sign in sign up
<<
Home , Organomercury, Sodium borohydride

870 Reactionsof Alkylmercuric Halides with in the Presenceof Molecular Oxygen'

Craig L. Hill and George M.Whitesides*

Contribution from the Department of Chemistry, Massachusetts Institute of Technolog)', Cambridge, Massochusetts 02 I 39. ReceiuedAugust 24, 1973

Abstract: Reaction of alkylmercuric halides with sodium borohydride in saturated with molecular oxygen producesalcohols and borate estersin good yields. The products obtained following reaction of neophylmercuric bromide (l), 1,7,7-trimethylbicyclol2.2.llheptyl-2-mercuricbromide (9), and endo- and exo-nor- bornyl-2-mercuric bromides (13 and l4) with borohydride in the presenceof oxygen are compatible with a reaction mechanism involving free, noncaged,alkyl radicals as intermediates. This mechanism finds further support in the observationsthat rcaction of I with borohydride and oxygen in solutions containing2,2,6,6-tetramethylpiperidoxyl radical leads to good yields of thc prodr-rctof coupling of neophyl radical with the nitroxyl. Reaction of a-alkoxyl alkylmercuric halides with borohydridc and oxygen generatcsa-alkoxyl in good yields; similar reaction of a-hydroxy alkylmercuric halides doesnot lead to vicinal diols.

[lkyl radicalsare establishedintermediates in the re- synthesis, provided that these alkyl radicals survive 11 ductiveclemercuration of alkylmercurichalides by sufficientlylong to be accessibleto reagentspresent in metal .2-a The loss of stereochemistrythat their solutions. One previous attempt to trap alkyl occursduring conversionof the - bonds radicals produced during reductive demercurationlead of diastereomeric2-norbornylmercury compounds into to ambigllous rr-sults: reaction of 2-norbornylmer- carbon- bonds, the characteristic structural curic bronride with sodium borohydride in the pres- rearrangementsthat accompany demcrcuration of ence of high concentrations of di-tert-butylnitroxyl nortricyclyhnercurycompounds, and the absenceof l,- (DTBNO) produced approximately 20% of N,lf-di- 2-phcnyl migration on reduction of neophylmercuric tert-bntyl-O-2-norbornylhydroxylarnine.2 This yield is bromide combine to define the lifetimesof the inter- lower than that expectedfor reaction betweenfree 2- rnediatealkyl radicalsin thesereactions to be short, norbornyl radicalsand DTBNO by analogywith other br"rtdo not diffcrcntiate between radical-cage mech- reactions involving this or similar scavengers,T'8but anisms(of which onc possiblesequence is reprcscnted steric hindrance may contribute to the apparent ineffi- by eq 2 and 3) and rapid radical-chain reactions (eq ciency of the coupling reaction in this instance. This 4 and 5). paper describesexperiments intended to establishcon- ditions under which alkyl radicals, generated from ]I}I RHgBr --> RHgH (l) alkylmercuric halides by reaction with sodium boro- , can be diverted from the normal path leading --> .HgH (2) RHgH R. + to hydrocarbon by an external reagent. Molecular R' + 'HgH ---> RH * Hg(O) (3) oxygen was chosenas radical scavengerin theseexperi- it is highly reactivetoward RHgH---> R. (-1t mentsfor severalreasons: alkyl radicalsl'but relativelyunreactive toward organo- -> (s) R' + RHgH RH + R. + Hg(O) mercurycompoundst' and borohydrideion; it has small ancl its successfulcoupling with Since alkylrncrcuric halides are among the most steric requirements: moieties derived from organomercury reagents, readily availableand most tractableof organometallic alkyl particularly those synthesized by oxymercuration, compounds,;'';and sincetheir reductivedemercuration provide Lrseflllnew methods of forming carbon- is a particularly facile process,the reaction of alkyl- wor-rld and of adding functionality to olefinic mercurichalides with rnetalhydridcs seems potentially oxygen bonds These experimentswere also intended to attractiveas a method of-generating alkyl radicalsboth rnoieties. help to diflerentiate between the two types of mech- for mechanisticstudics and for nossiLrleutilization in anistic schemesoutlined by eq 2-5, by qualitatively (l) Supportedby tlre NationalInstitutes of Hcalth, Grants No. GM- establishingthe rapidity with which the generationand 16020ancl HL-15029, ancl by' thc National ScierrceF'ounclation, Crant No. GP-28-5E6X. (7) J. R. Thorlas and C. A. Tolman, J. Amer. Chem.Soc.,84,2930 (2) G. M. Whitcsidesand J. San Filippo, Jr.,J. Antcr.Chem. 9oc.,92, (1962): S. F. Nelsortand P. D. Bartlett,ibid.,88,143 (1966). 661I ( 1970). (8) Butyl radicals.gencratcd by reactionof rr-butyl(tri-n-butylphos- (3) G. A. Gray ancl W. R. Jackson,.1. Anter. Clrem.Soc., 91, 6205 phinc)silver(I)rvith 2,2,6,6-tetramcthylpiperidoxyl(TMPO), are scav- (1969);D. J. PastoanclJ. Gont'trz, ibid.,9l,7l9 (1969). engeclalrnost quantitativclyby this nitroryl radical in solution (4) Recluctionlrsing othcr rcagcntsmay take an trnrelatedcoursc: when [BuAgPBu,r]o: 0.05 M and [TMPO]o : 0.05 M: P. E. I{cndall, c/. F. R. Jensen,J. J. Miller, S. J. Cristol, and R. S. Beckley, J. Org. D. E. Bcrgbrciter,and G. M. Whitesidcs,unpublished work. Butyl Chern.,37, 4341(1912). radicals producedby photolysisof cli-n-butylbis(triphcnylphosphine)- (5) L. G. Makarova arrclA. N. Ncsrncyanov,"Methods of Ele- platinr-rrrr(II)are scavengcdwith high elicicncy by DTBNO: G. M. rrrcrtto-OrganicChcnristry," Vol. 4, North-Hollancl Publishing Co., Whitcsides,J. F. Gaasch,and E. R. Stedronsky,"/. Amer. Chent.Soc., Arrsterdam, 1967: t.. G. Makarova, "Organomctallic Reactiorrs," 94. 5258(972). Vol. 1, E. I.Bc'ckcr ancl M. Tsr-rtsui,Ed., Wiley-Interscier.lce,Ncw (9) Thc rate constantsfor rcaction of alkyl radicals rvith oxygen are York, N. Y., 1970,p I l9 ff'; Vol. 2, p 335ff. >107 l. mol-r scc 1: c/. B. Smaller,J. R. Remko, and E. C. Avery, (6) W, Kitching,Orgunometal. Chem. Rea.,3,35 (1968); W. Kitch- J. Chetrt.Phys.,48, 5174 (1968); A. A. Miller and F. R. Mayo, ./' ing in "OrganometallicReactions," Voi. 3, E. L Bcckerand M. Tsutsui. Amer. Chettt.Soc.,78,1017 (1956); C. M. Bamford and M. J. S' De- Ed., Wiley-Intersciencc,New York, N.Y., 1972,p319 tr. war, Proc. Roy. Soc.,Ser. A, 198,252(1949).

[Reprinted from the Journal of the American Chemical Society, 96,870 (1974).] Copyright I974by the American Chemical Society and reprinted by permission of the copyright owner. 871

CHz --QHr (QFgnc'zo* ,r@g;n*.*-"3

@*,h:,9Hs I t-u

o ( J @[1.1-. q I o --@[5,. J tt3 lrl ; ADDITION TIME ( min )

Figure 1. Product yields from reaction of neophylmercuric bromide (1) with sodium borohydride in the presenceof oxygen as a function of the addition rate of l0 ml of a 0.05 M solution of 1 to a O.23M solution of borohydride in dimethylformamide: (t) yields of neophyl (2); (o) yields of benzyldimethylcarbinol (3); and (A) yields of tert-bulylbenzene. Oxidations were carried out using 0.5 mmol of L and 0.7 mmol of sodium borohydride. r --] / r -'l Lwq)/ [RHqBrJ consumption of radical intermediates in reductive demercurationoccurs. Figure 2. Product yields for reaction of neophylmercuric bromide (1) and trcrns-2-methoxycyclohexylmercuricbromide (5) with sodium Results borohydride in the presenceofoxygen, as a lunction ofborohydride. Yields of products are representedby: (n) neophyl alcohol (2); Products. Initial experiments establishedthat the (O) benzyldirnethl'lcarbinol (3); (A) tert-butylbenzene(4); (rt reaction of alkylmercuric halides with sodium boro- Iruns-2-methoxycyclohexanol (6) ; (o) cis-2-methoxycyclohexanol hydride in dimethylformamidesolution in the presence (17): (l) cyclohexylmethyl ether (8). of molecular oxygen does yield alcohols as significant products. Using similar reaction conditions, the organomercurycompounds were sensiblyinert to oxy- CH, I xatltlr,o: gen in the absenceof'borohydride . The mercury- Ph-C-CHrHg Br ------> (II) originally presentin the organomercuryreagent is I t)\rF' reducedto mercury(O)in high yield during reduction in CHt the presenceof oxygen,as it is during reductionsin the I absenceof oxygen. CH.' CHa CHt I part I I I The reaction conditionsused in the major of the ph-c= cH,roH + PhcH,-c-oH + Ph-c-cH3 (6) work reported in this paper are basedon the resultsof rll investigationsof the influenceof solvent,order and rate CHI CH' CH, 234 of addition of reagents,and work-up procedureon the 77i;$s\) 8%$%) 3%(377) yield of these alcohols. A number of dipolar aprotic solvents both dissolved and were chemically inert to yields shown without parenthesesin eq 6 are those ob- sodiumborohydride and alkylmercurichalides: DMF tained by glpc analysisof the reaction mixture without rather than dimethyl sulfoxide or hexamethylphos- work-up imrnediately after addition of the alkylmer- phoramide was selectedfor use on the basisof cost and curic halide had been completed; the yields in paren- convenienccof purification and removal in work-up. theseswere obtained by glpc analysisafter hydrolysisof Reactions were carried out by adding a solution of the reaction mixture (oide infra). Only the forn-rerare alkylmercurichalide in DMF at room temperatureto a reproduced in Figure 1. For the concentrationsand solutionof sodium borohydridein DMF through which quantities of reagents typically used in these experi- a stream of oxygen was passedrapidly; comparable ments ([RHgBr]s : 0.05 M (0.5 mmol), [NaBHrio : results were obtained by adding a solution of sodium 0.23 M (0.7 mmol)) the yields of alcoholsdid not in- borohydride to an oxygen-saturatedsolution of organo- creaseif addition of the alkylmercuric halide was car- mercury compounds. The rate of addition of the so- ried out over times longer than 6 min. The relative lution of alkylmercuric halide to the oxygen-saturated yields of alcohol and hydrocarbon products also de- borohydride solution significantly influenced the par- pend on the rate of flow of oxygenthrough the reaction titioning of products between alcohol and hydro- solution: underthese conditions, flow ratesgreater than carbon; representativedata usedto definereaction con- 200 ml/min maximizedyields of alcohol. The effective ditions are summarizedin Figure I for the conversion stoichiometry of the reaction with respect to boro- of ncophylmercuricbromide (1) to a mixture of neophyl hydride was establishedby examiningthe yieldsof prod- alcohol (2), benzyldimethylcarbinol (3), and tert-butyl- ucts obtained by reaction of typical alkylmercuric (4) (eq 6). Substitution of sodium boro- halideswith varied amounts of borohydride; Figure 2 deuteride for sodium borohydride results in similar shows data obtained with L and with trans-2-methoxy- yields; no deuterium is incorporated into 2. The cyclohexylmercuricbromide (trons-S)(eq 7). In this

Hill. WhitesidesI Reactionsof'Alkvhnercuric Halides with NaBHr 872

HgBr heptane (12) (eq 8); isocamphane and its alcoholic NaBH.,.o, {!ou tl derivatives were not detected among these products. Dl\,{F \-r'"'ocH, Since Wagner-Meerwein rearrangement is believed to be concerted with formation of a carbonium ion at the 6 2 position of 9, r0 the absenceof rearrangedmaterials among the products of its demercurationindicates that a carbonium ion is not an intermediatein this reaction, but is compatible with a radical intermediate. Sim- presence 78 ilarly, demercuration of 1 in the of oxygen yielded a mixture of 2,3, and 4 (eq 6). Migration of and other instances,1.2-1.3 mol of borohydride was the phenyl group of neophyl radical is a relativelyslow required to effect complete demercuration of I mol of process(k < 105sec-t at 100");111,2-aryl migration in alkylmercurichalide. neophyl carbonium is probably concertedwith forma- In addition to the expectedalcohols and hydrocar- tion of the positivelycharged center. r2 bons, product mixtures obtained under these reaction Demercuration of endo- and a mixture of endo- and conditions contained significant quantities of sub- exo-norbornyl-2-mercuricbromides (13 and 14), in the stancesinferred to be alkoxyboron compounds on the presenceof oxygen,yielded the correspondingalcohols basis of their reactivity. The product balance ob- exo- and endo-2-norborneol(15 and 16), in addition to served when reaction mixtures were analyzed imme- small amounts of norbornene(17) and norbornane (18) diately on conclusion of the addition of solutions of (eq 9). As previously, the yields without parentheses alkylmercury reagents to the sodium borohydride so- lution ranged from 50 to 90%; subsequenthydrolysis NaBH, reactions increased the product balances to 95*100%. -r The hydrolysis reactions were carried out either by treat- 02, D\IF' oH ing the initial product mixtures with water or aqueous HgBr H acid, extracting the resulting mixtures with , and analyzingthe etherealphase, or, in many in- l5 stances,simply by heating the initial product solutions 3e%(70%) in DMF for severaldays at -ll0o; each of thesepro- ---- ceduresled to indistinguishableproduct yields. Rep- .-f (e) resentativeyields obtained using each of thesework-up LY procedures following the borohydride-induced oxida- OH tion and subsequenthydrolysis reactions of 1,7,7-tri- methylbicyclol}.2.llheptyl-2-mercuricbromide (9) are 16 t7 l8 summarizedin eq 8; yields for 1 were summarized in 12%(179'o) 8%(r0%) Lo/o(r%)

NaBH, 62% 13 + HgBr o,. D\'lF 'HgBr H 10 14 15 16 17 18 37%(730/o) 110/i,(18?;) 970(ll%) l90 (10/0) were those obtainedat the conclusionof the demercura- tion reaction; those enclosedin parentheseswere ob- ++hE tained after hydrolysis. The oxidation of 13 and 14 OH occurs with the loss of stereochemistryat ll t2 C-2 expected to result from an intermediate free 2-norbornyl rad- 10 11 t2 ical. The ratios of exo to endo alcohol(76:24 from 13 yields(fr) immediately afterreaction 41 27 5 and 77: 23 from the mixture of 13 and 14) are within ex- yields(%) afterhydrolysis perimental error of one another, and very similar to (HrO, rt, 10min) 60 37 4 ratios for products from (aqHrSOr, pH 0, 10min) 58 36 5 observed other reactions in- yields(fi) afterheating v olvin g int ermed iate 2-norbornyl radicals.2' r 3,| 4 (3days, 110") 61 36 4 (10) J. A. Berson in "Molecular Rearrangements,"Vol. [, P. de- eq 6. The yields in the following sectioninclude data Mayo, Ed., Interscience,New York, N. Y., 1963, Chapter 3; H, C. obtained both before and after hydrolysis; the differ- Brown and H. M. Bell, J. Amer. Chem.Soc., 86, 5006(1964). (ll) R. Kh. Freidlina, Aduan. Free-RadicalChem., l, 2ll 11965); encesin theseyields suggeststhe extent to which alkoxy- C. Walling in "Molecular Rearrangements,"Part I, P. deMayo, Ed., boron compounds are formed in the demercuration Wilcy, New York, N.Y., 1963; C. Riichartand R. Hecht, Chem.Ber., reaction. 98,2460,2471 (1965); G. M. Whitesides,E. J. Panek, and E. R. Sted- ronsky,J. Amer. Chem. Soc.,94,232 (1972); E. J. Hamilton, Jr., and Detection of Intermediate Free Alkyl Radicats. Free H. Fischer, Helu. Chim. Acta,56,795 (1973), and referencescited in alkyl radicals were established as intermediates using each. (12) A. H, Fainbergand S. Winstein,J. Amer. Chem.Soc.,79, 1608 unexceptional stereochemicaltests. Demercuration of (1957); W. H. Saunders,Jr., and R. H. Paine,ibid.,83,882 (1961). a mixture of endo and exo diastereomersof 9 in the pres- (13) P. D. Bartlett, G. N. Fickes, F. C. Haupt, and R. Helgeson, enceof oxygenled to the correspondingexo and endo Accounts Chem. Res., 3, 177 (1970); A. G. Davies and B. P. Roberts, J. Chem. Soc. B, 3ll, 317 (1969); l8l5 (1970); R. Schimpf and P. alcohols L0 and 11, and 1,7,7-trrmethylbicyclol2.2.ll- Heimback, Chem. Ber., 103, 2122(1970).

Journal of the American Chemical Society 96:3 I February6, 1974 873 To establishthat the loss of stereochemistryobserved it is possibleto vary the chain length widely without in going from organomercurycompounds to products grossly influencingthe character of the reaction. It did not reflect epimerization of the mercury reagents seemsnlore probablethat the reaction either involves under the reaction conditions,pure 13 and a mixture of shortchains or no chains. L3 and 14 were allowed to react to approximately 50% In a related experiment,reaction of a mixture of L competition, and the remaining organomercury re- and 9.6 equiv of TMPO with borohydrideand oxygen agents were reisolated. Comparison of the melting generated 19, the substanceexpected to result from point and infrared spectra of recovered and starting couplingof neophylradical and TMPO (eq 10),in good in thesereactions, and pre- material demonstratedthat ('H sumablyin the other reactionsstudied, no loss of stereo- (.H Frc+_\ chemistryin the starting material accompaniesreaction. I -O-N A similar conclusion was reachedconcerning the anal- I + TMPO 2 + Ph-('-CH ) ogousreaction in the absenceof oxygen.2 ;.): ('HIY Hc' Taken together, these product studies are entirely I.*, with the hypothesisthat reductive demer- ( compatible 19 !)57;) (10) curation of alkylmercuric halides in the presenceof molecular oxygen generates intermediate free alkyl yield.tT This observation establishedthat essentially radicals,which are trapped in turn by molecular oxygen all of the radicals produced by reaction between1 and and convertedultimately to alcohols. Thesedata give borohydride ion can be scavenged,and indicatesthat no indication that the initial steps in these reactions neitherthe conversionof L to the alcohols2 and 3 nor leading to the intermediatealkyl radicals differ signifi- its conversionto 4 can be cagereactions.:r) cantly from the correspondingreactions in the absence Norbornene has been used effectirely to trap the of oxygen,although small yields of 3 are observedon diboraneproduced on reactionof borohydrideion with reduction of 1 in the presenceof oxygenwhile isobutyl- alkylpalladium(Il)halides.rt Reactionof 9 and of a benzenewas not observedon reduction of 1 in the ab- mixture of 13 and 14 with borohydrideand oxygenin senceof oxygen.2 However, they do not definethe ex- the presenceof largeexcesses of norbornadiene(20 mol tent to which the overall reduction-oxidation sequences per mol of alkylmercurichalide) resulted in only ca.6fl involveschain or nonchainprocesses, or detailsof the decreasein the yields of the correspondingalcohols. involvementof the boron hydride in the reaction. In Thus, the courseof the reactionis also relativelyinsen- an effort to resolvethese questions, we examinedqual- sitive to the presenceof materialsthat might compete itatively the influence of radical inhibitors and of with the alkylmercurichalides for any pres- norbornadiene,a scavengerfor diborane, on the course ent. For comparison,reaction of 1 with 1.2 equiv of of reduction-oxidation of representativealkylmercuric BHa-THF in DMF in the presencebf oxygen under halides. reactionconditions similar to thoseused for reductions Reactions in the Presenceof Scavengers. The reac- with borohydride yielded 2 (20-30'2,),3 (<57;), and tion of neophylmercuricbromide (1) with borohydride 4 (50*60%). These results are unfortunately not very and oxygen was carried out in solutions containing informativemechanistically. The observationthat the 2,6-di-tert-butyl-4-methylphenol,rr' hydroquinone,li' presenceof norbornadienedoes not significantlyalter benzoquinone,rr and 2,2,6,6,tetramethylpiperidoxyl the product distribution does not necessarilyindicate (TMPO).16 Very high concentrationsof these scav- that diboraneor complexesof BHr with DMF are not engersslowed but did not stop generationof alcohol. effectivehydride donors toward alkyrnercurichalides Thus, under conditions in which reaction of 1 with under the reactionconditions. but only if theseor other borohydride and oxygen led to 2 (777J, 3 (B%), and 4 diboranederivatil'es are important in thesereductions, (37;), reaction of a mixture composed of 1 and 0.2 they must have reactivitytoward alkylmercurichalides equiv of 2,6-di-terl-butyl-4-methylphenolyielded 2 that is at leastcomparable with their reactivitytoward (307,), 3 (3%), and 4 (10%); similar results were ob- norbornadiene. Correspondingly,the fact that neo- tained with hydroquinone and benzoquinone. These phyl alcohol is prodr-rcedon reaction of BHr with 1 observations indicate that the conversions of alkyl- in the presenceof oxygen does not necessarilyimply mercuric halidesto alcoholsare relativelyinsensitive to that the rate of reductionof 1 by BH, is sufhcientto fiee radical chain inhibitors, and suggestthat any chain component in thesereactions must involve short chain (17) In an ef}brt to dctcct interactiorl between I and TMPO beforc lengths,or that chain initiation is sufficientlyfacile that reaction with borohyclriclc, the esr spectrulr of a DMF solution of TMPO (-10-t M) was comparccl with that of thc samc solution after (14) Thc prcclominantlycxo oxidationin thesereactions is in agree- aclditior-rof 1; thc spectra were not detcctably diffcrent. Thus, al- nrerrtwith the expectationthat 2-norbornylradical should react with though a number of mctal havc been shown to complex with di- oxygenmorc rcadily fiom thc cxo siclc; cf. D. L Davies and S. J. Cristoi, terl-butylnitroxyl rt arrcl TMPO, 1e this and previously reportecl uv Adcun.Free-Rudical Chem.,l, 1-55(1965). observations2 suggcst that any irrteraction bctrvcetr stable nitroxyl (15) K. U. Ingold, Chem. Soc.,Spec. Pabl., No. 24, 285 (1970): radicals and alkylmercuric halides is wcak. Williarn A. Pryor, "Frcc Radicals," McGraw-Hill, New York, N. Y., (18) B. M. Hollman ar-rdT. B. Eames, J. Amer. Chent.Soc',91,5168 1966,Chaptcr 2l; C, Walling, "Free Raclicalsin Solution," Wiley, (1969); W. Beck, I{. Schrnicltner,and H. J. I{cller, Chem. Ber., 100,503 Ncw York, N, Y., 1957,pp 162-178,430-436; K. U. Irrgold, Chem. (1967): W. Beck and I(, Schmidtner, ibid., 100,3363 (1967); B. M. Reu.,61,563 (1961);L. Reich and S. S. Stivala,"Autoxidation of Hollnratrn ancl T. B. Etrmcs, J. Amer. Chent. Soc., 91, 2169 (1969). Hydrocarbons ancl Polyolefins," Marcel Dekker, New York, N. Y., (19) C. M. Palcos, N. M. I(arayarrnis, and M. M. Labes, Chem- | 969,Chapter 3, ar"rdreferences cited in each. Commun.,l95 (1970). (16) Stcrically hir-rdcrcclnitroxyl groups arc believed to be stable (20) The yield of 19 observcd also suggests that the rclatively low towarcl hybridic reclucing agents and molecr.rlaroxygen; c/. A. R. yield (20 i,,") of O-2-norbornyl-i/, N-di- tert- butylhyclroxyl formed Forrester,J. M. Hay, and R. H. Thomsorr,"Organic Chemistry of on rcduction of 13 or 14 in thc presence of DTBNO: reflccts a stcric StableFree Radicals,"Academic Press, New York, N.Y., 1968,Chap- ell'ect on the coupling rcactiort, rather than a protroutlcecl cagc com- ter 5; E. G. Rozantsev,"Free Nitroxyl Radicals," Plenum Press,Ncrv ponent to thc reaction. York, N. Y., 1970; D. J. I(osmanand L. H. Piette,Chem. Commut.,926 (21) E. Vcdejs and M. F. Salomon, J, Amer. Chem. 9oc.,92,6965 (1969). (te70).

HiU. Whitesides I Reactions of-Alkvlntercuric Halides with NaBHt 874 compete with reduction of 1 by borohydride ion. and consumptionof free,noncaged, alkyl radicals. A Thus, the importanceof BHa in thesereductions is not plausiblereaction sequencefor the conversionof alkyl- establishedby the availabledata. Nonetheless.the ob- mercurichalides to alcoholsby borohydrideion in the servationthat )1.2 equiv of borohydrideper equiv of presenceof oxygen would consist of a modification of mercurial is required for maximum yield of alcohol the sequenceinvolved in the absenceof oxygen'2'23in (Figure 1) suggeststhat borohydride rather than di- theseequations "BH" representsa borohydrideof un- borane or its derivativesis the predominant reductant specifiedstructure. The availabledata give no indica- in thesereactions. tion of the probable source of the hydrogen consumed Reactions of Oxymercurated Olefins. The most in reduction of ROO. to ROOH; alkylmercurichy- readily accessibleclass of organomercury compounds dride, borohydride,or solvent(eq 13-15,respectively) are those derived from olefins by oxymercuration and "nII" related reactions.;'6'22 Conversion of alkoxymer- RHgBr --> RHgH (1) curated olefins to B-alkoxy alcohols by reaction with RHgH --e ft. (4) borohydride and oxygen takes place in good yield (eq R. + O,:-----> ROO. (12) 11),and appearsto offer a convenientmethod for con- ROO. * RHgH --> ROOH + R. * He(O) ( 13) . H-Rt' nH ..8 BHq (), ,oH II'' l/">l"h"' /'\)'"' + ROO. ----> ROOH (14) tl D\{r' ll+ V'r,rn V"tlH "()lt SH ROO. _+ ROOH (1s) 6, 40ok(52"/,) 7. 220/n(359/.) 5.R:CH. "Bll" 20.R = CH2Ph 21. 490/o(57%) 22. 330/o(38%) ROOH -> ROH (16) + (r1)all seem possible. The observationthat ) I equiv of Goo O BHt- is required to achievemaximum conversionof alkylmercurichalide to alcohol suggeststhat "BH" in 8, 4% (4%) Sok (Sot,,) eq I must be borohydrideion; the weak influenceof 73. 41"(4Yo) 2o/,(2%) norbornadieneon the reactionsuggests that diboraneis not essentialto the reactionscheme. The observation into p-alkoxy alcohols,provided that a verting olefins that deuteriumis not incorporatedinto 2 derivedfrom mixture of products can be tolerated. diastereomeric reactionof 1 with sodium borodeuterideindicates that hydroxymercuratedolefins under similar Reaction of the conversionof ROOH to alcohol takes place by productthat we have been conditionsleads to no useful direct reduction of the oxygen-oxygenbond, rather for example,treatment of trans-Z-hy- able to detect; than by base-catalyzedelimination of water from the bromide with borohydride droxycyclohexylmercuric hydroperoxidewith formation of aldehydeor , yields (37iJ, cyclohexanol and oxygen cyclohexene followed by reduction of this substance by a boro- (4o,,(), (5%); the remainder of the and cyclohexanone hydride (or deuteride). polymeric material prodr"rctappears as an apparently In conclusion,the treatment of alkylmercuric bro- The factorsresponsible for of unknown composition. mides with borohydrideis a mild and convenientway in behavior of alkoxymercuratedand hy- the difference of producing alkyl radicals in solution. If the re- olefinsare not readily apparent. droxymercurated action medium is saturatedwith oxygen,these radicals yield. Discussion can be converted to alcohols in good This method is not necessarilyapplicable to other classesof establishthat the produc- Severallines of evidence reducing agents or organomercurycompounds; re- borohydride demercuration of tion ol' alcohols by ductions of vinylic2a and aromatic2:'mercury com- presence pro- alkylmercurichalides in the of oxygen pounds follow mechanisms that differ significantly free radicals. First, the ceedsby intermediate alkyl from that followed by alkylmercurichalides and boro- rearrangementdetected during oxida- small amount of hydride, and substitution of other reducing agentsfor bromide,and the tivc demercurationof neophylmercuric borohydrideion may also result in changesin mecha- rearrangementduring oxidative absenceof detectablc nism.a'2'l demercuration of bornylmercuric bromide, exclude carbonium ion intermediatesin thesereactions. The Experimental Section is un- involvement of intermediatealkyl carbanions General Methods. Melting points were obtained using a Thomas- Iikely. both by analogywith the reactionin the absence Hoover apparatus and are uncorrected. Boiling of oxygen,2and becausea carbanion would be ex- points are uncorrected. Magnesium sulfate was employed as a pectedto react to a detectableextent with DMF. The drying agent unless otherwise stated. Nmr spectra were run as carbon , or dimethyl-r/o sulfoxide loss of stereochemistryduring conversionof the car- carbon tetrachloride, solutions on a Varian T-60 spectrometer; chemical shifts are re- bon-mercury bond to carbon oxygen bonds argues ported in ppm downfield from tetramethylsilane and coupling con- again a concertedprocess. All of these observations, stants in hertz. Infrared spectra were taken in as well as the significantrearrangement of neophyl cellsor as potassiumbromide pelletson Perkin-Elmer Models 237 or moieties to benzyldimethylcarbinylmoieties observed 2378 grati-tg spectrometers. Mass spectra were obtained on a Hitachi Perkin-Elmer Model RMU-6D mass spectrometer. Ana- during reaction of 1, the efficienttrapping of neophyl radicals by TMPO, and the slight but detectablein- (23) the formation hibition of the reaction of 1 by radical scavengers,are Similar cotrclttsiotrshavc been drawrl cotlcerrlitrg of rearrangccl iilcohols in reactiou of 2,2,2-triphcnylethylmercuric compatiblewith a reactioncourse involving generation chloride with borohyciride ion in the preseuce of oxygell: R. P. Quirk, J. Org. Chem., 37, 3554 (1972). (22) N. S. Zcfirov, Russ.Chem. Rec.,34,527 (1965); R. C. Fahey, (24) J. San Filippo, Jr., and G. M' Whitesidcs, unpublished results. Top. St ereochem.,3, 2Ji (1968). (25) T. G. Traylor, Chem. Ind. (London), 1223 (1959).

Journal o/'the Arnerican ChenticalSocietv I 96:3 I Fehruarl'6, 1974 875

lytical glpc analyses were performed on F&M Model 810 and 2905(vs), 2850 (s), 1439 (s), 1335 (m), 1155(m), 1028(vs), and737 Perkin-Model 990 gas chromatographs equipped with flame ioniza- cm-r (m); nmr (DMSOT-d6)6 7.4 (5 H, s, aromatic),4.53(2 H, s, tion detectors-and Disc integrators. Response factors were ob- OCHTAT),3.2-3.6 (l H, broad singlet,R2C11OR), 0.9-2.8 (9 H, tained with authentic samples. Products were collected for mass complex). spectra by glpc using a Hewlett-Packard Model 700 thermal con- Anal. Calcd for ciuctivity instrument with a 12-ft, 207; UC-W98 column operated C. 33 37; H, 3.85. at 140'. All components in the reaction mixtures could be trans-Z-Hydroxycyclohexylmercuric Bromide (trans-[7). To a separated using a S-ft, 15)( Carbowax 20M on Chromosorb W stirred suspensionof 63.8g (0.2 mol) of mercuric acetate in 350 ml column operated at 120" (for hydrocarbon products) and at 190' of water was added 20 ml (0.2 mol) of cyclohexene. The resulting for alcohols. Microanalyses were performed by Midwest Micro- milky mixture was stirred for 15 min and then treated with 200 ml lab, Inc.. Indianapolis, Ind. (0.2 mol) of 1 rt/ aqueous solution. A white lumpy All demercurations were carried out in reagent grade N,N- solid precipitated immediately. This solid was recrystallized from dimethylformamide and employed U.S.P. oxygen. Tetrahyclro- ethyl acetate to yield 60 g (Zg %) of trans-2-hydroxycyclohexylmer- furan was distilled from a dark purple solution of sodium curic bromide: mp 150.5-151.5'; ir (KBr pellet) 3300-3600(s), benzophenone dianion. and diethyl ether were dis- 2919(vs),2850 (s), 1442(s),1349(m), 7247(m),1150 (s), 1102 (m), tilled from calcium hydride under a nitrogen atmosphere. Mer- 1050(s), 1034(s), and 950 cm-1 (s); nmr (DMSO3-d6)6 4.8 (1 H, curic bromide and mercuric acetatewere used without purification. s, OH), 3.3-3.9 (1 H. s, methine).0.9-2.8 (9 H, complex). Materials. Benzyldinrethylcarbinol (3), tgrt-butylbenzene (4), Anal. Calcd for CoHrrBrHgO: C, 18.97; H,2.92. Found: C, isoborneol (10), (11), norbornene (18). norbornane (19), 18.98: H, 2.89. nortrorneols (15 and l6), ancl the cyclolrexanediolswere commer- cis-Z-Benzyloxycyclohexanol (22) was prepared by treatment of cial sanrplesand were used without purification. Hydroquinone, the monosodium salt of cr.s-1,2-cyclohexanediolwith benzyl bro- Lrenzoquinone.2. 6-di-lcrl-butyl-4-methylphenol, and norbornadiene mide in DMF. r'i.r-l,2-Cyclohexanediol(9.1 g, 78.5 mmol), pre- were conlmerc'ial samples and were purilied before use. 1,7,7- pared by the method of Wiberg,38was dissolvedin 40 ml of DMF. Trimelhylbicyclo[2.2.l]heptane (12, bornane), prepared by hy- Sodium amide (3.06 g, 78.5 mmol) was added slowly to this stirred drolysis of bornylmagnesium bromide. had mp I 55- I 56" (lit. z0mp solution at 0o over a period of 2 min. After addition was com- 156-157"). 2,3.3-Trimethylbicyclo[2.2.1]heptane (isocamphane) ;llete, the resulting mixture was maintained under an argon atmo- was prepared by lryclrogenating camphene over platinum black in slrhere.stirred for 40 ntin. and then treated dropwise with 14.5 g (gS ellryl acetateat 25". A sample of pure endo-bicyclo[2.2.1]heptyl-2- rnmol) of benzyl bromide at 0o. The resulting yellow slurry mercuric bromide (13). obtained as a gift from D. Bergbreiter,had turned gray on heating to reflux temperature. After 2 hr of re- mp 119 120" (lit.zr ffrp 120-121"). Neophylmercuric bromide,2 fluxing, ca. 100g of ice and 100 mlof saturatedaqueous ammonium 1,7,7-trimethvlbicyclo[2.2.I ]heptyl-2-rnercuric bromide (9),'?bicyclo- chloride were added to the reaction mixture. The mixture was [2.2.1]heptyl-2-mercuricbromides (13 and 14)2,neophyl alcohol,28 extracted with four. lCX)-mlportions of ether. The ethereal phase camplrene h ydrate, 2e 2-rnethylcamphenilol,':0 | rutt s -2-methoxycyclo- was dried (MgSOr), cortcentrated,and distilled under reduced pres- lrexanol (6),"t cis-2-methoxycyclohexanol (7),"2 tarts-2-benzyl' sure to yielcl I1.3 g (-55mmol. 7O%) of a clear oil having bp 105' oxycyclohexanol (21),'r'rand 2,2,6,6-tetramethylpiperidoxyl(TM- (0.01 mm). Elution with ethyl acetate-hexane (1 :5) on a dry PO;rr were prepared using literature procedures. Cyclohexyl cclumn of silica gel isolated ca. 0.5 g of pure cis-2-benzyloxycyclo- methyl ether (8), bp 135-136" (lii.;;rffrp 135-136"),and cyclohexyl lrexanol(22): ir (CCl,) 3570 (s), 3300-3500(s), 3059 (m), 3023 (m), benzylether (23). bp 70'(0.07 Torr) (lit.36bp 88'(0.1 Torr)), were 2932(s),2851(s), 1730(m), 1497(m),1450 (s), 1175(s), and 1080 prepared by alkylating sodium cyclohexyl oxide in DMF. cm-t (s); nmr (CCl.,) 6 7.40 (5 H, s, aromatic), 4.51 (2 H, s, truns-2-Methoxycyclohexylmercuric Bromide (5). (30 OCH2AT), 3.3-3.9(2 H, ReCHO),2.9 (1 H, broad, OH), and 0.9- ml) containing 24.6S (0.3 mol) of cyclohexenewas added slowly to a 2.1 (8 H, complex). stirred suspension of 79.5 g (0.25 mol) of mercuric acetate in 4iX) ml Anal. Calcd for CnHraO'z: C, 75.69; H, 8.79. Found: C, of methanol. After 15 min the resulting solution became clear. 75.72: H, 8.71. It was treateclwith 300 ml of 10fl aqueouspotassium bromide. A Procedures for f)emercuration in tl{t Presence of Oxygen. Sim- white solid precipitatecl immediately. This solid was recrystallized ilar procedures were employed in all reactions of organomercury twice from methanol to yield 75 e Q6?(,) of trarrs-2-methoxycyclo- compounds with sodium borohydride in the presenceof oxygen. hexylmercuricbromide: mp 112.5-113'(lit.rr rrp 114 114.5'); All reactions were carried out at ambient temperature in dimethyl- nmr (DMSO-do) 6 3.4 (3 H, S), 3.2-3.5 (l H, broad), 0.9-2.9 (9 H, formamide solution using 0.50 mmol of alkylmercuric halide, 0.70 complex). mmol of sodium borohydride, and an oxygen flow rate of ca. 3ffi Anctl. Calcd for CzHuBrHgO: ml/min unless otherwise specified. A representative procedure C, 21.06;H, 3.50. follows. tuns-2-Benzyloxycyclohexylmercuric Bromide (20). To a stirred Demercuration of Neophylmercuric Bromide (1) in the Presence of suspension of 31.9 S (0.1 mol) of mercuric acetate in 150 ml of Oxygen. Dimethylformamide (3.0 ml) and 0.025 g (O.7 mmol) of benzyl alcohol was added 10.1ml (0.1 mol) of cyclohexene. After sodium borolrydride were placed in a 40-ml centrifuge tube which l0 min the resulting clcar solution was treated with 300 ml (0.1 was capped with a No-Air stopper containing a l2-in., 15-gauge mol) of warm 0.33 N methanolic sodium bromide. A white pre- stainless steel syringe needle as a vent. Oxygen was supplied to cipitate formed over a period of 15 min. After 30 min this pre- the centrifr"rgetube from a compressed gas cylinder through a brass cipitate was collected by suction filtration and recrystallized from needle valve manifold (Metaframe) and four, 8-in., 2O-gaugestain- 3:l heptane-benzene to yield 37 g Q97,0 of tans-2-benzyloxy- less steel syringe needles extending to the bottom of the centrifuge cyclohexylmercuricbromide: mp 87-88'; ir (KBr pellet) 3030(m), tube. The oxygen flow rate could be regulated precisely by regu- lating the pressure on the reducing valve at the oxygen tank and the setting of the needle valves on the manifold. Oxygen was (26) L. Wolfl, JustusLiebigs Ann. Chem.,394,86(1912). bubbled through the solution at a flow rate of -300 ml/min. The (27) S. Winstein, E. Vogelfanger, K. C. Pandc, and H. F. Ebel, dimethylformamide solution of borohydride was flushed with J. Amer. Chem.^Soc., 84, 4993(1962). oxygen for 2 min to saturate the solution and to remove all other (28) F. C. Whitmorc, C. A. Weisbcrgcr,and A. C. Shibica, Jr., gases from the centrifuge tube. Oxidation was accomplished by (1943). J. Amer. Chem..S'oc.,65" 1469 adding 10 ml of a dimethylformamide solution containing 0.5 (29) O. Aschan,Chem. Ber.,4l.l092(1908). mmol of 1 and 0.5 mmol of n-pentadecane (an internal glpc stan- (30) S. Moycho and F. Zienkowski,Justus Liebigs Ann. Chem.,340, period to the borohydride solution by a 58 fi905). dard) dropwise over a 6-min (31) S. Winstein ancl R. B. Hcnderson,J. Amer. Chem. Soc.,65, syringe equipped with a 6-in., 2o-gauge syringe needle inserted 2196(943). through the No-Air stopper. Elemental mercury precipitated (32) K. W. Buck, A. ts. Foster, A. Labib, and J. W. Webber, -/. over a period of 7 min. Two minutes after addition of the mer- Chem.Soc., 2846 (1964). curial solution was complete, all the needles were removed from (33) B. C. Mcl{usick,J.Amer. Chem.Soc.,70,1976(1948). the No-Air stopper and the reaction tube was centrifuged to settle (34) T. Toda, E. Mori, and I(. Murayama, Bull. Chem.Soc. Jap.,1904 in suspension. The supernatant was (t972t. any elemental mercury analyzed without further delay by glpc using an 8-ft, 15fi Carbo- (35) "Dictionary of Organic Compounds," Vol. II, Oxford Univer- sity Press,New York, N.Y., 1965,p 785. (36) T. A. Cooper and W. A. Waters,J. Chem.Soc. 8,455 (1961). (37) J. Romeyn aud G, F. Wright, J. Amer. Chem. Soc., 69, 697 (38) K. B. Wiberg and (1947). 2822(tes7).

Hill. Whitesides I Reactions of Alkylnlercuric Halides with NaBH t 876 wax 20M on ChromosorbW column operatedat 720" for hydro- basisof spectraldata: ir (CCl4)f080 (w). 3052(m), 2968(s), 2930 carbon product and 190' for alcohol product. Hydrolysisof the (s),2870 (s), 1944 (w), 1871(w), 1801(w), 1469(m), 1372(m), 1358 product mixture was accomplishedeither by sealingthe 40-ml (m), 1255(m), 1245(m), 1048(m),968 (m), and 915cm-r (m); centrifugetube with the No-Air stopper and placing the sealed nmr (CCl.)6 7.3(5 H, aromatic),3.70 (2 H, CH2O),0.9-1 .8 (24 H): tubein an oil bath maintainedat ca. 110ofor 3 daysor by adding massspectrum (70 eY) ntle 289(<1, M*), 274(<1), 177(3). 157 6 ml of 1.0 N aqueoussulfuric acid to the reactionmixture, and (22),143(8).142 (100), 9r (2r). extractingthe acidifiedsolution with diethyl ether. The yield of Anal. Calcd for Cr'HnHO: mol wt, 289.2397. Found: mol alcohol as a function of the rate of addition of the alkylmercuric wt.289.2420. halide solution to the borohydride solution was determinedby Demercurationof NeophylmercuricBromide (1) in the Presenceof following thesegeneral procedures but varying the time involved Oxygen and2,2,6,6-Tetramethylpiperidoxyl. To a solutionof 0.5 in the addition of the mercurial solution to the solution of the mmol of 1 and 0.5 mmol of n-pentadecaneinternal standard in l0 borohydride. The effective production stoichiometryof alcohol ml of DMF was added0.76 g (4.85mmol,9.6-fold molar excess with respectto borohy'dride ion was determinedby following the relativeto 1) of 2,2,6,6-tetramethylpiperidoxyl.This solutionwas generalprocedure but varyingthe startingamount of sodiumboro- added to the oxygen-saturatedsolution of 0.7 mmol of . borohydridein the usual manner. Elementalmercury was com- Demercurationof Neophylmercuric (1) Bromide in the Presenceof pactedby centrifugationafter 30 min and the red supernatantsolu- Oxygen and 2,6-Di-tert-butyl-4-methylphenol and Hydroquinone. tion was analyzeddirectly by glpc (6-ft, 107i on ChromosorbW The effectsof thesetwo inhibitors weredetermined by followingthe columntemperature programmed from 100to 230"and an S-ft,15 generalprocedure (20 i, exceptthat 0.1 mmol mol I relativeto 1) Carbowax20M on ChromosorbW column operatedat 240"). of inhibitor(0.22 of 2,6-di-terr-butyl-4-methylphenolor 0.11 g of e Demercurationof alkylmercuricbromides in the presenceof oxygen hydroquinone)was dissolvedin the mercurial solution before the and norbornadienewas carried out following the generalprocedure latter was addedto the oxygen-saturatedsolution of the reducing with I .01ml ( l0 mmol)of norbornadiene(a 20-foldmolar excess of agent. Glpc analysisof the product mixture was carriedout after norbornadienerelative to alkylmercurichalide) added to the DMF 20 min. After this length of time the reactionscarried out in the solutionof the mercurialbefore addition of the latter to the oxygen- presenceof inhibitor werenot yet complete. saturatedborohydride solution. O-Neophyl-2,2,6,6-tetramethylpiperid-l-ylOxide (19). To 20 ml of a tetrahydrofuransolution of 0.3 N neophylmagnesium Acknowledgments. We thank David Bergbreiter for chloride(6.0 mmol) at -50" wasadded l0 ml of a sample of endo-2-norbornylmercuric bromide (13). containing0.63 g (4 mmol) of 2.2,6,6-tetramethylpiperidoxyl.The resultingyellow solutionwas allowed to warm to room temperature We are also indebted to Larry Trzupek and Phillip overnight. The resultinglight orangemixture was extracted with Kendall for assistance in obtaining mass spectra. three 50-ml portionsof chloroform. The combinedorganic phase High resolution mass spectra were obtained through the was washed,dried, and concentrated. (6-ft, Glpc analysis 102,, good offices of Dr. C. Hignite, Mr. B. Andresen, and UC-W98on ChromosorbW columntemperature programmed from Professor K. Biemann, 100to 230'')of the resultingred concentrateshorved one major peak under National Institutes of and severalsmaller peaks of long retentiontime, in additionto tert- Health Grant No. RR003l7 from the Division of Re- butylbenzene. The major peak was assignedstructure 19 on the search Facilities and Resources.