (P)*, and Ladd (L)* Ionic Radii (A) lon G P He Jon G Pp L le 1.54 2.08 139 Poe 1.17 1.21 Fe 1.33 1.36 119 cr 181 181 1.70 Mn2* 0.91 0.80 0.9 Br 1.96 1.95 1.87 Fee 0.83 0.76 09 Ir 2.20 2.16 2.12 Co 0.82 0.74 08 Nit 0.68 0.69 oe 1.32 1.40 1.28 cut 0.72 - so 174 1.84 1.70 Se? 191 1.98 1.81 Bist 0.2 0.20 - Te? 20 2.21 1.97 ABY 0.45 0.50 Sc 0.68 0.81 - Lit 0.78 0.60 0.86 ys 0.90 0.93 Nat 0.98 0.95 112 Lat 1.04 115 kr 1.33 1.33 1.44 Gat 0.60 0.62 - Rot 1.49 1.48 1.58 Int 0.81 0.81 cst 1.65 1.69 1.84 T+ = 0.91 0.98 Cut 0.95 0.96 - Ast 113 1.26 1.27 Fe+ 0.53 oat Aut - 1.37 = cat 0.53 _ Tit 1.49 1.40 134 NH — 1.48 1.66 cH ous 0.15 - sit 0.38 0.41 Bet 0.34 0.31 = TH 0.60 0.68 Mgt 0.78 0.65 0.87 Zr} 0.77 0.80 Cat 1.06 0.99 118 Ce 0.87 1.01 - srt 127 113 1.32 Get 0.54 0.53 ~ Bat 1.43 135 1.49 Sat 0.71 071 Rat i 1.40 1.57 Pot 0.81 0.84 - Zn2* 0.69 0.74 = cat 1.03 0.97 1.14 Hg?+ 0.93 1.10 - # These radii are obtained by using the rock-salt type of structure as standard (i, small corrections can be made for other coordination numbers; see A. P. Sinha, Struct. Bonding, 15 25, 69. For effective ionic radii of M+ in corundum-type oxides of Al, Cr, Ga, V, Fe, Rh, Ti, In, Ti, see C. T. Prewitt et al., Inorg. Chem., 1969, 9, 1985, and for M** in rutile or closely related oxi of Si, Ge, Ma, Cr, V, Rh, Ti, Ru, Ir, Pt, Re, Os, Te, Mo, W, Ta, Nb, Sn, and Pb, see D. B. Roger al., Inorg. Chem., 1969, 8, 841. »M. F.C. Ladd, Theor. Chim. Acta, 1968, 12, 333. attainable. Thus when r~/r* becomes equal to 1.37, the competition between tractive and repulsive Coulomb forces is balanced, and any increase in the ré may make the CsCl structure unfavorable relative to a structure with a lo coordination number, such as the NaCl structure. In the NaCI structure, in order to have all ions just touching but not squeez with radius r~ for X~ and r+ for M* we have arr = Art +H)NONMOLECULAR SOLIDS 15 Cesium chloride'(CsCL} Zine blende (cubic ZnS} ‘Wurtzite (hexogana! 208) Fig. 1-9. Six important ionic structures. Small circles denote metal cations, large circles denote anions. which gives for the critical radius ratio If the ratio -~/r+ exceeds 2.44, the NaC] structure becomes disfavored, and a structure with cation coordination number 4, for which the critical value of epee is 4.44, may be more favorable. To summarize, in this simple approximation,16 INTRODUCTORY TOPICS packing considerations lead us to expect the various structures to have the following ranges of stability in terms of the r~/r* ratio: CsCl and CaF? structures. 1 > r* a structure such as that of ZnS will be preferred. The more common ionic crystal structures shown in Fig. 1-9 are mentioned re: peatedly throughout the text. The rutile structure, named after one mineralogica form of TiOz, is very common among oxides and fluorides of the MF, and MO: types (e.g., FeF2, NiF2, ZrO2, RuO2), where the radius ratio favors coordination number 6 for the cation. Similarly, the zinc blende and wurtzite structures, name after two forms of zinc sulfide, are widely encountered when the radius ratio favor four-coordination, and the fluorite structure is common when eight-coordinatios of the cation is favored. When a compound has stoichiometry and ion distribution opposite to that in on of the structures just mentioned, it may be said to have an anti structure. Thu compounds such as Li.O, Na2S, and KS, have the antifluorite structure in whic! the anions occupy the Ca2* positions and the cations the F~ positions of the CaF structure. The antirutile structure is sometimes encountered also. Structures with Close Packing of Anions. Many structures of halides and oxide can be regarded as close-packed arrays of anions with cations in the octahedre and/or tetrahedral interstices. Even the NaCl structure can be thought of in thi way (cep array of Cl~ ions with all octahedral interstices filled), although this i not ordinarily useful. CdCl; also has ccp Cl~ ions with every other octahedral hol occupied by Cd?+, and Cdl, has hep I~ ions with Cd?+ ions in half the octahedré holes. It is noteworthy that the CdCl; and Cdl structures (the latter appears i Fig. 1-10) are layer structures. The particular pattern in which cations occupy ha the octahedral holes, is such as to leave alternate layers of direct anion-anion cot tact. Corundum, the « form of AlOs, has an hep array of oxide ions with two-thir of the octahedral interstices occupied by cations and is adopted by many other oxid (e.g., Ti03, V203, Cr2O3, Fe203, Ga2Os, and Rh2Os). The Bil; structure has < hep array of anions with two-thirds of the octahedral holes in each alternate pa of layers occupied by cations, and it is adopted by FeCl, CrBr3, TiCl3, VCls, ar many other AB; compounds. As indicated, all the structures just mentioned a adopted by numerous substances. The structures are usually named in referen to one of these substances. Thus we speak of the NaCl, CdCl, Cdl2, BIs, and rundum (or a-Al,O3) structures. ‘Some Mixed Oxide Structures. There area vast number of oxides (and also sor stoichiometrically related halides) having two or more different kinds of catio Most of them occur in one of a few basic structural types, the names of which aNONMOLECULAR SOLIDS 17 re numa ipa i nba t i a oa Hagan Hutu ETE Fig. 1-10. A portion of the Cdl; structure. Smail spheres represent metal cations. derived from the first or principal compound shown to have that type of structure. Three of the most important such structures-are now described. 1. The Spinel Structure. ‘The compound MgAl,O,, which occurs in Nature as the mineral spinel, has a structure based on a cep array of oxide ions. One-cighth of the-tetrahedral holes (of which there are two Per anion) are occupied by Mg?* ions and one-half of the octahedral holes (of which there is one per anion) are oc- cupied by Al>+ ions. This structure, or a modification to be discussed below, is adopted by many other mixed metal oxides of the type MTM3!0, (c.g., FeCrOy, ZnAl,04, and CollCo}!'O,), by some of the type MIVM¥O, (e.g., TiZnyO4 and SnCo;0,), and by some of the type MJMV¥!0, (e.g., NayMoO, and Ag>MoO,). This structure is often symbolized as A[B3]4, where brackets enclose the ions in the octahedral interstices. An important variant is the inverse spinel structure, B[AB]Og, in which half the B ions are in tetrahedral interstices and the A ions are in octahedral ones along with the other half of the B ions. This often happens when the A ions have a stronger preference for octahedral coordination than do the B ions. As far as is known, all M!VM¥O, spinels are inverse {c.g., Zn[ZaTiJOy), as are many of the M™M3""O, ones (e.g., Fel [CoMFell"]Q,, Fell FellFell]Q,, and Fe[NiFe]0,), as well. There are also many compounds with disordered. spinel structures in which only a fraction of the A ions are in tetrahedral sites (and a corresponding fraction in octahedral ones). This occurs when the preferences of both A and B ions for octa- hedral and tetrahedral sites do not differ markedly. 2. The Hmenite Structure. This is the structure of the mineral ilmenite,18 INTRODUCTORY TOPICS © Small cation © Lorge cation © Oxide or halide ion Fig. 1-11. The perovskite structure. Fe!!Ti!¥O3. It is closely related to the corundum structure except that the cations are of two different types. It is adopted by ABO3 oxides when the two cations, A and B, are of about the same size, but they need not be of the same charge so long as their total charge is +6. Thus in ilmenite itself and in MgTiO3 and CoTiO; the cations have charges +2 and +4, whereas in a-NaSbO; the cations have charges of +1 and +5. 3. The Perovskite Structure. The mineral perovskite, CaTiOs, has a structure in which the oxide ions and the large cation (Ca?*) form a cep array with the smaller cation (Ti4+) occupying these octahedral holes formed exclusively by oxide ions, as shown in Fig. 1-11. This structure is often slightly distorted—in CaTiO3 itself, for example. It is adopted by a great many ABO3 oxides in which one cation is comparable in size to O2- and the other much smaller, with the cation charges variable so long as their sum is +6. It is found in Sr#!Ti!¥O3, Bal!TiVO3, La!'Ga!llQ, and Na!NbVOs, and KINbYO, and also in some mixed fluorides (e.g., KZnF3 and KNiF3). 1-6. Energetics of fonic Crystals One may define a model in which so-called ionic crystals are treated as purely ionic—that is, only Coulombic forces plus some higher-order repulsive forces arising at close distances, because of the interpenetration of electron clouds, are assumed to be operating. It is then possible to correlate a great deal of thermodynamic data, but, as will be shown later, it is certain that this model is a considerable oversim- plification and its success should not be taken as proof of its physical validity. Before developing this ionic model, a few points concerning the formulation of ions in the gas phase must be reviewed. To form cations, electrons must be detached from atoms; this process is characterized by a standard enthalpy AF inn. Similarly, to form an anion an electron must be attached to a neutral atom; this process also has a standard enthalpy AHfa, the enthalpy of electron attachment.* For use in subsequent calculations, these quantities are listed in Appendices 2 and 3. * Much of the chemical literature uses /, the ionization potential, in electron volts, which is equal to AH, if converted to kilojoules, and A, the so-called electron affinity, which is equal to ~AH, if put in the same units. We shall not use these nonstandard (though commonplace) quantities,NONMOLECULAR SOLIDS 19 TABLE 1-2 Madelung Constants for Several Structures Structure type M NaCl 1.74756 CsCl 1.76267 CaF, 5.03878 Zinc blende 1.63805 Wurtzite 1.64132 The enthalpy of vaporizing one mole of a crystalline, ionic compound to form an infinitely dilute gas consisting solely of the constituent ions has, of course, a positive sign. Traditionally this enthalpy has been designated the lattice energy; although this term lacks precision, we shall use it with the understanding that when precise definition is required it means the enthalpy of the process just described. The purely ionic description of “ionic” crystals, as stated above, allows an ab initio calculation of the lattice energy of such crystals, provided the structure of the crystal is known. Let us consider as an example sodium chloride, which has the structure shown in Fig. 1-9. The energy, in joules, required to separate two opposite charges, te and —e, each in coulombs (C) initially at a distance ro, in meters, toa distance of infinity is given by the equation e reoro The factor (47r€9)~' is required in the SI system of units; €9 is the permittivity (dielectric constant) of a vacuum that has the value 8.854 X 10-!2 C2 m7! J-!. Let us call the shortest Na+—CI~ distance in NaCl ro. Each Nat ion is sur- rounded by six Cl~ ions at the distance ro, in meters, giving an energy term 6e2/ 4réoro. The next closest neighbors to a given Na* ion are 12 Na* ions which, by simple trigonometry, lie V2r9 away. Thus another energy term, with a minus sign because it is repulsive, is —12e?/~/2ro4ze9. By repeating this sort of procedure, successive terms are found, leading to the expression 1_(6c2_ 120? | Be? _ be? o-oo) pers |g Ze Baan tara ) Frere V2" V3 2 V5 It is possible to derive a general formula for the infinite series and to find the numerical value to which it converges. That value is characteristic of the structure and independent of what particular ions are present. It is called the Madelung constant, Myaci, for the NaCl structure. It is actually an irrational number, whose value can be given to as high a degree of accuracy as needed, for example, 1.747 +, or 1.747558 ..., or better. Madelung constants for many common ionic structures have been evaluated, and a few are given in Table 1-2 for illustrative purposes. (ly)20 INTRODUCTORY TOPICS A unique Madelung constant is defined only for structures in which all ratios of interatomic vectors are fixed by symmetry. In the case of the rutile structure there are two crystal dimensions that can vary independently. There is a different Madelung constant for each ratio of the two independent dimensions. When a mole (N ions of each kind, where N is Avogadro's number) of sodium chloride is formed from the gaseous ions, the total electrostatic energy released is given by 2 E, = Natsu (1-2) a This is true because the expression for the electrostatic energy of one Cl- ion would be the same as that for an Na* ion. If we were to add the electrostatic energies for the two kinds of ion, the result would be twice the true electrostatic energy because each pairwise interaction would have been counted twice. The electrostatic energy given by eq. 1-2 is not the actual energy released in the process Na*(g) + Cl-(g) = NaCl(s) (1-3) Real ions are not rigid spheres. The equilibrium separation of Na*+ and Cl~ in NaCl is fixed when the attractive forces are exactly balanced by repulsive forces. The attractive forces are Coulombic and follow strictly a 1/r? law. The repulsive forces are more subtle and follow an inverse r” law, where n is >2 and varies with the nature of the particular ions. We can write, in a general way, that the total repulsive’ energy per mole at any value of r is Exp = MB where B is a constant. At the equilibrium distance, the net energy U for process 1-3 (where we now use algebraic signs in accord with convention), is given by NB =-NM, +B a eeeet bs 5 (1-4) Observe that the attractive forces produce an exothermic contribution and the repulsive ones an endothermic term. The constant B can now be eliminated if we recognize that at equilibrium, when r = ro the energy U is a minimum by definition. The derivative of U with respect tor, evaluated at r = ro must equal zero. Differentiating eq. 1-4 we get eu = Miosce- nNB 4 =r Areoro? 8 which can be rearranged and solved for B: eM NaCl pp EM NaCl py ‘Freon When this expression for B is substituted into eq. 1-4, we obtain 2 y= —NMacie? (-3 (1-5) 4reoro nNONMOLECULAR SOLIDS 21 The value of n can be estimated as 9.1 from the measured compressibility* of NaCl. In a form suitable for calculating numerical results, in kJ mol-!, using ro in angstroms, eq. 1-5 becomes Ua m1399 Money -}) en and inserting appropriate values of the parameters we obtain ep el aaa U = -1389 5 (I -3y U = -860 + 95 = —765 kJ mol“! Notice that the repulsive energy equals about 11% of the Coulombic energy. The net result is not very sensitive to the exact value of n. If a value of n = 10 had been used, an error of only 9 kJ mol™! or 1.2% would have been made. Refinement of Lattice Energy Calculations. As noted, the Madelung constant is determined by the geometry of the structure only. Thus for a case, such as MgO, where the structure is the same but each ion has a charge of +2, the only modifi- cation required is to replace ~e? by (2e)(—2e) = —4e? in the Coulombic energy term. In general, eq. I-5 becomes u= ~ Mmac2e ( ies 1 4xeoro ni for any structure whose Madelung constant is M with ions of charges Z+ and Zz. The value of m can be estimated for alkali halides by using the average of the following numbers: He 5 Kr 10 Ne 7 Xe 12 Ar 9 where the noble gas symbol denotes the noble-gas-like electron configuration of the ion. Thus for LiF an average of the He and Ne values (5 + 7)/2 = 6 would be used. To make very accurate calculations of crystal energies, sometimes called lattice energies, certain refinements must be introduced. The main ones are as follows. |. A more accurate, quantum expression for the repulsion energy. 2. A correction for van der Waals energy. 3. A correction for the “zero-point energy,” the vibrational energy present even atO° K. The last two corrections are opposite in sign and often of similar magnitude. Thus for NaCl a refined calculation gives Coulomb energy 860 Repulsion energy +99 van der Waals energy -2B Zero-point energy +8 =766 kJ mol~! * Fractional change in volume per unit change in pressure, that is, (AV/V)/AP.Enthalpy, KJ mol7! 22 INTRODUCTORY TOPICS — Crate aa Lio iar brane a ut i i eel eh Nene i eel Burne Lee Tiina Lit, i a Lane nt ney i aa Cre ur ire i reel ee 200 Ln Fig, 1-12. The Born-Haber cycle for NaCl. The Born-Haber Cycle, One test of whether an ionic model is a useful description of a substance such as sodium chloride is its ability to produce an accurate value of the enthalpy of formation. Note that we cannot make this test simply by mea- suring the enthalpy of reaction 1-3, or its reverse. The former is possible in principle but not experimentally feasible. The latter is not possible because sodium chloride does not vaporize cleanly to Na* and Cl but to NaCl, which then dissociates into atoms. To handle the energy problem, we use a thermodynamic cycle called the Born- Haber cycle (Fig. 1-12). The basic idea is that the formation of NaCl(s) from the elements, Na(s) + ¥Clo(g), whose enthalpy is by definition the enthalpy of for- mation of Na€l(s), can be broken down into a series of steps. If the enthalpies of these steps.are added, algebraically, the result must be equal to AH; according to the law of conservation of energy, the first law of thermodynamics. We thus have the equation AHS = AHay + Vp AM Gig + Allen + Atfion + U. (1-6) where the enthalpy terms are for vaporization of sodium (A#\ap), dissociation of Chg). into gaseous atoms (AM{j,), electron attachment to Cl(g) to giveNONMOLECULAR SOLIDS 23 Cl-(g)(AH ea), ionization of Na(g) to Na*(g) + e(AHinq), and the formation of NaCl(s) from gaseous ions (U). More generally, any one of these energies can be calculated if all the others are known. For NaCl all the enthalpies except U in eq. 1-6 have been measured inde- pendently. The following summation can thus be made: AH; -411 jAton S =108 —h AH =, —121 —AHEa 354 SAHinn =502 U = —788k).mol! This result is very close (within 2.7%) of the value of U calculated by using the re- fined ionic model (—766 kJ mol~!). Actually, even more precise calculations give a value that agrees to within less than'1%. This good agreement supports (but does not prove) the idea that the ionic model for NaCl is a useful one. Often, the Born-Haber cycle, or a similar one, is used differently. If it is assumed that U calculated on the ionic model is correct, the cycle can be used to estimate some other energy term. For example, there is no convenient direct way to measure’ the enthalpy of formation of the gaseous CN ion. From a Born-Haber cycle for Nan, where values of all the other enthalpies are available and U is calculated, it is found that AHy for CN~(g) is 29 kJ mol™!. Caveat. It is important to realize that because the purely ionic model of some: real compound such as NaCl affords a fairly accurate value for its lattice energy, this does not entitle us to conclude that the compound is literally ionic, Very refined X-ray diffraction studies indicate that the electron distribution does not correspond strictly to the requirements of Nat (10 electrons) and Clr (eighteen electrons), but shows somewhat greater density on Na and less on Cl. This is tantamount to the existence of some shared electron density, or, in other words, some covalence in the bonding. The purely ionic model is successful energetically mainly because the use of the experimental internuclear distances compensates for the assumption of exactly integral formal charges, +1, +2, and so on. Although we shall not go further into the matter here, the purely ionic model does not give structural pre- dictions that are always correct; that is, the rules developed in Section 1-5 on the basis of fixed radius ratios and assuming purely electrostatic forces are frequently violated. Thus the purely ionic model is just that—a model. It is convenient but not literally true. 1-7. Covalent Solids Elements. The elements that form extended covalent (as opposed to metallic) arrays are boron, all the Group IV elements except lead, also phosphorus, arsenic, selenium, and tellurium. All other elements form either only metallic phases or only molecular ones. Some of the elements above, of course, have allotropes of metallic or molecular type in addition to the phase or phases that are extended covalent arrays. For example, tin has a metallic allotrope (white tin) in addition to that with24 INTRODUCTORY TOPICS Vig, 1-13, The diamond structure seen from two points of view. (a) The conventional cubic unit cell (b) A view showing how layers are stacked; these layers run perpendicular to the body diagonals of the cube. the diamond structure (gray tin), and selenium forms two molecular allotropes containing Seg rings, isostructural with the rhombic form and the monoclinic form of sulfur. For tellurium we have a situation on the borderline of metallic be- havior. The structures of the principal allotropic forms of all the elements are discussed in detail as the chemistry of each element is treated. For illustrative purposes, we shall mention here only one such structure, the diamond structure, since this is adopted by several other elements and is a point of reference for various other structures. It is shown from two points of view in Fig. 1-13. The structure has a cubic unit cell with the full symmetry of the group Ty. However for some purposes it can be viewed as a stacking of puckered infinite layers. It will be noted that the zinc blende structure (Fig. 1-9) can be regarded as a diamond structure in which one-half the sites are occupied by Zn?* (or other cation) and the other half are occupied by S?> (or other anion) in an ordered way. In the diamond structure itself all atoms are equivalent, each being surrounded by a perfect tetrahedron of four others. The electronic structure can be simply and fairly accurately described by saying that each atom forms a localized two-electron bond to each of its neighbors. Compounds. As soon as one changes from elements, where the adjacent atoms are identical and the bonds are necessarily nonpolar, to compounds, there enters the vexatious question of when to describe a substance as ionic and when to describe it as covalent, No attempt is made here to deal with this question in detail for the practical reason that, very largely, there is no need to have the answer—even granting, for the sake of argument only, that any such thing as “the answer” exists. Suffice it to say that bonds between unlike atoms all have some degree of polarity and (1) when the polarity is relatively small it is practical to describe the bonds asNONMOLECULAR SOLIDS 25 polar covalent ones and (2) when the polarity is very high it makes more sense to consider that the substance consists of an array of ions. 1-8. Defect Structures All the foregoing discussion of crystalline solids has dealt with their perfect or ideal structures. Such perfect structures are seldom if ever found in real substances, and although low levels of imperfections have only small effects on their chemistry, the physical (i.¢., electrical, magnetic, optical, and mechanical) properties of many substances are often crucially affected by their imperfections. It is, therefore, ap- propriate to devote a few paragraphs to describing the main types of imperfection, or defect, in real crystalline solids. We shall not, however, discuss the purely me- chanical imperfections such as mosaic structure, stacking faults, and dislocations, all of which entail some sort of mismatch between lattice layers. Stoichiometric Defects. There are some defects that leave the stoichiometry unaffected. One type is the Schottky defect in ionic crystals. A Schottky defect consists of vacant cation and anion sites in numbers proportional to the stoichi- ometry; thus there are equal numbers of Na* and Cl~ vacancies in NaCl and 2CI~ vacancies per Ca?* vacancy in CaCl. A Schottky defect in NaCl is illustrated in Fig. 1-14a. When an ion occupies a normally vacant interstitial site, leaving its proper site vacant, the defect is termed a Frenkel defect. Frenkel defects are most common in crystals where the cation is much smaller than the anion, for instance, in AgBr, as illustrated schematically in Fig. 1-14. Nonstoichiometric Defects. These often occur in transition metal compounds, especially oxides and sulfides, because of the ability of the metal to exist in more than one oxidation state. A well-known case is “FeO,” which consists of a cep array of oxide ions with all octahedral holes filled by Fe2* ions. In reality, however, some of these sites are vacant, whereas others—sufficient to maintain electroneutral- ity—contain Fe?+ ions. Thus the actual stoichiometry is commonly about Feo.9sO. Another good example is “TiO,” which can readily be obtained with compositions ranging from Tio.74O to Ti, 67O depending on the pressure of oxygen gas used in preparing the material. Nonstoichiometric defects also occur, however, even when the metal ion has but one oxidation state. Thus, for example, CdO is particularly liable to lose oxygen when heated, to give yellow to black solids of composition Cd; 440. A comparable situation arises when NaCl is treated with sodium vapor, which it absorbs to give a blue solid of composition Naj+,Cl. An appropriate number (Vy) of anion sites are then devoid of anions but occupied by electrons. The electrons in these cavities behave roughly like the simple particle in a box, and there are excited states ac- cessible at energies corresponding to the energy range of visible light. Hence these cavities containing electrons are color centers, commonly called F centers, from Farbe (German for color). The existence of defects has a simple thermodynamic basis. The creation of a defect in a perfect structure has an unfavorable effect on the enthalpy; some26 INTRODUCTORY TOPICS Displaced Aneel Nes Vacant ° cr coven She Br gt (o) Fig. 1-14. (a) A Schottky defect in a layer of NaCl. (b) A Frenkel defect in a layer of AgBr. The defects need not be restricted to one layer, as shown. Coulombic or bonding energy must be sacrificed to create it. However the intro- duction of some irregularities into an initially perfect array markedly increases the entropy; a TAS term large enough to cancel the unfavorable AH term will thus arise up to some limiting concentration of defects. It is possible to write an expression for the concentration of defects in equilibrium with the remainder of the structure just as though a normal chemical equilibrium were involved. Finally, there are defects resulting from the presence of impurities. Some of these, when deliberately devised and controlled, constitute the basis for solid-state elec- tronic technology. For example, a crystal of germanium (which has the diamond structure) can be “doped” with traces of gallium or arsenic. A gallium atom can replace an atom of germanium, but an electron vacancy is created. An electron can move into this hole, thus creating a hole elsewhere. In effect, the hole can wander through the crystal and under the influence of an electrical potential difference, it can travel through the crystal in a desired direction. The hole moves in the same direction as would a positive charge, and gallium-doped germanium is therefore called a p-type (for positive) semiconductor. Whenever an arsenic atom replaces a germanium atom, an electron is introduced into a normally unfilled energy band of germanium. These electrons can also migrate in an electric field, and the arse- nic-doped material is called an n-type (for negative) semiconductor. Doped silicon and germanium are technologically the most important types of semiconducting material. To ensure reproducible performance, the type and level of impurities must be strictly controlled; thus superpure silicon and germanium must first be prepared (cf. Chapter 12) and the desired doping then carried out. In addition to silicon or germanium as the basis for creating semiconductors, it is also possible to prepare certain isoelectronic III-V or II-VI compounds, such as GaAs or CdS. Holes or conduction electrons can then be introduced by variation of the stoichiometry or by addition of suitable impurities.NONMOLECULAR SOLIDS 27 Semiconductor behavior is to be found in many other types of compound, for example, in Fe,-,O and Fe,_,S, where electron transfer from Fe? to Fe?+ causes a virtual migration of Fe** ions, hence p-type conduction. General References Addison, W. E., Allotropy of the Elements, Oldbourne Press, 1966. Dasent, W. E., Inorganic Energetics, Penguin Books, 1970. Excellent, readable coverage of basic principles. Donahue, J., The Structures of the Elements, Wiley, 1974. Galasso, F.S., Structure and Properties of Inorganic Solids, Pergamon Press, 1970. Greenwood, N. N., onic Crystals, Lattice Defects and Non-Stoichiometry, Butterworths, 1968. Ex- cellent short introduction, Hannay, N. B., Solid-State Chemistry, Prentice-Hall, 1967. Good general survey. Johnson, D. A., Some Thermodynamic Aspects of Inorganic Chemistry, Cambridge University Press, 1968. A good outline of fundamentals. Krebs, H., Fundamentals of Inorganic Crystal Chemistry, McGraw-Hill, 1968. Excellent discussion of structures and bonding. Kroger, F. A., Chemistry of Imperfect Crystals, North Holland, 1964, Ladd, M. F. C., Structure and Bonding in Solid State Chemistry, Horwood-Wiley, 1979. McDowell, C. A., in Physical Chemistry, Vol. Il, H. Eyring, D. Henderson, and W. Jost, Eds., Aca- demic Press, 1969, p. 496. Evaluation of electron affinities. Pearson, W. B., The Crystal Chemistry and Physics of Metals and Alloys, Wiley, 1972 Vedeneyev, V. I., et al., Bond Energies, lonization Potentials, and Electron Affinities, Arnold, 1965. Data tables. Wells, A. F., Structural Inorganic Chemistry, 4th ed., Clarendon Press, 1975, A fascinating book with superb illustrations and a wealth of data; 1068 pages of text.CHAPTER TWO Symmetry and Structure THE SYMMETRY GROUPS Molecular symmetry and ways of specifying it with mathematical precision are important for several reasons. The most basic reason is that al! molecular wave functions—those governing electron distribution as well as those for vibrations, nmr spectra, and so on—must conform, rigorously, to certain requirements based ‘on the symmetry of the equilibrium nuclear framework of the molecule. When the symmetry is high these restrictions can be very severe. Thus from a knowledge of symmetry alone it is often possible to reach useful qualitative conclusions about molecular electronic structure and to draw inferences from spectra about molecular structures. The qualitative application of symmetry restrictions is most impressively illustrated by the crystal-field and ligand-field theories of the electronic structures of transition metal complexes, as described in Chapter 20, and by numerous ex- amples of the use of infrared and Raman spectra to deduce molecular symmetry. Illustrations of the latter occur throughout the book, but particularly with respect to’some metal carbonyl compounds in Chapter 25. A more mundane use for the concept and notation of molecular symmetry is in the precise description of a structure. One symbol, such as Dan, can convey precise, unequivocal structural information that would require long verbal description to duplicate. Thus if we say that the Ni(CN)3" ion has Da, symmetry, we imply that (a) it is completely planar, (b) the Ni—C—N groups are all linear, (c) the C— Ni—C angles are all equal, at 90°, (d) the four CN groups are precisely equivalent to one another, and (e) the four Ni—C bonds are precisely equivalent to one an- other. The use of symmetry symbols has become increasingly common in the chemical literature, and it is now necessary to be familiar with the basic concepts and rules of notation to read many of the contemporary research papers in inorganic and, indeed, also organic chemistry with full comprehension. It thus seems ap- propriate to include a brief survey of molecular symmetry and the basic rules for specifying it. 28SYMMETRY AND STRUCTURE 29 2-1, Symmetry Operations and Elements When we say that a molecule has symmetry, we mean that certain parts of it can be interchanged with others without altering either the identity or the orientation of the molecule. The interchangeable parts are said to be equivalent to one another by symmetry. Consider, for example, a trigonal-bipyramidal molecule such as PFs (2-1). The three equatorial P—F bonds, to F), F2, and F3, are equivalent. They have ab oF: a i Re ey i. Fy en the same length, the same strength, and the same type of spatial relation to the remainder of the molecule. Any permutation of these three bonds among themselves leads to a molecule indistinguishable from the original. Similarly, the axial P—F bonds, to Fy and Fs, are equivalent. But, axial and equatorial bonds are different types (e.g., they have different lengths), and if one of each were to be interchanged, the molecule would be noticeably perturbed. These statements are probably self- evident, or at least readily acceptable, on an intuitive basis; but for systematic and detailed consideration of symmetry, certain formal tools are needed. The first set of tools is a set of symmetry operations. Symmetry operations are geometrically defined ways of exchanging equivalent parts of a molecule. There are four kinds which are used conventionally and these are sufficient for all our purposes. 1. Simple rotation about an axis passing through the molecule by an angle 2x/n. This operation is called a proper rotation and is symbolized C,,. If it is repeated am s SceanEEEESocee ZN fo) awe s. i 0% Na Fig, 2-1. The operation C2 carries H2S into an orientation indistinguishable from the original, but HSD g0¢s into an observably different orientation.30 INTRODUCTORY TOPICS Fig. 2-2. The effects of symmetry operations on an arbitrary point, designated 0, thus generating sets, of points. 1n times, of course the molecule comes all the way back to the original orienta- tion. 2. Reflection of all atoms through a plane that passes through the molecule. This operation is called reflection and is symbolized o. 3. Reflection of all atoms through a point in the molecule. This operation is called inversion and is symbolized i. 4. The combination, in either order, of rotating the molecule about an axis passing through it by 27/n and reflecting all atoms through a plane that is per- pendicular to the axis of rotation is called improper rotation and is symbolized Sn.SYMMETRY AND STRUCTURE 31 These operations are symmetry operations if, and only if, the appearance of the molecule is exactly the same after one of them is carried out as it was before. For instance, consider rotation of the molecule H2S by 27/2 about an axis passing through S and bisecting the line between the H atoms. As shown in Fig. 2-1, this operation interchanges the H atoms and interchanges the S—H bonds. Since these atoms and bonds are equivalent, there is no physical (i.e., physically meaningful or detectable) difference after the operation. For HSD, however, the corresponding operation replaces the S—H bond by the S—D bond, and vice versa, and one can see that a change has occurred. Therefore, for HS, the operation C; is a symmetry operation; for HSD it is not. These types of symmetry operation are graphically explained by the diagrams in Fig. 2-2, where it is shown how an arbitrary point (0) in space is affected in each case. Filled dots represent points above the xy plane and open dots represent points below it. Let us examine first the action of proper rotations, illustrated here by the Cy rotations, that is, rotations by 27/4 = 90°. The operation C, is seen to take the point 0 to the point 1. The application of C, twice, designated C3, generates point 2. Operation C} gives point 3 and, of course, C4, which is a rotation by 4X 2/4 = 2r, regenerates the original point. The set of four points, 0, 1, 2, 3 are permutable, cyclically, by repeated Cy proper rotations and are equivalent points. It will be obvious that in general repetition of a C,, operation will generate a set of n equivalent points from an arbitrary initial point, provided that point lies off the axis of rota- tion. The effect of reflection through symmetry planes perpendicular to the xy plane, specifically, o,, and oy, is also illustrated in Fig. 2-2. The point 0 is related to point 1 by the oy, operation and to the point 3 by the ox, operation. By reflecting either point | or point 3 through the second plane, point 2 is obtained. The set of points generated by the repeated application of an improper rotation will vary in appearance depending on whether the order of the operation, S,, is even or odd, order being the number . A crown of x points, alternately up and down, is produced for even, as illustrated for Sg. For odd there is generated a set of 2n points which form a right n-sided prism, as shown for S3. Finally, the operation i is seen to generate from point 0 a second point, 1, lying on the opposite side of the origin. Let us now illustrate the symmetry operations for various familiar molecules as examples. As this is done it will be convenient to employ also the concept of symmetry elements. A symmetry element is an axis (line), plane, or point about which symmetry operations are performed. The existence of a certain symmetry operation implies the existence of a corresponding symmetry element, and con- versely, the presence of a symmetry element means that a certain symmetry op- eration or set of operations is possible. Consider the ammonia molecule (Fig. 2-3). The three equivalent hydrogen atoms may be exchanged among themselves in two ways: by proper rotations, and by re- flections. The molecule has an axis of 3-fold proper rotation; this is called a C3 axis. It passes through the N atom and through the center of the equilateral triangle defined by the H atoms. When the molecule is rotated by 27/3 in a clockwise di- rection H; replaces H2, He replaces Hs, and H; replaces Hj. Since the three H32 INTRODUCTORY TOPICS fa) Fig, 2-3. The ammonia molecule, showing its 3-fold symmetry axis C3, and one of its three planes of symmetry 1, which passes through Hi and N and bisects the H>—Hs line. atoms are physically indistinguishable, the numbering having no physical reality, the molecule after rotation is indistinguishable from the molecule before rotation. This rotation, called a C3 or 3-fold proper rotation, is a symmetry operation. Rotation by 2 X 27/3 also produces a configuration different, but physically in- distinguishable, from the original and is likewise a symmetry operation; it is des- ignated C3, Finally, rotation by 3 X 22/3 carries each H atom all the way around and returns it to its initial position. This operation, Cj, has the same net effect as performing no operation at all, but for mathematical reasons it must be considered as an operation generated by the C3 axis. This, and other operations which have no net effect, are called identity operations and are symbolized by E. Thus, we may write C} = E. The interchange of hydrogen atoms in NH; by reflections may be carried out in three ways; that is, there are three planes of symmetry. Each plane passes through the N atom and one of the H atoms, and bisects the line connecting the other two H atoms. Reflection through the symmetry plane containing N and H, interchanges H2 and H;; the other two reflections interchange H, with H3, and H, with Hp. Inspection of the NH3 molecule shows that no other symmetry operations besides these six (three rotations, C3, C3, C3 = E, and three reflections, 6), 62, 03) are possible, Put another way, the only symmetry elements the molecule possesses are C3 and the three planes that we may designate 01, 02, and 03. Specifically, it will be obvious that no sort of improper rotation is possible, nor is there a center of symmetry.Fig. 2-4. The symmetry elements of the RezClf” ion. (a) The axes of symmetry. (b) One of each type of plane and the center of symmetry. 3334 INTRODUCTORY TOPICS ‘Asa more complex example, in which all four types of symmetry operation and element are represented, let us take the ReClf” ion, which has the shape of a square parallepiped or right square prism (Fig. 2-4). This ion has altogether six axes of proper rotation, of four different kinds. First, the Re;-Rez line is an axis of four-fold proper rotation, C4, and four operations, C4, Ci, Ci, Ci = E, may be carried out. This same line is also a C2 axis, generating the operation C>. It will be noted that the Cj operation means rotation by 2 X 27/4, which is equivalent to rotation by 2/2, that is, to the C2 operation. Thus the C2 axis and the C2 operation are implied by, not independent of, the C4 axis. There are, however, two other types of C2 axis that exist independently. There are two of the type that passes through the centers of opposite vertical edges of the prism, C, axes, and two more that pass through the centers of opposite vertical faces of the prism, C} axes. The Re2Cl}~ ion has three different kinds of symmetry plane (see Fig. 2-4b). There is a unique one that bisects the Re—Re bond and all the vertical edges of the prism. Since it is customary to define the direction of the highest proper axis of symmetry, Cs in this case, as the vertical direction, this symmetry plane is hor- izontal and the subscript h is used to identify it, o,. There are then two types of vertical symmetry plane, namely, the two that contain opposite vertical edges, and two others that cut the centers of the opposite vertical faces. One of these two sets may be designated a and o®), the v implying that they are vertical. Since those of the second vertical set bisect the dihedral angles between those of the first set, they are then designated o) and o, the d standing for dihedral. Both pairs of planes are vertical and it is actually arbitrary which are labeled o, and which oq. Continuing with ReCl}”, we see that an axis of improper rotation is present. This is coincident with the Cy axis and is an S4 axis. The Sg operation about this axis proceeds as follows. The rotational part, through an angle of 27/4, in the clockwise direction has the same effect as the C4 operation. When this is coupled” with a reflection in the horizontal plane, op, the following shifts of atoms occur: Re Re Ch—=Cle Cs + Chr Rey Re Ch Ch Cle + Cl Ch—+Cle Ch Cy Clu Cls Cle Ch Finally, the RepCl?" ion has a center of symmetry i and the inversion operation ican be performed. In the case of RezCl}~ the improper axis S4 might be considered as merely the inevitable consequence of the existence of the C4 axis and the op, and, indeed,SYMMETRY AND STRUCTURE 35 this is a perfectly correct way to look at it. However it is important to emphasize that there are cases in which an improper axis S,, exists without independent ex- istence of either C, or 6. Consider, for example, a tetrahedral molecule as depicted in 2-II, where the TiCl, molecule is shown inscribed in a cube and Cartesian axes, x,y, and z are indicated. Each of these axes is an S axis, For example rotation by 2/4 about z followed by reflection in the xy plane shifts the Cl atoms as fol- lows: Ch+Ch Ch +Ch Ch+Ck = Cl Ch Note, however, that the Cartesian axes are not Cy axes (though they are C2 axes) and the principal planes (viz., xy, xz, yz) are not symmetry planes. Thus we have here an example of the existence of the S, axis without C, or 64 having any inde- pendent existence. The ethane molecule in its staggered configuration has an S¢ axis and provides another example. 2-2. Symmetry Groups The complete set of symmetry operations that can be performed on a molecule is called the symmetry group for that molecule. The word “group” is used here not as a mere synonym for “set” or “collection,” but in a technical, mathematical sense, and this meaning must first be explained. Introduction to Multiplying Symmetry Operations. We have already seen in passing that if a proper rotation C,, and a horizontal reflection ¢» can be performed, there is also an operation that results from the combination of the two which we call the improper rotation S,. We may say that S, is the product of C, and 4. Noting also that the order in which we perform ;, and C,, is immaterial,* we can write: Cu X On = On X Cy = Sy This is an algebraic way of expressing the fact that successive application of the two operations shown has the same effect as applying the third one. For obvious reasons, it is convenient to speak of the third operation as being the product obtained by multiplication of the other two. The example above is not unusual. Quite generally, any two symmetry operations can be multiplied to give a third. For example, in Fig. 2-2 the effects of reflections in two mutually perpendicular symmetry planes are illustrated. It can be seen that one of the reflections carries point 0 to point 1. The other reflection carries point 1 to point 2. Point 0 can also be taken to point 2 by way of point 3 if the two re- flection operations are performed in the opposite order. But a moment’s thought will show that a direct transfer of point 0 to point 2 can be achieved by a C2 operation about the axis defined by the line of intersection of the two planes. If we call the two reflections o(xz) and o(yz) and the rotation C>(z), we can write (xz) X oz) = ayz) X o(xz) = C2(z) It is also evident that * This is, however, a special case; in general, order of multiplication matters (see p. 38).36 INTRODUCTORY TOPICS a(yz) X C2(z) = C2(z) X a(yz) = o(xz) and (x2) X Cx(z) = Caz) X ofaz) = (v2) It is also worth noting that if any one of these three operations is applied twice in succession, we get no net result or, in other words, an identity operation, namely, o(xz) X o(xz) = E E o(y2) x a(yz Co(z) X Cr(z) Introduction to a Group. If we pause here and review what has just been done with the three operations (xz), o(yz), and C2(z), we see that we have formed all the nine possible products. To summarize the results systematically, we can arrange them in the annexed tabular form. Note that we have added seven more multipli- cations, namely, all those in which the identity operation E is a factor. The results of these are trivial, since the product of any other, nontrivial operation with E must be just the nontrivial operation itself, as indicated. E Cxlz) o(xz) yz) E E x2) o(xz) oz) xz) Cz) E oz) o(xz) a(xz) o(xz) o(yz) E C22) olyz) o(yz) o(xz) Caz) E The set of operations E, C2(z), o(xz), and o(yz) evidently has the following four interesting properties: 1. There is one operation E, the identity, that is the trivial one of making no change. Its product with any other operation is simply the other operation. 2. There is a definition of how to multiply operations: we apply them successively. The product of any two is one of the remaining ones. In other words, this collection of operations is self-sufficient, all its possible products being already within itself. This is sometimes called the property of closure. 3. Each of the operations has an inverse, that is, an operation by which it may be multiplied to give E as the product. In this case, each operation is its own inverse, as shown by the occurrence of E in all diagonal positions of the table. 4. It can also be shown that if we form a triple product, this may be subdivided in any way we like without changing the result, thus o(xz) X o(v2) X Cxlz) = [a(x2) X o(92)] X Ca(2) = Cale) X Cale) = o(xz) X [a(yz) X C2(2)] = o(x2) X o(xy)SYMMETRY AND STRUCTURE 37 Products that have this property are said to obey the associative law of mult cation. The four properties just enumerated are of fundamental importance. They are the properties—and the only properties—that any collection of symmetry opera- tions must have to constitute a mathematical group. Groups consisting of symmetry operations are called symmetry groups or sometimes point groups. The latter term arises because all the operations leave the molecule fixed at a certain point in space. This is in contrast to other groups of symmetry operations, such as those that may be applied to crystal structures in which individual molecules move from one lo- cation to another. The symmetry group we have just been examining is one of the simpler groups; but nonetheless, an important one. It is represented by the symbol C2,; the origin of this and other symbols is discussed below. It is not an entirely representative group in that it has some properties that are not necessarily found in other groups. We have already called attention to one, namely, that each operation in this group is its own inverse; this is actually true of only three kinds of operation: reflections, twofold proper rotations, and inversion i. Another special property of the group Cy is that all multiplications in it are commutative; that is, every multiplication is equal to the multiplication of the same two operations in the opposite order. It can be seen that the group multiplication table is symmetrical about its main di- agonal, which is another way of saying that all possible multiplications commute. In general, multiplication of symmetry operations is nor commutative, as subsequent discussion will illustrate. For another simple, but more general, example of a symmetry group, let us recall our carlier examination of the ammonia molecule. We were able to discover six and only six symmetry operations that could be performed on this molecule. If this is indeed a complete list, they should constitute a group. The easiest way to see if they do is to attempt to write a multiplication table. This will contain 36 products, some of which we already know how to write. Thus we know the result of all mul- tiplications involving E, and we know that GXG=Q Cx C=CXC=E It will be noted that the second of these statements means that C; is the inverse of C} and vice versa. We also know that E and each of the o’s is its own inverse. So all operations have inverses, thus satisfying requirement 3. To continue, we may next consider the products when one 6, is multiplied by another. A typical example is shown in Fig. 2-Sa. When point 0 is reflected first through o(") and then through 0), it becomes point 2. But point 2 can obviously also be reached by a clockwise rotation through 27/3, that is, by the operation C3. Thus we can write: 0 X 92) = If, however, we reflect first through o'2) and then through o), point 0 becomes38 INTRODUCTORY TOPICS 1° o) Fig. 2-5. The mult'plication of symmetry operations. (a) Reflection times reflection. (b) Reflection followed by C3, point 4, which can be reached also by C3 X C3 = C3. Thus we write 2) X oD =C} Clearly the reflections 0) and o“2) do not commute. The reader should be able to make the obvious extension of the geometrical arguments just used to obtain the following additional products: oO) XO = aX a =Cy a) XG) =Cy aX 6 = There remain, now, the products of C3 and C3 with o(), 6), and o). Fig. 2-5b shows a type of geometric construction that yields these products. For example, we can see that the reflection o() followed by the rotation C; carries point 0 to point 2, which could have been reached directly by the operation 02). By similar proce- dures all the remaining products can be easily determined. The complete multi- plication table for this set of operations is given below. E G a ol) o) 0) 7 E G a oo of o®) G G 3 E oo oa) of Q G E G oe) Ce o oo a o®) o E G G of) of) of) oo) CG E CG ony mal) oo o) G. Gg E The successful construction of this table demonstrates that the set of six opera- tions does indeed form a group. This group is represented by the symbol C3,. TheSYMMETRY AND STRUCTURE 39 table shows that its characteristics are more general than those of the group C,. Thus it contains some operations that are not, as well as some which are, their own inverse. It also involves a number of multiplications that are not commutative. 2.3. Some General Rules for Multiplication of Symmetry Operations In the preceding section several specific examples of multiplication of symmetry operations have been worked out. On the basis of this experience, the following general rules should not be difficult to accept: 1. The product of two proper rotations must be another proper rotation. Thus although rotations can be created by combining reflections [recall: o(xz) X o(yz) = C,(z)], the reverse is not possible. 2. The product of two reflections in planes meeting at an angle @ is a rotation by 26 about the axis formed by the line of intersection of the planes [recall: 0!) X o) = C, for the ammonia molecule]. 3. When there is a rotation operation C,, and a reflection in a plane containing the axis, there must be altogether such reflections in a set of n planes separated by angles of 27 /2n, intersecting along the C,, axis [recall: 6) X C3 = o for the ammonia molecule]. 4. The product of two C2 operations about axes that intersect at an angle @ is a rotation by 2 about an axis perpendicular to the plane containing the two C2 axes. 5. The following pairs of operations always commute: (a) Two rotations about the same axis. (5) Reflections through planes perpendicular to each other. (c) The inversion and any other operation. (d) Two C; operations about perpendicular axes. (e) C, and op, where the C, axis is vertical. 2-4, A Systematic Listing of Symmetry Groups, with Examples The symmetry groups to which real molecules may belong are very numerous. However they may be systematically classified by considering how to build them up using increasingly more elaborate combinations of symmetry operations. The outline that follows, though neither unique in its approach nor rigorous in its pro- cedure, affords a practical scheme for use by most chemists. The simplest nontrivial groups are those of order 2, that is, those containing but one operation in addition to E. The additional operation must be one that is its own inverse; thus the only groups of order 2 are: CuEe CE, Cu, The symbols for these groups are rather arbitrary, except for C2 which, we shall soon see, forms part of a pattern.40 INTRODUCTORY TOPICS Molecules with C, symmetry are fairly numerous. Examples are the thionyl halides and sulfoxides 2-III, and secondary amines 2-IV. Molecules having a center of symmetry as their only symmetry element are quite rare; two types are shown as 2-V and 2-VI. The reader should find it very challenging, though not impossible, to think of others. Molecules of C> symmetry are fairly common, two examples being 2-VIT and 2-VIIf. cc S. N. eee x40 R44 H x R xv uh Iv) lan L Xx (2) x Y ae H H X eae o—o x Y x (Vn) NN (2V1) (2-VIIT) The Uniaxial or C,, Groups. These are the groups in which all operations are due to the presence of a proper axis as the sole symmetry element. The general symbol for such a group, and the operations in it, are Cui Cn C2 C2... CPNCE=E A Cy group is thus of order n, We have already mentioned the group C2. Molecules with*pure axial symmetry other than C2 are rare. Two examples of the group C3 are shown in 2-IX and 2-X. ON & (2X) (IX) The Cy, Groups. If in addition to a proper axis of order 1 there is also a set of n vertical planes, we have a group of order 2n, designated C,,. This type of sym- metry is found quite frequently and is illustrated in 2-X] to 2-XV, where the values of nare 2 to 6.SYMMETRY AND STRUCTURE 41 N a i I FL [ UF 0-0 Feel @x) oF F (2X) (XIN) HC CH, ° } HAC ©) ‘CH, Ni H.C Cr ‘cH, (2XIV) XV) The Cnn Groups. If in addition to a proper axis of order n there is also a hori- zontal plane of symmetry, we have a group of order 2n, designated Cyy. The 2n operations include S,” operations that are products of C” and o}, for 2 odd, to make the total of 2”. Thus for C3, the operations are G3, 6,6 oh an X C3 = C3 Xo, = S3 on X C} = CPX ay = S$ Molecules of C,, symmetry with > 2 are relatively rare; examples with n = 2, 3, and 4 are shown in 2-XVI to 2-XVIII. a c. 07 30 lei 7® a Mo ne x7] ac (2XVI1) CH, (2XV1) (2-XVIID) The D,, Groups. When a vertical C, axis is accompanied by a set of n C) axes perpendicular to it, the group is D,. Molecules of D, symmetry are, in general, rare, but there is one very important type, namely, the trischelates (2-XIX) of D3 sym- metry. The D,,, Groups. If to the operations making up a D, group we add reflection ina horizontal plane of symmetry, the group Dp is obtained. It should be noted42 INTRODUCTORY TOPICS that products of the type C) a» will give rise to a set of reflections in vertical planes. These planes contain the C2 axes; this point is important in regard to the distinction between Dy and Dag, mentioned next. The Dan symmetry is found in a number of important molecules, a few of which are benzene (Den), ferrocene in an eclipsed configuration (Dsj), RezClg”, which we examined above, (Day), PtCl- (Day), and the boron halides (D3,) and PFs(D3,). All right prisms with regular polygons for bases as illustrated in 2-XX and 2-XXI, and all bipyramids, as illus- trated in 2-XXII and 2-XXIII, have Dya-type symmetry. Ue (2XX) 2. 2-XXIIL} (2XX1) (2XXID)¢ ) The D,z Groups. If to the operations making up a D, group we add a set of vertical planes that bisect the angles between pairs of C axes (note the distinction from the vertical planes in Dyn), we have a group called Drg. The Dna groups have No gp. Perhaps the most celebrated examples of Dna symmetry are the D3q and Dsg symmetries of R3 W==WR; and ferrocene in their staggered configurations 2-XXIV. and 2-XXV. R Ray i Fe A, & Ro. (2XXV) (2-XXIV) Two comments about the scheme so far outlined may be helpful. The reader may have wondered why we did not consider the result of adding to the operations of Ch both a set of noy’s and a oy. The answer is that this is simply another way of getting to D,,,, since a set of Cy axes is formed along the lines of intersection of the 9, with each of the o,s. By convention, and in accord with the symbols used to designate the groups, it is preferable to proceed as we did. Second, in dealing withSYMMETRY AND STRUCTURE 43 the Dnr type groups, if a horizontal plane is found, there must be only the ” vertical planes containing the C2 axis. If dihedral planes were also present, there would be, in all 2 planes and hence, as shown above, a principal axis of order 2, thus vitiating the assumption of a D, type of group. The S,, Groups. Our scheme has, so far, overlooked one possibility, namely, that a molecule might contain an S, axis as its only symmetry element (except for others that are directly subservient to it). It can be shown that for n odd, the groups of operations arising would actually be those forming the group C,,,. For example, take the operations generated by an S33 axis: Comparison with the list of operations in the group C3, given on page 41, shows that the two lists are identical. It is only when n is an even number that new groups can arise that are not already in the scheme. For instance, consider the set of operations generated by an Sy axis: Se Sl=C St E This set of operations satisfies the four requirements for a group and is not a set that can be obtained by any procedure previously described. Thus Sg, S¢, etc., are new groups. They are distinguished by the fact that they contain no operation that is not an S¥ operation, even though it may be written in another way, as with S} = C) above. Note that the group S} is not new. A little thought will show that the operation Sy is identical with the operation i. Hence the group that could be called S. is the one we have already called C;. An example of a molecule with S4 symmetry is shown in 2-XXVI. Molecules with S, symmetries are not very common. ‘, B at | a» + x My Nal = x7 i R (2XXVI)44 INTRODUCTORY TOPICS Linear Molecules. There are only two kinds of symmetry for linear molecules, There are those represented by 2-X XVII that have identical ends. Thus in addition to an infinitefold rotation axis C.., coinciding with the molecular axis, and an infinite number of vertical symmetry planes, they have a horizontal plane of symmetry and an infinite number of C2 axes perpendicular to C.. The group of these operations is Day. A linear molecule with different ends (2-XXVIII), has only Cx and the »'s as symmetry elements. The group of operations generated by these is called Cap. A~B—C—B-A A—B—C—D Q-XXVI) (2-XXVIH1) 2-5. The Groups of Very High Symmetry The scheme followed in the preceding section has considered only cases in which there is a single axis of order equal to or greater than 3. It is possible to have sym- metry groups in which there are several such axes. There are, in fact, seven such groups, and several of them are of paramount importance. The Tetrahedron. We consider first a regular tetrahedron. Figure 2-6 shows some of the symmetry elements of the tetrahedron, including at least one of each kind. From this it can be seen that the tetrahedron has altogether 24 symmetry operations, which are as follows: There are three S4 axes, each of which gives rise to the operations S4, S$} = C2, Si, and Si = E. Neglecting the S!’s, this makes 3 X 3 = 9. There are four C3 axes, each giving rise to C3, C3 and C} = E. Again omitting the identity operations, this makes 4 X 2 = 8. There are six reflection planes, only one of which is shown in Fig. 2-6, giving rise to six og operations. Thus there are9 + 8+ 6+ one identity operation = 24 operations. This group is called Tg. It is worth emphasizing that despite the considerable amount of symmetry, there is no inversion center in Tz symmetry. There are, of course, nu- Fig. 2-6. The tetrahedron, showing some of its essential symmetry elements. All Sg and C3 axes are shown, but only one of the six dihedral planesSYMMETRY AND STRUCTURE 45 merous molecules having full Tg symmetry, such as CHa, SiFs, ClOZ, Ni(CO)«, and Ir4(CO) 2, and many others where the symmetry is less but approximates to it. If we remove from the Ty group the reflections, it turns out that the Sy and S} operations are also lost. The remaining twelve operations (E, four C3 operations, four C3 operations and three C operations) form a group, designated T. This group in itself has little importance, since it is very rarely, if ever, encountered in real molecules. However if we then add to the operations in the group 7 a different set of reflections in the three planes defined so that each one contains two of the C2 axes, and work out all products of operations, we get a new group of 24 operations (E, four Cs, four C3, three C2, three a, i, four Sg, four $3) denoted T,. This, too, is rare, but it occurs in some “octahedral” complexes in which the ligands are planar and arranged as in 2-XXIX. The important feature here is that each pair of ligands ‘on each of the Cartesian axes is in a different one of the three mutually perpen- dicular planes, xy, xz, yz. Real cases are provided by W(NMez)g and several M(NO3)é" ions in which the NO} ions are bidentate. > (2-XXIX) The Octahedron and the Cube. These two bodies have the same elements, as shown in Fig. 2-7, where the octahedron is inscribed in a cube, and the centers of the six cube faces form the vertices of the octahedron. Conversely, the centers of the eight faces of the octahedron form the vertices of a cube. Figure 2-7 shows one Fig. 2-7. The octahedron and the cube, showing one of each of their essential types of symmetry element46 INTRODUCTORY TOPICS (a) (b) Fig, 2-8. The two regular polyhedra having /, symmetry. (2) The pentagonal dodecahedron. (b) The icosahedron. of each of the types of symmetry element that these two polyhedra possess. The list of symmetry operations is as follows: There are three Cy axes, each generating Cy, C7 = C2, Ci, C} = E. Thus there are 3X 3 = 9 rotations, excluding Cis. There are four C3 axes giving four C3’s and four C?’s. There are six C) axes bisecting opposite edges, giving six y's, There are three planes of the type o, and six of the type oq, giving rise to nine reflection operations. The Cg axes are also S4 axes and each of these generates the operations Sq, Ss} = Cand Sj, the first and last of which are not yet listed, thus adding 3 X 2 = 6 more to the list. The C3 axes are also S¢ axes and each of these generates the new operations Sg, S}=i, and S§. The i counts only once, so there are then (4 X 2) + 1 = 9 more new operations. The entire group thus consists of the identity + 9 + 8 +6 +9 +6 +9 = 48 op- erations. This group is denoted Oj. It is, of course, a very important type of sym- metry since octahedral molecules (¢.g., SF¢), octahedral complexes [Co(NH3)@*, IrCl"], and octahedral interstices in solid arrays are very common. There is a group O, which consists of only the 24 proper rotations from Op, but this, like T, is rarely if ever encountered in Nature. The Pentagonal Dodecahedron and the Icosahedron. These bodies (Fig. 2-8) are related to each other in the same way as are the octahedron and the cube, the vertices of one defining the face centers of the other, and vice versa. Both have the same symmetry operations, a total of 120! We shall not list them in detail but merely mention the basic symmetry elements: six Cs axes; ten C3 axes, fifteen Cz axes, and fifteen planes of symmetry. The group of 120 operations is designated J, and is often called the icosahedral group. There is as yet no known example of a molecule that is a pentagonal dodecahe- dron, but the icosahedron is a key structural unit in boron chemistry, occurring in all forms of elemental boron as well as in the B)2H?3 ion. If the symmetry planes are omitted, a group called J consisting of only proper rotations remains. This is mentioned purely for the sake of completeness, since no example of its occurrence in Nature is known.SYMMETRY AND STRUCTURE 47 2-6. Molecular Dissymmetry and Optical Activity Optical activity, that is, rotation of the plane of polarized light coupled with unequal absorption of the right- and left-circularly polarized components, is a property of a molecule (or an entire three-dimensional array of atoms or molecules) that is not superposable on its mirror image. When the number of molecules of one type ex- ceeds the number of those that are their nonsuperposable mirror images, a net optical activity results. To predict when optical activity will be possible, it is nec- essary to have a criterion to determine when a molecule and its mirror image will not be identical, that is, superposable. Molecules that are not superposable on their mirror images are called dissym- metric. This term is preferable to “asymmetric,” which means “without symmetry,” whereas dissymmetric molecules can and often do Possess some symmetry, as will be seen.* A compact statement of the relation between molecular symmetry properties and dissymmetric character is: a molecule that has no axis of improper rotation is disspmmetric. This statement includes and extends the usual one to the effect that optical isomerism exists when a molecule has neither a plane nor a center of symmetry. It has already been noted that the inversion operation iis equivalent to the improper rotation S. Similarly, S, is a correct although unused way of representing o, since it implies rotation by 27/1, equivalent to no net rotation, in conjunction with the reflection. Thus o and i are simply special cases of improper rotations. However even when o and é are absent, a molecule may still be identical with its mirror image if it possesses an S, axis of some higher order. A good example of this is provided by the (-RNBX—), molecule shown in 2-XXVI. This molecule has neither a plane nor a center of symmetry, but inspection shows that it can be superposed on its mirror image. As we have noted, it belongs to the symmetry group S4. Dissymmetric molecules either have no symmetry at all, or they belong to one of the groups consisting only of proper rotation operations, that is, the C, or Dy, groups. (Groups T, O, and J are, in practice, not encountered, though molecules in these groups must also be dissymmetric.) Important examples are the bischelate and trischelate octahedral complexes 2-VIII, 2-X, and 2-XIX. MOLECULAR SYMMETRY 2-7, Coordination Compounds Historically it has been customary to treat coordination compounds as a special class separate from molecular compounds. On the basis of actual fact only (i.e., * Dissymmetry is sometimes called chirality, and dissymmetric chiral, from the Greek word xeup for hand, in view of the left-hand/right-hand relation of molecules that are mirror images.48 INTRODUCTORY TOPICS neglecting the purely traditional reasons for such a distinction), there is very little, if indeed any, basis for continuing this dichotomy. Coordination compounds are conventionally formulated as consisting of a central atom or ion surrounded by a set (usually 2 to 9) of other atoms, ions or smal] mol- ecules, the latter being called ligands. The resulting conglomeration is often called a complex or, if it is charged, a complex ion. The set of ligands need not consist of several small, independent sets of atoms (or single atoms) but may involve fairly elaborate arrangements of atoms connecting those few that are directly bound to—or coordinated to—the central atom. However, there are many molecular compounds of which the same description may be given. Consider, for illustration, the following: SiF, SIF Cx(CO), Co(NHa)3* Ses Peg CHNHSH Coc Conventionally the first two are called molecules and five of the other six are called complexes; Cr(CO)s can be found referred to in either way depending on the context. Obviously, one basis for the different designations is the presence or absence of a net charge, only uncharged species being called molecules. Beyond this, which is really quite a superficial characteristic as compared with such basic ‘ones as geometric and electronic structures, there is no logical reason for the division. The essential irrelevancy of the question of overall charge is well demonstrated by the fact that Pt(NH3)2Clo, Cu(acac)2, CoBr2(Ph3P)2, and scores of similar com- pounds are quite normally called complexes. The molecules are really only com- plexes that happen to have a charge of zero instead of +n or —m. Thus SiF2", PF, and SF¢ are isoelectronic and isostructural. Although the character of the bonds from the central atom to fluorine atoms doubtless varies from one to another, there is no basis for believing that SFe differs more from PF than the latter does from SiFZ-. It might be argued that the terms “complex” or “coordination compound” should be applied only when the central atom, in some oxidation state, and the ligands, can be considered to exist independently, under reasonably normal chemical con- ditions. Thus Cr+ and NHg would be said to so exist. However Cr** actually exists under normal chemical conditions not as such but as Cr3+ (aq) which is, in detail, Cr(H20)3*, another species that would itself be called a complex. Again, ina similar vein, the argument that PFs and SiFg” can be considered to consist, respectively, of PF; + F~ and SiF, + 2F~, whereas there is no comparable breakdown of SF6, is a poor one; once the set of six fluorine atoms is completed about the central atom, they become equivalent. The possibility of their having had different origins has no bearing on the nature of the final complex. The terms “coordination compound” and “complex” may therefore be broadly defined to embrace all species, charged or uncharged, in which a central atom is surrounded by a set of outer or ligand atoms. Having thus defined coordination compounds in a comprehensive way, we can proceed to discuss their structures in terms of only two properties: (1) coordination number, the number of outer, or ligand, atoms. bonded to the central one, and (2) coordination geometry, the geometric arrangement of these ligand atoms and theSYMMETRY AND STRUCTURE 49 consequent symmetry of the complex. We shall consider in detail coordination numbers 2 to 9, discussing under each the principal ligand arrangements. Higher coordination numbers will be discussed only briefly as they occur much less fre- quently. Coordination Number 2. There are two geometric Possibilities, linear and bent. If the two ligands are identical, the general types and their symmetrics are: linear, L—M—L, Day; bent; L—M—L, Cy). This coordination number is, of course, found in numerous molecular compounds of divalent elements, but is relatively uncommon otherwise. In many cases where stoichiometry might imply its occur- rence, a higher coordination number actually occurs because some ligands form “bridges” between two central atoms. In terms of the more conventional types of coordination compound—those with a rather metallic element at the center—it is restricted mainly to some complexes of Cu', Ag!, Au!, and Hg!!. Such complexes have linear arrangements of the metal ion and the two ligand atoms, and typical ones are [CICuCl]-, [Hs NAgNHs]*, [ClAuCl]-, and [NCHgCN]. The metal atoms in cations such as [UO2]?*, [UQ2]*, and [PuO2]2+, which are linear, may also be said to have coordination number 2, but these oxo cations interact fairly strongly with additional ligands and their actual coordination numbers are much higher; it is true, however, that the central atoms have a specially strong affinity for the two oxygen atoms. Linear coordination also occurs in the several trihalide ions, such as Ij and CIBrCl-. Coordination Number 3.1" The two most symmetrical arrangements are planar (2-XXX) and pyramidal (2-XXXI), with Ds, and C3, symmetry, respectively. Both these arrangements are found often among molecules formed by trivalent central clements. Among complexes of the metallic elements this is a Tare coordi- nation number; nearly all compounds or complexes of metal cations with stoichi- ometry MX; have structures in which sharing of ligands leads to a coordination for M that exceeds three. There are, however, a few exceptions, such as the planar Hels ion that occurs in [(CH3)3S*][Hgls], the MN3 groups that occur in Cr(NRz2)3 and Fe(NR2)3, where R = (CH3)3Si, and the CuS; group found in Cu[SC(NH2)2]3Cl and Cu(SPPh3)3C10x. In a few cases (¢.g., CIF; and BrF3), a T-shaped form (2-XXXII) of three- coordination (symmetry C3,) is found. B Ay ra B—At Bef B—A—B “p (2-XXX) 2X B 2XXXD) (2XXXII) Coordination Number 4. This is a highly important coordination number, oc- curring in hundreds of thousands of compounds, including, inter alia, most of those formed by the element carbon, essentially all those formed by silicon, germanium, and tin, and many compounds and complexes of other elements. There are three | P.G. Eller, D. C. Bradley, M. B. Hursthouse, and D. W. Meek, Coord. Chem. Rev., 1977, 24, | (a comprehensive review with 306 references).50 INTRODUCTORY TOPICS principal geometries. By far the most prevalent is tetrahedral geometry (2-XXXIll), which has symmetry Tq when ideal. Tetrahedral complexes or molecules are almost the only kind of four-coordinate ones formed by non-transition elements; whenever the central atom has no electrons in its valence shell orbitals except the four pairs forming the o bonds to ligands, these bonds are disposed in a tetrahedral fashion. With many transition metal complexes, square geometry (2-XXXIV) occurs be- B B B Nac i NZ ole ae ae ™~, Bee BB i B (2XXXIM1) ncaa (2XXXV) cause of the presence of additional valence shell electrons and orbitals (i.¢., partially filled d orbitals), although there are also many tetrahedral complexes formed by the transition metals. In some cases (e.g., with Ni!!, Co!!, and Cu! in particular), there may be only a small difference in stability between the tetrahedral and the square arrangement and rapid interconversions may occur. Square complexes are also found with nontransitional central atoms when there are two electron pairs present beyond the four used in bonding; these two pairs lie above and below the plane of the molecule. Examples are XeF, and (ICI3)2. Sim- ilarly, when there is one “extra” electron pair, as in SF4, the irregular arrangement, of symmetry Cz (2-XXXV) is adopted. More detailed discussions of these non- transition element structures will be found in Chapter 5. Coordination Number 5. Though less common than numbers 4 and 6, coordi- nation number 5 is still very important.'® There are two principal geometries, and these may be conveniently designated by stating the polyhedra that are defined by the set of ligand atoms. In one case the ligand atoms lie at the vertices of a tri- gonal bipyramid (rbp) (2-XXXVD), and in the other at the vertices of a square pyramid (sp) (2-XXXVII). The thp belongs to the symmetry group D3,; the sp belongs to the group Cy. It is interesting and highly important that these two (2XXXVID) (2-XXXVI) structures are similar enough to be interconverted without great difficulty. Moreover, a large fraction of the known five-coordinate complexes have structures that are intermediate between these two prototype structures. This ready defor- mability and interconvertibility gives rise to one of the most important types of stereochemical nonrigidity (cf. Section 28-13). th EL. Muetterties and R. A. Schunn, Q. Rev., 1966, 20, 245; B. F. Hoskins and F. D. Whillan, Coord. Chem, Rev., 1973, 11, 343.SYMMETRY AND STRUCTURE 51 A a Ny” é-\ \/ Tevrogmel, G+ 2, C3 ay ==) 2. G— 9 So — @ It is important to recognize that there exists a continuous range of structures between the trigonal prism (2-XX) and the trigonal antiprism (Fig. 2-9). This can best be considered in terms of the twist angle @, shown in Fig. 2-10, which is 0° for the prism and 60° for the antiprism. The drawing represents schematically a tris- chelate complex with an intermediate configuration. While there are some tris- (dithiolene)metal complexes with > = 0, there are others with angles between 08 and 60°. The tropolonato anion (Section 4-27) is another bidentate ligand prone to give intermediate angles. To explain the range of ¢ values observed, many factors have been considered.*3 With few if any exceptions, the metal ion itself will prefer $ = 60°, but geometrical and electronic properties of the ligand can override this. If the ligand has a short “bite” (i.e., a short distance between the two atoms bonded to the metal), low ¢ is favored. It is also believed that direct interactions between the sulfur atoms in certain |,2-dithiolene complexes favor low ¢. There are a few 3 A.A, Diamantis, M. R. Snow, and J. A. Vanzo, J. C. S. Chem. Comm., 1976, 264. + A. Avdeef and J. P. Fackler, Jr., Jnorg. Chem., 1975, 14, 2002. 5M. Cowie and M. J, Bennett, Inorg. Chem., 1976, 15, 1596.SYMMETRY AND STRUCTURE 53 cases in which the ligand is all one hexadentate unit so designed as to impose a small or zero ¢ angle.® Coordination Number 7.78 There are three important geometrical arrangements: the pentagonal bipyramid (2-XXXVIII) of symmetry Ds,; the capped octahedron (2-XXXIX), symmetry C3p, obtained by adding to an octahedral set a seventh li- (2XXXVII) (2XXXIX) (2XL) gand over the center of one face, which then becomes enlarged; and the capped trigonal prism (2-XL), symmetry C,, obtained by placing the seventh ligand over one of the rectangular faces of the trigonal prism. The available structural data suggest that Nature does not particularly favor any one of these except where a bias might be built into a polydentate ligand, and theory likewise implies that all three will, in general, have similar stability. Moreover, theory suggests that inter- conversions can occur without difficulty, so that seven-coordinate complexes should, in general, be stereochemically nonrigid. Coordination Number 8.°°" It is conceptually convenient to begin with the most symmetrical polyhedron having eight vertices, namely, the cube, which has Oy symmetry. Cubic coordination occurs only rarely in discrete complexes, especially in those of uranium and some other actinides, although it occurs in various solid compounds where the anions form continuous arrays, as in the CsCl structure. Its occurrence is infrequent, presumably, because there are several ways in which the cube may be distorted so as to lessen repulsions between the X atoms while main- taining good M—X interactions. The two principal ways in which the cube may become distorted are shown in Fig. 2-11. The first of these, rotation of one square face by 45° relative to the one opposite to it, lessens repulsions between nonbonded atoms while leaving M—X distances unaltered. The resulting polyhedron is the square antiprism (symmetry Daa). \t has square top and bottom and eight isosceles triangles for its vertical faces, ‘The second distortion shown can be best comprehended by recognizing that the cube is composed of two interpenetrating tetrahedra. The distortion occurs when the vertices of one of these tetrahedra are displaced so as to decrease the two vertical angles, that is, to elongate the tetrahedron, while the vertices of the other one are displaced to produce a flattened tetrahedron. The resulting polyhedron is called a dodecahedron, or more specifically, to distinguish it from several other kinds of © P.B. Donaldson, P. A. Tasker, and N. W. Alcock, J. C. S, Dalton, 1977, 1160. 7 R. Hoffmann, B. F. Beier, E. L, Muetterties, and A. R. Rossi, Inorg. Chem. 1977, 16, 511 ® M.G.B. Drew, Prog. Inorg. Chem., 1977, 23, 67 * M.G. B. Drew, Coord. Chem. Rev., 1977, 24, 179 (covers coordination numbers 8 to 14, with 295 references). ® J. Burdett, R. Hoffmann, and R. C. Fay, Inorg. Chem., 1978, 17, 2553. % A. R. Al-Karaghouli, R. O. Day, and J.S, Wood, Inorg. Chem., 1978, 17, 3702.354 INTRODUCTORY TOPICS 0) ) Fig, 2-11. The two most important ways of distorting the cube: (a) to produce a square antiprism:; (b) to produce a dodecahedron dodecahedron, a triangulated dodecahedron. It has D2g symmetry, and it is im- portant to note that its vertices are not all equivalent but are divided into two bis- phenoidal sets, those within each set being equivalent. Detailed analysis of the energetics of M—X and X—X interactions suggests that there will in general be little difference between the energies of the square antiprism and the dodecahedral arrangement, unless other factors, such as the Fig. 2-12. The structure of many nine-coordinate complexes.SYMMETRY AND STRUCTURE 55 istence of chelate rings, energies of partially filled inner shells, exceptional op- portunities for orbital hydridization, or the like, come into play. Both arrangements occur quite commonly, and in some cases [e.g., the M(CN)?" (M = Moor W:” = 3 or 4) ions] the geometry varies from one kind to the other with changes in the counterion in crystalline salts and on changing from crystalline to solution phases. A form of eight-coordination, which is a variant of the dodecahedral arrange- ment, is found in several compounds containing bidentate ligands in which the two coordinated atoms are very close together (ligands said to have a small “bite”), such as NO} and 03”. In these, the close pairs of ligand atoms lie on the m edges of the dodecahedron (see Fig. 2-11b); these edges are then very short. Examples of this are the Cr(O2)37 and Co(NO3)2" ions and the Ti(NO3)4 molecule. Three other forms of octacoordination, which occur much less often and are essentially restricted to actinide and lanthanide compounds, are the hexagonal bipyramid (Do) (2-XL1), the bicapped trigonal prism (D3,) (2-XLII) and the B (2XLD (XLID (2-XLII) bicapped trigonal antiprism (D3) (2-XLIII). The hexagonal bipyramid is restricted almost entirely to the oxo ions, where an OMO group defines the axis of the bi- pyramid, though it is occasionally found elsewhere. Coordination Number 9.9*!9 There are two regular structures, the tricapped trigonal prism (Fig. 2-12) and a capped square antiprism (Fig. 2-11a), where the ninth ligand lies over one square face. In both complexes and clusters (Section 2-8) the former is far more common; only three examples of the capped square antiprism have been reported, including one in the compound RbTh3F}3 where both types of nine-coordinate polyhedra otcur. Examples of coordination number 9 are re- stricted to lanthanide and actinide complexes and to [ReH]?~, and the majority are polymeric, not discrete. Specific examples of the tricapped trigonal prism are [ReHo]?- and the [M(H20)9]?* ions with lanthanide ions. The rarer capped square antiprism occurs in [ThFg]*~ and the Cl-bridged [LaCI(H 0);]$*. Higher Coordination Numbers.® These are rare and found only for the largest metal ions; the geometry is usually not easily, if at all, classifiable in terms of simple prototypal shapes. Ten-coordinate species include KsTh(O2CCO>)4-4H20, which has a bicapped square antiprism, and several others with capped dodecahedral geometry. Only two eleven-coordinate complexes have been clearly delineated. All ‘0 L. J. Guggenberger and E. L. Muetterties, J. Am. Chem. Soc., 1976, 98, 7221.56 INTRODUCTORY TOPICS but one of the known cases of twelve-coordination comprise M(LL)< species with distorted icosahedral arrangements. Examples are [Ce(NO3)¢]*~ and other [M(NO3)o]2> ions and [Pr(naph),]3+, where naph is 1,8-naphthyridine. 2-8. Cage and Cluster Structures!" The formation of polyhedral cages and clusters is now recognized as an important and widespread phenomenon, and examples may be found in nearly all parts of the Periodic Table. This section mentions each of the principal polyhedra and gives illustrations. Further details may be found under the chemistry of the particular elements and in the sections on metal atom clusters (Chap. 26) and polynuclear metal carbonyls (Section 25-1). A cage or cluster is in a certain sense the antithesis of a complex; yet there are many similarities due to common symmetry properties. In each type of structure a set of atoms defines the vertices of a polyhedron, but in the one case—the com- plex—these atoms are each bound to one central atom and not to each other, whereas in the other—the cage or cluster—there is no central atom and the essential feature is a system of bonds connecting each atom directly to its neighbors in the polyhedron. There are, however, some examples of clusters that also have a central atom, usually carbon. Thus they combine the features of both complexes and clusters, although the bonding between atoms in the polyhedron appears to be stronger and more important than bonds from outer atoms to the central atoms. Examples of these “filled clusters” are Rug(CO)\7C, Cog(CO))4C™, and Fes(CO);5C. The first two have an octahedral set of metal atoms with a C atom at the center, and the iron compound has a square pyramidal set of iron atoms with a C atom approximately in the basal plane (2-XLIV). To a considerable extent the polyhedra found in cages and clusters are the same as those adopted by coordination compounds (e.g., the tetrahedron, trigonal bi- pyramid, octahedron), but there are also others (see especially the polyhedra with six vertices), and cages with more than six vertices are far more common than coordination numbers greater than 6. It should be noted that triangular clusters, as in [Re3Clj2]~ or Os3(CO);9, though not literally polyhedra, are not essentially different from polyhedral species such as MogClg* or Ir4(CO)}2, respectively. Just as all ligand atoms in a set need not be identical, so the atoms making up a cage or cluster may be different; indeed, to exclude species made up of more than one type of atom would be to exclude the majority of cages and clusters, including some of the most interesting and important ones. Four Vertices. Tetrahedral cages or clusters have long been known for the Pa, 1 R.B. King, in Progress in Inorganic Chemistry, Vol. 15,S. J. Lippard, Ed., Wiley-Interscience, 1972 (transition metal cluster compounds); P. Chini, G. Longoni, and V. G. Albano, in Advances in Organometallic Chemistry, Vol. 14, F.G. A. Stone and R. West, Eds., Academic Press, 1976 (review of metal clusters of the metal carbonyl type); J. D. Corbett, in Progress in Inorganic Chemistry, Vol. 21, S. J. Lippard, Ed., Wiley-Interscience, 1976 (cluster ions formed by main group metals); M. G, B. Drew, Coord. Chem. Rev., 1977, 24, 179 (structures of high coordination number complexes); K. Wade, Adv. Inorg. Chem. Radiochem., 1976, 18, 1 (an authoritative treatment of structural bonding patterns in cluster chemistry).SYMMETRY AND STRUCTURE 37 sq, and Sb, molecules and in more recent years have been found in polynuclear metal carbonyls such as Cog(CO)12, Irs(CO)12, [15-CsHsFe(CO)]4,* RSiCo3- (CO)o, Fes(CO)#3, Res(CO);2H4, and a number of others; BsCla is another well-known example and doubtless many more will be encountered. Five Vertices. Polyhedra with five vertices are the trigonal bipyramid (rbp) and the square pyramid (sp). Both are found among the boranes and carbaboranes (e.g., the tbp in BxC2Hs and the sp in BsHo), as well as among the transition elements. Examples of the latter are the rbp cluster, PtsSn3, in (CgsH12)3Pt3(SnCl3)2 and the sp clusters in Fes(CO);sC (2-XLIV) and the Fe3(CO)9E (E = S or Se) species (2-XLV). The Oss(CO);¢ molecule provides a recent example of a thp. E=S or Se (2-XLIV) (2-XLV) Six Vertices. Octahedral clusters and cages are rather numerous. Several are polynuclear metal carbonyls, such as Rhg(CO);6 and [Cog(CO)14]4>. (2-XLV1) (XLVI) There is also an extensive series of metal atom cluster compounds formed by nio- bium, tantalum, molybdenum, and tungsten, all based on octahedral sets of metal atoms, the principal types being of MgXg (2-XLVI) and MgX12 (2-XLVII) stoi- + The 4° is a symbol denoting the number of atoms (5 C) involved in x-bonding to the metal.58 INTRODUCTORY TOPICS chiometry. The BsH?" and BgC2H, species are also octahedral. The borane BeHo, however, has a pentagonal pyramid of boron atoms; Bg octahedra also occur in the class of borides of general formula MB. Still other, less regular geometries also occur—for example, a capped sp in H20s¢(CO),s (2-XLVIII) and a bicapped tetrahedron in Osg(CO);— (2- WF * = 0s(CO), (2XLVIID # = 0s(CO), (2-XLIX) Seven Vertices. Such polyhedra are relatively rare. The isoelectronic BH} and BsC2H7 species have pentagonal bipyramidal (Ds,) structures. The Os7(CO)a1 molecule has a capped octahedron of Os atoms with three CO groups on each Os. Eight Vertices. Eight-atom polyhedral structures are very numerous. By far the most common polyhedron is the cube; this is in direct contrast to the situation with eightfold coordination, where a cubic arrangement of ligands is extremely rare because it is disfavored relative to the square antiprism and the triangulated dodecahedron in which ligand-ligand contacts are reduced. In the case of a cage compound, of course, it is the structure in which contacts between atoms are maximized that will tend to be favored (provided good bond angles can be main- tained), since bonding rather than repulsive interactions exist between neighboring atoms. = The only known cases with eight /ike atoms in a cubic array are the hydrocarbon cubane, CgHs, and the Cug(i- MNT)3" ion [- MNT = SxCC(CN)3"]. The other cubic systems all involve two different species of atom that alternate as shown in (2-L). In all cases either the A atoms or the B atoms or both have appended atoms or groups. The following list collects some of the many cube species, the elements at the alternate vertices of the cube being given in bold type. A (and appended groups) B (and appended groups) Ma(CO); SEt 0s(CO); ° PiMes or PtEty C1, Br, 1, OH CH3Za OCH; TI OCH; a-CsHsFe s MesAsCu 1 Phal NPh Co(CO)s Sb FeSR s RacecarSYMMETRY AND STRUCTURE 59 A My | ! 13 BW—« (2L) Although the polyhedron in cubane, or in any other (AR) molecule, may have the full O, symmetry of a cube, the A4Bg-type structures can have at best tetra- hedral, Ty, symmetry since they consist of two interpenetrating tetrahedra. It must also be noted that only when the two interpenetrating tetrahedra happen. to be exactly the same size will all the ABA and BAB angles be equal to 90°. Since the A and the B atoms differ, it is not in general to be expected that this will occur. In fact, there is, in principle, a whole range of bonding possibilities. At one extreme, represented by [(75-CsHs)Fe(CO)]4, the members of one set of atoms (the Fe atoms) are so close together that they must be considered to be directly bonded, whereas the other set (the C atoms of the CO groups) are not at all bonded among themselves but only to those in the first set. In this extreme, it seems best to classify the system as having a tetrahedral cluster (of Fe atoms) supplemented by bridging CO groups. At the other extreme are the A4B, systems in which all A—A and B—B distances are too long to admit of significant A—A or B—B bonding; thus the system can be regarded as genuinely cubic (even if the angles differ somewhat from 90°). This is true of all the systems listed above. However the atoms in the smaller of the two tetrahedra tend to have some amount of direct interaction with one another, thus blurring the line of demarcation between the “cluster” and “cage” types. Of the systems listed above those with RSFe and S groups are especially note- worthy because they occur in many of the biological electron carriers called fer- redoxins (see Section 31-5 for details) A relatively few species are known in which the polyhedron is, at least approxi- mately, a triangulated dodecahedron (Fig. 2-11b). These are the boron species BsHg-, BeC2Hg, and BgClg. Nine Vertices. Cages with nine vertices are rare. Representative ones are Bis* (in BzgClog), BoH3”, and B7C2Hs, all of which have the tricapped trigonal prism structure (Fig. 2-12), and Sn3~, which is a square antiprism capped on one square face!2, Ten Vertices. Species with ten vertices are well known. In BjoH?5 and BgC2H jo the polyhedron 2-L1 is a square antiprism capped on the square faces (symmetry D4q). But there is a far commoner structure for 10-atom cages that is commonly called the adamantane structure after the hydrocarbon adamantane (CjoH 6), which has this structure; it is depicted in 2-LII and consists of two subsets of atoms: a set of four (A) that lie at the vertices of a tetrahedron and a set of six (B) that lie at the vertices of an octahedron. The entire assemblage has the Tz symmetry of the tetrahedron. From other points of view it may be regarded as a tetrahedron with '2 J.D. Corbett and P. A. Edwards, J. Am. Chem. Soc., 1977, 99, 331360 INTRODUCTORY TOPICS (Lp a bridging atom over each edge or as an octahedron with a triply bridging atom over an alternating set of four of the eight triangular faces. The adamantane structure is found in dozens of A4Be-type cage compounds formed mainly by the main group elements. The oldest recognized examples of this structure are probably the phosphorus(IIl) and phosphorus(V) oxides, in which we have P4O, and (OP)4Og, respectively. Other representative examples include P4(NCH3)¢, (OP)4(NCH3)o, Asa(NCHs)o, and (MeSi)4S¢. include P4(NCHs)g, (OP)4(NCH3)¢, As4(NCH3)¢, and (MeSi) So. Eleven Vertices. Perhaps the only known eleven-atom cages are B,,H?y and ByQ2Hi1. Twelve Vertices. Twelve-atom cages are not widespread but play a dominant role in boron chemistry. The most highly symmetrical arrangement is the icosa- hedron (Fig. 2-8b), which has twelve equivalent vertices and J}, symmetry. Icosa- hedra of boron atoms occur in all forms of elemental boron, in B}2H?7, and in the numerous carboranes of the BjoC2H) type. A related polyhedron, the cubocta- hedron (2-LIII) is found in several borides of stoichiometry MB2. (2-L0T) General References Cotton, F. A., Chemical Applications of Group Theory, 2nd ed., Wiley-Interscience, 1971. Wells, A. F., Structural Inorganic Chemistry, 4th ed., Clarendon Press, 1975. Wyckoff, R. W. G., Crystal Structures, 2nd ed., Vols. 1-5, Wiley, 1963-1966. Encyclopedic and critical collection of structural data obtained by crystallography.CHAPTER THREE Introduction to Ligands and Complexes GENERAL REMARKS 3.1 Introduction It was known long before the Danish chemist S. M. Jgrgensen (1837-1914) began his extensive studies on the synthesis of such “complex” compounds that the metal halides and other salts could give compounds with neutral molecules and that many of these compounds could easily be formed in aqueous solutions. The recognition of the true nature of “complexes” began with Alfred Werner (1 866-1919) as set out in his classic work Neuere Anschauungen auf dem Gebiete der anorganischen Chemie (1906)!; he received the Nobel Prize for this work in 1913. Werner showed that neutral molecules were bound directly to the metal so that complex salts such as CoCl3-6NHs were correctly formulated [Co(NH3)6]3+Cly. He also demon- strated that there were profound stereochemical consequences of the assumption that the molecules or ions (ligands) around the metal occupied positions at the corners of an octahedron or a square. The stereochemical studies of Werner were later followed by the ideas of G. N. Lewis and N. V. Sidgwick, who proposed that achemical bond required the sharing of an electron pair. This led to the idea that a neutral molecule with an electron pair (Lewis base) can donate these electrons to a metal ion or other electron acceptor (Lewis acid). Well-known examples are the following: ' Braunschweig, 1905; English translation of second edition as New Ideas on Inorganic Chemistry, by E. P. Hedley, London, 1911. 6.62 INTRODUCTORY TOPICS & o N it ae Et. F H,N:—>Co+—:NH, No—scr ral Et F un f N Hs We can now define a ligand as any molecule or ion that has at least one electron pair that can be so donated. Ligands may also be called Lewis bases; in the terms used in organic chemistry, they are nucleophiles. Metal ions or molecules such as BF; with incomplete valence electron shells are Lewis acids or electrophiles. Although it is possible to regard even covalent compounds from the donor-ac- ceptor point of view—for example, we could regard methane (CH,) as composed of C4 and four H~ ions—it is not a particularly profitable or realistic way of looking at such molecules. Nevertheless, in inorganic chemistry, ions such as H~, F-, Cl-, NOj, and SO}, and groups such as CHy and C6H5, are commonly termed ligands even when they are bound in simple molecules by largely covalent bonds as in SF¢ or W(CH3)¢. Although SiF, is normally called a molecule and SiFR- a complex anion, the nature of the Si—F bonds in each species is essentially the same (See Section 2-7.). There are innumerable ways of classifying ligands. One useful approach is based on the type of bonding interaction between the central atom and its surrounding neighbor atoms. The bonding details are dealt with later, but the distinction between two major types of ligand can be illustrated by posing two questions. 1. Why do molecules like water or ammonia give complexes with ions of both main group and transition metals—for example, [Al(OH2)¢]** or [Co(NHa)6]>*, while other types of molecules such as PF3 or CO give complexes only with tran- sition metals? 2. Although PF; and CO give neutral complexes such as Ni(PF3)4 or Cr(CO)6, why do NH3, amines, oxygen compounds, and so on, not give complexes such as Ni(NH3)4? There are two main classes of ligands: (a) Classical or simple donor ligands act as electron-pair donors to acceptor ions or molecules, and form complexes with all types of Lewis acids, metal ions, or molecules. (6) Nonclassical ligands, 1 -bonding or 7-acid ligands, form compounds largely if not entirely with transition metal atoms. This interaction occurs because of the special properties of both metal and ligand. The metal has d orbitals that can be utilized in bonding; the ligand has not only donor capacity but also has acceptor orbitals. This latter distinction is perhaps best illustrated by comparison of an amine, :NR3, with a tertiary phosphine, :PR3. Both can act as bases toward H*, but the P atom differs from N in that it has vacant 3d orbitals of low energy, whereas inLIGANDS AND COMPLEXES. 63 N the lowest energy d orbitals are at far too high an energy to use. Another example is that of CO, which has no measurable basicity to protons, yet readily reacts with metals like nickel that have high heats of atomization to give compounds like Ni(CO)«. Ligands may also be classified electronically, that is, according to the number of electrons that they contribute to a central atom when these li: igands are regarded (sometimes artificially) as neutral species. Thus atoms or groups that can forma single covalent bond are regarded as one-electron donors—examples are F, SH, and CH3. Any compound with an electron pair is a two-electron donor (e.g.,:NHs, H20:). Groups that can form a single bond and at the same time donate can be considered to be three-electron donors. For example, the acetate ion can be either a one- or a three-electron donor, namely, A molecule with two electron pairs (¢.g., HXNCH2CH2NHz2) can be regarded as a four-electron donor, and so on. This classification method is useful in that it is an aid in electron counting, particularly for transition metal complexes whose va- lence shells contain eighteen electrons and whose stoichiometries correspond to what is called the eighteen-electron rule or noble gas formalism. This is merely a phenomenological way of expressing the tendency of a transition metal atom to use all its valence orbitals, namely, the five nd, the (n+ 1)s, and the three (n + 1)p orbitals as fully as possible in metal-ligand bonding. The sum of the number of valence electrons in the gaseous atom plus the number of electrons from neutral ligands may attain a maximum value of 18, This is illustrated by the following examples: Cr + 6CO = Cr(CO)« 6 + (6X 2) = 18 Ru + H + CO2CH; + 3PPhy = RuH(CO;Me)(PPhs)s 8+14+34+(3X2)=18 A third way of classifying ligands is structurally, that is, by the number of connections they make to the central atom. Where only one atom becomes closely connected (bonded) the ligand is said to be unidentate [e.g., the ligands in Co(NH3)é*, AICIz, Fe(CN)7]. When a ligand becomes attached by two or more atoms it is bidentate, tridentate, tetradentate, and so on, generally multidentate. Note that the Greek-derived corresponding prefixes (mono, ter, poly, etc.) are also used in the literature. Bidentate ligands when bound entirely to one atom are termed chelate, as in 3-I, 3-H, and 3-II1.64 INTRODUCTORY TOPICS th a a ee me \y Che M—0O Hi (31) (311) ($1) Another important role of ligands is as bridging groups. In many cases they serve as unidentate bridging ligands. This means that there is only one ligand atom that forms two (or even three) bonds to different metal atoms. For monoatomic ligands, such as the halide ions, and those containing only one possible donor atom, this unidentate form of bridging is, of course, the only possible one. A few examples are shown in 3-IV to 3-VI. Ligands having more than one atom that can be an electron donor often function as bidentate bridging ligands. Examples are shown in 3-VII and 3-VIII. H i 4 ' eee a Pd FE oe OC Mom Mo Alea Gv) @V) Vb (VID (vu A further classification is according to the nature of the donor atom of the ligand. ‘Thus we may have carbon, nitrogen, phosphorus, oxygen, sulfur, and so on, donor atoms. Ligands classed this way are discussed in Chapter 4. Under oxygen donors we can list not only HzO but also Ph3PO and SO}, and so on, whereas under carbon we can list CO, 7-CsHs, CH3CH=CHp, and so on. For nitrogen we have all the aliphatic and aromatic amines plus a host of others, such as NOZ and NO. Some unidentate ligands have two or more different donor sites so that the pos- sibility of linkage isomerism arises. Some important ligands of this type, which are called ambidentate ligands, are: ake Nitro (Moe Cyano M—ONO Nitrito M—NC Tsocyano {aes S-Thiocyanato RS—>M,RSO->M — Sand O-Bonded M—NCS N-Thiocyanato Dialkylsulfoxide I 0 The next sections discuss primarily the chemistry of simple donor ligands. Much of this chemistry applies equally well to main group metal ions (¢.g., Nat, Ca?*, Ga3*, or Cd2+) and to transition metal ions. It is a chemistry largely of aqua ions, 2 RJ, Balahura and N. A. Lewis, Coord. Chem. Rev., 1976, 20, 109; A. H. Norbury, Ade. Inorg. Chem. Radiochem., 1975, 17, 231 [NCO, NCS, and NCSe]; J. L. Burmeister, Coord. Chem. Rev., 1966, 1, 205; A. H. Norbury and A. P. Sinha, Q. Rev., 1970, 24, 69.LIGANDS AND COMPLEXES 65 nitrogen donor ligands, such as ammonia or ethylenediamine, and halide ions, and it is chemistry of metal ions in positive oxidation states, usually 2+ and 3+. Later sections consider complexes that have a-bonding ligands and also com- pounds that are called + complexes, which are those formed by unsaturated organic molecules. This is a chemistry largely of transition metals, often in formally low oxidation states such as ~1, 0, and +1. The borderline between x-bonding and non-7-bonding ligands is by no means clearly defined. Also the terms “nonclassical” versus “classical” are of limited validity, since Werner and his contemporaries studied complexes of cyanide ion and of tertiary phosphines; even pyridine, which they used extensively, is not solely a simple donor. STABILITY OF COMPLEX IONS IN AQUEOUS SOLUTION? 3-2. Aqua Ions In a fundamental sense metal ions simply dissolved in water are already com- plexed—they have formed aqua ions. The process of forming in aqueous solution what we more conventionally call complexes is really one of displacing one set of ligands, which happen to be water molecules, by another set. Thus the logical place to begin a discussion of the formation and stability of complex ions in aqueous so- lution is with the aqua ions themselves. From thermodynamic cycles the enthalpies of plunging gaseous metal ions into water can be estimated and the results, 2 X 102 to 4 X 103 kJ mol! (see Table 3-1 ), show that these interactions are very strong indeed. It is of importance in under- standing the behavior of metal ions in aqueous solution to know how many water molecules each of these ions binds by direct metal-oxygen bonds. To put it another way, if we regard the ion as being an aqua complex [{M(H20),]"*, which is then further and more loosely solvated, we wish to know the coordination number x and also the manner in which the x water molecules are arranged around the metal ion. Classical measurements of various types—for example, ion mobilities, apparent hydrated radii, entropies of hydration—fail to give such detailed information be- cause they cannot make any explicit distinction between those water molecules directly bonded to the metal—the x water molecules in the inner coordination sphere—and additional molecules that are held less strongly by hydrogen bonds to the water molecules of the inner coordination sphere. There are, however, ways of answering the question in many instances, ways depending, for the most part, on modern physical and theoretical developments. A few illustrative examples will be considered here. For the transition metal ions, the spectral and, to a lesser degree, magnetic * (a) J. Burgess, Metal fons in Solution, Horwood-Wiley, 1978 [covers most aspects of solution behavior (¢.g., thermochemistry, kinetics, solvation, nonaqueous solutions]; (b) C. F. Baes and R.E. Mesmer, Hydrolysis of Cations, Wiley-Interscience, 1976.66 INTRODUCTORY TOPICS TABLE 3-1 Enthalpies of Hydration? of Some Ions (kJ mol!) Ht 1091 Catt -1877 cat —1807 Lit ~519 srt -1443 Het —1824 Nat —406 Ba?* =1305 Sn2* -1552 K+ -322 crt -1904 Poet —1481 Rot -293 Mn?+ -1841 Apt —4665 cst -264 Fe? -1946 Fe+ —4430 Agt 473 Corr 1996 Fa 515 ai 326 Nitt =2105 cr —381 Be?+ —2894 cut =2100 Bro —347 Mg 1921 Zn -2046 is -305 2 Absolute values are based on the assignment of —1091 + 10 kJ mol~! to H* (cf. H. F. Halliwell and S. C. Nyburg, Trans. Faraday Soc., 1963, 59, 1126). Each value probably has an uncertainty of at least 10m kJ mol~!, where 1 is the charge of the ion. properties depend on the constitution and symmetry of their surroundings. For example, the Co!! ion is known to form both octahedral and tetrahedral complexes. Thus we might suppose that the aqua ion could be either [Co(H20)«]?* with oc- tahedral symmetry, or [Co(H20).]?* with tetrahedral symmetry. It is found that the spectrum and the magnetism of Col in pink aqueous solutions of its salts with noncoordinating anions such as CIO; or NO} are very similar to the corresponding properties of octahedrally coordinated Co!' in general, and virtually identical with those of Co! in such hydrated salts as Co(C1O4)2°6H20 or CoSOg-7H20 where from X-ray studies octahedral [Co(H2O)«]?* ions are known definitely to exist. Complementing this, the spectral and magnetic properties of the many known tetrahedral Co!l complexes, such as [CoCl4]?~, [CoBr4]?~, [Co(NCS)4]?~, and [py2CoCla], which are intensely green, blue, or purple, are completely different from those of Co! in aqueous solution. Thus there can scarely be any doubt that aqueous solutions of otherwise uncomplexed Co! contain predominantly* well- defined, octahedral [Co(H2O)¢]?+ ions, further hydrated, of course. Evidence of similar character can be adduced for many of the other transition metal ions. For all the di- and tripositive ions of the first transition series, the aqua ions are octa- hedral {M(H20)¢]2°" 3)* species, although in those of Cr!!, Mn!!!, and Cu" there are definite distortions of the octahedra because of the Jahn-Teller effect (see Section 20-18). Information on aqua ions of the second and third transition series, of which there are only a few, however, is not so certain. It is probable that the coordination is octahedral in many, but higher coordination numbers may occur. For the lanthanide ions, M3+(aq), it is certain that the coordination number is higher. For ions that do not have partly filled d shells, evidence of the kind mentioned is lacking, since such ions do not have spectral or magnetic properties related in a straightforward way to the nature of their coordination spheres. We are therefore not sure about the state of aquation of many such ions, although nmr and other 4 However, there are also small quantities of tetrahedral {Co(H20)4]?*; see T. J. Swift, Inorg. Chem., 1964, 3, 526.LIGANDS AND COMPLEXES 67 relaxation techniques have now supplied some such information. It should be noted that, even when the existence of a well-defined aqua ion is certain, there are vast differences in the average length of time that a water molecule spends in the coordination sphere, the so-called mean residence time. For Cr!!! and Rh" this time is so long that when a solution of [(Cr(H20)6]3* in ordinary water is mixed with water enriched in '$0, many hours are required for complete equilibration of the enriched solvent water with the coordinated water. From a measurement of how many molecules of H2O in the Cr!!! and Rh!" solutions fail immediately to exchange with the enriched water added, the coordination numbers of these ions by water were shown to be 6. These cases are exceptional, however. Most other aqua ions are far more labile, and a similar equilibration would occur too rapidly to permit the same type of measurement. This particular rate problem is only one of several that are discussed more fully in Section 28-3. Aqua ions are all more or less acidic3; that is, they dissociate in a manner rep- resented by the equation = LH*[M(H20),—1(OH)] (M(H20)) The acidities vary widely, as the following Ka values show: [M(H20)x]** = (M(H20)x—(OH)]°"9F + HF Ka, M in [M(H20)6]** Ka allt 1.12. 10-5 cen 1.26 X 10-4 Fell! 6.3 X 10-3 Coordinated water molecules in other complexes also dissociate in the same way, for example, [Co(NHs)5(H20)}>* = [Co(NH3)s(OH)]2+ + Ht K = 10757 [Pt(NH5)4(H20)2]** = (Pt(NH3)4(H20)(OH)}>+ +H* == K ~ 10-2 3-3. The “Stepwise” Formation of Complexes?S The thermodynamic stability of a species is a measure of the extent to which this species will form from, or be transformed into, other species under certain conditions 5 L.D. Pettit and G. Brookes, Essays Chem., 1977, 6, 1; M. T. Beck, Chemistry of Complex Equi- libria, Van Nostrand-Reinhold, 1970; 8. Fronaeus, in Techniques of Inorganic Chemistry, Vol. 1, Wiley, 1963; F. L. C. Rossotti and H. Rossotti, The Determination of Stability Constants, McGraw-Hill, 1961; L. G, Sillén and A. E. Martell, Eds., Stability Constants of Metal Jon Complexes, Chemical Society, London (volumes collecting stability constant data); R. M. Smith and A. E. Martell, Eds., Critical Stability Constants, several volumes, Vol. 1, 1974, Vol. II, 1976, Plenum Press; S. J. Ashcroft and C. J. Mortimer, Thermochemistry of Transitional Metal Com- plexes, Academic Press, 1970; J. J. Christensen, D. J. Eatough, and R. M. Izatt, Handbook of Metal Ligand Heats and Related Thermodynamic Quantities, 2nd ed., Dekker, 1975; J. Kragten, Atlas of Metal Ligand Equilibria in Aqueous Solution, Horwood-Wiley, 1977 (45 metals, 29 li- gands); D. D. Perrin, Stability Constants of Metal Ion Complexes, Part B, Organic Ligands, Pergamon Press, 1979 (5000 references).68 INTRODUCTORY TOPICS. when the system has reached equilibrium. The kinetic stability of a species refers to the speed with which transformations leading to the attainment of equilibrium will occur. This section considers problems of thermodynamic stability, that is, the nature of equilibria once they are established. If in a solution containing aquated metal ions M and unidentate ligands L, only soluble mononuclear complexes are formed, the system at equilibrium may be described by the following equations and equilibrium constants: M+L=ML Ki= IML} (MI{L] ML; oo [ML}IL} {ML} ML, + L = ML; K3= aS MEE) [MLy] MLy-; + L = ML, Pear eeieeeerereereeeseee a OSS Mint) There will be N such equilibria, where N represents the maximum coordination number of the metal ion M for the ligand L, and N may vary from one ligand to another. For instance, Al?+ forms AICI; and AIFg~ and Co®* forms CoCIz~ and Co(NH3)¢2*, as the highest complexes with the ligands indicated. Another way of expressing the equilibrium relations is the following: IML} M+L=ML = Osa) : IML} M+2L=ML, Bp iMIILE [M3] M + 3L = ML. i 0 TL it __[MLy] M+NL=MLy By TMy(L1 Since there can be only N independent equilibria in such a system, it is clear that the K;’s and the f;’s must be related. The relationship is indeed rather obvious. Consider, for example, the expression for 83. Let us multiply both numerator and denominator by [ML][ML»] and then rearrange slightly: = IMs) [MLI[MLa} {M][LP [ML]{ML3] _ IML (ML) IML) {M}{L] (ML][L] [ML2}{L] Bs It is not difficult to see that this kind of relationship is perfectly general, namely, Bu = KiK2K3..-LIGANDS AND COMPLEXES 69 The K;’s are called the stepwise formation constants (or stepwise stability constants), and the §;’s are called the overall. | formation constants (or overall sta- bility constants); each type has its special convenience in certain cases. Inall the equilibria above we have written the metal ion without specifying charge or degree of solvation. The former omission is obviously of no importance, for the equilibria may be expressed as above whatever the charges. Omission of the water molecules is a convention that is usually convenient and harmless. It must be re- membered when necessary. See, for example, the discussion of the chelate effect, in the next section. With only a few exceptions, there is generally a slowly descending progression in the values of the K;’s in any particular system. This is illustrated by the data* for the Cd!!-NH; system where the ligands are uncharged and by the Cd-CN- system where the ligands are charged. Ca? + NH3 = [Ca(NH3)}2* kK = 10265 [Cd(NHs)}?* + NH3 = [Cd(NHs)2]?* = 10210 [Cd(NHs)2]?* + NH3 = [Cd(NH3)3]?* kK = 101-44 (Cd(NH3)3]?* + NH3 = [Cd(NH5)3]* —_K = 10993 (8, = 107 i) Ca?+ + CN~ = [Ca(CN)]+ [Cd(CN)}* + CN~ = [C4(CN)2] [Cd(CN)2] + CN~ = [Cd(CN)3]- [Cd(CN)s]- + CN~ = [Cd(CN)4]?- = 10355 (8, = 10188) Thus, typically, as ligand is added to the solution of metal ion, ML is first formed more rapidly than any other complex in the series. As addition of ligand is continued, the ML2 concentration rises rapidly, while the ML concentration drops, then ML3, becomes dominant, ML and ML becoming unimportant, and so forth, until the highest complex MLy is formed, to the nearly complete exclusion of all others at very high ligand concentrations. These relationships are conveniently displayed in diagrams such as those shown in Fig. 3-1. A steady decrease in K; values with increasing / is to be expected, provided there are only slight changes in the metal-ligand bond energies as a function of i, which is usually the case. For example, in the Ni?+-NH; system to be discussed below, the enthalpies of the successive reactions Ni(NH3);-1 + NH3 = Ni(NHs3); are all within the range 16.7-18.0 kJ mol-!. There are several reasons for a steady decrease in K; values as the number of ligands increases: (1) statistical factors, (2) increased steric hindrance as the number of ligands increases if they are bulkier than the HO molecules they replace, (3) Coulombic factors, mainly in complexes with charged ligands. The statistical factors may be treated in the following way. Suppose, as is almost certainly the case for Ni?*, that the coordination number remains the same throughout the series ([M(H20)y] . .. [M(H20)v-nLa] ... (MLy]. The [M(H20)y-nLn] species has sites from which to lose a ligand, whereas the species [M(H2O)w-n+1Ly1] has *CA-NHs cofistants determined in 2M NHqNOs; Cd-CN~ constants determined in 3M NaCiO,.70 INTRODUCTORY TOPICS -8 6 4 ee oO tog {CN™] (a) (b) Fig. 3-1. Plots of the proportions of the various complexes [Cd(CN).]2-©1* as a function of the ligand concentration ee = (CA(CN).J/toral Cd Lae = E Can). =o [Reproduced by permission from F. J.C. Rossetti, in Modern Coordination Chemistry, J. Lewis and R.G. Wilkins, Eds., Interscience, 1960, p. 10.] (N —n + 1) sites at which to gain a ligand. Thus the relative probability of passing from [M(H20) y—n¢iLn—1] to [M(H20)v-nLnl is proportional to (N —n + 1)/n. Similarly, the relative probability of passing from [M(H20)y-nLba} to [M- (H20)y—-n—1Ln+i] is proportional to (N — n)/(n + 1). Hence on the basis of these statistical considerations alone, we expect N-n N-nt+h_ n(N- 7) ntl n (nF IN = 2+) In the Ni2+-NHs3 system (NV = 6), we find the comparison between experimental ratios of successive constants and those calculated from the formula above to be as shown in Table 3-2. The experimental ratios are consistently smaller than the statistically expected ones, which is typical and shows that other factors are also of importance. There are cases where the experimental ratios of the constants do not remain constant or change monotonically; instead, one of them is singularly large or small. There are several reasons for this: (1) an abrupt change in coordination number and hybridization at some stage of the sequence of complexes, (2) special steric effects that become operative only at a certain stage of coordination, and (3) an Kut /Kn = TABLE 3-2 ‘Comparison of Experimental and Statistical Formation Constants of Ni?+-NH3 Complexes Experimental Statistical KK, 0.28 0.417 KyK> 031 0.533 KyKs 0.29 0.562 KsJKy 0.36 - 0.533 KK 0.2 o4l7LIGANDS AND COMPLEXES. a abrupt change in electronic structure of the metal ion at a certain stage of com- plexation. Each of these is now illustrated. Values of K3/K are anomalously low for the halogeno complexes of mercury(II); HgXz species are linear, whereas [HgX4]?~ species are tetrahedral. Presumably the change from sp to sp? hybridization occurs on going from HgX2 to [HgX;3]~. K3/K7 is anomalously small for the ethylenediamine complexes of Zn", and this is believed to be due to the change from sp? to sp3d? hybridization if it is assumed that [Znen2]?+ is tetrahedral. For the Ag+-NH3 system K2 > Ki, indicating that the linear, sp-hybridized structure is probably attained with [Ag(NH3)2]* but not with [Ag(NH3)(H20)3¢or 5)]*. With 6,6’-dimethyl-2,2’-bipyridine (3-IX), many metal ions that form tris- 2,2’-bipyridine complexes form only bis or mono complexes, or, in some cases, no isolable complexes at all, because of the steric hindrance between the methyl groups and other ligands attached to the ion. SQ HC CH, (3-IX) In the series of complexes of Fe!! with 1,10-phenanthroline (and also with 2,2’- bipyridine), K3 is greater than K2. This is because the tris complex is diamagnetic (.e., the ferrous ion has the low-spin state 1$,—see Section 20-9 for the meaning of this symbol), whereas in the mono and bis complexes, as in the aqua ion, there are four unpaired electrons. This change from the t3,e? to the t§, causes the enthalpy change for addition of the third ligand to be anomalously large because the ee electrons are antibonding. 3-4, The Chelate Effect® The term “chelate effect” refers to the enhanced stability of a complex system containing chelate rings as compared to the stability of a system that is as similar as possible but contains none or fewer rings. As an example, consider the following equilibrium constants: Ni?+(aq) + 6NH3(aq) = [Ni(NH)6]?*(aq) log 8 = 8.61 Ni?*(aq) + 3en(ag) = [Ni ens]?*(aq) log 8 = 18.28 The system [Ni en3]?* in which three chelate rings are formed is nearly 10!° times as stable as that in which no such ring is formed. Although the effect is not always so pronounced, such a chelate effect is a very general one. To understand this effect, we must invoke the thermodynamic relationships: AG* = -RTInB AG = AH® ~ TAS? © (@)R.T. Myers, Inorg. Chem., 1978, 17, 952; (b) D. Munro, Chem. Br., 1977, 13, 100; (c) C. F. Bell, Principles and Applications of Metal Chelation, Oxford University Press, 1977.72 INTRODUCTORY TOPICS TABLE 3-3 ‘Two Reactions Illustrating a Purely Entropy-Based Chelate Effect Cd?*(aq) + 4CH3NH(aq) = [Cd(NH2CH3)4]?*(aq) log 8 = 6. Cd?+(aq) + 2H2NCH2CH2NH)(aq) = [Cd eng]?*(aq) log Ligands AH® (kJ mol!) AS® (J mol“! deg-!)_ —TAS°(kJ mol~!)_AG® (kJ mol~!) 4CH3NH> -373 -673 20.1 37.2 2en -56.5 +141 4.2 -60.7 Thus @ increases as AG® becomes more negative. A more negative AG? can result from making AH® more negative or from making AS° more positive. As a very simple case, consider the reactions, and the pertinent thermodynamic data for them, given in Table 3-3. In this case the enthalpy difference is well within experimental error; the chelate effect can thus be traced entirely to the entropy difference. In the example first cited, the enthalpies make a slight favorable contribution, but the main source of the chelate effect is still to be found in the entropies. We may look at this case in terms of the following metathesis: [Ni(NH5)6]?*(aq) + 3 en(aq) = [Ni ens}?*(aq) + 6NH3(aq) log 8 = 9.67 for which the enthalpy change is 12.1 kJ mol~!, whereas —TAS® = —55.1 kJ mol-!. The enthalpy change corresponds very closely to that expected from the increased CFSE* of [Ni en3]?+ which is estimated from spectral data to be -11.5 kJ mol! and can presumably be so explained. Asa final example, which illustrates the existence of a chelate effect despite an unfavorable enthalpy term, we may use the reaction [Ni ena(H2O)2]?*(aq) + tren(aq) = [Ni tren(H2O)2]?*(aq) + 2en(aq) log B = 1.88 [tren = N(CH2CH2NH2)s] For this reaction we have AH® = +13.0,—TAS® = —23.7, and AG® = —10.7 (all in kJ mol"!). The positive enthalpy change can be attributed both to greater steric strain resulting from the presence of three fused chelate rings in Ni tren, and to the inherently weaker M—N bond when N isa tertiary rather than a primary ni- trogen atom. Nevertheless, the greater number of chelate rings (3 vs. 2) leads to greater stability, owing to an entropy effect that is only partially canceled by the unfavorable enthalpy change. Probably the main cause of the large entropy increase in each of the three cases we have been considering is the net increase in the number of unbound mole- cules—ligands per se or water molecules. Thus although 6 NH; displace 6 H,0, making no net change in the number of independent molecules, it takes only 3 en molecules to displace 6 HO. Another more pictorial way to look at the problem is to visualize a chelate ligand with one end attached to the metal ion. The other end cannot then get very far away, and the probability of it, too, becoming attached * For meaning of CFSE see Section 20-19.LIGANDS AND COMPLEXES 23 TABLE 3-4 Factors Influencing Solution Stability of Complexes? Enthalpy effects Entropy effects Variation of bond strength with electronegativities of metal ions and ligand donor atoms Ligand field effects Steric and electrostatic repulsion between ligands in complex Enthalpy effects related to conformation of uncoordinated ligand Other Coulombic forces involving chelate ring formation Enthalpy of solution of ligands ‘Change in bond strength when ligand is charged Number of chelate rings Size of chelate ring ‘Changes of solvation on complex formation Arrangement of chelate rings Entropy variations in uncoordinated ligands Effects resulting from differences in configurational entropies of the ligand in complex compound Entropy of solution of ligands Entropy of solution of coordinated metal ions (same donor and acceptor atom) @ From R. T. Meyers, Inorg. Chem., 1978, 17, 952. to the metal atom is greater than if this other end were instead another independent molecule, which would have access to a much larger volume of the solution. The latter view provides an explanation for the decreasing magnitude of the chelate effect with increasing ring size, as illustrated by data such as those shown below for copper complexes of HyN(CH)2NH and H2N(CH2)3;NH2(tn): [Cu ena]?*(aq) + 2tn(aq) = [Cu tns]?*(aq) + 2en(aq) log B = -2.86 Of course, when the ring that must be formed becomes sufficiently large (seven- membered or more), it becomes more probable that the other end of the chelate molecule will contact another metal ion than that it will come around to the first one and complete the ring. Table 3-4 summarizes the factors influencing the sta- bilities of complexes. 3-5. The Macrocyclic Effect” Just as a chelating n-dentate ligand gives a more stable complex (more negative AG? of formation) than n unidentate ligands of similar type, a phenomenon just discussed under the name chelate effect, so an n-dentate macrocyclic ligand (see Section 3-6) gives even more stable complexes than the most similar n-dentate open chain ligand. For example: 7 F.P. Hinz and D. W. Margerum, J. Am, Chem. Soc.. 1974, 96, 4993; Inorg. Chem., 1974, 13, 2841; G. F. Smith and D. W. Margerum, J. C. S. Chem. Comm., 1975, 80774 INTRODUCTORY TOPICS No ON c. ’ N log K = 5.2 AG = -30 kJ mol"! at 300°K Just as in the case of the chelate effect, this so-called macrocyclic effect might result from either entropic or enthalpic contributions, or both, and similar con- troversy about their relative importance has arisen. The most recent results® indicate that entropy always favors the macrocycle, with a value of ca. 70 J mol~! deg-! for systems such as that shown above. The enthalpy contribution usually also favors the macrocycle, but by an amount that can vary a great deal from case to case. In the specific case shown, AH = —10 kJ mol~'. In general, the macrocyclic effect results from a favorable entropy change assisted, usually, by a favorable enthalpy change as well. TYPES AND CLASSIFICATION OF LIGANDS 3-6. Multi- or Polydentate Ligands Regardless of whether 7-bonding is involved, ligands can have various denticities, and we now illustrate some of the more important types. Bidentate Ligands. These are very common and can be classified according to the size of the chelate ring formed as in the following examples: R a NR, c ‘Three-membered ut : “I A(X) > ACSA) > AQAA) However this is not the actual order because enthalpy differences between diaste- reomers are rather small (2-3 kJ mol!), and an entropy factor must also be con- sidered. Entropy favors the 6A and 6A species because they are three times as probable as the 665 and NX ones. Hence the best estimate of relative stabilities, which in fact agrees with all experimental data, becomes A(55N) > A(55S) = A(SAA) > ACAAD)LIGANDS AND COMPLEXES 81 4H N, | S c H Fig. 3-5. The absolute configuration and expected conformation (i.e., with an equatorial CH 3group) for an M(J-pn) chelate ring. In crystalline compounds, the A(546) isomer (or its enantiomorph) has been found most often, but the other three have also been found. These crystallographic results probably prove nothing about the intrinsic relative stabilities, since hydrogen bonding and other intermolecular interactions can easily outweigh the small in- trinsic energy differences. Nor studies of solutions of Ru", Pt!¥, Ni, Rh!, Ir!ll, and Coll! [M en3]"* complexes have yielded the most useful data, and the general conclusions seem to be that the order of stability suggested above is correct and that ring inversions are very rapid. Both experiment and theory suggest that the barrier to ring inversion is only about 25 kJ mol~'. Thus the four diastereomers of each overall form (A or A) are in labile equilibrium. One of the interesting and important applications of the foregoing type of analysis is to the determination of absolute A or A configurations by using substituted ethylenediamine ligands of known absolute configuration. This is nicely illustrated by the [Co(/-pn)3]>* isomers. The absolute configuration of /-pn [pn = 1,2-di- aminopropane NHyCH(CH3)—CH»NH)] is known. It would also be expected from consideration of repulsions between rings in the tris complex (as indicated in Fig. 3-4) that pn chelate rings would always take a conformation that puts the CH group in an equatorial position. Hence, an /-pn ring can be confidently ex- pected to have the 5 conformation shown in Fig. 3-5. Note that because of the ex- treme unfavorability of having axial CH3 groups, only two tris complexes are ex- pected to occur, namely, A(655) and A(656). But by the arguments already ad- vanced for en rings, the A isomer should be the more stable of these two, by 5 to 10 kJ mol™'. Thus we predict that the most stable [Co(/-pn)3]>+ isomer must have the absolute configuration A about the metal. In fact, the most stable [Co(/-pn)3]3+ isomer is the one with + rotation at the sodium-D line, and it has the same circular dichroism spectrum, hence the same absolute configuration as (+)-[Co en3]3+. The absolute configuration of the latter has been determined, and it is indeed A. Thus the argument based on conforma- tional analysis is validated. -ACID OR 1-BONDING LIGANDS: 7 COMPLEXES The ligands for which -bonding is important are carbon monoxide, isocyanides, substituted phosphines, arsines, stibines or sulfides, nitric oxide, various molecules82 INTRODUCTORY TOPICS with delocalized -orbitals, such as pyridine, 2,2’-bipyridine, 1,10-phenanthroline, and with certain ligands containing 1,2-dithioketone or 1,2-dithiolene groups, such as the dithiomaleonitrile anion. Very diverse types of complex exist, ranging from binary molecular compounds such as Cr(CO)s or Ni(PF3)4 through mixed species such as Co(CO);3NO and (C¢Hs)3PFe(CO)s, to complex ions such as [Fe(CN)sCO}-, [Mo(CO)sI]-, [Mn(CNR)o]*+, [Vphens]*, and INi[S2C2(CN)2]2}?-. In many of these complexes, the metal atoms are in low-positive, zero, or negative formal oxidation states. It is a characteristic of the ligands that they can stabilize low oxidation states. This property is associated with the fact that in addition to lone pairs, these ligands possess vacant x orbitals. These vacant orbitals accept electron density from filled metal orbitals to form a type of « bonding that sup- plements the « bonding arising from lone-pair donation; high electron density on the metal atom—of necessity in low oxidation states—thus can be delocalized onto the ligands. The ability of ligands to accept electron density into low-lying empty x orbitals can be called + acidity, the word “acidity” being used in the Lewis sense. There are many unsaturated organic molecules and ions that are also capable of forming complexes with transition metals in low oxidation states, and these are called + complexes. There is a qualitative difference from 7-acid ligands. The latter form bonds to the metal involving o orbitals and 7 orbitals whose nodal planes include the axis of the o bond. For the 7-complexing ligands such as alkenes, arenes, and allyl groups, both the donation and back-acceptance (see below) of electron density by the ligand are accomplished using ligand 7 orbitals. The metal is thus out of the molecular plane of the ligand, whereas with 7-acid ligands the metal atom lies along the axes of the linear ligands or in the plane of planar ones. In a third class of ligand that involves bonding there are metal-oxygen or metal-nitrogen multiple bonds, as in O=VClz, MnOz, and N=OsO3. Here the electron flow is in the opposite sense to that in the bonding of w-acid ligands (i.e., from p orbitals on O or N to the metal d orbitals). We now consider bonding in the main classes of + bonding in a pteliminary, qualitative way; a more sophisticated approach is given in Chapter 20. a-BONDING LIGANDS 3-8. Carbon Monoxide Carbon monoxide is the most important 7-bonding ligand. Thousands of compounds ranging from pure carbonyls like Cr(CO), to mixed complexes like RuH(CO)- (PPh3)3 or [Ru(CO)Cls]?~ are known. The chemistry of carbonyls is discussed in extenso in Chapter 25. The bonding of CO (and of other similar 7 acids) can be regarded as involving the following contributions:LIGANDS AND COMPLEXES 83 Fig. 3-6. (a) The formation of the metal = carbon o bond using an unshared pair on the C atom. (b) ‘The formation of the metal —» carbon x bond. The other orbitals on the CO are omitted for clari 1, Overlap of a filled carbon o orbital with a o-type orbital on the metal atom as in Fig. 3-6a. Electron flow C — M in such a dative overlap would lead to an unacceptable concentration of electron density on the metal atom when the latter is not a +2 or more highly charged ion. The metal therefore attempts to reduce this charge (Pauling’s electroneutrality principle) by pushing electrons back to the li- gand. This of course is possible only if the ligand has suitable acceptor orbitals. 2. A second dative overlap of a filled dz or hybrid dp metal orbital with the empty, pm orbital on carbon monoxide, which can act as a receptor of electron density (Fig. 3-6b). This bonding mechanism is synergic, since the drift of metal electrons, referred to as “back-bonding”, into CO orbitals, will tend to make the CO as a whole neg- ative, hence to increase its basicity via the o orbital of carbon; at the same time the drift of electrons to the metal in the o bond tends to make the CO positive, thus enhancing the acceptor strength of the orbitals. Thus up to a point the effects of o-bond formation strengthen the 7 bonding, and vice versa. It may be noted here that dipole moment studies indicate that the moment of an M—C bond is only very low, about 0.5 D, suggesting a close approach to electroneutrality. The main lines of physical evidence showing the multiple nature of the M—CO bonds are bond lengths and vibrational spectra. According to the preceding de- scription of the bonding, as the extent of back-donation from M to CO increases, the M—C bond becomes stronger and the C=O bond becomes weaker. Thus, the multiple bonding should be evidenced by shorter M—C and longer C—O bonds as compared to M—C single bonds and C=O triple bonds, respectively. Actually very little information can be obtained from the CO bond lengths, because in the range of bond orders (2-3) concerned, CO bond length is relatively insensitive to bond order. The bond length in CO itself is 1.128 A, while the bond lengths in metal carbonyl molecules are ~1.15 A, a shift in the proper direction but of little quan- titative significance owing to its small magnitude and the uncertainties (~0.02 A) in the individual distances. For M—C distances, the sensitivity to bond order in84 INTRODUCTORY TOPICS © 0 N00 Hee oc } The poorer CS multiple bonding in M—CS compounds means that the C atom in M—CS is markedly more electrophilic than the C atom in M—CO. This leads to greater reactivity toward electrophiles at sulfur and to nucleophiles at carbon. There are reactions'® such as diphos;(CO)WCS + RSOSF = diphos,(CO)WCSR* + SO3F~ (CO)sWCS + H2NR = (CO)sWC=NR + H2S oN: a-CsHsRu(CO)2CS* + Ny = [1-CsHsRu(CO)C(S)Ns] — n-CsHsRu(CO)2NCS IrH(CS)(PPhs)3 + 2H2 = IrHx(SCH3)(PPh3)3 IrCly(CS)(CO)(PPhs)} + BHs — IrCl(SCH)(CO)(PPhs)2 Bridged CS complexes!” such as 3-XXVI and 3-XXVII have also been made as well as a bridge type not known for CO, that is, diphos»(CO) WCSW(CO)di- phos». s s Q i oc I No nae Ninn oc ec a “So on™ C7 SO I I 8 8 (3XXVD (-XXVI) 3-9. Dinitrogen!® The dinitrogen and carbon monoxide molecules are isoelectronic, but although metal carbonyls were discovered by Mond in 1890, the first dinitrogen complex '4 1M. Butler, Ace. Chem. Res., 1977, 10, 359; P. V. Yaneff Coord. Chem. Rev., 1977, 23, 183; G. Gattow and W, Behrend, Carbon Sulfides and Their Inorganic and Complex Compounds, Thieme, 1977, 15D. L. Lichtenberger and R. F. Fenske, Inorg. Chem., 1976, 18, 2015; M. A. Andrews, Inorg. Chem., 1977, 16, 496; I. S. Butler ef af., Inorg. Chem., 1976, 18, 2602. 16 R. J, Angelici et al., Inorg. Chem., 1978, 17, 1634; T. J. Collins and W. R. Roper, J. Orgomet. Chem., 1978, 189, 73; F. Faraone et al., J. C. S. Dalton, 1979, 931 17 RJ. Angelici er al., J. Organomet. Chem., 1978, 160, 231; Inorg. Chem., 1977, 16, 1173; A. Blraty et al., Inorg. Chem., 1977, 16, 3124; R. E. Wagner et al., J. Organomet. Chem., 1978, 148, C35. 18D. Sellman, Angew. Chem., Int, Ed., 1974, 13, 639; A. D. Allen, Chem. Rev., 1973, 73, 11; J. Chatt and G. J. Leigh, Chem. Soc. Rev., 1971, 1, 121; W.G, Zumft, Struct. Bonding, 1976, 29, 1; J. Chatt, J.R. Dilworth, and R. L. Richards, Chem. Rev., 1978, 78, 589.LIGANDS AND COMPLEXES 87 was discovered only in 1965 by Allen and Senoff. Some chemistry of dinitrogen complexes is discussed in Section 29-17. The bonding in linear M—N—N groups is qualitatively similar to that in terminal M—CO groups; the same two basic components, M «~ N2 o donation and M — Np @ acceptance, are involved. The major quantitative differences, which account for the lower stability of Nx com- plexes, appear to arise from small differences in the energies of the MO’s of CO and N>. For CO the o-donor orbital is weakly antibonding, whereas the corre- sponding orbital for N> is of bonding character. Thus Nz is a significantly poorer a donor than is CO. Now it is observed that in pairs of N2 and CO complexes where the metal and other ligands are identical, the fractional lowerings of Ny and CO frequencies are nearly identical. For the CO complexes, weakening of the CO bond, insofar as electronic factors are concerned, is due entirely to back-donation from metal d7-orbitals to CO 7* orbitals, with the o donation slightly canceling some of this effect. For Nz complexes, on the other hand, N=N bond weakening results from both o donation and m back-acceptance. The very similar changes in stretching frequencies for these two ligands suggests then that N2 is weaker than CO in both its o-donor and 7-acceptor functions. This in turn would account for the poor stability of Nz complexes in general. Terminal dinitrogen compounds have N—N stretching frequencies in the region 1930-2230 cm~! (N2 has v = 2331 cm-!). Bridging by the Nz molecule is different from that of CO, and the following types!? are known. Symmetric linear M—N—N—M Asymmetric linear = M!—N—N—M? Bent N Perpendicular mM—||—M N For dinitrogen there is a further binding mode that does not occur with CO, save possibly in transition states in reactions and that is sideways binding. This bonding ussed later when other molecules that bond sideways are treated (Section 3-10. Trivalent Phosphorus Compounds”? Compounds such as PFs, PCls, P(CsHs)3, and P(OCHs)s, as well as corresponding derivatives of arsenic and antimony are important 1-bonding ligands. In particular, 19M. Mercer, J. C. S. Dalton, 1974, 1637; R. D. Sanner et al., J. Am. Chem. Soc., 1976, 98, 8351, 8358; K. Jonas, Angew. Chem., Int. Ed., 1973, 12, 997; C. Kruger and Y.-H. Tsay, Angew. Chem., Int, Ed., 1973, 12, 998; K. Jonas et al., J. Am. Chem. Soc., 1976, 98, 74. 20 (a) C. A. Tolman, Chem. Rev., 1977, 77, 313, Extensive review with >!P nmr and other data; (b) J. Emsley and D. Hall, The Chemistry of Phosphorus, Wiley, 1976; (c) R. Mason and D. W. Meek, Angew. Chem,, Int. Ed., 1978, 17, 183. (See also references in Section 4-20).88 INTRODUCTORY TOPICS Empty dys orbital Zz Filled orbital Overlap Fig. 3-8. The back-bonding from a filled metal d orbital to an empty phosphorus 3d orbital in the PX; ligand taking the internuclear axis as the z axis. An exactly similar overlap occurs in the yz plane using the dy. orbitals. PF; gives many compounds comparable to those given by CO. Tertiary phosphines and phosphites are much stronger Lewis bases than CO or PF3 and will give com- plexes with non-transition-metal acceptors as well as being easily protonated. Tertiary phosphines and arsines also give many compounds with metals in high positive oxidation states [e.g., ReCl3(PPh3)3 or PtClg(PMe2Ph)2]. In such com- pounds M—P bond lengths show no evidence for 7-bonding, though the extent of x-bonding presumably increases as the oxidation state is lowered. The -bonding is shown in Fig. 3-8; it differs from that for CO in that the m-acceptor orbitals are the phosphorus 3d orbitals. Hence the bonding can be designated as da —d- whereas that of CO is dx—pz. The extent of both donation from the lone pair on the P atom and back-donation depends on the nature of the groups attached to P. For PH3 and P(alkyl)s, a-ac- ceptor ability is very low, but it becomes important with more electronegative groups. Analogous PX3, AsX3, and SbX3 compounds differ very little, but the li- gands having a nitrogen atom, which lacks orbitals, cause significantly lower frequencies for the CO vibrations, as indicated by the CO stretching frequencies (cm7') in the following series of compounds (PC1;)3Mo(CO)s 2040, 1991 (AsCls)sMo(CO); 2031, 1992 (SbCIy)3Mo(CO); 2045, 1991 dien Mo(CO); 1898, 1758 The pronounced effect of the electronegativity of the groups X is shown by the following CO stretching frequencies [(C2Hs)aP]sMo(CO)s 1937, 1841 [(CoHs0)5P];Mo(CO)3 1994, 1922 [Cla(C.Hs0)P];Mo(CO); 2027, 1969 (ClsP);Mo(CO)3, 2040, 1991 (FsP)sMo(CO)s 2090, 2055 The most electronegative substituent, F in PF3, will reduce very substantially the o-donor character so that there will be less P -> M electron transfer, and Mda —>LIGANDS AND COMPLEXES 89 Pdr transfer should be aided. The result is that PF; and CO are quite comparable in their 7-bonding capacity. Based on comparative spectroscopic data m ligands can be arranged in order of decreasing = acidity: PF; > PCI; ~ AsCly ~ SbCly > PCI(OR) > PCLR > PCIOR)2 ~ P(OR)s > PR3~ ARs ~ SbR3 X-Ray structure determination shows that in (PhO)3PCr(CO)s the P—Cr bond is 0.11 A shorter than in (Ph3P)Cr(CO)s, confirming the greater 7 acidity of phosphites. Spectroscopic data of other types have been used to confirm the extent of a bonding in PX3 compounds. Of at least as great importance to the chemistry of PX3 compounds as the elec- tronic factors are steric factors. Indeed these may be more important or even dominating in determining the stereochemistries and structures of compounds. Steric factors also affect rates and equilibria of dissociation reactions such as eq. 3-1 and the propensity of phosphine complexes to undergo oxidative-addition re- actions (Chapter 29). -PR =P. Pd(PR3)s — Pd(PR3)3 — Pd(PR3)2 GB-1) FPR) FPR, The stereochemistry of phosphine ligands is the prime factor in many highly se- lective catalytic reactions of phosphine complexes, such as hydroformylation and asymmetric hydrogenation (Chapter 30) The steric effects?0* can be correlated with an easily measured parameter—the cone angle @ defined by a conical surface as in 3-XX VIII assuming a metal-phos- phorus bond length of 2.28 A (these are quite constant) that can just enclose the van der Waals surface of all ligand atoms over all rotational orientations about the M—P bond.* Triphenylphosphine has @ = 184 + 2° and P(OCH3)3 has @ = 107 + 2°. It might have been expected that compounds with smaller cone angles would be better ligands, but since such compounds are stronger bases, it is not always easy to distinguish steric from electronic factors. However increasing the cone angle by having bulky groups tends to favor (a) lower coordination numbers, (b) the formation of less sterically crowded isomers, and (c) increased rates and equilibria ($XXVID. * A more sophisticated treatment calculates actual “ligand profiles,” since the ligands are not regular cones. See G. Ferguson et al., Inorg. Chem., 1978, 17, 2965; J. D. Smith and J. D. Oliver, Inorg. Chem., 1978, 17, 2585. For an attempt to separate electronic and steric factors, see M. Zahres et al., Angew. Chem, Int. Ed., 1979, 18, 40190 INTRODUCTORY TOPICS TABLE 3-5 Dissociation Constants and @ Values for Phosphine Complexes of Nickel L Ka,M ae) P(OEt); < 10-10 109 PMe3, <10-% 118 »(o> P(OMe); > PPh; > AsPh3 and for the reaction? x Nils 22 Nil; +L we have the values in Table 3-5. The dissociation of bulky phosphine ligands has profound consequences in catalytic reactions, since it can provide sites on a metal atom in which substrates can come together and react (Chap. 29). Very bulky phosphines (e.g., with t-butyl groups) can also promote other effects such as (a) barriers to rotation about M—P bonds, (5) stabilization of unusual low coordination numbers and valencies, and (c) propensity to undergo cyclometallation reactions (Section 29-6).?? The importance of acceptance in stabilizing bonds from poor ¢ donors to metal atoms is dramatically illustrated by the fact that (CH3)3S*, which is isoelectronic with (CH3)5P, but certainly an exceedingly poor donor because of the positive charge, forms bonds to metal atoms and appears comparable to PCI; in its 7 acidity.?3 3-11. Nitric Oxide?* The NO molecule is closely akin to CO except that it contains one more electron, which occupies a 1* orbital. Consistently with the general similarity of CO and NO, they form many comparable complexes, although, as a result of the presence 21 T, Takayanagi, H. Yamamoto, and T. Kwen, Bull. Soc. Chem. Jpn., 1975, 48, 2618. 22 For examples, see B. L. Shaw et al., J. C. S. Dalton, 1977, 2285; 1978, 257. 23, RD. Adams and D. F. Chodosh, J. Am. Chem, Soc., 1978, 100, 812. 24 (a) F, Bottomley, Ace. Chem. Res., 1978, 11, 158; (b) R. Eisenberg and C. D. Meyer, Acc. Chem. Res., 1975, 8, 26; (c) J. H. Enemark and R. D. Feltham, Coord. Chem. Rev., 1974, 13, 339; (d) N.G. Connelly, Inorg. Chim. Acta Rev., 1972, 6, 48; (¢) K. G.Caulton, Coord. Chem. Rev., 1975, 14, 317,LIGANDS AND COMPLEXES 91 of the additional electron, NO also forms a class (bent MNO) with no carbonyl analogues. Linear, Terminal MNO Groups. Just as the CO group reacts with a metal atom that presents an empty o orbital and a pair of filled dz orbitals, as illustrated in Fig. 3-6, to give a linear MCO grouping with a C > M o-bond and a significant degree of M - C z-bonding, so the NO group engages ina structurally and elec- tronically analogous reaction with a metal atom that may be considered, at least formally, to present an empty o orbital and a pair of dz orbitals containing only three electrons. The full set of four electrons for the Mda —> 7*(NO) interactions is thus made up of three electrons from M and one from NO. In effect, NO con- tributes three electrons to the total bonding configuration under circumstances where CO contributes only two. Thus for purposes of formal electron “bookkeep- ing,” the ligand NO can be regarded as a three-electron donor in the same sense as the ligand CO is considered a two-electron donor. This leads to the following very useful general rules concerning stoichiometry, which may be applied without specifically allocating the difference in the number of electrons to any particular {i.e., or 7) orbitals: 1. Compounds isoelectronic with one containing an M(CO), grouping are those containing M’(CO),—1(NO), M’"(CO),—2(NO)2, and so on, where M’, M”, and so on, have atomic numbers that are 1, 2,. . . ,etc. less than M. Some examples are: (n-CsHs)CuCO, (7-CsHs)NiNO; Ni(CO)4, Co(CO);NO, Fe(CO)2(NO)», Mn(CO)(NO)3; Fe(CO)s, Mn(CO)4NO. 2. Three CO groups can be replaced by two NO groups. Examples of pairs of compounds so related are Fe(CO),, Fe(CO),(NO), Mn(CO), NO, Mn(CO)\(NO), cr(CO), Cr(NO), Et gt s a Ns ONDF (Orr SFO, (OND, aon Et Et It should be noted that the designation “linear MNO group” does not disallow a small amount of bending in cases where the group is not in an axially symmetric environment, just as with terminal MCO groups. Thus MNO angles of 161° to 175° ‘may be found in “linear” MNO groups. Truly “bent MNO groups” have angles of 120 to 140° (see below). In compounds containing both MCO and linear MNO groups, the M—C and M—N bond lengths differ by a fairly constant amount, ~0.07 A, approximately equal to the expected difference in the C and N radii, and suggest that under comparable circumstances M—CO and M—NO bonds are typically about equally92 INTRODUCTORY TOPICS strong. In a chemical sense the M—N bonds appear to be stronger, since substi- tution reactions on mixed carbonyl nitrosyl compounds typically result in dis- placement of CO in preference to NO. For example, Co(CO)3NO reacts with a variety of R3P, X3P, amine, and RNC compounds, invariably to yield the Co- (CO)2(NO)L product. The NO vibration frequencies for linear MNO groups substantiate the idea of extensive M—N bonding, leading to appreciable population of NO 7* orbitals. Both the NO and Of species contain one 7* electron and their stretching frequencies are 1860 and 1876 cm”, respectively. Thus the observed frequencies in the range 1800-1900 cm, which are typical of linear MNO groups in molecules with small or zero charge, indicate the presence of approximately one electron pair shared between metal dx and NO r* orbitals. Bridging NO Groups. These occur less commonly than bridging CO groups, but there are the same two types, doubly and triply bridging. A triply bridging NO group occurs in the compound (7-CsHs)3Mn3(NO),, which also contains three doubly bridging NO’s.25 A symmetrical doubly bridging NO group occurs in (a-CsHs)(NO)Cr(¢-NO)(u-NH2)Cr(NO)(7-CsHs) and quite unsymmetrical doubly bridging NO groups occur in 9-CsHs(NO2)Mn(u-NO)2Mn(n-CsHs)NO. Just as with the corresponding types of bridging CO groups, the NO stretching frequencies decrease with the extent of the bridging. Thus in (7-CsHs)3Mn3(NO)4 there are two bands due to the doubly bridging NO groups at 1543 and 1481 cm7! and one from the triply bridging group at 1320 cm™!. In (7-CsHs)(NO)Cr(u- NO)(u-NH2)Cr(NO)(7-CsHs) the terminal NO groups absorb at 1644 cm7! and the bridging group has a frequency of 1505 em=!. Bridging NO groups are also to be regarded as three-electron donors. The doubly bridging ones may be represented as ‘Nu where the additional electron required to form two metal-to-nitrogen single bonds is supplied by one of the metal atoms. The situation is formally quite analogous to that for bridging halogen atoms. Bent, Terminal MNO Groups. It has long been known that NO can form single bonds to univalent groups such as halogens and alkyl radicals, affording the bent species N=O: and N=0: x R Metal atoms with suitable electron configurations and partial coordination shells may bind NO in a similar way. This type of NO complex is formed when the in- completely coordinated metal ion L,M would have a t§eg configuration, thus being prepared to form one more single o bond. Table 3-6 lists some compounds in which 25 RC. Elder, Inorg. Chem., 1974, 13, 1037.LIGANDS AND COMPLEXES 93 TABLE 3-6 Some Compounds with Bent MNO Groups 4£MNO (deg) ¥no (om) {Co en,Cl(NO)]* 121 1611 [IrC(CO)(PPh3)2NO]* 124 1680 IrCl(NO)(PPh3)2 123 1560. Irl(CH3)(NO)(PPhs)2 120 1525 [RuCl(NO)(PPh3)2NO]* 136 16872 CofS2CN(CH3)2]2(NO) 139 1626 # The other NO group is of the linear MNO type and its stretching frequency is 1845 cm~!. this type of M—NO structure has been demonstrated by X-ray crystallog- raphy. The NO stretching frequencies for the authenticated cases fall in the range 1525-1690 cm™!, that is, generally lower than those for linear MNO systems, except perhaps when the latter occur in anionic complexes, such as [Cr(CN)s(NO)]*- (yo = 1515 cm"). Tentatively, at least, this may be used as a criterion of structure type. In organic nitroso compounds, RNO, no is generally found in the 1500-1600 cm~! range. Finally it may be noted that although many CS compounds are known, there are so far only a few with M—NS links.”6 These have been made by nucleophilic attacks of nitrido complexes on Ss, for example, (Et,NCS;);Mo=N + S = (ENCSz);MoNS or by the reaction Na[CsHsCr(CO)3] + '4SsN3Cly = CsHsCr(CO),NS + CO + NaCl In (y-CsHs)Cr(CO) NS the Cr—N—S group is essentially linear. 3-12. Isocyanides?7 Tsocyanide complexes can be obtained by direct substitution reactions of the metal carbonyls and in other ways. They include such crystalline, air-stable compounds as red Cr(CNPh)¢, white [Mn(CNCHs3)¢]I, and orange Co(CO)(NO)(CNC;7H;)2, all of which are soluble in benzene. Isocyanides generally appear to be stronger o-donors than CO, and various complexes such as [Ag(CNR)a]*, [Fe(CNR)6]?*, and [Mn(CNR)«]** are known where x bonding is of relatively little importance; derivatives of this type are not known for CO. However the isocyanides are capable of extensive back-acceptance of 7 electrons from metal atoms in low oxidation states. This is indicated qualita- tively by their ability to form compounds such as Cr(CNR)g and Ni(CNR)4, analogous to the carbonyls and more quantitatively by comparison of CO and CN stretching frequencies. As shown in Table 3-7, the extent to which CN stretching 2 P., Legzdins er al, J. C. S. Chem. Comm., 1978, 1036; J. Am. Chem. Soc., 1978, 100, 2247; J. Chat et al., J.C.S. Dalton, 1979, 1 27 P.M. Treichel, Adv. Organomet. Chem., 1973, 11, 21; F. Bonati and G. Minghetti, Inorg. Chim. Acta, 1974, 9, 95.94 INTRODUCTORY TOPICS TABLE 3-7 Lowering of CO and CN Frequencies in Analogous Compounds, Relative to Values for Free CO and CNAr? Molecule? ‘Av (m=) for cach fundamental mode Cr(CO)s 4B 123, 160 Cr(CNAN) 68 140 185, Ni(CO)s 15 106 Ni(CNAF)« 70 125 # Ar represents CoHs and p-CHs0CcH« ® Data for isonitriles from F. A. Cotton and F, Zingales, J. Am. Chem. Soc., 1961, 83, 351. frequencies in Cr(CNAr)¢ and Ni(CNAr), molecules are lowered relative to the frequencies of the free CNAr molecules exceeds that by which the CO modes of the corresponding carbonyls lie below the frequency of CO. Structural evidence for the ability of isocyanides to form m bonds is provided by the Cr—C bond length? of 1.94 A in Cr(CNC¢Hs)6, which is very similar to the value of 1.91 A in Cr(CO)s. The same compound, however, also shows the ability, mentioned above, of isocyanides to function as good ¢ donors without ex- tensive 7-bonding. Although Cr(CO)< cannot be oxidized to Cr(CO)g* cations, the Cr(CNPh)?* ions can be isolated as PF; or BPhj salts that are thermally stable at room temperature. There are several compounds containing bridging isocyanide groups? of the types: oe M—M' Peete 7 — L. h. rn ve RY If the metals are different, there can be isomers, whereas for doubly bridging sys- tems there will be syn and anti isomers as found in (7-CsHs)2Fea(CO)2 (CNMe) 30; these may interconvert rapidly. R R n% a i i \ \ M—M — v7 \/ \, i 1 Nr RR” syn anti 28 BE, Ljunstrom, Acta Chem. Scand., 1978, 32, 47. 29 P.M. Treichel and G. J. Essenmacher, Inorg. Chem., 1976, 15, 146. 30 (a) A. L. Balch et al., J. Organomet. Chem., 1978, 159, 289; W. P. Fehlhammer et al., Angew. Chem., Int. Ed., 1978, 17, 866; (b) R. D. Adams and F. A. Cotton, Inorg. Chem., 1974, 13, 249.LIGANDS AND COMPLEXES. 95 Finally the iso form of hydrogen cyanide can form complexes. These are usually obtained by protonation of cyanide complexes, for example, Fe(phen)a(CN)2 + 2H* = [Fe(phen),(CNH)2]?+ 3-13. Orders of 7 Bonding We have dealt with the major 7-bonding ligands, yet there are others in which 3 bonding plays a part. These ligands are (a) compounds of sulfur donors, such as thioethers, (b) ligands with extended a systems such as 2,2’-bipyridyl, 1,2-di- thiolenes, and related sulfur ligands such as dithiocarbamates, and (c) compounds with multiply bonded groups such as MO, M==N, and M=NR. All these ligands are considered under their particular donor atoms in Chapter 4. It has been noted that from comparisons of CO stretching frequencies in sub- stituted carbonyls it is possible to arrange a variety of ligands in order of .-bonding capacity, as done earlier, for phosphorus, arsenic, and antimony donors. Orders based on other sorts of data can be derived; for example, that from trans effect studies (Section 28-7) gives CO, CN~, C2H4 > PR3, H~ > CH3, SC(NH2) > CsHs, NOz, Im, SCN~ > Br~ > Cl > py, NH3, OH~, OH, which also includes some non-7-bonding ligands. Such orders are semiquantitative at best, but there are evidently extremes of good and poor 7-bonding ligands. a COMPLEXES OF UNSATURATED ORGANIC MOLECULES Molecules that have multiple bonds (C=C, C=C, C=O, C=N, S=O, N=O, etc.) can form what are called + complexes with transition metals (Chapter 27) 3-14. Alkenes3! The most important + complexes are those of compounds with C=C bonds. The earliest known organotransition metal complex was discovered by Zeise in Co- penhagen in about 1845, but the true constitution was not recognized until the 1950s. Zeise’s salt, K[Pt(C2H4)Cl3], has ethylene bound to it as shown?24 in Fig. 3-9a, The key point is that the C=C axis of the coordinated alkene is perpendicular to one of the expected bond directions from the metal. The expected line of a bond orbital from the metal strikes the C=C bond at its midpoint (though for unsym- metrical alkenes, this need not be so). Certain other molecules may be bound sideways, the most important of these being dinitrogen and dioxygen. The structures? of the three square complexes, trans-RhCIL(PPr4)2, L = CzH4, No, and O» are shown in Fig. 3-9b, c, and d. *'S.D. Ittel and J. A. Ibers, Adv. Organomet. Chem., 1976, 14, 33: D. P. Mingos, Adv. Organomet Chem., 1977, 18, 1 * (a) R.A. Loveet al., Inorg. Chem., 1975, 14, 2653; (b) C. Busetto et al., J. C.S. Dalton, 1977, 1828,96 INTRODUCTORY TOPICS td) te) ig. 3-9. (a) The structure of the ion in Zeise’s salt. (b, c, d) The structures of ethylene, nitrogen, and oxygen complexes, trans-RhCIL(PPr4)2. [Reproduced by permission from C. Busetto et al., Ref. 326.) (e) The structure of the tetracyanoethylene complex IrBr(CO)|(CN)2C=C(CN)2}(PPha)2 (PPhs)2. The most generally useful description of the bonding was developed for cop- per-alkene complexes by M. J. S. Dewar and later extended to other transition metals. Fig. 3-10 illustrates the assumption that as with other a-bonding ligands like CO, there are nwo components to the total bonding: (a) overlap of the a-electron density of the olefin with a o-type acceptor orbital on the metal atom and (b)a “back-bond” resulting from flow of electron density from filled metal dy, or otherLIGANDS AND COMPLEXES 97 \/ Me c DC < 4 ~S. + SM os. 4 c tC - Sods NLS ue Donation from filled Back-bonding from filled 7 orbitals to vacant metal orbital to acceptor ‘metal orbital * orbitals Fig. 3-10. The molecular orbital views of olefin-metal bonding according to Dewar. dzx-pz hybrid orbitals into antibonding orbitals on the carbon atoms. This view is thus similar to that discussed for the bonding of carbon monoxide and similar weakly basic ligands and implies the retention of appreciable “double-bond” character in the olefin. Of course, the donation of 7-bonding electrons to the metal ¢ orbital and the introduction of electrons into the -antibanding orbital both weaken the x-bonding in the olefin, and in every case except the anion of Zeise’s salt there is significant lengthening of the olefin C—C bond. The important qualitative idea about metal-olefin bonding is that the bonding has dual character. There is donation of those electrons initially forming the C—C 7 bond into a metal orbital of suitable symmetry, and there is donation of electrons from filled metal orbitals of suitable symmetry back into the m-antibonding orbitals of the olefin. As in the CO case, the two components are synergically related. As One component increases, it tends to promote an increase in the other. On both theoretical and experimental grounds, it appears that the metal-alkene bond is essentially electroneutral, with donation and back-acceptance approximately balanced. Although as we have seen, in crystalline compounds the C—C axis of ethylene is perpendicular to the plane of the square coordination shell of the metal, in solution the nmr spectra show that there is rotation of the alkene about the metal-alkene axis. Extensive x-bonding would be expected to hinder the rotation as is the case for C=C bonds. However in some cases the barrier is relatively small; for example, in Os(C2H4)(CO)(NO)(PPh3)) it is about 40 kJ mol-!, whereas in other cases such as (7-CsHs)W(CO)2(C2H4)CH; there appears to be more hindrance.33 In complexes containing tetracyanoethylene (Fig. 3-9e) or F7C—=CF 34 the C—C bond is about as long as a normal single bond and the angles within the C(CN)4 or C2F, ligand suggest that the carbon atoms bound to the metal approach tetra- hedral hybridization. Indeed, it is possible to formulate the bonding as involving * H.G. Alt, J. A. Schwarzle, and C. G. Kreiter, J. Organomet. Chem., 1978, 183, C7; W. Porzio and M. Zocchi, J. Am. Chem. Soc., 1978, 100, 2048; C. Bachmann, J. Demuynck, and A. Veillard, J. Am. Chem. Soc., 1978, 100, 2366. 4 D.R. Russell and P. A. Tucker, J. C. S. Dalton, 1975, 1752,98 INTRODUCTORY TOPICS two normal 2e-2c metal-carbon bonds in a metallocycle (3-XXIXa) with ap- proximately sp} hybridized carbon. A number of other molecules that have multiple bonds and can be bound to metals in the 7? fashion can be regarded as forming metallocycles 3-XXIXb through 3-XXIXe. CFs ma o 1 ACF, 5 go a at uo ae wt mu | oe $ 0 0 R (@XXIXa)- (GXXIXb) —(BXXIXe)_—(3-XXIXd) (3XXIXe) Actually, the metallocycle view and the 7-donor view are neither incompatible nor mutually exclusive but are complementary, with a smooth graduation of one description into the other. The one to be preferred in any given case depends on the extent to which the double bond of the ligand has been reduced to a single bond. From a formal point of view, however, the metallocycle view entails a problem with oxidation state. For example, a compound such as Ni(C2F4)(CO)3 could be re- garded as a nickel(II) rather than a nickel(0) complex. Clearly, in a compound such as Pt(C2H,)3, it would be absurd to propose Pt¥!. It is best to regard molecules bound sideways as neutral ligands that do not alter the formal oxidation state. Conjugated Alkenes. When two or more conjugated double bonds are engaged in bonding toa metal atom the interactions become more complex, though quali- tatively the two types of basic, synergic components are involved. The case of the 1,3-butadiene unit is an important one and shows why it woild be a drastic over- simplification to treat such cases as simply collections of separate monoolefin-metal interactions. Two extreme formal representations of the bonding of 1,3-butadiene to a metal atom are possible (Fig. 3-11). The structure b would imply that bonds 1-2 and 3-4 should be longer than bond 2-3. In CsH¢Fe(CO); the bond lengths are approxi- mately the same and '3C-H coupling constants in the nmr spectra indicate that the hybridization at carbon still approximates to sp?. However in some other compounds of conjugated cyclic alkenes, the pattern is of the long-short-long type, indicating some contribution from this extreme structure. Alkynes. An alkyne RC=CR’ can use only one pair of 7 electrons and bond toa metal atom in the same way as does an olefin. However complexes of this sort Wy M 7) Fig. 3-11, Two extreme formal representations of the bonding of a 1,3-butadiene group to a metal atom: (a) implies that there are two more or less independent monoolefin metal interactions; (b) depicts 7 bonds to C-1 and C-4 coupled with a monoolefin metal interaction to C-2 and C-3.LIGANDS AND COMPLEXES 99 are rare, and acetylenes are most commonly found?’ in a bridging posture, using both pairs of m electrons, as in 3-XXX. There may be at least one case of an alkyne that may be thought of as a four-electron donor to one metal atom, but this is at best atypical. y: ww (3-XXX) 3-15. Aromatic Ring Systems?” Just as the x electrons of alkenes can interact with metal d orbitals, so can certain of the delocalized -electron ring systems of aromatic molecules overlap with dyz and dy, metal orbitals. The first example of this type of complex was the molecule Fe(CsHs)2, now known as ferrocene, in which the 67-clectron system of the ion CsH5 is bound to the metal. Other aromatic systems with the “magic numbers” of 2, 6, and 10 for the aromatic electronic configuration are the carbocycles: VEQ90O©0 2e be 10e The CsHs, CoHo, and CgHs rings are the most common in arene complexes, but the C7H; and C4H,?8 systems also occur frequently. It should also be noted that for purposes of electron counting the ring system and the metal atom may be con- sidered as neutral. For example, the total of eighteen electrons in ferrocene can be regarded as five per CsHs ring plus eight from Fe. Compounds are known that have only 7-bonded rings such as ferrocene (3- XXXI), dibenzenechromium (3-XXXII), or (CsHg)2U (3-XXXIII), but there are many compounds with one ring and other ligands such as halogens, CO, RNC, and R3P. Examples are 7-CsHsMn(CO)3 and 7-CsHsFe(CO)Cl. The symbol 7 is used to signify that all carbon atoms of the ring are bonded to the metal atom. There are also molecules in which two different types of arene ring are present, such 35 W. 1. Bailey, Jr., et al., J. Am. Chem. Soc., 1978, 100, 5764; V. W. Day et al., J. Am. Chem. Soc., 1976, 98, 8289. 36 L. Ricard et al., J. Am. Chem. Soc., 1978, 100, 1318. 37 (a) H. Werner, Angew. Chem., Int. Ed., 1977, 16, | (sandwich compounds); (b) W. E. Silverthorne, Adv. Organomet. Chem., 1975, 13, 48 (arene complexes); (c) R. E. Riley and R. FE. Davis, J. Am. Chem, Soc., 1976, 18, 2735 (heterocyclic systems); (d) J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc., 1976, 98, 1729 (bonding in CsHsM compounds); (e) K, R. Gordon and K, D. Warren, Inorg. Chem., 1978, 17, 987. Magnetic and spectroscopic data for (CsHs)2M. See also Chapter 21. 38 A. Efraty, Chem. Rev., 1977, 77, 691; W. Stallings and J. Donohue, J. Organomet. Chem., 1971, 139, 143.100 INTRODUCTORY TOPICS Fe Cr U (3-XXX1) (XXX) (3-XXXI) that the total number of z electrons they provide, plus those possessed by the metal atom itself, add to eighteen. For example, in 3-XXXIV,2? there are five 7 electrons irom CsHs, four from C4Ry, and nine from Co. Similarly, we have (q-CsHs)(n- CoHe)Mn. It is also possible for heterocyclic arene rings to form vomplexes,*° examples being (q-CgHgN)Mn(CO)3, (-C4HaS)Cr(CO)s, (n-CsHs)(n-CaH4N) Fe, (n-CsHs)- (9-C4HaP) Fe, and 3-XXXV. (3-XXXIV) (3XXXV) The basic qualitative features of the bonding in ferrocene are well understood, and will serve to illustrate the basic principles for all (7-C,H,)M bonding, although for (CgHg)M systems there are a few additional points that are covered later. The discussion of bonding does not depend critically on whether the preferred rotational orientation of the rings (see Fig. 3-12) in an (y-CsHs)2M compound is Fig, 3-12. Staggered and eclipsed configurations of an (7-CsHs)2M compound. In crystalline ferrocene there are molecules of different orientations randomly distributed throughout the crystal (P. Seiler and J. D. Dunitz, Acta Cryst., 1979, B35, 1068). Also, the H-atoms of the rings are bent towards the metal (F. Takasagawa and J. F. Koetzle, Acta Cryst., 1979, B35, 1074). P.-E. Riley and R. E. Davis, J. Organomet. Chem., 1976, 113, 157; A. Clearfield et al., J. Orga- nomet. Chem., 1977, 135, 229. © P.E. Riley and R. E. Davis, Inorg. Chem., 1976, 15, 2735; F. Mathey, A. Mitchler, and R. Weiss, J. Am. Chem. Soc., 1977, 99, 3537.LIGANDS AND COMPLEXES 101 ad 7 —S 4p1,,,6,) a A 4s(a,) 3 ste, 6,46.) : 1g? “, oe 19" ee “og eee Ca Fe Fig, 3-13. An approximate MO diagram for ferrocene. Different workers often disagree about the exact order of the MOs; the order shown here, especially for the antibonding MOs, may be incorrect, in detail, but the general pattern is widely accepted. staggered (Dsq) or eclipsed (Dsj); nor is that question unequivocally settled. It is experimentally certain that in ferrocenes the barrier to rotation is only about 8 to 20 kJ mol~!. The eclipsed configuration may be the more stable, but in condensed phases, especially crystals, where there are intermolecular energies of the same or greater magnitude than the barrier, either configuration may be found. The bonding is best treated in the linear combination of atomic orbitals (LCAO-MO) approximation. A semiquantitative energy level diagram is given in Fig. 3-13. Each CsHs ring, taken as a regular pentagon, has five + MOs, one strongly bonding (a), a degenerate pair that are weakly bonding (e1), and a de- generate pair that are markedly antibonding (e2), as shown in Fig. 3-14. The pair of rings taken together then has ten m orbitals and, if Dsy symmetry is assumed, so that there is a center of symmetry in the (7-CsHs)2M molecule, there will be centrosymmetric (g) and antisymmetric (u) combinations. This is the origin of the102 INTRODUCTORY TOPICS weokly bonding strongly bonding Atomic orbitals Molecular orbitals, Fig. 3-14. The a molecular orbitals formed from the set of px orbitals of the CsHs ring. set of orbitals shown on the left of Fig, 3-13, On the right are the valence shell (34, 4s, 4p) orbitals of the iron atom. In the center are the MOs formed when the ring x orbitals and the valence orbitals of the iron atom interact. For (7-CsHs)2Fe, there are eighteen valence electrons to be accommodated: five zr-electrons from each CsHs ring and eight valence shell electrons from the iron atom. It will be seen that the pattern of MOs is such that there are exactly nine bonding or nonbonding MOs and ten antibonding ones. Hence the eighteen elec- trons can just fill the bonding and nonbonding MOs, giving a closed configuration. Since the occupied orbitals are either of a type (which are each symmetric around the S-fold molecular axis) or they are pairs of e, or e2 type, which are also, in pairs, symmetrical about the axis, no intrinsic barrier to internal rotation is predicted. The very low barriers observed may be attributed to van der Waals forces directly between the rings. Figure 3-13 indicates that among the principal bonding interactions is that giving rise to the strongly bonding e,, and strongly antibonding e/, orbitals. To give one concrete example of how ring and metal orbitals overlap, the nature of this par- ticular important interaction is illustrated in Fig. 3-15. This particular interactionLIGANDS AND COMPLEXES 103 2 Yeo Fe Preeeeeseseeeeeaeeem oa! gps ©. Fig. 3-15. Overlapping of one of the e -type d orbitals dy: with an e1-type orbital to give a delocalized metal-ring bond: cross-sectional view taken in the xz plane. is in general the most important single one because the directional properties of the e;-type d orbitals (d,:. dyz) give an excellent overlap with the e)-type ring 7 orbitals, as Fig. 3-15 shows. Systems containing only one 7-CsHs ring include (n-CsHs)Mn(CO);, (n- CsHs)Co(CO), (n-CsHs)NiNO, and (7-CsHs)CuPR3. The ring-to-metal bonding in these cases can be accounted for by a conceptually simple modification of the picture given above for (n-CsHs)2M systems. In each case a principal axis of symmetry can be chosen so as to pass through the metal atom and intersect the ring plane perpendicularly at the ring center; in other words, the CsHsM group is a pentagonal pyramid, symmetry Cs,. The single ring may then be considered to interact with the various metal orbitals in about the same way as do each of the rings in the sandwich system. The only difference is that opposite to this single ring is a different set of ligands which interact with the opposite lobes of, for example, the de, orbitals, to form their own appropriate bonds to the metal atom. The MO picture of ring-to-metal bonding just given for ferrocene is qualitatively but not quantitatively applicable to other (n-CsHs)2M compounds, and with ob- vious modifications, to other (n-arene)2M molecules. Relative orbital energies can and do change as the metal is changed. Ferrocene itself has been treated by very elaborate MO calculations,*! and at the other extreme a ligand field model has also 4! P.S. Bagus, U. I, Walgren, and J. Almlof, J. Chem. Phys., 1976, 64, 2324,104 INTRODUCTORY TOPICS been applied to the (n-CsHs)2M compounds generally.4? Although it is clearly unrealistic to treat the bonding in these systems as electrostatic (i.e., M?* and two negatively charged rings), this approach is of some utility in interpreting electronic absorption spectra. It must also be noted that not all “(7-CsHs)2M” compounds are as simple structurally or electronically as might naively be expected. Those involving titanium and niobium have complex binuclear structures, as described in detail under the chemistry of these elements. In the case of (y-CsHs)2Ni there are two electrons in excess of those required for an eighteen-electron configuration; nonetheless, the molecule appears to have a structure with parallel rings and the two additional electrons have been assigned to the e7, orbitals where they are appreciably delo- calized onto the rings.*3 With manganese the situation is complicated but now understood: (CsHs)2Mn is a brown solid in which the Mn atoms have five unpaired electrons but show strong antiferromagnetic coupling. This coupling results from a chain structure“ in the crystal (Fig. 3-16). At 159°C the crystal structure changes so that (CsHs)2Mn becomes isomorphous with ferrocene, and the color becomes light orange. However there are still five unpaired electrons; thus the (7-CsHs)2Mn molecule can be considered to be mainly ionic. It is also ionic, with five unpaired electrons, in the gas phase. However (q-CsH4CH3)2Mn in the gas phase at about 25°C is a roughly 2-1 mixture of high-spin (ionic) molecules with Mn—C distances of 2.42 A and low-spin (covalent) molecules with Mn—C = 2.14 A.45 For a monocyclopentadienyl compound having two bulky substituents, (R'R2C5H3)Co(PMe3)2, rotational isomers have been observed.*® Compounds with two 7-CsHs rings that are not parallel are also numerous. They include a number of (CsHs)2MX2 compounds in which M = Ti, Zr, or Moand X represents a univalent group such as a halogen, H or an R group, as well as others such as (7-CsHs)2Mo(CO) and (n-CsHs)2NbH(CO). The angle subtended at the metal atom by the centroids of the two rings is generally 130° to 135°. oe Se tp Seta AOR eR AOS HE Ab Y y ib kG Fig. 3-16. The chain structure of (CsHs)2Mn in the low-temperature, antiferromagnetic crystal form, 42K. D. Warren, Inorg. Chem., 1974, 13, 1243. 43 W. T. Scroggins, M. F. Rettig, and R. M. Wing, Inorg. Chem., 1976, 15, 1381. 4 W. Biinder and E. Weiss, Z. Naturforsch, B, 1978, 33, 1235. 45 A. Almenningen, S. Samdal, and A. Haaland, J. C. S. Chem. Comm., 1977, 14. 46 W. Hofmann, W. Biichner, and H. Werner, Angew. Chem., Int. Ed., 1977, 16, 795.LIGANDS AND COMPLEXES 105 Fig. 3-17. Diagrams showing how p, and py orbitals are symmetry-adapted to overlap the e x orbitals of a CsHs ring One of the features of (n-CsHs)2ReH and (7-CsHs)2MH2, (M = Mo, W) is that in addition to the orbitals involved in bonding to the CsHs rings, there are three orbitals that may be used either for bonding to hydrogen or to hold lone pairs. Thus (n-CsHs)2W Hy? has one lone pair and can accept a proton giving (n-CsHs)2WH3, whereas the molecule (7-CsHs)2TaH3 does not act as a base. The electronic structure of such bent sandwich compounds has been worked out in some de- tail.47 Tt is also possible to have covalent (7-CsHs)M groups even when the metal atom has no valence shell d orbitals, provided it has p orbitals of suitable energy and size. As shown in Fig. 3-17, a pair of px and py orbitals can overlap with the e 7 orbitals of CsHs in much the same way as do d,, and d,, orbitals. The CsHsIn and CsHsT1 molecules are the best documented cases of this type of bonding. Bonding in (7-CgHg)2M Molecules. Although there are a few cases—for ex- ample, (n-CsHs)ZrCl,(THF)**—of one fully octagonal, planar CgHg ring being symmetrically bonded to a d-block metal and it may be safely assumed that only metal d orbitals are used, the (7-CgHs)2M compounds are formed only by metal atoms like uranium and thorium where the participation of f orbitals might rea- sonably be expected. The highest half-occupied 7 orbital of a planar octagonal CsHs ring is of type e2 and can overlap much better with an e2-type f orbital than with an €2 type d orbital. Nonetheless there has been considerable dispute over the extent of such f-orbital participation; photoelectron spectra and MO calculations have been used to support arguments for both much*? and little® f-orbital participation. It does seem likely that f-orbital participation is of some significance. 47 J. W. Lauer and R. Hoffmann, J. Am. Chem. Soc., 1976, 98, 1729; R. D. Wilson et al., J. Am. Chem. Soc., 1977, 99, 1775; A. J. Schultz et al., Inorg. Chem., 1977, 16, 3303 “8D. J. Brauer and C. Kruger, Inorg. Chem., 1975, 14, 3053. 4 N. Edelstein et al., Inorg. Chem., 1976, 15, 1397; J. P. Clark and J. C, Green, J. C. S. Dalton, 1977, 505; N. Rosch and A. Streitweiser, Jr., J. Organomet. Chem., 1978, 145, 195. $0 I. Fragola et al., J. Organomet. Chem., 1976, 122, 357.106 INTRODUCTORY TOPICS 3-16. Partially Delosalized Enyl Complexes Aromatic systems are fully delocalized, but other types of ligand are known that have delocalized open chain systems or cyclic systems with only partial delocali- zation. The simplest is the ally/ group. This can be bound as in 3-XXXVI and 3- XXXVII, the former being designated 73-allyl, the latter o-allyl or 9!-allyl. Note H H Socin uM” Nc 4 Hon 1? or w-Allyl 7 or o-Allyl (3-XXXV1) (3-XXXVII) that (1) the 73-allyl group is a three-electron donor, and (2) the hydrogen atoms of the methylene groups are not equivalent. There are syn-(Ha) and anti-(Hp) protons that can be distinguished by nmr. Many allyls show nonrigid behavior in solution! (Section 28-16). Other more complicated, partly delocalized enyl systems are: 1-1-3-Cyclohexenyl Three-Electron donor 111-5-Cyclohexadienyl ea Five-Electron donor Similar systems can be formed from seven-, eight-, and nine-membered ring hy- drocarbons. General References Jensen, W. B., Chem. Rev., 1978, 78, 1. Lewis acids and base behavior, hard-strong acids and bases. Tables of donor strengths, and so on, Hartley, F. R., Chem. Soc. Revs., 1973, 2, 163. Cis-trans effects of ligands. Hoffmann, R., and M. M.-L. Chen, Inorg. Chem., 1977, 16, 503. Bonding of diatomic molecules to metals, Malatesta, L., and C. Cenini, Zerovalent Compounds of Metals, Academic Press, 1974.CHAPTER FOUR Classification of Ligands by Donor Atoms Chapter 3 discussed the general features of ligands and the distinction between a-bonding and non-x-bonding types. This chapter deals with ligands classified according to the atom of the ligand bound to the metal. Detailed chemistry of certain important classes of ligands is described separately, notably for metal carbonyls and related compounds (Chapter 25), cluster compounds, many of which are carbonyls (Chapter 26), and transition metal hydrido and organo compounds (Chapter 29). HYDROGEN The chemistry of binary hydrogen compounds except for boranes (Chapter 9), is discussed in Chapter 6, and the transition metal compounds with M—H bonds are dealt with in Chapter 27, and also in Chapter 29. Here we discuss only the ligand behavior of the simpler complex hydrido anions such as BH,~, B3Hg~ and AIH,-. 4-1, Complex Hydrido Anions There are numerous complexes of tetrahydridoborate! (BH47) with main group elements, d-group transition metals, lanthanides, and actinides. Some examples are AI(BH4)3, [U(BH,)s]-, (R3P)2CuBHs, [Mo(CO)sBH,]-, and Y(BH4)3(THF)3. They are usually obtained by interaction of the appropriate halide with LiBH, or NaBH ina solvent like tetrahydrofuran or ether, or by interaction of B2Hg with metal alcoxides. 1 T. J. Marks and J. R. Kolb, Chem. Rev., 1977, 77, 263; B. G. Segal and S. J. Lippard, Inorg. Chem., 1978, 17, 844, 107108 INTRODUCTORY TOPICS The ligand is bound via hydrogen bridges: Few unidentate complexes are known, an example being (PhyMeP)3;CuHBH3,” Bidentate complexes are most common. Many compounds have been structurally characterized, but infrared spectroscopy is a useful criterion; for example, bidentate compounds have an absorption at ca. 2500 cm=!. Nuclear magnetic resonance spectra show that the complexes are commonly nonrigid. In bi- and tridentate species the bridge and terminal hydrogen atoms are undergoing rapid intramolecular exchange and so appear equivalent on the nmr time scale.? The octahydridotriborate ion B3Hg~ also forms complexes with M—H—B bonds,* for example, 4-I. H cO ae ™ pH, oc, | 4. | AB i ae | / oc | Su—p—— # co Nu (4D The more complicated borane anions also give complexes.> The tetrahydridoalu- minate ion (AIH4~) forms few complexes, but a number of compounds that have groups of the following type are known.® Hi errata we ATOMS OF GROUPS I TO III Compounds of the elements in Groups I to III do not commonly act as ligands except in special cases, which are discussed separately. 2 JL. Atwood et al, Inorg. Chem., 1978, 17, 3558, 3 P.L. Johnson et al., J. Am. Chem. Soc., 1978, 100, 2709; T. J. Marks, et al., J. Am. Chem. Soc., 1977, 99, 7539; S. W. Kirtley et al., J. Am. Chem. Soc., 1977, 99, 7154. 4 §.J. Hildebrandt, D. F. Gaines, and J. C. Calabrese, Inorg. Chem., 1978, 17,790; D. F. Gaines and S. J.Hildebrandt, Inorg. Chem., 1978, 17, 794. 3 N,N. Greenwood et al., J.C.S, Dalton, 1978, 237; Pure Appl. Chem., 1977, 49, 791; Chem. Soc. Rev, 1974, 3, 231; T. P. Fehlner, et al., J. Am. Chem. Soc., 1979, 101, 4390. 6 See, eg., T. J. McNeese, S. S. Wreford, and B. M. Foxman, J.C.S. Chem. Comm., 1978, 500.CLASSIFICATION OF LIGANDS 109 4-2. Group I and Group II There are essentially no cases in which compounds of the electropositive elements in Groups I and II act as ligands in the acknowledged sense. The smallest of the alkali atoms, lithium, can be bound directly to other atoms in molecular compounds; also in polymeric organolithium compounds (Chapter 7) there is strong interaction between lithium and hydrogen atoms of an alkyl group leading to abnormally low C—H stretching frequencies. For the zinc-cadmium-mercury group these atoms, especially mercury, can be bound in compounds such as those of metal carbonyls like Hg{Co(CO)4}>. 4-3, Group Ill The elements boron, aluminum, gallium, indium, and thallium form compounds of varying types. Those of indium and thallium are similar to those of zinc, cad- mium, and mercury, noted above. There are a few compounds, mostly of transition metals, with bonds to aluminum, but the major ligand behavior is shown by boron compounds. These are rather special and are sufficiently important to be dealt with separately in Chapter 27; briefly, however, they comprise 1. Boranes acting as ligands but giving M—B as well as M—H—B bonds. 2. Carbaboranes (Chapter 9) acting as ligands comparable to n-CsHs and giving sandwich-type complexes. 3. n-Complexes of borazine (B3N3Hs) and related compounds analogous to those formed by arenes. CARBON 4-4, Organometallic Compounds: General Survey of Types’ Organometallic compounds are those in which the carbon atoms of organic groups are bound to metal atoms. Thus we do not include in this category compounds in which carbon-containing components are bound to a metal through some other atom such as oxygen, nitrogen, or sulfur. For example, (C3H70)sTi is not consid- ered to be an organometallic compound, whereas CsHsTi(OC3H7)3 is, because in the latter there is one direct linkage of the metal to carbon. Although organic 7 G. Coates, M. L. H. Green, and K. Wade, Organometallic Compounds, 3rd ed., Vol. 1, Main Group Elements, 1961; Vol. 2, Transition Elements, 1968, Methuen; Organometallic Chemistry, Specialist Reports, Chemical Society, London; Advances in Organometallic Chemistry, Academic Press; J. Organomet. Chem. (annual reviewis); E. Maslowsky, Jr, Vibrational Spectra of Or- ganometallic Compounds, Wiley, 1977; J. D. Smith and D. R. M. Walton, Adv. Organomet. Chem., 1975, 13, 453 (guide to literature for main group elements); D. S. Matteson, Organome- tallic Reaction Mechanisms, Academic Press, 1974; Comprehensive Organic Chemistry, Parts 12-15, Vol. 3, Pergamon Press, 1979110 INTRODUCTORY TOPICS groups can be bound through carbon, in one way or another, to virtually all the elements in the Periodic Table, excluding the noble gases, the term organometallic is usually rather loosely defined and organo compounds of decidedly nonmetallic elements such as boron, phosphorus, and silicon are often included in the category. Specific compounds are discussed in the sections on the chemistry of the individual elements, since the organo derivatives are usually just as characteristic of any el- ement as are, say, its halides or oxides. However, it is pertinent to make a few general comments here on the various types of compound. 1. Ionic Compounds of Electropositive Metals. The organometallic compounds of highly electropositive metals are usually ionic. Thus the alkali metal derivatives with the exception of those of lithium, which are fairly covalent, are insoluble in hydrocarbon solvents and are very reactive toward air, water, and so on. The alkaline earth metals calcium, strontium, and barium give poorly characterized compounds that are even more reactive and unstable than the alkali salts. The stability and reactivity of ionic compounds are determined in part by the stability of the carb- anion. Compounds containing unstable anions (e.g., CnH3,+1) are generally highly reactive and often unstable and difficult to isolate; however where reasonably stable carbanions exist, the metal derivatives are more stable, though still quite reactive [e.g., (CoHs)3C~ Nat and (CsH3)2Ca?*]. 2. a-Bonded Compounds. Organo compounds in which the organic residue is bound to a metal by a normal two-electron covalent bond (albeit in some cases with appreciable ionic character) are formed by most metals of lower electropos tivity and, of course, by nonmetallic elements. The normal valence rules apply in these cases, and partial substitution of halides, hydroxides, and so on, by organic groups is possible, as in (CH3)3$nCl and CH3SnCls, for example. In most of these compounds bonding is predominantly covalent and the chemistry is organic-like, although there are many differences in detail due to factors such as use of higher d-orbitals or donor behavior as in R4Si, R3P, R2S, and so on, incomplete valence shells or coordinative unsaturation as in R3B or R2Zn, and effects of electroneg- ativity differences between M—C and C—C bonds. Although the existence of stable M—C bonds has long been regarded as a normal part of the chemistry of the non-transition metals and metalloids, compounds containing transition metal-to-carbon a bonds have only in recent years been made in substantial numbers. The reasons for the relative rarity of such compounds are still a subject for investigation, but several points seem clear. First, an important pathway for decomposition of M—R bonds is by a shift of a B-hydrogen atom, followed by olefin elimination, namely; | cH M—CH,—CHRR’ —> M+|| | —> MH + CH,=CRR’ CCR’ This pathway for decomposition can be substantially inhibited either by (a) using groups such as CH, or sterically bulky alkyl groups like CH2SiMes or CH2Ph that do not have 8-hydrogen atoms, or (b) blocking the sites required for the transferCLASSIFICATION OF LIGANDS 11 reaction by firmly held ligands as in the substitution-inert octahedral species [RhC2Hs(NHs3)s]**. Available bond energy data indicate that bonds between transition metals and carbon are comparable in strength to those with nontransition elements. Transition metal compounds are discussed in detail in Chapter 27. 3. Nonclassically Bonded Compounds. There are many compounds in which metal-to-carbon bonding cannot be explained as either ionic or covalent in the simple sense of a 2c-2e M—C bond or bonds. The largest and most important class of these “nonclassical” molecules comprises those formed primarily by the transition elements in which unsaturated groups are attached to metal atoms by interaction of the 7 electrons with metal orbitals. We have already discussed such ligands briefly in Chapter 3 and they are treated in more detail in Chapter 27. Another, smaller class of nonclassical compounds is made up of those with bridging alkyl groups. The elements boron, aluminum, gallium, indium, and thallium all form fairly stable but reactive trialkyls and triaryls, those of boron, gallium, indium, and thallium being monomeric in the vapor and in solution. (CH3)3In and (CH3)3TI form tetramers in their crystals, but the association is weak and uncertain and does not persist in other conditions. The aluminum compounds are unique in Group III in forming several reasonably stable dimers. Thus tri- methylaluminum is a dimer in benzene solution and, partly, even in the vapor phase. AIEt3 and AlPr3 are also dimeric in benzene solution, but are almost completely dissociated in the vapor phase. AIPri is a monomer in benzene. The molecules Alo(CH3)4(CoHs)2 and Alz(C¢Hs)s are also known, as well as the polymeric [Be(CH3)2]x. The structures of four of these substances are shown in Fig. 4-1. In all cases the bridging carbon atoms are equidistant from the metal atoms; in short, the AI—C—AI—C groups have D2y symmetry. The dimeric structure has also been established at low temperature in nondonor solvents by 'H and '°C nmr® spectroscopy, although at room temperature an ex- change process occurs, leading to apparent equivalence of the terminal and bridging groups (see below). The bridging is accomplished by means of AI—C—AI 3c-2e bonds where each Alatom supplies an s-p hybrid orbital and so also does the carbon atom. The sit- uation then is as depicted in Fig. 4-2a. In the case of Alp(CH3)¢ this view of the bonding has been strongly supported by a low-temperature structure determination. In the case of the bridging phenyl groups, which lie perpendicular to the AICAIC planes, the larger AI—C—AI angles and slight inequalities in the C—C distances about the rings have been taken to mean that the px orbital of the bridging carbon atom may also play some role in the bridge bonding, as indicated by the overlap depicted in Fig. 4-2b. The fact that none of the alkyls of the Group III elements except those of alu- minum are dimerized (except perhaps GaEt; and trivinylgallium, which appear to be dimers in solution) has not yet been satisfactorily explained. For the larger metals, the small M—C—M angles required to secure good overlap would intro- duce large repulsions between the bulky metal atoms, but this cannot explain why 8 GA. Olah et al., Proc. Nat. Acad. Sci, (U. S.), 1977, 74, 5217,112 INTRODUCTORY TOPICS ba [aUCeHs)s)2 oF = CHyor CeHs [Acs Cots} Fig. 4-1. The structures of [Be(CHs)2]» and several dimeric AIR3 molecules. B(CH3)3 does not dimerize, especially since hydrogen bridging is quite important in the boranes. An important feature of coordinatively unsaturated alkyls, such as those just noted or those of magnesium, zinc, and so on, is the moderately rapid exchange of alkyl groups. The exchanges can be readily studied by nmr methods and it appears fa) (by Fig. 4-2. Schematic indication of orbital overlaps in Al-C—Al bridge bonding: (a) for a methyl bridge; (b) possible component in phenyl bridging,CLASSIFICATION OF LIGANDS 113 that bridged transition states, or intermediates of the type 4-II and the like, provide H, M L.Me Sco Me “C7 ‘Me A, (41D the means for exchange. The rates of exchange reactions are usually slowed by the presence of donor ligands and, if the donor is strong enough to block the coordination sites, the exchange is stopped. i Bridging alkyl groups occur in lithium alkyls and are algo known in transition metal compounds such as {Mn(CH SiMes)2}, and Re3(CH3)o(PMe3)3, as discussed in Chapter 27. 4-5, Cyanide? Cyanide complexes are readily formed in aqueous solution by Zn?*+, Cd2+, and Hg?* Most transition metals of the d groups form complexes, a well-known ex- ample being ferrocyanide, [Fe(CN)o]*~. Some lanthanide complexes are known although other ligands are usually present. As far as is known, it is always the carbon atom of unidentate CN that is bound to the metal. The -acceptor behavior of CN~ does not seem to be nearly as high! as for CO, NO*, or RNC, which is, of course, reasonable in view of its negative charge. Since CN“ is a strong nucleophile, back-bonding need not be invoked to explain the stability of its complexes with metals in normal (i.e., II, 111) oxidation States. Types of Cyano Complex. The majority of cyano complexes have the general formula [M’*+(CN),]@~")~ and are anionic, such as [Fe(CN)e]4~, [Ni(CN),]2-, and [Mo(CN)s]?~. Mixed complexes, particularly of the type [M(CN)sX]"-, where X may be HO, NH3, CO, NO, H, or a halogen, are also well known. Cyanide can act as a bridge group usually of the end-on type M—C—N— M.!0b Linear bridges play an important part in the structures of many crystalline cy- anides and cyano complexes. Thus AuCN, Zn(CN)3, and Cd(CN)> are alll poly- meric with infinite chains, The free anhydrous acids corresponding to many cyano anions can be isolated, examples being H3[Rh(CN)¢] and H4[Fe(CN),]. These acids are thus different from those corresponding to many other complex ions, such as [PtClg]2~ or [BF,]~, ° A.G. Sharpe, The Chemistry of Cyano Complexes of Transition Metals, Academic Press, 1976 (1800 references); W. P. Griffith, Coord. Chem. Rev., 1975, 17, 177 (Ti, Cr, V, Mn); L. H. Jones and B. 1. Swanson, Acc. Chem, Res., 1976, 9, 128 (force constants). ‘02 M.S. Lazarus and T. S. Chow, J. Chem. Phys,, 1976, 64, 3544, 10D. N. Hendrickson et ai., Inorg. Chem., 1977, 16, 924; Inorg. Chem., 1974, 13, 1911; P. L. Gaus and A. L. Crumbliss, Inorg. Chem., 1976, 15, 2080; D. Gaswick and A. Haim, J. Inorg. Nucl. Chem., 1978, 40, 437.114 INTRODUCTORY TOPICS which cannot be isolated except as hydroxonium (H3O*) salts, and they are also different from metal carbonyl hydrides in that they contain no metal-hydrogen bonds. Instead, the hydrogen atoms are situated in hydrogen bonds between anions (ie., MCN +++ H--++NCM). Different sorts of structures arise depending on the stoichiometry. For example, in H[Au(CN)a] there are sheets. For octahedral anions there is a difference in structure depending on whether the number of protons equals half the number of cyanide ions. For H3[M(CN)g] compounds an infinite, regular three-dimensional array is formed in which the hydrogen bonds are perhaps sym- metrical, whereas in other cases the structures appear to be more complicated. Metal-Cyanide Bonding. The cyanide ion occupies a very high position in the spectrochemical series (Section 20-12), gives rise to large nephelauxetic effects and produces a strong trans effect (Section 28-7). Alll these properties are accounted for most easily by postulating M—CN 7 bonding, and semiempirical MO calcu- lations support this. From close analysis of the vibrational spectra of cyanide complexes, the existence of « bonding has been confirmed more directly, but it does not appear to be nearly as extensive as in carbonyls. However the cyanide ion does have the ability to stabilize metal ions in low formal oxidation states, and it presumably does this by accepting electron density into its xr* orbitals. The fact that cyano complexes of zerovalent metals are generally much less stable (in a practical as opposed to a well-defined thermodynamic or chemical sense) than similar metal carbonyls has often been taken to show the poor 7r-acidity of CN-, but it should be noted that the cyano compounds—for example, [Ni(CN)s]*"—are anionic and might tend to be more reactive for this reason alone, In some instances cyano complexes are known in two or even three successive oxidation states, [M(CN),}~, [M(CN)a]@*0>, [M(CN),]&*2)-, and in this respect CN~ resembles ligands such as 2,2/-dipyridyl and 1,2-dithiolenes (Sections 4-10 and 4-34). Finally note that some cyaho complexes may themselves act as li- gands, One example!! is that of [Ni(CN)4]?, which gives with 2,2’,2”- triami- notriethylamine (tren) the complex 4-III. NG CN : a : trenNi, Ni ‘Nitren ee ee (4m) Es 4-6. Carbon Disulfide: Carbon Dioxide Carbon disulfide complexes!? were discovered at the same time as thiocarbony! 1 K.R. Mann, D, M. Duggan, and D. N. Hendrickson, Inorg. Chem., 1975, 14, 2577. 12 PV. Yaneff, Coord, Chem. Rev., 1977, 23, 183; G. Gattow and W. Behrendt, Carbon Sulfides and Their Inorganic Complex Chemistry, Thieme, 1977; H. Huber, G. A. Ozin, and W. J. Power, Inorg. Chem., 1977, 16, 2234; |. S, Butler, Ace. Chem. Res., 1977, 10, 359; M. Herberhold, M Siiss-Fink, and C. G. Kreiter, Angew. Chem., Int. Ed., 1977, 16, 193, 194; H. Le Bozec et al., Inorg. Chem., 1978; 17, 2568; U. Oemichen et al., J. Organomet. Chem., 1978, 156, C29.CLASSIFICATION OF LIGANDS. 1s complexes (page 86). Although rather weak end-on S-bonded 7! complexes can be formed, most of the CS2 complexes are of the three-membered ring 7? type, examples being as follows: Ss Ph,P. ea qQ oo Pt! Mi Php~ QJ oc! \pes i Bridging CS: complexes may be of these types: I M—S—C—M M ie C=S— M LC The 7? complexes have strong CS bands ca. 1160 and 1140 cm-!. The bound CS) is quite reactive and can be attacked by electrophiles!3# and acetylenes!3® sometimes to give carbene complexes (Chapter 27), for example, sy + CHI — | (Pn,P)(CO, RU | r SO ‘Me. f aM SN a & Eli (PhiP){CO),0s< | + 2CH,I —* | (Ph;P),(CO),10s—C- T Ss ‘SMe (Ph,P).(CO),Ru: Ss s A SMe ee “y (Ph,P),Pt: | + 2CH —-| (Ph,P),IPt—C: Tes Ss Non A number of carbon dioxide complexes'* have been claimed, but some of these have turned out to be misnamed. Thus the action of CO7 on HRh(PR3)a gives a bridged carbonato complex 30D. H. Farrar, R, O, Harris, and A. Walker, J. Organomet. Chem., 1977, 124, 125. 180 H. Le Bozec et al., J. Am. Chem. Soc., 1978, 100, 3946, ' (a) S. Krogsrud et al., Inorg. Chem., 1976, 15,2798; M. E. Volp'in and I. S. Kolomnikov, in Or- ganometallic Reactions, E. 1. Becket and M. Tsutsui, Eds., 1975, 5, 313; G. Fachinetti et al., J. Am. Chem. Soc., 1979, 101, 1767; 1978, 100, 7405; V. D. Bianco et al., Inorg. Nucl. Chem. Leit., 1979, 18, 187; (b) G. R. John et al., J. Organomet. Chem., 1979, 169, C23; (c) M. Aresta and C.F. Nobile, Inorg. Chim. Acta, 1977, 24, L49.116 INTRODUCTORY TOPICS [oper De d/ (RaP)sRh—O Osmium cluster anions with p12-CO> bridges are also known.'4 However complexes such as (Cy3P)2Ni(CO>), and RhCl(CO2)(PR3)2' are authentic. The complexes appear to be of two types, one like the y?-CS» complexes, the other, C-bonded: 0 4 c 0 ue | m—cO Nb © The latter appear to have ir bands ca. 1550 and 1220 cm-!, whereas the 9?-type bands are ca. 1660 and 1630 cm-!; examples are respectively, probably, [Irdi- ars(CO3)]* and RhCK(CO2)(BusP)2. SILICON, GERMANIUM, AND TIN 4-7. Silicon, Germanium, and Tin(IV)!> The major types of complex containing silicon, germanium, or tin ligands are those with SiR3, GeR3, SnR3 groups that are similar in stoichiometry to those of carbon compounds. They can be obtained by reactions such as (-CsHs)(CO)sWNa + CISiMe3 = (7-CsHs)(CO);WSiMes + NaCl or, for silicon, by oxidative addition reactions (Section 29-3) of silanes, ¢.g., trans-Ir}C\(CO)(PPh3)3 + Me3SiH = Ir!!1CI(H)(SiMe3)(CO)(PPh3)2 Me, SiMe,H Si Fe(CO), + OL — OL DF e(C0), + He + CO SiMe,H Si Me, The SiR3 groups have a very high trans effect. There are also compounds with bridging groups! such as the following: 154, Bonny, Coord, Chem. Rev., 1978, 25, 229; E. H. Brooks and R. J. Cross, Organomet. Chem. Rev., A, 1970, 6 227; C. S. Cundy, B. M. Kingston and M. F. Lappert, Adv. Organomet. Chem., 1973, 11, 253; F. Héfler, Topics Curr. Chem., 1974, 50, 129. 16 See, e.g., G. Schmidt and G. Etzvodt, J. Organomet. Chem., 1977, 137, 367.CLASSIFICATION OF LIGANDS 17 0 [co CO—Co. R [ %co it j y we Ly pico (CO), (CO).C% Ee cx00), R, Si a (oo.co Soucy, Ea a (CO),Min ‘Mn(CO), I 0 4-8. Tin(l)'” Divalent tin compounds have a lone electron pair (Chapter 12) and consequently can act as donors. There are numerous complexes of the trichlorotin(II) ion SnClz '7° and Sn(n-CsHs)>; tin(I)carboxylates and 6-diketonates can also act as ligands. Some germanium(II) and lead(II) compounds behave similarly.'8 The SnCls~ can generally displace Cl“, CO, PFs, and so on; for example, eC 13 38nC15 PUCI> ——> PrCla(SnC 3) ——> Pr(SnC ls) sacl RhCls(aq) —> [(ClsSn)2Rh'(u-Cl)Rh!(SaChy)2]*~ Platinum metal complexes with SnCly dissolved in molten quaternary alkyl am- monium salts, RsN*SnCl; or ReN*GeCly, have been used as catalytic systems for hydrogenation of olefins and other reactions (cf. Chapter 30).19 Some metal-metal bonded complexes react with SnCl; in a type of “insertion reaction,” (Chapter 29) where SnCl, could be considered to be showing carbenelike behavior, but the resulting compounds are clearly compounds of tin(IV): a Sncl, (-CsHs).Fe{CO), ——* (7-C,Hs) (CO),Fe™ Sn ~aFe(CO),(74C,H;) cl "7 (a) J.F. Young, Ade. Inorg. Chem. Radiochem,, 1968, 11, 92: (b) T. Kruck et al, Z. Naturfursch., 1978, 33b, 129, P.G. Harrison et al., J.C.S. Dalton, 1975, 2017; 1976, 1054, 1608, 1612. 1° G. W. Parshall, J. Am. Chem. Soc., 1973, 94, 8716. 8118 INTRODUCTORY TOPICS NITROGEN The types of organic nitrogen compound that can act as ligands are innumerable, and only the more important are discussed here. The bonding of Nz and NO was discussed in Sections 3-9 and 3-11, respectively. 4-9, Ammonia and Amines”? Ammonia and substituted ammonias (including hydrazine and substituted hy- drazines: sce later) act as ligands toward both non-transition and transition metal ions, as well as giving adducts with Lewis acids. Coordinated ammonia and amine ligands can undergo oxidation and conden- sation reactions.2! Thus the trisethylenediamine complex [Ruen3]?+ may be oxi- dized, probably initially to a Ru!"! complex, whereupon intramolecular oxidative dehydrogenation of the ligand occurs, to give an a-diimine complex: ws u " 7X4 cH,| ae NSH en.Ru I — | en,Ru 1 + 4Ht Naot \yeoon There is also a similar reaction for o-phenylenediamine complexes: OG Fe(CN), | “+ Ces + 2Ht N q o-Benzoquinonediimine complex Diimine complexes can also be obtained in other ways—for example, by reduction of FeCl:(PMe3)2 in CH3CN by Na/Hg, the complex 4-1V is formed.?? MeC=N\ | Fe(PMe,)». MeCy/ (41V) BZ 2m 20K. H. Schmidt and A. Muller, Coord. Chem. Rev., 1976, 19, 41 (vibrational spectra of pure amines); W. W. Wendlandt and T. P. Smith, Thermal Properties of Transition Metal Amine Complexes, Elsevier, 1967; D. A. House, Coord. Chem. Rev., 1977, 23, 223 (pentammines of Coll! and Crlll, stereochemistry and reaction rates); S. T. Chow and C. A. McAuliffe, Prog. Inorg. Chem., 1975, 19, 51 (tridentate complexes of amino acids); D, A. Phipps, J. Mol. Catal., 1979, 5,81 (amino acids); E. Uhlig, Z. Chem., 1978; 18, 440 (chelating pyridine complexes). 21 CP. Guengerich and K. Schug, J. Am. Chem. Soc., 1977, 99, 3298; G. M. Brown et al., Inorg. Chem., 1976, 18, 190; G. G. Christoph and V. L. Goedken, J. Am, Chem. Soc., 1973, 95, 3869; D. F. Mahoney and J. K. Beattie, Inorg. Chem., 1973, 12, 2561; L. F. Lindsay and S. E. Living- stone, Coord. Chem, Rev., 1967, 2, 173: 1. P, Evans, G. W. Everett, and A. M. Sargeson, J. Am. Chem. Soc., 1976, 98, 8041. 22, W. Rathke and E. L. Muetterties, J. Am. Chem. Soc., 1975, 97, 3272.CLASSIFICATION OF LIGANDS 119 Finally, amine complexes may be oxidized to give nitrile complexes: [(NH3)sRuNH,CH)R]** — [(NH3)sRuN=CR J?+ Examples of condensation reactions are given later in this chapter when discussing template synthesis, but one example is the interaction of ammonia complexes with §-diketones in basic solution?’ to give nitrogen ligands that are comparable to B-diketonates (Section 4-27): [PUNH),J** + CH,COcH,cocH, “+ | (wH,),P¢ 4-10. 2,2/-Bipyridine and Related Ligands* The aromatic amines or rather polyimines, 2,2’-bipyridine (4-V), 1,10-phenan- throline (4-V1), and terpyridine (4-VII), differ considerably from aliphatic amines. in that they can form complexes with metal atoms in a great range of oxidation (4-V) (4-V1) (4-VII) states. For example, there is the redox series?5: [Crdipys]** = [Crdipys]** = [Crdipys]* = [Crdipys]° = [Crdipys}~! For metal ions in “normal” oxidation states, the interaction of metal dz orbitais with the ligand 7* orbitals is significant, but not exceptional. However, these ligands can stabilize metal atoms in very low formal oxidation states and in such complexes it is believed that there is extensive occupation of the ligand r* orbitals, so that the compounds can often be best formulated as having radical anion ligands L~. Most work has been carried out on bipy complexes, but it is apparent that phen and terpy afford very similar ones. The methods of preparation are varied. Complexes involving transition metal ions in “normal” oxidation states can usually be obtained by conventional reactions 23S. A, Brawner et al., Inorg. Chem., 1978, 17, 1304, 24 W. R. McWhinnie and J. D. Miller, Adv. Inorg. Chem., Radiochem., 1969, 12, 135; E. D. McKenzie, Coord. Chem. Rev., 1971, 6, 187; A. A. Schilt, Applications of 1,10-Phenanthroline and Related Compounds, Pergamon Press, 1969; W. A. McBryde, A Critical Review of Equi- librium Data for Proton and Metal Complexes of 1,10-Phenanthroline, 2,2’-Bipyridyl and Re- lated Compounds, Pergamon Press, 1978. 25M. C. Hughes and D. J. Macero, Inorg. Chem., 1976, 18, 2040,120 INTRODUCTORY TOPICS and then reduced with a variety of reagents such as Na/Hg, Mg, or BH;. The most general method employs Lizbipy: THE MX, + pLizbipy + n(bipy) —> M(bipy), + yLiX + yLi(bipy) It is also noteworthy that many highly reactive organometallic compounds can be stabilized against hydrolysis, for example, by addition of these ligands. This is particularly true of R2Zn, R2Cd, and RaHg species. The low-valent metal complexes are invariably colored, usually intensely so. For those containing transition metals, the bands responsible are believed to be mainly d > x* charge-transfer bands. In other cases + —> +* ligand bands may also be active. For the ML complexes of beryllium, magnesium, calcium, and strontium, electron spin resonance (esr) spectra show the presence of a ground, or low-lying excited, state that is a spin triplet. This can be best explained by postulating an M2+ cation and two radical anion ligands L~. Also for the tris- bipy complexes [Cr(bi- py)s]*, [V(bipy)3], and [Ti(bipy)3]-, esr data indicate that there is strong o in- teraction with metal 4s orbitals, while the unpaired electrons are extensively de- localized on the ligands. Aqueous solutions of bipy and phen complexes, such as [Fephens]+, [Rubipy2- Py2]**, or [PtpysClp]2+ often show unexpected kinetic, equilibrium, and spectral behavior especially with OH-, CN-, and OR~. This can be explained?® by nucleophilic attack on the 2-position of the heterocyclic ring: (naan +Ht one [Rhpy,Cl,]” —> N—Rhpy,Cl, H OH Although 2,2’-bipyridine and 1,10-phenanthroline usually give chelate complexes, unidentate complexes?’ can be formed as in {PtCl(PEts)2(7'-phen)]* and [Ir(bi- py)2In'-bipy)(H2O)]*; in solution they are fluxional and the platinum atom moves from one nitrogen atom to the other. This is so also for unidentate pyridazine and 1,8-naphthyridine complexes (see below). 26 R. D. Gillard et al., Trans. Met. Chem., 1977, 2, 47; J.CS. Dalton, 1979, 190, 193; M. J Blandamer, J. Burgess, and D. L. Roberts, J.C.S. Dalton, 1978, 1086. 27 R. J. Watis, J. 8. Harrington, and J. Van Houten, J. Am. Chem. Soc., 1977, 99, 2179; K. R. Dixon, Inorg. Chem., 1977, 16, 2618.CLASSIFICATION OF LIGANDS 121 4-11, Other Nitrogen Heterocycles” In addition to bipyridine and related heterocycles, there are numerous other N-heterocycles that give uni- or multidentate complexes. Some of the more im- portant are: oe he N-N K H Pyridazine Purine ree a NN YL NN’ Pyrazine 1,8-Naphthyridine e e N-N NA Pyrazolate Imidazolate One of the most important areas of concern for metal binding with nucleotides, purines, and pyrimidines arises because of their presence in nucleic acids.?° The action of certain metal complexes, notably cis-PtCl)(NH3)2, as anticancer agents, is believed to arise through binding to nucleic acids. Other aspects of the binding of metals to nucleic acid include the attachment of lanthanide ions as shift reagents and fluorescent probes and the use of heavy metals to assist in X-ray structural determinations. For unsubstituted purines, the most probable site for coordination is the imidazole nitrogen (N-9), which is protonated in the free neutral ligand. An example is the cobalt complex of adenine, [Co(ad)2(H20)s]* (4-VIII). HH, ] | OH, NAJN Z OH: (4-VI) 28M. Inoue and M. Kubo, Coord. Chem. Rev., 1976, 21, 1; B.C. Bunker et al., J. Am. Chem. Soc., 1978, 100, 3805; W. E. Hatfield, J.C.S. Dalton, 1978, 868 (pyridazine, pyrazine); J.G. Vos and W. L. Groenveld, Inorg. Chim. Acta, 1978, 27, 173 (pyrazolate). 29 B.E. Fischer and R. Bau, Inorg. Chem., 1978, 17, 27; D. J. Hodgson, Prog. Inorg. Chem., 1977, 23, 211 (stereochemistry of complexes); L. G. Marzilli, Prog. Inorg. Chem., 1977, 23, 256 (metal ion interactions); L.. G. Marzilli and T. J. Kistenmacher, Acc. Chem. Res., 1977, 10, 146; G. Pneumatikakis et al., Inorg. Chem., 1978, 17, 915 (Pa).122 INTRODUCTORY TOPICS If the 9-position is blocked, the other imidazole nitrogen, N-7, is coordinated. Binding appears somewhat less likely through N-1 than through N-7; but of the three complexes established by X-ray crystallography, two also involve binding with both N-1 and N-7. Imidazoles* have been widely studied. Although the binding is usually through the N atom (4-1X), in some Ru!!, Rull!, Fe®, and Cr° complexes it is possible to have C-bonded groups? (4-X). H 2NR, M— m—c€ sa " "= H (41x) (4X) (4xD (44x11) The C-bonded entity can be regarded as a carbene (4-XI) (see Chapter 27) or as a C-bound amidine?! (4-XII). An example of a C-bonded species is the rutheni- um(II) complex obtained as follows: 2 ((NH,),Ru" HOF + Nx NH — | sc oe N. omtione€ H The N-bonded imidazoles commonly form bridges between two metal atoms?? as in [Cu3(imH)s(im)2]**, and in (Mn(im)(TPP)THF},, where TPP is tetra- phenylporphyrin. Biimidazoles can act as mono or dianions,>3 for example, in rhodium(1) complexes: =, 5 Sey SoC < z La NON NON ™L YO 304 R. J. Sunderburg and R. B. Martin, Chem. Rev., 1974, 74, 471 (imidazole and histidine com- plexes) 30 R. J. Sunderberg et al., J. Am. Chem. Soc., 1974, 96, 381; Inorg. Chem., 1977, 16, 1470; S. S. Isied and H. Taube, Inorg. Chem., 1976, 15, 3070. D. J. Doonan, J. E. Parks, and A. L, Balch, J. Am. Chem. Soc., 1976, 98, 2129. 3G. Kolks et al., J. Am. Chem. Soc., 1976, 98, 5720; J. T. Landrum et al., J. Am. Chem. Soc., 1978, 100, 3232; M. S. Haddad et al., Inorg. Chem., 1979, 18, 141 3S. W. Kaiser et al., Inorg. Chem., 1976, 15, 2681 enCLASSIFICATION OF LIGANDS 123 4-12. Ligands Derived by Deprotonation of Ammonia and Amines: Dialkylamido, Nitrene, and Nitrido Complexes Ammonia can be deprotonated by alkali metals to give the anions NHz, NH?2-, and N3-, and all of these species can act as ligands. There are numerous examples of the amido ligand NH2 acting as a bridge, as in the ruthenium complex** i, M+ Pctitiin ica deal H, The imido ion NH2-, which is isoelectronic with O-, is not common as a ligand, although its alkyl and aryl derivatives NR are (see below). Some examples? of complexes, which have terminal or bent bridged NH groups, are the following species: pad N (B10) PS: hMoL Mel MOBY (diphos),C1,Mo=NH Nitrido Complexes have N3~ bound in the following ways: 1. Multiply Bonded Nitride M=N.*° Here the nitride ion is forming three covalent bonds to the metal; it is one of the strongest 7 donors known. The com- pounds are rather similar to those containing M=O groups (Section 4-23). The complexes are largely those of molybdenum, tungsten, rhenium, ruthenium, and osmium, examples being NReCls(PPh3)2, [NOsCls]2~, and [NOsO3]~. The M=N_ bonds are very short (ca. 1.16 A) and the M—N stretching frequencies are in the region 950-1180. cm=!. 2. N-Bridged Species. These are of the following types: 34M. T. Flood et al., Inorg. Chem., 1973, 12, 2153. 35. P. Cheney and J. N. Armor, Inorg. Chem., 1977, 16, 3338; A. W. Edelblut, B, L. Haymore, and R. A, D. Wentworth, J. Am. Chem. Soc., 1978, 100, 2251. W. P. Griffith, Coord. Chem. Rev., 1972, 8, 369. D. Pawson and W. P. Griffith, J.C.S. Dalton, 1975, 417; C.D. Cowman et al, Inorg. Chem., 1976, 15, 1747. 36124 INTRODUCTORY TOPICS L,.M—N—ML, L,.M-N—ML, Symmetric Asymmetric L. ou ET Triangular N-centered, where ~~ = bidentate chelate anion Ly) L= neutral ligand L L, i N L,MS=N. LM” | ML, s . L.M==N. Only a few singly bridged species are known.37 The bridges may be symmetrical (M—N—M) as in [Ru2NCls(H20).]3 and (TPPFe)2N (where TPP = tetra- phenylporphyrin), or asymmetric as in [ReN(CN)4]2-. Unlike oxo bridges (Section 4-23), there seem to be no bent MNM bridges. The linearity results from M—N—M 7 bonding. The ruthenium complex is used for electroplating of ruthenium. The triangular, N-centered complexes are much less common than O-centered triangles (Section 4-23) and no systematic way exists for making them. The metal atoms may be bridged by groups such as SO} or RCO}, and the charge on the complex depends on the anion. Since NM$"' has a 6+ charge, we can have complexes like [Ir3N(SOx)6(H20)3]*. Tetra-bridged (4*) species are very uncommon, and the only well-defined ex- ample is the tetrahedral ion [N(HgMe)s]*CIO;; this can be regarded as a type of quaternary ammonium ion. ‘The nitrogen atom in some of these compounds can be attacked. Thus tertiary phosphines react with OsNCI3(AsPh3)> to give phosphine imidate complexes R3PClOs—N=PR; in which there is d7-pa-dm bonding. The formation of a thionitrosyl complex with an Mo—N=S group by attack of S38 has already been mentioned (p. 93). An exceptional case of interaction is that with another metal atom as in the complex?? 4-XIII, where the Mo=N distance (ca. 1.65 A) is similar to that in N=Modtes, whereas the =N — Mo distance is 2.12 to 2.14 A. re Mo(S,CNR,), (R.NCS,)Mo==N~ (4-XII) 37D. A, Summerville and D. A. Cohen, J. Am. Chem. Soc., 1976, 98, 1747; D. Pawson and W. P. Griffith, J.C.S. Dalton, 1973, 1315. 38 J. Chatt and J. R. Dilworth, J.C.S. Chem. Comm., 1974, 508. 39M. W. Bishop et al., J.C.S. Chem. Comm., 1976, 781CLASSIFICATION OF LIGANDS 125 Dialkylamido ligands® are derived from secondary amines by deprotonation; for example, Et)NH + Bu"Li = LiNEt; + CaHio or by cleavage of a hydrazine [(MesSi)2N]2 + 2Li = LiN(SiMes)2 The complexes are obtained by reaction of lithium compounds with metal halides. Dialkylamides are closely related to alcoxides (Section 4-25) and alkyls, often having similar stoichiometries and structures [e.g., Cr(NEt2)4, Cr(OBu')s, Cr(CH2SiMe3)4]. The metal-nitrogen bond has some multiple bond character owing to flow to electrons from the lone pair into empty metal orbitals ~_ UR MN. SR This has the result that the M—NRz group is planar and that rotation of NR2 about the M—N axis is considerably restricted. This can lead in certain complexes to the inequivalence of R groups in nmr spectra. Dialkylamides can undergo insertion reactions (Chapter 29) as with CO, and CS) to give carbamates or dithiocarbamates; for example, MesTa(NMe2)2 + 2CS2 = MesTa(SxCNMe)2 Bulky dialkylamides are also very effective in preventing dimerization or poly- merization reactions, so giving compounds with unusually low coordination num- bers, such as two-coordination in Be[N(SiMe3)2]2 or three-coordination in Fe[N(SiMes)2]3. Alkylimido (Nitrene) Complexes.‘! By deprotonation of a primary amide we obtain the substituted imide RN2-. The ligand can be bound to a metal by a double bond, with px-dz overlap. Although in certain compounds like Ut PhP= the M—N—C moiety is bent, in transition metal complexes like O;Os—NBu! and (Me:N);Ta=NCMe;* there is a linear M—N—C unit. The differences in chemical reactivity between the nucleophilic bent compounds and the electrophilic 40 M.H. Chisholm, M. Extine, W. Reichert, in /norganic Compounds with Unusual Properties, ‘American Chemical Society, 1976; D, C. Bradley and M. H. Chisholm, Acc. Chem. Res., 1976, 9, 273; M. F. Lappert, A. R. Sanger, R. C. Srivastava and P, P. Power, Metal and Metalloidal Amides, Horwood-Wiley, 1979; D. C. Bradley, Adv. Inorg. Chem. Radiochem., 1972, 15, 259 41. Cenini and G. LaMonica, Inorg. Chim. Acta, 1976, 18, 279; J. R. Dilworth et al., J.C.S. Dalton, 1979, 914, 921; B. L. Haymore et al., J. Am. Chem. Soc., 1979, 101, 2063; W. A. Nugent and R. L. Harlow, J.C.S. Chem. Comm., 1979, 342. #21 W. A. Nugent and R. L. Harlow, J.C.S. Chem. Comm., 1978, 579.126 INTRODUCTORY TOPICS linear compounds reflects the increased m donation from the lone pair on nitrogen into low lying d orbitals of the metal, thus giving more triple-bond character to the M—N bond. Among the best known are the rhenium complexes, and these are made by condensation of oxorhenium species with aromatic primary amines such as aniline: O=ReCIs(PPH3)2 + ArNHo** ArN=ReCl3(PPh3)2 + H20 The NH*® or NR ligand can be referred to as a nitrene just as CH2 or CR3 li- gands can be considered as carbenes. There are a number of reactions in which nitrene complexes are believed to be intermediates. Isolation is usually impossible,*? as for example, in the photolysis of [Cr(NH3)sN3]2*, which forms Cr(NH3)sN2+, and in interaction of the complex [(NH3)sIrNH2Cl]3+ with hydroxide ion. How- ever in the reduction of RNO2 compounds by iron and other carbonyls (Chapter 30), intermediates with NR groups bound to metals have been isolated.*4 An ex- ample is the compound 4-XIV: Et N \ Fe(CO), (co) a co FCO), 4 0 (4-XIV) In some special cases, there may be other routes to nitrene complexes; for ex- ample, IrC(CO)(PMePha)2 + %CF3N==NCF3 = Ir( NCF3)C(CO)(PMePh,)2 The splitting of the N=N bond is doubtless favored by the electronegative CF3 groups, since az0 compounds generally retain the —N—=N— bond on reaction with metal complexes (see later). The reduction of NbCl, and TaCly in CH3CN by zinc gives complexes*5 with an unsaturated dinitrene ligand with linear M=C—N groups (4-XV) formed by dimerization of acetonitrile (cf. 4-IV). qaqa Me iy . Q ¢ MeCNNb=N—C. q aa Me cl 426 J. Chatt er al., Trans. Met. Chem., 1979; 4, 59. “3M. Katz and H. D. Gafney, Inorg. Chem., 1978, 17, 93; E. D. Johnson and F. Basolo, Inorg. Chem., 1977, 16, 554. 44 See, e.g... Aime et al., J.C.S. Dalton, 1978, 534. 4 PA. Finn et al., J. Am. Chem. Soc., 1975, 91, 220; F. A. Cotton and W. T. Hall, Inorg. Chem., 1978, 17, 3525,CLASSIFICATION OF LIGANDS 127 Methyleneamido (amino) and Related Ligands. Alkylimido compounds of the type just discussed that have alkyl groups can undergo the reaction*® L.M=Nq L,.M—Na cur Se Son os \ for example, y (PPh,Me),CLReNCH, = (PPh,Me)pyCl,Re—N, Hel s, CH, The ligand so formed, —N=CHb, methyleneamido, or the substituted groups ~-N=CRz, can be considered to resemble NO, since they can act not only as one-electron ligands but as three-electron ligands giving linear -bonded groups. The bonding modes are thus: Roe i—N- \, a. ma. UR of —R M—N=CO I R N, Ne wy Bent Linear Bridged There are a number of compounds with symmetrical bridges.4” In addition to the deprotonation reactions mentioned above, there is a more general method of synthesis involving reaction of metal complex halides with LiN=CR2,** which are in turn obtained from compounds such as PhyC—=NH or Bu3C=NH by action of Bu”Li. 4-13. Ligands with N—N Bonds‘? There are a large number of complexes of ligands with N—N bonds. They have been much studied in recent years because of their relationship, real or imagined, to the problem of the conversion of dinitrogen to ammonia or hydrazine and the reactions of coordinated N2 (Chapter 29). The major ligand types and the ways in which they can be bound are given in Table 4-1. Hydrazine usually acts as a reducing agent, but some compounds of NH, are known that have only one N coordinated, as in Zn(N2H4)2Cl2 or (7-CsHs)Re(C- O)2N2H4.5 Substitution of alkyl or aryl groups as in NH2NRz sterically inhibits 4 J, Chatt et al., JCS. Dalton, 1976, 2435. 47 G.P. Khare and R. J. Doedens, Inorg. Chem., 1976, 15, 86. 48M. Kilner et al, JCS. Dalton, 1974, 639, 1620. * J. R. Dilworth, Coord. Chem. Rev., 1976, 21, 29 (an extensive review); D. Sutton, Chem. Soc. Rev., 1975, 4, 443; A. Albini and H. Kisch, Topics Curr. Chem., 1976, 68, 105 (diazene and diazo complexes); D. L. DuBois and R. Hoffmann, Nouv. J. Chim., 1977, 1, 479 (MO picture of di- azenido, 1,2-diazene, imido, and nitrido complexes). SD. Sellmann and E, Kleinschmidt, Z. Naturforsch., 1977, 32b, 795.TABLE 4-1 Ligands with NN Bonds Ligand Structure Diazenido (diazenato)® 1,2-Diazene (a70)* 1,2-Diazene (hydrazido,2-)° Hydrazino M+—nR,—AR, Hydrazido(!-) MNHNR, M—NHNHR Triazenido M “G. Butler, et al., J.C.S. Dalton, 1979, 113; K. D. Schramm and J. A. Ibers, Inorg. Chem., 1977, 16, 3287; J. A. Carrol et al., Inorg. Chem., 1977, 16, 2462; D. Sutton, Chem. Soc. Rev, 1975, 4, 44 J.Chatt et al., J. Chem, Soc., 1977, 688; 1976, 1520; W. A. Hermann et al., Angew. Chem., Int. Ed., 1976, 164; M. Keubler et al., J.C.S. Dalton, 1975, 1081; W. E. Carron, M. E. Deane, and F. J. Lalor, JCS. Dalton, 1974, 1837, E, W. Abel and C. A. Burton, J. Organomet. Chem., 1979, 170, 229. + J, A. McCleverty, D. Seddon, and R. N. Whiteley, J.C.S. Dalton, 1975, 839, D. Sellman, J. Or- ganomet. Chem., 1973, 49, C22. © J.Chatt et al., J. Organomet. Chem., 1978, 160, 165; J.C.S. Dalton, 1977, 688; 1974, 2074; M. Veith, Angew. Chem. Int. Ed., 1976, 387; N. Wiberg, H, W. Haring, and O. Schneider, Angew. Chem., Int, Ed., 1976, 386; M. Herberhold and K. Leonhard, Angew. Chem. Int. Ed., 1976, 230. 128CLASSIFICATION OF LIGANDS. 129 the coordination of the substituted N atom, especially in the formation of octahedral complexes. However both nitrogen atoms of NoH4 and MeyNNH) can bridge two metals as in the complex 4-XVI_50 (4-XV1) Diazenido. Compounds containing the ligand NNR may also be referred to in the literature as arylazo, aryldiazo, or aryldiazenato, The aryl compounds are the most common; they can be obtained from diazonium compounds, ArN¢, and from hydrazines such as AYCONNH. They are also formed by electrophilic or nucleophilic attacks on dinitrogen compounds (Section 30-4); for example, Nets (diphos),Mo(N2) + RCOC! + HC] —> (diphos)CIMo(N2COR) n-CsHs(CO)2MnN3 > | inCAHLCO) MAN Ph ‘The ligand can be regarded as RN and to be the analogue of NO*, or as RN>, the analogue of NO™, or as neutral RN3, the analogue of CO, and there has been the same type of discussion concerning the extent of metal-ligand a bonding. As with NO compounds, the N=N stretches vary widely from ca. 2095 cm! (indi- cating M—N 7 bonding) down to ca. 1440 cm~! in bridging species.5! The main distinction is between the following types*? of complex: ae ine | @) (b) but other types of bonding’? of RNo groups are known namely: R ae N N i mC | and = M—N—N—R uM” R Type (a) is sometimes called linear or single bent, to distinguish it from (b) or double $06 TV. Ashworth ef al., J.C.S. Dalton, 1978, 1036. 51K. D. Schramm and J. A. Ibers, Inorg. Chem., 1977, 16, 3287. % D. T. Clark, et al., Inorg. Chem., 1977, 16, 1201; M. Cowie, B. L. Haymore, and J. A. Ibers, J. Am. Chem. Soc., 1976, 98, 7608. 53 B. L. Haymore, J. Organomet. Chem., 1977, 137, C11130 INTRODUCTORY TOPICS bent. Type (a), of which ArN2RuCl;(PPh3), is an example, can be regarded as derived from RN}. The M—N—N bonds are almost but not quite linear, the M—N-—N angle usually being ca. 170°. There is considerable M—N 7 bonding, but the character appears to be less than that in M—NO compounds.53 Type (b), of which [ArN2RhCltriphos]* is an example, show considerable variation in angles, but M—N—N is usually ca. 120°. They can be considered to be derived from RMN-. There are some iridium complexes of the neutral group CsCl4Nz acting as a two-electron donor [e.g., IrCl(N2CsCl4)(PPh3)2], which can be compared with IrCl(CO)(PPh3)2; these also have a type (a) bent structure.54 In the compound [PhN=NMn(CO)], the double bridge is asymmetric, with Mn—N—N angles of 134° and 119°.5> The bonding can be written as involving RNN as a neutral three-electron donor: The unidentate ligand RN# can be protonated (see below), and it can be hy- drogenated® to give a hydrazino complex; for example, H: (PhsP)PUNNAr* > (PhsP)2Pt(H)(NH2—NHA)* Diazene Ligands. Diazenes or az0 compounds RN=NR usually utilize the lone pairs on nitrogen as o donors rather than use the x electrons of the double bond in ethylenelike bonding. Azo compounds of many types include azo dyestuffs. Both cis and trans azo compounds can give complexes normally unidentate but the cis azo compound pyridazine gives only bridged complexes.5? A feature of aromatic azo compounds 1s that the C—H of the ortho position on the aromatic ring is reactive and can undergo the cyclometallation reaction (Section 29-6), with formation of M—H and M—C bonds. 4 DO : M, Mu NAW I fan ql *O N oO K. D. Schramm and J. A. Ibers, J. Am. Chem. Soc., 1978, 100, 2932. 58 M.R. Churchill and K.-K. G. Lin, Jnorg. Chem., 1975, 14, 1132. 56S. Krogsrud ef al., J. Am. Chem, Soc., 1977, 99, $277. M.N. Ackermann ef al., Inorg. Chem., 1977, 16, 1298.CLASSIFICATION OF LIGANDS 131 Unsibstituted diazene complexes can be made,°* for example, by oxidation of hydrazine complexes: Hanne x(00), #0, N= mCp(CO);MnNH,—NH;Cr(CO), —> 1-Cp(COMn’ In the monosubstituted aryldiazene derivative 4-XVII, which is in the cis form, the metal-nitrogen bond is largely °°: N—Ph PAO. H (4-XVII) Such compounds are related to diazenato complexes by acid-base equilib Nr 1, M—n7 — MON a + our H There are fewer examples of N,N-disubstituted diazenes, which are isomeric with RN=NR compounds. However the simple hydrogen compounds can be made, for example, by protonation of dinitrogen compounds HBF (diphos))Mo(N2)2 —> [(diphos)2F MoNNH2]* The M—N-—NH> group is essentially linear.®! Diazenes with both nitrogen atoms bound to a metal in a three-membered ring are not common, but one example is (R3P)2Ni(PhN=NPh).*? 1,3-Triazenido Complexes.©? The compounds AYNNHNAr are acids giving rise to a monoanion, ATNNNAr-~. These ligands can be unidentate, chelate, or Ho-bridging: R eee MN” i Sy iia \n Nye t R iN, Np 58D. Sellman and K. Jédden, Angew. Chem. Int. Ed., 1977, 16, 464. 5° B.L. Haymore and J. A. Ibers, J. Ant. Chem. Soc., 1975, 97, $369. 69 R. Mason et al., J. Am. Chem. Soc., 1974, 96, 261 6! M. Hidai et al., Inorg. Chem., 1976, 15, 2694; G. A. Heath, R. Mason, and K. M. Thomas, J. Am. Chem. Soc., 1974, 96, 259. © $.D. Ittel and J. A. Ibers, Inorg. Chem., 1975, 14, 1183. © N.G. Connelly and Z, Demidowicz, J.C.S. Dalton, 1978, 50 and references therein; P. I, van Vliet et.al, J. Organomet. Chem., 1976, 122, 99; L. D. Brown and J. A. Ibers, J. Amt. Chem. Soc., 1976, 98, 1597; Inorg. Chem., 1976, 15, 2788, 2794132 INTRODUCTORY TOPICS The structures of complexes closely resemble those of the considerably more im- portant carboxylates (Section 4-28). If the central N atom is replaced by CR’ (R’ = H, CHs, etc.), we obtain N,N’- disubstituted formamidino or alkyl- or arylamidino ligands®4 (4-X VIII). If re- placed by SR’, we obtain the sulfurdiimines® (4-XIX). RX Rr R | R R RR \AN/ \ow! (4-XVIIL) (4-XIX) The neutral ligand RN=S—=NR also gives complexes of various types.° Azide.S7 The azide ion N3 can give unidentate complexes and also those that are bridged®*: M. ‘Z—Zz—Z N—N—N. Sn—n—n uo = M SM Syon—n7 c™ MoM Examples of singly and doubly bridging azides are the nickel complexes [Nio(N- tetramethylcylam)3N3]* and [Niz(tren)9(N3)2]*, respectively; the symmetric bridging is found in Cu(N3)>. 4-14. Macrocyclic Nitrogen Ligands® The large ring compounds whose structures are such that several donor atoms can bind to a metal are most commonly nitrogen donors. However mixed nitrogen- oxygen, nitrogen-sulfur, and oxygen-sulfur donors are known, as well as macrocyclic oxygen and sulfur donors. Depending on the donor atoms, these can be designated Na, N202, Ox, and so on. 64 W.H. de Roode et al., J. Organomet. Chem., 1978, 184, 273; 1978, 145, 207; M.G. B. Drew and J.D. Wilkinson, J.C.S. Dalton, 1974, 1973; F. A. Cotton et al., Inorg. Chem., 1975, 14, 2023, 2027. 85K. Vrieze et al., J. Organomet. Chem., 1977, 142, 337; 1978, 144, 239. 6 R. Meij, JCS. Chem. Comm., 1978, 506. 67. Doriand R. F. Ziolo, Chem. Rev., 1973, 73, 247. ©% F. Wagner et al., J. Am. Chem. Soc., 1974, 96, 2625; C. G. Pierpont et al., Inorg. Chem., 1975, 14, 604; D. M. Duggan and D, N. Hendrickson, Inorg. Chem., 1973, 12, 2422. Synthetic Multidentate Macrocyclic Compounds, R. M. Izatt and J. J. Christensen, Eds., Academic Press, 1978 (this reference work discusses mainly oxygen macrocycles); D. H. Busch, Ace. Chem. Res,, 1978, 11, 392; L. F. Lindoy, Chem. Soc. Rev, 1975, 4, 421 (transition metal complexes of synthetic macrocycle ligands); A. M. Tait, D. H. Busch, ef al., Inorg. Synth., 1978, 18, Chapter 1: R.M. Izatt and J. J. Christensen, Eds., Progress in Macrocyclic Chemistry, Vol. 1, Wiley, 1979; G. A. Melson, Ed.. Coordination Chemistry of Macrocyclic Compounds, Plenum, 1979 “CLASSIFICATION OF LIGANDS 133 The heterocyclic compounds can be broadly classed into those with conjugated m™ systems and those without. The former, especially macrocycles giving a set of four essentially coplanar N atoms, have been extensively studied in part because these types of systems are involved in chlorophyll, heme, vitamin B,2 and other naturally occurring metal complexes (Chapter 31). Macrocyclic complexes characteristically have the following properties. 1. A marked kinetic inertness both to the formation of the complexes from the ligand and metal ion, and to the reverse, the extrusion of the metal ion from the ligand. 2. They can stabilize high oxidation states that are not normally readily at- tainable, such as Cu!!! or Nil! 3. They have high thermodynamic stability—the formation constants for Ng macrocycles may be orders of magnitude greater than the formation constants for nonmacrocyclic Ng ligands.” Thus for Ni?* the formation constant for the macrocycle cyclam (4-XX) is about five orders of magnitude greater than that for the nonmacrocycle tetradentate 4-XXI: * 2 gee oN ot \ 7 Pe of A ee wi Sw a a ae ea (4-XX) (4-XXI) This “macrocyclic effect” has been discussed thermodynamically in Section 3-5. Ligands with Conjugated 7 Systems Phthalocyanines.”! These were one of the earliest classes of synthetic Ny macro- cycles to be discovered. They are obtained by interaction of phthalonitrile with metal halides, in which the metal ion plays an essential role as a template. Complexes such as 4-XXII characteristically have exceptional thermal stabilities, subliming in vacuum around 500°C. They are also intensely colored and are an important class of commercial pigments. The solubility is usually very low, but sulfonated deriva- tives are soluble in polar solvents. 7° M. Kodama and E. Kimura, J.C.S. Dalton, 1978, 1081; A. Anichini et al., J.C.S. Dalton, 1978, $77; F. P. Hinz and D. W. Margerum, J. Am. Chem. Soc., 1974, 96, 4993, 7! A. B.P. Lever, Adv. Inorg. Chem. Radiochem., 1965, 7, 28.134 INTRODUCTORY TOPICS (4-XXII) Porphyrins.’? These ligands are derivatives of porphine (4-XXIII). They are especially important because many naturally occurring metal-containing natural products (chlorophylls, heme, cytochromes, etc.) contain related macrocyclic li- gands. GHs CH, Ph Hy CH, Ph Ph CH CoH, GH, GH Ph (4-XXID (4-XXIV) (4-XXV) Almost every metal in the Periodic Table can be coordinated to a porphyrin, The most widely used synthetic ligands apart from porphine itself are octaethylporphyrin (H2OEP) (4-XXIV) and meso-tetraphenylporphyrin (H2TPP) (4-XXV)73; in the latter the phenyl rings are free to rotate. Although the most common form of coordination is that with a metal atom in the center of the plane and bound to four nitrogen atoms, nevertheless porphyrins can act as bi-, tri-, tetra-, or hexadentate ligands in which the metal atom lies out of the Ng plane,”4 as discussed below. The metal atom is also 0.448 A out of the plane in the iron(III)protoporphyrin IX dimethyl ester p-nitrobenzthiolate (4- XXVI),”5 where we use a diagrammatic representation of the Ng prophyrin skel- eton. 7D. Dolphin, Ed., The Porphyrins, Academic Press, 1978. (an authoritative reference in seven volumes); J. F. Falk, Porphyrins and Metalloporphyrins, Elsevier, 1964; W. R. Scheidt, Acc. Chem. Res., 1977, 10, 339; K. M. Smith, Ed., Porphyrins and Metalloporphyrins, Elsevier, 1975; S.J. Chantrell et al., Coord. Chem. Rev., 1975, 16, 259 (MO calculations); J.-H. Fubrhop, Struct. Bonding, 1975, 18, 1; Angew. Chem., Int. Ed., 1974, 13, 321; 1976, 15, 648; J.-H. Fubrhop and K.M. Smith, Laboratory Methods in Porphyrin and Metalloporphyrin Research, Elsevier, 1975; P. Hambright, Coord. Chem. Rev., 1971, 6, 247; D. Dolphin and R. H. Felton, Acc. Chem. Res., 1974, 7, 26; J. W. Bichler, Angew. Chem., Int. Ed., 1978, 17, 407 (hemoglobin); F. R. Longo, Ed., Porphyrin Chemistry Advances; Wiley, 1979; J. W. Buchler et al., Siruct. Bonding, 1978, 34,79, 73.8, Eaton and G. R. Eaton, J. Am. Chem. Soc., 1977, 99, 6594, 74 G. A, Taylor and M. Tsutsui, J. Chem. Educ., 1975, 52, 715; D. Ostfeld and M. Tsutsui, Acc. Chem. Res., 1974, 7, 52. 75 S.C. Tang et al., J. Am. Chem, Soc., 1976, 98, 2414,CLASSIFICATION OF LIGANDS 135 cfty (4-XXVI) Complexes of synthetic porphyrins can be made by interaction of the ligand with a metal salt in a common solvent, usually dimethylformamide.’¢ Although HXOEP. has high solubility in organic solvents, H2TPP is less soluble, but sulfonated de- tivatives are soluble in both water and methanol. Complexes can also be obtained by interaction of H2porph with metal carbonyls, acetylacetonates, alkyls, hydrides, and so on. The radius of the central hole is, of course, fixed, although it can be altered to some extent by puckering of the rings, and it lies between 1.929 and 2.098 A. This means that many metal atoms cannot fit in the hole and must form out-of-plane complexes. Some examples are the following: 76 W. Schneider, Struct. Bonding, 1975, 23, 123 (kinetics and mechanism of metalloporphyrin formation).136 INTRODUCTORY TOPICS Polymeric compounds with metal atoms between porphyrin groups are also known. There have been recent attempts to synthesize porphyrin ligands that will act as models for hemoglobin (Section 31-3) in that absorption of oxygen is truly re- versible, rather than as is usual with Fe!! porphyrin complexes, leading to irre- versible oxidation to Fe!!! porphyrin. These attempts involving “picket-fence” porphyrins are discussed in Section 4-24. Another type of modification is the por- phyrin “crowned” by a cryptate (Section 4-26) ring that can accommodate two metal atoms”? (4-X XVII). A cap can also be provided by pyridine groups as indi- as W070 ‘0 bu, oN (4-XXVI1) (4-XXVIII) cated diagrammatically in 4-XXVIII. Yet another type is the strati-bisporphyrin”8 (4-XXIX). (4-XXIX) Reproduced by permission from Ref. 78 7 C.K. Chang, J. Am. Chem. Soc., 1979, 101, 3413; 1977, 99, 2819; D. A. Buckingham er al., J. Am. Chem. Soc., 1978, 100, 2899. 78 N.E. Kagan, D. Mauzerall, and R. B. Merryfield, J. Am. Chem. Soc., 1977, 99, 5484,CLASSIFICATION OF LIGANDS. 137 Other important conjugated macrocycles are the corrins (Section 31-9), which occur in vitamin B,2. Conjugated systems such as dibenzotetraaza-[14]-annulene (4-XXX)79 have a smaller coordination radius than porphyrins and usually form out-of-plane complexes. (4-XXX) Template Syntheses.*° These are reactions in which the presence of the metal ion controls the synthesis. We have already noted this feature in the synthesis of phthalocyanins from phthalonitrile. A great variety of nitrogen-containing macrocycles can be made by employing the Schiff base condensation reaction (eq. 4-1), often (but not necessarily) with R. R. Yo=o + H.NR” — Doan, +#H.O a) R R” R” a metal ion as a template, and with subsequent hydrogenation to obtain a saturated system not subject to hydrolytic degradation by reversal of reaction 4-1. Some representative preparative reactions are eqs. 4-2, 4-3 and 4-4, H TS N N N as Nit =] +4Meco— [ Ni# (42) wt’ a, N H HH, “pf +200 — (he us rN. 00 ry HY OMA BN Ma 79 V.L. Goedken et al., J. Am. Chem. Soc., 1976, 98, 8391 #0 M. de S. Healy and A. J. Rest, Adv. Inorg. Chem. Radiochem., 1978, 21, 1; SM. Peng et al., Inorg. Chem., 1978, 17, 119, 820; N. F. Curtis, Coord. Chem. Rev., 1968, 3, 1; J. J. Christensen et al., Chem, Rev., 1974, 74, 651; D. St. C. Black and A. J. Hartshorn, Coord. Chem. Rev., 1972-1973, 9, 219; T. J. Marks and D. R. Stojakovic, J. Am. Chem. Soc., 1978, 100, 1695.138 INTRODUCTORY TOPICS NONE+ON (44) Niw HN NH iH N. H Other good examples*! are as follows. 1. The condensation of o-aminothiophenol with pyridine-1-carboxaldehyde gives benzthiaazoline in absence of metal ions: SH S. N= Zt OC O-OL, 9-009 Benzthiazoline Schiff base In the presence of metal ions, any small amount of the Schiff base that may be in equilibrium with benthiaazoline will be removed to give a metal complex. 2. The self-condensation of o-aminobenzaldehyde in presence of BF3-OEt, in acetic acid gives a macrocycle8? and when metal ion is present also, a complex, for example, NH, No N= Ome CON ‘CHO, ne N 3. The reaction of chelate complexes as follows: N ZN\e NH \aZ Nye A Me — Be OH § 3 H, H, N, 8! D.H. Busch et al., Inorg. Chem., 1977, 16, 1716, 1721; J. Lewis and K. P. Wainwright, J.C.S. Dalton, 1978, 440. *2 J.D. Goddard and T. Norris, Inorg. Nucl. Chem. Lett., 1978, 14, 221.CLASSIFICATION OF LIGANDS 139 Finally, note that there must be some control of the reaction by the size of the metal used. If the ion is too small or too large, no macrocyclic complex may be formed.§? Ligands Without Conjugated 2 Systems. There are a large number of macro- cyclic N ligands that have double bonds in part of the ring only or are completely saturated.84 There is often a very substantial difference between the conjugated and saturated macrocycles with regard to the rates of substitution reactions, which may be up to 10!? times greater for the conjugated systems. The enormous rate enhancement can be correlated with the lifetime of the leaving ligand in the inner coordination sphere of the metal complex. This profound difference presumably explains why biologically active metal systems invariably have highly unsaturated macrocyclic ligands. Some examples of Ng macrocycles have already been shown, others are the fourteen-membered ring compounds®s 4-XXXJ and 4-XXXII; there are similar tetraenes with fifteen- and sixteen-membered rings.** B ei Me. u Me Me y=N Nae N ° N Me: * SMe 7 y ‘Me st a H ¥ Me,[14]-1,3,8,10-tetraene-N, Me,[14]-1-ene-N, (4-XXXI) (4-XXXII) 4-15. Schiff Base Ligands®’ Schiff base ligands are very diverse and usually contain both N and O donor atoms, although purely N donors are known. There are also N and S donors. 3D. H. Cook ef al, J.C.S. Dalton, 1977, 446; D. E. Fenton et al., J.C.S. Chem. Comm., 1977, 623. 84 See, e.g., A. M, Tait, F. V. Lovecchio, and D. H. Busch, Inorg. Chem., 1977, 16, 2206; R. F. Pasternack, M. A. Cobb, and N. Sutin, Inorg. Chem., 1975, 14, 866; T. N. Margulis and L. J. Zompa, J.C.S. Chem. Comm., 1979, 430. 85 A.M. Tait and D. H. Busch, Inorg. Chem., 1979, 18, 1555. 86D. D. Riley, J. A. Stone, and D. H. Busch, J. Am, Chem. Soc., 1976, 98, 1752. 87 R.H. Holm, G. W. Everett, Jr, and A. Chakravorty, Prog. Inorg. Chem., 1966, 7, 83; M. Calli- garis, G. Nardini, and L. Randaccio, Coord. Chem. Rev., 1972, 7, 385; U. Casselato, M. Vidali, and P. A. Vigato, Coord. Chem. Rev., 1977, 23, 31; D. G. Hodgson, Prog. Inorg. Chem., 1975, 19, 173 (Schiff base dimeric complexes of first row transition metals); C. M. Harris and E. Sinn, Coord. Chem. Rev.. 1969, 4, 391 (Schiff base complexes as ligands); L. Sacconi, Coord. Chem. Rev., 1966, 1, 192; $. Yamada, Coord. Chem. Rev., 1966, 1, 415 (stereochemistry); C. A. McAuliffe et al., J.C.S, Dalton, 1977, 1762; D. E. Fenton and S. E, Gayda, J.C.S. Dalton, 1977, 2095.140 INTRODUCTORY TOPICS One of the best known Schiff base ligands is bis(salicyclaldehyde)ethylenedi- imine®® (salyen): H. H a CHO So=ncH,cHn=c’ OL + NH,CH,CH,NE, —#2, B© HO ‘OH ‘OH This is a bifunctional (two OH groups), tetradentate (2N, 20) ligand. Other Schiff bases can be mono-, di-, or tetrafunctional and can have denticities of 6 or more with various donor atom combinations (e.g., for quinquedentate, N30; N03; N202P; N20)S, etc.). Complexes of un-ionized or partly ionized Schiff bases are also known’? (e.g., LaCl;salzenH-aq). Some representative types of complex that illustrate not only the formation of mononuclear but of binuclear and polymerig¢ species are 4-XXXIII to 4- XXXVI, CH,——CH, Ph, \ Q alte oo - OOO O30 (4-XXXIII) co (4-XXXIV) (4-XXXVI) ‘SM. D. Hobday and T. D. Smith, Coord. Chem. Rev., 1972-1973, 9, 311. 8% J. 1, Bullock and H.-A. Tajmir-Riahi, J.C.S, Dalton, 1978, 36.CLASSIFICATION OF LIGANDS 141 4-16. Polypyrazolylborate Ligands” The interaction of pyrazole (itself a ligand: see Section 4-11) and sodium borohy- dride or alkyl-substituted borohydrides leads to anionic ligands of the type 4 XXXVII and 4-XXXVIII designated R2Bpzz and RBpz;, respectively. The boron atom is tetrahedral, and bonding to metal is through only one of each of the pyrrole ring N atoms as shown, Q No ~ had RB. = \ as ee oO & (4-XXXVII) (4-XXXVII) The bridge in [Cu(HBpz3)]2, shown diagramatically in 4-XXXIX, so far is unique.9!* (4-XXXIX) The dipyrazolyl anions R2Bpzz have a formal analogy to the 8-diketonate ions (Section 4-27) and, like them, form complexes of the type (R2Bpz2)2M. However because of the much greater steric requirements of the R2Bpz7 ligand, such com- pounds are always strictly monomeric. For steric reasons it appears to be difficult to make tris complexes, and only one example®!® is known, namely, the anion [V(H2Bpz2)3) ~~ The RBpz; ligands give a number of unusual complexes. These ligands them- selves are unique in being the only trigonally tridentate, uninegative ligands. They form trigonally distorted octahedral complexes, (RBpz3)2M®*, with di- and tri- valent metal ions, most of which are exceptionally stable. At least to a degree, an 90 A. Shaver, in Organometallic Chemistry Reviews, Vol. 3, Elsevier, 1977; S. Trofimenko, Chem. Rev., 1972, 72,497. 91a C. Mealli et al., J. Am. Chem. Soc., 1976, 98, 711. 91> P. Dapporto et al., Inorg. Chem., 1978, 17, 1323.142 INTRODUCTORY TOPICS analogy can be made between RBpz3 and the cyclopentadienyl anion CsH;; both are six-electron, uninegative ligands. There are some mono-RBpz; complexes that bear considerable resemblance to half-sandwich complexes, CsHsML, (see Chapter 27); thus Mo(CO)< reacts with Na(RBpz3) and NaCsHs to give, re- spectively, (RBpz3)Mo(CO)3 and (7-CsHs)Mo(CO);. 4-17. Nitriles?? Acetonitrile and other organic cyano compounds are good donors and form com- plexes with most Lewis acids and metal ions. The bonding is usually end on through nitrogen (i.e, RCN — M). Although “side-on” 72 bonding (4-XL) has been pro- vk “et M. appears to proceed by nucleophilic attack of OH~: RT [am,CoN==CR]* + OH™ = | am;Co—N—C I 0 The catalytic hydration of nitriles to amides by rhodium and other transition metal complexes such as Rh(OH)(CO)(PPh3)2 probably proceeds by a cycle*® 92 B.N. Storhoff, Coord. Chem. Rev., 1977, 23, (transition metal complexes); R. A. Walton, Q. Rev., 1965, 19, 126. 93 B, Storhoff and A. J. Infante, Inorg. Chem., 1974, 13, 3044; W. J. Bland, R. D. W. Kemmitt, and R. D. Moore, J.C.S. Dalton, 1973, 1292; see also J. E. Sutton and J. 1. Zink, /norg. Chem., 1976, 15, 675. % D. J. Cole-Hamilton and G. Wilkinson, J.C.S. Dalton, 1979, 1283. 98 R. J. Balahura et al., J. Am. Chem. Soc., 1974, 96, 2739. 96 M.A. Bennett and T. Yoshida, J. Am. Chem. Soc., 1973, 95, 3030.CLASSIFICATION OF LIGANDS 143 Ly L,RhOH + RCN == L,RhN= H,0 RCONH, L,RhN—C, Nucleophilic attacks on coordinated nitriles by aromatic amines and by alcohols give, respectively, complexes of amidines and imidate esters,.” for example, hHN. q Ol NS c=N. | Lo Me Rec’ cv | YN NHPh a HS PI (CH,CN),ReCl, + 2PhNH, —> ‘Me We noted earlier in this chapter the reduction of nitrile ligands to give other types ‘of complex (pp. 118, 126). 4-18. Oximes?* and C-Nitroso Compounds Oximes are derived by condensation of aldehydes and ketones with hydroxylamine. The best known are the cis-dioximes such as dimethylglyoxime (4-XLI) which, with Ni?* in ammonia solution, gives the well-known red nickel complex 4-XLII. -Q qe ee Me. OH c=—N N= No=n7 V7 i i | .==N. Me on v \ Me OHO (4-XLI) (4-XLIL) ‘The important features here, apart from Ng binding, are first the strong O—H—O- hydrogen bonding, and second the stacking of the planar units parallel to each other in the crystal. Oxime complexes have formed the basis for the construction of clathrochelate or encapsulating ligands in which three oxygen atoms of oxime ligands are “capped” by a group such as BF as shown in 4-XLIII (see also 3-XXI and 3-XXII). 97 J. M. Castroand H. Hope, Inorg. Chem., 1978, 17, 1444; M. Wada and T. Shimohigashi, Inorg. Chem., 1976; 15, 954. See also A. M. Sargeson et al., Acta Chem. Scand., 1978, 32A, 789. 98 R.C. Mehrotra et al., Inorg. Chim. Acta, 1975, 13, 91; A. Chakravorty, Coord. Chem. Rev., 1974, 13, 1; B. Chatterjee, Coord. Chem. Rev., 1978, 26, 281 (hydroxamic acid complexes); A. Naka- mura et al., J.C.S. Dalton, 1979, 488 (chiral oximates).144 INTRODUCTORY TOPICS Me. Me ne ot NA FB: NS Dy BF oor C M s (4-XLIID) Such compounds can be obtained by interaction of dimethylglyoxime with a metal salt in the presence of BF; or B(OH);3%; an example is Fe(dmg)3(BOH)2. Some of these clathrochelates, for example; FB(ONCHCsH3N)3P~, can impose trigonal prismatic geometry. Instead of BF capping, clathrochelates with a metal complex cap {c.g., diethylenetriaminechromium(III)] can also be made.'0° Monooximes with a different functional group!®! [e.g., 2(2-hydroxyethy!) imino-3-oximobutane] can give complexes such as 4-XLIV. Me H 7 No=n—07 gut CHOON NON, C Y CH, cH,—0% Yo—Nn% 7 Me" “S N(CH,),OH : \ I co. Me“ Me (4-XLIV) Oxime complexes are used commercially for the extraction of metals by com- plexing and extraction into organic solvents.!9%@ For example, copper is so removed from dilute copper sulfate solutions obtained by wet microbiological leaching of low-grade ores. Oxime compounds that have the chelate function Oo N=o \ / 99 B, Larsen et al., Inorg. Chem., 1972, 11, 2652; B. B. Fleischer et al., Inorg. Chem., 1972, 11, 2775; D. R. Boston and N. J. Rose, J. Amt. Chem. Soc., 1973, 95, 4163; M. R. Churchill and A. H. Reis, Jt., Inorg. Chem., 1972, 11, 2239. 100 R. S. Drago and J. H. Elias, J. Am. Chem. Soc., 1977, 99, 6570. 101 J. A. Bertrand, J. H. Smith, and P. G. Eller, Inorg. Chem., 1974, 13, 1649. 102 A, W. Ashworth, Coord. Chem. Rev., 1975, 16, 285.CLASSIFICATION OF LIGANDS 145 occur in nature in the green iron(II) pigment ferroverdin.!02> Such compounds are best regarded as derived from hydroxamic acids. Monooximes of the type R2C=NOH can be bound also in different ways (4- XLVa,b)!®3; in type a the ligand can be regarded as a three-electron donor Mo. oJ Yeo N—-O Me—' of Me (4-XLVa) (4-XLVb) Nitroso alkanes or arenes,'°* RNO, can be bound in complexes in several ways: The complexes can be made directly from compounds like nitrosobenzene, PhNO, or from substituted hydroxylamines by reactions such as 0 oO } (Me,.NCS. 1 Mo + RNHOH = (Me.NCS,), hc} +HO 0 The three-membered rings have been termed! metallooxaziridines. Such compounds can undergo quite facile N—-O bond cleavage!®4+ and nitrene, 102 L.A. Epps et al., Inorg. Chem. 1977, 16, 2663. 103 (a) R. B. King and K. N. Chen, Inorg. Chem, 1977, 16, 1164; G. B. Khare and R. J. Doedens, Inorg. Chem., 1977, 16, 907. (b) S. Aime et al. Chem. Comm., 1976, 370; G. P. Khare and R. J. Doedens, Jnorg. Chem., 1976, 15, 86. 104 (a) K. B. Sharpless et al... Am. Chem. Soc., 1978, 100, 7061; F. Mares et al., J. Am. Chem. Soc., 1978, 100, 7063; D. B. Sams and R. L. Doedens, Inorg. Chem., 1979, 18, 151. (b) S. Otsuka e¢ al., Inorg. Chem., 1976, 15, 657. (c) D. Mansuy et al., J. Am. Chem. Soc., 1977, 99, 6441; J.J. Watkins and A. L. Balch, /norg. Chem., 1975, 14, 2720. (d) K. Wieghardt et al., Angew. Chem. Int. Ed., 1979, 18, 548, 549.146 INTRODUCTORY TOPICS :NR, groups can be transferred to alkenes, cyclohexanone, or isocyanides, for ex- ample, Ni(ButNC), PhNO + Bu'NC PAN = CN‘Bu oO 0 NHR 0 My ann WW lo. fessca lo + NS Nr RNO compounds are also intermediates in the insertion of NO into metal alkyls (Section 29-11), but these can react further depending on whether the initial product is paramagnetic, in which case an N-methy]-N-nitroso-hydroxylaminato chelate is formed: Me 7 O—N 7 No | WMe, + NO —> Me,W—O—N—Me => Me,W > N Nf or is diamagnetic, in which case MeN is transferred to NMe: 7; 2ReOMe, + 2NO —* 2ReX(ONMe)Me, — 2MeRew + MeN = NMe 0 A closely related complex made from hydroxylamine! can be regarded either as protonated oxaziridine or merely as a 7?-hydroxylaminato complex: 4-19. Other N Ligands Biguanide [HNC(NH)NHC(NH)NH3] has a large proton affinity in aqueous solution, its base strength being only slightly less than OH~ (pK = 13.25). It forms stable complexes!°S* as a bidentate ligand, has a strong preference for square coordination as in Ni(bgu)?*, and finally tends to stabilize high oxidation states such as Ag!!! in the complex 4-XLVJ. 1051 L. Fabbrizi et al., Inorg. Chem., 1978, 17, 494.CLASSIFICATION OF LIGANDS. 147 : a Ce (4-XLVI) Guanidines such as arginine and creatine can also give complexes.!°5° Thus a hydrogen atom of creatine can be replaced by PhHg to give the zwitterionic com- plex. PhHg. peae NS A oN: NH? CH,CO,- PHOSPHORUS, ARSENIC, ANTIMONY, AND BISMUTH!°% In this section we deal only with phosphorus compounds as ligands. In general the behavior of analogous trivalent arsenic and antimony ligands resembles that of the phosphorus compounds. Relatively few compounds of antimony and even fewer of bismuth are known, and none are of great importance. Little further need be said concerning the complexes of compounds of arsenic, antimony, and bismuth. The o-donor ability decreases in the order P > As > Sb > Bi, and bismuth is a very weak donor indeed. Steric effects due to the donor atom itself will increase as follows: 4036 A. J. Canty et al., Inorg. Chem., 1978, 17, 1467. 106 For phosphines see Organic Phosphorus Compounds, G. M. Kosolapoff and L. Maier, Eds., Wiley; see also under phosphorus x bonding, Chapter 3, page 87; O. Stelzer, in Topics in Phosphorus Chemistry, Vol.9, Wiley, 1977, p. | (extensive review); C. A. McAuliffe, Ed., Transition Metal Complexes of Phosphorus, Arsenic and Antimony Ligands, McMillan, 1973; C. A. McAuliffe and W. Levason, Phosphine, Arsine and Stibine Complexes of Transition Metals, Elsevier, 1978; W. A. Levason and C. A. McAuliffe, Ace. Chem. Res., 1978, 11, 363 (organostibine complexes); G. Booth, Adv. Inorg. Chem. Radiochem., 1964, 6, 1; W. Levason and C. A. McAuliffe, Coord. Chem. Rev., 1976, 19, 173 (P, As, and Sb complexes of main group elements); C. A. McAuliffe, Adv. Inorg. Chem. Radiochem., 1975, 17, 165 (complexes of open-cliain tetradentate ligands); W. A. Levason and C. A. McAuliffe, Adv. Inorg. Chem. Radiochem., 1972, 14, 173 (complexes of bidentate P ligands); A. N. Hughes and K. Wright, Coord. Chem. Rev., 1975, 18, 239 (cyclic phosphine complexes); P. Rigo and A. Turco, Coord. Chem. Rev., 1974, 13, 133 (phosphorus- cyanide complexes); J. Verkade, Coord. Chem. Rev., 1972-1973, 9, 1, 106 (spectroscopic studies of phosphite complexes); D. G. Holah, A. N. Hughes, and K. Wright, Coord. Chem. Rev., 1975, 15, 239 (complexes of cyclic phosphines, phospholes, etc.); P.G. Eller et al., Coord. Chem. Rei 1977, 24, | (three-coordination review includes RsP complexes of Group VIII metals); D. M. Roundhill et al., Coord. Chem. Rev., 1978, 26, 263 (complexes of phosphinates and secondary phosphites).148 INTRODUCTORY TOPICS P As> Sb. We have discussed the bonding of phosphorus trihalides, notably PF3, and tertiary phosphines, together with some other aspects of their ligand behavior in Chapter 3. We now consider additional types of P ligand. 4-20. Tertiary Phosphine Ligands In addition to the unidentate R3P ligands, there are a number of multidentate phosphines. Diphosphine ligands (diphos) such as PhyPCHCH2PPh; are almost invariably chelated, although there are examples [e.g., for W(CO)s(diphos)] of unidentate and of bridging behavior.!07 Bidentate phosphines with only one bridging group such as PhyP—CH,— PPh2!° and CH3N(PF3)2 tend to promote metal-metal interaction or bond for- mation because the two donor P atoms are so close together: cH, R.P7 oPR: 1 M----M The use of chelate phosphines with many bridging groups giving long flexible chains has quite a different effect. For example, the chelate phosphines Bu}P(CH))ioPBu} or Bus3PC=C(CH2)sC=CPBu} can give complexes!!® that have as many as 72 atoms in a ring of type 4-XLVII. Similar bidentates can span trans positions as shown diagrammatically in 4-XLVII.!!) Gree RP. c1—Pa—c ci—Pa—cl oS | | 7 ph, R,P—CH,— (CH;)s—CH;—PR, (4-XLVI1) (4-XLVII1) A rather special chelate!!? may form where triphenylphosphine is 7-bonded via a phenyl ring in a compound like 4-XLIX. Substituted ferrocenes (Chapter 27) like 4-L can also be made. 107 R.L, Keiter et al., J. Am. Chem. Soc., 1977, 99, $224; A. L. Balch, J. Am. Chem. Soc., 1976, 98, 8049, 108 A A.M. Ally et al., Angew. Chem. Int, Ed., 1978, 17, 125. 1 M.G. Newton et al., JCS. Chem. Comm., 1978, 514 NOB, L. Shaw er al., JCS. Chem, Comm., 1977, 311; J.C. Dalton, 1976, 322; A. R. Sanger, JCS. Dalton, 1977, 1971 "1 LOM. Venanzi et al, Helv. Chim. Acta, 1977, 60, 2804, 2815, 2824 (see also Chap. 3, Ref. 10), 12 €, Elschenbroich and F. Stohler, Angew. Chem., Int. Ed., 1975, 14, 174, For other references to PPh3 acting as a x-arene see D. J, Cole-Hamilton and G. Wilkinson, J.C.S. Dalton, 1976, 1995.CLASSIFICATION OF LIGANDS 149 Ge PR r Fe KR && (4-XLIX) (4-L) Heterocyclic phosphorus compounds!" such as phosphole (CsHsP) can bond either through P!!4 or be 7-bonded as an arene.!!4° The compound 4-L] also acts!'5 as a ligand in, for example, Fe(CO)4L*. Me > CH NO Me (411) It may be noted finally that tertiary phosphine ligands of the type R2PH or R2PCI can often undergo reactions whereby the H or Cl group is attacked.!!¢ Examples are CH.—! RNH2 cis-Mo(CO)4(PPh2Cl)2 —> cis-Mo(CO)4(PPh2NHR)2 MeOH Ni(PCl3)4 ——> Ni[P(OMe)3]4 Macrocyclic Phosphine Complexes.''7 These may have only phosphorus or have mixed donor atoms of phosphorus, nitrogen, sulfur, and so on. Examples are 4-LI1 and 4-LIII. (4-LI) (4-LI) 113 R.E, Atkinson, Rodd’s Chemistry of Carbon Compounds, Vol. IV, Part G, 2nd ed., Elsevier, 1978, 1141 B, W. Abel and C. Towers, J.C.S. Dalton, 1979, 814. 4b A. J. Ashe, HI, and J. C. Colburn, J. Am. Chem. Soc., 1977, 99, 8099. 1S R,W, Light and R. T. Paine, J. Am. Chem. Soc., 1978, 100, 2230; R.G. Montemayor, J. Am. Chem. Soc., 1978, 100, 2231. 116 G.M. Gray and C. $, Kraihanzel, J. Organomet. Chem., 1978, 146, 23; J. von Seyerl et al., Angew. Chem., Int. Ed., 1977, 16, 858. 17 RE. Davis et al., J. Am. Chem. Soc., 1978, 100, 3642; T. A. Delbonno and W. Rosen, J. Am. Chem, Soc., 1977, 99, 8051, 8053, J. de ©. Cabral et al., Inorg. Chim. Acta, 1977, 25, L77; M. M. Taqui Khan and A. E. Martell, Inorg. Chem., 1975, 14, 676.150 INTRODUCTORY TOPICS 4-21. Phosphorus and Phosphorus Oxides There are a few compounds in which P2, Ps, or Ps molecules act as ligands. Some rather unstable compounds of P, such as RhCl(PPh3)2(P,) and [Fe(CO)4]3P, are known, but the best characterized! 8 are those of a tripod ligand MeC(CH2PPh2)s (L) that contain the cyclotriphosphorus (5-P3) group either as an end group or as a bridge; the P atoms can act as ligands and bind to Cr(CO)s. L 2 Co at Lig A NiL & YW Some phosphido species containing only P atoms are noted later—these are con- sidered to be derived from P3~. Since phosphorus oxides (Chapter 14) have lone pairs, these can also act as ligands, and compounds such as Fe(CO)s(PsO¢), W(CO)s(P4O4) and Cr- (CO)s(P4O7) are known."!9 4-22. Phosphido-Bridged Species!° Phosphorus can act as a bridge between metal atoms in the form of P, PR and PR2 (where R = H, F, Cl, CH3, CoHs, etc.). The most common type of bridge is u-PR2. Phosphido-bridged species can be made directly by interaction of halide complexes with LiPR, or from sodium salts such as NaCo(CO), with PPhCl and in other ways. They are often obtained by reactions involving tertiary phosphine ligands in which a P—C bond of the ligand is cleaved. Some examples of such reactions are: Ph; PhP. P. Ph, 2PUPPhd. — Dal Sec" + Ph—Ph PhP P PPh, Ph, Ph, P. va 2HINCOMPPh,), — (Ph, PCO, Sate + 2CHe P Ph, 118 L., Sacconi et al., Inorg. Chem., 1978, 17, 3292; Angew. Chem., Int, Ed., 1979, 18, 72, 469; J. Am. Chem, Soc., 1978, 100, 2550; 1979, 101, 1757. 119M. L. Walker and J. L. Mills, Inorg. Chem., 1977, 16, 3033; J. Organomet. Chem., 1976, 120, 355. 120 C.M. Bartish and C. S. Kraihanzel, Inorg. Chem., 1978, 17, 735; J.C. Burt et al., J.C.S. Dalton, 1978, 1385, 1387; E. Keller and H. Vahrenkamp, Chem. Ber., 1977, 110, 430; J. K. Burdett, J.C.S. Dalton, 1977, 423; W. Clegg, Inorg. Chem., 1976, 15, 1609; W. Malisch et al., Angew. Chem., Int. Ed., 1977, 408; 1976, 769; J. Reed et al., Inorg. Chem., 1973, 12, 2949; N. J. Taylor, et al., ICS. Chem. Comm., 1975, 448; E. A. V. Ebsworth, et al., JC.S. Dalton, 1978, 272; D. R. Fahey and J, E, Mahan, J. Am. Chem. Soc., 1976, 98, 4491; G, W. Bushnell ef al., J.C.S. Chem, Comm., 1977, 709.CLASSIFICATION OF LIGANDS 151 Ph, P. “ Ru(NOXPMePh,) + 2CH, P Ph, a 2HRu(NOXPPh,CH,)y —> (Ph,MeP)(NO)Ru: These elimination reactions are considered to involve intramolecular oxidative addition of the ligand (Section 29-4) followed by reductive elimination, namely, Ph oS » iy yi "N vn a i, mi M"—PPh “oa? Ls Le or intermolecular transfers of the type Ph, 2\ Ph. mM” CH, _, Le a Ww M—H Although P—C bond cleavage also occurs in reactions of t-butyl phosphine with RhCls, for example, the product is a complex of BuSPH and is not a bridged species.'2! An example of a 44-PPh bridged compound !?? is 4-LIV. (4-LIV) (4-LV) There are finally a few compounds!?3* that have s-P as in the cubane-type complex [(9-CsHs)CoP], (4-LV), and encapsulated species such as [Cog(CO); 4(4-CO)2P]~ and [Rhs(CO)21P]?~ are known, !23% 121 R,G, Goel et al., J. Am. Chem. Soc., 1978, 100, 3629. 122° LF. Dahleral., J, Am. Chem, Soc., 1975, 97, 6904; 1976, 98, 5046, For 43-PH, see R. G. Austin and G. Urry, Inorg. Chem., 1977, 16, 3359. 1239 G. L. Simon and L. F. Dahl, J. Am. Chem. Soc., 1973, 98, 7175. 123 Pp, Chini et al., JCS, Chem. Comm., 1979, 188; J. L. Vidal er al., Inorg. Chem. 1979, 18, 127.152 INTRODUCTORY TOPICS OXYGEN 4-23. Water; Hydroxide and Oxide Ions In aqueous solution metal ions are surrounded by water molecules; in some cases, such as the alkali ions, they are weakly bound, whereas in others, such as [Cr(H20)¢] 3* or [Rh(H2O)g]3*, they may be firmly bound and exchange with solvent water molecules only very slowly (Chapter 28). The bound water molecules may be acidic, giving rise to hydroxo species,'?4 such as (M(H20),]"* = [M(H20).—1(OH)}@")* + H* The acidities of aqua ions can vary by orders of magnitude, e.g., (Pt(NH3)<(H20)2]4* = [PUNH3)a(H20)(OH)}* + Ht K ~ 10"? [Co(NH3)s(H20)]?+ = [Co(NH3)s(OH)}* + H* K~ 10-57 Some molten hydrates (e.g., ZnCl2-4H20) can act as extremely strong acids, even though the aqua ion [Zn(H2O)4]?* in aqueous solution is a very weak acid.!25 A common feature of hydroxo complexes is the formation of hydroxo bridges!?6 of the following types: H, H 4 "S 0. H WS ae Sa Mage wu] Nu 4 H H Double bridges are most common. Three p2-bridges are found in the m-arene complex [ArRu(OH);RuAr]*, whereas a triply bridging, 3-hydroxo group occurs in the cubanelike ions [(ArRuOH).4]*+ (4-LVI), [Pt(OH)Mes]4, and some others.!?7 Ng iS AsRu — "BE , IY | Ark (4-LVI) 124 C.F Baes and R. E. Mesmer, The Hydrolysis of Cations, Wiley-Interscience, 1976; V. Baran, Coord. Chem. Rev.. 1971, 6, 65. 1254. A. Duffy and M. D. Ingram, Inorg. Chem., 1978, 17, 2798 126 D. J, Hodgson, Prog. Inorg. Chem., 1975, 19, 173 (complexes of first-row transition ele- ments). 127 D, R. Robertson and T. A. Stephenson, J. Organomet. Chem., 1976, 116, C29; R. O. Gould et al., JCS. Chem. Comm., 1977, 222; 8. Merlino et al., Inorg. Chim. Acta, 1978, 27, 233.CLASSIFICATION OF LIGANDS 153 The loss of a second proton from water can lead to the formation of oxo com- pounds, !28 which can be of several types: Oo, 0. as AN Z™ M M M od M. MN 7 Es VA M—O—M Symmetric Asymmetric M oO MUM | ‘No On, An, M—0 o=M=0 ou” M4 YM M4 YM \ 0 M M M 0 on I O=M M=0 0=M—O—M=0 eeu 0 The double bridged species are almost invariably symmetrical, but in an osmium compound there is an asymmetric bridge!2 with distances 1.78 and 2.22 A. The multiply bonded oxo group M=O, is found not only in oxo compounds and oxo anions of non-transition elements such as O—SCL2, SO}, O=PCls, and POF, but also in transition metal compounds such as vanadyl (O=V?*), uranyl (O=U=0"+), permanganate (MnOj), and osmium tetraoxide (OsO,). In all these cases the bond distances (ca. 1.59-1.66 A) correspond to a double bond, and the M=0 infrared stretching frequencies usually lie in the 800 to 1000 cm! region for transition metal species. Protonation by acids will convert M==O to M—OH. In transition metal compounds, the component is best regarded as arising from Opa > Madr electron flow. Since this is the opposite of electron flow in 7-bonding ligands of the CO type, it is not surprising that the latter are most stable in low oxidation states whereas M=O bonds are most likely in high oxidation states. The M=O bonding is commonly affected by the nature of groups trans to oxy- gen—and oxygen has a strong trans effect (Section 28-7). Donors that increase electron density on the metal tend to reduce its acceptor properties, thus lowering the M—O multiple bond character, hence the M—O stretching frequency. Because of the strong trans effect, ligands trans to oxygen may be labile. Dioxo compounds may be linear (trans) as in O=U=O?* or angular (cis) as in some molybdenum complexes and in ReO2Mes. Singly bridged complexes may have either bent or linear bridges. The M—O—M angle can vary from ca. 140° to 180° and toa large extent the angle seems to be determined by the steric requirements of the other ligands attached to the metal. There are few bent bridges except for those in Cr;03~, W207;, and P)03”. Ex- amples are!30: 128 W. P. Griffith, Coord. Chem. Rev., 1970, 5, 459 (M—O, M=0); K.S. Murray, Coord. Chem. Rev., 1974, 12, 1 (u-0x0 complexes of Fel"), 129 B.A. Cartwright et af., J.C.S. Chem. Comm., 1978, 853 130 pT. Cheng and S. C. Nyburg, /norg. Chem., 1975, 14, 327; D. W. Phelps et al., Inorg. Chem., 1975, 14, 2486.154 INTRODUCTORY TOPICS (PPh C1.NO Ir ravocrrh.) eee (bipy {NO,)Ru Ru( NO, )bipy }** Linear M——O—M groups are found in some complexes of rhenium, iron, ruthenium, and osmium.'?! In the ruthenium and osmium ions, [M2OX,9]*~, the M—O—M unit forms an electronically unique independent chromophore. The linearity results from d7-p7 bonding through overlap of the p, and py orbitals on O with d,, and dy. orbitals on the metal atoms. Linear M—O—M groups have infrared vibrations lower than those in bent bridges, ca. 260 cm™! versus ca 570 cm ', In linear oxo species such as O=Re—O—Re=O again 7 bonding is im- portant. Some silicon compounds (e.g., Ph3Si—O—SiPhy), also have linear Si— O—Si bonds."3? The pyramidal M30 unit with y3-O occurs in the ion OHgs, in [W302- (O2CR)o(H20)3]?* ions,'?3* in OsyOa(CO)19, and in the [Re30(H)3(CO)o]2- anion,!33> Oxo-centered complexes can have w14-tetrahedral oxygen in the center of a tet- rahedron of divalent metal atoms as in MgO(O2CMe)a.'34 The best-known complex is BesO(O2CMe)a, but Zn"! and Co! analogues are known. An iron(II) complex, Fe}!'0(O2CMe)1o, also has 444-O, and higher oxo iron polymers may have trigonal bipyramidal coordination, as in [FesO(O2CMe),2]*.!35 Oxygen-centered triangles are found in the so-called basic carboxylates of tri- valent metals such as Cr3+, Mn3+, and Fe?+. They have the general formula [M30(CO2R)6L3]*, where L is a ligand such as H2O or py. The structures 4-LVII R | o~0- 070 ace Oo (4LVID) indicate that the MO group is planar on account of M—O = bonding. However in the pivalate [Fes;O(CO2Mes)(MeOH)s]* the O atom is 0.24 A out of plane,!3° ‘31 CC. Ouetal., J. Am. Chem, Soc., 1978, 100, 4717; J. San Fillipo, Jr., et al., Inorg. Chem., 1976, 15, 269; 1977, 16, 1016; R. J. H. Clark er al., J. Am. Chem. Soc., 1977, 99, 2473. 132 C. Glidewell and D. C. Liles, J.C.S. Chem. Comm., 1977, 632. 1338 A. Bino et al., Inorg. Chem., 1978, 17, 3245. 1336 G. Ciani et al., J.C.S. Dalton, 1977, 1667. 134 J. Charalambous et al., Inorg. Chim. Acta, 1975, 14, 53. "35" J, Catterick ef al., JCS. Dalton, 1977, 1420. 136. B. Blake and L. R. Frazer, J.C.S. Dalton, 1975, 193.CLASSIFICATION OF LIGANDS 155 and 13, out-of-plane oxygen atoms also occur in some molybdenum compounds that have a triangle of Mo atoms.'37 For ruthenium, the O-centered complexes can be reduced!38: [Ru30(CO2R)spys]* 2 [RusO(CO2R)epys] SErw(COmR)e In reduced species like Rus;O(CO2R)epys there are nonintegral oxidation states (2U11 + II = 234) and the metal atoms are equivalent. Reduced forms of manganese and iron containing formally M'!!""" are known—for example, Mn30(02C- Me)epys.'3° Although most oxo-centered species have carboxylate bridges, some examples with SO} and NO} bridges are known, as in the ion!° [Pts3O(NO2)6]?-. Finally amino acids can form oxo-centered species such as [Fe3O(ala- nine)o(H2O)3]7*, and this type of structure is a possible candidate for the coordi- nation of iron in the serum iron transport protein ferritin. '4! 4-24, Dioxygen, Superoxo, and Peroxo Ligands!4? Molecular oxygen can be reduced by two one-electron processes without the O-O bond being broken te te O20; OF and each of these species can act as a ligand toward transition metals. Molecular oxygen reacts reversibly with some metal complexes and such reac- tions play a key role in life processes, for example, in oxygenation of hemoglobin and myoglobin!43 (Chapter 31). Compounds with O2 groups can be obtained as follows: 1. From QO by reversible addition reactions with such coordinatively unsatu- rated complexes (Chapter 29) as trans-\t!'Cl(CO)(PPhs)3 + O2 = 1x!"(0,)CI(CO)(PPh)2 Coll(acacen) + O2 + MezNCHO = Co!#(O2)(acacen)(Me2NCHO) Note that in these two reactions the number of electrons formally transferred from 137A. Bino et al., J. Am. Chem. Soc., 1978, 100, $252. 138 TJ, Meyer et al., Inorg. Chem, 1978, 17, 3342; J. Am. Chem. Soc., 1979, 101, 2916. A.R.E, Baikie et al., J.C.S. Chem. Comm., 1978, 62. 140 A,B, Underhill and D. M. Watkins, J.C.S. Dalton, 1977, 5 441 E, M, Holt er al,, J. Amt. Chem. Soc., 1974, 96, 2621 M42 A. P.B, Lever and H. B. Gray, Acc. Chem. Res., 1978, 11, 348; J. E. Lyons, in Aspects of Ho- mogeneous Catalysis, Vol. 3, R. Ugo, Ed. D. Reidel, 1977; R. W. Erskine and B. O. Field, Struct. Bonding, 1976, 28, 3; L. Vaska, Acc. Chem, Res., 1976, 9, 175; G. Henrici-Olivé and S. Olivé, ‘Angew. Chem., Int, Ed,, 1974, 13, 29; F. Basolo, J. A. Ibers, and B. M. Hoffman, Chem. Rev., 1979, 79, 13%; J. Chem. Educ., 1979, 56, 157; Acc. Chem. Res., 1976, 9, 1384; G. McLendon and A.E. Martell, Coord. Chem. Rev., 1976, 19, 1; V. J. Choy and C. J. O'Connor, Coord. Chem. Rev., 1972-1973, 9, 145; J. Valentine, Chem. Rev., 1973, 73, 235; J. A. Connor and E. A. V. Ebsworth, Ado. Inorg. Chem. Radiochem., 1964, 6, 280 (peroxo) O. Hayaishi, Molecular Mechanisms of Oxygen Activation, Academic Press, 1974 (oxyge- nases), 19 3156 INTRODUCTORY TOPICS metal to ligand is 2 for the iridium complex and 1 for the cobalt complexes, corre- sponding to reduction of O2 to 03” and O3, respectively. 2. From Oy by irreversible reactions involving oxidation of the metal, notably cobalt complexes (Section 21-F-6), which form bridged complexes Cot + NH3 + O2 + [amsCo!!02Co!ams]5+ Some of the amine complexes may have hydroxo ‘ridges in addition to M— O—O—M bridges.'*4 Note, of course that metal complexes are often irreversibly oxidized by O2 without forming O2 complexes—in some cases the metal is oxidized; for example, iron(I1)tetrasulfophthalocyanin is oxidized to the iron(III) complex.'45 In others, the ligand may be oxidized and, for example, hydrogen atoms removed as in the conversion of ethylenediamine to imine complexes discussed earlier (p. 118). Destructive oxidation of organometallic and other complexes is a common feature. In most if not all these reactions peroxo and hydroperoxo intermediates are involved. 3. From hydrogen peroxide and metal aqua or other complex ions in aqueous solutions, ¢.g., HCrOj + 2H202 + H* —* CrO(O2)2(H20) + 2H20 Peroxo species of titanium, niobium, tantalum, chromium, molybdenum, and tungsten have been well studied. Sometimes the same complex can be obtained either from O2 or from HOo, &.g., or H202 [Coldiars2]* > (02Co!diarsa]* aa cis-[Co!(H20)adiars2]?+ oH Dioxygen complexes can broadly be classed in two groups: 1. Those containing peroxo (037) groups that may be (a) part of a three- membered ring, (b) bridging staggered, or (c) bridging symmetrical ° A 7O M M 1 I \ ~ M + M o (@) ) © There is only one example of a symmetrical bridge, in the uranyl compound [C1302U-4(0,)-UO,Cl3]*- obtained by action of O2 on uranyl sulfate (UO2SO,) in methanol.'46 In peroxo compounds the O—O bond distances are. fairly constant in the range 1.40-1.50 A (O3- = 1.49 A) and do not depend on the nature of the metal and its ligands. The O—O stretching frequencies are in the 790 to 930 cm~' region; for 144M. Zehnder and S. Fallab, Helv. Chim. Acta, 1975, 58, 13;G. A. Lawrence et al., Inorg. Chem., 1978, 17, 3318. 143 G. McLendon and A. E. Martell, Inorg. Chem., 1977, 16, 1812. 146 R, Haegele and J. C. A. Boeyans, J. C.S. Dalton, 1977, 648.CLASSIFICATION OF LIGANDS 157 the triangular species these frequencies are around 850 cm—!. It makes no difference whether compounds are made from O2 or H2O2, and there is no correlation between reversible oxidation by O2 and any bond parameters. Although the bonding in the three-membered ring is most easily described by localized bonding, it can also be described by an MO treament!” similar to that for the bonding of olefins or acetylenes (p. 95). Crudely, a « bond is formed by filled Op — Mdo bonding and back-bonding is due to Md — Oz. Represen- tative examples of three-membered ring compounds are oxygen adducts of planar d® metal complexes such as trans-IrCl(CO)(PPh3)2 (4-LVII) and chromium peroxo complexes such as the dodecahedral [Cr(O2)4}3~ ion 4-LIX. o—o 7 r—co cd Vopn, (4-LVII) (4-LIX) Ph,P- 2. The superoxo 03 ion'4® can be bound in the following ways: O—O, M—O, M Me Nu No ‘oO The bridged species are mostly those of cobalt(III) or rhodium(III) formed by oxidation of peroxo complexes, ¢.g., z 0». amc Sonn | amor Seam ‘NH,’ NH,’ The bridged superoxo complexes have O—O distances in the range 1.10-1.30 A, e.g., 1.24 A in [(CN)sCo(O2)Co(CN)s]*-,"4 which can be compared to that in O; (1.33 A). The O—O stretching frequencies lie in the 1075 to 1195 cm™! region (Oz, 1145 cm"). The unpaired electron is delocalized over the metal atoms, ac- cording to electron paramagnetic resonance (epr) studies and lies in an MO of 3 symmetry relative to the planar MO2M or MO> group. 147 8 Sakaki er al., Inorg. Chem., 1978, 17, 3183; Y. Ellinger et al., Inorg. Chem., 1978, 17, 2024; J.G, Norman, Jr., Inorg. Chem., 1977, 16, 1328. \48 A.M, Michelson, J. M. McCord, and I. Fridovich, Eds., Superoxide and Superoxide Dismutases, ‘Academic Press, 1977. '49 G. McLendon et al., Inorg. Chem., 1977, 16, 1551158 INTRODUCTORY TOPICS End-on, unidentate, bent superoxo groups are found mainly in complexes of rhodium(III)!°° or cobalt(II1),'5! such as Co(O2)acacen (DMF) noted above, or [Co(O2)(CN)s]*-. The oxygen adducts of complexes of tetrapyrrole and other macrocyclic N ligands have been intensively studied because of the relation to natural oxygen transport molecules containing iron and copper. Considerable ingenuity has gone into trying to make truly reversible synthetic models for heme.!5? The problem is to prevent irreversible oxidation of the iron atom in the macrocycle from Fel! to Fe!!l. One approach has been to construct what are termed “picket-fence” 1534 or basket handle!53> porphyrins. Here, the way whereby the oxygen molecule can approach and leave the iron atom axially is sterically restricted by bulky groups (Fig. 4-3). What appears to be a much more realistic model for reversible binding of molecular oxygen has a heme group attached to a non-cross-linked polymer that confers water solubility (Fig. 4-4). Here the O coordination site is shielded by both histidine and the polymer.'54 Heme systems are discussed further in Chapter 31. 4-25. Alcohols, Alcoxides,'*5 and Phenoxides In solution in alcohols, particularly methanol, metal ions may be solvated just as in water, but the solvent molecules are usually readily displaced by stronger donor ligands such as water itself. Just as coordinated water'can lose a proton to give hydroxo complexes, so can alcohols: In the deprotonated form RO, all hydroxo compounds can act as ligands, for example, Schiff bases derived from hydroxo compounds, hydroxo acids, and so on. Compounds derived from simple alcohols are called alcoxides. They are normally made by reactions of metal halides and alcohols in the presence of a hydrogen halide 150 R, D. Gillard et al., J.C.S. Chem. Comm., 1977, 58. 1S A, Dedieu et al., J. Am. Chem. Soc., 1976, 98, 5789; BK. Teo and W.-K. Li, Inorg. Chem., 1976, 15, 2005; M. Corrigan et al., JCS. Dalton, 1977, 1478; G. B. Jameson et al., J.C.S. Dalton, 1978, 191 182 J. W. Buehler, Angew. Chem., Int. Ed., 1978, 17, 407; R. W. Erskine and B. O. Field, Struct. Bonding, 1976, 28, 3; J.-H. Fuhrhop, Angew. Chem., Int, Ed., 1976, 15, 648; J. P. Collman, Acc. Chem. Res., 1977, 10, 265; F. Basolo, J. A. Ibers, and B, M. Hoffman, Ace. Chem. Res., 1976, 9, 384; W. M. Reiff er al., Inorg. Chim. Acta, 1977, 25, 91. 1349 J.P, Collman et al., J. Am. Chem, Soc., 1978, 100, 2761; G. B. Jameson et al., Inorg. Chem., 1978, 17, 850, 858. 1536 M. Momenteau ef al., Nouv. J. Chim., 1979, 3, 77. 's4 B, Bayer and G. Holzbach, Angew, Chim. Int. Ed., 1977, 16, 117. 135. C. Bradley et al., Metal Alkoxides, Academic Press, 1978; Adv. Inorg. Chem. Radiochem., 1972, 15, 259; Coord. Chem. Rev., 1967, 2, 229; Prog. Inorg. Chem., 1960, 2, 203.b n=l Fig, 4-3, Picket-fence porphyrins having an appended imidazole base. [Reproduced by permission from J.P. Collman, et al., Accounts Chem, Res., 1977, 10, 265.} 159160 INTRODUCTORY TOPICS Fig. 4-4. Water soluble heme complex that binds O2 reversibly. [Reproduced by permission from ref. 154] acceptor, €.8., TiCly + 4CgHsOH + 4EtsN = Ti(OC2Hs)4 + 4EtNH*Cl Although some alcoxides, particularly those with very bulky groups like tert-butyl can be monomeric, most alcoxides are polymeric with alcoxo bridge groups as in titanium tetraalcoxides (Fig. 4-5). Not only doubly bridging 4#2-RO groups are known, but also triply bridging y13- ones as in SngO4(OMe),, which has an ada- mantane SngOq skeleton.!5¢ The polymerization is usually such as to attain max- imum metal coordination. Alcoxides are readily hydrolyzed but are usually ther- mally stable, distillable liquids or volatile solids. Meti-O% FON LM MesMg—O% FON eMe, \ NI A / ea Ne va / Roy 9 RO OR RO OR RO OR a Me Me Me, (4:LX) (4LXD 136 p, G, Harrison ef al., JC.S. Chem. Comm., 1978, 112.CLASSIFICATION OF LIGANDS 161 Fig. 4-5. The tetrameric structure of crystalline [Ti(OC3Hs)4]s. Only Ti and O atoms are shown. Not only can the alcoxo groups act as bridges to the same metal in the polymers, they may act as donors to other metal compounds. Thus U(OPr‘) gives adducts with lithium, magnesium, and aluminum alkyls!5? as in 4-LX and 4-LXI. Aromatic hydroxo compounds such as phenol readily form complexes that may have unidentate groups as in W(OPh)g, or have phenoxo bridges as in (PhO)- ClTi-“-(OPh)2TiCl.(OPh). A few cases are known where the phenoxide ion is 7-bonded to the metal!58 as in 4-LXII. The C—O group becomes more like a keto group, and the bonding is delocalized and is similar to that in -1-5-cyclohexadienyls (Chapter 27). Thus the compounds are best regarded as 7-1-5-oxocyclohexadienyls rather than ar- enes, Ph [ cn 7 a CH)]3N, the structure of whose complex with Rb* is shown in Fig. 4-6. See also page 265 for further discussion. Crown ethers have particularly large complexity constants for alkali metals— equilibrium constants for cyclohexyl-crown-6, for example, are in the order Kt > Rbt > Cst > Nat > Lit. The cryptates also have high complexing ability espe- 13 R.M, Izatt and J. J. Christensen, Eds., Synthetic Multidentate Macrocyclic Compounds, Aca- demic Press, 1978 (a major reference, mainly on crown ethers and cryptates). C. J, Pedersen and K.K. Frensdorff, Angew. Chem., Int. Ed., 1972, M1, 16; D. J. Cram and J. M. Cram, Ace. Chem. Res., 1978, 11, 8; R. M. Izatt and J. J. Christensen, Eds., Progress in Macrocyclic Chemistry, Vol. 1, Wiley, 1979; G. W. Gokel and H. G. Hurst, Synthesis, 1976, 168 (18-crown-6). 164 JM. Lehn, Ace. Chem. Res. 1978, 11, 49; Coord. Chem., 17, IUPAC, Pergamon Press, 1977.