OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. COUPLING REACTIONS OF ENOLATE ANIONS By Robert S. Nygren A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1979 ABSTRACT COUPLING REACTIONS OF ENOLATE ANIONS By Robert S. Nygren Lithium ester enolates are dimerized by titanium tetrachloride to give symmetrical succinates. Yields are excellent for ethyl isobutyrate and ethyl phenyl— acetate but decrease as substitution at the alpha carbon of the ester decreases. The reaction is thus a complemen- tary procedure to the copper(II) bromide-promoted coupling of unsubstituted ester enolates. Where stereoisomerism in the product is possible, a mixture of gl_and mgsg. diastereomers is obtained. Lithium ketone enolates are dimerized by iron(III) chloride to give symmetrical l,h-diketones. The yield is excellent for diisopropyl ketone and decreases as substitution at the alpha_carbon decreases. The reaction is thus a complementary procedure to the copper(II) chloride-promoted coupling of unsubstituted ketone eno- lates. Where stereoisomerism in the product is possible, a mixture of diastereomers is obtained. Cyclohexanone Robert S. Nygren behaves anomalously as its enolate is dimerized in only thirty percent yield by iron(III) chloride but in sixty eight percent yield by silver acetate. With all other enolates examined iron(III) chloride gave higher yields of l,H-diketone than did silver acetate. Lithium ester enolates react with a-iodoesters gen- erated in_§itu_from ester enolate and elemental iodine to give unsymmetrical succinates. The reaction is thus a one-pot procedure for the cross-coupling of ester enolates. Lithium ketone enolates react with Eggtfbutyl haloacetates to give y-ketoesters. Once again, the highest yield was obtained with diisopropyl ketone. Yields de- creased as substitution at the alpha_carbon of the enolate decreased. To Pamela 11 ACKNOWLEDGMENT The author wishes to express deep appreciation to Professor Michael W. Rathke for his inspiration, guidance, patience, and perpetual optimism. The author considers it a privilege to have been associated with him over these last several years. The author further expresses his appreciation to his fellow students for both the professional and the personal interest and companionship which they have extended to him. The author also wishes to thank his family both in Alabama and in Michigan for their continued faith, encourage— ment, and interest throughout this effort. Finally, the author wishes to express his deep in- debtedness to his wife, whose love, encouragement, confi- dence, and support have been and continue to be irreplaceable. iii Chapter TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O O O 0 CHAPTER I - INTRODUCTION. . . . . . . . . . . . . CHAPTER III- DIMERIZATION OF a-SUBSTITUTED ESTER ENOLATES . . . . . . . . . . IntrOduction O O O O O O O O O O O O O O O O 0 Results Discussion. . . . . . . . . . . . . . . . . Experimental. . . . . . . . . . . . . . I. Materials . . . . . . . . . . . . . II. Reaction of Lithio Ethyl Isobutyrate with Iron (III) Chloride. . . . . . . A. Preparation of Lithium Di- isopropylamide (LDA). . . . . . . B. Reaction of Lithio Ethyl Iso- butyrate with Iron (III) Chloride in THF . . . . . . . C. Reaction of Lithio Ethyl Iso- butyrate with FeCl in Ether, Toluene, Tetramethylethylene— diamine, and Glyme. . . . . . D. Reaction of Lithio Ethyl Iso- butyrate with FeCl3 in Pentane with a Co-solvent . . . . . . . . E. Reaction of Lithio Ethyl Iso- butyrate with FeCl3, Amine- Free. 0 O O O O O O O O I O O O O F. Reaction of the Cobalt (II) Enolate of Ethyl Isobutyrate With F8013 in THF 0 o o o o o o 0 III. Reaction of Lithio Ethyl Cyclohexane- carboxylate with Transition Metal salts O O O O O O O O O O O O O 0 iv Page viii 21 31 31 32 32 33 33 3M 35 35 36 Chapter IV. VI. VII. A. Reaction of Lithio Ethyl Cyclohexanecarboxylate with CO— T o (1) CH3COCH2CH2COCH3 (NHu)2CO3\k ——-‘éf:f_—IE§>——' N H C) v; i \V CH C) NaOEt CH3COCH2CH2002Et -————+> (2) l Dicarbonyl compounds are usually formed by condensation reactions. However, the l,“ disposition of two carbonyl groups cannot be brought about by simple condensation of two carbonyl fragments. l,A-Dicarbonyl compounds have been synthesized by conjugate addition of acyl anion equivalents, such as nitro-stabilized carbanions,2’6 lithium di[bis(phenyl- thio)methyl]copper,7 and acyl carbonylnickelate,8 to a,B-unsaturated carbonyls (Eq. 3). _ TiC13 R-CH + CH2=CHCOR' + R-CHCH2CH2COR' + (3) N02 N02 RCOCHZCHZCOR' H09 reported an interesting variation of this Michael condensation method, using an enamine followed by oxida- tion (Eq. A). = ' = RCH CHNR2 + CH2 CH00202H5 + RECHZCH2COZC2H5 CH ' NR2 02 (u) +- RCOCH CH CO C H CuCl 2 2 2 2 5 KuwaJima10 used a ketal plus silylated cyclobutene diolate in a method which forms the bond between the Y and 6 carbons of a y-ketoester (Eq. 5). OR OSi(CH ) (CH3)3SiO 051(CH3)3 ' 3 3 ___. R R c R1R2C(OR)2 + 1 2 + SnClu + CHzcl2 0 05 P R i(CH3)3 ‘1 2C * R R COCH CH co R CH CH co R 1 2 2 2 2 11 wehrli and Chu reported the radical-initiated reac- tion of an aldehyde with diethyl maleate to form the bond between the B and Y carbons of a v-ketoester (Eq. 6). C) C) CO2Et R! RCHO + ___;5> R OEt B(OH)3 002Et > 002Et .' B \ O. 0 BO (5) -002 R O ———-> 110112002131: R H20 02Et L—————€> RCOCH20H2CO2Et 12 Pelter and coworkers combined a trialkyl borane, a lithium acetylide, andauia-halocarbonyl compound to form both the bonds between the a and 8 carbons and the y and 6 carbons (Eq. 7). This technique may be used not only to make y-ketoesters and l,A-diketones, but also to synthesize B,y-unsaturated ketones and esters. - BrCH2COZ R3B + LiCECR' + R3BCECR' I) R R' H202 ---€>RCOCHCH2002 R23 CHZCOZ R' (7) l R' >‘ RCH::1< CH2COZ Z = alkyl or alkoxide While all of these methods are useful for the synthesis of particular compounds, they all have limitations. Some require multistep procedures; others require starting materials not readily available from the carbonyl compound itself. Several of the methods place limitations on the substitution at the a or 8 carbon. From the standpoint of simplicity, one of the most attractive synthetic routes to l,H-dicarbonyl compounds is oxidative dimerization of enolate anions. Lithium ester and ketone enolates are readily available from reaction of the corresponding esters13 and ketoneslu with lithium diisopropylamide (LDA) (Eq. 8). O >__ | l \ /OL1 I -CH-C-Z + LDA + C=C + HN (8) / '\Z Z = alkyl or alkoxide Formation of a bond between the a carbons of two eno— lates, with concurrent removal of two electrons by an oxidizing agent, would lead to the desired product (Eq. 9). l O c'c'z \ OLi-2e"'- 2 C=C, 2 I o (9) /’ \Z ll -c-c-z I Z = alkyl or alkoxide Transition metal-promoted dimerization of carbanions has been a convenient method for the formation of carbon- carbon bonds in organic synthesis. Copper-promoted di- merizations of carbanions stabilized by sulfonyl, phos- phoryl, and imidoyl groups are known.15 Copper—promoted l6 dimerizations of ester and ketone17 enolates have been reported, but suffer limitations. Iron (III) promoted cross-coupling of a ketone and an ester enolate has been effected in moderate yield.18 We proposed to explore the transition metal-promoted dimerization of ester and ketone enolates as synthetic routes to symmetrically substituted succinates and sym- metric l,A-diketones. Such a study should not only be synthetically useful because of the wide availability of enolates, but should also add to our knowledge of the reactions of ester and ketone enolates and especially the behavior of transition metal enolates. CHAPTER II DIMERIZATION OF a-SUBSTITUTED ESTER ENOLATES Introduction Although the dimerization of carbanions in the pres- ence of copper salts is a well-established reaction in organic chemistry,15 until the 1970's there were relatively few examples of the copper-promoted dimerization of enolate anions. Vogler19 had reported the formation of dimers as a side product from the copper-catalyzed oxidation of a,B-unsaturated aldehydes and ketones and Kauffman had reported20 the dimerization of sodioacetophenone by copper(II) chloride. In 1971 it was reported16 that lithium ester enolates react with c0pper (II) salts to give the corresponding dimerized esters (Eq. 10). However, yields are reported to decrease drastically with increasing alkyl substitu- tion at the alpha carbon of the enolate. For example, the enolate of tertrbutyl acetate is dimerized in 85% yield by CuBrZ, while, under the same conditions, the enolate of ethyl isobutyrate is dimerized in only 25% yield. ,OLi CH200§+’ 2 CH2 = c + 2 CuBr + (10) \ 2 O-+- _+_ CH2C02 We proposed to develop a transition metal-promoted reaction to dimerize alpharsubstituted ester enolates in good yield, using ethyl isobutyrate as a model of a highly alphaysubstituted ester. We chose anhydrous iron (III) chloride as our initial one-electron oxidizing agent be- cause we had observed that it dimerized phenyllithium and lithium acetylides, suggesting that it might oxidize organolithium reagents to radicals (Eq. 11). ¢Li + FeCl3 + [¢'] + ¢-¢ (11) RCECLI + FeCl3 + [RCEC-] + RCEC-CEC—R If iron (III) chloride were to oxidize an ester enolate to a radical, one might expect that radical to react with another enolate to form a ketyl (Eq. 12). FeCl3 has been ' - - —c- o \ /O \ \ O C 2R C=C + FeCl + ‘;C—C02R ,c=c (12) shown to oxidize simple ketyls to ketones so one might expect a second FeCl3 to oxidize the ketyl to product (Eq. 13). + FeCl3 + | (13) The formation of the free radical in Equation 12 should be faVOred by increasing substitution on the alpha_carbon of the enolate since that would lead to a more highly substituted radical. If this were true, we might be able to couple alpha-substituted ester enolates in good yield. Results 1.0 molar solutions of lithio ethyl isobutyrate were prepared by adding one equivalent of ester to LDA in tetrahydrofuran (THF) at -78°C. FeCl3 was added to these solutions, which were then allowed to reach room tempera- ture over the period of approximately one hour. Analysis revealed the principal organic products to be ethyl iso- butyrate and diethyl tetramethylsuccinate. Yields of coupled product is. the ratio of enolate to iron are shown in Table I. 1.0 M solutions of lithio ethyl isobutyrate were reacted with one equivalent of FeCl3 in various solvent 10 Table I. Reaction of Lithio Ethyl Isobutyrate with Iron (III) Chloride in THF. 0L1 THF 1.0MA GogEt >k::< + Fec13 o / OEt ‘78 *RT 002Et Equivalents of Equivalents of c Enolate FeCl3 Yield (Z) 1 1/2 32a 1 1 usa 1 1 1/2 50a 1 2 39a 10 1 78b aYield based on amount of starting ester. ineld based on amount of FeCl3. CGLC yields. 11 systems. These results are shown in Table II. THF or THF/ pentane proved to give the best yields. Therefore all further reactions were done in THF. Removal of the diisopropyl amine from the enolate solution prior to addition of the FeCl3 reduced the yield from “5% to 3A%. Use of the cobalt(II) enolate in place of the lithium enolate reduced the yield to 27%, with almost “3% recovered ethyl isobutyrate. Inverse addition, 222;; adding a solution of the enolate to a solution of FeCl3 in THF, reduced the yield of coupling to 25% with almost AO% recovered ethyl isobutyrate. In an attempt to determine whether the recovered ester resulted from free radicals abstracting hydrogen from THF, the reaction was carried out in perdeuterated THF. Mass spectroscopic analysis of recovered ethyl isobutyrate showed no deuterium incorporation. At this point, we needed an ester with by-products more easily monitored on the gas chromatograph. Therefore, we changed our model from ethyl isobutyrate to ethyl cyclohexanecarboxylate. We were then able to separate and quantify coupled ester, original starting ester, and a third product; ethyl l- cyclohexenecarboxylate. The optimum yield of coupled product was obtained by direct addition of 1.1 equivalent of FeCl3 to a 1.0 molar THF solution of lithio ethyl cyclohexanecarboxylate at -78°C. This resulted in 53% coupled ester, 33% ethyl cyclohexanecarboxylate (starting 12 Table II. Reaction of LiC(CH3)2002Et + FeCl3 in Various Solvent Systems. OLi 1.0M CO2Et >fi:=< + FeCl3 > >*_ OEt T°C*25°C cozst 002Et A 8 Yield (%)° Solvent T (°C) 1. ‘2 THF —78 us 28 THF -100 53 29 Pentane/CH3CNa —38 35 --b Glyme —38 O 100 TMEDA -78 3A 2A Pentane/DMSOa -78 18 52 Pentane/THFa ~78 H9 8 Pentane/pyridinea -78 19 --b Pentane/DMFa -78 o -_b Etgo -78 32 us ¢CH3 -78 17 -—b aRatio pentane:cosolvent=9:l. bCould not separate from solvent by GLC. cGLC yields. l3 ester) and 13% ethyl 1-cyclohexenecarboxylate (unsaturated ester). We then surveyed various other anhydrous transition metal salts which could act as one-electron oxidizing agents. These results are shown in Table III. We found T101“ to give the best yield of coupled ester. Highest yields were obtained when TiClu was added dropwise to 1.0 molar solutions of ester enolate at 0°C. Reaction mixtures were allowed to reach room temperature over the period of about A5 minutes and then quenched with three equivalents of 5.0 M aqueous sodium hydroxide. Removal of diisopropyl amine from enolate solutions prior to addition of TiClu gave no improvement in yield. A tan precipitate forms upon addition of TiClu to the ester enolate solution. With warming and time, the reac- tion mixture darkens and the viscosity increases markedly. Then, as the reaction nears completion, the reaction mixture thins down to a dark brown, almost homogeneous solution. Quenching by aqueous sodium hydroxide followed by addition of pentane gives a colorless organic phase over a thick, blue aqueous layer reminiscent of aqueous TiClB. We discovered that the reaction was highly sensitive to benzophenone impurities in THF and to lithium butoxide in butyllithium. THF had to be distilled from lithium aluminum hydride and n-butyllithium had to be free from 1A Table III. Reaction of Lithio Ethyl Cyclohexanecarboxylate with Various Transition Metal Salts. c0221 Et CO Et 0L1 A-lz 02 2 <:>-——< ——> OEt c Yield (1) Metal Salt A B C FeCl3 53 3A 13 Fe(acac)3a 3a 22 13 Fe(och)3b o --d o Co(acac)3a 13 29 20 CrCl3 12 39 20 Cr(acac)3a 18 80 1 CuBrz A8 18 7 CeSOu o --d 0 MnBr2 o --d o Hg:2 o --d 0 1/2 I2 0 --d o T101“ 6n 8 5 Ti(0Et)u o 51 o Ti(O<< )u 0 8“ 0 ZrClu o --d o Ti(Cp)2C12 o --d o aacac-2,5-pentanedionate. bR--CH-(CH2)3-CH3. cCp-cyclopenta— dienate. C2H 5 Yield of B not determined. 15 any turbidity in order to maximize the yield. While TiClu gave only a modest improvement in yield for ethyl cyclohexanecarboxylate (TiClu: 6“%; FeCl3: 53%), it raised the yield for the coupling of ethyl isobutyrate from “5% with FeC13 to 89%. A series of ester enolates were coupled with TiClu. The results are shown in Table IV. Clearly a-branched and a-phenyl enolates gave the highest yield. The yield plummets dramatically as alpaa substitution decreases. A ketone enolate, lithio cyclo- hexanone, did not dimerize. When the reaction was carried out in the presence of anisole to trap the intermediate free radical, no adduct was observed. Several experiments were done in an attempt to deter- mine the source of the recovered ethyl cyclohexanecarbox— ylate. The coupling reaction was carried out in per- deuterated THF both at 1.0 molar and 0.1 molar. Recovered ester was analyzed by mass spectrometry and no deuterium was found to be incorporated. The reaction was carried out as usual and quenched with D20. Recovered ester was again measured by mass spectrometry and found to contain no deuterium. The reaction was carried out as usual and then carefully analyzed for ethylene to check the possibility of removal of a hydrogen from the ethoxy group of the enolate (Eq. 1“). 16 Table IV. Direct Addition of TiClu to Enolates. \ / I C=C,OLi TiClu/THEL -?-C02R ‘\0R T°C 25°C’2 -C-002R I Starting Carbonyl Compound T Yield (%)e cyclo-C6H11C02Et 0a 6“ a afC3H7002Et o 89 06H5CH2C02Et oa 9“f cyclo-CuH7CO2Et -78b 67 cyclo-C3HSCO2Et -78b 5“ CH3CH2C02Et —78C 0 CH3(CH2)uC02Et -78° 10-15f CH3CH=CHCO2Et -78C 0 CH3CON(CH3)2 —78C 0 C CH3COZC(CH3)3d -78 o CH3CO2C(CH3)3 -78 16 cyclohexanone -78c 0 aYields 5-10% less if addition is carried out at -78°. inelds 5-1o% less if addition is carried out at 0°. cYield o if addition is carried out at 0°. dAmine—free enolate used. eYields based on GLC analysis. fObtained as a mixture of stereoisomers. l7 \cqc’ Mn _’ \C C,o- + 011243142 0 - \ . \ / GLAD-0112- ”CHZU -\-c -002Et / o + H-C—COZEt (1“) Less than one mole percent of ethylene was found. The same amount of ethylene was liberated by adding the ester to LDA in THF and allowing it to warm to room temperature with no transition metal present. Further investigations were carried out to look for the intermediacy of d-chloroester. Lithio ethyl iso- butyrate was added dropwise to two equivalents of TiClu in THF. No chloroester was observed. Lithio ethyl iso- butyrate was coupled in the presence of one equivalent of ethyl a-chloroisobutyrate, yielding 58% coupled product and 79% recovered chloroester. Lithio ethyl isobutyrate was added to one equivalent of ethyl u-chloroisobutyrate, giving no coupled product. Finally, ethyl 8,8,8-trideu- teroisobutyrate was reacted with TiClu in the presence of one equivalent of ethyl a-chloroisobutyrate (Eq. 15). CD I 3 CD3 ,OLi 9H3 TiCl CH3"? "—002“ 7H3 CH3 OEt CH3 CD3(H3) CH3 (15) 18 Recovered chloroester showed no deuterium incorporation, though the coupled product showed some deuterium loss. Less substituted ester enolates were found to give increased yields of coupled products when an inverse addition procedure, where the enolate solution is added to a solution/suspension of TiClu in THF, was used. This procedure drastically reduced the yield with the highly substituted esters. Results of inverse addition of var- ious enolates to TiClu are shown in Table V. In an attempt to draw comparisons among the various metal coupling agents, lithio ethyl cyclohexanecarboxylate was coupled with CuC12, FeCl3, and TiClu in THF at 1.0 molar, 0.1 molar, and 0.01 molar concentrations. All reactions were done by adding the metal salt to the eno- late at -78°C and allowing the reaction to reach room temperature. No attempt was made to optimize the pro- cedure for each metal. CuC12 was used instead of CuBr2 in the interest of consistency. The results are shown in Table VI. Cu012 does not couple the enolate of ethyl cyclohexanecarboxylate well at any concentration. The yields with FeC13 appear to be unchanged from 1.0 to 0.1 molar, then to decrease drastically while the yields with TiClu seem to vary more or less smoothly with concentration. 19 Table V. Inverse Addition of Enolates to TiClu. I OLi -78° -c-co Et \ / 2 TiClu/THF + ,c=q\ ———€> OR THF -C—CO2Et I Ester Yield (%)c i-C3H7CO2Et “1 cyclo-C6H11C02Et 30 d CH3CH2002Et 28 a a Et02C(CH2)6C02Et 32 CH3C02C(CH3)3 27 b CH3C02C(CH3)3 61 aProduct is l,2-dicarboethoxycyclohexane. bAmine-free enolate used. cYield based on GLC analysis. dProduct is a mixture of stereoisomers. Table VI. 20 Effects of Dilution on Metal Coupling Agents. CO2Et 002Et 002Et 0L1 THF wan—e 0+ 6 OEt '782’25° A B 0 Yield (%)a Metal Salt Molarity A B C 011012 1.00 6 50 11 0.10 3 53 A 0.01 3 33 2 FeCl3 1.00 30 32 18 0.10 3“ 31 18 0.01 9 71 7 T101” 1.00 52 1“ 7 0.10 30 3“ 13 0.01 13 60 10 aYie1d based on GLC analysis. 21 Discussion Direct addition of TiClu couples alaaarsubstituted lithium ester enolates in THF solution. Yields decrease as substitution on the alpaa carbon of the enolate de— creases. Thus TiClu—promoted coupling of substituted ester enolates is a complementary procedure to the CuBr2- promoted coupling of unsubstituted ester enolates. While inverse addition increases the yields obtained with less substituted enolates, it drastically decreases the yield for highly substituted enolates. This suggests that the less substituted ester enolates couple by a dif— ferent mechanism than the highly substituted ester eno- lates. This dependence upon order of addition also suggests that coordination of the enolate to the titanium is im— portant to the coupling of substituted enolates. Direct addition of TiClu to the enolate solution allows the enolate and the solvent to compete for coordination sites on the titanium. Inverse addition allows the titanium to completely coordinate with the solvent before the eno- late is present and significantly lowers the yield. We observe that the addition of TiClu to pure THF is quali- tatively more violent than addition of TiClu to one molar ester enolate in THF. We also observe that titanium (IV) alkoxides do not dimerize ester enolates. Addition of titanium tetraethoxide or titanium tetraisopropoxide to 22 ester enolates gives a red solution but no brown color or precipitate. Upon quenching, one finds titanium dioxide but no blue material in the aqueous layer. The organic phase contains only starting ester and condensation products. The first three entries in Table IV, ethyl isobutyrate, ethyl cyclohexanecarboxylate and ethyl phenylacetate, show the highest yield when TiClu is added at 0°C, while the next two entries, ethyl cyclopropanecarboxylate and ethyl cyclobutanecarboxylate, show the highest yield when TiClu is added at -78°C. This may be rationalized by considering that the three and four membered rings have considerable strain introduced into them by enolization. Addition of TiClu at -78°C may slow down enolate decomposi- tion caused by hot spots in the exothermic reaction. Six- member ring and acyclic enolates may be stable enough to withstand addition of TiClu at 0°C. It is possible that lithium ester enolates do not compete with THF for co- ordination sites on titanium (IV) as well at -78°C as at 0°C, which would explain why addition of TiClu at ~78°C lowers the yield with the six member ring and acyclic enolates. The major side products of the reaction are starting ester and the corresponding a,B-unsaturated ester. There is always more of the former than the latter, which is not surprising since the unsaturated ester is susceptible 23 to Michael condensation as well as metal-promoted polym- erization. However, it is of both theoretical and prac- tical interest to know whether all of the saturated ester comes from formation of unsaturated ester or whether some significant portion of it has another source. This is equivalent to asking where the alpaa_hydrogen on the re- covered saturated ester comes from. Does it all come from the aaaa carbon of another molecule, or does part of it come from somewhere else? We are able to rule out the solvent, the amine, the work-up, and the ethoxy group as sources of hydrogen. The reactions in deuterated THF show that, at least at 1.0 and 0.1 molar, no hydrogen is abstracted from THF. Re- moval of diisopropyl amine does not affect the yield of coupled product or unsaturated ester, suggesting that the amine is not a hydrogen or proton source. Quenching the reaction with D20 shows that none of the saturated ester comes from unreacted enolate which is protonated during aqueous work-up. Abstraction of a hydrogen from the ethoxy group of the enolate should lead to the forma- tion of ethylene (Eq. 1“), yet we observe only less than 1% ethylene. Therefore, we must conclude that all of the saturated ester is formed by removal of a hydrogen from the 2223 carbon of another molecule, so that one molecule of unsaturated ester must be formed for every molecule of saturated ester. 2“ We felt that it was necessary to consider whether the reaction might proceed via an a-chloroester intermediate (Eq. 16). I \ ,0Li 110114 I \ ,0Li "0"002R . /c=c _———>01-c-002R C=C (16) \0R I ’ \OR \. —C-C02R I 16 d-Bromoester has been reported as a side product in the CuBr2-promoted coupling of unsubstituted ester enolates. However, it was shown that the bromoester was not a necessary intermediate by using copper (II) valerate instead of copper (II) bromide to effect the coupling. We were unable to do this since our titanium (IV) alkoxides failed to couple ester enolates. Although it seemed unlikely that the enolate of ethyl isobutyrate would substitute the tertiary chlorine of ethyl a-chloroisobutyrate, Brock— som22 had reported an 85% yield for the corresponding re- action with the iodoester and we had observed a “0% yield with the bromoester. We were unable to observe any sub- stitution when we reacted lithio ethyl isobutyrate with ethyl a—chloroisobutyrate. We also added the enolate dropwise to two equivalents of T101” and observed no chloroester. This was, in our estimation, the most likely procedure for maximizing the yield of any chloro- ester that might be formed. These results strongly 25 suggested that chloroester was neither formed nor con- sumed in the reaction, but did not prove that chloroester was not formed slowly and used up quickly under the re— action conditions. However, the TiClu-promoted coupling of lithio ethyl B,8,B-trideuteroisobutyrate in the pres- ence of non-deuterated ethyl a-chloroisobutyrate shows that while the enolate apparently does react with the chloroester in the presence of TiClu (Eq. 17), no chloro— ester is formed from the enolate under the reaction CD CD CD3 OLi CH I 3 I 3 \. / I 3 T101” CH3-C-002Et CH3-C-002Et ‘/c=c\ + Cl—C-CO2Et ____29 I + I CH3 OEt CH3 CH3-C-002Et CH3-f-C02Et D CH C 3 3 (17) conditions (Eq. 18). CD3 OLi CD3 \ _ I 8 //C-C + 1101“ 01-0-0022t (1 ) CH3 OEt CH3 We had originally conceived the coupling of substi- tuted enolates occurring through a ketyl mechanism (Eq. 12, 13), although we did not rule out the possible di- merization of two free radicals (Eq. 19). Either mechanism 26 would explain the by-products, either by disproportiona- tion of two free radicals (Eq. 19) or by abstraction .. _\_ 0 R 2 0 2M (C 2 — ——-—> 2>—C02R '9 \ ; C02R OR (19) 2>~C02R —-9 H>-‘C02R+ >~CO2R of a beta-hydrogen from an enolate (Eq. 20). In order to distinguish between dimerization of two free radicals O- \ O M w + a a new. + >3 ——> OR OR \>v-CO2R (20) and a reaction involving some metal—containing species, lithio ethyl cyclohexanecarboxylate was coupled at three concentrations and with three different metals. If all of the reactions involved disproportionation and dimeriza— tion of free radicals, then the yields of coupled product and saturated ester should vary with concentration in a 27 similar manner for each metal. If, on the other hand, the metal were involved in the coupling, the ratios of coupled and unsaturated ester should show a different variation with concentration for each metal. The results in Table VI show that CuCl2 does not couple the hindered ester very well at any concentration. The yields with FeCl3 are essentially unchanged going from 1.0 to 0.1 molar, then fall off sharply at 0.01 molar. The yields with TiClu, on the other hand, vary more or less regularly with concentration. This suggests that at the very most only one of the metals could promote coupling Xaa dimeriza- tion of two free radicals. The exact nature of the intermediate remains unknown. The failure of the intermediate to abstract hydrogen from THF and diisopropylamine or to form an adduct with anisole suggests something other than a free radical. 0n the other hand, yields are lowered by the presence of radical scavengers such as lithium butoxide and benzophenone. Because there is no clear evidence for free radicals, we considered mechanistic alternatives. One such alternative involves an intermediate having two enolates on the same titanium atom. This could account for all of the products but would involve two-electron oxidations (Eq. 21) either from an unknown titanium (V) species (Eq. 21, n=5), or to an unknown titanium (II) species (Eq. 21, n=“). This difficulty could possibly 28 reductive_.> Tin'2 + *f'C02R /’ elim " 0' -C-002R n (>0 d/ ) 5 ' 0- T1 = \OR 2—_——> \ 00 R + H—Tin (\C=c’ ) '* elim 2 / \OR reductive n-2 | A/ Ti + H-CCO2R (21) elim [ be rationalized by postulating either a transient titanium (V) which immediately goes to titanium (III) or a transient titanium (II) which immediately reacts with another titan- ium (IV) to form two molecules of titanium (III). It seems likely that, whatever the nature of the inter- mediate, the first step is transmetalation to form a titan- ium enolate23 (Eq. 22). The titanium (IV) enolate may v \ /0Li \ ,OTiI C=C + T101“ + ,c=c + LiCl (22) / \OR \OR next be oxidized to a radical plus titanium (III) and then couple (stepwise reaction) (Eq. 23), or couple and oxidize I \ ’0T11v \ In -C-C02R /c=c -> -C-C02R + Ti + + (23) -?-CO 2R 29 simultaneously (concerted reaction) (Eq. 2“). n \ fifoTiIV ‘ 0 4.1111 C=C -c—c /2 \OR ‘OR " (2“) A similar concerted mechanism could account for the by- products (Eq. 25). The concerted mechanism accommodates f" 0-... III IV Ti 91—“ 11> /< H'\> {OR + CH C OR 3 \ OR (25) ‘00 TiIV \O ...... T1111 the fact that the intermediate reacts to form a bond between alpha carbons or to abstract beta hydrogens but does not do other typical free radical reactions. It also accom— modates the fact that titanium (IV) is a relatively weak 3O oxidizing agent. In the concerted mechanism, the titanium (IV) is never required to generate a free radical or other high-energy species, but only to accept an electron while a stable molecule is being formed. Finally, we should consider why TiClu dimerizes highly substituted enolates but fails with unsubstituted enolates. Titanium (IV) is a weak oxidizing agent and may therefore require substituents on the alpaa_carbon to stabilize any radical character that may develop there over the course of the reaction. We typically observe unreduced titanium (IV) in the aqueous work-up of attempted couplings of un- substituted ester enolates, suggesting that the enolates have otherwise decomposed before oxidation could occur. It is tempting therefore to suggest that success of the reaction depends upon the stability of the enolate. One could argue that the more substituted enolates survive longer at the temperature and conditions at which coupling occurs. One must note, however, that the enolate on N,N- dimethylacetamide is undimerized by TiClu even though it is considered to be exceptionally stable.25 Since TiClu is a potent Lewis acid and the data indi- cate that coordination is important, one might want to consider an oxygen-bridged intermediate, possibly polymeric, where titanium enolates complex each other. This could account for the observation that the reaction mixture first thickens considerably and then thins out again as the reaction proceeds. A bridged intermediate could nicely 31 accommodate the concerted reaction mechanism, holding together the alpha carbons, or alpha carbon and beta hydrogen, thus facilitating the reaction. One possibility for a bridged intermediate is shown in Equation 26. How III ”(:22 <0.- T11 0,...- TiIV Tixan-n- 0T1 \ \j \ ” J\O’ /\/‘O /_______ (26) close together the alpha carbons could actually be in such a hypothetical intermediate would depend upon the number and nature of the other ligands (C1, THF, diisopropylamine) on the titanium. Experimental I. Materials peButyllithium, Aldrich, was obtained as a 1.6 M hexane solution and standardized by the method of Watson and Eastham.ll4 Diisopropylamine and dimethylformamide were distilled from calcium hydride and stored under argon. Tetrahydrofuran (THF) was distilled over lithium aluminum 32 hydride. Ethyl d-chloroisobutyrate was prepared from ethyl isobutyrate and carbon tetrachloride by the method of Arnold and Kulenovic.29 Ethyl 3,3,e-trideuteroiso- butyrate was prepared from ethyl propanoate and trideutero- methyl iodide by the method of Schlessinger.39 Other esters were obtained commercially and distilled from cal- cium hydride. Titanium tetrachloride, Alfa, was distilled prior to use. Anhydrous iron (III) chloride and chromium (III) chloride, Baker, were stored and handled in an argon atmosphere. Anhydrous copper (II) chloride, Mathe- son, Coleman and Bell, was stored in a drying oven at 130°C and handled briefly in the atmosphere. All other anhydrous transition metal salts were purchased from Alfa, stored in a desicator, and handled in an argon atmosphere. II. Reaction of Lithio Ethyl Isobutyrate with Iron (III) Chloride A. Praparation of Lithium Diisopropylamide (LDA) - A 50 ml round-bottomed flask equipped with septum inlet, magnetic stirrer, and mercury bubbler was flushed with argon, immersed in an ice water bath and charged with 3.3“ ml (5.5 mmole) 1.6 M_p7butyllithium solution and 2.5 ml arpentane. Stirring was initiated and 0.775 ml (5.5 mmole) diisopropylamine was added dropwise. The cooling bath was replaced by a warm (approximately “0°C) water bath and the mixture stirred for five minutes. The solvent was removed 33 under vacuum to yield LDA as a white powder. B. Reaction of Lithio Ethyl Isobutyrate with_I;pn (III) Chloride in THF - The following procedure is representative of the preparation of lithium ester eno- lates and the dimerization of lithium ester enolates by FeCl3 in THF: LDA (7.35 mmole), prepared as previously described, was dissolved in 7.35 ml THF and cooled to -78°C in a Dry Ice/acetone bath. Ethyl isobutyrate (0.98 ml; 7.35 mmole) was added dropwise and stirred for 15 minutes. FeCl3 (1.19 g; 7.35 mmole) was added through a powder funnel under a stream of argon. The reaction was stirred for 15 minutes at -78°C, then allowed to reach 25°C over the period of “5 minutes. afPentadecane was added as an internal GLC standard. The reaction was cooled to 0°C and 3 m1 saturated aqueous KH2POu and 10 m1 prpentane were added. The organic phase was dried over K2CO3 and analyzed by GLC (column temperature: 100°C for four minutes, then 175°C). Diethyl tetramethylsuccinate was observed in “5% yield. C. Reaction of Lithio Ethyl Isobutyrate with FeCl3 in Ether, Toluene, Tetramethylethylene- aiamine, and Glyme - The procedure previously described for THF was followed, except that the appropriate solvent is substituted for the THF. For example, 5.62 mmoles of LDA was dissolved in 5.62 ml diethyl ether and 3“ cooled to -78°C. Ethyl isobutyrate (0.75 ml; 5.62 mmole) was added dropwise. The reaction was stirred for 15 min— utes at -78°C, then 0.912 g (5.62 mmole) FeCl3 was added. The reaction was stirred for 15 minutes at —78°C then allowed to reach 25°C and analyzed as previously described. Diethyl tetramethylsuccinate was observed in 32% yield. D. Reaction of Lithio Ethyl Isobutyrate with Fe013 in Pentane With a Co-solvent - The follow- ing procedure is representative: a 50 m1 round-bottomed flask equipped with septum inlet, magnetic stirrer and mercury bubbler was flushed with argon, immersed in an ice water bath, and charged with 2.79 ml (“.33 mmole) agbutyllithium and 2.0 m1 aepentane. Stirring was initi- ated and 0.61 ml (“.33 mmole) diisopropylamine was added dropwise. The reaction was stirred for 15 minutes at 0°C, the ice water bath replaced by a Dry Ice/acetone bath, and 0.58 ml (“.33 mmole) ethyl isobutyrate added dropwise. The reaction was stirred for 15 minutes at -78°C, then 0.5 ml (10% of solvent) pyridine was added, followed by 0.70 g (“.33 mmole) FeCl3. The reaction was stirred for 15 minutes at -78°C then allowed to warm to 25°C and analyzed as previously described. Diethyl tetramethylsuccinate was observed in 19% Yield. 35 E. Reaction of Lithio Ethyl Isobutyrate with FeC13, Amine-free - The following procedure is representative of the preparation of amine-free lithium ester enolate solutions: a 1.0 M pentane solution of lithio ethyl isobutyrate was prepared as described in the preceding procedure. The cooling bath was removed and the solvent and amine evaporated under vacuum. After most of the solvent had been evaporated, a warm (approximately “0°C) water bath was placed under the reaction flask and the remaining solvent and amine evaporated, leaving the lithium ester enolate as a white powder. The enolate was dissolved in “.3 ml THF to make a 1.0 M solution and cooled to -78°C. FeCl3 (0.70 g; “.3 mmole) was added through a powder funnel. The reaction was stirred for 15 minutes at -78°C then allowed to warm to 25°C and analyzed as previously described. Diethyl tetramethylsuccinate was observed in 3“% yield. F. Reaction of the Cobalt (II) Enolate of Ethyl Isobutyrate with FeCl3 in THF - A 1.0 M THF solution (6.9 ml; 6.9 mmole) was prepared at -78°C as previously described. Cobalt (II) bromide (1.51 g; 6.9 mmole) was added through a powder funnel. The cooling bath was removed and the mixture warmed with stirring until the CoBr2 dissolved and an intense blue solution was formed. The Dry Ice/acetone bath was then replaced and 36 the reaction allowed to reach -78°C. CoBr2 did not pre- cipitate. FeCl3 (1.12 g; 6.9 mmole) was added through a powder funnel and the reaction carried out and analyzed as usual. Diethyl tetramethylsuccinate was observed in 22% yield. III. Reaction of Lithio Ethyl Cyclohexanecarboxylate with Transition Metal Salts A. Reaction of Lithio Ethyl Cyclohexanecarboxylate with Co(CH3COCHCOCH3)3 - The following procedure is representative of the reaction of lithio ethyl cyclo- hexanecarboxylate with transition metal salts: a 1.0 M THF solution of lithio ethyl cyclohexanecarboxylate (“.8 mmole) was prepared by the method previously described for lithio ethyl isobutyrate. Co(CH3COCHCOCH3)3 (1.88 g; 5.3 mmole) was added through a powder funnel. The reaction was stirred for 15 minutes at -78°C then allowed to warm to 25°C. agPentadecane (0.67 ml; 2.“ mmole) was added as an internal GLC standard. The reaction was cooled to 0°C and 5 ml saturated aqueous KHZPOu and 10 m1 arpentane were added. The organic phase was dried over K2C03 and analyzed by GLC (column temperature 100°C for 2 minutes, 150°C for 3 minutes, 170°C for 3 minutes, then 2“0°C). l,l'- dicarboethoxybicyclohexyl was observed in 13% yield. Ethyl cyclohexanecarboxylate was observed in 29% yield. Ethyl l-cyclohexenecarboxylate was observed in 20% yield. 37 B. Product Analya;a_- GLC analysis were performed on a Varian Model 920 gas chromatograph equipped with a “ ft x 0.25 in stainless steel column packed with 2.5% SE-30 on Chromosorb G. The flow rate was maintained at 1 ml per sec- ond. NMR spectra were determined on a Varian T-60 using tetramethylsilane as internal standard. Mass spectra were taken with a Hitachi/Perkin-Elmer RMU-6. 1,1'—Carboethoxybicyclohexy1 1H NMR (001,): 61.“0 (t,6H); 51.75 (m,l2H); 52.30 (m,8H); 6“.20 (q,“H) MS: m/e 262 (M+). Ethyl l-cyclohexenecarboxylate 1H NMR (001“): 51.15 (t,3H); 51.60 (m,“H); 52.20 (m,“H); 6“.10 (q,2H); 66.80 (m,lH). IV. Reaction of Lithium Ester Enolates with TiClu A. Reaction of Lithio Ethyl Cyclohexanecarboxylate with TiClg - The following procedure is repre- sentative of the dimerization of lithium ester enolates by direct addition of TiClu: A 1.0 M_THF solution of lithio ethyl cyclohexanecarboxylate (5.0 mmole) was prepared at -78°C as previously described. The Dry Ice/acetone bath was re- placed with an ice water bath and 0.605 ml (5.5 mmole) TiClu was added dropwise with a syringe. The reaction was 38 stirred for 20 minutes at 0°C, then allowed to warm to 25°C. agPentadecane (0.69 ml; 2.5 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C and 3.3 ml 5.0 5 aqueous NaOH and 10 ml arpentane were added. The organic phase was dried over K2C03. GLC analysis showed 1,1'—dicarboethoxybicyclohexyl in 6“% yield. B. Reaction of Lithio Ethyl Isobutyrate with TiClu via Inverse Addition - The following procedure is representative of the inverse addition of ester enolates to TiClu: A 50 m1 round-bottomed flask equipped with septum inlet, magnetic stirrer, and mercury bubbler was flushed with argon and immersed in a Dry Ice/acetone bath. Stirring was initiated and 0.605 ml (5.5 mmole) TiClu added. When the TiClu was frozen, 5 m1 THF was added at such a rate that the TiClu was not warmed. The cooling bath was then removed and the mixture allowed to warm until a yellow precipitate formed and all visible reaction had ceased. The bath was replaced and the suspension cooled to -78°C. In a second flask a 1.0 M THF solution of lithio ethyl isobutyrate was prepared as previously described. This enolate solution was added dropwise through a Teflon tube to the TiClu/THF suspension over the period of 15 minutes. The cooling bath was removed and the reaction allowed to reach 25°C over the period of “5 minutes. E?015H32 was added as an internal GLC standard. The reaction was cooled to 0°C and treated with 3.3 ml 5.0 N_NaOH and 10 m1 39 aepentane. The organic phase was dried over K2C03 and analyzed by GLC as previously described. Diethyl tetra- methylsuccinate was observed in “1% yield. C. Product Analysis - Products were analyzed by GLC, NMR and MS as previously described. IR spectra were determined on a Perkin-Elmer 327—B grating spectrophotom- eter using polystyrene film as reference. Diethyl tetramethylsuccinate 1H NMR (001“): 51.35 (5,12H); 61.“3 (t,6H); 5u.25 (QauH)° MS: m/e 230 (M+) Diethyl 2,3:diphenylsuccinate 1H NMR (00013): 51.0 (t,3H); 51.3 (t,3H); 63.6-“.“ (complex m,6H); 66.8-7.“ (m,lOH) MS: m/e 256 (M+) 1,1'-Dicarboethoxybicyclobutyl 1H NMR (001“): 51.2 (t,6H); 51.5-2.5 (m,l2H); 5“.0 (q,“H) IR (001“): 1730 cm”1 (0=0) MS: m/e 208 (M+) “0 1,1'-Dicarboethoxybi0yclopropyl 1H NMR (991“): 50.0-1.l (m,8H); 51.2 (t,6H); 5u.0 (q,“H) IR (001“): 1730 cm“1 (0=0) MS: m/e 178 (M+) l,2-Dicarboethoxycyclohexane 1H NMR (001“): 51.1 (t,6H); 51.5 (m,“H); 51.8 (m,“H); 02.6 (m,2H); 6“.1 (q,“H) IR (001“): 1730 cm“1 (0=0) MS: m/e 228 (M+) Di—tert-butyl succinate 1H NMR (001”): 51.“ (s,18H); 52.3 (s,“H) V. Attempted Observation of Ethylene in the Reaction of Lithio Ethyl Isobutyrate with TiClQ A 1.0 M THF solution of lithio ethyl isobutyrate was treated with TiClu as previously described. The re- action flask was sealed, 2 ml aypropane added as internal GLC standard, and the cooling bath removed. After the reaction had warmed to 25°C, a 1 m1 aliquot of the gas above the reaction mixture was withdrawn and analyzed by GLC using a 20 ft x 0.125 in stainless steel column packed “1 with Duropak (n-octane on Porasil) maintained at 60°C. Ethylene was identified by co-inJection with an authentic sample, but less than 1% ethylene was found. As a control, an enolate solution containing no TiClu was allowed to warm to 25°C in a sealed flask under the same conditions. The gas above the reaction mixture of this flask was ob- served to contain approximately the same (less than 1%) amount of ethylene. VI. Investigation of Ethyl a-Chloroisobutyrate as an Intermediate in the Reaction of Lithio Ethyl Isobutyrate with TiCly A. Attempt to Observe Ethyl d—Chloroisobutyrate in the Reaction of Lithio_Ethyl Isobptyrate with TiClQ - A 1.0 M THF solution of lithio ethyl isobutyrate (5.0 mmole) was prepared as previously described and added dropwise through a Teflon tube to a suspension of 1.21 ml (11 mmole) TiClu in 10 ml THF, pre- pared as described in the procedure for inverse addition of enolates to TiClu. The reaction was allowed to warm to 25°C over the period of “5 minutes. GLC analysis of aliquots of the reaction mixture showed no ethyl u-chloro- isobutyrate. Authentic ethyl a-chloroisobutyrate (0.71 ml; 5.0 mmole) was added to the reaction. The amount of ethyl a-chloroisobutyrate observable by GLC did not diminish during one half hour of stirring at 25°C. “2 B. Reaction of Lithio Ethylylgpbutyrate with TiClu in the Presence of Ethyl d-Chloroiso— butyrate - A 1.0 M_THF solution of lithio ethyl isobutyrate (5.0 mmole) was prepared as previously described and treated at -78°C with 0.71 ml (5.0 mmole) ethyl a-chloroisobutyrate followed by 0.605 ml (5.5 mmole) TiClu. The reaction was allowed to warm to 25°C over “5 minutes. a7C15H32 was added as an internal GLC standard and the reaction quenched as usual with aqueous NaOH and extracted with arpentane. The organic phase was analyzed by GLC (column temperature: 60°C for 3 minutes, 100°C for 3 minutes, then 170°C). Diethyl tetramethylsuccinate was observed in 59% yield and ethyl a-chloroisobutyrate was observed in 79% yield. 0. Reaction of Lithio Ethyl 8,8,B-trideuteroiso- butyrate with TiClu in the Presence of Ethyl d-Chloroisobutyrate - A 1.0 M THF solution of lithio ethyl B,8,B-trideuteroisobutyrate was prepared by the method previously described. Ethyl a-chloroiso- butyrate (0.71 ml; 5.0 mmole) was added at —78°C, followed by 0.605 ml (5.5 mmole) TiClu. The reaction was allowed to warm to 25°C over “5 minutes, then worked up with aqueous NaOH and pentane as usual. Diethyl tetramethyl- succinate and ethyl a-chloroisobutyrate were isolated from the organic phase by preparative GLC and analyzed by mass “3 spectrometry. The mass spectra for ethyl u-chloroiso- butyrate showed no deuterium incorporation. Mass spectra for diethyl tetramethylsuccinate showed some product con— taining six deuterium atoms and some containing three deuteriums, in a 2.5:1 ratio. VII. Attempts to Observe Ethyl a-Deuterocyclohexane— carboxylate A. Reaction of Lithio Ethyl Cyclohexanecarboxylate with TiClu in Perdeuterated THF - A 1.0 M THF solution of lithio ethyl cyclohexanecarboxylate (5.0 mmole) was prepared at -78°C as previously described, except that THF d8 was used as solvent. This solution was treated with 0.605 ml (5.5 mmole) T101“ and allowed to warm to 25°C. The reaction was then cooled to 0°C and treated with 5 ml H20, followed by pentane extraction. Ethyl cyclohexane- carboxylate was isolated from the organic phase by prepara— tive GLC and analyzed by mass spectrometry. No deuterium incorporation was observed. The same results were obtained when the reaction is carried out at 0.1 M in THF d8. B. Reaction of Lithio Ethyl Cyclohexanecarboxylate with TiClu Quenching with 929 - A 1.0 M_THF solution of amine-free lithio ethyl cyclohexanecarboxylate (5.0 mmole) was treated with 0.605 ml (5.5 mmole) and allowed to warm to 25°C as previously described. The ““ reaction was then cooled to 0°C, treated with 5 m1 D20, and extracted with pentane. Ethyl cyclohexanecarboxylate was isolated from the organic phase by preparative GLC and analyzed by mass spectrometry. No deuterium incorporation was observed. VIII. Dilution Study of Transition Metal-Promoted Coupling of Lithio Ethyl Cyclohexanecarboxylate A 1.0 M_THF solution of lithio ethyl cyclohexanecarboxyl- ate was prepared as usual and maintained at -78°C while being diluted to the desired molarity (1.00, 0.10, or 0.01 M). The transition metal salt was added and the cooling bath removed. The reaction was then allowed to warm to 25°C (“5 minutes for 1.00 M reactions, 2 hours for 0.10 M reactions, and 5 hours for 0.01 M reaction). £7015H32 was added as internal GLC standard. The reaction was cooled to 0°C, quenched (5.0 M NaOH for TiClu; saturated KHZPO“ for FeC13; 5.0 M NHuOH for CuCl2)and extracted with pentane. The organic phase was dried over K2003 and analyzed by GLC for ethyl cyclohexanecarboxylate, ethyl 1—cyclohexene- carboxylate, and 1,1'-dicarboethoxybicyclohexyl. CHAPTER III DIMERIZATION OF a-SUBSTITUTED KETONE ENOLATES Introduction Since copper—promoted dimerizations of carbanions stabilized by sulfonyl,15 phosphoryl,15 imidoyl,15 and 16 alkoxycarbonyl groups have been reported, it would seem that the oxidative dimerization of ketone enolates is the most straightforward approach to the synthesis of sym— metric l,“-diketones. However, prior to 1975 no success- ful dimerization of ketone enolates by metal salts had been reported. In 1975 Saegusa and co-workers reported17 that lithium ketone enolates, generated ap_situ from ketone and LDA in THF, are dimerized upon treatment with CuCl2 in dimethylformamide (DMF) (Eq. 27). Methyl ketones were OLi THF 2 | + 2 CuCl /DMF—>R000H CH COR (27) 2 2 2 R-C=CH2 dimerized to l,“-diketones in excellent to moderate yields, but increasing alkyl substitution at the coupling site led to "a remarkable reduction in the yield."17 For example, acetophenone was dimerized in 95% yield “5 40 while propiophenone was dimerized under the same condi— tions in 28% yield. Isobutyrophenone was dimerized in only 2% yield. Saegusa and coworkers reported the same trend in yields with the oxidative dimerization of silyl enol ethers by Ag20.zu We aspired to develop a transition metal-promoted coup- ling reaction for alpha substituted ketone enolates which would serve as a complementary method to Saegusa's CuClZ/DMF procedure, Just as TiClu is a complementary method to CuBr2 for the coupling of ester enolates. We chose 3-pentanone as our preliminary model of a substituted ketone. From the ketone itself we prepared the papprbutyl imine, the oxime, and the dimethyl hydrazone. We planned to attempt coupling of the anions of these four species (the dianion in the case of the oxime) with CuC12, FeCl3, and TiClu; the transition metal salts which had proven most effective for coupling ester enolates. Results 3-Pentanone was dimerized with a series of transition metal salts. The results are summarized in Table VII. TiClu yielded no 1,“-diketone but gave about equal amounts of dehydrated Aldol product and a symmetric furan which, could be formed by acid-promoted cyclization of the product diketone (Eq. 28). “7 Table VII. Dimerization of Lithio 3-Pentanone by Various Transition Metal Salts. 9L1 THF 8 n 2 CH3CH2C=CHCH3 + 2 MXn :;E::;;?CH3CH2CCH(CH3)CH(CH3)CCH2CH3 Metal Salt Yield (7)3"c 011012 “9 CuCl2 (amine free) trace Fe013 73 FeCl3 (amine free) “7 T101“ 0b AgOCOCH3 33 AgNO3 33 aProduct obtained as a mixture of al_and meso diasteriomers. bProduct apparently cyclizes to symmetrical dimethyl diethylfuran. cheld based on GLC analysis. “8 0L1 TiClu 0H2 =CHCH3 ——-——;>0H30H2000H0H3 + CH3CH2COCHCH3 CHBCHZCOCHCHB CH3CH2CCH20H3 OH O CH3CH2 \ / CH2CH3 (CH3CH2)2C=C(CH3)COCH2CH3 CH3 CH3 (28) CuCl2 without DMF dimerized the enolate of 3-pentanone in “9% yield. The major by-product was an unsymmetrical furan which must somehow be produced from the aldol product (Eqs. 29, 30). CuCl CH3CH2C=CHCH3 -EEE§>CH3CH2COCH(CH3)CH(CH3)COCH2CH3 O CH3CHf:§L-22:CH3 0H3 CH2CH3 (29) “9 02H5 ,// CH3 CH 02H5 Cu C2H5 \\ // CH3 CH3 2H5 (30) Anhydrous FeCl3, however, dimerized the enolate of 3- pentanone in 73% yield. The product is a mixture of al and meso diasteriomers. The highest yield is obtained by dissolving the FeCl in THE (0.5 M) and adding this solu- 3 tion dropwise to the enolate solution at -78°C, then allow- ing the reaction to warm to room temperature. The reaction is then quenched with a saturated aqueous solution of mono- basic potassium phosphate which, upon addition of pentane, produces a colorless organic phase over an aqueous layer containing a fine grey precipitate. Quenching with aqueous HCl does not appear to lower the yield but does not remove all of the dissolved iron from the organic phase. Removal of the diisopropylamine from the enolate prior to addition of the metal salt decreases the yield of coupled product from 73% to “7% in the case of FeCl3 and from “9% to almost zero in the case of CuClz. When the FeCl3-promoted dimerization of the enolate of 3—pentanone was carried out in the presence of anisole, the yield of l,“-diketone fell to 65% but no adduct with anisole was observed. 50 The papprbutyl imine of 3-pentanone was treated first with hzhutyllithium to form the anion then with TiClu at -78°C. The reaction was allowed to reach 25°C and quenched with NaOH. Although a mixture of products was obtained, more than 50% of the starting imine was recovered. Treat- ment of the anion with FeCl3 gave no recovered starting material but a complex mixture of products, none of which appeared to be formed in more than 20% yield. Since 3- pentanone itself gave good yields of dimerized ketone, further work with its derivatives was abandoned. We decided to try the dimerization of two other sub- stituted ketones: cyclohexanone and 2,“-dimethy1-3-penta- none (diisopropyl ketone). Fe013 dimerized cyclohexanone enolate in only 31%. CuC12 and TiClu dimerized cyclo- hexanone enolate in 30% yield each. Interestingly enough, Saegusa's method, CuClz/DMF, gave only a 15% yield. Quenching the T101“ reaction with aqueous NaOH showed some reduced titanium (blue color) but considerable amounts of unreduced titanium (TiO2 precipitate) remaining. At this point, we began a general survey of the transi- tion metal salts that were at hand. These results are shown in Table VIII. Anhydrous silver acetate proved to be the best coupling agent for cyclohexanone. The yield of l,“-diketone was 68% when the enolate solution was added to a solution of silver acetate in THF at -78°C, kept at -78°C for thirty minutes, and then allowed to 51 Table VIII. Dimerization of Lithio Cyclohexanone by Various Transition Metal Salts. OLi O 2 s + 2Mxn “(gin > (3 Metal Salt Yield (7)5"c FeCl3 31 FeCl3 (amine-free) 30 T101“ 30 TiClu (amine—free) 7 CuCl2 30 CuClZ/DMF 15 ZrClu 0 MnBr2 0 Co(acac)g 23 Cr(a0ac)§ 0 CrCl3 0 AgOCOCH3 68 AgNO3 “3 AgClOu 30 aProduct obtained as a mixture of stereoisomers. bacac=2,5-pentadionate. °Yie1d based on GLC analysis. 52 warm to room temperature over the period of about one hour. Anhydrous silver nitrate and silver perchlorate gave lowered yields. We attempted to dimerize the enolate of 3-pentanone with silver salts but found them to be inferior to FeC13 (Table VII). The enolate of diisopropyl ketone was dimerized with CuC12, FeC13, TiClu, and silver acetate. The results are summarized in Table IX. Once Again, FeCl3 gives by far the best yield of l,“-diketone. Also, once again CuC12 in THF gives as good or better yields than Saegusa's CuC12/ DMF method. Discussion With the exception of cyclohexanone, iron (III) chloride apparently dimerizes lithium enolates of alphafsubstituted ketones in good to excellent yields. Comparison of 3— pentanone to diisopropyl ketone indicates that the yield increases as alpha substitution increases. Since we were able to couple ketone enolates generated lh_alpa from ketones plus LDA in THF, we were able to dimerize ketones in one pot in approximately two hours. There was, therefore, no interest in pursuing the couplings of anions of the imines, oximes, and hydrazones. Prelimin- ary studies showing that TiClu did not dimerize ketone enolates were done by the optimal procedure for dimerizing 53 Table IX. Dimerization of Lithio 2,“-Dimethyl-3—Pentanone with Various Transition Metal Salts. OLi 0 2 .__ + 2 Mn A/ C) Metal Salt Yield (76)8 AgOCOCH3 35 CuC12 21 CuC12/DMF l8 FeCl3 91 TiClu 2“ aYield based on GLC analysis. 5“ ester enolates, llal, direct addition of TiClu to the eno- late. We found that inverse addition of ketone enolate solutions to TiClu already complexed with THF gave some coupling, though the yields were poor. While the yield with CuCl2/DMF did decrease dramatically with increasing alpha substitution, we found that CuCl2 in THF gave better yields in our systems. Most importantly, we discovered that reports18 indicating that F6013 would not appreciably dimerize ketone enolates were misleading because they were based upon results obtained with cyclohexanone. Saegusa repeatedly makes the point that "the use of DMF as a cosolvent is crucial in the oxidative dimeriza- tion of ketone enolates."l7 Our findings indicate the opposite. It is conceivable that CuCl2 dimerizes substi- tuted ketone enolates yla a different mechanism than the one operating for nonsubstituted enolates and that the results cannot be meaningfully compared. However, an alternative explanation may be found in Saegusa's experi- mental technique. Saegusa generates LDA by adding 27 butyllithium in hexane solution to diisopropylamine in THF, obtaining a THF solution of LDA containing considerable amounts of hfhexane. Our own procedure results in a THF solution of LDA containing no hfhexane. It is possible that Saegusa found it necessary to use DMF as a co-solvent only to offset the deleterious effects of hexane on both enolate and CuCl2 solubility. It is not clear why DMF 55 lowers the yields with our substituted ketone enolates though it may compete for coordination sites on the copper. It does seem clear, however, that molecules capable of complexing metals play a crucial role in the reaction. Saegusa found that, for whatever reason, DMF was essential in his system. We find that removal of the diisopropyl- amine from the enolate before reaction with the metal drastically lowers the yield for CuC12, FeCl3, and TiClu. Though the exact role of the amine is unknown, it may be important either as a ligand for an intermediate metal species or as a complexing agent for Li+ at some stage in the reaction. While the mechanism is unknown, the following comments are offered. There seems to be no reason to consider oligomeric bridged intermediates. The optimum yield for all three metals is obtained by dissolving/suspending the metal in THF prior to contact with the enolate. The order of addition does not appear to be important. During the course of the reaction color changes are observed but no phase change or large change in viscosity is apparent. The reaction is sensitive to the presence of benzophenone and lithium butoxide though no direct evidence of free radicals has been observed. The yield apparently increases with increasing alkyl substitution at the alpha carbon of the enolate. This may be rationalized by a radical-type mechanism which requires substituents on the alpha carbon 56 to stabilize radical character at that site. The by- products of the reaction are starting ketone, aldol pro- duct, and intractable high—boiling material. The latter may result from polymerization of hypothetical unsaturated ketone, which is never directly observed, or from condensa- tion of the enolate with the product l,“-diketone. The latter possibility, along with the fact that aldol product is observed, suggests that the more substituted ketone enolates may give higher yields simply because they are more sterically hindered and therefore undergo self- condensation and condensation with the product at a much slower rate. There is some evidence suggesting that the factor which limits the yield is decomposition of the enolate. When the coupling agent is TiClu, one can see qualitatively how much unreduced titanium remains after the reaction is quenched. After coupling a ketone enolate and quenching with NaOH, one observes much more titanium dioxide than blue reduced titanium. When the coupling agent is FeCl3, one can obtain high yields of coupling based on the amount of starting iron (III) if a lO-fold excess of ketone (or ester) enolate is used. This suggests that the product distribution is the result of a competition between oxidative coupling of the enolate and decomposi- tion of the enolate, possibly yla interaction with the metal as a Lewis acid but without reducing the metal. All of the data are consistent with either the 57 dimerization of two free radicals, the original ketyl mechanism suggested in the Introduction of Chapter II (Scheme I), or a two-electron oxidation mechanism similar to the one suggested18 for the FeCl3—promoted cross-coupling of pinacolone and tert-butyl acetate enolates (Scheme II). III OLi OFe 0 R t 0/+ F6013 R t 0’ R 8 0/ + F II -= .__ +_... '-—‘—€> \ e 8 + \0-E-— + xflx :0/. FeIII R-0-0 / - R R ‘/C\.‘7’R -—€> 0- 1" 1 0 /H\ \0’ R 3 ' ' a II R ’,0; + R-C-C-C- -R + Fe L OFeIII II SCHEME I 58 OLi O - I ’ FeCl3 ( $1 3 III a / II R-C=C R- :0 Fe R- -0. + p \——9 \ + \ e t 2 2 - 0 - -—————€> (81h2032FeIV F / (R—0=0 )FeIII \ Reductive elim I I I R-C—C-C- -R + FeII I 8.0 t 0? 2 <9“ 0 0 III |\-} Reductive H l I “ Fe (R-CiC )2FeIII a» R-C-C-C-C-R + FeI---€> 2 FeII \ elim I I SCHEME II There is, of course, no reason to rule out a concerted- type mechanism (Scheme III) similar to the one discussed in Chapter II. As a final comment about the transition metal-promoted dimerization of enolates, we caution that the mechanism or mechanisms actually in operation may well be somewhere in between the models we have suggested. If one conceives the ketyl mechanism, for example, so that the enolate lead- ing to the radical is an iron (III) enolate and the enolate 59 capturing the radical to lead to ketyl is an iron (III) enolate, the ketyl mechanism begins to look very much like the concerted mechanism (Scheme IV). Similarly, if an iron (III) enolate is a precursor to free radicals, then the free radical dimerization mechanism also begins to look like the other two (Scheme V). FeIII R \WWO‘) -——> RM}? + 2FeII SCHEME III Fiill FeII R' 0770 “KY/R "“0 \ __€> rv R/%HR' 0 (LS. FJZDO/FeIII —9 \\ FeIII R FeII 2‘0 R' Fe‘I‘I W "O R' R —9 MR' R t R 0111,6111 b R, O~‘~~ FeII SCHEME IV 60 self Fen RI R' U00 RCH R O R O 0 J2 ~> f.) __) R R R DCHR' 0 III R ' 0 Fe ‘a II R' C) RI Fe SCHEME V Experimental I. Materials heButyllithium, diisopropylamine, solvents, and transi- tion metal compounds were obtained and handled as des- cribed in Chapter II. All ketones were obtained commer- cially and distilled prior to use. The pahpfbutyl imine of 3-pentanone was prepared from the ketone, pappgbutylamine, and TiClu by the method of Weingarten, Chupp, & White.21 The corresponding oxime was prepared by treatment of the ketone with hydroxylamine hydrochloride and Na2CO3 by the method of Bousquet.28 The N,N-dimethylhydrazone was pre- pared by treatment of the ketone with unsymmetrical di- methylhydrazone in ethanol by the method of Newkome and Fishel.35 Silver salts were purchased from Matheson, Coleman & Bell, and used without further refinement. 61 II. Preparation of Ketone Enolate Solutions A. Preparation of Lithio 3:Pentanone - The follow- ing procedure is representative of the preparation of solutions of ketone enolate containing amine: A 1.0 M THF solution of LDA (5.5 mmole) was prepared as described in Chapter II and cooled to 0°C in an ice water bath. 3-Pentanone (0.53 ml; 5.0 mmole) was added dropwise. The reaction was then stirred for 15 minutes at 0°C, giving a 1.0 M_THF solution of ketone enolate containing one equivalent of amine. B. Preparation of Amine-free Lithio cyclohexanone - The following procedure is representative of the prepara- tion of amine-free ketone enolate solutions: A 50 ml round— bottomed flask equipped with septum inlet, magnetic stirrer and mercury bubbler was flushed with argon, immersed in an ice water bath and charged with “.97 ml (8.18 mmole) 2? butyllithium solution and 3.2 m1 hfpentane. Stirring was initiated and 1.16 ml (8.18 mmole) diisopropylamine was added dropwise. After stirring for 15 minutes at 0°C, 0.52 ml (7.““mmole) cyclohexanone was added dropwise. The reaction was stirred for 15 additional minutes, then the solvent and amine were removed under vacuum. When most of the solvent had evaporated, the cooling bath was replaced by a warm (approximately “0°C) water bath to facilitate the evaporation of the remaining solvent and amine. 62 Lithio cyclohexanone was obtained as a white powder which was dissolved in 8.18 ml THF to provide a 1.0 M_THF solu- tion of amine-free enolate. III. Dimerization of Lithium Ketone Enolates hy CuII, FeIII, and TiIV A. FeCl3-Promoted Dimerization of Lithio 3- Pentanone - The following procedure is repre- sentative of the iron (III) promoted dimerization of lithium ketone enolates: A 1.0 M THF solution of lithio 3-penta- none (“.5 mmole) was prepared as previously described and cooled to -78°C in a Dry Ice/acetone bath. A second round- bottomed flask equipped with septum inlet, mercury bubbler and magnetic stirrer was flushed with argon and charged with 0.66 g (“.6 mmole) FeC13 and 9.2 ml THF. Stirring was initiated. After all the FeCl3 had dissolved, the FeCl3/ THF solution was added dropwise to the enolate solution‘ through a Teflon tube. The reaction was stirred for 30 minutes at -78°C then allowed to warm to 25°C over the period of “5 minutes. hrC15H32 (0.635 ml, 2.3 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3 m1 saturated aqueous KH2POu, and extracted with pentane. The organic phase was dried over K2003 and analyzed by GLC (column temperature: 100°C for 3 minutes, 150°C for “ minutes, then 170°C). “,5-Dimethyl- 3,6-heptanedione was observed in 73% yield as a mixture 63 of two stereoisomers. B. TiCly-Promoted Dimerization of Lithio cyclo- hexanone - The following procedure is representa- tive of the reactions of lithium ketone enolates with TiClu: A 1.0 M THF solution of lithio cyclohexanone (5.0 mmole) was prepared by the method previously described for lithio 3-pentanone and cooled to -78°C in a Dry Ice/acetone bath. A second 50 ml round-bottomed flask equipped with septum inlet, magnetic stirrer and mercury bubbler was flushed with argon, immersed in a Dry Ice/acetone bath and charged with 0.605 ml (5.5 mmole) TiClu. When the T101“ had frozen, 5 ml THF was added and stirring was initiated. The cooling bath was then removed and the reaction allowed to warm until a yellow precipitate had formed and all visible reaction had creased. The cooling bath was replaced and the suspension allowed to reach -78°C. The enolate solu- tion was added to the suspension through a Teflon tube, stirred for 30 minutes at -78°C, then allowed to warm to 25°C over the period of “5 minutes. (0.69 ml; E"‘315HB2 2.5 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3.3 ml 5.0 M aqueous NaOH, and extracted with pentane. The organic phase was dried over K3CO3 and analyzed by GLC (column temperature: 100°C for “ minutes, 170°C for 5 minutes, 220°C for 3 minutes, then 2“0°C). Bicyclohexyl-2,2'-dione 6“ was observed in 30% yield. 0. QanQ-Promoted Dimerization of 2,“-Dimethy1-3: Pentanone - The following procedure is repre- sentative of the reactions of lithium ketone enolates with CuC12: A 1.0 M THF solution of lithio 2,“-dimethy1-3- pentanone (5.0 mmole) was prepared as previously described and cooled to -78°C in a Dry Ice/acetone bath. A second 50 ml round-bottomed flask, equipped as described in the preceding procedure, was flushed with argon, charged with 0.7“ g (5.5 mmole) CuC12 and 5.0 m1 THF and cooled to -78°C in a Dry Ice/acetone bath. The enolate solution was added to the CuClz/THF mixture through a Teflon tube. The reaction was stirred at -78°C for 30 minutes then al— lowed to warm to 25°C over the period of “5 minutes. hfcl6H3u (0.773 ml; 2.5 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3.3 ml 3 M HCl and extracted with pentane. The or- ganic phase was dried over K2003 and analyzed by GLC (column temperature: 150°C for four minutes, then 200°C). 2,“,“,5,5,7-Hexamethy1-3,6-octanedione was observed in 22% yield. D. Product Analysis - Products were analyzed by GLC, NMR, IR, and mass spectrometry as described in Chapter II. 65 “,5-Dimethy1-3,6-octanedione (obtained as a mixture of stereoisomers) 1H NMR (001“): 51.0 (m,l2H); 52.1-2.6 (m,“H); 52.6-2.9 (m,2H) IR (CClu): 1710 cm.1 (C=O) MS: m/e 170 (M+) Bicyclohexyl:2,2'-dione36 1H NMR (001“): 51.2-2.1 (m,l2H); 52.2-2.6 (m,“H); 52.6-3.l (m,2H) IR (neat): 170“ om‘1 (0:0) MS: m/e l9“ (M+) 2,“,“,5,5,7rHexamethyl-3,6-octanedione 1H NMR (001“): 51.0 (d,12H); 51.1 (s,12H); 52.9 (septet,2H) IR (001”): 170“ om’l (0=0) MS: m/e 226 (M+) 3-Ethy1—“-methyl-3-heptene—5-one 1H NMR (001“): 51.0 (t,9H); 51.8 (s,3H); 52.1 (q,“H); 52.5 (q,2H) IR (001“): 1690 cm”1 (0:0); 1605 om‘l (C=C) MS: m/e 15“ (M+) 66 2,5—Diethy1-3,“-dimethylfuran 1H NMR (001“): 51.0 (t,6H); 51.8 (s,6H); 52.“ (0.“H) MS: m/e 152 (M+) IR shows no C=O 2,“-Diethyl,3,5-dimethy1furan 1H NMR (001“): 51.1 (0,6H): 51.8 (s,3H); 52.1 (s,3H); 62.3 (quintet, “H) MS: m/e 152 (M+) IR shows no C=O IV. Reaction of Ketone Enolates with Other Metal Compounds The enolates of 3-pentanone, cyclohexanone, and 2,“- dimethy1-3-pentanone were reacted with various silver salts. Lithio cyclohexanone was also reacted with chromium (III), cobalt (III), manganese (II) and zirconium (IV) compounds. The following procedure, for the reaction of lithio cyclohexanone with silver acetate, is representative of all of these reactions: A 50 ml round-bottomed flask, equipped with a septum inlet magnetic stirrer, and mercury bubbler was flushed with argon, immersed in a Dry Ice/ acetone bath, and charged with 0.637 g (3.82 mmole) AgOCOCH3 and 3.9 m1 THF. The mixture was stirred at -78°C for 10 minutes. A 1.0 M_THF solution of lithio cyclohexanone (3.82 mmole) was prepared as previously described, cooled to -78°C, and added to the AgOCOCH3/THF mixture through a Teflon 67 tube. The reaction was stirred at -78°C for 30 minutes then allowed to warm to 25°C over “5 minutes. Ef015H32 (0.998 ml; 3.6 mmole) was added as an internal GLC stan- dard. The reaction was cooled to 0°C, treated with 3.3 m1 3 M HCl, and extracted with pentane. The organic phase was dried over K2003 and analyzed by GLC as previously described. Bicyclohexy1-2,2'—dione was observed in 68% yield. CHAPTER IV REACTIONS OF ENOLATES WITH a-HALOESTERS Introduction In 19“1, Hauser reported31 the iodine—promoted dimeriza- tion of sodio ethyl isobutyrate in very poor yield. Later he reported32 dimerizing the anions of either the ethyl ester or the nitrile of phenylacetic and diphenylacetic acids, using various vicinal alkyl dihalides or polyhalides. More recently, Brocksom and Petragnani22 have dimerized lithium ester enolates using elemental iodine (Eq. 31). The iodine is added to a THF solution of the enolate at \ ,OLi 12"0'002R 2 /C=C -—€> (31) \ OR -C-C02R I -78°C. The reaction gave excellent yields for isobutyrate esters, and in cases where R equals pappfbutyl, but poor yields with methyl or ethyl esters without alpharbranching. We have observed that the reaction fails with lithio ethyl cyclohexanecarboxylate. Lithium ester enolates have been shown26 to react with iodine or bromine at -78°C in THF to form the corresponding a-iodo or u-bromoester (Eq. 32). Brocksom22 has identified a-iodoester as the initial 68 69 \ . /0=0 .__; 1-0-002R (32) product formed in his reaction (Eq. 31) and suggests that the reaction proceeds yla a substitution of the iodine by another enolate anion (Eq. 33), citing successful alkylation of ester enolates by d-haloacetates33 and suc- cessful intramolecular substitutions of secondary a-halo- 3“ esters by ester enolates. If this were true, there I \ ,OLi I I I-C-COZR + /C=C\OR + R02000002R (33) would seem to be no reason why the iodoester, generated lh alpa by reaction of iodine with an ester enolate, could not be reacted in the same pot with a different ester enolate. We felt that this could be the basis for a simple preparation of unsymmetrical l,“-dicarbony1 compounds and so proposed to react a-haloesters with both ester and ketone enolates to synthesize, respectively, unsymmetrical succinates and Y-ketoesters. We proposed to first investigate whether unsymmetrical succinates could be made by adding one equivalent of eno- late to one equivalent of elemental halogen, allowing the u-haloester to form, then adding a second equivalent of a different enolate. We chose as our model esters ethyl 7O isobutyrate and tapprbutyl acetate. Since we felt that Y-ketoesters were more useful synthetic intermediates than succinates, we decided to devote most of our efforts to studying the reaction between ketone enolates and d-haloesters. We chose to begin with the reaction of ketone enolates with iodoacetates to yield y-ketoesters which would be precursors to cyclo- pentane-l,3-diones (Eq. 3“). (D I ?L1 I 8 I base H0-0=0- + ICH2002R -> HC-C-(lJ-CH2002R——> C) (3“) We chose cyclohexanone as our initial ketone because of the importance of the six-member ring in naturally occurring ketones and because the product was available from the cross-coupling of the enolates with Fe013 18 (Eq. 35). Thus not only was the product well characterized OLi C) OLi I CH2_—-< + ‘ 2FeCl3\ . 0020(0113)3 0+ 7 (35) 71 but there was an alternative procedure of roughly equal expense and convenience with which to compare results. Results We began by comparing the bromine and chlorine-pro- moted dimerization of lithio ethyl isobutyrate with the results of Brocksom reported for iodine. We found that the yield dropped remarkably as the halogen changed, as shown in Table X. We also found that lithio pappfbutyl acetate reacted with papprbutyl bromoacetate to yield di- pahpfbutyl succinate in only ““% yield (Eq. 36); consider- ably less than the 97% Brocksom had reported for the iodine- promoted dimerization (Eq. 37). /OLi CH2=C\O+ + Br0H2002+++0000H20H2002+ ““75 (36) OLi / 2 CH2=C\O+ + I2++OCOCH2CH2CO2+ 97% (37) Lithio ethyl isobutyrate reacted with tert-butyl bromoacetate to give an unsummetrical succinate in 80% yield (Eq. 38). 72 Table X. Reaction of Lithio Ethyl Isobutyrate with Ethyl d-Haloisobutyrates. CH I 3 (EH3 CH3\ / 0L1 CH3-C-C02Et X-C —CO2Et + /C=C + | CH CH \OEt CH3"?"702Et 3 3 CH 3 Halogen Yield (%)d I 85a Br “0b 01 ' 0° aReference 22, one equivalent 12 added to 2 equivalents enolate. bOne equivalent Br2 added to 2 equivalents enolate. cChloroester added dropwise to enolate solution. ineld based on GLC analysis. 73 CH3 0L1 >__< + BrCH2C02-|— + EtOCOC(CH3)2CH200é-"— All attempts failed to form the same unsymmetrical succinate from addition of a pappfbutyl acetate enolate to ethyl a-bromoisobutyrate formed 12.2222 by addition of lithio ethyl isobutyrate to bromine in THF at -78°C (Eq. 39). ,OM 0H3 OLi 0H3 \0~f— CH E CH 3 O t L 3 _ —><—> EtOCOC(CH3)2CH2C02-+- M=Li, CoII, CuI (39) However, replacing the bromine in Equation 39 with iodine resulted in a 6“% yield of the unsymmetrical suc— cinate when the lithium enolate of papprbutyl acetate was used. Besides product, the reaction mixture contained pappfbutyl acetate and a trace of iodoester. No sym- metrical succinates were observed. Thus we are effec- tively able to cross-couple two different ester enolates without isolating any intermediates. 7“ Alphaeiodoacetates are not readily available from ester enolates as addition of the enolate solution to iodine results in formation of succinate even at -100°C. Ethyl iodoacetate and pappgbutyl iodoacetate were prepared by treatment of the corresponding chloroesters with sodium iodide in acetone.27 Ethyl iodoacetate was found to be extremely sensitive to heating and air, decomposing with concurrent formation of elemental iodine. Ethyl iodo- acetate were also found to be an extremely potent lachrym- ator, which proved especially inconvenient because it is also essentially odorless. papprButyl iodoacetate, on the other hand, seemed much less inclined to decompose and was much more pleasant to work with. We were able to prepare pahpgbutyl iodoacetate on a 300 millimole scale in 83% isolated yield, taking only routine precautions to degas solvents, washwater, and drying agents. The lithium eno— late of cyclohexanone was added to one equivalent of pappfbutyl iodoacetate to give the corresponding y-keto- ester in excellent yield (Eq. “0). Use of the potassium enolate of cyclohexanone lowered the yield to 3“%. OLi O CO C CH ‘ ”macaw.” —> . 2< :>3 a.) Results of experiments to optimize reaction conditions 75 are shown in Table XI. While the reaction appears to pro— ceed slowly to completion at -78°C, the most convenient procedure is to cool the solution of ketone enolate (pre- pared by addition of ketone to LDA in THF at 0°C) to -78°C, add the iodoester, and allow the reaction to warm to 25°C over the period of approximately one hour. A survey was done of the reaction of various lithium ketone enolates with pahpfbutyl iodoacetate. The results are shown in Table XII. One notes that enolates with alkyl substitution at the alpha carbon give higher yields than the methyl ketone acetophenone. Also, we found that the reaction was somewhat sensitive to excess base. Use of ketone enolate containing ten percent excess LDA lowered the yield from 79% to 75% in the case of 3-pentanone and from 69% to 51% in the case of cyclopentanone. Removal of the diisopropylamine from the enolate prior to addition of the iodoester lowered the yield with 3-pentanone from 79% to 56%, while use of lithium diethylamide or lithium 2,2,6,6-tetramethylpiperidide in place of LDA had no effect upon the yield. Use of excess iodoester improved the yield somewhat, as shown in Table XIII for the case of 3-pentanone. Finally, we contrasted pappfbutyl iodoacetate with other a-haloacetates in the reaction with lithio cyclo- hexanone. The results, summarized in Table XIV, show that the reaction is not very sensitive to the nature of the halogen but is adversely affected by replacement of the tert-butyl group by an ethyl group. 76 Table XI. Reaction of Lithio Cyclohexanone with tert- Butyl Iodoacetate. OLi O l.M CH + 10H2co2c(CH3)3--—€> 2C020(CH3)3 Conditions Yield (7)a THF, -78°0 + 25°C 97 THF, 0°C + 25°C 87 DMF, 0°0 + 25°C 28 DMSO, 25°C 13 THF, -78°C, 3 hours 100 aYield based on GLC analysis. 77 Table XII. Reaction of Ketone Enolates with tert-Butyl Iodoacetate. OLi £ THF, 1.0 M R- =C-+ ICH lo I >R-d--0-01120020(CH3)3 000011 2 2 ( 3)3 -78°C*25°C I Ketone Yield (%)a Cyclohexanone 97 3-Pentanone 79 Cyclopentanone 69 2,“-Dimethy1-3-pentanone 100 Acetophenone “0 aYield based on GLC analysis. 78 Table XIII. Reaction of Lithio 3-Pentanone with Various Ratios of tert-Butyl Iodoacetate. OLi I H EtC-CHCH3 + ICH20020(CH3)3 + CH3CH2CCH(CH3)CH2002C(CH3)3 Ratio of Enolate: Iodoester Yield (%)a 1:1 79 1:1-1/2 82 1:2 96 aYield based on GLC analysis. 79 Table XIV. Reaction of Lithio Cyclohexanone with a-Halo- acetates. OLi O l.M CH 00 R s + XCH2C02R -——>THF 2 2 -78°+25° Haloester Yield (%)a ICH2C02(CH3)3 97 BrCH2C02(CH3)3 91 ClCH2002(CH3)3 87 ICH2C02C2H5 62 aYield based on GLC analysis. 80 Discussion Tertiary a-iodoesters and tertiary d-bromoesters ap- parently do undergo substitution by ester enolates. The reaction appears to be extremely dependent on the nature of the halogen and seems to work best when the alpha carbon of the enolate has alkyl substituents. When lithio pappgbutyl acetate is reacted with ethyl a-iodoisobutyrate, we observe pappgbutyl acetate in the product mixture. The absence of ethyl isobutyrate suggests that the halogen does not exchange (Eq. “1), even though we cannot separate /0L1 0H3 CH3 OLi 0H2-C\O+_ + 1-0 —C02Et 10H2002+ + >.___._.< CH3 CH3 0Et (“1) and distinguish pappfbutyl iodoacetate from ethyl a—iodo- isobutyrate in the trace quantities in which it appears in the reaction. The absence of di—pappebutyl succinate and diethyl tetramethylsuccinate also suggests that the halogen does not exchange. 22 only that he "prefers to consider Brocksom states an 8N2 mechanism" and makes no further comment about the path of the reaction. While it is conceivable that the influence of the carbonyl group speeds up substitution relative to elimination, this does not explain why lithio 81 ethyl isobutyrate gives a higher yield than lithio papa: butyl acetate when reacted with the same tertiary a-halo— ester. One would expect the more hindered enolate to be less nucleophilic and thus favor elimination over substitu- tion. One could rationalize this problem by proposing that the more hindered enolate is less likely to lower the yield by condensation with the product. Arnold29 has recently observed that lithium ester enolates react with carbon tetrachloride and carbon tetra- bromide to give the corresponding a-haloesters plus halo- form (Eq. “2). Meyers had previously reported30 that \ /OLi 1)THF I 0=0 + cxu"——9x-0-002R + ch3 (“2) / \OR 2)H3o+ I ketones and sulfones react with CClu-KOH—pappruOH to give a—chloro products or poly-u-chloro products or their derivatives. Meyers had also observed products derived from dichlorocarbene and thus proposed an electron-transfer mechanism shown in Scheme VI. A similar mechanism can be envisioned for the reaction of ester enolates with a-haloesters and is shown in Scheme VII. 82 \ /OLi \ . C=C + CX + ['C-CO R + CX- + / \OR “ / 2 “J I .. .. X-C-CO2R + .CX3 + X + .CX2 OX“ 2 'CX3 + X- -——> SCHEME VI \0 d/OLi + x d 00 R \0 00 R [x t 00 R]‘ = — -— -) ° — + _ .- I + I + X- SCHEME VII Arnold observed substantial amounts of haloform but no products derived from dihalocarbene and thus propose529 that the trichloromethyl anion is stable under the reaction conditions: enolate, product, and diisopropylamine in THF solution at temperatures up to 25°C. We have observed 83 that trichloromethyl anions decompose in THF solution even at —78°C. This suggests that we consider as an alternate possibility the formation of a different radical-radical anion pair (Scheme VIII). \ /OL1 /C=C\ + CXLI OR , ,OLi I—-——> x-0-0 + -0x I "OH 3 [O] I SCHEME VIII The equivalent pathway for enolate and a—haloester is shown in Scheme IX. While Scheme IX involves a more famil- iar radical anion, the ketyl; Scheme VI, involving elec- tron transfer from the enolate, offers a more attractive explanation for the higher yields with alphaybranched enolates. Whatever the nature of the intermediate, an electron-transfer mechanism is attractive because of the marked difference in reactivity of the various halides and the fact that alphafbranched enolates give the best yield. 8“ I— I '- \ ‘/OLi I 'C'CO2R /9=C\ + I-0-002R + + I' OH ' -C—C-OR I I _ OLi J I -C-002R -> + LII "(E‘COzR SCHEME IX The reaction of ketone enolates with a-haloacetates shows relatively little dependence on the halogen, suggest— ing that perhaps a Simple SN2 mechanism is in operation. While Table XII shows that alphaysubstituted enolates give better yields than acetophenone, one may rationalize that the more substituted enolates are more stable and thus reduce competing enolate decomposition. The more hindered enolate should also be less likely to reduce the yield by condensation with the product y-ketoester. Whatever the mechanism, the reaction is certainly sensitive to the nature of the alkoxy group on the iodo- ester. Brocksom suggests22 that the role of the £332: butyl group is to stabilize the enolate and thus prevent competing enolate decomposition. In our case the tert-butyl 85 iodoacetate is preformed and so no pappfbutyl enolate is involved in the reaction. It seems reasonable to suggest that the pappfbutyl group stabilizes the iodoester, how- ever, and prevents its competing decomposition. It is evident from our preparation of iodoacetate that the EEEEbeLyl group stabilizes the iodoester to heat and air and lessens its lachrymatory potency. Table XIII may be interpreted to suggest that decomposition of the iodo- ester independent of reaction with the enolate is a com- peting process. On the other hand, the pahpfbutyl group may serve only to protect the ester function from attack by the enolate and Table XIII may only indicate that increasing the concentration of iodoester speeds up the desired re- action and thus reduces both competing enolate decomposi- tion and competing condensation of enolate with the product. In conclusion, we have demonstrated that, in principle, unsymmetrical succinates may be conveniently prepared from the corresponding esters without isolation of intermediates. We have shown that tahtfbutyl iodoacetate, unavailable from the corresponding ester enolate, may be prepared in good yield, and stored without apparent decomposition and that this iodoester reacts with ketone enolates to give the corresponding y-ketoesters. Clearly, several directions for further research are indicated. The scope and limita- tion of the reactions of ester enolates with a-iodoesters 86 generated lh alph have not been fully explored. The re- actions of ketone enolates with a-haloesters other than acetate have not been studied. Finally, the most straight forward route to unsymmetrical l,“-diketones would appear to be a substitution reaction between a ketone enolate and an a-haloketone, yet to our knowledge this has not been attempted with lithium ketone enolates prepared from ketones plus LDA. Exparimental 1. Materials A. General - Diisopropylamine, hrbutyllithium, solvents, and esters were prepared or obtained as de- scribed in Chapter II. Potassium hydride, Ventron Corp., was obtained as a 25-30% mineral oil dispersion and stan- dardized by measuring H2 evolved when a known volume was added to water. ‘papprutyl bromoacetate and ethyl chloro- acetate, Eastman, were stored over 5A molecular sieves and used without further purification. Ethyl iodoacetate was prepared by treating ethyl Chloroacetate with Nal/ acetone by the method of Ashworth and Coller.27 Mono- chloroacetic acid, Fisher, was converted to the acid chloride by treatment with two equivalents of benzoyl chloride by the method of H. C. Brown37 and distilled at 105-107°C. tert-Butyl alcohol was stirred overnight 87 over CaO before use. Dimethylaniline was freshly dis- tilled before use. B. Pzeparation of tert-Butyl Chloroacetate38 - A 300 m1 round-bottomed flask equipped with a magnetic stirrer and a dropping funnel protected by a drying tube was charged with 50.5 ml (“00 mmole) N,N-dimethylaniline and “5.“ g (“00 mmole) monochloroacetyl chloride and cooled to 0°C. taptyButyl alcohol (37.5 ml, “00 mmole) was added dropwise over the period of one hour while the temperature was maintained between 10°C and 20°C. The reaction was then placed in an ice water bath and allowed to stir over- night while the bath warmed to 25°C. The reaction was treated with 100 ml H20. The layers are separated and the aqueous phase extracted with 3 x “0 ml diethyl ether. The combined organic phase was washed three times with 2 M HCl then with saturated NaHCO3. The ether was evaporated and the residue distilled at 56-9°C at 13 mm to provide 23.6 g (63%) product. 0. Preparation of tert-Butyl Iodoacetate27 - A 500 m1 round-bottomed flask equipped with septum inlet, magnetic and mercury bubbler was flushed with argon and charged with 300 m1 acetone. Argon was bubbled through the acetone with a gas dispersion tube for 20 minutes to remove all traces of dissolved oxygen. Sodium iodide 88 (“5 g, 300 mmole) was added through a powder funnel and the mixture stirred until the NaI dissolved. happyButyl Chloroacetate (32.5 g, 272 mmole) was added and the reaction stirred at 25°C under argon for 2“ hours. 100 ml H20, degassed with argon in the same manner as the acetone, was added. The solution was extracted with degassed diethyl ether, washed with saturated Na28203 solution to remove any I2, and dried with degassed MgSOu. The ethereal solution was filtered through Celite under argon in a Schlenk funnel. Evaporation of the ether and distillation at 53°C at 2 mm provided 5“.6 g (83%) product. II. Praparation of Ester Enolate Solutions Lithium ester enolates were prepared by the same procedure described in Chapter II. Cobalt (II) and copper (I) enolates were prepared by the reaction of CoBr2 and Cu2I2, respectively, with the lithium enolate.23 The preparation of the copper (I) enolate of happybutyl acetate is representative: A 1.0 M THF solution of lithio EEEEfDUtyl acetate (5.0 mmole) was prepared at -78°C as previously described. Cu2I2 (0.95 g, 2.5 mmole) was added through a powder funnel. The cooling bath was removed and the reaction allowed to warm until the Cu2I2 dissolved and an intense purple solution formed. The cooling bath was replaced and the reaction re-cooled to -78°C. The color remained and Cu212 did not precipitate, providing a 1.0 M 89 THF solution of the copper (I) enolate of tert-butyl acetate. III. Reaction of Ester Enolates with a-Haloesters A. Reaction of Lithio Ethyl Isobutyrate with tert- Butyl Bromoacetate - The following procedure is representative of the reaction of ester enolates with preformed a-haloesters: A 1.0 M_THF solution of lithio ethyl isobutyrate (5.0 mmole) was prepared at —78°C as previously described. pahprButyl bromoacetate (0.73 ml, 5.0 mmole) was added. The reaction was stirred for 5 minutes at -78°C, then allowed to reach 25°C over the period of 1.25 hours. h—CISH32 (1.38 ml. 5.0 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3.3 ml 3 M HCl, and extracted with pentane. The organic phase was dried over K2003 and analyzed by GLC (column temperature: 100°C for “ minutes, 150°C for 2 minutes, 175°C for 5 minutes, then 2“0°C). 2,2-Dimethylsuccinic acid 1-ethy1 ester 2-pappfbuty1 ester was observed in 80% yield. B. Reaction of Lithio tert-Butyl Acetate with Ethyl a-Iodoisobutyrate Formed in situ from 12 and Lithio Ethyl Isobutyrate - The following procedure is representative of the reaction of an ester enolate with an a-haloester formed lh situ from the 90 corresponding ester enolate and molecular halogen: A 50 ml round-bottomed flask equipped with septum inlet, mag- netic stirrer and mercury bubbler was flushed with argon, charged with 1.27 g (5.0 mmole) I2 dissolved in 5 m1 THF, and cooled to -78°C in a Dry Ice/acetone bath. A 1.0 M THF solution of lithio ethyl isobutyrate (5.0 mmole), prepared as previously described, was added dropwise through a Teflon tube. The iodine color was observed to disappear as the enolate solution was added. The reaction was stirred for 15 minutes at -78°C, then a 1.0 M_THF solution of lithio happybutyl acetate, prepared as previously described, was added dropwise through a Teflon tube. The reaction was stirred for 5 minutes at -78°C then allowed to warm to 25°C over 1.25 hours. h_-C15H32 (1.32 ml, 5.0 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C and worked up as described in the preceding procedure. GLC analysis showed 6“% 2,2—dimethy1succinic acid l-ethyl- 2-tert-butyl ester. 0. Product Analysis - Products were analyzed by GLC, NMR, and MS as described in Chapter II. 242-Dimethylsuccinic acid 1-ethy1-2-tert-buty1 ester 1H NMR (001“): 51.20 (s,6H); 51.21 (t,3H); 51.37 (8.9H); 62.35 (s.2H); 6“.00 (9,2H) MS: m/e 230 (M+) 91 IV. Reaction of Ketone Enolates withia-Haloketones A. Reaction of Potassium Ketone Enolates with d-Haloacetates - The procedure for cyclohex- anone and happybutyl iodoacetate is representative: A 50 ml round-bottomed flask equipped with a septum inlet, magnetic stirrer and mercury bubbler was flushed with argon, immersed in an ice water bath and charged with 5 ml THF and 1 ml (5.0 mmole) KH-mineral oil suspension. Stirring was initiated and 0.52 ml (5.0 mmole) cyclohexanone was added dropwise. The reaction was stirred for 15 min- utes at 0°C, then 0.77“ ml (5.0 mmole) papprbutyl iodo- acetate was added. The reaction was stirred at 0°C for 10 additional minutes then allowed to warm to 25°C over the period of one hour. 97016H3“ (1.“6 ml, 5.0 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3.3 ml 3 M HCl and extracted with pentane. The organic phase was dried over K2CO3 I and analyzed by GLC (column temperature: 100°C for 3 minutes, then 175°C). papprButyl 2-oxocyclohexy1acetate was observed in 3“% yield. B. Reaction of Lithium Ketone Enolates with c-Haloacetates - The reaction of lithio 3- pentanone with papa-butyl iodoacetate is representative: A 1.0 M_THF solution of lithio 3-pentanone (5.0 mmole) was prepared as described in Chapter III and cooled to 92 -78°C. papprutyl iodoacetate (0.77“ ml, 5.0 mmole) was added. The reaction was stirred at —78°C for 5 minutes then allowed to warm to 25°C over 1.25 hours. h7C15H32 (1.38 ml, 5.0 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C, treated with 3.3 ml 3 M HCl, and extracted with pentane. The organic layer was dried over K2C03 and analyzed by GLC (column temperature: 100°C for “ minutes, 150°C for 5 minutes, then 175°C). happeButyl 3—methy1-“-oxohexanoate was observed in 73% yield. 0. Reaction of Lithium Ketone Enolates with a-Haloacetates in Solvents Other Than THF - The reaction of lithio cyclohexanone with pappfbutyl iodo- acetate in DMF is representative: Lithio cyclohexanone (5.0 mmole) was prepared from LDA in hypentane and isolated as a white powder as described in Chapter III. The enolate was dissolved in 5 ml DMF. The reaction was cooled to 0°C and 0.77“ ml (5.0 mmole) happebutyl iodoacetate was added. The reaction was stirred at 0°C for 10 minutes then allowed to warm to 25°C. 27016H3“ (1.“6 ml, 5.0 mmole) was added as an internal GLC standard. The reaction was cooled to 0°C and worked up and analyzed as previously described. taptyButyl 2-oxocyclohexylacetate was observed in 29% yield. 93 D. Product Analysis - Products were analyzed by GLC, NMR, and MS as described in Chapter II. tert-Butyl 2-oxocyclohexylacetate18 1H NMR (001“): 51.“ (s,9H); 51.6—2.8 (m,llH) MS: m/e 212 (M+) tert-Butyl 3-methy1-“-oxohexanoate 1H NMR (001“): 51.0 (t,3H); 51.1 (d,3H); 51.“ (s,9H); 52.0-3.0 (m,“H) MS: m/e 200 (M+) tert-Butyl 2-oxocyclopentylacetate (GLC column temp. 200°C) 1H NMR (001“): 51.“ (s,9H); 51.6-2.6 (m,9H) MS: m/e 198 (M+) tert-Butyl 3y3i5-trimethy1-“—oxohexanoate (GLC column temp. 165°C) 1H NMR (001"): 51.05 (d,6H); 51.20 (s,6H); 61.“0 (s,9H); 62.35 (s,2H); 63.0 (septet, 1H) MS: m/e 2A0 (M+) 9“ tert-Butyl “-phenyl-“-oxobutanoate GLC column temp: 100°C for 3 minutes, then 175°C 1H NMR (001“): 51.“ (s,9H); 52.5 (t,2H); 53.1 (t,2H); 67.3 (m,3H); 67.8 (m,2H) MS: m/e 23“ (M+) BIBLIOGRAPHY 100 11. 12. 13. 1“. BIBLIOGRAPHY a) A. P. Dunlap and F. N. Peters, "The Furans", Do 35, Rheinhold Pub. Co., 1955; b) P. Bosshard and C. H. Eugster, Adv. in Hetero- cyclic Chem., 11, 377 (1966); c) P. Bosshard and C. H. Eugster, ibid, 11 38“ (1966). E. Baltazzi and L. I. Krimen, Chem. Rev., 03, 511 (1963). a) T. L. Ho, Synth. Commun., 3, 265 (197“). b) K. Oshima, H. Yamamoto, H. Nazaki, J. Amer. Chem. Soc., 95, “““6 (1973). R. A. Ellison, Synthesis, 397 (1973). Y. Yura and J. Ide, Chem. Pharm. Bull., ll, “08 (1969). J. E. McMurray and J. Melton, J. Amer. Chem. Soc., 23: 5309 (1971). T. Mukaiyama, K. Narasaka, F. Furusato, J. Amer. Chem. Soc., “, 86“1 (1972). E. J. Corey and L. S. Hegebus, J. Amer. Chem. Soc., 21: “926 (1969). T. L. Ho, Synth. Commun., 3, 135 (197“). E. Nakamura, K. Haskmoto, I. Kuwajima, J. Org. Chem., 5;, “166 (1977). P. A. Wehdi and V. Chu, J. Org. Chem., 3Q, 3“36 (1973). A. Pelter, C. R. Harrison, D. Kirkpatrick. Tet. Lett., ““91 (1973). M. W. Rathke and A. Lindert, J. Amer. Chem. Soc., 8. 0. Watson and J. F. Eastham, J. Organometal. Chem., 96 15. 16. 17. 18. 19. 20. 21. 22. 23. 2“. 25. 26. 27. 28. 29. 30. 97 a) T. Kauffman and D. Berger, Chem. Ber., 101, 3022 (1968). b) C. A. Maryanoff, B. E. Maryanoff, R. Tang, K. Mislow, J. Amer. Chem. Soc., 25, 5839 (1973). M. W. Rathke and A. Lindert, J. Amer. Chem. Soc., 23. “605 (1971). Y. Ito, T. Konoike T. Haradu, T. Saegusa, J. Amer. Chem. Soc., 92, 1&87 (1977). ‘— Alan A. Millard, Doctoral Thesis, Michigan State University, 1978, p. “9. H. C. Vogler, W. Brackman, J. W. F. M. Lemmers, Recl. Trav. Chem. Paya. Pas., 83, 1203 (1965). T. Kauffman, at al., Angew. Chem. Int. Ed. Engl., 1. suo (1968). H. Weingarten, J. P. Chupp, W. A. White, J. Org. Chem., 3a, 32“6 (1967). T. J. Brocksom, N. Petragnani, R. Rodrigues, H. LaScala, Synthesis, 396 (1975). J. A. Katzenellenbogen, A. L. Crumrine, J. Amer. Chem. Soc., 28, “925 (1975)- Y. Ito, T. Konoike, T. Saegusa, J. Amer. Chem. Soc., 21. 6A9 (1975). R. P. Woodbury and M. W. Rathke, J. Org. Chem., “2, 1688 (1977). M. W. Rathke and A. Lindert, Tet. Lett., 3995 (1971). J. Ashworth and B. A. W. Coller, Trans. Faraday. Soc., 01, 1077 (1971). E. W. Bonsquet, "Organic Synthesis Coll. Vol. II", p. 313 (19“3). R. T. Arnold and S. T. Kulenovic, J. Org. Chem., M3, 3587 (1978). a) C. Y. Meyers, A. M. Malte, W. S. Mathewes, l, Amer. Chem. Soc., gl, 7510 (1969). b) C. Y. Meyers and L. L. Ho, Tet. Lett., “319 (1972). 30. 31. 32. 33. 3“. 35. 36. 37. 38. 39. 98 c) C. Y. Meyers and V. M. Kolb, J. Org. Chem., M3, 1985 (1978). B. E. Hudson, Jr. and C. R. Hauser, J. Amer. Chem. Soc., Q3, 3156 (19“1). W. G. Kofron and C. R. Hauser, J. Org. Chem., 35, 2085 (1970). W. G. Kofron and L. G. Wideman, J. Org. Chem., 31, 555 (1972). R. N.)MacDonald and R. R. Reitz, Chem. Commun., 90 1971 . G. R. Newkome and D. L. Fishel, J. Org. Chem., 3l, 680 (1966). T. Hiyama, H. Taguchi, S. Fujita, H. Nozaki, Bull. Chem. Soc. Jap., “5, 1866 (1972). H. C. Brown, J. Amer. Chem. Soc., 50, 1325 (1938). R. Baker, "Organic Synthesis Coll. Vol. III", p. 1““ (1975). R. J. Cregge, J. L. Herrmann, C. S. Lee, J. E. Richman, R. H. Schlessinger, Tet. Lett. 2“25 (1973).