Pure & Appl. Czem., Vol. 49, pp. .839 — 855. Pergamon Press, 1977. Printed in Great Britain.

AROMATIC TRANSITION. METAL COMPLEXES THE FIRST 25 YEARS

Peter L. Pauson

Department of Pure and Applied Chemistry,

UNIVERSITY OF STRATHCLYDE,, Glasgow, Gl 1XL Scotland.

ABSTRACT

The progress of the chemistry of aromatic transition metalcomplexes is traced from the discovery of ferrocene and otherearly work on z—complexes. The gradual elaboration ofmetallocenes from most transition'metals, then of mixedcyclopentadienylmetal carbonyls and analogues with other oradditional ligands was followed by the recognition andsynthesis of bis—benzenechromium and in turn a varied rangeof benzenoid complexes. The importance of successivedevelopments in preparative methods in making possible therapid expansion of this field is stressed and it is shownhow in turn the benzenoid complexes were followed bytropylium, cyclobutadiene, cyclooctatetraene and cyclo—propylium complexes, thus completing a range with 3—8 memberedcarbocyclic aromatic systems. Knowledge of heterocyclicanalogues is much more restricted, but rapid expansion in thisand in the chemistry of triple—decker and other multinuclearcomplexes is foreseen. Photochemicál, electrochemical andmetal vapour methods of synthesis are expected to play 'a majorpart in such further developments. Finally a brief account isgiven of the increasing emphasis on understanding the effectsof transition metals on aromatic ligand reactivity, e.g. byfacilitating both electrophilic and nucleophilic reactions.Many new organic synthetic procedures are being evolved whichutilize such effects on reactivity or the stabilization incomplexes of structures which have only transient existencein the metal—free state; a' few selected examples of these newmethods are given.

Non—benzenoid aromatic systems were still a rather new field and one inwhich I had a strong interest, when, in 1950, H. D. Brown predicted thatfulvalene, 1, should belong in this class. I was fortunate therefore, whenearly in 1951, Tom Kealy agreed to work with me and to try, synthetic routesto this molecule. After some unproductive attempts along relatively lengthyroutes I suggested that he try a very simple approach: the reaction ofcyclopentadienylmagnesium bromide with ferric chloride. As an allylic

C5H5MgBr+ FeCL3—.(C5H5) Fe2

839

Grignard reagent C5H5MgBr was expected to undergo coupling when heated withtransitionmetal halides; FeC13, if sqmetimes less effective than CoCl2,had the advantage of being readily available in anhydrous form and mighteven cause immediate oxidation of the expected dihydrofulvalene to thedesired aromatic system. I first saw theresult of this experiment when a :

few drops of the ether solution ran down the outside of the separating funneland left orange crystals. They proved to have the solubility characteristicsand stability of the hoped—for aromatic hydrocarbon — but microanalysis show-ing 6L.6% of carbon could only be accommodated by assuming the formulaC10H10Fe, We could hardly believe that we had an organo—iron compound whenmuch of the product was recovered unchanged from solution in conc. sulphuricacid; but on prolonged boiling with nitric acid or fuming with perchioricacid the presence of iron was duly confirmed both qualitatively and quanti—tatively .

This compound, now known as ferrocene had in fact been made somewhat earlierby Miller, Tebboth and Tremaine2 by the vapour phase reaction of cyclopenta-diene with an iron—oxide catalyst,but their publication of this discoveryappeared after our own brief note.3 You may well ask why these accidentaldiscoveries should have led the organisers of this Conference to ask me tospeak on the 'first 25 years' of aromatic transition metal chemistry. Thisis tantamount to asking why the discovery of ferrocene led to such a suddensurge of interest in organo—transition metal chemistry. I believe theanswer lies in the timeliness which is an essential factor determining theimpact of any chance discovery. It may be helpful therefore to set thescene by summarising the state of the art in 1951.

Ferrocene was far from being the first organo—transition metal compound.The first organometallic compound' was the ethylene—platinum compound knownas Zeise's salt, K[C2H4PtC13] described in 1827 — long before Frankland'salkyl zinc compounds and other 'conventional' organometallics. Alkyls ofplatinum and gold were also well—known, but there had been so many failureswith other transition metals, that these substances were regarded as excep—tions to a genera). rule that 'typical transition metals do not form stableorganometallic compounds'. Indeed the coupling of Grignard reagents whichwe had tried to use had been one of the, byproducts of such unsuccessfulattempts.' The carbonyls were other exceptions, regarded as sui generis, ordespite their metal—carbon bonds not even thought of as organometallics.Reihlen's discovery5 in 1930 of butadienetricarbonyliron, (C4H6)Fe(C0)3, hadbeen largely ignored. But perhaps most sinificant, the one known class ofaromatic transition—metal complexes — Hem s so—called polyphenylchromium

is '[(C6H5) C.r]X' had obviously been discovered twenty years too soon?Their existence wa widely known and too well documented to be readily dis-missed as fictional, but they had been allowed to remain a class of verypuzzling compounds of unknown structure.

Against such a background we could have had no doubt that we had' a most noveland remarkable compound on our hands and one whose true structure might notbe what it appeared at first sight. We did not guess the correct structure— but we did initiate an attempt to have it determined crystallographically.The fact that that determination was delayed and later abandoned when threedifferent groups of workers undertook similar work is irrelevant. The pointI want to make is that fe'rrocene was discovered when X—ray techniques weresufficiently advanced and could give a rapid and firm answer — a situationvery different from that, which faced Hem in 1919.. I have mentioned else—'where that the correct structure of ferrocene. was, first suggested to me byW. E. Doering, to whom I described our compound shortly after we first madeit. It is. well known that the same structure was arrived at independentlyby E. 0. Fischer,7 by R. B. Woodward and by G. Wilkinson8 when they firstread our communication. Each sought and obtained evidence supporting thisstructure. Woodward went on to demonstrate9 the aromatic nature of thebeast and to coin the now familiar name1 ferrocene. Fischer and Wilkinsondirected their attention to ,the analogues of other tran,sition metals andduring the next few years such compounds were described in a rapid successionof communications, frequently simultaneously, from both their laboratories.There is ample evidence in the patent literature that several industriallaboratories in the U.S.A. were independently discovering many of the samecompounds — further attesting to the range of interest aroused.

Aromatic transition metal complexes—The first 25 years. 841

We had been lucky in our choice of the water—soluble ferric chloride. Theother simple halides we might have used were al]. too insoluble to reactunder the same conditions and our own first attempts to make cobalt andnickel analogues were therefore quite unsuccessful. But improved techniqueswere soon forthcoming, by way of more soluble metal derivatives and the morereactive alkali metal cyclopentadienides: Thus Wilkinson10 used the metalacetylacetonates, while Fischer11 employed sodium and potassium cyclopenta—dienide in liquid ammonia and the metal hexammine chlorides, These andsimilar innovations led to the whole series of bis—cyclopentadienyls of firstrow elements from Ti to Ni— compounds of widely varying.stabilities whoseredox and magnetic properties provided much of the initial interest andstimulus to the theoretician, For not only was ferrocene tinly in termsof structure determination, but also and perhaps more so in terms of under—standing. The early papers proposing the .cor'ect structure already con—tamed interpretations in M.O. terms and were soon followed by more refinedtreatments, H{ickel's theory of aromaticity and Dewar's suggestions concern—ing it—bonding provided some of the essential background for such treatmentsand the latter almost simultaneously led Chatt and Duncanson1 2 to interpretthe related structural problem posed by Zeise's salt. Computers were notyet available - theoretical treatments had to be largely qualitative andcontroversy ensued mainly about the extent of involvement of differentorbitals, Moffitt's treatment13 emphasising a single metal—ring bond wastoo great an oversimplification, but a useful stimulus to further experi-mental work. The Fischer—Ruch treatment1 based on the donation of 6it—electrons (3 pairs) from each cyclopentadienide and hence formation of atypical 6—coordinate metal complex, while also oversimplified, had theadvantage, in common with later treatments, of emphasising the attainmentof a formal inert gas electron configuration by the metal. This continuesto be a most useful guide as to which comVlexes are stable, Indeed as thefield has expanded the exceptions to the inert gas rule' have become adecreasing proportion.

Of the neutral bis—cyclopentadienyls of course only those of Fe, Ru and Oscould fit the rule. In the first row, manganese proved exceptional, formingan apparently ionic analogue with 5 unpaired electrons. The paramagnetic V,Cr and Co complexes have considerable thermal stability, but are verysensitive to oxidation, while Cp2Ni with two unpaired electrons showssurprising stability in both respects. The cobalt complex could attaininert gas configuration of electrons in the oxidised — cobaltocenium saltform — one of the most stable of all cyclopentadienyl complexes, Othermetals could attain inert gas configuration by incorporating additionalligands. Among early examples described the basic () rhenium hydride, 2,and the unique carbonyl of titanium, 3 stand out, Even higher coordination

Re-H °<:5: MV6: MTi,Zr,Hf

2 3 7

number was attained in the Nb compound, 4 and its Ta analogue while vanadium,presumably because of its smaller size, gave the paramagnetic dichioride, 5,The very stable dihalides of the Ti/Zr/Hf group, 6, are 16 electron complexes,but since all outer electrons are bonding they are necessarily diamagnetic.The titanium compound yielded1 the first of the new stable transition metalalkyls, Z.

842 PETER L. PAUSON

Alongside such compounds were discovered the first of the mono—cyclopenta—dienyl metal derivatives. The examples shown (Fig. 8) illustrate thevaried types differing in counterligand. We see that their wide rangeincluded cationic and anionic derivatives The carbonyls (Fig 9)

9ci'Ict Ci I'o ON l'CH3 OC rNOCi CI NO CO

9oc'I'Co CC' I CO C H NC 'CNC6H5

CI CO CO 6 5 CN6H5

9OC'I ,CH2(" H C.C6H5 ICH ' 5'

C.C6H5NO

o,Cq

CC I CO.CH3 NC I CN Br P(C H)CO CN 653

Fig.8were reached either from the biscyclopentadienyl derivatives with carbonmonoxide or from the simple carbonyls with cyclopentadiené and included thefirst binuclear cyclopentadienyls, Again I must emphasise the importantcontribution whicI the crystallographers made in establishing these struc-tures and in some cases even the compositions. Thus both the hexacarbonyl—molybdenum (analogous to Cr) and the iron compounds had been formulated byWilkinson as pentacarbonyls before structure determinations16'17 establishedthe compositions shown.

9 Rr_OC\CO 9OC2\CO "CO OC" I CO

CC CO CO CO

9 _OC' / CO

Fig.9

Bis (cyclopentadienylnickel_carbonyl) required a different preparative methodproving accessible1 8 only by reaction of nickelocene with nickel carbonyl(Fig. 10). The binuclear complex is somewhat unstable losing more carbonmonoxide to give the paramagnetic (C5H5Ni)3(CO)2 — the first trinuclearcomplex of this type.

Aromatic transition metalcomplexes—The first 25 years 843

+ Ni(CO)4

\NiFig. 10

if cyclopentadienyls seemed a unique class up to 1955 the dramatic develop—ments of that year transformed the picture. Apparently several workers hadbegun to speculate about the possibility of forming stable complexes frombenzene and its derivatives, Positive results derive from two of thesespeculations: Lars Onsager postulated that Hem's polyphenylchromiums werein fact compounds of this type and stimulated experimental work by H. Zeissand his coworkers which ultimately confirmed this,1 Zeiss presented theearly experimental findings which strongly supported Onsager's idea inseveral lectures given in Europe in the summer of 1955. But what appearedto be highly suggestive results only acquired the aura of proof after theindependently undertaken synthesis of the parent bis—benzenechromium byE, 0. Fischer and W. Hafner.20 I was fortunate to be a participant in ahalf—day meeting of the Chemical Society in London where this result wasfirst presented, If anyone listening to Professor Fischer on that occasionmight have doubted that a whole new dimension had just been added to organo—transition metal chemistry, there was, at the same meeting, the confidentprediction by Longuet-Higgin.s and Orgel that cyclobutadiene—metal complexesshould also be stable.21

So far as the benzenoid complexes are concerned intensive exploration inE. 0. Fischer's laboratory added complexes of many other metals (Fig. fl)in the next 3 years and in 1958 the synthesis of the arene—tricarbonyl-'chromiums, independently achieved from the arenes and Cr(C0)6 by his owngroup in an22 by Natta' s group in Italy23 and by Whiting and Nicholls

844 PETER L. PAUSON

666CO® CH3Re Fe Ru

CH3

CH3ØC H3

Fig.llin the TJ•K,2 provided facile access to much more stable complexes andopened up world—wide interest in the arene compounds. IndependentlyWilkinson showed that the same reaction works with cycloheptatriene25 andsince the complexes behaved like the free hydrocarbon in transferringhydride to trityl cations, the first complexes containing the aromatic 7—membered tropylium ring became available in this way,26

Development of suitable methods once again proved the key to many of theseadvances. Fischer and Hafner20 had invented the use of aluminium halides tocatalyse reaction between arenes and chromium(III) chloride and this provedsuitable for many other transition metals, . Application to metal carbonylhalides was first tried by Coffield and his coworkers at he Ethyl Corpora-tion27 and mixed cationic complexes such as [C6H6Mn(CO)3] and [C6H6FeC5H5]+resulted, Zeiss' group pioneered the formation of polysubstituted arene—chromium complexes direct from acetylenes1 while in E, 0, Fischer'slaboratory K. Ofele28 and, later H. Werner29 developed more and more sophisti-cated methods of replacing carbonyl groups by arenes using labile complexesof the type: Cr(C0)3L3 as intermediates from which the labile liand (L)could be displaced under sufficiently mild conditions to permit formationof e.g. the iodobenzene, thiophene and even borazole complexes. As a result,

I

CH=CH2 ç5 ç7eCr(CO)3 r(C0)3 Cr(CO)3 r

CH C6H5

C6H5-'TCH3 TC6H5Cr(C0)3 Cr(C0)3

RCH3,C2H5 etc.

Fig. 12

a very wide range of such complexes is now known including the examples inin Fig. 12.

While on the subject of heterocycles, I may mention that K, Joshi, workingin my laboratory was convinced that 7—pyrrolyl complexes comparable tocyclopentadienyls should exist and correctly chose C5H5Mn(CO)3 as the moststable mono—cyclopentadienyl to imitate, He therefore heated pyrrole withmanganese carbonyl to achieve the first success30 with a heterocycliccyclopentadienyl analogue, 13.

In the meantime, the predicted stability of cyclobutadiene complexes had notfailed to arouse interest; the dramatic proof of the correctness of thisprediction was obtained by Criegee and SchrSder,31 who in 1959 prepared the

Aromatic transition metal complexes—The first 25 years 845

NH + Mn2(CO)10cN/P1n\

,,,/IV'

OC CO

tetramethylcyclobutadiene nickel chloride complex by treating nickel carbonylwith dichlorotetramethylcyclobutene (Fig.n). Independently Hübel and Brayetentatively identified one of the products from iron carbonyls and tolane asthe tetraphenylcyclobutadiene iron tricarbonyl.32 These two approaches haveremained the principal routes. Thus Pettit returned to the former method to

Ni(CO)4 i- ML [C4Me4NiCI2]

Fig,14

obtain33 the unsubstituted iron complex in 1965 and the facile formation oftetraphenyicyclobutadiene palladium complexes from tolan, elucidated byBlomquist and Maitlis,3' gave much impetus to further studies in the fieldincluding similar reactions of other metals. Maitlis35 moreover developedan important secondary method when be showed that the Pd and Ni complexesreadily transfer their cyclobutadiene ligands to other metals including Co,Fe and Mo.

Complexes of the smallest aromatic, the cyclopropenyl system have provedelusive — with the exception36 of the single authenticated type exemplifiedby Ph3C3NiC5H5, , and its precursors [Ph3C3Ni(CO)Br]2 and [Ph3C3NiBr(py)2].

C6H5AC6H5—'C6H5

015At the other end of the scale C8H8 is the largest ring known to form stableand aromatic transition metal complexes. The first examples of cycloocta—tetraene—titanium derivatives, e.g. 16, with flat 8—membered rings weredescribed by Breil and Wilke,37 but even greater interest was aroused byStreitwieser's discovery38 of the uranium and related sandwiches, e.g. 17.

What fascinates the organic chemist in such systems is the formation of thestable planar aromatic system in cases where the free ligand is unknown,unstable, non—planar and/or non or even anti—aromatic. Before those ofcyclobutadiene and cyclooctatetraene the complexes of the potentially anti—arQmatic cyclopentadienone were known, while more recently, examples ofcomplexes derived from fulvalene and pentalene have become known in steadilygrowing numbers,

çN K + BrMn(CO)5

Me11—11 Me

Me4-1Me

CL—i——1

Cl

Fe2(CO)9 i- PhCCPh —Ph Ph

Ph4-11Ph

OC •COCo

I7

Equally surprising at first was the realization that metals can havethe converse effect of stabilising non—aromatic forms: fixing double bonds.A splendid illustration is provided by the stabilisation of the ketonic formof phenol in e.g. complex 18, but the most startling examples are the arenecomplexes in which benzene rings are bent and linked as Li—electron donorswith one clearly localised double bond, e.g. .

I have kept rather close to the term 'aromatic' in my title, but so manyinterrelated developments have marked the progress in this field that it is

0 CF

JCF)CF

18 19impossible •to draw the line. Perhaps my dilemma is best illustrated bywhat may be the most remarkable of all the complexes — the carboranederivatives. Surely no conventional definition of arornaticity is applicable— yet in many ways they are closely similar to the cyclopentadienyls. Theyexploded upon the chemical world with Hawthorne's first descriptions (inl965) and in the rapid stream of publications which has followed, what hasfascinated us is the architectural beauty of the structure of these molecules,their variety and their unprecedented rearrangements (Figs. 20, 21).

B-HC-H

PhPh _

Anti- isomer Syn—isomer

OB-HFig. 20Many other if less spectacular developments took place in parallelinfluenced by and in turn influencing the development of the aromaticcomplexes: The work on cyclopentadienyls led directly to our reformulation"0of Reihlen's butadienetricarbonyliron and in the same year Jonassen41 pre-pared the first 1—allyl complex from butadiene and HCo (co) , • This was soonfollowed by general routes which made such n—allyls widely available.'2 Alsoaround 1960, the first cyclohexadienyl complexes were prepared, both byhydride abstraction (22)from diene complexes'3 and by nucleophilic addition"to cationic arene complexes 23, They turned out to be merely the firstexamples of a range of dienyl complexes which so closely resemble the

846 PETER L. PAUSON

—I

oB-H oB-HC-H •C-H

0

OCH •BH

Aromatic transition metal complexes—The first 25 years 847

corresponding cyclopentadienyls thatsuggests itself for such ligands.

/Mn\oc Co

Co

f \11CH2

OCICO

And a little further away from the aromatic complexes there has been a greatvolume of intriguing literature, on complexes of alkenes and non—conjugateddienes and in the last 12 years the pioneering work of E.O. Fischer oncarbene and carbyne complexes and the discoveries, led by Wilkinson and byLappert of many remarkably stable simple alkyls. Throughout the 25 yearperiod under review some of the most unpredictable results have come fromstudies of the reaction of acetylenes with metal carbonyls. Indeed thesestudies go back to the early nineteen forties, when Reppe's group firstisolated several iron and cobalt complexes during their work on carbonylation.The astounding structural diversity which results from acetylenes on complex—ation seems to be unending — (i have already mentioned cyclobutadiene andcyclopentadienone complexes) — and is far beyond the scope of this account.Perhaps I may give a solitary example to show that almost every X—ray deter-mination reveals a new typ.e. The compound was first made 17 years ago,'5but its complete purification and structure analysis has only just beencompleted.'6 It is the product of successive treatment of Co2(CO)8 withPhCCH, then CO and then MeCCH and illustrates —bonding, x-olefin andx-allyl—bonding within the same molecule.

But what of the l970s and beyond? I should be foolhardy to risk wildpredictions and have no intention to gaze into a crystal ball. But a lookat recent advances and current trends may give us some hint of where rapidprogress is most likely. Each new method of preparing complexes has addedto the range and variety. In the seventies the most impressive example hasbeen the atomic vapour method based on P.S. Skell's work and brilliantly

Fig. 21

the description 'pseudo aromatic'

9/ Fe\

OC COCo

Ph3C® CIO /IOC co

22

23

2Li

developed by Timms and more recently also Green in the transition metalfield. Much is still to be expected not only from this method, but alsofrom the development and more systematic application of both electrochemicaland photochemical techniques. The power of the latter has long been recog-nised, but application has been sporadic despite the success in obtainingunusual structures not readily available otherwise. A few examples may serveto illustrate this:

Pr'MgX/hvFeC3 -'- LJJ

Fig. 25Fischer and Miiller' employed irradiation to promote reactions of metalhalides with unsaturated hydrocarbons in the presence of (CH3)2CHMgBr asreducing agent (Fig. i). Yet the way irradiation assists this versatilesynthetic reaction remains obscure.

hi' Rh"Rh

//Rh

26/Fe\

hi' _____

oc /1.1 0

848 PETER L. PAUSON

Ph

OC

p6

Qa

Pr'Mg X/hvFeC3+ 1- —0-

U

27

Aromatic transition metal complexes—The first 25 years 849

Many workers have used irradiation to induce dissociation of carbonyl groups.Results include the formation of cluster compounds, e.g.'8 26, themselves thecentre of much recent interest and surely offering great scope for furtherexploration. Other photolysis products include multiply metal—metal bonded

C6H8/ hi'

(i\J /Fe

systems, e.g.'9 27, and irradiation has made possible the simultaneoussubstitution of two or three CO groups from otherwise very stable complexese.g.50 (Fig.

If triple dekers have been the subject of speculation for 20 years theisolation of the first examples51 52 (Fig.29) undoubtedly opens up not onlythe likelihood of finding analogues with rings other than 5—membered — butthe possibility of multilayered structures.

[(C5H5)3N JtBF (C5H5)2Co2C2B3H5 isomers

Fig. 29Heterocyclic complexes are still rare, but I doubt that this reflects theirinherent instability rather than insufficient effort to make them both byapplying known techniques and by devising new ones specifically suited forthis purpose. In the former category application of the metal vapour methodmay prove rewarding, while a new approach has been pioneered by Herberich whoobtained a series of borabenzene complexes by ring expansion reactions,53e.g. 30. It is one example of bow studies of reactivity of complexes have infact contributed to the syntheses of new types of organometallics. There are

ANOC CO

CO

C H /h&'8

Fig. 28

+

Violet —black 1,7,2,3 red 1,7,2,4 green

850 V PETER L. PAIJSON

B BrC6H5

SnBr4 C6H5

a o 30

many others — including a variety of ring expansions and contractions.Fischer and Breitschaft's early observation5' that attempted Friedel-Craftsacylation of benzenecyclopentadienylchromium or —manganese gives thecorresponding tropylium complexes (Fig. 31) has remained one of the mostfascinating examples perhaps. not least because no satisfying mechanism hasever been advanced to rationalize it.

_cz? _I)AICI + NaSOCr •s- CH COd. Cr 2 2 Cr

3 2)k20

c::7— CH3 CH3

Fig.3lI have chosen to concentrate on preparative aspects in order to illustratethe multitude of structural types which have become known. Many otheraspects could and should perhaps have been stressed and I can only mentionthem very briefly. Stereochemical features of the structures have been apersistent source of interest and the elegant studies of Brunner and ofTirouflet have provided a multitude of examples of metal chirality and aidedstudy of reaction mechanisms. The observation of single proton resonancesin —cyclopentadienyls and in cyclooctatetraenetricarbonyliron — inapparent conflict with solid state structures — first drew attention to theremarkable fluxionalityof numeros complexes, so elegantly elucidated andwidely studied by F. A. Cotton.5'

The reactivity of the organic ligands in complexes rather than mere prepara-tion of the complexes has increasingly become the main centre of attentionand presents problems which will continue to offer an interesting challengefor many years. Ligands may be split off or exchanged; they may undergonovel additions while still bound to the metal or at the moment of liberationas in the elegant work of Pettit on the C4HFe(C0)3 system. They may beactivated towards electrophilic or nucleophilic attack. The former is thebasis of the extensive aromatic chemistry of ferrocene and of C5H5Mn(C0)3 —a subject which would have merited a review lecture of its own in a moreorganically oriented symposium. It has been contributed to by many, butmost consistently, indeed massively by the group at the Institute of Organo—Element Chemistry in Moscow under the leadership of Academician A. N.Ne smeyanov.

Nucleophilic attack has been equally widely studied - but even the simplestreactions still provide problems of interest. I shall illustrate thislargely from our work. Strong activation of halobenzenes by the Cr(C0)3group was first demonstrated by Whiting who showed that nucleophilicsubstitution by alkoxide occurs under mild conditions.2'55 Nesmeyanovdemonstrated that even greater reactivity is bserved in substituting thecationic complexes of the type [C5H5FeC6H5C1] (Fig. 32) and used a widevariety of nucleohiles to do so.56 Our recent studies57 of[(C6H5C1)Mn(CO)3] showed a marked further increase in reactivity emphasisingthe effect of the counterligand. With strong nucleophiles, e.g. hydride orcarbanions, addition is always preferred to substitution and occurs regard-less of the nature of the benzene substituents although these were shown toexert interesting directive effects, e.g. . The cyclohexadienyl complexesformed by addition of carbanions were shown58 to undergo hydrogen migrationson heating, e.g. , probably via Mn—H derivatives. Similar migrations hadpreviously been studied in the C7H8Cr(CO)3 series.

Aromatic transition metal complexes—The first 25 years 851

_ OxFe + X —' Fe + Ct

0 0Fe + X Fe + C1

e.g. x R0, RS, RNH2

Fig.32The Mn complexes also undergo ligand displacements with hydrogen transfer,

— perhaps via the same intermediate, 36. Simple ligand displacements,37, were long known in the ArHCr(CO)3 series,59 but displacement preferences

MeN MeN

Me(7NMe2 2I7® LitH4 Me MeMn Mn + Mn 33(CO)3 (CO)3 (CO)3

97% 3%

Me

H

- 1350e

+M

R(CO)3 (CO)3 (CO)3 JJ

Mn-H(CO)

çI7 MeOCI

Me Me 36

Mn _________ Cl Mn + Mn CL 35(CO)3 (CO)3 (CO)3

are unknown and are currently under investigation. If displacement of e.g.alkoxy—benzenes with chioro—or fluorobenzene can be effected sufficientlyefficiently, the cyclic process, , becomes an attractive method of phenyla..tion in which the relatively expensive chromium carbonyl component can nowbe seen in an essentially catalytic role.

In contrast to the catjonic complexes which add nucleophiles so readily, few

852 PETER L. PAUSON

ArH+ (Ar'H)Cr(CO)3±Ar'H+ (ArH)Cr (CO)337

QOR

Q ORRO Qx

38

Cr(CO)3 Cr(CO)3

X=C or F

additions to neutral complexes are known. But Semmelhack60 has shown thatmany carbanions add smoothly to C6H6Cr(CO)3 giving unstable adducts, buthigh yields of the free substituted arenes on oxidation. We have encount-ered another amusing example of addition to a neutral complex. Duringpreparation of the above cyclohexadienyl—manganese tricarbonyls we notedthat excess LiA1H4 gave two by—products.61 One of these had been isolatedby Winkhaus,62 who on i.r. evidence alone formulated it as 39. Its n.m.r.

_ c79LAHMn —'- Mn + Mn

OC CO OC CO OC CH3

CO CO CO

40

OC—Mn—HOCL.—H/\ 0C7INH

OC CO CO

39 41

spectrum in fact shows clearly that it is the dicarbonyl—methyl, Butthe second product is of greater interest. It can be obtained in high yieldby first preparing the cyclohexadienyl complexes and then subjecting them tofurther hydride reduction. 1H nmr suggests the formulation . Thespectrum shows coalescence of the signals on warming, corresponding to rapidinterchange of all the protons in two groups: an 'exo'—pool of six and an'endo' pool of three hydrogens - possibly via e.g. 39.

These rather arbitrarily chosen examples are designed to show bow much wecan still learn about apparently simple reactions. Other examples wouldprovide better evidence of the steadily increasing value of transition metalorganometallics to the synthetic organic chemist, although most are againfurther removed from the 'aromatic' class. I think particularly of thewell known and elegant work of Wilke on the cyclooligomerization of buta—diene and many related reactions of allyl—Ni complexes. I think of a widerange of fascinating condensation reactions studied by Heck63 in America, byChiusoli6' in Italy and by Tsuji65 and others in Japan. These reactionsin which quite simple molecules are combined in one step to produce muchmore complex products have already been exploited by Corey and others innatural product and other organic syntheses. In my own laboratory we arestill perfecting a cyclopentenone synthesis66 which, illustrates well howmuch can be achieved in one step (Fig. Methods such as these, or thereactions of cyclopentadienyidicarbonyliron complexes explored by

Aromatic transition metal complexes—The first 25 years 853

+ CH2CH2 ---Ph 0

oc"ICo Co. O _ Ph.

. RPh÷ .ø

Ph

.Fig L2

Rosenblum,67 the dienyliron complexes studied by Birch,68 arid thesynthetic methods using tetracarbonylferrate developed by Collman69are rapidly taking their place among the standard methods of theorganic chemist and I believe that we will soon see such organometallicreagents as widely accepted as say organolithium compounds. Finally theeffect which all this chemistry has had on the development of homogeneouscatalytic methods will I hope become clear from some of the later lecturesin this symposium since neither the time nor the scope of my lecture permitsme to touch on this vast field.

I have given what is clearly a highly personal account of the developmentsas they have impressed themselvon my memory in the last 25 years. I wouldlike to thank the organisers for inviting me to give this account and mycoworkers who have enabled me to play a small part in these developments

REFERENCES

R, D, Brown, Nature, 566 (1950).2 S. A. Miller, J. A. Tebboth and J. F. Tremaine, J.Chem,Soc. 632 (1952).T. J. Kealy and P. L. Pauson, Nature, 168, 1039 (1951).W. C. Zeiss, 9 632(i27).H. Reihlen, A. Gruhl, G. v. Hessling and 0. Pfrengle, Ann. , 161 (1930).

6 F. Hem, 52, 195 (1919).E. 0. Fischer and W. Pfab, Z.Naturforsch. 7b, 377 (1952).

8 G. Wilkinson, M, Rosenblum, M. C. Whiting and R. B. Woodward, J. Amer.Chem. Soc. j, 2125 (1952).R, B. Woodward, M. Rosenhiurn and M. C. Whiting, J. Amer. Chem, Soc. ,3458 (1952).

10 G. Wilkinson, J. Amer. Chem. Soc. j, 6146, 61L18 (1952).11 E. 0. Fischer and R. Jira, Z.Naturforsch. Sb, 217 (1953).12 J Chatt and L. A. Duncanson, J. Chern. Soc. 2939 (1953).13 W. Moffitt, J.Amer.Chern., 3386 (l95Lt).1 E. Ruch and E. 0. Fischer, Z,Naturforsch. 7b, 676 (1952); E. Ruch,

Sitzungsber bayer Akad. Wiss. jnchenmath1-turwiss. I. 347 (1954).1 L. Summers and R. H. Uloth, J. Amer. Chem, Soc. 76, 2278 11954).1 6 F. C. Wilson and D. P. Shoemaker, Naturwiss. , 57 (l956 J.Chem.s.

27, 809 (1957).17 0. S. Mills, 620 (1958).1 8 E. 0. Fischer, Angw. Chem. , 715 (1957); E. 0. Fischer and C. Palm,• Che., 91, 1725 11958).19 See e.g. H. Zeiss, Chapter 8 in H. Zeiss (Ed.) 'Organometallic Chemistry'

Reinhold, New York, 1960.20 E 0. Fischer and W. Hafner, rLgil. i2., 655 (1955).21 H. C. Longuet—Higgins and L. E. Orgel, J. Che. 1969 (1956).22 E. 0. Fischer, K. öfele, H. Essler, W. Fröhlich, J. P. Mortensen and

W. Semmlinger, Z•Naturforsch. l3b, 458 (1958); Chem.Ber. 91, 2763 (1958).23 G. Natta, R. Ercoli and F. Calderazzo, çhirn. elnd. Q, 287(1958).24 M. C. Whiting and B. Nicholls, Proc • Chem. Soc. 152 (1958); J.çhem. Soc.

551 (1959).

— -- --— (L6t) Lit 'Z

wetjo xeuiy f '3tITo 0 PU TCnj-IttBoA N 'ITCH J, H 'oiiTeunues j 09 (9c61) cooi '7 i---—ii:Eii5 'otoiqwueg pire 'tiooxa i 'v e

(9 L 6T) :i sac' U09fld 'j pu oIT1nw w v ç (cL6t) £9t 'LL9T 'itd—5:F te8eg v r Pu U0Sfld '1 d LS

(96T) iC ' 'oseo N • I pU A09I • S •'l ' nueoJ V N ' A0U(6WSON N •V (9 61) 9T '3tt8N111tTPt Tepkj 'U S S S )FLNP)WrOAZI 'Ae9I S '1 P A099t0fl

_____ N I 'flJT0A V N 'AorndCewseN N v ' (L96T) 909

' LI i S S S TToa 'Aoseo N I PtI nue TA v N 'A0UL0UISON N V

96T1 £l19 'Z (L961) gcLtcc 5 fi

'[(9961) i699 ' 9 •x:i:s:_riiieii] (c96t) TLOcC 5 a '9utttTj (L6i) 6 'i uiet awouiO F 'U00O V ,

- - - - — — ___________ (9961) CT '66

xea wot '(96I) i6 'L weij etxy S xetjost 0 a (TL6T)9CT •tmuiwo3 uietj3

'IeTInw F tO}{ H 'stexe e 'toteqxeH 0 '[sos ' uieiif:1

(oL6i) s:s ' 'reH t H PU sstexe e 'totieqJe a e (9L61) 6 -: 'sewiJe

N H ptre G9TeA i (CL6T) 91OC ' 'ue[ •f •e

pu •j4 'sewIJe N U 'uoppeug 3 'l '18111W 'H A 'eea 0 E [oc6 'r ——ij:ii (L6i) 6i6 '11 'ieu1eA H t xezis v '(L6T) 6C ' wet ei0u] tt.üX 'iezT5 v pu xeuxe

• (cL6i) CTO ' 53ewouio •f '-(eg r •f pu uosrwcI "-I d f(tiL6T) 69 •UTUUWOO •wt1O 53f

'PtL A •W PU epL a 'I 'uosetj •A L 'GçAe'1 •f 'uostxT.tof •ID iI 'a

_________ [tic '17

(iL6T) ç;; iiiito euy jxosn A ieuieo pu puiqueptH 'xepos •i t(iL6T) C9 unwwo3wetj3gQf S3{ N N 'IT03{

•w 'TM 'A 'TUTIOW 1 'ANOS0H J 'Tt0120ZThI 1 'T't'W I-S 6 . • uTaxet seoueeje pire (L961) TC ' weto reew0u1O I' 'snTnd pu TTThI o , —— —---— ______________

(C961) LT xea weti Zi 'ietTnw F pre ieijos 0 a I. ssaxd u ' wit' iewou'o9 ' sv.t a it pu uowA y ci 'U0Sfld 'l d

'xeiewoj •f 'xou j 'Lezjje y r 'extjn 5 f Q 'iePIa V d 9! (6c6i) 6CC 'i 305 uxegjewy 'iepuep

pue epei 'V •U '1Cq:re • 'euuo •a '0 's()rtuIS r 'IeqtLIe5 H , (1961L9o1i1__PUIPUU1tT2 'U0StITTTTA P" seuo [ (o961) 616 'L wet[3teuy 'xetjosj [ ptre ietosT o a

(1961) 1091 '? •005 wet ieuiy •f '0H A •;r pire seTieenw •'-I •a 'sddTlD N H'u'qeOH H H 'uTTeTooi

•u •A t(o961) oL ' °°S wet3Jewv 'f 'AoTsaX 5 Q pire )OJ[ •[ •U , (s6t) 99 ' •S wetj ieuiy •f • •v pt.re eoow t •c[ 'wueN •f 'suree5 •I •H 'ueseuo • H , 6fl19 •30s uietj' 'UOGfld •'-I c1 pre wf 'A •a , (cL6Tl L6 '001 'wettoTew1o --: 'etLI0tA} •I •H (c96T) SIST

7J oos wet ewy f 'Jeu2eA V d ptre &TXflOA 0 Q 'LI0tflI'tH 1 N '(96T) fli thiiiiÜii3 wé5 'sbtelpmj a , ptxe euiottj N

___________________ (s961)

''19CL '06 005 weT.toJeuIv f 'JJotxeseA—xeT1 fl pire • [9 6S 'c Pa • uxeui] 6T) 6 ' 'ejfl • f TT'H • H

(L6T) 6CT '1F 5 wetj xeuiy 'uox 'f j, pire iePtretO N t '(oL6T) 19611 '6 wet ewyf 'IeAep '1 ([ Pre 1flL N U 'TT3flU

(I N '(11961) 1109 '( wet I0UI 'e114eN V S 2u111't0f) t a (99611 6 't7 wetj 10W0U1O A 'StfltN N ci eec

(96T) 6C 'T 305 '9t1tN N ci ptre stnbwoT v , (c961) ICT 'L9 305 wetjJewf 'tied U S4 'uosxewa ci

(0961) 96L '991 0IflN 'Ie)p3tUOt35 A pu epo ' 'jo (6C6T),C '01 wetio TN 2xoui c 'eLei a

pire eqn A '(6c6-r) 1io ' iett 5iif 'puz00H 0 UtN a S f) 'U1'tOJ[ V a 'eNIetLIM f 'ssie a 'esnelo v 'ekeIa a 'eqnH t

(6C61) I 'C9 1iV '(6C61) oL 'Z iiitj0 AÜ 'zepoxtog ptre ee2etx0 u (2i) 9C S UIOtiO xci jØ9fl ' c Pu tt[SOf N N

(6961) C6 'o-r Iea'wetio 'unewleioea a ptre zuTJci 'xetLIet H '[16 ' a tLxeuI:1 (L961) ç9 ' wetç teuv 'Ieu1eA H pire ZUII

(9961) C1 ' Iea •uie5 'elejo N

-- (LC61) 99C

'6t oo wetio xewy 'uoceo a u pyre 1PUS A 'PTeTJJOO H J • (9c61) 0LCC ' oo'wetiieuiy 'ueuuoH pire ueqrie c 'H

______________ (6C61) oog wetiof

'(9c61) CT 305 w9140 ooici 'uosutft ptre 4euuea V N 'eqy t a

MOSflVd .i aLd

Aromatic transition metal complexes—The first 25 years 855

61 G. A. M. Munro and P. L. Pauson, unpublished work.62 G. Winkhaus, Z.anChem. 319, LO1 (1963),63 e.g. R. F. Heck in i. Wender and P. Pino (Editors) '0rg1c Srntheses via

Metal Carbonyls', Interscience, New York 1968.e.g. G. P. Chiusoli, 1139 (1969).

65 e.g. J. Tsuji, Accounts chem. Res, 2, 14CT969); 6, 8 (1973).66 , u. Khand and P. L, Pauson, J,Cbern.Res, 1977, in press.67 e.g. M, Rosenbium, Accounts Chem, Res, 7, 122 (197k),68 e.g. A, J. Birch, K, B. Chamberlain and D. J. Thompson, JCSPerkinI.,

1900 (1973); A. J, Birch and A, J. Pearson, Tetrahedron Letters 2379(1975).

69J• P. Collman, Accounts Chem, Res. 8, 3l2 (1975),