ABSTRACT THE PHOTOISOMERIZATION AND PHOTO-OXIDATION OF SOME SUBSTITUTED DIENONES by Robert H. Young The purpose of this research was to investigate the effect of substituents on the photolysis of 2,4-cyclohexa- dienones. Substituents used were acetyl and hydroxyl groups. Diacetylfilicinic acid (2,4-diacetyl-3,5-dihydroxy— 6,6-dimethyl-2,4-cyclohexadienone, l) was prepared from 1,3— diacetylphloroglucinol, g“ by treatment with methyl iodide and sodium methoxide in methanol. Phloroglucinol and acetic anhydride were refluxed with_an acid catalyst to give the 1,3-diacetylphloroglucinol. Acetylfilicinic acid (2-acetyl-3,5-dihydroxy—6,6-di- methyl-2,4-cyclohexadienone, g) was obtained by refluxing diacetylfilicinic acid in methanol-sodium methoxide or by allowing the reaction mixture to stand at room temperature for more than two weeks. A number of other related substi- tuted 2,4— and 2A¥cyclohexadienones were also prepared to aid structural assignments. Spectral data including uv spectra at different acid and base concentrations were used to assign definite structures to these compounds. Robert H. Young OH O \\ HO - OH HO OH 2H 2N) 2w Diacetylfilicinic acid, 1/ was inert when irradiated through Pyrex with a Hanovia L 450 watt or a Hanovia S 200 watt lamp in methanol under nitrogen. However, in alkaline methanol under oxygen it did undergo an interesting photo- oxidation and rearrangement to 2-acetyl-3,4-dihydroxy-4— carbomethoxy-5,5-dimethyl-2-cyclopentenone, 2; This com- pound was identified by its spectral properties and its conversion to the known 2-acetyl-3,4-dihydroxy—5,5-dimethyl— 2-cyclopentenone, Q, using 2N sodium hydroxide at 100°. Spectral data confirmed the structure of the latter compound. A mechanism was postulated for the formation of the photo- oxidation product, which involved the reaction of the enolate anion of l'with singlet oxygen as a key step. The photo-oxidation could be sensitized by dyes known to produce singlet oxygen, The same photo-oxidation product, 2, was also obtained thermally, using (singlet) oxygen from the peroxide of 9,10-dipheny1anthracene. Robert H. Young Acetylfilicinic acid, é/ was photo—oxidized in the same way as diacetylfilicinic acid to give not only the same product, 2, but in addition some of the decarboxylated compound, Q“ probably gig a new route. Thus the reactions of enolate anions with singlet oxygen may be general; the study of further examples seems desirable. 4 5 g Acetylfilicinic acid, g” underwent a photolytic rear- rangement in neutral media, unlike diacetylfilicinic acid. The product was a pyrone, 3—acetyl-4-hydroxy-61i-propyl-ZH- pyrone, Q; The formation of the pyrone may be rationalized by photolytic ring opening of g'to a ketene, followed by intramolecular attack by the 5-hydroxyl substituent on the ketene. Some kinetic experiments showed that the photo-oxida- tion was independent of the oxygen and base concentrations, except when the base concentration was low. The chemical rate depended only upon the efficiency with which the excited state was formed followed by reaction with oxygen resulting in singlet oxygen. THE PHOTOISOMERIZATION AND PHOTO-OXIDATION OF SOME SUBSTITUTED DIENONES BY Robert H? Young A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1967 ACKNOWLEDGMENTS The author expresses his sincere appreciation to Professor Harold Hart for his patient direction and help— ful suggestions both in the eXperimental work and in the writing of this thesis. He wishes to thank the National Science Foundation for financial support from April 1966 through June 1967. Appreciation is also extended to Phillips Petroleum Company for a Summer Research Fellowship during the Summer of 1967. ii TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . . . . . . 1. Preparation of Substituted Dienones II. III. IV. A. B. Preparation of Filicinic Acid and Related Derivatives . . . . . . . . . Preparation of Related Compounds from Resorcinol . . . . . . . . . . . . Discussion of the Structures of Filicinic Acid and Related Compounds . . . . . . . . A. The Structure of Filicinic Acid . . . . B. The Structure of Diacetylfilicinic Acid C. The Structure of Methylfilicinic Acid . D. The Structure of Acetylfilicinic Acid . E. The Structure of Acetylmethylfilicinic ACid O O O O O O O O I O O O O O O O O Photolysis and Photo-oxidation of Diacetylfilicinic Acid . . . . . . . A. B. C. D. E. F. Direct Photolysis of Diacetylfilicinic Acid . . . . . . . . . . . . . . . . Photo-oxidation of Diacetylfilicinic Acid . . . . . . . . . . . . . . . . Dye-sensitized Photo-oxidations . . . Mechanism of the Photo-oxidation of Diacetylfilicinic Acid . . . . . . . The Formation of Dimethylmalonic Acid The Methyl Ethers of the Photo-oxidation Products . . . . . . . . . . . . . . The Photolysis and Photo-oxidation of Acetylfilicinic Acid . . . . . . . . . . A. B. Photo-oxidation of Acetylfilicinic Acid Photoisomerization of Acetylfilicinic Acid . . . . . . . . . . . . . . . . Results and Disscussion of Kinetic Experiments . . . . . . . . . . . . . . . iii Page 27 32 34 39 45 45 49 51 51 52 61 66 73 74 79 79 84 '89 TABLE OF CONTENTS (Continued). Page EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . 104 I. General Procedures . . . . . . . . . . . . 104 II. Preparation of Filicinic Acid and Deriva- tives . . . . . . . . . . . . . . . . . . . 105 A. Preparation of 1,3—Diacetylphloro- glucinol . . . . . . . . . . . . . . . 105 1. Using Boron Trifluoride as Catalyst 105 2. Using Concentrated Sulfuric Acid as Catalyst . . . . . . . . . . . . . 106 a. Small Scale Reaction . . . . . 106 b. Large Scale Reaction . . . . . 107 B. Preparation of 2,4—Diacetyl—3,5-di- hydroxy-6,6-dimethyl-2,4-cyclohexadienone (Diacetylfilicinic Acid) . . . . . . . 107 1. Attempted Preparation of N-Phenyl- maleimide Derivative of Diacetyl- filicinic Acid . . . . . . . . . . 109 2. Attempted Acetylation of Diacetyl- filicinic Acid . . . . . . . . . . 109 a. Using Acetyl Chloride . . . . . 109 b. Using Acetic Anhydride . . . . 109 C. Preparation of 3,5-Dihydroxy-4,4- dimethyl-2,5—cyclohexadienone (Fili- cinic Acid) . . . . . . . . . . . . . . 109 D. Preparation of the Diacetates of rFilicinic Acid . . . . . . . . . . . . 110 E. Reaction of Diacetylfilicinic Acid with Diazomethane . . . . . . . . . . . 111 F. Preparation of Acetylphloroglucinol . . 112 1. From Phloroglucinol and Acetonitrile 112 2. From Phloroglucinol and Acetic Anhydride ... . . . . . . . . . . . 112 G. Preparation of 2:Acetyl-3,5-dihydroxy- 4,6,6-trimethyl-2,4-cyclohexadienone (Acetylmethylfilicinic Acid) . . . . . 113 H. Preparation of 2,4,4-Trimethyl-3,5-di- hydroxy-2,5-cyclohexadienone (Methyl- filicinic Acid) . . . . . . .-. . . . . 115 iv TABLE OF CONTENTS (Continued) TIII. Page I. Preparation of 2-Acetyl-3-hydroxy-5- methoxy—4,6,6—trimethy1-2,4-cyclohexa- dienone . . . . . . . . . . . . . . . . 116 J. Preparation of 2:Acetyl—3,5-dihydroxy- 6,6—dimethyl-2,4-cyclohexadienone (Acetylfilicinic Acid) . . . . . . . . 117 1. Standing Eighteen Days at Room Temperature . . . . . . . . . . . . 117 2. From Refluxing Methanol and Sodium Methoxide . . . . . . . . . . . . . 118 K. Preparation of 22Acetyl-3-hydroxy-5- methoxy-6,6-dimethyl-2,4-cyclohexa- dienone (Monomethyl Ether of Acetyl- filicinic Acid) . . . . . . . . . . . . 118 L. Preparation of 2,4- and 4,6—Diacetyl- resorcinols . . . . . . . . . . . . . . 119 M. Preparation of 2,4-Diacetyl-5-hydroxy- 6,6-dimethyl-2,4-cyclohexadienone . . . 120 N. Preparation of 2-Methyl-4,6-diacetyl- resorcinol . . . . . . . . . . . . . . 121 0. Preparation of 2,4-Diacety1-5-hydroxy- 6,6-dimethyl-2,4-cyclohexadienone from 2-Methyl-4,6-diacetylresorcinol . . . . 121 P. Preparation of 4eAcetylresorcinol . . . 122 Q. Preparation of 4eAcety1-5—hydroxy-6,6- dimethyl-Z,4-cyclohexadienone . . . . . 122 1. From 2,4—Diacetyl-5—hydroxy-6,6- dimethy1~2,4-cyclohexadienone . . . 122 2. By Methylation of 42Acety1resorcinol 123 R. Attempted Methylation of 2,4-Diacetyl- resorcinol . . . . . . . . . . . . . . 124 Photolyses and Photo-oxidations . . . . . . 124 A. The Photolysis and Photo-oxidation of 2,4-Diacety1—3,5-dihydroxy-6,6-dimethyl- 2,4-cyclohexadienone (Diacetylfilicinic Acid) . . . . . . . . . . . . . . . . . 124 1. The Photolysis of Diacetylfilicinic Acid in Ethyl Ether . . . . . . . . 124 V n 16 TABLE OF CONTENTS (Continued) 2. Photolysis of Diacetylfilicinic Acid in Methanol under a Nitrogen Atmosphere . . . . . . . . . . a. Small Scale . . . . . . . . b. Large Scale . . . . . . . . 3. Photolysis of Diacetylfilicinic Acid in Methanol with an Oxygen Atmosphere . . . . . . . . 4. Photolysis of Diacetylfilicinic Acid in Alkaline Methanol, Using a Nitrogen Atmosphere . . . . . . 5. Photolysis of Diacetylfilicinic Acid in Alkaline MethanOl With an Air AtmOSphere . . . . . . . . . . . 6. Photolysis of Diacetylfilicinic Acid . in Alkaline EthanOl with an Air AtmOSphere - ... . . . ... . . . 7. Photolysis of Diacetylfilicinic Acid in Water with an Air Atmosphere . 8. Dark Reaction of Diacetylfilicinic Acid in Alkaline Methanol in an Air Atmosphere . . . . . . . . . . ... Preparation of ZeAcetyl-3, 4-dihydroxy- 5, 5-dimethyl-2-Cyclopentenone . . . With 2N Sodium Hydroxide . . . . . With 2N Hydrochloric Acid . . . . . With Sodium Methoxide . . . . . From 2—Acety1-3, 4-dihydroxy-4-carbo- ethoxy-5, 5-dimethyl-2-cyclopentenone and 2N Sodium Hydroxide . . . . . . QWNH Preparation of 22Acetyl-3-hydroxy-4- keto-5,5-dimethyl-2-cyclopentenone 1. Oxidation of 2—Acetyl-3,4-dihydroxy- 5, 5-dimethyl-2-cyC10pentenone with Manganese Dioxide . . . . . 2. Oxidation of 2eAcetyl-3, 4-dihydroxy- 5, 5-dimethy1-2-cyc1opentenone with Bismuth Oxide . . . . . . . . . . Dye-sensitized Photo-oxidations . . . 1. Standard Reaction Without Dye a. In a Pyrex Photolysis Well b. In a 250-ml 3-Necked Flask vi Page 125 125 125 126 126 126 127 128 129 129 129 130 130 131 131 131 132 132 132 132 133 TABLE OF CONTENTS (Continued) 2. Dye-sensitized Photo-oxidations of Diacatylfilicinic Acid. . . . . . . a. With Fluorescein . . . . . b. With Methylene Blue . . . . . . C. With Rose Bengal . . . . . . . E. Reaction of Diacetylfilicinic Acid with Singlet Oxygen Prepared from 9,10-Di- phenylanthracene Peroxide . . . . . 1. Standard Experiment . . . . . 2. Preparation of ZeAcetyl- -3, 4-di— hydroxy-4-carbomethoxy- 5, 5-dimethyl- 2-Cyclopentenone from the Peroxide f 9,10-Dipheny1anthracene . . . . F. Preparation of Methyl Ethers of the Photo-oxidation Products . . . . . 1. Reaction of leAcetyl-3,4-dihydroxy- 4-carbomethoxy-5, 5-dimethyl-2-cyclo- pentenone with Diazomethane . 2. Reaction of 2-Acetyl-3, 4-dihydroxy- 4-carboethoxy-5, 5-dimethyl-2- cyclo- pentenone with Diazomethane . . . . G. Reaction of ZeAcety1—3,4-dihydroxy-5,5- dimethy1-2-cyclopentenone with Sodium Borohydride . . . . . . . . . . . . . H. Photolysis and Photo-oxidation of 2- Acetyl-3,5-dihydroxy-6,6-dimethy1-2,4- cyclohexadienone (Acetylfilicinic Acid) 1. Photo-oxidation of Acetylfilicinic Acid in Alkaline Methanol with an Air Atmosphere . . . . . . . 2. Photolysis of Acetylfilicinic Acid with One Equivalent of Sodium Meth- oxide and an Air Atmosphere . . 3. Photolysis of Acetylfilicinic Acid. a. In Ethyl Ether Under a Nitrogen Atmosphere . . . . . . . . . b. In Methanol under a Nitrogen Atmosphere . . . . . . . . . . I. Photolysis of 22Acetyl-3-hydroxy- -5- methoxy-6, 6-dimethyl-2, 4-cyclohexa- dienone . . . . . . . . . . . . 1. In Ethyl Ether. . . . . 2. In Alkaline Methanol with an Air Atmosphere . . . . . . . . . vii Page 133 133 134 134 135 135 136 137 137 138 139 139 139 140 141 141 142 143 143 143 TABLE OF CONTENTS (Continued) IV. Page J. Miscellaneous Photolysis Reactions . 143 1. Attempted Photolysis of 2-Acetyl- 4-hydroxy-61i-propyl-ZH-pyrone . 143 2. Attempted Photolysis of 2—Acety1- 3,4-dihydroxy-5,5-dimethyl-2— cyclOpentenone . . . . . . . . . . 144 a. In Methanol . . . . . . 144 b. In Alkaline Methanol Under an Air Atmosphere . . . . . . . . 144 c. In Alkaline Water Under an Air Atmosphere . . . . . . . . . . 144 3. Photolysis of 2-Acety1-3,4—dihydroxy- 4-Carbomethoxy-5,5-dimethyl-2- cyclOpentenone . . . . . . . . . . 144 Kinetic Reactions on the Photo-oxidation of Diacetylfilicinic Acid and Acetylfilicinic Acid . . . . . . . . . . . . . . . . . . . 145 A. Series I . . . . . . . . . . . . . . . 145 1. Photo-oxidation with Sodium Methoxide (0.0093 mole) . . . . . . 145 2 Photo-oxidation with Sodium Methoxide (0.037 mole) . . . . . . 146 3. Photo-oxidation with Sodium Methoxide (0.074 mole) . . . . . . 146 4. Photo-oxidation with Sodium Meth- oxide (0.0093 mole) Under Oxygen . 146 5 Photo—oxidation with Sodium Meth- oxide (0.0093 mole); Solution Sat- urated with Air Before Irradiation. 146 6. Photo-oxidation with Sodium Meth— oxide (0.0093 mole) and Diacetyl- filicinic Acid (0.000497 mole) . . 147 7. Photo-oxidation with Sodium Meth- oxide (0.0093 mole) and Acetyl- filicinic Acid . . . . . . . . . . 147 B. Series II . . . . . . . . . . . . . . . 148 1. Standard Photo-oxidation with a Hanovia L 450 Watt Lamp . . . . . . 148 2 Standard Photo-oxidation with a 200 Watt Tungsten Lamp . . . . . . 149 3. Dye- -sensitized (Rose Bengal) Photo- oxidation . . . . . . 149 4 Photo-oxidation of 2, 5-Dimethyl- furan . . . . . . . . . . . . . . . 149 viii TABLE OF CONTENTS (Continued) SUMMARY LITERATURE CITED APPENDIX I. APPENDIX II. A. B. C. 10. 11. 12. 13. 14. Photo-oxidation with 2,5-Dimethy1- furan as a Singlet Oxygen Trapping Agent . Photo- oxidation with 2, 5-Dimethy1- furan as a Singlet Oxygen Trapping Agent in a Dye-sensitized Reaction. Photo—oxidation with 2,5-Dimethyl— furan (0.0104 mole) as a Singlet Oxygen Trapping Agent, in a Dye- sensitized Reaction under Oxygen Photo-oxidation with Sodium Meth- oxide (0. 037 mole) in a Dye-sensi— tized Reaction . . . . . . . Photo- oxidation with 2, 5-Dimethy1— furan (0.104 mole) as a Singlet Oxygen Trapping Agent . . . . Photo- oxidation‘with 2, 5-Dimethy1— furan (0.104 mole) as a Singlet Oxygen Trapping Agent in a Dye- sensitized Reaction . . . . . . Photo—oxidation of Acetylfilicinic. Acid . . . . . . . . . . . . . Photo- oxidation of Acetylfilicinic Acid with a Dye-sensitizer .. . . Photo-oxidation of Acetylfilicinic Acid with One Molar Equivalent of Sodium Methoxide Present . . . Photo- oxidation of Acetylfilicinic Acid with One Molar Equivalent of Sodium Methoxide in a Dye-sensi- tized Reaction . . . . . . . Ir and Nmr Spectra Molecular Orbital Calculations (w-Technique) . . . . . . The Method .. . . . . . . . . . . . The Effect of Substituents on the Pho- tolysis of 2,4-Cyclohexadienones . The Relative Stability of Tautomeric Structures as Predicted by Molecular Orbital Calculations . . . . . . . ix Page 150 150 151 152 152 153 153 154 154 155 156 159 162 172 173 179 192 Table II. III. IV. VI. VII. VIII. XI. XII. XIII. LIST OF TABLES Page Nuclear magnetic resonance Spectra of filicinic acid and related compounds . . . 15 Nuclear magnetic resonance of spectra of acetylfilicinic acid and related compounds. 23 Nulcear magnetic resonance spectra of photo- oxidation products and related compounds . 54 Parameters for the oxygen heteroatom in molecular orbital calculations . . . . . . 68 Charge densities of the anion of diacetyl- filicinic acid . . . . . . . . . . . . . . 68 Charge densities of the anion of acetyl- filicinic acid . . . . . . . . . . . . . . 81 Series I reaction rates . . . . . . . . . 90 Series II reaction rates using diacetyl- filicinic acid . . . . . . . . . . . . . . 94 Series II reaction rates using acetyl- filicinic acid . . . . . . . . . . . . . . 99 Parameters for molecular orbital Calculations of 2,4-Cyclohexadienones . . . . . . . . . 182 Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of 2,4-Cyclohexa- dienone (w-technique) . . . . . . . . . . . 180 Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of acetylfiIiCinic' acid (w-technique) . . . . . . . . . . - . 185 Coefficients of the highest occupied and lowest unoccupied molecular orbitals and Charges on ring atoms of diacetyl- filicinic acid (w—technique) . . . . . . . 186 LIST-OF Table XIV. XV. -XVI. XVII. TABLES (Continued) Page Coefficients of the highest occupied and lowest unoccupied molecular orbitals and Charges on ring atoms of monoanion of acetyl- filicinic acid (w - technique) . . . . . . 189 Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of the monoanion of diacetylfilicinic acid (w - technique) . . 189 (Molecular orbital calculation results on tautomeric structures of substitued dienones 193 Molecular orbital calculation results on standard compounds containing oxygen atoms. 196 xi Figure II. III. IV. VI. VII. VIII. X. XI. XII. XIII. XIV. LIST OF FIGURES Page Outline of products synthesized from phloroglucinol . . . . . . . . . ... . . . 10 Outline of products synthesized from resorcinol . . . . . . . . . . . . . . . . 12 Ultraviolet absorption spectra of filicinic acid '; 2'. . . u L . . . . . . . . . . . . . 35 Ultraviolet absorption spectra of 3-hydroxy— 4,4,6,6—tetramethyl-2—cyclohexen-1,5-dione. 36 Ultraviolet absorption Spectra of diacetyl- filicinic acid . . . . . . . . . . . . . . 40 Ultraviolet absorption spectra of 2-acetyl- 3-hydroxy-4,4,6,6—tetramethyl-2-cyclohexen- 1,5-dione . . . . . . . . . . . . . . . . . 41 Ultraviolet absorption spectra of 2,4-di- acetyl-5-hydroxy-6,6-dimethyl-2,4-Cyclo- hexadienone . . . . . . . . . . . . . . . . 42 Ultraviolet absorption SpeCtra of methyl- filicinic acid . . . . . . . . . . . . . . 46 Ultraviolet absorption spectra of acetyl- methylfilicinic acid . . . . . . . . . . . 50 Graphic results of kinetics; Serieer . . . 91 Graphic results of kinetics; Series II, with Hanovia L 450 watt lamp . . . . . . . 93 Graphic results of kinetics; Series II, dye-sensitized photo-oxidations . . . . . . 96 Graphic results of kinetics; Series II, acetylfilicinic acid . . . . . . . . . . . 97 Plot of ultraviolet absorption of 3-hexene- 2,5-dione during the photo-oxidation of dimethylfuran . . . . . . . . . . . . . . . 98 xii LIST OF.FIGURES (Continued). Figure XV. XVI . Page Plot of absorption wavelength gs differ- ence in energy of highest occupied and lowest unoccupied molecular orbitals of substituted cyclohexadienones . . . . . . . 198 Plot of absorption wavelength XE difference in energy of highest occupied and lowest unoccupied molecular orbitals of standard compounds containing oxygen . . . . . . . . 199 xiii LIST OF S PECTRA Infrared Spectra: Page 1. ZeAcetyl-S, 4—dihydroxy-4-Carbomethoxy-5, 5- dimethyl-Z-cyc10pentenone . . . . . . . . . 163 2. 2-Acetyl-3, 4-dihydroxy-5, 5-dimethyl- -2-. cyclop entenone . .. .,. . . . . . . . . . . 163 3. 2-Acetyl—3, 4-dihydroxy-4—caflaoethOXy-Z-Cyclo- pentenone . . . . . . . . . . . . . . . . 164 4. Unknown Compound, 82' . . . . . . . . . . . . 164 5a. 32Acetyl-4-hydroxy4611fpropyl-2H-pyrone in KBr 165 5b. 3eAcetyl-4-hydroxy-6-i-pr0pyl-2H-pyrone in CCl4 . . . . . . . . . . . . . . . . . 165 6a. Dehydroacetic acid in.KBr . . . . . . . . . 166 6b. Dehydroacetic acid in CC14 . . . . . . . . . 166 Nuclear Magnetic Resonance Spectra: 1. 22Acetyl-3,4-dihydroxy-4-carbomethoxy-5,5- dimethyl-Z—cyclOpentenone . . . . . . . . . .. 167 2. 2-Acetyl-3,4-dihydroxy-4-carboethoxy-5,5- dimethyl-2-Cyclopentenone . . . . . . . . . . 168 3. 2~Acetyl-3, 4-dihydroxy-5, 5-dimethyl-2— cyclOpentenone . . . . . . . . . . . . . 169 4. Unknown Compound, 82' . . . . . . . . . . . . 170 5. 3-Acetyl-4—hydroxy46fii-propyl—ZH-pyrone . . . 171 xiv INTRODUCTION Until recently the normal photochemical reaction for 2,4—Cyclohexadienones in solvents containing a nucleophile was a ring opening to give esters, amides, or acids depend- ing upon whether the nucleophile was an alcohol, amine or water (1,2,3). The photochemical product is known to be a ketene, which then reacts thermally with the nucleophile. An example is outlined in the scheme: 0 O " R1 "R1 R1 0 / . \ n .RZJV—_{> / R2___‘}£__{> / c=CH-CH=CH—CH2 -C-X \ . R2 2H X: = nucleophile a. R1 = R2 = CH3 b. R1 = R2 = OAC c. R1 = CH3 ; R2 = OAC Alternate modes of rearrangement include expulsion of a heteroatom from the C-6 position or migration of a substit— uent from C-6 to C-5 to give a phenol (3). Recently Hart gt_§l. (4) have shown that a fourth alternative is possible. Hexamethyl-2,4-cyclohexadienone, 2, rearranged in the presence of ultraviolet light to a bicyclic isomer, 3” In a later paper (5) they discussed the mechanism which was elucidated through labeling ex- periments; no migration of methyl groups occurred. (N 203 3 Hart and Collins (6) prepared and photolyzed a series of methyl-substituted 2,4-Cyclohexadienones, and found that the reaction course depends critically on the number and positions of the substituents. A mechanism involving two separate routes with two different types of transition states to account for the two products was favored. They suggested that additional work is necessary to determine the nature of the excited state for each reaction. R2 II R3 0—D R5 CH30H E, (R2. 2, (R: E, (R2 E. (R2 2. (R2 CH3 H H O o . hV \ - I u R n (R5 /C=C - C = c - C-C--OCH3 + 2 n I I CH3 R5 CH3 R3 R2 ’ R3 2 2 =R3=R5-CH3) 11)) > 2' =R3—CH3,R5=H) 1” > g = H ; R3 = R5 = CH3)1—hx——> g. =R5—H:R3=cns> h" > g hv =R5=CH37R3=H)—-—> §+g The photoisomerizations of 2,5-cyclohexadienones have been studied more intensively than those of their 2,4-iso- mers. Chapman first prOposed dipolar excited states (7), and Fisch applied these to explain the photorearrangement of 2,5-cyclohexadienones to bicyclic compounds (8). Zim- merman later proposed a more plausible diradical excited state (9). Zimmerman's mechanism involved four steps: 4 (1) excitation, (2) rebonding, (3) demotion of the v* electron, and (4) rearrangement of the resulting Charge- separated species. This scheme is outlined below. 0 0° 0° ll 1. hV 0 ——>"* W a— a Exc1tat1on 1ng V C6H5 C6H5 C6H5 C6H5 C6H5 CsHs 3.Demotion of 7* electron V7 0 0" n C6H5 4.rear-[¢é\\+‘ I Grangaznt 8H5_ C6H5 C635 More recent experimental evidence favoring this mechanism has been advanced by Schuster and shows that radicals must be involved in the excited state in order to get the frag- mentation products in the photolysis of 12 (10). Frag- mentation must occur before the electron demotion step in the previously mentioned scheme. He also showed the import— ance of this mechanism in the photolysis of 11'(11). Photol- ysis Of 11 carried out in ether gave lz'by radical frag- mentation followed by hydrogen abstraction. Photolysis in benzene gave 12, OCH2CH3 9 OH CH3CHCH2CH2 hv Ether Jl> + CHZCHs OH 12. OH .CH3CHCHOCH2CH3 CH3 0 OH O CC13 12 13 6 Some work has also been carried out on the mechanistic interpretation of the photolyses of 2,5-cyclohexadienones by the use of substituent groups in other than the 4 posi- tion of the molecule. The reaction as outlined here was taken from Chapman's, Organic Photochemistry (3). When both R2 and R4 are hydrogens, then both lz'and lg'are formed. When R2 is electron donating (R2 = CH3 ; R4 = H), then the intermediate 12.15 favored and the product is 11; When R4 2 0H2 _-> 22. 0’. H3O R4 £2, '°H’ R2 p. —> 3 R4 HO is electron donating (R4 = CH3 ; R2 = H) then 18 is formed. ’w However, when R2 is electron withdrawing (R2 = CHO or COOH R4 = H) then 12'15 unfavorable as an intermediate and pro- ducts are formed from intermediate 12; I 7 Both of these intermediates could result from electron demotion after the initial photochemical reaction and not necessarily have any effect on the excited state of the molecule or the bond crossing mechanism. Except for the work on the different methyl-substituted 2,4-cyclohexadienones (6) little work has been done on the effects of different substituents in these compounds. For this reason 2,4-cyclohexadienones subStituted by electron— donating and/or electron-withdrawing groups were irradiated. Filicinic acid, which has been represented in the literature as 12'(12,13,14) and as 22'(13,15), was chosen as an initial example. Not only can the substituent effect of the hydroxyl groups be studied, but also the substance has the interesting option of behaving as a 2,4— or 2,5-Cyclohexa- dienone. Due to the initial difficulties in acquiring a quantity of filicinic acid and because its structure was found to be a 2,5-Cyclohexadienone, work was concentrated on the photolysis of 2,4-diacetyld3,5-dihydroxy-6,6-di- methyl-2,4-cyclohexadienone (diacetylfilicinic acid, 21) a readily available precursor in the synthesis of filicinic acid. 8 Although this particular compound, 21, did not undergo any direct photolysis reaction for reasons which will be discussed later, it did experience a novel and interesting photo—oxidation. The preparation and structural determina- tion of filicinic acid, 12, and its derivatives will be presented and discussed, as will those of a number of re— lated compounds. New photoisomerizations and photo-oxida- tions of two of these compounds, 2,4-diacetyl—3,5-dihydroxy- 6,6-dimethyl-2,4-Cyclohexadienone (diacetylfilicinic aCid, 21) and 2-acety1-3,5—dihydroxy-6,6-dimethyl-2,4-cyclohexa- dienone (acetylfilicinic acid, 22), were discovered and studied in detail. The mechanism and general implications 0 E? O; " ,2}, R = CHa-C- HO OH 22,: R = H R of these photoreactions will be discussed with the aid of kinetic results. RESULTS AND DISCUSSION I. Preparation of Substituted Dienones A. Preparation of Filicinic Acid and Related Derivatives 3,5-Dihydroxy-4,4-dimethy1-2,5-cyclohexadienone (fili- cinic acid, 12) and its diacetyl precursor, 2,4-diacetyl— 3,5-dihydroxy-6,6-dimethyl-2,4—cyclohexadienone (diacetyl— filicinic acid, 21) were prepared according to the method used by Riedl (12). Although definite structures are given HO OH 19 21 for these compounds at this time, other tautomers are pos- sible. Because of the interrelation of the structures of a number of these filicinic acids and related products, (Fig. I and Fig. II) proof of structures will be delayed until all the spectral evidence has been presented. Dry phloroglucinol, 22, in acetic acid was treated with boron trifluoride for about one hour. An 81.5% yield of 1,3-diacetylphloroglucinol, 22” was Obtained. This compound was identified by its mp of 167.5-169o after recrystallization and sublimation (lit., 168° (12)). Its 9 10 Figure I. Outline of products synthesized from phloro- glucinol, 22, OH O O o‘ CHzNa O’ _ ————{> HO OH HO OH HO OCH3 .253, 2.2., §§. Ac20(2 eq) NaOCHa H+ 18 days OH O o O i 0 6 CH 1 1 ° 2N HCl _3_:[> HO OH CHsONa HO OH HO OH “O 21%. 2.1.. 19. ACZO O I 0 + 9 9 /"\O OH A0 O)\ 11 Figure I. (Cont.) :OH O, O ' ll ACZO o 0 <1—— 22 £9. 5‘51 Aczo (1 eq) CH3I 2N HC1 H A CH30Na 0H ll O (::]| CH I HO O CH30Na OH HO 0 H ~‘i CHzNz 2N HC1 A ’ 9 o , HO CH3 12 Figure II. Outline of products synthesized from resorcinol. O OCH3 13 ir spectrum had bands at 1610 cm"1 for a conjugated intra- molecularly hydrogen-bonded (Chelated) carbonyl group, at 3000 cm"1 for chelated hydroxyl groups, and at 3450 cm-1 0 OH for the non-chelated phenolic hydroxyl group. This prepara- tion was found useful only for small scale reactions. The temperature required for the Fries rearrangement of the di- acetoxy; intermediate, 22“ to 22'was unObtainable for larger scale reactions with this method. A preparation related to that used by Israelstram (16) was found not only more suitable for larger scale reactions but also quicker and gave almost as good yields (61.5% yield on a scale four times as large as that used in the boron trifluoride reaction). In this method oven dried (110°) phloroglucinol was suspended in twice its molar equivalent of acetic anhydride. After the addition of sulfuric acid the phloroglucinol dissolved and the solution was refluxed at 135° for a few minutes. Work up afforded a compound which was identical to 24’by its mp after sublimation. Diacetylfilicinic acid, 21, was prepared from 22' by methylating with methyl iodide and sodium methoxide 14 in anhydrous methanol. It was important to use dry sodium methoxide and anhydrous methanol in order to obtain the product. Diacetylfilicinic acid, 21” was identified by its ir spectrum which had a band at 1560 cm-1 assigned to the conjugated, chelated carbonyls of the acetyl groups, and a band at 1665 cm_1 assigned to the conjugated carbonyl group of the ring. The nmr spectrum, given in Table I, indicates the presence of two acetyl methyl groups (T 7.47 and 7.37), a gem dimethyl group (T 8.59) and two chelated hydroxyl hydrogens (1 —8.45 and -8.85). The nmr spectrum also indi— cated the presence of about 15% of a tautomer (with methyl peaks at T 8.46 and 7.38; the relative amounts were deter- mined by integration of the peak observed for the gem di- methyl group at T 8.46 compared to T 8.59 for 21). Filicinic acid, 12) was prepared by refluxing diacetyl- filicinic acid, 21, with 2N hydrochloric acid, whereupon a good yield (67%) of filicinic acid was obtained. The compound was identified by its mp of 215-217° (sublimed 220-2210, lit., 212-2150 (12)). The ir spectrum had bands at 1630 cm”1 assigned to a conjugated carbonyl group and 1 for the intermolecularly hydrogen-bonded at 2500-3500 cm" enol hydroxyl groups. The nmr spectrum (Table I) in sodium deuteroxide and deuterium oxide had only one peak at T 8.83, the other hydrogens being exchanged very rapidly with deuterium. 15 mv.b oe.o Hm.e wo.m v3O mm oo.m ne.e mo.w «Sou .mm 0 oo.v Aeov 2;. do _ _om va.v me.s mo.m ooouood om m m0 om . Own .€( ov.e eo.w Aeovomzo nm.m- «a.» .mm n¢.wn S¢.S on.w .HOO Hmnuo .m Hmcw> mOI nmUOI OHUIMmD mmU\\ome ucm>aom UCSOQEOU mogsomeoo Umumawu Cam Cflum UHCHUHHHM mo *muuuwmm monocommu oaumcmmfi HMTHUSZ .H Canoe 16 .Omuoc mmH3Hm£uo woman: mumamcflm mum mxmmm HH¢* Ono .ca mo om Hagen: Scum nm.m ow.m oomz om _ _ .- o no.» p). no.4 me.s oo.w vSo do .<< on - m oe.m oonz mm ,, ,uo \\ o on.m )3. ev.e oo.m 1.So an Amends Scum om.m on.» mo.m Aeovomzo .mw Hmfluo 3 4159.2, $0! «$00! OHUlmmU nmU\Um/E.U UCO>HOm CCSOQEOU A.ucoov .H magma 17 This compound was further identified by preparing both the monoacetate and diacetate, by treatment of 12'with acetic anhydride and pyridine. The two acetates were separ- ated by recrystallization from carbon tetrachloride, from which the monoacetate precipitated (mp 142-144°). The ir spectrum was consistent with structure 22. It had bands at 1660 cm.1 (conjugated carbonyl group) and at 1775 cm-1 (acetyl carbonyl). The nmr spectrum is given in Table I and shows the presence of one acetyl methyl group at r 7.72, a gem dimethyl group at T 8.62 and two vinyl hydrogens at 0 9. 0 o o 0 9 HO 0/"\ /\O O/\ 252. 2.1 T 4.06 and 4.54. The residue from the carbon tetrachloride mother liquor contained largely the diacetate of filicinic acid, 21, This compound was recrystallized from a mixture of ethyl ether-petroleum ether and gave a mp of 85° (lit. 82-85° (17)). The structure was confirmed by its ir spec- trum which had absorption bands at 1660 cm"1 for a conjugated carbonyl group and 1775 cm—1 for the ester carbonyl groups. The nmr Spectrum (Table I) indicates the presence of two equivalent acetate methyl groups at T 7.75, a gem dimethyl group at T 8.68, and only one peak at T 3.99 for the two equivalent vinyl hydrogens. 18 An attempt to identify diacetylfilicinic acid, 21, in a similar manner by formation of its diacetate was un- successful. Attempts to prepare the methyl ether(s) by treatment of 21’with diazOmethane yielded a crude reaction product whose nmr spectrum (Table I) showed the presence of a methyl ether (peak at T 6.19). On work up, this com- pound, 22, hydrolyzed back to the more stable chelated starting material, 21, 28 (w In order to learn more abOut the structures of the previously mentioned compounds other similar compounds were prepared for use as spectroscopic models. Two selected were 2-acetyl—3,5-dihydroxy-4,6,6-trimethyl-2,4-cyclohexa- dienone (acetylmethylfilicinic acid, 22) and 2-methyl-3,5-di- hydroxy-Z,5-cyclohexadienone (methylfilicinic acid, 22). The precursor to 22, acetylphloroglucinol, 21, was prepared in two ways. The first was similar to the method developed by Robinson (18). Phloroglucinol was treated with acetonitrile using hydrogen chloride gas and zinc chloride as catalysts. This yielded (81.6%) the 19 acetophloroglucinol with a mp of 215-218° (lit. 218-219° (18))- ( 9 Q‘ OH H O l 0 9 HO /’ OH HO OH HO OH 29 30 31 A second method involved treatment of phloroglucinol with one molar equivalent of acetic anhydride under reflux for 2 min with concentrated sulfuric acid as a catalyst. Although the yield was not as good (41%) the method proved faster and more feasible. The ir spectrum of 21'had ab- sorption bands at 1610 cm.1 for the chelated hydroxyl group, and 3550 cm.1 for the unassociated hydroxyl groups. Acetylmethylfilicinic acid, 22, was prepared from.21’ by treatment with methyl iodide and sodium methoxide in methanol at room temperature using the method of Riedl (15). The product was identified by its mp of 157° (lit. 160- 161° (15)) and its ir spectrum which had bands at 1600 cm-1 for a conjugated carbonyl and 3300 cm"1 (broad) caused by an intermolecularly associated enolic hydroxyl group. The nmr spectrum (Table I) showed the presence of a gem dimethyl group with a peak at T 8.68, and acetyl methyl group at T 7.50, and a ring methyl group at T 8.20. Since the nmr Spectrum was carried out in dimethyl sulfoxide-d6 no hy- droxyl hydrogen signals were observed. 20 Riedl also isolated 2-acetyl-3—hydroxy-4,4,6,6-tetra- methyl-2—cyCIohexene-1,5—dione, 22, from the benzene soluble portion and this fact was verified. Compound 22 was identified by its ir Spectrum which had bands at 1560 cm”1 (conjugated carbonyl group) and at 1710 cm"1 (nonconjugated carbonyl group). The nmr Spectrum of 22,(Table I) indicated the presence of two different gem dimethyl groups at T 8.69 and 8.59. A peak at T 7.44 for one acetyl methyl group was also found. To further prove the identity of 22, the compound was refluxed in 2N hydrochloric acid for five hours. 3-Hydroxy- 4,4,6,6-tetramethyl-2-Cyclohexen-1,5-dione, 22, was obtained. This compound was identified by its mp of 192;193o (lit. 187-190° (15)) and its ir spectrum which had bands at 1610 cm"1 for the conjugated carbonyl group, 1700 cm.1 for the non—conjugated carbonyl group and 3000 cm.1 for the enolic hydroxyl group. The nmr spectrum in sodium deuteroxide and deuterium oxide (Table I) showed only one peak for the two gem dimethyl groups, as expected for the corresponding sym- metrical anion. 21 The acetate derivative, 22, (mp 54-57°) was formed from 22'and its structure confirmed by its ir spectrum which had absorption bands at 1660 cm.1 for the conjugated carbonyl group, 1710 cm-1 for the non-conjugated carbonyl group and 1770 cm.1 for the carbonyl of the ester group. The nmr spectrum of 22'(Table I) had peaks indicating the presence of two different gem dimethyl groups at T 8.69 and T 8.65, one acetate methyl group at T 7.73 and a vinyl hydrogen at T 4.02. From the original ether layer in the preparation of acetylmethylfilicinic acid, 22, methylacetylphloroglucinol, 32, was Obtained. The crude material had a mp of 185-205° (lit. 211° (15)). This compound was further treated with methyl iodide and sodium methoxide to give an additional quantity of 22, Methylfilicinic acid, 22, was prepared from 22'by refluxing in 2N hydrochloric acid for five hours. 'A yield of 64% was obtained. The cOmpound had a mp of 174-1770 (lit. 180° (15)). The ir spectrum of 22'had absorption bands at 1650 cm-1 (conjugated carbonyl group) and 3100 cm- 1 22 (broad); (intermolecular hydrogen-bonded hydroxyl groups). The nmr spectrum (Table I) in deuterium oxide and sodium deuteroxide had peaks at r 8.80 indicating the presence of a gem dimethyl group and T 8.35 for a methyl on a double bond. A diacetate of methylfilicinic acid, 36/ was prepared from 32,and acetic anhydride using concentrated sulfuric acid as a catalyst. The nmr spectrum (Table II) had peaks at T 8.77 showing the presence of a gem dimethyl group, 1 7.77 for two acetate methyl groups, T 8.42 for the methyl on the double bond and T 3.94 for the vinyl hydrogen. The monomethyl ether of acetylmethylfilicinic;aéid, §1, was prepared from acetylmethylfilicinic acid, 22“ using diazomethane. The ir spectrum of gz'indicated two conjugated carbonyls with bands at 1635 and 1655 cm-1. The nmr spec- trum (Table II) had peaks at T 8.67 assigned to the hydro- gens of the gem dimethyl group, T 8.08 for thq:methyl group on the double bond, 17.43 for the acetyl methyl and 1 6.08 for the.methoxyl group. The uv spectrum (and structure proof, i.e., position of the double bonds) is related to 23 HN.N m )3. oaumaoua Ammo m nbvwv.m om.vn om.s «Hon mm . I , mm.m mm.mu av.» Ammo m n no mw.m m Hmcfl> mm.v m.m- 5H.o sw.s o>.m .Hou .mm Aocv 2). m “V \om m Hmafl> av.v m.su on.» mm.m mcoumom an _ \o o 22 «moo \fom Hanna: mafia mo.m mo.m av.» sm.w «H00 an ,, . .o ,o. M. 0&0- ofi assumz.mcflm «v.m su.s s>.m VSo on \tw at m Hmca> am.” is Q 350 mo: «moo: Ouownmo n \on/mo ucmfiom. vasomsoo muuowmm QUCMGOmmH Uwumcmmfi Ammaosz .HH magma 24 mc.v m ofiumsou< m>.N so.v mm.mu 2;. uw.m cam mm.v oa.H om>ummno uoz mcoumua ma bN.v om.e >3. m Hmcfl> um.H om.s mo.w .Hoo av assumz mafia mo.m .cc m UfiumEon< wo.m o.mn wv.b vHUU NV uwnuo mo: m"moon ouounmo n \xow, ucm>Hom, vasomsoo _ .. . mo mo A.u:oov .HH magma Hmnumz.mcfim oo.w m on.m . sm.m y). oflumsoum Ammo m.m u no so.m mm.m- mH.o am.» «How we mm.m «moo assumz.msam mu.» m mm.m . ao.m .zc % 0.3982 8. a E. m- «w. a 3 Ammo m u so ma.¢ sm.m as w «Hoo .w« 3.. N :2 b . m Hmcfl> Ammo oH u a. m~.v mv. v H930 mo: . «moo: ouwl :8 «mo \ 0W6 Hawk/How 65:09:00. A.u:oov .HH magma 26 that of the methyl ether of acetylfilicinic acid, 38” which will be discussed later. 22Acetyl-3,5-dihydroxy-6,6-dimethyl-2,4-cyclohexa- dienone (acetylfilicinic acid, 22) was prepared by treat- ing diacetylfilicinic acid, 21” with sodium methoxide in methanol at reflux for eighteen hours or at room tempera- ture for eighteen days. The product was identified by its mp of 17o-172.5° (lit. 174-1760 (19) and 177-1780 (20)). The ir spectrum had absorption bands at 1540-1580 cm-1 for the conjugated chelated carbonyl group, 1630 cm-1 for the cyclohexadienone carbonyl and 2600-3000 cm-1 for the intermolecular hydrogen-bonded enol hydroxyl group. The nmr spectrum (Table II) had peaks at T 8.62 for the gem dimethyl, 1 7.50 for the acetyl methyl, T 4.49 for the vinyl hydrogen and T -7.6 for the chelated hydroxyl group. The hydrogen of the other hydroxyl was not located. The uv spectrum and further structure proof will be presented later. The monomethyl ether of acetylfilicinic acid, 38” was prepared from acetylfilicinic acid, 22, and diazomethane. It was identified by its uv spectrum (see later discussion) and its ir spectrum which had absorption bands at 1620 and 1660 cm.1 for the conjugated carbonyl groups. The large band (2600-3000 cm-l) for the intermolecularly hydrogen— bonded hydroxyl groups that was characteristic of Eggwas not present. The nmr spectrum (Table II) had peaks at 27 o o / ll , ' O I . CH2N2 O I 0 -——-> 0 HO OH HO OCH3 22, 38 T 8.70 for the gem dimethyl, T 7.47 for the acetyl methyl, T 6.17 for the;methoxyl methyl, T 4.65 for the vinyl hydro- gen and 1 -8.6 for the chelated hydroxyl hydrogen. See Fig. I for an outline of the products synthesized from phloroglucinol, 23' B.. Preparation of-Related Compounds from Resorcinol The reaction of resorcinol with acetic anhydride and concentrated sulfuric acid as a catalyst yielded a mixture of 2,4-diacetylresorcinol, 32, and 4,6-diacetylresorcinol, 42; These compounds were separated by recrystallization from acetone, from which pure ég'was Obtained (mp 179-181°) (lit. 182° (21)). v2,4-Diacetylresorcinol, 32, was obtained from the acetone mother liquor and after fractional sublimation had a mp of 87-89° (lit. 92° (21)). The nmr spectrum of 32,(Table II) had peaks at T 7.48 and 7.30.for the two dif- ferent acetyl methyl groups and an AB pattern for the two aromatic hydrogens centered at 1 3.05 (J = 9 cps). The low 28 OH g1 22. field doublet was assigned to the hydrogen adjacent to the acetylgroup. 2,4-Diacetyl45-hydroxy-6,6—dimethyl-2,4-cyclohexa- ‘ .dienone, 2;, was prepared from 22 by treating it with methyl iodide and sodium methoxide in anhydrous methanol for four days at room temperature. The methanol was evap- orated, water added, then the solution neutralized and extracted with ethyl ether. From a potassium bicarbonate extract of the ether a; 15.5% yield of gl’wss obtained.rw The compound had a mp of 134-137° after recrystallization. reprecipitation from base and sublimation. Its ir spectrum had absorption bands at 1560 cm.1 for the chelated car- bonyl and 1670 em-1 fOr the conjugated carbonyls. The nmr spectrum (Table II) had peaks at r 8.62 for the gem dimethyl group, 1 7.60 for one acetyl methyl group.r 7.50 for the chelated acetyl methyl group and r 1.27 for the vinyl hydrogen between the two acetyl groups. There was no sign of the chelated hydroxyl hydrogen. The compound ana- lyzed W811 for C138140‘- 29 From the ether layer 2~methyl—4,6—diacetylresorcinol, 42, (mp 132-133.5°) was isolated. This compound was identi— fied by its nmr spectrum (Table II) which had peaks at T 8.02 for the aromatic methyl group, T 7.48 for the two acetyl methyl groups, T 2.08 for the aromatic hydrogen and T -3.0 for the two hydroxyl hydrogens. This compound was also prepared by treating 2—methyl- resorcinol with acetic anhydride (two molar equivalents) using concentrated sulfuric acid as a catalyst. ‘A 63.6% yield of 42'was obtained. 2-Methyl-4,6—diacetylresorcinol, 42“ was treated with methyl iodide and sodium methoxide in methanol for one week and the reaction was worked up as in the preparation of 42; A 56% yield of 2,4-diacetyl-5- hydroxy—6,6-dimethyl-2,4-cyclohexadienone, 4;” was obtained. 4-Acetylresorcinol, 42, was prepared from resorcinol and acetic anhydride (one molar equivalent) by refluxing for ten minutes with concentrated sulfuric acid as a catalyst. A 63.6% yield of 42 was obtained with a mp of 145° (lit. 30 147° (18)). The nmr spectrum (Table II) had peaks in the aromatic region at T 4.27, 4.22, 4.07 and 4.03 (for two hydrogens) due to the hydrogen at atom 2 between the hydroxyl groups and the hydrogen at carbon 6 adjacent to one hydroxyl group. This latter hydrogen was coupled to the hydrogen ortho to it (J = 9 cps) and also to the hydrogen meta to it (J = 2.0-2.5 cps). The nmr spectrum also had a quartet at T 2.80 assigned to the hydrogen adjacent to the acetyl group and coupled to the adjacent hydrogen (J - 9 cps) and to the hydrogen para to it (J = 0.8-1.0 cps). Peaks for the hydroxyl hydrogens were found at r 1.10 and —2.59. 4-Acetyl—5-hydroxy—6,6-dimethyl—2,4-cyclohexadienone, 24, was prepared in two ways. 2,4-Diacetyl-5-hydroxy-6,6- dimethyl-Z,4-cyclohexadienone, 42” was refluxed in 2N hydrochloric acid to give 44. Alternatively, 22'was methyl- ated with methyl iodide in sodium methoxide and methanol. In neither case was the yield very good nor the product very pure. It was an oily substance which had ir absorption bands at 1605 and 1660 cm.1 for the chelated and conjugated 31 carbonyl groups respectively. The nmr spectrum had peaks at 1 8.71 for the gem dimethyl group, T 7.71 for the acetyl methyl (note the possibility of the acetyl car— bonyl being less chelated than the cyclohexadienones men— tioned previously), and a multiplet at 1 3.51 for an AB pattern caused by the vinyl hydrogens, with a coupling constant of 10 cps. The down field peak was assigned to the vinyl hydrogen adjacent to the acetyl group. Methylation of 42,gave largely two compounds, a sodium bicarbonate-soluble compound, 2-methyl-4-acetylresorcinol, 22, and an ether-soluble compound which dissolved in 2N sodium hydroxide with difficulty, identified as 2-methyl- 3-hydroxy-4-acetylanisole, 22, Compound 22 was identified OCH3 by its mp of 151-155° (lit. 156-157° (22)) and ir spectrum which had absorption bands at 1620 cm-1 for the acetyl car- -1 bonyl and 3000 and 3300 cm for the chelated and non-che- lated hydroxyl groups reSpectively. The nmr spectrum as given in Table II was consistent with this structure. Com- pound gg'was also identified by its mp of 78° (lit. 82—83° 32 (22)), nmr (Table II) and ir Spectra. Attempted methylation of 2,4-diacetylresorcinol, 22, with methyl iodide in sodium methoxide and methanol reé' sulted in recovered starting material. See Fig. II for an outline of the compounds synthesized from resorcinol. II. Discussion of the Structures of Filicinic Acid and Related Compounds The question of the structure of filicinic acid ac- tually began some time ago when Boehm suggested 22,0r 22' as possible structures (13,23). He found that oxidation 0 | II HO'OH HO I OH 12, 29. of filicinic acid with potassium permanganate gave car- bonic acid, isobutyric acid and dimethylmalonic acid. Since the tautomers can probably rapidly interconvert dur— ing a chemical reaction, the results obtained may not be reliable in determining which structure is better. With the use of more recent spectrometric methods of analysis such as ir, uv and nmr, it is possible to examine molecules without disturbing their structure appreciably. In order to have suitable standards, several compounds such 33 as 22, 22 and 42 were used for comparative purposes. There can be little doubt as to the structures of these compounds. The uv spectra of each of these, along with filicinic acid, diacetylfilicinic acid, methylfilicinic acid, and acetylmethylfilicinic acid, were determined at varying acid and base concentrations. Their changing absorption inten- sities at given wavelengths were plotted against the concen- tration of the acid or base added to the solvent (water). These graphs are presented in the next few pages. Each of these will be discussed in connection with other spectral evidence to assign a definite structure to each compound and its anion(s). The uv Spectra reported for these compounds in the literature vary considerably in wavelength and intensity of absorption. The reason for this is that different acid and base concentrations were used by different workers. The uv spectra they measured were of compounds for which different and undetermined concentrations of the various anions and neutral or protonated forms were present. By 34 carrying out the uv Spectral determinations at different base concentrations each absorption band was identified and assigned to a particular structure. Although there are structures in the literature for each of these compounds they vary depending upon the author. For this reason all of the compounds with questionable structures were given an independent proof. A. The Structure of Filicinic Acid The structure of filicinic acid has been given asygg (10,12,13,23,24,25) and 22 (23,25). For a recent review of related compounds see reference 25. AS the ir Spectrum did not indicate any non-conjugated carbonyls, poly-keto structures which could conceivably be present can be elim— inated from consideration. This applies also to the other related compounds to be discussed later. In all cases the fully enolized structures are the most probable. O 11 O. .33 47 The uv spectra at various concentrations of acid and base for filicinic acid are given in Fig. III. The uv spectrum of the neutral compound had a band at 250 mu. This compares favorably with the band at 258 mu for compound 22' (Fig. IV) and 260 mu for compound 41 (dihydroresorcinol)(26). x 10 (I Absorbance 35 275 300 Wavelength (mu) e x 10‘3 Absorbance 20 15 Fig. IV. 36 UV Spectra of 3-hydroxy—4,4,6,6—tetramethyl—2- cyCIohexen—1,5-dione. /3 225 250 275 3b0 325 350 “ Wavelength (mu) /, F A 258 / """" \. — ° \. / \ /. \// . /\\\ ./ / / - A 275 ' \ ..__. _--..__'/ 16% 16% 1' ‘ :3 ' 5‘ ‘ 7L :9 1: 1'3 37 This comparison indicates that filicinic acid has the cross conjugated structure, 22; The fully conjugated form, 22” would be expected to have a uv absorption band in the re- gion of 300 mu, as observed for other fully conjugated structures discussed later. The anion of filicinic acid can have two forms, ggland '42; Schwarzenbach (26) explained the changes and differ- ences in intensities of various uv bands at different base concentrations to different concentrations of the two possible forms of filicinic acid. He found that the uv spectrum of 52 had uv absorption bands at 360 mu and 290 mu. Na+—O- O O-Na+ I u l CH3-C=CH-C-CH=C-CH3 29. However, the uv Spectrum of the monoanion of filicinic acid (Fig. III) had a band a 278 mu, comparable to bands at 275 mu and 280 mu for the anions of 22'and 41 respec-' tively. This would indicate that gg'was the form of the monoanion. 38 In Figure III there was a sharp change in intensities of absorption wavelengths at pH of 10-11. When the base concentration became large enough, presumably the dianion of filicinic acid formed. It had uv absorption bands at 348, 250, and 231 mu. This ion can be represented by structure 52; The monoprotonated form 52 was formed in strong acid solutions (pH < 1) Some experimental evidence for the structure of filicinic acid came from the preparation of its mono- and diacetates, 2g and 21. If 20 were the structure, the di- acetate would be expected to have two nonequivalent acetyl methyl groups and two different vinyl hydrogens. Due to the symmetry of the diaéetate of 29, only one peak for the acetyl methyl groups and one for the vinyl hydrogens would be present in the nmr spectrum. This was found as the nmr spectrum (Table I) had one peak at T 7.75 integrating for six hydrogens and one peak at T 3.99 integrating for two hydrogens. This indicates that the structure of filicinic acid is £2. 39 B. The Structure of Diacetylfilicinic Acid There are two structural possibilities for diacetyl— filicinic acid. Structure 52'has been used recently (12) for this compound, although 21 is now believed to be the correct structure. Evidence for this form came from comparison of its uv spectrum (Fig. V) with those of 22” (A x 277 and 241 mu) (Fig. VI) and 41 (xmax 310-345 mu) ma (Fig. VII). The cross-conjugated form, 52, would be expected to have a uv spectrum Similar to that of 22, However the neutral compound actually had absorption bands at Amax 300— 315 mu, which suggests that the structure is the fully conjugated form, 21. 21 53 32 41 In addition the acetyl and hydroxyl groups are equiva- lent in structure 52” and would be expected to have only one peak in its nmr spectrum for the acetyl methyl groups and one for the hydroxyl hydrogens. The nmr spectrum actually had peaks for two different acetyl methyl groups (Table I) at T 7.47 and 7.34, and two peaks for two differ- ent hydroxyl hydrogens at T —8.45 and —8.85. This e x 10'-3 Absorbance Fig. V. 40 2%5 360 Wavelength (mu) A 340 20 . 15 _ ‘"x 1 300—315 “ \ \‘ 10 r 5- 50% 10% 1 L 5 7 9 11 13 %H2304 e x 10-3 Absorbance 20 15 Fig. VI. 41 UV Spectra of 2—acetyl-4,4.646-tetramethyl-2— cyclohexen-1,5-dione. 0—O—O—O_ 225 250 275 360 "-3325 350 Wavelength (mu) A 277 / _ _._._ _/ ./ / /. — — —. —'—'\ ./ // A 241 -\. /./ l 1 1 1 l l L L l I 1 J A 1 20% 10a 1 3 5 7 9 11 13 e x 10"3 42 Fig. VII. UV Spectra of 2,4-diacetyl—5-hydroxy-6,6-di- __ _- methyl-Z—cyclohexadienone. .9 D - OH to o” o o c m n 3 /’—~*\ ' *\ 3 ./ \ l4: '/ \ ./ /.»—\\ ./‘ \ / “ \ '\ 1 I I 1‘\‘ 250 275 300 325 350 400 450 Wavelength (mp) 20- 15' A 398 10’ /_ x 279 \. I 5- -..,_. I _ __ -..—\(\, 1 310-345 -— ————— / \ )\ 245 26% 16%1 '5 5 V 7 9 11 13 %sto4 pH 43 corresponded to 85% of the material with an additional 15% of another isomer, perhaps 22; The nmr spectrum thus con- firms the structure of diacetylfilicinic acid to be 22; The upfield acetyl methyl group (T 7.47) was assigned to the group adjacent to the carbonyl on carbon 1. The other acetyl methyl (1 7.34) was assigned to the group between the two hydroxyl groups, on carbon 4. These assignments were based on a comparison with the nmr spectrum of 22' which had its acetyl methyl group at T 7.44. Also a com- parison showed that the gem dimethyl group for 22“ between the carbonyl and hydroxyl groups, had an nmr peak at T 8.59 consistent with that for ZA'(T 8.59). The upfield nmr peak of §2’(T 8.69) was assigned to the gem dimethyl group be- tween the two carbonyl groups. The possibility for two different monoanions of di- acetylfilicinic acid also existed, as represented by 22 and 22; Anion 22 would be expected to Show a hypsochromic shift in the anionic form from the absorption of the neutral compound (see later arguments). Structure, 22” on the other hand would be expected to show a bathochromic shift from the neutral molecule in the uv absorption and this in fact was observed. There was a shift from 315 mu to 340 mu (Fig. V). Comparison of this to the uv spectrum of 21 (Fig. VII) shows agreement in the trend of the shift but not in its magnitude. That is, in going from 22 to 22 there was a bathochromic shift from 310-345 mu (broad) to 398 mu. The anion of diacetylfilicinic acid cannot be 24 since this 44 should have a uv spectrum similar to the monoanion of 22, which must have structure 21; This anion, however, had its long wavelength absorption at 277 mu. Additional evidence came from the nmr spectrum of 22' in dimethylsulfoxide-da, in which there was only one peak for the acetyl methyl groups (T 7.40). In this solvent 22'was probably present as its monoanion, due to the basicity O + + CH3 “S "CHA3 I OH of the solvent. Structure Qg'is symmetrical, whereas 2% might be expected to Show two acetyl methyl peaks. Other less likely explanations are possible; for example 22'might 45 be the major component in this solvent, or there might be rapidly interconverting structures between 22;and §§u causing the acetyl methyl groups to become equivalent. C. The Structure of Methylfilicinic Acid Methylfilicinic acid was assigned structure 22 because its uv Spectrum (Fig. VIII) was similar to that of filicinic acid, 22; There was a slight bathochromic shift from that of 22'because of the extra methyl group. o o It . HO'OH 0" OH 30 58 59 Methylfilicinic acid, 22” has three possible mono- anionic forms, but because the uv Spectra were similar to those of filicinic acid (Fig. III) the monoanion was as- signed a cross-conjugated structure. Of the two possible such anions (22'or 22), structure 22'would be more probable since there is little or no increase in negative charge on the carbon atom bearing the extra methyl group. D. The Structure of Acetygfilicinic Acid Acetylfilicinic acid may have a cross-conjugated structure and two different fully conjugated forms. e x 10"3 20 15 10 46 Fig. VIII. Absorbance Uv spectra of methylfilicinic acid OH O 1 227 ’ ,/7\255 / / f _ '— I,/ I 6 ‘1248 1285 p 1255-265 4 o ."-'\. I, ./ \ . . , ‘\. I u " \ ’’’’’ F' . . .\- ./ I/ . ‘\./ / o ./ \ — / .\s\\ / 72 ' A 350 \ "‘-. \ \ \. .# _“_ x 358 1 \--——- 20% 10% 1 3 5 7 9 11 13 % H2804 PH 47 Because 1ts uv spectra (Amax 330 mu (aCid) and}max 345 mp. (basic)) were similar to those of diacetylfilicinic acid, 22; (Kmax 300 mu (acid)and Amax 340 mu (basic)) its struc- ture was thought to be either 22'or 22, The latter has appeared in the literature (15) as has the cross-conjugated O HO OH HO OH 22 60 form (12). However, it is probable that 22'is the correct structure. The nmr spectrum of acetylfilicinic acid (Table II) has a peak for an acetyl methyl group at T 7.50 which agrees with other examples of acetyl groups between a carbonyl and hydroxyl group, chelated to the hydroxyl hydrogen. The vinyl hydrogen was at T 4.49. To Show further that structure 22'was correct, the monomethyl ether, §§x was prepared. Forsen and Nilsson (27) have previously prepared this compound and have argued that 22'was the structure. They based their choice on the fact that methylation of acetylfilicinic acid could give several methyl ethers, only two of which are capable of giving enols with strong intramolecular hydrogen bonds, 22 and 22, They are both interconvertable to other forms, 22. and 23.- 48 O HO OCH3 CH3O 38 62 My m Forsen and Nilsson say that 22'would be unlikely Since it does not have the additional stabilization to be gained from an intramolecular chelated hydrogen bond. However, they observed two distinctly different forms in the nmr which were interconvertable as shown by a ratio of 1:6 in carbon tetrachloride solution and a ratio of 1:8 in benzene solution. They therefore felt that the correct structure must be of the type 22 or 22. The uv spectrum of the methyl ether (Afing (acidic). 315 and 235 mu) compared with that of acetylfilicinic acid (1:::H (acidic) 330, 271, 235 and 204 mu), indicates that the larger fraction must have structure 22; Additional evidence for this structure was provided by its nmr spectrum 49 (Table II). The acetyl methyl (7 7.47) had not shifted from that of acetylmethylfilicinic acid, 22/ (T 7.50), but both the vinyl hydrogen (T 4.65) and the gem dimethyl group (1 8.70) had shifted (from T 4.49 and 8.62) indicating that the methoxy group must be between the vinyl hydrogen and the gem dimethyl group, as in 22; In addition the uv spectrum in base showed a hypso— chromic shift (from kmax 315 mu to Rmax 305-310 mu) as con- sistent for the anion 22, This was compared to the batho— chromic Shift for the anion of acetylfilicinic acid, 22, £5.95 £5.53, (kmax 345 mu) from that of the parent compound 22’(1max 330 mu). This would be consistent with structure-22 for the neutral compound and 22 for its monoanion. E. The Structure of Acetylmethylfilicinic Acid Acetylmethylfilicinic acid has three possible neutral structures. As the uv spectrum (kmax 334 mu) (Fig. IX). was like that of acetylfilicinic acid, 22” (Amax 330 mu) only the fully conjugated structures 22 and 22 will be 1X. 50 UV Spectra of acetylmethylfilicinic acid. Absorbance 1 1 I 1 ' 225 250 275 300 325 350 400 r_._ i 20- ( h 350 15. I \ 10' m 0 I a , 72-: .Z” \. A 279 X 5 .K4' 0 / \ 71-. 250 0 L1 L 1 1 J” 1 1 1 l J 1 J 1f 1 20% 10% 1 3 5 pH 7 9 11 13 95112304 51 considered. An nmr spectrum (Table I) of acetylmethylfil- icinic acid was consistent with structure 22/ indicating a gem dimethyl at T 8.68 and an acetyl substituent between a carbonyl and hydroxyl group at T 7.50. Structure 22’is the most probable since the monomethyl ether prepared from 22 can be assigned structure 21; This structure was con- 29 2.2 31 firmed by its nmr spectrum (Table II) which had peaks at T 7.43 for an acetyl methyl between a carbonyl and hydroxyl group. The uv spectra (xmax 323-340 mu (neutral) and 285- 310 mu (basic)) were Similar to those of 22, and show the bathochromic shift expected for 21, III. Photolysis and Photo-oxidation of Diacetylfilicinic Acid A. Direct Photolysis of Diacetylfilicinic Acid Recently Hart and co-workers (4,6) have Shown that some methyl-substituted 2,4-cyclohexadienones undergo a photo-rearrangement to give bicyclo[3.1.0]hexenones in much the same way as 2,5-cyclohexadienones (2,3). This 52 rearrangement was unusual, for 2,4—cyclohexadienones usually cleave to open chain compounds or phenols (1,2,3). Since diacetylfilicinic acid, 21/ was shown to exist in the 2,4-cyclohexadienone form, it was decided to study the photolysis of this and related compounds to ascertain the effect of acetyl and hydroxyl substituents on the reaction course. Diacetylfilicinic acid was photolyzed using a Hanovia L 450 watt lamp under a variety of conditions through Pyrex. No noticeable reaction in ethyl ether or in methanol under nitrogen was observed. There also was no reaction when 22'was photolyzed in methanol under oxygen, nor in the presence of alkaline methanol under nitrogen. Thus diacetylfilicinic acid appears to be photochem- ically inert, at least with respect to radiation > 300 mu, despite its absorption band in that region. It was noted, however, that irradiation of alkaline solutions of diacetyl— filicinic acid (243,, the monoanion) in the presence of air resulted in a gradual change in the absorption spectrum of the solution. Since no Similar change occurred in a nitrogen atmOSphere, it was concluded that a photo-oxidation was involved, and this reaction was studied in detail, as described below. B. Photo-oxidation of Diacetylfilicinic Acid Diacetylfilicinic acid, 22 (0.03 molar solution) was irradiated in about 0.1 N sodium methoxide in methanol, 53 while bubbling air through the solution. It was smoothly converted to a new crystalline product, mp 77.5-790, C11H1406, to which the structure 2-acetyl-3,4-dihydroxy— 4—carbomethoxy-5,5—dimethyl—2—cyclopentenone, 21) is as- signed. The reaction was followed by the appearance of new uv absorption bands at xmax 271 and 250 mu in place of the band at xmax 340 mu (due to the monoanion of diacetylfili- cinic acid, 22). O O” l \ coocna 67 55 Compound 21 was identified by its spectral properties and chemical behavior. It gave a parent m/e of 242 in its mass Spectrum. Its nmr Spectrum (Table III and nmr Spectrum I) indicated the presence of two methyl groups probably due to a gem dimethyl group, Since rearrangement of one of the methyl groups of 22'(or 22) seemed rather un- likely. These methyls had nmr peaks at 1.8.76 and 8.71. There might be four reasons for the presence of two methyl peaks rather than one: (1) restricted rotation in an open chain compound, (2) coupling with another hydrogen, (3) loca— tion adjacent to an asymmetric carbon, or (4) presence of tautomeric isomers. 54 mm.m Amvmh.w mm.o mm.s xmvvw.m 3.30 .mm m Amvfim.m >3. mo om sameness museums mm.n me.o sv.s Amvss.m sauce cs 0 m _ QUOHO Hmnum oo.m mm.m smoom: om ms.n ss.w AmmoHo 220> v o 2;. mm; _ Am.> n be ww.m mw.m mumamcflm 0331 H 0 mm m/ to mm.m No.0 Hm.s ow.m d mmo loot moooubuooom Amvsm.o mm.w 0coumo¢ sec .0.508 AmvH>.m .(c me.o mm.m me.s “moms.w «Home as 90:00 mo: «moon Ouoummo «mo\\owwo ucm>aom 02902500 mocsomaoo pmumamu Ucm muodooum soaumonOIouozm mo muadmws muuommm mosesomws caumcmmfi Hmmaosz .HHH wanme 55 ammouamm Hmca> mm.v ma.m A» u so “seem.s m.mu wv.s Asunvam.w ”Home .mm Amvms.w have .ca nm.m we.» Amvam.m muoumua mm msouw Hmfium umooo O mm.m oo.w Amvmw.w .cz mo 5 mm.m ms.m mm.m mm.» Amvmm.m 930 em 0 Am.s u so om.m mw.w fl mm.m mmoo QDOHG Hmsum me n Ho.w “moms.m 1). Anne n so ms.m ms.m em.m mm.s Amvmm.w wave mm sm.m mm.m sm.m Ho.m Amvew.w . a). mfi.w, mm.s Amvoo.m vH00 mm umguo mOI mmUOI ouUImmU mmo\\0m U ucm>aom UCSOQEOU - A.ucoov .HHH manme 56 Restricted rotation was unlikely because there was no change in the relative positions of the methyl resonance peaks as the temperature was varied from 35° to 60°. How— ever the peaks did look sharper and better separated at the higher temperatures (Nmr Spectrum 1). The methyls were probably not part of an isopropyl group, because the difference between the two peaks was only 3 cps instead of the expected coupling constant of 7—10 cps. Long range coupling was ruled out by determining the nmr spectrum at 100 Me. In this case the methyl peaks were separated by 5 cps; there would have been no change if the difference had its origin in coupling to another hydrogen. The conclusion was that the chemical shift dif- ference of these peaks was either caused by a difference in methyl environment due to a ring structure with an asymmetric: center, or by the presence of equal concentrations of two tautomeric forms. The nmr spectrum had a singlet at T 7.49 probably due to an acetyl methyl between two carbonyls of an enolized a-diketone (see Table I for other examples such as 22) 22) and 22). Therefore the photOproduct contained the partial structure given below. 0 0 ’% HO J\ Partial structure 32 57 This partial Structure was supported by the uv spectrum of the photoproduct which had absorption bands at Xmax 273 mu in its neutral form.and xmax 271 and 250 mu in its monoanion form. These Spectra were not unlike those of compound 22’(1max 277 and 241 mu (acid) and kmax 277 mu (basic)). /' Photo-oxidation of 22'in aqueous alkali yielded a small amount of a compound which had an ir spectrum identi- cal to that of dimethylmalonic acid and an nmr Spectrum (Table III) which substantiated this structure. This in- dicated that the gem dimethyl group remained intact during the photo-oxidation, and that the partial structure for the photo-oxidation product must have been as.shown, with two tautomeric forms pOssible. The existence of two tautomers could explain the presence of two methyl peaks in the nmr spectrum. However, an nmr spectrum of the photoproduct in deuterium oxide and sodium deuteroxide solution still indicated two distinct methyl groups (difference of 1 cps), rather than a combina— tion of these two to form one peak as expected for the anion of the two tautomeric forms, 22; Thus the difference in the methyls must be due to a difference 58 in environment caused by the groups on an adjacent carbon atom. The deuterium oxide, sodium deuteroxide nmr spectrum also indicated the presence of two easily exchangeable hydrogens, other than those of the acetyl methyl group which also exchanged, but slowly. This suggests that there are two hydroxyl groups in the photo-oxidation product. The nmr Spectrum of the photoproduct also Showed a three proton Signal at 7 6.25, assigned to the methyl hydrogens of a carbomethoxy group. The presence of an ester was confirmed by a band at 1740 cm.1 in the ir Spec- trum (Infrared Spectrum 1). By carrying out the photo-oxidation in slightly basic 95% ethanol a different photoproduct, mp 80-81°, C12H1606, was obtained to which the structure 2-acetyl—3,4-dihydroxy- 4-carboethoxy-5,5-dimethyl-2-cyclopentenone, EEK was ulti— mately assigned. The nmr spectrum (Table III and Nmr 0 [I °° fl OH HO .C-O-CHZCHa 0 69 59 Spectrum 2) indicated the presence of a carboethoxy group in place of the carbomethoxy group. This showed clearly that one acetyl group was lost from diacetylfilicinic acid, 22) and that a solvent molecule was added, probably as an eSter group. When either 22 or 22'was heated in 2N sodium hydroxide at 100° for sixteen to twenty hours a single product with the molecular formula C9H1204 formed in 79% yield. This represents the loss of C2H202 or C3H402 due to saponifica- tion of the ester and decarboxylation as.eXpected for a B-keto acid. Possibly the alkaline cleavage of the ester group was direct, rather than 223 saponification and de- carboxylation. This possibility was strengthened when 22' was prepared by refluxing 2Z'with sodium methoxide in meth- anol. An ir spectrum of the photo-oxidation product, 21' showed the presence of an unassociated hydroxyl at 3450 cm-1. All the evidence presented here was combined to de- duce structures 21'and 22,for the photo—oxidation products, and structure ZQ’for the degraded compound. .60 The nmr Spectrum (Table III and Nmr Spectrum 3) of 2-acetyl-3,4-dihydroxy-5,5—dimethyl-2-cyclopentenone, 12) indicated the presence of a gem dimethyl group adjacent' to an asymmetric center (T 8.73 and 8.61), an acetyl group (T 7.47), a methyl hydrogen (T 5.85) and an hydroxyl hydrcr— gen (7 0.45). The ir spectrum (Infrared Spectrum 2) was Similar to that of 2Z'except for the band at 1740 cm—1 which was absent. In addition the acetyl carbonyl of 22' was reduced with sodium borohydride to give a compound with a uv absorption (xmax (acidic) 248 mu, (basic) 276 mp) similar to the uv absorption of compound 22'(Amax (acidic) 258 mu, (basic) 279 mu). The nmr spectrum of Z2 indeuterium oxide and sodium deuteroxide indicated the presence of two easily exchanged hydrogens due to two hydroxyl groups. The peak at r 5.85 did not exchange and must be due to a methyl hydrogen ad- jacent to an electron withdrawing group(s). It appeared at too high a field to have been a vinyl hydrogen. There was no noticeable change in the relative positions of the two different methyl groups in this solution, as compared to the nmr in deuterated chloroform. This indicated, as 61 before, that the different methyl peaks were not due to two tautomeric forms. The mp of 22 was 104-105° which agreed with that of the known compOund (104-106°) which had been prepared by Stevens (28) in the following manner: 2N NaOH .. ; <9, ; MnO ”‘ 0 w ;fi\ 510/“— OH HO \‘0 H B1203 70 71 To further identify Z2'it was oxidized with manganese dioxide and also with bismuth oxide to Zl’in a manner simi— lar to that used by Stevens. A yellow crystalline compound was obtained with a mp of 87-88° (lit. 87° (28)). The uv spectrum was similar to that obtained by Stevens. C. Dye—sensitized Photo-oxidations The mechanism for the photo-oxidation of diacetyl- filicinic acid, 22“ was most likely via a self—sensitized 62 reaction which produced Singlet oxygen. Singlet oxygen has been postulated recently as being involved in numerous sensitized photo-oxidations (29,30,31). Schonberg (32) had originally proposed that self-sensi- tized photo-oxidations occur gig a "moloxide" intermediate ('S-O-O') which then transferred the oxygen to a second molecule of reactant (A). The existence of this "moloxide" was based partially on the fact that anthracene was photo- oxidized to a peroxide faster in the presence of diphenyl- anthracene than in its absence. -This would indicate that there was a transfer of oxygen from the diphenylanthracene "moloxide" to the anthracene molecule. This mechanism for photo-oxidation was adopted by Schenck (33) for dye-sensitized photo-oxidations, and can be summarized as follows: S + hv --> -S-* s * + 02 —> s-o—o-* -s-o—o-* +A —-> s + A02 In a solution without an added sensitizer, 8.: A. Schenck used kinetic, spectrosc0pic and chemical data to support this mechanism. Kasche and Lindquist (34) considered that the triplet state of the sensitizer was responsible for its photo- chemical activity, although possibly the sensitizer might react gig its singlet state. In addition to the short-lived "moloxide" they suggested two additional mechanisms. One 63 involved reduction of the photo—excited dye by the oxidiz- able reactant to a semireduced dye radical, followed by reoxidization to the ground state by the oxygen present in solution. The other is one in which the photo-excited dye is oxidized by oxygen to a semi-oxidized dye radical, which is then reduced to the ground state by the oxidizable re— actant. They based the latter mechanism on the formation of a semi-oxidized dye Species in the photo-oxidation of fluorescein. More recently it has been prOposed that singlet oxygen is formed during the reaction and this subsequently reacts with the reagent to give the photo-oxidation products. Foote (29) proposed the mechanism: hv Sens. > 1Sens. 2:22- 1Sens. > 3Sens. 3Sens. + 302 > Sens. + 102 102 + A > A02 (prOdUCt) He based this mechanism on the similarity of products ob- tained from dye-sensitized photo-oxidations and from oxida- tions of the same compounds by singlet oxygen formed from sodium hypochlorite. Corey (31) produced singlet oxygen using gaseous oxygen subjected to an electrodeless dis- charge at 6.7 Mc. He then bubbled this through various solutions and found oxidation products identical to those obtained from photo-oxidation reactions. Compounds com- pared included anthracene, 9,10—dimethylanthracene and some reactive 1,3-dienes. 64 McKeown (31) produced singlet oxygen by the reaction of alkaline hydrogen peroxide with either sodium hypochlorite or bromine and also by decomposition of alkaline solutions of organic peracids. He allowed the singlet oxygen thus produced to react with derivatives of anthracene and obtained peroxides identical to those formed by photo-oxidations. Thus the more recent evidence supports a singlet oxygen mechanism for these photo-oxidations. Similar techniques were used in the present work to explore the scope and mechanism of the photo-oxidations of the enolate anions of 1,3,5—cyclohexanetriones. Diacetylfilicinic acid, 2;“ was photo-oxidized by irradiating an alkaline methanol solution, under air, with a 200 watt tungsten lamp using a variety of dyes (Rose Bengal was found to be the best) as sensitizers. A good yield of the photo-oxidation product, QZ/ was obtained. Other dyes used besides Rose Bengal were fluorscein and methylene blue; however, these were found difficult to work with due to problems in separating the dye from the product. In a standard reaction, photo-oxidation of the mono- anion of diacetylfilicinic acid, 2g, without a dye sensi- tizer, gave only 4% reaction in two hours. On the other hand, in the presence of Rose Bengal (0.0002%) a 98.8% yield of QZ'(as measured by uv spectroscopy, 73.8% isolated yield) was obtained after thirty-five-minutes of irradiation. Additional evidence that singlet oxygen was responsible for the photo-oxidation of Qg’was obtained by allowing 65 chemically prepared singlet oxygen to react with the anion. Wasserman :35} decomposed the peroxide of 9,10-di- phenylanthracene by heating it in benzene or chloroform in the presence of such reagents as 2,3,4,5-tetraphenyl- furan. He postulated that the peroxide decomposed to give the parent hydrocarbon plus singlet oxygen; the products isolated were similar to those formed in typical photo- oxidation reactions. When an alkaline solution of diphenylanthracene per- oxide in methanol was refluxed with Eé'a good yield of QZ' was obtained. The 9,10-diphenylanthracene-9,10-cyclic peroxide was prepared by passing air through an ether solu- tion of 9,10-diphenylanthracene while irradiating through Pyrex with a Hanovia L 450 watt lamp. The reaction was followed by observing the disappearance of uv absorption bands in the region of R 350-400 mu and the appearance of a band at kmax 210 mu. On completion of the photo-oxida- tion the ethyl ether solution was concentrated and added to an alkaline methanolic solution of 21; When this solu- tion was refluxed in the absence of light, singlet oxygen was produced and subsequently afforded the photo-oxidation product, Q1, However, because the 9,10-diphenylanthracene peroxide did not seem to decompose very rapidly in a standard reaction (see also Ref. 35) it is possible that the reac— tion of Qfi with the peroxide might be proceeding other than via singlet oxygen. 66 In conclusion the function of the anion of diacetyl- filicinic acid, g2“ as as sensitizer can be replaced by dyes known to produce singlet oxygen, or £2 can be oxi- dized by singlet oxygen formed from diphenylanthracene peroxide. D. Mechanism of the Photo-oxidation of Diacetylfilicinic 5.21.9 The photo-oxidation of the anion of diacetylfilicinic acid, 66, probably occurred via a self-sensitized photo- oxidation. A reasonable scheme based on that presented by Foote (29) would be: 67 hv 5.2 > ”it” Main isc > 3(2g)* 3(55)* + 302 > £2 + 102 g + 102 > 67 rw In order to determine which position of anion Qé'would be most susceptible to attack by singlet oxygen, the charge densities in the anion were calculated using molecular orbital theory (w—technique). Molecular orbital calcula- tions were performed using a program written by Professor R. S. Schwendeman of Michigan State University (see Appendix II). This employed a variation of the w—technique which allows for the changes in molecular orbital energies caused by the introduction of the oxygen heteroatoms. Values for the coulomb and resonance integrals for the car- bonyl and hydroxyl oxygen atoms were taken from Quantum Organic Chemistry (36). When calculations were carried out on anions, appropriate ratios of these values were used depending upon the relative contributions of the carbonyl and hydroxyl groups. The parameters are given in Table IV. Molecular orbital calculations of the anion of diacetyl- filicinic acid, Q6“ showed a relatively high negative charge at carbon atoms 2 and 4 (which are equivalent). It is reasonable that attack by singlet oxygen, which can be represented by the structure below, should occur at the most negative site. This representation for the charged 68 Table IV. Parameters.for the oxygen heteroatom in molecu- lar orbital calculations Group(s) h k c=0 1.2 2.0 C-O-H 2.0 0.9 i 0363] 1.6 1 45 i ngH} 1.46 1.62 g ngH} 1.72 1.26 i ngH} 1.4 1.77 Table V. Charge densities of the anion of diacetyl- filicinic acid ,Atom Charge 1 0.1313 2 -0.2035 3 0.0605 4 -0.2035 5 0.1313 separated oxygen molecule has been used for the base catalyzed oxidations of enolizable ketones (37). A second formulation for singlet oxygen also has been used (31), and can be written as: 0:0 This would indicate its structural resemblance to ethylene, and explain the reactivity of singlet oxygen as a dienophile, capable of immediate addition to conjugated 1,3-dienes and to anthracenoid hydrocarbons. This formulation better explains the diradical character of singlet oxygen when reacting with olefins (33). The evidence indicates that singlet oxygen behaves very much like dienophiles and as such their reactions withcfifimes may involve a charge trans- fer process. Although the monoanion of acetylfilicinic acid, 62/ underwent a similar photo-oxidation to the monoanion of diacetylfilicinic acid, ééu it was shown that Qé'was not formed under the conditions of the photo-oxidation of 62' (but in the absence of air or oxygen) in the time that ii could photo—oxidize to 61; 70 A plausible mechanism for formation of the peroxide intermediate and the resulting photoproduct, 61, is given here. Rearrangement of intermediate zg'is similar to the rearrangement of humulone, 74, to isohumulone, lg” and to W rearrangements of related compounds (38). This rearrangement 71 of 12 to 12 has been carried out using either dilute base or uv irradiation. The latter method (39) required a long reaction time (thirtywsix hours). Due to the rapid rate of the present photo-oxidation and rearrangement, the re- arrangement of zg'to 61 was probably base-catalyzed and not light-catalyzed. 22. 21. This rearrangement is not unexpected in light of the work of Roberts (40). He found by using isotopic labeling that a benzoyl shift occurred rather than a phenyl shift in the decomposition of 1,3-diphenyl-triketopropane. This would be similar to the rearrangement of zg'to the photo- oxidation product, 62; 0 C6H5 C C6H5 " C6H5 " C6H OH 0 5 \C/ \ > \ I I) / c c -c —-c II II II V II 0 0 0 OH 0 1 0 OH 0 OH 0 OH II I "CO H I H+ II I C6H5-C-C-C6H5 <—-—2-06H5-c-c-06H5 <——- C6H5-C-C-C6H5 I I I H COOH c=o o-Na+ 72 A related photo-oxidation has been carried out by Matsuura and Saito (41). They photo—oxidized xanthine, 76a, and its dimethyl derivative, 222/ in an alkaline water solu- tion. After an initial Diels-Alder type addition of an oxygen molecule to the diene, cleavage of the perox— ide followed by a benzilic acid-type rearrangement was post- ulated to account for the product. In the absence of a sensitizer no product was formed. II ' II I I R \N/EEJ 02* R‘N $4311 ELK/10 NH2 0 "klI o‘kN _N 0” ‘ Nib ' R , N R R Zfig ; R = H OH 76b ; R = CH3 9.7 c=o 0 N Gill/N to NH2 \‘r OH NHz I Q—- . g/J90 g (EU/LO H R A suggested anion structure given here, 21, could be the species attacked by singlet oxygen. This would be con- sistent with the photo-oxidation of the monoanion of di— acetylfilicinic acid, £2. 73 O R\ ll - “ET I OCK‘W N R 77 55 Whether the intermediate peroxide in the photo—oxi- dation of ég'is 28 or 22 is inconclusive. However at this time structure 22,13 favored as cleavage of the C1-C2 bond in zg'might be favorable due to the stable anion, 82x.which would form. Products arising from this type of ring open- ing were not observed. E. The Formation of Dimethylmalonic Acid It was mentioned previously that dimethylmalonic acid formed when diacetylfilicinic acid, a}, was irradiated through Pyrex with a Hanovia L 450 watt lamp, under air, 74 in an alkaline water solution. The formation of dimethyl— malonic acid was used as evidence that cleavage of the C1—C6 and C5-C6 bonds of gl'did not occur in the photo- oxidation of gg'in alkaline methanol. Since the methanol photoproduct, 21' and the degraded compound 79.. were found to be inert under the conditions of the photolysis and photo-oxidation, they cannot be intermediates in the photo- oxidation of gg'to dimethylmalonic acid. Thus the forma- tion of dimethylmalonic acid must occur through a change in the mechanism prior to the rearrangement to 61; A sug- gested intermediate for the formation of dimethylmalonic acid could be 22, which could form if cleavage of the acetyl group was slow. This becomes probable when the con— centration of base is low. One possible mechanism for this reaction is outlined 0n the following page. F. The Methyl Ethers of the Photo-oxidation Products Some interesting results were obtained when thephoto— oxidation product , 61/ was treated with diazomethane. Two different methyl ethers, obtained in good and equal yield, had identical uv spectra. The two methyl ethers could not be separated by vapor phase chromatography, the peaks being only separated enough to tell that there were two compounds present. However it was possible to distinguish the two isomers by their nmr spectra (Table III). The two methyl ethers were identified as gl'and 82; The assignments of the nmr peaks were made on the assumption that the chemical 75 OH 2 <7 :9 CH3COOH + coon _ + €02 HO + coon co + CH3-C-CH3 O 76 9 0 ocn3 l 'l °' D h CH3O OH 03 OH coocss COOCH3 g; 22.2. shifts of the gem dimethyl group in §AI(T 8.84 and 8.75) would not vary much from those of EZ'(T 8.76 and 8.71), whereas those for §£'(T 9.00 and 8.84) would be expected to change due to the altered environment. The band at T 6.25 was assigned to the methyl ester group of gl'and the band at T_5.93 to the methoxy group on the double bond. For §§'the ester methoxyl was considered to give rise to the nmr peak at 1 6.19, whereas the methoxyl on the double bond was responsible for the peak at 1 6.01. The relative assignments were obtained by comparison not only with 62“ which showed the upfield methoxyl peaks to be due to the ester groups, but also by comparison with the two methyl ethers, gg'and 82, similarly obtained from Z}; Al- though §§Iand 82'were not separated, relative assignments of the peaks in their nmr spectra were possible because the compounds were produced in a 1:2 molar ratio. The compound in least abundance, 82“ showed the least change in chemical shift of the gem dimethyl group (T 8.85 and 8.73) in com- parison to the starting material, (1 8.80). Thus by taking 77 O 0CH3 I l COCHZCHa 0 83 84 rw I'W the lower intensity peaks throughout it was possible to assign peaks in the nmr spectrum as given in Table III. It was noted that the low field methylene hydrogens of the ethoxy group were due to compound_§2“ which had the methoxy adjacent to the gem dimethyl group. For this reason the compound having the low field peak due to the ester methoxyl (T 6.19) in the previous pair was taken as 82; In the same manner the position for the.methoxyl on the double bond was assigned as indicated in Table III. Since it has been previously shown (42) that an equi- librium such as that outlined here exists for this type of compound, it is not too surprising that two different methyl ethers were obtained in each case. For example, RI 78 Nilsson has shown (43) that of the four possible enol tautomers, shown below, of the 2-acety1 or 2-formy1cyclo— pentane-1,3-diones only §§,and 81 are probable when R =‘H. Because of the distinct AB pattern of the methylene O/Eio O"'H\. ll ,1, a R——» a” 0 o 252. s; 1T i (R = H, CH3) 1 C.) R (Q: R n °° ——>.___ 0”? o/H .~O...H 8.2. 8.2. hydrogens in the nmr spectrum, these two interconvert only slowly relative to the nmr time scale. Structures §§'and §§' were ruled out as there was no noticeable Coupling between the vinyl hydrogen and the expected hydroxyl hydrogen, when R ='H. Nilsson stated that the ir and uv Spectra of the compound where R = CH3 were similar to those where R = H; thus forms §§'and §Z’would also be preferable for the acetyl derivative. 79 Evidence that a similar equilibrium was also important for Qz'was the broadening of the nmr peaks for the methyl hydrogens of the gem dimethyl group (Nmr Spectrum 1). Be- cause the peaks for the chelated hydroxyl hydrogens could not be found in the nmr spectra of 61 and Qg'this indicated that these hydrogens might be broadened due to enol-keto tautomerism. When the methyl ethers, 81 and 82 or Qg'and 84” were placed in alkaline solution, 61 or 62 were recovered. This reaction occunnd very rapidly with only a trace of base in 95% ethanol. The driving force for this reaction could be the formation of a more stable compound since §Z,and 62, have an intramolecularly hydrogen-bonded hydroxyl group. O- , 9. °' a I. > CH30 OH > HO OH 0H COOR COOR g}; R : CH3 £1; R = CH3 ’8‘?) R =‘ "CH’CH3 227 R : ‘CH2CH3 IV. The Photolysis and Photo-oxidation of Acetylfilicinic~Acid A. Photo-oxidation of Acetylfilicinic Acid To extend the photochemical study to related substi- tuted 2,4-cyclohexadienones, the behavior of monoacetyl- filicinic acid, 22, was examined. When an alkaline solution 80 of gg'was irradiated through Pyrex using a Hanovia L 450 watt lamp, under air, a very rapid photo-oxidation occurred, Work-up afforded a 50:50 mixture of inand 12 as determined by nmr spectral analysis. Repetition of the reaction gave a mixture of 51 mole % Qz'and 35 mole % 12” Recrystalliza— tion from hexane-ethyl ether gave a 14% isolated yield of QZ identified by its ir spectrum. Acetylfilicinic acid, 22“ was also photo-oxidized us- ing a 200 watt tungsten lamp with Rose Bengal as the sensi- tizer. The photolysis was completed within an hour and an nmr spectrum of the product mixture indicated it to be 83% compound 61, The photo-oxidation product, 61“ is postulated to form in an analogous manner to that given before for the photo- oxidation of the anion of diacetylfilicinic acid, 52; In- stead of the cleavage of an acetyl group, a proton is lost. »The charge densities at the carbon atoms of acetyl- filicinic acid monoanion, 62“ were calculated using'the molecular orbital method described previously. The results are given in Table VI. This indicates that singlet oxygen should attack primarily at carbon atom 4, the most negative position (although carbon 2 carries nearly as much negative charge). The major product formed, 61, can be accounted for by the scheme shown, initial attack being at carbon 4. 81 O H " °’ I. 0.: D HO 0‘ 3 Ho OH COOCH3 COOCH3 67 Table VI. Charge densities of the anion of acetylfilicinic acid (65') Atom Charge 1 0.1288 2 ~0.1990 3 0.0420 4 -0.2082 5 0 1106 82 Although it was possible to prepare 22 from 61 under alkaline conditions it does not seem likely that the temperature of the photolysis (25-350) was high enough, the base concentration (0.1 N) strong enough, nor the reaction time long enough (two hours) for cleavage to have occurred to give 72 from 61 during the photo-oxidation. Also, com- pound Qz'does not photorearrange to ZQ’under similar condi- tions. Thus there must be a new mechanism Operating for the photo—oxidation of acetylfilicinic acid, not operative for diacetylfilicinic acid which allows the direct forma- tion of 12; One possible mechanism given here for the formation of zg'involves direct reaction of triplet oxygen with the MeOH 83 excited state of the monoanion of acetylfilicinic acid, 62“ since it Was found that fig'did not give zg‘on reaction with singlet oxygen formed in a dye-sensitized photo—oxidation. Still a third product was obtained when gz'was photo- oxidized through Pyrex using a Hanovia L 450 watt lamp in methanol containing only one molar equivalent of sodium methoxide. The reaction was slower than before. On work-up a small amount of a carbon tetrachloride-insoluble compound (mp 179—181°) was obtained, in addition to 62; The compound analyzed for the empirical formula C11H14O7. Its nmr spec- trum (Table III and Nmr Spectrum 4) indicated the presence of a gem dimethyl group at T 8.81 and 8.75, an acetyl group at T 7.42 and a methoxyl group at 1 6.65. ‘The uv spectra were similar to those of 61” The ir spectrum had absorp- tion bands at 1580 (broad), 1670, 1755 and 3000—3500 cm"1 (Ir Spectrum 4). The nmr spectrum of this compound in sodium deuteroxide and deuterium oxide still showed two distinct peaks for the gem dimethyl group. This would in- dicate that the two peaks were not due to the presence of two tautomeric forms but due to the gem dimethyl group being adjacent to an asymmetric center in a closed ring system. An unequivocal structure assignment cannot yet be made for this compound. From the combined evidence a partial structure can be drawn: | + —OCH3 + —OH + 0202 84 Structure gg'is one possibility for this compound which is consistent with the evidence. 0 - HO <:_ OH 0’] OCH3 52 B. Photoisomerization of Acetylfilicinic Acid When acetylfilicinic acid was irradiated in methanol or ether, under a nitrogen atmosphere, for 32-48 hours with a Hanovia L 450 watt lamp, a new compound was Obtained. This compound was identified as 3-acetyl-4-hydroxy-6fii- prOpyl-ZH-pyrone, 22, It analyzed correctly (C10H1204) for an isomer of 22; The nmr spectrum of 22'(Table III 90 91 92 and Nmr Spectrum 5) indicated that the compound had an acetyl methyl group (1 7.48), a chelated hydroxyl proton (T -6.9), a vinyl proton (T 4.28), and an iSOpropyl group (a six—pro- ton doublet at T 8.75, J = 7 cps and a one—ptoton multiplet centered at T 7.34, J = 7 cps). Its uv and ir spectra 85 resembled those of dehydroacetic acid, 91; The uv spectra had absorption maxima at A 309, 223 and 203 mu in acidic ethanol and at A 291, 229 and 204 mu in basic ethanol. “With dehydroacetic acid similar uv maxima appeared at A 310 and 225 mu in neutral solution, and at A 294 mu in basic solu— tion. This would not be likely for the other possible tauttr- meric structure 92, which should show a bathochromic shift as observed for the related compound 32, The ir spectrum of 29,1n KBr (1555, 1635, 1710, 1730 cm—1)(Ir Spectrum 5a) was different from that in carbon tetrachloride solution (1575 (small), 1610 (sh), 1635, 1710 (sh) and 1725 cm-1) (Ir Spectrum 5b), but almost identical to the ir spectrum of dehydroacetic acid both in potassium bromide (1550, 1635 and 1715 (broad) cm-l) (Ir Spectrum 6a) and carbon tetrachloride (1615 (sh), 1640 and 1724 cm-1) (Ir spectrum 6b). The difference in ir spectra of gg'in carbon tetrachloride and potassium bromide may be due to ' different concentrations of the two possible pyrone struc- tures, 92 and 22. The similarity of ir spectra of the pyrone to the ir spectra of dehydroacetic acid indicates that the . two must have almost identical structures. A possible mechanism for the formation of gg'is out- lined below, the key intermediate being ketene 93; The intramolecular reaction of the postulated ketene inter- mediate is apparently faster than attack by the nucleo— philic solvent, methanol, since no open-chain unsaturated esters which might result from such attack were isolated. 86 -Such products are common from the photolysis of 2,4-cyclo- hexadienones which do not have substituents which may act iJItramolecularly as nucleophiles (1,2,3,6). Barton (1) has shown that usnic acid, 22/ racemized by uv light in dioxane solution and he postulated a similar ketene inter- mediate which then reclosed in either of the two possible manners to give a racemic mixture of the starting compound. 87 vThe monomethyl ether of acetylfilicinic acid, 38, failed to photo-oxidize under conditions where acetyl- filicinic acid was readily oxidized to 61; Ether gg'also did not photoisomerize to a pyrone under neutral condi- tions. These facts lend some support to the proposed mechanisms for each of these photoreactions of acetyl- filicinic acid. Compound 3§'did not photo—oxidize as it could not give the required anion at carbon 4. Although ring-opening to a ketene (on irradiation of gg'in neutral solution) could have occurred, this must have reclosed as in the Barton photolysis of usnic acid, because isomeriza- tion to a pyrone would have required the loss of a methyl cation. It is perhaps surprising that a ring-opened, un- saturated methyl ester was not obtained by solvent capture of the ketene. A possibly related example for this type of ring-Open- ing and recyclization has been presented by Cookson.(44), who suggested that the photochemical rearrangement of 92' to gfi'involved a hydrogen migration, possibly gig a ketene. This reaction involves an a-cleavage (Type I) of a ketone, whereas the photoisomerization of acetylfilicinic acid, 22, can be visualized as involving either a Type I or a Type II cleavage. A rationalization is required for the observation that acetylfilicinic acid underwent photoisomerization to a pyrone in neutral solution, whereas diacetylfilicinic 88 H hv “' D’ . ma 0’ 0 Type I 95 Cleavage 22' acid, treated similarly, was photochemically inert. Perhaps cleavage to a ketene intermediate occurs in both cases. However, the rotation necessary for internal ring closure to give the pyrone is less likely with diacetylfilicinic acid, since it requires disruption of the internal chelated structure, 21; Ring closure to give the original starting compound would result in no observable change. Photoisomerization of mono- or diacetylfilicinic acid to a bicyclo[3.1.0]hexenone (the major photochemical path for hexamethyl-Z,4-cyclohexadienone; see page 3) was not observed. One possible reason could be the instability of 89 the intermediate, 98“ which would occur after electron de- motion. This intermediate would be unstable due to the location of the positive charge adjacent to the acetyl group(s). V. Results and Discussion of Kinetic Experiments A series of kinetic experiments was undertaken to as- certain the effect of various conditions upon the photo- oxidation rates. Such information would hopefully prove useful in understanding the photo-oxidation reaction mechan- ism. The reactions were followed using the uv Spectra of small aliquots. All reaction conditions were maintained 90 constant except the one being varied. In the first series of reactions 0.0105 molar diacetylfilicinic acid was used and the concentration of base was varied. The results of these experiments are plotted ianig. X, and the steady state reaction rates of these zero order reactions are given in Table VII. There was very little change in rate as the base concentration increased from 0.05 N sodium methoxide to 0.40 N sodium methoxide. Table VII. Series I reaction rates (methanol at room temp.) Experi- NaOCH3 . . Rate ment (molarity) Conditions (mole/l/hr) (x 10 3) IV A1 0.05 Air V 3 .3 IV A2 0.20 Air 3.6 IV A3 0.40 Air 3.6 IV’A4 0.05 Oxygen 3.3 IV A5 0.05 Saturation with air for 3 hrs 3.1 before irradiation IV A6 0.05 Air, 0.0026 mole of diacetyl- 4.3 filicinic acid IV A7 0.05 Air, 0.00255 mole of acetyl- 17.3 filicinic acid InH lBliBllmTl m4 >H Isle lol ms 3 l6 la lol we 5 .I..II.II md >H .InT-UIIDIa N¢K>H -|+u-+.|TI 2 En \\. \6\ cm 0 '¢ uoraoeau % O {D OOH 92 photo-oxidation as well as in the dye—sensitized photo-oxi- dation, then this would be evidence for the formation of singlet oxygen in both reactions. In these reactions an 0.00525 molar solution of diacetylfilicinic acid in 200 ml of methanol was photo-oxidized with a Hanovia L 450 watt lamp. Results were plotted (Fig. XI) and the rates are given in Table VIII. When dimethylfuran was added (EXp. IV B5, IV B9), an initial rapid trapping of singlet oxygen occurred. However, the subsequent photoreaction appeared to occur relatively quickly. The dimethylfuran, or its photo-oxidation product(s) were thought to react with the diacetylfilicinic acid under the influence of uv radiation. Schenck and Foote (45) have shown that the photo-oxidation of dimethylfuran yielded a cyclic peroxide which was hydro- lyzed in acid to 3-hexene—2,5—dione. A similar reaction was carried out to check the reac- tivity of singlet oxygen to 2,5-dimethylfuran in alkaline methanol. A change in the uv absorption from A 217 mu to A 213 and 282 mu occurred on photo-oxidation of the 2,5-di- methylfuran. This was followed by a decrease in intensity of these absorption bands which were replaced by an absorp- tion band at A 230 mu. This phenomenon is interpreted to mean that the dimethylfuran was photo—oxidized to gig-3- hexene-2,5—dione (uv xfing.223 (6600) and 292(174)mu (46)). which subsequently isomerized to the trans isomer (uv AEEEH 228(14,400) and 324 (70) mu (46)). Conceivably the gig or trans form of 3-hexen-2,5-dione was reacting with the 93 .mEmH pumz omv A mH>ocmm nufl3 .HH mwflnmm “moflumcflx mo muHSmmH UH£QMH0 .HX .mHm Aura mass GP 0 N lunlnfol mm a \o\ . . .1 mm 5 \D\ \ 1 I!0.l0|lo|:Hm.>H \\\\\\b \\\\\\\ \\.- o \\\\\\ \\\\\\\ I o ‘v uorqoeau % 1 O CO _ loos 94 cmummcsu 66.6 666 666.6 6666.6 6663 666 .6666I666V 666 >6 6663 Hm.o 666 woa.o mmoo.o omv .cmm. .mcmmGD mm >H :mummcsp m.oa 666 tho.o 6663 com .mcwmlmmn mm >H cmummcsu 6.6 666666 6666.6 6666.6 6663 666 .6666-666 66 >6 Gmummcsu 6.6 666 6666.6 6666.6 6663 666 .6666-666 66 >6 6663 66.6 666 6666.6 6666.6 666 .666 .666666 66 >6 cmpmmcsu 66.6 666 6666.6 6663 666 .6666u666 66 >6 cmummcsu 66.6 666 6666.6 6663 666 .666666 66 >6 6663 mm. H HHN mwoo. O omV . Gmm . mflmmCD, Hm >H A6 OH xv x66mwmmo6osc 6666666666 xwm6wwmmmw6mMmsv mhwmvmmwomm 6666 66666666 ”mmm HOCMQHGE CH 6mm .6606 06660666666666666 66666 66666 66666666 66 666666 .HHH>.OHQME 95 diacetylfilicinic acid to give unidentified product(s). For the dye-sensitized reactions using Rose Bengal a marked decrease in rate was noticed on the addition of 2,5-dimethylfuran (Exp. IV B6, IV BIO). The photoeoxidation rate in the presence of 0.0104 mole of the dimethylfuran was initially slow, then increased, then slowed down after approximately an hour. This would fit the tentative ex- planation that the rate increased as the concentration of dimethylfuran decreased by reacting with the singld:oxygen. The rate then decreased as the isomerization from the gig to trans 3-hexen-2,5-dione occurred (see Fig. XIV for a plot of the absorption at A 282 mu versus time). vThis isomerization may be causing the quenching of the triplet state of the dye, or the singlet oxygen moiety, resulting in a rate retardation. When a large amount of furan was used to trap singlet oxygen the rate was slower by a factor of 12 (Exp. IV 310). This indicates that the singlet oxygen was reacting with the furan rather than the diacetylfilicinic acid. Although furan caused a rate retardation in the dye- sensitized photo-oxidation, the same effect was not observed in the non-sensitized photo-oxidation of diacetylfilicinic acid (Exp. IV B5, IV B9). Singlet oxygen was trapped only in the initial stages of the latter reaction, as can be seen in Fig. XI. An increase in the base concentration caused only a Slight increase in the dye-sensitized photo-oxidation rate 96 .coflumUHXOIouozm U06666mcmmlmmo .Il allm.lsu.ll. mm >H IIO.I.O.E-®.|! OHm >H '10.... .O :1. 0 75 mm >H iI,-Q1 --Diiaial hm >H til+!|+i|.r x. mm >H . II . . mm >H 4) .HH moflnmm “moflumcflx NO 6663606 UHSQMHO A666 6666 .HHx .666 O N uorqoeau % ow loOH 97 + t .. -IlDiIIDIlDiJ va >H m .\ ION ... \ II TIL. II +1 66m >6 .. o\ .UHUm oacfloflaflmamumom .HH mmflumm “moflpmsflx mo m6HSmmH UHQQMHO .HHHX mudmflm Auzv 0868 N H "o 66 H 3 ...ii? .I 66m >6 \ 166m . u 1 \ l I. Isololloil :6 >6 \ o m \ . \ i \ \ tom 0. . o . \.. 10w 98 Fig-6- XIV . — (- y— Time (hr) Plot of ultraviolet absorption of 3-hexene-2,5- dione during the photo-oxidation of dimethylfuran (at Amax 282mu). of diacetylfilicinic acid (from 9.6 x 10—3_moles/l/hr to 10.9 x 10‘3 moles/l/hr)(Exps.-IV 33 and IV B8). The photo-oxidation of acetylfilicinic acid was faster by a factor of 6.2 than the photo-oxidation of diacetyl- filicinic acid, when using a Hanovia L 450 watt lamp (Exp. IV B11). 'This could result from a more favorable excited state for acetylfilicinic acid than diacetylfilicinic acid. Table IX. Series II reaction rates using acetylfilicinic acid in methanol Exp. . NaOCH3(mole Rate No. ReaCtlon Lamp in200 ml) (mole/l/hr) (x 10 3) IV 811 .Unsens. -Han. 450 0.0093 ‘ 8.1 watt IV 312 Dye—sens. 200 watt 0.0093 5.9 tungsten IV B13 Unsens. Han. 450 0.000128 2.35 watt IV 314 Dye-sens. 200 watt 0.000128 3.8 The rate of the dye-sensitized.photo—oxidation of 22’ using a 200 watt tungsten lamp (5.9 x10.3 moles/l/hr) was only slightly less than the equivalent rate for diacetyl- filicinic acid (9.6 x 10‘3 IV’B3). moles/l/hr) (Exps. IV 812 and These comparable rates might be expected as the reaction only depends upon the rate of production of singlet oxygen, which would be constant in both reactions, and the subsequent attack of this species on the enol anions. 100 When the base concentration was decreased to one molar equivalent a marked decrease in the rate (3.45 times slower) for the direct photo-oxidation of acetylfilicinic acid occur— red (EXp. IV B13). This reaction not only gave the normal photo-oxidation product, 61, but also a new product, 66; For the dye-sensitized photo—oxidation at low base conCen- tration, the rate was only slightly decreased (1.55 times) and only the normal photoproduct, 62, was obtained (Exp. IV 314). A question which remains regarding these reactions was the determination of the rate-determining step. If -S represents both the anion and sensitizer then the reaction can be expressed by the following series of equations: {D S + hv > 1S k. 15 __1_s_<.=_.> 3s 3s + 302 ,> 133-~302 35...302 .__§2_9 S + 102 k S + 102 ‘—_r'L> 802 kr _802 ---—3? products k (Furan + 102 > Furan peroxide > Furan products) Three factors are likely candidates for the slow.step in the reaction. These are the photochemical step, the rate of diffusion of oxygen in the solution, or the subsequent chemical reaction, one step of which is the cleavage of the 101 acetyl group from diacetylfilicinic acid or the hydrogen from acetylfilicinic acid. As the rate for the direct photo-oxidation of di- acetylfilicinic acid did nOt change very much with increas- ing base concentration, cleavage of the acetyl group by base could not be the rate-determining step (see Table VII). Also the dye—sensitized photo-oxidation rate was much faster than the self—sensitized rate, showing that cleavage by base was not rate-limiting. When oxygen was used in the place of air, effectively increasing the oxygen content by about five times, no major change in the rate of the photo-oxidation occurred. Thus it is unlikely that the rate of diffusion of oxygen through the solution is the controlling factor. The most probable rate-controlling step is the initial quantum yield, which would involve the initial excitation and intersystem crossing to the triplet state, in a manner suggested in Section III. For the previously mentioned series of equations it can be shown that in the steady state the rate of photo- oxidation can be represented by: d(product) =' dt 10” However if the base concentration were to become low enough the rate of cleavage of the acetyl group, or loss of a pro- ton might become slow enough to be the rate—determining step of the reaction. This was observed in the photolysis of acetylfilicinic acid (Table IX). ized are: 102 The results of these kinetic reactions can be summar— aS follows: 1. The Acetylfilicinic acid undergoes direct photo-oxida— tion faster than diacetylfilicinic acid. Diacetylfilicinic acid undergoes dye—sensitized photo—oxidation slightly faster than acetylfili- cinic acid. The effect of changing the base concentration was almost unnoticeable unless the basic concentration was low. The rate of photo-oxidation of diacetylfilicinic acid was decreased by adding 2,5-dimethylfuran. The rate of reaction was independent of the oxygen concentration. conclusions which can be drawn from these results The photochemical step(s) in the self—sensitized photo—oxidation of acetylfilicinic acid, 66” is better than in the self-sensitized photo-oxidation of diacetylfilicinic acid, 66“ perhaps because of a more favorable excited state. The cleavage of the acetyl group in the mechanism (page 70) is unlikely to be the rate-determining step of the reaction under the normal conditions. When dimethylfuran was added to the reaction it preferentially reacted with the singlet oxygen. 103 A mechanism which is consistent with outlined here. by ME 302 2?; > > slow or or $2.5. £32. O U" 01 w H O m w I 0: these results is EXPERIMENTAL I. General Procedures All ultraviolet Spectra were measured with a Unicam Model SP-800 recording Spectrophotometer in 95% ethanol solution unless otherwise stated. When Spectra are referred to as having been measured in base or acid, the basic solutions were obtained by addition of a small amount of sbdium methoxide, and the acidic solutions by the addition of one or two drops of 2N hydrochloric acid to the ethanol. The term pH refers to the value calculated from the amount of acid (H2804) or base (NaOH) added to a water solution. Unless otherwise stated,all uv Amax are reportedas mu, with the extinction coefficient, 6, given in parentheses. The infrared spectra were obtained on a Unicam Model SP—200 Spectrophotometer, calibrated with polystyrene. The nuclear magnetic resonance spectra were obtained on a Varian A-60, a Varian HA—100 or a Jeolco C-60H spectrometer, with tetra- methylsilane as an internal reference. Melting points were taken with a Gallenkamp melting point apparatus and are un— corrected. The photolyses were carried out with a Hanovia L 450 watt or a Hanovia S 200 watt lamp with a Pyrex filter. Photo-oxidationsunder air or oxygen were obtained by bubbling air or oxygen through the reaction solution. The dye-sensi- tized photo-oxidations were carried out with a 200 watt tungsten lamp in a 250—ml 3—necked flask equipped with a 104 105 reflux condenser and a gas inlet tube. Elemental analyses were performed by Spang Microanalytical Laboratories, Ann Arbor, Michigan. Sublimations were carried out at 0.5 mm of mercury a few degrees below reported melting points. Reactions followed by uv were sampled with a 50 ul syringe; a constant amount of reaction miXture was injected into 3 ml of solvent (95% ethanol) in a 1 cm quartz cell. The disappearance of absorption bands of the reactant and/or the appearance of absorption bands for the product(s) were monitored. For example the uv absorption wavelengths used for the monoanion of diacetylfilicinic acid, 66, was Am ax 340 mu: for the monoanion of acetylfilicinic acid, 66, Amax 345 mu: for acetylfilicinic acid, 66, Amax 330 mu: and for the photo-oxidation product 2-acetyl-3,4-dihydroxy—4-carbo- methoxy-5,5-dimethyl-2-cyc10pentenone, 61, Amax 270 and 250 mu in basic ethanol. When reaction percentages are quoted, these were calculated from the relative intensities of the longest wavelength absorption bands obtained from Spectra at the beginning and at the end of the reaction. II. Preparation of Filicinic Acid and Derivatives A. Preparation of 1,3-Diacetylphloroglucinol, ’26 1. Using Boron Trifluoride as Catalyst In a 100-ml 3-necked flask equipped with a gas inlet tulbe and a reflux condenser were placed 10 g (0.0795 mole) fo phloroglucinol and 50 ml of glacial acetic acid. The 106 phloroglucinol was dissolved by warming themfixmure. With- out cooling, boron trifluoride gas was bubbled through the solution for one hour with stirring. The reaction mixture was poured into about 400 ml of water and allowed to stand several hours. Filtration afforded 13.6 g (81.5%) of crude 1,3-diacetylphloroglucinol, 22; This crude material Was recrystallized from 50% aqueous methanol. After sublima- tion at 150° and 1 mm of mercury white crystals of the pro- duct were obtained (mp 167.5-169°, lit. 168° (12)). The infrared spectrum of gg'had absorption bands at —1 Vmax 1610, 3000 and 3450 cm 2. Usinq_Concentratedgulfuric Acid as Catalyst a. Small_§ga1e Reaction.- Ten grams (0.0795 mole) of oven dried (110°) phloroglucinol was added to 16.16 g (0.158 mole) of acetic anhydride. After the addition of 1.5 ml of concentrated sulfuric acid the temperature rose to 97°. The reaction mixture was refluxed at 135° for five minutes, then 40 ml of 30% concentrated hydrochloric acid in methanol was carefully added. The mixture was poured into 300 ml of cold water. Crystals precipitated immediately and filtration gave 12.2 g of crude product (mp loo-107°). Recrystallization from 50% methanol in ‘water yielded 6.3 g (37.7%) of an orange crystalline pro— duct (mp 160—166°). After sublimation at 150° and 1 mm of mercury the mp rose to 168-170°. 107 b. Large Scale Reaction.- In a 250-ml 3-necked flask equipped with a reflux condenser was placed 40 g (0.317 mole) of dry phloroglucinol. Acetic anhydride (64.3 g, 0.63 mole) was added followed by 4 ml of concentrated sulfuric acid. The temperature rose to 95° after which the mixture was re- fluxed at 135° for five minutes. A 50 m1 mixture of 50:50 concentrated hydrochloric acid and methanol was added care— fully to the hot reaction mixture. After the mixture was poured into one liter of water, the orange solid which sep- arated was filtered. Recrystallization from 50% aqueous methanol yielded 40.2 g (61.5%) of 1,3-diacetylphlorog1ucinol. It had a mp of 168-170° after sublimation. B. Preparation of 2,4-Diacetyl-3,5-dihydroxy-6,6-dimethyl- 2,4-gyclohexadienoné(Diacety1filicinic Acid, 21. A solution of 20 g (0.37 mole) of sodium methoxide and 200 ml of anhydrous methanol was cooled in an ice-water bath. Methyl iodide (40 g, 0.282 mole) was added slowly, followed by 20 g (0.095 mole) of 1,3-diacetylphloroglucinol in one portion. The mixture was allowed to stand with oc— casional stirring from two to four days at room temperature. The methanol was removed under water vacuum and the residue dissolved in 200 ml of water. After acidification, this mixture was extracted twice with ethyl ether. The ether solution was extracted several times with 50-ml portions of saturated potassium bicarbonate solution until a clear extract was obtained. The sodium bicarbonate solution was 108 neutralized and allowed to stand until the reddish-brown oil crystallized. After filtration, the reddish-brown solid was triturated with 30 m1 of cold benzene and filtered. The filtered solid was identified as recovered starting material by its uv spectrum. The benzene mother liquor was evaporated under water vacuum on a rotary evaporator and the residue recrystallized from methanol. After vacuum distillation (170° and 0.5 mm pressure) yields varying from 30% to 60%, based on reacted starting material, were ob- tained in a series of reactions with about 50% conversion. For example in one reaction 12.3 g of starting material was recovered and 5.7 g (65.5%) of recrystallized and dis- tilled diacetylfilicinic acid was obtained. After sublima— tion the diacetylfilicinic acid had a melting point of 63- 66° (lit. 65-66° (12)). vKBr max Infrared spectra: 1560 and 1665 cm—1 vCC141580 and 1660 cm"1 max vCHCla 1560 and 1660 cm”1 max . EtOH . . UltraViolet spectra: A max’ aCidic, 300(13,800) and 245(12,350) xEtOH, basic, 340(20,600), 275 max (11,600)and 220(15,200). The nuclear magnetic resonance spectrum is given in Table I. 109 1. Attempted Preparation of a N-Phenylmaleimide Deriva— tive of Diacetylfilicinic Acid {21 Diacetylfilicinic acid (1 g, 0.0042 mole) and 0.73 9 (0.0042 mole) of N-phenylmaleimide were refluxed in ethyl ether overnight. Only gl'was recovered as identified by its melting point and nmr spectrum. 2. Attempted Acetylation of Diacetylfilicinic Acid [2%. a. Using Acetyl Chloride.- Diacetylfilicinic acid (1 g, 0.0042 mole) and 1 g of acetyl chloride were heated on a steam bath for 15 minutes. After water was added, un— changed gg'crystallized from the solution. The yield of recovered starting material was 0.8 9; it was identified by its melting point. b. Using Acetic Anhydride.- Acetic anhydride (5 ml) and 0.8 g of diacetylfilicinic acid were mixed in a 4" test tube. After the addition of one drop of concentrated sul- furic acid the mixture was heated on a steam bath, then added to water. Only starting material was isolated, and identified by its melting point and nmr spectrum. C. Preparation of 3,5-Dihydr9xy-4,4-dimethyl-2,5-cyclohexa- dienone (Filicinic Acid, 12) Diacetylfilicinic acid (5 g, 0.021 mole) was refluxed for six hours in 120 ml of 2N hydrochloric acid. When most of the oil had dissolved the hot solution was filtered and concentrated to 10-20 ml under a water vacuum on a rotary 110 evaporator. A solid product crystallized from the solution on standing overnight. The precipitate was filtered and washed with a few ml of ethyl ether, yielding 2.7 g (67%) of filicinic acid (mp 215-2170, after sublimation 220-2210, lit. 212-2150 (12)). Infrared spectrum: viii, 1630 and 2500-3000 cm- 1 H20 max, pH = 11, 348(9340), 250 Ultraviolet spectra: X (18,000) and 231(21,300). H20 = - Amax, pH 5 10, 278(20,000) H2O - _ mmax, pH 1.5 3, 251(14,300) 13:3, 10% sulfuric acid, 343(7550) and 243(16,100). The nuclear magnetic resonance spectrum is given in Table I. D. Preparation of the Diacetates of Filicinic Acid One gram (0.0065 mole) of filicinic acid and 6.2 9 (0.0608 mole) of acetic anhydride with two drops of pyridine were heated in a test tube on a steam bath for 14 hours. The excess acetic anhydride and acetic acid was removed under water vacuum on a rotary evaporator. The residue was vacuum distilled at 185° and 0.5 mm pressure to give 0.95 g of crude material. One half of this material was recrystal- lized from carbon tetrachloride and 0.1 g of the monoacetate of filicinic acid, 26” was obtained (mp 142-144°). Infrared spectrum: v§2i4 1660 and 1775 cm—1. The nuclear magnetic resonance spectrum is given in Table I. 111 After the carbon tetrachloride was evaporated from the mother liquor of the filicinic acid monoacetate, the resi- due obtained was recrystallized from petroleum ether—ethyl ether to yield 0.1 g of filicinic acid diacetate, 21 (mp 85°, lit. 82-85° (17)). RBI Infrared spectrum: Vmax 1660 and 1775 cm- 1 The nuclear magnetic resonance spectrum is given in Table I. E. Reaction of Diacetylfilicinic Acid with Diazomethane A solution of 30 ml of ethyl ether and 12 ml of 30% potassium hydroxide was cooled in an ice-salt bath to less than 5° and 1.5 g N-nitrosomethylurea was added slowly. After this decomposed to form a yellow ethereal solution of diazomethane, the ether layer was decanted into an erlen— meyer flask containing 1 g of diacetylfilicinic acid in 10 ml of ethyl ether. When the evolution of nitrogen ceased and the solution remained yellow the excess ether was evapor- ated. An nmr spectrum of the residue indicated the presence primarily of the monomethyl derivative of diacetylfilicinic acid. A uv spectrum had absorption bands in neutral 95% ethanol at xmax 320, 278 and 235 mu and in basic ethanol at kmax 305, 278 and 215 mu. When an ether solution of this material was extracted with concentrated aqueous potassium bicarbonate and 10% aqueous sodium carbonate solutions, the product isolated had an nmr spectrum identical to diacetylfilicinic acid. 112 F. Preparation of Acetylphloroglucinol, 31' 1. From Phloroglucinol and Acetonitrile Anhydrous phloroglucinol (20 g, 0.158 mole) was mixed with 13 g (0.317 mole) of anhydrous acetonitrile, 80 ml of anhydrous ethyl ether and 4 g of anhydrous zinc chloride. The mixture was cooled in an ice-salt bath while hydrogen chloride was passed through for two hours. -The mixture was placed in a refrigerator for three days, then filtered and the solid washed twice with 20-m1 portions of ethyl ether. The solid was refluxed for 3.5 hr in 1 liter of water, then cooled to give 21.6 g (81.6%) of acetylphloroglucinol (mp 215-218°, lit. 218-219° (18)). 2. From Phloroglucinol and Acetic Anhydride Phloroglucinol (10 g, 0.079 mole) was added to 8 g (0.079 mole) of acetic anhydride. When 8 drops of concen- trated sulfuric acid were added in one portion the tempera- ture rose to 100°. The mixture was refluxed at 135—140° for 2-3 min, then 10 ml of concentrated hydrochloric acid in 30 m1 of methanol was carefully added. After the mix- ture was poured into 400 ml of water, 6.0 g of crude product precipitated (mp l95-208°). Recrystallization from 50% aqueous methanol yielded 5.5 g (41%) of orange colored crystals. Sublimation gave white crystals of acetylphloro- glucinol (mp 213—218°, lit. 218-219° (18)) Infrared spectrum: vgggl3 1610, 3200 and 3550 cm-1 113 G. Preparation of 2-Acetyl-3,5—dihydroxy-4,6,6-trimethyl- 2,4-cyclohexadienone (Acetylmethylfilicinic Acid, 22) Acetylphloroglucinol (18 g, 0.107 mole) and 17.2 g (0.319 mole) of sodium methoxide in 150 ml of anhydrous methanol was cooled in an ice bath while 91.4 g (0.64 mole) of methyl iodide was added slowly. The mixture was stirred for 5 days at room temperature. The methanol was removed under water vacuum on a rotary evaporator, and the residue dissolved in a mixture of 150 ml of water and 50 m1 of ether. After neutralization with hydrochloric acid the aqueous layer was extracted twice with ethyl ether. The ether portions were combined and extracted with eight 30- ml portions of saturated potassium bicarbonate solution. Neutralization gave 9.55 g of a solid material which was . washed with 180 ml of cold benzene to yield 8.2 g (36%) of acetylmethylfilicinic acid (mp 157°, lit. 160-161° (15)). Infrared Spectrum: Vii: 1600, 1660 and 3300(broad) cm-l. Ultraviolet spectra: 13:2, pH less than 2.5, 240(9120), 279(7000) and 335(12,100). H20 = “max' pH 4.5 to 10, 350(18,500), and 222(16,000). H20 kmax' pH greater than 12, 331(13,400) and 250(21,000). ' The nuclear magnetic resonance Spectrum is given in Table I. 114 The benzene mother liquor was evaporated and yielded 1.4 g of a material identified as 2-acetyl-3-hydroxy-4,4,6,6- tetramethyl-2-cyclohexen-1,57dione, 32; VCHC 1 3 1 Infrared spectrum: 1560, 1660, 1710 and 3400 cm‘ . H20 UltraVlolet spectra: ax , acidic, 277(10’700) and 241(8500). H20 7‘max’ basic, 277(17,400). The nuclear magnetic resonance spectrum is given in Table I. This compound was refluxed in 2N hydrochloric acid for 5 hr to give 0.6 g of 3-hydroxy-4,4,6,6-tetramethyl-2-cyclo- hexen-1,5-dione, §§'(mp 189°, after sublimation 192-193.5°, lit. 187—190° (15)). KBr Infrared spectrum: Vmax 1610, 1700 and 3000 cm-1 Ultraviolet Spectra: lggg, acidic, 258(16,200). AHZO, basic, 279(17,850). max The nuclear magnetic resonance spectrum is given in Table I. Compound §£'(0.8 g, 0.0036 mole) was mixed with 2 ml of acetic anhydride and 4 drOpS of pyridine in a small test tube, then heated on a steam bath for 5 min. The mixture was dissolved in ethyl ether and washed with aqueous solu- tions of saturated potassium bicarbonate and 2N hydrochloric acid. The ether solution was dried with sodium sulfate. Evaporation gave 0.55 g of the acetate derivative (56%). 115 After sublimation at 50°, 0.5 mm, a white crystalline solid was obtained (mp 54-57°). Infrared spectrum: VCHC13 1660, 1710 and 1770 cm"1 max The nuclear magnetic resonance spectrum is given in Table I. The original ether layer in the preparation of acetyl- methylfilicinic acid was extracted with eight 50-ml portions of 10% aqueous sodium carbonate. Neutralization of the basic solution gave 7.0 g of acetylmethylphloroglucinol, §§'(mp 185-205°, lit. 211° (15)). This material was treated with sodium methoxide and methyl iodide in methanol as be- fore to give, after work-up, an additional 2.6 g of acetyl- methylfilicinic acid (mp after sublimation, 158-162°). H. Preparation of 2,4,4-Trimethyl-3,5-dihydroxy-2,5-cyclo- hexadienone (Methylfilicinic Acid, 32) Acetylmethylfilicinic acid, 22'(8.2 g, 0.039 mole) was added to 100 ml of 2N hydrochloric acid. The mixture was refluxed for 5 hrs, filtered while hot and extracted with ethyl ether. After the ethereal layer was dried and evapor- ated 4.2 g (64%) of crude methylfilicinic acid was obtained. Sublimation afforded a white crystalline material with a mp 174-1770 (lit. 180° (15)). Infrared spectrum: vii: 1650 and 3100 (broad) cm-l. Ultraviolet Spectra: 0322' pH greater than 11, 358(10,100), 255(18,800) and 227(20,300). \HZO , , pH 5-10, 285(13,000). max 116 1:33, pH 0.8-3.3, 265(11,300) and 255(11,200). .HZO . h 10a f . Amax, greater t an p sul uric acid, 350(7030), 285(2580) 248(8300) and 208(19,100). The nuclear magnetic resonance spectrum is given in Table I. One gram of methylfilicinic acid was dissolved in 5 ml of acetic anhydride and 5 drops of concentrated sulfuric acid. The temperature rose to 70° and the mixture was al- lowed to stand for 5 min. After the mixture was poured into water and extracted with ethyl ether, the organic layer was washed with water and 10% aqueous sodium carbonate solution. The ether layer was dried and evaporated to give a small‘ amount of residue which, judging by its nmr spectrum (Table II), was the diacetate of methylfilicinic acid, 36. I. Preparation of 2-Acetyl—3-hydroxy—5-methoxy-4,6,6-tri- methyl-2,4—cyclohexadienone, 31; An ether solution of diazomethane was prepared by add— ing 2.5 g of N—nitrosomethylurea to a mixture of 15 ml of 30% potassium hydroxide and 30 ml of ethyl ether cooled to below 0° in an ice-salt bath. The yellow diazomethane- ethyl ether solution was decanted into a methanol—ether solution containing 1 g of acetylmethylfilicinic acid. The diazomethane solution was added until the nitrogen evolution ceased. After one hour the excess ether was evaporated. 117 The residue was found to be the desired monomethyl ether, EZX Infrared spectrum: v;::4 1635 and 1655 cm"1 Ultraviolet spectra: Agzgfl, acidic and neutral, 323- 340. AEESH, basic, 285—310. The nuclear magnetic resonance spectrum is given in Table II. J. Preparation of 2-Acetyl-3,5-dihydroxy-6,6—dimethyl-2,4- cyclohexadienone (Acepylfilicinic Acid, 22) 1. Standing Eighteen Days at Room Temperature Diacetylfilicinic acid (10 g, 0.042 mole) was added to 1 liter of methanol and 10 g (0.185 mole) of sodium meth— oxide. After 18 dayS at room temperature the methanol was evaporated under water vacuum on a rotary evaporator. The residue was dissolved in water, neutralized with concentrated hydrochloric acid and extracted with ethyl ether. The ether layer was dried with anhydrous sodium sulfate and evapor- ated. The residue was triturated with carbon tetrachloride and filtered to yield 5.27 g (94% based on reacted starting material) of acetylfilicinic acid (mp 170-172.5°, lit. 174- 176° (19)). Evaporation of the carbon tetrachloride yielded 3.23 g of recovered diacetylfilicinic acid. KBr v Infrared spectrum; x 1540-80, 1630 and 2600-3000 cm—1 118 Ultraviolet Spectra: kfiggH, acidic, 330(10,100), 271 (9,290), 235(15,500) and 204 (13,200). 1EtOH, basic, 345(18,000), 310sh max (14,200) and 218(18,600). The nuclear magnetic resonance spectrum is given in Table II. 2. From Refluxing Methanol and Sodium Methoxide Diacetylfilicinic acid (2.5 g, 0.0105 mole) was added to 3 g (0.0555 mole) of sodium methoxide in 600 ml of anhy- drous methanol. The mixture was refluxed for 18 hrs and worked up as before to give 1.45 g (70%) of acetylfilicinic acid. The product was contaminated with a trace of the completely deacetylated product, filicinic acid,as determined by removal of the acetylfilicinic acid by sublimation at 160-170° and 0.5 mm of mercury. The residual filicinic acid was identified by its mp of 205-215°. Higher base con— centrations or longer reaction times increased the amount of filicinic acid produced. K. Preparation of 2-Acetyl-3—hydroxy-5-methoxy-6,6-dimethyl- 2,4-pyclohexadienone (Monomethyl_Ether of acetylfili- cinic Acid): An ether solution of diazomethane was prepared by add- ing 2 g of N-nitrosomethylurea to an ice—cold mixture of 6 ml of 40% aqueous potassium hydroxide and 20 ml of ethyl ether. The diazomethane-ether solution was decanted into 119 an ether solution of 0.7 g of acetylfilicinic acid. After 20 hrs the ether was evaporated and gave 0.6 g of a sub- stance which was 97% one component as shown by gas liquid phase chromatography. This was identified as the mono- methyl ether of acetylfilicinic acid (mp 103°, lit. 107-109° (27)). Infrared spectrum: vggi4 1620 and 1660 cm-1 Ultraviolet spectra: liégfi, acidic, 315 and 235. iEtOH, basic, 305-10, 278(sh) and max . 235. The nuclear magnetic resonance Spectrum is given in Table II. L. Preparation of 2,4— and 4,6-Diacetylresorcinols Resorcinol (11 g, 0.1 mole) was dissolved in 25 g (0.245 mole) of acetic anhydride and 3 ml of concentrated sulfuric acid. The mixture was refluxed for 10 min, then treated with 40 ml of methanol and concentrated hydrochloric acid (2:1). On addition of water, a mixture of the two diacetyl- resorcinols crystallized out. The solid material was re— crystallized from about 30 ml of hot acetone to yield 1.7 g of crude brown crystals (mp 179-181°, lit. 182° (21)) of the 4,6-diacetylresorcinol. A second crop of 1.3 g was obtained from the mother liquor for a total yield of 3.0 g (15.5%). The residue obtained by evaporation of the acetone mother liquor yielded, after fractional sublimation, mainly 120 2,4-diacetylresorcinol, (1.0 g (5%), mp 87-900, lit. 92° (21)) as identified by its nuclear magnetic resonance Spec- trum (Table II). M. Preparation of 2,4-Diacetyl-5—hydroxy-6,6-dimethyl-2,4— cyclohexadienone, 41 Methyl iodide (7.1 g, 0.05 mole) was dropped into an ice cold solution of sodium methoxide (3.55 g, 0.066 mole) and 30 ml of anhydrous methanol. 4,6-Diacetylresorcinol (2.5 g, 0.05 mole) was added in one portion. »After 4 days at room temperature with occasional stirring, the methanol was removed under water vacuum on a rotary evaporator. The residue was dissolved in a mixture of 150 ml of water and 50 ml of ethyl ether, neutralized with concentrated hydro- chloric acid and extracted twice with ethyl ether. The ether layer was extracted eight times with 40-ml portions of saturated potassium bicarbonate solution. After evapOra- tion of the ether, 1.15 g of a light yellow material re- mained, identified as primarily 2-methyl-4,6-diacetyl- resorcinol by its nuclear magnetic resonance spectrum (Table II). Neutralization of the potassium bicarbonate fraction yielded 1.15 g (40%) of 2,4-diacetyl-5-hydroxy-6,6-dimethyl— 2,4-cyclohexadienone. The compound was recrystallized from ethanol-water, redissflyed in base and reprecipitated, and sublimed (mp 134-137°). Infrared spectrum: vii: 1670 and 1560 cm-1. 121 Ultraviolet spectra: zfiggH, acidic, 320(12,650) and 340(5,930). lEtOH, basic, 400(22,600) and max 278(15,900). The nuclear magnetic resonance Spectrum is given in Table II. Apal. Calcd. for C12H14O4: C, 64.85; H, 6.35. Found: C, 64.91; H, 6.24. N. Preparation of 2-Methyl-4,6-diacetylresorcinol, 42' 2—Methylresorcinol (5 g, 0.04 mole)(Aldrich Chemical Co.) was added to 8.22 g (0.08 mole) of acetic anhydride and 2 ml of concentrated sulfuric acid. The mixture was refluxed at 140° for 10 min then treated with 30 m1 of methanol and concentrated hydrochloric acid (1:1). This mixture was poured into 200 ml of water. After one hour, filtration afforded 5.3 g (63.6%) of a reddish brown solid. Sublimation yielded white crystals of 2-methyl-4,6-diacetyl- resorcinol (mp 132-133.5°). It was identified by the nuclear magnetic resonance Spectrum given in Table II. 0. Preparation of 2,4-Diacetyl-5-hydroxy-6,6-dimethyl-2,4- cyclohexadienone, 41, from 2-Methyl-4,6-diacetylresor- 2.21.21. I" Sodium methoxide (3.22 g, 0.06 mole) and 50 ml of an- hydrous methanol were cooled with ice water and stirred as 6.3 g (0.0445 mole) of methyl iodide was added dr0pwise. 122 2-Methyl-4,6-diacetylresorcinol (3.1 g, 0.015 mole) was added in one portion. After one week the mixture was worked up as in Experiment II-M. From the saturated potassium bi- carbonate fraction 0.65 g (56%‘ based on the reacted start- ing material) of 2,4—diacetyl-5-hydroxy-6,6-dimethyl-2,4- cyclohexadienone was obtained. This compound was identified by its nuclear magnetic resonance spectrum (Table II) as in Experiment II-M. Evaporation of the ether fraction yielded 2.0 g of recovered 2-methyl-4,6-diacetylresorcinol. P. Preparation of 42Acetylresorcinol, 42' Resorcinol (30 g, 0.272 mole), 12 g of anhydrous zinc chloride, 18 g (0.44 mole) of acetonitrile and 150 ml of anhydrous ethyl ether were stirred with a mechanical stir- rer as hydrogen chloride was bubbled in for 70 minutes. After 12 hrs the solid material was filtered, washed with ether and refluxed in 500 ml of water for 2.5 hrs. 4—Acetyl- resorcinol (17 g, 41%) precipitated from the aqueous solu- tion after several hours (mp 145°, lit. 147° (18)). It was identified by its nuclear magnetic resonance Spectrum, given in Table II. Q. Preparation of 4—Acety1-5-hydroxy-6,6-dimethy1-2,4- cyclohexadienone , 4,4, 1. From 2,4-Diacetyl-5-hydroxy-6,6-dimethyl-2,4-cyclo- hexadienone, 41 2,4-Diacetyl-5-hydroxy-6,6-dimethyl-2,4-cyclohexadienone 123 (1 g, 0.0045 mole), 3 ml of methanol and 30 ml of 2N hydro- chloric acid were refluxed for 3.5 hrs. :Starting material (0.2 g) crystallized on cooling the aqueous phase. Extrac— tion of the water with ethyl ether resulted in a small amount of crude oily material identified as 4-acetyl-5-hydroxy-6,6- dimethyl-2,4-cyclohexadienone. Infrared Spectrum: vizit 1660 and 1605 cm-1 The nuclear magnetic resonance Spectrum is given in Table II. 2. By Methylation of 4eAcetylresorcinol To an ice-cold solution of 23 g (0.425 mole) of sodium methoxide and 120 ml of anhydrous methanol was added 61 g (0.43 mole) of methyl iodide. 4—Acetylresorcinol (16.2 g, 0.108 mole) was added in one portion. After three days at room temperature the methanol was evaporated under water vacuum on a rotary evaporator. The residue was dissolved in a mixture of 150 m1 of water and 50 ml of ethyl ether, neutralized with concentrated hydrochloric acid, and ex- tracted twice with ethyl ether. The ether fraction was extracted twice with 40-ml portions of saturated potassium bicarbonate solution. Neutralization and extraction with ether gave 0.25 g (1.3%) of a crude oil which was identified as 4-acetyl-5-hydroxy-6,6-dimethy1-2,4-cyclohexadienone, 22x by its nmr (Table II) and ir spectra. The ether layer was extracted 8 times with 50—ml por- tions of 10% sodium carbonate solution. Neutralization 124 yielded 4.0 g of a material identified as 2-methy1-4-acetyl- resorcinol. It was recrystallized from carbon tetrachloride, and a sublimed sample gave a mp of 151-155° (lit. 156-7° (22)). Infrared Spectrum: vgggl3 1620, 3000 and 3300 cm-1. The nuclear magnetic resonance spectrum is given in Table II. R. Attempted Methylation of 2,4-Diacetylresorcinol A solution of 3.55 g (0.06 mole) of sodium methoxide and 50 ml of methanol was cooled in an ice bath as 7.1 g (0.05 mole) of methyl iodide was added. 2,4-Diacetylre- sorcinol (2.5 g, 0.013 mole) was added in one portion and the mixture stirred for 3 days at room temperature. The methanol was removed under water vacuum on a rotary evapor- ator. The residue was dissolved in water, neutralized with concentrated hydrochloric acid and extracted twice with ethyl ether. Two grams of starting material were recovered. III. Photolyses and Photo-oxidations A. The Photolysis and Photo-oxidation of 2,4-Diacetyl-3,5- dihydroxy-6,6-dimethyl-2,4-cyplohexadienone (Diacetyl- filicinic Acid, 21) 1. The Photolysis of Diacetylfilicinic Acid in Ethyl Ether Fifteen milligrams of diacetylfilicinic acid and 6 ml of anhydrous ethyl ether were placed in a small quartz test 125 tube. After one day of irradiation through Pyrex with a Hanovia L 450 watt lamp, a uv spectrum of a sample indicated only the presence of starting material. After irradiation for 52 hours with a Hanovia S 200 watt lamp a uv spectrum denoted only starting material. 2. Photolysis of Diacetylfilicinic Acid in Methanol under a Nitrogen Atmosphere a. Small Scale.- A solution of 30 mg of diacetyl- filicinic acid in 8 ml of methanol was placed in a small Pyrex test tube. Nitrogen was passed through for a few minutes and the test tube sealed with a rubber serum cap. The mixture was irradiated for 44 hours with a Hanovia L 450 watt lamp. A uv spectrum of a sample indicated no change in the reaction solution. b. Large Scale.— The above reaction was repeated us- ing 1 g of diacetylfikufihic acid and 1 g of sodium methox- ide in 700 ml of methanol. The solution was irradiated through Pyrex for 24 hours with a Hanovia L 450 watt lamp, under nitrogen. The methanol was evaporated and the resi- due was dissolved in water, neutralized with hydrochloric acid and extracted with ether. On evaporation of the ether the residue had identical ir and uv Spectra to those of di— acetylfilicinic acid. There was no indication in the ir spectrum of the presence of 2-acetyl-3,5-dihydroxy-6,6- dimethyl-2,4-cyclohexadienone (acetylfilicinic acid, 22). 126 3. Photolysis of Diacetylfilicinic Acid in Methanol with an Oxygen Atmosphere A solution of 30 mg of diacetylfilicinic acid in 8 ml of methanol was irradiated for 44 hours with a Hanovia L 450 watt lamp, under oxygen. A uv Spectrum of a sample indicated no change in composition of the reaction solution. 4. Photolysis of Diacetylfilicinic Acid in Alkaline Methanol, Usipg a Nitrogen Atmosphere One gram of diacetylfilicinic acid, 1 g of sodium‘ methoxide and 700 ml of methanol were placed in a 700-ml photolysis well. The solution was irradiated for 46 hrs with a Hanovia L 450 watt lamp using a Pyrex filter and under nitrogen. A uv spectrum of a sample indicated only the presence of starting material. Oxygen was passed through the solution for 30 min. The photolysis was continued by irradiation through Pyrex for 13 hours with a Hanovia S 200 watt lamp and 12 hours with a Hanovia L 450 watt lamp. A uv spectrum indicated the loss of the uv absorption band at l 340 mu and the ap- pearance of two absorption bands at A 272 and 250 mu. Work- up of the reaction as outlined in the next experiment af- forded 2-acetyl-3,4-dihydroxy-4-carbomethoxy-5,5-dimethyl- 2-cyclopentenone, EZ/ as identified by ir and nmr Spectra. 5. Photolysis of Diacetylfilicinic Acid in Alkaline Methanol, Air Atmosphere Diacetylfilicinic acid (5 g, 0.021 mole), 5 g(0.093 mole) of sodium methoxide and 700 ml of anhydrous methanol 127 were irradiated through Pyrex with a Hanovia L 450 watt lamp for 24 hours, under air. The methanol was removed under water vacuum in a rotary evaporator and the residue dissolved in water. The aqueous mixture was neutralized, extracted twice with ethyl ether and the ether evaporated. The residue was recrystallized from cyclohexane—ethyl ether to give 3.47 g (68%) of 2-acetyl-3,4-dihydroxy-4-carbo- methoxy-5,5-dimethyl-2-cyclopentenone, 62, A mp of 77.5- 79° was obtained after two sublimations.’ Infrared spectrum: VK:: 1560, 1640, 1705, 1740 and (Ir Spectrum 1) m _1 3450 cm . Ultraviolet Spectra: lfizgn, acidic, 273(9,800) and 225(9,960). xEtOH, basic, 271(15,700), 250 max (18,600) and 207(11,350). The nuclear magnetic resonance spectrum is given in Table II and Nmr Spectrum 1. Apal. Calcd. for C11H1406: C, 54.54; H, 5.83. Found: C, 54.56; H, 5.76. A mass Spectrum indicated a parent mass of m/e 242. 6. Photolysis of Diacetylfilicinic Acid in Alkaline Ethanol (95%)with an Air AtmoSphere The previous reaction was repeated with 2 g (0.0084 mole) of diacetylfilicinic acid and 2 g (0.037 mole) of sodium methoxide in 700 ml of 95% ethanol. The mixture was irradiated through Pyrex for 22 hrs with a Hanovia L 128 450 watt lamp under air. Work-up in a manner similar to that used in the previous experiment resulted in 1.1 g (51%) of 2-acetyl-3,4—dihydroxy-4-carboethoxy—5,5—dimethyl-2-cyclo— pentenone, 62 (mp 80-81° after sublimation). Infrared spectrum: vcgi4 1600, 1640, 1715, 1740 and (Ir Spectrum 3) m _1 3600 cm . Ultraviolet Spectra: XEEEH, acidic, 273(10,600) and 223(11,000). liggH, basic, 272(17,000) and 250(20,600). The nuclear magnetic resonance spectrum is given in Table III, and Nmr Spectrum 2. Anal. Calcd. for C12H1605: C, 56.24; H, 6.29. Found:€: c, 55.94; H, 6.23. 7. Photolysisp£_piacetylfilicinic Acid in Water with an Air Atmosphere Experiment III A5 was repeated with 1 g (0.0042 mole) of diacetylfilicinic acid, 1 g (0.0186 mole) of sodium methoxide and 700 ml of water. The solution was irradiated through Pyrex with a Hanovia L 450 watt lamp, under air. The reaction mixture was worked up as in previous experi— ments. Trituration of the ether residue with carbon tetra- chloride gave 0.67 g (12%) of dimethylmalonic acid (mp 188— 189°, lit. 186° (47), 192-193° (48)) as the only isolated product. Its ir spectrum was identical to that of dimethyl- malonic acid (47). 129 The nuclear magnetic resonance Spectrum is given in Table III. 8. Dark Reaction of Diacetylfilicinic Acid in Alkaline Methanol in an Air Atmosphere A solution of 1 g of diacetylfilicinic acid, 1 g of sodium methoxide and 700 ml of anhydrous methanol were placed in a 700-ml photolysis well, wrapped in aluminum foil and stored in the dark under air. After 6 days a uv Spectrum indicated little change, and that no photo-oxidation pro- ducts had formed. Work-up yielded 0.5 g (61%) of acetyl- filicinic acid (mp 174-176°, lit. 177-178° (20)). B. Preparation of 2-Acetyl-3,4-dihydroxye5,5-dimethyl-2- cyclopentenone, 22, 1. With 2N Sodium Hydroxide 2-Acetyl-3,4-dihydroxy-4-carbomethoxy-5,5-dimethyl-2- .cyclopentenone (1 g, 0.0041 mole) in 50 ml of 2N sodium hydroxide was heated on a steam bath for 16 hrs. The mix— ture was neutralized with concentrated hydrochloric acid and extracted twice with ethyl ether. Evaporation of the ether and recrystallization of the residue from carbon tetra- chloride gave 0.6 g (79%) of 2—acetyl-3,4-dihydroxy-5,5- dimethyl-Z-cyc10pentenone (mp 96-97°, and after sublimation 104-105°, lit. 104-106° (29)). VKBr 1570, 1630, 1705 and 3500 cm-1. max Infrared Spectrum: Ultraviolet Spectra: xfiggH, acidic, 264(10,500), and 223(13,250). 130 1EtOH, basic, 269(19,200), 249 max (21,700) and 205(14,400). The nuclear magnetic resonance spectrum is given in Table III (Nmr Spectrum 3). Anal. Calcd. for C9H12043 C, 58.69; H, 6.57. Found: C, 58.77; H, 6.44. 2. With 2N Hydrochloric Acid A solution of 0.5 g of 2—acetyl-3,4-dihydroxy-4-carbo- methoxy-5,5-dimethyl-2-cyclopentenone in 50 ml of 2N hydrochloric acid was refluxed for 3 hrs, then extracted with ethyl ether. The ether solution was evaporated and the residue fractionally sublimed. The first fraction con— tained both 2-acetyl-3,4-dihydroxy-5,5-dimethyl-2-cyclo— pentenone, 22x and starting material. The second fraction had a mp of 98° and an ir spectrum (KBr) identical to that of authentic 22; 3. With Sodium Methoxide A solution of 0.5 g of 2-acetyl-3,4-dihydroxy—4-carbo— methoxy-5,5-dimethyl-2-cyclopentenone and 2 g of sodium methoxide in 100 ml of methanol was refluxed for 12 hrs. The methanol was removed under water vacuum on a rotary evaporator and the residue dissolved in water. Neutral- ization and extraction with ethyl ether afforded on evapora- tion a compound (mp 96-1010) with an ir spectrum (CHClé) 131 Similar to that of 2-acetyl-3,4-dihydroxy-5,5-dimethyl—2- cyclOpentenone. 4. From 2-Acetyl-3,4-dihydroxy-4—carboethoxy-5,5- dimethyl-Z-cyclopentenone and 2N Sodium Hydroxide 2-Acetyl-3,4-dihydroxy—4-carboethoxy—5,5-dimethyl-2- cyclOpentenone (0.2 g) in 40 ml of 2N sodium hydroxide was heated on a steam bath for 20 hrs. The reaction mixture was neutralized and worked up as in the previous experi- ments. An ir Spectrum of the residue indicated only the presence of 22’ C. Preparation of 2-Acetyl-3—hydroxy-4—keto-5,5—dimethyl- -2-cyclopentenone, Z}, 1. Oxidation of 2—Acetyl-3,4-dihydroxy-5,5-dimethyl- 2-gyclopentenone with Manganese Dioxide A mixture of 2.0 g of activated manganese dioxide and 0.4 g (0.00217 mole) of 2-acetyl—3,4—dihydroxy-5,5-dimethyl- 2—cyclopentenone in 15 ml of chloroform were stirred for 45 min. The mixture was filtered and the chloroform evap- orated to give starting material. The filtered residue was washed with 50 ml of hot methanol. The methanol was evaporated and the residue recrystallized from petroleum ether. A small amount of yellow solid formed. A fractional sublimation yielded starting material, zg'(mp 99-1010) and an impure yellow crystalline solid. The manganese dioxide residue was rewashed with 200 ml of hot methanol. The methanol was evaporated and the 132 residue sublimed resulting in 30 mg (7.6%) of a compound (mp 82-86°, lit. 87° (28)) which had a uv spectrum identical to that reported for 2-acetyl-3-hydroxy-4-keto-5,5-dimethyl— 2-cyclopentenone, Zl'(28). EtOH Ultraviolet Spectra: Xmax . acidic, 282(7,150). lEtOH, basic, 323(7,880) and max 252(11,100). 2. Oxidation of 2-Acetyl-3,4-dihydroxy-5,5-dimethyl- 2—cyclopentenone with Bismuth Oxide A mixture of 90 mg of 2—acetyl-3,4-dihydroxy-5,5-di- methyl—Z-cyclopentenone, 180 mg of bismuth oxide and 10 ml of glacial acetic acid were refluxed for 24 hrs. The acetic acid was evaporated under water vacuum on a rotary evapor— ator and a small amount of dilute hydrochloric acid was added to the residue. The aqueous mixture was extracted with ethyl ether and the ether layer separated, dried and evaporated. Fractional sublimation yielded 2 mg of 11 (mp 87-88°). D. Dye-sensitized Photo-oxidations 1. Standard Reactions Without Dye a. In a Pyrex Photolysis Well.- Diacetylfilicinic acid (0.5 g), 0.5 g of sodium methoxide and 190 ml of methanol were irradiated with a 200 watt tungsten lamp un- der air, while cold water circulated through the inner well 133 jacket. After 18 hrs 92% of the starting material remained, as indicated by the difference in the uv spectra of ali- quots at the beginning and the end of the reaction. b. In a 250-ml 3—necked Flask.- Diacetylfilicinic acid (0.25 g, 0.00105 mole), 0.5 g (0.0093 mole) of sodium methoxide and 200 ml of methanol were placed in a 250-ml 3-necked flask equipped with a reflux condenser, a gas inlet tube and covered with aluminum foil. After one day of irradiation with a 200 watt tungsten lamp, under air, 90% of the starting material had reacted. A uv Spectrum indicated that a compound with a uv Spectrum identical to that of 2-acetyl-3,4-dihydroxy-4-carbomethoxy-5,5-dimethyl- 2-cyclopentenone, 61/ was formed in 75% yield. The meth- anol was evaporated, the residue dissolved in water, neutralized with hydrochloric acid and extracted with ethyl ether. Evaporation of the ether yielded 0.16 g of a yel- low solid. A compound with an ir Spectrum identical to that of 2-acetyl-3,4-dihydroxy-5,5-dimethyl-2-cyclopentenone, 70, was obtained after sublimation. (W 2. Dye-sensitized Photo-oxidations of Diacetylfili- cinic Acid a. With Fluorescein.- Fluorescein (0.120 g) was added to the solution of Experiment III D 1a which contained 92% unreacted diacetylfilicinic acid. The solution was ir- radiated with a 200 watt tungsten lamp, under air. After 22 hrs 83% of the diacetylfilicinic acid had reacted as 134 indicated by a uv spectrum. The reaction was worked up by evaporation of the methanol, addition of water to the resi- due, followed by neutralization and extraction with ethyl ether. The residue from the ether layer had uv absorp— tion bands at 250 and 271 mu in basic ethanol solution which was characteristic of 2-acetyl—3,4-dihydroxy—4—carbomethoxy- 2—cyc10pentenone, 61; Its ir Spectrum had absorption bands at 1600, 1640, 1710 and 1740 cm-1 Similar to the ir spec— trum of QZ/ but also Showed a band at 1660 cm_1 due to the dye and/or the starting diacetylfilicinic acid. b. With Methylene Blue.- A solution of 0.5 g of di- acetylfilicinic acid, 0.5 g of sodium methoxide and 0.20 g of methylene blue in 190 ml of methanol was irradiated with a 200 watt tungsten lamp, under air. After two hours 60% of the starting material had reacted, and also 70% of the methylene blue. Irradiation for an additional 2.75 hrs resulted in complete reaction of the methylene blue and no increase in the photo—oxidation of the diacetylfilicinic acid. The reaction was worked up as in the previous experi- ment. The residue had ir and uv Spectra Similar to a 60:40 mixture of 1-acetyl-3,4—dihydroxy-4-carbomethoxy-5,5—di- methyl-Z-cyclopentenone, 61/ and diacetylfilicinic acid. c. With Rose Bengal.- A solution of 0.5 g (0.0021 mole) of diacetylfilicinic acid, 0.5 9 (0.0093 mole) of sodium methoxide, 0.030 g of Rose Bengal and 200 ml of methanol was irradiated for 3.3 hrs with a 200 watt tungsten 135 lamp, under air. The reaction was followed by uv Spectra which indicated a 98.8% yield of 2—acetyl-3,4-dihydroxy-4- carbomethoxy-5,5-dimethyl—2-cyc10pentenone. The methanol was evaporated under water vacuum on a rotary evaporator, the residue dissolved in water, neutralized with acid and extracted with ethyl ether. The ethyl ether was evaporated and the residue recrystallized from cyclohexane-ethyl ether to give 0.375 g (73.8%) of 2-acetyl-3,4-dihydroxy-4-carbo- methoxy-5,5-dimethyl-2-cyclopentenone. This was identi- fied by its ir Spectrum. E. Reaction of Diacetylfilicinic Acid with Singlet Oxygen Prepared from 9,10-Diphenylanthracene Peroxide 1. Standard Experiment A solution of 0.33 g of 9,10-diphenylanthracene in 190 ml of ethyl ether was irradiated for 0.5 hr with a Hanovia L 450 watt lamp, under air. The photo—oxidation was followed by observing the decrease in uv absorption bands of diphenylanthracene at A 350-400 and 258 mu, and the increase in intensity of the uv absorption band at h 210 mu due to the diphenYkmthracene peroxide. The ether solution was concentrated to 10 ml and added to 0.075 g of diacetylfilicinic acid in 40 ml of methanol. The reaction mixture was refluxed for 4 hrs in the dark. There was no change in a uv Spectrum of an aliquot which indicated no change in the composition of the reaction mixture. Sodium methoxide (0.2 g) was added and after 0.5 hr reflux 78% of 136 the diacetylfilicinic acid had reacted to give 2-acetyl—3,4- dihydrOXy-4-carbomethoxy-5,5-dimethyl-2-cyclopentenone,.62, as identified by a uv spectrum. 2. Preparation of 2—Acetyl—3,4-dihydroxy-4-carbo- methoxy—5,5—dimethyl-2-cyclopentenone from the Peroxide of 9,10-Diphenylanthracene A solution of 0.5 g of 9,10-diphenylanthracene in 190 ml of ethyl ether was irradiated with a Hanovia L 450 watt lamp for 2.5 hrs, under air. The reaction was followed by observing the decrease in intensity of the uv absorption bands of diphenylanthracene at A 350-400 mu. The ether was concentrated to 10 ml and added to a solution of 40 ml of methanol, 0.075 g of diacetylfilicinic acid and 0.150 g of sodium methoxide. The uv absorption band at A 340 mu for the diacetylfilicinic acid monoanion disappeared after 1.75 hrs reflux. Subtraction of the uv absorption bands for diphenylanthracene and its peroxide from a uv Spectrum of the reaction solution resulted in bands at A.250 and 270 mu characteristic of the photo-oxidation product, 2- lacetyl-3,4-dihydroxy—4-carbomethoxy-5,5-dimethyl-2-cyclo- Iaentenone, 62, The methanol was removed under'water vacuum <1n.a rotary evaporator and the residue dissolved in water. flIhe solution was neutralized with acid and extracted with erthyl ether. Evaporation of the ether yielded 0.0564 g (’74%) of 61, It was identified by its ir and uv spectra. 137 F. Preparation of Methyl Ethers of the Photo-oxidation Products In all cases diazomethane was prepared by the addition of N-nitrosomethylurea to an ice cold 30-40% aqueous potas- sium hydroxide and ethyl ether mixture. When the N-nitroso- methylurea had completely decomposed, the yellow diazo- methane-ethyl ether phase was decanted into an ice cold ether solution of each compound. At least 70% conversion of the N-nitrosomethylurea was assumed in all cases and an excess of the N-nitrosomethylurea beyond this was used. 1. Reaction of 2-Acetyl-3,4-dihydroxy-4—carbomethoxy— 5,5-dimethyl-2-gyc10pentenone with Diazomethane An excess of diazomethane in ether was prepared from 2.5 g of N-nitrosomethylurea and 10 ml of 30% potassium hydroxide. The ether layer was decanted into 10 ml of ether containing 1 g of 2-acetyl-3,4-dihydroxy-4-carbomethoxy- 5,5-dimethyl-2-cyclopentenone, 61; After one hour the ether and excess diazomethane was evaporated. Vpc analysis indicated a 67% yield of equal concentrations of two methyl ethers, 2-acetyl-3-methoxy-4-hydroxy-4-carbomethoxy-5,5- <22<>1111.1<>d1 2%)71211... 11121.:1.2..11..1.»1 1111.3 311... . 1,111,111.11. 111.151.. .111 11.111 >21. e11aw21y .wcocmucmmoauwolmlahaumfiflplm.mlmxozuwfionumolfilhxovanflplv1mlamumodlm .H Efluuuwmm HEZ 168 Emu 01 a m e m man man ummmmo mmw3m é .waocmpcmmoaowolNiamnumfiflolm.mlmxonpwonnmolwlmxoHU>£HUI¢1Mlamuwo¢rw .N Eduuummm HEZ 169 Ema 0H $1111.11) _ - 1.1 1 . .. _. . 11131123511111... 1 E1113. .31....1111111111111233 111,111....11.111141112141711...11.1.1171...,.....1... .,. ....».........1. .. K1: 1.1.1 1 111.121 .maodwuawmmwowwlNIHMQMmEHUIm1nlwxouohnflvlv.MIamuwo1. .11..\,\.>>? 1. 3111\1111.1. .1 .1112. . ‘4‘“ 1. .1 \W1...1.1.. x. <2 151-u a fi . ......._.1 1111. 5% 11....w11111r . 1r d .1 . 1 1n. ,w._ an ...: 1 u ... ... amm .wczomEoo.czocch .v Eduuommm HEZ 171 8mm OH fi» ‘ - .1, — \I‘L . . i'.‘~|l.o|l.l'lh [gig jg SOCH .1 .mcoHamlmmlwmmoumimrolmxothglvlamumo¢rm .m Ednuommm HEZ APPENDIX II Molecular Orbital Calculations (w-Technique) 172 APPENDIX II Molecular Orbital Calculations (uwTechnique) Although much work has been done on the photolysis of 2,4-cyclohexadienones (2,3,9), no quantitative experimental work has been done to explain why some 2,4-cyclohexadi- enones give ring-opened products and others rearrange to bicyclo[3.1.0]hexenones when irradiated. In Sections III and IV it was found that molecular orbital calculations were useful in predicting the position of attack by singlet oxygen on two carbon anions. Molecular orbital calculations might also provide some explanation for the two different reaction paths observed in the photolysis of 2,4-cyclo- hexadienones. Also calculations could provide a rationale for the preferred stability of one tautomeric form of a 2,4- or 2,5-cyclohexadienone over another as assigned in Sections I and II. A. The Method Molecular orbital calculations were carried out with a 3600 Control Data Computer using a program written by Professor R. S. Schwendeman of the Chemistry Department of Michigan State University. The method used was a variation of the Huckel molecu- lar orbital method known as the "wrtechnique" (49). Nor— mally the values of the coulomb integral are taken as a 173 174 constant value, "a" for carbon. It is a function of the nuclear charge of the atom. The resonance integral, 6, is related to the degree of overlap and is constant for a given overlap between two carbon "p" orbitals at a given distance. A change in the atomic number of an atom effects both the coulomb and resonance integrals. The relationships for new values of these integrals can be expressed by: a do + hfio 1 B = k50 2 Where a0 and 50 are the values for carbon, and h and k are constants which depend upon the heteroatom which re- places carbon. Values of these constants for a carbonyl oxygen and a hydroxyl oxygen were taken from the text Quantum Organic Chemustry (36). Because the anions of some of the compounds in this research are resonance hy- brids containing several carbonyl and hydroxyl groups, a direct ratio of values for h and k was used. These values are given in Table IV in Section.III. The "artechnique" allows the coulomb integral to be adjusted for changes in the charge on each atom. It has been suggested (49) that the value of a should be linearly related to the charge and may be formulated as: a=ao+(1-qr)wfio 3 Where "w" is a dimensionless parameter whose value may be chosen (the recommended intermediate value of 1.4 was used here) to give the best agreement with experiments. The 175 charge on the atoms was taken as qr, defined in the usual way as: _ 2 Z NCi 4 qr Where N is the number of electrons in the molecular or- bital, and ci is the coefficient of the "i"th atom. The sum is taken over all the occupied molecular orbitals. After each calculation the charges can be recalculated from the coefficients. The new charges can then be put into equation (3) and the molecular orbital calculations repeated. This process is continued until the values for the molecular orbital energy levels and charges on the atoms converge to constant values. The program does this auto- matically, it being necessary only to put in the desired number of iterations. The input data were written up as: Input for ggowm Name Card: 1 CI A! RI BI OI NI Y! LI II GI RI OI UI PI 7 I Name can be from column 2 to 55 Parameter Card: I212 4 I T NO. NO. NO. Atoms No. Iterations Pairs Electrons 100 x w ‘ 176 Data Cards: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20,21 1L j k 1 1°?Mi' 100Mik 100Mi1 100Mii I y I I J I I I r I I I J I L i. 2 2 O 0 1 , 2 IbODNH An explanation of the input data follows. In the-Name Card, if column one is left blank, the program operates normally. This means that the atom-atom polarizabilities are calculated only once and these values are used in suc- cessive iterations. Although this procedure may cause slight inaccuracies, it saves computer time. When a 1 is placed in this column, the atom-atom polarizabilities are recalculated for each iteration. This may aid in convergence of the molecular orbital energy values and the charges on the atoms. Any desired name may be placed on the rest of the card. The second card is the parameter card. In the first two columns the number of atoms in the system is placed. The next two spaces are for the total number of electrons, and the following two spaces are for the desired number of iterations. Normally three to four iterations are suf- ficient for convergence to occur. The value for 100w is put in the next four places. Normally "w" equals 1.4 so 140 would placed here. This program is written such that Eva... .. 392.44ij I llv’l [I III Ii 177 if "b" equals 1.4 it is not necessary to put this value in. However, if a different value for "w" is used then it must be put in these columns. The next series of cards are the actual data cards, one for each atom of the molecule. In the first two columns the atom number of the molecule is placed. The next three places are for the atom number(s) that this atom is connected to. It is not necessary to indicate the interrelation between any two atoms more than once, al- though all atom numbers must appear in the first two col- umns. Consider, for example, an ethylene group. A number 1 would be placed in column 2 and a number 2 would be placed in column 4, signifying that atom 1 is connected to atom 2. In a second card a 2 would be placed in column 2 and noth- ing else is needed as it has already been indicated on the first card that atom one is connected to atom two. Under the columns labeled 100Mij, 100Mi and 100Mi k l are placed the relationship of the resonance integral be- tween atcms"i" and "j", atoms"i" and "k" and atoms “i" and "1" respectively, if different from one. This would be the value of k from equation 2, x 100. In column 19 the value of h (Equation 1) for the coulomb integral begins, if it differs from the normal 1 for carbon atoms. An example is written out for the car- bonyl group using the values of h and.1< taken from Table IV in Section III. Note that only two data cards are needed and only a 2 in the second column will be needed. 178 on the second card Calculations of compounds with a larger number of atoms can be carried out in a similar manner. At times it may be desirable to carry out simple Hfickel calculations with the same program. This will be done automatically if there is no charge on the molecule. However, the program was designed to spread any charge evenly over all the atoms in the first calculation so as to be one step towards the actual solution and save com— putor time. To obtain the simple Hfickel molecular orbital energy levels it is necessary to add or subtract from each molecular orbital energy level the amount which the diagonal value of the topological matrix deviates from the normal a value. This normal value would be 1 for a carbon atom and h for the heteroatoms. For charged hydrocarbons this would turn out to be the fraction of the charge ex— pected on each atom. The molecular orbital calculations were used to calcu— late the charges on the atoms, not only in the ground state but in excited states as well. The ground state charges were obtained directly from the output data sheet. To get the n-7* excited states, extra charges were added to each of the atoms using equation 4 and the coefficients of the lowest unoccupied molecular orbital. (”Jim “5:1"? 179 For the w - 0* excited states a difference of the squares of the coefficients for the lowest unoccupied mo- lecular orbital and the highest occupied molecular .orbital were added to each of the ground state charges. This gave the charges on each atom for this excited state. B. The Effect of Substituents on the Photolysis of 2,4— Cyclohexadienones It seemed desirable to try to use molecular orbital theory to examine the effect of substituents on the pho- tolysis of 2,4-cyclohexadienones in an attempt to under- stand why ring-opened products are obtained in some cases and bicyclic products in others. R3 R20 I I 9 9. R5 R5 R4 R2 h R2 n and/0r —‘c': - 6 = c - c - EOCH .__X__4> I’> fi 3 R3 5 R3 R4 R4 Hart (6) found that a methyl substituent at carbon 5 does not effect the course of the photolysis. When the substituents (R2, R3, R4) are all methyls then a bicyclic product was obtained. If R3 was a hydrogen then both the bicyclic and ring-opened products were obtained. In the case where R2 was hydrogen or R2 and R5 were hydro— gens then only the open chain compound was formed. \E-Mfihfi- I 53. . 180 Examination of the coefficients of the lowest unoc- cupied molecular orbital and the highest occupied molecular orbital (Table XI) of 2,4-cyclohexadienone showed that the greatest increase in electron density for n - 0* excita- tion would be at C2 and C5. Such an increase at these carbons should facilitate a-cleavage between the carbonyl carbon and C6 due to the increased ease with which double bonds might be formed at these positions. Table XI. Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of 2,4-cyclohexadienone (w-technique) Coefficients Charges Atom m + 1 m Ground n - 0* 0 - 0* State 0 0.3223 0.2785 -0.2119 -0.316(+0.684)* -0.2359 C1 -0.2183 -0.0340 0.1430 0.0955 0.0964 C2 -0.5024 -0.5726 -0.0066 -0.2586 0.0684 C3 0.4404 -0.3488 0.0399 -0.1521 -0.0321 C4 0.2790 0.3621 0.0079 -0.0699 0.0609 C5 —0.5694 0.5836 0.0277 -0.2953 0.0437 '*Indicates actual charge on the oxygen atom after account- ing for the loss of one "p" electron on n - 0* excitation. Most of the difficulty in predicting the photolysis course for 2,4-cyclohexadienones when substituents are present, is the result of a lack of knowledge regarding the photochemistry of the reaction(s). It is not known whether 0 - 0* excitation or the photolysis occurs via n - 0* or is a singlet or a triplet reaction. Hart (6) has shown 181 that the reactions are not quenched by piperylene, and says the question of singlet or triplet remains open for both types of photoisomerizations. For this reason both types of excitation were considered in the molecular orbital calc— ulations (see Table X). The ground state for 2,4-cyclo- hexadienone has the following molecular orbitals: 0+ 0— 0+ 0+ l + — — - I: ”L + {Kl ‘ KI i” 4' '+' + + -V+ + — + + + + m-2 m-l m m+1 V It can be seen that the m - 2 orbital favors the C1 - C5 bonding while the other two do not. 'The first excited state (m + 1) also favors bonding between C1 and C5. This is necessary if bond crossing at this point is involved in the mechanism as outlined here to give the bicyclic product. demotion 182 Table X. Parameters for molecular orbital calculations of 2,4-cyclohexadienones c c-c C-OH c-o' Compound k h k h k h k 0 II 2 1.2 2.0 1 1 3 5 2’1, 1.2 2.0 1 1 2.0 0.9 2.2» 1.2 2.0 1 1 2.0 0.9 65' 1.46 1.62 1 1 2.0 0.9 1.46 1.62 55 1.4 1.77 1 1 2.0 0.9 1.4 1.77 183 Zimmerman gives a similar argument in explaining the bond- crossing of 2,5-cyclohexadienones to give a similar bi- cyclic product (50). If the photo excitation is n - 0* to the m + 1 molecular orbital, then there is increased elec- tron density at C2 and C5 which is favorable for the forma- tion of ring-opened compounds and also for increased bond- ing between C1 and C5, which favors bicyclic products. However, it is known that bicyclic products are only formed when certain substituents are present. One possible ex- planation is that n - 0* excitation causes the formation of ring-opened products, whereas 0 - 0* excitation favors bicyclic products. This is consistent with the calcula- tions. An n - 0* excitation increases electron density at C2 and C5 Whereas 0 - 0* excitation would not change the electron density at these carbons very much. It can be seen from Table XI that the coefficients for the m and m + 1 orbitals at C2 and C5 are similar. Therefore 0 - 0* excita- tion would not be particularly favorable for the formation of ring-opened product. However, 0 - 0* excitation would be favorable for bond-crossing, leading to a bicyclic product. Not only are the signs of the coefficients correct for C1 - C5 bond formation in the m + 1 state but one electron would be removed from the m level,.which is un- favorable for such bonding. In all cases disrotatory bond formation of the cyclopropane ring is assumed. The n - 0* excitation results in increased negative charge at all the carbon atoms. Methyl substituents, which ‘3: a 184 would not favor such an increase, would be expected to de- stabilize the n — 0* excited state. However, in the 0 - 0* excited state there is an increase in positive charge at 02, C4 and C5, the largest increase being at 02 and C4. This is consistent with the observation that bicyclic pro- duct was formed only when methyls were at both of these positions (6). Methyl substituents should favor the n - 0* excited state and lower the energy of the m + 1 level, per- haps low enough so that the 0 - 0* absorption could occur at a wavelength (330 mu) longer than that of the n - 0* absorption. It was of interest to examine the effect of other sub- stituents on the 2,4—cyclohexadienone. If the above con- clusions are valid then the type of substituent is important in determining which excited state is obtained. It would be expected that electron—withdrawing groups at C5 and C4 should not favor 0 - 0* excitation. This expectation was supported when it was found that neither diacetylfilicinic acid, 21, nor acetylfilicinic acid, 22” gave a bicyclic. product in ethyl ether or methanol on irradiation under a nitrogen atmosphere. . ' ‘.' .‘fl‘l-I'TI‘JW'“ twig-a: . 185 Acetyl groups at C2 and/or C4 should favor the n - 0* excited state as they would stabilize the increased nega- tive charge. In order to examine this in more detail, molecular orbital calculations were carried out for both gl'and 22; The results are given in Tables XII and XIII. For acetylfilicinic acid, calculations show an in- creased electron density especially at carbons 2, 3 and 5 h. for n - 0* excitation. Increased electron density at C2 and C5 should favor bond-breaking to give ring-opened product. ~wfiar Table XII. Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of acetylfilicinic acid (w-technique) Coefficients Charges Atom m + 1 m Ground n - 0* 0 - 0* State 0 0.2250 0.2934 -0.2626 -0.3136(+0.6864)* -0.2276 C1 -0.1724 -0.0848 0.1214 0.0918 0.0990 C2 -0-3002 -0.5939 -0 1824 -0.2724 0.0796 C3 0.4780 —0.1330 0.0607 -0.1683 -0.1506 C4 0.1679 0.4036 -0 1163 -0.1445 0.0179 C5 -0.5682 0.3012 0.0637 -0.2593 -0.1697 * A Indicated actual charge on the oxygen atom after account- ing for the loss of one "p" electron on n -0* excitation. all the carbon atoms for n - 0* excitation. i . An overall increase in negative charge was found at This is favor- able at carbon 2 where the acetyl group is located but probably not favorable at the 3 and 5 carbons where there are hydroxyl substituents. Thus the n - 0* excited state 186 for acetylfilicinic acid may not be as stable as the n - 0* excited state of 2,4-cyclohexadienone. This may explain why the ring opening does occur for 22 but requires such long reaction times (32 to 48 hours). Table XIII. Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of diacetylfilicinic acid(w-technique) Coefficients Charges d Atom m + 1 m 3:2:2 n - 0* 0 - 0* 0 0.2317 -o.2275 -O.2416 -o.2951(+o.7049)* -o.2433 01 -0.1577 0.0557 0.1318 0.1070 0.1254 C2 -0.3556 0.4654 -0.1212 -0.2474 -0.QS04 C3 0.4327 0.1406 0.0768 -0.1102 -0.0904 C4 -0.4249 -0.5635 -0.1032 -0.2842 0.0338 C5 0.3384 —0.2239 0.0447 -0.0703 -0.0203 *- Indicates actual charge on the oxygen atom after account- ing for the loss of one "p" electron on n - 0* excitation. For the 0 - 0* excitation there is an increase in positive charge at carbons 2 and 4, and an increase of negative charge at carbons 3 and 5, both of which work against the stability of this excited state. In addition it can be seen that the electron density at carbon 2 would actually decrease and the density at carbon 5 would only increase slightly over that of the ground state, which would not favor ring cleavage. For these reasons the 0 - 0* excited state was thought to be unstable relative to the n - 0* excited state, which, via ring-opening to a ketene, 187 can give the pyrone, 22'. Calculations for the molecular orbitals and charge densities for diacetylfilicinic acid, 21, are given in Table XIII. It can be seen that although the n - 0* ex- cited state would be stabilized with negative charges on the atoms where the acetyl groups are located, there are nega- tive charges on the atoms containing hydroxyl substituents. The increase in electron density at carbon 5 (0.1150) is much less than that for acetylfilicinic acid, 22“ (0.319) and for 2,4-cyclohexadienone (0.323) both of which undergo ring-opening. This might be an explanation why diacetyl- filicinic acid, 22/ gives no ring-opened compound. For the 0 - 0* excitation the charges are unfavorable for stabilization of the excited state, as with acetyl- filicinic acid, and the increased electron density at car- bons 2 and 5 (0.0908, 0.0650) would not favor ring cleavage. Also in this case 0 - 0* excitation does not enhance the possibility of bond—crossing, which is seen to be unfavor- able both in the m and m + 1 molecular orbitals. On the whole there are only three molecular orbitals that favor bond-crossing for diacetylfilicinic acid, 21, whereas four do not. When one adds the fact that 0 - 0* excitation does not increase the possibility for bond crossing, it is not unreasonable that no bicyclic product is formed. In order to seek a reason for the relative rates of the self-sensitized photo—oxidations of acetylfilicinic 1|.I-uv‘imwkfl .(i.l.. Mali... all], .4 u ., .. . .. .. 188 acid, 22, (8.1 x 10.3 mole/l/hr) and diacetylfilicinic acid, 21“ (1.35 x 10-3 mole/l/hr) molecular orbital calculations were carried out on the mono-anions of these compounds (Table XIV and Table XV). + — + - .+O _ 0 +0 .0 " K " {W 4'. + + + + - - + + \/ r- + + — + m—6 m-5 m-4 m-3 l + - - + E +0 _+0 +0 _0 § _ +3 _ - + + - + + m-2 m-l m m+1 Molecular orbitals of Diacetylfilicinic acid, 21. The only major difference is the fact that in the 0 - 0* excited state there are two acetyl groups destabi- lizing the excited state for diacetylfilicinic acid and only one destabilizing the excited state for acetylfili- cinic acid. On this basis the 0 - 0* excited state for acetylfilicinic acid may be less unstable. The acetyl groups should stabilize the n - 0* excited states for both monoanions, whereas the hydroxyl groups should destabilize it. The only other factor which may be 189 important in connection with the faster photo-oxidation rate of acetylfilicinic acid is that the increase in elec- tron density at C5 for acetylfilicinic acid (0.192) is greater than the increase at the equivalent carbon atom for diacetylfilicinic acid (0.0628). Table XIV. Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of monoanion of acetylfilicinic acid (w-technique) Coefficients Charges Ground * Atom m + 1 m State n 0 0 0* 0 0.2216 -0.3055 -0.3923 -0.4413(+O.5587)* -0.3483 c1 -0.2475 0.1455 0 1288 0.0676 0.1688 c2 -o.0917 0.4889 -0 1990 -0.2074 0.0311 c3 0.5520 -0.0868 0 0420 -0.2630 -0.2555 04 -0.1847 -0.5918 -0 2082 -0.2422 0.0738 05 —0.4397 -O.1681 0 1106 -0.0814 -0.0536 * Indicated actual charge on the oxygen atom after account- ing for the loss of one "p" electron on n — =* excitation. Table XV. Coefficients of the highest occupied and lowest unoccupied molecular orbitals and charges on ring atoms of the monoanion of diacetylfilicinic acid (w-technique) Coefficients Charges Atom m + 1 m :Sf'tliztreld n - 0* 0 - 0* 0 0.2452 -0.3023 -0.3467 -0.4067(+0.5933)* —0.3157 C1 -0.2516 0.1250 0.1313 0.0685 0.0213 C2 -0.1610 0.5348 -0.2035 -0.2294 0.0317 C3 -0.6025 —0.0000 0.0605 0.0243 -0.0119 C4 -0.1610 -0.5348 -0.2035 -0.2294 0.0317 C5 -0.2516 -0.1250 0.1313 0.0685 0.0213 ‘Indicated actual charge on the oxygen atom after account- ing for the loss of one P electron on n - 0* excitation. 190 Schenck says (33) that in his opinion the singlet and triplet states of the sensitizer react with oxygen to form a sensitizer—oxygen adduct, rather than to produce excited oxygen molecules by mere physical energy transfer. This adduct may be interpreted as a Mulliken charge-transfer complex. He also suggested later that if the excited mole- cule has biradical character then it could behave chemically in the manner of free radicals. Therefore reactions with oxygen should not only be very efficient in quenching the excited state but also lead to a new biradical molecule in which the oxygen is attached to a former radical site by a normal chemical bond. It is possible that the two carbanions studied here might form singlet oxygen by the following mechanism(s). 191 If this is the case then the rate of singlet oxygen forma- tion may be related to the electron density at C5 in the excited state; this may be the basis for the rate differ- ence of the self-sensitized photo-oxidations of acetyl- filicinic acid, 22” and diacetylfilicinic acid, le’ The rate of the dye-sensitized photo-oxidation of acetylfilicinic acid (5.9 x 10"3 mole/l/hr) and diacetyl- filicinic acid (9.6 x 10.3 mole/l/hr) are more comparable and.the difference here could be due to the fact that there are two equivalent positions in the diacetylfilicinic acid anion which can be attacked whereas there is only one for acetylfilicinic acid anion. In conclusion molecular orbital calculations suggest that for an unsubstituted 2,4-cyclohexadienone, n - 0* excitation might lead to a ring-opened product. If the n - 0* excited state is destabilized and the 0 - 0* excited state is stabilized by substituents such as methyl groups, then the 0 - 0* excited state could become more favorable. Calculations show that the 0 - 0* excited state would favor bond-crossing to give a bicyclic product, in preference to a ring-opened product. On extending this idea to other substituents such as acetyl and hydroxyl groups, calculations show that acetyl- filicinic acid, 22“ would be more likely to form a re- active n - 0* excited state than diacetylfilicinic acid, a}, This offers an explanation for the fact that acetyl— filicinic acid undergoes ring-opening on photolysis, whereas diacetylfilicinic acid is photochemically inert. Molecular 192 orbital calculations also show that the 0~0* excited state in both cases are unfavorable for bond-crossing and give a reason for the lack of bicyclic compound formation. The reason for the more rapid photo-oxidation of acetylfilicinic acid over that of diacetylfilicinic acid is not as clear. The results indicated that acetylfili- cinic acid was more efficient in reacting with oxygen to produce singlet oxygen. This could be due to a more stable excited state or, as suggested by molecular orbital calcu- lations, due to a higher electron density at C5. This would enable more facile quenching of the excited state by triplet oxygen, if the quenching reaction was visualized as a free radical type reaction. C. The Relative Stability of Tautomeric Structures as Predicted by Molecular Orbital Calculations Molecular orbital calculations were carried out to determine relative stabilities of different isomeric struc- tures of the substituted 2,4- and 2,5-cyclohexadienones given in Sections I and II. The values of the energy of the systems (in terms of 5) obtained from simple Hfickel calculations and using the "w-technique" are recorded in Table XVI. Calculations for a number of possible structures were carried out for some compounds. Also recorded are the difference in 6 values between the highest occupied molecular orbital and lowest unoccupied molecular orbital for each structure. 193 Table XVI. Molecular orbital calculation results on tauto- meric structures of substituted dienones. _ Longest Devi- itruc Total 0 Energyjfi) (m+1)-(m) absorp- ation ure . w . Sym- Huckel tion from tech- (6) . bol nique wave Line length (5) (mp) 13 18 .6946 13 .7767 1.2347 251 0 .32 ' 22' 18.8341 13.8946 0.9757 251 0.57 0 3.5.3. 18 .3715 13 .5924 1.0059 278 ° 42' 18.4434 13.6649 0.9385 278 0 51‘ 18.1674 13.4411 0.9949 348 0‘2 ’5‘; 18 .1674 13 .4411 0 .9949 343 21' 30.2735 25.1604 0.8714 300-315 0.08 ° .533. 29.8206 24.7198 1.1800 300-315 0.23 D 52' 30.1330 25.2904 1.0841 340 0 {5‘4 29 .6742 24.8082 1.0572 340 d> ’2‘?" R=CH3 and 24.0557 19.6138 0.9536 -335 0.13 ‘ 22. R= 330 (6"6 R=CH3 and' 24 .0089 19 .5578 0 .9518 335 0 .13 '3 29. R= 330 R=CH3 99’ 24.4449 19.4572 1.1800 335 0.36 U R: 330 9'.“ . __._' 137.11?“ "“" 194 Table XVI. (Continued) Struc- Total 0 Energy(fi) Longest DeVl- u n g - absorp- ation ture HUckel w (m+1) (m) tion from Sym- tech- . bol ni ue (5) wave- Line q length (3) (mill R:CH3 fié. 24.0141 19.2077 1.2116 350 R=H 345 R=CH3 109, 23.4303 19.1367 1.0388 350 q (I R=H : 345 i 33, 12.0472 9.5472 1.3696 258 ° 7 101 11.6344 9.3015 1.2431 279 o t 32, 17.7960 15.2207 1.2948 277 241 41’ 26.0268 23.3759 0.8959 310-345 C 22. 25.8128 23.4274 0.9539 398 51, 17.2177 14.8503 1.4469 277 k 195 The calculations (Table XVI) agree extremely well with the structures assigned on the basis of spectroscopic evi- dence. The only exception is filicinic acid itself. In this case the fully conjugated structure, 22“ was predicted by calculations to be more stable than the cross-conjugated structure, 19} the latter was assigned on the basis of spectral evidence. For all the other compounds, that struc- ture predicted by molecular orbital calculations to be most stable was the structure which predominated, as deduced experimentally. The agreement for anion structures was ”It” “‘ yi)_ "(— :’- ~ "V also good, with the exception of the monoanion of filicinic acid. Additional evidence for structure assignments were obtained by plotting the difference in energy, Am, (in terms of 6) between the highest occupied and lowest unoc- cupied molecular orbitals against the longest uv absorp- tion wavelength for each compound (Fig. XV). These were compared to a curve obtained by plotting similar values for a series of related compounds with known structures, (Table XVII); see Fig. XVI. For most of these compounds there was quite good agreement between absorption wavelength and Am. The compounds from Table XVI were plotted in a similar way. The assigned structures for the neutral compounds gave good agreement with the line obtained in the previous graph. By taking the deviation from the line in terms of B, struc- ture preferences were obtained (Table XVI). The structure which had the smallest deviation would be preferred. As Table XVII 196 Molecular orbital calculation results on standard compounds containing oxygen atoms abggrp— Compound m tion‘ Symbols (mu) 0 CH3(CH=CH)n-C-H n = 1 1.7130 220 . n = 2 1.1106 271 . [1 n = 3 0.8202 315 . ( n = 4 0.6499 351(353) 3 n = 5 0.5381 377 1% n = 6 0.4591 399 n = 7 0.4003 424 OH 0 CH3-C=CH-C—OEt 1.4226 245 o 9 ,OH r' \( 0.8641 380 o \:/ i T( a 1.2635 274 o /2\\/. o ACO\ :7 /l\ H i 1.3107 240 o /\/ r“ 9 1.5032 270 x LL-TLCH 197 Table XVII. (Continued). UV Compound m aiig;p_ Symbols (W) U o (CH=CH)n—CHO n = 1 1.0131 .312 X n = 2 0.7613 346 X n = 3 0.6092 366 X n = 4 0.5076 389 X n = 5 0.4350 412 X n = 6 0.3806 429 X 198 .mmuosmfiomxmnoaomo Umusuflpmnsm mo mamuflnuo HmasomHoE flmfimsoooss umm3OH Usm Umflmsuoo ummgmflz mo mmumcm CH mocmHmMMHU.MN npmsmam>m3 coaumHOQO mo uoam .>x .38. 8:6 v.H m.H N.H H.H o.H m.o w.o v.0 ®.o m.o a. _ s a _ . . _ 4 4 _ -owm ‘ w . D o ... o // o e .28 -;/// / o ,// III coasmaa :/, - NO //// own 6 1|. GOHQ¢ .HoosH o o e .D . /// 0 0‘0 1/ // 0 .II- COHG¢ .Hoo //, :owm // . III. .502 .uoo ///// D .502 .HOUQH o .oov mmus0usupm (0m) qabuetaAeM 199 -.. -_.._..__..__ .mEOum cmmhxo mcflcflmucoo macsomEoo Unmosmum mo mamgflnuo HmasuwHOE Umflmsouoss umw3oH 6cm Umflmso loo pmmzmfln mo mmumcm CH mucmumMMHU MM.£umcmHm>m3 coaumuomnm mo uoam .H>x .mHm 0 iii: mumguo . III. omocAmoumovmmo C omo Amoumov o .X Ill. : : l ....--1 .-.l. -1... ___ ... ovm 0mm 0 N C") (hm) quuaIaAeM 200 can be seen from Table XVI structure 19’(for filicinic acid) was favored over that of 22; For diacetylfilicinic acid 21 was favored over 53; Both agree with the spectroscopic- ally assigned structures. For acetylfilicinic acid (and acetylmethylfilicinic acid) this method did not distinguish between the two fully conjugated systems, 22, 62 and 32x 66, but did favor these over the cross-conjugated structures. For the anions most points appeared to be above the line, and perhaps should have a relation of their own, different from the neutral compounds. Most of the calculations for the anions are inconclusive using this methoa as they do not distinguish between structures. On the whole, there was good agreement between the molecular orbital calculations and the structures assigned for the derivatives of filicinic acid and related compounds.