7 offprinted from the Discussionof The Faraday society,1962,No. 34 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY. STRUCTUREAND DYNAMIC CHARACTER OF ALLYLIC GzuGNARD REAGENTS SPectroscoPY'Structureand NuclearMagnettc Resonance * Grignard Reagents Dynamic Characterof Atlvlic D' Rosenrs Entc NonoleNDER AND JoHN Bv Geoncs M' Wstreslors' J' Instituteof Technoiosv' Gatesand Crellin Laboratories r"^l^::ilttil,|orl1,lfornia ReceiuedIst June' 1962 of several allylic Grignard magnetic resonance spectra have been made of the nuclear bestformulated studies ir"ti, r,i;qr*r,rt-r*i"'r'r'uiorn'c'.1ryru" reagenrs.The Grignar<r reagent o..ooit asan equiribrium "i;;;"';i-r,r-oi"ll,rjrr,,riri-"rii.:*',. tllnTi,#n:J[*1':',11f:i']'[ l,'""*'lJnilfi*:[;|"ff:',fi :',xfJ'""J.:"#""'H:";:::;tiJr''"'"^-io" rn- Theabiiityo|allylicGrignardreagentstoyield'p'o9.11:',^0.''u.ofromtheisomerhas stimuratedextensrve startinghalide and the .or..rponii"g-"iLvric vestigationintotr'. .on,tit'tion ;i,d:i{i{:::;:l'l*;'\lT;:"1fi[ii",,t:"",i,H thcse investtgatrons #t*'*:[tiilT1*lt'in';[iiiiti:':',i.!1,:',q:l$i*:i"treac'|i'ns'rg..tilf re'rielvsgive detaiis of i here. of equilibrium ;.J;T"i and'degree a*ylicGrignard 1r,1r'3^i-ltitnce(l): of the type shown in eqn' betweenallylic i**"t" (l) "#;;r^hF Z CH:-CH-CH:CHz CHr-CH:CH-CHz-MgX MgX gnesium bromide cr-methYlallY'ima crotylmagnesium bromlde butenes that the proportions.ofisomeric young co_workersr,vere able to shorv of the and pr.purld-.olplete^' iior r arious mixtures on rrydrolysisof Grignardreagenrs of com- obtarned ,".r. independenr un",l,-*.thyra*yr brornicle butenes isomericcrotyr_ the rtduciionoi thehalides to -r,4 simiiarresult *u, oUtoinedin position: a data *'ere originallyinterpreted of finerydivided metars.s-rThese in the presence of bute*yiorganometallic compounds' evidencefor a rapidlyequilibraffil]*,.rr" the extent as isomers_existedto aDproximately in which both crotyr-and a-me.trrvrulyr jormed by hydrolyri:' of the ..rp."tiu. butenes to th. .on,.position to reactwrtn corresponding crigllurd in iut.r" riuJi.r, trie uutJnvi- .reagen'.'lii,::,lnd However, n ti yield almosterclusively e-methylallyl carbon dioxide8 and phenyi ir";;;;;;. derivatives(eqn. (2)): !t, -+ Q) C+HsMd(+Co2 CHz:CH-in-CozH (- 100?;) d'B-unsatur- of-ald.ehydesand and simiiar resultsusing a variety !.,,^",i:t't-11rendering the These ilaiides12 were inter^oreted as ated carbonyicompounds,e u.ro-urrlii. Contributionno' 2847' of Naval Research' t * supportedin part by the Office 185 4 r86 ALLYLIC GRIGNARD REAGENTS hypothesis of a rapidly equilibrating mixture of Grignard isomers less satisfactor\. than an alternative formulation in which the Grignard reagent is taken to be almost exclusively the crotyl issmsl.ll' 13 In recent years, the problem of constitution of allylic Grignard reagents has been further complicated by discovery of a remarkable series of allyl-transition metal compounds, whose structures seem rvell established as n-compie,res of the metal with the aliyl moiety.14-16 The possibility that allylic Grignard reagents might be similar complexes of magnesium with the allyl residue had been suggested eailier r r but offered no special advantage in interpreting the chemrcal evidence then avaiiable. Despite the large amount of pertinent chemical evidence, the problem of the composition of allylic Grignard reagents could not be considered soived; indeed, Kharasch t has asserted that the problem is incapable of purely chemical solution. The work reported in this paper rvas initiated in the belief that nuclear magnetic resonance (n.m.r.) spectroscopy offered an especially appropriate, non-destructive physical method with which to investigate the constitution of simple allylic Grignard reagents in solution. EXPERIMENTAL The allylic Grignard reagentswere prepared in the cyclic reactor shorvn in fig. I through the reaction of a dilute solution of the corresponding halides in diethyl ether with a column of amalgamated magnesium turnings (prepared by treating the magnesium rn sirr with an equal weight of mercuric bromide dissolved in ether, follou,ed by thorough washing rvith ether). Amalgamation was found to reduce materially the yield of coupling p.odu.tt formed in the reaction. The magnesium used in these experimentsrvas turned from a bar of very pure resublimed metal: use of ordinary magnesium invariably gave samplesrvhose n.m.r. spectraappeared to be broadened by paramagneticimpuriries. The preparations normally were carried out with 0.01 moie of halide. The rotary selector of the cyclic reactor was initially positioned away from the flask with the sintered- glassdisc. After 0'015 g atom oI magnesiumhad beenamalgamated, activated rvith ethy4ene dibromide,lT and washed rvith ether, the flask containing the w'ashings\\'as replaced rvith a flask containing fresh ether, and the ether heated to reflux. The allyiic bromide was then added slorvly to the stream of condensing ether, and the top piecesof magnesium rvere punctured with the tungsten needleto initiate reaction. After an initial induction period, a vigorous reaction occurred : rvhen tiris had subsided,the florr' from the column rval directed into the second terminal flask. The reaction rvas allorved to proceed at such a rate that slight ebullition was always apparent in the column. When rhe addition was complete,the solution in the terminal flask was filtered through the sintered-gjassfrit and concentratedin a stream of dry nitrogen in the round-bottomed ffask equipped u'ith n.m.r. tubes as sidearms. The concentratedsolution was decanted into the n.m.r. tubes; these were broken off, charged with tetramethylsilane, sealed. and centrifuged to remove suspendedsolids. With this apparatus, it was possible to prepare allylmagnesium bromide * and buten-,-i- magnesium bromide essentially free of coupling products. It proved impossible to prepare y,7-dimethylalll'lmagnesium bromide by this method without serious contaminaiion by coupling products; holvever, the coupling products could be removed from saturated ethereal solutions of this Grignard reagent by extraction of the solution th-reetimes with an equal volume of dry iso-octane (or similar saturated aliphatic hydrocarbon). These ex- tractions were conveniently carried out directly in the n.m.r. tube. After addition of the iso-octane, the tube was stoppered and shaken vigorously. The layers were separated by centrifugation, and the upper (iso-octane) layer removed w.ith a dropper. After the final extraction, the thick residue was diluted with approximately four times its volume of fresh ether, mixed, and centrifuged. t In this paper, the solvatedorganometallic compound prepared from ail-"-lbromide and mag- nesiumin etherrviil be cailedallylmagnesium bromide, although investigations into the consriturion of Grignardreagents indicate that this representationis at bestincomplglg.ls, le J' D' ROBERTS 187 G. M. WHITESIDES'J. E' NORDLANDERAND l0 Grignard reagent; N.m.r. spectra rvere taken using solutions. approximately % pronouncedviscous broadening, particu- spectraof more concentratedsolutiois dispiayed to be.indefinitelystable; on larly at low tempe.ut.rr... The Grignard reagentsappeared bromide)frequently occurred standinga slow depositionof crystals"(presumablymagnesium in the simple tubes,but the spectradid not change' spectrorneterequipped with spectrawere taken at 60 Mc/s.. o" u vurian model v-4300B audio sidebandtechniques 20 a Superstabilizer. chemical shifts weremeasured by standard Model 521c frequencycounter' using a Hewlett,packardNlodel 200A8 audio oscillator and previously;21temperatures in the probe The low-temperatureapparatus has been described togetherwith a Leeds were measuredusing a caribratedcopper-constantan thermocouple and Northrup precisionpotentiometer' tltl tl TNCnCASEOGATt i ,l TUNGSTENNEEOLE lil HERSTIBERG ,'li DROPPTNGFUNNEL i,i'l r'll )iJ''t't- , ri I crH,o i I rl,t :: I i REFLUXCONDENSER *'J'J'v---/* NITROGENSY.PASS AND \ t . FLARE TO PREVENT \ srPHoNlNG ORIP TIP magnetlcresonance spectra Frc. 1.-Cy'clic l'eactor and accessoryapparatus Frc. 2.-Nuclear bromide (top) butenyl- in the preparation of the allyiic Grignard of ailylmag,nesium used (middle) and reagents' magnesiumbromrde v,v-di' meihy'iallyimagnesiumbromide (bottom) in diethyl ether at 33' and 60 N'lclsec' The intensequartet and tripiet in eachspecrrum are the resonancesof the dierhylether CHz - and CH3. The complexof linesat 55c,rsec in the bottom spectrumis due to iso-octane' Chemicalshifts ire in c'secfrom tetramethyl- silane(0 Nlc,'sec)as external (top) or internal (middleand bottom) standard' DISCUSSION 21 zz (fi1' 2); that Allylmagnesium bromide shorvsan AX+-type n.m.r. spectrum respectto the proton is, the four a and I protons are magnetically lqoiua.ot with B I88 ALLYLIC GRIGNARD REAGENTS and coupled to it with J : l}'clsec. This spectrum can be accounted for by postulatingthat the Grignard reagentis either a rapidly equilibrating(r140.001 sei) mixture of classical structures (I) or a bridged structure 'uvithmagnetically equiv- alent protons (II). Tire same interpretation is required for p-methylallylmagnerium bromide whose spectrum shows a methyl group and only one other kind of proton. CH CH CH /\ kr/\/,\ HzC CH2MgBr ;* BrMgCHz CH2 HzC CHz ,(_1 Mg Br II The n.m.r. spectrum of butenylmagnesiumbromide (fig. 2) sholvs the spin- splitting and pattern of chemical shifts consistentrvith crotylmagnesiumbromide. This spectrumindicates either that the substanceexists exclusiveiy as crotylmagnesium bromide or that the position of equilibrium of a raoidly