Perspectives in Organometallic Chemistry

Perspectives in Organometallic Chemistry Perspectives in Organometallic Chernistry Edited by C. G. Screttas and B. R...

Author: C.G. Screttas | B.R. Steele

31 downloads 1755 Views 20MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Perspectives in Organometallic Chemistry

Perspectives in Organometallic Chernistry

Edited by C. G. Screttas and B. R. Steele National Hellenic Research Foundation, Athens, Greece

advancing the chemical sciences

The proceedings of the 20th International Conference on OrganometallicChemistry held in Corfu, Greece on 7-12 July 2002.

Special Publication No. 287 ISBN 0-85404-876-6 A catalogue record for this book is available from the British Library 0 The Royal Society of Chemistry 2003

All rights resented. Apartfrom anyfair dealingfor the purpose of research or private study, or criticism or review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemistry, or in the case of reprographic reproduction only in accordance with the t e r n of the licences issued by the Copyright Licensing Agency in the UK,or in accordance with the terms of the licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be se. to The Royal Society of Chemistry at the address printed on this page. A

Published by The Royal Society of Chemistry, Thomas Graham House, Science Park,Milton Road, Cambridge CB4 OWF, UK Registered Charity No. 207890 For further information see our web site at www.rsc.org Printed and bound by by Athenaeum Press Ltd,Gateshead, Tyne & Wear

Preface The 20* International Conference on Organometallic Chemistry, which was held in Corfu, Greece in July 2002, provided an opportunity for organometallic chemists from all over the world to learn about the latest developments in the field from young and senior researchers alike. This series of conferences dates back to 1963 and has become a major event in the calendar of those who have an interest in this important area. Prominent organometallic chemists were specially invited to make presentations at the conference and a number of them kindly agreed to submit written accounts of their recent work to be published in this special volume. The chapters in this book are thus intended to reflect state-of-the-art developments in organometallic chemistry by some of the foremost groups in the field. The aim of the conference was to provide a forum for the presentation of the latest results in all areas of interest to organometallic chemists. Organometallic chemistry is an area which touches on and plays an active role in all of the traditional divisions of chemistry, inorganic, organic, physical and theoretical, while the field of bio-organometallic chemistry is also now attracting substantial interest and it was particularly pleasing that all of these areas were represented at the meeting. It is intended that the present volume should also reflect the many facets of organometallic chemistry and, in the knowledge that an organometallic chemist is required to have a broad knowledge of most, if not all of these areas, we hope that the contents of this volume will be of wide appeal. We are grateful to all the authors for preparing their contributions within a rather strict time schedule and hope that this book will constitute a useful source of information and ideas. C.G. Screttas B.R. Steele

V

Contents Group 15 element imido and phosphido cages; Coordination chemistry and synthetic applications E L Doyle, A.D. Hopkins, G.T. Luwson, M.McPartlin, A.D. Woods and D.S. Wright Neutral clusters EnRn of the monovalent elements gallium and indium. Recent results in synthesis and reactivity W. Oh1

1

16

New titanium imido chemistry with polydentate N-donor ligands P. Mounford

28

Organometallic complexes with 1,2-dichalcogenolate-o-carboranes Guo-Xin Jin

47

Synthesis and reactivities of multinuclear sulfur-bridged metal complexes ranging from dinuclear to hexanuclear cores M. Hidai

62

a,o-Bis [(triphenylphosphine)gold(I)] hydrocarbons K.A. Porter, A. Schier and H. Schmidbaur

74

Researches on non-classical organolanthanide chemistry P.B. Hitchcock, A. G. Hulkes, A. V. Khvostov, M.F. Lappert and A. K Protchenko

86

Hyper-structured alkynylruthenium complexes: Effect of dimensional evolution on NLO properties M.G. Humphrey, M.P. Cijientes, M. Samoc, T. Isoshimu and A. Persoons Cycloaddition of alkynes mediated by [RuCp(L)]+ (L = CO, NCH, PH3) and RuCpCl complexes - Metallacyclopentatrienes as key intermediates A DFT study M.J. Calhorda, K. Kirchner and L F. Veiros Selective C-C coupling reactions of Me2N-bC-NMe2 at iron(0) centers A.C. Filippou, T. Rosenauer and G. Schnaknburg Routes to fluorinated organic derivatives by nickel mediated C-Factivation of heteroaromatics T. Braun and R.N.Perutz

Vii

100

111

120

136

'

Viii

Contents

Novel q5 - q6 rearrangement of bis(fluoreny1)lanthanide complexes by the addition of A l R 3 H. Yasuda Results and perspectives of high oxidation state organomolybdenum chemistry in water E. Collange, F. Demirhan, J. Gun, 0.Lev, A. Modestov, R. Poli, P. Richard and D. Saurenz Modulation of electronic behaviour of metal carbonyl clusters D. Collini, C. Femoni, M.C. Iapalucci, G. Longoni and P. Zunello Interionic and intermolecular solution structure of transition metal complexes by N M R A. Macchioni Synthetic and mechanistic pathways in platinum(I1) chemistry R. Romeo and L. Monsu Scolaro New perspectives for olefin complexes: Synthesis and characterisation of stable rhodium(0)and iridium(0) complexes J. Harmer, G. Frison, M. Rudolph, H. Schonberg, S. Deblon, P. Maire, S. Boulmiiaz, F. Breher, C. Bohler, H. Riiegger, A. Schweiger and H. Griitzmacher Substitution and addition reactions catalyzed by transition metal complexes I. P. Beletskuya Late transition metal (CoyRh,Ir)-siloxide complexes - Synthesis, structure and application to catalysis B. Marciniec, I. Kownacki, M. Kubicki, P. Krzyzanowski, E. Walczuk and P. Btazejewska-Chadyniak

152

167

183

196

208

222

240

253

Cheap chiral ligands for asymmetric transition metal catalyzed reactions M.T. Reetz

265

Chiral metal complexes in asymmetric catalysis C.Moberg, 0.Belda, K.Hallrnan, R. Stranne, M. Svensson, J.L. Vasse, T. Wondirnagegn and R. Zalubovskis

275

In search of asymmetric propargylic substitution reactions mediated by optically active indenyl-ruthenium(II) allenylidene complexes V. Cadierno, S. Conejero, M. P. Gamasa and J. Gimeno

285

Recent developments on hydride iridium triisopropylphosphine complexes: [IrH2(NCCH3)3(PiPr3)]BF4as hydrogenation catalyst LA. Oro, E. Sola and J. Navarro

297

Contents

ix

Pd complex-catalyzed ring-opening polymerisation of 2-aryl- 1-methylenecyclopropanes S. Kim, D. Takeuchi and K. Osakada

306

Subject Index

3 17

GROUP 15 ELEMENT IMIDO AND PHOSPHIDO CAGES; COORDINATION CHEMISTRY AND SYNTHETIC APPLICATIONS

Emma L. Doyle,' Alexander D. Hopkins,' Gavin T. Lawson,' Mary McPartlin? Anthony D. Woods,' Dominic S. Wright

'

'

Chemistry Department, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK; e-mail dsw 1OOO@,cus.cam.ac.uk. School of Chemistry, University of North London, London N7 8DB, UK

1 INTRODUCTION This review details recent developments in the synthesis and coordination chemistry primarily of Group 15 imido and phosphido cages containing a variety of anionic arrangements. The review will concentrate on the applications of these Group 15 anionic ligands in organometallic chemistry, and essentially follows the theme of the lecture given at the XXth International Conference on Organometallic Chemistry (Corfu, 2002). Further aspects of this work have been published in separate review articles."2 2 MIXED (OR STEP-WISE) METALLATION In contrast to the alkali metal organometallics (such as the ubiquitous "BuLi), the organometallics of the later p block elements (E= Group 14, Sn, Pb; Group 15, As-Bi) are significantly less polar. As a consequence previous synthetic strategies to imido (RN2-) complexes of the later main group metals had been mainly limited to procedures involving condensation with Group 15 halides (eqn. l), desilylation with SiR3 reagents (eqn. 2), or (in rare cases) reactions of alkali metal RN2- reagents with p block element salts (eqn. 3). In view of this background it is perhaps not surprising that until fairly recently very few imido complexes of the later p-block metals had been structurally characterised. EX3

+

EX3

+

EX3

+

2

Perspectives in Organometallic Chemistry

The aim at the beginning of our studies in this area was to develop a range of p- block metal reagents which were strong enough bases to effect smooth deprotonation of primary amines, allowing direct access to complexes containing the RN2- dianion. We showed in preliminary studies that Sb(NMe2)3, which is readily prepared in high yield via the reaction of LiNMe2 with SbC13 (eqn. 4); will doubly-deprotonate a broad range of primary amines even at low temperature (eqn. 5): Similar dimers of the type [Me2NSb@-NR)]2 are isolated from these high-yielding reactions. -3LiCl

SbC13

+ 3Me2NLi

Sb(NMe&

A

(4)

-4Me2NH

2Sb(NMe2)3 + 2RNH2

A

[MezNSb(p - NR)]2

(5)

The key point as regards further developments in this field to p-block metal anion arrangements is that the extremely basic nature of reagents like Sb(NMe& contrasts with the lower basicity generally observed for alkali metal organometallics (RM). In the absence of conjugation within the organic group of the primary amine only single deprotonation will follow (giving RNHM). Thus, there was the possibility that step-wise deprotonation of the primary amine, first with R'M then with Sb(NMe2)3, would lead to mixed alkali metal/p-block element imido complexes. This scenario is shown in Scheme 1.

[MI= Alkali metal Reagent; [El= Group 15 Reagent

Scheme 1

In practice the strategy of step-wise (or mixed-) metallation of primary amines (and, indeed, of primary phosphines) employing Group 15 bases of the type E Me2 3 works the very well.' Alkali metal cage compounds containing the trianions [E(NR)3] - (type

P )3.'

dianions [E2(NR)4I2- (type ZI),6 and the monoanions [Me2NE(pu-NR))2E]-(type ZZI) can be readily obtained (Figure 1). A significant point in regard to the selection of a particular anion type is that the precise anion unit obtained by these reactions (the syntheses of which are discussed in more detail later) depends on the synthetic route used. In this sense, the reaction products are kinetically controlled and a particular desired ligand grouping can be targeted by the choice of reaction conditions.

Group 15 element imido and phosphido cages TYPE I

r -I3-

3

TYPE I1

TYPE 111

1

/\ / \

2-

iR'

""\

R E /E\ER

MQN,

/N\ R\

E\N/E

E\ R '

/E\

N ,M,

1-

/E

i

7

R

E = N; E= Sb R= Cy, 'Bu, 2,5-(MeO)2C6H3. CHzPh. E = N; E= As R= 'Bu, 2-pyridyl, CHzPh, 2-MeOC6H4. E'= P; E= Sb R= 'Bu, Cy.

N

R

E= Sb, A0

E= Sb

R= Cy E= Bi R= 'Bu

R= Cy

Figure 1 Principal Group 15 imido anion types.

3 STRUCTURE AND ORGANOMETALLIC REACTIONS OF IMIDO COMPLEXES The synthesis of trianionic frameworks of type Z is the most obvious reaction sequence, involving the 3:l reaction of a primary amido alkali metal precursor (RNHM) with Sb(NMe2)3 (eqn. 6a).5"bIn the case of the As(II1) analogues, a modified reaction sequence is required owing to the lower basicity of As(NMe2)3, involving reaction of the amine with As(NMe2)3 followed by deprotonation with "BuLi (eqn. 6b)." Several structurally characterised examples containing Li' have been ~ r e p a r e d .All ~ of these have similar structures in which two [E(NR)3I3- anions are associated by six Li' cation^,^ e.g., [ { Sb(NCy)3}2Li6.2Me2NH] (1) (Figure In most cases the central N6Li6 'stack' arrangement within these complexes is retained even in the presence of extensive Lewis base solvation. The [E(NR)3I3-anions(E= Sb, As) are valence isoelectronic with the Group 16 dianions [E(NR)3]2-.7

-3Me2NH

As(NMe2)3 + 3RNH2

---+

+3"BuLi

[As(NHR)3]

[As(NR)~]~- (6a)

Bearing out the view that the [Sb(NR)3I3- trianions are the robust chemical constituents of alkali metal cages of this type, these anions are transferred intact in reactions with a range of main group and transition metal precursors.' For example, the reaction of [Cp2Pb.TMEDA] (2) with 1 gives the heterometallic cage [{ Sb(NCy)3}2Pb3] (3) (eqn. 7) (Figure 3), in which the three Pb" centres replace the six Li' cations at the centre of the cage structure.* The predominant bonding within the SbzNbPb3 core of 3 is undoubtedly between the N and Sb and Pb centres. However, it is of value to note here, in relation to the later discussion of the behaviour of phosphide analogues, that the Pb3Sbz metal core of the complex would (if an isolated fragment) conform to Wade's rules (an n+l ,closo polyhedron).

Perspectives in Organometallic Chemistry

4

Figure 2

Structure of the trianion complex I .

Figure 3

Structure of 3

Dianions can be accessed via the 1:2 stoichiometric reactions of dimers of the type [Me2NE@-NR)]2 (E= As, Sb, Bi) with RNHM (eqn. 8).6 The structures of [{E2(NCy)4}2L&](E= Sb,6aAs6c) (4) (Figure 4) consist of two [E2(NCy)4I2- dianions that are associated by four Li' cations (adopting a tetrahedral Li4 arrangement at the centre of the cage). This arrangement can be described as arising from the association of two E2N4Li2 cubane units, a view that is supported by the dissociation of the com lexes into these units in arene solutions.6dUnlike the alkali metal complexes of [Sb(NR)3] trianions, Lewis base solvation of the Li' cations results in dissociation of the cubane constituents. This is seen in the formation of the discrete cubane [Bi2(NfBu)4Li2.2thfl(5) from the reaction of Bi(NMe2)3 with 'BuNHLi in thf as the solvent.6d

P

Group 15 element imido and phosphido cages

Figure 4

5

Structure of the dianion complexes 4.

The reactivity patterns of alkali metal cages containing [E2(NR)4I2- dianions mirror that of the trianion counterparts, the dianions being transferred intact to other metal ions. This is illustrated here by the organometallic example of the reactions of the complexes 4 with Cp2Mn leading to the cubanes [E2(NtBu)4(MnCp)2] (5) (Figure 5).' Investigation of the magnetic behaviour of these complexes reveals that they are predominantly high-spin, with the magnetic properties being subtly dependent on the Group 15 element (E). The origin of this dependence stems from the geometric constraints within the [E2(NCy)4I2ligands which effect the Mn.. .Mn separation and therefore the communication between the two Mn" centres.

Figure 5 Structure of 5 The ability for ligands of this type to influence the architecture of coordinated metal cores is illustrated most dramatically by the comparison of the structures of the two Cu complexes [{ Sb2(NtBu)4}2Cuq] and [{As2(NtBu)4}2C~](7)'Ob (Figure 6). The more compact As2N2 ring of the [AS~(NCY)~]~dianion leads to a different ligand coordination mode than in the Sb analogue, in which unfavourably close Cu.. .Cu contacts are avoided. Consequently, 7 has an unusual CQ butterfly arrangement at its centre whereas 6 has a square-planar CQ core.

Perspectives in Organometallic Chemistry

6

Figure 6

Structure of (a) 6 and (3) 7.

Monoanions of type IZI can be repared by the reaction of the salt Sb(NHR)4Li with the result is a bicyclic, spiro arrangement in which Sb(NMe2)3 (1 :2 molar equivalents)! the central Sb(II1) atom has a 10e, pseudo-trigonal bipyramidal geometry with the two terminal Sb(II1) centres being 8e, pyramidal (eqn. 9). The normal coordination mode observed in these species is illustrated in the structure of [{Me2NSb(NCy)2}2Sb]Li (8) (Figure 7), in which the alkali metal ion is coordinated by two bridging NCy groups and by the terminal Me2N groups of the monoanion ligand.5a The reactivity of the parent [{Me2NSb(NCy)2}2Sb]- anion is of some interest. Reaction with primary amines leads to replacement of the Me2N groups by RNH groups, with retention of the spiro structure of the original monoanion.* In the structure of [{CyNHSb(NCy)2)2Sb]K.toluene (9) (Figure 8) the low coordination number of the K' ion is made up for by agostic interactions with the Me groups of two toluene ligands.* -4Me2NH

[Sb(NHR)4]- Li'

Figure 7

+ 2Sb(NMe2)3

Structure of the monoanion complex 8.

[{Me2NSb(p -NR)2)2Sb]

-

(9)

Group 15 element imido and phosphido cages

Figure 8

7

Structure of 9.

However, the reaction of the [{Me2NSb(NCy)2}2Sb] - monoanion with 'BuOH leads to rearrangement of the spiro structure into a nido isomer." The structure of the product [{ Sb(,~-NCy)}3(p3-NCy)(O'Bu)2]K.toluene(10) is illustrated in Figure 9. One of the major reasons for this rearrangement is the greater Lewis acidity of the Sb(II1) centres in 10 compared to those in 9, leading to an overall desire to increase the coordination numbers of the Sb centres.

T27 Figure 9

Structure of the spiro-anion complex (10).

4 PHOSPHIDE ANALOGUES

As can be concluded from the previous section, the chemically robust nature of Group 15 imido systems gives them some potentially broad applications in various aspects of coordination chemistry. There is, however, a striking difference between these systems and their phosphide analogues2 Such phosphide cages decompose into Zintl compounds containing -E: anions at relatively low temperatures, via an apparent step-wise mechanism involving heterocyclic intermediates of the type [(RP)nE]- (Scheme 2). The extent of this decomposition process and whether or not it can be limited to the intermediate heterocycles depends on a number of factors, which include,

Perspectives in Organometallic Chemistry

the organic group (R) present within the phosphide groups (d9; aromatic groups accelerate the formation of Zintl compounds, whereas aliphatic groups result in the stabilisation of the [(RPbE] - heterocycles. the presence of Me2NH in the reaction, which encourages formation of the Zintl compound. the alkali metal present in the Group 15 cage; as Group 1 is descended the decomposition of the cage is encouraged. the Group 15 element present; as the group is descended (from As to Bi) formation of the Zintl phase becomes more favourable.

Scheme 2

Underlying the thermolability of the phosphide cages is the strength of single P-P bonds, which are the strongest homoatomic bond energies of all the Group 15 elements. The latter provides the thermodynamic driving force for the conversion of the Group 1Yalkali metal cages to (effectively) an alloy phase (the Zintl phase). The ultimate formation of cylophosphazanes, [RP],,, together with the Zintl compound suggests that the [(RP) ,,El - intermediates play the role of metal atom deliverers. Although the phosphide cage C ( S ~ ( P C Y ) ~ ) ~ L ~ ~ . ~(11) M ~(Figure ~ N H ] 1Oa) can be obtained from the reaction of Sb(NMez)3 with CyPHLi (eqn. 10),l2the complex is unstable above ca. 0°C. At 30-40°C 11 undergoes thermal decomposition into the Zintl compound [Sb7Li~.6Me2NH](12) (Figure 10b).13 Interestingly, if 11 is held under vacuum and the Me2NH solvation removed then the decomposition process no longer takes place (if MezNH is bubbled through a solution of the complex rapid decomposition to 12 ensues)? 2Sb(NMe2)3

Figure 10

+ 6CyPHLi

-

[(Sb(PCy)3}2Li6.6Me2NH]

(10)

11

(a) The cage 11 and (b) the Zintl compound 12.

The decomposition of 11 can be compared to that of the related complex [ { Sb(P'Bu)}zLi6.6thfJ (13) in which the final product isolated is the bicyclic, distibane [('BuP)3Sb]z (14) (Figure 1l), together with ['BuP]~.'~ This reaction can be monitored by

Group 15 element imido andphosphido cages

9

31

P NMR spectroscopy and occurs via the heterocyclic anion [(‘BuP)3Sb] -.14 The significance of the different reactivity of 13 to that of 11 is the suggestion that coupling of the heterocyclic [(RP)nSb]- anions via an oxidative process is probably the key metalmetal bond forming process in the ultimate formation of the Sb-: anion. The fact that the decomposition of 13 stops at 14 provides some circumstantial evidence for the importance of the presence of Me2NH bonded to the alkali metal within the cage precursor and intermediates (as implied by the activation and deactivation of 11, mentioned previously).

d Figure 11

Structure of the bicyclic distibane 14.

The in situ reaction of the dimer [Me2NSb@-PCy)]z (15) with CyPHNa was carried out in an attempt to obtain the phosphide analogue of the imido cage [{Sb2(NCy)4)2Na] (16), containing the Sb2(NCy)4I2- dianion and having a structure similar to the Cu complex 6 (Figure 6)J5 If the reaction mixture is heated to 6OoC in the presence of the Lewis base donor TMEDA (= Me2NCHzCH2NMe2) the Zintl compound [Sb7Na3.3thf.3TMEDA] (17) is isolated in almost quantitative yield (together with [CyP]4).’5 However, if held at 10°C then the product is [{ (CyP)4Sb)Na.TMEDA.Me2NH]2 (18) (Figure 12), containing the heterocyclic anion [(CyP)4Sb]-.I5 The different ring size of this heterocyclic intermediate to that involved in the formation of 14 suggests that the organic substituent present in the phosphine has the major influence over ring size in these species.

Figure 12 Structure of 18. Although the Sb(II1) heterocycles are thermally unstable and therefore cannot be used readily as ligands in their own right, the greater bond energy of As-P bonds compared to Sb-P bonds leads to greater thermodynamic stability of analogous [(RP)nAs]- anions.’’-” For reactions involving aliphatic phosphines (containing R-groups like ‘Bu,Cy or 1adamantyl), the heterocyclic anions are stable at room temperature and only decompose

10

Perspectives in Organometallic Chemistry

into Zintl compounds containing As-: anions on prolonged reflux in toluene. As with the Sb(II1) systems, the ring sizes of the heterocycles isolated depend on the organic substituent. This is seen most obviously in the formation of [(CyP)4AsLi.TMEDA.thfJ(19) (Figure 13a), containing the five-membered [(CyP)4As]- heterocycle, in the 3: 1 stoichiometric reaction of CyPHLi with As(NMe2)3 in TMEDA/thf.’5”6 In contrast, the reaction of ‘BuPHLi with As(NMe& under the same conditions gives [(‘BuP),AsLi.TMEDA.thfl (20) (Figure 13b), in which a four-membered heterocyclic ring is formed.17

Figure 13

Structures of the [(RP),,As]-- complexes (a) 19 and (b) 20.

For aromatic organic substituents we have so far been unable to isolate stable heterocycles. For example, the reaction of MesPHLi (Mes= 2,4,6-Me3C&) with As(NMe2)3 in the presence of TMEDA gives the Zintl compound [As7Li3.3TMEDA](21) after stirring at room temperature for only one hour.I6 Analogous reactivity is also seen in reactions involving arsines, the reaction of PhAsHLi giving 21 in the presence of TMEDA (22) (isostructural with 14) is obtained in whereas the bicyclic compound [(‘BuAs)~As]~ the analogous reaction of ‘BuPHLi with As(NMe2)3.I6 Another observation relating to the stability of [(RP)nE]- heterocycles is their increased tendency to decompose into Zintl compounds as the size of the Group 1 counter-cation is increased. For example, the Na complex [(CyP)4AsNa.TMEDA]z (23) (whose structure is related to the Sb complex 18) is considerable less thermally stable than the Li complex 19.17 Interestingly, the reactivity patterns found for related Group 14 phosphide compounds mirror those described above for the Group 15 systems. The reactions of Sn(NMez)2 with RPHLi (R= ‘Bu, Cy) give the cages [ { Sn(p-PR)}2(p-PR)}2Li4’4thfJ(24), containing metallacyclic [ { Sn(p-PR)}2(p-PR)}2]” tetraanions (Figure 14).18 The latter can be regarded as the Group 14 analogues of precursor cages such as 11 and 13. If the same reaction is undertaken using MesPHLi in the presence of TMEDA then coupling of two of the MesP groups occurs in the product [(Sn(pPMes)}2(MesPPMes)](Li.TMEDA)2 (25) (Figure 15), containing a [{ Sn(pPMes))2(MesPPMes)l2- dianion.18Further coupling of the phosphide groups ensues in the reaction of CyPHK with Sn(NMe2)2, giving [Sn2(CyPPCy)2(pu-PCy)](K.2thf)2(26) (Figure 16), containing the [Sn2(CyPPCy)2@-PCy)14-dianion. l 9

Group 15 element imido andphosphido cages

11

Figure 14 Structure of 24, and the metallacyclic [{Sn(p-PR))2(p-PR)}2J4 tetraanion.

2Mes,

P

1-p

/Mes

Figure 1$ Structure of 25, and the [{Sn(p-PMes}}2(MesPPMe~)]~ dianion.

1*I

Figure 16 Structure of 26, and the [Sn,(CyPPCy)2(p-PCy)/” dianion.

Perspectives in Organometallic Chemistry

12

5 LIGAND CHARACTERISTICS AND REACTIVITY OF [(RP),As]- ANIONS The stable As(II1) heterocyclic anions [(RP),As] - are of interest as new ligands to a range of main group and transition metals. Of particular interest is the potential for these heterocycles to behave as sources of As atoms, as illustrated in the above section in regard to the formation of Zintl compounds. The characteristics of [(RP),As]- anions as ligands are unusual, and appear to be dominated by their large steric demands. The reaction of [(‘BuP),As] - with [CpFe(C0)2Cl] leads to the expected substitution of the metal-bonded C1 ligand, the product being [CpFe(CO)2As(‘BuP)3] (27)?’ However, reaction with [CpM(C0)3Cl] (M= Mo, W) results (28) (Figure 1 7).21 in [ ((‘BuP)3AsC~H~>M(CO>~Cl]

9

Figure 17 Structure of 28. This formal H- substitution reaction of the Cp ring probably occurs via a mechanism involving addition of the [(‘BuP)3As] - anion to the metal centres, generating a $ecies like 27, followed by transfer of the ligand to the C ring. The high steric demands of the [(‘BuP)3As] group in 28 are also apparent in the P NMR spectrum of the complex which reveals restricted rotation of the [(‘BuP)3As] ring about the As-C bond of the [(‘BuP)3AsCJ&] ligand.2’ The reactivity of [(RP)nAs]- anions with electrophiles is also worthy of mention. The reactions of H20 or organic halides (RX) with the [(‘BuP)3As]- anion provide a very simple and highly efficient route to terminally substituted triphosphines of the type [(‘BuP)~R~] (R= H, or organic group) (29) (scheme 3). This reaction clearly relies on the polarity of the As-P bonds in the [(‘BuP)3As] - anion. Ab initio MO calculations reveal that this reaction is enthalpically driven. We have recently shown that new heterocycles can also be generated using this method. For example, the reaction of the five-membered [(CyP)4As]- anion with excess MezSiC12 gives the four-membered heterocycle [(CyP)3SiMez] (30).l7

P

Group 15 element imido and phosphido cages

i““

13

1-

‘Bu-

+

M

2

I

I

‘Bu

‘Bu

Scheme 3 Terminally substituted triphosphines like 29 are of some interest since there are few simple routes available to this class of ligands and their coordination chemistry has therefore not been studied extensively. The reaction of a solution of the triphosphine [(‘BuP)3H2] (generated in the above manner) with \W(CO)4.2thq give the simple (31) (Figure 18). phosphine complex [(‘BuP)~H~.W(CO)~]

Figure 18 Structure of the triphosphine complex 31 5 CONCLUSIONS AND CLOSING REMARKS The results presented in this short review show the breadth of new chemistry that can be accessed using Group 15 reagents of the type E(NMe2)3. Imido anions (like the tripodal [E(NR),]” trianion) are stable and have broad applications as ligands to a range of main group and transition metals. However, the analogous phosphide species are thermally unstable, decomposing via heterocyclic intermediates [(RP)nE]- into Zintl compounds. The As(II1) heterocycles are stable enough to be used as new ligands or reaction precursors in their own right, exhibiting unusual reactivity and coordination chemistry. References 1 2 3 4

M.A. Beswick and D.S. Wright, Coord. Chem. Rev., 1998,176,373; M. A. Beswick, Dalton Trans., 1998,2437. M.E.G. Mosquera and D.S. Wright, J. Chem. SOC. A.D. Hopkins, J.A. Wood and D.S. Wright, Coord. Chem. Rev., 2001,216, 155. A. Kiennemann, G. Levy, F. SchuC and C. Tanielian, J. Organomet. Chem., 1972,35, 143. A.J. Edwards, M.A. Paver, M.-A. Rennie, P.R. Raithby, C.A. Russell and D.S. Wright, . I Chem. SOC.,Dalton Trans., 1994,2963.

14

5

6

7

8 9 10

11

12

13 14

15 16 17 18 19

20 21

Perspectives in Organometallic Chemistry

(a) A.J. Edwards, M.A. Paver, M.-A. Rennie, P.R. Raithby, C.A. Russell and D.S. Wright, Angew. Chem., Int. Ed. Engl., 1994, 33, 1277; (b) M.A. Beswick, N. Choi, C.N. Harmer, A.D. Hopkins, M.A. Paver, M. McPartlin, P.R. Raithby, A. Steiner, M. Tombul and D.S. Wright, Inorg. Chem., 1998, 37, 2177; (c) A. Bashall, M.A. Beswick, A.D. Bond, S.J. Kidd, M. McPartlin, M.A. Paver, A. Steiner, R. Wolf and D.S. Wright, J Chem. SOC., Dalton Trans., 2002,343. (a) R.A. Alton, D. Barr, A.J. Edwards, M.A. Paver, M.-A. Rennie, C.A. Russell, P.R. Raithby and D.S. Wright, J. Chem. SOC., Chem. Commun., 1994, 148 1 ; (b) A. Bashall, M.A. Beswick, C.N. Harmer, A.D. Hopkins, M. McPartlin, M.A. Paver, P.R. Raithby and D.S. Wright, J Chem. SOC., Dalton Trans., 1998, 1389; (c) M.A. Beswick, E.A. Harron, A.D. Hopkins, P.R. Raithby and D.S. Wright, J Chem. SOC., Dalton Trans., 1999, 107; (d) D. Barr, M.A. Beswick, A.J. Edwards, J.R. Galsworthy, M.A. Paver, M.-A. Rennie, C.A. Russell, P.R. Raithby, K.L. Verhorevoort and D.S. Wright, Inorg. Chim. Acta, 1996,248,9. (a) R. Fleischer, S. Freitag, F. Pauer and D. Stalke, Angew. Chem., 1996, 108, 208; Angew. Chem., Int. Ed. Engl., 1996, 35, 204; (b) T. Chivers, X. Gao, M. Parvez and G. Schatte, Inorg. Chem., 1996,35,4094, and references therein. M.A. Beswick, C.A. Harmer, M.A. Paver, P.R. Raithby, A. Steiner and D.S. Wright, Inorg. Chem., 1997,36, 1740. A. Bashall, M.A. Beswick, E.A. Harron, A.D. Hopkins, S.J. Kidd, M. McPartlin, P.R. Raithby, A. Steiner and D.S. Wright, J Chem. SOC.,Chem. Commun, 1999, 1 145. (a) D. Barr, A.J. Edwards, S. Pullen, M.A. Paver, P.R. Raithby, M.-A. Rennie, C.A. Russell and D.S. Wright, Angew. Chem., Int. Ed. Engl., 1994, 33, 1875; (b) A. Bashall, M.A. Beswick, E.A. Harron, A.D. Hopkins, S.J. Kidd, M. McPartlin, P.R. Raithby, A. Steiner and D.S. Wright, J Chem. SOC., Chem. Commun., 1999, 1145. (a) M.A. Beswick, N. Choi, A.D. Hopkins, M. McPartlin, M.A. Paver and D.S. Wright, J. Chem. SOC.,Chem. Commun., 1998,261; (b) A. Bashall, M.A. Beswick, N. Feeder, A.D. Hopkins, S.J. Kidd, M. McPartlin, P.R. Raithby and D.S. Wright, J. Chem. Soc., Dalton Trans., 2000, 1841 . M.A. Beswick, J.M. Goodman, C.N. Harmer, A.D. Hopkins, M.A. Paver, P.R. Raithby, A.E.H. Wheatley and D.S. Wright, J. Chem. SOC., Chem., Commun., 1997, 1879. M.A. Beswick, N. Choi, C.N. Harmer, A.D. Hopkins, M. McPartlin and D.S. Wright, Science, 1998,281, 1500. A. Bashall, F. Garcia, G.T. Lawson, M. McPartlin, A. Rothenberger, A.D. Woods and D.S. Wright, Can. J Chem., in press. M.A. Beswick, N. Choi, A.D. Hopkins, M.E.G. Mosquera, M. McPartlin, P.R. Chem. Commun., 1998,485. A. Bashall, M.A. Beswick, N. Choi, A.D. Hopkins, S.J. Kidd, Y.G. Lawson, M.E.G. Mosquera, M. McPartlin, P.R. Raithby, A.E.H. Wheatley, J.A. Wood and D.S. Wright, J. Chem. SOC.,Dalton Trans., 2000,479. A. Bashall, F. Garcia, G.T. Lawson, M. McPartlin, A. Rothenberger, J.A. Wood, A.D. Woods and D.S. Wright, J. Chem. SOC. Dalton Trans., submitted. J.E. Davies, A. Hopkins, A. Rothenberger, A.D. Woods and D.S. Wright, J Chem. SOC.Chem. Commun., 200 1,525. A.D. Bond, F. Garcia, G.T. Lawson, M. McPartlin, A.D. Wood and D.S. Wright, J. Chem. Soc., Chem. Commun., submitted. J.A. Woods, Ph. D. thesis, Cambridge, 2001. A. Bashall, A.D. Hopkins, M.J. Mays, M. McPartlin, J.A. Wood, A.D. Woods and D.S. Wright. J. Chem. SOC..Dalton Trans.. 2000. 1825.

Group 15 element imido and phosphido cages

15

22 A.R. Amstrong, N. Feeder, A.D. Hopkins, M.J. Mays, D. Moncrieff, J.A. Wood, A. D. Woods and D.S. Wright, J. Chem. Soc., Chem. Commun., 2000,2483.

NEUTRAL CLUSTERS EnRn OF THE MONOVALENT ELEMENTS GALLIUM AND INDIUM, RECENT RESULTS IN SYNTHESIS AND REACTIVITY

W. Uhl Fachbereich Chemie der Philipps-Universitat Marburg, Hans-Meerwein-Str., D-35032 Marburg, Germany; e-mail: [email protected]

1 INTRODUCTION Organoelement clusters of the heavier elements of the third main-group exhibiting strong element-element interactions have been known since about ten years. This now well established class of novel compounds, which contain the elements in unusual low oxidation states, may fkther be separated into two sub-classes: (i) Metalloidal clusters which are often charged and in which the number of aluminium, gallium and indium atoms exceeds the number of ligands, and (ii) clusters EnRn, which have an equal number of cluster atoms and substituents. The first ones have been reviewed only recently.’ They comprise compounds such as [Ga84(N(SiMe3)2)2ol2-or [Al7{N(SiMe3)2)6]-, in which the arrangement of the atoms in the clusters often resembles the structural motifs found in allotropes of the corresponding elements. Also compounds such as Inl& (R = SitBu3)’ and In8h (R = 2,6dime~itylphenyl)~ may be included in this particular class, of which boron analogues have not yet been reported. In contrast, the second group of clusters is strongly related to compounds known with boron. Examples are [All2iBul2I2- or [Ga8R8]2- (R = fluorenyl),4” which in accordance with Wade’s rules may be described as closo-clusters. Furthermore, tetrahedral clusters6 of aluminium, gallium and indium have been synthesized which are similar to B4C14 or B4(CMe3)4.7This report is focused on some current aspects of the synthesis, physical properties and chemical reactivity of neutral derivatives EnRn (E = Ga, In) which have two electrons less than the closo-compounds and, therefore, were sometimes classified as hypercloso. To the best of our knowledge, no recent results are known with the corresponding aluminium compounds, which usually form tetrahedral clusters. Their syntheses and the few results of chemical reactions are summarized in a review article which was published in 1998.6 2 SYNTHESIS OF CLUSTER COMPOUNDS WITH GALLIUM AND INDIUM ATOMS In 1993 we reported on the synthesis of the dark red cluster compound Ga[C(SiMe3)3]4 1 by the reaction of Ga2Br4.2dioxane with solvent-free LiC(SiMe3)3.8 The mechanism leading to the formation of 1 was obscure, and the yield was very low (3 to 10%). Only recently, we found a very effective route to obtain 1 in about 70% yield by the reduction of

Neutral clusters E,R, of the monovalent elements gallium and indium

17

an alkyltrichlorogallate with Rieke magnesium in hot toluene (Scheme l).9 1 possesses an almost undistorted tetrahedral cluster of four gallium atoms in the solid state. Owing to the delocalized bonding situation in the cluster with four bonding electron pairs we observe Ga-Ga distances (268.8 pm on average) which are longer than Ga-Ga single bonds in organoelement digallium derivatives (<254 pm). Upon dissolution in benzene dissociation occurs, and in very dilute solution the monomeric formula unit was detected by molar mass determination. The formation of monomers may be caused by the strong steric interaction between the bulky substituents. The structure of the monomer was determined by electron diffiaction in the gas phase at 250 OC.l0 That temperature demonstrates the unexpectedly high thermal stability of the alkylgallium(1) compound 1 impressively. Other tetrahedral clusters with gallium atoms such as GQ[ Si(CMe3)3]4 or G~[Si(SiMe3)3]4were obtained by other groups on different routes.l 1 They will not be discussed here, because almost nothing is known about their chemical reactivity. 4 Li[C13GaC(SiMe2R)3]- xTHF + 4 Mg

-THF C(SiMe2R)3

I

A? + 4 MgCI2 + 4 LiCl (RMe2Si)$ (R = Me (l),Et) Scheme 1

We had the idea that larger clusters containing more than four atoms may be accessible when we employ smaller substituents. However, clusters sterically less shielded than those given in Scheme (1) could not be obtained by that route owing to decomposition of the products by the formation of elemental gallium. tert-Butyllithium was known to yield tri(tert-buty1)gallium upon reaction with gallium trichloride. The reaction is accompanied by a redox process and elemental gallium precipitated. Therefore, we changed the reaction conditions and hoped to stop the disproportionation on an intermediate oxidation state (Scheme 2). Indeed we succeeded in isolating dark green crystals of a new compound (2) in a low, but reproducible yield of about 5% after separation of the volatile main-product tri( tert-buty1)gallium.l3 Crystal structure determination revealed that a cluster of nine gallium atoms had formed and that as originally expected the smaller tert-butyl groups gave the larger cluster size. As schematically shown in Scheme 2, the molecular centre of 2 adopts a tricapped trigonal prismatic structure. The Ga-Ga distances along the edges of the deltahedron are quite different. Small separations from the capping gallium atoms Ga3 were observed (258.8 pm on average), while longer ones resulted along the edges of the triangles of the trigonal prism (Gal -Gal, Ga2-Ga2,267.0 pm). Very long Ga-Ga distances of 298.8 pm parallel to the threefold rotation axis (Gal-Ga2) indicate only very weak bonding interactions, they were plotted by dashed lines in the drawing of Scheme (2). All structural data are in excellent agreement with the results of quantum-chemical calculations (see Table 1, next section). Compound 2 is thermally quite stable and decomposes only above 228 "C.

Perspectives in Organometallic Chemistry

18

GaC13

+

3 LiCMe3

-LicI

+

Ga(CMe3)3 + Gag(CMe3)g 2

Scheme 2 The carbon atoms attached to the cluster gallium atoms show an unusual chemical shift of 6 = 100 in the 13C NMR spectrum, which may be due to several low lying excited states and to some influence of spin-orbit ~oupling.'~ The structure of 2 is similar to that of the neutral boron subhalide B9C1915and shows a similar chemical behavior as discussed below. Organoindium clusters similar to the tetragallium compound 1 were easily obtained by the reaction of fieshly sublimed indium(1) bromide with the THF adducts of the bulky alkyllithium derivatives LiC(SiMeR'R")3 (Scheme 3). l4>l6All compounds characterized by

4 InBr

+ 4 LiC(SiMeRR')3. xTHF

-

C(SiMeRR')3

I

+

4LiBr + xTHF

(R 'RMe Si)3C R = Me, R' = Me (3) R = Me, R' = Et R = Me, R' = n-Bu R = Me, R' = i-Pr R = Me, R' = Ph R = Et, R' = Et

Scheme 3 crystal structure determinations possess tetrahedral clusters of four indium atoms. Possibly, the sterically most shielded derivative with phenyl groups attached to silicon has mother

Neutral clusters E,,R,, of the monovalent elements gallium and indium

19

structure. The only evidence for that assumption is its colour, which is orange instead of deep violet as is characteristic for the cluster compounds. Regrettably, we did not succeed in growing single crystals of that particular product. The In-In distances in the tetrahedral clusters (300 to 315 pm) are longer than those of In-In single bonds of organoelement molecules (<284 pm). They strongly depend on the steric demand of the substituents, the shortest one was observed with the trimethylsilyl derivative 3, the longest one so far was detected with SiMeziPr groups. In contrast to the tetragallium compound 1 the similar indium compound 3 remains a tetramer even in diluted benzene solution. However, the sterically more stressed compounds with SiMez(C6H5), SiMeziPr or SiMeEts substituents dissociate in solution to give the monomeric fragments with coordinatively and electronically highly unsaturated indium atoms. The tetraindium(1) compounds are thermally stable and decompose between 65 and 150 "C depending on the size of the substituents. In boiling toluene slow decomposition of 3 was observed accompanied by the precipitation of elemental indium. A remarkable by roduct (4) of the synthesis of 3 was isolated in up to 24% yield only recently (Figure l).'f It possesses a chain of three indium atoms connected by two In-In single bonds. Both terminal indium atoms are bridged by two bromine atoms. A third bromine atom is coordinated to the inner indium atom and occupies a bridging position to the Li(:THF)3 counter ion. Such subhalides derived from the cluster compound 3 still containing indium in an unusual low oxidation state are discussed in more detail in chapter 4. C(SiMe3)3

I

4

Figure 1 Schematic drawing of the triindium compound 4

3 ELECTRON TRANSFER Cyclovoltammetry of both tetrahedral compounds 1 and 3 revealed a reversible transfer of one electron to the GQ and In4 clusters at quite similar potentials (-1.99 V versus ferroceni~dferrocene).'~ In the case of the gallium cluster 1 the radical could be determined by ESR spectroscopy. The hyperfine splitting of the resonance can be simulated perfectly on the assumption of complete spin delocalization and the coupling of the unpaired electron with four equivalent gallium atoms. Thus, the tetrahedral symmetry of the cluster seems to be retained after the addition of one electron, although the lowest unoccupied state of these clusters is formed by a degenerate set of three molecular orbitals. Up to now, we did not succeed in generating these radical anions on a preparative scale and in isolating these interesting paramagnetic products for a more complete characterization. Two further redox processes were detected for both clusters. A second reduction wave at -3.0 and -2.8 V, re-

20

Perspectives in Organometallic Chemistry

spectively, is irreversible, as is the electrochemical oxidation at about +0.8 and +0.25 V. The indium cluster is much easier to oxidize, which may be related to its lower stability. A similar result was obtained with the nonagallium cluster 2, which showed a reversible one-electron reduction wave at -1.74 V (versus ferrocene'"), an irreversible reduction at -2.7 V and an irreversible oxidation at +0.4 V.I3 The radical anion resulting from the first reduction step could not be generated by treatment of 2 with alkali metals or their activated forms. Metal exchange seems to be the main reaction in these cases, and elemental gallium precipitated almost quantitatively. We found a suitable reductant with decamethylcobaltocene, which gave the bright green nonagallium radical anion 5 by the transfer of one electron to the cluster in 83% yield (Scheme 4).18The arrangement of the gallium atoms in

Scheme 4

the cluster still corresponds to a tricapped trigonal prism, however, the Ga-Ga distances changed compared to the neutral starting compound 2 (Table 1). The very long Ga-Ga separations along the edges of the trigonal prism parallel to the threefold rotation axis in 2 (298.8 pm) became significantly shorter by 17 pm (281.9 pm), while the edges of the triangle of the prism were elongated from 267.0 pm in 2 to 274.7 pm in the radical 5 . Only the distances to the bridging gallium atoms remained almost unchanged (258.8 compared to 255.8 pm). These experimental results are in excellent agreement with the data obtained by quantum-chemical calculations (Table 1). The frontier orbitals of both compounds are almost identical. The unpaired electron of the radical is located mainly in p-orbitals of the gallium atoms (0.993 e) with the largest contribution by the p-orbitals of the capping atoms (80%). These orbitals form a belt at the surface of the cluster with some interactions to orbitals of the gallium atoms of the prism. These interactions cause the shrinkage of the cluster parallel to the threefold rotation axis.

Table 1 Important structural parameters of the nonagallium clusters in the neutral compound 2 and in the radical anion 5 (numbering scheme in Scheme 2)

Gal -Ga3, Ga2-Ga3 Gal -Gal, Ga2-Ga2 Gal -Ga2

Gag(CMe3)g Gag(CMe3)9 experimental ab initio 258.8 pm 256.3 pm

I

[Gag(CMe3)9]'ab initio 254.4 pm

[Gas(CMe3)9]'experimental 255.8 pm

267.0 pm

267.3 pm

275.1 pm

274.6 pm

1298.8 pm

298.2 pm

281.6 pm

281.9 pm

The reaction according to Scheme (4) impressively shows the narrow relationship between these gallium clusters and boron compounds such as B9C19. Also with the last one the transfer of one electron was observed yielding a nonaboron radical anion.15 The same alteration of distances in the cluster was observed with the important shrinkage of the cluster parallel to the threefold rotation axis. However, the reduction of the boron compound succeeded at milder conditions with a much lower reduction potential. Nevertheless,

Neutral clusters E,,R,, of the monovalent elements gallium and indium

21

in this case these reactions veriQ the strong similarity between boron and its heavier homologues. 4 INDIUM SUBHALIDES

Careful oxidation of the tetraindium cluster 3 should yield compounds which still have indium in low oxidation states. An example is the sulfur derivative 6, which was obtained in our group by treatment of 3 with propylene sulfide under thoroughly controlled reaction conditions (Figure 2).19 It has one sulfur atom bridging one face of the In4 tetrahedron, and the average oxidation state of the indium atoms is +1.5. Complete oxidation by chalcogen atoms yielded heterocubane type molecules E&&.6 Compounds similar to 6 with halide atoms attached to indium would be very helpfbl for secondary reactions and for the facile generation of further cluster derivatives by salt elimination. However, treatment of 3 with the free halogens bromine and iodine gave mixtures of several unknown products regardless of the stoichiometric ratio of the starting materials. As mild chlorine or bromine transfer reagents we finally employed 1,2-dibromoethaneand hexachloroethane. C(SiMe3)3

I

6

Figure 2 Schematic drawing of the In& compound 6 Hexachloroethane reacted with 3 in hot toluene to yield a pale yellow product in 49% yield (Scheme 5) which was identified by crystal structure determination as the dimeric dialkyldichlorodiindium derivative 7.2' Two molecules of chlorine were transferred to the cluster of 3, which is accompanied by an oxidation of the indium atoms from +1 to +2. This oxidation results in a more localized bonding situation, and the indium atoms form pairs connected by In-In single bonds (282.3 pm). The overall structure of 6 may be derived from that of the tetrahedral cluster of 3. Two opposite edges of the tetrahedron become InIn a-bonds, while the remaining four edges are bridged by chlorine atoms. Treatment of 3 with one equivalent of dibromoethane yielded a Unique product (8) in which the tetrahedral arrangement of the indium atoms is retained (Scheme 6):' One face of the tetrahedron is bridged by a bromine atom, which results in a considerable distortion of the cluster with very long In-In distances at the edges of that face (361 to 407 pm). One of these edges is bridged by an additional bromine atom. The three remaining In-In distances (284.2 to 290.4 pm) are in the range of In-In single bonds. As in compound 6 cited above, a mixed valent compound was formed with an average oxidation state at the indium atoms of +1.5.

Perspectives in Organometallic Chemistry

22

C(SiMe&

I

+ (Me3Si)3C

C2C16

-"C2C12"

-

C( SiMe3)3

I

I Scheme 5

Elemental bromine reacted with 3 to give an inseparable mixture of at least four unknown products. Decomposition was an important side reaction; elemental indium precipitated and HC(SiMe3)3 was detected by NMR spectroscopy in considerable concentration. The alkane derivative was possibly formed by a radicalic cleavage of In-C bonds and the reaction of the radicals with the solvent. Thus, the attack of bromine is not only directed to the cluster centre of the molecules, but gives an unspecific reaction characteristic

3

(Me3Si)3C I

In

8

Scheme 6

Neutral clusters E,,R, of the monovalent elements gallium and indium

23

of many processes involving radicals. We hoped to prevent such a course by polarizing the bromine molecules with equimolar quantities of aluminium tribromide. Indeed, we isolated yellow crystals of the product Ir~Br4[C(SiMe3)3]4in 28% yield nowyo which is an analogue of the tetrachloro derivative 7 described above (Scheme 5). Another, quite remarkable product (9) was isolated from the reaction of 3 with a mixture of iodine and aluminium triiodide (Scheme 7).219 has a chain of three indium atoms which are connected by In-In single bonds. The terminal indium atoms are bridged by iodine atoms and tetracoordinated, while the middle indium atom has the coordination number three. The oxidation states are +2 at the terminal indium atoms and +1 at the inner indium atom. Thus, 9 is an analogue of the recently isolated by-product of the synthesis of 3, which is depicted in Figure 1 (compound 4) and which has a third halogen atom coordinated to the inner indium atom. 9 was, however, formed in a much more specific reaction and was isolated in 73% yield. By-products are elemental indium and the alkane derivative HC(SiMe3)3. In4[C(SiMe3)3]4 + AlI&

-

3

C( SiMe3)3

I

I

C(SiMe3)3 9

Scheme 7 These organoindium subhalides form a very interesting series of novel compounds (see summary in Figure 3) which cover a broad range of oxidation states at their central atoms and which may be suitable as starting materials for secondary reactions such as substitution or cluster modification. Starting with the tetraindium cluster, which has an oxidation state of +l at its indium atoms, the first oxidation by dibromoethane gives the Ir4Br2 compound 8 with mixed valent indium atoms (average +1.5). An average oxidation state of +1.66 was observed with the In312 derivative 9, while the In4X4 compounds (e.g. 7) contain bivalent indium atoms. Up to now, a similarly fascinating variety of secondary products was not observed for the tetragallium compound 1. As the only derivative, Ga212[C(SiMe3)3]~was isolated in trace amounts by the reaction of 1 with iodine. In contrast to the analogous indium compound 7 it is a monomer in the solid state with coordinatively unsaturated gallium atoms. 5 HOMOLEPTIC TRANSITION METAL DERIVATIVES

As described above, the tetragallium cluster 1 dissociates into its monomeric fragments in the gas phase or upon dissolution in benzene. In contrast, the tetraindium analogue remains tetrameric in solution owing to the lower steric stress in the molecules. However, its monomer was detected in the mass spectrum or it could be trapped by cycloaddition reac-

24

Perspectives in Organometallic Chemistry

R

R I

R

I

I

Br-In-Br

I

Br-In-Br I R

In I R

Br

+I

R

I

+2

+1,66

+1,5

R Figure 3 Organoindium subhalides and the average oxidation states of their indium atoms [R = C(SiMe3)3] tions6 These monomers have coordinatively and electronically highly unsaturated gallium or indium atoms. They resemble carbon monoxide owing to their lone electron pair and to their two acceptor orbitals with x-symmetry (p-orbitals). Indeed, we succeeded in replacing bridging CO ligands in transition metal carbonyl complexes by the simple treatment of the carbonyl com lexes with the cluster compounds. These results have been summarized in a recent review!f and do not need a detailed discussion here, some examples are given in Fig-

10

0

C(SiMe3)j

I

Fe(C0)4 (Me3Si)jC

11

C( SiMe3)3

C(SiMe3),

I

(Me3Si),C

C( SiMe3)3

Figure 4 Bridging InR and GaR Iigands in transition metal carbonyl complexes

Neutral clusters E J , of the monovalent elements gallium and indium

25

ure 4. In most cases, the products obtained are isostructural to the pure carbonyls. Exceptions are for instance the gallium iron compound 10, in which the GaR ligands and one CO group bridge all edges of the Fe3 triangle, or the monosubstitution product of enneacarbonyl diiron (1 l), in which only the InR group is in a bridging position. Terminal GaR or InR ligands could not be introduced by the direct reaction of mononuclear carbonyl complexes with the clusters. However, cyclooctadiene derivatives of nickel and platinum proved to be suitable reagents that gave homoleptic gallium and indium complexes in high yield (65 to 76%, Scheme 8). Three compounds were isolated and completely characterized so far: Ni(ER)4 (E = Ga, In) and ~ ~ ( I I I RThe ) ~ transition .~~ metal atoms of all compounds have an ideally tetrahedral coordination sphere with very short metal-gallium (Ni-Ga 217.0 pm) or metal-indium distances (Ni-In 23 1.O pm; Pt-In 244.1 pm). The M-Ga-C and M-In-C groups are linear. All products are quite stable and decompose only above 325 "C (12) or 197 "C (14). They are remarkable analogues of tetracarbonyl nickel Ni(C0)4 or the platinum compound Pt(C0)4, which is known in matrix at low temperature only. As was shown by quantum-chemical calculations, the bonding situation, too, is quite similar to that of the carbonyl complexes with strong x-back bonding of electron density from nickel or platinum into the empty p-orbitals at gallium or indium. Thus, the monomeric fragments of the clusters 1 and 3, GaR and InR, make an important contribution to the well-known class of CO analogous ligands such as isonitriles or phosphorus trifluoride. E4[C(SiMe3)3]4 + M(COD)2 E=Ga: 1 E = In: 3

-

M = Ni, Pt

C(SiMe&

I

E

+ 2COD

M = N i , E = G a : 12 M = Ni, E = In: 13 M = Pt, E = In: 14 (COD = cyclooctadiene)

Scheme 8

6 CONCLUSION Cluster compounds of the heavier elements of the third main-group form a now well established class of novel compounds. Most derivatives have metalloidal structures in which the number of cluster atoms exceeds the number of the substituents and which up to now have no counterparts in boron chemistry. A minor group of clusters resembles compounds known with boron, such as dianionic closo-clusters or the neutral clusters with four or nine gallium or indium atoms. The chemical behavior of the last ones is singular in most cases. We observed complete degradation of the clusters, as was shown here with the terminal

26

Perspectives in OrganometallicChemistry

coordination of the monomeric fragments to transition metal atoms in coordination compounds, or we found larger di-, tri- or tetranuclear fragments with the subhalides, in which the indium atoms still possess low oxidagon states. On the other side, the electron transfer to the nonagallium compound and the influence of that reduction on the structural parameters give an impressive example for a strong analogy of that chemistry to the chemistry of the lightest homologue boron. One of the most fascinating future tasks in that field is the extension of our knowledge of differences and similarities in that family of elements. References

1 G. Linti and H. Schnockel, Coord. Chem. Rev., 2000, 206, 285; A. Schnepf and H. Schnockel, Angew. Chem., 2001, 113, 734; Angew. Chem. Int. Ed., 2001, 40, 712; A. Donchev, A. Schnepf, G. Stoljer, E. Baum, H. Schnockel, T. Blank and N. Wiberg, Chem. Eur. J., 2001, 7, 3348; M. Kehnvald, W. Kostler, A. Rodig, G. Linti, T. Blank and N. Wiberg, Organometallics, 2001,20, 860. 2 N. Wiberg. T. Blank, H. Noth and W. Ponikwar, Angew. Chem., 1999,111, 887; Angew. Chem. Int. Ed., 1999, 38, 839; N. Wiberg, T. Blank, A. Purath, G. Stoljer and H. Schnockel, Angew. Chem., 1999,111,2745; Angew. Chem. Int. Ed., 1999,38,2563. 3 B.E. Eichler, N.J. Hardman and P.P. Power, Angew. Chem., 2000, 112, 391; Angew. Chem. Int. Ed., 2000,39,383. 4 W. Hiller, K.-W. Klinkhammer, W. Uhl and J. Wagner, Angew. Chem., 1991, 103, 182; Angew. Chem. Int. Ed. Engl., 1991,30, 179. 5 A. Schnepf, G. Stoljer and H. Schnockel, 2. Anorg. Allg, Chem., 2000,626,1676. 6 W. Uhl, Rev. Inorg. Chem., 1998,18, 239. 7 T. Mennekes, P. Paetzold, R. Boese and D. Blaser, Angew. Chem., 1991, 103, 199; Angew. Chem. Int. Ed. Engl., 1991,30, 173. 8 W. Uhl, W. Hiller, M. Layh and W. Schwarz, Angew. Chem., 1992, 104, 1378; Angew. Chem. Int. Ed. Engl., 1992,31, 1364. 9 W. Uhl and A. Jantschak, J. Organomet. Chem., 1998,555,263. 10 A. Haaland, K.-G. Martinsen, H.V. Volden, W. Kaim, E. Waldhor, W. Uhl and U. Schiitz, Organometallics, 1996, 15, 1146. 1 1 N. Wiberg, K. Amelunxen, H.-W. Lerner, H. Noth, W. Ponikwar and H. Schwenk, J. Organomet. Chem., 1999,574,246; G. Linti, J. Organomet. Chem., 1996,520, 107. 12 R.A. Kovar, H. Derr, D. Brandau and J.O. Callaway, Inorg. Chem., 1975,14,2809. 13 W. Uhl, L. Cuypers, K. Harms, W. Kaim, M. Wanner, R. Winter, R. Koch and W. Saak, Angew. Chem., 2001,113,589; Angew. Chem. Int. Ed., 2001,40,566. 14 W. Uhl, A. Jantschak, W. Saak, M. Kaupp and R. Wartchow, Organometallics, 1998, 17, 5009. 15 H. Binder, R. Kellner, K. Vaas, M. Hein, F. Baumann, M. Wanner, R. Winter, W. Kaim, W. Honle, Y. Grin U. Wedig, M. Schultheiss, R.K. Kremer, H.G. von Schnering, 0. Groeger and G. Engelhardt, 2. Anorg. Allg. Chem., 1999,625, 1059. 16 W. Uhl, R. Graupner, M. Layh and U. Schiitz, J. Organomet. Chem., 1995,493, C1. 17 W. Uhl, F. Schmock and G. Geiseler, 2. Anorg. Allg. Chem., in press. 18 W. Uhl and L. Cuypers, unpublished. 19 W. Uhl, R. Graupner, W. Hiller and M. Neumayer, Angew. Chem., 1997, 109, 62; Angew. Chem. Int. Ed. Engl., 1997,36,62. 20 W. Uhl and S. Melle, Chemistry Eur. J., 2001,7,4216. 21 W. Uhl, S. Melle, G. Geiseler and K. Harms, Organometallics, 2001,20,3355.

Neutral clusters EnRnof the monovalent elements gallium and indium

27

22 W. Uhl, M. Pohlmann and R. Wartchow, Angew. Chem., 1998, 110, 1007; Angew. Chem. Int. Ed. Engl., 1998, 37, 961; W. Uhl, M. Benter, S. Melle, W. Saak, G. Frenking and J. Uddin, Organometallics, 1999, 18, 3778; W. Uhl and S. Melle, 2. Anorg. Allg. Chem., 2000,626,2043.

NEW TITANIUM IMIDO CHEMISTRY WITH POLYDENTATE N-DONOR LIGANDS

Philip Mountford Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR,UK. WWW: http://www.chem.ox.ac.uk/researchguide/pmountford.html

1 INTRODUCTION Only first fully authenticated in 1988, terminal Group 4 imido complexes of the type [(Ln)M=NR] (R typically is tert-butyl or aryl, (L)n is a supporting ligand or ligand set) have been the focus of sustained activity. One of the main points of interest for these systems is the chemistry of the polar and reactive M=NR multiple bonds themse1ves.l These can undergo a range of both stoichiometric and catalytic coupling reactions with unsaturated organic substrates, as well as C-H bond activation and other processes. This short review article focuses mainly on our own contributions to the chemistry of terminal imido titanium compounds, with a specific emphasis on reaction chemistry of the Ti=NR linkage itself. A brief account of our own early work in titanium imido chemistry up to mid- 1997 has been published;2 Wigley has provided a comprehensivereview of all transition metal terminal imido chemistry up to 1994.3 It is not the purpose of this contribution to review comprehensively the work of other groups in Group 4 imido chemistry. Nevertheless, it is important to provide the reader with selected milestones and highlights in order to set the scene and context for the remainder of this article. Burger and Wannagat reported a terminal imido compound of titanium, namely [Ti(NSiMe3)C12(py)2]as early as 1963.4 No NMR or structural data were provided and no further reports of the chemistry of this compound have appeared. However, the octameric, pyridine-free analogue [Ti(p-NSiMe3)C12]8 has recently been described5 and has proved to be a source of crystallographically authenticated terminal trimethylsilylimido complexes with macrocyclic ligands.6 However, attempts to prepare Burger and Wannagat’s compound by reaction of this octamer with pyridine gave illdefined mixtures and elimination of Me3SiC1.7 During the quarter of a century following Burger and Wannagat’s paper, a number of reports of Group 4 bi- and poly-nuclear. bridging imido compounds appeared, for example the compounds [M2( pNfBu)2(NMe2)4] (M = Ti, Zr)899 and [Cp2Ti2(p-NtBu)2C12].*o

New titanium imido chemistry with polydentate N-donor ligands

29

Interest in Group 4 imido chemistry was revitalised in 1988 with the independent reports by Bergman' 1 and Wolczanski12 of highly reactive terminal imidozirconium transients, namely [Cp2Zr(NfBu)](Scheme 1j and [Zr(NSitBu3)(NHSitBu3)2],by alkane elimination from amido-alkyl precursors. Trapping of the bis(cyclopentadieny1) compound with THF provided the first structurally characterised terminal imidozirconium compound [Cp2Zr(NfBu)(THF)]. Subsequently, reactive transient imidotitanium complexes were soon reported by Wolczanski13 (Figure 2) and then by Livinghouse. l4a Just prior to these reports the first crystallographically authenticated terminal imido complexes of titanium, [Ti(NPh)(0-2,6-CsH3Pri2)2(py')2] (py' = 4 pyrrolidinopyridine) and [Ti { NP(S)Ph2}C12(py)3] were simultaneously described by Rothwell and Roesky.

R

i

RC=CR CP~Z~~--.N-BU'

c

A

/"

cp2zr\

1

Y

Reactive transient

tBu

/

I

H

""'"\ TtBu

Scheme 1 Generation and selected reactions of the reactive transient [Cp~Zr(WBu)]ll

Schemes 1 and 2 summarise some of the exciting and novel chemistry first shown by the reactive zirconium11 and titanium13b imido transients of Bergman and Wolczanski, respectively.

Perspectives in Organometallic Chemistry

30

H

I

But3S{

But3si\

=/

-Ti-Me

But3Sio' *

*

But3Si0 R = H, Et, Bz, Ph

1

Bu'3SiO.. . ;Ti=N-SiBu Bu'3SiO'

'1

MeCzCMe

3

Reactive transient

0

\*

,,,,<

But3si\

+ SiBu'3 I

Scheme 2 Generation and selected reactions of the reactive transient (Ti(MSitBu3)(OSitBu3)2] I 3 As mentioned (and as is also the case for Wolczanski's diamide-imido zirconium species [Zr(NSifBu3)(NHSitBu3)2] 13a) the reactive species is formed by thermally induced alkane elimination from of an amido-alkyl precursor. The transients were shown

New titanium imido chemistry with polydentate N-donor ligands

31

to activate sp2-or even sp3-hybridised C-H bonds in the absence of other substrates such as unsaturated org anic molecules like alkynes, or Lewis bases such as tetrahydrofuran; self-trapping via dimerisation to form bridg ing imido compounds of the type [M 2 ( ~ NR)2(Ln)2] is also typically an option for these highly reactive species. Methane activation by Group 4 imido compounds has been addressed computationally by Cundari.17 In 1991 and 1992 half-sandwich cyclopentadienyl imido complexes of the Group 4 metals were reported by Roeskyls'and Wigley.19 These are the Lewis base adducts of the types of transient intermediates postulated by Livinghouse. Interestingly, halfsandwich imidotitanium complexes have recently been implicated in alkyne and allene hydroammination reactions 'catalysed' by the bis(cyclopentadieny1) compound [Cp2TiMe2].20 Our group was the first to report Lewis base-stabilised bis(cyclopentadieny1)- and indenyl-cyclopentadienyl-imidosandwich compounds of titanium,21°the analogues of Bergman's [Cp2Zr(NfBu)(L)]complexes. Andersen and Bergman described monomeric. base-free [Cp*2Ti(NPh)],22 (the homologous congener of the transient [Cp2Zr(NtBu)]) which undergoes cycloaddition and/or C-H bond activation reactions with certain alkenes and alkynes. However, of all the cyclopentadienyl-supported Group 4 imido compounds reported, it is the bis(cyclopentadieny1)zirconium system that Bergman,20a,23and also others,24,25 have most studied with respect to the fascinating reactivity of its Zr=NR linkage. Over the past ten years a range of other new classes of Group 4 terminal imido compounds have also been reported and (setting aside our own contributions mentioned later below) the supporting ligand or ligand sets include porphyrin,2 bis(benzamidinate),27calix[4]arene,28 and other tetradentate, dianionic m o i e t i e ~ .The ~~,~~ 'parent' titanium imide [Ti(NH)C12(0PPh3)2] in which the imido nitrogen bears only i hydrogen atom substituent has also be reported.31 While much of the emphasis has been on the reactions of Group 4 imides with organic substrates, Winter has shown that terminal titanium imido complexes are implicated in the MOCVD formation of titanium nitride.32

2 SYNTHESIS AND SCOPE In a preliminary communication2la in 1994 we reported an extremely general synthesis of titanium imido compounds starting from the imido-dichloride complex [Ti(NtBu)C12(py*)2] (py* = 4-tert-butylpyridine), itself readily made from TiC14, t B u N H 2 and py*. This approach was inspired by Winter's synthesis 01 [Ti(NtBu)C12(OPPh3)2] .32 The new examples of imidotitanium complexes first made by C1 and py* substitution reactions of [Ti(NtBu)C12(py*)2] were [Cp2Ti(NtBu)(py*)], [ T i ( N t B u ) ( M e n t a a ) ] (n = 4 or 8, H 2 M e n t a a = tetra- or octa-methyl dibenzotetraaza[ 14]annulene), [ T p * T i ( N t B u ) C l ( p y*)] (Tp* = tris(3,5dimethylpyrazolyl)hydroborate), the indenyl complex [(q-QH4Me3)Ti(NC)Cl@y*)], as well as the half-sandwich cyclopentadienyl compounds [(q-C5R5)Ti(NtBu)C(py*)] (R = H or Me) of the type previously reported by Roesky.18a Full papers on the syntheses of these initial systems have appeared.21b733934

32

Perspectives in Organometallic Chemistry

However, although useful, the 4-tert-butylpyridine-derived complex [Ti(NtBu)C12(py*)2] was not entirely satisfactory for a number of reasons: residual py* can be troublesome to remove from reaction mixtures, and complexes containing it tend to have very high solubility which can hamper clean isolation and crystallisation. We therefore developed the bis- and tris-pyridine homologues [Ti(NtB~)C12(py)~] (n = 2 or 3), the latter being easily made on 40 g scale.35 Moreover, treatment of [Ti(NtBu)C12(py)3] with anilines ArNH2 gives facile access to arylimido analogues [Ti(NAr)C12(py),] via a tert-butylimidolaniline exchange protocol of the type first reported by Bergman.36 Recently we developed a complimentary and, in some ways more flexible, route to imidotitanium complexes starting from [Ti(NMe2)2C12] and RNH2 (R = wide range of alkyl or aryl). This method gives the dimethylamino-substituted complexes [Ti(NR)C12(NHMe2)2], some of which feature interesting supramolecular interactions.37 Subsequent studies confirmed the usefulness of [Ti(NR)C12(py)n] (R = tBu or aryl; n = 2 or 3) for the synthesis of hrther families of titanium imido complexes with the incorporated ligands including: aryloxides,38 Schiff s bases,39 c a l i ~ [ 4 ] a r e n e s , ~ ~ cyclooctatetraenes,41 triazacyclo-nonanes and -hexanes and some thia-analogues,42 amidinates,43 tris(pyrazoly1)-methanes and -methides,44 and tri- and tetra-dentate diamide-derived systems.45 Screening of these complexes for reactions at the Ti=NR bond itself led us to focus on three different systems in particular, the syntheses and molecular structures of which are summarised in Scheme 3. All three systems (i.e., 1 and 2 , 3 and 4, and 5) have a well-defined supporting ligand set. The compounds [Ti(NR)(Me4taa)] (R = fBu 1 or aryl2)34 have a quite rigid supporting ligand in Meqtaa (we also studied some Megtaa homologues), and the titanium possesses a square base pyramidal geometry and a 14 valence electron count. These imides are isoelectronic and isolobal analogues of the terminal 0x0 complexes [Ti(O)(Meqtaa)] shown by Geoffroy to undergo cycloaddition reactions at the Ti=O linkage.46 The half-sandwich amidinates [(q-C5R34Me)Ti{ R2C(NR1)21(NR)] (R = tBu 3"or aryl 4; (R*, R2) = (SiMe3, Ph) or (iPr, Me); R3 = H or Me)43b,43c,47possess pseudo-octahedral, 1 6 valence electron titanium centres. The cyclopentadienyl-amidinate ligand set is easily modified to explore the importance of steric factors with regard to both the cyclopentadienyl ring and amidinate ligands. The diamido-pyridine compound [Ti(NtBu)(N2Npy)(py)] (5)45a has a trigonal bipyramidal geometry and, as the pyridine adduct, possesses a 16 valence electron titanium. However, it is thought that this compound reacts through its pyridine-free analogue [Ti(NtBu)(N2Npy)] which can be prepared quantitatively from 5 by high vacuum sublimation, and has been structurally characteri~ed.~S~ Four-coordinate [Ti(NtBu)(N2Npy)]has an approximately trigonal pyramidal geometry and a 14 valence electron count. The central diamide-imide trigonal planar 'core' of 5 and [Ti(NBu)(N2Npy)]is reminiscent of those in Wolczanski's transient titanium13 (Scheme 2) and zirconiuml2 imides. There are further significant differences between the Me4taa and cyclopentadienyl-amidinateligand sets and N2Npy. Firstly, the supporting ligands in 1-4 have fixed hapticities (although in principle cyclopentadienyl ring slippage could occur in 3 and 4), while the pyridyl group of N2Npy can readily decoordinate from the metal centre during the course of a reaction if this is necessary to stabilise the product. Secondly, there is negligible It-donation to titanium from the nitrogen donors in 1-4,

New titanium imido chemistry with polydentate N-donor ligands

33

whereas in 5 the amide nitrogens of N2Npyare known to be good .n-donors. The amides can therefore compete with the n-donor imido ligand for available dn acceptor orbitals on titanium; this competing n-donor effect is thought to destabilise the Ti=Nimide linkage, enhancing the reactivity of 5.

I

Li2 [Me4taa]

PY 5

(i) Li[C5R3@e]

(ii) Li [R2C(NR) 1

-*R 3R3

3R i*3R3 Arm2 *

-'BuNH2 n

3

Ar = 2,6-C6H3Me2or C6H4Me

R' 4

Scheme 3 General synthesis of [Ti(IVR)(Me4taa)] (R = tBu 1 or alyl 2), [(qC5R34Me)TiJR2C(NR1)2)(NR)](R = tBu 3 or aryl4; (R1, R2) = (SiMej, Ph) or tPr, Me); R3 = H or Me) and [Ti(NR)(N2NpY)@y)] (5)34~45443b~43c~47

34

Perspectives in Organometallic Chemistry

3

REACTIONS OF TITANIUM IMIDO COMPLEXES OF N-DONOR LIGANDS

3.1 Dibenzotetraaza[14lannulene (Me,taa) supported titanium imido complexes The compounds [Ti(NR)(Me4taa)] (R = tBu 2 or aryl3) undergo cycloaddition reactions with tBuNCO, ArNCO, C 0 2 and TolNCNTol (To1 = 4-CgH4Me).48 S o m e corresponding reactions of the terminal hydrazido complexes [Ti("Phz)(Me4taa)] and the zirconium imide [Zr(N-2,6,-CgH3iPr2)(py)(Me4taa)]were also studied and gave similar results. The reactions of [Ti(NtBu)(Me4taa)] (1) with isocyanates were illbehaved; with PhNCO no tractable product was obtained, and with tBuNCO the N,Obound ureate [Ti{N(tBu)C(NtBu)O}(Me4taa)] (Equation 1) could be made on an NMR tube scale but not isolated. Different behaviour of tert-butylimides compared with the aryl analogues is typical of much of the chemistry of the Me,taa and cyclopentadienylamidinate supported imido complexes. However, reaction of 1 with C02 did give the N,O-bound carbamate product [Ti{N(fBu)C(O)O} (Melrtaa)] in very good yield, just as for the arylimido systems discussed below. ,But

N

1

The reactions of the arylimido systems [Ti(NAr)(Me4taa)](2) were generally welldefined; specific examples for the p-tolylimido system are shown in Scheme 4. These were the first cycloaddition reactions of Group 4 imides with these substrates to lead to metal-bound coupled products. The formation of the N,N-bound ureate [Ti{N(Tol)C(O)N(Tol)}(Me4taa)] as opposed to an N,O-bound isomer as shown in Eqn. 1 for the tert-butyl homologue is attributed to an apparent absence of any adverse steric factors in the arylimide/arylisocyanate system. The C02 reactions were only the second reported for transition metal imides (the first being for an iridium and the guanidinate product [Ti(N(Tol)C(NTol)N(Tol)} (Me4taa)l was the first to be prepared this way. The carbamates [Ti {N(R)C(O)O}(Me4taa)l (R = tBu or Tol) are mildly light-sensitive, decomposing to the corresponding RNCO and [Ti(O)(Me4taa)]. The ureate and carbamate products in Scheme 4 could be made also independently from [Ti(O)(Me4taa)] and TolNCNTol or TolNCO, respectively. The former reaction (Equation 2) is a novel example of the formal insertion of an unsaturated substrate into a metal-oxygen multiple bond.

New titanium imido chemistry with polydentate N-donor ligands

35

,To1 N

/I

TolNCO

I

TolNCNTol

/

Scheme 4 Reactions of [Ti(NTol)(Me4tuu)] with TolNCO, C02 and ToINCNTOI~~

ArNCNAr

Ar

=

The ureate complex [Ti{ N(Ph)C(O)N(Tol)}(Me4taa)] (made from [Ti(NTol)(Mestaa)] and PhNCO) undergoes an isocyanate exchange reaction in the presence of an excess of PhNCO to yield TolNCO and [Ti{N(Ph)C(O)N(Ph))(Me4taa)] (which could be independently prepared from [Ti(NPh)(Me4taa)] and PhNCO and verified). This implies a reversible C-N bond forming process which, in principle, could proceed via an associative or dissociative mechanism. Crossover experiments suggested that the reaction in fact goes by an associative mechanism involving a biuret complex (Eqn. 3).

36

Perspectives in Organometallic Chemistry

R

,\ [Ti1

PhNCO -TolNCO*

1

R

Th sc/N\cyo

Ph/LITil/LToj

-TolNCO

/"\

TolNCO

[Ti] = Ti(Me4taa)

[Ti1

(3)

3.2 Cyclopentadienyl-amidinate supported titanium imido complexes

The second type of titanium imido complex that we turn to are the cyclopentadienylamidinate supported complexes of the type [(q-C5R34Me)Ti{ R2C(NR1)2}(NR)] (R = tBu 3"or aryl 4; (R1, R2) = (SiMe3, Ph) or (iPr, Me); R3 = H or Me) the synthesis of which is shown in Scheme 3 above. Like those in the [Ti(NR)(Me4taa)] systems 1 and 2, the Ti=NR linkages form isolable cycloaddition products only with activated substrates such as C02 and isocyanates as discussed below.47 Reactions of 3 or 4"(R3 = Me) with sulfur-containing substrates CS2, COS or RNCS yield sulfide-bridged dimers [Cp*2Ti2{ R2C(NR1)2}2(p-S)2] as the only isolable organometallic products, although short-lived intermediate cycloaddition products can be inferred from NMR tube scale reactions of the arylimides 4.47b There is a clear difference in the stability and reactivity of the tert-butylimido compounds [(q-CgR34Me)Ti{ R2C(NR1)2}(NfBu)] (3) and that of the arylimido homologues 4. This chemistry is summarised in Scheme 5 for the reactions of 3 and 4 with C02.47a Reaction of either [(q-CgR34Me)Ti{R2C(NR1)2}(NtBu)] (3) or the aryl analogue [(q-C5R34Me)Ti{R2C(NR1)2}(NAr)] (4) with C02 rapidly forms the corresponding N,O-bound carbamate complexes [(q-C5R34Me)Ti{ R2C(NR1)2}{N(R)C(O)O}]. To this extent the C02 reactions parallel those of the macrocycle-supported imides [Ti(NR)(Meqtaa)] (Scheme 4). The X-ray structure of the homologous complex [(qCsMeg)Ti( Me3SiNC(Ph)NCH2CH2NMe2}{ N(tBu)C(0)O}] has been determined.47b In contrast to the macrocyclic systems, however, the aryl carbamates [(qC5R34Me)Ti{R2C(NR*)2}{ N(Ar)C(O)O}] react smoothly with a second equivalent of C02 which inserts into the Ti-NAr bond forming the novel six-membered ring complexes [(q-C5R34Me)Ti{ R2C(NR1)2}{OC(0)N(Ar)C(O)O}]which have also been crystallographically characterised. The formation of these compounds from the overall reaction of 4 with two equivalents of C02 was hitherto unknown in transition metal imido chemistry. Further studies of this reaction with the para-substituted arylimido complexes [(q-CgMes)Ti{ MeC(NiPr)2}(N-4-CgH4R)] for R = NMe2, Me or CF3 have shown a small enhancement in the rate of the second C02 insertion for the more electronreleasing NMe2 substituent. The tert-butylimido carbamates [(q-C5R34Me)Ti{ R2C(NR1)2}{N(tBu)C(0)O}] do not react further with C02 and in fact are unstable (either in the dark or light) to a

37

New titanium imido chemistry with polydentate N-donor ligands

cycloreversion reaction forming tBuNCO and the 0x0-bridged dimers [(q CgR34Me)2Ti2 { R2C(NR1)2}2(p-O)2] over several hours. The stability of these carbamates towards 0x0 dimer formation increases with increasing steric demands of the cyclopentadienyl or amidinate ligands. This in turn implies that the rate of self-trapping by dimerisation of the supposed transient terminal 0x0 intermediate [(q C5R34Me)Ti{R2C(NR1)2}(O)] (not observed) is an important factor. Me

Me R3+R3

co2

$ _---

I

-T-- - - - -\ ";I /\/-o A N

R2

I

\

R

R'

1

-

'BUNCO

1

12

Scheme 5 Reactions of [(77-CgR3qMe)TiIR2C(NRI)~(NR)](R = tBu 3 or alyl4; (RI, R2) = (SiMe3, Ph) or t P r , Me); R3 = H o r Me) with CO2d7

38

Perspectives in Organometallic Chemistry

The reactions of the tert-butlyimides [ ( ~ - C S R ~ ~ M ~ ) T ~ ( R)(NtBu)] ~ C ( N R(3)~ ) ~ with isocyanates RNCO (R = tBu or aryl) are slow, taking 4 to 10 days to go to complete consumption of 3. As for the C02 reactions above, the ultimate organometallic products isolated are the 0x0 dimers, with the organic products being the corresponding carbodiimides tBuNCNR. In the reactions of the 3 with arylisocyanates, NMR evidence has been obtained only for the supposed N,O-bound ureate intermediate [Cp*Ti{PhC(NSiMe3)2} {N(tBu)C(N-2,6,-C6H3Me2)O}]. This N,O-coordination is mechanistically necessary for the final extrusion of carbodiimide products and has been crystallographically characterised in the reactions of certain arylimides 4 with aryl isocyanates as discussed below. Me

'Pr Ar'NCO (Ar = Ar' = Tol)

I

C02 (Ar = Ar' = Tol)

Me Me

a Me

Me M -e

Scheme 6 Reactions of [Cp*Ti{MeC(NiPr)2,' (NAr)] with aryl isocyanates Ar 'NCO (Ar or A r ' = 4-Ca4Me (Too or 2,6-C&~Me2)47b

New titanium imido chemistry with polydentate N-donor ligands

39

The {Cp*-MeC(NiPr)2} ligand set provided the best environment in the general class of arylimides 4 for the study of cycloaddition and insertion reactions of these compounds with isocyanates. The results for the reaction of [Cp*Ti{MeC(NiPr)2}(NAr)] with aryl isocyanates Ar’NCO (Ar or Ar’ = 4-C&Me o 2,6-C6H3Me2) are summarised in Scheme 6 below. Addition of one equivalent of Ar’NCO to [Cp*Ti{ MeC(NiPr)2}(NAr)] proceeds smoothly within hours to form quantitatively the crystallographically characterised N,O-bound ureates [Cp*Ti{MeC(NiPr)2}{N(Ar)C(NAr’)O}],in which the aryl group of the exocyclic iminc group comes exclusively from the isocyanate. Unlike the very slow isocyanate cycloaddition products of the tert-butylimides 3, the cycloaddition products in Scheme 6 are stable to cycloreversion reactions. Addition of a further equivalent of TolNCO to the cycloaddition product [Cp*Ti { MeC(NiPr)a} {N(Tol)C(NTol)O}] yields a C,-symmetric product [Cp*Ti{ MeC(NiPr)2}{ OC(NTol)N(Tol)C(NTol)O}] which is the product of TolNCO insertion in the Ti-NTol bond of the N,O-bound ureate. This parallels the behaviour of the arylimide-derived carbamates [(r)-CgR34Me)Ti{R2C(NR1)2}{ N(Ar)C(O)O}] summarised in Scheme 5 . As with the C02 reactions, the second isocyanate inserts exclusively into the Ti-N bond of the first-formed cycloaddition product. Interestingly treatment of [Cp*Ti{MeC(NiPr)2} {N(Tol)C(NTol)O}] with C02 instead of a second TolNCO affords exclusively [Cp*Ti{MeC(NiPr)2} { OC(O)N(Tol)C(NTol)O}], again with complete cleavage of the original Ti-Nimid, bond. 3.3 Diamide-pyridine (N2Npy) supported titanium imido complexes Reactions of [Ti(NtBu)(N2N py)(py)] ( 5 ) and its pyridine-free analogue [Ti(NfBu)(N2Npy)] show a wider variety than those of the macrocyclic or cyclopentadienyl-amidinatesystems 1-4 so far considered. Aspects of the chemistry of diamide ligand supported imido complexes have recently been reviewed.49 We first of all consider the reactions of 5 with OCN-2,6-CgH3iPr2,50 tBuCP and MeCN5 as summarised in Scheme 7. Compared to the reactions of the macrocyclic or cyclopentadienyl-amidinate systems 1-4 with heterocumulenes, those of [Ti(NtBu)(N2Npy)(py)] (5) are extremely fast.50 Indeed, apart from the reaction of the bulky OCN-2,6-CgH-jiPr2 (which must be carried out at low temperature), no other isocyanate examined, or indeed C02, afforded a tractable product. The ureate formed in the reaction of 5 with OCN-2,6-C6H3iPr2 (Scheme 7) features an N,O-bound coordination mode, presumably because of the sterically demanding nature of the aryl group. However, the pyridyl group of the N2Npy ligand itself remains bound to titanium in the six-coordinate product. This is not the case in the reactions of 5 with fBuCP and MeCN51 which form crystallographically characterised mono- and bi-nuclear products of cycloaddition, respectively, in which the pyridyl is detached from the metal centre. The reaction of 5 with tBuCP, carried out as part of the first study of reactions of this kind for Group 4 imides, gives a nitrogen-bound-to-phosphorus metallacycle in the The K2-coordinationmode clearly helps relieve product [T~(K~-N~N~~){N(C)PC(C)}]. the steric strain induced by the presence of two tert-butyl and trimethylsilyl groups in the immediate vicinity of titanium. The nitrogen to phosphorus addition in this product

40

Perspectives in Organometallic Chemistry

is favourable, and it also avoids the soft phosphorus atom having to coordinate to the hard titanium (+4) centre; also avoids placing two tert-butyl groups (that from the imide and that from the phospha-alkyne) adjacent to each other.

A,

mbar

- MeGN

5

Scheme 7 Reactions of ~i(WBu)(N2Npy)@y)](5) with ArNC0,jO tBuCP and MeCNjl (Ar = 2,6-CgHjiPr2) The reaction with MeCN gives a dimeric product [Ti2(~2-N2N,,)2{ pNC(Me)NC 121. This compound is formally the nitrogen-bridged dimer of the cycloaddition product of 5 in which the imido nitrogen now bonds to carbon of MeCN (avoiding an unfavourable nitrogen-nitrogen single bond in the other possible first-formed product). The nitrogen originating from MeCN bonds to titanium and also, presumably

New titanium imido chemistry with polydentate N-donor ligands

41

because it is quite sterically unprotected and nucleophilic, bridges to a second titanium. Remarkably, attempted high vacuum and high temperature sublimation of the pure dimeric compound quantitatively yields liberated MeCN and [Ti(NtBu)(N2Npy)]which can also be obtained independently by high vacuum sublimation of S'itself. This was the first cycloaddition of any transition metal imide with a nitrile, and does not occur for nitriles bulkier than MeCN, possibly for steric reasons.

R.HC=C=CH2 or MeCzCR 40-80 'C, 10-d R = Me or Ph 5

Bu

,C-R

1

H

Eqn. 4 summarises the reaction of [Ti(NtBu)(N2Npy)(py)](5) with the methyl- and phenyl-allene, and with but-2-yne and 3-phenyl-prop-2-yne, which give the same products. The compounds formed in Eqn. 4,namely [Ti(N2Npy){N(C)C(CHR)CH2)], are the expected ones for the straightforward addition of these allenes to the Ti=NfBu linkage of 5, and indeed this was the first reported example of the addition of an allene to any transition metal imide. However, they are evidently not the metalla-azacyclobutene products expected from the reaction of internal alkynes with a Group 4 imide.11313 The mechanism of the formation of the compounds [Ti(N2Npy){N(C)C(CHR)CH2)] from alkynes is not known50 but clearly must involve a hydrogen atom migration at some stage of the mechanism. The final set of reactions to consider are those of 5"with isocycanides RNC (Scheme 8).52 No reaction is observed with fBuNC, but with ArNC (Ar = 2,6,C6H3Me2) the compound [Ti(N2Npy){N(C)C(NAr)C(NAr))] is formed with the consumption of two equivalents of ArNC per titanium. It was not possible to identify the expected 1 :1 molar ratio reaction intermediate [Ti(N2Npy){q2-N(fBu)C(NfBu)}] (formally an $-bound di-tert-butyl carbodiimide compound) although Bergman has seen a species similar to this in the reaction of transient [Cp*2Zr(NfBu)]with f B ~ N c . 2 3The ~ compound [Ti(N2Npy){N(C)C(NAr)C(NAr)>]does not react hrther with ArNC, but does add EtNC (and a range of other comparatively small isocyanides), giving the ringexpanded product shown in Scheme 8. With isocyanides RR'CHNC (R,R' = alkyl or H) that have at least on H atom a tc the cyan0 group a completely different reaction sequence takes place forming the 1:3 molar ratio coupled product shown in Scheme 8 (these are dihydropyrimidine derivatives). The a-H atom of one of the RR'CHNC substrates has been activated. If the reaction is carried out between 5 and less than three equivalents of RR'CHNC then the same product is obtained, but in lower yield, and unreacted 5 remains.

Perspectives in Organometallic Chemistry

42

RR'CHNC Me&

R,R = alkyl or H

p'y

Me3Si'

5

1

*,-

Bu

\NCHlW

ArNC

R"NC b

R" = n-alkyl, CH,Ph, To1 Scheme 8 Reactions of [Ti(NtBu)(N2Npy)(py)] (5) with ArNC (Ar =2,6-C6H3Me2) and RR'CHNC (R,R' = alkyl or H)52 The mechanism summarised in Scheme 9 has been proposed to account for the formation of the dihydropyrimidine derivatives with RR'CHNC and the observation of the a-H atom activation, and the other products formed when the reaction is carried out with two ArNC, or with two ArNC followed by a smaller, second type of isocyanide. Thus the proposed q2-carbodiimide intermediate A (not observed as mentioned above) inserts a second equivalent of isocyanide to form B. If the isocyanides coupled so far are large and do not possess a-H atoms then the reaction stops here. If an equivalent of a (small) isocyanide is added then compounds of the type C (see Scheme 8) can be isolated. But if the first isocyanide added to 5 possesses an H atom a to the cyan0 group then it is positioned ideally for a sigmatropic rearrangement forming D which then ring closes to form the product observed with RR'CHNC substrates.

New titanium imido chemistry with polydentate N-donor ligands

43

IBLt

5

El

1

RR'cHNc

,Ti-

H-shift

CHRR

Scheme 9 Proposed mechanism for the reaction of [Ti(lvtBu)(N2Np,,)(py)] (5) with isocyanides52 4 CONCLUSION A facile route to a range of reactive titanium imido complexes is available starting fiom

compounds of the type [Ti(NR)C12(py)J. The judicious choice of supporting ligand set enhances or otherwise modifies reactions of the Ti=NR linkage in complexes with

Perspectives in Organornetallic Chemistry

44

macrocyclic, cyclopentadienyl-amidinateor diamide-pyridine supporting ligands. The choice of imido N-substituent can be critical, especially between tert-butyl and aryl groups. The Ti=NR group in the compounds discussed in detail here couples with a number of substrates: carbon dioxide, carbodiimides, isocyanates, isocyanides, acetonitrile, tert-butyl phospha-alkyne, allenes and alkynes. In some instances sequential, multiple coupling reactions are facile. The outlook for M h e r reactions and new applications of titanium and other Group 4 imido compounds continues to be very promising. References and footnotes 1

2 3 4 5 6 7 8 9 10 11 12

13

14 15 16 17 18

Although for ease of representation all titanium imido linkages are drawn "Ti=NR", the formal titanium-imido nitrogen bond order in the complexes described herein is probably best thought of as closer to three (pseudo-o* ,$; triple bond) rather than as two: T. R. Cundari, Chem. Rev., 2000, 100, 807; N. Kaltsoyannis and P. Mountford, J. Chem. Soc., Dalton Trans., 1999, 78 1, and references therein. P. Mountford, Chem. Commun., 1997, 2 127 (Feature Article). D. E. Wigley, Prog. Inorg. Chem., 1994,42,239. H. Burger and U. Wannagat, Montasch., 1963,94,761. W. Schick, R. Bettenhausen, and W. Milius, Chem. Eur. J., 1997,3,1337. J. D. Gardner, D. A. Robson, L. H. Rees and P. Mountford, Inorg. Chem., 2001, 40, 820. J. D. Gardner, Part 1 1Thesis, University of Oxford, 2000. D. C. Bradley and E. G. Torrible, Can. J. Chem., 1963,41, 134. W. A. Nugent and R. L. Harlow, Inorg. Chem., 1979,18,2030. C. T. Vroegop, J. H. Teuben, F. van Bolhuis and J. G. M. van der Linden, J. Chem. SOC.,Chem. Commun., 1983,550. P. J. Walsh, F. J. Hollander and R. G. Bergman, J. Am. Chem. Soc., 1988, 110, 8729. C. C. Cummins, S. M. Baxter and P. T. Wolczanski, J. Am. Chem. SOC.,1988,110, 8731. (a) C. C. Cummins, C. P. Schaller, G. D. Van Duyne, P. T. Wolczanski, A. W. E. Chan and R. Hoffmann, J. Am. Chem. SOC.,1991,113,2985; (b) later reports on a related trigonal plane bis(si1oxide)-imido system: J. L. Bennett and P. T. Wolczanski, J. Am. Chem. SOC.,1994,116,2179. (a) P. L. McGrane, M. Jensen and T. Livinghouse, J. Am. Chem. SOC.,1992, 114, 5459; (b) D. Fairfax, M. Stein, T. Livinghouse and M. Jensen, Organometallics, 1997,16,1523. J. E. Hill, R. D. Profilet, P. E. Fanwick and I. P. Rothwell, Angew. Chem. Int. Ed. Engl., 1990,29,664. H. W. Roesky, H. Voelker, M. Witt and M. Noltemeyer, Angew. Chem. Int. Ed. Engl., 1990,29,669. T. R. Cundari,J. Am. Chem. SOC.,1992,114,10557. (a) Y. Bai, M. Noltemeyer and H. W. Roesky, Z. Naturforsch., 1991,46b, 1357; (b) Y. Bai, H. W. Roesky, M. Noltemeyer and M. Witt, Chem. Ber., 1992,125,825.

New titanium imido chemistry with polydentate N-donor ligands

19 20 21 22

23

24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

45

D. J. Arney, M. A. Bruck, S. R. Huber and D. E. Wigley, Inorg. Chem., 1992, 31, 3749. (a) J. S. Johnson and R. G. Bergman, J Am. Chem. Sac., 2001, 123,2923; (b) E. Haak, I. Bytschkov and S. Doye, Angew. Chem. Int. Ed., 1999,38, 3389; F. Pohlki and S. Doye, ibid. 2001,40,2305. (a) S . C. Dunn, A. S. Batsanov and P. Mountford, J Chem. SOC.,Chem. Commun., 1994,2007; (b) S . C. Dunn, P. Mountford and D. A. Robson, J. Chem. SOC.,Dalton Trans., 1997,293. J. L. Polse, M. R. Smith, R. A. Andersen and R. G. Bergman, Abstr. Pap. Am. Chem. SOC.,1994,208,592-INOR; J. L. Polse, R. A. Andersen and R. G. Bergman, J.Am. Chem. Soc., 1998,120,13405. Selected examples: (a) P. J. Walsh, F. J. Hollander and R. G. Bergman, Organometallics, 1993,12, 3705; (b) K. E. Meyer, P. J. Walsh and R. G. Bergman, J. Am. Chem. SOC.,1994, 116, 2669; (c) R. L. Zuckerman and R. G. Bergman, Organometallics,2000,19,4795; (d) Z . K. Sweeney, J. L. Salsman, R. A. Andersen and R. G. Bergman, Angew. Chem. Int. Ed., 2000,2339. C . J. Harlan, J. A. Tunge, B. M. Bridgewater and J. R. Norton, Organometallics, 2000,19,2365. F. G. N. Cloke, P. B. Hitchcock, J. F. Nixon, W. J. D. and P. Mountford, Chem. Commun., 1999,661. L. M. Berreau, V. G. Young and L. K. Woo, Inorg. Chem., 1995, 34, 527; J. L. Thorman and L. K. Woo, Inorg. Chem., 2000,39,1301; J. L. Thorman, I. A. Guzei, V. G. Young and L. K. Woo, Inorg. Chem., 2000,39,2344. J. R. Hagadorn and J. Arnold, Organometallics, 1994,13,4670;J. R. Hagadom and J. Arnold, Organometallics, 1998,17, 1355. U. Radius and A. Friedrich, 2. Anorg. Allg. Chem., 1999,625,2154. M. J. Scott and S. J. Lippard, Organometallics, 1997, 16, 5857; D. P. Steinhuebel and S. J. Lippard, Inorg. Chem., 1999,38,6225. M. D. Fryzuk, J. B. Love and S. J. Rettig, Organometallics, 1998, 17, 846. P. J. McKarns, G. P. A. Yap, A. L. Rheingold and C. H. Winter, Inorg. Chem., 1996,35,5968. (a) C. H. Winter, P. H. Sheridan, T. S. Lewkebandara, M. J. Heeg and J. W. Proscia, J Am. Chem. SOC.,1992,114, 1095; (b) T. S. Lewkebandra, P. H. Sheridan, M. J. Heeg, A. L. Rheingold and C. H. Winter, Inorg. Chem., 1994,33,5879. S . C. Dunn, P. Mountford and 0. V. Shishkin, Inorg. Chem., 1996,35, 1006. D. Swallow, J. M. McInnes and P. Mountford, J. Chem. SOC.,Dalton Trans., 1998, 2253. A. J. Blake, P. E. Collier, S. C. Dunn, W.-S. Li, P. Mountford and 0. V. Shishkin, J. Chem. SOC.,Dalton Trans., 1997, 1549. D. S. Glueck, J. Wu, F. J. Hollander and R. G. Bergman, J Am. Chem. SOC.,1991, 113,2041. N. Adams, A. R Cowley, S. R Dubberley, A. J. Sealey, M. E. G. Skinner and P. Mountford, Chem. Commun.,2001,2738. P. E. Collier, A. J. Blake and P. Mountford, J. Chem. SOC.,Dalton Trans., 1997, 291 1. J. M. McInnes, D. Swallow, A. J. Blake and P. Mountford, Inorg. Chem., 1998, 37,5970.

46

Perspectives in Organometallic Chemistry

40 41

S. R. Dubberley, D.Phil. Thesis., University of Oxford, 2000. A. J. Blake, S . C. Dunn, J. C. Green, N. M. Jones, A. G. Moody and P. Mountford, Chem. Commun., 1998, 1235. P. J. Wilson, A. J. Blake, P. Mountford and M Schr der, Chem. Commun., 1998, 1007; P. J. Wilson, A. J. Blake, P. Mountford and M. Schr der, J. Organomet. Chem., 2000,600,71. (a) P. J. Stewart, A. J. Blake and P. Mountford, Inorg. Chem., 1997,36, 3616; (b) P. J. Stewart, A. J. Blake and P. Mountford, Organometallics, 1998, 17, 3271; (c) P. J. Stewart, A. J. Blake and P. Mountford, J. Organomet. Chem., 1998,564,209. S . A. Lawrence, M. E. G. Skinner, J. C. Green and P. Mountford, Chem. Commun., 2001,705. (a) A. J. Blake, P. E. Collier, L. H. Gade, J. Lloyd, P. Mountford, S. M. Pugh, M. Schubart, M. E. G. Skinner and D. J. M. Tr sch, Inorg. Chem., 2001,40,870; (b) S . M. Pugh, H. S. C. Clark, J. B. Love, A. J. Blake, F. G. N. Cloke, and P. Mountford., Inorg. Chem., 2000,39,2001; (c) M. E. G. Skinner, D. A. Cowhig and P. Mountford, Chem. Commun., 2000, 1167; (d) J. Lloyd, S. Z. Vatsadze, D. A. Robson, A. J. Blake and P. Mountford, J. Organomet. Chem., 1999,591, 114. C. E. Housemekerides, D. L. Ramage, C. M. Kretz, J. T. Shontz, R. S. Pilato, G. L. Geoffroy, A. L. Rheingold and B. S . Haggerty, Inorg. Chem., 1992,31,4453. (a) A. E. Guiducci, A. R. Cowley, M. E. G. Skinner and P. Mountford, J. Chem. SOC.,Dalton Trans., 2001, 1392; (b) A. E. Guiducci, C. L. Boyd and P. Mountford, Abstr. Pap. Am. Chem. SOC.,2001,222,245-NOR. A. J. Blake, J. M. McInnes, P. Mountford, G. I. Nikonov, D. Swallow and D. J. Watkin, J. Chem. SOC.,Dalton Trans., 1999, 379. L. H. Gade and P. Mountford, Coord. Chem. Rev., 2001,216-217, 65. D. J. M. Tr sch, P. E. Collier, A. Bashall, L. H. Gade, M. McPartlin, P. Mountford and S . Radojevic, Organometallics,2001,20, 3308. S . M. Pugh, D. J. M. Tr sh, D. J. Wilson, A. Bashall, F. G. N. Cloke, L. H. Gade, P. B. Hitchcock, M. McPartlin, J. F. Nixon, and P. Mountford, Organometallics, 2000,19,3205. A. Bashall, L. H. Gade, M. McPartlin, P. Mountford, S. M. Pugh, S. Radojevic, M. Schubart, I. J. Scowen and D. J. M. Tr sch. Oraanometallics. 2000.19.4784.

42

43

44 45

46 47

48 49 50 51

52

ORGANOMETALLIC COMPLEXES WITH 1,2-DICHALCOGENOLATE-OCARBORANES

Guo-Xin Jin Chemistry Department, Fudan University, Shanghai 200433, China

1 INTRODUCTION

Mononuclear Cp*M half-sandwich complexes are useful model complexes in which one hemisphere of the coordination shell is blocked by a 6-electron substituted cyclopentadienyl ring. In the protected space below the substituted cyclopentadien 1 ligands, chalcogenolato ligands can be accommodated to form a variety of complexes.l 2 1,2-Dicarba-closo-dodecaborane 1,2-dichalcogenolates (ortho-carborane dichalcogenolates) can be used as the voluminous and chemically robust chelate ligands in this regard. Dilithium dichalcogenolate carboranes L ~ ~ E ~ C ~ B (E~ =O S, H ~Se,O Te) can be e a d y prepared by the insertion of elemental chalcogen into the Li-C bonds of dilithium carborane which was produced by the reaction of carborane with butyl lithium (Scheme l).3d4Due to their rigid backbone, dichalcogenolate carborane ligands have allowed the synthesis of complexes that exhibited novel structure^.^

1

Scheme 1

The molecular structure of L ~ ~ S ~ C ~ in B ~theO solid H ~ Ostate is dimeric One lithium atom of dilithium dithiolate carborane links three sulfur atoms to form a dinuclear structure (Figiure 1).

Perspectives in Organometallic Chemistry

48

Figure 1 Structure of L~zS~C~BIOHIO 2 SYNTHESIS OF COMPLEXES

2.1

Titanium, Zirconium and Hafnium Complexes

Metallocene complexes Cp'MC12 (Cp'= q5-C5H5, qSJBuCs&, q5-1,3-tBu2CsH3; M=Ti, Zr, Hf) react with dilithium dichalcogenolate carborane L ~ ~ E ~ C ~ (E=S, B ~ OSe) H to ~O produce mononuclear half-sandwich complexes [Li(THF)4][CpfM(E2C2B10H10)2] (2) (Scheme 2).6

2 Cp'= q5C5H5, q5-'BuC&, q5-'Bu2C5H,; M =Ti, Zr Hf;E = S,Se

Scheme 2

Analogous complexes [L~(THF)~][C~*M(E~C~B~OH~O)~] (Cp*=q5-CsHs;M=Ti, Zr, Hf; E=S, Se) (3) can be directly obtained from the reaction of Cp*MC13 (M=Ti, Zr, Hf) with L ~ ~ E ~ C ~ B(E=S, I O HSe). ~ O The molecular structures of 2 (Cp'= q5-C5H5, qs2BuCs&, q51,3-'Bu2C~H3;M=Ti; E=S) and 3 (M=Zr; E=S, Se) have been determined by x-ray crystallographic studies6

r

9

x

1

3

49

Organometallic complexes with 1,2-dichalcogenolate-o-carbonates

Molybdenum and Tungsten Complexes

2.2

The reaction between Cp*MoC14 and L ~ ~ E ~ C ~ in B~ THF O Hsolution ~ O leads to the salts of [Li(THF)4][Cp*Mo( E2C2B10H&] (E=S, Se) (4), in the which molybdenum has been reduced from Mo(V) to Mo(IV) by the chalcogenolate ligands (Scheme 3).'

1

H-NMR spectra of both complexes (4) exhibited intense signals [6=1.85ppm (E=S) and 1.88ppm (E=Se)] for the methyl substituents of Cp* ring indicating the diamagnetic nature of the complexes. Cp*WC14 reacts with L ~ ~ E ~ C ~ (E=S, B ~ OSe) H ~under O the same reaction conditions to affclrd neutral symmetrical bis(dichalcogeno1ate carborane) tungsten complexes Cp4'W(E2C2B10H10)2(E=S, Se) (5) in good yields. These complexes (5) can be easily converted to corresponding terminal 0x0 com lexes Cp*W(O)(E2CzBloHlo) (6) in THF solution in the presence of water (Scheme 4)..pBoth complexes shown, (5) and (6), are paramagnetic as indicated by their NMR spectra. FP*

5

6

Scheme 4 The X-ray crystal structures of Cp*W ( E ~ C ~ B ~ O H (E=S, & Se) (5) (Figure 2) show that W(V) is in the center of a tetragonal pyramid and that they are 16-electron complexes. The X-ray crystal structure of Cp*W(O)(Se2C2BloHlo) (6) (Figure 3) indicates that the geometry of complexes can be described as a having distorted tetrahedral coordination sphere. The W=O distance is 1.754A. Complexes (5) (E=S, Se) which can be viewed as organometallic ligands react with [Cu(MeCN)4]PFb or AgBr in MeCN solution to form the hetero-tetranuclear complexes [Cp*W(E2C2BloH&Cu(CNMe)2]2 (E=S, Se) (7) and hetero-trinuclear complexes [Cp*W ( E ~ C ~ B ~ O H ~(E=S, O ) ~Se) ] ~(8) A ~respectively (Scheme 5 ) .

'

Perspectives in Organometallic Chemistry

50

Figure 2

Structure of Cp*W(Se2C2Bld10)2(5)

Figure 3

Structure of Cp*W(O)(S~~C~B~OH~O) (6)

5

Scheme 5

Organometallic complexes with 1,2-dichalcogenolate-o-carbonates

51

The molecular structure of [Cp*W ( S ~ C ~ B ~ O H~~)~CU(CNM~)~]~ (7) is shown in Figure 4. One Cu atom as a bridge is connected with 3 sulfur atoms of the dithiocarborane ligands and an acetonitrile in a distorted tetrahedral coordination sphere. Another Cu atom is coordinated with two sulfur atoms and an acetonitrile in a trigonal geometry. The molecule of complex is chiral. The molecular structure of [ C ~ * W ( S ~ C ~ B ~ O H (8)~ (Figure O ) ~ ] ~5A) ~ indicates that Cp*W(E2C2BloHlo)2units are linked through a Ag atom and Ag as a bridging atolm is connected via four sulfur atoms.

3

Figure 4

Structure of [Cp* W(S~C~B~OH~~)~CU(CNM~)~]~ (7)

H

Figure 5

Structure of [Cp*W(S2CzBiaHio)2/2Ag(8)

Interestingly, reaction of the binuclear complex [CpMo(NO)I2] with H2S2C2B 10Hloin the presence of Et3N leads to an 18-electron mononuclear complex wEt3H] [CpMo(NO)(S ~ C ~ B ~ O H O ) Imolecular ] .8 ~The structure of the complex indicated that the geometry at molybdenum approximates to square pyramidal with the centroid of the cyclopentadienyl ring axial. In contrast to cyclopentadienyl complexes, the mononuclear coordinatively unsaturated complex Tp*Mo(O)C12 (Tp* = Hydrotris(3,5-dimethyl pyrazolyl) borate) reacts with L ~ ~ E ~ C ~ (E=S, B ~ OSe H and ~ OTe) in THF solution to produce mononuclear five coordination, formally 13-electron complexes T~*MO(O)(E~C~BIOHIO) (E=S, Se, Te). The

52

Perspectives in Organometallic Chemistry

molecular structure of Tp*Mo(O)(Se2C2B1oHlo) reveals that the bonding distance between molybdenum and oxygen is shorter than the normal Mo=O double bond and that the interaction of Mo and one N located in the opposite of Mo=O bond is absent. 2.3

Rhenium Complexes

The reaction of 1,2-dilithiumdichalcogenolatecarborane L ~ ~ E ~ C ~ (E=S, B ~ OSe) H with ~ O the half-sandwich pentamethylcyclopentadienyl rhenium complex, Cp*ReC14, in THF solution leads to mononuclear salts [L~(THF)~][C~*R~(E~C~B~OH~O)~] (E=S, Se)(9)7 in which the rhenium atom has been reduced from ReV to Re". The violet salts are shown to have paramagnetic properties according to their 'H and 13C-NMR spectra. The salts [Li(THF)s][Cp*Re(E2CzB10H10)2] can be easily oxidized by AgBF4 to form green neutral products, Cp*Re(E2C2BloHlo)2 (E=S, Se) (lo), in very high yields (Scheme 6). The 'H and 3C-NMR spectra of the complexes display a strong signal for the methyl of Cp*.

9

10

Scheme 6 2.4

Iron, Ruthenium and Osmium Complexes

Half-sandwich ruthenium nitrosyl dichalcogenolate carborane complexes Cp*Ru(NO)(E2C2BloHlo) (E = S, Se, Te) can be prepared by the reactions of Cp*Ru(NO)C12 with L ~ ~ E ~ C ~ B H ~Se,O Te)'. These mononuclear complexes (E~=O S, contain a Ru" atom. The 16-electron half-sandwich q5-(4-isopropyl toluene) Ru" and 0s" complexes (p-cymene)M(S2C2BloHlo) (M = Ru, 0s) have been also synthesized. These coordinatively unsaturated mononuclear complexes can convert into the coordinative saturated 18-electron derivatives in the presence of an electron donor ligands, such as: triphenyl phosyhane (Ph3P), 'Butyl isonitrile ('BuNC) and carbon monoxide (CO).l o

Scheme 7 Attempts to combine the [ E ~ C ~ B I O (E H ~=OS,] Se) building block with two 17 electron fiagments such as [CpFe(CO)2] and [CpRu(PPh3)] were unsuccessful, although gold(1) complexes of the type [ S ~ C ~ B ~ O H ~ ~ ] [ A have U ( P Rbeen ~ ) ] ~obtained' and mercury(I1)-

Organometallic complexes with 1,2-dichalcogenolate-o-carbonates

53

bridged oligocarboranes are known.12 Instead, CpFe(C0)2Cl reacts with L ~ ~ S ~ ~ C ~ B inI O H I O a molar ratio of 2:l to give black prismatic crystals of C ~ ~ F ~ ~ ( C O ) ~ ( ~ - S ~(13) ~ C in ~BIOH~ high yield.’ The carborane-l,2-diselenolate bridge combines a [CpFe(CO)] and a [CpFe(CO)z] fragment (Figure 6); one iron atom carries only one carbonyl group in addition to a formal 3-electron chelate ligand [(B~oHIo)C~S~(S~-R)](R = CpFe(C0)z). 02

cu

Figure 6

Structure of C~~F~~(CO)~(~-S~~C~BI&IIO) (13)

The comparable reaction of CpRu(PPh3)2C1 with thedilithium dithiolate reagent L ~ ~ S ~ C ~ B did ~ O Hnot I Ogive a product analogous to corresponding iron complex C P ~ F ~ ~ ( C O ) ~ ( ~ - Sbut ~ ~instead C ~ B ~the O Hsymmetrical ~O) dinuclear complex CpzRu2(pS ~ C ~ B I O H(14) I O )with ~ two o-carborane dithiolate bridges (Figure 7). Each CpRu fragment is attached to one terminal and two bridging sulfur ligands. The Ru-Ru distance of 2.778181 corresponds to a single bond.

’

Figure 7

Structure of C ~ ~ R U ~ ( ~ - S ~ C(14) ~BIOHIO)~

With the chloro-bridged dimer [Cp*RuCl(p-C1)]2 as starting material, the reaction with dilithium diselenolate carborane Li&C2B10H10 in THF solution led to a black crystalline complex Cp*2Ru2(p-Se)(p-Se2C2B10Hlo) (15). In this complex, each Ru center is connected to three selenium atoms which are all bridging (Figure 8). The Ru-Ru single bond distance (2.7877 A) may therefore be compared to the corresponding distance in a dinuclear complex such as [Cp*2Ru2(p-SeC6&-Me(4))3]Cl (2.685(3) A).’ The remarkable formation of a monoselenide bridge can be ascribed to steric congestion caused by the bulky Cp* rings.

Perspectives in Organometallic Chemistry

54

Figure 8 2.5

Structure of CP*~RU~(~-S~)(~-S~~C~BIOHIO) (IS)

Cobalt, Rhodium and Iridium Complexes

The dilithium carborane dichalcogenolates L ~ ~ E ~ C ~ B(E=S, ~ O HSe) I Oreact with the halfsandwich iridium chloro-bridged dimer [Cp*IrCl( -C1)]2 to give green 16-electron CpIr complexes, Cp*Ir(E2C2BloHlo) (E=S, Se) (16).13-' These complexes can take up twoelectron ligands L, such as L = CO, CN'BU, pyridine, phosphanes, to produce the coordinatively saturated 18-electron complexes Cp*Ir(L)(E2C2BloH10) (E=S, Se; L = CO, CN'BU, Pyridine, PMe3, PPh3) (17) (Scheme 8).13

Scheme 8

The molecular structures of Cp*Ir(Se2C2BloHlo) (16) and Cp*Ir(PMe3)(Se2C2BloHlo) (17) (Figure 9 and 10) were determined by X-ray crystal structure analysis. The molecular structure of Cp*Ir(PMe3)(Se2C2BloHlo)corresponds to CzVpoint group symmetry with two perpendicular mirror planes. The short distances for the Ir-Se and C-Se bonds in the coordinatively unsaturated mononuclear complex are interpreted as being due to a x-type intera~tion.'~ Addition of PMe3 to form the 18-electron complex Cp*Ir(PMe3)(Se2C2BloHlo) leads to Cs symmetry with only one mirror plane. The iridadiselenolene heterocycle in Cp*Ir(Se2C2BloHlo) is bent in C~*I~(PM~~)(S~~C~B~O with a dihedral angle at the Se....Se vector of 156.1'. Apparently, the ligand L = PMe3 destroyed the pseudoaromatic IrSe2C2 system.I3

Organometallic complexes with I ,2-dichalcogenolate-o-carbonates

55

Y

Figure 9

Structure of Cp*Ir(Se2CZBloH10)(16)

Figure 10 Structure of Cp*Ir(PMe3)(Se2CzBloHlo)(1 I )

Analogous 16-electron half-sandwich cobalt and rhodium complexes, Cp "Co(S2C2B1oHlo) (Cp'= q5-Cp and llS-'Bu2C5H3)(18)16-17and Cp*Rh(E2C2BloHlo) (E=S, Se)(19)? have been also synthesized by the reactions of Cp'Co(CO)I2 with L ~ ~ S ~ C ~ B ~ O H I and [Cp*RhCl(p-C1)]2 with L ~ ~ E ~ C ~ (E B~ OH O = S, Se)~respectively. The corresponding 18electron complexes Cp'Co(L)(SzC2BloHlo) (20) and Cp*Rh(L)(E2C2BloHlo) (E=S, Se) (21) have been prepared by addition of two-electron ligands L (L = CO, CN'BU, PMe3).3 The dirhodium complexes with o-carborane dichalcogenide bridging ligands R ~ : ~ ( C O ) ~ ( ~ - E ~ C ~(EB = I OS,HSe) ~ O(22) ) have been recently prepared from the rhodium carbonyl complex, Rh2(CO)4(pL'C)2,with Li2E2C2B10H10(E = S, Se) in THF solution. The cornplexes were isolated as red crystals in good yields. 0-carborane dithiolate bridged two rhodium atoms (Figure 11). The Rh-Rh distance of 2.8900A corresponds to a single bond.

Figure 11 Structure of R ~ ~ ( C O ) ~ ( ~ - S ( Z22C )~ B ~ ~ ~ O )

56

Perspectives in Organometallic Chemistry

2.6 Lanthanide Complexes

Although some cyclopentadienyl transition metal complexes of the d-block elements with o-carborane dichalcogenolate ligands have been synthesized during the last four years, the series of related organolathanide dichalcogenolatecarborane complexes is underdeveloped. Reactions of [CpLn(p-C1)]2 (Ln = Nd, Yb, Dy, Gd, Er) with an equivalent of L ~ ~ E ~ C ~ B(E~ O =H S,~ Se) O in THF solution afforded the dinuclear o-carborane dichalcogenolate bridged complexes [Li(THF)4] [ C ~ L ~ ( E ~ C ~ B ~ (E=S, O H Se; ~ OLn ) ]= ~ Nd, Yb, Dy, Gd, Er) (23)(Scheme 9).18

E = S, Se; Ln = Nd, Yb, Dy,Gd, Er

Scheme 9

The complexes are salts and the molecular structures of [Li(THF)4] [CPL~(E~C~B~O O )Ln ] ~= Nd, Yb; E = Se, Ln = Yb) (23)have been determined (EH=~S, by X-ray crystal structure analysis (Figure 12). Each Cp2Ln fragment is connected to one terminal and two bridging chalcogen atoms of the o-carborane dichalcogenolate ligands. The central Ln2(p-E2) four membered ring is not planar. Direct metal-metal bonding interaction in these complexes is absent.

Figure 12 Structure of [ L ~ ( T H F ) ~ J [ C ~ ~ ~ ( S ~ ~ C(23) ~BIOHIO)]~

For the purposes of comparison, similar tert-butyl substituted cyclopentadienyl complexes [Li(THF)4] [Cp#Ln(E2C2B10H10)]2 (Cp' = qS-'BuCs&, q5-fBu2C5H3;E=S, Se; Ln = Nd, Yb, Dy, Gd, Er) (24) have also been investigated and structurally characterized in the present study.6

57

Organometallic complexes with 1,2-dichalcogenolate-o-carbonates

3 REACTIVITY STUDIES Although numerous organothiolate complexes of transition metals have been reported and in some cases their chemistry studied,” the chemistry of the complexes containing chelating 1,2-dicarba-closo-dodecaborane(1,2)-dichalcogenolate ligands has received only scant attention.

3.1 Coordination Reactions In the light of the nucleophilic behaviour of chalcogen ligands and easy formation of pchalcogen bridges in heterobimetallic complexes, it is not surprising that the neutral halfsandwich rhenium dichalcogenolate carborane complexes of the type [Cp*Re(E2C2BloHlo)2] (E = S, Se) (10) are able to act as organometallic bidentate ligands through their chalcogen atoms. [ C ~ * R ~ ( E ~ C ~ B ~react OH~ with O ) the ~ ] toluene-stabilised tricarbonyl molybdenum fragment (C7Hg)Mo(C0)3 to give the hetero-tetranuclear complexes [C~*R~(E~C~B~OH~O)~MO(C~)]~ (E = S, Se) (25) in which chalcogen atoms of o-carborane dichalcogenolate ligands assume the bridging function. However [Cp*Re(E2C+3loH10)2] reacts with (COD)2Ni to give [Cp*Re(S2C2B10H10)2Ni]2 (26) (Scheme 10).

26

I

25

Scheme 10 In the molecular structure of [C~*R~(S~C~BI~H~O)~MO(CO)]~ (25), the complex contains four o-carborane dithiolate chelating ligands. Two of these ligands are arranged so that S atom of each chelate ligands bridge Re and Mo atoms and another S atom bridges

58

Perspectives in Organometallic Chemistry

two Mo atoms (Figure 13). A carbonyl ligand is attached to each of the Mo atoms. The four-membered rings connecting the Re and Mo atoms through two S atoms of carborane dithiolate ligands are almost planar. The shorter distances between Re and Mo (2.5868(2)A) lie in the range for the triple bonds and thus each of Mo and Re centers in the complex is formally an 18-electron system. The analogous o-carborane diselenolate carborane complex [C~*R~(S~~C~B~OH~O)~MO(CO)]~ was also synthesized and characterized by X-ray structure analysis. A strong metal-metal interaction between Re and Mo atoms (2.6205(7) A) was observed.

Figure 13

Structure of [Cp *R~(S~C~BIOHIO)&~O(CO)]~ (25)

The molecular structure of the hetero-tetranuclear complex [ C ~ * R ~ ( S ~ C ~ B ~ O H & N ~ (26)was also determined by X-ray crystal structure analysis (Figure 14). The molecule of [CP*R~(S~C~B~OH (26) ~ Opossess ) ~ N ~ ]three ~ perpendicular mirror planes. The plane of Ni2(p-S)2 is in the middle and bridged via two S atoms of the carborane dithiolate ligands. The Re-Ni distances (2.4744(8) A) and the capping angles (Re-S-Ni: 66.52') in the another two planes indicate a strong direct interaction between Re and Ni atoms.

Figure 14

Structure of [Cp * R ~ ( S ~ C ~ B I O H I(26) O)~N~]~

59

Organometaliic complexes with I,2-dichaicogenolate-o-carbonates

3.2 Addition Reactions The mononuclear 16-electron dithio-o-carboranylcobalt complex, CPCO(S~C~B~OHIO), reacts with 1 equiv of CpCo(C2&)2 in toluene to give an addition product, ( C P C O ) ~ ( S ~ C ~ B ~(O O ) structure of ( C ~ C O ) ~ ( S ~ C ~ Bconsists 27H ).~The I O H ~ Oof ) binuclear (Cpco)~fragments in which each CpCo unit is attached to both sulfur atoms of a 1,2dithio-o-carboraneligand.l6 The addition reaction of C ~ C O ( S ~ C ~ B I Owith H I Oan ) alkyne forms alkyne adduct products CpCo(R1C = C R ~ ) ( S ~ C ~ B ~ (28), O H ~ while O ) the reaction of the cobaltadithio-ocarborane complex with trimethyl silyldiazomethylene in dichloromethane solution gives the trimethylsilyl methylene adduct (Scheme 11).l6

I

I

27

CO

)s

s’

I

‘c-c

@ \

\

a

28

29

Scheme 11 3.3

Carborane Substitution

In the 16-electron half-sandwich complexes, Cp*M(S2C2BloHlo) (M = Rh, Ir) (19,16) and (pcyrnene)M(S2C2BloHlo) (M = Ru, 0s) (ll),the metal centers, the metal s u l k bonds and the B(3, 6)-H bonds of o-carborane cage are reactive sites for interactions with unsaturated substrates?’ Reactions of Cp*Rh(S ~ C ~ B ~ Owith H I ~ methyl ) acetylene monocarboxylate or dirnethyl acetylene dicarboxylate form B(3,6)-substitution products of carborane cluster or acetylene additiodinsertion products (Scheme 12).2‘-23

Perspectives in Organometallic Chemistry

60

i @ii I

S / M \ S

\

Me02CCECCOZMe2

M = Rh

\

\

M = Rh, Ir

w

31

\ 32

Scheme 12

Acknowledgements I would like to thank my co-workers and collaborators on this project. I also wish to thank Professor K. Tatsumi and Professor M. Herberhold for their kind help. Our research was supported by The National Natural Science Foundation of China (29925 101) and by the Special Funds for Major State Basic Research Projects (G 1999064800).

References 1 2 3 4 5 6

L. C. Roof and J. W. Kolis, Chem. Rev., 1993,93, 1037 M. Herberhold, Guo-Xin Jin and W. Milius, Angew. Chem. Int. Ed. Engl., 1993,32,85. M. Herberhold, Guo-Xin Jin, H. Yan, W. Milius and B. Wrackmeyer, J Organomet. Chem., 1999,587,252. H. D. Smith, C. 0. Obenland and S. Papetti, Inorg. Chem., 1966,5,1013. S. Lu, Guo-Xin Jin, S. Eibl, M. Herberhold and Y. Xin, Organometallics, 2002,21, 2533. X. Y. Yu, Dissertation in Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, 2002,06.

Organometalliccomplexes with 1,Z-dichalcogenolate-o-carbonates

7 8 9 101

11 12,

13

14 15;

16 17 18 19 20 211 22 23

61

Guo-Xin Jin, K. Tatsumi and H. Kawaguchi, unpublished results. J. D. Mckinney, H. Chen, T. A. Hamor, K. Paxton and C. J. Jones, J. Chem. SOC., Dalton Trans., 1998,2163. Guo-Xin Jin, Abstract of 19&International Conference on Organometallic Chemistry, Paper S. 0 . 3 0 , Shanghai, China. M. Herberhold, H. Yan, W. Milius and B. Wrackmeyer, J Organomet. Chem., 2000, 604,170. a) 0. Crespo, M. C. Gimeno, P. G. Jones, B. Ahrens and A. Laguna, Inorg. Chem., 1997,36,495; b) 0.Crespo, M. C. Gimeno, P. G. Jones and A. Laguna, J. Organomet. Chem., 1997,547,89. M. F. Hawthorne and Z. Zheng, Acc. Chem. Res., 1997,30,267. M. Herberhold, Guo-Xin Jin, H. Yan, W. Milius and B. Wrackmeyer, Eur. J. Inorg. Chem., 1999,873. J. Y. Bae, Y. I. Park, J. KO, K. I. Park, S. I. Cho and S . 0. Kang, Inorg. Chim. Acta, 1999,289,141. D. Sellmann, M. Geck, F. Knock, G. Ritter and J. Dengler, J Am. Chem. SOC.,1991, 113,3819. D. H. Kim, J. KO,K. Park, S. Cho and S. 0. Kang, Organometallics, 1999,18,2738. Q. A. Kong, Guo-Xin Jin and Y. H. Lin, Chem. J Chin. Univ., 2002,23,410. X. Y. Yu, Guo-Xin Jin, L. H. Weng, Chin. J Chem., in press. a. J. Arnold, Progr. Inorg. Chem., 1995,43,353; b. J. R. Dilworth and J. Hu, Adv. Inorg. Chem., 1993,40,41 1. M. Herberhold, H. Yan, W. Milius and B. Wrackmeyer, Chern. Eur. J , 2000,6,3026. M. Herberhold, H. Yan, W. Milius and B. Wrackmeyer, Angew. Chem., Int. Ed. Engl., 1999,38,3689. M. Herberhold, H. Yan, W. Milius and B. Wrackmeyer, J Chem. SOC.,Dalton Trans., 2001,1782. J. Y. Bae,Y. J. Lee, S. J. Kim, J. KO, S. Cho and S. 0. Kang, Organometallics, 2000, 19, 1514.

SYNTHESIS AND REACTIVITIES OF MULTINUCLEAR SULFUR-BUDGED METAL COMPLEXES RANGING FROM DINUCLEAR TO HEXANUCLEAR CORES

Masanobu Hidai Department of Materials Science and Technology, Faculty of Industrial Science and Technology, Science University of Tokyo, Noda, Chiba 278-85 10, Japan

1 INTRODUCTION

We have long been interested in biological nitrogen fixation which proceeds on a metalsulfur cluster site in nitrogenase under mild conditions. Stimulated by the unique catalysis of the metalloenzyme, our attention has been paid to synthesis and reactivities of polynuclear transition metal-sulfur complexes containing noble metals because novel modes of activation and transformation of organic molecules on the well-defined multinuclear centers might lead to development of new organic syntheses which are inaccessible through conventional mononuclear complexes. In this paper, we summarize our recent study on synthesis and reactivities of polynuclear metal-sulfur complexes ranging from dinuclear to hexanuclear cores. 2 METHOD AND RESULTS

2.1 Novel Catalysis of Thiolato-Bridged Diruthenium Complexes

Three types of thiolato-bridged diruthenium complexes, [Cp*Ru(p2-SR)2RuCp*], [Cp*Ru(p2-SR)3RuCp*], and [Cp*RuCI(p2-SR)2RuCp*CI] (Cp* = q'-CsMes, R = alkyl, aryl) are readily available from [Cp*RuCI(p~-C1)2RuCp*C1] by treatment with appropriate thiolato compounds.' For example, the Ru(III)/Ru(III) complexes [Cp*RuCl(pzSR)2RuCp*CI] 1 are prepared in good yield by the reactions with RSSiMe3 (eq 1).

l a : (R = Me) lb: (R = Et) I c : (R = "Pr) Id: (R = 'Pr)

Complexes la, lb, and l c exhibit catalytic activity for the head-to-head Z dimerization of various terminal alkynes including aliphatic alkynes with functional groups

Synthesis and reactivities of multinuclear sulfur-bridged metal complexes

63

such as chloro, hydroxy, and ester groups (eq 2).2The reactions proceed at 60 "C with high regio- and stereoselectivity. Noteworthy is that the sterically demanding 'PrS-bridged complex Id show low activity, which is in sharp contrast to high activity of complex lc under the same conditions. The reactions can be extended to the intramolecular cyclization of a,o-diynes. Thus, this provides a novel synthetic route to endo-macrocyclic (Q-l-en-3-

2 RH -

(n = 6,8,10, 12)

ynes from a,o - d i y n e ~ ,whereas ~ the palladium-catalyzed cyclization of a, o-diynes to af'ford em-macrocyclic I -en-3-ynes was already reported by Trost and c o - ~ o r k e r s The .~ cyclization of a, a-diynes HC-C(CH&CrCH (n = 6, 8, 10, 12) is performed in the presence of complex la, lb, or l c and NH4BF4 under dilute conditions to obtain the endomacrocyclic ring products in high yield (eq 3). It is to be noted that the iridium complex [Ir(biph)(PMes)3Cl] (biph = biphenyl-2,2'-diyl), which is known to be effective for the selective head-to-head Z dimerization of aliphatic terminal alkynes,' does not show the catalytic activity for the cyclization of a,a-diynes under similar conditions. When the complex [C~*RUC~(~~-S'P~)~RUC~*(OH~)][OT~J (OTf = OS02CF3) l e available from complex Id and AgOTf is reacted with acetylene or terminal alkynes with electron-withdrawing substituents at room temperature, the terminal vinylidene complexes [C~*RUC~(~~-S'P~)~RUC~*(=C=CHR)][OT~J (R = H, COOMe, COMe) are formed (eq 4).6 On the other hand, treatment of complex l e with ferrocenylacetylene at ambient temperature affords the butenynyl complex [Cp*Ru{q1:q2-p2-C(=CHFc)C~CFc](p2S'Pr)zRuCp*J[OTfJ (Fc = ferrocenyl), the structure of which has been unequivocally determined by X-ray analysis (eq 5).7 The butenynyl complex catalyzes the selective dimerization of ferrocenylacetylene at 60 "C to form the head-to-head Z dimer CHFc=CHC-CFc.

le

(R = H,COOMe, COMe)

1OTf le

Perspectives in Organometallic Chemistry

64

Scheme 1 shows a plausible mechanism for the dimerization of terminal alkynes including the cyclization of a , a-diynes on the basis of the above results. Two terminal alkynes are incorporated at the diruthenium center to form a vinylidene-alkynyl intermediate, which- is then transformed into a butenynyl intermediate by nucleophilic migration of the alkynyl group to the a-carbon of the vinylidene ligand. Protonolysis of the butenynyl moiety gives rise to the formation of the head-to-head 2 dimer under the steric influence of the Cp* and thiolato ligands around the diruthenium center.

n

CP*,

I

MeS

CI

n

Ru-RU

I

CI

la

Scheme 1

Of great interest is that diruthenium thiolato-bridged complexes 1 catalyze the propargylic substitution reactions of propargylic alcohols with various heteroatom-centered nucleophiles such as alcohols, amides, amines, and thiols,* although conventional monoruthenium complexes such as [CpRuCl(PPh3)2] (Cp = q5-C5H5) and [RuC12(PPh3)3] are totally ineffective for the reactions. This chemistry is further extended to more valuable carbon-carbon bond formation reactions by employment of carbon-centered nucleophiles including not only P-diketones, but also simple ketones.' The reactions proceed with complete regioselectivities, and allenic by-products, which are always produced by the classical propargylic substitutions, are not formed at all. Interestingly, when unsymmetrical simple ketones are used as carbon-centered nucleophiles, the propargylic alkylation occurs at the more encumbered a-site of the ketones. Noteworthy is that the propargylic substitution of propargylic alcohols with ketones proceeds smoothly under mild and neutral conditions to give the corresponding y-keto acetylenes. These novel reactions catalyzed by complexes 1 are in sharp contrast to the Nicholas reaction which is known to be effective for the propargylic substitution, but requires a stoichiometric amount of CO2(CO)* .I" As typical examples, Scheme 2 summarizes the propargylic substitution reactions of 1-phenyl-2-propyn- 1-01 with various heteroatom- and carbon-centered nucleophiles which produce the corresponding propargylic derivatives in high yields with complete regioselectivities when performed at 60 "C in the presence of complex l a (5 mol %) and NH4BF4 (10 mol YO).

Synthesis and reactivities of multinuctear sulfur-bridged metal complexes

P P

h

Y

h

N

65

H

i)Et

H

t

ph
pTH 0

P

h

y

H

Y Y 0 0 Scheme 2

When complex l a is reacted with 1 equiv of 1,l-di-p-tolyl-substitutedpropargylic alcohol in the presence of NH4BF4 in EtOH at ambient temperature, the allenylidene complex [Cp*RuCl(pz-SMe)ZRuCp*{ C=C=C(Tol-p)2}][BF4] is isolated from the mixture in high yield, which liberates the corresponding propargylic ethyl ether in high yield by treatment with excess of EtOH at 60 "C (eq 6). Based on these results, a mechanism for the propargylic substitution reaction of a propargylic alcohol with an alcohol is proposed as shown in Scheme 3, which also explains the propargylic substitution reaction of a propargylic alcohol with other nucleophiles. The first step is the formation of a vinylidene complex 2, which is transformed into an allenylidene complex 3 via the dehydration. Subsequent nucleophilic attack of an alcohol on the y-carbon atom in the allenylidene ligand gives rise to the formation of another vinylidene complex 4, which isomerizes into an alkyne complex 5 where the C=C bond in a propargylic ether is bound to the ruthenium center. The ether is liberated from the metal center by excess of the propargylic alcohol to regenerate intermediate 2.

Tolfl"

la

OH NH4BF4 r.t.

'p*\

RUFRU

/

MeS'T.LaSMe

C p " 1 BF4 EtOH

II

C ToI'ToI

60 OC

t

TO1flH

To1

OEt

Perspectives in Organometallic Chemistry

66

la

OR A

Ry"-r

n l+

Ru-RU I .$OH

>*

OH

2

n

Ru-RU

b,

nl+

Ru-RU

i+

I

OR H& 5

"

fiC

R

R' 4

Scheme 3

2.2 Heterobimetallic IrzM (M = Pd, Pt) Trinuclear Sulfido Clusters and Their Catalytic Activity A series of dinuclear hydrosulfido-bridged complexes [Cp*MCl(p2-SH)2MCp*Cl] (6a : M Ru; 6b : M = Rh;6c : M = Ir ) are readily available from the reactions of [Cp*MCl(p2C1)2MCp*Cl] with excess H2S gas." A variety of tri-, tetra, and pentanuclear sulfido

=

clusters are derived starting from these complexes 6. Thus, treatment of complex 6c with [MCl2(cod)] (M = Pd, Pt; cod = cycloocta-1,5-diene) affords trinuclear sulfido clusters [(Cp*Ir)2(p-13-S)2MC12] 7 (eq 7).'* Although several complexes such as [PtClz(phosphine)2]13 have been reported to catalyze the addition of alcohols to alkynes, no regioselective catalyst has so far been developed for the reactions of internal alkynes.

7a (M = Pd) 7b (M = Pt)

Synrhesis and reactivities of multinuclear sulfur-bridged metal complexes

67

Interestingly, clusters 7 catalyze the regioselective addition of alcohols to internal 1-aryl1-alkynes. Thus, when 1-phenyl-1-propyne is treated with MeOH at 50 “C in the presence of cluster 7a, 2,2-dimethoxy-l-phenylpropaneis mainly formed with a small amount of 1,l-dimethoxy-1-phenylpropane in the ratio of 98 : 2 (eq 8). Complex 7b is a less regioselective catalyst for the reaction. The reaction is probably initiated by replacement of the chloro ligand at the palladium atom in 7a with an alkyne, which is followed by the regioselective nucleophilic attack of an alcohol on the coordinated alkyne. However, the detailed mechanism is not known.

O-Me

+

MeOH

-

M e 0 OMe

7a 50 OC

(98:2)

2.3 Synthesis and Reactivities of Cubane-Type Sulfido Clusters Containing Noble Metals’

Hydrosulfido-bridged dinuclear complexes with a M(pz-SH)2M’ core are versatile precursors for cubane-type sulfido clusters because a-elimination of hydrogen halide or hydrocarbons from the metal gives rise to the formation of coordinatively unsaturated sulfido-bridged dinuclear species with a M(p2-S)zM’ core and the resultant species dimerize to form cubane-type sulfido clusters with a MzM’2S4 core. Thus, treatment of complexes 6 with triethylamine affords a series of cubane-type clusters [(Cp*M)4(p3-S)4] (M = Ru, Rh, Ir). ” The heterobimetallic hydrosulfido-bridged complex [CpzTi(p~SH)2RuClCp*] 8, which is obtained from [Cp2Ti(SH)2] and [(Cp*Ru)4(p3-C1)4], is also transformed into the mixed-metal cubane-type sulfido cluster [(CpTi)2(Cp*Ru)2(p3-S)4] 9 by treatment with triethylamine.’5 The 60e- cluster 9 has only four Ru-Ti dative bonds

cp*’

a

9

Bu ~ NCI

cp*’

cI-II\ S3,i/JiCP

ClCHzCHzCl reflux

I\

RuCp”’ 10

Scheme 4

68

Perspectives in Organometallic Chemistry

because the Ru-Ru antibonding orbital lies below the Ti-Ti bonding orbital due to the high-lying titanium d orbitals. When cluster 9 is oxidized with HCI, the 62e- cluster [(CpTiCl2)(CpTi)(Cp*Ru)2(p3-S)4] 10 is obtained where the Ti atom with two chloro ligands extrudes significantly from the remaining three metals. Interestingly, the Cp ligand bound to the unique Ti atom in cluster 10 is substituted by a chloride anion to give the trichloro 58e- cluster [(TiC13)(CpTi)(Cp*Ru)2(p3-S)4] (Scheme 4). These unusual reactivities may be ascribed to the steric and electronic flexibility of the cubane-type sulfido core. Here, the metal-metal bonds in cubane-type sulfido clusters are not depicted for simplicity in the structural formulas. The dinuclear sulfido-bridged complexes with a syn-M2S~(p?-S)2 core are also employed as usefbl precursors for cubane-type sulfido clusters. Thus, the reactions of [M2S2(p2-S)2(S2CNEt&] (M = Mo, W) 11 with 2 equiv of [Pd(PPh3)4]I6 or 1 equiv of [{M'(cod)}2(p2-C1)2] (M' = Rh, Ir)I7 give cubane-type sulfido clusters 12 and 13, respectively. It is to be noted that when complex 11 is treated with 2 equiv of [M'CI(PPhl)3] (M' = Rh, Ir), the expected tetranuclear cluster is not obtained, but is observed the formation of the trinuclear clusters 14 with a M ' M Z ( ~ Z - S ) ~ ( ~core ~-S) (Scheme 5).

cod,

PhsP, PPh3

S-M-S'g

I/-s

CI

14

'

13

Scheme 5

Novel trimetallic cubane-type sulfido clusters are prepared by employment of dirhenium sulfido-bridged dithiolene complex [Ph4P]2[Re2S2(p2-S)2(S Z C ~ R ~15 ) ~(R ] = SiMe3), which equilibrates between sy17 and anti isomers in solution. Treatment of complex 15 with [Cp*IrCI(p~-C1)2IrCp*Cl] results in the formation of an incomplete cubane-type cluster 16 with a Re2lrS4 core, which fbrther reacts with [Pd(PPh3)4] to form a trimetallic cubane-type cluster 17 with a RezIrPdS4 core. Interestingly, the reaction of cluster 16 with

Synthesis and reactivities of multinuclear sulfur-bridged metal complexes

69

15 (R = SiM%)

,Pi, Ph3P 17

PPh3 18

Scheme 6

[Pt(PPh3)4] affords a tetranuclear cluster 18 with a raft-type Re2IrPtS4 core (Scheme 6). In cluster 18, the formal oxidation of the Pt atom may be +2 which favors a square planar structure. l 8 Shibahara and coworkers developed a usehl synthetic method for cubane-type sulfido clusters which is comprised of incorporation of heterometals into the incomplete cubane19 readily available from "H&[MoS4]. l9 We type cluster [ { MO(H~O)~)~(~~-S)(~Z-S)~]~+ have employed this synthetic route to prepare cubane-type sulfido clusters with PdMo3S4 and PtMo3S4 cores2' When cluster 19 is treated with Pd black in aqueous HCl, a cubanetype cluster 20 with a Mo3PdS4 core is obtained. The water molecules coordinated to the Mo atoms in 20 are readily replaced by 1,4,7-triazacyclononane (tacn) to afford a cubanetype cluster 21 (Scheme 7). The unique tetrahedral palladium atom in cluster 21 can bind various substrates such as olefins and CO. Noteworthy is that the PdMo& cluster 213'[PF& effectively catalyzes the stereoselective addition of alcohols to alkynic acid esters such as methyl propiolate (eq 9), although conventional mononuclear Pd complexes are not effective for the reactions. 14+ L-Mo-S :/L / S S-Mo-L /I L L-Mo-S

I/

I I/L

L/L

19 (L = H20)

13+

hL

L-MOS -//

Pd aqHCI

d$Pd

1)

I/s-I;Mp-L L

-

tacn

13+

\

tacn

L-Mo-S

L/[

tacn' 20

Scheme 7

21

Perspectives in Organometallic Chemistry

70

H-COOMe

+

MeOH

2 13+[PF& 40 OC

MeoMcooMe H H (9)

A plausible mechanism for the reaction of eq 9 is shown in Scheme 8, where the coordinated tacn ligands on the Mo atoms are omitted for clarity. The initial step is the coordination of the alkyne to the unique Pd atom and subsequent nucleophilic attack of methanol from the outer coordination sphere gives rise to a vinylpalladium intermediate 22, which liberates the trans addition product by protonolysis. Although the detailed mechanism is not known, it is of great interest that the cubane-type core is retained during the reaction and the catalysis is realized by the unique Pd atom embedded in a M03S4 aggregate.'Oa MeOH

cI

I

Pd

1 3 +

H+

H-COOMJ 4+ , r C O O M e1 I

P'd

/ I \

"

3 +

I

Pd

22

Meo)==(cooMe H H

H+, H-COOMe

Scheme 8

The stereoselective addition of carboxylic acids to alkynes with electron-withdrawing groups also smoothly proceeds under mild conditions in the presence of a catalytic amount of cluster 213+[PF~]3.'' When alkynoic acids such as 4-pentynoic acid are used, the cyclization occurs to give the corresponding enol lactones.22 In this case, conventional mononuclear Pd(I1) compounds such as [PdC12(PhCN)2] are used as the catalyst for the lactonization of alkynoic acids. However, the catalytic activity of cluster 21 is extremely higher than those of the mononuclear compounds. When 4-pentynoic acid is treated with cluster catalyst 21 for 19 h at 40 O C , the turnover number reaches 100,000(eq 10).

Synthesis and reactivities of multinuclear sulfur-bridged metal complexes

71

2.4 Pentanuclear and Hexanuclear Sulfido Clusters

A series of dicationic pentanuclear bow-tie clusters [(Cp*Ir)z(p3-S)zM( ~ ~ - S ) Z ( I ~ C ~ * ) Z ] ~ + (23a: M = Fe; 23b : M = Co; 23c : M = Ni) are readily obtained starting from complex 6c as shown in Scheme 9.23Reaction of complex 6c with excess FeClz gives the trinuclear cluster [(Cp*Ir)~(p3-S)zFeCl~],which is then transformed into the dicationic 78epentanuclear bow-tie cluster 23a by treatment with NaBPh. The 79e- and 80e- bow-tie clusters 23b and 23c are prepared directly from the reaction of complex 6c with CoClz or NiC12, respectively. X-ray analysis of bow-tie clusters 23 reveals that the structures significantly change depending upon the kind of the central metal atom.

23c

Scheme 9

A more sophisticated hexanuclear cluster is prepared in a stepwise way starting from the hydrosulfido-bridged complex 8. The reaction of complex 8 with [Pd(PPh3)4] gives a trinuclear cluster 24 with an unprecedented M3(13-S)(pz-S) core, which is then converted into a hexanuclear cluster 25 with a TizRuzPdzS40 core by treatment with water and a base.24In the X-ray structure of cluster 25, the two corner-voided TiRuPdzSzO cuboidal fragments share the PdzO face, where the metal-metal bonds are omitted for clarity (Scheme 10).

8

24

Scheme 10

25

72

Perspectives in Organometallic Chemistry

3 CONCLUSION

In summary we have developed usefbl routes to a variety of multinuclear sulfbr-bridged metal complexes ranging from dinuclear to hexanuclear cores and found some unique reactions catalyzed by polynuclear complexes. We expect that rational synthetic methods for more sophisticated metal sulfbr clusters will be further developed and the clusters prepared in the rational ways exhibit novel catalytic activities and other functions which are not realized by conventional mononuclear complexes. Acknowledgements 1 am deeply indebted to the enthusiastic efforts of the students and collaborators whose names are listed in the papers cited in the references. References 1 (a) M. Hidai, Y. Mizobe and H. Matsuzaka, J. Organomet. Chem., 1994, 473, 1; (b) M. Hidai and Y. Mizobe, in Transition Metal Si@r Chemistry: Biological and Industrial Significance, eds. E. I. Stiefel and K. Matsumoto, American Chemical Society, Washington DC, 1996, ch. 19, p. 3 10. 2 J.-P. QU, D. Masui, Y. Ishii and M. Hidai, Chem. Lett., 1998, 1003. 3 Y. Nishibayashi, M. Yamanashi, I. Wakiji and M. Hidai, Angew. Chem., Int. Ed. Engl., 2000,39,2909. 4 B. M. Trost, S. Matsubara and J. J. Caringi, J. Am. Chem. SOC.,1989, 111, 8745. 5 C.-H. Jun, Z. Lu and R. H. Crabtree, TetrahedronLett., 1992,33, 71 19. 6 Y. Takagi, H. Matsuzaka, Y. Ishii and M. Hidai, Organometallics, 1997, 16, 4445. 7 H. Matsuzaka, Y. Takagi, Y. Ishii, M. Nishio and M. Hidai, Organometallics, 1995, 14, 2153. 8 Y. Nishibayashi, I. Wakiji and M. Hidai, J. Am. Chem. Soc., 2000, 122, 11019. 9 Y. Nishibayashi, I. Wakiji, Y. Ishii, S. Uemura and M. Hidai, J Am. Chem. SOC.,2001, 123,3393. 10 (a) K. M. Nicholas, ACC.Chem. Res., 1987, 20, 207; (b) A. J. M. Caffyn and K. M. Nicholas, in Comprehensive Organometallic Chemistry II, eds. E. W. Abel, F. G. A. Stone and G. Wilkinson, Pergamon, New York, 1995, Vol. 12, ch. 7.1. 1 1 (a) K. Hashizume, Y. Mizobe and M. Hidai, Organometallics, 1996, 15, 3303; (b) Z. Tang, Y. Nomura, Y. Ishii, Y. Mizobe and M. Hidai, Organometallics, 1997, 16, 151. 12 D. Masui, T. Kochi, Z. Tang, Y. Ishii, Y. Mizobe and M. Hidai, J Organomet. Chem., 2001, 620, 69. 13 Y. Kataoka, 0. Matsumoto and K. Tani, Organometallics, 1996, 15, 5246. 14 M. Hidai, S. Kuwata and Y. Mizobe, Acc. Chem. Res., 2000,33,46. 15 S. Kabashima, S. Kuwata and M. Hidai, J. Am. Chem. SOC., 1999, 121, 7837. 16 T. Ikada, S. Kuwata, Y. Mizobe and M. Hidai, Inorg. Chem., 1998,37, 5793. 17 T. Ikada, S. Kuwata, Y. Mizobe and M. Hidai, Inorg. Chem., 1999, 38, 64. 18 H. Seino, T. Kaneko, S. Fujii, Y. Mizobe and M. Hidai, to be published. 19 (a) T. Shibahara, Adv. Inorg. ('hem., 1991, 37, 143; (b) T. Shibahara, Coord. Chem. Rev., 1993, 123, 73. 20 (a) T. Murata, Y. Mizobe, H. Gao, Y. Ishii, T. Wakabayashi, F. Nakano, T. Tanase, S. Yano, M. Hidai, I. Echizen, H. Nanikawa and S . Motomura, J. Am. Chem. SOC., 1994, 116, 3389; (b) D. Masui, Y. Ishii and M. Hidai, Bid/. Chem. SOC.Jpn., 2000, 73, 93 1. 21 T. Wakabayashi, Y. Ishii, T. Murata, Y. Mizobe and M. Hidai, Tetrahedron Lett.,

Synthesis and reactivities of multinuclear sulfur-bridged metal complexes

73

1995,36, 5585. 22 T. Wakabayashi, Y. Ishii, K. Ishikawa and M. Hidai, Angew. Chem., Inf. E d Engl., 1996,35,2123. 23 Z. Tang, Y. Nomura, S. Kuwata, Y. Ishii, Y. Mizobe and M. Hidai, Inorg. Chem., 1998,37,4909. 24 S . Kuwata, S. Kabashima, Y. Ishii and M. Hidai, J. Am. Chern. SOC.,2001, 123, 3826.

a,o-BIS [(TRIPHENYLPHOSPHINE)GOLD(I)]HYDROCARBONS

Keith A. Porter, Annette Schier and Hubert Schmidbaur* Anorganisch-chemisches Institut, Technische Universitat Miinchen, Lichtenbergstrasse 4, D-85747 Garching, Germany

1 INTRODUCTION While simple organomercury compounds of the types R-Hg-R or R-Hg-X were discovered already in the 19th century,' the first organogold(1) complexes of the type RAu-L with L representing a two-electron donor ligand and R an alkyl or an aryl group were prepared as late as 1959. Coates et al. used organolithium reagents for the first successful alkylations/arylations of @hosphine)gold halides.2 A large number of representative complexes have since been synthesized and carefully investigated in several laboratories. In the following two decades alkenyl- and alkynylgold(1) complexes were also prepared but remained smaller in number.3 The highly explosive gold(1) acetylide Au2C2 first described by Berthelot in 1866 was converted into a stable (phosphine)gold(I) complex also by Coates et al. in 1959.2y4y5 bdmmuku gold(1) organyls are among the most stable transition metal organometallics with discrete metal-to-carbon sigma-bonds. Thus the prototype species MeAu(PMe3) is not affected by water and cleaved only by aggressive acids and bases, and is thus comparable in chemical stability only to the classical organomercurials like MeHgC1.6 This stability is largely due to the low polarity of the Au-C bonds which is not only evident fiom the robustness of the compounds against nucleo- and electrophilic cleavage, but was also confirmed by physical measurements using a variety of techniques.'~*Gold is known to be the most electronegative metal with a very small covalent and ionic radius. State-of-the-art quantum-chemical calculations including relativistic effects have consistently demonstrated that Au-C bonds are strongly directional following the sp/dhybridization model and have a very low dipolar character.' It should be noted that the stability of Au-C sigma-bonds is in sharp contrast to the properties of the labile organosilver com ounds, where the metal atom has a significantly larger radius and lower electronegativity. The mechanistic pathways of the thermal and photochemical decomposition of compounds of the type R-Au-L were studied extensively, mainly by Kochi et al.," and shown to be based on a radical process which leads predominantly to R-R coupling of the groups R.3a Organogold complexes are therefore precursors for thermally and photochemically induced C-C coupling. lXnu&ar organogold compounds in which an organic group is acting as a spacer

Po

a,a~Bis[(triphenylphosphine)gold(I)]hydrocarbons

75

between bars gold(1) centres (A) should thus give access to s l i d i d species if the metal centres are well separated to rule out direct neighbowing group assistance in the decomposition process. The hydrocarbon spacer could be an amalkanediyl unit or a related difunctional olefinic, acetylenic or aromatic group. The diradicals may be expected to undergo internal stabilization or external coupling, as well as be substrates for trapping reactions with suitable reagents.

A

1-

hydrocarbon spacer

J

Dinuclear com ounds of this type (A) are very rare in the literature except for a series of P (poly)acetylides’2’ and isolated cases with highly substituted or branched bridging The present account is a report of our own efforts to provide a selection of simple a,w-digold hydrocarbons which will be studied in detail in a forthcoming mechanistic investigation of their thermal decomposition in the absence or presence of trapping reagents. 2 PREPARATIVE STUDIES Two examples of a,w-digolhave been prepared by the reaction of the corresponding Grignard reagents with (tripheny1phosphine)gold bromide in the molar ratio 1:2 in tetrahydrofuran as a solvent [Eq. (l)]. The yield of the n-butanediyl (1) and npentanediyl compound (2) was 41% (m.p. 152 “C) and 36% (m.p. 126 “C), respectively. The products were isolated as colourless, air- and water-stable crystalline solids, soluble in tetrahydrofuran and dichloromethane. Note that the stability of the compounds allows a convenient aqueous work-up of the reaction mixtures. B

r

wBr

1. Mg, THF c 2 . 2 Ph3PAuBr,THF

Ph3PAuwAuPPh3

(1)

ln=2 2n=3

Preparative efforts oriented at the synthesis of the corresponding ethane-, propane- and cyclohexane-diyls were not successful. Neither the Grignard nor the organolithium route gave any alkane-diyl product. It appears that the elimination of olefin leads to an immediate deorganylation after the first metallation step. Metallation of certain olefinic substrates also proved unsuccessful especially where benzylic or allylic precursors were employed. Metallation of 1,4-dibrom0-2-butene, and both 0- and pbis(bromomethy1)benzene resulted in the elimination of the respective olefins (possibly as dienes). Although benzylic and allylic type Grignard reagents are difficult to prepare, novel magnesium-anthracene compounds can offer a route to their synthesis. 9 Their use in the synthesis of a,wbis[(triphenylphosphine)gold]-olefms, however, proved also unsuucessfbl. The preparation of an a,w-digold&fh has been achieved by the two-step reaction of thiophene with ‘butyllithium followed by (tripheny1phosphine)gold chloride. Lithiation of thiophene in tetrahydrofurdn-hexane in the presence of tetramethylethylenediamine occurs regioselectively yielding the dilithio salt 3a (Eqn 2). The reaction of 3a with two

76

Perspectives in Organometallic Chemistry

equivalents of (Ph3P)AuCl yields solely the 2,5-bis[(triphenylphosphine)gold]-thiophene in 23% (3, m.p. 189 "C with decomposition). Compound 3 has been fully characterized by NMR, elemental analysis and by its mass spectra although no single crystal suitable for Xray diffraction could be obtained.

0

"BuLih-hexane, TMEDA

L

i

eLi

Ph3PAuCI

L

Ph3PAu

AuPPh3 (2)

3

3a

The attempted synthesis of a,o-digoldamxs from the corresponding unsubstituted benzene and naphthalene precursors was not successful, however, three digoldamxs have been synthesized from 1,4-xylene, biphenyl and anthracene substrates. The lithiumaryls were obtained by treatment of 1,4-dibrom0-2,5-dimethylbenzene,4,4'dibromobiphenyl and 9,lO-dibromoanthracene with 'butyl-lithium or "butyl-lithium in tetrahydrofuradhexane and reacted with two equivalents of (Ph3P)AuCl. Aqueous work-up gave the products in yields of 63% (4, m.p. 147 "C with decomposition), 76% (5, m.p. 175 "C with decomposition) and 50% (6, m.p. 207 "C with decomposition) [Eqs. (3) - ( 5 ) ] .

J$

1. 'BuLih-hexane, THF *

Y

2. P ~ ~ P A u C I

AuPPh3

Br

4

I. 'BuLiln-hexane, THF

PhlPAu-AuPPhl-

(4)

2. P ~ ~ P A u C I 5

Br I

fiuPPh3

1. 'BuLih-hexane, THF

*

2. PhSPAUCI

Br

AuPPh3

6

Compounds 1 - 6 have been characterized by their NMR spectra and by elemental analyses (Exp. Part). For compound 2 , 3 and 6 the molecular ion was observed in the mass spectrum, while all other compounds showed only fragment ions upon chemical ionization or fast atom bombardment. Single crystals suitable for structure determination by X-ray diffraction have been obtained for complexes 1 and 4 - 6.

a,m-Bis[(triphenylphosphine)gold(I)]hydrocarbons

77

3 CRYSTAL STRUCTURE ANALYSIS

3.1 1,4-Bis[(tripheny1phosphine)gold)-n-butane, 1. Crystals of { [(Ph3P)Au]CH2CH2}2, 1 (from dichloromethane/pentane at -30 "C), are monoclinic, space group P21/c, with Z = 2 molecules in the unit cell. The lattice is free of solvent and the asymmetric unit contains only one half of the molecule. The second half can be generated by the symmetry operations related to the centre of inversion situated in the middle of the central C-C bond [C2-C2A, Figure 11. Each of the two gold atoms is attached to one of the two terminal carbon atoms of the butane chain and has the usual linear two-coordinate environment: P-Au-C 1 177.58(9)", Au-P 2.3042(7), Au-C 1 2.081(3) A. For comparison, the data for the reference compound MeAu(PPh3) are 179.1", 2.279 and 2.124 A.

Fig. 1 The butane chain of the molecule (1) has C-C distances corresponding to standard single bonds [Cl-C2 1.523(4), C2-C2A 1.530(6) A]. The hydrogen atoms have been located in the structure determination and their positions refined with isotropic displacement parameters. The methylene groups form a gauche conformation in the two terminal CH2CH2 units and a trans conformation in the central CH2-CH2 unit. The carbon atoms of the butane chain are thus coplanar, but the gold atoms are located on different sides of the plane of the carbon atoms with a torsional angle Au-C 1-C2-C2A of 62.1(4)". With the butane chain fully extended and with large triphenylphosphine ligands shielding the gold atoms there are no discernable intra- or intermolecular aurophilic contacts between the molecules. The packing shows no z-z-stacking of phenyl groups and intermolecular contacts are limited to weak "phenyl-embracing"of neighbouring P b P ligands. Compound 2 is assumed to have a molecular structure similar to that of compound 1, but no single crystals could be obtained of this complex.

3.2 1,4-Bis[(triphenylphosphine)gold]-2,4-dimethyl-benzene,4. Crystals of 4 are triclinic, space group P i,with Z = 1 molecule in the unit cell. The lattice contains no solvent and the asymmetric unit features only one half of the molecule, the

78

Perspectives in OrganometallicChemistry

second half being related by a centre of inversion in the middle of the central benzene ring (Figure 2).

Fig. 2 The Au-P [2.289(1) A] and Au-C [2.046(4) A] distances and the P-Au-C angle [176.78(12)”]can be considered normal for an arylgold complex of a triarylphosphine. The C-C distances and C-C-C angles of the central ring show only small deviations from the standard dimensions of p-xylene indicating only minimal steric interactions between neighbouring (Ph3P)Au and methyl groups. The packing of the complex appears to be governed by the stacking of phenyl groups of neighbouring molecules with a combination of z-n- and phenyl-embracing contributions as shown in the projections of Figures 3a and 3b. Owing to steric hindrance there are no aurophilic contacts. An evaluation of potential extremely weak hydrogen bonds between phenyl hydrogen atoms and the remainder functions (Au, C=C etc.) is not within the scope of this study.

Fig. 3a

a,o-Bis[(triphenylphosphine)gold(I)] hydrocarbons

79

Fig. 3b 3.3 4,4'-Bis[(triphenylphosphine)gold]-biphenyl, 5.

Crystals of 5 are monoclinic, space group P21/n, with Z = 2 molecules in the unit cell. The asymmetric unit contains half of the formula unit related to the second half by a centre of inversion in the middle of the bond C M 4 A connecting the two phenyl parts of the biphenyl molecule (Figure 4). The Au-P [2.2914(9) A] and Au-C [2.051(4) A] distances and the P-Au-C angles [175.6(l)"] are similar to those of compound 4 (above) and of reference complexes.'* The two phenyl rings of the biphenyl unit are virtually coplanar with a dihedral angle C3-CM4A-C5A of only 1.1". This is a rare case in the structural chemistry of biphenyls, but the remaining details of the structure give no reason to assume any special chemical bonding, nor does the packing of the molecules show any anomalies or preferred orientations of the phenyl or biphenyl rings (Figure 5).

Fig. 4

Perspectives in Organometallic Chemistry

80

Fig. 5 3.4 9,10-Bis[(triphenylphosphine)gold]-anthracene, 6.

-

Crystals of compound 6 2(CH2C12) are monoclinic, space group P2dn with Z = 2 molecules in the unit cell. The asymmetric unit holds one half of the formula unit and one molecule of dichloromethane solvent. The second half of the complex and the second solvent molecule are generated via a centre of inversion in the middle of the central ring of the anthracene part (Figure 6). The Au-P [2.2936(9) 81 and Au-C [2.062(3) A] distances and the P-Au-C angle [177.56(9)0] show no significant differences from the data for 4 and 5 (above). The anthracene geometry exhibits some significant alternation of the C-C bond lengths making C3<4 and C W 7 the shortest and C4-€7A, C2<3, (25-426 and C 2 4 5 A the longer bonds in the outer rings [1.365(6), 1.356(6) versus 1.413(6), 1.441(5), 1.430(5), and 1.446(5) A]. C 1 4 2 and Cl-C5 are intermediate [1.406(5), 1.416(5) A]. This indication of some partial double bond localization at C3-C4 and C W 7 is typical for anthracene and has been verified by quantum-chemical calculations.20There is therefore nc specific effect of two gold substituents in 9,lO-position. The deviations of the anthracene part from planarity are also very small, the torsional angles not exceeding 1So.The packing of the molecules is efficiently space-filling and there is no evidence for orientational preference originating from phenyl interactions (Figure 7).

Fig. 6

a,w-Bis[(triphenylphosphine)gold(I)] hydrocarbons

81

Fig. 7 Owing to disorder of the dichloromethane molecule it remains an open question if the solvent molecule plays a role in determining the packing of the components. 4 CONCLUSIONS

The preparative part of this study has shown that tertiary phosphine complexes of simple a,o-digold-alkanes and -arenes are readily available via the reactions of (phosphine)gold halides with the corresponding organolithium or Grignard reagents. Like related mononuclear alkyl- and arylgold complexes the products are stable to daylight and to air and water and show high thermal stability. With these convenient general properties the complexes will be ideal model substrates for a mechanistic study of the controlled decomposition which is expected to follow a biradical pathway. The NMR-spectroscopic data and the results of structure determinations of four representative examples gave no indication for bonding anomalies related to the disubstitution with two gold atoms. Future investigation will answer the question if a communication between the radical centres generated in the thermal degradation will alter the decomposition pathway compared to the monoradical motif documented in the literature. Current studies focus on a,co-digold olefins which are much less readily available and therefore pose more of a synthetic challenge. 5 EXPERIMENTAL

All experiments were routinely carried out in an inert atmosphere of dry nitrogen. Solvents were purified, dried and saturated with nitrogen and glassware was oven-dried and filled with nitrogen. Standard equipment and instruments were used throughout. (Ph3P)AuCI2’ and ( p h 3 P ) A ~ B rwere ~ ~ prepared following literature grocedures, the Grignard and organolithium reagents synthesized via published methods. 724

82

Perspectives in Organometallic Chemistry

1,4-Bis[(triphenylphosphine)gold]-n-butane, 1 A Grignard solution was prepared from 1,4-dibromobutane (120 mg, 0.55 mmol) and magnesium powder (120 mg, 4.9 mmol) in 15 mL of tetrahydrofuran and added dropwise to (Ph3P)AuBr (300 mg, 0.56 mmol) dissolved in 5 mL of the same solvent at room temperature with stirring. Stirring was continued for 45 min and subsequently the solvent was removed in vacuu. The residue was taken up in dichloromethane (10 mL) and extracted with water (10 mL) to remove the magnesium bromide. The organic fraction was dried over MgS04, the solvent reduced in volume (ca. 2 mL) and the roduct precipitated ? by addition of n-pentane; yield 110 mg, 41%, m.p. 152 "C (decomp.). H-NMR (CDKL): 1.49 - 1.52, m, 4H, CH2; 1.99 - 2.01, m, 4H, CH2; 7.38 - 7.55, m, 30H, Ph. 13C{'H}-NMR (CD2Cl2): 31.9, d, 2 J c = ~ 94.6 Hz, 2C, CH2; 39.2, m, 2C, CH2; 129.2, d, 2 J c = ~ 10.1 Hz, 6C, U-C; 131.1, S, 3C, p-C; 132.5, d, 'Jcp = 44.6 Hz, 3C, @SO-C;134.7, d, 3 J c = ~ 13.9 Hz, 6C, m-C. 3'P{'H}-NMR (CD2C12): 45.9, S. MS (FAB): m/z 459 [P~~PAu]' (10%). Calcd. for C ~ O H ~ ~ ACU 49.30, ~ P ~H: 3.93; found: C 49.28, H 3.96%. 1,5-Bis[(triphenylphosphine)gold]+pen tane, 2 A Grignard solution was prepared fiom 1,5-dibromopropane (150 mg, 0.67 mmol) and magnesium powder (50 mg, 2.1 mmol) in 15 mL of tetrahydrofuran and added dropwise to (Ph3P)AuBr (300 mg, 0.56 mmol) dissolved in 5 mL of the same solvent at room temperature with stirring. Stirring was continued for 75 min and subsequently the solvent was removed in vacuo. The residue was taken up in dichloromethane (10 mL) and extracted with water (10 mL) to remove the magnesium bromide. The organic,fraction was dried over MgS04, the solvent reduced in volume (ca. 2 mL) and the roduct precipitated by addition of n-pentane; yield 100 mg, 41%, m.p. 126 "C (decomp.). H-NMR (CD2CL): 1.38 - 1.44, m, 4H, CH2; 1.56 - 1.64, m, 2H, CH2; 1.87 - 1.96, m, 2H, CH2; 7.38 - 7.56, m, 30H, Ph. 13C{'H}-NMR (CD2Cl2): 31.8, d, 2 J ~ = p 95.3 Hz, 2C, CH2; 32.1, s lC, CH2; i 44.3, S, 2C, CH2; 129.3, d, 2Jcp= 10.7 Hz, 6C, u-C; 131.1, S, 3C,p-C; 132.4, d, JCP = 44.6 Hz,3C, @SO-C;134.7, d, 3 J c ~= 13.9 Hz, 6C, rn-C. 31P{'H}-NMR(CD2Cl2): 45.9, S. MS (FAB): m/z 459 [Ph3PAu]+ (loo%), 721 [Ph3PAuPPh3]+ (61), 988 [MI+ (4). Calcd. for C ~ ~ H ~ O ACU 49.81, ~ P ~H: 4.08; found: C 49.51, H 4.10%.

s

2,5-Bis[(triphenylphosphine)gold]-thiophene, 3 To a solution of thiophene (34 mg, 0.4 mmol) and tetramethylethylenediamine(0.07 mL, 0.84 mmol) in n-hexane (3 mL) was added a 1.6 M solution of "butyllithium (0.53 ml, 0.85 mmol) at room temperature. The solution was stirred for 30 min at reflux and then an additional hour at room temperature. The resulting 2,5-dilithiothiophene suspension was then slowly added to Ph3PAuCl (380 mg, 0.8 mmol) in tetrahydrofiuan (10 mL). After an additional 40 min the solvent was removed in vacuu and the residue was taken up in dichloromethane (10 mL) and extracted with water (10 mL) to remove the lithium chloride. The organic phase was dried over MgS04, the solvent reduced in volume (ca. 2 mL) and the product precipitated by addition of diethyl ether; yield 90 mg, 23%, m.p. 189 "C (decomp.). 'H-NMR (CD2CL): 7.38 - 7.65, m, Ph, C4SH2). 31P{'H} NMR (CD2CL): 44.9, s. MS (FAB): m/z 459 [Ph3PAu]+ (35%), 721.3 [(Ph3P)2Au]+ (18), 1001 (3) . Calcd. for C~OHXAU~PZS: C 48.01, H 3.22; found: C 47.93, H 3.40%.

w]+

1,4-Bis[(triphenylphosphine)gold]-2,fi-dimethyl-benzene, 4 1,4-Dibrom0-2,5-dimethyl-benzene(78 mg, 0.30 mmol) was dissolved in tetrahydrofuran (10 mL) and a 1.6 M solution of kutyllithium in hexane (1.O ml, 1.5 mmol) was added at room temperature with stirring. The reaction mixture was stirred f6r 3 h to form a

a,oBis[(triphenylphosphine)gold(I)]hydrocarbons

a3

suspension of 1,4-dilithi0-2,5-dimethyl-benzene. To this suspension was added (Ph3P)AuCl (292 mg, 0.59 mmol). After 2 h of stirring at room temperature the solvent was removed in vacuo and the residue was taken up in dichloromethane (10 mL) and extracted with water (10 mL) to remove the lithium chloride. The organic phase was dried over MgS04, the solvent reduced in volume (ca. 2 mL) and the product precipitated by addition of n-pentane; yield 191 mg, 63%, m. . 147 "C (decomp.). 'H-NMR (CD2Cl2): 1.51, s, 6H, Me; 7.21 - 7.69, m, 32 H, ArH). P('H}-NMR (CD2CL): 44.2, s. MS (CI): m/z 721.3 [(Ph3P)zAu]+ (19 %). Calcd. for ( C U H ~ S A U ~ P ~ ) ~ ( CCH50.86, ~ C ~ ~H ) : 3.69; found: C 50.75, H 3.79%.

R

4,4'-Bis [(triphenylphosphine)gold]-biphenyl, 5 4,4'-Dibromo-biphenyl(lOOmg, 0.32 mmol) was dissolved in 3 mL of tetrahydrofuranand a 1.6 M solution of "butyllithium in n-hexane (0.88 ml, 1.41 mmol) was added at room temperature. The reaction mixture was stirred for 2 h to form a suspension of 4,4'-dilithiobiphenyl. To this suspension was added 302 mg (0.64 mmol) of (Ph3P)AuCl. After stirring at room temperature for an additional hour, solvent was removed in vacuo, the residue was taken up in dichloromethane (10 mL) and washed with water (10 mL) to remove lithium chloride. The organic phase was dried over MgS04, solvent reduced in volume (ca. 2 mL) and the product precipitated by addition of n-pentane; yield 260 mg, 76%, m.p. 175 "C (decomp.). 'H-NMR (CD2Cl2): 7.42 - 7.70, m, ArH. 31P{'HI NMR (CD2Cl2): 43.8, s. MS (FAB): d z 721 [(Ph3P)2Au]+(4%), 612 [ P ~ ~ P A u C ~ H ~ C (19), ~ H459 ~ ] '[(Ph3P)Au]+(100). Calcd. for C ~ S H ~ S A U C ~53.84, P ~ : H 3.58; found C 53.73, H 3.69%. 9,lO-Bis[(tripheny1phosphine)goldl-anthracene, 6 9,lO-Dibromoanthracene (98 mg, 0.29 mmol) was dissolved in 5 mL of tetrahydrofuran and a solution of 'butyllithium in n-hexane (1.0 mL, 1.5 mmol) was added at room temperature. The reaction mixture was stirred for 2 h to form a dark green suspension of 9,lO-dilithio-anthracene.To this suspension was added (Ph3P)AuCl (144 mg, 0.58 mmol) and the suspension turned orange. After stirring for an additional 45 min at room temperature the solvent was removed in vacuo, the residue taken up in 10 mL of dichloromethane and extracted with water (10 mL) to remove the lithium chloride. The organic phase was dried over MgS04, solvent reduced in volume (ca. 2 mL) and the Product precipitated by addition of n-pentane; yield 160 mg, SO%, m.p. 207 "C decomp.). H-NMR (CD2Cb): 7.28 - 7.32, 7.73 - 7.80, and 8.82 - 8.86, m, ArH. 31P{H}-NMR (CDzCL): 45.4, S. MS (FAB): d z 1095 [MI+ (7%), 833 [Ph3PAuK14Hs]+ (4), 721 [(Ph3P)2Au]+ (28), 637 [(P~~P)AUCMHS]+ (3), 459 [(Ph3P)Au]+ (46). Calcd. for C S O H ~ ~ A U ~ P ~ ( CCH50.61, ~ C ~ ~H)3.38; ~ S : found C 50.20, H 3.37%.

\

X-Ray crystallography Specimens of suitable quality and size of compounds 1, 4, 5, and 6 2 CH2C12 were mounted on the ends of quartz fibres in F06206R oil and used for intensity data collection on a Nonius DIP2020 difiactometer, employing graphite-monochromated Mo K, radiation. The structures were solved by a combination of direct methods (SHELXS-97) and differenceFourier syntheses and refined by fbll matrix least-squares calculations on (SI-IELXL-97).25 The thermal motion was treated anisotropically for all non-hydrogen atoms, except for those of the disordered solvent atoms in 6 which were refined isotropically in split positions (Hatoms neglected). All hydrogen atoms of compounds 4 and 6 and the Ph-H atoms of 1 were calculated and allowed to ride on their parent atoms with fixed isotropic contributions, ~

84

Perspectives in Organometallic Chemistry

whereas all hydrogen atoms of compound 5 and the C-H atoms of compound 1 were located and refined isotropically. Further information on crystal data, data collection and structure refinement are summarized in Table 1. Important interatomic distances and angles are given in the corresponding Figure Captions. Displacement parameters and complete tables of interatomic distances and angles have been deposited with the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 lEZ, UK. The data are available on request on quoting CCDS-XXX. Acknowledgements This work was generously supported by Deutsche Forschungsgemeinschaft, Fonds der Chemischen Industrie and Volkswagenstiftung. Donation of chemicals by Heraeus GmbH and Degussa AG is gratefully acknowledged. References and notes 1 a) E. Frankland, Philos. Trans. R. SOC.London, 1852, 142, 417. b) J. L. Wardell in Comprehensive Organometallic Chemistry, G. Wilkinson, F. G. A. Stone, E. W. Abel, editors, Pergamon, Oxford, 1982, p. 863 ff. 2 G. Calvin, G.E. Coates, and P.S. Dixon, Chem. and Ind., 1959, 1628. 3 a) H. Schmidbaur in Gmelin Handbook of Inorganic and Organometallic Chemistry, 8th edn., Au, Organogold Compounds, Springer-Verlag, Berlin, 1980. b) Gold: Progress in Chemistry, Biochemistry and Technology, H. Schmidbaur (editor), Wiley, Chichester, 1999, chapter 18, p. 648 ff., by H. Schmidbaur, A. Grohmann, M. E. Olmis. 4 M. P. Berthelot Liebigs Ann. Chem., 1866,139, 150. 5 A. Matthews, L. L. Watters, J. Am. Chem. SOC.1900,22, 108 6 K. Hartley, H. 0. Pritchard, H. A. Skinner, Trans. Faraday SOC.,1950,46,1019. 7 G. M . Bancroft, T. Chan, R.J. Puddephat, J. S . Tse, Inorg. Chem., 1982,21,2946. 8 A. Haaland, J. Hougen, H. V. Volden, and R. J. Puddephatt, J. Organomet. Chem., 1987,325,3 11. 9 R. L. DeKock, E. J. Baerends, P. M. Boerrigter, and R. Hengelmolen, J. Am. Chem. SOC.,1984,106,3387 10 G. van Koten, J. G. Noltes in Comprehensive Organometallic Chemistry, G. Wilkinson, F. G. A. Stone, E. W. Abel, editors, Pergamon, Oxford 1982, p. 709 ff. 11 A. Tamaki and J. K. Kochi, J. Organomet. Chem., 1973,61,441. 12 M. I. Bruce, K. R. Grundy, M. J. Liddell, M. R. Snow, E. R. T. Tiekink, J. Organomet. Chem., 1988,344, C49. 13 M. I. Bruce, E. Horn, J. G. Matisons, and M. R. Snow, Aust. J Chem., 1984,37, 1163. 14 T. V. Baukova, Y. L. Slovokhotov and Y. T. Struchkov, J. Organomet. Chem., 1981, 221 375. 15 C. J. Gilmore and P. Woodward, J. Chem. Soc., Chem. Commun., 1971,1233. 16 C. M. Mitchell and F. G. A. Stone, J. Chem. SOC.Dalton, 1972, 102. 17 D. E. Harwell, M. D. Mortimer, C. B. Knobler, F. A. L. Anet, and M. F.Hawthorne, 3: Am. Chem. SOC.,1996,118,2679. 18 T . V. Baukova, L. G. Kuz'mina, N. A. Oleinikova, D. A. Lemenovskii, A. L. Blumenfel'd, J. Organomet. Chem., 1997,530,27. 19 C. L. Raston, G . Salem, J. Chem. SOC.,Chem. Commun., 1984, 1702. 20 S . Li and Y. Jiang. J. Am. Chem. Soc.. 1995.117.8401

a,rnBis[(triphenylphosphine)gold(I)] hydrocarbons

85

21 N. C. Baenziger, W. E. Bennett, and D. M. Soboroff.Acta Crystallogr., Sect B, 1976, 32,962. 22 D. Westland, Can. J Chem., 1969,47,4135. 23 Y. Matano, H. Kurata, T. Murhji, N. Afluna, and H. Suzuki, Organometallics, 1998, 17,4049. 24 D. J. Chadwick and C. Willbe, J Chem SOC.Perkin I, 1977,887. 25 G. M Sheldrick, SHELX-97, Program for Solution and Refinement of Crystal Structures, UniversiGt Gottingen, Germany, 1 997.

RESEARCHES ON NON-CLASSICAL ORGANOLANTHANIDECHEMISTRY

P.B. Hitchcock, A.G. Hulkes, A.V. Khvostov, M.F. Lappert and A.V. Protchenko The Chemistry Laboratory, University of Sussex, Brighton, BN1 9QJ, UK

1 INTRODUCTION This paper deals with the following five themes, with emphasis on unpublished work. The following abbreviations are used in the text: R = SiMe3; Cp" = q5-C~H3(SiMe3)2-1,3; Cp" = q5-CsH3Bu'2-l ,3; Cp' = q5-Cs&SiMe2But; L = [{N(R)C(Ph)}2CH]. 1. Approaches to organoLn(I1) cyclopentadienyls(Ln = Sc, Y, La, Ce, Pr, Nd) 2. Reactions of a mononuclear arene (C6H6, PhMe, C6&R2-1,4, C6H2R4-1,2,4,5) with K in the presence of (18-crown-6) 3. Cationic organoLn(I1) complexes (Sm, Yb; Cp", NR2) 4. New Ce(IV) molecular complexes 5. Subvalent Sm and Yb P-diketiminates 2 RESULTS 2.1 Approaches to OrganoLn(I1) Cyclopentadienyls(Ln = Sc, Y, La, Ce, P r, Nd) 2.1.1 Background. A range of compounds [LnCpX3](Ln = La, Ce, Pr, Nd; Cp" = Cp", Cp", Cp') was used in studies reported in the period 1995 - 2000;' - these were recently reviewed.' The more significant observations, illustrated for Ln = La, are summarised in Scheme 1 (similar results have also been reported for Ln = Ce, Pr, Nd). In the cyclic voltammetric studies, it was also shown that El12 for the reduction of [LnCp'3] is - 3.10V, relative to the [FeCp2]+/[FeCpz]couple in thf at 25 0C.3It is noteworthy that the bis(sily1)substituted substrate [LaCp"3] was more readily reduced than the carbon analogue [LnCp'3], consistent with the notion that the silyl substituents help to delocalise the negative charge in [LaCp"s]-.

'

2.1.2 Synthesis and Characterisation of Tri(cyclopentadienyl)lanthanate(I.s. The reactions of a [LnCp"3] complex with potassium in the presence of a crown ether were carried out using either benzene or toluene as "solvent" (items 3 - 5 in Scheme 1):. 4v We now show that changing the reaction medium to diethyl ether or thf resulted in a different outcome, affording the crystalline compounds 1 - 3 in 50 - 80% yields (eq. (1)). Their

Researches on non-classical organolanthanide chemistry

87

structures were confirmed by X-ray crystallography for 1 and 2 (Figure 1 for 2; average a-Cp" centroid, 2.62 A, cf. av. Ce-Cp" centroid, 2.57 9), EPR spectroscopy for 2 (octuplet with gav= 1.996, u ( ' ~ ~ = L 133.5 ~ ) G in EtzO) and magnetic moment for 2.*

l.3 CV

-

rel. to [FeCp2lf/[FeCp2]):

[LaCp"3] -2.80; [LaCp'3] -3.10 V

2.'2 [LaCptt3]+ 2 K + dme [{LaCp"2(p-OMe))2]+ C2H4 + 2KCp" via [La1'Cp"3]-, [La"c~"~] : EPR ('39La, I = 7/2)

Scheme 1 Selected results appertaining to reduction of various compounds [LaCp'J.

+ ]K [Lnc~"~

+ L*

-

[K(L*)][LnCptt3] 1 Ln = La, L* = (18-crown-6), blue-violet 2 Ln = La, L* = cryptand-2,2,2, blue-violet 3 Ln = Ce, L* = (18-crown-6), violet

Figure 1 The X-ray structure of [K(cryptand-2,2,2)][LaCp'$12

(1)

88

Perspectives in Organomerallic Chemistry

2.1.3 Evidence for Yttrium and Scandium Ln(I4 Transients. X-Ray-characterised [YCp"3] 4 was prepared from [{YCp"2(p-I)}2] and KC in refluxing toluene, but the corresponding Sc(II1) iodide 5 proved to be unreactive.g)'kttttempts were made to reduce each of 4, 5 or [{YbCpf'2(p-C1)}2] with K in the presence of an arene at ambient

K + 18-crown-6 + ArH

[YbCp"2(Cl)(p-CI)K( 1 8 - ~ r o w n - 6 ) ] ~Y(II1) ~ hydrides 6 Orange, 43%, X-ray

J('H-89Y) 32.2 Hz ( t )

[(SCCP"2)2(P-C2H,)I 7 Brown, < 5%, X-ray

Scheme 2 Reactions of LnCp '5Xsubstrates with K,(18-crown-6)and an arene

temperature, Scheme 2. The ytterbium compound yielded the heterobinuclear Yb(I1) complex 6." We deduce that both 4 and 5 were at least in part reduced to form transient Ln(I1) intermediates, which in the Y case abstracted a hydrogen atom to yield a mixture of Y(II1) hydrides.* Compound 7, isolated in low yield, was diamagnetic and hence is regarded as Sc(II1) complex containing a bridging [C2&l2- ligand, which is presumed to have arisen from radical cleavage of the crown ether. Its X-ray structure is shown in Figure 2, Z(C23-C23') 1.465(5) A, Z(Sc-C23) 2.314(3) A. There are other examples of bridging [CZH$ complexes; the first was [{ZrC13(PEt3)2}2(p-C2&)], having Z(C-C) 1.469(3) A, obtained from [{ ZrC13(PEt3)2}2] + C2H4.l 1

Figure 2 X-Ray structure of [(ScCp '5)2(p-C2H4)] 7

Researches on non-classical organolanthanide chemistry

89

2.1 Reactions of Mononuclear Arenes with Potassium in the Presence of (18-crown-6)

2.2.I Background. It was established in the 1990s that two poly(trimethylsily1)benzenes yielded X-ray-characterised crystalline roducts: [Li(thf)12[C6%]~'~ [Li(dme)l2[C~H2h-1,2,4,5]! and [Na(drne)3][C6H2&-1 ,2,4,5].'4 We recently reported the synthesis and X-ray structures of [K( 18-crown-6)(PhMe)] 8 and [{K( 1~-C~OWII-~)}~(~A-I~~:?~~-C~H~-C~H~)] 9.15Although ["'" the Csp3-CSp3 bond in 9 was short, 1.46(1) A, nevertheless 9 0 9 in benzene solution was in equilibrium with [K( 18-crown6 ) ( c 6 ~ ).I51 2.2.2 Reactions of Ca2&-1,2,4,5 and c&4&-1,4.' Some reactions involving C&h-1,2,4,5 (= ArH) are summarised in Scheme 3. X-Ray-quality crystals of the [ArHI2- salt 10 were not obtained, but 11 (anion in Figure 3) and 12 were X-raycharacterised. From C6H4R2-Iy4with potassium and (18-crown-6) in toluene, crystalline [K( 18-crown-6)(C6H&-l,4)] 13 (Figure 4) was produced. The anion in 11 appears to have its negative charge delocalised in the 1, 6, 5 and 2, 3, 4 allylic fragments (numbering as in Figure 3).

* [K( 18-~rown-6)]~[ArH]

2K + ArH + 2( 18-crown-6)

Recrystn. fiorn THF-Et20

[LaCp"2(C1)(p-Cl)K(18-crown-6)I 12 [K( 18-~rown-6)(thf)~][ArH] i[K( 18-crown-6)][LaCp"2(K2-C6H6)] 11, X-ray -t[K( 18-CrOWn-6)(~2-c6~)2][(Lacpt'~)2(jA-~6:r)6-c~~)] Scheme 3 Reactions involving Ca2&-1,2,4,5 (=ArH)

1.462

Me3%

I

I C6

Figure 3 Selected bond lengths (A) and angles 11 (th~~][CSH2R4-1,2,4,5]

51me3 (4 in the anion of [K(l8-crown-6)-

Perspectives in Organometallic Chemistry

90

Figure 4 X-Ray structure of the [K(18-crown-6)(C~H~R2-lt 4)] 13

2.3 Cationic OrganoLn(I1) Co mplexes (Sm, Yb; Cp", NR2) 2.3.1 Background. Treatment of [LnCp"l] with (18-crown-6) in benzene at 20 "C afforded the crystalline complexes [SmCp"( 18-crown-6)][SmCp"3] for Ln = Sm and [YbCp"(l8-crown-6)][Cp"] for Ln = Yb; the latter in toluene solution showed fluxional behaviour probably involving neutral [YbCp"2( 18-~rown-6)].'~ An attempt to reduce the former with potassium in toluene at -30 "C afforded [K(18-crown-6)][SmCp"3].

2.3.2 Approaches to Cationic Yb(ll) Amides." The reactions of two ytterbium(I1) amides with (18-crown-6) are summarised in Scheme 4. The salt 15 is of particular interest, being the first cationic Yb(I1) amide. Also shown in Scheme 4 is the structure of the crystalline [K(NPh2)(18-crown-6)] 16, obtained from K(NPh2) and( 18-crown-6) in thf; the interesting feature of the structure of 16 is that the K ion coordinates to one of the Phrings rather than the amido nitrogen atom.

[Yb(NR'2)2(thf)J + (18-crown-6) R' = S M e 3 = R, n=2

[Yb(N(SiMe,),}( 18-~rown-6)][Yb{N(SiMe,)~}~] 15 blue, (X-ray);

14 yellow, X-ray

Researches on non-classical organolanthanide chemistry

91

KNPh, + (18-crown-6)

\

[K( 1 8-crown-6)NPh2] 16

Scheme 4 The synthesis of Yb(I4 amides 14 and 15 and the synthesis and structure of the potassium amide 1617

2.3.3 A Remarkable Reaction of [YbCp'yl8-crown-6)][Cp '7 and Other Observations on the Facile Desilylation of a [Cpy- Ligand.8 The reaction of [YbCp"(lS-crown6)][Cp"] l6 with [YbCp"2] in benzene, unexpectedly, afforded not only [YbCp"( 18-crown6)][YbCp"3] but also as a minor product the complex, containing the unsubstituted bridging Cp-ligand, [YbCp"( 1S-crown-6)][(YbCp"2)2(p-C~H5)] 17 (Figure 5). The X-raycharacterised compound 18, containing the same unexpected anion, was obtained in acceptable yield as shown in eq. (2), and on hydrolysis gave Cp"H + C5H6.

18 In order to shed light on this apparently facile desilylation reactions, it was shown by NMR spectroscopic experiments in C6D6 that whereas the salt KCp" in presence of a trace of water yielded only Cp"H, the crown ether adduct [K(18-crown-6)Cp"] gave C5Hs + C5H5SiMe3 + (Me3Si)zO.

x

03 Figure 5 The X-ray structure of [YbCp'~l8-crown-6)][(YbCp'i).&K~H~)] 1 78 2.3 New Cerium(1V) Complexes 2.4.I Background. The majority of well characterised Ce(1V) organic compounds have Ce-0 bonds. Examples include [Ce(OBut)2(N03)2(HOBu')] and [Ce{(cC6H11)gSig0,3}2(~y)3].~~ They have rarely been obtained by oxidation of a Ce(II1) substrate, but an important exception is {Ce(OCBUt3)3}2(~-OC6H40-1,4), obtained by

92

Perspectives in Organometallic Chemistry

oxidation of [C~(OCBU‘~)~] with 1,4-ben~oquinone.~~ Although the inorganic Ce(1V) salt [NH&[Ce(N03)6] (CAN) is used as a reagent in organic synthesis?l the similar use of Ce(1V) organic compounds has not yet been explored. The first Ce-0-free Ce(1V) organic compound was A (R’ = SiMe2But),prepared from the tris(amido)aminecerium(III) complex [Ce(N”3)] p 3 = N(CH2CH2NSiMe2But)3]and 12, whereas use of Cl2 or Br2 generated the mixed valence compound R\ ,R’ [{ Ce(N”3)2}2(p-Hal)].22 By contrast [Ce(NR2)3] was not oxidised by N-.,:ce-~ Cl2, but crystalline [CeCl(NR&] B was obtained in modest yield, Scheme 5 , using TeC14.23The geometric data for B were in reasonable RlJ$&l A agreement with those computed for the model compound [CeCl{N( SiH3)2}3].23

1

2.4.2 Syntheses and Structures of New Amidocerium(IV) Complexes. Further oxidation reactions of [Ce(NR2)3] with T e Q (X = C1, Br) or PBr2Ph3 have been explored, yielding the bromocerium(1V) amide 19 and the new X-ray-characterised crystalline Ce(II1) compounds 20,21 and 22, Scheme 5?4 It may be significant that the rather unusual oxidants Te& and PBr2Ph3 dissociate in solution to form halogenonium ions [TeX3]+and [PBrPh3]+. The X-ray structures of the isomorphous complexes B 23 and 19 24 are illustrated in Figure 6. Two oxidative reactions of [Ce(NR2)3] with molecular oxygen and a bulky quinone are illustrated in Scheme 6, which yielded the novel Ce(1V) complexes 23 (Figure 7) and 24 (Figure 8).26It is possible that an intermediate along the pathway to 24 was Ce(NR2)3(OH). X-Ray structures of compounds 23 and 24 are shown in Figures 7 and 8. In 23 the average

[CeBr(NR2)3] + CeBr3(0Et2), 19, purple, 30%

TeBr4

B ,23 purple, 24%

t [CeBrdthf)d 20, colourless, 33%

19 (low yield)

21, colourless, 27%

+ [CeBr2(NR2)(thf13] 22, colourless, 12%

Scheme 5 Reactions between [Ce(NRj3] and TeX4 (X = Cl, Br) or PBr2Ph3 24

Ce-WA Ce-NIA N-Ce-N’/” Si-N,/A

Figure 6 The X-ray structures of [CeX(NR2)3 X selectedparameters with those for [Ce(NR2)3] 15 (

B 23 2.597(2) 2.217(3) 117.3(4) 1.75l(3)

=

19 24 [Ce(NR2)3] 25 2.766(2) 2.219(7) 2.320(3) 117.25(12) 118.24(4) 1.751(8) 1.702(2)

Cl B 23, Br 19 2.‘) and comparison of

Researches on non-classical organolanthanide chemistry

93

Ce-0 and Ce-N and 0-0 distances are 2.334, 2.261 and 1.47(2) A, respectively, while in 24, the Ce-0 and Ce-N average bond lengths are 2.093 and 2.268 A, respectively, and the average 0-Ce-0’ and Ce-0-Ce’ bond angles are 143.7 and 95.8”, respectively. But

23, black

Scheme 6 Syntheses of the amidoceriurn(IV) 0x0 complexes 23 and 24 26

Figure 7 The X-ray structure of [{Ce(”R$$2(p-02)] 23 26

Figure 8 The X-ray structure of [{Ce(NR2)2(p-0))3]24 26 2.4.3 Syntheses and Structures of Ce(II4 and Ce(Iv Dithiocarbarnates.27Treatment of [Ce(NR2)3] with tetraethylthiuram disulfide yielded the Ce(II1) dithiocarbamate 25, which with 2,2’-bipyridine or 0 2 afforded the crystalline bipy adduct 26 or the novel Ce(1V) dithiocarbamate 27 (Figure 9), Scheme 7, which also shows a superior method of generating 25. The average Ce-S bond lengths and av. S-Ce-S’ bond angles in 25 or 26 are 2.812 or 2.931 A and 63.05 or 60.8”,respectively.

94

Perspectives in Organometallic Chemistry

biPY

[Ce(~~-S~CNEt~),(bipy)] 26, brown

27, black

Scheme 7 Syntheses of the Ce(II0 and Ce(Iv) dithiocarbamates 25 - 2727

Figure 9 The X-ray structure of [ce(d-S~Ch?EtsJ2727

2.5 Subvalent Sm and Yb P-Diketiminates 2.4.1 Background. The area of metal P-diketiminates is generating much current interest.28From small beginnings in 1960s, the area is burgeoning, there having been 38 publications in 2001 out of total to that date of ca. 140 and 28 in the first R3 C half of 2002. About 500 metal p-diketiminates have now been reported \CR4 for 43 different metals and in all, except one the ligand has been monoanionic, i.e. I, although it has been observed in diverse NR5 bonding modes;' not only in the n-delocalised form shown in I. R'N The complex [Yb((N(R)C(Ph))2CH}2]has been reported 30 and was I used as a starting material for one of the experiments to be described in Section 2.5.2. Its preparation from YbI2 and the potassium p-diketiminate is unexceptional.

'

"in I

2.4.2 Syntheses and Structures of Unusual Bi- and Trinuclear Sm and Yb pDiketiminates. The P-diketiminates to be discussed in this section have the formula [ (N(R)C(R1)}2CH]" (R' = Ph or CsH4-Ph-4, abbreviated as L"- or L'"-, respectively). Evidence will be presented to show that p-diketiminates can be found when metal-bound not only as monoanionic ligands I, but also as di- and trianionic analogues I1 and I11 (shown for L), respectively. The initial objective of this study was to gain access to Ln 9diketiminates for Sm and Yb in oxidation states <+2.

Researches on non-classicalorganolanthanide chemistry

95

H p

T

w

R

1

a c & , P h

I

RN

NR I1

I -

I

RN

NR

I

NR

I11

The preparation of the novel P-diketiminato complexes of samarium and ytterbium 28 - 32 is summarised in eqs. (3),29(4)," and (5).* The X-ray structures of compounds 28,30, 31 and 32 are illustrated in Figures 10 - 13, respectively.

29 Lx=L'

- KL, EtZO

KL', Et,O

[(SmL')2(p-L')] 30

Sm12

[(KSmLL)21 31

(4)

lo

In 28 and 29 the average C-N and C-C bond distances of the P-diketiminate backbone are 1.413 8, for Cl-Nl (or C3-N2) and 1.427 A for Cl-C2 (or C2-C3). These bond lengths may be compared with those in [YbL2] of 1.326 8, (C-N) and 1.412 8, (C-C), respectively." This is consistent with the notion that 28 is best formulated in terms of Yb2+bonded to L2- (cf. 11), with the x* orbital IV of the NCCCN moiety being sufficiently accessible to accept an electron (L- 3 L2-), resulting in a longer C-N bond than in IV L- but a "normal" C-C bond.

w

Figure 10 The

96

Perspectives in Organometallic Chemistry

Alternative assignments for 28 and 29 as Yb(0) and L- is also negated by the observation that they are paramagnetic; Yb(0) would have the s2 configuration and hence these assignments would require 28 and 29 to be diamagnetic.

Figure 11 The X-raystructure of [(SmL 32(p-L3 130 lo

Figure 12 The X-raystructure of [(KSmLj2] 31 lo

Researches on non-classical organolanthanide chemistry

97

Compound 32 consists of an (YbL)2 dimer with both of the diketiminato ligands bridging the two Yb atoms (Ybl and Yb2) and an YbL(thf) moiety (Yb3) connected to the former via q5-coordination to a Ph-ring (Figure 13). Evidence for the assignments (Yb oxidation states and formal charge on ligands) shown in Figure 14a is based on the paramagnetism of 32 and the bond length data as summarised for the L- and L3(exemplified by the N1 nN2 and N5 nN6 ligands in Figure 14b). Moreover the Cg rings ato N1 and N3 atoms are almost coplanar with the adjacent Nl-Cl-C2 and N3-C22-C23 fragments, respectively, whereas those a-to N2 and N4 atoms are close to orthogonal (the torsion angles are 83 and 53O, respectively).'

Figure 13 A schematic representation of the X-ray structure of [YbL(thj(YbL)2] 32 for clarity some substituents are omitted)

Perspectives in Organometallic Chemistry

98 1.392

Phe+fph RN:

1.365

N~R

/

1.309

I N6

N5

1.394

1.401 1.386

Ybl: Yb”

NlflN2: L3-

Yb2: Yb”

N3flN4: L3-

Yb3: Yb”

N5nN6: L-

1.412

(a)

I

N2

N1

(b)

Figure 14 Bonding (a) and selected bond lengths (b) in [YbL(thj(YbL)2] 32 A recent report has described the synthesis and structure of a compound of composition [{ Mg(Br)L”)zScBr] (C)formulated as a Sc’Br species bridged by two neutral Mg(Br)L” moieties, while an alternative assignment [Mg(Br)L”]-2Sc3+Br-was regarded as unlikely; L“ = [{N(CH2CH2NEt2)C(Me)}2CH].31 We agree with the latter conclusion but suggest that the assignment [ M ~ B ~ ] ~ [ S C ” ’ B ~ ( is L ”worthy ~ - ) ~ ] of consideration.

c 31 Acknowledgement We are grateful for the award of fellowships to EPSRC (A.G.H.), the Royal Society (A.V.K. and Z.L.) and the Leverhulme Trust (A.V.P.).

Researches on non-classical organolanthanide chemistry

99

References

Yu.K. Gun'ko, P.B. Hitchcock and M.F. Lappert, J. Organomet. Chem., 1995,499, 213. M.C. Cassani, Yu.K. Gun'ko, P.B. Hitchcock and M.F. Lappert, Chem. Commun., 1996,1987. M.C. Cassani, M.F. Lappert and F. Laschi, Chem. Commun., 1997, 1563. M.C. Cassani, D.J. Duncalf and M.F. Lappert, J. Am. Chem. SOC., 1998, 120, 12958. M.C. Cassani, Yu.K. Gun'ko, P.B. Hitchcock, M.F. Lappert and F. Laschi, Organometallics, 1999, 18, 5539. Yu.K. Gun'ko, P.B. Hitchcock and M.F. Lappert, Organometallics, 2000,19,2832. M.C. Cassani, Yu.K. Gun'ko, P.B. Hitchcock, A.G. Hulkes, A.V. Khvostov, M.F. Lappert and A.V. Protchenko, J. Organomet. Chem., 2002,647,71. P.B. Hitchcock, M.F. Lappert and A.V. Protchenko, unpublished work. S.D. Stults, R.A. Andersen and A. Zalkin, Organometallics, 1990,9, 115. 10 P.B. Hitchcock, A.V. Khvostov and M.F. Lappert, unpublished work. 1 1 F.A. Cotton and P.A. Kibala, Inorg. Chem., 1990,29,3192. 12 A. Sekiguchi, K. Ebata, C. Kabuto and H. Sakurai, J. Am. Chem. SOC.,1991, 113, 1464. 13 A. Sekiguchi, K. Ebata, C. Kabuto and H. Sakurai, J. Am. Chem. SOC.,1991, 113, 708 1. 14 H. Bock, M. Ans ari, N. Nagel and Z. Havlas, J. Organomet. Chem., 1995,499,63. 15 P.B. Hitchcock, M.F. Lappert and A.V. Protchenko, J. Am. Chem. Soc., 2001,123, 189. 16 Yu.K. Gun 'ko, P.B. Hitchcock and M.F. Lappert, Chem. Commun., 1998, 1843. 17 P.B. Hitchcock, A.V. Khvostov, M.F. Lappert and A.V. Protchenko, J. Organomet. Chem., 2002,647,198. 18 W. J. Evans, T. J. Deming, J. M. Olofson and J. W. Ziller, Inorg. Chem. 1989, 28, 4027. 19 Yu.K. Gun'ko, R. Reilly, F.T. Edelmann and H.-G. Schmidt, Angew. Chem., Int. Ed. Engl., 2001,40, 1279. 20 A. Sen, H.A. Steche r and A.L. Rheingold, Inorg. Chem., 1992,31,473. 21 V. Nair, J. Mathew and J. Prabhakar an, Chem. SOC.Rev., 1997,26, 127. 22 C. Morton, N.W. Alcock, M.R. Lees, I.J. Munslow, C.J. Sanders and P. Scott, J. Am. Chem. SOC.,1999,121,11255. 23 0. Eisenstein, P.B. Hitchcock, A.G. Hulkes, M.F. Lappert and L. Maron, Chem. Commun., 200 1, 1560. 24 P.B. Hitchcock, A.G. Hulkes and M.F. Lappert, unpublished work. 25 W.S. Rees, 0. Ju st and D.S. Van Derveer, J. Muter. Chem., 1999,9,249. 26 P.B. Hitchcock, M. F. Lappert and Z. Li, unpublished work. 27 P.B. Hitchcock, A.G. Hulkes, M.F. Lappert and Z. Li, unpublished work. 28 L . Bourget-Merle, M.F. Lappert and J.R. Severn, Chem. Rev., 2002,102,3031. 29 A.G. Avent, A.V. Khvostov, P.B. Hitchcock and M.F. Lappert, Chem. Commun., 2002,1410. 30 P.B. Hitchcock, M.F. Lappert and S. Tian, J. Chem. Soc., Dalton Trans., 1997, 1945. 31 A.M. Neculai, D. Neculai, H.W. Roesky, J. Magull, M. Baldus, 0. Andronesi and M. Jansen, Organometallics, 2002,21,2590.

HYPER-STRUCTURED ALKYNYLRUTHENIUM COMPLEXES: EFFECT OF DIMENSIONAL EVOLUTION ON NLO PROPERTIES Mark G. Humphrey,l Marie P. Cifuentes,l Marek Samoc> Takashi Isoshima? and h d r k ~ersoons4 1 Department of Chemistry, Australian National University, Canberra, ACT 0200, Australia (Email: [email protected]) 2 Australian Photonics Cooperative Research Centre, Laser Physics Centre, Research School of Physical Sciences and Engineering, Australian National University, Canberra, ACT 0200, Australia 3 RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan Centre for Research on Molecular Electronics and Photonics, Laboratory of Chemical and Biological Dynamics, University of Leuven, Celestijnenlaan 200D, B-3001 Leuven, Belgium

1 INTRODUCTION Materials with desirable nonlinear optical (NLO) properties can provide an efficient means of controlling and modifylng the nature of light beams used in a variety of photonic applications (optical processing, switching, frequency generation, data storage, optical communication, optical computing). As a consequence, the NLO properties of new materials have come under increasing scrutiny. Early attention was focussed on inorganic materials, which have high thermal stability, are optically transparent, and are readily available as large crystals or glasses. However, they can have slow response times (where the NLO effect arises from lattice distortions or where the dynamics of charge carriers are involved, this can be of the order of nanoseconds), can be difficult to process (in instances where thin film integrated structures are desirable), and have little architectural flexibility. In comparison, organic materials are attractive: The NLO effects arising from electron polarization have response times of the order of femtoseconds, organic materials can be readily processed to be cast as thin films, and organic compounds possess structural diversity and architectural flexibility which is useful for optimizing NLO responses. Organic compounds suffer from some disadvantages, though, including lower thermal stability than inorganic complexes, and (for dipolar examples) an optical transparency / NLO efficiency trade off. Increasing attention has recently been given to organometallic complexes.1-3 Organometallic complexes have the advantages of organic compounds, but possess significant additional design flexibility (the possibility of varying metal, oxidation state, ligands, and coordination geometry). The most intensively studied organometallic complexes to date are ferrocenyl derivatives and alkynylmetal compounds, both of which are thermally robust, oxidatively stable, and accessible in high yields by well-established methodologies. Examination of ferrocenyl complexes preceded evaluation of alkynyl compounds, impetus to investigation of the alkynyl complexes being given from the suggestion that incorporation of the ligated metal into the plane of the organic Ic-system, and the presence of some Ic-bonding between metal and organic Ic-ligand, should result in complexes with enhanced quadratic NLO merit.43 A brief outline of NLO properties is presented here to clarify the subsequent discussion. At the molecular level, NLO responses can be attributed to changes in the electron polarization of the medium. Molecular polarization is described by the molecular dipole moment expansion,

Hyper-structured alkynylruthenium complexes

p=

101

+ a E + pE2 + y E 3 + ...

where is the intrinsic dipole moment of the molecule and E is the oscillating electric field strength. The coefficients a,p and y are the linear polarizability and the first (quadratic) and second (cubic) hyperpolarizabilities (first-, second- and third-order molecular responses), respectively. This explanation is somewhat simplified: The a,p and y terms are tensors, and the p and E terms are vectors, and all of the variables should be treated as either time-dependent or frequency-dependent. The p term is important for frequency mixing (e.g. second-harmonic generation, sum and difference frequency generation) and electro-optic applications (e.g. photorefiactive applications such as real-time holography). The y term should in general be treated as complex, and can be deconvoluted to real (ymal) and imaginary (%mag) components, with the former corresponding to nonlinear refraction and the latter to nonlinear absorption. Nonlinear refraction is of significant interest for alloptical switching of signals, while nonlinear absorption is of interest in optical limiting (e.g. the protection of sensors from high-energy laser pulses). Two-photon absorption (TPA) is an important nonlinear absorption process of interest in multiphoton microscopy and optical data storage. The TPA performance of molecules is generally quantified as the TPA cross-section 02, the trends in which follow those of yimag for the results summarized herein. Both quantities are proportional to each other, assuming that one is concerned only with the true, instantaneous two-photon absorption and not with other processes mimicking the behaviour of TPA on a longer time scale, such as excited state absorption. A two-level model, involving the ground state and a single charge-transfer excited state, has been developed to understand trends in p for systematically varied organic molecules.6 Although it is no longer considered completely adequate for organometallic complexes, this model does reveal the structural features which are of importance for p. In this model,

where pge is the transition moment, k - pgg is the difference in dipole moments of the ground and excited states, and Ege is the difference $I energy between the ground and excited states. Thus, if this model has some utility, complexes with an intense (large pge) low energy (small Ege) charge-transfer (large k - pgdtransition are expected to have significant p values. Not surprisingly, dipolar complexes with a donor-bridge-acceptor composition have been the focus of most studies. A related expression for a three-level model, involving the ground state and two excited states, has been developed to understand trends in y

The two-level and three-level models outlined above suggest dipolar compounds should have significant NLO properties. Our earlier studies focussed on developing structure / NLO property relationships for dipolar alkynylmetal complexes.7-10 Trends observed with organic systems are also seen with donor-bridge-acceptor alkynyl complexes. Thus, introduction of strong acceptor group and bridge lengthening and modification (progressing

Perspectives in Organometallic Chemistry

102

from 2 to E stereochemistry, and imino-linked to ene-linked to azo-linked) result in an increase in nonlinearity. Replacing coligand PMe3 by PPh3 at the donor metal centre results in an increase in nonlinearity; the additional donor strength of the former is less important than the additional n-delocalization possibilities of the latter. Of greatest interest for organometallics is that quadratic nonlinearity can be increased by increasing metal valence electron count and ease of oxidation; nonlinearities of 14 valence electron (phosphine)gold alkynyl complexes are lower than those of analogous 18 valence electron cyclopentadienyl(phosphine)nickel or cyclopentadienylbis(phosphine)ruthenium alkynyl complexes, the last-mentioned being the easiest to oxidize and having the largest p coefficients. The ligated ruthenium centre is a more potent donor and affords compounds with larger nonlinearities than those containing the classical organic donor groups, dialkylor diarylamino. Cubic nonlinearities for these organometallic complexes are in most instances small, the most important determinant of their magnitude being the extent of the n-electron delocalization.9,11,12 As mentioned above, dipolar molecules suffer from an NLO efficiency / transparency trade-off (increasing nonlinearity is associated with a red-shift of the important linear optical absorption band). However, as odd higher-order multipoles may also contribute to p, octopolar compounds can also possess significant NLO properties while maintaining optimum transparency. 1,3,5-Trisubstituted aromatic ring systems have a three-fold symmetry axis and are octopolar; we have therefore commenced a study of the NLO properties of octopolar and dendritic alkynylruthenium complexes derived from a 1,3,5triethynylbenzene core, the results from which are summarized herein (ruthenium has been chosen as it had afforded complexes which were the most NLO-efficient in our dipolar studies). Also reported herein are linear analogues of the octopolar compounds, the data from which are contrasted with that from the octopolar examples to assess the effect of "dimensional evolution" (progressing from 1-D linear to 2-D octopolar) on NLO properties. Finally, we also summarize here our preliminary studies of electrochemical and protic "switching" of quadratic and cubic nonlinearity in these complexes. 2 METHOD AND RESULTS 2.1 Syntheses of Octopolar and Dendritic Alkynylquthenium Complexes

Our initial attempts to prepare octopolar alkynylmetal complexes utilized 1,3,5triethynylbenzene (Scheme 1). However, while chloro(triphenylphosphine)gold(I) reacted cleanly to afford 1, reaction with excess cis-[RuC12(dppm)2]proceeded to afford the bisadduct 2 only. X-ray structural studies confirmed that steric constraints had restricted the extent of reaction.13 PPh3

xs [AuCI(PPh3)] Au PhBP'

(I)

xs cis-[RuClp(dppm)d IRul

CI'

PPh3 Scheme 1

(2)

H

[Ru] = Ru(dppm)a

We therefore inserted "spacer" units between the 1,3,5-triethynylbenzene core and the ligated metal units.Sonogashira coupling of 1,3,5-triethynylbenzenewith three equivalents of 1-iodo-4-(trimethylsilylethynyl)benzene afforded 3 in good yield, subsequent

103

Hyper-structured alkynylruthenium complexes

desilylation with fluoride giving 4 (Scheme 2).14 The "ene-linked" analogue 6 was prepared from 5 by an Arbuzov, Emmons-Horner, Sonogashira, and desilylation sequence (Scheme 2). 15,16 Linear alkynes were prepared similarly. H

Sik3

1

p$

H+

(4)

3H

Scheme 2

Coupling the alkynes to bis(bidentate ph0sphine)ruthenium cores was readily accomplished by Dixneufl7 protocols, to give the octopolar alkynylruthenium complexes 7-10.14-16,18 Bis-alkynyl complexes are readily accessible when the bidentate phosphine is dppe, and examples with phenylalkynyl incorporating electron donating (NEt2) and accepting (N02)substituents have been prepared (Scheme 3).14-16

n

P P =dppm,dppe

I

CI (i) 6, NH4PF6, CH&Iz (ii)NEb

Scheme 3

104

Perspectives in Organometallic Chemistry

The facile bis-alkynyl complex synthesis of 11 - 13 suggests that 8 can function as the core of an alkynylnithenium dendrimer. Dendrimers can be prepared by convergent or divergent procedures, but the former generally results in increased ease of purification and is synthetically preferable. The convergent approach involves constructing appropriate "wedge" compounds (dendrons) and coupling them to the core at the final step. Dendrons for the present studies have been assembled using the aforementioned synthetic methodologies, the reaction sequence being displayed in Scheme 4. This has afforded the two first-generation dendrimers 15 and 16.19 The lengthy syntheses in Scheme 4 are not ideal when the intent is to promulgate physical property investigations, so we have recently developed a new and rapid

pa,,

Mg(C4H)Br Zn& c-Pd", NEb

I

9 Br &s,i

H+

.[Rf

Scheme 4

105

Hyper-structured alkynylrutheniumcomplexes

methodology for dendron synthesis. Although an undesired product from an attempt to trimetallate triethynylbenzene, complex 2 can also be considered as an archetypal organometallic dendron, with the "AB2" 1,3,5-trisubstituted benzene composition required for alkynylruthenium dendrimer construction. Replacing the bidentate phosphine dppm with dppe has permitted the reaction sequence shown in Scheme 5 , and thereby rapid access to nanometre-sized It-delocalized octopolar complexes.24 The utility of dendrons such as 17 in dendrimer synthesis has recently been confirmed20

I

scheme 5

2.2 Dimensional Evolution of Linear and Nonlinear Optical Properties

Access to this series of systematically-varied n-delocaliied linear, octopolar and dendritic alkynylruthenium compounds has permitted examination of the effect of dimensional evolution on NLO properties. Linear and nonlinear optical data are presented in Table 1. There is little loss of optical transparency in progressing fiom the linear hgments 18 - 21 to the two-dimensionally It-delocalized octopolar complexes 11 - 14; indeed, Lax for the nitro-containing example 12 is at lower wavelength than that for its linear fragment 19. Extending the n-system in proceeding from 8 to 11or 10 to 14 results in a small blue-shift in optical absorption maximum and increase in extinction coefficient. Increasing the size of the complex in proceeding fiom 11 and 12 to the first generation dendrimers 15 and 16 does not reduce optical transparency significantly; the small blue shift observed in proceeding fiom 11to 15 may indicate a lack of co-planarity through the dendritic structure of 15.

106

Perspectives in Organometallic Chemistry

Yimag 310 f 70 -760 f 120

300*loo

70i30

530i 100 700 f 120 1300*200 looOf200

All measurements in tM except where indiied. Complexes are optically tramparent at 1064 and 800 nm. a (nm). (104 M-1 awl). b (10-30 esu), hyper-Rayleigh scattering at 1064 nm, values f 10 %, referenced to p-nitroaniline (p = 21.4 x 1030 esu). c (10.36 esu), 2-scan at 800 nm, referenced to the nonlinear refractive index of silica (n2 = 3 x 1016 cm* W-I). d (10-50 Cnrc s), cdarlated using the equation 0, = where p is the Wo-ph&m absorption coeffiaent.21 e CH2C12solvent. f Not measured due to insufficient solubility.

trw,

Interestingly, the magnitude of the extinction coefficient, E, increases more than threefold in proceeding from linear fragments 18 - 21 to the three metal centre-containing octopolar complexes 11 - 14, and more than three-fold in progressing from the three metal centre-containing complex 11 to the nine metal centre-containingcomplex 15. The non-nitro-containing complexes have absorption bands far from the second harmonic wavelength of 532 nm, permitting assessment of the impact of structural variation on quadratic NLO merit. Proceeding from the linear fragment 18 to the octopolar complex 11 results in a three-fold increase in quadratic NLO merit, with little loss of optical transparency accompanying the large increase in p. Extending the delocalized nsystem through the metal in progressing fiom 8 to 11 is ineffective in increasing p, indicating that the trans-phenylalkynyl ligand is acting largely as a n-donor pseudochloride. The absolute values of p for 8, 11 and 15 are amongst the largest thus far for octopolar compounds optically transparent at the second-harmonic, for which resonance enhancement is much less important. They are also amongst the largest thus far for octopolar compounds lacking a formal acceptor moiety. The blue shift observed in the linear optical spectra in progressing from 8 to 11 to 15, and from 10 to 14, are suggestive of onset of lack of coplanarity. This can be probed by examining the depolarization ratio,

in which the numerator is the second-harmonic intensity parallel to the incident beam and

Hyper-structured alkynylruthenium complexes

107

the denominator is that perpendicular to the incident beam. The depolarization ratio depends on the symbetry of the molecule, with p values of 5 and 1.5 expected for linear (dipolar) molecules and three-fold symmetric (octopolar) molecules, respectively. Although complex 8 has the p value expected for its molecular symmetry (1.4 f 0.2), the depolarization ratio of complex 11 (2.1 f 0.1) clearly deviates fiom this symmetryidealized value, consistent with lack of coplanarity of the peripheral arylethynyl groups with the central triruthenium unit. It seems likely that the three-fold symmetric bis-alkynyl and dendritic ruthenium complexes are distorted from the idealized D3h symmetry in solution. The nitro-containing compounds 12 and 16 exhibit very large (resonanceenhanced) quadratic nonlinearities for three-fold symmetric molecules. Molecular cubic NLO parameters are also listed in Table 1. The significant 'yimag values for all complexes are indicative of two-photon absorption, which becomes important as Laapproaches 20. The negative 7-1 values for these compounds are therefore likely to result from two-photon dispersion effects, rather than being indicative of zero-frequency negative cubic hyperpolarizabilities. The data reveal a significant increase in upon progressing fiom the linear compounds 18 - 21 to their three-fold symmetry analogues 11 - 14. A further significant increase in nonlinearity is observed on proceeding from 11 and 12 to their dendritic analogues 15 and 16. The yred value for 16, ybg values for 15 and 16, and lg values for 15 and 16 are the largest for organometallic complexes thus far. The large increase in ynal,'yimg and 1in proceeding fiom the linear compound 21 to the branched compound 14 is particularly impressive.l5,16 The two-photon absorption cross-sections 0 2 of these complexes have also been calculated and are listed in Table 1; as the measurements were performed using 100 fs pulses, excited state absorption should be neghghle. The dendritic complexes 15 and 16 have the largest 0 2 values for organometallic compounds thus far, data which are comparable in magnitude to the largest values reported for organic molecules: it is likely that the size and two-dimensional nature of the n-delucalized systems in 15 and 16, combined with the strong MLCT transition, contribute to the large observed 0 2 values.19 We have applied electroabsorption spectroscopy to organometallic complexes to evaluate cubic nonlinearities for the first time, selected data (those for 10, 14 and 21) bang listed in Table 2.15,18 Small blue-shifts in Laare observed on proceeding from the solutions in thf, discussed above, to solid solutisns in poly(methy1 methacrylate) (PMMA). The imaginary components of electroabsorption-derived nonlinear susceptibility have been normalized by the number density of molecules in PMMA to cancel the influence of concentration.These Im(x(3))Mvalues attain a negative peak 20 nm Table 2 Experimental cubic nonlinear optical response parameters in solid solution in

PMUA measured by electroabsorption. limaxa

conpand R

Strucbe

I ~ ( x ( ~ ) ) MA, b (negPew

I ~ ( x ( ~ ) ) M1,b (positive Peak)

H

21

400

31m,420

36ooo,452

CI

lo

420

-120000,438

100000,472

CICP

14

418

-95o00,m

m,472

ph.c,k+Gc-(Ru~GcPt H I

-g

RW

'.R

h

a (nm). b (1W a) (nm); . in order

molecule N (M) in PMMA matrix.

to caned the iniluence of concentrabon * , XC" was normalized by the number density dbfl

108

Perspectives in Organometallic Chemistry

red-shifted from L a x , followed by a positive peak ca 50 nm red-shifted from Lax, the magnitudes increasing with the number of ruthenium centres. A detailed analysis of spectral profiles and anisotropy for 8, 10 - 16, 18 - 21 will be presented elsewhere; preliminary analysis indicates that the important excited states responsible for NLO properties in these complexes are localized for linear compounds and those possessing polarizing (NEt2, N02, Cl) groups, but are delocalized and octopolar in character for those without such groups.22 2.3 Switching Nonlinear Optical Properties An understanding of how to design molecular materials with significant NLO properties is

emerging. Attention has now moved to examining ways to control by external means or reversibly modulate optical non1inearities.Fromthe discussion in the Introduction, it is clear that procedures that modify L a x , E, and/or - clgs should modify NLO coefficients. Such modifications can conceivably be accomplished protically, electrochemically, or by photoexcitation. Metal complexes can frequently exist in more than one oxidation state, so should be good candidates for electrochemical switching. Alkynyl complexes are prepared by deprotonation of vinylidene complexes, a process that can be reversed to regenerate the vinylidene complexes. The protic and electrochemical routes to switching optical nonlinearity in alkynyl complexes are indicated below (for an o d o f f situation), both possibilities having been investigated in our preliminary studies (Scheme 6).

-

H+

-8

[M]-C=C-R1+ NL0"off"

[MI-C-C-R

R [M&C=C<~ NLO "off

NLO "on" Scheme 6

The facile interconversion of alkynyl and vinylidene ligands by protonatioddeprotonation sequences may offer a route into pH switching of both the quadratic and cubic NLO response. Examination of the linear optical spectra of an extensive series of dipolar complexes revealed that proceeding from a vinylidene complex to the analogous alkynyl complex results in little change in Lax (a small red-shift usually being observed), but an increase in E, the latter suggesting that protic switching of nonlinearities should be achievable.23 Decrease in nonlinearity upon protonation was observed for the majority of complexes studied, illustrative examplesbeing shown in Scheme 7:

b E

Y

Scheme 7

473nm 1.8 (104 M" an-') 290*60 (lO-=eSu)

Hyper-structured alkynylruthenium complexes

109

Chemical transformation (either protonatioddeprotonation as above, or chemical oxidation and reduction) needs to be performed remote from the optical bench, a less-thanideal approach to switching nonlinearity. Very recently, we have demonstrated in situ switching utilizing an optically-transparent thin-layer electrochemical (OTTLE) cell.18 Compound 8 undergoes reversible oxidation in solution, assigned to metal-centred RuImII processes. The UV-vis-NIR spectroelectrochemistry of 8 (Figure 1) reveals the resting state of the complex to be essentially transparent at frequencies below 20 x 103 cm-1 (wavelengths greater than 500 nm), whereas the oxidized form 83+ has a strong absorption band at 11 200 cm-1 (893 nm); oxidation therefore "switches on" a transition with appreciable intensity at the frequency of a Ti-sapphire laser.

10

20

30

40x103

Wmnmbw

Figure 1 W-vis-NIR spectra of a CH2Cl2 solution of 8 in an OTTLE cell, path length 0.5 mm, during oxidation at Eappl = 0.80 V a t 248K. As summarized above, octopolar and dendritic compounds such as 8 have significant 02 values at 800 nm. The sign of the imaginary part of the third-order nonlinearity is positive for TPA but, under strong one-photon absorption conditions in the oxidized form, absorption saturation is possible, reversing the sign of the nonlinearity, as displayed below:

a (nm).b (104 M-1 m 1 ) . c (1Wesu),Z-scan at 800 nm, referenced to the nonlinear refractiveindex

d silica (% = 3 x 1 W MI?

wl).d (I@ aTj,s), CalaJated usingthe equationq =AafYN, where p isthe hw-phbnabsorption coeffident.21

Oxidation therefore results in changes in both imaginary (absorptive) and real (rehctive) parts of the nonlinearity. Cycling between 8 and 83+ is facile. Smaller alkynylruthenium complexes with no detectable third-order NLO response have their nonlinearity "switched on" on oxidation. Current studies are addressing the demonstration of this effect at telecommunications wavelengths.

3 CONCLUSION The studies summarized above have involved development of synthetic procedures to rigid n-delocalizable dendrimers containing metals in the repeating units, the recent results suggesting that a more facile synthesis is achievable. The optical nonlinearities of these complexes i n c m significantly on proceeding from lincar to octopolar and then to

110

Perspectives in Organometallic Chemistry

dendritic complex, while little or no loss of optical transparency is observed; nonlinearities (quadratic and cubicJ are some of the largest for organometallic complexes thus far. The facile and reversible protonation and oxidation of these alkynyl complexes has been used to demonstrate switching of the NLO response, an effect of significant current interest. The electrochemical switching in the current work offers significant advantages over previouslydemonstrated approaches to NLO switching. References I.R. Whittall, A.M. McDonagh, M.G. Humphrey and M. Samoc, Adv. Organomet. Chem., 1998,42,291. 2 I.R. Whittall, A.M. McDonagh, M.G. Humphrey and M. Samoc, Adv. Organomet. Chem., 1999,43,349. 3 T. Verbiest, S. Houbrechts, M. Kauranen, K. Clays and A. Persoons, J. Muter. Chem., 1997,7,2175. 4 J.C. Calabrese, L.-T. Cheng, J.C. Green, S.R.Marder and W. Tam, J. Am. Chem. SOC., 1991,113,7227. 5 M.G. Humphrey, A.M. McDonagh, S. Houbrechts, I. Asselberghs, A. Persoons, T. Wada, H. Sasabe, M. Samoc and B. Luther-Davies, in Hyper-Structured Molecules III Chemistry, Physics and Applications, ed. H. Sasabe, Gordon and Breech, Reading, U.K. 2002, ch. 5,90. 6 J.L. Oudar and D.S. Chemla, J. Chem. Phys., 1977,66,2664. 7 I.R. Whittall, M.G. Humphrey, D.C.R. Hockless, B.W. Skelton and A.H. White, Organometallics, 1995,14,3970. 8 I.R. Whittall, M.G. Humphrey, S. Houbrechts, A. Persoons and D.C.R. Hockless, Organometallics, 1996,15,5738. 9 1.R Whittall, M.P. Cifuentes, M.G. Humphrey, B. Luther-Davies, M. Samoc, S. Houbrechts, A. Persoons, G.A. Heath and D. Bogsanyi, Orgunometallics, 1997, 16, 263 1. 10 I.R. Whittall, M.G. Humphrey, A. Persoons and S. Houbrechts, Organometallics, 1996,15,1935. 1 1 I.R. Whittall, M.G. Humphrey,’M. Samoc, J. Swiatkiewicz and B. Luther-Davies, Organornetallics, 1995,14,5493. 12 I.R. Whittall, M.G. Humphrey, M. Samoc and B. Luther-Davies., Angew. Chem., Int. Ed. Engl., 1997,36,370. 13 I.R. Whittall, M.G.Humphrey, S. Houbrechts, J. Maes, A. Persoons, S. Schmid and D.C.R. Hockless, J. Organomet. Chem, 1997,544,277. 14 A.M. McDonagh, M.G. Humphrey, M. Samoc, B. Luther-Davies, S. Houbrechts, T. Wada, H. Sasabe and A. Persoons, J. Am. Chem. Soc., 1999,121,1405. 15 S.K:Hurst, M.G. Humphrey, T. Isoshima, K. Wostyn, I. Asselberghs, K. Clays, A. Persoons, M. Samoc and B. Luther-Davies, Organometallics, 2002,21,2024. 16 S.K.Hurst, N.T. Lucas, M.G. Humphrey, T. Isoshima, K. Wostyn, I. Asselberghs, K. Clays, A. Persoons, M. Samoc and B. Luther-Davies, Inorg. Chim. Acta, 2002, in press. 17 D. Touchard, C. Morice, V. Cadiemo, P. Haquette, L. Toupet and P. H. Dixneuf, J. Chem. SOC.,Chem. Commun., 1994,859. 18 M.P. Cifuentes, C.E. Powell, M.G. Humphrey, G.A. Heath, M. Samoc and B. LutherDavies, J. Phys. Chem. A, 2001,105,9625. 19 A.M. McDonagh, M.G. Humphrey, M. Samw and B. Luther-Davies, Organometallics, 1999,10,5 195. 20 C.E. Powell and M.G.Humphrey, unpublished results. 21 Handbook of Nonlinear Optics, ed. R.L. Sutherland, Marcel Dekker, New York, 1996. 22 T. Isoshima and M. G. Humphrey, unpublished results. 23 S.K.Hurst, M.P. Cifuentes, J.P.L. Morrall, N.T. Lucas, I.R. Whittall, M.G. Humphrey, M. Samoc, B. Luther-Davies, I. Asselberghs, A. Persoons and A.C. Willis, Organometallics,2001,20,4664. 24. S.K. Hurst, M.P. Cifuentes and M.G. Humphrey, Orgunometallics,2002,21,2353. 1

CYCLOADDITION OF ALKYNES MEDIATED BY [RuCp(L)]+ (L = CO, NCH, PH3) AND RuCpCl COMPLEXES - METALLACYCLOPENTATRIENES AS KEY INTERMEDIATES - A DFT STUDY

Maria JosC Calhorda,' Karl Kirchner: and Luis F. Veiros3 1

ITQB, Av. da Republica, EAN, Apart. 127, 2781-901 Oeiras, Portugal, Departamento de Quimica e Bioquimica, Faculdade de Cihcias, Universidade de Lisboa, 1749-016 Lisboa, Portugal Institute of Applied Synthetic Chemistry, Vienna University of Technology, Getreidemarkt 9, A- 1060 Vienna, Austria Centro de Quimica Estrutural, Instituto Superior TCcnico, 1049-001 Lisboa, Portugal

1 INTRODUCTION Ruthenium halfsandwich complexes,of the type [RuCp(COD)Cl] and [RuCp*(COD)Cl] have been recently recognized as efficient pre-catalysts for [2+2+2] cycloadditions of 1,6diynes with a variety of unsaturated substrates such as terminal alkynes,' olefins: CSY (Y = S, NPh, NCy)? or tricarbonyl compounds O=CE2 (E = COOEt) (in a simplified manner shown in Scheme 1):

I

- COD

D

Scheme 1

These reactions are believed to proceed via the intermediacy of a metallacyclopentadiene species (I) as a result of an oxidative coupling with the 1,6-diyne

112

Perspectives in Organometallic Chemistry

molecule. Owing to the unsaturated nature of the metallacyclopentadiene, addition and subsequent insertion of unsaturated organic compounds into the ruthenium-carbon single bond eventually leads to the formation of cycloaddition products. This is in line with mechanistic proposals made for analogous reactions mediated by other transition metal complexes, e.g., [CoCp(CO)2].5 Interestingly, besides the [2+2+2] cycloaddition products (see paths a-d in Scheme l), cyclic olefins form also unusual double cyclopropanation products (path e in Scheme 1). It is therefore suggested that carbene intermediates may play a role in this process as proposed by Itoh et a2.2 In this context, it is important to mention that both [RuCp(COD)Cl] and [RuCp*(COD)Cl] do not promote the catalytic cycloaddition of terminal alkynes in the absence of 1,6-diynes. In fact, [RuCp(COD)Br] and the RuCp* analog react with HC=CPh to give in stoichiometric fashion a metallacyclopentatriene complex featuring a biscarbene functionality which is no longer catalytically active (Scheme 2).6 The 18e metallacyclopentatriene reacts readily with nucleophiles to give a metallacyclopentadiene comple~.~%~

Scheme 2 We have previously shown* that terminal and internal alkynes as well as diynes react with cationic complexes of the types [RuC~(ER~)(CH~CN)~]+ (E = P, Sb) to give either ally1 (11) or butadienyl carbene (111) complexes (Scheme 3). In these reactions, a cationic metallacyclopentatriene species is a common key intermediate which has been spectroscopically characterized and, in some cases, even isolated.

I

--I+

H3CwCH: I11

H,CCN'**'pePh3 H3CCN E=P,Sb

\

@ EPh3 migration c

E=P \

I1

CH3

Scheme 3 Herein we investigate by means of DFT calculations the oxidative coupling of acetylene mediated by RuCpCl and [RuCp(L)]+ (L = CO, NCH, PH3) fragments. Further support of the important role played by metallacyclopentatriene complexes will be

Cycloaddition of alkynes mediated by [RuCp(L)]+and RuCpCl complexes

113

presented, providing perspectives for unusual and unexpected reactions as yet to discover. In this context, it may be mentioned that in the [RuCp*(COD)Cl] mediated reaction of alkynes with carboxylic acidsg as well as in the [RuCp(CH3CN)3If mediated cycloisomerization of alkynes and propargyl alcohols" metallacyclopentatriene intermediates have been invoked. 2 METHODS AND RESULTS 2.1

The Precursors

All the available experimental evidence and theoretical calculations indicate that the first step in all these reactions consists of coordination of two molecules of acetylene to the [RuCp(L)]' fiagment (complex A; z = 0, 1) to form [RuCp(L)(HC=CH)2IZ(complex B), followed by coupling of the two acetylenes yielding a metallacyclopentatriene (complex C), as shown in Figure 1.

NC

11.5 10.0

AE2 -38.7 -25.3 -31.0

H PH3

10.3

-32.4

AE1

71 2 H Ru.

C1

CO

AE3 13.2 9.9 10.0 7.7

A

Figure 1 Energy profile (kcal mot') for the formation of the acetylene complex and the oxidative coupling of acetylene mediated by the [RuCp(L)]" (L = Cl, z = 0; L = CO, NCH, PH3, z = +I)@agment. In a previous work,*d dealing with the theoretical study of the mechanism of the reactions when L = PH3, we showed that this step is energetically unfavorable (for details on calculations, see Section 3). The energy of [RuCp(PH3)(HC=CH)2]' plus two free nitriles is 10.3 kcal mol-' higher than that of [RuCp(PH3)(CH3CN)$ and two acetylenes, while the activation energy for substitution of the first nitrile is 16.7 kcal mol-'. These energies show that bis(acety1ene)ruthenium complexes are not stable species, and indeed no examples could be found in the CSD. In order to check the calculated geometry, it was compared with that of molybdenum derivatives [MoCp(L)(q2-RC=CR)2]'(R = Me, L = I, CO, NCMe; R = Ph, L = CO)" and it was noticed that the M-C distances for Mo (ranging between 2.038 and 2.268 A) are shorter than for Ru (2.320, 2.322 A). This difference is larger than what would be expected on the grounds of different metals. Also, the C-C bond is much longer in the Mo complexes (1.259-1.268 A when L = CO; 1.267-1.280 A in other species) than the calculated value for ruthenium (1.246 A), suggesting less efficient

Perspectives in Organometallic Chemistry

114

backdonation in the Ru complex (weaker Ru-C bonds, stronger C-C bonds), although formally d6 Ru should be more electron rich than d4 Mo, and both complexes are monocations. In this work similar calculations were performed for related complexes with L = CO, NCH, C1. The formation of the bis(acety1ene) complexes from the nitrile ones is also energetically unfavorable for the carbonyl and nitrile species (AE1 in Figure 1). The synthetic route is different for the chloride. The geometries of all the acetylene derivatives are shown in Figure 2.

2.271

2.236

Figure 2 Some relevant distances (A) in [RuCp(L)(HCZH)2]' (L [R UCPCI(HC-
=

CO, NCH, PH3) and

The C-C distances are short and very similar to those found for the phosphine derivative (1.239 A for CO; 1.243 A for NCH; 1.247 for A Cl). An analogous behavior is observed for the Ru-C bonds which are also long (between 2.271 and 2.278 A for CO; between 2.236 and 2.250 A for NCH; between 2.200 and 2.203 for A Cl). The coligand does not have a strong effect on the geometry of the bisacetylene complexes, although it changes from a strong o-donor, to a weak n-donor and a strong n-acceptor. The Ru-C(Cp) distances are, in the whole family, typical of a slightly distorted q5-cyclopentadienlyring with three shorter and two longer bonds. 2.2

The formation of the metallacyle

The two molecules of acetylene coordinated to ruthenium couple in a concerted way to form the metallacycle. This step has been discussed before in ref. 8d for the phosphine system, and the calculations have now also been performed for other coligands L, namely CO, NCH, and C1. The energetics of the process is shown in Figure 1. Notice that when L = C1, the complexes are neutral, instead of cationic as in all the other cases. The calculated activation energy is slightly lower for L = PR3 (7.7 kcal mol-'), being comparable in the other three systems (Cl, CO, NCH: 13.2, 9.9, 10.0 kcal mol-', respectively). The more exothermic reaction is found for C1(-38.7 kcal mol-') and becomes less exothermic as one moves to NCH, CO, and PH3 (-31.0, -25.3, -32.4 kcal mol-l). Although the formation of the bis(acety1ene)complex from the bis(acetonitri1e) is unfavorable, the global reaction from this complex to form the metallacycle, given by AE2-AE1 (-13.8, -21.0. -22.1, for CO, NCH, and PH3, respectively) is exothermic. This reaction appears very similar when the coligand is CO or NCH, although the metallacycle containing CO has the highest energy in the series. In Figure 3, the geometries of the transition states for the conversion of bis(acety1ene) complexes into metallacycles are compared. The transition states for L = CO, NCH, C1 are also analogous to that calculated for the phosphine complex. The Ru-C bonds start to become different, two longer and two shorter, while the C-C distance between adjacent coordinated acetylenes has decreased from ca.

Cycloaddition of alkynes mediated by [RuCp(L)]+and RuCpCl complexes

115

2.787 A in complexes B to ca. 2.105 A in the TS, for L = CO, NCH, and C1. In order to form the metallacycle, the Ru-C, bond still shrinks to 1.938, 1.967 A (CO), 1.946 A (NCH), and 1.945 A (Cl), while a strong C-C bond forms between the two Cp of adjacent acetylenes (1.381 A, CO; 1.390 A, NCH; 1.402 A, Cl).

2.107

2.106

2.133

2.049

Figure 3 The transition states for the conversion of [RuCp(L)(HC=CH)*]+(L = CO, NCH, PH3) and [RuCpCl(HC-
The bonding of the C4H4 group to the [RuCp(PH3)]+fragment was studied in detail8d and it was found that there were four bonding electron pairs, resulting from donation from the symmetric and antisymmetric combination of C lone pairs, and n donation to Ru, as well as Ru backdonation to an empty n* orbital of C4H4. Therefore, it seems more appropriate to describe the metallacycle as a metallacyclopentatriene, formulated as sketched in Figure 1 (C). The Ru-C bonds have carbenic character, which fits the known reactivity of these species. While this metallacycle is a common intermediate in the reactions mediated by these RuCpL fragments, the steps following its formation differ, namely in the case of L = phosphine, when the phosphine migrates to the a carbon. In the other systems, another acetylene molecule approaches (or, in a more general way, another unsaturated molecule) and couples to the metallacycle affording benzene, other rings, other species, depending on the type of molecule. The first question that might be asked concerning the common intermediate is whether the description proposed for L = phosphine can be extended to other systems as well. In order to answer it, we analyse in detail the four metallacycles under study (Figure 4).

1.437 1.430 1.381

1.390

Figure 4 The geometry of the rnetallacyclopentatrienes [RuCp(L)(HCCHCHCw]”o (L CO, NCH, PH3, Cr) (relevant distances in A).

=

There is a remarkable similarity between the C-C and Ru-C bond distances within the metallacycles. The Ru-C distances range from 1.938 to 1.967 A. The C-C bond distances exhibit a long-short-long pattern with the Cp-Cp bonds being the shortest (1.371 to 1.382 A). The C,-Cp bonds do not differ much (1.399 to 1.446 A), but are longer for CO, NCH, and PH3. On the other hand, for the chlorine, the trend seems to be the opposite, but in this

116

.

Perspectives in Organometallic Chemistry

case the differences are even smaller (1.402 and 1.399 A), the distances becoming equivalent in practical terns. It is interesting to notice that the main structural features of the metallacycle, namely its metallacyclopentatriene, rather than metallacyclopentadiene, character persists despite the coligand. Although our series of ligands is not very extensive it covers a wide range of bonding possibilities, from the n-acceptor CO, to the n-donor C1, the weak n-donor NCH, and the strong o-donor PH3. The Ru-C bonds to the cyclopentadienyl ring do not fall in two groups (three longer plus two shorter) as observed in the acetylene complexes, but must be arranged in three groups (two long, two medium, one short). In the CO complex these distances are 2.348, 2.326; 2.308, 2.302; 2.278 A, respectively. They are comparable to those observed in the phosphine derivative. On the other hand, these bonds become longer for the chlorine and nitrile species. The lengthening is specially significant for C1, with distances 2.456, 2.458; 2.309, 2.310; 2.230 A. The weakening of the Ru-C bonds can be traced to the n-donor character of the ligands, and is more relevant when the stronger Cl is coordinated. It is also interesting to note that the angle formed by the Ru-L vector and the plane defined by the five-membered metallacycle Ru-Cl-C2-C3-C4 varies between 95", 92", 88", and 103" for L = CO, NCH, PH3, and C1, respectively. Another indication of the strengh of Ru-C and other bonds is given by the Wiberg indices. Some of them are shown in Figure 5 for the four metallacycles along with the charges (italics) resulting from a natural population analysis (NPA).

-0.103

Figure 5 Wiberg indices for the Ru-C, Ru-L, C-Cbonds, and charges in the atoms Ru, C,L the metallacycles [RuC~(L)(HCCHCHCH)/'~'(L = CO, NCH, PH3, CI). The Wiberg indices give information which parallels the bond lengths, for each type of bond. It becomes more striking that the Cp-Cp bond is stronger than the C,-Cp one, with the exception of the chlorine where the strength is comparable. The Ru-C carbenic bonds are weakest in the CO derivative, although the differences are small, as discussed above. In what concerns charges, the metal is always positively charged, the highest charge being found for the chlorine complex. This fact, accompanied by the absence of strong steric effects, may ex lain the tendency of the chlorine species to be attacked by nucleophiles at the ruthenium.6 5 The HOMO and the LUMO of the four metallacycles are shown in Figure 6. The HOMO is essentially a metal based orbital, with a bonding contribution from the metallacyclic carbons, almost negligible for the phosphine and the chlorine, but significant for CO and nitrile. When L = Cl, there is also a strong antibonding contribution from the chlorine. The cyclopentadienyl ring contributes very moderately to the HOMO, except in the nitrile complex, giving rise to a bonding interaction. The LUMOs, on the other hand, are mainly the antibonding combination of the second n* orbital of the diene ring with a ruthenium orbital. They differ because in the carbonyl complex there is also a bonding contribution from the CO n* orbital, while in the chloride a Ru-Cl antibonding interaction

Cycloaddition of alkynes mediated by [RuCp(L)]+and RuCpCl complexes

117

occurs. In the other two complexes, the PH3 and the nitrile ligand do not interact with the metal. Despite these differences, it is easy to see how similar the fiontier orbitals are for these metallacycles.

Figure 6 HOMO (top) and LUMO (bottom) of the metallacyclopentatrienes [RUC~(L)(HCCHCHCH)/'.~ for L = CO, NCH, PH3, and Cl @om left to right).

2.3

The reactivity of the metallacyclopentatriene complexes

The metallacyclopentatriene reacts in different ways depending on the coligand (L = CO, NCH, PH3, Cl). As was studied already in detail previously,sd in the presence of a phosphine the usual pathway involves migration of the phosphine to the a carbon of the metallacyclopentatriene. The metallacycle is inert towards another acetylene molecule. The first step in the migration, which is the rate determining one according to our calculations, consists of a distortion of the planar metallacyclopentatriene (see previous figures) towards a non-planar arrangement better described as a metallacyclopentadiene, with concomitant approach of the phosphine to the carbon atom. The complexes with L = NCH and Cl are reactive toward unsaturated molecules such as acetylene, olefins, and CS2, and the reaction proceeds to afford 4 +2 cycloaddition products containing six-membered rings. These reactions will be described in detail in a forthcoming publication. With L = CO, cyclopentadienone complexes are formed as a result of CO insertion in to the metal carbon double bond of the metallacycle. l2 3 COMPUTATIONAL DETAILS All calculations were performed using the Gaussian98 s o h a r e package13 on the Silicon Graphics Cray Origin 2000 of the Vienna University of Technology, at IST and ITQB. The geometry and ener y of the model complexes and the transition states were optimized at the B3LYP level' f with the Stuttgart/Dresden ECP (sdd) basis set'5 to.describe the electrons of the ruthenium atom. For all other atoms the 6-3 1g** basis set wai employed.'6 Frequency calculations were performed to confirm the nature of the stationary points, yielding one imaginary frequency for the transition states and none for the minima. Each

Perspectives in Organometallic Chemistry

118

transition state was further confirmed by following its vibrational mode downhill on both sides, and obtaining the minima presented on the reactions energy profile. All geometries were optimized without constraints (Cl symmetry) and the energies were zero point corrected. A Natural Population Analysis (NPA)" and the resulting Wiberg indices" were used for a detailed study of the electronic structure and bonding of the optimized species. 3D representations of orbitals were drawn with MOLEKEL.19 4 CONCLUSION The formation of a metallacycle from RuCp(L) fragments in the presence of certain acetylenes has been well documented. In this work, we analyzed in detail the energetics of the reaction for L = CO, NCH, PH3, C1, as well as the structural features of the intermediates and the transition state for coupling of the coordinated acetylenes, extending a previous detailed study of the reactivity for L = PH3.8d The metallacycles can be classified in all cases as metallacyclopentatrienes, with double bond character in the Ru-C bonds. The geometric and electronic characteristics of the metallacycles appear to be remarkably similar, although in the neutral C1 derivative, for instance, steric constraints are less important, as the wide angle formed by the Ru-L vector and the plane defined by the five-membered metallacycle decreases from 103" to only 95O, 92O, and 88", respectively, for L = CO, NCH, PH3. The C1 complex also differs from the others in the composition of the frontier orbitals, which exhibit a strong Ru-Cl antibonding character. In the carbonyl, the L ligand does not contribute to the other HOMOS and appears as the Ru-CO bonding component of backdonation. Also, the HOMO of the nitrile metallacycle is the least localized in the metal. Interestingly, these two .n-donor ligands are associated with the reactivity toward acetylene or other unsaturated molecules to give 4 + 2 products. However, these details do not shed all the light on the reactivity of the metallacycles when in presence of unsaturated molecules and a detailed study of the reaction pathways, to be published soon, is required. References

Y. Yamamoto, R. Ogawa and K. Itoh, Chem. Commun. 2000,549. Y. Yamamoto, H. Kitahara, R. Hattori and K. Itoh, Organometallics 1998,17, 1910. Y. Yamamoto, H. Kitahara, R. Ogawa, H. Kawaguchi, K. Tatsumi and K. Itoh, J Am. Chem. SOC.2000,122,4310. Y. Yamamoto, H. Takagishi and K. Itoh, J Am. Chem. SOC.2002,124,28. Y. Yamamoto, H. Takagishi and K. Itoh, J Am. Chem. SOC.2002,124,6844. (a) D. B. Grotjahn, in Comprehensive Organometallic Chemistry 11, ed. L. S. Hegedus, E. W. Abel., F. G. A. Stone and G. Wilkinson, Pergamon, Oxford, 1995, vol 12, p. 741. (b) N. E. Shore, in Comprehensive Organic Synthesis, ed. B. M. Trost and I. Fleming, Pergamon, Oxford, 1991, vol. 5, p. 1037. (a) M. 0. Albers, P. J. A. deWaal, D. C. Liles, D. J. Robinson, E. Singleton and M. B. Wiege, J Chem. SOC.,Chem. Commun. 1986, 1680. (b) C. Gemel, A. LaPensee, K. Mauthner, K. Mereiter, R. Schmid and K. Kirchner, Monatsh. Chem. 1997,128, 1 189. (c) C. Ernst, 0. Walter, E. Dinjus, S. Arzberger and H. Gorls, J Prakt. Chem. 1999, 341, 801. C. S. Yi, J. R. Torres-Lubian, N. Liu, A. L. Rheingold, I. A. Guzei, Organometallics 1998,17, 1257, and references therein.

Cycloaddition of alkynes mediated by [RuCp(L)l+ and RuCpCl complexes

8

9 10 11 12 13

14

15

16

17

18 19

119

(a) K. Mauthner, K. M. Soldouzi, K. Mereiter, R. Schmid and K. Kirchner, Organometallics 1999, 18, 4681. (b) E. Ruba, K. Mereiter, R. Schmid and K. Kirchner, Chem. Commun. 2001, 1996. (c) E. Becker, E. Ruba, K. Mereiter, R. Schmid and K. Kirchner, Organometallics 2001, 20, 3851. (d) E. Ruba, K. Mereiter, R. Schmid, V. N. Sapunov, K. Kirchner, H. Schottenberger, H.; M. J. Calhorda and L. F. Veiros, Chem. Eur. J., 2002,8,3948. J. La Paih, S. Derien and P. H. Dixneuf, Chem. Commun. 1999,1437. B. M. Trost and M. T. Rudd, J. Am. Chem. SOC. 2002,124,4178. F. H. Allen, 0. Kennard, Chem. Design Autom. News, 1993,8,31. E. Ruba, K. Mereiter, K. M. Soldouzi, C. Gemel, R. Schmid, K. Kirchner, E. Bustelo, M. C. Puerta and P. Valerga, P. Organometalfics2000,19, 5384. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, 0. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Peterson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle and J. A. Pople, Gaussian 98, revision A.7 Gaussian, Inc.: Pittsburgh, PA, 1998. (a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648. (b) B. Miehlich, A. Savin, H. Stoll and H. Preuss, Chem. Phys. Lett 1989, 157, 200. (c) C. Lee, W. Yang and G. Parr, Phys. Rev. B 1988,37,785. (a) U. Haeusermann, M. Dolg, H. Stoll and H. Preuss, Mol. Phys. 1993, 78, 121 1. (b) W. Kuechle, M. Dolg, H. Stoll and H. Preuss, J. Chem. Phys. 1994, 100, 7535. (c) T. Leifinger, A. Nicklass, H. Stoll, M. Dolg and P. Schwerdtfeger, J. Chem. Phys. 1996, 105,1052. (a) A. D. McClean and, G. S. Chandler, J. Chem. Phys. 1980, 72, 5639. (b) R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys. 1980, 72,650. (c) A. . H. Wachters, Chem. P h p . 1970,52, 1033. (d) P. J. Hay, J. Chem. Phys. 1977,66, 4377. (e) K. Raghavachari and G. W. Trucks, J. Chem. Phys. 1989,91, 1062. (f) R. C. Binning and L. A. Curtiss, J. Cornput. Chem. 1995,103,6104. (g) M. P. McGrath and L. Radom, J. Chem. Phys. 1991,945 1 1. (a) J. E. Carpenter and F. Weinhold, J. Mol. Struct. (Theochem) 1988,169, 41. (b) J. E. Carpenter, PhD thesis, University of Wisconsin (Madison WI), 1987. (c) J. P. Foster and F. Weinhold, J. Am. Chem. SOC.1980, 102, 721 1. (d) A. E. Reed and F. Weinhold, J. Chem. Phys. 1983, 78, 4066. (e) A. E. Reed and F. Weinhold, J. Chem. Phys. 1983, 78, 1736. (f) A. E. Reed, R. B. Weinstock and F. Weinhold, J. Chem. Phys. 1985, 83, 735. (g) A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev. 1988, 88, 899. (h) F. Weinhold, J. E. Carpenter, The Structure of Small Molecules and Ions, Plenum, 1988,227. K. B. Wiberg, Tetrahedron 1968,24, 1083. S. Portmannr H. P. Luthi, Chimia 2000,54,766.

SELECTIVE C-C COUPLING REACTIONS OF Me2N-GC-NMe2 CENTERS

AT IRON(0)

Alexander C. Filippou*, Torsten Rosenauer and Gregor Schnakenburg Institut fur Chemie, Humboldt-Universit zu Berlin, Brook-Taylor Str. 2, D- 12489 Berlin

1 INTRODUCTION The reactions of alkynes with transition metal compounds have been extensively investigated since the pioneering work of W. Reppe et al.' Numerous transition metalmediated C-C coupling reactions were discovered in the course of these studies and used to produce a variety of organic compounds. Illustrative examples are the cyclooligomerization and [2+2+2] cycloaddition reactions: the Pauson-Khand reaction3and the Dotz benzoannulation? In comparison, the transition-metal chemistry of alkynes, which bear (CO)&r =C

Ph

,S=C

+W(CO)s(S=C(Ph)H)

I + ('BuO)sWMe,

I+quin

t (OC)$e

v I

NEt2

NEt2 EtzNh N E t 2

n

EtzNVNEt2 NEt2

Scheme 1. Selected reactions of 2 with transition metal complexes (quin dppm = Ph2PCH.PPh2).

=

quinuclidine;

on both carbon atoms strong n-donor substituents as the bis(N,N-dialky1amino)acetylenes R~N-CEC-NR~(1: R = Me; 2: R = Et), remains little explored? In fact, only a few reactions of 1, 2 with metal carbonyls: some insertion reactions of 1, 2 into the metalcarbon double bond of carbene complexes7 and the C=X bond (X = s, Se, Te) of metal-

121

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0)centres

coordinated heteroalkenes,' a few [2+2] cycloadditions to the C=C bond of vinylidene complexes: a metathesis reaction of 2 with the carbyne complex (*BuO)3W=CMeIoand a palladium-catalyzed cyclotrimerization of 2 to hexakis(diethy1amino)benzene have been reported so far (Scheme l)." The lack of attention that the transition metal chemistry of bis(N,N-dialky1amino)acetylenes has received in the past is surprising, given the strong influence exerted by amino groups on the electronic properties and reactivity of 'IIsysterns.I2 This can be in part attributed to the difficulty in preparation and handling of these "electron-rich", very air-sensitive alkynes.6c."-'3 Our interest in the chemistry of diaminoacetylenes (ynediamines) arose from a series of electrophile-promoted metal-centered C-C coupling reactions of C1-ligands, which we found in the course of mechanistic studies of the reductive isocyanide-isocyanide coupling '' reaction of [MX(CNR)6]+ complexes (M = Mo, W; X = halide; R = a l k ~ l ) . ' ~ ~These include the proton-induced isocyanide-isocyanide'4e~14fy15b~*5c and carbyne-isoc anide coupling reaction (Eq. 1)16 or the oxidative coupling of two carbyne ligands (Eq. 2).' In all these reactions ynamine or ynediamines are formed, which are tightly bound to the metal center and act as four-electron donor ligands.

Y

4X

%+ cN

?J

%

I'

C

N / ",

x-kgj

+ HX

Cl'

L

\

%N R = Ph, NEt2; X = Br, I; L

c o , 'BUNC

Et2N

The high nucleophilicity of unligated ynediamines, which is best evidenced in the fast protonation and methylation reactions of 1 (Scheme 2)," let us suggest that coordination of 1 or 2 to electrophilic transition metal centers would rapidly occur under mild conditions to generate novel metallacycles displaying new reaction modes. In addition, stabilization of Me2N-C--L; *Me2

+ (NEt3H)Br

1

Scheme 2. Selected reactions of 1 with electrophiles.

122

Perspectives in Organometallic Chemistry

reactive intermediates was conceivable taking advantage of the strong +M effect of the dialkylamino groups and a wealth of chemistry was anticipated upon electrophilic functionalization of the "electron-rich" metallacycles. Experimental support for these hypotheses is presented in the following work describing the unique reaction mode of 1 with [Fe(C0)5] and the unprecedented reactivity patterns of the resulting ferracycles. 2 RESULTS AND DISCUSSION 2.1 Reaction of [Fe(C0)5] with Bis(N,N-dimethy1amino)acetylene The reactions of [Fe(C0)5] with alkynes are in general complex and not selective affording a wealth of organometallic complexes and carbocycle~.'~The tricarbonyl( q4cyclopentadienone) iron complexes are widely known among the products6a,6c,20 due to their application in organic synthesis?' All reactions of [Fe(C0)5] with alkynes reported to date seem to follow a dissociative pathway and need thermal or photochemical activation of [Fe(C0)5]19722 because of the large Fe-CO bond dissociation energy?3 The alkyne complexes [Fe(C0)4( #-alkyne)] have been postulated as intermediates in these reactions and were recently isolated as thermally sensitive s0lids.2~By contrast, the reaction of 1 with [Fe(C0)5] in THF follows an associative pathway and affords at -50 "C the 18electron ferracyclobutenone 3, which was identified in solution by IR and NMR spectroscopy (Scheme 3)?5 The conversion of [Fe(C0)5] into 3 resembles the reaction of the isosterous carbonylrnetallate [Re(CO)5]- with the activated alkynes RC=CC02Me (R = C02Me, Me, H) to give the rhenacyclobutenones [Re(C0)4{ q':qlC(R)C(C02Me)C(O)

3

4

Scheme 3. Associative reaction pathway of [Fe(CO)j] with 1. Complex 3 is thermolabile and loses CO above -30 "C to give the ferrac clobutenone 4 (Scheme 3), which was isolated as an orange solid in overall 79% yield?'Complex 4 is indefinitely stable in solution at ambient temperature, but eliminates CO either in refluxing toluene or upon melting at 124 "C to yield selectively the red alkyne complex 9 (Scheme 4). Complexes analogous to 4 have been so far postulated as intermediates in metalcentered alkyne-CO coupling reactions, but no experimental evidence could be found for their e~istence.2~ From that point of view the thermal stability of 4 is astonishing. The structural parameters and spectroscopic data of 4, as well as the theoretical calculations of the model compound [Fe(C0)3{ q' :q1-C(NH2)C(NH2)C(0)>] (4a) reveal extensive n-electron delocalization over the atoms Fe, C,, N, and CB, and show that no bonding interaction exists between the iron center and the Cp atom despite their close distance (Figures 1 and 2).

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0)centres

123

The molecular structure of 4 shows a distorted square-pyramidal complex, in which the Cacarbon atoms (C4, C8) of the ferracycle and two carbonyl ligands (Cl-01, C3-03) occupy the basal coordination sites (Figure 1). The four-membered ferracycle is puckered (folding angle Fe,C4,C8/C4,C5,C8 = 128.4"), with the Cp ring-carbon atom being located at a close distance from the iron center (Fe***C5 2.21 5(2) A) in an opposite direction to the axial CO ligand (Figure 1). This distance is, however, considerably longer than those

Figure 1. Lefr side: Diamondplot of the molecular structure of 4 with thermal ellipsoids set at 50% probability. Hydrogen atoms are omittedfor clarity. Selected bond lengths (A) and bond angles Fe-C4 1.933(3), Fe-*-C5 2.215(2), Fe-C8 1.896(3), C4-C.5 1.469(3), C5<8 1.400(4), C4-04 1.210(3), Fe-C4-C5 80.0(2), C4-C.548 103.2(2), C5-C8-Fe 83.0(2), C4-Fe-C8 71.9(1); Right side: Contour line diagram of the Laplacian distribution V 'Hr) and atomic interaction lines of 4a at the B3LYPh-311 G* level (projection in the plane defined by the iron center and the bond critical points (m) oj the C5-C8 and the C 4 - 0 bonds). There is only one ring critical point (A) in the ferracycle and no bondpath between the atoms Fe and C5.

r):

observed in ($-vinyl-aminocarbene)Fe(CO)3 complexes, such as [Fe(C0)3{ $C(NMe2)C(OEt)C(H)C(0)tBu}] (Fe-Cp 2.063(3) and [Fe(C0)2(PMe3){ $C(NiPr2)C(NMe2)C(H)NMe2}] (Fe-Cp 2.086(2) A)?8b The latter distances compare well with those observed for Fe-C(sp3) single bonds (2.07 - 2.09)?9 In addition, only one ring critical point was found in the four-membered ferracycle and no bond path exists between the atoms Fe and Cp (Figure l).30 The Fe-CDMe2) bond of 4 (Fe-C8 1.896(3) A) compares well with that of the related PMe3-substituted ferracyclobutenone 8 (Fe-C 1 1.888(2) A; Scheme 4, Figure 4) and with those of metallacyclic iron-carbene complexes?' This bond is, however, considerably shorter than the Fe-CaMez) bonds of the 18electron ferracyclobutenones 5a ({Fe-Ca(NMe2)},V 2.029(4) A) and 7 (Fe-C3 2.03 l(3) A; Scheme 4, Figures 3 and 4), the latter corresponding to Fe-C(sp2) single bonds (1.99 2.03 A). In addition, the atoms Fe, C8, C5, N2, C9 and ClO are almost arranged in one plane (interplane angle Fe,C8,C5/N2,C9,C10 = 6.1°), the CCbonded amino group is planar

124

Perspectives in Organometallic Chemistry

(sum of angles at N2 = 359.6'), the C8-N2 bond is short (1.306(4) A), and the C5-C8 bond length of 1.400(4) 8,has an intermediate value between that of a C(sp2)-C(sp2) single All these structural parameters indicate a n(1.48 .$) and a C-C double bond (1.32 electron delocalisation over the atoms Fe, C8, N2 and C5 as observed in aminocarbene complexes. Additional evidence for this is given by the isosurface plot of the delocalized

Figure 2. ACIDplot of the delocalized electron density of 4a (isosurface value 0.065 a.u.) and NBO charges.

electron density (Figure 2 p 3 which moreover shows no delocalization between the atoms Fe and C4. Furthermore, the 'H NMR spectrum of 4 reveals that rotation of the Ccbonded amino group is frozen at ambient temperature (AG' > 14.4 kcal mol-I), the I3C{'H} NMR spectrum (THF, 25 'C) displays a low-field shified signal for the C a M e 2 ) atom at S 216.4 ppm, and the IR spectrum of 4 shows a t(CgC,-N) absorption band at high wavenumbers (1632 cm-' in THF; Table l).25In comparison, the Cpbonded amino group is not planar (sum of angles at N1 = 343.l'), the Cp-NMe2 bond (C5-Nl 1.384(3) A) is only slightly shorter than a C(sp2)-N(sp3) single bond (1.42 A),32and the amino group is rotated out of the plane of the trigonal-planar coordinated Cp atom (dihedral angle C6,Nl,C7/C4,C5,C8,Nl = 47.1') (Figure 1). All these data suggest a small degree of n;bonding in the C r N bond. Consequently, the activation barrier for rotation of the amino group about the C r N bond (9.95 kcal mol-') is smaller than the barrier for rotation about the C,-N bond (vide supra). Notably, the Cg NMR signals of 4 (S59.3) and of the PMe3substituted analogue 8 (659.0) appear at considerably higher field than those of the 18electron ferracyclobutenones 5a/5b (6 124.6, 126.6) and 7 (S124.4) (Table 2), and the Fe-C,] bonds of 4 (Fe-C4 1.933(3) A) and of 8 (Fe-C3 1.928(2) A; Figure 4) are considerably shorter than those of the 18-electron ferracyclobutenones 5a ({ Fe-C,I}, 2.095(8) A; Figure 3) and 7 (Fe-Cl 2.077(4) I$; Figure 4) or those of the ferracyclopentenedione 6 (Fe-C1/C4 1.996(2)/1.978(2) A; Figure 3). All these data emphasize the difference in bonding between the 16-electron ferracyclobutenones 4 and 8 and their 18-electron counterparts 5a and 7.

125

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0) centres

2.2 Reactions of the Ferracyclobutenone 4 with PMe3 The ferracyclobutenone 4 is quite reactive towards nucleophiles due to its coordinative and electronic unsaturation. Two types of reactions with nucleophiles were observed so far. The first type of reactions is described in the following and involves nucleophiles, which can not insert into Fe-C bonds, such as PMe3 (Scheme 4).

4

W5b

i

- CO vac., 20'C

vNMg

Me2N

10 m3p

+ PMe3

w%..

pentane, 20°C

9

8

&'

I

NMe2

F ..**\ e k M e Z

I 7

Scheme 4. Reactions of the ferracyclobutenone 4 and otherferracycles with PMeJ. Complex 4 reacts rapidly with one equivalent of PMe3 in THF or Et20 at -78 "C to afford an isomeric mixture of the 18-electron ferracyclobutenones 5a and 5b, which was isolated as a yellow solid in 92% yield. The mixture consists according to NMR spectroscopy of the facial isomer (5a) and one meridional isomer (5b) in the molar ratio of 1 : 0.94. The rneridional isomer is suggested on the basis of steric arguments to contain the PMe3 ligand in trans-position to the acyl-carbon atom. Evidence for the presence of two isomers is also provided by the IR spectrum of 5d5b in pentane, which displays two t(C=O) absorptions bands for the facial isomer at 2043 and 1970 cm-' and three < G O ) absorptions bands for the meridional isomer at 2035, 1987 and 1956 cm-' (Table 1). In addition, the isomers 5a and 5b are distinguished by two IR absorption bands of medium intensity at 1688 and 1546 cm-I, which are assigned to the t(C=O) and t(Cp-Ca-N) vibrations of the ferracycle, respectively. These vibrations have considerably lower frequencies than those of complex 4 (t(C=O) 1733 cm-'; t(CgCa-N) 1632 cm-I), which underlines the considerable change in bonding of the ferracycle upon PMes-coordination. This is further confirmed by the 13C{'H}NMR spectrum of 5d5b (THF, -84 "C), which reveals a large high-field shift of the Cwl and C m M e 2 ) resonances (S 176.9, 178.7, 182.1, 185.6), but a large low-field shift of the CflMe2) resonances (6124.6, 126.6) with respect to the ring-carbon resonances of 4 (Cacyl,8220.8; CmMez), 8216.4; CB(NMe2), S 59.3; Table 2). The molecular structure of the facial isomer 5a was determined by X-ray crystallography and reveals a distorted octahedral arrangement of the ligands around the iron atom (Figure 3).34 The distortion arises from the small bite of the unsaturated C3

Perspectives in Organometallic Chemistry

126

ligand (C3-Fel-Cl 64.49(9)") and the inclination of the axial ligands PMe3 and CO (Cll-01) to the ferracycle (PI-Fe-Cll 168.19(7)'; Cll-Fe-Cl 83.8(1)'; PI-Fel-Cl 85.09(7)'). The four-membered ferracycle is planar, the CB atom is located at a considerably larger distance from the iron center (2.616 A) than in 4, and the mean Fe-Ca(NMe2) and Fe-Cacyl bond lengths of 2.029(4) and 2.095(8) A, respectively, compare well with Fe-C(sp2) single bond lengths (1.99 - 2.03 A)?9 The Cpbonded amino group is not planar (sum of the angles at Np = 335.9'), and is arranged perpendicular to the ring-plane (interplane angle N1 ,C7,C8/Fe,C1,C2,C3 = 89.5'). This orientation reduces the steric repulsion between the vicinal dimethylamino groups. Consequently, the mean CFN bond length of 1.442(4) A compares well with a C(sp2)-N(sp3) single bond length (1.42 A) and rotation of the amino group about the C r N bond has a low activation barrier. In comparison, the CKbonded amino group is planar (sum of angles at N, = 359.5") and slightly rotated out of the ring plane (interplane angle N2,C9,ClO/Fe,Cl,C2,C3 = 14.9'). In addition, the C,-N bond is short (mean value 1.339(2) A), but longer than that of 4 (1.306 A). All these data suggest a reduced double-bond character of the C c N bond in 5a in comparison with 4.

=11

6

C13

Figure 3. Diamond plot of the structure of one of the four independent molecules of 5a found in the unit cell (left side) and of 6 (right side). Thermal ellipsoids are set at 50% probability and hydrogen atoms are omitted for clarity. Selected bond lengths (A) and bond angles (") of one molecule of 5a: Fel-Cl 2.104(2), Fel-C3 2.024(2), Cl-C2 1.420(3), C2<3 1.393(3), C1-04 1.220(3), C2-Nl 1.438(3), C 3 4 2 1.337(3), Fel-Cl-C2 94.0(2), C l -C2<3 103.1(2), C243-Fel 98.4(2), C3-Fel -Cl 64.49(9); 6: Fe-Cl 1.996(2), Fe-C4 1.978(2), Cl-C2 1.489(2), C 2 X 3 1.353(3), C3<4 1.530(2), C1-01 1.226(2), C4-02 1.221(2), C2-Nl 1.421(2), C3-N2 1.379(2), Fe-Cl-C2 115.0(1), Cl-C2-C3 114.9(2), C2<3-C4 113.1(2), C3-C4-Fe 114.5(1), C4-Fe-Cl 81.11(8), PI -Fe-P2 174.45(2). Additional evidence for this is given by the temperature dependent 'H NMR spectra of 5a (300 MHz, [DslTHF), which show that the activation barrier AGf for rotation of the amino group about the C,-N bond is lower (1 1.O kcal mol-I, Tc = 234 IS,Av = 123 Hz) than in 4 (AG' > 14.4 kcal mol-I). This barrier is, however, larger in the facial isomer 5a than in the meridional isomer 5b (9.8 kcal mol-*, T, = 210 K , Av = 120 Hz). The reason for this is probably the increased steric repulsion between the cis-positioned PMe3 ligand and

Selective C-C coupling reactions of Me,N-C=C-NMe2 at iron(0) centres

127

the C,-bonded amino group, which evolves in 5a during the out-of-plane rotation of the amino group. Additional experimental support for a steric influence on this barrier provides a comparison of 5a with the related 18-electron ferracyclobutenones 3 and 7, which reveals an increase of AG' (3: 9.5 kcal mol-I, Tc = 204 K, Av = 123 Hz;7: A@ > 13.8 kcal mol-I, Tc > 293 K, Av = 144 Hz) as the axial CO ligands are successively replaced by the sterically more demanding PMe3 ligands. Some n-electron delocalization is evidenced in 5a by the mean C@iMeZ)-Cp (1.393(2) A) and C B C ~bond ~ ~ length (1.4 14(4) A). Table 1. IR absorptions of the complexes Fe(C0)j and 3 - 10 in the region 2200 - I500 cm-' and "P NMR data of 5 - 8 and 10.

1 Fe(CO), 3 4 5a

5b

[a'

[bl

[I' [I'

6

7 8 9 10

[I' la'

I

IR [cm-'] &PPml L(C=O) t(C=O) c(C-C-N) 2019(s), 1993(vs) 2086(w), 2034(s), 20 1~ ( v s ) 1723(m) 1563(m) 2035(vs), 1963(vs, sh), 1955(vs) 1733(m) 1632(m) [dl 2035(w), 1987(m), 1956(s) . 1688(m) 1546(m) 16.6,20.4 2043(m), 1970(vs) 1993(vs), 1935(vs) 1602(w), 1582:m), 1549(w) Iel 19.7 1967(vs), 1898(vs) 1637(m) 1530(m) 24.7 1978(s), 1971(vs), 1921(s), 191~ ( v s ) 1701(m) 1607(m) [rl 24.9 1698(m) 20 11(s), 1927(vs) 1936(s), 1871(s) 1674(m) [gl 36.0

spectrometer in solution at ambient temperature unless otherwise stated. a: THF; b: THF (-40 "C); c: pentane. Abbreviations used for the intensity and shape of the IR bands: w, weak, m, medium; s, strong; vs, very strong; sh, shoulder. "P('H) NMR spectra were recorded on a Bruker AM-300 spectrometer in solution at 25 "C unless otherwise stated. d: [D8]THF, -84 "C; e: [D8]toluene; f: [DslTHF; g: CDCI,.

Table 2. 13CNMR data of the complexes 3 - 10.

I +

#

204.7, 208.2, 220.8

I I 218.5 (22.0)

9 ~

10

I

216.4

59.3

44.1,46.7 40.5 42.1,48.644.8 (2.5), 16.0 (30.5) 45.2 (3.4)

176.9 (23.7), 178.7 (7.6) 124.6 (13.6), 126.6 182.1 (8.5),185.6 (13.6) 49.2 (br)

159.8 (1.1) - 41.7(0.6) 17.4(m) 262.8 (23.8) 211.7 (12.8) 201.3 (17.5) 182.8 (27.6; 124.4 (9.2) 41.0 (br) 45.2 (3.6) 17.7 (m) 48.3 (br) 225.9 (19.9) 219.8 (4.6) 59.0 (4.2) 42.4,46.5 41.1 20.5 (27.5) [ I 218.1 (8.7) 45.1,47.1 1["]1 220.5 193.8 45.5,47.5 - 22.8(26.5) 190.5 (4.0) 226.0 (12.0)

%{'H) NMR spectra were recc ,ded on a Bruker AM-300 spectrometer in solution at 25 "C unless otherwise stated. a: [D8]THF; b: [D8]THF, -50 "C; c: pE]THF,-84 "C; d: [DEJtoluene;e: [D8]toluene, -79°C; f CDC13, -59 "C; g: averaged signals of the three rapidly interconverting isomers. Abbreviations used for signal shape and multiplicity: br, broad; m, multiplet. The 13CNMR signals of 4, 7 and 8 were assigned by HMBC and C,H COSY experiments.

128

Perspectives in Organometallic Chemistry

Mononuclear metallacyclobutenones, such as 5d5b or 3 and 7, are very rare and only a few studies have been carried out on their These studies would improve the understanding of the role of metallacyclobutenones in metal-centered alkyne-carbonyl coupling reactions. In this context, we investigated the reaction of 5a/5b with PMe3. Treatment of the complexes 5a/5b with 1.1 equivalents of PMe3 in THF at -78 "C followed by slow warming of the reaction solution to ambient temperature affords selectively the ferracyclopent-3-ene-2,5-dione6, which was isolated as a red solid in 75% yield (Scheme 4). IR monitoring of the reaction reveals that conversion of 5a/5b into 6 starts at ca. -15 "C and proceeds rapidly at ambient temperature. Complex 6 is also selectively formed upon treatment of 4 with an excess of PMe3 (8 equivalents) in THF at 78 "C followed by slow warming to ambient temperature. This reaction proceeds via the ferracyclobutenones 5a/5b, which are rapidly built from 4 and PMe3 at low temperature (vide supra). Complex 6 was hereby isolated in 93% yield. The nucleophile-induced CO insertion reactions of 4 and 5a/5b to give 6 are in so far remarkable, as these provide for the first time direct experimental evidence for the intermediacy of 16- and 18-electron ferracyclobutenones in iron-centered alkyne-carbonyl coupling reactions to give ferracyclopent-3-ene-2,5-dione~?~~ Complex 6 is thermally stable at ambient temperature, but eliminates CO and PMe3 either upon heating in refluxing toluene or upon melting at 127 "C to afford selectively the alkyne complex 10 (Scheme 4). The molecular structure of 6 reveals a slightly distorted octahedral complex, in which the two PMe3 ligands are transarranged and the five-membered ferracycle is planar (Figure 3). No n-bond delocalisation is evident in the ring: the Fe-CacyI (Fe-Cl 1.996(2) A; Fe-C4 1.978(2) A) and C,I-C~ bond lengths (Cl-C2 1.489(2) A; C3-C4 1.530(2) A) compare well with Fe-C(sp2) and C-C single bond lengths, respectively, and the C2-C3 bond (1.353(3) A) is a C-C double bond. The difference in the two CacyI-Cp bond lengths arises from the different orientation of the amino groups, which results from their mutual steric repulsion. Thus, the C3-bonded amino group is planar (sum of angles at N2 = 359.0") and is less twisted with respect to the ring plane (interplane angle N2,C7,C8/C 1,C2,C3,C4,Fe = 30.2') than the C2-bonded amino group, which is arranged almost orthogonally to the ring plane (interplane angle N1 ,C5,C6/C1,C2,C3,C4,Fe = 81 .Oo). Furthermore, the C2-bonded amino group is not planar (sum of angles at N1 = 345.7") and the methyl groups (C5, C6) point away from the vicinal dimethylamino group (Figure 3). Complex 6 has similar spectroscopic features with other ferra~yclopent-3-ene-2,5-diones,~~ the most conspicuous feature being the low-field shifted NMR signal for the acyl-carbon atoms (6262.8; Table 2). The selectivity of the reaction of 4 with PMe3 depends on the ratio of the reactants and the reaction conditions. As described above, the ferracyclopent-3-ene-2,5-dione6 is exclusively formed, when an excess of PMe3 is used (ratio PMe3 : 4 = 8) and the reaction allowed to warm slowly from -78 "C to ambient temperature. However, as the ratio PMe3 : 4 is decreased and the reaction temperature is raised, the concomitant formation of the 18electron ferracyclobutenone 7 is observed (Scheme 4). For example, treatment of 4 with two equivalents of PMe3 in THF at -10 "C followed by warming to room temperature afforded the complexes 6 and 7, which were separated by column chromatography on basic alumina at 10 "C and isolated in 60 and 20% yields, respectively. The thermal behaviour of the 18-electron ferracyclobutenones 5a/5b was studied to account for these observations. Complexes 5a/5b are thermolabile and decompose slowly in the solid-state at 20 "C under vacuum to give selectively the ferracyclobutenone 8, which was isolated as an orange solid in 98% yield (Scheme 4). 8 reacts rapidly with one equivalent of PMe3 in THF at 0 - 20 "C

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0) centres

129

Figure 4. Diamond plot of the molecular structure of 7 (left side) and of 8 (right side). Thermal ellipsoids are set 50% probability and hydrogen atoms are omitted for clarity. 7: Fe-CI 2.077(4), Fe-C3 2.031(3), Selected bond lengths (A) and bond angles CI-C2 1.422(5), C2€3 1.395(5), CI -01 1.229(4), C2-NI 1.442(4), C3-N2 1.344(4), Fe-CI 4 2 95.2(2), CI -C2€3 102.2(3), C2-C3-Fe 98.1(2), C3-Fe-Cl 64.5(1), P'-Fe-P 174.15(5); 8: Fe-CI 1.888(2), Fe--.C2 2.218(2), Fe€3 1.928(2), CI 4 2 1.417(3), C 2 4 3 1.471(3), Cl-Nl 1.316(2), C2-N2 1.403(2), C3-01 1.223(2), Fe-CI-C2 83.0(1), C l - C 2 1 1 3 103.1(2), C2-C3-Fe 80.2(1), CI-Fe-C3 72.67(8).

r):

to afford exclusively the yellow ferracyclobutenone 7 in 78% yield (Scheme 4). In comparison, the decomposition of 5d5b in solution is not selective and follows two competitive pathways. The predominant pathway involves the decarbonylation of 5a/5b to afford the 16-electron ferracyclobutenone 8 and the minor pathway the elimination of PMe3 to give the 16-electron ferracyclobutenone 4. IR and NMR monitoring reveals that the decomposition of 5a/5b starts at ca. -5 "C, but is fast at ambient temperature being completed in.1 h. The product ratio 8 : 4 increases with decreasing temperature and is 4 : 1 at 20 "C and 8.3 : 1 at 0 0C.37On the basis of these results the observed change in the product selectivity of the reaction of 4 with PMe3 can be rationalized as follows: Complex 4 adds rapidly PMe3 at -78 "C to give 5a/5b, which then undergoes a nucleophile-induced CO insertion reaction with PMe3 at ca. -15 "C to afford the ferracyclopent-3-ene-2,5-dione 6. At slightly higher temperatures (ca. -5 "C) and lower PMe3 concentrations, the decarbonylation of 5a/5b to give 8 becomes competitive to the CO insertion reaction. Complex 8 reacts then rapidly with PMe3 to afford 7. An alternative method to prepare complex 7 involves the reaction of the alkyne complex 9 with two equivalents of PMe3, which proceeds rapidly in pentane at ambient temperature to afford 7 in 95% yield (Scheme 4). This reaction occurs probably in two steps. The rate determining step involves a PMe3-promoted migratory CO insertion reaction into the metal-alkyne bond of 9 to give the 16-electron metallacyclobutenone 8, which is then rapidly trapped by PMe3 to give the complex 7. Such migratory CO insertion reactions into metal-alkyne bonds have been often implicated in mechanisms involving the coupling of alkynes with CO to give cyclic organic products:* but have been rarely ~bserved?'~Complex 7 is stable in solution at ambient temperature, but decomposes in refluxing toluene or upon melting at 1 18 "C to afford selectively after elimination of PMe3

130

Perspectives in Organometallic Chemistry

and CO the purple alkyne complex 10 (Scheme 4). In comparison, complex 8 melts at 84 'C without degradation, but decomposes very slowly in pentane or THF solution at ambient temperature to give a mixture of 7 and 9. The decomposition of the 16-electron ferracyclobutenone 8, and the above mentioned reaction of the alkyne complex 9 with PMe3 to give 7, suggest the presence of an equilibrium between 7 and 8, that is dependent on the PMe3-concentration. Complex 7 displays similar spectroscopic and structural features with 3 and 5d5b (Tables 1 and 2, Figure 4). The IR spectrum of 8 in THF displays a pair of very strong t(C=O) absorption bands at 1971 and 1912 cm-' and a pair of less intense *CEO) absorption bands at 1978 and 1912 cm-I, which suggests the presence of at least two stereoisomers in solution (Table 1). In addition, the IR spectrum of 8 shows two absorption bands of medium intensity at 1701 and 1607 cm-I, which are assigned to the t(C=O) and t(CgCa-N) vibrations of the ferracycle, respectively. These bands appear at lower wavenumbers than those of 4. A shift of the t(C=O) and t(CgC,-N) absorption bands to lower wavenumbers is also observed in the 18-electron ferracyclobutenones (3 > 5d5b > 7), as the electron density is increased at the metal center (Table 1). The 31P{1H)spectrum of 8 in THF at -89 'C shows three singlet signals at 824.6, 27.7 and 30.4 in the integral ratio of 10 : 1 : 3.5. This indicates that all conceivable square-pyramidal stereoisomers of 8 with the C3 ligand occupying two basal coordination sites, are present in sol~tion.~' This is As the also confirmed by the low-temperature 'H and I3C{'H) NMR spectra of temperature is raised the 31PNMR signals of 8 first broaden, then coalesce at Tc 205 K and appear in the fast exchange limit spectrum at 25 "C as one singlet at 824.9 (Table 1). This process is reversible and indicates a rapid interconversion of the stereoisomers in solution on the NMR time scale. The molecular structure of one of these stereoisomers was determined by X-ray crystallography and shows a distorted square-pyramidal complex with the PMe3 ligand occupying the trans position to the C m M e 2 ) atom in the basal plane (Figure 4). The four-membered ferracycle is puckered (folding angle Fe,Cl ,C3/C1,C2,C3 = 129.4") and displays the same structural features as in 4. Complexes 4, 6 and 7 are useful starting materials for the synthesis of $-alkyne complexes. Thus, thermal decarbonylation of 4 in refluxing toluene affords the red alkyne complex 9 in 95% yield (Scheme 4). Similarly, thermal decarbonylation of 6 or 7 in refluxing toluene gives exclusively the purple alkyne complex 10, which was isolated in 95 and 60% yield, respectively. Both alkyne complexes are thermally stable solids, which melt at 71 "C (9) and 84 "C (lo), respectively. Their thermal stability is remarkable taking into consideration, that the analogous complexes [Fe(C0)3-,Ln($-RC=CR)] (n = 0, 1; L = phosphane; R = H, alkyl) have been so far suggested to be reactive intermediates in carbon 1 substitutions reactions of the alkyne complexes [Fe(C0)4-,Ln( $-RC=CR)] (n = 0, l).4 The thermal stability of 9 and 10 provides another striking example for the stabilising +M effect of the amino groups. Both alkyne complexes are distorted square pyramidal with two CO ligands (C3-01, C4-02) and the alkyne occupying the basal coordination sites (Figure 5). Distortion results from the small bite angle of the alkyne ligand (9 43.8(1)"; 10 43.6(1)") and the bending of the apical ligand away from the alkyne moiety (9: C,-Fe-C5 114.9'; 10: C,-Fe-P 110.9'). The short Fe-Calkyneand Ca1hne-N bonds of 9 and 10, which range from 1.844(3) to 1.85l(3) 8, and from 1.318 to 1.336(3) A, respectively, the long (C-C)dhne bond of 9 (1.375(4) A) and of 10 (1.374(3) A), as well as the planarity of the coordinated bis(N,N-dimethy1amino)acetylene give evidence for the presence of a four-electron donor alkyne ligand and an extensive n-electron delo~alization.'~~ This is further confirmed by the spectroscopic data of 9 and 10, such as the high frequency of the t(N-C-C-N) vibration (1698 and 1674 cm-I, respectively), or

-

v

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0) centres

131

the downfield-shifted resonance signal for the alkyne-carbon atoms at S 193.8 and 190.5, respectively (Table 2). In addition, the temperature dependent ‘H NMR spectra of 9 in D8toluene and of 10 in CDC13 reveal that rotation of the dimethylamino groups about the (2dhne-N bonds is restricted on the NMR time scale giving rise to two methyl resonance signals in the slow-exchange limit spectra (9: Ac‘ > 13.3 kcal mol-’, T, > 293 K, Av = 306 Hz;10: A@ = 11.4 kcal mol-*, T, = 241 K, Av = 96 Hi).

Figure 5. Diamond plot of the molecular structure of 9 (lep side) and of 10 (right side). Thermal ellipsoids are set at 50% probability and hydrogen atoms are omitted for clarity. Selected bond lengths (A) and bond angles p): 9: Fe-CI 1.845(3), F e 4 2 1.844(2), CI -C2 1.375(4), Cl -NI 1.318(4), C2-N2 1.333(4), CI -Fe-C2 43.8(1), C3-Fe-C4 . 5 114.9; 10: Fe -Cl I , 85 I (3), 99.0(2), C3-Fe 4.599.7(2), C4-Fe -C5 99.0(2), C,,, -Fe 4 Fe 4 2 I . 850(3), CI 4 2 1.374(3) CI -NI I . 336(3), C2-N2 I . 328(4), CI -Fe € 2 43.6(1), C3-FeX4 101.4(1), C3-Fe-P 93.27(8), C4-Fe-P 97.14(9), C,-Fe-P 110.9 (C, denotes the midpoint of the alkyne C-C bond). Ongoing studies show that the complexes 4 - 10 can be functionalized by various electrophiles to give n-electron delocalized systems, and that the ferracyclobutenone 4 undergoes with isocyanides and alkynes selective insertion reactions into the Fe-C bonds to afford a multitude of “electron-rich” ferra~ycles.2~

2.3 Summary The reaction of Fe(C0)S with Me2N-C=C-NMe2 follows an associative reaction pathway to afford the ferracyclobutenone [Fe(C0)3{ q1:q1-C(NMe2)C(NMe2)C(0)}](4). A variety of unprecedented ferracycles have been selectively prepared and fully characterized taking advantage of the high reactivity of 4 and the stabilizing +M effect of the dimethylamino groups. These include the 18-, and 16-electron ferracyclobutenones [Fe(C0)4,(PMe3)n{ q’:q1-C(NMe2)C(NMe2)C(0)}] (n = 1, 2) and [Fe(C0)2(PMe3){ql: qlC(NMe2)C(NMe2)C(O))], respectively, or the #-alkyne complexes [Fe(C0)3,,(PMQ)~{#-C2(NMe2)2}] (n = 0, 1). Their interconversion provide for the first time direct experimental evidence for the intermediacy of metallacyclobutenones in metal-centered alkyne-carbonyl coupling reactions.

132

Perspectives in Organometallic Chemistry

3 ACKNOWLEDGEMENTS We thank the Humboldt Universitat zu Berlin and the Deutsche Forschungsgemeinschaft (Graduiertenkolleg "Synthetische, mechanistische und reaktionstechnische Aspekte von Metallkatalysatoren" (GK 352)) for financial support, Dr. B. Ziemer, Dr. G. Kociok-Kohn and P. Neubauer for the single-crystal X-ray diffraction studies, T. Nadulski for assistance in the experimental work, and Prof. J. Takats for fruitful discussions. 4 REFERENCES

( a ) W. Reppe, 0. Schlichting, K. Klager and T. Toepel, Justus Liebigs Ann. Chem., 1948, 560, 1; ( b ) W. Reppe and W. Schweckendick, Justus Liebigs Ann. Chem., 1948, 560, 104; (c) G. Wilke, Angew. Chem., Int. Ed. Engl., 1988, 27, 185, and references therein. Selected review articles: (a) P. M. Maitlis, Acc. Chem. Res., 1976, 9, 93; ( b ) K. P. C. Vollhardt, Acc. Chem. Res., 1977, 10, 1; (c) K. P. C. Vollhardt, Angew. Chem., Int. Ed. Engl., 1984,23, 539; (d) H. Bonnemann, Angew. Chem., Int. Ed. Engl., 1985,24,248; ( e ) W. J. Youngs, C. A. Tessier and J. D. Bradshaw, Chem. Rev., 1999,99,3 153; v) S. Saito and Y. Yamamoto, Chem. Rev., 2000,100,2901. Selected review articles on the Pauson-Khand reaction: ( a ) N. E. Shore, Chem. Rev., 1988, 88, 1081; ( b ) 0. Geis and H.-G. Schmalz, Angew. Chem., Int. Ed. Engl., 1998, 37, 91 1; (c) Y. K. Chung, Coord. Chem. Rev., 1999,188,297; (d)K. M. Brummond, and J. L. Kent, Tetrahedron, 2000, 56, 3263; ( e ) T. Sugihara, M. Yamaguchi and M. Nishizawa, Chem. Eur. J., 2001, 7, 1589; v) N. Jeong, B. K. Sung, J. S. Kim, S. B. Park, S. D. Seo, J. Y. Shin, K. Y. In and Y. K. Choi, Pure Appl. Chem., 2002, 74, 85; (g) M. A. Pericas, J. Balsells, J. Castro, I. Marchueta, A. Moyano, A. Riera, J. Vazquez and X. Verdaguer, Pure Appl. Chem., 2002,74, 167. Selected review articles on the Dotz benzoannulation reaction: ( a ) K. H. Dotz, Angew. Chem., Int. Ed. Engl., 1984, 23, 587; ( b ) W. D. Wulff, in Comprehensive Organometallic Chemistry 11, eds. E. W. Abel, F. G. A. Stone, G. Wilkinson, and L. S. Hegedus, Pergamon, New York, 1995, vol. 12, p. 469; (c) R. Aumann and H. Nienaber, Adv. Organomet. Chem., 1997, 41, 163; (d) A. de Meijere, H. Schirmer, M.. Duetsch, Angew. Chem., Int. Ed. Engl., 2000,39, 3964. Other names used for 1 and 2 in the literature are: Bis(dialkylamino)ethyne, Tetraalkylethynediamine or N,N,N',N-Tetraalkylethynediamine. (a) R. B. King and C. A. Harmon, Inorg. Chem., 1976,15, 879; ( b ) G. G. Cash, R. C. Pettersen and R. B. King, J. Chem. SOC.,Chem. Commun., 1977,30; (c) C. Wilcox and R. Breslow, Tetrahedron Lett., 1980, 21, 3241; (d) T. LePage, K. Nakasuji and R. Breslow, Tetrahedron Lett., 1985, 26, 5919; (e) J. Heck, K.-A. Kriebisch, W. Massa and S. Wocadlo, J. Organomet. Chem., 1994,482,81. ( a ) K. H. Dotz and C . G. Kreiter, Chem. Ber., 1976, 109, 2026; (b) K. H. Dotz, D. Neugebauer, Angew. Chem., Int. Ed. Engl., 1978, 17, 851; (c) H. Fischer, T. Meisner and J. Hofmann, J. Organomet. Chem., 1990,397, 41; (6)C. Hartbaum, E. Maw, G. Roth, K. Weissenbach and H. Fischer, Organometallics, 1999, 18, 261 9. (a) H. Fischer and I. Pashalidis, J. Organomet. Chem., 1988,348, C1; (b) H. Fischer, J. Hofmann, U. Gerbing and A. Tiriliomis, J. Organomet. Chem., 1988, 358, 229; (c) H. Fischer, K. Treier and J. Hofmann, J. Organomet. Chem., 1990,384,305.

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0) centres

9 10 11 12

13

14

15

16

17

133

( a ) H. Fischer, 0. Podschadly; A. Friih, C. Troll, R. Stumpf and A. Schlageter, Chem. Ber., 1992, 125, 2667; (b) H. Fischer, 0. Podschadly, G. Roth, S. Herminghaus, S. Klewitz, J. Heck, S. Houbrechts and T. Meyer, J. Organomet. Chem., 1997,541,321. M. L. Listemann and R. R. Schrock, Organometallics, 1985,4,74. S. R. Klopfenstein, C. Kluwe, K. Kirschbaum and J. A. Davies, Can. J. Chem., 1996, 74,2331. Selected references for the effects of amino group substitution in msystems: (a) B. Cetinkaya, G. H. King, S. S. Krishnamurthy, M. F. Lappert and J. B. Pedley, Chem. Commun., 1971, 1370; (b) K. Elbl, C. Krieger and H. A. Staab, Angew. Chem.,Int. Ed Engl., 1986, 25, 1023; (c) H. Bock, W. Kaim, M. Kira, L. RenC and H.-G. Viehe, Z. Naturforsch. Teil B, 1984, 39, 763; (d) J. M. Chance, B. Kahr, A. B. Buda, J. P. Toscano and K. Mislow, J. Org. Chem., 1988, 53, 3226; (e) H. Bock, K. Ruppert, K. Merzweiler, D. Fenske and H. Goesmann, Angew. Chem., Int. Ed. Engl., 1989, 28, 1684; v) K. Elbl-Weiser, C. Krieger and H. A. Staab, Angew. Chem., Int. Ed. Engl., 1990, 29, 21 1; (g) H. Bock, K. Ruppert, C. Nather, Z. Havlas, H.-F. Herrmann, C. Arad, I. Gobel, A. John, J. Meuret, S. Nick, A. Ruaschenbach, W. Seitz, T. Vaupel and B. Solouki, Angew. Chem., Int. Ed. Engl., 1992, 31, 550; (h) C. A. Zificsak, J. A. Mulder, R. P. Hsung, C. Rameshkumar and L.-L. Wei, Tetrahedron, 2001,57,7575. (a) H. G. Viehe and M. Reinstein, Angew. Chem., Int. Ed Engl., 1964, 3, 581; (b) S. Y. Delavarenne and H. G. Viehe, Chem. Ber., 1970, 103, 1209; (c) L. RenC, 2. Janousek and H. G. Viehe, Synthesis, 1982, 645; (d) R. van der Heiden and L. Brandsma, Synthesis, 1987, 76; (e) L. Brandsma and H. D. Verkruijsse, Synth. Commun., 1991,21, 81 1. ( a ) C. T. Lam, P. W. R. Corfield and S. J. Lippard, J. Am. Chem. SOC.1977,99, 617; (b) C. M. Giandomenico, C. T. Lam and S. J. Lippard, J. Am. Chem. SOC.,1982, 104, 1263; (c) R. Hoffmann, C. N. Wilker, S. J. Lippard, J. L. Templeton and D. C. Brower, J. Am. Chem. SOC., 1983, 105, 146; (d) R. N. Vrtis and S. J. Lippard, Isr. J. Chem., 1990, 30, 331; (e) E. M. Carnahan and S. J. Lippard, J. Chem. Soc., Dalton Trans., 1991, 699; v) E. M. Carnahan, J. D. Protasiewicz and S. J. Lippard, Acc. Chem. Res., 1993,26,90; (g) J. A. Acho and S. J. Lippard, Organometallics, 1994,13, 1294. ( a ) A. C. Filippou and W. Griinleitner, J. Organomet. Chem., 1990,398, 99; (b) A. C. Filippou and W. Griinleitner, J. Organomet, Chem., 1990,393, C 10; (c) A. C. Filippou and W. Grunleitner, Z. Naturforsch. Teil B, 1991, 46, 216; (d) A. C. Filippou, C. Volkl, W. Griinleitner and P. Kiprof, J. Organomet. Chem., 1992,434,201. (a)A. C. Filippou and W. Griinleitner, Z. Naturforsch. Teil B, 1989, 44, 1023; (b) A. C. Filippou, Polyhedron, 1990, 9, 727; (c) A. C. Filippou, C. Volkl, W. Griinleitner and P. Kiprof, Angew. Chem., Int. Ed. Engl., 1990, 29, 207; (d) A. C. Filippou, C. Volkl, W. Griinleitner and P. Kiprof, Z. Naturforsch. Teil B, 1990, 45, 351; (e) A. Mayr and C. M. Bastos, J. Am. Chem. Soc., 1990, 112, 7797; v) A. Mayr and C. M. Bastos, Prog. Inorg. Chem., 1992, 40, 1; (g) A. Mayr and H. Hoffmeister, Adv. Organomet. Chem., 1991, 32, 227; (h) A. C. Filippou, B. Lungwitz and G. KociokKohn, Eur. J. Inorg. Chem., 1999, 1905; ( i ) A. J. L. Pombeiro, M. F. C. Guedes da Silva, R. A. Michelin, Coord. Chem. Rev., 2001,218,43. (a) A. C. Filippou, W. Griinleitner, C. Volkl and P. Kiprof, Angew. Chem., Int. Ed. Engl., 1991,30,1167; (b) A. Mayr, C. M. Bastos, N. Daubenspeck, G. A. McDermott, Chem. Ber., 1992, 125, 1583; (c) A. C. Filippou, P. Hofmann, P. Kiprof, H. R. Schmidt and C. Wagner. J. Orpanomet. Chem.. 1993, 459,233. (d)R. A. Henderson,

134

18 19

20

21

22

23

24 25 26 27 28 29

30

Perspectives in Organometallic Chemistry

A. J. L. Pombeiro, R. L. Richards, J. J. R. Frausto da Silva and Y. Wang, J. Chem. SOC.,Dalton Trans., 1995, 1193. A. Bouvy, Z. Janousek and H. G. Viehe, Synthesis, 1983,718. (a) H. W. Sternberg, R. Markby and I. Wender, J. Am. Chem. SOC.,1958,80, 1009; (b) W. Hubel, H. Braye, A. Clauss, E. Weiss, U. Kruerke, D. A. Brown, G. S . D. King and C. Hoogzand, J. Inorg. Nucl. Chem., 1959, 9, 204; ( c ) G. N. Schrauzer, J. Am. Chem. Soc., 1959, 81, 5307; (d) E. Weiss, W. Hubel and R. Merenyi, Chem. Ber., 1962, 95, 1155; ( e ) W. Hubel, in Organic Syntheses via Metal Carbonyls, eds. I. Wender and P. Pino, Wiley-Interscience, New York, 1968, vol. 1, p. 273 and references therein; v) W. R. Fehlhammer and H. Stolzenberg, in Comprehensive Organometallic Chemishy, eds. G. Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon Press, Oxford, 1983, vol. 4, p. 545, and references therein. Selected examples of tricarbonyl( v4-cyclopentadienone) iron complexes are found in: ( a ) E. Weiss, R. MCrenyi and W. Hubel, Chem. Ber., 1962,95, 1170; (b)J. L. Boston, D. W. A. Sharp and G. Wilkinson, J. Chem. SOC., 1962, 3488; ( c ) C. G. Krespan, J. Org. Chem., 1975, 40, 261; (d) D. Fornals, M. A. Pericas, F. Serratosa, J. Vinaixa, M. Font-Altaba and X. Solans, J. Chem. SOC.,Perkin Trans., 1987,2749. ( a ) A. J. Pearson and A. Perosa, Organometallics, 1995, 14, 5178, and references therein; ( b ) H.-J. Knolker, E. Baum, H. Goesmann and R. Klaus, Angew. Chem., Int. Ed. Engl. 1999, 38, 2064, and references therein; (c) H.-J. Knolker and S. Cammerer, Tetrahedron Lett., 2000, 41, 5035; (d) J. E. Imbriglio and J. D. Rainier, Tetrahedron Lett., 2001,42,6987. ( a ) R. B. King, in The Organic Chemistry oflron, eds. E. A. Koerner Von Gustorf, F.W. Grevels and I. Fischler, Academic Press, New. York, 1978, vol. 1, p. 397; (b) M. Poliakoff and E. Weitz, Acc. Chem. Res., 1987, 20, 408; (c) N. Leadbeater, Coord. Chem. Rev,, 1999,188,35. ( a ) K. E. Lewis, D. M. Golden and G. P. Smith, J. Am. Chem. SOC., 1984,106, 3905; ( b ) J.-K. Shen, Y.-C. Gao, Q.-Z. Shi and F. Basolo, Inorg. Chem., 1989, 28, 3404; (c) J. Li, G. Schreckenbach and T. Ziegler, J. Am. Chem. SOC.,1995,117,486; (d) A. W. Ehlers and G. Frenking, Organometallics, 1995,14,423. J. Cooke and J. Takats, J. Am. Chem. SOC.,1997,119,11088. A. C. Filippou and T. Rosenauer, Angew. Chem., Int. Ed., 2002,41,2393. L. L. Padolik, J. C. Gallucci and A. Wojcicki, J. Am. Chem. SOC.,1993, 115,9986. (a) T. Mao, Z. Zhang, J. Washington, J. Takats and R. B. Jordan, Organometallics, 1999,18,2331; ( b ) M. Barrow, N. L. Cromhout, A. R. Manning and J. F. Callagher, J. Chem. SOC.,Dalton Trans., 2001, 1352. (a) J. Park and J. Kim, Organometallics, 1995, 14, 4431; ( b ) A. C. Filippou and T. Rosenauer, unpublished results. ( a ) A. G. Orpen, L. Brammer, F. H. Allen, 0. Kennard, D. G . Watson and R. Taylor, J. Chem. SOC.,Perkin Trans., 1989, S1 - S83; ( b ) A. G. Orpen, L. Brammer, F. H. Allen, 0. Kennard, D. G. Watson and R. Taylor, in Structure Correlation, eds. H. B. Biirgi and J. D. Dunitz, VCH, Weinheim, 1994, vol. 2. 4a was calculated using different methods (RI-MP2, B3LYP, RI-BP86) and basis sets (TZVP, 6-3 1 1 G* and cc-pVTZ). The Atoms in Molecules calculation (Figure 1, right side) was Derformed usine MORPHY98 (P. L. A. PoDelier and R. G. A. Bone).

Selective C-C coupling reactions of Me2N-C=C-NMe2at iron(0) centres

135

31 ( a ) R. Schobert, J. Organomet. Chem., 2001,617-618, 346; ( b ) N. Le Gall, D. Luart, J.-Y. Salaun, H. des Abbayes and L. Toupet, J. Organornet. Chem., 2001, 617-618, 483. 32 CRC Handbook of Chemistry and Physics, 82ndedn., ed. D. R. Lide, CRC Press, Boca Raton, 2001-2002,9-1 - 9-14. 33 ( a ) R. Herges and D. Geuenich, J. Phys. Chem., 2001, 105, 3214; (b)NBO 5.0 (E. D. Glendening, J. K. Badenhoop, A. E. Reed, J. E. Carpenter, J. A. Bohmann, C. M. Morales and F. Weinhold). 34 Four independent molecules with marginally different bonding parameters were found in the unit cell. The discussion is based on the unweighted mean values xu. The standard deviation 0 of xu is given in parentheses and was calculated by c? = E(x, xu)2/(n-1),x, = individual value, n = 4. 35 (a)R. Burt, M. Cooke and M. Green, J. Chem. SOC.A , 1970,2981; (b)W. Wong, S. J. Singer, W. D. Pitts, S. F. Watkins and W. H. Baddley, J. Chem. SOC., Chem. Commun., 1972,672; ( c ) P. A. Corrigan, R. S. Dickson, G. D. Fallon, L. J. Michel and C. Mok, Aust. J. Chem., 1978,31, 1937; (6) P. A. Corrigan and R. S . Dickson, Aust. J. Chem. 1979, 32, 2147; (e) J. Foerstner, A. Kakoschke, R. Wartschow and H. Butenschon, Organometallics, 2000,19,2 108. 36 M. Barrow, N. L. Cromhout, D. Cunningham, A. R. Manning and P. McArdle, J. Organornet. Chem., 2000,612,6 1 , and references therein. 37 Two minor products, complex 7 and the ferracyclopent-3-ene-2,5-dione [Fe(C0)3(PMe3){q' :q1-C(0)C(NMe2)C(NMe2)C(O)}](1l), are also formed upon the decomposition of 5a/5b in solution: Complex 7 results from the reaction of complex 8 with the released PMe3 and complex 11 from the carbonylation of 5a/5b. The latter reaction was observed only, when the decomposition of 5a/5b was carried out in a sealed tube. Formation of both products is suppressed, if the volatiles PMe3 and CO are being removed during the decomposition of 5a/5b. 38 ( a ) C. W. Bird, E. M. Briggs and J. Hudec, J. Chem. SOC.C, 1967, 1862; (b) P. Pino and G. Braga, in Organic Synthesis via Metal Carbonyls, eds. I. Wender and P. Pino, Wiley-Interscience, New York, 1977, vol. 2, p. 419; ( c ) S. Otsuka and A. Nakamura, Adv. Organomet. Chem. 1975, 14,245; (d) J. Washington, R. McDonald, J. Takats, N. Menashe, D. Reshef and Y. Shvo, Organometallics, 1995,14,3996. 39 We suggest on the basis of steric arguments that the PMe3 ligand of the minor stereoisomer is located in the basal plane and is cis-arranged to the Ca(NMe2) atom. 40 I3C{IH} NMR data of the major stereoisomer of 8 in [DsITHF at -94 "C: 6 Cppm] 19.6 (d, 'J(I3C,3'P) = 27.5 Hz, PMe3), 40.1 and 42.5 (br, CpNMe2), 42.0 and 46.5 (C,NMe2), 58.0 (d, 3J( 13C,31P) = 4.2 Hz, Cp), 215.1 (d, 2J(13C,31P)= 15.9 Hz, CO), 219.0 d, 2J('3C,3'P) = 12.2 Hz, CO), 219.8 (d, 2J(13C,31P) = 13.8 Hz, Ca), 226.5 (d, 2 13 J( C, 'P) = 22.2 Hz, Cacyi). 41 ( a ) J. Pearson, J. Cooke, J. Takats and R. B. Jordan, J. Am. Chem. SOC.,1998, 120, 1434; ( b ) S. A. Decker and M. Klobukowski, J. Am. Chem. SOC.,1998,120,9342.

ROUTES TO FLUORINATED ORGANIC DERIVATIVES BY NICKEL MEDIATED C-F ACTIVATION OF HETEROAROMATICS

Thomas Brauna and Robin N. Perutzb a Fakultat fiir Chemie, Universitat Bielefeld, Postfach

100131, 33501 Bielefeld, Germany, e-mail: [email protected] b Department of Chemistry, University of York, York YO10 5DD, UK, e-mail: rnp [email protected]

1 INTRODUCTION

The introduction of fluorinated groups into organic molecules can cause a dramatic change in their physical properties, chemical reactivity and physiological activity. This is illustrated by the application of fluorinated pyrimidines or pyridines as liquid crystals, herbicides, insecticides, anti-cancer agents and antibiotics.' However, it is still a challenge for synthetic chemists to prepare the desired molecules. The typical synthetic routes to fluorinated azaheterocycles involve introduction of fluorine at key positions or functionalisation of the fluorinated aromatics e.g. by nucleophilic substitution of a flu~rine.''~ Our strategy for the synthesis of a polyfluorinated aromatic molecule is totally different and is initiated by the selective replacement of a fluorine atom by a transition metal.2 Once the aromatic ring is attached to the metal centre, the fluorinated organic ligand can then be derivatised to yield new fluoro-organic molecules.

2 C-F ACTIVATION REACTIONS OF FLUORINATED HETEROCYCLES AT NICKEL AND THEIR MECHANISMS Several methods have been reported for the activation of a carbon-fluorine bond at appropriate transition metal centres? Some of the discoveries are summarised in thorough reviews: One approach we have studied in the last few years is the fast oxidative addition of fluorinated heteroaromatics such as pentafluoropyridine, 2,3,4,5-tetrafluoropyridine, 2,3,5,6-tetrafluoropyridineor 2,4,6-trifluoropyrimidine at a nickel centre giving truns[ N ~ F ( ~ - C S N F ~ ) ( P1,E truns-miF(2-CsNF3H)(PEt3)2] ~~)~] 2, 3 and trans-[NiF(4-C4N2F2H)(PEt&] 4 in high yield (Scheme l).5-7The reactions are carried out in a non-polar solvent, typically hexane, at room temperature. The intermolecular reactions are regioselective and chemospecific for C-F over C-H activation. The specificity is particularly striking in the reaction to form 3 (Scheme 1). The reactions also proceed far more rapid1 than the anaIogous activation of hexafluorobenzene yielding truns-[NiF(csFs)(PEt3)2] 5J8The role of the nitrogen atom in the heterocycles in accelerating the reactions is not fully understood.

Routes to Jluorinatedorganic derivatives

137

Complexes 1 - 4 are representative of the class of nickel (aryl) fluoride complexes, that were unknown prior to our work. The X-ray structures of 2 and 4 show that nickel is square planar with the aryl group perpendicular to the coordination plane of nickel. The nickel-fluorine bond length is ca. 1.86 A, close to expectations from well-known Ni-0 bond lengths. The nickel-C(ary1) bond length is almost identical in length to the Ni-F bond. The most important solution characteristic is the fluoride "F NMR resonance that lies at high field, ca. 6 -370, and is coupled to the 31Pnuclei (ca. 47 Hz) and the 19Fnuclei (ca. 9 Hz) on the aryl carbon atoms ortho to nickel.

F Fy$; F Et3P-Ni-PEt3

I

F l:R=F

F

2:R=H 5

F F

F

F$$F

F F

F

Si(COD),]

+

-

PEt3

F

3 6

F

F

-

Et3P-Ni-PEt3

F$I; F

I

F

F

4

Et3P-Ni-PEt3

I

COD = 1,5-cyclwctadiene

F 7

Scheme 1 C-F activation ofjluorinated aromatics and heteroaromatics at nickel

Before considering the mechanism of reaction of the fluorinated heterocycles, we discuss the corresponding reactions of hexafluorobenzene and octafluoronaphthalene. There is strong evidence that precoordination of the aromatic compounds at the nickel centre is a crucial step in the activation of a C-F bond in fluorinated aromatic systems. This is indicated by the observed coordination and intramolecular activation of

138

Perspectives in Organometallic Chemistry

octafluoronaphthalene at (Ni(PEt&) yielding trans-[NiF(2-C10F7)(PEt3)2]7 (Scheme l).9 The crystal structure of the intermediate mi(q2-1,2-C10F8)PEt3)2] 6 shows asymmetric q2coordination of the aromatic system at nickel. The C-F bonds of the coordinated carbon atoms are extended and lie out of the octafluoronaphthalene plane. In addition, the Ni-C(2) is appreciably shorter than the Ni-C(l) bond (1.899(4) and 1.959(4) respectively, Fig. 1). Moreover, the rates of loss of 6 and formation of 7 are compatible with a concerted intramolecular oxidative addition of the octafluoro-naphthalene ligand forming the Ni(I1) C-F activation product 7. Although the structure suggests an incipient transition state for concerted C-F activation, DFT calculations indicate that the potential for distortion of the coordination geometry is very soft. It should be mentioned that Crespo et al. have obtained kinetic evidence for a concerted oxidative addition of fluoroaromatic substituents in imines at platinum.lo The hexafluorobenzene compound [Ni(q2-C6F6){ tBu2P(CH2)2PtBu2)] has also been synthesised and it has been shown that on heating it reacts to form [NiF(C6Fs){ ~ B u ~ P ( C H ~ ) ~ P ~but BU no~kinetics )], are reported."

c22

C18

Figure 1 An ORTEP diagram of 6. Ellipsoids are drawn at the 50 %probability level Tsou and Kochi studied the reactions of mi(PEt3)4] with Ar-X (X = I, Br, C1) and showed that there are two corn eting pathways leading to [Ni"(PEt3)2(Ar)X] and [Ni1(PEt3)3X]+ ArH, respectively." The second pathway is of major importance when X = I, contributes 4 0 % of product when X = Br, and is not observed for X = C1. Tsou and Kochi postulated that a tight ion-pair (Ni(PEt3)PArX-} precedes both products on the basis of solvent effects, substituents effects and deliberate addition of Ni(I).'* In our reactions with fluoroaromatics, there is direct evidence for a [Ni(PEt3)2(q2-arene)] intermediate, but we cannot exclude involvement of an ion pair in addition. There is no evidence for Ni' products at all. We note that C6F6- would be short-lived since this species dissociates fluoride in solution leading to C6F5*.13 Theoretical studies of the reaction of {Ni(PH&) with hexafluorobenzene show that product formation becomes increasing energetically favourable in the order [Ni(PH3)2(q2-c#6)] < cis-[Ni(PH3)2(CsFs)F] < trans[N~(PH~)~(C~FS)F]. l 4 In contrast to C-F oxidative addition, the corresponding reaction of benzene to form trans-[Ni(PH3)2(CgHg)H] is conspicuously unfavourable. The activation energy for conversion of [Ni(PH3)2(q2-C6F6)] to C ~ S - [ N ~ ( P H ~ ) ~ ( Cis~calculated F ~ ) F ] to be 97 ICJ IIIOI-~.~~ The reactions of fluoropyridines with [Ni(COD)2] proceed rapidly in a non-polar solvent with two equivalents of triethylphosphine and a slight excess of fluoropyridine, but

139

Routes to fluorinated organic derivatives

no intermediates have been observed. The heteroaromatic systems may undergo C-F activation via q2-coordination of the aromatic system or via nitrogen-coordination. The former coordination mode has been observed in [(q5-C5H5)Rh(PMe3)(q2-C5F5N)], while the latter was reported by Bercaw et al. in the cationic complex [(trneda)Pt(CH3)(NCsFs)lBAr’4 [Ar’ = 3,5-C&(CF3)2]. l5,l6 Density functional calculations on coordination modes of pentafluoropyridine at (Ni(PH&) indicate that q2-coordination via an aromatic C=C bond is preferred in this case. l4 The observed preference for C-F activation at the 2-position of pentafluoropyridine provides indirect evidence for concerted oxidative addition of the azaheterocycles via a three-centred transition state (Scheme 2, a). An alternative electron transfer reaction pathway via a tight ion pair (Scheme 2, b) would lead to a reaction in the 4-position as has been established for other such reaction^.^ A SNAr type nucleophilic mechanism via a Meisenheimer intermediate (Scheme 2, c) would probably also result in an attack at the 4position of pentafluoropyridine as has been observed in countless reaction^.^ Exceptionally, the nucleophilic attack of the phosphine PHfBu2 at pentafluoropyridine takes place at the 2-~0sition.l~ This regioselectivity has been explained by increased steric hindrance, but bulky anionic transition metal complexes such as [Co(CO)2(PPh3)2]- or [Rh(C0)2(PPh3)2]-react at the conventional 4-position of the heterocycle.

Concerted

(4

Electron transfer/ ion pair (b)

SNAr/Meisenheimer (c)

Scheme 2 Possible intermediates and transition state for the C-F activation of pentajuoropyridine at nickel Further evidence for a concerted pathway for the reaction of azaheterocycles at nickel is derived from competition experiments. They show that the nickel system reacts 4.5 times faster with pentafluoropyridine than with 2,4,6-trifluoropyrimidine,yet the pyrimidine undergoes nucleophilic attack thousands of times faster than the pyridine.lg Thus, we have strong evidence for a concerted oxidative addition, although nucleophilic attack of the nickel centre at the heterocycle remains a possibility which we cannot exclude entirely. 3 CHEMOSELECTIVITY OF C-F ACTIVATION REACTIONS Fluorinated heterocycles also bearing a chlorine atom generally undergo C-CI activation at (Ni(PEt3)2}. This has been demonstrated by the insertion of nickel in a C-Cl bond in 3chlorotetrafluoropyridine, 3,5-dichlorotrifluoropyridine and 5-chloro-2,4,6trifluoropyrimidine (Scheme 3).5120,2’ However, the activation of a C-F bond in the presence of a much weaker C-C1 bond in 5-chloro-2,4,6-trifluoropyrimidinecan be accomplished using the sterically more hindered tricyclohexylphosphine yielding trans-

140

Perspectives in Organometallic Chemistry

DiF(4-C4N2ClF2)(PCy3)2] 11 together with a minor product (18%), which was assigned as

tran~-[NiC1(4-C4N~ClF2)(PCy3)2].~~ Such an activation of a C-F bond in the presence of a C-Cl bond in the same ring has never been observed before. For comparison, Crespo et al. reported the C-F activation of the imine (C6Fs)CH=NCH2(2-clC6H4)at a Pt(I1) centre, but with the C-F and C-Cl bonds on different rings.22

(i) PEt3

p. F

F‘

I E~~P-P~~-PE~~ I CI

8:R=F 9: R = CI F

Et3P-Ni--PEt3

I

CI 10

Cy3P-Ni4Cy3

I

F 11

Scheme 3 Activation of azaheterocycles bearing a chlorine atom Preference for C-F bond activation over C-H bond activation is critical to the development of applications since tolerance of C-H bonds is essential. As demonstrated by the DFT calculations, the reactions at {Ni(PH3)2) are energetically unfavourable for C-H bond activation but kinetically and energetically favourable for C-F bond activation. l4 The observed preference for C-F activation over C-H activation at nickel contrasts with observations at a rhenium centre, {(q5-CsMes)Re(C0)2} . For instance, photochemical reaction of [(q5-C5Mes>Re(CO)2(N2)] with 1,4-difluorobenzene yields the C-H activation product [(q5-CsMes)Re(H)(C6H3F2)(CO)2]12 (Scheme 4).23 Comparable results are obtained with the more fluorinated benzenes C6HF5 and 1,2,4,5-C&F4. Thus, UV irradiation of [(q5-CsMes)Re(CO),] in the presence of C6HF5 affords [(qsCsMes)Re(H)(CsFs)(C0)2] 13 as the principal product.24However, it is C-F activation in combination with intramolecular C-H activation that dominates on photolysis of [(q5C5Mes)Re(CO)3] in neat C6F6 yielding the tetramethylfulvene complex [(q 6C ~ M ~ ~ C H ~ ) R ~ ( C ~ F14.25 S ) ( The C ~ )C-H ~ ] activation products 15 and 16 are generated

Routes to fluorinated organic derivatives

141

with the cyclopentadienyl analogue [(q5-C5Hs)Re(CO)3] and C6HFs or CsH2F4. There are minor by-products including bis(ary1) complexes produced by C-F activation of a second aromatic molecule as well as the binuclear complexes 17 and 18. F

qMe5

U

0

#Me5

1 hv, CO, -HF 14

H

15:X=F

17:X=F

16:X=H

18: X = H

Scheme 4 C-F and C-H activation reactions at rhenium Other transformations with a preference for C-H activation over C-F activation have While the complex [(q5-CsMe5)Rh(PMe,)(C2&)] and C6F6 been described at can be converted to the fluoro complex [(.rl5-CsMes)Rh(F)(C6F5)(PMe3)]by photochemical means in liquid hexafluobenzene, the thermal reaction of 1,4-difluorobenzene with [($CsMes)Rh(H)(Ph)(PMe3)] gives the C-H activation product [(.r15-CsMes)Rh(H)(C6F2H3)(PMe3)I. The cyclopentadienyl complex [(q5-C~Hs)Rh(PMe3)(CzH4)] shows a preference for C-H activation too reacting photochemically with 1,4-difluorobenzene to form [($C ~ H S ) R ~ ( H ) ( C ~ F ~ H ~ ) Density ( P M ~ ~functional )]. calculations for the oxidative addition of 1,4-difluorobenzeneat [(q5-C5H5)Rh(q2-C6F2H4)(PH3)] show that both C-F and C-H bond activation are energetically favourable (contrast nickel). They support a mechanism with

142

Perspectives in Organometallic Chemistry

concerted oxidative addition to the 16-electron fragment { (q5-CsH5)Rh(PH3)} and show that the preference for C-H activation is of kinetic origin.27 Preference for C-F over C-H bond activation can be achieved by reaction of some dihydride complexes with fluorobenzenes. Cis-[RuH2(drnpe)z] yields products of C-F bond activation with pentafluorobenzene, tetrafluorobenzenes or 1,2,3-trifluorobenzene.l 3 The reaction is unaffected by added fluoride. As postulated for other transformations at rhodium or iridium, this reaction is thought to proceed by an electron transfer mechanism rather than by a simple oxidative a d d i t i ~ n . ~ ' A ~ ~base-catalysed -~' mechanism has been postulated by W. D. Jones et al. in the C-F bond activation of fluorinated benzenes using [(q5-CsMes)Rh(H)2(PMe3)] as substrate.31This nucleophilic mechanism includes an attack of [(q5-C5Me5)Rh(H)(PMe3)]- at C6FsH and explains the observed preference for C-F over C-H activation. These reactions are less suited to formation of new organic products than the reactions at nickel because of the trans-octahedral structure of the ruthenium products and the difficulty of reductive elimination at { (q5-C5Me5)Rh(PMe3)). 4 REACTIVITY OF NICKEL FLUORIDES

Transition metal complexes bearing a fluoro ligand are increasingly regarded as valuable compounds in organometallic chemistry with interesting properties as catalysts or synthetic precursors.4932One other special feature of metal-fluoride complexes is that they are capable of coordinating hydrogen fluoride via hydrogen bonds, thus forming coordinated bifluoride (FHF).6y7*33 The bifluoride complexes trans-mi(FHF)(2-C5NF4)(PEt3)2] 19 and ~~-U~S-[N~(FHF)(~-C~N~F~H)(PE~~)~] 20 have been prepared by reaction of Et3No3HF with nickel fluorides and characterised in solution (Scheme 5).6y7An X-ray structural analysis of 20 suggests that the FHF interaction is best described as a hydrogen bond between a NiF moiety and HF.6 The Ni-F bond length is 1.908(3) A compared to 1.856(2) A for trans[N~F(~-CSNF~H)(PE~&] 2 indicating that the hydrogen bonding causes some lengthening of the Ni-F bond.

I

Et3P-Ni-PEt3

Et3P-Ni-PEt3

I F

I

1

f

FHF

FHF

4

Scheme 5 Formation of nickel bijluorides

19

20

I43

Routes to Juorinated organic derivatives

The reactivity of nickel fluoride compounds bearing polyfluoropyridyl ligands has also been inve~tigated.~ Fluoride may be abstracted with BF3 or with Me3Si derivatives. Thus, treatment of 1 with BF3-OEt2 in the presence of acetonitrile yields the cationic compound trans-Bi(2-CsNF&NCMe)(PEt3)2IBF4 21. W ~ ( O T ~ ) ( ~ - C S N F ~ ) (22 P Ecan ~ ~ )readily ~] be synthesised from 1 and Me3SiOTf.34Similarly, the chloride 23 can be formed by reaction of 1 with Me3SiC1. The reaction of 1 with HCI provides an alternative route to 23 (Scheme 6).7

Et3P-&-pEt3

I EtsP-Ni-PEts

Et3P-Ni-PEt3

I

CNtBu

29

28

CH3CN

I

$F

Et3P-Ni-PEt3

I Me

I OPh

I

Ph

25

24

NaBAr'4

'BAr'4

F

BF4

NaBAr'4

Et3P-Ni-PEt3

I

I

kCH3

27

NCCH3

21

26

Scheme 6 Reactivity of I Although free tetrafluoropyridine reacts rapidly with nucleophiles, the nickel complexes react with nucleophiles at the metal resulting in replacement of the fl~oride.~ Thus, we have successfblly replaced Ni-F in trans-[NiF(2-CsNF4)(PEt3)2] 1 by Ni-C bonds by reaction with Me2Zn or PhLi yieldin truns-WiMe(2-CsNF,+)(PEt3)2]24 and transwiPh(2-C,NF4)(PEt3),] 25 (Scheme 6)b4 Some of these reactions can be used in the synthesis of new non-metallated heterocycles as described below.

144

Perspectives in Organometallic Chemistry

The chloride and triflate derivatives are useful precursors in their own right. Treatment of trans-[NiC1(2-CsNF4)(PEt3)2] 23 with HBF4 abstracts PEt3 to afford the binuclear complex [NiCI{ ~-K~(C,N)-(~-C~NF~)}(PE~~)]~ 26.7 The X-ray crystal structure of 26 reveals a “butterfly”-shaped dimeric complex with square-planar coordination at both nickel atoms (Scheme 6, Fig. 2). Reaction of trans-[Ni(OTf)(2-CsNF4)(PEt3)2] 22 with NaBAr’4 [Ar’ = 3,5-CsH3(CF3)2] and acetonitrile or CNtBu gives trans- i(2-CsNF4)(NCMe)(PEt3)2]BAr’4 27 and trans-[Ni(2-C5NF4)(CNtBu)(PEt3)2]BAry428.73p The triflate complex 22 can also be converted into the phenoxy compound trans-wi(OPh)(2-CsNF4)(PEt3)2] 29 on treatment with NaOPh.

Figure 2 An ORTEP diagram of 26. Ellipsoids are drawn at the 50 %probability level

5 NICKEL-MEDIATED SYNTHESIS OF NEW HETEROCYCLES

The reactions of fluorinated precursors at nickel provide access to fluorinated heterocycles, which are otherwise inaccessible. Overall, we start with a commercially available fluorinated heterocycle, selectively remove a fluorine from it by reaction at nickel, and then fhnctionalise further. The unusual substitution patterns in the final product arise from the initial chemo- and regioselective attack by nickeL2 Note that no tetrafluoropyridyl complexes with the metal in the 2-position had been described p r e v i o u ~ l y The . ~ ~ new heterocycles can usually be obtained in an overall yield of 20-50% based on the organofluoro starting compound. This strategy is demonstrated by the activation of pentafluoropyridine in the 2-position followed by the sequential methylation of trans-[NiF(2-CsNF4)(PEt3)2] 1 and reaction with CO, which affords the ketone 2-C5F4NC(=O)Me by elimination (Scheme 7).34 It is normally very difficult to prepare tetrafluoropyridines substituted in the 2-p0sition.2’~ Complex 1 can be used to synthesise a variety of these compounds: for instance, reaction of 1 with iodine affords 2-CsFsN1, while prolonged treatment of 1 with HCl gives 2C S F ~ N H .On ~ . ~admission of air to a solution of the methyl complex trans-piMe(2CsNF4)(PEt&] 24, the reductive elimination product 2-CsF4NMe is formed.34

Routes to fluorinated organic derivatives

145 F

F

0

Ni(CODhlPEt3

ZnMe2 Et3P-Ni-PEt3

Et3P-Ni-PEt3

I

[Nil

I

F

F F

Me 1

24

- [Nil

- [Nil

F

Scheme 7 Nickel-mediated derivatisation of pentajluoropyridine The nickel-mediated approach provides an unusual entry to halopyrimidines bearing three different substituents by removal of a fluorine from 5-chlorotrifluoropyrimidine (Scheme 8).2*Treatment of trans-[NiF(4-C4N2ClF2)(PCy&] 11 with HCl or iodine affords 5-chloro-2,4-difluoropyrimidine and 5-chloro-2,6-difluoro-4-iodopyrimidine.

F

Cy3P-Ni-pcy3

I

F

Al2 11

- [Nil

F

F

Scheme 8 Nickel-mediated derivatisation of 5-chlorotrijluoropyrimidine In another intriguing example, the metal-mediated C-F activation of 2,4,6trifluoropyrimidine again has the attraction of producing different regiochemistry from the typical organic route (Scheme 9).6 Treatment of trans-[NiF(4-C4N2F~H)(PEt3)2]4 with CsOH in the presence of 2,4,6-trifluoropyrimidine affords a nickel derivative of a pyrimidin-4-one 30 with the heterocyclic unit bound as an anionic ligand via a nitrogen atom at the metal. On treatment of 30 with HC1 the free difluoropyrimidin-4-one can be obtained. Note that the reaction of 2,4,6-trifluoropyrimidinewith NaOH results in the formation of the difluoropyrimidin-2-one.

Perspectives in Organometallic Chemistry

146

y

F N F

E~JP-N~--PE~~

I

F

-

6

I

F 1

I

(1

F N F

0

HCI

Et3P-Ni-PEt3

30

(i)NaOH

F

(ii) HCI

Scheme 9 Synthesis offluorinatedpyrimidinones

6 CATALYTIC CONVERSIONS BY C-F ACTIVATION AT NICKEL Catalytic C-F activation of polyfluoroaromatics has become a reality, but the examples are sparse and have been limited to the formation of new C-H or C-Si bond^?*'^,^^ Milstein observed the catalytic conversion of hexafluorobenzene to pentafluorobenzene (Scheme 10) using hydrogen and [HRh(PMe3)4] as cataly~t.2~ Murai and coworkers reported the rhodium-mediated silylation of pentafluoroacetophenone?6 Other research groups demonstrated catalytic conversions of monofluorinated aromatics forming new C-H or C-C We achieved the catalytic conversion of pentafluoropyridine and 2,3,5,6tetrafluoropyridine to their 2-vinyl derivatives by cross coupling reactions with H2C=CHSnBu3 using a nickel catalyst (Scheme They represent the first catalytic C-C coupling reactions involving C-F activation of a polyfluorinated molecule. We also found that the cross-coupling reactions are likely to proceed via the formation of the q2vinylpyridine complex [Ni{q2- 2-CsNF4(CH=CH2))(PEt3)2] 31, which was observed during the stoichiometric reaction of trans-[NiF(2-CsNF4)(PEt3)2] 1 with H2C=CHSnBu3 (Scheme 11). However, 31 is not stable in solution and two fbrther compounds are observed after 1 d of reaction. They were assigned as the C-F activation product 32 and the divinylpyridine complex [Ni{q2 - 2,6-CsNF3(CH=CH2)2)(PEt3)2]33. At present, the catalytic reactions are limited to a few turnovers, possibly because of competing decomposition pathways during the oxidative addition under catalytic condition^.^^

Routes to fluorinated organic derivatives

1

147

Me3P-Rh-PMe3

I

PMe3 H2, 85 psi 100°C, Et3N

F

F&F

F

10 % [Rh(COD)2]BF4

F

F

F

Me3SiSiMe3 13OoC

F

2o %

I

F

F

I F

F F

EtgP-Yi-PEtg

FqF F

F6Sitvle3

'

*

Fx F

PEt3, H2C=CHSnBu3 5OoC,THF, C S ~ C O ~

F

Scheme 10 Catalytic conversions offluorinated aromatics by C-F

y F

\

Bu3SnCH=CH2

F

Et3P

Et3P-Ni--fEt3

F

I

F

31

i

l

y F

'Ni-w\

F

Bu3SnCH=CH2

Et3P

F

F

Et3P-Ni---PEt3

I

F

33

Scheme 11 Reaction of 1 with BujSnCH=CH2

32

Perspectives in Organometallic Chemistry

148

7 OUTLOOK Future goals in the area of C-F activation at transition metals still involve the development of new fluorinated building blocks, which are not accessible by current technology. To find catalytic transformations with high turnover numbers will certainly be one of the major challenges. One of the next stages is to synthesise some compounds with biological activity via metal-mediated selective removal of fluorine from polyfluorinated precursors. The special properties of anionic fluorocarbon ligands bound to a transition metal centre will lead to new and unexpected reaction pathways. For instance, Hughes et al. have already demonstrated that fluorinated alkyl ligands at iridium can be hydrogenated (Scheme 12).4'942Another example is represented by the nucleophilic substitution of a fluorine atom by a phosphine in a perfluorovinyl ligand at a cationic nickel complex (Scheme 12).43

1

PEt3 Et3P-Ni-pEt3

I

CNtBu

A NaBAr'4

Et3P-Ni-PEt3

I

CNtBu

Scheme 12 Reactivity ofjluorinated l i g a n d ~ ~ " ~ ~ One other demanding goal is the activation and selective functionalisation of fluorinated alkenes or even alkanes in the coordination sphere of a metal. The heterogeneous, catalytic conversion of fluorinated alkanes and cycloalkanes to alkenic and aromatic compounds has alread been documented by the research groups of Richmond and Crabtree (Scheme 13)?vJ The homogeneous reduction of hexafluoropropene to propane or 1,l ,1trifluoropropane has been achieved recently using zirconium or rhodium complexes (Scheme 13).45B46 The zirconium-mediated conversion of 1 -fluorohexane into hexane has also been rep0rted.4'~~~ These examples show that the activation and functionalisation of fluorinated olefins and alkanes in the coordination sphere of a metal certainly holds out promise of krther surprising and exciting results in the near future.

Routes toJuorinated organic derivatives

F3cxF

F27

7F2

F2c\ CHCF2

149

Cp'ZFe, hv

F2FX

>

-

Zn, 2 LiO3SCF3, 2 LiF, - Zn(03SCF3)2

F2C,

F2

-

C IF ,CF2 C F2

[Cp42ZrH21

F2C=CF(CF3)

7 [Cp'2ZrH21

- 6 [Cp'2ZrHF]

t

[Cp'2Zr(n-propyl)H]

H2

+

H3CCH2CH3

Scheme 13 Metal-mediatedderivatisation of perfluorinated alkenes and alkanesu4' Acknowledgement

The work described in this article represents the work of members of the groups in York and Bielefeld who have contributed by their experimental work, their ideas and their enthusiasm. Important contributions have been made by C. L. Higgitt, R. Karch, D. Noveski, M. Reinhold, V. Schorlemer, M. I. Sladek and M. K. Whittlesey. We also have benefited fi-om collaborations with J. E. McGrady (York), A. H. Klahn and B. Oelckers (Valparaiso), 0. Eisenstein and F. Maseras (Montpellier) and S. Parsons (Edinburgh). We are indebted to the EPSRC, the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie for financial support. T. B. thanks P. Jutzi for his continuous support in Bielefeld. References

1 2

T. Hiyama, Organofluorine Compounds, Springer, Berlin 2000; R. E. Banks, B. E. Smart and J. C. Tatlow (eds.), Organofluorine Chemistry: Princijdes and CommercialApplications, Plenum, New York 1994. R. Dagani, Chem. Eng. News,2001,79,40.

150

3

4

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

26

Perspectives in OrganometallicChemistry

G. M. Brooke, J. Fluorine Chem., 1997, 86, 1; P. L. Coe and A. J. Rees, J. Fluorine Chem., 2000, 101, 45; A. J. Adamson, W. J. Jondi and A. E. Tipping, J. Fluorine Chem., 1996, 76, 67; H. Benmansour, R. D. Chambers, P. R. Hoskin and G. Sandford, J. Fluorine Chem., 2001,112, 133. J. Burdeniuc, B. Jedlicka and R. H. Crabtree, Chem. Ber./Recl., 1997, 130, 145. J. L. Kiplinger, T. G. Richmond and C. E. Osterberg, Chem. Rev., 1994, 94, 373. E. F. Murphy, R. Murugavel and H. W. Roesky, Chem. Rev., 1997, 97, 3425. T. G. Richmond in S. Murai (eds.) Topics in Organometallic Chemistry, Vol. 3, Springer, New York 1999, pp. 243-269. L. Cronin, C. L. Higgitt, R. Karch and R. N. Perutz, Organometallics, 1997, 16, 4920. T. Braun, S. P. Foxon, R. N. Perutz and P. H. Walton, Angew. Chem. Int. Ed. Engl., 1999,38,3326. S . J. Archibald, T. Braun, J. F. Gaunt, J. E. Hobson and R. N. Perutz, J. Chem. SOC., Dalton Trans., 2000,2013. D. R. Fahey and J. E. Mahan, J. Am. Chem. SOC.,1977,99,2501. T . Braun, L. Cronin, C. L. Higgitt, J. E. McGrady, R. N. Perutz and M. Reinhold, New J. Chem., 2001,25,19. M. Crespo, M. Martinez and E. de Pablo, J. Chem. SOC.,Dalton Trans., 1997, 1231. I. Bach, K.-R. Porschke, R. Goddard, C. Kopiske, C. Kriiger, A. Rufinska and. K. Seevogel, Organometallics, 1996, 15,4959. T. T. Tsou and J. K. Kochi, J. Am. Chem. SOC.,1979,101,63 19. M. K. Whittlesey, R. N. Perutz and M. H. Moore, J. Chem. SOC.,Chem. Commun., 1996,787. J. E. McGrady, R. N. Perutz and M. Reinhold, unpublished results. R. N. Perutz and M. K. Whittlesey, unpublished results. M. W. Holtcamp, J. A. Labinger and J. E. Bercaw, J. Am. Chem. SOC., 1997, 119, 848; M. W. Holtcamp, L. M. Henling, M. W. Day, J. A. Labinger and J. E. Bercaw, Inorg. Chim. Acta, 1998,270,467. Y. A. Veits, N. B. Karlstedt, A. V. Chuchuryukin and I. P. Beletskaya, Russ. J. Org. Chem., 2000,36,750. B. L. Booth and R. N. Haszeldine, J. Chem. Soc., Dalton Trans., 1975, 1843; I. P. Beletskaya, G. A. Artamkina, A. Y. Mil’chenko, P. K. Sazonov and M. M. Shtern, J. Phys. Org. Chem., 1996,9,3 19. R. D. Chambers, Y. A. Cherbukov, T. Tanabe and J. F. S. Vaughan, J. Fluorine Chem., 1995, 74, 227; R. D. Chambers, P. A. Martin, J. S. Waterhouse, D. L. H. Williams and B. Anderson, J. Fluorine Chem., 1982,20, 507. M. I. Sladek, T. Braun, B. Neumann and H.-G. Stammler, submitted. M. I. Sladek, T. Braun, B. Neumann and H.-G. Stammler, J. Chem. SOC.,Dalton Trans., 2002,297. M. Crespo, M. Martinez and J. Sales, J. Chem. SOC.,Chem. Commun., 1992, 822. J. J. Carbo, 0. Eisenstein, C. L. Higgitt, A. H. Klahn, F. Maseras, B. Oelckers and R. N. Perutz, J. Chem. SOC.,Dalton Trans., 2001, 1452. F. Godoy, C. L. Higgitt, A. H. Klahn, B. Oelckers, S. Parsons and R. N. Perutz, J. Chem. SOC., Dalton Trans., 1999,2039. A. H. Klahn, M. H. Moore and R. N. Perutz, J. Chem. Soc., Chem. Commun., 1992, 1699. S. T. Belt, M. Helliwell, W. D. Jones, M. G. Partridge and R. N. Perutz, J. Am. Chem. SOC.,1993,115, 1429; A. D. Selmeczy, W. D. Jones, M. G. Partridge and R. N. Perutz, Organometallics, 13, 1994, 522.

Routes to fluorinated organic derivatives

27 28 29 30 31 32

15 I

R. Bosque, E. Clot, S. Fantacci, F. Maseras, 0. Eisenstein, R. N. Perutz, K. B. Renkema and K. G. Caulton, J. Am. Chem. SOC.,1998,120,12634. M. Aizenberg and D. Milstein, Science, 1394, 265, 359. M. Aizenberg and D. Milstein, J. Am. Chem. SOC.,1995,117, 8674. 0. Blum, F. Frolow and D. Milstein, J. Chem. SOC., Chem. Commun., 1991,258. L. Edelbach and R. W. Jones, J. Am. Chem. SOC.,1997,119,7734. N. M. Doherty and N. W. Hoffman, Chem. Rev., 1991,91,553; B. L. Pagenkopf and E. M. Carreira, Chem. Eur. J., 1999, 5, 3437; V . V. Grushin, Chem. Eur. J., 2002,8, 1007.

45 46

J. Gil-Rubio, B. Weberndorfer and H. Werner, J. Chem. SOC,Dalton Trans., 1999, 1437; N. A. Jasim and R. N. Perutz, J. Am. Chem. SOC.,2000, 122, 8685; N. A. Jasim, R. N. Perutz, S. P. Foxon and P. H. Walton, J. Chem. SOC, Dalton Trans., 2001, 1676; D. C. Roe, W. J. Marshall, F. Davison, P. D. Soper and V . V. Grushin, Organometallics, 2000, 19,4575; M. K. Whittlesey, R. N. Perutz, B. Greener and M. H . Moore, J. Chem. SOC., Chem. Commun., 1997, 187; M. C. Pilon and V . V . Grushin, Organometallics, 1998, 17, 1774; V. J. Murphy, T. Hascall, J. Y. Chen and G. Parkin, J. Am. Chem. SOC.,1996, 118, 7428; V. J. Murphy, D. Rabinovich, T. Hascall, W. T. Klooster, T. F. Koetzle and G. Parkin, J. Am. Chem. SOC.1998, 120, 4372; H. W. Roesky, M. Sotoodeh, Y. Xu, F. Schrumpf and M. Noltemeyer, 2. Anorg. Allg. Chem., 1990,580, 13 1. T. Braun, S. Parsons, R. N. Perutz and M. Voith, Organometallics, 1999, 18, 1710. T. Braun and R. N. Perutz, unpublished results. Y. Ishii, N. Chatani, S. Yorimitsu and S . Murai, Chem. Lett., 1998, 157. R. J. Young and V . V . Grushin, Organometallics, 1999, 18, 294; H. Yang, H. Gao and R. J. Angelici, Organometallics, 1999, 18,2285. V . P. W. Bohm, C. W. K. Gstottmayr, T. Weskamp and W. A. Herrmann, Angew. Chem. Int. Ed. Engl., 2001, 40, 3387; Y. Kiso, K. Tamao and M . Kumada, J. Organomet Chem., 1973, 50, C12; D. A. Widdowson and R. Wilhelm, Chem. Commun., 1999, 2211; R. Wilhelm and D. A. Widdowson, J. Chem. Soc., Perkin Trans. I , 2002,3808. T . Braun, M. I. Sladek and R. N. Perutz, Chem. Commun., 2001,2254. A. L. Casado, P. Espinet, A. M. Gallego and J . M. Martinez-Ilarduya, Chem. Comm. 2001, 339; A. Jutand, K. K. Hii, M. Thornton-Pett and J. M. Brown, Organometallics, 1999,18, 5367. T . G. Richmond, Angew. Chem. Int. Ed., 2000,39,3241. R. P. Hughes and J. M. Smith, J. Am. Chem. SOC.,1999,121,6084; R. P. Hughes, S. Willemsen, A. Williamson and D. Zhang, Organometallics, 2002,21, 3085. T. Braun, B. Bliicker, V. Schorlemer, B. Neumann, A. Stammler and H.-G. Stammler, J. Chem. SOC.,Dalton Trans., 2002,2213. J. L. Kiplinger and T. G. Richmond, J. Am. Chem. SOC., 1996, 118, 1805; J. Burdeniuc and R. H. Crabtree, J. Am. Chem. SOC., 1995, 117, 10119; J. Burdeniuc and R. H. Crabtree, J. Am. Chem. SOC.,1996, 118, 2525; J. Burdeniuc and R. H. Crabtree, Organometallics, 1998,17, 1582. B. M. Kraft, R. J. Lachicotte and W. D. Jones, J. Am. Chem. SOC.,2000,122,8559. T . Braun, D. Noveski, B. Neumann and H.-G. Stammler, Angew. Chem. Int. Ed.,

47

B. M. Kraft, R. J. Lachicotte and W. D. Jones, J. Am. Chem. SOC.,2001,123, 10973.

33

34 35 36 37 38

39 40

41 42 43 44

2002,41,2745.

NOVEL q5- q6 REARRANGEMENT OF BIS(FLU0RENYL)LANTHANIDE COMPLEXES BY THE ADDITION OF AlR3

Hajime Yasuda Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, HigashiHiroshima 739-8527, Japan

1 INTRODUCTION

The metallotropic q5-q6 tautomeric equilibrium of late transition metal tricarbonyl complexes of substituted fluorenes and indenes has been reported for Mn, Cr and Mo complexes. For example, the abstraction of proton from [Mn(q6-C13Hl~)(C0)3]PF6by using bases such as potassium tert-butoxide or triethylamine generates Mn(q5-C13Hg)(C0)3 irreversibly.' The corresponding reaction using [Fe(q6-C13Hlo)(C0)3]PFgdoes not produce

I Mn(CO2 PF6

Figure 1

r/5-v6Rearrangementfor an Mn complex

an q5-coordinated product but maintains the q6-coordinated structure, Fe(q6C&s)(C0)3? Raising the temperature of 9-substituted q6-fluorenechromium tricarbonyl R

Figure 2

qs-$ Rearrangementfor a Cr complex

Novel

775 -

776 rearrangement of bis(juoreny1)lanthanide complexes

153

complexes to >O"C also provides an anionic q5-(fluorenyl)Cr(CO)3 as evidenced by 'H NMR.3 The reaction of alkyl halide with anionic q5-(indenyl)Cr(C0)3complex also gave q6-(indene)Cr(C0)3 irrever~ibly.~ The complex q5-(indznyl)Mo(C0)3H is reported to be less stable than q6-(indenyl)Mo(C0)3.'

oc/l0\CO CO

Mo

oc/

I

'(20

CO

Figure 3 $-q6 Rearrangement for a Mo complex

In sharp contrast to this behavior, the q5-q6rearrangement has not been reported for early transition metal indenyl and fluorenyl complexes. The reaction of AIR3 with q5(CsMes)Ln(THF)z generally affords q5-[(C5Me5)2Ln(p-R)2AlR& (M = Sm, Y)6 which exist in equilibrium with monomeric q5-(C5Me5)2Ln(p-R)2AlR2or a 1:1 A& adduct, q5(CgMe5)2Ln(pL-R)AlR2(THF)(Ln = Yb).7 The complexes q5-(C5Me5)2Ln(p-R)2AlR2can

R'

'R

Figure 4 Modes of addition of AIR3 to (C&fej)2Sm

also be formed by reaction of q5-(C5Me5)2LnX2with MeLi/AlR3.* We describe herein the first examples of the q5-q6 rearrangement of q5-bis(Me3Si-fluorenyl)Sm(THF)2 to q6bis(Me3Si-fluorenyl-AlR3)Sm by reaction with AlR3 and the conversion of q5-bis(Me3Sifluorenyl)Yb(THF)2 to q6-bis(Me3Si-fluorenyl-A1R3)-q5-(Me3Si-fluorenyl)Ybby reaction with AlMe3. The q6-complexes can convert to q5-species by the successive addition of excess THF. 2 METHODS AND RESULTS All operations were performed under an argon atmosphere by using standard Schlenk techniques. 'H NMR spectra were recorded on a Bruker AMX 400wb spectrometer (400.13 MHz), chemical shifts were calibrated using benzene (6 7.20 ppm). The Mn and Mw/Mnvalues of poly(s-caprolactone) were determined by gel permeation chromatography (GPC) on a Tosoh SC-8010 using TSK gel G2000, G3000, G4000 and G5000 columns in

154

Perspectives in Organometallic Chemistry

chloroform at 40°C. The Mn and MJMn values of polyethylene were determined by GPC on a Waters 150C instrument using Shodex AT-806MS column in 1,2,4-trichlorobenzene at 140°C. The M,, and M,IM, values were calibrated using standard polystyrene. Elemental analyses were performed on a PE 2400 series I1 CHNS/O analyzer. The samples were sealed in tin foils in an argon stream using a dry-box. EIMS spectra were recorded on a JEOL JMS-SX-102A spectrometer, and the samples which were sealed into glass capillaries in argon were opened just before analysis.

Synthesis of q5-(Me3Si-fluorenyl)2Sm(THF)2,1 To a THF solution (90 mL) of fluorene (25.0 g 151 mmol) held at 0°C was added dropwise butyllithium in hexane (1.61M, 93.6 mL, 151 mmol) with a dropping funnel. The reaction mixture was stirred at room temperature for 5 h. The resulting orange solution was added to a THF solution of trimethylchlorosilane (28.6 mL, 226 mmol) at 0°C and the mixture was stirred for 3 h at room temperature. The mixture was then poured into an aqueous solution saturated with NaHC03 (500 mL) and the orange layer was extracted with 3 portions of THFkexane (1:3) mixed solvent (30mL x 3) to giveMe3Si-fluorene in 89.6% yield (32.1 g). 'H NMR (400MHz, CDC13) 6 -0.07 (s, 9H, SiMe3), 3.87 ( s , lH, Cp-H), 7.30 (t, 2H x 2, fluorenyl-H2, 3,6, 7), 7.49, 7.85 (d, 2H x2, fluorenyl-H1, 4, 5, 8). To a 300mL round-bottomed flask equipped with a reflux condenser and a three way stopcock were placed potassium hydride (washed with hexane) (1.94 g 48.3 mmol) and Me3Sifluorene (12.7 g 48.3 mmol). After the addition of 75 mL of THF, the mixture was refluxed for 12 h to give Me3Si-fluorenylpotassium. To a 500 mL round-bottomed flask were placed Sm turnings (3.55 g, 23.6 mmol) and diiodoethane (6.64 g, 23.6 mmol). After the addition of THF (160 mL), the resulting deep blue solution was stirred for 3 h and Me3Sifluorenylpotassium in THF was added to the solution at ambient temperature. The stirring was continued overnight and the mixture was evaporated to dryness to give black oily product. After separation of the soluble part from black oily product by extracting with hexane, the insoluble part was recrystallized from THFkexane to yield 8.07 g (45%) of (Me3Si-fluorenyl)zSm(THF)2as black crystals. 'H NMR (400MHz, CDC13) 6 0.67 (bs, 8H, THF), 1.06 (bs, 18H, SiMe3), 7.80 (bs, 8H, THF), 8.06, 8.39 (m, 4H x 2, fluorenyLH2, 3,6, 7), 11.46, 11.72 (m, 4H x 2, fluorenyl-HI, 4, 5 , 8). Anal. Calcd for C40H5002SmSi2: C, 62.47; H, 6.55; Sm, 19.55. Found: C, 62.33; H, 6.78; Sm, 20.10 (oxidation method as Sm203). EIMS for lS2Sm:m/z (relative ratio), 627 (M-2THF, 39), 390 ( M - ~ T H F - C I ~ H ~ ~ S ~ , 100). Synthesis of q6 -(Me3Si-fIuorene-AlMe3)2Sm, 2 To a stirred solution of (Me3Si-fluorenyl)2Sm(THF)2(0.55 g 0.71 mmol) in toluene (75 mL) was slowly added excess AlMe3 (0.35 mL, 3.5 mmol). The color of the solution turned to dark red immediately after the mixing. After stirring for 12 h, the solution was evaporated to dryness and the residue was washed with excess hexane (30mL x 3) to remove AlMe3. After the separation of the hexane solution by centrifugation, the resulting solid was recrystallized from toluenekexane (1:3 vol ratio) to give (Me3Si-fluoreneAlMe3)2Sm, 2, as dark-red crystals in 57% yield (0.32 8). Anal. Calcd for C3gH~A12Si2Sm: C, 59.32; H, 6.81. Found: C, 59.05; H, 6.64. EIMS for ''*Sm: d z (relative ratio), 626 (M-2AlMe3, 14), 389 (M-2AlMe3-C13H&Me3,41), 238 (C13HgSiMe3,lOO). Synthesis of q6-(Me~Si-fluorene-AlEt~)2Sm, 3 To a stirred solution of (Me3Si-fluorenyl)$Sm(THF)2 (1.66 g 2.2 mmol) in toluene (75 mL) was slowly added excess AlEt3 (1.48 mL, 10.8 mmol). Immediately after mixing, the color

Novel

f75

- 776 rearrangement of bis(Juoreny1)lanthanidecomplexes

155

of the solution turned dark red. The stirring was continued for 12 h and the solution was evaporated to dryness. The residue was washed with excess hexane to remove AlEt3. After separation of the solid by centrifugation, the resulting solid was recrystallized from toluenekexane (1 :3) to afford (Me3Si-fluorene-AlEt3)3Sm, 3 as dark red crystals in 25.3 % yield (0.47 g). Anal. Calcd for C~H62A12Si2Sm:C, 62.06; H, 7.34. Found: C, 62.00; H, 7.41. EIMS for '52Sm: d z (relative ratio), 626(M-2AIEt3, 21), 389 (M-2AlEt3C13H&Me3,41), 238 (C13H89SiMe3, 100). Synthesis of q5-(Me3Si-fluorenyI)2Yb(THF)2,4 To a 300 mL round-bottomed flask equipped with a reflux condenser and a three way stopcock were placed potassium hydride (0.95g, 23.6 mmol) and Me3Si-fluorene (5.6 g, 23.6 mmol). After the addition of THF (75 mL), the mixture was refluxed for 12 h to give Me3Si-fluorenylpotassium. To a 500 mL round-bottomed flask were placed Yb turnings (2.05 & 11.8 mmol) and diiodoethane (3.3 g, 11.8 mmol). After the addition of THF (180 mL), the resulting light green solution was stirred for 12 h and Me3Si-fluorenylpotassium in THF was added at a stroke to this solution at ambient temperature. Stirring was continued overnight at ambient temperature and then the mixture was evaporated to dryness. The residual red oil was dissolved in toluene (90 mL) and the KI salt was removed by centrifugation. The toluene solution was evaporated to dryness and the resulting red oil was washed with hexane to give q5-(Me3Si-fluorenyl)2Yb(THF)2, 4, as a red oil in 91.1% yield (8.4 8). 'H NMR(400MHz) 6 0.65 (bs, 18H, SiMe3), 1.17 (bs, 8H, THF), 2.87 (bs, SH, THF), 6.87, 7.16 (m, 4H x 2, fluorenyl-H2, 3, 6, 7), 7.78, 7.83 (m, 4H x 2, fluorenyl-H1,4, 5, 8). EIMS for 174Yb:m/z (relative ratio), 648 (M-2THF, 25), 41 1 (M2THF-C13H&Me3,41). Synthesis of q6-(Me3Si-fluorene-A1Me3)- q5-(Me3Si-fluorenyl)Yb,5 To a stirred solution of (Me3Si-fluorenyl)2Yb(THF)2, 4, (4.5 g, 5.6 mmol) in toluene (75mL) was added excess AlMe3 (2.75 mL, 28.0 mmol). After stirring the mixture for 12h, the solution was evaporated to dryness and hexane (100mL) was added. The red hexane soluble part was allowed to stand at -25°C for 5 days without concentration of the solution and orange crystals were precipitated to provide q6-(Me3Si-fluorene-A1Me3)-q5-(Me3Sifluorenyl)Yb, 5, as red crystals in 32% yield (0.30 g). 'H NMR(400MHz) 6 -1.24 (bs, 9H, AlMe3), 0.15, 0.53 (bs, 9H x2, SiMe,), 6.84, 7.10,7.24 (m, 4H, 2H x2, fluorenyl-H2, 3, 6, 7), 7.45, 7.82, 7.95 (d, 2H x4, fluorenyl-H1, 4, 5 , 8). Anal. Calcd. for C35H43Si2AlYb: C, 58.39; H, 6.02. Found: C, 58.35; H, 6.04. EIMS for 174Yb:d z (relative ratio), 648 (M2AlMe3, 1S), 41 1 (M-2AlMe3-C13H&Me3,49), 238 (C13H&Me3,100). Synthesis of (fluorenyl)Yb(THF)4/A1Me4,6 To a stirred solution of q6-(Me3Si-fluorene-AlMe3)-q5-(Me3Sifluoreny1)Yb (0.10 g, 1.O mmol) was added excess THF (30 mL). After evaporation of the solution, the residue was washed with hexane (20 mL) and the resulting solid was recrystallized from hexane/THF to give (fluorenyI)Yb(THF)dAIMe4,6, in 15% yield. 'H NMR(400MHz) 6 -0.33 (bs, 12H, AlMe4), 1.29 (bs, 16H, THF), 3.28(bs, 16H, THF), 6.23 (bs, lH, Cp-H), 7.04-7.07 (t, 2H x 2, fluorenyLH2, 3,6, 7), 7.72, 8.08 (d, 2H x2, fluorenyl-H 1,4, 5, 8). Anal. Calcd for C33H5304AlYb: C, 53.51; H, 7.21. Found: C, 53.46; H, 7.22. Synthesis of ('Pr-indenyl)~Yb(THF)z,7 To a solution of indene (30.0 mL, 258 mmol) in THF (75 mL) held at 0°C was dropwise added butyllithium in hexane (2.52 M, 102 mL, 258 mmol) via a dropping funnel. The

156

Perspectives in Organometallic Chemistry

reaction mixture was stirred for 5 h at ambient temperature to give indenyllithium. The color of the solution turned to orange during the stirring To a 500 mL round-bottomed flask equipped with a three way stopcock was placed the THF solution of isopropyl bromide (28.8 mL, 309 mmol). The solution was cooled to 0°C with vigorous stirring and the THF solution of indenyllithium was added to this solution. The mixture was stirred for 3 h and then poured into an aqueous solution saturated with Na2C03 (500 mL). After the separation of organic layer, the aqueous layer was extracted with 3 portions of hexane (each 30 mL) and the combined organic layer was washed with water (30mL x3). Distillation of the solution (2 torr/55-58"C) afforded isopropylindene (33.9 g) in good yield (83.2%). 'H NMR (400MHz) 6 0.62, 1.13 (d, 3H x 2, CH-Mez), 2.30 (m, lH, CHMez), 3.42 (d, indenyl-Hl), 6.50, 6.83 (d, 1H x2, indenyl-H2, 3), 7.16,7.23 (t, 1H x2, indenyl-H5, 6), 7.34, 7.41(d, 1H x 2, indenyl-H4,7). To a 300 mL round-bottomed flask equipped with a reflux condenser and a three way stopcock were placed potassium hydride (0.95g, 23.4 mmol) in THF (75 mL). After the addition of iPr-indene (3.27 mL, 23.4 mmol), the mixture was refluxed for 10 h to give iPrindenylpotassium. To a 500 mL round-bottomed flask were placed Yb turnings (2.03g, 11.7 mmol), THF (180 mL) and then dibromoethane (1.01 mL, 11.7 mmol). The mixture was stirred for 12 h during which time the color of the solution turned green. The above mentioned iPr-indenylpotassium was added to the resulting YbBr2 at ambient temperature and the stirring was continued overnight. The solution was evaporated to dryness and the resulting red oily product was dissolved in 90 mL toluene. KI salt was separated by centrifugation and toluene was removed by flash distillation. The oily product changed to powder by the addition of hexane (40 mL). Crystallization of the solid from THFhexane (1 : 5 ) afforded ('Pr-indenyl)2Yb(THF)~(1S8g) in 2 1.4% yield, 'H NMR (400MHz, C6D6) 6 1.26 (bs, 8H, THF), 1.40, 1.44, 1.51, 1.67 (d, 3H x4, CHMe2), 3.1 1 (bs, 8H, THF), 3.37 (m, 2H, CHMe2), 5.93, 6.42, 6.84 (m, 1H x 12, indenyl-H). Anal. Calcd for C32H4202Yb: C, 60.86; H, 6.70; Yb, 27.40. Found: C, 60.78; H, 6.67; Yb, 27.33 (oxidation method as Yb203). EIMS for 174Yb:m/z (relative ratio), 488 (M-2THF, 18), 331 ( M - ~ T H F - C I ~ H ~ ~ , 36), 157 (C12H13, 100). X-ray Analyses of 1,2,3,5,6 and 7 All the diffkaction data were collected on a Rigaku AFC-5R diffractometer with graphitemonochromatized MoKa radiation (Table I). As the complexes are all air-sensitive, crystals were sealed in thin-walled glass capillary tubes under argon atmosphere. The Xray data were collected at room temperature using 01-28 scan techniques to a maximum 28 value of 55.0". The data were corrected for conventional absorption, Lorentz and polarization effects. The crystal structures were solved by the heavy-atom method and were expanded by successive Fourier syntheses. The non-hydrogen atoms were refined anisotropically by the full-matrix least-squares methods except for 3 and 5 (only metal atoms were refined anisotropically),while the hydrogen atoms were fixed at their standard geometries and were not refined. All the calculations were performed by the use of the texsan crystalographic software package (texsan: Crystal Structure Analysis Package, Molecular Structure Corporation, 1985 & 1992).

Novel

775 -

157

776 rearrangement of bisCfluoreny1)lanthanide complexes

Table I. Crystal Data for 1,2,3,5,6 and 7 1

formuta

2

3

5

C40H500SmSi2 C3sH52SmSi2A12 CUH&rnSi2Al2 C;5Hj;YbSi2Al

6

7

C;,Hs jOjYbAl

CjzHjzOzYb

fw

769.40

769.33

853.52

7 19.92

713.80

63 1.72

System

orthorhombic

tetragonal

monoclinic

monoclinic

monoclinic

monoclinic

Space group

p212121

P421c

Pz1/a

P2 ,/a

P2Ja

P2,la

aIA

16.951(3)

14.424(7)

15.427(3)

9.681(4)

16.94(7)

12.98(2)

blA

20.536(4)

1 1.63(1)

A

10.718(3)

18.74(1)

Bldeg

18.2l3(3)

14.322(4)

10.71 (6)

17.148(3)

24.713(5)

2 1.20(8)

19.06(I )

114.68(1)

94.32(3)

1 1 1.2(3)

92.45(7) 2874(4)

Vf A’

373 1( I )

3898(3)

4377(1)

34 I 6( I )

3584(26)

2

4

4

4

4

4

4

1.370

1.307

1.295

1.119

1.323

1.460

D d e

(3113’

F(OO0)

1584

1576

1776

1456

1464

1280

AMoK()/cm-’

16.73

16.39

14.67

28.54

26.64

32.78

no. of meads rtlns

4793

2624

4467

6963

8245

6980

no. of obsd rtlns

33 12

1802

2083“

1365’

5553

3465

R‘(R,)b

= 0.060(0.087), 0.033(0.046), 0.098(0.12), 0.080(0.077), 0.050(0.093),0.071(0.10)

’

‘R = L‘llFol-IF,IIILlFol.bRw= (~~IIFg(-IFcl)2/zi)((Fo~)”z ;o = l/~*(Fo).I > I.00(1), I > 1.50(1)

2.1 Reaction of AlMe3 with Bis(Me3Si-fluorenyl)Sm(THF)2,1 Reaction of b i s ( f l u o r e n y l ) S m ( T H F ) 2 with AlR3 should give rise to the formation of bis(fluorenyl)Sm(p - R ) 2 A l R 2 or b i s ( f l u o r e n y l ) S m . A l R 3 according to the literatwe.6-8 However, A l M e 3 did not react with bis(fluorenyl)Sm(THF)z because of its low solubility in toluene. Therefore, we improved its solubility by introducing the trimethylsilyl group into the fluorenyl group. Addition of the potassium salt of Me&-fluorene, prepared from potassium hydride and Me&fluorene, to S m I 2 ( T H F ) 2 generated b i s ( M e 3 S i f l u o r e n y l ) S m ( T H F ) 2 , 1 (Fig 5).

SiMe3

Figure 5 Bis(MejSi-fluorenyl)Sm(THF)2 1

Figure 6 X-ray analysis of 1

Perspectives in Organometallic Chemistry

158

The 'H NMR spectrum reveals the formation of the desired bis(q5-Me3Sifluorenyl)Sm(THF)z(Fig 7) and the molecular structure was determined by X-ray analysis. Fig 6 shows the ORTEP drawing of 1. The five membered ring in the fluorenyl group coordinates to the Sm metal and the Sm-Cp(centroid) distances (2.66-2.77A) are nearly equal to the 2.633A of (CsMe&Srn(THF)? and 2.629A of (fluoreny1)2Sm(THF)2.'O As a whole this complex assumes a coordination geometry of C2-symmetry, while C2,-symmetry was reported for bis(fluorenyl);!Sm(THF)2. The dihedral angle between the two Cp planes is ca. 45.3" (Table 11).

SiMes

benzene

'i'

THF

Figure 7 IHNMR

spectrum of1 in ~

6

~

6

The reaction of an excess amount of AlMe3 (5 equivalents) with complex 1 gave

bis(Me3Si-fluorene-AlMe3)Srn, 2 (Fig 8) as revealed by the 'H NMR spectrum (Fig 9).

I

15.0

Figure 8 Reactions of AIR3 with his($fluorene)$'m (THO2

10.0

5.0

0.0

Figure 9 ' H NMR spectrum of 2 in C&

m

Novel

q5

- 776 rearrangement of bisCfluoreny1)lanthanidecomplexes

159

Complex 2 promptly decomposes in a moist-air. Its exact structure was determined by Xray analysis using the sample sealed in a thin glass capillary (Fig 10). The most important finding of this complex lies in the q6-coordination of the Sm atom towards the phenyl group, while in the initial complex the q5-coordinationof the Sm atom to the Cp group was observed. The distances of Sm-Cp(centroid) and Sm-Ph(centroid) are 3.59 and 2.7481, respectively. The complex assumes a chiral racemic structure, rather than a meso structure as a result of steric repulsion (Fig 12). The AlMe3 molecule assumes o-bonding with C(l) atom of the 5 membered ring and an agostic interaction was observed between the Sm metal and one of the Me groups of AIMe3. The Sm-Me bond length is 2.86A. The AlMe3 molecule is located near the Sm atom, while the Me3Si group is far from the Sm atom. The C(l)-C(2) and C(l)-C(13) distances are, by 0.09A and 0.0381, longer than those of complex 1. The dihedral angle of the two Cp planes is ca 55.4", 10.1" larger than that of complex 1.

Figure 10 X-ray structure of 2

I

.

Figure 11 X-ray structure of 3

1 -

racemic Figure 12 Geometry of q6-bis(Me3Si-jluorene-A1Rj)Ln

meso

160

Perspectives in Organometallic Chemistry

2.2 Reaction of AlEt3with Bis(Me3Si-fluorenyl)Sm(THF)2,1 The reaction of excess AlEt3 (5 equivalents) with complex 1 affords complex 3 in good yield as revealed by the 'H NMR spectrum. Its ORTEP drawing is shown in Fig 11. The coordination geometry of 3 resembled that of 2.It assumed a racemic structure. One of the Et groups exhibits an agostic interaction with the Sm via its CH2 group. The Sm-CH2 distance (2.92A) is a little longer than that (2.87A) of 2,reflecting the bulkier AlR3 group substituent (Table 11). The dihedral angle of the two Cp rings is 65.7', much larger than that of complex 2.The distances of Sm-Cp(centroid) and Sm-Ph(centroid), 3.56 and 2.76A respectively, are also consistent with those of 2. When complexes 2 and 3 were dissolved in THF, the initial complex 1 was formed in a quantitaive yield immediately after mixing with THF. Thus, alternation between the q5-and q6-bonding mode was observed by the reaction of A1R3 and the successive addition of a donor molecule.

Table II.Selected Bond Distances(A) and Angles (deg) for Complexes 1,2,3,5 and 7with Estimated Standard Deviations (parentheses) 2

1

7

5

3

C( 1)-C(2)

1.37(2)

1.46(1)

lSO(5)

1.52(5)

1.40(3)

C(2)-C(7)

1.43(3)

1.42(1)

1.45(4)

1.38(5)

1.46(3)

C(7)-C(8) C(8)-C(13)

1.46(3)

1.43(1)

1SO(4)

1.41(5)

1.40(3)

1.38(3)

1.41(1)

1.36(6)

1.41(5)

1.38(4)

C( 1)-C(13)

1.45(2)

1.48(1)

1.48(4)

1.48(4)

1.42(3)

C(2 1)-C(22) or C( 1*)-C(2*)

1.43(3)

1.46(1)

1.45(4)

1.43(5)

1.41(3)

C(22)-C(27) or C(2*)-C(7*)

1.41(2)

1.42(1)

1.37(4)

1.46(5 )

1.47(3)

C(27)-C(28) or C(7*)-C(8*)

1.49(2)

1.43(1)

1.47(4)

1.43(5)

1.43(3)

C(28)-C(33) or C(8*>C( 13*) 1.45(3)

1.41(1)

1.42(4)

1.44(5)

1.34(3)

C(21)-C(33) or C(l*)-C(13*)

1.46(3)

1.48(1)

1SO(4)

1.48(5)

1.47(3)

M-O(1)

2.58(1)

2.47( 1)

M-O(2)

2.51(1)

2.77(2)

M-Cp(l)(centroid)

2.70

3.59

3.56

3.35

2.50

M-Cp(2)(centroid)

2.67

3.59

3.65

2.41

2.48

M-Ph(l)(centroid)

3.43

2.74

2.75

4.94

M-Ph(2)(centroid)

3.41

5.08

5.10

2.53

M-Ph(3)(centroid)

3.66

2.74

2.77

3.34

M-Ph(4)(centroid)

3.58

5.08

5.15

3.30

118.6

113.5

168.1

Cp(centroid)-M-Cp(centroid) 142.4 O( 1)-M-0(2)

86.6

CP(l)-CP(2)

45.3

123.5 120.5

55.4

65.7

40.2

120.7

Novel

175 -

776 rearrangement of bis(jluoreny1)lanthanide complexes

161

2.3 Reaction of AlPr3, Al'BU3 or BEt3 with bis(Me3Si-fluoreny1)samarium The reaction of AlPr3 or Al'Bu3 with complex 1 was carried out to find a new type of complexation due to its large steric bulkiness. The reaction proceeds smoothly to produce complexes that are very soluble even in hexane at low temperatures. Therefore we could not identifjr the mode of complexation in detail. However, 1) the initial complex 1 is insoluble in hexane, but the resulting complex is freely soluble in hexane, 2) the 'H NMR spectra of the adducts are nearly identical with those of complexes 2 and 3, regarding the absorptions for Me3Si-fluorene group (10.0-15.0 ppm). Therefore we can readily deduce the structure to be as shown in Fig 13. In the same manner, BEt3 reacted with complex 1 to produce an identical complex as determined by 'H NMR, although we could not succeed in isolating the respective complexes (Fig. 13).

Figure 13 Mode of addition of A1R3 and BEt3 to bis($-Me3Si$uorenyl)Sm, 1 2.4 Reaction of AIMe3 with Bis(Me3Si-fluoreny1)ytterbium TO understand the unique reactivity of the corresponding bis(Me3Si-fluorenyl)Yb(THF)2, bearing a metal of small ionic radius compared to that of Sm, toward AlR3, we have prepared bis(Me&-fluorenyl)Yb(THF)2, 4, starting from (Me&-fluoreny1)K and YbI2. Complex 4 was obtained as red crystals and its 'H NMR spectrum (Fig 16) reveals the q5coordination of the fluorenyl group. The reaction of excess AlMe3 with 4 gave a mixture of hexane soluble complex 5 and hexane insoluble complex 6 in a 9: 1 ratio (Fig 14).

4

5

Figure 14 Mode of reaction of AIR3 with bis($-Me&-jluorenyl) Yb(THF)z 4 The 'H NMR spectrum of the hexane soluble complex indicates the formation of complex 5 which consists of bis(Me3Si-fluorenyl)Yb/AlMe3 in a 1:l and not a 1:2 ratio (Fig 17). The molecular structure of 5 was finally determined by X-ray crystallography and the

162

Perspectives in Organometallic Chemistry

resulting ORTEP drawing is shown in Fig 15. The complex 5 exhibits an unsymmetrical structure, where AlMe3 bind to the Me3Si-fluorenyl group at its C(l) position, while the other Me3Si-fluorenyl group is free from coordination by AlMe3 (it keeps the q5coordination). The Yb atom is tetra-coordinated and we cannot observe any coordination of THF in this molecule. An agostic interaction exists between Yb-C(17) and Yb-C(19) whose bond distances are 2.80 and 2.70A, respectively, a little shorter than those of 2 and 3, reflecting the small diameter of the Yb atom as compared with Sm. The geometry of the present coordination should originate from the preferential formation of the agostic interaction between Yb-C( 17) and Yb-C( 19), which prevents the further coordination of AlMe3 to another Me3Si-fluorenyl group. The dihedral angle of the two Cp planes is 40.9", the smallest angle among 1,2,3 and 5.

Figure 15 X-ray structure of q6-(Me&jluorene-AlMeJ)-q 5 - ( M e & j l u o r e n yYb ~ 5

Figure 16 'HNMR of 4

Figure 17 'HNMR of 5

We can readily estimate the structure of complex 6 based on the 'H NMR spectrum. The Me3Si group is absent in this complex and the signal of AlMe3 or AlMes is observed. The X-ray analysis of 6 reveals the presence of only one fluorenyl group, 4 coordinated

Novel

775

-

163

rearrangement of bis(Jluoreny1)lanthanide complexes

THF molecules, and one AlMe4- group (Fig 18). This complex is also obtained in a low yield by the addition of excess THF to 5. The reaction pathway for the formation of 6 is unclear at present. However, we could obtain (indenyl-d)Yb(THF)dAlMe4with a d-labeled indenyl group at the C(l) position, when we used THF-dg in place of THF. Therefore, the Me& group should be liberated from the indenyl group by the attack of the THF-dg molecule.

Figure 18 ORTEP drawing of 6

2.5 Reaction of Bis('Pr-indenyl)2Yb(THF)2 with AIR3 To understand the role of the Me$i group bonded with the fluorenyl group and the role of the fluorenyl ring, we have explored the use of bis('Pr-indenyl)Yb(THF)2, 7, and examined the reaction with excess AlMe,. The 'H NMR spectrum of 7 indicates that all the indenyl protons appear at different positions and the signals of the 'Pr groups are split into 4 peaks to indicate that this complex exhibits an unsymmetrical structure due to the restriction of free rotation around the Yb-Cp(centroid) axis. The final molecular structure of 7 was determined by X-ray analysis and Fig. 20 shows its ORTEP drawing The molecular structure of the present complex resembles that of 1. The addition of excess AlMe3 to 7 produced a toluene insoluble compound 8 in quantitative yield, which readily affords initial 7 by the addition of excess THF. Although the molecular structure is unknown due to its low solubility to toluene, we can readily deduce the structure 8a or 8b based on the reactions similar to 1 (Fig. 19). ,iPr

iPr

AIR3

8a

Figure 19 Mode of reaction of AIR3 with $-bis(Pr-indenyl) Yb(THF)2 7

This result suggests that the q5-q6rearrangement can occur even when we use the indenyl ring bearing an 'Pr group. Thus, the addition of AIR3 to bis(q5-fluorenyl)Ln(THF)2or bis(q5-indenyl)Ln(THF)2brings about the formation of bis(q6-fluorene-A1R3)Lnor bis[q6indene-(AlR&,]L,, (n = 1or 2).

Perspectives in OrganometallicChemistry

164

Figure 20 X-ray structure of $-bis(Pr-indenyl) Yb(THq2 7

2.6 Catalytic activities of Complexes 1,2 and 3 for Polymerization of Ethylene and Ecaprolactone The Kaminsky" and Brookhart12 catalysts which are known to be effective homogeneous catalysts for the polymerization of ethylene and 1-olefins generally require the presence of cocatalysts such as methylaluminoxane (MAO) or modified methylaluminoxane (MMAO). In sharp contrast to these catalyst systems, rare earth metal complexes exhibit high catalytic activity towards the polymerization of ethylene13 and polar monomer^'^ in the absence of any cocatalyst. We have examined here the catalytic activity of 1, 2, and 3 for the polymerization of ethylene and some polar monomers. The result of the polymerization of ethylene is summarized in Table 111. Complex 1 shows very low activity, while complex 3 has relatively high catalytic activity towards the polymerization of ethylene. Every catalyst provides polyethylene whose molecular weight exceeds 50,000 with rather narrow polydispersity. However, their catalytic activities are lower than those of racemic Me2Si(2SiMe3-4-'Bu-CsH&Sm(THF) and meso Me2Si(SiMe2OSiMe2)(C5H2-3-tB~)2Sm(THF)2.'~ Table 111. Catalytic Activities of 1,2,and 3 for Polymerization of Ethylene Complex Polym.time/min Polym.temp./"C Activity(g/mol.h.atm)

Mn/lO4

Mw/Mn

1

60

25

32

25800

2.03

2

20

25

125

43600

2.56

3

60

25

7556

53200

1.98

20 65 1719 45700 Polymerization conditions: solvent toluene. Ethylene was added at 1 atm.

2.23

The complexes 1-3 also showed good catalytic activity toward the polymerization of c-caprolactone (Table IV). The conversion is quantitative and the molecular weight exceeds 50,000 with a rather narrow polydispersivity. However, their catalytic activities are lower than those of racemic Me2Si(2-Me3Si4-'Bu-CsHz)Sm and meso Me2Si(SiMe20SiMez)(CsH2-3-'Bu)2Sm( THF)2.

Novel

~5

- 776 rearrangement of bis(Jluoreny1)lanthanidecomplexes

165

Table IV. Catalytic Activities of 1,2, and 3 for Polymerization of E-Caprolactone

Complex Polym.time/h

Polym.ternp./”C

[M]o/[I]o

Yield/%

M,/l O4

MwIM,,

1

5

25

176

97

6.43

1.20

2

5

25

150

99

5.45

1.18

3

5

25

176

100

7.28

1.17

Polymerization conditions: solvent toluene. [solvent]/[monomer] = 2.0 (vol/vol), [M]o/[I]o initial ratio of monomer to the initiator (mol/mol). 3 CONCLUSION

q5-Bis(Me3Si-fluorenyl)Sm(THF)2 1 was prepared by reaction of Me&fluorenylpotassium with SmI2(THF)2. 1 gave q6-bis(Me3Si-fluorene-A1Me3)Sm2 upon reaction with excess AlMe3. The corresponding reaction of excess AlEt3 with 1 gave q6-bis(Me3Si-fluorene-A1Et3)Sm3. The reaction of bis(Me3Si-fluorenyl)Yb(THF)~4 with an excess amount of AlMe3 gave q5-(Me3Si-fluorenyl)-q6-(Me3Si-fluorene-A1Me3)Yb 5, where one equimolar AlMe3 coordinated to the Yb atom through its two Me groups via agostic interaction. One of the Me3Si-fluorenyl groups assumes an $-coordination while the other Me3Si-fluorenyl group shows q6-coordination. The addition of excess THF to 5 produced (fluorenyl)Yb(THF)dAlMe46 in low yield. \the catalyses of 1, 2, and 3 for polymerizations of ethylene and E-caprolactone were examined. Acknowledgement This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas (NO. 283, “Innovative Synthetic Reactions”) from Monbusho. References

P. M. Treichel and J. W. Johnson, Inorg. Chem. 1977, 16,749 J. W. Johnson and P. M. Treichel, J. Chem. SOC.Chem. Commun., 1976,688. a) N. A. Ustynyuk, Yu. F. Oprunenko, S. G. Malyugina, 0. I. Trifonova and Yu. A. Ustynyuk, J. Organornet. Chern., 1984,270,185. b) Yu. F. Oprunenko, Yu. N. Lizikov, Yu. A. Ustynyuk and N. A. Ustynyuk, J. Organornet. Chem., 1982,231, 137. c) A. Ceccon, A. Gambaro, G. Agostini and A. Venzo, J. Organornet. Chem., 1981, 217,79 a) N. A. Ustynyuk, L. N. Novikova, Yu. F. Oprunenko, S. G. Malyugina and Yu. A. Ustynyuk, J Organomet. Chem., 1985,294,31. b) N. A. Ustynyuk, Yu. F. Oprunenko, S. G. Malyugina, 0. I. Trifonova and Yu. A. Ustynyuk and, J. Organornet. Chem., 1984,270,185. G. L. Kubas, G. Kiss and C. D. Hoff, Organomefallics, 1991,10,2870. a) M. A. Busch, R. Harlow and P. L. Watson, Inorg. Chim. Acta, 1987,140, 15. b) W. J. Evans, L. R. Chamberlain, T. A. Ulibarri and J. W. Ziller, J. Am. Chem. SOC.,1988, 110,6423. H. Yamamoto, H.Yasuda, K. Yokota, A. Nakamura, Y. Kai and N. Kasai, Chem. Left., 1988,1963. J. Holton, M. F. Lappert, D. G. H. Ballad, R. Pearce, J. L. Atwood and W. H. Hunter, J. Chem. SOC.Dalton Trans., i1979, 54.

166

9 10 11 12 13

14

Perspectives in Organometallic Chemistry

W. J. Evans, J. W. Grate, H. W. Choi, I. Bloom, W. E. Hunter and J. L. Atwood, J. Am. Chem. SOC.,1985,107,941. W. J. Evans, T. S. Gummersheimer, T. J. Boyle and J. W. Ziller, Organometallics, 1994,13,1281. a) H. Sinn, W. Kaminsky, H. J. Jollmer and R. Woldt, Angew. Chem. Int. E d Engl., 1980,19,390. b) W. Kaminsky, M. Miri, H. Sinn and R. Woldt, Makromol. Chem. Rapid Commun., 1983,4417. a) L. K. Johnson, S. Mecking and M. Brookhart, J. Am. Chem. SOC.,1996,118,267. b) L. K. Johnson, C. M. Killian and M. Brookhart, J. Am. Chem. SOC.,1995,117, 6414. c) B. L. Small andM. Brookhart,J. Am. Chem. SOC.,1998,120,7134. a) H. Yasuda and E. Ihara, Adv. Polym. Sci., 1997,133,53. b) G Leske, L. E. Schock, P. N. Swepston, H. Schumann and T. J. Marks, J. Am. Chem. SOC.,1985,107,8103. c) E. Ihara, M. Nodono, K. Katsura, Y. Adachi, H. Yasuda, M. Yamagashira, N. Kanehisa and Y. Kai, Organometallics, 1998, 17,3945. a) H. Yasuda, H. Yamamoto, K. Yokota, S. Miyake and A. Nakamura, J. Am. Chem. SOC.,1992,114,4908. b) E. Ihara, M. Morimoto and H. Yasuda, Macromolecules, 1995,28,7886. c) M. Yamashita, Y. Takemoto, E. Ihara and H. Yasuda, Macromolecules, 1996,29, 1798. d) H. Yasuda, M. Furo, H. Yamamoto, A. Nakamura, S. Miyake and N. Kibino, Macromolecules, 1992,25, 51 15.

RESULTS AND PERSPECTIVES OF HIGH OXIDATION STATE ORGANOMOLYBDENUM CHEMISTRY IN WATER

’,*

Edmond Collange,’ Funda Demirhan, Jenny Gun: Ovadia Lev: Alexandre Modestov? Rinaldo Poli,*’ Philippe Richard,’ Dirk Saurenz’ 1

Laboratoire de Synthese et d’Electrosynthese Organometalliques,Universite de Bourgogne, Faculte de Sciences (( Gabriel D, 6 boulevard Gabriel, 2 1000 Dijon, France 2 Celal Bayar University, Faculty of Sciences & Liberal Arts, Department of Chemistry, 45030, Muradiye-Manisa, Turkey. 3 Div. of Environmental Sciences, Fredy and Nadine Hermann School of Applied Science, The Hebrew University of Jerusalem, Jerusalem, Israel

1 INTRODUCTION

Organometallic chemistry has traditionally been developed in non aqueous media, especially when involving odd-electron carbon ligands such as alkyl or aryls (1-electron), allyls (3electron), cyclopentadienyl (5-electron), and so forth, because the negatively polarized carbyl ligands are generally susceptible to hydrolytic attack. This rule has many exceptions, however. In addition, the investigation of organometallic compounds in aqueous media has recently attracted great interest for a variety of reasons. From the practical oint of view, the use of water as a solvent is attractive for homogeneous catalytic applications?-’ Separation and recycling of water-soluble transition metal catalysts is easier under aqueous biphasic conditions and the polarity effect of water may result in usehl modifications of activities and selectivities. In addition, water is better than most solvents for heat exchange purposes. In many cases, the knowledge acquired for catalytic processes operating under anhydrous conditions can be extrapolated to the aqueous medium by the simple decoration of the metal coordination sphere with hydrophilic subsituents such as sulfonates, hydroxo, amine or ammonium groups, and so forth, on phosphine or cyclopentadienyl ligands.’. Water is also more attractive fiom the economic and environmental points of view, being a readily available, inexpensive and non toxic liquid. Furthermore, water is the major component of physiological fluids, justifying a growing research activity in the “bioorganometallic” chemistry area? Fundamental interest in aqueous organometallic chemistry includes, among other things, the study of the fr ility of metal-carbon bonds towards protonolysis, the involvement of water as a ligand and metal-catalyzed transformations using water as a reagent.

7

Perspectives in Organometallic Chemistry

168

Most aqueous investigations carried out so far involve low to middle-valent transition metal complexes. Organometallic chemistry in the highest oxidation states has been intensively developed in the last 20 years but most of this research work has been confined to non aqueous media. The greater metal electronegativity in the higher oxidation states confers a greater degree of covalency to the resulting metal-carbon bonds, which consequently may become quite resistant to hydrolytic conditions. It is therefore somewhat surprising that the physical behavior and chemical reactivity of high oxidation state organometallics is not systematically investigated in water, although aqueous reagents are sometimes used for their syntheses. Indeed, high oxidation state complexes are often stabilized by the highly ndonating and electronegative 0x0 ligand, which is nothing more than a doubly deprotonated water molecule. The relationship between the aqua, hydroxo, and 0x0 ligands is shown in Scheme 1 and it is obvious that an increase of the metal oxidation state increases the acidity of the oxygen-bound protons, stabilizing the 0x0 form. Aqua-complexes will be more favored, on the other hand, in the lower oxidation states. -H+

-H+

~

M-OH2

+H+

-

0

M=O ~ s

+H+

Scheme 1

The systematic investigation of high oxidation state organometallic complexes in water can open new perspectives for aqueous catalysis and, when a highly redox-active metal is selected, also for electrocatalysis. For instance, one can envisage the combined reduction and protonation of 0x0 complexes to yield labile aqua ligands, making coordination sites available for substrate coordination. The activated substrate may then be capable of accepting electrons from the metal and protons from the medium, yielding a hydrogenated product by use of proton and electrons (instead of molecular hydrogen), while regenerating the high oxidation state 0x0 complex. One of the oldest high oxidation state organometallic complexes is Cp2Mo205, first reported by Malcolm Green in 1964.7 The Cp* analogue was first described b Herberhold in 19858 and structurally characterized later in several different polymorphs, 10, 1 1 always showing a symmetric Mo-0-Mo bridge. A related anionic 0x0 complex of Cp*Mo(VI), [Cp*MoO$, also exist.12' l3 Some aspects of the synthetic procedures leading to these complexes involve water as a solvent and/or as a reagent. For instance, CpMoCl4 and Cp*MoC14 are hydrolyzed in air-free water leading to [CpMoO2], and Cp*MoOCL, respectively. Subsequent aerial oxidation and/or basic hydrolysis leads to the dinuclear Mo(V1) products. These products, however, have not been systematically investigated in water. No knowledge was available on the stability of the CpMo or Cp*Mo bond toward hydrolysis, nor on the nature of the Moo3 moiety as a function of pH. The redox behavior of the molecules in an aqueous environment (as well as in non-aqueous media) was also unknown. Because of the availability of these materials and of the expected rich redox activity of the molybdenum atom, we have selected these systems, in particular the Cp*containing compound, for our initial studies in this area. '

9

Results and perspectives of high oxidation state organomolybdenum chemistry

169

2 METHODS AND RESULTS 2.1 Improved syntheses of (Rin&MozOs Compounds

The literature syntheses of the Cp and Cp* 0x0 derivatives of Mo(V1) suffered from drawbacks. The best synthesis of Cp2MozOs involves aqueous hydrolysis of CpMoCl4, giving orange [CpMoO2]2 via an isomeric red, apparently tetranuclear intermediate. The orange dinuclear Mo(V) species is oxidized to the desired product, but only slowly (> 1 day) by the unusual Ag20 reagent in refluxing cH~C12.l~Under the assumption that this sluggish reactivity is caused by the strong metal-metal bond (whose presence is shown" by an X-ray structural study), and that the red and unstable tetranuclear isomer does contain such bond, we have isolated the red isomer and tested its susceptibility to oxidation. As predicted, this compound is oxidized rapidly at room temperature to Cp2Mo205 by the readily available oxygen transfer agent PhIO. The overall synthesis can now be accomplished in high yield from CpMoCl4 in one-half day.16 Concerning the Cp* analogue, a similar basic hydrolysis route starting fiom Cp*MoC14 in acetone-water (carried out in an open flask) gives Cp*M002Cl,'~which had also been obtained earlier by aerial oxidation of [Cp*Mo(COh]2 in chloroform? Partially contradicting reports indicated that Cp*M002Cl is transformed to Cp*2M0205 upon treatment with excess NaOH,13 whereas the hydrolysis of Cp*MoC4 in the presence of the weaker base Bu'NH2 leads to [Cp*M003]-.'~ The relationship between the dinuclear neutral compound and the mononuclear trioxo anion is now fully understood (see following section). At any rate, we have developed an improved rational synthesis fiom Cp*MoC14 under the assumption that Cp*2M0205 is quantitatively transformed to the [Cp*MoO3]- ion by OH- (see Scheme 2). Aerial basic hydrolysis of Cp*MoCL with aqueous NaOH (6 equivalents) yields water-soluble Na[Cp*Mo03]. After filtration, Cp*2Mo205 is separated from the NaCl by-product and recovered as a precipitate from the aqueous phase by acidification with glacial acetic acid.l6

(Ring)Mo02Cl

Scheme 2 The understanding of the chemistry shown in Scheme 2 has allowed a fine tuning of the synthetic strategy. The oxodichloro Mo(V) intermediate undergoes competitive hydrolysis (faster for the Cp system, even under neutral conditions) and oxidation (faster for the Cp* system, but only when no excess of NaOH is present). Thus, the initial use of just three base equivalents minimizes the formation of the [Cp*Mo02]2 by-product (which is resistant to aerial oxidation). The procedure can be stopped at this point to recover Cp*Mo02C1 or additional hydrolysis leads to Cp*2Mo205 after the acetic acid treatment. Formation of the

170

Perspectives in Organometallic Chemistry

Mo(V) product, on the other hand, can be maximized by use of an inert atmosphere and 4 base equivalents. This synthetic strategy has also been extended to systems containing bulkier cyclopentadienyl rings, namely C5H2But3 and CsHPr'4, whose structure shows a linear and symmetrical Mo-0-Mo bridge like the analogous Cp* compound.l6

2.2 Nature of Cp*MoV1in water at pH 0-14 The above described synthetic work indicates that the (Ring)MoV1moiety resists hydrolytic splitting at least down to pH 4 and that the dinuclear (Ring)2Mo205 converts reversibly to [(Ring)Mo03]-upon increasing the pH. A number of interesting questions that we have asked ourselves are: (i) down to what pH is the (Ring)-Mo bond chemically inert? (ii) What is the mechanism of transformation of the mononuclear anion to the dinuclear neutral compound (is this occurring via a yet unobserved dioxo-hydroxo complex, as proposed earlierI3)?(iii) What is the pK, of such hydroxo complex? (iv) Do other (Ring)Mo"' species exist under any given pH conditions? These questions were initially addressed by a kinetic approach, using stopped-flow mixing techniques and UV-visible detection.' For reasons of both UV-visible absorption intensity and solubility, the study was restricted to the Cp* compound. In fact, even for this compound a small amount (20% v/v) of methanol has to be used in order to avoid the precipitation of Cp*2Mo205. This, however, has no effect on the solution chemistry. The kinetic approach was based on the assumption, then proven valid, that the yet unobserved dioxo-hydroxo complex could be formed from the anion by protonation and then rapidly decompose. The results, however, have turned out different and more interesting than expected. To cut a long story short, [Cp*Mo03]- is stable at high pH (> ca. 6) and its protonation affords Cp*Mo02(0H) in an immediate reaction (relative to the stopped-flow dead time, ca. 1 ms). The latter species, however, evolves at low pH by proton addition. A new mononuclear species is obtained quantitatively at pH < 2, while an equilibrium is established between pH 2 and 6. The mononuclear nature of the new species is shown by the kinetics being strictly first 1.2

1 09

A

1

08

07

0.0

.-ti 2

0.6

0.4

06

05 04

03 02

0.2

0.1

0

0

325

1

375

425

475

2

3

4

5

6

7

PH

A/ nm

Figure 1 UV-visiblespectra ofpure c p * ~Figure o 2 Distribution Of cp*hdov' species in 20% methanol-water as a function ofpH. 0x0 species: [Cp *MoOj]-at pH > 6; [Cp*MoO2]' at pH 1; Cp*M002(0H) upon protonation of the anion at pH 1, after 1 ms.

171

Results and perspectives of high oxidation state organomolybdenum chemistry

-1

order in metal in the entire pH range. The lOgk+Z= -3.30 7CP*M003H2+] interpretation of this species as a cationic dioxo complex, [Cp*Mo02]+, follows in a Cp*Mo02+ 10gk-2/KB1= +1.24 PKal < 0 straightforward manner. The spectra of the lOgk+l = +3.51 key species are shown in Figure 1 and their logk.l = -6.20 Cp*Mo02(0H) equilibrium distribution as a fbnction of pH (&OH-) pKa2 = 3.65 is shown in Figure 2. The speciation results from the complete determination of rate and Cp*MoO$ equilibrium constants as shown in Scheme 3. Scheme3 The exact nature of the [Cp*Mo02]+ species deserves careful consideration. The protonation of [Cp*Mo03]- immediately affords, as stated above, Cp*Mo02(0H), followed by conversion to the new mononuclear species, this being first order in H'.The latter species is unlikely to be a dihydroxo complex, because the consecutive pK, values of E(OH)2 oxoacids (E = any element, e.g. S in H2S04) usually differ by at least 4 units and pK,1 should therefore be lower than 0, see Scheme 3. However, this dihydroxo species could be formed in a pre-equilibrium step followed by a slow, irreversible transfonnation which could be either an intramolecular proton transfer to yield a dioxo-aqua species (hypothesis a in Scheme 4), or water dissociation to yield an unsolvated dioxo complex (hypothesis b). Either way, the transformation is remarkable because a tautomerization exchanging a proton between two oxygen sites is not expected to be such a slow process, whereas an unsolvated dioxo species is electronically unsaturated (though 5coordinated dioxomolybdenum(V1) complexes are precedented!). * * Preliminary DFT studies show that water coordination to Cp*MoO> is highly exothermic and experimental studies aimed at fully elucidating the nature of this complex are ongoing. We shall keep using the Cp*MoO> formulation for this species, though it should be kept in mind that an additional water molecule may be coordinated to it.

[

1

-

hypothesis a

hypothesis b

I

o&"yo-

o'/"\o

Scheme 4

The presence of a cationic Cp*MoV1complex in solution is confirmed by additional physical studies, as follows. The implication of Figure 2 is that Cp*2Mo205 should self-ionize in an aqueous environment to yield Cp*MoO2+ and Cp*MoO3- (plus a small equilibrium amount of Cp*Mo02(0H)). Indeed, whereas Cp*2Mo205 does not conduct electricity when dissolved in all common organic solvents (including methanol), the corresponding solutions in 20% methanol-water are good conductors. Furthermore, a certain cation fraction (which is a function of the concentration) should hydrolyze to yield the anion plus protons, lowering the pH (this is best appreciated by inspecting Figure 2, see also Figure 3(a)). Indeed, pH

172

Perspectives in Organometallic Chemistry

determinations for solutions of various concentrations afford values that agree with those predicted on the basis of Scheme 3, see Figure 4. Detailed concentration dependent conductivity studies yield results that are in perfect agreement with the model of Scheme 3, see Figure 5. There are, therefore, two independent verifications of the consistency of Scheme 3. 41

0

31

A

1 2 3 [Cp*,Mo,Od (M)xlO4

/

1

0

4

2

3

4

[Cp*Mo,CI] (M)xW

Figure 3 Concentration of ionic species derivingfiom the dissolution of Cp*2Mo205 (a) or Cp *MoO2Cl (b) in 20% methanol-water. 5 1

Cp*Mo02Cl

A J

3 4 0

1

1

3 c (nqx.104

2

4

5

Figure 4 pH of solutions of compounds Cp *2M02O5 (squares) and Cp *M002C1 (triangles) in 20% MeOH-H20. The curves are calculated on the basis of Scheme 3.

0

1

2

3

4

c (M)xlW

Figure 5 Conductivity of solutions of compounds Cp *2M02O5 (diamonds) and Cp*MoO2Cl (triangles) in 20% MeOHH20. The curves are calculated on the basis of Scheme 3.

Like compound Cp*2Mo205, the well knowng complex Cp*MoO2C1 had apparently not been investigated before in an aqueous medium. We have found that this compound also behaves as a strong electrolyte in 20% MeOH-water. The Cp*Mo02f ion hydrolyzes as discussed above and the measured pH and electrical conductivity at various concentrations, once again, correspond to those predicted by Scheme 3, see Figures 3(b), 4 and 5 . The speciation which is illustrated in Figure 2 is, of course, valid only in the given solvent mixture. The electrical conductivity study shows a steady decrease of conductivity as the percent methanol is increased, until a negligible conductivity is obtained in pure methanol. We have evidence, in fact, for the presence of higher nuclearity species in methanol-richer mixtures (vide inza).

Results and perspectives of high oxidation state organomolybdenurn chemistry

173

2.3 Coupled electrochemistry- electrospray ionization mass spectrometry studies As stated in the introduction, molybdenum is a highly redox active metal and a long term goal of this research is to develop potentially interesting electrocatalytic processes based on watersoluble organomolybdenum species. Therefore, it is interesting to examine the electrochemical activity and the nature of the new products that might result from the reduction of Cp*Mov* in water. We have found that on line electrospray ionisation (ESI) mass spectrometry coupled to the electrochemical cell is increadibly powerful for the analysis of the complex mixture of products which is obtained from the reduction of Cp*2M0205.~~ First of all, we have carried out standard electrochemical studies (polarography, cyclic voltammetry) and found that these can yield only very limited information because several irreversible processes overlap in the same potential region. Carrying out the electrochemical process in a flow-through cell2' with direct injection of the electrolyzed feed into the ESI mass spectrometer, however, has permitted the determination of all individual products from their characteristic isotopic distribution. In addition, the experiment has provided qualitative indication of the potential at which each species starts to form. It must be pointed out that, amongst the various possible mass spectrometric ionisation methods, electrospray is the least likely one to yield fragmentation phenomena and the most likely one to detect the solution-borne species. The various spectrometer conditions (cone voltage, capillary temperature, nature of solvent and buffer, etc.) have been carefully optimized in order to minimize fragmentation processes and to insure that the detected species do not derive from electrochemical events induced by the ionisation process. We must add a word or warning, however. Whereas certain fragmentation processes can be easily detected by the very nature of the products (e.g. loss of an H atom from Cp* to yield a CsMe4CH2containing fragment), it is never possible to unambiguously establish whether a species having an innocent looking chemical formula is in fact solution borne. Unfortunately, the ESI-MS investigation could not be carried out under solvent conditions identical to those of the stopped-flow investigation because of low instrumental sensitivity. A sufficient substrate concentration could only be achieved by using 5050 methanoVwater mixtures. The positive ion ESI-MS spectra recorded at pH 4 (acetic acid buffer) and at pH 1.8 (trifluoroacetic acid) in the absence of electrochemical processes (see Figure 6), show a number of mono-, di-, tri- and even tetranuclear species. The main species correspond to: Cp*MoOzf (256), Cp*Mo03Hzf (277), Cp*Mo02(MeOH)+ (29 l), Cp*2Mo2OsH+(535), Cp*3Mo30; (793), Cp*4Mo401IHC (1090). Each species is identified in parentheses by the m/z value of the lightest isotopomer in each envelope. The ESI-MS study confirms the presence of mononuclear cationic species as indicated by the stopped-flow and conductivity studies, although both solvent-free and solvated (water and methanol) dioxo species are detected. The mass spectrometric investigation does not allow us to conclude which of these species (or all) are solution borne, because either solvent loss or solvent addition processes may occur in the electrospray chamber. In addition, the m/z 277 species may also be generated by proton addition to the Cp*MoO2(OH) which is present in solution. The large amount of Cp*zMo205H+probably derives from the proton addition to Cp*2Mo205, which is certainly present in the MeOH-rich solvent mixture used (vide supra). Very small (under optimized conditions) amounts of species that manifestly derive from fragmentation processes are also observed. The more surprising finding, however, is the detection of heavier species. Whereas oligomerization processes may be induced by the

174

Perspectives in Organometallic Chemistry

ionization in the MS injection chamber, a number of observations are consistent with these oligonuclear species being solution borne, at least the trinuclear one. The most convincing indication is the observed response to the electrochemical reduction, vide infia. 542.9

200

400

600

800

1000

1200

1400

Figure 6 Electrospray mass spectrum of a 0.1 mMsolution of Cp*2Mo205 in 1:1 H2O/MeOH Heated capillary temperature = 100°C. (a): pH 4, E = -0.35 V; (b): pH 1.8, E = 0 K The essence of the coupled electrochemistry - mass spectrometry study is as follows. A continuous solution flow is passed through the electrochemical cell via a system of coaxial inlet and outlet capillary tubes whose orifices are in close proximity to the working electrode surface. Mass spectra are continuously recorded while the potential is scanned at constant rate from an initial to a final chosen value and back, just like in a cyclic voltammetric experiment. The dead time of the transfer line (ca. 20 s) requires, however, very slow scan rates (ca. 0.1 V s-*). As the MS analysis is always carried out on freshly electrolyzed solution, the reverse scan carries no information on chemical dynamics. However, it provides information on the history of the electrode (e.g. film deposition). A thee-dimensional space, namely peak intensities vs. time (or electrode voltage) and vs. m/z, results from the experiment. The integration of this space in a narrow time (potential) range gives the average MS in the chosen time range. An example is Figure 6 (starting potential, no electrochemical activity). Integration is a narrow m/z range, i.e. a range corresponding to the isotope envelope of a given species, gives the time (potential) evolution of the relative abundance of such species. Figure 7 shows the potential evolution of selected starting species at pH 1.8:l These plots carry a considerable amount of useful information. First, the observed decrease after reaching a certain potential value signals the occurrence of an electrochemical process which consumes either the species itself (if this is solution borne) or its precursor (if the species is generated in the mass spectrometer chamber). Second, this potential value qualitatively correlates with the standard potential for the redox process of the given species (no thermodynamic information can be extracted from this value, however!). Third, the symmetry of the plots indicates the absence of phenomena such as film deposition that would alter the nature of the electrode surface during the potential scan. Fourth, the comparison between the shapes of different plots can reveal chemical relationships between the different species. For instance, if the trinuclear species Cp*,M030? were generated from mono- and dinuclear precursors (see Scheme 4) by condensation in the MS chamber, such species could not disappear at a less negative potential relative to the proposed precursor species. Thus, Figure 7 demonstrates that such species, whose likely genesis is as shown in Scheme 4, is already present in the analyte. It can furthermore be concluded that the trinuclear species is reduced more easily than the dinuclear and mononuclear species and that its formation equilibrium is slow relative to the residence

Results and perspectives of high oxidation state organomolybdenum chemistry

175

time in the transfer line from the electrode surface to the MS instrument (ca. 20 s). The plot of the Cp*MoO2+ species (not shown) follows closely the shape of the Cp*M003Hc plot. Therefore, if only one of these two species is solution borne, the process generating the other one must be rapid in the time scale of the experiment. d z = 276-285

Cp*MoO,H,+ I

0

-0.5

-1

d z = 535-548

0

-0.5

0

-0.5

0

Cp*,Mo,O,H+

-1

m/z = 795-8 13

-0.5

-0.5

0

Cp*,Mo,O,+

-1 -0.5 E N vs. AgIAgC1

0

Figure 7 Potential dependent abundance oj selected species during a linear potential sweep (scan rate 0.5 mV s-' from 0.0 V to -1.0 Vand back: nH = 1.8).

Scheme 4

The m/z plot corresponding to the most negative potentials and the time evolution plots of the newly identified reduction products carry the most valuable information of this experiment. The main new species observed during the experiment carried out at pH 4 are shown in Figure 8. The molecular assignment is supported in all cases by MS" fragmentation studies, the details of which will not be reported here." The majority of the reduction products (all except Cp*2M0204Hf, which is the protonated version of the known22 Cp*2Mo204) are unprecedented and their detection by this technique means that they survive under the given experimental conditions for at least minutes. It must be again emphasized that each observed species is not guaranteed to be solution borne. However, given the well known mild ionization conditions of the ESI-MS technique, most of them probably are. Therefore, the power of this analytical technique in terms of identifjkg new synthetic targets can immediately be appreciated. Other undetected species may, of course, be present in solution. Neutral species are not revealed by the MS experiment whereas negatively charged species would be revealed by operating in negative detection mode. Species with multiple positive charge have not been detected in our experiments on this precursor, although their presence and rapid equilibria with the observed monocations through coordinatioddissociation of the acetate anion cannot be excluded.

Perspectives in Organometallic Chemistry

176

(b) Dinuclearspecies

(a) Mononuclearspecies

035

-10

035

340

Potentiat' V vs Ag/AgCI

350

360

d.

(c) Trinuclear species

Potential V vs Ag'AgCI

d z

Potentid V vs Ag'AgCI

d~

I

'

Figure 8 Potential dependent abundance for reduction products formed at pH 4 (left), corresponding isotope envelope (center) and tentative structuralformula (right).

A few interesting observations can be summarized as follows. Within each nuclearity, more reduced species start to form at a more negative potentials, e.g. the initial potential Ein for the generation of dinuclear species (this is the intercept of the tangent line drawn at the inflection oint of the potential rise curve) are: -0.55 V for C P * ~ M O ~ ~ ~ ~ O ~-0.74 (OAC V)for +, -0.78 V for Cp*2M02'>'~O(OAC)~+.Within a given oxidation state, Cp*2Mo? B7V02(OA~)2+, oligonuclear products are formed at less negative potentials relative to mononuclear products: e.g. for Mo(V) species: -0.55 V for Cp*~Mo203(0Ac)' vs. -0.76 V for Cp*MoO(OAc)+. Amongst the dinuclear species, Cp*zMo204H+ is generated at the least negative potential (Ein = -0.48 V), which corresponds to the Ein for the consumption of the Cp*zMo205H+peak. The experiment carried out at pH 1.8 (in the presence of trifluoroacetic acid) gives further interesting information. A selected number of products are shown in Figure 9. Some of these are related to those obtained in the presence of acetic acid at pH 4, for instance Cp*2Mo20(02CCF3)<. Others, however, are new, for instance C ~ * ~ M O ~ O ( O H ) ( O ~ C C F ~ ) ~ + whereas certain species that are observed at pH 4 do not form at pH 1.8. These observations are useful for deducing a possible reduction mechanism. Also interesting is the fact the a greater degree of reduction can be achieved at lower pH. For instance, the oxidation state I11 is achieved within the dinuclear framework with species C~*~MO~(OH)(O~CCF~);, whereas at pH 4 the reduction stops at the oxidation state IV. Analogously, the trinuclear Cp*3Mo30; (oxidation state +4) product is obtained at pH 1.8, whereas the reduction stops at Cp*3Mo306+ (average oxidation state +5.33) at pH 4 at the achievable potentials. This pH dependence is expected because the reduction processes necessitate the protonation of 0x0 ligands. It is interesting to note that no acetate nor coordinated solvent molecules has ever been observed for these reduced trinuclear clusters (cJ:chemical reduction section below). Finally, we note that the Cp*/Mo ratio is always 1 except for species that are proven to result fiom fiagmentation processes by the MSn experiments. Therefore, the Cp*Mo bond resists protonolysis, even under strongly acidic conditions, for all observed reduction products as well as for the starting material.

Results and perspectives of high oxidation state organomolybdenum chemistry

(a) Dinuclear

177

(b) Trinuclear

m / ~ =531.8-533.4

0

0.5

0.5

0

1

0.5

0

,

,

1

05

1

m/z =6 15631

2

3

5

m/== 768-779

-

0

0.5

0 ----

0

05

4 5

0.5

0

= 75 1-761

= 697-7 1 1

& 0

4.5

0 &Z

&Z

-1

1

0.5

0

45

-1

45

0

Potential vs AgIAgCl

0

Potential vs Ag/AgCl (111,111)

Figure 9 Potential dependent abundance for reduction products formed at p H 1.8 (“e#), and tentative structural formula (right). The m/z integration range in selected cases is limited because of peak overlap

2.4 Chemical reduction studies The next step in our project is to verifj whether any species which is observed by the combined electrochemical reduction and ESI mass spectrometric analysis can be chemically generated, isolated and studied. We chose Zn as the reducing agent for reasons of compatibility with the aqueous methanol medium at low pH. Reduction turns out to be sluggish, with the solution showing a variety of color changes through brown, red and green, but eventually a blue precipitate is obtained after a few days stirring at room temperature in an acetic acid medium (pH 4). In a separate experiment, an intermediate red reduction product was isolated and shown to correspond to the already known Cp*2M0204, confirming the ECESIMS indications. The final blue product turns out to be diamagnetic Cp*zMo202(02CCH3)2, as shown by NMR and X-ray crystallographic analyses, see Figure This compound cannot obviously be revealed by the electrochemical/mass spectrometric analysis. It is, however, related to the mixed-valence cation Cp*zMo202(02CCH3)2+ (see Figure 8). An independent cyclic voltammetric study of the isolated compound in CH2C12 shows that it undergoes a reversible one-electron oxidation process at Eln = -0.50 V, confirming the existence and stability of the cationic species.

178

Perspectives in Organometallic Chemistry

Figure 10 An ORTEP view of the Cp*2M0202(02CCH3)2molecule.

The structure shows two 0x0 and two acetato groups bridging the two metal atoms that are, in addition, held together by a direct metal-metal interaction (Mo-Mo: 2.5524(3) A). This structure is based on the four-legged piano stool ligand arrangement which is typical of CP*MO'~dinuclear species, as seen for instance in [Cp'Mo(S)(SMe)]2 and in other similar derivative^:^ but it is the first example of a dinuclear Cp*Mo" compound which is fully supported b 0-donor ligands. It can be considered as an example of a stabilized aqueous Cp*2Mo202I+ species. The zinc reduction carried out at lower pH (trifluoroacetic acid) under otherwise identical conditions is accompanied by analogous color changes, yielding a green-blue precipitate. From this solid, however, only dark blue trinuclear reduction products could be crystallized. The nature of the isolated product delicately depends on the crystallization conditions. From THF-heptane, crystals that can be formulated as [CP*3M03(p-o)6-,-,(p OH),,I2+ x 2 CF3COO- were recovered. High disorder in the trifluoroacetate and Cp* fiagments, confirmed on several crystals from different batches, hampered a satisfactory refinement of the X-ray data, thus the intimate structural details cannot be obtained. However, the basic equilateral triangular Cp*3Mo3(p-0)6 core (Mo-Mo: 2.79 A) and the trifluoroacetate arrangement as depicted schematically in Figure 11 are unambiguously shown by the data refinement?' The two trifluoroacetate ions establish close contacts between both oxygen atoms and different p-0x0 ligands above and below the M03 triangle, as shown in the Figure, the 0-0distance being typical of hydrogen bonds (2.6 A). Therefore, the value of n in the chemical formula is most likely 4, corresponding to an average oxidation state of +4.33 for the metal atoms. It is relevant to note that a related structure, also characterized by severe disorder, has previously been reported for a compound formulated as [Cp*3Mo3(pL-0)s-,,(p0H),,l2' x 2 Cl-, this being obtained by a synthetic strategy similar to ours (zinc reduction of Cp*Mo02C1 in CHCI3 in the presence of concentrated HC1)?6 In that case, however, the most likely value of n was proposed as 5 . When the crystallization was carried out from THFdiethyl ether, a different compound, identified as [Cp*3M03(p3-O)(p-O)3(p02CCF3)3]2[Zn2(02CCF3)6],was crystallized. The dianion has a lantern type structure, similar to other previously reported neutral Zn2(02CCR)L2 c ~ m p l e x e s . ~The ~ - ~structure ~ of the unprecedented monocation is shown in Figure 12.3

Results and perspectives of high oxidation state organomolybdenum chemistry

179

0 T 0 CF3

Figure 11 Structural arrangement oj compound [Cp *~Mo~O~-~(OH)~][CF~COO]~ The cation has a few novel and interesting features. No trinuclear Mo complex, either with or without Cp* ligands, in an oxidation state as high as +5 has been previously reported. The Mo304 core is typical of the oxidation state +4, and few examples in more reduced states are also known, but none in higher ones.31The metal atoms are not directly bonded to each other (average Mo-Mo = 3.1287(15) A). This can be readily rationalized from simple considerations based on the coordination geometry and on the electronic structure?' The combined EC-ESIMS and chemical synthesis results lead to the proposal that the reductive chemistry leading to mononuclear, dinuclear, and trinuclear reduction products are probably independent. However, acid/base equilibria can probably convert one type of structure into another with the same nuclearity via oxohydroxo/aqua tranformation and water dissociatiodassociation equilibria. Furthermore, the anions provided by the supporting acid, especially the weakly coordinating trifluoroacetate ion, may also establish facile associatioddissociation equilibria. It is clear that this promises to be an extremely rich chemistry and the synthetic studies that we have carried out so far have probably just uncovered the tip of the iceberg. For instance, the current picture on the trinuclear species is summarized in Scheme 5, where the oxidation state decreases fiom right to left and the oxygen content decreases fiom top to bottom. The circled species have been observed as such in the EC-ESIMS experiment, while the boxed species are related to the isolated compounds shown. More reduced species could perhaps also be synthesized under superacidic media, and a wealth of compounds where the bridging oxohydroxo ligands are replaced by ligands based on other donor atoms may also be reasonable synthetic targets. Indeed, an example of this kind, complex [Cp*3Mo3S4If (this being unsupported by additional ligand, just like the Cp*3Mo30; species observed by EC-ESIMS), has previously been rep~rted.~' 3 CONCLUSION AND PERSPECTIVES Although aqueous organometallic chemistry on one side, and high oxidation state (mostly 0x0-based) organometallic chemistry on the other side, have been intensively investigated and both are now mature areas, the combination of them, i.e. high oxidation state aqueous organometallic chemistry, has not so far attracted a lot of attention. With our initial

180

Perspectives in Organometallic Chemistry

investigations of the speciation and reductive chemistry and electrochemistry of Cp*2Mo205 that are described in this report, we hope to have demonstrated the enormous opportunities that exist in this area. These opportunities encompass fundamental studies of the synthesis, characterization, redox and speciation of novel compounds, as well as coordinatiodactivation chemistry studies of these compounds with a variety of ligands, and applied studies aiming at the functionalization and the catalyzed or electrocatalyzed transformation of different substrates in an aqueous environment. Relative to inorganic substances, the Cp*Mo-based complexes present the advantage of a well-defined coordination environment, three coordination positions being blocked by the robust Cp* ligand. This limits or avoids oligomerizatiodpolymerizationphenomena and makes model and mechanistic studies more easily approachable. Indeed, only compounds with nuclearities up to four have been observed in this study (we note, however, that film formation phenomena have been observed by electrochemical reduction at high pH, a phenomenon that we are also further investigating). We further remark the exceptional resistance of the Cp*Mo bond against hydrolysis even at pH 1 for extended periods (several days). Our studies have also provided strong indications for the facile dissociation of water and/or weakly coordinating anions, suggesting that the reduced species should be able to easily coordinate and activate a variety of multifunctional substrates. We look forward to our exploration of this area. Finally, it is quite clear that an equally rich chemistry and electrochemistry should exist for a large variety of other high oxidation state organometallic systems.

Scheme 5 Acknowledgement Different aspects of this research have been sponsored by the CNRS, the French Ministery of Education, the Conseil RCgional de Bourgogne (purchase of the stopped-flow apparatus and post-doctoral fellowship to DS), the French and German Ministeries of Foreign Affairs (Procope program, Dijon-Kaiserslautern), the French Embassy in Israel (Arc-en-CielKeshet program, Dijon-Jerusalem) and NATORUBITAK (travel grant to FD). We are grateful to

Results and perspectives of high oxidation state organomolybdenum chemistry

181

these organization for financial help, to Dr. Mikhail Vorotyntsev for helpful discussion, and Mr. Juan Garcia for technical assistance. References 1 2 3 4 5

6 7 8 9 10 11 12

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

P. Kalck and F. Monteil, A h . Organometal. Chem., 1992,34,219. D. Sinou, Topics Curr. Chem., 1999,206,41. B. E. Hanson, Coord Chem. Rev., 1999,186,795. C. Muller, D. Vos, and P. Jutzi, J. Organometal. Chem., 2000,600, 127. G. E. Jaouen, J. Organomet. Chem. (special issue: bioorganometallic chemistry), 1999,589. U. Koelle, Coord. Chem. Rev., 1994, 135,623. M. Cousins and M. L. H. Green, J. Chem. SOC.,1964, 1567. M. Herberhold, W. Kremnitz, A. Razavi, H. Schollhorn, and U. Thewalt, Angew. Chem. Int. Ed. Engl., 1985,24,601. J. W. Faller and Y. Ma, J. Organometal. Chem., 1988,340,59. P. Leoni, M. Pasquali, L. Salsini, C. di Bugno, D. Braga, and P. Sabatino, J. Chem. SOC.Dalton Trans., 1989, 155. A. L. Rheingold and J. R. Harper, J. Organometal. Chem., 1991,403,335. M. S . Rau, C. M. Kretz, L. A. Mercando, G. L. Geoffroy, and A. L. Rheingold, J. Am. Chem. SOC.,1991,113,7420. M. S . Rau, C. M. Kretz, G. L. Geoffroy, and A. L. Rheingold, Organometallics, 1993, 12,3447. M. J. Bunker and M. L. H. Green, J. Chem. SOC.,Dalton Trans., 1981,847. C. Couldwell and K. Prout, Acta Crystallogr., 1978, B34,933. D. Saurenz, F. Demirhan, P. Richard, R. Poli, and H. Sitzmann, Eur. J. Inorg. Chem., 2002, 1415. E. Collange, J. Garcia, and R. Poli, New J. Chem., in press. R. H. Holm and J. M. Berg, Acc. Chem. Res., 1986,19,363. J. Gun, A. Modestov, 0. Lev, D. Saurenz, M. Vorotyntsev, andR. Poli, Eur. J. Inorg. Chem., submitted. A. D. Modestov, S. Srebnik, 0.Lev, and J. Gun, Anal. Chem., 2001,73,4229. J. Gun, A. Modestov, 0. Lev, and R. Poli, in preparation. H. Arzoumanian, A. Baldy, M. Pierrot, and M. Petrignani, J. Organometal. Chem., 1985,294,327. F. Demirhan, P. Richard, and R. Poli, Inorg. Chim. Acta, in press. M. Rakowski DuBois, M. C. VanDerveer, D. L. DuBois, R. C. Haltiwanger, and W. K. Miller, J. Am. Chem. SOC.,1980,102,7456. F. Demirhan, R. Poli, and P. Richard, unpublished observations. F. Bottomley, J. Chen, K. F. Preston, and R. C. Thompson, J. Am. Chem. SOC.,1994, 116,7989. W. Clegg, I. R. Little, and B. P. Straughan, J. Chem. Soc., Dalton Trans., 1986, 1283. W. Clegg, P. A. Hunt, and B. P. Straughan, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1995,51(pt.4), 613. B. Singh, J. R. Long, F. F. deBiani, D. Gatteschi, and P. Stavropoulos, J. Am. Chem. Soc., 1997,119, 7030.

182

30 31 32

Perspectives in Organometallic Chemistry

F. Demirhan, J. Gun, 0. Lev, A. Modestov, R. Poli, and P. Richard, J. Chem. SOC, Dalton Trans., 2002,2 109. D. T. Richens, in ‘The Chemistry of Aqua Ions ’, Wiley: Chichester, 1997. P. J. Vergamini, H. Vahrenkamp, and L. F. Dahl, J. Am. Chem Soc., 1971,93,6327.

MODULATION OF ELECTRONIC BEHAVIOUR OF METAL CARBONYL CLUSTERS

D. Collini,' C. Femoni,' M. C. Iapalucci,' G. Longoni' and P. Zanello2 Dipartimento di Chimica Fisica ed Inorganica, UniversitA di Bologna, viale Risorgimento 4, 401 36 Bologna, Italy.

* Dipartimento di Chimica, UniversitA di Siena, via A. Moro, 53 100 Siena, Italy.

1 INTRODUCTION

It has been suggested that ligand shell stabilized metal clusters are valid candidates to assemble fbnctional devices for data storage and could potentially represent the ultimate solution for miniaturization in microelectronics and nanolitography. l6 Could the carbonyl metal clusters play a role in this area? In the following text we will analyse the nature of transitiofi-metal carbonyl clusters in the attempt to verifL whether or not these molecular compounds might be worth of investigation for the above purposes. In principle, a molecular carbonyl metal cluster is a potential molecular capacitor, as in general is based on a globular kernel of metal atoms shielded by a shell of carbonyl ligands. Although the effectiveness of the carbonyl shielding is unknown, a comparison between certain behaviour of carbonyl metal clusters and bare post-transition metal clusters or fullerenes suggests that the carbonyl shell might effectively insulate the metallic core. Indeed, the EPR spectra of pePt3(CO)5]- and [Agl~Fe8(CO)3$ are unaffected by the presence of their respective [Fe#t3(CO)15]o - and [Agl3Feg(C0)32l3-- diamagnetic congeners at all ratios and temperatures, both in solution and the solid state. Their EPR pattern is like a fingerprint that enables their detection even in a very minute amount in solution or doped crystals. The metal core acts as a quantum dot, in which the cluster valence electrons (CVE) are confined.' Coulombic repulsion may generate an energy barrier for intermolecular exchange of electrons also in ligand-fiee post-transition metal clusters, e. g. M9"- (M= Ge, Sn, Pb; n = 2 4 , and fullerides. However, at least in the case of C60b fullerides, the energy barrier is not sufficient to hinder intermolecular electron exchange in solution, which is fast on the time scale of 13CNMR. Mixtures of differently charged C6c and Cm3-hllerides display 13CNMR chemical shifts intermediate between the values of their respective pure anions.lo Moreover, exchange of electrons in the solid state cannot be ruled out. In order to make a molecular cluster formally assimilable to a spherical metal capacitor, at least two other conditions ought to be preliminarily fullfilled: 1) the molecular cluster should be able to reversibly accept and release electrons. In practical terms, it means that it should either be multivalent (by multivalent clusters we mean

184

Perspectives in Organometallic Chemistry

clusters which are stable on the work-up time-scale with different free charges and remain structurally invariant) or, at least, exhibit electrochemically reversible redox properties, 2) the metal core of the molecular cluster should undergo a transition fiom an insulator to a semiconductor or a metallic regime, Moreover, for some applications, a size of the metallic core of a few nanometers in diameter will probably be necessary. This last condition will likely be interconnected to items (1) and (2) in most cases, but not necessarily always. Indeed, several recent low-valent metal clusters, stabilised mainly or uniquely by carbonyl ligands, are well-defined molecules or molecular ions trespassing in the field of nanomaterials (1-1000 nm) from the lowest limit. For instance, structurally characterised cluster compounds such as [&nNi38Pf6(C0)48]n- (n = 4-6), 14 IN~z~.P~~o(CO l2 ) ~ H1 ~2P]d~28,Ptl3(C0)27(PMe3)(PPh3)12, l3 ~d33Ni9(C0)41(Pph3)6]4-, 15 Pd59(C0)32(PMe3)21, and the giant 3-shell Pd14~(CO)@Et3)30 (x - 60) l6 display metallic cores with diameters in the '1-2 nm range. However, while [&,,Ni3~&(C0)~]~fulfils item (I)," all other compounds do not. To our knowledge, all Pd-containing clusters only display irreversible reduction and oxidation waves, probably owing to a too low Pd-Pd, Pd-L and MPd binding energy. Most carbonyl metal clusters displaying multivalence or showing electrochemically reversible redox behaviour have so far been fortuitously obtained." To develop molecular clusters as molecular capacitors, it would be desirable to know how to tailor their geometry and composition in order to induce redox propensity and multivalence. The availability of strategies to modulate the electronic behaviour of a cluster will eventually allow tailored tuning of its redox potentials. In the following pages we will review our present state of knowledge regarding items (1) and (2). Before doing that, however, we will briefly address the question why metal carbonyl clusters are not necessarily close-shell and could be tailored to display redox propensity. 2 CLUSTER VALENCE ELECTRONS AND METAL ARCHITECTURE It is a widespread opinion that metal carbonyl clusters adopt close-shell electronic configurations and exhibit a quasi-bijective mapping between number of cluster valence electrons (CVE) and geometry of the metal frame, as a result of the relatively wide HOMOLUMO gap induced by the carbonyl ligands. Countless examples of structurally-characterised low-nuclearity clusters lend support to the above belief l9 The exceptions are relatively few. For instance, all octahedral or trigonal antiprismatic clusters display 86 CVE (e. g. pi6(CO)12]2- 24. However, 86 CVE do not necessarily mean octahedral or trigonal antiprismatic geometry, as shown by the trigonal prismatic [Pf6(C0)12l2-21 congener and the planar [Fe3Pt3(CO)15l2-.zz Most exceptions have been convincingly interpreted as due to "ad hoc'l conditions arising from different combinations of electronic and steric effects. l9 The fortuitous occurrence of miscellaneous ''ad hocl' conditions has also been invoked to interpret the rare examples of open-shell electronic configurations, l8 and the few scattered reports of significant electron-sink behaviour of metal carbonyl clusters. 18, 23 As a consequence of the above situation, a series of rules have been put forward, l9 which were able to rationalise the electron count and geometry of most clusters and proved to be extremely useful in designing and making new compounds. However, implicit in multivalence and, to a lesser degree, in electrochemically reversible redox behaviour and open-shell configurations, is the absence of any propension of the metal

185

Modulation of electronic behaviour of metal-carbonyl clusters

cluster for bijective mapping between C W and structure of the metal frame. A given cluster stereochemistry should remain almost unaltered upon addition or subtraction of electrons. Probably, the first hint of a possible strategy to override one of the above rules has been provided by complementary works carried out in Dahl and our laboratories on non-centred and Ni-centred Nil,& (E = post-transition element or molecular moiety) icosahedral carbonyl clusters. As a result of these investigations, non-centred (Figure 1a) [ N ~ ~ o ( E R ) ~ ( C ~(E ) I= S]~' P, As, Sb, Bi; R = alkyl or aryl substituent), 24-28 Ni-centred (Figure lb) [NilOE2(p12Ni)(CO)lg]"- (E = Sb, Bi, Se), 2p31 and Ni-centred (Figure lc) [Nilo(SnRh(p12-Ni)(CO)1s]~(R = alkyl substituent) 32 and [Nilo(Sb+Ni(C0)3)2(p12-NI)(C0)1g]"- (n = 2-4) 33 derivatives became available. The non-centred nickel clusters display conventional electron counts and only exhibit irreversible redox behaviour. In contrast, the Ni-centred species show anomalous number of CVE (8-10 additional electrons with respect to the non-centred congeners), reversible electrochemical redox behaviour and multivalence. 29-33

P 8 a

U

b

C

Figure 1 Schematic structures of non-centred (a) [Ni1dER)2(CO)18J2-,Ni-centred (6) [Nil&(p12-Ni)(CO) 181" and Ni-centred (c) m i l dER)~(plrNi) (CO)181" clusters (E atoms are shown as blackened spheres).

EHMO analysis of Ni10E2(p12-M)(C0)18 model compounds with different interstitial M atoms pointed out that the out-of-phase combinations of the five d A 0 of the interstitial atom with the suitable MO of the NiloE2 cage are not sufficiently destabilised when M=Ni and fall in the 33 fiontier rather than the antibonding region. This result suggested that interstitial metal atoms, at difference fiom interstitial main-group elements, 34 do not necessarily act simply as an internal source of electrons but may alter the number of cluster valence molecular orbitals as a function of the relative energies of their valence orbitals. The presence of these additional weakly-antibonding orbitals in or close to the HOMO-LUMO gap increases the electron count and promotes redox aptitude, whereas a purported strengthening effect of the N-E moieties enables multivalence in these clusters.

186

Perspectives in Organometallic Chemistry

Similar conclusions stem from analysis of non-centred M&4-E)&g (M = Fe, Co, Ni) and M-centred Mtt(~-E)(pg-M)Lg (M = Ni, Pd) hexacapped cubic clusters. ” The latter exhibit miscellaneous electron counts in the 121-1 30 CVE range, while those of the former are comprised in the 99-120 range. Their miscellaneous CVE have been rationalized by EH, DFT and Xu calculations, that pointed out the relevance of both E and the n-donor or n-acceptor properties of L in determining the electron count within each series. However, the observation that the M-centred are sistematically more electron-rich than the non-centred clusters leads once again to the conclusion that the inclusion principle, 34 so nicely effective with interstitial main-group elements, can be violated when the interstitial atom is a suitable metal. Interestingly, several of the above clusters display open-shell electronic configurations but, unfortunately, their redox aptitude has only received occasional attention. 3 MODULATION OF ELECTRONIC PROPERTIES OF A METAL CARBONYL CLUSTER AND PROMOTION OF CAPACITOR BEHAVIOUR.

To develop carbonyl metal clusters as molecular nanocapacitors to apply in molecular electronics and lithography it would be desirable to know how to design cluster composition so to modulate the electronic behaviour of a given cluster geometry. Such a modulation could in principle induce redox aptitude whenever is lacking and tune the redox potentials of those species already multivalent or displaying redox aptitude. Following the observations of the previous heading 2, the first strategy we decided to explore consisted in the substitution with nickel of the interstitial atom of a suitable M-centred cluster. The first targets have been the body-centred cubic [Rh14(CO)25]~‘ 36237 and [Rhl~(C0)30]~38 clusters. The attempt to substitute a Rh interstitial atom with Ni in the architecture shown in Figure 2 could appear awkward, since Ni has a lower heat of atomisation than Rh and the Ni-CO bond is energetic. We speculated that a Rh12Ni(pg-R.h) Rh&g-Ni) isomerisation could become thermodynamically favoured, in case of contraction of the Ni-centred Rhs cube owing to the consequent gain of the energy of four extra Rh-Rh interactions (AH = 4 I3~h-w+ 2 ENi-CO - 4

-

Figure 2 The metal architecture of [ R h , ~ ( ~ N ~ ( C O (Ivi ) ~ jisf shown as a blackened sphere)

Modulation of electronic behaviour of metal-carbonyl clusters

187

ENi-a - 2 E~aco).The above crude reasoning turned out to be effective, as we successfully synthesised a series of Ni-centred [Rh14-xNix(CO).Z5]"(x = 1, n = 5 ; x = 2, n = 4; x = 5 , n = 2, 3) bimetallic clusters 39 isostructural to m14(CO)25]4-,as well as the hexacapped body-centred cubic [Rh14Ni(CO)zg]~derivative, related to [Rhl5(CO)30]~-. Notice that substitution of an interstitial Rh with a Ni atom only brings about minimal variations of the number of CVE, in comparison with those (8-10 CVE) triggered by interstitial occupation of an empty cluster cavity. Thus, the pentacapped Ni-centred cubic NiRh clusters feature 2-4 additional CVE with respect to the Rh-centred w14(CO)25]c. The related Nimh exchange in the hexacapped body-centred cubic architecture of [Rh15(CO)30]~seems to correspond to a minor requirement of 2 CVE, as shown by the formula of the bimetallic [Rh14Ni(CO)2g]d cluster. However, EHMO calculations confirm that a hypothetical [Rhl5(C0)28lw,isostructural with [Rh14Ni(C0)28]4-,would display a close-shell configuration with n = 3 and, therefore, would be 2 electrons short with respect to the latter. Preliminary experiments indicate that all above tetradecanuclear clusters display up to two oxidation and two reduction steps, several of which have features of electrochemical reversibility. Furthermore, progressive substitution of Rh caps with Ni triggers multivalence. Accordingly, both the close-shell [Rh9Ni5(C0)25l3- and the open-shell [Rh9Nis(CO)z5Izspecies have been isolated in the solid state and structurally characterised. The relevance of the peripheral metal atoms in determining the most favorable number of CVE of a given metal architecture is also proved by the isostructural and non-isoelectronic [pt38(C0).ul2-41 and [Ni24ptl4(CO)~~]~42 pair of clusters. The latter formally derives from the former by substitution of all corner Pt with Ni atoms and his metal fiame is shown in Figure 3. The above findings regarding Ni-centred icosahedral [Nil1E2(C0)1gln-and Ni- or Rh-centred cubic [Rh14-xNix(C0)25]w(x = 0-5) suggested that the presence of interstitial metal atoms could induce significant reversible redox behaviour also in metal carbonyl clusters. Indeed, platinum carbonyl clusters containing one and two interstitial platinum atoms such as [Pt19(C0)~]~and [Ptz4(C0)30]~were reported to show several redox waves with features of electrochemical reversibility. 43* In contrast, spectroelectrochemical experiments on pt26(cO)32l2-and [Pt3g(cO)~]~-, 439 44 which respectively contain three and six interstitial Pt atoms, mainly disclosed irreversible features. This somehow contradictory situation prompted a systematic investigation of all other carbonyl clusters known to contain either Pd or Pt interstitial atoms. The results have been rewarding. First of all, [I6-,,Ni36Pt4(C0)45IRand [& nNi3gPt,j(C0)4g]w,which respectively contain four and six interstitial platinum atoms, display electron-sink features. 17*45 Generally, they exhibit one oxidation and up to four reduction steps with features of electrochemical reversibility, and their cyclic voltammetric profile displays almost equally-spaced waves. Moreover, it turned out that protonation of the parent anion (the hexaanion in both above compounds) is a further possibility to modulate the electronic properties and tune the redox potentials of a given cluster. 17, 45 As a corollary, these results provided circumstantial evidence of the formation of hydride derivatives upon protonation of [Ni36Pt4(CO)45]6-and [Ni3gPts(CO)4g]6-,even if 'H, 13C and 195 Pt N M R experiments hiled in giving a direct or indirect proof of their presence. It appears conceivable to suggest that these rather big anions tumble in solution so slowly to give rise to exceedingly broadened 'H, as well as 13C and lg5Ptsignals due to anisotropy. However, this conclusion is contrasted by the observation of well-defined 'H and 31PNMR signals of the interstitial hydride atoms and peripheral phosphine ligands of the comparable-in-size HlzPd2gPt13(CO)n(PMe)(PPh3)n derivative. l3

188

Perspectives in Organornetallic Chemistry

Figure 3 The metalframework of [Ni29t14(C0)44]4(Pt atoms and Pt-Pt bonds are blackened). The presence of interstitial palladium atoms, as well as the nature and energy separations 8), 46 [Ni16Pd&0)40]~, 47 of their frontier orbitals, candidate [Ni~-~PddC0)48]~’ (x [Ni22Pd20(C0)48]~46 and [Ni%Pd20(C0)54l6-47 as potential electron-sinks. Indeed, EHMO calculations suggest the absence of a well-defined HOMO-LUMO gap and preliminary SQUID and Cantilever measurements on [Ni16Pd16(C0)40]~ indicate the presence of several unpaired electrons per cluster ion. 48 However, none of them and neither the phosphinesubstituted [NisPd33(C0)41(PPh3)6I6- hexaanion l4 withstands redox conditions. It appears probable that the weakness of all bonds involving palladium could favour fragmentation or condensation processes upon reduction or oxidation, respectively. This shows that a relative robustness of the ligand-stabilised metal cluster is a fbndamental prerequisite for developing nanocapacitor behaviour. In keeping with the above suggestion, the [Ni32c6(co)36]6- and [Ni3&6(C0)42]6- carbide clusters, whose metal fiamework is strengthened by the presence of six interstitial p8-c atoms, exhibit electron-sink behaviour comparable to that of [Ni38Pts(C0)48]6‘.49,50 Strictly speaking there are no interstitial nickel atoms in the above two compounds. However, eight nickel atoms of [Ni32c6(cO)36l6 are 12-coordinated (9 Ni + 3 carbide atoms), whereas 6 out of the above 8 nickel atoms increase their coordination to 13 in [Ni38C6(C0)42I6-(10 Ni + 3 carbide atoms). The cyclic and hydrodynamic voltammetric behaviour of [Ni32C6(C0)36]6- in acetonitrile solution is reported in Figure 4 as an example. On the left side of Figure 4 we report the energy levels at the frontier region of the cluster obtained with CACAO. 51 Observation of four reduction Steps up to [Ni32C6(C0)36]’@appears to be a direct consequence of filling all energy levels below a small but sufficiently defined ga of ca. 0.5 eV. The first two reduction products, vzz. the [Ni32C6(CO)36]” and [Ni32c6(co)36] % anions, are also readily

-

available upon chemical reduction with sodium-naphthalenide in dimethylformamide and appear to be stable in solution for several hours.

189

Modulation of electronic behaviour of metal-carbonyl clusters

From a formal point of view, the voltammetric profile of these Ni-Pt and Ni-C clusters closely resembles that exhibited by c60. As major differences, they exhibit slightly smaller spacings between consecutive redox couples than C60 (for instance, ca. 0.28 V in [Ni32C6(C0)36]&vs. ca. 0.5 v in c60) and redox potentials shifted toward less negative values.

'*

1889s

E (VOLT)

1872s 1851s

Figure 4 Thefrontier molecular orbitals and the voltammetric behaviour of [NijJC&O)j6/6-. Besides, there is also a remarkable similarity between the electrochemical behaviour of the above molecular high-nuclearity metal carbonyl clusters and mono-dispersed ligand-stabilised gold nanoparticles. Thus, the cyclic and differential pulse voltammetric profiles, as well as the separations between the current peaks (0.2-0.4V), of an ensemble of hexanethiolate quasimolecular Au1& nanoparticles (1.64nm in diameter) in solution "are strictly comparable with those shown by the high-nuclearity Ni-Pt and Ni-C carbonyl metal clusters. 17* 457 'O Incidentally, it is also worthwhile to take note of the remarkable correspondence between the electrochemical response in solution of the above ensemble and the tip-based tunneling spectroscopy of an individual hexanethiolate quasi-molecular Au16 cluster physisorbed on a gold-on-mica support. The STM tip experiment on the latter displa s a reversible coulombic staircase with six charging steps constantly separated by ca.0.34V. Such a correspondence seems rather relevant for candidating carbonyl metal clusters for potential applications in microelectronics. As a conclusion of this heading, several miscellaneous carbonyl metal clusters behave as molecular electron sponges capable to reversibly accept 2-8 electrons, each at a distinct potential. Indeed, the observed potentials are not sufficiently close to give rise to an unique envelope, as it occurs in molecules featuring several non-interacting electro-active sites. See, for instance, some nanometer-size organic macromolecules such as tetraphenylmethane and hexaphenylbenzene molecules funcionalised with four and six 2,5-dimethoxy-4-methyl-phenyl electro-active moietes, '4 res ectively, or the tetrakis { 4-[N,N-di(4-methoxyphenyI)amino] phenyl} phosphonium cation. However, the experimental CV profiles of the high nuclearity carbonyl metal clusters are suggestive of their entrance in the antechamber of a regime in which the fiontier energy regions are characterised by a quasi-continuum of energy levels.

"

"

190

Perspectives in Organometallic Chemistry

4 ON THE SIZE-INDUCED SEMICONDUCTOR-TO-METAL TRANSITION OF

MOLECULAR CARBONYL METAL CLUSTERS. How many metal atoms ought to be assembled before a discrete metal cluster could develop metal-like behaviour is a much debated question. 55-57 We will try here to address this point in the particular case of molecular carbonyl metal clusters. As partially anticipated in the previous paragraph, the voltammetric profiles of the redoxactive carbonyl clusters change as a function of nuclearity. A tipical profile of a low-nuclearity cluster with extensive redox aptitude is exemplified by that of [Rh14(C0)25]~ and is shown in Figure 5 . The difference with that of a high-nuclearity cluster, as exemplified by that of I N ~ ~ ~ C S ( C Oreported ) ~ ~ ] ~ in - Figure 4, is rather evident. The former is suggestive of a closeshell cluster displaying a rather wide HOMO-LUMO gap of ca. 1.5 eV (as inferable from the ]~-’~ difference in Volts between the formal redox potentials of the [ R ~ I ~ ( C O ) ~ Sand [Rh14(C0)&’~- pairs) and a pairing energy of ca. 0.3 eV in the related orbitals. The latter might be due to an open-, as well as a close-shell cluster, and the interpretation of the spacing may change as a consequence. In any case, the average separations between the formal electrode potentials of consecutive redox couples of carbonyl metal clusters displaying multiple reversible redox changes decrease as a function of the nuclearity. Indeed, the

I

1.0

.

l

0.5

.

l

0.0

.

l

.

-0.5 -1.0 Volt vs. SCE

l

.

-1.5

l

~

l

‘

-2.0

Figure 5 Cyclic and hydrodynamic voltammetric proJilesof [IUI~~(CO)ES]Iin dimethylformamide s o h tion consecutive one-electron transitions progressively become almost equally-spaced and tight as the nuclearity of the cluster increases. Average AEs down to 0.25-0.3 V are observed, strictly comparable with those displayed by the alkanethiolate quasi-molecular Au14 cluster in solution (0.2-0.4 V). 53 Miscellaneous plots of average AE vs. nuclearity (abscissa) for different collections of metal carbonyl clusters, in spite of the dispersion of the AE between consecutive redox couples and unrespectfidly of their composition, display satisfactory linear fits with correlation coefficients up to a ca. 99 % level. All plots agree in showing extrapolated intercepts at ca. 0.75 V in ordinate and 65-70 in abscissa. An example limited to clusters

Modulation of electronic behaviour of metal-carbonyl clusters

191

containing at least one 12-coordinated metal atom and featuring a correlation coefficient of 93 % is reported in Figure 6. Once one realizes that the separations between the formal electrode potentials of consecutive redox couples (AE) mainly mirrors the differences between their ionization energies, these AE may be extrapolated to quantify the energy separation between the fiontier one-electron energy levels. Within this premise, the present one-electron energy

5

5:1

-

10

15

20

25

30

35

40

45

50

55

60

65

70

75

NUCLEARITY Figure 6 Experimental average separation of consecutive redox couples (AEm in y) as a firnction of the nuclearity of the cluster. 1) [NillSb2(CO)18]*(n = 2-4); 2) [Ni11Bi2(CO)l8]" (n = 2-4); 3) [Ni,3Sb2(CO)24ln-(n = 2- 4); 4)[Rh14(CO)25ln-(n= 2-5); 5) [Ni5Rhs(CO)25ln'(n= 1-4); 6)[Pt19(C0)22In-(n= 3-7); 7) [Pt24(C0)30ln-(n = 1-5); 8) [Ni32C6(C0)36Iw(n = 5-10); 9 ) [HNi38C6(C0)42In-(n = 4-8); 10) [Ni38C6(Co)42lb (n = 5-9); 11) [HNi3&Pf4(CO)45Jn(n = 4-7); 12) [Ni36Ph(CO)45In-(n = 5-9); 13) [N&8Pf6(C0)aln(n = 5-9); 14) [HNi38pf6(CO)48Iw(n = 3-7). separations suggest that the above clusters still have a semiconductor rather than a metal-like nature. However, according to Figure 6, the size-induced metallisation process of the carbonyl metal cluster is well underway and the transition could be expected to occur for nuclearities above a likely threshold value of ca. 65-70. It is worth stressing that such a nuclearity is well within reach of molecular chemists, as shown by the recent isolation of the carbonyl-substituted Pd~(CO)x(PEt3)30(x 60) l6 derivative, that represents the present world-record of low-valent organometallic molecular clusters characterised by X ray studies. Furthermore, it is rewarding to note that the above conclusion is in good agreement with miscellaneous attempts to experimentally determine the

-

192

Perspectives in Organometallic Chemistry

size-induced metal-insulator transition. For instance, spectroscopic studies on CO adsorbed on palladium islands deposited on alumina films point out that aggregation of ca. 100 Pd atoms is necessary in order to observe transition from molecular to metallic features. 59 Moreover, very similar size regimes have been observed to determine a size-induced metal-insulator transition for gas-phase bare Hg and supported bare Au clusters. ”-”Finally, photoelectron spectroscopy studies on mass-selected bare Au, clusters suggest that bulk properties begin to emerge for n > 40. 6o 5 CONCLUSIONS AND PERSPECTIVES

On the basis of the results described in the previous paragraphs, we may conclude that molecular carbonyl clusters are truly assimilable to molecular nanosized capacitors, since: 1. the metal core is in the semiconductor regime and on the way to undergo a size-

induced semiconductor-metal transition, 2. metal cores of nanometer diameters (1-2 nm) are already available and cluster chemists can provide even more interesting examples in the future, 3. the surrounding shell of carbonyl ligands probably provide a sufficiently efficient intermolecular shielding, 4. the electronic properties of the molecular cluster can be modulated by subtle variation of the metal composition, as well as by involvement of interstitial metal and non-metal atoms, 5 . as a direct consequence of 4), the chemical behaviour and, above all, the redox potentials of a given cluster geometry can also be tuned to a certain degree. Therefore, molecular carbonyl metal clusters appear to deserve consideration as possible valuable building materials to be investigated in the assemblage of molecular devices for nano-electronics and re-writable memories, as well as in advanced nano-lithography or molecular lithography (vide infra). Possible drawbacks arise from their air sensitivity and limited thermal stability. Therefore, their handling and study require strictly anaerobic conditions. However, some advantages arising fiom their nature and properties can be envisaged. For instance, one could profit of their anionic nature in preparing surfaces nanostructured with clusters to investigate in nano-lithography. The nano-structured surfaces so far prepared have been obtained via weak van der Waals interactions between cluster molecules 61 and the substrate (e.g. F&(C0)12 on highly oriented pyrolytic graphite, Au~~(PPh3)&16on poly(p-phenylene ethynylenes polymers, 62 silica and macromolecular templates 63-65). At difference, it appears conceivable to suggest that surfaces hnctionalised with cationic nitrogen moieties could be nano-structured by simple dipping into solutions of anionic carbonyl metal clusters via ionic interactions. In addition to catalysis, hybrid materials such as the above may have a fallout in advanced nanolithography. It has been demonstrated that nanoscale patterning of Langmuir-Blodgett films of gold nanoparticles by electron-beam lithography is viable. A related patterning should be possible also onto monolayers of anionic carbonyl metal clusters anchored onto cation-derivatised s u h c e s as the above, once given fer granted that the exposed and decomposed molecules will graft onto the surface and the unexposed molecules could be washed away with suitable solvents. At this stage it is only possible to warrant that most carbonyl clusters will be burned by exposure to an electron beam. Indeed,

Modulation of electronic behaviour of metal-carbonyl clusters

193

several years ago it was shown that [HNi38Pt6(co)48]5- salts undergo a series of transformations up to the Ni3Pt superalloy by varying the intensity of the electron beam. 66 Moreover, if one could deliberately drop one or more electrons in single selected molecules of the above mono-layer of clusters, an image of charges could be produced by acting on their reduction potentials. Such an image could be readable, modifiable and erasable at their corrispondent oxidation potentials. In such a way, molecular lithography and data storage could be pursued. Finally, it was cited that some organic molecules display electron-sponge behaviour (see the bottom of heading 3) complementary to those of the Ni-Pt and Ni-C carbonyl cluster anions. The above electro-active organic cations and cluster anions show opposite-signed and widely-separated redox potentials, that should stir up equilibrium reactions all the way shifted towards the right-hand side. However, tuning of the redox potentials of the cation with those of the anion seems worth of pursuing and appears possible with other kind of organic cations. A fast anion-cation electron transfer could enable the assembly of new charge-transfer saltbased materials.

''

Acknowledgements We wish to thank the University of Bologna and the MURST (Cofin2000) for a grant. References 1 U. Simon, in "Metal Clusters in Chemisny", eds. P. Braunstein, L. A. Oro, P. R. Raithby, Wiley-VCH, Weinheim, 1999, Vol3, p. 1342. 2 G. Schmid, Y-P. Liu, M. Schumann, T. Raschke, C. Radehaus, Nuno Lett., 2001,1,405 3 S. Hoeppener, L. Chi, H. Fuchs, Nan0 Lett., 2002,2,459. 4 M. H. V. Werts, M. Lambert, J-P. Bourgoin, M. Brust, Nan0 Lett., 2002,2,43. 5 A. C. Templeton, W. P. Wuelfing, R. W. Murray, Acc. Chem. Rex, 2000,33,27. 6 L. F. Chi, M. Hartig, T. Drechsler, T. Schwaack, C. Seidel, H. Fuchs, G. Schmid,Appl. Phys. Lett., 1998,66, S187. 7 J. Sinzig, L. J. de Jongh, A. Ceriotti, R. Della Pergola, G. Longoni, M. Stener, K. Albert, N. Rosch, Phys. Rev. Lett., 1998,81,3211. 8 S. Ulvenlund, L. Bengtsson-Kloo, in "Metal Clusters in Chemistry", eds. P. Braunstein, L. A. Oro, P. R. Raithby, Wiley-VCH, Weinheim, 1999, Vol 1, p. 561. 9 C. A. Reed, R. D. Bolskar, Chem. Rev., 2000,100,1075. 10 P. D. W. Boyd, P. Bhyrappa, P. Paul, J. Stinchcombe, R. D. Bolskar, Y. Sun, C. A. Reed, J. Am. Chem. Soc., 1995,117,2907. 11 A. Ceriotti, F. Demartin, G. Longoni, M. Manassero, M. Marchionna, G. Piva, M. Sansoni, Angew. Chem. Int. Ed. Engl., 1985,24,696. 12 C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, J. Wolowska,Angew. Chem. Int. Ed. Engl.,2000,39, 1635 . 13 J. M. Bemis, L. F. Dah1,J. Am. Chem. SOC., 1997,119,4545. 14 M. Kawano, J. W. Bacon, C. F. Campana, B. E. Winger, J. D. Dudek, S. A. Sirchio, S. L. Skruggs, U. Geiser, L. F. Dahl, Inorg. Chem., 2001,40,2554.

194

Perspectives in Organometallic Chemistry

15 N. T. Tran, M. Kawano, L. F. Dahl, J. Chem. SOC.Dalton Trans., 2001,2731. 16 N. T. Tran, M. Kawano, D. R. Powell, L. F. Dahl, Angew. Chem. Int. Ed. Engl., 2000,39, 4121. 17 F. Fabrizi de Biani, C. Femoni, M. C. Iapalucci, G. Longoni, P. Zanello, A. Ceriotti, Inorg. Chem., 1999,38,3721. 18 G. Longoni, C. Femoni, M. C. Iapalucci, P. Zanello, in "MetalClusters in Chemistry",eds. P. Braunstein, L. A. Oro, P.Raithby, Wiley-VCH, Weinheim, 1999, Vol2, p. 1137. 19 D. M. P. Mingos, A. S. May, in "The Chemistry of Metal Cluster Complexes': eds. D. F. Shriver, H. D. Kaesz, R. D. Adams, VCH Publishers, New York, 1990, p. 1 1 . 20 J. C. Calabrese, L. F. Dahl, A. Cavalieri, P. Chini, G. Longoni, S. Martinengo, J. Am. Chem. SOC.,1974,96,2616. 21 J. C. Calabrese, L. F. Dahl, P. Chini, G. Longoni, S. Martinengo, J. Am. Chem. SOC.,1974, 96,26 14. 22 G. Longoni, M. Manassero, M. Sansoni, J. Am. Chem. SOC.,1980,102,7973. 23 D. F. Rieck, R. A. Montag, T. S. McKechnie, L. F. Dahl, J. Am. Chem. SOC.,1986, 108, 1330. 24 R. E. Des Enfants, J. A. Gavney, R. K. Hayashi, A. D. Rae, L. F. Dahl, J. Organomet. Chem.,l990,383, 543. 25 D. F. Rieck, J. A. Gavney, R. L. Norman, R. K. Hayashi, L. F. Dahl, J. Am. Chem. SOC., 1992,114,10369. 26 P. D. Mlynek, L. F. Dahl, Organometallics, 1997,16, 1656. 27 P. D. Mlynek, L. F. Dahl, Organometallics, 1997, 16, 1641. 28 A. J. Kahaian, J. B. Thoden, L. F. Dahl, J. Chem. SOC.Chem. Comm., 1992,353. 29 J. P. Zebrowski, R. K. Hayashi, L. F. Dahl, J.Am. Chem. SOC.,1993, 115, 1142. 30 V. G. Albano, F. Demartin, M. C. Iapalucci, G. Longoni, A. Sironi, V. Zanotti, J. Chem. SOC.Chem. Comm., 1990,547. 31 V. G. Albano, F. Demartin, M. C. Iapalucci, F. Laschi, G. Longoni, A. Sironi, P. Zanello, J. Chem. SOC.,Dalton Trans., 1 99 1,739. 32 J. P. Zebrowski, R. K. Hayashi, L. F. Dahl, J. Am. Chem. SOC.,1993,115, 1142. 33 V. G. Albano, F. Demartin, M. C. Iapalucci, G. Longoni, M. Monari, P. Zanello, J. Chem. SOC.,Dalton Trans., 1992,497. 34 D. M. P. Mingos, 2. Lin, J. Chem. SOC.Dalton Trans., 1988, 1657. 35 R. Gautier, J-F. Halet, J-Y. Saillard, in "MetalClusters in Chemistry", eds. P. Braunstein, L. A. Oro, P. R. Raithby, Wiley-VCH, Weinheim, 1999, Vol3, p. 1643. 36 S. Martinengo, G. Ciani, A. Sironi, P. Chini, J. Am. Chem. SOC.,1978,100,7096. 37 G. Ciani, A. Sironi, S. Martinengo, J. Chem. SOC.,Dalton Trans. 1982, 1099. 38 J. L. Vidal, L. A. Kapicak, J. M. Troup, J. Orgunomet. Chem. 1981,215, C l l . 39 D. Collini, C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, P. Zanello, submitted for publication; abstr. XX ICOMC 40 D. Collini, C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, submitted for publication to Inorg. Chim. Acta. 41 A. Ceriotti, N. Masciocchi, P. Macchi, G. Longoni, Angew. Chem. Int. Ed. Engl., 1999,38, 3724. 42 C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, Chem. Comm., 2001,1776. 43 P. Zanello, in "Stereochemistry of Organometallic and Inorganic Compounds", ed. P, Zanello, Elsevier, Amsterdam, 1994,5, 163.

Modulation of electronic behaviour of metal-carbonyl clusters

195

44 J. D. Roth, G. J. Lewis, L. K. Safford, X. Jiang, L. F. Dahl, M. J. Weaver, J. Am. Chem. SOC.,1992,114,6159. 45 F.Demartin, C. Femoni, M. C. Iapalucci, G. Longoni, P. Zanello, J. Cluster Sci., 2001,12, 61 46 D. Collini, C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, submitted for publication; abstr. XX ICOMC. 47 C. Femoni, M. C. Iapalucci, G. Longoni, P. H. Svensson, J. Wolowska, Angew. Chern. Int. Ed. Engl., 2000,39,1635. 48 M.ficco, work in progress 49 F. Calderoni, F. Demartin, M. C. Iapalucci, G. Longoni, Angew. Chem. Int. Ed. Engl., 1996,35,2225. 50 F. Calderoni, F.Demartin, F. Fabrizi de Biani, C. Femoni, M. C. Iapalucci, G. Longoni, P. Zanello, Eur. J. Inorg. Chem., 1999,663. 51 C. Mealli, D. M. Proserpio, J. Chem. Ed. 1990,67,399. 52 Q.Xie, E.Perez-Cordero, L. Echegoyen, J.Am. Chem. SOC.,1992,114,3977. 53 R. S.Ingram, M. J. Hostetler, R. W. Murray, T. G. Schaaff, J. T. Khouxy, R. L. Whetten, T. P. Bigioni, D. K. Guthrie, P. N. First, J. Am. Chem. SOC.,1997,119,9279. 54 R.Rathore, C.L. Bums, M. I. Deselnicu, Org. Lett., 2001,3,2887. 55 C. Lambed, G.Noll, F. Hampe1,J. Phys. Chem., A 2001,105,7751. 56 P. P. Edwards, in "Advances in Physical Metallurgy", Eds J. A. Charles, G. C. Smith, London, Institute of Metals, 1990,93. 57 R. L.Johnston, Phil. Trans. R. SOC.Lond. A, 1998,356,211. 58 P. P. Edwards, R. L. Johnston, C. N. R. Rao, in "Metal Clusters in Chemistry", eds. P. Braunstein, L. A. Oro, P. R. Raithby, Wiley-VCH, Weinheim, 1999,Vol3, p. 1454. 59 K. J. Taylor, C. L. Pettiette-Hall, 0. Cheshnovsky, R. E. Smalley, J. Chem. Phys., 1992,96, 3319 60 A. Sandell, J. Libuda, P. A. Briihwiler, S. Andersson, M. Baumer, A. J. Maxwell, N. artensson, H.-J. Freund, J. Phys. Rev. B, 1997,55,7233. 61 T. Fujimoto, A. Fukuoka, J. Nakamura, M. Ichikawa, J. Chem. SOC. Chem. Commun., 1989,845. 62 T.Sawitowski, S.Franzka, N. Beyer, M. Levering, G. Schmid, Adv. Funct. Mater., 2001, 11, 169. 63 G.Schmid, M.Baumle, N. Beyer, Angew. Chem., Int. Ed., 2000,39,181. 64 S.Hoeppener, L.Chi, H. Fuchs, Nan0 Lett., 2002,2,459. 65 G. Schmid, Y-P. Liu, M. Schumann, T. Raschke, C. Radehaus, Nan0 Lett., 2001,1,405. 66 B. T. Heaton, P. Ingallina, R. Devenish, C. J. Humphries, A. Ceriotti, G. Longoni, M. Marchionna, J. Chem. SOC.Chem. Comm., 1987,765. 67 P.Cassoux, J. S. Miller, in "Chemistry of Advanced Materials", Eds. L. V. Interrante, M. J. Hampden-Smith, Wiley-VCH, New York, 1998,19.

INTERIONIC AND INTERMOLECULAR SOLUTION STRUCTURE OF TRANSITION METAL COMPLEXES BY NMR

A. Macchioni Department of Chemistry, University of Perugia, Via Elce di Sotto, 8 - 06123 Perugia, Italy

1 INTRODUCTION It is well known that non-covalent interactions' are of fundamental importance in many areas of chemistry. In particular, the structure and reactivity of transition metal complexes are profoundly affected by weak intermolecular interactions, but while most of the structural information is usually obtained in the solid state, changes in reactivity, induced by non-covalent interactions, are more often observed in solution. This could lead to the risky procedure of interpreting the effect of intermolecular interactions on the chemical reactivity based on structural data in the solid state that, due to the little amount of energy that is at stake, could be different from that in solution and could, consequently, lead to wrong conclusions. On the other hand, the determination of intermolecular structural parameters in solution is a challenge that can be carried out with some possibilities of success only using NMR techniques. Two main types of information are crucial for investigating intermolecular interactions in solution: (1) the relative orientation of the interacting moieties, and (2) their degree of aggregation. In order to obtain information (1) the most suitable NMR techniques are those based on NOE (Nuclear Overhauser Effect)' that acts directly through space and, consequently, indicates the spatial proximity of the interacting nuclei independently of their belonging to intramolecular or intermolecular3 moieties. PGSE (Pulsed-Field Gradient Spin Echo) NMR experiments4provide information (2) allowing the determination of the translational diffusion coefficient that is strictly connected to the hydrodynamic volume of the difhsing particles. The detection of an NOE, i. e. of an intensity variation of the resonance of I-nucleus by perturbing the resonance due to S dipolarly coupled to it, qualitatively indicates that the distance between I and S is less than about 5 A. By performing a quantitative analysis of the kinetics of NOE build up, internuclear distances can be measured if a reference distance between two other nuclei (A and B) is known. A and B necessarily belong to the same moiety and, consequently, rm is an intramolecular reference distance. On the other hand, I and S are found on different moieties and rIs is an intermolecular distance. In order to have accurate intermolecular distances, the rotational correlation times of A-B and I-S vectors in solution must be estimated. In the case of PGSE experiments, the evolution of the resonance intensities as a function of the intensity of pulsed-field gradients is monitored. There is a linear dependence of ln(I/Io)

Interionic and intermolecular solution structure of transition metal complexes

197

on G2 (where I = resonance intensity, 10 = resonance intensity without gradient and G = gradient strength) and the slope is directly proportional to -Dt (diffusion coefficient). The latter is inversely proportional to the hydrodynamic radius of the diffusing particles. As a consequence, when X has a lower absolute value of slope than Y this simply means that X diffuses more slowly than Y and, consequently, is larger than Y. This is true if X and Y have similar shapes and are both larger or smaller than the solvent molecules. Furthermore, attention must be paid when comparisons of Dt values relative to different solutions are made because the diffusion coefficients depend on the viscosity of the solution. In the last few years we have studied the intermolecular structure in solution of several transition metal complexes taking advantage of the complementarity of information of the two above-mentioned NMR techniques. Most of our studies were dedicated to the 1D- and 2D-NOE, homo- and hetero-nuclear, qualitative investigation of transition metal complex ion pairs.5 On the other hand, we showed that the quantification of interionic NOES can be successfdly carried out with the consequent possibility of estimating average interionic distances. The aggregation of transition metal complex ion pairs to ion quadruples and higher aggregates was also investigated. Finally, very recently we started to explore the aggregation of neutral transition metal complex units. The general requirements for the systems to be investigated are (1) the presence of NMR active nuclei in both molecular moieties, (2) the presence of magnetically non-equivalent nuclei, which spatially disperse around the molecular moieties, making it possible to discriminate different relative orientations of the moieties, and (3) the presence in solution of a high percentage of intimate adducts. The latter condition is particularly difficult to respect for neutral intermolecular adducts. The main results and conclusions of our studies are here summarized with particular attention to systems in which the intermolecular interactions affect the structure andor the reactivity of transition metal complexes.

2 TRANSITION METAL COMPLEX ION PAIRS 2.1 Interionic Structure by NOE NMR Studies

The investigation of the compounds illustrated in Scheme 1 , in solvents (methylene chloride or chloroform) and at concentrations (1 O-* - lo-' M) that ensure the predominance of intimate ion pairs, allowed the identification of some general features concerning the interionic structure of transition metal complex ion pairs: (1) in most cases a well defined relative anion-cation orientation is present in solution; (2) the anion approaches the complexes from the side of N,X-ligands due to an accumulation of positive charge at the junction of the moieties containing the N- and X-donor atoms; (3) in square planar complexes, the apparent favored apical positions for the anion often are protected due to electronic or to a mix of electronic and steric factors; furthermore, the anion accessibility to the metal center can be evaluated and appears to be finely modulated by the steric hindrance in the apical positions. The anion has a high tendency to locate close to the N,X-11 and independent of its nature and that of the solvent both in the case of unsaturated (1: 2,.g4-10) and saturated (3)' N,Xligands. The accumulation of positive charge at the junction of the moieties containing the N- and X-donor atoms was confirmed by means of quantum mechanical calculations on 16' and 49 compounds. In the latter compounds the anion prefers to locate close to the junction

Perspectives in Organometallic Chemistry

198 OCTHAEDRAL COMPLEXES

1

2

M = Fe, Ruor 0s;R = H orpz Y = PMe3, CO or pz A- = BPh4-. OTf, BF4-or PF6-

M = Fe or R u R = Me, Ph or pz

H 4 A- = B P h j , OTf, BF4' or PF6' L = PPh3 or PPh2Me

SQUARE PLANAR COMPLEXES

5

6

7

8

M = Pd or R; A- = BPh4-, OTf, BF4-, PFS or BArF-; R = H,Me, Et or LPr; R'2 = Me2 or An; R = Me or Ph

9

10

R = H,Me, Et or LPr; R 2 = H2 or Me2; R = H, Me, or CO2Me

Scheme 1

11

Interionic and intermolecular solution structure of transition metal complexes

199

of the pyridine rings despite the fact that, on the side of the hydride, it could have the possibility to stay much closer to the metal center. In square planar complexes the apparent favored positions above and below the plane are in reality occupied by the anion only in the cases in which there is no steric hindrance in such positions or where there is no accumulation of positive charge in a particular position of the li ands. As a matter of fact, we found the anion in the apical positions only for compounds 5Fb even if, also in this case, the anion was slightly shifted toward the bipy ligand and, in particular, it preferred to stay close to the N-pyridine ring trans to the Pd-C 0 bond (Figure 1).

8.0

3 3

8.5

Figure 1 Section of the l9F,lH-H0ESY NMR (376.65 MHz, 21 7 K, methylene chloride-d2) spectrum showing that the interionic contact between 6' proton and the anion is weaker than that with 6proton (the FI trace relative to the " B F i fluorine resonance is reported on the right) Clear electronic protection of the apical positions was encountered in investigating compounds 611 and lo.';! For such complexes the anion specifically interacts with C-H, CH;! or N-H and the protons that are close to them. A mix of electronic and steric protection was observed for complexes 713 and 914 in which a synergistic action of the accumulation of positive charge on the diimine carbons and the axial steric protection of R groups denies the access of the anion to the metal center. The only observed interionic interactions in the 19F,'H-HOESYNMR spectra, for large R groups, are between A' and R and R' protons. The presence of an accumulation of positive charge is of fundamental importance in order to have a specific anion-cation relative orientation. In fact, in complexes 7 (R'2 = An (9-Anthr~l))'~ and 11,15where the accumulation of positive charge is difficult to reach for the anion, the interionic interactions are less specific. In the case of complexes 8l6 bearing hemi-hindered a-iminoketone N,O-ligands the anion, remaining on the side of this ligand, has the possibility to choose between the least crowded side close to the 0-arm or the favored position close to the imine carbon. The observed interionic

200

Perspectives in Organometallic Chemistry

structure suggests that both the positions are present in solution with almost the same abundance. Complexes 5, in which M = Pd, are active catalysts for the CO/styrene alternating copolymerization carried out in mild condition^'^ and the A- counteranion strongly affects the catalytic performance. Higher catalytic efficiency is obtained by using less coordinating anions, i. e. those that were found to show weakest interionic interactions in ‘H-NOESY or 19 1 F, H-HOESY low temperature NMR spectra. The observation of the same anion-cation orientation in solution by changing the anion features (size and electrondensity distribution) and the correlation between the catalytic performances of the complexes with the non-coordinating power of the anion, clearly indicate that the anion competes with the substrates for the coordination to the metal. Interestingly, we found that the interionic solid state structure is different from that observed in solution. In fact, due to the stacking between two cation moieties, in the solid state the counteranions are located sideways running around the coordination plane. In the case of complexes 7 and 9, bearing a-diimine ligands of general formula (2,6-(R)2C6H3)N=C(R7)-C(R’)=N(2,6-(R)2-C6H3), the accessibility of the counteranion to the metal center, as a function of R groups, was investigated by looking at the B F i - H interionic interactions in the 19F,’H-HOESYNMR spectra. It was found that the anion position is very sensitive to the steric hindrance of such groups: only when R is small (R < Et for 9 having R ’ = Me; R < i-Pr for 9 having R” = H; R < Me for 7), besides giving strong contacts with R’ and R protons, the anion also interacts with protons that belong to q ,q25-methoxycyclooctenyl (7) or olefin and Pt-Me (9) ligands. Such results are strictly connected with the performances of compounds bearing a-diimine ligands as catalysts for the olefin polymerizations and CO/olefins co-polymerizations where the amount of steric hindrance in the apical positions is of fundamental importance. In the polymerization of olefins, the presence of encumbered axial positions is to be hoped for in that it slows the chain transfer rate relative to propagation resulting in the formation of polymers with higher molecular weight. l9 Instead, in the copolymerization reactions it inhibits the coordination of CO in the apical position that seems to be the first step of the copolymerization reaction, and reduces the catalytic performances.20The direct knowledge of the steric hindrance necessary to completely protect the axial positions, obtainable by studying the interionic structure, can be usehl to design new catalysts in both cases. Complexes trans-[Ru(COMe) { (Pz~)CH~>(CO)(PM~~)~]BP~~R (R = Me, n-Bu, n-Hex or Ph; pz = pyrazol- 1-yl-ring), bearing unsymmetrical counteranions, were investigated in order to understand whether, besides locating on the side of the N,N-ligand the anion has a preferential orientation. Indeed, already from qualitative ‘H-NOESY NMR studies, we found that the anion orients two phenyl rings almost parallel to the two pz-rings and the aliphatic chain (R = n-Bu or n-Hex) distant from the metal center.21When R = Me the qualitative approach did not give a definitive answer in that both B-Me and ortho protons interact with the CH2 and H5 protons of the cationic moiety. We decided to quantifj the NOES by recording the kinetics of NOE build up by inverting several target resonances.22 An example of intramolecular and interionic NOE build up is showed in Figure 2. By taking the OH-mH distance (2.47 A) as reference and after having measured both the interionic and intramolecular correlation average interionic distances were estimated. From such distances it could be concluded that also in the case in which R = Me the aliphatic chain points away from the metal center. In fact, the distance between CH2 and OH protons resulted about 0.7 A longer (see Figure 2) than that relative to CH2/B-Me protons in perfect agreement with what expected from molecular models. The observed counteranion preference (A- = BPh3R- with R = Me, n-Bu or n-Hex) of orienting two

Interionic and intermolecular solution structure of transition metal complexes

201

phenyl groups almost parallel to the two pyrazolyl-rings and the B-Me group far from the metal center seems to be a consequence of a maximization of the lipophilic interactions that should produce a gain in energy of about 4 kJ/m01.~~ 6NOE(%) .

A

54-

3-

HJ

0

2

4

6

8

1

0

1

w

2.47 A

2

Figure 2 Experimental data relative to the %NOE as a function of zfor the irradiation of oHprotons of the BPhi (T = 302 K) for complex in which R = Ph (Reproducedfiom ref. 22 with the permission of the American Chemical Society) A recent 19F,'H-HOESYNMR in~estigation~~ of compounds 9 in which R" = Ph, 4-Me-Ph or 4-CF3-Ph, R = i-Pr and R' = Me led to the observation, for the first time, of the equilibrium between two anion cation relative orientations illustrated in Scheme 2 in which the anion stays in Pseudo-Trans (PT) or Pseudo-Cis (PC) position with respect to the styrene substituent. Ar

'X Pseudo-Trans (PT)

Pseudo-Cis (PC)

Scheme 2

The anion locates again close to the a-diimine carbons but, in this case, the two positions above and below the plane are not equally populated. From a quantitative analysis of the heteronuclear F-H NOES it was found that the PC ion pair is mainly present in solution. Furthermore, the equilibrium constants and the consequent variation of fiee energy could be estimated. Selected data are reported in Table 1. The preference for the PC ion pair was explained with the help of quantum mechanical calculations and seems to be due to the slight polarization of the styrene moiety that lead to a partial accumulation of positive charge, higher for X = CF3, in its ortholmetha positions that can be appreciated by the anion also being found on the side of N,N-ligand in the PC position.

202

Perspectives in Organometallic Chemistry Table 1. PC and PT abundance, equilibrium constants and AGO (kJ mof') in CDC13 estimated by the quantification of interionic heteronuclear NO€

Ph

4-Me-Ph

4- CF3-Ph

277K

298K

277K 81(4)%

PC ion pair

71(4)%

72(4)%

PT ion pair

29(4)%

28(4)%

K

2.4 f 0.5

2.6 f 0.5

4.3 f 1.I

AGO

-2.0 f 0.4

-2.4 f 0.5

-3.4 f 0.9

19(4)%

In the case of the compound having R ' = 4-CF3-Ph, due to the presence of the intramolecular reference distance between CF3 and the meta protons (3 .O A), interionic distances were estimated. They confirmed that the anion found on the side of the N,Nligand is able to perceive the ortho protons of the styrene moiety (average distance between B F i and o-H = 4.4 A). A system in which the ion-pairing phenomenon directly affects the structure of the compounds was recently found25 by studying the interionic structure of hemi-hindered palladium(I1) complexes, shown in Scheme 3. Depending on the relative orientations of the R substituents (Me, Et, i-Pr or CF3), the four conformational isomers reported in Scheme 3 could exist. All the four isomers where observed in solution, at low temperature, in percentages different from case to case, with isomer c always predominating. In our opinion, the predominance in solution of c with respect to isomers a and b is determined by the possibility for the anion to closely approach the diimine carbons, that in c (and d) are not protected by the R substituents, maximizing the favorable ion pair interactions. Isomer c is favored compared to d (and a) by the minimization of the steric repulsion between R and OMe groups. As a confirmation, from a mixture of isomers a-d with R = iPr, suitable crystals for X-ray studies were obtained that were found to be due to isomer c; close interionic contacts between the anion and diimine Me groups and isopropyl protons that point toward them were observed. R

a

b

n

d

C

Scheme 3

By considering that (1) the hemi-hindered complexes shown in Scheme 3 are active catalysts for the COlp-Me-Styrene copolymerization carried out under mild conditions (room temperature and PCO= 1 atm) and (2) the stereochemistry of the product copolymers depends on the relative orientation of R groups, this could be one of the few examples26in

Interionic and intermolecular solution structure of transition metal complexes

203

which the anion-cation interactions in the catalysts are directly related to the stereochemistry of the catalytic product. 2.2 Aggregation by PGSE N M R Studies Several ionic organometallic compounds are active catalysts for reactions carried out in solvents with low dielectric constants. For this reason, it is of fundamental importance to understand whether the species that really act as catalysts are free ions, ion pairs, ion quadruples, etc. Such information can be obtained by performing PGSE NMR experiments that have been known for several decades but their application to organometallic compounds is still limited.27 By carrying out PGSE NMR studies on compounds 1:* in which M = Ru, we demonstrated that the type of ionic species present in solution is strongly affected by the solvent and concentration. As an example, ion quadruples are the predominant species in saturated solutions of trans-[Ru(COMe){( P z ~ ) C H ~ } ( C O ) ( P M ~ ~(A) ) ~ ]in B chloroform-d P~~ while free ions are mainly present in nitromethane-d3 solutions (Figure 3). These results can be deduced by comparing the slopes of the straight lines reported in Figure 3 relative to compound A and to trans-[Ru(COMe){(pz2)BH2}(CO)(PMe3)2] (B) that is isosteric and almost isomass with the cation moiety of A (A'). B was used as an internal reference in order to be sure that the results were not affected by a change of the viscosity of the solutions. In the case a of Figure 3 (chloroform-d), the ratio between the slopes (that corresponds to the ratio of the diffusion coefficients DB/DA)is equal to 1.6. By assuming A and B to be spherical, the volume ratio VANBis equal to 4.1 and, due to the fact that Aand A' have almost the same volume, this ratio indicates the main presence in solution of ion quadruples. On the other hand, the ratio of the slopes in case b (nitromethane-d3)is 1.1 with a consequent volume ratio of 1.4 that can be explained with the main presence of free ions that have a volume a little higher than B due to the solvent that they drag during the translation. In between these two extreme cases, in which ion quadruples (a) or free ions (b) predominate, there are a continuous situations corresponding to the contemporary presence of ions, ion pairs and ion quadruples. '7

'7

-5-

'7:

0.0

a 0.1

0.2

0.3

0.4

@(a.u .)

0.0

0.1

0.2

0.3 G7a.u.)

0.4

Figure 3 Plot of ln(II0) versus G2 (a. u. arbitrary units)for the PMe3 resonances ofA' and B and for the mH resonance of A- for the PGSE experiments carried out in (a) saturated solution of A in chloroform-d and (h) nitromethane-d3 (4 , I 0-3ly, (Reproducedfromref 28 with the permission of the American Chemical Society)

Perspectives in Organometallic Chemistry

204

3 INTERMOLECULAR ADDUCTS As mentioned in the introduction the rather obvious difference between the investigation of neutral intermolecular adducts compared to ion pairs is the difficulty of having a reasonable percentage of intimate adducts in solution. The latter condition is fundamental especially if we are interested in the relative orientation of the interacting units. We have started our investigations on intermolecular interactions by considering the possible stacking in solution of trinuclear cyclic basic Au' compounds [Au(p-C2,N3bzim)]3 (bzim=l -benzylimidazolate) (17) or [Au(p-C,N-C(OEt)=N-CbHKH3)]3 (18) with the trinuclear H$' acid complex [Hg(p-C,C-CsF4)]3 (19) (Scheme 4).

18

I9

Scheme 4 only indirect and While it is well known that similar compounds stack in the solid uncertain indications of the stacking process in solution were reported. By combining the results coming from l9F,'H-H0ESY and PGSE NMR experiments we demonstrated that the stacking process is also present in s~lution.~' F3

F2

-120

-140

ppm

-160 19F

Figure 4 Section of a I 9 F , 'H-HOESY NMR spectrum of 17/19 adduct (9.6 mM in THF-& at 376.63 MHz. Intermolecular NOES are shown between F2 and the CH2, H4, H5 and o-H

Interionic and intermolecular solution structure of transition metal complexes

205

protons; and between F3 and the CH2, H4, and H5 protons. The 1D-traces relative to the F2 column and CH2 row are shown on the right and top of the section, respectively (Reproducedfiom ref 33 with the permission of the American Chemical Society)

The 19F,‘H-HOESY NMR measurements showed the presence of intermolecular cross peaks for solutions of both 17/19 and 18/19 adducts (Figure 4) indicating that a reasonable percentage of intimate adducts are present in solution. On the other hand, PGSE measurements allowed to quantify the size of the adducts: for saturated solutions of 17 and 19 complexes an equilibrium between “Au3Hg3” and “Au3Hg3Au3” adducts is present while for 18 and 19 complexes, free molecules and “Au3Hg3” adduct are contemporary present in solution. Finally, we have recently found that intermolecular adducts can even derive from the aggregation of cationic moieties. This was met by investigating the palladium(1) dinuclear derivatives (20) shown in Scheme 6 in which A- = CH3COO-, PFi, B(3,5(CF3)2-CsH3)~3l

Scheme 6

These and analogous compounds, in which the N,N-bidentate ligand is substituted by P,Pligands,32 often form during the carbonylation processes catalyzed by palladium(I1) complexes carried out in protic solvents. The complexes bearing N,N-ligands are usually very insoluble and amorphous and, consequently, hard to fully characterize. To the best of our knowledge only in one case (N,N = bipy) was it possible to obtain a crystal structure probably due to the presence of HOs4(C0)1; as a ~ounteranion.~~ The compound shown in Scheme 6 is soluble in some solvents (acetone, THF, nitrobenzene, etc.. .) only when A- = B(3,5(CF3)2-C6H3)4‘. PGSE measurements were carried out in nitrobenzene-ds and aCetOne-ds as a function of the concentration using (Phen)PdMeCl as standard substance. The ratio between the volume of the adducts coming from the aggregation (Vaggr)of the dinuclear species and that of the reference compound (Vsmd) was determined. It was found that 20 has a remarkable tendency to aggregate: the VaggrNstand ratio goes from 8 at the lowest concentration used (0.4mM) up to 41 for the highest concentration (55.2 mM). In the meantime, the anion does not seem to participate in the aggregation process in that its apparent volume only increases slightly enhancing the concentration of 20. 3 CONCLUSION The selected results reported here indicate that the relative orientation and the aggregation of non-covalently bonded transition metal moieties can be successfully investigated in solution by means of NOE and PGSE NMR experiments, respectively. The deduced information assumes a relevant meaning in several areas of chemistry, ranging from supramolecular coordination chemistry to homogeneous metallorganic catalysis, especially because it refers to the environment where non-covalent interactions affect the chemical

206

Perspectives in Organometallic Chemistry

reactivity more heavily, i. e. in solution. The increasing number of examples in which the reaction products of either stoichiometric or catalytic reactions are determined, in terms of chemo-, regio- or stereo-selectivity, by weak non-covalent interactions seems to suggest that such solution information will have to be taken into account to a greater extent in the future. Acknowledgments

I would like to thank all the co-workers who participated in obtaining the results here reported. This study was supported by grants fiom the “Minister0 dell’Istruzione, dell’universith e della Ricerca (MIUR, Rome, Italy), Programma di Rilevante Interesse Nazionale, Cofinanziamento2000-200 1”. References J. Israelachvili, Intermolecular & Surface Forces, Academic Press, London, 1992. D. Neuhaus and M. Williamson, The Nuclear Overhauser Eflect in Structural and Conformational Analysis, VCH, Weinheim, 1989. H. Mo and T. C. Pochapsky, Prog. NMR Spectrosc., 1997,30,1. P. Stilbs, Prog. Nucl. Magn. Res. Spectrosc., 1987, 19, 1. C. S. Johnson Jr., Prog. Nucl. Magn. Res. Spectrosc., 1999,34,203. W. S . Price, Concepts Magn. Res., 1997,9,299. W. S. Price, Concepts Magn. Res., 1998,10, 197. A. Macchioni, Eur. J. Inorg. Chem., submitted for publication. (a) G. Bellachioma, G. Cardaci, A. Macchioni, G. Reichenbach and S . Terenzi, Organometallics, 1996, 15, 4349. (b) A. Macchioni, G. Bellachioma, G. Cardaci, V. Gramlich, H. Riiegger, S. Terenzi and L. M. Venanzi, Organometallics, 1997, 16, 2139. (c) A. Macchioni, G. Bellachioma, G. Cardaci, G. Cruciani, E. Foresti, P. Sabatino and C. Zuccaccia, Organometallics, 1998, 17, 5549. (d) G. Bellachioma, G. Cardaci, A. Macchioni, F. Valentini, C. Zuccaccia, E. Foresti and P. Sabatino, Organometallics, 2000, 19,4320. G. Bellachioma, G. Cardaci, V. Gramlich, A. Macchioni, M. Valentini and C. Zuccaccia, Organometallics, 1998,17, 5025. G. Bellachioma, G. Cardaci, F. D’Onofrio, A. Macchioni, S. Sabatini and C. Zuccaccia, Eur. J. Inorg. Chem., 2001, 1605. A. Macchioni, C. Zuccaccia, E. Clot, K. Gruet and R. H. Crabtree, Organometallics, 2001,20,2367. 10 A. Macchioni, G. Bellachioma, G. Cardaci, M. Travaglia, C. Zuccaccia, B. Milani, G. Corso, E. Zangrando, G. Mestroni, C. Carfagna and M. Formica, Organometallics, 1999,18,3061. 11 B. Binotti, G. Bellachioma, G. Cardaci, A. Macchioni, C. Zuccaccia and E. Foresti, P. Sabatino, Organometallics,2002,21,346. 12 R. Romeo, N. Nastasi, L. Monsu Scolaro, M. R. Plutino, A. Albinati and A. Macchioni, Inorg. Chem., 1998,37,5460. 13 G. Bellachioma, B. Binotti, G. Cardaci, C. Carfagna, A. Macchioni, S. Sabatini and C . Zuccaccia, Inorg. Chim. Acta, 2002,330,44, 14 C . Zuccaccia, A. Macchioni, I. Orabona and F. Ruffo, Organometallics, 1999, 18, 4367. 15 R. Romeo, L. Fenech, L. Monsu Scolaro, A. Albinati, A. Macchioni, C. Zuccaccia, Inorg. Chem., 200 1,40,3293.

Interionic and intermolecular solution structure of transition metal complexes

207

16 A. Macchioni, C. Zuccaccia, B. Binotti, C. Carfagna, E. Foresti and P. Sabatino, Inorg. Chem. Comm., 2002,5,3 19. 17 B. Milani, G. Mestroni, A. Sommazzi and F. Garbassi, Italian Patent N. MI 95/A 000337,1995; European Patent N. 96101967.6-2102,1996. 18 A. Macchioni, G. Bellachioma, G. Cardaci, M. Travaglia, C. Zuccaccia, B. Milani, G. Corso, E. Zangrando, G. Mestroni, C. Carfagna and M. Formica, Orgunometullics, 1999,18,3061. 19 L. K. Johnson, C. M. Killian and M. Brookhart, J Am. Chem. SOC., 1995,117,6414. D. J. Tempel, L. K. Johnson, R. L. Huff, P. S. White and M. Brookhart, J Am. Chem. Soc., 2000, 122, 6686. L. Deng, T. K. Woo, L. Cavallo, P. M. Margl and T. Ziegler, 3: Am. Chem. SOC., 1997,119,6177. R. D. J. Froese, D. G. Musaev and K. Morokuma, J. Am. Chem. SOC.,1998,120, 1581. 20 C. Bianchini and A. Meli, Coord. Chem. Rev., 2002,225,35. 21 C. Zuccaccia, G. Bellachioma, G. Cardaci and A. Macchioni, Orgunometullics, 1999, 18, 1. 22 C . Zuccaccia, G. Bellachioma, G. Cardaci and A. Macchioni, J Am. Chem. SOC., 200 1, 123, 11020-1 1028. 23 H.-J. Schneider, T. Schiestel, P. Zimmermann, J Am. Chem. SOC., 1992, 114,7698. H.J. Schneider, Chem. SOC. Rev., 1994,227. 24 A. Macchioni, A. Magistrato, I. Orabona, F. Ruffo and Ursula Rothlisberger, C. Zuccaccia, unpublished results. 25 B. Binotti, G. Bellachioma, C. Carfagna, G. Cardaci, A. Macchioni and C. Zuccaccia, unpublished results. 26 A. Lightfoot, P. Schnider and A. Pfdtz, Angew. Chem. Int. Ed. Engl., 1998,37, 2897. G. Desimoni, G. Faita, A. Martoni and P. Righetti, Tetrahedron Lett., 1999, 40, 2001. M.-C. Chen and T. J. Marks, J Am. Chem. SOC., 2001,123,11803. 27 M. Valentini, H. Riiegger and P. S . Pregosin, Helv. Chim. Actu, 2001, 84, 2833 and references therein. 28 C. Zuccaccia, G. Bellachioma, G. Cardaci and A. Macchioni, Orgunometallics, 2000, 19,4663. 29 A. Burini, J. P. Fackler Jr., R. Galassi, T. A. Grant, M. A. Omary, M. RawashdehOmary, B. R. Pietroni, R. J. Staples, J. Am. Chem. SOC., 2000, 122, 11264. A. Burini, R. Bravi, J. P. Fackler Jr., R. Galassi, T. A. Grant, M. A. Omary, B. R. Pietroni, R. J. Staples, Inorg. Chem., 2000,39,3 158 and references therein. 30 A. Burini, J. P. Fackler Jr., R. Galassi, A. Macchioni, M. A. Omary, M. A. RawashdehOmary, B. R. Pietroni, S. Sabatini and C. Zuccaccia, J Am. Chem. SOC., 2002, 124, 4570. 3 1 R. Bortolo, G. Guglielmetti, A. Macchioni, C . Querci and C. Zuccaccia, unpublished results. 32 V. V. Grushin, Chem. Rev., 1996,96,2011. 33 S . Chan, S. -M. Lee, Z. Lin, W. -T. Wong,J. Orgunomet. Chem., 1996,510 219.

SYNTHETIC AND MECHANISTIC PATHWAYS IN PLATINUM(I1) CHEMISTRY

R. Romeoa and L. Monsu Scolaroa-b aDipartimento di Chimica Inorganica and ISMN-CNR, Sezione di Messina, Universita di Messina, Salita Sperone 3 1, Vill. S. Agata 98 166 Messina, Italy. Kafxomco 'i2:chcm.unirnc.i t bIstituto Nazionale di Fisica della Materia, INFM, Unita di Messina, Salita Sperone 3 1 , Vill. S. Agata 98 166 Messina, Italy. 1 INTRODUCTION

We have a longstanding interest in the kinetics and reaction mechanisms of fourcoordinate coordination and organometallic square-planar complexes, especially those of platinum(II).'.' Our main efforts were and are directed (i) to intercept and recognise species having coordinatively and electronically unsaturated metal sites (tri-coordinate asymmetric 14-electron species) originated through dissociation of square-planar organometallic platinum(I1) substrates, and (ii) to define the structural properties of these transient species and the role they play as key intermediates in a number of fundamental organometallic processes. This work prompted us to perform a rational design and synthesis of new organometallic species containing very weakly coordinated ligands. The knowledge of their mechanistic pathways was exploited as a useful guide to new synthetic procedures. In the first part of this paper we report on the use of some complexes, particularly [Pt(CH3)Cl(Me2S0)2] or [Pt(CH3)(N-N)(Me2SO)lf(N-N = a diimine), as synthons for accessing to a variety of interesting mononuclear and binuclear compounds or to more complex systems containing metal derivatives of porphyrins. The kinetic part will deal with the following topics: (i) the demonstration that cyclometalation in platinum(I1) complexes, contrary to previous claim^,^,^ does not play any particular role in controlling rates and mechanisms and, (ii) a further insight into the factors that promote dissociative ligand exchange and substitution at square-planar platinum(I1) complexes. 2 INSTRUMENTATION AND METHODS 2.1 Syntheses

All syntheses were performed under a dry, oxygen-free nitrogen atmosphere using standard Schlenk-tube techniques. Solvents were distilled under nitrogen from appropriate drying agents. Reagents were purified through distillation or crystallisation, when necessary. The compounds were characterised through their IR and one- and twodimensional IH, I3C, 31Pand I9F NMR spectra. The interionic structure of some ion-pairs was studied by detecting interionic dipolar contacts in the 'H-NOESY or I9F('H}HOESY NMR spectra, at low temperature. Structural studies in the solid state were performed through X-rays. The complexity of the platinated porphyrin systems required

Synthetic and mechanistic pathways in platinum(II) chemistry

209

the use of a variety of spectroscopic techniques to get: (a) local information (UV/VIS spectroscopy, fluorescence, circular dichroism, NMR or IR) and, (b) network and morphological information (elastic and dynamic light-scattering, resonance lightscattering, electron and optical microscopy). 2.2 Kinetics

Only for very few systems the reaction rates could have been followed through conventional UV/VIS spectroscopy or isotopic exchange. Ligand exchange and substitution reactions of very labile compounds required the use of stopped-flow techniques or of fast dynamic ‘H or 31PNMR techniques, such as magnetisation transfer experiments or line-shape analysis. This last method was applied to the study of the fluxional motion of some chelated dinitrogen ligand. 3 Trans- CHLOROMETHYLBIS(DIMETHYLSULFOXIDE)PLATINUM(II).(1)

3.1 Synthesis and Characterization. About twenty years ago Eaborn et al. reported the synthesis of the complex [Pt(CH3)Cl(Me2S0)2] (l)? Later on we showed that, in order to improve the quantitative yield, a number of precautions must be adopted.6 The original assignment’ of the trans geometry was based on the presence of a single band for the vptcl stretching frequency in the IR spectrum (275 cm-I) and of a simple pattern for the coordinated dimethylsulfoxide in the ‘H NMR spectrum. We were able to grow up single crystals of an unusual cocrystallisation product between an organotin compound and the sulfoxide complex and its trans geometrical configuration has been proven through an X-ray diffraction study.7 Despite the large amount of data available for the structural characterisation and the reactivity of complexes containing two sulfoxide molecules spanning mutual cis positionsY8this seems to be the unique example of crystal structure of an organometallic platinum(I1) complex containing two sulfoxide ligands in trans configuration. Usually trans isomers prefer to interconvert into the more stable cis species. Apart from the significant lengthening of the Pt-C1 bond (2.416 A) in trans to the strong electron o-donor methyl group, nothing of particularly unusual relates to bond lengths and bond angles. Inspection of the Pt-S bond distances of almost fifty compounds having only one S-bonded sulfoxide per metal atom and different trans-activating groups, allowed for a trans-influence order for platinum 0 < N = C1< S << C to be established.

3.2 Behaviour in Solution The overall ‘H NMR pattern of 1 can be strongly dependent on the temperature and the water content in the solvent (CDC13). This is a straightforward indication that more species are in equilibrium. ‘H, 13C and 19’Pt NMR measurements revealed that in chloroform solution complex 1 gives a mixture of four different species, which have been unambiguously identified as the starting complex 1 in equilibrium with c i s [Pf(CH3)Cl(Me2S0)2] and the two corresponding isomeric aqua-species cis and trans[Pt(CH3)C1(Me$30)(H20)].7 19’Pt NMR magnetisation transfer technique allowed the

210

Perspectives in Organometallic Chemistry

Me2(O)S-

re

Me

I

Me2(0)S- Pt -CI

Pt-S(O)Me2

I

I

CI

Me2(0)S-

1""

Pt-OH2

I

S(OW2

- 148

Me2(O)S-

Ye

pt -CI

I

94

CI

OH2

Scheme 1. Results of the 19jPt Magnetisation Transfer Measurements on Complex 1 at 298 K in CDC13 (rates are in s-') determination of the rate of interconversion among the various complexes, showing that the direct trans-cis isomerism between the [Pt(CH3)Cl(Me$30)2] species is negligible and that the geometrical interconversion occurs through a water-catalysed pathway. The significance of the result is not the absolute rate constants but that Me2SO exchange and isomerisation of 1 is strongly catalysed by adventitious water in the solvent. The absolute rates will depend on the quantity of water present. The additional information we get from NMR magnetisation transfer experiments is that the most labile Me2SO group is that in trans to the strong 0 donor methyl group, followed by that in trans to MezSO, while that in trans to chloride is hardly removed.

4 USE OF 1 AS PRECURSOR TO ORGANOAMINE COMPLEXES. 4.1 Reactivity Toward Monodentate Nitrogen Ligands. The reaction of 1 with a large variety of amines or heterocyclic nitrogen bases L, in chloroform solution, leads to the formation of uncharged complexes of the type [Pt(CH3)C1 (Me2SO)(L)], containing four different groups coordinated to the metal Only two out of the three different possible isomers were detected in solution through 'H NMR spectroscopy. According to the well known hi h trans-influence of the methyl group, the values of the coupling constants with '95Pt ( J P t H ) for the protons of dimethylsulfoxide or pyridine coordinated in trans position to the methyl are expected to be much lower than those found when the same ligands are in trans position to a chloride or an amine." Isomer C can be ruled out as the 3 J p t ~coupling constants for the coordinated dimethylsulfoxide are in the range 25-40 Hz, which is well above the usual

K

Scheme 2. Possible Different Isomersfor Reaction of 1 with Nitrogen Bases L.

Synthetic and mechanistic pathways in platinum(II) chemistry

21 1

values observed for a Me2SO ligand trans to a methyl group (= 12 Hz). The remaining two trans(C,N) (A) and cis(C,N) (B) species can be unambiguously identified on the basis of the magnitude of the values observed for the 3Jptbl coupling constants of the protons of Me2SO and, in the case of pyridine or substituded pyridines, of the H6 hydrogen of the aromatic ring. We exploited successfully this method in a number of cases to distinguish between two different pyridine moieties.' ' , I 2 For the trans(C, N) isomers (A), average values of 3 J p t = ~ 36 4 Hz and of 12+ 3 Hz have been observed for the coordinated Me2SO and pyridines ligands, respectively. In the case of the cis(C,N) isomers, these values change to 3 J p t H = 26 f 3 Hz and 3 J p t ~= 42 f 2 Hz due to the mutual change of ligands in trans position to CH3 and Me2SO. NMR experiments have shown that the first reaction product is the trans(C, N) isomer, as a consequence of the very fast removal from 1 of the ligand trans to the methyl group by the nitrogen bases. Only in the case of the most sterically hindered ligands, such as 2methyl quinoline, acridine orange and 2,6-dimethylpyridine, the trans(C,N) isomer is quite stable and the compounds could have been isolated as solid in high yield. The trans kinetic products undergo a geometrical conversion into the more stable cis(C,N) isomers through the intermediacy of fast exchanging aqua-species. The rate of isomerisation and the relative stability of the two isomers depends essentially on the rate of aquation and on the steric congestion imposed by the new L ligand on the metal. Complex 1 is fairly soluble in water and it reacts easily with molecules of biological relevance such as simple amino acids (NO) to yield neutral species of the type [Pt(CH3)(Me2SO)(N-O)].'3

*

4.2 Reactivity Toward Short-Bite Bidentate Nitrogen Ligands The reaction of 1 with a series of chelating diamines or diirnines (N-N) of widely different steric and electronic characteristics yielded monoalkyl square-planar complexes of the type [Pt(N-N)(CH3)(Me2SO)]PF6, which were synthesised and fully characterised as solids and in solution." The series of ligands N-N includes, among others, 1,2diaminoethane (en), N,N,N',N'-tetramethyl- 1,2-diaminoethane (Me4en), pyridine (py), 2(aminomethylpyridine (2-ampy), N,N'-dicyclohexylethylenediimine (Cy2dim), N,N'diisopropylethylenediimine ( i - P ~ d i m )2,2'-dipyridyl , (bpy), 2,2'-dipyridylamine (dpa), 2,2'-dipyridylsulfide (dps), 1,lO-phenanthroline (phen) and substituted phenanthrolines. The assignment of the iminic protons of the chelated diimine ligands coordinated to an asymmetric platinum metal centre in such complexes was facilitated by the presence of largely different coupling constants with the isotopically abundant "'Pt (33%, I = 1/2), the lowest 3 J P l H being that of the iminic proton trans to the methyl group while the corresponding proton trans to Me2SO exhibits a much higher coupling constant. Dimethyl sulfoxide exchange with all the complexes has been studied as a function of ligand concentration by H NMR line-broadening, isotopic labelling and magnetisation transfer experiments in acetone as the solvent. Second-order rate constants were obtained from linear plots of kobs versus [Me2SO] and activation parameters were obtained from exchange experiments carried out at different temperatures. Second-order kinetics and negative entropies of activation indicate an associative mechanism. The lability of dimethylsulfoxide in the complexes depends in a rather unexpected and spectacular way upon the nature of the coordinate N-N ligands, the difference in reactivity between the first ( N-N = Me4en, k2298= (1.1550.1) x 1Oe6 mol-' s-I) and the last (N-N=2,9-dimethyl-l, 10-phenanthroline (dmphen), k?98 = (3.8 1*0.005) 1O4 mol-' s-')

212

Perspectives in Organometallic Chemistry

members of the series being greater than ten orders of magnitude, as result of a wellknown phenomenon of steric retardation (for the first complex) and an unprecedented case of steric acceleration (for the last complex). Other factors of primary importance in controlling the reactivity are (i) the presence of an extensive IT system on the ligand N-N, (ii) the ease with which this x system interacts with non-bonding d electrons of the metal and, (iii) the flexibility and ease of elongation of the chelate bite distance. The basicity plays a somewhat minor role, except in the restricted range of the same class of compounds such as substituted phenanthrolines. The overall structure-reactivity relationship for these complexes, having the same array of donor atoms around the metal, has been rationalised on the basis of an interplay of electronic and steric effects. It is worth of interest to note that a fine tuning of the properties of the "spectator" ligands can lead to platinum(I1) substrates of very high reactivity, comparable or even by far greater (as for the dmphen complex) to that of similar palladium(I1) species. 4.2.I Steric Effects. The difference of reactivity between the first and the last members of the series of [Pt(N-N)(CH,)(Me2SO)]' complexes is greater than 10 orders of magnitude. These complexes are formed by the most sterically demanding ligands Meqen and dmphen. Thus, the steric bulk due to alkyl substituents at the nitrogen atoms of the chelated diamine or in the 2, 9 positions of the phenanthroline, affects the lability of the corresponding complexes in opposite directions. For Me4en a decrease of rate of ten times is observed with respect to that with en, the complex formed by unsubstituted ethylenediamine, in agreement with similar results obtained in a study of the solvolysis of alkyl substituted ethylenediamine platinum(I1) c ~ m p l e x e s . 'The ~ interpretation is straightforward. The moderate decrease of rate can be ascribed to the additional difficulty of the nucleophile in approaching the metal centre and in forming the new bond or, in other words, to a fairly modest destabilisation of the 5-coordinate transition state. In contrast, the substrate containing dmphen exchanges the coordinated Me2SO at a rate that is 5 orders of magnitude higher than the complex containing unsubstituted phen. This behaviour finds its origin in the specific relevant non-bonding repulsions by the methyl groups in positions 2,9 with the two cis groups in the square-planar configuration, as seen in the congestion and distortion of the molecular structure of [Pt(CH3)(dmphen)(PPh3)]+.There is significant steric relief when the metal adds a fifth ligand in the formation of a 5-coordinate trigonal-bypiramidal structure, where the phenanthroline occupies two positions of the equatorial ~ 1 a n e . In l ~this new arrangement, the ligand is completely planar with the methyl groups far away from short range steric repulsions. Under these circumstances the free energy of activation for ligand exchange is markedly reduced and we get an unprecedented example of steric acceleration, as a result of steric destabilization of the square-planar ground state.

@

Figure 1. ORTEP Druwing of the Complex [Pt(CH3)(drnphen)(PPh~)]PFb.

Synthetic and mechanisticpathways in platinum(1I)chemistry

213

4.2.2 Fluxionalify. No evidence for fluxionality of the N-N ligands was found in [Pt(CH3)(MezSO)(N-N)]PFs complexes except for the case of that formed by dmphen. In solution this complex is fluxional with the phenanthroline oscillating between nonequivalent bidentate modes. The mechanism proposed' involves rupture of the metalnitrogen bond and rapid interconversion of two coordinatively unsaturated T-shaped 14electron three-coordinate molecular fragments. Rates of this fluxional motion were measured by NMR spectroscopy from the exchange effects on the 'H signals of the methyl and aromatic hydrogen atoms. Since the extreme lability of Me2SO interferes with the 'H NMR measurements of the flipping we decided to block the reaction site occupied by Me2SO with triphenylphosphine or with other neutral ligands L and to start a detailed investigation on this system. The fluxional behaviour in solution that equilibrates the phenanthroline environment in the ionic complexes [Pt(CH3)(L)(dmphen)]X (X = PFC, BF4-, CF3SO3-, c104-, B(C6H5)<, [ B ( ~ , ~ - ( C F ~ ) ~ C S is H ~strongly ) ~ I - ) affected by the nature of the coordinated ligand L.16 When L = CyNH2, i-PrNH2, 2,6-Me2py, EtNH2, AsPh3, dimethylthiourea, the compounds feature a static structure in solution. When L = PPh3 (see Figurel), the fluxional motion is affected by the coordinating properties of the solvent, of the counterion X- and of nucleophiles added in solution. The mechanism is switchable between dissociative and associative pathways, depending on the structural features of the complexes and the operation of potential nucleophiles. The high trans effect and trans influence of the methyl and phosphine groups, together with the remarkable steric congestion at the square-plane seen in the molecular structure of the complex, favour either a facile Pt-N bond dissociation or the addition of a fifth group to form a trigonal bipyramidal five-coordinated species, where steric congestion by the phenanthroline is relieved. Dissociation can be prevalent within the ion-pairs formed by "non-coordinating" anions with the metallic cationic fragment in a non-polar solvent. Specific interionic dipolar interactions, such as those detected in the 19F('H) -HOESY NMR spectra for the corresponding four-coordinated [Pt(Me)(PPh3)(dmphen)]+X - (X = PFC and BF4-) complexes, could well stabilize a three-coordinated T-shaped transition state formed upon Pt-N dissociation and explain the counterion effect on the fluxionality.. Addition of external weak nucleophiles to a chloroform solution of the cationic complex ion [Pt(Me)(dmphen)(PPh3)]+accelerates the fluxional motion of the symmetric chelating ligand dmphen between non equivalent exchanging sites.

X-

Me

Scheme 3. Dissociative Mechanism Involving Cationic T-shaped Three-Coordinate Intermediates and Ionic Interaction with the counterion X .

214

Perspectives in Organometallic Chemistry

Concentration-dependent measurements were carried out with a series of sulfoxide ligands of widely different electronic and steric properties. The values of the rate constants for the processes were resolved quantitatively into steric and electronic contributions by use of QALE (Quantitative Analysis of Ligand Effects).” Inhibitory steric effects are linearly operative for the entire set of ligands, the rates of reactions are enhanced with increasing the electron donor capacity of sulfoxides, and there is a small but significant E,, effect that enhances the reactivity of the aryl sulfoxides. There is a strict analogy between the dependence of fluxional motion of dmphen upon the stereoelectronic properties of the ligands and the pattern of substitution reactions. This similarity, combined with the intrinsic lability of the platinum-nitrogen bonds, would suggest the operation of a traditional associative mechanism involving stereospecific consecutive ring-opening and ring-closure steps for the fluxional motion of dmphen. However, the available evidence does not allow to rule out alternative mechanisms involving intramolecular rearrangements of the five-coordinate intermediate, such as Berry pseudorotation or a turnstile rearrangement. In conclusion, observed consequences of the great steric destabilization of the square planar configuration of dmphen complexes are (i) a marked acceleration of the rate of ligand exchange, (ii) the easy uptake of an additional ligand (especially olefins or alkynes)15 yielding fairly stable 5-coordinate species and, (iii) a fluxional motion of the ligand (flipping). 4.2.3. Anion Recognition . The complex [ PtMe(dpa)(Me2SO)]+(CF3S03)-(dpa= bis(2pyridy1)amine) shows a strong hydrogen bonding interaction involving the amine hydrogen (N3) and a triflate anion oxygen 03, 2.898(5) A.18The tendency by the NH group of the ligand moiety to attract anions is maintained in solution of non polar solvents. Tight ion pairs of structure similar to that in the solid state are formed with socalled “uncoordinating” anions such as PFs-, BF4-, CF3S03- and with C1- in chloroform, as shown by the strong dependence of the chemical shifts of the NH, H(3) and H(3’) protons of the dpa ligand on the nature of the counterion. I9F{’H)-HOESY experiments on [PtMe(dpa)(Me2SO)]+(PF6)- in CD2Cl2 confirmed that the preferential position of the counterion is close to the NH proton. The absorption spectra are also strongly affected by the nature of the counterion. The PF6- for C1- exchange rate at the NH site, which is a bimolecular process with k2 = 96.4 f 4 M-’ s-’,was slow enough to be measured through stopped-flow experiments. Ion-pairing and full deprotonation of the amine ligand have a remarkably little effect on the reactivity.

Figure 2 . ORTEP Drawing of the Complex [ P ~ ( C H ~ ) ( ~ ~ U ) ( M ~ $ ~ O ) ] C F ~ S O ~ .

Synthetic and mechanistic pathways in platinum(II) chemistry

215

4.3. Reactivity Toward Long-Chain Bidentate Nitrogen Ligands.

A series of new organometallic binuclear platinum(I1) complexes of the type

[(PtMeCl(Me2S0)}2(p-N-N)] (N-N = H2N(CH&NH2 with n = 4, 6 , 8, 10, 12; 4,4'bipyridy 1 and 1,2-bis(4'-pyridyl)ethane) has been synthesised and fully characterised through elemental analysis, 'H and I3C NMR spectros~opy.'~ The interaction of the mononuclear model complex [PtMeCl(Me2SO)(BuWH2)1 and of the binuclear complex [ (PtMeCl(Me2S0))2(p-H2N(CH2)6NH2]with different nucleosides has been investigated by 'H NMR spectroscopy, revealing a binding preference toward N7 position in inosine and guanosine and N3 in cytidine. The compound containing 1,6-diaminohexane is interesting because it is analogous to bimetallic coordination complexes, whose antitumor activity is already known. A detailed biophysical study has evidenced that this particular organometallic compound interacts with DNA, leading to simple monofunctional adducts, in contrast with classic coordination compounds.20 4.4 Reactivity Toward Tri-Dentate Nitrogen Ligands.

In order to obtain inert species to be exploited as intercalators of nucleic acids, we synthesized organoplatinum(I1) complexes of the type [Pt(tpy)(R)]X (tpy = 2,2':6',2"terpyridine, R = Me632'or Ph"). These compounds are (i) fairly soluble and stable in aqueous solutions, (ii) relatively inert to nucleophilic substitution, and (iii) possess spectroscopic features suitable to the study of aggregation phenomena. The X-ray structure of the complex [Pt(tpy)(Me)]+BPh[ evidences the presence of n-n stacking interactions between two [Pt(tpy)(Me)]+ units in an head-to-tail di~position.~~ The Pt-Me bond is remarkably stable upon acidification and carbonylation. The planarity and electron delocalization of the terpy moiety leads to extensive stacking interactions in aqueous solution forming dimers or even higher aggregates. The interaction of the cationic complex with calf thymus DNA was investigated by UV-VIS, CD spectroscopy and gel electrophoresis mobility assays. At high rfratios (rf = [Pt]/ [DNA] the complex seem to form extended aggregates on the surface of the nucleic acid but at lower rf evidence was obtained for intercalation.

Figure 3 . ORTEP Drawing of the Dimer Formed by the Cationic Complex [Pt(Me) (tPY)l+.

216

Perspectives in Organometallic Chemistry

5 PLATINUM(I1) COMPLEXES CONTAINING PORPHYRINS.

The formation of organised supramolecular structures of chromophores having well defined shape and size is a topic of relevant interest.24Their potential application in energy storage and conversion, supramolecular catalysis, optics and electronics can be easily recognised. In this respect, porphyrins and the relative metal-derivatives are well suited compounds due to their favourable photophysical, electronic and catalytic properties. These features are strictly related to their aggregation state and strongly depend on the microstructural e n ~ i r o n m e n t The . ~ ~ introduction of properly designed substituents at the periphery of the porphine core can alter their propensity to interact selectively with organized media, also introducing rigidity and directionality in the binding mode. We took advantage of the knowledge of the lability of some platinum(I1) complexes using them as monofunctional buildind block for introducing the platinum molecular fragment and for accessing specific tailored porphyrins. The elective porphyrin was TpyP = 5,10,15,20-tetrakis(4-pyridyl)-2 1H,23H-porphine and the platinum complexes used so far are [Pt(CH3)(Cy2dim)(Me2SO)lf(Cy2dirn= dicyclohexyldiimine)26

Scheme 4. Ball and Stick Representation of the TpyP Porphyrin [Pt(CHj)(Lauzdim) (Me 2SO)]+ (Lauzdim = didodecyldiimine)~’ [Pt(tpy)(CH3CN)] (CF3S03)2 (tpy =,2’:6’,2”-terpyridine)2 * and [Pt(P-P)Me(Me2SO)]+ (P-P = various chelating phosphines containing redox active groups or chiral residues).28 The sites of attack are the meso substituting pyridines at the periphery of the TpyP porphyrin and the reaction between the two reagents leads almost quantitatively to the formation of a tetranuclear complex. The gradual platination of the porphyrin can be monitored easily in situ through ’H NMR spectroscopy, by titrating a solution of TpyP with the platinum complex. The aromatic region of the spectrum evidences a decrease of the signals relative to TpyP and a matching increase of peaks due to intermediate mono, di and tri-substituted species, affording eventually the required tetra-substituted complex at a metal to porphyrin ratio of 4:l. The metal fragment provides the function of (i) increasing the aromatic surface, consequently enhancing the self-aggregating properties of the species (e.g. [Pt(terpy)] fragment), (ii) introducing redox-active groups, able to influence the photophysical properties of the porphyrin ligand, and (iii) tuning the extent of hydrophobicity of the system. The new formed species exhibit interesting features. Slow evaporation of an acetone or chloroform solution of the novel tetranuclear platinum(I1)

Synthetic and mechanistic pathways in platinum(II) chemistry

217

Scheme 4. Space Filling Molecular Model of the [(Pt(Lau,dim)Me) JTpyP) ] 4+ Cation.

complex [(Pt(Lau,dim)Me) ,(TpyP) ](CF3S03), on a glass surface affords globular micrometric sized aggregates as shown by Scan Electron Microscopy (SEM). The formation of porphyrin rings in the mesoscopic range has been achieved through simple solvent evaporation of the [Pt(P-P)Me],(TpyP)(CF, SO,),salt.

6 REACTION PATHWAYS OF CYCLOMETALATED Pt(I1) COMPLEXES 6.1 Reactivity of [Pt(N-N-C)CI] Complexes : Associative Mechanism.

Some years ago van Eldik et a13,4raised the important question as to whether the considerable acceleration in rate observed for the substitution of the aqua ligand in the complexes [Pt(N-C)(N)(H20)] (N-CH = N,N-dimethylbenzylamine,N = pySO3-3) 334 and [Pt(N-C-N)(H20)]’,29 (N-CH-N = 2,6-bis((dimethylamino)methyl)phenyl) stems from a significant contribution of back donation into the empty x*-orbitals of the in-plane aryl ligand. Our opinion was that it is simply the result of the well known strong trans labilising effect of the Pt-C bond.30However, we were aware that the problem could not find a definite resolution in the lack of a proper comparison of reactivity between cyclometalated and uncyclometalated substrates having a strictly similar set of donor atoms. Thus we carried out kinetic studies on cyclometalated compounds of the type [Pt(N-N-C)Cl], containing a number of terdendate N-N-C anionic ligands derived from deprotonated alkyl, phenyl or benzyl 6-substituted-2,2’-bipyridinesand exhibiting different structural properties only at the periphery of the substrate. l 2 The reaction studied was chloride for phosphine substitution and these rates have been compared with those of the corresponding [Pt(N-N)(C)Cl] (N-N = 2,2’-bipyridine, C = CH3 or CsHs) complexes having the same set of donor atoms but less constrained arrangements of the ligands. All reactions are first-order with respect to complex and phosphine concentration, obeying the simple rate law kobsd = k2 [ PPh31. The values of second-order rate constants k2 do not seem particularly sensitive to the nature of the bonded organic moiety (alkyl or aryl), to its structure (cyclometalated or not), to the size of the ring or to the number of alkyl substituents on it. The effects are those foreseen on the basis of an associative mode of activation. Factors of primary importance in controlling the reactivity and the mode of activation are the presence of a planar a,a’diimine, such as 2,2’-bipyridine, the presence of an extensive n: system on the chelating

Perspectives in Organometallic Chemistry

218

CI

OH2

A

B

ligand and the ease with which this 'x: system interacts with the nonbonding d electrons of the metal, thereby increasing the electrophilicity of the metal centre and assisting the nucleophilic attack. The sequence of rate data along the examined series of complexes strongly suggests that the introduction of a cyclometalated ring has little effect on these factors and on the reactivity. It is of interest to compare briefly the behaviour of complexes containing N-N-C and N-C-N anionic ligands, at least as far as the lability of the fourth coordinated group is concerned. Both the cyclometalated compounds A and B exhibit high reactivity. The relief of electron density from the metal through n-bonding by the planar di-imine is crucial in controlling reactivity and activation mode of compound A. The substitution rate is comparable to that of the uncyclometalated [Pt(bpy)(CH3)Cl] ( k2 = 1046 *6 M" s" vs 1854 f 37 M-*s-') The introduction of a cyclometalated in-plane aryl ring in A, separated through a spacer methylene group from the bipy moiety, hardly affects the rate of chloride substitution. This fact rules out any effective back-donation towards the aromatic n* orbitals in the 18-e five-coordinated transition state. This should be true also for compound B, even though some flow of electron density from the metal to the in-plane aryl ring cannot be excluded. 6.2 Reactivity of [Pt(C-C)(SR2)2] Complexes :Dissociative Mechanism

A conclusion of the previous work was that cyclometalation does not cause a changeover of mechanism or particular acceleration on the reactivity of complexes containing a bipyridyl fragment. In other words, 7c back-bonding from the metal to the in-plane aryl rings does not come into play. In particular, there is no appreciable change of the electrophilic character of the platinum(I1) centre with respect to uncyclometalated complexes. The obvious next step was to analyse the behaviour of a carbon analogous of 2,2'bipyridine and therefore we carried out comparative kinetic and theoretical studies on the complexes [Pt(bph)(SR2)2](bph2-= 2,2'-biphenyl dianion) and ~is-[PtPh2(SR2)2].~'In the first complex the aryl groups lie approximately in plane with the complex and photochemical studies32suggested an effective electron density transfer from filled d orbitals of the metal into empty n* orbitals of the cyclometalated ligand. The second compound, in which the two monodentate o-bonded aryl rings are almost perpendicular to the plane of the complex, is known to exchange the thioethers with a dissociative mechanism.33 The reactions studied were the temperature and concentration dependencies of ligand exchange of [Pt(bph)(SR2)2] (R = Me or Et) with *SR2 and of [PtPh2(SMe&]

219

Synthetic and mechanistic pathways in platinum(II) chemistry

with *SMe2 and the kinetics of displacement of the thioethers with the dinitrogen ligands 2,2'-bipyridine and 1,lO-phenanthroline. Theoretical ab-initio calculations were performed for both [Pt(bph)(SMez)z] and cis-[PtPh2(SMe2)2], as well as for their threecoordinated T-shaped derivatives upon the loss of one SMe2 ligand. Calculations were performed in the HF approximation and refined by introducing the correlation terms. The assessment of a dissociative mechanism has been made essentially on the basis of: (i) the independence of the rates of ligand exchange on the nature and concentration of the entering thioethers, (ii) the saturation profile observed in the reactions with diimines, as a result of mass-law retardation produced by the leaving thioether group, (iii) the identity of the rate of dissociation with the rate of solvent exchange, (iv), the sign and the magnitude of the entropy of activation always positive, (v) positive values of the volume of activation. For example, a variable-pressure study on cis-[PtPh*(SMe2)2],assumed in this work as paradigmatic of a dissociative behaviour, has already led to a positive volume of activation AV' = + 4.7 A0.5 cm3 mol-' in toluene. 34

8;s

AH& 52.3 kJ moll

AH& 72.2 kJ moll

Figure 4. Geometries for Starting Square-Planar Complexes and T-shaped Molecular Fragments Formed Upon Ligand (SMe2) Dissociation. The activation enthalpies from the optimized vacuum-phase geometries for [Pt(bph)(SMe&] and cis-[PtPh2(SMe2)2] are 52.3 and 72.2 kJ mol-', respectively, and compare well with the experimental values obtained in CDCl3 solution. Thus, the complexes [Pt(bph)(SR2)2], which combine cyclometalation and a favourable in-plane disposition of the aryl groups, show reaction trends which differ substantially from those of other species carrying ligands such as chelating di-imines. The crossing of the negative electrostatic potential which surrounds the 16-electron square-planar substrate by the nucleophile seems to be a prerequisite to start the associative mode of activation. When the lone pair of the entering ligand does not find a positive electrostatic channel to pursue, dissociation takes-over as an energetically favoured reaction pathway. Factors which concur in promoting dissociation are, (i) bond weakening at the leaving group due to the trans influence of strong o-donors (in the present cases, the organic carbanions which form strong Pt-C o-bond) and, (ii) the stabilisation by the remaining set of three in plane ligands of a three-coordinated 14-electron intermediate without changing the singlet ground state. These favourable factors do not apply in the presence of ligands which exhibit major n-acceptor capabilities. By relieving the excess of electron density at the metal they not only destabilise the 14-electron 3-coordinated intermediate but facilitate the addition of the axially incoming nucleophile. Evidently, the cyclometalated

220

Perspectives in Organometallic Chemistry

2,2'-biphenyl dianion either does not possess nor can exert a similar behaviour in spite of its presumed .n-acceptor capabilities. References 1 R. Romeo, Comments Inorg. Chem., 1990, 11,21. 2 R. Romeo, Comments Inorg. Chem., 2002,23, 79. 3 M. Schmiilling, A. D. Ryabov and R. van Eldik, J. Chem. SOC.Chem. Commun., 1992, 1609. 4 M. Schmiilling, A. D. Ryabov and R. van Eldik, J. Chem. SOC.Dalton Trans., 1994, 1257. 5 C. Eaborn, K. Kundu and A. J. Pidcock, J.Chem.Soc. Dalton Trans., 1981,933. 6 R. Romeo and L. Monsu Scolaro, Inorg. Synth., 1998,32, 153. 7 R. Romeo, L. Monsu Scolaro, N. Nastasi, B. E. Mann, G. Bruno and F. Nicolo, Znorg. Chem., 1996,35,7691. 8 M. Caligaris and 0. Carugo, Coord. Chem. Rev., 1996, 153, 1 . 9 L. Monsu Scolaro, A. Mazzaglia, A. Romeo, M. R. Plutino, M. Castriciano and R. Romeo, Znorg. Chim. Acta, 2002,330, 189. 10 P. S. Pregosin, Annu. Rep. NMR Spectrosc., 1986, 17, 285. 1 1 R. Romeo, L. Monsu Scolaro, N. Nastasi and G. Arena, Znorg. Chem., 1996,35, 5087. 12 R. Romeo, M. R. Plutino, L. Monsu Scolaro, S. Stoccoro and G. Minghetti, Inorg. Chem., 2000,39,4749. 13 L. Monsu Scolaro, A. Mazzaglia, A. Romeo and R. Romeo, J. Inorg. Biochem, 2002,in the press. 14 F. P., Fanizzi, F. P. Intini, L. Maresca, G. Natile and G. Uccello-Barretta, Inorg. Chem., 1990 ,29,29.§ 1 5 V. G. Albano, G. Natile and A. Panunzi, Coord. Chem. Rev., 1994,133,67. 16 R. Romeo, L. Fenech, L. Monsu Scolaro, A. Albinati, A. Macchioni and C. Zuccaccia, Inorg. Chem., 2001,40,3293. 1 7 R. Romeo, L. Fenech, S. Carnabuci, M. R. Plutino and A. Romeo, Inorg. Chem., 2002, 41,2839. 18 R. Romeo, N. Nastasi, L. Monsu Scolaro, M. R. Plutino, A. Albinati and A. Macchioni, Inorg. Chem., 1998,37,5460. 19 G. Cafeo, C. Lo Passo, L. Monsii Scolaro, I. Pernice and R. Romeo, Inorg. Chim. Acta, 1998,275-276, 141. 20 V. Marini, J. Kasparkova, 0. Novakova, L. Monsu Scolaro, R. Romeo and V. Brabec,J. Biol.Inorg. Chem., 2002, in the press. 2 1 G. Arena, L. Monsu Scolaro, R. F. Pasternack and R. Romeo, Inorg. Chem., 1995,34, 2994. 22 G. Arena, G. Calogero, S. Campagna, L. Monsu Scolaro, V. Ricevuto and R. Romeo, Inorg. Chem., 1998,37,2763. 23 R. Romeo, L. Monsu Scolaro, M. R. Plutino and A. Albinati, J. Organomet. Chem., 2000,593-594,403. 24 J. M. Lehn, Supramolecular Chemistry ;VCH: Weinheim, 1995. 25 W. I. White, in The Porphyrins, ed. D. Dolphin, Academic Press, New York, 1978, vol. 5, Chapter 7. 26 L. Monsu Scolaro, C. Donato, M. Castriciano, A. Romeo and R. Romeo, Inorg. Chim. A c ~ u2000,300-302,978-986. , 27 M. Castriciano, A. Romeo, R. Romeo and L. Monsu Scolaro, Eur. J. Inorg. Chem., 2002,53 1. 28 Unpublished work.

Synthetic and mechanistic pathways in platinum(II) chemistry

22 I

29 M. Schmulling, D. M. Grove, G. van Koten, R. van Eldik, N. Veldman and A. L. Spek, Organometallics, 1996, 15, 1384. 30 L. I. Elding and R. Romeo, J. Chem. SOC.,Dalton Trans., 1996, 1471. 3 1 M. R. Plutino, L. Monsu Scolaro, R. Romeo and A. Grassi, Inorg. Chem., 2000,39, 27 12. 32 G. Y. Zheng and D. P. Rillema, Inorg. Chem., 1998,37, 1392 and references therein. 33 R. Romeo, A. Grassi and L. Monsu Scolaro, Inorg. Chem., 1992,31,4383 and references therein. 34 U. Frey, L. Helm, A. E. Merbach and R. Romeo, J. Am. Chem. SOC.,1989,111,8161.

NEW PERSPECTIVES FOR OLEFIN COMPLEXES: SYNTHESIS AND CHARACTERISATIONOF STABLE RHODIUM(0) AND IRIDIUM(0) COMPLEXES

J. Harmer', G. Frison2, M. Rudolph3, H. Schonbert, S. Deblon2, P. Maire2, S. Boulmk, F. Breher2,C. Biihle? ,H. Riiegger2,A. Schweiger and H. GrWmacher2 1

Laboratory of Physical Chemistry, ETH-Honggerberg, CH-8093 Ziirich, Switzerland, E-mail: [email protected] Laboratory of Inorganic Chemistry, ETH-Honggerberg, CH-8093 Ziirich, Switzerland, E-mail: [email protected] Faculty of Chemistry, Friedrich-Schiller-University of Jena, August Bebel Str. 2, D-07743 Jena, Germany, E-mail: [email protected] 1 INTRODUCTION

About five years ago we discovered accidentally the di(tert.butyl)tropylidenyl phosphane 3 [IUPAC: (dibenzo[a,d]cyclohepten-5-yl)-di(tert.-butyl)phosphane] as the minor component (40%) of the reaction mixture when the cyclic hosphonium salt 1 was treated with sodium bis(trimethylsily1)amide 2 (eq. (I), Scheme 1).

P

"-

- NaCl - HN(SiMe,), 3

1

5

4

7

6

HPR,

E:P,As A R -1, aryl, NR',,OR',etc. R': alkyl, aryl, NR',, OR', F,Cl, Br, CN, e X:F, C1, Br, I, CN,SO,R', NO,, NR',, etc. Scheme 1 Synthesis of troppRtype ligands

p

~8

R,

tropp'

8

223

New perspectives for olejn complexes

The other major product of this mechanistically not understood reaction is the phosphane 4 (60%). The potential of phosphane 3 as a ligand was recognized by the result of an X-ray structure analysis which revealed that the phosphanyl group and the olefin moiety of the central seven membered ring are arranged such that a concave binding site is obtained. Steric considerations show that a bulky group will bind axially to the 5-position of the dibenzo[a,d]cycloheptenyl ring and NMR spectroscopic investigations indicate that no ring inversion is observable on the NMR time scale up to T = 120°C. An alternative simple and straight forward synthesis of these chelating phosphanes which we named troppR (the superscript R denotes the substituent bonded to the phosphorus atom) was subsequently developed as shown in eq. (2) in Scheme 1. The commercially available dibenzosuberenone 5 is first reduced to the corresponding hy droxyl compound 6 with sodium boronhydride in methanol and subsequently converted to the chloride 7 with freshly distilled SOC4 in toluene or methylene chloride as solvent. Addition of phosphanes, HPR2 (R = alkyl or aryl), in toluene solution leads to a precipitation of the corresponding hydrophosphonium salts, [R2HP-trop]+C1-, which upon gentle warming in presence of an aqueous deoxygenated solution of mco3 cleanly give the troppRcompounds 8 usually in over 80% yield.2 As indicated in the general formula A in Scheme 1, the donor ER2 group can be varied broadly. Additionally, further substituents R' at the olefinic C=Cvopbond of the central seven-membered ring and/or X at the benzogroups can be easily introduced into the ligand framework, which allows the sterical and electronical fine-tuning of the ligands. 2 RESULTS AND DISCUSSION

2.1 Synthesis of 16 electron and 18 electron rhodium(1) and iridium(1) tropp complexes The synthesis of metal complexes with troppR type phosphanes as ligands is straight forward and employs either metal salts delivering an electronically and coordinatively unsaturated metal fragment [like Cu(1)Cl or Ag(I)03SCF3I1 or a metal complex bearing labile ligands like [M(cod)z]+X(9: M = Rh, 10: M = Ir; cod = q4-1,5-cyclooctadiene;X- = CF3 03-, BF4-, PFs-) as precursors.

M = Rh trans- 11 M = Ir trans- 12

Cis- 11

cis-I2

X- = CF,SO,- or BF,- or PF,

Scheme 2 Synthesis of cationic I6 electron rhodium and iridium troppphcomplexes

Perspectives in Organometallic Chemistry

224

Especially the cationic 16 electron rhodium(1) and iridium(1) troppR complexes were of interest to us and an example of their synthesis for R = Ph is shown in eq. (3) in Scheme 2. The reaction of two equivalents of dibenzo[a,d]cyclohepten-5-yl-diphenylphosphane, troppph 8, with either 9 or 10 gives quantitative yields of red air stable crystals of compounds [M(troppPh)2]+X11 (M = Rh)3 or 12 (M = Ir): Both complexes are obtained as mixtures of trans- (11: 80%; 12: 88%) and cis-isomers (11: 20%; 12: 12%) which do not interconvert on the NMR time scale in non-coordinating solvent^.^ This observation is in accord with a square-planar coordination sphere of the metal centers in 11 and 12 and the prediction that a cis-trans-isomerism is a thermally forbidden process for planar tetracoordinated 16-electron complexes.6 In the case of tetra-coordinated rhodium [Rh(troppR)2]' complexes with various substituents R, the trans-isomers are generally characterized by positive lo3Rh NMR shifts [6(lo3Rh) 11 ppm - 1164 ppm] and small 1 103 J( Rh3lP) coupling constants (I 130 Hz) while the cis-isomers show frequently negative lo3RhNMR shifts [S('03Rh) -8 ppm - -372 ppm] and 'J('03Rh3'P) coupling contstants > 170 Hz. The 13C NMR shifts of the coordinated C=CwOp bond varies between 6 70 (for tran~-[Rh(~tropp~~)2]+) and 6 145 (for ~is-[Rh(~~tropp~~)(tropp~~)]+). When two isomers are formed, the trans-isomer shows the lowest frequency (high field) shifted 13C resonances (&is > 10). 2.2 Reactivity of 16 electron rhodium(1) and iridium@)tropp complexes

In a coordinating solvent like tetrahydrofuran (thf) or acetonitrile (acn) or in the presence of a coordinating anion like C f , the tetracoordinated complexes 11 and 12 give penta-coordinated complexes. The formation can be easily followed by 'H and 31PNMR spectroscopy. These complexes [M(L)(troppPh)2]"have a ground state structure with the phosphorus centers in the axial and the olefin units in equatorial positions of a trigonal bipyramid.' As is indicated in Scheme 3 by differently shaded circles, the 'H and 13CNMR signals of the coordinated olefin units of the tropp ligands allow to distinguish and to assign the structures of the species involved in the equilibrium indicated in eq. (4). P troppph =

M Rh trans- 11 Ir trans- 12

n=O

n=+1

13-CI 14-CI

13-acn 14-acn

cis- 11 cis- 12

Scheme 3. Formation of penta-coordinated 18 electron [A4L(troppp~2/complexes. The grey, white and black circles indicate the ' H and 13C nuclei of the coordinated olefinic units, respectively. while these are equivalent in the tetra-coordinated species trans/ cis11,12 (grey circles) they become inequivalent in the penta-coordinated complexes 13 and 14.

New perspectives for olefin complexes

225

Line shape analyses (LSA) of the temperature dependent 31Pand 'H NMR spectra allow to estimate constants for the formation (kf) and dissociation (kd) of some of the complexes involved in (4). In particular, we investigated the equilibria given in eqs. (5) - (9). Selected data are listed in Table 1.

Table 1. Dissociation rate constants kd and activation energies AG#(T)for equilibria (5) (9). equilibium kd [s-'1 (T [K]) AG' kJ mol-'(T [K]) method (5) 8640 (295) 50 (295) LSAa (6) 238 (295) 58 (295) LSAa (7) 6b (298) 69d(298) SI (FHMT) (8) 8 (298) 68 (298) LSA' (9) 3700 (298) 53 (298) LSA' kf'=2.3 x lo6 a The data were determined by a line shape analyses (LSA) of the 31PNMR resonances using the MEXICO program packagesa which gives k i at different temperatures. The activation energies were obtained from Eyring-plots [Rln(k: h/kbT) vs. 1/T]. The process 7 is slow on the NMR time scale and a LSA cannot be preformed. A spin inversion (SI) 'P NMR experiment [Forsen-Hoffman ma netization transfer (FHMT)9was used and the data fitted with the CIFIT program packagefb9'to determine the rate constant k&s = k: of 14-C1. The activation energy of the process, 14-C1 + 14-acn [see forward reaction in eq. (7)] was obtained from an Eyring-plot. ' The data were determined by a line shape analyses (LSA) of the coordinated olefinic resonances using the MEXICO program packagesa which gives k: at different temperatures. The activation energies were obtained from Eyring-plots.

P

(5) [~h~l(tropp'~>21 s trans-[~h(tropp'~>21+ + CI13-C1 trans- 11 (6) [RhCl(troppPh)2]S cis-[Rh(troppph)2]' + C113-C1 cis-11 (7) [1r~l(tropp'~)21+acn s [1r(acn>(tropp'~)2]+ +~ 1 14-C1 14-acn

(8) [IrCl(troppPh)z]S trans/cis-[Ir(troppph)~]++ C114-C1 trans/cis-12 (9) trans/cis-[1r(tropp'~)21+ + acn s [1r(acn)(tropp'~)21+ trandcis- 12 14-acn In the dissociation process of the 18-electron rhodium chloro bis(troppPh)complex 13C1 into the tetra-coordinated 16-electron complexes trans-11 and cis-11 all species are observed directly by NMR spectroscopy. Hence, the data for both equilibria (5) and (6) can be obtained. The dissociation of 13-C1 to give the trans-complex is about 40 times faster than to the cis-complex. The higher activation barrier [A(AG#,is-wms = 8 kJ mol-'1 for the formation of the latter is likely due to the fact that apart from Rh-Cl bond rupture also a considerable rearrangement of the ligand sphere, i.e. a type of Turnstile-rotation, must

226

Perspectives in Organometallic Chemistry

occur. On the other hand, the formation of the trans-isomer trans-11 is pre-organized in the structure of the 13-C1and follows simply the Rh-C1 bond stretching mode. The dissociation of the penta-coordinated iridium complex 14-C1 into the ions transhs[Ir(troppPh)z]' (trandcis-12) and C1- is not directly observed by NMR spectroscopy. Therefore the mutual substitution reaction shown in eq. (7) occurring when 14-C1 is dissolved in CHzC12 containing various amounts of acetonitrile was investigated. Under these conditions, only the penta-coordinated complexes 14-Cl and 14-acn are observed. The rate constant for the forward reaction [IrCl(troppPh)2] (14-C1) + acn -+ [Ir(acn)(troppPh)2]' (14-acn) + C1- was obtained using a spin inversion 31P NMR experiment (Forsen-Hoffman magnetization transfer) at various temperatures. From an Eyring-plot, the activation energy AG'(298 K) = 69 kJ mol-' is obtained for reaction (7). It is assumed that the dissociation of the chloro complex 14-CI is the rate determining step in reaction (7) and that the observed rate constant, &bs, equals the rate constant kd of the dissociation to the tetra-coordinated complexes trans/cis-12 which cannot be directly observed.lo This assumption is supported by the data which are obtained when line shape analyses of the olefinic proton resonance signals for 14-C1 and 14-acn are performed. These give separately access to the dissociation parameters for the reaction 14-C1 + trandcis-12 + C1[eq. (S)] and 14-acn + trans&-12 + acn [eq. (911. Clearly, the activation energies for the processes (7) and (8) are very similar while the dissociation of the acetonitrile complex is much faster (4.6 x lo2). A UVNIS titration of a CH2C12 solution of the tetra-coordinated complexes trans/cis-[Ir(troppPh)2]+(PF6)- with acetonitrile, furnishes the equilibrium constant, K = 626, for reaction (9). With kf = K x kd, the rate constant kf = 2.3 x lo6 M-' s-l for the formation of the iridium acetonitrile complex 14-acn is calculated. This rate constant is six orders of magnitude larger than the dissociation rate constant of the chlorocomplex which is thus rate determining in the ligand exchange reaction (7). From a van'tHoff plot (-RlnK vs. 1/T) the thermodynamic data (AH", ASo) for the substitution reaction 14-C1+ acn -+ 14-acn + C1-were obtained [see forward reaction in eq. (711. The data show that the cationic acetonitrile complex 14-acn is preferred over the chloro-complex 14-C1 by AH" = -26 kJ/mol. However, the entropy for this process is also negative, AS" = -116 J mol-' K-', whereby the substitution of the chloride becomes endogonic by AGO = 8.4 kJ mol-' at T = 298 K. A negative reaction entropy is expected for a process where a neutral complex reacts to give charged products and consequently a higher-organized solvation sphere is created. The reactions (5) - (9) give fundamental insight into the stability of 18 electron [M(L)(troppPh)2]"complexes containing different ligands L and provides valuable data elucidating the reaction behavior of [M(tropp)z]complexes (vide in@u). 2.3. Synthesis and EPR data of neutral paramagnetic 17 electron rhodium(0) and iridium(0) troppRcomplexes The cationic 16-electron complexes trans&-1 1 and transkis-12 are reduced either electrochemically or simply by alkali metals like Li, Na, or K on a reparative scale to give the neutral paramagnetic 17-electron complexes trans/&-[M(tropph! )2]0 (trans/cis-15: M = Rh; trans/cis-16: M = Ir) with a d9 valence electron configuration at the metal centers in high isolated yields (Scheme 4)?-5 Further reduction leads to the d" valence electron configured metalates [M(troppPh)2]-(M = Rh: 17; M = Ir: 18) which were also isolated and fully characterized.

ax-

New perspectives for olefin complexes

227

Ph2

Ph2

\

M = Rh M = Ir

M = Rh M = Ir

Ph* trans-I1 trans-12

+

c

cis- 11 cis- 12

trans- 15 trans-I6

cis- 15 cis- 16

M = R h 17 M=lr 18

Scheme 4. Synthesis of neutral paramagnetic [ ~ ( t r o p p ~compIexes y2~

8

[M(troppph)2Jo and anionic dl'

Recently, Longato et a1."712and Le Floch and Mathey et al.I3 succeeded in isolating comparable d9 and d" rhodium and iridium complexes with either 1,l'bis(dipheny1phosphino)ferrocene (dppf), or 3,3',4,4'-tetramethylbiphosphinine (tmbp) as ligands with electron acceptor properties. The redox potentials of [M(R'troppR)2]' complexes are remarkably low (see Table 2) when compared to other rhodium and iridium c~mplexes.'~ The stabilizing effect of the tropp ligands on formally low metal oxidation states is manifested also in the straight forward synthesis of the analogous cobalt(0) complex [Co(troppPh)2]Ofrom CoBr2, troppph, and zinc dust. The electronic and geometric structures of the complexes [M(troppPh)2I0(15: M= Rh, 16: M = Ir) were investigated by CW and echo-detected EPR in combination with pulse ENDOR and ESEEM techniq~es.~ The resulting experimental hyperfine coupling constants are given in Table 3.

'

228

Perspectives in Organometallic Chemistry

ds, 2%

Table 2. Half-wave peak potentials and of various [MtltroppR)z/+complexes (M = Rh, Ir; R1 = H, Me, R = Ph, Cyc) versus [Ag/AgCq at a scan rate of v = I00 mV sec-'. Working electrode: Pt-wire; electrolyte: thj70.I M nBuaPF6; T= 20°C.

[Rh(troPPY")i+ [wMetroPP )21 [Rh(troPPh"y921+ [Ir(tropp )2]+a a Electrolyte: CH2C12/0.1 M nBmNPF6 +

E'% (V) -0.917 -0.892 -1.189 -0.650

EL% (V) -1.308 -1.302 -1.532 -0.920

Table 3. Experimental hyper$ne coupling constants [in MHz] of [M(troppph)J complexes trans/cis-lS (M = Rh) and trans/cis-16 (M = Ir). trans-[Ir(troppY")2](trans-16) trans-[~h(tropp")2] (trans-15) Nuclei Aiso A1 A2 A3 Aiso A1 A2 A3 31pa 45 55 45 45 -430 -40 4 0 103wa 20 23 19 19 cis-[Ir(troppYh)2](cis-16) cis-[~h(tropp'~)2](cis-15) Nuclei Aiso A1 A2 A3 Aiso AI A2 A3 31pa 69.5 80 65 65 -40 4 0 <80 1 0 3 m 17 20 16 16 a Absolute values are given for the experimental data. Errors were estimated to be h1 MHz. The trandcis-[Rh(troppPh)2]0 complexes show only a small g anisotropy (trans-15: giso = 2.037, g1=2.030, gll = 2.050; cis-15: giso = 2.019, g1=2.013, gll = 2.030). The g matrix of Ph 0 [Rh(tropp )2] is axially symmetric with gil > gL indicating either a distorted square planar structure (SOMO essentially):.d: or a compressed tetrahedron (SOMO essentially dxy).The irdium(0) complexes trandcis-16 show comparatively large g anisotropies (trans16: giso = 2.093, g1=2.150, gll = 1.980; cis-16: 2.050, g1=2.060, gll = 2.030) and a reverse ordering with gL > gll is found, which cannot be explained by simple ligand field arguments. The experimental hyperfine interactions of the unpaired electron with the metal (Io3Rh) and the surrounding nuclei (31P, 'H, I3C) were found to be small which by comparison with atomic constants, implies small spin densities on these centers. However, especially the good agreement of the distance (2.65 A) between the olefinic protons and the metal centers determined from the dipolar coupling parameter indicates that the unpaired electron is primarily located at the metal center. This is supported by DFT calculations on model c~mplexes'~ which show that about 60% of the spin density is located on the rhodium center in 15 and 50% on the iridium center in 16. It must be noted, however, that the structural agreement between the experimentally determined structures (vide inJi.a)and the calculated model complexes which differ from for cis-15 and trans-16 by replacing all arene groups (i.e. the annulated benzo groups and P-phenyls) by hydrogen is poor. The structures of the models show distorted structures close to tetrahedral structure of the cobalt(0) complex [Co(troppPh)2]and as in this one any differentiation between cisand trans-isomers becomes meaningless. However, cis- and the trans-isomers were detected for [Rh(tr0pp~~)2]~ and [Ir(troppPh)2]0 for which a dynamic equilibrium was established. The thermodynamic data show that the cis-isomer is slightly preferred by AH" = -4.7 k 0.3 kJ/mol (M = Rh) and AH" = -5.1 +_ 0.5 kJ/mol; (M = Ir). Because the entropies for the process tran~-[M(tropp~~)2]~ S cis-

New perspectives for olejin complexes

229

[M(troppPh)2l0are also negative [AS" = -5 f 1.5 J/mol (M = Rh); ASo = -17 & 3.7 J/mol (M equilibrium constants of Kbms+cis = 3.65 for M = Rh and Kms+cis = 1.03 for M = Ir at T = 298 K, respectively, become small at T = 298 K. = Rh)] the

2.4. Reactivity of 16-, 17- and 18-electron iridium(0) tropp complexes Tetra-coordinated 17-electron and 18-electron complexes merit attention because they have a promising potential in bond activation ~hemistry'~-'~ and were discussed in the context of the photocatalytic H202' and HB? splitting. We investigated the reactions of the iridium complexes [Ir(troppPh)2]' (trans/&-la), [Ir(troppPh) 2 ] 0 (trans/cis-16), and [Ir(troppPh)2]- (18) with hydrogen and protic reagents (various acids, H20) which are summarized in Scheme 5.' P troppph =

trans/cis-[~r(troppPh)),l+ trandcis-I 2 I9

+ H+l

I

1

base, - BaseH'

20

18

Scheme 5. Reactions of 1 6 , 17-, and 18-electron [Ir(tr0pp~92]~ (n = +I, 0, -I) complexes with H2 andprotons. As a result, we obtained the reaction diagram shown in Scheme 5 in which two electrons [steps (a) and (a)] and subsequently two protons [steps (c) and (43 are added "dropwise" to produce finally dihydrogen [step (e)]. The latter reaction is remarkably clean and 19 decomposes quantitatively within hours in solution (days in the solid state) to give analytically pure starting material trans/cis-l2. This cycle includes all chemical steps which are necessary to produce dihydrogen from proton sources. A key-point for the rationale design of such cycles is a better knowledge of the thermodynamic and kinetic parameters which interconnect all participating species. Especially, an evaluation of the reactivity of the 17-electroncomplexes 16 was of interest. To this end, we investigated the reaction with HZand protons with trans/cis-16 and observed in both cases the quantitative formation of the very stable 18-electron monohydride complex 20. However, also the 16electron complexes [Ir(troppPh)2]' (trans/cis-12) and the 18-electron iridate [Ir(troppPh)2]-

230

Perspectives in Organometallic Chemistry

(18) react rapidly and exothermically with H2. The latter reaction is remarkable and produces one equivalent of “inorganic” hydride in the form of LiH or NaH ([Li(thf)4]+or [Na(thf)6]+serve as counter cations for 18). Thus it is possible that the “reactivity” of the 17-electron complexes trans/ci~-[Ir(tropp~~)2]~ is evoked in reality by the diamagnetic closed-shell species trandcis-12 and 18 which may be present in the equilibrium (12) trans/cis-[1r(tropp~~)21++ [1r(tropp~~)215 2 trans/cis-[Ir(troppPh1210, and characterized by the disproportion constant Kdls,,. In order to gain more insight in this process, square wave (SW) voltammograms of trans/cis-[1r(tropp~~)21+(trans/cis-12) in ~ ~ 2 solution ~ 1 2 containing various amounts of acetonitrile (0.0375 - 0.73 mol L-’) were taken. A typical SW voltammogram is shown in Figure 1. I/pA 30 25 20

15 10 5

E(V): -0.6 -0.7 -0.8 -0.9 -1 Figure 1. Comparison of measured (-------) and calculated (0 o o o 0) square wave (SW) voltammograms of trans/cis-[Ir(troppPh)2]+(trans/cis-l2) in CH2Cl2 containing 0.73M L of acetonitrile atpequencies of 25, 100, 200, 350 and 700 Hz. The curve obtained with the lowestpequency (25 Hz) is indicated as ( 0 0). The mechanism used to simulate22the curves is given in the equations (9) - (15): (9) trans/cis-[~r(tropp~~)~]+ + acn 5 [1r(acn)(tropp~~>21+ trans/cis-12 16acn (1 0) trans/cis-[1r(tropp~~>21+ + etrandcis- 12

s trans/cis-[~r(tropp~~)~~~ trans/cis-16

(I 1) trans/cis-[1r(tropp~~)21~ + e- 5 [1r(troppPh)2]trans/cis-16 18 (12) trans/cis-[1r(tropp~~)2]+ + [1r(tropp~~)215 2 trans/cis-[1r(tropp~~)21~

trans/cis-12

18

trans/cis-16

23 1

New perspectives for oleJn complexes

(13) [1r(acn)(tropp~~>21+ + e- + trans/cis-[1r(tropp~~)21~ + acn

14-acn

trans/&-16

(14) trans-[~r(tropp~~)z]' s cis-[1r(tropp~~)2]~

trans-16

cis-16

The following data given in Table 4 were obtained. Table 4. Simulated thermodynamic and kinetic data >om square wave voltammograms of trans/cis-[~r(tropp~~J (trans/cis-12). reaction K kf[M-' S-'1 kb (9) trans/cis-l2 + acn 5 14-acn 620 - 520 (3.6-2.5)x106 (5.8-4.8)x103[s-'1 (12) trans/cis-12 + 18 S 2 trans/cis-16 3.7 x lo4 =5 x 10' =i.4 x lo4[M-' s-'] (14) trans-16 S cis-16 =l 1 x lo4 cS-l] For the reaction (9), K corresponds to the equilibrium constant for the formation, kf to the rate constant for the formation, and kb = kd to rate constant for dissociation of the pentacoordinated acetonitrile complex [Ir(acn)(troppPh)2]' (14-acn) and they agree satisfactorily with the values determined by the UVNIS and NMR experiments discussed above ( K = 626, kf = 2.3 x lo6 [M'' s-'1, kd = 3.7 x lo3 [s-'1. The equilibrium constant for the trans/cis isomerisation of the 17-electron complexes [Ir(troppPh)2I0given by eq. (14) is found to be close to 1 which is also in accordance with the value determined by EPR-spectroscopy. The rate constant kf = kiso for this process is obtained by the simulation of the SW voltammogram and shows this process to be quite fast (1 x lo4 s-'). This indicates that the energy difference between square-planar and tetrahedral structures for tetra-coordinated rhodium(0) and iridium(0) complexes is small. Because of the good agreement of the simulated and independently determined data for reactions (9) and (14), the data for the synproportiod disproportion equilibrium (12) can be regarded with some confidence. Noteworthy, the hump which is clearly observed between the first and second wave in the SW voltammograms at low frequency (25 Hz) can only be reproduced when the homogeneous electron transfer reaction (12) is taken into account. A rate constant kf = khomo= 5 x 10' [M-' s-'1 is obtained.23The equilibrium constant K for reaction (12) is simply extracted from the potential difference AE = Eo2- E"' = -0.27 V for the two redox processes given by eqs. (10) (E"' = -0.65V) and (11) (Eo2 = -0.92 V) (i.e. 1nK = -(nF/RT)AE). The disproportion constant Kdjsp (see Scheme 5), given by 1K = 2.7 x is small. However, the electron transfer between 12 and 18 is sufficiently fast, kb = kdisp = 1.4 X 104 [M-' s-'1 allowing the assumption that the paramagnetic 17-electron complexes [M(troppPh)2] 0 are not directly involved in the reactions shown in Scheme 5 but their closed-shell equilibrium partners are. A similar conclusion, that d9-[ML4 complexes are quite unreactive indeed, has been drawn by Eisenberg et al. some time ago.

lp,

2.5. Structures of [M(R'troppR)2]and [ML(troppR)2]complexes. As representative examples, structures of tetra-coordinated 16-electron, 17-electron, and penta-coordinated 18-electron complexes are shown in Figure 2. Selected bond lengths and angles are compiled in Table 5 and Table 6, respectively.

232

Perspectives in Organometallic Chemistry

Figure 2. (A): Structure of trans-[Rh(tropppvJ'; (B): Structure of cis-fRh~troppph)2]'; (C): Structure o IrCl tropppy2];.(D): Structure of cis-[Rh(tropppv2] (E): Structure of trans-[Ir(troppP"0 )2] . ( As the data show, the tropp ligand system is remarkably rigid. Neither the M-P nor the metal to olefin distances, nor the inner-ligand angles a1 and a3 vary significantly with the formal oxidation state or the coordination number at the metal center. Small differences follow the expected trends, i.e. shorter metal to olefin distances M-Ct' and M-Ct2 (Ct = centroid of the coordinated C=Cw,,,,, bonds) go hand in hand with longer C=Cbopbonds indicating a higher degree of metal to olefin back-bonding. Notably, this effect is as

New perspectives for oleJin complexes

233

pronounced in the penta-coordinated 18-electron com lex [IrCl(troppPh)2](Figure 2C) as in the tetra-coordinated 17-electron complex [Ir (troppPR)21° (Figure 2E) (see entries 5 and 7 in Table 5 ) . Table 5. Selected bond lengths [A] for tetra-coordinated 16-electron [My'tropppp2]' complexes (M = Rh, Ir), the 18-electron complex [IrCl(tropppv2], and the 17-electron complexes [M(tr0ppph)27(M = Rh, Ir). The centroids Ct' and C? indicate the midpoints of the coordinated C' =C? and C 3 = 6trop bonds, respectively. Fig. 2 compound M-P', M-Pz M-Ct', M-C? C'=Cz, C3=C4 1. tranS-[Rh(trOpp~h)2]PF63 2.30,2.30 2.15,2.15 1.40, 1.40 2. A) tran~-[Rh(tropp~~)2]BF4 2.30,2.30 2.09,2.09 1.41, 1.41 3. B) ~ i s - [ R h ( ~ ~ t r o p f ) 2 ] P F2.23,2.26 ~~ 2.25,2.33 1.37, 1.38 2.31,2.30 2.07,2.07 1.40, 1.44 2.32,2.34 2.04,2.08 1.45, 1.43 2.27,2.27 2.13,2.09 1.41, 1.43 2.06,2.05 1.46, 1.46 2.28,2.27 Table 6. Bond angles a1 - a around the metal center M (M = Rh, Ir) in [Mp'tr0pp~~)2]~ complexes (n = +I, 0). Under the constraint that two of these six angles are confned to 90" due to the rigid chelating ligand, Zsv O = (4 x 900) + (2 x 1800) = 720"for a square planar (sqp) and Ztet O = (2 x 900) + (4 x 1200) = 660"for the flattened tetrahedron (tet) with D2d symmetry. The degree of distortion p PA) by which the planar structure approaches the flattened D2d structure is given by: p (%) = [(720" X 100]/60"; with 2& O - &el O = 604 The inner-chelate angles a1 and a3 are shaded in grey.

trans

Compound 1. kanS-[Rh(trOppph)2]PF6 2. trans-[Rh(troppph)2]BF4 3. cis-[Rh(MetropP h ) 2 1 ~ ~ 6 4. trans-[Ir(trop gh)2]PF6 5 . cis-[Rh(trop h! )2] 6. cis-[Ir(troppF h)2] al: Ctl-M-P1;a2: P1-M-Ct2;

cis

91.9 92.4 150.0 93.6 151.3

180 171.1 94.2 169.7 98.4

180 146.7 101.0 143.5 95.3

720 (0.0) 683.0 (61.7) 675.5 (74.2) 679.2 (68.0) 673.3 (77.8)

Larger variations are observed in the degree of distortion from a square planar (sqp) towards a distorted tetrahedral structure. Taking into account that two of the six angles which define the coordination sphere around the metal are restricted to approximately 90°, the highest symmetrical tetrahedral structure which is possible is one O f D2d symmetry with four 120" and two 90" bond angles (see the Figure above Table 6 for details),. This allows to express the degree of distortion cp as is indicated in the Table caption. Note that the structure of the 16-electron cation trans-[Rh(troppPh)2]' is perfectly planar in the crystal lattice of its PF6- salt (not shown in Figure 2) while changing the counter d o n to BF4-

234

Perspectives in Organometallic Chemistry

leads to a distorted structure (9= 61.7%) shown in Figure 2A (see entries 1 and 2 in Table 6). An even stronger distortion (Table 6, entry 3, cp = 74.2%) is observed in the cisconfigured cation ~is-[pRh(~~tropp~~)2]+ in its PFC salt (Figure 2B). Also the iridium complex trans-[Ir(tropp h)2]' (Table 6, entry 4, not shown in Figure 2) shows a structure which is closer to a tetrahedral form than to a square plane and the structural difference to the paramagnetic 17-electron [Ir(troppPh)2I0being even slightly flatter is very small. Note in this respect, that the anion [Ir(troppPh)2]-shows a distortion of cp = 82.7%, i.e. the distortion towards a tetrahedral structure is more pronounced, but still the relatively similar structures of all [Ir(tr0pp~~)2]'>~,complexes may explain the high rate of the electron transfer reaction (I 2).

2.6. Paramagnetic [16+1] electron rhodium and iridium tropdad complexes One reason for the low reactivity of the 17-electron [M(troppPh)2]0 discussed above may be the fact that they are electronically saturated and the addition of a further ligand to give a 19-electron complex is strongly disfavored. Indeed from the SW experiments described Ph 0 above, it can be concluded that the equilibrium constant for the reaction [Ir(tropp )2} + acn rS [Ir(a~n)(tropp~~)2]~ must be smaller than K << [see eq. (14) in 2.41. We focused on the synthesis of compound 23 (Scheme 6) which contains both, the trop unit and a diazabutadiene moiety (dad) as a "non-innocent" ligand, i.e. a ligand which contributes actively to the redox state of the complex.26

&

0?)q2, NH,(l)/thf

-400c- r.t.

NH2

H

(p

O

@ HCOOH (cat.)

-

-

7

21

'f-H

EtOH

23

- 2 Cod

-

-wH H 24 M = R h 25 M = l r

1 models

for DFT

Zn W Zn(O,SCF,), H H 26 M = R h 27 M = l r

I

['(chtdad)l M = Rh, Ir

Scheme 6.Synthesis of [M(tropdad)]" complexes 24 - 27 (M = Rh, Ir; n = + I , 0).

New perspectives for olefin complexes

235

This potentially tetradentate ligand, named 5ropdad hereafter [IUPAC: 1,4-bis(5Hdibenzo[a,d]cyclohepten-5-yl)- 1,4-diazabuta-1,3-diene], is easily obtained in excellent yields by a condensation reaction of tropamine 21 [IUPAC: (5H-dibenzo[a,d]cyclohepten5-yl)amine] with glyoxal 2Z2' Upon reaction with the complexes [M(cod)2]03SCF3 (9: M = Rh; 10: M = Ir) in thf, deep red solutions are obtained from which the 16-electron 9ropdad complexes 24 and 25, respectively, precipitated as intensely red colored crystals in almost quantitative yields. Both complexes, 24 and 25, are (quasi)reversibly reduced to the neutral complexes 26 and 27, respectively, at the lowest potentials reported to date for any 16-electron rhodium or iridium complex (26: E1x = -0.56 V; 27: E 1 x = -0.35 V referred vs. Fc/Fc+). Chemically, the reduced 17-electron species can simply be prepared in almost quantitative yields by stirring a thf solution of the cationic precursors with an excess of zinc powder. No single crystals suitable for an X-ray analysis were obtained and only micro-crystalline powders were obtained so far. However, information about the electronic structure of 26 and 27 was obtained from continuous wave (CW) and pulse EPR spectroscopy, combined with DFT calculations.l 5 The X-band (9.7 GHz) Davies-ENDOR spectra of [Rh(%opdad)] 26 (b) and [Ir(?ropdad)] 27 (a) are shown in Figure 1 for two observer positions where orientations around gl are selected (see inserts).

'H 1

I

0

5

10

15

20

25

30

Figure 1: X-band Davies-ENDOR spectra measured at 20 K of the (a) Ir complex 27 and (b) Rh complex 26. Marked in (b) is the signalfiom Io3Rhwhich is centerd at half the hyperfine coupling A/2 (=6 MHz) and separated by twice the Larmor fiequency 2 v (strong coupling case 1A1>21 vv, and the ' H signal which is separated by A ( = I 5 3 MHz) and centerd at VI (weak coupling case IA1<21 v&. Inserts: EPR spectra showing the field positions where the ENDOR spectra were measured.

236

Perspectives in Organometallic Chemistry

Both spectra show a large proton hyperfine coupling (A = 15.3 MHz) which is assigned to the diazadiene protons. Note that in addition the [Rh(Htro dad)] complex 26 shows a hyperfine coupling of A = 12 MHz which is assigned to thehR'" nucleus. A representative Q-band (35.3 GHz) HYSCORE spectrum of 27 obtained when orientations close to g l are selected is shown in Figure 2.

\ I

-20

B

I

-10

0

I

I

10

20

Figure 2: @band HYSCORE spectra of the Ir complex 27 measured at 20 K in thJ Insert: EPR spectrum showing the field position where the HYSCORE spectrum was measured.

Since the nitrogens are strongly coupled, the cross-peaks representing the correlations of the nuclear frequencies in the two electron spin manifolds are predominantly in the second quadrant, as shown on the left side of the Figure.28The hyperfine coupling is found to be strongly anisotropic (Al= 34 MHz, A2,3 = 0.9 MHz) which is manifested by two long ridges.

Table 7. Experimental hyperfine coupling constants [in MHz] of [M(%-opdad)]complexes 26 and 27 and calculated-values of [M(chtdad)] (M = Rh, Ir) 28 and 29 (see Scheme 6). DFT [Rh(chtdad)J (28) Experimental [Rh(Htropdad)J (26) Nuclei Aiso A1 A2 A3 Aiso A1 A2 A3 4.8 4.6 43.8 17.7 11.9 34 0.9 0.9 -18.4 -4.8 -15.4 -23.0 'HIa1 14.5 20.0 15.5 8.0 -0.01 0.05 0.00 -0.04 12 -6 10 103Rh 5.3 DFT [Ir(chtdad)] (29) Experimental [Ir(%opdad) J (27) 11.9 34 0.9 0.9 15.5 39.2 3.9 3.5 lHa 14.5 20.0 15.5 8.0 -15.7 -23.4 -18.6 -5.2 a absolute values are given for the experimental data. Errors were estimated to be A1 MHz. 26: giso = 2.0022, gi =1.9977, gll = 2.01 13 g values: 27: giso = 2.0024, g1=1.9870, gll = 2.0332 = 2.1% DFT spin densities: 28: p(N) = 29.5%(x 2), P(C'Cdad) = 16.1% (x 2), P(C=C~op) (x 4), p(Rh) = 2% 29: p(N) = 26.3%(X 2), P(C=Cdad) = 16.6%(X 2), P(C=C,p) = 3.2% (x 4), p(1r) = 3%

'w

New perspectives for olefin complexes

237

The HYSCORE spectra of [Rh(%opdad)] 26 (not shown) exhibit virtually the same nitrogen signals. Note that in addition signals assigned to rhodium couplings are present in the spectrum of 26. Combining the results from the CW EPR, ENDOR and HYSCORE experiments, the hyperfine couplings of nitrogens, protons, and the rhodium were obtained (see Table 7). These and the isotropic g values (26:2.0022;27: 2.0024) are typical for paramagnetic complexes with the unpaired electron located on the dad moiety.29 DFT calculations also show that the spin density is mostly located on the dad unit (see values listed in the bottom of Table 7) and is not significantly influenced by the metal. Consequently, the [M(%opdad)]’ complexes 26,27 (M = Rh,Ir) are best described by the Lewis-structure M” [Htropdad’] and their valence electron configuration is assigned as [16+1]?6hImportantly, 26 and 27 are stable in deoxygenated thf/ H20 for at least several days. Hence, the Htropdad ligand should allow to investigate the reactivity of an organic radical coupled functionally to an electronically and coordinatively unsaturated metal center, here a 16-electron rhodium or iridium center, even under protic conditions, a concept used widely by nature.30

3 CONCLUSIONS Ligands prepared on the basis of the 5H-dibenzo[a,d]cyclohetene-5-y1 (trop) “platform” proved to be ideally suited for the preparation of rhodium and iridium complexes with less commonly known valence electron configurations. The combination “trop-plus-phosphine” allows the synthesis of genuine rhodium(0) and iridium(0) complexes. To present, our and studies by others show that these species are less reactive than anticipated. Probably the electronic saturation of the metal centers impedes classical organometallic transformations. However, other applications - especially electron transfer chemistry - remain to be explored. The combination “trop-plus-diazabutadiene” allows the synthesis of the first water-stable neutral rhodium and iridium complexes with a [16+1]electron configuration. In these species an organic radical is functionally coupled to a transition metal center which opens up possibilities for bond activation chemistry. Although not yet fully understood, the olefinic units of the trop ligand play an important role for the stabilization of these formally low valent metal complexes. Possibly the rigidity of the ligands discussed in this work enforces the electron acceptor properties of the olefinic moieties such that all trop type complexes have significantly lower reduction potentials that the corresponding classical cyclooctadiene (cod) complexes. The trop unit also proved to be a valuable tool for mechanistic studies using NMR techniques. These aspects are under active investigation. This work was supported by the Swiss National Fund and the ETH Ziirich. We acknowledge generous support with chemicals by the Haarmann&Reimer GmBH.

References

1 J. Thomaier, S. Boulmak, H. Schonberg, H. Ruegger, A. Currao, H. Griitzmacher, H. Hillebrecht and H. Pritzkow, New J. Chew., 1998,21,947. 2 H. Griitzmacher, S.Deblon, P. Maire and H. Schonberg, DP 10159015.6. 3 H. Schonberg, S. Boulmak, M. Worle, L. Liesum, A. Schweiger and H. Griitzmacher, Angew. Chem., 1998,109,1492;Angew. Chem. Int. Ed. Engl., 1998,37,1423. 4 H. Griitzmacher, H. Schonberg, S. Boulm&, M. Mlakar, S. Deblon, S . Loss and M. Worle, 3: Chem. SOC.,Cem. Commun., 1998,2623.

238

Perspectives in Organometallic Chemistry

5 S. Deblon, L. Liesum, J. Harmer, H. Schonberg, A. Schweiger and H. Griitzmacher, Chem.

Eu. J., 2002,8,601. 6 T.A. Albright, J.K. Burdett and M.-H. Whangbo, in Orbital Interactions in Chemistry, Wiley, New York, 1984,304. 7 C. Bolder, N. Avarvari, H. Schonberg, M. Worle, H. Ruegger and H. Griitzmacher, Helv. Chim. Acta, 2001,84,3 127. 8 For the MEXICO program package see, A. D. Bain, G. J. Duns, Can. J. Chem., 1996, 74, 819; for the CIFIT program package see, r8b1 A. D. Bain, J. A. Cramer, J. Magn. Reson., 1996, A 118, 21, and 8c D. R. Muhandiram, R. E. D. McClung, J Magn. Reson., 1987,71, 187. 9 9a S. Forsen, R. A. Hoffman, J. Chem. Phys., 1963,39,2892; 9b R. A. Hoffman, S. Forsen, Prog. NMR Spectrosc., 1966, 1, 15. 10 That square planar tetra-coordinated trans/cis-[Ir(troppph)2]+ are very likely the intermediates in the ligand displacement reactions, indeed, is indirectly indicated by the exchange of the diastereotopic olefinic proton resonances in 14-C1 and 14-acn (see the white and black circles in Scheme 3) which is observed in the 'H NMR spectra. This observation requires either an intermediate with a mirror plane to participate in the equilibrium (7) or a highly fluxional behaviour of the penta-coordinated species [IrL(troppph)2](i.e. a double Turnstile rotation plus a Berry rotation). However, no such behaviour is seen in solutions of [IrCl(troppPh)2]or [IrH(troppph)2]in non-coordinating solvents. 1 1 B. Longato, L. Riello, G. Bandoli, G. Pilloni, Inorg. Chem., 1999,38,2818. 12 B. Longato, R. Coppo, G. Pilloni, R. Corvaja, A. Toffoletti, G. Bandoli, J. Organomet. Chem., 2001,637 - 639,710 - 718. 13 N. MCzailles, P. Rosa, F. Mathey, P. Le Floch, Organometallics, 2000,19,2941. 14 Only a selected number of references is given here. For a more complete list see ref.. 'la Reduction of [RhCl(PPh3)3]: D.C. Olson and W. Keim, Inorg. Chem., 1969, 8, 2028; Reductions of [MCl(dppe)2], [MX(dppe)COn], and [M(dppe)z]+ (M = Rh, Ir): ' l b J.A. Sofranko, R. Eisenberg and J.A. Kampmeier, J. Am. Chem. SOC. , 1980,102, 1 163 and cited lit.; G. Pilloni, G. Zotti and M. Martelli, Inorg. Chem., 1982, 21, 1283 and cited lit.; 'Id Reduction of [Rh(P(OiPr3)4]+:G. Pilloni, G. Zotti and S. Zecchin, J. Organornet. Chem. 1986, 317, 357; 'le [M(PPh3)3(L)]': G. Zotti, S. Zecchin and G. Pilloni, J. Electronanal. Chem., 1984, 175, 241 and cited lit.; 'lf Reduction of [Rh(cod)z]+: J. Orsini and W.E. Geiger, J. Elektroanal. Chem., 1995,380,83. 15 G. Frison, unpublished results. The calculations were performed with the methods described in: 15a Lee, C.; Yang, W.; Pan, R.G. Phys. Rev. 1988, B 37,785; 15b Becke, A.D. J. Chem. Phys. 1993,98,5648; 15' C. Adamo, V. Barone, J. Chem Phys. 1996,664. 16 Chemical COZ reduction with [Rh(Ph2P-(CHz)n'PPh2)2]- (n = 2, 3): B.BogdanoviC, W. Leitner, Ch. Six, U. Wilczok and K. Wittmann, Angew. Chem., 1997, 109, 518; Angew. Chem. Int. Ed Engl., 1997,36,500. 17 Hydroformylation: A.S. Chan, H.4. Shieh, Inorg. Chim. Acta, 1994, 218, 89, and cited lit.. 18 Electrochemical C02 reduction with [MCl(CO)(PPh3)2](M = Rh, Ir): A. Szymaszek, F.P. Pruchnik, J Organomet. Chem., 1989,376, 133, and cited lit.. 19 C-H bond activation: J.A. Sofranko, R. Eisenberg, J.A. Kampmeier, J. Am. Chem. Soc., 1980,102,1163. 20 Photochemical H20 reduction: S. Oishi, J. Mol. Cat., 1987,39,225. 21 A bimetallic LRhO-Rho catalyst has been used to generate H2 from HX: A.F. Heyduk, D.G. Nocera, Science, 2001,293, 1639. 22 M. Rudolph, J. Electroanal. Chem.,2001 ,503,15.

New perspectives for olejin complexes

239

23 Rate constants in this range are not unusual for electron transfer reactions between strcuturally closely related organometallic species, see for example: C.F. Harlan, T. Hascall, E. Fujita, J.R. Norton, J. Am. Chem. SOC.,1999,121,7274. 24 K.T. Mueller, A.J. Kunin, S. Greiner, T. Henderson, R.W. Kreilick, R. Eisenberg, J. Am. Chem. Soc., 1987, 109,6313, and cited lit.. 25 S. Deblon, H. Ruegger, H. Schonberg, S. Loss, V. Gramlich and H. Griitztnacher, New. J. Chem. 2001,23,83. 26 Typical "non-innocent" ligands are quinones, Q, their diimine derivatives, N-heterocyclic chelates, NHchel (bipyridine, phenanthroline, etc.), and 1,4-diazabutadienes, dad = RN=CR'-CR'=NR, There is an extensive literature on this topic and only a limited selection of papers is cited here: For reviews see:26aC.G. Pierpont, C.W. Lange, Prog. Inorg. Chem. 1994, 41, 331; 26b G. van Koten, K. Vrieze, A h . Organomet. Chem. 1982, 21, 151. - Electrochemical studies with N-heterocyclic chelate complexes: 26c [M(bipy)J and [M(phen)2] complexes (M = Rh, Ir): H. Ciildsiraru, K. DeArmond, K.W. Hanck, Y. Em. Sahini, J Am. Chem. SOC.1976, 98, 4455, and lit cited therein; 26d J.L. Kahl, K.W. Hanck, K. DeArmond, J Phys. Chem. 1978, 82, 540, and lit cited therein; 26e J.L. Kahl, K.W. Hanck, K. DeArmond, J. Phys. Chem. 1979, 83, 2611; and lit. cited therein; [M(NHchel)(diene)]+complexes:26fW.A. Fordyce, K.H. Pool, G.A. Crosby, Inorg. Chem. 1982, 21, 1027; and lit cited G. Costa, C. Tavagnacco, G. Balducci, G. Mestoni, G. Zassinovich, J. Elekfroanal. Chem. 1989, 261, 189; and lit cited therein. For the recent use of dad in stabilizing low-valent [16+1] Nb and Ta complexes, see:26hP.J. Daff, M. Etienne, B. Donnadieu, S.Z. Knottenbelt, J.E. McGrady, J. Am. Chem. SOC. 2002, 124, 38 18; and lit. cited therein. 27 C. Bohler, D. Stein, N. Donati, H. Griitzmacher, New. J. Chem., 2002,24, 1291. 28 A. Schweiger, G. Jeschke, Principles of Pulse Electron Paramagnetic Resonance, Oxford Press, 2001. 29 See the discussion about addressing the problem of assigning the redox state in the dad complexes [Ir"3Cp*Cl(dado)]'PF6 and [Ir+3Cp*(dad2-)]and citing previous work in the field: S. Berger, F. Bawnann, T. Scheiring, W. Kaim, Z. Anorg. Allg. Chem., 2001, 627, 620. 30 For related biologically relevant copper mediated reactions see: W. Kaim, J. Rall, Angew. Chem., 1996,107,47; Angew. Chem. Int. Ed Engl., 1996,35,43.

SUBSTITUTION AND ADDITION REACTIONS CATALYZED BY TRANSITION METAL COMPLEXES

I.P. Beletskaya Department of Chemistry, Moscow Lomonosov State University, Leninskie Gory, MOSCOW, 119992, Russia

1 INTRODUCTION Transition metal catalyzed reactions of aryl(viny1) halides or triflates and nucleophiles create not only a new way to form carbon-carbon bonds but also carbon C(sp2)-element bonds. It is particularly important for the synthesis of different kinds of phosphorous compounds and amines. 2 SYNTHESIS OF ORGANOPHOSPHORUS COMPOUNDS BY THE SUBSTITUTIONPROCEDURE 2.1

Formation of Awl Phosphonates by the Reaction of ArHal with Trialkyl Phosphites

The reaction of (RO)3P with AlkHal is a well known Arbuzov reaction leading to phosphonates AlkP(O)(OR)2 as a result of the rearrangement. In the presence of nickel catalyst this reaction can be performed with aryl or vinyl halides (Scheme 1)

Scheme 1 The reaction requires harsh conditions but gives high yields of the product. We used in this reaction not only (R0)3P but also (Me3SiOhP that makes the procedure easier.2 The reaction of perchloropyridine proceeds selectively to give only corresponding yphosphonate. In the case of P-cyanotetrachloropyridine the reaction proceeds also very selectively with the formation of a-phosphonate (Scheme 2).

Substitution and addition reactions catalyzed by transition metal complexes

24 1

1) NiC12

2) MeOH

N

N

+ ( Me3SiO),P or (AlkOhP Alk = Et, Bu

N

1) NiClz

2)MeOH

C1

Scheme 2 It is interesting to mention that the reaction (Scheme 1) had been discovered before the cross-coupling reaction between ArHal and organometallic compounds which was found in 1972. In spite of some common features between phosphorylation and cross-coupling reactions there are yet essential differences. It is known that Ni2+ forms the phosphite complex Ni[P(oR)3]~,3oxidative addition of aryl iodide to it gives ArNi[P(OR)3]21 from which reductive elimination produces the phosphoniwn salt ArP(OR)3+1 which is transformed into the product ArP(O)(OR)2. One can see that in this mechanism there is no transmetallation step and that the oxidative addition is followed by reductive elimination. 2.2

The Formation of Aryl Phosphonates by the Reaction of ArHal with Dialkyl Phosphites; Reactions of other Hydrophosphoryl Derivatives

A simple and easy way to the synthesis of dialkyl phosphonates is the Pd-catalyzed reaction of dialkyl phosphites with ArHal proposed by Hirao! We have shown that the ary lation of the hydrophosphoryl derivatives can be performed under phase-transfer conditions without any solvent for liquid ArHal and in MeCN for solid The reaction with ArI can be performed in phosphine-free manner but for ArBr it is better to use phosphine ligands (PPh3 or PFur3). The use of the water-soluble ligand Ph2P(C&S03Na-m) in the arylation of (Et0)2P(O)H also gives excellent yield of the product (Scheme 3).

Scheme 3 The reaction time can be decreased substantially if the reaction is carried out in aqueous solution' (e.g., MeCN/H20 = 1:l) or even in neat water with a water-soluble ligand.*For the water-soluble ArHal the reaction needs no phosphine ligand (Scheme 4).

Perspectives in Organometallic Chemistry

242

This reaction is certainly closer to the cross-coupling reaction with a traditional catalytic cycle including oxidative addition, transmetallation, and reductive elimination (Scheme 5 ) .

I '

ArPd-P(0)(OR)2

'

ArPdX

?+' HX base '

HP(O)(OR)2

Scheme 5

This mechanism proposes the formation of the anion P(O)(OR)Y under the action of the base.

2.3

The Formation of Vinyl Phosphonates by the Ni- and Pd-Catalyzed Reactions of Vinyl Halides with (RO)sP and (ROhP(0)H

The Ni- and Pd-catalyzed reaction of vinyl halides or triflates with (R0)3P or (R0)2P(O)H leads to the formation of vinyl phosphonate~.~~'~ We used this reaction for the synthesis of 1- and 2-alkoxy- and dialkylaminophosphonates (Scheme 6)."

Hal = Br, CI R=H,Me

R = H, Me, Br

The reaction of (EtO)*P(O)H with vinyl halides (even chlorides) having halogen at aposition to the OR- or NR2-group proceeds very easily even at room temperature and affords high yields of vinyl phosphonates.

Substitution and addition reactions catalyzed by transition metal complexes

243

The reaction of vinyl halides having halogen atom at the P-position needs high temperature and the yields are modest. The reaction with (Et0)3P in both cases requires high temperature (120-160°C) but the yields of both a- and P-phosphonates are high. 2.4

The Formation of Secondary and Tertiary Aryl Phosphines

The synthesis of phosphines can be performed using free starting phosphines or their Sn or Si derivatives.I2 We synthesized a variety of the secondary and tertiary aryl phosphines including water soluble phosphines using this procedure (Scheme 7).13 PdC12(MeCN)2 h(Het)x +

R:p-siMe3 H

20 - l0O0C

t

R, ,P-Ar(Het) H

70 - 90%

PdC12(MeCN), Ar(Het)X +

R'\ R2/

100 100°C

P--SiMe3

~ 1 ,

R2/p-Ar(Het)

75 - 90%

Scheme 7

Bis- and tris-silylsubstituted phosphines can be used in this reaction (Scheme 8). RF'(SiMe,)2

-t

R02C+13r

100 - 12OOC (R02C@2PR

-

H20

2.5

The Formation of Vinyl Phosphines

Functionalized vinyl phosphines were synthesized by the Pd-catalyzed cross-coupling reactions of a- and P-alkoxy (or dialky1amino)vinyl halides with Ph2PH or Ph2PSiMe3 (Scheme 9).'4715 E-bromostyrene reacts with Ph2PH with net retention of configuration. The a-halogen isomer is much more reactive than the p-isomer and gives high yields of the product (including a-chloro derivative) at room temperature. The P-bromo isomer needs 70- 120°C and a substantially longer time for the completion of the reaction.

244

Perspectives in Organometallic Chemistry

R' R2-0Alk

Hal

or B r ~ R 2 (or NR32) R OAlk (or NR32)

Ph2PH pdC~,(pph,),

*

benzene, 2OoC or 70 - 12OoC ___)

R' R' R' R'

= R2 = H,

Me R2 = Me = Me, R2 = Me3Si = H, R2 = Me3Si = H,

R2 = H, Ph

Scheme 9 3 THE SYNTHESIS OF PHOSPHORUS COMPOUNDS BY THE TRANSITION METAL CATALYZED ADDITION REACTION OF P-H AND P-P BONDS TO THE TRIPLE AND DOUBLE BOND

The transition metal catalyzed reaction of E-H and E-E' bonds to unsaturated compounds is the simplest and most efficient route to new carbon-element bonds. The addition reactions become more and more important because they principally can proceed without formation of wastes (even without salt formation as in the substitution processes) and thus meet the demands of green chemistry (100% atom efficiency). In the case of the addition across the triple bonds, the possibility to form tri- and tetrasubstituted olefins with new C(sp2)-E bond and syn-selectivity of addition is realized.I6 3.1

The Addition of (EtO)ZP(O)H to the Triple Bond

Tanaka was the first to show that dialkyl phosphite reacts with terminal alkynes catalyzed by the cis-Me2Pd(PPh2Me)~complex with the formation of a-substituted pho~phonates.'~ We showed that the reaction could be performed with simple Pd(0) complexes and could be a plied to the synthesis of a variety of a-aryl (hetaryl) vinyl phosphonates (Scheme 10). ~

1 9

=-

+

Ar (Het)

HP(O)C(OEt),

Pd2(dba)3CHC1, THF, 67OC

Ar (Het)

= Ph, pMeC6&,

=(p(o)(oEt)2 Ar (Het)

pPhC6&, pClC6&, 1-Nf, 2-Nf, 6-Me0-2-Nf, 2-Py,

4 -pY

Scheme 10 The reduction of these compounds leads to 1-aryl ethylphosphonates or phosphonic acids including a P-analog of naproxene (Scheme 11).

Substitution and addition reactions catalyzed by transition metal complexes

245

HCOONH,, 5% Pd/C

Ar (Het)

refl. in MeOH or H20

R = Et, H

Scheme 11 The asymmetric reduction waas also studied.20.21

3.2

Hydrophosphination of the Triple Bond

Recently we have shown that hydrophosphination of the triple bond that proceeds by the radical mechanism non-selectively can be performed regio- and stereoselectively using Pd or Ni catalysis?2 A surprising phenomenon was observed in the beginning of the study when we found that the application of Pd(0) or Ni(0) led to the formation of p-isomer (as a major or the only one isomer) but the use of Pd" or Ni" as precursors of the catalyst led to a-isomer (again as a major or the only one product). One exception is tert-butylacetylene where p-isomer is formed. We proposed that the difference in regioselectivity is due to the formation of HX during the reduction of M2+under the action of PhZPH. And indeed, in the presence of the catalytic amount of the acidic compounds such as HOAc for Pd(0) or H(O)P(OEt)2 in the case of Ni(0) we changed completely the regioselectivity fiom p- to aisomer (Scheme 12). R=

+ PhzPH

- -(""' [Nil or [Pdl

+

R = Ph, n-Pr, n-Bu Ph-Ph

+

Ph2PH

[Nil

/dPPh2 R

R

H Ph

Scheme 12 We think that, in the case of p-isomer formation, we observe the oxidative addition of the phosphine to M(0) with the formation of phosphide complexes. The insertion of the alkyne into the M-H bond gives the intermediate with the MPPh2 fragment in the pposition for steric reasons and the p-isomer results after the reductive elimination. In the presence of HX another catalytic cycle operates including oxidative addition of HX to M(0) and Markovnikov addition of HMX to the triple bond, which is followed by the cross-coupling reaction with Ph2PH and the formation of the a-isomer after reductive elimination. It is worth mentioning that the reaction proceeds in the absence of base. In the majority of cases Ni catalysts demonstrate better results than Pd catalysts (Scheme 13).

Perspectives in Organometallic Chemistry

246

4-

Ph2PH

Ph2PH

Hx

’

4 Ph2P R *

Scheme 13 3.3

Diphosphination of the Triple Bond

A new reaction of diphosphination of terminal alkynes by Ph2PPPh2 catalyzed by Ni or Pd complexes was proposed by us for the synthesis of alkylidene bi~(phosphine).~~ Chelating

bisphosphine ligands play a very important role in transition metal catalyzed reactions. Of substantial interest is the synthesis of alkylidene cis-bisphosphines having different groups in a backbone. We have shown that a variety of terminal alkynes react with Ph2PPPh2 in benzene or toluene in a sealed tube at 130°C using NiBrz(PPh3)~or PdC12(PPh3)2 as precursors of the catalyst. Ni complex is more efficient: the reaction with phenylacetylene was completed in 45 h (95%) in benzene while the reaction with Pd complex needs 125 h for completion (92%). It is interesting to note that CoBr2 or Co(acac)3 can also catalyze the reaction. The addition proceeds in a syn-manner to give high yield of the products with a number of aryl (hetaryl) acetylenes. The only exceptions are p-dimethylaminophenylacetylene and a-pyridylacetylene which afford low yields (20% and 40%, resp.). Alkylacetylenes are less reactive than aryl acetylenes. No reaction was observed for internal acetylenes (Scheme 14). R-

+

Ph2PPPh2

[Nil benzene or toluene 80 - 13OOC

ph2pPpph2 R

R = Ph, O-MeOCgHq, p-MeOC6H4, m-CF,C&, p-CNC&, p-Me2NC6H,, a-naphthyl, a-thienyl, 2-Py, 4-Py,6quinolyl, n-C5H,

Scheme 14 A catalytic cycle including the oxidative addition of P-P bond to M(0) with the formation of diphosphide intermediate and with the insertion of the alkyne to M-P bond can be proposed. Reductive elimination leads to the product (Scheme 15).

Substitution and addition reactions catalyzed by transition metal complexes

247

+I

Ph2P-WPPh2

I R

e

Scheme 15 3.4

Hydrophosphination of the Double Bond

The double bond is less active than the triple bond in the transition metal catalyzed addition reaction. Metal-catalyzed hydrophosphination is known only for Michael olefins such as acrylonitrile or ethyl a ~ r y l a t eWe . ~ ~have found that styrenes or their heteroanalogs, which can be considered as weakly activated olefins, react with Ph2PH in the presence of Pd or Ni catalyst. The best catalyst is Ni[P(OEt)3]4 in benzene or toluene to give the addition product in a high yield (Scheme 16).25 Ar(Het)+

+ HPPh,

[Nil or [Pd] benzene or toluene, 13OoC

Ar (Het) ePPh2

Ar = Ph, o-M&C&, p-MeOCbH4, p-ClC& Het = 2-Py, 4-Py,2-Me-3-4.

Scheme 16 The reaction leads to the formation of only the p-isomer. A side reaction is the oxidation of Ph2PH into Ph2PPPh2 which can be suppressed by the addition of Et3N. The formation of Ph2PPPh2 means that even with M(0) the reaction of the following type proceeds (Scheme 17):

Scheme 17 Simultaneously we can observe the reduction of styrene into ethylbenzene. One can suggest a catalytic cycle with the oxidative addition of Ph2PH to M(0) and the insertion of styrene into the M-H bond in a more sterically favorable manner (Scheme 18).

Perspectives in Organometallic Chemistry

248

Ph2P-M-H

ArpMPPh2

Ar-

Scheme 18 4 THE FORMATION OF C-N BONDS

4.1

Arylation of Benzotriazole and Tetrazoles

Earlier we have shown that benzotriazole which forms a rather stable N-anion can be directly arylated in the Pd-catalyzed process b aryl iodides and aryl bromides under phase-transfer conditions or by Ar2IBF4 in The yield is high but two isomers are formed. In the presence of a catalytic amount of Cu" only the N'-isomer is formed selectively (Scheme 19).

+

a : J N

Ar(Het)Hal or Ar21BF4

-

Ar (Het)

I

WO), Cu" Ph,P or TPPTS

80 - 95%

Scheme 19 The arylation of tetrazole is even more difficult because the anion is more stable. Nevertheless, we have found that selective N'-arylation can be carried out using a diaryliodonium salt and a Pd 0 -Cu" catalytic system in the presence of BINAP as a supporting ligand (Scheme 20).

is>

tun=

(ph/do) 2 cu2+

Scheme 20

Substitution and addition reactions catalyzed by transition metal complexes

249

The corresponding Sn salt can be arylated in the same manner using only copper as a catalyst (Scheme 2 l)F9

65 - 73%

I

SnBu3

I Ar

R = Ph, p-MeC6H4,p-BrCsH4, 3 4 lH-indol-3-yl) no reaction with R =p-N02C6H4, 4 - Py

Scheme 21 4.2

Vinylation of Monoazoles and Phenothiazine

N-vinylazoles can serve as monomers for the synthesis of poly(N-vinylmoles) which can be used as semiconducting and photosensitive materials. However, while the arylation of moles according to the Buchwald-Hartwig procedure is well documented, the corresponding vinylation is problematic because of competitive elimination under the action of a strong base t-BuONa. Nevertheless, in some cases the reaction proceeds quite smoothly,3° such as the reaction of trans-P-bromostyrenewith indole (Scheme 22).

F Y" + N H

R' BrAA

Pd(dbah 12L

*

t-BuOM toluene I DME 8OoC

r3R N $!a

R'

Scheme 22 The reaction proceeds with the retention of configuration. To avoid elimination, we used tBuOLi as base. This allows one to carry out vinylation of moles (pyrrole, indole, carbazole, and their derivatives) and phenothiazine by a variety of alkenyl bromides with the yields of the products fiom good to excellent. 4.3

The Formation of New Polyazacycles through the Arylation of Polyamines

Recently we have shown that polyamines of the type H~N(CH~)INH(CH~)JW~(CH~)~NH~ or H2N(CH2)lO(CH2),O(CH2)nNH2 react with 1,s- or 1,5-dichloroanthracenes and with 1,8- or 1,5-dichloroanthraquinonesusing the Pd(dba)a/BINAP - t-BuONa system in dioxanee31732 A direct route to polyaza and polyoxapolyazamacrocycles presents itself by this reaction (Schemes 23 and 24).

Perspectives in Organometallic Chemistry

250

+

or

w 0

Pd(dba)2 / BINAP 4 - 8 mol% / 4 - 8 mol% NaOtBu or Cs2C03 reflux in dioxane 21 - 36%

Scheme 23

w 0

Pd(dba), / BINAP 4 8 mol% / 4 8 mol%

-

-

NaOtBu or Cs2C03 reflux in dioxane

Scheme 23 Amination of 1,8-dichloroanthracene by cyclic tetramine - N,N',N"-trimethylc yclam and 1 -a1 5-crow-5 gave corresponding bismacrocyclic molecules with an anthracene spacer which possess interesting complexing abilities (Scheme 24)?3

Substitution and addition reactions catalyzed by transition metal complexes

25 1

Me

("."")

u

N H N Me'

+

NaOtBu dioxane

n

or

n

C"3 A

'Me

Pd(dba)2/BINAP 8 - 16 mol%

Pd(dba)*/BINAP NaOtBu dioxane

Scheme 24 References 1 2

3 4 5

6 7

8 9 10 11 12 13 14 15 16 17 18 19

P. Tavs, Chem. Ber., 1970,103,2428. N.N. Demik, M.M. Kabachnik, Z.S. Novikova and I.P. Beletskaya, Im. AN SSSR, Ser. Khim., 1991, 1461. N.N. Demik, M.M. Kabachnik, Z.S. Novikova and I.P. Beletskaya, Im. AN SSSR, Ser. Khim. , 1992,2432. T.M. Balthazor and R.C. Grabiak, J. Org. Chem., 1980,45,5425. T. Hirao and T. Masunga, Synthesis, 1981,56. M.M. Kabachnik, M.D. Solntseva and I.P. Beletskaya, Russ. Chem. Bull., 1997, 46, 1491. I.P. Beletskaya, M.M. Kabachnik and M.D. Solntseva, Zh. Org. Khimii, 1998,34, 106. The arylation (Et0)2P(O)H in MeCN-H20 (1:l) was demonstrated using 10 mol% Pd[Ph2P(m-C&S03Na)]3: A. Casalnuoro and J. Calabrese, J. Am. Chem. SOC., 1990, 112,4324. M.M. Kabachnik, M.D. Solntseva and I.P. Beletskaya, Zh. Org. Khimii, 1999,35,79. P. Tavs and H. Weitkamp, Tetrahedron, 1970,26,5529. T. Hirao, T. Masunga, J. Okshiro and Tagawa, Tetrahedron Lett., 1980,21,3595. M.A. Kazankova, I.G. Trostyanskaya, S.V. Lutsenko and I.P. Beletskaya, Tetrahedron Lett., 1999,40, 569. S.E. Tunnay and J.K. Stille, J. Org. Chem., 1987,52,748. Yu.A. Veits, N.B. Karlstedet and I.P. Beletskaya, Zh. Org. Khimii, 1994,30,66. M.A. Kazankova, E.A. Chirkov, A.N. Kochetkov, I.V. Efimova and I.P. Beletskaya, Tetrahedron Lett., 1998,39,573. Yu.A. Veits, N.B. Karlstedt and I.P. Beletskaya, Tetrahedron Lett., 1995,36,4121. I.P. Beletskaya and C. Moberg, Chem. Rev., 1999,99,3435. L.-B. Han and M. Tanaka, J. Am. Chem. SOC., 1996,118,1571. N.S. Gulyukina, T.N. Dolgina, G.N. Bondarenko, I.P. Beletskaya, J.-C. Henry, D. Lavergne, V. Ratovelomanana-Vidal and J.-P. Genet, Zh. Org. Khimii, 2002,38, 600. A. Alleen and D.R. Manke, Tetrahedron Lett., 2000,41, 151. It has been shown that the reaction of alkynes having an aryl group with electronwithdrawing properties with (EtO)2P(O)H leads to bis-hydrophosphorylation.

252

20 21 22 23 24 25 26 27 28 29 30 31 32 33

Perspectives in Organometallic Chemistry

However, we have found that the reaction can be performed as monophosphorylation with the formation of a-isomer as a main product. N.S. Goulioukina, T.N. Dolgina, I.P. Beletskaya. J.-C. Henry, D. Lavergne, V. Ratovelomanana-Vidal and J.-P. Genet, Tetrahedron: Asymmetry, 2001,12,3 19. J.-C. Henry, D. Lavergne, V. Ratovelomanana-Vidal, J.-P. Genet, I.P. Beletskaya and T.M. Dolgina, Tetrahedron Lett., 1998,39, 3473. M.A. Kazankova, I.V. Efimova, A.N. Kochetkov, V.A. Afanas'ev, I.P. Beletskaya and P.H. Dixneuf, Synlett, 2001,497. M.A. Kazankova, V.A. Afanas'ev and I.P. Beletskaya, unpublished results. D.K. Wicht, I.V. Kourkine, B.M. Lew, J.M. Nthente and D.S. Glueck, J. Am. Chem. Soc., 1997,119, 5339. M.O. Shulyupin, M.A. Kazankova and I.P. Beletskaya, Org. Lett., 2002,4, 761. I.P. Beletskaya, D.V. Davydov and M. Moreno-Manas, Tetrahedron Lett., 1998, 39, 5617. I.P. Beletskaya, D.V. Davydov and M. Moreno-Manas, Tetrahedron Lett., 1998, 39, 5621. I.P. Beletskaya, D.V. Davydov and M.S. Gorovoy, Tetrahedron Lett., 2002, 43, in press. I.P. Beletskaya, D.V. Davydov, B.B. Semenov and Y.I. Smushkevich, Tetrahedron Lett., 2002,43, in press. A.Y. Lebedev, V.V. Izmer, D.N. Kazyul'kin, I.P. Beletskaya and A.Z. Voskoboinikov, Org. Lett., 2002,4,623. I.P. Beletskaya, A.D. Averin, A.G. Bessmertnykh and R. Guilard, Tetrahedron Lett., 2001,42,4983. I.P. Beletskaya, A.D. Averin, A.G. Bessmertnykh and R. Guilard, Tetrahedron Lett., 200 1,42,4987. I.P. Beletskaya, A.D. Averin, A.G. Bessmertnykh, F. Denat and R. Guilard, Tetrahedron Lett. ,2002,43, 1 193,

LATE TRANSITION METAL (Co, Rh, Ir) - SILOXIDE COMPLEXES - SYNTHESIS, STRUCTURE AND APPLICATION TO CATALYSIS

B. Marciniec'*, I. Kownacki', M. Kubicki2,P. Krzyzanowski', E. Walczuk' and P. Blaiejewska-Chadyniak' 'Department of Organometallic Chemistry, Adam Mickiewicz University, 60-780 Poznah, Poland 2 Department of Crystallography, Adam Mickiewicz University, 60-780 Poznah, Poland

1 INTRODUCTION Molecular compounds incorporating TM-0-Si groups (M = TM) are of great interest, particularly as models for metal complexes immobilized on silica and silicate surfaces known to catalyze a variety of organic transformation^.'-^ According to a general idea presented by Wolczanski,4 alkoxide and siloxide ligands are alternative to cyclopentadienyl complexes of transition metals. Alkoxide or siloxide ligands bind preferably through a o-type orbital such as the sp hybrid and via x-donation of two px orbitals, perpendicular to the M-0 direction as illustrated in Fig. 1 '.

E = C,Si M = transition metal

Figure 1 Bonding of transition metal with alkoxide and siloxide ligands However, the strength of both CF and x-interactions depends on the electrophilicity of the metal center. Thanks to a more positive character of a silicon atom when compared to a carbon one, siloxide ligands are generally less basic than alkoxides and therefore, can bind to a TM with slightly more ionic character. However, the well known d,-pn (or o*-p,) bonding as an interaction of fairly low-lying empty 3do* molecular orbitals with px orbitals of oxygen describes the x-accepting capability of silicon in the Si-0 bonding, which can weaken the p,-d, bonding of 0-TM? It is worth emphasizing that inductive and steric effects of substituents at silicon can additionally influence the stereoelectronic properties of a TM-0-SiR3 system. The latter properties of siloxide as ancillary ligand in the system TM-0-SiR3 can be effectively utilized in molecular catalysis particularly by early transition metal complexes. A combination of high reduction potential and high electrophilicity of TM and steric properties of the siloxy ligands is responsible for the

254

Perspectives in Organometallic Chemistry

unusual reactivity of low-valent TM-siloxide derivatives. Mono- and di-substituted branched siloxy ligands (.e.g. incompletely condensed silsesquioxanes) have been employed as more advanced models for the silanol sites on silica surface for catalytically active centres of early TM (Ti, W, V) which could be effectively utilized in polyrneri~ation,~metathesis,6 epoxidation of alkenes' and dehydrogenative coupling of silanes.' Contrary to the situation for early TM-siloxides, the respective information on the late TM-siloxy complexes is scarce.' Therefore, the aim of this paper is to give an overview of the synthesis, structure, reactivity and some applications to catalysis of group 9, i.e. cobalt(I), rhodium(1) and iridium(1) - siloxide complexes. 2. SYNTHESIS, STRUCTURE, REACTIVITY AND CATALYTIC ACTIVITY OF DIMERIC M-SILOXIDE COMPLEXES (M = Co, Rh, Ir) The reaction of [(diene)MC1]2 (where M = Rh, Ir) with sodium trimethylsilanolate has proven to be a general method for the preparation of the corresponding siloxide complexes, according to the following equation: [(diene)MCl]z + 2 R3SiONa + [(diene)M(pOSiR3)]2 + 2 NaCl R = Me, Ph, M=Rh(I), Ir(I)?l2 diene = 1,5-cyclooctadiene'~'oy12or norborna-2,5-diene.''

(1)

Carbonyl derivatives were synthesized by replacement of cod ligand in the complex according to the equation,' [Rh(cod)(pOSiPh3)]2 + 4 CO

+ [Rh(CO)~(pOSiPh3)2]2+ 2 cod

(2)

The complexes were characterized by 'H, 13C,31Pand 29SiNMR spectroscopies. All the complexes prepared have A-frame bis-square planar geometry see exemplary iridiumsiloxide complex, whose structure has been determined - Fig.2.(2

@Ck

Figure 2 Crystal structure of (cod)Ir((+OSiMe3)2

Late transition metal (Co, Rh, Ir)-siloxide complexes

255

It is worth noting that also cobalt (yet Co(I1)) forms dimeric species with siloxide ligands. l3 [Co(p-OSiPh3)(0SiPh3) THF]2 was synthesized via a cobalt silylamide complex according to the following equation (yield 47%).'3a [pN(SiMe3)2CoN(SiMe3)2]2 + 4 Ph3SiOH -+ [{Co(p-OSiPh3)(0SiPh3)THF}2]+ 4 HN(SiMe3),

(3)

The TM-O(Si) bond in the bridging siloxy groups in these complexes is always longer than in the terminal ones. In Table 1 the distances and the M-0-Si bond angles in dimeric siloxides of 9 group are presented. There are only few reports of structure determinations of dimeric M-siloxide complexes (M = Rh,Ir, Co). In four-coordinated cobalt complexes the Co have a distorted tetrahedral ge~metry,'~ while the coordination of iridium and rhodium in all known structures is square-planar.87'o-'2 Table 1. Dimeric siloxide complexes Complex

Bond lengths [A]

A

PhSi

Co-O(Si) terminal 1.858(4), 1.854(4) bridging 1.993(3), 1.991(2) Si-O(Co) 1.630(2), 1.644(2), 1.647(2) Si-0 terminal 1.589(4), 1.595(4) bridging1.63 1(3), 1.633(3) Co-O(Si) terminal 1.793(7), 1.781(8) bridging 1.945(7), 1.957(7) Si-O(Co) 1.644(8), 1.64l(8) Si-0 Rh-O(Si) bridging 1.630 0-Si 1.637

Ref.

Bond angles

["I

13a

Co-0-Si terminal 161.3(2)

13b

Co-0-Si terminal 160.9(5)

8

Rh-0-Rh82.3 Rh-0-Si 136.1 0-Rh-0 79.1

'H 6 = 7.30-7.80(m)

10

Rh-0-Rh

Me3Si

I

Me3Si

Me3&

I1

Rh-O(Si) bridging 2.087(6), 2.096(6) 2.075(6), 2.088(5) Si-O(Rh) 1.617(6), 1.632(6)

Rh-O(Si) bridging 2.082(5), 2.073(5) Si-O(Rh) 1.632(5)

Ir-O(Si) bridging 2.033(3), 2.153(2) 2.141(2), 2.173(3) Si-O(1r) 1.651(4), 1.55l(4)

91.6(2), 90.9(2) Rh-0-Si 128.2(4), 139.4(3) 0-Rh-0 78.5, 79.0(2) Rh-0-Rh 85.2(2) Rh-0-Si 136.4, 127.9 0-Rh-0 79.4(3) Ir-0-Ir 80.40(9) Ir-0-Si 145.1(2) 0-Ir-0 78.96(10)

'H 6 = 0.15(s)

11 'H 6 = 0.31 (s) 29Si6 = 10.46 (s)

12 'H6 = 0.31(s) 29Si{'H} (INEPT) 6 = 17.18(m)

256

Perspectives in Organometallic Chemistry

The central M202 (M=Co, Rh, Ir) four-membered rings have a roof-shaped geometry. Generally, the roof angle (defined as the angle between the OM0 planes) decreases with the growing atomic number, the mean values are ca. 170' for Co (such large angle is probably connected with the tetrahedral coordination of Co), 130' for Rh and 120' for Ir complexes. This geometry leads to relatively short intramolecular M-M contacts, of 2.75 - 2.9& much shorter than for similar p-Cl complexes (3.5A and more). The data in Table 1 show that when the atomic number increases, also M-0 bond length increases. The differences in bond angles are not that easy to rationalize, as the bond angles are, in general, 'softer' parameters than the bond lengths, and therefore many additional factors may influence the actual values of bond angles. The two dimeric trimethylsiloxide complexes of rhodium (I) and iridium (11) were used in homocoupling of vinylsubstituted silicon compounds and their heterocoupling (silyl group transfer, trans silylation) with olefins, e.g. styrene. The catalytic, synthetic and mechanistic studies of the process catalysed by ruthenium, rhodium and cobalt complexes have shown that in the presence of the catalysts containing initially M-H or M-Si bonds (or those in which these bonds can be generated in situ) the reaction proceeds through the cleavage of the =C-Si bond of the vinyl-silicon compounds and the =C-H bond of the olefin (also the vinyl-silicon compound in the homocoupling). The reaction occurs as follows:l4 -CH -SiR3

R

H.~-c-~-H

F H ~

II

HCC ' I R

+

I

-

M-H,M-Si

.I. iH2

EH

+

+

YH SiR3

CH2

,ct,H 2 R

siR3

(€+a (4)

The catalytic cycle of this new type of silylolefin conversion involves the mi ratory insertion of olefin (or vinylsilane) into M-Si bond, where M = Ru,",'~ Rh,17 and Co * (and vinylsilane into the M-H bond), followed by P-H and -Si) transfer to the metal atom with elimination of phenyl(sily1)ethene (and ethene)368 (Scheme 1).

9

-p' R3si

+

where: M = Ru, Rh. Co; R = alkyl, aryl, alkoxyl, siloxyl R = Ph, alkyl, silyl

Scheme 1 Mechanism of silylative coupling of vinylsilanes with alkenes catalyzed by M-H and M-Si complexes

Late transition metal (Co, Rh, Ir)-siloxide complexes

257

The complex I was tested in catalytic homocoupling of vinylsubstituted silanes which occur in a wide temperature range, according to equation 5.19

The catalytic study, has shown that complex (I) appeared an active catalyst in the reaction of the majority of the vinyltrisubstituted silanes used, and yielded two isomers (1) and (2). The yield of bis(sily1)ethenes [(l) + (2)] was 60-71% (60°C, 24h) where: R3 = Me(OEt)z, Me2Ph, Me3 and 50% (9OoC,24h) where R3= Me(OSiMe3)z. Although I is less effective in the disproportionation of vinyltris(trimethylsi1oxy)silane than [RuC12(PPh3)3] and [RuHCl(CO)(PPh3)3], the yields of bis(methy1,siloxy) ethenes are similar to those reported for ruthenium catalyzed reactions, 2o but the relevant reactions can occur effectively at milder conditions (even 6OoC). The fact that the transformation of vinylmethyldi(trimethy1siloxy) silane gives almost exclusively E-isomer can be valuable information to establish the real conditions for crosslinking of poly(methy1,vinyl)siloxanes via the catalytic disproportionation. Results of the experiments performed by Brookhart et. a1.21 using CsMe~(cH2=CHSiMe3)2complex under thermolysis conditions (14OoC, cyclohexane-dl2) and in the 10-fold excess of vinylsilane provided the evidence for hydrovinylation of one of the coordinated molecule of vinylsilane (see eq. 5 ) followed by the insertion of the second molecule in the generated Rh-H bond, subsequent elimination of ethylene and reductive elimination of the two types of bis(silyl)ethenes.21

7-

-

Me3Si R [h M 1 /p.e3

/'

(C5Me5)RhY SiMe3

-

Me3Si/IRh1TSiMe3

ySiMe3 + =

Me3Si

Me3Si

+

=

(6)

On the basis of the Brookhart experiment, a catalytic scheme for disproportionation (homocoupling of vinylsilanes) was presented.l9 However, in order to find evidence for catalysis with I and I1 precursors initially having no M-H or M-Si bond, the reactions of s ene (andp-styrenes) with various vinylsubstituted silicon compounds have been tested.l2? The reactions occur according to the following equations:

T

F +4 SiR3

Ph

I

-

- R = allyl, aryl, alkoxy, siloxy

I1 -R3 = SiMe2Ph, Si(OEt)3, SiMe3

(I),(11)

FSiR3 + = Ph

Perspectives in Organometallic Chemistry

25 8

However, contrary to the I catalyzed process, instead of the silylative coupling with evolution of ethylene, iridium complex (11) catalyzes hydrovinylation (co-dimerization) of styrene with CH2=CHSi(OR)3 (where R = SiMe3, t-Bu) according to the equation:

/=

+

. ( I I ) _ Phw \ SiR3

JiR3

Ph

Steric hindrance of three bulky substituents at silicon stops the silylative coupling and promotes the co-dimerization (hydrovinylationprocess). In order to find the mechanism of activation of =C-H in styrene, some experiments with deuterium-labeled reagents were performed. Analysis of the reaction of HzC=CHSiMezPh with styrene-d8 catalyzed by both I and I1 complexes has allowed us to exclude the metallacarbene mechanism of the process examined, since the results of GCMS of the reaction mixtures in a very early stage showed exclusive formation of silylstyrene-d;r (and ethylene that was not analyzed) according to the following equation :12,16h22

M'=M+6

+

H\

/H

/c=c,,'

H

,' SiR3

hydride or silyl

-

complexes

D/c=c\ /D

D5C6,

SiR3

+

H\/c=c,/H H

D

M'=M+7

(9) Moreover, the reaction between styrene do and styrene d8 tested in the presence of the catalytic amounts of I and I1 yielded a mixture of styrenes (do, dl, d2, d3, d5, dg, d7) whose presence was confirmed by the GC-MS and 'H NMR methods. These experiments clearly showed that WD exchange took place. The effect of p-substituents in styrene was checked in the presence of I catalyst to reveal a promotion of the process by electron-withdrawing substituents. Yet, a general effect at the p-position of styrenes is combined with electronic, mesomeric and steric effects of the substituents at silicon. Nevertheless, a quantitative transformation of pchlorostyrene in the reaction with all trisubstituted vinylsilanes catalyzed by I (at 90°C) is a key point for the mechanistic consideration22of the silylative heterocoupling process of vinylsilanes with styrenes. To characterize the catalytic performance of M-0-SiMe3 complex, a series of pseudostoichiometric reactions between the initial dimeric (I, 11) complexes and both substrates was analyzed.'2,22 In this overview we present the reaction illustrating two experiments of I1 with vinylsilanes followed by introduction of styrene to such a mixture at the next step.5

[Ir] : [CH2=CHSi=]: [styrene] = 1 : 10 : 10

Late transition metal (Co, Rh, Ir)-siloxide complexes

259

[Ir] : [CH,=CHSi=] : [styrene] = 1 : 10 : 10

(1 1) All pseudostoichiometric reactions of I and I1 with styrene and vinylsubstituted silane provide evidence that the mechanism of the two reactions observed, i.e. silylative coupling (trans-silylation, silyl group transfer) and hydrovinylation (occurring only in the presence of I1 with three bulky substituents at silicon), involves metal-hydrogen intermediates. On the other hand, labeling study have provided a convincing proof for oxidative addition of the =C-H of styrenes and vinylsilane to the metal-siloxide complexes I and 11. A general mechanism of catalysis by I and I1 can be summarized in Scheme 2 I-

I

(cod)WSiR,

I

A M = Rh. Ir

M = Ir

OSiMe,

'P-

Scheme 2 Catalytic activity of =C-H of styrene and vinylsilane The very fast H/D exchange process occurring in I and I1 with styrene has allowed us to assume that the coordination of styrene to metal is responsible for the cleavage of dimeric metal siloxide complex to form (1) followed by an oxidative addition of =C-H of styrene to the metal. In the presence of vinylsilane as a second substrate, the latter initiates a catalytic process observed by GC-MS (see scheme - cycle A). The insertion of vinylsilane into M-H bond of ( 5 ) is followed by elimination of ethylene (p-transfer) to give the M-Si intermediate (7). According to the dissociative mechanism, the reductive elimination of (E)-silylstyrene takes place, regenerating (in an excess of styrene) complex (1). The pseudostoichiometric studies of I1 with vinylsilanes enabled us to consider a pathway via I1 + (2) + (8) + (9) initiating the hydrovinylation process (cycle B), (with R = bulky substituent at silicon), but a sequence of reversible reactions (2) + (8) + (9) + (3) -+( 5 ) must also be regarded when other vinylsilane is added prior to styrene. However, in the presence of I1 and a bulky vinylsilane, the reaction of formation of (9) is followed by insertion of styrene into the Ir-H bond and elimination of the product of hydrovinylation (cycle B). We can finally conclude that the reaction proceeds according to the non-metallacarbene mechanism involving M-H 16e intermediate, which is generated in situ via oxidative

260

Perspectives in Organometallic Chemistry

addition of the =C-H bond of the styrene andor vinylsilane coordinated to the metal atom. When rhodium complex (I) is used or when phenyl, methyl and alkoxy substituted vinylsilanes take part in the reaction, in the presence of iridium complex (11), the stereo-, regio- and chemoselective syntheses of styrylsilanes via the silylative coupling are observed. When bulky siloxy substituents at silicon in vinylsilane are present, than the reaction catalyzed by Ir complex (11) leads to stereo- and regioselective formation of (a-4 phenyl, 1-silyl- 1-butene via the hydrovinylation (co-dimerization). The stoichometric study of I1 with vinyltriethoxysilane and vinyltrimethoxysilane allowed us to reveal a new type of the reaction, i.e. alkoxy group transfer from silicon to iridium with a simultaneous transfer of a siloxy group from iridium to silicon. The reaction occurs according to the following equation:23

[Ir] : [CH2=CHSi(OEt)3] = 1 : 10

(12)

GC-MS analysis indicates a formation of vinyltriethoxytrimethyldisiloxane but additional experiments with higher concentrations of the substrates allowed us to isolate dimeric iridium-ethoxide complex whose structure was determined by the X-ray m e t h ~ . ~ ~ The dimeric siloxide complex I was used as precursor of the catalyst in the hydrosilylation of 1-hexene by triethoxysilane showing much higher catalytic activity than the respective chloro-rhodium(1) complexes.24Kinetic dependence of the reaction rate on the initial concentration of [Rh] and stoichiometric reactions of Rh-complexes with triethoxysilane and 1-hexene allow a distinction between the catalytic cycles of the reactions occurring in the presence of siloxy-rhodium vs. chloro-rhodium complexes. Direct reaction of triethoxysilane with I was followed by next elementary steps, in particular reductive elimination of disiloxane (Et0)3SiOSiMe3 which generates hydridorhodium(1) complex. The latter was suggested as an intermediate for the catalysis of the hydrosilylation by siloxy-rhodium complexes.24 3. SYNTHESIS, STRUCTURE, REACTIVITY AND CATALYTIC ACTIVITY OF MONOMERIC LATE TM SILOXIDE COMPLEXES (TM = Co, Rh, Ir) Two types of reactions can be used for the synthesis of the corresponding metal-siloxide complexes' [{M(cod)(pOSiMe3)}2]

+

2 PR3

r.t., 24h

2 [M(~od)(PCy,)(oSiMe3)~]

M = Rh, Ir [Ir(CO)CI(PPh&] + NaOSiMe,

(13) C6H12

6OuC, 5,7 [Ir(Co)(PPh,)2(osiMe,)1

+

NaCl

(14) In all complexes the coordination of rhodium25 and iridium26 is square plane, which is evidenced by the inspection of the bond angles and the least square plane ciculations. An exemplary molecular structure of an analog of the Vaska complex, i.e., [Ir(CO)(PPh3)20SiMe3]with the atom numbering scheme is depicted in Fig.3.26

Late transition metal (Co, Rh, Ir)-siloxide complexes

26 1

5

Figure 3 Crystal structure of Ir(C0) (PPh3)2OSiMe3

The cobalt(1) siloxide complex was synthesized in the reaction of [CoCl(PPh3)3] with sodium trimethyl~ilanolate?~This complex (orange-brown colour) occupies a special position of the space group P3 on the three-fold axis passing through the Co, 0 and Si atoms (see Fig.4). The coordination of cobalt is tetrahedral. The complex is extremely sensitive in solution towards traces of air (oxygen) and gets easily oxidized to Co(I1)-blue complex.

Figure 4 Crystal structure of Co(PPh3)OSiMe_t

There have been even fewer attempts at structural determinations of monomeric TM siloxide complexes (TM = Co, Rh, Ir) than of dimeric ones. In fact, besides our recent communications there has been only one report on the structure determination of (bis(odimethylaminomethylphenyl)- methylsilanol - N, 0) - dibromo -cobalt(II) (however, it is a chelate).229 28 Anyway, even from this small body of data similar conclusions can be drawn as for dimeric complexes. The Co complexes are tetrahedral, while Rh and Ir complexes

262

Perspectives in Organometallic Chemistry

are square-planar. Also, the M-O(Si) bond distances follow the same general rule, that they grow with increasing atomic numbers. The other bond lengths and angles of M-0-Si fragments are rather typical, however, in the case of [Co(PPh3)3(0SiMe3)]?’ due to the molecular symmetry (see Fig. 4) the Co-0-Si bond angle has an extremely large value of 1800. Siloxide complexes particularly of rhodium and iridium are effective in the hydrosilylation of allyl ether^^'"^ and esters3 with triethoxysilane and polyethers and a l k e n e ~with ~ ~hydrosiloxanes. In the overview an exemplary reaction of hydrosilylation of allyl glycidyl ether by triethoxysilane leading to glycidoxypropyltriethoxysilane - commercially important silane coupling agent - is presented. The reaction is catalyzed by Rh and Ir siloxides giving the hydrosilylation products with a very high yield, (with I the reaction occurs at room temperature) accompanied by traces of unsaturated silane - a product of the dehydrogenative silylation according to the following equation: 0

bo+ HSi(OEt)3 [cat] o b O - S i ( O Et ) 3

(15) The stoichiometric reactions of rhodium and iridium siloxides with substrates (hydrosilanes and vinylsilanes) and preliminary kinetic measurements are the basis of the mechanistic scheme proposed for the catalysis by examplary catalyst (~od)Ir(PCy3)0SiMe3.~~ OSiMe,

HSiMeR,

Scheme 3 Mechanism of catalysis of hydrosilylation by (cod)Ir(PCy3)OSiMe3

Similarly to the hydrosilylation by dimeric rhodium siloxides, monomeric phosphine complexes undergo oxidative addition with trisubstituted silanes followed by elimination

Late transition metal (Co, Rh, Zr)-siloxide complexes

263

of disiloxane. This reaction occurs even at room temperature and has been confirmed by GC-MS analysis of the product. However, the oxygenation of phosphine proceeded at enhanced temperatures to generate [(cod)M(H)(alkene)] (see Scheme 3). This 16e hydridemetal complex with already coordinated molecule of alkene seems to be a key intermediate in all catalytic transformations involving hydrosilanes, e.g. hydrosilylation, dehydrogenative silylation, etc. References

1 B. Marciniec; B, Maciejewski, H. Coord.Chem.Revs, 2001,223,301. 2 Y. Iwasawa, ed., Tailored Metal Catalysts, Reidel, Boston, 1986. 3 F. J. Feher, J Am. Chem. SOC.1986,108,3850. 4 P. T. Wolczanski, Polyhedron, 1995, 14,3335. 5 F. J. Feher, R. L. Blanski, J. Am. Chem. SOC.1992,114,5886. 6 F. J. Feher, T. L. Tajima, JAm. Chem. SOC.1994,116,2145. 7 A. Choplin, B. Coutant, C. Dubuisson, P. Leyrit, C. McGill, F. Quignard, R. Teissier in Stud. Surf. Sci. and Catal., Heterogeneus Catalysis and Fine Chemicals IV, (H. U. Blaser, A. Baiker, R. Prins, eds) Elsevier (Amsterdam), 1997, p. 353. 8 A. .J. Vizi-Orosz, R. Ugo, R. Psaro, A. Sironi, M. Moret, C. Zuchi, F. Ghelfi, G. Palyi, Inorg. Chem. 1994,33,4600. 9 B. Marciniec, P. Krzyianowski, J Organometal. Chem. 1995,493,261. 10 P. Krzyianowski, M. Kubicki, B. Marciniec, Polyhedron, 1996,15, 1. 1 1 B. Marciniec, P. Krzyianowski, M. Kubicki, Polyhedron, 1996,15,4233. 12 B. Marciniec, I. Kownacki, M. Kubicki, Organometallics,2002,21,3263. 13 a) G. A. Siegel, R. A. Bartlett, D. Decker, M. M. Olmsted, P.P. Power, Inorg. Chem. 1987,26,1773; b) T. A. Chesnokova, E. V. Zhezlova, A. N. Kornev , Y. V. Fedotova, L. N. Zakharov, G. K. Fukin, Y. A. Kursky, T. G. Mushtina, G. A. Domrachev, J Organometal. Chem., 2002,642,20. 14 a) B. Marciniec, NewJChem., 1997,21,815; b) B. Marciniec in Applied Homogeneous Catalysts with Organometallic Compounds,Cornils B .&Hermann W.A. eds, Verlag Chemie, Weinheim, 2002, Chapter 2.6; c) B. Marciniec, Appl. Organometal. Chem., 2000,14, 527; d) B. Marciniec in Ring Openning Metathesis Polymerization and Related Chemistry,E. Koshravi& T. Szymahska-Buzar Eds. Kluwer Acad,Publ;., 2001, p.391. 15 Y. Wakatsuki, H. Yamazaki, M. Nakano, Y. Yamamoto, J Chem. SOC.,Chem. Commun. 1991,703. 16 a) B. Marciniec, C. Pietraszuk, J. Chem. SOC.,Chem. Commun. 1995,2003; b) B. Marciniec, C. Pietraszuk, Organometallics, 1997, 16,4320. 17 B. Marciniec, E. Walczuk-GuSciora, C. Pietraszuk, Catal.Lett. 1998,55, 125. 18 B. Marciniec, I. Kownacki, D. Chadyniak, Inorg. Chem.Commun. 1999,2,581. 19 B. Marciniec, E. Walczuk-GuSciora, P. Blaiejewska-Chadyniak,J Mol.Catal., 2000, 160, 165. 20 B. Marciniec, C. Pietraszuk, M. Kujawa, J Mol. Catal., 1998,133,41. 21 C. P. Lenges, P. S. White, M. Brookhart, J Am. Chem. SOC.1999,11,4385. 22 B. Marciniec, E. Walczuk-GuSciora, C. Pietraszuk, Organometallics,2001,20,3423. 23 I. Kownacki, B. Marciniec, M. Kubicki, Chem. Commun., (submitted for publication). 24 B. Marciniec, P. Krzyianowski, E. Walczuk-GuSciora, W. Duczmal, J. Mol. Catal. 1999,144,263. 25 P. Blaiejewska-Chadyniak, M. Kubicki, B. Marciniec (unpublishedresults). 26 B. Marciniec, I. Kownacki, M. Kubicki, Inorg.Chim.Acta.2002,334,301.

264

Perspectives in Organometallic Chemistry

I. Kownacki, M. Kubicki, B. Marciniec, Polyhedron, 2002,20,3015. R. F. Baggio, M. T. Garland, J. Manzur, E. Spodine, Acta CrystaZZogr.1995, C51,602. Pat.Po1. P-351 449 (2001). Pat.Po1. P-351 451 (2001). B. Marciniec, E. Walczuk-GuSciora, P. Blaiejewska-Chadyniak, D. Chadyniak, M. Kujawa-Welten, S. Krompiec, 1'' European Silicon Days, Munich 617 Sept. 2001, P-73. 32 I. Kownacki, B. Marciniec (unpublished results). 33 Pat.Po1. P-351 450 (2001).

27 28 29 30 31

CHEAP CHIRAL LIGANDS FOR ASYMMETRIC TRANSITION METAL CATALYZED REACTIONS

M.T. Reetz Max-Planck-Institut f%rKohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 MulheirdRuhr, Germany

1 INTRODUCTION The stereoselective synthesis of chiral organic compounds using catalytic methods is of substantial academic and industrial interest,"2 as evidenced inter uliu by the award of the Nobel Prize for Chemistry 2001 to K.B. Sharpless: R. Noyori4 and W.S. Knowles.' Two major options are available, namely chiral synthetic catalysts such as transition metal c ~ m p l e x e sor ' ~ biocatalysts ~~~ such as enzymes.2 In the latter area we have proposed and developed a fundamentally new approach to asymmetric catalysis, namely the directed evolution of enantioselective enzymes.6 It is based on the proper combination of gene mutagenesis,' gene expression and high-throughput ee-screening: the "Darwinistic" process being independent of any knowledge of the structure or mechanism of the enzyme. Nevertheless, this novel concept has limitations because it is restricted to the types of transformations that enzymes are known to catalyze. For this reason we have proposed the concept of directed evolution of hybrid catalysts? Accordingly, an appropriate protein is chosen as a host for a synthetic transition metal catalyst, its wild-type gene is mutated and expressed in bacteria, and the encoded mutant proteins are modified chemically en masse so as to contain catalytically active transition metal centers. Following ee-screening for a given reaction of interest, the best "hit" is identified and the process is repeated using the corresponding mutated gene.' While these are ongoing efforts, we are also concentrating on more traditional research directed towards developing novel chiral ligands for application in transition metal catalysis, which is the subject of this overview. The main thrust is to design and prepare ligands which are particularly easy to prepare, i.e., those which are cheap while being highly enantioselective. This approach may make more industrial applications possible. The examples summarized here pertain mainly to asymmetric hydrogenation, although some results regarding conjugate addition reactions are also included. 2 CHIRAL DIPHOSPHONITES AND DIPHOSPHITES At the beginning of our studies no examples of the use of chiral diphosphonites or diphosphites as ligands for efficient Rh-catalyzed asymmetric hydrogenation were

266

Perspectives in Organometallic Chemistry

k n ~ w n .We ' ~ were therefore pleased to discover that the readily accessible BINOL-derived diphosphonites 1 - 4 are in fact excellent ligands.'

','*

bn=4

The best ligand turned out to be the ferrocene-based diphosphonite 3 (Scheme 1). In all cases precatalysts were prepared by reacting Rh(COD)2BF4 with a diphosphonite which displaces one of the COD-ligands.".12

HZ-

ee > 99.5 % 3 H C:C ?

HZ-

J+ H J ,!3 H3C02C

ee = 99.5 % H

phlN'cH

H3C02C

HZ_

H

HZ_ H02C

H

phlNi,H3 ee = 99 %

H02C

H

HP_ Ph N '

HAc

PhANAc H

ee = 96 %

Scheme 1 Typical Rh-catalyzed as mrnetric hydrogenationreactions (precatalystprepared @om Rh(C0D)ZBFd and ligand 3)$12

The nature of the chiral P-heterocycle is crucial for obtaining high ee-values in hydrogenation, the use of other chiral diols such as TADDOLS or hydrobenzoin leading to poor enantioselectivities(Scheme 2).

Cheap chiral ligands for asymmetric transition metal catalyzed reactions

e%

267

.Ph PC12

I

I

HO OH 4.4 NEt3, Et2O

/4&

Ph

ClpP

Ph'

H2 Rh/ligand

(89 %)

C02CH3 H3C--CC02CH3

ee=5%

Scheme 2 Synthesis and application of a hydrobenzoin-baseddiphosphonite'2b Force field calculations show that in the case of hydrobenzoin the P-heterocycle takes on a twist-envelope geometry which places the phenyl groups relatively far away from the rhodium coordinated at the lone electron pair of phosphorus (Scheme 3, lefi). In contrast, in the case of the BINOL-derived analogs, the aryl moieties (naphthyl) are closer and may thus provide greater differential shielding in the respective transition states (Scheme 3, right). Ph

twist-envelope (gauche)

i Ph

/'" iCR

Scheme 3 Stereochemical properties of hydro benzoin- and BINOL-based diphosphonites CfcR' =ferrocene plus another P-heterocycle);Rh has been omittedfor clarity Diphosphonites 1 - 4 are also excellent ligands in the asymmetric Rh-catalyzed conjugate addition of arylboronic acids,13 a reaction that had previously been described by Hayashi and Miyaura using BINAP as the chiral ligand.I4 Exploratory experiments were performed using cyclohexenone (5) and phenylboronic acid (6a). Table 1 shows that the ethano- and phenylene-bridged ligands l a and 2 are clearly superior in the present reaction, resulting in ee-values of 95% and 99%, respectively, which compare well with the enantioselectivity obtained in the BINAP-based Rh-catalyst (97% ee).l4 Interestingly, the diphenyl etherderived diphosphonite 4 leads to a complete reversal of enantioselectivity (ee = 99%), which is of considerable theoretical interest. An important advantage of the diphosphonitesystem relates to the fact that a five-molar excess of arylboronic acid is not necessary (as in the case of BINAP),14 because undesired competitive protonation of, the Rh-wl intermediates does not Moreover, substituted phenyl derivatives 6b - 6g also add to 5 enantioselectively (ee = 89 - 95%). Other cyclic and acyclic enones react similarly.

268

Perspectives in Organometallic Chemistry

The diphosphonite ligands 1 - 4 are also useful in the Cu-catalyzed conjugate addition of diethylzinc.l 5

A'

r 7

Table 1 Rh-catalyzed conjugate addition of 6a to 5 Parallel to these efforts we also studied chiral diphosphites, e.g., derivatives 9 prepared from dianhydro-D-mannite (8):16

OH 8

9

High ees were observed in the hydrogenation of itaconic acid ester (88% ee with (9BINOL and 95% ee with (R)-BINOL at phosphorus in ligands 9).16 These are the first cases of acceptable enantioselectivity in olefin hydrogenation using a chiral disphosphite. The mismatched and matched cases differ slightly, and it is the chirality of BINOL which determines the final outcome. Moreover, it was observed that the mismatched case reacts more slowly. Indeed, a type of non-linear effect using non-diastereopure catalyst mixtures was observed in a detailed kinetic study.17 The results offer an explanation why the use of mixtures of conformationally fluxional diasteromers lead to high enantioselectivities (Scheme 4). l6

Cheap chiral ligandsfor asymmetric transition metal catalyzed reactions

R,R-diastereomerI Rh

J fast product (ee = 99%)

R,Sdiastereorner / Rh

1

269

S,Sdiastereorner I Rh

slow

product

1

slow

product

Scheme 4 Use of a mixture of three conformationally fluxional diasteromeric ligands in the Rh-catalyzed hydrogenation of itaconic acid ester (Rh omittedfor clarity)16 It is clear that the unusually high enantioselectivity is not due to the chiral backbone alone, but to the local chirality in the P-heterocycles. Since these are prepared from achiral o,o'dimethyldiphenol, the system benefits from the fact that one of the three diastereomers is much more reactive than the others. This interesting principle was observed independently by Noyori in a different system composed of two fluxional diastereomers," and other examples have since f~llowed.'~ It simply means that in a ligand system it is possible to use one portion which is configurationally stable but not effective in conjunction with another moiety which is chirally fluxional and highly effective in inducing enantioselectivity. Moreover, the effect has been observed by Walsh in "intermolecular" catalyst systems using chirally fluxional meso-type compounds as additives?' 3 CHIRAL MONOPHOSPHITESAND MONOPHOSPHONITES

In the wake of studying the above chiral bidentate diphosphites, we made an unexpected discovery, namely that BINOL-derived monophosphites are better and even cheaper ligands for Rh-catalyzed hydrogenation.21In a mechanistic study aimed at the synthesis of a chiral diphosphite 9 in which one P-moiety is attached to (R)-BINOL and the other to (9-BINOL, an intermediate compound was prepared in which one hydroxy function of 8 bears an (R)-BINOL-containing P-heterocycle and the other hydroxy function is protected as the 0-benzyl ether. This monophosphite was routinely tested in Rh-catalyzed hydrogenation. To our surprise this catalyst showed ee-values of > 95%! Therefore, simple alcohols were used as building blocks in the synthesis of a number of chiral monophosphites. Scheme 5 shows only a few select examples as well as the result of using these simple ligands in the Rh-catalyzed hydrogenation of itaconic acid ester.21As can be seen, the ee-values range between 29% and 99%, depending upon the nature of the alcohol which is used in the synthesis of the ligand.

270

Perspectives in Organometallic Chemistry

mo' mo' \

\

(S)

\

\

(S)

Scheme 5 Modular construction of BINOL-derived mono hosphites and their use as ligands in Rh-catalyzed hydrogenation of itaconic acid ester2 r A variety of 2-actamido acrylic acid esters are likewise hydrogenated with high degrees of enantioselectivity, as in the case of the parent substrate shown in Scheme 6.21Rh: substrate ratios of 1 : 1000 but as low as 1 : 5000 can be used in these reactions at room temperature (1.3 bar H2)?l Conversion is usually complete. C02CH3 +c02CH3 NHCCH3

H2 Rh 1 ligand ~

H3Ciiiii(

I1

NHCCH3

It

0

0

ligands:

,P-OR

ROH used in ligand-prep.

ee (% )

CH3OH

72.8

(CH3)zCHOH

94.8

C&OH

80.6

HO L P h

Scheme 6 Ligand eflects in asymmetric hydrogenation2'

93.3

Cheap chiral ligands for asymmetric transition metal catalyzed reactions

27 1

Finally, industrially important chiral mines 11 can also be prepared enantioselectively, in this case N-acyl enamines 10 being the precursors?2 The ee-values range between 90% and 98% depending upon the nature of the alcohol component in the chiral monophosphite. Again, these results demonstrate the power of the modular nature of these readily accessible ligands.

H2 * Rh(L)(COD)BF4

10 (R' = awl; R2 = H, Me)

11

In independent work Feringa, de Vries and co-workers have shown that the monophosphoramidite 12 is also an excellent ligand in Rh-catalyzed olefin hydr~genation?~ In contrast, P r i ~ ~ gand l e ~we25 ~ have reported that the analogous BINOLderived monophosphonites 13 are generally less efficient (higher activity but lower enantioselectivity).

12

13a b c d

R=Me R=Et R=t-Bu R=Ph

It is clear that in hydrogenation the long-standing dogma that chelating (bidentate) ligands are necessary in order to observe high enantioselectivities in a general way no longer However, the source of high enantioselectivity has not yet been l l l y illuminated. We have examined the NMR spectra of precatalysts Rh(L*)2(COD)BF4 and have proven that two chiral monophosphites (L*) are bonded to rhodium. Moreover, strong non-linear effects are observed (Scheme 7), suggesting that two monophosphites are attached to the metal in the transition state of hydrogenation and/or in a pre-equilibrium step?6 Preliminary kinetic studies show the following order in rate in Rh-catalyzed hydrogenation: monophosphonites > monophosphites > monophosphoramidites. In the case of monophosphites turnover numbers of 200 000 have been achieved?6

272

Perspectives in Organometallic Chemistry

100

90 80 9

70

<2 60 0

9 50

40

8:

30 20 10

0 0

10

20

30

40 88

50 60 [“rc] (catalyst)

70

80

90

100

Scheme 7 Non-linear efects in the Rh-catalyzed hydrogenationof itaconic acid ester using the monophosphite derivedfiom BINOL and isopropano12‘

In spite of these mechanistic advances, they do not explain the source of enantioselectivity on a molecular basis. Thus, further studies are necessary before final conclusions can be reached. Unfortunately, we have not been able to obtain adequate crystals of relevant Rhcomplexes. However, the X-ray crystal structure of the first metal complex of a BINOLderived phosphorous acid ester was recently analyzed in our laboratories?’ It is a Ptcomplex in which the acidic OH-functions are H-bonded to triethylamine. The structure has a cis-arrangement of the two chiral ligands (Scheme 8). It is currently unclear to what extent Pt-P bond rotation is reduced or inhibited in complexes of this type, which may bear some resemblance to the Ph-complexes used in hydrogenation. If restricted rotation around P-metal bonds pertains, then the two mono-P-units simulate a chelating bidentate ligand system, a phenomenon which would explain the high enantioselectivity in Rh-catalyzed hydrogenation.

w Scheme 8 X-ray crystal structure of a chiral phosphorous acid ester Pt-complex2’

Cheap chiral ligands for asymmetric transition metal catalyzed reactions

213

4 CONCLUSIONS BINOL-derived chiral diphosphites and diphosphonites are readily accessible ligands which show high degrees of enantioselectivity in a number of olefin hydrogenation reactions. Surprisingly, BINOL-derived monophosphites, which are even cheaper, are even more efficient. Moreover, since the modular nature of these compounds allows for easy ligand tuning, they are attracting the attention of industrial chemists. The question remains as to their utility in other types of transition metal catalyzed processes. References

1

(a) Comprehensive Asymmetric Catalysis, eds. E.N. Jacobsen, A. Pfaltz and H. Yamamoto, Springer, Berlin, 1999, Vol. 1-111. (b) H. Brunner and W. Zettlmeier, Handbook of Enantioselective Catalysis with Transition Metal Compounds, VCH, Weinheim, 1993, Vol. 1-11. (c) R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, New York, 1994. (d) R.A. Sheldon, Chirotechnology: Industrial Synthesis of Optically Active Compounds,Dekker, New York, 1993. 2 (a) H.G. Davies, R.H. Green, D.R. Kelly and S.M. Roberts, Biotransformations in Preparative Organic Chemistry: The Use of Isolated Enzymes and whole Cell Systems in Synthesis, Academic Press, London, 1989. (b) C.H. Wong and G.M. Whitesides, Enzymes in Synthetic Organic Chemistry (Tetrahedron Organic Chemistry Series Vol. 12), Pergamon, Oxford, 1994. (c) Enzyme Catalysis in Organic Synthesis: A Comprehensive Handbook, eds. K. Draw and H. Waldmann, VCH, Weinheim. 1995, Vol. 1-11. (d) K. Faber, Biotransformations in Organic Chemistry, 3rd Edn., Springer, Berlin, 1997. 3 K.B. Sharpless, Angew. Chem., 2002, 114, 2126; Angew. Chem. Int. Ed., 2002, 41, 2024. 4 R. Noyori, Angew. Chem.,2002,114,2108; Angew. Chem. Int. Ed., 2002,41,2008. 5 W.S. Knowles, Angew. Chem., 2002, 114, 2096; Angew. Chem. Int. Ed., 2002, 41, 1998. 6 (a) M.T. Reetz, Pure Appl. Chem., 2000, 72, 1615. (b) M.T. Reetz, S. Wilensek, D. Zha and K.-E. Jaeger, Angew. Chem., 2001,113, 3701; Angew. Chem. Int. Ed., 2001, 40, 3589. (c) D. Zha, S. Wilensek, M. Hermes, K.-E. Jaeger and M.T. Reetz, Chem. Commun. (Cambridge), 2001, 2664. (d) M.T. Reetz and K.-E. Jaeger, ,in Directed Molecular Evolution of Proteins, eds. S . Brakmann and K. Johnson, Wiley-VCH, Weinheim, 2002, p. 245. 7 (a) K.A. Powell, S.W. Ramer, S.B. del Cardayr6, W.P.C. Stemmer, M.B. Tobin, P.F. Longchamp and G.W. Huisman, Angew. Chem., 2001,113,4068; Angew. Chem. Int. Ed., 2001,40,3948. (b) F.H. Arnold, Nature (London),2001,409,253. 8 (a) M.T. Reetz, Angew. Chem., 2001, 113, 292; Angew. Chem. Int. Ed., 2001, 40, 284. (b) M.T. Reetz, Angew. Chem., 2002, 114, 1391; Angew. Chem. Int. Ed., 2002, 41, 1335. 9 (a) M.T. Reetz, in Pharmacochemistry Library, Vol. 32 (Trends in Drug Research 111), ed. H. van der Goot, Elsevier, Amsterdam, 2002, p. 27. (b) M.T. Reetz, Tetrahedron, 2002, in press. (c) M.T. Reetz, patent application DE-A 101 29 187.6 (June 19,2002). 10 Brunner and Wink used chiral diphosphite ligands based on carbohydrates and tartaric acid derivatives in Rh-catalyzed hydrogenations and obtained ee values of 1 34%: (a) H. B m e r and W. Pieronczyk, J. Chem. Res., Synop., 1980, 76. (b) D.J.

274

11 12 13 14

15 16 17 18 19 20 21 22 23 24

25 26 27

Perspectives in OrganometallicChemistry

Wink, T.J. Kwok and A. Yee, Inorg. Chem., 1990, 29, 5006. (c) 0. P h i e s , G. Net, A. Ruiz and C. Claver, Tetrahedron: Asymmetry, 2000,11, 1097. M.T. Reetz, A. Gosberg, R. Goddard and S.-H. Kyung, Chem. Commun. (Cambridge), 1998,2077. (a) M.T. Reetz, Pure Appl. Chem., 1999, 71, 1503. (b) A. Gosberg, Dissertation, Ruhr-UniversitZit Bochum, 2000. M.T. Reetz, D. Moulin and A. Gosberg, Org. Lett., 2001,3,4083. (a) Y. Takaya, M. Ogasawara, T. Hayashi, M. Sakai and N. Miyaura, J. Am. Chem. Soc., 1998, 120, 5579. (b) T. Hayashi, T. Senda, Y. Takaya and M. Ogasawara, J. Am. Chem. SOC.,1999, 121, 11591. (c) T. Hayashi, T. Senda and M. Ogasawara, J. Am. Chem. SOC.,2000, 122, 10716. (d) S. Sakuma, M. Sakai, R. Itooka and N. Miyaura, J Org. Chem., 2000,65,5951. (e) Y. Takaya, T. Senda, H. Kurushima, M. Ogasawara and T. Hayashi, Tetrahedron: Asymmetry, 1999, 10, 4047. ( f ) M. Kuriyama and K. Tomioka, Tetrahedron Lett., 2001,42, 921. (g) T. Hayashi, Synlett, 2001, 879-887. (h) R. Itooka, Y. Iguchi and N. Miyaura, Chem. Lett., 2001, 7 , 722723. M.T. Reetz, A. Gosberg and D. Moulin, Tetrahedron Lett., 2002,43, 1 189. M.T. Reetz and T. Neugebauer, Angew. Chem., 1999, 111, 134; Angew. Chem. Int. Ed., 1999,38, 179. D.G. Blackmond, T. Rosner, T. Neugebauer and M.T. Reetz, Angew. Chem., 1999, 111,2333; Angew. Chem. Int. Ed., 1999,38,2196. K. Mikami, T. Korenaga, M. Terada, T. Ohkuma, T. Pham and R. Noyori, Angew. Chem., 1999,111,517; Angew. Chem. Int. Ed., 1999,38,495. See for example: (a) 0. P h i e s , M. DiCguez, G. Net, A. Ruiz and C. Claver, Chem. Commun. (Cambridge), 2000, 2383. (b) W. Chen and J. Xiao, Tetrahedron Lett., 2001,42,8737. J. Balsells and P.J. Walsh, J. Am. Chem. SOC., 2000,122, 1802. (a) M.T. Reetz and G. Mehler, Angew. Chem., 2000, 112, 4047; Angew. Chem. Int. Ed., 2000, 39, 3889. (b) M. T. Reetz and G. Mehler, patent application DE-A 100 27 505.2 (June 6,2002). M.T. Reetz, G. Mehler, A. Meiswinkel and T. Sell, Tetrahedron Lett., submitted. M. van den Berg, A.J. Minnaard, E.P. Schudde, J. van Esch, A.H.M. de Vries, J.G. de Vries and B.L. Feringa, J. Am. Chem. SOC., 2000,122,11539. C. Claver, E. Fernandez, A. Gillon, K. Hes1op;D.J. Hyett, A. Martorell, A.G. Orpen and P.G. Pringle, Chem. Commun. (Cambridge), 2000,961. M.T. Reetz and T. Sell, Tetrahedron Lett., 2000,41,6333. M.T. Reetz et. al., unpublished data. T. Sell, Dissertation, Ruhr-Universitat Bochum, 2002 (structure solved by R. Goddard).

CHIRAL METAL COMPLEXES IN ASYMMETRIC CATALYSIS

Christina Moberg,' Oscar Belda, Kristina Hallman, Robert Stranne, Mats Svensson, JeanLuc Vasse, Tebikie Wondimagegn, Raivis Zalubovskis Department of Chemistry, Organic Chemistry, Royal Institute of Technology, SE 100 44 Stockholm, Sweden

1 INTRODUCTION

Asymmetric metal catalysis constitutes one of the most important methods for the preparation of chiral compounds in enantiomerically pure form.' The preparation of chiral ligands with ability to transfer their chiral information from the catalyst to the substrate undergoing reaction is therefore an important issue.2 To achieve high enantioselectivity in the catalytic reactions, control of the electronic and steric properties of the catalytically active metal complexes is crucial. For the design of efficient ligands, several factors therefore need to be taken into account. These include the choise of donor atoms, which affect the stability and reactivity of the complexes and influence properties such as the trans influence and the trans effect. The cone angle3of monodentate ligands and bite angle4 of bidentate ligands as well as the size of chelate rings are additional factors influencing the reactivity and selectivity of the catalytic species. Ligands which are conformationally rigid frequently result in more efficient chirality transfer than those which are conformationally labile, although strong conformational preferences, resulting from weak interactions such as hydrogen bonding, van der Waals interactions and n-stacking or from stereoelectronic effects may add suMicient rigidity to otherwise conformationally labile ligands. Another factor that may play a crucial role in catalytic reactions is the symmetry properties of the ligands and their metal complexes. C2-Symmetric metal complexes have been frequently employed in asymmetric metal catalysis, and in a large variety of processes such ligands have resulted in formation of products with high enantioselectivity.' The success of ligands with rotational symmetry originates in a reduction of the number of catalyst-substate interactions and, as a consequence, the number of competing reaction pathways, thus increasing the probability of a successful result.6 Occasionally, however, ligands devoid of symmetry may result in superior enantioselectivity.

To whom correspondence should be addressed

276

Perspectives in Organometallic Chemistry

2 PALLADIUM-CATALYZEDALLYLIC SUBSTITUTIONS One process where asymmetric ligands as well as ligands with twofold rotational symmetry have been shown to induce high enantioselectivity is the palladium-catalyzed asymmetric nucleophilic substitution of allylic acetates and carbonates (eq l).' In the process, a prochiral substrate may be transformed into a chiral product or a racemic substrate may undergo deracemization.

Allylic substitutions of carbonates and acetates are catalyzed by several metals, palladium, molybdenum and tungsten being the most common examples. The allylic substitutions are synthetically highly versatile processes whereby formation of carbon-carbon and carbonheteroatom bonds may be achieved in a highly enantioselective manner. The mechanism of the palladium-catalyzed process has been studied in great detail, and has been shown to consist of the coordination of the unsaturated substrate to Pd(O), followed by oxidative addition to form a Pd(I1) ally1 complex which undergoes attack by the nucleophile forming a Pd(0) complex with the product olefin. Ligand exchange completes the catalytic cycle (Scheme 1).*

Scheme 1

Catalytic cycle for the palladium-catalyzed substitution of an allylic acetate

Depending on the substrate, the enantiodetermining step is either the oxidative addition or the nucleophilic addition. The attack of the nucleophile on the allylic moiety occurs trans to palladium and thus outside the coordination sphere of the metal. For symmetrical allylic substrates, the enantioselectivity is determined by the regiochemistry of the nucleophilic addition (Figure 1). This is governed by ligand-substrate interactions resulting from the steric and electronic properties of the catalyst. Different trans influence of the donor atoms in bidentate ligands may favor attack at one of the allylic positions.

211

Chiral metal complexes in asymmetric catalysis

L'

Nu

Figure 1

Nucleophilic attack on a symmetric r -ally1 complex

C&mmetric bidentate ligands must necessarily have identical donor atoms and the enantiodiscrimination in asymmetric reactions involving such ligands relies essentially on the steric properties of the enantiocontrolling catalytic intermediates. Introduction of electronic dissymmetry destroys the rotational symmetry, but may lead to superior properties in catalytic processes due to the different trans influence and different n-back donation ability of the donor atoms. A large variety of P,N-ligands have indeed proven to induce high enantioselecitity in the palladium catalyzed allylati~n.~ Control of the steric as well as electronic properties is thus crucial for high selectivity. 2.1

Sterically symmetric complexes

During studies of 2-( 1-methoxyalkyl)pyridinooxazolines (1) and 2-( 1-hydroxyalkyl)pyridinooxazolines (2) as ligands we found that the two types of ligands exhibited different behavior, the ligands 1 with (R*,R*) configuration resulting in higher enantioselectivity than the diastereomers having (R*,S*) configuration, whereas the reverse behaviour was found for ligands of type 2.''

In order to get an insight into the reason for the different behaviour of the two types of lignads, the structure of the metal complexes was studied using experimental, including NMR spectroscoy and X-ray crystallography, as well as theoretical methods.' In the free ligands, the two types of compounds adopt different conformations. Due to hydrogen bonding, 2-(hydroxymethy1)pyridine adopts a syn planar conformations, whereas the methoxy ligand adopts an anti planar conformation with a N-C-C-0 dihedral angle close to 180". The reason for this conformational preference is the repulsion of the lone pairs on the nitrogen and oxygen atoms and, to a minor extent, a stereoelectronic effect resulting fiom donation of electrons fiom the nitrogen lone pair into the antibonding carbon-oxygen bond.12 A plot of the energy versus the dihedral angle shows that the energy decreases continously upon rotation from a syn-planar conformation to an anti-planar struture (Figure 2, blue curve). A slight elongation of the C-0 bond was found to accompany the rotation (pink curve; that the bond length has a maximum close to 90" is probably due to a stereoelectronic effect involving the n-orbitals).

'

Perspectives in Organometallic Chemistry

278

30

6o

w

120

1%

l(0

Energy and C - 0 bond length as a function of the N-C-C-0 dihedral angle Figure 2 in 2-methoxymethylpyridine

In their x-allylpalladium(I1) complexes, the two types of ligands adopt similar trans planar conformations, with N-C-C-0 dihedral angles close to 180". For Pd(I1) complexes containing ligands 2 a second conformational minimum with a dihedral angle close to 70" was found as well, although the anti conformation was found to be the most stable one. Attack of the nucleophile, which is accompanied by the reduction of Pd(I1) to Pd(O), does not result in any major conformational change in ligands 1, whereas for the hydroxycontaining ligands 2, the conformation with a smaller dihedral angle (close to 60°, Figure 3) becomes more stable than the anti conformation. The energy difference for the complexes resulting from attack of fluoride ion on a simple propenyl group was found by DFT calculations to amount to about 9 kJ/mol. This conformational change was found to occur already in the transition state, and is therefore expected to affect the stereochemical outcome of the catalytic reaction.

Figure 3 Conformational minimum found by DFT calculations for Pd(0) olefin complex obtained by nucleophilic attack ofjluoride ion on a sally1 Pd(I4 complex

This conformational change results in different symmetry properties of the transition state complexes containing ligands 1 and 2. Figure 4 shows the position of the sterically bulky groups viewed along the coordination plane of the metal. The pseudo-C2 symmetry (A) of the starting ally1 Pd(I1) complex changes to apseudo-C, symmetry (B). Ligands with conformation A lead to high enantioselectivity in the nucleophilic substitution of 1,3-

Chiral metal complexes in asymmetric catalysis

279

propenyl acetate, whereas those adopting conformation B result in inferior stereoselectivity. The conformational change from A to B for alcohols 2 with (R*,R*)configuration is thus the reason why these ligands induce lower enantioselectivity than ethers 1 with the same absolute configuration, and vice versa.

View along the coordination plane of the metal of conformational change Figure 4 occurring in (R *,R *)-2 during the catalytic reaction

The driving force for the conformational change was found to originate in an interaction between the elecron-rich palladium atom of the olefin complex and the hydrogen atom of the hydroxy group. DFT calculations showed the Pd-H distance to be short, around 2.4 A. The interaction observed thus involves an electron-rich metal center acting as an acceptor towards a proton donor, and has therefore the characteristics of a hydrogen bond. It is thus different from an agostic interaction. Whereas the agostic interaction involves an electron deficient metal center and represents a three-center twoelectron bond, this type of weak hydrogen bonding involves a basic metal center and represents a three-center four-electron bond. The present type of interaction has some precedence in the literature. It has been shown by X-ray crystallography and NMR spectroscopy to exist in complexes containing electron-rich late transition centers, typically in d8 and d’’ compexes with Ni(II), Rh(I), Pd(II), Ir(I), Pt(II), and Ni(0) interacting intramolecularly of intermolecularly with acidic prot~ns.’~ As expected, the interaction becomes weaker with less acidic proton donors and with less basic metal centers, as well as in complexes with less basic ligands surrounding the metal. Although a large number of examples of M...H-X interactions (X being an electronegative atom such as N or 0)have been found during the last decade, the catalytic reactions involving ligands 2 is the first example where this type of interaction has been shown to affect the stereochemistry of a chemical reaction. 2.2 Electronic dissymmetry In order to exploit the behavior of bidentate ligands with ”steric symmetry” (possessing a mirror plane or a twofold rotational axis) which are being electronically desymmetrized, ligands with different donor atoms having identical substituents on the two donor atoms were prepared. Examples of such ligands are 3 and 4.14

Ligand 3, with the sterically bulky groups on different sides of the coordination plane is sterically Cz-symmetric and electronically asymmetric, whereas 4 is sterically a meso compound, the chirality originating solely from the presence of different donor atoms. The

Perspectives in Organometallic Chemistry

280

new types of ligands were assessed in the palladium-catalyzed allylic alkylation of rue- 1,3diphenyl-2-propenyl acetate and rue-2-cyclohexenyl acetate with dimethyl malonate (Scheme 2).

[(q3-C3Hs)Pd Cllz. lba nd

(MeOCO),CH,, BSA, KOAc

Med Palladium-catalyzed allylic alkylation of rac-I,3-diphenyl-2-propenyl Scheme 2 acetate and rac-2-cyclohexenyl carbonate with dimethyl malonate It was found that reaction of rac-l,3-diphenyl-2-propenylacetate in the presence of ligand 3 resulted in full conversion to the product having (S) absolute configuration in 98% ee within six hours at room temperature, whereas pseudo-meso ligand 4 exhibited lower selectivity and lower reactivity, affording 95% of the product with opposite absolute configuration with merely 37% ee within three days at the same temperature. Interestingly, in reactions with the cyclic substrate, ligand 3 proved to be inferior to 4, both giving the (R)-product but with 12 and 24% ee, respectively. Ligand 3 also exhibited lower reactivity than 4 with this substrate, requiring 24 hours for 40% conversion, as compared to 70% conversion within the same period of time for 4. Due to the asymmetry of the ligands, metal complexes with ex0 and endo configuration are expected to form from each substrate. From the non-cyclic substrate ally1 moieties with syn-syn, syn-anti, anti-syn and anti-anti configuration may form. The synsyn complexes are expected to have highest stability and are probably those principally leading to product. The situation with the cyclic substrate is simpler, as only complexes with anti-anti configuration of the allylic moeity may form. The attack of the nucleophile is governed by steric as well as electronic factors and expected to occur trans to phosphorus due to the larger trans influence of phosphorus than of nitrogen in such a way as to form the most stable Pd(0) olefin c~mplex.'~ In Scheme 3 the ex0 and endo complexes for each type of ligand for the two substrates are shown, along with the Pd(0) olefin complexes obtained by attack of the nucleophile at the allylic position trans to phosphorus. The olefin complexes expected to suffer from least steric hindrance are indicated. These complexes are favoured over other complexes, as shown by the experimental results. It is interesting to note, that for the two types of substrates, different types of ligands should thus be selected, although the meso ligand affords a modest enantioselectivity.

Chiral metal complexes in asymmetric catalysis

28 1

A: C2;syn-syn

Nu

I

B:meso;syn-syn

Nu

2

Nu

CC,; anti-anti

QF?*

D:meso;anti-anti

Nu

-

2% -

N

UU

N

J

UU 1

Scheme 3 Pd(I4 ally1 complexes and Pd(0) oelfn complexes involved in catalytic reactions with I , 3-diphenyl-2-propenylacetate and 2-cyclohexenyl carbonate employing ligands 3 and 4 3 MOLYBDENUM-CATALYZEDALLYLIC SUBSTITUTIONS

Allylic alkylations catalyzed by molybdenum complexes have also been the subject of recent studies.16~17Interestingly, the molybdenum-catalyzed processes have been found to have a complementary behaviour compared to those catalyzed by Pd, since when using Mo complexes as catalysts, the nucleophilic attack takes place preferentially at the most substituted carbon atom when non-symmetrical allylic substrates are used," as opposed to most Pd catalyzed allylations (Scheme 4). Nu

Nu

Scheme 4

Mo

R A O A c

Nu

Pd

W"

Regiochernistry of Mo- and Pd-catalyzed allylic substitutions

3.1 Pyridylamides as ligands

The coordination chemistry of bispyridylamides of type 5 to metal ions has been widely investigated mainly by Vagg and coworker^.'^ Coordination of the pyridine nitrogen atoms as well as the carbonyl oxygen or the deprotonated nitrogen atoms of the amide groups is commonly observed. Although structural aspects of the complexes have been studied in detail, metal complexes with chiral analogues have only been employed in asymmetric catalytic reactions to a limited extent?'

Perspectives in Organometallic Chemistry

282

n

5

6

Trost reported the first asymmetric catalyzed allylic substitution using a complex of Mo(0) and bispyridylamide 6 in which excellent enantioselectivities and regioselectivities were obtained for a variety of substrates.16Later, we developed a robust and convenient reaction procedure for the microwave accelerated (Smith CreatorTMfrom Personal Chemistry) reaction of cinnamyl carbonate with sodium dimethylmalonate using Mo(CO)~, ligand 6 and N, 0-bis(trimethylsily1)acetamide (BSA) under air that afforded the product exhibiting basically the same regio- and enantioselectivity(Scheme 5)F1

OC02Me

4 % Mo(C0)e , Ligand CHz(C02Meh, CHNa(CQMe), BSA THF, p-waves

branched

Inear

19:l branched tolnear 98% ee

Scheme 5

Mo-catalyzed substitution of cinnamyl carbonate

With this reaction procedure in our hands we decided to investigate the electronic and steric effects of the ligand on the outcome of the reactionF2 We found that electronic effects were important and thus, by introducing an electron-donating group (4-methoxy) in the pyridine rings the product obtained presented higher regio- (40:l branched to linear) and enantioselectivity (>99% ee). On the other hand, an electron-withdrawing group (4nitro) in the pyridine rings resulted in a less selective catalyst (16: 1 branched to linear and 97% ee). Steric factors also played an important role since when a ligand with a methyl group in the 6-position of the pyridine rings was used, a catalyst exhibiting poor activity and selectivity was obtained (13:1 branched to linear, 79% ee). Recently, we prepared new derivatives of ligand 6 in order to study the generality of these results and to investigate if a more selective catalyst could be obtained by further modification of the electronic properties of the liga11d.2~ Again, electron-donating groups in the 4-position of the pyridine nuclei gave rise to catalysts that exhibited higher selectivity. Thus, a ligand having the pyridine rings substituted with pyrrolidine groups in the 4-position gave the products with 88:l branched to linear ratio and the branched product was formed with 96% ee. If instead ligands with 4-ChlO1-0substituents were used, the products had a 74: 1 branched to linear ratio and 96% ee. Furthermore, we explored the reaction outcome with other allylic substrates when using the easily obtained 4-chloro substituted ligand and compared to the reaction outcome when using the parent ligand 6. In all cases, the catalyst formed with the 4-chloro substituted ligand afforded the product with higher regioselectivity and comparable enantioselectivities.

Chiral metal complexes in asymmetric catalysis

283

New 6-pyridine subtituted bispyridylamides derived from 6 (6-bromo and 6-methoxy) were also prepared, but use of these ligands in the catalytic reaction did not lead to product. We also investigated the coordination chemistry of the metal complex formed b microwave heating the chiral bispyridylamide 6 and Mo(C0)6 by NMR spectroscopy.K The complex gave rise to two sets of signals for the protons in the pyridylamide ligand. The complex observed most probably contains only one molecule of the ligand coordinating in an unsymmetrical manner, however, since identical 'H NMR spectra were obtained from complexes prepared from enantiomerically pure and racemic ligands. Recently, a detailed investigation of the binding mode of a molybdenum allyl complex of 6 was published in connection to a study of the mechanism Mo-catalyzed asymmetric allylic alkylati0n.2~The allyl molybdenum complex observed was indeed shown to coordinate in an unsymmetrical manner. Further details about the mechanism of the Mocatalyzed process may be helpful for the design new ligands efficient for a larger variety of substrates. 4 CONCLUSION

Ligands exhibiting high enantioselectivity in palladium and molybdenum catalyzed nucleophilic substitutions of allylic acetates and carbonates were designed. For the molybdenum catalyzed process, ligands resulting in highly regioselective formation of products from unsymmetrical allylic substrates were achieved. The electronic and steric properties as well as the symmetry of the ligands and metal complexes were crucial for the selectivity in the catalytic reactions. References 1

I. Ojima, Catalytic Asymmetric Synthesis; VCH Publishers: New York, 1993; R. Noyori, Asymmetric Catalysis in Organic Synthesis; John Wiley & Sons, Inc.: New York, 1994; E. N. Jacobsen, A. Pfaltz, H. Yamamoto, Comprehensive Asymmetric Catalysis, Vol 1-3; Springer: Berlin, 1999. 2 J. Seyden-Penne, Chiral Auxiliaries and Ligands in Asymmetric Synthesis; John Wiley & Sons, Inc.: New York, 1995. 3 C. A. Tolman, Chem. Rev. 1977,77,313-348. 4 P. Dierkes and P. W. N. M. van Leeuwen, J Chem. SOC., Dalton Trans. 1999, 15191529; P. W. N. M. van Leeuwen, P. C. J. Kamer, J. N. H. Reek and P. Dierkes, Chem. Rev. 2000, 100,2741-2769. 5 J. K. Whitesell, Chem. Rev. 1989,89, 1581-1590. 6 C. Moberg, Angew. Chem. Int. Ed. Engl. 1998,37,248-268. 7 C. G. Frost, J. Howarth, J. M. J. Williams, Tetrahedron: Asymmetry 1992, 3, 10891122; B. M. Trost, C. Lee, Catalytic Asymmetric Synthesis, 2ndEd.; I. Ojima, Ed.; Wiley: New York, 2000; pp 593-649. 8 B. M. Trost, D. L. VanVranken, Chem. Rev. 1996,96,395-422. 9 For a recent discussion of this type of ligands, see: M. Widhalm, U. Nettekoven, H. Kalchhauser, K. Mereiter, M. J. Calhorda and V. FClix, Organometallics 2002, 21, 3 15-325. 10 K. Nordstrom, E. Macedo and C. Moberg, J Org. Chem. 1997, 1604-1609. 1 1 M. Svensson, U. Bremberg, K. Hallman, 1. Csaregh and C. Moberg, Organometallics 1999,18,4900-4907.

284

.

Perspectives in Organometallic Chemistry

12 C. Moberg, H. Adolfsson, K. Wzirnmark, P.-0. Norrby, K.-M. Marstokk. and H. Mdlendal, Chemistry: A European Journal 1996, 2, 516-522; C. Moberg, U. Bremberg, K. Hallman, M. Svensson, P.-P. Norrby, A. Hallberg, M. Larhed and I. Csoregh, Pure Appl. Chem. 1999,71, 1477-1483. 13 S. G. Kazarian, P. A. Haley and M. Poliakoff, J. Chem. Soc., Chem. Commun. 1992, 994-997; S . G. Kazarian. P. A. Haley and M. Poliakoff, J. Am. Chem. SOC.1993,115, 9069-9079; L. Brammer, D. Zhao, F. T. Lapido and J. Braddock-Wilking, Acta. Cryst. 1995, B51, 632-640; D. Braga, F. Grepioni, E. Tedesco, K. Biradha and G. R. Desiraju, Organometallics, 1997,16, 1846-1856; A. Milet, A. Dedieu and A. J. Canty, Organometallics 1997, 16, 5331-5341; G. R. Desiraju, J. Chem. SOC., Dalton Trans., 2000,3745-3751;M. J. Calhorda, Chem. Commun. 2000,801-809. 14 R. Stranne, J.-L. Vasse and C. Moberg, Org. Lett., 2001, 2525-2528; J.-L. Vasse, R. Stranne, R. Zalubovskis, C. Gayet and C. Moberg, To be submittedfor publication. 15 A. Saitoh, K. Achiwa, K. Tanaka and T. Morimoto, T. J. Org. Chem. 2000,65,42274240. 16 B. M. Trost and I. Hachiya, J. Am. Chem. Soc., 1998,120, 1104-1 105; B. M. Trost, S. Hillbrand and K. Dogra, J. Am. Chem. Soc., 1999,121, 10416-10417. 17 F. Glorius and A. Pfaltz, Org. Lett., 1999,1,141-144; F. Glorius, M. Neuburger and A. Pfaltz, Helv. Chim. Acta, 1999,84,3178-3 196; A. V. Malkov, P. Spoor and V. Vinader, ComprehensiveAsymmetric Catalysis, 1999,2,833-884. 18 J. W. Faller and K. H. Chao, J. Am. Chem. Soc., 1983,105,3893-3898; B. M. Trost and M. Lautens, J. Am. Chem. Soc., 1982,104,5543-5545; B. M. Trost and M. Lautens, J. Am. Chem. SOC.,198,105,3343-3344. 19 R. L. Chapman and R. S. Vagg, Inorg. Chim. Acta, 1979,33,227-234;F. S . Stephens and R. S. Vagg, Inog. Chim. Acta, 1988,142,43-50 and references therin. 20 H. Adolfsson and C. Moberg, Tetrahedron:Asymmetry, 1995,6,2023-203 1; C. Moberg, H. Adolfsson and K. Wzirnmark, Acta Chem. Scand., 1996, 50, 195-202; R. Halle, A. Brehdret, E. Schultz, C. Pinel and M. Lemaire, Tetrahedron:Asymmetry, 1997,8 21 N-F. Kaiser, U. Bremberg, M. Larhed, C. Moberg and A. Hallberg, Angew. Chem.,Int. Ed., 2000,39,3595-3598. 22 0.Belda, N-F. Kaiser, U. Bremberg, M. Larhed, A. Hallberg and C. Moberg, J. Org. Chem.,2000,65,5868-5870. 23 0.Belda and C. Moberg, Synthesis, In press. 24 0.Belda and C. Moberg, Unpublishedresults. 25 B. M. Trost, K. Dogra, I. Hachiya, T. Emura, D. L. Hughes, S. Krska, R. A. Reamer, M. Palucki, N. Yasuda and P. J. Raider, Angew. Chem, Int. E d , 2002,41,1929-1932.

IN SEARCH OF ASYMMETRIC PROPARGYLIC SUBSTITUTION REACTIONS MEDIATED BY OPTICALLY ACTIVE INDENYL-RUTHENIUM(I1) ALLENYLIDENE COMPLEXES

V. Cadierno, S. Conejero, M. P. Gamasa and J. Gimeno Departamento de Quimica Orghica e Inorghica, Instituto Universitario de Quimica Organometiilica "Enrique Moles" (Unidad Asociada al C.S.I.C.), Universidad de Oviedo, E-33071 Oviedo, Spain

1 INTRODUCTION Within the context of transition metal complexes containing unsaturated carbene ligands, allenylidene (propadienylidene) derivatives [M]=C=C=CR'R2 have attracted a great deal of attention in recent years as a new type of intermediates with applications in organic synthesis via stoichiometric and catalytic transformations.' In 1982, a simple synthetic approach to allenylidene complexes, based on the spontaneous dehydration of propargylic alcohols HC=CC(OH)R' R2 u on coordination to unsaturated metallic fragments, was reported by Selegue (Chart 1).F

/ OH I H-C'=-C-G,,,

A R R'

H I [ M I A

0

R'

R~\\'~-O~\-H

H20

4

\'

R2

[M]=C=C=q

/

[M]=C=C,

H ,OH

c. [MI,

R'

H (/

,H

c*

*R2

4R3 : 1

c\

,OH

R' dCG0R2

[M]=C=C,/ Rl'

Chart 1 General methodfor the preparation of allenylidene complexes

R3

c=<

R4

286

Perspectives in Organometallic Chemistry

Following this general method, a wide range of allenylidene complexes have been synthesized and characterized with a variety of metal fragments.' Different types are presently known including square-planar, penta-, hexacoordinate and half-sandwich derivatives of Group 6-9 transition metals. However, the presence of hydrogen atoms in p position with respect to the hydroxy group seems to be a s nthetic drawback since Y tautomeric vinylvinylidene species, i.e. [M]=C=C(H)C(R')=CR R3, can also be formed (Chart l).' The reactivity of allenylidene complexes is clearly marked by the strong polarization of the three carbon atoms of the cumulenic chain which, as evidenced by theoretical calculations, are alternatively electron-deficient and electron-rich when moving, starting from the metal: [M]=C$=C:-=Cy6+.3 Thus, while electrophiles add selectively to the Cp carbon atom yielding vinylcarbyne derivatives [M]=CC(E)=CR'R2 (A), the nucleophilic attacks can take place both at the C, or C, atoms affording metal-allenyl [MI-C(Nu)=C=C R'R2 (B)or metal-alkynyl [M]-C=CC(Nu)R'R2(C)complexes, respectively (Chart 2).' Nu [M]-C\ (B)

I

Nu

c\\

7-2

R'

/

R'

[M]=C=C=C,

R2

IE

[M]=C-

Nu

?/

R2

[MI-GC-C,

(C)

Nu

Nu = nucleophile E = electrophile

EI C ' -RZ

(A)

-

R'

I R'

Chart 2 Typical nucleophilic and electrophilic additions on transition metal allenylidenes In the context of our studies in the chemistry of indenyl-ruthenium(I1) complexes: we have reported a wide series of cationic allenylidene derivatives [Ru(=C=C=C R'R2)(q5 -C9H7)(PPh3)2]+. In accordance with the typical reactivity of cationic complexes,' the chemical behaviour of these derivatives is governed by the nucleophilic additions.' The regioselectivity of the nucleophilic attacks (C,vs C,) seems to be controlled by the electronic and steric properties both of the metallic fragment, the nucleophile used and the substituents on the unsaturated hydrocarbon chain.' Nevertheless, as a general trend, regioselective C, additions are usually observed when electron-rich andor bulky metallic fragments are used.' In particular, we have demonstrated the ability of the moiety [Ru(q5-C9H7)(PPh3)2]+ to protect sterically the electrophilic C, atom of the cumulenic chain allowing the regioselective nucleophilic addition of a large variety of organic substrates at the more accessible C, atom! In this way, we have been able to prepare broad series of both neutral and cationic 0-alkynyl complexes of general formula [Ru{C=CC(Nu) R'R2}(q5-C9H7)(PPh&ln+ (n = 0, 1). On the basis of these regioselective nucleophilic attacks, we have recently developed an efficient synthetic procedure for the propargylic substitution of 2-propyn-1-01s mediated by the metallic fragment [Ru(q5-C9H7)(PPh3)2]+ (Chart 3; Method Thus, in a first step, allenylidene complexes A are formed and subsequently transformed into the

In search of asymmetric propargylic substitution reactions

287

corresponding 0-alkynyl derivatives B (Nu- = HC-C-, R3C(=O)CH2-, (CO)sW=C(OMe)CH2-, H~C=CH(CHZ),CH~(n = 0, 1)) which undergo a selective Cp protonation to afford vinylidene complexes C.7Finally, demetalation of C with acetonitrile leads to the functionalized terminal alkynes D in excellent yields. This synthetic methodology constitutes an alternative to the well-known Nicholas reaction (Chart 3; Method B)8 in which propargylic alcohols are easily functionalized via dicobalthexacarbonyl-x-alkynecomplexes containing coordinated propargylic cations (E).9 Although both synthetic methods require the same number of steps, the quantitative recovery of the metal fragment as the acetonitrile complex [Ru(N=CMe)(q5C9H,)(PPh&]+ (Chart 3; Method A) represents a major advantage compared with the classical Nicholas reaction in which the metal auxiliary can not be recovered after the oxidative decomplexation step. Method A: Mediated by [Ru(q5-CgH7)(PPh3h]+

R'

c1- J H20

Nu

[RuI-N-CMJ

@

Method B: Mediated by [CO~(CO)~] (Nicholas reaction)

Chart 3 Progargylic substitutions mediated by [ R u ( $ - C ~ H ~ ) ( P P ~ and ~ ) ~[cO~(co)8] ]+ The efficient access to hnctionalized alkynes D prompted us to use chiral substrates in order to obtain novel optically active terminal alkynes. To achieve this, novel indenylruthenium(I1) allenylidene complexes bearing chiral auxiliaries have been prepared and

288

Perspectives in Organometallic Chemistry

their reactivity towards nucleophiles has been explored. A short account of the most relevant results in this area is presented here. 2 RESULTS AND DISCUSSION 1.1

Activation of Optically Active Propargylic Alcohols by [RuC1(q5-CgH~)(PPh&]

Following the well-known Selegue synthetic protocol: the chiral allenylidene derivatives [Ru{=C=C=C(C9H 16)] (q -C9H7)(PPh3)2][PFs] (C(C9H16) = (1R,4S)- 1,3,3-trimethylbicyclo[2.2.1]hept-2-ylidene (3), (1R,4R)-1,7,7-trimethylbicyclo[2.2.1] hept-2-ylidene (4)) have been easily prepared (ca. 80% yield) by refluxing [RuCl(q5-C9H7)(PPh3)2],NaPF6, and 2-exo-ethynyl-173,3-trimeth1-2-endo-norbornanol (1) or 2-endo-ethynyl-1,7,7Bin methanol (see Scheme 1).6g,’1 trimethyl-2-exo-norbornanol(2),’

C I

H

Scheme 1 Synthesis of the chiral indenyl-ruthenium(I0 allenylidene complexes 3 and 4

Spectroscopic data for compounds 3 and 4 are similar to those reported for related indenyl-ruthenium(11) allenylidene complexes [Ru(=C=C=CR’R2)(q5C9H7)(PPh3)2][PF6]?,e,h! Significantly, the presence of the allenylidene moiety was clearly identified on the basis of: (i) a strong t(C=C=C) IR absorption (asymmetric stretching vibration) at 1963 (3) and 1948 (4) cm-I, and (ii) typical low-field resonances, in the 13C{ ‘H) NMR spectra, for the Ru=C=C=C carbon nuclei (3: & 305.06 (dd, 2J(CP) = 18.9 Hz, RU=Ca), 183.24 (Cp) and 202.01 (C,) ppm; 4: & 304.38 (dd, 2J(CP) = 19.3 Hz, RU=Ca), 185.76 (Cp) and 191.58 (C,) ppm).6g’” Moreover, an unambiguous characterization of complex 4 by single-crystal X-ray analysis was undertaken (Figure 1).l2 Structural

In search of asymmetric propargylic substitution reactions

289

parameters (see caption of Figure 1) are in accord with the well-known description of the metal-allenylidene bonding.' Thus, the structure shows a nearly linear allenylidene fragment with typical Ru-Ca, C,-Cp and Cp-C, bond lengths, the observed differences from the values expected for double bonds indicating an important contribution of the o-alkynyl canonical form [M]-C=C-C+R'R2. C40 c73

Figure 1 ORTEP view of the structure of the optically active allenylidene complex 4. Selected bond lengths (A) and angles (9:Ru(1)-C(1) = 1.875(4), C(I)-C(2) = 1.235(6), C(2)-C(3) = 1.342(6), Ru(l)-C(I)-C(2) = I71.0(4), C(I)-C(2)-C(3) = I76.8(5)

In contrast to the afore-mentioned results, chloride complex [RuCl(q5-C9H7)(PPh3)2] reacts with the optically active propargylic alcohols ethisterone (5a), 17a-ethynylstradiol (Sb) and mestranol (Sc), in methanol and in the presence of NaPF6, to give mixtures containing the desired allenylidene derivatives 6a-c and their vinylvinylidene tautomers 7a-c (Scheme 2).6hThe outcome of these reactions shows the two competitive pathways in the activation of 2-propyn- 1-01s by transition-metal complexes (Chart 1). 2.2. Reactivity of Optically Active Indenyl-Ruthenium(I1) Allenylidene Complexes towards Nucleophiles As expected from our previous studies,6 the allenylidene complex [ R ~ ( = C = C = C ( C ~ H I ~ ) } ( ~ ~ - C ~ H ~ ) ( P P ~ ~(C(C9H16) )~][PF~] = (1R,4S)-1,3,3trimethylbicyclo[2.2.l]hept-2-ylidene (3)) regioselectively reacts at C, with a large variety of anionic nucleophiles (i.e. lithium triethylborohydride, methyllithium, sodium cyanide, lithium phenylacetylide, lithium methyl-enolates and allylmagnesium bromide), in tetrahydrofuran at -2O"C, to afford neutral o-alkynyl derivatives [Ru{C=CC(NU)(C~HI~)}(~~-C~H~)(PP~~)~] (8a-h) in 68-91% yield (Scheme 3).6g'7'1 Analytical and spectroscopic data of compounds 8a-h support the proposed formulations. In particular, the formation of a a-alkynyl chain was clearly confirmed on the basis of ( i ) the presence of a typical ~ C E C absorption ) band in the IR spectra at 2069-2089 cm-l, and (ii) characteristic resonances in the I3C-{IH) NMR spectra for the Ru-CGC carbon nuclei at 84.36 (dd, 2J(CP) = 22.0-23.8 Hz, Ru-C,) and 108.82-116.40 (Cp) ppm.

Perspectives in Organometallic Chemistry

290

Me f,i-C=C-H

&c=c-H&b /

0

/

RO

R = H (5b), Me (5c)

(5a) HO

m i

C

//

H [Ru]- C1

{B

rH

Me

\C=C=[Ru]

+

H

H (6a-c) (7a-c) Scheme 2 Activation of propargylic alcohols derivedfiom hormonal steroids

[RUI

(8a-h)

Nu = H ( 8 4 , Me (8b), CkN (8c), G C P h @a),CH2C(=O)Me(8e), CH,C(=O)Ph (80, CH2C(=O)'Pr (8g), CH2CH=CH2(8h)

Scheme 3 Regio- and diastereoselective nucleophilic additions on allenylidene complex 3 Formation of complexes 8a-h involves the generation of a novel stereogenic center at C,. Remarkably, all the nucleophilic additions proceed in a diastereoselective manner since only one diastereoisomer was detected by NMR spectroscopy. The X-ray crystal structure of complex 8g (Figure 2) shows that these nucleophilic attacks take place on the less sterically congested exo face of the allenylidene chain in [ R U ~ = C = C ' C ( C ~ H ~ ~ ) > ( ~ ' C9H7)(PPh3)2][PF6] (3)-6g The efficient access to a-alkynyl complexes 8a-h in optically pure form prompted us to use prochiral nucleophiles which, by addition to the allenylidene chain in 3, could be able to generate two new diastereogenic centers. With this in mind, the reactivity of 3 with the lithium enolates derived from cyclopentanone and cyclohexanone was explored.6'

In search of asymmetric propargylic substitution reactions

29 1

Thus, whilst 3 remains unchanged upon addition of a large excess of the cyclohexanoneenolate (probably due to steric reasons), it readily reacts with a slight excess (ca. 3:l) of the cyclopentanone-enolate affording the corresponding 0-alkynyl derivative 8i (Scheme 4). Although the cyclopentanonic fragment adds stereoselectively through the ex0 face of the chiral auxiliary CgH16, complex 8i has been obtained as a mixture of two diastereoisomers (ca. 2:l ratio) which can be easily separated by column chromatography. The moderate overall diastereoselectivity (33.3% de) observed for this nucleophilic addition clearly indicates that the allenylidene substituent (1R,4S)-1,3,3trimethylbicyclo[2.2.l]hept-2-ylidene presents a poor chiral induction with respect to the incoming cyclic enolate. C25 C24

c19 \I

C14

c10

II

c7

Figure 2 ORTEP view of the structure of a-alkynyl complex 8g (Nu = CHK(=O)'Pr). Selected bond lengths (A) and angles (g:Ru(1)-C(1) = 2.03(3), C(l)-C(2) = 1.21(3), C(2)-C(3) = 1.55(4), Ru(l)-C(l)-C(2) = 173(3), C(l)-C(2)-C(3) = 168(3)

Scheme 4 Nucleophilic addition of the cyclopentanone-enolate on allenylidene complex 3

292

Perspectives in Organometallic Chemistry

The reactivity of the allenylidene complex [Ru{=C=C=C(C9H16)}(q5C9H7)(PPh&] [PFs] (C(C9H 16) = (1R,4R)- 1,7,7-trimethylbicyclor2.2.13 hept-2-y lidene (4)) towards anionic nucleophiles has been also explored.l 1 Unfortunately, deprotonation of one of the acidic methylenic protons at CS seems to be favoured versus the nucleophilic C, addition since the o-enynyl complex 9 has been, in all cases, selectively obtained (87% yield; Scheme 5). The related chiral o-enynyl derivatives 10a-c (Figure 3) were also selectively obtained (7 1-83% yield) when the mixtures containin allenylidenes 6a-c and 6F vinylvinylidenes 7a-c were treated with anionic nucleophiles.

&’

[Ru]=C=C=

PF,I *.

THF / -20°C NuH

(4)

(9)

Scheme 5 Reactivity of the allenylidene complex 4 towards anionic nucleophiles

( W

R = H (lob), Me (10c)

Figure 3 Structure of a-enynyl complexes I Oa-c derivedfiom hormonal steroids

2.3. Synthesis of Optically Active Terminal Alkynes from Functionalized IndenylRuthenium(I1) a-Alkynyl Derivatives Demetalation of the functionalized a-alkynyl complexes [Ru{c ~ c c ( h ) ( c ~ H ~ 6 ) } ( ~ 5 C9H7)(PPh&] (C(C9H16) = (1R,4S)-1,3,3-trimethylbicyclo[2.2.l]hept-2-ylidene) (8a-h) using our two-step method (Chart 3; Method A) proceeds cleanly, yielding the optically

In search of asymmetric propargylic substitution reactions

293

active terminal alkynes HCrCC(Nu)(C9H16) (12a-h) in good overall yields (Scheme 6).6gJt1 Thus, in a first step vinylidene derivatives [Ru{=C=C(H)C(Nu)(C9H16)}(q5CgH7)(PPh3)2][BF4] (1la-h) were prepared (78-96% yield), via selective Cp protonation of 8a-h with HBF4: and fully characterized by means of standard spectroscopic techniques and elemental analyses. The most relevant spectroscopic feature of these complexes is the presence in the 13C-{lH}NMR spectra of highly deshielded signal at & 328.25-345.48 (dd, 2J(CP) = 14.1-19.5 Hz) ppm typical of a carbenic [M]=C, carbon atom.7 In a second step vinylidene complexes lla-h were treated with acetonitrile, resulting in the liberation of the free chiral terminal alkynes Hc=CC(Nu)(C9H16) 12a-h (77-99% isolated yields) and the quantitative recovery of the metal auxiliary as the cationic solvato complex [Ru(N=CMe)(q5-CgH7)(PPh&][BF4] (13).

WI

"'6 Nu\

,c

#

[4 C / I /

[RUI

HBF4 P

Et2O HBF4 / -20°C

1

(8a-h)

N=CMe

(lla-h)

[RUI

+

>

r.t.

H'

(12a-h)

Nu = H (a), Me (b), 6 N (c), G C P h (a), CH2C(=O)Me(e), CH2C(=O)Ph(0, CH2C(=O)'Pr (g), CH2CH=CH2(h)

Scheme 6 Synthesis of optically active terminal albnes @om o-albnyl complexes 8a-h This demetalation process proceeds through the initial tautomerization at the ruthenium center of the q'-vinylidene group to the corresponding q2-terminal alkyne (Figure 4) and subsequent elimination of the organic fragment from the metal by exchange with acetonitrile.6d

Perspectives in Organometallic Chemistry

294

/"

L

H

II

[Ru] Figure 4 Proposed ?f-vinylidene-#-alkyne tautomerizationprocess Alkynes 12a-h were easily purified from the reaction mixture by column chromatography on silica gel, after filtering off the unsoluble nitrile complex 13, and spectroscopically characterized. In particular, characteristic acetylenic CGCH proton and carbon resonances are observed in the NMR spectra at ca. & 2, and & 74 (ECH) and 87 (EC) ppm, respectively. Similar reactions conducted with each one of the two diastereoisomers of o-alkynyl complex 8i (Scheme 4) allow the selective formation of the optically active y-ketoacetylenes 12i containing four stereogenic carbon atoms (Figure 9.62

/L

H

IL

H

Figure 5 Structure of the optically active y-keto-acetylenes12i

3 CONCLUSIONS We have recently demonstrated that allenylidene chains stabilized by the electron-rich and bulky indenyl-ruthenium(I1) fragment [Ru(.a5-C~H7)(PPh3)2]+can be considered as synthons of propargyl cations HC3X+R'R2. dJg,'J This fact allowed us to develop an efficient synthetic methodology for the propargylic substitution of 2-propyn-1-01s HC=CC(OH)R'R2 with nucleophiles leading to the corresponding functionalized terminal alkynes HC-CC(Nu)R'R2 (Chart 3; Method A), which constitutes an alternative to the well-known Nicholas reaction (Chart 3; Method B).I3 Although this method may not be quite as accessible as the Nicholas reaction to which it is compared, it is straightforward enough to be a valid alternative for synthetic targets for which the Nicholas reaction cannot be readily adapted. This is illustrated by the coupling of homoallyl Grignard reagents with ruthenium allenylidene complexes', i.e. 8h, since the corresponding Nicholas reaction fails due to the decomposition of cobalt-stabilized propargylic cations by Grignard reagents.l4 In this paper the asymmetric version of this synthetic route has been shown starting from allenylidene complexes bearing stereogenic centers, readily available from chiral

In search of asymmetric propargylic substitution reactions

295

propargylic alcohols. This methodology discloses an efficient approach for the synthesis of a large variety functionalized optically active terminal alkynes in good yields. Acknowledgements. Authors gratefully acknowledge the financial support provided by the Ministerio de Ciencia y Tecnologia (MCyT) of Spain (Project BQU2000-0227) and the Gobierno del Principado de Asturias (Project PR-01-GE-4). Special gratitude is expressed to all the co-workers whose names appear in the reference list.

References For reviews see: (a)H. Le Bozec and P. H. Dixneuf, Russ. Chem. Bull., 1995,44,801; (b) H. Werner, Chem. Commun., 1997, 903; (c) M. I. Bruce, Chem. Rev., 1998, 98, 2797; (d)D. Touchard and P. H. Dixneuf, Coord. Chem. Rev., 1998,178-180,409; (e) V. Cadierno, M. P. Gamasa and J. Gimeno, Eur. J Inorg. Chem.,2001,571. J. P. Selegue, Organometallics,1982, 1,217. (a)B. E. R. Schilling, R. Hoflkann and D. L. Lichtenberger, J Am. Chem. SOC., 1979, 101,585; (b) H. Berke, G. Huttner and J. Von Seyerl, 2. Naturforsch, Teil B, 1981,36, 1277; (c) V. Cadierno, M. P. Gamasa, J. Gimeno, M. C. Lopez-Godlez, J. Borge and S. Garcia-Granda, Organometallics, 1997, 16, 4453; (d) A. J. Edwards, M. A. Esteruelas, F. J. Lahoz, J. Modrego and L. A. Oro, Organometallics, 1996, 15, 3556; (e) M. A. Esteruelas, A. V. Gomez, A. M. Lopez, J. Modrego and E. Oiiate, Organometallics, 1997, 16, 5826; v) N. Re, A. Sgamellotti and C. Floriani, Organometallics, 2000, 19, 1115; (g) M. Baya, P. Crochet, M. A. Esteruelas, E. Gutierrez-Puebla, A. M. Lopez, J. Modrego, E. Oiiate and N. Vela, Organometallics, 2000, 19, 2585; (h) R. F. Winter, K. W. Klinkharmner and S. Zalis, Organometallics, 2001,20,1317. V. Cadierno, J. Diez, M. P. Gamasa, J. Gimeno and E. Lastra, Coord. Chem. Rev., 1999,193-195, 147. The protonation of compounds [Os(=C=C=CPh2)($-CSHS)(P'P~~>~] [PF6], [Ru{=C=C=C(R)Ph}($-CsMes)( d-P,P-'Pr2PCH2CH2PiPr2)][BFq] (R = H, Ph), [RuCl(=C=C=CPh2)(d-P,O-Cy2PCH2CH20CH3)2][PF6] and [Re(=C=C=CPh2)(C0)2 { A?-P,P,P-M~C(CH~PP~~)~}] [CF3S03] to give the corresponding dicationic vinylcarbyne derivatives, i. e. [Os{-CC(H)=CPh2}( $-CSHS)(P'P~~)~][PF~]~, [Ru{rCC(H)=C(R)Ph}( $-CsMes)( d-P,P-'Pr2PCH2CH2PiPr2)][BF&(R = H, Ph), [RuCl(=CC(H)=CPh2)(2-P, O-CY~PCH~CH~OCH~)( K' -P-CY~PCH~CH~OCH~)] [PF& and [Re{rCC(H)=CPh2)(C0)2{I~?-P,P,P-M~C(CH~PP~~)~}][CF$~O~]~, are the only examples of electrophilic additions in cationic complexes reported to data: (a) S. Jung, C. D. Brandt and H. Werner, New J Chem., 2001, 25, 1 101; (b) E. Bustelo, M. JimCnez-Tenorio, K. Mereiter, M. C. Puerta and P. Valerga, Organometallics, 2002, 21, 1903; (c) N. Mantovani, L. Marvelli, R. Rossi, C. Bianchini, I. de 10s Rios, A. Romerosa and M. Peruzzini, J Chem. Soc., Dalton Trans., 2001, 2353; See also ref. 3g (a) V. Cadierno, M. P. Gamasa, J. Gimeno, M. Godlez-Cueva, E. Lastra, J. Borge, S. Garcia-Granda and E. PCrez-Carreiio, Organometallics, 1996, 15, 2137; (b) V. Cadierno, M. P. Gamasa, J. Gimeno, J. Borge and S. Garcia-Granda, Organometallics, 1997, 16, 3178; (c) V. Cadierno, S. Conejero, M. P. Gamasa, J. Gimeno, I. Asselberghs, S. Houbrechts, K. Clays, A. Persoons, J. Borge and S. Garcia-Granda, Organometallics, 1999, 18, 582; (d)V. Cadierno, M. P. Gamasa, J. Gimeno, E. PCrezcarreiio and S. Garcia-Granda, Organometallics, 1999, 18, 2821; (e) V. Cadierno, M.

296

7

8

9 10

11 12 13

14

Perspectives in Organometallic Chemistry

P. Gamasa, J. Gimeno and E. Lastra, J. Chem. SOC.,Dalton Trans., 1999,3235; (f) V. Cadierno, S. Conejero, M. P. Gamasa and J. Gimeno, J Chem. Soc., Dalton Trans., 2000, 451; (g) V. Cadierno, S. Conejero, M. P. Gamasa, J. Gimeno, E. P&ez-Carreiio and S . Garcia-Granda, Organometallics,2001,20, 3 175; (h) V. Cadierno, S. Conejero, M. P. Gamasa, J. Gimeno and M. A. Rodriguez, Organometallics,2002, 21, 203; ( i ) V. Cadierno, S. Conejero, M. P. Gamasa, J. Gimeno, L. R. Falvello and R. M. Llusar, Organometallics,in press; 0') V. Cadierno, S. Conejero, M. P. Gamasa and J. Gimeno, Organometallics,in press; (k) S. Conejero, J. Diez, M. P. Gamasa, J. Gimeno and S. Garcia-Granda, Angew. Chem.,Int. Ed., in press. For reviews on the synthesis and reactivity of vinylidene complexes see: (a) M. I. Bruce, Chem. Rev., 1991, 91, 197; (b) H. Werner, J Organomet. Chem., 1994, 475, 45; (c) C. Bruneau and P. H. Dixneuf, Acc. Chem. Res., 1999, 32, 3 1 1 ; (d) M. C. Puerta and P. Valerga, Coord. Chem. Rev., 1999,193-195,977. For reviews on the Nicholas reaction see: (a) K. M. Nicholas, Acc. Chem. Res., 1987, 20, 207; (b) W. A. Smit, R. Caple and I. P. Smoliakova, Chem. Rev., 1994, 94, 2359; (c) G. G. Melikyan and K. M. Nicholas, in Modern Acetylene Chemistry, eds. P. J. Stang and F. Diederich, VCH, New York, 1995, p. 118; (d) A. J. M. C a m and K. M. Nicholas, in Comprehensive Organometallic Chemistry II, eds. E. W. Abel, F. G. A. Stone and G. Wilkinson, Pergamon, New York, 1995, vol. 12, p. 685; (e)J. R. Green, Curr. Org. Chem., 2001,5,809; v) B. J. Teobald, Tetrahedron,2002,4133. For a review on transition-metal stabilized propargyl cations see: H. El h o u r i and M. Gruselle, Chem. Rev., 1996,96, 1077. Propargylic alcohols 1 and 2 have been prepared starting from the commercially available chiral ketones (-)-fenchone and (+)-camphor as described in: (a) M. M. Midland, J Org. Chem., 1975, 40, 2250; (b) D. S. Keegan, M. M. Midland, R. T. Werley and J. I. McLoughlin, J Org. Chem., 1991,56, 1 185. S. Conejero, Ph.D. Thesis, University of Oviedo, 2001. S. Conejero, M. P. Gamasa, J. Gimeno, E. Perez-Carreiio and S. Garcia-Granda, unpublished results. Related ruthenium-catalyzed propargylic substitutions of monosubstituted propargylic alcohols, via allenylidene intermediates, have been recently reported: (a) Y. Nishibayashi, I. Wakiji and M. Hidai, J. Am. Chem. SOC., 2000, 122, 11019; (b) Y. Nishibayashi, I. Wakiji, Y. Ishii, S. Uemura and M. Hidai, J Am. Chem. SOC.,2001, 123,3393; (c) Y. Nishibayashi, Y. Inada, M. Hidai and S. Uemura, J: Am. Chem. SOC., in press. S . Padmanabhan and K. M. Nicholas, J Organomet. Chem., 1981,212,115.

RECENT DEVELOPMENTS ON HYDRIDE IRIDIUM TRIISOPROPYLPHOSPHINE COMPLEXES: [IrH2(NCCH3)3(PiPr3)]BF4AS HYDROGENATION CATALYST

Luis A. Oro, Eduardo Sola and Janeth Navarro Departamento de Quimica Inorghica, Instituto de Ciencias de Materiales de Aragh, Universidad de Zaragoza-C.S.I.C.,5O0O9-Zaragoza7Spain

1 INTRODUCTION Osborn and co-workers discovered cationic rhodium and iridium complexes of general formula [M(diene)L,]+, which were recognised as excellent hydrogenation catalyst precursors for a variety of L ligand combinations.’ Some iridium derivatives of this type, specially those containing both a phosphine and a N-donor ligand, give rise to exceptional catalytic activities, being also able to reduce very hindered olefinic substrates.2 Under catalytic conditions, these precursors are known to undergo diene hydrogenation, affording coordinatively unsaturated active species stabilised by solvent co-ordination. As a consequence, the catalytic activity of such systems is maximum in weakly co-ordinating solvents, even though, in the absence of substrates, such solvents are unable to protect the active species from aggregation processes leading to non-active c1uste1-s.~ The hydrogenation reaction leading to these solvated active species has been used by Crabtree and others with synthetic purposes, in the preparation of bis-solvato-compounds of general formula [IrH2(S)2(PR3)2]+, starting from bis-phosphine-diene precursors, or in the presence of an excess of a bulky ph~sphine.~ The bis-water or bis-acetone derivatives of this type are good starting materials for the synthesis of man other iridium compounds, and are also effective in a variety of catalytic Following and extending this synthetic strategy, we have found that the direct hydrogenation of the cationic precursors [Ir(diene)(S)(PiPr3)]+,in the absence of added phosphine, leads to the formation of related tris-solvato-complexes of formula [IrH2(S)3(PiPr3)]+. The tris-acetonitrile derivative IrH2 CCH3)3(PiPr3)]BF4 (l),initially synthesised through this hydrogenation procedure, can also be prepared by reaction of a conventional iridium starting material, the dimer [Ir(pOMe)(cod)]2, with the phosphonium salt [HPiPr3]BF4 in the presence of dihydrogen and acetonitrile (Scheme 1).6 A similar synthetic procedure allows the preparation of the arene compound [I~Hz(~~-C~H~)(P~P~~)]BF~, in which the three fac solvent positions are occupied by a q6-benzeneligand.’ The complex 1 has been found to be a good hydrogenation catalyst, and has demonstrated a high value for mechanistic investigation. In fact, the spectroscopic studies of this compound in the presence of alkenes, alkynes, and dihydrogen have allowed the observation and characterisation of reaction intermediates likely involved in homogeneous hydrogenation, as well as the identification of other feasible side reactions which compete with those leading to hydrogenation. This has allowed direct comparison among most

transformation^!^

I ”

298

Perspectives in Organometallic Chemistry

elementary organometallic reactions, and the conception of new catalytic transformations under hydrogenation conditions. Most noticeable conclusions of these studies are presented below. H2

[Ir(cod)(CHgCN)(PiPrg)]BF4

1

CH3CN

r CH3CN

1

BF4

/

,

iPr3P

H

Scheme 1 2 ALKENE HYDROGENATION CATALYSED BY [IrH2(NCCH3)3(PiPr3)]BF4 The three acetonitrile ligands of complex [IrH2(NCCH3)3(PiPr3)]BF4 (1) are labile, being readily replaced by acetonitrile-d3 in solution at room temperature. A detailed kinetic study of these substitution reactions revealed that any of the acetonitrile ligands co-ordinated trans to hydride can be easily dissociated (AI$ = 26.6 f 1 KcaVmol and AS' = 32 f 2 eu.) to give a five-co-ordinate species. This unsaturated species is fluxional, allowing the formation of an intermediate with the co-ordination vacancy trans to phosphine (AH' = 30 f 3 KcaVmol, ASs = 34 f 4 eu.). The life-time of this latter intermediate is long enough to allow substitution reactions at the position trans to phosphine (Scheme 2).

1

Scheme 2 Given that the activation energies required to generate these two co-ordination vacancies are small and rather similar, the selectivity of substitution reactions is most likely dictated by thermodynamic factors. In fact, the substitution reactions of Scheme 3 suggest that small incoming ligands would prefer co-ordination at the most labile position, trans to hydride, whereas bulkier ones substitute at the less hindered position, trans to phosphine. The spectroscopic observation of the reaction of complex 1 with ethene reveals consecutive formation of several products, as illustrated in Scheme 4. The complex

Recent developments on hydride iridium triisopropylphosphine complexes

299

[IrH2( #-C~&)(NCCH~)~(P~PT~)]BF~ (2) is observed after a slow stream of ethene is passed through a solution of 1 in CDCl3 at 233 K. On raising the temperature to 273 K, and in the presence of an excess of dissolved ethene, complex 2 is converted into the diethyl complex [Ir(Et)2(NCCH3)3(PiPr3)]BF4 (5), as a result of two ethylene insertions into the Ir-H bonds. The spectroscopic control of this reaction reveals the appearance and subsequent disappearance of two intermediates, which can be identified as compounds 3 and 4. The formation of complex 5 from 1is a reversible process. Thus, if argon is bubbled through a solution of 5 at 273K, a quantitative formation of 1occurs. However, if solutions of 5 are warmed to room temperature, the diethyl-complex disappears and the elimination of ethane is observed. Since the formation of butane can not be detected, it is possible to conclude that a pelimination from an ethyl ligand of 5 is kinetically favoured over the C-C reductive coupling of the mutually cis-disposed ethyl ligands.

I

CH3CN

Scheme 3

iPr3P, CH3CN' CH3CN

Irt H,

I

CH3CN

3

1

CH3CN

-Ir -NCCH3

I

CH3CN

6

tl

l+

pi~r3 CH3CN

-I Ir

I

II

CH3CN

7

Scheme 4

NCCH3

e

300

Perspectives in Organometallic Chemistry

In the presence of an excess of ethene, the NMR spectra of the white solution resulting upon the reductive elimination of ethane from 3 display several broad signals. After removal of the ethene excess, the colour of the solution changes to orange, and the observed NMR spectra indicates that the solution contains a mixture of the complexes [Ir (NCCH3)3(PiPr3)]BF4 (6) and [Ir(#-C2J&)(NCCH3)2(PiPr3)]BF4 (7), in a ratio which depends on the time employed for the removal of ethene. These species readily react with H2 to regenerate the starting complex 1, closing a cycle for ethene hydrogenation in which most of the participating species have been spectroscopicaly characterised. The ratedetermining step of this catalytic cycle is the reductive elimination of ethene, which is the only reaction that requires temperatures above 273 K to proceed. The reactions of complex 1 with propene reveal some noticeable differences compared to the processes involving ethene. As shown in Scheme 5 , the initial step consists of the formation of the complex [IrH2( $-C3Hb)(NCCH3)2(PiPr3)]BF4 (8), which is observed upon treatment of 1with propene at 233 K. In 8, the alkene ligand co-ordinates trans to the phosphine, in contrast to the cis mutual position of these ligands found in the analogous ethene complex 2. On raising the temperature to 273 K, the propene ligand inserts into one of the Ir-H bonds to yield the complex [IrH(n-Pr)(NCCH3)3(PiPr3)]BF4 (9), which does not undergo any observable reaction at this temperature. It appears that the regioselectivity of this insertion is very high, since a species containing the isomeric 2-propenyl ligand can not be detected.

I

l+

i P r 3 P Ir ~ /H CH3CN(

I\/rCH3

CH3CN

1

H2 CH3CN

10

tl CH3CN

11

Scheme 5

Recent developments on hydride iridium triisopropylphosphine complexes

301

On warming to room temperature, compound 9 eliminates propane; however, the expected iridium(1) species resulting fiom this abstraction can not be observed. Instead, a mixture of allylic C-H activation products is formed, namely the allyl-hydride derivatives [IrH(+-C3H5)(NCCH3)2(PiPr3)]BF4 (10) and [IrH($-C3H5)( 7jZ-C,Ha)(NCCH3)(PiPr3)] BF4 ( l l ) , respectively. Both complexes 10 and 11 have been isolated and fully characterised by analytic and spectroscopic methods, and the structure of 11 has been determined by X-ray diffraction. The cycle for the hydrogenation of propene outlined in Scheme 5 is closed by the reaction of complex 10 with dihydrogen in the presence of one equivalent of acetonitrile. At room temperature and one atmosphere of dihydrogen, the formation of complex 1 goes to completion within a few minutes. The NMR spectra of the reaction mixtures formed with substoichiometric amounts of H2 indicate that the formation of 1 occurs with the simultaneous release of propene (not propane). Under the same experimental conditions, the reaction with deuterium leads to the formation of [IrHD(NCCH3)3(PiPr3)]BF4 (1-6) as the major product, together with very small quantities of complexes 1 and 1-d2, the latter deuterated at both hydride positions. With regard of this lack of H/Dscrambling, it appears likely that the elimination of propene fiom the allyl-hydrido complex 10 results from a 0bond metathesis process in a transient dihydrogen complex. The features of the hydride-ally1 complex 10 merit further comment. Along with the CH activated propene, this complex presents two labile acetonitrile ligands. In this case, the kinetic studies on acetonitrile substitution processes indicate that both acetonitrile ligands can be dissociated (Scheme 6), with activation parameters of AHs = 18.4 f 1 Kcal/mol and ASt = 8 f 2 eu., and Ms= 23 f 3 KcaVmol and ASt = 8 f 4 eu. for the ligand trans to hydride and trans to allyl, respectively. The simultaneous presence of a C-H activated alkene and two readily accessible co-ordination positions points to the potential use of such compound in the activation and subsequent functionalisation of alkenes. Further studies in order to develop this methodology are in progress.

10 Scheme 6 3 1-ALKYNE HYDROGENATION CATALYSED BY [IrH2(NCCH3)3(PiPr3)]BF4 A noticeable feature of the cycle in Scheme 3 is the double alkene insertion leading to the bis-ethyl compound 5. Unfortunately, with ethylene as substrate, such double insertion is non-productive, given that the hydrogen pelimination from 5 to reform 2 seems to be kinetically favoured over the reductive elimination of butane. This unfavourable relationship between pelimination and C-C reductive elimination rates can be overcome by using 1-alkynes as substrates. Thus, the reaction of 1 with an excess of 1-alkynes such as PhC=CH or tBuC=CH afford butadiene complexes (Scheme 7).

Perspectives in Organornetallic Chemistry

302

R=

1

mu,12 R = Ph,13b

Ph,13a

Scheme 7 The stereoselectivity of these reactions is dependent on the 1-alkyne substrate. Thus, the reaction with tBuC=CH selectively affords the compound [Ir(q4-1,3-(tBu)2C4&) (NCCH3)2(PiPr3)]BF4 (12), whereas the reaction with PhCECH is less selective, leading to the compound [Ir(( q4-(Ph)2C4H4)(NCCH&(PiPr3)]BF4 (13) as a 7: 1 mixture of isomers which contain 1,3- (13a) and 1,4-diphenylbutadiene (13b) ligands.’ Monitoring of these reactions by NMR spectroscopy, in CDCl3 at low temperature, allows the observation of alkenyl-hydride intermediates which are consistent with the expected double insertiodc-C coupling reaction sequence (Scheme 8).

a-insertion

P-insertion

R = mu,15

R = Wu, not observed

14a (50%)

RC&H

RG CH

not observed ‘1-

I

I

/R

7

R = mu,12 (100%) Ph, 13a (87%)

rI: -

R

R

w

R = Ph, 13b (13%)

Scheme 8 The treatment of 1 with one equivalent of PhC=CH at 253 K affords an equimolar mixture of two isomeric alkenyl-hydride complexes, 14a and 14b, which have a-and (2)-alkenyl ligands, respectively. Both isomers are stable towards the reductive elimination of styrene at room temperature. The reactivity of these two isomers towards a second equivalent of the alkyne is different. Thus, 14b readily reacts with an excess of the alkyne already at 253 K giving 13, whereas the disappearance of the a-alkenyl-hydride 14a is slow even at temperatures above 273 K. In both cases, the spectroscopic observations do not allow the detection of additional intermediates of this second reaction step. This suggests that, once the second insertion takes place, the subsequent C-C reductive coupling

Recent developments on hydride iridium triisopropylphosphine complexes

303

is fast. The isomeric distribution resulting from the evolution of the alkenyl-hydride intermediates suggests that a sequence of two a-insertions is not favoured. Moreover, after an initial p(Z)-insertion, a subsequent insertion of a-stereochemistry seems to be preferred. The reactions of the butadiene compounds 12 and 13 with dihydrogen, in CDC13 and in the presence of an excess acetonitrile, afford alkenes, regenerating the starting complex 1 (Scheme 9). This reaction closes a possible cycle for hydrogenative dimerisation of terminal alkynes which may operate at room temperature at atmospheric pressure of dihydrogen.

12

1

I

Scheme 9 Interestingly, the reaction sequence leading to products 16-18 is also the major reaction path when tBuC=CH and dihydrogen are treated with catalytic amounts of complex 1 in 1,2-dichloroethane (Figure la). In contrast, under the same catalytic conditions, phenylacetylene affords, almost exclusively, simple hydrogenation products: styrene and ethylbenzene. With regard to the aforementioned experimental observations and mechanistic proposals, this dependence of the selectivity on the alkyne may arise from the competition of substrates for the alkenyl-hydride reaction intermediates. Thus, in the case of tBuC=CH, the second alkyne insertion in 15 seems to be kinetically favoured over its reaction with dihydrogen, whereas this latter reaction seems to be the fastest alternative for intermediates 14. In agreement with such rationalisation based on substrate competition, it can be observed that the reaction selectivity depends upon the relative concentration of the reactants. Thus, the proportion of tBuCH=CH2 and tBuCH2CH3 hydrogenation products has been found to increase at low alkyne concentrations (Figure lb). Under the latter conditions, the lower substratekatalyst ratio leads to a faster isomerization of the kinetic product 16 into its internal isomers 17 and 18, as well as to faster tBuCH=CH2 to tBuCH2CH3 hydrogenation. The decrease of the initial alkyne concentration has also been found to result in higher initial hydrogenation rates. Similar substrate inhibition effect has been observed with PhC=CH as substrate, revealing a complex dependence of the hydrogenation rate upon the alkyne concentration, currently under study. To the best of our knowledge, a catalytic hydrogenative dimerisation such as that leading to 16-18 has not been previously reported.

4 CONCLUSION The labile [IrH2(NCCH&(PiPr3)]BF4 complex has allowed identification of several reaction intermediates, and observation of various elementary reactions involved in the catalytic hydrogenation of alkenes. From the structures deduced for these intermediates, it can be concluded that the steric properties of the alkene substrates are important to define

Perspectives in Organometallic Chemistry

304

the position at which the substrate co-ordinates to the metal. From the elementary reactions observed, it can be concluded that the C-H reductive elimination of the alkane is the ratedetermining step of the hydrogenations catalysed by this iridium complex. Other processes involved in the deduced cycles such as oxidative additions or a-bond metathesis, and insertions have been found to require lower activation barriers. This dihydride complex has allowed the observation of unusual double insertion reactions. Such processes do not contribute to the reaction outcome in the catalytic hydrogenation of alkenes, but lead to unusual hydrogenative dimerisations when 1-alkynes are used as substrates. The labile species generated in the hydrogenation cycles are convenient precursors for the formation of Ir(1) complexes, which are active in the oxidative addition of C-H bonds. Interestingly, the products resulting from the C-H activations are still capable of generating new co-ordination vacancies with low kinetic barriers, which suggests the potential use of these compounds in C-H activatiodfunctionalisation sequences.

(a)

3000

2000

time Imin)

4000

16-18 product distribution

/ -

.

18

- "*

E

2.0

1000

3000

2000

(b)

300

200 1

I

4000

.

1

time (min)

400 .

1

I

time (min)

.

'I0.4

16-18 product distribution

b

100

200

300

400

time(min)

Figure 1. Reaction projles for the hydrogenation of tBuC€H catalysed by 1.Conditions: 1,2-dichloroethane (8 mL), 293 K , P(Hd = 1.1 bar; (a) [ l ] , = 0.010 mmol,[tBuC€H], = 5.0 mmol;(b) [ l ] , = 0.015 mmol,[tBuC=CH], = I . 5 mmol.

Recent developments on hydride iridium triisopropylphosphine complexes

305

References

(a) R.R. Schrock and J.A. Osborn, J. Am. Chem. SOC., 1976, 98, 2134. (b) R.R. Schrock and J.A. Osborn, J. Am. Chem. SOC.,1976,98,4450. (c) R.R. Schrock and J.A. Osborn, 3: Chem. Soc., Chem. Commun., 1970, 567. (d) P.A. Chaloner, M.A. Esteruelas, F. Joo and L.A. Oro, in Homogeneous Hydrogenation, Kluwer Academic, Dordrecht, 1994 (a) R,H. Crabtree, H. Felkin, T. Fillebeen-Khan and G.E. Morris, J. Organomet. Chem., 1982,168, 183. (b) R,H. Crabtree, H. Felkin and G.E. Morris, J. Organomet. Chem., 1977,141,205.(c) L.A. Oro, J.A. Cabeza, C. Cativiela, M.D. Diaz de Villegas and E. MelCndez, 3: Chem. SOC.,Chem. Commun., 1983, 1383. (d) J.A. Cabeza, C. Cativiela, M.D. Dim de Villegas and L.A. Oro, J. Chem. SOC.,Perkin Trans. I, 1988, 1881. (a) R.H. Crabtree, Acc. Chem. Res., 1979,12,331.(b) R.H. Crabtree, in Homogeneous Catalysis with Metal Phosphine Complexes, ed. L.H. Pignolet, Plenum Press, New York, 1983,p. 285-316. (a) O.W. Howarth, C.H. McAteer, P. Moore and G.E. Morris, J. Chem. SOC., Dalton Trans.,l981,1481.(b) R.H. Crabtree, G.G. Hlatky, C.P. Parnell, B.E. Segmuller and R.J. Uriarte, Inorg. Chem., 1984,23, 354. (c) A.M. Mueting, P.D. Boyle, R. Wagner and L.H. Pignolet, Inorg. Chem. 1988,27,271.(d) A. Habib, R.S. Tanke, E.M. Holt and R.H. Crabtree, Organometallics, 1989,8,1225.(e) M.A. Esteruelas, M.P. Garcia, M. Martin, 0. Niirnberg, L.A. Oro and H. Werner, J. Organomet. Chem. 1994,466, 249. E. Sola, V.I. Bakhmutov, F. Torres, A. Elduque, J.A. Lopez, F.J. Lahoz, H. Werner and L.A. Oro, Organometallics, 1998,17,683. E. Sola, J. Navarro, J.A. Lopez, F.J. Lahoz, L.A. Oro and H. Werner, Organometallics, 1999,18,3534. (a) F. Torres, E. Sola, M. Martin, J.A. Lopez, F.J. Lahoz and L.A. Oro, 3: Am. Chem. SOC., 1999, 121, 10632. (b) F. Torres, E. Sola, M. Martin, C. Ochs, G. Picazo, J.A. Lopez, F.J. Lahoz and L.A. Oro, Organometallics, 2001,20,2716. J. Navarro, M. Sagi, E. Sola, F.J. Lahoz, I.T. Dobrinovitch, A. Katho, F. Joo and L.A. Oro, submitted manuscript.

Pd COMPLEX-CATALYZED RING-OPENING POLYMERIZATION OF 2-ARYL- 1-METHYLENE-CYCLOPROPANES

S. Kim, D. Takeuchi and K. Osakada

Chemical Resources Laboratory, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8503, Japan

1 INTRODUCTION Methylenecyelopropanes, which are the cyclic isomers of 1,3-dienes, have high potential as monomers of the polymerization promoted by transition metal complexes. Transition metal complex-promoted addition of H-X (X = OR, NR;, SiR3 etc) to methylenecyclopropanes often causes ring opening of the substrates. These reactions, involve insertion of the C=C double bond into M-H, M-0, and M-N bond (M: transition metal) and subsequent C-C bond cleavage of the three-membered ring. On the other hand, there have been a few number of reports of the ring-opening polymerization of these molecules triggered by insertion of the C=C double bond into the M-C b ~ n d . ~ Marks - ~ and Jia and their co-workers found that cationic metallocenes of Zr and Lu promote the ringopening polymerization of methylenecyclopropane to produce a new polymer, (CH2-CH2C(=CH2)-)", and the polymer containing partial polyspiropyrane structural units formed via rapid back-biting of the growing polymer end during the polymerization.' The highly strained monomer structure renders insertion of the C=C double bond and subsequent ringcleavage thermodynamically favorable. This paper reports that Pd (11) complex-catalyzed ring-opening polymerization of 2aryl- 1-methylenecyclopropanes leads to a polymer with em-methylene group on the main chain, in which the head-to-tail sequence is well regulated. A part of this work has been published preliminarily.' 2 RESULTS AND DISCUSSION

Ring-opening Polymerization of 2-Phenyl-1-methylenecyclopropabeby Pd Complexes Various palladium complexes with chelating diimine ligands catalyze ring-opening polymerization of 2-aryl-1-methylenecyclopropanes to afford polymers having CH2C(=CH2)-CH(C6&-X)- repeating units in the polymer chain as shown in eq 1.

2.1

Pd complex-catalyzed ring -opening polymerisation

307

I: X = H, 11: X = Me, 111: X = F, IV: X = CI,V: X = OMe

'

Chart 1 summarizes the Pd complexes used as the polymerization catalyst of this study. Cationic n-allylpalladium complexes 1 and 2 catalyze polymerization of 2-phenyl-1methylenecyclopropane at 80 "C in MeCN to give (CH2-C(=CH2)-CHPh-), (I) with M, = 4000 and 5900, respectively. The polymer structure was determined by means of 'H and 13C{'H} NMR spectroscopy. The 'H NMR signals at 6 = 4.3 - 5.0 and the 13C{'H}NMR signals at 6 = 110.6 and 111.2 correspond to the hydrogen and carbon atoms of the =CH2 group. The peak positions are at similar positions to those of a model compound (CH2=CH- CHPh-CH2-CH3) and indicate the presence of the vinylidene group in the polymer. Signals corresponding to -CH=CR (R = H, Ph) and the cyclopropyl group, which should exhibit the the 'H peaks at 6 = 5.5-6.5 and 6 = 1.5-0.0, respectively, are entirely absent in the spectrum. The other NMR signals are assigned to the CH, CH2, and Ph groups of I. The 13C{'H} NMR signal of the ips0 carbon atom of the phenyl group appears as four resonances in the range 6 = 148.3 - 149.6. They are assigned to four possible triads (rr, rm, mr, and mm) with regard to the chiral benzyl carbon atom in the repeating unit (Figure 1(a)). A similar poly(styrene-co-CO) with a regulated head-to-tail linkage of the monomer units also exhibits four distinct 13C{'H} NMR peaks at 6 = 134.8 - 136.5 for the ips0 phenyl carbon atom, whereas the polymer with a disordered linkage shows complicated peaks in a wider region (Figure l(b)).9 These NMR data indicate that the produced polymer I has a well-regulated head-to-tail sequence of the single repeating unit, -CH2-C(=CH2)-CH(Ph)-.

Chart 1 Pd Complexes Used as the Polymerization Catalyst

Perspectives in Organometallic Chemistry

308

(a)

(b)

I

H-T Syndiotactic

H-T Atactic Y

H-T, H-H, T-T I

I

I

150

149

148

ppm

137

136

PPm

135

Figure 1 13C{'H) NMR spectrum of (a) I and (b) poly(styrene-co-CO). (b) is taken fiom ref: 9.

Polymer I with a lower molecular weight was prepared from the reaction with a smaller monomerPd ratio in order to characterize the structure of the polymer end. The 13C{'H} NMR spectrum of the polymer, prepared from the reaction of 2-phenyl-lmethylenecyclopropane promoted by 1, contains small signals at 6 138.4, 114.4, and 34.1. They are assigned to =CH2, =CH-, and CH2 carbons of the initiation end group, CH2=CHCHz-, of the polymer. Figure 2 depicts the 13C{'H) NMR spectrum of I obtained from the polymerization catalyzed by 2. Small signals at 6 141.7, 114.3, and 47.8 are attributed to =CH2, =CH2-, and CH carbons of the CH2=CH-CH(Ph)- end group based on comparison with peak positions of the model compounds. These results indicate that the polymer contains the n-ally1 ligand of the catalyst as the end group and which is incorporated at the initiation step of the polymerization. It dso suggests that insertion of the monomer into the 1-phenylallyl ligand of 2 forms the bond between the olefinic carbon of the monomer and the benzylic carbon of the ligand. model compounds

6 11

6139

1 140 120 100 70 50

160

30

10

Figure2 13C{1H) NMR of I (a low molecular weight) obtained from the polymerization of 2-phenyl-1-methylenecyclopropanecatalyzed by complex

Pd complex-catalyzed ring-opening polymerisation

309

Table 1 Polymerization of 2-phenyl-1-methylenecyclopropane promoted by Pd complexa

run complex additiveb solvent time(h) conv(%) M: 1 2 3 4 5 6 7 8e

9 10 11 12 13 14 15 16 17

1 2 3a 3a 3a 3a 3a 3a 3a 3a 3b 3c 3d 4a 4a 4b 4c

MeCN MeCN MeCN MeCN THF NMP MeCN MeCN MeCN MeCN MeCN MeCN MeCN MeCN MeCN MeCN MeCN

12

10 24 6 60 10 5 15 4 5

24

24 24 42

6 5 20

92 95 93 95 97 88

94 89 95 75 86 >85

95 84 94 91 99

Mw/Mc 1O3kbsd(s-')

4000 5900 5700 4900 1300 3900 6500 1 1000 4200

3600 5400 4300 5000 4300 5700 5100 6300

1.37 1.46 1.44 1.33 1.72

0.59 0.39 1.28 0.12

1.42

1.39 1.59 1.70 1.23 1.49 1.54 1.53 1.44 1.46 1.46 1.70

1.53

1.93 0.79 0.41 0.34 0.34 1.21 1.31 0.59

Reaction conditions: [Pd] = 25 mM, [monomer] = 1.8M at 80 "C unless stated otherwise. [additive]/[Pd] = 3.0. GPC in THF with polystyrene standards. Estimated by change of amount of the consumed monomer by NMR. [monomer] = 5.0M. a

Table 1 summarizes results of the polymerization under various conditions. Neutral palladium complexes with diimine ligands, 3a-3d and 4a, also bring about the polymerization of 2-phenyl-1-methylenecyclopropane in MeCN to produce polymer I in high yields (run 3, 11-14in Table 1). These polymerization reactions obey first-order kinetics to concentration of the monomer (kobsd = 0.34-0.41x s-l with [Pd] = 25 mM). The structure of the polymers produced does not depend on the catalyst used. The molecular weights Mn of I obtained from these reactions range from 4300 to 5700 with MJMn ratio 1.44-1.54 (GPC, polystyrene standards). Addition of AgBF4 to the reaction s-l mixture using complex 3a increases the polymerization rate (run 4:kobsd = 1.28 x with [Pd] = 25 mM). Addition of AgPF6, AgBARF (BARF = tetrakis-(3,5bis(trifluoromethyl)phenyl)borate), and NaBF4 to the reaction catalyzed by 3a also enhances the polymerization of 2-phenyl-a-methylenecyclopropane. Since these salts convert the chloropalladium complex into the cationic palladium complexes having PFa, BARF, or BF4- as the counter anion, the polymerizations promoted by these cationic Pd complexes proceed more smoothly than those by the neutral complex 3a. The solvents MeCN and NMP are suited for the polymerization catalyzed by the cationic Pd complexes (run 4 and 6), whereas the reactions in THF and PhCN give the polymer in low molecular weights (1300-1700). The polymerization promoted by the catalyst prepared from the reaction of AgPF6 with 3a produces I with Mn = 11000,which is the highest among the experimental results conducted (run 8). Figure 3 shows change of GPC traces of the polymer in the polymerization promoted by the catalyst prepared fiom 4a and AgPF6 in

310

Perspectives in Organometallic Chemistry

MeCN, suggesting increase of the molecular weight by conversion of the monomer. Complex 3a also catalyzes the polymerization of 2-aryl- 1-phenylmethylenecyclopropanes under similar conditions (eq 1). Table 2 summarizes the NMR data of polymers I-V.

M,IM,

= 1.39

I

26

I

2.8 10" M,

I

0.3

Figure 3 GPC profiles of the reaction mixture of the polymerization of 2-phenyl-lmethylenecyclopropane by 4a/AgPFs (1/3) catalyst ([monomer]/[4a] = 70) in MeCN at 80 "C. Monomer conversion: (a) 28% and (b) 94%. Table 2 NMR Data of the Polymersa Polymer

CH

=CH2

=C

Aryl

I

'H 2.0-2.5(br) 2.8-3.4(m) 4.4-4.9(m) 13C 39.2 48.8,49.8 110.6 142.6 126.1, 39.6 50.0, 50.8 111.2 149.6

7.4 (s), 6.97(s) 127.9 148.5, 148.9, 143.1 128.3 149.4,

I1

'H 1.9-2.4(br) 2.6-3.4(m) 4.4-4.9(m) 13C 39.2 48.1,48.9 110.2 139.7 135.5,39.5 49.3, 50.1 110.7 149.6

6.8 1,7.00,2.13(Me) 127.9 148.9, 149.3, 140.2 128.3 149.4,

'H

6.9 114.6, 114.9, 147.8, 148.4 129.4, 129.5, 149.0, 149.1 149.2, 159.7, 162.9

I11

13C

IV

V

a

CH;!

1.8-2.4(br) 2.6-3.2(m) 4.3-4.9(m) 39.0 48.1,48.3 110.0 137.8 39.5 48.7,49.0 111.4 138.0 49.4, 50.2 138.3

'H 1.6-2.4(br) 2.5-3.2(m) 4.2-4.9(m) 13C 39.0 48.2,48.5 111.3 140.5 49.0,49.2 111.7 140.6 39.3 49.7,50.4 141.0

6.8,7.1 128.2 147.5, 129.4 147.9 148.6

'H 1.9-2.5(br) 2.6-3.4(m) 4.3-4.9(m) 13C 39.2 48.1,48.9 110.1 134.7 39.5 49.1, 50.0 110.8 135.1

6.5-7.2,3.77(OMe) 113.2, 148.9, 149.4, 157.7, 129.6, 149.6, 55.1(s, OMe)

'H (400 MHz) and 13C{'H) (125 MHz) in CDCl3.

132.0,

Pd complex-catalyzed ring-opening polymerisation

31 1

Glass-transition (Tg)and melting temperatures of olymer I were estimated by DSC in nitrogen atmosphere at a heating rate of 10 "C-min-. Polymer I demonstrates a glasstransition at 62 "C. The polymer of 2-phenyl-1-methylenecyelopropane with another structure, (CH2-CH2-C(CH2)2)n was recently obtained by our group fiom the polymerization promoted by a Ni complex.lo This polymer containing a three-membered ring in every structural unit exhibits a much higher glass transition temperature than I due to the rigid conformation of the main chain of the polymer. Thermogravimetric analysis (TG) revealed high stability of I up to 250 "C under nitrogen. The temperature of 5% weight loss of I (obtained in run 1 of Table 1) is 287 "C.

P

2.2. Kinetic Results of Ring-opening Polymerization of 2-Aryl-1-methylenecyclopropanes As shown above, the polymerization obeys first-order kinetics with respect to the concentration of the monomer. Figure 4 depicts the results of the kinetic measurement of the polymerization promoted by 3d at 80 - 110 "C in NMP. The first-order plots of the reaction in this temperature range as well as Arrhenius plots of the rate constants show good linearity. The activation energy of the reaction, Ep,is determined as 67.6 kl mol-', which is smaller than many of the reactions promoted by transition metal complexes. The polymerization accompanies ring opening of the monomer molecules and is highly exothermic, due to release of the large ring strain energy during the polymerization. These reaction features probably render the transition state of the polymer growth product-like and decreases activation energy of the total polymerization reaction. Figure 5(a) shows time-yield curves of the polymerization of 2-aryl-1methylenecyclopropanes catalyzed by the neutral Pd complex 3d in MeCN. The polymers obtained in the reactions show molecular weights of Mn 5600-7200 with Mw/M,, = 1.341.56. The polymerization is completed within 30 h in four of the five monomers

0.0r.

0

*

5

-

'

-

10 tl h

*

15

-

*

20

-

' 25

-c "0

5

10

15

20

25

tlh

Figure 4 (a) Time-conversion plots and (3) Jirst-order plots of the polymerization of 2phenyl-I-methylenecyclopropne by 3d in NMP ([3d] = 25 mM, [monomer]/[3d] = 70). Arrhenius plot of the reaction is in inset of (b)

312

Perspectives in Organometallic Chemistry

n

r

3

:-o.6 0 -1.0 w

-

-1.4

tl h

\;

Figure 5 Polymerization of 2-aryl-1-methylenecyclopropanescatalyzed by 3d at 80 "C in MeCN ([3d] = 25 mM, [monomer]/[3d] = 70). (a) Time-conversion curve and @) Hammett plots of the polymerization examined, while the polymerization of 2-(4-chlorophenyl)-1-methylenecyclopropane requires a much longer time for completion. Thus, the polymer growth is not retarded by the presence of such functional groups as F and OMe. From the time-conversion curve of the polymerization of as well as Hammett plots of the rate constants of the reactions, a relationship is observed between the substituent of the phenyl ring of the monomer and the polymerization rate. The OMe group-substituted monomer polymerizes more rapidly than the other monomers, while the monomer with C1 substituent undergoes slower polymerization and requires longer polymerization time than the other monomers. Random copolymerization of these monomers promoted by the Pd complex was conducted in order to reveal more details of relative reactivity of the monomers for the polymerization. The five monomers shown above undergo smooth copolymerization to produce the random copolymers as shown in eq. 2.

hi$&-

q$ \

n

(2)

X

Table 3 summarizes the random-copolymerization of several series of the monomers. The reaction was quenched before conversion of the monomers exceeded 36% so as to estimate the relative reactivity of the monomers correctly. 2-(4-Methoxyphenyl)-1methylenecyclopropane and 2-(4-methylphenyl)-1-methylenecyclo-propane react with the monomers whose polymerization rates in Figure 5 are smaller than those. The composition of the random copolymers indicated that these two monomers are incorporated more smoothly than the other monomers. The ratios of the monomer structures of the copolymers show that the ratios of the monomer unit with OMe and Me substituents to the other monomers are higher than those expected from the respective homopolymerization rates.

313

Pd complex-catalyzed ring-opening polymerisation

Table 3 Random-copolymerizationof 2-aryl-1-methylenecyclopropanes" R U ~monomer 1 X

1 2 3 4 5 6

OMe OMe OMe OMe Me Me

monomer2 Y

conv(%)

Mn

H Me F

33 30 36 25 22 22

5600 6100 6400 5000 5000 5600

c1 H

c1

MJMn

1.45 1.46 1.47 1.41 1.35 1.35

Polymer composition monomer1 : monomer2 93: 7 83:17 90:lO 100: 0 74:26 73:27

Polymerization was carried out for 3 h (or less) at 80 "C in MeCN. [3d] = 25 mM, [monomerl] = [monomer21 = 1.8 M. Substituents of the aryl group of the monomer, X and Y, are shown. Ratio of the monomers incorporated into the copolymer determined by peak area ratios of the 'H NMR spectra. a

The substituents of the aryl group of the monomers influence both reactivity and stability of the growing polymer end and reactivity of the monomer for the insertion because both the polymer end and monomer have the aryl groups. Generally, the electron donating substituents of the substrates of the reaction enhances the polymer growth more significantly than the electron withdrawing substituents. The dual effects of the substituents to activate both the polymer end group and the monomer were observed in the above study. 2.3 Polymerization Mechanism The polymerization of 2-phenyl- 1-methylene-3['3C]-cyclopropane produces polymer I-13C whose I3C 'H NMR spectrum shows signalscat6 = 110.4 and 110.9. These signals arise from the C-enriched carbon atom and correspond to the =CH2 carbon of the polymer chain. This result indicates selective cleavage of the distal C-C bond during the polymerization as shown in eq. 3.

5 )

Based on the results of the end-group analyses of the polymer, the polymerization is considered to be initiated by insertion of the methylenecylcopropane into the Pd-x-ally1 bond accompanied by its ring opening. The existence of the CH*=CH-CH(Ph)- terminal group of I, formed from the polymerization promoted by 5, and the well-regulated head-totail sequence of the repeating unit indicate that insertion of the monomer occurs preferentially into the Pd-CHPh bond of the q3-1-phenyl-allylpalladium species both in the

Perspectives in Organometallic Chemistry

314

initiation step and in the propagation step of the polymerization. Formation of n-ally1 Pd species by the insertion of methylenecyclopropanes into H-Pd bonds has been proposed to account for results of Pd complex catalyzed coupling reaction of methylenecyclopropanes with malonates or The reaction of 1,3-diene with n-ally1 Pd complexes was also reported to cause C-C bond formation between a vinyl carbon of the diene and a sterically hindered n-ally1 carbon of the complex.12 Their results are rationalized by taking 2,linsertion of a double bond of butadiene into the more sterically hindered Pd-C bond of n-ally1 palladium complex into account. Scheme 1(a) illustrates a plausible mechanism of the polymerization initiated by complex 2.

(ii)

Ar > -

-

-+

CPdj+J

'

"R.;"4

CP*

\Ar Ar

-+

Scheme 1 Proposed polymerization mechanisms for (0 ring-opening polymerization of 2aryl-1 -methylenecyclopropanes catalyzed by Pd complexes and (ii) that of methylenecyclopropane catalyzed by Zr and Lu complexes The 2,l-insertion of the monomer into the Pd-CHPh bond of 1 leads to the formation of a cyclopropylpalladium intermediate which undergoes rapid p-alkyl activation, which regenerates a n-allylpalldium complex. Repetition of the above procedure accounts for the smooth polymerization that gives the product in a regulated head-to-tail sequence of monomer units. This polymerization mechanism contrasts with that proposed for the metallocene-catalyzed polymerization of methylenecyclopropane [Scheme 1(ii)]: in which polymer growth takes place by repetition of the 1,2-insertion of the monomer into the metal-alkyl bond and subsequent p-alkyl elimination. Thus, these two ring-opening polymerization reactions take place via different insertion mode of the monomers, while their products have a common structure with the vinylidene group. 3 CONCLUSION We have demonstrated the Pd complex-catalyzed ring-opening polymerization of 2-aryl-1methylenecyclopropanes to give structurally regulated polymers. The overall activation energy of the ring-opening polymerization was estimated to be 67.6 kJ-mol-'. The polymerization was found to be tolerant with such functional groups as F, C1, and OMe.

Pd complex-catalyzed ring -opening polymerisation

315

The mechanistic study has elucidated the polymer growth which involves new C-C bond formation between the =CH2 carbon of the monomer and benzylic carbon of the n-ally1 ligand of the growing polymer and selective C-C bond activation of the cyclopropane ring. 4. EXPERIMENTAL SECTION 4.1 Preparation of Monomers. 2-Phenyl- 1-methylenecyclopropane was synthesized according to a reported procedure. To a 500-mL round bottomed flask attached to a three-way stopcock, containing a toluene solution (180 mL) of sodium bis(trimethylsily1)aide (300 mmol, 55 g), anhydrous styrene (150 mmol, 17.2 mL) and a magnetic stirring bar under N2 was added dropwise 1,ldibromoethane (135 mmol, 12.1 g) at 0 "C, and the mixture was stirred at 25 "C. After 24h stirring, the reaction mixture was treated with NH4Cl aq., washed with water, and extracted with ether. After drying over MgS04, the volatile fraction was evaporated and the residue was distilled to afford synlanti- 1-bromo- 1-methyl-2-phenylcyclopropanein 81.7% yield (21.9 g). The latter was converted into 2-phenyl-1-methylenecyclopropane as described below. To a 200-mL round bottomed flask attached to a three-way stopcock, containing a dimethylsulfoxide solution (80 mL) of potassium t-butoxide (63.0 -01, 7.07 g), and a magnetic stirring bar under N2 was added dropwise synlanti- 1-bromo-1 -methyl2phenylcyclopropane (58.0 mmol, 12.2 g) at 0 "C, and the mixture was stirred at 25 "C. After stirring for 24 h, the reaction mixture was treated with N b C 1 aq., washed with water, and extracted with ether. After dried over MgS04, the volatile fraction was evaporated and the residue was distilled to afford 2-phenyl-1-methylenecyclopropane in 90% yield (6.82 8). 'H NMR (CDC13): 6 1.32 (lH, m, CH3), 1.80 (lH, m, CH2), 2.69 (lH, m, CH), 5.65 (2H, m, =CH2), and 7.2-7.4 (5H, m, Ph). Other 2-aryl-1-methylenecyclopropanes were prepared analogously. 2-(4-methylphenyl)-1-methylenecyclopropane:83 % yield, 1H NMR (CDC13): 6 1.28 (lH, m, CHz), 1.81 (lH, m, CH2), 2.40 (3H, s, 4-CH3), 2.67 (lH, m, CH), 5.67 (2H, m, =CH2), and 7.1 1-7.21 (5H, dd, 0-,m-Ph). 2-(4-fluorophenyl)- 1-methylenecyclopropane: 77 % yield, 'H NMR (CDC13): 6 1.17 (lH, m, CH2), 1.75 (lH, m, CH9, 2.60 (lH, m, CH), 5.61 (2H, m, =CH2), and 6.907.19 (5H, m, 0-,rn-Ph). 2-(4-chlorophenyl)-1-methylenecyclopropane: 63 % yield, 'H NMR (CDCl3): 6 1.19 (lH, m, CH2), 1.77 (lH, m, CHz), 2.58 (lH, m, CH), 5.60 (2H, m, =CH2), and 7.137.32 (5H, m, 0-,rn-Ph). 2-(4-methoxyphenyl)-1-methylenecyclopropane: 63 % yield, 'H NMR (CDC13): 6 1.20 (lH, m, CHz), 1.74 (lH, m, CHz), 2.62 (lH, m, CH), 3.82 (3H, s, 4-OCH3), 5.64 (2H, m, =CH2), and 6.84-7.22 (5H, dd, 0-, rn-Ph). 4.2 General procedure for polymerization. Typically, to a 25-mL round-bottom flask attached to a three-way stopcock, containing an acetonitrile solution (1 mL) of Pd catalyst (0.025 mmol), co-catalyst (0.075 mmol), and a magnetic stirring bar under Ar, was added 2-phenyl-1-methylenecyclopropane (Ia, 1.75 mmol, 227.5 mg) by a syringe, and the mixture was stirred at 80 "C. An'aliquot of the polymerization mixture was periodically taken out from the flask, and subjected to 'H NMR spectroscopy (CDC13) and GPC to determine monomer conversion and average

316

Perspectives in Organometallic Chemistry

molecular weights (M,,, M,) of the produced polymer, respectively. Monomer conversion was calculated fiom the relative intensity of the signals at 65.55-5.58 and 64.40-4.80 due to C=CH;! in Ia and IIa, respectively. For isolation of the polymer, the polymerization mixture was poured into a large amount of methanol, and the white precipitates formed were collected and dried in vacuo at 25 "C. References

Recent review of coordination polymerization by transition metal complexes. (a) H. G Altand A. Koppl, Chem. Rev. 2000, 100, 1205. (b) H. H. Brintzinger, D. Fischer, R. Miilhaupt, B, Rieger and R. M. Waymouth, Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (c) S, D, Ittel, L. K. Johnson and M. Brookhart, Chem. Rev. 2000, 100, 1 169. (d) G. J. P. Britovsek, V. C. Gibson and D. F. Wass, Angew. Chem., Int. Ed 1999,38, 428. (e) S. Mecking, Coord. Chem. Rev. 2000,203,325. Reviews and leading references. (a) A. Brandi and A. Goti, Chem. Rev., 1998, 98, 589. (b) P. Binger, and H. M. Busch, Top. Cur: Chem., 1987, 135, 77. (c) T. Ohta and H. Takaya, In Comprehensive Organic Synthesis; B. M. Trot Ed., Pergamon: Oxford, 1991; Vol. 5, p. 1183. (d) M. Lautens, C. Meyer, and A. Lorentz, J. Am. Chem. SOC. 1996,118,10676. (e) N. Tsukada, A. Shibuya, I. Nakamura and Y. Yamamoto, J. Am. Chem. SOC. 1997, 119, 8123. ( f ) D. H. Camacho, I. Nakamura, H. Itagaki, and Y. Yamamoto, Angew. Chem., Int. Ed. 1998,38, 3365. (g) I. Nakamura, S. Saito, and Y. Yamamoto, J. Org. Chem. 1998,63,6458. (h) M. Suginome, T. Matuda, and Y. Ito, J. Am. Chem. SOC.2000, 122, 11015. (i) A. G. Bessemertnikh, K. A. Blinov, K. Yu, N. A. Donskaya, E. V. Tveritinova, N. M. Yur'eva, and I. P. Beletskaya, J. Org. Chem. 1997,62,6069. (i) T, Ishiyama, S . Momota, and N. Miyaura, Synlett 1999, 1790. Examples of radical and cationic ring-opening isomerization polymerization: (a) L. A. Errede, J. Polym. Sci., 1961, 49, 253. (b) T. Takahashi, J. Polym. Sci., Part A, 1968, 6,403. (c) W. J. Roberts and A. R. Day, J. Am. Chem. SOC.,1950,72,1226. Anionic ring-opening polymerization of cyclopropanes having electron withdrawing group: J. Penelle, and T. Xie, Macromolecules, 2000,33,4667 Coordination polymerization of small-ring molecules without ring opening: (a) S. Rush, A. Reinmuth, W. Risse, J. O'Brien, D. R. Ferro and I. Tritto, J. Am. Chem. SOC. 1996, 118, 12230. (b) C. Mehler and W. Risse, Macromolecules 1992,25,4226. (c) A. L. Sdir and B. M. Novak, Macromolecules 1995,28, 5396. (d) A. Sen, T.-W. Lai, and R. R. Thomas, J. Organomet. Chem. 1988,358,567. Ring-opening polymerization of oxygen containing cyclic monomers catalyzed by Pd complexes: N. K. Lim and B. A. Arndtsen, Macromolecules 2000,33,2305. (a) Y. Yang, A. M. Seyam, P. F. Fu and T. J. Marks, Macromolecules 1994,27,4625. (b) L. Jia, X. Yang, S. Yang and T. J. Marks, J. Am. Chem. SOC.1996,118, 1547. (c) L. Jia, X. Yang, A. M. Seyam, I. D. L. Albert, P. F. Fu, S. Yang, amd T. J. Marks, J. Am. Chem. SOC.1996,118,7900. D. Takeuchi, S. Kim and K. Osakada, Angew. Chem., Int. Ed. Eng. 2001,40,2685. (a) M. Brookhart and M. I. Wagner, J. Am. Chem. SOC.1996,118,7219. (b) A. Abey, A, Gsponer, and G. Consiglio, J. Am. Chem. SOC.1998,120, 11000. 10 D. Takeuchi and K. Osakada, Chem. Commun. 2002,646. 1 1 R. P. Hughes and J. Powell, J. Am. Chem. SOC.1972,94,7723.

Subject Index

acetonitrile, reactions of titanium imido complexes with, 39-40 activation of C-F bonds, mechanism of, 139 by nickel(0) complexes, 136ff in the presence of C-Cl bonds, 139-140 in the presence of C-H bonds, 140-142 addition reactions, of acids to alkynes, sulfido cluster complex catalysed, 70 of alcohols to alkynes, sulfido cluster complex catalysed, 66,67, 69 of dithio-o-carboranylcobalt complex, 59 of P-H and P-P bonds to triple and double bonds, Ni and Pd catalysed, 244-247 agostic interactions, 6 alkenes, binding of, to palladium cluster, 69 polymerisation of, 164,200 reactions with iridium hydride complexes, 298-301 silyl group transfer to, 256 alkynes, preparation of, 103 coupling reactions of, 120ff hydrogenation of, iridium complex catalysis of, 298-301 optically active, synthesis of, 293 reactions with iron pentacarbonyl, 122 ruthenium complex catalysed Qmerisation of, 6 2 4 6 allenes, reactions of titanium imido complexes with, 41 allenylidene complexes of ruthenium, 6546,285ff amines, chiral, preparation of, 271

primary, deprotonation of, 2 anion recognition by platinum complexes, 214 antimony, imido cages of, Iff phosphido cages, Iff mixed imido cage with lead, 3 aqueous organomolybdenum chemistry, 168ff Arbuzov reaction, nickel complex assisted reaction of aryl and vinyl halides in, 240 arsenic, imido cages, Iff phosphido cages, Iff bismuth, imido cages, Iff phosphido cages, Iff cages, group 15 imido, Iff group 15 phosphido, Iff alkali metal, with group 15 imido and phosphido anions, Iff calculations, ab initio, of gallium clusters, 20 8-caprolactone, polymerisation of, 164, 165 carbodiimides, reactions of titanium imido complexes with, 34-35 carbon dioxide, reactions of titanium imido complexes with, 34-39 carbon disulfide, reactions of titanium imido complexes with, 36 carbon-fluorine bond activation, 136ff carbonyl, equivalents, gallium and indium alkyls as, 25 carbonyl sulfide, reactions of titanium imido complexes with, 36 carboranes, dichalcogenolate,complexes with, 47ff catalysis, of alkene hydrogenation, 298-30 1

318

of alkyne hydrogenation, 301-303 of allylic substitutions, 276ff of asymmetric hydrogenation, 265-267,270-272 of asymmetric conjugate addition of arylboronic acids, 267-269 of coupling reactions of vinylsilanes, 256-260 of ethylene polymerisation by fluorenyl lanthanide complexes, 164 of hydrosilylation of alkenes and ally1 ethers and esters, 262-263 of polymerisation of E-caprolactone by fluorenyl lanthanide complexes, 164-165 cerium, amido complexes, 92-93 cyclopentadienyl complex, 87 dithiocarbamates, 93-94 oxidation of complexes of, 92 chelates, phosphine-olefin, 222ff chiral allenylidene ruthenium complexes, 285ff chiral alkynyl ruthenium complexes, 290-294 clusters, gallium, 16ff indium, 16ff iridium-palladium, 66-67 metal carbonyl, electronic structure of, 184-186 modulation of electronic properties of, 186-1 89 and applications to nanotechnology, 192 cobalt, dichalcogenolate carborane complexes, 55,59 siloxide complexes, dimeric, 254, 255 siloxide complexes, monomeric, 26 1-262 sulfido cluster complexes, 7 1 conformation of Pd complexes, 278 copper, complexes with tungsten dichalcogenolate carboranes, 49-5 1

Perspectives in Organometallic Chemistry

group 15 imido cage complexes, 5-6 coupling reactions, of alkenyl bromides with monoazoles and phenothiazine, 249 of aryl halides with benzotriazole and tetrazoles, 248 of fluorinated heterocycles with vinyl tributylstannane, nickel catalysed, 146 of vinylsilanes, Ir and Rh catalysed, 256-260 cyclisation of a,wdiynes, ruthenium complex catalysed, 63 cycloaddition reactions, of dichalcogenolate carborane complexes, 59-60 of alkynes, ruthenium complex catalysis of, 11Iff of titanium imido complexes, 34ff cyclopropanes,ring-opening polymerisation of, 306ff cyclovoltammetry, of gallium clusters, 16ff of indium clusters, 16ff of lanthanide cyclopentadienyls, 86 of metal carbonyl clusters, 189, 190 dendrimers, ruthenium alkynyl complexes as, 104-105 desilylation, of substituted cyclopentadienyl, 9 1 DF" calculations, for activation of C-F bonds, 140, 141 in Pd complex catalysed allylic substitution, 278-279 of Ru complex mediated oxidative coupling of acetylene, 112-1 17 digoldalkanes, 75,77 digoldolefins, 75 digoldarenes, 76-8 1 P-diketiminato complexes, 94-98 dimerisation of terminal alkynes, 64

Subject Index

diphosphonites, chiral, in asymmetric catalysis, 266-269 DNA, interaction with platinum complexes, 2 15 dysprosium, dichalcogenolate carborane complexes of, 56 electroabsorption spectroscopy, 107 electrochemistry, of rhodium and iridium complexes, 228,230-231 of molybdenum cyclopentadienyl complexes, 173-177 of metal carbonyl clusters, 187189 of ruthenium alkynyl complexes, 109 electron transfer in cluster compounds, 19ff electronic dissymmetry of bidentate ligands for asymmetric catalysis, 279-28 1 EPR of rhodium and iridium complexes, 226228,234-236 erbium, dichalcogenolatecarborane complexes of, 56 ferracycles, formation and reactivity of, 122ff fluoride complexes, fluoride abstraction from, of nickel, 143 substitution of fluoride in, of nickel, 143-144 fluoropyridines,oxidative addition to nickel of, 138-139 fluxionality, in platinum complexes, 213 of ytterbium cyclopentadienyl complex, 90 gadolinium, dichalcogenolatecarborane complexes of, 56 gold organyls, 74ff hafnium, dichalcogenolate carborane complexes, 48

319

halopyrimidines, nickel mediated synthesis of, 145 heterocycles, fluorinated, nickelmediated synthesis of, 144-146 hexafluorobenzene, oxidative addition to nickel of, 137-138 homoleptic transition metal derivatives, 25 hydride complexes of iridium, 229,258, 259,297ff hydrogen fluoride, coordination to fluoride complexes of, 142 hydrosilylation, Rh and Ir catalysed 262, 263 hyperpolarisability, 101 indium subhalides, 21ff indium clusters, halogenation of, 21-23 ion pairs, transition metal complex, 197203 iridium, alkene hydrogenation catalysed by complexes of, 298ff dichalcogenolate carborane complexes, 54-55,5940 olefin-phosphine chelate complexes, 223ff paramagnetic 17-electron complexes, EPR of, 226-229, 234-236 pentacoordinated complexes, formation of, 224-226 reaction with hydrogen and protic reagents with complexes of, 229-230 reduction of 16-electron complexes of, 226 siloxide complexes, dimeric, 254ff siloxide complexes, monomeric, 260-262 iron, alkyne complex, 122, 125, 130 complex with arsenic phosphido anion, 12 dichalcogenolatecarborane complexes 53 gallium cluster with, 24 indium cluster with, 24 metallacycles, 122ff

320

sulfido cluster complexes, 71 isocyanates, reactions of titanium imido complexes with, 3 4 4 0 isocyanides, reactions of titanium imido complexes with, 4 1 4 3 isothiocyanates, reactions of titanium imido complexes with, 36 lanthanides, dichalcogenolate carborane complexes, 56 cyclopentadienyls, 86-88 lanthanum cyclopentadienyls, 86-87, 89 lithium dichalcogenolate carboranes, 47, 48 manganese cyclopentadienyl cage complex, 5 manganocene, reaction with group 15 imido anions, 5 mass spectrometry of molybdenum cyclopentadienyl complexes, 173177 mechanism of catalybc alkene hydrogenation, 297-300 metallacyclopentatriene complexes, 112ff metallation, mixed, Iff molybdenum, asymmetric catalysis of allylic substitution, 28 1 complex with arsenic phosphido anion, 12 complexes with rhenium dichalcogenolate carboranes, 57-58 dichalcogenolate carborane complexes, 49-5 1 high oxidation state cyclopentadienyl complexes, conductivity of, in aqueous solution, 172 nature of, in aqueous solution, 170 synthesis of, 168-169 sulfido cluster complexes, 68-70 monophosphites, chiral, in asymmetric catalysis. 269-272

Perspectives in OrganometallicChemistry

monophosphonites, chiral, in asymmetric catalysis, 269-272 neodymium, dichalcogenolate carborane complexes of, 56 nickel, addition of P-H and P-P to triple and double bonds catalysed by complexes of, 245-247 Arbuzov reaction of aryl and vinyl halides catalysed by complexes of, 240 carbonyl clusters, 186-1 88 complexes with rhenium dichalcogenolate carboranes, 57-58 fluoride complexes, 136ff formation of vinyl phosphonates catalysed by complexes of, 242 homoleptic gallium and indium derivatives, 25 sulfido cluster complexes, 7 1

NMR, and catalysis, 200 of ferracycles, 125-127, 130-131 of fluorenyl and indenyl lanthanides, 158,161-162 HOESY, 199-20 1,204-205 and intermolecular interactions in organometallic compounds, 204-205 and dissociation of iridium complexes, 226 of organogallium clusters, 13-C, 18 pulsed-field gradient spin echo, 203 of platinum complexes, mechanistic studies by, 209ff 103-Rh, of Rh complexes, 224 of transition metal complex ion pairs, 197ff nonlinear optical properties, of ruthenium complexes, lOOff dimensional evolution of, 105-108 switching of, 108-109 nucleophilic substitution in perfluorovinyl Iigand, 148

321

Subject Index

octafluoronaphth alene, oxidative addition to nickel of, 137-138 optical transparency of ruthenium alkynyl complexes, 105 osmium, dichalcogenolate carborane complexes of, 52,59 oxidative addition, of fluorinated heteroaromatics to nickel, 136-139 palladium, carbonyl clusters, 188 complexes for asymmetric catalysis, conformation of, 278 sulfido cluster complexes, 68-7 1 palladium complex catalysis, of addition of P-H and P-P to triple and double bonds, 244247 of allylic substitution, 276ff of arylation of benzotriazole and tetrazoles, 248-249 of arylation of polyamines, 24925 1 of cyclopropane ring-opening polymerisation, 306ff of formation of aryl and vinyl phosphonates, 241-242 of formation of aryl and vinyl phosphines, 243-244 of vinylation of monoazoles and phenothiazine, 248-249 phospha-alkyne, tert-butyl, reactions of titanium imido complexes with, 39 40 phosphine, tropylidenyl, as chelating ligand, 223ff phosphonates, aryl, formation of, 240-242 vinyl, formation of, 242 platinum, carbonyl clusters, 187-188 cyclometallated complexes, 2 17, 218 dimethylsulfoxide complexes, solution behaviour of, 209ff fluxionality in complexes of, 213 homoleptic indium derivative, 25 interaction of complexes of, with nucleosides and DNA, 2 15

porphyrin complexes of, 216 thioether complexes of, 218-219 plumbocene, reaction with antimony imido anion, 4 polymerisation, of alkenes, 164,200 of ecaprolactone, 164-165 of cyclopropanes, ring-opening, 306ff porphyrins, platinum complexes containing, 216-217 potassium, salts with arene anions, 89, 90 pyridylamides, asymmetric catalysis with, 28 1-283 rearrangement, of antimony imido cage, 7 of bis(fluoreny1)lanthanide complexes, 152ff reduction, of organomolybdenum compounds, 177-179 restricted rotation of arsenic phosphide heterocycle, 12 rhenium, C-F and C-H bond activation by complexes of, 140-141 dichalcogenolate carborane complexes, 52,57-58 sulfido cluster complexes, 68-69 rhodium, C-F and C-H bond activation by complexes of, 141-142 carbonyl clusters, 186-1 87 complexes for asymmetric catalysis, 265-272 dichalcogenolate carborane complexes, 5 5 , 5 9 4 0 olefin-phosphine chelate complexes, 223ff paramagnetic 17-electron complexes, EPR of, 226-229, 234-236 pentacoordinated complexes, formation of, 224-226 reduction of hexafluoropropene using complex of, 149 reduction of 16-electron complexes of, 226

322

siloxide complexes, dimeric, 254ff siloxide complexes, monomeric, 260-262 sulfido cluster complexes, 67-68 ruthenium, allenylidene complexes, 285ff carbene complexes, 1 12 C-F and C-H bond activation by complexes of, 142 dichalcogenolate carborane complexes, 52-54,59 half-sandwich complexes, 11 Iff indenyl complexes, 285ff octopolar and dendritic alkynyl complexes, 102-105 sulfido cluster complexes, 67-68, 71 thiolato-bridged complexes, catalysis with, 62-66 vinylidene complexes, 108 samarium, cyclopentadieny 1 complex, 90 P-diketiminates 94-96 fluorenyl complexes, synthesis and rearrangement of, 157-161 scandium, cyclopentadienyl complex, 88 silacycle, formation of, 12 siloxide ligands, bonding to transition metals, 253 silver complexes with tungsten dichalcogenolatecarboranes, 49-5 1 spectroelectrochernistry of ruthenium alkynyl complexes, 109 steroids, ruthenium substituted derivatives of, 292 substitution reactions, allylic, of carbonates and acetates, 276,281 in platinum complexes, 210ff propargylic, ruthenium complex catalysed, 64-65,287

Perspectives in Organometallic Chemistry

tin, phosphido cages, 10-1 1 titanium, dichalcogenolate carborane complexes of, 48 titanium imido complexes, cyclopentadienyl-amidinate supported, 36 synthesis of, 31ff with polydentate ligands, 32ff sulfido cluster complex, 71 triphosphine, terminally substituted, 13 tungsten, complex with arsenic phosphido anion, 12 dichalcogenolate carborane complexes, 49-50 sulfido cluster complexes, 68 triphosphine complex, 13 vinylidene complex of ruthenium, 63-66 vinylsilanes, silyl group transfer with, 254ff Wiberg indices, 116 ytterbium, cyclopentadienyl complex, 88,90, 91 dichalcogenolatecarborane complexes, 56 P-diketiminates 94-98 fluorenyl complexes, synthesis and rearrangement of, 161-164 Zintl compounds, 7ff zirconium, dichalcogenolatecarborane complexes of, 48 reduction of hexafluoropropene using complex of, 149