Theoretical and experimental nonaromatic heterocyclic chemistry by Kenneth Barry Lipkowitz A thesis submitted in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY in Chemistry Montana State University © Copyright by Kenneth Barry Lipkowitz (1975) Abstract: New methodologies have been developed for the synthesis of the 6,8-dioxabicyclo[3.2.1]octane substructure; particularly the seven substituted derivatives. The unusual intramolecular ketalization induced by hydrogen over palladium of a substituted pyran has been rationalized in terms of effective hydronium ion character. An isomer enrichment scheme using a pseudo-surface of Titanium Tetrachloride in a Carbon Tetrachloride matrix reveals the unusual stability of these dioxabicyclic ketals. Carbon 14 labeling indicates the migrating group plays a passive role during the pyrolysis of N-acyllactams over calcium oxide. A dual reaction pathway, dependent on reaction conditions, has also been elucidated. This methodology permits one to prepare 2-substituted cyclic imines and heretofore difficult to prepare pyrrolidenes have been synthesized in this manner. An empirical and theoretical discussion of through-space participation in the cis-8-hetero[4.3.0]non-3-ene substructure is presented. Conformational and configurational analysis of various related heterocycles and the analysis of the stereospecificity of 2,3-disubstituted pyrrolidine quaternizations has been scrutinized. The observation that paramagnetic shift reagents induce proton shifts in sulfonium ions and the implication that these pseudocontact shifts are due to complexation of the counterion is presented. An attempt to analyze anchimeric participation with respect to binding constants of shift reagents has also been looked at.  ito Professor Richard F . Smith ii "He who has achieved oneness Should move on to twoness... Ken Kesey ITHEORETICAL AND EXPERIMENTAL NONAROMATIC HETEROCYCLIC CHEMISTRY by KENNETH BARRY LIPKOWITZ A thesis submitted in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY in Chemistry Approved: Head, Major Department MONTANA STATE ^UNIVERSITY Bozeman, Montana August, 1975 /ACKNOWLEDGMENTS In particular I wish to thank John Theodore and Gary Dirks for providing me with entertainment in the lab; Arnold Craig and Patrik Callis who were intimately associ­ ated with various portions of my research; and Brad Mundy for not only being an inspiration but for. letting me do it V my way. TABLE OF CONTENTS Page- D E D I C A T I O N ............................... i F R O NTISPIECE.......... ii VITA . '.......................... iv ACKNOWLEDGMENTS v TABLE OF CONTENTS............. vi ABSTRACT . . . . ............ ................... .. . viii Chapter 1. SYNTHESES AND INATE CHARACTERISTICS OF ■ THE 6/8-DIOXABICYCLO[3.2.1]OCTANE SUBSTRUCTURE . . . . . . . . ............. I List of Tables ............................... .2 List of Figures ........................... . . 3 List of Graphs ................. . . . . . . 6 Introduction . . . . ........................ 8 Experimental ................................. 67 Literature Cited . .............. 101 2. ON THE MECHANISM OF THE N-ACYLLACTAM . ’ CYCLIC !MINE REARRANGEMENT . . . . . . . . 109 List of T a b l e s ................... H O List of Figures . . . . . ..............■ 111 Introduction . . . . . . . . . ........... 112 vi vii D i s c u s s i o n .............................. .. . 116 Experi m e n t a l ............................ 132 Literature Cited .............. 140 3. EMPIRICAL AND THEORETICAL DISCUSSION OF THE CIS-8-HETEROBICYCLO[4.3.0]NON- 3-ENE SUBSTRUCTURE........................ 143 List of Tables . . . ............... 144 List of Figures............................... 145 List of Schemes .......................... 146 ,Introduction.......................... • • • 147 Discussion . . . . . . ............. . . . . 150 Experimental ................................. 179 ■ Literature Cited ............ 190 Page viii ABSTRACT New. methodologies have been developed for the syn­ thesis of the 6,8-dioxabicyclo[3.2.1]octane substructure; particularly the seven substituted derivatives. The un­ usual intramolecular ketalization induced by hydrogen over palladium of a substituted pyran has been rationalized in terms of effective hydronium ion character. An isomer en­ richment scheme using a pseudo-surface of Titanium Tetra­ chloride in a Carbon Tetrachloride matrix reveals the un­ usual stability of these dioxabicyclic ketals. Carbon 14 labeling indicates the migrating group plays a passive role during the pyrolysis of N-acyllactams over calcium oxide. A dual reaction pathway, dependent on reaction conditions, has also been elucidated. . This meth­ odology permits one to prepare 2-substituted cyclic imines and heretofore difficult to prepare pyrrolidenes have been synthesized in this manner. An empirical and theoretical discussion of through- space participation in the cis-8-hetero[4.3.0]non-3-ene substructure is presented. Conformational and configura­ tional analysis of various related heterocycles and the analysis of the stereospecificity of 2,3-disubstituted pyrrolidine quaternizations has been scrutinized. The ob­ servation that paramagnetic shift reagents induce proton shifts in sulfonium ions and the implication that these pseudocontact shifts are due to complexation of the coun­ terion is presented. An attempt to analyze, anchimeric participation with respect to binding constants of shift reagents has also been looked at. Chapter I SYNTHESES AND INATE CHARACTERISTICS OF THE 6,8-DIOXABICYCLO[3.2.1]OCTANE SUBSTRUCTURE 2LIST OF TABLES Table Page 1. Effects of Concentration . . . . . . . . . . 18 2. Monoalkylation Optimization . . . . . . . . . 29 3. Wurtz Coupling Conditions........................ 42 4. Isomerization Conditions . . . . . . . 44 5. Electrophiles Used in Pseudobrevicomin Synthesis................. 46 6. Possible TiCllf • Pseudobrevicomin Octahedral Complexes . 57 3Figure Page LIST OF FIGURES 1. Periodic' Cleavage of a 1,2,3-triol to a 6.8- Dioxabicyclo[3.2.1]octane . . . . . . . 11 2. Lead Tetraacetate Conversion of a Pyranyl Carbinol to a 6.8- Dioxabicyclo[3.2.1]octane . . . . . . . 12 3. Original Silverstein Brevicomin S y n t h e s i s ............................. 13 4. Current Synthesis of Brevicomin Used by United States Forest Service ........... 14 5. Mundy Synthesis of Brevicomin ............. 15 6. Formation of 6,8-Dioxabicyclo[3.2.1]octane Contaminants . ............................... 17 7. Cope Rearrangement to the Synthetically Useless D i m e r ................... 20 8. Proposed Synthesis of Brevicomin Starting With Methylvinyl Ketone and Acrylonitrile . . . . . . . . . . . . . 22 9. Proposed Synthesis of Brevicomin Starting With Methylviriyl Ketone and Methylmethacrylate................. .. . 23 10. Dimerization of Methylvinyl Ketone With Ethylvinyl Ketone to Yield the Synthetic­ ally Undesirable Methyl Ketone ............. 24 11. Appropriate.Refunctionalization of Methylvinyl Ketone Dimer .................... 25 12. Mannich Alkylation of Methylvinyl Ketone D i m e r ............................ .. • 26 413. Monoanion Alkylation of Methylvinyl Ketone Dimer ................................. 27 14. Synthesis of Brevicomin Precursor via Dianion Alkylation ...................... 28 15. Predicted Equilibrium of Kinetic and Thermodynamic Anions ........................ 31 16. Proposed Kinetically Controlled Alkylation of Methylvinyl Ketone D i m e r ................................ 32 17. Actual Alkylation of Methylvinyl Ketone D i m e r .......... . . . ................ 33 18. Reaction of Trityl Anion With Methylvinyl Ketone Dimer . : ............... 36 19. A Reasonable Ring Opening of , Ketone 62 to Produce a S u b a r o n e ........... 37■ rIi 20. Dithiari Coupling Sequence . . . . . . . . . 38 21. Model System for I,3-Dithian Ring Closure................... 38 22. Rationalization of Dithian Coupling . . . . 39 23. Double Displacement Using Diethylmalonate Anion ...................... 40 24. Configurational Change of Ketal §2 With a Lewis A c i d .......................... 43 25. Synthesis of Pseudobrevicomin . . . . . . . 45 26. Hydrogenation of Alcohol 89 to a Bicyclic Ketal ............................... 47 27. Hydrogenolysis Products of Pseudobrevicomin ............................. 50 Figure . Page 5Figure Page 28. Multiple Pathway to Hydrogenation Product . ..................................... 51 29. Unusual Grignard Cyclization of Methylvinyl Ketone Dimer to Bicyclic Ketals ............................ 58 30. Cyclization of Methylvinyl Ketone Dimer to a Bicyclic Ketal With Methyllithium ............................... 60 31. Meerwein-Pondorff-Veerley Cyclization of Methylvinyl Ketone Dimer to Pseudobrevicomin ............................. 61 32. ■ Effect of a Heteroatom...................... 62 33. Grignard Addition to Methylvinyl Ketone Dimer in Which the Heterofunctionality Has Been Replaced With a Carbon and the Vinyl Methyl Group Replaced With a P r o t o n ....................................... 63 34. Identical Reaction as in Figure 33 in Which the Carbonyl Methyl Has Been Replaced With a Slightly More Bulky Ethyl G r o u p ............................ .. . 63 35. Grignard Addition to Methylvinyl Ketone Dimer Where the Alkyl Substituents Have Been Replaced by Protons............... .. . 64 36. Grignard Addition to Methylvinyl Ketone Dimer That Has Had its Hetero Function Replaced by Carbon and Alkyl■Substituents Replaced by Protons . . . . . . . . . . . . 64 37. Proposed Synthesis of Methylvinyl Ketone Dimer Sulfur A n a l o g ................. .. 66 6LIST OF GRAPHS Graph . Page 1. Time Study of Methylvinyl Ketone-acrolein Dimerization . . „ . , . . . 19 2. . Hydrogenolysis With 10% Pd/C ......... 49 3. Hydrogenolysis With 5% R h / C ................. 49 iI .,'i I INTRODUCTION 6,8-dioxabicyclo[3»2.I]octanes in Natural Products1 Apart from sugars, in which the 6,8-dioxabicyclo- [3.2.1]octane structure I is found in abundance, it has% been noted that these bicyclic ketals are present in a variety of natural products. / X . )-- I One of the earliest references to this dioxabicyclo- [3.2.1]octane substructure was associated with a constitu­ ent of Japanese hop oil from Humulus lupulus.2 Spectral analysis, coupled with a low-yield, but unambiguous syn­ thesis, proved the constituent to have structure 2. 2 Real impetus towards a better understanding of this skeletal system came after the careful isolation and ele­ gant structure assignment of the aggregating sex pheromone 8of the female pine bark beetle, Dendroetonus breviaomis. 3 This compound, assigned structure 3, was given the trivial name, "brevicomin." The exo isomer, 3a, is the active pheromone for D. breviaomis. The endo isomer has been impli­ cated as an attractant for D. frontalis. endo A second ketal was isolated from the male pine bark beetle and was demonstrated to be the same as the aggre­ gating sex pheromone of Dendroatonus frontalis.1* This was given the trivial name, "frontalin," and was assigned structure 4. % Several bicyclic ketals, including 4,9-dioxabicyclo- [3.3.1]octanes and 6,8-dioxabicyclo[3.2.1]octanes have been isolated from tobacco.5 In the latter series, ketals of the general structure, 5, have been found. 9 R= -C -C H 3ii J O R = -CHCH3 OH 9 H3 R= -C -C H 3 OH The formation of a 6,8-dioxabicyclo[3.2.I]octane derivative during fatty acid metabolism in yeast, eqn. I, has been reported. (eqn. I) The 6,8-dioxabicyclo[3.2.1]skeletal arrangement has recently been found in a diterpene,7'ti(8). An introductory comment, albeit dilatory, is now in order. From informal discussions with Dr. William Bedard9 10 and Professor Robert Silverstein,10 the urgency for a suit­ able large-scale synthesis of earo-brevicomin was recognized. Thus, although the emphasis of this thesis is not directed towards the synthesis of brevicomin, per se, a large por­ tion of my efforts have been in that direction. I now pre­ sent a discussion of synthetic methodologies directed to­ wards the synthesis of brevicomin followed by some unique bicyclic ketal chemistry developed by myself and co-workers during my three year apprenticeship with Professor Mundy. Syntheses The 6,8-dioxabicyclo[3.2.1]octane substructure I has been synthesized both by accident and design. Peri­ odic cleavage of the triol, (9), did not give the antici­ pated produce, 13; but rather the bicyclic ketal, (1,2) .11 A rationalization for the observed product can be seen in Figure I. 11 Figure I: Periodic Cleavage of a 1,2,3-Triol to a 6,8-Dioxabicyclo[3.2.1]octane As part of a study on the rearrangement of carbonyl epoxides, Wasserman12 noted the facile formation of bicyclic ketals, exemplified by the conversion of 14 -> 15 (equation 2). This prompted Wasserman to apply the methodology to the synthesis of brevicomin (equation 3). C H 3 o i4 N 15 (eqn. 2) Q 12 (eqn. 3) A low yield (15%) cyclization of a pyranyl carbinol, (^7), to 6,8-dioxabicyclo[3.2.1]octane (I) by lead tetra­ acetate has been postulated as occurring by a radical mechanism. Figure 13. Pb(OAc) O-CH2 Figure 2: Lead Tetraacetate Conversion of a Pyranyl Carbinol to a 6,8-Dioxabicyclo[3.2.1]octane By acknowledging the carbonyl functionality as a ketal synthon, Silverstein14 was able to synthesize both 13 isomers of brevicomin by a sound synthetic scheme delin­ eated in Figure 3. An improvement currently being used by the United States Forest Service is shown in Figure 4.15 0 Ii O Il DHnlEtOH 0Ii Br(CH2)3BrtCH3CCH2 COEf CH3C(CH2)lfBr V /— ^ y /— X M OH C1V-CC.,!, Sr -gj- V e f g ^ aio O ' ' . a . i t o P r H u ml s m l m Hp mlu.H H mluH s 2 m l Deu fesfesfeRfec rXj HaSOif HzO CH3COCH3 3a and 3b 25%CiS 75%, Trans 90% Figure 3: Original Silverstein Brevicomin Synthesis 14 Intramolecular cyclization can be envisioned as an alternative method for ketal formation. Thermally4 or electronically16 induced cyclization of vinyl ethers pos­ sessing structure 29 have in fact been used as a synthetic methodology for ready access to the 6,8-dioxabicyclo- [3.2.1]octane framework. TsCI CzH5CH - CH (CH2 )z OH ~ g6o/o C2 H5 CH = IOHIO Dnl 2%R I 0 I 1 C2H5 CH = CH(CH1)2 OTS + CH3 C CH2^ OC1H5 I1 " 'I 2 37% MCPBA =-CH3 = OI i (CHz)3 C CHd D Therm^earransement I 4 H)QfH f IS0C. j 3a and ^b Reflux &55% Figure 4: Current Synthesis of Brevicomin Used by United States Forest Service 15 In his original synthesis of brevicomin, Mundy16 cyclized pyranyl carbinol (37) with an oxymercuration- demurcuration procedure shown in Figure 5. The hhao d hhF. cycloaddition of methyIvinyl ketone and acrolein, followed by addition of ethyl Grignard afforded ^7. No attempt at isolating intermediates was made, which unknowingly side­ stepped some serious problems, vida infra, facilitating a 9% yield of exo- and endo-brevicomin. Since the synthesis of brevicomin using this meth­ odology was established, it was for me to investigate how to modify the existing methods to improve yield. It was 0 EtMq Br OH h H g ( Q A c ) ^ 3a and 3h Il) Na BHif % Figure 5: Mundy Synthesis of Brevicomin reasoned that cycloaddition of methyIvinyI ketone and acro­ lein would produce four major products, shown in Figure 6, which when carried through the reaction sequence would give 16 rise to variously substituted 6,8-dioxabicyclo[3.2.1]octane contaminants; each of which exist as two (exo and endo) isomers. To negate this problem, reactant conditions were varied so as to reduce self dimerization products, 41 and 44, and increase cross dimer products, 36 and 3B (of course only one of these, 36, is of immediate synthetic value). Four test tubes containing different ratios of methylvinyl ketone and acrolein were sealed and let react at 120°C for twenty-four hours. Effects of concentration are listed in Table I. The maximum concentration of mixed dimer was obtained by starting with a 2:3 (methylvinyl ketone:acrolein) volume ratio. 17 + I CHO v OH OxyMERCURATlON- DEMERCURATlON Figure 6: Formation of 6,8-Dioxabicyclo[3.2.1]octane Contaminants Though I was not able to obtain reproducible effects from varying the temperature, an analysis of several dupli­ cate runs suggest 148-158°C as the most ideal temperature range for cross dimerization. Time plays an important role in these reactions as is shownin Graph I. One observes an increase in all dimers for the first ten hours followed by a continued increase of mixed dimer and loss of self dimer. 18 Effects of Concentration Ratio * Ac, MVK-Ac+ MVK2 + (methylvinyl ketone:acrolein) Table I 1:4 . r —— 1.4 1.0 2:3 — No Data — 3:2 1.0 3.5 — 4:1 120°C, 1.0 IOml total volume COO — t Ac2=Acrolein dimer; MVK-Ac=ItiethyIvinyI ketone-acrolein • cross dimer; MVK2 =methylvinyl ketone dimer Were the two disappearing self dimers eyeloreverting to methylvinyl ketone and acrolein then recombining to form the mixed dimer? A mixture of pure acrolein dimer and methylvinyl ketone was heated in a sealed tube for 14 hours at 140°C and by glc analysis produced no mixed dimer. A brown tar suggesting thermal decomposition of the self dimers, but no cycloreversion followed by cross dimer for­ mation, is implicit in the argument. Utilizing conditions of concentration, time and temperature which should result in optimum yield of methyl­ vinyl ketone-acrolein dimer, a large scale preparation was attempted. Mhs-S' spectral analysis of the crude product 19 Graph I: Time Study of Methylvinyl Ketone-Acrolein Dimerization Z H- t> 8 IO IZ IH 16 18 2D 22 24 TIME IN HOURS 2:3 ratio of MVK:Ac, components identified by mass spectra, reaction temperature 148°C suggested that I was, indeed, getting a product with the correct molecular weight; so a preliminary purification was in order. This was achieved by high vacuum distillation since distillation at aspirator pressures had been pre­ viously demonstrated to result in undesirable tars. Treating the distilled product with ethyl Grignard afforded a product whose mass spectrum gave the necessary molecular weight. Encouraged by this, the final cycliza- tion to brevicomin was attempted without purification or additional structure analysis. Very low yield of a 20 product similar to brevicomin was obtained; but spectral analysis assured me it was not brevicomin. After considerable "back-tracking" the problem with this synthetic scheme was recognized. This type of pyran can and does undergo a [3,3] sigmatropic shift to give the "wrong" mixed dimer as shown in Figure 7. Thermal energy * in excess of room temperature is enough to cause this Cope rearrangement; so now we more fully understand why Mundy1s original synthesis of brevicomin was of such low yield. The thermal energy needed to form the dimer is also enough to rearrange it to the "wrong" product. Processes of this sort have literature precedence.17 Figure 7: Cope Rearrangement to the Synthetically Useless Dimer This synthetic shortcoming voids this particular approach to the problem yet the methodology remains an * * Rearrangement was observed at 120°C. 21 * attractive one. If the mixed dimer, 4J/ could be formed by a Tr4 S + 3321 cycloaddition, routine modification of the nitrile functionality would yield alcohol 37 as shown in Figure 8. Attempts at making 4J at high temperature failed even with Lewis Acid catalysts. This result is surprising in light of the electron withdrawing effect of the nitrile group compared to the Z ketone. One would expect the complexed acrylonitrile to be a far better dienophile than methylvinyl ketone yet no adduct was formed. Almost any functionality which can undergo easy modifications can conceivably replace the nitrile group. A Diels-Alder adduct of methylvinyl ketone and methyl meth­ acrylate could be routinely hydrolized to the corresponding * * One reasonable approach to preventing this rear­ rangement is to modify the carbonyl group. The following reaction scheme is feasible but the ketal of methylvinyl ketone could not be formed.18 I /VWW^ 0 22 TT^S +Tr2S ' O ^ C N 4J ii) Hydrolysis %iB Na BH^ I) EtM9Br % OH Figure 8: Proposed Synthesis of Brevicomin Starting with Methylvinyl Ketone and Acrylonitrile acid and converted to 37 by addition of a two mole equiva­ lent of ethyllithium followed by reduction with sodium borohydride at low temperatures to minimize the possibility of a [3,3] sigmatropic shift. This methodology is pre­ sented in Figure 9. Unfortunately the self dimerization of methyIvinyI ketone was too competitive a reaction and only minor amounts of 48 were formed. methyIvinyl ketone-acrolein cross dimer was that the smaller group (Me vs H) ended up on the pyran ring. One might then expect the ethyl ketone (SrI) to be the more stable Diels-Alder adduct between methyIvinyI ketone and An observation in the Cope rearrangement of the 23 Figure 9: Proposed Synthesis of Brevicomin Starting with Methylvinyl Ketone and Methyl methacrylate ethylvinyl ketone. Figure 10 describes a proposed synthe­ sis of brevicomin based on the concept that the equilibri­ um for the Cope rearrangement will favor the ethyl ketone. Experimental results indicated no ethyl ketone was present. The larger group (ethyl vs methyl) ends up at­ tached to the pyran ring. These conflicting "migratory aptitudes" of "large" groups during thermal rearrangement of 36 and 51 is undoubtedly due to differences in stability of aldehydes and ketones.* The conclusion extracted from Direct formation of (5$), however, can not be pre­ cluded. In fact, the cross dimer 50 probably is formed since the ethylvinyl ketone is a better diene than methyl- vinyl ketone. 24 this experiment was that once the ketone, (51), is formed the Cope rearrangement will destroy it by rearrangement to the thermodynamically more stable, yet synthetically unde­ sirable pyran 50. Figure 10: Dimerization of Methylvinyl Ketone with Ethyl- vinyl Ketone to Yield the Synthetically Un­ desirable Methyl Ketone In the aforementioned attempts of brevicomin synthe­ sis, the facile formation and thermal stability of methyl- vinyl ketone dimer, (^1), was enough impetus to consider it a useful intermediate. Since the Cope rearrangement of this system is degenerate, no contaminants (wrong dimers) would be present. The necessary refunctionalization is shown in Figure 11. This could conceivably be carried out } 25 by the Mannich reaction, a general one-carbon alkylation procedure displayed in Figure 12. Elaboration of the methyl group to an ethyl group is envisioned as occurring by the enol of methylvinyl ketone dimer, (^2), reacting with the iminium ion prepared in situ. Mannich base (^ 3) could be quaternized then reduced to an Figure 11: Appropriate Refunctionalization of Methylvinly Ketone Dimer alcohol with concomitant removal of trimethyl amine. This resulting alcohol (3,7) could then be cyclized. Of course there was an alternative plan where the carbonyl function in 53 could be reduced and cyclized to the corresponding amino-substituted ketal which then could be quaternized and eliminated to give the same product. Though mass spectral information assured me that the Mannich alkylation had taken place, subsequent reactions did not yield brevicomin. It was finally determined that the major product from this sequence was the ring alkylated 26 Figure 12: Mannich Alkylation of Methylvinyl Ketone Dimer pyran, (54), which appears to be reflecting the greater enol stability of (SrxS) . This is not at all unreasonable in view of the greater stability generally observed in highly substituted olefins as compared to terminal olefins (viz. $5 vs. ^2). 0 27 The direction of alkylation (controlled by the sta­ bility of the enol) was indicative that anion alkylation would follow the same path and alkylate on the ring rather than the methyl group. Dr. R. Otzenberger52 tried the monoanion alkylation and indeed got ring methylation via the more stable enolate anion (^6) depicted in Figure 13. Figure 13: Monanion Alkylation of Methylvinyl Ketone Dimer Though enamine alkylations of ketones generally give the lesser substituted alkylation product (for a complete discussion on enamine alkylations see B. P . Mundy1s Canaepts in Organia Synthesis, Chapter 9),19 no enamine formation could be achieved when tried by Dr. Mundy. An alternative method of forming the kinetically stable alkylation product is the method advanced by Hauser and Harris20 in which a dianion, ($8), is produced. The second anion formed is not expected to be stabilized to any great extent, thus reactivity at the methyl group is expected to be higher than at the ring. 28 After alkylation, the remaining anion, (^9), can be neutra­ lized by any protic source such as ammonium chloride. The proposed reaction scheme is presented in Figure 14. minor Figure 14: Synthesis of Brevicomin Precursor via Dianion Alkylation Formation of ^8 followed by alkylation, protonation and subsequent cyclization yielded 6.6% exo- and endo- brevicomin as well as polyalkylated products. The major product, however, was not polyalkylated contaminants, rather unreacted methylvinyl ketone dimer which underwent further reaction as it was carried through the process. Attempts at contaminant separation by methods other than glc did not prove to be fruitful. As before, I hoped to establish conditions for maximal alkylation on the methyl group so enolates were prepared from a host of materials 29 and alkylations run in representative solvents. Di- and poly-alkylated species were always produced even under op­ timum monoalkyration conditions. Conditions are expressed in Table 2. Table 2 Monoalkylation Optimization Metal Solvent * Time % Mono- . alkylation % Di- alkylation 2K NH3 I hr 39 17 2Na NH3 . I hr 36 13 2Li N H 3 , I hr 34 25 2Na NH3 2 hr 6.1 — 2NaH . 0H I hr — - - NaH 0H . I hr . — — 2K+tBuo_ 0H I hr 27 4 * - time allowed for the dianion to form before alkyl halide was added + - obtained by triangulation.methods as % total glc peak areas ( t - Dianion polyalkylated products identified;by mass spectral analysis Alkylation on the methyl group and subsequent reac­ tions yielding brevicomin has proved itself a viable prep­ arative route. The dianion alkylation scheme, as attrac­ tive as it appears, is useless on a large scale operation 30 since attendant polyalkylated products can only be sepa­ rated with small scale glc techniques. Monanion formation, as we have seen, produces the thermodynamically stable, but synthetically undesired, alkylated product. If a proton from the carbonyl methyl group can be removed instead of * the methine proton, the less stable anion should form. Once this kinetically produced anion is in existence, no excess methyIvinyI ketone dimer should be present since one would expect this anion to abstract a proton (methine) from a methylvinyl ketone dimer molecule to form the * * thermodynamically more stable anion. An equilibrium would then be reached in which most anions were of the thermodynamically stable type. This equilibrium is demon­ strated in Figure 15. Thus, anion ((^ 1) must be prepared in a kinetically controlled experiment where methylvinyl ketone dimer (^1) is titrated into the base so an end-point can be determined. Three requirements of the base were ob­ vious : The base must be strong enough to abstract a proton, * I recognize that the methine proton could be both the thermodynamic and kinetic proton. * * The idea of an intramolecular proton abstraction is unreasonable in that the system would have to proceed by a high energy symmetry imposed barrier during the [1,3] shift. 31 it must be sterically large enough so as not to be able to come in contact with the methine proton, but rather, the methyl proton, and it must be colored so that when used up the color disappears and end point can be observed. Figure 15: Predicted Equilibrium of Kinetic and Thermo­ dynamic Anions Dr. Norman Holy (University of Western Kentucky) suggested that the anthracene radical anion was a base which fulfills all the requirements. The proposed reaction scheme is pre­ sented in Figure 16. To the prepared radical anion, kept under an inert atmosphere, was titrated methylvinyl ketone dimer until the solvated blue electron color disappeared. To this was added a large excess of alkylating agent and an immediate color change along with slight warming of the 32 reaction mixture was observed. After workup, a compound with the correct molecular weight as well as having an odor similar to encfo-brevicomin was isolated. Further spectral analysis proved the compound to be the bicyclic ketone (^2) (see experimental section for structure proof). The actual chemical sequence is shown in Figure 17. It is interesting to note that two things have hap­ pened in a one pot reaction: alkylation and cyclization to form 6^2 in 31% yield. The following observations may shed some light on the still unknown mechanism. Figure 16: Proposed Kinetically Controlled Alkylation of Methylvinyl Ketone Dimer 33 Figure 17: Actual Alkylation of Methylvinyl Ketone Dimer i 3S mole equivalent methylvinyl ketone dimer is neces­ sary to reach the end point of anthracene radical anion. ii the experimental procedure was repeated where as soon as the alkylating agent was added, one half of the mixture was removed and worked up to yield 30% (based on total glc peak areas) cyclic ketone. After two hours mixing time, the second half was worked up in the same way and it too produced 30% yield of cyclic ketone. Thus there is no long term time dependence in the alkylating step. It seems that as soon as the alkylating agent is added, the reaction goes to completion. iii replacement of MeI with EtI as alkylating agent gave the corresponding product, 63. iv if the end point is reached and let stir for three hours before alkylating agent is added, the same re­ sults are obtained as if the endpoint were reached and immediately alkylated. The kinetically formed enolate anion (if indeed that is what it is) seems quite stable. v if the methylvinyl ketone dimer is titrated into the anthracene radical anion solution and protonated (worked up without alkylating agent) a qualitative 34 yield of methylvinyl ketone dimer is recovered. * Without alkylating agent, the ketone can not cyclize. vi if the end point is reached as near as possible, af­ ter workup some unreacted methylvinyl ketone dimer is detected. If the end point is not reached, methylvinyl ketone dimer remains which suggests the mechanism in Figure 15 if operative is slow. vii the same products are obtained using the anthracene radical anion prepared from Li, Na and K metals. viii also observed in the reaction are side products with molecular weights 168 and 182. These products are presumed to be the di and trialkylated cyclic ke­ tones (§4) and (65) analyzed only by mass spectrom­ etry. 0 - ^ 0tv Mechanistically the reaction may be proceeding via radical, radical anion or a full anionic mechanism. To gain further insight as to the type process we are dealing with, another base (fulfilling the requirements) which is known not to possess radical character was used. The *This reaction should be contrasted with the Grig- nard additions to methylvinyl ketone dimer presented near the end of this chapter. 35 reaction involves preparation of trityl anion in diethyl ether using Hauser's method.21 Methylvinyl ketone dimer was titrated into the anion solution until the red color dis­ appeared (in this case 1/3 mole equivalent was necessary to reach an end point). Immediately, a four molar excess (based on trityl anion) alkylating agent was added and the reaction mixture was allowed to react thirty minutes. Af­ ter workup, a compound was isolated in 20% yield with the structure (SrJ) , as shown in Figure 18. Confirmation of this structure was fulfilled by the reaction of ketone (517) with methyl Grignard to form the tetramethyl-6,8-dioxabi- cyclo[3.2.1]octane (^6). More concerning this unique cy- clization will be presented later. Also, reduction of the carbonyl followed by electrophilic cyclization to SrJ lends unambiguous proof. Trityl anion had removed a proton (methine) produc­ ing the thermodynamically more stable enolate anion which eventually alkylated on the pyran ring. This clearly de­ fined anion reaction stands in sharp contrast to the radi­ cal anion reaction. The reaction of methyIvinyl ketone dimer with anthracene radical anion proceeding via a full anion mechanism is now viewed as an unlikely process. 36 Figure 18: Reaction of Trityl Anion with Methylvinyl Ketone Dimer Ketone (f^ 2) in the presence of HCl or BF 31OEt2ZHOAc reacted exothermically to give an intractable black tar. The product took on the distinct sickly-sweet smell of a subarone which in light of rearrangement of 68 with a Lewis acid to subarone 69 is not unreasonable.22 A feasible 37 mechanism is presented in Figure 19. Though the work in this area has not been done, these observations and ideas are presented merely to act as a progenitor for future research. Figure 19: A Reasonable Ring Opening of Ketone 62 to Produce a Subarone All attempted and actual syntheses of the dioxabi- cyclic ring system presented, have as the final step the formation of the ketal. An attractive alternative method is presented in Figure 20, where the ketal is formed early in the synthesis. The use of the dithian anion as a link­ ing device has gained sudden popularity due to the efforts of Corey et al.23 To test the feasibility of this 38 I,3-dithian ring closure, a model system was prepared from epichlorohydrin (Figure 21). 3 Figure 20: Dithian Coupling Sequence l > / c , M l A CtCH2CH(OBt)2. 4*- Cl (PH TsOH^A OH eSrm "Hfw Cl 3.5 * l.0(Trans: CIS) H % Figure 21: Model System for I ,3-Dithian Ring Closure The reaction can be imagined (Figure 22) to proceed as follows: Dithianyl anion (^3) is prepared from 1,3- dithian and a strong base. To 8^ 3, the dihalide (^ 2) is 39 added, forming the coupled product (§4). The process is repeated to displace the second halide, finally forming 85. Desulfurization with Raney Nickel is the final reduction step as previously shown in Figure 20. Figure 22: Rationalisation of Dithian Coupling With the aid of Dr. Otzenberger, it was demonstrated that formation of the anion and addition of halide resulted in a quantitative yield of I,3-dithian upon workup, even after 44.5 hours reaction time. Twenty-four hours of gentle heating likewise had no effect nor did twenty-one hours of vigorous refluxing. Obviously the first anion * formed was not displacing the halide. As a remedy an * * Using hindsight reflection, it is a possibility that the ketal proton in §2 had been abstracted by 8^3 to form dithian, rendering this reaction sequence useless. 40 attempt to replace chlorine by iodine (^2b) in hopes that a better leaving group would assist the reaction was tried but no chlorine displacement was observed and §2b was not formed. The Perkin type double displacement seemed too rea­ sonable to abandon so another anion was tried (Figure 23). EtOi "0 +SCH(OOtEf)2. O - ^ / C I K2 ^ C O 2Ef Figure 23: Double Displacement Using Diethylmalonate Anion The diethylmalonate anion formed with sodium hydride was reacted with dihalide §2 and refluxed twenty-four hours; however, no reaction was observed. 41 The classical method of connecting two halides is the Wurtz reaction. Though Wurtz coupling would produce the wrong size ring as compared to the Perkin cyclizations, an investigation, nevertheless, was in order. 'T T If this type of coupling had been successful, elong­ ation of one chain from chloromethyl to chloroethyl so as to produce the correct ring size was considered trivial. In general the Wurtz reaction was unsuccessful. Reaction conditions and observations are listed in Table 3. The possibility of two distinct isomers of the di- chloro ketal 8^2 exists and in fact were both observed. During the coupling reactions where some product was ob­ tained, only one peak (longer glc retention time) decreased in size. The short retention isomer was assigned trans 42 Table 3 Wurtz Coupling Conditions Metal Solvent Remarks Zn EtOH-H2O NR see Li Et2O N R (6 hr) Na xylene <2% Na benzene NR K xylene NR Na diglyme <0.5% Na diglyme(anh) <2% Na THR(anh) NR Na toluene NR Na dioxane(anh) <5% reference 24 stereochemistry and the longer retention isomer as cis. This is based solely from molecular model studies where the trans dihalide, if coupled, would place an unbearably high ring strain on the system whereas the cis would not and should preferentially cyclize. The mixture of isomers (originally a 3.5:1.0 trans:cis mixture) can be modified to a synthetically more useful ratio by a Lewis acid. Interconversion of cis-trans isomers occurs by complexation 43 of Lewis acid with an oxygen to form a planar carbonium ion, 88.25 Recyclization followed by loss of catalyst re­ sults in a change of configuration described in Figure 24. H ^ H | 'o o C H zCI f-H H C O % ,CH2Cl TTans H ««• as % vt n t CIS 87 Figure 24: Configurational Change of Ketal 82 With a Lewis Acid Isomer ratios are listed in Table 4 with reaction conditions. 44 Table 4 Isomerization Conditions Conditions Neat, RT I volume equiv. BF3 eOEt2,RT I volume equiv. BF 30OEt2, warmed I volume equiv. BF30OEt2, refluxed I.0/3.5 I.0/1.2 1.0/1.06 1.0/0.81 * - Ratios obtained by glc analysis on the assumption the longer retention isomer has cis.stereochemistry since no coupling procedure has been advanced, are pre­ sented as stimulus for future work. A perusal of the aforementioned attempts of high yield brevicomin syntheses makes one acutely aware of the difficulties involved in the refunctionalization (Figure 11) of methyIvinyl ketone dimer. At this time I would like to "shift gears" and discuss some novel 6,8- dioxabicyclo[3.2.1]octane syntheses in general as well as some miscellaneous reactions these ring systems undergo. which we gave the trivial name pseudobrevicomin, is quite easily synthesized (Figure 25). Ketal (90) was originally thought of as a cheap replacement for brevicomin itself. These results, though seemingly of no consequence Though the ethyl substituted ketal (^)(brevicomin) is hard to prepare in high yields, the nor-ketal (9^ 0) , J 45 Field tests with D. Frontalis and D. Brevioomis have been car­ ried out by the Forest Service group at Berkeley. These tests with D. brevioomis demonstrate no activity. 26 Figure 25: Synethesis of Pseudobrevicomin27 Reduction of methylvinyl ketone dimer with sodium borohydride or lithium aluminum hydride results in a 50:50 (glc analysis) mixture of alcohols (8^ 9) . These alcohols in the presence of an electrophile cyclize to ^O.28 A list of electrophilic reagents used and exo:endo ratios of pseudobrevicomin are presented in Table 5. Most unusual is the last entry in Table 5. When the enol ether (89) is placed over 10% Pd and 40 pounds hydrogen pressure, exo-5,7-dimethyl 46 Electrophiles Used in Pseudobrevicomin Syntheses Table 5 Reagent % yield Exo:Endo Hg(OAc)2 75 50:50 p-TsOH 87 64:36 (I)COCl** 92 82:18 Transition metal 13 100:0 * - The exo isomer has the methyl group proximal to the ketal "jaws"; the endo isomer is distal with respect to the "jaws" EXO H E N D O Mechanistically this can be occurring as follows: 47 ketal (%p) is obtained in 13% yield along with the ex­ pected hydrogenation product (^3)(Figure 26). Further ex- perimentaion shows that SK) is not at all stable to these conditions and extended reaction results in the hydroge- nolysis of the C 5-O8 bond to form ^3. A more detailed analysis of this ring cleavage was pursued. Beginning with a mixture of 60% exo and 40% endo * isomers of pseudobrevicomin,. (90), I observed the fast Figure 26: Hydrogenation of Alcohol 8<) to a Bicyclic Ketal *Pseudobrevicomin was not the only ketal that under­ went hydrogenolysis to form a tetrahydropyran. Ketal 94 also formed pyran. 48 hydrogenolysis of the endo isomer to the extent that in less than three hours reaction time this isomer was no longer detectable by glc. The e%o isomer, however, is much less reactive and even after thirty hours- reaction time was still detected. Molecular models demonstrate that inimical approach of the proximal isomer towards the metal surface can be rationalized as the reason for such slow reaction times. Graph 2 summarizes the hydrogenolysis results.29 It is interesting to note that of the carbon-oxygen bonds which can be cleaved, only the C 5-O8 ketal bond is ruptured specifically giving pyranyl derivatives rather than oxepins.30'54 Generally the rate of hydrogenolysis is increased as the bond being cleaved is a better leaving group supporting the preference of C 5-O8 bond cleavage to form pyran. 31 Under the same conditions, replacing the catalyst with 5% rhodium on carbon, identical products were obtained, albeit at a slower rate. As expected, the endo isomer was cleaved much faster than the exo isomer but an unexplainably increased reaction rate of exo isomer was evident between 27 and 30 hours reaction time. This less uniform reaction profile is plotted on Graph 3. The catalytic 49 hydrogenolysis of C-O bonds in cyclic ethers and ketals is well documented. 32 The formation of ketals, however, under catalytic conditions is very rare indeed 33 and this is the first bicyclic ketal to be prepared in such a fashion. Graph Graph 3 20 3< lime (hr) 7 IO 15 20 25 time (hr) -Si rSl -Qt 50 The sterochemistry about the hydroxy carbon is derived from the structure of the ketals. Two isomers, however, were separable on a 20% carbowax column. This undoubtedly arises from separation of cis and trans isomers about the pyran ring. The four isomers possible from the hydrogenolysis reaction are listed in Figure 27. The determination of ring sterochemistry is being pursued by means of computer- assisted lanthanide shift reagent studies. EfOH fT (R) Trans Figure 27: Hydrogenolysis Products of Pseudobrevicomin Formation of isomeric alcohols is thus shown to arise from a multiple path reaction as depicted in 51 Figure 28, where some alcohol is formed by the more cir­ cuitous path involving ketal formation followed by hydro- genolysis. Metal/C C-CH Figure 28: Multiple Pathway to Hydrogenation Product A relevant question which now arises is, how did the bicyclic ketal form in the first place? In a recent study by Nishimura, 33a it was possible to classify two groups of metals as to their ability to form acetals. Os­ mium, ruthenium and iridium belong to the group which cat­ alyze acetal formation weakly while rhodium, palladium and platinum belong to the group which catalyzes it efficiently. In total agreement with this, I observed no ketalization of ^9 using 5% Ru on carbon as catalyst. It is also known 52 that hydrogen dissolved in or chemisorbed on palladium and platinum has been found to be positively charged. 34 This charge is weak, corresponding to 1/15 of an electric charge per. atom but in light of the propensity for SrS to ketalize, it is not at all unreasonable to conclude that these metals are charged enough so as to act as an electro­ phile. Needless to say, when these experiments were re­ peated without catalyst, ketal formation was undetectable. The question of the electrophilicity of metal catalysts has been discussed before and it was maintained that there is enough hydronium ion character available to permit a pinacol type rearrangement of deoxy-dihydrowithaferin A to an A-nor-5-formyl derivative.35 Likewise, I feel enough hydronium ion character is present to allow the metal catalysts to act as an electrophile resulting in ring d o - sure of e n d ether (8^ 9) to the 6,8-dioxabicyclic ketal, pseudobrevicomiri. An alternative cyclization might be con­ sidered to result solely from interaction of the palladium with the 7r-bond of the enol ether. 55 However, this was dem­ onstrated not to be the case by shaking ^9 with Pd/C under a laboratory atmosphere. No cyclization was found; thus necessitating the role of hydrogen. An explantion of why 53 only exo isomer is found can not, at this writing, be ad­ vanced. Since it is well documented that powerful synergis tic effects are noted for compound mixtures in testing,36 an effective method to remove unwanted isomers is quite important. Presented below is an isomer (isomeric about C 7) enrichment scheme making use of the unusual stability of the 6,8-dioxabicycld[3.2.1]octane ketals. Substituted I,3-dioxalanes and dioxane derivatives have been generally observed to suffer cleavage and rear­ rangement to an ester upon hydrolysis of the complex formed with titanium tetrachloride.37 And, although this has been a well known Lewis acid, capable of complexing ligands having a heteroatomic functionality,3 8 no bicyclic ketal complexes have been reported. I observed, however, that titanium tetrachloride readily forms a complex with pseudobrevicomin and this complex can be hydrolyzed to recover .93% of the initial ketal. As mentioned above, hydrogenolysis of the endo pseudobrevicomin (9^ )) proceeds much faster than the exo isomer due to a steric effect of the exo methyl group on the catalyst surface. I reasoned that this effect could be used to my advantage by preparing a titanium 54 tetrachloride "surface" on which one isomer may selec­ tively interact. This was accomplished by preparing a di­ lute titanium tetrachloride-carbon tetrachloride matrix at liquid nitrogen temperature. Typically 0.005 mole of titanium tetrachloride in 20ml carbon tetrachloride was frozen. To this "surface" was added 0.01 mole ketal (>2 molar excess) and the solution was allowed to warm, unper­ turbed, to ambient temperature. The complex was filtered through a fritted glass filter and the filtrate was reduced in volume. Glc analysis indicated that, as expected, the endo isomer was selectively complexed. Hydrolysis of the filtered complex followed by extraction with methylene . chloride yielded an enriched endo isomer mixture. Experi­ mentally if no solvent is present, selectivity is de­ creased. This enrichment procedure can be repeated as many times as necessary to reach a desired isomeric purity. Starting with pseudobrevicomin having an exo:endo ratio of 15.9:9.4, three cycles increased the ratio to 16:2.7. This constitutes an enrichment of 71.5% by an experimentally simple procedure. One might expect the 7-phenyl-6,8^ dioxabicyclo[3.2.1]octane (^ 6) to demonstrate even greater selectivity due to the bulky phenyl group, but experimental 55 evidence proved otherwise. The endo isomer preferenti­ ally complexes but isomer enhancement is only 28%. This can be rationalized by the fact that the tentatively as­ signed endo isomer is solid and the exo isomer is liquid at room temperature. At low temperatures, both are solid and are not expected to behave as the liquid pseudobrevicomin. Titanium tetrachloride in the presence of tetrahy- drofuran and similar ethers usually complexes in a 2:1 ratio of ligand to metal with cis stereochemistry.39 Sev­ eral octahedral complexes can be envisioned for TiCl4*- pseudobrevicomin. The ketal function, though having a small bite, is capable of acting as a mono or bidentate ligand.5 3 A list of possible octahedral complexes with calculated elemental analyses is presented in Table 6. *Configuration of phenyl isomers are as yet unknown, but on the basis of glc retention times, etco-dioxabicyclics are generally shorter than endo, and from the titanium tet­ rachloride complexation studies, the long retention isomer is tentatively assigned as the endo isomer. 56 T i a^ +ZTHI a - Z ? ™ " V Mass spectrometry was inconclusive in molecular weight determination as was NMR spectroscopy for struc­ tural analysis. None of the presented structures account for the actual experimental analysis. The method of prep­ aration was that of Muetterties' second procedure without further purification.40 It seemed quite possible that com­ plex inhomogeneity could account for the reason poor ele­ mental analyses were observed. Several different ratios of pseudobrevicomin and titanium tetrachloride were com- plexed and dried for different lengths of time at high vacuum in an Abder-Holden drying pistol. Halogen titra­ tions run by Raima barter consistently gave the same per­ centage halogen indicating purity was not the problem. Higher alkylated (at the 7-position) 6,8- dioxabicyclo[3.2.1]octanes are possible by a novel cycli- zation developed by roommate and lab cohort, Gary Dirks. 4 57 Possible TiCl4'Pseudobrevicomin Octahedral Complexes %C %H %C1 Table 6 MW = 474 40.51 5.91 29.96 MW = 332 28.92 4.22 42.77 MW = 403 47.64 6.95 17.62 MW = ? 29.65 4.48 21.35 It was demonstrated that a typical Grignard addition to the methyl carbonyl of methyIvinyI ketone dimer does not give the corresponding tertiary alcohol, rather as shown in Figure 29, the appropriate bicyclic ketal. The follow­ ing 7-disubstituted ketals were prepared in this fashion: 58 Figure 29: Unusual Grignard Cyclization of Methylvinyl Ketone Dimer to Bicyclic Ketals R \ ^ \ R * SG R R1 IlCMOS R 3 = C H 3 %%% R R 3 CH 3; R1 IlKXIl C 6H %%% R R 3 CH 3; R1 = H ; R2 = Et R R 3 CH3; R1 = H ; R2 = iPr I4 R = R2 = R 3 = CH3; R1 = H * For the synthesis of exo -7-phenyl-6,8-dioxabicyclo- [3.2.1]octane by a novel photochemical reaction see reference 41. 59 m Each ketal is capable of exo-endo isomerism at C-I and as expected are formed and separable by glc. The iso­ propyl compound , however, exhibits only one glc peak with all columns tried. It is unknown at this writing whether one or two isomers have been prepared. The "homo- brevicomin" as of yet has not been field tested for biological activity. Mechanistically we are unsure of exactly when the ketal forms; before or after hydrolysis. The instability of the Grignard salts precludes easy preparation for spec­ troscopic and X-ray analysis but these problems are still being worked on. What is causing this cyclization to occur, the at­ tacking organo-metal or the inherent nature of the system itself? When methylvinyl ketone dimer is treated with methyllithium, the same product as the methyl Grignard re­ sults (Figure 30). Meerwein-Pondorf-Veerley reduction of methylvinyl ketone dimer does not result in alcohol (89), 60 rather, pseudobrevicomin. This is in sharp contrast to the sodium borohydride and lithium aluminum hydride reac­ tions mentioned earlier (Figure 31). The addition of organonickel ir-complexes to ketones to form the same prod­ uct as Grignard additions, albeit by a completely different and as yet unknown mechanism, has been brought to my at­ tention. 42 An attempt to add an organonickel of type to the methylvinyl ketone dimer carbonyl will be under­ taken shortly. R Figure 30: Cyclization of Methylvinyl Ketone Dimer to a Bicyclic Ketal With Methyllithium 61 Figure 31: Meerwein-Pondorf-Veerley Cyclization of Methylvinyl Ketone Dimer to Pseudobrevicomin A prediction as to the type of product (cyclization or simple addition) will not be given at this writing. The presence of a heteroatom, because of its dif­ ferent nuclear kinetic and electron Coulombic interactions, changes not only the inherent physical properties of a system but the way that system will react when perturbed. This thesis, in part, is a study of the role of heteroatoms from both a theoretical and experimental point of view. As such, I found it necessary to compare carbocyclic and heterocyclic analogs as well as their substituent func­ tionality so as to gain further insight into the role heteroatoms play in synthesis. 62 An earlier discussion of the reaction of 2-acetyl- 3,4-2H-dihydropyran (4J-) with organometallies to form cy­ clic ketals has been presented. Removal of the oxygen atom in 4J- (i.e. , compound , under the same reaction con­ ditions does not result in cyclization to an ether (Fig­ ure 32), rather, the simple 1,2 addition product 109. MeMq Br Figure 32: Effect of a Heteroatom The presence of an alkyl substituent, because of steric and/or electronic effects can also alter the sys­ tem. In Figure 33, replacement of the vinyl methyl group with a proton under the same conditions is seen not to cause any substantial difference in the mode of 63 reaction than that of . The change of functionality from a methyl ketone 4^ to a slightly more bulky ethyl ketone, (^13), likewise does not affect the reaction to a noticeable extent (Figure 34). Me MgBr Figure 33: Grignard Addition to Methylvinyl Ketone Dimer in Which the Heterofunctionality Has Been Re­ placed With a Carbon and the Vinyl Methyl Group Replaced With a Proton Figure 34: Identical Reaction as in Figure 33 in Which the Carbonyl Methyl Has Been Replaced With a Slightly More Bulky Ethyl Group 64 Removal of both methyl groups shown in Figure 35 results in an appreciable difference in products. Not only is the usual 1,2 addition observed but both isomers of bicyclic acetal (115) as well. Once again, to demon­ strate the importance of heteroatoms, the removal of both alkyl groups and heteroatoms results in simple addition with no trace of cyclization observable (Figure 36). M e M g Br Figure 35: Grignard Addition to Methylvinyl Ketone Where the Alkyl Substituents Have Been Replaced by Protons Grignard Addition to Methylvinyl Ketone Dimer That Has Had Its Heterofunction Replaced by Carbon and Alkyl Substituents Replaced by Protons Figure 36: 65 (The synthesis of is described in the experimental section under the preparation of acetylcyclohex-3-ene (111).) (For similar reactions involving Grignard addi­ tions to substituted pyrans see reference 43.) Obviously a requirement for cyclization is that an oxygen be present in the 6-membered ring. Whether the ketalizations are a consequence of an inherent property of the starting material or (as noted earlier) the extraor­ dinary stability of the 6,8-dioxabicyclic substructure or a combination of these two is still unknown. Though not paramount, the methyl alkyl groups play an important role in the reaction of these systems. If the oxygen heteroatoms in methylvinyl ketone dimer are replaced by sulfur, the above reactions could be repeated, and from information gathered, an assessment of how different heteroatoms react under the same conditions can be ascertained. Figure 37 shows a proposed synthesis of the sulfur analog (^ .2^ .) of methylvinyl ketone dimer. The reaction of the ketone to the thioketone could possibly be followed by a [3,3] sigmatropic shift where a carbonyl is produced to once again be converted to a thiocarbonyl. It is conceivable that from this thiocarbonyl the 6,8- dithiabicyclo [3.2.1] octane (^^) could be prepared. In 66 light of the work presented by Eleil concerning structure, reactivities and stabilities of I ,3-dioxolanes vs 1,3- dithiolanes,44 this heretofore unknown I ,3-dithiolane would be a most welcome synthesis. If the reaction could be properly controlled, the mixed heterocyclic ketals (^ > and might be prepared. Figure 37: Proposed Synthesis of Methylvinyl Ketone Dimer Sulfur Analog 0 5 R m hR EXPERIMENTAL \ Methylvinyl ketone and acrolein were sealed in pyrex test tubes and heated to 120°C for 24 hours. The total volume in each tube was IOml so a 1:4 ratio methyIvinyl ketone:acrolein implies 2ml methyIvinyl ketone mixed with 8ml acrolein. Upon cooling, the tubes were broken and analyzed on a 10% UCON 50 HB2000 chromosorb W column by glc (triangulation). Methylvinyl Ketone - acrolein Dimerization (Graph I) On the same basis as above, a series of test tubes containing a 2:3 methylvinyl ketone:acrolein mixture was heated. At various reaction times a tube was removed and analyzed by glc. Each of the following peak areas was ob­ tained by triangulation. All injections were one micro­ liter. Compounds were identified by MS only. Acrolein Dimer Relative glc peak areas 0 8.3 20.2 25.6 28.0 46.8 57.7 Methylvinyl Ketone - acrolein'Dimerization (Table I) Time in hours 0 1 2 3 4 7.5 9 IIV 68 Time in hours 11 13.5 ■24 Relative glc peak areas 64.1 42.5 17.5 Methylvinyl ketone-acrolein dimer 0 1 2 3 4 7.5 9 11 13.5 24 0 15.2 39.9 49.0 54.6 97.1 109.0 129.0 131.1 141.9 Methylvinyl ketone dimer 0 1 2 3 4 7.5 9 11 13.5 24 0 16.3 18.8 23.8 29.9 32.1' 27.3 7.6 7.6 69 Attempted Synthesis of 2-cyano-6-methyl-3,4-dihydropyran Acrylonitrile (5.3g; 0.1 mole) in 40ml CH2Cl2 was added dropwise to BF 3eOEt2 (14.2g; 0.1 mole) in 110ml CH2Cl2. No precautions were taken to remove oxygen and water from the reaction flask before conducting the expert ment, but the flask was isolated from the atmosphere by a drying tube. Methylvinyl ketone. (7.Og; 0.1 mole) in 20ml CH2Cl2 was added dropwise. No warming of the mixture was noted. The reaction mixture was stirred 25 hours at room temperature. The resulting brown solution was poured into 600ml of ice. A considerable amount of brown residue re­ mained on the walls of the reaction flask but this was re­ moved with acetone and added to the ice mixture. The aqueous mixture was extracted with CH2Cl2 followed by EtOAc. After concentration the viscous residue was run through a 3cm x 18cm florisil column with EtOAc. This re­ moved the bulk of the dark brown color. Glc analysis showed no product as having been formed. Attempted-Synthesis of 2-ethyl-2-vinyl dioxolane (F)45 In a round bottom flask equipped with a Dean Stark water trap, 50ml benzene containing 0.1 mole methyIvinyl ketone and 0.1 mole diol with a catalytic amount of toaic 70 acid was refluxed 24 hours. i.Spl H2O,was collected and the reaction mixture was black. Concentration of this yielded a black tar which distilled in a range of 33-95°C at 0.1cm Hg. Glc analysis showed the major component as unreacted ethylene glycol. Preparation of 6-methyl-pyranyl ester (48) A) The first reaction was carried out in a three necked, 500ml flask equipped with condenser, mechanical stirrer and addition funnel. The system was maintained under an inert gas blanket. Methyl methacrylate (8.,6g; 0.1 mole) was added dropwise to a room temperature slurry of AlCl3 (13.Sg; 0.1 mole) in 200ml dry benzene. Slight exothermicity was noted and the resulting solution was colorless. Methylvinyl ketone (7.0g; 0.1 mole) in 50ml dry benzene was added dropwise upon which the mixture turned amber then brown and after 11 hours dark purple. The mixture was poured slowly into 600ml of ice water and extracted with methylene chloride (400ml CH2Cl2). The methylene chloride extracts were washed with.2x200ml H2O, dried over anhydrous MgSOlt, filtered and concentrated. Distillation yielded a colorless distillate (60-65°C at 71 IOmm Hg) which upon standing turned brown. A Beilsteiri test indicated the presence of halogen. B) Due to - the ease of methyivinyl ketone dimeri­ zation an excess methylmethacrylate was used. In a high pressure bomb were placed 235ml methylmethacrylate and 25ml methyivinyl ketone and heated at 150°C for 18 hours neat. Removal of excess methyl methacrylate and distilla­ tion (range of 107-117°C at 0.1cm Hg) produced very little cross dimer. No material balance was done. MVK2 :.MVK-methyl- Run # MVK Me-methacrylate methacrylate I 4ml Iml - . 5.5 :1.0 2 3ml 2ml I.1:1.Q 3 2ml 3ml I. 0:1.1 4 Iml 4ml I.0:4.5 Percentage of self and cross dimerization was obtained by glc analysis on a. 5' 10% SE 30 column. IR: 1745cm -I -I(ester carbonyl); 1690cm (vinyl ether); 1175cm (methyl ester); 1300-1050cm ^ (ester). NMR: 64.5 (multiplet, 2H) 63.8 (singlet, 3H); 62.2-1.9 (multiplet, 4H); 61.8 (broad singlet, 3H). Elemental analysis: calcd C; 61.54 H; 7.70 found C ; 61.70 H; ,7.91. 72 Attempted synthesis of 2 (1-oxopropyl)-6-methyl-3,4-di- hydropyran (^ ) 5ml portions of methyIvinyI ketone and ethylvinyl ketone were sealed in pyrex tubes and heated in an oven at 130°C for 18 1/2 and 24 hours. GlO analysis using a 5% SE 30 on carbowax column showed three products as being methyIvinyl ketone dimer, methyIvinyI ketone-ethylvinyl ketone cross dimer and ethylvinyl ketone dimer with reten­ tion times in that order. Mass spectral analysis of the three peaks are peak I - M W 140 MVK dimer; peak 2 - MW 154 MVK-EVK dimer; peak 3 - MW 168 EVK dimer. Methylvinyl ketone: ethylvinyl ketone Time in hours' MVK2 MVK-EVK EVK2 1:4 18.5 1.0 6.8 11.8 1:4 24 1.0 11.2 23.8 2:3 18.5 1.0 2.6 1.3 2:3 24 1.0 2.0 1.2 1:1 . 18.5 . 1.0 ' 1.4 0.6 1:1 24 1.0 . 1,6 0.7 3:2. 18.5 1.0 0.7 • 0.2 3:2 24 1.0 0.8 0.2 Glc collection of the MVK-EVK cross dimer gave a positive haloform reaction. 73 Synthesis of 2 (1-oxo-ethyl)-6-methyl-3,4-dihydropyran (^ 1) The best synthesis of methyIvinyI ketone dimer is the dimerization of methylvinyl ketone neat at 150°C for 21 hours reaction time giving 88-90% yield after high vacuum distillation. NMR 64.5 (broad triplet, (I H, vinyl)); 64.2 (broad multiplet, (I H , methine)); 62.2 (sharp singlet, (3 H , ketone-methyl)); 61.8 (methylene envelope., 4 H) ; 61.7 (broad singlet, (3 H vinyl methyl) ) . -I -IIR 1665cm (enol ether), 1715cm (carbonyl). Mass Spec­ trum found and calculated 140. Attempted synthesis of 2-(l-oxo-3-NN-dimethyl - propylamine)-3,4-dihydropyran (^ 3) The Mannich alkylation was tried following the method of Maxwell.46 In a 100ml round bottom flask (5.4g; 0.039 mole) methylvinyl ketone dimer, (I.26g, 0.016 mole) paraformaldehyde and (4.53g; 0.05 mole) dimethylamine hy­ drochloride were added to 15ml absolute ethanol and 2ml acetic acid. The mixture was heated at reflux for two hours. Extractionwith diethyl ether, drying over anhy­ drous sodium carbonate and concentration resulted in a brown liquid which produces a positive haloform test. Molecular weight found and calculated 198. 2,4-DNP mp = 262-264 °C 74 Dianion alkylation procedure and synthesis of brevicomin A 500ml flask equipped with condensor, mechanical stirrer and addition funnel was charged with 200ml liquid ammonia. Potassium (7.Sg; 0.2 mole) was added slowly fol­ lowed by a trace of ferric chloride. When the blue colored solution had cleared to a pale yellow, methyIvinyl ketone dimer (14.Og; 0.1 mole) was added and the reaction mixture was allowed to stir for 30 minutes. After this time, an equivalent of methyl iodide (14g) in 50ml anhy­ drous ether was slowly added over 20 minutes. An addi­ tional 30 minutes of Stirring was followed by the addition of ammonium chloride. The ammonia was allowed to evapo- 1 rate continuously replacing lost solvent with anhydrous methanol. The resulting methanol solution was treated with a 2 molar excess NaBH4 for I hour. Addition of water and extraction with CH2Cl2 followed by drying over anh. MgSO4 and reduction in volume resulted in 88% (based on expected alcohol of crude alcohol. A portion of this alcohol mixture was submitted to the Oxymercuration- demercuration procedure47 to yield brevicomin in 6.6% yield. Product obtained this way had identical glc reten­ tion times and mass spectral fragmentation patterns as an ■ original sample, supplied by the Berkely group. Glc 75 analysis was done on an apiazon column and peak areas were obtained by triangulation. Synthesis of I,5-dimethyl-8-oxabicyclo[3.2.1]octane-6-one In a 500ml round bottom flask equipped with addi­ tion funnel, magnetic stirrer, condenser and nitrogen bub­ bler was added 400ml freshly distilled tetrahydrofuran and anthracene (6.85g; 0.035 mole). The solution was thorough­ ly purged with dry nitrogen after which a small positive pressure was maintained. Sodium metal (0.81g; 0.035 mole) was added and the reaction mixture was allowed to stir for 24 hours. Methylvinyl ketone dimer (2.7g; 0.018 mole) was. added into the reaction vessel via addition funnel, result­ ing in a color change from blue to light brown. At this point, methyl iodide (25g) was added and the solution changed to a slightly lighter shade of brown. The reac­ tion mixture was reduced in volume and filtered through celite. Most of the anthracene could be removed by addi­ tion of ether and cooling for selective crystallization. Short path distillation, 70°C at 10mm Hg, yielded a mixture of products of which 31% was (52. The major by-products from the reaction were unreacted starting material (22%) and a monoalkylated product of molecular weight 168 (37%) . 76 Separation was effected on a 10', 1/4" 10% UCON-50 column operating at 140°C. IR: 1745cm 1 sharp and strong, -I _i 1735cm sharp and medium, 1360cm sharp and medium, -I -I1240cm sharp and medium, 1120-1090cm broad and strong. NMR: 64.5 (singlet, I H); 62.3 (singlet, 3 H); 62.1 (sing­ let, I H); 61.8 (methylene envelope, 6 H); 61.3 (singlet, 3 H(B)). The assignment of the upfield three proton singlet as B was made on the observed shifts of the two methyl singlets in the presence of Eu(fod)3. The downfield methyl group shifted at a much faster rate upon incremental addi­ tion of shift reagents. This methyl singlet was assigned as being nearer the carbonyl. MS calculated and found 154. Elemental analysis calculated: C; 70.14 H ; 9.08, found C; 70.04 H; 9.29 77 Synthesis of l-methyl-5-ethyl-..8-oxabicyclo [ 3.2.1] octane- g-qne-J^) The procedure for this synthesis is the same as for 62, substituting a four molar excess of ethyl iodide as the alkylating agent. Glc analysis indicates two major compo­ nents with increasing retention' time, as' methyIvinyl ketone dimer (MW = 140) and (MW = 168). Separation is af­ fected by a 101 1/4" 10% UCON-50 column operating at 152°C. -I -IIR: 1670cm (strong sharp doublet); 1615cm (strong and - I - sharp); 1360cm (strong and sharp). NMR: 62.8-2.3 (CH2 buried under a quartet, 4 H , J=4H ); 62.1 (singlet, 3 H); 62.0 (methylene envelope, 4 H); 61.9-1.5 (CH2 buried under a triplet, J=4H 3); spin decoupling collapsed the triplet as expected. MS: calculated and found 16 8. Elemental anal­ ysis calculated: C; 75.30 H; 9.56, found: C; 75.23 H; 9.37. Synthesis of 2-(1-oxoethyl)-2,6-dimethyl-3,4-dihydropyran E t ; . ; ~ A 500ml round bottom flask equipped with condenser, mechanical stirrer and addition funnel was charged with 300ml of liquid ammonia. Potassium (I.37g; 0.035 mole), was added and. allowed to react. Triphenylmethkne (8.45g; 0.035 mole) in 150ml anhydrous diethyl ether was added LL 78 : dropwise over a five minute period. The reaction mixture was allowed to stir and warm to room temperature. Anhy­ drous ether was added to replace the lost ammonia. After reaching room temperature, the orange colored mixture was refluxed for two hours resulting in a deep red suspension. Methylvinyl ketone dimer (I.Sg) in 15ml anh. ether was ti­ trated until a dark green color was noted. Immediately an j excess of methyl iodide was added (20ml). The reaction- mixture was allowed to stir at reflux for thirty minutes after which it was cooled. The solid was filtered and washed with ether. The combined ether extracts were re- duced in volume yielding a semisolid mass containing tri- pheny!methane and reaction products. Distillation of the crude product (85°C at 0.5mm Hg) yielded I.Sg clear liquid. Glc analysis indicated three components in the ratio of -it (increasing retention times) 0.6:1.5:1.0. The first peak was characterized by glc coinjections and mass spectral analysis as unreacted methyIviriyl ketone dimer, the second j peak as ketone (EjJ) and last peak as a polyalkylated spe- -1 , cies with molecular weight of 168. IR: 1715cm (car- — 1 ' ■’ bony I) ; 1675cm (enol ether). NMR: 64,. 5 (broad, I H) ; 62.2 (singlet, 3 H); 61.8 (singlet and methylene envelope, ■ I ■7 H); 61.3 (singlet, 3 H). MS calculated and found 154; I 79 M/e (111) indicates loss of acetyl as very prominent. Ele­ mental analysis calculated: C; 70.1 H ; 9.1, found: C; 69.4 H; 9.0. ' Synthesis of 1,5,7-trimethyl-6,8-dioxabicyclo[3.2.1]octane To 0.43g in 15ml isopropanol was added 0.057g NaBH4 as a typical borohydride reduction procedure. After stirring 30 minutes, water was added and product extracted with methylene chloride. Removal of excess solvent fol­ lowed by a typical solvomercuration-demercur.ation process4 7 produced a clear liquid product separated by glc. IR: 4 bands 1200-1050cm ^ (6m cyclic ketal). NMR: 61.7-1.5 (methylene envelope, 7 H); 61.4 (singlet overlapping a multiplet, 6 H); 61.2 (doublet, 3 H). MS calculated and found 156. Elemental analysis calculated: C; 69.2 H; 10.3, found: C ; 68.9 H; 10.1. Synthesis of I,5,7,7-tetramethyl-6,8-dioxabicyclo[3.2.1]- octane (6%) In a typical small scale Grignard reaction, (0.3g; 0.002 mole) glc pure pyran BrJ is slowly added to methyl Grignard reagent and let react 15 minutes. The salts are hydrolyzed with H2O and organics extracted with 2x15ml chloroform, dried over anhydrous MgSO4, filtered and 30 reduced in volume to yield 0.27g (79%) product. One peak is distinguishable on a 20% carbowax column operating at IOO0C. IR: 4 bands 1200-1050cm ^ (cyclic ketal). NMR: 61.66 (methylene envelope, 6 H); 61.43 (singlet, 3 H); 61.3 (singlet, 3 H); 61.16 (broad singlet, 6 H). This last singlet is resolvable into two methyl singlets at 100 MH2,. MS: calculated and found 170. Elemental analy­ sis calculated: C; 70.95 H; 10.95, found: C; 70.35 H; 10.38. ; Synthesis of cis and trans 2,4-di(chloromethyl)dioxolane (&%) To 41.25g (0.5 mole) epichlorohydrin in a 500ml round bottom flask is added 200ml 1% H2SO4 and let reflux 3. hours. The solution is cooled and adjusted to pH 8 with potassium carbonate. Three 50ml diethyl ether extractions were dried over anhydrous MgSO4., filtered and reduced in volume. Distillation, 105°C at 0.75cm Hg, yielded 10.Ig 3-chloro-l,2-propanediol. IR: 3325cm (large and broad) Beilstein test was positive. 229g (0.65 mole) chloroace- taldehyde diethyl ketal and 166g (1.51 mole) 3-chloro-l,2- propanediol were refluxed seven hours in 500ml benzene with a trace of tosic acid. Ethanol and water were azeo- troped off. The resulting solution was washed with 10%. 81 potassium bicarbonate, dried over anhydrous MgSOlf and re­ duced in volume. The product was distilled at 86°C, 2mm Hg pressure but no material balance taken. A 10' 5% SE 30 column operating at IOO0C was able to separate the two -I -Iisomers. IR: 2650cm (small, sharp); 1430cm (small, -I -Isharp); 1175cm (large, sharp); 1050cm (large broad doublet); 780-740cm (large, broad). NMR: 65.4-5.1 (multiplet, I H , J=3.5H3); 64.6-4.2 (multiplet, I H , J=6H3); 64.1-3.9 (multiplet, 2 H , J=6H3); 63.7-3.5 (multi­ plet, 4 H). This is the spectrum of the cis-trans isomer mixture. Mass spectrum calculated 171; both isomers found 171. Attempted synthesis of I,3-dithia-3-oxa-51,21-epoxyspiro 5.6 dodecane (££) N2 Septum t o in t roduce reagents via syringe N H3(I) The reaction was run in a 500ml round bottom flask (above illustration) under an inert atmosphere and main­ tained at -33°C. To 6.Og (0.05 mole) 1,3-dithiane in 200ml dry THE, 5.4g (0.075 mole) n-BuLi in 90% hexane was added 82 dropwise. The resulting mixture was stirred 1.5 hours. 8.Sg (0.05 mole) dioxolane, £2, was injected neat and let stir for 6 hours. 50ml dry THF was added and cooled to dry ice/isopropanol slush temperature. '4.Sg of the second equivalent of n-BuLi in hexane was added and this mixture maintained at -33°C for 14 hours and 0°C for 4 hours. 5ml H2O was cautiously added (no reaction observed after. 2ml) and reaction mixture reduced in volume. To this, SOOrnl H2O was added and extracted with pentane which was subse­ quently washed with NaHCO3 twice and dried over anhydrous potassium carbonate. Column chromatography using Florisil as the stationary phase and CCllf as solvent yielded 5. Sg ■ recovered 1,3-dithiane. Attempted synthesis of 2-iodomethyl-4-chloromethyl-l,3- dioxolane (j3^ b) 17.Ig (0.1 mole) dichloro compound, (§2), and S3.2g (0.2 mole) KI were refluxed 24 hours in 250ml acetone. \ Workup and distillation yielded quantitative yield of. starting material. Attempted synthesis of 4,4-di(ethyloxycarbonyl)-6,8- dioxa-bicyclo [3.2.1] octane (8(5) : To I.6g NaH in 50ml diethoxyethane was added 9.6g (0.06 mole)'diethylmalonate. One mole equivalent 83 dioxolane, (^2), was added dropwise and let stir and reflux 24 hours. The reaction yielded a solid mass of salts to which water was added and organics extracted with diethyl ether. Glc analysis using a 101 5% SE 30 column and 10' UCON 50 column both operating at 150°C indicated no reac­ tion had taken place. Attempted preparation of 5,7-dioxabicyclo[3.2.1]heptane A) Zinc metal.24 In a 100ml round bottom flask equipped with condenser, dropping funnel and mechanical stirrer was placed 36ml EtOH7 3.6ml H2O and 25g (0.38 mole) zinc powder. The mixture was continuously stirred to pre­ vent caking of the metal and brought to reflux. At this temperature, 15g (0.1 mole) dihalide was added dropwise over a 20 minute period and let reflux 24 hours. The liquid was decanted and zinc powder washed with EtOH. Fil­ tration and concentration followed by glc analysis with a 151 5% SE 30 column showed no reaction had taken place. B) Group IA metals. In a 100ml round bottom flask equipped with addition funnel, condenser and magnetic stirrer, 3g (0.02 mole) dihalide in 20ml solvent was added dropwise to 0.04 mole (0.92g Na7 0.78g K 7 0.28g Li) reflux­ ing metal in the same solvent. The mixture was allowed to t! I 84 reflux 24 hours where upon cooling, water was added cau­ tiously and layers separated. Concentration was followed by glc analysis on a 151 5% SE 30 column. BF ?*OEt, isomerization of cis/trans dioxolan (^ 12) To approximately one ml of cis/trans mixture, one ml BF 3*OEt2 was ,added. The mixture was heated and solution turned black. Glc analysis was taken neat from solution and analyzed on a 101 5% SE 30 column operating at 120°C. Areas were obtained by triangulation methods. Synthesis of 2 (1-hydroxyethyl)-6-methyl-2 > 3,4-trihydropyran (&%) To 3.78g (0.1 mole) NaBH4 stirred in anhydrous MeOH (O0C), 14g (0.1 mole) methylvinyl ketone dimer, ^l, was slowly added and let stir 30 minutes. Water was added and product extracted with CH2Cl2. Drying over anhydrous MgSO4, reduction in volume and distillation, 45-50°C at 0.5mm Hg, produced a clear liquid alcohol. Two peaks are resolvable on a 15' 10% UCON 50 column operating at 125°C. Phenyl urethane mp = 150-152°C. IR: 3700cm ^ (broad alco­ hol stretch).. MS: calculated and found 142. I 85 Synthesis of exo and endo 5,7-dimethyi-6,8-dioxabicyclo- [3.2.Ijoctane (j^jg)(pseudobrevicomin) A. To 15.95g (0.95 mole) mercuric acetate stirred in 50ml anhydrous THF, 7.Ig (0.95 mole) alcohol 89 was added and let stir 30 minutes. 52ml 3N NaOH was added and the . yellow solution allowed to stir 15 minutes. I .14g NaBH4 in 60ml 3N NaOH was added and let stir 10 minutes as Hg pre­ cipitates. A saturated salt solution was added and Hg de­ canted. Three 25ml CH2Cl2 extractions were combined and dried over anhydrous MgSO4 and concentrated. Distillation 45-46°C at 12mm Hg, yields a 75% yield of exo and- endo (50:50) pseudobrevicomin; isomers were separable on a 10% UCON 50 column. IR: 4 peaks 1200-1050cm (cyclic ketal) . NMR: (exo isomer) 64.05 (quartet I H, J=6cps); 63.9 (sing­ let, I H); 61.55 (multiplet, 6 H); 61.3 (singlet, 3 H); 61.1 (doublet,. 3 H, J=6cps) , (endo isomer) 54.05 (multi­ plet, I H); 64.03 (singlet, I H); 61.55 (methylene enve­ lope, 6 H); 61.33 (singlet, 3 H); 61.28 (partially buried . doublet, 3 H). MS both isomers calculated and found 142. Elemental analysis calculated: C ; 67.56 H ; 9.94, found C; 67.16 H; 9.63. B. To a catalytic amount of tosic acid stirring in 30ml benzene was added I.42g (0.01 mole) alcohol 86 Slight exothermicity was noted and the reaction mixture was allowed to stir 30 minutes, upon which time the ben­ zene was removed to yield I.24g (87% yield) pseudobrevi- comin. All spectra fit those previously described. C. On a 1.42g (0.01 mole) basis, alcohol !3)3 was treated with 2,4-dinitrobenzoyl chloride as described in a text.48 Potassium carbonate was added and organics ex­ tracted 2xlOml with CH2Cl2. This was dried over anhydrous MgSO4 and concentrated to give a 92% yield of pseudobrevi- comin spectroscopically identical to the product obtained in method A. D . Transition metals - experimental described be­ low. The bicyclic ketal was spectroscopically the same as above. . E . Meerwein-Pondorf-Veerly method - experimental explained later. Transition metal ketalization and hydrogenolysis studies All reactions involving metal on solid support were run as follows. Approximately one gram of olefin was added with 0.25g to 0.50g catalyst in 20ml anhydrous ethanol. The reactions were carried out in a low pressure Parr shaker at ambient temperatures. When necessary, catalyst was filtered with the aid of celite and reduced in volume. 87 Synthesis of 2-(1-hydroxyethyl)-G^methyl-tetrahydropyran <23) A. Alcohol was submitted to the above hydroge­ nation conditions for 24 hours. After workup, a clear liquid was recovered in 80% yield. Exo pseudobrevicomin was a side product. IR: 3600-3200cm (large/ broad, -1hydrogen-bonded OH stretch), 1090cm (cyclic ether). NMR: 63.95-2.95 (broad multiplet, 3 H); 62.9 (broad singlet, I H); 61.9-1.1 (methylene envelope, 6 H); 61.19 (doublet, 3 H, J=2H g) ; 61.08 (doublet, 3 H, J=2 .SHg). MS: Two iso­ mers were separable on a 10% UCON-50 column, obviously cis and trans about the pyran ring. Each isomer had a calcu­ lated and observed molecular weight 144. Elemental analy­ sis: (Both isomers collected together) calculated, C ; 66.6 H; 11.11, found, C; 66.50 H ; 10.98. B. Methylvinyl ketone dimer 41 was hydrogenated in the typical reaction procedure.. Usual workup followed by distillation, 97°C at 2cm Hg, gave one major product ob­ served on a 10% UCON-50 column as being tetrahydropyran W in 32% yield. 88 IR: 1715cm 1 (carbonyl), 1140-1070cm 1 (6m ether). NMR: 64.0-3.2 (multiplet, 2 H); 62.2 (singlet, 3 H); 62.0-1.4 (methylene envelope, 6 H) ; 61.2 (doublet, 3 H , J=6H 3) . MS: calculated and found 142. Elemental analysis: calculated C; 67.61 H; 9.86, found C; 67.23 H ; 9.41. This tetrahydro- pyran was subjected to a two molar excess NaBHll. Typical reaction and workup resulted in alcohol ^ . All spectra were identical as were glc retention times with the sample produced by method A. No material balance was attempted. C. One gram of exo and endo 5,7-dimethyl-6,8- dioxabicyclo[3.2.1]octane (pseudobrevicomin) under normal hydrogenolysis conditions quantitatively produced alcohol ^3. All spectra and glc retention times were identical to the product, synthesized by method A. Synthesis of 2-(1-methyl-l-hydroxy)-6-methyl tetrahydro- PYra.n. In a typical hydrogenation procedure, 0.Sg tri- methylketal 9£ was hydrogenated 24 hours to quantitatively JJ 11 (I 89 -I form pyran . IR: 3450cm (broad alcohol stretch), -I -I1320-1400cm (3° alcohol), IllO-IOSOcm (ether). NMR: 60.99 (multiplet, 9 H).; 61.4 (methylene envelope, 4 H) ; 61.75 (methylene envelope, 2 H); 63.0 (doublet of doublets, I H, J=5.5, 1.0); 63.3 (multiplet, 2 H). MS: calculated ■ and found 158. Elemental analysis: ' calculated C; 6 8.35 H; 11.39, found C; 67.98 H; 11.29. . Attempted cleavage of 5.7-dimethyl-6,8-dioxabicyclo[3.2.1]- octane C^ O) carbon-oxygen with TiClli Following the procedure of Mastagli37 I .56g (0.01 mole) ketal £^ 4 was added dropwise to 1/2 mole equivalent (Ig) TiCl4 in CH2Cl2. The reaction was stirred over an ice.bath for 5 hours and hydrolized with H2O . Extraction with CH2Cl2, drying over anhydrous MgSO4 and glc analysis on a 20% carbowax column indicated no change. This failure for substituted dioxolan (£0) to react like other ketals attests, to its unusual stability. Synthesis of 5,7-dimethyl-6,8-dioxabicyclo[3.2.1]octane TiClk Complex To TiCl4 in CH2Cl2, CCl4, or any other suitable sol­ vent was added directly ketal ^ . Complex immediately forms a precipitate where upon the solid was filtered to remove most of the. solvent. Without further hesitation. 90 the complex was placed in. an Abder-Holden drying pistol under high vacuum in the presence of P2O 5 for 24 hours. . This sufficiently dries and removes any TiCl4 and solvent absorbed to the complex. Analytical samples were taken without further purification. Elemental analysis of TiCll complexes Halogen analysis. Titration for halogen was the standard Mohr procedure.49 . . Attempted titanium analysis. No end point was ob­ served for this titration. The procedure, presented by Welcher was used.50 Mass spectrometry was inconclusive since no parent ion was observed. A huge peak at 140 (MW of pseudobrevi- comin) was evident. Elemental analysis: C; 29.65 H ; 4.48 Cl; 21.35. The complex was very hygroscopic and within hours on a bench top can become a gum like resin. Warning - TiCllt is a severe mucous irritant and potentially lethal. Synthesis of exo and endo 6,7-dimethyl-7-phenyl-6,8- dioxabicyclo [3.2.1] octane ( ) This synthesis was done by G. W. Dirks on a 0.05 mole basis. The phenyl Grignard reagent was prepared in usual manner. 7g (0.05 mole) methyIvinyl ketone dimer 41 i i u r 91 in IOml ether was added dropwise with stirring. Stirring was continued until glc aliquots indicated the methyIvinyl ketone dimer was used up. Cautious addition of H2O caused immediate precipitation from the clear yellow solution. These salts were filtered with Celite filteraid and the two resulting layers separated. The aqueous layer was ex­ tracted with chloroform, organic layers combined and re­ duced in volume. Glc analysis on an SE 30 column operating at 200°C indicated mainly two peaks with long retention times in the ratio 3:1 (short retention time:long retention time) by triangulation methods. The resulting thick oil had a sweet odor, possibly remnants of the bromobenzene. A glc pure sample of the first isomer, however, was a wax like solid. This solid could be obtained much faster and with less trouble than glc collection methods by high vacuum -Isublimation of the oil neat. IR: 1200-1050cm , 4 bands (cyclic ketal). NMR: (liquid isomer) 67.24 (aromatic mul- tiplet, 5 H); 64.33 (broad singlet, I H); 61.42 (sharp sing­ let centered over methylene envelope, 10 H); (solid isomer) 67.2 (aromatic multiplet, 5 H); 64.3 (broad singlet, I H); 61.66 (methylene envelope, 6 H); 61.49 (singlet, 3 H); 61.41 (singlet, 3 H). MS both isomers calculated and found 218. Elemental analysis: short retention isomer calculated f ■ . 92 C; 77.03 H;' 8.31, found C ? 76.93 H; 8.35; long retention time isomer calculated C; 77.03 H; 8.31, found C; 77.10 H; 8.14. Synthesis of 6,7-dimethyl-7-ethyl-6,8-dioxabicyclo[3.2.1]- octane (^ ) Ethyl Grignard was prepared on a 0.007 mole scale with 0.168g Mg turnings and enough ethyl bromide to dis­ solve all the metal. I.Og (0.007 mole) methyIvinyl ketone dimer in diethyl ether was slowly added by addition funnel. Salts precipitated and the reaction was allowed to stir I hour. This mixture was hydrolyzed with H2O and extracted with CH2Cl2. Organic layers were combined, dried over an­ hydrous MgSO4 and concentrated, to give an 82.4% yield crude product. Glc analysis demonstrated two peaks, un­ reacted methyIvinyl ketone dimer and a product with longer retention time which consists of 87% of the mixture. This represents a 71.6% glc yield. This product can exist as. .. - exo and endo isomers but were not separable on 5% SE 30, . . 10% UCON 50, 20%. carbowax nor Apiazon columns. IR: 1200^ ■ _ i , 1050cm 4 bands (cyclic ketal). NMR: 63.95 (singlet, I H); 61.7 (methylene envelope, 6 H); 61.4 (singlet, 3 H); 61.3 (singlet, 3 H); 60.05 (broad ethyl multiplot, 5 H). 93 MS: calculated and found 170. Elemental analysis.: calcu­ lated C ; 70.59 H ; 10.59, found C; 70.63 H ; 10.38. Synthesis of 6,7-dimethyl-7-isopropyl-6,8-dioxabicyclo- [3.2.1] octane This work was done by G. W. Dirks. On a 0.011 mole scale, 0.267g Mg turnings and 1.35g 2-bromopropane were condensed in 50ml anhydrous diethyl ether. To this.was. added 1.54g methylvinyl ketone dimer and let stir. The reaction was hydrolized with water and worked up in the usual fashion. 1.34g crude product (71.2%) was obtained. Glc analysis on a 5% SE 30 column showed six peaks includ­ ing methylvinyl ketone dimer. NMR: 63.88 (broad singlet, I H); 62.0-1.3 (methylene envelope, 7 H); 61.28 (singlet, 3 H); 61.10 (singlet, 3 H); 60.83 (two overlapping doub­ lets, 6H, J=7H 3) . MS: calculated and found 184. Synthesis of 2,7,7-trimethyl-6,8-dioxabicyclo[3.2.1]octane (%&) This work done by G. W. Dirks. To 2.4g (0.1 mole) Mg turnings in 300ml anhydrous diethyl ether was added dropwise 14.2g MeI to prepare the methyl Grignard. 14.Og (0.1 mole) methylvinyl ketone dimer in ether was slowly added and let react. H2O was added cautiously and the ether layer was separated. The aqueous layer was extracted 94 with CH2Cl2 and organic layers combined, dried and reduced in volume. Two major peaks were observed by glc, one of which proved to be unreacted starting material. IR: 1380cm (geminal dimethyl), 1200-1040cm 1 (cyclic ketal) . NMR: 63.88 (broad singlet, I H); 61.65 (methylene enve­ lope, 6 H);1 61.40 (singlet, 3 H); 61.38 (singlet, 3 H); 61.28 (singlet, 3 H). MS:, calculated and. found 156. Elemental analysis: calculated C;. 69.20 H; 10.20/ found C ; 69.03 H ; 9.91. Meerwein-Pondorf-Veerley synthesis of exo and endo 5,7- dimethyl-6,8-dioxabicyclo [3.2.1] octane (<)0.) (pseudobrevi- comin) In a 100ml 2 neck flask equipped with magnetic stirrer and condenser was added I.40g (0.01 mole) methyl- vinyl ketone dimer, 2.04g (0.01 mole) A l (0-i-pr)3 and 15ml isopropanol. The mixture was refluxed and stirred for 6 hours and quenched with water. The salts were filtered and washed with diethyl ether. The aqueous layer was extracted with CH2Cl2 and organic layers combined. After drying over anhydrous MgSO4., 0.7g (50% yield) of pseudobrevicomin was collected. Glc analysis on a 20% carbowax column indicated exo and endo isomers in the ratio of 71.4:28.6. Glc 95 coinjection and spectral analysis are identical in all ways with the product prepared and described earlier. Preparation of l-acetyl-3-methyl-3-cyclohexene (^ p8) Methylvinyl ketone dimer was converted to the cor­ responding olefin with a modified Wittig reagent as de­ scribed by Buchi51 in 91% yield. Thermal rearrangement in sealed tubes quantitatively formed ketone All spectra are identical with the literature reported spectra. Preparation of acrolein dimer (4.^ ) This work was presented earlier for another area of study. For the Grignard additions to carbonyls, however, the acrolein dimerization by Buchi was used.51 Synthesis of acetylcyclohex-3-ene ( ) This compound was prepared by the following scheme: CH3 96 22g (0.2 mole) commercially available 1,2,3,6-tetrahydro- benzene aldehyde ^ is treated with a 0.2 mole equivalent methyl Grignard reagent in diethyl ether. This was allowed to stir 2 hours and quenched with water (during the hydrol­ ysis I lost most of my product). The product distilled at 75°C (asp. pressure) and 8.2g (33% yield) clear liquid was obtained. In a typical Jones oxidation, 2.52g (0.02 mole) al­ cohol was oxidized to ketone. Workup consisted of add­ ing enough H2O to dissolve the salts and extract with CH2Cl2. Upon drying and concentration 98% yield was rea­ lized. Glc analysis on a 20% carbowax column showed one major peak. Mass spectral analysis calculated and experi­ mental 124. No further analysis was deemed necessary. Preparation of cyclohex-3-ene-ethyl ketone (^ ^^) This ketone was prepared by the following synthetic scheme: 97 A 0.1 mole (2.4g Mg) ethyl Grignard reagent was pre­ pared in diethyl ether. To this was added tetrahydrobenz- aldehyde, A, (11.Og; 0.1 mole) dropwise and let stir 20 minutes. Hydrolysis followed by extraction with CH2Cl2 yielded 8.73g (67.2% yield) clear product. Glc analysis on a 20% carbowax column operating at 150°C shows one major product (B1 ) . IR: 3600-3110cm (large OH stretch) . NMR: 60.94 (triplet, 3 H, J=3Hz ); 61.46 (methylene envelope, 5 H); 61.90 (methylene envelope, 5 H); 63.26 (broad sing­ let, I. H) ; 65.60 (singlet, 2 H) . B1 was oxidized on a 0.02 mole scale (3.Og) by Jones reagent in acetone at 0°C. H2O was added to dissolve all salts and organics extracted 3x50ml CH2Cl2 which in turn was dried over anhydrous MgSO4. Concentration produced 99% yield.of ketone 113 which when analyzed on a 20% carbowax column was 95% pure. IR: -X 1715cm (carbonyl). NMR: Integration is unintelligible, however, quartet and triplet are clearly present J=7.5Hz . MS: calculated and found 138. Synthesis of 2-methyl-4-(methylhydroxyethyl)cyclohexene A typical methyl Grignard was prepared on a 0.01 mole basis in diethyl ether. Ketone 108.was added slowly and let react 30 minutes. Upon workup, 1.42g (94% yield! 98 alcohol was realized. Glc analysis on a 20% carbowax col­ umn indicated the reaction went to 96% completion. IR: 3600-3200cm (alcohol stretch). No further analysis was deemed necessary. Synthesis of expand endo 7-methyl-6,8-dioxabicyclo[3.2.1]- octane ( ^ ) , . Acrolein dimer, (44), as prepared by Buchi was dis­ tilled only to produce serious polymerization. A second preparation was undertaken in which the dimer was immedi­ ately added to a 2 molar excess (1.0 mole) methyl Grignard reagent. (A 1:1 mole equivalent of methyl Grignard leaves some acrolein unreacted and this eventually polmerizes making the workup difficult). Typical Grignard workup in­ cluding a diethyl ether extraction gave 20.60g (33% yield based on acrolein monomer) product. Vpc analysis on a 20% carbowax column showed four major peaks the smallest of which (shortest retention time) was unreacted acrolein - dimer. Two peaks (54% of the mixture, short retention time:long retention time = 1.28:1.00) were identified as acetal NMR: (shorter retention isomer) 65.39 (sing­ let, I H) ; 64.10 (quartet, I H, J=3.25) ; 63.85 (broad sing­ let, I H); 62.0-1.2 (methylene envelope, 6 H); 61.1 (doub­ let, 3 H, J=3.25Hg). MS: calculated and found 138. 99 Elemental analysis: calculated C; 65.63 H ; 9.38, found C; 65.62 H; 9.19. NMR: (longer retention isomer) 3.'58 (singlet, I H) ; 4.0 ..(multiplet, 3 H , J=SE^). MS: calcu­ lated and found 138. Elemental analysis: calculated C; 65.63 H; 9.38, found C; 65.68, H;. 9.38. Synthesis of 2-(l-hydroxyethyl)-3,4-dihydropyran ( ^ )^ This compound is a component of the previous experi­ mental (115). In the mixture these two isomeric alcoholsrVrVrV had the longer retention times and comprised 38% of the total mixture. The ratio of these alcohols (shorter reten­ tion time:longer retention time) is I.8:1.0. NMR: (shorter retention isomer) 66.38 (doublet split into trip­ lets, I H, J=SH2 and 0.75HZ); 64.7 (multiplet, I H); 64.0- 3.45 (multiplet, 2 H); 62.48 (broad singlet, I H); 62.2- 1.55 (methylene envelope, 4 H); 61.19 (doublet, 3 H , J=3Hz). MS: calculated and found 138. Elemental analysis: calcu­ lated C ; 65.63 H; 9.38, found G ; 65.73 H ;■ 9.44. NMR: (longer retention isomer) 66.39 (doublet almost split into triplets, I H, J=3HZ); 64.7 (multiplet, I H); 64.15-3.5 (multiplet, 2 H); 62.2-1.6 (methylene envelope, 5 H); 61.19 (doublet, 3 H, J=3HZ). MS: calculated and found 138. 100 Elemental analysis: calculated C ; 65.63 H; 9.38, found C; 65.84 H;. 9.1,7. Synthesis of 4-(1-methyl-l-hydroxypropyl)-cyclohexene (114) Methyl Grignard was prepared on a 0.007 molar basis in diethyl ether. One gram (0.007 mole) ketone was added and let react 5 minutes. Hydrolysis was followed by CH2Cl2 extraction to provide 93% of the theoretical yield. Glc analysis on a 20% carbowax column shows only 57% of the reaction mixture as being product 114 with 37% unreacted starting materials. Total glc yield = 53%. IR: 3600- 3200cm (alcohol) . NMR: 65.59 (singlet, 2 H) ; 61.7 (methylene envelope, 6 H); 61.45 (quartet, 2 H , J=3.5); 61.20 (singlet, I H); 61.04 (doublet, 3 H , J=IHz); 60.86 (triplet, 3 H, J=3.5). LITERATURE CITED ' LITERATURE CITED 1. Much of the work presented in this chapter has been supported by the United States Forest Service. 2. Y. Naya and M. Kotake, Tet. Lett, 2459 (1967) . 3. R.M. Silverstein, R.G. Brownlee, T.E. Bellas, D.L. Wood and L.E. Brown, Soienoe, 159, 889 (1968). 4. G.W. Kinser, A.F. Fentiman, Jr., T.F. Page, R.L. Foltz, J.P. Vite" and G.B. Pitman, Nature, 221, 477 (1969) . 5. E. Demole, C. Demole and D . Berthet, Helv. Chim. Aota, 57, 192 (1974) . 6. R.J. Light and J.S.V. Hunter, Biochemistry, 9^, 4289 (1970), have shown that this ketal substructure is formed in yeast from properly functionalized fatty acids: OAc This may be significant for the biosynthesis of similar ketalic pheromones of insects. 7. U . Scheideggar, K . Schaffner and 0. Jeger, Helv. Chim. Aota, 4j8, 400 (1962) and references therein. 8. R.C. Cambie, A.F. Preston and P.D. Woodgate, Aust. J. Chem., 26(8), 1821-5 (1973) and references therein. 9. B.W. Bedard, USFS Experimental Station, Berkeley, California. 10. R.M. Silverstein, S.U.N.Y. at Syracuse, Syracuse, NY. 103 11. J . S . McConaghy and J.J. Bloomfield, J. Chem. Soo., c, 7 (1968). 12. H.H. Wasserman and E .H . Barber, J. Am. Chem. Soo., 91, 3674 (1969). 13. M.L. Mihailovic, A. Mihailovic, S . Konstatinovic, J. Jankovic, Z. Cekovic and R.E. Partch, Tett., 25^ 3205 (1969). 14. T .E . Bellas, R.G. Brownlee and R.M. Silverstein, Tett., 25, 5149 (1969). 15. J.O. Rodin, C.A. Reece, R.M. Silverstein, V.H. Brown, and J. I. Degraw, J. Chem. Eng. and Data, 16, 380 (1971). 16. B .P . Mundy, R .D . Otzenberger and A.R. DeBernardis, J. Ovg. Chem., 2390 (1971). 17. R.P. Lutz and J.D. Roberts, J. Am. Chem. Soo. , 83, 2198 (1961). These workers looked at the rearrange­ ment : Their results may have little significance for a system that does not quite give a degenerate rearrangement. This problem certainly warrants further investigation. 104 18. Mr. Glen Giacoletto (undergraduate research student from 1974-1975) attempted this.ketalization under several different conditions, including as catalysts: molecular seive, p-toluenesulfonic acid, sulfuric acid and ion-exchange resin. No evidence for the desired product was seen after any of these attempts. 19. To be published by Marcel Dekker Publishing Company. 20. C.R. Hauser and T.M. Harris, Am. Chem. Soo. , 80, 6360 (1958). 21. R. Levine, E. Baumgarten and C.R. Hauser, V. Am. Chem. Soo. , 6$_, 1230 (1944). 22. R. Noyori, S. Makino and H. Takaya, Tet. Lett., 20_, 1745 (1973) . 23a. E.J. Corey and D. Seebach, Angew. Chem. Intevnat. Ed., 4 (12) , 1075 (1965). b. D. Seebach, N.R. Jones and E.J. Corey, J. Org. Chem., 33, 300 (1968). 24. R.W. Shortridge, R.A. Craig, K.W. Greenlee, J.M. Derfer and C.E. Boord, J. Am. Chem. Soo., 70, 946 (1948). 25. For a closely related acid epimerization of dioxanes see J.C. Jochims and Y. Kobayashi, Tet. Lett., 575 (1974) . 26. Personal communication from Dr. W. Bedard, PWS Experimental Station, Berkeley, California. 27. J. Renwick, Contrib. Boyoe Thompson Inst., 2_3, 355 (1967). 105 28a. For literature precedence of acid catalyzed cycliza- tions of pyrans to bicyclic ketals, see R .R . Whet­ stone , Shell Development Co., US Patent 2, 511, 891 (1950). CA44;8961 b. F . Tamura, K . Uehra, Y . Kubota and N. Murata, Kogyo Kagaku Zasshi, 67 (10) , 1566 (1964). CA61:10400 (1965) 29. The hydrogenolysis results with 5% Rhodium were ob­ tained by T.H. Matsko, MSU undergraduate, 1975. 30. An exception is presented by A. Lakodey and F . Weiss, Fr. I, 526, 149 (Cl. Co7c). RCOa(CHMeCH2O)n , 'x O' R“/e RCOaCHaIH^H(CH2) ltOH IZS11C SObnrs 31. H. House, Modem Synthetic Reactions 2nd Ed., W.A. Benjamin Inc., 1972, p. 24. 32a. M . Bartok, I. Torok and I. Szaba, Acta. Chim. (Budapest), 76 (4) , 417 (1973). b . V.M. Shostakovskii, M. Ya Samoilova and O.M. Nefedev, Izv. Akad. Nauk. SSSR3 Ser. Khim, IlO , 2382 (1973) . CA, 73;P29607z. c . N .I. Shuikin, R.A. Karakhanov and I. Ibrakhimov. ibid, 10, 1807 (1966). CA, 60:P10650f d. E. I. Mistrik and T . Rendko, Chem. Tech., 19(3) , 154 (1967). e. Kh. I. Areshidze and G.O. Chivadze, Soobsch. Akad. Nauk. Gvuz. SSR., 56(1), 97 (1969). CA, 72:31417q. 106 f. H.A. Jung, Brit. I, 320, 188, CA, 79:P92003k. g. S . Kyo and T. Yasui, Japan Kokai 74 11,806 CA, 8Ck 145381m, 11,810 CA, 80:145383p, 11,811 CA, 8Ckl45390p. h. W.D. Schaeffer, U.S. 3, 285, 967 CA, 66:37437m. i. F .C . Canter and A.G. Robinson III, U.S. Pat. Off. 882, 016, Off. Gas. U.S. Pat. Off., 882 (4) 1345 (1971). j. G. Geisler and F.J. Baumeister, Ger.I , 297, 089, CA, 70:P46832w. 33a. S . Nishimura and I. .Itaya, Chem. Comm., 422 (1967). b. ________ , M. Katagiri, T. Watanabe and M. Uramoto, Bull. Chem. Soo. Japan3 44, 166 (1971) . c . L.A. Hamilton, U.S. 3, 454, 596. 34. D.D. Eley, Quart. Rev. , 3_, 209 (1949) and references therein. 35. D. Lavie, Y. Kashman, E . Glotter and N. Daniel, J. Chem. Sda., (C) , 1757 (1966). 36. M. Jacobson, Inseot Sex Pheromones, Academic Press, New York, 1972. 37. P . Mastagli and M. DeNanteuil, C. R. Aoad. So. (Paris) t268, 1970 (1969) 38a. R. Feld and P . Cowe, The Organic Chemistry of Titanium, Butterworths Washington, 1965. b. P.C. Wailes, R.S.P. Coutts and H. Weigold, Organo- . metallic Chemistry of Titanium Zirconium and Hafnium, Academic Press, N.Y., NY, 1974. 39. R.S. Borden and R.N. Hammer, Inorg. Chem., 99_, 2004 (1970) . E.L. Muetterties, J. Am. Chemt Soc. 82, 1082 (1960) .40. 107 41. C . Bernasconi and G. Descotes, C. R. Aoad. So. (Paris), t280, 469 (1975). 42. . Personal Communication with Dr. L. Hegedus, Colorado State University, 1975. 43a. J. Cologne and P . Jeltsch, Bull. Soo. Ckim. Fr., 1288- . 1298 (1963). b. ________ , J. Buendia and H. Guignard, Bull. Soo. Chim. Fr., 3[, 956 (1969) and references therein. 44. For leading references to other review articles see E.L. Eleil, Bull. Soo. Chim. Fr., 2, 517 (1970). 45. T. Bruice and P . Szkiewicz, J. Am Chem. Soo., 89:14, 3568 (1973). 46. C.E. Maxwell, Organic Synthesis Coll. Vol. III. edited by Horning, J. Wiley and Sons, New York, 1955, . p. 305. 47. B.P. Mundy, R.D. Otzenberger and A.R. DeBernardis, J. Org. Chem., 36., 3830 (1971) . 48. R.L..Shriner, R.C. Fusbn and D.Y. Curtin, The Syste­ matic Identification of Organic Compounds, 4th edition, John Wiley and Sons, 1956. 49. D .N . Grindley, An Advanoed Course in Practical Inorganic Chemistry, London Butterworths, 1964. 50. F . J . Welcher, The Analytical Uses of Ethylenediaminetetra- acetic Acid, Van Nostrand Co. , Inc., Princeton, NJ 1958, p. 184. 51. G. Buchi and J. Powell, Jr., J. Am. Chem. Soc., 92:10, 3126 (1970) . 52. R.D. Otzenberger, Ph.D. Thesis, Montana State Univer sity, (1971), Part III. 53. Personal Communication with Dr. K. Emerson, Montana State University, 1974. . . 54. The numbering system used in this text (as pointed out by a reviewer of a paper) is wrong. It is, however, the most common numbering found in the literature. ^ 108 55. Personal Communication with Professor James Collman, Summer 1975. Chapter 2 ON THE MECHANISM OF THE N-ACYLLACTAM- CYCLIC IMINE REARRANGEMENT = X r , n______________ill i, H O LIST OF TABLES Table Page 1. Literature Perusal of the Ubiquitous Cyclic !mines . . . .. . . . . ■ . . . . . . . 112 . 2. Molecules Potentially Available by an N-Acyllactam Rearrangement . ........... 120 Ill u l . hK yT T l E n O I . Figure Page 1. Synthesis of Nicotine Alkaloids (after Larsen) .............................. 113 2. General Synthesis of Cyclic !mines ........... 117 3. Grundon-Reynolds Synthesis of 2-tert-butyI-1-piperidine............. H S 4. Synthesis of 2-tert-butyl-l-pyrrolidine . . . . H 9 . 5. Basic Similarities of Nigrifactin and Pomegranate Alkaloid.S y n t h e s i s ...............125 6. Attempted Synthesis of a Model Pomegranate A l k a l o i d ..................................... 127 7. Attempted Diels-Alder Trapping . . ... . . . . 127 8. Retention of 14C During Pyrolysis...............128 9. Partial Mechanism of N-Acyllactam Rearrangements .............................. 129 10. Two Possible N-Acyllactam Pyrolysis R o u t e s ......................................... 129 11. Proof of the Dual Thermal Pathway for N-Acyllactam Pyrolysis ...................... 120 12. Pyrolysis of Alkyl Substituted N-Acyllactams . . . . . . . . . . . . . . . . 121 INTRODUCTION The chemical literature is replete with 1- pyrrolines and piperidines, covering such diverse areas as natural products, proposed biosynthetic intermediates, pharmaceuticals and intermediates of various industrial processes. An incomplete, though representative, list is presented in Table I. These heterocyclic molecules have been of interest particularly after the discovery of a class of alkaloids isolated from the tobacco plant, Nioo- tiana, at the turn of the eighteenth century. Several syn­ theses of nicotine and related alkaloids are presented in Figure I. For a rather complete historical discussion with leading references see Brent Larsen's Master's Thesis (Montana State University, 1972) Table I. Literature Perusal of the Ubiquitous Cyclic !mines ' „ Dye intermediates, surfactant intermediates Precursors of plastics2 Heat Resistant polymers 3 Bread Aroma4 Antihistimines 5 Antiacetylcholinic. agents6 Antispasmodic agent, Amoebocides7 Substituted piperidines and their quaternary salts have muscarinic activity8 113 4 5 ( o r ^ — 6 7 Figure I . The Pictet Synthesis of Nicotine 11 7 Figure 2. Spath & Bretschneider Synthesis of Nicotine Figure I: Syntheses of Nicotine Alkaloids (after Larsen)1 114 ^ ( g r ^ — ' i ° / ? — » 17 18 ( s r ^ 7 Figure 3. The CraIg Synthesis of Nornicotinc and Nicotine Figure 4. Synthesis of Myosmine by Spath & Mamoli Figure I cont. 4o 115 Flgurr 5. Synlhnsis of Kortc & Schiilze-Steiner Figure 10. Synthesis of Myosmine Figure I cont DISCUSSION Germaine to this thesis is the fact that the metho­ dology developed by Mundy involves a simple condensation followed by a pyrolytic rearrangement. Generally, it was observed that coupling of an acid halide, where R=alkyl, aryl, strained cyclic or heterocyclic, with a 5-, 6-, or 7-membered lactam forms the corresponding N-acyllactam (^). Pyrolysis of this product over calcium oxide results in a loss of CO2 and H2O concomitant with rearrangement to cyclic imine 2 as presented in Figure 2. The following molecules have been prepared by this rearrangement previous to my engagement in this research area. The original concept was precipitated by I. Mura- kashi10 where N-benzoyl-4-aminobutyric acid A was pyrolized over CaO to form phenylpyrroline C. This is now presumed to go through an N-acyllactam intermediate B . 117 1 2 2 I 9 RCOCI + H n=l,2,3 Figure 2: General Synthesis of Cyclic !mines The synthesis of otherwise difficult to prepare molecules can be accomplished in this fashion. For ex­ ample, the synthesis of 2-tertbutyl-l-piperideine (^ ) is rather complicated11 (Figure 3). 118 Figure 3: Grundon-Reynolds Synthesis of 2-tert-butyl-l- piperideine Since the preparation involves the prior alkylation of an aromatic system, obvious restrictions for the general preparation of 2-tert-butyl-l-pyrrolidines and tetrahydro- azepines can be imagined. The methodology used in these laboratories has no such restrictions. Figure 4 demon­ strates the simple synthesis of a 2-tert-buty!pyrrolidine (6). Condensation of pivaloyl chloride with 2-pyrrolidine produces N-acyllactam (^), which subsequently gives rise to the otherwise difficult to prepare imine Q . This rear­ rangement has not been tried with six and seven membered rings but no major obstacles are envisioned. 119 Figure 4: Synthesis of 2-tert-butyI-1-pyrrolidine That the thermally labile cyclopropane ring rear­ ranges without opening means the energy of activation necessary to reach a transition state for thermal rear­ rangement is lower than that necessary to open the strained cyclopropane ring.12 Thus, I believe some of the more "fragile" natural products could be achieved by this metho­ dology. Table 2 lists molecules which potentially can be synthesized by an N-acyllactam rearrangement. 120 Molecules Potentially Available by an N-Acyllactam Rearrangement I. Simple alkaloids of the type characterized by Hygrin Table 2 Hygrin Cuscohygrin N N-methy lpelletierine CwtA Sedamine 121 Table 2 (cent.) II. General Alkaloid Synthesis &T" This 2-substituted pyrroli­ dine is a critical intermed­ iate in the synthesis of elaeocarpine and isoelasocar- pine.13 Routes to the prere­ quisite acid are found in reference.14 An intermediate necessary for the synthesis of dihydrode- oxyprallocerniune, a complex lycopodium alkaloid.15 Prec­ edence for this type of mole­ cule has been set in these laboratories where the fol­ lowing has been observed: CHjf CqO ™~5 * X A/ Pinidine 122 Table 2 (cent.) HCX N ^ (CHa) CH- Cassine1.6 Precedence for large chain migrations have been set in these labora­ tories by the C g migration: OaR This is an intermediate for a securinine type alkaloid.17 H 123 Ill. Alkaloids and Related Compounds of Pharmacological Interest Table 2 (cont.) Histrionicotoxin. Toxin from the skin of the South African arrow poi­ son frog, Dedrobates his- trioniaus.18 W ^ Ha)hCH3 n. Ioj lajif '""(CM1)Z^z(cH3)nCH3 n = 3 . $■ Venom of the "fire ant" Solenopsis saevissima.13 Pyracrimycin. An antibi­ otic from S. eridani.20 It has recently been shown that pyracrimycin is the same as desdanine (5. oaelestis) .21 For a recent synthesis see reference 2 2. 124 Table 2 (cont.) An unstable constituent of immature pomegranate leaves.2 3 Nigrifactin, An antibi­ otic from S. nigrifaoiens. The biosynthesis of this antibiotic has recently been demonstrated. 24 For a recent synthesis see reference 25. IV. Preparation of Biosynthetic Intermediates Hemlock alkaloids26 For a review on the bio­ synthesis of hemlock alkaloids see reference 27. A proposed biosynthetic intermediate of shihu- nine. 28 The antibiotic nigrifactin, if prepared by our methodology, involves a very similar rearrangement as the pomegranate alkaloid (Figure 5). 125 o.f\r o I 8 £o i2 ♦ 10 Figure 5: Basic Similarities Envisioned for Nigrifactin and Pomegranate Alkaloid Syntheses Although the synthesis of an antibiotic appears more useful than a minor alkaloid constituent both from a philosophical and economic viewpoint, the availability of crotonic acid suggested an attempted synthesis of the for­ mer in favor of the latter. The fact that ring size being of minor importance in these reactions as well as the ex­ pense of the six membered lactone suggested that I use pyr- rolidone as a paradigm. The synthesis of this model sys­ tem is presented in Figure 6. Transformation of crotonic acid (]^ 1) to crotonyl chloride (^2) occurred endothermi­ cally in diethyl ether. Condensation of the acid halide with 2-pyrrolidone to form the N-acyIlactarn (]^ 3) was 126 accomplished in the usual manner although in quite low yield. The literature reports the conjugated imine (^ 0) reacts to form ketone (10a)2 3 after a short period of exposure to the atmosphere. To be sure ]^4 does not have time to do this, the product from the pyrolysis of \3 was allowed to distill directly into a chloroform solution of maleic anhydride to trap the product as the Diels-Alder adduct ]^ 5 shown in Figure 7. Of course, this procedure assumes an s-cis arrangement of olefin. The actual pyroly­ sis was carried out and added to the anhydride solution. No Diels-Alder adduct was observed and in fact, the major product was 2-pyrrolidone! The reaction sequence was re­ peated using 2-piperidone. Pyrolysis of N-acyllactam £ resulted mainly in 2-piperidone! The results from these experiments indicated to me that the migrating groups in fact do play an active role during the migration. An in­ vestigation of the mechanism was promptly initiated. O K0 127 Figure 6: Attempted Synthesis of a Model Pomegranate Alkaloid Figure 7: Attempted Diels-Alder Trapping During the rearrangement, CO2 and H2O are lost. Un­ doubtedly one of the carbonyl carbons is gaining an oxygen and being extruded as a stable molecule. The use of ratio isotopes, as shown in Figure 8 indicated that the pyrro- lidone carbonyl is being lost. 14C labeled benzoic acid was converted in routine fashion to the labelled benzoyl 128 chloride, (]^ 6) , which subsequently was condensed with 2- pyrrolidone to yield N-acyllactam (17)(specific activity 246dpm/m-mole). Usual pyrolysis conditions produced 2- phenylpyrroline (specific activity 233dpm/m-mole). Spectra of this product were in every way similar to those reported in the literature.29 Within the experimental error of our counting, one could suggest that all the carbonyl carbon of the N-acyl group was maintained in the product. A par­ tial mechanism consistent with this observation 0C O aH 0C O C I % Figure 8: Retention of 14C During Pyrolysis is delineated in Figure 9. If this proposed mechanism has any merit, it implies that the migrating group is not ac­ tive but plays a passive role during the rearrangements. On one hand I have evidence suggesting active participation and on the other, non-participation of the migrating group. After considerable thought and consultation I realized that 129 two possible reaction paths are available for the pyrolysis of N-acyllactams. One is the rearrangement (route a) and the other is the thermal cleavage (route b) as shown in Figure 10. One path, (a), occurs primarily with moderate heat and the other, (b), at higher temperatures and shorter contact time. Indeed, the thermal rearrangement of N- accyllactam (£) at high temperatures and short contact time Figure 9: Partial Mechanism of N-Acyllactam Rearrangements Figure 10: Two Possible N-Acyllactam Pyrolysis Routes 130 results not in pyrroline (£) as before but in cleavage product (4) (Figure 11). Figure 11: Proof of the Dual Thermal Pathway for N-Acyllactam Pyrolysis Both methyl and ethyl N-acyllactams, (^9) and (20) , respec­ tively, were observed to behave in a similar manner (Fig­ ure 12). Slow pyrolysis produces the corresponding pro­ ducts which in the presence of methyl iodide quaternerize as all the cyclic imines should. Fast pyrolysis resulted in 2-pyrrolidone, lending support to the dual reaction pathway and to the observation that the migrating group plays no active role during pyrolysis. H 131 Figure 12: Pyrolysis of Alkyl Substituted N-Acyllactams Thermal rearrangement of the model pomegranate alkaloid has not been finished, but in view of the above findings, no problems are envisioned. EXPERIMENTAL Synthesis of 2-tert-butylpyrrolidine (^ ) A solution of 2-pyrrolidone (17g) and pyridine (7.9g) in 100ml benzene were stirred at room temperature while pivaloyl chloride (12g) in 25ml benzene was slowly added. The reaction mixture was diluted with 100ml CH2Cl2 and washed with dilute hydrochloric acid until the washings were acidic. The methylene chloride layer was separated and washed with saturated bicarbonate. This solution was dried over anhydrous sodium carbonate, filtered and reduced in volume. Distillation, 62-65°C at 0.2mm Hg, gave 13.7g (81% yield) of the desired N-acyllactam. . An equal weight mixture (6.72g) of this product and calcium oxide was heated in an open flame to yield the crude pyrroline. Distillation, IlO0C at 1.5mm Hg, gave 2.Og (40% yield) of 6. NMR: g c 63.8 (triplet of triplets . / \ protons A, 2 H , J=7.5, 1.75); 62.52 (triplet of triplets, protons C , 2 H , J=7.5, 1.75); 61.88 (multiplet, protons B , 2 H , J=7.5); 61.19 (singlet, 9 H). -I IR: 2900cm (huge C-H stretch), 1640cm (sharp C=N spike) MS: calculated and found 125. Elemental analysis: 11 > I calculated C ; 76.8 H ; 12.0 N ; 11.2, found C ; 76.8 H; 12.1 N; 10.9. 133 Attempted synthesis of 2-(prop-2-ene)-piperidine (1^ 0) IOg (0.1 mole) valerolactam and 10.4g crotonyl chlo­ ride were mixed with 7.9g (0.1 mole) pyridine in a typical condensation procedure. Normal workup (some loss due to lab technique) yielded, after distillation, 145°C at 10mm Hg, 2.9g yellow liquid. Glc analysis using a 20% carbowax column indicated two peaks. The short retention peak upon coinjection proved to be unreacted starting material and the long retention peak proved to be N-acyllactam ^ . Eighty-six percent of the mixture was product so a total yield of 15% was realized. NMR: 66.83 (multiplet, 2 H); 63.7 (broad multiplet, 2 H); 62.3 (broad multiplet, 2 H); -I -I61.82 (broad multiplet, 7 H). IR: 1670cm and 1630cm (broad doublet for two C=O stretches). I.Og of this N-acyllactam was thoroughly mixed with I.Og CaO and pyrolized. The product was collected in a vial flushed with N 2 and capped with a serum stopper to prevent air oxidation. Glc analysis on a base column.in­ dicated one major peak. Mass spectral analysis shows a 134 parent ion at 99. IR is identical with valerolactam (MW 99). Attempted synthesis of 2-(prop-2-ene)-pyrrolidine (^ 4) 2-pyrrolidone and an equalmolar volume of pyridine were added to one mole equivalent crotonoyl chloride in benzene. Usual workup gave a very low yield yellow tinted liquid product, N-acy I lactam !^ 3. A mixture of unreacted lactam and acid halide always remain. The method for sepa­ ration of N-acyIlactarn from lactam is to separate them by means of a column chromatagraphic procedure. The solid support is silica gel and the solvent is diethyl ether. The N-acyllactam eludes rapidly whereas the lactam has a longer retention time. NMR: 67.3 (multiplet, 2 H); 63.9. (triplet, 2 H, J=3.5HZ); 62.58 (multiplet, 2 H );'62.3-1.9 -I -I(multiplet, 5 H). IR: 1736cm 1685cm ( C=O groups). This N-acyllactam was pyrolized directly into a solution (one mole equivalent) of maleic anhydride in CHCl3 . This was transferred to a round bottom flask and refluxed under a nitrogen blanket 2 hours. No Diels-Alder adduct was observed. Concentration of this mixture resulted in a black-brown tar. Glc analysis on a basic column indicated one major product. Mass spectral analysis demonstrated a 135 parent ion at 85. The IR was identical (as was the glc retention time) with 2-pyrrolidone (MW=85). Synthesis of phenylpyrroline (3^ 8) In a IOOml round bottom flask, 121g (1.0 mole) ben­ zoic acid was dissolved in 300ml anhydrous ether. To this mixture was added a O.lmC sample of carboxyl-labeled ben­ zoic acid and thoroughly mixed. The mixture was concen­ trated to yield 121g of a labeled acid. This acid was con verted to a acid chloride by refluxing with SOCl2 followed by removal of excess thionyl chloride. As described ear­ lier, 1 4C-benzoyl chloride and pyrrolidone were condensed to form the pheny1-N-acyllactam (^7) which had spectral characteristics the same as reported in the literature.29 The N-acyllactam was mixed with an equal weight CaO and rearranged. The crude mixture was distilled (no material balance) and glc analysis showed one major product which was identical to that reported.29 136 Scintillation counting - sample procedure

Perhydrophthalan .......................... 151 2. Ortep Side View of the Four Possible C o n f o r m e r s ................................. 153 3. Attempted Alkylation of N-methyl Amine ^ .................... ............ .. 170 4. Preparation of Sulfonium Salt QQ . . . . . . 171 5. Lanthanide Induced Shifts of a Sulfonium Salt . . . ...................... 172 6. Possible Complexation with Sulfonium Sait Counterion............................ 174 7. Evaluation of Anchimeric Assistance During Shift Reagent Complexation .......... 177 LIST OF SCHEMES Scheme Page I. Synthesis of N-methyl and N-benzyl Tertiary A m i n e s ........................ 160 II. Benzylation Studies ........................... 163 III. Synthesis of Amine 19 . .............. 168 rXirXi INTRODUCTION The fortuitous observations and naive assumptions concerning heteroatoms, which inevitably led chemists to wrongly equate heterocyclic rings with their carbocyclic counterpart, has only recently been partially corrected. From a particle in a box point of view, replacement of one or more carbon atoms by another nucleus changes the size of that box; and the consequence of such an act, no matter how small, is that an entirely different system is being observed. The inherent physical properties of the new system as well as its ability to react when perturbed are different because a heteroatom has different nuclear kinetic and coulombic values than a carbon. The world of chemistry, as we know it thus becomes quite complex and we as scientists are obliged to simplify these complexities. Let me present an example of how this is done. If we walk along the ocean for a breath of fresh air we see sand and rocks jutting up about the beach, water, the moon, a board­ walk and so on. To simplify matters we could categorize what we see ... sand, rocks and the moon. To proceed a step beyond, we could even say these are different forms of the same thing!28 Of course, the basic text of 148 individuality must be maintained but categorization, as­ sumptions and the like are of great importance. This part of my thesis deals with studies directed towards understanding the inate properties of heteroatoms in conformationally heterogeneous systems. I shall pre­ sent some generalities as well.as point out the individu- ality of each molecule. An assumption made in this work is that nuclear kinetic energies play a very minor role and the observed results are a direct consequence of elec­ tronic effects. By electronic effects I mean distortion of molecular geometries resulting from different bond lengths and angles, resonance, inductive, neighboring group participation, dipolar and so on effects. The rela­ tive importance and contribution of these effects on the conformation and reactivity of non-aromatic heterocyclic molecules undoubtedly depends on. what heteroatom is pres­ ent, type of perturbation, electronic state of the system and factors of the like. There is a wealth of available literature which can be used to evaluate or demonstrate the existence of various isolated effects; however, there is a paucity of data to assist in efforts to distinguish how these various effects interact. Molecules of the 149 type are interesting in that all of these effects can be studied, eventually sorted out and weighted. DISCUSSION As part of a continuing study in Professor Mundy1s lab, I have been particularly interested in the conforma­ tional analysis of type ^ molecules. It is maintained that rates of electrophilic attack on the olefin portion of per- hydrophthalan (z = oxygen) proceed faster than the carbo- cyclic system via intermediate ^ which ultimately lowers the energy of activation by delocalization of charge.1 Only an eight fold increase in reactivity was observed during oxymercuration but serious questions concerning The solvolysis of a closely related series of molecules has been r e p o r t e d :1b,c * Relative rates of solvolysis 1.0 0.8 0.7 1.00 151 charge densities during these reactions have been raised.2 Such a small change of rate could easily be accounted for Figure I: Through Space Oxygen Participation During Electrophilic Addition to Perhydrophthalan 03 Co by torsional and angular distortions transmitted through the sigma framework from the oxygen. I found it necessary to evaluate with the aid of the CNDO/2 version of semiem- pirical calculations whether the molecule in fact can ex­ ist in the puckered form (Figure 2). If this puckered form were of such high energy, the above argument would be invalid since the heteroatom would not be close enough to 152 effectively overlap its lone pair electrons with the incip­ ient carbenium ion. Examination of molecular models show these 6,5-bicyclics as capable of existing in two major conformations; open and closed, in which the open form for z = oxygen and (Me)2N+ have been demonstrated as being the major conformer. 3'4 Actually four conformers are possible since in each of these major conformations exist two minor conformations; pucker of the five member ring up and pucker of the ring down. Open I Closed * I Ortep sideviews of the four conformations of jJS are pre- * sented in Figure 2. Coordinates for the CNDO/2 program were obtained directly from the X-ray structure of ^E.4 This quaternary structure was found to be in the com­ pletely open (0-0) form but evaluation of the remaining * * Quantum Chemistry Program Exchange 141 was modified for use on an XDS Sigma-7 computing facility. For program parameters and derivations of equations, see reference 5. 153 Figure 2: Ortep Sideviews of the Four Possible Conformers 15.4 Figure 2 (cont.) 155 three conformational isomers was of minor difficulty. A reflection6 of atoms 4,5 and appropriate hydrogens through the plane defined by carbon atoms 2,3,6 and 7 results in conformation C-D. Likewise the heteroatom can be reflected through the xy plane to change the pucker of the five member ring. Z Reflect y Through xy P la n e ' o-o Z y Combinations of these two reflection operations eventually gives rise to all four conformer's cartesian coordinates. Exchange of heterofunctionality with con­ comitant bond lengthening or shortening produces the ap­ propriate molecule in any one of the four conformations. 156 The results of these semiempirical SCFMO calcula­ tions show that as expected for the bicyclic system the order of stabilities is 0-0>0-C>C-O1VC-C. Total energies and differences are presented in Table I. The difference in energy of a completely opened conformer and a com- —1pletely folded one is only 4.44 Real mol .; comparable to the barrier of rotation of ethane at room temperature. One must, however, be cautious of CNDO energies since this em­ pirical scheme, is notorious for poor energies7; yet these can be considered safe "ball park" estimations. The fact that these are gas phase calculations applied to solution chemistry has not been considered. The relatively small change in energy going from the open to the closed con­ formation lends credence to the original hypothesis that: when called on to donate electrons from lone pair orbitals, the molecule does, not have an extremely big energy hill to climb over during conformational change and the heteroatom can readily be positioned close to the positively charged center. That heteroatoms can assist by a through space mechanism, albeit not explicitly argued, seems to be gain­ ing popularity in the literature (for a reasonably good 157 Conformer Table I CNDO/2. Results Total Energy . in Hartrees Difference in Kcal mol""I 0-0 -96.302808 0.00 O-C -96.29869.2 2.54 C-O -96.295737 ■ 4.44 0 1 O -96.295348 4.67 review of through space heteroatom participation, see ref erence 8) . One might ask if an inverse relation; where the olefin affects the heteroatom is operative. To answer this, the following experiment was designed. Competitive quaternizations of N-methyfamines ^ and 4 were carried out One can expect the unsaturated amine to quaternize faster than the saturated amine since the incipient ammonium ion could be stabilized as in 7^ Experimental results, in fact, bear this out. The unsaturated amine quaternized approximately seventeen per cent faster than the satur­ ated amine (the methodology of this is presented in the experimental section). Through space participation of heteroatoms thus can . lend electron density and borrow it when necessary. 158 To cast some light on the quaternizations of the 8-aza-bicyclo[4.3.0]non-3-ene series, a discussion of the stereochemistry of quaternization is now presented. There has been considerable activity in studies relating to qua­ ternization of tertiary amines, particularly in attempting to understand the factors which control stereoselectivity.9 Most of the contributions, to date, have been concerned with the piperidine system,10 and the little attention fo­ cused on the pyrrolidine system has not considered 3,4- disubstituted pyrrolidines. 159 In analyzing the steric course of quaternization of 2-methylpiperdines 7 McKenna concluded that N-alkylation occurred preferentially cis to the 2-methyl group.11 Anal­ ysis of the chemical shift data demonstrated that the N- alkyl groups cis to the 2-alkyl group in quaternary salts of the 2-alkylpiperdine series were, found at higher field.12 A recent examination of the quaternization of 2- pheny!pyrrolidines has resulted in the conclusion that al­ kylation with small alkylating groups will generally occur cis to the phenyl group with larger groups alkylating trans.13 The nmr data are consistent With data in the other 2-alkylpiperidine series.12 As part of the continuing study of the effects of heteroatoms in a five-membered ring on conformation and reactivity,1 ' 3'4 quaternizations in the cis-8-azabicyclo- ■ [4.3.0]non-3-ene series was analyzed. These compounds are prepared according to the reactions delineated in Scheme I. The available tetrahydrophthalimide (£) was reduced to £ with lithium aluminum hydride. Methylation of ^ was accom­ plished in good yield to give 3.■ The N-benzyl amine 12 was prepared from I^ L by first converting the anhydride to the imide 11 which was in turn reduced to 12 with lithiumrV ■ aluminum hydride. • 160 Scheme I Synthesis of N-Methyl and N-Benzyl Tertiary Amines Quaternizations were carried out by mixing the ter­ tiary amine with the alkyl halide in ether. The salts from these reactions were then subjected to the scrutiny of nmr spectroscopy. Methylation of ^ gave a quaternary salt having dis­ tinguishable methyl signals at 63.52 and 3.60. Substitu­ tion of deuteriomethyl iodide as the alkylating agent re­ moved the upfield methyl signal by 93%, demonstrating con­ siderable stereoselectivity in the alkylation. However, the reasons for this high stereoselectivity were not ob­ vious, since Drieding models did not appear to suggest any special factors which might be controlling the direction of addition. 161 In the piperidine series, higher field nmr signals have been assigned to axial substituents; however, it is sometimes more difficult to consider axial or equatorial positions in the five-ring series. These structural prob­ lems were alleviated with the completion of an X-ray structure analysis of the methyl quaternary salt of I7E . This structure, completed as part of a separate problem,4 clearly demonstrated the nonidentity of the two methyl groups. A model, prepared by adjusting bond lengths and bond angles to fit those determined from the X-ray struc­ ture, is represented by I7E. From this structure, it is evident that we can con­ sider an axial and an equatorial methyl group.14 Continu­ ing this train of thought, the methyl signal at 63.53 can be tentatively assigned as the axial methyl, which re­ quires that methylation of ^ must occur primarily as an axial alkylation. This is consistent with observations in M 162 the 2-alky!pyrrolidine series where preferential alkyla­ tion cis to the 2-alkyl group can most readily be appreci­ ated as axial alkylation. If one suggests treating quaternization of pyrroli­ dines to be similar to the piperidine series, several ad­ ditional tests are possible. As a first test, benzylation has been demonstrated to be less stereoselective than methylation, and an increasing percentage of equatorial alkylation is observed.15 Treating 3 with benzyl halide gave two products, 3^ 3 and ]^ 4 (Scheme II). . In the quater- nizations of ^ with benzyl chloride, two distinct methyl­ ene signals from the benzyl groups could be observed in the nmr spectrum of the isomer mixture at 65.03 and 5.13 (assigned to 3^ 3 and 3^ 4, respectively). However, no sepa­ rate, distinguishable methyl signals could be seen. The iodide salts mixture, on the other hand, exhibited dis­ tinct methyl signals at 63.32 and 3.27 and methylene sig­ nals at 64.95 and 5.10 (for 3^3 and 3^ 4, respectively). 163 Scheme II Benzylation Studies 3 + PhCH2Cl 84 : 16 + ^ ^ 3 - C H aPk 5* 3 CH1PI1 5* 4 CH, ^ + PhCh2I 86 : 14 These data again show consistency with the assump­ tion that the pyrrolidine salts can be analyzed similarly to piperidine salts in that (a) higher field signals cor­ respond to axial substituents and (b) an increased percent­ age of equatorial alkylation is observed with the benzyl halides. These data are also consistent with the studies reported for benzylation of the 2-phenylpyrrolidines.1 3 It has been demonstrated that a quaternary salt can be thermally equilibrated.16 Thus, a piperidine quaternary salt possessing an axial N-benzyl group can be isomerized to a mixture enriched in the isomer having an equatorial N-benzyl group. Attempting this isomerization on the chloride salts of 1^ 3 and I^ (84:16 mixture) resulted in considerable decomposition; however, the resulting mixture was 62:38 of I^ and 14. Whether this was an artifact due 164 to preferred decomposition of could riot be determined. A similar study on the iodide salts resulted in much less decomposition, and the 60% recovered salts were shown to consist of a 27:73 mixture of 1^ 3 and I^. These results can only reflect isomerization. A similar analysis with pure 14 resulted in no isomer change. These results, too, support the assignments of structure for 1^ 3 and 14, and lend additional credibility to treating pyrrolidines with methods successfully used for piperidines. A last test is possible. The direction of alkyla- tiori of tertiary piperidines has been suggested by exam­ ining the stereoselectivity of "inverse alkylations."15 Results from our studies are presented in Table 2. In general, if the.methylation is more selective than another alkylation, the major product from the methylation reac­ tion is presumably derived from axial alkylation. This analysis is consistent with our studies, since the steri- cally smaller methyl group should give more axial attack. These combined studies allow us to consider that alkylation of tertiary pyrrolidines can be studied using the methodologies.which have proven so effective for the piperidines. Further, it has not escaped our attention that these results can be applied to the tfopane series ^5. 165 When considered as a substituted piperidine l£a, N- alkylation has to be considered as primarily equatorial,17 / % in contrast to the wealth of data demonstrating that pi­ peridines undergo preferential axial alkylation. This anomaly can be eliminated, however, by considering tropane as a 2-substituted pyrrolidine I^b. In this series we can expect preferential axial alkylation. Table 2 Inverse Alkylation Studies PhCH I 2 I 86:14 % % CH3I 12 'X, 8:92 ^ % Consideration of a model of tropane reveals that the steric interactions on the pyrrolidine ring side of 166 the molecule are considerably less than on the piperidine side. Since quaternizations are affected by subtle steric interactions, it is not surprising that the pyrrolidine ring dominates the tropane chemistry.18 Many methods have been used to probe the mobile in­ version equilibria at the nitrogen atom in piperidine and its N-alkyl derivatives. Recently, a reliable method for studying conformational equilibria in piperidines has been Table 3 Physical Properties of Quaternary Salts Amine Alkyl Halide Yield % 3 C H 3I 78 3 CD3I 71 3 . PhCH2Cl 90 3 PhCH2I 74 12 . CH 3I 73 Mp (crude isomer 0C mixtures) 215- 216 . 216- 218 , 175-177 130-145 200 (pure 9) 143-148 (50:50, 8:9) devised.20 The above discussion, in which quaternizations of pyrrolidines were shown to be similar in piperidines, implies we can generalize a bit further and extend this. 167 method of obtaining conformational equilibria from six to five member rings. Basically the method involves kinetically controlled protonation of tertiary amines in which three criteria must be met: (a) protonation must be much faster than inver­ sion of nitrogen so that the Curtin-Hammett principle does not apply, (b) protonation must be stereospecific with re­ tention of configuration and (c) protonation must be ir­ reversible. These requirements are satisfied by adding a large excess of sulfuric acid to a dilute amine solution, shaking and separation of the layers. The ratio of enanti­ omeric ions can then be analyzed by magnetic resonance and conformational free energies obtained. N CH3 2 Ha. S O ^ Hevane 1:1 The protonation of the N-methylamine ^ was observed by John Theodore as being nonstereospecific and resulting in a 50:50 mixture of product ions. This is in direct conflict with the aforementioned quaternizations in which 168 ^ demonstrates preferential axial alkylation. Obviously, at this point we see that our generalization breaks down and we must consider each reaction as somehow being differ­ ent from each other. Explanations of differing results from quaternizations and protonations are not germaine to this thesis and shall not be discussed here. The addition of extra steric restraints can con­ ceivably make protonation stereospecific. I believed this goal achievable by protonation of the sterically congested N-methylamine 1,9. Synthesis of this molecule is deline­ ated in Scheme III. Scheme III Synthesis of Amine 19 169 Examination of Drieding models shows that the equatorial methyl group sits directly over the middle of a benzene ring and therefore in the middle of the paramagnetic cone induced by the aromatic portion of the molecule. The axial conformer, on the other hand, sits di­ rectly over the edge of the aromatic ring and therefore in the diamagnetic region of the benzene cone. The N-methyl group is found diamagnetically shifted 101 Hz with respect to the N-methyl group in implying exists in an axial conformation. Following the same protonation scheme as before, a dilute solution of 19 was mixed with concentrated sulfuric acid. Immediately a red-brown color appeared and in less than a minute a black tar had formed making NMR measure­ ments impossible. Whether competitive sulfonation of the 170 aromatic ring or an interaction between ammonium ion and aromatic ring occurs is as yet unknown. This amine, 1^ 9, in the presence of MeI was unreactive as was the case where it was reacted at reflux with "Magic Methyl" (Figure 3). No changes of the steric restraints (they presently seem too prohibitive) were tried in an attempt to increase the stereospecificity of alkylation and protonation of pyrrol­ idines. Though generalizations can be made, we must once again reckon with the fact that each of these systems are in their own small ways different from each other and should not be misconstrued as belonging to a mold merely because some observations are the same. Figure 3: Attempted Alkylation of N-methyl Amine 3^9 171 By virtue of alkylating the N-methyl amine ^ , all available electrons (source of Lewis Basicisity) are used. Moving diagonally across the periodic table, exchange of nitrogen by sulfur followed by methylation produces a very similar salt, 2^ 0 as depicted in Figure 4. This methylated sulfur compound unlike the analogous amine ^ , has an extra pair of electrons. Though one might expect reduced Lewis Basicisity due to the positive charge, it is conceivable that these lone pair electrons still make the sulfonium ion a suitable base. (I might interject at this juncture that no literature precedence for shift studies on sulfon­ ium salts is available.) Indeed, a CDCl3 solution of 20 in the presence of Eu(fod) 3 shifts proton absorbances in the expected direction. The methyl sulfonium protons were Figure 4: Preparation of Sulfonium Salt (2^ 0) . 172 xie'2 Proton A I IlltlM XlB'2 MBB 1.0000 2 0000 3.0000 5OT0 I 5000 2.5000 Xie'2 Proton B Proton C Proton D Figure 5: Lanthanide Induced Shifts of a Sulfonium Salt 173 0000 I 0006 2 0000 3.0000 5000 I 5000 2 5000 Proton E .BBBB .5000 I .0000 I 5000 .2500 7500 1.2500 XlB'2 Proton G eeee I Beee 2 0000 3. .5000 I 5000 2 5000 XlB'2 Proton 7 # Figure 5 (cont.) 174 shifted approximately 140 Hz and the methylene protons on the five member ring were shifted approximately 80 Hz be­ fore any serious line broadening was observed. Figure 5 presents a line of standard A S vs [L]/[S] plots with proton assignments tentatively made. Whether the anion is actually acting as a Lewis base rather than the sulfur lone pair electrons has not escaped my attention. The lanthanide could be complexing with the halide, which, if exists as a tight ion pair (good chance since the solvent medium is CHCl3) would bring the shift reagent within operative range of the ring protons. A pictorial representation of this is presented in Figure 6. Figure 6: Possible Complexation with Sulfonium Salt Counterion 175 Although one might not give this line of reasoning much value at first glance because (i) the proton induced shifts are very large implying the Europium must be quite close to the heteroatom and (ii) in solution one expects the anion to be almost randomly moving about (much like a dog on a long leash) making the shift reagent an isotropic species, the following has been observed. The dimethyl ammonium ion, Ir^, which has no place for the Europium to complex demonstrates diamagnetic shifts under the same conditions! Experimental results show the dimethyl protons shift much faster (this molecule generally shifts much slower than the sulfonium ion) than the other protons implying that the shift reagent is somehow anisotropic near the nitrogen. Whether this is in fact due to complexing with the counter­ ion is as yet unknown but the possibility of a new method for evaluating counterion positions in solution can be foreseen, particularly with the aid of computer assisted lanthanide induced shift programs.21 In sulfonium ions, of course, there is a possibility that both the sulfur lone pair and counterion participate in alignment of the shift reagent. This effect may be thought of as a type of neighboring group participation which forces me to ask: if neighboring group participation 176 is operative in shift reagent studies, how much so? In an attempt to answer this and some unusual europium-oxygen psuedocontact distances observed by Donna Bennett in l£ and various derivatives if l£ the following study was de­ vised . 22 If neighboring group participation is available (for example, the olefin assists during quaternization of N-methyl-8-azabicyclo[4.3.0]non-3-ene), one would expect a larger binding constant for the molecule which can assist than for one which can not. To test this, a molecule, in a known conformation which places the heteroatom near the olefin (e.g., endo ether 2^ .) was compared to a similar exo ether which has been demonstrated to exist primarily in the open conformation. 3 The endo ether can donate elec­ tron density from the olefin portion to the electron de­ ficient ether moiety whereas the exo ether can not and binding constants should differ as demonstrated in Figure 7. Binding constants were obtained using the second method described by Armitage23 with Eu(fod)3 the shift reagent. Calculated binding constants were so large that no distinc­ tion could be made between the two ethers. Replacing the electronegative fluorines by hydrogen on the europium lig­ ands should decrease the binding constants. This is borne 177 Figure 7: Evaluation of Anchimeric Assistance During Shift Reagent Complexation out by the same analysis as above using Eu(thd)3 as a shift reagent. Much smaller induced shifts were, in fact, ob­ served. Using this method, however, proved unsatisfactory since the chemical shifts were so small that much error was introduced in calculating binding constants and the results proved to be statistically useless. On one hand the shifts are big enough to give statistically useful plots but binding constants are too high to be evaluated and on the other, the induced proton shifts are so small that the appropriate plots are statistically useless . . . robbing Peter to pay Paul! No doubt this problem can be 178 solved using a shift reagent which is fluorinated some­ where between these two extremes. Conformational analysis and reactivities of type I nonaromatic heterocycles are by no means completed. Work in these and other laboratories is continuing at a furious pace. EXPERIMENTAL A solution of cis-1,2,3,6-tetrahydrophthalimide (30.2g, 0.2 mole) in dry THE was added dropwise to a mix­ ture of lithium aluminum hydride (15.6g, 0.4 mole) and dry THE (500ml). The resulting reaction mixture was refluxed , for 7 hr. Work-up and distillation gave 10.Og (41%).of the colorless cis-8-azabicyclo[4.3.0]hon-3-ene, bp 77-80° (10mm) (lit.24 bp 79-81° ClSmm)). Synthesis of 8-methyl-cis-8-azabicyclo[4.3.0]non-3-ene (3) The amine (I.23g, 0.01 mole) from the previous re­ action was treated with 90% formic acid (2.6g) and 37% for­ maldehyde. The reaction mixture was heated at 90-100°. Evolution of carbon dioxide was observed, and heat was re­ moved until this ceased, after which time the reaction mix­ ture was heated for an additional 8 hr. After this time, the product was removed by extraction with 5ml of 4N hy­ drochloric acid, and evaporation of water, yielded the crude hydrochloride salt. The free base was liberated by adding aqueous base to the salt and extracting with benzene. Dis­ tillation gave 0.9g (66%) of the colorless N-methyI amine (3), bp 68-71° (10mm) (lit.25 bp 62-88° (10mm)). Synthesis of cis-8-azabicyclo [4 „ 3.0] non-3-.ene (9) I I 180 In a 20OOml, 3 neck flask equipped with mechanical stirrer, reflux condenser and addition funnel, which had been flushed with dry nitrogen and connected to a mercury bubbler was placed 15.6g (0.4 mole) lithium aluminum hy­ dride in 500ml anhydrous THF. 30.6g (0.2 mole) cis- 1,2,3,4-tetrahydrophthalimide in 200ml dry THF was added dropwise and let reflux seven hours. A saturated solution of Rochelle salts was added, liquid decanted and salts washed with several portions diethyl ether. Concentration yielded 6.Og (24%) colorless amine which distilled at 54-80°C, 7mm Hg. IR: no carbonyls; 3250cm ^ (N-H stretch). 3.Og (0.024 mole) amine was reacted with 6.25g 90% formic acid and 4.8ml 37% formaldehyde according to the procedure described in the synthesis of . Workup was car­ ried out using 12ml 4N HCl, 12ml H 2O and 6ml 18N NaOH. Ex­ traction and distillation yielded 2.3g (69%) of the methyl amine. The amine distills at 59-62°C with 6mm Hg pressure. IR: N-H stretch is absent. MS: calculated and found 139. Synthesis of 8-benzyl-cis-8-azabicyclo[4.3.0]non-3-ene (12) cis-1,2,3,6-tetrahydrophthalic anhydride (20g) was Synthesis of 8-methyl-cis-8-azabicyclo[4.3.0]nonane (4) mixed with benzyl amine (15ml) in 40ml of xylene, 181 resulting in an instantaneous exothermic reaction. This reaction mixture was heated for an additional 2 hr, after which time it was cooled. Product (16.Sg) crystallized from the reaction mixture (m/e 241), and this crude pro­ duct was treated, without further purification, with 5.Sg of lithium aluminum hydride in dry THF (175ml). The re­ duction mixture was refluxed for 18 hr, hydrolyzed with water and, after cooling, the tetrahydrofuran was decanted. The salts were washed with ether and the combined organic layers were concentrated. Distillation yielded 11.Sg (79%) of the N-benzyl amine (1^2,) , bp 140-160° (4mm, Hg) , characterized as the methiodide salt, mp 200°. Anal, calcd for C16H22IN; C , 54.09; H , 6.25. Found: C t 54.35; H, 6.38. Preparation of quaternary amine salts The salts were prepared by adding the alkyl halide to an ether solution of the amine. Reactions using methyl iodide and benzyl chloride required refluxing for 5 hr. The crude salt mixtures precipitated from the. ether as they were formed. These salts were filtered, washed with ether, ■ and dried. The benzyl chloride salts were hygroscopic and required drying in a vacuum desssicator. Where necessary. 182 recrystallization was from methylene chloride. See Table 3. Isomer ]^ 4 could be separated from ]^ 3 by recrystal­ lization from methylene chloride. Thermal equilibrium of quaternary salt isomers A solution of the mixed iodide salts 3^ 3 and 14 (O.IO6g) in chloroform (IOml) was sealed in a glass tube and heated at 75-80° for 138 hr. Reisolation yielded 60%. recovery16 of the salt mixture, whose nmr spectrum indicated a 27:73 ratio of ]^ 3 and 14. Similar treatment of pure 14 (75° for 120 hr) resulted in a 50% recovery of 1^ 4 with no y . Synthesis of 4-methyl-h,k,-dibenzotricyclo[5.2.2.O2' 3j-4- aza-undecane (19) Anthracene, 89.Ig (0.5 mole) and maleic anhydride, 49g (0.5 mole) were dissolved in a mixture of xylene (400ml) and benzene (IOOml). This mixture was refluxed six hours (the initially green mixture turns amber). The dark grey crystalline product was filtered, dissolved in hot acetone, decolorized and recrystallized. The resulting colorless crystals were filtered and air dried to give the 183 Diels-Alder adduct in 76% yield. mp = 266-269°C. MS: Calculated and found 276. IR: 1770cm ^ and 1830cm (anhydride). 47g (0.17 mole) of the anhydride and 30ml 28% NH4OH (22.2ml = 20g - 0.33 mole N H 3) were mixed with a 500ml flask equipped with an air condenser. The condenser was large enough in diameter to allow a glass rod to pass through its entire length . . . this is necessary to keep the condenser from clogging with sublimate during the re­ action. The mixture was heated to fusion (300°C) with a Fisher burner and maintained at a molten state until, the evolution of H2O had ceased (approximately 10 min). The molten brown liquid was then poured into an exaporating dish to cool. A brown solid was obtained which was de­ colorized and recrystallized from 95% ethanol (1000ml) to yield 17g product 3^ 7. IR: 3330cm-1 (N-H stretch); 1770 and 1700cm 1 (5 member imide ringe). MS: calculated and found at 275. As described for the synthesis of 4, the imide 17 was lithium aluminum hydride reduced with no difficulty to the 2° amine 20. mp = 224-225°C. IR: No carbonyl pres- ent; N-H stretch present. NMR: Not soluble enough in any suitable solvent. 184 This amine, , was N-methylated by the Leuckart re­ action described in the synthesis of 4. In this case the reaction is slightly different since the amine is so in­ soluble . Upon heating, the 2° amine eventually dissolves and CO2 was slowly liberated. Further refluxing (24 hr) caused the product to form and precipitate out of solution in quantitative yield. This 3° amine, ^9, was filtered and dissolved in CHClg (the 2° amine was totally insoluble in the solvent) and filtered to remove unreacted starting material. Addition of hexane knocked the product out of I solution which upon filtering and air drying gave a fluffy white product with a mp 225-226°C. A mixed mp of 2° and 3° amines depressed the phase transition by 12°. IR: N-H stretch not observed. MS: could not accurately be ob­ tained due to such strong metastable peaks. NMR: 67.16 (aromatic, 8H); 64.9 (doublet, 2 H , J=VH^); 64.16 (broad multiplet,. 2 H); 64.0 (broad singlet, 3 H); 63.15-2.3 (broad multiplet, 4 H). Elemental analysis: calculated C ; 87.36 N ; 5.36 H ; 7.28, found C; 81.62 H ; 6.43 N; 0.10. 185 ID as prepared by B. P . Mundy was condensed with MeBr filtered, and air dried. Synthesis of endo-4-oxatricyclo[5.2.I.02'6]dec-8-ene (21) See reference 29. Competitive quaternization of 3 and 4 Unsaturated amine ^ (0.2859g, 2.087mmole) and 0.2901g (2.08Vmmole) saturated amine 4 were added to form a stock solution in acetone. Glc analysis on a basic col­ umn (described in experimental section of Chapter 2) op­ erating at 170°C, 181 lbs He, cleanly separated both com­ pounds. By method of co-injection the short retention peak was found to be the unsaturated amine. All numbers are an average of three injections and peak areas were determined by triangulation. In that 4.ISmmole amine was in acetone, a limiting amount of methyl iodide, 0.284g (2.OOmmole) was used. The stock solution was divided into three equal portions as was the methyl iodide. The alkyl halide was added neat to the amine mixture via IOOyl syringe in which it was care­ fully weighed. The ammonium salts which formed were let Synthesis of 8-methyl-cis-8-thiabicyclo[4.3.0]non-3-ene bromide (20) 186 Vpc run# Compe- . tition run# Area 3 Area 4 3+4 % 3 I 4 Ave .4 Ave 3 I I 26.65 42.90 65.55 38.32 61.68 2 I 26.00 41.07 67.07 38.77 61.23 39.24 60.76 3 I 26.76 39.10 65.86 40.63 59.37 I II 18.13 20.60 47.73 37.99 62.01 2 II 18.48 29.76 48.24 30.31 61.69 38.58 61.42 3 II 18.23 28.00 46.23 39.43 60.57 I III 20.76 33.00 53.76 38.62 61.38 2 III 17.95 29.76 47.71 37.62 62.38 38.80 61.20 3 III 20.72 30.88 51.60 40.15 59.84 settle out and aliquots of the remaining solution were analyzed. The average of the 3 averages are: % ^ = 38.87 % ^ = 61.13. This represents a change of 8.70% in each peak for a total of 17.4%. Determination of lanthanide shift reagent binding constants Procedure adapted from Armitage23 of constant (Lq ), varying (Sq ). Preparation of shift reagents 0.5ml of 0.17633mole/liter Eu(fod)3 in CCl4-II TMS (I.8294g/10ml CCl4) was diluted to IOml with CC14/1% TMS to obtain a Eu(fod) 3 solution (8.8165mmole/liter). — 50.0681g (9.71x10 moles) Eu(thd)3 was dissolved in IOml CCl4 to obtain a Eu(thd) 3 solution. The shift reagent 187 was not sufficiently soluble to be used so 0.0634g -5(9.044 xlO mole) was dissolved in CDCl3 and volume made up to IOml. Final concentration was 9.044mmole/liter. Preparation of standard substrate solutions 0.816g endo ether 21 (MW = 136) was diluted to IOml in a volumetric flask with CCl4 to form a 0.6M solution. 0.743g cis ether I1C (MW = 124) was diluted to IOml with CCl^/1% TMS to produce a 0.6M solution. 0.0815g endo ether 21 was diluted to 10ml with CDCl3/l% TMS. Final concentration was 0.6M. All shift studies were carried out at 60 MH2,. The following data has been reduced via the inhouse program BOJAC. 26 21:Eu (Thd)3 Plot nd —1 I / ^ Proton Slope Sigma . Intercept ■ Sigma Weight A .0.547 0.099 -0.0514 -0.0406 606.1 B 1.144 0.382 -0.0756 -0.0824 147.2 C 1.159 0.261 -0.0257 -0.0454 485.8 D 3.235 0.621 -0.0648 -0.456 480.7 E 1.141 0.203 -0.0925 -0.0470 453.1 188 21:Eu(fod)_ Plot S vs I/. d O Zi Proton Slope Sigma Intercept Sigma Weight A 1.50 0.086 -0.0051 -0.0065 23.7 Xio3 B 1.97 0.279 -0.0282 -0.0228 I.92x103 C 3.98 0.529 -0.0223 -0.0203 2.43 xio 3 D 8.28 -.595 -0.0102 -0.012 9.61x10 3 E 2.29 0.263 -0.0053 -0.0158 4.01x10 3 1C:Eu (Thd) 3 Plot VS V a A 1.38 0.115 0.0148 0.0109 8.42 xio 3 B 1.57 0.163 0.0263 0.0125 6.40x10 3 C 4.42 0.627 0.0143 0.0186 2.89x10 3 D 4.54 0.400 0.0155' 0.0115 7.56x10 3 E 0.78 0.097 0.0063 0.0172 3.38x10 3 1C:Eu(fod) „ Plot S 3 O VS V a A 2.15 0.195 0.0023 0.0167 3.59 xio 3 B 1.97 0.279 0.0282 0.0228 1.92x10 3 C 7.89 0.516 0.0006 0.0128 6.10 xio 3 D 9.07 ■ 0.705 -0.0021 -0.0162 3.81 xio 3 E 0.91 0.278 0.0212 0.0474 0.45 xio 3 Mean values were calculated as presented in Statistics in Vhysical Science.'11 x = XiWixi where x = mean, wLN = weight for ith observation Z . S . x = observation. i i a is the standard deviation for the new mean x(!/ZiWi) 189 w . = 1/a? where a. is the standard deviation for the i^ *1 i l l observation Mean Compound LSR Intercept CT , ro Range 21 Eu(fod)3 -0.0058 0.0059 (liter/ mole) 714 158 to +” 21 Eu(thd)3 -0.0588 0.0215 18.4 13.2 to 30.5 IC Eu(fod)3 0.0041 0.0079 CO — IC Eu(fod) 3 0.0129 0.0056 OO — Kg was calculated from the plot of oD —1 1/^ K = I/ - (L + intercept) where K is the binding constant B O . B L is the initial LSR concen- ° tration LITERATURE CITED -LI LITERATURE CITED la. B.P. Mundy, A.R. DeBernardis and R .D . Otzenberger, J. Org. Chem., 36, 3830 (1971). b. P. Wilder, Jr. and C.V.A. Drinnan, J. Org. Chem., 39, 414 (1974). c . R.F. Gratz and P . Wilder, Jr., Chem Comm., 1499 (1970). 2a. R.D. Bach and R.F. Richter, J. Org. Chem., 3S_, 3442 (1973). b. and P.A. (1972). Scherr, J. Am. Chem. Soc. , 9jl, 220 c . and R.F. Richter, Tet. Lett, 4099 (1973). 3. B.P. Mundy, K.R. Sun and R.D. Otzenberger, J. Org. Chem.., 37, 677 (1972) . 4. G.D. Smith, R.D. Otzenberger, B.P. Mundy and C.N. Caughlan, J. Org. Chem., 39_, 321 (1974) . 5. J.A. Pople and D .L . Beveridge, Approximate Moleaular Orbital Theory, McGraw-Hill, 1970. 6. All programs used to reflect, rotate and make nuclear co-ordinates were written by Dr. G. David Smith, MSU, 1972. 7. R.C. Bingham, M.J.S. Dewar and D.H. Lo,' J. Am. Chem. Soo.,, 97 (6), 1285 (1975) . 8. R.D. Otzenberger, Ph.D. Thesis, MSU, 1970, Part I. 9a. J . McKenna in Conformational Analysis - Scope and Present. Limitations, G. Chiurdogiu, Ed., Academic Press, New York, NY, 1971, pp. 165-176. b. ________ , Top. Stereoohem., 5, 275 (1970). IOa. R.A.Y. Jones, A.R. Katritzky and P.G. Mente, J. Chem. Soc. , B, 1210 (1970) . 192 IOb. H .O . House, B.A. Tefertiller and C.G. Pitt, J. Org. Chem. , 31, 1073 (1966) . c . T.M. Bare, N.D. Hershey, H.O. House and C.G. Swain, J. Ovg. Ohem., 31_, 997 (1972) . 11. J.K. Becconsall, R.A.Y. Jones and J . McKenna, J. Chem. Soa. , 1726 (1965). 12. J. McKenna, J. McKenna, A. Tulley and J . White, J. Chem. Soa. , 1711 (1965) . 13. A. Solladie-Cavallo and G . Solladie, Tetrahedron Lett., 4237 (1972). 14. It has been suggested by a referee that one might bet­ ter consider these as exo and endo methyl groups, and that further consideration of the preferred course of alkylation might better be discussed in these terms. 15. D .R . Brown, R. Lygo, J . McKenna, J.M. McKenna, and B.G. Hutley, J. Chem. Soo., B , 1184 (1967). 16. See Reference 15. Decomposition is noted to accompany isomerization, so that it is essential in these studies to demonstrate the per cent of recovered product so that there can be assurance that a rel­ ative increase of one isomer is not, in fact, a ■ preferential decomposition of the other isomer. 17a. For a recent review of some problems in the tropane series, see G. Fodor, D . Frehel, M.J. Cooper and N . Mandavar in Conformational Analysis - Sooipe and Present Limitations, G. Chirudogiu,.Ed., Academic Press, New York, NY, 1971, pp. 73-92. b. A recent X-ray analysis demonstrates the propensity of tropanes toward equatorial alkylation. See V.O. de la Camp, A.T. Bottini, C.C. Thut, J. Gal . and A.G. Bellettini, J. Org. Chem., 37_, 324 (1972) and leading references. 193 18. This suggestion is compatible with the observed domi­ nating effect of the cyclopentane ring in control­ ling the course of reduction of bicyclo[3.2.1]- octane-8-one . 9 ) 19. A. C . Cope, J.M. Grisar and P.E. Peterson, J. Am. Chem. Soo. , 82^ , 4299 (1960). 20. P .J . Crowley, M.J.T. Robinson and M.G. Ward, Chem. Corm., 82 5, (1974). 21. For example PDIGM; R .E . Davis, U . Texas (Austin) and M .R . Willcott III, U . Houston. 22. Donna J . Bennett, Senior Thesis, MSU, 1975. 23. I. Armitage, G. Dunsmore, L .D . Hall and A.G. Marshall, Can. J. Chem., July (1972). 24. K . Murayama, S . Morimuna, Y . Nakamura and G. Sunagawa, Yakugaku Zasshi, 8_5, 130 (196 5) CA, 62;16173f. 25. R . A . Schmidt, Ar oh. Bioohem. Biophys., 8_3, 233 (1959). 26. Written by Dr. R.A. Howald, MSU. 27. C . W . Hamilton, Statistics in Physical Science, Ronald, 1964, pp. 40-43. 28. R . P . Feynman, R . B . Leighton and M. Sands, The Feynman Lectures on Physics, Addison-Wesley Publishing Co., 1963, Chapter 2. 29. B . P . Mundy , O.P.P.I. , in Press (1975). 194 XT BoTHEBi M t THAT I MAY NoT Bt r O( O) t E f l' . a f ' l 5 ) l OS r WHAT BoTHEBi M t EVtN MORE I i THAT MAYBE X AM . © 1976 Oy NEA inc T M Reg U S Pu 01' 8-20 CO—4TI 3 1762 D 3 7 8 L 6 6 k L i p k o w i t z , K e n n e t h B c o p . 2 T h e o r e t i c a l a n d e x ­ p e r i m e n t a l n o n a r o m a t i c h e t e r o c y c l i c c h e m i s t r y D A T E I S S U E D T O c ,2' / - / i JAN 3 I i. < I « r A S ' I/ “-*■ ) v . J U i o , / / / / OEC t J