Advances in ORGANOMETALLIC CHEMISTRY
VOLUME 14 CONTRIBUTORS TO THIS VOLUME
V. G. Albano Kenneth P. Callahan P. Chini Ernst Otto Fischer M. Frederick Hawthorne James A. lbers Steven D. lttel M. F. Lappert P. W. Lednor G. Longoni Yoshio Matsumura Akira Nakamura Rokuro Okawara Sei Otsuka V. S. Petrosyan 0. A. Reutov Hubert Schmidbaur Dietmar Seyferth N. S. Yashina Advances in Organometallic Ch emis tr y
EDITED BY
F. G. A. STONE ROBERT WEST
DEPARTMENT OF INORGANIC CHEMISTRY DEPARTMENT OF CHEMISTRY THE UNIVERSITY UNIVERSITY OF WISCONSIN BRISTOL, ENGLAND MADISON, WISCONSIN
VOLUME 14 @ 1976
ACADEMIC PRESS New York * San Francisco * London A Subsidiary of Harcourt Brace Jovanovich, Publishers COPYRIGHT0 1976, BY ACADEMICPRESS, INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS. ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.
ACADEMIC PRESS, INC. 111 Fifth Avenue, New York. New York 10003
United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road. London NWI
LIBRARYOF CONGRESSCATALOG CARD NUMBER: 64-16030
ISBN 0-12-031114-3
PRINTED IN THE UNITED STATES OF AMERICA Contents
LIST OF CONTRIBUTORS . . ix PREFACE xi
On the Way to Carbene and Carbyne Complexes ERNST OTTO FISCHER
I. Introduction . .1 11. Transition Metal-Carbene Complexes . . .3 111. Other Syntheses of Cnrbene Complexes . .6 IV. Reaction Possibilities of Carbene Complexes . .8 V. Transition Metal-Carbyne Complexes . . 21 VI . Reaction of Other Peptacarbonylcarbene Complexes with Boron Trihalides . . 24 VII. Reaction of Pentacarbonylcarbene Complexes with Halides of Aluminum and Gallium . . 27 VIII. Reaction of Lithium Benzoylpentacarbonyltungstate with Triphenyldibromophosphorane . . 27 IX. Reactivity of the Carbyne Ligarid . . 28 References . . 29
Coordination of Unsaturated Molecules to Transition Metals STEVEN D. ITTEL and JAMES A. IBERS
I. Introduction . . 33 11. Theoretical Models . . . 35 111. Structural Results . . 37 IV. Summary . . 59 References . . 60
Methyltin Halides and Their Molecular Complexes V. S. PETROSYAN, N. S. YASHINA, and 0. A. REUTOV
I. Introduction . . 63 11. Methods of Study . . 64 111. Structures of Methyltin Halides . . 68 IV. Molecular Complexes of Methyltin Halides . . . 76 V. Conclusion . . 91 References . . 92
V vi CONTENTS
Chemistry of Carbon-Functional Alkylidynetricobalt Nonacarbonyl Cluster Complexes DIETMAR SEY FERTH I. Introduction : General Properties of Alkylidynetricobalt Nonacarbonyl Complexes . . 98 11. Synthesis of Alkylidynetricobalt Nonacarbonyl Complexes . . 100 111. Chemistry of the Trirobaltcarbon Decacnrbonyl Cation . . 110 Iv. Highly Stable Nonacarbonyl Tricobaltcarbon-Substituted Carbonium Ions . 119 v. Decomposition Reactions and Derived Synthetic Applications of Alkylidynetricobalt Konacarbonyl Complexes . . 135 VI. Concluding Remarks . 138 References . . 141
Ten Years of Metallocarboranes KENNETH P. CALLAHAN and M. FREDERICK HAWTHORNE
I. Introduction . . 145 11. Metallocarboranes: Synthetic Methods , . 150 111. Twelve-Vertex Metallocarboranes . . 155 IV. Thirteen-Vertex Metallocarboranes . . 167 V. Fourteen-Vertex Metallocarboranes . . 171 VI. Eleven-Vertex hletallocarboranes . . 171 VII. Ten-Vertex Metallocarboranes . . 175 VIII. Nine-Vertex Metallocarboranes . . . 178 IX. Oxidative Addition to B-H Bonds . . 180 X. Metallocarboranes in Homogeneous Catalysis . . 182 References . . 183
Recent Advances in Organoantimony Chemistry ROKURO OKAWARA and YOSHIO MATSUMURA I. Introduction . . 187 11. Hexacoordinate Mono- and Diorganoantimony Compounds . . . 188 111. Triorganostibine Sulfide . . 192 IV. Tertiary Stibines . . 197 References . . 202
Pentaalkyls and Alkylidene Trialkyls of the Group V. Elements HUBERT SCHMIDBAUR I. Introduction . . 205 11. Simple Nitrogen Ylides . 207 CONTENTS vii
111. Phosphorus Ylides and Pentaalkylphosphoranes . 209 IV. Arsenic Ylides and Pentaalkylarsoranes . 224 V. Antimony Ylides and Pentaalkylstiboranes . . 23 1 VI. Bismuth Compounds . 236 VII. Related Compounds of Vanadium, Niobium, and Tantalum . . 236 References . 240
Acetylene and Allene Complexes: Their Implication in Homogeneous Catalysis SEI OTSUKA and AKIRA NAKAMURA
I. Introduction . 245 11. Acetylene Complexes. . 246 111. Allene Complexes . 265 References . 279
High Nuclearity Metal Carbonyl Clusters P. CHINI, G. LONGONI, and V. G. ALBANO
I. Introduction . 285 11. Structural Data in the Solid State . 286 111. Structural Data in Solution . 306 IV. Syntheses. . 311 V. Methods of Separation . 316 VI . Reactivity . 317 VII. Iron Derivatives . 323 VIII. Ruthenium Derivatives , 324 IX. Osmium Derivatives . . 325 X. Cobalt Derivatives , 325 XI. Rhodium Derivatives . 327 XII. Iridium Derivatives . . 332 XIII. Nickel Derivatives . 333 XIV. Platinum Derivatives . 334 xv. Bonding Theories . 336 References . 34 1
Free Radicals in Organometallic Chemistry M. F. LAPPERT and P. W. LEDNOR
I. Introduction . 345 11. Metal-Centered Organometallic Radicals . 349 111. Other Organometallic Radicals . . 367 IV. Bimolecular Homolytic Substitution (SH~)at the Metal Center of an Organometallic Substrate . 370 viii CONTENTS
V. Addition or Elimination Radical Reactions , . . 381 VI. Appendix . . 390 References . . 392
SUBJECTINDEX . . 401
CUMULATIVELIST OF CONTRIBUTORS . . 410
CUMULATIVELIST OF TITLES , . 412 fist of Contributors
Numbers in parentheses indicate the pages on which the authors’ contributions begin.
V. G. ALBANO(ass), Istiticto de Chtmicu Generale dell’ Universitd, Milano, Italy KENNETH1’. CALLAHAN(143), Metcalj Research Laboratory, Department of Chemistry, Brown University, Providence, Rhode Island P. CHINI(as,;), Istituto de Chirnica Generale dell’ Universitd, Milano, Italy ERNST OTTO FISCHER(1), Inorganic Chemistry Laboratory, Technical University, Munich, West Germany
M. FREDERICKHAWTHORNE (143), Department of Chemistry, University of California, Los Angeles, Caltjornia JAMESA. IBERS(33), Depnrtntent oj Chemistry, Northwestern University, Evanston, Illinois STEVEND. ITTEL(33), Central Research and Development Department, E. I. du Pont de Nemours and (’onipany, Wilmington, Delaware
M. F. LAPPERT(345), School of Molecular Sciences, University of Sussex, Brighton, England
P. W. LEDNOR*(345), School of Molecular Sciences, University of Sussex, Brighton, England G. LONGONI(ass), Istituto de Chiniica Generale dell’ Universitd, Milano, Italy
YOSHIOMATbUMURAt (187), Department of Applied Chemistry, Osaka University, Yamadakartzi, Suita, Osaka, Japan
AKIRA NAKAliURA (24,5), Department of Chemistry, Faculty of Engineering Science, Osaka Universit!y, Toyonaka, Osaka, Japan
ROKUROOKAWARA (187), Department of Applied Chemistry, Osaka Uni- versity, Yamadakami, Suita, Osaka, Japan
* Present address: Institut fur Anorganische Chemie der Universitat Munchen, 8 Miinchen 2, Meiserstrasse 1, Germany. t Present address: Japan Synthetic Rubber Co., Ltd., Research Laboratory, 7569 Ikuta, Tama, Kawasaki, Japan.
ix X LIST OF CONTRIBUTORS
SEI OTSUKA(245), Department of Chemistry, Faculty of Engineering Science, Osaka University, Toyonaka, Osaka, Japan V. S. PETROSYAN(63), Chemistry Department, M. 8. Lomonosov Moscow State University, MOSCOW,USSR 0. A. REUTOV(63), Cheniistry Department, M.V. Lomonosov Moscow State University, MOSCOW,USSR
HUBERTScmmBAm (205), Anorganisch-chemisches Laboratorium, Tech- nische Universitat Munchen, Munich, West Germany
DIETIIARSEYFEwrn (97), Department of Chemistry, Massachusetts Institute oj Technology, Cambridge, Massachusetts
N. S. YAsnIxA (63), Chemistry Department, M.V. Lomonosov Moscow State University, Moscow , USSR Preface
The first volume of Advances in Organometallic Chemistry was published early in 1964, and twelve other volumes have appeared since that date. The Editors have sought to produce a series of books containing specialist articles on all aspects of this field. The success of the series, as judged by the reviews of the books published in the journal literature, is due in large measure to the cooperation and help we have received from some one hundred and ten contributors. However, the demand for authoritative surveys of topics in organometallic chemistry derives mainly from the continued resilience of this area of endeavor, one measure of which is the annual appearance of over 2000 primary journal articles. After a little over a decade of publication it seemed to the Editors that we should arrange for the appearance of a commemorative Volume con- taining articles by distinguished chemists which would emphasize both the wide scope of organometallic chemistry and its international character. The number of contributors was necessarily limited by the need to keep the book to a reasonable length. This presented a problem in relation to selection of authors. Our choice is, therefore, a personal one guided to some degree by geographical distribution and the desire to balance transition metal chemistry versus main group metal chemistry. F. G. A. STONE It. WEST
xi This Page Intentionally Left Blank Advances in ORGANOMETALLIC CHEMISTRY
VOLUME 14 This Page Intentionally Left Blank On the Way to Carbene and Carbyne Complexes* ERNST OTTO FISCHER
fnorganic Chemistry labardory Technical University Munich, West Germoqy
I. Introduction . 11. Transition Metal-Carbene Complexes . . A. Preparation of the First Carbene Complexes . B. Bonding Concepts and Spectroscopic Findings . . 111. Other Syntheses of Carbene Complexes . IV. Reaction Possibilities of Carbene Complexes . A. Addition and CO Substitution . B. Transition Metal-Carbene Complex Residues aa Amino-Pro- tective Groups for Amino Acids and Peptides . . 11 C. Addition-Rearrangement Reactions. . 13 D. Substitution of Hydrogen at \:- r-Carbon Atom . 13 E. Liberation of the Carbene Ligand . 14 V. Transition Metal-Carbyne Complexes . . 21 A. Preparation of the First Carbyne Complexes . 21 B. X-Ray Structural Analyses . 22 VI. Reaction of Other Pentacarbonylcarbene Complexes with Boron Trihalides . . 24 VII. Reaction of Pentacarbonylcarbene Complexes with Halides of Aluminum and Gallium . 27 VIII. Reaction of Lithium Benzoylpentacarbonyltungstate with Tri- phenyldibromophosphorane . . 27 IX. Reactivity of the Carbyne Ligand . 28 References . 29
I INTRODUCTION
In 1960 I had the honor to lccturt: at this University about our investi- gations in the area of sandwich complexes. Today I do not wish to return to the results of these earlier studies, but instcad I would like to report on an area of work that has occupied us intensively for several years, namely that of carbene and, more recently, carbyne complexes.
* A Nobel lecture translated by P. Legzdins and G. 0. Wiedersatz, Technical Uni- versity, Munich. Copyright @ The Nobel Foundation, 1974. 1 2 ERNST OTTO FISCHER
If one formally replaces one of the hydrogen atoms in an alkane hydro- carbon, such as ethane, by a metal atom (which may also have other ligands attached to it), one obtains an organometallic compound in which the organic entity is bonded to the metal by a r-bond (Fig. la). Such compounds were first prepared more than 100 years ago by Bunsen who obtained cacodyl (tetramethyldiarsine) (1) , as well as by Frankland who prepared various dialkylzincs (2). Later, Grignard succeeded in synthe- sizing alkylmagnesiurn halides by the treatrncnt of magnesium with alkyl halides (3), and for this awompliahment he was awarded thc Nobel Prize in 1912. In addition, thc organoaluminum cornpounds (4) of Ziegler, which made possible the low-prcssurc polymerization of olefins such as ethylene, should be remembered. Professor Zieglcr, together with Natta, were honored with the Nobel Prize in 1963. If one next considers a system with 2 carbon atoms bonded to each other by a double bond, i.e., an alkene molecule, one recognizes a number of separate paths leading to organometallic derivatives. One path involves the replacement of a substituent by a metal atom in a manner that we have just seen and leads to u compounds that are exemplified by vinyl- lithium derivatives. Alternatively, only the a-electrons of the double bond need be employed to bind the organic molecule to the metal. One, thus obtains a-complexes (5, 6) (Fig. lb), the first example of which, Zeise’s salt, Ii[PtC13(C2H4)], was prepared as early as 1827 (7). Such metal r-complexes involving olefins are preferentially formed by transition metals ; the main group elements, on the other hand, are less able to form such bonds. In this class of compound one can also include sandwich complexes (8, 9) in which the bond between the metal and the ligand is no longer formed by just 2 *-electrons but by a delocalized, cyclic ?r-electron system. A particular example is dibenzenechromiuin(0) (10) in which the chro- mium atom lies between two parallel and eclipsed benzene rings. One attains the third variant when one formally cleaves the double bond and attaches one of the resulting halves of the olefin molecule to a
(a) (b) (C)
FIG.1. Production of organometallic compounds from hydrocarbon derivatives, M = a metal or a metal-complex fragment. Carbene and Carbyne Complexes 3 transition metal component (Fig. lb). This concept is realized in the transition metal-carbene coniplcxes in which carbenes ( :CRIX’ that are short-lived in the free stat(, arc‘ stabilized by bonding to the metal. The first part of my account will bt. dcvoted to complexes of this kind. Finally, if one considers molccules with a carbon-carbon triple bond of the type that exists in alkyncs, OIW realizes that there also are three possible paths to metal-cootairiing derivatives (Fig. lc) . One may con- struct a-compounds as discussed in the previous cases or one may use both a-bonds to synthesize. complexes (11) in which the two metal-ligand bonds exist at a right angles to each other. Let us finally consider a cleaved triple bond in which one-half is replaced by a metal complex fragment; thus we come to the carbytic mmplcxes about which I shall report in the second part of this review.
TRANSITION METAL-CARBENE COMPLEXES
A. Preparation of the First Carbene Complexes
In a short communication in 1964, Maasbol and I (12) reported for the first time stable carbene complexes. We had treated hexacarbonyltungsten with phenyl- or methyllithium in ether with the intention of adding the carbanion to a CO ligand at the carbon which, relative to the oxygen, is the more positively charged ligand site. Indeed, we did obtain the lithium acylpentacarbonyltungstates, which were converted to pentacarbonyl[hy- droxy (organo) carbene]tungstc~i(0) complexes by subsequent acidification in aqueous solution (Scheme I). We quickly established that) these complexes are not particularly stable. They tend to split off tho carbene ligand with a simultaneous hydrogen shift thereby liberating aldchydcs, a fact that Japanese investigators also discovered (13). Very recent13 , we learned how to prepare these hydroxy- carbene complexes analytically pure (14). Previously, these complexes, without isolation, had been successfully converted to the substantially more stable methoxycarbene (.ompounds by treatment with diazomethane (12). We soon found a more elegant route to the latter complexes involving the direct alkylation (25) of the lithium acylcarbonylmetalates with tri- alkyloxonium tetrafluoroborates (16) (Scheme 1) which can be prepared 4 ERNST OTTO FISCHER co OC,I co oc~T=co co
according to the method of ilIecrwein et al. This preparative route to the methoxycarbene compounds possesses the advantage of being simple and straightforward and of leading to the desired conipounds in very high yields. There arose, therefore, the possibility of synthesizing a broad spectrum of carbene romplexes. Instead of phcnyllithium many other organolithium reagents (17-23) can be employed. Moreover, hexacarbonyl- chromium (17), hexacarbonylmolybdcnum (17) , the bimetallic deca- carbonyls of manganese (24,25), technetium (25) and rhenium (25), and pentacarbonyliron (28) as well as tetracarbonylnickel (27) may be used instead of hexacarbonyltungsten, but the resulting carbene com- plexes become increasingly more labile in thc indicated order. Finally, substituted metal carbonyls (27-30) can also be subjected to the carbanion addition and subsequent alkylation. The carbene complexes are generally quite stable, diamagnetic, soluble in organic solvents, and sublimable. Before we consider their reactions in more detail, I want to discuss briefly carbene ligand-metal bonding.
B. Bonding Concepts and Spectroscopic Findings The first X-ray crystal structure determination, carried out by Mills in cooperation with us (31) on pentacarbonyl[niethoxy (phenyl)carbenel- chromium (0), confirmed our originally postulated bonding concept. Ac- cording to this concept, the carbene carbon atom is sp2 hybridized. It should therefore possess an empty p-orbital and be electron-deficient. Substantial compensation for this strong electron deficiency is provided by a pr-pa bond between one of the free elcctron pairs on the oxygen atom of the methoxy group and the unused p-orbital of the carbene Carbene and Carbyne Complexes 5 carbon. To a lesser but no less certain extent, there is also a &-pn back- bond from a filled central metal orbital of suitable symmetry to this empty p-orbital of the carbenc carbon. These bonding features affect the distances of the carbene-carbon atom from the oxygen atom on the one side, and from the central chromium atom on the other side: the Ccarbeno-O distance, for which a value of 1.33 A was found, lies between the values for a single (1.43 in diethyl ether) and a double (1.23 A in acetone) bond. Although the average Cr-Cco distance in the carbene complex amounted to 1.87 8, the Cr-Ccarbene distance was 2.04 A. How- ever, according to the considerations of Cotton (32), a distance of 2.21 d would have been expected for a pure chromium-carbon a-bond. Conse- quently, the bond order of the Cr-Ccarbene bond is distinctly less than that of the Cr-Cco bonds in thc same complex, but it is greater than that of a single bond. That the phenyl group, at least in the crystal lattice, does not engage in pr-pn bonding with the carbene carbon can be seen by the significant twisting of the plane containing the Cr, C, and 0 atoms relative to the one formed by the phenyl ring. At the same time, it should be recognized that the double-bond character of the Ccarbene-O bond is so substantial that cis and trans isomers can easily exist relative to this bond (Fig. 2). In the case of pentacarbonyl[methoxy (phenyl)carhenel- chromium(O), only molecules of the trans type are found in the lattice, but at low temperatures lH NMR spectroscopy reveals in solution the coexistence of the cis isomers (33,34). Further important insights into the bonding relationships of the carbene complexes are made possible by a consideration of the vco bands of vi- brational spectra (20, 35-37). As we know, the carbonyl ligands in metal- carbonyl complexes may be considered as very weak donor systems. They donate electron density from the carbon's free electron pair to unused orbitals on the metal atom, a process that formally leads to a negative charge on the metal. This is reduced primarily by a back donation
I 0 0
trans CiS
FIG.2. The structure of pentacarbonyl [methoxy(phenyl)carbene]chromium(0) ; bond lengths in angstroms. 6 ERNST OTTO FISCHER of charge density from the metal to the carbon monoxide via a dr-p backbond. Therefore, carbon monoxide has simultaneously an acceptor function as well as a donor function. The a-donor/*-acceptor ratio of the CO ligand in a complex is a very sensitive probe of the electronic character of the other ligands bonded to the metal. It can be qualitatively estimated by determination of the CO-stretching frequency. Let us coniparc carbon monoxide with methoxy(phcny1)carbene as a ligand in complexcs of the type (CO) &rIl (whcre L = CO or [C (OCH3)- CsHs], respectively) by considering the vco absorptions. While the totally symmetric Itaman-active vco-stretching frequency in Cr (co)6 appears at 2108 cm-I (A,g) (38), we find the absorption of the CO group that is situated trans to the carbcne ligand to be shifted drastically to lower wave numbers and to occur at 1953 cn1-l (A,) (17).This means that the carbene ligand possesses a substantially larger a-donor/a-acceptor ratio than CO. In other words, the entire rarbene ligand is positively polarized with the @r(CO)6 part bcing negative. Thus, the dipole moments of the complcxrs (ca. 5 D) are also relatively large. In the remainder of this account I shall not consider further purely spectroscopic studies. Hoivever, special attention should be briefly given to I3C NMR measuremcnts as they represent extraordinarily valuable aids for the devdopment of this area of chemistry. In the first study of this kind, Kreiter (39) succceded in showing that in pentacarbonyl[methoxy- (phenyl)carbenc]chroniiuni (0) the carbene-carbon atom is very positively charged. The observed chcmical shift of 351.42 ppni lies well within the range of shifts exhibited by the carbo-cations of organic chemistry. Thus, this modern investigative method confirmed once again our original concept. With its intensely positively charged character, the carbene carbon behaves as an elcctrophilic center, a feature of paramount importance in the reactivity of such compounds. Wc shall return to this point.
111 OTHER SYNTHESES OF CARBENE COMPLEXES Since the time that our first paper about metal carbene complexes was published in 1964, this area of research has expanded quickly. Today there are available several major review articles (40-43) dealing with the chemistry of carbene coniplexes. Therefore I want now only to single out particularly interwting syntheses. In our laboratory in 1968, Ofele (44) treatcd l,l-dichloro-2,3-diphenyl- 2-cyclopropene with disodium pentatcarbonylchromate and, with the con- comitant elimination of sodium chloride, he obtained pentacarbonyl(2,3- diphenylcyclopropenyliderie) chromium (0). This compound is stable up Carbene and Carbyne Complexes 7 to 2OOOC and is notable for thci fact that the carbene ligand does not contain any heteroatom. The elrctronic requirements of the cwbene carbon here are satisfied by the three-nicinbtrcd cyclic a-system :
X-Ray structural analysis (45) showed that the three C-C distances of the ligand are not identical: the distance between the 2 carbon atoms bearing the phenyl substituents is shorter than the other two distances. The carbene carbon-chromium distance of 2.05 A lies within the range of values found for our carbene complcxcs, thereby indicating that in this case as well an authentic clarbcne complex exists. A vcry interesting synthetic mcthod was published in 1969 by Richards and co-workers (46). They found that, in the reaction of alcohols with certain isocyanide complexes [such as those of platinum( 11) I,an addition of the alkoxy group to thc (.arbon atom as well as of the hydrogen to the nitrogen atom of the isocyanidc ligarid occurs, and one thus obtains the corresponding carbene cornplcxcs :
c1 P(C,H5)3 + C,H50H ~ pt/P ( cZ H5) 3 ’.t: Cl’ CeN\ Cl’ 1, 7 (2) C-N-C,H, C6H5 II OCZH5 This procedure has led subsequently to many compounds of this kind. The relationship between the complcx chemistry of isocyanides and of carbon monoxide comes to mind at this point. In 1971 we succeeded, again for the first time, in transferring a carbene ligand from one metal atom to another (26,47). For example, if one irradi- ates a solution of cycloprntadic~nyl(carbonyl) [methoxy (phcnyl)carbcnel- nitrosylmolybdenum (0) in the prcscricc of an cxccss of pentacarboiiyliron, one obtains tctracarbonyl[mothoxy (phcnyl)carbencliron(0) with the simultaneous formation of c yclopcntadienyl (dicarbonyl) nitrosylmolyb- denum : a ERNST OTTO FISCHER
Finally, an additional synthesis of recent times (1971) comes from Lappert and his group (48).They treated an electron-rich olefinic system, such as 1, 1‘ ,3,3’-tetraphenyl-2, 2‘-biimidazolidinylidene, with a suitable complex compound. In this manner, they attained cleavage of the C=C double bond and attachment of the carbene fragment to the metal:
CH CH lB l6 (C,H5),P\ /C1\ /C1 Pt Pt /\/\ C1 C1 P(C,H,), N N I I C6H5 C6H5 xylene 140°C (4) CH l6
c1 I C85 This brief sketch summarizes some of the other independently dis- covered methods for synthesis of carbene complexes.
IV REACTION POSSIBILITIES OF CARBENE COMPLEXES
I now wish to confine my attention to carbene complexes of “our type” and to show with recent examples the kinds of reactions we were able to produce. We have already established that the carbene carbon is an electrophilic center and, hence, it should be very easily attacked by nucleophiles. In most reactions we believe that the first reaction step probably involves attachment of a nucleophile to the carbene carbon. In some cases, for instance with several phosphines (49) and tertiary amines (50), such addition products are isolable analytically pure under certain conditions (1 in Fig. 3). For the second step there exists the possibility that the nucleophilic agent may substitute a carbon monoxide in the complex with preservation of the carbenc ligand (2 in Fig. 3). One can also very formally think of the carbene complex as an ester type of system [X=C(R’)OR with X = bf(C0) 6 instead of X = 01, because the oxygen atom as well as the metal atom in the A3 (0)6 residue are each missing 2 electrons for attainment of an inert gas configuration. So, it is not surprising that the Carbene and Carbyne Complexes 9
0
carbene
Liberation of the carbene ligand FIG.3. Reaction possibilities of alkoxycarbene complexes.
OR group can be replaced by amino, thio, or seleno groups (3 in Fig. 3). In this way, the amino- (36, 51-64), thio- (51, 55), and seleno(organ0)- carbene complexes (56) are accessible, but the synthesis of the two latter species requires a special experinirntal skill. We can also observe reactions that lead to a more stable arrangement of the whole system very probably via primary addition and subsequent rearrangement (4in Fig. 3). In addition, it can be established that because of the electron withdrawal of the lf(C0)S moiety, hydrogen atoms in a-alkyl positions to the carbcnc carbon develop such an acidic character that their acidity corresponds to that of nitromethane (5 in Fig. 3). Finally, by cleavage of the carbene ligand from the metal complex, path- ways in synthetic organic chemistry are opened (6 in Fig. 3).
A. Addition and CO Substitution If one treats trialkylphosphinm with pentacarbonyl[alkoxy (organo) - carbelie] complexes of chroiniuni (0) arid tungsten (0), typically in hexane at temperatures below - 30"C, the corresponding phosphorylide complexes (addition compounds) can be isolated analytically pure and studied (49). The formerly carbene carbon is now sp3 hybridized and exhibits only a a-bond to the central metal. In the case of triaryl- and mixed alkylaryl- phosphines, the addition-dissociation cquilibriurn (5'7) (Fig. 4) lies largely on the side of the starting materials, and so the ylide complexes can merely be detected spectroscopically. Figure 4 shows the reaction scheme for pentacarbonylCmethoxy (methyl)carbcne]chromium(O) and tertiary phosphines. Upon irradiation of solutions of these ylide complexes in hexane-toluene mixtures at -15"C, a CO ligaiid is eliminated from the cis position and thereby the cis-tetracarbonylCalkoxy (organo) carbenelphosphine com- plexes are obtained (58). The phosphine that was initially added to the carbene carbon of the starting material thus takes the place of a CO 10 ERNST OTTO FISCHER
<-30°C 7% * (OC),Cr-C-CH, @ 12- hexane I
hexane/toluene I !l cis-(OC),Cr=C I n-hexane, +70"C co \ j 1>+200c
R = alkyl, aryl
FIG.4. Reactions of pentacarbonyl[methoxy(methyl)carbene]chromium(0) with tertiary phosphines.
ligand at the metal atom, at which site the carbene group is re-formed. In addition, but to a lesser extent, the carbcne ligand is also rcplaced by the phosphine. One arrivcs at the samc products if, instead of the isolated ylide complcxcs, one simply employs the equilibrium mixtures of penta- carbonylcarbene complexes arid phosphines under slightly different con- ditions (at -20°C in tetrahydrofurari) (58). By contrast, if one carries out the reaction thermally (at 70°C in hexane) the pure cis-tetracarbonyl- carbenephosphine complexes arc no longer formed but instead mixtures of the cis and trans isomers result (59, 60). We succeeded in isolating both forms in a pure state (60). When solutions of each of these components are warmed, isomerization occurs until an equilibrium is attained (61). We were particularly inter- ested in the mechanism of this process, and we found that the isornerization reaction (6%') follows first-order kinetics, that the reaction speed is not affected by the preseiicc of free ligands such as phosphine or carbon monoxide, and that the isomerization velocity of tetracarbonyl[methoxy- (methyl) carberic]triethylphosphi~iechromium (0) is greater than that of the corresponding tricyclohexylphosphine complex. How do we now visu- alize the progression of this isomerization? The findings speak for an intra- molecular mechanism in which none of the bonds of the six monodentate ligands to the metal is broken nor any new bonds are formed while the transitions from the cis to the trans isomer and vice-versa are occurring: Instead, a rotation of two planes, each containing three ligands, by 120" relative to each other, could take place (twist mechanism) (Fig. 5). Carbene and Carbyne Complexes 11
Cis trans FIG.5. Hypothesis for the isomerization of tetracarbonyl[methoxy (methy1)carbenel- phosphinechromium(0).
Since a trigonal-prismatic transition statc with greater steric hindrance must be traversed, it thus becomes understandable why the compound with the very bulky tricyclohcxylphosphine ligand isomerizes more slowly than does the corresponding trietliylphosphine complex.
B. Transition Metal-Carbene Complex Residues as Amino-Protective Groups for Amino Acids and Peptides If we treat alkoxycarbcric complcxcs not with phosphines but with primary or secondary amincs, we observe a new kind of reaction, remi- niscent of the reactions of csttm. This observation led us into peptide chemistry along a path that proved to be quite surprising to a coordination chemist. We could show that the alkoxy group of alkoxy(organo) carbene complcxcs can be substitutrd not only by mono- or dialkylamino residues but also by free amino groups of amino acid and pcptidc esters (63, 64). The principle of this reaction is shown in Scheme 2.
tB 40 CF,C00@ + H,N-CH-C, + M(CO), + R'CHO + ... I OR^ RZ
M = Cr, W; R' = C,H,, CH,
,O &N-CH-C' = GlyOMe, AlaOMe, ValOMe, PheOMe, SerOMe, MetOMe, & \OR' LeuOMe, Glu(OMe),, TrpOMe, (LysOMe, ProOMe), Leu- LeuOMe Scheme 2 12 ERNST OTTO FISCHER
At 20°C in ether, not only simple but also multifunctional amino acid esters react with alkoxy (organo) carbene complexes partially without pro- tection of the third function. The organometallic residue thus shows itself to be a new and interesting protective group, particularly because it can be very easily removed again by treatment with trifluoroacetic acid. In some cases this removal can be effected under still milder conditions with acetic acid. A further example shows that one may add more amino acids to an amino acid ester-carbene complex by employing the customary methods of peptide chemistry (64).Starting from pentacarbonyl[GlyOMe( pheny1)- carbene]chromium(O) and using the N-hydroxysuccinimide/dicyclohexyl- carbodiimide) (NHS/DCCD) method, we succeeded in synthesizing the sequence 14 to 17 of human proinsulin C-peptide (Scheme 3).
1. dioxane - NaOH 2. HCI 3. ProOMe NHS/DCCD 1. dioxane - -;;Gly- Gly - Pro- GlyOMe NaOH .Gly-Gly-ProOMe (OC)SCr=C + (OC) Cr- C2 I \C6H5 2. HC1 ‘C85 3. GlyOMe NHS/DCCD
Scheme 3 In working together with Wiinsch in this area, we have concluded that the use of such carbene complexes offers a series of possibilities and ad- vantages to the peptide chemist. 1. Amino acid as well as peptide derivatives of this kind are yellow and, hence, can be easily recognized when using methods such as chroma- tography. 2. This new protective group can be removed under mild conditions so that, in addition to the amino acid or peptide esters, respectively, the reaction products that originate (mostly aldehyde and hexacarbonylmetal) are volatile and hence can be easily separated. 3. Most carbene complexes of amino acid esters, as well as of some di- peptide esters, are volatile and, therefore, can be analyzed by mass spec- trometry. 4. This is a method by which heavy metal atoms, such as tungsten, may be introduced into peptides and in this way free amino groups can be labeled. Carbene and Carbyne Complexes 13
C. Addition-Rearrangement Reactions Two recent examples belong to this class of reaction for carbene com- plexes. Pentacarbonyl[mcthylthio (methyl) carbene]chromium(O) and -tungsten(O) react at low temperatures with hydrogen bromide to give pentacarbonyl[ ( 1-bromoethyl) methylsulfide] complexes (65) : ACH, pentane 7, (OC),M-=C” + HBr-(OC),M-S H Tt \CH, (5)
M = Cr, W In this reaction the original carbene carbon loses its bonding mode to the metal and sulfur takes over this function. The second example shows that such a reaction need not always lead to an uncharged system. If, instead of thiocarbene complexes, amino- carbene complexes are treated with hydrogen halides, products of saltlike character are isolable (66). One finds the halogen at the metal and the hydrogen at the removed carbene ligand, and one obtains iminium halo- genopentacarbonylmetalatcs :
Consequently, there is available a synthesis for cations of this kind that permits multiple variations not otherwise easily attainable.
D. Substitution of Hydrogen at the a-Carbon Atom Icreitcr (67) first demonstrated by ‘H NMR spectroscopic studies the acidity of hydrogen atoms that are bonded at the a-C atom of alkoxy- (alkyl)carbene complexes. In CH30D solution, in the presence of catalytic amounts of sodium methanolate, pentacarbonyl[methoxy (methyl)car- benelchromium (0) exchanges for deuterium all hydrogen atoms of the methyl group that is situated next to the carbene carbon: ,OCH, + 3 CH,OD ,&CH, (OC),M-CT L (OC),M--C, + 3CH,OH CH, CH,ONa, cat. @ CD3 (7) M = Cr, Mo, W
R 14 ERNST OTTO FISCHER
Obviously, the base is able to form an anion in this situation by reversible cleavage of a proton from the position a to the carbene carbon. This reaction may also be used to introduce new groups into the carbene ligand at this position (40, 68). By its nature, the a-CH acidity is closely connected with the strong positively charged character of the carbene-carbon atom. Let me, therefore, return once more to the unusual 13C NMR data which are available for this atom in some characteristic chromium(0) complexes (Table I).
TABLE I 13cNUCLEAR MAGNETIC RESONANCE SHIFTS FOR THE CcsrbneATOM IN SOMECHROMIUM(O)-CARBENE COMPLEXES
Complex 6 (PPm)"
0 &Values are relative to internal TMS. * In (CU3)zCO. In c6z)6.
If we start with the methoxy(pheny1)carbene complex, for which 351.4 ppm was measured (39), and we replace the methoxy group by the better- stabilizing dimethylamino group, wc find an expected decrease of the &value to 270.6 ppm (66). On the other hand, a value of 399.4 ppm is found for the diphenylcarbene complex (69). In this compound, with which we have been occupied just recently, there exists, therefore, a carbon atom that, even according to the standards of organic carbo-cation chemistry, is extremely positively charged. Both phenyl groups are, thus, barely able to remedy the electron deficiency at the carbene carbon. This carbene chromium complex is much more labile than the homologous tungsten compound which has been briefly described by Casey (70).
E. liberation of the Carbene ligand
1. Reaction with Acids I suppose that the time is past when one, as a chemist, still was permitted to think in terms of the separate notions of "inorganic" and "organic." Carbene and Carbyne Complexes 15
Nowadays one should rather keep in view all the possibilities offered by nature. Our carbene complexes should open the way to new organic chemistry if the carbene ligand can be successfully cleaved from the metal under not too severe reaction conditions. This expectation can be realized, for in- stance, by utilizing hydrogen halides in methylene chloride at -78°C (71):
,OCH, CH C1 (0c),w-cV + HX- (OC),XWH + ‘‘6H5 -18°C (c’ZrH:} (9) X = I, Br, C1 In this manner are formed pentacarbonylhalogenotungsten hydrides which, to the best of our knowledge, are new, in contrast to the corre- sponding anions. The neutral hydride complexes are very unstable and in aqueous solution they dissociate completely into hydronium and penta- carbonylhalogenotungstate ions. The latter can be precipitated as tetra- methylammonium salts (71):
We obtained from other studies hints concerning the fate of the liberated carbene ligand. Namely, if one treats tetracarbonylCmethoxy (organo) - carbene]triphenylphosphinechromium (0) with benzoic or acetic acid in refluxing ether, one can isolate a-methoxyorgano esters of these acids (72):
ether 2-6 hr reflux OCH, (11) I trans -[(C,H,),PI,Cr(CO), + R’-C-0-C-H I1 I OR R, R’ = CH,, C,H,
This “trapping” reaction formally corresponds to an insertion of the carbene into the OH group of the carboxylic acid. The triphenylphosphine that is added to the reaction mixture merely facilitates separation of the metal-complex fragments as the slightly soluble tetracarbonylbis (triphenyl- phosphine) chromium (0). The reaction with hydrogen chloride proceeds in an analogous manner, but the insertion products (a-halogenoorgano- (methyl)ethers) formed react at once with the phosphine present to form the corresponding phosphonium salts (72). 16 ERNST OTTO FISCHER
In this connection, the question immediately arises: What happens to the liberated carbene ligand when no suitable reaction partner is at its disposal? In this case the selected reaction conditions are of crucial im- portance.
2. Reaction with Pyricline
Already at the beginning of our studies involving carbene complexes, we observed that the carbene ligand can be easily rcleased from the metal with pyridine and that the nictal fragment can bc isolated in the form of carbonylpyridinechromium complexes (73). In the carbene fragment of alkoxy (alkyl)carbene complexes, there ensues, with the cooperation of the base, the shift of a hydrogen atom to the original carbene-carbon atom, and enol ethers are thus formed (73, 74):
,y OCH, ,OCH, (OC),Cr=C\ ,R1 90°C, 1.5 hr C pyridine H' \R2 (12)
R': H H H CH, R2: H CH, C2H, CH,
3. Thermal Decomposition
In order to test whether the base exerts an essential influence on the subsequent reaction of the carbene ligand, we also decomposed penta- carbonylCmethoxy (methyl) carbenelchromium (0) purely thermally at 150°C in Decalin. Under these conditions wc obscrvcd exclusively the formation of carbene dimers-as a mixture of cis and trans isomers (74)- thereby confirming the influence of the base:
;QCH, 150"C, 12 hr H,CO, ,OCH, H,C\ ,OCH, (OC),Cr=c; /c=c, + + ... (13) CH, decalin H3C CH, HsCO/C=C\ CH,
Since the shift of a hydrogen atom is not possible with the methoxy- (phenyl) carbene ligand, it can only dimerize in reactions with bases and in thermal liberations (75). Carbene and Carbyne Complexes 17
4. Reaction with Elements of the Sixth Main Group Xaturally, reactions are esperially interesting to us if the products obtained are not readily accrssiblr by the methods of classic organic chem- istry but are rasily preparable with our complexes. We found one such ex- ampic upon treatment of pmtacarbonyl[methoxy (aryl)carbenelchrom- ium(0) complexes with oxygen, sulfur, or selenium (76). In this way one obtains convenicntly the Corresponding methyl cstcrs of arylcarboxylic acids and the 0-methyl esters of arylthio- as well as arylselcnocarboxylic acids; this serms to us to be syrithrtically useful in the last two cases:
OCH, I 02,hexane 68"C, 1-6 hr "='aR
OCH, OCH,!I I (OC),Cr--r-Cl------S, , ether 34"C, 24 hr s=cmR
?CH3 Se= C Se, , dioxane 101"C, 24 hr - mR R = H, CH,, OCH3, C1
5. Reaction with Vinyl Ethers and iC'-VirLUlpyn.olidones
Already at a vcry early stag(. of the studies on carbene complexes, we considered that these compounds only properly merit their name when they engage in reactions that are also typical for carbenes. In organic chemistry one surely thinks at once of the construction of cyclopropane derivatives from olefiris and carbencs. Indeed, it has been shown that this also is possible with our complcxes and with C=C double bonds that are electron-poor and arc either polarized or easily polarizable (77-81). As an example of this, I would like to cite the reaction of penta- carbonyl[niethoxy (phenyl)carbc1~c]chromium(0), -molybdenum(O) , or -tungsten (0) with ethyl vinyl ether (79). One obtains the corresponding cyclopropane derivatives in this rase, however, only when one removes 18 ERNST OTTO FISCHER the carbene ligand under CO pressure of 170 atm in an autoclave at 50°C:
50°C, 65 hr
(15) H OCH, H + H C,H, H + M(CO), + H,' "OC,H, H,,' "OC,H5 (a) (b) M: Cr Mo W (a) 76 80 64 (b) 24 20 36 As expected, there arise two isomers [(a) and (b) in Eq. 151 whose proportions depend on the choice of the central metal atom under other- wise similar reaction conditions. This seems to us to be a rather important hint that the reaction does not proceed via a "free" methoxy(pheny1)- carbene but that the metal atom participates in the decisive reaction step. When we attempted the analogous reaction with N-vinyl-2-pyrrolidones as the olefin components under similar conditions, we obtained surprisingly l-[4-methoxy-4-phenyl-3-oxo-l-butenyl]-2-pyrrolidones instead of the expected cyclopropane derivatives (82):
I C,H,, 80°C CH + 2co - II f 150 atm CO, CR 60 hr c=oI CHR CH~O-CH I C6H5 R = H, CH, How can one explain the origin of these products in which one finds, aside from the former carbene ligand and the pyrrolidone, just an ad- ditional carbonyl group? It seems plausible that the carbene ligand first of all reacts with carbon monoxide to form methoxy(pheny1) ketene. This, in turn, forms with the polarized olefin a cyclobutanone derivative which by ring opening goes over to the observed product (Fig. 6). Moreover, this hypothesis is supported by the fact that with N-(0- methylvinyl) -2-pyrrolidone the postulated four-membered ring system was isolated, in addition to the open-chained end product (8%'). Corbene and Corbyne Complexes 19
OCH, Q0 + pH5I
I II C HC HC II II 0 RHC RHL :
R = H, CH,
= CH3 J /
00I CH II CR c=oI I CH,O-CH I C,H5 R = CH,, H
FIG.6. Hypothesis concerning the course of the reaction during the treatment of pentacarbonyl[methoxy(phenyl)carbene]chromium(O) with N-vinyl-Zpyrrolidones un- der a CO pressure of 150 atm.
Our original idea to employ carbon monoxide solely for the removal of the carbene ligand led, thcrcforc, to an unexpected result. From this observation, it is clear that the potential of the reaction of carbon monoxide with organic systems must not be ncglectcd. Scvcrtheless, to arrive at the sought-after cyclopropane derivatives, we treated the same react!ants with each other thermally in benzene in the absence of carbon monoxide. In this reaction as well we did not get the desired compounds, but obtained instead, again surprisingly, the corre- sponding substituted a-methoxystyrenes (83) (Scheme 4). A possible reaction course conies to mind. The N-vinyl-2-pyrrolidone also possesses at the oxygen a nucleophilic center which could attack the electrophilic carbene carbon and could release the carbene ligand from the metal. The intermediate product formed-irrespective of whether it is an open chain or a six-membered ring-then undergoes a heterolytic fragmen- tation by splitting similar to that, observed by Grob (84) (Fig. 7). 20 ERNST OTTO FISCHER
0 80°C, 7 hr I C,H,
R‘, ,OCH, R2,c=c, + Cr(CO), + ... Rl&HR2 m C,H, R’ = H, CH,; R2 = H, CH, Scheme 4 6. Reaction with Electrophilic Carbenes
As mentioned earlier, the carbenc ligand in our complexes shows ‘hucleo- philic” charactcr with respect to the metal fragment. Therefore, we decided to combine it with an electrophilic carbenc. For this purpose we treated pentacarbonylCmethoxy (phenyl)carbene]chromium (0) with phenyl (tri- chloromethyl) mercury (85). Compounds of this kind have been studied intensively by Seyferth et al. (86) arid arc known as a source of dihalogeno- carbenes. The carberic complex reacted with the carbenoid compound at
FIG.7. Hypothesis concerning the course of the reaction during the treatment of pentacarbonyl[methoxy(phenyl)carbene]chromium(0) with N-vinyl-2-pyrrolidone and p-substituted N-vinylpyrrolidones under normal pressure. Carbene and Corbyne Complexes 21
80°C in benzene to give @,B-dichloro-a-methoxystyrene (85) : ,;OCH, C,H, H,CO, ,C1 (OC),Cr-C, + C,H,HgCCl, 80"c, H5c,'C=C, + C,H,HgCl + Cr(CO), + ... C,H, c1 24 hr
1:l H,CO, ,Br + Br, ,Br ,c=c, ,C=C, + C,H,HgBr + Cr(CO), + ... H,C, Br Br Br This combination reaction is very sensitive to the temperature conditions. We encountered complications when using phenyl (tribromomethyl) - mercury since mixtures of olefins arose. I hope that I have been able to demonstrate, with this small selection of our newest research results, what a variety of reaction possibilities the chemistry of transition metal-carbenc complexes display. In the following I review an area whose devc.lopmcnt we have made most recently our special task, namely that of transition metal-carbyne complexes.
V TRANSITION METAL-CARBYNE COMPLEXES
A. Preparation of the First Carbyne Complexes In order to fathom the entire range of reactions of transition metal- carbene complexes, we had undertaken years ago experiments to treat our complexes not only with nueleophilic but also with electrophilic reagents. It was our intention to exchangc the methoxy group of methoxy(organ0)- carbene complexes with a halogen with the aid of boron trihalides, and so to arrive at halogeno (organo) carbrne complexes. Indeed, we initially ob- served a quick reaction but found only decomposition products. Just a short time ago, when Kreis attempted this reaction at very low temper- atures, we were able indeed, to isolate definite but rather thermally labile compounds (87).Their composition corresponded to the sum of a tetra- 22 ERNST OTTO FISCHER carbonylmetal fragment, a halogen and the carbene ligand less the methoxy group :
.OCH, pentane
(OC),M--CT R + ” {B%ocH,i (18) (OC),M(X)CR M = Cr,-c Mo, W X = C1, Br, I R = CH,, C,H,, C,H, The IR spectra indicated the presence of disubstituted hexacarbonyl- metal derivatives with two different ligands in trans positions (trans- (CO).&flL11t2). Morcover, the cryoscopic molecular weight determination proved that the complexes must be monomers. Together with further spectroscopic findings, especially from 13C and lH NMIt studies, this could only be interpreted to mean that besides the four CO ligands a halogen atom and a CR group are bonded to the metal (Fig. 8). For this new type of compounds we would like to propose the name “carbyne complexes” for two reasons: (i) in analogy with carbene com- plexes, (ii) because similarly to alkyne, on the basis of the diamagnetism of these compounds, a formal metal-carbon triple bond had also to be postulated.
B. X-Ray Strucfural Analyses The proposed triple bond should have a very short distance between the metal and the carbyne carbon. Only X-ray structural analyses could clarify this question as well as provide definitive confirmation of our structural proposal, and these studies were then carried out in our Institute by Huttner and his co-workers on three carbyne complexes. The first successful X-ray study involved trans- (iodo) tetracarbonyl- (phenylcarbyne) tungsten(0) (87, 88) (Fig. 9). It confirmed in essence our concepts, and yielded an extremely short tungsten-carbon distance of
oc, ,co X--MC-R oc’ ‘co
FIG.8. Structural and bonding concepts for (C0)rXMCR. Carbene and Carbyne Complexes 23
0 0 \/cc 2.845 i 0.005 \ / 1.90 f 0.05 i I c I.40 162*4b\ H, C C-C, ,H ;/" \0 \c 0 H/C\~--\C, 'H H
W-Cco: 2.07 f 0.06 8, [in W(CO),: 2.058 h1 W-Cspz-single bond [in C,H,W(CO),C,H,]: 2.32 i W-Csp-single bond (estimated from rCspz = 0.74 h and ycSp = 0.69i): 2.27
FIG.9. Molecular structure of trans-(iodo)tetracarbonyl(phenylcarbyne)tungsten(O).
1.90 8. Instead of a linear arrangcmcnt of metal, Ccarbyne, and Cl.l(phenyl) atoms, mc found a distinct bend (ca. 162'). Since we cannot as yet explain whether this is attributable to electronic or lattice effects, we undertook the examination of a second complcx. Figure 10 shows the result, namely, the structure of trans- (iodo) trtrararbonyl (methylcarbyne) chromiuni (0) (88). In this compound one finds not only the expected linear arrangement of chromium, carbon, and thc methyl group, but also the shortest chro- mium-carbon distance (1.69 d) known at present. This value is distinctly shorter than that for the Cr-Cco distance in the same complex (1.946 d) or in hexacarbonylchromium (1.91 8). Of subsequent interest to us was the question of whether a third kind of ligand in the starting carberic romplex could influence the orientation of the halogen in the eventual carbyne complex. Therefore, we first treated cis-tetracarbonylCmethoxy (methyl)rarbcneltrimethy lphosphine-, -arsine-,
0 \ Z.18 i 0.Olh cc 2.792 f 0.002 h \ /1*946 * 0.0°9 1.49 i 0.02 8, I Cr C CH3 / \ 1.69 * 0.01 C 0/" \0 Cr-C,pz-single bond: 2.22 h Cr-Csp- single bond: 2.17 h Cr-CCo in Cr(CO), : 1.91 A
FIG.10. Molecular structure of trans-(iodo) tetracarbonyl(methy1carbyne)chro- mium(0). 24 ERNST OTTO FISCHER
0 \ C P(CH,), 2.604 + 0.006i \ 1 1.69 + 0.04 1.49 * 0.05 A Br G CH3 177 i 3" C 0/" \0 Cr-CCO : 1.93 * 0.04 ;i Cr-P: 2.40 * 0.01 FIG.11. Molecular structure of mer-(bromo)tricarbonyl(methylcarbyne)trimethyl- phosphinechromium (0).
and -stibinechromium (0) likewise with boron trihalides (89).
-OCH, pentane cis -(OC),Cr[Y(CH,),]C': + B&- (OC),[Y(CH,),](X)Cr=CCH, + (BqOCH,) + CO CH, ~ 10°C X=CI, Br, I; Y = P, As, Sb (19) The reaction progressed just as smoothly as those previously studied, but among the products of composition (CO) 3[Y (CH3) 3](X) Cr=CCH3 (X = C1, Br, I and Y = P, As, Sb) the relative positions of the ligands could not at first be unequivocally clarified. Therefore, an X-ray structural analysis was also carried out on a representative compound of this type (88) (Fig. 11). For (bromo)tricarbonyl (methylcarbyne) trimethylphos- phinechromium(O), it showed a meridional arrangement of the three substituents and, once again, a trans positioning of halogen and carbyne ligands. How a starting carbene complex possessing a trans configuration behaves in such reactions with boron trihalides is presently still under investigation.
VI REACTION OF OTHER PENTACARBONYLCARBENE COMPLEXES WITH BORON TRIHALIDES
It was also of interest to us how changes in the organic residue of the carbyne ligand influence the stability and the behavior of carbyne com- plexes. Hence, we treated with boron tribromide a series of pentacarbonyl- [methoxy (aryl)carbeneltungsten (0) complexes which were substituted at Carbene and Carbyne Complexes 25 the phenyl residue (90) OCH, (OC),W& f BBr,-- pentanet Br-WOC\ $’--(@-C t {BBr,OCH,} f CO oc’ ‘70 / R R R= p-CH,, p-OCH,, P-CF,, 2,4,6-(CH,), Also here 13C NMR spectroscopy should be suitable as a probe for electronic changes. The chemical shifts of the carbyne-carbon atoms of the trans- (bromo) tetracarbonyl (arylcarbyne) tungsten (0) complexes ob- tained are presented in Table I1 (87, 91). Contrary to expectation, in this series one finds for the pCF8 derivative the lowest &value, i.e., the strongcst screening of the carbyne-carbon atom, whcreas for the 2,4,G-trimcthyl derivative, by comparison, the screening is distinctly weaker. For 311 exact interpretation of these results, it is apparent to us that further experiments (currently in progress) are necessary.
TABLE I1 l3C NUCLEARMAGNETIC RESONANCE SHIFTS FOR THE CoarbyneATOM IN SEVERALtruns-Br(CO),WC-Ar COMPLEXES‘
Complex 6 (PPd
266.15
271.30
ECa CH, 271.43
273.16
EZC CH, 275.13
H,C
a CHzClz: &values are relative to internal TMS. 26 ERNST OTTO FISCHER
We could show further that not only methoxy (organo) carbene complexes react with boron trihalides in the above-mentioned sense. For instance, it was found that trans- (bromo) tetracarbonyl (phenylcarbyne)chromium (0) and -tungsten (0) are also accessible from pentacarbonyl[hydroxy (phenyl) - carbenc]chromium (0) (92) and from pentacarbonyl[methoxycarbonyl- methylamino (phenyl)carberie]tungsten (0) (the glycine methyl ester de- rivative) (93),respectively:
,;NHCH,COOCH, (oc),w--c, + BBr, CbH5 I CH,Cl, - 25°C (22) OC\ ,co t HzO Br - W EC-CC,H, + CO + (BBr,NHCH,COOCH,)- HBr .NH,CH,COOH + . . . oc’ ‘co I would emphasize that the reaction of the amino acid-carbene complex with boron tribromide represents a good possibility of again cleaving the “carbenyl” protective group under extremely mild conditions at - 25OC. That experimental results cannot always be generalized is shown by the treatment of cis- (bromo) tetracarbonyl[hydroxy (methyl) carbenelniangs- nese with boron tribromide. This procedure does not lead to the analogous carbyne complex but rather to a produrt in which the hydrogen atom of the hydroxy group is substituted by a BBrz residue (94):
Br\-/Br
0 n hexane (23) OC-Mn-C‘ + BJ3r3- - OC-Mn---C + HBr ‘CO \ CH, -20°C 06 OC’ \CO ‘CH,
Here the conditions for thc formation of a carbyne complex apparently are not available because of the fixation of the OH group by formation of a bridge to the cis-situated broniinc ligand. An open question was also how pentacarbonylCethoxy (diethylamino)- carbeneltungsten(0) would react with boron trihalides, since, as we have learned previously, in principle both the alkoxy and the amino group are removable. The answcr was given by the exclusive formation of trans- (bromo)tetracarbonyl (dicthylaminocarbyne)tungsten (0) (96), a com- pound that is comparatively easy to handle. Its stability can be attributed to the interaction of the metal-carbon bond with the free electron pair of Carbene and Carbyne Complexes 27 the nitrogen. This interpretation is supported by 13C NMR findings:
OC, ,;OCZH, pentane ,co (oc),w-=c. + BBr,-Br-W EC-N(C,H,), + {BBr,OC,H,) + Co 10°C /\ K~ (c,H,), oc co (also with BI,)
VII REACTION OF PENTACARBONYLCARBENE COMPLEXES WITH HALIDES OF ALUMINUM AND GALLIUM
In the extension of our synthctic methods, aluminum trichloride and tribromide as well as gallium trichloride may also be employed instead of boron trihalides (96) :
Br (oc)5w-c-, ;OCH, -Br-W-C-C,H,Al,Br, OC\ /co + IBr\AI/ + CO C,H, pentme, /\ benzene, OC co Br’ ‘O(H3 ‘Br 30°C
OC, ,co GOCH, MCI, (OC),W==C, Cl-WSEC-C,H, + (MCI,OCH,} + CO C,H, toluene, /\ -30°C oc co M = Al, Ga Also in these cases we obtained cnrbyne complexcs in good yields.
Vlll REACTION OF LITHIUM BENZOYLPENTACARBONYLTUNGSTATE WITH TRIPHENYLDIBROMOPHOSPHORANE
In principle, a new synthetic route resulted from the treatment of lithium benzoylpentacarbonyltungstate with triphenyldibromophosphorane 28 ERNST OTTO FISCHER at low temperature in ether (9’7)
oc co ;OLi + Br,P(C,H,)3~Br-W~C-C,H,ether \/ + LiBr + CO (26) (OC),W--C~ C,H, oc/ ‘co + OP(C,H,), + ...
One may surely suppose that the first step involvcs the creation of a C,,,~,,,-O--P linkage by formation of lithium bromide. The resultant intermediate product could stabilize itself by attack of a second bromine atom at the metal, elimiriation of a CO ligand, and liberation of the thermo- dynamically favored triphenyloxophosphorane, thereby forming the car- byne complex.
IX REACTIVITY OF THE CARBYNE LIGAND
Also with carbyne complexes we do not want to confine ourselves only to preparing and spectroscopically examining new variants of this type of compound, and thus we have already begun to study their reaction be- havior. Initially we looked for a possibility to compare such a metal-carbon triple bond with a carbon-carbon triple bond. In this connection, it seemed appropriate to rcact dimethylamine 11 ith trans- (halogeno) tetracarbonyl- (phenylethyriylcarbync)tungsten( 0) (98) ; the latter compounds is ac- cessible from pentacarbonylCethoxy (phenylcthynyl) carbcneltungsten (0) (21) and boron trihalides. Wc found that, at -40°C in ether, only ad- dition to the “organic” triple bond took place, whereas the carbyne-metal bond remained unchanged (99) (Scheme 5).
OC, ,CO,;;OC*H, -45°C OC\ oc- w-c + BX,- X-WEC-CEC pentane oc’ ‘co a - 40°C X = C1, Br, I ether +(CH,),NH
OC, /CO1 \ X- W -C- C(H)=C @ oc’ ‘co N(CH3), X = Br
Scheme 5 Carbene and Carbyne Complexes 29
Simultaneously, we have been occupied with the question as to how the carbyne ligand behaves when it is split off from the metal. Analogous to the carbene complexes, in the absence of a suitable reaction partner, one also observes dimcrization, in this case to alkynes (100).The conditions for the removal are very mild. In nonpolar solvents, diphenylacetylene or dimethylacetylene are accessible in this way at 30°C. One arrives at the same result when the solid mcthylcarbync complex is licatcd to 50°C :
OC, ,co hexane, Br-CrEC-CC,H, : C,H,-CC_C-CC,H, + ... oc’ ‘co +30”C, 1.5 hr
... CH,-C=C-CH, + .’. L, 3 nr
Thus, the way appears to br open to make carbync complexes for use in the syntheses of organic compounds. Since to our knowledge there is no specific and selective “carbyirc source” available for preparative pur- poses, presumably thcre arises hcw a wide field of interesting possibilities of application, especially because of thc mild conditions required to transfer the carbyne ligand. It is hoped that this report his shown that thcre are still many exciting possibilities open in the field of organometallic chemistry.
ACKNOWLEDGMENTS The chemistry which 1 have had the honor to report is in large part the work of my co-workers. I would like once more to thank very heartily Miss Dip1.-Chem. K. Weiss as well as Dr. K.H. Dotz, Dr. 11. Fischer, Dr. F.R. Kreissl, Dr. S. Itiedmuller, Dr. K. Schmid, Dr. A. de Kenzi, Dip1.-Chem. B. Dorrer, Dip1.-Chein. W. Kalbfus, Dip1.-Chem. 13.-J. Kalder, Dip].-Chem. G. Kreis, 1)ipl.-Chem. E.W. Meineke, Dipl.-Chem. D. Plabst, Dip1.-Chem. K. Richter, Dipl.-Chem. U. Schubert, Dip1.-Chem. A. Schwanzer, Dip].- Chem. T. Selmayr, Dip].-Chem. S. Wale, and Cand. Chem. W. Held for their collabora- tion. At the same time, that sentiment applies also to the colleagues of our Institute and their co-workers: Dr. J. Muller for the evaluation of the mass spectra, Dr. C.G. Kreiter for the 13C NMlt studies, and Dr. G. Iluttner in collaboration with DipLChem. W. Gartzke as well as Dip1.-Chem. H. Lorenz for the X-ray structural analyses.
REFERENCES 1. Bunsen, It., Liebigs. Ann. Cheni. 42, 41 (1842). 2. Frankland, E., Liebigs Ann. Chem. 21, 171, 213 (1849); 95, 36 (1855). 3. Grignard, V., C. R. Acad. Sci. 130, 1322 (1900). 30 ERNST OTTO FISCHER
4. Ziegler, K., Angew. Chem. 76, 545 (1964). 5. Fischer, E. O., and Werner, H., “hletnl-x-Complexes,” Vol. I: Complexes with IX- and Oligo-olefinic Ligands. Elsevier, Amsterdam, 1966. 6. Herberhold, M., “Metal-dhmplexes,” Vol. I1 : Complexes witah Monoolefinic Ligands. Elsevier, Amsterdam, 1972. 7. Zeise, W. C., Poggendorfs Arm. 9, 632 (1827). 8. (a) Fischer, 2;. O., and Fritz, H. P., Adrmn. Zi~org.Chem. Radiochem. I, 55 (1959); Angeu. Chem. 73, 353 (1961); (b) Fischcr, E. O., Angew. Chem. 67, 475 (1955). 9. Wilkinson, G., arid Cott,on, F. A,, Z’rogr. Inorg. Chem 1, 1 (1959). 10. Fischer, E. O., and Hafricr, W., Z. Natr~rforsch.B 10, 665 (1955). 11. Hubel, W., in “Organic Syntheses via Alecal Carbonyls” (I. Wender and P. Pino, eds.), Wiley (Interscience), New York, 1967. 12. Fischer, E. O., and Rlaasbijl, A,, Ange~r.Chem. 76, 646 (1964); Angew. Chem., Znt. Ed. Eizgl. 3, 580 (1964). 13. Ilyang, RI., Ithee, I., and Tsutsumi, S., Bull. Chem. Soc. Jap. 37, 341 (1964). 14. Fischer, E. O., Krcis, G., and Kreissl, F. It., J. Organometal. Cliem. 56, C37 (1973). 16. Aumann, It., and Fischer, E. O., Chem. Uer. 101, 954 (1968). 16. Meerwein, H., Hinx, (i., Hofmnnn, P., Kroning, E., and Pfeil, R., J. Pralct. Chem. CZ], 147, 257 (1937); Meerwein, H., Bettenberg, E. Gold, H., Pfeil, E., and Willfang, G., ibid. [2] 154, 83 (1940). 17. Fischer, E. O., and Maasbd, A., Chern. Zkr. 100, 2445 (1967). 18. Fischer, 15. O., and Kollmeier, H. J., Angetu. Chem. 82, 325 (1970); Angew. Chem., Znt. Ed. Engl. 9, 309 (1070). 19. Fischer, E. O., Winkler, I<., Kreiler, C. G., Ilut,tner, G., and Krieg, B., Angew. Chem. 83, 1021 (1971); Angew. Chem., Znt. Ed. Eiigl. 10, 922 (1971). 20. Fischer, 13. O., Kreiter, C. G., Kollmeier, 1%.J., Muller, J., and Fischer, It. D., J. Organometal. Chem. 28, 237 (1971). 21. Fischer, E. O., and Kreissl, F. It., J. Organometal. Chem. 35, C47 (1972). 22. Fischer, E. O., Kreissl, F. It., Kreiter, C. G., and Meineke, E. W., Chem. Ber. 105, 2558 (1972). 23. Wilson, J. W., and Fischer, E. O., J. Organometnl. Chem. 57, C63 (1973). 24. Fischer, E. O., and Offhaus, E., Chem. Ber. 102, 2449 (1969). 25. Fischer, E. O., Offhaus, E., Muller, J., and Nothe, D., Chem. Ber. 105,3027 (1972). 26. Fischer, E. O., Beck, 13.-J., Kreiter, C. C., Lynch, J., Muller, J., and Winkler, E., Chem. Ber. 105, 1F2 (1972). 27. Fischer, E. O., Kreissl, F. It., Winkler, E., and Kreiter, C. G., Chem. Ber. 105, 588 (197’2). 28. Fischer, E. O., and Itiedel, E., Chem. BET.101, 1% (1968). 29. Fischer, E. O., and Aumann, It., Chem. Ber. 102, 1495 (1969). 30. Fischer, E. O., and Heck, 11.-J., Chem. Ber. 104, 3101 (1971). 31. Mills, 0. S., and Itedhouse, A. I)., J. Chem. Soc. A 642 (1968). 32. Cotton, F. A., and Richardson, D. C., Znorg. Chem. 5, 1851 (1966). 33. Moser, E., and Fischer, E. O., J. Organometal. Chem. 13, 209 (1968). 34. Kreiter, C. G., and Fischer, E. O., Angew. Chem. 81, 780 (1969); Angew. Chem., Znt. Ed. Engl. 8, 761 (1969). 35. Kreiter, C. G., and Fischer, E. O., Chem. Ber. 103, 1561 (1970). 36. Connor, J. A,, and Fischer, E. O., J. Chem. Soc. A 578 (1969). 37. Fischer, E. O., and Kollmeier, H. J., Chem. Rer. 104, 1339 (1971). 38. Hawkins, N. J., Mattraw, H. C., Sabol, W. W., and Carpenter, D. R., J. Chem. I’hys. 23, 2422 (1955). Carbene and Carbyne Complexes 31
39. Kreiter, C. G., and FornibEek, V., Angew. Chem. 84, 1.55 (1972); Angew. Chem., Int. Ed. Engl. 11, 141 (1972). 40. Fischer, E. O., Pure Appl. Chem. 24, 407 (1970); 30, 353 (1972). 41. Cardin, 11. J., Cetinkaya, B., and Lnppert, M. F., Chem. Rev. 72, 545 (1972). 42. Cotton, F. A., and Lukehart, C. bl., Progr. Inorg. Chem. 16, 487 (1972). 43. Cardin, D. J., Cetinkaya, B., Ihylr, M. J., and Lappert, R.1. F., Chem. Soc. Rev. 2, 99 (1973). 44. Ofele, K., Angew. Chem. 80, 1032 (1968); Angew. Chem., Int. Ed. Engl. 7, 950 (1968). 45. Hatt,ner, G., Schelle, S., and Mills, 0. S., Angew. Chem. 81, 536 (1969); Angetu. Chem., Int. Ed. Engl. 8, 515 (1 46. Badley, E. M., Chatt, J., Richards, lt. L., and Sim, G. A,, Chem. Commun. 1322 (1969). 47. Fischer, E. O., and Beck, H.-J., Angew. Chem. 82, 44 (1970); Angew. Chem., Int. Ed. Engl. 9, 72 (1970). 48. Cardin, 1). J., Cetinknya, B., Lnppert, M. F., ManojloviE-Muir, Lj., and Muir, K. W., Chem. Commun. 400 (1071). 49. Kreissl, F. It., Fischer, E. O., Kreiter, C. G., and Fischer, 11. Chem. Ber. 106, 126%(1973). 50. Kreissl, F. R., arid Fischer, E. O., (‘hem. Ber. 107, 183 (1974). 51. Klabunde, U., and Fischer, E. O., J. Amer. Chem. SOC.89, 7141 (1967). 52. Fischer, E. O., Heck], B., and Werner, I-I., J. Organometal. Chem. 28, 359 (1971). 53. Fischer, I-:. O., and Leupold, RI., (‘hem. Ber. 105, 599 (1972). 54. Fischer, E. O., and Fontana, S.,-1. Organometal. Chem. 40, 367 (1972). 55. Fischer, E. O., Leupold, M., Kreiter, C. G., and Muller, J., Chem. Ber. 105, 150 (1972). 56. Fischer, E. O., Kreis, C., Kreissl, F. H., Kreiter, C. G., and Muller, J., Chem. Ber. 106, 3910 (1973). 57. Fischer, H., Fischer, E. O., Kreit,er, C. (;., and Werner, IT., Chem. Ber. 107, 2459 (1974). 58. Fischer, H., Fischer, E. O., and Kreissl, F. R., J. Organometal. Chem. 64, C41 (1974). 59. Werner, H., and Rascher, H., Inorg. Chim. Acla 2, 181 (1968). 60. Fischer, E. O., and Fischer, II., Chem. Ber. 107, 657 (1974). 61. Fischer, I<., and Fischer, E. O., Chcm. Ber. 107, 673 (1974). 62. Fischer, H., Fischer, E. O., and Werner, H., J. Organometal. Chem. 73, 331 (1974). 63. Weiss, K., and Fischer, E. O., Chem. Ber. 106, 1277 (1973). 64. Weiss, K., and Fischer, E. O., unpublished. 65. Kreis, G., and Fischer, E. O., Chem. Ber. 106, 2310 (1973). 66. Fischer, E. O., Schmid, K. R., Kalbfus, W., and Kreiter, C. G., Chem. Ber. 106, 3893 (1973). 67, Kreiter, C. G., Angew. Chem. 80, 402 (1968); Angew. Chem., Int. Ed. Engl. 7, 390 (1968). 68. Casey, C. P., Boggs, R. A,, and Anderson, It. L., J. Amer. Chem. SOC.94, 8947 (1972). 69. Fischer, E. O., Held, W., Riedmuller, S., and Kohler, F., unpublished. 70. Casey, C. P., and Burkhardt, T. .J., J. Amer. Chern. SOC.95, 5833 (1973). 71. Fischer, E. O., Walx, S., and Kreis, G., unpublished. 72. Schubert, U., and Fischer, E. O., Chem. Ber. 106, 3882 (1973). 73. Fischer, E. O., and Maasbiil, A., J. OrganomelaZ. Chem. 12, P15 (1968). 32 ERNST OTTO FISCHER
74. Fischer, E. O., and Plabst, D., Chem. Ber. 107, 3326 (1974). 75. Fischer, E. O., Heckl, B., Dotz, K. H.,Muller, J., and Werner, H., J. Organometal. Chem. 16, P29 (1969). 76. Fischer, E. O., and Riedmuller, S., Chem. Ber. 107, 915 (1974). 77. Fischer, E. O., and Dotz, K. H., Chem. Ber. 103, 1273 (1970). 78. Dotz, K. H., and Fischer, €3. O., Chem. Ber. 105, 1356 (1972). 79. Fischer, E. O., and Diitz, K. H., Chem. Ber. 105, 3966 (1972). 80. Cooke, M. D., and Fischer, E. O., J. Organometal. Chem. 56, 279 (1973). 81. Cf. Fischer, E. O., Weiss, K., and Burger, K., Chem. Ber. 106, 1581 (1973). 82. Dorrer, B., and Fischer, E. O., Chem. Ber. 107, 2683 (1974). 83. Fischer, E. O., and Dorrer, B., Chem. Ber. 107, 374 (1974). 84. Grob, C. A., Angew. Chem. 81, 543 (1969); Angew. Chem., Znt. Ed. Engl. 8, 535 (1969). 85. De Renzi, A., and Fischer, E. O., Znorg. Chim. Ada 8, 185 (1974). 86. Seyferth, D., Burlitch, M., Minasz, It. J., Yick-Pui Mui, J., Simmons, H. D., Jr., Treiber, A. J. IT, and Dowd, S. It., J. Amer. Chem. SOC.87, 4259 (1965). 87. Fischer, E. O., Kreis, G., Kreiter, C. G., Muller, J., Huttner, G., and Lorenz, H., Angew. Chem. 85, 618 (1973); Angeu:. Chem., Znt. Ed. Engl. 12, 564 (1973). 88. Iluttner, G., Lorenz, II., and Gart,zke, W., Angew. Chem. 86, 667 (1974); Angew. Chem,., Znt. Ed. Engl. 13, 609 (1974). 89. Fischer, E. O., and Richter, K., unpublished. 90. Fischer, E. O., and Schwanzer, A., unpublished. 91. Fischer, E. O., Schwanzer, A., and Kreiter, C. G., unpublished. 92. Fischer, E. O., and Kreis, G., unpublished. 93. Fischer, E. O., and Weiss, K., unpublished. 94. Fischer, E. O., and Meineke, E. W., unpublished. 95. Fischer, E. O., Kreis, G., Kreissl, F. It., Kalbfus, W., and Winkler, E., J. Organo- metal. Chem. 65, G.3 (1974). 96. Fischer, E. O., and Walz, S., unpublished. 97. Fischer, H., and Fischer, E. O., J. Organometal. Chem. 69, C1 (1974). 98. Kreis, G., Thesis, Technical University, Munich, 1974. 99. Fischer, E. O., Kalder, H. J., and Kiihler, F. €I.,J. Organometal. Chem. 81, C23 (1974). 100. Fischer, E. O., and Plabst, D., unpublished. Coordination of Unsaturated Molecules to Transition Metals STEVEN D. ITTEL Central Reseorch 8. Development Department E. I. du Pont de Nemours and Company Wilmingfon, Delaware JAMES A. IBERS
Deportment of Chemistry Northwestern University Evonston, Illinois
I. Introduction . . 33 11. Theoretical Models . . 3.5 111. Structural Results . . 37 A. Geometry of the Coordinated Olefin . . . 38 B. Geometry of the Complex . . 53 C. Complexes of Nonolefinic Unsaturated Molecules . 55 IV. Summary . . 59 References . . . A0
I INTRODUCTION
The interaction of unsaturatd molecules, for example olefiiis and acety- lenes, with transition metals is of paramount importance for a variety of chemical proccsses. Includvd among such processes arc stercospecific polymerization of olcfin moriomers, thc production of alcohols and alde- hydes in tho hydroformylation rcaction, hydrogenation reactions, cyclo- propanation, isomcrizations, hydrocyanation, and many other reactions. Many of thew processes takc place under both homogencous and hcter- ogeneous conditions. Because of their apparently greater simplicity, ho- mogeneous rcactions have servcd incwasingly as models for heterogeneous rc,actions. Many of these proceed catalytically in solution, and the direct methods available to characterize such reactions arc ncccssarily limited to kinetic and spectroscopic invest igat ions. As a result, unequivocal informa- tion on the nature of the all-important interaction of the unsaturated molecule with the transition metal is necessarily limited. A very useful modeling approach to gain more information on the metrical and bonding aspects of this interaction is the isolation and structural characterization 33 34 STEVEN D. ITTEL AND JAMES A. IBERS of related, inore stable complexes. Although such an approach must ncces- sarily be used with caution, since, for examplc, one might succeed in isolat- ing the least soluble rather than the most important spccies or since the Ir analogue of a Rh catalyst may not he a perfect one, the approach has yielded considerable information rrcontly on thr. nature of the interaction between unsaturated molecules and transition mctals. In this review we will concentrate on the metrical aspects of this inter- action as derived from a considerable body of structural data. WP will not attempt a comprchensivc. rcvicw; rather, \ic use sclectcd examples and discuss the results within the framework of current theories of bonding in these systems. Thcse cxamplcs will bv limited to noriring molecules inter- acting through onc double bond, that is, olefins, acetylenes, diazenes, ketones, and imincs. Coiisequc,ntly, the cmphasis in this review diffcrs from that of other rcccnt rc.vic.ns (40, 46,55, 58, '72) on mvtal-olefin and rdatcd complexes. In order to provide an overview of the interaction of unsaturated mole- cules with transition nwtals, the following remarks may prove helpful. If we take a simple olefin as a11 examplc, then the major characteristics of its interaction with a transition metal arc that (a) the C atoms are essentially equidistant from the mctal center, arid (b) the formcrly planar olefin bc- comes noriplanar mith the substitumts on the carbon atoms bending away from thc metal. The orientation of the C'=C bond with respect to the coordination geomc%ry surrounding the metal shon s some variation but generally is that shown in Fig. 1. (In Fig. 1 and throughout this review we considcr the olefin to br a nioiiodcntate ligand.) Figure 1 illustrates the observation that in squarc-planar (SP) complexes the C=C bond is ap- proximately perpendicular to the planc, in trigonal complcxes (TR) it is
SP TR
FIG.1. Four-coordinate square-pyramidal (SP), three-coordinate trigonal (Tlt), and five-coordinate trigonal-bipyramidal (TBP) olefin complexes. Unsaturated Molecule-Metal Complexes 35 approximately in the plan(>,and in trigonal-bipj raiiiidal complexes (TBP) the olcfin occupies an equatorial sit(. M it11 the C=C bond approximatcly in the equatorial plane. We now turn to a discussion of thc~orctical modcls for and structural information on this interaction Iwtv wn an unsaturated molecule and a transition metal.
II THEORETICAL MODELS
Thc iiiost n-idcly acwpted niotlc~lfor th(h patterns of bonding illustrated in Fig. 1 is that proposc>dby Ihm(2
The rchtive iiiiportanccl of (r and a contributions to the overall bonding is unclrar, but stvxal different combinations of relativc strengths lead to “limiting case” models. When thrre arc 2 dcctrons in the forward u-bond and 2 dwtrons in the a-backbuild, thcw arc 2 bonding clectrons for each metal-carbon bond. This is mathomatically equivalent to Zu-bonds and a mctallocyclopropanc structurcl (7‘2). This modcl does not ncccssitate strict sp3 hybridization at thr carbon atoms. Molecular orbital calculations for cyclopropane (15) indicate, that thc C-C bonds have higher carbon atom p character than do thc C-H bonds. Thus, thc metallocyclopropane inodcl allows a interactions with substituent groups on the olefin (G8). 36 STEVEN D. ITTEL AND JAMES A. IBERS
FIG.2. The Dewar-Chatt-Duncanson model for olefin bonding showing the u forwardbonds (left) and T backbonds (right).
The Walsh model (73),which has been uscd to explain the conjugation of cyclopropane with a-bonding substitucnts, can be uscd to gain an insight into the a interaction in olefin complexes. Thcrc is close similarity betwccn the orbitals involved in classic a-bonding (see Fig. 2) and the highrst occupied molecular orbital of the proper symmetry to interact with a sub- stitucnts (Fig. 3) in the Walsh model. Examination of known cyclopropane structures (49) demonstrates that a-withdrawing substituerits cause a shortening of the bond across the ring from the substituent. Calculations have shown (41) that removal of a-clectron density from the orbital weakens the two adjacent bonds while strcngthrning thc opposite bond. This model has important implications for the intcnctions of olcfins \\ ith transition metals drspitc the fact that thc many extra d orbitals available complicatc the situation. Finally, when the u forward bond makes a negligible contribution to the bonding and 2 electrons have been a-backbondrd, the metal is oxidized and the olefin functions as a bidentate dicarbanion with two o-bonds to the mctal. This model is consistent with the conventional practicc of assign- ing a ncgative charge to coordinated alkyl groups. Although somr justifica- tion for this model will be presented, it has not proved to be as useful as either thc Dewar-Chatt-Duncanson or the Walsh models. Recently, there have been several detailed calculations on the bonding
FIG.3. The highest occupied molecular orbital capable of r-bonding 'in the Walsh model (73) for cyclopropane. Unsaturated Molecule-Metal Complexes 37 of unsaturated molecules to transition metals. The results for Zeise's anion,' [PtCL( CzH4) ]-l, are morc complex than the simple a-r type inter- action (67). When the results are interpreted in terms of the Dcwar- Chatt-Duncanson model, the a-bonding effects are found to be consider- ably more important than the a-backbonding effects. When a similar calculation is performed for a complex of the ethylene-like molecule, oxy- gen, in the complex Pt (PHS) (02) (by),it is found that a back donation is now equal in importance to a forward bonding, but it takes place by a reorganization of electrons throughout the complex as it is formed. These two calculations demonstrat(., as cbxpcctcd, that as the electron density on the metal increases and the dcct roncgativity of the unsaturated molecule increases, a-backbonding bccomcs more important.
111 STRUCTURAL RESULTS
There have been too many crystallographic studies of transition metal- olefin complexes to present a. comprehcnsivc survey in this limited space. Therefore, only representative structiircs of major classes of compounds will be discussed, drawing on pertinent structural determinations as they are needed. Many of the important fcatures of olefin bonding can be illus- trated in the d10 system wh(w most of the complexes are approximatcly trigonal-planar (Fig. 4). Before proerrding with a discussion of bond lengths and related details, it is essential to consider th(. inhcwnt accuracies of structure determina- tions. In the usual X-ray diffraction experiment the scattering power of
FIG.4. A d10 trigonal-planar complex. L1 is considered to be trans to the X sub- stituents and Lz trans to the Y substituents.
For purposes of clarity, we depart from the usual IUPAC nomenclature rules and place the unsaturated molecule at the end of the formula. 38 STEVEN D. ITTEL AND JAMES A. IBERS
an atom increases with atomic number. Clearly the location of the carbon atoms of an olefin can be accomplished with greater precision the greater the contribution these atoms make to the total scattering power of the molecule. Thus greater precision is possible if the transition metal is lighter (first-row versus third-row) and if the other ligands are simpler [methyl isocyanide versus tri (cyclohexyl) phosphinc]. In most X-ray diffraction determinations the final parametcrs are obtained by a noriliiiear least- squares fit of calculated structure amplitudes to observed struct,ure ampli- tudes derived from the intensity data. Standard deviations on the derived atomic parameters are obtained by standard mathematical techniques in this process. In doing so it is assumed that all crrors rest with the observa- tions and none nith the theorctical model. Despite the fact that this as- sumption is not strictly valid, it is gericrally found that, for example, by comparing presumed equivalmt bond distanccs within the molecule with their errors obtained from the lrast-squares procdure, the cstirnated standard deviations are reasonable. But because of inherent limitations in the overall process, there is considerable evidence, based, for example, on comparisons of crystal structures pcrfoimcd on the same substance in different laboratories, that thcse standard deviations should be doubled or trrbled brfore applying the usuaI statistical significance tests. There is no doubt that considcrablc overinterpretation of the significancc of small diffcrcnces has occurrrd whcn thcsc facts have riot been kept in mind.
A. Geometry of the Coordinated Olefin
1. C=C Bond Lengths and Spectroscopic Manifestations The aspect of coordination geometry most readily explained by thco- rctical models is the lcngthening of the olefin bond. An examination of the data listed in Table I reveals that the bond length of a coordinated olefin is significantly longer than that of a free olefin [CZH4, 1.337(2); CZ(CN),, 1.34(2); CzF4, 1.31(4) A] but it is rather insensitive to the nature of the substitucnts on the olcfin or to thr othrr ligands trans to the olefin. This insensitivity has prompted the hypothesis (2)that in a series of complexes where thc trans ligands are held constant, the separation of the olefinic carbon atoms will remain constant for all simple olefins. This hypothesis, although lacking in a theoretical basis, does have limited validity. Yet a clcar-cut failurc. is found in the structurc of (C5H5)Rh(C2H4)(CZF4J (XXIV, Table I) where t he C=C distances are 1.405 (7) and 1.358(9) A for CzF4 and CZH1, respectively. These distances are significantly different, Unsaturated Molecule-Metal Complexes 39
i 15401 E I I I I
8 [L
r 2 cn I4*O t "I& 1460 130 135 1.40 1.45 1.50 C=C BOND LENGTH (A)
40 I I I I 1 9, I I I I -35- - -2 - 130- + - -L- - $25- - i- t- - LL ?20 - z7- g In -c- z 15- +=w+ -1 0 3 I= l0- 3 -- 6- z 5- -v$ P I I I I -I I I I 0 5
FIG.5. (A) The C=C stretching frequency, (B) the metal ionization potential, and (C) the chemicsu shift (T) plotted versus the C=C bond length for coordination complexes of ethylem.
even if one trebllls the estimated standard deviations. It is, therefore, clear that there a= some significant variations in C=C distances. For the detcctbn of subtle differences within a given series of closely related compounds, spcctroscopic studies are generally more useful than diffraction studiea. An inverse correlation between the C=C bond lengths and the infrared stretching frequencies, VC=C, in a series of ethylene com- plexes has been ooserved (63) (Fig. 5A). There is at present some discus- sion about the nature of the infrared band observed around 1500 cm-' [(70) and rcfereqces therein]. It is found that the normal mode contains TABLE I 0p.
STRUCTURAL PARAMETERS FOR SELECTED TRANSITION&f ETAL COMPLEXES OF UNSATURATED >fOLECULESa
*c hI+ M:Ld." ,I pd,f 6d.l ,,&I ed.f h-0. Compoundb Structure' (9) (A) (.i) (ded (des) (ded (ded (ded (ded Ref.
Olefin complexes
I Ni(PPhdz TR 1.43(1) 1.99(1) 2.152(5) - (HzC=CHz) I1 NI[P(O-O-TOI)I]? TR 1.46(2) 2.02(2) 2.005(2) - (HzC=CHz) 2.016(10) 2.121(4) I11 Ni[P(O-o-Tol) 212 TR - (HzC=CHCN) 1'46(2))& I.9Il(l%j 2.096(4) IV iSi[P(p-Tol)lz TR 1.471(19) 2.019(13) 2.181(4) - (PhHC=CPhH) V Ni(DCPE) TR 1.421(3) 1.081(2) 2.156(6) 54.5 (&fe?C=CMezj VI Ni(t-BuN=C:)z TR 1.47G(5) 1.954(4) 1.866(5) 56.8(c5) [(CN)zC=C(CN)z] VII Ni(PCy8) TR 1.401(14) 2.014(11) 2.196(2) - (HzC=CHz)z VIII Ni(BCH)a TR 1.40 2.063 - - IX Ni(CDT) TR 1.372(5) 2.024(2) - - X Ni(COD)z TET 1.39(1) 2.12(1) - - XI Ni(DUR) (COD) TET 1.325(13) 2.10(1) - - XI1 Ni(TDPME) TET 1.37(3) 1.86(2) 2.23(1) 84 (FzC=CFz) XI11 Pt (PPha)z TR 1.43(1) 2.11(1) 2.268(2) - (HzC=CHz) XIV Pt(PPh3)z TR 1.53(4) 2.11(2) 2.287(5) - [(CN)HC=C (CN) HI XV Pt(PPha)z TR 1.49(5) 2.11(3) 2.290(9) 6G.1 141.3 - - [(CN)zC=C(CN)zl XVI Pt(PPhs)z TR 1.62(3) 2.04(3) 2.285(8) 81.3(22) 132(2) - 78.1(15) ~ClzC=cC12) XVII Pt(PPhr)z TK 1.42(3) CN 2.10(2) 2.260(6) 62.0(26) 68.9(33) 131(3) 41 (ClrC=C(CN)z) )Cl 2.00(2) 2.330(6) ) 48.4(23) 153(3) XVIII Pt(PPha)z TR 1.416 (15) 2 [(PNP)HC= C(PNP)HI XIX Pt(PPhr)z TR 1.429(14) - 10.8(7) 1 [(CFdFC= 132(1) C(CFdF1 xx K[PtCla(HzC=CHz)I SP 1.37(3) 2.127(19) 2.327(5) - - - 45 1.354 (15) 2.139(10) 2.33 34.7 72.7 38 XXI Pt(NH:CaHu)Clz SP 1.36(6) 2.17(5) - - - l50(5) 3 (MeHC=CMeH) XXII Rh(ACAC) SP\CzHa 1.42(2) 2.19(1) 2.027(8) - - - 28 (HzC=CHz) JC2F4 1.40 (2) 2.01(1) 2.047(8) - - - (FzC=CFz) XXIII Rh(ACAC) SP 1.41 (3) 2.14(2) 2.03(1) - - 28 (HzC=CHz)z XXIV Rh(Cp) (HzC=CHz) -\C2H4 1.358 (9) 2.167(2) 2.171(6) 43.4 69.1 154.4 37 (FzC=CFz) JGFr 1.405 (7) 2.024 (2) 2.244 (7) i4.3 32.8 131.4 xxv Ir(PPhdzH(C0) TBP 1.431(20) 2.110(9) 2.317(3) - - 136(1) 53 [(CN)HC= C(CN)Hl 2.148(11) 2.400(3) 70.1(13) 54.8 - 54 XXVI Ir(PPhr)zBr(CO) TBP 1.506 (15) 136 [(CN)zC=C(CK)zl XXVII Ir (AsPhdzCl(C0) TBP 1.447 (23) 2.107(18) 2.480(2) 68.5 56.0 - 27 [ (CN) zC=C (CN)zl XXVIII Ir(PPhah(C6HNd TBP 1.526(12) 2.166(15) 2.392(7) 67.4(12) 56(2) 64 (CO) [(CN)zC=C(CN)zl XXIX Ir(PMezPh)z(Me) 1.386 (28) 2.191 2.323 (9) - 11 (COD) 1 ,362 (27) 2.224 - XXX Ir(DPPE) (Me) 1.459(21) 2.139 2.308 (3) - 12 (COD) 1.374(21) 2.215 - - 1.84(4) 154- (3) 62 XXXI Fe(C0) I[(HOG) HC= TBP 1 1.40(5) 2.09(3) - CH(COzH)] TBP 2 1.30(4) 2.06(3) 1.85(3) 153(3) TBP 3 1.40 (4) 2.03(3) 1.81 (4) 143(3) XXXII Fe(C0)r TBP 1.40 2.10 1.77 - 50 [HzC=CH(CN)I XXXIII Fe(C0)r(HzC=CHd TBP 1.46(6) - 22
(Conlinued) XXXIV Ni(t-BuN-C)z TR 1.284(16) 1.899(19) 1.832(28) 31.4(14) - 8(4) 172(4) 87.8(7) 2.6(7) 24 (PhC-CPh) XXXV Pt(PPhr)z TR 1.32(9) 2.03 2.28 40 - - - - 14 32 (PhC-CPh) 3.7(4) 20 XXXVI Pt(PPhdz[(CFa)CP TR 1.255(9) 2.028(5) 2.281(2) 39.9(5) m C(CFJ)I Z XXXVII 1r(PPha)z(C1HNz) TBP 1.29(2) 2.09(1) - 40 (1) - 47 CJ (CO) [(CN)CEC(CN)I =i 86.5 (15) 19 XXXVIII [Pt(Me)(PMezPh)z SP 1.22(3) - - ;;I (MeC=CMe)lPFs r 90 21 XXXIX Pt(C;H;NHz)CIz SP 1.24(2) 0 180 (t-BuC-C-t-Bu) 1.8(2) 18 XL Pt(TPB)(Me) TBP 1.292(12) 0 180 (CFaC=CCFz) Heteroatom complexes XLI Ni(t-BuN-C)z TR 1.898(4) 1.841(5) - 153.2 (4) 103.4 (3) 88.8 (3) 1.2(3) 25 ? (PhN=NPh) XLII Ni(P-pTols)n TR 1.930(5) 2.198(3) - 156 .5(3) 101.7 (3) 89.1 (3) 7.6(3) 42 (PhN=NPh) XLIII Nl(PPha)z TR 1.89(2) 2.249(7) - - 6.9 16 [(CFa)zC=Ol 1.87(1) 2.175(6) XLIV Ni(t-BuN-C)z TR 1.855 (4) 1.878 (4) 35.8 (2) 11.9 (10) 179.6(8) 82.5 (3) 7.8(3) 75 [t-BUN= 1.843(3) 1.819(5) 41.3t.7) - 168.5 (2) C=C(CN)zl XLV Ni(CsH8NO) TR 1.917(14) - - 77 4.4 52 [ (CsHsO)HC= 1.867(10) - 55 NHMel XLVI Pt(PPha)z[(CFi)z TR 2.02 - - 127 117 13 C=N[N=C(CFa)zlI 2.11 - a The complexes in this table have been selected from the many structures available to illustrate pertinent details of coordination. Cyclic olefins and olefins wth substituents which would complicate the interpretation have not been included except when necessary. * The compounds have been named with the ligands ordered as in Fig. 1, with the olefin last. Abbreviation are Ph = phenyl. To1 = tolyl, Me = methyl, t-nu = tert-butyl, Cy = cyclohexyl, DCPE = 1,2-bis(dicyclohexylphosphiuo)ethane,TDPME = tris(diphenylphosphinomethyl)ethsne,BCH = bicycloheptene, CDT = cyclododecatriene, COD = 1.5-cyclooctadiene. DUR = duraquinone, PNP = para-nitrophenyl, ACAC = 2.4-pentanedionato, Cp = cyclopentadienyl, DPPE = 1 ,2-bis(dipheny1phosphino)ethane.TPB = hydrotris(1-pyrasolyl)borato. .4bbreviations: TR = trigonal planar, TET = tetrahedral, SP = square planar, TBP = trigonal bipyrsmidal. These values represent averages except in special cases where individual values are given to illustrate differences. Values given are for the trans ligands. I These angles are defined in the text. Note thatol is defined differently for double- and triple-bonded molecules. 44 STEVEN D. ITTEL AND JAMES A. IBERS a significant portion of CH2 scissoring deformation. A lower-energy, strongly polarized Raman line may contain more C=C stretching char- acter, but this band is found to be rather inscnsitive to the nature of the complex. The variation of thc 1500 cm-’ band, which may rcsult from differences in the mixing of the C=C stretching arid the CH2 scissoring motions, is thus a more sensitive probr of the nature of the complex. The ground-state ionization potentials of the mc.tals invo1vc.d in the ethylene complexes arc found to corrclatr directly with thc C=C stretch- ing frclquencies or inversdy with the (’=C bond lengths (Fig. 5B). These ionization potentials of the free metals am an indication of thc ability of the complex to modify the s-electron density to thc olefin, although they arc undoubtedly aff ected by the varirty of remaining coordinated ligarids. This modified elcctron density on thr olcfin is also reflected in thr chemical shift (7) of thc olefinic protons in the NMIL sprctra (Fig. 5C). Larger values of 7 indicate a greater elcctron density on the olefin *-orbitals re- sulting in increased magnetic shielding. Although these vibrational and NMR effccts would br rxprrtcd to hold for substituted olrfins, it is diffi- cult to predict the additional perturbations on the spectra caused by substituents. It is evidtmt that thr elcctron density on the olefin and thr resultant changes in bond lrngth, nuclear shirlding, and vibrational niodcls are re- lated to thc ionization potential of th(. metal atoms. It is not clca nhethrr this effect is causcd by inhibition of thc u forward donation or enhance- ment of the a back donation as the ionization potrntial is lowered. When elcctronegative substituents are placrd 011 the olvfin to “activate” it, there is also the question of the extrnt to which the backbondrd electron density rcmairis localized on the double bond or is drlocalizcd to the substituent groups. Any means of probing the electron drnsity on the metal or olefin would be useful in ascertaining the relative importance of the u and ?F effects. Elrctron spectroscopy for chemical analysis (ESCA) has been employed to measure the elcctron drnsity crnttwd on thc mctal in a series of plati- num complexrs (14).Thr binding oricrgies of the mctal rlectrons were interpreted on the assumption that vithin a serks of complexes the binding energies of a given level are dependent only on the net electronic charge transferred from the mrtal to the ligand. On this basis it is possible to assign a dcgree of oxidation to the mcxtal. In thc platinum srries, oxidation states of 0.0 and 2.0 were assigned to Pt ( PPh,) arid Pt (PPh,) 2C12, respec- tively. It was then fourid that in the complex Pt(PPh,)zL, 0.7, 0.8, and 1.8 electrons wer(1 transferred to L whcn L was diphrriylacetyleric, ethylene, and dioxygen, respectively. These rrsults indicate that there has been a net transfer of electron density to the unsaturated molecules. Thus, as Unsaturated Molecule-Metal Complexes 45 expected, the high e1ectronegativit)y of dioxygen, which is bonded to the metal in a sideways a fashion, rcsult,s in a complex in which almost 2 elec- t'rons have been removed from the inctal center. The degree of platinum oxidation in the olefin and acct'ylcnc complexes is not as great but it is, nonetheless, significant,. Another probe of the electronic propchm of a metal center is the stretch- ing frequency of a carbonyl or isocyanidc (42) ligand in the complex of interest. The limitations of this tcchniqun are the constraints on the nature of the complexes studied and thc unccrt,ainty of the relation between the stretching frequencies and the actual clcctron density on the metal, but these are more than offset by tho grttatc.r ease of measurement and the greater resolution of this tcchniquc when compared with ESCA measure- ments. Table I1 presents the isocyanidc st,retchingfrequencies for a variety of complexes, Ni(t-BuN=C)J, (35, 4.2,44, 6O), listed in order of increas- ing froqucncy. The complexes at the t,op of the list arc considered to con- tain nickel(0) , whercas thosc: at thc: bottom contain nickel(I1). In the intermediate complexes, varying d(?grc:cs of clectrori transfer from the nickcl atom have occurred. Alt,hough c.thylcnc does not appear on the list, some substituted olefins do. If on(' again adopts the proceduro of assigning degrees of electron transfer, thcn vN=C around 2000 cm-' indicates no elec-
TABIAE I1 ISOCYANIDESTRETCHING FREQUENCIES FOR COMPLEXES Ni (t-UriN-C)Z (Un).
Un
(t-BuN=C)n 2000 - (PEta)z 2023 1970 trans-PhHC=CPhH 2100 2057 CHF=CH(CN) 2138 2102 PhC=CPh 2138 2110 trans- (CN)HC=CH (CN) 2164 2138 PhN=NPh 2168 2140 (CF3)CEC (CF3) 2169 2134 (CFs)FC=CFz 2170 2130 (CF$)zC=NH 2182 2150 ~~u~s-(CF~)FC=CF(CF~) 2185 2152 (CN)&=C(CN)z 2194 2179 04 2 196 2178 (CF3)ZCeO 2199 2183 (CF&C=C (CN)z 2210 2190 -(CFz),- 2230 2219
Un = an unsaturated molecule. 46 STEVEN D. ITTEL AND JAMES A. IBERS tron transfer and vN=C above 2200 rcprescnts the 2+ oxidation state. Itj is then found that a-bondcd molecules accr.pt bctwecw 1 and 2 electrons, depending on the substituciit groups. It is again evident that a-bonding can make a substantial contribution to the overall bonding. a-Halogcn substitucnts, which may be considcrcd to be a-acceptors and a-donors, apparently rcwlt in lcss hackbonding when compared with the cyario group, which is a wtak u- arid a strong a-acceptor. This is in contrast to the P-halogen suhstitumt, CF3, nhich sccnis to rcsult in greatc.r removal of electron density from thc metal than docxs the cyano group. The classic activation of thc doublc bonds by introduction of electronrgative substitu- ent groups can also ti(> accomplished by introducing idrogcn or oxygen into the double bond. When similarly substituted, the clwtron-\tithdran ing powcrs of the various 2-atom bridges incrmsc as C=C < C-C: < N=N, and (’=C < C-N < (‘=O. Substituting one cyano group onto a bridging carbon atom makes it approximatt.1y as vff(v3ive in c~lcctron-withdrawing power as a bridging nitrogen atom, and tno cyano groups make it approxi- mately as effective as a bridging oxygcln atom. Thus intrinsic activation by modification of tho bridging double bond has the same effect as classic extrinsic activation by substituent groups (43). A probe of the electron density on the double bond is discussed in Section III,C,2.
2. Metal-Carbon Bond Lengths The metal olcfin bond lengths prcsentcd in Tablr I exhibit two inttwst- ing trends. The first is that the shortest &I-C bond lengths arc associated with olcfins twaring halogen suhstituents. Although it is well known that cyano groups make olefins much httcr a-acids, halogen substituents rcsult in shorter metal-carbon bond lcngths. This cffcct has to be rationalized on the basis of the U-a bonding scheme. The recent calculations on the bonding for [PtCld(C2H,)]- show that u forward bonding is the major contributor (67) . Whrrcas in a zero-valent platinum complex a backbond- ing would increase in importancc because of the higher electron density on thc metal, wc would still expect u forward bonding to make an important contribution. Introduction of cyano groups lowers both the T- and a*-olcfin orbital energics. This rcsults in reduced u forward bonding and greatly increased a-backbonding. There is a net increase in the strcngth of the metal olefin bond, and the bond becomes more polar or ionic. Introduction of halogen substitucnts, specifically fluorine, docs not change the level of thc a-orbital, but the mergy of the a*-orbital increases greatly (70).This should have the effect of reducing the overall bonding by reducing the a-backbonding contribution. Although this has been observed for mono- Unsaturated Molecule-Metal Complexes 47
2.20 1 I
- 2.10 '4 u 1 nL 2.00 ttt I I 2 5 2.30 2.35 Pt-PG,
FIG.6. Plot of Pt-C bond distances versus trans-Pt-P bond distances for a series of olefin(L) complexes, Pt(PPhs)z(l,). and disubstitution (70),it is not observed for the tetrasubstituted olefins studid crystallographically. Ti\ o explanations have been given for this seemingly anomalous result. l'hv first is that there must be an extensive reorganization of molecular orbitals ming to the change in geometry of the olcfin, and this rcarrangcnmit rendcrs use of uncoordinated molecular orbitals inappropriate (70); the rcsultant molcculc has a high dcgrcc of cyclopropanc character. The scwnd explanation is based on a thrre-center bonding molecular orbital (2, 51) in nhich the a-donating properties of halogens incrcascl the electron dcnsity in that orbital. This would result in a more covalent bond. Whrrchas both of these explanations are useful, it may be convenient to attribut(. the enhanced bonding to a purely inductive enhancement of thc a-backbonding based on the clectroiicgativity of the substiturnts. A second trend citrd in nic>tal-olefinbond lengths is their inverse corre- lation with trans metal-ligand bond Imgths (3). This phenomenon has been observed in several structures in which the 2 olefinic carbon atoms are at different distances from the metal. A vivid illustration is found in the platinum complex Pt (PI'h,) 2[('12C'=C (CN),] (XVII, Table I). The effect was attributed to the strong u-donation capability of the CCl, group. A plot of Pt--C distances vcrsiis Pt-P distances for the serics of com- plexes I't (PPh3)2 (olcfin) (Fig. 6) revcals a poor correlation. It is found that several complexes in Table I show a direct rdationship between metal-olefin and metal-ligand bond distances. For example, in Ni[P (O-o- Tol)3]2[H&=CH (CN)] (111) t he cyaiio end of thc olefin is significantly closer to the metal, and the tra,u-phosphitc ligand is closer to the metal than is the cis-phosphite. This can bc rationalized on the basis of the stronger *-accepting capability of the cyano end of the olefin. The metal- 48 STEVEN D. ITTEL AND JAMES A. IBERS
FIG.7. A trigonal complex displaying displacement of the olefin along the double bond. carbon bond is shortened by cnhancvd T-backbonding, and the trans- phosphite can then form a strongcr and thus shorter o-bond to the metal. Again we see a difference betwren the activation of an olefin by r-with- drawing or by donating substitumts. This discussion of the metal-carbon distances should include another closely related effect: sliding of the olefin along the carbon-carbon bond. This effect, pictured in Fig. 7, involves a displacement of thc centroid of the olefin double bond in thc coordination plane relative to the bisector of the ligand-metal-ligand angle. This dfect is apparently not a steric one, as both Ni[P (0-o-Tol) 3]2[H2C=CH (CN)] (111) and Pt (PPh,) 2- [Cl,C=C(CN)J (XVII) display a marked shift to bring the cyano end of the olefin closer to tlw biscctor, cvcn though in the platinum complex the chloro end is closer to the nirtal. Thc samr effect is observed in Fe(CO) *[H2C=CH (CN)3 (XXXII). The shift of the cyanoolefins is perhaps caused by intcmction of the a*-orbital associatcd with the cyano groups with the d orbitals of the nictal.
3. Xonplanarity of the Bound Olefin We next discuss the nonplanarity of the bound olcfin. As we indicated in Scction I, this nonplanarity manifests itself in the bending back of sub- stituent groups away from the metal. Thcre arc a number of ways to de- scribe this bending back. The mmsure easiwt to picture but least informa- tive is the olefin doublr bond to substitucnt angle, which would be 120" for an idealized sp2 hybridized carbon atom, and 109.5" for an idealized sp3 hybridized carbon atom. This measurc will bc used in the description of metal-acctylencl bonds. A more informative measure of the bending back are the a and p angles defined in Fig. 8, which have come into common usage (69). The anglc (Y is the anglc between the normals to the planes defined by thc substituent groups; p and p' are the angles between the olefin bond and the plane normals. As bending back of the substitucnts occurs, (Y increases from 0" and the p angles decrease from 90". Thc sum Unsaturated Molecule-Metal Complexes 49
FIG.8. Illustration of angles a, 0, P', v, and 7'. of the three angles need not lw 180" unless the bending back of the sub- stitucrit planes takes place so that thr two plane normals and the C=C bond all remain in the samc plan(%.It is gmcrally found that the sum of the anglrs is nrar 180". Insprctiori of t,hc values of a and p in Table I rcvcals that bcnding back is lvast for hydrogon atoms, increases for cyano groups and alkyl groups, and is grrat (1st for halogen substitucnts. This parallels thc trvnd observed for metal-rarl)on distances (Section 111, A, 2). If one fashions a geomctricd ~notldbased on the "mctallocyclopropane" model for bonding, the sul)st itucwt, atoms on the olcfiri should exhibit trtrahedral geometry. They cannot do this because of the severe steric and rlectronic nontctrahrdral rest raints imposed by a three-membered ring. A geometrical modcl proposcd (37) to account for the simplest of steric considerations is constructrd I)y bisecting the M-C=C angle, 1 (the dottrd line in Fig. 8). The suhstituriit atoms are required to lic on a plane pcrprndicular to thc M--C=(' plan(. and containing the bisector. This construction results in thc now "t(~trah(~dra1"values of 01 and p defined by
The only distanw criterion in asscwirig the geonictrirs is fixed by the ratio of the M-C to C=C distancm. Thrrr are several possible tvsts for this assumed tetrahedral geometry. The gc.omctry of ethylene oxidr lias been determined (71).The observed valucis of a and p of 47.8" and (i8.G" cornpared with thr calculated values of 39.2" and 60.4"indicatr that thv rnodcl prrdicts cxcrssivc. bonding back. In the molecule rthylcne sulfide,, t h(1 ring shapc more closcly rcwmblcs an olefiri complex. The obscrvcd valucls of 01 and p of 56.6" and 61.7" (17) are closer to thr calculated valum of 65.8" and 57.1", but again tho model pre- dicts excessive bcnding back.
When the measured valuw of (Y and p are compared with the rcqective values of 01y and pr calculatcd for thc various complexes listed in Table I, 50 STEVEN D. ITTEL AND JAMES A. IBERS
FIG. 9. Illustration of angles y, 6, and 6'. it is found that thc only complcxes in which thc hending back is grcater than predicted are those of halogen-substituted olefins. Thus, again wc ob- serve that halogen substituents cause greater changes in the olefin geom- etry upon coordination than do other substituents. In complexes of olefins where the positions of all four substituents can- not be determined, another measure of the bending back must be used. The measure commonly cmploycd involves thc torsional angles about the carbon-carbon bond (42).This is particularly useful in the case of trans- disubstituted olefins. Figure 9 illustrates the angles for a tram-olefin. Therc is an M-C=C-R torsional angle, 6, for each substituent and an R-C=C-R torsional angle, y, for each pair of trans substituents. The angles 6 and 6' in Fig. 9 incrcascl from 90" as bending back occurs and y decreases from 180". The sum of the three angles must nccessarily equal 360". For complexes with no trans substitucnts, cis-disubstituted olefins, or monosubstitutcd olefins, only the 6 angles can be measured. As might be expccted, thcre is a good correlation between 6 and a,which was defined in Fig. 8, for tetrasubstituted olefins as shown in Fig. 10. Thus 6, or the related y, is a good measure of the bending back of substituent groups. In complexes of symmetrical trans-olefins where the positions of all four substituents have bccn determined, it is possible for one set of trans sub- stituents to be bent back more than the other set. This twist of the CR2 groups about the C=C bond would not be obvious from an inspection of the a and @ angles. This is truc in compound XIX (Table I) where the CF3 groups arc bent back more than the F atoms; y = 132 (1)O and 120 (1)" for F and CF3, respcctivcly. It is difficult to ascribe this effect solely to either steric or electronic causes. It seems reasonable that certain groups bend back more for elcctronic reasons, but in this instance it is the bulkier group that is bent back more. It is not possible to predict on electronic grounds whether F or CFS should bend back morc. In the Ni complex of tetrainethylethylene (V, Table I), where there is no difference in the size of the four substituents, the two observed values Unsaturated Molecule-Metal Complexes 51
116 I I I I I 114 - 112 - y 110 - - $108- t t+ - - f t - 102 - t - - 100- , I I I I 1
FIG.10. Plot, of 6 versus a for olefin complexes. of y are 139" and 152". This twist of the two C(CH8)a groups about the olefin C=C bond has the dfcct of orienting the r-orbitals of the carbon atoms slightly toward the Ki,P,1' plane. The opposite is found in the octa- fluorobutene structure (XIX, Table I) and more importantly in the tetra- chloroethylene structure (X6'1). I'cdiaps the twist is stcric in nature, having little significancc. in th(1 dcscription of the bonding, but this remains to be detrrmined.
4. Twist of the OleJin out of the Coodination Plane A twist of the olcfiri out of thc coordination plane is common in trigonal- planar d10 structures. Two intlrpendc.nt measures of this twist have been used somcwhat interchangmbly although they are not necessarily equiva- lent. The more common nwasurc is the angle 0 between the normals nl to the L-M-L plane and Q to the C-M-C plane, illustrated in Fig. 11. A sccond measure is the angle p between the C=C vector and the normal nl to the L-M-L plane. Thcsc two angles are complementary only when the rotation of the olefin takvs place about the line D, bisecting both Q LML and Q CMC (Fig. 11A). Frequently this is not the only distortion
FIG.11. Illustration of the two possible modes of twisting of the olefin out of the P-M-P plane. The angles are, of course, measured between the normals to the planes. 52 STEVEN D. ITTEL AND JAMES A. IBERS of the coordination plane, but there is also a trigonal pyramidal distortion with the metal moving out of the ligand plane (Fig. 1lB). The most marked example of this pyramidal distortion is found in the nickel TCNE complex VI (Table I). Thc intcrplanar angle 0 of 23.9 (2)O is one of thc larg~stobscrvcd for a simple oldin complex, but the vector-plane normal anglc cp of 82.2(2)O is not, cxccptional. The 16.1” deviation from comple- mentarity results from the pyramidal distortion reflected in the deviations of the 5 atoms from a Icast-square: planc.. The nickel atom is above and the rmt of the atoms avwagc. 0.25 ii below the planc. This extrcme pyra- midal distortion has not hecln observed in any other complvx; it may be attributable to intermolecular packing forcm. The more commonly obscrwd distortion, purc rotation of the olcfin about axis D,rcmlts in compl(1mcntary anglcs. There have been scvclral conflicting thcorctical calculations dcaling with this distortion, predicting a potcmtial minimum for both planar (74) or slightly twisted (39) corn- plexes. Thesc calculations arc based on Cz, symmetry and have not con- sidered thc reduction of symmetry to Cz oftcn brought about by the pack- ing of the bulky groups on the trans ligands. It is now wll-wtablishcd that in solution olefins exhibit a very low barrier to rotation about the mctal-olcfin bond in a variety of complexes. Hcricr a slight rotation about the axis D rquircs little energy and prohably occurs frequently in order to lessen nonbondcd contacts.
5. Orientation of x Subslitue,its Anothm aspcct of the gcomcltry of the bound olcfin that has bcen barely studied is thv orimtation of substitumts capable of x interactions with the olefin. The most studid x substituent is th(1 cyano group, but its linearity precludes discussion of the naturc. of th(k intcmction. The structures of two complcxcs of diphmylcthylcnes (IV arid XVIII, Table I) haw bcen de- termincd. On the basis of electronic cffrcts om would expect a phcnyl ring either to bc coplanar with the oldin doubl~bond for better conjugation or to be perpendicular to the nic.tal-olefin planc for greatest x overlap (in the cyclopropane model) . Thc limit cd evidcnce favors thc second orienta- tion. Howcver, structural studics of oldms with substituent groups such as -COH, -COOR, or -NOz would be useful for the further definition of the orimtation of x suhstitucnts.
6. “Poiriting” of the Trans Liganils One final effect which has bccri riotrd (5) is the variation of the direction in which the trans ligands “point.” This cffcct is not peculiar to olefin Unsaturated Molecule-Metal Complexes 53
FIG.12. Illustration of the “pointing” of the trans-phosphorus ligands. complexes but is observrd in all types of complexes involving phosphorus and other ligands. There arc’ two manifestations of this cffcct that can be measured. It is very commonly noted that, whereas thr three carbon- phosphorus-carbon anglrs :m cqual, the three metal-phosphorus-carbon angles for phosphine ligands arc not equal; thcrcfore, the ligand is not pointing directly at the mc\tal atom. Another indication of this cffcct is found in the anglc betwc.cn thr normals to the planes described by the 3 a atoms of each phosphorus ligand. As Fig. 12 indicates, the normals may intersect in front of or behind thc mctal atom, resulting in an interplanar angle greatcr or less than t hr pliosphorus-metal-phosphorus angle, re- spectively. It is found that in complcxes containing bidentate phosphorus ligands, such as V (Tablr I),thc phosphorus atoms point behind the metal owing to the structural constraints of thc chrlate ring. It is interesting that for thc rrmainder of tlw applicable complcxes listed in Table I the phosphorus ligands generally point in front of the metal atom, although thtw are also distortions ahovc and below thc coordination plane. These effrcts most probably arise from stcric causes, with the phosphorus ligand turning toward the greatcr rionboridt3d interaction. Nonethclcss, these distortions should be considered in discussions of olefin coordination.
B. Geometry of the Complex The foregoing discussion &alt primarily with dl” complexes, and, al- though deviations from thc coordination plane were discussed, the overall geomcltry of the complexes was not dcalt with in detail. A recent calcula- tion (66) predicts that tris( cthylcrw) nickel(0) will be a trigonal complex with all 3 rthylrne molcculcs lying in thc coordination plane. The structures of nickel (bicycloheptcnc) (VIII, Table I) and Pt ( C2Hd)2 ( C2F,) (34) are consistent with the calculation. This same geometry is found when one 54 STEVEN D. ITTEL AND JAMES A. IBERS
A 0
FIG.13. The so-called cubic and dodecahedra1 modifications of tetrahedral four- coordinate olefin complexes. olefin is rrplaccd by a phosphorus ligand; thus, complex VII (Table I) has both ct hylcnc molrculcs in t ht. coordination planci. The structures of many complcxm h’i (PR,) (oldin) arc kiimvn, and in each the olcfin is found in the coordination plane, so that a planar or nciarly planar gcomctry rcsults. In many of these complcxcs thew is littlc stvric intwaction bctwccn the two substitumt ligands and the olcfin. In othrrs there is pronounced stcric interaction and yet thc complexes rcmain (wcntially planar. Four-coordinate complcxcs can assume tit o gc.omctries-square-planar or tetrahedral. It is found that dl” complcxcs arc usually tctrahcdral. A tctrakisolcfin complex can assumc one of two modifications of this geom- etry, &hcr ‘[cubical” or “dodecahedral,” as shown in Fig. 13. It has been predicted that the dodecahedra1 conformation would be more stable (66), but the known d’O four-coordinate compl(.xcs (X and XI, Table I) involvc cyclic olcfins that lirriit the geometries to cubical. The only structure (XII) of a tc%rahcdral complex in which the olcfin is free to assume its prcfcrrcd geometry involves a tridcntatc phosphorus ligand. Details on this structure are limited, but it appcars that the olcfiri lies close to a plane defined by 1 phosphorus atom and thc bisector of thc P-Ni-P angle between the other 2 phosphorus atoms. If so, the tctrafluoroethylene has assumed the dodecahedra1 geometry. This conclusioii is tenuous, and other comp1cxc.s of this type should bc invmtigatcd. Four-coordinate complexes involving d8 metals are found to have a square-planar geometry. An olcfin could potentially assume a geometry parallel or pcrpcndicular to the coordination plane; it is found in all cases that the olcfin is pc.rpcndicular to the plane. This can be rationalized par- tially on stcric grounds; niolecular niodels indicate impossibly close non- bonded intc.ractions between an olefin in the equatorial plane and the ligands cis to thc olcfin. Of course, thcsc arguments do not eliminate modest variations from the perpendicular position. In addition, calculations (66) have pr0vidt.d cvidence on electronic grounds that the perpendicular con- forniation is more stable than the parallel conformation. Unsaturated Molecule-Metal Complexes 55
Five-coordinate complexrs an' goncrally either trigonal hipyramidal or square-pyramidal, with the former gcoinrtry favored by d8 metals and the latter by CP metals [ (30) and rc~f(~rcwcc~stherein]. The known five-coordi- nate monoolefin complexcs arc. trigonal hipyramidal, presumably because d*, rather than d6, systems havc. bwn extensively investigated. In these trigonal bipyramidal nionool(6n complexes, the olcfin is equatorial and lies in the trigonal plane. l'hc nonbonded interactions between thr olefin and the axial ligands would lw cquivalcnt to the interactions between the olcfin and the cis ligands 90" away, noted above for square-planar com- plcxcs. Thc nonbonded contacts in the equatorial plane an' much lrss be- cause the ligaiids are 120" away. Thus, again the observed geometry of these trigonal bipyramidal coinpI(ws can bc rationalized on stcric grounds alone. A rcccnt calculation on tllo five-coordinate complexes (66) has shown that a a-acceptor ligand in an (quatorial position nil1 prefer a configura- tion in which its a-acceptor orbital is in the planc. Although there are no knomn fivv-coordinatc> d'" olrfin cwmplexes, the results arr applicable to d8 coniplcxrs also. Thus both stwic and electronic effects again predict the sanic conformation. Several five-coordinat e coniplr,xc.s involving the cyclic diolefin, 1 ,5-cyclo- octadienc (COD), have brcn invcstigatcd (XXIX and XXX, Table I). The olefiri coordinates in an (quatorial-axial manner, thus allowing the equatorial double bond to lit. ill thc, trigonal plane. If the olefin had co- ordinated in an cquatorial~cciuatorialfashion, the t\\ o double bonds would necessarily be pcrpcndicular to tho trigonal plane. This prcferred orienta- tion may also be influencrd by the approximately 90" bite of COD. In kwping with other five-coordinatr complexrs, the axial hl-C distances are greater than the correspoiiding equatorial distances. The stronger equatorial interaction also rcwilts in a longer C=C distance.
C. Complexes of Nonolefinic Unsaturated Molecules
1. Acetylene Coniplexes A modest numbvr of acrtylcw complexes has bccri invrstigated struc- turally. Marly of the features of olcfin complexes are also observed in acetylenc complexes, the major difference being the change in geometry of the coordinated acetylenc. The acctylene molecule approaches the geometry of a cis-olefin with the C=C-R angle deviating frorii lS0" by the angle a. From Table I one finds that, for a variety of complexes, a ranges from 12" to 40". There 56 STEVEN D. ITTEL AND JAMES A. IBERS again seems to be no correlation between C=C distances and the bending back of substituent groups, but the number of structures is limited. It should be noted that metal-carbon distances are about 0.07 8 shorter in acetylene complexes than in related olcfin Complexes. Much of this differ- ence could be attributed to the 0.04 8 change in the carbon single-bond radius on going from sp2 to sp hybridization. Yet the carbon-carbon dis- tance between the bridge and ring carbon atoms of coordinated diphenyl- acetylene and trans-stilbene are not appreciably different. Thus, if one equates bond length with bond strength, then one concludes that acetyl- me's intckract mom strongly with metal complexes than do similarly sub- stituted olefins. This conclusion is in accord u ith theoretical predictions (56). When the usual description of metal-olefin bonding is extended to metal-acetylene systems, the same orbitals are used in the model. But acetylenes possess an additional set of T- and a*-orbitals orthogonal to the metal-olcfin plane. A calculation indicates that interaction of this T*- orbital with additional metal d orbitals could be significant. Thus the metal-acetylene bond could be strengthened beyond that of a metal-olefin bond. In general it is found that thr carbon-carbon bond of acetylenes is lengthened less upon coordination than that of olefins. This might be in- terpreted as an indication that acetylenes interact to a lesser degree. Yet, note that the correlation between bond length and bond strength is not linear. Thus the increases in going from a triplc bond to a double bond and then to a single bond arc 0.13 and 0.21 8,respectively. A smaller lengthen- ing of an acetylenc. molrculc relative to an olefin nioleculc can still indicate a comparable change in bond order. As was noted in the discussion of olcfin complexes, a twisting of the C-C bond is usually observed in acetylene Complexes. This twisting is mani- festc.d in a nonzero R-C-C-R torsion angle. Angles y and 6, defined the Sam(' way as for olefin complexes, should ideally be 0" and 180", respec- tively. Va1uc.s of y up to 9" have been observed, and the two 6 angles are not necessarily equal (Tablr I). As for olefin complexes, the differences in the 6's can be attributed to nonbonded intmactions, both intra- and intcrmolccular. The variations in y for various diphenylacetylene complexes can bc attributcd to minimization of rionhondcd contacts between the two phenyl rings. The contacts between ortho-hydrogen atoms would be rather close in a strictly planar molecule, and conjugation between the phenyl rings and the acetylene bridgcx is interfered with if the rings twist too much. The remaining method for relicf of the hydrogen atom contacts is twisting about the acetylene bond, Although this argumcnt seems plausible, a twist of 9" is observed in a dicyanoacetylenc complex (XXXVII, Table I), Unsaturated Molecule-Metal Complexes 57 where steric crowding is minimal, and 0" (by symmetry) is observed in a bis (tert-butyl) acetylene complex (XXXIX), where nonbonded contacts might become important.
2. Diazene Complexes The structures of two complexes (XU and XLII, Table I) of the di- phmyl-substituted diazenc., azobciizcwe, have been determined. Azo- benzene is found to be capable of stronger a-backbonding than the iso- electronic traizs-diphenylethyleri(~. This cffect is manifested in several structural aspects. After thr diffcrcncc. in nitrogen and carbon atomic radii is considered, it is found that thr N=N bond is longer and the M-N bond is shorter than the respectivc hotids in the olefin complex. Angle y is
6.2( 13)O Icss in the azobenzme compl~~x,indicating a greater bending back of the phenyl rings. Molccular orbital calculatioiis (43) show that the highest occupied molecular orbitals of azobcnzenc and trans-stilbene are at approximately the same energy, but the lowest unoccupied molecular orbital of azobcnzene is much lower than that of trans-stilbene. Thus azobenzene should be capable of better a-backbonding. The higher clcctroncgativity of the N=N bridge compared with that of a C=C bridge would also argue for better A-backbonding for a diaz(~iw. In the Ni(0) complexes of azobc~nz(mm,the n -+ a * transition observed in the visible spectrum provides an additional probe of the multiple bond (42, 44). It is found that the transition shifts to higher energy as the elec- tron density on the doublc bond is incrcased. When electron density is in- crcascid by putting more electron-donating ligands on the nickel atom, the nickcl-azobcnzcnc bond is strcngthcncd. If, however, the electron density on the N=N bond is increased by putting elcctron-donating substituents on the phcnyl rings of the azob~nz(~nr,the coniplcx is destabilized.
3. Ketone Conzple.res
Several complexes of hexafiiioroacc.toiie with dlo metals have berii prc- pared. Based on 19FNMR data, thew complexes were thought to involve a ketonic C=O group involved in a a intcraction with the metal. This was confirmed by a structural detcrniination of the nickcl(0) bisphosphine complex (XLIII, Table I). In this structure thc CF, groups are bent back about as far as in typical halogenated olefin complexes. Interestingly, the oxygen atom and the carbon atom (with a larger atomic radius), are ap- proximately equidistant from thc nickcl atom, whereas the phosphine ligand trans to the carbon atom shows a significant trans cffcct. This sug- 58 STEVEN D. ITTEL AND JAMES A. IBERS gcsts that a CF3-substituted carbon atom may be a better ?r-acceptor than an oxygen atom, consistent with the spectroscopic results of Section 111, A,1.
4. Imine Complexes Few Ir-bonded imine complexes have been prepared and investigated structurally. The only complexes of nonbridging ?r-bonded C=N investi- gated structurally are a suhstitutcd ketenimine complex (XLIV), an iminium complex (XLV) , and a complex of the azine [ (CF,) 2C=N-)z (XLVI) (see Table I). The ketenimine complex is unusual in that angle y for the complex is 11.9(10)" and the average 6 is 174". These angles are indicative of a coordinatcd triple bond, rather than thc expected double bond. This scem- ingly anomalous result can bc rationalized on the basis of a large con- tribution of resonance form B to the structure of the ketenimine. This
N N *c e/ -C\/ I C 111 +N f -4, A B form would be enhanced by the highly electron-withdrawing cyano groups. The -C=N-C (CN) portion of the complex is very nearly planar, and this should enhance electron withdrawal by the two cyano groups. The deviations from linearity, a, at either end of the C-N bond are similar to thosc observed for acetylcnc complexes, thus lending additional support to the triple-bond model of the kctcnimine. The ?r-iminium complex (iV-methylsalicylaldiminato)- (N-methylsalicyl- aldiminium) nickel(0) (XLV) is the first example of a nonchelating salicyl- aldimine complex. The nitrogen atom of the C=N double bond is made more electron-withdrawing by introduction of the positive charge. Thus, the greater bending back of the nitrogen substituents relative to the carbon substituents (p = 55" and 77", respectively) is reasonable. The azine complex (XLVI), formed by the reaction of bis(trifluoro- methyl) diazomethane with Pt (0) is a rather simple imine complex. The feature of note in this complex is angle y that represents one of the greatest observed deviations from planarity (y = 127"). Unsaturated Molecule-Metal Complexes 59
IV
SUMMARY
The broad geometrical f(~aturc~sof the interaction bet\\ ccn unsaturated molecules and transition mvtal.; arc’ now well-drfincd as a result of a large number of structural studicxs. 111 a grncral way these results can be ration- alized by current, crudr bonding inodds. But if the ovcrall understanding of thr bonding of unsaturated mol(1culvs to transition mctals is to bc im- proved, additional cxpwimentnl arid thcoretical work nerds to bp done. On the exprrimcntal side, a iiumhrr of factors must be considered. The diffraction results on a givcn conipouiid could be improved considerably if thci cxpcrirncnt wcw done at low tclmperatures so that the smearing effects of thermal motion could tw minimized. As yet, very few such studies have bcen performrd. Now that marc powerful nrutron sources have been developed, thrre is a wry iniportant nwd for extensive neutron diffraction experiments oil these typcs of cwniplrxcs The scattering of neutrons docs riot follow a simple pattcrn \\ it11 atomic numbcr and, generally speaking the location of both the C or N atoms and the H atoms of an unsaturated molcculc in th(x presence of a traiisitiori metal could be made with much grc.atc.r precision using neutroiis. This is particularly true now that ligand systrms involving fewrr atoms are available (r.g., t-BuNC with 19 atoms versus PPh3 with 34 atoms). But cvm the usual diffraction studicxs at room temperature could yield valuable information through n tcmatic approach. On(. nccds systematic studirs of thc samr urisaturatd molcculc bound to an ML, system in which first M and then L is varied. Id(dly one ncclds far more studies of systems in which two or mow different ol(4iiis are bound simultaneously to the samr ML, system. Clearly such systcmatic studies are dependent on new syntheses. Although improvcmclnt of thv tlicwrctical models, especially in the direc- tion of a przori calculations, prcvrrts formidablr computational problems, thew is the grneral trend that if r~liablc,interesting, and perhaps tantaliz- ing observations are availabl(~,thcsc will serve as an incentive for theo- reticians to procwd. Finally there is a dcspcratc. need for greater corrrlations of thrsc struc- tural and theoretical studivs of thv nature of thc metal-unsaturated mole- cule interaction with other cxpcrirnentally derived quantities. We have in mind corrclations with spectroscopically derived quantitirs, such as stretching frcqucncies and Nh4R shielding parameters. But we also have in mind the most important problcm of the correlation of the metrical de- tails of the bonding with the. reaction chcmistry. If the discovery and 60 STEVEN D. ITTEL AND JAMES A. IBERS utilization of ncw catalyst systems, so essential today in view of shifting patterns of frcdstocks and of encrgy considcrations, is to be anything but empirical, thrn an undm-standing of the relation betwwn the metal-un- saturatcd molccule interaction and thc rcaction chrmistry is of paramount importance.
REFERENCES 1. J. M. Baraban and J. A. McGinnety, Inorg. Chem. 13, 2864 (1974). 2. J. M. Baraban and J. A. AZcGinnety, J. Amer. Chem. Soc. 97,4232 (1975). 3. E. Benedetti, P. Corradini, and C. Pedone, J. Organomelal. Chem. 18, 203 (1969). 4. G. Bombieri, E. Forsellini, C. Panattoni, R. Graziani, and G. Bandoli, J. Chem. SOC.A 1313 (1970). 5. I). J. Brauer and C. Kriiger, J. Organometal. Chem. 44, 307 (1972). 6. D. J. Brauer and C. Kriiger, .I. OrganonietnZ. Chem. 77, 423 (1974). 7. J. Browning and B. It. Penfold, Chem. Commun. 198 (1973). 8. J. Chatt and L. A. lhncanson, J. Chem. Soc. 2939 (1953). 9. P. T. Cheng, C. I). Cook, C. 1-1. Koo, S. C. Nyburg, and M. T. Shiomi, Acta Cr?lstaZlogr., Sect B 27, 1904 (1971). 10. P. T. Cheng and S. C. Nyburg, Can. J. Chem. 50, 912 (1972). 11. M. R. Churchill and S. A. Bezman, Znorg. Chem. 11, 2243 (1972). 12. M. 11. Churchill and S. A. Bemian, Inorg. Chem. 12, 260 (1973). 13. J. Clemens, It. E. Ihvis, 11. Green, ,J. I). Oliver, and F. G. A. Stone, Chem. Conzmun., 10% (1971). 14. C. D. Cook, K. Y. Wan, U. Gelius, K. IIamrin, G. .Johansson, E. Olsson, €1. Siegbahn, C. Nordling, and K. Sieghban, J. Amer. Chern. Soc. 93, 1904 (1971). 15. C. A. Codson and T. H. Goodwin, J. Chem. Soc. 3161 (1963). 16. It. Countryman and B. I<. Penfold, Chew. Cornnitin. 1598 (1971); J. Cryst. MoZ. Struct. 2, 281 (1972). 17. G. L. Cunningham, A. W. Boyd, R. J. Rleyerx, W. D. Gwinn, W. I. LeVan, J. Chem. Phys. 19, 676 (1951). 18. B. W. Davies and N. C. Pnyne, Inorg. Chem. 13, 1843 (1974). 19. B. W. Davies and N. C. Payne, Can. J. Chetn. 51, 3477 (1973). 20. B. W. Davies and N. C. Payne, Inorg. Chem. 13, 1848 (1974). 21. G. R. Davies, W. IIewertson, It. H. B. Mais, P. G. Owston, and C. G. Patel, J. Chem. SOC.A 1873 (1970). 22. M. I. Davis and C. 8. Speed, J. Organometal. Chern. 21, 401 (1970). 23. M. J. S. Dewar, Bull. Soc. Chim. Fr. 18, C 71 (1951). 24. It. S. Dickson and J. A. Ibers, J. Organomelnl. Chem. 36, 191 (1972). 25. It. S. Dickson and J. A. Ibers, J. Amer. Chem. Soc. 94, 2988 (1972). 26. H. Dierks and 11. Dietrich, Z. Kristullogr. 122, 1 (1965). 27. J. B. I<. Dunn, It. Jacobs, and C. J. Fritchie, J. Chem. SOC.,Dalton, Trans. 2007 (1972). 28. J. A. Evans and 1). 13. Itussell, Chem. Commun. 197 (1971). 29. K. Fischer, K. Jonas, P. Misbach, It. Stabba, and G. Wilke, Angew. Chem. Znt. Ed. 12, 943 (1973). 30. B. A. Frenz and J. A. Ibers, M.T.P. Int. Rev. Sci., Phvs. Chem. Ser. 1, 11,33 (1972). 31. J. N. Francis, A. McAdam, and J. A. Ibers, J. Organometal. Chem. 29, 131 (1971). 32. J. 0. Glanville, J. M. Stewart, and S. 0. Grim, J. Organometal. Chem. 7, 9 (1967). 33. M. I>. Click arid L. F. Dahl, J. Orga?~ometnZ.Chem. 3, 200 (1965). 34. M. Green, J. A. K. Howard, J. L. Spencer, F. G. A. Stone, private communication. Unsaturated Molecule-Metal Complexes 61
33. &I.Green, S. K. Shakshooki, :~ndIT. G. A. Stone, J. Chem. Soc. A 2828 (1971). 36. L. J. Gugyenberger, Inorg. Chern. 12, 499 (1973). 37. L. J. Guggenberger and R. Cr:inwr, J. Awicr. Chem. SOC.94, 3779 (1972). 38. W. C. Hamilton, K. A. Klandrrinitn, and 12. Spratley, Acta Crystallogr., A 25, S172 (1969). 39. F. R. Hartley, Angew. CAeni., Int. Etigl. Ed. 11, 596 (1972). 40. F. R. Harl.ley, Cheni. Reu. 69, 709 (1969). 41. 11. Hoffmann, Tetrahedron Lett. 33, 2907 (1970). 42. S. D. Itt,el and J. A. Ibers, J. Orgnnotnetcd. Chem. 57, 389 (1973). 43. S. I). Ittel and J. A. Ibers, J. Orgniiometal. Chem. 74, 121 (1974). 44. S. D. Ittel, iinpublislied results. 45. J. A. J. Jarvis, B. T. Kilbourn, :ind P. C;. Owston, Acta Crystallogr., Sect. B 27, 366 (1971). 46. RI. M.T. Khan arid A. 13. Maricll, “Homogeneous Catalysis by Metal Complexes,” Vols. I and 11. Academic Press, New York, 1974. 47. li. &I. Kirchner and J. A. Ibers, J. Ainer. Chem. Soc. 95, 1095 (1973). 48. C. Kriiger and Y. H. Tsay, J. Orgat!ometol. Chcni. 34, 387 (1972). 49. J. W. 1,auher and J. A. Ibers, tJ. Amer. Chon. SOC.,97, 561 (1975). .j0. A. lt. Luxmoore and 11. Trritei, Acta (‘rystalZogr. 15, 1117 (1962). 51. A. RlcAdam, J. N. Francis, iind .J. A. Ibers, J. Organometal. Chem. 29, 149 (1971). Ii2. RI. Matsrimoto, K. Nakatsu, K. Timi, A. Nakamura, and S. Otsuka, J. Amer. Chem. Soc. 96, 6777 (1974). 53. K. W. hluir and J. A. Ibers, J.Orguriornetal. Chem. 18, 175 (1969). 54. L. M. Muir, K. W. Muir, and ,J. A. Ibers, Z~iscuss.Faraday SOC.47, 84 (1969). 55. J. H. Nelson arid H. B. Jonassen, Coortl. Chem. Rev. 6, 27 (1971). 56. J. €I. Nelson, K. S. Wheelock, I,. C. Cusachs, and H. B. Jonassen, J. Amer. Chem. SOC.91, 7005 (1969). 57. J. G. Norman, J. Anier. Cherri. Soc. 96, 3327 (1974). 58. M. Orchin, Aditan. Catal. 16, 1 (1966). 59. L. E. Orgel, “Int,roduction to Transition-Metal Chemistry,” p. 137. hleuthen, London, 1960. 60. S. Otsuka, A. Nakaniura, Y. Tatsiino, J. Amer. Chem. Soc. 91, 6994 (1969). 61. C. Panattoni, 11. Graziani, C. Uandoli, 13. A. Clemente, U. Belluco, J. Chem. SOC.,B 371 (1970). 62. C. Pedone and A. Sirigu, Z720rg. Clheni. 7, 2614 (1968). 63. €1. W. Qiiinn and J. H. Tsai, Adu(ztL. Inory. Chem. Radiochem. 12, 217 (1969). 64. J. S. Ricci and J. A. Ibers, J,Atiicr. Chern. Sor. 93, 2391 (1971). 65. G. B. Robertson and P. 0. Whiinp, J. Organometal. Chem. 32, C69 (1971). 66. N. ltosch and It. IIoffmann, Iiiorg. Chem. 13, 2656 (1974). 67. N. Itosch, 11. P. Messmer, and K. €I. Johnson, J. Airier. Chem. Soc. 96, 3855 (1974). 68. P. van R. Scheyer and V. Bass, J. Anzer. Chem. SOC.91, 5880 (1969). 69. ,J. K. Stalick arid J. A. Ihers, J. Artier. Chem. SOC.92, 5333 (1970). 70. C. A. Tolman, J. Atner. Chein. Sor. 96, 2780 (1974). 71. T. E. Turner and J. A. Howe, J. Chern. E’hys. 24, 924 (1956). 72. 11. Ugo, Coord. Chem. Rev. 3, 319 (1968). 73. A. Walsh, Trans. Faraday SOC.45, 179 (1949). 74. K. S. Wheelock, J. 13. Nelson, I,. C. Ciisachs, and H. 73. Jonassen, J. Amer. Chem. Soc. 92, 5110 (1970). 75. D. J. Yarrow, J. A. Ibers, Y. Tatsuno, and S. Otsuka, J. Amer. Chem. SOC.95, 8590 (1973). This Page Intentionally Left Blank Methyltin Halides and Their Molecular Complexes V. S. PETROSYAN, N. S. YASHINA, and 0. A. REUTOV Chemistry Deporfment M. V. Lomonosov Moscow Sfafe University Moscow, USSR
I. Introduction . . 63 11. Methods of Study 64 A. Infrared and Raman Spectroscopy . . 64 B. Xuclear Magnetic 1teson:inre . 65 C. Nuclear Quadrupole Resonancr 66 D. hliissbauer Spect,roscopy . 66 E. I'kctron and X-Ray Diff r:ictiori 67 F. Other 124ethods . 67 111. Structures of Slethyltin Halides . 68 A4.Trimethyltin Halides . . 68 B. Dimethyltin Dihalides . 71 C. Methyltin Trihalides . 72 IV. Molecular Complexes of Methylt in Halides 76 A. Complexes of Trimethyltin Halides . 77 13. Complexes of Dimethyltin Dihalides . 84 C. Complexes of Methyltin Trilhalitlrs . . 90 V. Conclusion . . 91 References . . . 92
I INTRODUCTION
Of the various organotin compounds which have been studied (118),the most intensive research has bwn carried out on organotin halides. The synthetic and mechanistic significancc of these species was surveyed by Clark and Puddephatt (as),whose review covers the literature up to 1968. The simplest representatives, mcthyltin halides, are, in turn, the most im- portant for structural studies forming the subject of more than a hundred papers during the last 6 years. The present review aims at a detailed, up-to- date discussion of methyltin halides and their molecular complexes with organic ligands. The halides and their complexes are very promising for the study of the chemistry of organometallic and complex compounds, throw- ing light on the problem of solvent effects on the kinetics and mechanisms of organometallic reactions (110, 117). 63 64 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
METHODS OF STUDY
To study the electronic structures arid stereochemistries of mcthyltin halides and their molecular complexes, infrared and Raman spectroscopy, nuclear magnetic resonance, nuclear quadrupolc resonance, Mossbauer spectroscopy, diffraction, and other tcchniqucs arc usually employcd today. Consequently, thc application of these mrthods to the objects of interest, and the information to be gained will bc discussed in this section. The conclusions made will help us to understand the bchavior of specified compounds.
A. Infrared and Raman Spectroscopy Thc most important structural information may be obtained by analyz- ing intensities and frrquencics associated with tin-carbon, tin-halogen, and tin-ligand stretching vibrations. Let us first discuss intcnsitics. A C&i site in MeaSnX givcs rise to vibrations of two types, v, (Sn-C) and v,,(Yn--C). If v,(Sn-C) is absent or its intcrisity is low compared with that of Yas(Sn-C), thr geometry of the site can he interprcked as planar or close to planar. Similarly, the absence of vS(Sn-C) or the fact that it is markcdly less intense than is v,,(Sri-C) shows that a C2Sn site in MczSnXz is linear or close to linear. Intensity analyses of tin-halogen or tin-ligand stretching vibrations pcrmit geometries of the sites to bt assigned. The frequency variation occurring on going from one molcculc to another or to the molecular complex depends on forct constants in noncomplexcd molecules and reflects, in the final analysis, the ionic nature of the bonds. Whcii no spectra can be obtained in solution or in thc gas phase, thc intcr- molecular contributions should bc taken into account, especially in intcr- preting the Sn-X patterns. It should also he rcmcmbcrcd that Sn-X frequcncies are markedly dependent on the mass of the halogen involvcd and increase with atomic number. Neverthclcss, the frequcncics, which are a function of the naturcl and the number of halogcns and vary over a wide range, are a good reflection of thc bond ionicitics and of the coordination number. The Sn-C bond frequencics arc rather less sensitive to electronic effects in methyltin halides and depend weakly on either the number or the nature of the halogcns. The increases associated ivith complex formation ale also insignificant. Vibrational spectroscopy is a good tool not only for studying the spatial arrangemcnt and ionicities of Sn-X bonds but also for nwasur- Methyltin Halides 65 ing stability constants of the coniplcxcs (96,15, 91). The method is based on estimating the intensity due to the uncomplexed ligand as a function of the concentrations of the tnctliyltin halide and of the ligand.
B. Nuclear Magnetic Resonance The application of NMIt to a study of the structurc and complexation of tin compounds was rcportd k)y us at the 11th European Conference on Molecular Spectroscopy (114). In the present review we \vould likc to emphasize that a study of clcctronic and spatial structures of methyltin halides rcquircs, in the first plaw, a study of spin-spin coupling between “’Sn and lH and 13C in methyl groups. In 1961, Rurkc and Lautcrbur (21) showed that the constants J(llqSn-C--’H) in mcthyltin lialidcs depend significantly on both the number of halogens and thcl solvcnt uscd. Later work (66, 148), dealing with solutions of methyltin halidc in watw, postulated a linear dependence of J(llgSn-C--’H) on the s contribution of spn hybrid tin orbitals in Sn--C bonds. The dcpcndenw found sccmcd very attractive since it would allow, i)iter dial thc assignmcwt of the geometries of molecular methyltin halide complexes in solution (In, 137, G1, 24, 4, 71) by quantitatively estimating the s contribution in Sn-C bonds. However, in 1965 Vcrdonck and van der Kelen (150) studicd the proton magnetic resonance (PMR) spcctra of the ethyltin halides showing that the depcndencc postulated (66,148) should be viewed with great caution and that the variations in constants J ( llSSn--C-lH) divcrgvd from a simple correlation involving the s contribution. Similar conclusions were made by McFarlane (9?‘), who showed that J(119Sn-(”- ‘H) did vary as J(119Sn-13C) but that the straight line did not pass through the origin, and by Lorberth and Vahren- kamp (87) who studied thc spc’ctral characteristics of methyltin and cthyltin halides in detail. In 1972 we found (109)that J(lYYHg-C-lH) increased from 98.0 to 104.5 He, whercas J(lg9Hg-Cl-C-lH) (127.5 Hz) was unaffected, on going from diethylmcrcury solutions in incrt solvents to solutions in solvat- ing solvents. These findings and the fact (46) that J(lg9Hg-C- ‘H) and J ( lg9Hg-C-C-lH) behave similarly to J (lgSHg-W) and J ( lg9Hg-C- 13C) suggested (109,114) that thtw constants, analogous to other hetero- nuclear constants, depended on the relative content of s electrons in the rcspective sites (Hg-C, Hg-V-H, Hg-C-C, Hg-C-C-H) rather than on just the s contribution of sp” hybrid orbitals of the metal. That this conclusion applies to mrthyltin halidrs and their molecular complexes is substantiated by a thorough study (to be discussed in the following) of 66 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV concentration and tenipcrat urc dcpcndcnces of J ( llYS~i-C-lH) measured for the chlorides and bromides in various solvents (116). Constants J(l19Sn-W) have been studied less fully and will be dealt with when discussing the structures of individual compounds of interest. A review on ll9Sn chemical shifts has appcarc>drrcently (13/t).
C. Nuclear Quadrupole Resonance In 1973, Rryukhova, Scmin, and thc present writers (112) and, later, van der Kelen and his co-workers (149) showed that nuclear quadrupole reso~iancc(NQR) spectroscopy, especially of TI,*IBr, and 12’1 is highly applicable to the study of the structure and complex formation of mrthyltin halides in the solid phase. The most important XQ12 characteristics are resonance frequency, multiplicity of signals, and relative intensities. The frequency shifts arc due to thr alteration of thc ~lectricfield gradient on the halogen atoms, that is, to alteration of the tin-halogen bond ionicities (130). Conscqucntly, KQE frequcncics of 35Cl,81Rr, and 12’1 may provide information on the tin-halogrn bond naturr as a function of spatial and electronic factors. At thci sanie time, if two or more halogens arc present in a molecule, the resonarm’ frcqumcies may differ and thus reflect (88,89) a difference in the clwtronic cnviroiiment (chrmical nonequivalcnce) or a difference in positions of the nuclei in the crystal lattice (crystallographic nonequivalencc) . Crystallographic splittings [ ( Av/vaV) - l00%] are, as a rule, of about 2-3Cj,, J+ hcreas the chcmical nonrquivalcnce may be sig- nificantly greatcr (130). Analysis of the splittirigs and relative intrnsities often allows one to depict rtlliably the spatial arrangement in solid methyl- tin halides arid their molecular complexcs (112).
D. Mossbauer Spectroscopy More than 10 years ago, Gol’danskii (55) had shown how promising lL9SnMossbauer spectroscopy is for thc study of tin compounds. The data obtained up to 1971 for various organotin compounds have been surveyed by Smith (132), Zuckernian (l54),Parish (101), and Bancroft and Platt (1). Mossbauer spectroscopy of monoorganotin (IV) derivatives has been reviewed very recently (7). These data, as analyzed in the light of the recent results (119, 51, 8, 83, 33, 17, 120, 84, 58, 113, 128, 63, 70, 107, 108, 121, 2, S), demonstrate that the principal Mossbaucr parameters, isomer shift (IS) and quadrupole splitting (QS) , arc highly informative of chcmical structurr and coordina- tion. For example, IS is a measurc’ of the total electron density at a tin Methyltin Halides 67 atom (53) and depends mainly oti thc. population of the valciice shell, that is, it reflects those altwations of thv structure affecting the s-electron dcnsity (82). The alterations ar(' chic4ly duc. to alterations of tin-atom or till-ligand bond ionicities (102), although a contribution to the s-cllectron dcmity may sometimes 1)~provitlod by the shielding associated with d orbitals (53). Itcgardlcss of th(5 mc~lianisirithat governs the IS behavior, the IS pattern in any methyltiti halid(. coniplcxcd with inonodcritate elcc- tron-donor ligands reflects charge transfcr, i.c., the rclative donor ability of tlic ligands (113). Kccciit data (113, 3) show that QS valucs depcnd, first of all, on the tin coordination number in thc coinpounds and the stcreocheniistry. They also dcpcnd on thc nature of thligsrid in isostructural molecules. This picture, illustrated in original papers (64,104) arid survcycd in reviews (1, 101), gives a clue to assigriiiig spatial arrangements of ligands, L, in complexrs of thc type RSnX3- ILL (113), ftzSnXz.2L (1IS), and R3SnX- L (3) as wcll as to coiistructing donor ability series for ligands in thc com- pounds.
E. Electron and X-Ray Diffraction Diffraction methods, which arc hcyoird doubt the most informative ap- proach, arr at the same timc. thc inost cumbersome. It should also bc stressed that, although this approach may allow the complete structure of a molecule to br given in thc gas or solid phase, in practice nonvolatility, instability, and difficulties inhcrent in crystal growth, of the samples may interfere significantly. Wherc, howvcr, thcsc difficulties arc ovcrconie the results obtained are profitable nhatcw~thr effort invested. Reccntly, Ho and Zuckerinair Iiavc. published an extensive rwiew (65) on structural organotin chemistry and discussed the experimental cvidcncc accumu1atc.d up to 1971. Thr niiml)t.r of papers devoted to organotin structures has incrrascd (13, 19, 20, 23, 59, 60, 74, 125, 127) lately, the increase being markedly favorrd by progrrss in electronic and X-ray hard- ware. Thc bond lengths and anglcs obtained are very important for corrc- latioiis nith the spectral cvidcnw (57).Such correlations may make other spectral cvidencc quite reliable (wm iri thr absence of the rcqectivc diffrac- tion data.
F. Other Methods Other approaches include UV sprctroscopy (151, 73,92,93, 133),photo- electron spcctroscopy (106, 55, 18, II), conductometry (78, 79, 140, 141, 68 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
47, 24), thermochemistry (14, 15), and dipole moment measurements (144, 67, 56). The works cited have demonstrated (54, 122), however, that these methods are markedly less applicable.
111 STRUCTURES OF METHYLTIN HALIDES
In this section, structures in the gas phase, in solution, and in the solid state are discussed. Let us first give a classification of the solvents used for studying structures in solutions. These may be divided into (i) poorly solvating and nonionizing solvents (cyclohcxane, CC14, benzene, chloro- form, dichloromethane, nitromethane, nitrobenzene, and others), (ii) strongly solvating and nonionizing solvents [pyridine, acetone, tetrahydro- furan, dioxan, N,N-dimethylformamide (DMF), dimethyl sulfoxide (DMSO) , hexamrtapol (HMPT), and others], and (iii) strongly solvating and ionizing solvents such as water, alcohols, amines, and others. Methyltin halides dissolved in solvents of type (i) do not enter into a noticeable donor-acceptor interaction with the solvent; the tin coordina- tion number is, therefore, unaffected. At the same time, conductivity techniques (78, 79, 140, 141, 47, 24) show that methyltin halidrs do not ionize in solvents such as nitromethane and nitrobenzene whose dielectric permeabilities are high (35.9 and 34.8, respectively). In this section the halide structures are discussed only in solvents of type (i). When they are dissolved in solvents of type (ii) , the halides enter into complex formation; these systems are dealt with under complexation of methyltin halides (see Section IV) . Organotin cations formed through dissolving the halides in solvents of type (iii) are not discussed in this review. Their recent study has been rather limited, but they have been reviewed by Tobias (142).
A. Trimethyltin Halides
1. Me3SnF In the gas phase, no structure is known for MesSnF. In the solid state, IR spectroscopy suggested (100, 80) a structure of the Me3Sn+F- type. X-Ray studies (26,152) demonstrate, however, that orthorhombic MesSnF crystals have a covalent structure in which fluorine atoms alternate with Methyltin Halides 69
Me& groups in an infinite chain, the tin being five-coordinate and ap- proximately trigonal-bipyramidal. Thcrc are two sorts of fluorines: Sn-17, 2.15 8 and Sn-- -F, 2.45 8. It is bcdiwcd (26,49) that the elt1ctron density distribution data may be intcrprrtrd in tc~nisof either planar or pyramidal structures for the SnMea fragment, but a weak band found at 513 cm-I is thought to indicate that the structure is pyramidal. The planarity, how- ever, has been proved unambiguously in studies (44, 136, 83) of S11-c stretching vibrations in the 111 atid Raman spcctra. Thc Mossbaurr param- eters [S, 1.28 mm/sec; A, 3.86 mm/scc (62, 34, 103)] agrce with the fivc- coordinate structure in terms of corrdation (62) or point charge (104) approaches. The Debye-Wrllrr facxtor as a function of tcniprraturc (15'5) also suggests that Me3SnF is a strongly bonded polymer. The compound is irisolublr in inrrt organic solvrnts, hrrice no structure in solution has been obtained. As for thc PMR parameters found in meth- anol solution (87),thc peculiaritivs of mrthanol (sre above) suggest that the paramcters [S('H), 0.45 ppni; J(lH-C-llSSn), 69.0 Hz] hear no rcla- tion to the structure of hle~SnF.
2. MeZSnX (X = CZ, BY,I) Elrctrori diffraction showd (131) that in the gas phasr the compounds are distorted tctrahedra with tlw Sn-( ' and Sn-X bond lengths roughly corresponding to the sums of thc covalmt radii: Mv3SnC1 (Sii-C, 2.11 8; Sn-Cl, 2.36 A), rvlc3SnBr (Sn-C, 2.17 8; Sn-Br, 2.49 A), Mc3SnI (Sn-I, 2.72 8).No X-ray studicls have been publishcd. 1nfrarc.d (Table I),NQR (112,149), and Mdsshnucr s1wctroscopy data (Tablc 11) demon- strate that the solid compomitls nrr associates containing fivr-coordinate tin, the associates bcing dcconiposd on mrlting or on dissolving in inclrt solvmts. The increase in v (Sn-X) across thc scrics solid-melt-solution suggmts that thc covalcncch of the bond increasrs across the same series, i.e., with a decrease in association. The fact that v(Sn--CI) varies morc than v(Sn-Br) shows (SO) that Mc3SnC1is associated mow strongly than is MesSnHr, nhcreas the constnncv of v(Sn-I) suggests that there is no significant association in M($hT. That association decreases in the scrips Mc&iF >> Mr3SnC1 > Mr3SnIJr > hfe3SIiI agrccs also with thr )loss- bauw parametrrs (Table 11) arid cqwially with the quadrupole splittings. When trimethyltin chloride is dissolved in inert solvrnts, v (Sn-(21) r('- mains essentially invariant: C'('Ii, 330 cn-1 (8, 9) ; benzrnc, 331 en-' (S, 9) ; cyclohcxane, 331 cm-I (32); carbon disulfidc, 331 cm-I (80). Com- pound MtGnC1 may, therefore, be assumed to be a tetrahedral moiionier in these solvcnts. A similar conclusion was arrivcd at in a study of thc concentration de- TABLE I TIN-HALOGEXASD TIN-CARBONSTRETCHING FREQL-ENCIES OF MeaSnS IN THE SOLIDAAD LIQL-ID STATESASI) IU SOLUTION"
MeaSnF MesSnCl RleaSnBr MesSnI
Solution z Solution Solution in in cyclo- v, Solid Solid Liquid in CCL Solid Liquid cyclohexane Solid hexane Vibration IR It IR R Ilt R IR It IR R R IR It IR R IR
~ ~~ v(Sn-X) 335vs ~ - 288 325 315 336 331 - 199 219 234 229 - 177 189 vS(Sn-C) - 521s 514w 513 514 514 513m 514 512m 512 512 511m 512 509 511 - v,,(Sn-C) 555s 559w 543vs 546 545 545 543s 544 54lvs 545 543 539s 541 540 538 P - ?
a Values are expressed in cm-l. Methyltin Halides 71
6 (mm/sec) 1 28 1 42 1.49 1 48 A (mm/sec*) 3 80 3 41 3.25 3 05
pcndcncvs of thr %n chemical shifts (69, 134, 149) and J(lH-C-llgS n) constants (116) in nonpolar solvcmts. The NMIt data (45, 25, 126, 87, 99, 98, 146, 147) listed in Tablc 111 for trimethyltin halides in CCll show that the most informativc arc' the J ( lH--C'--"gSn) values which increase slightly ii ith thc clcctroncgativity of thr halogen. Their behavior, which reflects thc gro\\th of thc s contribution in the Sn-C-H site, fits well with thc Bent rehybridization thcory (10). On the othcr hand, th(. fact that thc J(lH-C-ilgSn) valuvs cvnbrace a rather small range, whercas thcl J (iH-ldC) couplings arc' constant across thc series Mc3SnCl > Me3SnBr > M(>#%iI,points out that rehybridization of tin orbitals is rather insignificant in this sckric>\.
B. Dimethyltin Dihalides Elcctron diffraction tmhniqncs showcd (131, 50) that in the gas phase McaSnXz (X = CI, Br, I) arc soincwhat distorted trtrahedra: MczSnClz (LCSnCl, 109.5"; LClSriCI, 107.5"; Sn-C, 2.11 A; Sn-CI, 2.33 A), MczSnBrz (Sn-C", 2.17 A; Sn-Br, 2.45 A), MczSnIz (53-1, 2.69 A). X-Ray analysis of crystallinc~hZ(~zSnFz showed that (129) in thc solid stat(, the compound is an associate containing six-coordinate tin. The structure is a two-dimcnsional infinitc net in which cvery tin atom is
6(CH,) 6("9Sn) ./(~L~SII-C;-'H) J(ll9Sn-l3C) J(X-1II) lTeoSnS (ppin) (pprn) (frZ) (lfz) (HZ)
lIeaSriCI 0.61 158.0 58.1 386 131 .6 MerSnBr. 0.73 128.0 57.8 372 131.8
MeuSnI 0.88 38.6 57.2 - 132.1 72 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV bonded to 4 other tiri atoms by fluorines situated symmetrically in between. Methyl groups lie under and above the resulting plane and, thus, complcte thr octahcdral structure. The Sn-C distance of 2.08 f 0.01 A is the short& knon n in organotin compounds and mas ascribed (129) to ionicity of the (SnF2) site. Irifrarcd spectra of the compound have shown, how- ever, that v,(Sn-C') cannot be detected, whereas v(Sn-l?) is observed. Thus, covalcncc1 of the octahedral po1yInc.r ith trans-methyl groups is bcyond doubt (Table IV), (86, 22, 61, 83). The structure fits ell with the Mosshauer data (Table V) processed via a correlation (62) or a point- chargc mrthod leading to A values for the trans-Ii2SnLe structures (48). The IIt data of Tabl(. Ic' and the Mossbaucr data of Table V show that the association falls across thc. scrics M(+hiF2 > MczSnClz > MczSnBr2 > Me2Sn12 (105). A sharp differcnce bctn-(.en the Ric2SnF2structure and the structures of the other dihalidcs is vcxrificd by an X-ray study of MezSnC12 (41). The tin environment 11 as shown to be intc.rmediatc between a tetrahedron and an octahedron, ou irig to association of adjacmt molecules through Sn- (3...Sn bridges. The structure consists of molecular chains, with tin and chlorine atoms h~ingcoplanar in vach of thr chains. Mcthyl groups lie under and ahov(1the. plane. 'I'hc cliain has a zigzag shape. The bond lengths and anglrs arc listed in Tablc VI. The distorted octahrdron of ;1!I(+hiC12 agrcw also with the hiossbaucr quadrupolc splitting found for the compound (Table V) , lying between the tc.trahcdra1 (2.3 mni/scc) and octahrdral (4.1 mm/sec) valurs. Nuclear quadrupole r(wmancc 79Br arid *IBr spectra of solid Me2SnBrz display (119) n cakly split doublcts which suggest that u-electron density is equal in both the broinirics accurately to within the crystallographic splitting. The associatcb assumed on the basis of Mossbauer data (105) is thus not a fivr-coordinate bpccies, because othcrwisc a doublet of consider- able splitting, with onc~of the components corresponding to a bridgc bro- mine and the othor to a tcmiinal hrominc, would haw arisen. Thc "IR parameters in Table VII, especially the J(lH-C-llgS n) couplings, imply that thc tiri rtihybridization in thc halides is significant and drpcndcnt on thc halogcm. As in the trimcthyltin halides, the s-electron contcmt in the Sn-C-H sit(. incrcasm across the series I < Br < C1.
C. Methyltin Trihalides Electron diffraction showd (131) that, in the gas phase the trihalides have a slightly distorted tetrahedral structure: McSnC13 (Sn-C, 2.19 A; Sn-Cl, 2.32 A), 1CI(C3nBr3 (Sn-Br, 2.45 A), MeSn13 (Sn-I, 2.68 A). No Solution Solution Solution Solution Solution Z in in in in in sJ hen- ryelo- hen- cyrlo- cyclo- z. Solid Solid Liquid zene hexane Sofid zcne hexane Solid hesane -i3 VibriLt,ion IR rc IT{. R. IR 111 I II 1It I11 1I<, 11j 11 Llt I -.n Q V*(STl.-X) 3rin - 307 344 320 350 :mi - - 240 - 1x2 IXG 3 Vns(Sn-X) 360 - 332 344 320 350 361 - - 2% - 197 204 .,(SIl-C) - 536 515 531 515 521 524 514 518 - 511 513 - v,,(Sn-C) 5% - 567 566 566 559 560 563 554 - 547 54-1 -
a Values are expressed in crn-1. 74 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
crystal structures have been reported (ware studying them now in col- laboration with L. A. Aslanov), but a natural assumption is that the greater Lwis acidity of the trihalidrs (compared with MezSnXz) will cause the molecules to associate in thcx solid phase. For MeSnF3, this as- sumption agrees with the III spectrum (Table VIII) and with the tempera- turc dr.pcndence of the 1Cldssbauc.r spectrum n hich contains a well-resolved doublet at 298°K. The rwulting assigrinicnt (83) leads to a polymeric structurc containing six-coordinate tin atoms, two bridging fluorines, and one terminal fluorine. It also agrees with the rclative intensitics of the IR signals Y (Sn-F) and v (Sn-F) b. With MeSnCI3 and MeSnBr3, the six-coordinate association in the solid statc is verificd by 3iCl, 79Br,aid *lBr NQIl data (112).The spectra con- tain strongly split doublcts, with the intensity ratio bhg2: 1 for th(1 low- and high-frequency componcmts. ('onsequcntly, there are tu o bridging halogens and one tu-rninal halogen. That the association is stronger in MeSnCls than in MoSnBr3 is suggested by the temperature dependence of the NQR spectra and by a comparison of the Mossbauer spectra of the compounds with those of thc molecular complexes containing various elec- t,ron-donor ligands (113). The isomer shifts demonstrate that McSnC'la (6, 1.32 mm/secj is a stroiigcr Lewis acid than is MeSnBrs (6, 1.44 mm/ sec) . As with MezSnXz,the J(lH--C--"'Sn) values (116) and the 6(l19Sn) chemical shifts (134) found for MeSnXBdissolvcd in solvents such as bcn- zene or dichloromethane drpcnd neither on conccntration nor on tempera-
Bond Lcngth (I) Fragment .\ngle
Sll-C 2.21 CI-S~l-Cl 93'0' SII-Cl 2.40 C-Sn--C 123"30' Sn- - .C1 3.64 C-Sn-C1 109%' Methyltin Halides 75
T.kl3IJl~; VII xI.C'LEAR hlAGSETIC ItESONASi('E 1'ARAMETERS FOR SOL1.TIOSS OF Rle?SrlS2 IN CCl4
6 (C€I,) 6 ("9Rn) J (119Sn-C--'II) J ('T--LII) SlesSnSs (pptn) (ppm) (Rz) (I&)
RIe2SnC12 1.15 140 69.0 136.2 31rzSnRr2 1.33 70 66.3 136.6 MezSn12 1.63 -157 62.4 136.8 turc. This may be intcrpretcd tiy assuming that there is no significant association in the systems. Molten MeSnCls has 6(11%n) of 6.0 ppm (42),markedly different from tlir value found for thc bcnzmc. solution (Table IX) . Consequently, there is somc association in thc mdt, which increases on going to the solid state. Nuclrar magnetic resonanccl param(tcw, especially the J ( 1H-C--"9Sn) valucs, found for solutions of RlrSnXa in CCl, shorn that the s-electron contcnt in thc Sn-C-H sit(. is higher than in M&nX or MezSnXz,and dt.pends markedly on the naturc of the halogen, in agreement with the Bent theory (10). On summarizing thc structurw of mcthyltin halides (55,36), we may say that, in the serics Mc3SnX < Rlr$hXz < MeSnX3, on the one hand, the increase in the NMR constants J ( lH-C--"gSn) reflect an increase in the s-clcctron contmt in thc Sri-C-H site and, on the othcr hand, the increases in the IR v (Sn-Hal) frcqucncics and the NQR Y (Hal) frcquen-
Solution in cyclo- Solution in Solution in Vibration Solid Solid Liquid liesane ryclohexarie cyclohexane
~(SII-C) 548~s 535s 542 546 551 539m 527w v(Sri-X) 646vs 629s 425vs 384 360s 382vs 264vs 2351n 207vs 174w 76 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
TABIJE IX NUCLEAR MAGNETICRESONANCE PARAMETERS FOR SOLUTIOXS OF'hleSn>is IN CCla-
G(CH3) 6(Il9Sn) J(llgSn-C--IH) J(11gSn-13C) J(13C-lH) MeSnXt (ppm) (PP) OIZ) (Hz) [Hz(in C6H6)]
MeSnCls 1.69 21 100.0 - 141.2 hleSnBr3 1.85 - 1G5 88.6 - 640 141.2 (in CsH6) MeSnIa 2.32 - GOO 73.4 - 141.2 cies both reflect a decrease in the Sn-X bond polarity. These experi- mental data draw attention to a quantum chemistry calculation (56) of electric dipole moments and u and T charges in Mc,SnCL,. The calcula- tion has shown that thc Sn-C1 bond polarity decreases across the series Me3SnC1 < MezSnClz < MeSnC13 (Table X) and that the Sn--C polarity falls in the same direction. Sign inversion of the dipole occurs in MeSnCls which is why the MeSnC13 dipole moment is greater than that of Me3SnC1. Thus, the methods discussed in the forcgoing allow all the necessary in- formation on the structures of methyltin halides in the gas phase (electron diffraction), in the crystal state (X-ray techniques, NQR, Mossbauer spectroscopy, IR spectroscopy), and in solution (NMR and IR methods) to be obtained.
IV MOLECULAR COMPLEXES OF METHYLTIN HALIDES
The data reported in the literature on molecular complexes of trimethyl- tin halides demonstrate that these have a 1:l composition with mono-
TABLE X DIPOLEMOMENTS AKU BONDPOLARITIES I& hZle,SnCL, (n = 1-3)
Polarity (yo)
Me,SnClr-, Calrulated Observed Sn-Cl Sn-C C-H
MerSnCl 3.46 3.46-3.52 38.68 10.66 3.30 MezSnClz 4.04 4.14-4.21 33.70 4.48 3.64 MeSnCla 3.63 3.62-3.77 28.07 -2.77 4.05 Methyltin Halides 77 dentate ligands and a 2: 1 composition with bidentate ligands. On the other hand, dimethyltin dihalides form, as a rule, 1:2 complexes with mono- dentate ligands and 1 :1 complexes with bidentate ligands. Under certain conditions, however, 1: 1 complexcs with monodentate ligands may be iso- lated for Me2SnX.L and MeSnX3; their formation in solution may also be possible. In the following discussion, 'we denote the number of organotin molecules by the first figure and thc number of coordinated donor centers by the second figure, regardless of whether the ligand is mono- or bidentate. Molecular complexes of methyltin halides may exist in various stereoiso- meric forms whose types are listed in Table XI together with the notation to be employed below. It is also noteworthy that there are many complexes whose isolation in the individual state is impossible. However, methods such as NQR or Mossbaucr spectroscopy allow one to study the complexes in frozen solutions. As for the spectral parameters [c.g., v(Sn-X) or J(IH-C-ll9S n)l of methyltin halides dissolved in clectron-donor solvents, they should be tcsted in each individual case to scc whether thcy correspond to nondis- sociated complexes or to an equilibrium between the complex and the uncomplexed halide. Together with stereochemical problcms, the ways in which the Sn-C and Sn-X bonds are affected on going from a methyltin halide to its molecular complex are of interest. Such information is quite helpful in interpreting the reactivities of mcthyltin halides in various solvents (110, 117). More than twenty molecular complexes of mcthyltin halides with elec- tron-donor solvents (111) were rtlported in 1973 and their structures were studied by NQR (112), Mossbauer (llS),and NMR (116) techniques. These and other data will be discussed under individual types of the complexes.
A. Complexes of Trimethyltin Halides In 1963, Gielen and Nasielski (52) studied the NMR spectra of tri- methyltin bromide in various solvents and showed that J (1H-C--"9S n) increased across the series CCl4 < MeCOOH < dioxan < acetone < MeOH < pyridine < water < DMSO < DMF. Simultaneously, Beattie and McQuillan (8)found that v(Sn-Cl) in the IR spectra of trimethyltin chloride is also solvent-dependent and increases as follows: CCl, < ben- zene < acetonitrile < pyridinc. In both cases the data were interpreted as indicating formation of 1: 1 complexes. Finally, Hulme (68) found by 78 V. S. PETROSYAN, N. S. YASHINA, AND 0.A. REUTOV
TABLE XI THE BASICTYPES OF STRbCTVRES FOR hfOLECULAR COMPLEXES OF ~ZETHYLTIS TIALIDES
Me s Me s 1, Me S Me X L hle x \I \I \I/ \I \I/ Sl1-hle Sn-L s11 Sn-1, s I1 /I /I /I\ /I /I\ lle I, 3le x I> IIe x x x L x x (Val
hle hIe Me X x Me I’ s x 1, Me X \I \I \I/ \I \I/ Sn--X Sn--X Sn Sn-Me Sn /I\ x x 1, 0-b)
kle >; Y Me 1, Me Me S Me S 3le X I/ \I \I/ \I \I/ Ale-Sn Sn-I, s n Sn-X Sn I\ /I /I\ /I /I\ hZe I, S Ile L s X X I, LLX (Ic) (IIC) (II I c) (IVC)
lle Me Me 1le I, hle hle ?i X I/ \I \I/ \I Ale-Sn Sn-L s11 Sn-L I\ /I /I\ /I XL sx S S I, S Me (Id) UId) (Illd) (IVd)
S XIe X hle \Ie \I \I/ Sh-Me Sn
X-ray methods that the trimethyltin chloride-pyridine complex has a 1: 1 composition. It is a molecular adduct of trigonal bipyramidal structure, with three methyls lying on the equatorial plane and with the pyridine and a chlorine atom at the axial positions. The Sn-CI distance found in the complex exceeds that in the initial halide. Methyltin Halides 79
Y~,.(811-C) us(Sn-C) v(Sri-X) Compound (c.nP) (em-’) (en-’)
These pioneer works were follon-c3d by numerous studies on complexes of trimethyltin halides, the data froni which are discussed in the following.
1. Coin pleaes with Pyridine and LY-Oxopyridines The preceding results for thc trimcthyltin chloridc-pyridine complex (68) fit well with IR (8,9)and RIossbaucr data (1,101).Thc u,(Sn-C) : v,,(Sn-C) intensities ratio (Tablti XII) suggests a planar arrangement of the methyl groups, whereas u(Sn--Cl) suggests that the bond polarity is markrdly higher than in uncomplcxcd R/Ie3SnCl. The chloride, bromidcb, and iotlidc complexes probably possess similar structures sincc all three havc idcntical Sii-C stretching frequency ranges (Table XII). An NQR 81Rrspectrum of MeXSnBr-Py points to a noticc- able crystallographic nonc~pival(mcein the crystal cell and to a higher Sn-Br bond ionicity comparc.d with McXSnBr. The PMR spectra of McsSnX - I’y complexes dissolved in CHC13 implied that the compounds dissociated completely (145) in solution. Wc showed (116),however, in a study of thr concentration and tcmperaturc depend- ence of J (lH-C-llgSn) in n4esSnX-pyridine-CHzClz mixtures, that the 80 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV equilibrium MeqSnX + Py z====MeaSnX-Py depends on the pyridine-to-CHzCl2 ratio and on the sample temperature, and is totally shifted toward the complex at higher pyridine concentra- tions and lower temperatures. The limiting J(lH-C--”gSn) values ob- served (Table XIII) in the system reflect the s-clectron density content in the Sn-C-H site and allow one to deduce a series for the electron-donor ability of coordinating solvcnts. Pyridine occurs between weak and strong electron donors. Complexes of the type
(X = C1, Br; X’ = H, CH3, OCH3, C1, NOz) were studied both in the solid state and in solution (73, 76). Their Y(N- 0) frequencies are lower than in the free ligands, implying coordination at the oxygen. The Y (Sn-X) frequencies (Table XII) show that complex formation raises the bond polarity.
J (llSSn-C--’H) (Hz)
L in ILleaSnX-La MesSnCl MeaSnRr
MeCN 66.7 65.1 THF 64.5 - Acetone 66.1 66.0 DMTAA 65.7 - l’yridine 68.0 67.6 DMAA 68.8 TM15U 69.7 68.5 I1 M F 70.0 69.6 DMSO 70.1 69.6 kI M 1’T 71.8 71.3
Abbreviations: THF, tetrahydrofuran; DlIF, N,N-dimethylforinamide; DMSO, dirnethylsulfoxide; HMPT, hrxametapol; DMTAA, diniethylthioacetarnide; DMAA, di- rnethylacetamide; TMED, tetramethylethyl- enediamine. Methyltin Halides 81
Spectrophotomctric data for the stability constants of complexes of Mc3SnCl with substituted pyridinc. N-oxides in acetonitrile show (76) that these constants correlate linearly with the n parameters of the y sub- stituent s. No V,(Sn-C) was found in McSnBr-PyO. It was, therefore, inter- preted (38) in terms of planarity of the C3Sn site, namely, structure Ia (Table XI).
2. Complexes with Phosphine or Arsine Oxides
Conductivities of the complexes in absolute ethanol were shown (38) to be very low when compared ~ithtcbtramethylammonium bromidc as a reference. This finding was intorprc+d to indicate that molecular adducts exist in solution. The absence of v,(Sn-C) from the IR spectra and the J(lH-C-llSSn) magnitude (ca. 70 Hz) were interpreted in terms of planarity of thc C3Sn site and a noticeable rehybridization of the tin orbitals.
3. Complexes with Sulfoxicles Trimcthyltin chloride and bromide form DMSO complexes of 1: 1 com- position, melting at 49°C (73) and 63°C (112) respectively. The dccrease in v(S=O) (Av is 95 em-' for R4v3SnC1.DMS0, and 45 cm-' for MesSnBr. BzzSO) may show (38,77) that the coordination is via the oxygen, whereas the absence of v,(Sn-C) from the complexes of Me3SnBr (Table XII) suggests a structure of type Ia. As the NQR *lBr frequency is s1iiftc.d on going from Me3SnBr to its com- plex with DMSO (IIg),a strong charge transfer and an increase in the Sn-Br bond polarity may bc assumed. Studies of the concentration and tctmpcrature dcpendences of mixtures of DMSO and Me3SnX (X = ('1, Rr) in CH2Clz gave limiting values for the J ('H-C-llgSn) constants (Table XIII) and the MeSSnX. DMSO stability constants (Table XIV). Thcsc~results suggcst that the values of 62.0 and 63.0 Hz, obtained csrlicr for CHC13 solutions of Me3SnBr.DMSO (38) and Mc3SnBr-BzzS0(77), rc,spcctivcly, are nothing but the equi- librium values and do riot reflect structural specificity of the complexes formed. The J ('H-C-llgSn) valucis found for Me3SnX solutions in DMSO (4) exceed our va1uc.s (Table XIII) by 1.8 and 2.4 Hz, but the couplings reported (4) for Rlr3SnX solutions in CDC1, or for other methyl- tin halides are also 1.5 to 3.8 Hz grclatcr than the rrspective couplings re- ported elscwherc (66, 87, 116, 144). 82 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
TABLE XIV STABILITYCONSTANTS FOR Me3SnX.L COMPLEXES
KP AH" (molcs/ (kcal/ Complexa Solvent LZethod t("C) liter) mole)
Me3SnC1.acetone acetone HNDMR 113-{119sn) +37 1.1 4.2 Acetone N Mll (ll9Sn) +20 7.0 - Isooctane Calorimetry +26 0.9 -6.0 CCla PhfR +26 0.4 -5.7 CH,Cl2 PMR - 30 0.8 - Me&& .acetone Acetone NMR(lIgS11) +20 3.0 - CFI,Cl, rim - 30 0.7 - MenSnCl. MeCN CCl, 1'MR +26 0.5 -48 MeCN NMR(1LgSn) +20 2.7 - MeCN HNDMR 'H- {119Sn] +20 2.1 4.1 Me3SnBr - MeCY MeCN NRfR(llgSn) 120 3.5 - MesSnCl -DMTAA CCl4 PRlR +26 0.5 -5.9 CH,C12 PMR - 70 1.5 - MesSnCl - Py cc1, IR +27 1.8 -6.5 CII &I2 PMR - 30 36.0 -
MesSnBr - Py CHZC12 PhlR - 30 28.0 - Me3SnCl.DMAA CCl, IR $27 3.7 -7.9 cC14 PhlR +34 3.1 -7.9 Me3SnC1-DMS0 cC14 Calorimetry +26 9.1 -8.2 CCl4 IR $27 8.3 - CCl* PMR +34 6.2 - CH zC12 PMR +28 2.8 - MeaSnBr-DMSO CH 2C12 PMR +28 3.6 - MeaSnCl.HMPT Isooctane PMR +26 384 -10.1 CHzCI, PMlt +28 102 - MerSnBr-HMPT CHzC12 I'M R +28 99 - MerSnCl-dioxan Dioxan HNDMR l€l-(IlsSn) +20 2.1 5.1 MesSnCl .DMF CHsC12 PMR -30 2.5 - Me3SnBr .DMF CHZC12 PMR -30 3.1 -
Abbreviations: DMSO, dimethylsulfoxide; HMPT, hexametapol; DMF, N,N- dimethylformamide; DMAA, dimethylacetamide; DMTAA, dimethylthioacetamide; HNDMR, heteronuclear double magnetic resonance. Methyltin Halides 83
The data obtained, when :malyxcd as a whole, show that Mc3SnBr- DMSO may be the strongest cortiplcx among molecular coniplexcs of tri- methyltin halides with monodrntat r ligands.
4. Other Co,nplP.ces The molecular complexes of nirthylt in halides with various electron- donating solvents haw hccn givm a grmt deal of attention by a number of workers. Drago et al. (96, 14-16) eniplo~rd calorimetry arid I11 and PMR meth- ods to mrasurr the stability const:mts and licats of formation of McBnCI complexcs with dinicthgl sulfoxidc, diintxthylacctamidc, DMF, acetonitrile, pgridine, and HMPT (see T:hlc Xl\'). Okavara and co-workers (91),\vho studied solvent effects on the PMR spectra, and on the ,,(Sn--C) :Vcls(Sn--<') intensities ratio in the 113 spec- tra, of Mc3SnCl in various solvents concluded that the coordination ability of the solvents niay bc reprc~scntrd by the series DMSO - DMF - DMAA > pyridine > acetoric > dioxan > MeCN > PhCN. We isolated (111) molcculnr complexes of Me3SnBr, studied their struc- ture by NQR in the solid stat (> (112),and showed that complcxrs of type Ia arc formed it ith monodcntat (, rlcctron-donor solvmts, whereas com- plexes with bidcntate solvents ni:ty havc structure Ib. The latter assump- tion agrees with the fact that thc IiniitingJ (lH-C-llgS~i) values measured in dioxan and DME at differrnt tcniptmtures are different; they may re- flect the relative percentages of Ia and It) complexes in solution. The 81BrNQR frequencies of thr coniplcxcs lead (112) to the following series of electron-donor ability nith respclct to hlcaSnBr: DEE < dioxan < acetone < DRlE < THF < IIMF < HMPT < pyridinc < DMSO < TMED. This serirs fits, on tlic wholc, with the one based on NMR data obtained for Me3SnBr iri thv s:+iiw solvrnts (116). The strongest electron donor is HMPT, hich givrs n inolwular coniplcx cvcn with Mc3SnCF3 (114, 115), in contrast to thc. commonly hcld belief that organotin com- pounds containing four C-Sn honds cannot form molecular coinplexes (4). An X-ray study showed (20) that Mc3SnC1.Ph3P=CH-COCH3 has a Ia type structure, with the oxygw atom in the srcond axial position. In conc1usion, trimethyltin halides give Ia comp1cxc.s with all mono- dentate ligands. All three 611--C: bonds arr coplanar and contain more s-electron density than does thr uncomplcxrd McaSnX. The donor atom of the coordinating ligand, and thr halogen, occupy axial positions in the structure, with thc Sn-X bond in the complex being more polar than in the frce halide. The s-t+.xtrori content in the Sn--C bond and the Sn-X bond polarity increase with the ckctron-donor ability of the ligand. 84 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
B. Complexes of Dimethyltin Dihalides The complcxcs of dimethyltin dihalides were readily studied since they may be obtained easily, owing to thc high acceptor ability of the starting halides, and may be purified by simple rccrystallization or sublimation.
1. Complexes with Xulfozides Many sulfoxide complexes, mainly with MczSnC12, have been reported. The simplest, MezSriClz.2MezS0, was dcscribed for the first time by Langer and Blut (81) and studied in great detail by various physical methods. Infrared spectroscopy led to the conclusion (27) that the two methyls (as well as the two chlorines) wcw trans-arranged (structure IIIb in Table XI), but more sophisticated IR (84, 85) and Raman (1.23, 124-6) studies of MezSnCl2complexes with sulfoxides have shown that the Tanaka (137) IIIa structure is correct. The trans-arrangemcnt of the methyl groups agrees also with the Mossbaucr quadrupolc splitting data (4.40 mm/scc) (40,84, 113) and with an X-ray study (153) that revealed a slightly dis- torted octahedron with the atom-atom distances listed in Table XV. The J(lH-C--"gSn) valucs (113.0-117.1 Hz) (75, 91, 116) obtained for the complex in excess DMSO suggest that the trans-methyl configura- tion is rctained in solution and reflect a strong increase in the s-electron content in the Sn-C-H site. As for J(lH-C--"gSn), equal to S6.0 Hz for the solution in CHCL (58), it reflects nothing but an equilibrium bctmcn the complexcd and uncom- plexed MczSnClz rnolccules. The absence of v,(Sn-C) from the 111 spectrum of MezSnBrz-2DMS0, and a high value (430 cm-l) for v(Sri-0) suggested structure IIIb for the complex (137).This agrees with the SIBr NQR singlet (112) which,
Sn-CI1 2.53 C11--Sn-C12 96" Sn--C12 2.48 CI1--Sn-O1 92" Sn--0 I 2.32 (I1 1-Sn-0 2 175" Sn-Os 2.38 c11-Sn-cI 90" Sr1-C 1 2.08 CII-Sn--Cx 94" Sn-Cp 2.07 C12-S11-0 1 172" CI-SlI-C2 170" Methyltin Halides 85 among other things, shows the great& charge transfer among MeSnBrz complexes with electron-donor solvents such as pyridine, DMF, and HMPT. The sulfoxide complrx also has the highest (113) quadrupole splitting (4.49 mm/sec). Thc J(lH--C‘-llgSn) of 115.5 Hz found for the solution in rxcess DMSO (116) shows that the trans-dimethyl arrangement is retained in solution. Complexes of Mr2SnClzwith TCtzSO, PrzSO, BuZSO, Ph,SO (84) BzZSO, and BzMcSO (77) rwre characterized by their PhlR (77), IR (77, 84), and Mossbauer (84) paramc.tcm. nilany 1: 1 and 1: 2 complexes were iso- lated. The decrease in v (S=O) shows coordination via oxygen in all cases. A strong V(Sn-C) and a complex niultiplct in the Sn-C1 and Sn-0 region found in the 1: 2 compl(,xcs led to a structure of type IIIa. A trans- dimethyl arrangement in the solid statc fits well with the Mossbauer quad- rupole splittings (3.94-4.10 mm,/scc) . The J ( *H-C--”gSn) values for solutions of ?clc2SnClg- 2Et,SO, hZe2811(’l2-ZBzMeSO, and &le2SnC1z- BzzSO in CH2Clzarc 88.5, 85.5, and 80.0 Hz, respectivdy. In our opinion they reflect the complexed-uncomplc.xed quilibrium rather than the geometry of the complcxes in solution. The fact that BzzSO forms a 1 : 1 complex, whereas Mc2SO and McBzSO form both 1: 1 and 1: 2 complt-ws may imply an important role for steric factors on coordination. Thc 1 : 1 complcxes have two intense v(Sn-C) vibrations and a multiplrt in the Sn-C1 region; therefore, they were ascribed (84) a IIa structuw.
2. Complexes .zcith Pyriclzne, Phosph ine, and Arsine Oxides Pyridine, phosphinc, and arsine oxide complexes have also been studied extensively. In 1966, Clark and Wilkins (29) studied the IR spectra of Me,SnC12 complexes with pyridirir and triphcnylphosphine oxides (Table XVI) and assumc.d a structurcl of the IIIa type (see Table XI). Later, howcvcr, it was shown (SO, 73, 38, 84) that the Sn-C multiplicity in Me&Clz. PyO docs not significantly differ from that in MczSnClz.2DMSO. These facts, mhcn collated with the absence of vs(Sn-C), the singlet of v(Sn-0), and the magnitudc of A (3.96 mm/sec), lead to the IIIb struc- ture. This assignment was verified by an X-ray study (12) which rcvcaled a slightly distorted octahedron with the distances and angles as listed in Table XVII. The IR spectrum of Mc2Sn(”l2.PyO (Table XVI) contains two intense v(Sn-C) vibrations and a V(Sn-(’l) vibration. The complcx was, there- fore, assigned (84) a IIa structure, in agreement with the rule of prcdomi- nantly axial arrangement of elrctron-donor groups in a trigonal bipyramid (99). In Me2SnClzcomplexes with substituted pyridine N-oxides, a linear 86 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
h3ezSnCl2-2 I'y 0 574In - 233s 228s 223sh 324vs 54 1s - 229vs 234sh -
MczSnC1?. 1'yO 570111s 508111s 252ms 324vs 320sh hleeSnCle~21'hsl'0 577111 - 261s 246s 323m 305s 575111s 507w 248s 242Sll 264sh 258s
~ierSnCl2.2l'h3AsO 571 - 245-220s 37Sm
corrrlation was found (76) between v(Sn-0) and the u constants of the y substitucnts. Thc J(lH-C-llYSn) values found (51) for the complcxcs in solution are inscnsitivc to the geometry but are very sensitive, unlike Mr3SnBr complexes, to the nature of X in
No ,,(Sn-C) was observed in Mc2SnC12.2Ph3P0 or in Mc2SriClz. 2PhsAsO (29, 38, 84) arid a A value of 4.23 mm/sec was found for the former (84); both facts indicate, a tl-a?is-dimethyl arrangement. A multi- plicity of bands found in the Sn-0 frequency region suggests a IIIa structure.
Bond Length (8) Fragment Angle
Sn--CI 2.58 C1--SIl-C1 89.5" Sn-0 2.25 ("--St1-0 89.5" Sn--C 2.22 cI~-srl-o 95.6" 0-N 1.37 Sn-0-N 117" Methyltin Halides 87
3. E'yridine Coiriplexes
In the Ill spectra of pyridincx complcxcs (29, 27, 139, 85), there is only one v,,(Sii-C') vibration Iwt\iwn .XO and 564 cm-'. This fact, when col- lated nith the A values of 3.0s to 4.03 nim/scc (34, 113) suggests the trans-dimcthyl arrangement. l'hc1 fact, that the Sn-C frequencies pattcrii in MczSriClz.2Py closely rcwnil)l(~sthat found for b4czSnCIz.bipy (43) mas interpreted in terms of a IlIa structuw for both thr coniplexes, and for Me2SnBr2- 2Py, n hich \I as shuwn to IIP isoinorphous \\ ith MczSnCls- 2Py in an X-ray study. The prcwncc of v,(Sn-C) in Mc2Sn12.2Pgwas as- signed to a distortion of thr octahcdral structure. These assumptions fit well with thc 81Rr NQR spoctrurn of RI~~Snl3r~.2Py which contains just one resonaiicc signal (112). Proton inagnctic resonancc spcctroscopy \\as of no avail in the study of complcx formation of hlczSn('l2 \\ it 11 pyridinc in solution siiicc insoluble spccics arc formed lvith any ratio of rvactants. An IR spectrum of the MenSnCle.r-picoliric.complcx contains (61) a strong singlet for v,,(Sii-C), at 560 cm-l, and a broad Sri-('l multiplct at 233 cm-I, suggcsting the IIIa structure. The J ('H-('-I1''Sn) value. of 102.0 Hz found for thc. solu- tion in CHC& (I&), and corrcctcd for the possibility of dissociation, may confirm that the configuration is retaincd in solution.
4. Conzple.t-es toath DJ/F and Othei, L410iiodeiitate Liyands
In 1968, Okaiiara et al. (92) isolatcd thc 1: 1 arid 1: 2 R4e2SnX2(X = CI, Br) comp1cxt.s nith DMF and thc I :1 complexes 11 ith aromatic carbonyl compounds. In thc IIL spwtra of tliesc complexes (92), there is a shift of v (0)toward lower frquen&>s, suggclsting coordination via the carbonyl oxygen. For MezSnXs.2IlRIF, containing weak vS(Sn-C) vibrations in their IR spchctra, a structurc of the IIIa type (see Tablc XI) was proposed, although angle CSnC \\as adniittcdly below 180". The geometry was be- licvcd to be due to the low(~rcoordinativc ability of IIMF compared with, for example, DMSO. Indrcd, v(Sii-0) values are lower, and v(Sn-CI) values higher, in the DMF coniplcxc~compared M ith RleeSnXz-213RIISO. A IIb structure correctc)d via the. Ptliicttcrties rule (99) was proposed for MezSnXz.Il?vlF sincc thcw arc' t\\o intcnsc v(Sii-C) vibrations arid two intense v(Sn-Cl) vibrations that imply nonlinearity of the CSnC and ClSnCl skclctons. Complexes of type Me2SnXz- I, (L is a para-substitutcd carbonyl com- pound) were also assigned a IIb structure on the basis of similar IIt data. The application of IR and UV tecliniqucs to solutions of the complexes in 88 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV
1, 2-dichloroethane gave (92) stability constants lower than the constants for the respective bipyridyl complexes discussed in the following. In 1968, Tanaka and Kamitani (138) synthesized, and obtained IR spectra of the 1: 1 and 1: 2 MezSnClz complexes with dimethylselenoxide. Complex formation shifts v(Se-0) toward lower frequencies, suggesting probable coordination via the oxygen. The magnitude (60-100 cm-') of the shifts observed is, however, markcdly lower than that (150-200 cm-I) in similar complexm with DMSO. The presence of only one v(Sn-C) vibration and of two well-resolved v (Sn-X) vibrations suggests the IIIa structure. Complex MezSnCl2.McpStO shows two v(Sn-C) bands and one v(Sn-Cl) band, indicating a IIa structure.
5. Complexes with Bi- and Polydentate Ligands Complexes of MezSnXz with bipyridyl and o-phenanthroline have been studied extensively (47, 31, 29, 32, 37, 85). In nitrobenzene, they have very low conductivities (47) showing that no ionization occurs in solution. The presence of only one ~(Sii-C) vibration (ca. 570 cm-I) in the IR spectra of the complexes (Table XVIII) and the Mossbauer quadrupole splittings C4.03 nim/stc for MezSnClz pheri; 4.09 mm/sec for hlezSnClz. bipy, (34)] imply the trans-dimcthyl arrangemcnt. The presence of two v(Sn-N) frequencies and a broad Sn-CI multiplet suggest a IIIa struc- ture. Complex formation lowers the V(Sn-Cl) frequencies by 80-100 em-', so that bond polarity riscs considerably. Spectrophotometry gave (9Oj the stability constant for RilczSnClz-bipy in acetonitrile (log K = 3.36 f 0.04 liters/molc) and thermodynamical parameters of its formation (AH = 17.25 kcal/mole, AG = 4.50 kcal/mole, and AS = 24.8 eu). The v(C0) vibration found for the 1: 1 MezSnClz complex with N ,N- dimethylpicolinamide (94) is lower, whereas the ring-bending frequencies
TADLP; XVIII TIX-CARBOXAND TIN-HALOGENSTRETCHING FHEQUESCIES OF Rle2Sn)i,-L COMPLEXES@
bipy phen
Vibration C!l Br I C1 Hr I
v(Sn-C) 575s 571w 569m 578 - 572m 551w 560m 554sh v(Sn-X) 244s 151s 145m 247s 239sh 157m 149s 139m 126s
a Values are expressed in cm-1. Methyltin Halides 89 are higher, than in the frcc ligaiicl. This \\as explained by assunling that both the carbonyl oxygen and tho ring nitrogen participate in coordination. The prcscricc of two strong v(Sti--Cl) vibrations arid of one v,(Sn-C) and on(' V,,(Sii-C) vibratioii suggmts a IIIa structure with the CSnC backbone slightly nonlinear. 'l'hr fact that v(C=S) is unaffcctcd in the corrcsporidirig complcx with A', S-diniethylthiopicolinamide was explained on the basis of the loxcr donor ability of a C=S group compared with that of C=O.
In complexes of hlczSnClz ith A\7, ,~-diiiiethylnicotiriamideand the iso- nicotinamide, V(C=O) falls hut the ring-bending frcquericics rise. The structures were assumed (94) to bc. polymeric with ligarid bridges, since chelate structures arc impossil)l(. in this casc. Recently, complexation of orgat iotin halides with polydentatc ligands has been intcwsively studied in variou5 countrim (71,125, 94, 95, 72, 107, 108). An X-ray study madc in 1973 by Ikmdaccio (125) showed that the molecu- lar complcx of bis (snlicylald(~hyd(~)cthylc~wdiiniine \T ith MczSiiClz is, in the solid stat(., a polymer. 'l'h tin atom, trans-dimethyl groups, and trans-chlorincs lic on a plan(, of :in octahc~drori,whose axial positions are occupicd by the oxygen atoms of thv ligarid (a IIIb structure). A similar conclusion \\as arrived at in :tii IR arid Miissbaucr study (5) of R2SnX2 coiiiplcxes (R = CH3, C&, (',,HI;X = C1, Br) with bis(accty1acctonc)- ethylcnc~diimiiic.Thc same explanation probably holds for the NQR spectra of MezSiiXz complcxcs with dioxaii (112), although a IIIc structure is possiblc for the complexes uith DME. The PRlIt spectra of solutioiis of Me2SnC12 complcxc~sn it h various picoliiic.aldimincs showed (94) that the iininc nitrogen also takes part in the. coordination since thc irniric proton- tin couplings were observed. 'l'h~prrsei1c.c. of a strong v (Sn-C) vibration (575 cm-') and two intcnsc ~(Sii-('l) vibrations (240 and 253 cm-I) sug- gwt thc IIIa structure. The J (1H--C:--"9SSn) value of 111 to 115 Hz shows linearity of the C-Sn-C skd(~tori.The concentration depcndcnce of the couplings rcsultd in a dc%mriin:itiori of thv stability constants of the com- plexes, and thermodynamic p:ir:ttiitltcm for the coniplcxation were obtained calorimetrically. The interaction of orgaiiotin halides with tri- and tctra- dentate Schiff bases more oftcbti than not leads to cleavage of the Sn-X bonds, resulting in ionic, rat1ic.r than molecular complexes (71, 72, 95). These will not be discussed hcrc.
6. Complexes with Electron-Dotior Solverits In order to study the interaction of MctSnXz with electron-donor solv- ents, we have recorded and ailalyzc~d the 81Br, NQIt spectra (112) of RlczSnBrz complexes with dioxaii, THF, acetone, DME, pyridine, HMPT, 90 V. S. PETROSYAN, N. S. YASHINA, AND 0.A. REUTOV
DMF, and DMSO and found that the donor ability of the solvents in- creases across the wries given. The singlet pattern obtained suggested structurcs of types IIIa or IIIb which agreed with the A values of 3.85 to 4.40 mm/scc found from the Miissbauer spectra (113). A good correlation was found betwccm v (*'13r) and A. Both of thmc quantitim vary slightly (within 5 to 15y0and 8 to 13%, rcspcctivclly) in R/lezSn13rzcomplrxes with weak c>lwtron donors such as THF, acetonv, dioxan. In complexes with strongc.r donors (DMF, DMPO, HhlPT) , howvcr, the variations are rather high (40-45y0 and %5-:3O7,, rcqwctivcsly). It is interesting to note that the NQIt spectra of the 1:2 MczSnClz complexes with ac(toiie, and thc I :I complexes with DhIE, differ sharply from the respective complexes of MeoSnBrz in that thcy arc clear-cut doublets. This is readily vxplained in terms of a IIIc structure for the former complcxcs. A study of the concentration arid temperature dependences of J(IH- C--"%n) in MezSnClo/C:HzClz/electron-donorsolvent systems showed (116) that rcgularitics found in thc solid stat(>also operate, on the whole, in solution. A collation of NQR (llb),Mijssbauer (113), and PMR (116) data suggcxsts that in thc MezSnClzcomplrxcs the s-electron content in the Sn--C-H site, arid thc Sn-CI bond polarity arc greater than in the frec halides. As for the stcrcochemistriw of thc 1: 2 complexes, thcy are mgular or slightly distortctd octahedra of thc IIIa or IIIb types, depending on the solvent applied.
C. Complexes of Methyltin Trihalides In 1963, Beattie and McQuillan (8) studic.d th(. IR spectra of MeSnCI3 complexes with pyridinc, bipyridyl, and o-pheiianthrolinc, the compositions being 1 :2, I :1, and 1: I, rcspcctively. They obtained a rather complicated pattern of Sn-Cl vibrations and could make no structural assignment. Later it was shown (31, 81, 32) that IR spectroscopy is hardly applicable to the study of complexes of this type. In 1967, Wardell (151) studied the UV spectra of diethyl ether solutions of MeSnCl3 complexes with nitro- anilines and measured the equilibrium constants. The data were inter- preted in terms of 1 :I complexes containing a five-coordinate tin (with 2,s- or 2,4-diaminonitrobcnzcne) or a six-coordinate tin (with 3,4-di- aminoriitrobcmzencx). In 1974, Barbieri and co-workers (107) synthesized the 1:1 complex of McSnC13 with N ,IT-cthylencbis (salicylidcneiminato) nickel and studied its Mossbauer, IR, and electronic spectra in the solid state, and its elec- Methyltin Halides 91 tronic and PMR spectra in solutioii. Tlic data led to a structure of the Va or Vc type (Tablr XI). In 1973, we described riglit c~oniplcxc~s(111) of the type McSnX3.2L (X = C1, Br; L = Py, DMF, I)hISO, IIRIPT) as 1~11as hleSnCl3-dioxan and 2McSnBr3.dioxan. Wr stutlicd (1IS) the RIossbaucr spectra of these coniplexc~sand of frozcn solutioiih of RId3riX3 in DEE, DIME, THF, ace- tone, arid TMED. The isomw shifts \\('re found to be very scnsitive to complex formation, uiilikr the. casrs of hlczSriXz arid AIr3SriX, and led to a scries for thr clcctron-donor ability of the solvrnts ton ard RilcSnX3. The quadrupole splitting pat tcm found for lLlcSnXJ cornplexes is also very diff(wiit. Urilike Mc2SnX2,in \\ hich thc splitting increascs with electron- donor activity of the ligands, iii hl&i& it is greater for thc complexes with weaker donors. Thcsc. fact 5 ngrc'c 1% ith the present-day quadrupole splitting thcories (see Scctioii 11, 1)) and with thc NQEt spcctra of the complvxw (112). The pattc.rii of V( V CONCLUSION The data rcported in the litrraturc and analyzed in this review demon- strate that when we deal with thc twelve methyltin halides, which have bccri studied for more than 125 yars and have constituted the subject of 92 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV hundreds of papers, we are still far from a complctc understanding of the structures and, even more so, the properties of these rather elementary but, nevertheless, extremely important compounds. A promising fact, however, is that thc numbrr of rescarches devoted to the subject has risen steeply during rrcent ycars, as is clcar from the refer- ences cited. Unfortunately, the numbcr of electron diffraction and X-ray studies, which givr the most unambiguous information on the structures in the gas and solid statrs, is still rather low. As for solution studics, the use of IR and PMR methods has givcn a satisfactory picturc of the clcc- tronic structures and stcreochcmistrics of the molecules and their com- plexes. In this field as well, however, the numbrr of thorough studies is not high and it is possiblc that thc main problrms still await attack. .~CKNO~).I,EL)GMEUT The authors are greatly iiidchted to Ilr. A. V. Grib who kindly translated this review. REFXRE. 1. Bancroft, G. hI., and l’l:ttt, R. H., Advan. Inorq. Chem. Rudiochem. 15, 59 (1972). 2. Bancroft, G. M.,Das, V. G. K., and I3ut,ler, K. D., J. Chem. SOC.,Dalton Trans. 2355 (1974). 3. Bancroft, G. XI., Das, V. G. li., Sham, T. li., and Clark, M. C., Chem. Cornmiin. 236 (1974). 4. Barbieri, G., and Taddei, F.,J. Chrm. SOC.,Prrkin Trans. ZI 1327 (1972). 5. Ihrbieri, It., Cefald, R., Chandra, S. C., and Herber, R. H., J. Organometal. Chem. 32, 97 (1971). 6. Barbieri, It., and Herber, It. II., .I. Organonwtul. Chrnz. 42, 65 (1972). 7. Barbieri, R., Pellerito, I,., Bertazzi, N., and Stocro, G. C., Znorg. Chini. Ada 11, 173 (1974). 8. Beattie, I. R., and McQuillan, G. l’., J. Cheni. SOC.1519 (1963). 9. Beattie, I. It., and Ozin, G. A,, J. Chern. Soc. L4 370 (1970). 10. Bent, 11. A,, Chem. Rev. 61, 275 (1961). 11. Bischof, P. K., Dcwar, XI. J. S., Goodman, D. IV., and Jones, T. B., J. Organormetal. Chem. 82, 89 (1974). 12. Blom, l<.A,, Penfold, 13. R., and Robinson, W. T., J. Chern. SOC.A 913 (1969). 13. Bokii, K. G., Struchkov, Yu. T., and I’rokof’yev, A. I<., Zh. Strukf. Khim. 13, 665 (1972). 14. Bolles, T. F., and Drago, I?.. S., J. Amcr. Chem. Soc. 87, 5015 (1965). 15. Bolles, T. F., and Drago, R. S., J. Alnrr. Chem. SOC.88, 3921 (1966). 16. Bolles, T. F., and Drago, R. S., J. Amer. Chem. Soc. 88, 5730 (1966). 17. Brown, J. M., Chapman, A. C., Harper, R., Mowthorpe, D. J., navies, A. G., and Smith, 1’. J., J. Chem. Soc., Dalton Trans. 338 (1972). 18. Brown, 12. S., ISaton, n. F., Hosomi, A,, Traylor, T. G., and Wright, J. M., J. Organomelal. Cheni. 66, 249 (1974). 19. Buckle, J., Harrison, 1’. G., and Das, >I.I<., Inorg. Chim. Ada 6, 17 (1972). Methyltin Halides 93 20. Huckle, J., ITnrrison, 1’. G.,King, T. J., ant1 Itirhards, J. L4., Chern. Colnmun. 1104 (1972). 21. Burke, J. J., and Lauterbur, 1’. C‘., J. Airier. Chenz. Soc. 83, 326 (1961). 22. Butcher, F. I<., Gerrard, \Y., 1looney, b:. F.,Itees, R. G., \Villis, H. A,, Anderson, A., and Gebbe, H. A., J. Organottwtul. (‘hem. I, 431 (1964). 23. Calligaris, M., Itandaccio, I,,, Ihrbicri, I<., and Pellerito, I,., J. Orgnnonzetal. Chem. 76, C56 (1974). 24. Cefalli, It., I’ellerito, L., and Rrrhiwi, It., .I. Organomelal. Chem. 32, 107 (1971). 25. Chan, S. O., and Reeves, I,. \Y., Inorg. (‘hti. 12, 1704 (1973). 26. Clark, H. C., O’Brien, It. .J. O,, and Trotter, J., J. Chetn. Soc. 2332 (1964). 27. Clark, H. C., and Goel, It. C., J. Orgariomclal. Chem. 7, 263 (1967). 28. Clark, H. C., arid Puddepliatt, It. J., iti “Organometallic Compounds of Group IV Elements” (A. G. Rlcl)iariiiitl, ed.), \rol. 2, Part 11. Dekker, New York, 1972. 29. Clark, J. l’., and \\.‘ilkins, C. ,I,, .I. (,‘herti. Soc. A 871 (1966). 30. Clark, J. l’., Langford, V. If., and \\ilkins, C. J., J. Chem. Soe. A 792 (1967). 31. Clark, It. ,J. H.,and \Villiams, C. S., Spectrochint. Acta 21, 1861 (1965). 32. Clark, It. J. IT., Davies, A. G., :mtI l’udtleyhatt, It. J., J. Chem. SOC.A 1828 (1968). 33. Clark, 11. G., l\ladtlock, .4. (;., and l’latt, It. II., J. Chenr. Soc., Dalton Trans. 281 (1972). 34. Cordey-TIaycs, )I., I’eacork, It. I)., :mtl Vuceli6, M.,J. Inorg. iVucl. Chem. 29, 1177 (1967). 35. Crasnyanskii, 31. E., Litinskii, A. O., mid Schifrovich, E. J., ‘I‘eor. Ezp. Khim. 10, 536 (1974). 36. Crasnyanskii, M. E., Lysrnko, Yu. A,, I,it,inskii, A. O., and Schifrovich, 1;. J., Zh. Stritkt. Khinz. 15, 711 (1971). 37. Das, V. G. K., Znorg. 1Vitel. rhoti. IAt. 9, 1.55 (1973). 38. Das, V,G. K., and Kitchirig, \i’.,J. Orgcmornetal. Chem. 13, 523 (1968). 39. navies, .4. G., Harrison, 1’. (;., I 55. (:rutsch, 1’. A,, Zellcr, 11. V., and Felilner, T. l’., Iuorg. Chrm. 12, 1431 (1973). 56. Gupta, It., and hliijee, 13., J. 0rganonzdo.l. Chent. 33, 169 (1971). 57. Gupta, It., and Ilajce, 13., J. Orqanonicfa2. Chctn. 36, 71 (1972). 58. Gupt,a, It., and X‘lajcr, H.,J. Orgononzrfd. Chem. 49, 203 (1973). 59. IIarrison, P. G., King, T. J., atid Ilicliards, J. A4.,J. Chrni. Soc., Dalton Trans. 1723 (1974). 60. TIarrison, 1’. G., and King, T. J., J. Chrni. Soc., Dalton Trans. 2298 (1974). 61. Ilendrickcr, 1). (i.,Inorg. >Vitcl. (‘hem. Lrff.5, 115 (1969). 62. Iltvber, I?. IT., Stiicsklcr, 1-1. A,, :~tidl~eiclrle, \V. T., J. Choin. E’hys. 42, 2447 (1965). 63. lIcrhrr, 11. II., J. Iriorg. Sircl. ()‘hem.35, 67 (1973). 64. IIill, .I. C.,Dingo, I<. S.,:tnd Herber, 11. €I.,J. Amw. Chwni. Soc. 91, 1644 (1969). 65. Ho, 1%.I. I<., and Zuckerman, J. J., J. Orgnnonzrttc/. Chern. 49, 1 (1973). 66. IloltneP, .I. I<,, :md I<:Lesz, €1. U., ./. Ilnier. (“heni.Soc. 83, 3903 (1961). 67. IIuang, IT. H., ITui, I<. RI., and Chiu, I<. I<., J.Organonicfcil. Chenz. 11, 515 (1968). 68. IIulnie, It., J. (‘heni. Soc. 1524 (1963). 69. Ilunter, 13. li., arid Iteevrs, 1,. \V., Con. J. (VLCVZ.46, 1399 (1968). 70. Jaurii, I<. I,., and Vernia, V. K., J. Inorg. ,Vud. Chem. 35, 2361 (1973). 71. Iiawakami, I<., :id Taiiaka, T., J. Organornrlal. ~”hent.49, 409 (1973). 72. Iiawak:iini, K.,ATiy:i-lT(>tii, hl., arid T:in:Lka, T., J. Organonicftrl. (.“hem. 70, 67 (1974). 73. Kaiv:isaki, Y.,Hori, ll., anti l-eiiak:t, K., nil[/.Chcnf. SOC.Jap. 40, 2463 (1967). 74. King, T. J., :ind IIai on, 1’. G., CRcrn. (‘oninzttn. 815 (1972). 75. liitcliing, \V., Telrahdron Lctt. 3689 (1966). 76. Kitching, I!.., antl Ihs, V. (;. I<., Ai~sf.J. (‘hn~.21, 2101 (1968). 77. I 95. lIatsubayaslii, G., Okunaka, AI., :tnd ‘r‘anaka, T., J. Orqarrottwlal. Chon. 56, 215 (1973). 96. 12Iatwiyoff, p\T. A,, and Drago, l<.S., 111o~g.Chrttt. 3, 337 (1964). 97. RZcFarlane, \V., J. Chcm. Soc. ;I :)2S (19671. 98. lIit,chell, T. S.,J. Orgmot,rdtr/. (‘hctrt. 59, 189 (1973). 99. Muetterties, I<. I,., Mahlrr, \!’,, I’:ic*ker. I<. J., and Schniutzler, H., Inorg. f’hmi. 3, 1298 (1964). LOO. Okawara, lt., \\-chster, 1). I?,, :tiit1 Itoc.lron., E. G,, J. .ltti(,r. Chrttt. Soc. 82, 3287 (1960). 101. Parish, I<. V., Z’roqr. Inorq. Chc.ttr. 15, 101 (1972). 102. I’arish, R. V., and Johnsoil, (:. (‘hcttf. I’h!/s. Letf. 6, 239 (1970). 103. Parish, R. V., and l’latt,, R. II., wi. (‘on!tt/tlti.1118 (1968). 104. Parish, R. V., and l’lstt, It. II., .I. (“hcttz. Sor. 11 2145 (1969). 105. Parish, It. V.,and I’latt, R. ]I,, Itiouq. Vhirti. :lcfa 4, 65 (1970). 106. Parshall, G. IV., Inorg. (Thew. 11, 373 (1972). 107. I’ellerito. I,., Cefal~,R., Giairgrizz:~,.I., :tiid Rarbieri, R., J. Orgnnotnefal. (‘hotl. 70, 303 (1974). 108. l’ellerit,o, I,,, Cefalh, It., Silvwtri, !I., I)i Hianca, F., Uarhirri, R., H:tupt, lI.-J., l’reut, TI., and ITuher, F., J. Or(qtrttot~tclnl.(‘hem. 78, 101 (1974). 109. I’et,rosyari, V. S., Yasliina, N. S., :urtl Itrutov, 0. A,, Izv. ;Ikad. Nairk SSSX, Ser. Rhim. 1018 (39721. 110. I’etrosyaii, V. S., :~ndRcntov, 0. .I.,J. Orqanonz~tcd.Chotz. 52, 307 (1973). 111. I’etrasyari, V. S.,Yashina, 3. S.,:md Itrutov, 0. A,, .I.Organottzeld. Chcttt. 52, 315 (1973). 112. l’etrosyan, V. S., Y‘ashina, S. S.,ltcutov, 0.A,, Brucliova, I<. V., and Sernin, G. I<., J. Orgnnoitietcd. (’hon. 52, 321 (I 973). 113. l’rt,rosyan, V. S., Yashina, N. S.,S:wli:~rov, S. G., Iteutov, 0.A,, Rachev, V. Ya., arid Gol’danskii, 1‘. I., J. Oq/citiortc~,/cd.(‘hptn. 52, 333 (1973). 114. Petrosyan, V. S., and Iteutov, 0. ~4.,t’irrt Appl. Chem 37, 147 (1974). 115. Petrosyan, V. S., I’crniin, A. H., :iriti Ilrutov, 0. :I., IZP.Akad. Saul; SSSIZ, Ser. Khittz. 1303 (1974). n, V. S., Yasliina, S.S., I3:iklimutov, V. I., l’crniin, A. B., and Rcutov, 0. II.,J. Organotttctcil. (’hciii. 72, 71 (1974). 117. Petarmyan,V. S., and Permin, -1. l%,,Irtf. Conf. Orgcmottielal. Chrni., 7fh,Ahslr. I’npcrs, 1975 2.50. I 18. l’oller, It. C., “Tlic Chemistl~yof 0rg:inotin Compounds.” Logos Prrss, Tmdon, 1970. 119. l’oller, It. C., and Ttuddick, J. N. It., .I. Orqanottidal. Chmt. 39, 121 (1972). 120. Pollei-, R. C., and Ruddick, J. S. lt,, .I. (’hem. Soc., Dallon Trans. j55 (1972). 121. Potts, I)., Sliarrna, 1%. D., (’arty, ~1.J., :md JValkcr, A., Inorg. Chetn. 13, 1205 (1974). 122. I’udovik, .4. N., lluratovu, A. .I., blctlvcdeva, 11. I)., arid Yamaliyeva, I,. S., Zh. Obsch. Khint. 42, 2402 (1072). 123. Rsmos, 1‘. 13., arid Tobias, R. S.,Irrorq. (’hcm. 1 I, 2451 (1972). 124. Itamos, V. 13., and Tobias, R. S., Spw/ror.hinr. Ada 30.4, 181 (1974). 125. Itandaccio, I,., J. Orgnnomeln/. (‘hotr. 55, C58 (1973). 126. Iledl, G., and lfkkur, M.,J. Or~qccnot,ieltal. Chon. 26, C36 (1971). 127. Iionova, I. A,, Sinitsina, Y, :I., Strurltkov, Yu. T., Okhlobystin, 0. Tu.,and I’rokof’yev, A. I<., Zh. Sfrukt. Khitti. 13, 15 (1972). 128. Ruddick, J. N. It., and Sams, J. It., J. Orgnnornrtal. Chern. 60,233 (1973). 96 V. S. PETROSYAN, N. S. YASHINA, AND 0. A. REUTOV 129. Schlemper, E. O., and IIamilton, \\:. C., Inorg. Chem. 5, 995 (1966). 130. Semin, G. K., Babushkina, T. A,, and Yacobson, G. G., “Application of Nuclear Quadrupole Resonance in Chemistry” (in Russian), Moscow, 1972. 131. Skinner, H. A., and Sutton, L. I<., ‘I’rans. Paraday Soc., 40, 164 (1944). 132. Smith, P. J., Organonietal. Chem. lieu., A 5, 373 (1970). 133. Smith, P. J., and Dodd, U., J. Organornetal. Chem. 32, 195 (1971). 134. Smith, 1’. J., and Smith, I,., Inorg. Chim. Actu lieu. 7, 11 (1973). 135. Stockler, H. A., and S:tno, I%.,Chrrn. Cornmiin. 954 (1969). 136. Taimsalu, P., and LVood, ,J. L., Speclrochirn. Acta 20, 1043, 1357 (1964). 137. Tanaka,’T., Inorg. Chinz. Actu 1, 217 (1967). 138. Tanaka, T., and liamitani, T., Inorg. Chint. Actu 2, 175 (1968). 139. Tanaka, T., Matsumura, Y., Okawara, R., Muslia, Y., and Kinumaki, S., Bull. Chena. Soc. Jap. 41, 1497 (1968). 140. Thomas, A. B., and Rochow, E. G., J. Amer. Chrm. Sor. 79, 1843 (1957). 141. Thomas, A. B., and Rochow, E. G., J. Inorg. Nrcrl. (Them. 4, 205 (1957). 142. Tobias, R. S., Organomdul. Cheni. Rev. 1, 93 (1966). 143. Torocheshnikov, V. M., Tupciauskas, A. P., Sergeev, S.If., and Ustynyuk, Yu. J. Orgartometal. Chern. 35, C25 (1972). 144. Van den Berghe, 13. V., and Van der Kelen, G. l’., J. Organometal. Chent. 6, 515 (1966). 145. Van den Berglie, E. V., and van der Kclen, G. P., J. Orgunomelal. Chem. 11, 479 (1968). 146. Van den Berghe, E. V., and van der Kelen, G. l’., J. Organomelal. Chem. 59, 175 (1973). 147. Van den Berghe, E. V., and van der Kelen, G. €’., 1.Organomelul. Chern. 72, 65 (1974). 148. Van der Kelen, G. k’., Nature (London) 193, 1069 (1962). 149. Van den Vondel, U. F., \rillernen, H.,and van der Kelen, G. P., J. Organomctal. Chem. 63, 205 (1973). 150. Verdonck, L., and vaii der Kelen, G. l’., Rer. B~tsenges.l’hys. Chem. 69, 478 (1965). 151. IVardcll, J. L., J. Org~nomelnl.Chem. 9, 89 (1967). 152. Yasuda, K., Kawasaki, Y., Kossi, N., and Tanaka, T., Bull. Chem. SOC.Jap. 38, 126 (1965). 153. Isaacs, N. W., and Kennard, C. €1. L., J. Chem. Soc. A 1257 (1970). 154. Zuckerman, J. J., Advan. Organometal. Chent. 9, 21 (1970). Chemistry of Carbon-Functionul Alkylidynetricobalt Nonucarbonyl Cluster Complexes* DIETMAR SEY FERTH Department of Chemistry Massachusetts Institute of Technology Cambridge, Massachusetts OR I’ll call it “Fred” ‘cause alkylidynetricobalt nonacartmnyl’\ too ciatiin long tor me. ’ I. Introduction: General Properties of Alkylidynetricobalt Nonacarbonyl Complexes . . 98 11. Synthesis of Alkylidynetricobnlt, Nonacarbonyl Complexes . . 100 111. Chemistry of the Tricobaltcarbon Decacarbonyl Cation . . 110 IV. Highly Stable Nonacarbonyl Tricobaltcarbon-Substituted Carbonium Ions . . . 119 * This review is based in large part on research carried out at the Massachusetts Institute of Technology by Ralph J. Spohn, John E. Hallgren, Anthony T. Wehman, Gary €I. Williams, Paul L. K. Hung, C. Scott Eschbach, Mara Ozolins Nestle, Cynthia L. Nivert,, and Ying-Ming Cheng. 1 Claiming that it took longer to say “alkylidynetricobalt nonacarbonyl” than to make one, R. J. Spohn, who began our research program in this area, informally chris- tened this compound class “Fred.” ‘l’hus C1CCo3(CO)9is chlorofred, (OC)&o,CC02II, fredoic acid, (OC)&o&C(O)CHa, aret,ofrcdonr, etc. 97 98 DIETMAR SEYFERTH V. Decomposition Reactions and Derived Synthetic Applications of Alkylidynetricobalt Nonacarbonyl Complexes . . . 135 VI. Concluding Remarks . . 138 References . . . 141 I INTRODUCTION: GENERAL PROPERTIES OF ALKYLIDYNETRICOBALT NONACARBONYL COMPLEXES In 1971 we reviewed the chemistry of alkylidynetricobalt nonacarbonyP complexes (1).We have continued our rrsearch in this area and have made substantial progress in the development of the organic chemistry of this interesting class of organometallic compounds. In this article we review principally the chemistry of carbon-functional alkylidyrietricobalt nona- carbonyl complexes. In view of the unusual structurc of these complexes, it will be useful to begin with a general review of their properties, structure, and formation. The serendipitous preparation of the first member of this class of com- pounds was reported in 1955 by workers at the Bureau of Mines labora- tories at Bruceton, Pennsylvania (2): H *Many of the other workers in this field call these complexes methinyltricobalt enneacarbonyls. However, we are not Greek scholars and prefer the more prosaic Latin numerical prefixes. Alkylidynetricobalt Nonacarbonyl Complexes 99 FIG.1. Structure of CH~CCO~(CO)~.[From P. W. Sutton and L. F. )ah], J. mer. Chern. SOC.89, 261 (1967).] After initial confusion coricerniiig thc structure of this compound (9), chemical studies by othcr norkrrh (sa, 36) and, quite conclusively, an X-ray crystal structurc detcrinination hy Sutton and Dahl, (4)established thr prcscricc of a tetrahedral ('Co?clustvr in this compound. This structure is shoivn in Fig. 1. Thc apical carbon atom is coordinatcd symmetrically, apparently via u-bonds, to 3 cobalt atoms. Each cobalt atom may be re- gardcd as achieving a closrd-shell configuration by v-bonding to 2 othcr cobalt atoms, the apical carbon atom, and three terminal carbon monoxide ligands. The C--CH3 distanw is 1.53 A, the normal Csp3-Csp3 distance. The Co-C(apica1)-Co bond nnglos average 81.1",the H3C--C-Co angles averagc 131.3" (4).The arrangrmcnt of thc carbon monoxide ligands in this complcx merits special noticc for it has important consequences with regard to the chemistry of this class of compounds. Six of the ninc CO ligands are disposrd upward in the grnrral direction of the apical carbon atom and its substituent. As a wsult, any rractions at the apical carbon atom or at its substitucnt group will be subject to substantial steric hin- drancc. Not only is thc backsid(. of thc apical carbon atom well shielded, but attack from other directions also will be hindered by steric intcrfcrcnce of thcse six CO groups. Thus, stcric hindrance plays an important role in the chcmistry of alkylidynctricobalt nonacarbonyls. The alkylidynetricobalt nonacarbonyl complexes all are highly colored, with colors ranging from red to purplc to brown to black, depending on the apical substiturnt. Their thermal stability also depends on the apical sub- stitucnt; some survive heating to 100"-185" and many rnay be sublimed in vacuum at 5O"-8O0C. Most, but riot all, are air-stable (in contrast to their pyrophoric parent, dicobalt octacarbonyl) in thc solid and in solution. 100 DIETMAR SEYFERTH Most of the complexes that we have worked with are crystalline solids, usually (but not always) with well-defined melting and/or decomposition points. However, when the apical carbon atom bears a long carbon chain substitucnt, the melting points of these complexes are lowered and oils, rather than crystalline solids, can result. Most of the RCCO~(CO)~com- plexes are soluble in the usual organic solvents. In view of their color, stability, and solubility properties, thin-layer and column chromatography are well suited for their detection, purification, and isolation. Their com- bustion analysis presents no difficultirs. The general chemical reactivity of the alkylidynetricobalt nonacarbonyls includes instability toward attack by oxidizing agents and many haws and nuclcophiles. Thus, the action of ceric ammonium nitrate or potassium pc'rmanganate on an RCc'o3( C0)g compound results in evolution of carbon monoxide and thc formation of inorganic cobalt salts and the carboxylic acid derived from tht apical car- bon, RCOzH (5). Many alkylidynetricobalt nonacarbonyl complexes, how- ever, are stable toward protonic and Lewis acids. This property provides the basis for much of the chemistry which we have developed since. 1970. II SYNTHESIS OF ALKYLIDYNETRICOBALT NONACARBONYL COMPLEXES Procedures are available for the prcparation of RCCo3(CO) complexes with a wide variety of substituents, R, at thc apical carbon atom. The original preparation of CHKCo3(CO) via acetylcriedicobalt hcxacarbonyl [Eq. (l)]can be extcndcd to dicobalt hcxacarbonyl complexes of terminal acetylenes, (RC~H)CO~(CO)~(I), but the effect of It on the success or failure of this reaction has not been explored with a systematic variation of R. In any case, the (RC2H)Co2(CO)6+ ItCHzCCo3(CO)9conversion is not always successful. Furthermore, it is limited to thr synthesis of com- plexes with a CH2 group attached to the apical carbon atom. Complexes of the typc (OC)9C03CCHzRhave bcm prrparc.d by this rout(. with R = H, CH3, CF3, MeOzC, CzH5, HO(CH2)2, ?z-C~H~,MczCH, PhzCH, McZC, CCHb, p-BrC6H4, C6F, (6).Some of these products wcre formed directly in RC=CH/COZ(CO)~reactions (c.g., R = Ph, CF3); presumably thc acetyl- ene in these cases is a strong enough acid to induce conversion of thc ini- tially formed acetylenedicobalt hcxacarbonyl to the alkylidynetricobalt nonacarbonyl. Another more general preparative route to such complexes was devel- Alkylidynetricobalt Nonacarbonyl Complexes 101 oped by other rcscarch groups (36,7,8) : RCX, + Cor(C0'is- I1(:cO3(Co)g + (:ox2 This procedure may be used to prcparc' a great diversity of alliylidynetri- cobalt nonacarbonyl cornplexcs, including RCCo3( CO) compounds where 11 = H, halogen, alkyl, and aryl. Sonic. dihalomethyl compounds also may be used as starting materials, c.g., l'li('HCl2, CH30CHC12, arid [MeZN= CHClICl, which react \I ith Co2(('0) to givc €'hC('oa(C0)9, CH30CCo3- ('20)9, and Mc2NCCo3(CO) 9, rcywctivcly (9). Of particular interest are those reactions that introduce rcwtivc organic functional groups at the apical carbon atom: CC13C0,R + ("O2((Y ))& -(0c')9CO~CCO2IL (Yb, 7) (3) CCIj-C-R + CiJr(C'0)s -(OC)gCo,CCR (M) (4) II II 0 0 CCI3COSR, + ('i),((Y))" A (oC)9CoIcCosI~~(9) (-5) CC'ljl'(0)(011)~ + Co,(('O)x --c(OC)gC'i)3(:r(O)(oR), (9) (6) C"'?l~CI~d"IIl~'+ CO:IC*O)~ --c (OC)&o,CCR=CIIIt' (10) (7) The mechanism of thc reaction of organic di- and trihalo compounds with dicobalt octacarbonyl has not lx~wvlucidatcd yet. It seems cwtain that the active reagent is the Co(('0) 1- anion. This sprcics can bc used instead of dicobalt octacarbonyl, as can C'o, (('0)12. Some thoughts concerning the mrchanisni of RC1X3/Co2 (CO) rcact inns wcrc givm in our previous re- view (I). If the atom transfw/radical nicchanism ~thich wc suggested is, indeed, opcrative, it is not surprising that the €~CCO~(CO)~product yields are, in gencral, only low to modcmt(~,as shown in Table I, which presents some yield data from our on n nork. In the inc.antimc, it has become clcar that some attcmptcd RC'X,/t '02 (('0)reactions are unsucccwful not be- taus(' of difficulties iiith this rcwtiori pc'r se, but brcause of the instability of the product to the reaction and/or \I ork-up conditions. For instance, CCl&R=CHR'/Co2 (CO)8 rcmtioiis must bt. carried out undcr rigorously anhydrous conditions and hydrolytic uork-up must be avoided or rlse rcduccd products, (0'2)9C03(Y'HltC'H2H', arc ohtaincld instead of the de- sired vinylic products (10).Thus w~ \I crc surprised that our first attempted preparation of Mc3SiCH=('HC'C"03 (('0) by reaction of R4e3SiCH= CHCBr3 nith Co~(c'0)gave RI(~3Si('HzCHzCCo3(CO), as the sole product. Maintenance of scrupulously anhydrous conditions throughout reaction and product isolation steps gave' th(x desired product. Attempted prc'para- tion of a-hydroxy dcrivativcw of typcx ( OC)9C03CC(OH) RR' has bccn 102 DIETMAR SEYFERTH TABLE I PREPARATION OF ALKYLIDYNETRICOBALTNONACARBONYL COMPLEXES BY REACTIONSOF POLYHALIDES WITH DrCoBALT OCTACARBONYL" Starting halide R in ItCCoa(C0)P produced yo Yield cCl4 c1 46 CBr4 Br 43 CIIBr3 H 34 CH3CC13 CH3 43 C6H6CClz CfiHs 29 CeH5CHClz C6Hj 34 Me3SiCC13 MeaSi 41 PhMenSiCC13 PhMesSi 50 (EtO),P(O)CCl3 (EtO)zP(O) 42 CCl3COzEt EtOzC 53 CClpCOzSiMea Me,SiOzC 38 CClaCONEtz EtzNC(0) 19 CCl,COCH3 CHaC(O) 40 CClaCOC4Hg-n n-CIHsC(0) 49 CC13COCMe3 Me3CC(0) 24 CCIzCOPh PhC(0) 33 CClaCHZOH CHzOH 0.8 CBr3CHZOSiMea CIIZOH (after hydrolysis) 5 CHaOCHC12 CIIaO 27 [ClCH=NMez]Cl MezN 26 CIIz=CHCC1ab CHz=CH 45 CHz=C (CH3)CClp' CHz=C (CIIa) 59 Me3SiCN=CHCBr3b MeSSiCH=CH 28 CHaC(O)CH=CHCCla' CHaC(O)CH=CH 19 a From Seyferth et al. (9). Seyferth et al. (10). quite unsuccessful to date, giving either none of the desired product at all (11 ) or only very low yields ( < 1yo) (9) when trichloroethano1 derivatives of type CC13C(OH)RR' were allowed to react with dicobalt octacarbonyl. The major products of such reactions were the alkyl derivatives, (OC)J2osCCHRR', but the yields of these also were low ( < 10%) (21). Conversion of the starting alcohols to their trimethylsiloxy ethers prior to the reaction with dicobalt octacarbonyl resulted in an increase in the (OC)&03CC (OH)RR' yield to 4-5% (9). To cite one more example, the aldehyde (OC)&o&CH=O was not obtained when chloral was allowed to react with dicobalt octacarbonyl, but it could be isolatcd in 3% yield when the ethylene glycol-derived acctal of chloral or bromal was used in place of the aldehyde (9). In all of these reactions we believe that it is the Alkylidynetricobalt Nonacarbonyl Complexes 103 formation of cobalt tetracarbonyl hydridc, HCo (CO)4, a strong protonic acid, during the reaction or during work-up, that results in diversion of the desired product; in fact, we have shown that HCo (CO) 4 or Co2 (CO)8- strong acid systems will convert (Or)gCosCH=CH2 to (OC)&oZCCHzCH3 (10). One may also postulate reasonable reaction courses for HCo(C0)4- induced destruction of (OC) g<‘orK (OH)RR’ and (OC)9Co3CCH=O, especially with after-thc-fact ltnou Icdgc. of our later research results in the area hich we review here. Interconversions of one alkylidyiietricobalt nonacarbonyl complex to anothcr also are possible but arc’ rathcr limited in scope. First, as already mentioned, many strong bases and nucleophilcs ultimately destroy the RCCo3(CO) cluster. In a halogm derivative, XCCoS( CO) 9, any nucleo- philic attack in which the halogcw atom is displaced must occur by way of nucleophilic attack at halogen or I)y an electron transfer mechanism, since, as already mcntioncld, the backsidc of the apical carbon atom is very cffec- tively shielded. A usually iionproductivc. alternative which very likely re- sults in ultimate destruction of the cluster is nucleophilic attack at the carbon atoms of coordinated (‘0 ligands. 13romo- and chloromethylidynetri- cobalt nonacarbonyl are arylatcd in moderate yield by the action of an aryl Grignard reagent in considcrablc CXCPSS (Table 11) (12).Primary and secondary alkyl Grignard reagmts, on thc other hand, destroy rather than alkylatcl these cluster halides. It \\as suggested that either course of reac- tion possibly begins with RhlgX attach at coordinated carbon monoxide (12), but, in our opinion, rcaction via an electron-transfer mechanism seems a reasonable alternative.. Organolithium reagents react with CICCoa(CO) 9 to form a radical anion of some stability, [C1CCo3 (CO)9]1, which is dctcctablr by ESR (IS) [also formed by the action of an alkali metal on thc chloro conipound (14)1, but subsequent events do not lead to alkylation of the cluster, but rat1~c.rto its destruction. A very useful reaction for thc conversion of HCCo3(CO)g to ArC- Cos (CO)9 compounds has been cl(~vc1opt.din these laboratories (5) : benzene or THF, reflux ArLHg + HCCCJ,(CO), + ArCCo,(CO)s + ArH + Hg (8) GO atmosphere, 2-8 hr Arylmcrcuric halides niay bc used in place of diarylmcrcurials. Excc.llent yields, ofttw 90-98%, were obtainc.d in many cases. Table I11 lists some examplrs. Oxidative degradation studies of selected products showed that the aryl carbon atom originally attached to mercury was thc one that bc- came bonded to the apical c:trbon atom of the cluster. This reaction was much less useful in the preparation of alkyl derivatives, because of very long reaction times (1-2 wec.ks) giving at best moderate yields (Table 111). The reaction of cr-haloalkylnicrcurials with HCCo3(CO) 9 occurred TABLE I1 GRIGNARDSYNTHESIS OF SUBSTITUTEDBENZYLIDYNETRICOBALT NONACARBONYL COMPLEXES FROM BROMOMETHYLIDYNETRICOBALTNONACARBONYLa Ar in ArMgBr' Product % Yield @c co, ( CO) 40 70 H,d H,C @- CC0,(CO~, '? OCH, OCH, CH,O 40 25 @- 20 10 23 25 21 0 30 a From Dolby and Robinson (fd). * More than 9 moles of ArMgBr per mole of BrCCoa(C0)s is required. Alkylidynetricobalt Nonacarbonyl Complexes 105 TABLE I11 PREPARATION OF RCcOs(C(~)sCOMPLEXES BY REACTIONSOF ORGANOMERCURI.~LSWITH HCCo3(C0)ga R in RCCo,(CO), Organomercurial’ produced % Yield 93 64 92 96 49 83 93 57 85 86 51 69 75 58 57 26 54 11 Fe 14 37 oc+co c 0 CH,OCH,CH,HgCl 32 (~-c,H,,),H~ 32 a From Seyferth et al. (5). Reactions were carried out under an atmosphere of carbon monoxide unless otherwise specified. Reaction carried out under an atmosphere of dry nitrogen. D. Scyferth and C. S. I~hhhacli,J. Oryanonwtal. (‘hem. 94, C5 (1975). 106 DIETMAR SEYFERTH with concomitant reduction of the carbon-halogen bond but provcd to be an unexpectedly eff ective synthcsis of alkyl-substitutcd cluster complcxes (5) I s (It = I€, S = Br: 59% product yield) (It = >le3Si, = 13~:70% product yield) C6Hs (RCHI)LHK + IICCo,(CO) g -ItCIHiCC0~(CO)g (It = TI, 77% product yield) (I?. = CH,, €%yoproduct yield) This react,ion also m-as uscxful in the preparation of specific deuteratcd derivatives : Hg(CH2Rr). + I)CCos(CO)g -CHII)CCO~(CO)~ (11) Hg(<'l>J). + IICc03(C~o)g- cIIL>fc03(Co)g (12) Reactions of organomcrcurials with HCCo3(CO) appear to bc subject to steric hindrance. As sccn in Table 111, ortho-substituted arylmcrcury compounds usually give. lower product yiclds than those with substitucnts in mcta or para positions. The prcscncc of two ortho substitucnts largcr than fluorine causrs such reactions to fail coniplctely. Thus, no organo- cobalt cluster product was obtain~din reactions of HCCO~(CO)~with bis (ptmtachlorophenyl) mercury and dimwitylmercury. Unfortunately, nothing is known at this time concerning the mechanism of thesr alkyl- ation and arylation rmctions, arid thc reduction process accompanying thc rcactions of Lu-haloalkylmc.rcurials with HCCo3(CO) is at present still a complete mystery. Anothcr synthcsis of aryl-substitutc.d methylidynctricobalt nonacar- bonyls was developed by Dolby and Robinson (15) who found that chloro- methylidynctricobalt nonacarbonyl alkylatvs aromatic compounds in a Fricdcl-Crafts-type reaction. High product yields wcre obtained when equimolar amounts of C1CCo3(CO) and aluminum chloride and an cxccss of the arcne were' stirred at 60"-70°C for 2 hours. When the arene was a solid, the reaction was carried out in dichlororncthanc solution. Both ortho and para substitution was encountcrcd, thc size of the substituent(s) al- ready present on the benzene ring appearing to dctcrmirie the position of substitution (Table IV) . Noteworthy is that milder temperature conditions affected the position of substitution; thus, reaction of chlorobenzene with chloromethylidynetricobalt nonacarbonyl in dichloromethane at 42" gave Alkylidynetricobolt Nonacarbonyl Complexes 107 TABLE IV FRIEDEL-CRAFTSSYNTHESIS OF !SUBSTITUTED BENZYLIDYNETRICOBALT NON.\C.IRBONYLCOMPLEXES FROM CHLOROMETHYLIDYNETRICOBALT NONACARBONYL" Reaction Arene Solvent temp. ("C) Ar in ArCCo3(C0)9 produced % Yield the p-C1-substituted product rather than the ortho isomer obtained at 70". Deactivatrd arenes, such as nitrobenzene, and hindered and poly- substituted arenes did not rmct with ClCCo3(CO) 9/A1CI3. Although the Friedel-Crafts procedurc appc'ars to be useful for the preparation of vari- ous substituted benzylidynctricobalt nonacarbonyl complexes, one is at the mercy of dircctive effects and thc arylmercurial/HCCo3(CO)g reac- tion has a much broader scopc of itpplication. An intriguing feature of the aluminum chloride-induced rractions of CICCo3(CO) is that they may involve the unusual apical carhoniurn ion (I) or at least a species with a great deal of positive charge at thc apical carbon atom. We will rncounter this species again later. The addition of H-CCo,(CO)g to an olefinic C=C bond is a newer method of introducing organic functionality at the apical carbon atom of a CCO~(CO)~cluster and was reporled first from these laboratories in 1971 (1). Ally1 acetate and ally1 ethyl ether were found to react with HCCo3(CO) to give (OC)9Co3C CCH2) 302CCH3 and (OC)9Co3C (CH2)3- OC2H5,respectively, when the oldin was used as solvent. This addition occurred only in the presence of about 20 mole % of a free radical initiator, 108 DIETMAR SEYFERTH azoisobutyronitrilc (16): 0 II ~H,CH,CH,OCCH, The implication that the apical radical (11) is an intermediate in these reactions is of some interest and suggests that other radical reactions of such organocobalt cluster complcxes may be found in the future. Results reported by Japanese workers (17, 18) confirnicd that such reactions arc free-radical proccsses and provided more examples of this addition reac- tion. Unfortunately, the synthetic potential of such C-H addition seems limited in view of the low product yields generally obtained (Table V). Some of the niorc reactive functional groups were rcduced in such reac- tions. Thus the rcactioiis of HCCO~(CO)~with ally1 bromide and with vinyl acetate gave C'H2=CHCH2CCo3(CO) and C'2HsCCo3(CO) 9, re- spectively, instead of the expected functionally substituted clustcr com- plcxes. + (CO), (CO), (1) (11) These then are the more generally applicable routes by which such RCCo3( CO) 9 clustrr complcxes may deliberately bc prepared. At times such complexes are unexprctcd main products or by-products in reactions involving cobalt carboriyl complcxes. For instancr, synthesis of succinic acid from acetylene, carbon monoxide, and water in the presence of a cobalt carbonyl catalyst is accompanied by formation of organocobalt carbonyl by-products, including the organo-functional cluster complex ( 0C),Co3CH2CH2C02H(19~). Related to this transformation is the acid- induced conversion of acetylrnedicobalt hcxacarbonyl carbonylation prod- ucts to unsaturated acid drrivativcs of the (OC)&O~Ccluster (19a, b) : RO \// H,SO,/Me,CO -H,O (OC),Co,CCH=C(R)CO,H (14n) ,=\ ,=\ H,SO,/MeOH-H,O (oc),co~co(co), * (OC),Co,CCH=C(R) C0,Me (14b) Alkylidynetricobalt Nonacarbonyl Complexes 109 TABLE V REACTIONSOF METHYLIDYNETRICCIBALTNONACARBONYL WITH OLEFINS~. Olefin Product % Yield CH,=CH (at 130°C) C',H,CCo,(CO), 20 CH,CH=CH,(at 130°C) C',H,CCo,(CO), ( 5: In liso ) 10 11 0 4 0 26 45 CCO,(CO), 0 11 CCO,(CO), 0 25 CH,=CH-CH =CH, (:H,CH=CHCH,CCO,(CO), 32 CH,=CH -CH ECHCH, :H,CHCCO,(CO), I 11 CH =C HCH, CH,=CHCO,CH, CH,O,CCHCCo,(CO), 19 CH, a From Sakamoto et al. (17). * Reaction for 5 hours at 100°C. Other reactions that one may cite iricludc the following: l0OT (OC)4CoCF2CF2Co(CO)a- CFJ<:Co3(C0)S (20) 110 DIETMAR SEYFERTH THF l\Ze3SiCo(CO)r-1\fedSiOCCo3(CO), (23) (18) EtzO Co,(CO), + 52 ---(OC)9C03CO-M+ (>I = IA, Sa, K) (24) (1 9) Finally, the formation of a "supcrclustd' by the reaction of dicobalt octacarbonyl with hexachlorocyclopropane in THF solution may be noted (25) : A single-crystal X-ray diffraction study of its benzene solvate was required to establish its identity. CHEMISTRY OF THE TRICOBALTCARBON DECACARBONYL CATION One of the major goals in our study of the chrmistry of the alkylidynetri- cobalt nonacarboriyl complexes was to obtain cxperimental information concerning the clcctronic effects of the nonacarboriyl tricobaltcarbon sub- stituent. In order to obtain such information, it was necessary to have available some RCCo3(CO) complexes in which the substituent R con- tained a conventional, well-studied organic functional group. At the time we began these studies, the only such derivatives in the literature were the esters, (OC)9CosCCO2R (R = Me, Et) . These could be prepared by direct reaction of the appropriate ester of trichloroacetic acid with dicobalt octacarbonyl (Sb, 7)or by the methanolysis of bromomethylidynetricobalt nonacarbonyl at 50" (7,26). These, we thought, should be a good starting point, since directly, or by way of the free acid and its various derivatives, they could be converted to many other organic functional groups. The first conversion of (OC)&03CC02Et which we attempted was its hydroly- Alkylidynetricobalt Nonacarbonyl Complexes 111 sis to the acid, (OC) gCo3C('O2H (27'). Base-catalyzed hydrolysis using sodium carbonate in aqueous THF resulted in destruction of the complex within 30 minutes. Our initial attrmpts to effect acid hydrolysis of the ethyl ester also were unsuccrssful; thr use of catalytic, stoichiometric, or above stoichiometric amounts of mineral acids in aqueous THF was with- out effect, and the ethyl estrr as rrcovcrcd unchanged. Why, we askcd ourselves, is this cluster-substituted c1stc.r so resistant to acid hydrolysis? One could envision that both rlrctronic rffects and steric effects are re- sponsible. We have already notrd in our initial discussion of the structure of RCCo3(CO)g complexes that stcric hindrance is cxpected to play an important role in their chemistry, and our further efforts to effect the hydrolysis of (OC)gCo3CC02Et n ere bnscld on the assumption that rvc were dealing with a problem of steric hindrance. Some observations by Harnnictt et nl. (28) on solutions of carboxylic acids in concentrated sulfuric acid had 1c.d to a procedure which served in thr cstdication of hindered ncak acids and in the hydrolysis of their esters (28, 29). It had been found that most carboxylic acids dissolved in concentrated sulfuric acid to givr tlw rnonoprotonated species. However, some highly hindered carboxylic acids reacted further to give the much less hindered acylium ion : [RC When a sulfuric acid solution containing such a reactivc acylium ion was poured onto ice, the carboxylic acid was produced; when it was treatcd with an alcohol, the cstw was formcd. The magnitude of the steric factor is critical; in contrast to 2,4,6-trimcthylbenzoic acid which formed the acylium ion in concentrated sulfuric acid, 2,4,5-trimrthylbenzoic acid did not. We decided to try thr Hamnicbtt procedure for the hydrolysis of (OC)gCo&COJCt, although it was not at all certain that this compound would survivc trc.atment with such strong acid. However, the ester dis- solved in concentrated sulfuric acid without evolution of carbon monoxide, giving a yellow-brown solution which was stable at room temperature when maintained under a dry nitrogen atmosphere. When such a solution was poured onto cracked ice, tho dcsircd (OC)&03CC02H was obtained in virtually quantitativc yield. When such a sulfuric acid solution of the acid or of the ethyl ester was poured into a large excess of another alcohol, the ester of that alcohol was obtaincd, generally in high yield (Table VI). 112 DIETMAR SEYFERTH TABLE VI PREPARATIONOF ESTERS OF (OC)9C03CC02HBY THE SULFURICACID PROCEDURE^ ItOH It in (OC)9C03CCOzR % Yield 9.5 96 39 99 85 a From Seyfcrt,h el al. (27). This procedure, however, was not applicable to the preparation of aryl esters. Of greater interest to 11s than the availability of the desired acid was the fact that ( OC)9<:03CC"02Rcomplexes dissolved in concentrated sulfuric acid, apparently to give the tricobaltcarbon decacarbonyl cation, (OC)&o&CO+, by the reaction sequericc shown in Eqs. (21) and (22). This novel acylium ion would be expected to bc more reactive than any other dcrivativc of thc acid, e.g., the acid halides. Sulfuric acid, of course, was not the solvent of choice in which to study the chemistry of the (OC)&o&CO+ ion since it is not compatible with so many of the nucle- ophiles with which on(' might vant thc acylium ion to react. Maximum development of the chemistry of this spwies required its generation in an inert organic medium, and well-known techniques from carbonium ion chemistry made this possiblc. Solution of eithrr carboxymethylidynetri- cobalt nonacarbonyl or one of its estcrs in a minimum volume of propionic anhydride followed by addition of an exc('ss of 65yo aqueous hexafluoro- phosphoric acid resulted in the immediate precipitation of a black solid. This material was moisture sensitive but stable to oxygen and could be handled using Schlcnk tube techniques. It was insoluble in and unaffectcld by ethers, chlorinated hydrocarbons, acetone, and many arenes; it dissolved in nitromethane without reaction. An analytically pure sample was obtained simply by washing it with dichloromethane. The analysis, as well as all of the chemical conversions of this solid, was compatible with its formulation as the salt (OC)9C03CCO+PFC-. In most of its reac- tions which we studied the solid salt was allowed to react directly with the nucleophilic substrate or it was used in the form of a dichloromethane slurry. In a few cases, nitromethane solutions of the salt were used. The chemical reactions of (OC)&03CCO+PFC- are summarized in Fig. 2. Individual examples are given in Table VII. As expected this reagent Alkylidynetricobalt Nonacarbonyl Complexes 113 (OC),Co,CCO,H 0 II (OC),Co,CCSR t (OC),Co,CCO,R (OC),Co,CC(O)R (M= ZnBr (OC),Co,CC(O)Ar J RR'NH (OC),Co,CC(O)NRR' FIG.2. A general survey of the reactions of (OC)~C~~CCO'PFB-. acylates primary, secondary, and tertiary alcohols; it also rcacts with phenol, a reaction that could riot be accomplished in sulfuric acid medium. Acylatioii of ammonia, primary arid secondary amincs gave amide deriva- tivcs of thc (OC)9C03Ccluster which wcre not very stablc thermally. Even less stablc werp thc imides obtainrd by acylation of formamidc and aceta- mide with (OC)&03CCO+PFb-; these decomposcd slowly even at 0°C. The reaction of (0C)9C03 Nuc leop hi le R in (OC),Co,Cr product % Yield CH,OH 82 C,H,OH 96 (CH,),CHOH 89 (CH,),COH 84 HOCH,CH,OH 35 HOCH,C CCH,OH 52 C,H,OH 72 C,H,SH 80 C,H5SH 80 NH, (g) 89 CH,NH, (g) 96 (CH,),NH (g) 84 (C,&),NH 93 On- On- C(0)N-0 52 L/ LJ C6%NH2 C(O)NHC,H, 73 C,H,NHCH, 66 42 HC(O)NH, 20 C(O)NHC(O)H 42 c (0)NH c (0)CH, 65 C(O)NHCH,CO,Et 63 C(O)NHCH,C(O)NHCH(CH,Ph)C (O)NHCH,CO,Et 58 80 70 69 31 58 9 54 a From Seyferth et al. (87). Alkylidynetricobalt Nonacarbonyl Complexes 115 or S-H functions. Application\ in polypc~ptid~X-ray crystallography and dcctron microscopy arc conct4v:il)lc. Attempted alkylation of (0(')9('odC'CO+PF6- with organolithium or organomagncsium reagents as not successful. However, mild alkylating agents did scrvr in the prqxuation of ketones from this salt. Ethylzinc halides rc.acted rapidly in THF vith the PF6- salt to givr(OC)gCo3- CC(0)Et in variable (20-55 ) yicdd. Tetramcthyl- and tetrarthyltin also wer(' found to rcact mith thv clustcr acylium ion, but much less rapidly, to give thc cxpcctcid kctotws. Thc nrylation of (OC)gCo3CCO+PF6- with nuclrophilic aromatic substratcii as (Jf spccial inter& since it provided clear rvidcnce that thc c1ustc.r acyliuin ion was rolativr~lypoorly rractive as an clcctrophile and that this \\:is due to t+ctronic factors. The acylation of diphcnylaminc has hccw nwiitioncd already. Only highly nuclcophilic aromatic substrates wew found to undrrgo such reactions : S , ,V-dimethyl- aniliiw, pyrrole, indolr, and f(wwcene, but iiot anisole or bcnzofuran. The fact that anisole, which is fairly high on the Friedrl-Crafts reactivity scale, does not rcact must be duck to :m clcctronic factor, since S,-L'-dimrthyl- anilinc docs rcact . At the time the chemistry of (O(')9Co3CCO+PF6- was being drveloped (27, 30),anothcr route to this iiovc~lacylium ion was found in thew labora- tories (31, 32). This discovery \\as a result of our intention to adapt the Friedel-Crafts synthesis of bcrizj lidynct ricobalt nonacarbonyl complexes of Dolby and Robinson (15) to the preparation of alkyl derivatives of mcthylidynetricobalt nonacarl)onyl, whose general preparation was not well in hand. Thr reaction chosrn for investigation was thP aluminum chloridr-induced interaction of trtraalkyltin compounds with ClCCoX(CO)9 In analogy to a known kctonc. synthesis (M), AlCll RaSn + l<'(*O('l -1t'COR + liaSnC1 (24) this reaction might have bwn c>xprctrdto give thc alkyl-substituted clus- trr, RCCoa(C0)9. Howevvr, uh(m this reaction was carried out in di- chloromethanc solution with tc,tramrthyltin, the product obtained (in 26% yield) was the acetyl derivat ivr, <'H3CI (0)CCo3 (CO)9, rather than the mrthyl compound, CH3CCo3(('0) 9. Since wr had found tttratiwthyltin to rract with (oc)gCo3CCO+PF6- to give this acetyl dmivativc., it u as a reasonable conclusion that thc action of aluminum chloride on cliloron~cthylidynetricobaltnonacarbonyl under our reaction conditions had lcd to formation of an acylium ion salt, per- haps (OC)9C03CCO+AlCli-. Furthcr investigation of this system indicated that this actually was what, was occurring (54).Thr chemistry of the formation of the acylium ion in this system must be somewhat complicated. When CICCo3 (CO) and a two- to thrcrfold excess of aluminum chloride arc mixcd under nitrogen in dichloromethane, an initially purple solution 116 DIETMAR SEYFERTH is formed as most of thc solids dissolve. (It is such solutions that can alkylate aroniatic substrates.) When such a solution is stirwd under nitro- gen for 20 to 30 minutes, the color changes from purplc to an intense yellow-brown. At this point, thin-laycr chromatography shows that the CICC'03(CO) 9 has bccm consunicd complctclly; thc reagent formed is com- pletcly dissolved or alinost so. Such rcagrnt solutions arc stable at room tcmprraturc for at lcast 2 days and probably longcr. Trcatmmt of such ycllow-brown solutioiis with any of thr niiclwphilrs knom n to react ith (OC) gCoaCC'O+PF,,-, at rooin temprraturr, rrsults in forniatiori of the expected acylated product, gcncrally in good yield, in what apprar to be nmrlg instantaneous rmct ions as indicated hy thc obswvcd color changes. Table VIII shows some of thr reactions that \I(TC carricd out with the (OC) gCo3CCO+A1C14-reagent . Acylation of alcohols, phcriol, thiols, ain- monia, and aminm could be accomplishrd; reactions with tetraalkyltins gave ketones; highly iiuclcophilic aromatic compounds were acylatcd. Especially noteworthy is thr rcaction with tricthylsilanc which gave the aldchydc (OC) gCo3CCH=0. By contrast, (OC) gCo3CCO+PF6-reacted with tricthylsilanc in dichloronicthanc to givc a mixture of CH3CCo3(CO)9 (30y0)and HCC'O~(('O)~(43y0). Honcvcir, when 3 molar equival(~ntsof A1C13 was added to a slurry of (OC)gC'o~('CO+PFG- in dichloroimthane, the bro~vn,iirarly homogrncous solution which formed reacted R ith an cx- crss of tricthylsilanr to givc thc clustc>ralddiyd(1 in 74% yidd. Wr suggest that aluminum chloridc effects rclcasc. of t he aldchgdc from thc intcirmcdi- at(* in thc first step in the rrduction of thc1 acylium ion, (OC)gCo3- CCHOSiEt3+,via Cl- attack at silicon. In thc1 absence of aluminum chloride, furthcr rcductiori to givc the methyl dcrivativr is possible. Thc conversion of chlororiicthylidynetricobalt iionacarbonyl to the acylium ion by the actiori of aluminum chloride is a remarkable proccss. The rcaction docs not require rxt crnal carbon monoxide; it proceeds per- fcctly satisfactorily undcr a nitrogcm :Ltniosphcrc. Tho CO function at the apical carbon atom of the products obtained thus was derived from carbon monoxide ligarids on cobalt in C1C'Co3(C0)9.This transfer of CO to the apical carbon atom is vcry c+Kcicnt. Thr yic.lds given in Tahlc VIII are based on the amount of elCC~(co)gchargc>d. If one rccalculates thc yield of (OC)9C03CCOzCIIjin thr first critry in Tablc VIII based on the available (OC)gC'03CCO+ from thr C'l('Co.~(CO) used [assuming the de- struction of an amount of C:1C<'03 (C'O) cquivalmt to the thcorctically requirrd amount of CO for the trsnsfcr to carbon], it is 92cjl,. No improvrmrnt in yicld was achicvrd in a reagent prrparation carried out under an atmosphcw of carbon monoxide rathcr than nitrog~n.The mechanism of this corivcrsion of C1C('03 (('0) to thc. (OC)gCo3CCO+ ion is of inhest, rspecially sincc, this rcagrnt is prcparativcly useful. The aluminum chloride is essential in order to obtain good product yields. Alkylidynetricobalt Nonocorbonyl Complexes 117 TABLE VIII SYNTHETICAPPLICATIONS OF THE (OC)9C03CCO+A1Clr-REAGENT' CH,OH CO,CH, 83 C,H,OH CO,C,H, 78 Me,CHOH CO,CHMe, 78 Me,COH CO,CMe, 77 CH,=CHCH,OH CO,CH,CH -CH, 66' HC ECCH,OH CO,CH,CECH 62' CCI,CH,OH CO,CH,CCI, 62' C,H,OH C02C8H5 66 p - CH,OC,H,OH CO, C, H,O C H, - p 39c Me,CSH C(O)SCMe, 51 PhSH C(0)SPh 58 NH, (g) C(O)NH, 64 PhNH, C(0)NHPh 76 Me,NH C(O)NMe, 55 Et,NH C(O)NEt, 75 LNH 60 n 64 ' "WNH (CH,),Sn 61 (C,H,),Sn 66 Me,N+@ 69 6 Fe 41 Et,SiH 63 ' Seyferth and Williams (34). Yields are based on the quantity of ClCCo3(CO)s charged. Unpublished work, C. L. Nivert. Chloromethylidynetricobalt nonacarbonyl docs react with alcohols, but only very slowly. Thus its rcaction with an excess of ethanol at room tem- perature for 15 days gave (OC)&03CC02Et,but only in 6% yield. [The bromo cluster is more reactive, giving (OC)~CO~CCO~MCin 59% yield on reaction with methanol at 53" (7,26) .] To obtain the high yields shown in 118 DIETMAR SEYFERTH Table 1'111, at least 2 (and preferably 3) moles of AICls per mole of C1CC'03(CO)g are required. The acylating agent is not formed instantane- ously; only whrn the purpltx to yellow-brown color change has taken place is the acylium ion species present in high yicld. When a mixture of 2 molar equivalents of Al("1a and 1 of ClCCo3 (CO) 9 in dichloromethanc was quenched with methanol immediately after mixing, no ester was formed and a high recovery of the chloro cluster was obtained. At thr presmt time, both the cbxact constitution of the final acylating agent and the mcchanisni of its formation remain unclear. The picture which wc favor is admittc>dly incomplete and very possibly subject to change. We suggest that initially a (OC)gCoaCC1.A1C13complex is formed in which substantial polarization of the C-C1 bond has occurred and which is capablc of Fricdrl-Crafts substitution on aromatic systems. Subsequent complexation of a second molccule of AlC'l, st>a carbon monoxide ligand [a known procms in nictal carbonyl chemistry (35, SS)] provides the activation for CO migration from cobalt to the elcctron-drficient apical carbon atom. This is not unrcasonablc sincc. it is known that in binuclear metal carbonyls t he bridging carbonyl ligands are stronger Lewis basic sitm than arc the trrmirial carbon monoxidcs (35, 36). In fact, aluminum alkyls haw bccn rcxportc>dto promote a terminal to bridging carbon monox- ide ligarid shift in a binuclear ruthcnium complex (37).Such a cobalt-to- carbon CO transft.r in our system, occurring eithw intra- or intrrmolecu- larly, would leave coordinativcly unsaturated cobalt atoms in cluster acylium ions which would require clfficient C'O transfer from other molecules in order to obtain the. (Oc')gCo,JXO+ species in high yield. We are still working in this area in the hop(. of achieving a better understanding of this process. We not(. that systems that give the cluster acylium ion more rapidly are available. Thus, BrCCo3(CO) is converted to the yellow-brown acyliuni ion solution in dichloromethanc by an excess of aluminum chloride at a faster rat(. than is ClCX'03(CO)9.This transformation of ClCCo3(C;O)9 can be accelerated by using a larger excess of aluminum chloride (10 molar equivalents of AlC13 rather than 2 or 3) or by carrying out the ClCCo3- (CO)9/3A1C13 reaction in thc presence of iodomethane. In the latter case, it appears that thc rate accclcration is duc. to the formation of the more reactive aluminum iodide rather than of ICCo3(CO) 9. The preparative utility of these systems has not yet bccn assessed in detail, but in all cases the yields of (OC)gCo3C"COzRon treatment of the solutions with methanol or ethanol were good. Robinson arid co-workers have claimed, without providing any experi- mental details, that boron trifluoride also converts BrCCoS(CO)g to the cluster acylium ion under a carbon monoxide atmosphere (38). They also Alkylidynetricobalt Nonacarbonyl Complexes 119 report that the ClCCog(CO)g/AlClg reaction can be carried out in situ in the presence of the iiucleophilr (o.g.,1120, EtOH, PhOH) (38). IV HIGHLY STABLE NONACARBONYL TRICOBALTCARBON-SUBSTITUTED CARBONIUM IONS Our discovery of the easily formed, very stable, and preparatively very useful (0C)gCo3CCO+ion snggclsted to us that carbonium ions of type (OC)9C03CCItR’+ niight provide another fruitful area of study. The elec- tronic effects of the (OC)gCo,C clustcr were riot at all clear and one could entertain the possibility thatj this organometallic substituent might stabil- ize an adjacent positive. charge. A suitable carboiiiuni ion precursor was necdcd in ordcr to carry out such an investigation and at that time no such (OC)9C03C derivativc w:ts available in useful amounts. Alcohols of type (OC) &oSCC (OH)RR’, as mcntioncd already, were available only in yields of 5% or less, and halidcs of typc (OC)gCo&C (X)RR’ were (and still are) unknown. The first problem, then, was to dcvthlop a more useful, high-yield alcohol synthesis. The reduction of a ketonc1 to a11 alcohol is a well-known organic reaction, and since ketones of typc (OC)9C03CC (0) R and the aldrhyde (OC) gCos CCHO could be preparcd in good yield, wc chose to cxaminc this route TABLE IX REDUC,TIONOF RC(O)CCO~(CO)~ TO RCH2CCoa(C0)9BY Et3SiH/CF3C02Ha R in I~C(O)CCO~(CO)~Et3SiH CFPCOzH RCHzCCo3(CO)g (mmole) (mmole) (mmole) (5% yield) 7 6 CZHSCCO~(CO)~(90) 7 6 n-C3H;CCo3(CO)g (92) 7 6 n-C4HyCCo3(CO)g (87) 7 6 n-CaHiiCCos(C0)~(80) 8 8 ~-C;I-I~,CCO~(CO)~(85) 8 9 cyclo-cG€111CI-12cco3(co)9(75) 5 6 (CH~)&HCH~CCO~(CO)~(81) 5 5 C~H,CHZCCO~(CO)~(82) 5.2 6 ~-CII~CJI,CH&CO~(CO)~(78) 7 6 p-BrCGHICH2CCo3(CO)~(67) (1 From Seyferth et al. (39). 120 DIETMAR SEYFERTH (39).The initial results were not encouraging. Attempted conversion of formylmethylidynetricobalt nonacarbonyl to the primary alcohol was un- successful. Treatment of this aldehyde with sodium borohydride in THF at reflux gave a mixture of CH3CCo3(CO)Yand HCCO~(CO)~,whereas re- action with lithium aluminum hydride or sodium borohydride in benzene gave only HCCo3(C0)9.In view of the stability of some of these cluster complexes to strong acids, we turned our attention to an acidic reducing system, tricthylsilane/trifluoroacetic acid, that had been developed by Russian workers [see Kursanov et al. (40)for a recent review]. This rc- duction proceeds by initial protonation of the carbonyl compound, followed by reduction, via hydride transfer from the silicon hydride, of the prot- onated species. If one of the carbonyl substituents can stabilize an adjacent positive charge (e.g., an aryl group), the reduction proceeds past the alcohol stage to give a hydrocarbon: 7' + R-C-OH, A R>CH, I R H A number of clustrr-substituted ketones and the aldehyde all reacted with the Et3SiH/CF3C02Hto give the alkyl derivatives rather than the expected alcohols: Et&lH/CFCOzH (OC)YCOICC (0)It + (OC) yCojCCH2R (27) The examples studicxd are given in Table IX. The product yields were ex- cellent (75-927,) and, in fact, this reaction is the best available procedure for the preparation of such alkyl derivatives. The R in Eq. (27) may be primary or srcondary alkyl or aryl; when R becomes more bulky (e.g., Me&, (OC)gCo3C) the reduction fails. Although this reaction did not rc- sult in the alcohols we rcquired for our furthrr studies, the results were of considrrable interest to us since they provided good indication that the ( OC)yC~3C-~~bstitutedcarbonium ions would be rather stable species. As mentioned, the complete reduction observed occurs only when at least one of the substituents on the ketone carbonyl function is capable of stabilizing an adjacent positive charge. In those cluster complexes in Table IX where R = H or alkyl, it must be the (OC)yCo3C substituent that is providing such stabilization. However, in order to pursue this question, we still re- quired (OC)9Co3CC (OH)RR' compounds in useful amounts. Hydrosilylation of ketones and aldehydes converts these to silyl ethers whose hydrolysis gives alcohols. Phenylsilanes were found to add to benzo- Alkylidynetricobalt Nonacarbonyl Complexes 121 phenone, but the reaction conditions (reaction temperatures of 220’-270°C) were not attractive (41).Catalyzed hydrosilylations proceeded under much milder conditions, e.g., zinc chloride (&), HzPtCl6.6H2O (43), (Ph3P)3- RhCl (44). The hydrosilylntion of our cluster-substituted ketoncs was found to occur under surprisingly mild conditions. It was sufficient to heat equimolar quantities of tric*thylsilanc and the ( OC)gCo3CC(0) R com- pounds in benzene at reflux, under a carbon monoxide atmosphcre for about 8 hours. The crude silyl cthw was converted to the alcohol by solution in concentrated sulfuric acid and subsequmt hydrolysis by pouring into an ice-water mixture: 0 It /I I (OC)~CO~C-CI?+ 1 :t 8111A (OC)yC03CCHOStlCt3 (28) I( It I +/ (OC)Y(’O,C(‘I 1oSIl’:t3 ~(0c)yCO3cc (29) \ I1 It It +/ I (OC),C~i,CC’ (OC)~C~~CCH--OH (30) \ [I It was found important to carry out the hydrosilylation under an atmos- phere of carbon monoxidc in ordw to obtain the good yields given in Table X for a number of these reactions. Thc quc.stion remains why these hydro- silylation reactions proceed so rcmlily under mild conditions in the absence of a catalyst. One possibility vhich we considered was that the cluster ketone provided its omii catalyst by way of minor dccomposition to give mononuclear cobalt carbonyl intermc.diatcs which could bc thc actual cata- lytic species. However, RCCo,(CO), were not found to catalyze the hydro- silylation of cyclohcxanone. A second possibility is that the C=O bond in these (OC)gCo3CC (0)R compounds is exceptionally reactive. Table XI lists the ketonic C=O stwtching frequencies of some of the (0C)9c103- CC (0)R compounds that n-c’ have prepared. These are found to be in the range of 1560 to 1645 cm-l, xt~much low-cr frequency than in dialkyl ketoncs (1725-1705 cm-l) or aryl kctoncs (1660-1700 cm-l), and this implies greater carbon-oxygen bond polarization : + (oc)gC(i~<‘-(’-It (oc)yC<&-(;-It (31) I1 I 0 0- 122 DIETMAR SEYFERTH TABLE X REDUCTIONOF RC(O)CCO~(CO)~TO RCH(OH)CCO~(CO)~~ R in RC(O)CCo3(CO)g EtsSiH RCH(OH)CCoa(CO)g (mmole) (mmole) (% yield) 17.5 CH3CH(OH)CCoa(CO)g (84) 20.8 (90) 3.2 C~H~CH(OH)CCO~(CO)~(73) 3.2 ~-C~H~CH(OH)CCO~(CO)g (75) 3.4 n-C,HgCII(OH)CCo3(CO)g (81) 3.4 ~-C~H~~CH(OH)CCO~(CO)~(81) 3.0 (CH~)~CHCH(OH)CCO~(CO)~(80) 3.4 CYCIO-C~H,ICH(OH)CCO~(CO)~(52) 87.5 Ce,HsCH(OH)CCo3(CO)g(87) 3.4 P-CH~C,H&H(OH)CCO~(CO)~(70) 3.4 p-BrCsH4CH(OH)CCo3(C0)9(84) 31.4 IIOCH&Co3(CO)g (46) a From Seyferth et al. (,99). This should facilitatc attack by the silicon hydride, Et3Si6+-H6-, if a polar mechanism is operative. With the (OC)9C03CC (OH) RR’ alcohols available, experiments whose goal was the generation of cluster-substituted carbonium ions were now possible. Our initial cxpcrimrnts were carried out with (OC)9Co3C- CH(OH)R, where It = H, CH3 and C6H5(45). Thc procedure used in the conversion of (OC)&o3CC02Et to (OC)yCo3CCO+PF6- was found to be applicable to cluster-carbonium ion synthesis. Thus, treatment of (OC)gCo3CCH (OH)CH3 in propionic anhydridt. solution with a small cx- cess of 657, aqueous hexafluorophosphoric acid under nitrogen resulted in precipitation of a black solid whose analysis after several washes with dichloromethanc indicated the constitution (OC)gCo3CCHCH3+PFti-. Similar reactions gave (OC)9C03CCH2+PFti- arid (OC)&o3CCHPh+PFe-. These salts were quite stable in the absence of air and moisture. Reactions at carbon were observed with alcohols, a thiol, aniline, and N,N-dimethyl- aniline with all three salts, e.g., (OC)gCo&CH2+PF,- + CH3OR -(OC)~CO?CC€IJ~CH~ + H+PFC (32) An exception was the reaction of (OC)9C03CCHCH3+PF6- with N ,N-di- methylaniline, in which the basic substrate abstracted a proton to give the vinyl-substituted cluster. The results of these reactions are summarized in Table XII. The ( OC)gCo3CCHR+PFti-salts, like the cluster-substituted acylium ion, are rather weakly elcctrophilic. Although they alkylatc N,N- Alkylidynetricobalt Nonacarbonyl Complexes 123 TABLE XI FREQUENCIESOF KETONIC CARBONYL VIBRATIONl3 IN (OC)&o&C(O)R COMPOUNDSO R in (OC),Co,CC(O)R u (c=O) (cm-', in CCL) 1625' 1640 1645 1640 1635 1618 1635 1610 1611 1610 1582 1586 1555 @H Id 1573 aH Fe 1590 a From Seyferth et al. (9). In CHC1,. *Williams (32). Seyferth (27). dimethylaniline when R = H or Ph, neither of these salts will react with anisole. It seems clear that clustcr-substituted carbonium ions are easily gen- eratc.d, are very stable thermally, and are of sufficient electrophilic reac- tivity to be useful in the synthesis of many new functional cluster com- pounds. Before we consider further their exceptional stability and the mode of their stabilization, it is of interest to mention some chemical con- sequences of the high stability of a positive charge generated a to the apical carbon atom of the (OC)sCosCcluster. For instance, one might expect the addition of the positive part of an ionic or polar reagent, X+Y-, to the 124 DIETMAR SEYFERTH TABLE XI1 ORCANOCOBALT CLUSTER nERIV.YrIVES PREPARED FROM THE (OC)9C03CCH1L+PFe-S ILTS~ ~ ~~ It in (OC)&03CCHR+PFG- Reachnt Product (% yield) H MeOII (OC)9C03CCH20Me (85) H EtOTI (oc)9co,ccezoEt(76) H PhNHz (OC)&o3CCHZNHPh (67) H CGIfaNMez (OC)~C~~CCHZCG€I,NR~~-~(49) Ale MeOII (OC)gCo3CCH(Ale)OAIe(83) Me E t,O €I (OC)&o3CCII(Ale)OEt (82) Me PhSH (OC)gCoaCCII(Me)SPh (42) Me PhNHz (OC)&03CCEI(Me)NHPh(73) hZe CaHBNAfe2 (OC)gCo&CH=CHz (68) Ph MeOII (OC)sCo3CCH(Ph)OMe (59) Ph EtOH (OC)gCoaCCH(Ph)OEt (73) Ph PhSH (OC)~CO~CCII(P~)SP~(38) Ph PhNHt (OC)sCo3CCII(PhINHPh (59) Ph CsH5NMe2 (0C)&o3CCII(Ph)C&NMez-p (54) a From Seyferth et al. (45). terminal carbon atom of a C=C substitueiit attachcd to the apical carbon of the (0C)ycoae cluster to bc a rathcr favorablc process, so that the direction of addition would bc I/ II (Oc)gCOjCc=C + >;+y-_t ((~<‘i,CO,C~(I--C--S,Y- \ +I II -(oc)9(’0,~~--c-c-X (33) I1 Y All of the reactions of such vinylic derivatives that we have studied proceed in this rnaniicr (46).This rcaction is of spclcial utility in the generation of tertiary cluster-substituted carbonium ions: (EtCO)20 (OC)&o,CC=CH> -tTT’l’Fe- -(OC)~C(~~CC((’H,)L’~’~~:B- (34) I (‘Ha The carbonium ion salt obtained in this manncr rcactcd with methanol to give the methyl ether, (OC)&o&C (CHI)20CH3, in 86% yield and with aniline to produce (OC)9C03CC (CHI) zNHC61-Is in 49y0 yield. At,tempted purification of the methyl cthcr by chromatography on pH4 silicic acid re- Alkylidynetricobalt Nonacarbonyl Complexes 125 sultcd in formation of the alcohol, (OC)9C03CC (CH,) LIH, which is a furthcr indication of the easy accessibility of the tertiary carbonium ion. Reduction of isopropenyl-substituted cluster, by way of the carbonium ion, to (0C)&oJCCH (CH,) could be effected with zinc amalgam in tri- fluoroacctic or concentrated hydrochloric acid. Similar protonation of (OC)&'oJCCH=CH2 ith HPV6/ ( EtCO) *O gave the (OC)&'03CHCH3+- PF6- salt which we had prepard previously from the alcohol. In the case of ( 0<:)SCo3CCH=CHSilll(.3, protonation resulted in drsilylation : + ~IeJSi('II=C11('Co3(CO)s + H2S04- I\lejSiCII,-(:€I~CCoJ(~O)s, HSOi- - OH Sincc1 olrfin dimination of this tjpc from a species with a positive charge p to a silyl group is well lziov ti (.i?'),this result was iiot uiwxpcLctcd. Elect ropliilic l+idel-Craft s acylation of (0C)&03CCH=CH2 with acetyl chloridc/aluiniiiuni cliloricle under a carbon monoxidc atmosphere gave thrce products (all of which had the acctyl group attached to the terminal carbon atom of thc vinyl group, as cxpectcd) : (OC)gCo3CCH2- CH2C (0)CH3 (13(x), tr airs- (O(')&'o?CCH=CHC (0)CH3 (127,), and (OC)~Co~CCH(OH)CH2C(0)('H3(Gf,;). All thrw arc drrivablc from the initially fornicd cation, (Or)y( 'oJC('HCH2C (0)CH3+, through reduction, proton loss, and hydrolysis, rclspectivc.ly. Similar addition of CH3CO+A1C14- to the isopropciiiyl-suhstitutcd c,luster gave (OC)yCosC(CH3) HCH2C (0)- CH3. The oxymercuration of allylid~iietricobaltnonacarbonyl also pro- ceeded in the direction expcctcd : + (:ir2=~~tIc(~0j(~'o~g+ (q1cjumg+ -C'F3c~~Lirg(:IiL('Ir(:Co109 lCII&H C'FjCOLHgC HLC I ICCOj(C'0) g (36) I O('Ii3 The radical-initiated addition CJf CBrC'I:, or CBr4 to olefiris followed by base-induced dchydrobrominatiori of the adducts can be uwd to prepare a wide variety of RH'C=C (R")CX3-type compounds mhosc. reaction with dicobalt octacarbonyl will givcl the rcspectivc vinylic cluster derivatives. The reactions of the lattcr nitl: appropriate reagents will extend further the number and types of organo-functional cluster complexes. Also, their 126 DIETMAR SEYFERTH catalytic hydrogenation or their rcduction with systems that generate HCo (CO) 4 can make available as well simple or functional alkyl deriva- tives that are difficult or impossible to prepare by other routes. It may be noted that two examplcs of thc catalytic hydrogenation of unsaturated cluster complexes haw bcen reported : H*/CO (260 atm) (OC)&o$XII=CI-ICO,H t (OC)gCo&CH,CH,C02H (37) [co2(Co)s].IlOOC (18) €I? (-10 PSI) CH, (The 2-mcthylallylidynctricobalt nonacarbonyl used in Eq. (38) was pre- pared by the novel reaction of dimethylketene with dicobalt octacarbonyl.) The stabilization of a positive charge on a carbon atom a to a (OC)9C03C cluster appears to play an important rolc in the chemistry of aryl-substi- tuted cluster complcxc>s.During an investigation of the chemical trans- formations of brnzylidyrietricobalt nonacarbonyl complexes, we discovered that C&CCo3 (CO)9 and its 0- and m-mcthyl and chloro derivatives roact with acetyl chloride/aluminum chloride to give p-acetylated cluster com- plexes in good yield (Scheme 1). Similar Friedel-Crafts acylation was ob- served with the benzoyl chloride/aluminum chloride reagent, but not in the case of the C1-substituted cluster complcxes (49). Formylation of C6H5CCo3(CO) 9 in the para position in low yield could be achieved with CH30CHClZ/AlC13. The acetylation reactions, in particular, proceeded rapidly in high yield and under rathcr mild conditions. In contrast to the 0 CH3COC1 (Oc)9C03Co! -CH3 93% I C,H,COCl (0C),Co3C 71% &CH,OCHCl, (OC),CO,C 01-H 21% Scheme 1 Alkylidynetricobalt Nonacarbonyl Complexes 127 high reactivity of C6HjC(('03(('0)9, thc benzyl-substituted complex, C6HjCHZCCo3(CO)y, did not npp(m-to react with CH,COCl/AlC13. These reactions all were complicated hy the ability of the Lewis acid catalyst to coordinate to carbon monoxide ligands of the benzylidynetricobalt nona- carbonyls. Thcreforc, the rclaction conditions used were of critical impor- tanccb to the succcss or failurv of thc rcaction studird (496). Thcsc, reactions reprcscmt anothcr route to organo-functional cluster compl(1xcs whose scope of application, however, is rather limited. Ncvcr- thclcss, we were struck by thci facility of these rractions in those cases whcre they proceeded in high yidd, in particular, the acctylation of C6H5CCo3(CO)9and o- and p-('H3('6H&Co3(C0)9. In order to obtain a morr quantitative estimate of thrir rractivity, we carried out competition experiments in which a mixturc of ('6H5CC03( CO) 9 and anothcr aromatic substrate was allowed to rract nith a deficiency of the CH$OCl/AIC13 reagent in dichloromcthanci. The, high reactivity of benzylidynetricobalt nonacarbonyl was irnmrdiatcly apparent when it was found that in its compctition rcaction nith anisolc, a rather nucleophilic benzene deriva- tive, no acetglation product of tht. latter was formed. In a competition experiment 1% ith N ,h'-dinwth? lanilinc, a more potent nuclcophilc, the acctylation product yirld5 cst ablishcd that k[PhCCos (CO)9]/ k(CGHjNMe2) for reaction n ith ("H3COC1/AlC13under the conditions used is 1.3. A similar competition cy)cirinirnt with ferrocenc, known as a super- nucleophile in the Fricdcl-Crafts reaction (50),showed C6HjCCo3(CO)9 approximately as reactive as this orgarloiron complex { k[PhCCo3( CO)9]/ k(fcrrocene) values of 0.9-1.2, dcpcnding on reagent ratios and reaction times, assuming also in thc calculations that fcrroccne has ten reactive positions to only one in bcnzglidyrietricobalt nonacarbonyl) (49). Thus benzylidynctricobalt nonacarhonyl is one of the most, if not the most, reactive monosubstitutc,d hizriirs in the Friedel-Crafts reaction. This observation may be rationalizcd in terms of the possibilities for the stabilization of the charged iiitcmicdiatcs in such reactions. The attack of the CH3COCl/AlC13 rvagent (as ('H3CO+AlClI-) at the para position of CGHjCCo3(CO)9 would lead to an ionic intermediate in which the positive charge can be displaced to thr. ring carbon which is bonded to the apical carbon atom of the cluster, as shown in structure 111. Thus, further de- localization into the cluster would bc possible. The net stabilization gained must be considerable, in view of the high reactivity of benzylidynetri- cobalt nonacarbonyl. Ortho substitution in these systems would lead to a cluster-stabilized intermediate. (IV). Howevcr, the steric effect of the six carbon monoxide ligands disposed in the general direction of thc apical carbon atom and its substiturnt would bc expected to hinder or possibly completely prevent attack at an ortho position. In any case, meta substi- 128 DIETMAR SEYFERTH tution is not expected. As a result of thesc electronic and steric factors, reactions of ArCCo3((20) complcxes with the RCOCI/AICI3 reagents is limited to para-substitution, irrespective of electronic effects due to other substituent groups on thc benzene ring. Stabilization of the type shown in structure 111 is, of course, not possible in the caso of CeH:CH2CCo3(CO)a, which explains its lack of reactivity toward CH3COCI/AlCI3. 0 VCH3II P,-CH3 oc\ / \ /co 0% /c\ /co oc-co------co-co oc--0------co-co oc’ hC(f ‘co oc’ \I/ ‘co oc I co oc’yco co co (In) (N) The carbonyl stretching frequency of p-CH3C‘(0) CeH4CCo3 (C0) in dilute carbon tetrachloride solution was found to be 1685 cni-1. Fuson et al. (51) had found that a good linear correlation existed bctwcen Ham- mett u constants and v (C‘=O) of substitutcd acctophcnoncs in dilute CC12. From the position of thc v (C=O) of p-CH3C(0) CGH1CC(03 (CO) on their v (C=O)/u plot, we could estimate the u constant for the (OC)gCo3C: substitumt to be about -0.35. That this substituerit is an electron donor with respect to a p-acctgl group in an attached benzene ring is not unex- pectcd since charge dclocalizatiori into t hc clustclr via complex V should bc possible. Other observations, which WP and others have made during the course of investigations of the chemistry of alkylidync.tricobalt nonacarbonyl complcxcs, may have been conscqucnces of the high stability of a positive charge generatcxd at a carbon atom a to the (0C)9C03Ccluster. For in- stance, uncxpecttbd reduction of certain functional groups during organo- cobalt cluster preparations when thesc would be expected to appear on the a-carbon atom of the resulting complrx could be a result of the generation of acidic conditions during work-up or even during the reaction itself. This might result in formation of a cluster-substituted carbonium ion whose ultimate fate (reduction, hydrolysis, destruction of the cluster) will be Alkylidynetricobalt Nonacarbonyl Complexes 129 dictated by thv further reaction and work-up conditionq. In all such coii- siderations, it must be kept in mind that HCo (CO)4 could be formed dur- ing rclaction or \vork-up and that this compound, in aqueous medium, is as strong an acid as HC1 (52). The fact that all attempts thus far to prcparcl alk3’lidynctricobalt nona- carbonyl complcxcs with an a-halopn suhstitucnt haw been unsuccessful may also bc a rmult of the oasc~of forrnatioii of clustcir-substituted car- bonium ions. Evc~ivrry inild proccdurcs, c.g., the action of lithium chloridc 011 (OC)SCo3(’CH (CH3) 0802(’gH1(’H~p in THF, resulted in decoinposi- tion of the cluster (39). Posiibly thv cluster carbonium ion is generated and is dwtroycd by way of halitlc ion attack at cobalt, which, as we show in the following discussion, niiiit l)(lar a substantial portion of thr positive charge. Aiiothcr rcaction suggwt iirg that the carbon atom at which C-0 bond lictcmlysis in the alcohol lint1 occurrcd was not the only sit(. of elrc- trophilic reactivity in thcw cnrl)oniuni ions was that with triphcnylphos- phiiic. One might have cxxpc tl this nuclcophilc to form stable phosphon- iuiii salts by rcactioii with tho carbonium ion hexafluorophosphatcs at carbon, but its addition to R slurry of such a salt in dichloromc$hanc rrsultcd in vigorous gas cvolutioii md complete decomposition. Thc carboniuiii ion saltq that had prepared and studird could pos- sibly be pictuwd as structuw TI, hut this rcprescntation do~snot seem adcquate in view of thcir (’as(’ of formation, high stability, arid rathw low reactivity. These propc1rtic.s suggested that cxtcnsive dclocalization of the positivr charge originally gencr:ttcd at t hc carbon atom had occurrc.d. Such delocalization, hoivcver, could irot he accominodatcxd by conventional in- ductive and rcwnancc, (+f(’cti. To obtain mow information on the question of structure and bonding of t hcw novd organomc~talliccarbonium ions, we began a study of thcir NXllt qwctra (53).This investigation is not yet completed, but the prdimiiiary rcwlts providc intcrcsting and uscful information. (VIa) R = H (VIb) R = CH, (VIc) R = C,H, Table XI11 summarizes proton chemical shift data for the thrre car- bonium ion salts (VIa, b, and c) and for the alcohols from which they were prepared by trcatmeiit with conccritrated sulfuric acid or trifluoroacetic DIETMAR SEYFERTH acid. It can be seen that the hydrogcn atoms on the carbon CY to the (OC)9C03C cluster become less shielded on going from the alcohols to the carbonium ions, but the shifts, As, are not at all large, compared to those observed in other less stabilized systems [e.g., Mc2CHOH to Me2CH+, A6 = 8.5 ppm (54)]. The 13C NMIL spectra of compounds VIa, b, and c were more informa- tive. The chemical shifts of carbon resonances in 13C NMR spectra have been used as a measure of the electron densities of the carbon atoms being studied (55)) and so a comparison of the carhinyl carbon resonances of compounds VIa, b, and c with thosp of the alcohols from which they were TABLE XI11 PROTON NUCLEARMAGNETIC 1bX5ON.INCE SPECTR.4 OF (OC)gCo3CCH (R)OII AND (OC),Co3CCHR+X-a~ Alcohol, 6 Cation, 6 (solvent) (PPd (PPd As (PPd (OC)&o&CHzOH CHz 5.21 -0.6 -0.5 (OC)&O~CCH(CHI)OH (OC)~CO~CCHCH 3+ CH 5.4 6.7 (HzSO4) -1.3 6.9 (CF3COzH) -1.5 6.9 (PF6- salt in CH~NOZ) -1.5 CH3 1.8 2.4 (HzSOI) -0.6 2.5 (CF3CO2H) -0.7 2.4 (PI?,- salt in CHINO,) -0.6 (OC)gCo3CCH(CsII,)OH (OC)oCo3CCHCsI€jt CH 6.2 7.7 (HzSO4) -1.5 8.2 (CFsCOzH) -2.0 CsHs 7.4 7.2 (HzSO4) +0.2 7.6 (CF~COZH) -0.2 From Seyferth et al. (53). Alcohol spectra were obtained in chloroform-d and are referenced to internal tetramethylsilane. Cation spectra are referenced to tetramethylsilane contained in a capillary inside the NMR tube. The methyl group signal for the methyl-substituted carbonium ions (in nitromethane, sulfuric acid, and trifluoroacetic acid) was a doublet (J = 7 Hz). The methyne proton for these same carbonium ions appears as a quartet (J = 7 Hz). For the phenyl-substituted and unsubstituted carbonium ions, the methyne proton and methylene proton signals (respectively) were singlets. Alkylidynetricobalt Nonacarbonyl Complexes 131 derived should give an indication of tht. dcgrccx of charge delocalization in thcsc carbonium ions. The data obtaincd are given in Tablc XIV. Notc- worthy is horn small the changw in th(1 carbinyl carbon atom chemical shifts are on going from thv alrohol to the. carbonium ion. By way of com- parison, A6 for the carbinyl carh~iiatom when Mc2CHOH is converted to hfe2CH+SbF6- is 233.3 ppni (Lj). A comparison with similar clat a for :in alcohol/carboniuni ion sct where the latter is highly stabilized n ould bc of interest. Fcrrocenc chemistry provides such examples. Thc c~xtraordinarilyhigh stability of fcrrocenyl- methyl carbonium ions is wll docunic~ntedand for some years has bcm the subject of much discussioii and some coritrovcrsy (56). We measured thc 13C NMR spectra of fcrrocc,iiylcarbirlol, 1-fcrrocenylethanol and ferro- cenylphcri~lcarbirioland of thvir dcrivckd carbonium ions (VIIa, b, and c) obtained by dissolving the alcohols in conccntratcd sulfuric acid (57).The A6 values for each are shown with the structures. They arc slightly larger than thosc observed for the an:dogous cluster-substituted carbonium ions and so the latter are certainly of comparable stability to the fcrrocenyl- methyl carbonium ions and pcdiaps cv(w more stable. Fe Fe Fe A6 =-26.2 ppm A6 =-52.1 ppm =-49.4 ppm (VIIa) (VIIb) (VIIC) We note upon inspection of the data in Table XIV that the chemical shifts in these carbonium ions (VIa, b, and c) follow the normal order ob- served when substituents on a fully substituted carbon atom are varied from R = H (most shielded) to CH3 to CeHj (least shielded), rather than for substitucnts on an clectron-dcificicnt trivalent carbon atom, where the CH3-substituted carbon atom is less shidded than the analogous phenyl- substitutcd carbon atom (58). This is a rather clcar-cut indication that the carbinyl carbon atoms in conip1rxc.s irIa, b, and c are nearly fully bonded and only slightly clcctron-dt.ficic.iit. The slight increase in shielding of the CH, carbon when VIb is formcd from the alcohol and of the C-1 carbon of the phenyl group of VIc provides further confirmation of this. Thr data in Tablt. XIV suggc3st that thc positive charge in complexes VIa, b, and c has bcrn ddocalizcld to a large extent into the cluster sub- stitucnt. The observcd slight iricrcasc in shidding of the carbon atoms of the carbon inonoxidc ligands when the alcohols are converted to thc car- bonium ions speak in favor of this view. If the cobalt atoms are more 132 DIETMAR SEYFERTH TABLE XIV C.ZRBON-13 NUCLEARMAGNETIC ItESONANCE SPECTRA FOR (OC)gCoaCC€I(R)OH AND (OC)&O~CCHI~+HSO~-a-* Alcohol (ppm) Cation (ppm) A (PPk) (OC)9Co3CCI120H (OC)gCo3CCHz+HSO4- Carbinyl carbon 77.6 91.1 -13.5 CEO 200.6 192.7 +7.9 (OC)~CO~CCHCH~+€ISO~- Carbinyl carbon 82.5 119.9 -37.4 CH3 28.5 26.2 +2.3 CEO 200.5 193.2 f7.2 (OC)~CO~CCH (CgH s)+HSOr- Carbinyl carbon 88.8 124.5 -35.7 Ph-Ci 146.2 135.4 +10.8 Ph-Cz 128.8 129.9 -1.1 Ph-C3 126.8 129.3 -2.5 Ph-Cd 128.6 132.6 -4.0 C=O 200.0 192.4 +7.6 a From Seyferth et al. (55). b Alcohol spectra are referenced to internal tetramethylsilane. Cation spectra are referenced to external tetramethylsilane through the lgFlock signal. All cation spectra were obtained in concentrated sulfuric acid as solvent. All spectra were proton decoupled. electron-deficient in VIa, b, aiid c than in the corresponding alcohols, thcn the carbon monoxide ligands would be expected to be bonded more tightly in the cations. Thc consequent niovemcnt of thc CO carbon atoms to a position closer to the cobalt atoms might be expectcld to result in in- creased shielding due to the diamagnetic anisotropic shiclding effect of the cobalt atom. In our early efforts (53) we had becn unable to see a 13C signal of the apical carbon atom in the alcohols arid the carboiiium ions. In inore recent work, we have located the I3C signals due to the apical carbon atom in the carbonium ions in thc region 6C 255-275 ppm [VIa, 286.2 ppm; VIb, 273.5 ppm; VIc, 267.0 ppm; (OC)9C03CCh/Ic2+, 257.8 ppm; (OC)gCo&XgH19-n+, 273.3 ppm (59)]. Most of the alcohols from which thesc carbonium ions were derived were not soluble enough to mablc the apical carbon atom Alkylidynetricobalt Nonacarbonyl Complexes 133 rcsonancc to tw seen. Thc 1ong-ch:iin alcohol, (OC)gCo3CC'H (OH)CgHI9-n, however, was sufficiently solublc in CH2('I2,arid for its apical carbon atom 6c = 306.0 ppm. Thc 13C PU'hIli specxtrum of a mat sainple of the mter (0C)9C:03CC02Et above its mclting point of 46" showid 6c = 25S.4 ppm for its apical carbon atom. Thrscl 13C NhIR shifts are far downfic.ld from those of most carbon atoms in organic and orgaiionictallic cornpouiids with the exception of alkyl-substitutcd carbonium ions (GO), sonic transition metal carbenc complexes [cg., (O(')j<'rC*(OILle)Ph, 6c* = 351.4 ppm (61) and (OC)~WC*(OEt)C=('I'h, 6r* = 286.1 ppin (fit?)] and carbyne complexes [e.g., CH3C*=W(C'0) J'l, 6c* = 2883.8 ppm (63) and C6H,C= C-C*=W(CO)4Br, 6c* = 230.6 ppni (62)].In our previous review on RCCoJ(C'0)g complexes (1) v P .;uggcTstcd that thew compounds could, in principle, be regarded as adducts of a c*arbyncintrrmr.diatc, RC:, aid the Co3(('0)9 unit, i.e., as triply britlgcd cnrhync coniplrxrs. The similarity of thc I3C NMR chemical shifts of the apical carbon atom in R<'Co3(C'O)g complexw and thosc. of the car1)giic~carbon atoms in thr iioiihridged car- byne complexes of type ItC=RI (CO) 3,which were first rc.portcd by Fischer in 1973 (CS), is striking and vvr> likrly not coincidental. How then is the positive charges in thcs (OC)gC'03C-suhPtitut(d carboiiiuni ions delocalizcd into thc c1ustc.r sul)rtituc.iit '? In our first prcliniiriary rcprt on this ncu class of carboniuin ioris (45), \\ (1 suggested that thvir structure presents an cspc.cially favorabhl opportimity for U-P ~iypcircoi~jugation.It is this type of bonding, lateral ovcd:\p of n filld ~-lmidingorbital of a inctal- carbon bond arid a vacant p orbital on an clcctron-dcficic.nt carboniuni ion center p to the metal (VIII), that is bctlicvcd rwponsihlc for the high (VIII) stability of carbonium ion ceritcw p to both main group and trailsition metal atoms (64). In our clustrr-siil~stifUt(idcarbonium ions, such stabili- zation could make substantial contribution to their stability. It should, howcver, be recognized that thc ItC:-C'-Co bond angle in the rwutral clusters of about 131" is too largcl to allow very eflcctivv latc.ra1 overlap of thr type shown in complex VIII. To acliic.vc lwtter ovcdap, thc suhstitmnt at the apical carbon atom in thv car1)oiiiuin ion might bend down touard the cobalt atom, as shown in coinpoui~iIX. With 3 cobalt atoms giving equally good opportunity for suc.1~u-P ovvrlap, a wry stable species, with 134 DIETMAR SEYFERTH a high concentration of positive charge at the cobalt atoms would be cx- pected. Thcre also is a good possibility that carbonium ions of type IX would be fluxional species. An alternate way of regarding these carbonium ions, (X) suggests itself when one considcrs that the 13C NMR data for the a-carbon atoms in complexes VIa, b, and c show trends typical of olefinic carbon atoms (39). A structure of this type (XI) was suggested for the protonation product of the unsaturated osmium cluster (XII) by Deeming et al. (65). The proton NMR sprctrum of complex XI showed two high field singlets of relative intensity 12 in CDCly-CF3C02H at - 10" and this required an unsymmetrical structure. H H The a-carbon atom resonances in the I3C NNIR spectra of the carbonium ions VIa, b, and c fall within the range reported for the sp2 carbon atoms of olefins complexed to transition metal centers (66), but those of the apical carbon atom in these ions are far downfield from this range. In the case of the n-CgH19-substitutedalcohol and carbonium ion cluster systems, ASc on going from the alcohol to the carbonium ion is only 33 ppm and the SC values for all of the apical carbon atoms in the carbonium ion systems which we have studied are still within the range of SC values for apical carbon atoms in iwutral RCCO~(CO)~complexes. Thus far we have no spectroscopic cvidencc against the symmrtrical arrangement shown in structure VI. However, intuition suggmts that an unsymmetrical arrange- ment (IX or X) is more rcasonablc since it would optimize bonding in these carbonium ions. This question is receiving further attention. Alkylidynetricobalt Nonacarbonyl Complexes 135 V DECOMPOSITION REACTIONS AND DERIVED SYNTHETIC APPLICATIONS OF ALKYLIDYNETRICOBALT NONACARBONYL COMPLEXES One might hope that some of the transformations dcscribcd in this re- view could be useful in synthetic chemistry outside thc area of cluster chemistry. This brings up the qudoii of how best to rclcase the apical carbon atom and its substituent from tlw complex. Sincc the alkylidynctri- cobalt nonacarbonyls can be vicincd as complexes of a “carbyne,” Re:, with thc CO~(CO)~moiety, it is of intcrchst to see if thcir controlled decom- position could give. an intrrceptahlc carhyne fragment. Fischcr’s ncwly dis- covered carbynr complexes of typ(’ RC‘-M (CO)4X (X = halogen, RI = Cr, Mo, W) dccomposc to givc an acctylene, RC-CR, as the organic product (67).It is, therefore, n vcry intriguing fact that thc decomposition of some ItCCoa ((33) complexcis gives acctylenes (68) or acctylenrdicobalt hcxacarbonyl complexes (I, 69) : dlglynir, reflux Ph CCo j (CO)y - C’o iiietal + some l’hC--Cl’h (39) MeOII, rcflux ArCCor(CO) (40) The preparative scqucnce shown in Scheme 2, bawd on the latter reaction, Ph,Hg + HCCo,(CO), CH COCl or C,H,CCo,(CO) p - CH,C(O) C,H, CCo,(CO), -AlC1, PhCC1, + Co,(CO), MeOH, i70”C, 24 hr (0- CH,C(O)C,H,C,C,H,C(O)CH,-~ )co,(co), 39% yield p- CH,C(O) C,H, C=CC,H,C(O) CH,- p Scheme 2 was carried out in our laboratories (1).However, we know of no successful trapping of the Re: fragmcnt with an external reagent. Oxidation of two RCCo3 (CO) complexes with a limited amount of ceric ammonium nitrate 136 DIETMAR SEYFERTH in ethanol gave substantial quantities of acetylenic product (68) : EtOII, room lJhCII,CCoa(CO)g + tCe(1V) * l’hCH,CECCH?Ph + I’hCII,COJCt + temp (3 parts) (2 parts) I’hCII,CHO (41) (1 part) EtOfI, reflux 1’1iCCo3(CO)g + ICe(1V) * I’hCECPh + l’llCO>Et (42) 25% However, as we havr shown previously (5, 49a), when an excess of ceric ammonium nitrate in aqueous acetone is used, oxidation of RCCO~(CO)~ complexes proceeds cleanly to the carboxylic acid, RC02H. Some substi- tuted ArCCo3 (CO) wrre resistant to such complete oxidation by Ce (IV) , and in those cases potassium permanganate in aqueous acetone was found to produce the carboxylic acids in 50-60% yields (5).In one of the earliest reports on RCCO~(CO)~complrxcs (3a),their oxidation to thc carboxylic acid \I ith hydrogen pcmxidc was mentioned. An interesting oxidative degradation of RCCo3(CO) 9 complexes was reported by Japanese workers (70): It’OII, 70°C, 02bubbled in ItC‘Co,(CO)g * ltCII(C0LTt’)i + CO&~ (43) 40-80% This reaction has potmtial for useful application in organic synthesis. Oxidation of alkylidyrictricobalt rionacarbonyl complexes with halogens was the key reaction that led Kruerkc and Hubel to the correct assignment of their structure (3a): CClr I’hCI12CCo3(CX))9+ 13rL -I’hCI-12C13r~ + 3CoBrz + 9CO (44) The action of iodine also causes nearly quantitative release of carbon monoxide from such complexes, but the organic triiodides presumably produced wre not isolated (7). A few other reactions of these cluster complexes in which the Re: frag- ment is converted to organic products have bteri reported and are sum- marized as follolvs : CH3CCoa(CO)g [The action of methanolic sodium hydroxide on RCCo3(CO)9 compounds Alkylidynetricobalt Nonacarbonyl Complexes 137 had been shown previously to result in formation of the Co(C0)d- ion (2, 711. Hz, CO (elevated P) CI13CCoa(CO)g * CLHSCHO (70) (46) c61iS 50% n-BurP, MeOH CH3CCo3(C0)g C2HsCOsMe+ 100 atm CO 14% C€IdCH(CO,Me)L+ (CHLCO2Me)? (70) (47) 61 % 5% DhfE PhCH2CCo3(CO)o+ excess ?;a1311i -I’hCH2CH3 (68) (48) H20 (OC)9~03~~~,~~,~~i2~ --t I ro,c (CI r2),coLH(7 I) (49) IIO2CCH&II=CI€CO&1 (?f) (50) The limited number of reactions in which the Ilc: fragment is converted to organic products, which has heen studied thus far, consists of rather simple ones. This area of RCCo3(CO)gconversions merits further detailed exploration on a more sophisticatcd level since its development will provide the basis for synthetic applications that will make these complexes useful to others. Catalytic applications of KCC‘o3 (CO) complexes have been sought, and two such reactions have been reported. It was found that disubstituted acetylenes are trimerized to hc.xasubstituted benzenes when heated in the presence of alkylidynctricobalt nonacarbonyls (72) : lBO°C, 35 hr l’llCCo~(cO)g+ CH,C‘=CCTI, * (CK)GCG (51) hexane 20 % 160°C, 14 lir CHICCoa(C0)g + I’hC~Cl’h Ph&6 (52) 72% The RCCo3 (CO) “catalysts” in t hme reactions, however, only appear to be sources of simpler cobalt carhonyl species through their thermolysis, and it is the latter which cause the acetylene trimerization through well- established routes (73).Indccd, in some cases acetylene-derived cobalt carbonyl complexes are obtainrd (72): IOOT, 4 hr (OC)gCo3CC0,Me+ CIT,C-C‘H c (CH,CLH)COL(C~)~ (53) Various RCCO~(CO)~complexes have been shown to be initiators of olefin polymerization (74, 75).With the complexes in which R = H, C1, 138 DIETMAR SEYFERTH or Br, one might speculate that the initiating process involves reaction at the apical carbon atom, in vicw of the kiiown HCCo3 (CO) g/olefin reactions (16, 17) and reported reactions of BrCCo3(CO)g with olcfins (l7),e.g., 131("Coj(C'O)g + (";TI4 -t C'IH cwdmr=r(m3)cro3(co) (54) 1hcco3(Co)9+ ('IIL=C:HCO;l\lr --t (O(')9Col('('11LCII,CCoLlle (35) However, the fact that complexes in 11 hich It = Ph, F, NczCH, and CzFj also initiatc polymerization of acrylonitril(. suggests that chr3mistry at cobalt may be involvvd instcad. The displaccnient of carbon monoxide ligaiids by diolefins to givc stablc complcxcs of typc BCCo3 (CO)7 (nor- bornadienc) in the case of norbornadienc. has been rcported (18, 76), so that such a proccss is entirely possible. VI CONCLUDING REMARKS This revicw cannot bc cxhaustivc in vicw of space limitations. One would like to discuss other rcactions of alk~lidynctricobaltnonacarbonyl com- plexes occurring at the apical carbon atom, such as the remarkable con- version of CICCo3 (CO)9 to (OC) 9('oS('-C('o3 (C'O) by its reaction with triphenylarsine at 100" (77), thr rcwtion of C'lCCo3(CO)g with cobalt tetracarbonyl anion in THF at 70"C', 1% hich gives the [Co6 (CO)dZIz- carbidc cluster complcx (78),thc Ullmann-type synthesis of brnzylidyne- tricobalt nonacarbonyls from BrCCoj (C'O)9, which unfortunately gives only moderate yiclds (5),arid others. Completcly omittod from this re- view have been the discussion of chemistry at cobalt in these complcxes and more detailed consideration of the spectroscopic, mass spectrometric, electrochemical, and structural studics of the RCC'o3( CO j9 complexes that havc been reported. Howevcr, these topics haw been covered adequately in reviews by others (79, 80). We have mentioned only in passing other cluster complexes in which a tetrahedral core of 1 carbon and 3 metal atoms is present. Such complexes in which the nictal atoms are nickel, ruthmium, and osmium havc been prepared: XI11 (81),XIV (82, 83), and XV (65, 82). Their chemistry remains largely unexplored, except for the transformations of compound XV in strong acid medium which we mentioned in the prcvious section. Related structurally to the RCCo3( CO j complexes are the acetylcnedi- cobalt hexacarbonyls M hich havc the structure with a pseudotetrahedral Alkylidynetricobalt Nonacarbonyl Complexes 139 CzCoz core shown in complex XVI (84).These have becii covcrd quite thoroughly in the review litcmturc (5, 85), but the similarities between some of thc chemistry of the Il('c'o~(c'O)~and (RCZR)Coz(CO),com- plexes are worth noting. Thus, (HC~H)C~Z(CO)~can be converted to a mixture of (PhCZH)Co2 (CO) and ( PhCsPh) Coz(CO) 6 by rcaction with diphcnylmercury (5),in a rc.ac.tioii completely analogous to the phenyla- tion of HCCo3(CO) 9 with this rclagrnt. Diphenylacetylenedicobalt hexa- carbonyl, like P~CCO~(CO)~,is rmdily acetylatcd in the para position of both of its phcnyl rings to givv (.ithcr [p-CHsC (0)C6H&zCsH5]Coz (CO) or [p-CH3C (0)C6H&2C6H4(' (0) CHs-plCoz (CO) 6 in high yield, dcpend- ing on reaction stoichiometry (.$9a, 86). Finally, evidence has been pre- scnted by Nicholas and Pcttit (87) that carbonium ions dcrivcd from dico- balt hexacarbonyl complcxcs of propargylic alcohols, c.g., compound XVII, are much morc readily formed and morc stable than analogous carbonium ions derived from the frcc ligand. Thus, the Coz(CO)6 group stabilizes a-carbonium ion centers, but not as cffcctively as the (OC)&o3C group, since compound XVII and othcxr ('02 (CO)6-complcxcd carbonium ion salts cannot be isolated. The analogy b(%w:rn RCCo3(CO) 9 and (RCZR) Coz- (CO)6 complexes, however, should not bc overstressed. The acetylenedicobalt hexncart)onyls have been dcscribcd in terms of Me H \ 140 DIETMAR SEYFERTH complexes of the first excited state of the acetylene and the Co~(c0)~unit (88),and we note that the 13C NhIR chemical shifts of the cluster carbon atoms of these complexes are quite different from those of the apical carbon atom in the RCCo3(CO)g cluster compounds. As we havc seen, 6~ for the apical carbon in neutral alkylidynetricohalt nonacarbonyls is in the range 250-310 ppm. Diphcnylacctylenedicobalt hexacarbonyl, on the other hand, shows 6~ for the carbon atoms in its C2C02cow at 89.6 ppm (89), which is within the range of acctylcnic carbon atom shieldings (58). From the discussion of the organic chcmistry of the alkylidynetricobalt nonacarbonyl complexes given in this review, it is apparent that this has becn a fruitful area of research. We think that it shall continue to be so. Admittedly, the sensitivity of these clustsr complrxcs to diverse bases, nucleophiles, and oxidizing agents will seriously limit the chemistry that can be carried out, but even with these limitations it should be possible to continue a broad development of the organo-functional interconversions of these complexes. At this point, most of the functional substitucnts have been introduced into the cluster complexc>seither at the apical carbon atom, (OC)9C03CZ, or at the carbon atom a to it, (OC‘)gCo3CCZILR’.Any functional group attached to the apical carbon atom is very sterically hindered and as a result will have rather limited chemical reactivity. Functional groups at the a-carbon atom mill be less hindrred, but sonic may be easily lost due to the case with which a-carbonium ions are formed. Clearly, an investiga- tion of the organo-functional chemistry at carbon atoms more remote from the cluster (OC)gCo3C-Y-CZRlt’ (Y is a difunctional organic unit) should be of intersst and, perhaps, more useful in terms of possible ap- plications. As has becn pointed out, very littlc is known concerning reaction mecha- nisms in the area of organocobalt cluster chemistry. The main reactions by which these complexes am formed are only poorly understood, as are the reactions by which mcathylidyne- and halomethylidynetricobalt nonacar- bonyls are alkylated and arylated. We can only guess about the mechanism of the remarkable aluminum chloride-induced conversion of (OC)gCo3CC1 to (OC)&~CO+A1Cl4-, and the mechanisms of the decomposition and the oxidation of these clusters are not known. A b&tcr knowledge of mecha- nisms in organocobalt cluster chemistry most certainly would facilitate the development of the chemistry of these complexvs. Wc wers first drawn into studies on alkylidynetricobalt nonacarbony1 complexes because in thvse one is dealing with a carbon atom in a most unusual environment. Wc felt that this novel class of complexes would show some rather interesting organic and organoinetallic chemistry and we Alkylidynetricobolt Nonacorbonyl Complexes 141 have not been disappointed in this cxpcctation. We believe that more inter- esting chemistry of the RCCo3 (CO)9 and related cluster complexes remains to be uncovered and we are continuing our efforts in this area. ACKNOWLEDGMENTS My pre- and postdoctoral co-workers who carried out the research reviewed here, are listed on the title page of this rhapter. I am indebted and grateful to them for their dedicated, enthusiastic, and skillful efforts and for their important contributions of original ideas that resulted in rapid development of organo-functional organocobalt cluster chemistry. My co-workers and I are grateful to the National Science Foundation for generous support of this work (NSF Grant GP 31429X). REFERENCES 1. D. Seyferth, J. E. Hallgren, It. ,J. Spohn, A. T. Wehman, and G. H. Williams, Spec. Lect., dSrd Int. Congr. Pure Appl. Chem., 1971 6, 133-149. 2. R. Markby, I. Wender, It. A. Fricdel, F. A. Cotton, and H. W. Sternberg, J. Amer. Chem. SOC.80, 6529 (1958). 3a. U. Kruerke and W. Hiibel, Chem. Ind. (London) 1264 (1959). 3b. W. T. Dent, L. A. I~uncanson,It. G. Gliy, H. W. B. Reed, and B. L. Shaw, Proc. Chem. SOC.169 (1961). 4. P. W. Sutton and L. F. Dahl, J. Amcr. Chem. Soc. 89, 261 (1967). 5. I). Seyferth, J. E. Hallgren, R. .J. Spohn, G. H. Williams, M. 0. Nestle, and P. L. K. Hung, J. Organometal. Chem. 65, 99 (1974). 6. R. S. Dickson and P. J. Fraser, Advan. Organometal. Chem. 12, 323 (1974) (review). 7. R. Ercoli, E. Santambrogio, and G. Tettamanti-Casagrande, Chim. Ind. (Milano) 44, 1344 (1962). 8. G. Bor, L. Mark6, and B. Mark6, Acta Chim. Acad. Sci. Hung. 27, 395 (1961); Chem. Ber. 95, 333 (1962). 9. D. Seyferth, J. E. Hallgren, and P. I,. K. Hung, J. Organometal. Chem. 50, 265 (1973). 10. D. Seyferth, C. S. Eschbach, G. H. Williams, P. L. K. Hung, and Y. M. Cheng, J. Organometal. Chem,. 78, C13 (1974). 11. G. Phlyi, F. Piacenti, M. Bianci, and E. Benedetti, Acta Chim. Sci. Hung. 66, 127 (1970). 12. R. Dolby and B. H. Robinson, J. Chem. SOC.,Dalton Trans. 1794 (1973). 13. P. J. Krusic, private communication. 14. T. W. Matheson, B. M. Peake, B. H. Robinson, J. Simpson, and D. J. Watson, J. Chem. SOC.,Chem. Commun. 894 (1973). 15. R. Dolby and B. H. Robinson, J. Chem. SOC.,Dalton Trans. 2046 (1972). 16. D. Seyferth and J. E. Hallgren, J. Organometal. Chem. 489, C41 (1973). 17. N. Sakamoto, T. Kitamura, and T. Joh, Chem. Lett. 583 (1973). 18. T. Kamijo, T. Kitamura, N. Sakamoto, and T. Joh, J. Organometal. Chem. 54, 265 (1973). 19a. G. Albanese and G. Gavezzotti, Chim. Ind. (Milano)47, 1322 (1965). 142 DIETMAR SEYFERTH 191,. G. Plilyi and G. VBradi, J. Organometal. Chetn. 86, 119 (1975). 20. B. L. Booth, R. N. Haszeldine, P. 13. Mitchell, and J. J. Cox, J. Chem. SOC.A 691 (1969). 21. E. Klumpp, G. Bor, and L. RIarkd, Chem. Ber. 100, 1451 (1967). 22a. F. Klanberg, W. B. Askew, and L. J. Giiggenberger, Inorg. Chem. 7, 2265 (1968). 22b. G. Schmid and V. Biitzel, J. Organottietal. Cheni. 46, 149 (1972). 23. W. N. Ingle, G. Preti, and A. G. MacDiarmid, J. Chem. SOC.,Client. Commun. 497 (1973). 24. S. A. Fieldhouse, B. 11. Freeland, C. L). WI. Rlann, and 11. J. O’Brien, J. Chem. SOC., Chetn. Cornmun. 181 (1970). 25. 11. Seyferth, It. J. Spohn, M. It. Churchill, K. Gold, and F. R. Scholer, J. Organo- metal. Chetn. 23, 237 (1970). 26. R. Ercoli, Chim. Znd. (Milono) 44, 565 (1962). 27. D. Seyferth, J. E. ITallgren, and C. S. Eschbach, J. Amer. Chem. SOC.96, 1730 (1974). 28. (a) L. P. Haminett and A. J. Ueyrup, J. Amer. Cheni. SOC.55, 1900 (1933); (b) H. P. Treffers and L. P. Hammett, ibid. 59, 1708 (1937). 29. M. S. Newman, J. Amer. Chem,. SOC.63, 2431 (1941). 30. J. E. Ilallgren, Ph.D. Thesis, hlass. Inst,. of Technol., 1972. 31. 1). Seyfert,h and G. IT. Williams, J. Organomtnl. Chem,. 38, C11 (1972). 32. G. €1. Williams, Ph.D. Thesis, Mass. Inst. of Technol., 1973. 33. K. Iky, C. Eaborn aiid 1). 11. hl. Walton, Organometul. Chem. Syn. 1, 151 (1970/ 1971). 34. I>. Seyferth and G. H. Williams, J. Organometal. Chem. 38, Cl1 (1972). 35. J. C. Kotz and D. G. Pedrotty, Organometal. Chetn. Res., Sect. A 4, 479 (1969). 36. 11. F. Shriver, Accounts Chem Res. 3, 231 (1970). 37. A. Alich, N. J. Nelson, aiid 1). F. Shriver, J. Chem. SOC.,Chem. Commun. 254 (1971). 38. R. I)olby, T. W. hlatheson, U. K. Nicholson, B. H. Robinson, and J. Simpson, J. Organometal. Chem. 43, C13 (1072). 39. D. Seyfert,h, G. H. Williams, P. L. K. Hung, and J. E. Hallgren, J. Organometal. Chem. 71, 07 (1974). 40. L). N. Kursanov, Z. N. Parries, and N. 31. Loirn, Synthesis 633 (1974). 41. H. Gilmaii arid D. Wittenberg, J. Org. Che7n. 23, 501 (1958). 42. 11. Calas, E. Frainnet arid J. Bonastre, C.R. Acad. Sci. 251, 2987 (1960); R. Calm, Pure Appl. Chem. 13, 61 (1966). 43. S. I. Sadykli-Zade and A. D. Petrov, Zh. Obshch. Khim. 29, 3194 (1959). 44. I. Ojima, M. Nihonyanagi, and Y. Nagai, J. Chem. SOC.,Chem. Commun. 938 (1972). 45. 11. Seyferth, G. H. Willi:tms, and J. E. Hallgren, J. Amer. Chem. Soc. 95, 266 (1973). 46. I). Seyferth, C, S. Eschbach, G. H. Williams, P. L. K. Hung, and Y. &I. Cheng, J. Organornelal. Chem. 78, C13 (1974). 47. C. Eaborn, “Organosilicon Compounds,” pp. 140-143. Butterworths, London, 1960. 48. D. A. Young, Znorg. Chem. 12, 482 (1973). 49. (a) 1). Seyferth and A. T. Wehman, J. Amer. Chem. SOC.92, 5520 (1970); (b) D. Seyferth, G. H. Willi:Lms, A. T. Wehman, and M. 0. Nestle, ibid. 97,2107 (1975). 50. M, Itosenblum, “Chemistry of the Iron Group Metallocenes,” Chapter 4. Wiley (Interscience), New York, 1965. 51. N. Fuson, M.-L. Josien, and E. M. Shelton, J. Amer. ChenL. SOC.76, 2526 (1954). 52. H. W. Sternberg, I. Wender, It. A. Fricdel, and hl. Orchin, J. Amer. Chem. SOC. 75, 2717 (1953). Alkylidynetricobalt Nonacarbonyl Complexes 143 53. 1). Seyferth, G. H. Williams, arid I>. 1). Traficante, J. Amer. Chem. Soc. 96, 604 (1974). 54. G. A. Olah, E. B. Baker, J. C. ICvans, W. S. Tolgyesi, J. S. Rfclntyre, and I. J. Bastien, J. Amer. Chem. Soc. 86, 1360 (1964). 55. G. A. Olah, Science 168, 1298 (1970). -56. J. J. Dannenberg, RI. K. Levenlxrg, :tiid J. H. Richards, Tetrahedron 29, 1575 (1973), and earlier references cited therein. 57. G. IS. Williams, D.1) . Traficante, arid 1). Seyferth, J. Organometal. Chem. 60, C53 (1973). 58. .J. B. Stothers, “Carbon-13 NAZI?. Spectroscopy.” Academic Press, New York, 1972. 59. 1). Seyferth, C. H. I,;sctibsrh, nnd 11. 0. Scstle, b. Orgcc.nonieln/. (‘hciii. 97, Cll (1975). 60. See Stothers (58), pp. 217-23s. 61. C. G. Kreiter and V. Forniacrk, Aicgew. Cliem. 84, 155 (1972). 62. F. H. Kijhler, 1%.J. Kalder, and 15. 0. Fischer, J. Organometal. Chem. 85, C19 (1975). 63. 13. 0. Fischer, G. Kreis, C. (4. Kreiter, J. Rliiller, G. Huttner, and H. Lorenz, Angelo. Chem. 85, 018 (1973). 64. T. G. Traylor, 11. J. Uerwin, .J. Jerkunicx, and hl. 1,. Hall, Pure Appl. Chern. 30, 599 (1972). 65. A. J. Deeming, S. ITasso, M. Underhill, A. J. Canty, 13. F. G. Johnson, W. G. Jackson, J. Lewis, and T. W. hlatheson, .I. Chem. Soc., Chem. Commun. 807 (1974). 66. B. E. Rlann, Advan. Organomelnl. (‘hem. 12, 135 (1974). 67. E. 0. Fischer, Angela. Chem. 86, (is1 (1974). 68. I. U. Khand, G. R. Knox, P. L. Pauson, and W. E. Watts, J. Organometal. Chem. 73, 383 (1974). 69. A. T. Wehman, unpublished work. 70. K. Tominaga, N. Yamagami, and II. Wakamatsu, Tetrahedron Lett. 2217 (1970). 71. G. Albanesi and E. Gavezotti, Atti Accad. Naz. Lincei, Rend., CI. Sci. Fis. Mat. Nut. 41, 497 (1966). 72. R. S. Dickson and G. R. Tailby, Amt. J. Chem. 23, 229 (1970). 73. C. Hoogzand and W. Hubel, in “Organic Syntheses via Metal Carbonyls” (I. Wender and P. Pino, eds.), Vol. I, pp. 343-371. Wiley (Interscience), New York, 1968. 74. C. H. Bamford, C. G. Eastmond, and W. It. Maltman, Trans. Faraday Soc. 60, 1432 (1964). 75. G. PAlyi, F. Baumgartner, and I. Czajlik, J. Organometal. Chern. 49, C85 (1973). 76. P. A. Elder and B. FI. Robinson, J. Organometal. Chem. 36, C45 (1972). 77. T. W. hlatheson and B. 1%.Robinson, J. Chem. Sac. A 1457 (1971). 78. V. G. Albano, P. Chini, S. RIartinengo, M. Sansoni, and I). Strumbolo, J. Chem. Soc., Chem. Commun. 299 (1974). 79. G. P&lyi, F. Piacenti, and L. Mark6, Itiorg. Chim. Acta Rev. 4, 109 (1970). 80. B. R. Penfold and 13. 13. Robinson, Accounts Chem. Res. 6, 73 (1973). 81. T. I. Voyevodskaya, I. hl. Pribytkova, and Yu. ,4. Ustynyuk, J. Orgariometal. Chem. 37, 187 (1972). 82. A. J. Deeming and M. Underhill, J. C‘hern. Soc., Chem. Commun. 277 (1973); J. Chem. Soc., Dalton Trans. 1415 (1974). 83. A. J. Canty, B. F. G. Johnson, J. Lewis, and J. R. Norton, J. Chem. Soc., Chem. Commun. 1331 (1972). 84. W. S. Sly, J. Amer. Chem. Soc. 81, 18 (1959). 144 DIETMAR SEYFERTH 85. W. Hubel, in “Organic Syntheses via Metal Carbonyls” (I. Wender and P. Pino, eds.), Vol. 1, pp. 273-342. Wiley (Interscience), New York, 1968. 86. D. Seyferth, M.0. Nestle, and A. T. IVehman, J. Amer. Chem. Soc. 97, 7417 (1975). 87. K. M. Nicholas and R. Pettit,, J. OrganometaL Chem. 448, C21 (1972). 88. (a) Y. Iwashita, F. Tamura, and A. Nakamura, Inorg. Chem. 8, 1179 (1969); (b) Y. Iwashita, ibid. 9, 1178 (1970); (c) Y. Iwashita, A. Ishikawa, and M. Kaino- sho, Spectrochim. Acta, Part A 27, 271 (1971). 89. L. J. Todd and J. R. Wilkinson, J. Organometal. Chem. 80, C31 (1974). Ten Years of Metallocarboranes KENNETH P. CALLAHAN Metcalf Research laboratory Departrnenf of Chemistry Brawn Universify Providence, Rhode Island M. FREDERICK HAWTHORNE Departrnenf of Chernisfry* Univerrify of California Lor Angeles, Californio I. Introduction . . 14.5 11. h3etallocarboranes: Syntiiet,ic h1t:tliods. . 150 A. Preparation from nitlo-Cnrt)orane Anions . 150 B. Preparation by l’olyhetlral l+:xpnrision . 151 C. Preparation by I’olyhedrd Contraction . 152 1). Preparation hy l’olyhedrnl Sul,rog:Lt,ion . 153 E. Preparation by Thermal llct.al Trmsfer . . 153 111. Twelve-Vertex ~\letallocarborltii(,s 155 -4.llonometallic Complexes with Ttlcntictd Carborane Ligands . 155 B. Monoinetallic Complexes \\it11 1)iffcrent Carborane Iigands . 161 C. Mixed-Ligand Complexes . 163 D. Bimetallic Complexes . 166 IV. Thirteen-Vertex ~.letallocarl)or:Lllrs . 167 V. Fourteen-Vertex L~etsllocarhoraties 171 VI. Eleven-Vertex lletallocsrhoratIcs 171 A. Llonometallic Complexes 171 R. Bimetallic Complexes 173 VII. Ten-Vertex Met:illocarboraites 175 VIII. Nine-Vertex nZetsil1ocarhoranc.s . 178 IX. Oxidative Addition to B-H Ilonds . 180 X. Metallocarboranes in Homogcricvus (htdysis . . 182 References . 183 I INTRODUCTION In the early 1960s, the chemistry of the boron hydrides had been ex- tended not only to include a remarkable number of new parent boranes having diverse structures, but, the polyhedral BloHlo2-and B12H12~- ions * Contribution So. 3453.. 145 146 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE and two isomeric CzBloHIzcarborancs as well.' Both BI0Hlo2-and 1,2- (or ortho-) C~BloHlzwere obtained from decaborane (14), BI0Hl4,whereas the 1,7- (or mela-) isomer of CzBloHlz was prepared by thermal rearrange- mcnt of the 1,2-isomer at 400" to 600C" (Fig. 1) (31,39,57, 65). As later work would show (47, 77),an entire serics of polyhedral B,,Hn2- and isoelectronic CzB,-zH, carboranes were synthetically accessible for n = 6 through 12. The generally decreased chemical reactivity of these polyhedral species over that of thc boron hydridcs suggested that they 1.12 -C2BDH12 OBHOCH OH FIG.1. Structures and numbering of the three isomeric irosahedral carboranes, and the degradation of 1,2-C2Bl0H12 to 7, 8-CzBgH1z-. The bridging hydrogen is shown in one of the two equivalent bridging positions. Numbering of polyhedral positions follows the latest IUPAC-approved scheme as published in Pure Appl. Chenr., 30, 683 (1972). The following definitions apply to the various descriptions of polyhedra used in the text: closo refers to a borane, carborane, or metallocarborarie polyhedron that has a closed, fully triangulated geometry ; nido refers to a polyhedron the geometry of which can be described as a closo polyhedron from which one vertex (frequently one of high roordination number) has been removed; commo refers to metallocarborane complexes in which one vertex, generally a transition metal, is shared between t,wo polyhedra. Coordination numbers of polyhedral vertices are calculated from the number of nearest-neighbor atoms and do not imply the presence of discrete 2-electron chemical bonds between these atoms. Ten Years of Metallocarboranes 147 were probably stabilized by t hrcic-dirnciisional electron delocalization and hence were representative “aromaticJ’ members of the boron hydridc family. Indeed, Lipscomb (71, ’72) arid co-workers have carried out a series of molecular orbital treat mmts of selected polyhedral ions and car- boranes which adequately espleins t hv bonding present in these species. In all cases, a polyhedron having 11 number of vertices requires n + 1 clectroii pairs d(1localizcd in ail cqual number of extended bonding orbitals to achieve polyhedral cage bonding. Although the icosahcdral 1 ,2- and 1 , 7-(?2B1&12 carborancls n-erc’ found to be quite stablc at high teinpcmturcs and toward most common reagents (31, 41, 57), strong base in thc~prcscmcb of a protonic solvciit caused a specific degradation reaction (97) which cleanly removed a BH vertex from the icosahedron to produccx thv corresponding (’ZBSH12- ion (Fig. 1) : 1,2-C?BloHIr + 110- + 21tOH ----t7,8-C2139Hl?- + I3(OIL)~+ H? (1) It was correctly assumcd and latcr drtclrmined unequivocally (64) that in each of these ttto reactions thr BH vdex that mas removed was always one of the two equivalent vclrtices found as the nearest neighbors of the two equivalent C“H vertices. Thc fact that the carbon atoms in polj hcdral surfaces of C2B,-2Hn carboranc1.s are electronic counterparts of boron atoms in the corresponding isoelectronic B,,H,*- polyhedral ions (62) requires these carbon atoms to resemhlv C+, a species present in carbonium ions. Consequently, the BH vertices which are nearest ncighbois of two such carbon atoms will be activatcd for nucleophilic attack by base through the advent of a strong inductivcl effcct. The assumption that the isomeric 7 , 8- and 7 ,9-C2B9Hle-ionsstructurally resembled eleven-particle fragments of an icosahedron, coupled with their known empirical formulas, suggclsted that the twelfth hydrogen atom was prescnt in a three-center B-H-B bridge bond located in the periphery of the open five-membered fact. (see Fig. 1). Simplified molecular orbital considerations suggested that the rcinoval of this bridge hydrogen atom as a proton would generate a 7,8- or 7 J9-C2BgH112-ion having G delocalized electrons in the open five-mcbmbcred face. The orbitals in which these G electrons were distributed could, to a first approximation, be considered as sp3-like and pointed toward the unoccupied vertex of the original car- borane icosahedron. This disposition of dc1ocalizc.d electrons ( Fig. 2) should closely resemble the ubiquitous cyclopcntadicnide ion, a constituent in a large number of organomctallic compounds. As a result of this observation, the twelfth hydrogen atom in the 7 ,8-C2BgH12-ion was successfully re- moved by treatment nith strong bases such as sodium hydride (55). Re- action of the resulting 7 ,S-C’2ByHn2-ion with iron(I1) chloride produced 148 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE -2 FIG.2. Schematic representation of the spa-like bonding orbitals in (a) 7, 8-C2B9111,*- and (h) 7,9-C2B91IIlZ-. the first metallocarborarie in a manner analogous to the synthesis of ferro- cene (56) : NaH Fe(I1)., 7,8-CsB91T12- -7,8-C2T39H112- - il ,2-C2B~Hll)2Fe(II)2-(2) --H2 By using the C5H5-ligand in conjunction with the 7,8-C2&,H112- ligand, the mixed-ligand species CjHjFc(1 ,2-CzBgHll)- was obtained (44). One- electron oxidation of this anion formed the unchargcd C',H5Fc ( 1,2-C'&&1) species containing formal iron(III), an analog of the ferriciriium ion. Subsequent single-crystal X-ray diffraction studies on this derivative (104) substantiatcd the bclicf that the iron atom was indeed occupying the twelfth vertex of an icosahedron. In this manner metallocarborane chemistry was brought into bcing and thc first wedding of transition metals with carborancs into hybrid cluster compounds was accomplished. Following thew events, the chemistry of the metallocarboranes was rapidly expanded through the use of many of the transition metals and certain main group mrtals such as Be (82, 83),Al, and Ga (102).A boron vertex may be reinserted as well (53). In addition, carborane ligands other than the original 7,s- and 7,9-CzB9Hll2- ions were attached to metals, often in conjunction with a wide varic.ty of truly organic ligands (47'). More recent developments have led to the synthesis of metallo- carboranes that contain more than a single transition metal in a polyhedral surface, and these transition metals nccd not be identical (3). With very few exceptions, the gross geometries of polyhedral metallo- carborancs may be correlated with the total number of vrrticrs present in the polyhedron. Metal, carbon, boron, or other nonhydrogen elements are counted as vertices and thc.ir total number equated with the value of n in R,Hn2- ions. In nearly every case the approximate geometry of thr poly- Ten Years of Metallocarboranes 149 hedral metallocarborane will coincide with that of the corresponding B,H,2- ion, when known (4).Table I lists these geometries as a function of n. The thermal polyhedral rearrangement of 1 ,2- to 1 ,7-CzB,,HI2 already mentioned (see Fig. 1) is but one rxample of a reaction commonly observed throughout the polyhedral carborane and metallocarborane families. At the present timc, the mechanism of tlicse interesting rearrangements rc- mains obscure, although several schemes have been advanced to explain experimental results (46).Polyhedral rearrangements in metallocarboranes occasionally occur with great facility in comparison to the energy required to effect the 1,2- to 1,7-C2BloHlz isomerixation, and are important aspects of the physical and chemical propcrtics of metallocarboranes. In this review, we treat in dcpth thc synthesis, structures, properties, and reactions of 7-bonded metallocarboranes. Our survey is restricted to complcxcs of 2-carbon carborancs and to specics that havc between nine and fourteen total polyhedral vchces. Coverage of metal complexes of other hetrroboranes is availahh. in Grimes’s book (41) and in Todd’s re- view (9.3). The recent work of Grimes and his group has concentrated on metallocarboranes having fewer than nine vertices (42, 75, 76). Our approach to the subject has been to divide thc metallocarboranes according to the size of the polphcdron. Starting with twclvc-vertex com- pounds, which constitute the majority of the effort, we proceed to the larger polyhedra, so far unknown in the B,Hn2- and C2B,-2H, series, and then to the lower polyhedra. Furthcr subdivisions within each polyhedral size include synthesis, structurcs, and propertics of monometallic com- plexes, rcactions of monometallirs, bimetallic preparations and reactions, and, in two instances, trimetallic compounds. TAT3T,I< I CORRELATIOVOF (:Ro\\ I’OI.YHEDRAI, GEOMETRYWITH TOT~LYr MDCR OF VERTI~ES Total vertice5, n Observed geometry 12 Icosahedron 11 (ktadecahedron 10 I3ieapped square antiprism 9 Tricapped trigonal prism 8 I hdecahedron 7 Pen tagonal bipyramid 6 ()ctdiedron 150 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE Metallocarborancs have only been known for 10 years, and research into their synthesis and characterization has involved a small number of workers. Conscquent?ly,practical applications of these unique compounds havc not been rapidly forthcoming. Itcccnt work has shown catalytic activity in certain of t hex compouiids, however, arid may signify future commercial value and industrial importance of mctallocarboranes. I1 METALLOCARBORANES: SYNTHETIC METHODS Five major synthetic routes are now available for the preparation of metallocarborancs, although only oiic was well established in 1969. The recently developed synthetic methods have allowed the preparation of more complex and diverse compounds and have greatly expanded the field of metallocarborsne chemistry. Thcse preparative methods are discussed in some detail in this section, for the synthesis of all the known metallo- carboranes have been accomplished by one or more of these routes. A. Preparation from nido-Carborane Anions As mentioned previously, the first metallocarborane synthesized, (1,2-C2BBHl1) zFe (11) 2- (Fig. 3), was prepared in a manner similar to the synthesis of ferrocene, e.g., reaction of anhydrous FeClz with the nido- carborane dianion 7,8-CzB9H1I2-,which itself was formed from 1,2-CzBloH12 2- FIG.3. Structure of the first metallocarhorane ever synthesized, (1, 2-C2B9Hll)zFe(II)2-. Ten Years of Metallocarboranes 151 by degradation with strong base [Eq. (2)I.This synthetic method, with some modifications, has been used to prepare a wide varicty of mono- metallic twelve-vertex metalloc,dr b oranes. The reaction conditions for this type of preparation generally involve nonaqueous solvents, such as tctrahydrofuran (THF) or diethyl ether, and rigorous exclusion of air and watcr. Some metallocarboranes, however, may be prepared in high yicld by reaction of a mctal salt and the nido monoanion, CzBsH12-, in strong aqueous base : 7,S-C213JIll- + 011- -H20 + 7,8-C2BgHIl2- (3) 2 (7,&C2B9111,2-1 4- 11'" -(1,2-CJ39Hil) 211n-' (4) In these instances, it is belicwd that the strong base deprotonatcs the monoanion to a small extent, permitting complexation to occur. This synthetic approach proved valuable for the preparation of lower monometallocarboranes as wcll : the C&H1I2- ion, prepared from 6,8- C2B7H13,was found to react with mctal ions, losing one equivalent of hy- drogen gas, to form metallocarboranes of the type (CZB~H~)~M"-~(38). B. Preparation by Polyhedral Expansion Polyhedral expansion, which was first reported in 1970 (Fig. 4) (16), entails the reduction of a closo-carborane with a strong reducing agent, FIG.4. Polyhedral expansion of 1,7-C&,H,. 152 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE such as an alkali metal, followed by reaction with a transition metal re- agent. All reactions must be performed in nonaqueous solvents under nitrogen. It is believed (4) that the reduction step results in an opening of the closed carborane polyhedron, producing a nido dianionic species, which then reacts with a transition metal ion in a fashion similar to that just discussed : C?Bn--2Hn+ 2Sa -CzB,-zHn2- (5) 2 (C?Bn-?ITn?-)+ >In+- (CzB,-,H,) 231n-4 (6) Although a number of different metallocarboranes have been synthe- sized in high yield by this reaction (18, 94),the actual chemistry is fre- quently much more complex than implied by Eqs. (5) and (6). For ex- ample, products containing greater and fewer numbers of boron atoms than present in the carborane starting matwials have been isolated from polyhedral expansion rcactions. The polyhedral expansion reaction appears to be a general synthetic method for metallocarboranes; all the known closo-carboranes have been found to produce metal-containing compounds when subjected to the re- duction-complexation operations of this synthetic scheme. Moreover, metallocarboranes containing more than one transition metal may be pre- pared by the polyhedral expansion of monometallocarboranes. Examples of this synthetic route will be described in following sections. It should be noted that in the polyhedral expansion process, as idealized in Eqs. (5) and (6), the product metallocarborane has one more vertex than was present in the carborane starting material-hence the origin of the descriptive phrase “polyhedral expansion.” By contrast, when metallo- carboranes are prepared by reaction with CzBgH112- ions, which are pre- pared from the icosahedral C2BlOHl2carboranes, twelve-vertex metallo- carboranes result. C. Preparation by Polyhedral Contraction The polyhedral contraction route to metallocarborancs consists of the degradative removal of a polyhcdral boron atom of a metallocarborane followed by oxidative closure of the resulting nido-metallocarborane com- plex to a closo species having one fewer vertex than present in the starting material (68): 1 on- C,H~CO(I,~-C~R~H,,)C,H.CO(~,~-CPRHH~O) (7) -1x202 2 The polyhedral contraction process is thus complementary to polyhedral Ten Years of Metallocarboranes 153 expansion, in that the formcr d(.creascs polyhedral size (Fig. 5) whereas the latter iiicrcascs the number of polyhedral vertices. Polyhedral contrac- tion is not as general a synthrtic method as is polyhedral expansion, since some metallocarboranes urdlrgo complete decomposition upon attempted partial degradation, and side rcactioiis are frequently a difficulty in poly- hedral contraction. NCvert8hdvssJthis is a valuable route to new complexes if the proper reaction conditions can effected. D. Preparation by Polyhedral Subrogation Synthetic polyhedral subrogation for the preparation of polymetallo- carboranes from mononictallocarboranes is an off shoot of polyhedral contraction in that, after degradativc removal of a BH vertex, a transition metal ion is reacted with the niclo-metallocarboranc produced rather than with an oxidizing agent. In this way, a new transition metal vertex is in- corporated into the polyhedral framcwork without a change in the number of vertices between reactant and product (Fig. 6) : 1. OH This method is useful for the synthesis of metallocarboranes containing two similar or differcnt transition mrtal vertices (19, 23) but has not yet been explored to determine its full potential. E. Preparation by Thermal Metal Transfer A newly discovered syrithctic route (27) to bimetallocarboranes, thermal metal transfer may prove to be a highly valuable synthetic method but, as 154 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE EtOH I I1 II 12 OCH Ow FIG.6. I'olyhedral subrogation of C&Co(l ,~-CJ~I,H,Z)with Fe(I1). yet, has been solidly established for only one system (Fig. 7). The method, which involves the pyrolysis of a metallocarboranc through a hot tube in vucuo, was found to produce bimctallocarborancs having one vertex greater than was prescnt in the starting material: heat C~EI,CO(CZB,Hio) -(CX,) ZCOZ(CzBJIm) (9) Even more intriguing was the observation that the same products were formed upon pyrolysis of ionic cobalticinium salts of commo-metallocar- boranc ions : (CjHg) &of (C,BJI,,) 2Co-kICjHs) ZCOL(C~U~H~O) (10) This production of neutral bimctallocarboranes from ionic monometallic FIG.7. Thermal metal transfer as a synthetic route to bimetallocarboranes. Ten Years of Metallocarboranes 155 precursors was unprccedcnted, and further examples of this type of rcac- tion are being sought. Spencer, Green, and Stone (58) have rccently described a new synthetic approach to metallocarboranc~sin which an organometallic transition mctal complrx is thermally reacted \I itli a neutral closo-carborane, resulting in the incorporation of the metal iiito the polyhedron: 1 ,8- (('FI1) ,C2U,H, + (1 , 3-<'JII2)."I'i -( 1 , 5-CaHI~)Ni (1 ,7-(CHJ) 2'2~13~H~) (1 1) Grimes and co-workers (7'6) suhsequcntly reported the synthesis of several othcr metallocarboranes in a similar fashion. This method may proceed in a manner similar to thcrmal mdal transfer and is, thus, included here. A ncw synthetic classification may h arrantcd if this prcparativc route proves to have widespread utility. 111 TWELVE-VERTEX METALLOCA RBORANES Although several thirtcen- and fourtccn-vertrx mctallocarboranes have been prepared and investigatrd, wr have chosen to begin our closcr ex- amination of mctallocarborancs \\ ith twclve-vertex specics for several reasons. The first mc~tallocarl~orancprcpared was a twelve-vertex complex (5C), and much of our uridcrstanding of metallocarboranes in general has arisen from studies on tv clvtl-vc'rtc.x compounds. In thv early days of metallocarboranc research, almost all the. work was performed on twclvc- vertex systrms (54), and thcrc have bccn more "icosahedral" metallo- carborancs reported than of any othw geometry. Thus the foundations of metallocarborane chemistry arc twclve-vertex systems, and the propcrtics of mc%allocarborancs of nonicosahedrnl geometry may be predicted on the basis of the behavior of tmelvc~-vrrtc~xanalogs. A. Monometallic Complexes with Identical Carborane ligands 1. Synthesis, Structure, and Physical Properties The d3 chromium( 111) complcx (1, 2-CzB9Hll) zCr- was prepared (85) from 7 ,8-CzR9HIl2-and CrC13 in THF solution. An X-ray crystal structure determination on thc Cs+ salt, performed by St. Clair, Zalkin, and Temple- ton (90)showed it to have a sytnirietrical sandwich structure in which the 156 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE metal atom is nearly equidistant from 3 boron atoms and 2 carbon atonis of each icosahedron but is distinguished by relatively long metal-ligand bond distances. No other significant difference in structure between this overall 15 valcnce-electron complex and 18 valence-electron complexrs such as CjHjFe ( 1,2-C2BsHll) (104) wrrc noticeable. This Cr (111) com- pound was the first electron-deficient metallocarborane prepared. The first mctallocarborancs ever synthcsized were the iron (11) and iron (111) species, (1,2-C&BsHll) ~Fc?-*~-, which were prepared in the man- ner described previously (56). The pink diamagnetic Fe (11) compound was observed to be air-sensitive; it is rapidly oxidized to the maroon, paramagnetic Fe (111) monoanionic complex (1, 2-C2BsHll)2Fe-. By con- trast, the fcrriciniurn ion, ( C5H5)2Fe+,exhibits a limited lifctimr (84).This stabilization of a high metallic oxidation statc by carborane ligands is ob- served in other nictal systems (see below), and may be due to the ability of the “aromatic” carborane ligarids to donate electron density to the metal. Tho rlectrostatic cffccts due to thr presence of diariionic carborarie moieties rather than monoanionic cyclopentadienyl ligaiids may also be important in this stabilization of high metallic oxidation states. The yellow, diamagnetic, 1s-electron complex anion (1,2-CzBgH11)2Co- may be prc- pared from 7,S-(’zBsH112- and CoC12 by either aqueous or noriaqurous routes; the yields arc comparable. The 1 ,7-isonier may be prepared from 7,9-C2B9H,2- in a similar fashion (54).An X-ray diffraction study of the Cs+ salt of this latter coniplcx confirmed SJ mmetrical sandwich bonding, although the polyhedral carbon atonis could not be precisely located due to somc disordering (103). A significant change is observed in the analogous complexes of nickel, palladium, copper, and gold. Rraction of 7 ,S-C2B9H12- with Ni(11) salts, using either aqueous or nonaqueous tcchniqucs, results in thc isolation of a d? Ni (111) complex when the workup is performed in air (54).This para- magnetic complex exhibits a geometry similar to that obsc.rvcd in the Co(II1) and Fc(I1) complrxcs (43). It undergoes rcversiblr 1-electron oxidation and reduction reactions, forming complexes of the formulas (I ,2-C2BsH11)2Ni, which contains formal d6 Ni(IV), and (1,2-CzB9Hll)2Ni2-, which contains formal d8 Ni(1I). Thc 1attc.r compound is highly susceptible to air oxidation, forming the Ni(II1) spccirs. Analogous compounds arc formed with palladium (95).Nonaqueous reaction of 7 ,8-CzBsH112- with cupric salts produces complexes of Cu(I1) and Cu(III), e.g., (1,2- CzBsHll)2Cu2-, which contains formal ds Cu (11), and (1,3-CzBsH1l) 2Cu-, which contains formal c18 Cu(II1). Similar compounds are formed using gold salts (94). Complexes of Ni (11) , Ni (111), Pd (11), Pd (111), Cu (11), Cu(II1), Au(I1) , and Au (111) all have more than lS valence electrons. Ten Years of Metallocarboranes 157 Crystallographic studies pc.rformed by Wing (99, 100) have shown that these electron-rich metallocarboranes exhibit severe distortions from the symmetrical sandwich structurw of thc d3 Cr(II1) (go), dj Fe(II1) (IOS), d6 Co(II1) (104,and cl? Ni(II1) (43) compounds. The d8 complexes of Ni(11) , Cu (111) , Pd (11), and Au (111) arc isomorphous (99) and have a “slipped sandwich” structurc in which the metal atom is in a position closer to thc 3 boron atoms in the bonding face of each carborane ligand than to the 2 polyhedral carbon atoms also nearest neighbors to the metal vertex. The d9 compounds exhibit similar distortions: the Cu (11) complex, for example, shows an averagc. (‘u-B distance of 2.20 8 and an average Cu-C distance of 2.57 A whcrras for the Cu (111) species, Cu -B averages 2.11 8 and Cu-C is 2.52 (99, 100). Two explanations for the adoption of “slipped sandwich” structures by these d8 and d9 metal comp1cxc.s liavc been offered. Wing suggested (99, 100) that thc complcxps he considered as carborane analogs of organo- metallic r-ally1 complexes, in 11 hich 3 ligand atoms are coordinated to the metal, thus reducing the effrctive atomic number at the metal vertex. Warren and Hawthorne (95) have suggested that the slipped sandwich modc of coordination is requircd hy orbital symmetry considerations, similar to those invoked (15, 78) in explaining the slipped configuration of the Ag (I)-benzene complrx.? Comp1exc.s of d6 Ni (IV) and Pd (IV), although isoclectronic with the Co(II1) and Fe(1I) species, exhibit structures different from these latter anionic compounds (91). Although there is no slip distortion, as present in the electron-rich d8 and c19 coniplcxcs, the Ni (IV) and Pd (IV) species are distorted from the icosahcdral gmmetries of the other d6 metallocar- borancs. The polyhedral carlion atoms are arranged in a cisoid fashion in the Ni(1V) and Pd(IV) compounds and are tucked in toward the center of the icosahedron rather than occupying the normal positions in a twelve- vertex closed polyhedron. This gtmnetrical configuration produces a sizable dipole moment : the value mcmurcd for (1,2-CzBgH11) ,Ni (IV) is 6.16 D in cyclohexane solution (95). The polyhedral C-H protons in the Ni(1V) and Pd(IV) compounds show a grratcr degree of acidity than normally observed in metal1ocarboranc.s (95) ; electronic attraction by the highly positive transition metal atom is doubtless a major influence in producing this increase in acidity. Thc structural changes occasioned by change in oxidation state are shown in Fig. 8 for the nickel complexes. The most intriguing propmty of these tetravalent metal complexes was * llolecular orbital theory has rrcently been employed by Wegner to account for these structural changes [l’. A. JVqqwi, Inorg Chpm., 14,212 (1975)l. 158 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE I- 2- - a b FIG.8. (a) Symmetrical sandwich structure of (I ,2-~~~91~11)2~i(111)-ion; (h) risoid sandwich structure of (1,2-C:21391€ll)2X(IV) ; and (c) slipped sandwich structure of (1 , 2-C2R9H11)4Si (11) z-. observed in the compounds in which all four of the polyhedral carbon atoms were substitutrd with mrthyl groups (95). The stcric hindrance caused by the proximity of these bulky substitucnts in cisoid positions produced th(. first example of a lowenergy polyhedral rearrangement in a, metallocarboranc. Reaction of nickcl(1I) acetylacrtonate with 7 , S- (CH,) 2-7, S-CJ39H92- under rigorous exclusion of air and watcr produced an Ni(I1) complex, denoted as [ ( Me2)zNi( 11) I>-, which could be rtvrsibly oxidized to a nickel (111) complex, [ (Mez)2Ni(III)I-. This latter compound underwent irreversible oxidation to the neutral (Mez) ,Ni (IV) , which upon reduction afforded an Ni (111) species different from that originally prepared. Fur- ther study showed the existence of anothcr isomeric ( Mez)zNi( IV) com- plex, two additional Ni(I1) isomers, and a third Ni(II1) spccirs. A com- bination of spectral and crystallographic information determined that these isomers arose by migration of polyhedral C-CHs vertices to positions distant from each other, thereby decreasing steric crowding. These poly- hedral migrations occur in addition to the structural changes occasioned in the Ni( 11)-Ni (111)-Ni( IV) oxidation series, in which the complexes have slipped sandwich, symmetrical transoid, and distorted cisoid con- figurations, respectively. The structures established (10, 95) for the three positionally isomeric { (Mez)*Ni)series are depicted in Fig. 9, where the migration of polyhedral C-CH3 vertices on each icosahedron explain the existence of the three isomeric families. That steric hindrance is responsible for the low activation cnergy requirrld to effect the observed rearrange- ments was shown by attempts to effect similar migration of polyhedral C-H vertices in the unsubstituted complex (1 ,2-C2B9H11)qNi;a mixture Ten Years of Metallocarboranes 159 FIG.9. The three posit,ioii:d isomers of [(CIIa)2C2B9H~]2?;i(III)-. of isomeric compounds could be detected only after pyrolysis in the vapor phase at 360" to 4OO0C, whercas rearrangements of the methylated com- pounds occurred between room t empcrature and 100°C. 2. Reactions Limited investigations of the chemical reactions of metallocarboranes have been reported. In general, these complexes are much less reactive than the analogous metallocenes, yet several unique new species have been prepared from the unsubstituted metallocarboranes. It has been known for a considerable time that ferrocene is protonated by strong acids to produce the cationic species (GH,j 2FeH+; there is some evidence that the proton is directly coordinated to the iron atom, although some interaction of this proton with the cyclopentadienyl rings must be invoked to explain the rapid acid-catalyzed deuteration of ferrocene (IS). The metallocarborane analog of protonated ferrocene may be prepared by reaction of (1 ,2-CzBgHll),I+ (11)*- with strong acids such as HC1 or HC104, and the resulting spocies has been isolated as an air- and water- sensitive crystalline solid (52).A band in the infrared spectrum of this compound at 1885 cm-I has becri tentatively assigned as vFeE, although no signal attributable to this proton mas observed by 'H NMR. This protonated metallocarborane is unique in that it undergoes ready substitution at polyhedral boron atoms, whcrcas the unprotonated species is unreactive in the absencr of acid. Rcaction of (1,2-CzBgH11)2FeH- with Lewis bases such as dialkyl sulfidcs results in the loss of Hf and the forma- tion of boron-substituted complexes in good yields : €IFe(1,2-CzBJI,,)2- + ILS -(1,2-C2B9Hll) (1,2-C2B9H10SIt2)Fe-+ Hz (12) 160 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE It is not necessary to start with the protonatcd Fe(I1) complex to isolate these substituted products, however. Reaction of a slurry of the unsub- stituted Fe(II1) complex with anhydrous HCl in diethyl sulfide produced a mixture of substitution products, including the species shown in Eq. (12). Although no evidence has been found for a protonated cobalt metallo- carborane analog, boron-substituted complexes of cobalt (111) may also be obtained by reaction of the (1,2-C2B9Hll)2Co- ion with R2S and HC1 (52). The cobalt (111) complex, (1 ,2-C2B9Hll) 2Co-, undergoes broinination in glacial acetic acid solution to afford a hexabromo dcrivativo, (1,2- C2BgHsBr3)~CO- (54). An X-ray crystallographic study of this product (14) showed that bromination occurred on the boron atoms farthest from the polyhedral carbon atoms, those boron atoms expected to have the greatest electron density, as predicted for an clectrophilic at tack mcch- anism. Treatment of the potassium salt K[( 1,2-CzB9Hll)zCo] with CS2 and HCl in the presence of AlC13 affordcd a novel neutral compound, Co(l,2- + CzBgHlo)*S2CH in which boron atoms on the two icosahedra are substituted + by sulfur and the polyhedra are linked by thc S2CH moiety (11). The structurc of this compound is depicted in Fig. 10. It contains an elcctron- deficient carbon atom presumably stabilizcd by the iionbondirig electron pairs of its adjacent sulfur atoms. A similar complex may be formed using FIG.10. Zwitterionic structure of (1,2-C,D,H,,),Co(S,~H). Ten Years of Metallocarboranes 161 acetic acid-acetic anhydride iii placv of CS,; in this case, the intercage + bridging group is the H3C--C'02 nioiet y (32). The diazonium salts ( CGH~IZN2)+ (1, 2-C2B9Hll)2C~- undergo an inter- esting free-radical reaction upon pyrolysis. Nitrogen gas is liberated and the product is found ti) consist of a synimctrical sandwich comnzo-metallo- carboranc with an ortho-disu1)stitutc.d phenyl ring bonded to polyhedral boron atoms arid bridging betn ovn thc two icosahedra (35). The neutral complex (1, %-C'?139H11)zNi (IV) behaves as a rather strong Lewis acid and forms stable crystalline 1: 1 addition complexes with clec- tron-donating moleculcs and ions. ('omplexes with halide and pseudohalide ions, as wcll as with certain largci aromatic hydrocarbons, have been re- p0rtc.d (95). Preliminary crystnllographic rclsults (101) indicate that the donor molrcule approaches the mr~tallocarboraneon the side holding the cisoid polyhedral carbon atoins, :is might be expectrd for charge-transfer interactions. B. Monometallic Carboranes with Different Carborane Ligands Degradation of the icosnhcdral 1,2- and 1,7-C&o& isomers x\ith strong base to producc. the niclo c~lcvt~ii-vc~texanions 7,8- and 7,%C&H12- has licc~ipreviously discussed and is an important route to thc prcparation of twclve-vertex mononictallocarboranes. The discovery that similar reac- tions could be performed on mctallocarboranes led to the isolation of novel chains of metal atoms bridgd hy rarborane groups and to the development of the polyhedral contraction and poll hcdral subrogation reactions. The first indication that nwt allocarborancs could undergo base degrada- tion was found upon investigation of thc colored products remaining after isolation of the (1 ,2-C2BgHll) :( lo- ion from an aqueous preparation (32, 34). The high molecular wight, complrxes found in the reaction residue contained species forrned by partial dcgradation of the icosahedral com- plex, which had undergone flirt hvr I (.action to produce polymctallic com- pounds. The structurcs (12, 89) of the two reaction products isolated are shown in Fig. 11. The following reaction srqucwcc I~adingto these complexes is believed to bc similar to the stc.pwise rcactions in which metallocarborme complexes are formed from neutral closo-carboranes : (1,2-C213911,,) KO- + 20H- + H 2O t (I ,2-('21391311) (C,I&Hlo)Co3- + H2 $- II(0II)y (13) (1 ,~-C~~SIII,)(CL&~TIO)CO~- f ("IJ(11) $- 7,8-Ci13gFI112- Lo1 t (1 ,2-C213qITl,) CO(C',I3,H1,) C,)(l,2-C?B9Hii)?- (14) 162 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE 2 OH- co2+ + 3 CH 'BH FIG.11. Degradation of (1 ,2-C2BgR11)2C~)-to bi- and trinietallic compounds. (1 ,2-C%B~TIll)Co(C&€llU)CO( 1 ,%c%ngH~,)~-+ OH- $- H20 . ( 1 ,2-CzH 9Hi1)CO (C LBBHIU) CO (C2BSHin) 4- ( 15) (1,2-CsRgH11) CO(CzRxR in) CO(C2B8H10)4- + CO(I1 1 + 7,8-C2Bd-11i2- [ol (1,2-C2139Hii)C0 (CLUBHIU)CO(CLB,H~U)CO (I ,2-C~BgHii)~- (16) This reaction sequence may be considered the first example of polyhedra1 subrogation in that a boron vertex of a metallocarborane has been degrada- tively removed from the polyhedron and replaced by a transition metal vertex. No overall change in size of the polyhedron occurs in the reaction, however. Evidence substantiating this proposed mechanism has been gathered by the success of polyhedral contraction and subrogation reactions on this same substrate (66-68).Further examples of polyhedral subroga- tion reactions in the synthesis of polymetallocarboranes will be discussed in Section 111, D. Mixed commo-metallocarboranes of thc general formula (CzBgH11) Co- ( C2BXH10)- have been prepared by the polyhedral contraction of isomeric (CzBgHll)2Co- ions (68). These species react with pyridinc and other Lewis bases to afford nido-metallocarboranes that have the Lewis base attached to a boron atom on thc 5-boron cage. A crystallographic structure determination has been prrformcd on one of thcsc ligand adducts (9). Spectral evidence indicatrs that the carbon atom locations in the CzBsHloCo polyhedron of the (1 ,2-CzBgHll) Co ( CJ3XHlo)-ion are idcritical with those observed in the product from polyhedral contraction of CSH&o( 1,Z- C2BgH1,) (Srction VI, A) (66). Reaction of this Bs-Bg closo-metallocar- boranc anion with FeCI3 in ethanol affordcd (1,2-C,BgH,1>Co(1,6- Ten Years of Metallocarboranes 163 CzB7Hg)-, hicli could bc rcmraiigcd to ( 1 ,2-C‘zBgHll) (”0(1 ,10-C‘zB,Hg)- (66). KOintcrmcdiatc nido compl(w,s\vcrc noted in this latter transforma- tion, in contrast to thr cyclopc,iitntlic.ii~lanalog (sw Scction YII) although these species haw been prcpmd (67) and appear to have both B-H--R arid B-H-C‘o bridge bonds. C. Mixed-Ligand Complexes The class of compounds know11 as nioiioinetallocarboraries is not limited to spccics that contain tu o carhorancl ligands bonded to a trailsition mctal, but also encompasses compoiintls that contain both carboraric fragments and other, predominantdy organic, ligarids. Th(. first cyclopc.ntadic.riy1 nwt allocnrborane to be synthesized empha- sized the analogy betwwn mc~tallocmcchcmistry and metallocarborane chemistry. This preparation simply involved the reaction of stoichiometric amounts of thr two ligands with u transition mctal halide: 7,8-(”213$Hii2- + r,FI,- + vP(I1)L C,H,l”c(l ,2-(’213J11,,) + (C.H,)LFe+ (I ,2-C2B9HI1)2l”e- (17) As would be expected from this synthttic route, considerable amounts of syinmetrically substitutcd products \wre obtained along with the desired mixcd-sandnich complex (49).This synthetic route was succrssful for the preparation of other mixd-sand\\ irh CJ clopmtadienyl met allocarborancs with the transition metals cobalt (111) (54) and nickel(II1) (98),although yields were relativrly low. Modifid preparative schrmm, developed ovcr thc past several years, now pc.rmit thc synthesis of C5H,Co (1, 2-CzB9H11) in high yield, starting with tit1ic.r thv potassium salt of 7,X-CzB9H12- or with 1,2-C2B,,Hl2 itself (69,81). These iicw routcs have becn of great value in recent dcvclopments of mctnllo(,:ir~)orancchemistry, since neutral dia- magnetic CjH,Co( l ,2-CaB9Hll)is an ideal starting material for the prcpa- ration of other mctallocarboranc~s. As observed with the icosahrdral carborane 1,2-CzBloH1z, the mix(& sandw ich metallocarboritn(:~rane C ’BH.,(‘o( 1 ,2-CzBgH11) undergom thermal polyhedral rearrangemcnt to producc eight other isomeric species. All of thew possible positional isomers of this complcx have becn prepared and isolated by pyrolysis and chromatographic srparation (70). The thermal rearrangements arc dcpictcd in Fig. 12. Isomeric idcntification was effected primarily by ”13 NMR spectroscopy and clcctrochemical and chromato- graphic behavior. As was found in the rclarrangemcnt of 1,2-C2BloH1z, the polyhedral carbon atoms in C,H,<’o (1 ,2-C2B9Hll) migrate away from each 164 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE FIG.12. Isomers of C~I13Co(C2T3911,~)prepared by thermal rcarrangciuent. other during thcrmal rcarrangcnicnt. In order to isolate examples of all thc possible isomers, it was, thcrcfore, necessary to link the carbon atoms with a trimethylene bridge, forcing them to stay adjacent whilc still permitting migration over the polyhedral surface. Another routc to some of these isomeric complexes is through thc polyhedral expansion reaction : two C5H5CoCzByHnisomers were prepared by polyhedral expansion of 2 ,3- CzB&n (24). Although mixed carboranc-cyclopentadicnyl cornplexcs, prcdominantly those of cobalt, arc tho metallocarboranes that have received the most Ten Years of Metallocarboranes 165 study (due primarily to their air aid uatvr stability, neutrality, and dia- magnc6sm) , other mixc.d-ligand I tic~tallocarborancsmay be prcxparcd as well. Metal carbonyl compounds rwct \t ith carboraiie dianions to afford inctallocarboranc carbonyl complrws (45,54), for example, 7,8-C213911112-+ 131Mn(CO)4-13t- + (~'L1~~~Hl11\1ii(C0)>-) t (I,~-C~I~~H~~)~~I~(CO)~-+ 2C0 118) 7,8-CJ39IIllZ- + J\ (CO)G&L~('O+ (I ,~-CL~~~III,)\~(CO)~'-1 19) Thc cxxistcncc of the lLln(C0) ,-cmhrairc. complex was inferred (54) by the observation of rapid prccipitnt ion of NaBr from th(1 rcaction mixture, whcrcas CO cvolution did not occiir lint il thr rcwlting solution was hcated to rcflux. The photochemical synthcsis of tliv t ungstcn carboiiyl metallocarborane [Eq. (18)] is also effective in tlio prqxtration of the molybdcnuni carbonyl analog. Thf rcsulting air-srnsitivc coniplcxcs show chcmical bchavior simi- lar to tht. cyclopcntadienyl aiialogy, C',,HjM( C0)3-, in that they undergo protonation with anhydrous H( '1 :tnd imthglation with CH3I. They also react further with metal hcxacnrboiigls to afford biiiictsllic conipl(,xcs (5$): (~,2-~~l~~€~ll)~~o~~~~)~~-+ JV(CY)),>-(I ,2-C?1~9~Il]~~l~)(C(~)~\\ (CO) ,2- + co 120) Scvrral other twlve-vertex i~i(~t:~llocar~)oraii(,carbonyl coniplcxcs have bccm prcparcd (51). In gencr:d, thc. csh(1mistry and structures of thcw species, whcn investigated, haw hii found to parallel the analogous cyclopc~ntadicnylmetal carboiij Until rcwntly, little work had hccw rcyortcd on the preparation of mctallocarborancs containing morc' divcme counterligands. The tctra- phcii~lc~clohutadicnccoinplcbx [((',$Id) IC"t]Pd (1, 2-C&,H11) was reportcd in 1966 (9(i),and a cyclooctadioiic~roniplcx of platinuin(I1) was suhsc- qucntly prepared (95). Only in thcb past 2 years have thc cxistcncc of phosphine-substituted cZoso-mc.tnllocnrl)oiarics been established, and a rich ncw area of metallocarboraiici clicaiiiistry thereby developed. Rcwtion of the Iziclo-carboraiic nionoaiiio~i7 ,8-CzB9H1a- n ith phosphinc- substitutcld transition metal coniplrxw afforded novel hydridophosphino- metallocarborancs (80) according to [(C"~II,)~l']~RllCl+ 7,8-C>13glIl2- - IIII~[(C~H~)~~']?(~,~-C~B~II~I) (21) This complex, along with its iridiim arinlog, has provm to hr. an effective catalyst for the isomerizatioii aiid hydrogc~nationof olcfins (80),for H-D exchange at BH vertices (GO), and for othw organic reactions. Catalysis by nictallocarboranes is discussed in dvt ail in Section X. 166 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE Spencer, Green, and Stone (88) have prepared nickelacarboranes with phosphine and other counterligands by reacting the closo-carborane 2,3- CzBsHll with organornetallic nickel compounds. Grimes (76) found that the reaction of 1 ,2-CzBloH~2with CsHjCo(CO)2in a sealed tube afforded a mixture of the isomeric C6H5Co( C2BgHll)compounds. D. Bimetallic Complexes Examples of bimetallic twelve-vertex mctallocarboranes have bccn pro- vided by various synthetic efforts. The preparation of bi- and trimetallic chain complexes by polyhedral subrogation of ( 1 ,2-C2&Hll) 2Co- has been mentioned earlier (33, 34), and the { Co(C2BsHlo)Co) fragment present therein may be considered as an example of this type of bimetallic complex. Cyclopentadienyl analogs of the bi- and triinetallic chain complexes were prepared recently (69).Thcse Compounds, as well as the carborane- capped compounds, exhibit revcrsible 1-electron redox couples, indicating that the electrons arc added to and removed from delocalized molecular orbitals rathcr than localized metal centcrs. A second routc to bimetallic twelve-vertex complexes is via polyhedral expansion. The starting material may be either an cleven-vertex mono- metallocarboranc, 1. 3e- C jH :CO (CzBgHio) + (C,H,)zCoX2138Hlo + other products (22) 2. CoUI), CsHa-, [O] or a ten-vertex ncutral carboranc, 1 2e- 1,6-CzlhH10 + (C,H,) zCoK'2B,FI,o + other products (23) 2 Co(II), C,H5-. I01 but yields are substantially higher in the former case. Both routes have been used (24) to prepare the icosahedral complex 2,3- (CsH,) 2-2, 3-Co2- 1,7-CzBsHlo, which has been shown byoX-ray crystallography to have the metals in adjacent positions only 2.39 A apart (7).At 250°C, this complex thermally rearranges to an isomer in which the metals are no longer adja- cent (50).The identical product is obtained upon pyrolysis of the bimetallic product prcparcd by double polyhedral subrogation (69) (Fig. 13). The discovery of the thermal metal transfer reaction (27) afforded a third preparative routc to bimetallic twelve-vertex metallocarboranes. This method involves the pyrolysis of eleven-vertex closo-metallocarboranes or cobalticinium salts of eleven-vertex commo-mctallocarborancs and re- sults in thc production of sevcral isomeric, closed, neutral, twelve-vertex bimetallocarboranes [Eqs. (9) and (10) ; Fig. 71. Yields are reasonable in Ten Years of Metallocarboranes 167 FIG;. 13. Structures and interconversions of three isomeric (CsHj)2CoLC2B8H10 com- pounds. this reaction, and thermal metal transfer could prove to be a valuable new synthetic tool. Only one trimetallic twelve-vertex metallocarborane has been reported. This species, ( C5H5)3C03C2B7H9, arose as a side product during the poly- hedral expansion of 2-C5H5-2-Co-1 ,6-CzB7H9 with Co (11) and C5H5- (25, 28). The isolation of this trimetallic complex suggests that the polyhedral expansion reaction may be extended to bimetallic substrates to produce novel metal-rich polyhedra. IV THIRTEEN-VERTEX METALLOCARBORANES The first supraicosahedral monoiiietallocarborane was prepared by poly- hedral expansion of 1 ,2-CzBloH12with CoClz and C~HF,-(17, 18) : 1. 2e- 1,2-CrBioH1r + C,HsCoC2BioEIiz (24) 2. Co(II), CKHS-,lo1 Three isomeric complexes of this formula have been prepared, which are 168 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE BH FIC;.14. Structure of the red isornw of C,H:CoC,B,,Hn. colored red, orang(', and amber. All three cxxhibit the same melting point, 250"-251"C, but upon heating, the red complex is wen to change color to orang(', then ambcr before mdting, and the orange isomer turns amber before it melts. The red isomer is quantitatively converted to the orange isomer by boiling in hcxanc (68"C), and thc orange complex rearranges to the ambcr isomer on licating in benzcnc solution (80°C). These rearrangc- ments occur at surprisingly low tmipcratures: recall that rearrangement of 1,2- to 1,7-CzBloHlz rcquired a tcmpcirature in excess of 400°C. The thcrmochcmical paramctm of those rearrangements have been drter- mined, and rdlect an intramolccular mechanism, as expected (20). The structurc of thc red isomer of CjH&oCzBloHlz (Fig. 14), determined by Churchill (8),indicatc.d that the complex was asymmetric, although "I3 NMK spcctroscopy indicated mirror symmetry for this species in solu- tion at room temperature.. Low-tcmpcrature IlB NMR spmtral invcstiga- tiom showed that the molecule was fluxional in solution at room tempera- ture. It has bccn postulated (28) that the. complex undergoes rapid dia- mond-squarc-diamond rcarrangemcnt (Fig. 15) in which cnantiomcric complexes intcrconvert. Thirteen-vertex complt.xrs of metals other than cobalt have been pre- pared, as have comnio t hirtccn-vertex anions and metal carbonyl deriva- tives (18).In gcmcral, these coniplcxes cixhibit stabilities similar to or less than thcir twelve-vertcxx countcxrparts. Thc chemistry of the thirtccn-vertex monometallocarboranes, although less cxtensivcly studied, is similar to that of the twclve-vertex analogs. Thus, CjH5CoC2Rl,,H12undergoes polyhedral subrogation to produce thirtwn-vcrtcx bimct allocarboranes [the metals may be idrntical (23) or Ten Years of Metallocorboranes 169 =%- a FIG.15. Proposed merhanisrn of i.:tc.c’iiiiZ;itioi~ of the rrd isomer of C,IIiCoC2BlJI12 through a diamoiid-square-diamond iiit rrnirtliate. they may differ (19)]: C,H,Co(’21310111~+ 011- + I:t.(LI) + ~:,lI,-IO]((C,HB)~CoE’eC,BsHll (25) Polyhedral expansion of the thirttwi-vertex monomctallocarborane with Co (11) and CjHj- produced thv novd fourteen-vertex bimctallocarboranc (C‘jHj)2CoGBloHlz (29).Thrsc spc1cic.s are discussed in Section V. Attempted polyhedral contraction of (’.,H6CoC2Bl0HlZ,expected to pro- duce one or more of the isomeric tw~lvc~-vertcxCjHjCoCzB9Hll complexes, instead resulted in extensive disruption of thc polyhedron. Three boron vertices and one carbon vcrtcx wrc extruded from the starting material, and a novel monocarbon mctallocarborane ion, CjH,CoCB7H8-, was iso- lated (21,22). As expected for a nino-vertex complex, this species has the geometry of a tricapped trigorial prism, with the carbon atom in a low- coordinate capping site (Fig. 16) (6). When subjcctcd to the action of alcoholic base and excess Co(I1) in the abscncc. of additional cyclopcmt adic.nidc ion, the thirteen-vertex mono- mctallocarborane reacts to produce 5 trimetallic complex that contains tu o bridging thirteen-vertex framcwwrks ($5). The proposed structure of this species is shown in Fig. 17. This reaction is similar to that used to prepare the twelvc-vertex bi- ant1 trimctallic chain complexes discussed in Section 111, B. Polyhedral expansion of 1,2-(12B10H12 in the presence of early transition metal halides has been found to producc. the first examples of metallo- carboranc complcxcs of these elcnicmts. Expansion using TiC1, afforded a titariacarboraric in which thc inctal has it formal oxidation state of 2+ and only 14 valence electrons (86). 170 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE (9- B6H 84 H FIG.16. Structure of C5HjCo('B7Hs- This electron deficiency is not structurally significant, however; crystallo- graphic investigation of the complex formed between 1,2- (CH3)2-1, 2- CnBloHlo and TiCL showed the thirteen-vertex polyhedra to be symmetric- ally disposed about the metal (74) and to exhibit less distortion than ob- served in the CsHsCoCzBloHl~polyhedron (8). Complexes of V(I1) and FIG. 17. Proposed structure of the trimetallocarborane prepared by polyhedral subrogation. Ten Years of Metallocarboranes 171 Zr (11) have been similarly prepared and appear spectrally comparable to the Ti (11) compound (86). V FOURTEEN-VERTEX METALLOCARBORANES The polyhedral expansion of the orange and amber thirteen-vertex closed metallocarborane isomers with Co(I1) and CsH,- was found to afford complex mixtures of products, the primary species in each case being the desired isomeric (C5H5)2C02C2B10H12complexes (29).Spectral data are consistent with a closed fourteen-vertex geometry for these compounds, but definitive structure determination by X-ray crystallography will be required to determine the overall geometry of this large polyhedron. A possible structure of one of the isolated isomers which is consistent with the spectral properties (29) is prcsented in Fig. 18. VI ELEVEN-VERTEX METALLOCARBORANES A. Monometallic Complexes The first reported preparation of an eleven-vertex closed monometallo- carborane involved the polyhedral expansion of the ten-vertex closo-car- OCH OBH FIG. 18. Preparative method and proposed structure of one isomer of a fourteen- vertex metallocarborane. 172 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE FIG.19. Structure of C,H:CoC2B8H,, prepared by polyhedral expansion borane 1,6-CzBxH,o with Co (11) and CEH- (24,26): 1 2e- 1,6-C2n~Hi0 + C,H .CoCJ3,Hjo (27) 2 Co(II),C5H5-, I01 The structure of this compound, inferred from IlB NMR spectroscopy (26) is shown in Fig. 19. The metal resides at the high-coordinate vcrtcx, while the 2 polyhcdral carbon atoms occupy the two low-coordinate positions in the closed eleven-vertex geometry. The analogous Fe (111) compound was formed using Fe(I1) in the polyhedral cxpansion reaction (24). Another eleven-vertex complex, isomeric with the cobalt compound just mentioned, may be prepared (66) by the polyhedral contraction of C5H5Co(1, 2-CzB9Hll). Reaction of this substrate with strong aqueous hydroxide ion in the absence of additional Co(II), followed by hydrogen peroxide oxidation, resulted in the isolation of a new C5H5CoCzB,H10 isomer in which both polyhedral carbon atoms occupy adjaccnt sites, one in a low- coordinate position and one in a six-coordinate site (68). This isomer is thermally unstable and rearranges (26) to thc isomer produced in the polyhedral expansion of 1,6-CzB8Hlo with Co (11) and CSHS-, when heated to 150°C. The corresponding commo complexes have also been isolated as side products from the polyhedral expansion of 1 ,6-CzBeH10 (24). The nido eleven-vertex complex, CSHsCoCzBxH12,has been prepared, and spectral data indicate the prcsence of B-H-B and B-H-Co bridges (67). The commo eleven-vertex metallocarborane, (CZBxHlo)2Co-, pre- pared by polyhedral expansion of 1,6-CzBxHlo with Co (11) in the absence of cyclopentadienide ion, has carbon atoms in positions identical to the sites in the C5HSCoCzBxH,isomer prrpared by polyhcdral expansion in the presence of CsHS-, i.e., at low-coordinate vertices (26). Ten Years of Metallocarboranes 173 Reactions of monometallic eleven-vertex metallocarboranes have been discussed in previous sections and may be summarized briefly as (a) poly- hedral expansion to bimetallic twelve-vertex complexes and (b) thermal metal transfer to bimetallic twelve-vertex compounds. Polyhedral con- traction to ten-vertex monometallocarboranes is discussed in Section VII. 6. Bimetallic Complexes The eleven-vertex bimetallocarboranes have primarily been synthesized by application of polyhedral expansion techniques. One obvious synthetic route, polyhedral expansion of a ten-vertex closo-monometallocarborane, has been investigated in detail (64).The products arising from this reaction are shown in Fig. 20. Three isomeric eleven-vertex cobaltacarboranes have I 3ea 2 CO .C5H5- + 3 [OI + + + + FIG.20. Polyhedral expansion of CbH5Co(l,6-C2B,Hs). 1 74 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORN€ been isolated in addition to the trimetallic icosahedral complex mentioned previously, as well as a positional isomer of the starting ten-vertex mono- metallic species and a bimetallic nine-vertex compound resulting from polyhedral degradation. Polyhedral expansion of the same ten-vertex starting material, employ- ing Fe(I1) rather than Co(II) as the entering transition metal, led to the preparation (19) of the heterobimetallocarborane (CjHs) &oFeC2B7HS: The structure of this complex has recently been determined by X-ray crystallography (73) and shown to be that depicted in Fig. 21. Interest- ingly, the iron atom occupies the high-coordinate vertex, a result consistent FIG.21. Structure of (C6H&CoFeC2B7Hs. Ten Years of Metallocarboranes 175 with the fact that it formally possesses a dj electronic configuration and is thus more electron-deficient than the ti6 Co(II1) vertex. Other isomeric eleven-vcrt ex binwtallocarboranes have been prepared, and the thermal rcarrangcmcints of these compounds have bccn studied in depth (30). Thcrc appears to h a major driving force, in all cases, for the polyhedral carbon atoms to occupy the low-coordinate vertices of the octa- decahedron and for a metal atom to occupy the high-coordinate site. Once this arrangement has been achicved, f urthcr thermal isomerization results in migration of the second mttal vertex over the polyhedral surface to posi- tions at greater distance from the high-coordinate metal (Fig. 22). These rearrangement tendencies haw b(wi rationalized in terms of decreasing electrostatic repulsions and more favorable electronic environments (46). VII TEN-VERTEX METALLOCARBORANES Thc original preparation of ten-vcrtc.x metallocarboranes involved the deprotonation of the nido-carboranc 6,S-C2B7H13, followed by reaction with a transition metal halide. Additional hydrogen gas was liberated during the complexation, and closo-met allocarboraric compounds were formed (38): 2NaH Co(I1) 6,8-CzB,H,,- CLR~1III'- (C&IIg) LCO-+ €12 (29) 101 Mixed-sandwich complexes could he synthesized by adding cyclopentadie- nide ion to the reaction mixturc (38, 48). These ten-vertex monomctallic complexes exhibit thermal polyhedral rearrangements similar to those mentioned previously. In fact, they were 176 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE Frc. 23. Polyhedral rearrangements in the CJIZC~CIB~H~system the first metallocarboranes found to isomerize thermally (37). Both the commo salts and the cyclopentadienyl compounds underwent carbon atom migration when heated to 315°C. Rearrangements in thisgeometry (20) con- sist of movements of the polyhedral carbon atoms to the low-coordinate capping positions of the bicapped square antiprism. The metal has always been found to occupy a high-coordinate equatorial position in these com- plexes (Fig. 23) (46). Another route to ten-vertex monomctallocarboranes is through poly- hedral expansion of 4,5-C2B7Hg. A new isomer of C5H5CoC2B7Hgwith car- bon atoms in positions 3 and 10 is produccd in this reaction. Side products included the nine-vertex species C5H5CoC2B6H,(Section VIII) and the eleven-vertex C5H5CoC2B8Hlo(24). Polyhedral contraction of CsH5Co(2,4-CzBsHlo) proceeds through an unusual intermediate on the way to the closo ten-vertex complex C5H5Co- CZB7Hg (66). This intermediate has been shown (5) to be a nido-metallo- carborane, C5H5CoC2B7Hll,which has both B-H-B and B-H-Co bridge bonds (Fig. 24). On heating this compound, hydrogen gas is evolved and the polyhedral contraction product CjH5CoCZB7H9is isolated. Hbl Hb2 CP I A P, (I: C CP4 Frc. 24. Structure of the nido-metallocarborane CjH,CoCnB,H,I. Ten Years of Metallocarboranes 177 A novel ten-vertex birnrtallocarborarie, (C,Hj) 2C02C2B6Hs, is produced upon the polyhedral expansion of 1,7-C&& (16). The expected mono- metallic nine-vertex species, (?&c'oC2R6Ha, is also produced in this reac- tion (see Fig. 4). An X-ray crystal structure of the bimetallic product (61) showed the 2 cobalt atoms to occupy adjacent positions on the two equa- torial belts of the hicapped squarch antiprism. The carbon atoms occupy the lowcoordinate caps. Thermal rearrangc.ment (SO) produced an isomeric complex in which the metals arc srparatrd from each other, but the carbon atoms remain in the low-coordinate sites. Polyhedral expansion of 4,5-C's137H9with Fc(I1) and CjHj- led to the isolation of the ten-vertex binictallic iron complex ( CjHj)2Fc2C2B6Hx(2). This compound differs from thc cobalt complex discussed above in that it has 2 fewer clcctrons [cl" Fe(II1) as opposed to d6 Co(III)]. A profound structural change rcsults from this dcctron deficiency : the complex adopts a geomrtry different from thv bicapped square antiprism, although it is still classified as a cZoso-mc.tallocarborarie since it has a fully triangulated closed polyhedral structure. The iron atoms are bonded to each other at a distance of 2.571 A (Fig. 23). A paramagnetic isomer of this compound has been reported, and it was suggrstc,d that it may have a nido geometry in which the iron atoms are no longer interacting, but definitive structural data are still lacking. Ten-vertex meta1locarboranc.s containing both cobalt and nickel in the same polyhedron have becn synthesized from the nine-vertex anionic species CjH5CoCB7Hx- (Srction VIII) by polyhedral expansion (87). The resulting neutral complexcis, of formula ( C5Hj)&oNiCB7Ha and con- Fe I CP c2 Fe 6 FIG.25. Structure of (CjH&Fe2C2B8H8. 178 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE taining formal Ni (11') , exhibit thermal rearrangements to isomeric species. The structure of one isomer has been crystallographically determined : the cobalt and nickel atoms occupy positions on the samc' equatorial bclt, whereas the polyhedral carbon atom is at a capping site (73).These corn- plexcs are isoclectronic with the ( C5H,)sCozC'zB6Hg isomers previously mentioned. By hcating 2,4-C2&H7 at 260" in the presence of CjHjCo(CO)z,two isomeric trimctallic complexes formulatd as ( CjHj)3C03C2B5HT were iso- lated (76).One isomer has one Co-Co interaction with a high-coordinate carbon atom, whereas the proposed structure of the other isomer has the 3 cobalt atoms positioned on the sanic equatorial bclt of the bicapped square antiprism, bound to each other, and thc carbon atoms occupying the apical vcrtices. Polyhedral expansion of 2,4-C2R3H7 afforded, among other products, a new (C5Hj)zCozCzB6H8isomer in which thc cobalt atoms occupy nonadja- cent verticc.s on the same belt of the tcn-vertex polyhedron. One of the carbon atoms occupied an (quatorial position bctu cen the cobalt vertices (42).The other carbon atom is located at the apical position most distant from the metal-containing belt. The chemical reactions of monometallic ten-vertex metalloearboranes have been examined (40).Friedcl-Crafts acylatiori of CjHgCo( 1,6-CzB7Hg) produced a monosubstitution product. Attack occurred on the boron atom farthest removed from the polyhedral carbon vertices. No substitution on thc cyclopentadicnyl ring was observed. Vlll NINE-VERTEX METALLOCARBORANES The first metallocarborane of this gcomctry to be synthesized was an unexpectcd product. In an attempt to prepare a tcn-vertex manganese carbonyl complex, the C2B7HllZ-ion, discussed in Section VII, was reacted with BrMn (CO) 5. Surprisingly, the only metal-containing compound iso- lated from the reaction mixture had just 6 boron atoms (36, 50). The course of the reaction may be outlined as follows: 6,8-C,H7Hla 2N&H.C2137H1,2- BrMn(Co)LCL136H8XZn (CO) 3- 130) Similar results were observed when ClMn (CO) or Mn2(CO) were used as metal sources and when C-phenyl- or C ,C'-dimethyl-substituted car- boranes were used as starting materials. The structure proposed for this complex and later confirmed by a crystallographic study (63) was based Ten Years of Metallocarboranes 179 on tricappcd trigonal prismatic geometry with the Mil (CO)3 group occupy- ing a high-coordinate position and the 2 carbon atoms in positions adjacent to the metal in low-coordinatc capping vertices. The first reported examplc of polyhedral expansion also resulted in the isolation of a nine-vc'rtex nionomctallocarborane (16).Polyhedral expan- sion of 1,7-C2BGHs in the prcw.iicc1 of C'ocl2 and C&- resulted in the isola- tion of C,HjCoC2B,Hs as wdl as the ten-vertex bimetallic (CjHj)ZCo2- C2BsHsmentioned in Section VII. Thc nionoInetallic complex thus prepared has both carbon atoms in low-coordinate capping positions but only 1 carbon atom is adjacent to thc mdal vertex. The isomeric species with both carbon atoms adjacent to the cobalt atom was isolated from the mix- ture of products obtained from thc polyhedral cxpansion of 4,5-CzB7Hs with CoCl2 arid CjH5- (24) arid was also prepared by thermal rearrange- ment of the first isomer (20). Anothcr isomer is formed by expanding 2,4-C2R jH7 (75). A more unusual structurcl has hwn established for the complex C5H5Fe- CZBGHs, which was synthcsizd by polyhedral expansion of I,7-C&Hs in the presence of FcClz (24).'I'he llB NMR spectra of the Fe(II1) and Fe(1I) compounds [the 1attc.r gcncrated by Na/Hg reduction of the Fe (111) coniplcx in silu] and thv *HNMR spectrum of the Fc (11) species indicate that the carbon atoms occupy high-coordinate positions and the iron vertex is at a low-coordinate sit(.. Thc structure proposed on this basis is shown in Fig. 26. An isoniclric product, believed to have a structure simi- lar to the C2B6HsMn(0) ,3- iun discussed in the foregoing, was also iso- lated from this reaction mixture. Commo complexes containing two fused nine-vertex polyhedra may be prepared by the polyhedral cxpansion of 1,7-C2BsHs with CoClz in the 4 FIG.26. Proposed struc%ure of the ferracarborane C;H5FeC2B6H8. 180 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE absence of cyclopentadicriide ion (24).The carbon atom positions in the two polyhedra arc identical to those in the product generated in the prcs- ence of C5H,-. Although it is not within our limitation of two-carbon metallocarboranes, mention has been made of the C5H5CoCB7Hg- ion, vhich results from at- tempted polyhedral contraction of orange or amber CjHjCoCJ310Hlz (21, 22). The structure of this anion (6),prwiously prcscntcd in Fig. 17, is again based on tricapped trigonal prismatic geometry. The lone carbon atom is not adjacent to the mcltal but occupies a low-coordinate capping vertex. Two isomeric bimetallic. ( C5H5)&02C2B,H7 compounds wcre formed by heating 2,4-C2B5H7 at 260°C in the presence of C,HsCo (CO)2. One of these complexes exhibits a metal-metal interaction (76). The other isomer, origi- nally prepared by polyhedral expansion of 2,4-CzBaH7 with CoClz (75),is also formed in the polyhedral expansion of C5H5Co(1,6-CzBTHg) (25). Reaction of the C2BIH7- monoanion with NiBrz and C5H5- afforded, among other products, the dinickelacarborane ( C5H5)zNizC2B5H7 (42). This electron-rich complex is diamagnetic, indicating intermetallic inter- action, but a lengthening of the Ni-Ni bond was suggested as a possible geometric distortion, as expectcd in this compound because of its excess electron dcnsity. IX OXIDATIVE ADDITION TO B-H BONDS Early in the development of polyhrdral boranc and carborane chemistry the need arose to degrade these materials under protolytic conditions for analytical purposes. Such reactions, for example, B,Hn2- + 2H+ + 3n-ltC00HPt-n-B(OC011)3 + (2n + 1)-111 (31) were' useful in establishing complete empirical formulas based on total hy- drogen evolution. Oddly enough, reactions of this sort would proceed at a reasonable rate only in the presence of metallic platinum or palladium catalysts. It appears as though the protolytic degradation reactions may proceed through a series of oxidative addition reactions in which the metal catalyst is inserted into B-H bonds and, possibly, into the borane cage itself. The recent results described in the following support this view. The now well-known ort ho-metalation reaction observed in certain low- valent transition metal complexes of arylphosphincs (I, 79) provided a model for the observation of oxidative addition to terminal B-H bonds. Ten Years of Metallocarboranes 181 H.6 IIo L=Io,h=H b L;Ib,h:H c L=Ic,h=H0-11 FIG.27. Oxidative atldition to a polyhedral HH vertex. The phosphine 1- (1, 2-CzBloHll)P (CHS) was prepared and reacted with an iridium (I) complex, as shown in Fig. 27. Spectroscopic evidence strongly supported the view that the complexed Ir(1) had inserted into a B-H bond with the formation of an iridium-boron bond and an iridium-hydride link (58). Specific deuterium labels attached to the carborane moiety of the phosphine clearly proved this point. Thus, the B-H vertices of the icosahedral carborane group that arc nearest the carbon atom bearing phos- phorus were shown to be involvcd, although a distinction could net be made between the 3,6 or 4,5 sets of BH groups. In another series of experiments (59), toluene solutions of 1-(1,2- CzBloH11) P ( CHs) were reacted with deuterium gas in the presence of (Ph3P) 3RuHC1. Mass spectral evidence proved that up to eight deuterium atoms had been introduced into the phosphine. Since only four BH ex- change sites could possibly be involved in deutcrium exchange which pro- ceeded only through ortho-metalation intermediates, an intermolecular, reversible oxidative addition of catalyst to terminal B-H bonds was indi- cated. Indeed, further work proved that under the same conditions 1,2-, 1,7-, and 1,12-CzBloHl2 carboranes could be totally deuterated at boron 182 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE Ph3PL H‘Ir H‘Ir /PPh3-CI I R3 FIG.28. Structure of‘ the pi’oduct of oxidative addition to 1 ,2-C2B1~)H12. vertices, as could the imtallocarboranc C5HjC‘o(1, 2-CzB9Hll) ~ In fact, every B-H-containing compound examined to date was found to be sus- ceptible to catalyzed deuterium exchange [i.c., Bl0HI4, Bl0HlO2-, and (CI33)&BHs], and no cxchange was observed at carbon. Thcsc results lead one to the somewhat surprising conclusion that terminal BH groups are much more reactive with low-valcnt transition rnctal complexes than arc ordinary CH groups. With the foregoing results in hand the rcactions of the 1,2-, 1,7-, and 1,12-CzBloHlz carboranes were examined with Ir(1) compounds. In each case iridium was inserted into terminal B-H bonds to form stable a-bonds betwen boron and iridium (60): +[Ir(C,H,,) 2Cl3 + 21’hJ’ T 1 ,2-C2TL,H,,-3-[ il’hsP)2IrHCl]-I ,2-C2B10Hll (32) The relative reactivities of thc various BH groups present in each carborane were such that those BH groups nearest the carbon vertices were favored for attack. Figure 28 illustrates this point. Reaction of the B-a-bondcd metallocarboranes with CO led to the regeneration of the carborane and the formation of Vaska’s compound, trans-IrC1 (CO) (PPhz)z, 3-[ (I’hJ’) ,111ICl]-l ,2-C2B,,J€,, + CO -1, 2-C213,011,2+ trans-IrCl(C0) (I’Phs) 2 (33) X METALLOCARBORANES IN HOMOGENEOUS CATALYSIS Although the chemistry described in the foregoing deals with the oxida- tive addition of low-valent transition metal complexes to terminal B-H bonds, we now describe the formal addition of the same reagents to the Ten Years of Metollocarboranes 183 open face of the isomeric CZRJl12- ions with the formation of an Ir-H bond (80). In a scnse this rcyrcsents oxidative addition to a B-H-B bridge linkage. The most thoroughly studied system of this sort was gen- erated from Rh(1) : (1’l~al’)~lthCl+ 7,8-<‘pBy1112-- (‘1- + 3,3-(1’haI’) 2-3-11-3-Itll-1 ,2-C,ByII,l (34) The structure of the product has been confirmed by an X-ray diffraction study (92). The iridium analog was also prcpared in a similar fashion. Although the hydridorhodHcarboraii(1 is formally a rhodium (111) deriva- tive, it functions as a facile catalyst in alkenc isomerieation, hydrogenation, hydroforniylation, and hydrosilglation reactions (80). This catalyst sys- tem is extremely stablc and maj‘ be recovered quantitativcly from alkene isomerization and hydrogenation rractions. In addition to these reactions, the hydridorhodacarborane is vcry effcctivr in the catalysis of deutcrium exchange at terminal 13H positions (59). These discoveries may soon lead to industrially useful mctallocarboranc catalysts. I. L(erinet,t,hl. A,, and hfilner, 11. I,., J. :lvwr. Chem. Soc. 91, 6983 (1969). 2. Callahan, K. l’., Evans, W. ,J., 1’0, F. Y., Strouse, C. E., and I-Iawthorne, A1. F., J. A nier. C‘hem. Soc. 97, 296 (1975). 3. Callahan, K. l’., and Hawthoriicx, 11. F.,J. Orgunotrietul. Cheni. 100, 97 (19751, and references therein. 4. Callahan, I<. P., and IIawtIiornr, 11. F., 1’m-e dppl. Chenz. 39, 473 (1974) and ref- erences therein. 5. Call:than, I<. l’., Lo, F. Y., Sinis, A. I,., Strouse, C. E., and Hawthorne, XI. F., Inorg. Chern. 13, 2842 (1974). 6. Callahan, K. l’., Sinis, A. I,,, Strouse, C. E., and Hawthorne, XI. F., Inory. Chem. 13, 1393 (1974). 7. Callahan, I<. l’., Sinis, A. I,., Strouse, C. I<:., and IIawthorne, &I.F., Inorg. Chenz. 13, 1397 (1974). 8. Churchill, 11. R., and DeBoer, 13. G., Inorg. C‘henz. 13, 1411 (1974). 9. Churchill, 31. It., and Gold, I<., Inorg. (“hem. 12, 1157 (1973). 10. Churchill, R.I. It., and Gold, I<., J. Amfw. Chern. Soc. 92, 1180 (1970). 11. Churchill, If. It., Gold, K., Francis, J. N., and fIawthorne, 31. F., J. Amer. Chetn. Soc. 91, 1222 (1969). 12. Churchill, M.R., Iteis, .4. H., Jr., Francis, J. N., and Httxthorne, 3f. F., J. Amer. Chrrti. Soc. 92, 4993 (1970). 13. Curphey, T. J., Santer, J. O., Roscnblum, hl., and Richards, J. IT., J. Amer. Chem. Soc. 82, 5249 (1960). 14. DeBoer, 13. G., Zalkin, -4.,IIopkins, T. E., and Templeton, D. H., Inorg. Chem. 7, 2288 (1968). 15. Dewar, M. J. S.,Ri~ll. Soc. Chitti. France 18, 79 (1951). 16. Dunks, G. B., and IIawthortie, 11. F., J. Anter. Chem. Soc. 92, 7213 (1970). 184 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE 17. Dunks, G. B., McKown, M. M., and Hawthorne, M. F., J. Amer. Chem. Soe. 93, 2541 (1971). 18. Dustin, D, F., Dunks, G. B., and Hawthorne, M. F., J. Amel. Chem. Sor. 95, 1109 (1973). 19. Dustin, D. F., Evans, W. J., and Ilawthorne, M. F., J. Chem. Soc., Chem. Commun., 805 (1973). 20. Dustin, D. F., Evans, W. J., Jones, C. J., JViersema, It. J., Gong, H., Chan, S., and Hawthorne, M. F., J. Amer. Chem. SOC.96, 3085 (1974). 21. Ihstin, D. F., and Hawthorne, h4. F., Inorg. Chem. 12, 1380 (1973). 22. Dustin, D. F., and Hawthorne, M. F., J. Chem. Soc., Chem. Commun. 1329 (1972). 23. Dustin, D. F., and Hawthorne, M. F., J. Amer. Chem. Soc. 96,3462 (1974). 24. Evans, W. J., Dunks, G. B., and Hawthorne, 31. F., J. Amer. Chem. SOC.95, 4565 (1973). 25. Evans, W. J., and Hawt,horne, M. F., Inorg. Chem,. 13,869 (1974). 26. I’hans, W. J., and Hawthorne, M. F., J. Amer. C‘hem. Soc. 93,3063 (1971). 27. Evans, W. J., and Hawthorne, M. F., J. Amer. Chem. SOC.96,301 (1974). 28. Evans, W. J., and Hawthorne, M. F., J. Chena. SOC.,Chem. Commun. 706 (1973). 29. Evans, W. J., and Hawthorne, M. F., J. Ch.em. SOC.,Chem. Cornmiin. 38 (1974). 30. Evans, W. J., Jones, C. J., Stibr, B., and Hawthorne, M. F., J. Organometal. Chem. 60, C27 (1973). 31. Fein, M. M., Bobinski, J., Mayes, Y., Schwartz, S.,and Cohen, RI. S., Inorg. Chem. 2, 1111 (1963). 32. Francis, J. N., and Hawthorne, hf. F., Inorg. Chem. 10,594 (1971). 33. Francis, J. N., and Hawthorne, M. F., Inorg. Chem. 10, 863 (1971). 34. Francis, J. N., and Hawthorne, M. F., J. Amer. Chem. Soc. 90, 1663 (1968). 35. Francis, J. N., Jones, C. J., and Hawthorne, M. F., J. Amer. Chem. SOC.94, 4878 (1972). 36. George, A. D., and Hawthorne, M. F., Inorg. Chem. 8, 1801 (1969). 37. George, T. A., and Hawthorne, RI. F., J. A7ner. Chem. Soc. 90, 1661 (1968). 38. George, T. A., and Hawthorne, RI. F., J. Amer. Chem. Soc. 91, 5475 (1969). 39. Grafstein, D., and Dvorak, J., Inorg. Chem. 2, 1128 (1963). 40. Graybill, B. M., and Hawthorne, M. F., Inorg. Chena. 8, 1799 (1969). 41. Grimes, R. N., “Carboranes.” Academic Press, New York, 1970, and references therein. 42. Grimes, It. N., Beer, D. C., Sneddon, L. G., Miller, V. It., and Weiss, R., Inorg. Chem. 13, 1138 (1974). 43. Hanson, F. V., Hazell, It. G., IIyatt, C., and Stucky, G. D., Acta Chem. ScantE. 27, 1210 (1973). 44. Hawthorne, M. F., and .4ndrews, T. D., Chem. Commun. 443 (1965). 45. Hawthorne, M. F., and Andrews, T. D., J. Amer. Chem. Soc. 87, 2196 (1965). 46. Hawthorne, M. F., Callahan, K. l’., and Wiersema, It. J., Tetrahedron 30, 1795 (1974), and references therein. 47. Hawthorne, M. F., and Dunks, G. 13., Science 178, 462 (1972), and references therein. 48. Hawthorne, M. F., and George, T. A,, J. ilmer. Chem. Soc. 89, 7114 (1967). 49. Hawthorne, M. F., and I’illing, R. L., J. Amer. Chem. Soc. 87, 3987 (1965). 50. Hawthorne, M. F., and l’itts, A. n., J. Amer. Chem. SOC.89,7115 (1967). 51. Hawthorne, M. F., and Ruhle, H. W., Inorg. Chem. 8, 176 (1969). 52. Hawthorne, M. F., Warren, L. F., Jr., Callahan, K. P., and Travers, N. F., J. Amer. Chem. Soc. 93, 2407 (1971). Ten Years of Metallocarboranes 185 53. IIawthorne, 31. F., and Wegner, 1’. A,, J. Amer. Chenz. Soc. 90, 896 (1968). 54. Hawthorne, 11. F., Young, D. C., Andrcws, T. D., Howe, I). V., l’illing, R. I,., I’itts, A. D., Iteintjes, M., Warren, I,. I!’., Jr., and Wegner, 1’. A,, J. Amer. Chem. Soc. 90, 879 (1968). 55. Hawthorne, XI. F., Young, D. C., Garrett, 1’. M., Owen, D. A,, Schwerin, S. G., Tebbe, F. N., and Wegner, P. A4.,J. Arrtc~.Chem. Soc. 90, 862 (1968). 56. Hawthorne, M. F., Young, 1). C., and \\’egner, 1’. A,, J. Amer. Chem. SOC.87, 1818 (1965). 57. Heying, T. L., Ager, J. IV., Jr., Clark, 8. L., Mangold, 1). J., Goldstein, H. I,., Hillman, M., Polak, It. J., and Rzyrnanski, J. \V., Inorg. Chent. 2, 1089 (1963). 58. Hoel, 13. I,., and Hawthorne, 3f. F., J. Anier. Chem. SOC.95, 2712 (1973). 59. Hoel, E. L., antl Hawthorne, M. F., .I. iimer. Chem. Soc. 96, 4676 (1974). 60. Hoel, 15. L., and Hawthorne, Rf. F., J. Aiizcr. Chxm. Soc. 96, 6770 (1974). 61. Hoel, E. L.,Strouse, C. R., and Hawthorne, M. F., Inorg. Chem. 13, 1388 (1974). 62. Hoffmann, R.and Lipscomb, \V, N., J. Chrm. Php. 36, 3489 (1962). 63. Hollander, F. J., Templeton, 1). H.,and Zalkin, A., Inorg. Chem. 12, 2262 (1973). 64. Howe, D. V., Jones, C. J., IViersenia, 13. J., and Hawthorne, M. F., Inorg. Chen. 10, 2516 (1971). 65. Hughes, R. I,., Smith, I. C., antl Lawless, E. IV., “Production of the Boranes and Related Itesearrh” (R. T. liolzmann, ed). Academic Press, New York, 1967. 66. Jones, C. J., Franris, J. N., arid Hawthorne, M. F., J. Amer. Chem. Soc. 94, 8391 (1972). 67. Jones, C. J., Francis, J. X., and Hawthorne, 31. F., J. iliner. Chenz. SOC.95, 7633 (1973). 68. Jones, C. J., Francis, J. N., and IIa\rthornc, hl. F., J. Chenz. Soc., Chem. Conzmun. 900 (1972). 69. Jones, C. J., and Hawthornc, XI. I!’,, Inorg. C‘hem. 12, 608 (1973). 70. Kaloustian, M. K., JVierscrna, 11. ,J., and Hawthorne, M. F., J. Amer. Chem. Soc. 94, 6679 (1972). 71. Iipscomb, W. S.,Accounts Chcrn. Ih. 6, 257 (1973), and references therein. 72. Lipscomb, IV. N., “Boron Hydrides,” Henjamin, New York, 1963. 73. Lo, F. Y., and Strouse, C. E., persorid communication. 74. Lo, F. Y., Strouse, C. E., Callahan, K. P., Knabler, C. B., and Hawthorne, XI. F., J. Amer. Chent. Soc. 97,428 (1975). 75. Miller, V. R., and Grimes, It. K.,J. Amer. Chent. Soc. 95,2830 (1973). 76. Miller, V. It., Sneddon, L. G., Beer, I). C., and Grimes, R. N., J. Amer. Chem. Soc. 96, 3090 (1974). 77. Muetterties, I?. I,., and Knoth, \\-, TI., “l’olyhedral I3oranes.” Dekker, New York, 1968. 78. llulliken, It. S., J. Anm. ChfJm.Soc. 74, 811 (1952). 79. Parshall, G. W.,Accounts Chcm. ltra. 3, 139 (1970). 80. l’axson, T. I<>., and Hawthorne, Rl. F., .I. Anter. Chent. Soc. 96,4674 (1974). 81. PleBek, J., Stibr, IJ., and HehBnek, S., Syn. Inorg. Metal-Org. Chem. 3, 291 (1973). 82. l’opp, G., and IIawthorne, M. F., Inorg. Chem. 10, 391 (1971). 83. Popp, G., and Hawthorne, M. F., d. Anw. Cheni. Soc. 90, 6553 (1968). 84. Itosenblum, M., “Chemistry of the Iron Group Metallocenes.” \Viley, Nen- York, 1965. 85. Ruhle, €I. IV., and Hawthornc, Ll. F.,Inorg. Chem. 7, 2279 (1968). 86. Salentine, C. G., and Hawthorne, 32.1. F., J. Amer. Chent. Soc. 97, 426 (1975). 87. Salentine, C. G., and Hawthorne, M. F., J. Chem. Soc., Ch.e7n. Commun. 560 (1973). 186 KENNETH P. CALLAHAN AND M. FREDERICK HAWTHORNE 88. Spencer, d. L., Green, &I., and Stone, F. G. .4.,J. Chem. SOC.,Chem. Commun 1178 (1972). 89. St. Clair, D., Zdkin, A,, and Templeton, D. II., Znorg. Chem. 8, 2080 (1969). 90. St. Clair, D., Zalkin, A,, and Templeton, D. H., Inorg. Chem. 10, 2587 (1971). 91. St. Clair, D., Zalkin, A,, and Templeton, D. H., J. Amer. Chem. SOC.92, 1173 (1970). 92. Strouse, C. E., and Hardy, G., personal communication. 93. Todd, 1, J., Arivan. Orgnnometnl. Chtm. 8, 87 (1970). 94. \Varren, L. F., Jr., and Hawthorne, M. F., J. ilnzer. Chem. SOC.90,4823 (1968). 95. \4-arren, L. F., Jr., and Ilawthorne, hf. F., J. Amcr. Chein. Soc. 92, 1157 (1970). 96. FVegner, 1'. A4.,and Haethorne, 11. F., Chem Cornmun. 861 (1966). 97. IVeisboeck, It. A,, and liawtliorne, h1. F., J. Amer. Chem. Soc. 86, 1642 (1964). 98. Wilson, li. J., \Varren, 1, F., ,Jr., and Hawthorne, M. F., J. Amer. Chem. Soc. 91, 758 (1 969). 99. lying, R. M., J. Amer. Chenz. SOC.89, 5599 (1967). 100. Wing, R. M.,J. Amer. Chem. SOC.90, 4828 (1968). 101. N'ing, R. M., personal communication. 102. Young, I). A. T., iViersenia, H. J., and Hawthorne, 31. F., J. Amer. Chem. SCC. 93, 5687 (1971). 103. Zalkin, A4.,Hopkins, T. I<.,and Templeton, D. H., Znorg. Chem. 6, 1911 (1967). 104. Zalkin, A,, Templeton, I). H., and Hopkins, T. I<., J. Amer. Chenz. Soc. 87, 3988 (1965) Recent Advances in Organoantimony Chemistry ROKURO OKAWARA and YOSHIO MATSUMURA* Departmenf of Applied Chemistry Osaka University Yamodakami, Suita, Osaka, Japan I. Introduction . 187 11. Hexacoordinate Mono- and Uiii,gnnoantiinoiiy Compounds . 188 111. Triorganostibine Sulfide 192 A. Ssture of the Stibine-Sulfur Ihnd 193 13. Reactions . 195 I\’. Tertiary Stibines . . 197 A. C1e:tvage Ileactions of the I’lirii~~I~AiitiinonyIbnd 197 13. Properties of Asymnietrir:~lTertiary Stibines and Resolution of Quaternary Stibonium Iodidtt 198 c. Coordination Behavior with Tiatisitioii llet,al CarhCJllylS 200 Referelices . . 202 I INTRODUCTION Organoantimony compounds wre synthesized for the first time by Lowig in 1550, and bctwem 1910 and 1930 many new compounds were prepared, largely in the hop(, of finding pharmacologically active sub- stances similar to those of srsvnic. However, further progress was slow, and orgarioantimony chemistry has bwn one of the backward areas com- pared with the remarkable progrcss since 1960 in other fields of organo- metallic chemistry. In 1970, thcrc. appeared an extensive monograph on organometallic compounds of arscmic, antimony, and bismuth by Doak and Freedman (1)) covering the pvrtincnt literature on organoantimony compounds through the end of 1967. In the past 5 years) there has been increased activity in this area of chemistry. Here we describe some recent advances in the field, emphasizing mainly structural aspects and interesting new reactions. * Present address: Japan Syntlwtic Rubber Co., Ltd., Itesearch Laboratory, 7569 Iknta, Tama, Kawasaki, Japan. 187 188 ROKURO OKAWARA AND YOSHIO MATSUMURA II HEXACOORDINATE MONO- AND DIORGANOANTIMONY COMPOUNDS Pentavalent compounds, R$b, R4SbX, and R3SbX2,are the best-known organoantimony compounds, and the structure of these stable compounds has been the subjwt of several invcstigations (1-4).On the other hand, mono- and dialkylantimony halidrs, RSbX, and R2SbX3,are thermally unstable and decompose cven at room temperature with evolution of RX forming SbX3 and RSbX2, respectively, so that very few studies had been done on these compounds (5,6).However, we have succeeded in stabilizing these halides by substituting an X atom with a chelating acctylacctonate group or by adding an oxygen donor molecule, such as hcxamethylphos- phoric triamide, arid have studied thc structures of resulting hexacoordinate alkylantimony compounds (7-9) togcthcr with those of corresponding arylantimony derivatives (8-11). At almost the same time a research group in Utrecht independently studied hexacoordinate organoantimony compounds including acetylacetonatc (12-14) , oxinate (15), carboxylate (16))and alkoxidc (16) derivatives. One of the most interesting features in these hexacoordinate conipounds is the existence of tn o isomers for dichlorodiaryl (acctylacctonato) antimony in solution, deduccd from the IR and PMIL spectra of (CbH,) 2SbC1, (acac) in solution. The research group in Utrecht (12) initially suggested that two forms existed in clquilibrium in CHCla: a hexacoordinatc trans-dichloro structure with a chelating acetylacrtonatc ligand and a pentacoordinate structure with thc antimony atom bonded to thr y-carbon atom of the acctylacctonate group. Howver, w(1 proposcd (10) that the two isomers in various solvents might be cis- and trans-diphenyl forms. This proposal was confirmcd by detailed studies (11) of dihalodiaryl (acetylacetonato)- antimony in solution. Recently, in agrccwicnt with our conclusion, the Utrccht group reported furthrr studies (1’7) on the trans-cis isomcrization process and the influence> of these two forms on 0-diketonate ligand ex- change reactions in dichlorodiphcnyl (P-dikctonato) antimony. Some new mono- and diorganoantimony chloride derivatives wre ob- tained according to the following reactions.? Acac = acetylacetonate group. 2 Compounds (p-YC6FIl)zSbnrp(acac)(Y = SO1,C1, 13, CHI, CHdI) were prepared from p-YC6H&bO(OEI)z (11, 180) and p-YC6H4ShC13(acsc)(Y = NOZ,If, CHa) were obtained from corresponding stibonic acid (18n, b). IlRlP.4 = hexamethylphosphori~ triamide; DNSO = dimethyl sulfoxide; I’yO = pyritline A-oxide; 4-CH,l’yO = y- picoline N-oxide; T1’1’0 = triplienylpliosphine oxide. Organoantimony Chemistry 189 I sop (It = CH3, CsHs, p-CH3CsFIr; 1, = ~I~~l'A,DiLISO, I'yO, 4-CH3PyO, TPPO) These compounds arc stablc at room temperature in the solid state, but in solution somc monomcthylantimony tetrachloride adducts, CHZSbC14L (L = HMPA, at room temperature; L = PyO or 4-CH3PyO above 70"), decompose into CH3Cl and Sb('13L. All thcse compounds mcntioncd contain a hexacoordinate antimony atom. From the PMR spcctrum of C6HSSbC13(acac)in CHC13, we have suggested a structure I in which thc phenyl and the acetylacetonatc groups are in the same plane (18~).Hcccntly, a similar structure has been estab- lished for the methyl analog, C>H3SbCI3(acac)(19), by X-ray crystal- lography. Configurational isomcrs in cyuilibrium in solution werc also found for the two series of hexacoordinatc antimony compounds, RzSbXz (acac) (It = aryl) and CH3SbC14L (I, = PyO or 4-CH3PyO). The PMR spec- trum of (C6HS)2SbC12(acac)at 22O in CDC13 shows two sets of acetyl- acetonate methyl (A,B) and y-protons (C,D), as shown in Fig. 1. How- ever, the spectrum at -30" of thc solution freshly prepared below -30" in dichloromcthane shows only one methyl resonance (A in Fig. 2). With increasing temperature, a new peak (B in Fig. 2) begins to appear and the intcnsity ratio B/A increases rapidly to the equilibrium value. From these observations, togcthcr with t,hc results of benzene-induced solvent shifts of methyl protons of both tolyl- and acetylacetonate groups in (p-CH3CsH4)z- 190 ROKURO OKAWARA AND YOSHIO MATSUMURA B 6.5 5.5 5.0-' 2.5 2.0 1.5 b(ppm) FIG. 1. Proton magnetic resonance spectrum of (CGH5)2SbC12(acac)in CDCIBat 22". SbClz(acac) , we proposed that isomerixation of the trans-phenyl (11) (peak A) to cis-phenyl (111) (peak B) isomer takes place and that the two isomers are in equilibrium in ~olution.~In the solid state, this compound gave two crystal forms. Both forms show the same PMR spec- tral properties on solution as described in the foregoing. The X-ray struc- ture determination revealed, however, that both have trans-phenyl structures. The only difference is in the dihedral angle of the two phenyl ring planes as shown in Fig. 3 (20, 21). A trans-methyl structure was also indicated by an X-ray structural study for the methyl analog, (CH3)2- SbBrz (acac ) (22). 3 The effect of substituents X and Y on the equilibrium was also studied for (p-YCBHa)zSbX2(acac)(X = F, C1, Br; Y = NO2, C1, CH3, C&O) (11). Organoantimony Chemistry 191 A 22" 4 FIG.2. hlethyl proton resonanws of (C611j)&Clz(acac) in dichloromethane at vari- ous temperatures. (a) The solution \V:E freshly prepared below -30". The figures in parentheses represent the time (in minutes) elapsed from the preparation of the solution to the measurements. (b) The solut,iori was kept at room temperature for 24 hours, and ample time mas allowed for each mwsurement. For monomethylantimony adducts, CH3SbC14L (L = PyO or 4- CH3PyO) in solution, two isomeric forms (VI) and (VII) were also sug- gested to be in equilibrium from the solvent-dependent PMR spectra (9). From the PMR and IR spectra of (CH3) &3bC13L (L = DMSO, HMPA, TPPO, or PyO) in solution, an octahedral geometry with trans-methyl configuration (VIII) was suggested (8). 192 ROKURO OKAWARA AND YOSHIO MATSUMURA FIG.3. The molecular structure of (C61-I~)2ShC1~(acnc).The dihedral angle of two plienyl ring planes: (IV), 84.6" (20);(V), 38" (21). (VI) (VII) (VnI) It is notable that the existence of configurational isomers in solution has been established for these hexacoordinatc organoantimony compounds, since among hexacoordinate organotin compounds, the structures of which have been extensively studied, there are few reports of such isomer^.^ 111 TRIORGANOSTIBINE SULFIDE In contrast to the well-studied triorganophosphine oxide or sulfide, very little work had been done on the analogous organoantimony compounds. Recently two configurational isomers have been isolated from dimethyltin N,N'- bis(salicyla1dehyde)ethylenediiminate (63). Organoantimony Chemistry 193 250 260 270 280 290 300 310 320 330 Wavelength (nm) FIG.4. Ult,raviolet spectra of l<&S and IlaSbX,. Dashed curve, in n-hexanc; solid curve, in acetonitrile. (1) (CH3)3SbS (A,,,, 279 nm, because of its poor solubility, the supernatant solution obtained after shaking for 30 minutes was used); (2) (CII,),SbS (1.10 X mole/liter, A,,, 267 nm, 10g e 3.6); (3) (C6!Il1)3SbS (0.893 X mole/ liter, A,,, 282 nm, log o 3.7); (4) ((:6f111)38bS(0.839 X lo-' mole/liter, Amax 274 nm, log c 3.8); (5) (CH3)aSbCb (1.27 X mole/liter); (6) (CHa)3SbBrz (1.06 X mole/liter); (7) (C6Hl,)rSb13rz(1.01 X 10-4 mole/liter). Zingaro and Cheremos (24) reported that triorganostibine oxide in solu- tion is an equilibrium mixture of monomer and polymer. However, tri- organostibine sulfide is monomoric and is expected to have a semipolar bond like that of R3P0 or R3PS. We investigated the nature of this bond by spectroscopic methods (25, 26). Also, in the hope of finding a particular reactivity of this bond, the reactions with alkyl halides (25) , acid halides (2?'),or some organometallic compounds (28-30) were carried out. A. Nature of the Stibine-Sulfur Bond The UV spectra of trimcthyl- and tricyclohexylstibine sulfides have absorption maxima at ca. 280 rim in n-hexane which give a blue shift in acetonitrile, as shown in Fig. 4. These bands may be assigned to intra- 194 ROKURO OKAWARA AND YOSHIO MATSUMURA Compound Solvent J(I€Z)" fpcb (CH8)aSI) Nrat 130.5 0 (CH,) 3SbS CIICll, 138. <5 0.67 (CII,),Sb+ TI20 139.5 1 (CHd31) Neat 127d 0 (CH,) 31'0 DZO 129" 0.29" (CH,)rP+ D20 134" I a Values are conpidered reliable to kO.5 Hz. b fpc = formal positive rtiarge. c (CH,)rSbI was employed :is :I source of (C€I,),Sh+. Haake et ccl. (;3f 1. This value is comparable with that obtained from a dipole moment measurcnient (0.36) [we I'hillips ef a!. (32)l. molecular charge transfer from the semipolar structure (IX) to the double bond structure (X). c- (Cli,)BSb--S (CH,)Bb=S (IS) (S) This suggests that thc sc.mipolar structure makes a major contribution to the resonance hybrid of the sulfide in its ground state (25) .5 In Table I the l3C-H coupling constants of three methylantimony compounds are shown, togcthcr with those of methyl phosphorus analogs for comparison. From the formal positivc charge, calculated on the as- sumption that the l3C-H coupling constants arp dependent linearly on the amount of positive charge on the central atom, the Sb-S bond in (CH3)SSbS may be regarded as a semipolar bond (IX), although some contribution of thc dr-pr bonding may also be involved (26).6 This is con- sistent with the results obtaincd by the UV spectral investigation. 6 The Sh-S stretching frequency of (CH,),SbS is observed at 433 rrn-l. The esti- mated frequrncies for Sb-S pure single and double bonds are 337 and 483 cm-I, re- spectively (28). 5The P-0 bond in (CII3)3PO has been reported to be well described as a double bond, since the '3C-H couplmg constant of this compound is closer to that of (CH3)3P than to that of (CHI)d'+ (31). Organoantimony Chemistry 195 B. Reactions 1. With Alkyl and Acid Halides Reactions of trimethylstibine sulfide with alkyl halides proceed smoothly under mild conditions to give dialliyl disulfides and organoantimony com- pounds in quantitative yields (226): 2 (CH,),SbS + 2RX -2 [(CH,),Sb-S :]---+ 2 [(CH3),SbYsR] X-R ‘X (RX = CH,I, C,H,I, C,H,CH,I, C,H,CH,Br) The semipolar Sb-S bond may easily allow formation of trimethylanti- mony halide thioalkoxide (XII) via a four-centered intermediate (XI). In the reaction of trimethylstibine sulfide with acid halide, trimethyl- antimony halide thiocarboxylatcs (XIII), corresponding to compound XI1 in Eq. (1) were isolated (a7): (CH3)3SbS + IICOX -(CH,) 3SbX(SCOR) (2) (XIII) (R = CH3, or C6H5; X = C1, or Br) The difference in stability between compounds XI1 and XI11 may be attributed to a less reactive Sb-S bond in XIII. This is consistent with the fact that trimethylantimony bis (thiocarboxylates) is thermally more stable than the bis(thioa1koxidcs) (33, 34). 2. With Some Organometallic Compounds The rcactions of RJSbS (R = CHs, C,Hll) with organotin compounds arc summarized in Table 11. Thc rcsults can be classified into three typcs: recovery of the reactants, coniplcx formation, and sulfur-halogen exchange (68, 35). Complexes RzSnXz-2(CH3)3SbS(R = CH3, CzH5;X = C1, or Br) were obtained only in a few cases. These complexes exist as monomcric species in toluene, but in CHCl, the IIt and PMR spectra indicate that these complexes change into thrw species, (CHs)3SbS, ( CH3)8bX2,and [(CH3)zSnS]3. Since the complexes can be prepared from the mixture of 196 ROKURO OKAWARA AND YOSHIO MATSUMURA TABLE I1 I~EACTIONSOF RzSbS ANT) TIN COMPOVSDSIN hfETIIANOL, ACETONE, OR CHLOROFORM Tin chlorides (CJIll)BbSb (CHd3SbSb 13 x Complex Complex EX 1:,X EX (1 The results obtained with bromide are quite simi- lar to those of chloride. Dash (-) = starting materials are recovered. Ex = exchange reaction. these three in CHCI,, the presence of the following equilibrium has been confirrncd: R~S~XL.~(CH~)3SbS (CH3) 3SbS + (CIT3),SbX, + f (R,SnS) 3 (3) Reactions of trimethylstibine sulfide with alkylindium halides are quite different from those with organotin compounds described above (SO). As shown in the following equation, reaction of R21nX with (CH3)3SbS gives stibonium salts: (CH3) 3SbS + lizInX -R (CII3) 3SbX + RInS (4) (R = C113, C2€15; X = C1, Br, I) Under mild reaction conditions, complexes R21nX. (CH3)i3bS (X = C1, R = CH3, C2HS; X = Br, R = C2H5) were isolated. Alkyl migration from indium to antimony probably occurs via this type of intermediate com- plexes. Triorganostibine sulfide reacts with some organotin compounds con- Organoantimony Chemistry 197 taining a stable tin-tin bond (29) : RsShS + R3’SnSnltl’ -lt3Sb + (1t~’Sn)PS (5) (11 = CH3; It’ = C6H,, (‘61€d(“I12 or 11 = C,H,; R‘ = C,HJ This result is particularly intervst ing since no reaction occurred when such ditins were heated with elenicntal sulfur under similar conditions. The enhanced reactivity of the triorganostihine sulfide in these reactions may bc attributed to the semipolar Sb--8 bond. IV TERTlA RY STlBlNES Various tertiary stibines, 1<&3b,are easily obtained by the reaction of SbC13 with the Grignard reagmt. However, it is difficult to prepare RItz’Sb or RR’1t”Sb by this method, for the materials R2’SbC1or R’II’’SbC1 are not easily obtained. We havc succecdd in preparing these types of com- pounds in fairly good yields via cleavagc of the phcnyl-antimony bond of (C&)sSb or (C6Hs)zRSb (I< = alkyl) with sodium in liquid amrn~nia.~ By this method, some asymmc4rical tertiary stibines (CH,) (C6H5)RSb were prepared (37, 38), and optically active stiboriium conipounds have been obtained for the first timr from these asymmetrical stibines (39). A selective cleavage rcaction of thc phrriyl-antimony bonds of ( C6H5)zRSb with dry hydrogen chloride to givc HSbC12 was also found to be applicable to the preparation of compounds ltRz’Sb (40,41). One of the notable featurw of tertiary stibirics is that they have co- ordinating ability toward transit ion mvtals. Howcver, the coordination behavior of RaSb had been invc.stigatcd only in connection with analogous phosphine or arsine ligands. W(>haw prcparcd some interesting tertiary stibines, such as (ItzSb) 2CH2 (41-43) and RR’ ( CH2=CHCH2) Sb (44), and studied their coordination bchavior (44-47). A. Cleavage Reactions of the Phenyl-Antimony Bond 1. By Sodium in Liquid Anivioiiia As shown in the following rclaction scheme, tertiary stibines R(C&) zSb and [(c&)zSb]z(CH~.). were obtained in good yields by a cleavage re- ’I Recently, Xeinema el al. (36) reported that an alkyl-antimony bond of It3Sb (R = alkyl) IS also cleaved by sodium in liquid ammonia. 198 ROKURO OKAWARA AND YOSHIO MATSUMURA action of one phenyl-antimony bond of (C6H,)3Sb by sodium in liquid ammonia and subsequent reaction with RCl and (CH2) ,C12,8respectively: RC1 R (CeH 5) zSb (R = CH,, CzHa, i-C3H7, CfiHsCHz) (6) (CHz).Cl? [(C,Hs)zSblz(CH*), II = (n 1,3) Selective cleavage of one phenyl-antimony bond of CH$(C&)2Sb was also found to occur,g and, from the sodium-containing intermediate, hitherto unknown'O asymmetrical tertiary stibines, (CH3) (C6H5) R'Sb (R' = C2H5, i-C3H7, C6H5CH2),and [(CH3) (C6H5)Sb]2(CH2). (n = 1, 3,4) can be prepared easily in about yield ($7'). 2. By Dry Hydrogen Chloride It was also found that the phenyl-antimony bonds of (CH3)(C6H5)2Sb were cleaved sclectively by dry hydrogen chloride in chloroform to give CH3SbC12 (40). This cleavage reaction was successfully applied to the preparation of new ligands," bis(diorganostibino) methanes (41), as fol- lows : HCI RMgX [(CeH5)2Sb]zCH2 -(ClZSb)&H, -(RzSb)*CHz (7) in CHCh (R = CH3, CzHs, P-CH~C~H~) B. Properties of Asymmetrical Tertiary Stibines and Resolution of Quaternary Stibonium Iodide As shown in Table 111, the PMR spectra of pheriylmethylbenzylstibine and phenylmethylisopropylstibine in CDCl, at 23" show an AB quartet * Reaction of (C6H5)eSbNawith (CHz)2C12gave the distibine [(CsH&Sb]z (48). In the reaction of R(CfiH5)2Sb(R = C2Hs,i-CzH7, G,€IsCH*),cleavage of the alkyl- antimony bond took place (38). lo Asymmetrical triarylstibines had been prepared (49, 60) by Campbell from the reaction of phenylmagnesium halides with diarylchlorostibirles RR'SbCI, which were obtained by the diazonium reaction. However, the alkyl-substituted analogs were not identified. l1 Compound [(CII,)zSb]zCHzcould be obtained from the reaction of (CHa)zSbNa with CHzC12 in liquid ammonia, but the purification is difficult (49). Organoantimony Chemistry 199 TABLl< 111 PROTON hIACNETIC RESONANCIGDATA OF ASYMMETRICALTERTIARY S'I'IRINES" ~~ Tertiary stibines 6 (ppm) J (Hx) ilssignment 0.90 (s) - SbCH3 1.30 (m) - SbC2HS 0.95 (s) - SbCH3 I .21 (cl) 7.5b 1 C(CH& 1.26 (d) 7.56 J 1.90 (in) - SbCIl C,H,CHz 0.95 (s) - SbCH3 2.96 (ti) 12.0" 7 } SbCHz 3.06 (d) 12.00 J 0.91 (s) - SbCH3 1.45 (d) 11.3" I .50 (d) I .50 (s) 3 0.83 (s) - SbCII3 1.64 (m) - Sb (CH2)3Sb 4 0.84 (s) - SbCH3 1.63 (m) - Sb(CH,),Sb a In CDCL at 23"; 6(ppm) tiownfield from internal TMS. Aromatic protons were observed at about 7.0-7.6 ppm. Multiplicity: (s) singlet; (d) doublet; (m) multiplet. ' J(CII3-CH). J(gem). At 100 MHz. for the benzyl methylene group and two doublets for the isopropyl methyl groups, which are expected for the asymmetrical compounds. In bis (phenyl- methylstibino) methane, a singlet and an AB quartet for the methylene group are observed. These arc reasonably assigned to the racemic (XIV) and the meso (XV) forms, respectively. These observations suggest that the rates of the pyramidal inversion of our asymmetrical tertiary stibines 200 ROKURO OKAWARA AND YOSHIO MATSUMURA meso racemic (xv) are very slow on the PMR time scalc.12This is consistent with a lower limit to the barrier of pyramidal invcrsion for diisopropyl-p-tolylstibine, which was detcrmincd to be 26 kcal/mole from the PMR spectrum (53). The resolution of phosphonium (54) and arsonium salts (1) containing four diffcrcnt organic groups has bcen the subject of considerable investi- gation. We have succeeded in resolving a quaternary stibonium iodide for the first time (39). A racemic mixture of thc quaternary stibonium iodide, (CH,) ( C>2H5)(i-C3H7) ( CsHj)SbI was obtained by the reaction of (CH3)- (i-C3H7) ( C6H5)Sb with ( C2Hj)30BF4 in methylcne chloride, followed by treatment with KI in methanol. Onc cnantiomcr of this stibonium iodide was 0btainc.d by way of Ag-( - ) -dibenzoylhydrogcntartrate (DBHT): (CH,) (C2115)(z-C~II~) (C811r,)ShI Ag--(-)-DBHT - [aID18- 66.5" (c, 1.02 in inc~thanol)3 [a]oZ6 + 4.10" (c, 0.63 in methanol) (8) This cnantiomcr is optically stable both in thc solid and in solution. C. Coordination Behavior with Transition Metal Carbonyls 1. Reaction of Bis(c1iorqanostibino)methane with Metal Carbonyls In the substitution reactions of the metal hcxacarbonyls, M (CO)6 (M = Cr, Mo, W) , with bis (diorganostibino)mcthane, we found that bis (diaryl- stibino) mcthanc acts as a monodentatc ligand and bis (dimcthylstibino) mcthanc as a bridging ligarid to form pcritacarbonyl complexes XVI and XVII (45). Furthermore, doubly bridged tctracarbonyl complexcs (XVIII) were obtained from the rcaction of M(CO)r(dicne) with bis- (diarylstibino) methanc (46).I3 This coordination behavior of bis (diorgano- stibino)methane sccms to be different from that of the phosphinc analog l2 The calculated barrier to the pyramidal inversion for trimethylstihine was reported to he 26.7 (61) and 25 kcal/mole (62). l3 Diene = 1,h-cyclooctadiene (C8H12)or 2,5-norbornadiene (C~HS). Organoantimony Chemistry 20 1 (R = C,H, , P-CH3C,H4 ; M = Cr, Mo) (xvII1) in the sense that bis(diphcnylphosphin0) methane acts as a chelating ligand in Mo(C0)4[ (CsHj)d']L'H2 (55). However, a recent report (56) suggests the possibility that onr bis (diphenylstibino) methane acts as a chelating ligand in new monomrric compounds, Mo (CO) 4[ (C6H5)zSb]zCHz and MO(CO)~([(C6H5)zSb]zC:€Iz]z. Bis (diorganostibino) methane rcacted with dicobalt octacarbonyl, under mild conditions, to give a dinuclcar complex (XIX) containing both bridg- ing carbonyl groups and bridging stibinomethane. The substitution reac- tion of compound XIX with diarylacetylene gave complex XX (47). 00 cc R'C'CR' oc, I I ,co oc, A co ,co-co>< ,co-co' oc I I'CO oc I I \co ,SbCH,Sb, R' $bCHzSb, RI \R \R R R R R (R = CH, , C,H, , C,H, , P-CH,C,H, ; R' = C,H, , p- CH,C,&) Recently, Hartwcll (57) reportled a rhodium complex, { [ ( C6H6)zSb]z- CHzRhCOCl ) 2, in which stibinomethane also acts as a bridging ligand. 202 ROKURO OKAWARA AND YOSHIO MATSUMURA 2. Reaction of Tertiary Allylstibirzes with Cp(CO) 2FeCl In the reaction of tertiary allylstibiries with Cp (CO) zFeC1,14a cleavage of one stibine-ally1 bond took place and yellow unstable compounds (XXI) were obtained (44): Cp(C0)2FeC1 + (CI1~=ClICII,)ItR’Sh - { [Cp(CO) LFe]?SbRR’}Cl + CIIi=CIICHdJ + trans-C€I,CH=CIICl (9) (XXI) Compounds XXI were charactcrixed as stable tetraphenylborate salts, which are thought to be a type of stibonium that contains Fe-Sb a-bonds (XXII) . Similar compounds containing halogen groups bonded to the antimony atom were prepared recently by the reaction of [Cp (CO)zFe]z or Cp (CO)zFeNa with SbX, (58,59). (R = R’ = CH, , C,H, , CH,=CHCH, ; R = CH,=CHCH,, R’= CH,, C,H,) Reaction of (CH2=CHCH2) (CGHB)2E(I3 = P, As) with Cp(C0)2FeC1 gave the product [Cp (CO) ( CH2=CHCH2) (C6H5)JC1, and, even in the presence of excess Cp (CO) 2FeC1, cleavage of the E-ally1 bond was not observed. The difference in behavior of these allyl-substituted Group V elements toward Cp (CO) 2FeCl may be related to the strength of the bond from these elements to carbon. REFERENCES 1. Doak, G. O., and Freedman, L. D., “Organometallic Compounds of Arsenic, Anti- mony and Bismuth.” Wiley, Ncw York, 1970. 2. Shindo, M., and Okawara, R., J. Orgunometa1. Chem. 5, 537 (1966). 3. Shindo, hf., and Okawartt, R., Inorg. Kuc1. Chem. Lett. 5, 77 (1969). 4. hfatsumura, Y., Shindo, hf., and Okawara, It., J. Organometul. Chem. 27, 357 (1 971). 5. Morgan, G. T., and Davies, C. R., Proc. Roy. Soc., Ser. A 110, 523 (1926). 6. Doak, G. O., and Long, G. G., Trans. N.Y. Acud. Sci. 28,402 (1966). 7. Nishii, N., Shindo, M., Matsumura, Y., and Okawara, R., Inorg. Nucl. Chem. Letl. 5, 529 (1969). ‘4 Cp = r-CsHs. Organoantimony Chemistry 203 8. Kishii, N., Matsumura, Y., antl Okawara, R., J. Organonzetal. Chem. 30, 59 (1971). 9. Nishii, N., Hashimoto, K., and Okawara, It., J. Organometal. Chern. 55, 133 (1973). 10. Nishii, N., Matsumura, Y., and Okawara, It., Znorg. Niccl. Chem. Lett. 5, 703 (1969). 11. Sishii, N., and Okawara, R., J. Orgcrnort~etul.Chem. 38, 335 (1972). 12. hleinema, H. A,, and Noltes, J. G., J. Orgniiometal. Chem. 16, 257 (1969). 13. Meinema, H. il., and Noltes, J. G,,J. Organonzetal. Chen~.37, C31 (1972). 14. lleinema, H. A,, and fioltes, ,J. G., J. Organometal. Chem. 37, 285 (1972). 15. Meinema, 1%. A., Rivarola, I;., itrid Noltes, J. G., J. Organonzetal. Chern. 17, 71 (1969). 16. Meinema, H. A,, and Noltes, .J. G., J. Orgunometnl. Chem. 36, 313 (1972). 17. Meinema, H. A,, and Noltes, J. G,,J. Organonietal. Chenz. 55, C77 (1973). 18a. Kawasaki, Y., and Okawara, It., Hirll. Chrm. SOC.Jap. 40, 428 (1967). 18b. Kawasaki, Y., Ito, T., and Okawrnra, It., I7it. Symp. Deconzpositioti Organomekd. Cotnp. Refractory Ceranzics, ,IIetul utid Illrial Allo?js, 1967, p. 47. 19. Kanehisa, N., Kai, Y., and %mi, N., Iriorcq. Nuel. Chem. Lett. 8, 375 (1972). 20. Onuma, K., Kai, Y., and Iiasai, N., Inorg. AV~~cl.Chem. Lett. 8, 143 (1972). 21. Kroon, J., Hulscher, J. 13., antl I’cerdeman, A. F., J. Organometal. Chem. 37, 297 (1972). 22. Uda, S., Kai, Y., Yasuoka, N., tttiti Kasai, N., Cryst. Strurt. Comrrian. 3, 257 (1974). 23. Kawakami, K., Miyaurhi, M,, :tntf Tatxtko, T., J. Organomrtal. Chem. 70, 67 (1974). 24. Cheremos, G. N., and Zingaro, It. A,, J. Organonwfal. Chem. 22, 637 (1970). 25. Otera, J., and Okawara, R., J. Or~ganonzetrcl.Chem. 16, 335 (1969). 26. Otera, J., and Okawara, It., lwrg.Nitcd. Chem. Lett. 6, 855 (1970). 27. Otera, ,I., and Okawara, It., J. Orguuometul. Chem. 17,353 (1969). 28. Shindo, M.,Matsumura, Y.,atid Okawara, It., J. Organomelul. Chrm. 11,299 (1968). 29. Otera, J., Kadowaki, T., and Okawara, It., J. Organometal. Chenz. 19, 213 (1969). 30. Maeda, T., Yoshida, G., and Okawara, It., b. Organometd Chem. 44, 237 (1972). 31. Haake, l’., Miller, W. B., and Tyssee, 11. A., J. Amer. Chem. Soc. 86, 3577 (1964). 32. Phillips, G. M., Hunter, J. S., antl Sutton, 1,. 15., J. Chem. Soc., 146 (1945). 33. hlatsumura, Y., Shindo, M., and Okawara, It., Znorg. Nuel. Chem. Lett. 3,219 (1967). 34. Schmidtbauer, €I.,arid Mitschkr, I<. H., Chenz. Ber. 104, 1842 (1971). 35. Shindo, hl., hlatsumura, Y., and Okawara, It., Rztll. Chem. SOC.Jap. 42, 265 (1969). 36. Meinema, H. A4.,hlart,ens, 11. F., and Noltes, J. G., J. Organometal. (’hem. 51, 223 (1973). 37. Sato, S., llatsumura, Y., and Okawara, It., J. Organometal. Chem. 43, 333 (1972). 38. Sato, S., Matsumura, T., and Okawxra, It., Inorg. Nucl. Chem. Left. 8, 837 (1972). 39. Sato, S., Matsumura, Y., and Ok:iwnla, It., J. Organometal. Chern. 60, C9 (1973). 40. Matsumura, Y., and Okawara, R., unpublished work. 41. Matsumura, Y., and Okawara, lt., Inorg. Sucl. (%em. Lett. 7, 113 (1971). 42. hlatsumura, Y., and Okawara, I<,,Znorg. Nud. Chem. Lett. 5, 449 (1969). 43. Matsumura, Y., and Okawara, R., J. Organometul. Chem. 25, 439 (1970). 44. Matsumura, Y., Harakawa, M., and Okawara, It., J. Organometal. Chem. 71, 403 (1974). 45. Fukumoto, T., Matsumura, Y., and Okawara, It., J. Organotrietal. Chem. 37, 113 (1972). 46. Fukumoto, T., Matsumura, Y., and Okawara, It., Znorg. Nucl. Chem. Lett. 9, 711 (1973). 47. Fukumoto, T., hlatsumura, Y., arid Okawara, It., J. Organometal. Chem. 69, 437 (1974). 48 Hewertson, W., and \Vatson, €1. R., 1. Chem. Soc., 1490 (1962). 204 ROKURO OKAWARA AND YOSHIO MATSUMURA 49. Campbell, I. G. M., J. Chem. SOC.,3116 (1955). 50. Campbell, I. G. M., and White, A. W., J. Chem. SOC.,1184 (1958). 51. Weston, R. E., Jr., J. Amer. Chem. SOC.76, 2645 (1954). 52. Keopple, G. W., Sagatys, D. S., Krishnamurthy, G. S., and Miller, S. I., J. Amer. Chem. SOC.89, 3396 (1967). 53. Jacobus, J., J. Chem. SOC.D, 1058 (1971). 54. Gallagher, M. J., and Jenkins, I. D., in “Topics in Stereo-Chemistry” (E. L. Eliel and N. L. Allinger, eds.), Vol. 3, p. 1. Wiley (Interscience), New York, 1968. 55. Cheung, K. K., Lai, T. F., and Mok, K. S., J. Chem. SOC.A, 1644 (1971). 56. Beall, T. W., and Houk, L. W., J. Organomelal. Chem. 56,261 (1973). 57. Garrou, P. E., and Hartwell, G. E., Private communication. 58. Trinh-toan and Dahl, L. F., J. Amer. Chem. SOC.93, 2654 (1971). 59. Cullen, W. R., Patmore, D. J., and Sams, J. R., Znorg. Chem. 12,867 (1973). Pentaalkyls and Alkylidene Trialkyls of the Group V Elements HUBERT SCHMIDBAUR Anorganisch-chemisches Loboratorium Technische Universitat Miinchen Munich, West Germany I. Introduction . . 205 11. Simple Nitrogen Ylides . . 207 111. Phosphorus Ylides and Pentaalkylphosphoranes . . 209 A. Alkylidene Trialkylphosphoranes . . 209 B. Pentaalkylphosphoranes . . . 214 IV. Arsenic Ylides and Pentaalkylarsoranes . . . 224 A. Alkylidene Trialkylarsoranes . . 224 B. Silylated Ylides of Phosphorus and Arsenic . . 228 C. Pentamethylarsorane . . 229 V. Antimony Ylides and Pentaalkylstiboranes . . 231 A. Antimony Ylides . . 231 B. Pentaalkylstiboranes . . 232 VI. BismuthCompounds . . . 236 VII. Related Compounds of Vanadium, Niobium, and Tantalum . 236 A. Vanadium Ylides? . . . 236 B. Niobium Compounds . . 237 C. Tantalum Pentaalkyls and Tantalum Ylides . . 238 D. Metal Ylides or Metal Carbene Complexes? . . 239 References . . . 240 I INTRODUCTION With the exception of nitrogen, all of the Group Vb elements are ex- pected to form pentacoordinate compounds in their 5+ oxidation state, and this is, indeed, the case with some of the halides, alkoxides, etc. It was not until the pioneering work of Georg Wittig and his collaborators, however, that the first examples of pentaorganyls of these borderline elements between metals and metalloids were reported (95-99, 102, 104). In this early investigation, a complete set of the pentaphenyls could be obtained and characterized (95-99, lo??),but apart from the pentamethyl- antimony case, all attempts for the preparation of pentaalkyl derivatives failed (104). 205 206 HUBERT SCHMIDBAUR Wittig’s work was later confirmed and extended by Russian, Dutch, and Japanese groups, who were able to prepare other pentaalkylantimony com- pounds using his method, which consistcd of the reaction of trialkyl (or tetraalkyl) antimony halides with alkyllithium reagents. Table IV lists some of the important examples. When this method is applied to compounds of phosphorus and arsenic, the components undergo an entirely different reaction, and alkylidene trialkyl derivatives of these elements (ylides) are obtained (102, 104). Thus, instead of the introduction of an additional alkyl substituent at the central element, as obscrved with antimony, a deprotonation at one of the a-carbon atoms adjacent to phosphorus and arsenic occurs: LICH~ LiCHa f [(CH3),.\s]C1--\ -LiCl Although the reaction path of Eq. (23,) is typical for phosphorus, and pentaalkylation is characteristic for antimony [Eq. (3) 1, arsenic has recently been found to be the only element to supply both the alkylidene- trialkyl (73) and the pentaalkyl (37) derivative, depending very critically on the reaction conditions [Eq. (2b)l. The revived interest in this area led to new efforts for improved methods of synthesis for pure alkylidene trialkyls (ylides), for the detection and possible characterization of a pentaalkylphosphorane species, and for a better understanding of the state of structure and bonding in ylides and pentaalkyls. This recent work is summarized here with the main focus directed to some of the very simple compounds and their silicon deriva- tives, which have turned out to be of special significance. These derivatives have not only provided new synthetic pathways but have shown improved thermal stability (as in the case of arsenic ylides) and a modified pattern of chemical reactivity. The donor properties of ylides (55, 24), and most of their synthetic applications (103),have been covered in other reviews and articles (3, 26) and are not duplicated here. The general organometallic chemistry of arsenic, antimony, and bismuth is the subject of the invaluable monograph by Doak and Freedman (11). The broad scope of phosphorus ylide and pentaorganophosphorane chem- istry was covered in the leading multivolume series on organophosphorus chemistry edited by Kosolapoff and hlaier (3, 21). Finally, the recent Group V Pentaalkyls and Alkylidene Trialkyls 207 success of efforts to synthesize pentaalkyls [Nb(CH,) 5 (SS)] and cven ylides ([ ( CH3),CCH2l3TaCHC (CH,) (@a) ) of the subgroup metals va- nadium, niobium, and tantalum add a new dimension to the scopc of this article. II SIMPLE NITROGEN YLIDES In an attempt to demonstrate the existence of pentavalent nitrogen, Schlenk and Holtz studied the rtwtion of triphenylmethyl sodium with tetramethylammonium chloride (52). The highly colored material was strongly conducting in polar solvents and could be identified as a salt, the stability of which is due to the resonance stabilization of the triphenyl- methide anion. In the absence of such stabilizing substituent effects (53), as with n-butyl or another alkyllithium reagent, a metalation of the tetra- methylammonium cation occurs, which leads to type I products (18) : --RH 63 (CH3)aPBre + Illi -(CH3)3N-CH2LiBr- (5) (1) These species can be formulated as alkali halide complexes of the nitrogen eo ylide (CH3)3N-CH2, and inany of their chemical reactions have been interpreted according to this structure (9, 15, 40, 91-93, 100-102, 105). Thc addition compounds (I) arc insoluble in diethyl ether, and the slurries obtained are quite stablc. In more strongly solvating media, such as tetrahydrofuran or dimethoxyethane, the compounds are soluble but show rapid decomposition, with trimethylaniine and polymethylene as the main products. These experiments indicate (9, 40) that when the lithium salt is trapped by donor solvent molecules, the free ylide quickly undergoes decomposition (40). No free trialkylammonium ylide has yet been prepared, even under very mild conditions (35).On the other hand, it has been shown, that the tctr:imcthylammonium cation can even be metalated twice by organolithiuni reagents (102) to afford dimethyl- ammonium bismethylidcs : ‘?€I3 CH,--I,i PC6fI6L1 \e/ (CH,)J@Bre - N I3r@ -2C6H6 /\ CII3 CH2-Li (11) 208 HUBERT SCHMIDBAUR Chemical reactions of these materials have also been carefully investi- gated (101, 102), and their properties closely resemble those of other typical carbon nucleophiles, as those of the monometalated analogs do. From this work it is apparent that these ammonium ylides, which are only stable in the presence of lithium salts, are similar in their chararter- istics to other organolithium reagents and have very little in common with other ylides (26).The nitrogen plays no role beyond stabilizing the neighboring carbanion, largely through its inductive effect. As will be mentioned again later, the introduction of trialkylsilyl groups as substituents on the carbanion of an ylide generally leads to significant stabilization. Inherently unstable ylides, e.g., of arsenic, could thus be converted into strikingly stable derivatives through silylation (see Section IV,B). However, similar efforts have not been successful with nitrogen ylides. The landmark experiments by Miller (35) have shown that no ylide (111) can be isolated as the product of a reaction according to Eq. (7).Instead decomposition products (111’), Eq. (8),are observed (Scheme 1).As a consequence the desilylation process (74),normally an important method for generating free ylides, is not applicable with nitrogen ylides. The situation is somewhat different in the case of pyridinium ylides CsH6N@-CH2e, but even compounds of this type are not very stable thermodynamically in the absence of substituents offering some resonance participation to the carbanion or to the onium group (26). CH, Scheme 1 As already mentioned (105), no pentavalent nitrogen compound NR5 has ever been reported. In view of the present knowledge of the valence Group V Pentaalkyb and Alkylidene Trialkyls 209 properties of second-row elements generally, and of CX, and related isosters in particular, this is riot surprising. 111 PHOSPHORUS YLIDES AND PENTAALKYLPHOSPHORANES Alkylidene triorganophosphoranes (phosphorus ylides) are known with a large variety of substituents both at the phosphonium center and at the carbanionic function : Although it is not the purposc of this chapter to review the chemistry of this important class of compounds, it is necessary to recall the main methods of synthesis and the characteristic properties of these species in order to point out some relationships between these and the more elusive penta- alkylphosphoranes. A. Alkylidene Trialkylphosphoranes 1. Preparalion Virtually all of the important methods (3) of synthesis for phosphorus ylides are based on the reaction of a phosphonium salt with a strong base: it4 R4 @/ base / (R1R2R3)1’-CH Xe -(RlRZR3)1’=C (10) \ -1IX \ R5 115 In the presence of electron-withdrawing substituents ( R1-R5) , even rela- tively weak deprotonating agents may be used. If however R1-R3, and in particular R4-R5, are alkyl groups, only the strongest bases are able to produce the ylides (102). Typically, alkali alkyls or ainides are used. With these reagents an alkali halide is, thus, formed, and alkane or am- monia is evolved. The presence of the alkali salt in the reaction mixture was a major problem in these syntheses because most of the ylides were found to form 210 HUBERT SCHMIDBAUR very stable coordination compounds with the smaller alkali cations. Two procedures are now available to avoid this difficulty. The first one was derived from the observation (34, 72) that silyl substituents preclude the formation of addition compounds with the salt through steric and electronic effects. Therefore salt-free silylated ylidts ran be readily obtained, purified, and finally desilylated by alcohols (74) [Eqs. (lla,b)]. The sccond method is based on the fact that sodium chloride does not form ylide complexes in tetrahydrofuran (28),in which sodium amide reacts suffi- ciently smoothly with phosphonium salts [Eq. (la)]: @ + LiR [(CH~)~P-CHLSI(CI-IS)J]C~~ * (CHS)3P=CH-Si(CH3)3 (llft) -LiCI. -RH With bulky substituents attached to the phosphorus, none of these special measures are necessary because complex formation is precluded through steric effects (74).Therefore ethylidene triethylphosphorane may be obtained in high yield without complications from the tetraethyl- phosphonium salt and an alkyllithium reagent (7‘4): -CH4, -LiCl r-* r-* (C2€Ij) ,P=CHCH (13a) (CzH,)d‘CH, (13h) - LiCl Although the reaction conditions have been greatly modified, no evidence whatsoever has been found for the generation of pentaalkylphosphorane according to Eq. (13b). This is also true for other reactions of noncyclic alkylphosphonium salts with lithium alkyls. 2. Properties, Structure, und Bonding Alkylidene trialkylphosphoranes RaP=CRz’ (R = alkyl, R’ = hy- drogen, alkyl) are colorless mobile liquids that can be crystallized at low temperatures in most cases and distilled under reduced pressure (74, 28) (Table I lists melting and boiling points.). The compounds are extremely reactive, spontaneously igniting when finely divided on paper, and vigor- ously fuming when exposed to air (74). Miscibility with and solubility in most aprotic organic solvents is excellent and the materials dissolve as monomers. The reaction with alcohols (57), which leads to tetraalkyl- alkoxyphosphoranes and related secondary products (70, 71), is highly exothermic; even in cold water hydrolysis occurs explosively. Group V Pentaalkyls and Alkylidene Trialkyls 21 1 TABLE I PHYSICALCONSTANTS OF SOMESIMPLE ALKYLIDENE TRIALKYLPHOSPHORANES hp mp Compound ("C/mm IIg) ("C) Ref. 122/760 13-14 28, 74 143-145/760 (-16)-( - 14) 74 60-62/12 ( -45)- ( - 43) 74 26/0.2 (-17)-(-15) 28, 74 80-83/12 25/0.2 28 38/0.001 28 58/0.001 - 28, 6C - 128 28 2S/O. 01 (-38)-( -36) 28, 74 86-87/12 (CzHS)*(n-C3H7)P=CH-CH3 30/0.001 28 (i-C 3H 7) 3P=C H-CH 3 45/0.001 28 (CzHs)aP=CHCzHs 30/0.001 28 (~-C~II~)~P=CHCZHB fi0/0.001 28 (n-C4H9) aP=CHCzHs 80/0.001 28 (CZH~)~P=CHC~H? 33/0.001 28 CzHS (n-C4Hg)zP=CHC3H7 65/0.001 28 (CH3)3P=CH--CH=CHz 62 Crystal structure determinations have not been carried out for a simple peralkylated phosphorus ylide, but a whole series of arylated species and more complicated systems have been carefully investigated (2, 6-8, 32, 84,85, 94). The results of these studies can be summarized as follows: 1. Phosphorus atoms in ylides largely retain the tetrahedral ligand geometry of the phosphoniurn cation precursor, but the bond to the ylene-ylide carbon atom is shortened, indicating an increase in bond order. 2. The ylide-ylene carbon atom is found to be the center of a trigonal- planar arrangement of ligands. This planar configuration may be part of a P system, thus giving rise to restriction in bond rotation and to the existence of geometrical isomers. Infrared and Raman spectroscopy has been applied primarily to provide additional information on thc nature of the ylidic function. An evaluation of the data obtained through a normal coordinate analysis gave a bond order of 1.65 in (CH,),PCH2 (51) and of 1.3 in (CeH,),PCH, (30). This result complements the bond shortenings observed by X-ray analyses. 212 HUBERT SCHMIDBAUR Dipole moments of ylides, although measured only with some of the more complicated species (31), indicate a high degree of charge separation in the ylidic function, ranging between 5 and 9 D. These results strongly favor the zwitterionic (ylide) over the double-bond (ylene) formula. Nuclear magnetic resonance data, which are now essentially complete, also clearly support this description of the bonding in alkylidene phos- phoranes. Early ‘H NNR results had already shown a very considerable shielding of hydrogens attached to the ylidic function, and 31P NRIR values pointed to a quasi-onium-type character of the phosphorus atoms (22,28, 58, 75,80).Whereas all the chemical shifts and coupling constants of the alkylidene groups differed completely from those of the alkyl ligands attached to phosphorus, the shielding of the phosphorus and its coupling to the alkyl hydrogens were similar to that observed for related phos- phonium salts. The more recently reported I3C NRlR spectra (22, 58, 80) have finally shown that the ylidic carbon itself is very strongly shielded and that its couplings with the phosphorus and hydrogen atoms indicate an sp2 state of hybridization. On the other hand, the values for the alkyl carbon atoms and for the phosphorus exhibited the expected LLnormal” parameters for a tetrahedral sp3 configuration. Some of these results are summarized in Table 11. 3. Theory and Photoelectron Spectra The set of physical data compiled so far for the alkylidcne trialkyl- phosphoranes seems to allow us to draw a consistent picture of the structure and bonding in these molecules. A valence bond description thus still has to contain the classic Wittig formulas, but it is clear that the ylide form is the most significant for the ground-state phenomena of these molecules. Semiempirical, extended Huckel molecular orbital calculations on the hypothetical H3P=CH2 molecule with and without the inclusion of d orbitals have shown, according to this model (25), that both the charge distribution and the P=C bond order arc affected significantly by d-orbital participation. The extended Huckel method employed in this LCAO study still gave a high negative charge at the ylidic carbon even when d orbitals were included ! The extended Huckel calculations also indicate a preference for a planar geometry at the ylidic carbon, but it is not obvious whether this result is sufficiently reliable because other geometries have been wrongly predicted by this method. As (CH3)3P=CH2 is still too formidable a problem for theoretical chemists, the related but hypothetical HjP=CH2 molecule was the subject of the ab znitio LCAO-MO-SCF studies published in 1972 (1).Using two uncontracted Gaussian basis sets, one with and the other without phos- TABLE I1 l3C NUCLE.4R MAGNETICRESONANCE SPECTRA OF THREESIMPLE ALKYLIDENEPHOSPHORANES" \ / C= < Compound n (CHs)aP=CH* - 2.3 90.5 149.0 - 18.9 56.0 127 3IT P (CzH,),P=CHz -14.2 86.8 147.5c 6.6 2.9 129.5 20.3 55.9 125 -k (CpHs)aP=CHCH3 - 7.9 113.2 -d 6.0 - -d 27.6 54.4 -d rh (CJM 5 9.5 - - b --I (CZH4) 1. kP a From Schmidbauret al. (80). r ur b Value of 6 is measured relative to CGDC, and converted relative to TMS. Broad;=t3 Ha. Not observed. Off-resonance:dd. 214 HUBERT SCHMIDBAUR phorus d orbitals, two conformations of this molecule have been investi- gated. An essentially zero barrier to C-P bond rotation was calculated in either basis set. Allowing d orbitals to participate, the energy change was such that the picture is consistent with n feedback in the ylidic func- tion. Again, however, a high negativc atomic charge at the ylidic carbon was obtained. This quantity “charge on the atom” is sensitive to the choice of basis set and is to be taken as a formal rather than as a physical property. Inspection of the various orbital sets shows that the ylide has s-type orbitals even when d character is disallowed, but when d character is included the bonding in these molecular orbitals taken as a group is particularly enhanced. Photoelectron studies of (CH3) 3P=CH2 have provided data on the energy of the highest occupied molecular orbital and some of the lower- lying states (49).The frontier orbital energy of 6.81 eV is very lorn and seems to illustrate the carbanionic nature of the carbon, where this orbital is largely localized. This is borne out by semiquantitative CSDO/2 calcu- lations on this molecule. The orbital sequence obtained is satisfactory according to ab initio results. B. Pentaalkylphosphoranes It has already been pointed out in the Introduction that no simple PR6 compound, in which all R’s were independent alkyl groups, has been re- ported to date. However, mixed species, e.g., Hellwinkels CHJ?( C&) 4, have been synthesized (19) as has the perarylated analog P (C&,) 6 (102). These partly or fully arylated compounds are very stable thermally, and their properties and reactions have been studied (21). From these investi- gations it seems that further attempts to prepare a pentaalkylphosphorane might be promising. The successful synthesis of pentamethylarsenic ($7) and of some derivatives of pentamethylphosphorane, such as (C&) 4POCH3 (71),provided the necessary stimulus, but the ultimate goal has not yet been reached.* In the meantime, new concepts for the synthesis of cyclic pentaalkyl- phosphoranes have been developed (88),which have proved the existence of a phosphorus atom bearing 5 aliphatic carbon atoms, 2 of which are * Note added in proof: In May 1975 The and Cave11 have reported (86%)the successful synthesis of dimethyltris(trifluoromethy1) and trimethylbis(trifluoromethy1) phospho- rane from the reaction of tetramethyllead with suitable fluorophosphorane precursors. The compounds are stable white solids of trigonal bipyramidal structure. Group V Pentaalkyls and Alkylidene Trialkyls 215 part of a ring system. In the pioneering work by IZatz and Turnblom (27, 87),the concept of ring strain was employed and ingeniously applied to this problem. These authors obtained the first stable pentaalkylphos- phorane. The investigations of thc group of the present author led to cyclic penta- alkylphosphoranes through inschon of ylides into the strained silacyclo- butane system (78), but the initial products of this reaction could not be isolated. Their transitory existcncc has been demonstrated, however, by a careful study of the reaction mechanism and the products of thc insertion and rearrangement processes. 1. Polycyclic Pentaalkylphosplioraiie It was assumed (80) that there wre two possible reasons for the for- mation of a pentaalkylphosphorane: either there must be a peculiar insta- bility preventing formation of the ylide, or there must be a peculiar stability enhancing formation of the pentavalent phosphorus. Inspection of certain bridged model phosphonium salts has shown that a bridged ring system might, indeed, enhance formation of the pentavalent phosphorane by straining the internal angle at phosphorus in the phosphonium salt. This effect is analogous to that suggested to explain the accelerated rate of hydrolysis of small-ring phosphate esters (90). The following reaction scheme summarizes a series of reactions, in which various phenyl, phenyl/ methyl, and fully methyl-substituted homocubylphosphoranes are gener- ated (27,87, 88) : 216 HUBERT SCHMIDBAUR In all of these examples the phosphorus atom is a member of two doubly fused five-membered rings in the phosphonium salt. This configuration gives rise to considerable strain because the five-membered rings are themselves part of the honiocubyl system. This strain would not be changed if one of the exo-methyl groups was converted into an ylidic function as this basically does not alter the geometry at the phosphorus center (88). However, the strain is relieved if the coordination number at phosphorus is increased to five, because the trigonal-bipyramidal and square-pyramidal polyhedra can accommodate a whole range of valence angles between 120" and 90". Previous work by Hellwinkel had already made use of this and related concepts in his synthesis of the methyl bisbiphenylenephosphorane (19) : The structure of pentaalkylphosphorane has been proved by a complete set of spectroscopic and analytical data, including 'H, 13C, and 31PNMR and IR and mass spectroscopy (88). Among these the chemical shift of the phosphorus atom is particularly diagnostic as it is very close to the values reported for authentically pentacoordinate species. The IH-13C, 1H-31P, and 13C-31P coupling constants arc also characteristic of methyl groups attached to pentacoordinate phosphorus. Therefore there is no doubt that compound IV "is a reality." Its existence has encouraged efforts to synthesize related compounds including pentamethylphos- phorane. Furthermore, the general properties of compound IV indicate that pentaalkylphosphoranes, once prepared, can be quite stable and easy to handle. The compound is a colorless liquid (bp. 20°C at lop6 mmHg) that fragments thermally to trimethylphosphine and cyclooctatetraene at 75°C : Group V Pentaalkyls and Alkylidene Trialkyls 217 After 10s hours at this temperature, only 50% of the material was decom- posed. The thermal decomposition of the arylated species is much faster and, hence, the pentaalkylphosphorane seems to exhibit the most satis- factory properties within this series [Eqs. (14)-(16)]. This result is in agreement with the striking properties of pentamethylarsenic (37). 2. Evidence for Existence of Pentaalkylphosphorane Intermediates The failure of attempts to generate a PIL species via routes according to [R,P]X + RT,i -kL, LiX + PR, (18) i.e., through alkylation of a phosphonium salt,' led to reconsideration of various other methods of synthesis. A promising alternative was found in the insertion of alkylidene phos- phoranes into small carbocyclic systems, such as cyclopropane or -butane : Both of these reactions were expected to profit from relief of strain in the starting material and from a strain-free heterocycle in the product. Owing to the polar nature of the ylidc it was desirable, however, to introduce at least 1 heteroatom into the cyclobutane ring in order to facilitate heterolytic cleavage of the system. The silacyclobutanes seemed to be an excellent choice, and, consequently, a project on ylide cleavage reaction of mono- and disilacyclobutanes was initiated. Although no stable cyclic pentaalkyl- phosphorane was obtained, it was possible to confirm the appearance of Ylidea are obtained instead [see Eq. (l)]. 218 HUBERT SCHMIDBAUR these species as primary products in these reactions : si (No silacyclopropane is available for similar studies at the present time?) a. Ylicle Cleavage of 1 ,3-Disilacyclobutanes (78). 1 ,1 ,S ,3-Tetramethyl- 1, 3-disilacyclobutane is readily cleaved by trimethyl and triethyl methyl- enephosphorane to yield silylated ylides of type V (a-c) (Scheme 2). The 7 I L (BJ 74 7% R,P=CH-Si-CH,- Si-CH, I I CH, CH, R: CH, C,H, i-C,H, (V) (V): (a) (b) (c) Scheme 2 reaction with (i-C3H7)3PCH2is slower due to a steric effect, but an ana- logous product is obtained. It is not known at this stage if this reaction actually involves an insertion product of the pentaalkylphosphorane form (BI) because compound V may be formed either via (A1) alone or via both (A1) and (Cl). If, however, (B1) is formed as an intermediate, it obviously undergoes facile ring opening, whereby silicon-stabilized car- 2 Seyferth et al. (29) have been able to synthesize a first member of this series, but its set of substituents was unfortunately not suitable for our purposes. Group V Pentaalkyls and Alkylidene Trialkyls 219 banioris (A,) and (C,) can be generated with equal probability. These are then isomerized through proton migration, resulting in the silylated ylides (60). Similarly, no decision as to the existence of a pentaalkyl- phosphorane intermediate could be made in the corresponding reaction with 1,3-dimethyl-1, 3-disilacyclobutane (Scheme 3). The final product, r t L HI 7% R,P=CH-SiL CH,-Si- CH, I I - CH, H R: CH, C,H, i-C,H, (VI) (VI): (a) (b) (c) Scheme 3 isolated in quantitative yield, again consisted of a silylated ylide only, and there was no evidence for n PlZs species. b. Ylicle Cleavage of a Il.lonosalacyclobutanes (78). The situation was much more favorable in the inonosilacyclobutanes, where possible inter- mediates no longer have any symmetry with respect to the distribution of the heteroatom. Thus 1, 1-dimethylsilacyclobutane, although also yield- ing only an ylide (VII) as thc sole product of the reaction with (CH3) 3- PCH2, makes it possible to demonstrate clearly the transitory existence of a cyclic pentaalkylphosphorane B3 (Scheme 4). The structure of the material obtained can only be explained on the basis of the equilibria (A3) (13,) 6 (C,) because there is obviously no mechanism for a direct conversion of AS into compound VII. The B3 differs significantly from the intermediates Bz and B1 as in this case it is only carbon atom 5 that may form a silicon-stabilized carbanion, whereas carbon atom 4 lacks this possibility. The pentaalkylphosphorane ring B3 will, therefore, open selectively to Ca, avoiding the forniation of the “hot” carbanion A3. 220 HUBERT SCHMIDBAUR The observation that the proton migration of reaction C3 4 (VII) affects one of the methyl hydrogens is in agreement with previous findings (74).An inspection of formula VII shows that one of the methyl carbons attached to silicon in the product is supposed to originate from the alkylene group of the ~lide.~In order to provide further support for this conclusion, which was based on mechanistic grounds only, it was desirable to introduce different substituents both at the ylide and at silicon, as well as to locate their position in the final product. The combinations of the simple dihydro- silacyclobutane and various methyl/ethyl ylides were considered to be an attractive selection of compounds for this purposc. With these components the reactions were complicated by an unexpected cyclisation process, which tended to obscure the course of the reactions. However, unambiguous conclusions were reached, offering convincing proof for the above mecha- nism. c. Cleavage with Successive Recyclization. Monosilacyclobutane reacts with (CH3)3PCH2 even at temperatures as low as -6OOC. An NMR study of the reaction mixture shows the appearance of a product analogous to compound VI, but this compound undergoes a further reaction with 3 Similarly, it was assumed that the ylidic carbon and the carbon of one of the methyl groups of the terminal silicon atoms should stem from equal parts of carbon atom 5 of the RlPCHl starting material and from one of the ring carbons (4C) in the product of the reaction according to Eq. (21). Group V Pentoolkyls and Alkylidene Triolkyls 22 1 evolution of hydrogen above -50°C. The final product isolated in high yields was shown to be the cyclic ylide (VIII) , which contains the moieties of the starting materials in a very peculiar arrangement, as indicated by the numbering of the carbon and silicon atoms (Scheme 5). The structure of this ylide was proved by detailed analytical and spectroscopic studies. c Scheme 5 Again the silicon atom of the ylide obtained bears a methyl group, which must originate from the ylide carbon of (CH,) 3PCH2if the reaction scheme (Scheme 5) is correct. In agreement with this assumption it was demonstrated that (C2Hs)3PCHZ and HzSi3 form a product with the three ethyl groups in the expected positions on the six-membered ring: (IN From ethylidene triethylphosphorane and H2Si3 , a cyclic ylide was obtained, in which an ethyl group was, indeed, attached to the silicon atom. 222 HUBERT SCHMIDBAUR J The formation of this species cannot be explained on the basis of any other obvious mechanism. Neither does the initially considered alternative, a C-C bond cleavage of the silacyclobutane instead of a Si-C bond cleavage (77), account for the ethylation of silicon that occurs in this crucial test experiment [Eq. (26)]. The structure of compound X is of primc importance in understanding a mechanism involving a pentaalkylphosphorane, so that some of its reactions wcre studied in addition to its physical and spectral properties. Among these the reaction with methanol was very valuable because this process first yields a well-defined methoxy derivative XI and thcn leads to a complete degradation with formation of ethyltrimethoxysilane : (CZHJZ (CZH5)2 H3C4 CH,OH H3CygP H,C,-Si(OCH,), excess + N3qi CH,OH (27) H5CZ>C 3- H5CI J-- by-products (XI) All of these substitution and degradation products were identified by their 'H, 13C, 31PNRlR and mass spectra, with special decoupling techniques being employed in the former, in order to confirm the assignments. The cleavage of Si-C-P bonds in ylides by methanol is in agreement with previous findings (74),and the same is true for the degradation of the SiCCCP linkage (79). 1-Monomethyl-1-silacyclobutane undergoes analogous reactions with ylides, which may be summarized as follows: Group V Pentaalkyls and Alkylidene Trialkyls 223 (m) There can be no doubt that the products are formed via the same mecha- nism as proposed for the earlier examples. Detection of the CH3(C2H6)Si moiety in compound XIV again demonstrated the partition of moieties of the starting materials in the final product. Still another group of reactions, the mechanism of which requires the intermediacy of a pentaalkylphosphorane, was found with difluorosila- cyclobutane. The cleavage of the ring by an ylide and the closing and re- opening of the cyclic phosphorane are followed by a dehydrofluorination caused by a second ylide equivalcnt. The HF abstraction leads to a recycli- zation analogous to the H2 elimination in the previous examples (Scheme 6).The HF is trapped by (CzHb)3PCHCH3 to form the covalent tetraethyl J H3 c J (xv) Scheme 6 fluorophosphorane, ( C2H5)4PF (64).In the reaction with (CH3)3PCH2 the 224 HUBERT SCHMIDBAUR ionic tetramethylphosphonium fluoride is isolated (67): (XVI) With the corresponding chlorosilacyclobutanes the substitution of chlorine predominates and no ring cleavage is observed (78). In conclusion, the reactions of silacyclobutanes with ylides have pro- vided unambiguous evidence for the existence of cyclic pentaalkylphos- phorane intermediates. The presence of silicon atoms in an a-position to the phosphorus atom of these ring systems does, however, facilitate the reopening of the heterocycle to such an extent that there is little chance for an isolation of these intermediates. It is characteristic of the ring reopening that it occurs exclusively through heterolytic breaking of the P-C-Si linkage, which leads to the silicon-stabilized carbanion. The synthesis of a simple phosphacarbocyclic system without silicon may be possible, perhaps, by using reaction conditions under which the nonpolar cyclobutane itself can be attached to an ylide without inducing secondary reactions [Eq. (19b)1. The reopening of the pentaalkylphosphorane ring must be of much higher activation energy than in the sila-substituted case and the desired species should be more easily trapped, although there is little doubt about the unfavorable thermodynamics of pentaalkylphos- phoranes with respect to the ylidic analogs. IV ARSENIC YLIDES AND PENTAALKYLARSORANES A. Alkylidene Trialkylarsoranes 1. Preparation Only one alkylidene trialkylarsorane has been prepared in a pure state and fully characterized (73).This compound, (CH3) 3AsCH2, can be ob- tained through desilylation of a trimethylsilyl precursor (34) (first synthe- Group V Pentaalkyls and Alkylidene Trialkyls 225 sized by Miller in 1965) : (CH~)~AS=CHZ (CH3)3As + ClCH,Si(CH,), (33) CHIOH I ~L-CIHOI,~ [ (CH3),4sCH1Si(CH3)3]C1 - (CH&As=CH-Si 1 (CH3) 3 Methylene trimethylarsorane (the term “trimethylarsonium methylide” is equally correct) had already been formulated in 1953 by U’ittig and Torssell (104) who studied the reaction of tetramethylarsonium salts with organolithium compounds. Although this method cannot be used for the preparation of the salt-free matcrial due to the strong complexation by the lithium cations, it is clear from reactions of the product mixture that the ylide is present in solution. The formation of arsonium salts upon addition of alkyl halides is a typical example: Li It [(CH,),i\s]Br -(CHJ)3.4~CH2.Li@Br@ - RII (34) [(CH,),ASCHZCH,]I + LiBr 2. Properties, Structure, and Boiitlzng At room temperatixc, methylene trimethylarsorane (73) is a colorless crystalline compound, mp 33”C, which sublimes in a vacuum at 20°C/0.1 mm Hg. It is rapidly decomposed above 33OC; trimethylarsine and poly- methylene are the main products formed. Blue and brown colors develop and some gas is evolved, predominantly ethylene. With trimethylphos- phine, a mcthylene t,ransfer reaction takes place yielding trimethylarsine and methylene trimethylphosphorane : (CH~)~AS=CH~---~(CIT~)~.~S+ (CH,), (35) (CH3)JA~+ (CII~),I’=CHZ (36) The compound is extremely sensitive to oxygen and moisture and must be handled with great care. Thick fumes are observed when the material is exposed to air and spontaneous ignition may occur. A vigorous reaction with water leads to a strongly alkaline solution of some tetramethylarsonium hydroxide, but trimethylarsenic oxide is also formed ( 104) . Tetramethyl methoxyarsornne can be obtained in a quantitative yield (68) with methanol under carefully 226 HUBERT SCHMIDBAUR controlled conditions : H2O (CH,),As=CH, -[ CH,),As] OH, (CH,),AsO (37) CH,OH \1 YH3.,CH, H,C-As,' I .CH, H3C/0 Compound XVII is a colorless distillable liquid of surprisingly high thermal stability. Low temperature lH and 13C NMR spectroscopy revealed a trigonal-bipyramidal geometry of this molecule, which is nonrigid at room temperature on the NMR time scale (68). As expected from current theory, the methoxy group occupies an axial position at the low-tempera- ture limit. With traces of protic species (HZO, CH30H) , a rapid proton exchange is induced in (CH3) 3AsCH2 samples (73),rendering all hydrogen atoms equivalent on the NMR time scale: 7% FH3 7% TH2 CH3-As=CH2 CH3-As-CCH3 CH,=As--CH, CH,--As-CH, I II I I CH, CH2 CH, CH3 (39) The same phenomenon is well documented for phosphorus ylides (75). Methylene trimethylarsorane is a monomer in benzene solution. Study of its lH and 13C NMR spectra led to some unexpected conclusions, dis- tinctly different from the findings with the phosphorus analog (80). As shown in Table 111, the 6 and J characteristics of the CH2 nuclei are not nearly as different from those of the CH3 groups as are those of CH2 and CH3 in the phosphorus ylide (58).Whereas J(H2C) = 149 of the latter indicates an sp2 carbon (the value for ethylene is 150 Hz) , J(H2C) = 131 in the former rather points to an sp3 carbon for the carbanion in (CH3)3- AsCH2. (Methane, 125 Hz; CH3 in the ylide, 133 Hz!) In the absence of X-ray data for an arsenic ylide, it was, therefore, predicted that the carbanion in arsenic ylides should possess pyramidal geometry as opposed to the planar geometry in the phosphorus ylides. ee &AS-CR~' -R3As=CR2' (40) ylide ylene This implies that the ylene formula is even less important for arsenic TABLE I11 NUCLEARMAGNETIC RESONSNCE SPECTRAOF (CH3),P=CH2, (CH3)3As=CHz, AND THEIR SILYLDERIVATIVES~ A 0 (CHI)~P=CHZ -2.3 90.5 149 18.9 56.0 127 - - - n3 (CHI)3P=CH-Si(CH3) +0.7 88.2 134.6 20.0 57.3 128 +4.8 4.4 117.7 2 k (CHd3P=CCSi(CH3)312 +0.3 63.3 21.4 57.3 127 +6.7 4.4 118 5 - Q-. (CH3)3As=CHz 7.6 130.9 15.6 133.8 2 - - - - - 0 (CHI) 3As=CH-Si (CHI) 8.0 141.2 17.1 - 133.0 +5.0 - 116.2 2 (CH~IAS=CCS~(CH~)IIZ 9.3 19.1 134 +6.9 kP - - - - 118 w a From Schmidbaur et al. (69). - b Value of 6 is measured relative to CeD,, and converted relative to TMS. 228 HUBERT SCHMIDBAUR then it was for phosphorus. The very limited stability of arsenic ylides and their high chemical reactivity are in agreement with this proposal. The photoelectron spectrum of (CH3)3AsCH2 shows an energy of only 6.72 eV for the highest occupied molecular orbital, which is predominantly a lone pair localized at tJhe carbanion (69). This energy is lower than in the phosphorus compound (6.81 cV) . This result is at least qualitatively in agreement with the picture drawn above. Very little is known about chemical reactions of arsenic ylides, and the examples reported in the litcrature have been conducted with arylated species often bearing stabilizing substituerits (11, 26). Even from these examples it appears that oftcn the course of thr reaction is entirely different from analogous processes with phosphorus ylides (50). B. Silylafed Ylides of Phosphorus and Arsenic It has been pointed out previously that silylation of ylides leads to stabilizcd products and that this is only one example of the very general phenomenon of carbanion stabilization through silicon (34, 61, 72). This effect was also found for arsenic ylides (34, 73),and is the basis for the preparation of other compounds of this series. Tlie influence of silicon is by no means solely an electronic effect. In many cases, where alkylsilyl substituents are introduced, a steric effect inay wcll dominate, which inay reduce lattice energies for salts in transylidation reactions, preventing intermolecular contacts in decomposition processes, and rendering the for- mation of salt adducts unfavorable. This steric effect is reduced to a minimum, but not eliminated, if simple SiH3 groups are employed (61). Even then, however, a pronounced silicon effect is found, which must be based on electronic influences (49, 60, 61). Of the two explanations advanced, one is the usual ?r-bonding concept involving d orbitals of the third-row elements (34), whereas the other refers to a more electrostatic picture (60). The lone pair of electrons at the carbanion is supposed (a) to intcract with the suitable empty orbitals of silicon and (b) to suffer much less repulsive interactions from bond pairs in the presence of the larger elements. Both arguments call for the planar or quasi-planar geometry of the carbanion, which is observed in phosphorus ylides. In arsenic ylides, As-C multiple bonding scems to be strongly reduced and the nonbonding interactions are also likely to be less important. The proposed nonplanar geometry may well account for this. Because in silylated arsenic ylides the silicon atoms would meet with unfavorable Group V Pentaalkyls and Alkylidene TI ialkyls 229 bonding conditions, it was, thcrcforc, intriguing to find that the 'H and 13C NMR data and the photoelt.ctron spectra again indicated a different situation (69, 80). Whereas with (CH3)3PCH2 the silylatiori is known to cause a decrease in IJ(IH-l3C) in the charbanion by 15 Hz, this coupling constarit was found to increase by niorc than 10 Hz (Table 111) in (CH3)3- AsCH2. A first plausible explarintioii for these opposite cffects has been offcred by postulating a flattening of tlic carbanion pyramid in the arsenic ylide upon silylation, with favorable roiisequericcs for the electronic inter- actions of the carbon with silicoii : Photoelectron spectra have shown tliat the cnergy of the HOMO in (CH,) 3AsCH2 is, indeed, lowered murh inorc by silylation than that in (CH3) ZCHz (69). C. P entarnethylarsorane 1. Preparation In his study of the reaction of tetrainethylarsonium salts with dimethyl- zinc, Cahours (5) originally clairried to have obtained pentamethylarsorane. However, subsequent attempts to repeat this preparation have met with failure (16, lO4). Very recent experiments have led to successful synthesis of this compound through reaction of ( CH3)3AsC12 with inethyllithium under very mild conditions (37).If these precautions arc taken no ylide is formed in this reaction, which was first tried by Wittig and Torssell (104). Deprotonation seems to be a very slow process at low temperature in ether as a solvent so that the attack of the rnet>hylcarbanion at the ar- 230 HUBERT SCHMIDBAUR sonium center can compete successfully. Once formcd, pentamethylarsenic is stable and easily recovered in the work-up of the reaction mixture: LiCHz LiCHx - LiCl 2. Properties, Spectra, and Structure At room temperature, pentamethylarsorane is a colorless liquid of a characteristic odor which resembles the analogous antimony compound. It crystallizes below -6°C and can bc sublimed under reduced pressure at -10°C. The (CH3)gAs is a monomer in benzene solution and shows a molecular ion in the mass spectrum with very low intensity. The vibra- tional spectra, infrared and Raman, could be assigned to a trigonal- bipyramidal skeleton. There are striking similarities to the spectra of Sb(CH3)b (13). The proton ISMR spectrum appears as a singlet over a temperature range of +35" to -95°C and indicates nonrigid behavior. This pseudo- rotation phenomenon is charactcristic for pentacoordinate molecules. The lH signal of the methyl resonance of pentakis-p-tolylarsorane also remains unsplit even at -90°C (20). Compound (CH,) 5A~is slowly hydrolyzed by water, and tetramethyl- arsoniurn hydroxide arid some trinicthylarsenic oxide are formed (37). With methanol, tetramethylmethoxyarsorane is generated with evolution of methane (68). It is iritercsting to note that (CH3) AS and (CH3) 3AsCH2 both lead to identical products with these protic reagents: Hzo - ((CH,),As]OH (42a) (cH,),As j[:H (cH,),As =a, (42~ CH,OH (CH,),AsOCH, - CH, More generally speaking, these are only two examples for a larger series of reactions which are currently under investigation (89). - CH4 (CH3)sAs + €IX [(CH,),As]X (43) (CI-I~)~ASCH~+ HX t Group V Pentaalkyls and Alkylidene Trialkyls 23 1 The thermal decomposition of (CHs)& at 100°C leads to quantitative yields of trimethylarsine, methane, and ethylene, as followed by gas chromatography. Only traces of ethane were detectable. It is, therefore, assumed that the compound is decomposed via the ylide, which is known to be unstable under these conditions: heat ~(CH~)SAS-2(CH3)3;2~=CH2 + 2CI-I4 (44) fast very slow heat I- 2(CII,)3As + C,IL (45) 2C2He Thus, (CH3) AS is at least kinct,ically much more stable than the corre- sponding ylide, arid this relationship should give pentamethylphosphorane a good possibility for isolation if it can be prepared. There is no information available, however, on whether the decomposition occurs via a polar or a radical mechanism. v ANTIMONY YLIDES AND PENTAALKYLSTIBORANES A. Antimony Ylides It is obvious from the literature summarized in recent reviews (11, 86) that very little information is available on ylidic compounds of antimony. Moreover, the fcw compounds synthesized were all taken from the aryl series, whereas those of the alkyl series appear to be completely unknown. Even the aryl antimony ylides were of very limited thermal stability and could only be isolated in special rases (I1 ) . It is, therefore, to be expectctl that a compound such as trimethyl- methylene stiborane, (CH3) 3SbCH2, should be relatively unstable and attempts to synthesize this spcvies would have to be carried out at low temperature. Another difficulty arises from the experimental fact (104) that all conventional methods of synthesis for stibonium ylides lead to pentaalkylstiboranes instead of ylides (see Introduction). Thus mcthyl- ation of (CH3)J3bX or (CHI) 3SbX2 halides by orgariometallic reagents of lithium, magnesium, aluminum, or zinc invariably yield (CHs) &b as the sole product. In a search for other possible routes to trimethylmethylene stiborane 232 HUBERT SCHMIDBAUR the transylidation method in the combination with phosphorus ylides was investigated very recently (69). Although (CH3) sPCH2 was found to convert tetrainethylstibonium salts into ( CH3) Sb, the corresponding ethylidene phosphorane reacted in a different rnanner-(CH,) 3Sb was the only antimony-containing product. As expected for a transylidation re- action, tetraethylphosphonium salt was found in a quantitative yield. This result can be explained in terms of ((CH,)aSb]Cl + (CzH,)31’=CHCH, (CH,),Sh=CH, + [(C2H,),P]Cl (46) heat L(CH&Sb + (CH,), in which the transitory cxistencc of antimony ylide is proposed. Unfortu- nately it has not yet been possible to obtain inore direct evidence for this intermediate, even whcn the reaction was carried out at -25°C. In prin- ciple, hov-ever, processes of this type seem to be very promising because they use ylides as very strongly basic but nonalkylating reactants, which are to be preferred over the strongly alkylatirig organometallic reagents. Further studies arc required, at even lower temperatures, with othcr solvents and a variety of ylides. This section is not cornplctc without mentioning the hydrogen-deuterium exchange experiment by Doehring and Hoffmann (la),which indicated an enhanred acidity of the hydrogens in a tetramethylstibonium salt. In the corresponding equilibrium, stiboriium ylides play the role of strong bases : (CIT,),Sb@-CH,e (CI-I,),Sb@= 5 ]+H@ (47) (ClTs)jSb=CH2 As would be expected, thc proton exchange is much slower than in tetra- methylphosphoniuni and -arsonium salts (la). B. Pentaalkylstiboranes 1. Preparation Pentamethylantimony was the first RsSb species to be obtained by Wittig and Torssell in 1955 (104).Various other derivatives of this type have since been synthesized, and a range of preparative methods has been described. Among these are the alkylation of R4SbX, RSSbX2, and SbX6 species with organometallic compounds of lithium, magnesium, aluminum, Group V Pentaalkyls and Alkylidene Trialkyls 233 and zinc, as follows: KSbX 1<’1,1, Rz’Rlg, Ri’Al RaSbX? R,R’S-nSb (48) ShXs 1WZn I X = halogen R,R’ = alkyl If R = R’, isoleptic pentaalkylstiborancs are obtained, but if It # R’, a whole range of mixed alkylated compounds is present in the product. This is due to a facile alkyl exchange process between R5Sb compounds, as demonstrated by AZeincma and Soltes (33).This exchange is probably catalyzed by organometallic rotgcnts. Pentamethylantimony is also generated in the reaction of pentaaryl- stiborancs with methyllithiuni ; niixed species are the intermediates (54) :* The chemistry of peritaalkeiiq.1stibor:tncs has been carefully studied by Xcsmeyanov et al. (41-48).Thus, pentavinylantimony was prepared from trivinylantimony dichloride arid vinylinagnesiuin bromide (42). The syn- thesis of these compounds obviously prcwmts no problem, even when more complicated alkenyl groups arc to be introduced (44). 2. Properties Pentaalkylstiboranes are palc yc~llowliquids that can be distilled under reduced pressure (Table IV) . Although distillation at atmospheric pressure is also possible, it is to be avoided. A violent explosion was reported in one case when a larger quantity of (CH,) sSb was rectified (still temperature 160°C) (33).When small amounts of It5Sb compounds are heated, slow thermal decomposition is observed, which leads to trialkylantimony com- pounds, alkane and alkene (%a) : (CIH5)bSb +(C?II,),Sb + CzHs + CiH, (50) * Noted added in proof: More recent experiments have shown (59) that up to five tri- methylsilylmethyl groups can also t)c introduced at the antimony-V center, and, e.g., compounds of’the type (CH3),Sb[CIIBSi(CI-Ia)r]s-ncan be easily obtained. Among the decomposit,ion products of pentakis-triniethylsilylmethyl-stiborane(n = 0) a high yield of tetramethylsilane has been detected. This observation suggests that an antimony ylide may be among the primary products of this decomposition, but such a species could not be isolated: [(CH,) 3SiCH&Sb -+ (CH3),Si + [(CH3) 3SiCH2]3Sb= CIISi(CH3) 3 1 decomposition 234 HUBERT SCHMIDBAUR The mechanism of this reaction is not yet fully understood, but it is likely that ylidic species are involved. Vibrational spectra of (CH3) 5Sb have been measured and assigned on the basis of a Da,,structure for this molecule (IS).Valence force constants have later been calculated and large differences are found for axial (1.40 mdyn/b) and equatorial (1.90 mdyn/b) Sb-C bonds of the trigonal- bipyramidal structure (Ira, 108). The IR spectrum of (CZH5) 5Sb has also been reported (Ma). By using proton magnetic resonance, holyever, one is not able to dis- tinguish between axial and equatorial methyl groups, and only one singlet has been recorded even at - 100°C. The usual "pseudorotational" process is invoked to account for this result (39). Pentaethylantimony and mixed methyl-ethyl species show similar Plf R characteristics. The intermolecular alkyl exchange is slow on the NR3R time scale, however, and separate signals are recorded in the mixture of (CHB),(C2Hb) s-,Sb compounds (33). The electronic absorption spectrum of cyclohexane solutions of (CH3) sSb reveals at least two absorption bands (IS). 3. Reactions The chemical reactions of pentaalkylstiboranes have only recently been studied by a number of groups. The bulk of the reactions, some of which TABLE IV BOILINGPOINTS REPORTEDFOR SOME PENTAALKYLSTIBORANES Compound bp ('C/mm Hg) Ref. 126-127/76O 104 130-131/760 33 [mp -16 to -181 104 53-54/16 33 71-74/16 3s 42/0.03 3s 55/0, 4 SS 64/0.4 3s 55.8/0.19 86a 68/5.5 59 65/0.1 59 30/0.001 (subl) 69 60/0 .OO 1 69 50/0.001 (suhl) 69 [mp 931 Group V Pentaalkyls and Alkylidene Triolkyls 235 had already been considered by Wittig and his collaborators (104), may be classified into three or four groups. a. With BrGnsted Acids, pcntaalkylst iboraries undergo cleavage of one or two Sb-C bonds, with evolution of alkane and formation of It,SbX or R3SbX2 drrivatives, respectivcly : Any further reaction of the RZSbX, products is very slow even with strong acids. In the first step in Eq. (51), very weak acids can be employed. Thus, alcohols, water, mercaptans, arid carboxylic acids are now known to react smoothly with pentaincthylantiniony under mild conditions. The following tabulation cites compounds prepared from (CH3) 5Sb and free acids : 1IX (CH6)iSh -(CH,),SbX (52) -CII, x: IT0 RO RS 1tCOO C,HjO F S3, SCN, etc x: 02I’Itz OzPFz 0,I’IIL OSI’R, SOSRz OSlRa Ref.: (8.9) (89) (89) (89) (36) (-56) Depending on the nature of X, the (CH3)lSbX products were found to be either covalent molecular spccic3s (OH, OR, SR, F) or saltlike materials (azide, thiocyanide, chloride, diincthylphosphinate) . The state of bonding may be different in solution from the solid state, and it may vary with the nature of the solvent (carboxylates) . Only the stronger acids, if employed in excess, will afford the R3SbXzspecies. This is true with the halogen hydrides, halosulfuric or -phosphonic acids, ete. b. With oxidizing reagents, a similar sequence of reactions is observed, which may be represented by A typical example (95) is the reaction of (CH3) sSb with bromine, which leads to (CH3)hSbBr or (CH3) sSbBrz, depending on the molar ratio of the reagents. This process has riot yet bccn widely used for preparative purposes. c. With Lewis acids, an alkyl carbanion may be abstracted from the RsSb species with formation of stibonium salts. Such a reaction was demonstrated with aluminum alkyls R3A1 (and alkyl halides R,AlXa-,) 236 HUBERT SCHMIDBAUR and with antimony pentachloride (36, 86) : (CJXdsSb + Al(C,Hd, -[(CzH,),Sbl+[Al(CzH5)41- (54) (CH3)sSb + SbCl5 --+ [(CH1)4Sb]+[CH3sbCls]- (55) With thallium tribromide, a tetrabromothallate complex was obtained with formation of an alkyl bromide and thallium(1) bromide (47): R5Sb + TlBr3 -[R4Sb]+[RT1Br3]- TlBrr I (56) [R4Sb]+[T1Br4]-- [RaSb]+Br- + RBr + TlBr Only toward the strongest nucleophiles do the pentaalkylstiboranes act as Lewis acids. This reaction principle is involved in the interaction of pentamethylantimony with methyllithium, which is believed to yield a hexamethylantimonate complex : (CH3)sSb + LiCH, -I,i+[Sb(CH3)6]- (57) The structure of this product is not known, however, and may well be much more complex (104). VI BISMUTH COMPOUNDS No pentaalkylbismuth compounds R5Bi and no trialkylbismuthonium ylides RgBi=CR2’ have been recorded in the literature (11, 26). If there have been attempts for the preparation of these compounds, they must have met with complete failure. Although the strong oxidizing properties of bismuth(V) , and the extremely unfavorable situation for a interactions with carbon, make it highly unlikely that such compounds can ever be isolated, very recent findings with niobium and tantalum call for caution in making such a prediction. VII RELATED COMPOUNDS OF VANADIUM, NIOBIUM, AND TANTALUM A. Vanadium Ylides? Recent investigations directed toward the synthesis of homoleptic va- nadium alkyls (10) have not been successful for VR5 species, and only Group V Pentaalkyls and Alkylidene Trialkyls 237 VR4, VR3, and VR30 compounds have been obtained (38), even when the promising novel ligand systems R = (CHI)3SiCH2, (CH3) 3CCH2, or [ (CH3) 3Si]2CH were employed. It might have been expected that the reactions could lead to vanadium ylides instead of the pentaalkyls, but this was obviously not the case. From the literature it is not clear whether or not true vanadium(V) compounds have been used as starting materials in at least some of the reactions to warrant the appropriate oxidation state. The problems associ- ated with the vanadium pentahalides (except for the fluoride) and other v& compounds may well be the reason for the difficulties arising in the syntheses. If these difficulties can be overcome, there should be a chance for a successful preparative procedure according to one of the following pathways : RCHiM (RCH,) SV (584 RCHzM VX, -(RCH2)dVX - RH M = metal X = halogen, etc. R = alkyl, silyl, hydrogen Considering the presently available knowledge on vanadium (V) com- pounds, reaction (58b) seems to be more attractive, as compounds of types (RO)aVO, (RO),V=NR, ctc., are well documented for both R = alkyl and R = silyl (4,81), whereas only very few v& species are known. The pertinent chemistry of the congeners niobium and tantalum offers examples for pentacoordination as well as for ylide formation. B. Niobium Compounds It was only in a 1974 report (83) that some evidence was provided for the possible existence of pentamethylniobium through the isolation of its addition compounds with a ditertinry phosphine and with methyllithium. These compounds were synthesized according to the following scheme : 3 LiCH, (CH,),NbCl, (CH,),Nb (not isolated) -~ 3 LiCl /\ 238 HUBERT SCHMIDBAUR From these results it appears that the stepwise alkylation of (CH3) ZXbCls, which should involve the tetrainethylniobium chloride stage, leads filially to the polyrnethylated species Sb(CH3)t-’)e where n = 5, 6, 7, and does not yield ylidic species. (CI1J)sYb LiCHa LlCHj / (C‘H3):XtKi13 A (CII3):SbCl, (CII3)rKhCl (60) - % (CH3),Sb=CHz + CH, In this respect Kb resembles Sb, which is also exclusively converted into Sb (CH,) ?-’I0 structures. With the niore bulky (CH,) zSiCHzligands, completely different products are obtained (26). From thc reaction of NbC15 with (CH3) 3SiCH2MgC1, a binuclear complex of formula [(CHI) 3SiCHz]4Kbn[CSi(CHs) 312 was isolated. The crystal structure of this complex has been deterniined and the niobium centers werc shown to be bridged by tn o trimethylsilyl- mcthylidyrie moieties. The diamagnetism of the material requires spin pairing through direct or indirect metal-metal iritersctions to account for the formal cl’ configuration of the niobiurri atoms : (CH,),Si-CH,, /--hc ,CH,-Si(CH,), Nbl/,”b, (CH,),Si-CH,/ \c/ CH,-Si(CH,), From this result it may be concluded that successive alkylation of Nb(V) with (CH,) 3SiCH2groups does not proceed to the SbRs stage, :M observed in the methylation, but that deprotonation of the CH2 groups does occur, and that ylidic species may therefore be intermrdiates. liltimately, oligo- merization with loss of alkyl and silyl groups leads to dimcrs, however, and no mononuclear compounds have been characterized to date. C. Tantalum Pentaalkyls and Tantalum Ylides Tantalum is the first element of subgroup Va for which both ptritaalkyl and ylidic compounds could bc completely cliaracterized and fully investi- gated. Not only was it possible to prove the existence of Ta(CH3)s and some of its complexes (83), but the “tantalum ylidc,” trisneopmtyl neopentylidrne tantalum, could also be isolated (82a). This ylide is the product of a reaction between trisneopentyltantaluni dichloridc and neo- Group V Pentoalkyls and Alkylidene Trialkyls 239 pentyllithiuin in pentanc at rooin tcinperature. Two equivalents of LiCl and 1 equivalent of neopentane are observed as the sole by-products. The ylide was identified by its ‘H and 13C NAlR and mass spectra, by chemical analysis, and by a cryowopir determination of molecular inass in solution. It could also be obtained, in lower yields, from TaClS and 5 equivalents of the neopentyl Grignard reagent. It has a melting point of 71°C and can be distilled (!) at 75°C under vacuum. The compound is very sensitive to oxygen and moisture, but may be stored indefinitely at room temperature in an inert utmosphcrc. The mechanism of the formation of the ylide has been studied by deuterium-labeling experiments, and it is assumed that the pentakisneo- pentyl tantalum, which is fornicd first, is “decomposed” with elimination of neopentane. This idea parallels the findings for pentamethylarsenic (37),which has also been found to decompose via an ylide intermediate. In the light of these results it, is surprising that the reaction of the silaneopentyl analogs with TaC15 does not afford a similar trimethyl- silylmethylide species. Again, a binuclear complex, as already described for niobium, is formed (26): TaCI, + 5(CH,)1SiCHJlgC1 -[(CI13)3SiCI12]4Ta2[CS~(CH3)3]L (62) It is not clear how this difference in reaction behavior arises as both ligands have at least similar stcric requirements.* D. Metal Ylides or Metal Carbene Complexes? In this article the author has been using the traditional ylide-ylene nomenclature as well as the modcrn phosphorane-arsoranestiborane formalism. No attempt has been made to discriminate between one or the other of these modalities because it is felt that both offer a sufficiently clear description of at least those species that are derived from the main Group Vb elements. (Fortunately no NR5 or RiR, had to be given names.). However, when this procedure \{-as followed with the vanadium, niobium, and tantalum compounds, an interesting problem arose because the authors * Kote added in proof: Very recent work by Schrock has extended this tantalum ylide chemistry very considerably and some cyclopentadienyl tantalum methylenes and benzylidenes could be isolated and fully chararterized (82b).Among the new compounds the complex (CbH,CH3)2Ta(CH3)(CHn)is particularly noteworthy. Its crystal structure has been determined and the temperature dependence of its nmr spectra has been care- fully investigated (1%). The ylidic carbon is in a trigonal planar configuration! 240 HUBERT SCHMIDBAUR of the original communications had baptized some of their compounds “carbene complexes” (8.2).Indeed, an R3Ta=CR2’ compound may well be thought of as a carbene adduct of a (hypothetical) RsTa component. The niobiuni and tantalum compounds of this type would thus merely add to the ever-increasing plethora of transition metal carbene complexes (14). It is perhaps inopportune to elaborate on the nomenclature, but some of the data reported for the “tantalum ylide” indicate that there may be a fundamental difference between this transition metal compound and the formally related ylidcs of the Group Vb elements. The most significant discrepancy is found with the I3C NMR shift of the carbene/ylide carbon atoms, which typically is downfield for the Va element, but upfield for the Vb element derivatives. Ylidic carbon atoms may, therefore, possibly bear a much higher negative charge. Apart from this feature there are many similarities between ylides and carbene complexes, primarily among the structural criteria. The carbene carbon may be, but not necessarily, in a planar configuration, and the M-C bonding indicates some multiple bonding character just as in most of the ylides. On the other hand, carbene transfer rcactions have been observed with ylides [e.g., Eq. (36)], indicating that the carbene complex formalism can, indeed, be successfully applied with ylides. There is hope, therefore, for a fruitful symbiosis of ylide and carbene complex chemistry, which may soon become complementary as more data become available from this new area of transition nietal chemistry. REFERENCES 1. Absar, I., and Van Wazer, J. R., J. Amer. Chem. SOC.94, 2382 (1972). 2. Bart, J. C. J., J. Chem. SOC.B 350 (1969). 3. Bestmann, H. J., and Zimmermann, R., in “Organic Phosphorus Compounds” (G. M. Kosolapoff and L. Maier, eds.), Vol. 3, p. 1. Wiley (Interscience), New York, 1972. 4. Burger, H., Sawodny, W., and Wannagat, U., Monatsh. Chem. 95, 292 (1964). 5. Cahours, A., Justus Liebigs Ann. Chem. 122, 329 (1862). 6. Chioccola, G., and Daly, J. J., J. Chem. SOC.A 568 (1968). 7. Daly, J. J., J. Chem. SOC.A 1913 (1967). 8. Daly, J. J., and Wheatley, P. J., J. Chem. SOC.A 1703 (1966). 9. Daniel, H., and Paetsch, J., Chem. Ber. 98, 1915 (1965). 10. Davison, P. J., Lapperl,, M. F., and Pearce, It., Accounts Chem. Res. 7, 209 (1974). 11. Doak, G. O., and Freedman, L. D., “Organometallic Compounds of Arsenic, Antimony and Bismuth.” Wiley, New York, 1970. 12. Doering, W. v. E., and Hoffmann, A. K., J. Amer. Chem. SOC.77, 521 (1955). 13. Downs, A. J., Schmutzler, R., and Steer, J. A., Chem. Commun. 221 (1966). 14. Fischer, E. O., Angew. Chem. 86,651 (1974); Angew. Chem., Znt. Ed. Engl. (1974). 15. Franzen, V., and Wittig, G., Angew. Chem. 72, 417 (1960). Group V Pentaalkyls and Alkylidene Trialkyls 24 1 16. Friedrich, M. E. P., and Marvel, C. S., J. Amer. Chem. SOC.52, 376 (1930). 17a. Goel, R. G., Maslowsky, E., and Senoff, C. V., Znorg. Chem. 10, 2592 (1971). 17b. Guggenherger, I,. ,J., and Schrock, R. R., J. Amer. Chem. Soc. 97, 6578 (1975). 18. Hager, F. D., and Marvel, C. S., J. Amer. Chem. SOC.48, 2689 (1926). 19. Hellwinkel, I)., Chem. Ber. 98, ,576 (1965). 20. Hellwinkel, D., Angew. Chem. 78, 749 (1966). 21. Hellwinkel, D., in “Organic Phosphorus Compounds” (G. M. Kosolapoff and I,. Maier, eds.), Vol. 3, p. 185. Wiley (Interscience), New York, 1972. 22. Hildenbrand, K., and Dreeskamp, H., 2. Naturforsch. B 28, 226 (1973). 23. Hoffmann, R., Boyd, D. B., and Goldberg, S. Z., J. Amer. Chem. SOC.92, 3929 (1970). 24. Houssain, S., and Schmidbaur, H., J. Organometal. Chem. in prep. 25. Huq, I?., Mowat,, W., Skapki, A. C., and Wilkinson, G., Chem. Commun. 1477 (1971). 26. Johnson, A. W., “Ylid Chemistry.” Academic Press, New York, 1966. 27. Katz, T. J., and Turnblom, E. W., J. Amer. Chem. SOC.92, 6701 (1970). 28. Kost,er, It., Grassberger, M. A,, and Simi6, D., Justus Liebigs Ann. Chem. 739, 211 (1970). 29. Lamberth, R. L., and Seyferth, D., J. Amer. Chem. SOC.94, 9246 (1972). 30. Liittke, W., and Wilhelm, K., Angew. Chem. 77, 867 (196.5). 31. Lumbroso, H., Bertin, D. M., and Froyen, P., Bull. SOC.Chim. FT. 819 (1974), and earlier papers quoted therein. 32. Mak, T. C. W., and Trotter, J., Acta Crystallogr. 18, 81 (1965). 33. Meinema, H. A., and Noltes, J. G., J. Organometal. Chem. 22, 653 (1971). 34. Miller, N. E., J. Amer. Chem. Soc. 87, 390 (1965). 35. Miller, N. E., Znorg. Chem. 4, 1458 (1965). 36. Mitschke, K. H., Dissertation, Universit,y of Wiirzburg, 1972. 37. Mitschke, K. H., and Schmidbaur, H., Chem. Ber. 106, 3645 (1973). 38. Mowat, W., Shortland, A. J., Yagupsky, G., Hill, N. J., Yagupsky, M., and Wilkinson, G., J. Chem. SOC.,Dalton Trans. 533 (1972). 39. Muetterties, E. L., Mahler, W., Packer, K. J., and Schmutder, R., Znorg. Chem. 3, 1298 (1964). 40. Musker, K. W., and Stevens, It. It., Inorg. Chem. 8, 255 (1969), and earlier papers quoted therein. 41. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Zzvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 147 (1960). 42. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Zzvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 952 (1960). 43. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Zzvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 612 (1961). 44. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Izvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 730 (1961). 45. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Zzvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 1578 (1961). 46. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Izvest. Akad. Nauk SSSR, Otdel. Khim. Nauk 1197 (1964). 47. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Iztrest. Akad. Nauk SSSR, Otdel. Khim. Nauk 1202 (1964). 48. Nesmeyanov, A. N., Borisov, A. E., and Novikova, N. V., Tetrahedron Letters 1960, 23. 242 HUBERT SCHMIDBAUR 49. Ostoja-Starzewski, K. A., Tom Dieck, I%., and Bock, H., J. Organometal. Chem. 65, 311 (1974); see also Ostoja-Starzcwski, K. A,, and Bock, H., J. Amm. Chem. Soc. submitted. 50. Piskala, A., Zimmermann, M., Fouquet,, G., and Schlosser, M., Collection Czech. Chem. Soc. 36, 1482 (1971). 51. Sawodny, W., 2. Anorg. Allg. Chem. 368, 284 (1969). 52. Schlenk, W., and Holtz, J., Chem. Ber. 49, 603 (1916). 53. Schlenk, W., and Holtz, J., Chgm. Ber. 50, 274 (1917). 54. Schlosser, M., Kadibelban, and Steinhoff, G., Justus Liebigs Ann. Chem. 743, 25 (1971). 55. Schmidbaur, H., Accounts Chem. Res. 8, 62 (1975). 56. Schmidbaur, H.,Arnold, H. S., and Beinhofer, E., Chem. Ber. 97, 449 (1964). 57. Schmidbaur, H., Buchner, W., and KBhler, F. H., J. Amer. Chem. Soc. 96, 6208 (1974). 58. Schmidbaur, H., Buchner, W., and Scheutzow, D., Chem. Ber. 106, 1251 (1973). 59. Schmidbaur, H., and Hasslberger, G., unpublished results, 1974. 60. Schmidbaur, H., and Malisch, W., Chem. Ber. 103, 97 (1970). 61. Schmidbaur, H., and Malisch, W., Chem. Ber. 103, 3007 (1970). 62. Schmidbaur, H.,Malixh, W., and Rankin, D., Chem. Ber. 104, 145 (1971). 63. Schmidbaur, H., and Mitschke, K. H., Chem. Ber. 104, 1837 (1971). 64. Schmidbaur, H., Mitschke, K. H., Buchner, W., Stuhler, H., and Weidlein, J., Chem. Ber. 106, 1226 (1973). 65. Schmidbaur, H.,Mitschke, K. €I., Weidlein, J., and Cradock, S., 2. Anorg. Allg. Chem. 386, 139 (1971). 66. Schmidbaur, H., Mitschke, K. H., and Weidlein, J., 2. Anorg. Allg. Chem. 386, 147 (1971). 67. Schmidbaur, H., Mitschke, K. H., and Weidlein, J., Angew. Chem. 84, 166 (1972); Angew. Chem., Int. Ed. Engl. 11, 14.5 (1972). 68. Schmidbaur, H., and Richter, JV., Angc~.(‘hem. 87, 204 (1975); Angew. Chem., Znt. Ed. Engl. 14, 183 (1975). 69. Schmidbaur, H., Richter, JV., and Ostoja-Star~e\\.ski,K. A,, Chem. Ber. 109, 473 (1976). 70. Schmidbaur, H., and Stiihler, H., Chem. Ber. 107, 1420 (1974). 71. Schmidbaur, H., Stiihler, IT., and Buchner, W., Chem. Ber. 106, 1238 (1973). 72. Schmidbaur, H., and Tronich, W., Chem. Ber. 100, 1032 (1967). 73. Schmidbaur, H., and Tronich, W., Inorg. Chem. 7, 168 (1968). 74. Schmidbaur, H., and Tronich, W., Chem. Ber. 101, 595 (1968). 75. Schmidbaur, H., and Tronich, W., Chem. Ber. 101, 604 (1968). 76. Schmidbaur, H., Weidlein, J., and Mitschke, K. H., Chem. Ber. 102, 4136 (1969). 77. Schmidbaur, H., and Wolf, W., Angew. Chem. 85, 345 (1974); Angew. Chem., Iflt. Ed. Engl. 12, 321 (1973). 78. Schmidbaur, H., and Wolf, W., Chem. Ber. 108, 2834, 2842, 2851 (1975). 79. Schmidbaur, H., and Wolf, W., unpublished results, 1974. 80. Schmidbaur, H., Wolf, W., Richter, W., and Kohler, F. H., Chem. Ber. 108, 2649 (1975). 81. Schmidt., M., and Schmidbaur, IT., Angew. Chem. 71, 220 (1959). 82a. Schrock, R. It., J. Amer. Chem. Soc. 96, 6796 (1974). 82b. Schrock, R. R., J. Amer. Chem. Soc. 97, 6577 (1975). 83. Schrock, R. R., and Meakin, P., J. Amer. Chem. Soc. 96, 5288 (1974). 84. Speziale, A. J., and Ratts, K. W., J. Amer. Chem. Soc. 87, 5603 (1965). Group V Pentaalkyls and Alkylidene Trialkyls 243 85. Stephens, F. S., J. Chem. Soc. 5640 and 5658 (1965). 86a. Takashi, Y., J. Organometal. Chem. 8, 225 (1967). 86b. The, K. I., and Cavell, It. C., J. (‘hrm. Soc., Chem. Commun. 716 (1975) 87. Turnblom, E. W., and Katz, T. J., J. Amer. Chem. Soc. 93, 4065 (1971). 88. Turnblom, E. W., and Katz, T. J., J. Amer. Chem. Soc. 95, 4292 (1973). 89. Weidlein, J., Ott, X., Mitschkr, Ii.-H., and ICberwein, B., Chimia 29, 262 (1975); in press. 90. Westheimer, F. H., Accounts Chem. Res. 1, 70 (1968). 91. Weygand, F., and Daniel, H., Chem. Ber. 94, 3147 (1961). 92. Weygand, F., Daniel, H., and Simon, H., Chem. Ber. 91, 1691 (1958); Justus Liebigs Ann. Chem. 654, 111 (1962). 93. Weygand, F., Daniel, H., and Schroll, A,, Chem. Ber. 97, 1217 (1964). 94. Wheatley, P. J., J. Chem. Soc. A 5785 (1965). 95. Wittig, G., and Clauss, K., Justus Liebigs Ann. Chem. 577, 26 (1952). 96. Wittig, G., and Clauss, K., Justus Liebigs Ann. Chem. 578, 136 (1952). 97. Wittig, C., and Geissler, G., Justus Liebigs Ann. Chem. 580, 44 (1953). 98. Wittig, G., and Hellwinkel, D., Angew. Chem. 74, 76 (1962). 99. Wittig, G., and Kochendorfer, I<;., Angew. Chem. 70, 506 (1958). 100. Wittig, G., and Krauss, D., Justus Liebigs Ann. Chem. 679, 34 (1964). 101. Wittig, G., and Polster, R., Justus Liebigs Ann. Chem. 599, 1 (1956); 599,13 (1956). 102. Wittig, G., and Rieber, M., Nuturwissenschuften 35, 345 (1948); Justus Liebigs Ann. Chem. 562, 187 (1949). 103. Wittig, G., and Schollkopf, U., Chem. Ber. 87, 1318 (1954). 104. Wittig, G., and Torssell, K., Actu Chem. Scad. 7, 1293 (1955). 105. Wittig, G., and Wetterling, M., Justus Liebigs Ann. Chem. 557, 193 (1947). 106. Woods, J., and Long, G. G., J. MoZ. Spectrosc. 38, 387 (1971). This Page Intentionally Left Blank Acetylene and AIlene Complexes: Their Implication in Homogeneous Catalysis SEI OTSUKA AND AKIRA NAKAMURA Department of Chemistry Faculty of Engineering Science Osaka Universify Toyonaka, Osaka, Japan I. Introduction . 245 11. Acetylene Complexes . 246 A. Structure and Bonding 246 B. Insertion Reactions . 251 C. Catalytic React,ions 26 1 111. Allene Complexes. . 265 A. Structure and Bonding 265 B. Cyclooligomerization 270 References . . 279 I INTRODUCTION The field of acetylene complex chemistry continues to develop rapidly and to yield novel discoveries. A number of recent reviews (1-10) covers various facets including preparation, structure, nature of bonding, stoichio- metric and catalytic reactions, and specific aspects with particular metals. The first part of this account is confined to those facets associated with the nature of the interactions between acetylenes and transition metals and to the insertion reactions of complexes closely related to catalysis. Although only scattered data are available, attempts will be made to give a consistent interpretation of the rcactivities of coordinated acetylene in terms of a qualitative molecular orbital picture. The second part deals with the complex chemistry and catalytic oligo- merizations of allene. We emphasize the importance of the role played by auxiliary ligands of transition metals in determining the paths of catalytic oligomerizations. Recent reviews (11-12?) covering most of the literature published up to 1972 lack a pcrspective view on catalysis. The most recent review (13) describes mainly oligomcrization and cooligomerization. 245 246 SEI OTSUKA AND AKIRA NAKAMURA II ACETYLENE COMPLEXES A. Structure and Bonding The sidc-on coordination of acetylene may be conveniently described by the four orbital interactions shown in Fig. 1. These interactions have al- ready been discussed by Maitlis (7) and Jonassen et al. (14). The extent of overlap and the energy level diff ercnce between these interacting orbitals are important in determining the nature of bonding. The overlap decreases in the order: (a) > (b) > (e) > (d). The energy levels are functions of the effective oxidation state of the metal, the nature of auxiliary ligands, and the substitucnts on thc acetyleriic carbons. Interaction (a) is usually bonding. Interaction (b) is in most cas(xs bonding, since most transition metal ions or atoms have d-dectrons which may occupy ?rll-orbitals. In early transition rnetal cornplexcs such as CpzW=O(RC=CR) (15) or CpNb(C0) (PhC=CPh)2 (16) having vacant d?r,-orbitals interaction (c) is bonding, but the contribution to the bond strength is less than that of (a) or (b). In later transition metal complexes, especially d10 complexes the interaction should be repulsive and antibonding (c’) . Interaction (d) can be neglected because of the poor ovcrlap. We consider that interactions (a), (b), and (c) or (c’) are important in determining thc overall bond strength. The qualitative. inolecular orbital scheme in Fig. 1, although M +c (b) @ Filled orbital 0 Vacant orbital (C) (d) FIG.1. Molecular orbital interactions for coordinated acetylene. (a) Metal u-orbital (s, pu, or du) and qlb; (b) metJal nil-orbital (pall or dq)and all*; (c) metal nl-orbital (pn~or dn~)and Tlb; and (d) metal &orbital (dS) and al*. Acetylene and Allene Complexes 247 naive, is useful for our present purpose in discussing the reactivity of co- ordinated acetylenes. A number of acetylene cornplcxes have been prepared in recent years and structural and spectroscopic data accumulated. Here the nature of bonding will be discussed on the basis of inter alia structural parameters and infrared frequencies. The C-C bond lengths in various types of acetyl- ene complexes arc summarizcd in Tablc I. The C=C bond elongates upon a-coordination. The extent of the lengthening and of the bending of C=C-C angle reflects a major contribution by a* back donation [inter- action (b) in Fig. 13 and a minor contribution of u and a donation [inter- actions (a) and (c)]. The coordinated C-C length of 1.24 in a typical Pt (11) complex with an clcctron-donating acetylene is definitely shorter than the valucs in Pt (0) complexes with a-acidic acetylenes, implying im- portance of back donation. Thc data in Table I also indicate that the majority lies in a relatively small range, i.c., 1.25 + 0.02 A. In an electron- deficient h'b complex (16),howver, a long C-C bond is observed (1.35 A). This may be accounted for by an effective a-donor interaction [(c) TABLE I THECEC BONDLENGTHS AND DEFORMATIONANGLES IN ACETYLENECOMPLEXES (Y Acetylene complex CeC (A)a C=C--R Ref. Pt(cyclo-CTHio)(PPh3)z 1.294 (17) 137", 143" 17 Pt(Cyclo-CsHs) (PPh3)z 1.289(17) 127", 128" 17 Pt(CF&-CCF,) (PPh3)z 1.255 (9) 140.1' (5) 18 Pt(PhCECPh) (PPh3)z 1.32(9) 140" 19 Pt(NCC-CCN) (PPh3)z 1.40 140" 20 Pt(CH3)C1(CF3C=CCF3) (AsMe3)~ 1.32(4) - 21 Pt(CH3)(tripyrazolylborate)(CF3C=CCF3) 1.292 (12) 145.6" 22 PtClz(p-toluidine) (t-BuCeC-t-Bu) 1.24(2) 162"-165" 25 Pd(CH3OzCC-CCO?CII3) (PPh3)z 1.28(1) 145" 24 [Pd(PPhzC=CCF3) (PPh3)Iz 1.28(2) 25 1.29(2) 138" (1) Ni(PhC-CPh) (t-BuNC)z 1.28(2) 149" 26 Nb(CsH6) (CO) (PhC=CPh)z 1.35(2) 138" 16 Nb(C5HS)(CO) (PhC=CPh) (Ph4C4) 1.26(4) 142" 27 WO(CbH,)z(PhCECPh) 1.29(3) 143.5"(4) 15 Ir(trans-C(CN)=CHCN) (CO)(PPh& (NC-CEC-CN) 1.29(2) 139" 28 W (PhCsCPh)3(CO) 1.30 140" 29 Standard deviation in parentheses. 248 SEI OTSUKA AND AKIRA NAKAMURA in Fig. 11. A further structural study of related electron-deficient acetylene complexes is desirable to confirm this inference. The deformation angles are also in a small range, 143" f 5". Larger values (162"-165") in [PtCln- (toluidine) (t-BuC-Ct-Bu) 3 suggest an extensive ?r-donor interaction [(a) in Fig. 11. Quantitative interpretation of these structural parameters seems difficult at the present stage (28). The infrared frequencies (SO) associated with metal-acetylene bonding (MC2) can be factored into three fundamentals (2al + b1) if the system is regarded as a vibrationally isolated, triatomic, isosceles ( CzV local sym- metry) : I v,(a,) v,( UJ v,(b I) Coupling is expected between these vibrations and other vibrations of the rest of the molecule. Fundamentals of the essentially metal-carbon vibra- tions (v2 and ~3)have never been assigned due to the extensivc coupling. Vibration vl(al) is remarkably high (1600-2050 cm-l) and relatively iso- lated (Tables I1 to IV) . Therefore, v1 values can be used in discussing metal-acetylene bonding. Thus, Pt (11) Clz(RC=CR) L (38) or [Pt (11) (CH;) (RC=CR) L]PF6 (38) exhibits a band at about 2000 cm-' (ca. 250 cm-l lowcr than for the free acetylene; see Table IV) , whereas Pt (0) (RC=CR)L2 shows a strong band in the range, 1680-1850 cm-' (see Tables I1 and 111). The MC2 frequency (al) in isostructural iridium com- plexes decreases with an increase in the electron-donating ability of the auxiliary ligands, as clearly shown in [Ir (CO) & ( RCzCR) ]+ (see Table 111); however, the trend in Pt(CF3C=CCF3)L is obscure (Table 111). It appears that the extent of ?r back donation is so great that the MCP frequencies become insensitive to a change in the phosphine ligands. The data in Table I1 suggest that the apparent thermal stability of the isostructural acetylene complexes of zero-valent metals (Pt > Ni > Pd) is mainly governed by the backbonding effect as reflected in the decrease ( Av) in v ( MC2) ( v1 vibration) frequencies ( Av: Pt > Ni > Pd) . This cor- relation docs not occur among complexes of different structures (see Tables I1 to IV) . Because of the varying extent of mixing, a meaningful compari- son of metal-acetylene bond strengths requires, first of all, assignment of the al vibrations, particularly v~,through isotopic (e.g., 13C) infrared data. The IR N=C frequencies of a series of complexes, M(Un) (t-BuNC)z (M = Ni, Pd), are useful for discussing the nature of metal-acetylene Acetylene and Allene Complexes 249 TABLE I1 COMPARISONOF THE MC2(al) FREQUENCIESIN ISOSTRUCTURALACETYLENE COMPLEXES Ph COtMe CN CF3 M (R-CEC-R) (PPha)za (cm-I) (cm-1) (cm-I) (cm-l) M = Pt 1740, 1768 1782 1682 1775 Pd 1830, 1845 d 1811, 1838 Ni 1800 d d 1790 Free acetylene ligand 2223 2256 2218 2300 Ph C02Me M (R-C-C-R) (t-BuNC)z” (cn-l ) (cm-I) M = Ni 1810 1830 Pd 1825 1804 M = Mo 1613 1830 1774 1782 - v - - - 1800 1821 Pt(RCECR’) (PPh3)il: PtCz(a1) PtCZ(ad RCECR’ (cm-1) RC=CR (cm-I) MeCECH 1712 EtCECEt 1805 EtCECH 1705 ptolyl-C~C--p-tolyl 1755 PhCECD 1642 Data from Davidson (30) and Wilke and Herrmann (31). b Data from McClure and Baddley (52) and Baddley (33). The acetylene complex is too unstable to be readily isolated. Thermal oligomerisation of the acetylene prevents characterization. (I Data from Otsuka et al. (34). I Data from Nakamura and Otsuka (55),from Brintzinger and Thomas (36),and from Tsumura and Hagihara (37). 0 Data from Mann et al. (38). bonding. The correlation bctwcen the N=C stretching values and the electron density on the metal provides a measure of metal-acetylene dona- tion and/or back donation. In general, high NrC stretching values are associated with a high effectivc oxidation state of the metal. Table V ap- pears to suggest that dimethyl acetylenedicarboxylate and dimethyl 250 SEI OTSUKA AND AKIRA NAKAMURA TABLE I11 EFFECTOF THE PHOSPHINESUBSTITUENT ON THE LIGANDINFRARED FREQUENCIESIN ISOSTRUCTUR.4L ACETYLENECOMPLEXES (a) [Ir (C0)ZLz(MeOzC-C=C-COzMe)]+ a : vco VIrCn VC-0 of COlMe L (cm-’) (cm-l) (cm-1) PPhi 2091, 2049 1814 1708 PMePhz 2086, 2038 1808 1704 PEta 2088, 2044 1791 1697 PCya 2068, 2018 1794 1703 (b) PtLz(CFiC=CCF,)b: VPtC2 VPtC, L (cm-I) L (cm-l) PPha 1765, 1775 PMezPh 1767 PEta 1771 P (*Bu) 3 1758 a Data from Church et al. (39). b Data from Davidson (SO) and Cherwirski et al. (40). maleate in nickel complexes are comparable in dal ,-accepting property. This is rather peculiar since an acetylenic carbon is more electronegative than the corresponding olefinic carbon. In analogous Pd complcxcs (Table V), we observe the expected trend for the N=C values. The relatively lower YN=C values in the nickel complexes arc explicable in terms of an enhanced interaction (e‘) that donates electrons to the isocyanide bonds. TABLE IV SELECTEDVIBRATIONAL FREQUENCIES OF tram-[PtCH 3 ( RCEC 12’) Lz] ‘PF 8- 5 RC=CR’ L VPt-CHJ VPtCz(a1) CH~CECCH~ PMezPh 547 2114 CHiC4XzHs PMetPh 55 1 2116 CzH,C=CCzH, PMezPh 547 2101 PhC-CPh PMezPh 556 2087 PhC=CCH(OH)Ph PMezPh 547 205 1 PhCECPh AsMeo 549 2024 a Data from Chisholm and Clark (41). Acetylene and Allene Complexes 25 1 TABLE V COMP 4RISON OF NEC s rRE I'CHING FREQUENCIES IN 11 (UII)(t-l%uX( ')p(,ZI =Xi, I'd)' VNZC RI Un (cm-') AVb Ni (c~s-CH~O~C-CH=CH-CO~CH~) 2154, 2120 - - (cI~3o,c-c~c-co~crl3) 2160, 2123 +6 +3 (PhCECPh) 2138, 2100 -16 - 20 Pd (C~S-CH~O~C--CH=CH-CO~CH~) 2160, 2140 - - (CH~OZC-C~C-COZCH~) 1177, 2158 +17 + 18 (PhCECPh) 2150, 2125 -10 - 15 a Data from Otsuka et al. (34). b Difference in the frequencies fiom thc values of the coirespmdmg dimethyl maleate complexes. B. Insertion Reactions Insertion of acetylenes into transition metal-hydrogen, -@-carbon, -+ acetylene, or -halogen bonds is an important elementary step in catalytic hydrogenation, oligomerization (lincar or cyclic), or polymerization. In recent years, considerable information has become available in this field, particularly with regard to stercorhcmistry and mechanism. Kinetic data, however, still remain scarce. Thc~bonding between metals and acetylenes is considerably complicated by thc existence of the two mutually ortho- gonal s-orbitals on the ligand, and insertion reactions exhibit complex features reflected in the varic.d rclgio- and/or stereoselectivity. A short ac- count of the stercocht.mistry arid mechanism of the insertion into an H-M bond will be given to illustrate the situation. A delicate dependence of thr &/trans stereochemistry on the identity of the metal, oxidation state, auxiliary ligands, and the nature of the sub- stituents on the acetylene has already been observed (42, 43-65). The complexity of the problem, however, has prevented general interpretation of the mechanism (4.2). The stcrt.ochemistry determined after the decom- position by hydrogen (or sonietimc,s by water) of a labile a-alkenyl com- plex is sometimes ambiguous (43)bccause of possible geometrical isomeri- zation during decomposition. Fluoroacetylenes, CFsC=CH or CF&= CCF,, may be used with advantage because of their ability to form stable u-alkenyl complexes which can be easily examined by lH and I9F NMR spectroscopy. The observed coupling constants [e.g., J (CF3-CF3) ] un- ambiguously reveal the stereochemistry about the double bond in the 252 SEI OTSUKA AND AKIRA NAKAMURA a-alkenyl complexes. The stereochemistry of insertion determined with the fluoroacetylenes may be depicted as follows. CFSC-CH Trans insertion to HML,: ML,:Re(CO)s (52,53), FeCp(C0)z (53) Cis insertion to HML, : ML, :Mo (H)Cpz (47) In addition to these, a-trifluoromethyl products were obtained regioselec- tively from CFaC=CH and PtHC1(PEt3), (53) or MnH(CO)5 (53) with- out stereochemical information. For the insertion of diphenylacetylene or dimethyl acetylenedicarboxyl- ate the steric course is sometimes uncertain because of the meager struc- tural information available for the alkenyl complex. Hopefully I3C NMR data will provide the necessary information. PhCECPh Cis insertion to HML, : MLn = Mo (H)Cpz (467, IrClz (dmso)z (54), PtCl ( PEt3) (55),Co (dmg)2 (56) Trans insertion to HML, : ML, = Rh( CO) (PPh3)3 (52), COCP (PPh3) (57) MeOzCC=CC02Me Cis insertion to HML, : ML, = Mo (H)Cpz (46), W(H)Cpz (46) Mn(C0)sPPha (58),IteCpz (591, RuCp(PPh3)2 (5O), Rh(CO) (PPh3)3 (43), Pd(CzCPh) (PEt3)z (GO), PtCI(PEt3)a (60) Acetylene and Allene Complexes 253 Trans inscrtion to HML,: ML, = Mri(CO)5 (58, 61), Rh(C0) (PPhO3 (51) Various mechanisms for thr insertion reaction are conceivable: (a) ionic stepwise, (b) radical (chain or nonchain) , and (c) concerted. Generally, ionic or radical mechanisms give a mixture of products with cis and trans stereochemistry. In some special cases of the ionic reaction, however, ex- clusive formation of a trans product has been observed (62, 66). Therefore, stcreoselcctivity does not necessarily imply a concerted mechanism; other evidmee, e.g., regioselectivity, kinetic data, solvent effects, and substituent effects, must be sought out. The addition of fluoroacetylenes to hydrogen compounds of some typical elements, e.g., R2EH (E = N, P, As) (66) or RXH (X = 0, S) (67), proceeds via an ionic stepwisc or radical mechanism; for example, E = N, P, As The distribution of the products depends on the electronic properties of R2Eand the reaction conditions (66).By contrast, the reaction of CF,C=CH with Cp2MoHnproceeds with high rcgio- and stereoselectivity to give cis insertion over a range of reaction temperatures (47): F,C\ ,H CF,C-CH + Cp,MoD, ,c=c -Cp,Mo\ \D D Similar cis insertion was observed for diphenylacetylene or dimethyl acetylene dicarboxylate (47). A mechanism analogous to the rclated olefin cis insertion (68) has been proposed for the reaction with Cp2MHz (M = Mo, W) (Scheme I), on the grounds of the same stereochemistry and comparable reaction tempera- ture (47).In view of the failure of CF&=CH to react with the moderately a-basic Cp2WHz (pKb = 8.6) at ambient temperature, the facile reaction of the weakly a-basic Cp2MoH2 (p& = 12.5) with CF&=CH cannot be ascribed to the a-basicity at the metal. Intervention of a thermally excited, parallel metallocene molecule with high a-basicity is thus assumed. The excited molecule will then receive a a-acidic acetylenic bond forming an acetylene a-complex that smoothly gives the observed cis-insertion product. The stereoselective cis insertion is best accounted for by a four-centered 254 SEI OTSUKA AND AKIRA NAKAMURA I CF,CECH Scheme 1 transition state with some polarity: The polarity and contribution of d-orbital character seems to assist this formally forbidden [azs+ T~,]reaction, and the process may not be rigor- ously synchronous (69). Reccntly, trans insertion of hexafluorobutyne into one of the M-H bonds in some metallocene hydrides, CpZMH,, was studied in some detail (47).Experiments carried out in the presencc of various radical-sensitive reagents such as N-phenyl-a-naphthylamine suggested that a free radical mechanism was unlikely. A stepwise ionic mechanism, involving a zwitter- ionic intermediate, CpZ(Hz)M+-C (CF3)=cCF3, is improbable, since (i) the stereochemistry and the apparent rate are not influenced by the polar- ity of the solvents, (ii) no deuterium is incorporated in the reaction in EtOD, and (iii) the trend in reactivity (Mo > W) does not reflect the trend in a-basicity or M-C bond stability (W > Mo). An essentially concerted trans-insertion mechanism is inferred, which is supported inter alia by the low kinetic deuterium isotope effect (kH/lcD = 1). Acetylene and Allene Complexes 255 The concerted trans insertion formally belongs to a thermally allowed + ma] reaction utilizing the acetylene a, orbital. A nonpolar four- centered heteroatomic transition state with a skewed disposition of par- ticipating U- and ?r-bonds may bc postulated: The geometry in the transition state readily explains the preferential formation of the conformational isomer a (46) (Scheme 2). Y-,,,, Isomer a Isomer b Scheme 2 It is interesting to observe different mechanisms for the apparently simi- lar insertion reactions betwccn Cp,MoH2 and fluoroacetylenes. Thus, Cp2MoHz acts as a a-base for CF3C=CH with a polar triple bond, but in the essentially concerted trans insertion it may behave as a c2*component against a nonpolar bulky fluoroacetylcne, CF3C=CCF3. The observed dis- criminating behavior toward these fluoroacetylenes contrasts sharply with the nondiscrimination of typical N-bases ( R2NH or &PH) (66) in reac- tions of the fluoroacetylenes. Specificity of transition metal complexes in reactions with apparently similar organic substrates is thus of interest and deserves further study. 256 SEI OTSUKA AND AKIRA NAKAMURA 1. Metalocyclization Metal +acetylene complexes react with further molecules of acetylenes in two different ways, namely ligand exchange or substitution (Scheme 3). R' R' C- CR' R' Scheme 3 Insertion initially gives metalocyclopentadicnes which may further give rise to larger metalocyclic complexes. Factors determining the reaction paths are not clear at present. In general many metal acetylene complexes of dad0metals (70-89), e.g., Fe (3,YO), Co (10,7l),Ni (72),RU (73,74), Rh (75-78), Pd (83-84), and Ir (86),react with excess acetylene to give metalocyclopentadicne complcxes or acetylene oligomers (87). The follow- ing are some examples: PPhs RCECR CpC( R \ c' R=CO,CH, c+ R R' R &Pd---lllCYR R=CO,CH,RCGCR = &PdgR c 'R R R In these complexes, repulsive interaction (c') is operating in addition to interactions (a) and (b) (see Fig. 1).Thermal excitation then either causes Acetylene and Allene Complexes 257 thcsc complexes to lose the acetylene or gives rise to a thermally excited molecule in which a change in coordination state (to a monohapto state) may occur (73, 89). A similar thermal activation has been proposed (90) for soine insertion reactions of dioxygen complexes, L2MO2 (M = Ni, Pd, Pt), where the orbital interaction scheme is similar except for the occu- pancy of the dioxygen rr*orbital: / c thermal + +c- . ,c- M-C -M-C' M---''' excitation = I I ti I The radical or ionic character dcpcnds on the identity of the metal, the effective metal oxidation statc, and the auxiliary ligands. These combined effects determine the reactivity, stereo- and regioselectivity toward acety- lene insertion. For example, thc reaction of CpCo(PPh3) (RC=CR') with asymmetric acetylenes, RC=Clt', gives a mixture of isomeric products: PPh, R-CGC-R' cpco I R R R' R' R R' R Isomer a Isomer b Isomer c (i) R = CO,CH, , R' = CH,: a. 9%; b. 50% (ii) R = Ph, R' = CO,CH,: a. 13%; b. 20% A predominance of isomer b and the absence of isomer c indicates the direc- tion of polarization in the metal +acetylene moiety (91). A zwitterionic intermediate, Cp ( PPh3)Co+--C (CHI)=e-CO2CH3, is implied for case (i) ; much less polarization with some radical character would account for the isomer distribution in case (ii) . Recently, a similar polar monohapto- acetylene intermediate was invoked to explain the novel addition of 258 SEI OTSUKA AND AKIRA NAKAMURA CF3C=CCF3 to a C-H bond of an Ru-alkenyl complex (92)l: I R R R = CO,CH, For reactions of cationic Pt-acetylene complexes (93), another polar monohaptoacetylene complex (C) may be postulated: / L3Pt 'I + 'C\ \ ( C) ( D) L L Y CYR I+ Me-Pt--ill - 1 c\R R L I ,OMe Me -Pt -C, I CHR, L Thus, most of the electrophilic reactions of [PtCHS(L)z(RC=CR)]+ can be explained with the intermediate mechanistically indistinguishable from the platinieed carbonium ion model (D) proposed by Chisholm and Clark. Since the positive @-carbonhas a vacant pr-orbital, the rearrangement to alkoxycarbene complexes can be regarded as a carbonium ion rearrange- ment. In contrast to the later transition metal complexes, electron-deficient complexes of the earlier transition metals, e.g., CpV(CO), (RC=CR) (94), are mostly inert to acetylene cyclization. Thus, bis- or trisacetylene com- 1 The molecular structure of the product [Ru.C(C02Me) :C(CO2Me)C(CF3)k(CFs)H- (PPhs)(+CaH5)]has been fully confirmed by an X- ray diffraction study (L. E. Smart, J. Chem. SOC.,Dalton Trans., in press). Acetylene and Allene Complexes 259 plexcs, e.g., M (RCECR) 3( CO) (95-97), are prepared by thermal and photochemical reactions. The absence of stable bis or tris complexes of d8-d10metals may be attributed to the facile metalocyclization. Thus, inter- action (c) (see Fig. 1) or (c') appmrs to impart this distinction. In molyb- denum acetylene complexcs, ('pJ10 (RC=CR) (46, 47), intermediate between the two cases, the relevant d~,orbital is partially occupied by interaction with the El, (Cp) orbital, resulting in a weak attractive interaction (c) . Indeed, these acetylene complexes are inert to acetylene oligomerization. A further reaction of metalocyclopentadiene complexes with acetylenes leads to metalocycloheptatriene complexes by metalocyclic enlargement (3, 10, 98) or to benzene derivatives by reductive elimination (57, 70, 73, 7'7, 8,2,98): or M$ M$ metalocyclicenlargement ~ M$ 1 cis, cis,cis cis, trans, cis \aromatization isomer isomer For example, reaction of excess CFBC-CCR with Pt ( PEt3) gave ( Et3P)%- Pt[q2-Cs (CF3)61, which probably formed from a +acetylene complex through a platinacyclopentadiene complex (98).2 Rerently, Stone et al. [J.Chem. Soc., ('herti. Commm., 723 (1975)] have established a reverse pathway for reactions of this kind. Certain complexes of Pt(0) react with Ce(CF& with cleavage of a CCbond of the benzene derivative to give a platinacyclo- hepta-cis, cis, cis-triene. 260 SEI OTSUKA AND AKIRA NAKAMURA A related reaction with Ni[P (0Me)J.i gave unexpectedly a cis,trans, cis- nickelacyclohcptatriene complex that must be constructed with a trans- insertion process at some stage of its formation (98) : R Ni[P(OMe),], + CF,C-CCF, -[(MeO),P],Ni SR RR (R = CF,) 2. Metalococyclization of Acetylenes with CO or RNC Metalocycles arc also formed by the reaction of acetylenes with metal carbonyls or with isonitrile complexes (3, 73, 99-102). Their formation may involve monohaptoacetylene intermediates. 0 0 II I1 RCZCR M(C0) - /I 0 NR Some of these metalocycles have been confirmed by spectroscopic data and by an X-ray analysis (103) and are important intermediates in catalytic cocyclization with carbon monoxide or with isonitriles (see following). Acetylene and Allene Complexes 26 1 NN / \ R‘ R‘ C. Catalytic Reactions 1. Activation of Acetylene by Coinplexatioii Studics on elementary reactions of acctylcnes with metal complcxcs are now beginning to shed some light on tlw nature of “activation” caused by complexation. This activation is not a simple process. Many low-valent d*-dIO metal complexes and also som(’ rarly transition metal compounds with higher oxidation state ( &d2 complexes) are capable of activating acetylenes. As already describcd, in thc former complexes, interaction (c’) would lead to activation of an +acctylcne ligand to an 7’-acetylenc having some radical as well as some anionic character: $-Acetylene ql-Acety lene complex complex In the latter complexes, strong r-donor interaction (a) and weak r-back donation (b) (see Fig. 1) would lrad to thc formation of apparently similar +acetylene complexes by thrmmal activation. Here the species, however, have some cationic character as manifrstcd by their preferential reactions with electron-donating acety1tmc.s (63, 104) : In sharp contrast to thew activations, an q2-acetylme complex is stabil- ized when all the interactions [(a), (b), arid (c) ] arc bonding, as in some electron-deficient d6 complexcs, c’.g., W (RC=CR) (CO) (95-97). 262 SEI OTSUKA AND AKIRA NAKAMURA 2. Catalytic Cyclooligomerization Acetylenes are catalytically cyclized to benzenes and cyclooctatetraenes (70, 105-107). Small amounts of styrenes, vinylcyclooctatetraenes, naph- thalenes, and azulenes are also formed in some instances (108-110). Some elementary steps in these reactions have already been discussed. A plausible reaction path for the cyclization is in Scheme 4 (111). Scheme 4 3. Linear Oligomerization Acetylenes are also oligomerized to mono- or divinylacetylenes, or dienyl- acetylenes by Ni(0) (lid), Rh(1) (IIS), or Pd(1I) (114) complexes (Scheme 5). Meriwether et al. (119) proposed hydrido-u-alkynylnickel complexes as active intermediates in the catalytic linear oligomerization. Subsequent insertion of acetylene into an M-u-alkynyl bond has been assumed. RCGCH R'C thermal ML?I - III---MLm excitation HNC I q'-Acetylene complex Scheme 5 Acetylene and Allene Complexes 263 Conversion of an +acetylene complex to the hydridoalkynyl complex will lead to linear oligomerization or polymerization. The tendency of some Iih or Pd complexes to form hydridooalkynyl complexes explains their cata- lytic activity toward linear oligomcrization. Recently, Hagihara et al. (115) examined the reaction of preformed hydrido-a-alkynyl complcxes, MH (U-CECPh) Lf,with Mc02C'C-CC02Me and found cis inscrtion into the M-H bond. It R II tians-~I€I(C~C~'h)LL+ ItC'-C11 -Iruns-M(C=CHh) (-C=CII)L, 31 = Pd, Pt; 1, = 1'b:t3; 12 = CO&le They also found a novel stercosclective linear trimerization of PhCFCH with a Pd (Ph) (PBu3)2 catalyst. 4. Catalytic Cocyclixation with Isocyanides Cocyclization of acetylene with isocyanides gives interesting new cyclic compounds (103, 116). The reaction patterns are generally similar to the cocyclization with carbon monoxide which is already known (103, 11 7). Low-valent nickel, palladium, or cobalt complexes are active in the follow- ing reactions (102, 10.9) for which intervention of acetylene complexes has been suggested : Ni, Pd, or RCECR + R'NC Co complexes = -N EN-+\NgN/ +q N\ N- Recently, Yamazaki et al. (10.9) carried out stoichiometric reactions of cobalt-acetylene complexes with isocyanides and isolated the expected intermediate metalocyclic complexes (Scheme 6). Another interesting rcactiori is the formation of aminopyrrolc deriva- tives from t-BuNC and various acctylenes (118). The catalysts include various Ni(11) and Ni (0)phosphine complexes: t-B" Based on the formation of a C~CO(CNR')~(RC~CR)complex from acetylene and isocyanides (103), the paths shown in Scheme 7 are proposed. 264 SEI OTSUKA AND AKIRA NAKAMURA RCACR I Y R R RR R'R = Ph2,6-Dimethylphenyl R'-N RlNqNRl ~l-~,N I R' Scheme 6 5. Catalytic Cocyclizatioit with Heterounsaturntion Yamazaki et al. (91, 119) and Bonnemann et al. (120) have recently reported catalytic syntheses of substituted pyridines from acetylenes and nitrilcs. Various cobalt complrxrs serve as active catalysts, in particular, cpC0(PPh3)~(91) 01 Co(C8HI2)(C8HI3) (120). Similar reactions of acety- lenes with CS2 or RNCS also give new heterocycles (91) : 1 R'NC Ni(R'NC),(RC-CR) - \ N &LN. Ni NH-R' R' Scheme I Acetylene and Allene Complexes 265 RCECR + R’CN A RCGCR + CS, -)I+f C’ /I S RCECR + R’NCS - /I S Thc following intermediate coniplcx has also been isolated (121) R This structure gives support for thc proposal of an ionic monohaptoacety- + lcne complex, Cp (PPh3) Co--C (R)=(%, as an activated precursor for the reaction with CS2. 111 ALLENE COMPLEXES A. Structure and Bonding Since the compilation of cornplcxcs in 1972 (11-13), a few have been reported: M(PPh3)2(allenc) (122, 123) (M = Ni, Pd, Pt; allene is CHZ=C=CH2, CH~=C=CMP~, PhCH=C=CHPh) ; CpzFc (CO)- [CHF=C=C ( CH3)SnCl] (1.24, 125) ; and PtCH3( HBpz3) (&C=C= CR’2) (HBpz3 = tripyrazolylborato; It = R’ = CHI or R = H, R’ = CH3) (126). Single-crystal X-ray diffraction data arc shown in Table VI. As in olcfin complexes, there arc two types of coordination-one containing an allcnc double bond perpendicular to the equatorial molecular plane and the other containing the ligand in the plane. In the former type, we find labile complexes, e.g., RhI ( PPh3),(CH2=C=CH2) (136) and fluxional tu 0. TABLE VI 0. STRUCTURAL PARAMETERS OF TRANSITIONMETAGALLENE COMPLEXESa C (1);C(2) C (2)-C (3) C (1)-C (2)- Compound Structure 01) (A) C(3)(”) M-C(l) M-C(2) Ref. Perpendicular b 1.40 (1) 1.30 (1) 153.3 (6) 2.13(1) 2.07(1) 127 1.41(1) 1.29(1) 132.6(6) 2.13(1) 2.06(1) Perpendicular* 1.373 (8) 1.325 (8) 147.2 (6) 2.177 (6) 2.027(5) 128 1.377 (8) 1.321 (9) 148.9 (6) 2.176(6) 2.033 (5) Rhz(acac)z(CO)z(CHz=C=CHz)c Perpendicular * 1.37 (1) 144.5 (6) 2.12(1) 2.05 (1) 127 1.41(1) 2.14(1) 2.06 (1) R~I(PP~,),(CHZ=C=CHZ)~ Perpendicular* 1.3.5 (6) 1.34(7) 158 (4) 2.17(4) 2.04 (4) 129 [PtC12(RiezC=C=CMe,)]2 Perpendicularb 1.37(3) 1.36(3) 151 (2) 2.2.5(2) 2.07(2) 128 Pt (PPh3)z(CII?=C=CHz) In-plane* 1.48(5) 1.31(5) 142 (3) 2.13(3) 2.03 (3) 130 Pt(PPh3)z(CHz=C=CHMe) In-plane* 1.44(4) 1.32(4) 146 (3) 2.12(3) 2.05(3) 131 Pt(PPh3)z(CHz=C=CMez) In-plane * 1.430 (11) 1.316 (11) 140.8(8) 2.107 (8) 2.049 (7) 132 Pd (PPh3)s(CIIz=C=CH?) In-plane* 1.44(2) 1.32(2) 148 (1) 2.12(1) 2.07 (1) 1.33 S(l)-y2) C(2),S(3) s (1)-C (2)- 0%) (11) s (3) (“1 blGS(1) IiGC(2) 1.72(5) 1.54(5) 136 (1.5) 2.33 (1) 2.06 (4) 1.34 1.65 (3) 1.63 (3) 140 (2) 2.305 (11) 2.00 (3) 135 a Carbon disulfide complexes are included for comparison. The C=C bond length of free molecules ranges from 1.300 to 1.312 ,i. *Perpertdzcular refers to the molecular structure containing a coordinated double bond perpendicular to the niolecular plane, and in-planeto that containing the double bond lying in the molecular plane. c Each of the two double bonds acts as a monodentate olefin and thus the bent allene bridges the two nietal atoms. The bromo analog, RhBr(PPh~)z(CHz=C=CIIz),has quite similar structural parameters (Kasai et al., unpublished). Acetylene and Allene Complexes 267 ones, e.g., Pt2C14(Me2C=C=C.Rle2)2 (1S7, 138). In the latter are found both labile, e.g., P~(PP~~)~(C'HF==C'=CH~)(139) and inert compounds, e.g., Pt (PPh,),(CH2=C=CH2) (1.29) (see following). In general the degree of clongation of the double bond upon coordination parallels the degree of bending of the allene molecule. The Dewar-Chatt- Duncanson molecular orbital model of the mctal-olcfin bond would ac- count for these features. The central carbon-metal distance is shorter than the other carbon-metal distance and may be explained by an additional interaction between a filled metal d-orbital and the olefiri ?r*-orbital with the uncoordinated double bond. However, this view, although attractive, appears not to be supported by the fact that the M-C (2) distance remains nearly constant, within standard (wm, rc.gardless of the number of methyl substitucnts at the uricoordiriat cd double-bond carbon in Pt (PPh3)2- (allene). The X-ray bond distanccs may not be sensitive to the electronic variation or they may simply bc a reflection of the atomic radii susceptible to the change in hybridization. In thr case of Pt ( PPh3) ( CH2=C=CMe2), a nonbonding rtpulsion exists htwwn thc phosphine ligand and one of the methyl substitucnts, as the diffcrcnce Fourier map indicates the particular methyl group to be a hindered rotator (132). The repulsion is also reflected in the two P-metal-C angles. This stcric factor may be responsible for the apparent irregularity in the bmding [C (1)-C (2)-C (3) angle] and, hence, for the absence of a linear corrcllation between the angle and the distance of the coordinated double-bond [C' ( 1)-(" (2)] in Pt (PPh3) (allene) . A series of nickrl triad complexes ML2(allene) (Table VII) were pre- pared (192, 193, 139) and studicd in solution by means of 'H NMR spec- troscopy. Consistent with a planar niolecular structure, the 1H NMR spectrum shows three signals with the proton signal at site a highest field, and the signal at site c lowest field. The following features are conspicuous: (i) large Pt-H coupling constants, (ii) fairly strong P-H coupling, in particular the long-range coupling J~-HC(23-37 Hz) ; (iii) Jpa-Ho + Jpb-~" (-10 Hz) , Jpb-Hb ( < 3 He) < Jpa-~b (10-20 Hz) , and Jpb-~c ( <3 Hz) << Jpa-~c; arid (iv) JHa-Hb (complexes) < JHa-Hb (free allene) . The 'H NMR spectra revealed that monosubstituted allenes form only one isomer in which the substitucnt occupies site c. 1,3-Disubstituted TABLE VII ALLENECOMPLEXES OF THE NICKELTHAW M PR3 Allene MP("C) Color Ni PPh3 PhCH=C=CHPh 150-152 (dec) Yellow PPh3 (CHa)zC=C=CHz Yellow PhCH=C=CHPh 138-141 (dec) Yelloa Me (CHa)zC=C=CHz 117-120 (dec) Yellow p(o+$3)Me Pd PPhj CHz=C=CHz 83-85 (dec) Colorless PPh3 PhCH=C=CHPh 160-162 (dec) Pale yellow P(OPh)a CHz=C=CHz llO(dec) Colorless P(OPh)3 PhCH=C=CHPh 134-137 (dec) Colorless P(OM43 PhCH=C=CHPh Liquid Pale yellow Pt PPha CHz=C=CHz 152-154 (dec) Colorless PPh3 CHz=C=CHCH, 142-146 (dec) Colorless PPh3 CHz=C=C(CHj)z 130-135 (dec) Colorless PPh, CHz=C=CHPh 170 (dec) Yellow PPh3 PhCH=C=CHPh 170-182 (dec) Yellow PPh3 (CH,)E=C=C(CH3)2 118(dec) Colorless P(OPh)3 CHz=C=CHz 89-91 Pale yellow P(OPh)3 PhCH=C=CHPh 167- 168 Yellow Data from Otsuka et al. (122,123, lS9). Acetylene and Allene Complexes 269 allenes also form only one isomer with the two substituents occupying site c and one of the a positions. The Pt(0) complexes are in gentd rather stable, except for Pt(PPh3)2- (TMA) (TMA = tctramethylallcntx). In solution, the latter shows a sig- nal due to free TMA in addition to those of the complexed ligand, indicat- ing dissociation : Pt(PPh3)2(TRIA) 1’t(PPh3)2 + TLTA A dissociation constant of about>1.5 X (25°C) was assessed from the intensity ratio. The complex docs not show intramolecular fluxional be- havior up to +5O”C. Of the thrw methyl resonances the chemical shift of that at position b occurs at highest field owing to the influence of the phenyl ring current. Consistent with a corollary of the X-ray study (13.2) on Pt(PPh,)z(CHZ=C=CMez), wc may infer the existence of a steric com- pression between the methyl and the phenyl groups which may be primarily responsible for the ready dissociation of TMA. The Pd (0) complexes are rathm labile. Complex Pd (PPh3) (CH-C= CHZ) shows some broadening of the three proton signals even at -74”C, and above 22°C complcte equilibration of the signals to a sharp singlet is observed. Complex Pd ( PPh3) (PhCH=C=CHPh) shows a limiting spectrum of the olefinic protons below -40°C [Ha,6 = 6.04 (m); Hb, 6 = 4.75 (d)]. On raising the temperature, a coalesced signal appears (6 = 5.45) which becomes a vcry sharp singlet (6 = 5.50) above 60°C. The disappearance of the coupling u ith the phosphorous atoms excludes an intramolecular process for the equilibration. An exchange process be- tween the complex and the free allcnc is indicated by broadening of the signal of the free PhCH=C=C‘HPh upon addition of the allene. Since the half-height widths of the two signals before the coalescence are not affected by addition of free PhCH=C=CHPh, a dissociative mechanism is con- cluded : l’d(PPh~)~(PhCH=C=C€II’h) l’d(1’l’hj)~ + PhCH=C=CHPh Addition of free PPh3 to the complex lcads to the formation of Pd(PPh3),. Dissociation of the allene ligarid is also observed for Pd[P (OPh) 312- (PhCH=C=CHPh) and Pd[P (OMe)3]2 (PhCH=C=CHPh) , but dis- sociation of phosphine or phosphite ligands is not observed for any of these Pd(0) compounds. The Ni(0) compounds are in general reactive. It seems impossible to iso- late Ni (PPh3)2( CHz=C=CH2). Complex Ni (PPh,) (PhCH=C=CHPh) is isolable and shows a limiting spectrum for the olefinic protons below -15”; i.c., 6 = 6.15 (Hbcomplex doublet, J = 8.9 Hz) and 6 = 4.17 (Ha multiplet) . At higher temperature (+76”C), coupling with the phosphor- us atoms is lost; two doublets at 6 = 6.07 and 6 = 4.13 (J = 4 Hz) indicate 270 SEI OTSUKA AND AKIRA NAKAMURA only coupling between the allcnic protons, implying an intermolecular phosphorus ligand exchange. The widths of the two signals due to Ha and H* become narrower upon addition of free PPh,. In this case the exchange of phosphine ligands occurs through an associative mechanism : The exchange of Ha and Hb does not occur up to 76°C. The two methyl rrsonanccs of the phosphite ligands in Ni[P (O-o-tolyl) 312 (PhCH=C= CHPh) coalcscc at low temperaturr (22.5"C) retaining the P-H coupling with allenic protons. This indicates an intramolecular exchange of thr phosphite ligands prior to the intermolecular exchange. At higher tempera- tures ( >.50°C) the 1attt.r process is clearly shown by a sharpening of the Ha signal. In summary, (i) PtLz(al1ene) has a rigid planar structurr in the tcm- peraturc range -50" to +50"C and, exccpt for PtLZ(TMA), no dissocia- tion of ligands is obsrrvnblr, (ii) PdLz(allenc.) assumes a similar structure at low temperature although rapid allcne exchange, via a dissociative mechanism, may bc obscrvcld at high temperature, and (iii) NiL, (allcne) assumes a planar structure at low temperature but undergoes a rapid con- figurational change at high temperaturr lrading to equilibration of the two phosphorus ligands. Dissociation of the allene ligand dom not occur before the decomposition of the complex. Addition of excess PPh? causcs rapid ligand cxchangr via an associative mechanism, whereas addition of allene lcads to oligomerization. B. Cyclooligomerizution In general the thermal rcaction of allene gives a complex mixture of dimcrs, trimers, and highrr oligomers including small amounts of spiro compounds (140). A highly selective dimerization to 1,2-dimethylene- cyclobutane is achieved by thermal reaction of dilute solutions (141 ). Theoretically the process [as + 2a] of allencs and related cumulenes may bc facilitated by participation of the orthogonal pr-orbital in the addition (14.2).However, the concertedness of the cyclodimerization is still in dis- pute. Selective cyclotrimerization and pentamcrization of allene have not Acetylene and Allene Complexes 27 1 been achieved thermally, although the concerted processcs are “thermally allowed.” In metal-catalyzed rcactions, allene is cyclizcd selectively to methylcne-substituted compounds ranging from four- to twelve-membered rings. Studies of the mechanism arc of interest both for its own sake and for its pertincncc to gcneral qucstions regarding the role of ligands in transition metal-olefin catalysis. 1. A’iclcel Oligomerization Nickel (0)-allcne complexes arc charact erizcd by configurational insta- bility and a propensity to assume a high coordination number. It may not be surprising to find that the Ni (0) spc.cics is the most catalytically active of the triad. The cyclic dimers, 1 ,3- and 1 ,2-dimcthylcnccyclobutanc, arc formed only in the vapor phase rraction of allene with Ni(C0),(Ph2PC6- H4PPh2) (149). The liquid phase reaction with Ni(0) complexes selec- tively produces the trimer (I),tetramer (11),and pcntamer (111) (Table VIII) (123).Several intermediate Ni(0) complexes (IV-VI) were isolated. (1) (11) (111) Their structures and relative reactivities provide most important informa- tion as to mechanism. The allenc-trimer complex, NIL (CsH,,) (IV) (123, 144,145) is readily obtained as rather stable crystals by treating a mixture of Ni (cod), and a phosphorus ligand (e.g., PPh3, P (OPh)3, or P ( OC~HX- o-Ph),) with allene in solution. The structure was deduced from IR and NMR data and confirmed by a singlc-crystal X-ray diffraction study (145). The intcrniediacy of compound IV (L = PPh3) in the formation of compound I1 is confirmed by monitoring (‘H NMR) thc reaction of IV with allene at 50°C : TABLE VIII CYCLOOLIGOMERIZATIONOF ALLENE WITH NICKEL(O)CATALYSTS IN BENZENE-. Selectivity (%) Reaction Higher Concen- conditions oligomers trationd (t/"C; Conversion Trimere Tetramer Pentamer and Catalyst Added ligandc (mole %) time/hr) (yo) (1 ) (11) (111) polymer [Ni(cod)z]f - 0.20 40, 40 100 Trace Trace 52 48 [Ni(cod)z]f 4PPh3 2.5 70, 20 93 28 66 6 Trace [Ni(cod)z]f PBua" 2.5 .50, 72 100 13 60 18 10 [Ni (cod),]' 2P (OPh)3 0.25 70, 20 15 48 21 Trace 31 [Ni (cod)z]f P(0-0-tolyl) 3 2.5 70, 24 100 38 9 20 33 [Ni(cod)z]f P (0-o-biphenylyl) 3 2.5 70, 24 96 39 11 5 44 [Ni(CsHd (PPhd21 2.5 70, 26 100 20 53 10 17 [Ni(cod) (PBu3")Jf 2.5 60, 48 100 17 83 Trace Trace CNi (P(OPh) d41 2.5 70, 20 81 72 25 3 Trace a Data from Otsuka el al. (1.23). Allene (2.1 gm,52.5 moles); benzene (10 cm9. Added in silu to the reaction mixture at -78°C. d Mole '% of Ni(0) complexes based on allene. c About 5% of 1,3,5-trimethylenecyclohexanewas present. f Cod = cyclooctacl ,5-diene. Acetylene and Allene Complexes 273 Allene-tetramer complexes, NIL( C12H,,) (V) (123) and Ni2L3( C,Hl,) (VI) (Ids),were isolated from the low-temperature treatment of Ni(cod)2 with allene followed by addition of PPh3. These tetramer complexes, being more reactive than complex IV, readily react with allene even at room temperature in benzene to give IF' and tetramer 11: It is noteworthy that the reaction docs not produce pentamer 111. Their structures are deduced from the IR and NMR data and support is obtained from the following reactions: ePPh5+ CS, -40°C e,+ PPh, (VI) J. -4OOC [Ni(C,,H,,) (CS,)] + % [Ni(CS,) (PPh,)], f 2 PPh, 1- 10°C [Ni(PPh$CS,)], + (11) + PPh, Kinetic studies (123) on the three types of oligomerization indicate that the rate of reaction increased in the order Ni(0)-P(OPh), < Ni(0)- PPh3 < Ni(0). Interestingly, the rates for both pentamerization and tetra- merization are first order with respect to the allene concentration, whereas the trimerization rate is nearly zero ordcr. These results are accommodated by the following reaction sequences. 274 SEI OTSUKA AND AKIRA NAKAMURA Trimerization : The reaction of Ni(cod)2 with allene below -30°C produces a mixture of Ni( CgH12) and Ni ( C12H16)in roughly equal amounts, indicating that the potential barriers are quite low up to the trimer and to the tetramer stage in the absence of a phosphorus ligand. At higher temperature the reaction mixture yields pentamer 111. The qualitative reaction potential profile may be deduced as the solid line in Fig. 2. In the presence of 1 mole of PPhx, the low-temperature reaction of Ni (cod) 2 with allene produces exclusively the trimer complex (IV) . At 5OoC, reaction (1) takes place producing tetramer 11. Apparently the role of PPh3 is to facilitate ring closure of the linear tetramer ligand in complex V or VI. These results suggest the potential energy profile depicted (the broken line in Fig. 2). The Ni(0)-phosphite system yields the most stable trinier complex (IV) and, as the kinetics indicate, the unimolecular thermal decomposition of complex IV constitutes the rate-determining step. All attempts to identify the initial stage of the reaction have failed due to the fast ratc even at vcry low tempwaturc. Appropriate substituents on allene are effective in retarding the first few steps, and the monomer com- plexes NiL2(allene) are isolated using both 1 ,3-diphenylallene and 1,l- dimethylallene. A linear dimerization involving hydrogen migration was observed for the reaction of 1,1-dimethylallenc with NiLz (L = 3 Ni*: labile ligands such as cycloocta-l,5-diene are abbreviated. Acetylene and Allene Complexes 275 P-Ni-’C, PNi + CI2 Reaction coordinate FIG.2. Partial reaction potential profiles for cyclooligomerization reactions. P (OCsH1-o-Ph) 5) (149): Not only this Ni(0)-catalyzed reaction but also all reported allene dimer complexes, e.g., hexacarbonyl-p[1-3 : 1’-3’-7- (2,2’-biallyl) ldiiron (Fe-Fe) (146), hexacarbonyl-p-[l-3 : 1’-3’-7- (lfl’-diphenyl-2,2’-biallyl)ldiiron (Fe-Fe) (14?‘),and di-p-acctato-p-[1-3 : 1’-3’-7- (2,2’-biallyl) ldipalladium (148) point to the formation of 2,2’-biallyl. A mononuclear Rh (I) com- plex containing this ligand was rrccmtly isolated (149). Accepting this biallyl formation, then the ncxt stcp is the insertion to form the trimer ligand in complex IV. Thus thc entire reaction paths leading to complexes I, 11, and I11 may be depictcd (Scheme 8). It is now possible to give a rcasonable interpretation for the role of phos- phorus ligands in this catalysis. The stability of complex IV indicates that phosphorus ligands are effmtivc in stabilizing bisallyl coordination to Ni(0) [IV is the first examplc of a stable bisallyl phosphine nickel(0) complex]. The inertness is particularly enhanced with a triarylphosphite, which apparently is due to the c,lcctron-accepting property stabilizing the xero-valency state. Addition of phosphine to bis (allyl)nickel (0) species generally leads to the formation of biallyl, or to insertion reaction if olefin or dime molecules are present in thc system. Thus, the important role of phosphitc ligand here is to prevent further insertion. 276 SEI OTSUKA AND AKlRA NAKAMURA Ni(0) fast c3H4 > slow g; c3*/ fast \Iow [EN:-j (1:T 2C,H, fast t (1) Ni-L Scheme 8 The monomer insertion was assumed to take place via electrophilic attack of allene at the o-allyl-metal bond rather than at the n-ally1 end. In support of this notion is thc preferential formation of Complex V from IV and complex VIII from VII (Scheme 8), indicating that the insertion site is the carbon end of an extended pn-system in complexes IV and VII, where negative-charge localization is enhanced compared to the isolated allyl system. The ligand effect may also be reasonably interpreted assum- ing participation of the anionic end of the allyl group in the insertion. It is known that a-donor ligands stabilize high oxidation states of metals. Here the role of a tertiary phosphine ligand is, contrary to that of electron- withdrawing phosphites, to shift the n-a allyl equilibrium toward the Acetylene and Allene Complexes 277 o-form, requiring an increase in formal oxidation statc of the metal. Elec- trophilic attack of allene is thcn facilitated to give higher oligomers. How- ever, the predominant formation of complex I1 from V and of complex I11 from VIII suggests that hlocltage of a coordination site by a ligand L in complex V is effective in raising the potential barrier for allene coordina- tion (hence the insertion). In addition the free energy of activation for thermal decomposition of complex V should be low as the kinetics indicate. Thus, further insertion of allcnc into V to produce pentamcr I11 becomes a minor reaction path (Tablr VIII) (123) compared to the formation of complex 11. The foregoing interpretation is consistent with the observed relative rate, i.e., pentamerizatiori with nakcd Ni (0) > tetramerization with Ni(0)-PR3 > trimcrization with Ni(O)-P(OR)3. 2. Rhodium Oligomerixation The reaction path of alleiici oligomerization on Rh (I) differs somewhat from that on Ni(0) species. The rnoiiomer complex can be isolated. Com- plex RhCl (PPh,) 3 with a stoichionictric amount of allcne gives RhCl- (PPh3)2(C,H4) (X) (136). Thc homo or iodo analog is prepared from complex X. The structure of the iodo compound has been determined (129). In the absrnce of strongly coordinating substances such as phos- phorus ligands, RhCl speciw take up 5 moles of allcnc to give the cyclo- pentamer complex RhCl(('ljHzo) (XIV) (150). For example, [RhCl- (CzH4)2]2 is a good sourcc for thr production of RhCl species. Lower allene oligomers could not be detectc,d in this reaction. Dimer and tetramer com- plexes havc been obtained with Rh(1) having a chelating anion, e.g. acetylacetonate(acac) (149). Thus thc low-temperature ( - 78°C) reaction of allene with Kh (acac) ( C2H4) prodiiccs unstable Rh (acac) (C3H4) 3 of unknown structure which givcrs, upon treatment with pyridine, very stable Rh (acac)py2 ( C6H8) (XII). Thr structure of complex XI1 has been cstab- lished by an X-ray study. Hcnce the unstable compound is believed to have a rhodacyclopentanc unit (XI). The unstable five-membered ring (XI) is apparently stabilized in complex XI1 which may be regardcd as an octahcdral Rh(II1) (d6) complex. The unstable Rh(acac) (C3H4)I is a pre- cursor of the tctramer complcx Rh(acac) (C12H16) (XIII), which is also directly obtainable from the room-temperature reaction of allerie with Rh (acac) ( C1H4) in pentanc. Thc other p-diketonato complex, Rh (dpd)- 2 78 SEI OTSUKA AND AKIRA NAKAMURA (C12H16) (dpd = 1,3-diphenylpropane-l, 3-dionato) mas also made. The structure of complex XI11 was established by an X-ray study (151). The reaction path from the unstable complex containing (XI) to complex XI11 remains to be elucidated. The structure of pentamer ligand of XIV, de- rivable from the tctramcr ligand of XIII, is different from the nickel pentamer 111. The key step determining the pentamer structure is thus the tetramer stage. Complex XI11 is best described as an Rh (111) complex. The Rh(II1) ion apparently prefers coordination of the allylic group con- jugated with a double bond so that the negative charge localization at the ally1 end matches with the high metal oxidation state. Rhodium (I)-phosphine systems lead to catalytic tetramerization. For example, the system [RhC1(C2Ha)2]2 with 1 to 2 moles of PPh3 is effective in the selective formation of an interesting spiro compound (XV) (152) free from other isomers. Although the detailed reaction path is unknown due to the inaccessibility of the intermediate complexes, the formation of (XV) may be visualized from a tetramer complex as follows: Acetylene and Allene Complexes 279 REFERENCES 1. Bowden, F. L., and Lever, A. B. P., Organometal. Chein. Rev. 3, 227 (1968). 2. Kemmitt, R. D. W., MI‘P Znt. Ref).Sci., Znorg. Chein., Ser. 1 6 (part 2), 227 (1972). 3. Hiibel, W., in “Organic Syntheses via RIetal Carbonyls” (I. Wender and P. Pino, eds.), Vol. 1, p. 273. Wiley (Interscience), New York, 1968. 4. Hartley, F. 11., Chenz. Reu. 69, 799 (1969); 73, 163 (1973); Angew. Chenz. 84, 667 (1972); “Chemistry of Platiiiuni and Palladium,” p. 361. Applied Science Publ., London, 1973. 5. Heck, It. F., “Organotransition XIetaI Chemistry,” pp. 167-200. Academic Press, New York, 1974. 6. Green, Pvl. L. I%, iii “0rganomet:tllic Compounds” (G. E. Coates, K. Wade, and M. L. H. Green, eds.), Vol. 2, pp. 288-333. Methuen, London, 1968. 7. Rlaitlis, P. M., “Organic Chemistry of Palladium,” Vol. 1, pp. 110-130; Vol. 2, pp. 31, 47-58. Academic Press, New York, 1971. 8. Jolly, P. W., and Wilke, G., “0rg:rnic Chemistry of Nickel,” Vol. 1, pp. 305-315. Academic Press, New York, 1974. 9. Pettit, L. l)., and Barnes, D. S., Fortschr. Chem. Forsch. 28, 85 (1972). 10. I)ickson, It. S., and Fraser, P. .J., L4duai~.Organometal. Chenz. 12, 323 (1973). 11. Shaw, B. L., and Stringer, A. ,J., Inorg. Chim. Acta Rev. 7, 1 (1973). 12. Baker, lt., Chem. Reu. 73, 487 (1973). 13. Jolly, P. W., and Wilke, G., “The Organic Chemistry of Nickel,” Vol. 2. Academic Press, New York, (1975). 14. Nelson, J. H., Wheelock, K. S., Cusachs, L. C., and Jonassen, H. B., J. A,mer. Cheni. SOC.91, 7005 (1969); Inorg. Cheni. 11, 422 (1972); Wheelock, K. S., Nelson, J. II., Cusachs, L. C., and Jonassen, H. B., J. Amer. Chem. Soc. 92, 5110 (1970); Nelson, J. I%., and Jonassen, 11. I<., Coord. Chem. Rev. 6, 27 (1971). 15. Bokiy, N. G., Gatilov, Yu. V., Struchkov, Yu. T., and Ustynynk, N. A., J. Organometal Chein. 54, 213 (1973). 16. Nesmeyanov, A. N., Gusev, A. I., Pnsynskii, A. A., Anisimov, K. N., Kolobova, N. E., and Slrachkov, Yu. T., C‘hem. Commm. 277 (1969). 17. Bennett, M. A., Robertson, G. B., Whimp, P. O., and Yoshida, T., J. Amer. Chem. Soc. 93, 3797 (1971); Robertson, (;. U., arid Whimp, P. O., J. Organometal. Chem. 32, C69 (1971). 18. Davies, B. W., and Payne, N. C., Inorg. Chem. 13, 1848 (1974). 19. Glanville, J. O., Stewart, J. &I.,and Grim, S. O., J. Organometal. Chem. 7, P9 (1967). 20. I’anattoni, (>.,and Graziarri, It., it,” I’rogrrss ill Coordination Chemistry” (%i.Cttis, ed.), p. 310. Elsevier, London, lO(i8. 21. Ilavies, B. W., Puddephatt, li. ,J., and Pnyne, N. C., Can. J. Chem. 50, 2276 (1972). 22. Davies, B. W., and Payne, N. C., Inorg. Chern. 13, 1843 (1974). 23. Davies, G. It., Hewertson, W., hhis, 11. I€. B., Owston, P. G., and Patel, C. G., J. Chem. Soc. A 1873 (1970). 24. McGinnety, J. A., J. Chem. Soc., Dalton Trans. 1038 (1974). 25. Jacobson, S., Carty, A. J., Mathieu, &.I., and Palenik, G. J., J. Amer. Chem. SOC. 96, 4330 (1974). 26. Dickson, It. S., and lbers, J. A., J. Organometal. Chem. 36, 191 (1972). 27. Nesmeyanov, A. N., Gusev, A. I., Pasynskii, A. A,, Anisimov, K. N., Kolobova, N. E., and Struchkov, Yu. T., Chein. Coinmun. 739 (1969). 28. Kirchner, R. M., and Ibers, J. A., J. Amer. Chem. Soc. 95, 1095 (1973). 29. Laine, R. M., Moriarty, It. E., and Bau, li., J. Amer. Chem. SOC.94, 1402 (197’2) 30. Davidson, G., Organometal. Chem. Rev. 8, 342 (1972): see Maitlis (Y), Vol. I, pp. 120-121. SEI OTSUKA AND AKIRA NAKAMURA 31. Wilke, G., and Herrmann, G., Angew. Chem. 74, 693 (1962). 32. McClure, G. L., and Baddley, W. IT., J. OrganometaE. Chem. 27, 155 (1971). 33. Baddley, W. II., Znorg. Chem. Acta Reu. 2, 1 (1968). 34. Otsuka, S., Yoshida, T., and Tatsuno, Y., J. Amer. Chem. SOC.93, 6462 (1971). 35. Nakamura, A., and Otsuka, S., J. Amer. Chem. SOC.94, 1886 (1972). 36. Brintzinger, H. EL, and Thomas, J. L., J. Amer. Chem. Soc. 96,3694 (1974); Thomas, J. L., ibid. 95, 1838 (1973). 37. Tsumura, R., and Hagihara, N., Bull. Chem. SOC.Jap. 38, 861 (1965). 38. Mann, B. E., Shaw, B. I.., and Tucker, N. I., J. Chem. SOC.A 2667 (1971). 39. Church, M. J., Mays, M. J., Simpson, It. N. F., and Stefanini, F. D. S., J. Chem. Soc. A 2909 (1970). 40. Cherwirski, W. J., Johnson, B. F. G., and Lewis, J., J. Chem. SOC.,Dalton Trans. 1405 (1974). 41. Chisholm, M. H., and Clark, H. C., Znorg. Chem. 10,2557 (1971); Kong, P. C., and Theophanides, T., Can. .I. Chem. 45, 3193 (1967); 48, 1084 (1970). 42. Green, M., MTP Int. Rev. Sci., Znorg. Chem., Ser. 1 6, 198 (1972); Heck (5),p. 168. 43. Schwartz, J., Hart, D. W., and Holden, J. L., J. -4mer. Chem. SOC.94,9269 (1972). 44. Treichel, P. M., Pitcher, E., and Stone, F. G. A., Znorg. Chem. 1, 511 (1962). 45. Wilford, J. B., and Stone, F. G. A,, Znorg. Chem. 4, 93 (1965). 46. Nakamura, A., and Otsuka, S., J. Amer. Chem. SOC.94, 1886 (1972). 47. Nakamura, A., and Otsuka, S., “Prospects in Organotransition Metal Chemistry” (hl. Tsutsui and Y. Ishii, eds.), Plenum Press, New York, 1975; J. Mol. Calal. To be published. 48. Clark, H. C., and Tsang, W. S., J. Amer. Chem. SOC.89, 533 (1967). 49. Clark, H. C., and Mittal, R. K., Can. J. Chem. 51, 1511 (1973). 50. Blackmore, T., Bruce, pvl. I., Stone, F. G. A,, Davis, R. E., and Garza, A., Chem. Commun. 852 (1971); J. Chem. Sac., Dalton Trans. 106 (1974). 51. Booth, B. L., and Lloyd, A. I)., J. Organometal. Chem. 35, 195 (1972). 52. Bruce, M. I., Harbourne, D. A., Waugh, F., and Stone, F. G. A., J. Chem. SOC. 895 (1968). 53. Harbourne, D. A., and Stone, F. G. A,, J. Chem. SOC.A 1765 (1968). 54. Trocha-Grimshaw, M. J., and Ilenbest, €1. B., Chem. Commun. 757 (1968). 55. Eaborn, C., Farrell, N., and Pidcock, A., Chem. Commun. 766 (1973). 56. Schrauzer, G. N., and Windgassen, R. J., J. Amer. Chem. SOC.89, 1999 (1967). 57. Yamazaki, H., and Hagihara, N., J. Organometal. Chem. 21, 431 (1970). 58. Booth, B. L., and Hargreaves, G., J. Chem. SOC.A 2766 (1969). 59. Dubeck, M., and Schell, R. A., Znorg. Chem. 3, 1757 (1964). 60. Hagihara, N., et al., unpublished result. 61. Booth, B. L., and Hargreaves, G., J. Organometal. Chem. 33, 365 (1971). 62. Baddley, W. H., and Fraser, M. S., J. Amer. Chem. SOC.91, 3661 (1969); J. Organometal. Chem. 36, 377 (1972). 63. Wailes, P. C., Weigold, H., and Bell, A. P., J. Organometal. Chem. 43, C29 (1972). 64. Baddley, W. H., and Tupper, G. R., J. Organometal. Chem. 67, C16 (1974). 65. Zanella, R., Canziani, F., Ros, R., and Graziani, M., J. Organometal. Chem. 67, 449 (1974). 66. Cullen, W. R., Dowson, D. S., and Styan, G. E., Can. J. Chem. 43,3365 (1965). 67. Truce, W. E., and Goldhamer, D. L., J. Amer. Chem. SOC.81, 5795, 5798 (1959). 68. Nakamura, A,, and Otsuka, S., J. Amer. Chem. SOC.95, 7262 (1973). 69. Epiotis, N. D., J. Amer. Chem. SOC.95,1191 (1973); Pearson, R. G., Accounts Chem. Res. 4, 125 (1971). Acetylene and Allene Complexes 28 1 70. Hoogzand, C., and Hiibel, W., “Organic Syntheses via Metal Carbonyls” (I. Wen- der, and P. Pino, eds.), Vol. 1, p. 343. Wiley, New York, 1968. 71. Yamazaki, H., and Hagihara, N., J. Organometal. Chem. 7, P22 (1967). 72. Zeiss, H. I%, and Tsutsui, M., J. Amer. Chem. Sac. 81, 6090 (1959). 73. Burt, R., Cooke, M., and Green, M., J. Chem. SOC.A 2981 (1970). 74. Sears, C. J., Jr., and Stone, F. G. A., J. Organometal. Chem. 11, 644 (1968). 75. Collman, J. P., and Kang, J. W., J. Amer. Chem. Sac. 89, 844 (1967). 76. Brateman, L. R., Maitlis, P. M., and Dahl, L. F., J. Amer. Chem. Sac. 91, 7292 (1969). 77. Kang, J. W., Childs, R. F., and Maitlis, P. M., J. Amer. Chem. Sac. 92, 721 (1970). 78. Clarke, B., Green, M., and Stone, F. G. A., J. Chem. SOC.A 951 (1970). 79. Mague, J. T., and Wilkinson, G., Znorg. Chem. 7, 542 (1968). 80. Collman, J. P., Cawse, J. N., and Kang, J. W., Inorg. Chem. 8, 2574 (1969). 81. Mague, J. T., Znorg. Chem. 9, 1610 (1970); Mague, J. T., Nutt, M. 0. and Gause, E. H., J. Chem. Soc., Dalton Trans. 2578 (1973). 82. McVey, S., and Maitlis, P. M., J. Organometal. Chem. 19, 169 (1969); Kang, J. W., McVey, S., and Maitlis, P. M., Can. J. Chem. 46, 3189 (1968). 83. Moseley, K., and Maitlis, P. M., J. Chern. Sac., Dalton Trans. 169 (1974). 84. Ito, Ts., Hasegawa, S., Takahashi, T., and Ishii, Y., Chem. Commun. 629 (1972); Roe, I>. M., Calvo, C., Krishnalnnchari, M., Moseley, K., and Maitlis, P. M., J. Chem. Soc., Chem. Commun. 436 (1973). 85. Ashley-Smith, J., Green, M., and Wood, 1). C., J. Cheni. Sac. A 1847 (1970). 86. Collman, J. P., Kang, J. W., Little, W. F., and Sullivan, M. F., Irwrg. Chem. 7, 1298 (1968). 87. Singer, H., and Wilkinson, G., J. Chem. SOC.A 849 (1968). 88. Kern, R. J., Chem. Commun. 706 (1968). 89. Clemens, J., Green, M., Kuo, RI. C., Fritchie, C. J., Mague, J. T., and Stone, F. G. A., Chem. Commun. 53 (1972). 90. Otsuka, S., Nakamura, A,, Tatsuno, Y., and Miki, M., J. Amer. Chem. SOC.94, 3761 (1972). 91. Wakatsuki, Y., and Yamazaki, H., Tetrahedron Lett. 4549 (1974). 92. Blackmore, T., Bruce, M. I., Stone, F. G. A,, Davis, It. E., and Gartza, A., Chem. Commun. 852 (1971). 93. Chisholm, R.I. H., and Clark, H. C., J. Amer. Chem. SOC.94,1532 (1972); Chisholm, M. H., Clark, H. C., and Hunt,er., 1). II., Chem. Commun. 809 (1971). 94. Tsumura, R., and Hagihara, N., Ui~11.Chem. SOC.Jap. 38, 1901 (1965). 95. King, R. B. and Fronzaglia, A,, Cheni. Commun. 547 (1965). 96. Tate, D. P., Augl, J. M., Ritchey, W. M., Ross, B. L., and Grosselli, J. G., J.Amer. Chem. SOC.86, 3261. 97. Strohmeier, W. and von Hobe, I)., Z. Naturforsch. B 19, 959 (1964). 98. Browning, J., Green, M., Perifold, B. It., Spencer, J. L., and Stone, F. G. A., J. Chem. Soc., Chem. Commrcn. 31 (1973); J. Chem. Sac., Dalton Trans. 97 (1974). 99. Baddley, W. H., Chem. Comnaun. 762 (1972). 100. Greatrex, It., Greenwood, N. N., and Pauson, P. L., J. Organometal. Chem. 13, 533 (1968). 101. Kaska, W. C., and Kimball, M. E., Inorg. Nucl. Chem,. Lett. 4, 719 (1967). 102. See Heck (b), pp. 238-255. 103. Yamazaki, H., Aoki, K., Yamamoto, Y., and Wakatsuki, Y., Abstr. 22nd Symp. Oignnonzetal. Chenr. Jnp., p. 20813 (1974); J. Amer. Chem. Soc. 97, 3546 (1975). 104. Hubert,, A. J., and Dale, J., J. Chem. SOC.3160 (1965). 282 SEI OTSUKA AND AKIRA NAKAMURA 105. Schrauzer, G. N., Aduan. Organometal. Chem. 2, 1 (1964). 106. Bird, C. W., “Transition Metal Intermediates in Organic Syntheses,” pp. 1-29. Academic Press, New York, 1967. 107. lleppe, W., Schlichting, O., Klager, K., and Toepel, T., Ann. Chem. 560, 1 (1948). 108. SchrBder, G., “Cyclooctaletraen,” pp. 12-15. Verlag Chemie, Weinheim, 1965. 109. Cope, A. C., and Fenton, J. Amer. Chem. Soc. 73, 1195 (1951). 110. Craig, L. E., and Larrabee, L., J. Amer. Chem. Soc. 73, 1191 (1951). 111. Naknmura, A., Mem. Inst. Sci. Ind. Res., Osnka Univ. 19, 81 (1962); Chem. Abstr. 59, 8786 (1963). 112. Meriwether, L. S., Colthup, €3. C., Kennerly, G. W., and Iteusch, 11. N., J. Org. Chem. 26, 5155 (1961); Meriwether, L. S., Leto, &I. F., Colthup, E. C., and Kennerly, G. W., J. Orq. Chem. 27, 3930 (1962). 113. Brown, C. K., Georgion, I)., and Wilkirison, G., J. Chem. SOC.A 3120 (1971). 114. Tohdn, Y., Sonogashira, K., and Hagihara, N., unpublished result. 115. Tohda, Y., Sonogashira, K., and Hagihara, N., J. Chem. SOC.,Chem. Commun. 54 (1975). 116. Suzuki, Y., and Takixawa, T., J. Chem. SOC.,Chem. Commun. 837 (1972). 117. See Bird (106), pp. 174-1‘31; Thompson, D. T., and Whyman, It., in “Transition Metals in Homogeneous Catalysis” ((2. N. Ychrauxer, ed.), p. 147. Dekker, New York, 1971. 118. Jaut.elat, M., and Ley, K., Synthesis 593 (1970); Otsuka, S., Nakamura, A., and Yamagata, T., presented at Symp. Org. Synthetic Chenz., 1971. 119. Yamaxaki, II., and Wakatsuki, Y., Tetrahedron Lett. 3383 (1973). 120. Uonrieniann, H., Angew. Chem. 85, 1024 (1973); Bhnnernann, H., Brinkmann, R., and Schenkluhn, H. Synthesis 575 (1074). 121. Wakatsuki, Y., Yamazaki, H., and Iwasaki, H., J. Amer. Chem. Soc. 95, 5781 (1973). 122. Otsuka, S., Nakamura, L4.,and Tani, K., J. Organometal. Chem. 14, P30 (1968). 123. Otsuka, S., Tani, K., and Yamagata, T., J. Chem. SOC.,Dalton Trans. 2491 (1973). 124. Lichtenberg, 1). W., and Wojcicki, A,, J. Amr. Chem. Soc. 94, 8271 (1972). 125. Benaim, J., Merour, J. Y., and lioustan, J. L., Compt. Red. Acad. Sci., Ser. C 272, 789 (1972). 126. Clark, I-I. C., arid hlanzer, L. E., J. Amer. Chem. Soc. 95,3812 (1973); Inorg. Chem. 13, 7996 (1974). 127. Itacanelli, P., Psntini, G., Immirzi, A,, Allegra, G., and Porri, L., Chem. Commun. 361 (1969). 128. Hewitt, T. G., and De Bocr, J. J., J. Chem. Soc. A 817 (1971). 129. Kashiwagi, T., Yasuoka, N., Kasai, N., and Kakudo, hf., Chem. Commun. 361 (1969); Technology Rep. Osaka Univ. 24, 3.55 (1074). 130. Kadonaga, ll., Yasuoka, S.,and Kasai, S.,Chem. Commtn. 1597 (1971); Kashi- wagi, T., Yasuoka, X., Kasui, S.,and Iiakudo, >I., 2’whnol. Rep., Osaka I’nizl. 24, 1188 (1974). 131. Okamoto, K., Yasuoka, N., and Kasai, N., unpublishcd. 132. Yasuoka K.,RPorita, hl., Kai, Y., and Kasai, N., J, Organotrcctal, Chem. 90, 111 (19751, 133. Okamoto, K., Kai, Y., Yasuoka, N., arid Kasai, N., J. Organometal. Chem. 65, 427 (1974). 134. Mason, It., and Rae, A. I. hl., J, Chem, Soc. d 1767 (1970). 135. Kashiwagi, T., Yasuoka, N., Ueki, T., Kasai, N., Kakudo, M., Takahashi, S., and Hagihara, N., Bull. Chem. Soc. Jap. 41, 296 (1968). Acetylene and Allene Complexes 283 136. Otsuka, S., Nakamura, A., nntl Tani, K., Kogyo KagaAu Zasshi (J. Chem. Soc. Jap., Ind. Chern. Sect.) 70, 2007 (1967). 137. Vrieze, K., Volger, H. C., Gronert, PIT., and Praat, A. P., J. Organometal. Chem. 16, P19 (1969). 138. Vrieze, K., Volger, II. C., and Praat, -4.P., J. Organornefal. Chevz. 21, 467 (1970). 139. Otsuka, S., and Tani, K., unpublished. 140. For cxaniplc see Weinstein, B., atrd Fcnsclau, A. II., J. Chem. SOC.368 (1967); J. Org. Chew. 32, 2278, 2988 (1067); Fischer, H., “The Chemistry of Alkenes” (S. Patni, ed.), p. 1025. Interscience, New York, 1964. 141. Dolhier, W. It., ,Jr., and Ilai, S.-II., J. Amer. Chem. Sac. 92, 1774 (1970). 142. Woodward, It. B., and Hoffmanii, It., “The Conservation of Orbital Symmetry,” pp. 163-166. Academic Press, New York, 1970. 143. Hoover, F. W., and Lindscy, li. V., J. Org. Chem. 34, 3039 (1969). 144. Otsuka, S., Nakamura, A,, Uetla, S.,and Tani, K., Chern. Comnzun. 863 (1971). 145. Englert, M., Jolly, P. W., arid Wilke, G., Angew. Chern. 83, 84 (1971); ibid. 84, 120 (1972). 146. Nakamurz, A,, Hull. Chem. SOC.,Jnp. 39, 543 (1966); Xakamura, A., and Hagi- tiara, N., J. Organotnetal. Cheru. 3, 480 (1965). 147. Otsuka, S., Nakamura, A,, and Tnni, K., J. Chem. Sac. A 2248 (1968). 148. Hughes, 11. P., and Powell, J., J. Organotnetal. Chem. 20, P17 (1969). 149. Ingrosso, G., Immirzi, A,, and Porri, I,., J. Organometul. Chem 60, C35 (1973). 150. Otsuka, S., Tani, K., and Nakamura, A., J. Chem. SOC.A 1404 (1969). 151. Pantini, G., Itacanelli, P., Immiwi, A,, and Porri, L., J. Organometal. Chern. 33, C17 (1971). 152. Otsuka, S., Nakamura, A., and hIinamida, H., Chem. Commun. 191 (1969). This Page Intentionally Left Blank High Nuclearity Metal Carbonyl Clusters P. CHINI, G. LONGONI, and V. G. ALBANO lsfifufo di Chimico Generole dell'Univerrif6 Milono, lfoly I. Introduction . . 285 11. Structural Data in the Solid State . . 286 111. Structural Data in Solution . . 306 IV. Syntheses . . . 311 V. Methods of Separation . . 316 VI. Reactivity . . 317 A. Reduction . . 319 B. Oxidation . . 320 C. Ligand Substitution . . 322 D. Oxidative Addition . . 322 VII. Iron Derivatives . . 323 VIII. Ruthenium Derivatives . . 324 IX. Osmium Derivatives . . . 325 X. Cobalt Derivatives . . 325 XI. Rhodium Derivatives, . . 327 XII. Iridium Derivatives . . . 332 XIII. Nickel Derivatives . . 333 XIV. Platinum Derivatives. . . 334 XV. Bonding Theories . . 336 References . . 341 I INTRODUCTION In the last 5 years at least eight reviem concerning polynuclear metal carbonyls have been publishcd (1G,18,34,4l , 86, 87, 91, 108),and it may appear unlikely that any new niaterial could be added at this time. Never- theless, we have willingly acccytcd thc. invitation of the Editors to present a comprehensive review which, besidc the material published up to June 1974, also summarizes the most, rclevant of our recent results, as yet only published as preliminary notcs. In order to present a fresh view, we have confined ourselves to compounds containing 5 or more metal atoms. The slow rate of publication in this area is mainly due to the number of steps required by this research, i.e., synthesis, crystallization, structural identification, and chemical char- acterization. 205 286 P. CHINI, G. LONGONI, AND V. G. ALBANO In 1943, Hieber and Lagally reported that the reaction of anhydrous rhodium trichloridc with carbon monoxide at SO'C, under pressure, and in the presence of silver and copper as halogen acccptors, gavc a black crystal- line product which, on the basis of clcnieiital analysis, was formulated as Rh,(CO)11 (75). Thc exact nature of this compound was established 20 years latcr by Dahl using three-dimensional X-ray analysis which led to its reformulation as Iih6(CO)16(53). This discovery can bc rcgardcd as the birthday of the chemistry of high riuclearity clusters. In 1962, Dahl had also structurally characterized Fez(CO) 15C,the first high riuclcarity carbide ( 26). This compound was originally prepared in cxtrcmcly low yields (0.57,) by the reaction of FC3(C0)12 with substituted acetylrnes, and, probably dur to the pcwliarity of this synthesis, was considered for some hie much more a curiosity rather than bring recog- nized as the precursor of today's large fanlily of carbidc-carbonyl clusters. Finally the hexanuclcar diariion [C'O~(CO) ,512- was the first anionic high nuclearity cluster to bc isolated (33).Its discovery in 1967 prompted cx- tension of such investigations to other transition mrtals and originated the present chemistry of thc high nuclearity anionic clustcrs. More than fifty different examples of high nuclcarity carbonyl clusters (HNCC) arc prcscntly kiiou-n, all of which contain Group VIII transition metals (Table I). In post-transition metals thc incrcased separation be- tween the (n - 1)cl and ns-np orbitals is probably responsible for thc low stability of their bonds with the highly a-acidic carbon monoxidc ligand; a number of high nuclrarity clusters with lcss a-acidic ligands, such as tertiary phosphines (17, 50) or organic donor groups (51, 72),is known, however. Approximate calculations on some of the more crowded clustcrs, such as [Fe, (CO) 13]'-, Fe6('20) l;C, and Ilus(CO) 1BH2, show that, at the Icvc.1 of the carbon atoms, about 96'g of the available surface is occupied (95). This figure seems very high particularly if one takes into consideration that thc distribution of the carbonyl groups is not homogeneous. The high nuclcarity clusters of the transition metals that precede Group VIII are, thercfore, expected to be destabilized by steric crowding, although some carbides and mixed nitrosyl-carbonyl derivatives should bc sterically possible. II STRUCTURAL DATA IN THE SOLID STATE The main bond distances found in HNCC are reported in Table 11, which has been divided into three sections, corresponding to the three High Nuclearity Metal Carbonyl Clusters 287 transition pcriods, in order to niakc a comparison of the data easier. Such comparison should be done with sonir caution since the data reported are mean values and since there arc' thc usual uncertainties due to discrepancies among sterically equivalent intcractions (packing cffects, thermal motion, and crystal disorder) . Moreovcr, bccausc thc expanded orbitals of the low- valcrit transition metals can suff'er considerable dimcnsional variation on changing the electron density at thc m&l atoms, only large differences, or comparison between strictly rclated species, are meaningful. TABLE I1 BONDDISTANCES IN HIGH NUCLEIRITYMETAL CARBONYL CLUSTERS (A) ac-o(8) W Idealized and n crystallographic Edge Face Edge Face E Nuclearity Formula symmetry ~AI-V(Ax) Teririinai bridging bridging Terminal bridging bridging Ref. 5 0 (a) Iron, cobalt, and nickel 5 5 2.64 1.76 - 1.17 - - 26 0 6 2.67 1.70 - 1.18 - - 49 6 2.51 1.74 1.90 1.15 1.17 1.19 2 6 1.21 3 > 2.50 1.70 - 1.17 - Z 6 2.50 1.76 - 1.13 - 1.15 8 CJ 8 1.72 1.94 1.16 1.18 - IS .< 5 (1. 76a 1.84 (1.16“ 1. 13a - 95 \1.86b \1.04b 0 5 1.s9a 1.91n 1.01a 1.07O 13. loc 5 2.34a 1.85~ 1.82~ 1.050 1.14a 3.10c 6 2.38d 1.75 1.90 1.13 1.17 - 28 2.770 9 2. 44d - - - - - 95 2.70‘ (b) Ruthenium and rhodium 6 cz 2.89 - 115 105 6 CS 2.88 - 6 Ci 2.91 1.90 - 1.14 - - 48 1.16 1.20 65 6 Cz 2.78 1.86 2.17 - 1.15 1.19 6 6 - 2.75 1.85 2.17 - 7 C. 2.76 ------4 2.77, 2.93 1.82 2.00 2.19 1.17 1.14 1.17 11 3 7 - ln 1.17 1.19 5 12 Ci 2.79 1.87 2.00 2.19 1.15 J 15 13 - 2.80 - - 1.13 1.15 7 f 6 Cz 2.79 1.89 2.09 - - 2.81 1.91 2.23 1.13 1.15 1.15 12 E 8 - 2.06 13 2.80 1.87 2.08 2.18 1.09 1.15 1.20 :. 12 - 1.12 1.17 9 7 15 CZ 2.87 1.89 2.02 - - 3 nz - 2.80 - - 2.86 - cz 2. 66d 1.77 2.03 3.04' D3 cz 2.66d 1.79 2.00 3.Ose CB - 2. 66d 1.80 2.03 3.08' a Values in the equatorial fragment. i, i, Values for the apical groups. c M-Ni values. Intratriangular values. Intertriangular values. - 290 P. CHINI, G. LONGONI, AND V. G. ALBANO TABLE I11 CALCULtTED VALUES FOR TEIE c \VITY OF REGULARPOLYHEDRA Calcu1at:d hole Comparison with (in A) some cov@nt radii Polyhedron” Circumradius for d = 1.4 A (11) Tetrahedron I.225d 0.315 H = 0.37 Square pyramid 1.414d 0.58 Octahedron 1.414d 0.58 Trigonal prism 1.523d 0.733 C = 0.77 Square antiprism 1.64d 0.90 Triangular dodecahedron 1.701d 0.9s Cube 1.732d 1.02 Si = 1.17 a Regular polyhedron edge = 2d. b Average value for Rh, the central element of Group VIII. A peculiarity of the three-dimensional clusters, containing more-or-less regular polyhedra of metal atoms, is the existence of a central cavity whose dimensions, as shown in Table 111, are a function of the particular gcom- etry of the polyhedron. The existence of such a hole is confirmed by the formation of a large number of carbide derivatives. The known structures of pentariuclear HKCC are based both on the square pyramid and on the trigonal bipyraniid, and are illustrated in Fig. 1. The square pyramid of iron atoms in Fcj(CO) (26) is essentially rrgular and corresponds to one-half of an octahedron. By contrast, the trigonal bipyramids found in the bis( triphenylphosphino) iminium (PPXj salts of anions [n’i5(CO)12]2- (95) and [R42Ki,(CO)16]2- (M = &lo, W) (111) arc not only elongated along the threefold axis but also show significant diff erenccs in the distances betwecn thc central Ni, (CO)a(p?-CO) 3 triangle, of D3hsynimetry and thc apical groups. These diffcrcnces arc on avzrage 0.03 8 for apical Ni((’O), groups of CJulocal symmetry, and 0.1 A for apical Mo (CO) 5 groups of C1, local symmctry. In the ccntral triangle, all the Xi-(2 distanccs arc comparativc>ly long, and the C--0 distances cor- respondingly short, indicating predoniinantly dative bonding. Deformation and other bonding peculiarities arc often found in HPU’CC, and our present ability mcwly to describe most of these phenomena gives an idea of the present, state of thc theory. Prcliminary data 011 the salt [NEt&Mo2Nil( CO) iridicatcl a struc- ture with DShsymmctry bastid on an cquatorial Nii (CO) rhonibus with two Mo(COj4 groups placed above and below the plane (111). It is prob- able that thcre arc only tcrniinal carbonyl groups. High Nuclearity Metal Carbonyl Clusters 29 1 Q 0 0 FIG.1. Schematic molecular structrires of the pentanuclear high nuclearity metal carbonyl clusters and of OS~(CO)~S. As shown by the structurm so far discussed, and as previously pointed out by King (go),a triangular network of metal atoms is the most common basic unit in transition metal clust(m, suggesting that bonding between metal atoms in a triangular network is not confincd to the edges but can also occur within the trianglc itself. This hypothesis is reasonable bccause some orbital overlap could still occur at the center of the triangles (1.155 times the metallic radius), whereas there can be little bonding interaction 292 P. CHINI, G. LONGONI, AND V. G. ALBANO FIG.2. Schematic molecular structures of some high nuclearity metal carbonyl clusters based on octahedra of metal atom (section a). through the ccnter of a square, where the vertex-center distance is 1.414 times the metallic radius. A good example of the tendency toward triangulated polyhedra is given by the bicapped tetrahedron present in OS~(CO)~~(104). In this cluster (complex IV in Fig. l),each of the two osmium atoms shared by the basal triangles is directly bonded to 5 other osmium atoms. In spite of theoretical considerations (go),the presence of pentaconnected vertices is common in HNCC, indicating bonds of metallic type and implying considerable de- localization. The 0s-0s distances orthogonal to the symmetry axis, be- tween the atoms of greater metallic character (tetra- and pentazonnected vertices), are considerably shorter (2.74 8 as compared to 2.84 A). High Nuclearity Metal Carbonyl Clusters 293 b d FIG.3. Schematic representation of t,he molecular structure of the dianion [Rhn(CO) 301~. The octahedron, another triangulated polyhedron, is the most common type in HNCC; about half of the compounds reported in Table I contain octahedra or deformed octahedra of metal atoms. The structures of some octahedral HNCC are reported in Figs. 2-4. The high symmetry of octahedral RuG(CO)~,H,is illustrated in Fig. 2 by structure V (48).Tricoordinatiori toward carbon monoxide is maintained in Rus(CO)17C, structure VI (llS), and in the analogous RUG(CO)14 (nirsitylene) C (103) by formation of a carbonyl bridge. The same tend- ency is also evident in structurc VII, which has been found in the anion [FeG(CO) &]'- (49),although hcrc the bridges are very unsymmetrical and one of the iron atoms is bondcd to 4 carbon monoxides. All of the other octahedral clusters bearing sixtc~mcarbonyl groups, and also some of their derivatives, have the much more syinmctrical structurc VIII. Typical examples of this stereochemistry are RhG(CO)16(53) and the anion [Rhs(CO)i,I]- (6). Comparison among the octahedral structures VI-XI (Figs. 2-4) shows that the number of CO ligarids coordinated to each metal atom increases from three to five. This indicates that the metal-ligand and metal-metal systems of bonds are largely independent, a fact which is of considerable theoretical interest (see Section XV) . The formal coordination numbers of the metal atoms in these clustcrs 294 P. CHIN!, G. LONGONI, AND V. G. ALBANO Q PP n 0 CNi6(C0)6(pz-CO163Z~ (C3vl (XIII FIG.4. Schematic molecular structures of some high nuclearity metal carbonyl clusters based on octahedra of metal atom (section b). are not only variable but are also unusually high (34, 106). However, their stereochemical significance cannot be compared with that usually accepted in simple compounds, because in HNCC part of the bonds are metallic in character and cannot be represented as simple clectron pair bonds. The related unsymmetrical distribution of this high number of bonds along the directions of the quaternary axes of the octahedra is readily High Nuclearity Metal Carbonyl Clusters 2 95 apparent in structurc VIII, nhcw thcrc arc two bonds (terminal CO’s) pointing out from the cluster and six bonds (two bridging CO’s and four mctal-metal bonds) pointing ton ard thc cluster. In structure XI the cor- responding distribution is 1: 8. Thcsc unsymmetrical distributions of the bonds obviously require the prtwncc. of some counterbalancing electron density in the proper directions. The bond distances in specios llhG((’O)16and [Rh6(CO)ljl]- are very similar, the only exception bciiig the terminal CO group accompanying thc iodide ligand. The lth-C distance is 0.02 LL greater, and C-0 length 0.08 8 shorter, than in the remailling invtal carbonyl fragments, in agrcc- merit nith a locally lolwr backdotration. Although the dioctahcdral anion [llhls( CO) 30]2-, structure IX of Fig. 3, exhibits mean values for the I)ontlitig interactions similar to those found in the two preceding species, an c.xatnination of the individual distances shows considerable deformations (5). For example, pcrpcndicular to the twofold axis joining thc two tnoivtic~s,the equatorial planes of t,he octa- hedra are rectangularly dcformcd nith mean edges 2.68 and 2.54 A. Thcsc deformations haw been tentativ~ly~xplained in several ways (5, 40). The infrared spcctra of the solids suggest that Co6(CO) 16 and IrG(CO)16 arc isostructural with lthG(Co)Ib (36, 96) , and for the former compound this hypothesis is confirmed by thc isomorphism of the crystals (I). The same structure is probably prcwnt it1 the dianion [Hh,(CO) 1,(CN)J2- (45). Tetracoordinatioii to\\ ard carbon monoxide is niaintained in the [co,(c0)1,]’- dianion (compkx X in Fig. 4) (2) and, probably, in the homologous anions [Rh6(CO) j]2- (99) and [IrG(CO) 153’- (96). The progressive lengthening of thc Co-C distances (1.74-1.90-2.00 8) for the sequence, term inal-etl!le brzdying-face bridging, and tlic parallel increase in the C-0 values (1.15-1.17-1.19 A), are in agreement both 11 itli the increasing mu1ticentc.r character and related steric request of the mvtal-carbonyl interaction, aid with the lowering of thc C=O stretching absorptions obscxrved in the Ilt qwctra (34). Structure XI (Fig. 4) is conimoii to the isoclectronic anions of the salts [NMe~]s[CoaNiz(CO)ll] (8),I 90 80 I0 40 30 20 +o. 1 0 - 0.1 - 0.2 - 0.3 - 0.4 A FIG.5. Relationship between percent terminal CO and average negative charge per carbonyl group. groups and the average negative chargc per carbonyl group (74a). This plot not only proves that dissipation of ricgative charge is mainly respon- sible for the interconversion of terminal into bridging carbonyl groups but shows also that no significant deviations due to steric effects are present in this particular serics of compounds. The same plot allows limited predic- tions to be made concerning other members of this family. For example, the hypothetical species [Kh6(C0)17]2+and [RhG( CO) 13]6- should have fifteen and three terminal groups, respectively. A significant and uricxpected tendcncy of the octahedral framework to degenerate into a trigonal antiprism can already be noticed in the distribu- tion of the M-M distances of anions [COG(CO) 15]2-, [Co6( CO) 1114-, and [Co1Ni2 (CO) 14]2-; howsver, the best example of such a deformation is in the dianion [Nis(CO)12]2- (structure XI1 in Fig. 4), which is correctly described as a trigonal antiprism with mean basal and interbasal Ni-Ni distances of 2.38 and 2.77 A, respectively (28). The decrease in the intra- triangular distances can bc rclatcd to the presence of bridging carbonyl groups spanning all six edges of these faces. In fact, the carbonyl bridge represented as a polycentric bond generates di per se a bonding component between the metals (34). Furthcr twisting of the trigonal antiprism along the ternary axes gener- atcs a trigonal prism. Such psculiar packing has been found in the dianions [Pt3(CO)6]n2- (n = 2, 3, 4, 5) (29, 95),which contain a repeated trigonal High Nuclearity Metal Carbonyl Clusters 297 prismatic stacking of metal triangles along the pseudo threefold axis (structures XIII-XV in Fig. 6). Why the platinum atoms in [I't,(CO)12]2- prefer a prismatic arrange- ment is not clcar. The antiprismatic packing found in [Ni6(CO)12]2- should be the more favorable in both casrs as such packing requires less steric in- CPt (CO ) ( p2- co)611 - ( D3h ) (XIII) FIG.6. Schematic representation of the molecular structures of some [Pts(CO)a],,* (n = 2, 3, 4, 5) anions. 298 P. CHINI, G. LONGONI, AND V. G. ALBANO teractioiis between the two halves of the dianion and formally allows a gain of three n1-M bonds. Probably the prismatic geometry rwults from a total energy minimization in M hich olcctronic r(Lasons, such as the rcluc- tancc of platinum to use all of its valence orbitals, are the key factors. In all of tliesc platinum clustcrs the intratriarigular mctal-metal bonds (2.66 A) are supported hy carbonyl bridges and arc' significantly shorter than the intertriangular bonds (3.04-3.0s -%). The observed deviations from a regular prismatic stacking of platinum atoms, which include a translational sliding by ca. 0.51 along one edge in [pt6(co)12]2- and a top-bottom hclical twisting along the psvudo thrccfold axis in [Ptg(CO)1$- and [Pt1s(CO)30]2- (by ca. 26" :\rid 64.1", rcspcctivdy), may result from a tmdcncy to minimize nonbonding repulsions, niaiiily among the carbonyl groups of adjacent layers. The same kind of structure has alw bwn found in thr dianion [Pt 12 (C'O) 24]?- (95) and is most probable for thc dianion [PtlB( CO) ?GIL-. Diariion [Nis(CO)18]2- has a structurc very similar to that of thc con- gener [Pte(CO) & hut in this case the top-bottom helical tnistiiig amounts to a total of ca. 60" (95). Thr iritratriangular distances (2.44 A) are once again much shorter than the intertriangular ones (2.70 -4). Only three structurcs of hcptanuclear clusters arc' known, all containing metal fraincw orks based on the monocapped octahedron. Figure 7 5hows the structurcs of [Rh, (CO) (XI'I) and of [Kh7(CO) 16I]'- (XYII) (4,12). Structure XVI forninlly dvrives from that of [Kh6(C0)15]'-, reprc- sented by structure X in Fig. 4, by insertion of an lXki(C0)- unit into the uiibridgcd lower face of the octahedron, 11 hcrcas structure XVII dc- rives from that of Rh6(C:O)16by inschon of an 1LhP unit into one of the four unbridgcd faces. In both caws thcre ~~ouldbe subsequent formation of threc. edge carboiiyl bridges along the ii(w tetrahedral edges. In the octahedral part of the two clusters, the average 1x11-Rh distances arc very similar (2.76 versus 2.77 A). By contrast the average distances along thc new tetrahedral edges differ considerably in the two C~S('S(2.76 versus 2.93 A), in agrwnicnt with the presence of on(' extra electron pair in the substitutvd dianion. This sevms to bc another examplc, as previously pointed out by Dahl (Z."), of the cffcct of an excess of electron density in MO's mainly aritibonding with respect to the metals. The capptid octahedral structure of Os7(CO) 21J which is isoelcctronic with [Rh, (CO) has only been preliminarly rcportcd (104); all the Os-0s distances are similar (61). Despitc the c.xistencc1 of 86-electron octahedral carbide clustws, such as RuG(C0) 17C, thr diariiori [Ith6(CO) &]*- has 90 valence elcctrons and a trigonal prismatic array of nictal atonis (structure XVIII in Fig. 8) (7). This fact could indicate that the three 31--M bonds formally lost in the High Nuclearity Metal Carbonyl Clusters 299 PR (XVII FIG.7. Schematic representation of the molecular structures of the heptanuclear clusters. 300 P. CHINI, G. LONGONI, AND V. G. ALBANO P FIG.8. Schematic representation of different molecular structures in carbide high nuclearity metal carbonyl clusters. High Nuclearity Metal Carbonyl Clusters 30 1 octahedron-trigonal prism transformation are energetically compensated for by bonding two more carbonyl groups and by allowing more space for the central carbide (see later). Such an interpretation appears reasonable because the DRh-CO (ca. 39 kcal molt-') (27') is expected to be almost twice as much as t,he &-Rh (ca. 27 kcal mole-'). The congener [CO~(CO) &I2-, whose trimethylbenzylammonium (NMeSBz) salt is isomorphous with the corresponding salt of [Rhs (CO)&Iz-, appears to have the same structure (9). In the carbide rhodium clusters, thc Rh-C distances of the terminal carbonyl groups are significantly longcr than those of the noncarbide de- rivatives, whereas the corresponding C-0 distances are shorter (see Table 11). This trend indicates a decrease in backbonding to the carbonyl 302 P. CHINI, G. LONGONI, AND V. G. ALBANO TABLE IV EXPEI~IMENTALVALUES OF THE C.\RBIDECARBON R.LDIUS IN HIGHNCCLE \RITY METLL C LRBONYL CLUSTERS Polyhedron Square pyramid Fe5(CO)ljC 1.89-2.64/2 = 0.57 26 Octahedron 1.91-2.67/2 = 0.57 49 2.05-2.89/2 = 0.61 113 2.04-2.88/2 = 0.60 103 2.04-2.87/2 = 0.60 9 Half trigonal prism Co3 (CO),CX 1.92-2.48/2 = 0.68 109 Trigonal prism 2.13-2.79/2 = 0.74 7 2.13-2.81/2 = 0.72 1.2 Square antiprism (distorted) [CO~(CO)ISC]' 1.99-2.52/2 = 0.73 13 groups and, in the particular caw of [Rh6(CO)1,C]2-, is in agrccineiit with the shift to highcr frequenciw of thc carhonyl absorptions in the IR (see Scc- tioii 111). However in this case, th(. 13C KhIR of thc carbide carbon atom falls at very low fidd 264.7 ppin (10) (sw rmonanco Il in Fig. 13) suggest- ing that the carbidc carbon could br positivdy polarizcd. The M--C carbidc distances found in som(' HNCC arc summarized in Table IV. It seems probablv that in the smaller octalwdral cavities (com- pare Tables 111 and IYj thc positive chargc on the carbide atom will b~- come higher to allow the nccvssary contraction. A trigonal prismatic basic unit, containing a carbidc atom, is still present in Rhs(CO)& (structurt. XIX in Fig. 8) (12j. The prcscnce of a rhodium atom in an c~xcc~ptionalbonding situation, spanning just one edge of the original prism, explains 1 he particular reactivity of this compound. Struc- ture XIX does riot possc~ssany symmetry dement, even considering the metal skeleton alone, and challenges thc common opinion that clusters have high symmetry. Dianion [CoB(CO) is isoclcctronic with Rh8(CO)UC but presents a very different stcreochrmistry as shown in structure XX (Fig. 8) (IS). Its geometry can bc described as a deformed square antiprism with a car- bide atom in the ccntcr of the polyhcdron. This structure can bc derived from that of thr bicappcd trigonal prism by stretching the common edge of the two capped square faces, as indicatrd in Fig. 8 by thc dotted line. In such a process thc loss of an M-hZ bond formally generates a square face. The idealized symmetry of a tctragonal antiprism is D4d. Homever, High Nuclearity Metal Carbonyl Clusters 303 since the cluster is stretched along onc of the twofold axes, only Dzsym- metry is retained. One reason for this deformation is that the squarc anti- prismatic cavity is too largc. to allow sufficient bonding interactions bc- tivecn the central carbon arid thcb cobalt atoms, and the most favorable arrangcmcrit leads to the 4 colxdt atoms moving nearer to the central carbon. The structurc of the monoanion [lihl,(CO)zsCz]-, which is one of the biggest known clusters, is reported in Fig. 9 (9).The metal atom framework can be described as a ccntercd tvtracappcd pentagonal prism in which the bases and two side faces are capped. The pentagonal prism is not regular and the Rh-Rh distances betweon the two capped side faces are too long to be considrred bonding distnnrcs (3.33 A). In agrccincnt with the high average value of all the other Hh-Ilh distances (2.87 the chcniical reactivity of this compound is high. The polyhedron of [Rhlj (CO)z8C'2]- is chcmically derivcd from that of [Rh6 (CO)&I2-, and thc presence of two diffrrciit carbide atoms in octahcdral cavities strongly suggests that the two octahedral fragments derivc from the condensation and isomerization FIG.9. Schematic representation of the molecular structure of the anion CR~U(CO)B(C)~]-. 304 P. CHINI, G. LONGONI, AND V. G. ALBANO b FIG.10. Schematic representation of the molecular structure of Rh12(C0)26(C2). of two formerly prismatic units. Theocentral rhodium atom is coordinated to 12 external rhodiums (d = 2.90 A) and to the 2 carbides and, hence, can be considered to be in a true metallic situation. This special structure is probably due to the additional bonding contribution of the 2 carbide heteroatoms; such an energetic contribution seems able to stabilize cluster geometries that arc likely to be unstable in a noncarbide analog. The structure of Rhlz(CO)25(C2),reported in Fig. 10, is a further ex- ample of carbide stabilization (1.3). This is the most irregular cluster as yet characterized; like Rhs( CO)1&, it has no symmetry element. This cluster is also derived chemically from the oxidation of [Rhs(CO)&I2-. The 2 central carbide atoms are definitely bonded together (1.47 8) and lead to High Nuclearity Metal Carbonyl Clusters 305 the classification of this derivative as an ethanide. The metal atoms, al- though irregularly bonded, can bc considered as distributed on three differ- ent layers. Such a layered arrangement is similar to hexagonal packing with 2 atoms missing from thc cmtral layer and 1 extra atom in the lower layer (compare Figs. 10 and 11). The tendency of the finite clusters toward close packing of metals is clearly evident in the structure of the anions [Rh1S(CO)2iH5--n]n-, (n = 2, 3) , reported in Fig. 11 (13).The central metal atom is dodecacoordi- nated, whereas the 12 rhodium atoms on the surface are pentacoordinated toward the other metal atoms and tricoordinated toward carbon monoxide. The average Rh-Rh distance is 2.81 without significant differences be- tween internal and surface bonds. The high symmetry of the metal skele- ton, D3h, decreases to apparent C, in the molecule; it has not been possible so far to determine thc positions of the hydrogen atoms. Structure XXIII (Fig. 11) corresponds to the smallest possible unit of d FIG.11. Schematic representation of the molecular structure of the dianion [Rhia (CO)zJLY. 306 P. CHIN!, G. LONGONI, AND V. G. ALBANO a close-packed metal and conclusively demonstrates that the metal skelc- ton of a cluster can be regarded as a “rouiid surface.” Such a relationship between round surfaccs and HNCC also providcs a formal rationalization of the unsymmetrical distribution of bonds around the metal ccnters (as, for example, has becw pointcd out beforc. for octahedral clusters). Formally, this is simply a consequence of the fact that on a surface the distribution of bonds can ncvw be symnirtrical rdativc to the plane of the surface itself. Therefore, important connections bctwecm HNCC and surface chem- istry are expccted, and HNCC could be a reliable model for carbon monox- ide absorbed on mctallic surfaces. 111 STRUCTURAL DATA IN SOLUTION A simple technique. that providcs structural information in solution is infrared spectroscopy, particularly in the carborigl stretching region (2100-1600 cin-l) (25). Unfortunately this method dovs not generally provide sufficient information for a complcte structural characterization of HNCC. This limitation is mainly due to the prc.dominancc of local coupling between carbonyl groups; in other words, local symmetry is mainly re- sponsible for the spectrum. Wcvertheless, the infrared spectrum very often allows an unequivocal decision about the presence of bridging carboriyl groups, although further diffcrentiation between edge and face bridging groups is sometimcls very difficult. Generally a difference of 150 to 210 cm-l between the main absorptions of terminal and bridging carbonyl groups indicates edge bridges, whcreas a diffcrcnce of 210 to 250 cm-’ (but 400 em-’ in [Co,(CO)lo]- (66) } indicates face bridges. The assign- ment is complicatcd both by thc possible prrscnce of asymmetric bridges and by the broadness and consequent low resolution of these absorptions. The salts of carbonyl metalates arc generally soluble only in polar solv- ents, in which absorptions are broader due to the dipole interactions of the solvent. The spcwtra are also often considcrably dependent on the particular cation and solvcmt, due to the formation of ion pairs. The largest differences have been observed on going from alkali salts to salts of the large tetra- substituted ammonium and phosphonium cations; the latter generally provide simpler spectra that indicate minor formation of ion pairs. A similar trend is observed in going from less polar (e.g., THF) to more polar solv- ents (e.g., CHsCN). In the carbonylmetalatrs the main absorption of terminal carbonyl groups is fairly well related to the ratio of the number of metal atoms to MA/NC Ad I I I 2080 2040 1875 205 I [I I 1 6 2070 2040 2$07 1823 1771 275 I I I I I1790 6 2055 2022 1825 1795 262 __.~~.~~ --______- I I1 I I I 1 4 2060 2025 1990 1960 1887 1842 1655 1113 2010 I II I1 255" 3.5 2058 2018 1996 1777 1763 I II I I 212 3 2045 1980 1960 1790 1768 I I 11 154 3 2030 19$0 1883 1843 1831 155 2.3 1d87 19'50 1820 17'75 * I 1 I 2 1975 1925 1810 115 r I I I I I 1.5 1985 1945 1890 1735 1700 1650 155' I 1 1900 I II I 9 2060 1900 1875 1840 185 r I1 11 185 7.5 2055 1890 1840 1870 1825 I I 185 6 2045 1880 1840 I I1 188 4.5 2030 1855 1835 I II 195 3 1990 1818 1795 a CHaCS. CI&O-CH,-CH,OH. CHaCN, potassium salt. Separation between strongest terminal and bridging absorptions. 0w FIG. 12. Infrared spectra of rhodium and platinum polynuclear carbonyl anions in the carbonyl stretching region (tetrasubstituted U ammonium or phosphonium salts in THF). MA/NC, ratio of number of metal atoms to the number of negative charges. 308 P. CHINI, G. LONGONI, AND V. G. ALBANO the number of negative charges (MAINC) (94),as shown for the carbonyl- rhodates and platinates in Fig. 12. This is a very useful relationship, be- cause it is often possible to interpolate with reasonable accuracy the MA/NC ratio of a new and unknown species just from the IR spectrum. Obviously, the rdiability of this information is higher when comparing carbonylmetalatts of thr same metal, and when both local symmetries and the ratio between different types of carbonyl groups arc similar (as with platinum). A surprisingly reasonable agreement is, however, often obtained in less related cases. Introduction of a carbide atom results in a lowering of the stretching absorptions of the terminal groups (e.g., [co6(co)15]2- 1980 cm-l and [Co6(Co)&]'- 1975 cm-') in agreement with a positive character for this central heteroatom (10). Sometimes, however, the opposite shift is observed (compare [Rh6 (0)l5I2-and [Ith, (CO) 1d3-J- in Fig. 12). Very useful information on the actual structure in solution can be ob- tained from 13C NMR spectra (118),particularly when using samples en- riched in TO(now available commercially). Enrichment can often be accomplished by direct exchange at 25°C and 1 atm. This 13C NMR tech- nique is particularly useful with the salts of polynuclear carbonylmetalates, because their solubilities in polar solvents (such as CDaCOCD3) are gen- erally sufficient even at low temperatures (-70°C). The use of a low temperature not only generally reduces the fluxional character of these molecules but often rrsults in a marked sharpening of the signals when metal atoms of high nuclcar magnetic spin are present (119). A typical 13C NMR spectrum is shown in Fig. 13 for the case of the [Rh6(CO)&]2- dianion (lo),whose structure is shown in Fig. 8. Here SOLVENT D C B 'I 'I J L &-+.+bIdd- II8 "& It 264 240 220 200 PPm FIG.13. The 13C NMR spectrum of (a) [R~~(CO)IPC]~(ca. 90% l3C) and (b) [Rhs(WO)laCP (ca. 92y0 WO) at -70°C in perdeuterioacetone solution in the presence of Cr(acac)a. High Nucleority Metal Corbonyl Clusters 309 the coupling with the rhodium atoms (I = f) splits the terminal carbonyl resonance into a doublet (A), and the two different sets of bridging car- bony1 groups into two triplets (B and C). Similarly, it has also been pos- sible to show that at low temperature anions [Rhlz (CO) 3Ol2--, [Rh7(CO)16]3-, [Rh,(CO) J]-, and [Rh6(CO)14]4- all maintain the struc- ture found in the solid state (47, 74, 74a). Generally absorptions at about 180 to 210 ppm [low field from tetra- methylsilane (TMS) ] are characteristic of terminal carbonyls, whereas those at about 220 to 260 ppm are typical of bridging carbonyl groups. Figure 14 confirms that the main variable responsible for the 13C0chem- ical shift toward low-field positions is the net partial negative charge that can be transferred to the carbonyl group (74).This has previously been shown for monomeric carbonyl derivatives (18~). The main limitation of the method derives from the fluxionality of the carbonyl groups, which, on the other hand, is itself a potential source of basic information. The low-temperature fluxional behavior of the carbonyl groups can be ascribed to the nucleophilicity of the delocalized skeleton electron density, which can accumulate on particular metal centers giving rise to inter- mediate carbonyl bridges. I'luxionality is, therefore, expected to increase on increasing the negative charge, or on decreasing the MA/NC ratio. This 0 *V 0 A FIG.14. Relationship between weighted average of WO chemical shift (SCO') and fraction of negative charge per carbonyl group (A) in rhodium high nuclearity metal carbonyl clusters. 310 P. CHINI, G. LONGONI, AND V. G. ALBANO TABLE V COMPARISONOF FLUXIONIL BEH LVIOR AND RE wrIvIm OF SOMERHOIIIUM C YRBOUYL CLUSTERS Temp. Relative NO. Compound (“C) Behavior PPm “reactivity” 1 li114(CO)lz +63 Fluxional 189.5 (quintet)“ 1 > 2* 2 IthG(CO),c +70 Not fluxional 231.6 (quadruplet) 180.1 (doublet) 3 [I a Data from Cotton et ul. (56). Reactivity toward HZ+ olefins, ILO, NazC03 + CIT30H, HzS04+ CH,CN. Reactivity toward 011- + CO, [Iih(CO)a]-, IihZ(CO)&. Reactivity toward CO and HZO. is not the whole story, however, as differencrs in conformational energies, due to the different gcomctrical distribution of t8hecarbonyl groups, could also affect such procrsscs. Although only liniited information is prcsently available on thc fluxioriality of HNCC (74), a low activation energy conccrted mechanism, involving formation and breaking of bridges, secms generally reasonable. SevcLral similar mechanisms have already been dis- cussed in the literature for less complicated clusters (55, 64). An rmpirical relationship bctwcon fluxionality and “rc.activity” is pres- ently cmerging (Table V) in agrccmrnt with the common observation of the ease in changing the original dist ribution of the ligands and in produc- ing vacancies in the barrier of carbonyl groups. Although information on the forcc constants of metal-metal bonds can be obtained from Raman spectra (115), no study of HNCC has yet been reported, probably becausc thcse clusters arc highly colored and readily decompose in the 1asc.r beam. Chmparison between several colorless mono- nuclear carbonyls and thc strong colors of HNCC shows that the metal- High Nuclearity Metal Carbonyl Clusters 31 1 nietal bond is a chromophorc., probahly because of the decreased scpara- tion between the frontier orbitals. The obvious liniit to this separation is apparent in the continuous absorption found in pure mctals (41). At- tempts to use the clcctronic spectra of HKCC for identification purposes have given poor results, mainly due to the broad character of thcse ab- sorptions (95). IV SYNTHESES Condensation rmctions, nhicli arc’ at the very hrart of the synthetic methods for HKCC, arc cxpcctcd to I)(> subjc,ct to basic ewrgctic condi- tions. The available cxperimc1it:il dnt a show that condensation reactions such as ~CO~(CO)~-CO~(CY))~> + 3c‘O AH ‘u 33 kcal (62) (1) 3l (a) Fe3(C0)]? - &3(co)12 (e) [(h6(co)15cl2-- [Rh,(CO)isCI2- (b) Coa(CO)lz - Itha(CO)12 (f) [~16(Co)12I2- - [1't6(CO)1212- (C) CO6(CO)i, - Rh,(c())i~ (g) "ls(CO) 181'- - [l'tdCO) 1d2- (d) [C06(CO)i.5l2- - [Rhe(Co)is12- This difference in behavior reflects diff erent erithalpic conditions, because for each pair one could assume very similar contributions of the AS term (although in the a, f, and g pairs there is some approximation since the species are not strictly isostructural). Moreover, the ready exchange be- tween some species, such as [Rh,(CO),,]2-, and WO shows that kinetic factors may probably not be responsible for this behavior. The relative increase in Dh1-M in descending a subgroup is of general synthetic significance, bccause, whereas synthesis of HNCC of first-row transition metals is generally possible only in the absence of carbon mo- noxide, in the second and third rows thcre is less dependence on carbon monoxide pressure. In the first row it is, therefore, necessary to start from preformed carbonyl species and to condense them in a second stage, whereas in the second and third rows reduction to carbonyl spccies and condensa- tion to HNCC can oftcn be carried out in one step. This direct preparative method is illustrated by the nearly quantitative reactions (95, 102) : 25OC. 1 atm + 42KC1 + 12KzC03 f 24HzO (3) 25°C. 1 atm 151Ta2PtC16+ 84NaOH + 61CO + Na~[Pt~dC0)301 MeOH + 9OSaCl + 22XaHCO3 + 9C02 + 31H20 (4) In the preparation of the HNCC of rhodium, decomposition by carbon High Nuclearity Metal Carbonyl Clusters 313 monoxide always becomes a limiting condition when the ratio MA/NC is low. This is shown by the behavior of the hexanuclear clusters (38,99,102) : -70°C. 1 atin 5[Rh6(CO)16]'- + 17CO + * 2[Ithi,(CO)3alZ- + 6[Rh(CO)r]- (6) 25"C, 1 atrn 25%. 1 atni (Rh,(CO)irI4- + 5CO * [lll~a(CO)ii]~-+ 2[Rh(CO),]- (7) Neutral Rh6(CO) 16 is perfectly stable under a carbon monoxide atmos- phere at 25"C, and the dianiori [Rh6(CO)15]2- is still stable at 25°C (al- though degraded at - 70"C), but thc trtraanion [Rhs (CO)11]4- is readily demolished under the samc conditions. The interpretation of this behavior is that the tcrm DM-c, is vcry drpcndent on the MA/NC ratio, due to the relatcd changc in back donation. Probably this is a very general effect and most of the HKCC are expocted to become increasingly unstable to- ward carbon monoxide on increasing the reduction of the cluster. On the basis of mechanism, condensation processes can be divided into two broad categories: (a) reactions induced by coordinatively unsaturated species; and (b) reactions betwen coordinatively saturated species in diff crent oxidation state (redox condrnsation) . Among the usual methods for generation of coordinatively unsaturated species (34), only pyrolisis has been applied to thc synthesis of HNCC. The main examples arc in ruthcnium and osmium chemistry (58) : 195°-2000C 0b;(CO)z1 + Os,(CO)z, + Oss(C0)ziC (8) Pyrolisis of a Rh, (CO)12 solution at about 60"-80"C reprcscnts one of the best methods for obtaining purr nh6(CO),,, whereas formation of the metals is observed at the high tcinperatures required for the thermal de- composition of Cor(CO)lz (-100°C) and Ir4(CO)12 (-160°C). In all of these cases, photochemical mvthods (122) could both significantly im- prove the selectivity of thr syntheses arid allow the isolation of new HNCC not accessible by thermal routts The relevance of redox condrnsation, which by contrast generally re- quires very mild conditions, began to rmerge slowly after 1965 (15, 110) when Hiebcr and Shubert reported the first example of such a reaction ('78): 25'C [Fe3(C0)l,]2-+ Fc((D)~- [Fe4(CO)13]2- + 3C0 (9) THF Here we mention only some recent applications of similar pairs of reagents 314 P. CHINI, G. LONGONI, AND V. G. ALBANO to the synthesis of HNCC, such as (13,102) since more examples arc presented Iatcr. Redox condensation is not rcstrictcd to the reaction of a carbonylmetal- ate with a neutral carbonyl, and it is possible to condense the carbonyl- metalate with other simple cationic species; for example (35, 102), 6OOC. EtOII ~[CO(IC~OH),][CO(CO)~]~ ~3[Co(~:tOl~),]lCos(C0),51+ llCO (13) vacuum Redox condensation bctwcen anionic and cationic species is often highly dependent on the medium, and the use of less basic solvents of high steric requirements, which do not coordinate strongly to the cationic center, may result in a marked improvement in thc ease of condensation ($5). The following are further examples of redox condensations, showing the general possibility of such a reaction between species in different oxidation states (99,101): 25°C [IthG(cO)ia]*- $- [I The great utility of redox condensation is due both to the ease with which the reactions take place, probably because of the redox character of the reactions themselves, and to the high number of possible combinations of reagents. This number is much higher than suspected at first glance because it is often unnecessary to start from a preformed pair of reagents. It is sufficient to generate in situ a second oxidation state by addition of a suitable reducing or oxidizing agent. Probably most of the syntheses of HNCC, and generally of polynuclear carbonylmetalates (34), in reducing conditions are of this type. A detailed discussion of some redox condensa- tions induced by simple reduction or oxidation is presented for rhodium and nickel in Sections XI and XIII, respectively. In the HNCC carbides the prcsencc of a central carbidc atom is expected to contribute significantly to t8hebonding encrgy of the cluster and, there- fore, to result in less dependence on the partial pressure of carbon monox- High Nuclearity Metal Carbonyl Clusters 315 idr. Unfortunately, the prcscmt stag(. of our understanding of these syn- thetic reactions is very primitive. The most common synthetic mrthod (Fe, Ru, 0s) is via pyrolysis of noncarbide metal carbonyl dcrivativcs. It seems probable that the high temperature (140"-260°C) is riot only responsible for formation of coor- dinativcly unsaturated spccicw arid rdatcd condensations but also for the necessary disproportionation of carbon monoxide (58) : 2CO ----+C + ('02 AG:98 = -28.64 kcal mole-' (10) An example of such a reaction is (85) n-BuzO, 142T llU3(CO)I? *Ru + lt~4(('O)iill4 + I?,u~(CO)~Z€€~+ ILU~(CO)~~C (17) G hr, Hz0 53% l(l[Z 3% 30% A differ~ntapproach, based on the us(' of carbon halides, has becn dis- covered niorc rccontly for cobalt and rhodium HNCC carbides, but its extension to other metals has not ytbecn successful. For cobalt the reac- tion is corivenicritly carried out in two different steps; first, the well-known chloromethynyl derivative, C'03 (('0)y('C1 is prepared (63'), 40°C ,iCo,(CO), + 3CCI4-2Cu3(COi9~C'1 + ~C'OCI~+ 22CO + CC14 $CI,C=CCI, (yield - 90%) (18) and then the transformation itito a carbide is completed via the tctra- carbonylcobaltate ( - 1) anion (9), %-re10 Co,(CO),CCl + 3X~a[Co(C0)~]---+ Xa,(Co6(CO),5Cl + SsCl + 6C0 (19) 23'C Otherwise oxidation of the [CO(CO)~]-anion to Co2(CO)8by the CC14 is observed (IS). For rhodium the reaction can be conveniently carricd out in one step, 4[NMe3Uz]CI + 9CO (20) but high yields (-90%) are obtained only when thc [Rh,(co)&y- dianion separates as an insoluble salt during thc reaction. Using I3C C14, it has been possible to determine unequivocally the source of the central carbide atom by the presence of a rrsonance at 264.7 ppm (resonance D in Fig. 13) (10). Resistance of thc rhodium HNCC to degradation by carbon monoxide also allows, in this caw, a more convenient direct preparation from KsRhClo and CHCI, [see Eq. (60)1 (7). Reasonable hopes of extending this method of synthesis to other halides of nontransition elements are presently under experimental verification 316 P. CHINI, G. LONGONI, AND V. G. ALBANO (IS).A preliminary claim (46) of a related [Rh6(CO)15Si]2-dianion was consequently proved incorrect by full X-ray and elemental analysis (13). The observed formation of the [Rh6(C0)uC]2- dianion, from SiCL and [Rh(C0)4]- anion, can be explained on the basis of the known trans- formation of MeaSiCo(C0) 4 into MeySiO-CCo3( CO) (80). Insertion of a Si-Co bond into a carbonyl group is believed to be due to the high stability of the resulting Si-0 bond (80) and may prove to be a serious limitation. v METHODS OF SEPARATION The synthesis of HNCC is rarely very selrctive; more often a mixture of high molecular weight compounds is obtained and the separation prob- lem becomes crucial. Separation is carried out using various techniques, depending on the nature of thr compounds. For nonionic species the solubilities of the corn- pounds are generally low and similar, and it has been necessary to use either fractionation by continuous extraction with low boiling solvents ($6) or thin-layer chromatography (58, 60). The first method has been used with air-sensitive compounds, whercas the second has bccn applied only to air-stable substances. In both cases, it is possible to separate only limited amounts of compounds, whose characterization is, therefore, car- ried out using particular techniques such as mass spectroscopy (58-60). Separation of salts of carbonylmetalates is generally easier, because solubility of thr salts of anions of similar molecular weight is mainly de- pendent on the MA/NC ratio. For instance, at MA/NC of 6 the sodium salts generally separate from aqueous solutions by simple addition of ex- cess NaCl {e.g., [Ith12(C0)30]2- and [Rh6(CO)151]-], wherras at an MA/NC value of 3, precipitation of the potassium salts is possible by addi- tion of potassium ions (e.g., [Co6(co)15]2-, [Ith6(CO)&y-, and [Ni6(C0)12]2-]. At MA/NC of 1.5 the potassium salt can be obtained only from very concentrated solutions of potassium ions {e.g., [c06(Co) and [Rhc(CO) 1414-), and thcse anions arc often more con- veniently precipitated as tetraalkylammonium salts. Great difficulties in separation have been found only in cases when the MA/NC ratios are very close. Sometimes it is not possible to use an aqueous medium because of the reaction between water and the carbonylmetalate. In thcse cases, separa- tion can be attempted using salts of large cations, such as “Bud]+, High Nuclearity Metal Carbonyl Clusters 317 [PPh4]+, and [PPN]+, and organic solvents such as THF and isopropanol (102). The preparation of HNCC in the form of well-shaped crystals suitable for diffractomctric analysis is cxtrc.mely important because of the basic need for this type of charactwization. In the last few years, we have used cxtensivcly crystallization methods involving slow diffusion of solvents in which the complex is insoluble. Gmcrally, it is sufficient to stratify slowly a lower-density solvent ovpr thc solution of the compound (e.g., i-PrOH on acetone, i-PrzO on McOH, tCJl1K’nP on THF) ; crystallization occurs by simple diffusion for a couple of days. Unfortunately, with salts of carbonyl- metalates the case of formation of crystals is generally unpredictable and dependent on the particular cation. Often it is necessary to try a whole series of cations, which may tw vwy time-consuming, in order to obtain propc’r crystals. The elimination of solvcmt coordinated to the alkali cations may disrupt the original crystal structurv during drying of the crystals; in this case, the apparent morphology of thc crystals may remain unchanged, but only a powder pattern is observcd on X-ray diffraction. Sometimes, similar be- havior has becn observed with large. cations, such as “Me4]+ and [PPh4]+, possibly due to the original prcscmrc of clathrated solvent niolecules (95). VI REACTIVITY It is well known that the carbon atoms of carbonyl groups are readily attacked by nucleophilic agents (41, X),whrreas the oxygen atoms of the same groups are available to clcctrophilic attack (93). It, therefore, seems reasonable to envisage thc prcwnccx of an alternatively polarized double barrier of negative oxygen and positive carbon atoms. unfortunately at the present 1cvt.l of knowledge, it is difficult to say under which conditions the high nuclcarity clusters rcact by an associative or a dissociative mechanism. Howcw,r, since the number of carbonyl groups bonded to each metal atom is rather small, it is reasonable to assume that the dissociative energies should bc high and that an associative mechanism should, therefore, predominate. Associative reactions are expcctcd to take place either indirectly, by prior attack on thc external barrier and transfer onto the core of the cluster, when the ligands around the cluster are crowded, or directly on the cluster core when the crowding is irregular. Figure 15 shows the polyhedra described by the oxygen atoms of some hexanuclear clusters. It is evidrnt that, on increasing the number of car- 318 P. CHINI, G. LONGONI, AND V. G. ALBANO 0 Terminal carbonyl groups @ Edge bridging carbonyl Kroups 0 Face brldging carbonyl groups FIG.15. Idealized polyhedra resulting from the distribution of the external oxygen atoms in some hexanuclear clusters. bony1 groups around the metal core, the packing of the oxygen atoms be- comes more and more compact. Whereas the polyhedra of the less crowded species, such as [N6(CO) 12y-and [Co6( CO) 14]4-, have hexagonal and quadrilateral faccs, the more crowded Rho(CO) 16 has only triangular oncs. The extcrnal geometry can greatly affect thc reactivity because the pres- ence of hexagonal or quadrilateral faces is expected to facilitate the direct attack of a reagent on the metallic core. However, the absolute significance of these steric considerations is difficult to assess not only ducx to the lack of knowledge about associative or dissociativc character, but also due to the dynamic character of the external barrier of ligands. It has already been pointed out that fluxionality and “reactivity” present similar trends, and High Nuclearity Metal Carbonyl Clusters 319 it seems probablc that dynamic rffccts controlling thc distribution of the ligands can often dominate the properties of thc cluster. Onirig to this basic lack of kiionlcdgr we have prcfcrred a morc formal classification of the reactivity of I-IN(’C based on the product obtained from the reaction itsrlf. Rcactioiis of I-INCC have therefore been classified in four main typcs: reductioii, oxidatioii, substitution, and oxidative addi- tion. Simple addition of neutral ligantls, which results in destruction of the cluster, is a common secondary procrss; it is mainly discussed with sub- stit ut ion. A. Reduction Itcduction of HNCX is srldoni a siiiiple process in which there is sub- stitution of a carbon monoxidc ligand, or of a metal-metal bond, nith an electron pair. A rare cwunplr of this type is (102) Io[<>OG(Co)]S]*-$- 22Xa- <~[~()CJ((~(~)I~]~-+ 6[Co(Co)4]- + 22s&+ (22) takes place with formation of carbon monoxide (a), and this, as shown by separate cxpcrimcnts, adds to the rcducrd cluster with formation of the [Co (CO)4]- anion (b) (37). A further complication is due to bond redistribution reactions as, for example, in the following sequrncc : [l’tB(Co),,12- + 2121 -ll’16(c~)l*l~- + [I’t,(CO)cl2- + 2IA+ (a) [l’t,(CO) 612- + [I’t 8(CO),42- -2ll’t a(C0)1?12- 0)) 2[l’tB(Co)18]2- + 2LI ----t3\I’tG(c(l)1?]2- + 2LI’ (23) Similar bond redistribution reactions have been carried out starting from preformed reagents, and have great significance in the chemistry of the oligomeric carbonylplatinatw (95). The formation of free carbon moiloxide, and consequcnt degradation reactions can be avoided using alkali hydroxidcs or alkoxides as reducing 320 P. CHINI, G. LONGONI, AND V. G. ALBANO agents. In these cases there is simultaneous oxidation of the carbon monox- ide to carbon dioxide. For example, the reaction between Iih, (CO)lfi and concentrated aqueous potassium hydroxide takes place according to the stoichiometry (39) 25OC Rhe(CO)i6 + The mechanism of this type of reaction, originally elucidated by Kruck (94), is related to the simple addition of the nucleophile observed with alkali alkoxides. InL the present case the following reactions, in which the saponification step is also of preparative significance (99, I@?), have been observed (45) : 25% Rh,(CO)ls + OR- -[I~~,(CO)I,(COOT~)]- ROH (R = hie, 13,) (25) 25oc [Rh,(CO),,(COOR)]- + 2X2tOH [lths(CO)Is]2- + Na+ + NrtHCOj + ItOH (26) Often the carboalkoxy group can hc hydrolyzed by siniple reaction with water, and the case of this hydrolysis is higher at lower MA/NC ratio, e.g., in the series [Rhfi(CO) lj( COOMc) 1- < [Rh, (CO) 11 ( COOMc) 1- < [Rh,(CO) lr(COOMe)2]L-. Somctimc.s, when the carboalkoxy group is not extremely sensitive. to hydrolysis, thr carboalkoxy anion has been obtained by simple use of anhydrous sodium carbonatc suspended in alcohols (99). Electrochemical reduction has bwn applicd to a number of low nu- clearity carbonyl clusters (57, 65), but to date there has been no applica- tion to HNCC. An interesting reducing agent is cobaltocene (S7), although its use is somewhat limited by the low solubility of thr cobalticinium salts. B. Oxidation Simple oxidation, resulting in substitution of negative charges by carbon monoxide or by metal-metal bonds, is not a common process. Complica- tions due to redox condensation reactions, to decomposition from the oxidizing agent, and to ticgradation by the evolved carbon monoxide, are frequent. For example, in the reaction (X), HeZ+. 25OC High Nuclearity Metal Carbonyl Clusters 32 1 formation of CoG(CO) 1G is accompanied by large amounts of Coq (CO)12 and by some cobalt(I1) cation. The best results have generally been ob- tained using stoichiometric amounts of oxidant or by working under condi- tions in which the reaction product is insoluble and not very sensitive to the excess of oxidizing agent. Often oxidation reactions arc' possible via intermediate formation and decomposition of hydride derivatives; for example (96, ll7), The relative acid strength in a series of anions [Mn(CO)z]- is expected to increasc on increasing the numbor of mt%alatoms n, because of an increased ability to accommodate the dolocalized negative charge. Significant devia- tions from this simple behavior may be predicted, however, due to the need for new coordination positions for the hydrido groups and to the related change in the overall distribution of ligands. Moreover, as pointed out by Kaesz (BY), both protonation and dcprotonation are generally slow and show high kinetic isotope effrcts; this behavior is considered to be char- acteristic of a tunneling mc.chanism imposed by the barrier of carbonyl groups. Therefore it is not surprising that formation of hydride dcrivatives of HNCC often requires reaction with concentrated acids. This behavior should be compared with the. result obtained on addition of a stoichio- metric amount of dilute sulfuric acid to the barium salt of the [Rhlz (CO) 30]2- dianion : In this case the IR spectrum of the dianion remains unchanged, showing that there is no detectable formation of Rh-H bonds but only of hy- dronium ions (43). Finally, selective oxidations, in which the oxidant is added to the poly- nuclear species, arc sometimes possible using halogens or pseudohalogens in stoichiometric amounts, for example (45, lOl), 322 P. CHINI, G. LONGONI, AND V. G. ALBANO C. ligand Substitution Simple substitution rcactions of carboriyl groups with tertiary phosphines generally occur only with the most robust species; for example (85)~ n-hexane I~usico),iC+ I, -1 (hlX = Sal, Ynu,I, NEtr13i, N11c3B,C‘1, 4d’IirCl, KCN, RSCN, yieltli 50-80%) compares with the dcgradation (101), 4[Rhi?(CO)?o]2-+ 1‘LClk -t 7[I D. Oxidative Addition Oxidative addition is a particular addition that results in an increase in the formal oxidation state of the metal atoms and that is known to occur High Nuclearity Metal Carbonyl Clusters 323 readily on metal-mctal bonds. Although it is expected to bc a common reaction in HNCC chrmistry, only a fcw cxaniplcs arc known, arid most of th(1rn are rc1atc.d to the dianioii [Ith12(CO)30]2- (45, 109) : [I~~~,~((YN~~I~-+ x2- ~~i~iI6(~~o),,~~- (s,= L,ir!) (38) It is probable that the hydrogvii li:tlid(,s also react in a similar way, al- though in this case the cxpoct od hydridc monoanion has ncver been ob- served (45).At ambicint pressurci thc rapid addition of hydrogen requires about 50°C, and the resulting hydrid(x anion is more readily isolated as an acyl anion after furthrr reaction Jvith cbthylenc and carbon monoxide (102) : H2 [1 VII IRON DERIVATIVES The reaction in rcfluxing diglymc of Fe(CO)5with a wide variety of carbonylmetalates, such as [Co (PO)4]-, [Fe (CO)412-, [Mn (CO) 5]-, and [Ir(CO) a]-, gives the hexanuclcar dianion [FP~(CO) &]*-in good yields (49, 117). Under analogous conditions, however, [CpMo(CO) 3]- reacts with Fe(CO) to give thc p(~ntaiiuc1car dianion [Fej(CO)1&]2- (79), which has not yet bcm structurally characterized. The extrcme dependence of t hew rcactioris on the experimental condi- tions is well cxcmplificd by thci follm irig syntheses: * [I'c(CO)r]'- + lln2(CO)lo (41) refluxinn THF (78) refluxing diglyrne (24) Fe(C0). + [nln(CO)sl- * [FeJ'In (CO),?I- (42) 5 min reaction refluxiiig dlalylne (49) * [ Fe6( c'o),&]'- (43) I hr reaction 324 P. CHINI, G. LONGONI, AND V. G. ALBANO The hexanuclear carbide dianion [Fe6(CO)&y- reacts either with strong acids (98y0HzS04, 85% H3PO1) or with oxidizing agents (AgBF4, Ph3CPFG) to give low yields (up to 20%) of neutral Fej(CO)& (11'7). So far all attempts to prcparc the hypot>hcticalFeG( CO) 17C have failed. Finally, 1-electron oxidation of [Fe2Rc(CO) 12]- with tropiliuni bromide has been reported to give a neutral mixcd-metal cluster formulated as [FC~R~(CO)~~]~on the basis of elemental analyses (67);however, its IR spectrum, which shows carbonyl absorptions quit(. similar to those of the starting material, is inconsistent with such a formulation. Vlll RUTHENIUM DERIVATIVES Reduction of Ru3(CO) 12 depends primarily on the experimental condi- tions. Reduction with (>ithersodium borohydridc or sodium amalgam in THF, or with methanolic potassium hydroxide, followed by acidification, has been rrported to give thc tetranuclear species Ru4(CO)13H2 and RU.,(CO)~~H~(82). On thcx other hand, reduction in THF with carbonyl- mctalatcs such as [Rlii (0)5]- and [Cpl'"l(CO) 3]- gives, aftcr acidifica- tion, a more complicated mixture, which has hensuccessfully separated into its components by srlcctivci rxtraction. After elimination of some tctra- nuclear hydridc derivatives by dissolution in light petroleum, the sparingly solublc hexaniiclear Rug(CO) 18H2 was isolatcld by extraction in dichloro- methane (48). Its parcnt dianion, [Ru6(CO)18]2-, although prcdictcd theorctically (120),has been neither isolatcld nor observed even in solution. Ruthenium carbide-carbonyl clusters haw been obtained through py- rolysis reactions. Thus, by heating 12u4(CO)12H1at 130°C in the prescrice of ethylene (10-12 atin), trace quantities of Ituj(CO)& have been ob- tained along with higher yields (30%) of I~U~(CO)~~C(59). Thc latter is, however, bcttcr synthesized by direct pyrolysis of Ru3(CO) 12 in di-n- butyl ether (85). Heating Itu3(CO)l2in aromatic hydrocarbons, such as benzene, toluene, 1?1-~ylcn(~,or nirsitylenc, gives a mixture of RUG(CO) 17C and of thcrAr-arene derivatives HuG(CO)14(arene)c', which in the last solv- ent is the major product (58,83). Monosubstitutcd derivatives RuG(CO) 16( L) C [L = PPh3, P (p-FC'eHd)s, AsPha] have bwn obtained by boiling Ru6(CO) 17C in hcxanc with cxc(w tertiary phosphincs or arsiiics (85). High Nuclearity Metal Carbonyl Clusters 325 IX OSMIUM DERIVATIVES Pyrolysis at 200°C of Osa(C'O)12 in a scaled, evacuated tube afforded a mixture of at least sewn diffcrcwt carbonyl clusters which could be scpa- rated by thin-layer chroniatographj . In addition to some unrcactcd OS~(CO)~~,the new compounds, OS~(CO)~~,OS~(CO)~~, OsB(CO)18, Os8(CO)Z~, and Os8(CO)zlC', w(w identified by mass spectroscopy (58); the last compound was originally formulated as Os5(CO) (61).Further pyrolysis of Os6(CO)18 at 255°C' givcs thc pentanuclear carbide dcrivativc, Os5(CO) in 40y0 yicld (59). In the presence of trace quatititics of water, pyrolysis of Osl(CO)12 at 230°C results in a mixture of hkdrid(1 carbonyl clusters. Thc new dcmva- ' tives, OS;(CO)~~H~,Os5(CO)16He, OS~(CO)~~H~, and OS~(CO)~~(C) Hz, along with the known compl(bxc1i Ohl (('0)loH (OH) , OsI(C0) 13H2, and Osl(C0) 12H4 (GO), have bccm scparatcd by thin-layer chromatography and identified by mass spcctroscopy, as twforr. The IR spcctra of all of th(w (,ompounds do not shorn absorptions due to bridging carbonyl groups, and thc NRIR spectra indicate that th(1 hy- drogen atoms arc always in hritlging positions. Future structural deter- minations of such an imprcssivc, swics of polynuclcar derivatives will make a considcrablc contribution to tlw chmiistry of high nuclcarity clusters. Until now only thr structurcs of Os,,(('O) 18 and OS~(CO)~~have been dcter- mined (104). X COBALT DERIVATIVES Scheme 1 summarizes thr syrithcws, arid the most significant reactions, of the [C'O~(CO)~~]~-dianion, which can he considered the key species in cobalt HNCC chemistry. 326 P. CHINI, G. LONGONI, AND V. G. ALBANO Yellow-green [Co, (CO)1j]2- is conveniently synthesized in high yields (80-90%) by merely heating under vacuum an cthanolic solution of [Co ( EtOH) z][Co (CO).J2, obtained by the reaction of Coz (CO) with ethanol (35). An altcrnativc route to this could be the reduction of Co4(CO) 12 with alkali metals (37), although this is less coiivmient both because of the simultaneous forma- tion of a large quantity of tctracarbonylcobaltate and of the further facile reduction to the tetra-anion [Co,(CO) ,1y- according to Eq. (22). Cobaltocene behaves as a toluene-soluble pseudoalkali metal, and in this nonpolar solvent ionic products precipitate out as soon as they are formed. Thus, the reduction of Coa (CO)12 with cobaltocenc in toluene gives an intermediate that analyzes as [CoCp2][Co4 (CO) 10-111, and which is probably dimcric. Unfortunately, its lability has prevented further characterization (37). Protonation of [Co6 (CO) 15]2- at - 70°C gives a new unstable species, for which thc analyl ical and spectral data agree with the formula [Co,(CO)i&]- (33). A logical extension of the synthesis of [Co,(CO) 13-J- from [Co(EtOH).] [Co (CO),I2 has been successfully applied to the preparation of mixed- metal clusters. Thermal decomposition of [Ni( EtOH) z][Co (CO)q]t pre- pared in situ gives the red hexanuclcar dianion [Ni2C04(CO)1J2-, through the following redox condensation and redistribution processes (44): ~[K~(E~OH),][CO(CO)~]~+ 2[Co(CO)a]---t2[?;1C03(CO)ii]- + 22:IZtOH + 2CO (45) 2[NiCo3(CO),,]-= [N12C04(CO),4]z- + Co2(CO)s (46) The whole process can be represented schematically by 2L1*+ + GR.l--2Rf4-~;ZlaZ- + $12 (47) The deep red tetra-anion [Co,(CO) 14]4- is conveniently synthesized by direct reduction of Co4(CO) 12 (37): (11 = Id, Sa, I<) The potassium salt of this anion could be isolated only in 45% yields owing to its high solubility. Thc corresponding sodium salt has been isolated in High Nuclearity Metal Carbonyl Clusters 327 two isomeric forms that differ in their solubility in THF. This finding, together with the extreme depciidcncc of the IR spectra of [co6(co)14]4- on the cation and on the solvent, :m in agreement with the existence in solution of an easily reversible cquilibriuni of the type: 1 HF form A form I3 Form A, as confirmed by an X-ray diffraction study (S), has structure XI (sc~Fig. 4) with six terminal and cight face-bridging CO groups (vC0 = 1640-1680 em-l), whereas form I3 should have some edge-bridging car- bony1 groups (vc0 = 1710-1760 cni-l) (37’). On oxidation both [Co6 (CO)15]2- and [Co6 (CO),*I4- givc. the ncutral hcxanuclcar cluster Co6(CO)16. A pcdiar nietal atom redistribution process was found in the pyrolysis of C’oJthz( CO) 12, which gives C‘oJihA ((‘0)16 (100). This result should apparmtly rcyuirc the intcrmcdiatc, formation of an octanuclear species, 1% hich rearranges to CoJLh4 (CO)l6 11y dimination of Coz(CO) 8. Finally, cobalt carbidc-carl)oril\.l clusters have rcccntly becn isolated through a two-step synthesis. First of all, the w.ell-known Co3(CO) ,CC1 is preparc.d from Coz(CO) and ( ’C‘1 and then the hexanuclear carbide di- anion [COG (CO) &I2- is obtairicd in good yields (9) by further reaction with Na[Co (CO)41 in diisopropylc%hcr [see Eqs. (18) and (19) 1. Further rcdox condcnsatiori between [<‘06(C’O) 1j(’]2- and Co4(CO) 12 [we Eq. (11)3 gives the square antiprismatic. ortaiiuclear cluster [Co8( CO) &I2- (1.3). Both these carbide derivatives, as wrll as all of the other cobalt high nu- clcarity clusters, are sensitive to air and react with carbon monoxide at atmospheric pressure. XI RHODIUM DERIVATIVES Since the time of thc original prciparation due to Hieber and Lagally (75) the synthesis of Rh6(CO) 16 has been considerably improvcd, and the best mcthods now available arc shomii in Scheme 2. This hexanuclear compound is sparingly soluble in organic solvents and its purification is rather difficult. However, thermal decomposition of Rh4(CO) 12 in solution affords analytically pure microcrystals (98). 328 P. CHINI, G. LONGONI, AND V. G. ALBANO In spit(. of its low solubility, It1ib(C'Ojl6 reacts nith a widc variety of iwutral ligands to give hJth iubititution and dcgradatioii products. In 1970, the synthcscs of thcx spc1cic.s Rh6(C10j7LS(L = PPh,, AsPh3j and ltlic,(CO j lo(1,-I,) 3 (L--I, = 1 ,2-bisdiphniylphosphinorthaiic.) wcrc rc- port c.d (84).Both compounds rcmlt froin th(. rcaction of Rh6(CO) 16 ith cxccss ligand in rcfluxirig chloroform. Howwcr, the first formnlation is cspccially puzzling-it i.: hard to visualize how niiic bulky ligands, such as tripliciiylphospliirie or triphenyl:trsiiic, could be accommodated around an octahcdral c1ustc.r. hlorc rcwwt ly, othw hubstitutcld compouiids, formu- lated as Rh6(C'O)loIIG [ I, = l'l'h3, P (OMcj3, arid 4-(thyl-2,6,7-trioxa-l- phosphabicyclo[:!.2.']oc.tanc~ (ETPO) } have been isolated by reacting Rhb(CO)lGwith ('XCPSS ligand in hizcnc at 80°C in a scaled tubc (21). A black, sparingly solu1)lc dwivativca, t mtatively formulatchd as Kh6(COj8(PH3jshas bwn isolstcd from the rcaction of [ltli(CO)zC1]2 \\ith phosphinc (89). Ihdy, hot h ((20) 12 mid Itlit,(('0 j rvact with dicncs to give the spccic3s Rh6(Co) 11 (1)icwv j (Ijicnc. = 1 ,j-c).clr)octadiene, norhornadicbne, 1,4-cycloh(~xadicnc~,and 2,5-dimct hyl-1 ,3-butadicnc) . Tho molecular wight of lth6(CO j 11 ( 1,5-COIjj has bc~nconfirmed by mass spectroscopy (92). Thr rcaction of Rho(('0) 16 with a primary aminc gives the monoanionic dcrivativc containing n cwboamidc group, 11 hcrcas the analogous carbo- alkoxy derivatives liavc: hccn obtaincd by reduction with alkali metal all I~~~+[l~l~6(~~~l,(~o~IIIt)]-(I1 = 2-1'1, WI3U) 2KNIlr Ill)6(Co) I F, (30) Olt- .r:[ltll6(cY))~ ,(cOoR)]- (It = llc, 15t) Halides or pscwdohalidcs give the. analogous substitution products [1th6(COj1-,X]- (X = C'l, fir, I, CN, WNj [see Eq. (35)]. High Nuclearity Metal Carbonyl Clusters 329 In these cases also, initial iiuclcophilic attack on the carbonyl group sccms probable. Thr least stahlc of thcw derivatives is [Rhs (C'O) &I]- from which the chloride ion rail br rc\adily displaced by mctathcsis with other anions such as SCN- (45): [Iih,(eo,,,cl]-+ S('\ -[lihG(co),,(scs)l- + c1- (51) Thc corresponding acyl derivat ivos, [Ilhh(CO) COR) 1- (It = Et, Pr) , wwe originally obtain(-d as st a1)lv solutions of the hydroniuin salts by re- action of Rh4(CO)12with watrr in thr presence of a mixturc of carbon rnonoxidr and an olcfin such as c~thyleiicor propylene (43): CHaCOCIIa ~iii~,(co),~+ 2~~1-1, + 21~0 '2(11~~r)lv)++ 2[lrtr,(Co),,(Col~t,]-+ 2c02 + 2co (52) It SCC~Sprobablc that their fOrnlatio11 dcrivcs from olcfin addition to the intcmicidiate [Ith6(CO)15H]- [sw Eq. (Sl))]. On increa4ng thc halide coiicc.ntratioii, further substitution occurs to give thc disubstitutcd diaiiioris [Iihb(('0) lrX.3 (45, 101).Although there is spcctroscopic cvidmcc for t hr prcwncc in solution of thc species [Ilh,(~~l)1412~--,[RhG(CO)lI((~Oonlr)2]2-, and [Rh6(CO)11(CN)2]2-, only thc last has bren isolatrtl as a pure. crystalline product. The stability of the dicyanide drrivativc. prolxhly ariscs from thr baclibondiiig propcr- tics of such a substitucnt. Crc~ncmlly,thc disubstitutcd derivatives arc un- stable and undergo further traiwforniation to the heptanuclrar species [lih.,(<;O)16X]2- (x = Br, I) (101). Schcine 3 briefly summarizrs tht. synthcscs of thcsc substitutrd anions: Thc. chemistry of unsubstit ut rd polynuclear carboiiylrliodates is fasci- nating, but, although it has 1wn under investigation for sevcral years, it is as yet, clarified only in part. Thc dcrivatives best characterized arc [Rh12(CO)30l2- and [Rh,(CO),j]'-, Jvhich seein to be the focal points of such a chemistry. Thc violct dianion [Rh12 (( '0)3"12- is obtained in high yields (SO~l)07~) by reduction under carbon monoxide of [Rh(CXl)2C1]2 or Ith4(C0)12in 330 P. CHINI, G. LONGONI, AND V. G. ALBANO buffered alltaliiie solutioii, such as potassium acctatc (40).The reduction proceeds in agrecrneiit with the following stoichionir%riw (38,40) : 211h?(C0)4Cl~+ GCO -1 21120 -t Iilli(CO)12 + 2C02 + 4HCl (33) 311211(CO)1r + 2011- -lRI~i~(~O)~ij]'- + ('O? f 5C0 + IT& (54) Equation (54) corrcsponds to a inultistcy~ mc~chaiiisniin hich the kvy step is the redox condciisation [rtll,(co),,l~-f 2lillr(C'0),2 -[Itll,2(CoL,,1~- + co (3.5) This reaction has lieen thoroughly verifichd starting from the piire rcagcnts, but no c~xperimcntalevidcncc~ for the formation of the. cxpcctcd octaiiuclcar intermediate could be obtairicd (102). The tetraiiuclcar dianioii [lihl( CO) Ill2- has bccw prcpard from Sh4(CO)12 using a sequriicc of rcactions that is believd to lie aiialogous to that involved in thc process rcprcsmtcd by 1Sq. (34) : 11114(~Yl)12+ OCli 4- -[nll.l(co,,,(eoocr-r,~l- (56) [Itl~r(CO)II(COO('H1)1- + OIIP --t [Iih~(~'O~~~\'~+ (I02 f HLO (57) In the solid state, the dianion [Rh, (C'O) ll]?- po sscs a distorted tetra- hedral geoinctry with four tcrrninal arid s(ven cidgc.-bridging carboriyl groups, whereas in solution it is fluxional at -70°C (13, 74). An important, but riot yct undcrstood, roaction of thc dianion [Rh12(("0) 30]2- is the revrrsiblc addition of carbon monoxide. Accurate mcasurcments of the amount of absorbed gas indicatc. the stoichioinctry (40,102) : &"C, 1 ntm, TIIF [R1112(CO)?"12- + x:o - - [Itlll'(r(l)-3i1~- (58) v'iruum Although dianion [Rh12(CO)-3J- has never bwn isolated in the solid state owing to its lability, it can I)(> coiisidcrcd a definite intvrmediate. In fact, it has a characteristic and rc~produciblcIR spectrum, and at -70°C it has a peculiar magiictic resoriancc spectrum consisting of two doub- lets (191.7 and 208.3 ppin) and a rnultiplct (247 ppm) (74). High Nuclearity Metal Carbonyl Clusters 33 1 As shown in Scheme 4, thc chemistry of the violct dianion [Rh12(CO)30]2- is relatively simple: the main fcaturcl is the easr of rupture of the Rh-Rh bond connecting the two octahcdrn. The best synthesis of the green hexa- nuclear dianion [Kh,j( CO) 1512- involvos the reduction of ILh6(CO)16under nitrogrn with a stoichiometric amount of alkali (99) : Nz, Me011 Iil16(~o),fi+ 40H- - [lths(C0)1a12-+ cor + 2H,O (59) This should bc contrasted with the reduction undrr carbon monoxide, which is complicated because of concurrent degradation and redox con- densation reactions [reprcscnted by Eqs. (6) and (14)] arid which gives rise to the heptanuclear anion [I& (C'O) 16y-. Dianioii [Rh6(CO)15]2- is qnitcl rvactive, as shown in Scheme 5. Most of thrsc reactions, as wcll as thr. prckparations of the [Rh7(CO)16?- and [Ith6(CO) 14y-anions, Eq. (3) and (24), haw already bccn discussed in previous parts of this review. Although the [Rh12(CO)so]z- dianion is modcratcly air-stable, solutions of thc. more reduced anions [Rh6(CO) 13y-, [Rh7(CO) 16]3-, and [1Zh6(CO) 1414- arc cxtrernely air-sensitive. Occasionally during the synthesis of the [IthT(CO) trianion, low yields of a yellow anion, which was formulated on thc basis of elemental analyses as [Rh3 (CO)lo]-, ww also obtained (39). Subsequent X-ray structural investigation showcd it to bv [Ithe( CO) &I2- and, furthermore, suggested that its casual formation could be dur to the accidental presence in the reaction medium of trace amounts of chloroform (7).As a result, this compound is now readily available in high yields (80-90%) by de- liberate addition of small amounts of chloroform to the reaction mixture. The following reaction accounts for thc apparent stoichiometry : 332 P. CHINI, G. LONGONI, AND V. G. ALBANO The mechanism probably involves a multistep reaction, with initial formation of [Rh7(C0)l6-J3-, degradation to [Rh(CO) 4]-, and condensa- tion of this anion with the chloroform (10, 23). The [Rh6(CO)&]2- dianion is the precursor of a wide series of carbide-carbonyl derivatives. Scheme 6 shows the compounds so far structurally characterized and the necessary conditions for thcir syntheses (9: IS). Several other species have been isolated in the crystalline state and are presently awaiting definite characterization. Great difficulties arise from the fact that these derivatives are only sparingly soluble in inert solvents and often react with polar solvcnts [see for instance Eq. (37)]. The octa- nuclear carbide Rh,(CO j l& polymerizes THF at room temperature (102). XI1 IRIDIUM DERIVATIVES Unlike cobalt and rhodium, the chemistry of polynuclear iridium car- bony1 derivatives has not been studicd in detail (15a). Reduction of Irq(CO) 12 under carbon monoxide with K&O3 in methanol gives the yellow tetranuclear hydride derivative [Ir4 (CO)llH]-, whereas under nitrogen the brown dianion [Irs(CO)20]2- has been isolated as a tetraalkylam- monium salt (97). It has been suggestcxd that the structure of the dianion could result from the linking of two iridium tetrahedra, although its formu- lation so far is based only on elemental analyses. Clearly such an interest- ing compound deserves further chemical and structural characterization. The reduction of Ir4(CO)12by sodium metal in THF under carbon monoxide gives hexanuclear [Ire( CO) Its IR spectrum compares well with those of the analogous dianions [CO~(CO)~S]~-and [Rh6(CO)1512-. As previously shown [see Eq. (29) 1, dianion [IT6( CO) 15]2- reacts with acetic High Nuclearity Metal Carbonyl Clusters 333 acid under carbon monoxide to givc red crystals of Ir6(CO)16(96). This hexanuclear species is much less stable than Irl (CO) 12, and its chemistry has not yet been studied. Xlll NICKEL DERIVATIVES There is spectroscopic evidencc that the reduction of Ni(CO)4 gives rise to a large number of products, most of which have still to be isolated and characterized. Results obtaiiwd so far from a recent reinvestigation of this chemistry (95) suggest that all of the formulations previously reported in the literature are incorrect (76, 77, f 16). Reduction under nitrogen of tc%racarbonyl nickel with alkali metals or sodium and lithium amalgams (7f) in THF, or with alkali hydroxides in methanol, gives a mixture of the dianions [Ni,(CO)12]2- and [NiG(CO)1J2- (68, 95). The final composition of thv reaction mixture greatly depends on the experimental conditions owing to the easily reversed equilibrium : "1,(C0),2]2- + NI(C'o)4=== [S1,(CO),,]'- + 4co (61) The yellow pentanuclear dianion, [Ni, (CO)12]2-, is rather labile and has been isolated in a pure state only as the bis (triphenylphosphine) iminium salt by crystallization under carbon monoxide in anhydrous solvents. In wet solvents, it reacts readily with carbon monoxide to give a mixture of tetracarbonylnickel and an unstable hydride derivative presently formu- lated as [Ni(CO)3H]- (T = 18.3),by comparison of its IR spectrum with those of [Ni (CO)3x1- (X = C'I, Hr, I) (31). This reaction contrasts with that of [Ni,(C0)12]2- with water under nitrogen when the red dianion [Nis(CO) 1212- is formed. Thc. rvaction proceeds with formation of tetra- carbonylnickel, hydrogen and traces of carbon monoxide and its stoichi- ometry is believed to be the following: 3[N1s(C0)1,]'- + 2H2O -2[Y16(COj,i]'- + Hi + 20H- + aNl(co), (62) The mechanism should involve an ~asyinitial protonation to give a hy- dride intermediate which thm loscs hydrogen and condenses to [Nis(CO) 1~]~-. Equation (62) explains why prcxcipitation from the original reaction mixture by addition of an aqueous solution of tctraalkylammonium salts always gives the red hexanuclcar dianion [NiG(CO) 12]2- (yields up to 60%). A comparison of its IR spoctrum with that reported in the litera- ture'for [Ni4(C0)9]2- (76) strongly suggests that the latter should be re- 334 P. CHINI, G. LONGONI, AND V. G. ALBANO formulated as [&(CO) 12]*-. Our present kriowlcdge of the reactivity of this hexanuclcar dianion is summarized in Scheme 7. Dianion [Ni6(CO) 127- reacts slowly under vacuum with Ni(CO)4, to give the dark-red cnneanu- clcar [Ni,(CO) 1J2-. The latter may, however, be better synthesized by oxidation of the hcxanuclear complex with a stoichiomctric amount of Ni(Et0H)&'lz. Hydrolysis of [Ni6(cO)12]2- takes place only in an acidic medium arid gives, depending on thc pH, two hydride derivatives, presently formulated as [Nix(CO)n4H2]2-and [Nill(C'O)zoH2]2-, but not yet struc- turally characterized. Such formulations are bascd on elemental analyses and, in the case of violrt [Nill(CO)20Hz7-, also on a molecular weight calculated from unit cell md density measurements. It is also worth noting that the ease of hydrolysis decreases in the series: Nis > Ni6 > Nis > Nill, as expected for a progressively higher delocaliza- tion of charge and consequent lowering of the nucleophilicity of the cluster. Finally mixed-metal clusters with formulas [M2Ni3(CO)16]2- (M = Cr, Mo, W) and [M2Ni4(CO) 13 (M = Mo) have been isolated by Ruff by condensation of the corresponding dianions [Mz (CO) with Ni (CO) in refluxing THF (yields 30-687,) (1I1 ) . XIV PLATINUM DERIVATIVES An insoluble compound formulated as [Pt (GO)Jn has been obtained by Booth and Chatt both by hydrolysis of Pt (CO)&lz in benzene and by carbonylation of Na2PtCI4in ethanol (19, 20). More recently, reductive carbonylation of Na2PtCl6.6HzO in methanol, in the presence of alkali acetates or hydroxide, has been shown to proceed High Nuclearity Metal Carbonyl Clusters 335 according to the following reaction schc.me (29,95) : OH-, CO ow, CO [PtCl$ -[rt(co)cl,]- - [Pt(C(>)L]n or [i’t3(CO)6ln2- (7L > 6) OIL-, CO OII-, co OH-, CO -[PtlS(C0)38]2- -[€’t15~~Y))dol~- - olive-green yc~llo\\-grceii ow, CO OH-, CO [l’tl?(co)2412-- [ I’t ,((Y )), s]z- -[l’t ,(CO) 1212- blue-green Vl(J1f’t-I CO, MeOII 6[~t,~(co)~,,12-+ [~t~i~p- * 5[l’t1~((‘0)361~-+ [Pt(CO)Clr]- + 3C1- (63) The orange-red dianion [PtG(CO)12]2-, which has also been obtained by reductive carbonylation in concentrated mcthanolic sodium hydroxide solution, is better synthesized by reducing preformed [Pt3(CO)6In2- (n = 3, 4, or 5) with sodium or lithium metal in THF. Using Na/K alloy, the reduction goes still further to givc the unstable pink dianion [Pt3(CO)6]2- (VCO = 1950, 1745 cm-l), as yet not isolated in the solid state. The most striking peculiarity of all of these derivatives is their ability 336 P. CHINI, G. LONGONI, AND V. G. ALBANO to de- and reoligomerize by breaking or forming Pt-Pt bonds. In fact all of these clusters are drgraded not only by halides and by tertiary phos- phines [see Eq. (34)] but also, when n = 4, 5, or 6 by molecular hydrogen [see Eq. (40)]. This last reaction can be reversed by oxidation with air and the cycle repeated several times with simple formation of water. Finally, we should mrntion that, in thc light of these results, which established the first contact points between the chemistry of the carbonyl derivatives of nickel and platinum, some analogous palladium derivatives could be expected. Until now, howevw, all attempts to prepare them have failed giving only palladium metal under a variety of conditions (95). xv BONDING THEORIES The metal frameworks in high nuclearity clusters can be considered as finite parts of closc-packed metallic structures stabilized by thc external ligands and the ncgativt. charges. The apparent simplicity arising from the fact that in a mcJtal cluster there is a finitr number of metal atoms is offset by thc complication of considrring thr interactions among the combinations of U, a, and P* orbitals of the 'CO groups with suitable combinations of metal orbitals. For example, as pointrd out scmd times (24, %), metal- metal bonds are strongly mixed with the bridging carbonyl groups. Owing to the complexity of thme intwactions, the number of independent pa- rameters to be used in a sufficicntly accurate LCAO-MO calculation is generally so high as to be well out of the rangr of the most modrrn com- puters. Other more advanccd calculations, particularly devised for clusters of pure metals (114),have not yrt been used in this area. This situation has forced us to compare the experimental data with morc-or-less empirical hypotheses, each one of which has its advantages and has enjoyed some popularity. So far the following four different ap- proaches have been proposed: (a) noble gas rule (112); (b) topological hypothesis (88); (c) relationship between metal clusters and polyboranes and carboranes (120, 121); and (d) LCAO-MO treatments limited to the metalatoms (24,26,34,35,105,111). The rare gas rule is based on the assumption that bonding interactions are localized, or, more exactly, localizable, along the directions of minimum distance, namely, thc edges of the polyhedron; therefore, M-M bonds can be represented by two centcr-one electron pair bonds. For a metal cluster the rule then takcs thc form High Nuclearity Metal Carbonyl Clusters 337 where X3is the numbcr of valoncc~shell electrons, given by the number of outer electrons of all of the inrtal atoms plus the number of electrons do- nated by the ligands and the cvtwtual negative chargr (when a hcteroatom is present in the cage, it is considcrcd to donate all of its outer cllectrons to the cluster). The iVl is thc nunilwr of metal atoms, and dY2 is the number of polyhrdral edges. First, it should be noted that, although almost all clusters with 3 or 4 metal atoins have the favored 18 outcir c.lcctron, inert gas configuration on each metal atom (34),Table VI shows that only a few penta- and hexa- nuclear clusters and almost norw of higher nuclearity obey this rule. Second, t he rule requires both an unoxplained topological correspondence between the number of polyhedral cdgw and the number of low-energy metallic molecular orbitals, and a bond order of 1 for each metal-metal bond. We gain, therefore, a picturc. of nwtal cluster bonding quite different from that envisaged in an infinite. mcltal structure, even if the M--M bond distances found in such a strricture arc' sometimcls shorter by only a few hundredths of an Angstrom from those found in clusters that obey the rule. It obviously follows that th(1 noblc gas rule is an empirical and formal one and that the electron drnsity efcctively concentrated along a poly- hedral edge must be less than rcquircd by the rule. Neverthelcss, deviations are quantitatively small and thc rule is often a useful formalism. The second approach has brcn proposed by Kettle (88) and is based on an empirical topological correspondence between thc sum of edges and tri- angular faces and the number of bonding skeletal orbitals. This approach tries to take into account, in a qualitative way, the additional possibility of interactions between skclrtal and ("0orbitals, although in practice it does not provide a rdiable picture of the cluster. For example, Kettle found RhG(C0)16 to be 2 elcctrons short and formaw the possibility of synthesizing the anion [Rh,(C'O) IGl2-, a prediction that has not been fulfilled (99). Recently, Wade has pointed out a formal analogy between the elec- tronic structures of carboranes and polyboranes and those of metal carbonyl clusters based on the assumption that certain triangulated polyhedra re- quire the same numbt.r of skeldal orbitals whether there are BH or CH units as well as transition metal atoms at their apiccs (120). This assump- tion is quite reasonable as the syntlmis of a large number of polyboranes and carboranes in which a transition mdal atom takes the place of a BH skeletal unit may be carried out (70). According to this approximation, in a metal carbonyl cluster vach metal atom, like a BH unit, may effectively contribute only three orbitals to the skeleton, whereas the remaining six are left for bonding with the external ligands and for lone pairs, and arc assumed to be completely filled. The TABLE VI 0 a3 A COMPARISONOF THE NOBLEGAS RULE WITH WADE’SRULES FOR HIGHNUCLEARITY METAL CARBONYL CLUSTERS Noble gas Wade’s Ni = rule, relation- Wade’s Geometry of the cluster metal Nz = N3 = 18N1 - ship bond (from X-rays) Example Ref. atoms edges electrons 2N2 1411‘1 +z a order nP Trigonal bipyramid [Ni5(COhIz- 95 5 9 76 72 72 - Square pyramid Feb(CO)15C 26 5 8 74 74 74 0.87 5i (nido) Bicapped tetrahedron Oss(CO)la 104 6 12 84 84 84 0.50 0 (capped) Bicapped rhombus [MozNic (CO)I,Iz- 111 6 13 82 82 86 P Trigonal prism C&(CO)i&7”- 7 6 9 90 90 86 0 CPt6 (co)1’21’- 29 6 9 86 90 86 0.78 Octahedron 1Lh6(CO)16 53 6 12 86 84 86 0.58 @ Capped octahedron CRh, (CO)16 13- 4 7 15 98 96 98 0.47 Z (capped) 0 100 96 98 0.53 5 (capped) 0 Bicapped trigonal prism Rhs(C0)lgC 12 8 15 114 114 114 0.6 Square antiprism [COS (CO),aC1’- IS 8 16 114 112 114 0.56 F Condensed trigonal prism [Pt9(Co!l,p- 29 9 15 128 132 128 0.67 P CPt1z(CO)2412- 95 12 21 170 174 170 0.62 0 [Pt15(CO)3032- d9 15 27 212 216 212 0.59 Two connected octahedra [Rhiz(CO) 301’- 5 12 25 170 166 170 0.52 Distorted tetracapped cube lEhl~(CO)2~(C-C) 13 12 27 164 162 170 Pentagonal tetracapped prism Rh,,(CO)~~(C)~- 9 15 31 200 208 198 (Ni = 14) Closo, z = 2; Dido, z = 4; arachno, x = 6; capped closo, x = 0. High Nuclearity Metal Carbonyl Clusters 339 analogy with the triangrilat cdpolyboranes arid carborancs then requires that, in the case of a “closo” cluqtcr containing lYl metal atonis, only -I’l + 1 electron pairs are accomniodatrd in thc skeletal bonding orbitals obtained from the 3N1 AO’s; but accornniodatioii of AT1+ 2 or Sl+ 3 electron pairs should cause cagc opening to “nido” and “arachiio” polyhedra. hIoreovcr capping of a triangular facci to give :I cappcd “closo” polyhedron (for in- stance [Ith7(CO) 16]3- of struct urv XYI in Fig. 7) would rcquirc. only Llrl skeleton electron pairs (l2ln).Thus, thr Wadc rc.latioiiship for siiiiple "close"-triangulated polyhcdra rcwilts in the niathtmatical expression LV, = 12,Vl + 2S1 + 2 = 14N1 + 2 (65) Unfortunately, application of this rulc is severely limited by the fact that the corrcspondencr of polj hdra hetwccn polyborancs and carborancs with metal clusters is confined to the square pyramid, the trigonal bipyra- mid, and the octahedron; morcovc’r, the rule fails to account for the 76 clcctrons of the known trigoiial bipyramidal clusters (Table VI) . Indiscrimiiiatc application of thr ride shows it to be in fairly good agree- ment with most of the high nuclmrity clusters (Table VI) . Furthcrmore, the metal-metal bond ordtm calculated in this way compare much niorc reasonably with that expectcd in t he purr metals. However, aftcr a first glancc, it 1)ecomes evidciit that this generalized version of the rulc also dom iiot match the experimental data sufficiently well. For cxaiiiple, the bond orders derived from it do riot agrrc ~iththe constant Pt-Pt distance in thcl scrips [Pt3(CO)6]$- (n = 2, 3, 4, 5) (29, 95); this behavior apparently rcquircs that at least one of the AV1 + 1 electron pairs allocated to tlic mctal-metal bonding framework be trans- ferred to a nonbondirig situatioii. Still more relevant is the fact that the gmwalized form of the rulc, bviiig dcpendent only on the whole riuniber of metal atoms, does not discriminat(. between different geometries, mid, therefore, is of very limited predictive capacity. For instance, with S metal atoms and 114 clectrons both the bicappcd trigonal prism arid the square aiitiprisni arc known, and with 12 mctal atoms and 170 clectrons both the condensed trigoiial prism and thcx douhlr octahedron have been found. The fourth approach, the LC’AO-RIO treatment, may potentially give the best insight into thc mrtal rlustcr bonding problem, but its accuracy is limited by the usual assumption of indrpcmdcnccl bctwwn skclctal and ligand orbitals. In favor of this assuinptioii thrrc. is, as already discusscd, thr constant number of val(wrc~dectrons for the octahedral clusters (86 elcctrons) , which is maintaincd throughout a concurrcnt variation in car- bon monoxide coordination (from thrw to five). This behavior shows that, as a first approximation igrioririg the significant distorsions of the octa- hedral core, the separation of MO’s at the frontier is independent of both 340 High Nuclearity Metal Carbonyl Clusters TBBLE VII EXAMPLESOF THE SIMPLESTLCAO-hfO APPROXIMATIONIN HIGHNUCLE \RITY METi~ CARBONYLCLUSTERS A0 for lone AO for pairs and A0 for bonding back skeletal Compound with CO’sa donation bonding Ref. Rh6(C0)16and [coG(co)lS]z- SP3 cL,,dz2-y~ dzz, duq dn 34, 35 CCos(CO)lr14- asp3 dZy(ord12-u2) dzz,d,,, dz2 24 Skeletal MO’s in Oh symmetry : dsz = A1g + En + TI, d,,,, = TI, + Tzg + Tau + TI, These arbitrary divisions have been made in agreement with the stereochemistry of the bonds around each metal center and with the particular choice of Cartesian axes. the number and type of carbonyl groups. It seems, therefore, reasonable to assume indcpendencc between skeleton and ligand MO’s if only a first-order justification of thc electronic distribution is dcsiwd. In the simplest form of LCAO-MO treatment the metal orbitals are arbitrarily divided into thrce groups, as shown in Table VII. The A1,(z2), T2g(zz,yz) and Tlu(xz,y;.) of Table VII are bonding combinations (54)’ and accommodate 14 electrons. The remaining 72 electrons are accommo- dated in the thirty-six othcr atomic orbitals (or thcir suitable com- binations). A similar qualitativc approach, although with a less rigid separation be- tween the starting atomic orbitals, has been applied by Dahl to Fe5(CO)& and [M,N&(CO)~G]’-(h1 = C‘r, NIo, w) (26,111). More rcccntly, Mingos has gone further in treating the isolated COGcage of the [Co6(CO) 14]4- anion in a rigorous way (105). On the basis of orbital overlap calculations, the fifty-four cobalt AO’s arc’ combined to give thirty- one bonding or weakly antibonding and twenty-three strongly antibonding skeletal orbitals. From thr point of view of the isolatcxd CoGcage this order- ing is both interesting and probably accurate, but, from the [COE,(CO)U]~- cluster point of vim therc arc similar limitations as before. In fact, metal MO’s and ligands MO’s arc’ still treated scparately and their reciprocal interactions have been estimated only by simple inspection. The final ordering of MO’s prcdicts a frontier separation of about 80,000 cni-’, which is a rather surprising result for a dark red-brown colored cluster. Finally, we should notc that thc 1,CAO-MO approximations profit from the high symmctry of some species and that application to less symmetrical High Nuclearity Metal Carbonyl Clusters 34 1 or uiisymmctrical clusters such as R h, (CO) and Rhlz (CO)25 (C-C) is quite difficult. ACKNOWLEDGMENTS We wish to acknowledge the contribution of all the co-workers cited in the references, and to mention particularly Dr. S. Martinengo, Prof. L. F. Dahl, and Dr. B. T. Heaton. Direct financial support was obtained from Consiglio Xaxionale delle Iticerche, Accad- emia dei Lincei, the Royal Society, SATO, and the University of IVisconsin. I<.EFERENCES 1. Albano, V. G., Chini, P., and Scatturin, V., J. Chem. Soc. Chem. Commun. 163 (1968). 2. Albano, V. G., Chini, P., and Scatturin, V., J. Organometal. Chem. 15, 423 (1968). 3. Albano, V. G., Bellon, P. L., Chini, P., and Scatturin, V., J. Organometal. Chem. 16, 461 (1969). 4. Albano, V. G., Bellon, P. I,., and Ciani, G. F., J. Chem. SOC.,Chem. Cornmun. 1024 (1969). 5. Albano, V. G., and Bellon, P. I,., J. Organometal. Chem. 19, 405 (1969). 6. Albano, V. G., Bellon, P. L., and Sansoni, M., J. Chem. SOC.A 678 (1971). 7. Albano, V. G., Sansoni, M., Chini, P., and Martinengo, S., J. Chem. Soc., Dalton Trans. 651 (1973). 8. Albano, V. G., Ciani, G. and Chini, P., J. Chem. Soc., Dalton Trans. 432 (1974). 9. Albano, V. G., Chini, P., Martinengo, S., Sansoni, M., and Strumolo, D., J. Chem SOC.,Chem. Commun. 299 (1974). 10. Albano, V. G., Chini, P., Martinengo, S., McCaffrey, U. J. A., Strumolo, D., and Heaton, B. T., J. Amer. Chem. SOC.96, 8106 (1974). 11. Albano, V. G., Ciani, G., Martinengo, S., Chini, P., and Giordano, G., J. Organo- metal. Chem. 88, 381 (1975). 12. Albano, V. G., Sansoni, M., Chini, P., Martinengo, S., and Strumolo, D., J. Chem. Soc., Dalton Trans. 305 (1975). 13. Alhano, V. G., Chini, P., Ciani, (i,, Martinengo, S., Sansoni, pvl. and Strumolo, D., unpublished results. Albano, V. (i., Ccriot,ti, A,, Chini, P., Ciani, G., Martinengo, S., and Anker, M., J. Chrm. Sor., Cheln. Commzin. 859 (1975). 14. Anders, U., and Graham, W. A. (+.,J. Chem. Soc., Chem. Commun. 291 (1966). 15. Anders, U., and Graham, W. A. G., J. Amer. Chem. SOC.89, 539 (1967). 15a. Angoletta, M., Malat,esta, L., and Caglio, G., J. Organometa2. Chem. 96,99 (1975). 16. Baird, M. C., Progr. Inorg. Chenz. 9, 1 (1968). 17. Bellon, P. L., Manassero, M., and Sansoni, M., J. Chem. SOC.,Dalton Trans. 2423 (1973) and references t,herein. 18. Biryukov, B. P., and Struchov, Y. T., Russian Chem. Rev. 39, 789 (1970). 18a. Bodner, G. M., and Todd, L. J., Inorg. Chem. 13, 1335 (1974). 19. Booth, G., Chatt,, J., and Chini, P., J. Chem. Soc., Chem. Commun. 639 (1965). 20. Booth, G., and Chatt, J., J. Chem. SOC.A 2131 (1969). 21. Booth, B. L., Else, M. J., Fields, R., and Haszeldine, R. N., J. Organometal. Chem. 27, 119 (1971). 22. Booth, B. L., Else, M. J., Fields, It., Goldwhite, H., and Haszeldine, R. N., J. Organometal. Chem. 14, 417 (1968). P. CHINI, G. LONGONI, AND V. G. ALBANO 23. Boon Keng, T., Hall, &I. B., Fenske, It. S., and Dahl, L. F., J. Organometal. Chem. 70, 413 (1974) and references therein. 24. Biaterman, P. S., Structure Bondzng (Berlzn) 10, 57 (1971). 25. Braterman, P. S., “hZetal Carbonyl Spectra.” Academic Press, New York, 1975. 26. Braye, I<. Il., IMil, I,. F., Hubel, W , and Wampler, 11. L., J. Amer. Chem. SOC. 84, 4633 (196%). 27. Brown, D. L. S , Connor, J. A., and Skinner, 11. A , ,J. Chem. SOC.Faraday Trans. Z 71 (3), 699 (197-5) 28. Cal:tbrese, J. C , Ilshl, L. F., Cavalieri, A, Chini, P., Longoni, G., and hlar- tmcngo, S., J. Amer. Chem. SOC.96, 2616 (1974). 29. Calabiese, J. C., Ualil, L. F., Chini, P., Longoni, G., and hlartinengo, S., J. Amer. Chetu. SOC.96, 2614 (1974). 30. Caitner, -4.,Itobinsori, U., and Gardner, P., J. (’hem. SOC.,Chem. Commun. 317 (1973). 31. Cassar, L., and FoA, hf., Inorg. Nucl. Chem. Lett. 6, 291 (1970). 32. Chaston, S. H , and Stone, F. G. A, J. Chem. SOC.A 500 (1969). 33. Chini, P., J. Chem SOC.,Chem. Commun. 29 (1967). 34. Chini, P., Znorg. Chzm. Acta Rev. 2, 31 (1968). 35. Chini, P., and Albano, V. G., J. Orgnnotnetal. Chem. 15, 433 (1968). 36. Chini, P., Znorg. Chetn. 8, 1206 (1969). 37. Chini, P., Albano, V. C+, and Martinengo, S., J. Organometal. Chenz. 16, 471 (1969). 38. Chini, P., and Martinengo, S., Znorg. Chzm. Acta 3, 315 (1969). 39. Chini, P., and hlartinengo, S., J. Chem. Sac., Chem. Commun. 1092 (1969). 40. Chini, P., arid Rlartinengo, S., Inorg. Chim. Acta 3, 299 (1969). 41. Chini, P., Pure Appl. C‘hem. 23, 489 (1970). 42. Chini, P., and Longoni, G., J. Chern. SOC.A 1342 (1970). 43. Chini, P., Rlaitinengo, S., and Garlaschelli, G., J. Chem. SOC.,Chem. Commun. 709 (1972). 44. Chini, P., Cavalieri, A., and Martinengo, S., Coord. Chem. Rev. 8, 3 (1972). 45. Chini, P., hlartinengo, S., and Giordano, G., Gazz. Chzm. Ztal. 102, 330 (1972). 46. Chini, P., Martinengo, S., Strnmolo, U., Albano, V. G., and Sansoni, hf., Abstr. 165th Meeting Amer. Chem. SOC.,57, 1973. 47. Chini, P., hlartinengo, S., McCaffrey, D. J. A,, and Heaton, B. T., J. Chem. SOC., Chenz. Commun. 310 (1974). 48. Churchill, M. It., Worinald, J., Knight, J., and Mays, M., J. Cheni. Soc., Chem. Comw~m.438 (1970). 49. Churchill, M. It., and Wormald, J., J. Chem. SOC.,Dalton Trans. 2410 (1974). 50. Churchill, pvl. lL, Bemian, S. A., Osborn, J. A,, and Wormald, J., Znorg. Chem. 11, 1818 (1972). 51. Churchill, M. R.,and Bezman, S. A., Znorg. Chem. 13, 1418 (1974). 52. Connor, J. A,, Skinner, €1. A., and Virmani, Y., J. Chem. SOC.,Faraday Trans. I 1574 (1972). 52a. Connor, J. A., Skinner, H. A,, and Virmani, Y., Faraday Sump. Chem. SOC.8, 18 (1974). 53. Corey, E. R., Dahl, L. F., and Beck, W., J. Amer. Chem. SOC.85, 1202 (1963). 54. Cotton, F. A., and Haas, T. E., lnorg. Chem. 3, 10 (1964). 55. Cotton, F. A,, Kruckzinski, L., Shapiro, B. L., and Johnson, I,. F., J. Amer. Chem. Soc. 94, 6191 (1972). High Nuclearity Metal Carbonyl Clusters 343 56. Darensbourgh, D. J., Conder, TI. I,., Darensbourgh, RI. Y., and Hasday, C., J. Amer. Chem. SOC.95, 5919 (1973). 57. De Beer, J. A,, Haines, 11. J., Grralrex, I<., and vanWyk, J. A., J. Chetn. Soc., Dalton Trans. 2341 (1973) and references therein. 58. Eady, C. I<., Johnson, B. F. G., and Lewis, J., J. Organometal. Chem. 37, C 39 (1972). 59. Eady, C. li., Johnson, B. F. G., Lcwis, J., and Ahtheson, T., J.Organometal. Chem. 57, C82 (1973). 60. Eady, C. 12., Johnson, B. F. G., :ind Lewis, J., J. Organonzetal. Chenz. 57, C84 (1973). 61. Eady, C. R., personal comniunicaliorr. 62. Ercoli, R., and Rarbieri-Heririitte, F., Red. .4cad. Lincei, C1. Sci. Fis. Il’at. 16, 249 (1940). 63. Ercoli, I<., Santanibrogio, I<>., :mtl Tetlanirtnti-Casagraiide, G., Chim. Ind. (Milano)44, ,565 (19G2). 64. Evans, J., Johnson, 13. F. G., Jmvis:, .J., Norton, J. IL., and Cotton, F. A., J. Ciiem. SOC.,Chem. Commun. 807 (1973). 65. Ferguson, J. A,, and Meyer, T. ,J., J. rlwer. Chem. SOC.94, 3409 (1972) and references therein. 66. Fieldhouse, S. A,, Freeland, €3. H., RZiinn, C. D., and O’Brien, 1:. J., J. Chetn. Soe., Chew Commitn. 181 (1970). 67. Flitcroft, N., and Leach, J., J. Oiycrnotnetal. Chem. 18, 367 (1969). 70. Grimes, li. N., Ann. N. Y. Actcd. Sci. In press. 71. Guerrieri, F., and FOB,M., personal coriiinrinication. 72. Cuss, J. M., Mason, It., Thoinas, K. RI., van Koten, G., and Noltes, J. G., J. Organometal. Chenz. 40. C79 (1072). 73. Haines, L. I. B., and Poe, A. .I., J. Cheni. Soe. A 2826 (1969). 74. Heaton, B. T., Towl, D. C., Clhi, P., Fumagalli, A,, McCaffrey, D. J. A., and Martinengo, S.,J. Chem. Soc., Chetn. Comnzun. 523 (1975). 74a. Heaton, B. T., Chini, P., kIartinengo, S.,McCaffrey, D. J. A., and Fuinagalli, A., unpublished results. 75. Hieber, W., and Lagally, IT., Z. Atiorg. Allg. Chenz. 251, 96 (1943). 76. Hieber, W., and Ellennann, J., Z. A‘citirrforsch. B 18, 595 (1963). 77. Hieber, W., Ellermann, J., and Ziilin, I<;., Z. Naturforsch. L3 18, 589 (1963). 78. Hieber, W., and Shubert, I?. II., Z. Aiiorg. AZZg. Chem. 338, 32 (19G5). 79. Hsieh, A. T. T., and Mays, Ill. J., ,I. Organontetal. Chern. 37, C53 (1972). 80. Ingle, W. M., Preti, G., and M:tcl)iarniitl, A,, J. Chert?,. SOC.,Chern. Commun. 497 (1973). 81. Johnson, €3. F. G., Johnston, 11. l),, and Lewis, J., J. Chem. Soe., Chem. Conzmun. 1057 (1967). 82. Johnson, B. F. G., Johnston, It. U., Lewis, J., Robinson, 13. H., and Wilkinson, G., J. Chenz. SOC.A 2856 (1968). 83. Johnson, B. F. G., Johnston, Lt. l)., aiid Lewis, J., J. Chem. Soc. A 2865 (1968). 84. Johnson, B. F. G., Lewis, J., and Iiobinson, P. W., J. Chenz. Soc. A 1100 (1970). 85. Johnson, B. F. G., Lewis, J., and Willixms, I. G., J. Chem. Soc. A 901 (1970). 86. Johnston, 11. D., Advan. Inorg. Cheni. Rndiat. 14, 471 (1971). 87. Kaesz, 11. I>., Chem. Brit. 9, 344 (1973). 88. Kettle, S. F. A., J. Chem. SOC.A 314 (1967). 89. Klanberg, F., and Muetterties, Id. I,., J. 47ner. Cheni. Soe. 90, 3296 (1968). 344 P. CHINI, G. LONGONI, AND V. G. ALBANO 90. King, R. B., J. Amer. Chem. Soc. 94, 95 (1972). 91. King, R. B., Progr. Inorg. Chem. 15, 288 (1972). 92. Kitamura, T., and Joh, T., J. Organometal. Chem. 65, 235 (1974). 93. Kristoff, J. S., and Shriver, D. F., Inorg. Chem. 13, 499 (1974) and references therein. 94. Kruck, T., Hofler, M., and Noack, M., Chem. Ber. 99, 1153 (1966). 95. Longoni, G., Chini, P., Martinengo, S., Dahl, L. F., Lower, L. and Calabrese, J. C., unpublished results; Longoni, G., Chini, P., Lower, L. D. and Dahl, L. F.,J. Amer. Chem. Soc. 97, 5034 (1975). 96. Malat,esta, L., Caglio, G., and Angoletta, M., J. Chem. Soc., Chem. Commun. 532 (1970). 97. Malatest,a, L., and Caglio, G., J. Chem. Soc., Chem. Commun. 420 (1967). 98. Martinengo, S.,Chini, P., and Giordano, G., J. Organometal. Chem. 27, 389 (1971). 99. Martinengo, S., and Chini, P., Gazz. Chi?. Ital. 102, 344 (1972). 100. Martinengo, S., Chini, P., Albano, V. G., Cariati, F., and Salvatori, T., J. Organo- metal. Chem. 59, 379 (1973). 101. Martinengo, S., Chini, P., Giordano, G., Ceriotti, A,, Albano, V. A. G., and Ciani, G., J. Organomeld. Chem. 88, 375 (1975). 102. Martinengo, S., Chini, P., Funiagalli, A,, and Strumolo, D., unpublished results. 103. Mason, R., and Robinson, W. It., J. Chem. Sac., Chem. Commun. 468 (1968). 104. Mason, It., Thomas, K. M., and hlingos, D. M. P., J. Amer. Chem. SOC.95, 3802 (1973). 105. Mingos, D. M. P., J. Chem. SOC.,Dalton Trans. 133 (1974). 106. Muetterties, E. L., and Wright, C. M., Qxart. Rer. 21, 109 (1967). 107. Noak, K., J. Organometal. Chem. 7, 151 (1967). 108. Penfold, B. R., Perspect. Struct. Chem. 2, 71 (1968). 109. Penfold, B. It., and Itobinson, B. H., Accounts Chern. Res. 6, 73 (1973) and references therein. 110. Ruff, J. K., Inorg. Chern. 7, 1818 (1968). 111. Ruff, J. K., White, It. P., and Dahl, L. F., J. Amer. Chem. Soc. 93, 2159 (1971). 112. Sidgwick, N. V., and Bailey, 11. W., Proc. Roy. Soc., Ser. A 144, 521 (1934). 113. Sirigu, A., Bianchi, M., and Benedetti, E., J. Chem. Soc., Chem. Commun. 596 (1969). 114. Slater, J. C., and Johnson, K. H., I’hys. Rev. B5, 844 (1972). 115. Spiro, G. T., Progr. Inorg. Chem. 11, 1 (1969). 116. Sternberg, H. W., Markby, It., and Wender, I., J. Amer. Chem. Soc. 82, 3638 (1960). 117. Stewart, It. P., Anders, U., and Graham, W. A. G., J. Organometal. Chem. 32, C49 (1971). 118. Todd, L. J., and Wilkinson, J. It., J. Organometal. Chem. 77, 1 (1974). 119. Todd, L. J., and Wilkinson, J. R., J. Organometal. Chem. 80, C31 (1974). 120. Wade, K., J. Chem. Soc., Chem. Commun. 792 (1971). 121. Wade, K., Znorg. Nucl. Chem. Lett. 8, 559 (1972) and references therein. 121a. Wade, K., personal communication. 122. Wrighton, M., Chem. Rev. 74, 401 (1974). Free Radicals in Organometallic Chemistry M. F. LAPPERT and P. W. LEDNOR* School of Molecular Sciences Univerrify of Sussex Brighfon, England I. Introduction . . 345 11. Lletal-Centered Organometallic Radicals . . 349 A. Maiu Group Element Compounds , 3.52 B. Transition Metal Compounds . , 363 111. Other Organometallic Radicals . . 367 IV. Bimolecular Homolytic Substitution (Sn2) at the Metal Center of an Organometallic Substrate . . 370 A. Main Group Element Compounds as Substrates . . . 371 B. Transition Metal Compounds as Substrates . . 373 V. Addition or Iclimination Radical Reactions . , 381 ,4. Oxidative Addition of Alkyl Halides or Related Reagcnk and Rcductive Elimination , 381 B. Metal Alkyl Photolysis or Ttiermolysis . 388 VI. Appendix . . 390 References . . . 392 I INTRODUCTl ON It is becoming apparent that much organometallic chemistry overlaps with that of free radicals. In this context, an organometallic compound is taken as having at least one metsl-carbon bond. The object of this review is to provide a classification of the area, discuss briefly in very general terms the experimental methods, and offer a guide to the relevant litera- ture. We do not claim to cover this comprehensively but aim to give a representative overview, with emphasis on features of interest to the or- ganometallicist rather than the ESR or kinetics specialist. We propose to deal with a relatively small number of systems in detail and point to others more superficially. In keeping with the aim of this volume, we shall draw extensively on examples from niir own work; we do this with some temerity, because our direct involvemrrit in this field is of recent origin (since 1971). We propose the following classification : (i) organometallic paramagnetic compounds (Sections I1 and HI), and (ii) organometallic reactions, in * Present address: Institut fur ..\norganisclie Chemie der Universitat hlunchen, 8 Munchen 2, Meiserstrasse 1, Germany. 345 346 M. F. LAPPERT AND P. W. LEDNOR which paramagnetic transient species, usually organic free radicals, are implicated in the reaction pathway (Sections IV and V) . Class (i) may be divided into compounds in which the unpaired electron is localized principally on the metal M (I) (see Section 11); or remote from the metal (11) (see Section 111). The former belong to a wider group, the metal-centered radicals, encompassing not only, for example, the or- ML L’M-. (1) (11) ganometallic &Mes, but also the inorganic $iC13. Corresponding para- magnetic transition metal complexes (I) are rarely named as radicals because they are often stable, and the term “radical” has traditionally had the connotation of a species of low kinetic stability. Class (ii) is conveniently separated into free radical reactions involving (a) substitution at the metal center, SH2, or bimolecular homolytic sub- stitutions at M (by analogy with SN2, etc.) (see Section IV), for example, LM-R + 2-LM-X + ft (1) (b) addition or elimination at M (see Section V, A and B) , for example, ML + R-LMR (2) and (c) a site remote from the metal. Although the first direct demonstration of the existence of a transient alkyl free radical involved an organometallic system of the type shown in Eq. (a), [Paneth’s experiments on the pyrolysis of PbMe4 and the rccom- bination of Me with metallic Pb, Zn, Sb, Bi, or Be (158)], the topic of free radical organometallic chemistry remained largely neglected except by the gas kineticist until the last decade. This may be demonstrated by reference to the standard general text (1967-1968) on organometallic chemistry (35, 36). There is no subject index entry for “radical,” “free radical,” or “homolysis,” although Paneth’s and related experiments were discussed, as were also, in free radical terms, the mechanism of thermal decomposition of HgRz OT of alkyls or acyls of Cu, Ag, and Au, and the addition of R3Sn-H (via hR3) to olefins or other unsaturated substrates. Even the autoxidation of metal alkyls MR, was then regarded as polar in nature, although the intermediacy of peroxides, such as ROOMR,-I, was recognized. Radical anions, such as CloHs’, are not within our terms of reference unless they are metal-centered, e.g., the stable [BAr3]: (S5). The existence of nymerous transient metal-centered radicals, such as &Me, PbMe3, or SbMez, was inferred from gas kinetic data, e.g., on thermolysis of HgMez, PbMesH, or SbMe3 (74, 75). Electron spin resonance spectroscopy has made a considerable impact on the study of organometallic free radical chemistry. Four significant de- velopments may be identified : (i) the introduction of steady-state photoly- Free Radicals 347 sis techniques for the detection and identification of transient radicals; (ii) various methods for determining the rates of radical reactions in solution, by measuring relative radical concentrations at different times using com- petition with a radical reaction of known rate; (iii) the use of spin traps, such as Bu'NO, to intercept a transient radical R as the kinetically more stable spin-trapped derivative, such as the nitroxide (111); and (iv) the attachment of a stable organic frcc radical, usually a nitroxide, a spin label, that can interact with a reaction center, as in structure IV for co- enzyme BIZ-controlled enzymatic isomerizations (19.3).A different mag- But R 'N' 'N' I' 0 netic resonance technique, which has some application in organometallic free radical chemistry, employs lH or, in principle, 13C NMR spectroscopy and the observation of enhanced absorption or emission as a consequence of chemically induced dynamic nuclear polarization. Proper interpretation of CIDNP (Chemically induwd dynamic nuclear polarization) spectra allows thc detection of radical pair processes. Other experimental methods, diagnostic for establishing radical path- ways, are broadly kinetic or stereochemical in origin, being based on regio- or stereoselectivity. The effect of the addition of a known radical initiator, such as azobisisobutyronitrile, AIBN (V) , or of a radical inhibitor, such as galvinoxyl (VI) , upon the rate of a reaction may be informative. Sim- But Bu' Me,C--N=N-CMe, I I CN CN But (V) (VI) ilarly, the composition of a copolymer obtained by addition of two mono- mers, such as PhCH=CH2 and CH.F==C(M~)CO~MC,to the reaction 348 M. F. LAPPERT AND P. W. LEDNOR mixture may implicate free radicals as initiators for the copolymerization since the copolymer composition is related to the relative propagation rates for each homopolymerization, and t,hese rates are distinct and differ- ent for R, R+, or R- initiation. In such experiments, as with spin trapping, great care must be taken to establish by control experiments that the added chemicals do not alter the course of the reaction that takes place in the absence of such addenda. The participation of a, specific free radical in a particular reaction may be inferred from product analysis: for instance, many alkyl radicals h undergo competing processes of radical combination and disproportiona- tion, e.g., for R = Et, and the ratio of appropriate constants (IcJkbj for gas phase reactions of many radicals is known (122). Trivalent, neutral, carbon-centcrcd radicals arc generally either planar or slightly pyramidal (unless thcre are highly clectroncgative substituents, as in the pyramidal F&j with a low inversion barrier. Consequently, most radical reactions, in which the alkyl reactant is optically active with chirality at the a-carbon, lead to raccmic products bccausc free radicals have relativrly long lifetimes comparcd with their relaxation with respect to nearest neighbors. Whereas thr formation of a raccmic product is con- sistent with a radical mechanism, it is not a necessary precondition; the possibility of stercoselectivity via radical reactions has long been recog- nized, especially in the context of geminatc combinations of “caged” radical pairs which are among the fastest of chemical reactions (131~).A further caution relates to the temptation to generalize from an observation on an optically active cornpound to a whole class of reaction. In the context of organometallic chemistry, optically stable chiral centers other than carbon are at this time found only for silicon, germanium, and the Group V elements. The corresponding trivalent radicals are pyramidal and, hence, a radical reaction involving such a metal-centered radical is expected to afford the products having the same configuration as the re- actant; this is rarely of diagnostic value. However, radical VII readily racemizes (1‘75) . Ph,? s1. /\ Ph Me WII) Free Radicals 349 Evidcncc for tlic participation of a free radical in an organometallic reaction, using one or more of the above methods, may still leave two major questions. The first is wh(>thcrthc radical pathway is a principal or a side reaction. Rather than discuss this generally (organomctallic reactions are, after all, exceedingly diverst. and complex), it is suggested that as many as practicable of the exprrimmtal methods be brought to bear on the study of a particular reaction. The smond is whether the mechanism of the reaction is a chain or a nonchain process. This distinction, in an organo- metallic context, may not rcadily bc made and requires detailed kinetic data. The major secondary SOU~C(~Sof literature are listed in Table I. The single most important, itrm is t hc tno-volume compendium edited by Kochi (126). It will be noticctl that thc rolr of organic free radicals in transition metal chcmistry, cxccyt for (10 (11) derivatives, is a rather new area and constitutes a major themc of our article (Sections 111-V) . With regard to mrtal-centered radicals (Scction 11) , we concentrate particularly on those that have considerabl(~kinetic stability, e.g., b[CH (SiMes)~]~, another recent development (48). II METAL-CENTERED ORGANOMETALLIC RADICALS The study of metal-centered radicals has advanced significantly during the 1970s. Much of thc carlior \\ ork was concerned with long-lived species, particularly radical anions, or with the interpretation of kinetic data from which the existence of various organomctallic radicals was inferred (74, 75, ISla). The emphasis has shifted to the characterization of such transient species by ESR spectroscopy, and to the study of their structure, stability, and reactivity. Initial experinwrits wrre with the main group (Group IV) elemcLnts but have bccn cxtendcd to horon and the other Group I11 and V clcmmts [phosphorus (17) including thc phosphoranyl radicals (VIII), (VIII) as well as oxygen- (128) and sulfur- (123) centered radicals are outside our scope]. Much attention has bwn given to the synthesis of stable transi- 350 M. F. LAPPERT AND P. W. LEDNOR TABLE I REVIEWSRELEVANT TO ~)RGANOMETALLIC FREERADICAL CIIEMISTRY Section of this review Subject/title" Rcferences I1 Organotin radical chemistry 135 I1 Characterization of organometallic compounds by ESR- 174 mainly concerned with line widths TI Reactivity, selectivity, and polar effects in hydrogenation 170 transfer reactions, including hydrogen abstraction from PhsSi-H by Ph I1 Atom transfer and substitution reactions, including 163 SnR3/RrHal and R/RjSn-€I reactions I1 Thermochemistry of free radicals, including SiIta 155 I1 The shape of inorganic paramagnetic compounds, 6 principally from ESR parameters I1 Organometallic electrochemistry (this topic is not 55 discussed extensively hare) I1 ESR of transition organometallic complexes (this topic 79, 80, 84. is not disrussed extensively here) II/III Organometallic radicals of B, Al, Ga, In, TI, Si, Ge, 74 Sn, and Pb II/III Organometallic radicals of As, Sb, and Bi 75 II/III Organometallic radicals of Si, Ge, Sn, and Ph and 74, 110, 13,5 their reactions IT/III Free radical rearrangements, inc!uding SizMet, 4 191 Me3SiSiille2CI12-+ Me3SiCITpSiMez II/III Addition of radicals, %In3(M = Si, Ge, or Sn) 1 to multiple bonds II/III Structure of free radicals by ESR spectroscopy, 71 including spin t,rapping (no direct organonietallic content) II/III/IV Orgarlometallic radicals, principally of main group 102 elements, literature survey 1971 to mid-1972 II/III/IV Organometallic radicals, principally of main group 103 elements; literature survey 1972-1973 II/III /IV Organometallic radicals and reactions, excluding 111 transition elements and radical ions; literature survey 1970-1971 II/III/IV Organometallic radicals and reactions, excluding 112 transition clernents and radical ions; literature survey October, 1971, to December, 1973 [II/IV Organophosphorus radicals (this topic is not within 1'7 our scope) II/IV Rate constants for free radical reactions in solutions, 107 including MR, (hl = Si, Ge, or Sn) I1 /IV /v ESR studies in nonaqueous solutions of organometallic 131 complexes, particularly Si, Ge, Sn, and Ph Free Radicals 35 1 TAliLTS I-Continued Section of this review Subjrct/title” References II/V The homolytic piithwiy for thermal or photochemical decomposition of metal hydrocarbyls MR, (R = alkyl or aryl); arid stable paramagnetic species MR,, III/V Application of magnet,ic resonance techniques, including 193 the use of a spin lahcl IV S112 reactions at sonic main group metal centers 50, 51, 108, 151 IV $3~2reactions at boron centers 85 IV Electron transfer reactions of free radicals and metal 127, 129 complexes, including reactions of it with M n+, e.g., oxidation by Cu2+or reduction by Cr2+ (considered to be outside our scope unless Mnf represents an organometallic compound), and reactions of RMgX with transition metal complexes IV/V Autoxidation of organomet,allic compounds of maiu 2 group elements IV/V Synthetic aspects of free radical reactions of 23 organoboranes V The kinetics and mechanism of metal alkyl 164 decomposition ; a p:trticularly detailed account of HgR, V CIDNP, e.g., RHal/I,ilt’ or R’MgX systems 186 V Electron transfer reactions of organic anions, including 76 photolysis of LiAr and RHal/LiR’ reactions V Spin trapping 11s V Oxidative addition reactions of coordinatively 89 unsaturated, low oxidation state, transition metal complexes, including established radical reactions especially of Co (I1) V Organocobalt chemistry, including one-electron transfer 59 reactions, especially of Co(I1) V Organometallic transition metal complexes as initiators 8 of free radical polymerization a Entries in this column are intended to indicate the scope of the review in so far as it is relevant to organometrtllic chemistry. tion metal (paramagnetic) slkyls, (wentially by judicious choice of alkyl ligands. These, by their bulk and/or by the absence of 0- or a-hydrogen, e.g., MesSieHz,PhCH2, or 1-norbornyl-, may insure that normally facile decomposition pathways, such as p or (Y elimination, become energetically unfavorable (48, 49). Additionally, photolysis (and thermolysis) of metal 352 M. F. LAPPERT AND P. W. LEDNOR alkyls has been examined critically: homolysis is an uncommon route for metal alkyl decomposition (see Section V, B) . Electrochemical techniques have been used in organometallic chemistry; organometallic derivatives of main group elements tend to extrude carbanions or radicals, e.g., HgR or P;/IRI from HgRCl or (MR3)z (M = Si, Ge, Sn, or Pb), respectively (55). A. Main Group Elernenf Compounds Because of the existcnce of the excellent recent reviews listed in Table I (102, 103, 110-112, lyl),it would be inappropriate to deal with this topic extensively. We confine ourselves, therefore, to drawing attention to a selection of such radicals in Table I1 [section (a)], providing an introduc- tion to the literature on Group IV elemcnt-centered radicals [excluding their reactions (110, 171 11,in order to place in context our studies on the unusually stable species M[CH (%Mes) (M = Si, Ge, or Sn) (44,46). 1. Transient Group IV Elerfient-Centered Radicals a. Formation of Radicals. It is advantageous to examine radicals in solution since the ESR spectra are usually isotropic, which results in an average g value, average hyperfine coupling constants, and narrow lines. Assignment is normally much easier than for radicals in the solid state in which anisotropy is generally found. The disadvantage is the high rate constant for combination (typically loJ liters/mole/second) that occurs in solution, making it difficult to achieve a steady-state concentration of radicals sufficient for direct detection by ESR. In situ UV irradiation of a sample containing (ButO)z in the cavity of an ESR spectrometer has proved valuable for obtaining Group IVB radi- cals from the parent hydridcs, (Bu'O) zhv2BlltO (4) Bu'O + It3MII-MR3 + ButOI-I (5) This led to the first observation in solution of radicals centered on Si (1969) (16, 130), Ge (1969) (16),and Sn (1972) (188). Hyperfine coupling to the central atom was only observed for Si. A Pb-centered radical has not yet been detected in solution, although the existence of such species has been inferred from ESR results. For example, photolysis of a solution of PbzMe6 in the cavity of a Spectrometer gave a lead mirror but no signal attributable to f'bMe3 (weak &e was observed). In the presence of an TABLE I1 REPRESENTATIVEORCIASO ~IAIS GRO~-P ELEMEYT RADICALS CHARACTERIZED BY I<%R Compounds References Compounds References (a) Metal-centered radicals 102, 103 [BAra]: 35 103 [AIMea]: 102 103 SIR3 102 103 SIM~,(SIP~I~~)~-, 102 Si[CH(SiMe3)?]3 44 (c) Radicals centered B to a metal SiMe2F 102 CR1R20BR:] 102 SIM~,C~~-,, 102 C(CF3)0SiEt3(Me) 102 [PhC=CSil\le?]: 102 CPhzOSiPhp 102 ID7 [SIYX’X”~”’]; 103 CH?CH?PllR3 (11 = Si, Gr, Sn, or Ph) 102 ID GeR3 103 001111e3 (21 = SI, Ge, Sn, or Phj 10.3 W 0 G;e[CH(S~lle~j~]~ 4: CBu :CHiSiXz 10.3 a SnR3 103 CA~?OS~>I~~ 10.3 n 0- Sn[CH(SiRfe3)J3 46 CH,CH~ASR~ 103 VI PbR3 103 OOAsPha(OBul) 103 AsR? 102,103 [o-C 6% (AsPrle?) f 102 (d) Radicals centered y to a metal, or still more [ASR~I? 103 remote AsMe3SBu t 102 CH2CH20SiXle3 _____ 105 AsPhsOBu 105 CH2CHCHCHzSnIt3 102 [HgAmI; 58 (b) Radicals centered a to a metal ButSnH& n CH2Li(Li;\Ie3)3 10s 105 CH2MgMe 103 H,’ ’ W CH2BMez 102 (M = Si,Ge, or sn) CH?A~M~? 102 C(si~e3j3 10s w cn CH2MMe3(11 = Si, Ge, Sn, or Pb) 102 105 w CMe2MMe3(hl = Si, Ge, Sn, or Pb) 102 118 CHsMMe22fMe3(RI = SI, Ge, Sn, or 1%) 102 43 354 M. F. LAPPERT AND P. W. LEDNOR alkyl halide RX t>helead mirror was not formed, but a white precipitate, MeaPbX, separated and the spectrum of R was obtained (42).The prob- lem of radical recombination is overcome if the radicals are isolated from one another by an inert matrix. This usually lcads to anisotropic spectra, but the adamantane technique, in which the trapped radicals are free to rotate and thereby give isotropic spectra, has been successfully applied to the study of &n(Me).C13-, (n = 0, 1, 2, or 3) and &Me3 (143).Othcr solid state techniques that have brrn used to generate Group IV radicals include y irradiation for 6eMcs using GcMe4 in a matrix (1457, and a procedure involving a rotating cryostat, for &Me3 and i)bMe3 from ClMMe3 and Na (IS). Transicnt cationic species, c.g., [Pbhhlt, are ac- cessible by one-electron oxidation, e.g., from PblLlc4-[IrC16]z- (129). b. Structure of Radicals. Whercas the mcthyl radical 6Ha is planar, there are now much data suggesting that all other Group IV radicals show varying degrees of deviation from planarity. The main line of evidence uses the hyperfinc coupling of the unpaired electron to those isotopes of the central atom that possess nonzero nuclear spin (29Si,73Ge, lI7Sn, and lI9Sn). For a planar radical, the unpaired electron occupies a pure p, orbital which has a node at the ccntral atom, whereas for a pyramidal radical the odd clcctron is in a hybrid orbital. The magnitude of the isotropic coupling depends on the amount of s character in the orbital containing the unpaired electron. Consequently, the more pyramidal the radical, the more s char- acter in the orbital of the odd electron and, hence, the greater is the hyper- fine coupling. In general the greater the diff erence in electronegativity between the central atom and thc atoms bonded to it, thc more pyramidal the radical. For cxample, in the series $iMe,(SiMe3)3-,, as methyl groups are replaced by less electroncgative SiMe3 groups there is a trcnd toward the more planar structures (41).From thc a(M) data of Table I11 and on MH3,$1hTe3, or MCl,, trends are summarized in Scheme 1. c. Stability of Radicals. The trityl radical, 6Ph3, is perhaps the best- known stablc radical of Group IV and was long thought to cxist in solution equilibrium with its symmetrical dimcr, hexaphenylethane. It has now been established (138),however, that therc is the following equilibrium: Dimerization is presumably prevented by crowding around the ccntral carbon. Morc reccnt results, including those of Sections 11, A, 2 and 111, demonstrate the importance of steric hindrance to dimerization in con- Free Radicals CH3,CP\le3 All show marked deviations from planarity ilpproximately Iiicreabiiigly pyramidal Approximately planar tetrahedral a Electronegativity ads in opposition to a steric effect; hoa- ever the former is clearly dominant despite the considerable bulk of It and It'. SCHEME1. Variations in geometry around t,he central metal M in some Group IV radicals &lX3[X = 11, CI, R, or XR;;It = (lIe3Si)2CII;R' = MeaSi] tributing to radical stability. 0thc.r factors, such as the nonavailability of disproportionation pathways and thr possibility of delocalization into silicon d orbitals, may contribute to stability. 2. Stable Group IV Element-Centered Radicals This work (44,46, 142) originated in an attempt to establish whether the interesting compound SriItz [R = (MeaSi)&H] (47) (formally analo- gous to a carbene) exists in a singlet or triplet ground state. No evidence was found from ESR spectra of liquid or frozen ( - 110"*) solutions for the latter, but a weak signal near g = 2 was detected. It had quartct structure, suggesting $nR3, since the threc equivalent a protons of 8n[CH(SiMea)2]3 would give rise to a 1: 3: 3 : 1 quartet. Unambiguous assignment, through detection of I17Sn and "!'Sn satellites, was not then possible due to the low concentration of the species. The presence of the radical was initially attributed to reaction of Snlh with traces of oxygen, but later work showed that irradiation of the solution with UV or visible light caused ft dramatic * Trmperatures are all Centigr:de unless otherwise noted. TABLE I11 ELECTRONSPIN RESONAXEPARAhlETERS FOR THE STABLE GROUPIv TRIS(ALKYL)ASD TRIS(A4hlIDO) R.4DICALSa r Synthesis 4HY 7 or Q5 Radicalb Reactantsc Radiationd Solvent g a(?l')e u(M)cJ.g Stability at 20" in solutionh rnQ --IW SiRB uv C,H, 2.0027 0.48 19.3 tli2 - 10 min at 30" e GeC12 -diox/LiR CsHs 2.0078 0.38 9.2 Unchanged after 4 months SnR3 SnC12/LiR j uv Or ViS. CsHs 169.8 (ll'Sn) tllz- 1 year 177.6 (119Sn) Ge(NR;)3 GeClz-diox/LiiYR; UV n-CeH1, 1.9991 1.06 17.1 tliz> 5 months Sn(SR;) 3 SnC12/LiNR; uv n-CsH14 1.9912 317.6 (llFh) f112 - 3 months {342.6 (1lg9n) Ge(SBu1R')3" Ge(ICButR')2 uv n-C,Hi, 1.9998 1.29 17.3 tliz - 5 minp Sn(NButRf)p Sn(NBulRf)z uv n-CsHi, 1.9928 1.27 f1/2 - 5 minp SnR3 SnR2 + 6Sn(NRi)zq UV n-CsHi4 2.0094 169.8 ("7Sn) 177.6 (Il9Sn) a Data from Refs. 44,46, 14%. bR = (Me3Si)&H; R’ = MeaSi. cR = (Me3Si)zCH, R’ = MeaSi. d Under Ar or a vacuum. In mT. f BSi, 76Ge, 1%n, or 119Sn. 0 Corrected for second-order shift (Eq. (7). Based on ESR signal strength of a light-protected, sealed sample. i None required. j Or SnR2. Or Ge(NR:)2. Or Sn( N&’)2. 81-2.016; 911-1.994. n Reference 14%. O Reference 95. s The reason for the lower stahility of these radicals compared with M(SRL), is uncertain; from 9 and a(Ge) values on the Ge radicals, bond angles are probably similar; the generation of &I(SBU~R’)~was not as clean as of &I(YRiIa, and other paramagnetic. species were 0 sometimes detected, possibly KUutR’. p.n This was an attempt at a crossover experiment, possibly relevant to the mechanism of photolysis, Eqs. (8)-(11); ii trace of Sn(SR;)a was also present, but not detected with 1 : 1 SnR2:Sn(NR:)z, nor any mixed alkyl-amido-radical (142). Irradiation of Sn(XR:)2 with Sn(CsH,-q)z yielded Sn(SR:)(C;H;-q), but no paramagnetic species. Radical formation was also not detected by UV irradiation of Sn(CsH6-q)z,SnIz, SnCl(NR:), or Zn(NR,’)z. 358 M. F. LAPPERT AND P. W. LEDNOR increase in the signal strength. Neither type of irradiation caused deposi- tion of a tin mirror, and it was also demonstrated that (i) heat did not cause any increase in signal strength (irradiation of a sample without concomitant cooling raises the temperature considerably), and (ii) a sam- ple prepared in the dark gave no ESIt signal, but irradiation generated the radical. The intensity of the signal obtaincd on irradiation allow-1.d idcntifi- cation of satellite pcaks due to 117Snand l%n, confirming formulation of the radical as $11113 (46). Estcnsioii to thc related fie& and to isoelcctronic amidcs, ii1 (Mi:) (M = GP or Sn; R' = McBSi) \\as carricd out by reacting thc nietal(I1) chloride nith the apprcq)riatc. lithium rwgent and irradiating a solution of the product or, alternatively, GrIZ2 (93) or M (SK32 (92). For the Si- cc.1itcw.d radical, 8iR3 (S(Y Fig. 1),a diffcrent routc was rquird sinw suit- ablr Si (11) species are unknown, excclpt as short-lived intermediates. Compound Si2CI, was reacted with LiR nith the view to forming K3SiSiX, (X = R3, C13, %PI, or RC12) which would thcn be expected to fragment readily to gilt3. This radical n as obtained from SizClsand Lilt, followed by irradiation, but the compound isolated from thr wactioii, (SiC'12R)2,sug- gestcd a diff (writ mechanism for radical formation, pwhaps via R2SiC1SiC'Is. IN situ UV irradiation of a solution of PbH2 (47) at 20" gave a complex spectrum containing lines attributable to ( Mc3Si)&H, other paramagnctic species, and a lead mirror. [Assignment was confirmed by generating the samc radical from the low-tcnipcrature ( -40") irradiation of ( ButO)2 and (Me3Si)2CH2:doublct of multiplcts, a(a-H) = 1.89 mT, a(-y-H) = 0.037 niT.1 Irradiation of a solution of Pb(NKi)2 (92) with visiblc or UV light at 20", or UV at -40", gave no signals. Some decomposition of the sample appeared to occur. The main feature of thc ESR spectra (e.g., Fig. 1) of these metal-centered radicals is a inultipht arising from the coupling of the unpaircd electron to three equivalent protons (quartet) or three equivalent nitrogen riuclci (septet). For dilute solutions of the amido radicals, the septets showed further structure, attributed to partially resolved proton coupling. Under conditions of higher gain, satellite lines from those isotopes of the central atom that possess nonzero spin were obsc>rved(nucleus, percent abundance, nuclear spin: "Si, 4.7, 3; 73Ge,7.6, 4;117Sn, 7.7, +; and l19Sn, 8.7, $). The low abundance of thcsc: isotopes makes detection of the satellite lines diffi- cult, but the intensity of these was increased by using high microwave poww (c.g., 50 mW) and high modulation amplitude (e.g., 0.5 mT). (For C-centered radicals, values such as these can lead to saturation or loss in resolution, rcspcctivcly.) For hR3,the satellite lines were very broad but could be sharpened by an increase in temperature. The width of the lines Free Radicals 359 is attributed to incomplete averaging of the anisotropic contribution to the and hyperfine tensors, caused by slow tumbling of the radical. Raising the tenipcrature increases the rntc of tumbling so that the anisotropy is averaged to zero and the spectrum becoines isotropic. Measurement of the q va1uc.s [rclativc to polycrystallirie diphenyl- picrylhydraxyl (DPPH)] and thck CY proton [a(H)] or nitrogen [a(N)] couplings, was straightfor\\ ard, hiit determination of the central atom hyperfine coupling was not, and requires further comment. It is swn from the spectra (c.g., Fig. 1) that the satellite lines are not synimctrical about the central niultiplct; the satellites are shifted down- field, but not by equal amounts. This is a second-order effect and results from thc breakdown of the “high-field approximation” in the theory of coupling constants. This approximation assumes no coupling between the spin of the electron and the spin of the nucleus, so that when the formcr is reversed, the latter rclmains unc.haiiged. Under conditions of low magnetic field or large hyperfine coupling, this is no longer true and results in a non- linear divergence of energy levcls as the field increases and, hence, an un- equal separation of lines in thc spectrum. (More rigorous explanations of this effect are given in Refs. 84 and 8.) Corrected values of the coupling constants arc obtained from the observed spacings by application of the Breit-Rabi equation (for I = i), 2Ho(Ho - Hd 2Ho(Hz - Ha) n(M) = a(J1) = (7) 2Ho - Iik 2Ho - HI where a(M)is the corrected coupling constant, Ho the field position of the central line, and Hk or Hl the field position of the low- or high-field satellite, respectively (104).Since each nucleus (*%, “’Sn, and lI9Sn) gives rise to a pair of lines (at Hk and Hl), two values of the couplings are obtained in 360 M. F. LAPPERT AND P. W. LEDNOR each case, which serves as an intrrnal check. For the germanium-centered radicals, thc Brcit-Rabi equation was used in the form of Ref. 168. Com- puter analysis providcs the position of thc satellite lines using a trial value of thc coupling constant and the observed ficld position of thc central multiplet. Thc value of the coupling constant is varied until the calculated line positions agrce with the measured ones. The results are included in Table 111. Stability nieasuremcnts on thc radicals AX3 were madc using samples in sealed tubes, protected from light and stored at ambient temperature. Spectra wcrr rccordcd pcriodically, and stabilitics cstimatcd from the de- crease in signal strength (we Table 111). Thr radical &It3 decayed in benzene with a half-life of about 10 minutes at ca. 30". The decay curve was measurcd using the spectrometer to plot peak hcight against time. Thc value of tllz remained constant over five half-lives, thereby showing that the radical decayrd with first-order or pseudo-first-ordcr kinctics. The decay of &R3 appeared to bc revcrsiblr in benzcnc but not in hexane. Irradiation of the sample in C6H6caused formation of the radical up to a constant maximum intmsity; shutting off thc light led to complete decay. This cycle of formation-dccay was repeated several times but does not prove unambiguously that the radical decayed rcversibly. Electron spin resonarm' data obtaincd on MR3 and a(NR4)3arc listed in Table 111. For comparison, rcsults on MMe3 are: &Me& solution, g,,, = 2.0031, a(H) = 0.634 mT, a(M) = 18.3 mT (15);GcMe3, matrix, g,,, = 2.0101, a(H) = 0.53 mT, a(M) = 8.47 mT (143);&Me3, matrix, g,,, = 2.0163, a(H) = 0.25 mT, a(M) = 153.0 (lI7Sn) and 161.1 (Il9Sn) mT (IS);i)bMe3, matrix, g,*, = 2.0389, a(M) = 185.0 mT (IS). The mechanism for 3x3 formation from photolysis of MX2 [MXz = GeR2, SnIh, Gc(NRi)p, or Sn(NR:)2] may follow cither of two routes (shown for MX2 = SnItz) (46), SnIt* hv R + SnIt It + SnR, -SnR3 or SnRz (SnRz) * (10) (Snit,)* + SnRz -SnR3 + SnR (11) In the first mechanism, an M-X bond is homolyzed and the resultant radical X is trapped by another molecule of MXz. [Unsuccessful attempts to provide evidence for this proposition were experiments (i)-(iii) ; the radical precursors (BU~ON)~and AIBN, as well as the inhibitor galvinoxyl, reacted with SnRz at 20" to give diamagnetic solutions; whereas, in ex- Free Radicals 36 1 pcrimcnt (iv), Ph3e did not react (1.42).] Alternatively, MX3 is formed from a bimolecular reaction bctwecn an excited state of MXZ (possibly triplet) and a ground state MX2. Both mechanisms require formation of a M (I) species: since no metal devrlops on photolysis, MX must react with solvent or form a soluble diamagnetic oligomer. It is possible that SIR, is formcd according to the following reactions: c1 c1 c1 c1 I1 I/ SiiC16 + GLiR-R-Si-Si-It + It-Si-Si-Cl (+other products?) (12) II 11 (‘1 c1 R C1 isolattd pxtulsted c1 (‘1 I, I/ 11 c:1 2SiR1”Y-SilL + -1 (Sin), n Supporting evidence for Eq. (13) is provided by (109) /(si’fe*)4\ xsiMez+ (SiMe2) I) (15) MezSi- Si?fe2 The postulated SiMcz was trappcid as an insertion product, but SizMeswas not isolated (which might haw brcn expected if the silylene itself photo- lyzed to &Me3). In situ irradiation of (SiMez)6(toluene, -60”) did not lead to ESR detection of $iMc3 (142). However, for Sinz the bulky R groups may stabilize the silylenc sufficiently for photolysis to biR3 to be favored over alternative reactions. There is precedent for such rcdox reactions in transition metal chemistry, e.g. (S), 1 (TiX3)2-TiS4 + - (TIX~)~-X = NMei n although photochemical disproportions are less common (25) (but see Section 11, B) . The g values for MRs increase down the group (Table 111) , the same trend as is found for &IMe3 and MH3. It is attributed to the parallel in- crease in the M-spin-orbit coupling constant, because, in general, Ag is proportional to E/AE, where Ag is the difference between the measured g and the free-spin value of 2.0023, E is the spin-orbit coupling constant, 362 M. F. LAPPERT AND P. W. LEDNOR and AE is the diffcrcncc bctwccli ground-state and excited-state energies. However, Q[&(NK~)~]> Q[~(NR.~)~](Table 111) ; cf., y(6eC13) > g($nC13). These findings may be due to delocalization of the unpaired electron into ligand T-type orbitals (142,243). Anisotropy in valuchs was only found for $nlt3 (Table 111). The Q values of the aniide radicals arc less than those of corresponding alkyls, coiisistcnt with the former having niorc pyramidal structures; bcnding of an xfX3radical mixes exeittd states into the Lvave function for the unpaired electron and, thus, loncm the Q valuc.. The methine splittings a(H) for the radicals ~‘I[CH(Sih4e3)2]3arc close to those for XI (CH3)3. The niost notable fcaturc is the large diffcrcncc be- tween C(CH3)3[a(H) = 2.25 mT] and the analogous Si, Gr, and Sn species [a(H) = 0.634, 0.53, and 0.275 inT, rcspectivcly; see Tablc 1111. This differcncc is attrihutd to (i) the niorc pyramidal gwmctry of the Si, Ge, Sn, (and Pb) radicals; (ii) th(x reluctance of thc heavier clcnicnts to form niultiplc bonds; and (iii) thc greater size of the heavier elements cornpared with C:. All t hrec factors reduce hyperconjugative coupling, which is believed to bc thc main cause of such splittings in simple alkyl radicals (6). 111 this procms there is a contribution to thc bonding from a structure in which thc uripaircd clcctron couples with onr of the electrons in a C-H bonding orbital, H H. (16) M-C -M=C The extreme longevity of radicals 6TR3 and Fh(NRi)3 (Tablc 111) must be mainly due to stcric hindrance to dimerization (44,46). The low values of the M-H bond strengths [D(M-I-I): C‘, 104; Si, 81; Ge, 73; Sn, 70 kcal mole-’] (110) do not favor H abstraction from the C-H bonds of the solvent, but the shorter lifetimes of the $11 (NR:) and Ge (NK;) radicals in hcxane compared to hnK3 and &R3 in beiizerie may reflect the ease of H abstraction from thc two solvents. A third factor, for thr tris(alky1) radicals is the low probability of disproportionation, such as that shown for a carbon-centered radical, because stable, double-bonded compounds of the heavier Group IV clc- ments arc unknown. Free Radicals 363 The first-order decay of radical &R3 in benzene implies reaction with solvent or an intramolecular rearrangement, such as SiMe, I CH,-Si(Me,)-C(H) (SiMe,)-Si[ CH(SiMe,),], Nonradical products It is interesting that the stability, unlike the geometry, of these radicals kX3 is so sensitive to steric effrcts. Bulky groups X thus hinder the forma- tion of a four-coordinate X3hl-MX3 or X3M-H and favor the three- coordinate AX3 (44).A similar vffect has been noted for transition metal MR3 and M (NR:) 3 complexes (48). B. Transition Metal Compounds Many paramagnetic organo-transition metal complexes are stable under ambient conditions. Relevant ligands include CO, It-, olefin, q-C5Hs-, or q-arene, and compounds may bc iirutral, t'.g., [V(CO)J, [Cr (CH2SiMe3)4], [Cr { CH ( SiMe3)2 131, and Ta('lz ( C5H5-q)21, anionic, c.g., [Os3 (CO) &, [Cr ( CH2SiMe3)J-, and [Tir12 ( (':H5-q) 2]-, or cationic, e.g., [Cr (ArH- q)2]+. These owe thcir stability, in part, to electron delocalization for .rr-bonded ligands; whereas for tliv alhyls, it is duc to stmic effects (effec- tively making thr mrtal coordinativrly saturated) and the use of ligands that do not allow normal dccomposition pathways to br accessible (e.g., p elimination from M--R if R- Iias no P-hydrogen). Various recrnt reviews are available (48, 49, 55, 79, 80, 84), but a representative serirs of com- pounds is included in Table IV. 13xpcriincntal results are conccrncd with preparative methods (e.g., all.;ali-nictal or elcctrochemical reduction in donor solvents for radical anions), and bulk magnetic susceptibility, ESlt measurements, or clrctronic spvctra. Thcsc have, in grneral, becn unexcep- tional. For vxample, many of thch alkyls of first-row transition mctals have magnetic moments clos~to spin-only valurs, and ESIt spectra arc' consist- ent n ith clectronic structure and gcomctry, e.g., tetrahedral and trigonal for thc local Cr environment of ('rl<4 (19,153, 185) or CrR3 (10); the tetra- 1-norbornyls of Fe(1V) and C'o(IV) are low spin with perf = 0 and 2.0 BM, respectively (19). TABLE IV 0.0 REPRESENTATIVEORGANO TRANSITION METALLIC PARAMAGNETICSPECIES P Compounds References Compounds References d' Ti(II1): d2 V(II1): [V(C~H~-V)ZXI(X = C1, Ph, 66, 84 SR, or Tol) [VICH(S~~~)Z)31 10 Cr (IV) : [CrR,] (R = Me, But, 19, 153, 185 3 MerSiCH2, i\.le,CCHz, n I'h3CCH2, PhMe2CCHZ or I-norbornyl) Ws Zr(II1) : W W(1V) : [w(CHzPh)rl 49 rn V(1V): -i5u b d3 V(I1): complexes containing 67, 84 z0 (1-C5H 5)- Cr(II1) : [Cr(CHzSihles)J 121 P [CrR3][R = (Me3Si)ZCH; 10, 19, 49 .r 2,2,3-Me3-l-norbornyl, r F MezP(CH2)Z; CHz(PPhzC6H,) o-RzNCHZC~H~] ; z Nb(1V): Mo(II1) : complexes containing (1-CsHs)- 84 Mn(1V) : [Mn(l-norbornyl) .,I 19 Ta(1V): d4 Fe(1V): [Fe(1-norbornyl),] 19 Mo(V) : d6 V(0): W(V): Xb(0): M(1): (M = Cr, Mo, or \V) 84, 97, 62 [Fe(C 5H4C &XO~-p-77) 21- 45 [M(arene-v)~I+, [Fe(C5H5-1){C5H4CO(COR) 11- 146 [M(C5H5-7) (CO)31, and ferrocenebenzo- [M(CO) 4(PMe3)1- semiquinones Cr(1): [Cr (C5Hm) (c~.Hm)l, 62 [Fe(C&-tl) (C0)dNOAr)l 105 [Cr(CsH 5-1)(C7H.r-v) 1- Co(I1) : [Co(diphos)R(H)] 95 [Cr(CJIS-v) (CsHs-v)I- 8 (IV) 19s [Cr(CHzPh)L]+ 129 Mo(1) : [Mo(C5H5-v)(C0)dh'OAr)l 105 dg Cr-(111): [Cr(CloH&P- 97 [hlo(CO)4(diacetylanil) 1- 56 Co(0) : [Co3(CO)9CX]- (X = R, Hal, 141 Mn(I1): [Mn(C6H4Me-v)(C0)z- 145 or Sihlea) { C(0Me) (1-ferrocenyl) )I+ [Co(C0)4021, [CO(CO)41 69, 120 [Mn(THF)#+ 105, [Co(CO) aP(0Et) a(NOA4r)] 105 Fe(II1) : ferrocenium salts 165 [Co(CO)(?;O)phen]- 57 11 [Co3(CO)9CI 172 9 [Fe(olefin)13+ (D 178 Co(1V) : [Co(l-norbornyl)41 19 Fe-(I) : [Fe(C&) (COMPPbdI- m 0 [Fe(CO) 51- 159 n 97 Fe(0) /Fe-(I) [Fe~(C0)91-,[FedCO) 121-, 159 5. d? Cr-(I): [Cr(C~oHdzl- : n Mn(O)/Mn(I) : [Mnz(CO)lol- 4 [Fe3(CO)lJ'(OPh) a]-, z Mn(0) : [Mn(CO)sBr]-, 4 [RuFedCO)121-, [Mn(C0)a(PPh3)~1 154 [Fez(CO)sPt{P(OPh)3)21- [Mn(CO)50~1,[Mn(CO)51 69, 101 Ru(O)/Ru(I) : [Rua(CO)izl-, [RudCO)i1Cl- 159 [Mn(CO) 4(P)( NOAr) 1 105 Os(O)/Os(I): ~Osr(CO)d- 159 (P = CO or PPhMez) Ir(O)/Ir(I): [Ir4(CO)1~1- 159 Tc(0) : [Tc(CO)zI 116 Ni(1): [M(PR3)3 or 4(E-)l 64 Re(0) : [Re(C0)3(P)2] (P = PPh3 or rs, 152 [E = Cz(CN)4, Clz(CN)z- PMeZPh) Pt(1): benzoquinone, or chloranil; 64 [Re(CO) &VOAr)l 105 M = Ni or Pt] Fe(1) : [Fe(CsH5-7) (C5H4NO(BuL)-s)l 72 Cu(I1) : [Cu(~lefin)]~+ 11 [Fe(CsFI,-v) (CSH,CO(Ph))]- 61 Ag(I1) : [.4g (olefin) 12+ 77 w 0. cn 366 M. F. LAPPERT AND P. W. LEDNOR Numerous transient paramagnetic compounds are known and some of these are also shown in Table IV; there is an overlap with Section V, and Table IV does not duplicate material which is more conveniently treated later. The distinction is arbitrary, but we shall defer consideration of transient transition metakentered radicals, e.g., Pt (I),if their formation is primarily of interest in connection with an organomctallic mechanistic study, e.g., the oxidative addition of an alkyl halide to a Pt(0) substrate. The designation of metal oxidation state in Table IV is somewhat formal; in many cases it might be more appropriate to describe a complex as de- rived from a paramagnetic ligand, such as a nitroxide or ketyl. The saga of neutral paramagnetic metal carbonyl transient species is somewhat confused. Claims to having obtained [Co (CO)4] (120) or [Mn(CO)5] (70) by sublimation of the dimer onto a cold (77 K) finger have been shown to be in error for the latter. In the absence of dioxygen, the Mn condensate is diamagnetic, but in the presence of O2 the radicals were trapped and identified (ESR) as [Co (CO)402] or [Mn (CO) respectively (69). Similarly, photolysis of metal-metal bonded dimers in the presence of nitrosodurene in chloroform at -30" yielded the spin- trapped mctallonitroxides (ESR) [ML,( NO(Ar) ]] [ML = Mn(C0)j, Re (CO)6, Mn (CO) ( PPhMr2), Co (CO)BP ( OEt) 3, Fe ( C5H5-s)(CO) 2, or Mo(C5H5-v)(CO),] (105) (sce Sections I11 and V, B). Radical [Re(CO)5] has been obtaincd by photolysis (194) and by co-condensing rhenium vapor and CO in an inert matrix and 11%analysis (200). (Several other metal carbonyl paramagnetic spccics havc been obtaincd by matrix isola- tion techniques, e.g., [Mn(CO)J (101).1 Irradiation at 350 nm of a dc- gassed THF solution of [Mn2(CO)lo] produced an ESR signal (88); this almost certainly is due lo a Mn(I1) species (105, 105~)rather than to [Mn (CO) 5]. Kinetic studies of CO displacement by tertiary phosphinc or phosphite from [Mn(CO)4(PPh3)I2, [Mn(C0)4{P(OPh)3)]2,[Re(C0)4 (PPh3)]2, or [Ru( CO)4SiMe3]2point to a small steady-state concentration of the corresponding monomc'r during thc course of these reactions (68). 7-Irradiation of [Mn2(CO) 10J providrd the radical anion [Mnz (CO) 101- in which the unpaired rlectron was believed to be principally in the Mn-Mn u* orbital (4). Similarly, [Mn(CO)5Br]- resulted from the neutral precursor, and the Mn-Br U* orbital was implicated. A series of stable (11,~- several hours) polynuclcar metal carbonyl radical anions has been obtained by alkali-metal and/or electrochemical reduction of the neutral parent, in O2-frcc cthcr solvents (159), which revert to the starting material by quenching with MeI. Lowfield ESR signals were assigned to the nonbridged species, believed to be in temperature-dependent equi- librium with the carbonyl-bridged isomer. Free Radicals 367 111 OTHER ORGANOMETALLIC RADICALS Representative lists of radicals in which the unpaired electron is mainly localized remote from the metal are included in Tables I11 and IV. The distinction can readily be made cxpcrimentally by ESR provided that the metal has an isotope with a nonzero nuclear spin. For transition metal species, when r-bonded ligaiids are often implicated, or radical ions, the demarcation may become somewhat blurred; for this reason Table IV is not subdivided according to whether a particular compound is best formu- lated as of type I or I1 (see Section I). The reviews cited in Section I1 are also relevant for type I1 radicals. Metal alkyl or alkoxy radicals in which the unpaired electron is localized on carbon, a or to the metal, are known for Li, Mg, B, Al, Si, Ge, Sn, Pb, and As (for a selection, see Tablc 11) but not at present for a transition metal. Methods of gencrating such species are by yirradiation of the solid at low tempcrature or by hyclrogcn abstraction, for example by Ph&b or Bu@ radicals, or the addition of a metal-centered radical (e.g., by photolysis of the metal hydridc with RulOz) to an unsaturated substrate; the following reactions are some examples: (LiMe), + But; : kH,Li(LiMe), (19) (73) [metal atom substituent eclipses odd-electron n-orbital (86)l 368 M. F. LAPPERT AND P. W. LEDNOR Reactions such as that shown in Eg. (19) are unusual. However, another example is the formation of the stable C(SiMes)3from (Me&i)&H and But6 (12,150) (this decays by a first-order hydrogen abstraction from the silane). More typically, But6 attacks the metal center, displacing the radi- cal R;this is an cxamplc of an SH~reaction discussed in Section IV [but see XVI (104a)l.Thc rhoice of a particular pathway is govemod by the activation energies for the competing processes, and relevant bond strengths provide a useful guidr. The ESR spectra of arylsilane radical anions show that the spin density is associated essentially with the aromatic rings. At low temperature, the unpaired electron in [Ph(SiRz)zPh17 is delocalized over one of the phenyl rings but, at higher temperatures, over both (118). In species such as [Mez%- ( C6HI- Ph-p) the MczSi moiety bridges essentially independent ?r systems (43); in the neutral analogs, the triplet electrons are similarly distributed. Nitroxides derived from organometallic Group IV elements have been obtained by reactions such as (189) 1 -H+(LiR); RMe %SiSHOSi A Ie AR *li~~~2Si~OSihIe2R~(RR'Iekh) *NO 2 electrolytic or 02ovldation 1- le- (25) (1t~e2~12)30 They show low values of a(N), indicating localization of the unpaired electron on the oxygen and strong h-M (Si > Ge) T bonding, consistent also with higher y values than in &NO. Transition metal paramagnetic complexes that warrant further com- ment are the radical anions derived from metallocene- (especially ferro- cene-) ketyls and related species, such as compounds IX (GI), X (45),or XI (146), and metallonitroxides. From the ESR spectra of compounds pC,H, No2-p Fe I such as IX-XI, generated from alkali-metal and the neutral precursor, there is some evidence for participation of metal orbitals in stabilizing the lone electron, and the analogy has been made with the unusually stable a-ferrocenylcarbonium ions. A similar picture emerges from an ESR and X-ray ( !) study of the ferroccnylnitroxide (XII) (72),obtained by autoxi- Free Radicals 369 dation of the corresponding hydroxylamine, which was, in turn, prepared from the fcrrocenyl-Grignard rcagcrit and ButNO. The metallonitroxides, such as compound XIII, were obtained by spin- 0 .I p,,,‘I N-Mn(CO), Fe 0 I Me Me (XI) (xm) trapping the metal-centrred prcwirsors [which, however, were not de- tectcd by ESR in the absence of spin trap (see Section II,B)] and identified by ESR (Table V and Fig. 2) (105). The similarity of a(N) in all the species and the g factors suggcst, that little spin density is associated with the metal. The nitroxidc (IV) (sw Section I) was used as a spin label in connection with a study relcvant to vitamin B12coenzyme, which is of type Co (111)-R (R = 5’-dcoxyadenosyl) and thus an analog of IV, but with a different axial group R (19.9).Compound (IF’), which we may abbreviate Co (111)- R’, was prepared from a Co(1) precursor BIZs and R’Br, and was found to bind to the enzyme ethanolaminc-ammonia-lyase (in this system ethanol- aminc is converted to acetaldrhydr). By ESR it was concluded that the activc site is relatively close to the enzyme surface (see also Section V,B). Another type of complex rcl(1vant here is that formed from a stable or- ganic radical and a metal complcx. An cbxample of the former is a nitroxide, a Data from Ref. 105. a In CHC13 or THF solution at ca. -30”. I = 5/2. I = 9/2. 370 M. F. LAPPERT AND P. W. LEDNOR I 1 mT c? FIG.2. The ESR spectrum of M&~rM;:'"l. in CFICI, at ca. -30". Me Me such as Bu$TO, which complexes Lewis acids such as SnC14, TiC14, or Bu$n-Cr(CO)6 (S7),or a nitroxide related to compound IV which can bind to Co (Section V,B) (193). IV BIMOLECULAR HOMOLYTIC SUBSTITUTION (S,2) AT THE METAL CENTER OF AN ORGANOMETALLIC SUBSTRATE Bimolecular hcterolytic substitution (SN~and SE~)at a saturated center has been extensively investigated during the formative years of physical-organic chemistry, and the SE~mechanism has an important role in the context of organometallic chemistry. The SH2 mechanism involving a metal center M is defined by Eq. (1), and has been established for main group elements M since 1966. The well-known free radical process of atom (H or Hal) abstraction is an example of the sH2 mechanism at a terminal Free Radicals 371 atom, e.g., Eq. (19). Relevant reviews, coiifincd to niaiii group metal centers, arc in Refs. 23, 50, ,51, 102, 103, 108, 111, 112, 131, 151, and this aspect is therefore not discussd in detail (Section IV,A) . We shall limit ourselves to a brief introductioii to SI12 processes, without indication of primary sources [but refer(wc(>should also be made to later papers by Davies et al. (5.3) and KH% reactions at Si and Gc (99),and Mg (32)3 (Section IV,A), follo\\-c~lby a discussion of the situation involving a transition metal center Rf, invoking observations made since 1973 (Section IV,B) . A. Main Group Element Compounds as Substrates Some indication of the scopc. of SH2 reactions is gleaned by considering the types of attacking radicals X in Eg. (1), which may be 1x6, Rb, R&, RCOfiR, R$O,, R, or Hal. Most, of the work involving alkoxyl radicals has used But6. This may I)(> generated photochemically from BuiOz, which is particularly convenient for ESR studies. For cxample, in 1969, several groups reportcd that the iu sztu photolysis of a static solution con- taining Bu402 and a substrate gave rise to radicals dcrivcd from thc latter (131). High-intensity UV light is focused onto thc solution, causing the formation of But6 radicals from symmetrical cleavage of the peroxide. The But6 radicals arc not dvtcctcd, due to extreme width of the signal. The two most comnion rc:wtioiis arc hydrogen atom abstraction, e.g., Eq. (19) , or an sH2 displacemcmt, i.c., iwo + NIL -~WOMIL + it (26) The sensitivity of thc techniquc is indicated by the detection of natural abundance satellites in tlw spcctrum of 6H3, derived from BMe3 and (ButO)z (131).The principal limitations of the technique arc thc solubility of the organonirtallic sample, its photochemical lability, and the efficiency of the light-generating and focusing apparatus. An intcresting example is in the ButO/CIMgPrz systcin whcn the SH2 process (to givr PrZ)was ac- companied by H abstraction [to yield (ESR) CNIe2MgC11 (32). Di-t-butyl hyponitritc is a suitable thermal sourcc of Bu@ and nitrogen (tlp = 29 minutes at 65") (124): (I%u'OS)~-2Bu'O + Nz (27) It has been used in the E:SR detection of alkyl radicals, as in Eq. (26), whcn the substrate is photosensitive. As a thermal source of But6, (ButON)z is usually more convenient than ButOOBut (ill2 = 1 hour at 150"), or the more labile arid somewhat dangerous di-t-butyl peroxyoxalate (calculated tlp = 6.8 minutes at GO"). 372 M. F. LAPPERT AND P. W. LEDNOR A chain reaction, involving SH2 attack by But6, is illustrated by the action of 2-butyl hypochlorite on a metal alkyl, BufO + MIX, -BU~OMR,~-, + R (28) R + ButOC1 -RC1 + BufO (29) This process has been demonstrated for MR, = BRB or R,SnCl4-, (n = 1-3). A synthetically useful SH~reaction is found in the conjugative addi- tion of organoboranes to cnones: .\ I R + C=C--C=O-lt--C--C=C-O (30) / II Ill I I It-C-C=C-~~ + BIta-lt--C-C=C--OBRz + R (31) Ill Ill The triplet kctoncs, RzC-O, are a special case of alkoxyl radicals. Thus UV irradiation of acetone and tri-n-butylborane in the cavity of an ESR spectrometer led to the observation of two radicals, eMezOBBu;, Eq. (21), and Bun. On a preparative scale the coupled products as well as BuaMezCOBBu,"were isolated. Both chain and nonchain reactions involving an SH~attack are known for thiyl radicals. An example of the latter is the photolysis of ButSSBut in the presence of an organoborane or BulMgC1, when an alkyl radical is displaced by But$ and can be detected by ESR. The former, known for B, Sb, and Bi, arc illustrated by R'S + BR, -R'SBRz + R (32) R + R'SH-RH + It's (33) Dimethylaminyl radicals may be produced photolytically from MezN.N=N - Mez, and the SH2displaced k has been observed for BR3 and SbR3; the evolution of nitrogen can interfere with the recording of satis- factory ESR spectra. A chain reaction has been observed between MezNCl and BR3, cf. Eqs. (28) and (29), but a competing nonradical process yielding MezNRand CIBRa is also found. N-Halogenosuccinimides undergo a chain S132 reaction with SnR4, in contrast to MezN (where there is no reaction) or But6 (where there is hydrogen atom abstraction). Another radical chain process is shown in the following propagation sequence: PhSOz + BBu;zl'hSOzBBu; + Bun (34) Bum + I'hSOz13r-BunBr + Phh02 (35) It has been suggested (50) that the widespread insertion reactions of sulfur Free Radicals 373 dioxide into metal-carbon bonds may also involve an sH2 reaction, KS02 + MR,, -ItS02MR,-l + R (36) R + SO,---RSO, (37) a scqumce similar to that found for the autoxidation of organoboranes. There are isolated reports of SH~reactions involving attack by C-centered radicals. These are typically gas-phase reactions using the photolysis of acetone (sometimes the per-deutero or per-fluoro analog) as a source of he, which gives CD, with Hg (CD,) 2; examples involving B and Sn have also been reported. Reactions involving halogm atoms are poorly dcfined, due to competing polar processcs from the parcrit Xz, or to alternative reactions by the halo- gen atom (e.g., hydrogen atom abstraction). However, the reaction of a dialkylmcrcury with a dihalogcn in a nonpolar solvent can be accelcrated by light or peroxide, and retarded by oxygen, so that the chain sequence, X + 1IgR2-XHgR + R (38) R + sz-ItX + x (39) may be operative. The reaction between Si2R6and (BrCH2)2, yielding BrSiRa and CHz= CHz, is inhibited by galvinoxyl but initiated by BzzOz and, hence, appears to be the first example of SrI2 at Si in solution (99, 171); GezMes behaves similarly. The sH2 reaction can, in principle, be stepwise or synchronous; if the former, an intermediate (XIV) may be detectable by ESR, X + hlR, -XkR, -XhIR,-l + R (40) (XIV) but has so far only been found in the reaction of Bu@ with trimethyl- phosphine or with phosphitcs, to give a phosphoranyl radical (VIII) . Many of these sH2 processes are fast, thc fastest approaching the dif- fusion-controlled limit, c.g., for But6/BBu,”, k = 3 X lo7 M-I second-’ at 30”. The factors favoring an S112 reaction are the presence of energetically low-lying M acceptor orbitals, the formation of a strong bond relative to the bond being broken (e.g., B-C-B-0), and the absence of steric effects that hinder access of 2 to M. B. Transition Metal Compounds as Substrates At the start of our initial work, published in 1973-4 (29, I@), there was no definite example of an sH2 reaction at a transition metal. We shall, there- fore, describe in detail SH2 displacement (-A) reactions of Butb (non- 374 M. F. LAPPERT AND P. W. LEDNOR chain) and Phh (chain) and cis-[PtRz(PIt:)2]. Our entry into this area was accidental and arose from an attempt to gcnerate a radical with an un- paired electron in an alkyl chain and not at a transition metal center [ (11), which are still unknown (sccScction TII)]. We chose to attempt to abstract a hydrogen atom from a methyl group bound to a metal, using Bu@. Thc metal compound selected was a Pt (11) methyl because platinum(I1) alkyls combine case of synthesis with stability to air and moisture; the nuclear spin of platinuni might aid ESIt idcntification of Pt-cH2; and the methyl group should present a unique site toward H abstraction. How- ever, three other modes of attack by But6 are in principle, also possible (XV). Actually, we have mainly observed S112 at Pt(I1) but, in onc case probably at silyl-substituted-Me (29, lo.&). R H abstraction H,C-pt-x from Me I H abstraction from R: much less likely I for R = phenyl than - for R = alkyl (X = Me (05-isomer) R’PR or Hal) (xv) Initial expcrimcnts with [PtBrMe (PPh,) 21 wcre hampered by its low solubility. For cis-[PtMcz (PEt3)2] the possibility of competing H abstrac- tion from Et existed but did not occur. Instead, ESR expcrirnents involving in situ irradiation of this compound in BuiOz, suggested an SH~reaction at platinum. To confirm and extend this evidence, a number of other platinum( 11) alkyl complcxcs wcrc synthesized and the following typcs of experiment carried out : 1. Dircct observation of the displaccd alkyl radical, using But6 radicals generated photochcmicslly from (BuQ) 2. 2. Spin trapping of the displaced alkyl radical, using Bu16radicals gen- erated thermally from (BulON)2. 3. Attempts to prepare a complcx containing the Pt-OBut group by an SH2 reaction. 4. Attempts to find an SH2chain reaction between a platinum alkyl and a thiol or disulfidc. As for experiment 1, thc ESR signals obtained were weak and lasted only a few minutes, making dctcction difficult. The reasons for this were Free Radicals 375 probably low solubility of the organonwtallic, competing reactions (see structure XV), and efficiency of the apparatus (increasing the light in- tensity, say from 230 W to 1 kW, Irads to strongcr signals, but of shorter duration j . Low solubility means rapid sample depletion and a corrcspond- ingly short-lived signal. Even the strongest signal obtained from cis- [PtMe2(PEt3),] decayed considcrably over the time taken to record a spectrum. Furthcrmorc, it is not possihlc significantly to reduce the teni- perature without extensive prccipitation. In general, the lower the tcmpera- ture the better, because the reactivity of the radicals is reduced, and the population difference between thv spin states is increased with consequent increase in the intensity of ahsorption. [A reduction in teniperaturc is not always beneficial-it can katl to hrondcned lines, particularly for large radicals, due to slower tumbling (sw Scction II,A,2) .] The best conditions wcw usually found by dissolving the sample in Bu.33, and irradiating at -45" (the pure peroxide freezcs at -40" and the presence of other matcrial reduces this to ca. -SO", but, by using a solvent such as cyclopropanc, much lower temperatures could be achieved). The signal ohtained by irradiating a mixture (ca. 1: 1) of dimc.thoxymcthane and (BuQ)~at -43" (which is then honiogcneous and liquid) was of good strength and lasted for about 1 hour; the spectrum is due to CHsOeHOCH, (doublet of sep- tets) and eH20CH20CH3(trigl(1t of triplets) and the intensity of these lines was used as a standard, with th(. optical apparatus adjusted to maxi- mize the signals. In thc work nith Pt (11) comp1cx(~s,conditions of solvent and tcmpcraturc. depended mainly on thtl phosphinv. Dcoxygrnatcd I3u$& was usrd cithcr neat or as a 4oIutioii in hcnzcnc and krpt under argon or nitrogen. Solutions of the Pt (11) cvniplcx were normally saturated at room temperature but a small amount of solid deposited by cooling did not interfere uith the signal. The 1I:Sli results are summarized in Scheme 2. The bis (trialkylphosphinc)Pt (11) complexes were studied to obtain better solubility than with czs-[PtMv3 ( PPhs),I. Complcx cis-CPtEt2- (PPry)2] was used (fit detectcd) in attempting to eliminate the possi- bility that the source of Ae formed froin cis-[PtMe2(PEt3),] was scission of Buf6catalyzed by Pt (11); and ~is-[PtMe~(PBu~)~]was employed in atternpting to obtain kinetic data with a highly soluble complex and moni- toring the relative decay of $1~and GsH9 when using cyclopcntane as solvmt (finally, this was not pursued because the concentration of M in pure Bui02 at -40" nas low). With cis-[Pt(CH2SiMe3)2(PMezPh)2] (58)in BuiOz, irradiation at -20" gave a strong spectrum (tip ca. 5 sec- onds at -20") (14.2) consisting of four multiplcts, each seemingly a 1 :4:6:4: 1 quintet which agrcwl reasonably with a computer-simulated spectrum of the platinum-c0ordiiiatc.d methylcnc. radical (XVI) , using a(P) = 5.6, a(Pt) = 14.9, a(H) = l.G niT, and q = 2.021 (10.4~).Attack, 376 M. F. LAPPERT AND P. W. LEDNOR hv BU~O-OBU~ z 2 ButO saturated solution in Bu;O2 2 R(ESR) hv, -40" R I P-Pt-R I P (A) solution in THF L No ESR signal hv, -40" (A): R = Me and P= PEt,, Me detected; R = Et and P =PPr:, it detected; R = Me and P = PBut, Me detected; R = PhCH, and P = PEt, , broad signal; R = Me,SiCH,, and P = PMe,Ph, four multiplets, the two outer being weaker than the two central multiplets, and each a 1 :4 :6 :4 :1 quintet, assigned to Complex XVI. The SH~at Pt(I1)-R by Bu'O, generated photolytically ; summary of ESR data (29,142 ). SCHEME2. The Sa2 at Pt(I1)-R by BulO), generated photolytically; summary of ESR data (29, 142). . in this instance, at Si rather than at Pt may be due to steric hindrance at the metal center. However, this is a tcntative assignment. PMe,Ph PhMe,P -F!t-bH, I CH,SiMe, (WI) Spin trapping is an ESR technique for identifying free radicals in reac- tions in which their steady-state concentration is inadequate for direct detection. The spin trap, commonly a nitroso compound or a nitrone, and in this work (29,142) Bu'NO or nitrosodurene, when added to the reaction mixture couples with free radicals to generate the paramagnetic spin adduct of much greater stability. The concentration of the latter, therefore, builds up and is detected by ESR. ButNO and 2,3,5,6-Me4C6HN0 (ArNO) have no a-H and, hence, in the spin adduct But(R) fi0 or Ar (R)fi0 there is no hyperfine coupling from But or Ar, so that R gives a distinctive coupling. The spectrum then observed consists of a 1 : 1 :1 triplet (due to I4N, 99.6% abundance, I = 1) further split by magnetic nuclei that are one or two bonds removed from nitrogen. For example, for R = Me, Et, Free Radicals 377 or MezCH the pattern observvd is a triplct (1:l:l) of quartets (1:3:3:1), a triplct (1: 1: 1) of triplets (1 :2:1), or a triplet (1: 1: 1) of doublets (1:1), respectively. A disadvantagc of BuLNOis that it is photochemically (vis- ible) more labile than ArNO and in this way or by heating gives Bu$O [on the other hand, it is adequatdy soluble in water for use in aqueous media (see Section V,B) 1. Our spin-trapping experiments involved mixing [PtR2P2], Bu%O or ArNO, and (BU'OS)~(as a thermal source of BuL6) in benzene solution in an ESIt tube, and warming to 40" in the cavity of the spectrometer. For ButNO, this temperature was a compromise bc- tween minimal dccay of BuLNOand maximal production of Buf6. (For the BuINO experiments, thc thrrc solid components were introduced in a foil-wrapped ESIt tube uiidcr argon, and deoxygcnated benzene was added by syringe; for ArNO, solutions of known concentration of each component in C6H6were introdurcd into the tube by means of a calibratrd pipette. The foil was removvd iinmcdiatcly before placing thc tube in the cavity.) Spcctra usually grcv in iritcnsity over a few minutes at 40" and for BulNO \{ere recorded whcn thr signal for But(R)%O tias at a maxi- mum with rcspect to that for Bu$O. For the nitrosodurene (ArNO) experiments, equal volumes (0.1 nil each) of solutions in CsH6 of the Pt compound (lo-' M), (Bu'ON)~(lo-' M),and ArNO (lo-* M) were em- polycd. The results are summarized in Scheme 3, together with details of the necessary control experimcwts. A typical spectrum, with assignments, is shown in Fig. 3. The ESR results summarizctl in Schemes 2 and 3 provide evidence for the reactions, h;O>*213LltO (41) RU~O+ [i'tn,(i~~r:)~l-[i't~t(o~~~)(PR:) + iz (42) with the latter representing an SH2 process at Pt (11). Confirmation of this result was sought by attempting to synthesize a Pt(1I)-t-butoxide, as in Ey. (42), in various ways: photolytically from Bu402 using trans-[PtMe( I3r) ( PPh3)2] or cis-[PtMe2 ( PEt3)21; thermally from (BU~ON)~and ~is-[PtMc~(Pll~)~](R = Et or Ph); or by an at- tempted chain reaction using BulOCl and cis-[PtMez(PR3)2] [cf. Eqs. (28) and (29)1. These were unsuccessful. However, compounds containing Pt (11)-0 bonds are rare and largely limited to complexes having chelating ligands (14),which may achicve kinetic stability by imposing a conforma- tion on the alkoxide that is unfavorable to normal facile decomposition pathways, such as p elimination. Clearly further work is required using a system suitable for monitoring preparative-scale experiments with ESR studies. 378 M. F. LAPPERT AND P. W. LEDNOR C I FIG.3. The ESIt spectrum generated from eis-[Pt;1\len(PBu;)~], (pu QN)z, and BuLNO in CJI, at 40" [a = Buf(fiO)Me,b = Bu'(NO)OBu', c = Bu:NO]. ' PhS + [Pt (CR2SiMe3)2(P;1IeJ'h) 23 t [Pt (CH2SMe3)SPh (I'MezPh) 2] + hk3fhCHZ (44) (XVII) Propagation % hle3S1&, + Ph2Sz-Me3SiCH2SPh + PhS (45) PhS + (XVI1)-[Pt (Sl'h) 2(I'MezI'h) 21 + hk3SICfTZ (46) , hfe3Sl&~ + PII~S~-M~~SICH~SP~+ PhS (47) This mechanism was established by identifying the products of the reaction from [Pt ( CH2SiMe3) ( PMc2Ph)2] and PhzS2 at 60" in C6H6 as trans- [Pt (SPh)2 (PMezPh)z] and MeaSiCHzSPh and showing that the reaction only took place undcr similar conditions at a reasonable rate in the pres- ence of (Bu~ON)~as a free-radical initiator; additionally, in the cis- [PtMe2(PEt3)&Ph& system, in the absence of initiator but with nitro- sodurene as spin trap, a weak signal of the spin adduct Ar(Me)fiO was Free Radicals 379 BJO-N=N-OBU' kT r2ButO + N, 1. solution (ca. 0.02 mM) in C,H, Bu<. ,NO (BU~ON),(ca. 0.02 mM), BU~NO R (ca. 0.004 mM), 40", dark [+ But0(Buf) h0 + ButhO 1 2. same as (l), but no (Bu'ON), Bu$O ~.R I 3. 0.1 ml each of solutions in C,H, of P-Pt-R = Ar(RjN0 + Ar(BufO)hO I A (lO-'M), (Bu'ON), (10-'M), and P 2,3,5, 6-Me,C,HN0(10-2M), 40', dark (A) 4. same as (3), but no (Bu'ON), very slow development of Ar(Me)hO Ar(Buf0) NO The SH2 at Pt(I1)-R by Bu'O, generated thermally; summary of ESR data(29,142) SCHEXE3. The SI12 at Pt(I1)-It by 13~~0,generated thermally; summary of ESIt spin-trapping data (99, 142). detected (87).Nuclear magnetic resonance experiments on thc latter sys- tem indicated that rcarrangcment takes place during the first Me displace- ment, generating trans-[PtMc(SPh) (PEt3)z] (142), but such an inter- mediate was not dctccted for the former. The initiation may involve Ru@, fie (from p scission), or Me3SicHz from SH~at Pt(I1) (formation of compound XVI takes place at ca. 100' lower). An altcrnative mechanism may include oxidative addition to form [Pt(CHzSiMe3)z(SPh)z(PMrzPh)z][this may involve a radical chain mechanism having a five-coordinate Pt (111) intermediate] and subsequent reductivc elimination { cf., oxidative addition of p-TolSOzBr to cis-[PtMez- (PMcZPh)z] (see Section V,A) in which fie is detected, suggestive of SH~ as a minor pathway (87) 1. Benzenethiol gave methane and [Au (SPh) L], cis-[AuMez(SPh) L], and trans-[PtMe( SPh) Lz] or trans-[Pt (SPh)zLJ, with respectively, [AuMeL], [AuMc3L], and cis-[PtMezLz] (L = PMe3, PMezPh, PMePhz, 380 M. F. LAPPERT AND P. W. LEDNOR SPh 1. 0.10 mM in C,H,, I PhMe,P-Pt -PMe,Ph + Me,SiCH,SPh (ButON), (0.001 mM), I Ph,S, (0.20 mM) at 60" SPh (ca. 100% reaction by 'H NMR) CH,SiMe, I PhMe,P-Pt-CH,SiMe, I PMe,Ph only ca. 2% reaction (by 'H NMR) SH2 at Pt (II)-R by PhS, generated thermally; summary of preparative-scale experiments (also, independently by NMR) (29,142) SCHEME4. The S1,2 at Pt(II)-It by Phb, generated thermally (29, 149) Additional control experiments: (i) [Pt(PPh3)s] + 13uiS0, (ii) PPh3 + BulXO, (iii) PPhZ + BulSO + 1LIeI. Additional variations: 02,solvent, concentrations. or PPh,) in CHzClzat ca. 20", and the Au(1) and Pt(I1) reactions were formulatcd as involving a frcc-radical chain mcchanism (90). The evidence was based on thc observation (NMR) of an induction period, inhibition by galvinoxyl (l%),accclcration by AIBN, and ESR detection of But(M~)fiOwhen the reaction was carried out in the presence of ButNO, which has little effect on the rate. Two mechanisms were considered, both involving initial attack by PhS at the metal center, PhSH We have independently examined a related Pt(I1) system, with cis- [PtMe2(PR3)2] (R = Et or Bun) (142) and note that galvinoxyl readily reacts with PhSH; AIBN is not cxpectcd to be an efficient initiator under the conditions used, tllz ca. 20 hours at 60" (184) (our reaction is very fast at 20" without initiator) ; and the lack of effect of ButNO on rate may be due to an insufficient amount present as the active monomer (ButNO exists as dimer in the solid state). Free Radicals 38 1 The autoxidation of TiR:, Zrlt:’, ( MoIt:”) 2, or [W (CH,Ph) 3]2 (R’ = Mc3SiCHz,R” = Mc3SiCH2or l’hCII2, and R’” = Me3SiCHzor Me3CCH2) was studied in hydrocarbon solution at 20” or -74“ (22). The amount of oxygen consumed (0.5 molv per M-C bond) indicated M(OR)4 as the end-product but only a trace of inetal peroxide MOOC was detected. Galvinoxyl rcacts with sonw of the alkyls and, therefore, its use as an inhibitor is mechanistically dubious; however, phenothiazine (and other inhibitors) were free from this complication and retarded the reactions. The analogy with autoxidatioii of main group clement alkyls, such as BR3, is clear, for which the follou ing propagation sequence is established (50) : I? + 02-1wo (50) It00 + ]%I?I -1tOOBR2 + R (51) v ADDITION OR ELIMINATION RADICAL REACTIONS Wc are hcrc conccrricd u it11 various organomctallic reactions for which there is evidence that organic frw radicals are implicated in the reaction pathway. Many of these arc forrnally two-electron oxidative additions or their retrogressions, the reductive eliminations (Section V,A) . We shall focus attention on systems in which transition metal Group VIII complexes are involved. A. Oxidative Addition of Alkyl Halides or Related Reagents and Reductive Elimination The reaction between a low-valcnt Group VIII metal complex and an alkyl halide belongs to the class known as oxidative addition and has at- tracted much study and controversy as to the mechanism. Recent evidence suggests free radicals as intermediates in many cases. The oxidative- addition reaction is of widespread occurrence and importance in transition metal chcmistry, due in part to its iise in synthesis and to its implication in many catalytic systems. In one of its forms it is described by [Lhl] + A-13 -[LA{ (A)B] (52) Related reactions involve addition of AB without fragmentation (e.g., AB = 02)or with fragmentation into three parts [e.g., AB = MezNCHClz (SO)]. The metal complex [LM] may be cationic, anionic, or neutral, and 382 M. F. LAPPERT AND P. W. LEDNOR the addendum A--13 also covers a wide range, including H2, 02, RX, ItsSiH, or RHgX. (Reviews covering both mechanistic arid preparative aspects of the reaction arc in Refs. 24, 39, 40, 54, 59, 81, 82, 89, 119, 167, 182.) We shall not duplicate this matclrial, but aim to present a brief ac- count of the main dcvclopmcnts in the mechanistic work and then to dis- cuss recent cvidcnce implicating radicals. 1. Afeckanistic Ideas Piior to I972 The addition of an alkyl halide to the d7 complex [CO(CN)~]~-results in a one-electron oxidation of the nwtal and procecds by homolytic ab- straction of halogen. For a mctthyl or bcrizyl halide, an organocobalt product is formed, [Co(CS) ,]I- + Irx-[(’o(S) (CN).]7- + R (53) “’0 (CN) 4]9- + R- [co (it)(cs) J- (54) whereas for other halidcs olcfiiis may be formed in a competing reaction, w-7 [(~o(~~)J--I- P2ir4-[v~(~~) (w) ,]3- + CJI, (55) These and othrr additions to cZ7 coinplcxes are revieacid in Refs. 59 and 89. Most of the mechanistic v:orl< on oxidative addition to cZ8 complexes has been concerned n ith Ir(1) and in particular with Vaska’s compound, trans-[Ir(CO) (Cl) (PPh3)2], or its analogs, r.g., [It (C‘O) ((’I) (1’1’1i3)~~] + lIeI-[Ii (CO) (Xle) (CI) (I) il’l’hj)1] (56) An early kinctic study (34, S(T also I77a) revcaled features similar to that of the Mcnschutkin rclaction (RSN + R’X-R3R’N+X-) , and it was suggested that thc Ir(1) coniplex behaved as a nucleophilc in an SN~type- mechanism, ,Me LM + Me1 - LM6+ .H\c’f.. 16- - [ 1 LM.. I (37) LM = [Ir(CO) (C‘l) (PPh,),] Latcr work showrd that no incorporation of added anions occurred during thr addition of MrI to lrans-[Ir(CO) (C‘l)P,] (I’ = PPh, or PMcPh2) and that solid Ir(1) complcxcs addcd gaseous MPI or McBr, suggestive of a nonionic mechanism (161) ; a molecular process, involving thc transition state (XVIII) was proposrd. Evidence was prcwnted for retention of con- figuration at C1 of optically active MeCHBrC02Et, accompanying the Free Radicals 383 (XVIII) addition (161). However, objc>ctions to this [later experimentally con- firmed (137)] and also to a claim (136) for inversion [later withdrawn (21)]have been published (114). A kinetic study of alkyl halide addition to Ir (I) has favored a polarizcd or unsymmetrical, three-center transition state (f83). Kinetic studies have shmvn that addition to the ,Po [Pt (PPh)n] (n = 3 or 4) or to [Pt(CzHa) (PPli3)z]are first-order in both addendum and metal complex; the lather may initially lose ethylcne or phosphine (89,162). 2. Recent Developments In 1972, a radical chain Incwhanisin was suggested for addition of PhCHFCHzBr (or dcutero aiialogs) to trans-[Ir (CO) C1 (PMe3) 23 to give [IrBr(CO)Cl(CHzCHFPh) (I’MC~)~](21). [A radical nonchain mecha- nism had been accepted for oiw-elcctron oxidation of Co(I1) (89).] The evidencc for radicals was (i) initiation by 02, AIBN, or BzzOz; (ii) retarda- tion by duroquinoric or hydroquinonc (radical inhibitors) ; (iii) loss of a specific stereochemistry at (deduced from NMR spectra), [the first positivc demonstration of rawmixation during oxidative addition involving Xi, Pd, or Pt was in thc rc:tction of (+) RCHBrCOzEt (It = Mc or Ph) to Pd(BulNC)z (156)];and (iv) a rcwtivity order in which thc ratc of addition of halide was incrcascd by its having clrctronegative substituents. The following mechanism was proposed, by analogy with additions to Cr (11) and Co (11) (cf. Rcfs. 54 and 89) : Q + [I1 (111- CIr (II)--QI (58) [I1 (11)-Q] + IU~~-[BII~(III)-~] + R (59) [Ir(I)] + R-[Ir(II) - R] (60) [Ir (11)-R] 3- Rlh- -[BrIr(III)-R] + R (61) Equations (58) and (59) rcprcsent initiation by an unknown Q, and Eqs. (60) and (61) the propagation sequcnce. This work was extended as follows (133),(137). The addition of optically activc MeCHBrC02F,t citlic~to [Ir (CO)CIPz] (1’ = PMc3, PMczPh, or YMcPhz) or to Pt (PPh3)3(IS?‘), or the chloride to [M (PEt3)J (M = I’d or Pt) (133) gave the rawrnic adduct [Ir (Br) (CO) (Cl) (CHMeC02Et)- Pz], trans-[Pt (Br) (CHMeCOzEt) (I’I’h3) 21, or trans-[M (CI) (CHMeC02- 384 M. F. LAPPERT AND P. W. LEDNOR Et) (PEt3)21, respectively. Inhibition by galvinoxyl was found, which was taken as evidencc for a chain mcchanism. The oxidativc addition rcaction may bc a complex composite. For cxamplc, [Pt(PEt3)3] and Bun& in PhMe at 25" reacted (133') as follons trans-[l't (Br)Bun(PEtp) 21 PhMe. 23' 95yo [Pt (PEts) J] + Bu7'Br tmns-[l't (Br)€I (PEt3) (62) 4% truns-[Pt13r2(PEt3)2] 1% during the first 2 hours; further reaction affordrd the dibromide at the expense of [Pt (Br)Bun( PEt3) 21, whereas the hydridc incrcascd to a maxi- mum and then decreased, by reacting with BunBr to yicld the dibromide. For a see-alkyl bromide, the hydridc was the major product in the absence of excess RBr. With neopentyl bromide in toluene, trans-[Pt (Br)CHzPh- (PEt3)23 was obtained in high yield, whereas with BrCH2(CH2)3CH=CH2, the ratio of cycloalky- to n-alkenyl-Pt (11) complex was 3: 1. Evidence for a radical mechanism is accordingly considerable, but indication sup- porting a chain mechanism was limited to inhibition by galvinoxyl or duro- quinone. Since these scavengers react with the Pt(0) complex, caution is required in interpreting thc observation. We are not convinced that the demonstration of inhibition, cven without this caveat, would inevitably support a chain, rather t)han a rionchain, mechanism; for this to be more persuasive a zeroth-order dcpcndencc on inhibitor would bc helpful. The next major line of evidencc for radicals as intermediates in related reactions came from our laboratory, whrre it was found that the ESR technique of spin trapping provided for the first time spectroscopic evi- dence for the intermediacy of free radicals (see Schemc 5) (139).When the addition of C&I, CnsI, EtI, PhCH2Br, or Ph2CHBr to [Pt(PPh3)3] was carried out in the presrnce of Bu'NO, strong signals for thc corrcspond- ing nitroxide But(R) $TO were obtained; analogous experiments with 2,3,5,6-Me4C6HN0 as t hc spin trap have yielded similar results (142). Control experiments showed that thc signals were not derived from either of the reactants, nor from either of the products (metal complex and phos- phonium salt), nor from the reaction of alkyl halide with phosphine. Furthermore, CPt(PPh3)z-J with Me1 in the presence of Bu"0 still gave [PtMe(I) (PPh3)2] as the major product. Reaction of [Pt(PPh,)3] with benzhydryl bromide proceeded according to Eq. 63 below and with Ph3CC1 gave [PtC12(PPh3)z]and Ph&; the latter is stable and was identified by ESR without a spin trap. {An attempt to observe the moderately stable Free Radicals 385 C6H6,dark immediate But [Pt(PPh,),] + RX t BU~NO ‘NlO 20” ESR siKnal R/ 5 X lo-* M (R = CH,, CD,, Et, 5 x 10- M or PhCH,) 5 X M either: C,H,, dark, 20” or: + PPh,(5 x 10-,M), C,H,, dark, 20” L 1 I t I 1. + excess of RX until no further color change No immediate ESR signal, 2. then f BdNO but slow (0.5. hour) develop- ment of BU$NO SCHEME5. Summary of ESI1 spiii-t,rapping experiments (similarly for 2,3,5,6- !tIc4C6HXO)for the [l’t(PPh3),]-RX wartinn (159, 14%’) ally1 radical directly in the reactlion of [Pt (PPh,) 3]/C3H5Br in dilute PhMe solution at low temperature (when PhMe is rather viscous) failed (142).] The mechanism we proposed is rionchain (139): [l’t (I’l’ha)3]-[1’t,(PPha)2] + PPhs (64) RlOW [Pt(PI’h3)2] + RXe[l’tX(PPh3)2] + R (65) (XIX) This mechanism readily accounts for formation of [PtXzP2] or [PtX(H)P,], compounds often found in these reactions, e.g., Eq. (62), by H or X abstraction by the Pt (I) complex (XIX), and is consistent with the second-order kinetics (89, 16%). Similar spin-trapping experiments have been carried out for the addition of PhCH2Cl or PhMeCHBr to CPd(PPh3) 4) (192).Nitroxides were generated, and their most likely origin was the metal product, since the combination [Pd(Cl) (CH2Ph)(PPh3)J and Bu‘NO generated a signal for PhCH2(But)fi0. However, we had previously reported that the analogous experiment with [Pt (PPhB)31, PhCH2Br (or other RX) , and ButNO does not give an ESR signal (139) ; this difference between the Pd and Pt systems is not surprising in view 386 M. F. LAPPERT AND P. W. LEDNOR of the greater lability of Pd-C than Pt-C bonds (94).Thus, cis-[PtMez- (PEt3)J may be distilled at S5°/10-4 mniHg (31), but cis-[PdMe:!(PEta)z] decomposes completely at 100". Palladium (11) alkyl complexes with Ph3P as a ligand are even less stable,e .g., cis-[PdM~z(PPh3)~]decomposed at 35" to 40" in solution (ZG).I Other oxidative addition reactions investi- gated by the spin-trapping technique using nitrosodurcne Kcre on (i) [Co (11) (dmg)zPPh3]/C'H2=C:HCHzBr (ding = dimethylglyoximato) and (ii) cis-[PtMe:!(PMezPh)z]/p-TolSOzBr (to give [Pt(IV)BrMc2- ( SOzTol-p) ( PMe2Ph)2] in (ii) } . Compounds Ar (Me)fi0 and Ar (p- TolS02)fiO were observed at +30°, but only Ar (Me)$TO possibly from SH~at -30", which lost intensity at thc expense of Ar(p-TolSOz)RO at +30° but reappeared with concomitant collapse of thr sulfonyl nitroxide (perhaps from reductive dimination) at +40", and control experiments ruled out alternative sources of these nitroxidm (87).In the [Pt (PPh3),]- CCI, system (see below), a weak signal [about lY0 of the intensity of Ar (Me)fi0 from Pt (0),/Me11 duc to Ar (CI3C)fi0 was detected (142).In the [RhCl (PPh,) $]-Me1 or trans-[Ir (CO) C1 (PPh3) 2]-McI systems, Ar(Me)fiO was not observed after 1 hour (142). The proponents of the chain mechanism have revised their proposal (132): they now believcl that for [I't(PEt3)3] as substrate, a radical pair is formed, as in Eq. (65) (139) which may collapse to the adduct [PtX( R)- (PEt3)2] as in Eq. (66) (139),but the adduct may alternatively (e.g., for PhCHzCl but not PhCH2Br) be produced by an sN2 displacement of X by Pt(0) from RX (cf. Ref. 141).For a "reactive" halide (c.g., an a-Br-estcr, PhCH2Br, or sec-RI) , t h(1 dihalide [PtX2(PEt3) 21 is no longer thought (132) to arise via [PtX(H) (PEts)2] (itself formed by a radical-chain process) but by halogen abstraction from [Pt (I) (R) (PEtz] (139), and this concept was supported by CIDNP data (132) on the coproduct (e.g., 'H NMR enhancements in thc resonances for CH2=CHMe and PrzI in the Pt (0)/Pr21system). These CIDNP efTects were observed only when the dihalidc was rapidly produced in thc cmly stages of the reaction, e.g., for PrzIbut not PrzBr.For an a-chlorocstcr, the original chain mechanism, in which Eqs. (63) and (64) (139) provide the initiation, is still held (132) to proceed, cf. Eqs. (60) and (61) (21). Evidence for inversion at C1 in thc addition of optically active PhCHDCl (192) or PhMcCHBr (141) to [Pd(PPh3)4] has been cited in support of an sN2 mechanism. However, racemic adducts have been obtained in the 1 An alternative, but coilsidered (192) a less likely possibilit,y, was that RutNO induced radical decomposition; the rase of an organosilver compound ~vascited as precedent (190). IIowever, we note that in that work a nitroride and iiot a spin trap was used. Free Radicals 387 addition of MeCHBrCOzEt or PhC‘HDrCO2Et (this rules out the possi- bility of a racemate being fornicd via a u-T rearrangement) to [P~(CNBU~)~](156) and of Ph(’HC1(:F3 to [Pd(PPh3)4] (I41),for uhich the reaction (SN2 by Pd(0) at Cil of l’d(I1)) was proposed as a possible alternative to a free radical proccw: It The secondary alliyl halides EthlrC‘HBr, EtMcCHI, PhRlcCHBr, or MrCHBrCOzEt and [PtP,] (I’ = l’Ph3, PMczPh, or PEtJ; 7) = 3 or 4) gavc. halide complexes CPtX21’21, rat1ic.r than a 1: 1 adduct (160). Although one-electron changes havc. only recently been rccognizcd in the contcxt of oxidative-addition rwctioris, thr ability of a metal, or complex, to abstract halogen honio1ytic:rlly appvars to be widespread. Thus, reactions of Na nith 1IX in the gas phas(i (It?’?’), of Mg with RX (CIDNP cxperi- ments) (18),of Ag(0) or dg(1) with Alr-chloramines(GO), and the transi- tion mctal-catalyzed addition of (‘(’1) to olefins (148),all procced, at lcast in part, 1)y halogen abstraction. Additionally, the combination of a low- valcnt trarisitioii mctal complvx arid an organic halide has bccn cxtensively investigated as a free radical-initiating system for vinyl-monomer poly- merization (8).The initiating stcy was considered to be electron transfcr to the halide, followed in sonic CYMCS by a choice of pathways (9a),such as (4 For [Pt (PPh3)4] and CC14, rcaction (a) was a minor pathway (cf. spin- trapping, abovc.) (96) ; it., only a small amount of CCl, was formed (96) with a molccular mechanism as tlic. major route. Further support for clcc- trori transfcr from dlo species \\’as providcd by identifying thc organic rad- cal anion obtained by mixing a Ni(0) or Pt(0) complex, such as [Ni(PEt3)4], with an electron arcvptor, such as C2(CN)4(64). Finally, it is noted that polar c+fccts oftcn found in oxidative additions (usually such that increasing thc elcctron density at the metal, or dc- creasing it at C1 of a halide, facilitates the reaction) are not inconsistent 388 M. F. LAPPERT AND P. W. LEDNOR with radical reactions. Similar features have been found and rationalized in the homolytic abstraction of halogen by tin-centered radicals (149). It is likely that many reductive eliminations also proceed by radical pathways, but few studies have been reported. The reaction of [PtRzL,] (R = Me, CD,, or Et, and LS = bipy; or R = Me and Lz = phcn or 1,5- cyclooctadiene) with diethyl fumarate or maleatc in the dark at room tcmperature in the presence of BulNO lcd to thc detection (ESR) of com- plex XX, which was interpreted in terms of the following sequence (91) : [l’tRzLz] + (CH(C0iEt)] 2-[Ptll*L2(CI1(COzl:lt) ).I t [l’tLrlCII(COslSt) 121 + 2R (69) 211CH(C02Et)CII(COiEt)-rczs- 01 f~ccns-(CH(COJCt)IL 1 (SX) However, a Pt (I) species could be an int ermediatc. The insertion reaction of trans-[Pt(X)MeLz] (X = C1, Br, or I; and L = PMczPh) with RCE CCOzMe in CHCI, afforded trans-[PtXLz(C(COzMe)=C(C1)R)] in the presence of RzzOz as a radical initiator (R = C02Me) or by addition of HCl (R = COzMe, Ph, Me, or H) (5);thc proposed mechanism involved initial formation of the 1: 1 adduct with the acetylene, followed by nucleo- philic attack by HCl generated by a radical process. The homolysis of a metal alkyl may bc regarded as a reductive elimina- tion; however, it is convenient to consider such rclactions separately (Sec- tion V,B) . Electron-transfer mechanisms for organometallic intermediates in catalytic rc.actions have been reviewd (129) ; examples are in thr forma- tion of transient RCu(1) or RCr(II1) in oxidation (by Cu(I1)) or reduction (by Cr (11)) of A, and in the role of Fe in the Kharasch-Grignard reaction (e.g., Fe catalysis of disproportionation of EtMgBr + EtBr - CzHs + CzH4,via Fe(1) + RBr-Fe(I1) Br + A). B. Metal Alkyl Photolysis or Thermolysis It was formerly assumed that homolysis is a common mode of decom- position for metal alkyls. In fact it is rather rare (49).A detailed survey of thc topic is available elsewhere (49).For main group element alkyls, the clearest case is the thermolysis of Hg (11) or Pb (IV) compounds, and other examples are for alkyls of Zn, Cd, Si, Ge, Sn, and methyls of Ga, In, T1, As, Sb, and Bi (also a minor pathway for B and Al). In the transition metal series, data are available for Mn (I), Ni (11), Pt (IV) , Cu (I), arid Ag (I), with homolysis playing a minor role for Ti(1V) and Zr(IV). Evidence is Free Radicals 389 based on stercochcmical argiinicmts, tlir isolation of products appropriate for radical reactions (c.g., meth~lcS.clop(.ntancfrom M--CH2CHnCH2CH= CHMc compounds), the cffcct of radical trarisfcr agrnts (c’.g., aminc.) or inhibitors, dcuterium-labeling, and C‘IDNP or EST1 rxpcrimcnts. Therc swms to be only oiir rxrtmplc. of direct ESIl ohscrvation of an organic radical dcrivcd from pliotoly& of a mctal alkyl, the stable c ( SiMc3) from [Hg { C ( SiMv3) ] 2] (12). Howevrr, spin trapping by nitrosodurcnc (ArNO) has hrm uscd, for UV irradiation of [Mn (CO)5- CH2Ph], whcn compound XI11 (Scctions II,B and 111) and Ar(PhCH2)k0 (105) wcrc drtected (ESR). The acylpcntacarbonylmanganrsc(I) com- pounds [Mn (CO)5( COR) ] (11 = PhCHz or Ph2CH) undcrwcnt similar homolysis, although the spin adduct formcd was gcnrrally dcrivcd from ft rathcr than ftC0. Other cxaniplc~sof spin trapping the photolysis products of metal alkyls rclatc to Sn(IV), Pb(IV), and Hg(I1) (cf. Ref. 11.9). Homolysis of M-R bonds may well play a role in a numbcr of catalytic proccsscs (1.29). The photolysis of c‘o (111) alkyls has attractcd considerable attrntion (59),becausc homolysis of such a bond may wrll bc implicated in eiizym- atic isomcrization reactions catalyzcd by vitamin BIZcocnzymr (scr Refs. 27, 98, 115a, 193). Early evid(mcc was basrd on characterization by UV or ESR of the d7 Co(I1) corriiwid. Hotwvcr, recently the organic fragment has brcn spin-trappcd and idcntifird by ESR for the case of (i) the co- enzyme (5’-dcoxyadcnosylcobalamin) ( M) with BuWO ( lop2 M) and photolysis in water at 50” to yield Ar (j’-deoxyadenosyl) fi0 (XXI) HO OH Ar-N-C0I1 ylGH2 (XXI) [Fig. 4; a(N),1.64 mT; a(H1),1.41 mT; a(H2),0.81 mT; and a(H3),0.06 mT] and Co(I1) (g - 2.2); and (ii) cthylcobalamin, which, under similar conditions but at 20°, gave But(Et)fiO [a(N) 1.71 mT and a(H) 1.13 mT] and Co(I1) (g - 2.2) (115). The anaerobic photolysis of alkyl- cobaloxime-pyridine adducts has been studied by ESR [Co (11) identifi- cation] and shown to be a homolytic process for Pri, Bui, n-CjH11, or cyclo-CcH1l, but for Me or PhCHz it is a one-electron transfer process in- 390 M. F. LAPPERT AND P. W. LEDNOR -1 rnT Fro. 4. The ESR spectrum of I3u t(ri'-deoxy~denosyl)~Ogenerated by photolgsis of vitamin Blz-coenzyme in the presence of 13urSOin €120 at 50". volving the equatorial ligands or solvent (78) . Spin-labeled 5'-deoxy- adenosylcobinamide, shown schematically in XXII, has been used as a co- (XXII) factor for the enzyme ethanolamine-ammonia-lyase and the ESR spectrum followed during catalysis (193). The ESR signal disappeared upon adding the substrate ethanolaminc, which indicates that the Co (111)-R bond may well cleave homolytically upon adding the substrate. VI APPENDIX Bricf reference is made here to some papers which have appeared since the submission of the manuscript in February 1975; this section was added at the proof stage, in December 1975. Free Radicals 39 1 Section 11,A : The photolysis of hydrides of XX’X”Ge-H in prcseiicc of BLI~O~has been reported (1710); 1,hcnyl-substitutrd radicals may be more planar than alkyl analogs, and th(w is significant dclocalization of the un- paired rlcctron into the aromttic ring. Further examples of bulky stable Group IV metal-centcrcd radirals have been found: &I(CsH2Mc3-2 ,4,6) (M = Ge or Sn), ?&[N((;chh)~]~(M = Ge or Sn), and Sn[N- (GrEt3)2I3(86c) ; the aniido-metal compounds were obtained by photolysis of corresponding M[N (GeK3)2]2dcrivatives [cf. Eqs. (8)-(11)], but this method failed to yield radicals froni Cr[S (GeEt3)2]2 or M[N (GcPh3)2]2 (M = Gc or Sn) , possibly for stcric rcasons (86~).A ncm high-yield method has been discovered (8th) for tho fnination of stable metal-centered radi- cals from a halide and an clec.tron-rich oldin, e.g., (~Gc), + 6e(C,jIIJIc3-2,4,6)3 Section II,B: Paramagnctica organometallic actinide and lanthanide complexes u ere omitted froni thci prcvious discussion and Table IV. How- cvcr, reference may be mado to rcvicn5 (49, 86d, 94a, 1460, 1466) ; corn- pounds includc [M(C5H5-vj2X] (c.g., X = C1, OPh, or Oh; M = a lanthanidc or U), [M (C5H,-q)3] (M = :t lanthanidr, U, Pu, Am, Cni, Bk, or Cf), [M(CsH5-q)2] (hl = Eli, Sin, or Yh), [M(C~H.,-~)C~Z.~THF] (M = Sm, Eu, Gd, Dy, Ho, or )*\I), [AT (C8H8-q),liqd. NHJ (M = Eu or Yb) , [U ((’>Hs-v)J, [M (C8H8-q)2] (M = Th, U, or Np), [M (CJL-v)3XI (X = OK, BHI. 11, or C1; M = ‘rh, U, or Np). Further intcrcsting species arc the alkyl-bridged XXIII (hlr = Yb, Er, Ho, or Dy) (986) and XXIV (ill = l’b, Er, Ho, Dy, or (:(I) (98~)and [Y~I{CH(S~M~~)~]~.~THF] (98a); other lanthanidc coniplcxes [M (CjH5-v)zR]n have been described (64-4a). Me Me /\ /\ (v-C&,),M M(C,H,-?), (v-C5HJ2M AlMe, ‘Me/ ‘Md (XXIII) (XxrV) A further contribution has been inadc to the controversy relating to the nature of thc long-lived par:trnagnetic species obtained by photolysis of [Mii2(CO),] in THF, suggesting this to br [Mn (CO),] (135~); however, detailed ESR line-shape analysis, Ill detection of [Mn (CO)5]-, and anal- ogy with the Fe (CO) b/basc\ tern supports the [Mn (THF)J2~[hlri1- (CO) 512- assignment (1050, 105b). Further photolysis studim rcifcr to [Mo (CO)3(C5Hj-v)2] (105c, 19/tb),[Mn2 (CO)l0] and [Re2 (CO)l~]and re- 392 M. F. LAPPERT AND P. W. LEDNOR lated dcrivativcs (194~).l’araniagnctic intermediates were proposed in kinctic schemes rolating to ligand (phosphine or phosphite) substitution in [Mn (CO) (PI’h3) l2(0’70) or [IlrH (03)J (25~).A Co (IV) complex re- sulted from c~lcctrochmnicalor Cc (IV) oxidation of a Co (III)-R cobalox- ime spccies (8%). Sections IV,A and IV,B: A CIDNP study of SH~reactions has been re- ported (110‘~)for [M (L) It]/ (PhC0)202 systcms, by heating an appropri- ate mixturc { [M (1,) I3 = SnlCt4, SnMc3C‘1, SnBii:Br, PbEt,, PbMesC1, AuMe (PPh,), or czs-PtMc21J;) in an NMlt tube and monitoring thc PhC021t, I’hK, or C2H6 (from It = Etj signals; the reactions afforded a metal bcnzoate, COz, and products from i.11 or il. Section V,A: Halide exchange, c.g., M(.I/CH2C12 at 100” under 150 lb/sq. inch of Ar or (20, was catalyzed by [IiuC12(l’Ph3)3], [RhCl- (PPh3j3], [Ith (CO)C1 (1’1’hJ)2], or [Ir (CO) C1 (l’Ph3)J, and this was ascribed to sequmtial oxidativc,-additioiis and reductive climinations (144a); but anothvr study attributw catalysis to halide ion formed by quartcrnization of frw PPh3 (72b). Halide ion catalysis was observed in the addition of alkyl halides to somv Hh (I) substrates, e.g., for [Rh (C0)- Cl(Q)z]/McI nith [Bu;“N]+I- for Q = As, Sb, (72a) but not P. Alkyl or aryl halidcs R”X oxidativdy add to (a) SnItn or (b) Sn(NK;)2 [lt = (Me3Si)&H, It’ = MrJSi] to yidd SnRz(It”) X or Sn (NR;)z (It”)X (86b). These rmctions arc’ radical in character as shown by ii) the spin trapping and ESlt charact crization of Ar (R”)fi0 (Ar = 2,3,5,6-Me~C,H)for (a), (ii) thv ESR dctcction of a tin-cchnttwd radical, probably hRz(X) for (a) and (iii) thc obtaining of a raccmic product in (a) when R”X = (+)-n- C6H13(Me)C‘HCl, and (iv) tlw catalytic effect of a trace of EtBr in the oxidative addition of PhBr, for (a) or (11). Section V,B:A furthrr ESR papclr has appeared on the photolysis of a Co (111)-R cobaloximc. (78a) and similar ohswvations have hccn madc on the corrcsponding vitamin BIZcocnzymc in propane-1 ,3-diol (28). Photoly- sis of a dimethylcobalt (111) chelatc. gave CH4 and a methylcohalt (11) coniplcx which was bdievcd to disproportionatr to a Co (I) and a Co (111) specim (191~). ACKNOWLEDGMENTS W’e thank Messrs. T. I,. IIall and J. J. hIacQuitty, and Drs. 31. J. S. Gynane, A. Hudson, R. A. Jackson, and B. I<. Nicliolson for useful discussions. REFEREKCES 1. Abell, 1’. I., in “Free Radicals” (J. I<. Kochi, ed.), Vol. 2, Chapter 13. Wiley (Interscience), Kew York, 1973. 2. Alexandrov, Yu. A., J. Organometal. Chenz. 55, 1 (1973). Free Radicals 393 3. Alyea, 15. C., Bradley, I$. C., I>:ippert, 11. F., and Sangrr, A. R., J. Chem. Soc., Chem. Comnzm. 1064 (1969). 4. iinderson, 0. P., and Symons, 31. C:. lt., J. Chem. Soc., Chenz. (‘ommiin. 1020 (1972). 5. Appleton, T. G., Chistiolm, 3f. II., and Cl:trk, H. C., J. Amer. ChenL. Soc. 94, 8912 (1972); Appleton, T. G., Chisholni, AT. II., Clark, €1. C., and Yasufuku, I<., ibid. 96. 6600 (1974). 6. Ayscough, 1’. B., “F2lectron Spin 1tesoii:rncc in Chemistry.” >Irthuen, London, 1967; Atkins, F’. IT., and Synions, 11. C. I<., “The Structureof Inorganic Radicals.” I~~lsevier,.4msterdam, 1967. 6a. Uaallen, A. van, Groenenboorn, C. *J., i~nd1,iefde Meijer, H. ,I. de, J. Organometal. Chem. 74, 245 (1974). 7. Bakalik, 1). P.,and Hayes, R. G., Inorg. (‘hrm. 11, 1734 (1972). 8. Barnford, C. H., in “Reactivity, lIcc~lianisin,and Structure in Polymer Chemistry” (.4. D. Jenkins and A. I,edwith, eds.), Vol. 3, p. 52. JVilcy (Interscience), New York, 1974. 9. (a) Barnford, C. TI., I~~astmorid,(;. C., :~nd\Thittle, D., J. Orgenometal. Chem. 17, 1’33 (1969); (b) Barnford, C. H., I~hstinond,G. C., and Hargreaves, K., Trans. Frcrad. Soc. 64, 175 (1968). 10. Thrker, G. K., and Lappert, Rl. F., J. Or(qnnometal. Chena. 76, C45 (1974). 11. Bartelink, II. .I. M.,Ostendorf, 1%.Ti., Itoest, B. C., and Schepcrs, H. .4. J., J. Chem. Soc., Chem. Comnzun. 878 (1971). 12. Bassindale, A. R., Bowles, A. J., rook, hl. A,, Eaborn, C., Hudson, A,, Jackson, It. A4.,and Jnkes, A. I<., J. Chenr. Soc., (‘hem. Commzin. 559 (1970). 13. Bennett, J. I<., and IIoward, J. A,, Chprr,. I’hys. Lett. 15, 322 (1972). 14. Bennett, M. A,, Robertson, C:. H., \!‘tiimp, 1’. O., and Yoshida, T., J. Amer. Chem. Soc. 95, 3028 (1973). 15. I3ennett, S. II-., Eaborn, C., IIurlson, A,, Jackson, 11. A,, and Root, I<. D. J., J. Chant. Soc.., A 348 (1970). 16. Bennett, S. IV., Eaborn, C., 1-Iudaon, I\., Hussain, 31. A., and Jackson, It. A., J. Organometd. Chena. 16, 1’36 (1969). 17. Bentrude, IV. G., in “Free Radic:~ls” (J. K. Kochi, ed.), Vol. 2, Chapter 22. IViley (Interscience), Sew York, 1973. 18. Uodewitz, FT. IV. H. J., J’Jombrrg, C., and Bickelhaupt, F., Tetrahedron 29, 719 (1973). 19. Bower, B. K., and Tennent, 11. (i.,d. A4rn~~r.Chem. Soc. 94, 2512 (1972). 20. Bowles, A. J., Hudson, A., and Jaekson, 11. A,, J. Chem. Soc., B 1947 (1971). 21. Bradley, J. S., Connor, I). E,, Ilolphin, I),, Labinger, J. A,, and Osborn, J. A., 1.Anzer. Clheni. Soc. 94, 4043 (1972). 22. Rrindley, 1’. B., and Hodgson, J. C., J. Organometal. Chem 65, 57 (1974). 23. 13r(iwn, 11. C., and Jlidland, L1. AT., Artpio. Chem., Int. Ed. Engl. 11, 692 (1972). 24. Burgess, J., Inorg. React. Afrch. 1, 311 (1071). 25. Burkett, A. R., Meyer, T. J., and \Vhittt:n, D. G., J. Orgunometal. Chem. 67, 67 (1974). 25a. Byers, 13. II., and Brown, T. l,.,J. Anzcr. Chem. Soc. 97,947 (1975). 26. Calvin, G., and Coates, G. K., J. (‘hem. Soc. 2008 (1960). 27. Cardin, D. J., Joblin, K. N., Johnson, A. W.,Lang, G., and Lappert, Rf. F., Biochim. Biophys. Acta 371, 44 (1974). 28. Cardin, D. J., Joblin, K. N., and Lowe, 1). J., unpublished work (1975). 29. Cardin, D. J., Lappert, M. F., and Lednor, 1’. W., J. Chem. Soc., Chem. Commun. 350 (1973). 394 M. F. LAPPERT AND P. W. LEDNOR 30. Cetinkaya, U., IAappert,31. F., lZcLaughlin, G. AT., and Turner, I<., J. Chem. Soc., Dulton ’I’rtrns. 1501 (1974). 31. Chatt, J., antl Sliaw, II. L.,J. Chm. SOC.705 (1959). 32. Chen, I<. S., T3atlioni, J.-P., and Kochi, J. I<., .I. A4?7s~r.Soc. 95, 4439 (1973). 33. Chen, I<. S., T3ertini, I?., and Kochi, J. I<., J. .4mer. Chertz. Soc. 95, 1340 (1973). 34. Chock, 1’. I<., and Hslpern, J., J. Bnicr. Chern. Soc. 88, 3511 (1966). 35. Coates, G. I<., Green, If. I,. H., and \\-adc, K., “Organometallic Compounds,” 3rd ed., Vol. 1: “The Main Group 1‘:lenients’’ by G. 13. Coates and li. \Vade. Methuen, I,ondon, 1967. 36. Coates, G. I,;., Green, 11. 1,. H., and \Vade, K., “Organometallic Compounds,” 3rd ed., Vol. 2: “The Transition ISlenicnts” by 11. L. €1. Green. Methuen, London, 1968. 37. Cohen, A. €I., and Hoffman, 13. J1., Inorg. Ch~m.13, 1484 (1974). 38. Collier, 11,It., bMorn, C., Jovanovi6, B., Lappert, 11. F., Manojlovii.-Aluir, Lj., Muir, K. JV., and Truelock, M. hl., J. Chenl. Soc., Chem. Comnaun. 613 (1972). 39. Collman, J. l’., Accorrritx Chrnz. Res. I, 136 (1968). 40. Collman, J. l’., and Roper, JV. R., Arlzw~.Orgnno?rletal. Chem. 7, 53 (1968). 41. Cooper, J., lIudson, A,, and .Jackson, R. A,, Mol. Phys. 23, 209 (1972). 42. Cooper, J., Iludson, A,, antl Jackson, R. A,, J. Chcni. Soc., Perkin Trans. I1 1056, 1933 (1973). 43. Corrca-Duran, F., Allred, A. I,., and IJoyd, R. J., J. Organonzotal. Chem. 49, 365, 373 (1973). 44. Cotton, J. I)., Chndg. C. S., Harris, D. €I., Hudson, A,, Lappert, M.F., and Lednor, 1’. \\r., .i.(“hem. Soc., Cheni. Cornrnun. 651 (1974). 45. Danen, \V. C., and \Vest, T. C., l’efrahcdron Lett. 219 (1970). 46. lhvidson, 1’. J., Hudson, A,, I,nppert, 11. F., and I,etlnor, 1’. \V., J. Chent. Soc., Cheni. Coinmun. 829 (1973). 47. Davidson, 1’. J., and I’appert, 32.1. I?., J. Chert[. Soc., Chem. Conitn?tn. 317 (1973). 48. Davidson, 1’. J.,I,appert, XI. F., and l’earce, R., Accounts Chern. XES.7, 209 (1974). 49. Davidson, P. J., Lappert, hl. F., and I’earce, It., Chern. Rev. 76, (1976)., in press. 50. Dsvics, A. G., and Roberts, U.l’., Accounls Chem. 12e.s. 5, 387 (1972). 51. Davies, A. G., and Roberts, H. P., in “Free Radicals” ( 62. J;lsc.henhroich, C., and I, F., .I. Orgunonzetnl. Chenc. 49, 445 (1973); I.:lsclienhroich, C. H., Gre ., aiid St,ohler, F., J. Anzer. Chenz. Soc. 95, 6956 (1973); Karthe, \\-., and ;~c.htcr,\V., Z. I‘h,ys. Clhcm. (Leipzig) 247, 2.11 (1971). 62:~.I 89a. Halpern, J., Chan, M. S., Hanson, J., Roche, T. S., and Topich, J. rl., J. Amer. Chem. Soe. 97, 1606 (1975). 90. Hargreaves, N. G., Johnson, A., Puddephatt, R. J., and Sutcliffe, L. H., J. Organo- metal. Chem. 69, C21 (1974); Johnson A,, and Puddephatt, R. J., J. Chem. Soc., Dalton, Trans. 115 (1975). 91. Hargreaves, N. G., Puddephatt, R. J., Sutcliffe, L. H., and Thompson, P. J., J. Chem. Soc., Chem. Commun. 861 (1973). 92. Harris, D. H., and Lappert, M.F., J. Chem. Soc., Chem. Commun. 895 (1974). 93. Harris, D. H., and Lappert, M. F., unpublished work (1974). 94. Hartley, F. R., “The Chemistry of Platinum and Palladium.” Applied Science Publ. London, 1973. 94a. Hayes, R. G., and Thomas, J. L., Organometal. Chem. Rev. 7A, 361 (1971). 95. Henrici-Olivit, G., and OlivB, S., J. Chem. Soc., Chern. Commun. 1482 (1969). 96. Henrici-Olivit, G., and OlivC, S., J. Organometal. Chem. 23, I55 (1970). 97. Henrici-Olivi., G., and Olivit, S., J. Amer. Chem. Soc. 92, 4831 (1970). 98. Hill, H. A. O., in “Inorganic Biochemistry” (G. I. Eichhorn, ed.), Chapter 30. Elsevier, Amsterdam, 1973. 98a. Holton, J., and Lappert,, M. F., unpublished work (1975). 98b. Holton, J., Lappert, M. F., Ballard, D. G. H., Pearce, R., Atwood, J. L., and Hunter, W.E., J. Chem. Soc., Chem. Comm7~n.in press (1976). 9%. Holton, J., Lappert, M. F., Scollary, G. R., Rallard, D. G. H., Pearce, R., At- wood, J. L.! and Hunter, W.E., J. Chem. Soc., Chem. Commun. (1976). 99. Hosomi, A,, and Sakurai, H., J. Amer. Chem. Soc. 94, 1384 (1972). 100. Huber, II., Kundig, E. P., and Ozin, G. A., J. Amer. Chem. Soc. 96, 5585 (1974). 101. Huber, H., Kiindig, E. P., Ozin, G. A,, and Poe, A. J., J. Amer. Chem. Soc. 97,308 (1975). 102. Hudson, A,, Electron Spin Reson. 1, 253 (1973). 103. Hudson, A., Electron Spin Reson. 2, 270 (1974). 104. Hudson, A., personal communication (1973). 104a. Hudson, A., Lappert, 111. F., Lednor, P. W., MacQuitty, J. J., and Nicholson, B. K., unpublished work (1975). 105. Hudson, A., Lappert, M. F., Lednor, P. tk-., and Nicholson, B. K., J. Chem. SOC., Chem. Commun. 966 (1974). 105a. Hudson, A., Lappert, 11. F., and Nicholson, B. K., J. Organometal. Chem. 92, Cl1 (1975). 105b. Hudson, A., Lappert, M. F., h’icholson, B. K., and MacQuitty, J. J., unpublished work (1975). 105c. Hughey, J. L., Bock, C. R., and Meyer, T. J., J. Amer. Chem. SOC.97,4440 (1975). 106. Ibekwe, S. D., and Myatt, J., J. Organometal. Chem. 31, C65 (1971). 107. Ingold, K. U., in “Free Radicals” (J. K. Kochi, ed.), Vol. 1, Chapter 2. Wiley (Interscience), New York, 1973. 108. Ingold, K. U., and Roberts, B. P., “Free Radical Substitution Reactions.” Wiley (Interscience), New York, 1971. 109. Ishikawa, M., and Kumada, M., J. Chem. Soc., Chem. Commun. 612 (1970). 110. Jackson, R. A., Advan. Free Radical Chem, 3, 231 (1969); Chem. SOC.Spec. Publ. 24, 295 (1970). 111. Jackson, R. A., MTP Znt. Rev. Sci., Org. Chem., Ser. 1 10,205 (1973). 112. Jackson, R. A., MTP Int. Rev. Sci., Org. Chena., Ser. 2 10, 205 (1975). 113. Janzen, E. G., Accounts Chem. Res. 4, 31 (1971). Free Radicals 3 97 114. Jensen, F. R., and Knickel, B., J. Amer. Chem. SOC.93, 6339 (1971). 115. Joblin, K. N., Johnson, A. W., Lappert, M. F., and Nicholson, B. K., J. Chem. SOC.,Chem. Commun. 441 (1975). 115s. Joblin, K. N., Johnson, A. W., I,appert, M. F., Hollaway, M. R., and White, H. A., FEBS Lett. 53, 193 (1975); and references cited therein. 116. Jong, I. G. de, and Wiles, D. R., Inorg. Chem. 12, 2519 (1973). 116a. Kaptein, R., Leuwen, P. W. N. hl. van, and Huis, R., J. Chem. SOC.,Chem. Commun. 568 (1975). 117. Kawamura, T., and Kochi, J. K., J. Organometal. Chem. 30, C8 (1971). 118. Kean, E. S., Fisher, K., and West, R., J. Amer. Chem. SOC.94, 3246 (1972). 119. Kemmitt, R. D. W., and Burgess, J., Inorg. React. Mech. 2, 350 (1972). 120. Keller, H. J., and Waversik, H., 2. Naturforsch. B 20, 938 (1965). 121. Kenworthy, J. G., Myatt, J., and Symons, M. C. It., J. Chem. SOC.,A 1020 (1971). 122. Kerr, J. A,, in “Free Radicals” (J. K. Kochi, ed.), Vol. 1, Chapter 1. Wiley (Interscience), New York, 1973. 123. Kice, J. L., in “Free Radicals” (J. K. Kochi, ed.), Vol. 2, Chapter 24. Wiley (Interscience), New York, 1973. 124. Kiefer, H., and Traylor, T. G., Y’etrahedron Lett. 6163 (1966). 125. Kinsella, E., Smith, V. B., and Massey, A. G., J. Organometal. Chem. 34, 181 (1972). 126. Kochi, J. K., ed., “Free Radicals,” Vols. 1 and 2. Wiley (Interscience), New York, 1973. 127. Kochi, J. K., in “Free Radicals” (J. K. Kochi, ed.), Vol. 1, Chapter 11. Wiley (Interscience), New York, 1973. 128. Kochi, J. K., in “Free Radicals” (J. K. Kochi, ed.), Vol. 2, Chapter 23. Wiley (Interscience), New York, 1973. 129. Kochi, J. K., Accounts Chem. lips. 7, 351 (1974). 130. Kochi, J. K., and Krusic, P. J., .I. Amer. Chem. SOC.91, 3938 (1969). 131. Kochi, J. K., and Krusic, P. J., Chem. SOC.Spec. Publ. 24, 147 (1970). 131a. Koenig, T., and Fischer, H., in “Free Radicals” (J. K. Kochi, ed.), Vol. 1, Chapter 4. Wiley (Interscience), Sew York, 1973. 132. Kramer, A. V., and Osborn, J. A., J. Amer. Chem. SOC.96, 7832 (1974). 133. Kramer, A. V., Labinger, J. A,, Bradley, J. S., and Osborn, J. A,, J. Amer. Chem. SOC.96, 7145 (1974). 134. Krusic, P. J., and Kochi, J. K., J. Amer. Chem. SOC.93, 846 (1971); Krusic, P. J., Meakin, P., and Jesson, J. P., J. Phys. Chem. 75, 3438 (1971). 134a. Krusic, P. J., Stocklosa, H., Manzer, L. E., and Meakin, P., J. Amer. Chem. SOC. 97, 667 (1975). 135. Kuivila, H. G., Accounts Chem. Res. 1, 299 (1968). 135a. Kwan, C. I,., and Kochi, J. K., J. Organometal. Chein. 101, C9 (1975). 136. Labinger, J. A., Braus, R. J., Dolphin, D., and Osbarn, J. A., J. Chem. SOC.,Chem. Commun. 612 (1970). 137. Labinger, J. A., Kramer, A. V., and Osborn, J. A., J. Amer. Chem. SOC.95, 7908 (1973). 138. Lankamp, H., Nauta, W. T., and Maclean, C., Tetrahedron Lett. 249 (1968). 139. Lappert, M. F., and Lednor, 1’. W., J. Chem. SOC.,Chem. Commun. 948 (1973). 140. Lappert, M. F., and Sanger, A. It., J. Chem. Soc., A 847, 1314 (1971). 141. Lau, K. S. Y., Fries, R. W., and Stille, J. K., J. Amer. Chem. SOC.96,4983 (1974). 142. Lednor, P. W., D. Phil. Thesis, Sussex, 1974. 398 M. F. LAPPERT AND P. W. LEDNOR 143. Lloyd, R. V., and Rogers, 3f. T., J. Anier. Chcin. Soc. 95, 2459 (1973). 144. T,yons, A. R., and Symons, hl. C. R., J. Chenz. SOC.Farutlay Trans. II 68, 502 (1972). 144a. Lyons, J. E., J. Cheni. Soc., Chetn. Conatticin. 418 (1973). 145. McCleverty, J. A,, Orchard, 11. G., Connor, J. A., .Jones, E. hl., Idoyd, J. l’., and Rose, P. D., .I. Orgwno?rirtal. Cheni. 30, C75 (1971). 146. McDonnell, J. J., and l’ochopien, D. J., J. Org. C‘hern. 36, 2092 (1971); 37, 4064 (1972). 146a. hlarks, T. J., J. Organonze/a(. Chena. 79, 181 (1974). 146b. >larks, T. J., J. Organometal. Chctn. 95, 301 (1975). 147. Matlieson, T. IV., Peake, I3. M.,Itobinson, B. H., Simpson, J., and \Vatson, D. J., J. Cheni. SOC.,Chem. Coninrun. 894 (1973). 148. hlatsumoto, II., Sakano, T., and Nagai, Y., Y’etrahedron Let/. 5147 (1973). 149. Menapace, I,. \V.,aid Iiuivila, 11. G., J. Anaer. Chem. Soc. 86, 3047 (1964). 150. Mendenhall, G. I]., and Ingold, I<. U., J. .?mer. (Ihenr. Soc. 9.5, 3422 (1973). 151. Milovskaya, 13;. 13., 1Ziiss. Chewi. Reu. 42, 384 (1973). 152. Moelwyn-Hughes, J. T., Garner, A. I\. B,, and Gordon, S.,J. Organometal. Chem. 26, 373 (1971). 153. XIoivat, \V., Shortland, A,, Tagupsky, G., Hill, N. J., Yagupsky, IT.,and Xlkin- son, G., J. Chem. Soc., Dtil(on Trans. 533 (1972); >lowat, \V., Short.hnd, A. J., Hill, N. J., and \Vilkinson, G., J. Chenz. Soc., Dalton Trans. 770 (1973). 154. n’skttyima, I[., Hull. C‘hrni. Sac. Jnp. 43, 2057 (1970). 155. O’Seal, 11. I<:., aid Renson, S. IV., in “Free I~adicals”(%J. K. Kochi, ed.), Vol. 2, Chapter 17. \\-iley (Interscience), New York, 1973. 156. Otsuka, S., Nakamura, A,, Yoshida, T., Xaruto, M., and Atak:i, K., J. ilnaer. Chrm. Soc. 95, 3180 (1973). 157. Ovcii, H. 0. van, Groencnboom, C. tJ., and Tiefde Rleijer, IT. J. de, J. Organomelal. Chem. 81, 379 (1974). 158. l’aneth, F. A., and HoFedit,z, lV., Her. 62, 1335 (1929). 159. I’eake, 13. M.,Robinson, B. H., Simpson, J., and \\‘7atson, I). J., J. Ch.em. SOC., Chem. Co~tzmicn.945 (1 974). 160. I’earson, R. G., Louw, JV., and Itajaram, J., Inorg. Chim. Ada 9, 251 (1974). 161. l’earson, 11. (;., and Muir, \V. R., J. A4rrarr. Chena. Sor. 92, 5519 (1970). 162. Pearson, It. (+.,and Rajaram, J., Inorg. Chert/. 13, 246 (1974). 163. Poutsma, hl. l,.,in “Free Radicals” (J. I<. liochi, ed.), Vol. 2, Chapter 14. \Viley (Interscience), Sew York, 1973. 164. Price, S. J. IV., in “Comprehensive Chemicnl Kinetics,” (C. H. Ramford, and C. F. 13. Tipper, eds.), VOI. 4, p. 197. Elsevier, Amsterdam. 165. See, e.g., l’rins, It., and liortbeck, A. G. T. G., J. Organonzetal. Chem. 33, C33 (1971) ; Horsefield, A,, and \Vassermann, A,, J. Chem. SOC.,Dalfon Trans. 187 (1972); Cowan, I). O., Candela, G. A., and liaufman, F., J. Amer. Chem. SOC. 93, 3889 (1971). 166. Rettig, RI. I?., Stout, C. D., Klug, .4.,and Farnllam. P., J. Amer. Chem. SOC. 92, 5100 (1970). 167. Robinson, S. D., Organometal. Chrm. 3, 376 (1975). 168. ltoncin, J., Debuyst, R., J. Chem. Phys. 51, 577 (1969). 169. Ilowe, M. I)., Gale, It.: and hlcCafferg, A. J., Chena. Phys. Lett. 21, 360 (1973). 170. Russell, G. A,, in “Free Radicals” (J. K. liochi, ed.), Vol. 1, Chapter 7. FViley (Interscience), New York, 1973. Free Radicals 399 171. Sakurai, II., in "Free Itadicalh" (.J. li. Korhi, ed.), 1'01. 2, Clinptei. 26. \Viley (Interscience), Xew Pork, 1973. 17111. Sakurai, H., lIochitia, I<.,a~ttl liii.:L. ll.,.I. An~er.Ch 712. Sac. 97, 929 (1976). 172. Seyferth, D., and IIallgren, J. I<;., -1. Orgntmmetul. Chem. 49, C41 (1973). 173. Sicgcrt, F. IV., and Liefde 1Tvijcr, It. J. dc, J. Orgunonietal. Ch~m.23, 177 (1970). 174. Sment,owski, F. J., Chprn. ATUJ!.26, 481 (1971). 175. Sommer, 1,. €I., and IJlland, I,. h., .I. Or!/. (,'hein. 37, 3878 (1972). 176. Starliie, H. C., and Syinons, 11. C'. I{., J. Phetn. Soc., Dalton Trans. 731 (1974). 177. Strnart, C. I'., and Porte, ,4.I,., J. (yh(,tti. Soc., Iltrlton l'runs. 722 (1973). 177:~.Sticger, II., and Kelm, II., J. I'//!/,s. Chrw. 77, 290 (1973). 178. Symons, XI. C. R., arid Andr~iw)n,0. l'., Iuorg. Chew 12, 1932 (1973). 179. Tcuben, J. II., and Liefde lI(.ijcr, 11. ,I. dr, J. Organonietul. Chem. 17, 87 (1969). 180. Tliiclc, I<.-IT., and \Vagner, S., .I. Olgtrnontc.tal. Chem. 20, 1'25 (1969). 181. Thomas, J. I'., and Hayes, lt. C;., Inorg. ('/wn. 11, 348 (1972). 182. Ugo, It., Coord. Chetn. ZZri,. 3, 310 (1968). 183. Ugo, It., l'asini, .I.,Fusi, .i.,:m(I ('cwini, S., J. Amer. Cheni. Soc. 94, 7364 (1972). 184. \Vdling, C., and liuyser, 11:. S., f)r!y. It'c~rd. 13, 91 (1963). 185. IV:trd, (;. -4.,Rrusc, \V., 13owcr, I{. I<., and Chien, J. C. If-., J. Orgarroirreful. Chem. 4,2, CX3 (1972); Kruse, \V.,ihitl. 42, (39 (1972); Gramlich, V., and PfefTerkorn, K., ihirl. 61, 247 (1973). 186. IVard, II. It., ir, "Free 1tadic~:il~"(d. I<. liochi, ed.), Vd. 1, Ckaptcr 6. \\ley 188. \Vntt,s, G. U., and Ingold, I<. LJ., ,I. Amr. Cheni. Soc. 94, 491 (1972). 189. \Vest, It., and Uoudjouk, l'., J. Attti~r.Chc.m. Soc. 95, 3983 (1973). 190. JVliitesides, G. ll., Berghreiter, I). 15., and Iiendall, 1'. I<., J. .'tftier. Chem Soc. 96, 2806 (1974). 191. IVilt, J. IV., in "Free Itrtdicds" (J. I<. Iiochi, ed.), Vol. 1, Chapter 8. IViley (Int~erscience),Sew York, 1973. 191:~\\.itm:~n, hl. \V., and \\eher, J. ll., I~iorq.Nucl. ChenL. Lett. 11, 591 (1975). 192. \Vong, 1'. Ti., TAU,S. R. Y., :~11(1Sit,illc, J. I<., J. Amer. Ch~tn.Sor. 96, 5956 (1974). 193. \Void, J. If., and Brown, D. (i., Str'itct. noriding 11, 47 (1972). 194. \\;righton, >I., and Bredesen, I)., .I. Orgntiornetal. Chern. 50, C35 (1973). 194a. tt'righton, RI. S., and Ginley, 1). S., J. ,t vier. Chenz. SOC.97, 2005 (1975). 19lb. \Vrigliton, RI. S.. and Ginley, 1). S., J. Avirr. Chein. Soc. 97, 42.26 (1975). This Page Intentionally Left Blank Subject Index A Acetylene trimerization, ratalysis by alkyl- idynetricobaltnonacarbonyls, 137 Acetylene complexes, 5.5-57, 245~26.5 Acylation reactions, with tricobaltcar- catalytic reactions, 261-265 bon decacarbonyl cation, I11-1 19 activation of acetylene on coordina- Alkyl compounds tion, 261-262 of antimony (V), 232-236 cocyclization with isocyanides, 263- of arsenic (V), 229-231 265 indium halides, reaction with triorgano- cYrlo-oligorncrizat,ion, 262 stibine sulfide, 196 linear oligomerization, 262-263 of lithium, in reduction of CICCo3(CO)9, electron-defirient species, 258- 259 103 in homogeneous ratalysis, 2455265 of niobium, 237-238 via insertion reartions, 25-261 of phosphorus (V), 209-224 geometry of transition st,at,e, 235 of tantalum, 238-239 merlianism, 253-255 of tin, reaction with tricobaltcarbon S11lt studies, 251 decacarbonyl ration, 115 stereochemistry of product, 251-253 of zinc, in alkylation of tricobaltcarbon metalococyclization reactions, 260-261 decacarbonyl cation, 115 with carbon monoxide, 260 Alkylidene trialkylarsoranes, 224-228 with isocyanides, 260 Alkylidene trialkylphosphoranes, 209-214 metalocyclization reactions, 256-260 Alkylidynetricobalt nonacarbonyl com- formation of metalorycloheptatI.icnrs, plexes, 97-114 259-260 analogy with aretylenedicobalt hexa- formation of metalocyclopentatlienes, carbonyls, 138-140 2.56-257 as catalysts, 137-138 structure and bonding, 246-231 rarbon-funct ional derivatives, 97-1 10 MO scheme, 246 esters, 110-1 12 nature of interacting orbitals, 246-247 catalytic hydrogenation of unsaturated twisting of C=C bond, 56 derivatives, 97-1 10 variation of C=C bond longt.h, 56, decomposition reactions, 135-138 247 to acetylenes, 135-136 variation of C-C-C bond angle, 247 to acetylene complexes, 135-136 thermal stability and extent of hark- with metlioxide ion, 136-137 bonding, 248 oxidation by ceric ion, 135-136 variation in stretching frequencies of Hammett u-constant for CCo3(CO)g meta-acetylene unit, 248-251 group, 128 effect of other ligands, 250 infra-red spectra of ketone derivatives, Acetylenedicobalt hexacarbonyl, 138-140 123 analogy with alk~lidynetricohaltnon- mechanism of formation, 101 acarbonyls, 139-1 40 merlianisni of reactions, 140-141 40 1 402 Subject Index preparation of :tlcohol derivatives, 119 ylitles, 231 -232 using triethylsilane, 121-1 22 aryl deriv:Ltives, 231 reduction of ketone derivatives, 119-122 i\nt.irnony (V)nlkyls, 232-236 by triethylsilane and trifluoroacetic Arsenic ylides, 224-228 acid, 120 effect of silylation on stability, 228-229 reduction hy orgtanolithium reagents, SXIlt spr’(’tRt, 226-227 103 photorlertron spertr:i, 228 stahle carbonium ions from, 119-134 reartions, 225 sterir hindrance in, 99, 1 I1 structure, 226 structure, 99 synthesis, 224-225 synthesis, 100--110 t1ierm:il stability, 226 from acetylenedico1):tlt 1iex:icar- Arsrnir (V) pentamet.liyl, 229-231 t)onyls, 100 I\zobene via addition to olefinir bonds, 107-108 a-tiackhonding ability, 57 using arylmerrury compounds, 103- 110 r:tlcul:ttions on, 57 106 Ni(0) cwniplexrs, 57 from robalt carboiiyl and tdocar- .~xobisisohut~~.ollitrile, as r:i.dical init,intor, bons, 101-102 317, 380 via Friedel-Crafts re:ict,ioiis, 106-107 B using GIYIUIJI11 halides, 109-110 via ol-haloalkylmercaI.ials, 103-106 I3cnt rc,hyhritfix:Ltion t.litory, application via radical reactions, 107--108 to nirtliylt,in ldides, 71 via substitution at, apic.al carbon liimolec~ul~rhomolytic substitution reac- atom, 103 tions involving free radicals, 370-381 tricohalt,rarbon decwarhoriyl c:ition with mttin group compound substrates, from, 110-119 371-373 Allene, oligomerization re:ictions, 270 with transition nietnl complcs suh- Allene complexes, eyclo-oligomrrizatioii strntes, 373-381 react,ions, 270-278 I3oron halides, reactions catalysis by Si(0) romplcxrs, 271- with c:trhene complexes, 21-27 277 with CICCO~(CO)~,118-1 19 catalysis by rhodium romplexcs, 277- with cobalt carbonyl, 109 278 Boron Iiydridcs, 145-150 kinetic studies, 273 st>ructurnl relationship with high nu- role of phospliorus lig:mtir, 275-277 clearit,y metal carboiiyl clusters, dissociation of, 269-270 337-339 effect of coordination on allene bond lengths, 267 c in homogeneous catalysis, 265-278 1H NhlR studies, 267 Carhene complexes, 2-20, 24-28 relative reactivities of Ni, Pd, and Pt acidity of a-rarhoii atom of alkoxyalkyl- complexes, 269-270 c:irbenes, 13 X-ray structural data, 265-267 addition-iearrangement reactions, 13 Alumiriu~iihalides, reactions with hydrogen halides, 13 with carheiic complexes, 27 honding, 4-6 with CICCO~(CO)~,116-1 19 JR spectral studies, 5-6 with cobalt carbonyl, 110 liberation of rarbenc ligand, 14-21 Antimony by arid, 14-1 6 organonietalhc chemistry, 187-204 by 0, S, or Se, 17 Subject Index 403 by pyridinr, 16 (':i talysi s t Iicrni:iIly, 16 by :illi~litl~tietricobaltnnnacathnyls, trapping of lig:iiid, 1.5 137-138 I3C SNIt studies, A, 133 awtylenc trimeiht~ioti,I37 olrfin I)(,lynieriz:ttioii, 137-138 positive ch:trgc on a-c~:irl)i)iiatoiir, I4 rr:ict,ions witii ncids, 14-16 :itldition :it c:irbenr c:irIion, 9 with :mines, 11-12 hy Si(0)complexc.s, 271-277 c:iri)enr displarcmrtit, 10 by rliodiuni complexes, 277-278 carbony1 suhstitut ion, 9-1 0 of oligomcrimtion (if acetglenw, 262- with rlcctropliilic. r:Lrbenes, 20-21 263 with phosphines, 9-10 by t,liodac,:irhor:inc complex, 183 with pyridinc, 1 (i ('licniically induccd dynamic nuclear suhstitution :tt c*arIwnr c:d)ott, 1 1-12 po1ariz:ition (CIDNI') sprrtra, and with vinyl etlirrs, 17-18 orgatioriirt:lllit. r:tdicnls, 347, 392 nith ~~~-vi~i~1-2-~~gr~olidoncs,18-1 9 ('I)l)dt i.c,lat ionsliip \vit,li ylidc c~omplrws,240 :ilkylidyne tioii:ic~:t~hoti~lclustws, 97- suhstitution of hydrogen :Lt. a-v:ir.l)on 144 :ttotli, 13-14 liigli nuck:tritg r:trbonyl clustcm, 287- tlcuteriuin excliangc~,13 28X, 325-327 syntliclsis, 3-4, 6-8 r:trl)i(lo-derivatives, 327 from nlcohols :ind isiicy:uii(Ic coni- IIL sprct r:i, 326-327 plexrs, 7 react ions, 325-327 vi:i carhetic ttxiisfrr, 7 struvtur:Ll d:Lt:f, 288 via clravage of clcct,ron-ricli (~l(ifiirs,8 syntlresis, 325-327 from I , 1 -dic~hlo1~i~-2,3-tli~ilic~tiyl-2-cy-(:yclo-oligonictrix:ttion clopropcne, (i of acetyl(~nes,262 vi:L organolithiuni rengrnt.s, 3-4 ' :illcti(~S;,270-278 X-ray struct,urnl studies, 4-5, 7 Carhoniuin ions honding, 36 tiori:~c*arl)otiyltrici~i~:~lt cnt~l)on siihst it u- froin c*arl)mc complexes and et,liylvinyl td, 119-134 ether, 17 st:hility of ferrorctiylmctli~1,1 i31 Carhorancs D tIegrsd;ition reactions, 147 met:il complexes, 145-1 86 De\\.:ir-Clintt-l)unr:inson model piilylietlral ~'earrsngrineritf,146, 149 for :illcne-metnl bonding, 287 Cnrbyne complexes for olefin-nict:il bonding, 35 cle:tv:tgc. of c*arhyncligand, 28-29 Diazrrir-t!.atisition niet,:tl complcxcs, S311t, 22 structurrs, 57 c~hcmiid s1iift.s of c:it,Iiyne c:it,t)on Dihalogenorarheties, reaction \vith (xr- atom, 133 i~enerarbonylcornplcxes, 20-21 re:wtions, 28-29 13IUL' thy1ti ti dihalides, 7 I -72, 84-90 ?yntliesis, 21 -28 electron diffraction studies, 71 via aluminurn Iialides, 27 molecular cornplcxes, 84-90 via hornn halides, 21, 24-26 with triorganostihine sulfides, 195- via. gallium haliclcs, 27 196 X-ray strurtuml studirs, 22-23 \liksIxiuer stutlic:s, 72 404 Subject Index NQR studies, 72 osmium, 325 X-ray diffraction studies, 71-72 platinum, 334-336 Dipole moments, of methyltin halides, 68, reactions, 317-323 76 effect of core enclosure by carbonyls, 317-318 E ligand substitution, 322 Electron diffraction studies, 67, 69, 71, 72 oxidation, 320-321 Electron spectroscopy for chemical analy- oxidative addition, 322-323 sis (ESCA) reduction, 319-320 measurement of electron density in rhodium, 327-332 platinum complexes, 44 ruthenium, 324 Electron spin resonance (ESR), of organo- separation, 316-317 metallic radicals, 346-391 solid state structures, 286-306 carbido-carbon radius, 302 F of carbido-species, 300-302 effect of charge on bond lengths, 295 Free radicals, in organometallic chemistry, of heptanuclear species, 299 345-392 metal coordination numbers, 293-295 Friedel-Crafts reaction occurrence of triangular metal arrays, acylation 291-292 of a cobaltacarborane, 178 octahedron-trigonal prism transforma- of H,C=CHCCo3(CO)g, 125 tion, 298-301 in synthesis of alkylidynetricobalt nona- tendency toward close packing of carbonyls, 106-107 metals, 305-306 G solution state structures, 306-311 13C NMR studies, 308-310 Gallium halides, reaction with carbene IR studies, 306-308 complexes, 27 synthesis, 31 1-316 Galvinoxyl, as radical inhibitor, 347, 380 bond energy considerations, 31 1-313 Grignard reagents in synthesis, of carbido-complexes, 314-315 of alkylidynetricobalt nonacarbonyls, by condensation reactions, 311-313 103-104 dependence on CO pressure, 312-313 of tertiary stibines, 197 by pyrolysis, 313-315 Group V elements, penta-alkyls and by redox condensation, 313-314 alkylidenetrialkyls, 205-243 Homogeneous catalysis involving acetylene complexes, 245-265 H allene complexes, 265-278 Hammett a-constant, for CCO~(CO)~Huckel calculations, on phosphorus ylides, substituent, 128 212 High nuclearity metal carbonyl clusters, Hydroformylation, catalyzed by rhoda- 285-344 carborane, 183 bonding, 336-341 Hydrosilylation, catalyzed by rhoda- analogy with polyboranes, 337-339 carborane, 183 application of noble gas rule, 336 LCAO-MO theories, 33S341 I topological theories, 337 Jmines, transition metal complexes, 58 cobalt, 325-327 Infra-red studies iridium, 332-333 on acetylene complexes, 248-251 iron, 323-324 C=C bond length versus YC-C in olefin nickel, 333-334 complexes, 39, 44 Subject Index 405 on high nuclearity metal cwhnyl Metallocarboranes, 145-186 clusters, 306-308 with eleven vertices, 171-175 on methyltin halides, 64-65, 68-75 bimetallic species, 173-175 molecular complexes, 79-91 monometallic species, 171-173 Sn-C bond frequencies, 64-65 polyhedral rearrangement, 175 structural studies, 64-65 reactions, 173-175 variation of YCN in isocyanide complexes, structures, 171, 174 45-46 synthesis, 171- 175 Insertion reactions with fourteen vertices, 171 of acetylenes, 251-261 geometry and number of polyhedral mechanisms of reactions, 253-255 vertices, 148-149 into metal-acetylene bonds, 251 in homogeneous catalysis, 182-183 into metal-o-carbon bonds, 251 alkene isomerization, 183 into metal-chlorine bonds, 251 deuterium exchange, 183 into metal-hydrogen bonds, 251 with nine vertices, 178-180 metalococyclization, 260-261 bimetallic species, 180 metalocyclization, 256-260 NMIt spectra, 179 stereochemistry of products, 25 1-253 structures, 178-179 or iridium, into B-H bonds, 181-182 synthesis, 178-180 of ylides, into silacyclobutanes, 215 oxidative-addition to R-H bonds, 180- Iridium, high nuclearity carbonyl clusters, 182 287 in deuteration, 181-182 structural data, 288 by iridium complexes, 181-182 synthesis, 332-333 relative reactivity of BH groups, 182 Iron stabilization of high oxidation states, carbene complexes, 4, 7 156 high nuclearity carbonyl clusters, 287 synthesis, 150-155 reactions, 324 from nido-carborane anions, 150-151 structural studies, 288, 290, 293 via polyhedral contraction, 152-153 synthesis, 323-324 via polyhedral expansion, 151-152 Isocyanide complexes via polyhedral subrogation, 153 carbene formation with alcohols, 7 by thermal metal-transfer, 153-154 variation of VCN with metal oxidation with ten vertices, 175-178 state, 44-45, 248-251 bimetallic species, 177-178 Isomerization, of carbenecarbonyl eom- mixed sandwich complexes, 175 plexes, 9-11 polyhedral rearrangements, 175-176 reactions, 178 K synthesis, 175-1 78 Kinetic studies trimetallic species, 178 oligomerization of allenes by nickel with thirteen vertices, 167-171 complexes, 273 polyhedral contraction, 169 isomerization of carbenecarbonyl com- polyhedral expansion, 169 pIexes, 10-1 1 polyhedral subrogation, 168-169 Ketone-metal complexes, 57-58 rearrangement reactions, 167-168 structure of nickel compound, 57 synthesis, 167-171 with twelve vertices, 155-167 M bimetallic species, 166-167 mixed ligand complexes, 163-166 Mercurials, in synthesis of alkylidynetri- monometallic species, 155-163 cobalt nonacarbonyls, 103, 105 polyhedral rearrangement, 158-159 406 Subject Index re:ictions, 169-1 61 vi:L protomttioii of vinyl derivatives, structures, 1.57 125 synthwis. 155-159. 161-16'7 :is n.c:tk c~lcctt~opliilt:~,122-123 RZethyltiu li:~litles,63-96 Suclrar nugnetic resonmice studies IIi~l~cul:~~COIII~I<~~~S, 76-92 arsenic ylitles, 226 s;pectr:tl studies, G4 -67 carbenc complexes, 11 structur:tl studies, 68-76 carhynr complexes, 25 i\Zethyltiti trildidcs, 72-76 lrigtr nuclc:irity niet:tl c:trbonyl clusters, electron tliffr:tctioii studies, 72 308-3 10 molecular cotnplrses, 90-91 clicmicnl shift and cliargc on CO, 309 SJIIt d:ttn, 74--75 fiusion:il bcliavior :rnd reactivity, NQR &t:t, 74 309-310 structures, 72-76 iroti:rc.arbotiylt~ricobaltcarbotic:trbonium Molecul:ir orbital calculations ions, 130-134 on :Lzobeirzen(:, 57 pmi ta-:tlk~lpliosplior;tncs,21 6 nil boron hydridc dwivativcs, 147 pliospliorus ylidrs, 21 2-21 3 for coorc1iti:rted :tcctylenes, 246-247 111 Suclwr niagnetic resonance studies, on high nuclc:trit,y nict:tl carhongl :trsciiic ylitlcs, 226 cluStC'rs, 339-341 nirtliyltiti li:ili(les, 65-66, 69. 71-72, on phosphorus ylidcs, 2 12 75-76 MBssbauer studies dcpendmcc of J(ll9Sn-C--'II) on of nicthyltin Iinlidrs, 66-67, Ci9, 71-72, solveti t, 85-66 74 inctliyltiii Iialidt: complexes, 77, 79-81, molecular complcsrs, 85-87, 89, 91 83-4, 86-01 tloii;tcttr~,oiiyltricolxtltcalbon carbon- N iurn ions, 130-134 Nickel olrfitr c~omplcxes,41 azobeiizenc complex, 57 orgniio:ttiti~iio~iycompounds, 189-191 carbetie complexes, 4 ~~cntn-:tlkyIpIiospho~~ncs,2 1 6 high tiucle:trity carhorij-l clusters, 287 pliospliorus ylidcs, 212 reactions, 333-334 311' Suclc:tr magnet.ic resonance studies structural data, 288, 290, 287-298 pen t:i-al ky1 phosplroran es, 2 1 6 synthesis, 333-334 pliosph(~~~sylides, 2 12 Xiobium, pcnt:dkyls, 207, 237--238 Sucl~ir quadrupole resonance (KQR) Nitrogen ylides, 207-209 studies, methyltiti halides, 66, 72 renctions, 207 molecular co~~iplcxt.~,79, 81, 83, 87, synthesis, 207 89-9 1 Kohle gas rule, and high tiuclcarity metal c:arhonyI (.lusters, 336-338 0 r\Tonnc:trbongltric~~~~altc,zrbori-su~~stitute~~ rarbonium ions, 119-134 Olefin lrytlrogeri:ition, catalyzed by rhoda- cotisequences of charge distribution, carborane, 183 123-128 Olefin isomcrizat,ion, catalyzed by rhodu- fluxitmtl behavior, 134 carhorane, 183 u-T Iiyperconjug:ttioti, 133 Olefin polymerization, cat:tlyzcd by alkyli- KlIR Spcctla, 129-134 tlytic~tric~ob:~ltn~~~i:teart~nnyls,137-138 reactions st carbon, 122-124 Olefin tr:tirsition-metal complexes structure, 129, 134 bonding, 34-37 synthesis, geometry, 53-55 from alcoliols arid Hl'F,, 12'2 five coordinate metal, 55 Subject Index 407 four c<)i)rdin:ttcnirtnl, ,-A iti rc.tluctive-c,lii~iiii:itioii i~eactioiir;,388 ttircc eoi)i~din:ttc~nic~tal, 33-51 tl.:tii.;it,ioii-inc.t:il compountfs, 363-366 noiiplan:irity of hound olcfin, 48 >I cnrhoiryl speries, 3GG st ructui~:ilstudirs, 37 -55 1~:511spcrtr:r, 363 twist of olcfin, 51-53 hy photolysis of dinicrs, 366 oririit,ation of r-sulistitucwts, .52 0rg:tiiotin conipounds, 63-96 pointing of trans-lignnd, 52 -53 re:tction ivitli triorganostihinc sulfides, variation of C=C hoiitl lengt I), 3s-46 195-197 correht ion trith clieniic:tl hliift of Osmium, high iiiic~lr:trity metal carhonyl olcfitiir protons, 44 c'lusters, 287 coridution \Y i t 11 iiir t a1 ioni z:i. t ii I i po- IIt stuclics, 325 tetiti:d, 14 strurtur:1l dnt:t, 288, 292 c.ortx.l:ttioii \rit.ti Y~.-(-, 39, 44 synthrsis, 325 v:i,riation of 1I-C: hond Icngtli, 46-38 0sicl:tt ivr :iddition ti-itli olrfin substituc~nt,4-47 of :ilkyI 1i:ilitiw to (;roup VIII metal ivi t I1 t 1~:tris-lig:tiid-inct:~1boir c I I(xngt11, POIIII)~C~X(~S,381-388 47-48 niwli:iiii.sni, 382-388 Organoant inioiiy conipounds to 1%--11 h~licl~,180-182 1iesaroolditi:ttc bpwirs, 188-1!)1 to 1%-11-B l)ri(lges, 183 ris-t,i~ansisonicrisni, 188-189 0xyniorcur;ttion, of alkylidynet ricotnlt- TI< studic~,189-191 Iloll:lcurt)onyls, 125 S1IR studies, 189-191 syntlirsis, 188-189 P X-ray stiidirs, 189-190 pcntavalcnt sprcics, 188 I'cnt:i-:~lk~lplio~~~lior:tn~~s,214-224 tertiary stihines, 197-202 cyclic spevics, 213 triorgnnostihiiie sulfides, 192-107 via insertii)n of ylides into silacyclo- Orgnno~nc~t:tllirratlicds, 345-392 but:tncs, 215 :wt,inide species, 391 :ts intermedi:rtes, 217-224 gener:~tion, in honiolytic srit)i.titutiorr in ylide clc:tv:tge of 1 ,3-tlisilacyclo- react ions, 370-381 1,1Itnlit~s,218-219 of main group species, 371-373 in ylidc c-lr:tvagc of monosilacyclo- of plat i tiuni dwivat ives, 373 --3X1 t)ut:tti<;s, 219-224 of Group IV elenicnt,s, 352-3G3 polycyclic spec,ics, 21 5-217 17S1t spectra, 350-362 spc:ctr:tl studirs, 216 mechanism of fortnation, 300-3G1 syntlirsis, 216 nt:iblc spccies, 355-3633 tlirrnial st:ihility, 21 6-217 st ruct wr, 354-355 I'etit:t:ilkylstit~tir:~iies, 232-236 transient, species, 352-355 "11 sprctr:r, 234 via UV irradiation, 3.52-351, 358 reactions, 234-236 lant1i:tnide species, 391 with 13ronstcd acids, 235 in met:tl-alkyl phot,ol with Lewis :tcid,q, 235 in met:tI-alkyl tlicrmolysis, 388-389 wit,li oxidizing :gents, 233-236 nietal-centered species, 346, 3.10-36(i synthesis, 232-233 not ccnt.crrd on nict:tl, 367-370 of :ilkeiiyl species, 233 14:SIt spectra, 367-369 of met,liyl cwnpound, 232 metnllocenes, 368 via orgatior~ic~tallics,232-233 nitroside derivatives, 368-369 of trinictli~lsilylmethyls,233 in oxidative-additioti reactiotrs, 383-388 tlicrnial tleeomposition, 233-231 RSlt evitlence, 384 vi1)r:ttionnl spectra, 231 408 Subject Index Pentaarylphosphoranes, 214 Reductive elimination reactions, involving Pentamethylarsorane, 229-231 radicals, 388 NMR spectra, 230 Rhodium, high nuclearity carbonyl reactions, 230 clusters, 287 structure, 230 anionic species, 329-332 synthesis, 229-230 reactions, 328-332 thermal decomposition, 231 structural studies, 288, 295-296, 298, Peptide synthesis, via carbene complexes, 302-30.5 11-12 synthesis, 328-332 Fhosphorus(V) alkyls, 214-224 Ruthenium, high nuclearity carbonyl Phosphorus ylides, clusters, 287 bonding, 212-214 structural studies, 288, 293 d-orbital participation, 212 synthesis, 324 dipole moments, 212 effect of silylation on stability, 228-229 5 NMR data, 212 photoelectron spectra, 214 Spin traps, 317 properties, 210-211 use in 8~2reactions at Pt(II), 376-378 reactions, 210 Stability constants, of methyltin halide with silacyclobutanes, 218-220 complexes, 8 1-82 structures, 21 1 synthesis, 209-210 T vibrational spectra, 21 1 TantaIum Phosphorylide complexes, from carbene pentaalkyls, 238-239 complexes and phosphines, 9-1 1 ylide complexes, 207, 238-239 irradiation, 9-10 mechanism of formation, 239 Photoelectron spectra X-ray structural studies, 239 of arsenic ylides, 228-229 Tertiary stibines, 197-202 of methyltin halides, 67 asymmetric species, 198-200 of phosphorus ylides, 214 rate of pyramidal inversion, 199-200 Platinum cleavage of Sb-phenyl bonds, 197-198 carbene complexes, 7 quaternization, 200 high nuclearity carbonyl clusters, 287 resolution of asymmetric salts, 200 reactions, 335-336 reactions, structures, 288, 296-298 of ally1 compounds with synthesis, 334-338 CpFe(CO)&l, 202 radicals, from homolytic substitution with metal carbonyls, 200-202 reactions, 373-383 synthesis, 197-198 Tin, methylhalides and their molecular R complexes, 63-96 Topology, and high nuclearity metal car- Radical anion, of ClCCo3(C0)9, 103 bony1 clusters, 337 Radical reactions, rate determination by Transition metal complexes of unsaturated ESR studies, 347 molecules, 33-61 Raman studies Tricobaltcarbon decacarbonyl cation, on high nuclearity metal carbonyl 110-119 clusters, 310-311 electrophilic nature, 115 on methyltin halides, 64-65 mechanism of formation, 116-119 Reaction mechanisms, in organocobalt reactions cluster chemistry, 140 acylation, 11 1-119 Subject Index 409 with alkyltin compounds, 115-1 16 alkyltin halides, 67-69,71-72, 78, 83-86, with triethylsila.ne, 116 89 synthesis allene complexes, 265-267, 271 via aluminum chloride, 115 criteria for accuracy, 37-38 from (CO)9C03CC02R,I1 1-1 12 diazene complexes, 57 of hexafluorophosphate salt, 112 high nuclearity metal rarbonyl clusters, Trimethyltin halides, 68-71 286-306, 327 fluoride, 68-69 iniine complexes, 58 structural studies, 68-69 ketone complexes, 57-58 molecular complexes, 77-83 metallocarboranes, 148, 155-156, 168, spectral studies, 68-71 174, 178 Triorganostibine sulfides, 192-197 olefin complexes, 37-55 nature of Sb-S bond, 193-194 phosphorus ylides, 211 *-bonding, 194 six-coordinate methylantimony com- spectral studies, 193-194 pound, 189-190 reactions, 195-1 97 tantalum-ylide complex, 239 U Y UV spectra, of triorganostibine sulfides, Ylides 193-194 ammonium, 208 V antimony, 231-232 arsenic, 224-229 Vanadium, ylides, 236-237 donor properties, 206 Vitamin 1312 coenzyme nitrogcn, 207-209 ESR studies of active site, 369, 389-390, phosphorus, 209-214 392 tantalum, 238-239 homolysis of Co-alkyl bond, 389-390 transition-metal, relationship with carbene complexes, 240 X vanadium, 236-237 X-Ray crystallographic studies acetylene complexes, 55-57, 247 Z alkylidynetricobalt nonacarbonyls, 99, 110 Zeise’s salt, 37 Curnulafive Lisf of Confribufors Ahel, E. W., 5, 1;8, 117 Aguilo, A,, 5, 321 Alhnno, V. G., 14, 285 Arniitagc, 1). ;\., 5, 1 Ittel, s...\., 1.2, 33 Atwell, \v. I-I., 4, 1 Jolly, 1’. \\’., 8, 29 Bennett., hl. A,, 4, 353 I3irini1igh:i~ii,rJ., 2, 365 Brook, A. G,, 7, 95 Bro\vii, 11. C., 11, 1 13ro\vn, rr. I,,, 3, 365 Kiliirr, 11.. 10, 11.5 I3ruc.r, \l. I., 6, 273; 10, 273; 11, 447; 12, Kitig, lt. B., 2, 157 379 Kingston, l<. AT., I I, 253 , 11., 8, 211 I Chllniaii, J. l’., 7, 53 I~ongoni,(;.? 14, 285 Corry, J. Y., 13, 139 l>uijtrii,J. (;. A., 3, 397 (>outts,It. S. l’., 9, 13.5 Lupin, 1r. S., 8, 21 1 Coylc, T. I)., 10, 237 SIcliiIl~ip,A,, I I, 117 Craig, 1’. .J., 1 I, 331 1l:ttltIox, Sl. I,., 3, 1 Cullell, \v, It., 4, 115 1I:LitIis; I). ir,, 4, 05 Cuntly, C. S., 1 I, 253 1l:iiin, 13. I.:., 12, 135 de I3oer, I<:., 2, 115 ~l:ll~uel,T. ;I.,3, 181 Dessy, It. I<;., 4, 267 1lasoll, It., 5, 93 I>ickson, I{. S., 12, 323 Kmerson, (;. F.,I, 1 ISrrist, <’. I{., 10, 79 ~l(lrg:lll,G. L., 9, 195 Fischer, I<. O., 14, 1 1li,owc:i, .J. .I., 7, 157 E’rnscsr, 1’. J., 12, 323 K‘;1gy, 1’. I,. I., 2, 325 Fritz, If. P., I, 239 S:ik:rmura, A., 14, 245 Furukawa, J., 12, 83 Kc~srilcyallov,i\. N., 10, 1 Fuson, It. c.,1, 221 Scurrrsml, \\.. l’., 7, 241 Gilman, H., I, 89; 4,, 1 ; 7, 1 Ok:i\var:i, It., 5, 137; 14,, 187 Grecn, 31. I,. lI., 2, 325 Oliver, J. l’., 8. 167 Griftith, \Y. P., 7, 21 I Oiittk, T., 3, 263 Gubin, S. l’., 10, 347 Otsuk;t, S., 14, 24.5 (;ysling, H., 9, 361 I’~irsli:tll, Ci. \V.,7, 157 Ilarrod, J. F., 6, 119 l’aul, I., 10, lY9 IIawttiorne, hl. F., 14, 145 Iieck, lt. F., 4, 243 l’ettit, It., 1, 1 Heirnbach, l’., 8, 29 l’ol:intl, J. S., 9, 397 410 Cumulative List of Contributors 41 1 l’ratt, J. XI., 11, 331 Tain:to, Ti., 6, 19 Prokai, 13., 5, 225 Taylor, 1.:. C., 11, 147 Iteutov, 0. A,, 14, 63 Tliayer, J. S., 5, 169; 13, 1 Rijkens, F., 3, 397 Todd, 1,. %J., 8, 87 Ititter, J. J., 10, 237 Treicliel, P. ll.,I, 143; 11, 21 Itocho\v, 15. G., 9, 1 Tsutsui, LI., 9, 361 Roper, \V. R., 7, 53 Tyfield, S. l’., 8, 117 Roundhill, D. Jl., 13, 273 van dcr Kerk, 0.J. M.,3, 397 Itubezliov, A. Z.,10, 347 \\.'ads, Jl., 5, 137 Schmidbaur, H., 9, 259; 14, 205 \\-altoii, D. R. If., 13, 453 Sclirauzer, c:. N., 2, 1 \\.‘ailex, I>. C., 9, 135 Scliwebke, G. I,., 1, 89 \Vest, It., 5, 169 Seyfertli, D., 14, 97 \Vilcs, D. It., 11, 207 Silvertliorn, \\-. E., 13, 47 \\.‘ilk<,, G., 8, 29 Skinner, I€.A, 2, 49 \Vojcic>ki,A,, 11, 87; 12, 31 Slocwm, D. \\.’., 10, 79 Yitshina, S. S., 14,, 63 Smith, J. I)., 13, 4.53 %ieglrr, li., 6. 1 Stafford, s. I,., 3, 1 Zuckermsn, J. J., 9, 21 StoIiC, 1’. U. A,, I, 143 Cumulative List of Titles Acetylene and Allene Complexes: Their Implication in Homogeneous Catalysis, 14,, 245 Alkali Metal Derivatives of Metal Carbonyls, 2, 157 Alkyl and Aryl Derivatives of Transition Metals, 7, 157 Alkylcobalt and Arylcobalt Tetracarbonyls, 4, 243 Ally1 Metal Complexes, 2, 235 r-Allylnickel Intermediates in Organic Synthesis, 8, 29 Applications of llQmSnhlossbauer Spectroscopy to the Study of Organotin Compounds, 9, 21 Arene Transition Metal Chemistry, 13, 47 Boranes in Organic Chemistry, 11, 1 Carbene and Carbyne Complexes, On the \Vay to, 14, 1 Carboranes and Organoboranes, 3, 263 Catalysis by Cobalt Carbonyls, 6, 119 Catenated Organic Compounds of the Group IV Elements, 4, 1 Chemistry of Carbon-Functional Alkylidynetricobalt Nonacarbonyl Cluster Complexes, 14, 97 13C NMR Chemical Shifts and Coupling Constants of Organometallic Compounds, 12, 135 Compounds Derived from Alkynes and Carbonyl Complexes of Cobalt, 12,323 Conjugate Addition of Grignard Reagents to Aromatic Systems, 1,221 Coordination of Unsaturated Molecules to Transition Metals, 14,33 Cyclobutadiene Metal Complexes, 4, 95 Cyclopentadienyl Metal Compounds, 2, 365 Diene-Iron Carbonyl Complexes, 1, 1 Electronic Effects in Metalloccnes and Certain Related Systems, 10, 79 Electronic Structure of Alkali Metal Adducts of Aromatic Hydrocarbons, 2, 115 Fast Exchange Reactions of Group I, 11, and 111 Organometallic Compounds, 8, 167 Fluorocarbon Derivatives of Metals, I, 143 Free Radicals in Organometallic Chemistry, 14, 345 Heterocyclic Organoboranes, 2, 257 a-Heterodiazoalkanes and the Rcactions of Diazoalkanes with Derivatives of Metals and Metalloids, 9, 397 High Nuclearity Metal Carhonyl Clusters, 14, 285 Infrared Intensities of Metal Carbonyl Stretching Vibrations, 10, 199 Infrared and Raman Studies of *-Complexes, 1, 239 Insertion Reactions of Compounds of Metals and Metalloids, 5, 225 Insertion Reactions of Transition Metal-Carbon u-13onded Compounds I. Carbon Monoxide Insertion, 11, 87 Insertion Reactions of Transition Metal-Carbon u-Bonded Compounds 11. Sulfur Diox- ide and Other hlolecules, 12, 31 Isoelectronic Species in the Organophosphorus, Organosilicon, and Organosluminum Series, 9, 259 Keto Derivatives of Group IV Organometalloids, 7, 95 Lewis Base-Metal Carbonyl Complexes, 3, 181 Ligand Substitution in Transition hletal r-Complexes, 10, 347 41 2 Cumulative List of Titles 41 3 Literature of Organo-Transition Metal Chemistry 1950-1970, 10, 273 1,iteratui-e of Organo-Transihn ,2let,al Chemistry 1971, 11, 447 Literature of Organo-Transition hlct.al Chemistry, 1972, 12, 379 Mass Spectra of Xletallocenes and ltelated Compounds, 8, 21 1 Alass Spectra of Organometallic Compounds, 6, 273 Metal Carbonyl Cations, 8, 11 7 Metal Carbonyls, Forty Years of Research, 8, 1 Metal *-Complexes formed by Scvcn- and Eight-Membered Carbocyclic Compounds, 4,353 Metallocarboranes, Ten Years of, 14, 145 Methyltin Halides and Their Mo1ccwl:tr Complexes, 14, 63 Nitrogen Groups in Illetal Carboriyl :ind Itelated Complexes, 10, 115 Nitrosyls, 7, 21 1 Pl‘uclear Magnetic Resonance Spectra of Orgarlometallie Compounds, 3, 1 Of Time and Carbon-Metal Uonds, 9, 1 Olefin Oxidation with I’alladium Chtalyst, 5, 321 Organic and Ilydride Chemistry of Transition Metals, 12, 1 Organic Chemistry of Copper, 12, 215 Organic Clieniistry of Lead, 7, 24 I Organic Complexes of Lower-valcrlt, Titanium, 9, 135 Organic Substituted Cyclosilanes, 1, 89 Organoantimony Chemistry, Iteccnt Advances in, 14, 187 Organoarsenic Chemistry, 4, 145 Organoberyllium Compounds, 9, 195 Organolanthanides and Organoactinides, 9, 361 Organometallic Aspects of Diborori Chemistry, 10, 237 Organometallic Benzheterocyclc:s, 13, 139 Organometallic Chemistry: A Historicd Perspective, 13, 1 Organometallic Chemistry: A Fort,y Years’ Stroll, 6, 1 Organometallic Chemistry, hly \Yay, 10, 1 Organometallic Chemistry of Nickcl, 2, 1 Organometallic Chemistry of the Ilain Group Elements-A Guide to the Literature, 13, 453 Organometallic Chemistry, Sonic I’ersoilal Notes, 7, 1 Organometallic Complexes wit11 Silic:on-Tr:trisition Metal or Silicon-Carbon-Transition Metal Bonds, 11, 253 Organometallic Nitrogen Compounds of Germanium, Tin, and Lead, 3,397 Organometallic l’seudohalides, 5, 169 Organometallic Reaction hlech:inisms, 4, 267 Organometallic Reactions Involving IIydro-Nickel, -Palladium, and -Platinum Com- plexes, 13, 273 Organopolysilanes, 6, 19 Organosulphur Compounds of Silimn, Germanium, Tin, and Lead, 5, 1 Organothallium Chemistry, Rcccmt Advances, 11, 147 Organotin IIydrides, Reactions \vit,ll Organic Compounds, 1, 47 Organozinc Compounds in Synt Iirsis, 12, 83 Oxidative-.4ddition Iteactions of da Complexes, 7, 53 I'alladiurn-Catalyzed Organic Reactions, 13, 363 I’entaalkyls and Alkylidcne Trialkyls of the Group V l’kments, 14, 205 I’reparation and Reactions of Olg:lnorobalt(III) Complexes, 11, 331 414 Cumulative List of Titles Itedistribut,ion Equilibria of Organometallic Compounds, 6, 171 Radiochemistry of Organometallic Compounds, 11, 207 Strengths of Metal-to-Carbon Bonds, 2, 49 Structural Aspects of Organotin Chemistry, 5, 137 Structural Chemistry of Orga.iio-Transition hktd Complexes, 5, 93 Structures of Organolithiuni Compounds, 3. 365 Transition Metal-Carborane Complexes, 8, 87 Transition Metal-Isocynnide Complexes, 11, 21 A6 87 C8 u9 EO F1 G2 H3 14 15