PART I A STUDY OF SOME IOTA LKOXYCARBONIU M IONS PART SYNTHESES FROM 5 6 - HEXAMETHYL - 2, '5 - ‘CYCLOH'EXADTENONE 2, 3, 4, 4, Thesis for the Degree of Ph. D. MTCHTGAN STATE UNIVERSTTY MONICA VERMA 19.70 f. 3 ”me-AAwrfybfi = 2. b $7 4%! W .2 a {HT-5” 0-169 LIBRARY Michigan Sn University . t .‘ This is to certify that the thesis entitled Part I. A Study of some Dialkoxycarbonium Ions. Part II. Syntheses from 2,3,4,4,5,6-Hexamethyl- 2,5-cyclohexadienone. presented by Monica Verma has been accepted towards fulfillment of the requirements for Ph.D Chemistry degree in :HW "LAW Major professor 7 Date "larch 4, 19 0 ABSTRACT PART I A STUDY OF SOME DIALKOXYCARBONIUM IONS PART II SYNTHESES FROM 2,3,4,4.5.6-HEXAMETHYL-2,5-CYCLOHEXADIENONE BY Monica Verma The purpose of this investigation was to determine the chemical shifts for the aromatic protons of 2-phenyl- 1,3-dioxolenium cation, and to correlate them with the charge densities at the corresponding positions on the phenyl ring. For this purpose we prepared 2-(phenyl-3,5— d2)-1,3—dioxolenium cation and 2-(phenyl-4-d)—1,3-dioxolen— ium cation and the chemical shifts of the ortho, meta, and para protons were uniquely assigned. The v-electron densi- ties were calculated by the "dwtechnique" which is a varia— tion of the Hfickel molecular orbital method. The calculated charge densities were plotted versus the proton chemical shifts and a qualitative correlation between charge densities and proton chemical shifts was observed. In addition to this nmr-charge density study, five new dioxolenium salts were synthesized and characterized. They are the monocationic 2—mesitoyl—1,3-dioxolenium tetrafluoro— borate and the four dicationic salts: 1 Monica Verma 2.2'-(gfphenylene)bis-1.3-dioxolenium tetrafluoroborate. 2,2'- (trans-1.4-cyclohexane)bis-1,3—dioxolenium tetrafluoroborate. 2.2'-(1,8-naphthalene)bis—1,3—dioxolenium tetrafluoroborate and 2.2'-(1.4-naphthalene)bis-1.3-dioxolenium tetrafluoro- borate. The nmr spectra of these new cations were compared with those of known dioxolenium cations. For the second part of His work several new small ring and bicyclic compounds were synthesized from 2,3.4.4,5.6- hexamethyl-Z,5-cyclohexadienone (£2)and their chemistry was investigated. The first was the spiro hydrocarbon 4.5.6.6, 7,8-hexamethylspiro[2.5]octa-4.7-diene (QZ). Dienone fig had already been converted to 1-methylene-2,3.4,4.5.6-hexa— methyl-2.5-cyclohexadiene (g2) in good yield.1 Addition of dibromocarbene to Q2 gave 2,2-dibromo-4.5,6,6.7,8-hexa- methylspiro[2.5]octa-4.7-diene (1;) which, by reduction with lithium and liquid ammonia. gave g1; 2,2-Dichloro-4.5.6,6, 7.8-hexamethylspiro[2.5]octa—4.7-diene (21) was also prepared by the addition of dichlorocarbene to 22; Compound §Z was found to rearrange to hexamethylbenzene in solution, both thermally and with acid. 0 I CH3MgBr CH3Br /NH t-BuO rw f3\l‘9\0 Monica Verma The chemistry of the precursor Z; was also investigated. It was found to polymerize on photolysis. Treatment of 1; with methyllithium did not yield an allene. When 11 was heated in solution. it gave three isomeric products, a— bromo-pentamethylstyrene (Q2). gigffi-bromo-pentamethylstyrene (§1) and trans—B-bromo—pentamethylstyrene (§g). H H Br Br L\/’ Br L\//}I L\//Ii Br ' H Br H 71 80 81 82 m NV In deuterated solvents, very little deuterium was incorpor— ated into the products. A plausible free radical mechanism is proposed for this reaction. The rate of rearrangement of Zl'was accelerated in the presence of a radical source (benzoyl peroxide) and was found to be proportional to the concentration of benzoyl peroxide. Dienone gé'was oxidized with mfchloroperbenzoic acid and two new compounds, 2.3-epoxy-2.3.4,4.5.6-hexamethyl-2.5- cyclohexadienone (22) and 2.3;5.6—diepoxy-2,3.4.4.5,6-hexa— methyl-2,5-cyclohexadienone (2Q) were obtained. 0 C1 0 0 I In H co3 _________’ O O O 35 94 95 Monica Verma Their photochemistry was investigated and 22 was found to rearrange to 6-acetyl-2,3.4.4.5-pentamethyl-2-cyclopentenone (102) on photolysis whereas 22 remained unchanged. 0 O 0 H n ‘hv \iIIII!!6f\\t 0 ———————» g; 102 We also prepared l-methylene-3-isopropyl-2,4.5.6.6— pentamethyl-Z.4-cyclohexadiene (107) from dienone 22 and isopropylmagnesium bromide. Triene 107 formed an adduct (Q2) with dimethyl acetylenedicarboxylate. o. // )—MgBr eozC-E-COZMe ____. —-————+ COZMe \ / \ COzMe $3.52. 10 64 W W When dienone g2 was reduced with lithium aluminum hydride three products were detected: 1-methylene-2.4.5.6,6- pentamethyl-Z.4-cyclohexadiene (lgg). 1,3,4.5.5,6-hexamethyl- 1,3-cyclohexadiene (122) and 1.2.3,3.4,5-hexamethyl-1.4-cyclo— hexadiene (110). Monica Verma O H H H H H mum. ————a + + \ H 35 108 109 110 “N W M W Triene 108 formed an adduct (111) with maleic anhydride which on hydrolysis gave the diacid 112. // o A H ‘50 COzH H 14" H Pb(0Ac), 11\1§; H H 1.2.9:. COzH 112 Diacid 112 was found to rearrange to a y-lactone thought to be 113. by oxidation with lead tetraacetate. REFERENCES 1. J. D. DeVrieze. Ph.D. Thesis. Michigan State University. E. Lansing. Michigan 1968. PART I A STUDY OF SOME DIALKOXYCARBONIUM IONS PART II SYNTHESES FROM 2,3.4.4.5.6-HEXAMETHYL-2.5-CYCLOHEXADIENONE BY Monica Verma A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1970 G- 61/440 /0'Q3i3'1i9 PLEASE NOTE: Some pages have indistinct print. Filmed as received. UNIVERSITY MICROFILMS. To Harish ii ACKNOWLEDGMENT The author wishes to express her sincere appreciation to Professor Harold Hart for his guidance and encouragement throughout the course of this investigation. and for his help in perfecting this thesis. Appreciation is also extended to the National Science Foundation and National Institutes of Health for providing financial support. iii TABLE OF CONTENTS PART I A STUDY OF SOME DIALKOXYCARBONIUM IONS INTRODUCTION . . . . . . . . . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . . . . . . A. B. The NMR Spectrum of and(flwmge-Distribution in the 2-Phenyl-1,3-dioxolenium Cation . Preparation of some New Dioxolenium Cations EXPERIMENTAL . . . . . . . . . . . . . . . . . . . A. General . . . . . . . . . . . . . . . . . 1. Microanalyses . . . . . . . . . . . . 2. Nuclear Magnetic Resonance Spectra . 3. Infrared Spectra . . . . . . . . . . 4. Solvents . . . . . . . . . . . . . . 5. Miscellaneous . . . . . . . . . . . . Preparation of 2-(Phenyl-3.5-d2)-1,2- dioxolenium Cation . . . . . . . . . . 1. proluidine—2.6-d2 (Z) . . . . . . . 2. Toluene-3.5-d2 (Q) ‘. . . . . . . . . 3. Benzoic-3.5-d2 Acid (9) . . . . . . . 4. 2-Hydroxyethyl Benzoate-3.5-d2 . . . 5. 2-(Phenyl-3.5-d2)-1.3-dioxolenium Cation. Preparation of 2-(Phenyl-4-d)-1.3-dioxo- lenium Cation . ... . . . . . . . . . . . 1. Toluene-4-d (12) . . . . . . . . . . 2. Benzoic-4-d Acid (12) . . . . . . . . 3. Z-Hydroxyethyl Benzoate—4—d . . . . . 4. 2-(Pheny1-4-d)—1.3-dioxolenium Cation iv 20 33 33 33 33 33 33 33 34 34 34 35 36 36 36 36 37 37 37 TABLE OF CONTENTS (Cont.) Page D. Preparation of 2-Mesitoyl-1.3-dioxolenium Tetrafluoroborate (14) . . . . . . . . . . 37 1. Mesitoyl Chloride . . . . . . . . . . . 37 2. Attempted Pre aration of 2-Hydroxyethyl Mesitoate (22' . . . . . . . . . . . . 38 3. 2-Bromoethyl Mesitoate (17) . . . . . . 38 4. 2-Mesitoyl-1.3-dioxolenium Tetra- fluoroborate (£4) . . . . . . . . . . . 39 5. 2-Hydroxyethyl Mesitoate (32) . . . . . 39 E. Preparation of 2,2'-(97Phenylene)bis-1,3- dioxolenium Tetrafluoroborate (24) . . . . 40 1. Phthaloyl Chloride .. . . ... . . . . . 40 2. 2-Bromoethyl Phthalate . . . . . . . . 40 3. 2.2'-(9:Phenylene)bis-1,3-dioxolenium Tetrafluoroborate (24) . . . . . . . . 40 F. Preparation of 2.2'—(trans-1,4-Cyclohexane)bis -1.3-dioxolenium Tetrafluoroborate (31) . . 41 1. trans-1,4-Cyclohexanedicarboxylic Acid Chloride . . . . . . . . . . . . . . . 41 2. 2-Bromoethyl-1,4—cyclohexane Dicarboxy- late . . . . . . . . . . . . . . . . . 41 3. 2.2'—(trans—1,4-Cyclohexane)bis-1.3- dioxolenium Tetrafluoroborate (31) . . 42 4. 2-Hydroxyethyl trans—1.4-Cyclohexane- dicarboxylate . . . . . . . . . . . . . 42 G. Preparation of 2.2'—(1.8-Naphthalene)bis-1,3- dioxolenium Tetrafluoroborate (33) . . . . 43 1. 1,8-Naphthoyl Chloride . . . . . . . . 43 2. 2—Bromoethyl 1,8-Dinaphthoate . . . . . 43 3. 2.2'-(1,8-Naphthalene)bis-1.3-dioxo- lenium Tetrafluoroborate (32) . . . . . 44 4. 2-Hydroxyethyl 1.8-Dinaphthoate . . . . 44 5. Attempted Preparation of 32 Using 'Magic' Acid . . . . . .A. . . . . . . 44 TABLE OF CONTENTS (Cont.) Page H. Preparation of 2.2'-(1.4-Naphtha1ene)bis—1,3— dioxolenium Tetrafluoroborate (32) . . . 45 1. Z-Bromoethyl 1.4-Dinaphthoate . . . . 45 2. 2,2'-(1.4-Naphtha1ene)bis-1,3-dioxo- lenium Tetrafluoroborate (32) . . . . 45 NMR SPECTRA . . . . . . . . . . . . . ... . . . . 47 SUMMARY . . . . . . . . . . . . . . . . . . . . . 53 PART II SYNTHESES FROM 2.3,4.4,5.6-HEXAMETHYL-2,5-CYCLOHEXADIENONE INTRODUCTION . . . . . . . . . . . . . . . . . . . 55 RESULTS AND DISCUSSION . . . . . . . . . . . . . . 65 A. Preparation of 4.5.6.6.7.8-Hexamethylspiro- [2.5]octa-4.7-diene . . . . . . . . . . . 65 B. Preparation and Photolysis of 2.3-Epoxy- 2.3.4.4.5,6-hexamethyl-2,5-cyclohexa- dienone and 2,3;5.6-Diepoxy-2.3.4,4.5.6- hexamethyl-Z.5-cyclohexadienone . . . . . 87 C. Attempted Preparation of 2—Methylene 1,3. 3.4,5-Pentamethylbicyclo[2.2.2]octa-5.7- diene . . . . . . . . . . . . . . . . . . 96 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . 105 1. Microanalyses . . . . . . . . . . . . 105 2. Nuclear Magnetic Resonance Spectra . 105 3. Infrared Spectra . . . . . . . . . . 105 4. Ultraviolet Spectra . . . . . . . . . 105 5. Mass Spectra . . . . . . . . . . . . 105 6. Melting Points . . . . . . . . . . . 105 7. Preparation of 37 from 1-Methylene- 2. 3. 4. 4. 5, 6- -hexamethyl- -2, 5-cyclohexa- diene (39) . . . . . . . . . . . . . 105 8. Preparation of 2.2-dibromo-4,5.6.6.7,8- hexamethylspiro[Z.5]octa-4.7-diene (13) 106 vi TABLE OF CONTENTS (Cont.) 10. 11. 12. 13. 14. 15. 16.) 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. Photolysis of Z}, . . . . . . . . . . Reaction of Zl'with Methyllithium . . Pyrolysis of Zl'in Solution . . . . . Preparation of.32,. . . . . . . . . . Preparation of 1—Dideuteromethylene- 2. 3. 4. 4. 5, 6-hexamethyl-2. 5—cyclo- hexadiene (84) . . . . . . . . . . . Effect of Benzoyl Peroxide on the Rate of Decomposition of 2, 2-Dibromo-4. 5, 6. 6, 7 .8-hexamethylspiro[2. 5]octa-4. 7- diene (71) . . . . . . . . . . . . . Page 107 108 108 109 110 111 Preparation of 2. 2-Dichloro-4. 5. 6, 6. 7. 8- hexamethylspiro[2. 5]octa-4, 7-diene (91) 111 Reduction of 71 to 4, 5. 6. 6. 7, 8+Hexa—.. methylspiro[2~5]octa-4, 7- diene (37) . Photolysis of 37 . . . . . . . . . . Acetone- -sensitized Photolysis of 31'. Mercury-sensitized Photolysis of 31 . Pyrolysis of 3Z’in Solution . . . . . Acid-catalyzed Rearrangement of 31 . Low Temperature Nmr Study of 37 . . . Attempted Preparation of 94 Using Sodium HydroXide and Hydrogen Peroxide Attempted Preparation of 32'Using Sodium Tungstate and Hydrogen Peroxide Preparation of 94 Using m—Chloroper- benzoic Acid . . . . . . . . . . . . Photolysis of 2.3-Epoxy-2.3.4.4.5.6- hexamethyl—Z.5-Cyclohexadienone (32). Preparation of 6-Acetyl-d3-methylvd3- 2. .4, 4. 5- -tetramethyl—Z-cyclopentenone 103) I O O I I O O O O O O O O O 0 Preparation and Photolysis of 3,5- Dimethyl-d6-2.3—epoxy-2.4.4,6—tetra- methyl-2,5-Cyclohexadienone (31) . . Attempted Hydrogenation of 64Acetyl- 2.3.4.4.5-pentamethyl-2-cyclopentenone (102) O O O O O O O O O O I O O C O 0 vii 112 112 113 113 113 114 114 115 115 116 117 118 118 118 TABLE OF CONTENTS (Cont.) Page 30. Attempted Photolysis of 2,3;5,6—Diepoxy- 2,3,4,4,5,6-hexamethyl-2,5-cyclohexa- dieneone (92) . . . . . . . . . . . . . 119 31. PreparatiOn of 1-Methylene—3-isopropyl- 2,4,5,6,6,-pentamethyl—2,4-cyclohexadiene (107) . . . . . . . . . . . . . . . . . 119 32. Preparation of Adduct 64 of 107 with Di- methyl Acetylenedicarboxylate . . . . . 120 33. Preparation of 1-Methylene-2,4,5,6,6- pentamethyl-Z,4-cyclohexadiene (108). . 121 34. Preparation of Adduct 111 of 108 with Maleic Anhydride . . ._. . .‘. . . . . 122 35. Attempted Oxidation of Adduct 11 . . . 122 36. Preparation of Diacid 112 . . ... . . . 123 37. Oxidation of Diacid 112 . . . . . . . . 123 38. Reduction of y—Lactone 113 . . . . . . 124 39. Attempted Preparation of Adduct 117 of with 2-Butyne . . . . . . . . . . . . 124 SPECTRA . . . . . . . . . . . . . . . . . 126 SUMMARY . . . . . . . . . . . . . . . . . . 140 LITERATURE CITED . . . . . . . . . . . . . . . . . . 142 viii TABLE II. III. IV. LIST OF TABLES Proton nmr chemical shifts of some phenyl carbonium inns . . . . . . . . . . . . . . Calculated charge densities and proton chemical shifts of some ions . . . . . . . Proton nmr chemical shifts for 2-aryl-1,3— dioxolenium cations . . . . . . . . . . . . Proton nmr chemical shifts of some dioxo- lenium dications . . . . . . . . . . . . . Results of pyrolysis of 2,2-dibromo—4,5,6, 7,8-hexamethylspiro[2.5]octa-4,7-diene (21) in deuterated solvents, at 150° . . . ... . ix Page 11 16 23 28 74 FIGURE 1. 10. 11. 12. 13. 14. LIST OF FIGURES Calculated charge densities versus chemical shifts of some ions . . . . . . . . . . . . . NMR spectrum (CCl4) of 2-bromoethyl mesitoate NMR spectrum (CH3CN) of 2-mesitoyl-1,3-dioxo- lenium tetrafluoroborate (£4) . . . . . . . . NMR spectrum (CCl4) of 2—bromoethyl phthalate NMR spectrum (FSO3H) of 2,2'-(97phenylene)bis- 1,3-dioxolenium tetrafluoroborate (24) . . . NMR spectrum (CC14) of 2-bromoethyl 1,4-cyclo- hexanedicarboxylate . . . . . . . . . . . . . NMR spectrum (CF3C02H) of 2,2'-(trans-1,4- cyclohexane)bis-1,3-dioxolenium tetrafluoro- borate . . . . . . . . . . . . . . . . . . . NMR spectrum (CCl4) of 2-bromoethyl 1,8-di— naphthoate . . . . . . . . . . . . . . . . NMR spectrum (CH3CN) of 2,2'-(1,8-naphthalene)- bis-1,3-dioxolenium tetrafluoroborate (33) . NMR spectrum (CH3CN) of 2,2'-(1,4-naphthalene)- bis-1,3-dioxolenium tetrafluoroborate (34) . Effect of benzoyl peroxide on the pyrolysis of 2,2-dibromo-4,5,6,6,7,8—hexamethylspiro[2.5]- octa-4,7-diene (11) . . . . . . . . . . . . . Kinetics of the thermal rearrangement of 11 in benzene, at 110° . . . . . . . . . . . . IR spectrum of 2,2-dibromo-4,5,6,6,7,8-hexa- methylspirolZ.5]octa—4,7-diene (11) . . . . IR spectrum of 2,2—dichloro-4,5,6,6 7,8—hexa- methylspiro[2.5]octa-4,7-diene (91).. . . . . X Page 18 48 48 49 50 50 51 80 81 127 127 LIST OF FIGURES (Cont.) FIGURE 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. IR spectrum of 4,5,6,6,7,8-hexamethylspiro- [2.5]octa-4,7-diene (91) . . . . . . . . . IR spectrum of 2,3—epoxy-2,3,4,4,5,6—hexa— methyl-2,5-cyclohexadienone (94) . . . . IR spectrum of 2,3;5,6-diepoxy-2,3,4,4,5,6- hexamethyl-Z,5—Cyclohexadienone (99). . . . IR spectrum of 6-acetyl—2,3,4,4,5,-penta- methyl—Z-cyclopentenone (102 . . . . . . . IR spectrum of 1-methylene-3-isopropyl- 2,4,5,6,6—pentamethyl-2,4-cyclohexadiene (127) . . . . . . . . . . . . . . . . IR spectrum of adduct 94' . . . . . . . IR spectrum of 1, 3, 4, 5, 5,6-hexamethyl- 1, 3- -cyclohexadiene (109) . . . . . . . IR spectrum of 1, 2, 3, 3, 4, 5-hexamethyl-1, 4- cyclohexadiene (110). . . . . . . . . IR spectrum of adduct 111 . . . . . . IR spectrum of y-lactone 113. . . . . NMR spectrum of 2,2-dibromo—4,5,6,6,7, 8- hexamethylspiro[2.5]ocat-4,7—diene (7) NMR spectrum of 2,2—dichloro-4,5,6, 6, 7,8- hexamethylspiro[2.5]octa-4,7—diene (91) . . rw NMR spectrum of 4,5,6,6,7,8-hexamethylspiro- [2.5]octa-4,7-diene (91) . . . . . . . . . NMR spectrum of 2,3—epoxy-2,3,4,4,5,6-hexa- methyl-2,5-cyclohexadienone (94) . . . . NMR Spectrum of 2,3;5,6-diepoxy-2,3,4,4,5,6- hexamethyl—Z,5—cyclohexadienone (99) . . . NMR spectrum of 6-acetyl-2,3,4,4,57penta- methyl-2—cyclopentenone (102) . . . . . . . xi Page 128 128 129 129 130 130 131 131 132 132 133 133 134 134 135 135 LIST OF FIGURES (Cont.) FIGURE 31. 32. 33. 34. 35. 36. Page NMR spectrum of 1~methylene—3~isopropyl-2,4,5, 6,6-pentamethyl-2,4-cyclohexadiene (107). . . 136 NMR spectrum of adduct 94'. . . . . . . . . . 136 NMR spectrum of 1,2,3,3,4,5-hexamethyl-1,4- cyclohexadiene (110) . . . . . . . . . . . . 137 NMR spectrum of 1,3,4,5,5,6-hexamethyl—1,3— cyclohexadiene (109) .. . . . . . . . . . . . 138 NMR spectrum of adduct 111 . . . . . . . . . 138 NMR spectrum of y—lactone 11 . . . . . . . . 139 M xii PART I A STUDY OF SOME DIALKOXYCARBONIUM IONS INTRODUCTION As early as 1939 Tipsonl, observed that in general halogenoacetal sugars, on treatment with silver acetate in glacial acetic acid, dry toluene or a similar solvent, yield a sugar acetate having t£§g§_acetate groups at C1 and C2 regardless of whether the original sugar itself had - gi§_or ££§g§_hydroxyl groups at C2 and C3. Isbell2 later .observed that this result could only be obtained if the C2 acetate group was Egggg to the halogen, as it can partici- pate in the displacement only when it is in the Egagg con- figuration. However, it was not until 1942 when Winstein3 postulated dioxolenium cationic intermediates, that these results were explained with precision. Winstein observed that trans-2-acetoxy-cyclohexyl bromide (l), on treatment with silver acetate, gave exclusively a Eraggfdiacetate (9). He explained these results by postulating a 2-substituted- 1,3-dioxolenium cation (9) as an intermediate for this reaction. Me Me F- ‘1 / ’z \ Me Q:C\ O O \ 4O ' _ / .____>. C,’ OAC -————> /'+ J ° 033 "‘ d O r I B l. 52. 8 2. 2 O 3 By now 1,3-dioxolenium cations have been well charac- terized. They occur as intermediates in a variety of reac- tions and have also been synthesized and isolated. The preparation of these ions can be grouped into three general reactions. 1. Displacement of a B-substituent by Participation of a Carboxinroup. O - O u _ r , R‘C'OCH2CH2X _....X_.> R" <+ j nucleoph1le 4’ Products \ O When X is a halogen the reaction is brought about by silver salts, organometallic reagents or a Lewis acid. When X is an ether, alcohol or ester function, the reaction is usually acid catalyzed, with a protic or Lewis acid. If X is a labile group such as a tosylate or a brosylate, solvol- ysis yields the 1,3-dioxolenium cationic intermediate. Re- actions of this type have been reviewed by Pacsu4 and Lemieux.5 The preparation of the dioxolenium cations for this work also falls under this category. 2. Acid-catalyzed Hydrolysis of Cyclic Ortho Esters. R0 0 O 2 II o X :l > R—<+ ] nucleoph11e> Products R o o 4 The reaction is brought about when Z is a protic or a Lewis acid.“‘10 3. Oxidation of aygyclic Acetal. H o ,0 >x< ‘::l x R"<&' nucleophIle> Products R c) O In this case X is C12, Bra, BrCCla or NBS.7:11‘15 The preparation and isolation of 1,3-dioxolenium cations by the reaction of 2-haloethyl esters with silver tetra- fluoroborate was devised by Meerwein and coworkers.8 Sub- sequent work showed that Lewis acids such as boron trifluoride and antimony pentachloride can also be used.9:11 Four dis- tinct families, outlined below, of these cyclic dialkoxy- carbonium ions (1,3-dioxolenium cations) have been made by Hart and Tomalia”:17 by Meerwein's method. 1. 2—Alkyl-1,3-dioxolenium cations: 5 2. 2,2'-Alkylene-bis-1,3-dioxolenium dications: O O \ ’ .- ‘: +’)>—(CH2);1—<(\f :‘ 2x 0 O 3. 2-Aryl-1,3-dioxolenium cations: RM?) 4. 2,2' and 2,2',2"-Aryl-1,3-dioxolenium dications and trications: __ W __ /___\ - O [a] 2X- and 3X 0. X . / < +‘». 0 °‘\.\ ’0 0 - fi — # Recently Beringer and coworkers18 used the same procedure as Hart and Tomalia to make 2—(2',6'-dimethoxyphenyl)-1,3- dioxolenium tetrafluoroborate. They have also studied the reactions of this cation with various nucleophiles. Tomalia and Hart17 examined the nmr spectra of the 2-aryl-1,3-dioxolenium cations and found a quantitative cor— relation between the prouxrnmr chemical shifts, of meta and 6 p§£g_substituted 2—phenyl-1,3-dioxolenium cations and Ham- mett 0 values. The chemical shifts of the dioxolenium ring protons of these ions were plotted against Hammett o and 0+ values. A better linear correlation was obtained with 0 than with 0+ values, which suggests that the major contributor to the resonance hybrid is g'rather than 9; RP? ‘WRQ-éjj ~ As 6, the chemical shift, is directly related to the Hammett 0 values, it should also be proportional to log k/ko, where k and k0 are rate constants for the ring opening reaction of substituted and unsubstituted 2-aryl-1,3- dioxolenium cations. This represents a quantitative correla— tion, by magnetic resonance, of charge densities in this carbonium ion system. In this work we have determined the chemical shifts for the aromatic protons of 2-phenyl—1,3—dioxolenium tetra- fluoroborate by preparing deuterium labeled tetrafluoro- borates. Both 2—(phenyl—3,5-d2)-1,3-dioxolenium tetra- fluoroborate and 2-(phenyl-4-d)-1,3—dioxolenium tetrafluoro— borate were prepared for this purpose. We have calculated the charge densities at various positions on the phenyl ring of 2-phenyl-1,3-dioxolenium ion and have correlated these with the chemical shifts. We have also determined the 7 chemical shifts for the benzoylium (or phenyloxocarbonium) ion by using EEEE and meta_deuterated benzoic acid. In addition to this nmr-charge density study, five new dioxolenium salts were synthesized and characterized. They are the monocationic 2-mesitoyl-1,3-dioxolenium tetrafluoro- borate and the four dicationic salts 2,2'-(g-phenylene)bis- 1,3-dioxolenium tetrafluoroborate, 2,2'-(1,8-naphthalene)- bis-1,3-dioxolenium tetrafluoroborate, 2,2'-(1,4-naphthalene)- bis-1,3-dioxolenium tetrafluoroborate and 2,2'-(E£ag§g1,4- cyclohexane)bis-1,3-dioxolenium tetrafluoroborate. The reasons for synthesizing each of these particular dioxolenium compounds will be presented in the Results and Discussion Section. RESULTS AND DISCUSS ION A. The NMR Spectrum of and Charge Distribution in the 2-Phenyl—1,3—dioxolenium Cation. We wanted to determine the proton chemical shifts for the aromatic protons of the 2-phenyl-1,3-dioxolenium cation, and correlate them with the charge densities at the corre- sponding positions on the phenyl ring. For this purpose we decided to prepare 2—(phenyl-3,5—d2)-1,3-dioxolenium cation and 2-(phenyl-4-d)-1,3-dioxolenium cation, so that the chemical shifts of the ortho, meta and BEES protons could be uniquely assigned. The first step in the synthesis of these ions was the preparation of benzoic-3,5—d2 acid and benzoic-4-d acid. The protons 9££h2_to the amino group of pftoluidine (9) were easily replaced by deuterium when 9 was heated with a 50:50 mixture of phosphoric-d3 acid and deuterium oxide, yielding pftoluidine-2,6-d2 (2). Compound Z was reduced to toluene— 3,5-d2 (9) with sodium nitrite and hypophosphorous acid. Benzoic—3,5-d2 acid (9) was obtained by the oxidation of 9’ with potassium permanganate. The nmr spectrum of 9'con— sisted of two singlets at T 2.54 (1H) and 1.96 (2H). The 8 9 areas allow these shifts to be assigned to the para and ortho protons respectively. CH3 H3 CH3 COZH D3PO4/ NaNO [::::]\D KMnO4 3P0 2 D//\ NH2 NHZ Q. 2, 8 9 To prepare benzoic—4-d acid (19) we started with 27 bromotoluene (19) which was converted to the corresponding Grignard reagent (11). This was decomposed by deuterium oxide to yield toluene-4—d (19) which on oxidation with potassium permanganate gave 19; Compound 19 was character- ized by its nmr spectrum which consisted of two doublets at T 2.62 (2H, g,- 8.25 Hz) and at 1.94 (2H, g_= 8.25 Hz), which can be assigned to the meta and ortho protons respectively. I CH3 THS // |JS_>/|22L>/‘m>/‘ \\ \\ \\ \\ l l l | Br MgBr D D 12, $1 12 13 The assignments of chemical shifts for benzoic acid are therefore as shown: 10 COZH 1.95 2.62 2.54 The deuterated benzoic acids were converted to the acid chlorides by reaction with thionyl chloride. The cor— responding 2-hydroxyethyl esters were prepared by refluxing a solution of the acid chlorides, in methylene chloride, with ethylene glycol until the evolution of hydrogen chlor— ide had ceased. The 2-hydroxyethyl esters of benzoic acid,“ on addition of fluorosulfonic acid, form the 2-phenyl-1,3- dioxolenium cation instantly. l“\ o /,OCH3CH30H Oxfxo C03H COC l §C OH Br 0 I - _ FSO SOClz > CH3 CH3 FSO3H 3 > The exact chemical shifts for the orthp and pa£a_protons on the phenyl ring were obtained from the nmr spectrum of 2- (phenyl-3,5-d3)-1,3-dioxolenium cation and the chemical shifts for the m§£§_and 9££h9_protons were obtained from the nmr spectrum of 2-(phenyl-4-d)-1,3—dixolenium cation. The chemical shifts are given in Table I. They were found to be almost identical to those of protonated benzoic acid, as reported by Farnum.19 11 Table I. Proton nmr chemical shifts of some phenyl carbonium ions. Chemical Shifts, T values Compound ortho meta para f-\ a 1.72 2.24 1.97 0450 / U1 b HQ~+,OH f 1.72 2.24 1.94 c 6% +C 1.44 2.01 1.13 aObserved. bD. G. Farnum, J. Am. Chem. Soc., g2, 2970 (1967). CFormed in fuming sulfuric acid containing 20-25% 503. It has been reported by Traficante20 that benzoic acid, when dissolved in 95-100% sulfuric acid forms protonated benzoic acid but when it is dissolved in sulfuric acid con- taining 10-35% sulfur trioxide it forms the corresponding acylium ion. The chemical shifts for the aromatic protons 12 of this acylium ion have not yet been reported. Since we had benzoic-3,5-d2 acid (2) and benzoic-4-d acid (12) on hand, we decided to determine the exact chemical shifts of the ortho, meta and para protons of this acylium ion. The labeled benzoic acids were dissolved in sulfuric acid con- taining 20-25% sulfur trioxide to form 5% solutions. The nmr Spectra were recorded. The chemical shifts obtained are given in Table I. It was not surprising to find that the chemical shifts for the aromatic protons of the 2-phenyl-1,3-dioxolenium cation were nearly identical to those of protonated benzoic acid. Both of these cations are dioxycarbonium ions. The difference between the two is that the protons on the oxygen in protonated benzoic acid are replaced by an ethylene bridge in the dioxolenium cation. This change apparently has little effect on the aromatic protons. It was our goal to calculate the charge densities at the various positions in the phenyl ring using molecular orbital methods, and to correlate these calculated parameters with the observed proton nmr chemical shifts. Tomalia and Hart17 have already demonstrated a quantitative relationship between the proton nmr chemical shifts of the aliphatic protons and the corresponding Hammett 0 values for meta and page substituted 2-pheny1-1,3-dioxolenium cations. Since Jaffe21 has shown the Hammett 0 values are proportional to the electron charge densities, there should be a correlation between the proton nmr chemical shifts for the aromatic 13 protons of 2-phenyl-1,3-dioxolenium cation and the w electron densities at the carbon atoms to which they are attached. Such correlation has been demonstrated beforefizpzs.24 The F electron densities were calculated by a variation of the Hfickel molecular orbital method known as the "ab technique.", The calculations were carried out with a 3600 Control Data Corporation computer using a program written by Professor R. S. Schwendeman of the Department of Chemistry, Michigan State University. The coulomb integral, which is a function of the nuclear charge of the atom, was taken as a constant value, do, for the carbon atom. The resonance integral is related to the degree of overlap and is constant for a given overlap between two carbon "p" orbitals at a given distance. The resonance integral was taken as a con- stant, 60, for the carbon atom, A change in the atomic number of an atom affects both the coulomb and resonance integrals. The relationship for the new values of these integrals Can be expressed in terms of a0 and 50 as C10 "l” hBO (1) Cl B kfio (2) where h and k are constants that depend upon the hetero- atom which replaces carbon. It was necessary to select ap— prOpriate values of h and k for the oxygen atoms and the bonds to oxygen in the 1,3-dioxolenium ions. There is some question about what are the most suitable values to use. 14 Streitwieser's25 recommended value for h is 2.5, for posi- tively charged oxygen. This value is suggested for use with the HMO method only. The use of the w~technique requires this value to be adjusted for the best fit. The value of h used for this work is 3.7. The value of k was taken as 0.9. This value is intermediate between the value recom— mended for a carbon-oxygen single bond (0.8) and that for a carbon-oxygen double bond (1.0).25 This choice is based on the suggestion of Sandorfy26 that the value of k is a simple function of bond length. The "artechnique" was used here because it aflows the coulomb integral to be adjusted for changes in charge on each atom. Wheland and Mann27 proposed that the value of a should be linearly related to the charge and may be formu- lated as: C! = a0 + (1 ‘ qr)wBo (3) where "w" is a dimensionless parameter whose value may be so chosen as to give best agreement with experiment. The value a)“ 1.4 seems to be well accepted and was used here. The total electron density at an atom is taken as qr and is defined as: 2 qr = 32- nj er (4) .th c is the coefficient of atom r in the 3 MO, which jr is occupied by nj electrons. The sum is taken over all the molecular orbitals. In an all-carbon system each atom 15 contributes a nuclear charge of +1 to the w system; hence the net charge Cr is 2;r=1-qr (5) In the artechnique the calculated w—electron densities are substituted in equation (3) and the molecular orbital calcu- lations repeated. This process is iterative, and is con- tinued until the values for the molecular energy levels and charges on the atoms converge to constant values. The pro- gram does this automatically, it being only necessary to specify the number of iterations. The details for punching' the data cards have been described by Young.28 The results are recorded in Table II. Fraenkel et, 21,22 demonstrated a linear relationship between the proton chemical shifts of cyclopentadienyl anion, benzene and tropylium cation and the net charge on the carbon atom. Benzene was chosen as the reference base, both for the proton chemical shifts, 5, and for the net charge density C. Thus since the v—electron density on the carbon atom in benzene is unity, C = 0. We adopt the further convention that 6 is positive for proton resonances ap— pearing at higher field than the benzene resonance and negative for resonances at lower field. Since benzene was the reference, the chemical shifts of cyclopentadienyl anion and the tropylium cation had to be corrected for ring cur- rent effects. A correction of -0.11 ppm and +0.6 ppm was calculated for cyclopentadienyl anion and the tropylium cat- ion respectively, using Pople's model.22 16 Table II. Calculated charge densities and proton chemical shifts of some ions. Charge Proton Compound Density Chemical Shift (6) Cyclopentadienyl Anion -0.2000 -2.22 Benzene 0 0 Tropylium Cation +0.1428 1.60 2-Phenyl-1,3-dioxolenium Cation 1,2 -0.8516 1 0:1-102 3 0.3912 2 4 0.0180 9 x’ ‘5 5,9 0.0797 0.91 8 \\ 6 6,8 0.0282 0.39 7 7 0.0780 0.69 2,2'-prhenylenebis—1,3-dioxolenium Cation 1,2,9,10 -0.8392 3, 8 0.4238 4, 7 0.0643 5,6,11,12 0.0952 1.35 2,2',2"-Phenylenetris-1,3-dioxolenium cat1°n 1,2,8,9,13,14 -0.8415 3,7,12 0.4200 4,6,11 0.0515 5,10,15 0.2115 2.21 Benzene is taken as the reference (0 = 0) for these chemical shifts. ‘ 17 The chemical shifts of the cyclopentadienyl anion and the tropylium cation should also be corrected for the effect of the charge on the carbon atom to which the hydrogen atom is bonded. This charge effect changes the electron density on the hydrogen atom. Musher29 has corrected the chemical shifts for this charge effect, by calculating the change in nuclear magnetic shielding (A0) using equation 6. A0 = -2.9 x 10‘12 E2 —7.38 x 10‘19 32 (6) Where E is the electric field and E2 is the field along the bond axis. When the corrected chemical shifts were plot- ted versus the net charge, a straight line with a slope of 11.2 ppm/electron was obtained. From this linear correla- tion we obtain: a = ck (7) When the calculated charge densities for the 2-phenyl- 1,3-dioxolenium cation were plotted versus the proton chemi- cal shifts (see Figure 1) they fitted the plot for cyclopenta- dienyl anion, benzene and tropylium cation only moderately well. The deviation can be explained by the fact that be- sides the altered electron density other factors can also contribute to the proton resonance shift. These additional contributions may be due to: a. the magnetic anisotropy of substituents or hetero atoms in the aromatic ring; Charge Densities .Tricat 0.20“ 0.15-4 0.10— 0.05— 18 T l -2 -1 -0.05_. Chemical Shift (0) —0 .10.. -O.15— '0 .20— Figure 1. Calculated charge densities versus chemical shifts. 19 b. ring current effects of neighboring rings in the polycyclic aromatic compounds; c. ion association effects of aromatic ions, and d. solvent effects. The magnitude of the proton resonance shift arising from the magnetic anisotropy of the substituent groups in individual molecules cannot be estimated with sufficient accuracy at present to permit a satisfactory correction with measured shifts. Since this contribution has a 1/R3 de- pendence, it will effect primarily those protons in the molecule that are close to the anisotropic center. Thus the anisotropic effect of the dioxolenium ring would mainly ef- fect the ortho aromatic protons. The ring current effects22 can be dealt with more readily. Since benzene has been chosen as the reference, no correction is required for benzenoid systems. With aromatic positive and negative ions, an associa- tion with counterions in solution may be anticipated. The counterion causes a shift in proton resonance of the aromatic ion. Introduction of bulky groups which would hinder as- sociation with counterions would lower this effect. Solvent effects also contribute to the observed chemical shift.30 By the use of a solvent with a high dielectric constant and dilute solutions counterion and solvent effects can be mini- mized. A 5% solution of 2-phenyl-1,3-dioxolenium cation in fluorosulfonic acid was used to avoid solvent effects. 20 The charge densities of the dication and trication were also calculated and plotted against the proton chemi- cal shifts (Table II and Figure 1). A qualitative correla- tion between charge densities and proton chemical shifts was observed. However, until precise calculations can be done to determine the degree to which other factors contrib- ute to the observed chemical shifts a quantitative correla- tion of charge density with the chemical shift will not be obtained. B. Preparation of Some New Dioxolenium Cations. The preparation of some new aryl dioxolenium cations with ortho substituents was undertaken to study their effect on the geometry of the dioxolenium cations. Others were prepared to compare their nmr spectra with those of known dioxolenium cations. The first ion studied was the 2- mesitoyl-1,3-dioxolenium cation (14) (as the tetrafluoro- borate), because we expected that the two ortho methyl groups might restrict the rotation of the dioxolenium ring and prevent it from becoming coplanar with the aromatic ring. The noncoplanarity of the two rings (structure 12) might encourage the concentration of the charge in the dioxolenium ring and this would be reflected in a lower field shift for the proton resonance of the dioxolenium ring protons. In fact there was some doubt that the dioxolenium cation would form if coplanarity was not possible. 21 CH3 1:: For the preparation of 14” mesitoic acid (15) was con- verted to mesitoyl chloride (12) by treatment with thionyl chloride. An attempt to make 2-hydroxyethyl mesitoate, a possible precursor of the desired cation,ty refluxing £6 with ethylene glycol was unsuccessful; instead, the dimesito- ate of ethylene glycol was obtained. However, 2—bromoethyl mesitoate 11 was readily prepared by refluxing a solution of 16, in carbontetrachloride, with 2-bromoethanol. Cation £4 was made from lz'according to Meerwein's procedure,8 by the addition of anhydrous silver tetrafluoroborate to a vigor— ously stirred solution of 11.1“ methylene chloride. A yellow precipitate, which consisted of silver bromide and the salt of 12, formed immediately. The precipitate was suspended in acetonitrile; silver bromide, being insoluble, was re- moved by filtration. When anhydrous ether was added to the acetonitrile filtrate,the salt of cation 14 precipitated out of solution as a white solid. 22 /—”\ CO H c0c1 0s H Br o‘j,o l 2 | COCHzc 2 V \\ // OH Br \\ I // \ / ' ' / - _ 'A BF _ [::::] soc12> CH2 CH3 9 4 > BF4 15 16 17 14 m m w ‘UU The other dioxolenium cations were made by the same procedure used for the preparation of 14; In all the cases we started with the corresponding aromatic acids. Different solvents were used for the separation of the dioxolenium cations from silver bromide in each case. These are reported in the Experimental section. The nmr spectrum of 14'(Figure 3) consisted of singlets at T 7.70 (3H,‘p§£§ methyl group), 7.48 (6H, pggpp_methyl groups), 4.67 (4H, ring protons), and 2.94 (2H, aromatic protons). Comparison with related ions17 (Table III) shows that the chemical shift of the dioxolenium ring protons of lg'is not affected in any unusual way by the two pgghp_methyl groups. The chemical shift of dioxolenium ring protons is directly proportional to the net charge in the dioxolenium ring, which in turn depends on the nature and the position of the aryl substituent. If the substituent has a positive inductive effect and is substituted in the pagg position where a resonance contribution is also possible, it tends 23 Table III. Proton nmr chemical shifts for 2-aryl-1,3-dr- oxolenium cations. Chemical Shifts, T Values Compound Solvent Ring Aromatic Protons Protons Other Protons a @433 (13) FSO3H 4.58 2.45-1.53 -— 0 a 0 Q93 (133,) FSO3H 4.60 2.57-1.76 CH3- 7.53 Me 0 Me-.—<.:) (2,0,) FSO3H 4.63 2.55 CH3- 7.45 0 1.80(d) a 0. MeO __<;+J (2,1,) FSO3H 4.67 2.73 CH3O- 5.92(s) 0 1.69(d) Me 0 MG 0 {+3 (134‘) CF3C02H 4.67 2.94(S) E-CH3- 7.70(S) O g-CH3- 7048(5) Me OMe b O a] (2323) CF3C03H 4.77 2.27(s) CH3O- 6.055(5) O OMe aD. A. Tomalia and H. Hart, Tetrahedron Letters, 22, 3383 (1966). bF. M. Beringer and S. A. Gatton, J. Org. Chem., 32, 2630 (1967). 24 to reduce the charge in the dioxolenium ring. This would result in a higher field shift of the dioxolenium ring pro- ton resonance. If the substituent is in the £353 position on the phenyl ring, only the inductive effect is significant. This explains the higher field appearance for the ring pro— ton resonance of 22 compared to 12'(Table III). A similar conclusion is reached for ion 21; methoxyl is a better electron donor than methyl, if the resonance effect is sig- nificant. It is noted, however, that the magnitude of these substituent effects is very small. Therefore delocalization of the charge into the aryl ring cannot be an overriding factor in the stabilization of these ions. For this reason, it is not surprising that one can prepare ions such as $4 and 22, with two pgghp substituents. Due to the posi- tive inductive effect of the methyl groups of £4, even the mgpg aromatic protons of this compound (T 2.94) are more shielded than those of 2-pheny1-1,3-dioxolenium cation (18” T 2.24). Cation 14 was hydrolyzed with ice cold water in order to find out if the nucleophile would attadk £4 at C3 or at either C4 or C5 to give the ring opened product, 2-hydroxy- ethyl mesitoate (23). Ester 23 was obtained on hydrolysis. It was characterized by its nmr spectrum (CC14) which con— sisted of a singlet at T 7.82 (9H, methyl groups) and two triplets at 6.64 (2H, g.= 4 Hz) and 5.81 (2H, g_= 4 Hz) for the two nonequivalent methylene groups. The aromatic protons were observed at r 3.33 (s, 2H). The ir Spectrum (CCl4) had 25 1 bands at 1721 (C:O) and at 3580 cm- (OH). The chemical analysis was in good agreement for a molecular formula of C12H1603° 5 4 10 who 3 O. . QC—OCHZCHZOH BF; .322__> l \\ 1:2. 23 The preparation of 2,2'-(pfphenylene)bis-1,3-dioxolenium tetrafluoroborate (24) was undertaken because the correspond- ing pggg and mgpg diCarbonium ions are known17 and we were interested in finding out if 24 could be made, because di- cation gé'has two positively charged moieties on neighboring carbon atoms which might have a destabilizing effect. Some difficulty was anticipated in the formation of 24 because a difference of 106 in the pKR+ of formation of dication fig as compared to the m§£§_and pggg analogs has been reported.31 The reasons given for this observation were that the steric C6H5 + // l 6H5 \ C6H5 + C6H5 26 effects limit the geometries available to 22; also, the proximity of two positive charges in 22 may have a destabil- izing effect. The formation of dication 22 occurs in two steps from tetraphenyl-pfxylene glycol (22). First the monocation 2Z'is formed. The major factor for the difficulty C6H5H c6115 WH C6H5 + 6H5H ‘/)LC H C6H5 -C6H5 \lf C6H5 C6H5H C6H5 C3H5 2.3 21. 22. in forming 22 is attributed to the larger energy difference between the monocation 2Z'and dication 22'as compared to the energy difference between the monocations and dications obtained from the tetraphenyl p§£3_and‘mgpg xylene glycols. This larger energy difference is due to oxygen participation, which is possible only in the pggpp_monocation. There was also some question that the two 23329 substi- tuted dioxolenium rings of 22 may interact with one another and thus restrict the geometries available to 22, In case the geometry of 22'approaches structure 22 we w0uld have two sets of equivalent protons (H and H') on the dioxolenium rings and one would not expect to observe a singlet in the nmr spectrum for the dioxolenium ring proton resonance. 27 H H' HI HI H 23.3. No difficulty was experienced during the preparation of 22; The dioxolenium ring proton resonance was observed as a sharp singlet at T 4.33, indicating that no serious interactions of the dioxolenium rings was present. As mentioned earlier, the proton resonance of the di- oxolenium ring protons depends on the nature and the posi- tion of the aryl substituent. In the case of EEK 22, and 22'(Tab1e IV) let us consider one of the dioxolenium rings to be part of the 2-aryl-1,3-dioxolenium cation structure and the other dioxolenium ring to be the substituent. The substituent in this case has a negative inductive and resonance effect, which increases the charge in the dioxo— lenium ring. For 22'only the inductive effect would be significant whereas in 22,both the inductive and resonance effects can operate. This is reflected by the lower field proton resonance of the dioxolenium ring of 22'compared to that of 22; For 22, the inductive effect would be stronger than in 22'as this effect is inversely proportional to the distance of the substituent from the reaction site. For 28 Table IV. Proton nmr chemical shifts of some dioxolenium dications. Chemical Shifts, T Values Compound Solvent . Ring Protons Other Protons a O i :] (29) FSO3H 4.40 1.97-0.74(m) O y w . 0 [:'%*<::>* I .// 522. 22. Dienone 33'is a white crystalline solid, very stable and easy to handle and because it was readily available to us, it was used as the starting material for the synthesis of some new compounds. The work done with dienone 33 can be divided into three parts. A. Preparation of 4,5,6,6,7,8-hexamethylspiroL2-5]octa- 4,7-diene One purpose of making 4,5,6,6,7,8-hexamethylspimfl2.5]- octa-4,7-diene (31) was that a hydrocarbon with the spiro— [2.5]octa-4,7-diene structure had not yet been isolated out of solution. In fact the only closely related known com— pound to this system is spiro[2.5]octa-4,7-dien-6-one (33).4° 55 56 Other derivatives of this ketone are also known. V V (.00 37 w The main purpose for making 3Z'was to investigate the thermal reaction of this compound. It was expected to exhibit some thermal reactions because a closely related compound, 1- methylene-Z,3,4,4,5,6—hexamethyl-2,5-cyclohexadiene (33) undergoes a thermal rearrangement to ethylpentamethylbenzene (4,9,) .41 39 40 w m The thermal conversion of 33 to 42 was found to pro- ceed by a radical chain mechanism, and is initiated either by the cleavage of 33'to pentamethylbenzyl radicals and methyl radicals (equation 8) or by a radical initiator (E) to produce methyl radicals (equation 9). 3.2 —————-—> + ~CH3 (8) \ it? 57 I (al+R)- + ~cH3 (9) V 3 + n w Similarly 3Z'might cleave on pyrolysis to give methyl radi- cals and radical gl which may rearrange further to give other products such as pentamethyl styrene (23). ' i 22, :1. 22. We also hoped to investigate the photochemistry of 31; Several examples of the photochemical isomerization of 1,4- cyclohexadienes to bicyclo[3.1.0]hex-2-enes have been re- ported.42 For example 3,3,6,6-tetramethyl-1,4-cyclohexa- diene (23) on photolysis in the vapor phase with mercury sensitization gave mainly 4,4,6,6-tetramethy1bicyclo[3.1.0]— hexene (23).43 A di—w-methane mechanism has been proposed ——hu 0 H g v 4:: > A 45 rw 25?. for this reaction. Accordingly 31 was expected to isomerize to structure 43'or 23: on photolysis. Based on the photo- chemical results of other nonconjugated dienes such as 58 norbornadiene44 and 1,5-cyclooctadiene45 compounds gz'and g3'may also be postulated as possible photoproducts. v \ v A .7 /, V 46 4 ' 47 48 NV m w m Compound 3Z'was successfully synthesized. Its prepara- tion and chemistry as well as that of the precursors obtained during the preparation of 31 will be presented in the Results and Discussion section. B. Preparation and Photolysis of 2,3-Epoxy-2,3,4,4,5,6- hexamethyl-2,5-cyclohexadienone and 2,3;5,6-Diepoxy- 2,3,4,4,5,6—hexametnyl-2,5—cyclohexadienone The preparation of 2,3-epoxy-2,3,4,4,5,6-hexamethyl- 2,5-cyclohexadienone and 2,3;5,6-diepoxy-2,3,4,4,5,6-hexa— methyl-2,4-cyclohexadienone was undertaken because they seemed to be interesting subjects for a photochemical investigation. The photochemical behavior of a,B-epoxy- ketones has been studied by various investigators. Even as early as 1918 Bodforss46 had observed the photochemical rearrangement of the a,B—epoxyketone chalcone oxide (13) to dibenzoylphenylmethane (32) on irradiation with ultra— violet light. 59 o o 0 0 n " " can5 hv ;7é%:%<:fi\\\C6Hs ““> CeH5//g\\ar/fl\\C6H5 H C6H5 C6H5 42. 22. The reaction was thought to proceed in two steps; Ca-O bond fission of the oxirane ring followed by 1,2-shift of the B—hydrogen to the a-position. Recently Reusch and coworkers47 and Jeger and coworkers48 found that B-alkyl groups also shift in a similar manner. Both Reusch47 (equation 10) and Zimmerman49 (equation 11) have shown independently that a B-methylene or methyl group migrates preferably over a B-' phenyl group. 0 o C6H5 //O H I (10) ._22__1 can5 it .52. 0 0 0 CH3 II n u 1 ) ( 1 CH3 hv > C635//A\W//~\\ CH3 can5 CH3 2‘2. 22. They also proposed that the photorearrangement of a,p-epoxy- ketones takes place by an n,v* excitation of the carbonyl group which causes the cleavage of the Ca-O bond of the 60 oxirane ring to an intermediate A, followed by rearrange- ment of the B-substituent as shown below. O O O 0* R3 / \ II R / \ \\ hv-n'r* 3\\ ' C"—'C-C'R1 (—"—‘——’. C—C—C-R1 Rz/fi . a . Rz/fi CI. * 1 O O 0* 0* I:___ * __ :I R3\ 1 : R3-C-C-'C-R1 <————> /c—c:c-R1 *R2 R2 6 0. 2 A O O —->R3 —C-C—C-R1< R2 * = ° or i From this mechanism it appears that the migratory aptitude of the B-substituent depends on the stability of the cor- responding radical intermediate. The migratory aptitude of several B-substituents has been studied by Markos and Reusch5° and they found: benzhydryl and benzyl > hydrogen > methylene > methyl >> phenyl. Jeger and coworkers‘”:51 have also studied the photo- chemistry of several steriodal epoxyketones, among which there are examples of Az-unsaturated epoxyketones, three of which are presented below; OAc E hv H -—-> II 0’ 0 £52. 2.9 .521. 61 R 0: W H K -—-——**” R=H R - CH3 R = CH3 22. 22. Jeger et. al. found that saturated epoxyketones rearranged on irradiation with light of wavelengths greater than 3100 3, therefore probably rearrange by an n,w* excitation of the carbonyl group whereas Azqunsaturated epoxyketones were found to be stable to n,v* excitation and rearranged only by exposure to 2537 A light. The photoreaction in this case was brought about by a w,v* excitation of the carbonyl group. The following mechanism was proposed by the authors for the photoreaction of Az-unsaturated epoxyketones. IN, N 1. * R \ hV;W.v* -————a -————9 ¢ * *0 O O 1 xj §_____ <34; fl *0 1 0 O 62 For our work we prepared 2,3-epoxy-2,3,4,4,5,6-hexa- methyl—2,5-cyclohexadienone which is an a,fi—unsaturated epoxyketone and 2,3;5,6—diepoxy-2,3,4,4,5,6—hexamethyl-2,5- cyclohexadienone from dienone 33; The preparation and the photochemical results will be presented in the Results and Discussion section. C. Attempted Preparation of 2-Methylene-1,3,3,4,5-p§nta- methylbicyclo[2.2.2]octa-5,7-diene. Lastly an attempt was made to synthesue 2—methylene— 1,3,3,4,5-pentamethylbicyclo[2.2.2]octa—5,7-diene from dienone 33'because this compound has three double bonds so oriented that any two of them can participate in a di-w— methane rearrangement on photoexcitation. The chemistry of related ketones, the bicyclo[2.2.2]octadienones, has been studied by Murray.52 Benzobicyclo[2.2.2]octadienone (33) on irradiation in ether gave 1,2,3,4-tetramethylnaphtha- lene (31) by the photoelimination of dimethyl ketene and benzobiCyclo[3.2.0]heptene (33) by the photoelimination of carbon monoxide. However, acetone-sensitized irradiation of 33'gave ketone 33'in addition to 33 and 33; Ketone 33’ was shown to be formed by a 1,2-acyl migration from C1 to C6. 63 On the basis of these results 2-methylene—1,3,3,4,5-penta- methylbicyclo[2.2.2]octa-5,7-diene (33) was expected to I eliminate dimethylallene or form a semibulvalene product corresponding to ketone 33” on photoexcitation. we had successfully synthesized compound 33 which contains the 2-methylene[2.2.2]octadiene structure, but this molecule has various substituents and a complicated nmr spectrum. Since the photoproducts are identified by their nmr spectra we decided to make a simpler molecule which would have a less complicated nmr spectrum. The preparation of bicyclo- octadiene 33'was undertaken because a possible route for its preparation from dienone 33'could be formulated. // 02Me 64 C02Me 64 While this work was being written, Cristol and Mayo53 pub— lished their work on the photoisomerization of 2-methylene- dibenzobicyclo[2.2.2]octadiene (33) to dibenzotricyclo- [4.2.1.01:3]nonadiene (33). This reaction is also thought to take place yla a di-w-methane rearrangement. This is the first example of a photochemical rearrangement of a 2-methylenebicyclo[2.2.2]octadiene system. hv a O acetone I “r Attempts to make 33'were unsuccessful but an interesting rearrangement was observed in the process. Details for the preparation of 32'and the attempts to make 33 will be dis- cussed in the next section. RESULTS AND DISCUSSION A. Preparation of 4L5,6,6,7,8—Hexamethylspir012.510cta-4,7- diene The synthesis of 4,5,6,6,7,8-hexamethylspiro[2.5]octa- 4,7-diene (31) was undertaken because we were interested in investigating its thermal and photochemical behavior. Since 1-methylene-2,3,4,4,5,6-hexamethyl-2,5-cyclohexadiene (33)' had already been obtained in good yields from dienone 33f1, we hoped to obtain 3Z'from 33'in a one-step synthesis by the Simmons-Smith reaction.’ Methylene iodide together with the zinc-copper couple made by LeGoff's54 procedure was used for the reaction. When methylene bromide was used instead of methylene iodide the reaction did not always take place. 0 n V CH MgBr _ CH I IéyO "::]I!!'[::Zn(Cu) “I'll 9.2. $22 :31. The reaction mixture was analyzed by vpc using a 5 ft SE—30 column at 150°. The vpc trace showed three peaks with re- tention times of 3, 5 and 14 min for methylene iodide, 65 66 starting material 33'and 3Z'respectively. From the areas of the peaks a 61.5% conversion of triene 33'to give 3Z'in a yield of 95% was calculated. The nmr spectrum (CC14) of 3Z'consisted of singlets at T 9.28 (4H), 8.90 (6H), 8.74 (6H) and 8.40 (6H) which are in accord with structure 31, In addition to these signals a singlet at T 7.78 was also observed which is probably caused by the presence of hexa- methylbenzene as an impurity. Hexamethylbenzene seems to be forming in small amounts during the course of the reac- tion either from 3Z'or 33, This method was not found suit- able for the preparation of 31 in large amounts, since it was difficult to separate hexamethylbenzene from 31 by VpC. Compound 33 has both an exocyclic and endocyclic double bonds. Hence by subjecting 33'to a Simmons-Smith reaction two possible products (32 and 13) may be obtained. 39 37 70 w I‘w fw However, only 32 was obtained. Attack at the highly sub- stituted endocyc1ic bond is probably sterically hindered A better route was sought for the preparation of 31, A dibromocarbene adduct of 33'was prepared which was ex- pected to give the hydrocarbOn 3Z'by reduction with a suit- able reagent. Dibromocarbene was generated by reacting 67 bromoform with potassium nfbutoxide in the presence of 33' and 2,2-dibromo-4,5,6,6,7,8-hexamethylspiro[2.5]octa-4,7— diene (23) was obtained in 52.7% yield. H ‘\\ CHBra > KOEBu 39 ' 71 W W The uv spectrum (hexane) of Zl'had maxima at 225 nm (6 1470) and 244 nm (e 955). The nmr Spectrum (CCl4, Figure 25) con- sisted of singlets at T 8.86 (3H), 8.78 (3H), 8.38 (6H), 8.20 (6H) and 7.81 (2H). The mass spectrum showed a parent peak at mass 346 (Br = 79) and the chemical analysis was also in accord with structure 23, The mass spectrum will be discussed later, in greater detail in this section. In order to investigate the photochemical behavior of 23, a 2% solution of 23 in pentane was irradiated first through a vycor filter and then through a corex filter, using a 450~watt Hanovia lamp. In both cases Zl'underwent polymerization and a black residue deposited on the walls of the test tube containing the pentane solution. 1,1-Dibromocyclopropanes form allenes in reasonable yields with magnesium or highly dispersed sodium,55 and good yields of allenes are obtained with butyllithium and methyl- lithium.°°'57 An attempt was made to prepare allene 13 from compound Zl'by treatment with methyllithium. 68 Br “" Br ll , ll 0 CH3L1 / I __x l l ’ 2.1. 22. When Zl'was treated with methyllithium at —78° the reaction did not occur because the nmr spectrum of the recovered material showed that starting material was unchanged. The reaction was repeated at 0° and a very complicated nmr spectrum was obtained for the reaction product. The ab- sence of peaks in the vinylic region indicated that allene zg'had not formed. The vpc analysis of the product mixture showed that the volatile components consisted mostly of starting material and another compound which was not present in sufficient amount to be collected. Compound Zl'gave a very interesting mass spectrum. When the mass spectrum was first obtained the parent peak atxmms 346 was not present. The temperature of the ioniza- tion chamber was as usual at 250°. Two sets of peaks in an almost 1:1 ratio were observed at mass 267 and 269 and at 252 and 254. However, when the ionization chamber was cooled to 100° and the mass spectrum retaken peaks at 346, 348 and 350 mass units were observed in a 1:2:1 ratio. This pattern is typical of a molecule containing two bromine atoms since 69 the relative abundance of the heavier isotope of bromine (atomic wt 81) is 98%. Peaks at mass 331, 333, and 335 in the same ratio are attributed to ion Z3'which can be formed by the loss of a methyl group (15 mass units). Once again peaks at mass 267 and 269 and also at 252, 254 were observed in an almost 1:1 ratio. This type of pattern suggests frag- ments containing one atom of bromine. These are thus at- tributed to fragments Z3 or 23] and z3'respectively. Br / "" Br 252. 2,2. 74' 75 «IN: AN The absence of the parent peak in the mass spectrum at higher temperatures indicated that compound Zl'was probably undergoing thermal rearrangements in the ioniZation chamber. In fact Weyerstahl and coworkers58 have observed that 1,1- dihalocyclopropanes undergo thermal rearrangements. They found that by treating tetraalkylated ethylenes with tri- halomethane, ethylene oxide and tetramethylammonium bromide at 150-170° the gem-dihalocyclopropane was not obtained but instead resulted in ring opening and hydrogen halide elimina— tion giving a 2-halo-1,3-alkadiene as illustrated below. 70 CH5\5__1FH3 CHx3 ,0: Me4N Br:3 3 3 CH ‘ EH 3 3 CH3 CH3 23: 11 9‘3 9H3 CH2=C - c=c - CH3 + CHz-CHZ x OH x 22. 22. Trialkylethylenes gave 1,14dihalocyclopropanes, 2-halo-1,3- alkadienes, or their mixtures, depending on the reaction temperature. Dialkylated 1,1-dihalocyclopropanes were stable in the presence of ethylene oxide even at elevated tempera- tures, but could be converted to 2-halodienes by pyrolysis or with quinoline catalyst. In order to study the thermal reaction of Zl it was pyrolyzed in solution in decalin, nitrobenzene or toluene. Three major thermal products were detected, in a total yield of 90-95% by vpc. The chemical analysis and the mass spec- trum of these products showed that they were isomeric and had the molecular formula C13H17Br. On the basis of their nmr spectra these products have been assigned structures 3.9.. 3.1. and i2- 82 rw 80 W The spectrum of compound 33 showed singlets at T 7.84 (9H) and 7.74 (6H) which are assigned to the aromatic methyl groups. The vinylic protons were observed at T 4.42 (d, 1H, £_= 1.25 Hz) and 4.10 (d, 1H, 3': 1.25 Hz). The coup-' ling constant of the vinylic protons suggests the presence of two nonequivalent hydrogens located at one end of a double bond, and justifies the assignment of structure 33, To further confirm this assignment compound 33 was synthesized independently by treating pentamethylacetophenone (33) with phosphorus tribromide. Compound 33 was obtained in a 59.5% yield. The nmr spectrum was identical with the one reported above. H B;\\/J 0% o ———» o §.§. 80 m H The nmr spectrum of compound 33'showed singlets at T 7.95 (9H) and 7.90 (6H) and doublets at T 3.62 (1H. 1 = 7-5 HZ) 72 and 2.99 (1H, g_= 7.5 Hz). The coupling constant of the vinylic protons suggests a gla_configuration for the protons and therefore justifies structure 33, The nmr spec- trum of 33'showed a singlet at T 7.89 (15 H) and doublets at T 4.15 (1H, a_= 14 Hz) and 3.11 (1H, 3': 14 Hz). The EEEEE configuration is justified by the coupling constant of the vinylic protons. It is clear that a hydrogen migration is involved in the formation of 33'and 33, In order to investigate the mechanism of this thermal reaction an attempt was made to prepare 2,2-dibromo-3,3-dideutero-4,5,6,6,7,8-hexamethyl- spiro[2.5]octa-4,7-diene (33). For this reason 1—dideutero- methylene-2,3,4,4,5,6-hexamethyl-2,5-cyclohexadiene (33) had to be prepared first, from methyl-d3 iodide and dienone 33, O * *VBr Br I H c031 CHBr3 D20 I KOtBu ii. 3:. 4:». But we found, as was previously reported by DeVrieze,41 that some label is lost during workup and only 72.5% deuterium label was obtained in the methylene group. Exchange is thought to occur yla_the heptamethylbenzenonium cation (ii-11+) - 73 D.\]r/D :13? or “ff-7 84 84-H+ fW When we tried to exchange the methylene protons by treating a solution of unlabeled triene Qg'in carbon tetrachloride with deuterium oxide containing a few drops of sulfuric-d2 acid. The mass spectrum of the recovered material showed that not only the methylene protons but also the protons of the methyl groups at C3 and C5 were exchanged to some extent. For our work we required 100% deuterium label in the methylene group because we wanted to calculate the deuterium content of the products from their mass spectra and the presence of hydrogen atoms in the methylene group would compli— cate the calculations. To overcome this difficulty un- labeled 2; was pyrolyzed in deuterated solvents. The results ‘are summarized in Table V, and a sample calculation for the mole per cent of the singly labeled species is given below. P = mass 346, P+1 = mass 347, P+2 = mass 348, P+3 = mass 349. Peak intensities in reference compound obtained in nondauterated solvents Mass: P P+1 P+2 P+3 Intensity: 1.0 0.244 0.977 0.232 (A) 74 Rob.m Rum.m Rum.m Rm mclmcmncmnouuflz R m.HH R o.w RNo.NH Rm ocimcmsaoa Rvmv.o R o R 0 Ron ovlmcmnsmm RNo.N R o In Rom nvlmcmdaoa Mm mm mm //// m us0>aom ucm>aom Hm // [MMX\// may cfl.flm mo muospoum CH cofiumnomuoocH Esflnmusma mo R sowumuusmosoo coma um .mucw>aom Umumumusmc as dflMv msmflclb.vlmuoo IHm.Naoufimmamgumsmxmnum.h.m.m.m.¢UOEounHvIN.N mo mammaonmm mo muHSmmm .> magma 75 Peak heights (scale divisions) in labeled species Mass: P P+1 P+2 P+3 Intensity: 20 6 19.5 4.5 (B) The entire peak at mass P, 20 div, must be due to the unlabeled species. To compute the contribution of the unlabeled species to P, P+1, P+2 and P+3, the peak height at mass P is multi- plied with the abundance of P, P+1, P+2 and P+3 in standard. 20x1 = 20 20x0.244 4.88 20 x 0.977 = 19.5 (c) 20X0.232 = 4.65 Subtract (C) from (B) 20 6 19.5 4.5 20 4.88 19.5 4.65 0 1.12 0.0 -.15 (0) The peak height due to the singly labeled species is thus 1.12. The sum of corrected intensities is 20 + 1.12 = 21.12. The distribution in mole per cent is 1.12/21.12 x 100 = 5.32 mole perCent of singly labeled speCies. The results of pyrolysis of 11 in deuterated solvents show very little deuterium incorporation into the products, especi- ally in concentrated solutions. Based on these results the following mechanism is proposed tentatively. (See Scheme 1.) It is proposed that compound Z} first loses a methyl radi— cal to give pentadienyl radical gé'which may give the ring— opened aromatic radicals gg'or Qz'by step a_ or b. respectively. 76 Br Br H C~C/\ H C—-C 2 Br 2 \Br ll. §.§. B — I .19. H . I j Br H \ / C \ /C\' Br If’c Br 86 7 T H d l f c e ._ I ' ._ —- fr *1—- Br H3C Br Br H H I C . C/ /H \c/ H/\\C/ \c’ \Br ...}c/C\ \Br H Br 222, 4.9. 42, 90 H Br H \ c C‘ H \:éc1. H 29, £51, or 22,2. 431, or 42, Scheme 1. 77 Intermediate gg'could lose a bromine atom to give 32 (step .Q) or it could pick up an hydrogen atom from the sclvent to give §§‘(step g) which can give §2 by elimination of hydrogen bromide. Since there is very little deuterium incorporation into the product on pyrolysis in deuterated solvents, step g_seems to be favored over step 2. Intermediate §Z on the other hand probably forms §2 by an intramolecular 1,2-hydrogen migration (step 3). Compounds §1'or gz'could be formed from §2'by the elimination of a bromine atom. Intermediate Q1 could also pick up an hydrogen atom from the solvent to give 92 (step 2) which could give §1'or §£'by elimination of hydrogen bromide. Again step 3 seems to be favored over step f_because of low deuterium incorporation in the products. A second possible route for the thermal rearrangement of 11' is shown in Scheme II. In this case it is suggested that the cyclopropane ring of Zl'opens to give two diradi- cals §§fand §ZJ which give radicals §§.and §Z respectively by the loss of a methyl radical. Radicals §§ and QZ'then proceed to the thermal products as in Scheme I. The effect of a radical source on the reaction of 11 in solution was investigated by pyrolyzing ll in the pres- ence of benzoyl peroxide. Duplicate samples of benzene solutions containing 0.20g_Zl'and varying concentrations of benzoyl peroxide, 0.01g, 0.02% and 0.05fl, were placed in nmr tubes. The nmr tubes were degassed to 0.2 mm of Hg, sealed and pyrolyzed at 110°. Two nmr tubes containing 78 71 glorféfi Scheme II. 79 0.20M_solution of 21 in benzene were prepared in the same way and used as the reference solutions. The reaction was followed by nmr. Since the reaction was carried out in an ordinary oil bath the temperature varied by 11°, which probably accounts for some difference between the two runs of each solution. The mean of the two runs was plotted and the results are shown in Figure 11. Benzoyl peroxide accelerates the decomposition of 11 in solution. (The half-life of benzoyl peroxide at 110° is approximately 30 min59.) A plot of ln(% of undecomposed 21) is time is shown in Figure 12. Since the conditions were erratic (see the Experimental Section for the conditions which were used) some of the points did not fit the straight line plots. The first order rate constants for the different runs were calculated by the least squares treatment of the data and are shown in Figure 12. The rate of decomposition of 11 at 110° seems to be increased by the increase in concentra— tion of benzoyl peroxide. The benzoyloxy or the phenyl radicals (2;) obtained by the decomposition of benzoyl per— oxide may abstract a bromine atom from Zl'giving the ring- opened radical Z}: which can form gg'bythe loss of a methyl radical. However, the other two thermal products gg'and gg' cannot be formed by this route. 80 100 80-q U 53 rd 0 m a 60-! 8 R 0 m - '0 c: ” I M - O 40 81 R _ $2 I 20" o 55 » l T l r T l r r 0 30 90 150 210 Time, min Figure 11. Effect of benzoyl peroxide on the pyrolysis of 11 R Reference solution, 0.2ggzg'in benzene. 31 Sample solution, 0.2g 2}. and 0.01! benzoyl peroxide in benzene. 82 Sample solution, 0.2;! “'1' and 0.025 benzoyl peroxide in benzene. 85 Sample solution, 0.2g Z} and 0.05§_benzoyl peroxide in benzene. 4.8 Q .5 .5 0 ln(% of undecomposed 71) c0 0: 00 N 2.8 Figure 12. 81 L .A ' 0 . 7 . k=4.44x10‘f5 A i0.09x10 d ..0 d -5 K=7.28x10 _5 4 ' i0.48x10 * k=9.24x10'5_ i0.43x 10 q A A j ' .~ ' I ' '7 r k=13.84x10:: 0 60 120 180 240 iO'QGXIO Time, min Kinetics of the thermal rearrangement of Zl'in benzene at 110°. 0 0.20fl Z}. in benzene. 0.20_y_1_ 2.1. and 0.011\_a benzoyl peroxide benzene. 0.20M 71 and 0.02fl benzoyl peroxide benzene. 0.20111 2}. and 0.05gl_ benzoyl peroxide benzene. in 5 82 2,2-Dichloro—4,5,6,6,7,8-hexamethylspiro[2.5]octa- 4,7-diene (9;) was prepared in the same way as the dibromo- carbene adduct Z}, using chloroform, potassium tfbutoxide and triene 32; 'The nmr spectrum (Figure 26) of 91 was very similar to that of Z}, It consisted of singlets at T 8.91 (3H), 8.73 (3H), 8.45 (6H), 8.30 (6H) and 7.98 (2H). The chemical analysis and the mass spectrum were also in accord with the structure of 21- Cl H‘ I V C1 CHC13 ——-———» KOEBu. 83 Compound Zl'was reduced to the parent hydrocarbon 31' with lithium and liquid ammonia by the procedure used by Birch°° to reduce 1,1-dihalocyclopropanes. Compound 31' was obtained in a 48% yield. Br V Br V Li/NH3 ———) 21 22. It is thermally unstable and decomposes to form a yellow oil when exposed to the atmosphere for a few hours. Due to its instability a good elemental analysis could not be obtained. It can be stored in 95% ethanol solution in the refrigerator. The spectral data are also in accord with the proposed structure of 31; The nmr spectrum (Figure 27) consisted of singlets at T 9.28 (4H), 8.93 (an), 8.74 (an) and 8.40 (6H). The ir spectrum (Figure 14) had bands at 3070 (cyclopropane ring-hydrogen band) and 1635 cm.1 (C=C). The uv Spectrum (cyclohexane) had a maximum at 213 nm (a 11,630). This ab— sorption could be compared to the intense absorption of spiro[2.4]hepta-4,6—diene (92) at 223 nm (EtOH, e = 6310) which appears to arise from the interaction of the 200 nm Rydberg diene transition (R <—-N band) with the 190~210 nm cyclopropane transition.61 It is suggested that with diene gg'the in-plane cyclopropane electronic transition has a 84 direction such that it can couple strongly with the diene 2p-—> 3s out-of—plane Rydberg transition of appropriate symmetry. Such an interaction is also possible for 31 and the higher intensity of the absorption in Qz'may be due to the greater flexibility of the six-membered ring allowing greater interaction of the cyclopropane ring. V V 92 37 w rw Since we had compound gz'in reasonable amounts we could study its chemistry further. For a photochemical investigation a 1% solution of gz'in ether was irradiated through quartz with a 450awatt Hanovia lamp. The nmr spectrum of the material recovered after 22 hr was identi- cal with the starting material. Next an acetone—sensitized photolysis was tried. This time a 1% solution of gz'in acetone was irradiated under similar conditions through Pyrex. Once again the nmr spectrum of the recovered material showed mostly starting material, but a small peak was observed at T 7.88 which may have been caused by the formation of a small amount of hexamethylbenzene. When the photolysis of 31 was carried out in the vapor phase with mercury sensitization using 2537 A resonance lamps in a Srinivasan-Griffin reactor for 21 hr at 40° still no change was observed in the starting material. The problem 85 here is that gz'has a very small vapor pressure. Therefore a very small amount of this compound is present in the vapor phase. The reaction was repeated at 71°, hoping that more of compound QZ’may be present in the vapor phase at this temperature. The nmr spectrum of the product showed mostly starting material and a large peak at 1 7.88 indi- cating the presence of hexamethylbenzene. It is not known whether hexamethylbenzene is a thermal or photo product, because compound gz'had been found to rearrange thermally to hexamethylbenzene. To investigate the thermal reaction of gz'a 10% solue tion of gz'in tetrachloroethylene (bp 121°) Was refluxed for 36 hr. This solvent was used because it would not inter- fere with the proton nmr spectrum of the thermal products. The solution was scanned by nmr, which showed the presence of a small amount of starting material and a strong peak at T 7.84 indicating the presence of hexamethylbenzene. In order to carry out a quantitative determination of the hexa- methylbenzene formed, a solution of 0.5 g of 31 in 10 ml of decalin was refluxed for 10 hr. The solutiOn was then concentrated and chromatographed over alumina. A white residue (0.2 g) was obtained. The nmr spectrum showed a singlet at T 7.84 and signals for the starting material. From the nmr spectrum,a 2.3% contamination of starting material was calculated. The mass spectrum of this residue showed the major peak at 162 (MW of hexamethylbenzene 162) and a small peak at 190 (MW of 31 190). From the relative 86 heights of the peaks it was confirmed that hexamethyl- benzene, obtained in 47.5% yield,was contaminated with 2.3% of 31. The mechanism of this reaction was not investigated. To investigate the behavior of compound gz'in an acidic medium it was added to a 2% solution of'pftoluene- sulfonic acid in carbon tetrachloride and refluxed for 15 hr. The solution was then concentrated and the residue was chromatographed over alumina. A 60% conversion and a 78.7% yield of hexamethylbenzene (93) was obtained. The rest of the material contained polymeric products and was not investigated further. When an nmr spectrum of gz'in trifluoroacetic acid was obtained, a signal at T 7.84 was. observed instantly indicating the rapid formation of hexa- methylbenzene. The following mechanism is proposed for this conversion. 3% 1 (~r~r H ————-5 —————-> CH2=CH2 g; 4 87 In order to justify this mechanism a low temperature study was undertaken. It has been reported that cyclo- propane is protonated and cleaved in FSOaH—SbFs-SOZClF solu- tion above -80° to Efbutyl and tfhexyl cations in varying amounts.62 A solution of 9Z/in methylene chloride was added slowly to fluorosulfonic acid at ~78°. An nmr spectrum was obtained at -80° of the resulting solution and it con- sisted of multiplets at T 8.50 and 7.08. There was not much change in the Spectrum by raising the temperature from -80° to 45°. Spectra were obtained at various temperatures between these two limits. It appears that the carbonium ion initially formed undergoes dimerization and trimeriza- tion, giving a complicated nmr spectrum. We had hoped to observe one of the intermediates at -80° or even a mixture of them. They should be stable since they are tertiary carbonium ions. Also because hexamethylbenzene is formed instantly at room temperature in strong acid solutions, we had hoped to observe the formation of hexamethylbenzene by gradually raising the temperature from -80° to 45°. No sig- nal was found to develop at T 7.84, indicating that start- ing material had probably polymerized. B. Preparation and Photolysis of 2,3-Epoxy-2,3,4,4,5,6-hexa- methyl-2,5-cyclohexadienone (94) and 2,3:5,6-Diepoxy-2,3, 4,4,5,6-hexamethyl-2,5-cyclohexadienone (99) Though some chemistry of 2,3,4,4,5,6-hexamethyl-2,5- cyclohexadienone (99) had been well studied in our laboratory, it has not been epoxidized before. The resulting 88- 2,3-epoxy-2,3,4,4,5,6-hexamethyl-2,5-cyclohexadienone (92), a Az- unsaturated, a,B-epoxyketone, should make an interesting subject for a photochemical study. The sodium salt of hydrogen peroxide or the sodium salt of tfbutyl hydroper- oxide are two reagents which are recommended for the epoxida- tion of a,B-unsaturated ketones.63 The reaction is believed to proceed by nucleophilic attack of the hydroperoxide ion on the B-carbon of the unsaturated ketone. But the attempt to prepare 94'from dienone 99'with sodium hydroxide and hydrogen per0xide was unsuccessful. Next hydrogen peroxide with sodium tungstate was used as the epoxidizing reagent. This is a very potent reagent for epoxidation. Payne and Williams64 found that maleic, fumaric and crotonic acid were very resistant to attack by peracetic or perbenzoic acid but were converted to their epoxides in good yields by hydrogen peroxide-sodium tungstate. However, 99'was not epoxidized by this reagent. Therefore mrchloroperbenzoic acid was used next. A vpc analysis of the product mixture showed that it contained two new products. When a mass spectrum was obtained of these compounds, one had a parent peak 16 units above that of the dienone, at mass 194, which indicated it might be epoxide 92; The second compound had a parent peak at mass 210 indicating it might be the diepox- ide 29. 89 O O 0 II II H C03H 35 94 95 W W w The uv spectrum (cyclohexane) of 92 had maxima at 323 nm (e 86) and 246 nm (e 8330). The infrared spectrum (Fig— ure 16) had bands at 1655 (conjugated c:0), 1625 (C2C) and at 1025 cm-1 (oxirane ring C-O). The nmr spectrum of 94 (CC14, Figure 28) showed gem dimethyls at T 8.91 (s, 3H)‘ and 8.71 (s, 3H). The allylic methyls were seen at T 8.26 (s, 3H) and 8.19 (s, 3H), whereas the methyls at C2 and C3 appeared as a singlet at T 8.60 (6H). In order to confirm these assignments 3,5-dimethyl-d5-2,4,4,6-tetramethyl-2,5- cyclohexadienone (99) was made as described previously.39 Epoxide 9Z'was obtained by the oxidation of 99 with m: chlorOperbenzoic acid. The nmr spectrum of 91'showed no 0 0 Cl 0 H I 03 Na/CH30D * «x- * -x- 22 22. 22: peak at T 8.19; therefore this signal can be assigned to the allylic methyl at C5. The signal at T 8.60 integrated for only 3 hydrogens instead of 6 hydrogens; therefore this signal is rightly assigned to the methyl groups at C2 and C3. 90 2,3-Epoxy-4,4-dimethyl-2,5-cyclohexadienone (99) is a known compound65 and the nmr spectra of 92 and gg'are given below for comparison. 0 O H 8.26 H H H 4.2 8.60-{() 6.6 O 8.19 H H 3.6 8.71 8.91 8.7 94 98 rw m The groups at C5 are deshielded more than the groups at C6 in both cases. Also the groups at C2 and C3 have identical chemical shifts. The mass spectrum of 92 showed the parent peak at mass 194 and a base peak at mass 152. The latter is probably formed from epoxide 92 by the elimination of propy- lene (42 mass units) as shown below, and may have structure 199, Ion 199'would be expected to be very stable because it is aromatic according to Huckel's 4n + 2w electron rule of aromaticity. O ‘0‘ :0: H 0 —e 0 CH3 CH3 CH3 CH3 2.2. :O: /’C \\ \\ CH3 CH3 + I /'O+ 91 The uv spectrum of diepoxide 99'had maxima at 240 nm (a 607) and at 217 nm (e 1960). The ir spectrum (Figure 17) had bands at 1690 (C=O) and 1025 cm-1 (oxirane ring C-O). The nmr spectral data for diepoxide 99'and diepoxide 99) which is a known compound,65 are given below: o H " H 6.9 AB quartet O 0 i=5 Hz H H 7.3 9.5 .87 92 9.93. Since the gem dimethyls of diepoxide 99'do not have identical chemical shifts, the compound probably has the syn config— uration. If 99'had the-anti configuration the geminal methyls would be expected to have identical chemical shifts. O The mass spectrum of 99'showed the parent peak at mass 210 and a base peak at m/e 140, which may represent ion 101 which can be obtained from diepoxide 99'by the elimination of carbon monoxide and propylene (28 +'42 mass units). H +10 :0: 1 1 92 For the photochemical study a 1% solution of 92 in ether was irradiated through Pyrex with a 450—watt Hanovia lamp. The reaction was monitored by vpc but no change was observed even after 6 hr. When the reaction was repeated using a 1% acetone solution of 94 under similar conditions, still no change was observed. The experiment with a 1% solu- tion of 94,in ether was repeated but with a vycor filter. The reaction was followed by vpc. As the reaction pro- ceeded the peak for 92 became smaller and a new peak which appeared grew larger.’ The reaction was complete after 10 hr. The photoproduct was collected by preparative vpc as a colorless oil. Elemental analysis and the mass spectrum of the photoproduct showed that it was isomeric with starting material. The photoproduct was identified as 6- acetyl-2,3,4,4,5,-pentamethyl-Z-cyclopentenone 199'on the basis of its spectral data. O 0 ll “ II N O hV -x- > > ether Na/CH30D *- 2; 102 103 The ir spectrum of 199;(Figure 18) showed bands at 1698 and 1647 cm.1 which are assigned to the B—diketonic groups. The uv spectrum (cyclohexane) had a maximum at 233 nm (8 14750). According to the empirical generaliza- tions derived by Woodward“:67 for the effect of substi- tuents on the position of w, w* absorption in the uv 93 spectrum of a,B-unsaturated ketones, 199'should have an absorption band at 247 nm. This predicted value is for spectra obtained in ethanol. In order to compare the ob- served value with the predicted value a correction has to be made for the solvent effect. A correction of +7 is rec- ommended for converting spectra from hexane to ethanol.66 Assuming that this correction should hold for cyclohexane,the observed 7, w* absorption band would appear at 240 nm in ethanol. Though this value is lower than the predicted value (247), it is in keeping with the observations of Gilliam and West68 that a,B-substituted cyclopentenones absorb at lower wavelengths than predicted by Woodward's rules. Gilliam and West found that substituted cyclopentenones absorb around 232-237 nm, and this value is in better agreement with our observation. The nmr spectrum (CC14, Figure 30) of 199'showed the geminal methyls as two singlets at T 9.0 (3H) and 8.94 (3H). A singlet at T 8.88 (3H) is assigned to the methyl group at C5. A multiplet at T 8.34 (3H) is assigned to the allylic methyl at C2 and the multiplet at T 8.06 (6H) represents the allylic methyl at C3 and the acetyl methyl group. In order to confirm the last assignment 199 was dissolved in 5 ml of 95% methanol-d containing sodium and refluxed for 6 hr to exchange the hydrogens of the allylic methyl at 03 and the acetyl methyl group. The nmr spectrum of the product showed no signal at T 8.06; therefore this signal is rightly assigned to the two groups that are capable of exchange. 94 The results of the photolysis of 94'are in accord with the observations of Jeger and coworkers51 for the photorear- rangement of Az-unsaturated a,B-epoxyketones which were men- tioned in the Introduction. Accordingly when 94 was irradi- ated through Pyrex (>2800 A) no photoreaction Was observed but when it was irradiated through vycor (>2100 A) rear- rangement took place. Therefore this reaction probably goes through a v, w* excited state, most likely the singlet state, because the reaction was not sensitized by acetone. 0* 11V; 7T, 7T* 9 0 O ) ° 0 )— O* o ‘1 0* L92. ‘ The preferential migration of the methylene group over the methyl group from the B-position to the a-position is in agreement with the migratory aptitude of the B-substituents found by Reusch and Markos.5° To further support the above mechanism and to show that the methyl at C5 in 94'ends up on the acetyl group of 199) we photolyzed 3,5-dimethyl-d6-2,3-epoxy-2,4,4,6-tetra- methyl-2,5-cyclohexadienone (99) which had been prepared 95 earlier. When 99'was irradiated under the same conditions as 92) the photoproduct obtained had an nmr spectrum identi— cal to that of 199; This supports the proposed mechanism for the photorearrangement by methylene migration. 0 0 l ' “ n 0 by 9(- —————+ * *- *- gg 103 In order to demonstrate the presence of an unsaturated five membered ring in 199'an attempt was made to hydrogenate 199'to prepare the correSponding saturated cyclopentanone. All attempts to hydrogenauathis compound were unsuccessful. Even when 199'was taken in an acidic ethanol solution with palladium on charcoal as the catalyst and subjected to hydrogen at 6 atm pressure, it still could not be hydrogenated. Diepoxide 99 was irradiated through various filters and finally through quartz using a 450dwatt Hanovia lamp, but it remained unchanged. When an acetone solution of 99' was irradiated through Pyrex under the same conditions for 16 hr, the nmr spectrum of the recovered material showed that starting material was still unchanged. Compound 99' seems to act like an ordinary saturated ketone. 96 C. Attempted Preparation of 2eMethylene—l,3,3L4,5-penta— methylbicyclo[2.2.2]octa-5,7-diene (99) Devrieze41 found that when dienone 99 was treated with methylmagnesium bromide only hydrocarbon'99'was isolated because the initially formed tertiary alcohol 104 dehydrates to give 99. u H / CHaMgBr —H20 _._____.) _____—.__, H20 L. i ii 104 22 Similarly when we treated 99 with isopropylmagnesium bromide a quantitative yield of hydrocarbon with a molecular formula C15H24 was obtained. The tertiary alcohol 105 can dehydrate to give two possible products, 106 or 107. — OH _ / O n W I )hMgBr ‘HZO, and/or H20 Q L 1 32.5.. 10 106 107 The nmr spectrum of the product showed that only 1- methylene-3—isopropyl-2,4,5,6,6—pentamethyl-2,4-cyclohexa— diene (107) was obtained. The nmr spectrum (Figure 31) in- cluded a Singlet at T 8.94 (6H) for the gem dimethyls. Two 97 peaks at T 8.85 (3H) and 8.71 (3H) probably represent the two methyls of the isopropyl group. A multiplet at T 8.33- 7.71 (9H)represents the three allylic methyls. The multie plet at T 7.06 (1H) represents the hydrogen of the iso- propyl group and the vinylic protons appear at T 5.10 (2H, g'= 2.3 Hz) as an AB quartet. In case structure 199 had been obtained, vinylic protons would not be present in the nmr Spectrum and six methyl groups would have been obtained in the allylic region instead of the three that were ob- served. Also the multiplet at T 7.06 for the tertiary pro- ton of the isopropyl group would not be observed if 199 had been obtained. The uv spectrum (cyclohexane) had maxima at 310 nm (e 3260), 255 nm (87050) and 220 nm (e 6800). Using the rules proposed by Fieser and Fieser69 for the calculation of uv spectra of steriodal dienes and polyenes a value of 318 nm is calculated for structure 191 and 279 nm for 199; The observed maximum at 310 nm is Consistent with structure 19Z'rather than structure 199, The ir spectrum (Figure 19) had bands at 1655 and 1625’c‘2m'1 (C:C). To further confirm that 107 had a conjugated cyclo- hexadiene structure, an adduct Of 107 was made with dimethyl acetylenedicarboxylate. It is assigned structure 99; COzMe I c *8 > 47 I COzMe H O q E 98 The chemical analysis and mass spectrum confirmed that 99 had a molecular formula of C21H3004. The nmr spectrum (Figure 32) consisted of singlets at T 9.03 (3H), 8.95 (6H), which may be due to the methyl groups at C3 and one of the methyls at C9. A doublet was observed at T 8.84 (3H, g_= 1.75 Hz) and probably represents the other methyl group at C9. Two singlets at T 8.56 (3H) and 8.45 (3H) are in the range of bridgehead methyls and therefore are assigned to the methyls at C1 and C4. The singlet at T 8.20 (3H) is assigned to the allylic methyl at C5. The methyls on the ester groups were seen as a singlet at T 6.40 (6H) and the two vinylic protons appeared at T 5.30 (s, 1H) and 5.14 (s, 1H). Compound 99'has a 2-methylenebicyclo[2.2.2]octadiene structure and we were interested in studying the photochem- istry of such a system. Since 99'is a very complicated molecule with various substituents we decided to prepare a simpler molecule with the basic 2-methylenebicyclo[2.2.2]- octa-5,7-diene structure. We chose to attempt the syn- thesis of 2-methylene-1,3,3,4,5-pentamethylbicyclo[2.2.2]- octa—5,7-diene (99) because a possible route for its prepara— tion from 2,3,4,4,5,6-hexamethyl-2,5-cyclohexadienone (99) could be formulated. The first step for the synthesis of this bicycloocta- diene was the preparation of 1-methylene-2,4,5,6,6-penta- methyl-2,4-cyclohexadiene (999) which had been made in this laboratory by Dr. A. Sheller by the reduction of dienone 99’ 99 with lithium aluminum hdyride.7° H H H O H l -.., + + % H 99’ 108 109 110 Dienone 99'on reduction probably forms the corresponding tertiary alcohol which dehydrates to give triene 199; When we repeated the reduction of 99 with lithium aluminum hydride, the vpc trace of the product mixture showed three peaks with retention times of 6.5, 13 and 17 min. The peak with a retention time of 17 min was the required triene 199, The nmr Spectrum of the collected material consisted of peaks at T 8.9 (s, 6H, gem dimethyls), 8.33 (s, 6H, methyls at C4 and C5), 8.13 (d, 3H, g.= 1.5 Hz, methyl at C2), 5.10 (S, 2H methylene group at C1), 4.56 (m, 1H, vinylic proton at C3). The nmr spectrum was identical to that obtained by Dr. Sheller for triene 199; The chemical analySis and the mass spectrum of the com— pound with a retention time of 6.5 min showed that it had a molecular formula of C12H20; that is there were two more hydrogens than in triene 999; This compound is assigned structure 199; The nmr spectrum (Figure 34) of this com- pound is conSistent with this structure, as it showed a doublet at T 9.21 (3H, q_- 7 Hz), two singlets at T 9.13 (3H) and 9.07 (3H), a multiplet at 8.38-8.30 (10H) and 100 another multiplet in the vinylic region at T 4.73 (1H). The ir spectrum (Figure 21) had a band at 1663 cm-1. Though double bonds are generally not reduced by lithium aluminum hydride, a,B-unsaturated carbonyl compounds form an excep- tion to this rule.71 They are first reduced to the cor- responding allylic alcohols which are further reduced to saturated alcohols gig organoaluminum intermediates. A similar mechanism is used below to explain the formation of 109 from dienone 3;. R H O)2A1H2 O —-> \\ )1- 22. H H H OH 0’ H20 (2529. H H 109 The compound with a retention time of 13 min had an nmr spectrum (Figure 33) which consisted of singlets at T 8.99 (6H), 8.43 (12 H), 7.60 (2H) and is assigned structure 112; Mass spectrum showed parent peak at 164. This com- pound was not investigated in great detail because it was a minor side product. Triene lgg'is very reactive and turns yellow on keep- ing. Therefore it cannot be stored for a long time and 101 must be used as quickly as possible. On adding maleic an- hydride to triene 128 an adduct was obtained in 93.5% yield. The adduct was assigned structure lll'on the basis of its nmr spectrum (Figure 35) which consisted of singlets at T 9.05 (3H), 8.89 (3H), 8.54 (6H) and doublets at T 8.28 (3H, g.= 1.7 Hz), 7.34 (1H, g_= 8.5 Hz), 6.90 (1H, g_- 8.5 Hz), two more singlets in the vinylic region at T 5.25 (1H) and 5.06 (1H), and a broad singlet at T 4.54. A band at 1763 cm.1 (c=0), was observed in the ir spectrum (Fig- ure 23). By oxidizing 111 with lead tetraacetate we hoped to obtain the desired compound, 2-methylene-1,3,3,4,5-penta- methylbicyclo[2.2.2]octa)5,7—diene (65). However, when 111 was oxidized with lead tetraacetate no reaction was observed and the nmr spectrum of the recovered material showed that it was unchanged starting material. H 0 I! H + 0 -—-¢ \\ \\ o 108 Another attempt was made to prepare 62 by hydrolyzing 111 62' “Al’to the diacid 112 and oxidizing it with lead tetra- aCetate. The major product, which formed 75% of the vola- tile products of this reaction, was not the desired compound Q2). The ir spectrum (Figure 24) of this product had a band at 1770 cm-1, which is typical of a y-lactone.72 The uv spectrum (cyclohexane) had maxima at 240 nm (a 8370), 220 102 (s 7940) and 214 nm (e 8370). The chemical analysis and mass spectrum showed that the lactone had a molecular form- ula of C15H2002. The nmr spectrum (Figure 36) consisted of singlets at T 9.06 (3H), 8.93 (3H), 8.78 (3H), 8.64 (3H), doublets at T 8.30 (3H, g_= 1.5 Hz) and 7.51 (1H, g_= - 6 Hz) and additional singlets at T 7.38 (1H), 5.38 (1H), 5.20 (1H) and a multiplet at T 4.66 (1H). On the basis of the spectral data and mechanistic considerations, the lac- tone has been assigned structure 113. j::zéj 111 112 rwv W // 0‘ ...; H The signal at T 7.38 is assigned to the proton at 0;. Com- pound 112 has been used as a model for this assignment. The tertiary proton in this compound has a chemical shift of T 7.31.73 This eliminates structure 112.1“ which the ter- tiary proton is at the C3 position and Should have a chemi- cal shift of about T 6.2. Structure ll§.is eliminated on the basis that the proton at C1 which Can also be assigned the chemical shift of 1 7.38 should be split into a doublet by the proton at C8 whereas the signal at T 7.38 appeared as a singlet in the nmr spectrum. OMeI “'Hm 11 i 115 116 A possible mechanism for the formation of 113 from 112 is presented below. ,// / H cozpb(0Ac)3 + AcOH H C02Pb(OAc)3 H Q + 2-Pb(0Ac)3 + C02 do2 o d.‘11,2-shift <5 <5 0 H =:O H P- I: H e. > \\ H... H H \v/ h 0 \ H (H H o ‘0 H 11 A 1,2—shift from C1 to C7 in step d would result in compound to i 116 which has been eliminated on the basis of the nmr spectrum. 104 A y-lactone can be reduced to the corresponding ether with sodium borohydride and boron trifluoride.74 When 11% was reduced by these reagents the ir spectrum of the pro- duct showed that the carbonyl function was not present. The molecular weight of the product was determined to be 220 from the mass spectrum. This indicated that not only the carbonyl function of the lactone group had been reduced but also one of the double bonds, probably the exocyclic double bond, had been hydrogenated. Since we had only a few milligrams of the y-lactone an appreciable amount of the reduction product was not obtained to carry out a com- plete structure determination. Since the 2-methylenebicyclo[2.2.2]octa—5,7-diene system was not obtained an attempt was made to form an adduct of triene 128 with 2-butyne to form this bicyclic system directly. All attempts to form this adduct 111 were unsuccessful. H CH3 // + ,é ——x+—) H §> f /’ CH3 108 117 EXPERIMENTAL 1. Microanalyses: Microanalyses were carried out by Spang Microanalytical Laboratory, Ann Arbor, Michigan. 2. Nuclear Magnetic Resonance Spectra: These spectra were recorded on a Varian Model A—60 using CCl4 solutions with tetramethylsilane as an internal standard. 3. Infrared Spectra: Infrared spectra were obtained with a Unicam Sp—200 spectrophotometer in CCl4 solution (unless otherwise stated). 4. Ultraviolet Spectra: These spectra were obtained with a Unicam SP—800 spectrometer. 5. Mass Spectra: Mass spectra were obtained by Mrs. R. Guile with a Hitachi-Perkin—Elmer RMU-6 mass spectrometer. 6. Melting Points: Melting points were determined with a Gallenkamp Melting Point Apparatus and are uncorrected. 7. Preparation of £1 from lenethylene-Z,3,4,4,5,6-hexa- methyl-2,5-cyclohexadiene (32): To a vigorously stirred solution of 0.25 g of cupric acetate monohydrate in 15 ml of glacial acetic acid was added 4 g of zinc dust. The Zn—Cu couple formed at once and was allowed to settle for 1 min. 105 106 Glacial acetic acid was decanted and the couple was washed with more glacial acetic acid (2 x 15 ml) followed by an- hydrous ether (3 x 12 ml). More anhydrous ether (10 ml) was added to the couple. The erlenmeyer flask which contained the couple was fitted with a reflux con- denser and a dropping funnel. A few drops of methylene iodide were added to start the reaction. A solution of 4.56 g (0.026 mole) of gg'and 21.44 9 (0.0352 mole) of methylene iodide in 25 m1 of anhydrous ether was added to the Zn-Cu couple over 1 hr. The resulting solution was refluxed for 30 hr. The ethereal layer was decanted and washed thrice with a solution of 13 HCl containing ice cubes and twice with water. The washed ethereal layer was dried (NaZSO4), concentrated and analyzed by vpc using a 5-ft 20% SE—3O column at 1500 with the flow rate of Helium at 75 ml/min. Three peaks with a retention time of 3, 5 and 14 min repre- sented methylene iodide, triene 32 and compound gz'respec- tively. From the areas under the peaks a conversion of 61.5% was calculated and gz'was obtained in 95% yield: ir (CC14) 1635 cm.1 (C:C); nmr (CC14) T 9.28 (s, 4H), 8.90 (s, 6H), 8.74 (s, 6H), 8.40 (s, 6H) and a singlet at T 7.78. 8. Preparation of 2,2-Dibromo-4,5,6,6,7,8-hexamethyl- spiro[2.5]octa-4,7-diene (Zl):fk3a.well stirred solution of 15.84 g (0.09 mole) of triene 32 and 15.75 g (0.108 mole, assuming it contains 38%‘E-BuOH) of potassium Efbutoxide in pentane (100 ml) at 0° was added bromoform (27.9 g, 107 0.108 mole) over 10 min. The mixture was stirred for 4 hr at room temperature. water (50 ml) was added to the mixture which was stirred for 10 min. The organic layer was separ- ated and washed with water whereas the aqueous layer was washed with two 30-ml portions of pentane. The combined pentane layers were dried (Na2804) and concentrated to yield impure product (36 g). The crude product was treated with methanol (10 ml). When the methanol solution was cooled, the dibromocarbene adduct crystallized out to give 14 g (44.7%) of crystals. Since the methanol solution still contained some un— reacted triene 32, it was concentrated and subjected to the reaction again uSing 12 g of bromoform and 7 g of potassium Efbutoxide. This time 2.5 g of product was obtained giving an overall yield of 52.75%: mp 88°; ir spectrum is shown in Figure 13; nmr (CC14, Figure 25) T 8.86 (s, 3H), 8.78 (s, 3H), 8.38 (s, 6H‘, 8.20 (s, 6H‘, and 7.81 (s, 2H); mass spec- trum (70 ev, inlet temperature 100°) m/e (rel intensity) 350(10), 348(20), 346(10), 335(4), 333(7), 331 (4), 269(3), 267(3), 254(8), 252(9), 175(100); mass spectrum (inlet temp 250°) m/e (rel intensity) 269(5), 267(6), 254(12), 252(12), 171(100). ' ' ' Aggi, Calcd for C14H208r2: c, 48.29: H, 5.79; Br, 45.91. Found : C, 48.29; H, 5.86; Br, 44.57. 9. Photolysis of 71: A 2% solution of Zl'in pentane (15 ml) was taken in a quartz test tube. The Solution was 108 irradiated with a 450—watt Hanovia lamp through vycor for 2 hr. At the end of this time the solution had become dark yellow and a black residue had deposited on the walls of the test tube. The reaction was repeated using a corex filter but the black residue was observed even after 1/2 hr. 10. Reaction of Zl'with.MethylLUflium: To a solution of 0.348 g (0.001 mole) of Zl'in 10 ml of anhydrous ether at -78° was added a 0.5 ml Of 2M_solution of methyllithium (0.027 g) in anhydrous ether. The solution was stirred for 30 min and the excess of lithium salt was decomposed with water. The ethereal layer was separated, dried (Na2804), and concentrated to yield a white residue. The nmr spec— trum of the residue showed peaks at T 8.86 (s, 3H), 8.78 (s, 3H), 8.38 (s, 6H), 8.20 (s, 6H) and 7.81 (s, 2H). The reaction was rerun at 0°. The nmr spectrum of the product had no peaks in the vinylic region. A vpc analysis with a 5-ft 20% SE-30 column at 160°, with helium flowing through at a rate of 75 ml/min showed a small peak at a retention time of 3 min and a large peak with a retention time of 6 min. 11. Pyrolysis of zg'in solution: A solution of Zl'in the required solvent waS'taken in a Pyrex tube which was degassed, sealed and heated in an oil bath for 3 or 8 hr. The tube was then cooled to -78° and opened. The solution was concentrated and analyzed by vpc using a 5-ft 20% SE-52 column at 180° with helium flowing through at a rate of 75 ml/min. 109 Compound 82; mp 74-760; retention time 37 min; ir (CC14) 1625 cm’l; uv max (cyclohexane) 227 nm (8 1090): nmr (CCl4): T 7.84 (s, 9H), 7.74 (s, 6H), 4.42 (d, 1H, g_= 1.25 Hz), 4.10 (d, 1H, g.= 1.25 Hz); mass spectrum (70 ev) m/e (rel intensity)254(7), 252(7), 172(100), 157(100). Compound 8;; mp 74-75°; retention time 44 min; ir (CCl4) 1615 cm’l; uv max (hexane) 214 nm (5 16580); nmr (cc14): T 7.95 (s, 9H), 7.90 (s, 6H), 3.62 (d, 1H, g_= 7.5 Hz) 2.99(d, 1H, J = 7.5 Hz); mass spectrum (56 ev) m/e (rel intensity) 254(45), 252(46), 172(86), 158(100). Anal, Calcd for C13H17Br: c, 61.66; H, 6.77; Br, 31.56. Found: C, 61.94; H, 6.82; Br, 30.41. Compound 82; mp 76-77°; retention time 51 min: ir 1 (C=C): uv max (hexane) 214 nm (6 13900); (cc14) 1610 em’ nmr (cc14): T 7.89 (s, 15H), 4.15 (d, 1H, q_- 14 Hz), 3.11 (d, 1H, g.- 14 Hz); mass spectrum (56 ev) m/e (rel intensity) 254(27), 252(28), 172(69), 157(100). Aggl, Calcdfor C13H17Br: c, 61.66; H, 6.77; Br, 31.56. Found: C, 61.20; H, 6.67; Br, 30.74. 12. Preparation of 82; Pentamethylacetophenone (1.0 g, 0.0053 mole) was refluxed With PBr3 (4.0 ml, 11.4 g, 0.041 mole) for 5 hr. The mixture was poured into a beaker con- taining water and ice cubes, and extracted with ether. The ethereal extract was dried (Na2504) and concentrated to yield 0.7 g of a solid. The residue was dissolved in a small amount of acetone and analyzed by vpc using a 5-ft 110 20% SE-30 column at 135°, with helium flowing through at a rate of 75 ml/min; the retention time of the main peak was 12 min and it amounted to a 59.5% yield: mp 74-76°; ir 1 (C=C); uv max (hexane) 227 nm (a 1090); cc14) 1625 cm’ nmr (cc14): T 7.84 (s, 9H), 7.74 (s, 6H), 4.42 (d, 1H, g_= 1.25 Hz); 4.10 (d, 1H, _q_ = 1.25 Hz). 13. ggeparation of 1—Dideuteromethylene—213,4,4,5,6- hexamethyl-Z,5—cyclohexadiene (82): a) Magnesium turnings (1.34 g, 0.055 mole) were taken in a 3-necked 100-ml flask fitted with a condenser and an addition funnel. Methyl-d3 iodide (8 g, 0.055 mole) in 50 ml of ether was added slowly. The resulting solution was stirred for 2 hr at room tempera- ature. A solution of 6 g (0.0338 mole) of dienone gé'in 25 m1 of ether was added to the Grignard reagent over 30 min and the solution was refluxed for 1.5 hr, cooled in an ice bath and hydrolyzed by 10 ml of D20 containing 6 drops of D2804. The ethereal layer was separated, dried (MgSO4) and concentrated to yield 6.0 g of crude product which was crystallized from methanol-d to give 3 g of pure 82; nmr (CC14): T 8.9 (s, 6H), 8.17 (s, 12H) and 5.24 (s,'0.55 H). b) ImMethylene-2,3,4,4,5,6-hexamethyl-2,5-cyclohexa- diene (32“ 1 g) was dissolved in 10 ml of carbon tetra- chloride and 10 ml of D20 containing 6 drops of D2804 was added to this solution and the mixture was stirred for 45 min. At the end of this time the D20 layer was replaced by a fresh amount of D20 containing D2804. The reaction was 111 continued for 5 more hours and the D20 layer was replaced again. The reaction was stopped after being stirred for 17 more hours, the carbon tetrachloride was separated, dried (M9804), and concentrated: mass spectrum (70 ev) m/e (rel intensity) 162(100), 176(12), 177(24), 178(25), 179(16), 180(9), 181(4), 182(2), 183(1). A 14. Effect of Benzoyl Peroxide on the Rate of Decompo— sition of 2i2-Dibromo-4L5,6,6;7,8-hexamethylspiro[2.5]octa— 4,7-diene (11): Duplicate samples (0.5 ml) of 0.20M 71 con- taining 0.01M, 0.02M_and 0.05M benzoyl peroxide in benzene were taken in nmr tubes and degassed to 0.2 mm of Hg and sealed. These were heated in an oil bath at 110° 1 1°. Every 30 min the tubes were removed from the oil bath and the reaction was quenched by cooling them in a dry ice-acetone bath. The nmr spectrum was recorded and the tubes reheated for another 30 min. The results are shown in Figure 11. Two 0.5-m1 samples of 0.20M_Zl'in benzene were taken in nmr tubes which were degassed to 0.2 mm of Hg and sealed. These were heated in an oil bath at 110° 1 2° and the re- action was followed by nmr as above. The results are shown in Figure 11. 15. Preparation of 2,2-Dichloro-4,5,6,647,8-hexamethyl- spiro[2.5]octa-4,7—diene (21): Compound 21 was made by the same procedure used for the dibromo compound 21“ using triene §2, chloroform and potassium Efbutoxide: mp loo-101.50; yield 44%; the ir spectrum is shown in Figure 14; uv max (cyclohexane) 112 234 nm (e 5850) and 206 nm (e 5540); nmr (CC14, Figure 26): T 8.91 (s, 3H), 8.73 (s, 3H), 8.45 (6H), 8.30 (s, 6H) and 7.98 (s, 2H). ' ‘ Anal, Calcd for C14H20C12: C, 64.86; H, 7.78; Cl, 27.32. Found: C, 64.29; H, 7.66; Cl, 26.40. 16. Reduction of Zl'to 4,5,6,6,7,8-Hexamethy1spiro— [2.5]octa-4,7-diene (31): To a blue solution of 2.0 g (0.284 atom) of lithium in 300 ml of ammonia was added a solution of 8.0 g (0.023 mole) of Zl'in 80 ml of anhydrous ether, over 30 min. The resulting Solution was stirred for an additional 2 hr. Then 14.98 g (0.28 mole) of ammonium chloride was added to the solution, whereupon the blue color disappeared. Ammonia was evaporated and the residue was treated with 30 ml of water and extracted with three 20-ml portions of ether. The ethereal extract was dried (Na2804) and concentrated to give 3 g of crude product which was crystallized from 95% ethanol to give 2.1 g (47.9%) of pure product: mp 65—67°; ir spectrum is shown in Figure 15: uv (hexane) 213 nm (a 11630); nmr (CCl4, Figure 27): T 9.28 (s, 4H), 8.93 (s,6H), 8.74 (s, 6H), and 8.40 (s, 6H); mass spectrum (70 ev) m/e (rel intensity), 81(22), 119(39), 133(25), 147(65), 161(42), 162(46), 175(100), 190(81). 17. Photolysis of 31; A 1% solution of Qz'in anhydrous ether was irradiated with a 450~watt Hanovia lamp through quartz for 22 hr. An nmr spectrum of the solid recovered from the ether solution showed singlets at T 9.28 (s, 4H), 8.93 (s, 6H), 8.74 (s, 6H), and 8.40 (s, 6H). 113 18. Acetone-sensitized Photolysis of 31; A 1% solu- tion of gz'in acetone was irradiated through Pyrex under the same conditions as above for 16 hr. The nmr spectrum obtained was the same as above except for an additional peak at T 7.88. 19. Mercury-sensitized photolysis of 31; Compound 31' together with a drop of mercury, was placed in a long quartz tube. This was cooled to -78° and evacuated. Next dry nitrogen was let into the tube and it was allowed to warm up slowly to room temperature. This procedure was repeated two more times and the quartz was finally evacuated to a 0.1 mm pressure. The evacuated tube was irradiated in a Srinivasan-Griffin reactor using 2537 A resonance lamps for 21 hr at 40°. At the end of this time the tube was removed from the reactor and cooled to -78°. Dry air was let into the tube and the residue was extracted with pentane. The nmr spectrum of the product showed singlets at T 9.28 (s, 4H), 8.93 (s, 6H), 8.74 (s, 6H), 8.40 (s, 6H) and 7.88 (s). The whole procedure was repeated but this time the re- actor temperature was maintained at 71°. The nmr spectrum obtained was identical with the one obtained when the reac- tion was carried out at 40°. 20. Pyrolysis of gz'in Solution: a) A 10% solution of u in tetrachloroethylene was heated for 36 hr at 121°. This solution was scanned directly by nmr and the spectrum showed small singlets at T 9.28, 8.93, 8.74, 8.40 and a very large singlet at T 7.88. 114 b) A solution of 0.5 9 (0.0026 mole) of gz'in 10 ml of decalin was refluxed for 10 hr. The decalin'solution was concentrated and chromatographed over 25 g of alumina. The column was eluted with 120 ml of pentane. A white resi- due (0.2 g) was obtained from the pentane solution; nmr (CC14): small peaks at T 9.28, 8.93, 8.74, 8.40 and a large peak at T 7.88; mass spectrum (70 ev) m/e (rel intensity) 147(100), 162(53), 190(2). 21. Acid-catalyzed Rearrangement of 31; Compound 31 (0.5 9, 0.0026 mole) was dissolved in 25 ml of a 2% solution of pftoluenesulphonic acid in carbon tetrachloride. The resulting solution was refluxed for 15 hr, and then washed with two 10-ml portions of water, dried (Nazso4) and concen- trated to yield 0.6 g of a black residue. The residue was chromatographed over a column of alumina which was eluted with pentane. Hexamethylbenzene (0.2 g) and 21 (0.2 g) were obtained from the pentane solutions giving 60% conver- sion and 78.7% yield of hexamethylbenzene. 22. Low Temperature nmr Study ofgg; To 1 m1 of FSO3H at -78° was added a solution of 0.10 9 Of 22,1“ methylene chloride dropwise and the solution was stirred vigorously. The acid layer was pipetted into an nmr tube containing 0.006 g tetramethylammonium fluoroborate. The nmr spectra were obtained at -80° to 45° at 10° intervals: nmr (PS03H) T 8.50 (m), 7.05 (m). 115 23. Attempted Preparation of gé'Using Sodium Hydroxide and Hydrogen Peroxide: In a 3-neckéd 100-m1 flask equipped with a mechanical stirrer, thermometer and dropping funnel was placed a solution of 1.78 g (0.01 mole) of dienone 32' and 2.9 ml (0.03 mole) of 30% hydrogen peroxide in 10 ml‘ of methanol. The flask was cooled to 15° and a solution of 0.2 g (0.005 mole) of sodium hydroxide in 0.5 ml of water was added slowly over 15 min. The mixture was stirred for an additional 3 hr at 20-25°, then poured into 200 ml of cold water. The aqueous solution was extracted with three 50-ml portions of ether. The ethereal solution was dried (Na2804) and concentrated. The residue consisted of 1.78 g of a white material: nmr: T 8.82 (s, 6H), 8.18 (s, 6H), 8.05 (s, 6H). 24. Attempted Preparation of 94 Using Sodium Tungstate and Hydrogen Peroxide: A solution of 1.78 g (0.01 mole) of dienone gg'and 1.65 g (0.005 mole) of sodium tungstate di— hydrate in 20 ml of ether was placed in a 3-necked, 100-ml flask fitted with a reflux condenser, dropping funnel and a magnetic stirring bar. The solution was stirred vigorously and 20 ml (0.208 mole) of 30% hydrogen peroxide was added to it over 45 min at room temperature. The resulting solu- tion was refluxed for 45 min, then poured into 100 ml of cold water. The aqueous solution was treated with ether as above and 1.78 g of a white material was recovered from the ethereal extract: nmr (CC14): T 8.82 (s, 6H), 8.18 (s, 6H), 8.05 (s, 6H). 116 25. Preparation of 94 using mfchloroperbenzoic Acid: To a well-stirred solution of 3.58 g (0.02 mole) of dienone Qé'in 60 ml of benzene at 0° was added 5.75 g(0.03 mole) Of mrchloroperbenzoic acid. The solution was further stirred for 1 hr at 0° and then at room temperature for 24 hr. At the end of this time the benzene solution was washed with three 30-ml portions of 10% sodium hydroxide solution and with three 30-m1 portions of water. It was then dried (Na2804) and concentrated to give 3.2 g of a white solid. This solid was dissolved in a small amount of acetone and analyzed by vpc using a 10-ft column of 20% DEGS on chromo- sorb 60/80 at 150°. The flow rate of helium was 170 ml/min and the injector temperature was 180°. Three peaks in a ratio of 1:19:10 were observed. The first peak had a retention time of 8 min which is identical with that of dienone fig; The second peak had a retention time of 11 min and was found to be the epoxide 22“ while the third peak had a retention time of 18 min and was found to be the di- e poxide 9,5,. SpeCtral data and analysis of 22; mp 48—48.5°; the ir spectrum is shown in Figure 16; nmr (CC14, Figure 28): T 8.91 (s, 3H), 8.71 (s, 3H), 8.26 (s, 3H), 8.19 (s, 3H), 8.60 (s, 6H): uv max (cyclohexane) 323 nm (a 86) and 246 nm (a 8330); mass spectrum (70 ev) m/e (rel intensity) 194(5), 188(4), 152(100), 137(78). Anal, Calcd for C12H1802: C, 74.19: H, 9.34. Found: C, 73.93; H, 9.29. 117 Spectral data and analysis of 92; mp 80-81°; the ir spectrum is shown in Figure 17; nmr (CCl4, Figure 29): T 8.96 (8, 3H), 8.59 (S, 3H), 8.68 (s, 6H), and 8.63 (S, 6H): uv max (cyclohexane): 240 nm (e 607), 217 (e 1960); mass spectrum (70 ev), m/e (rel intensity): 210(4), 140(100), 168(28). ‘ Anal. Calcd for C12H1803: C, 68.54; H, 8.63 Found: C, 68.39; H, 8.59. 26. Photolysis of 2,3-Epoxy-2,3,4,4,5,6-hexamethyl— 2,5-cyclohexadienone (94): A 2% solution of 22 was prepared by dissolving 0.6 g of gé'in 30 ml of ether. The solution was irradiated through vycor with a 450-watt Hanovia lamp. The photolysis was followed by Vpc. A 5-ft column of 20% SE-30 on chromosorb 60/80 at 144° was used for vpc analysis. The flow rate of helium was 150 ml/min. As the reaction proceeded the peak for the starting material at a retention time of 9.5 min decreased and a new peak appeared at a re- tention time of 7.5 min. After 10 hr the starting material had disappeared and the photoproduct was collected by prepar— ative vpc: the ir spectrum is shown in Figure 18; nmr (CCl4, Figure 30): T 9.0 (s, 3H), 8.94 (s, 3H), 8.88 (s, 3H), 8.34 (m, 3H), and 8.06 (m, 6H): uv max (cyclohexane): 233 nm (8 14750): mass spectrum (70 ev) m/e (rel intensity): 194(3), 152(82), 137(100). "Aggi. Calcd for C12H1302: c, 74.19; H, 9.34. Found: C, 74.23: H, 9.33. 118 27. Preparation of 6-Acetyl-d3-3-methyl-d3-2,4,4,5- tetramethyl-Z-cyclopentenone (103): To a solution of 100 mg of 6-acetyl-2,3,4,4,5-pentamethy1—2-cyclopentenone (122) in 5 ml of 95% methanol—d was added 50 mg of sodium. The solu- tion was refluxed for 6 hr under anhydrous conditions. At the end of this time the solution was analyzed by nmr which showed that the multiplet at T 8.06 (6H) had disappeared; nmr (CC14): T 9.0 (s, 3H‘, 8.94 (s, 3H), 8.88 (s, 3H), and 8.34 (m, 3H). 28. Preparation and Photolysis of 3,5-Dimethyl-d6— 2,3-epoxy-2,4,4,6-tetramethy1-2,5-cyclohexadienone (21):. 3,5-Dimethyl-d6-2,4,4,6-tetramethyl-2,5-cyclohexadienone (96) was prepared as described previously.39 Next, 26 was oxidized by mfchloroperbenxoic acid as described for'dienone 32“ and gz'was collected by vpc: nmr (CC14): T 8.92 (s, 3H), 8.71 (s, 3H), 8.60 (s, 3H), 8.26 (s, 3H). Compound gz'was photolyzed under the same conditions as those used fOr the nondeuterated compound. The photo- product was collected by vpc: nmr (CC14): T 9.0 (s, 3H), 8.94 (s, 3H), 8.88 (s, 3H), and 8.34 (s, 3H). 29. Attemptedgydrogenation of 6-Acetyl-2,3,4,4,5—penta- methyl-Z-cyclOpentenone (102): To a solution of 0.4 g (0.00206 mole) of ggg'in'4'm1 of ethanol was added 3 drops of hydrochloric acid. The solution was taken in a Parr hydrogenator and subjected to hydrogen at a pressure of 6 atm for 24 hr. The ethanol solution was neutralized (NaHCOa), dried (Na2S04) and filtered. A vpc analysis with 119 a 5—ft column of 20% SE-30 on chromosorb 60/80 at 150°, with a flow rate of helium at 200 ml/min showed that the filtrate contained only starting material. 30. Attempted Photolysis of 2,3;5,6—Diepoxy-2,3,4,4, 5,6-hexamethyl-2,5—cyclohexadienone (22): A 1% solution of gé'in ether (100 mg in 10 ml) was irradiated through quartz With a 450-watt Hanovia lamp for 9 hr. The ethereal solution was evaporated and the residue was dissolved in carbon tetra- chloride. An nmr analysis of this solution showed that the starting material remained unchanged. A 1% solution of gg'in acetone was irradiated for 16 hr through Pyrex with a 450-watt Hanovia lamp. When the solution was analyzed by nmr the starting material was found to be unchanged under these conditions. 31. Preparation of 1-Methylene-3-isopropyl-2,4,5,6,6- pentamethyl-2,4—cyclohexadiene (107): In a 3-necked 100-ml flask fitted with a condenser and an additional funnel was placed 0.972 g (0.04 atom) of magnesium. A solution of 4.92 g (0.04 mole) of isopropyl bromide in 20 ml of anhydrous ether was added slowly. The resulting solution was stirred for 2 hr at room temperature. Next, a solution of 3.56 g (0.02 mole) of dienone gé'in 50 ml of anhydrous ether was added slowly to the Grignard reagent. The solution was re- fluxed for 1.5 hr, cooled in an ice bath and hydrolyzed with 20 ml of 10% aqueous ammonium chloride solution. The ethereal layer was separated, washed with water, dried 120 (Na2804) and concentrated giving 4.08 g (100%) of 127; A Small amount was purified by vpc using a 5-ft column of 20% SE-30 on chromosorb 60/80 at 145° with the flow rate of helium at 150 ml/min: uv max (cyclohexane): 310 nm (e 3260), 255 nm (8 7050), 220 nm (8 6800); ir (CC14, Figure 19): 1655 and 1625 cm-1, (c=C); nmr in col4 (Figure 31): T 8.94 (s, 6H), 8.85 (s, 3H), 8.71 (s, 3H), 8.33-7.71 (m, 9H), 7.06 (m, 1H), and 5.10 (AB q, 2H, g_= 2.3 Hz). Anal, Calcd for C15H24: C, 88.22; H, 11.84. Found: C, 88.04; H, 11.91. 32. Preparation of the Adduct ggLof 107 with Dimethyl Acetylenedicarboxylate: A solution cf 2.04 g ((L01 mole) of lgl'and 1.42 g (0.01 mole) of dimethyl acetylenedicar- boxylate in 25 ml of benzene was refluxed for 12 hr. The solution was concentrated and distilled giving 1.8 g (52%) of the adduct. A sample was prepared for analysis by in? jecting a small amount of the adduct on to a 5-ft column of 20% SE-30 on chromosorb 60/80 at 2050 with helium flowing through at a rate of 75 ml/min: bp 137-1390 (0.07 mm); ir (CCl4, Figure 20): 1615 (c=0), 1630 (c=C), 1715 (c=0) and 893 cm-1 (=CH2); nmr in CCl4 (Figure 32): T 9.03 (s, 3H), 8.95 (s, an), 8.84 (d, 3H, g_# 1.75 Hz), 8.56 (s, 3H), 8.45 (s, 3H), 3.20 (s, an), 6.40 (s, 6H), 5.30 (s, 1H) and 5.14 (s, 1H): mass Spectrum (70 ev) m/e (relative intensity): 346(16), 246(100). ' Anal. Calcd for C21H3004: c, 72.80; H, 8.73. Found: C, 72.71; H, 8.68. 121 33. Preparation of 1-Methylene—2,445,§,6epentamethyl- 2,4-cyclohexadiene (108): In a 3-necked, 100-ml flask fitted with a mechanical stirrer, reflux condenser and dropping fun— nel was placed 0.6 g (0.0158 mole) of lithium aluminum hy- dride. A solution of 5.0 9 (0.0281 mole) of dienone 32 in 50 ml of ether was added over 30 min to the lithium aluminum hy- dride. The solution was further refluxed for 1 hr and then cooled in an ice bath. Water (1.3 ml) was added to the solu- tion to destroy the excess of lithium aluminum hydride. The ethereal solution was dried (Na2804) and concentrated to yield 4 g (79%) of a colorless oil. The oil was analyzed by vpc using a 10-ft column of 20% Carbowax on chromosorb 60/80 at 180° with helium flowing through at a rate of 75 ml/min, injector temp 240°. Three peaks with a retention time of 6.5, 13 and 17 min with a relative ratio of 2.8:1:2.3 were observed. The first peak was due to a new compound 109 which was charac- terized by spectral data and chemical analysis. The second peak was due to diene llg'and the third peak was due to the desired triene 128; Spectral data and analysis of 122; ir (CC14, Figure 21): 1663 cm”1 (c=c); nmr in ccl4 (Figure‘34): T 9.21 (d, 3H, g = 7 Hz), 9.13 (s, 3H), 9.07 (s, 3H), 8.38-8.30 (m, 10H), 4.73 (m, 1H); mass spectrum (70 ev) m/e (rel intensity): 164(67), 149(77), 124(83), 123(100). 522$: Calcd for C12H20: C, 87.78; H, 12.27. Found: C, 87.81; H, 12.12. Spectral data of $12; ir (CC14, Figure 22): 3400 (OH) and 1640 cm-1 (C=C); nmr'(CCl4, Figure 33): T 8.99 (s, 6H), 8.43 (s, 12H), 7.60 (broad s, 2H); mass spectrum (70 ev) m/e (rel intensity): 164(6), 149(100), 134(16), 133(17), 119(13). 122 Nmr spectrum of 108: (CC14): T 8.9 (s, 6H), 8.33 (s, 6H), 8.13 (d, 3H,'J'= 1.5 Hz), 5.10 (5, 2H), 4.56 (m, 1H). 34. Preparation of Adduct 111 of 108 with Maleic Anhydride: A solution of 0.2 g (0.00124 mole) of triene lgg'in 15 ml of carbon tetrachloride was placed in a 100—ml 3-necked flask at 0°. A solution of 0.121 g (0.00124 mole) of maleic anhydride in 5 ml of carbon tetrachloride was added slowly with the help of a dropping funnel. The re- sulting solution was stirred forCL5 hr, and concentrated giving 0.3 g (93.5%) of residue. The residue was crystal- lized from pentane: mp 125-127°; ir (CC14, Figure 23): 1763 (c=0), 1635 (c=c), and 940 cm’1 (=CH2); nmr (cc14, Figure 35): T 9.05 (s, 3H), 8.89 (s, 3H), 8.54 (s, 6H), 8.28 (d, 3H, g_= 1.7 Hz), 7.34 (d, 1H, g_= 8.5 Hz), 6.90 (d, 1H, g_= 8.5 Hz), 5.25 (s, 1H), 5.06 (s, 1H), and 4.54 (broad s, 1H). Anal, Calcd for C16H2003: C, 73.82; H, 7.74. Found: C, 73.70; H, 7.71. 35. Attempted Oxidation of Adduct 111: Oxygen was bubbled through a solution of 0.5 9 (0.00192 mole) of lll in 5 ml of pyridine. To this solution was added 1.28 9 (0.00288 mole) of lead tetraacetate and the solution was heated in an oil bath at 67° until the evolution of carbon dioxide ceased. The flask was then immersed in an ice bath and 5 ml of cold methylene chloride was added to the solution. 123 The resulting solution was poured into a 200-ml beaker and 15 ml of 50:50 mixture of nitric acid and water was added, This solution was extracted with five 10-ml portions of methylene chloride. The methylene chloride extract was washed with 50 ml of 5% aqueous sodium bicarbonate solution, dried (Na2SO4) and concentrated giving 0.2 g of material: nmr (CC14): T 9.05 (s, 3H), 8.89 (s, 3H), 8.54 (s, 6H), 8.28 (d, 3H, g_= 1.7 Hz), 7.34 (d, 1H, g.= 8.5 Hz), 6.90 (d, 1H, a.= 8.5 Hz), 5.25 (s, 1H), 5.06 (s, 1H), and 4.54 (broad s, 1H). 36. Preparation of Diacid llg; A suspension of 1.0 g (0.00384 mole) of anhydride lll'in 100 ml of 5% aqueous sodium bicarbonate was refluxed for 24 hr. At the end of this time the insoluble anhydride was filtered (0.5 g), the solution was acidified and the diacid llg (0.535 g, 50% conversion, 100% yield) precipitated from the solution: ir (nujol), 1730 (c=0), 1690 (c=0), and 895 cm"1 (=CH2). 37. Oxidation of Diacid 112: Compound 112 (0.5 g, 0.001798 mole) was oxidized with lead tetraacetate by the same procedure described for anhydride lll, The product was a yellow oil (0.3 g) and was analyzed by vpc using a 5-ft column of 20% SE—30 on chromosorb 60/80 at 175° with a flow rate of helium at 75 ml/min. The vpc chromatogram indicated that 75% of the product consisted of a compound with a retention time of 6 min. The spectral data and chemical analysis indicate that this compound is 113: ir 124 (CC14, Figure 24): 1770 cm-1 (C=O); nmr in CC14 (Figure 36): T 9.06 (s, 3H), 8.93 (s, 3H), 8.78 (s, 3H), 8.64 (s, 3H), 8.30 (d, 3H, a_= 1.5 Hz), 7.51 (d, 1H, g_= 6 Hz), 7.38 (s, 1H), 5.38 (s, 1H), 5.20 (s, 1H), 4.66 (m, 1H); uv max (cyclohexane): 240 nm (a 8370), 220 nm (8 7940), 214 (e 8370); mass spectrum (70 ev) m/e (relative intensity): 232(2), 188(10), 173(100). Anal, Calcd for C15H2002: C, 77.55, H, 8.68. Found: C, 76.82, H, 8.59. 38. Reduction of y-Lactone 113: y-Lactone 113 (0.030 g, 0.129 mmole) in 1 ml of boron trifluoride etherate was added to a well—stirred solution of 0.02 g (0.555 mmole) of sodium borohydride in 10 ml of anhydrous ether, at 0°. The solution was stirred further for 45 min at 0° and then refluxed for 2 hr. The ethereal solution was washed with 10 ml of 5% aqueous sodium bicarbonate solution, dried (Na2804) and concentrated giving a small amount of white solid: ir (cc14): 2960, 1380, 1260, 1100, 1020 and 823 cm-1; mass spectrum (70 ev), m/e (relative intensity): 220(13), 147(100). 39. Attempted Preparation of Adduct 117 of 108 with 2-Butyne: Triene 108 (0.5 9, 0.00308 mole) together with an excess of 2-butyne (0.65 g, 0.94 ml, 0.0154 mole) were placed in a metal bomb at -78°. The two compounds were allowed to react at 0° for 24 hr. At the end of this time the bomb was opened and 2-butyne was allowed to evaporate. The nmr 125 spectrum of the product showed signals at T 8.9 (s, 6H), 8.13 (d, 3H, g_= 1.5 Hz), 5.10 (s, 2H), 4.56 (m, 1H). The reaction was carried out at room temperature for 24 hr and the same results obtained. The reaction was repeated once again at 160° for 8 hr. A black tar was obtained as the product. SPECTRA NMR Spectra The nmr spectra presented here were obtained using CC14 solutions with tetramethylsilane as an internal standard. IR Spectra The ir spectra presented here were also obtained using CCl4 solutions. 126 transmittance " 'n' C, * O O Figure 13. mmmhr Figure 14. 5000 X0 wavenumber _ _ 127 D‘s-- IR spectrum of 2,2-dibromo-4,5,6,6,7,8-hexa— methylspiro[2.5]octa—4,7-diene (ll) 650 ...-...n Ir spectrum of 2,2-dichloro—4,5,6,6,7,8—hexa- methylspiro[2.5]octa-4,7-diene (El). 128 tnumhfl? 6 O \uvonumber Figure 15. IR spectrum of 4,5,6,6,7,8—hexamethylspiro— [2.5]octa-4,7-diene (31). 8 o E, I? .3 ‘. || . an 5000 0000 wavenumber 8 (luminance 6 0 Figure 16. IR spectrum of 2,3—epoxy-2,3,4,4,5,6-hexamethyl- 2,5—cyclohexadienone (94) 129 Figure 17. IR spectrum of 2,3;5,6-diepoxy-2,3,4,4,5,6— hexamethyl—Z,5—cyclohexadienone (99). 8 8 6 8 transmittance 6 O 5000 ‘000 5° wavenumber' .— .. . Figure 18. IR spectrum of 6-acetyl—2,3,4,4,5-pentamethyl— 2-cyclopentenone (102). 130 I” no no so no 70 yo co no so no no 00 33o :0 .5. "no a 0 5000 4000 wavenymber Figure 19. IR spectrum of 1-methylene-3-isoprop l—2,4,5, 6,6-pentamethyl-2,4-cyclohexadiene ( 07). Figure 20. IR spectrum of adduct 94. 131 8‘8 6 transmittance 6 O , 5000 4000 wavenumbegr ... -.. Figure 21. IR spectrum of 1,3,4,5,5,6-hexamethyl-1,3- cyclohexadiene (109). 0 so H H ‘ no can I. ' O ., ,,._._. ”00 4000 wavgnumber ....-. Figure 22. IR spectrum of 1,2,3,3,4,5-hexamethyl-1,4- cyclohexadiene (110). 132 8 8 6 S transmittancé 6 5000 C000 wavenumbe; Figure 23. IR spectrum of adduct 111. Figure 24. Ir spectrum of y-lactone 11 . 133 . Esme mooao . 38 0588.738 E. 3 no.8-ouoo_m.n_ouamuashuoeoxosum.u.o.o.m.v aouaouamnuoeoxosum.o.o.o.o.vuoeoun IouoanoHoIN.N mo Esuuowmm mzz .mN ousmfim IflvIN.N mo Ednuommm mzz .mm munmflm w. ... .m m.) .... .1. ...o on J J 3); 134 .«wwc ooooono umxoaoaomoum.unamzuosmxonno.mtv.o.m.m Imxomolm.w mo Ednuoomm mzz .wN wusmflm 0. 0+0 0% 08 _ _ L p .Ammc mooaons.vuouooHo.miouamoasgoosmxon 1w.u.o.o.m.v mo Esuuowmm mzz .FN ousmam 0. 0.7 p — all}: 3 » p 7.2255?) .ANOHV .dmwv mcocmfiomxmnoaomo mcocmucmmoaoholmuamsumamucmmlm.v.v.m.N In.Nnamzumfimxoaum.m.v.v.m.mn>xomoflp 1H>u00m1® mo Esnuommm mzz .om ousmflm I®.mum.m mo Eduuoomm mzz .mN ousmflm O- 0.7 0.. nwbuhi O. 0.7 0.“ Opfi L b F _ p P 135 )4 33:1... 4 136 // 1 T 1 ‘1”""""" 17"“- ‘ 7" "7 ‘ ‘ v ». ‘ no so 0-0 Figure 31. NMR spectrum of 1-methylene-3-isopropy1- 2,4,5,6,6-pentamethyl-2,4-Cyclohexadiene (107). Figure 32. NMR spectrum of adduct 64” 137 .AOHHV mcmwomxmnoHomolv.Hlahsume Imxonlm.v.m.m.m.fl mo Esnuowmm mzz .mm ousmflm u.-- 138 Figure 34. NMR spectrum of 1,3,4,5,5,6-hexamethyl— 1,3-cyclohexadiene (109). i I \\ 9 59° N l. \./o Figure 35. NMR spectrum of adduct 111. .MHfi ocouomank 139 mo Esuuoomm mzz .mm musmflm . #7 oh 0. use ow u»J L D l _ r _ i. . meLv. (la) SUMMARY 2,2-Dibromo-4,5,6,6,7,8-hexamethylspiro[2.5]octa- 4,7—diene (Zl) was prepared by the addition of dibromo— carbene to lémethylene-2,3,4,4,5,6-hexamethyl-2,5- cyclohexadiene (99). Compound Zl'underwent polymeriza- tion on photolysis. Treatment of Zl'with methyllithium did not yield an allene. Pyrolysis of Zl'in solution gave three isomeric products, d-bromopentamethylstyrene (99),.alafa-bromo-pentamethyl- styrene (9l), and Elana—B-bromo-pentamethylstyrene (99). Pyrolysis of Zl'in deuterated solvents showed very little deuterium incorporation into the products. Based on these results a possible mechanism was proposed for this reaction. Benzoyl peroxide accelerates the decomposition of ll, The rate of decomposition was found to be proportional to the concentration of benzoyl peroxide. Addition of dichlorocarbene to (99) gave 2,2-dichloro- 4,5,6,6,7,8-hexamethylspiro[2.5]octa-4,7-diene (9l). Reduction of Zl'with lithium and liquid ammonia gave 4,5,6,6,7,8-hexamethylspiro[2.5]octa-4,7-diene (91). Pyrolysis of 9Z'in solution produced mainly hexamethyl- benzene. Compound 9Z'was observed to rearrange to hexa— methylbenzene in acidic media. 140 5. 141 Oxidation of 2,3,4,4,5,6-hexamethyl-2,5-cyclohexadi- enone (99) with nfchloroperbenzoic acid gave 2,3-epoxy- 2,3,4,4,5,6—hexamethy1—2,5-cyclohexadienone (92) and 2,3;5,6—diepoxy-2,3,4,4,5,6-hexamethyl-2,5-cyclohexa- dienone (99). Photolysis of a 1% solution of epoxide 92'in ether with 450~watt Hanovia lamp through vycor for 10 hr gave 6-acetyl—2,3,4,4,5-pentamethyl-2-cyclo- pentenone (999). Diepoxide 99 was found to be photo- chemically stable. 1