PART I THE SYNTHESIS OF ET HYL 3,3-Dl- (4 - HYDROXY - 3,5 - D! - TERT - BUTYL) - PHENYL-Z-AMINOPROPIONATE PART H NOVEL REACTTON PRODUCTS 0F 4-HYDROXY-2,3,4-TRTPHENYL-2- CYCLOPENTEN - 1 - GNE Thesis for the Degree of Ph. D. MTCHIGAN STATE UNWERSITY MARY ELEZABETH CONNER 1969 Mun-:1. ant-35'5" - LIBERiRl’ y A'Ticlfigd :1 State / Unix :- rsi Cy . .. Jam”. .lvr.:1 .- . . This is to certify that the thesis entitled PART I THE SYNTHESIS OF ETHYL 3,3-DI-(4-HYDROXY-3,5—DI- TERT-BUTYL)-PHENYL-2-AMINOPROPIONATE PART II NOVEL REACTION PRODUCTS OF 4—HYDROXY—2,3,4- TRI PHENYL -2 - presented by CY CLO PENTEN -1 -ONE Mary Elizabeth Conner has been accepted towards fulfillment of the requirements for PhD degree in Chemistry erg/W / Major proflsgor Date M 0-169 amuse 3y MAB & SUNS' 9995-318115" JNL‘. 1 l em" ----:'-----'--:- ."~ .-u- A. n . .- mail I a. :u—n-n. THESYNT NOV] ABSTRACT PART I THE SYNTHESIS OF ETHYL 3.3-DI-(4-HYDROXY-3,5-DI-IEBI- BUTYL )PHENYL-Z -AMINOPROPIONATE (2,31) PART II NOVEL REACTION PRODUCTS OF 4-HYDROXY-2.3.4- TRIPHENYL-2-CYCLOPENTEN-1-ONE BY Mary Elizabeth Conner For the purpose of studying conductivity properties the synthesis of a stable free radical substituted amino acid was attempted. The precursor for this amino acid was ethyl 3,3-di—(4-hydroxy-3,5—di-tezt4butyl)phenyl-Z-amino- propionate (22). Amino ester 22 was synthesized by the Raney nickel catalyzed reduction of ethyl 3,3-di—(4-hydroxy-3,5- di-tgzt-butyl)pheny1-2-nitropropionate (1§)' The nitro ester (lg) was obtained by the addition of ethyl nitroace- tate to 2,6,3',5'-tetra:;§;;-butyl-4'-hydroxyphenyl-4- methylene-2,5-cyclohexadien-1-one (l1) (1) or to 4,4'-di- hydroxy-3,3',5,5'-tetra-§g§t-butyldiphenylethoxymethane (12). When oxidation of the side chain of amino ester gg'was attempted the amino ester moiety was cleaved giving only unsubstituted free radical. The desired free radical sub— stituted amino acid could not be obtained. When e in acetic a sulfuric ac benzophenon preparation Highly from novel CYClOpente: 1:2.3~trip} hydmxy ketc in tOIuene for its ide DehydJ m prOduCt the red 3’: Mary Elizabeth Conner When either cyclohexadienone lz'or ethoxy compound 12’ in acetic acid solution was treated with several drops of sulfuric acid, 4,4'-dihydroxy-3,3',5,5'-tetra-tg£t-butyl- benzophenone (fig) was obtained. Standard methods for the preparation of benzophenone 28 were unsuccessful. PART I I Highly colored reaction products were found to result from novel rearrangements of 4-hydroxy—2,3,4-triphenyl-2- cyclopenten-l-one (22) (2). A green crystalline compound, 1,2,3-triphenyl-4-azazulene (g2), resulted from reaction of hydroxy ketone 22' with pyrrolidine and p—toluenesulfonic acid in toluene. The spectral and chemical properties necessary for its identification are discussed. Dehydration of hydroxy-ketone gg'also results in color- ful products. In pftoluenesulfonic acid catalyzed reactions the red 3,3',4,4',5,5'-hexaphenyl-3,4-dihydro-2,2'-biscyclo- pentadienone (2g) was formed. In 10% sulfuric acid-acetic acid the yellow (§)-3,3',4,4',5,5'-hexaphenyl[bi—3-cyclo- penten-l-ylidenel-2,2'—dione (g1) is a minor product. The formation of both cyclone 2g and bicyclopentenylidene 21. can be rationalized gig rearrangement of a dimer of 2.3.4- triphenylcyclopentadienone (Q2). In the 10% acid solution a blue compound (g2) is the major product. On the basis of available evidence it was suggested to be 6-hydroxy-5,5 ,8, 9,10-pentapheny1benzo[ggjcyclopenth]azulen-4-(5gj—one. A new 1 pentadienom triphenyl-2 exist as a 1 precursor f 2,2'-biscyc cyclone kno The ch reactions 0 1' H-S.K ( 957) Z'C.p - x (1938), Mary Elizabeth C cnner A new cyclopentadienone, 2,3,4—triphenyl-5—bromocyclo- pentadienone (g2) was obtained from the bromination of 2,3,4- triphenyl-2-cyclopenten-1-one (22). Cyclone Q2 appears to exist as a monomer. Bicyclopentenylidene §l'served as a precursor for the preparation of 3,3',4,4',5,5'-hexapheny1- 2,2'—biscyclopentadienone (21), which is the simplest bis- cyclone known. The chemical and spectral data as well as further reactions of these compounds are discussed. REFERENCES 1. M. S. Kharasch and B. S. Joshi, J. Org} Chem., 22/ 1435 (1957). 2. C. F. Koelsch and T. A. Geissman, J. Org, Chem., g, 480 (1938). THE SYNTHESI NOVEL REAC in PART I THE SYNTHESIS OF ETHYL 3,3-DI-(4-HYDR0XY-3,s-DI-TERT-BUTYL)- PHENYL-Z—AMINOPROPIONATE PART II NOVEL REACTION PRODUCTS OF 4-HYDROXY-2,3,4-TRIPHENYL-2- CYCLOPENTEN-l-ONE BY Yak-9 Mary Elizabeth Conner A THESIS submitted to Midhigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1969 I wish for his pati this researc through his also like tc °f a Summer In add: 7- /~7<> ACKNOWLEDGMENTS I wish to express sincere appreciation to Dr. E. LeGoff for his patience and helpful suggestions in the course of this research and for the financial assistance provided me through his PRF. NSF, and NIH research grants. I would also like to thank NSF for financial assistance in the form of a summer fellowship in 1966. In addition, I would like to express appreciation to Mr. Eric Roach of Michigan State University for running 100 Mhz nmr spectra, and to Mrs. Lorraine Guile of Michigan State University and Rodger L. Foltz of Battelle Memorial Institute for providing mass spectral data. Sincere thanks also to Dr. A. A. Bothner-By of CarnegieaMellon University for sending us the LAOCOON III program description and FORTRAN deck, and to Kurt L. Loening of Chemical Abstracts Service for his assistance in naming several compounds. ii INTRODUCTION RESULTS AND EXPERIMENTAL 1. 2. Ger. 4,4 dig Etl phe Etl Phi OX 4. be Tr TABLE OF CONTENTS PART I INTRODUCTION . . . . . . . . . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . 1. 2. 5. 6. General Procedures . . . . . . . . . . . . 4, 4'-Dihydroxy-3, 3' ,5 5'—tetra-tLrt-butyl- diphenylethoxymethane (19). . . . . . Ethyl 3 ,3-di-(4-hydroxy-3, 5- di-tLrt-butyl)- phenyl—2-nitr0propionate (18) . . . . . . . Ethyl 3,3-di-(4-hydroxy—3,5-di-tert—butyl)- phenyl-Z-aminopropionate (22) . . . . . . . Oxidation of Amino Ester 20 . . . . . . . . 4 ,4'-Dihydrox -3, 3' 5, 5'-tetra- -tLrt—butyl- benzophenone 28) . . . . . . . . . . . . Trimethylsilyl Derivative of 28'. . . . . . PART II INTRODUCTION . . . . . . . . . . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . 1. 2. 3. 4. General Procedures . . . . . . . . . . . . 1,2,3-Tripheny1-4-azazulene (32) . . . . . 1, 2 ,3-Triphenyl-4-azatetrahydroazulene . . (E) )-3, 3' ,4 4‘ 5, 5'-Hexaphenyl[bi-3-cyclopent- 1-ylidene]- -2, 2'-dione (31) . . . . . . iii Page 23 23 23 24 25 26 26 27 28 35 72 72 72 73 73 TABLE OFC 5. R P 6. 3 P 7. R E 8. 'r 9. 3 b 10. R 11. 2 12. 6 c 13 . T 14. A 15. p BIBLIOGRAP TABLE OF CONTENTS (Cont.) Page 5. Ruthenium Tetroxide Oxidation of Bicyclo- pentenylidene (31) . . . . . . . . . . . . 75 6. 3,3',4,4',5,5'4Hexaphenyl—2,2'biscyclo- pentadienone (41) . . . . . . . . . . . . . 76 7. Reaction of Biscyclone 42 with Acetylenic Ester . . . . . . . . . . . . . . . . . . . 77 8. Thermal Rearrangement of Biscyclone 41' . . 77 9. 3,3',4,4',5,5'-Hexaphenyl-2,3-dihydro-2,2'- biscyclopentadienone (48) . . . . . . . . . 78 10. Reaction of Cyclone gg'with Acetylenic Ester 79 11. 2,3,4-Triphenyl-5-bromocyclopentadienone (52) 80 12. 6-Hydroxy-5,5 ,8,9,10-Pentaphenylbenzolggj— cyclopent[§]azulen-4-(5§)-one (32) . . . . 81 13. Trimethylsilyl Derivative of 32’. . . . . . 82 14. Acetate of 32’. . . . . . . . . . . . . . . 82 15. Potassium Permanganate Oxidation of 32’ . . 82 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 84 APPENDIX . . . . . . . . . . . . . . . . . . . . . . 86 iv flflmfi 1. Calcula 4-aza21 2. Compar tetrac 3- Comput triphe TABLE 1. LIST OF TABLES Page Calculated nmr arameters of 1,2,3-triphenyl- 4-azazulene (22', pyridine, and azulene . . 38 Comparison of absorptions of biscyclones and tetracyclones (20) . . . . . . . . . . . . . 55 Computer analysis of the mass spectrum of 1,2,3- triphenyl-4-azazulene (22) . . . . . . . . . 105 LIST OF SCHEMES Reaction Schemes, Part I 1. 2. 3. 4. Proposed reaction of 2,6-di-tert-butylphenol with 4-hydroxy-3,5-di-tert-butylbenzilidenes 10 Proposed reaction of 2,6-di-tert-butylphenol with 4-(3,5-di-tert-butyl-4-h droxybenzylidene)— 2-phenyl-2—oxazolin-5-one (12) . . . . . . . 12 Reaction of cyclohexadienone 11 with ethyl nitroacetate . . . . . . . . . . . . . . . . 15 Proposed incorporation of two amino acid groups on ga1v1noxyl precursor . . . . . . . 20 Mechanistic Schemes, Part II 1. Formation of 1,2,3-triphenyl-4-azazulene (22) 41 Formation of (§)-3,3',4,4',5,5'-Hexaphenyl- [bi-3-cyclopenten-1~ylidene]-2,2'-dione (22) 49 Formation of 6-Hydroxy-5,5 ,8,9,10-pentaphenyl- benzo[g§]cyclopent[£]azulen-4-(5§)-one (22). 70 FIGURE 10. 11. 12. 13. 14. 15. 16. 17. 18, 19. 20. Die! Lit Ext Int Sta Cle Din Cor of EX< FIGURE 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. LIST OF FIGURES Diagram of semiconductor and metal . . . . . Little's model of an organic superconductor . Extreme resonance forms of side chain . . . . Interaction of side chain and Spine . . . . . Stabilization of benzilidene anions . . . . . Cleavage of amino ester moiety . . . . . . . Dimerization of cyclopentadienones . . . . . Comparison of steric effects on dimerizations of triphenylcyclones . . . . . . . . . . . . Excited states of tetracyclones . . . . . . . Infrared spectrum of ethoxy compound 12’. . . Infrared spectrum of nitro ester 12'. . . . . Infrared spectrum of amino ester 22'. . . . . Infrared spectrum of benzophenone 22’ . . . . Infrared spectrum of azazulene 22’. . . . . . Infrared spectrum of bicyclopentenylidene 31. Infrared spectrum of biscyclone 33’ , . . , . Infrared spectrum of acetylenic ester adduct of biscyclone 22' . . . . . . . . . . . . . . Infrared Spectrum of cyclone 22'. . . . . . . Infrared spectrum of acetylenic ester adduct of cyclone 22, . . . . . . . . . . . . . . . Infrared spectrum of triphenylbromocyclone 52, vi 12 18 31 33 53 86 87 88 89 9O 91 92 93 94 95 96 LIST OF FIGI HQRE 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Infra Nmr : Nmr : Nmr: Nmr : Comp. of H Nmr Nmr Mass Mass LIST or FIGURES (Cont. FIGURE 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Infrared spectrum of azulene 22' . . Nmr spectrum of nitro ester 12’ . . . Nmr spectrum of amino ester 22’ . . . Nmr spectrum of biscyclopentenylidene 2’1, . . Nmr spectrum of azazulene 22’ . . . . Comparison of calculated and observed of H1 and H4 of azazulene 22' . . . . Nmr spectrum of cyclone 22' . . . . . Nmr Spectrum of azulene 22’ . . . . . Mass Spectrum plot of azazulene 22' . Mass spectrum plot of azulene 22' . . vii spectra Page 97 98 98 99 99 100 101 101 102 103 I NTRODUCT ION The preparation of an organic molecule having conduc- tivity properties comparable to those of metallic conductors presents a challenge to the synthetic chemist. Various models have been proposed but no working model has yet been synthesized. In contrast to metallic conductors, semiconductors have a finite energy gap or forbidden hand between the valence band or ground state and the conduction band (Figure 1). v f \\ " 2:212:12: 5 Forbidden Band 7/ r/7//// /// / I .Valence band Valence Band //////////////A //////J//J /1 Semiconductor Metal Figure 1. Diagram of semiconductor and metal. (1) Thermal excitation will occasionally raise an electron from the ground state to the excited conducting state. As more electrons obtain sufficient energy to cross over the for- 'bidden band the conductivity increases. 1 .. _.‘ ’1‘ L In org; is generallj of one mole: cules. The tions resul1 transfer be free electn to observe 7 molecular o: In Spi‘ in? organic Mile], of a 1 conductivit: Conductors zero, Littl. Organic Com room temper a SuPeI‘Cond to travel t 2 In organic crystals the overlap of molecular orbitals is generally very small, even to the extent that orbitals of one molecule may be isolated from those of other mole- cules. The relative weakness of the intermolecular attrac- tions results in low electron mobility. Since electron transfer between molecules must occur by tunneling, any free electrons formed may be very localized (2). In order to observe enhanced conductivity, intermolecular mixing of molecular orbitals is required. In spite of the difficulties encountered in synthesiz- ing organic semiconductors, W. A. Little has presented a model of a molecule which he predicts will possess super— conductivity properties (3,4,5). Whereas metallic super- conductors function as such only at temperatures near absolute zero, Little suggests the possibility that certain types of organic compounds may be superconducting near or even above room temperature. In simplified terms, when a metal enters a superconducting state it gives up electrons which are free to travel through the ionic lattice. An electron passing an ionic center, and that ionic center are attracted to each other. Due to the greater velocity of the electron, it is beyond the ionic center when the ionic center has reached maximum displacement. However, before the ionic center has returned to its initial position, a second electron passing this center is also attracted to it. The ionic center has, in effect, caused a binding or pairing of the two electrons. {Ehis pairing is the basis for quantum mechanical considerations of superco where the than offse dam of the too high, pairing ar Litt? main featx states am 0f electn the addit, tromelec e“emetic. The backb. Chains are In the Si can reson {Figure 3 effect On conduCtOr the metal 3 of superconductivity. Pairing occurs at low temperatures where the increase in energy resulting from pairing more than offsets the disadvantages involved in the loss of free- dom of the individual electrons. If the temperature becomes too high, thermal agitation will eventually break up the pairing and disrupt the superconducting state. Little‘s model of an organic system consists of two main features: a spine in which electrons fill the various states and a series of side chains which provide interaction of electrons in the spine. Even if the spine is an insulator, the addition of side chains can conceivably increase elec- tron-electron interaction to a point where it becomes energetically favorable to enter a superconducting state. The backbone of Little's model is a polyene and the side chains are molecules of the dye, diethylcyanine iodide (Figure 2) . In the side chain the positive charge is delocalized and can resonate from one end of the side chain to the other (Figure 3). This ionic site should have the same pairing effect on electrons as the ionic center in a metallic super- conductor. The difference, however, is that in contrast to the metallic case where a metal ion is being displaced, it is the motion of an electron which causes the ionic movement in the side chain. The temperature below which superconduc- tivity can occur is dependent upon the mass displaced. Since the mass of an electron is very small compared to that of a metal ion, Little's calculations predict superconductivity for this polymer to occur below 2000°K (5). Figure 2 . Ceasfli‘ Figure 3 . Figure 2. Little's model of an organic superconductor. (3) O 1' 0+ ‘9 CH \ / -0235 Figure 3. Extreme resonance forms of side chain. (3) 5 The accuracy of his calculations and the value of his model are at best only of theoretical interest. The syn- thesis of this particular molecule with its rigid specifica- tions is at present too demanding. However, it may be pos- sible to determine the validity of Little's model by incor- porating electron deficient side chains on a known spine and determine the effect of the side chains on its conduc- tivity relative to the unsubstituted compound. Greatly enhanced properties would warrant further attempts at syn- thesizing organic conductors having such features. Experimental evidence suggests that a polypeptide might be a useful backbone. As early as 1941 Szent-Gybrgyi sug- gested that proteins have energy levels associated with the whole protein rather than with any particular unit within the protein (6). In his laboratory the conductivity of protein films was found to increase considerably by irradia- tion (7). The conductivity of proteins could also be in- creased by the addition of small amounts of chloranil (2). In terms of acceptance of electrons by the chloranil, . transitory cations were left in the valence band. The incor- poration of an electron deficient side chain on a polypeptide spine should show the same effect as chloranil in accepting electrons from the spine. One useful side chain might be the galvinoxyl free radical (I). Galvinoxyl is an electron deficient free radical having an intense blue color (8,9). If this radical could be attached as a side chain on an amino acid the energy gap may be decreased. In addition, the O O *‘X I ,INH2 on 0-03 :: ‘\ooen ° 0 ,1. E, polypeptide which could then be formed would be an inter— esting polymer for conductivity studies. The electron deficient side chain may accept electrons from the spine, thus increasing the conductivity of the polypeptide. This interaction may occur between the radical site and the amide bond (Figure 4). The actual effect of the side chain could then be determined by comparison with the unsubstituted polypeptide. Figure 4. Interaction of side chain and spine. RESULTS AND DISCUSSION The synthesis of f,4'-dihydroxy-3,3',5,5'-tetra-tert- butyldiphenyl methane (3), the precursor to galvinoxyl, involves the condensation of 2,6-di-tert-butylphenol ‘with formaldehyde (8). OH OH + RCHD -——-——- > can OK i R=H /NHCOPh 2. R= \ coast COCH e. = 3 Xcozst)2 By choosing the appropriate aldehyde in this conden- sation compounds 4' and Q, might be obtained. However, the desired aldehydes could not be obtained. Several 8 9 attempts at synthesizing ethyl a-formylhippurate (Q) by con- densation of ethyl hippurate and ethyl formate failed. CHO I phcouchacoant + HCOzEt -)é—¢- PhCONHCHCOzEt E, Likewise, formylation of ethyl acetamidomalonate under Vilsmeier conditions failed to give ethyl a-formylacetamidomalonate (Z). In this case N—formylation rather than C-formylation apparently occurred. COgEt Et02C\ v I ,CHNHCOCHa + POC13 + DMF —§é—v OHCCHNHCOCH3 Eto c ' 3 COgEt 1 Another compound which appeared useful was 3,5-di- tert-butyl-4-hydroxybenzaldehyde (10). Benzaldehyde §. has been condensed with various active methylene compounds to give benzylidenes g, 12, and ll'(10). .Midhael addition of 2,6-di-tert-butylphenol to the benzylidene would give a galvanoxyl precursor 13, Thus a benzylidene having substit- uents readily convertible to an amino acid was desired. 10 OH X + CH2 \Y > HO Q g X=CN,Y=C02Et 1g x=Y=co,Et 1; X=Y=CN Lg X=C03Et, Y=N03 OH OH H OH 13 Reaction Scheme 1. Proposed reaction of 2,6-di-tert-butyl- phenol with 4-hydroxy-3,5-di-tert- butylbenzYlidenes. Condensation of ethyl nitroacetate with benzaldehyde 3 should give nitro ester 12; Although condensation was 11 attempted using piperidine in benzene, potassium fluoride in dimethylsulfoxide, and potassium fluoride in refluxing benzene suitable conditions could not be found for this condensation. It is possible that condensation occurs, but due to the strong electron attracting nitro and ester groups, retroreaction to starting materials becomes favor- able. Another potential precursor, 4-(3,5-di-Eg£§-butyl-4- hydroxybenzylidene)-2-phenyl—3-oxazolin-5—one (14), re- sulting from the condensation of benzaldehyde .§' with hippuric acid in acetic anhydride has been reported (11). g + PhCONHCHzcozH ———> 12, Michael addition of 2,6-di-tert-butylphenol to lé'would give a convenient precursor l§,to an amino acid. When such Michael condensations were attempted using benzylidenes g, 12, and ll'as model compounds, addition did not occur. Likewise, Michael addition to lé,did not occur. Since these condensations were attempted under basic conditions the failure to add is probably due to the enhanced acidity of I Fm: i '5! Ll... )l—A Reactior the benz formatio due to d 12 on a 0 0 1:1, + ———> on ' N on 15 Reaction Scheme 2. Proposed reaction of 2,6-di-tert-butyl- phenol with 4-(3,5-di-tert-butyl-4— hydro benzylidene)-2-phenyl-3-oxazolin- 5-one E4) the benzylidenes relative to 2,6-di-tert-butylphenol. Thus formation of anions from the benzylidene compounds is favored due to delocalization of the charge into the substituents. 2.22.22, 14 Figure 5. Stabilization of benzylidene anions. Em. In of a pr tion of silyl )a wise, B densati 13 In an attempt to prevent anion lg'from forming the use of a protecting group was investigated. Trimethylsilyla- tion of 2,6-di-tgrt-butylphenol with N,0-bis(trimethyl- silyl)acetamide (BSA) occurs in excellent yield (12). Like- wise, BSA gave excellent results with 12; Subseguent con- densation of 2,6-di-tert-butylphenol with lfi'by using either 31(033)3 /031(OH§)3 1’4" + OH30\ _"—'> 0 \N31( CH3 )3 /"\\ H \IZ‘Z: Ph 16 w an alkoxide or 1,5-diazbicyClol4.3.0]non-5-ene (13) as base, gave a deep red solution from which the unsilylated compound 14,was isolated. Apparently exchange or displace- ment of the trimethylsilyl group with the base occurred, producing the red anion of 14, The use of milder bases was ineffective. Another approach to galvinoxyl precursor lg'which appeared promising was a 1,6-addition across 2,3',5',6- tetra-tsrt-butyl-4'—hydroxyphenyl-4-methylene-2,5-cyclo- hexadiene—l-one.(11) (8). Condensation in basic solutions could not be achieved due to formation of the purple anion of 11, Even when the hydroxyl group was protected with a trime th base. resulte ethyl n: EVer , a 33-di-{ (Hi): We butyldip 14 O /X OH + CH2 > 13 \Y MI I OH 17 "W trimethylsilyl group, the protecting group was cleaved in base. Purple solutions characteristic of the anion of 11' resulted. Some addition did occur when cyclohexadienone lz,and ethyl nitroacetate were allowed to reaCt in benzene. How- ever, a better precursor for the condensation product, ethyl- 3,3-di-(4-hydroxy-3,5-di-Egrt-butylphenyl)-2-nitro-propionate (1§)' was found to be 4,4'-dihydroxy-3,3'.5,5'—tetra-tert- butyldiphenylethoxymethane (19). Ethoxy compound 12.was 3 H + EtOH -—————> 033’ OHOEt on on 19 obtained w diphenylbr ture of th analysis, . infrared a‘ with ethyl nitro estei 15 obtained when 4,4'-dihydroxy-3,3',5,5'-tetra-tert-butyl- diphenylbromomethane (8) was heated in ethanol. The struc- ture of the ethoxy compound (12) was supported by C, H analysis, and the presence of an ether absOrption in the infrared at 1075 cm-1. When ethoxy compound lg'was reacted with ethyl nitroacetate in ethanol excellent yields of H i 1102 nitro ester 1§,resulted. llor 12 + ochnzcoaet ——> O COgEt OH 18 m Reaction Scheme 3. Reaction of cyclohexadienone lZ/with ethyl nitroacetate. The structure of lg'was substantiated by C, H analysis and spectral data. In the infrared a sharp peak is present at 3630 cm-1, indicative of a hindered phenol. In addition, absorptions appear at 1750 and 1560 cm.1 corresponding to an ester carbonyl and a nitro group, respectively. The ether ab- 1 sorption at 1075 cm- is no longer present. (In the nmr there are overlapping singlets present at 6 7.15 and 7.11 (phenyl hydrogens) as well as at 6 5.15 and 5.11 (hydroxyl hydrogens). The nonequivalence of the phenyl hydrogens as well as the 16 nonequivalence of the hydroxyl hydrogens may be attributed to the presence of an asymmetric center in the molecule. Two doublets are centered at 6 5.80 and 4.79 (J = 12 Hz) corresponding to the methine hydrogens. The ester quartet is centered at 6 3.92 (J = 7.5 Hz) and the triplet at 6 0.89 (J = 7.5 Hz), whereas the tert-butyl hydrogens all appear as a singlet at 6 1.41. The next step in the synthesis was the reduction of the nitro ester to an amino ester. Some problems were en- countered in this reduction. Adams catalyst failed to ef- fect reduction after 12 hours at atmospheric pressure or after 5 hOurs in a Paar apparatus at initial pressure of 50 psi. 'Reduction'with iron in hydrochloric acid or with aluminum amalgam in ethanol gave complex mixtures of reduc- tion products. Finally, it was found that the use of a Raney nickel catalyst in a Paar apparatus at initial pres- sure of 50 psi, gave after 6 days nearly quantitative reduc- tion of the nitro group. H 18' + Raney Ni ‘——————> OH-dfi::n2 OzEt OH 20 17 The structure of ethyl-3,3-di-(4-hydroxy-3,5-di- Eggt-butylphenyl)-2-aminopropionate (22) is consistent with the C, H analysis as well as with Spectral data. In the infrared a carbonyl absorption is present at 1730 cm-1 and a hindered phenol at 3640 cm-1. Instead of the nitro absorp- tion at 1560 cm.1 there is an amine absorption at 3310 and 3375 cm-1. In the nmr overlapping singlets are present at 6 7.15 and 7.11 (phenyl hydrogens). The Egrt-butyl hydro- gens give rise to a singlet at 6 1.42. The ester consists of a quartet centered at 6 3.87 (J = 7.5 Hz) and a triplet at 6 0.87 (J = 7.5 Hz). The methine protons are buried under the ester quartet. The only other absorption present is a very broad one at 6 5.13. This is probably due to chemical exchange occurring between the nOnequivalent hydroxyl and amino hydrogens (13). It was anticipated that oxidation of amino ester 22' would give the desired galvinoxyl amino acid (2;). O 20 Oxidation ) 43/“? “W \00213 21 18 Unfortunately, when this oxidation was attempted by using potassium ferricyanide or lead dioxide, the amino ester moiety was cleaved giving only the unsubstituted galvanoxyl radical (I). This result is reasonable considering a free radical oxidation mechanism. Initial hydrogen abstraction gives a phenoxy radical. The radical site at the para position can undergo cleavage of the most stable a-sub- stituent. The final oxidation product is thus galvinoxyl. oa—orilm2 ”/1132 0°21"t .JOH‘ooaxt Figure 6. Cleavage of the.amino ester moiety. This problem might be circumvented if two amino ester groups could be substituted at the benzilic positions. The proposed scheme for accomplishing this consisted of a) NBS bromination of nitro ester 18, b) elimination of hydrogen bromide, c) addition of ethyl nitroacetate, d) reduction of the nitro groups, and e) oxidation to the galvinoxyl amino ester. C) 19 e+ms > g}; + ozncnzcozst OH | /N 02 Bro-0&0 I 02313 20 OH . H N /N d) 22’ Reduction > 2 \UH U oq\oflé BtOQO/ 02m OH 25 e) 22' Oxidation > El. Reaction Scheme 4. Proposed incorporation of two amino acid groups on galvanoxyl precursor. The initial NBS step again resulted in cleavage of the amino ester group giving cyclohexadienone 11;. Apparently the same type of free radical side chain cleavage occurred. Since no further ways of avoiding this cleavage could be envisioned this synthesis was abandoned. In the course of the preceding synthesis an interest- ing reaction was discovered. Protonation of cyclohexadi- enone 11,in acetic acid -- sulfuric acid should give carbonium ion 26, It was anticipated that phenyl alanine would under- go alkylation with this carbonium ion giving the substituted phenyl alanine (21). However, phenyl alanine did not par- ticipate in the reaction. Instead, a colorless compound which analyzed for C29H4203 was obtained. The hindered phenol 1 absorbs at 3610 cm- in the infrared and the carbonyl absorbs at 1660 cm-1. The infrared absorption as well as the fact that the Of a benz singlets tiri‘bu t} 21 9" on + /NH A 30 + @032 on 2 > no 03203 112 l \ oo 3 \oo 3 2 2 on on 22. 21 that the compound formed a red 2,4—DNP indicates the presence of a benzophenone-like carbonyl. The nmr consists of three singlets at 6 7.75 (2H), 5.70 (1H), and 1.48 (18H). These data are consistent with 4,4'-dihydroxy-3,3' ,5,5'-tetra- tert-butylbenzophenone (2,8) . OH 28 Th formed ‘ methyls absor pt nmr, as 22 The bistrimethylsilyl derivative of 28 was readily formed by using BSA. The analysis indicated that two tri— methylsilyl groups are present. In addition, the hydroxyl absorptions are no longer present in the infrared. The nmr, as expected, consists of three singlets at 6 7.75 (25), 1.44 (18H), and 0.43 (9H). Benzophenone 18 was obtained in about 80% yield when a solution of either cyclohexadienone lz'or ethoxy compound lg'was stirred in acetic acid - sulfuric acid solution in an open flask. The reaction apparently proceeds xi§_oxida— tion of carbonium ion 26; An alternate synthesis involving Friedel-Crafts reac- tions of 2,6-di-Egrtfbutylphenol with carbon tetrachloride or phosgene using catalysts such as aluminum chloride, or stannic chloride, resulted in cleavage of the Eggtfbutyl moiety. In view of the failure of conventional synthetic methods, this carbonium ion oxidation represents a valuable synthetic preparation of benzophenone 28, ,. , —. 1.. "V'L u was. . 1. Gene The Model 23 A-60, 56 tions of Ultravio instrume Low Hitachi resoluti Institut eter. Mel Gamma: Mic Laborato t°ries i EXPERIMENTAL 1. general Procedures The infrared spectra were recorded on a Perkin-Elmer Model 2373 spectrometer. The nmr were obtained on a Varian A-60, 56-60D, or HA-100 spectrometer. Computer calcula- tions of nmr spectra were done on a CDC 3600 computer. Ultraviolet spectra were recorded on a Unicam Model SP—800 instrument using 1 cm quartz cells. 0 Low resolution mass Spectra were obtained with a Hitachi Perkin-Elmer RMU-6 mass Spectrometer. The high resolution mass Spectra were run at Battelle Memorial Institute on an A.E.I. MS-9 double-focusing mass Spectrom- eter. Melting points were determined on a Thomas Hoover capillary melting point apparatus and are uncorrected. Microanalyses were performed by Spang Microanalytical Laboratory, Ann Arbor, Michigan, or by Galbraith Labora- tories in Knoxville, Tennessee. 2. 444'-Dihydroxyr3L3'l5,5'—tetr§:;§£;:butyldiphenylethoxyr methane (12) A solution of 3 g of 4,4'-dihydroxy—3,3',5,5'-tert- butyldiphenylbromomethane (8) in 20 ml of absolute ethanol 23 and 1 Upon c Two re gave a (hinde 5 7.11 J-71 24 and 1 ml of hydrochloric acid was refluxed for 2 hrs. Upon cooling, 1.5 g (54%) of yellow product was obtained. Two recrystallizations from ethanol and one from pentane gave a colorless product: mp 162-1630; ir (CHC13) 3640 (hindered phenol) and 1075 cm.1 (ether); nmr (CDC13) 5 7.11 (s, 4H), 5.21 (s, 1H), 5.67 (s, 2H), 3.51 (q, 2H, J - 7 Hz), 1.41 (s, 36H) and 1.25 (t, 3H, J = 7 Hz). 5231, Calcd. for C31H4803: c, 79.49; H, 10.25. Found: C, 79.38; H, 10.23. 3. Ethy1»343-di—(4-hydroxy-3J5-di- -butyl)phenyl-2- nitro propionate (18) a) From cyclohexadienone 12; To 3 g (0.0071 mole) of cycldhexadienone in 20 ml of benzene was added 3 g of ethyl nitroacetate (0.0213 mole) and the mixture heated to reflux for 24 hours. The benzene was removed on a rotary evaporator and the residue dissolved in ethanol. The first crop was starting material (1.0 g). The second crop con- tained 0.2 g of product: mp 153-1540; ir (CHCla) 3620 (hindered phenol), 1745 (ester c=o) and 1560 cm.1 (nitro); nmr (CDC13) 6 7.15, 7.11 (overlapping S, 4H), 5.80 (d, 1H, J = 13 Hz), 5.15, 5.11 (overlapping S, 2H), 4.79 (d, 1H, J = 12 Hz), 3.92 (q, 2H. J = 7.5), 1.41 (s, 36H), and 0.89 (t, 3H, J = 7.5).. ‘ - Anal, Calcd. for C33H49N06: C, 71.23; H, 8.89. Found: C, 71.32; H, 8.77. b) 1 diphenylei in 200 ml acetate . the etham in an eva; from etha to nitroe 25 b) From 4,4'—dihydroxy-3,3',5,5'-tetra-tg£§-butyl- diphenylethoxymethane (12): To 37 g of ethoxy compound 12’ in 200 ml of absolute ethanol was added 15 g of ethyl nitro- acetate. The solution was heated to reflux for 2 hrs and the ethanol solution then allowed to evaporate to dryness in an evaporating dish. The residue was recrystallized from ethanol to give 40.7 g (93.5%) of product identical to nitroester prepared in procedure a. 4. Ethyl 3L3-di-(4-hydroxy-3ygedi-terL-butyl)phenyl-Z- amino:pr9pionate (22) A mixture of 10.7 g of nitroester 12'and 2 g of freshly prepared Raney nickel catalyst in 200 ml of isopropanol was hydrogenated in a Paar apparatus at initial pressure of 49 psi. After 6 days the pressure was 37 psi, and no further uptake was evident. The mixture was filtered to remove the catalyst and the filtrate evaporated to dryness (reduced pressure). The residue was crystallized from ethanol to give 9.94 g (98%) of reduction product: mp 141- 1420; ir (c3013) 3610 (on), 3370, 3290 (N32), and 1725 cm'1 (ester 0:0); nmr (CDClg) 6 7.15, 7.11 (overlapping S, 4H), 3.87 (q, 2H, J ' 7.5 Hz), 3.87 (2H buried under quartet), 1.42 (S, 36H). 0.87 (t, 3H, J =(7.5 Hz) and 5.3-4.9 (very 'broad absorption). 8 A 5231, Calcd. for C33H51NO4: C, 75.38; H, 9.78. Found: C, 75.45; H, 9.80. 3.94. SOdlUJ added perio for 3 washe solve ident and 1 Over: the : SOlv. form of g 26 5. Oxidation of Amino Ester 22’ a) Potassium ferricyanide: To a stirred mixture of 3.94 g (0.012 mole) of potassium ferricyanide, 0.6 g of sodium hydroxide, 15 ml of water and 80 ml of benzene was added 1.73 g of amino ester 22'in 15 ml of benzene over a period of 30 minutes. The mixture was stirred under nitrogen for 30 minutes. The layers were separated and the benzene washed with water and dried (M9304). Upon removal of the solvent 0.9 g of purple crystals remained. The ir was identical with that of galvinoxyl. b) Lead dioxide: A mixture of 0.3 g of amino ester 2g and 1.5 g of lead dioxide in 20 ml of benzene was stirred overnight. The mixture was filtered through filter cel and the solvent removed from the filtrate. The residue was dis- solved in chloroform and filtered. Removal of the chloro- form gave a purple residue with an ir identical with that of galvinoxyl. 6. 4J4'-Dihydroxy-3i3'-5,5'—tetra;;§;;:butylbenzophenone (22) a) From ethoxy compound 12; A solution of 0.5 g of ethoxy compound 12'in 5 ml of acetic acid containing 3 drops of concentrated sulfuric acid was stirred in an open flask for 24 hrs. The solid which separated was isolated by fil- tration. Water was slowly added to the filtrate to give additional product. The combined solids were dried and recrysi (81%) < 182°; : ((3:0): (5, 18} b} procedt acetic acid ga Tc acetonj (0.0075 12 1/2 27 recrystallized from carbon tetrachloride to give 0.38 g (81%) of benzophenone 22; mp 221-2220; 2,4-DNP mp 180- 182°; ir (CHCl3) 3620 (hindered phenol), and 1635 cm-1 (c=0); nmr (CDC13) o 7.75 (s, 2H), 5.70 (s, 13) and 1.48 (s, 18H). ' ' ' . Agg1, Calcd. for C29H4203: C, 79.50; H, 9.59. Found: C, 79.45; H, 9.65. b) From cyclohexadienone 11; Following the same procedure as in a, 2 g of cyclohexadienone in 25 ml of acetic acid containing 10 dr0ps of concentrated sulfuric acid gave 1.6 g (78%) of benzophenone 22; 7. 4i4'-Di:(trimethylsilyl)—3,3',5,5'-tetra—§§;;-butyl- benzophenone ‘ To 1.09 g (0.0025 mole) of benzophenone 22'in 8 ml of acetonitrile (freshly distilled from P205) was added 1.5 9 (0.0075 mole) of BSA. The mixture was heated to reflux for 12 1/2 hrs. The solution was allowed to cool to room temper- ature during which time colorless crystals separated. After filtration and washing with solvent 1.0 g (80%) of product 'was Obtained: mp 183-1840; ir (CHcla) 164s (c=o), 1260, 1225, 835 and 750 cm.1 (trimethylsilyl); nmr (CDC13) 6 7.75 (s, 23), 1.44 (s, 18H), and 0.43 (s, 93). - 2331, Calcd. for C35H538i303: C, 72.10; H, 10.03. Found: C, 72.10; H, 10.06. PART II NOVEL REACTION PRODUCTS OF 4-HYDROXY-2,3.4- TRIPHENYL-Z-CYCLOPENTEN-1-ONE INTRODUCTION The synthesis of 4-hydroxy-2,3,4-triphenyl-2-cyclopent- en-1-one (22) via condensation of phenylacetone with benzil was first reported in 1934 by Dilthey and Hurtig (15). 0 2h 0 o Phcnacocn3 + Ph-C-C-Ph > In OH Ph 29 w Such structural features as the carbonyl, active methylene, and hydroxyl groups make this compound a valuable starting material for the synthesis of theoretically interesting molecules. Fused ring systems might be obtained by conden- sation across the carbonyl and active methylene of hydroxy- ketone 22; The hydroxyl group would then provide a means for introducing another double bond into the system. 28 29 29 One precursor for accomplishing such cyclizations would be the pyrrolidine enamine of hydroxy-ketone 22; With such cyclizations as the objective the synthesis of the enamine was attempted. An attempt to prepare this enamine from pyrrolidine and 22 resulted instead in the formation of a dark green crystalline compound (22). Ph Ph TSOH 2.2+ (3%96 00 N Ph OH H One reaction which might have occurred is a simple dehydration by‘p-toluenesulfonic acid, which is used as a catalyst in the reaction. However, the acid dehydration of hydroxy-ketone 22'has been reported (16) and none of the products are identiCal with green compound 22; These de- hydration products are interesting Since they are also highly colored. The minor product (21) is yellow and the 30 major product (22) is blue. A trace of red product was also reported. These same products also resulted from the hydrolysis of Sfpfdimethylaminoanilino-l,2,3-triphenyl- cyclopentadienone (22) (16). Ph 0 Green Product 22 TSOH ,’«~*”’* 0 / Pyrrolidine 1h 2h GI \\\\\\\ 10% H2S04-H0Ac 22. \\\\\\\’ Yellow Product 21 92 ////2 Blue Product 22’ HCl‘HzO Red Product (trace) An obvious common reaction product which could result from either the dehydration or hydrolysis reaction is 2,3,4-triphenylcyclopentadienone (22). In fact, cyclone 22 was the structure assigned to the blue product. The yellow product was reported to be a dimer but not the usual cyclopentadienone dimer. V An investigation of the properties of cyclopentadi- Eanones (1) (or cyclones) casts doubt on the assignment of 31 a cyclone structure to either blue compound 22'or green com- pound 22; ° 1 p. O 205 Ph Ph 3 4 3.3.4. I, Cyclones are in general very reactive compounds. Molecular orbital calculations predict a very reactive 2,3 double bond as well as strong dienic ability (17). In the absence of other reactants, cyclones may dimerize ylg_a DielSeAlder reaction. This dimerization is sterically dependent upon the bulk of the substituents on the cyclone ring. 34 Figure 7. Dimerization of cyclopentadienones. The parent cyclopentadienone exists exclusively in the dimeric form and has eluded all attempts at isolation (18). 32 On the other hand, dimerization is completely inhibited in the case of tetraarylcyclones, although they still func- tion as reactive dienes (19). Steric factors also inhibit the dimerization of 2,3,5-tri(tert-butyl)cyclone, 2-methyl— 3,4,5-triphenylcyclone, and tetra(trifluoromethyl)cyclone (20). Certain other cyclones such as 2,5-dimethyl-3,4-diphenyl- cyclone, 2-methyl-3,4-diphenyl-5-ethylcyclone, and 2,3,5- triphenylcyclone exist as temperature dependent dissociating dimers (20). For example, the hydrolysis of ngfdimethyl- aminoanilino—l,2,4-triphenylcyclopentadiene (22) is reported to give 2,3,5-triphenylcyclone dimer (222), which is in thermal equilibrium with its monomer (22) in solution (21). The colorless dimer can be isolated but when solutions of the dimer are heated the characteristic red color of cyclone monomers results. Ph Inez . 0 D . .. .. Q , > —> <— Fh 2h 22 1:52, is, In view of the steric factors influencing dimerization, intuitively it might be predicted that 2,3,4-triphenyl- cyclone (22) should dimerize more readily than 33 2,3,5-triphenylcyclone (22). Addition of the diene portion of one molecule is sterically more favorable across the less substituted double bond of a second molecule. In the case of 2,3,4-triphenylcyclone the addition should be more favorable because the termini of the diene portion are substituted with one hydrogen and only one phenyl group. The diene portion of 2,3,5-triphenylcyclone, however, has two phenyl substituents at its termini. In each case the steric effects of the dienophilic double bonds are comparable. Figure 8. Comparison of steric effects on dimerizations of triphenylcyclones. It appears that the existence of 2,3,4-triphenylcyclone in exclusively the monomeric form is unlikely. If, however, the monomer could exist it would be expected to be the char- acteristic red color of cyclone monomers. If a cyclone 34 structure were to be suggested for either blue compound 22’ or green compound 22; a satisfactory explanation for the large bathochromic shift would have to be made. There is no pecularity in the structure of 2,3,4-triphenylcyclone which could account for this Shift. Thus, a cyclone struc- ture for either green compound 22,or blue compound 22’is discredited because of color and because of steric effects which influence dimerization. Whereas one of the standard methods for the preparation of cyclopentadienones involves the dehydration of 4-hydroxy- 2-cyclopenten-1-ones it is apparent that dehydration of hydroxy-ketone 22’occurs in a completely different manner. The reaction products are unusual not only because they are not normal products, but also because they are so intensely colored. This suggests that the most favorable reaction paths available are ones which lead to highly stable con— jugated products. In order to determine why these reactions occur as they do it was necessary to attempt to identify these novel reaction products. It was anticipated that once the structures of the products were known a reasonable ex- planation of the reaction paths could be made. The discussion which follows concerns the identification of these products based on spectral and chemical evidence. RESULTS AND DISCUSSION Identification of 2,3,4-Triphenyl—4-azazulene (30) The green product resulting from the reaction of hydroxy-ketone 22'with pyrrolidine was found to be 1,2,3- triphenyl-4-azazulene (22). This novel 4-azazulene struc- ture is well substantiated by spectral data. Its intense 30 rw color and absorptions in the visible region at 610, 654 and 710 mu (acetonitrile) are immediately suggestive of an azulene type structure. The molecular formula C37H19N, de— rived from the elemental analysis and mass Spectrum, is also consistent with the proposed structure. A computer analysis of the high resolution mass spec- trum (22) shows initially only fragmentation of hydrogens from the parent molecular ion at mass 357. The most favor- able fragmentation is the loSS of a phenyl group. In 35 36 addition, there iS an abundance of peaks attributed to doubly ionized fragments. In general such an abundance of doubly ionized fragments is found only in highly stabilized molecules such as fused ring aromati:s(23). Furthermore, the absence of functional groups in this molecule is evi- dent from the infrared spectrum. The nmr Spectrum (Figure 25) is also consistent with an aromatic compound. (The nmr consists of two low field multiplets at 6 8.33-8.08 and 8.07-7.82, each due to one hydrogen, and a higher field multiplet at 6 7.32-6.82 due to 17 hydrogens. These chemical shifts are similar to the chemical Shifts of the hydrogens on azulene (Table I) (24). Of the 17 hydrogens in the highest field multiplet of azazulene 22; 15 are due to the phenyl hydrogens. The re- maining two hydrogens in this region are due to two hydro- gens on the azazulene ring. Since no significant spin interaction could occur between the phenyl hydrogens and the hydrogens on the azazulene ring, the nmr can be analyzed as a four spin case. This was accomplished by use of the LAOCOON III computer program (25). In order to determine the relative chemical shifts and coupling constants to be used as input in LAOCOON III, the two low field hydrogens were decoupled from each other. When each of the two low field hydrogens was decoupled from the other, the patterns of these two hydrogens coalesced to four lines each. Thus, they appeared to be weakly coupled to each other and each, in turn, coupled more strongly to the two higher field 37 hydrogens. Because the weakest coupling occurred between the two low field hydrogens they were assumed to be H1 and H4 (Table I). The relative positions of H1 and H4 were based on comparison with pyridine and azulene (Table I). The adjacent nitrogen should cause H1 to be shifted down- field relative to the corresponding hydrogen on azulene (Table I). In addition, H4 is in the shielding region of the nearby phenyl group and should appear at higher field than the corresponding hydrogen on azulene. Thus, H1 was assumed to be at lowest field. The two higher field absorptions, which are obscured by the phenyl multiplet, are due to H2 and H3. Of these two, Hz was assumed to be at lower field by comparison with the corresponding hydrogens on azulene. Using these relative assignments, the estimated chemical shifts and coupling constants were refined until a good match of the calculated and experimental Spectra of the low field region (460-510 Hz ) was obtained (Figure 26). The refined values of the chemical Shifts and coupling constants are listed in Table I. The values for pyridine and azulene are also given for comparison. By comparison of the calculated values of the chemical Shifts of the azazulene hydrogens with the corresponding hydrogens on pyridineand azulene, the relative positions of the azazulene hydrogens appear reasonable. For example, the relative positions of H3 and H3 might be obtained by comparing the difference in chemical shifts of H2 and 38 Table I. Calculated nmr spectra of 1,2,3-triphenylazazulene (22), pyridine (25), and azulene ( Hydro- Chem gen Shift (ppm) (Hz) (Hz) ‘ 1 495.8 8.26 1,2 6.52 2 425.9 7.10 1,3 1.11 3 420.3 7.05 1,4 0.267 4 478.4 7.97 2,3 9.46 2,4 1.30 3,4 8.96 22' 1 516.5 8.6 1,2 4.88 3 4.5 4.88 2 427.4 7.12 1,3 1.84 4 ,I’ 2 3,5 1.84 3 450.1 7.5 1,4 0.995 5 \\ 1 2,5 0.995 N 4 427.4 7.12 1,5 -0.132 5 516.5 8.6 2.3 7.67 3,4 7.67 2,4 1.37 8 1 7 1 7.27 1,2 4.0 2 7.80 2,3 4.0 ‘4‘!!! 3 7.27 4,5 9.5 2 6 4 8.22 7,8 9.5 5 7.04 5,6 10.3 5 6 7.47 6,7 10.3 3 4 7 7.04 4,6 1.5 8 8.22 6,8 1.5 39 H3, with the difference in chemical Shifts of the corre- sponding hydrogens on pyridine and azulene. An estimate of the difference in the chemical Shifts of H2 and H3 might be: AH2,H3(azazulene) = AH2,H3(pyridine)+-AH5,H5(azulene) By substituting the values from Table I for pyridine and azulene in the lefthand side, a value of 0.05 ppm is obtained. This same value is obtained by subtracting Hz-H3 for the azazulene hydrogens. This result lends some support to the relative assignments assumed when making the calculations. In any case, a plot of the calculated downfield portion of the nmr (1224, H1 and H4) of azazulene using the param- eters in Table I results in an excellent match with the observed spectrum. However, since the lines for H2 and H3 are buried under the phenyl region it is not known whether they coincide exactly with the calculated lines in that region. Thus, the calculated parameters may not be exact. The important result iS that the observed spectrum must arise from four interacting adjacent hydrogens. This eliminates from consideration any structures such as the isomeric heterofulvalene below. In this compound all of the ‘Ph Ph "\H 1m /3 n 40 hydrogens do not interact with each other. Overall, the establishment of a four spin system of the type ABCD, in conjunction with the starting materials, supports the assignment of 1,2,3—triphenyl-4-azazulene (22). The formation of azazulene 22'can be rationalized by a logical mechanistic scheme (I): (a) initial formation of the pyrrolidine enamine of hydroxy—ketone 22; (b) acid catalyzed loss of the hydroxyl group, (c) isomerization to a more favorable intermediate, (d), Vilsmeier type addition to the cyclopentadiene ring, (e) ring expansion to tetrahydroazazulene isomers, and (f) air oxidation dur- ing chromatography. . Evidence in support of the oxidation during chroma- tography is that an initial yellow band began to separate on an alumina column. Before it could be completely eluted the band turned green. No yellow product was subsequently eluted. If a nylon column (26) was used it could be cut apart when sufficient separation had occurred. The yellow compound could then be extracted from the alumina before oxidation occurred. Due to the many different tetrahydro isomers possible, the nmr was useful only in establishing the number of aliphatic hydrogens relative to phenyl hydro— gens in this compound. This ratio, 8:15, as well as the elemental analysis, is in accord with a tetrahydroazazulene. The ease of oxidation reflects the stability of the hetero- aromatic system which is formed. When alkylation of azazulene 22 with methyl 41 Ph 11 Other / \ tetrahydro (f) Eh isomers > / Rh (3,9,) Mechanistic Scheme I. The formation of 1,2,3-triphenyl-4- azazulene (22). 42 lithium in ether at room temperature was attempted, no addition occurred. However, when benzene was used as the solvent and the temperature increased to 60°, the green solution turned yellow and a mixture of products resulted. Although these attempts failed to give useful alkylation products, it should be possible to find proper conditions under which controlled alkylation can occur. Likewise, an attempt to prepare the N-oxide of azazul- ene 22'by using hydrogen peroxide in acetonitrile was unsuc- cessful. Only unreacted starting material was isolated. However, when gfchloroperbenzoic acid was used as the oxi- izing agent the reaction mixture faded to light yellow within 15 minutes. The failure of the N-oxide to form is probably due to the steric effects of the nearby phenyl group on the nitrogen. The stronger reagent apparently reacts with a double bond in the aromatic system. The destroyed conjugation results in loss of the green color. This particular system is one of only a few azazulenes known in which the heteroatom is present in the seven mem- bered ring. The parent 5—azazulene has been prepared by Hafner (27). The reaction scheme involves a Vilsmeier addi- tion to 6-N,N—dimethylaminofuvene (22) followed by hydrol— ysis and condensation with ammonia. The parent 5-azazulene is reported to have its long wavelength absorption in the visible region at 652 mg. The only other example of a 4-azazulene known is 1-cyano-2,6-dibenz-4-azazulene (22) (28). Azazulene 22’ 43 111162 030 680 + H 0 O + Me2N=CHCH=CH -NMe2 2'95 NH3 37 resulted from the condensation of 1-cyano-2-indanonimine with phthaldehyde. This product was a red crystalline solid which absorbs in the visible region at 438 mg. The hypsochromic shift of azazulene 22'relative to azazulene 22'is due to the presence of the benzo groups which remove some of the electron density from the azazulene system. 44 Several attempts were made to synthesize 1,2,3-tri- phenyl-4—azazulene (Q2) by an alternate route. A procedure analogous to the scheme used in the preparation of 4-azazulene gg'was attempted. Triphenylphosphinimine (29) has been reacted with benzophenone to give benzophenonimine (30). It was anticipated that the reaction of triphenyl- phosphinimine with 2,3,4-triphenyl-2-cyclopenten-1—one (22) (31) would give the corresponding imine. Condensation of the resulting imine with diacetoxydihydrofuran (32) might give the desired azazulene. However, reaction of ketone 22’ with triphenylphosphinimine failed to give the imine of ke- tone 22” This may be due to enolization occurring rather than addition of the imine to the carbonyl. 0 Ph /]"K Ph Ph Ph . Eh .‘ — 9 0 211 km 45 0 D |+ Ph 3 P =N H -————> Ph 0 Ph 2h Ph -. 40 w Although the alternate synthesis was not achieved the proposed structure has been well substantiated by spectral data. The reaction of pyrrolidine with hydroxyéketone 22' represents a novel synthesis of a new theoretically inter- esting molecule. Structure Determination of (§)—3,3',4,4',5,5'-hexaphenyl- [bi-3-cyclopenten-1-y1idene]-2,2'-dione (fig) The minor, yellow product (21) which resulted from the 10% sulfuric acid-acetic acid dehydration of hydroxy-ketone gg'was identified as (§)-3,3’,4,4',5,5'-hexaphenyl[bi—3- cyclopenten-l-ylidene]-2,2'-dione (g1). The dimeric nature 46 of biscyclopentenylidene 21,13 apparent from the elemental analysis and the parent peak at mass 616 in the mass spec- trum. The nmr of compound gl’consists of an aromatic multiplet (153) at a 7.24-6.91 and a singlet at 6 5.19 (1 H). The aliphatic singlet is suggestiveof a symmetrical mole- cule. The presence of identical carbonyls in the two halves of the molecule is supported by the presence of a single carbonyl absorption in the infrared spectrum. The position of this absorption at 1674 cm-1 is best rational- ized by a carbonyl conjugated with both an a,B-endocyclic double bond and an a.B-exocyclic double bond. For compari- son, the absorption of the carbonyl of 2,3,4-triphenyl-2- cyclopentenone (42) appears at 1702 cm-1. 4O Biscyclopentenylidene §l’is also supported by its re- action with ruthenium tetroxide. Ruthenium tetroxide has been employed for the oxidative cleavage of double bonds to carbonyl compounds. For example, this reagent has been used for the oxidation of 3-alkylidene-2'—gresenes (4}) to grisene-3-ones (42) (33). When biscyclopentenylidene g; was 47 ° Ruo4 033 on on; 002m 3 41 42 W (W allowed to react with ruthenium tetroxide, an orange crystal- line product was obtained. A comparison of infrared Spectra showed that the degradation product was identical with a known compound, 3,4,5-triphenyl-3-cyclopenten-1,2-dione (4g) (31). In addition, a mixed melting point of an authen- tic sample of dione 4g and the degradation product was un- 0 Ph I D-o > 211 Ph 43 m depressed. Biscyclopentenylidene §l’+ Ru04 The identification of the degradation product and the spectral data, which establish the symmetrical nature of compound g}, are consistent with the proposed bicyclopenten— Ylidene g1; ' The formation of bicyclopentenylidene gl'can be ration- alized by acid catalyzed rearrangement of the normal cyclone 48 dimer. In strong acid a facile reaction which hydroxy- ketone gg’might undergo is dehydration to 2,3,4-triphenyl- cyclone (g2). As discussed previously, cyclone gg’would be expected to exist either as a dimer or an equilibrium mixture of monomer and dimer. Subsequent rearrangement of dimer ggb’under the strongly acidic conditions (Mechan- istic Scheme II) would result in the formation of bicyclo- pentenylidene g1, Since there is free rotation about the central bond in some of the intermediates, the sterically most favorable product should result. Since bicyclopentenylidene §l was formed as a minor product in this reaction, an alternate synthetic route was sought. It was anticipated that dione 42 would be a useful starting material for this attempt. If.dione gé could be condensed with 2,3,4-triphenyl-2-cyclopenten-1-one (42) the desired product would result. 1m Ph P 0 D + ‘ Ph Ph 0 1’“ 2h 40 43 31 Condensation of these two compounds in strong acid solutions (trifluoroacetic acid or sulfuric acid -- acetic acid) failed to give any condensation product. If such condensation were to occur it would require the attack of the enol of 0 €03 run 0 Ph ’ a 2n Ph Hkl’h .‘ H 9h 23’- 21, Mechanistic Scheme II. Formation of (g)-3,3',4,4',5,5'- hexaphenyl[bi-3-cyclopentenylidene]-2,2'-dione. 50 ketone gg'on the carbonyl of dione 43, However, in strong acid both ketone gg’and dione 43’should exist in the most stable protonated forms. Protonation of the ketone should give rise to the stabilized carbonium ion 42’H+ whereas the dione should exist as a similar carbonium ion (43’H+). Condensation of these carbonium ions is unlikely. OH .. 6 Ph H 2h 2h 22. 22, 2.1a“ 22H” Condensation also failed when base catalyzed reactions were attempted. Green solutions resulted when base was added to a solution of the two reactants. The color faded as the reaction progressed. Only ketone gg’was isolated from these reactions. No evidence for the presence of either dione gg’or a condensation product was apparent. Dione 43’ apparently forms unstable green salts which decompose in basic solutions (32), thus removing the dione from further reaction. Appropriate conditions to accomplish the desired condensation could not be found. Preparation of 3,3',4,4',5,5'-Hexaphenyl-2,2'-biscyclo- pentadienone (44) Removal of two hydrogens from bicyclopentenylidene 31’ would give rise to the theoretically interesting 51 3.3',4,4',5,5'-hexaphenyl-2,2'-biscyclopentadienone (44). This conversion was, in fact, accomplished in low yield by 0 Pk “o "‘ .. 0 2h . Ph 44 m reaction of bicyclopentenylidene gl'with NBS. When 34 was reacted with chloranil the yield of biscyclone 44 increased to about 60%. Mass spectral and elemental analysis data were consistent with the dehydrogenation reaction. Biscy- clone 44 exists as an almost black cryStalline solid. Solu- tions of 44 are red-purple and absorb at 342 and 546 mu in the ultraviolet and visible regions (benzene solvent). Biscyclone 44'represents the simplest biscyclone known. The simplest biscyclone previously reported is one in which the two cyclopentadienone rings are connected to each other through a phenyl ring (20). ‘Phg—g g—grh + PhCH2CCH2Ph Q \ rh 52 In addition, several biscyclones of the general type 42' are known (34). oo oo o PhC-CPh-Y-PhC-CPh + PhCHz-CCHzPh Ph 2h ,0 O 0 2h 4g, Biscyclones 4§'in benzene have ultraviolet and visible absorptions at about 370, 463 and 505 mu. In order to rationalize these absorptions of the biscyclones it is necessary to consider the absorptions of tetraarylcyclones. Benzene solutions of tetracyclone absorb at 342 and 512 mu. The absorption maximum at 342 mu has been attributed to an excited state of type II whereas, the absorption at 512 mu has been attributed to an excited state such as III. This has been determined by placing various substituents at the para positions of the phenyl rings on tetracyclone (20). The 342 mu absorption was affected most by R3 and R3 sub- stituents whereas the 512 mu absorption was dependent upon R1 and R4 substituents. The 341 mu absorption was shifted to shorter wavelengths by electron withdrawing substituents at R2 and R3 and to longer wavelengths by electron releasing 53 II III Figure 9. Excited states of tetracyclones. 54 substituents, but was relatively unchanged by varying R1 and R4. Electron withdrawing groups at R1 and R4 caused hypsochromic shifts in the 512 mu band and electron re- leasing groups caused bathochromic shifts. The 512 mu band was only slightly affected by R2 and R3 substituents. ‘ The bathochromic shift in the lower wavelength absorp— tion of biscyclones 4§’(Y = O, 3) relative to the 342 mu band of tetracyclone isconsistent with these substituent effects. Table II compares the values for the biscyclones with the corresponding tetraarylcyclones. The additional absorption at 463 mu in the biscyclones was attributed to simultaneous excitation of both cyclone moieties. This band decreased as atom Y became less effective as a trans- mitter of electronic effects between the two groups. A similar effect is apparent with biscyclone 22: The absorption at 342 mu is unchanged but the long wavelength maximum has shifted to 546 mu. This indicates a stabilizing 55 Table II. Comparison of absorptions of biscyclones and tetraarylcyclones (20). Y uv max (mu) S 375 463 502 O 368 463 507 9‘1 - o 358 0 509 Q! C C, s 516 56 effect such as IV in which the second cyclone ring is more effective than a phenyl group in stabilizing the positive charge in the excited state. IV Biscyclone 44’was found to undergo some interesting reactions. It was anticipated that biscyclone 44'would react with two moles of diphenylacetylene giving octaphenyl- quaterphenyl (42), the same product which results from re- action of two moles of tetracyclone with diphenyldiacetyl- ene (35). However, the reaction did not occur as anticipated. When biscyclone 44 and diphenyl acetylene were heated to- gether in refluxing benzophenone, a dark blue solution resulted. The same results were obtained when biscyclone 44’ was heated alone in refluxing benzophenone. Within 15 min- utes a blue solution resulted, from which a 70% yield of blue compound was obtained. It was found to be identical with blue compound 22'isolated from the dehydration of hydroxy-ketone 22, This facile rearrangement will be con- sidered further in the discussion of blue compound 22, 57 0. Ph Ph 2n + 2 PhCECPh 2h / Pb ' 2h 0 3 2h 1h Eh Ph Ph Ph Ph Ph 0 22, an ah i 2 D + PhCEC-C'ECPh 3“ Ph In refluxing benzene, biscyclone 44 readily reacted ‘with dimethyl acetylene dicarboxylate giving tetraaryl- cyclone 41, Tetraarylcyclone 41 is a red crystalline com- pound which absorbs at 480 mu in the visible region. The 0 Ph I'll», Ph M9020 0°21“ Ph [y * 1. . HBQ20—0:=O-002H0 elemental analysis and mass spectral data are consistent 58 with the addition of one mole of acetylenic ester to bis- cyclone 44 with subsequent loss of one mole of carbon mon- oxide. In addition to the ester absorption at 1735 cm-1 in the infrared another absorption was present at 1705 cm-1. The addition of a single mole of acetylenic ester is reason- able considering the steric effects involved in the addition of a second mole of dienophile. The most favorable arrange- ment of cyclone 4Z’is one in which the plane of the cyclone ring is perpendicular to the plane of the hexasubstituted phenyl ring. The resulting effect is that the diene system is completely blocked to further addition of a dienophile. The ester group interferes with addition On one face and a phenyl group interferes on the other face. The hyposochromic shift of the 480 mu band of 4Z’rela- tive to the 512 mu band of tetracyclone may be due to two factors. Perhaps the more important factor is that steric crowding forces the cyclone ring to approach a position in which its pi—orbitals are orthogonal to those of the phenyl group. Thus, very little, if any, interaction occurs. If, L" 59 however, any overlap is possible the presence of the elec- tron withdrawing carbcuuethoxy groups on the phenyl ring . will cause destabilization of the excited state. On the basis of these observations, cyclone 4Z'is a reasonable reaction product. Its stability to further at- tack of a dienophile is readily apparent from steric con— siderations. Identification of 3,3',4,4',5,5'-Hexaphenyl—3,4-dihydro- 2,2'-bicyc10pentadienone (48). In order to determine whether 2,3,4—triphenylcyclone (24) is an intermediate in the formation of 22'from the dehydration of hydroxy-ketone 22, the preparation of cyclone 24'was attempted. When dehydration was accomplished in benzene with pftoluene-sulfonic acid or in 2% sulfuric acid- acetic acid a red compound was obtained as the major pro- duct. There was no evidence for the presence of bicyclo- pentenylidene 24'and only a trace of blue compound 22 was obtained under these reaction conditions. In addition an unidentified colorless compound resulted. At first the major product 42'was impure and appeared to be pink. Its solutions, however, were intense red. This ‘was suggestive of the monomer-dimer equilibrium which exists in solutions of 2,3,5-triphenylcyclone. However, further purification gave a single, red crystalline product. Al- though it is tempting to suggest a 2,3,4-triphenylcyclone dimer or a monomer-dimer equilibrium, the spectral data lead to inconsistencies in interpretation. 60 On the basis of spectral and chemical data, 3,3',4,4', 5,5'-hexaphenyl-2,3-dihydro-2,2'-biscyclopentadienone (42) is a more reasonable structure. The mass spectral data and elemental analyses suggest a dimeric structure of molecular weight 616. The nmr in deuterochloroform consists of a 30 hydrogen multiplet at 6 7.93-6.4 and two doublets at 6 4.73 and 3.61 (J - 4 Hz) each representing one hydrogen. The nmr can be interpreted by either cyclone 42'or 2,3,4- triphenylcyclone dimer 24b, Dimer 242, however, is the less favorable cyclone.dimer. If cyclone 24'is present exclusively as a dimer, the more favorable dimer 242'would be more reasonable. The nmr is not consistent with dimer £22.. The infrared and ultraviolet spectra also favor cyclone 42'over 2,3,4-triphenylcyclone monomer 24 and dimers 242’ and 242, In the infrared only one carbonyl is present. This absorption is at 1700 cm-1. If either cyclone dimer 24g'or 242'is present a second carbonyl should absorb at 1 1775 cm- . In all dimers previously reported the bridge carbonyl appears at 1775 cm.1 (36). In fact, this carbonyl 61 Ph Ph Ph 8» Ph ° 3 0 21: Ph 0 u 211 Ph 44' 34b absorption has been used as a diagnostic test for bridge carbonyls. The absence of the second carbonyl indicates either a structure such as 42'in which the similar car- bonyls cannot be distinguished in the infrared, or an ex- clusively monomeric 2,3,4-triphenylcyclone. Cyclohexane solutions of 42'absorb in the visible at 480 mu with log 6 equal to 3.06 as compared with log 6 equal to 3.20 for the corresponding absorption of tetracyclone. The similar extinction coefficient is reasonable for a cyclone of molecular weight 616 such as cyclone 42, If in— terpreted in terms of 2,3,4-triphenylcyclone the extinction coefficient would require the presence of essentially all 62 monomer. As previously discussed, existence of 2,3,4- triphenylcyclone in exclusively the monomeric form is very unlikely. Thus, it is seen that in assuming 2,3,4-tripheny1cyclone as the structure of 42, inconsistencies arise in the inter- pretation of the spectral data. In particular, extreme interpretations result from the nmr and ir spectra which were taken in essentially the same solvent. Cyclone 42, on the other hand, is totally consistent with all of the spectral data. The presence of the cyclopentadienone portion of com- pound 42 was established by reaction of cyclone 42'with acetylenic ester. Addition of one mole of dienophile oc- curred with subsequent 1oss of one mole of carbon monoxide. The structure of the resulting cyclopentenone (42) is sup- ported by its nmr spectrum which consists of an aromatic multiplet at 6 7.3-6.5 (30 H) and two doublets centered at 6 4.88 and 4.14 (J = 4 Hz) each representing one hydrogen. Two different methyl groups also appear at 6 3.62 and 3.40. -1 In addition to the ester carbonyl at 1730 cm another car- 1 bonyl is present at 1710 cm_ in the infrared. The formation of cyclone 42'can be rationalized by a carbonium ion rearrangement of the normal cyclone dimer (Mechanistic Sequence II) and it is a reasonable intermed— iate in the formation of bicyclopentenylidene 24, Appar- ently the weaker acid solution was ineffective in isomer— izing cyclone 42 to bicyclopentenylidene 31, However, the intermediacy of cyclone 42' 63 + M902 CCECCO 2M6 Ph m 11 Ph 48 in the formation of 24 is supported by the conversion of cyclone 42’to the yellow bicyclopentenylidene 24’ in 10% sulfuric acid-acetic acid. There was no evidence for the formation of blue compound 22 in this transformation. In addition, although cyclone 42 can be purified by chroma- tography on silicic acid, it is completely isomerized to the yellow bicyclopentenylidene (24) when chromatographed on alumina. In fact, a greater yield of the yellow com- pound could be obtained by chromatography of cyclone 42’ than from dehydration of hydroxy-ketone 22, The same trans- formation also occurs in low yield when cyclone 42'is heated to 220°. Cyclone 42’appears to be a reasonable structure to account for these facile rearrangements to bicyclopen- tenylidene 24, In! 4. v 531.; .IIIC. (txrfl‘oflug . in (Na .. .. HE EV- 64 If a 2,3,4-triphenylcyclone dimer is the precursor to cyclone 42 and bicyclopentenylidene 24, its isolation from acid solutions is unlikely due to facile carbonium ion re— arrangements. A synthetic route for preparing 2,3,4-tri— phenylcyclone in a nonacidic solution was sought. This attempt involved the bromination of 2,3,4-triphenyl-2- cyclopenten-l-one (42). If the monobromide could be formed, the elimination of hydrogen bromide would give the desired cyclone. O o 0 Br 9 1m Ph Ph Ph 2h Ph 29. :22, However, when either NBS in carbon tetrachloride or bromine in acetic acid was used a dark brown-red crystalline product was obtained. This compound has a very sharp carbonyl ab- 1 in the infrared and has an absorption sorption at 1715 cm- at 490 mu in the visible. These results are suggestive of a cyclopentadienone. The molecular weight of 387 and ele- mental analyses give a molecular formula of C23H15Br0. A compound whose formation is readily rationalized is 2,3,4- triphenyl-5-bromocyclopentadienone (22). Either a dibromide is formed initially and spontaneously loses hydrogen bromide, or a monobromide forms with loss of hydrogen bromide to 65 give the desired cyclone which is then brominated. The value of log 6 equal to 2.87 for the 490 mu absorption suggests the presence of monomer. O NBS in CC14 Eh Br or > Bra in HUAC 1’11 Ph 50 If 2,3,4-triphenylcyclone is formed in any of the re- actions discussed, it is either very reactive or its dimer is very reactive under the reaction conditions. No further attempts at its synthesis were made. Identification of 6-Hydroxy-5,5,8,9,10-pentapheny1benzo- [ggjcyclopent[§]azulen—4-(5g)-one (22) The major product resulting from the 10% sulfuric acid- acetic acid dehydration of hydroxy—ketone 22'was blue com- pound 22, The structure previously reported for this com- pound was 2,3,4-triphenylcyclone (24) (16). On the basis of spectral and chemical data now available, a more reasonable structure appears to be 6-hydroxy-5,5,8,9,10-pentapheny1— benzo[ggjcyclopent[§]azulen-4-(5§)-one (22). The blue color of compound 22’is suggestive of an azulene chromophore. solutions of compound 22'in cyclohexane absorb at 620 mu in the visible region as compared to the parent azulene which 32 m absorbs at 670 mu in the visible region (37). The presence of the hydroxyl and carbonyl groups on hydroxy-azulene 22’are apparent from the infrared spectrum. The carbonyl absorption is present at 1681 cm-1 and the hydroxyl at 3150-3300 cm-l. The presence of a relatively acidic hydroxyl is also evident from the nmr spectrum. There is a singlet (1H) at 6 10.2 and a multiplet at 6 7.60- 6.92 (29H). The low field resonance can be washed out with D20. ' .This hydroxyl group was useful for converting hydroxy- azulene 22'to an acetate (24) or a trimethylsilyl derivative (22). The trimethylsilyl derivative showed a hypsochromic shift in the visible region to 532 mu relative to the 620 mu band of hydroxy-azulene 22, This substantial shift is in agreement with the proposed structure. Stabilization of hydroxy-azulene 22’by the free electrons on oxygen is de- creased when the enol is converted to the trimethylsilyl derivative. The trimethylsilyldErivative was used to obtain mass spectral data since the extremely low volatility of hydroxy- azulene 22'prevented its use. A computer analysis of the high 67 resolution mass spectrum shows that the molecular ion is present at mass 686 and corresponds to a molecular formula of C49H38028i. In addition to the fragments due to the trimethylsilyl group, the loss of carbon monoxide and a phenyl group are also apparent. A C13H10 fragment which may be attributed to loss of a diphenylmethyl from hydroxy- azulene 22'is also present. Several reactions were attempted in order to gain sup— port for the suggested structure. However, many failed to give identifiable products.. Reduction of the carbonyl of hydroxy—ketone 22 was attempted using sodium methoxyethoxy- aluminum hydride, but a complex mixture of products resulted. Hydroxy-azulene 22’might be expected to give such results. In addition to the carbonyl, the enol might be reduced by way of its keto form. Furthermore, the fulvenoid portion of compound 22’is also susceptible to hydride attack. Oxidation with ruthenium tetroxide also gave a complex mixture of products. This, too, is not surprising consider- ing the many different double bonds which are available for oxidative cleavage. Catalytic hydrogenation with a platin— um catalyst led to excessive hydrogen uptake. Partial reduction of the phenyl groups could account for such re— sults. On the other hand, hydrogenation employing a paladium on carbon catalyst, led to negligible hydrogen consumption. The isolation of oxidative degradation products was then attempted. The use of chromic acid as the oxidant gave benzoic acid as the only identifiable product. Several 68 unidentified products appeared to be only partially oxi- dized fragments. When a two phase oxidation was carried out using basic potassium permanganate as the oxidant, a second product was obtained. This product proved to be benzophenone by comparison of infrared Spectra. In addi- tion, the melting points and the infrared spectra of an authentic sample of benzophenone 2,4-DNP and the 2,4-DNP of the degradation product were identical. Benzophenone would be expected to result frOm oxida— tive cleavage of the diphenylmethyl fragment on hydroxy- azulene 22, However, this evidence is ambiguous since benzophenone might also result from the oxidative cleavage of a group such as 24'if such a group is present in the molecule. Oxidation of fragment 24 should give benzil. In basic solution benzil can rearrange to benzilic acid which would give benzophenone on oxidation. Ph “a. o 0 OH 0 m1 I ll 9 -——9-> Base > Ph-C-COaH —> Ph-C-Ph I Ph Ph Ph Ph Ph 54 However, the presence of the diphenylmethyl group on hydroxy-azulene 22 is also suggested from the fragmentations in the mass spectrum of trimethylsilyl derivative 22, 69 Although a conclusive structure determination cannot be made from the available evidence, hydroxy-azulene 22’ is fully consistent with all of the data presented. The formation of hydroxy-azulene 22'can be rationalized by mechanistic scheme III. In acidic solution, hydroxy— ketone 22'may undergo the following conversions to blue com— pound 22; (a) a reversible aldol reaction, (b) condensa- tion of the retro-aldol product with either hydroxy-ketone 22'or 2,3,4-triphenylcyclone 24, (c) an acid catalyzed con- densation of the benzoyl group with a phenyl ring in the formation of a five membered ring, (d) protonation of the remaining carbonyl with subsequent phenyl migration, (e) conversion of the enol to the keto form, (f) protonation of the ketone,followed by acid catalyzed addition of the fulvenoid system to the phenyl ring, and finally (9) oxida- tion to azulene 22, No reaction path for the isomerization of biscyclopenta- dienone 44’to azulene 22'is readily apparent. If the blue compound is indeed azulene 22, this conversion represents an unusual rearrangement. Since no blue compound was found in the 2% sulfuric acid-acetic acid reaction or in the conversion of cyclone 42’ to bicyclopentenylidene 22, the stronger 10% sulfuric acid- acetic acid solution must provide conditions for an additional reaction path. In both cases the reverse aldol probably could occur. In the weaker acid ring closure to starting material may be favored. On the other hand, use of stronger Mechanistic Scheme III. The formation of 6—hydroxy-5,5,8,9,10- pentaphenylbenzo [ 951] cyclopent [£1 azulen -4- ( 5g; ) -one . 71 acid may provide an additional path for the condensation of the retro-aldol product with either hydroxy-ketone 22’or ‘with 2,3,4-triphenylcyclone (24). Due to the uncertainty of the structure of hydroxy- azulene 22'the suggested mechanism for its formation is at best speculative. The identification of the other novel products are supported by chemical and spectral data but the mechanisms for their formation are also speculative. However, these routes do show that reasonable reaction paths are available which can account for the novel reaction products of hydroxyeketone 22, 71a After discussing the structure of hydroxy-azulene 22' with Dr. D. G. Farnum, another possible structure, 3- hydroxy-1,1,4,5,6-pentaphenylbenzo[2jcyclopent[cd]azulen-2- (1§)-one (22;) was apparent. This structure is consistent with the spectral data discussed previously for hydroxy-azulene 22, In addition, the carbonyl absorptions in the infrared spectra of the acetate and trimethylsilyl derivatives of 22 appear at 1701 cm.1 as compared to the 1681 cm-1 carbonyl absorption of the parent hydroxy-azulene 22, This shift could be rationalized by the loss of the chelate type structure in conversion of the hydroxy group to the derivatives. Two different approaches may be useful in attempting to distinguish between structures 22'and 229. If the chelate structure is present it may be possible to con- vert it to a copper chelate by use of copper acetate. 71b A second approach involves the reaction of 22? with phenylhydrazine. This may result in condensation of the phenylhydrazine with both carbonyl positions in the form— ation of a lfigepyrazole. If either of these two reactions is successful 22; would be a more satisfactory structure.) However, failure of either of these reactions could result from steric factors. Thus, failure of either reaction would not elim- inate structure 22? from consideration. The formation of 22? from biscyclopentadienone 4Z'can be rationalized by the sequence of the following page. 71c Rearrangement of biscyclone 4Z'to azulene 22p. EXPERIMENTAL 1. General Procedures These are the same as those used in Part I. 2. 1,2,3-Triphenyl-4-azazulene (22) A mixture of 1.7 9 (0.0052 mole) of 4-hydroxy-2,3,4- triphenylcyclopent-2-en-1-one, 0.7 g (0.01 mole) of pyr- rolidine and 0.1 g of toluenesulfonic acid in 10 ml of toluene was refluxed overnight with continuous removal of water by means of a Dean-Stark trap. The solvent was then removed under reduced pressure to give a tarry black resi- due which was chromatographed on Alcoa F-20 alumina with benzene. Initially a yellow band began to separate but before it could be eluted it had turned green. Upon removal of the solvent from this green fraction a dark green solid remained which was crystallized from heptane giving 80 mg of azazulene 22; mp 224-2260; uv max (CH3CN) 285 (log 6 4.51), 325 (log 6 4.48), 380 (log 6 3.84), 610 (188 e 2.35), 654 (log e 2.43e, and 710 mu (109 e 2.35); ir (ch13) 1535, 1570, 1590 (arom), and (hexachlorobutadiene) 3020 cm-1 (arom C-H); nmr (CDc13)‘6 8.33-8.08 (m, 1H). 8.07-7.82 (m, 1H), and 7.32-6.82 (m, 17H); mass spectrum (70 eV) m/e 357 (Parent) ' 72 73 Anal.calcd for c27H19N: c, 90.80; H, 5.33. Found: c, 90.70; H, 5.50. 3. 1,2I3-Triphenyl-4-azatetrahydroazulene A mixture of tetrahydroazulenes could be obtained if the tarry mixture from the previous mixture was chromato— graphed on Baker neutral alumina packed in a nylon column (26). In packing the column the best results were obtained when the nylon column was supported in a glass sleeve and packed wet. When sufficient separation had occurred the column was cut apart and extracted with ether. The ether solution was concentrated to about 10 ml. After the solu- tion was cooled in a freezer, yellow crystals appeared. The solution upon standing at room temperature decomposed but the crystals when isolated were stable. In this manner from 30 g of hydroxy-ketone 22, 100 mg of the tetrahydro- azulene was obtained. This low yield was due to the many difficulties in isolation: nmr (CDC13) was not straight- forward due to the many isomers which are poSsible but the ratio of phenyl protons to all others is 15:8; mp — not sharp. ABEL: Calcd for C27H33N: C, 89.70; H, 6.38. Found: C, 89.60; H, 6.64. 4. (E)—3, 3' ,4 4', 5, 5'-hexaphenyl[bi—3-cyclopenten—1—ylidene]— 2 ,2'-dione (31) a) From hydroxy-ketone 22’ ‘Compound 24 was prepared by the method of Pauson (16); mp softens at 256—2580, then solidifies and does not melt 74 up to 310°; uv max (cyclohexane) 267 (log 6 4.46), 342 (log e 4.13), and 460 mu (log 8 2.60); ir (CHc13) 1674 cm"1 (Czo); nmr (CDCls) 6 7. 24- 6 .91 (m, 15) and 5. 69 (s, 1), mass spec- trum (70 eV) m/e 616 (Parent). 5334, Calcd for C43H3302: C, 89.60; H, 5.30. Found: ‘ C, 89.60; H, 5.34. b) Thermal decomposition of 3,3',4,4',5,5'-hexaphenyl- 3,4-dihydro-2,2'-biscyclone (42). When a 500 mg sample of cyclone 42 was heated to 220° for 5 minutes gas evolution occurred from the melt and then the melt resolidified. After cooling the sample to room temperature 10 ml of benzene was added. The yellow solid which separated was filtered and washed with benzene to give 300 mg of a highly insoluble yellow solid which did not melt up to 310°. From the benzene filtrate was isolated 60 mg of yellow compound having an ir identical with that of compound 24, c) 10% Sulfuric acid-acetic acid reaction of cyclone 42' A solution of 100 mg (0.33 mmol) of cyclone 42'in 5 ml of 10% sulfuric acid in acetic acid was refluxed for 30 min and then poured into water and extracted with benzene. The ben— zene was washed with sodium carbonate solution and dried with MgSO4. The yellowish residue was chromatographed on alumina ‘with benzene to give 80 mg (80%) of yellow compound 24, The ir was identical with that of compound 24, No blue product was found. 75 d) From chromatography of cyclone 42‘ A sample of 0.5 g of cyclone 42'was chromatographed on an Alcoa F-20 alumina column with benzene. The only fraction eluted gave 0.45 g (85%) of bicyclopentenylidene 31. ' ’ 5. Ruthenium Tetroxide Oxidation of Bicyclopentenylidene (31) fees... A solution of 300 mg (0.5 mole) of yellow compound in 20 m1 of chloroform (washed well with water to remove traces of alcohol) was added dropwise to a mixture of 10 m1 of chloroform, 30 m1 of water, 10 gm of ruthenium dioxide, and 0.64 g of sodium periodate. The resulting mixture was stirred overnight at room temperature.' The resulting black ruthenipm dioxide was filtered and washed with chloroform. The water and chloroform layers of the filtrate were separ- ated and the combined chloroform layers dried (MgSO4). The solvent was removed under reduced pressure and the residue chromatographed on Mallinckrodt silicic acid with benzene- methylene chloride (10:12). From the first yellow fraction 140 mg (47%) of starting material was recovered. From the second fraction 80 mg (49% based on unrecovered starting material) of orange needles was obtained. The ir Spectrum was identical with that of 3,4,5-triphonylcyclopent—3-ene— 1,2—dione (42) and a mixed melting point was undepressed; mp 160-162°: ir (cuc15) 1705 (c=o) and 1760 cm"1 (0:0); mass spectrum (70 eV) m/e 324 (Parent). 76 6. 3,3',4,4',5,5'—Hexaphenyl-2,2'-biscyclopentadienone (47) a) From reaction of bicyclopentenylidene 24'with chloranil. A mixture of 100 mg (0.17 mmol) of yellow compound 24 and 200 mg (0.1 mole) of chloranil in 10 ml of xylene was refluxed for 4 hrs. The solvent was removed under reduced pressure and chloroform added to the residue. The solution was filtered from insoluble material and the solvent re- moved from the purple solution. The residue was chromato- graphed on Baker neutral alumina with benzene to give 60 mg of black solid (60%). An ir and melting point showed this to be identical with the product from the NBS reaction. b) From reaction of bicyclopentenylidene 24'with NBS. A mixture of 200 mg (0.34 mmol) of yellow compound 24' and 400 mg (0.3 mole) of NBS was refluxed in carbon tetra- chloride for 20 hrs. The solution was then cooled, washed with cold water, and dried (M9804). The solvent was removed and the residue was chromatographed on Baker neutral alumina with benzene. The first fraction gave 130 mg of yellow starting material. A purple-black band was then eluted from which was isolated 20 mg of black solid which was crystal- lized from chloroform-isopropanol: mp 288-289°; uv max (ben- zene) 332 (log 4.23), and 546 mu (log e 3.46); ir (CHClg) 1705 cm-1 (c=0); mass spectrum (70 eV) m/e 614 (Parent). 5234, Calcd for C46H3002: ‘C, 90.00; H, 4.90. Found: C, 90.15: H, 5.09. 77 7. Reaction of BiscyclOpentadienone 4Z'with Acetylenic I Ester A mixture of 50 mg (0.08 mmol) of bistriphenylcyclo- pentadienone and 200 mg of dimethyl acetylenedicarboxylate in 5 ml of benzene was refluxed for 24 hrs. The solvent was removed from the bright red solution to give a red oil which was chromatographed on Mallinckrodt silicic acid with benzene—methylene chloride (10:12). A red oil was obtained which crystallized after addition of methanol giving 47.5 mg (82.5%) of red crystals. Recrystallization from methanol gave bright red crystals: mp 174—175°: uv max (cyclohexane) 248 (log 8 4.63), 325 (log 6 393), and 480 mu (log 6 3.01); ir (CHC13) 1710 (c=o), 1730 cm.1 (ester c=o); mass Spectrum (70 eV) mlg_(rel intenS) 728 (8) (Parent), 697 (100). - 5&34, Calcd for C51H3505: C, 84.10; H, 4.95. Found: C, 83.91; H, 5.06. 8. Conversion of Biscyclopentadienone 4Z'to Blue Compound 22’ A mixture of 100 mg (0.33 mmol) of biscyclone 41 in 500 mg of benzophenone was heated to gentle reflux. Within 15 min an intense blue solution resulted. The reaction flask was allowed to cool and 10 ml of methanol was added. Upon standing overnight the solution deposited blue crystals. The solution was filtered and 70 mg (70%) of product ident- ical with blue compound 22 (ir and melting point) was obtained. 78 9. 3,3',4,4',5,5'-Hexaphegyl-2,3—dihydro-2,2'-biscyclo- pentadienone (48) Am. a) From p-toluenesulfonic acid dehydration of hydroxy- ketone 224’ A mixture of 2 9 (0.0015 mole) of 4-hydroxy—2,3,4-tri- phenylcyclopent-Z-en-l-one and 1 g of pftoluene sulfonic acid in 25 ml of benzene was refluxed 1 hr. The solution was washed well with water and dried (M9804). The benzene was removed under reduced pressure and the residue chromato- graphed on Mallinckrodt silicic acid with chloroform. Only a trace of blue compound was eluted followed by 180 mg of an unidentified yellow compound, 140 mg of another unidenti- fied yellow compound, and 350 mg of a colorless compound which analyzed for a dimer: mp 264°. 2&34, Calcd for C46H3202: C, 89.60: H, 5.20. Found: C, 89.55: H, 5.41. Finally 320 mg of red compound was obtained. Recrystal- lization from chloroform-methanol gave a red crystalline product: mp 264-265°: uv max (cyclohexane) 245, 270, 332, 460, and 490 mu: ir (cac13) 1702cm"1 (c=o); nmr (CDCla) 5 6.4-7.41 (m, 303 ), 4.72 (d, 1, g_= 4 Hz), and 3.61 (d, 1, g_= 4 Hz); mass spectrum (70 ev) £12 (rel intensity) 616 (100), 598 (9), 588 (13), 570 (4), 539 (6), 525 (4), 521 (3), 511 (23), 510 (3), 498 (8), 493 (4), 433 (4), 308 (46). 280 (12), 267 (12) and 178 (19). ' ’ .2234, Calcd for c45H3202: c, 89.60; H, 5.20. Found: C, 89.47; H, 5.26. 79 b) From 2% sulfuric acid-acetic acid dehydration of hydroxy-ketone 22, To a solution of 10 g (0.03 mole) of 4-hydroxy-2,3,4- triphenylcyclopent-2-en-1-one in 50 ml of acetic acid was added 1 ml of concentrated sulfuric acid and the mixture heated to boiling for 5 minutes. The resulting solution was poured into water. The solid which separated was fil- tered and dried (M9304). When ether was added 2 g of unidentified yellow ccmpound separated. The remaining ether solution was evaporated and the residue chromatographed on silicic acid. The first fraction (carbon tetrachloride— benzene, 10:2) was blue and gave 0.3 g of product identi- cal with compound 22, The second fraction (methylene chlor- ide-benzene, 10:12) gave 3.4 g (35%) of cyclone 42, The final fraction consisted of an additional 1.4 g of unidenti- fied yellow product. The cyclone was identical with the product from the toluene sulfonic acid dehydration. 10. Reaction of Cyclone 42‘with Acetylenic Ester A mixture of 200 mg (0.33 mmol) of cyclone dimer and 300 mg of dimethyl acetylenedicarboxylate in 10 m1 of benzene was refluxed overnight. The resulting solution was Slightly yellow. The solvent was removed under reduced pressure and the residue chromatographed on silicic acid ‘with benzene-methylene chloride (10:12). An oil was ob- tained which crystallized upon addition of methanol. 80 Recrystallization from methanol gave 150 mg (62%) of color- less crystals: mp 264-265°: ir (CHCla) 1735 (eSter c=o) and 1705 cm“1 (c=o); nmr (cools) 6 7.32-6.23 (m, 30), 4.86 (d, 1, g_= 4 Hz), 4.15 (d, 1, g_= 4 Hz), 3.62 (s, 3), and 3.40 (s, 3): mass spectrum (70 eV) EZ§_730 (Parent). 523;. Calcd for €51H3305= c, 83.70; H, 5.2. Found: C, 83.45; H; 5.20. 11. 2,3,4-Triphenyl-5-bromocyclopentadienone (50) a) Bromination of 2,3,4-triphenylcyclopent-2-en-1-one. To a stirred suspension of 9 g (0.03 mole) of 2,3,4- triphenylcyclopent-2—en-1-one in 150 ml of acetic acid was added 6 g (0.035 mole) of bromine in 15 ml of acetic acid. The mixture was heated to dissolve the starting material and then allowed to stand at room temperature overnight. The solvent was removed under reduced pressure leaving an oily red residue which was chromatographed on silicic acid with carbon tetrachloride-benzene (10:2). The product was eluted first as a dark red band. Upon removal of the sol- vent the remaining red solid was crystallized from ethanol giving 2.8 g (41.5%) of dark red crystals: mp 188-1890: uv max (cyclohexane) 245 (log 6 4.19), 270 (log 6 4.24), 332 (log 6 3.93), 460 (log 6 2 90), and 490 66 (log 6 2.87); ir (CHC13) 1715 cm.1 (c=o); mass spectrum (70 eV) £19 (rel intensity) 388 (100), 386 (100), 360 (36), 358 (36), 307 (45.5), 279 (100), 278 (73), 277 (54.5), 276 (54.5), 202 (32), 178 (45.5), and 139 (50). 81 Anal. Calcd for C23H153r0: C, 71.40: H, 3.88. FOund: C, 71.36: H, 4.06. b) N-bromosuccinimide bromination of 2,3,4—triphenyl— cyclopent-Z-en-l-one. A mixture of 0.9 g (.003 mole) of 2,3,4-triphenyl- cyclopent-2-en-1-one and 0.5 9 (0.0028 mole) of N-bromo- succinimide in 25 ml of carbon tetrachloride was refluxed for 1 hour during which time the solution became dark red. The mixture was cooled in an ice bath and then filtered. The filtrate was washed with cold water and then dilute sodium thiosulfate and dried (MgSO4). Upon removal of the solvent a red oily residue remained. The remaining work—up ‘was the same as in part a giving 120 mg (11%) of product having the same mp and ir as in part a. 12. 6-Hydroxy-5,5',8,9,10epentaphenylbenzongjcyclopentlf|- azulen-4-(5§)-one (22) This was prepared by the method of Pauson (16): mp 294-295°; uv max (cyclohexane) 265 (log 6 4.20), 305 (log 6 4.25), 345 (log 6 3.34), and 620 mu (log 6 3.34); ir (cuc13) 3150-3300 (on) and 1681 cm-1 (0:0); 6m: (CDcls) 6 10.2 (s, 1), and 7.60-6.92 (m, 29). Agg4, Calcd for C46H3002: C, 90.00: H, 4.90. Found: C, 90.38: H, 5.23. 82 13. Trimethylsilyl Derivative of 22‘ A mixture of 100 mg of compound 22'in 8 ml of aceto- nitrile (freshly distilled from P205) and 1 ml of bistri- methylsilylacetamide was refluxed under nitrogen for 30 minutes during which time the color changed from blue to purple. Upon standing at room temperature the trimethyl- silyl derivative crystallized: mp 257-258°; uv max (cyclo- hexane) 255 (log 6 4.57), 295 (log 6 4.73), and 532 mu (log 6 3.72): ir (CHCla) 1701 cm.1 (c=o); mass spectrum (70 eV) gig 686 (Parent). ‘ 2 5324, Calcd for C49H388102: C, 85.50: H, 5.54. Found: C, 85.41; H, 5.57. 14. Acetate of 22' A mixture of 100 mg of compound 22, 5 ml acetic anhy- dride, and 2 ml pyridine was refluxed under nitrogen for 30 minutes during which time the color changed from blue to purple. When poured into water the acetate crystallized and was filtered. Recrystallization from hexane gave 60 mg (62%) of acetate: mp 300-301°; ir (CHCla) 1701 (c=O) and 1760 cm-1 (ester c=o): UVmax (cyclohexane) 525 and 290 mu. Anal. C31Cd for C43H3203: C; 87.70; H: 4.87. Found: C, 87.59: H, 5.02. 15. Potassium Permanganate Oxidation of 22' A mixture of 0.6 g (0.001 mole) of blue compound 22’in 50 ml of heptane and 20 ml of water in a three—necked 83 (stirrer, thermometer, dropping funnel) flask was heated to 100°. To this was added dropwise a hot solution of 4 g of KMnO4 and 3 g of sodium carbonate in water. When the per- manganate was consumed another 2—g portion (in water) was added. This process was repeated until a total of 8 g was added. The heptane layer was drawn off and the sol- vent removed. The residue was chromatographed on silicic acid with CCl4-benzene. The third and fourth fractions collected contained a total of 75 mg of an oil. The ir was identical with that of benzophenone: mp 2,4-DNP, 236-2380: benzophenone 2,4—DNP 237-238°. The ir spectra of the 2,4- DNP'S were identical. The aqueous layer from the oxidation was saturated with sulfur dioxide until all of the black umoz was gone. Acid was added to dissolve most of the salts and the acidic solu— tion was extracted for three days in a continuous extractor with ether. The ether was dried and the solvent removed. The residue was chromatographed on silicic acid with CCl4- benzene. All factions eluted with CCl4-benzene or with chloroform were found to contain benzoic acid (350 mg unpuri- fied). The remaining fraction gave oils which appeared to be only partially oxidized fragments. BIBLIOGRAPHY 7. 8. 9. 10. 11. 12. 13. 14. 15. BIBLIOGRAPHY R. R. Neuman and R. E. Johnson, 4nt. Sci. and Tech., 68 (May, 1964). Y. Okamoto and W. Brenner, "Organic Semiconductors", Reinhold Co., New York, 1964. W, A. Little in "Electrical Conduction Properties of Polymers", A. Rembaum and R. F. Landel, Ed., Inter- science Publ, New York, 1967. W. A. Little, "Superconductivity at Room Temperature", Scientific American, 212, 21 (Feb., 1965). w. A. Little, Phys. Revs., 134, 1416 (1964). A. Szent-Gyfirgyi, Science, 22, 609 (1941). A. Szent—Gyorgyi, Nature, 161, 875 (1941). M. S. Karasch and B. S. Joshi, J. Org. Chem., 22, 1435 (1957). G. M. Coppinger, J. Am. Chem. Soc., 12, 501 (1957). D. Lauerer, M. Coenen, M. Pestemer and G. Scheibe, Z. Physik Chem., 42, 236 (1957). G. A. Nikiforov and K. M. Dyrumaev, Chem. Abstr., 24, 7094g (1964). J. K. Klebe, H. Finkbeiner and D. M. White, J. Am, Chem. Soc., 22, 3390 (1966). H. Oeliger, H. Kobbe, F. Muller and K. Eiter, Chem. Ber., 22, 2012 (1966). J. D. Roberts, "Nuclear Magnetic Resonance-eApplications to Organic Chemistry", McGraw-Hill, New York, 1959, p 63. W, Dilthey and G. Hurtig, Chem. Ber., 22, 2004 (1934). 84 85 16. P. L. Pauson and B. J. Williams, J. Chem. Soc., 4162 17. R. D. Brown, J. chem. Soc., 2670 (1951). 18. C. H. DePuy, M. Isaks, K. L. Eilers and G. F. Morris, J. Org. Chem., 222 3503 (1964). 19. L. F. Fieser and M. J. Haddadin, J. Am. Chem. Soc., 86, 2081 (1964) . w 20. M. A. Ogliaruso, M. G. Romanelli and E. I. Becker, Chem. Revs., 22, 261 (1965). 21. W. Dilthey, W. Schommer, J. Prakt. Chem., 136, 293 (1933). 22. R. L. Foltz, Battelle Memorial Institute, personal communication. 23. F. W, McLafferty, "Interpretation of Mass Spectra", W, A. Benjamin, New York, 1967. 24. D. Meuche, B. B. Mollay, D. H. Reid and E. Heilbronner, Helv. Chim. Acta., 42, 2483 (1963). 25. A. A. Bothner-By, CarnegieeMellon University, personal communication. 26. B. Loev and M. M. Goodman, Chem. and Ind., 2026 (1967). 27. K. Hafner, et al., Angew. Chem. Int. Ed., 2, 123 (1963). 28. W. Triebs and W. Schroth, Angew. Chem., Z4, 71 (1959). 29. R. Appel and E. Guth, Z. Naturforsch., 15B, 57 (1960). 30. R. Appel and A. Hauss, Chem. Ber., 22, 405 (1960). 31. C. F. Koelsch and T. A. Geissman, J. Org. Chem., 2, 480 (1938). 32. N. Clausen-Kaas, Acta. Chem. Scand., 4, 379 (1947). 33. F. M. Dean and J. c. Knight, J. Chem. Soc., 4745 (1962). 34. M, A. Ogliaruso, L. A. Ahadoff and E. I. Becker,,l; Org. Chem., 22, 2725 (1963). 35. M, A. Ogliaruso and E. I. Becker, J. Org. Chem., 22, 3354 (1965). 36. c. F. H. Allen, Chem. Revs., 22, 653 (1962). 37. D. Ginsburg, "Non-benzenoid Aromatic Compounds", Interscience, New York, 1959. APPENDIX 86 .Anaomo Gav ma ossomaoo mxoguo mo Eduuommm pmumuusH .oH ousmflm o as 39 08. 3! coo. 89 88 on: 88 080.. i as an i m V o. . 9m. _ M n V m on 003 8 Fri/Anon l; 8. , 8. 0.2 of 0.: Cd. mZOmUE‘ 0.x ex 04.. On On mam 87 . A938 5.; 2m)”. H0930 can“: no gunman consume; .3 0.39.6 o 8o 89 82 8: 82 com. 88 8mm 8.” . m . , j .. “-9 I W cm 1 48 -4 111194. 41 14 - 1-1 A 3.8an _ new :- 1.1m.| 110-). .IIJi . m - 1.. .1111.-. u -. , 1; 1.83 . ow . - -1: .-l11:.-1 « 1 -11-!1 8. od. 02 o: no. 3208.: i an o. nzomoi 0.4. . nu ON 0v oo .AnHUmU :3 01m kumo 9580 no Eauowmm ponmuqu ON 0 V (“H SDNVllIWSNVUL O O 8.11.11: .jL1illJ8. 0 m 89 a 06» 000. -1111mmw. m i as _ _ . .._ m ~ Q .) ; {2.11 1...... , .Anaomo :3 .mm osocosmoucon mo Sauoam USHMHMSH 8v. 8.: 89 - 68m 8.: -. 018m .mn Gasman oonm ooov '- .III.|||I .lllu'll'll Cl" ,9 O O (%1 aanuIWSszu O on oo— .¢ . ,. . J.» o 3 . . .11 .(n . 1. a. ( .4 .1 |. v n& 90 .Anaomo :3 mm mamaauanm mo gunman UoumquH .wn ouamfih 3 8m 08» OON— 00: Muco— L ‘ :02”. w OOON OOWN 000m Down 000W 8 ‘ ‘ : 2 .Mr ‘ . I. . a ,V ~ V ov fl . , 9W. _ _ m _ V w w. 00 ,4 oo 3 A m ‘‘‘‘‘ w ‘ l . ‘ fl . . _ 0” +14 L. .+ , 1°C m w I, x. 4 g .7 ,. < O. O N— O 2 O O_ wZO¢U_<< O m 0 K 0.0 Qm. mZOmQZ Oi wfi O m W N .Anaumo c3 0mm oawvflahqmunomoaoxngfi mo guuoomm Umumnmcn .mu mun—mam 91 a Sm 82 8a. 8: 82 8m: 82 82 08... 83 88 < H _ u v _ _ c ,_ ltlr‘\b‘ » . . ID.» 1 b ‘7‘ ilk/ il‘l o 1 .L:1.l:iTiI-:l,3o~ E u an e u a .‘q... C1113! m V e V , .1 ‘ : 13%. m V w ‘ 003 W , om Ill ‘l--| ’| 1.. .I oo— (3 vi mZOmU_<< o‘v mam o.m 92 .Aaaomo a3 NW mcoaomouwn mo Esauommm umnmumcH .ofi 0.25am o. 08 89 3s. 8! 03. 08. 83 83 88 ) , 20,4 ON T on m V 2 2w .w. n V N o? D 8 *4 J 0.0— . . — o . 8—r...l.+i ...lui.-...- ; . - p . ea. c... 0.9 306.2 cm (7 O P OK 0. 93 .Anaomu 5.; NW anoaomumwn no 90969.... kumo- Dado-"mum"; mo guuoomm Uwumuqu .bH munmfih 3 00m 08— OON- 8: 000— 009 OCON 8mm 08H 82” 08W . , _ k . H m . u , , . u M . . c -----62-; m U M 1 , , p a _ m . _ - - ._ -11 -- F 1.11 T1.) I > . ,. 1 . . - a - M . . m _ _ 1 n, _, _ n m _ . - W 91m- '11 - + - - m M m m OVN x v . u . m 00 1 1- j k V 00.” 8 \I, . dm r 1% - . ( . _ u 8VWI_:H1..W H on m n . - H +1-17;- , . A . 8~.l.ul<+ .10 11A .-. w— ! v o, > . . u‘ll.11|||'l ~ 1 - . . - I . - I [‘1‘1 1 8— 0.0. ON— o.: 0.9 mZOaU=2 O-m OK 06 Oh m20¢U=2 0.1 MM QM , nu 94 00m— mzouui Qm 00v— QR .AaHUH-U :3 MW mqoaoho mo guuummm umumumcH .wH ousmwm 000— 90.2 000m comm 00mm 000v o om mzoaui 3. 000m 0 V (N) aanmwsuvm O O -om 95 .AnHUmU 93 NW- ocoaokfiv mo uusuvm Hmumm owcwamuwom mo Supp-5mm UmumumcH 00m 000— 003 003 002 009 Doom Gown 000m mflh ah Gm \ m n» ...-O 9 Ha osmoo owes. o. w :r mzoaui o.» ox 0.0 o.“ mzoxui 3. 3 3.. .mfi wuamfim 1%) aDNvmws‘Nvm 96 .Anaomo cavd‘m oGOHUhooEounamcmnmflHu mo Eduuommm UmumumcH .oN munmflm o o8 89 8w. 8! 8: 08. 38 82 88 8mm 8&9. _ m _ H 03 . . .. - - . . .ON ~ a _ .8 :l E H _ .3 \ / rm I. a __ v on. o OvN - S _ M m v . N 8. H83 __ w.-.“ 8 m8 . m co.” So. 0.2 of 0.: 9.9 501.54. 8 ON 0.» Rn u..,...nha..,..__: 0+ hd p m Om 97 1%) SDNVHIWSNVM . A «Homo :3 Mm mamasum mo Esuuoomm comma-nag” . an mug-Tm a . 8m 82 82. 8: 82 28. 88 82 OS,” 82 89‘ c H M . m M , . . . . H c . 1.. 111% V H1 _ _ . . 1. 11. - - V 1 1 1 1 y 1 11 111111 r .r _ . . . . m . on H :7 1. . 8 u w _. h . . .r m U m 1 . . 11. _ 1 * 0V1 -_ '1‘. 11 1 1.”.1 W 1 0‘ 1. 1 . w 1 8 ._ - .8 81; 1.. . .- 8 11 _ , . .. 87111.1... . 1; 11.11.... -. . 1 1 - ..... - 11 8. 0.0— 0.2 O.: 06— mZOmUmE Ow OK 0.0 Om mZOmBS‘ O.v mtm O.m ON ' I u 7. u I. {Ln193 .' fvvr‘ 7* 7' ""V v *VVVVVV~r""'VVf'v-vvvv"* vvv fi—V"'fi vvvvvvvvvvvvvvvvvv V—V 'Vfi I. a - ’ . 02’ I. II I' j '0 cn—c( 3’ : man 0' k W A4 Al 1. l - 1* IA; 1 - IA 1 Al 1- gel; 1‘s--JAAL-AA- ILA LAL 1 - Nmr spectrum of nitro ester 1§’(in CDC13). I! u u I. u +nutu ——'-T iv v 1 -vv' v v‘vva 'fvv r -V‘ iv vv ' Tffifv' wfivT vvv‘ v [Via vvvvv r1*f7*vv v v rvvvvwr ' --.vvv v'v'vrvvvv fry—rwr ivv V‘vvyv vrvwvwvwvvvv.,vv—r II H) II no ‘ - ‘ 003 7. ll ii I! 4 l A. - , -.A---l 81‘4- l 1 l -1 L -l -L- l -.--A-14- -.-- -1; Figure 23. Nmr spectrum of amino ester gQ'(in CDC13). 99 +~fll vwfivv v v 'wvv1vv wvvvv'v'v'vvrvaVfif v vi I. I - vvvv'v fl v u V--Tvvv-'--v- --v , vf'fi v ‘rvwv vvvvvvvvv u 7 m-rvvvvtvvww --fv'- 1 -- .Ilii- .A-L.. - ' 0.011 '|-1-“Il Iii“ i1 _AA.-A O .1 .‘liiiJiLL ? 1 L'.I .‘1 -.... -...-.4LAM- --.----1-L--.-i;-1 fl. II 1' CDC13 ). in w Nmr spectrum of biscyclopentenylidene 31 ( Figure 24. +“(l u —T V viff‘ v f vwvawv' v v v v V. vv v1 7 v v iv v v v v. 1v V .. vv v v v V 1 v 1 v v v fi fl '1 v v v7 ..V v v v v r ,l v v . v v v v. If v‘ L iv v. r v va Y O - 1-4.-L_8LFL _‘1-;.;; AAAAI--A Al II I. 1. Nmr spectrum of azazulene g2 (in CDC13). Figure 25. 100 Figure 26. Comparison of calculated and observed spectra of H1 and H4 of azazulene Q2, 101 m . g. v A f7 V .7 vv I. 1 vfi A . a ..- . a v I. j . vv i . v i 1' . A v. vaNll-lkll fill: 1115'... v.. v n A Y n l v, . A v. . TV . L .Y: . ~ I I v. . H . vv111...1- n 11 if _ A vs. . n n v . a .v t . . a . A Y. ....... n . . 4 v. .. . m ... v . . V hl‘lil‘l nil I'll .ll‘l 'll.«‘|. . -v M w . A v. . < a .o L V. _ r ._ . r v n _ Y. .0. a . w o v o . . J . . Y... 0‘- t 1 Us . vv u v. m . v r H Y.. 1 . A r _ _ . w. A v . - v . _ 31.-- 1 1 s. 111 1 - 1 v ... . 7v. . . . vi 1 vv 7 .v A 'v. A v. «T. l a. A V7 . v. A w A to 1 f. . r r. . . t. v A "T 1 v. A Z . v A v A r i L . v v A v . v . u. 1 v v v I I I I U I I J Nmr spectrum of cyclone g§,(in CDC13). Figure 27. 1-m I8» v. v. i v.» ‘1, Y... v.. .111 10.. 7.. 1.1» r ......... L-14 Jrnr Jilii 1AAA-L LIA -1-+--l,4 l --lkA- O 1. II II azulene g2 (in CDCl3 at 1000 Hz Nmr spectrum of sweep width). Figure 28. 102 J\: .3 mamaanmnm no uon Eauommm mmmz .mm musmflm em as aw 9m cm cm 9» ON 9N om. om 0v ON 0: om on De ON 4“ a : a . 1.0 -3 s rram é ..a é é W mow .om T 69 AlISNBlNI BMW—138 103 M\Z . ANNV. )Nsm mamasum mo uoam Efiuuommm mmm: . om musmfim mm form . EN , 0N, room? pm r 90$va PM bow? pm Jam, r rah _ UN. .02? Pm“; . Dim , DN P pm d . d d _ .. q : ‘1 fl. _. .. r09 ALISNELNI BAIlUWBH 104 “wo‘s‘o M/E 0 1 02 VjV 3 _(Cont.). Figure 30. ham pmo»oao»uhumzuhz~ IIIII lo..- .11.. ll. :0 I II' :- o O I I WMomv mamflsumnm a Ivua swamauflum N H.... . a- nun mo Enuuoomm mmmE ms» mo.me%Amnm HousmEou .m manna om" onhzwczau «Ohuzo: xazux¢ 01 no. 0 >uo chm o . ~m o.m~ 0. w.— oo- N.» msm ~.o ago" oacn m.~ >.on not" Nomn o. 0.0— noow ~.¢ nomm tan macaw hzuu awn oso hzuu awn can. a m gonoona ccmo.~o fiono.om acco.m~ Aomccsh Mano.oh mnmo.mh omqo.¢s ~03o.- memo.ho uono.mo mnmo.no «ascohm amoo.om memo.mm nuncomm mmmco~m om—c.om someone mvmcomo toqoono ooco.~c noncoqc mnmc.an ~3no.0~ m~no.3~ Dewyosm mn~9.5~ OcHOosw cmnooom co~Ooa~ -oo.s~ n—mo.¢~ moroomn mnwc.m~ mwh~mzwhz~ .Jmm ON. “WON-30H co.fi rmm:.rc~ o". moms.mo am. ”cmo._o em. name-95 ch. xoec.nh mo.~ mxmo.- cc. nomo.o~ mm. memo.ms um. mofio.es ms. Nsxc.~s NM. ammo.~o "n. coco.mo mm. ; ~n~o.no oo. ~o~o.sm co. enoo.om 0o.“ mmmo.mm mm. -mo.~m mm.“ osmo._m as. oodc.em o~.n cocoon: NN.~ hemocmc cs. omncomo cc. nscc.~¢ m».~ none..c ss.~ Nmmc.on ow. scso.mn w~.~ como.o~ co." MNMc.m~ $0.0” omao.>~ so." memo.- a». nfimo.- scog co~coo~ "c.mo oe~o.m~ ~m.o~ «woo.- nn. ocmo.o~ co. mcoo.md on. ammo.m~ .zomr» 44:»04 mqu owaamdwx umzhumzm»zn w>~h<4wa e am e c u ~Noocoo m o z zonxu mh~z~4 zouh<4304x o nuc oz~p4omua J as smoquo> oznufizog <_> auuzoomhzu mustcm .u cow umaah Ilylltnll ii..-»l. I ‘II ‘1‘. I:.II-II".I§‘ ol,il;.O i.--‘ ..I '!:i.‘. ..{xl . tw!cl|AO.‘lll.... - It- n»o~.~mm ham b~m hon~.~¢n newnogcn ~o-.ocn ucuuoonn nnogoomn but com—.own oo-om~m ~n.~.s~n; camcommn ~coo.o~n hat but ooo~.o~n s~o~.m~n -o~oo~n Nnoaonun pun but swunonon scouo~on ooooo—cn n~oc.oon mascoaom nmuug~o~ owedoao~ omoo.os~ chocosmw hum ocuuooa~ uOO¢¢0h~ cocooohw mono.~h~ nusoomsm but mono.no~ but but mooooomm “:s:.;onoo.un~ oquon >Jnnoo »xo»u10hUJ®Doc >Jraoa >Jmaoo >Jcaoo >Jmaoa >ao»uJmaoc >Jmaoa >choo >szoo >szo: rzobuzohuJmaoo >Jmaoo >Jmao: >Jmao: >zcwuxo»u¢umu»¢m oz .ou~.zo~ ou-zo~ ou~.zo~ cu-zo~ ou-zou --au-20~. au-zo~ owuuzou ouNuzou >zohu¢umnkJmDoc >Jmaoa >Jzao= >Joaoc >szoa >Jmaoo >Jmaoo >JQDOG >Jmao: >Jnnoa >Jmaoo >Jmaoo eraoo >Jmaos >Jaaoo >zohu1opuJmnoo .>zo»uJmaon; j-.§:»~u >¢oru¢um~>~ nmmm.oo~ ofimc.oo~ nsem.co~ “coo.mo~ uuom.oo~ -mo.con enmm.mo~ camo.no~ mocm.~ou once.~o~ mmem._o~ ocom.cm~ ~omc.2mH ommm.~m~ #omo.~m~ comm.om— neco.cm~ omcm.mm~ coco.mm~ ..emmm.~m_ oomo._m~ mmem.om~ mseo.omn “ocm.oo~ coco.oe~ ommm.mo~ chem.ec~ ohec.eo~ n~em.ne~ oomm.~c~ .mcmc.oe~ «mmm.on~ ofime.ann ~¢¢m.mn~ cseo.mn— omcm.bn~ ~ono.~n~ mono.on~ mocm.~n~ some.~nu omnm.en~ oomo.-~ soco.o~_ momm.m- osmo.m- nmwo.n- ammo.m- ~cqo.¢- csno.n- n~na.~”~ -CII‘I'I II. | II. {fir emu no .-z -z c~u no «a: «No no mu: «No no mgr nmu no huh >ao~u¢o»u¢opuJmaoo owNuzou >meoo ouNazoa >Jmaoa ouNuon >Jmaoa ow-zo~ >Jm3co ou-zo~ >Jmaoa ou-zo~ >Jmaoo uaokomu «A: z -z NA: 2 our 2 n“: mac»om~ 2 NA: n—r weapon" machomu tux ~_: 2 or m“: weapon” nu: -r maopom~ z «a: z a": uaopomu z -x 2 as: 2 $1 o~I Max 2 ca: - «a: uaohomn weapon" 2 -1 N“: a“: weapon" o~x a: -r ““1 ad: or a: CI 2. omu mmu tau ONO emu mmu ONO emu nmu nmu nmu nwu “Nu "mu "Nu -u “No «No "No ONO «we can o~u emu owu oNu o—u o—u emu emu onu o—u nuu coo.” coo." coo.“ coo.» cocoa mnmo .ooo.—_ coo.“ m~o. oo=.~ coo." ace.“ nmm. ooo.~ coo." coo.— N. com o.- Moo n.o~ ooc coon «cm .A~.u~‘ coc— o.» o.» o.e~ Mom ooo oosu o.~ m.o~ scmu Noh coda n.~— coma Gown coo" Now" oooN 9.0— not” con naa~.cmm “vno.nmm 9mac.¢~n cwauooun x¢caom~n asao.¢~m s~o~.mdn moa~.~on onao.ocm mnho.ao~ chaao~a~ -o~.am~ onaooQDN omnugom~ accuoasw cure-sum acmcossm mnho.mh~ ~x~o.c- s~=~omom n~zo.¢o~ doso.nom chocoomw @nao.~m~ «oncoumm Noncoomm cmoo.ro~ onwcocem "oaooonm NQNOoONN {ascommw omoo.c- omoo.oo~ ochoooou cm." ~co ac. ac. om. m~.o~ coon 9N. co. om. c3.~ ~(.¢ me. om.~ co.“ Hnom om.~ »~.c~ ~o.oc «mom x¢.~— omon ~e.~ a“. me. No. ~3. mo. nu. on.~ an.” snot on.” {com Nu. co.~ _o.~ moon oc.¢ cm.m oooh ~o.m oc.o mmfifi.hmm roo~.omm momoommn ~cooocmm e¢o~.o~n -o~.m~m -o~.¢~n ~no~.m~n smuaonon noodomon coco-«on nmoc.ccn mono-com nmnuouom so-goom omOa.ow~ omao.xm~ cmooosmm ~c~nomm~ om—uofixw coagoorw ~9oo.onN cano.nh~ mooo.-~ nsmo.o- nonc.m- coho.¢- neoo.mom homeoecm N¢L0.MON macaommw Naac.om~ cocoonnw ooQOonN onto-"mm ONmoocmN mmoc.mo~ OQOOoNcN NoncocoN raccoonm mnbooom~ mascommN momc.o- smho.~o~ some.oo~ moccoomn ohhcoahn Desmoahu 22$: coom.-~ ~no:.-~ _Hom.osn OhWOoOhu 108 Fun Fun but but 2 our swu mo >mohuzo»uzobu