.,_-_..-..-.—.,--u¢u-uv}=—ng. - EXPLORATION 0F FACIORSIAFFECTING THE-'15 LIB R A R Y Michigan State University This is to certify that the thesis entitled Exploration of Factors Affecting Asymmetric Induction in Additions to Carbonyls Directly Bonded to Asymmetric Centers presented by Thomas Henry Althuis has been accepted towards fulfillment of the requirements for __P:1'_ILD__.degree in ChemiSt 6‘1/ fitméqfif Major professor Date W 0-169 ABSTRACT EXPLORATION OF FACTORS AFFECTING ASYMMETRIC INDUCTION IN ADDITIONS TO CARBONYLS DIRECTLY BONDED TO ASYMMETRIC CENTERS BY Thomas Henry Althuis The recently proposed Karabatsos model1 for addi- tions to carbonyls directly bonded to asymmetric centers offers not only a means of predicting the major diastere— omer but also the diastereomeric product ratio in these reactions. These predictions are made on the basis of differences in carbonyl eclipsing interactions, M 4-9 0 vs. L 4—9CL inreactant like diastereomeric transition states A and le'z O 3 L“ "‘O --R' RI 2 To examine the usefulness of this model under a variety of variables--nucleophilic reagent, temperature, solvent and Thomas Henry Althuis concentration of reactants--several reactions were in— vestigated at various temperatures: additions of lithium aluminum deuteride in ether and sodium borodeuteride in tetrahydrofuran and 2—pr0panol to 2-phenylpr0panal-2-d and 3-methyl-2—pheny1butanal-2—d; additions of methylmagnesium halides in ether and tetrahydrofuran and methyllithium in ether and pentane to 3-phenyl—2—butanone-l,l,l,3-d4 and 4-methyl—3-phenyl—2-pentanone—l,l,l,3-d4; and additions of benzylmagnesium halides in ether and tetrahydrofuran to l,3-diphenyl-2—butanone—l,l,3-d and 1,3—dipheny1—4-methy1— 3 2—pentanone—l,l-d2. The ratio of equal energy isotopically labelled diastereomeric alcohols was determined from their n.m.r. spectra. Differences in the free energy of the diastereomeric transition states for these and related re- actions were calculated in accord with the Curtin—Hammett . . 3 princ1ple , +__ AAGAB — RTln(A/B). The following observations were made: a) AAGiB values are temperature independent. Thus, entropy effects are negligible; b) product stereospecificity is independent of the concentration of the reactants; c) the stereospeci- ficity of alkylmagnesium halides usually increases as the halogen of Grignard reagents is varied from iodine to chlorine; d) stereospecificity may change with solvent; Thomas Henry Althuis e) stereospecificity of alkyllithiums often differs from that of Grignard reagents. In ethereal solutions additions of lithium aluminum hydride and alkylmagnesium iodides to ketones of the type LMsC—C(CH2R")O give AAGiB values which are in good agree— ment with those predicted by Karabatsos' modell. Increasing the size of R and R' results in greater stereospecificity than predicted by the model. This is explained on the basis of differences in R' 6—9 R, R' 6—9 M and R' 6—9 L interac— tions in the diastereomeric transition states A and B. If R is branched at the alpha position, interactions between it and the substituents on the asymmetric center may become more significant than the M 6—9 0 and L e—e O interactions. J. Karabatsos, J. Am. Chem. Soc., g3, 1367 (1967). 2G. J. Karabatsos and N. Hsi, J. Am. Chem. Soc., fiz, 2864 (1965). 3D. Y. Curtin, Record Chem. Progr., 15, 111 (1954). See also E. L. Eliel, "Stereochemistry of Carbon Compounds,’ MCGraw—Hill Book Co., Inc., New York, 1962, pp. 151—156. EXPLORATION OF FACTORS AFFECTING ASYMMETRIC INDUCTION IN ADDITIONS TO CARBONYLS DIRECTLY BONDED TO ASYMMETRIC CENTERS BY Thomas Henry Althuis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1968 To Rose Marie ACKNOWLEDGMENTS The author expresses his appreciation to Professor G. J. Karabatsos for his inspiration, patience and guidance during the course of this investigation. Special thanks are given to the National Institutes of Health for a Pre-doctoral FellowShip (1967-68) and to the Dow Chemical Company for a Dow Summer Fellowship (1967). Financial assistance also provided by the National Science Foundation and the Petroleum Research Foundation is gratefully acknowledged. iii TABLE OF CONTENTS ACKNOWLEDGMENTS . . . . . . . . . . . . LIST OF TABLES . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . A. Errors in the Calculated AAG+B Values B. Entropy Effects . . . . . . C. Solvent Effects . . . . . . . D. Effect of the Nucleophile . . . . . E. Effect of the Symmetrical R Group of MLsC- CRO . . . . . . . . . . . F. Conclusion . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . A. General Procedures . . . . . B. Preparation of 2-Pheny1propanal—2-d C. N. M. R. Solvent Studies of 2- Phenylpro— panol— 2, 0— —d2 . . . . . . D. Additions to 2—Phenylpropanal-2-d E. Preparation of 3—Pheny1—2-butanone . F. Preparation of 3- -Phenyl- 2— buta— none-1, 1,1,3—d4 . . . . . . . G. N.M. R. Solvent Studies of 2— —Methyl— 3- -phenyl— 2— butanol . . . . . H. Additions to 3-Pheny1—2—buta— none—1,1,1,3—d_4 iv o Page 19 20 25 27 3O 32 32 33 34 34 37 37 38 38 Page I. Preparation of 1,3-Diphenyl-2-butanone . . . 40 Preparation of 1,3-Dipheny1-2-buta- none-1,1’3—g3 o o o o o o o o o 0 o o o o 4]. K. N.M.R. Solvent Studies of 2-Benzyl- l I 3-dipheny1‘2-bUtanOl c o o o o o o o o 41 L. Additions to 1,3-Diphenyl—2—buta- none—1,1,3-é3 o o o o o o o c o o o o o o 42 M. Preparation of 3—Methyl-2—phenylbutanal . . 44 N. Preparation of 3-Methyl-2-pheny1butanal-2-d. 44 O. N.M.R. Solvent Studies of 3-Methyl- 2—phenyl-l-butanol-2, O—d2 . . . . . . . 45 P. Additions to 3— —Methyl- 2-pheny1bu- tanal— 2— d . . . . . . . . . . . . . . 46 Q. Preparation of 4— —Methyl- 3- -pheny1— 2-pentanone . . . . . . . . . . . . . . 46 R. Preparation of 4-Methyl-3—phenyl- 2-pentanone—l,1,l,3-d_4 . . . . . . . . . 47 S. N.M.R. Solvent Studies of 2,4—Dimethyl— 3-phenyl-2-pentanol . . . . . . . . . . . 47 T. Additions to 4-Methy1-3-pheny1- 2-pentanone-1,l,l,3-d_4 . . 48 U. Preparation of 1,3-Diphenyl-4—methyl— 2—pentanone o o o o o o o o o o o o o o o 48 V. Preparation of 1,3-Dipheny1—4—methy1- 2-pentanone-l,l—d . . . . . 49 —2 O O O O O O W. N.M.R. Solvent Studies of 2-Benzy1- l,3—diphenyl—4-methyl-2-pentanol . . . . 49 X. Additions to 1,3-Dipheny1-4-methy1— 2-pentanone-l,l—d2 . . . . . . BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . 51 50 Table II. III. IV. VI. VII. VIII. IX. LIST OF TABLES Page Additions to 2-pheny1propanal and 2-pheny1propanal-2—d . . . . . . . . . . . 10 Additions to 3-phenyl-2-butanone and 3-pheny1-2—butanone-1,1,l,3-_d_4 . . ll Additions to 1,3-dipheny1-2-butanone and l,3-diphenyl—2—butanone—l,l,3—d_3 . . . . . 12 Additions to 2-pheny1-3—pentanone, l,2-diphenyl-l—pr0panone, 2-methy1- 4-pheny1—3-pentanone and 2,2-dimethy1- 4—phenyl-3-pentanone . . . . . . . . . . . l3 Additions to 3-methy1-2—pheny1butanal and 3-methy1-2—pheny1butana1-2-d . . . . . . . l4 Additions to 4—methyl-3-phenyl—2—penta— none—1,1,1,3—d_4 . . . . . . . . . . 15 Additions to 1,3-dipheny1-4-methy1- 2—pentanone—l,l-d_2 . . . . . . l6 Additions to 4-phenyl—2-pentanone and 1,3—dipheny1—l-butanone . . . . . . . . . 21 Additions to phenylacetoin and methylbenzoin . . . . . . . . . . . . . . 24 vi LIST OF FIGURES Figure Page 1. "Neat" n.m.r. spectra of (A) 2-pheny1- propanol—2,0—g2 and (B) 2-phenylpro- panOl—l’2,0—Q3 o o o o o o o o o o o o o o 35 2. N.m.r. spectra of 10% solutions of (A) 2—methyl—3-pheny1-2-butanol and (B) 2-methy1—3-pheny1-2-butanol—l,l,l,3—_d4 in pyridine . . . . . . . . . . . . . . . 39 N.m.r. spectra of 0.03 g. of (A) 2—benzy1- l, 3——dipheny1-2-butanol and (B) 2—benzyl— 1, 3- -diphenyl- 2— butanol— l, l, 3- d in 0.30 ml. —3 Of pyridine 0 O o o o o o o o o o o 43 3. vii INTRODUCTION The formation of new asymmetric centers or of en- tire dissymmetric molecules in unequal amounts due to the presence of asymmetry in the catalyst or reactants is the essence of asymmetric induction or asymmetric synthesis. This phenomenon is observed in and reviewed for a variety of reactions including pyrolytic eliminations and carbonyl, imine, and alkene additions.l—5 Of these, additions to carbonyls are perhaps the most widely studied. Yet, even in simple acyclic carbonyl systems it is still not possible for the physical organic chemist to fully account for the unequal amounts of diastereomeric products formed and to predict with assurance the product ratio of similar reactions. The importance of several factors has been realized in the development of empirical models which qualitatively account for observed results. The role of steric factors in determining the course of additions to these systems was realized by Cram and Elhafez6 in the presentation of an em— pirical rule which generally predicts the major diastereo- meric product resulting from additions to carbonyls directly bonded to asymmetric centers. If the sterically small (5), medium (M) and large (L) groups attached to the asymmetric carbon atom are Viewed in a conformation in which the car— bonyl is flanked by the small and medium sized groups, the major diastereomer is considered to arise from attack of the reagent (R'X) occurring mainly on the sterically least hindered face of the carbonyl: This rule applies only to kinetically controlled non- catalytic reactions in which the groups attached to the asymmetric carbon do not coordinate with the metal atom of the attacking reagent. The terms "steric approach control" and "product development control"7 are used to account for the influ— ence of product and reactant stability in additions to cyclohexanones. Therefore, "with respect to the reaction coordinate two extreme situations with a broad spectrum in between are conceivable: reactant like transition states marked by little bond breaking and making, and product like transition states marked by extensive bond breaking and making."8 Recently Karabatsos8 described a simple empirical model which not only predicts the major diastereomeric product in additions to carbonyls directly bonded to asymmetric centers but also permits semiquantitative pre— dictions of product stereospecificity. In the model the two low-energy diastereomeric transition states which con- trol product stereospecificity have the smallest group (s) of the asymmetric carbon closest to the incoming bulky group (R'): MO OL ' i A (— R-'- --R' ——§ B L M S s R R If little bond breaking and making is assumed to mark the diastereomeric transition states, the arrangement of the three groups (5, M and L) of the asymmetric carbon atom with respect to the carbonyl is the same as in aldehydes and ketones, that is, one group eclipses the carbonyl. The diastereomeric product ratio, A/B, is then evaluated from the relative magnitudes 0 vs. L 6—9 0. Of course, even with isotopic species these gauche interactions are not exactly the same because R' is associated with a metal atom in the transition state. Nevertheless, these interactions would be smaller than in non-isotOpic systems where R' does not approach the size of R. The effect of differences in solvation energies of the diastereomeric transition states cannot be predicted because of the uncertainty of complexation of various groups with the solvent. Any contribution to the AAGiB values would be expected to be equal to or less than 50 to 100 cal./mole.8 In additions of methylmagnesium halides and methyl- lithium to methylbenzoin and of phenylmagnesium halides and jphenyllithium to phenylacetoin, for which a rigid model12 applies, the diastereomeric product ratio and the major diastereomer produced depend on the reagent employed.l3 Methyllithium exhibits less stereospecificity in pentane than in ether.13 A systematic study of the effect of chang— ing solvent and varying slightly the reagent, that is, changes from alkyllithium to Grignard reagents and changes of halide in the Grignard reagents, remains to be reported in systems for which the Karabatsos model applies. There— fore, to determine the general applicability of the AAGiB 8 9 values calculated from the Karabatsos model ’ such an in- vestigation is necessary. In addition, these isotopic systems should lend iflaemselves to simple analysis since the magnetic non- emquivalence of the diastereotopicl4 groups of these diastere— cnners should be observable in their n.m.r. spectra.15 Inte- gyration of the areas of these diastereotopic absorptions tflien would enable the determination of the diastereomeric pumoduct ratio, A/B. Therefore, in order to consider the inqoortance of the previously mentioned factors controlling true difference in free energy of the two diastereomeric treansition states, reduction of 2-pheny1pr0panal-2-d and 3-Tnethy1—2—pheny1butana1-21i with lithium aluminum deuter— idae enui sodium borodeuteride, additions of methylmagnesium hefilides and methyllithium to 3-pheny1-2-butanone—l,l,l,3-d4 arufl 4-methyl-3-phenyl-2-pentanone-l,1,l,3-d4 and additions off benzylmagnesium halides to l,3-dipheny1-2-butanone—l,l,3—_d_3 arui l,3-diphenyl—4-methyl—2-pentanone—1,l-d2 have been in— xnestigated as described in the experimental part. RESULTS AND DI SCUSSION The results of the additions of metal deuterides, Inethyllithium, methylmagnesium halides and benzylmagnesium halides to carbonyls directly bonded to asymmetric centers kxaaring methyl, phenyl and deuterium, or iSOpropyl, phenyl eund deuterium, are summarized in Tables I-III and V-VII. Iflor purposes of comparison all other known additions of nuatal hydrides and organometallics to aldehydes and ketones ir1 which the carbonyl is directly bonded to an asymmetric cennter bearing the aforementioned substituents are also in- CIJJded in Tables I—VII. In all these systems the difference ir1 free energymxfthe two diastereomeric transition states hmis been calculated in accord with the Curtin-Hammett principle:lO + _ AAGAB _ —RT1n(A/B). A. Errors in the Calculated AAGiB Values For deuterated systems, except 1,3-diphenyl—4—methyl- 2-pentanone-l,l-_d_2 (Table VII), the error in the percentage Of each diastereomer reported is 11%, that is, a diastere- omeric product ratio of 60/40 cannot be considered more 10 s ahqmcm rahnmmw mGOHumHOQO camom Umcflanmpmp was m4 .mHoE\.Hmox om.on u 0 msflummn Hmucmo oauumflfi mum so Mom mmumummae Amv.o so uoqfla mam Amm.o so plotN.Hlmaocmmonmlalamcmnmlm oaumfiomnwpmmwp mo capmu onem .muopxfla << muuapmum HmUOE mempmnmnom may cmmoupmc Ugo .mHOE\.Hmox CH pmmmmnmxm Q : ”~2me : QED. CH HOnME map mo coflumummucfl an AH N.H- m.aa\m.mm 0mm serum Homszoa o mm.onv H\v A 0mm umeum “mass m mm.on H\m 0mm umcum ummZDm SH mm.o- m.mm\m.oe com- ma mv.ou om\oe 00 serum Amazes m Ha.ou H\m 0mm scrum Hose: omomwmmo e mo.fimm.o- om\os we 0am- I s mo.a~m.o- em\mo we 0mm moumfl ammz mo.HmH.ou me\sm mo .0 a mo.aaa.o- mexmm mm owe mes ammz mo.fima.o- os\oo mm .sm. 4 m mo.HmH.ou ma\mm mm 0mm umnum mamaq omoow mo 9 .mmm anO<< mm\¢ macaw w .mEmB ucm>aom maflzmooaosz mumuumnsm .mrmlamammoumahamnmlm mam Hmsmmoumamcmnmlm on mCOADH©U¢II.H magma 11 .mGOHpomoH nguo map mo coaumuucmocoo map cucmplmco pm cmEHomumm was coHpommH mazeo m.m.mHos\.Hmox SS.o- n aw mfi ahsmcm samnumfi mcflumon Hmucmo venue Ahh.m ev ROCHE paw A0h.m Ev << mpowpmum HmUoE mempmnmumm mag cmmoupma Ucm Mme cm mom .mHOE\sHmox CH pmmmmumxm Q .mcflpflumm :H mGOHpSHOm woa CH mGOHumHOQO camouomnmummflp HommE egg mo moons may mo coaumummucfl mg Umcfifiumgmp mos Ulm.H.a.HumaocmusnlmIawcmnmlmlenumEIm OHHmEomHmHmeU mo OHHMH mafia so.Hmo.Hn S\Nm SS oSS- So.Hsm.ou sH\mS SS omm serum HHS: SS.HHR.S- HN\SR as oSm mamucmm HHS: SS.HSS.H- SH\RS mS om ems SS.HNS.H- SH\SS SS 0mm Hmaum So.Heo.Hu sH\mS Se oSS ems Homsms «S.HmS.oa omH\HS SS .Sm HS.HRS.S- stmS SS om SS.HHS.S- Hm\ms SS .Sm Hmapm “moses HS.HSS.S- SH\NS mm .mm. maouooommo mo.HSS.o- Sm\ms em .mm Harem Hoses = _ o S SH 05.0- SH\SS cos- SH SS.o- S.N~\S.ss .S m m S mm.ou H\m.m 0mm mouomo mo SH mS.o- H\S.m .mm a % _ SH mm.o- S.smxm.ms .Sm “warm mHHHH S .mmm QmMo<< mm\¢ chHM w .mEmB ucm>Hom maflcdooHosz opmupmnsm .wl Ulm.H.H.Hlococmudnlmuamcm£mlm paw mcocmusnlmlahcmnmlm 0p mcoprUUHom mHflgmomHosz mpmupmnsm . mum.H.lenocmpsnnmuHscmamHeum.H Saw maoempsnlmlenmadHeumrH on meoHuHSSSII.HHH mHnSe l3 m4 m.m.oHOE\.Hmox om.ot n ~+HW<< mgoflpmHm HoUOE mOmumnmmmm ecu cmmoupmn pom Hammad .Hwagmfi mcHHmmn Hmpcmo OSHum and cm Mom .mHoE\.Hm0M GS Ummmmnmxmm SH HS.S- S.S\S.SS 00S- S I S SH SS.S- S\SS oSS Hague mHSHH smummw mo 0 S SH SS.S- H\S.S oSS Swarm SSHSHH HummmwSmo o S S SS.o-v H\SA SS oSS “Seam SSHSHH SwmoSmo 1 SH HS.S| H\S.S oSS S S S S S SS.o| H\S oSS Harem mHSHH mo SQflmw mo 0 S .mmm SSMSSS m\S SHSHS S .QESS ucm>Hom SHHadomHosz mnmupmnsm .maocmwchIMIawcwnmISlawnumfiflclm1N paw GQOCMchQIMIHmchQIS IHmaumEIN rmcocmmoumlHlenmcmSpim.H .mcocmpGQQIMIchocmlm cu mGOHpHUUSII.>H oHQMB J1 14 m. .wHoE\.Hmox om.OI H m£ Ucm Shamam SHSQOHQWmS mcflnmmn mHmpcmo 0S Dwfifiwwm mom .mHoE\.Hmox CH Ummmmumxmn .mcmamnpmwcoEOSQIH GS mQOHSSHOm wow mo mQOSHQHOQO DamouowHWHmmHU ANS.w SV HOQHE paw Amm.m SV Momma map mo QOHumummch an pocHEHmwmp mma mplorN.HImaocmgdnlaIamcmgmlmlawnpwfilm UHHwEomuwumeU mo OSuMH mafia SH So.o+ SS\SS SS °SS msmncmm HHHHH SH SH.SI H\S.H HS .SS “Seem SHSSES + Hmmzuma SH SS.o- H\S.H SS oSS Hanna HSSSHHS omomwumm S So.SSS.oI SS\SS SS oSSI I So.HSS.oI SS\SS SS oSS mouSH SS.HSS.SI SS\SS SS 0H S So.HoS.oI SS\HS SS oSS See ommz SS.HSS.SI SS\SS SS oSSI SS.HSS.SI SS\SS SS 0S S I So.HNS.oI SS\oS SS oSS “mane QHSHH omoowumH S .mom Qme<< mm\< UHwHw w .man pco>aom oHflnmomHosz opmupmnom .mINIHMGMDDQH%£@SQINIahnpmfilm paw Hmcmgsnawcwflmlmlfl>£pmfilm Op mcoHuHUU manme 15 S.S.mHos\.HSox SS.o- ahawam .meOHQOmH msSHmoQ mumpcmo 0S uofifiwmm Mom m¢w<< mHOSUmHQ Hmpoa mOmHMQMHMM msv cmmowpwn Ugo .mHoE\.Hmox GS pcmmmnmxm Q .USom UHEHOM QH mQOHHDHOm wow GS mcoHHQHOQO 0HmouomuopmmS© Amm.m Sv HOQSE one AwS.m SS Momma msv mo mmmHm may mo coHpmHmmpcS SQ Umcflfinogwo mos Ulm.H.H.HImHocmvcwmINIHmcwngmIchumfiflolvtN UHHmEomHmmeHU mo OHHMH wake SS.HSS.HI S\SS SS oSSI SS.HSS.H- SH\SS SS °S umanm HHS: SS.HSS.SI SS\SS SS °SS ocmuamm HHS: SS.SSS.SI SS\SS SS 0S SSS SS.HHS.SI SS\SS SS oSS “Seem So.SSS.o- SS\SS HS oSS See HoS2mz So.HHS.oI SS\SS SS 0S SS.HSS.SI SSxSS SS oSS swapm “Seam: SS.SSH.SI HS\SS SS .HSI So.SSH.o- SS\SS SS 0S S I SS.HSH.SI SS\SS SS .SS “Seem HSZSE ooflowumH o S Qmwo<< mm\< UHmSM S .mEmB ucw>Hom mHngooHodz mumupmnsm .SI UIS.H.HrHuococmpcmoumIHmcondImIHmcuoEIS 0p mQOSpflopmlI.H> oHQmB 16 .oSumH oEmm exp mw>Ho COSDSUUM meo>cHo .®HOE\.Hmox ow.0I u mmw<< muoprHm HGUOE mOmqumme map smmouphc mam \ Shawna SH>QOHWOMS mCSHmwQ mHopcwo OHM mafiwmm Mom .®H0E\.Hmox QH Ummmwgmxmn .mcflpfluhm CH mcoSuDHOm wow GS mQOSHQHome UHQOHOmeummHU Hoo.S So HOQSE pom Ahw.o So HoflmE w:u mo mcmHm wQH mo QOHHSMmmHGH an oocflfinmgmp mos UIH.HImaocmvsnlwIawgpmfilvlamcmgmflpImSHIHSNGmQIw OSMQEomepmMHU mo OSSMH mafia S SH SS.H- USH\SS SS °SS “Seem mHSHH sawmmwumm o S SS.HSS.SI HSxSS SS .S S SS.HSS.SI SS\SS SS oSS “Sana HUS: moS So.HHS.oI SS\SS SS °S S SS.HSS.SI SS\HS SS oSS Hwapm Hmms moS So.HSS.oI SS\SS HS oS S S I SS.SSS.SI SS\SS SS oSS “Seem HS: moS S QUASWHSH f .Smm anSSS Sm\< SHSHS S .dsme Hcm>Hom SHHadomHosz mnmupmnsm wl .wcocmxngMIamcmnmIS -HSSSSEHSIS.S SSS SIH.HlmgoemuemmISIHSSHSEISIHSQSSSHSIS.H ou mquuHSSSII.HH> SHQSS 17 accurate than 59/41 or 61/39. Due to poorer resolution of the diastereotopic absorptions of the two diastereomeric 2—benzy1—1,3-diphenyl—4—methy1-2-pentanols—1,1—d2 (Table VII) the error in the percentage of each diastereomer re— ported is estimated to be i3%. The errors which are re— ported for the AAGiB values in Tables I—III and V—VII were calculated on the basis of the errors in the diastereomeric product ratios. Since these errors in the diastereomeric product ratios cause errors ranging from five to twenty percent of the AAGiB values a temperature variation of 12°, or approximately one percent, does not contribute signifi— cantly to the total error in AAGiB. The accuracy of AAGiB values calculated for some systems previously reported in the literature may be ques- tionable. Whereas errors in diastereomeric product ratios 17,19 determined by infrared and vapor phase chromatogra- 16’18 techniques should be comparable to or smaller phic than those of n.m.r. analyses, the errors in diastereomeric product ratios determined by actual separation of the di— astereomers by crystalization of derivatives6 may be con— siderably larger. Also, except for the studies of Felkin 16,18 6,17,19 do not and co—workers previous investigators appear to consider a temperature dependence of the diastere— omeric product ratio and, therefore, the reaction tempera- tures could only be estimated from experimental descriptions. Recently Felkin and co—workers18 reported that the results 18 of Gault and Felkinl6 were obtained by a faulty technique and are slightly in error. These factors should be con— sidered in the comparison of AAGiB values. Also questionable may be the validity of comparing systems in which the asymmetric centers differ isotopically. No evidence exists as to the degree in which isotopes may alter the diastereomeric product ratio in additions to these carbonyl systems. In a different system, the alcoholysis of d—phenylbutyric anhydride with (S)-2—propanol—l,1,l-d3, a small but detectable optical yield of 0.4 to 0.5% was observed in the a-phenylbutyric acid obtained after hydrolysis of the J unreacted anhydride.20 However, since secondary deuterium isotope effects are small, the extent of asymmetric induction due solely to isotopic differences in additions to these car— bonyl systems is expected to be small and beyond the detection of the analytical methods employed in this investigation. The stereochemistry of the diastereomeric products formed in additions to these deuterated aldehydes and ke- tones and in additions of lithium aluminum hydride to 1,3—dipheny1-2—butanone and benzylmagnesium chloride to 2—pheny1propanall7 has not been proven. However, in other investigations6’l6’l8’19 (see Tables I—VII) the stereochem- istry of the diastereomeric products has been proven and the major diastereomer corresponds to that predicted by the Karabatsos model. Due to the similarities of the systems of proven and unproven stereochemistry the major l9 diastereomer of the latter is presumed to be the one pre— dicted by the Karabatsos model. B. Entropy Effects Prior to the start of this investigation the only study of the temperature dependence of diastereomeric pro— duct ratios in additions to carbonyls directly bonded to asymmetric centers was performed by Gault and Felkin.l6 Additions of methylmagnesium bromide to 2-pheny1propanal (Table I) and lithium aluminum hydride to 3-pheny1—2—buta— none (Table II) revealed variations in AAGiB of 0.07 kcal./ mole over a 50 degree range and 0.11 kcal./mole over a 105 degree range respectively.l6 Similar variations of AAGiB values with temperature have been observed in the lithium aluminum hydride reduction of 3—methy1-2-pentanone.l6 Of the twenty studies of the temperature dependence of AAGiB values performed in this investigation with sys- tems of differing substrate, nucleophile or solvent, only one—-the reduction of 3—methy1—2-pheny1butanal-2—d with sodium borodeuteride in 2—propanol in which AAGiB varies 0.14 kca1./mole over an 86 degree range—-shows AAGiB values which are outside the error range of the average AAGiB value. The consistency of the values in the other nineteen studies warrants a consideration of the possibility that the difference in the AAGiB values in this reduction may arise from a random error. Calculation of AAGiB from the 20 recently reported diastereomeric product ratios18 ob— tained on reduction of 2,2-dimethy1—4—pheny1—3-pentanone with lithium aluminum hydride (Table IV) reveals a vari- ance in AAGiB of 0.13 kcal./mole over a range of 105 degrees. The temperature dependence of diastereomeric product ratios in additions to carbonyls beta to an asymmetric center,21 1,3—asymmetric induction (Table VIII), also reveals entropy effects of the same magnitude as in 1,2-asymmetric induction. Thus, the observed increase in diastereomeric pro- duct ratios with decreasing temperature does not result from entropy effects. These results constitute evidence 4. AB trols AAGiB. Thus, since AAGiB values are a correction that in cases where Karabatsos' model applies AAH con— of diastereomeric product ratios for temperature differ- ences, a comparison of AAGiB values offers an easier com- parison of the stereospecificity of these reactions. C. Solvent Effects Although an extensive study of the influence of solvent on diastereomeric ratios has not been performed, sufficient evidence indicates that the product stereospeci— ficity of these reactions depends on the solvent as well as on the reactants. The most striking effects of changing Solvent occur in the addition of 3—pheny1—2-butanone- 1r1,1,3—d4 (Table II) to solutions of methyllithium in -il 21 .H.E.c SQ UmcSEHmvmo mm OHHSH poopoum OHHmEomeumoHo .mHOE\.HmoM CS commommxmo Q .Hw mocmumwog Eonw Umuomnvmnm mumom SH.o+ SS\SS SSH co Hompm Home: SH.o+ SS\HS HS oSSI SH.o+ SS\SS SS 00 SH.S+ SS\SS SS 00 SS.o+ SS\SS SSH oSS Hmmum “mazes SH.S+ SS\SS SH 00 Serum Homzmm SS.S- SS\SS SS 00 Smmnm HHS: SmSmomwSmo o S HS.SI SH\SS SS ooHHI SS.o- SS\SS SSH oSS- SS.S- SS\SS SS 00 SS.SI HS\SS SS °SS mmmmm SSSSS HS.o- SS\SS SSH oSSI SH.SI SS\SS SS 00 S S S SH.oI SS\SS SSH oSS Smem HHS mom momo mo 0 .1 omMSSS nm\< SHSHS S .mfime umm>Hom mHHmmomHosz mumupmnsm w.wcocmu59IHIHmco£mHoIm.H cam macadacmmeIHmcochS ou msoSpSUU oHQme 22 pentane (AAGiB = —0.71 kcal./mole) and ether (AAGiB = —l.02 kca1./mole) and in the addition of 4—methy1—3—pheny1— 2—pentanone—l,l,l,3—d_4 (Table VI) to solutions of methyl— lithium in pentane (AAGiB = —0.44 kcal./mole) and ether (AAGiB = —l.05 kcal./mole). Similar decreases in pentane vs. ether have been observed13 in additions of methyllithium to methylbenzoin and phenyllithium to phenylacetoin (Table IX). Under the reaction conditions methyllithium is only slightly soluble in pentane and appreciably more so in ether. These decreases in stereospecificity could con— ceivably arise from differences in concentration or structure 1 of methyllithium in these solvents, or from a heterogeneous vs. a homogeneous reaction. It is not believed that the carbonyl group eclipsing interactions will change signifi— cantly in pentane vs. ether.9 Changes of solvent from tetrahydrofuran to 2—propanol result in an increase of 0.2 kcal./mole in the difference in free energy of the diastereomeric transitions states for additions of sodium borodeuteride to the deuterated aldehydes (Tables I and V). Since an increase in solvent polarity favors phenyl eclipsing relative to methyl or isopropyl eclipsing of the carbonyl,9 the difference in free energy of the diastereomeric transitions states might be predicted to be smaller in 2-propanol than in tetrahydrofuran. How- ever, the opposite result is observed. This could be due to the formation of different alkoxyborohydrides as the 23 .GOSwHopm owuo>ch .oHoE\.Hmox SS Ummmmumxmo .OSva pospoum OSHwEomuwpmmHQn .MH mocmumme Eoum pmwomupmnm mgmom SS.HI H\S.S Sommm SS.SI H\S.S mmmpmmm HHS: SS.o- H\SS.S meum HUSSSE SS.S- H\SS.S ummum SSSSSS Smo _ SS.oI H\SS.H SomoS SS.SI H\SS.H Hmmpm Haze: @ _ om SS.o- H\HS.S SS.oI H\SS.S Smmnm Hosz SS.o- H\SS.S SS.oI H\SH.S SS.S- SH\SH.S Homom SisS SS.S+ oo.S\H Hmmmm HSzS SS.S+ SH.S\H Swarm SsSS Smo HS.H+ S.S\H mcmmmmm _ SS.H+ SS.S\H SmoomoS SS.H+ S.S\H % _ SS.H+ S.oH\H Smmnm HHS om mm o +oSS nHS\omSz mmm>HoS mHHmmomHosz ammunmnsm m.omm um QSONGMQHSQUQE mam QSOHwomHmcmnm ou mQOSuSUUssh1B may not be significantly different. Third, whereas the interactions of R' 6—9 5 should be equal in both transition states, the difference in interactions of R' <—9 M in A and R' 6—9 L in B would become larger, thus increasingly favoring A over B as R' becomes larger. This is generally observed in ether solutions of hydride and alkylmagnesium bromide and iodide additions to the system where the asymmetric carbon bears methyl, phenyl and hydro- gen (Tables I—III). In additions, however, where the asymmetric carbon bears isopropyl, phenyl and hydrogen (Tables V—VII) only small differences in stereospecificity are observed as the size of R' increases. These observa- tions are not unreasonable, since, in the vicinity of the 27 carbonyl, the steric effects of phenyl and methyl obviously differ whereas those of phenyl and isopropyl, both branched groups, may not differ significantly. Thus, a difference in R' 6—9 M vs. R' 6—9 L interactions is expected if the asymmetric center has methyl, phenyl and hydrogen substitu- ents but is not expected if it has isopropyl, phenyl and hydrogen substituents. E. Effect of the Symmetrical R Group of MLsC—CR Increasing the size of the R group results in in— creased stereospecificity. Observation of the data from hydride reductions in which the asymmetric center has methyl, phenyl and hydrogen as substitutents (Tables I—IV) indicates four levels of stereospecificity depending on the degree of branching in R. In the aldehyde reduction the stereospecificity is low. When R is CH2R" such as methyl, ethyl or benzyl, AAGiB values range from -0.6 to -0.7 kcal./mole as predicted by Karabatsos' model. When R is CHRE, such as isopropyl and phenyl, AAGiB values are about —l.0 kcal./mole. Further increases in stereospecifi- city are observed when R is CR3 such as p—butyl. For the system in which the asymmetric center contains isopropyl, phenyl and hydrogen as substituents only a limited amount of data are available. Lithium aluminum hydride reduction of the aldehyde (Table V) shows less stereospecificity than the reduction of the isopropyl ketone (Table VII). 28 The low stereospecificity in hydride additions and some Grignard additions to aldehydes may be due to exten— sive bond breaking and making marking the diastereomeric 11’25 With sterically small R groups transition states. (hydrogen) and small nucleophilic reagents the substrate and nucleophile may easily approach the carbon—carbon bond distance of the product. Increasing the size of R and R' makes extensive bond breaking and making more difficult. Increases in the degree of branching of R will re— sult in greater interactions between R and R': A B This could sterically crowd both transition states to re— sult in a greater manifestation of the interactions between R' S—a M and R' 6—9 L further favoring A over B. When R becomes very large, such as Efbutyl, pre— dictions stemming from carbonyl group interactions require revision because the R 4—9 s and R S—S L interactions in A and the R S—9 5 and R S—> M interactions in B may alter the minimum energy conformations of the two transition states. These interactions may be such that the conformations hav- ing M and L eclipsing the carbonyl might be energetically 29 unfavorable. The importance of these interactions in de— termining the minimum energy conformation of carbonyl compounds has been inferred by the low field chemical shift of the isopropyl methine in isopropyl p—butyl ketone rela— tive to those of the methines of other less substituted isopropyl ketones. In isopropyl p—butyl ketone the methyls of the isopropyl group do not eclipse the carbonyl in the most stable conformation:26 0 CH3 / Even if R is only disubstituted, complexation of a metal with the oxygen of the carbonyl may cause interactions between the metal and the groups eclipsing the carbonyl and between R and the substituents of the asymmetric center, thus, altering the minimum energy conformations from those predicted by the model. Expecting the stable carbonyl com- plexed conformations to be other than that depicted below in somewhat analogous systems (Z = alkyl, OR, NR2)27 is unwarranted: 30 Thus, the following conformations may be more representa- tive of the transition states: In I complexation forces R" away from an eclipsing posi— tion, thus, increasing the unfavorable interaCtions between R" and L and s. Complexation as illustrated in II diSplaces 4 or L from an eclipsing position. Thus, when R is branched, the Karabatsos modeléusformulated previously may not be ap- plicable. Sufficient data are not available for such xranched systems to draw definite conclusions regarding the applicability of the model. ‘ Conclusion This study of additions to carbonyls directly bonded r>tasymmetric centershas revealed that entrOpy and concen- .r€rtion of reactants do not influence product stereospecifi- itqf in these reactions. Stereospecificity depends strongly n syibstituents of the carbonyl compound.on nucleophiles-- ruxluding differences in the halogen of a Grignard reagent-— rui CH1 solvents. In ethereal solutions lithium aluminum ydixhde reductions of ketones of the type MLsC—C(CH2R")O 31 amiadditions of methylmagnesium iodides to methyl ketones ammed very good agreement between the observed differences hifree energy of the diastereomeric transition states and Hume predicted by the Karabatsos model. Increasing the sfize of the nucleophilic reagent and the degree of branch- ing in the symmetrical alkyl group of the ketone from those mentioned above generally resulted in stereospecificity greater than that predicted by the model. Stereospecifi— city usually increased with alkylmagnesium bromides and chlorides. Methyllithium was found to be more stereospecific in ether than in pentane solution. EXPERIMENTAL A. General Procedures In the following experiments described in these sections, the apparatus for the addition reactions was standard for Grignard reactions: a 100 m1. flask was equipped with either a magnetic or overhead stirrer, ther- mometer, condenser with calcium chloride drying tube, and an addition funnel or septum cap for additions from a syringe. The temperature of the reaction mixtures was maintained by dry ice-2-propanol, water-ice, or water baths or by refluxing the solvent. The maximum deviation of temperature in the addition reaction vessel was 12° in the water and ice-water baths and 13° in the dry ice-2-propanol bath. Solutions of methylmagnesium chloride in tetra- hydrofuran, and methyllithium in ether were obtained from Alfa Inorganics, Inc. When reactions with these reagents were performed in solvents other than these, the ether or tetrahydrofuran was removed under vacuum and the desired solvent (pentane or ether) added to the solid material under a nitrogen atmosphere. Methylmagnesium bromide in ethem was obtained from Arapahoe Chemicals. "Domal High 32 33 Purity Magnesium Granules" were used in the preparation of all other Grignard reagents. N.m.r. spectra were obtained using a Varian A—60 n.m.r. Spectrometer and the diastereomeric product ratios were determined by three or more integrations of the areas of the diastereotopic absorptions at a sweep width of 50 c.p.s. and a sweep time of 100 seconds. The average of these integrations rounded off to the nearest percent.is reported. Concentrations of solutions for n.m.r; spectra are expressed as percent by volume, unless otherwise noted. All melting and boiling points are expressed in degrees centigrade and are uncorrected. B. Preparation of 2-Pheny1propanal-2-Q Twenty grams (0.149 moles) of 2-phenylpropanal (Aldrich Chemical Co., P-3160-4) was placed in a 100 ml. flask equipped with a magnetic stirrer and a condenser with a calcium chloride drying tube and treated succes- sively at 100° with 50 ml. of 90% deuterium oxide for one day, 33 ml. of 99% deuterium oxide for two days, and 15 ml. of greater than 99% deuterium oxide for one day. The pH of the deuterium oxide was maintained at 9-11 by the addi- tion of anhydrous potassium carbonate. After each treat- ment the aldehyde was extracted with ether which was then removed under vacuum. On distillation five portions of oil totaling 15.6 g. (77%) and having the following range 34 of prOperties were collected: b.p. 88-89° at 16.5 mm.; n35 1.5154-1.5156; n.m.r. spectrum, three singlets at T 0.36, 2.72, and 8.68 (ratio 1.0 : 5.0 : 3.0). C. N.M.R. Solvent Studies of 2-Pheny1propanol-2,0e§ 2 2-Pheny1propanol-2,0-d (b.p. 101-103° at 12 mm.; 2 n§5 1.5279-l.5283), prepared in 78% yield from 2-pheny1- prepanal—Z-d and lithium aluminum hydride by procedures similar to those described in the next section, exhibited a singlet at T 2.88,an AB quartet centered at T 6.50, and a singlet at I 8.80 (ratio 5.0 : 2.0 : 3.0) in the n.m.r. spectrum of the "neat" compound (see Figure l). The non- equivalence of HA and HB was not observed in 20% solutions of this alcohol in carbon tetrachloride, chloroform, carbon disulfide, formamide, dimethylformamide, dimethylsulfoxide, acetonitrile, benzene, chlorobenzene, toluene, nitrobenzene, pyridine, or thiophene. D. Additions to 2—Pheny1pr0panal-2-Q A solution of 2.00 g. (14.8 millimoles) of 2-pheny1- propanal-Zji dissolved in 5 m1. of ether was added over a ten minute period to a stirred mixture of 0.20 g. (4.77 millimoles) of lithium aluminum deuteride and 20 ml. of ether maintained at the reaction temperature (25° or -57°). After stirring an additional four hours at this temperature, 1.0 ml. of deuterium oxide was added drOp-wise with 35 l T = 6.50 8.80 10.00 M, I l l T = 6.46 6.58 8.80 10.00 Figure l.--"Neat" n.m.r. spectra of (A) 2—pheny1propanol—2,0—d_2 and (B) 2—phenylpropanol-l,2,0-d3. 36 stirring. The ether layer was decanted from the granular salts and the ether was removed by vacuum evaporation. On was obtained in n25 D distillation 2-pheny1-1—propanol—l,2,0-d3 85 to 92% yield: b.p. 97—98° at 11 mm.; 1.5230-1.5234. The n.m.r. spectrum of this alcohol consisted of four sin— glets at I 2.86; 6.46 and 6.58; and 8.80. The ratio ob- tained by integration of the areas of these absorptions was 5.0 : 1.0 : 3.0. The diastereomeric product ratios deter— mined from integrations of the major (I 6.58) and minor (T 6.46) diastereotopic absorptions (see Figure l) are reported in Table I. A solution of 3.10 g. (23.0 millimoles) of 2—pheny1— propanal-Z—d dissolved in 5 ml. of solvent (tetrahydrofuran or 2—propanol) was added over a ten minute period to 0.30 g. (7.1 millimoles) of sodium borodeuteride in 75 ml. of tetra— hydrofuran or 35 m1. of 2—propanol. During the addition and for twelve to twenty—four hours thereafter, the reaction mixture was stirred and maintained at the desired tempera- ture. Work-up was accomplished by adding 10 m1. of 2 M hydrochloric acid, removing the organic solvent under re- duced pressure, adding 30 m1. of ether and separating the aqueous layer. The ether layer was washed with 10 m1. of water before removing the ether under vacuum. The residual oil was stirred with 10 ml. of deuterium oxide, extracted 25 with ether and distilled: b.p. 99-101° at 11 mm.; nD 1.5226-1.5246. The n.m.r. spectrum of this 37 2—pheny1—1—propanol—l,2,0—d3 showed that the 2—position was no longer fully deuterated and the resolution of the absorptions at T 6.46 and 6.58 was not good. Therefore, the diastereomeric product ratios were determined by cut— ting and weighing tracings of these absorptions. The diastereomeric product ratios thus obtained and the yields of these reactions are reported in Table I. E. Preparation of 3—Phepy1—2-butanone 3—Pheny1—2-butanone (b.p. 87—88° at 16 mm.; n35 l.5063—1.5065) was prepared in 68% yield by the alkylation of phenylacetone by the procedure of Suter and Weston28. Vapor phase chromatography revealed no phenylacetone in the product under conditions which the two could be resolved. 28 20 Literature value: b.p. 106—107° at 22 mm.; nD 1.5092. F. Preparation of 3-Pheny1—2—butanone—1,1,1,3—d_4 The BABES protons of 17.2 g. (0.116 mole) of 3—pheny1-2—butanone were exchanged in the presence of basic deuterium oxide as described for the preparation of 2—pheny1- propanal—Z—d (section B). Vacuum distillation yielded three portions of 3—pheny1-2-butanone-l,l,l,3—-d4 totaling 15.4 g.(88%) and having the following range of properties: b.p. 88—90° at 18 mm.; nSS l.5049—1.5052; n.m.r. spectrum, two singlets at T 2.96 and 8.72 with an integrated area ratio of 5.0 : 3.0. 38 G. N.M.R. Solvent Studies of 2—Methy1-31phenyl-2-butanol Via a Grignard synthesis as described in the next section, 2—methy1—3—pheny1-2—butanol (b.p. 103—104° at 12 mm.; n35 l.5139-l.5142) was prepared in 80% yield from 3—pheny1-2—butanone. The n.m.r. spectrum of the "neat" compound consisted of a singlet, T 2.82; quartet centered at I 7.27; singlet, T 7.35; doublet centered at T 8.68; and a singlet, T 8.88 (ratio 5.0 : 1.0 : 1.0 : 3.0 : 6.0). The six proton singlet assigned to the two diastereotopic methyl groups was also observed as a singlet in 10% solu— tions in benzene, chlorobenzene, toluene, formamide and chloroform. It was resolved into two equal intensity sin— glets at T 8.67 and I 8.73 in a 10% solution in pyridine (see Figure 2). Lesser resolution of these singlets was observed with 10% solutions in acetonitrile, dimethylsulf- oxide, dimethylformamide, acetone, cyclohexane and carbon tetrachloride. H. Additions to 3—Pheny1-2—butanone—1,1,1,3—d4 A solution of 1.00 g. (6.58 millimoles) of 3—pheny1- 2—butanone—l,1,l,3—d in 3.0 ml. of solvent (ether, tetra- 4 hydrofuran, or pentane) was added over a ten to fifteen minute period to a stirred mixture of 10 millimoles of Organometallic reagent (methylmagnesium halide or methyl— lithium) in 15 ml. of solvent maintained at the desired 39 l T = 7.07 8.50 8.73 10.00 8.o7 B 7‘ l I l T = 8.57 8.76 10.00 8.70 Figure 2.-—N.m.r. spectra of 10% solutions of (A) 2—methy1- 3—pheny1-2—butanol and (B) 2-methyl—3-pheny1— 2-butanol—1,l,l,3—d_4 in pyridine. 40 temperature. After stirring another five to twelve hours at the desired temperature 17 ml. of 1% ammonium chloride solution was added. When tetrahydrofuran was used as the solvent 30 m1. of ether was added to extract the alcohol from the aqueous medium. The aqueous layer was separated and the organic layer washed with 10 m1. portions of water and saturated sodium chloride solution before drying with anhydrous sodium sulfate. Upon removal of the solvent 2—methy1—3—pheny1—2—butanol—l,1,1,3—_d_4 with the following range of properties was collected: b.p. 78—79° at 2.5 mm.; n35 1.5090-l.5134. Yields of these reactions and the ratios of diastereomers as determined by integration of the areas of the major (T 8.70) and minor (I 8.76) diastereo- topic methyl absorptions (see Figure 2) of 10% solutions in pyridine are summarized in Table II. I. Preparation of l,3—Dipheny1—2-butanone First, l,3-dipheny1-2-butanol (m.p. 72.5-74.0°) was prepared in 41% yield by the addition of 160 g. (1.19 mole) of 2—pheny1propanal to 2.00 moles of benzylmagnesium chlo— ride as described by Julien and Kayserzg. By use of the procedure of Sandborn30 125 g. (0.554 mole) of this alcohol was oxidized with acidic aqueous sodium dichromate. Traces of unreacted alcohol were removed by column chromatography on silica gel using carbon tetrachloride as the eluant. After removal of the solvent 55.7 g. (45%) of 41 1,3—dipheny1—2-butanone was obtained upon distillation: n25 29 D b.p. 130° at 1.2 mm.; 1.5587-l.5589. Literature value: b.p. 192° at 21 mm. J. Preparation of l,3—Diphenyl-2—butanone—l,l,3-d_1 Deuteration of the elppe positions of 33.2 g. (0.148 mole) of 1,3-dipheny1-2—butanone with alkaline deuterium oxide as described in the preparation of 2-pheny1- propanal-2—d (section B) yielded 23.1 g. (69%) of l,3—di— phenyl—2—butanone—l,l,3-d3: b.p. 121—122° at 0.8 mm.; n35 1.5580-l.5583; n.m.r. spectrum of the "neat" compound, singlets at T 2.80 and 8.70 (ratio 10.0 : 3.0). K. N.M.R. Solvent Studies of 2—Benzy1-1,3-dipheny1- 2—butanol Reaction of l,3—dipheny1—2—butanone with benzyl— magnesium chloride as described in the next section yielded 89% of 2-benzy1—1,3—dipheny1-2-butanol: m.p. 103-104°. For n.m.r. spectra 0.03 g. of this alcohol was dissolved in 0.30 ml. of solvent. In carbon tetrachloride the spec— trum consisted of a singlet at T 2.90, quartet centered at T 7.15, singlet at T 7.27, singlet at T 7.33, doubled cen— tered at T 8.63, and a singlet at T 8.82 (ratio 15.0 : 1.0 : 2.0 : 2.0 : 3.0 : 1.0). The two singlets at T 7.27 and 7.33 assigned to the two diastereotopic benzylic groups were also observable in chloroform (T 7.15 and 7.18), 42 nitrobenzene (I 7.12 and 7.17), and pyridine (T 6.92 and 7.02, see Figure 3). However, in benzene and thiophene only a singlet at I 7.23 was observed. The diastereotopic benzylic protons appeared as two AB quartets centered at T 7.02 and 7.18 in methylene bromide and acetonitrile and at T 7.10 and 7.20 in benzonitrile. L. Additions to 1,3—Dipheny1—2—butanone-1,l,3-g1_3 The Grignard reagents were prepared by adding in one portion 10.3 millimoles of benzyl halide (2.25 g. of freshly distilled benzyl iodide, 1.76 g. of benzyl bromide or 1.31 g. of benzyl chloride) dissolved in 10 ml. of sol- vent (ether or tetrahydrofuran) to 0.25 g. (10.3 millimoles) of magnesium. The mixture began to reflux and was main— tained at reflux for ten to fifteen minutes before bringing it to the desired reaction temperature. Within twenty min— utes from the start of the Grignard formation, 1.13 g. 3 in 2.5 ml. of solvent was added over a ten to fifteen minute (5.00 millimoles) of l,3—dipheny1—2—butanone-l,1,3-g period while stirring and maintaining at the desired tem— perature. After stirring an additional two to seven hours at the desired temperature, 1.0 ml. of 10% ammonium chloride solution was added. The inorganic salts formed were re- moved by filtration and the ether was removed under reduced pressure to leave a crude solid which upon recrystallization from 10 m1. of 95% ethanol yielded white crystals of A II I | T: 6.92 7.02 8.47 10.00 B I I I I T = 6.90 7.02 8.47 10.00 Figure 3.--N.m.r. spectra of 0.03 g. of (A) 2—benzy1- l,3—dipheny1—2—butanol and (B) 2-benzy1- 1,3-dipheny1—2—butanol—l,l,3—d_3 in 0.30 ml. of pyridine. 44 2—benzyl-l,3-dipheny1-2—butanol—l,1,3—d3: m.p. 102—104°. The ratio of diastereomeric alcohols was determined by integration of the areas of the major (I 6.90) and minor (I 7.02) diastereotoPic absorptions of 0.03 g. of alcohol in 0.30 ml. of pyridine (see Figure 3). These ratios and the yields of these reactions are collected in Table III. M. Preparation of 3-Methy1—2-phenylbutanal A Darzens glycidic ester condensation involving the reaction of 166 g. (1.36 moles) of ethyl chloroacetate with 200 g. (1.35 moles) of isobutyrophenone in the presence of sodium ethoxide was performed exactly as described by Cram.19 Saponification and acidification of the glycidic ester yielded 3-methy1—2—pheny1butanal. Upon distillation six portions totaling 67.7 g. (31%) and having the following range of properties were collected: b.p. 69-72° at 2 mm.; 25 19 nD l.5044-l.5048. Literature value: b.p. 72—73° at 1 mm.; n35 1.5051. N. Preparation of 3—Methy1-2—pheny1butana1—2—d Preparation was effected by deuterium exchange of the d-proton of 19.9 (0.123 mole) of 3-methy1—2-pheny1— butanal as described for the preparation of 2-pheny1pro- panal—Z-Q (section B) except that the reaction was performed at 70° for a total of twenty days and under a nitrogen atmosphere. Three portions of 3—methy1—2-pheny1butana1—2—d 45 I I I 1 I i totaling 15.5 g. (78%) were collected upon vacuum distil- lation: b.p. 93-94° at 7.0 mm.; n35 1.5027-1.5034; n.m.r. spectrum of the "neat" compound, singlet, T 0.10; singlet T 2.80; septet centered at T 7.65; doublet centered at T 9.03; and doubled centered at T 9.32 (ratio 1.0 : 5.0 : 1.0 : 3.0 : 3.0). O. N.M.R. Solvent Studies of 3-Methy1-2-pheny1-1—butanol— 2,0‘d2 Reaction of 3-methyl—2—phenylbutanal-2-§ with lithium aluminum hydride yielded 79% of 3-methy1—2—pheny1— 1—butanol-2,0—d2: b.p. 78—80° at 1.0 mm.; n35 1.5120; ‘ n.m.r. spectrum of a 20% solution in carbon tetrachloride, singlet, T 2.85; singlet, T 6.37; septet centered at T 8.08; doublet centered at T 9.07; and doublet centered at T 9.30 (ratio 5.0 : 2.0 : 1.0 : 3.0 : 3.0). The singlet at T 6.37, assigned to the two diastereotopic protons, was resolved into an AB quartet centered at T 6.35 in a 20% solution in 1—bromonapthalene. However, in the "neat" and in 20% solu- tions in chloroform, methylene bromide, methyl iodide, neOpentyl bromide, cyclohexane, benzene, chlorobenzene, o-dichlorobenzene, toluene, nitrobenzene, benzonitrile, acetonitrile, pyridine, thiophene, methanol, formamide, hexamethylphosphoramide, formic acid, dimethylformamide, dimethylsulfoxide, and l,4-dioxane,the absorption due to the two diastereotopic protons was observed as a singlet. 46 P. Additions to 3-Methy1—2—pheny1butanal-2-d A solution of 1.00 g. (6.14 millimoles) of 3-methy1- 2—phenyl—butanal—2-d dissolved in 3.0 m1. of solvent (ether, tetrahydrofuran, or 2—propanol) was added over a ten to fifteen minute period to a stirred mixture of 2.6 millimoles of metal deuteride (0.11 g. of lithium aluminim deuteride or sodium borodeuteride) and 8.0 m1. of solvent which was maintained at the desired reaction temperature. The reac— tions were stirred an additional five to twenty hours at the desired temperatures. Work—up was as previously de- scribed for additions to 2—pheny1propanal—2—g (section DL On distillation 3—methy1—2—pheny1—1—butanol—l,2,O—Q.3 of b.p. 78-79° at 0.9 mm. and n35 l.5108—1.51l8 was collected. Table V contains the yields and diastereomeric product ratios determined by integration of the areas of the major (T 6.32) and minor (T 6.42) diastereotopic absorptions in 20% solutions in l-bromonapthalene. Q. Preparation of 4—Methyl-3—phenyl—2—pentanone Alkylation of 67.0 g. (0.500 mole) of phenylacetone with 85.0 g. (0.500 mole) of isopropyl iodide as described by Schultz and co—workers31 yielded, upon distillation with a spinning band column (Precision Distillation Apparatus Company, # 593), five fractions of 4-methy1—3—pheny1-2— pentanone totaling 39.4 g. (45%) and having the following 47 25 range of properties: b.p. 229-232° at 743 mm.; nD 1.4973— 1.4964. Literature vaiue:3l b.p. 109-114° at 18 mm.; 20 nD 1.5000. R. Preparation of 4—Methy1-3-phenyl—2-pentanone—1,l,1,3-d4 A 24.9 g. portion of 4-methyl—3-pheny1—2-pentanone was treated with deuterium oxide until the elppe protons exchanged as described in section B. Distillation yielded 21.2 g. (83%) of the deuterated ketone: b.p. 115—118° at 23 mm.; n35 1.4958; n.m.r. spectrum, singlet, T 2.73; septet centered at T 7.53; doublet centered at T 9.16; and doublet centered at T 9.33 (ratio 5.0 : 1.0 : 3.0 : 3.0). S. N.M.R. Solvent Studies of 2,4—Dimethy1—3fipheny1—2— pentanol Reaction of methylmagnesium iodide with 4-methy1— 3-pheny1—2—pentanone, as described in the next section, yielded 2,4-dimethy1—3—phenyl—2-pentanol: b.p. 79—80° at 1.0 mm.; nés 1.5080. The n.m.r. spectrum of a 20% solution of this alcohol in carbon tetrachloride consisted of a singlet, T 2.80; a multiplet, T 7.48 to 7.85; a singlet, T 8.13;asing1et, T 8.85;adoublet centered at T 9.02; and a doublet centered at T 9.16 (ratio 5.0 : 2.0 : 1.0 : 6.0 : 3.0 : 3.0). In a 20% solution in formic acid the six proton singlet at T 8.85 was resolved into two equal intensity singlets at T 8.63 and 8.73. These absorptions were not 48 very well resolved in 20% solutions in the following sol— vents: benzene, chlorobenzene, o-dichlorobenzene, toluene, nitrobenzene, benzonitrile, acetonitrile, thiophene, pyri— dine, acetone, dimethylsulfoxide, methyl formate, dimethyl— formamide, hexamethylphosphoramide, methylene bromide, chloroform, acetic acid and methanol. T. Additions to 4—Methy1—3—pheny1-2—pentanone—1,1,1,3—d_4 A solution of 1.00 g. (5.56 millimoles) of 4-methy1— 3—pheny1—2-pentanone—1,l,1,3-d4 dissolved in 3.0 m1. of solvent was added to 10 millimoles of organometallic re- agent in 12 ml. of solvent as described in the additions to 3—pheny1-2-butanone-l,1,l,3—d (section H). On distillation 4 2,4—dimethy1-3—pheny1-2—pentanol-l,1,1,3-d4 was collected: b.p. 78—79° at 0.9 mm.; n35 1.5040—1.5060. In Table VI the diastereomeric product ratios, as determined from integra— tion of the areas of the major (T 8.78) and minor (T 8.63) diastereotopic absorptions, and the yields of the reactions are reported. U. Preparation of 1,3—Dipheny1-4—methy1—2-pentanone The procedure of Schultz and co—workers31 was applied to the alkylation of 83.3 g. (0.397 mole) of di— benzyl ketone with 81.0 g. (0.476 mole) of isopropyl iodide. The 011 thus obtained was purified by column chromatography 49 on silica gel using carbon tetrachloride as the eluant. Removal of the solvent yielded 34.9 g. (35%) of 1,3-di- phenyl-4-methyl—2—pentanone. The n.m.r. spectrum of a solution of this ketone in carbon tetrachloride consisted of two singlets, T 2.80 and 2.87; a singlet, T 6.45; a idoublet centered at T 6.57; a multiplet, T 7.13-8.00; a doublet centered at T 9.15; and a doublet centered at T 9.40 (ratio 10.0 : 2.0 : 1.0 : 1.0 : 3.0 : 3.0). V. Preparation of 1,3—Dipheny1—4—methy1-2:pentanone-l,1—d9 The elppe protons of a 34.9 g. (0.139 mole) portion of 1,3—dipheny1—4—methy1-2-pentanone were exchanged over a period of two months with deuterium similar to the procedure described in section B. Distillation yielded 27.3 g. (77%) of l,3-dipheny1—4-methy1-2—pentanone-l,1-g 22 at 1.5 mm.; n35 l.5440-1.5446. The n.m.r. spectrum of a b.p. 130-132° carbon tetrachloride solution consisted of two singlets, T 2.80 and 2.90; a doublet centered at T 6.53; a multiplet, T 7.13-8.00; a doublet centered at T 9.15; and a doublet centered at T 9.40 (ratio 10.0 : 1.0 : 1.0 : 3.0 : 3.0). W. N.M.R. Solvent Studies of 2—Benzyl-1,3—diphenyl—4— methyl—Z—pentanol Reaction of l,3—dipheny1-4—methy1-2-pentanone with benzylmagnesium chloride as described in the next section yielded 45% of 2-benzy1-1,3-dipheny1—4—methy1-2—pentanol: 25 n b.p. 169—171 at 0.15 mm.; D 1.5808-1.5812. The n.m.r. 50 spectrum of a 20% solution of this alcohol in carbon tetra- chloride consisted of three singlets, T 3.00, 3.03 and 3.07; an A'B' system observed as a singlet, T 7.20; an AB quartet centered at T 7.23; a multiplet T 7.33 to 7.83; a singlet, T 8.73; and two doublets centered at T 9.10 and 9.15 (ratio 15.0 : 2.0 : 2.0 : 2.0 : 1.0 : 6.0). The best resolution of the A'B' and AB systems was observed in a 20% solution in pyridine, a singlet at T 6.87 and a AB quartet centered at T 7.00. Lesser resolution of these diastereotopic ab— sorptions was observed in 20% solutions in nitrobenzene, benzonitrile, methylene bromide, chloroform, toluene and benzene. X. Additions to l,3—Dipheny1—4—methy1-2—pentanone—1,1—d_7 A solution of 1.70 g. (6.70 millimoles) of 1,3-di- phenyl—4-methy1—2—pentanone—l,1-d_2 dissolved in 5.0 ml. of ether was added over a ten minute period to 15.0 milli— moles of benzylmagnesium halide (3.27 g. of freshly distil- led benzyl iodide, 2.57 g. of benzyl bromide, or 1.90 g. of benzyl chloride and 0.37 g. of magnesium) in 15 ml. of ether as described in section L. Distillation yielded 2-benzy1- l,3—dipheny1-4—methy1—2—pentanol—1,1—d2: b.p. 168—172 at 0.17 mm.; n35 1.5795-1.5812. Table VII contains the yields Of these reactions and the diastereomeric product ratios as determined from integrations of the major (T 6.87) and minor (centered at T 7.00) diastereotopic absorptions of 20% solutions of this alcohol in pyridine. 10. 11. 12. 13. J. BIBLIOGRAPHY R. Boyd and M. A. McKervey, Quart. Rev., 11, 95 (1968). Pracejus, Fortschr. Chem. Forsch., 8, 493 (1967). Velluz, J. Valls and J. Mathieu, Angew. Chem. inter- nat. Edit., 8, 778 (1967). Weinges, W. Kaltenhauser and F. Nader, Fortschr. Chem. Forsch., 8, 383 (1966). D. Morrison in "Survey of Progress in Chemistry" Vol. III., edited by A. F. Scott, Academic Press, New York, 1966, p. 147. J. Cram and F. A. Abd Elhafez, J. Am. Chem. Soc., 18, 5828 (1952). G. Dauben, G. J. Fonken and D. S. Noyce, J. Am. Chem. Soc., 18, 2579 (1956). J. Karabatsos, J. Am. Chem. Soc., 88, 1367 (1967). J. Karabatsos and N. Hsi, J. Am. Chem. Soc., 81, 2864 (1965) and references c1ted there1n. Y. Curtin, Record Chem. Pr9g£,, 15, 111 (1954). See also E. L. Eliel, "Stereochémistry of Carbon Compounds," McGraw—Hill Book Co., Inc., New York, 1962, pp. 151—156. C. Brown and J. Muzzio, J. Am. Chem. Soc., 88, 2811 (1966). J. Cram and K. R. Kopecky, J. Am. Chem. Soc., 81, 2748 (1959). H. Stocker, P. Sidisunthorn, B. M. Benjamin, and C. J. Collins, J. Am. Chem. Soc., 81, 3913 (1960). 51 14. 15. 16. 17. 18. 19. 20. 21. 22. 23‘ 24. 25. 26. 27. 28. 52 The terms enantiotopic and diastereotopic, whose use is strongly recommended, have been introduced by K. Mislow and M. Raban, "Topics in Stereochemistry," Vol. I, N. L. Allinger and E. L. Eliel editors, John Wiley and Sons Inc., New York, N. Y., 1967, Ch. I. M. van Gorkom and G. E. Hall, Quart. Rev., 11, 14 (1968). Y. Gault and H. Felkin, Bull. Soc. Chim., 1342 (1960). R. Perez-Ossorio and L. Vargas, An. Real Soc. Espan. Fis. Quim. Ser. B, 88, 809 (1964). Chem. Abstr., 88, 11320 (1965). M. Cherest, H. Felkin, and N. Prudent, Tetrahedron Letters, 2199 (1968). D. J. Cram, F. A. Abd Elhafez and N. L. Nyquist, J. Am. Chem. Soc., 18, 22 (1954). A. Horeau, A. Nousaille and K. Mislow, J. Am. Chem. Soc., 81, 4957 (1965). T. J. Leitereg and D. J. Cram, J. Am. Chem. Soc., _8, 4013 (1968). H. 0. House, "Modern Synthetic Reactions," W. A. Benjamin Inc., New York, 1965, pp. 24—26 and ref- erences cited therein. D. J. Cram and F. D. Green, J. Am. Chem. Soc., 18, 6005 (1953). For a review of problems and developments in the deter— mination of the composition of Grignard reagents see E. C. Ashby, Quart. Rev., 11, 259 (1967). G. J. Karabatsos and T. H. Althuis, Tetrahedron Letters, 4911 (1967). G. J. Karabatsos, G. C. Sonnischen, N. H51 and D. J. Fenoglio, J. Am. Chem. Soc., 88, 5067 (1967). (a) G. J. Karabatsos and R. A. Taller, Tetrahedron, 88, 3347 (1968). (b) G. J. Karabatsos and R. A. Taller, Tetrahedron, 18, 3923 (1968) and references cited therein. C. M. Suter and A. W. Weston, J. Am. Chem. Soc., 88, 533 (1942). 53 29. P. Jullien and F. Kayser, Bull. Soc. Chim. (5), 8, 700 (1937). 30. L. T. Sandborn in "Organic Synthesis," Coll. Vol. I, edited by H. Gilman and A. H. Blatt, John Wiley and Sons, Inc., New York, 1941, p. 340. 31. E. M. Schultz, J. B. Bicking, S. Mickey and F. S. Crossley, J. Am. Chem. Soc., 18, 1072 (1953).