lReprintedfrom the Journalof the AmericanChemical Society, 95,8118 (1973).] Copyright 1973by the American Chemical Societyand reprinted by permissionof the copyright owner.

Mechanism of Reaction of Carbon Monoxide with Phenyllithium'

Larry S. Trzupek, Terry L. Newirth,2 Edward G. Kelly, Norma Ethyl Sbarbati,sand George M. Whitesides*

Contribution from the Department of Chemistry, MassachusettsInstitute of Technology,Cambridge, Massachusetts 02139. ReceiuedJuly 14, 1973

Abstract: The product mixtures obtained by reaction between phenyllithium and carbon monoxide in diethyl ether, followed by hydrolysis, include benzophenone (1). a,a-diphenylacetophenone(2), benzil (3), a,a-diphenyl- a-hydroxyacetophenone (4), benzpinacot (5), a-hydroxyacetophenone (6), 1,3,3-triphenylpropane-1,2-dione(7), 1,3,3-triphenylpropan-1-one-2,3-diol(8). and benzhydrol (9). Compounds 1, 2, 6,7, and 8 are produced in signifi- cant yields i 3,4,5, and 9 are produced in trace quantities. Spectroscopicstudies establish dilithium benzophenone dianion (18) as the first long-lived intermediate formed in this reaction; qualitative correlations betweenthe basicity of a number of organolithium reagentsand their reactivity toward carbon monoxide suggests,but does not prove, that benzoyllithium is a precursor of 18. Labeling experiments indicate that the products ultimately isolated following hydrolysis of the reaction mixture are derived from at least two pathways which compete for the initially formed 18. One involves combination of l8 with 1 equiv of phenyllithium and 1 equiv of carbon monoxide, followed by elimination of I equiv of lithium oxide. yielding 17, the lithium enolateof 2; a secondinvolves com- bination of 18 with 1 equiv of phenyllithium and 2 equiv of carbon monoxide. yielding 22,the trilithium trianion of 8. Hydrolysis of l7 yields 2 directly. Hydrolysis of 22 yields 8; reversealdol reactionsinvolving 8 or its pre- cursors generate1 and 6. The mechanism proposed to account for the major products of the reaction of phenyl- lithium and carbon monoxide is outlined in SchemeIII" On the basisof this scheme,plausible paths to the minor products of the reaction are proposed.

he addition of nucleophilesto carbon monoxide, catalyzed oxidative coupling of amines with oarbon activated by coordination to metal ions, forms the monoxide,e and metal-catalyzedoxidation of carbon basisfor a thoroughly explored and useful class of re- monoxide in aqueoussolution,l0 each involve the attack actions. Transformationsrelated to the hydroformyla- of formally anionic groups on metal-coordinated tion process,4additions of organolithium reagents to carbon monoxide. Acylmetalliccompounds haveeither metal carbonyls,scarbonylation of mercury(ll) salts,6 been inferred or identified as intermediatesin many of organoboranes,Tand organocopper reagents,8metal- thesereactions, and this classof substanceprovides the foundation for discussionsof their mechanisms.l'5 (1) Supported by the National Institutes of Health, Grants GM 16020and HL-l5029, and by the U. S. Army ResearchOffice (Durham), The reactionsof nucleophileswith free or weakly co- Grant ARO-D-3 I -l 24-G69. ordinated carbon monoxide are less well understood. (2) National ScienceFoundation Trainee, 1968-1969. and (3) Fellow of the Consejo Nacional de InvestigacionesCientificas y Base-catalyzed carbonylation of alcohols Techrricas. | 965 -l 966. amines,11,12and the reactions of organolithium,r3 (4) M. Orchin and W. Rupilius,Cutal. Reu.,6,85(1972); J. Falbe, "Carbon Monoxide in Organic Synthesis,"C. R. Adams, translator, (9) W. Brackman, Discuss.Faraday Soc., No. 46, 122 (1968); K. Springer-Verlag,Berlin, 1970,and referencescited in each. I(ondo, N. Sonoda,and S.Tsutsumi, Chem.Lett.,373(1972). (5) E. O. Fischer,and A. Maasbijl, Chem. Ber., 10O,2445(1967), (10) A. C. Harknessand J. Halpern,J. Amer. Chem.Soc.,83, 1258 M. Ryang, L Rhee, and S. Tsutsumi, Bull. Chem. Soc. Jap.,37,341 (1961); S. Nakamura and J. Halpern, ibid.,83,4102(1961). (1964); S. I(. Myeong, Y. Sawa, M. Ryang, and S. Tsutsumi, ibld., (11) For examples,see H. Winteler, A. Bieler,and A. Guyer, Hela. 38, 330 (1965); W. F. Edgell, M. T. Yang, B. J. Bulkirr, R. Bayer.and Chim.Acta,37,237O ,1'954); J. Gjaldback,Acta Chem.Scand.,2.683 N. I{oizumi,J. Amer. Chem.9oc.,87,3080 (1965); D. J. Darensbourg, (1948)r R. Nast and P. Dilly, Angew. Chem.,Int. Ed. En91.,6,357 and M. Y. Darensbourg,Inorg. Chent.,9, 169l (1970); N{. Ryang, (1967). Organometal.Chem. Reu., Sect. A,5,67 (1970),and referencescited in (12) Thc reverse reaction, decarbonylation of formate esters and cach. formarnideswith base,has been examined by J. C. Powers,R. Seidner, (6) W. Schoeller,W. Schrauth,and W, Essers,Chent.Uer.,46,2864 f. G. Parsons,and H. J. Berwin,J. Org. Chem.,31,2623(1966). A (1913); T. C. W. Mak and J. Trotter, J. Chem. 9oc.,3243 (1962): formally related transformation, the McFadyen-Stcvens reaction, J. M. Davidson,,I.Chem. Soc. A,193 (1969). has been reviewcd by M. Sprecher,M. Feldkimel, and M. Wilchek, (7) M. E. D. Hillman, J. Amer. Chem.Soc., 84, 4715(1962); H. C. ibid.,26,3664(1961);M. S. Newman and E. G. Caflisch,J. Amer. Chem. Brown,Accounts Chem. Res.,2,65 (1969). Soc.,80, 862 (l 958). (8) J. Schwartz, TetrahedronLett.,2803 (1972). (13) M. Ryang and S. Tsutsumi,Bull. Chem.Soc. Jap.,34, 1341

Journal of the American Chemical Society I 95:24 I Nouember 28, 1973 81l9

-magnesium,11-17-sodium,18'le and -zinc2oreagents with benzene in pentane solution.28gave similar product carbon monoxide require nucleophilic attack on carbon distributionsbut were heterogeneousin the latter stages monoxide in the absenceof the obvious activation pro- of reaction. The quantity of carbon monoxide ab- vided by coordinating transition metal ions. Stabie sorbed when the reactionswere allowed to proceed to acyl derivatives of group Ia and IIa metal ions are not completionvaried with reactionconditions, but typically well-characterized entities, although they have been im- ca.0"8-1.0 equiv of carbonmonoxide was consumed per plicated in reactions betweencarbon monoxide and tert- equivalent of phenyllithium. butyllithiunl,2l trimethylsilyllithium,22 and various Hydrolysis and glpc analysisof the reaction mirtures lithium amides.23'24 led to identificationof a number of products: benzo- The work reported in this paper was initiated in the phenone(1), a.a-diphenylacetophenone(2). benzil (3). hope that a study of the reaction of carbon monoxide a.a-diphenyl-a-hydroxyacetophenone(4), and traces with phenyllithium would provide evidencebearing both of benzpinacol(5) and benzhydroi (9). Careful thin- on the stability of acyllithium compounds and on the layer chromatography of the nonvolatile products of mechanisms of the reactions between organometallic the hydrolysis permitted isolation of a-hydrox\aceto- derivativesof group [a and Ila metal ions and that it phenone(6) and two compoundsassigned the structures might suggest new approaches to the generation of 1.2,3-triphenyl-1,2-propanedione(7) and l.i.-1-triphe- nucleophilic acyl groups.25 In fact, no evidence for nyl-2,3-dihydroxy-1-propanone(8) on the basis t'ri eri- longJived acyllithiun compounds as intermediates in dence outlined below. Certain of these products. t)t this reaction has been obtained, although the existence oo of these substancesas transitory intermediates is sug- 1. CO(1 atm), Et'.,O ll il phl-i ------PhcPh + PhccHPhz * gestedby certain of the data that follow. Nonetheless, 2'Hro thesedata establish the basic sequencesfollowed in the r z reaction between phenyllithium and carbon monoxide, o o ooH oHoH outline a number of interesting transformations involv- phc-Cphlllllll ing benzophenone ketyl, benzophenone dianion, and + PhCCPh2+ Ph2C-CPhr* 34s related materials, and suggestnew ways in which carbon monoxide might be used in synthesis. ooo ptr$cu,ou+ pn[ScnPhz* Results 67 Products. At -78o, phenyllithium in diethyl ether OOH OH OH solution reacts over ca.6 hr with carbon monoxide at I llll PhCCH-CPh,+ PhCHPhrl t reaction at 0o is complete in 3-4 hr. atm pressure; 89 The phenyllithium solutions used in most of the reac- tions in this study were obtained by reaction of bromo- their analogs,have beenreported in previousinvestiga- benzene with lithium metal and contained lithium tions of the reactions of carbon monoxide with marn bromide:26 the reactions of these solutions with carbon group organometallics.13-22 If reaction mixture S 3rc' monoxidewere homogeneousthroughout. Carbonyla- quenched with acetic anhydride before hydrolysis, onlr tion of analogous lithium halide free solutions of two major products are isolated: l-acetoxy-1,2,1-tri- phenyllithium, prepared by transmetallation of di- phenylethylene (10) and l-benzoyl-1-acetoxy-2'2-di- phenylmercury(I[) with lithium metal2T or by metal- phenylethylene (11). The combined yield of thesc' halogen exchange between r-butyllithium and iodo- o (1961); (1962)i Trsns. N. Y. Acad. 9ci.,27,724 (1965); Ph... 35, ll2l r c'o(r atm).Et,o '' ,Ph Ph-4 ,Ph Nippon Kagaku Zasshi, 82, 880 (1961); Chem. Abstr., 58' 12587b phli (l ):/ + tr< 963). 2AcrO / \ / \ (14) F. G. Fischer and O. Staffers,Justus Liebigs Ann. Chem.,500, AcO Ph AcO Ph 2s3 (1e33). (15) K. V. Puzitskii, Y. T. Eidus, and K. G. Ryabova, Izu. Akad. l0 1l Nazk SSSR,Ser. Khim., 1810(1966); Chem.Abstr.,66,9463lu(1967). (16) M. S. Kharasch and O. Reinmuth, "Grignard Reactions of materials is approximately 50-60/": 10 dominates at Nonmetallic Substances," Prentice-Hall, New York, N' Y., 1954, pp room temperature,Ll at low temperature. 910-9 13. 7 was prepared (17) M. Ryang and S. Tsutsumi, Nippon Kagaku Zasshi,82' 878 An authentic sample of compound (1961); Chem.Abstr.,58, 11387 (1963). using a scheme based on the directed aldol reaction2e (1964); (18) M. Schlosser,Angew. Chem.,Int. Ed. En91.,3,287 M. and in low yield by oxidation of 1,3,3-triphenyl-1- Ryang, H. Miyamoto, and S. Tsutsumi, Nippon Kagaku Zasshi, 82, 1276(1961);Chem. Abstr.,58, 11387(1963). propene with potassium permanganate in acetic an- (19) G. Wittig, L. Gonsior, and H. Vogel, JustusLiebigs Ann. Chem., hydride (SchemeI).30 This substancecould most con- 688,1 (1965). obtained by the hydrolysis of the LL, iso- (20) M. W. Rathke and H. Yu, ./. Org. Chem.,37,1732 (1972). veniently be (21) P. Jutzi and F. W. Schrtjder, J. Organometal. Chem., 24, I lated on quenching the reaction of phenyllithium and (1970). monoxide at room temperature with acetic an- (1970). carbon (22) P. Jutzi and F. W. Schrijder,ibid.,z4,C43 preparing a sample (23) P. Jutzi and F. W. Schrbder,Angew. Chem.,Int. Ed. Engl.,lO, hydride. We were not successfulin 339 (1971): U. Schtjllkopf and F. Gerhart, ibid., 6,805 (1967). See of compound 8 by independent synthesisbecause of the alsoU. Wannagatand H. Seyffert,ibid.,4,438 (1965). ease with which this compound undergoes a reverse (24) The structurally related lithium aldimines are stable at low temperatures: H. M. Walborsky and G. E. Niznik, J' Amer. Chem. Soc., 91, 7778 (1969); H. M. Walborsky, W. H. Morrison, III, and (28) M. Schlosser and V. Ladenberger, J. Organometal. Chem,, 8, G. E. Niznik, ibid., 92, 6675 (l 970). r 93 fl967). (25) D. Seebach,Angew. Chem.,Int. Ed. Eng1.,8,639(1969). (29) H. Reiffin "Newer Methods of PreparativeOrganic Chemistry," (26) H. Gilman, E. A. Zoellner, and W. M. Selby, J. Amer. Chem. Vol. VI, W. Foerst, Ed., Academic Press,New York, N. Y.' 1971' Soc.,55,1252(1933). p48tr. (27) G, Wittig, F. J. Meyer, and G. Lange, JustusLiebigs Ann. Chem.. (30) K" B. Sharpless,R' F. Lauer, O' Repic' A. Y' Teranishi, arrd 571,,t67 (1951). D. R. Williams.J.Amer. Chem. Soc.,93,3303 (1971).

l4thitesides, et al. f Mechanism oJ' Reaction of Carbon Monoxide with Pheny'llithiwn 8120 SchemeI. Synthesisof AuthenticSamples of Products meric iorm having the carbonyl group at C'2 and a hydroxyl at C-l. ( NI,i Table I lists representative product yields observed t'-1.- o Ph OAr 'Y4 2. AcrO \/ \-/ TableI. ProductYields ( [) from the Reactionof s2% /\ Phenyllitniumwith ClarbonMonoxide in DiethylEther" Ph Ph Ph I Temp.'C

Ph I cH3Li 21 :6509 z. ph2CO 16 43 t I \ \ nr r 0 + //-\rt -2('r 3; 4 O 0 1 LX 3 H'so{ -70 -1r600 \.J 460/; -7(t 26d7A0 fid 7cl 12 -0.5 " Phenyllrthiurnsolutions had concentrationstretween and Ph -1"0 -l equiv of lithium brornideper equiv of \. Hroz Ph lV and contained phenyllithium Blank entrieswere those not measured" Yields /t \\ )---c b U \L- Px 56qr //v -Ph were determined by glpc, except when noted otherwlse. Deter- oi o mined as krenzpinacol.following acid-catalyzedpinacol rearrange- Ph Pf ment. . Anal-vsesfor 6. 7, and I were only carried out in the ex.- 14, periment listed rn the lasi row of this table. d Obtained using isotopicdilutron techniques; see the text for a discussion.

i -7\.}-l.i in the reaction of phenyllithium with carbon monoxide Ph OAc Ph Ol,i in ether. Yields of 1-5 in the first four entries in the \./ tabie were obtained using unexceptional glpc techniques. i co -70- )< Ac"O ,/--ir I'hLi ((/ Ph +-- - ,! \>- Ph The yielcisof 1 and 6 in the last entry were obtained by 2 Ac"o (r )- 6,1o,l t,r i' usrng a combination of tlc and mass i, ptr isotopic dilution. \. spectroscopir techniques. Compounds 6, 1, and S ll \. HCI 15 'De could not assavcddirectly by any glpc technique we \lH3oI! ir-r,() 60". \. I ,i" were able to devise. Instead,phenyl-d5-lithium was al- \rr lowed to react to compietion with carbon monoxide- Pli I'h 0 Knowrr KM rr0o '\ and th'i reactlon mixture was hydrolyzed. F-r // quantities oi nondeuterated1,6, 1, and S vlerr' added At:-C) /i to the reaction mixture, and the resulting rrrixtttre ol i'HPh i,'t, (J ('lJI)i: oeutcratec analdol reaction,3t nonetheless.the spectralcharacterls' wer(' cstrnrated on the basis of the results of mass tics of this substance,comblnec: with its ongin ani spectraianaivsrs oi the isotopic composition of these chemistry.leave no doubt concerningthe essentraieie- rrrixtures Tire ciiffcrencein the yields of 1 observedat ments of its structure. Its nmr spectrum shows tw(-, ---?0uusing. gipc anti rsotopicdilution analysesis easily drstinctOH resonances,one broad at"6 5.6, onc sharp at rationaiizec on trtc assumption that 8 decornposesln fi 3.8, and one aliphatic CF{ singletat I -1,.I2,in additrorr part to I in the glpc rnjection port (oide infra\. to aromatic absorptions; the presenceol the hydroxyi The necessrtylor using isotopic dilution techniques is confirmed by its infrared spectrum. The mass tci determtne yrelds of 6, 7. and I rendered product spectrum of 8 (70 eV) has no detectable parent"ion but ! distribution stuciies too laborious. to be a practical displays prominent peaks at 183 (Ph2COH ), l8Z rnethocioi approacnrng a study of the mechanism ol (PhrCO. +), and 105(PhCO+). The one puzzlingfeature this reaction. Further, product distributions were dii' of the speatral data for this substanceis a carbonyl in- ficult to reproduce in detail, and product balances frared absorption at 17l0 cm--i. The unexpectedlv greater than 80:,; were achieved only under special high frequency of this absorption may be due to a pref'- circumstances. Nonetheless, two qualitative features erenceof 8 for a conformation in which a carbonyl C-l of the observeciproduct distributrons are experimentally is twisted out of conjugation with the geminal phenyl significant and pertinent to the mechanistic discussions group or, lessprobably, to the existenceof I in the tauto- that foliow" First. the yields of 1 and 2 vary in a re- (31) Compounds i and ii were prepared in attempts to synthesize producible way with temperature: Z dominates at room temperaturs; L dominates at low temperature. 0H ph ,,oH Pir Second, the yields of 1-5 (and presumably 6-8) are \/"/ HO-)--CH Pii relatively insensitive to changes in reaction conditions AcO-+-CH-\- / // \__ DL V_" other than tentperature: addition of dimethoxyethane PhPh*,i (DME) or .lV,lf,N',lf '-tetramethylethylenediamine(TM- o'J EDA) to these ethereal reaction solutions results in an I lt increasein the yield of benzil (3) to ca. lof with de- 2, 9. Efforts to remove the protecting groups fron: thescsubstances under creasesof the same magnitude in the yields of 1 and a variety of conditions led only to fragmentation to l, 6, and other substituting solutions of lithium halide free phenyl- products. Details of these proceduresand of other attempted routes lithium for the lithium halide-containing solutions to I may be found in the Ph.D. Thesis of T. L. Newirth, Massachusetts lnstitute of Technology,Cambridge, Mass., 1971. usually used or changing the concentration of phenyl-

Journal of the American Chemical Society I 95:24 I Nouember 28, 1973 8121 lithium by a factor of 5, resulted only in small changes from spectroscopicstudies of the striking color changes in yields. In no caseswere new products detected. that characterize the reaction. Contact of carbon In an effort to establishthe fate of the phenyl groups monoxide with phenyllithium in diethyl ether or ether- that do not appear among the characterizedproducts of DME solutions produces an immediate intense red the reaction, several reaction mixtures were explicitly coloration in the solution, which persists throughout examined for the presenceof a number of compounds the greater part of the reaction. The final stages of which might plausibly have been produced. Careful reaction are characterizedby a transition of the solution analyses of representative reaction mixtures demon- color from red to greenishbrown. strated that benzene, benzoic acid, benzaldehyde, Our initial attempts to identify the substance(s) benzyl alcohol, biphenyl, triphenylmethane, triphenyl- responsible for the red color were guided by the hy- carbinol, and 1,1,2-triphenylethane-i,2-diolwere not pothesisthat benzophenoneand other products contain- produced in detectablequantity. The observation that ing diphenylmethyl moieties might be derived in some triphenylcarbinol is not a reaction product is of particu- fashion from dilithium benzophenone dianion (18), lar mechanistic significance. Since phenyllithium re- acts readily with benzophenone, yielding triphenyl- (PhLi), ---g-----.. o l- o l'?- carbinol, under the conditions of these reactions, the :r,t* --r* absenceof triphenylcarbinol as a product indicates that ^, . . co -, Ji. PhL, I ll | -r'i* PhLi ---- PhCLi : -- [encnnJ benzophenonedoes not appeor in the reaction mixture until phenyllithium has been completely consumed. l8 o This conclusion was strengthened and extended by l- o l- tl I ll I u*-:+ PhCPh determining indirectly the quantity of benzophenone Lpn('pr'l present in solution on completion of the reaction be- I tween phenyllithium and carbon monoxide before l9 hydrolysis. An aliquot of a reaction solution that had which forms deep red-purple solutions in ether solu- been allowed to react to completion with carbon tion, or from lithium benzophenone ketyl (19). In monoxide at room temperature was treated with lithium principle, carbon monoxide insertion into a phenyl- aluminum hydride, then hydrolyzed. The yield of lithium dimer would yield 18; alternatively, carbon benzophenoneobserved in the hydrolyzed solution was monoxide insertion into phenyllithium monomer could indistinguishablefrom that observedin a secondaliquot yield a transitory benzoyllithium species which could of the original reaction mixture that had not been be converted into 18 by reaction with a second molecule treated with lithium aluminum hydride. Since anv of phenyllithium.s2 One-etectronoxidation could con- o vert 18 to 19 and ultimately to benzophenone. In order to test this hypothesis, it was clearly necessaryto r. LiArHi ll PhCPh (490/0)* others have available authentic samples of 18 and 19. A1- / both 18 and 1,9have been thoroughly studied in " "i- o though other circumstances,the techniquesused for preparation \ll for their characterization and L"'* PhCPh (480/0)* others of these species, and identification under the reaction conditions encountered benzophenone present before hydrolysis would have in these studies, were not entirely straightforward and been reduced to benzhydrol by LiAlH4, this experiment are outlined here. establishesthat the benzophenoneobserved as a product Solutions of dilithium benzophenone dianion in in these reactions is not present in the reaction mixture DME were preparedby reduction of benzophenonewith before hydrolysis. an excessof lithium dispersion.33,34Solutions of 19 One additional useful datum can be derived from in DME were obtained by reduction of I equiv of studiesof the products of the reaction; uiz, since 1,2,2- benzophenonewith I equiv of lithium dispersion,35-38 triphenyl-1-acetoxyethylene (10) replaces 2 in room or by reaction of benzpinacol with excessmethyl- or temperature reaction mixtures quenched with acetic an- phenyllithium. Preparation of these materials in di- hydride, the immediate precursor of 2 before hydrolysis ethyl ether solution were carried out using similar is probably the corresponding lithium enolate 11. (32) The state of aggregationof phenyllithium under the conditions of these experiments is unknown. Vapor-phase osmometry indicates Ph OAc Ph OLi Phrl that halide-freephenyllithium is dimeric in ether [P. West and R. Waack, ,\c,O \ ,/ g \"u-!" J. Amer. Chem.Soc., 89, 4395(1967); seealso G. E. Hartwell and A. /-\ Allerhand, ibid.,93, 4415(1971)l; lithium nmr studiessuggest that it is ,}:( ,/\ \ /\ monomeric [J. A. Ladd and J. Parker, J. Organometal. Chem., 28, I Ph Ph Ph Ph pi, Ph (1971)1. Regardless,tithium halides and lithium ethoxides are in- t0 77 corporatedinto alkyllithium clusters: L. M. Seitzand T. L. Brown, 2 J. Amer. Chem.Soc., 88, 2174(1966); D. P. Novak and T. L. Brown, ibid.,94,3793(1972). Similarly, the observation that 1, 6, and 7 are replaced (33) A. Le Berreand P. Goasguen,Bull. Soc.Chim. Fr.,1838 (1963). by 11 on treatment of product mixtures obtained at low (34) For alternativeprocedures, see (a) S. Selmanand J. F. Eastham, temperature suggeststhat these materials share a com- J. Org. Chem.,30,3804(1965); (b) E.L. Andersonand J. E. Casey,Jr', ibid.,30,3959 (1965),and rcferencescited in each. A lessconvenient mon precursor. Confirming this suggestionand identi- process,based on experimentsdescribed by A. Maercker arld J. D. fying the structure of the precursor is one focus of the Roberts lJ. Amer. Chem. Soc,, 88, 1742 (1966)1,involves reaction of work that follows. dipotassium benzophenonedianion with ar-rhydrouslithium bromide. (35) W. Schterrkand A. Thal, Chem.Ber.,46,2840 (1913). Dilithium BenzophenoneDianion (18) and Lithium (36) W. E. Bachmann,J: Amer. Chem.Soc., 55, 1179(1933). BenzophenoneKetyt (19). The first substantialevidence (37) N. Hirota in "Radical lons," E. T. I{aiser and L. I{evan, Ed., Interscience,New York, N. Y., 1968,p 53 ff. concerning the nature of the intermediates in the reac- (38) B. J. McClelland,Chem. Reu.,64,301 (1964); see also M. tion of phenyllithium and carbon monoxide emerged Szwarc,Progr. Phys. Org. Chem.,6,323(1968).

Whitesides, et al. I Mechanism of Reaction of Carbon Monoxide with Phenyllithium 8122 TableII. RepresentativeAnalyses of Product MixturesObtained on Oxidation or Hydrolysis of Solutionsof (PhrCO)-M + and (PhzCO)z-ZM

Method of --Yield," 7o- Compound Concn, M Solvent preparation Quenching agent PhrCO Ph,CHOH Ph2coli2 0.067 DME aq KOH' I 90 Os 90 0 PhrcoLi 0.022 DME aq KOH. 47 49 Or 100 0 Ph2coLi' 0.006s Et20 HrO 0 70 Oz 74 (f PhrcoLi o.o12f EtrO CH3OH-KOH" 53 47 Os 100 0 PhzCONaz 0.067 EtrO aq KOH" 1 100 Oz 95 0 PhzCONa 0.20f EtzO aq KOH" 47 41 Or 9l 0 I " Yieldsare lrased on eitherstarting benzophenone or benzpinacol and areprecise to i,5'o. Preparedby reactionof PhsCOwith a dispersionof alkalimetal. ' ThepH of thesesolutions was cc. 13. d Preparedby reactionof 2 equivof methyllithiumwith I equivof benzpinacol." Saturatedmethanolic KOH. / Somefraction of theseketylsmay have been insoluble under the reaction conditions; these concentrationsshould be taken as upper limits (cf, footnote 39).

techniques,although the low solubilities of 18 and 19 the yield of benzhydrol obtained on basic hydrolysis is in this solvent make their solutions particularly sus- probably best considereda measureof the total quantity ceptibleto accidentalhydrolysis or oxidation.3e of monomericand dimeric19 rather than a direct mea- Solutions were analyzedfor 18 and 19 by taking ad- sure of monomeric l9 alone. Second,since it has not vantageof the characteristicbehavior of thesematerials been possible to prepare samples of either 18 or 19 on oxidation and hydrolysis. Both dilithium benzo- unambiguously.it has not been possibleto prove that phenone dianion and lithium benzophenoneketvl are either hydrolysis or oxidation lead quantitatioely to oxidized to benzophenoneby molecular oxygen.33'34 the indicated products. Thus the concentrations for Hydrolysis of dilithium benzophenonedianion yields 18 and 19 inferred from these analyses should be benzhydrol: h-vdrolysisof benzophenone ketyl with taken as minimunr values. Nonetheless, several ex- aqueousbase (pH > 13) yieldsa I : 1 mixture of benzo- perimental observattonssuggest that determination of phenone and benzhydrol. Thus, in principle. com- the relative,vields of benzophenoneand benzhydrol ob- parison of the yields of benzophenoneand benzhydrol tained following hydrolysisand oxidation of a solution obtained on oxidation and hvdrolysis of aliq-uotsof a containingl8 and 19 doesin fact provide a reasonably solutioncontaining 18 and 19 shouldprovide a straight' accuratemethrtd for determiningthe concentrationsofl tT- thesecompounds. First, the yields of benzophenone I,hzco .o' ,* pnrcHoH and benzhl'droiobserved on oxidation or hydrolysisof O: lI:C) solutions oi' l8 and 19 preparedby reduction of benzo- PhrCO<- 19,u*-"r O.5PhrCO+ 0.5PhrCHOH phenonewith appropriatequantities of lithium metal accoLrnr.for 85 100% of the starting material (Table forward method of deterntining the concentrationsof II). Further. the productbalances obtained on hydrol- thesespecies: the yield of benzophenoneobtained on ysis anci oxidation of aiiquots of a common solution oxidation of the solution would rellect the combined agreewell. Tirus. it appearsthat both reactionstake quantities of 18 and 19 present; the yield of benz- placein high vield. Second,the ratio of benzhydrol to hydrol obtained on alkaline hydrolysis would cor- benzophenonc obtained on hydrolysis of solutions respond to one-half the quantitv of 19 present. This ostensiblycontaining benzophenonedianion is (0.01, analytical technique suffers, however, from two an1- whiie this ratio for solutions ostensibly containing biguities. First, the extent to which the ketyl 19 is benzophenoneketyl rangesfrom 0.87-0.95, in reasonable dimerized in solution, particularl;,-at the relatively high agreementwith the expectedvalue of 1.00. Third, the concentrationsused in this work. is unclear;40 thus. spectroscopicexarninations described below establish that reduction oi benzophenonewith I equiv of lithium (39) Solubility of both 18 and 19 in ethereaisolvents is limited: 18 ketyl appearsto precipitatefrom solutionsof DME at ambicnt temperatures yields a solution which contains benzophenone at corlcentrations;0.067 M, and from soiutionsof EtrO at concerltra- but little if any benzophenonedianion and that con- tions 10.0065 M; thelimiting soiubility of l9 in DME is ? 0.06 M,' tinued reduction using an excessof lithium yields solu- in Etro it is t 0.003M. (40) Although I9 is apparently monomeric in dioxerr-resolution at tions containing benzophenonedianion but little benzo- 10-3-10-4 M concentrations,atits visible absorption in dicthyl cther phenone ketyl. Taken together, these results indicate showssignificant cieviations from Beer'slaw.a: No benzpinacolis ob- reduction oi even high concentrationsof benzo- servedon hydrolysis of l9 under basic conditions,in trgrecmcntwith that earlier work;35'3{'nolletheless, because be.zpirlacol is rapidly cleaveci phenonewith appropriatequantities of lithium metal in to benzophelloneand benzhydrolby base,this obscrvatior.ris not use- yields solutions containing 18 and 19 36,38 etherealsolvents ful in determiningan equilibriurnconstant for dimer formation. quantities (41) H. V. Carter,B. J. McClelland,and E. Warhurst,Trans.Faraday in good yield, uncontaminatedby important Soc.,56,455(1960); for relatedstudies ofsodium benzophenoneketyl, of other materials, and that comparison of the yields seeD. J. Morantzand E. Warhurst,ibid.,5l ,l375 (1955);N. Hirota -l0r and S. I. Weissman, J. Amer. Chem.Soc., 86, 2538 (1964), Seealso D. equilibrium corlstitntfcrr dimerizationin the latter solvent is l. H. Eargle,Jr., and R. Emrich,J. Org. Chem.,35,3744(1970), for an mol-r: J. Garst, D. Walmslcy,-C. Hcwitt, W. Richards, and E. ir-rterpretationof the ir spectrumof l9 in DME. Zabolotny, ibid., 86, 412 (1964). Magnetic susceptibilitymeasurements (42) H. E. Bent and A. J. Harrison,J. Anter. Chem.Soc.,66,969 inclicatethat sodium benzophenoneketyl is >99f associatedto a (1944). The visible spectrum of sodium benzopherroneketyl obcys diamagneticform in benzcne: R. N. Doescherand G. W. Wheland, Beer'slaw at -10-{ M in DME but deviatesin diethyl ether; the ibid..s6,201 I (1934).

Journal of the American Chemical Society I 95:24 f Nouember 28, 1973 8123 Table III. Absorption Maxima and Extinction Coefficientsfor OrganometallicSpecies

Compound Method of preparation Solvent" ,\ma*, flflt o ,/\ Phcoc(oLiFCPh, (1s) Ph,C--.--CHCOPh + LiTMP EtrO 417 DME 437 PhrC:C(OLi)Ph (17) PhrEC(Oec)Ph + PhLi" DME 288 440,000 Phecoli, (18) Phrco + Li EtrO 494 25,000 DME 535 Phrcol-i (19) (PhrCOH), + PhLi" EtzO 600 DME 634 Phc(oli):c(oli)c(cH3)3 (20) PhCOCOC(CH')' f excesslithium DME 300 Ph3CLi Ph3cH + PhLi" DME 435 , " Solutions were 0.1 N in phenyllithium. The literature valuesfor the extinction coefficientof trityllithium in letrahydrofuran is log G 3.97(425 nm)}t " The phenyllithium usedwas that presentin the solvent.

of benzophenone and benzhydrol obtained on oxida- tion and hydrolysis of these solutions provides a practical method of analyzing for these substances. Spectroscopic Studies. The procedures described were used to prepare samplesof L8 and L9 whose uv and visible spectracould be compared with those from solu- o phenyllithium o tions obtained on reaction of with carbon 0 o monoxide. The spectra of 18 and 19 are strongly o dependent on concentration, on the nature and concen- ! tration of associatedmetal ions, and on the polarity of the solvent.sT'38'42'43In order to obtain spectra of 18 and 19 that could be compared directly with spectra taken in the initial stagesof reactions between phenyl- lithium and carbon monoxide, we carried out spectro- scopic studies of authentic solutions of 18 and 19 in

ethereal solutions containing approximately 0.1 M A'nm phenyllithium. The use of phenyllithium solution as Figure1. Ultraviolet-visiblespectra of organometallicspecies in solvent has the , additional practical advantage of DME solution: (a) the reactionmixture obtainedby adding minimiiing the effectsof adventious traces of oxygen or phenyllithiumto an excessof carbonmonoxide dissolved in DME; water on the spectraof the benzophenoneketyls and di- (b) thereaction mixture obtained by addingcarbon monoxide to a anions by acting as a scavenger for materials which large excessof phenyllithiumdissolved in DME; (c) PhrCLi; (1s)PhCOC(OLi):6phz: Ph'C:C(OLi)Ph;(20) PhC(OLi): would react with these organometallic reagents. The Gi) c(oli)c(cH,)a; (18) Phzcolir; (19) Ph,coLi. Sinceonly the spectrum of phenyllithium itself does not interfere with absorptionmaxima and peak shapes could be determinedwith any those of the compounds of interest at wavelengths accuracyfrom thesespectra, they are not drawnto scale,and the longer than approximately 300 nm. Visible spectra for absorbancesof the various spectra are arbitrary. dilithium benzophenone dianion and lithium benzo- phenone ketyl in 0.1 M solutions of phenyllithium in tinction coefficients (when estimated) obtained from given DME are in Figure 1. The spectra of 18 were thesespectra are summarizedin Table III. obtained with solutions prepared by dilution of con- These spectra were compared with spectra of solu- -10-3 of centrated stock solution to M; the spectra tions obtained by reaction of phenyllithium and carbon 19 were obtained from solutions prepared by a similar monoxide under conditions approximating those ob- dilution procedure or directly by reaction of benz- taining during the product studies summarized in pinacol with phenyllithium solution in the cuvette. Table I. Addition of 0.16 pmol of carbon monoxide to gives For comparison, this figure also spectra of several 4 nl of 0.1 N etherealphenyllithium solution in a Pyrex compounds that either are themselvesformed in signifi- cuvette at room temperature produced the red colora- (the cant co.ncentration during the reaction lithium tion characteristic of the initial stagesof the large scale (17) enolates of a,a-diphenylacetophenone and 1,1,3- reactions(Figure 2, curve b); a preciselyanalogous re- triphenylpropane-1,2-dione(15)) or that are spectro- action carried out in DME yielded the spectrumshown scopic models for compounds that are formed during in Figure 1, curve b. The similarity of the peak shapes (the the reaction dilithium dianion of l-phenyl-3,3-di- and absorption maxima of the light absorbing material methylbutane-I,2-dione (20) which is used as a model produced in these experiments and of authentic di- for the trilithium salt of 8, Ph2c(oli)c(oli):c(oli)- lithium benzophenone dianion 18 strongly suggests Ph (22) (uide inJra)).aa Absorption maxima and ex- that the compound formed initially in the reaction of

(43) For a study of the influenceof lithium bromide on the spectrum lithium dispersionin DME (se€the ExperimentalSection). Compound of lithium benzophenoneketyl, see D. G. Powell and E. Warhurst, l5 was obtained by reaction of 1,1-diphenyl-2-benzoylethylencoxide Trqns.Faraday Soc.,58,953 (1962). with lithium 2,2,6,6-tetramethylpiperidide(LiTMP).nr The spectrum (44) Compound 17 was prepared by reaction of 1,2,2-triphenyl-l- of lithium tripher-rylmethide,prepareC according to the procedure of acetoxyethylene(10) with phenyllithium in the cuvette. Lithium 1,1- R. Waack and M. Doran, J. Amer. Chem.Soc.,85, 1651 (1963), is also diphenyl ethoxide formed as the other product of this reaction is trans- includedin Figure l. parent in the spectral region of interest. Compound 20 was prepared (45) R. A. Olofsonand C. M. Dougherty,J.Amer. Chem.Soc.,95, by reduction of 1-phenyl-3-3-dimethylbutane-2,3-dione(21) with excess 583(1973).

iAhitesides, et al. I Mechanism of Reaction of Carbon Monoxide with Phenyllithium 8124

experiments is that dilithium benzophenonedianion 18 is formeci at the outset of the reaction of phenyllithium i with carbon monoxide. Lithiurn benzophenone ketyl a.rL may be present in the reaction mixture late in the re- action. Reactions of Dilithium Benzophenone Dianion with Carbon Monoxide in the Presence of Phenyllithium. Although these spectroscopic studies implicate di- lithium benzophenone dianion as the principal sub- stance generatedin the initial stageof these reactions, it remained to establish the role of this compound in the formation of the products actually isolated as their conclusiorr. In particuiar, it was important to know whether th e lithium enolateof cc,a-diphenyl ac':tophenone (17) observed as the major reaction product at room temperature was derived from 18 or from pirenyllithium by sorne reaction path not involving 18 and to deter- Figure2. L/ltravioiet-visiblespectra of organometallic speciesin mine whether 18 could plausibly give rise to benzo- diethylether solution: (a)the reaction mixture obtained by adding phen-vllithrumto an excessof carbonmonoxide dissoived in ether; phenone, the major iow-temperature product, under the (b) thereaction mixture obta.ined by addingcarbori monoxide to a reaction conditions usedin this work. largeexcess of phenS,llithiumdissolved in ether;(15) IThCOC{OLi)-= It was possible to establish by simpie labeling experi- CPiiz;(18) ['hrCOI-ir;(t9) PhzCOl-i.Tlie absorlrancesof the ments that the benzophenonemoiety spectraare arbitrar,"-. of benzophenone dianion was incorporated into both high- and iow- temperature products on reaction with phenyllithium phenyllitirium and carbon monoxide is in tact diiithium and carbon monoxide, For this purpose.we usecl4,4'- (23) Lrenzophenonedianion; none of the di-terr-butyli:enzophenone as the latreled benzo- eithercomrrounds 2F{- 1aC-labeled examined (1.5, 17, or 20) overlapped ttre absorption ohenone rather than or material, be- region for this clianionsufficientiv to hamper its identifi- cause thc dilithium dianion, 24., of this substituted cation. Comparisonof the absorptivity observeciin" the L,enzophenonewas lnuch more soiuble in ethereal sol- soiution resulring frorn reacrion or phenyliithium and vents,and henceeasier to prepareand manipuiate,than carbon monoxide wit!: that estimatedi'rom the extinc- was dilithiunr benzophenonedianion itself. Reaction tion coefficientof 18, determineciindepencjentil', and the of 7A with carLronrnonoxide in the preseneer'f a tenfoid quantity of carbon monoxicle used in the reaction ir-l- excesscrf phenyllithiuni. foiiowed by queni:rr';lg of the dicates that the amount of, lB fbrmecl acoounts for reaction rnrxture with acetic anhydride, leci to sub- .'*,50%of the carbon monoxide adcleo. stantial incorporations (-60 %) of the iabeled benzo- phenone ln an effort to generate spectra characteristic ol" thc moiety into the products 25 and 26, the ex- pected yields cornpoundspresent late in tire reaction betweencari:on anaiogsof 10 and 1L;an of 10, 11, 25, and 26 monoxide and phenyllithiurn, a small ciuantirv of obtained in severai reactions are summarized in phenyliithiun,was added to cuvetres conrainingtiMl *Ph-- or ether saturated with carbon monoxide (F'igure j., tcH,,t,c{==-),- - \_l curve a and Figure 2" curve a, respectively)" Tl.re r' co Et2o> maxinra of two prominent long waveiengtir oands ob- i*ph.col'-2i,i- + 10phLi served in the visible spectrum in ether 2. Ac"O corresponct 24 rougirly to those of 18 and 1g; under thesecondiiions 0 spectra characteristic of L7 and ZZ were *pi, otrscured bv "Ph Pi' overlap and possibly by other speciesin solution. The \_r,, .\/ spectrum in DMF. is not simply interpreted. Estirnates + +lo+ll *Phl-i *Ph/<\ of the conversionsof phenyllithium lB and 19 under OAc oA. these conditions, if indeed these were the compounds 25 26 responsiblein the region between 450 and 650 nrn, Table IV. An authentic sample of 26 was prepared would not be significanr, sinceimpurities present in the by substitution of 4,4'-di-terr-butylbenzophenone ether soiution before addition of phenyllithiurn de- for benzophenon€in the proceduressummarized stroyed appreciable quantities of the initially formed in Scheme I; authentic 25 was svnthesizedby the procedure organometallic materials. Control experiments estab- out- lined in SchemeIl. lished that neither l8 nor 19 reacted with phenyrlithiurn in theseor more concentratedsoiutions" (461 To establish that To lend further supporr theseresults were independent of the method to the contention that 1g is of labeiing the reaction partners, similar:although lessaccurate studies present in reacting mixtures of phenyllithium ancl were carried out involving ethereal solutions containing 1 equiv of carbon monoxide, a I M solution of phenyllithiurn dilithium benzophenonedianion and 5 equiv of p-ethylphenyllithium. was Exposure stirred ol' these solutions to carbon monoxidc, followed by hy- under I atm of carbon monoxide for 20 min and drolysis, led to a,a-diphenyl-p-ethylacetophenone(iii) in 24f yield, then quenctredby addition of NaOD in DzO. Analysis t'nr tt'9 by glpc and mass spectral analysisafter the usual work- [PheC:O]r-21-i+ * SLiPh-p-E, up indicated the presence of 0.8 /" benzhydrotr,con- PhTCHCOPh-p-Et * other products taining >0.8 deuterium atoms per molecule. iii The important conclusion from these spectroscopic basedon 18"

Journal of the American chemical society I 95:24 I Nouember 2g, IgT3 8125 Table IV. ProductsObserved Following an Acetic Anhydride olioo Quench of Product Mixtures from Reactionsof -Di-terl-butylbenzophenone (24), riltill Dilithium 4,4' Dianion *Ph"cli+c+C+ [,iPh - Phenvllithium. and Carbon Monoxide 24 PhLi A (concn, (concn, Temp, --yislfl,o /- oli oliol-i OHOH O OO oc M) M) 10 11 25 26 tlt H'o. | | ll -',, llll *Phrc-c:cPh *Phi-cu-CPh - -Ph,CHCCPh 0.50 0 25 39 14 00 0.25 0.05 25 32 16 7 (36)b 28 0.50 0.05 25 34 16 6 (62\b Trace a:B 0.50 0 -78 060 00 ,.nI AcO O o.25 0.05 -78 044 o 4 (z0)b f Aco oAco-lcl *Ph,c:c-cPhlll 0.50 0.05 -78 047 0 6 (59)b L- | \l I l) | + l*Ph.c-c:cPh 26 " Yields were determinedby glpc and are basedon phenyllithium l unlessnoted. b Yield basedon24. 29 inally present in24, would in turn yield 26. Since the SchemeII. Synthesisof 25 acetoxy group required to leave in this last step is 1. PhMeCl 1 CH?CO1[I activated toward heterolysis by two adjacent phenyl *Ph,co. ., - - 2. SOCI2, Pyr 2. HN{PA, A groups and one oxygen-substituted B-styryl moiety, ' 79o/o 36% sufficiently rapid conversion of 29 to 26 to preclude iso- lation of 29 would not be surprising; however,alterna- o *Ph oAc ,, tive schemesfor conversionof 28 to 26 involving, €.8., *Ph:CHCPh^^^------'ll l[(cH3)2cH]2NLi \ / - >< monoacylation of the oxygen derived from 27 followed 2.Ac"O / \ *Ph 28o^ Ph by rapid loss of acetateion, can of course also be writ- ten. We have no direct evidence for 29 or related -, :. 25 speciesin these reactions,but the postulation of 28 as A-plausible reaction sequencefor the transformation an intermediate is immediately compatible with the of 24 to 25 involves initial transformation to an inter- formation of compounds l, 6, 7, and 8 as hydrolysis mediate 21, either by insertion of carbon monoxide into products from the reaction of phenyllithium and carbon a mixed organolithium cluster containing units of both monoxide by protonation, dehydration, and reverse phenyllithium and dianion 24in a manner analogousto aldol reactions,starting from the analogous triamon 22. that suggestedfor formation of 18, by nucleophilic addi- The differencein the observedvields of 1 and 6 from tion of 24 to an initially formed benzoyllithium, or by ot,i ot,i ol.i nucleophilic reaction of phenyllithium with an adduct lll of ?A and carbon monoxide.aT Elimination of the ele- 2PhLi + :lco + I,h_(._(':cPh lithium ments of lithium oxide from 27 would vield the 22 enolate of 25.4e +lH'o o]-i O AoLioli \t H,O OH O- ,.OH OH O;H I lll I -"Li \-l *Phcl,i * *Ph-l ,o" rr i ii) I 11 + c * LiPh -{ry-CPh -- \i-Ph PhC Ph:c1!:cPh I *Ph "Ph Li 8 -u'n 24 27 tf | *tn..t": *on)a: 'tl I O OHOH :t rJi------> ' ,zoA'' oHo *ph.' "aott- .rn ph llll til ,n Ph,C+ CH:CPh PhrC:C -CPh ;- It I tl The mechanismfor conversion of the diarylmethylene tl {l o oo moiety of ?A to 26 is clearly more complex than that for tl llll the transformation of 24 to 25. Combination of I HOCH:CPh PhrcHCCPh equiv of dianion?A,2 equiv of carbon monoxide, and 1 ., 6 t equiv of phenyllithium in a process analogous to that required to form 27 would generatethe trianion 28, the the reaction presumablyreflects the sensitivity of 6 to labeled analog of 22. Acylation of this substancewith the conditions encounteredduring the hydrolytic work- acetic anhydride, followed by loss of the oxygen orig- up. Thus, examinationof productsderived from reactions (47) The ability of disodium benzophenonedianion to act as a nucleophiletoward carbon is well established'3aa'48ths reactivity of 18 of the labeled dilithium diaryl ketone dianion 24 with should be similar. phenyllithium and carbon monoxide fully supports the (48) H. E.Zaugs and R. J. Michaels,J. Org. Chem.,33,2167(1968). 18 is an inter- (49) Attcmpts to preparean authenticsamplc of 27 werenot success- hypothesisthat dilithium benzophenone ful. Nonetheless, reduction of a,cr-diphenyl-a-hyclroxyacetophenone mediate in the reaction of phenyllithium and carbon (4) with excessdilithium benzophenonedianion did yielcl 2 (-70%) monoxide and providesa unifyingmechanistic rational- after hydrolysis. Thus it apparently is possibleto eliminatc lithium oxide from a speciesresembling 27 under reducing conditions. A ization for most ot' the products of this reaction based similar eliminationof lithium oxide is proposcdto occur during forma- on conversionof 18 to intermediateshaving structures tion of 1,2-di-tert-butylcthylenefrom the reaction of /erl-butyllithium t7 and 22. with terf-butylethyleneoxicle; cf. I. K. Crandall and L. H. C. Lin, J. Amer. Chem.Soc.. 89. 452'l(1967). Oxidation of Dilithium Benzophenoneand Lithium

Whitesides.et al. I Mechanisntof Reactionof'Carbon Monoxide with Phenyllithium 8126

Benzophenone Ketyl by Carbon Monoxide. Benz- Qualitatively similar conclusions were reached on pinacol (5) and benzil (3) are the sole isolatedproducts examination of the changesinduced in the visible spectra of the reaction of phenyllithium and carbon monoxide of solutionsof 18 in ether and DME following exposure not easily generated from l7 or 22. One possible to carbon monoxide. In DME, this exposure results rationalization for the formation of 5 rests on the in disappearanceof the band due to 18 and appearance demonstration that carbon monoxide is a sufhciently of a band at 634 nm attributable to 19; in ether, addi- strong oxidizing agentto convert benzophenonedianion tion of carbon monoxide results in complete bleaching, to benzophenone ketyl (and benzophenone ketyl ulti- with the only new peak appearingat -350 nm (benzo- mately to benzophenone). Carbon monoxide is an ef- phenone). fective one-electronoxidant toward a variety of aro- Reactions with Other Organometallic Reagents. We matic radical ions and dianions,s0toward metallicpotas- have examined briefly the reactivity of several other sium,sr and toward solutions of potassium in am- types of organometallicreagents toward carbon mon- monia,52yielding in each case substancesthat can be oxide. These experiments, together with results derived from a hypothetical "alkali metal carbonyl" obtained by others, indicate qualitatively that only M+CO.-. derivatives of strongly basic anions will react with Chemical and spectroscopicdata demonstrate that carbon monoxide under conditions similar to those carbon monoxide is also capable of oxidizing 18 and used in the work describedin this paper. Thus, /ert- 19 the nature of the productsderived from the carbon butyllithiuffi, 2I phenyllithium, n-butyllithium,r,amethyl- monoxide was not establishedin these experiments. lithium, trimethylsilyllithium,22 and cyclopropyl- Thus, Table V summarizesthe yields of benzophenone lithiums3readily absorb carbon monoxide, as do dimsyl sodium and several lithium amides;23 the less basic lithium phenylacetylide, lithium n-butylacetylide, TableV. Yields of Benzophenoneand BenzyhydrolObtained lith- by Oxidationof Dilithium BenzophenoneDianion (18) and ium diphenylamide, lithium cyclopentadienide,l3and Lithium BenzophenoneKetyl (19)with Carbon Monoxide, lithium diphenyl phosphideappear not to react. These Followedbv Hvdrolvsis" data suggest that in classifying anions according to Compound Product(%vield) their reactivity toward carbon monoxide it may be (concn,M) Solvent Ph2CO PhZCHOH heuristically valuable to consider these reactions as proceeding through acyl (30) 18 (0 06) DME 50 45 anions whether or not 18 (0.004) EtzO 20 6 le (0.05) DME 52 38 o 19 (0.003) EtzO 95 0 lll [ fl] B-+('.- ' Yields are basedon the initial concentrationof 18 or 19. The Lu/ ) fate of the organic moieties not detectedas benzophenoneor 30 benzhydrolin theseexperiments was not established. substancesof this classare true intermediates. Values for pKr,'s of speciescontaining the acyl anion moiety and benzhydrol observedfollowing hydrolysis of solu- are not well-established,although that of CHaOzC- tions of 18 and 19 that had been allowed to reacr ro (30, B : OCH3) has been estimatedto be ca. 35;sr this value seemsconsistent with what little is known l9 about the chemistry of acyl anions. These organo- "5^'N,. lithium reagents that have so far been found to react with carbon monoxide all are derivatives of anions l8 I having pKr's of this magnitude or greater, while those \o t, co. e/ 0,,\ that have proved unreactiveare characterizedby pKu's l,,o that are smaller.ss Thus, it may prove possibleto pre- l9 dict the facility of addition of organic anions to carbon completion with carbon monoxide under conditions monoxide on the basis of their basicitv relative to that approximatingthose utilized for the reactionsof carbon of the appropriate acyl anion 30. monoxide with phenyllithium. The observationof ap- Discussion proximately equal yields of benzophenoneand benz- hydrol, characteristicof 19, after reaction and hydrol- Three principal kinds of data bear on the mechanism ysis of either 18 or 19 with carbon monoxide in DME of the reaction of phenyllithium with carbon monoxide. contrastswith the detectionof benzophenonealone after First, spectroscopic studies indicate that dilithium reaction and hydrolysis of these substancesin diethyl benzophenonedianion is formed in the early stagesof ether. The inference from these experiments is that the reaction. Second,examination of the products of carbon monoxide will oxidize 18 to 19 in either DME reactionsof carbon monoxide with a mixture of labeled or ether but will oxidize 19 to benzophenoneat a signifi- (53) E. G. I(elly, Ph.D. Thesis,Massachusetts lnstitute of Technology, cant rate only in ether; this inferenceis in qualitative Cambridge, Mass,, 1968. The major products from the reaction of accord with the supposition that lg should be less ,l-butyllithiumwith carbor-rmonoxicle in hexaneresemble those produced from phenyllithium, riz. dibutyl (3-071, stronglyreducing in the more stronglysolvating DME ketone 6-butyldecan-5-one (60-70\), and 7-butylundccanc-5,6-dione (l 0 %). than in the lessbasic ether. (54) K. P. Butin, I. P. Beietskaya,A. N. I(ashin, ar-rdO. A. Reurov, J. Organometal.Chem., 10, 197 (1967). (50) W. Blichner,Chem. 9er.,99,1485 (1966). (55) For discussionsof the pK,,'sof weak organic acids, seeH. F. (51) W. Btichnerand E. A. C. LLrckcn,Helc. Chim.Actu,47,2l13 Ebel in "Methodcn dcr OrganischenChemie," (Houben-Weyl),Vol. (1964); W. F. Serger,A. Fatiadi, P. C. Parks,D. G. White. anclT. p. l3/1, 4th cd, GcorgThieme Verl;rg, Stuttgarr, 1970, p 27 tr: J. R. Jones, Perros, J.Inorg. Nucl.Chem.,25,lS7 (1963). Quart.Rec., Chem. Soc., 25,365 (1971): D. J. Cram, "Furrdamentals (52) . W. Blichner,Hela. Chim. Acto, 46,2111(1963), and refcrc'ccs of CarbanionChemistry," Academic Press, New York. N. Y.. 1965. citedthcreir-r. ChapterI.

Journal of the Anterican Chemical Societlt I 95:24 I Nouember 2g. 1973 8127 dilithium benzophenone dianion and phenyllithium and one another at rates that are comparable with or establishes that the diarylmethylene moiety of the faster than the initial rate of formation of dilithium former is effectively incorporated into products char- benzophenonedianion 18. In particular, our inability acteristic of the reaction of carbon monoxide with to generatemore than traces (ca. I f) of benzhydroi on phenyllithium. Third, examination of the influence of hydrolysis of reactive mixtures that have absorbed only temperature and reaction composition on the distribu- small quantities of carbon monoxide indicates that the tion of products indicates that at. least two related concentration of 18 is never high and that its subsequent processes compete in these reaction mixtures: one! reactions with carbon monoxide and phenyllithiurn (or taking place at room temperature, converts 3 equiv benzoyllithiurn, whichever is actually involved) are of phenyllithiurn and 2 equiv of carbon monoxide tcr fasterthan its formation. Further, sincethe particular the lithium enolate of a,o-diphenyiacetophenonet17); combinations of benzophenone dianion, carbon a second, dominating at -78o, involves 3 equiv of rnonoxide, and phenyllithium shown in SchemeIII are phenyllithium and 3 equiv of' carbon monoxide and certainly not the only ones that can be written, the low generatesthe trilithium trianion 72. The products of product yields observed under many conditions may the reaction isolated following quenching with water or result i-rom condensations of these materials in dii- acetic anhydride are derived in straigtrtforwarci ways ferent combinations to yield higher molecuiar weight from 11 and22. The mechanismimplied by thesedata products" The factors influencing the partitioning of is summarized in SchemeIII, together with specuiations intermeciiatesbetween the reaction paths so far identt- fieci have not been established. One attempt to test SchemeII[. SummarizingReactrr:n Sequence fbr tl"reConversron of Phenyllithiumand Carbon N{r.rnoxide to Producis the hypothesis that the relative importance of the low- and high-temperature paths for the reaction reflected PhLi =* (PhLi), the concentration of carbon monoxide in solution by \co studying the reaction at high carbon monoxide pressure ro f! was f,rustratedby a drop in the prodtrct balance; further LiO--\^ Ph iti lr ."./ ? !tl + studieshave not beencarried out" l:i --ot-; li Ph't Lr"-t' -.1 ici) T'ire most important mechanistic question un- iph LUI still -r,'lr resolvedconcerning this reaction is the rmportance of Y I I i o\,..".-Fh I benzoyllithiurn in these transformations" Quaiitative t i t\o jllli F'hi'ft( ii i correiatron of' the reactivill, toward carbon rnonoxide ! nnit ''1' i /' v cri a nrrm"nerof" cirganolithium reagents with theii Y -z'u ,, 'rl' -// basicitvis comoatible with, but does not demand, pro- Lio oli i-o LiO t)L,i duction oi transit.oryacyllithium compounds as the 0Li lil i- li i tl / iii I '"-"-pl,t; Phr.i .lI:1 first intermediatein thesereactions; the ptrausibilityot PhzC-C:C +--'zco IC | -" Pir',C--Cf'h | ct) previous \ I thesemateriais as intermediatesis increasedby Ph Phj li LPh lf,i work in which thelr have also been implicate6[.e1'2'i 22 /l IB / Nonetheless. tirere is no unambiguous methocl of' -u.:ti l-r. / r,t: ? 1,.,- distinguishing between,e.9., conversion o'i monomenc t' ./l * l.-1"- F! phenyllithiurn to dilithium benzophenonedianion L8 r\I: \/ri0[l ,] o i" 1lo i Pil olr (i i[l b-v wav of benzovllithium and conversion of dimeric ti tti lil\t'l//! It:i {ltl iLi*Li* C:g Ph,{.;-_i--'.Ph,{.;-_i-l phenvllithium to 18 by a concertedinsertion of carbon Ir\l i !:{1. t\ - 'r. | l''\/ \ i / monoxide into ari aggregate;similarly, there is presently Ph Ph Ph_lPh ! Ph Ph Pir LPh no way of deciding whether or not t.ransformationsof I l9 17 18 to 22 or l7 involve benzoyllithium. o l*,u I H.O ll Y v- In a more generalvern" the reaction of organic anions OFi OH Ha with carbon rnonoxidewould seemto hold promlse as a t// method of generatrnga. number of unusual types of bH,oH I I Ph,c-c Ph:c-cPh: -\ organic anions--especialiydilittrium dianions of sym- J Ph metrical ketones and possibly acyllithium reagents- oPh , \/ provided that ways can be devised to inhibit the con- C-C /\ densation reactions that lead to formation of the PhrcH o mechanistically significant but synthetically unappeal- 7 ing types of products encounteredin thesestudies. HO Ph \/ CH-C] Experimental Section /\ GeneralMethods. Reactionsinvolving organometallic reagents Ph,C O werecarried out usingstandard techniques for manipulationof OH oxygenand water-sensitivecompounds.so Melting points were 8 determinedusing a Ttromas-Hoovercapillary melting point ap- paratus and are uncorrected. Boiling points are uncorrected. T-60 concerningthe derivation of certain of the minor prod- Nmr spectra were run on Varian 4-60 or Varian spectrom- eters; chemical shifts are reported in pnm downfield from TMS and ucts and the structures of undetectedintermediates be- coupling constants in Hz. Infrared spectra were taken in sodium tween establishedstructures. chloride cells using a Perkin-Elmer Model 2378 grating spectro- The complexity of this mechanism, and of the result- photometer" Llitraviolet spectra were determrned on a Cary 14 ing product mixtures, suggeststhat many of the various intermediates formed during reaction the course of the (56) D. F. Shriver,"The Manipulationof Air-sensitiveCompounds," take part in further reactions with carbon monoxide McGraw-Hill, New York, N. Y., 1969,Chapter 7.

Whitesides, et al. f Mechanism of Reaction of Carbon Monoxide with Phenyllithium 8128 spectrophotometer. Mass spectrawere determinedon a Hitachi- componentusing preparativetlc coatedwith 2 mm of silica gel pF- Perkin-ElmerRMU-6f) or RMU-6E mass spectrometer. Glpc 254,using 2 : I methylenechloride : benzeneas an eluentafforded analyses were performed gl0 the on F & M Model instruments desiredproduct as an oil: ir (CHCIB)3030, 2900,1715,1670, 1600, ecluippedwith flameionization detectorsand Discintegrators. An- 1500.1450, 1100 cm-r; nmr (CDCla)6 7.1-8.0(m, l5), 6.0 (s, 1)j alyticalthin-layer chromatography (tlc) wasperlbrmed using Baker_ massspectrum (70 eV) (rel intensity)300 (6.4, M+). 167 (36), 105 flex preparedanalytical plates coatedwith silicagel I B-F or 2B-F. (100),77 G7); calcdmol wt for CzrHroOz300.1150, founcl 300.1149. Preparativetlc plates were either Analtech plates precoated with Hydrolysis of 17 afforded 7 in better yield. A 50-ml round- silicagel GF or home-madeplates coated pF- with Mercinitrogen atmosphereimmediateiy before tetramethylpiperidineand l5 ml of THF. The flaskwas swept t-rse; 1,2-dimethoxyethane(DME), with tetrahydrofuran(THF), and nitrogenand cooledto 0". Methyllithium(2.2 ml of a 2.4 N ether benzenewere purified by distillation under nitrogen from dark solution, 5.2 mmol) was added to the solution by syringe. The re- purple solutions or suspensions of disodium benzophenonedi- action mixture was stirredfor 10 min, and a solution of 1.5g of 14 anion. Hexanewas treated with concentratedsuifuric icid, washed (5 mmol) in 25 ml of THF was acidedby cannula. The reiulting with water, dried (caclz), and distilled from a suspensionof di- deep red solution was stirred at 0o for 30 min and l0 ml of i sodiumbenzophenone dianion undernitrogen. carbon monoxide saturatedaqueous solution of ammonium chloride was added by (c. P., Matheson)was routinely passed through a columnof calcium syringe. After the reaction mixture had stirred for l0 min. an sulfate before use. However, no significantchanges in product additional 20 ml of water was added and stirring was continuedfor yieldswere observedusing carbon monoxide pass.d over calcium an additional20 min. The layerswere separated, the aqueouslayer sulfate,passed through a solution of disodium benzophenonedi- wasextracted with 30 ml of ether, and the organic layerswere com- anion in DME, or taken directly from the cylinder without treat- bined, dried (NasSon),and concentratedby rotary evaporation. ment. Glpc analysesfor compounds t and2 werecarried out using The orangeoil remainingwas distilledthrough a short-paihstill to an 8-ft column packedwith 5% DEGs on chromosorb w or a 2-f7 yield I .I e Q.l mmol, 737i.)of I as a yellowoil, bp I 55. 5 Torr). columnpacked with 5 Carbowax gG-100 a0.l "fl 20M on meshchromo- The most convenientprocedure for the preparation of 7 involved sorb P using ,-alkanes as internal standards. Analysesof 3, 4, hydrolysisof 11, isolatedfrom an aceticanhydride quench of the and 5 werecarried out using 2-ft a column packedwiin s7 sE-30 low temperaturereaction of phenyllithiumwith carbon monoxide. on 80-100mesh chromosorb P usingeicosane as an internal stan- Typically,a solutionof 1.5g(4.4 mmol) of ll in 50 ml of methanol dard. Analysesfor 10, ll,2s, and26were carriedout with eithera was treatedwith 0.5 ml of concentratedhydrochloric acid. The 4-ft or 8-ft columnpacked with 3 g0-100 %ov-17 on meshchromo- solutionwas allowed to reflux for 4 hr and was then saturatedwith sorb using dotriacontaneas an internal Q standard. Benzhydrol sodium chloride. The organicproducts were extracted with three was analyzed on a 4-ft, 5/. carbowax 20M on chromosorb w 30-ml portions of ether. The combinedether phasewas washed column usingoctacosane as internal phenyllithium standard. was with aqueoussodium bicarbonatesolution, witer, and aqueous preparedusing standard procedures; z-butyllithiumin hexanewas sodiumchloride solurion, dried (MgSoq),and concentrated give purchased from to the Foote Mineral corp. concentrations of 1:3_3g of a crude yellow oil. Compound organolithium 7 was isolatedtn AO% reagentswere determineduiing a double titration yield from this oil by preparativetlc. method.s? organolithium reagentsolutions were discardedwhen 1,1,2-Triphenylethane-1,2-diol.a,a-Diphenyl-a-hydroxyaceto_ the concentrationof residualbase in theseanalyses exceeded 10% phenone(0.28 g,0.97 mmol) in 3 ml of ethylether was added slowly of that organolithium reagent; however, residual basetiter had with stirringto lithiumaluminum hydride (0.138 g, 3.65mmol) in l5 no obvious influence on product distributions. Microanalyses ml of ethyl ether under nitrogen at room temperature. After 15 wereperformed by Midwest Microlabs, Inc. min, the reactionwas quenchedby the slow addition of ethyl ace- a,a-Diphenyl-a-hydroxyacetophenone(4), prepared in lg ii yield tate. Severalmilliliters of ethyl ether was added, followed by accordingto the procedureof Greene g4-g5. and Zook, had mp severalmilliliters of water. The organic layer was separatedand [it.5amp 84.5-85']. dried (Na2SO,). The solventwas removed,leaving a solid which a-Hydroxyacetophenone(6), prepared in 30il yield by the pro_ was recrystallizedfrom absoluteethanol giving 0.20 g (72%) of cedureof Tsuji,sehad mp 84-85.. 1,1,2-triphenylethane-1,2-diol,having mp 165-166. [lit.or mp 1,3,3-Triphenyl-1-propene(16). Benzyltriphenylphosphonium 168'1. chloride (4.3 g, 15 mmol) in 50 ml of ether was added fo a dry, 1,2,2-Triphenyl-1-acetoxyethylene(10) was prepared from 2 by 200-ml, three-neckedflask equipped with a condenser,dropping themethod of Houseand Trost62 in 65ft yieldor in 52f yieldby a funnel, and nitrogeninlet tube. ,-Butyllithium (10 ml, t.S l,r, t5 procedureanalogous to that describedfor 2s (exceptthat LiTMp mmol) was addedslowly to the vigorouslystirred solution and the was used in place of lithium diisopropylamide): mp 104-105.: resultingorange mixture was allowed to stir at 25ofor 3 hr. To nmr (CClr) 6 7,2 (m, 15),and 1.85(s, 3); ir (CCl.r)3030,2915, thissolution was added g (20 3.8 mmol) of a,a-diphenylacetaldehyde 1760, 7600, 1 500, I 440, 1360,1205, 1 I 70, 1050, 1020, r in ether. 7 30,6g0 cm- ; The resultingmixture was allowed to reflux overnight, massspectrum (70 ey) mle (rel (16, yielding intensity) 314 M+), 273(22), an almostcolorless solution and a white precipitate. The 272 (r0O),l6s (34), 105(29). mixture was hydrolyzedand the precipitate white *u, i.pu.uted by Anal. Calcd lbr CszHraOz:C, 84.05: H, 5.77. Found: C. filtration and washedwith ether. The combinedether layer was 83.78:H. 5.93. concentratedand passedthrough a gel g silica G column (2 X cm) 1,1,3-Triphenyl-2-acetoxy-prop-2-en-3-one(11) was isolated as elutingwith Z:1 hexane:methylene chlorideto yield 0.g e Gg%) the major product on quenchingthe mixture obtainedby reaction of 1,3,3-triphenyl-1-propeneas white a crystallini solid haiing mp betweenphenyllithium and carbon monoxideat -70o with acetic 93.5_-95.5'[it.oo mp gg-gg.]: ir (cHCl,) :050, 2870,1600, 1450, anhydride. An ether solution of 30 ml of phenyllithium(0.g05 N, 970 (trans olefin) cm-'; nmr (CDC!) (m, 6 7.2-7.4 l5), 6.4_6.g 25 mmol) was allowed to stir under an atmosphereof carbon (m,2), and4.9 (d,1. J : 7 Hz\. monoxideat -70" until reactionwas complete. Freshlydistilled 1,3,3-Triphenyl-1,2-propanedione (7) prepared , was in lorv aceticanhydride (10.2 g, 100 mmol) was added by syringeto the yield by the procedureof Sharpless, et ctl.;30O.OO g (2.44 mmol) vigoror,rslystirring solution. The solvent was removed in uacuo of I,3,3-tripher-ryl-I-propenein 10 ml of DME wasadded slowly to a (1.0Torr) leavinga white solid which wasfiltered and recrystallized stirringsolution of 1.58g (9.76mmol) potassium of permanganare from ethanolto give1.34 g(4877 of a whitecrystalline solid, 1,3,3- in l5 ml of 99.91[acericanhydride, keptat 0'by immersionrn an triphenyl-2-acetoxy-2-propen-l-one(fl): mp 140.5-142"; ir icebath. The solutionwas allowed to stir at 0'ior 1.5hr and was (CHCIB)3025, 3010, 1755,1655 m, 1600,1575, 1500, 1450, 1375, then hydrolyzed and titrated witrr a saturatedsodium bisulfite 1330,1270, 1180, 1150, 1020,980,950,880 cm-r; (CDCI3) solutionto nmr 6 destroyexcess permanganate. Etherwas added anci the 7.0-7.95(m, 15),and2.l (s,3). solution was washedwith crz.l0 volumesof water,and three20-ml Arrul. Calcd for portions CzrHrgOa: C. 80.68; H. 5.30. Found: C, of 10"1potassiumhydroxide. The ether was dried (Mg- 8O.42:H.5.44. SOo)and concentratedto yielda yellow oil. Isolationof the vellow This materialwas also synthesizedby a route analogousto that describedfor 26. (57) N-(1-Phenylethylidene)cyclohexylamine(12) was prepared G. M. Whitesides,C. p. Casey,ancl J. I(. I(rieger, J. Amer. by the Chem.Soc., 93, l31.g(l9i.l), end referenlescited therein. (58) J. L. Greeneand H. D. Zook, ibid.,gO,3629(195g). (61; g. A. Guthrie,E. Y. Spencer, (59) T. and G. F. Wright, Can. J. Chem., Tsuji, TetrahedronLett.,2413 (1966). 35,873 (1957). (60) J. H. Burckhalter zrndS. FL Johnson,Jr., .1.Amer. Chent. Soc., (62) H. O. Houseand 73,4830(1951 B. M. Trost, J. 0rg. Chem.,30, 1341, 2502 ). (1e65).

Journal of rhe American Chernical Society I 95:24 I Nouember 2g. Ig73 8129 method of Norton.63 Acetophenone(18 g, 0.15 mol). 17.4ml of initiated, it was added at such a rate that refluxing was maintained cyclohexylamine(15 g, 0.15mol), and l0 ml oi benzenewere placed at a steadylevel. The pale greensolution was stirred 1 additional in a 50-ml round-bottomed flask equippedwith a Dean-Starktrap, hr and then transferredby cannula to a flame-driedstorage flask. reflux condenser,Teflon-coated magnetic stirring bar, and outlet to Titration against2-butanol, using 1.10-phenanthroline as an indica- a bubbler. The apparatuswas swept with nitrogen and the re- tor,6eyielded a valueof 0.67M for the 105ml of benzylmagnesium (ca. action mixture refluxedfor 16 hr with continuousremoval of water chloridesolution 89%). (31) and then concentratedat 12 Torr to remove all material boiling l,l-Di($-t e rr-butylphenyl)-2-phenylethanol was synthesized below 98'. The pot residuewas distilledtrap-to-trap at I Torr to by a procedureanalogous to that describedby Adkins and Zart- yield 18 g of a clearliquid (90 mmol, 67%): nmr (CCl,')6 1.57(b' man.6E To a solutionof 105ml of C.67M benzylmagnesiumchlo- l0), 1.87(s, 3), 3.47(b, l),7.30-7.67(m, 5). This materialwas not ride in ether(70 mmol) in a 250-ml.three-necked. round-bottomed purified further. flask equippedwith a Teffon-coatedmagnetic stirring bar, a reflux 1,3,3-Triphenylprop-2-en-1-one(13) was preparedby a procedure condenser,and No-Air stopperswas added by cannulaa solutionof analogousto that describedbelow for 1-phenyl-3'3-di(lerl-butvl- 20.6e of 23 (70 mmol) in 100ml of ether. Each drop of ketone phenyl)prop-2-en-1-one:mp 84-85'[lit.0a mp 8G87"]; nmr turned a deeppurple as it struck the solutionof Grignard reagent, (CDCla)67.13 (s, l), 7.27-8.00(m. 15). the color fadedwith stirring. After addition wascomplete. the re- 1,1-Diphenyl-2-benzoylethyleneoxide (14) was preparedusing the action mixture was stirred for an additional0.5 hr. The solution procedureof House:6t mp I l5-119' [lit.esmp 123.5-124.5'],nmr was transferredby cannula to 50 ml of stirred ice water. To the re- (CDCh)6 4.70(s, 1).7.23-8.00 (m, l5). sulting slurry was added 35 ml of cold 2O'l sulfuric acid. The l-Bromo-4- tert-butylbenzene.In a 500-ml, three-necked.round organiclayer was separated.the aqueouslayer extractedwith two bottomedflask equipped with a Teflon-coatedmagnetic stirring bar, 50-ml portions of ether, and the organic layers combined,dried a refluxcondenser, addition funnel, and a gasoutlet tubeconnected (MgSOr), and concentratedto yield 25.5 e O4%) of crude 3l: from hexane; nmr (CCl4) 6 to a bubblerwas placed 100 C of lert-butylbenzene(0.75 mol). The mp 109-ll0' after recrystallization (m, (s' addition funnel was chargedwith 38.5ml of bromine (120 g' 0'75 1.23 (s, 18), 1.93 (s. 2), 3.40 (s, 1), 6.72-7 .OZ 5),7.17 8). mol). and the brominewas added to the stirredreaction mixture at a Anal. Calcd for CzrHarO: C, 87.04; H, 8.80. Found: C, fairly rapid rate" Evolution of hydrogen bromide proceededfor 86.98;H, 9.06. (32) prepared ca.2hr. at which time the reactionmixture was irradiated with a l,l -Di(4-t e rr-butyIpheny l)-2-phenylethylene was by pro- flood lamp and was stirredfor an additional48 hr. The red reac- thionyl chloride-pyridinedehydration of 31 according to a tion mixture was washedwith 100ml of 10f aqueoussodium bi- ceduremodeled on that of Newman.?o In a 250-ml erlenmeyer sulfitesolution, 100 ml of water,100 ml of l0i4 aqueoussodium bi- flask equipped with a Teflon-coatedmagnetic stirring bar was pyridine. carbonatesolution. and an additional100 ml of water. The cloudy placed25.5 g (66mmol) of crude31 and 150ml of freshlv yellow organicliquid was dissolvedin 300 ml of ether,dried (Mg- distilledfrom barium oxide. The reactionmixture was sweptwith SOr),and concentratedand the resultingiiquid distilledat aspirator rritrogen,and the ffaskcapped with a No-Air stopperand cooledin vacuum to yield 90 g of product (0"42mol, 56",;): bp 108" (14 an ice bath. Thionyl chloride(10 g, 80 mmol) was addedslowly Torr),lit.66 bp 230o,nmr (CClr)6 1.42 (s, 9), 7.0-1 (q.4i. to the stirreclmtxture lry syringe. The mixture was stirred for 1 4,4'-Di-tert-butylbenzophenone(23). The procedureused for the hr, poured into 400 ml of water. and extractedwith three 50-ml preparationof 23 was modeledon that describedb1' Jorgenson.6T portlonso! ether. The combirredether extracts were washed with ln a flame-dried. 500-ml. three-necked,round-bottomed flask three50-ml portions ol' 10"; aqueoushydrochloric acid and three (MgSOr) equippedwith a Teflon-coatecimagnetic sttrring bar, a rellux con- 50-mlportions of water. The organicmaterial was dried (849il paleyellow denser,an additionfunnel, and No-Air stopperswas placed I '5 g of andconcentrated to 1uig1620.5 g of crude32 as a iithium hydride (0.19mol) and 50 ml of dimethoxyethane,freshl.v solid. Recrystallizattonfrom hexaneyielded large white crystals: (rn. distilled from sodium benzophenoneketyl. The reaction vessel rnp78-81'; nmr(CCii)6 1.21(s. 9).6.72-7.23 l4). wasswept with nitrogen,and a solutionof 20 g of 4'tert-bstylbenzoic Anal. Calcd for CzsH:r:lC. 91.30; H, 8.69. Found: Cl. acid (0.11 mol) in 80 ml of freshlydistilled dimethoxyethanewas 91.14 H, 8.76. - - -butylphenyl)-2-acetoxyethan- I -ol (3f addedthrough the addition funnel. After addition was complete, 1 Phenyl-2,2 di(d-t ert i. (irude g the reaction mixture was refluxedfor 6 hr, then cooled. To the 32 (20.5g. 56 mmol), 125ml of chloroform,and 0.5 of placed flask cooled (0") slurry of lithium carboxylatewas added slowly by sodiumacetate trihydrate were in a 250-mlerlenmeyer mixture was cannula1 10 ml of a 1.14N solutionof 4-tert-butylphenyllithiumin with a Teflon-coatedmagnetic stirring bar. The peracetic (56 ether,in turn preparedfrom 1-bromo'4-tert-butylbenzeneusing un- cooledin an ice bath.stirred. and 7.4ml of 10il acid the exceptionalprocedures. Most of the solid materiaipresent in the rnmol) was added" The rce bath was removed after t hr: reactionmixture dissolvedin the courseof the addition. After all mixture was stirredfor 60 hr, pouredinto 200ml of ice water"and the lithium reagent had been added. the reaction mixture was extractedwith two 50-ml portions of chloroform' The combined stirred for I hr at room temperature,and was then quenchedby organic layers were washedwith 50 ml of 1073aqueous sodium transferringthe solutionlry cannulato a well-stirredsolution of ice carbonateand two 50-mlportions of water. The organicsolution water (500 ml) in a 1-1.erlenmeyer flask. The organic layer was wasdried (MgSOr) and concentrated to yielda gumm)'whitesolid. g (49 203- separated,washed with two 100-ml portions of I A sodium Recrystallizationfrom hexaneyielded 12.2 72,'Sof 33: mp (s, (s, (s' hydroxide solution,dried (MgSO-a),and concentratedon a rotary 20-5'; nmr (CClr)3 1.23(s. 9), i.35 9).2.48 l), 7.00 8)' (C:O). evaporator to yield a white crystallinesolid. which on recrystalliza- 7.33(s,5)t ir (CCli) 1755cm-r Thissubstance was used tion from hexane gave 26.Ag of 23 (0.88 mol, 79o,,;)t mp 132- without furtherpurification in the followingstep. (34). 10.2g 133'; nmr (CClo)6 1.37(s. l8),7.43(q,8); ir (CCl4)1675 cm-t a,orDi(4-tert-butylphenyl)acetophenone A solutionof (C:O). of 33 in 50 ml of dry hexamethylphosphoramidewas placedin a mag- Anal.. Calcd for 100-mlround-bottomed flask equippedwith a Teflon-coated outlet to a nitrogen 85.44; H, 8.92. nettcstirring bar and reflux condenserwith an Benzylmagnesiumchloride was prepared using a modified version bubbier. The mixture was flushedwith nitrogen,then stirred and and of the procedure described by Adkins and Zartman.6s In a 250-ml, heatedat 210' for 3 hr. The deep red solution was cooled yellow whichr three-necked, round-bottomed flask equipped with a Teflon-coated poured into 400 ml of stirred ice water. The solid was dis- , magnetic stirring bar, reflux condenser, and No-Air stoppers was precipitatedwas isolatedby suctionfiltration' This solid with placed 2.0 g of magnesium turnings and 100 ml of freshly distilled solvedin 200ml of chloroform; the chloroformextracted once (MgSOr), yield a yellow ether. Benzyl chloride (10 g,79 mmol) was added to the well- 100ml of water,dried and concentratedto yielded6.6 stirred mixture by syringe; the benzyl chloride was added in I ml solid, which on recrystallizationfrom hexane e Qa%) (CCl4)6 1.23(s, 18), portions at the beginning of the reaction; once the reaction had of 34 as whiteneedles: mp 153-.154';nmr 5.77(s,l ), 6.90-8.01(m, I 3); ir (CClr)1700srn-1 (Q:Q). Anal. Calcd for CrsHnO: C. 87.50; H, 8.33' Found: C, (63) D. G. Norton, et ql.,J. Org. Chem.19' 1054(1954). 87.21; H" 8.46. (64) P. L. Southwick, L. A. Pursglove,and P. Numerof, J. Anter. 1-Acetoxy - I -phenyl-2,2- di(4-t er t -butylphenyl)ethvlene (25). Chem.Soc., 72, 1600(1950). Freshlydistilled diisopropylamine (1.2 ml,0.8 g, 8 mmol) and 20 (65) H. O. Houseand D. J. Reif,ibid.,77,6525 (1955). (66) C. S. Marvel, M. B. Mueller, C. M. Himel, and J. F. I(aplan, ibid.,6l,277t (1939). (69) S. C. Watsonand J. F. Eastham,J.Organometctl' Chem.,9, 165 (67) M. J. Jorgenson,Org. React., 18, I (1970). (1967). (68) H. Adkins and W. Zartman, "Organic Syntheses,"Collcct. Vol, (70) M. S. Newmarr,A. Arkell, and T. Fukunaga,J' Anter' Chem. II, Wiley, New York, N. Y., 1943,p6A6. 9oc.,82,2498(1960).

Whitesides,et al. I Mechanism o.l'Reaction of Carbon Monoxide with Phenyllithium 8130

ml of ether were placedin a flame-dried,125-ml, round-bottomed solid. The filtrate was treatedwith an additional 9 ml of hydrogen flask equipped with a Teflon-coatedmagnetic stirring bar and peroxide,stirred another 12 hr, treatedonce again with 9 ml of No-Air stopper. The flask was sweptwith nitrogenand cooledin hydrogenperoxide, stirred a final 12hr, and then filtered to yield an an ice bath and 5.0ml of a 1.6M solutionof methyllithiumin ether additional0.8 g of white solid. The solid materialwas combined was addedby syringe. The mixture was stirredfor 10 min, and a and treated with 100 ml of ether. The insoluble solid remaining suspensionof 3.1 I (8 mmol) of 34 in 50 ml of etherwas addedby was removed by suction filtration; the filtrate was concentratedby cannula. The solutionturned a deepyellow, and stirringwas con- rotary evaporationto yield 2.32 g (5671 of 37: mp 149-151'; tinued for l0 min. The resultingenolate solution was added by nmr (CCl.r)61.27 (s,9), 1.35 (s, 9),4.35 (s, 1),7.30 (m, 13); ir (CClq) cannulato a stirred solution of 5 ml of freshly distilledacetic an- 1700cm_r (C:O). hydridein 100ml of etherin 250-ml,round-bottomed flask. This Anal. Calcd for CrsHezOz:C,81.46; H,7.76. Found: C, solutionwas treatedwith 100ml of saturatedaqueous sodium bi- 84.39:H. 7.90. carbonate. The organiclayer was separated,dried (MgSOr), and l,l-Di(4- te rl-butyl phenyl)-2-acetoxy-3- phenyl prop-1-en-3-one concentratedto yield a yellowoil which solidifiedon standingover- (26). Freshlydistilled THF (20 ml) and 2,2,6,6-tetramethylpiperi- night. This materialwas taken up in hot hexaneand recrystallized. dine (0.,12g, 3 mmol) were placedin a flame-dried100-ml, round- Unreacted ketone precipitatedfirst; crystallizationof the con- bottomedflask equipped with a Teflon-coatedmagnetic stirring bar centratedmother liquor yielded L0 g (28%) of 25: mp 144-146"; and No-Air stopper. This mixture was swept with nitrogen and nmr (CClr)6 1.23(s,9), 1.33(s,9), 1.80(s, 3), 6.87-7.23(m, 13); cooledto 0". A solutionof methyllithiumin ether(1.8 ml of a ir (CClq)1745 cm-t (C:O). 1.6 M solution)was addedby syringeto the stirredmixture. Stir- Anal. Calcd for CsoHsaOr:C, 84.50; H, 7.98. Found: C, ring wascontinued for 5 min, and'a solutionof 1.05g of the sub- 84.25; H, 8.28. stitutedethylene oxide preparedin the precedingstep in 20 ml of N-(1-Phenyl-3,3- di(4-terr-butyl phenyl)-3-hy d roxy)propylidenecy- THF was added by cannula. The deepred reactionmixture was clohexylamine(35) was synthesizedby a directedAldol condensa- stirred for I hr at 0" and t hr at 25o and was then quenchedby tion using a proceduresimilar to those describedby Wittig and addition to a stirredsolution of l0 ml of aceticanhydride in 75 ml f{s55s.2e'7tFreshly distilled diisopropylamine(5.05 g, 50 mmol) of ether. The resulting suspensionwas treated with 100 ml of and 40 ml of freshlydistilled THF wereplaced in a flame-dried,300- saturatedaqueous sodium bicarbonatesolution and the organic ml, round-bottomedflask equippedwith a Teffon-coatedmagnetic layer was separated,dried (MgSOn),and concentratedto give an stirring bar and No-Air stopper. The reactionvessel was flushed orangeoil. This oil was distilledtrap-to-trap at I Torr to remove with nitrogen and cooled to 0o, and 30 ml of a 1.6 M solution of the last tracesof aceticanhydride and 2,2,6,6-tetramethylpiperidine. methyllithium in ether was added by cannula. Evolution of The orange solid remaining was recrystallizedfrom ethanol to methanewas evident during the addition. The mixturewas stirred yield0.75 8((571 of 26: mp 168-169o;nmr (CCl4)6 1.13(s, 9), for 10 min, and a solutionof 10 g (50 mmol) of 12 in 80 ml of THF 1.38(s, 9),2.13 (s, 3), 6.87 (m, 13); ir (CClr)1760 and 1675 cm-r. was addedby cannula. This yellow solutionwas stirredat 0" for Anal. Calcdfor CnHarOa: C, 81.93; H,7.49. Found: C, l0 min, cooledto -78o, and a solutionof 15.1g (50 mmol) of 23 81.89;H,7.62. in 80 ml of THF was added by cannula to the stirred solution. a,a-Diphenyl-p-ethylacetophenone.DiphenylketeneT2 (9.7 g, 50 After addition wascomplete, the coolingbath wasremoved and the mmol) in 30 ml of ethyl ether was addedslowly to a solution of mixture was allowedto stand for 20 hr; it was then cooled to 0" p-ethylphenyllithium(75.0 mmol) in 120ml of ether at room tem- and treated with 100 ml of water. Stirring was continuedat 0o peraturewith vigorousstirring. After the addition was complete, for 0.5 hr. The cold mixture was filteredto yield the ketimine35 the reaction mixture was held at reflux temperaturefor 30 min, asa white solid. The filtratewas diluted with 50ml of etherand the cooledto 0 o,and hydrolyzedwith 20 ml of a saturatedaqueous solu- organiclayer was separated,dried (MgSOr), and concentratedby tion of ammonium chloride. Water (30 ml) was added and the rotary evaporation to yield additional product. The combined ether layer was separated. The aqueouslayer was extractedwith crude product was usedwithout further purificationor analysisin two 100-mlportions of ether. The etherealportions were com- conversionto 36. bined, washedwith water, and dried (CaClz). The ether was re- l,l-Di(4-terr-butylphenyl)-3-phenylprop-1-en-3-one(36). The moved, leaving a solid which was dissolvedin absoluteethanol, hydrolysisof the crude ketimine 35 was carried out by a method treatedwith decolorizingcharcoal, and recrystallizedto give 1.57g analogou.sto that of Reiff.2e In a 500-ml,round-bottomed flask (10.5%) of white crystals: mp 102-103"; ir (CClr) 1690cm-r equipped with a reflux condenser and Teflon-coated magnetic (C:O); nmr (CCln)6 7.17(m,14, aryl-H),5.88 (s, 1,CH), 2.50(q, stirring bar was placed the total quantity of crude ketimine de- 2,J : 8 Hz,CHz), andl.22(t,3,J: 8 Hz,CHs). scribedabove and 250 ml of 2 N sulfuric acid. This mixture was Anal. Calcd for CzzHroO:C, 87.96; H, 6.71. Found: C, heatedat reflux temperaturefor 2 hr. The cooled reaction mixture 87.91;H, 6.74. was extractedwith two 100-mlportions of ether. The ether ex- 3,3-Dimethyl-1-phenyl-2-hydroxy-1-butanone(38). To a solution tractswere combined, dried (MgSOe),and concentratedto yield a of 3 g (15mmol) of 2-phenyl-1,3-dithiacyclohexanein 100 ml of dry yellow oil. Analysis of this yellow oil by nmr and tlc indicated tetrahydrofurancooled to -60" wasslowly added l0 ml (15mmol, that it was a mixture of a,B-unsaturatedketone and p-hydroxy 1.5N) of n-butyllithiumby syringe.TaTbe solutionwas allowed to ketone. This crudeoil wasthen dissolvedin 150ml of ethanoland stir at -60" for I hr; then 1.6g(17.5 mmol) of 2,2-dimethylpro- treatedwith 15 ml of concentratedhydrochloric acid at reffux for panal was slowly added and the solutionwas allowedto warm to 4 hr. The reactionmixture was cooled and extractedwith two 100- -28" over a 4-hr period. The solution was hydrolyzedand the ml portions of methylenechloride. The organic portions were organic layer was separatedand washed with aqueoussodium combined,washed with 100 ml of saturatedaqueous sodium bi- chloridesolution, dried (NarSO,r),and concentratedto give3.5 g of carbonate,100 ml of water, dried (MgSOa),and concentratedto a mixture of the startingthioketal and anothercompound. With- yield a yellow solid. Recrystallizationfrom ethanol afforded 5.4 out further purification,the 3.5g of white solid wasadded to a 500- g (28%) of product. Treatmentof the mother liquors with addi- ml, round-bottomedflask containing 200 ml of 9:1 methanol:water tional hydrochloric acid, followed by further recrystallization, and 3.24g (15 mmcl) of mercuricoxide. The mixture was heated yieldedan additional3.6 g of product; the total yield of 36 was9.0 to refluxunder nitrogen; then8.13 g (30 mmol) of mercuricchloride eg6%): mp 104-106o;nmr (CCl+)d 1.30(s,9), 1.37 (s,9), 7.03- in 40 ml of the samesolvent was added. The mixture wasallowed 7.87(m,14); ir (CCl1)1670 cm-1 (C-O). to reflux under nitrogenfor 4 hr. The resultingwhite precipitate Anal. Calcd for CzsHezO:C, 87.87; H, 8.08. Found: C, wasseparated by filtration and washedwith three50-ml portions of 87.98:H. 8.22. methylenechloride. The filtrate was dilutedwith an equalvolume l,l-Di(4-terr-butylphenyl)-2-benzoylethyleneoxide (37) was pre- of water and extractedwith three 50-ml portions of methylene pared by epoxidationof 36 by a proceduremodeled on that de- chloride. The combinedmethylene chloride phase was washed with scribedby Houseand Reif.65 In a l-1.erlenmeyer ffask equipped water,aqueous ammonium acetate, and water,dried (MgSO.r),and with a Teflon-coatedmagnetic stirring bar was placed 3.96 g (10 concentratedto give an cil which was a mixture of benzaldehyde mmol) of a,B-unsaturatedketone 36 in 300ml of absolutemethanol and anothercompound. The oil in ca. 100ml of methylenechlo- and 50 ml of methylenechloride. To the stirredsolution was added ride and a few drops of absoluteethanol was allowedto stir over- 6 ml of 6 iV aqueoussodium hydroxide and 12 ml of 30f hydrogen night with 3.5 g of pulverizedcalcium chloride.Ta The calcium peroxide. This mixture was stirred for 36 hr; additional 9-ml portionsof 30'/" hydrogenperoxide were added at 6, 10,and 24 hr. (72) U. Staudinger, (1911). After 36 hr, the reactionmixture was filtered to yield 3.5g of a white Chem.Ber.,44,1619 (73) D. Secbach,Syrtthesis, 17 (1969). (74) Isolation of 37 by adsorptionon calciumchloridc was suggested (71) G. Wittig and A. Hesse,Org.5yn.,50,66 (1970). to us by ProfessorBarry Sharpless.

Journal oJ'the American Chemical Society I 95:24 I Nouentber 28, 1973 8131 chloride was separatedby filtration and washedwith small portions 0.313on silicagel 2B-F analyticalplates using 4:1 methylenechlo- of methylenechloride and then was added to a mixture of 100 ml ride:etheras eluent. This compound(6), identifiedas a-hydroxy- of water and 100ml of benzene. The benzenelayer was separated, acetophenone,sewas isolated from a 2O X 20-cm preparative tlc washedwith aqueoussodium chloride, dried (MgSOo),and con- plate coatedwith 2 mm of silica gel PF-254using 2: I methylene mp centratedto give0.9 g (lZ %) of 3,3-dimethyl-1-phenyl-2-hydroxy-1- chloride:ethereluent, as a white crystallinesolid: 84-85''; butanone: mp 33-33.5'; ir (CClr) 3500,1570, 1600' cm-r; nmr mmp 84-85". The ir spectrumof 6 is indistinguishablefrom the (CCln)67.25-7.9 (m,5),4.68 (d, l,.I :8 Hz),3.3(d, l,J : 8 Hz)" publishedir spectrumof a-hydroxyacetophenone'76 0.85(s, 9). A secondcomponent appearedas a yellow spot having Rr 0.79 Ana!" Calcd for CrzHreOz: on analytical tlc silica gel 2B-F plates using methylenechloride: 74.97:'H, 8.40. benzenein the ratio of 4: i as an eluent. To isolatethis substance. 1-Phenyl-3,3-dimethyl-1,2-butanedione.To a solution of 0.558 which was a minor product, an aliquot of the reaction mixture was g (3.14 mmol) of l/-bromosuccinimidein 60 ml of 90f aqueous sublimed(70', 0.05Torr) to removebenzophenone, leaving a resid- acetonewas added 100mg (0.522mmol) of 1-phenyl-3,3-dimethyl- ual yellow oil, enrichedwith this yellow substance,which was suh- 2-hydroxy-l-butanone in 20 ml of acetone. The solution was jected to preparative tlc From a 20 X 20-cm plate coated with allowedto stir for 2 hr at 25' then diluted with 30 ml of I :1 meth- 2 mm of silicagel PF-254"eluting with 2 :1 methylenechloride : ben- ylenechloride:pentane and 30 ml of saturatedsodium bicarbonate. zene, was isolated a yellow oil (7), which gave only one spot by The organic layer was washedwith live 30-ml portions of saturated analytical tlc using 4:l methylenechloride:benzene as eluent. sodium bicarbonate.with saturatedsodium chloride. dried (Mg- This substancehad ir spectrum. mass spectrum,and R; value SOr), and concentrated,yielding 0.79 g (80%) of 1-phenyl-3,3-di" (methylenechloride: benzene 4: 1) indistinguishablefrom that of an methyl-l,2-butanedioneas a yellow oii. The oil can be purified by authentiesample of 1,3,3-triphenyl-1,Z-propanedione. preparativethin-layer chromatography using 6 : 1 cyclohexane:ethyl f'he Rr value for a third compound (8) was A"47on analytical acetateas eluent. This substancehas nmr (CCln)6 7.3-8.0(m, 5, silica gel 2B-F tlc platesusing 3:1:12 benzene:ether:methylene aromatic),1.3 (s,9, C(CH3)a),and ir (C.lCl4)1725,1695 cffi-1, with ngas adsorption apparatus or to a 40-ml heavy walled phenyllithium-d;77(31 ml, 0.52N. l6 mmol), preparedfrom bromo- centrifugetube with stirring bar capped with a No-Air stopper, benzene-r/,?8and lithium wire, was allowed to stir under carbon -70" The stirred solutions were exposedto carbon monoxtde (in the monoxideat untii reactionwas complete. T'hesolution was former casefrom the gasabsorption apparatus, in the latter front a quenchedwith saturaledaqueous ammonium chloride solution and carbonmonoxide cylinder connecteci to the centrifugetutre through concentratedhydrochloric acid. To the resulting mixture was a length of Tygon tubing and hypodermicneedle). When gasatr- added authentic benzophenone(187 mg, 1.03 mmol), 6 (53 mg" sorptionhad ceased(4 hr at 0" and 1 atm pressure,2-3 hr at 0" and 0"39mmol),7 (13 *g,0.0434 mmol),and 8 (20 m9,0.063mmol). 2 atm)"the resultingdark brown reactionmixture was quenchedbv- The organiciaver wasseparated and washedwith water and aqueous the addition of 0.5 ml of'distilted water. Saturatedaqueous am- socjiumchloride solution' clried(MgSOn), and concentratedto an monium chloridesolution (l ml) and ether(15 ml) wereadded, and oil. Samplesof benzophenone,6. 7, and 8 wereisolated from this the ether layer was analyzedbl' gipc. oil for massspectroscopic analyses using preparativetlc" Benzo- Isolation and Characterizationof Products by Glpc. Benzophe- phenoneand 14 were isolatedusing Analtech tlc platesprecoated none and benzil were identifiedby comparisonof glpc retention with i.5 mnr of silicagel GF with 2:1 methylenechloride:cyclo- times.mass spectra. and rnfrareclspectra of coliectedsamples witii hexaneeluent" The Rr valuesfor benzophenoneand 7 under these thoseof authenticsamples. Compouncis2 and 4 were isolatedb-v' conditionsare 0.43 and 0"545.respectively. Using the sametype combiningseveral hydrolyzed reaction mixtures. The solventwas of piate and I :2 : i methyienechioride : benzene :ether eluent, 6 and yalues removedleaving a yellow oil which was dissolvedin a minimunl 8 wereisolated. Tire Rr for 6 and8 uncierthese tlc conditions volurneof hot benzene^The solutionwas allowed to cool to rootn ar:e0.40 and 0.514,respectively. Isotopicanalyses of eachof these temperature,and the resultingcrystals were collected and recrystai- rsolatedisotoprc mixtures were carrled out usinga nominal ionizing lized from absoluteethanol to give ct.a-diphenylacetophenone(2): voltageof l5 eV. In eachcase, the ratios reportedare the average -'(C--:Cl); peak heightsoi mp 137.5-138"5"ili1.zs mp 137-.139'llir (CCl4)1695 crr' oi'severalspectra For benzophenonethe ratio ol nmr (CCln)6 5'86(s, 1),7.17--8.(n (m, 15;; massspectrum (70 ev) the l92ll83 ions was 1"94+, 0.03. For 6 the ratio of peakheights mle(relintensityi 272(2.4. M*), 167(30). 105 (100),'!7 (14). of the 1101105ions was4.33 + 0.0-5. hor 7 the ratio of the peak Ana!. Calcd for C-zoH.roO:C, 8t1.20; H, 5.92. Found: C. herghtsr:f the 177!167and the l10i 105ions was 8.2 + 0.1" For I 88.12,H, 5.95. the ratio of the peak heightsof the I 10/105ions was 33 -t 0.6. a,a-Diphenyl-a-hydroxyacetophenone(4) was identified in the The product yieldsderived frorn thesedata are outlinedin Table i. hexanesolution by comparisonof its glpc retentiontimes on two Partial Reactionof Phenyllithiumand Carbon Monoxide in Ether columns(8-ft 5% DEGS on ChromosorbW at 200oand 8-ft 139; and Quenchwith SodiumDeuteroxide. A solutionof phenyllithium SE-52"on silanizedChromosorh W at 225') and its tlc rti valueson rn ether(5 ml, 1.0M,5 mmol) wastransferrecl by synngeto a 25- two adsorbents(silica Gel and Alumina eluting with chloro- ml. round-bottomedflask equipped with a No-Air stopper anC form:methylenechloride 4:1) with those of an authenticsample. T'eflon-coatedmagnetic stirring bar. This reaction mrxture was Benzpincol(5) wasassayed by treatingan aliquot of the originalre- sweprout with a streamof carbonmonoxide and then stirredunder action mixture with saturatedaqueous ammonium chloriciesolu- an atmosphereof carbon monoxide for ca. 20 min. The reaction tion, followed by hydrochloricacid and aceticacid, and analyzing mixture was cooledin an ice bath and 5 ml of sodium deuteroxide the resultingsolution for benzpinacoloneby glpc. in degasseddeuterium oxide (preparedby the addition of 0.i g of Isolation and Characterizationof Products Using Preparative sodium to 20 ml of deuterium oxide) was added cautiously by Thin-Layer Chromatography. Phenyllithium(50 ml, 0.6 li, 30.0 syringe. The quenchedreaction mixture was aliowed to stir for mmol) was allowed to stir under a carbon monoxide atmosphere cla.30 min and was then diluted with 50 ml of water and 50 ml of at -70" for 6 hr. The resulting green reaction mixture was ether. The ether layer was separated,washed with two 50-ml por- quenchedwith 2 ml of saturatedaqueous ammonium chloride solu- tion and 0.25 ml of concentratedhydrochlonc acid. The ether layer was separated.dried (MgSOa),and concentrated,yielding (76) "The Sadtler Standard Spectra," Sadtler RcsearchLaboratories, a viscousyellow oil" This oil was subjectedto thin-layerchroma- Philadelphia,Pa., 1970, ir prism spectrum9770. phenyl-dr-lithium was quenched with water; tography (tlc). The major unidentifiedspot had an Ri value of (7?) An aliquot of the resulting benzene, collected from atr 9-ft 20% UC-W98 columu and analyzed by mass spectroscopy,was found to be 97.3f( benzene- dt 2.7 7, benzene-dr. (75) A. C. Cope, P. A. Trumbull, and E. R. Trumbell. J. Amer. (78) G. M. Whitesidesand W. J. Ehmann, J. Amer' Chem.9oc.,92, Chem.Soc., 80, 2844 (1 958). 5625(1970).

Whitesides, et al. I Mechanism of Reaetion of Carbon Monoxide witlt Phenyllithium 8132 tions of water, (MgSor), dried and analyzedusing a coupled glpc through a serum cap into 0.1 N solutions of phenyllithium massspectrometer. con- in 1-cm Pyrex spectrophotometercells. solutions of trityl- Reaction of Phenyllithium and lained carbon Monoxide in Ether and lithium and of 17 were prepared by analogousprocedures starting Treatment of the ResultingReaction Mixture with Lithium Aluminum with tripheni,lmethaneand 10, respectively. Samplesof dilithium Hvdride. A solution of 14 ml of phenyllithium 0.6 N (g.4 mmol) benzophenonewere prepared by addition of small amounts of lg in ether and0.27g(0.975 mmol) of eicosane in a flame-dried40-ml to 0.1 N solutionsof phenyllithiumcontained in l-cm pyrex cells. centrifugetube, equippedwith a No-Air stopperancl a magnetic Samplescf 20 were preparedby reducingT4 mg of l-phenyl-3,3- stirring bar, was allowed to stir at -7oo under an atmosphireof dimethyl-1,2-butanedionewith excesslithium dispersionin 25 ml of carbon monoxide for 4 hr. As usual, during the course of the DME and diluting aliquots of this stock solution approximately reactionthe solution first becamered green. and then when the l:2.5 with DME. Solutionsof 15 wereobtained by reactionof f .i reactionwas complete.one 1-ml quenched aliquot was with I ml equiv of 37 with 1.0 equiv of LiTMp in ether,as describedin the of aqueouspotassium hydroxide solution and anotherl-ml aliquot preparationof 26. was quenchedwith I ml of saturated ammoniumchloride solution samplesused in studiesrequiring 18 or 19 in sorutionsfree of and 0.1 ml of concentratedhydrochloric acid. The remainingre- phenyllithiumwere obtainedby diluting concentratedsolutions of action mixture was addedto a solutionprepared by mixing 95 mg thesesubstances in closedsystems using standard techniques.a2 (2.5mmol) of lithi umaluminum (LiAlH4) hydride and20ml6f etherl Experimentsinvolving the reactionof carbon monoxidewith lg, refluxingfor 0.5 hr, and coolingto room temperature. The mixture 19, or phenyllithiumwere carried out by mixing the reaction was allowed com- to stir at 25" under nitrogenfor 30 min and was then ponentsdirectly in the spectrophotometercell. cautiouslyhydrolyzed with 3 ml of water. Each of the hydrolyzed Dilithium 4,4'-l)i(terl-butyl)benzophenoneDianion (24). In a solutions were analyzedfor benzophenone,benzhydrol, and benzil flame-dried,100-ml, round-bottomed flask equipped with a Teflon_ by glpc. As controls, thesereactions were repeatedusing 20 ml of coatedmagnetic stirring bar and a No-Air stopperwas placed0.93 phenyllithium(0.5 N,9.9 mmol) in ether. when each of the re- g of 23 (3.2 mmol) and 60 ml of freshlydistilled ether. The reac- actionswas complete,a solution g (0.g3 of 0.15 mmor) of benzo- tion mixturewas swept with argon,and 0.40g of freshlycut lithium phenonein 5 ml of ether was added uia a cannulato each reaction wire (57 mg-atoms)was addedto the reactionflask in a streamof mixture. The resultingmixtures were hydrolyzedand treatedwith argon. The flask was sealedwith a No-Air stopper and LiAIH{ as desciibed stirred abbve. The hydroiyzed ether solutions were overnight. The solution turned a deep blue, purple, analyzed glpc. then and by Analysis of the glpc data indicated that the finally dark red as the reactionprogressed benzophenonewhich had beenadded prio rto treatmentwith LiAlHa Using a syringe,20 ml of this red solution was withdrawn and had beenreduced to benzhydrol. added to a 50-ml erlenmeyerflask containing a Teflon-coatedmag- Dilithium BenzophenoneDianion (18) from Benzophenone and netic stirring bar. The solution was stirred in air until the red Lithium. Benzophenone(r .82 g, l0 mmol) and rinonadecane color had faded to pale green-yellow. This solution was treated (0.824g,3.07 mmol) weredissolved in 150 ml of anhyrJrousDME. with 20 ml of water; the organiclayer was separated,the aqueous Lithium wire (1.0 g, 140 mg-atoms, cut into 0.25-in.pieces; or layer wasextracted once with l0 ml of ether,and the organicpor- lithium dispersionwas added and the mixture was allowed to stir tions were combined, dried (MgSO.r),and analyzed uy gtpc uiing for I day at room temperatureunder nitrogen to yield a dark red tetracosaneas an internal standard. Glpc analysis revealed the solution of 18. Dilithium benzophenone dianionLan also be pre- presenceof 23 in 50i( yield,with a small amount (ca. l0\) of the pared in ethyl ether using an analogousprocedure. A pale blue correspondingbenzhydrol as well as a small amount of i higher precipitateof lithium benzophenone ketyl forms initiaily; this boiling material. The authentic sample of the rerl-butylated precipitateis subsequentlyreduced to precipitate. a red vigorous benzhydrolwas prepared by the reduction of 23 with lithium alu- stirringis requiredto keepthe precipitatesfiom seitling. minum hydride. Lithium BenzophenoneKetyl (19) from Benzophenoneand Lithium. using a syringe,another 20 ml of the red solutionof 24 waswith- Benzophenone(0.4.8 g,2.64 mmol) wasdissolved in 50ml of DME. drawn and was added to a well-stirredsolution of l0 ml of 2 N Lithium wire (0.0187g,2.64 mg-atoms) wasadded and the mixture potassium hydroxide solution. After stirring for 10 min, the wasallowed to stir at room temperature undernitrogen for 24hr. layerswere separated, the aqueouslayer wasextracted with 20 ml of Lithium BenzophenoneKetyl (19) from Benzpinacotand Methyl- ether, the organic portions were combined, dried (MgSOr), and lithium. Benzpinacol( g, 1.05 2.87mmol) and n-nonadecane(0.366 analyzedby glpc. Glpc analysisrevealed only 4,4,-di(tert-butyl)- g, 1.36mmol) were dissolvedin 20 ml of DME and were slowly benzhydrolas a product,in 88f yield. addedto methvllithiumin ethylether (3.90 ml,1.62 N, 6.31mmol) Reactionof Phenyllithiumwith carbon Monoxide in the presence uiacannula. An intenseblue color formedimmediately. of 24. In a typical experiment,a flame-dried,100-ml. round- The oxidation of Dilithium Benzophenone Dianion with carbon botlomedflask equipped with a Teflon-coatedmagnetic stirring bar Monoxidein Ether. Ten millilitersof an ethersolution of dilithium and No-Air stopperwas charged with 0.69 g of 4,4,-di(terr-butyl;- benzophenonedianion (ca. 0.004 M), which containeda known benzophenone(23) (2.5mmol), 40 ml of freshlydistilled ether, and amount of nonadecane,was transferredto a flame-driedcentrifuge 0.1 g of lithium wire (14 mg-atoms)in an argonatmosphere. This tube equippedwith a No-Air stopper r:ia cannula. The solution mixture was stirred overnight, during which time the solution was exposedto carbon monoxide and stirred occasionally. After turned a deep red. Into another flame-dried 100-ml, round- 20 min the deepred solution of 18 had turned green. An aliquot bottomed flask equippedwith a Teflon-coatedmagnetic stirring of the solution was quenched with degassedmeihanolic potassium bar and No-Air stopperwas transferredby cannula20 ml of a 1.25 hydroxideand analyzedusing glpc (Table V). N solution of phenyllithiumin ether (25 mmol). To this solution The oxidation of Lithium BenzophenoneKetyl with carbon was added by cannula the solution of 24 describedabove. This Monoxidein Ether. Benzpinacol(79.2mg,0.217 mmol), nonadec- reactionmixture was stirred for 3 hr at 25" (or for 5 hr at -7g.) ane(58.5 mg,0.218 mmol), andca. 25 ml of ether,freshly distilled under an atmosphereof carbon monoxide. The reactionmixture from sodium benzophenone ketyl, were addedto a flame-dried40- was then transferred by cannula to a stirred solution of 20 ml of ml centrifugetube equipped with a No-Air stopper and a glass- aceticanhydride in 75 ml of freshlydistilled ether cooled in an ice coveredstirring bar. To this solutionmethyllithium (ca.0.5 mmol) bath. The resultingsuspension was washedwith 100ml of a sat- was slowly added by syringewith immediateformation of a light urated aqueoussolution of sodium bicarbonateand dried (Naz- blue precipitate. After 2 hr, this precipitatewas compactedby so,). This solution was diluted to 300 ml with ether,dotriacon- centrifugationand the blue supernatantsolution was tiansferred tane was added as an internal standard,and the reactionmixture by'cannulato another40-ml centrifuge tube, which had previously wasanalyzed by glpc. been flame-dried and washedwith a solution of methyliithium in The Reaction of Dilithium BenzophenoneDianion with p-Ethyl- ether. A 5-ml aliquot of this supernatantsolution wai transferred phenyllithiumin the Presenceof carbon Monoxide. A mixture of to a similarlydried centrifugetube and wasquenched with degassed dilithium benzophenonedianion (0.093M, 1.3 mmol, 14.0 ml) methanolic potassium hydroxide. The remaining solution was in DME containinga known amountof n-nonadecaneas an internal allowedto react with carbonmonoxide for 5 hr, duiing which time standardand p-ethylphenyllithium(0.08 N, 1.00mmol, 1.25ml) the clear blue solution becamecolorless. The colorlesssolution in ethyl etherwas stirred under I atm pressureof carbonmonoxide was quenchedwith potassium methanolic hydroxide. Analysisof at room temperaturefor 2 hr. The reactionmixture was hydrolyzed both quenchedsolutions was carried glpc(Table out by V). with 0.2 ml of aqueousporassium hydroxide solution (pH 13).5 ml oxidations of 18 and 19 with carbon monoxidein DME were of etherwas added, and the reactionmixture was analyzedby glpc. carriedout usingprocedures analogousto thoseused in ether. a,a- Dip henyl- p-ethylacetophenonewa s i dentified by comparison of spectroscopicstudies. solutionsof lithium benzophenoneketyi its physicalproperties with thoseof an authenticsample. were obtained by injection of known quantitiesoi benzpinacol The Reduction of a,a-Diphenyl-a-hydroxyacetophenonewith

Journal of the American chemical societl, I 95:24 f Not,ember 2g, I9T3 8t33 Dilithium BenzophenoneDianion. Butyllithium in hexane (6.0 phenylphosphide was prepared by the reactionofdiphenylphosphine mmol) was addedslowly to 1,2-dibromoethane(0.957 g, 5.1 mmol) with n-butyllithium in THF. Lithium diphenylamidewas pre- in a 40-ml centrifuge tube capped with a No-Air stopper. The pared by the reaction of diphenylaminewith methyllithium in resulting precipitate of anhydrous lithium bromide was washed THF. Sodium dimsylatewas prepared as describedby Greenwald, four times with 20-ml portions of pentane,dissolved in 7 ml of hot Chaykovsky, and Corey.80 All of the above were reacted with DME, and added to 2 mmol of dipotassiumbenzophenone dianion carbon monoxide at 20Opsi; only sodium dimsylate took up an in 30 ml of DME containing0.1534 g of n-nonadecaneas internal appreciable quantity of carbon monoxide (0.5 equiv of carbon standard in a stoppered40-ml centrifuge tube. The solution was monoxideper equivalentof dimsylateover a 24-hrperiod). thoroughly mixed by shaking and the precipitate of potassium bromide separatedby centrifugation. The resulting solution of Acknowledgment. F. R. Koeng carried out pre- dilithium benzophenone?ewas added to a solution of a,a-diphenyl- Iiminary experimentsin this problem. This work was a-hydroxyacetophenone(0.0967 g, 0.336mmol) in 10 ml of DME. assisted substantially by a grant from the Research persisted The red color of dilithium benzophenone after the addi- Corporation for the purchaseof glpc equipment. Dr. tion. After stirring at room temperature for 2 hr, the reaction mixture was hydrolyzed with aqueouspotassium hydroxide solu- Charles Hignite, Mr. Brian Andresen, and Professor tion; the aqueouslayer was separatedand extractedwith one 50-ml Klaus Biemann generously provided us with high- portion of ethyl ether, and the organic layers were combined and resolution spectra and help with glc-mass spectra on dried (NazSOr). Analysis by glpc indicated the presenceof a,a- several occasions; this assistancewas provided under diphenylacetophenone (72%) and c,a-diphenyl-a-hydroxyace- Research tophenone(9%). National Institutes of Health Grant No. Preparation of Other Anions. Lithium phenyl acetylide and RR00317 from the Division of ResearchFacilities and lithium butyl acetylidewere preparedby reaction of the correspond- Resources. Professor Barry Sharpless was helpful in ing acetylenewith methyllithium in ether solution. Lithium di- suggestingtechniques for isolation and preparation of certain of the compounds of interest in this work. (79) Solutions of 18 prepared using this procedure were analyzed ProfessorRonald Breslow made severaluseful criticisms for potassium by hydrolysis and treatment with lithium tetraphenyl- borate: c-f.D. N. Bhattacharyya, C. L. Lee, J. Smid, and M. Szwarc, of our initial mechanistichypotheses. J. Phys. Chem.,69, 608 (1965). Lessthan l% of the potassiumorigi- nally present as dipotassium benzophenoneremained in solution after (80) R. Greenwald, H. Chaykovsky, and E. J. Corey, J. Org. Chem., addition of the lithium bromide. 28,1128(1963).