Extension of the Woodward-Hoffman Rules toOrganomet allic SystemsBY R. PETTIT, H. SUGAHARA, J. WRISTERS AND W. MERKDept. of Chemistry, The University of Texas at Austin, Austin, Texas 78712Received 3 1st January, 1969It has been demonstrated that in the presence of catalytic amounts of certain metal ions or metalcomplexes, various strained derivatives of cyclobutene can undergo an extremely facile disrotatoryring opening to yield derivatives of butadiene. An explanation for the " apparent " break-down ofthe Woodward-Hoffman rules is offered in terms of the formation of intermediate organometallic picomplexes and consideration of the subsequent energetics involved for the ring opening reaction ofthese complexes.Our interest in the application of the Woodward-Hoffman rules to organometallicsystems arose from the observation that dimerization of benzocyclobutadiene proceedsin different directions depending on whether the reaction is carried out in the presenceor absence of silver i0ns.I Cava and Nenitzescu and their coworkers hadshown that dimerization of benzocyclobutadiene proceeds to give the C1 6H12 hydro-carbon (I) and the plausible pathway below was proposed.However, we observedthat when benzocyclobutadiene was liberated from its iron carbonyl complex in thepresence of silver ions, then the isomeric material (11) was then only C1 6H12 productformed. Similar reactions when conducted in the absence of silver ions gave the" normal " dimer (I). Further experiments along these lines indicated that thesilver ion was allowing a disruption in the pathway leading to (I) and the nature ofits involvement became the subject of this study.(1) UWOODWARD-HOFFMAN RULESThere have been developed a set of powerful rules for determination of thenature of organic reactions which proceed in a concerted manner.3 Three suchreactions which are of present interest are (a) the thermal isomerization of cyclobutene772 EXTENSION OF THE WOODWARD-HOFFMAN RULESto butadiene, (b) cyclic addition of two ethylene units to give cyclobutane derivativesand (c) 1,3-hydrogen shifts in olefin isomerizations.The following discussionindicates that the rules governing these reactions are to be modified when the electronicrearrangements also include participation by metal atoms.CYCLOBUTENE-BUTADIENE INTERCONVERSIONSThe Woodward-Hoffman rules state that the thermal isomerization of cyclobuteneto butadiene will proceed preferentially via a conrotatory process (eqn.(1)) ratherthan by a disrotatory process (eqn. (2)). The reason for the distinction is that theRRsymmetry of the systems involved is such that one cannot proceed smoothly fromthe electronic ground state of cyclobutene to that of the ground state of butadienein process (2); whereas it is possible for process (1). The fact that the disrotatoryprocess is forbidden is evident in the following argument developed by Longuet-Higgins and Abraham~on.~ In the disrotatory process the molecular species involvedin the rearrangement at all times possess a plane of symmetry bisecting the molecule.Those molecular orbitals of cyclobutene and butadiene which becomes involved inthe chemical change are classified as to whether they are symmetric or antisymmetricwith respect to this plane and listed in fig. 1.The electronic ground state of cyclo-butene is 02z2 and if the symmetry of the orbitals is to be maintained, this wouldSymmetric 6,a *Antisymmetric 6, 7T 4 - %?a*)/ k2' P4FIG. 1 .-Classification of the molecular and atomic orbitals involved in cyclobutene-butadienecatalyzed isomerizations.lead an electronic state of butadiene ($ f $ $) which is of much higher energy than theground state ($ f $22). The energetics of this disrotatory process are schematicallygiven in fig. 2. The reaction, if forced to follow a disotatory process (dotted line,fig.2a) will involve at least a high activation energy.Consider now the isomerization of the silver complex of cyclobutene to the silvercomplex of butadiene via a disrotatory process. The ground state of the cyclobutene-Ag+ complex can be satisfactorily described as oz, (7c - ~ 4 2 ) ~ (n* -pXdx,J2 while that oR . PETTIT, H . SUGAHARA, J . WRISTERS AND W. MERK 73thesilver complex of butadienewillbe($l -s&)~, ($2 -pxdxz)4, (t+h3 - p , ~ 5 ) ~ ( $ ~ - dxy).2The atomic orbitals of the silver ion which are involved are also classified as to theirFIG. 2a FIG. 2bsymmetry properties in fig. 1. It is seen from fig. 1 that the transformation 02, (n - ~ d , 2 ) ~(n* -pxdxy)2-+($1 - ~ 4 2 ) ~ ($2 -p,dx,)2 ($3 -p,,dJ4 (+ dXJ2 is a symmetry allowedprocess.Now the difference between this allowed product and the description ofthe ground state of the butadiene-Ag+ complex mentioned above is whether twoelectrons are placed in a pydyz or a pxdxz hybridized atomic orbital of silver. Atworst this difference in energy will be very small; hence, as indicated in fig 2(b), theactivation energy will be small and the process becomes " allowed ". Argumentsalong similar lines have also been developed by Mango and Schachtschneider toindicate that the thermal cleavage of cyclobutanes to two olefinic units is also allowedin the presence of metals.Similar application of orbital symmetry rules indicate that the disrotatory con-version of the silver complexes of benzocyclobutene to o-xylylene (eqn. (3))'.is anallowed process ; whereas in the absence of metal ions, it is forbidden.-c b+CR (3)RConfirmation of these arguments are found in the isomerization of dibenzotri-cyclooctadiene(II1) to dibenzo-cyclooctatetraene(1V). The direct thermal ringopening of the central cyclobutane ring in (III) to yield(1V) directly is a forbiddenH(I11 1 CIV)process. Likewise ring opening to yield the oxylylene derivative 0, and subsequentisomerization to (IV), would have to be a disrotatory process and this also is forbidden.Hence despite the large amount of strain present in the system, the molecule onlyslowly undergoes thermal isomerization, the activation energy being approximately23 kcal/rnol. However, if silver nitrate is added to solutions of (III), the isomeriza-tion is complete almost instantly at room temperature ; even at - 24" the reaction i74 EXTENSION OF THE WOODWARD-HOFFMAN RULES50 % completed within 30 sec.The activation energy of the silver catalyzed reactionis found to be approximately 8 kcal/moI.Addition of maleic anhydride to compound (111) produces no reaction ; however,if silver nitrate is then added to the mixture, there ensues immediate precipitationof the Diels-Alder adduct (VI). This confirms that the isomerization of (Ill) to(IV) proceeds via the o-xylylene derivative (V). The role of the Ag+ is to complexwith the benzene ring and make the disrotatory ring opening to the hydrocarbon (V)an allowed process. The driving force behind the facile reaction is the relief ofsteric strain.co(VI)Similar types of catalyzed isomerizations have been observed with other strainedcyclobutene derivatives.One of particular interest is that of the isomerization ofbenzotricyclooctadiene (VI) to benzocyclooctatetraene (VII). As indicated, the metal-catalyzed isomerization of (VI) to (VII) can in principle proceed by two differentcr X Ipathways, one involving a benzocyclobutene-o-xylylene isomerization (VIII) whilethe other a cyclobutene-butadiene isomerization (IX) ; both processes involve dis-rotatory ring openings. Again the isomerization is strongly catalyzed by silver ions.When conducted in the presence of maleic anhydride, this reaction leads to thecoformation of the adduct (X), thus indicating the intermediacy of the o-xylyleneintermediate (VIII).Thus although the benzene ring is destroyed in this pathway,whereas it would not be if the reaction proceeded via (IX), there are presumablysteric factors present in the molecule which make this the lower energy process.However, the situation appears to be reversed when nickel is used as the catalyst.The isomerization of compound (111) to compound (IV) is not catalyzed by biscyclo-octadienyhkkel, this being consistent with the observation that aromatic compoundsdo not tend to displace the olefinic ligands from (XI). However, the isomerizationof the olefinic compound (VI) to (VLI) is catalyzed by the complex (XI). Presumablythen the isomerization with nickel proceeds via the intermediate (IX).Ring openinR. PETTIT, H . SUGAHARA, J . WRISTERS AND W. MERK 75from this same end is also indicated in the thermal isomerization of the solid palladiumcomplex (XU) to the palladium complex of benzocyclooctatetraene.A limited attempt has been made so far to determine the rmge of metal atomswhich can influence the disrotatory ring openings discussed above. As well assilver ions, we find that cuprous ions and the complexes PdC12(4CN), and[C2H4PtCl2I2 and metallic palladium on charcoal also have a catalytic influence onthe isomerization of benzocyclooctatriene to benzocyclooctatetraene. Silver ionsare also found to catalyze the analogous reactions indicated in eqn. (4), (5) and (6).__I_c IlllnSyn and Anti(5)The results now allow for a reasonable explanation for the effect of silver ions onthe dimerization of benzocyclobutadiene noted earlier.The intermediate species(Ia), in the presence of Ag+, undergoes rapid isomerization to (XIII) before itrearranges to (1). The intermediate (XIII) isomerizes further to (XIV) which thenundergoes an intramolecular Diels-Alder type addition to yield the observed product11.-wCYCLOBUTANE-ETHYLENE INTERCONVERSIONSThe Woodward-Hoffman rules allow one to conclude that concerted cyclobutane-ethylene interconversions (eqn. (7)) are forbidden. Hogeveen and Volger discoveredRho complexes enhance greatly the rate of isomerization of quadricyclane (XV) to/I + I1 (7)norbomadiene and Mango and Shachtschneider have produced analogous argumentsto those given above showing that in the presence of metal atoms the cyclobutaneto diolefin rearrangement now is an allowed process.One particular reaction of interest in this area concerns the metal-catalyzeddismutation of olefins (eqn.(8)). Thus treatment of 1-butene in benzene with atungsten catalyst (produced by reduction of WC16 with Et,Al) gives within secondsat room temperature a mixture of ethylene and 3-hexene. The presumed mechanis76 EXTENSION OF THE WOODWARD-HOFFMAN RULES(XV)of this unusual reaction involves the " allowed " metal-cycloadditon reaction to givea cyclobutane intermediate, reversal of this process giving the observed dismutationproducts. In order to test this hypothesis we have treated all cis-tetramethylcyclo-butane (XVI) under the same condition with the tungsten-containing catalyst.CHS CH, CH2 = cw2CH =cy II + I1CH CHI IE t Et W(8)+Et Et/ - --H-- - Et 5 EtHowever, no reaction occurred, in particular 2-butene was not formed (eqn.(9j).One possible explanation for this failure is that whereas the cyclobutane ring ineqn. (8) is formed in intimate contact with the metal, under the conditions used ineqn. (9), the preformed cyclobutane would need to displace a molecule of coordinatedbenzene before the cyclobutane-metal interaction could give rise to olefinic products.me me(9) - )1 + ,[meme meW(XVI)In agreement with this we find that the gas phase reaction of tetramethylcyclobutanewith molybdenum on alumina, a catalyst system which also effects the dismutationof olefins,8 readily yields 2-butene and subsequent products derived therefrom.Weconsider then that the dismutation of olefins does involve an allowed metal-participating cycloaddition reaction of two ethylenic units and that the failure of thecyclobutane to cleave in the presence of the tungsten resulted from the inability toposition the metal atom and the saturated hydrocarbon in close proximity.1 : 3 SIGMATROPIC SHIFTSThe above discussion suggests that there may well be other types of reactions inwhich the usual orbital symmetry rules become altered upon involvement of a metalatom. One such reaction would appear to be the metal-catalyzed reactions ofolefins. As indicated in eqn. (lo), the thermal supraf'acial concerted 1,3-hydrogenshift is not an allowed reaction according to the Woodwad-Hoffman rules.-4-Earlier suggestions for the mechanism of metal-catalyzed isomerization of olefinshave involved either addition of metal hydride bonds across the olefinic linkage(eqn.(1 1)) or involvement of intermediate n-ally1 metal hydrides (eqn. (12)).M-H -MHR--CH2-CH=CH2+R-CH2-CH-CH2 +R--CH=CH--CH2 (11)1R . PETTIT, H . SUGAHARA, J . WRISTERS AND W. MERK 77While this may be correct in some cases it cannot be in all, for Orchin has shown thatthe isornerization of 3-phenyl propene to 1-phenylpropene with DCO(CO)~ leadsto a product with only very small incorporation of deuterium. A mechanism suchCH /---\IR-CH2-CH=CH2-+R-CH' CHZ+R-CH=CH-CHs (12)M-H1Mas in eqn.(1 1) or (12) should lead to large incorporation of deuterium. Perhapsthen a meclianism such as indicated in eqn. (13) where the presence of the metalatom now allows the suprafacial 1,3-hydrogen shift to occur is operative. An addedH + rn mfeature to be then considered is whether the hydrogen atom on the same side of themetal is the one which shifts or whether it is the one on the opposite side. A recentexperiment by von Rosenberg and coworkers may here be significant. Theseworkers have found that Fe(CO), catalyzes isomerization of the alcohol (XVIl) tothe ketone (XVlII) but the epimeric alcohol is unaffected. It is reasonable to assumethat the iron atom only interacts with the olefin through the outside face of the mole-cule ; hence the hydrogen which is on the same side of the olefin as the attached metal( X I X Iatom is the one which migrates.A transition state as indicated in formula (XIX)is then suggested in such a concerted reaction78 EXTENSION OF THE WOODWARD-HOFFMAN RULESCONCLUSIONThere is now good evidence to suggest that involvement of certain metals candrastically affect the rules governing the concerted reactions of otherwise purelyorganic reactions. The implications of this effect could have important consequencesin synthetic organic chemistry as well as being of interest in its own right and thiscould develop into another important aspect of organometallic chemistry.W. Merk and R. Pettit, J. Amer. Chem. SOC., 1967,89,4787.M. P. Cava and D. R. Napier, J. Amer. Chem. SOC., 1957,79,1701.C. D. Nenitzescu, M. Avram and D. Dinu, Ber., 1957, 90,2541.R. Hoffmann and R. B. Woodward, Acct. Chem. Res., 1968,1,17.H. C. Longuet-Higgins and E. W. Abrahamson, J. Amer. Chem. SOC., 1965,87, 2045.F. D. Mango and J. H. Schachtschneider, J. Amer. Chem. SOC., 1967, 89,2484.H. Hogeveen and H. C. Volger, J. Amer. Chem. SOC., 1967,89,2486.1967,34, 3327.R. L. Banks and G. C. Baile, I. and EC. Product Res. Dez)., 1964,3 (3), 170.J. L. von Rosenberg, personal communication.' N. C. Calderon, E. A. Ofstead, J. P. Ward, W. A. Judy and K. W. Scott, Tetrahedron Letters