V4— ms svmuesas OF KEY INTERMEDIATE ma me mammon or cvcmacmmmE-m-Qmms f Disseriation for the Degree of Ph, D. MECHEGAN STATE UNEVERSITY CHARLES LEE GERACL JR. 1975 LI 1" 72/; R Y Mic ligan hate Univer: .ty This is to certify that the thesis entitled 3' 5- _- The Synthesis of Key Intermediates for the Preparation of Cyclooctatetraene-l,5-Quinone presented by Charles Lee Geraci, Jr. has been accepted towards fulfillment of the requirements for Ph.D. Chemistry degree in Major professor ABSTRACT THE SYNTHESIS OF KEY INTERMEDIATES FOR THE PREPARATION OF CYCLOOCTATETRAENE-l,S-QUINONE By Charles Lee Geraci, Jr. Cyclooctatetraene-l,S-quinone has been predicted to be a ground state singlet and may be depicted as its dipolar resonance form lg. Appropriate substitution may stabilize the system to the point of being isolable. Past experience with the cyclobutadiene-l,3-quinone system x has indicated that electron releasing groups lend stability to such systems. Synthesis of the substituted quinone lQa has Ar = p-methoxy- phenyl been considered feasible. The quinone lga possesses two strain free valence tautomers g3 and g§, and synthetic approaches to these valence tautomers have been investigated. Charles Lee Geraci, Jr. Ar Ar £9, £5 Bicyclo[3.3.0]octane-2,6-dione was arylated with p- methoxyphenylmagnesium bromide and dehydrated. Investiga- tions into the synthetic utility of 2,6-bis(p-methoxy- phenyl)bicyclo[3.3.0]octa-l,5- and -2,6-diene were made. The latter diene on hydroboration and oxidation produced 2,6-bis(p-methoxyphenyl)bicyclo[3.3.0]octane-3,7-dione. Arylation of l,5-dimethylbicyclo[3.3.0]octane-3,7- dione with p-methoxyphenyllithium and dehydration produced a mixture of 1,5-dimethyl-3,7-bis(p-methoxyphenyl)bicyclo- [3.3.0]octa-2,6-diene, fig, and l,5-dimethyl-3-p-methoxy- phenylbicyclo[3.3.0]oct-2-en-7-one, gg. Attempts were made at arylating Qg and converting it into diene gé, which is a possible source of methylated quinone precursor fig by direct oxidation of its allylic carbons. THE SYNTHESIS OF KEY INTERMEDIATES FOR THE PREPARATION OF CYCLOOCTATETRAENE-l,5-QUINONE By Charles Lee Geraci, Jr. A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1975 TO CONNIE ii ADKNOWLEDGMENTS The author wishes to thank Dr. Donald G. Farnum for his guidance and assistance, as well as the freedom he af- forded during the course of this research project. His friendship throughout the work is also greatly appreciated. Gratitude is also expressed to the Department of Chemistry at Michigan State University for providing finan- cial support in the form of teaching assistantships from September 1970 to August 1975. In addition, some very special thanks are extended to my fellow graduate students and workers, especially to the members of the "Farnum Group." My character has truly been tempered by my close association with this unique col- lection of people. Finally, I cannot close without thanking Connie for trying to understand it all. 111 TABLE OF CONTENTS Page DEDICATION. . . . . . . . . . . . . . . . . . . . . . ii ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . iii LIST OF TABLES. . . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES . . . . . . . . . . . . . . . . . . . viii INTRODUCTION. . . . . . . . . . . . . . . . . . . . . 1 RESULTS AND DISCUSSION. . . . . . . . . . . . . . . . 18 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 55 EXPERIMENTAL. . . . . . . . . . . . . . . . . . . . . 57 General Procedures . . . . . . . . . . . . . . . 57 Tetramethyl Hexane-l,3,”,6-tetracarboxylate €§° Free Radical Induced Coupling of Dimethyl Glutarate. . . . . . . . . . . . . . . . . . 57 Bicyclo[3.3.0]octane-2,6-dione éé° Dieckmann Cyclization of Tetraester fig and Hydrolysis— decarboxylation of the Product, Ketoester g2 . . . . . . . . . . . . . . . . . . . . . 59 2,6—Bis(p-methoxyphenyl)-2,6-dihydroxybicyclo- [3.3.0]octane gé. Addition of p-Methoxy- phenylmagnesium Bromide to Dione §Q° . . . . 61 2,6-Bis(p-methoxyphenyl)bicyclo[3.3.0]octa-l,5- diene 31. Dehydration of Diol %Q to Conju- gated Diene g1 . . . . . . . . . . . . . . . 62 iv Page 2,6-Bis(p-methoxyphenyl)bicyclo[3.3.0]octa-2,6- diene. Dehydration of Diol 3Q to NonconJu- gated Diene 36 . . . . . . . . . . . . . . . 62 2,6-Bis(p-methoxyphenyl)bicyclo[3.3.0]octane @Q. Catalytic Hydrogenation of g1 to Produce saturated %§ 0 o o o o o o o o o o o o o o o 63 2,6—Bis(p-methoxyphenyl)-3,7-dihydroxybicyclo- [3.3.0]octane 3%. Hydroboration of Diene 36 to Produce Diol £3 . . . . . . . . . . . . . 6H 2,6-Bis(p-methoxyphenyl)bicyclo[3.3.0]octane- 3,7—dione. Oxidation of Diol 5% to Dione flité O O I O O O O O O I O O O O O O O O O O C 65 Dimethyl Acetonedicarboxylate g1 . . . . . . . . 66 1,5-Dimethyl-2,“,6,8—tetrakis(carbomethoxy)bi- cyclo[3.3.0]octane-2,6-dione fig. . . . . . . 67 l,5-Dimethy1bicyclo[3.3.0]octane-2,6-dione. Hy- drolysis and Decarboxylation of Tetraester g3 67 Reaction of p-Methoxyphenyllithium with Dione ED. Mixture of l,5-Dimethyl-3,7-bis(p- methoxyphenyl)-3,7-dihydroxybicycloE3.3.0]- octane g; and 1,5—Dimethyl—3-p-methoxy— phenyl-3-hydroxybicyclo[3.3.0]octan-7-one fig . . . . . . . . . . . . . . . . . . . . . 68 1,5-Dimethyl-3,7-bis(p-methoxyphenyl)bicyclo- [3.3.0]octa-2,6-diene fig and -2,7-diene $2 . 69 Page l,S-Dimethyl-3-p-methoxyphenylbicyclo[3.3.0]oct— 2—en-7-one gé. Dehydration of Hydroxy Ketone fig. . . . . . . . . . . . . . . . . . 71 Arylation and Dehydration of a Mixture Contain- ing Keto Olefin i3 and Diene éé. . . . . . . 72 LIST OF REFERENCES. . . . . . . . . . . . . . . . . . 73 vi LIST OF TABLES Table Page 1. Summary of Dehydrations of 3Q. . . . . . . . . 22 2. Chemical Reductions of Diene 31. . . . . . . . 3A vii LIST OF FIGURES Figure Page 1. Possible Interconversions of meta-Quinone 11. 16 2. Preparation of Dione 26 . . . . . . . . . . . l9 3. Proposed Synthesis of Diaryl—meta-quinone @Q. 20 A. Scheme for Producing 38 Under Birch Condi- tions . . . . . . . . . . . . . . . . . . 31 . Alternate Method of Preparing Diaryldione 3%. 3S 5 6. Scheme I for Utilizing D101 3%. . . . . . . . 39 7. Scheme II for Utilizing Diol Q; . . . . . . . 39 8 . Synthesis Scheme Using Bicyclo[3.3.0]octane— 3,7-d10ne o o o o o o o o c o o o o o o o ’43 9. Preparation of 1,5-Dimethy1bicycloE3.3.0]- octane-3,7-dione. . . . . . . . . . . . . an 10. Scheme for Utilizing Products Arising from Dione ég O O O O O O O O O O O O O O O O O l‘8 11. Chemical Shifts of Compounds éé, §6, and fig . 51 12. Allylic Oxidation Methods . . . . . . . . . . 53 13. Summary of Bicyclo[3.3.0]octane Transforma- tions . . . . . . . . . . . . . . . . . . 56 viii INTRODUCTION The introduction of molecular orbital theory by l Hfickel in 1931 was doubly beneficial to the organic chemist because it gave both a quantitative explantion of the concept of aromatic stability and a means to make some intriguing predictions regarding aromaticity. Simple Hfickel molecular orbital (H.M.O.) theory places the n-electrons of a cyclic conjugated system into orbitals which are a representation of the electronic structure of the entire molecule. The filling of bonding molecular orbitals results in enhanced stability to the system. Using this concept, it is possible to predict that compounds whose bonding molecular orbitals are com- pletely filled should possess the enhanced stability as- sociated with aromaticity. A simple statement of these conditions is contained in the Hfickel "An + 2" rule,2 which contends that "those monocyclic 00planar systems of trigonally hybridized atoms which possess 4n + 2 n-electrons will possess relative electronic stability." A corollary to this rule would state that cyclic conjugated systems containing An n-electrons do not possess electronic stability. These systems are referred to as antiaromatic systems. The declaration of Hfickel's rule set organic chem- ists in motion to test the concept of aromaticity on a much broader and certainly more exciting level. A wealth of novel and sometimes unexpected systems found both their origin and an explanation of their stability or instabil- ity in this simple concept. Another important facet of molecular orbital theory is the calculation of a system's delocalization energy (DE). The DE is defined as the calculated addi- tional bonding energy which results from delocalization of electrons originally imagined to be constrained to isolated double bonds.3 H.M.O. calculations of delocal- ization energies use a number of inaccurate assumptions, yet they are of some value when dealing with a system in a qualitative sense. A large number of neutral compounds present them- selves--many in retrospect--as proof of Hfickel's theory. Such traditional aromatic species as benzene, pyridine, furan, and pyrolle appear in this grouping. For a theory of this type to be functional, it should have the ability to predict both stable and un- stable systems. These would be the systems belonging to the An n-electron class. Cyclobutadiene is considered to be the classic four n-electron system. Many approaches have been taken “’5 but none has shown any evidence of the to the system, bare diene. In 1965 though, Pettit6 was able to prepare a very reactive compound by decomposing the cyclobutadiene- iron tricarbonyl complex. The compound was not isolable, but gave reaction products consistent with a brief exist- ence. CIv l|‘_'>’e 28‘ l l l I Fe(CO)3 In the class of eight n-electron compounds there is cyclooctatetraene,7 which exhibits antiaromatic char- acter in its reactions. The tub-shaped conformation8 of cyclooctatetraene does not allow any significant overlap of the n-electron systems. Even if overlap were possible, however, the fact that the molecule is a An system would probably cause more instability than in its nonoverlap con- formation. In fact, placing one or two more electrons into the system causes the cyclooctatetraene radical anion and dianion to flatten so that the n-systems are co- p1anar.9’lo’ll . The theory also sparked the generation of many novel compounds that are either neutral, charged, or electron deficient. An intriguing possibility to predict whether a charged system exists as a ground state singlet or triplet also presents itself in H.M.O. This predictive ability has generated interest in An n-electron systems and is based on the degeneracy of the highest occupied molecular orbitals, which are half filled. An inspection of the stable cyclic systems pre- pared or proposed should start with those containing two n-electrons. Breslow's triphenyl-cyclopropenium ion12 was the first of a series of this stable ion,13 with the 1U,15 The DE of the 6 parent ion being prepared recently. cyclopropenium ion is calculated to be 2.008.1 Cyclobutadiene dication would also be a two n- electron system, and its preparation has been attempted by 17’18’9’10’ll Though the parent cation has several groups. not yet been prepared, the tetramethyl derivative was re- ported by Olah in 1969.19 The next higher set of stabilized systems contain six n-electrons. The cyclopentadienyl anion is a 6-h sys- tem which exhibited its unusual stability rather early.2O Another 6-" system was introduced when Doering and Knox21 prepared the cycloheptatrienyl (trOpylium) cation. H.M.O. calculations give the tropylium cation a DE of 2.998.16 One other 6-n system of interest is the cycloocta- tetraene dication. There seems to be some evidence for its existence as an intermediate in the anodic oxidation 22’23 although its actual preparation of cyclooctatetraene, has so far eluded the organic chemist. Cyclooctatetraene dianion is a member of the class of compounds with ten n-electrons. Prepared by Katz in 1960, this ion proved to be planar and to possess the sta- bility shown in other An + 2 ions.9’10’ll The cyclopropenyl anion showed a reluctance to form which Breslowzu attributed to the antiaromaticity of this four n-electron system. KOt-Bu ‘ \ Very SlowL7 The cyclopentadienyl cation is a four h-electron system that has proved to be interesting. H.M.O. predicts that this cation should have a pair of degenerate highest occupied orbitals and should exist as a triplet. The 25 pentachloro cation seems to adhere to this prediction. The pentaphenyl derivative, however, is a ground state singlet, though only 0.55 kcal away from a triplet state.26 ¢ Cl Cl . Cl ¢ 0 l Cl 0 Q “ I ¢ ¢ Cl C1 C1 Cl Singlet Singlet Triplet Breslow attributed the behavior of the pentaphenyl ion to molecular distortion.27 The benzene dication is predicted by H.M.O. to be a triplet diradical, yet the possibility for distortion of this system is greater than for other four n-electron sys- 27 This fact may allow the benzene dication to pos- tems. sess a singlet ground state. The preparation of this par- ticular An system seems a rather formidable task since one is fighting the stability of the classic six n-electron benzene system. The cycloheptatrienyl anion is an eight n-electron system that is quite unstable.28 H.M.O. predicts a triplet ground state for the anion, but the heptamethyl derivative 29 Once again, this is con- has been shown to be a singlet. sistent with the idea of distortion of the molecule to destroy the degeneracy of the highest occupied molecular orbitals. The reduction of triphenylbenzene to its dianion 30 This system presents another eight n-electron system. was found to exist as a ground state triplet in accordance with simple H.M.O. predictions. The testing of Hfickel's theory with cations is sometimes an arduous task due to the nature of the cation itself. The introduction of a group that will mimic a cation center allows the easy generation of some interest- ing systems. Due to the dipolar nature of its bonding, the carbonyl is such a group. 5+ 5‘ \\ //C-—-O 4: :>' ::C _0 Carbonyl containing conjugated ring systems fall into two classes: (1) those in which the conjugated sys- tem contains an odd number of carbons and is neutral, and (2) those that have an even number of carbons in the ring and must be positively or negatively charged. The cycloprOpenone system does not qualify as an aromatic compound according to Hfickel's rule, yet it is 31,32,33 quite stable, making it analogous to a two n- electron system. The high dipole moment of 5.08D3u in diphenylcyclopropenone points out the importance of the resonance form 2. Resonance form 2 allows the system to attain a state that mimics the cyclopropenium ion. This argument is reinforced by the high reactivity of the less strained cyclopropanone system.35 Also, the fact that cyclOpropenones react with strong acids to form hydro- xycyclopropenium salts lends even more weight to resonance form 2.36 The DE of cyclOpropenone itself is calculated to be 1.363.16 Cyc10pentadienone follows the analogy to the cyclopentadienyl cation in that it is an extremely reac- tive molecule. When cyclopentadienone is generated, only its dimer is isolated.37’38 By analogy to both the cycloheptatrienyl cation 39,AO and to cyclopropenone, cycloheptatrienone may be considered to be aromatic in nature. A resonance form giving rise to aromatic tropylium ion can be envisioned. <——> A dipole moment of H.3Dul--as compared to 3.0AD for cyclo- heptanoneu2--indicates a small contribution from the di- polar resonance form. H.M.O. calculations give tropone a DE of 2.55s.“3 The introduction of both a carbonyl and a cationic center into an even numbered ring gives a system analogous to a dication. The much sought after cyclobutadiene dication fo- cused a large amount of attention on cyclobutenones as precursors to the dication system. The treatment of 2—bromo-3-hydroxy-2,M-diphenylcyclobutenone with sulfuric acid gave what appeared to be the dication, %.uu The ¢ /() ¢ OH / H so \ 2 ’4 \ /‘\ I 1+:I ¢ HO ¢ it cyclobutenonyl anion is expected to be quite unstable since it is a An system, and attempts at its formation bear this out.)45 The six-membered ring cation is analogous to the benzene dication and is, therefore, expected to be quite unstable. The anion, however, is the well known and quite stable phenoxide ion. Substituted cyclooctatrienones could act as pre- cursors to the corresponding cyclooctatrienonium cation as well as the hydroxycyclooctatetraene dication. O X 0 OH Systems possessing two carbonyl groups fall into the class of compounds known as quinones. 10 In the four-membered ring series there are two pos- sible isomers-—the 1,2- and 1,3-cyclobutenedione. The 1,2 isomer, ortho—quinone, is analogous to the cyclobutadiene dication, and being a two n—electron system it has possi- bilities of being stabilized. Several cyclobutene—3,U- A6,A7 diones have been prepared, and the parent system, though not prepared, has a DE of 1.2118.16 The stability of the 1,2 isomer is attributed, once again, to the con- tribution of resonance form A, which contains an aromatic two h-electron moiety. In this system, however, cross-ring :as :23: resonance forms 5 and 6 are possible. A dipole moment of 5.62D for the cyclohexene substituted 1,2 dioneu6 makes the dipolar resonance forms seem significant. The 1,3-cyclobutenedione, meta-quinone, does not possess a simple uncharged structure, and dipolar struc- tures Z and 8 may be drawn. Though the parent system has O R 0" R 0- R o / + I,“ 9 J2 L‘ 0 R o" R o" 0 8 10 ’V'b R Z rt 2, 11 not been prepared, MO calculations predict it to have a singlet ground state with final charge distribution indi— cating a significant contribution from 9.u8 Any signifi— cant contribution by the bicyclic valence tautomer lg in the parent system is not consistent with extended Hfickel calculations“8 and is ruled out in the diaryl substituted A9 50 meta-quinone by crystallographic analysis. There are three possible quinone isomers in the six-membered ring series. The 1,2—, the 1,3-, and the 1,U-cyclohexadienediones are known as ortho-, meta-, and para-quinone, respectively. Despite the fact that they are analogous to the benzene dication, the ortho- and para-quinones have been known a long time and are rather stable. Benzene meta-quinone, however, is unknown. The dipolar resonance contributors to this quinone, ll, are essentially related to the antiaromatic four w-electron benzene dication. If benzene meta-quinone were to exist as 0_ is}. R as a ground state singlet, it could well collapse to structure l2 and undergo further reactions of l2 cyclopropenones. However, since the two highest bonding orbitals are separated by 0.1A58,u3 it is possible that they may actually exist as a degenerate pair and that they may give rise to triplet species 1%. Recent work favors the singlet meta-quinone.51 O o - R R R R R ‘I <—-> 0' o O. R R R $2? The eight-membered quinones are those of cyclo— octatetraene. Here there are four possible isomers. The 1,2- and 1,U-quinones are the neutral species while the 1,3- and 1,5-quinones must be represented as either a dipolar or diradical system. The neutral 1,2— and 1,4-quinones are expected to show aromatic character by virtue of the possible dipolar resonance forms l5 and l5, which contain the potentially aromatic cyclooctatrienium dication moiety, 1%. The 0 o 0" 0- ‘ o ‘ Q 0’ 0 «Ft ti l3 dipolar resonance form 13 may be subject to electrical limitations. The electrostatic repulsion between the negatively charged oxygen atoms may raise the energy of la and thus decrease its importance as a resonance con- tributor. This argument could also be used against the cyclobuladiene-l,2-quinone. The high dipole moment of the cyclobutenedione system, however, tends to signal the im- portance of the dipolar resonance form. The problem of electrostatic repulsion is not as critical in the 1,4— quinone. The strain imposed upon the cyclooctatetraene—1,2- quinone and, to a lesser extent, the 1,“ system by the dipolar resonance forms may be relieved somewhat by cross- ring resonance. This type of resonance would generate the forms 15 and %6 from the 1,2- and 1,A-quinones, respective- ly. This type of cross—ring resonance is analogous to 0- O 0 JR té that shown for the cyclobutadiene—1,3-quinone. As stated earlier, cross-ring resonance in the cyclobutadiene quinone system has been shown to be of little importance. Hfickel molecular orbital calculations on the 1,2— 52 quinone show it to have a DE of 3.078. This prediction In .of aromatic stabilization has prompted numerous synthetic attempts at both 1,2- and l,A-quinones. Though the qui- nones themselves have not yet been prepared, their imme- diate precursors have been recently reported.53’5u A very important factor in attaining aromatic character for these systems is the ability to adopt a planar conformation. The resulting ring strain would be similar to that incurred in the flattening of cycloocta- tetraene, which is estimated to be 27 kcal/mole.55 The H.M.O. DE for the 1,2-quinone is 55.3 kcal/mole; this pre- dicts that attaining a flat conformation should not be a prohibitive factor regarding the aromaticity of a system such as the 1,2-quinone. The cyclooctatetraene—1,3- and -l,5-quinones like the benzene meta-quinone cannot be drawn as an uncharged species. These quinones must be drawn either as dipolar forms or as diradicals. O. 0 O 0 ‘ 0 Go 00 0 0 u The goal of the research presented in this thesis is the preparation of a derivative of the unknown cyclo- octatetraene-1,5-quinone, l7. Molecular orbital l5 calculations on the system indicate a singlet ground state.“8 The first excited state is predicted to be a triplet species with an Ea of 7.85 kcal/mole. The final charge distribution makes the dipolar resonance form 18 significant. 305, An important question is whether the planar, open ring molecule is more stable than its bicyclic isomers. If the 1,5-quinone is drawn as l9, one bicyclic isomer is 38. Extended Hfickel calculations show t3, and its many reso- nance forms, to be more stable than 2g. 0 O A 0 #2 89 A rather strain-free valence tautomer of 17, how- ever, is the bicyclooctadienedione 21. The orbital 16 8% symmetry requirements for the interconversion of 17 and 2% give rise to the sequence shown in Figure 1. This Q 6..) El§7%z Elérék MD NA 0 trans-21 trans—17 Figure 1. Possible Interconversions of meta-Quinone 11. sequence allows the interconversion of the 1,5 and 3,7 positions in 21 through the open quinone 11. Recent work by Pupko,56 however, has shown that deuterium labeled 21 does not show any conversion to 22 on photolysis; instead, the lactone 23 is formed. This product is very likely derived from d-bond cleavage to give a diketene 1? 09 we Pvt fit 8% intermediate. Several attempts to raise the energy of 57 and to the (n+w*) transition to eliminate a-cleavage allow B-cleavage through the (n+h*) transition failed in the photolysis of 21. An appropriately substituted 21 may make the (n+n*) state more photochemically accessible by lowering its energy. Assuming that this energy lowering will favor B-cleavage, it would seem that 2% or 25 would be logical precursors to the 1,5-quinone. Ar 0 O 00 Ar 00 Ar 0 A1“ 0 84‘. éé RESULTS AND DISCUSSION A synthetic approach to a derivative of cycloocta- tetraene—1,5-quinone through the valence tautomers 2% and 25 seems rather reasonable. The bicyclo[3.3.0]octane skel- eton of 25 and 25 was the most strain free of the possible valence tautomers of 11 and should, therefore, have offered the most accessible route to the quinone. Also, the cleav- age of a single carbon-carbon bond to produce a species such as 11, diradical or zwitterion, was probably more ef- ficient than other means. An entry into the bicyclo[3.3.0]octane system was readily available through bicyclo[3.3.0]octane-2,6-dione, 26. This dione was obtained in good yield according to the scheme in Figure 2, which was developed through collabora— 58 59 tion with Hagedorn by modifying Simpson's synthesis of a- Tetramethyl hexane-1,3,N,6-tetracarboxy1ate, 28, was prepared by the free radical induced coupling of di- methyl glutarate, 21. The Dieckmann cyclization of 28 to the bis(ketoester) 29 was best accomplished using sodium methoxide in dimethylsulfoxide (DMSO). The hydrolysis and decarboxylation of 29 to dione 26 resulted in a 70-85% yield--compared to 60% by Simpson--from the tetraester 28. 18 19 CO CH CH O CO CH 2 3 3 2 2 3 t-BuOOt-Bu \ NaOCHB \ ’ DMSO ’ CH3O2C CO2CH3 002w3 Pol éé O 0 HCl/H2O CH302 .. CO2CH3 > .. 70-85% 0 from fig 0 82 9M6. Figure 2. Preparation of Dione 26. Substituting cyclooctatetraene—l,5-quinone with electron—releasing groups should stabilize the system. This reasoning was based on the success encountered in the cyclobutadiene-l,3-quinone system when electron-releasing aryl groups were used as substitutents.l49 Also, the di- polar nature of the parent system 28 would have indicated the use of electron-releasing groups for any gain in sta- bility. A proposed synthesis of an aryl substituted 22 is given in Figure 3, with Ar = p-methoxyphenyl. Some reactions on the systems in Figure 3 have been reported by Wawzonek,6O who prepared only the phenyl deriv- atives. The di-p-methoxyphenyl systems, however, showed markedly different behavior chemically, Vida infra. 2O O ‘r Ar $9 at it H2,Pd/C V 0 Ar r Ar .r / NBS 1 [0] .J. Ar Ar 0 Ar 0 Figure 3. Proposed Synthesis of Diaryl-meta-quinone 35. 21 Treatment of dione 26 with p-methoxyphenylmagnesium bromide gave diol 22 in 75% yield as a white crystalline solid, mp 187-190°. The diol was quite insoluble in most solvents and was best purified by a Soxhlet extraction using refluxing ethyl acetate. The ir spectrum of 22 show- ed a hydroxyl absorption at 3150 cm_l. The dehydration of diol 32 to 2,6-di-p-methoxy- phenylbicyclo[3.3.0]octa-l,5-diene, 22, failed using sup- posedly standard methods. Treating diol 22 with a number of dehydrating reagents used under normal conditions re- sulted only in black, uncharacterized material. This par- ticular occurance gave the first indication of the reactiv- ity of the di-p-methoxyphenyl system. When treated with a catalytic amount of toluenesul- fonic acid, diol 3Q dehydrated readily. The dehydration produced not only the conjugated diene 32, but also diene 36, depending on the reaction conditions. Table 1 gives the various dehydration methods, conditions, and products. Ar Ar Table 1. Summary of Dehydrations of 32. Dehydrating EXP Agent Conditions Results 1 Formic Acid/Acetic Reflux 2 hours Very crude Acida product 1:3 ~30% 31 2 TsOH (0.6 equiva- Reflux in benzene Very crude lents) product Mix of 32 + 26 3 KHSOua 160° 1 hour Very crude product ~20% g; u TsOH (uz molar .05 g 30 in ben- not sic quantity) zene, reflux 12 60% 3Q hoursb 5 TsOH (14% molar 0.1 g 353 in ben- 71:: ad quantity) zene, reflux 12 20% 36 hours 6 TsOH (uz molar 0.15 5,3g in ben- 89% $23 quantity) zene, reflux 1/2 11% mix 22 + 36 hr., seed with 22, continue re- flux 2 hours 7 TsOH (catalytic .02 M 29 in ben- 21% @2f amount) zene 12 hours 79% 26 a. See reference 60. b. H20 removed in large scale reactions. c. Product yield 85%. d. Product yield 80% e. Product yield 90%. f. Product yield 98%. 23 As seen in Table l, diol $0 dehydrated quite easily to dienes $1 and $Q, which were readily separated because of their different solubilities in benzene. Allowing the dehydration mixture to cool in reactions U through 7 in Table 1 caused conjugated diene 31 to crystallize. The more soluble éé stayed in solution, allowing separation of «a. Conjugated diene %l was identified by its spectro- scopic properties. The uv spectrum showed an absorption maximum at 350 nm (log a 4.63), with strong absorptions at 33“ nm and 370 nm. This was consistent with the uv spec- trum of trans, trans-l,“ diphenyl-1,U-butadiene, 31, which had absorptions at 33“ nm (log a H.68) and 352 nm (log 5 5.38). It can easily be seen that @Z was incorporated into «3%- ¢ Ar / l"‘ \ I ‘I \ / ~ I, / \ Ar ¢ ~31 5% The presence of the methoxy on the aryl group caused an absorption shift to a longer wavelength, as seen in a comparison of the uv absorptions of styrene with p- methoxystyrene. 2M // © Amax 2115 log 8 11.18 // © Amax 260 1go e 14.17 292 3.30 OCH3 The nmr spectrum of 31 showed an AA'BB' aromatic region, a methoxy signal, and two symmetrical multiplets in the methylene region at 6 3.5-3.2 and 6 3.1-2.7. The signals integrated for eight, six, four, and four protons, respectively. The mass spectrum of 31 showed a strong parent peak at m/e 318. The above data were consistent with the structure assigned to 31. Diene 36 was also assigned its structure based on spectroscopic data. The uv of 36 showed a Amax at 268 nm (log 6 “.53), which was consistent with that of p-methoxy- styrene. The nmr spectrum of 36 showed an AA'BB' pattern in the aromatic region, multiplets at 6 5.9 and 6 u.l-3.9, a methoxy signal at 6 3.8, and a multiplet at 6 2.9-2.u. The signals integrated for eight, two, six, and four protons, respectively. The mass spectrum also showed a strong par- ent peak at m/e 318. These data were consistent with the structure assigned. 25 The physical properties of 33 and 36 made distin- guishing them rather easy. Conjugated diene 33 was a yellow—green solid with a melting point of 233-235°, while 36 was a salmon solid with a melting point of l90—193°. Diene 33 was rather insoluble in benzene, but in solution it had a slight orange-yellow color and fluoresced strong- ly. Diene 36 was very soluble in benzene and gave a deep orange solution with little fluorescence. With diene 33 then available, the next task was its reduction to 2,6-bis(p-methoxyphenyl)-Al’5-bicyclo[3.3.0]— octene, 33. This conversion required the l,h~reduction of the diene system in 33. Wawzonek was able to reduce the diphenylbicyclo- [3.3.0]octadiene to the central olefin with 40% sodium amalgam in ethanol.60 His yield, however, was rather low (20%). When diene 33 was treated with u0% sodium amalgam in ethanol, a crude product mixture was obtained in 40% yield. The nmr spectrum of the product was inconclusive, but the uv showed the presence of both 33 and 36 in the product. Inspection of the product mixture by tlc showed it to consist mainly of dienes 33 and 36. It appeared that an isomerization of 33 to 36 had taken place; this was the first indication of the sensitivity of conjugated diene 33, which later became important. Because of the apparent lack of success with the sodium amalgam reduction, an attempt was made to 26 catalytically add one mole of hydrogen to 33.61 When diene 33 was allowed to absorb one mole of hydrogen in the pres- ence of 5% palladium on carbon, a mixture of products was obtained. Approximately 50% of the product mixture con- sisted of dienes 33 and 36, and the rest of the product was a white crystalline solid, mp l20-122°. The uv spectrum of this white product showed Amax 225 nm (log 5 H.53) and 275 nm (log 6 3.77). These absorp- tions were characteristic of an anisole moiety. The nmr spectrum showed an AA'BB' pattern at 6 7.2-6.6, a methoxy signal of 6 3.7, and multiplets at 6 3.2-2.8 and 6 1.9—1.2 in a ratio of eight, six, four, and eight protons, respectively. The mass spectrum showed a strong parent peak at m/e 322. These data were consistent with the structure 36. Ar Ar 3% As supporting evidence for structure 36, a sample of diene 33 was hydrogenated until two moles of hydrogen were absorbed in order to produce a sample of 2,6-bis(p— methoxyphenyl)bicyclo[3.3.0]octane. The hydrogenation product was a white crystalline solid, mp l20-l22°, with 27 identical spectral properties and go retention time as those of the earlier obtained product. In addition, a go of a mixture of the two showed only one peak. The comparative study showed that diene 33 chose to go to a totally saturated system rather than the central olefin 33. This was reinforced by the recovery of dienes %l and aa- At this point it is of value to talk about the sus- ceptibility of conjugated diene 33 to isomerization. The first attempts to reduce 33 resulted in the production of 36, as well as other products. This bit of information, plus the short study that follows, gave some insight into the sensitivity of diene 33. When diol 36 was originally dehydrated, all the spectral data for the crude yellow-green product indicated structure 33. Since the product was crude and appeared off-color, it was dissolved in benzene, treated with Norit, and allowed to crystallize. The solution and, eventually, the crystalline product took on an orange color. This was assumed to be the actual color of pure diene 33. The melting point of this material, however, had a range of 180-220°. The uv spectrum showed not only the characteristic triplet centered at 350 nm for conjugated 33, but also the 268 nm peak later assigned to 36. A sample shown by nmr to be essentially 100% 33 was dissolved in benzene, treated with Norit, and allowed to crystallize. The now orange solid was checked by nmr 28 and a peak then appeared in the vinyl region, along with the previously symmetrical methylene region being quite disrupted. A careful integration of the spectrum showed that diene 33 had been converted to a mixture of 36% 33 and 6U% 36. If exposure of 33 to Norit caused its isomerization to 36, the same action was anticipated when 33 was exposed to a carbon supported catalyst. This could explain the complete hydrogenation of 33, as well as the production of 36, when the attempt was made to add one mole of hydrogen using a palladium on carbon catalyst. When diene 33 was left in solution for any length of time it began to take on an orange color. The longer it was left in solution, the deeper the color became and the stronger the spectral evidence for the presence of 36. It was also found that the rate of isomerization of 33 to 36 was greatly accelerated with the use of acid- washed glassware. A solution of 33 maintained its integ- rity much longer in base-washed glassware than in acid- washed glassware. The acid sensitivity of 33 was demon— strated when it was exposed in a benzene solution to tolu- enesulfonic acid. The solution quickly developed a deep orange color, and only diene 36 could be identified by nmr. Any attempts to isomerize 36 back to 33 resulted in the production of a black amorphous solid plus some recov- ered 36. This could well indicate that the conjugated diene 33 is the kinetic product of the dehydration of diol 29 36 and that the nonconjugated diene 36 is the thermodynamic product. All of the information presented above must be kept in mind for future dealings with diene 33. In another attempt to generate the central olefin 33, a 5% palladium on barium sulfate catalyst was pre- pared.62 It was hoped that this less active catalyst would promote the 1,D-addition of hydrogen to the diene system.63 The only reaction product isolated, however, was the satur- ated 36. It is quite possible that the finely divided catalyst surface promoted the isomerization of 33 to 36. At this point, it was decided to investigate the reduction of 33 by chemical means. To this end, several dissolving metal Birch-type reductions were attempted.624 When diene 33 was treated with lithium in liquid ammonia, a product mixture consisting mainly of some re- covered 33, a larger percentage of 36, and the saturated 36 was obtained. There was, however, 20% of another product formed. The additional product had a gc retention time less than that of 36. The peak was actually an unresolved pair. A study of the reduction products after removal of dienes 33 and 36 was made, and a gas chromatogram of the reduction products showed them to consist of 62% totally saturated 36 and 38% unresolved products. The area ratios of the two peaks showed them to represent approximately 13% and 25% of the total product. 30 A mass spectrum of the doublet peak collected from the go showed a parent ion of m/e 320. This indicated that some type of mono-olefin had been produced. The uv spec- trum did not show a peak at 268 nm, thus eliminating a structure such as 33. Ar Ar .32 The only peaks seen in the uv spectrum were char- acteristic of the anisole moiety. The nmr and ir were in- conclusive. The unresolved doublet of peaks in the gc could be explained by the presence of the cis and trans isomers of 328° Ar Ar trans 3g gl§_33 Very similar results were obtained using a sodium/ liquid ammonia system. In this reduction, though, the m/e 31 320 product was slightly increased to “2% of the reduction products. The use of calcium in liquid ammonia,6u a milder reducing medium, gave results almost identical to the other dissolving metal reductions. The presence of 36 in these reductions could be a result of the scheme presented in Figure A. A A H 93 ——> CD ——> CD —-> CI Ar Ar Ar HAI’ H Ar o ——>! .99. ——>» 0% Figure 4. Scheme for Producing 36 Under Birch Conditions. Reduction of the anisole groups, giving a structure such as A , was ruled out by the absence of vinyl absorp- tions in the nmr spectrum. 32 O H 03 Ar $3.9 Several factors were considered in determining the synthetic utility of the above Birch-type reductions. First of all, the yields of reduction products were low-- 20% to UO%. The product that may well be 33 constituted at most “2% of the product mixture. This indicated an actual formation of 33 in only 8% to 16% yield. The second consideration was that the prospects for increasing the yield or quantity of 33 produced by the above methods were rather poor. Diene 33 is not soluble in liquid ammonia. When THF was used as a cosolvent, the diene fell out of solution rather quickly at liquid ammonia temperatures. These solubility factors alone allowed Birch-type reductions on 33 to be done only on a 10-3 molar scale. It appeared that the liquid ammonia-dissolving metal reduction was just not a feasible synthetic method for this system. This prompted the investigation of other 33 chemical means of reduction; these methods, along with those already mentioned, are presented in Table 2. The sodium napthalenide reduction showed some promise in that the ratio of the 320 mw compound to 3% was greatly increased. The 320 mw component was separated from the product mixture on an alumina column and was still an unresolved pair of peaks by gc. Unfortunately, the small amount of this compound collected, 0.03g, turned brown on standing and did not resemble the original mate— rial by gc. At this point it was decided to introduce function- ality other than the central double bond into the bicyclo- [3.3.0] system. At the same time though, it did not seem feasible to abandon the diene system 3%. Brown68 has reported the oxidation of bicyclo- [3.3.0]octane-l-boronic acid to bicyclo[3.3.0]octan-l—ol. It appeared that a reaction of this type at both bridge- [ 0 J \/ B(OH)2 OH heads of the [3.3.0] system could be of value, as shown in Figure 5. The production of the diaryldione a; would still allow entry into the original synthetic scheme (Figure 3, p. 20). 3H Table 2. Chemical Reductions of Diene 3%. Reducing System Conditions Results NA/Hga Refluxing benzene, Recovered 3% + 36 Ethanol 2 hours Pd/C THF, Room temp., 1 atm 90% 3Q H2 Pd/C THF, Room temp., 1 mole 50% 3% + 36 & 50% 3Q H2 Pd/BaSOu THF, Room temp., 1 mole 50% §% + éé & 50% ;§ H2 Li/NH3 THF Cosolvent, Excess Li 50% 3% + 36, 31% 3Q & 19% mw 320 product Na/NH3 THF Cosolvent, Minimum HOZ 3% + 36, 35% 3Q ammonia & 25% mw 320 product Ca/NH3 THF Cosolvent, Excess Ca “0% 3% + 36, 40% gQ & 20% mw 320 product Li/THF/ROHb Reflux in THF then add 3% + 3g only alcohol Na/ THF solvent, alcohol Minimum 3% produced Napthalene added & mw 320 in better yieldc a. See reference 60. b. See reference 66. c. See p. 31. o 35 Ar \ / ‘r 0 Ar 0 Ar We ”W. Ar Ar Ar OR /B\ Ar 0 it 3% %% Figure 5. Alternate Method of Preparing Diaryldione 33. The hydroboration of p-methoxystyrene produced the l-alcohol and 2-alcohol in relative ratios of 98 to 2, OH HO A o + o 2. [o] ” OCH OCH3 OCH3 3 2% 98% respectively.69 This directive ability made diol 3% seem a rather attractive synthetic objective. Diene 3% was treated with excess diborane in THF for twelve hours. The reaction was monitored by the dis- appearance of the three absorbances in the uv spectrum cen- tered at 350 nm. The reaction was considered complete when 36 the 350 nm absorbance was replaced by the lower intensity 225 and 275 nm absorbances. Oxidation and work—up of the borane resulted in a 95% yield of a semisolid oil. Treating the oil with etha- nol and hexane resulted in the formation of a white cryst- alline material, mp l60-l65°, in 50% yield. The broad melting point range indicated the product was not pure. The solid product consisted of one major and one minor component by tlc. The initial oil contained the same two components plus three minor components. Recrystallizing the solid several times afforded a solid that was homogeneous by tlc and had a melting point of 17u—175°. This compound, the major component in the product mixture, showed a hydroxyl absorption at 3225 cm-1 in its ir spectrum. The nmr spectrum showed an AA'BB' pattern in the aromatic region at 6 7.2-6.7, a multiplet at 6 u.3-3.9, a methoxy singlet at 6 3.7, a multiplet at 6 3.2-2.8, and a multiplet at 6 1.7-1.4. These signals in- tegrated for eight, two, six, four, and six protons, re- spectively. The multiplet at 6 1.7-1.H integrated for four protons after a D20 exchange. The mass spectrum showed a parent ion at m/e 354. The nmr spectrum for 5% was expected to show a two proton multiplet for the benzylic positions and an eight proton multiplet for the methylene region. This prediction was based on the nmr spectrum obtained for 2,6-bis(p- methoxyphenyl)bicyclo[3.3.0]octane, 3Q. 37 There was little comparison between the hydrobora- tion product and the 2,6—diol 30. Also, the physical prop— erties were quite different. These data eliminated the 1,5-diol Ql and the 2,6- diol 39 as possible structures for the major hydroboration product. Elimination of structure Ql was further reinforc— ed by the reluctance of the material to oxidize under 1,2- diol fragmentation conditions.70 The possibility existed that an isomeric diol such as 52 had been formed. The spectral data, however, were not consistent with this structure. OH ArOH The susceptibility of the 1,5-diene 3% to isomerize to the 2,6-diene 36 was again surfacing as an important consideration. At this point it seemed necessary to in- spect the hydroboration product or products of the 2,6- diene 36. Diene 36 was subjected to the same hydroboration conditions as 3i. The reaction was monitored by following the disappearance of the 268 nm absorbance and the emer- gence of the 225 and 275 nm bands in the uv. 38 Oxidation of the borane gave a white crystalline solid, mp l7u-l75°, in 90-99% yield. All the spectral data for this compound were identical to those obtained from the hydroboration of conjugated diene 3%. The tlc of both com- pounds showed identical Rf values either individually or mixed. A mixed melting point showed no depression. All of the above information was consistent with structure g3. Ar HO OH Ar It was now obvious that the conjugated 1,5—diene 3% was going to react as it chose. The choice, it appear- ed, was to react through the isomeric nonconjugated 2,6- diene 36. The acceptance of this fact prompted an investi- gation into the potential synthetic utility of 2,6-bis(p- methoxyphenyl)-3,7—dihydroxybicyclo[3.3.0]octane, Q3. One synthetic scheme envisioned is presented in Figure 6. Unfortunately, this sequence had several serious faults. The dehydration of Q3 had to proceed through one of three possible directions, but the hydration of fig had no strong directing factors to produce the correct diol of the three possibilities. Also, the aryl groups would end 39 up in the 2,6 positions rather than the preferred 1,5 or 3,7 positions. Ar AP 0 AP 0 Ar Howe—>9 Ar Ar Ar 0 Ar 0 U3 an "0 ’V'b Figure 6. Scheme I for Utilizing Diol 33. Another possible use for 33 is presented in Figure 7. The key factor in this scheme was the potential for Ar 0 ——> 0 Ar £33 it fit AI‘ 0 Al" 0 co <~ <— 0 Ar 0 Ar at Figure 7. Scheme II for Utilizing Biol 53. MO dione £5 to be opened to the diarylcyclooctadienedione 56. One possible method for this conversion would be a base in- duced fragmentation of the brominated dione fix. The A (”Br A 0 ——> O ‘§IU° Ar #1 it closure of 36 could be accomplished by some method similar to Henry's closure of cycloocta-l,5—diene.71 It might also OAc PdCl2 \ Pb(0Ac)u OAc be possible to close 56 by using an agent such as bromine. 55 v K __ Ar Br “1 This particular method suffers from the possibility of ring closure to the [U.2.0] system 3%. Br é AI'BP 36,1 The oxidation of diol £3 to 2,6-bis(p-methoxy- phenyl)bicyclo[3.3.0]octane-3,7-dione, 35, was accomplished by using chromium trioxide-pyridine in methylene chloride. The success of this oxidation was rather dependent on at- taining a complete solution of 33 in the methylene chlor- ide. The diol Q3 was not readily soluble in small amounts of the solvent. The first attempt at oxidation of 53 led to a mix- ture of dione Q5 and the hydroxy ketone £8. These were HO 0 Ar easily separated tflunuyl because of their different solu- bilities in benzene. Any isolated 38 could be oxidized to dione Q5. U2 Dione $5, a white crystalline material, mp 190- l93°, was produced in a 70% yield from the oxidation of diol £3. The ir spectrum showed a carbonyl absorbance at l7UO cm-l. The nmr spectrum was not very conclusive, show— ing rather broad multiplets in the methylene region. The mass spectrum showed a parent ion of m/e 350 and no evi- dence of diol 33 or hydroxy ketone $8. In an attempt to increase the yield of dione $5, the direct oxidation of the borane produced was investi- 7H gated using both chromium trioxide and sulfuric acid and 75 chromium trioxide-pyridine. The latter gave better re— sults in terms of carbonyl produced, but both gave low yields. A major problem was the insolubility of the or- ganoborane. Also, the Jones oxidation of the borane seemed to be producing some carboxylic acid products. This direct method did not seem adequate when compared to the sequen- tial approach. When diene 36 was hydroborated and oxidized to diol 33, the isolation of Q3 allowed for greater control of the oxidation. A typical run through using the separate syn- thetic steps produced dione $5 in 84% yield from diene 36. Further investigations into this particular bicyclo[3.3.0] system were temporarily suspended in favor of the bicyclo- [3.3.0]octane-3,7-dione £2. The potential for correct placement of aryl substitutents seemed greater in 5%. The M3 moon £2 series of synthetic steps necessary to use $2 is given in Figure 8. HO OH A 003:0 r2033. 9 <33 0 O o A Wm. item 0 O OH éé Figure 8. Synthesis Scheme Using Bicyclo[3.3.0]octane-3,7- dione. Upon investigating the reported syntheses for bicyclo[3.3.0]octane-3,7-dione, it became apparent that the most accessible system was l,5-dimethylbicyclo[3.3.0]- octane-3,7—dione, 50. UM CH CH 29 3 Dione 50 was easily prepared?!4 by the sequence pre— sented in Figure 9. Dimethyl acetonedicarboxylate, 5%, was it 2% 39 Figure 9. Preparation of l,5—Dimethylbicyclo[3.3.0]octane- 3,7—dione. prepared by the sulfuric acid oxidation of citric acid.75 The tetraester dione 50 was prepared in 85% yield by al- lowing 5l and 2,3-butanedione to stir in aqueous solution for five days. A yield of 95% was obtained overnight by buffering the solution to pH 5. The hydrolysis and decar- boxylation of 52 was accomplished by refluxing in 6 M HCl for five hours and was essentially quantitative. Yields of 50 from 5% were in the range of 8U-95%. 45 The yield of diaryl diol 53 when dione 50 was treated with phenylmagnesium bromide was a rather'disappoint— ing 20%. The diol had to be separated from a product mix- ture that was initially a viscous oil. When purified, diol 53 was white crystalline solid which sublimed at 136° in a sealed melting-point tube. The ir spectrum of 53 showed a hydroxyl absorption at 317“ cm-1. The nmr spectrum showed an AA'BB' aromatic pattern at 6 7.3-6.6, a broad singlet at 6 U.“, a methoxy singlet at 6 3.7, an AB quartet centered at 6 2.38, and a singlet at 6 1.1. The AB quartet had a coupling constant of 14 Hz. These signals integrated for eight, two, six, eight, and six protons, respectively. The singlet at 6 u.u washed out with D20. A symmetrical structure such as 53 would give rise to a single peak for the methyl groups. The mass spectrum showed strong peaks at m/e 364, 3&6, and 331, corresponding to two losses of water and loss of methyl. All the spectral data were consistent with structure 53. A6 A second consistuent of the Grignard reaction mix- ture was a white solid which melted at 123-127° in a sealed melting-point tube. The ir spectrum of this product showed a hydroxyl absorption at 3333 cm"1 and a carbonyl absorp- tion at 1724 cm-1. The product was homogeneous by tlc. The nmr spectrum exhibited an AA'BB' aromatic pattern at 6 7.A-6.8, a methoxy singlet at 6 3.7, an AB quartet centered at 6 2.A6, a singlet at 6 2.17, a broad singlet at 6 1.93, and a singlet at 6 1.13. The coupling constant for the AB quartet was 16 Hz. The broad singlet at 6 1.93 was washed out with D20. The mass spectrum showed a molecular ion at m/e 27“, with strong peaks at m/e 256 and m/e 2A1 corres- ponding to loss of water and methyl, respectively. On the basis of these data, the hydroxy ketone 55 was proposed for the structure of this product. CH Ar The AB quartet in the nmr spectrum was consistent with the data obtained on diol 53 for the methylene pro- tons adjacent to the hydroxyl. The singlet at 6 2.17 was also consistent with nmr data obtained from dione 55 for the protons adjacent to the carbonyl. “7 It appeared that the Grignard reaction on dione 55 was rather difficult and resulted in a low yield. The dif- ficulty of the reaction was also reflected in the fact that a large percentage of the product isolated was the hydroxy ketone 55. A significant factor in the low yield may well have been the formation of the cyclopentanone enolate. This was supported by the detection of anisole early in the reaction. Longer reaction times and higher temperatures did not affect the yield greatly. If a synthetic scheme through the diol 55 was still to be considered, some means of increasing the yield became imperative. To this end the dione 55 was treated with p— methoxyphenyllithium. A product mixture of hydroxy ketone 55 and diol 55 was thus obtained. The combined yield, how- ever, was greater than that obtained with the Grignard. The reaction showed a 66% reaction based on the amount of dione that reacted. It was later found that the product consisted of 16% yield of diol 55 and a 50% yield of hy- droxy ketone 55. It was not difficult to devise a synthetic scheme that made use of the products obtained thus far. The main idea was to put the aryl groups on in separate steps. A series of arylations and dehydrations demonstrating this sequence is given in Figure 10. The diaryldienedione 55 was a possible cyclooctatetraene-l,5—quinone precursor. U8 CH CH3 C HO 3 O H—> 0 Ar —>Ar .9 Ar Ar CH CH3 3 CH3 7% 2% t1 CH O O CH 3 % H3 CH3 C 3 0 HO OH firms-e“ CO A1“ A1” AI’ CH3 C 0 CH3 H 2; it 3 2% Figure 10. Scheme for Utilizing Products Arising from Dione 55. Care had to be taken in the dehydrations since 55 was not the only diene capable of forming in this particu- lar synthetic scheme. The 2,7-diene 59 could also have been generated. This might not have been detrimental, however, since it could conceivably have been used to gen- erate the precursor 55 to cyclooctatetraene—1,3—quinone. CH3 CH Ar Ar Ar Ar CH3 CH3 0 22 29 “9 Diol ég was the first compound to be dehydrated, and, depending on the reaction conditions, diene 3% and/or diene gé were the products isolated. Diol g3 was dissolved in benzene at a high dilution and was refluxed overnight with a catalytic amount of toluenesulfonic acid. An off-white solid with a melting point of 113-117° was isolated. The ir spectrum showed no evidence of hydroxyl. The nmr spectrum showed an AA'BB' pattern at 6 7.3—6.6, a broad singlet with multiple split- ting at 6 5.73, a methoxy singlet at 6 3.7, a broad singlet with multiple splitting at 6 2.68, and a singlet at 6 1.18. These peaks integrated for eight, two, six, four, and six protons, respectively. The six proton singlet at 6 1.18 arose from the two methyl groups and was very good evidence for the symmetrical diene QQ. The mass spectrum showed a strong parent ion at m/e 3H6 and a strong peak at m/e 331. The above data were consistent with structure 36. When diol §3 was dehydrated under conditions of higher concentration using 4% toluenesulfonic acid, an oil that gradually solidified was obtained as the product. This product had a melting range of 112-170°, indicating a mixture. The ir spectrum showed no evidence of hydroxyl. The nmr spectrum was essentially the spectrum of éé with a second set of peaks being present, those being: a singlet with fine splitting at 6 6.1, a broad singlet with fine splitting at 6 2.“, a singlet at 6 1.28, and a singlet at 6 1.10. These peaks integrated for two, four, three, and 50 three protons, respectively. The two three-proton singlets corresponded to the two nonequivalent methyls in diene 5%. The mass spectrum showed the same major peaks as 56. The relative ratio of 56 to 53, based on integration of the vinyl peaks in the nmr, was 5 to l, and it appeared that it was possible to generate only 56 or a mixture of 56 and 5% based on reaction conditions. Subjecting hydroxy ketone 5% to similar dehydrating conditions gave a solid off-white product, mp 103-105°, in 92% yield after only two hours of refluxing. The ir spectrum of this product showed no hydroxyl and had a carbonyl absorption at l7u0 cm-l. The mass spec— trum showed a good molecular ion at m/e 256. The nmr spec- trum contained an AA'BB' aromatic pattern at 6 7.3-6.7, a broad singlet at 6 5.72, a methoxy singlet at 6 3.7, a singlet with fine splitting at 6 2.67 (J = 2 Hz), a compact multiplet at 6 2.35-2.27, a singlet at 6 1.18, and a sing- let at 6 1.12. These signals integrated for four, one, three, two, four, three, and three protons, respectively. The spectral data were consistent with structure 55. The nmr chemical shifts used for the identification of pure compounds and the composition of mixtures are given in Figure 11. The chemical shift differences between 56 and 52 made it very easy to detect either of them in a product mixture. There were enough different peaks and sufficient chemical shift differences to make rather accur- ate estimations of mixture composition. 51 .18 1 28 35- 2 057 /5 73 AI.-‘3\73 /1.18 ' 1 /o 2.HO $2 Qé 3'32 Figure 11. Chemical Shifts of Compounds 55, 55, and 55. It has been demonstrated that 55 can be dehydrated quickly and efficiently. The dehydration of a mixture of 55 and 55 may, therefore, be designed to accommodate the desired dehydration product of the diol. When a mixture of diol 55 and hydroxy ketone 55 was subjected to high dilution dehydration conditions, the product was an oil that quickly solidified when scratched. The nmr spectrum indicated the presence of 55 and diene 55 (in a ratio of 82:18). No isomeric diene 55 was detected. Using this product ratio and the weight of the product ob- tained, compound 55 was formed in 96% yield and compound 55 was formed in 99% yield. The exclusive formation of diene 55 from the mixture was very encouraging. In order to complete the synthetic sequence to diene 55, the carbonyl in 55 had to be arylated. The re- sulting hydroxy olefin 55 was then subjected to dehydration conditions that favored exclusive formation of 55. 52 CH CH HO AP —9 Ar Ar Ar CH3 CH «‘él éé When a mixture of keto olefin 55 and diene 55 was 3 treated with p-methoxyphenylmagnesium bromide in THF, alco- hol 55 was produced in 79% yield. Dehydrating the product mixture at high dilution produced diene 55 in 95% Yield. The yield obtained from the arylation of 55 was rather en- 76 couraging considering the low yield Askani reported in the preparation of 55. The overall yield of diene 55 from dione 55 was increased to 5U%. CH MgBr ¢ 0 3 3, ¢ OH 58% CH3 CH 3 CH3 é}, Having diene 55 now available in feasible amounts, a couple of methods present themselves for its transformation into 55. One route is presented in Figure 8 (p. “3). This pathway does not seem practical at this 53 point though, partly because of the inefficiency of re- moving and then regenerating the diene. A more direct approach would involve oxidation of the allyl carbons in 55 to 55. CH Ar Ar CH 2% 3 A number of of allylic carbons, presented in Figure CH 0 [0] Ar \/ Ar 0 CH W 3 methods exist for the direct oxidation and a few of the more useful ones are 12. CrO3°Py2 Figure 12. CH2Cl2 Allylic Oxidation Methods. 5h The presence of the methyl groups in 55 is rather beneficial. Most of the oxidative methods are basic in nature. It has been shown that the dienedione 55 is ex- tremely base sensitive with the bridgehead protons being 58 The bridgehead the apparent source of sensitivity. methyl groups should remove the base sensitivity and allow more vigorous oxidation of the system. A small amount of diene 55 was reacted with chrom- ium trioxide-pyridine in methylene chloride for four days, and a gummy solid was collected after work—up. The ir spectrum showed a rather broad and weak carbonyl absorption. The nmr spectrum indicated very little change in the ratio of vinyl to allyl protons. If allylic oxidation had taken place, it was only to a small extent. SUMMARY The purpose of this research was to lay the ground- work for the preparation of a 1,5-diarylcyclooctatetraene- 1,5-quinone. Once it had been decided that a bicyclo- [3.3.0] system was the most logical precursor, it became necessary to develop efficient synthetic manipulations of the system. The work presented here was an initial effort in the direction of preparing precursors to the quinone. With the knowledge gained regarding the various bicyclo- [3.3.0] systems, the ultimate preparation of a 1,5-diaryl- cyclooctatetraene-l,5-quinone seems within the realm of possibility. A schematic representation of major synthetic transformations is presented in Figure 13. The yields for each step are includedenui,where possible, the yields of reactions similar in nature are given in parentheses for comparison. Oxidation of the allylic sites in 55 seems to hold the greatest potential for the generation of the corres- ponding quinone precursor. 55 56 HO A AP I‘ (72%)6 O. (52%) 0 g Ar HO Ar $8 36 Big/:69%W AP AP <1IIII‘;% Ho.4‘llll.}_OH 192;; (ww‘lllliyzo AP AP $9 £3 £2 CH3 CH3 HO CH OH 96 a: 192., 0 Ar Ar ”(r (70%)76O Ar)“ CH3 C 3 CH3 25 33 31 50% (50%)76 % 20% 9/ CH3 CH HO OH 2% CH3 5% CH3 £2 CH3 a. 99% ___ (25%)76 b. 86% 1M% Figure 13. Summary of Bicyclo[3.3.0]octane Transformations. EXPERIMENTAL General procedures. Infrared spectra were record— ed on a Perkin—Elmer 137 spectrophotometer and were re— ferred to the 1602 cm"1 peak of polystyrene. The nmr spectra were obtained using a Varian T-6O and were taken at the ambient temperature using tetramethylsilane as an internal standard. The uv spectra were recorded on a Uni- cam Model SP-BOO spectrophotometer using 1 cm quarts cells. Mass spectrawere obtained with a Hitachi Perkin-Elmer RMU-6 mass spectrometer. Melting points were determined using a Thomas-Hoover Uni-Melt and were uncorrected. Tetramethyl hexane—1,31516-tetracarboxy1ate 55. Free radical induced coupling of dimethyl glutarate. This procedure, a slight modification of that used by Simpson,59 was developed in collaboration with Hagedorn58 and with the assistance of a number of other people. The reaction ves- sel was a 5R three-necked flask equipped with a mechanical stirrer, a constant addition funnel with nitrogen inlet, a thermometer, and a distillation apparatus for removal of t- butyl alcohol. The flask was charged with 2.52 of dimethyl glutarate (DMG), which was heated to boiling while nitrogen was passed through the apparatus. After fifteen minutes 57 58 boiling, the DMG was cooled to 175° and a mixture of di-t- butyl peroxide (Columbia, 400 ml, 23° 2.17 mole) and DMG (160 ml) was added to the vigorously stirred DMG. The ad- dition rate was 1 ml/min. and the temperature was maintain- ed at 170-175°. After addition was completed, heating was continued until gas evolution ceased; usually this required an additional hour. When no more gas evolved, the contents of the flask were rapidly heated to boiling until the DMG (bp 214°) began to distill over. The reaction was allowed to cool overnight and the mixture was transferred to a 5% round—bottom flask. Unre— acted DMG (3a. 21) was recovered by vacuum distillation through a 50 cm Vigerux column; everything boiling up to 1u00/17 mm was collected in one fraction. The residue in the pot was cooled to around 80° and was then transferred to a l£ round-bottom flask. Distillation was then contin- ued through the 50 cm Vigerux. Three fractions were col- lected-~fraction 1, 20g, bp 80-135"/0.02 mm; fraction 2, 290g, bp l35°/0.02mm; and fraction 3, 20g, bp 155—180°/ 0.03 mm. All of the fractions eventually solidified to give mushy white crystals. Further purification was a- chieved by recrystallization from methanol. Cooling to -10° and -60° was necessary for a second and third crop, respectively. The total recrystallized product was 236g. This corresponded to a yield of approximately ”6% based on the DMG consumed. 59 The recrystallized tetraester had a melting point of U8-59°, which was undoubtedly due to the presence of a mixture of diastereomers. The other data were: ir (Nujol), 17M0 cm-1 ; nmr (6 CClu); 3.70 (6H, s), 3.65 (6H, s), 2.68 (2H, broad t, J = 4 Hz), 2.50-2.06 (4H, m), 1.84 (UH, broad m). Bicyclo[3.3.0]octane-2,6-dione g5. Dieckmann cy- clization of tetraester 55 and hydrolysis-decarboxylation of the product, ketoester 55. A 1% three-necked flask was fitted with a mechanical stirrer, an addition funnel, a thermometer, and provisions for a nitrogen atmosphere. The apparatus was dried by flaming while evacuated and by flushing with dry nitrogen several times. Sodium methoxide (60g, 1.1 mole), freshly opened, and dry DMSO (22' 300 ml) were then added. Tetraester g5 (159g 0.50 mole), dissolved in 3a. 150 ml warm DMSO, was next added to the stirred slurry of sodium methoxide. Addition took fifteen minutes and resulted in an increase of temperature to 50-60°. An additional portion of sodium methoxide (UOg, 0.79 mole) was added after all the ester had been added. The deep orange reaction mixture was heated at 70-80° for ninety minutes. The by now nearly black reaction mixture was then cooled to 20-25° with an ice bath, and ice cold 6M hydrochloric acid (320 ml) was added slowly while stirring continued. Ice bath cooling was used to keep the temperature below 30°. As the acid was added, the mixture lightened in color, finally becoming pink, and a finely divided solid 60 precipitated towards the end of the addition. The final slurry was poured into 2.5% of ice water to which 9a. 50g of salt had been added. The solid was collected by filtra- tion through a large Buchner funnel, washed several times with cold water and once with cold methanol-water (1:1), and sucked as dry as possible. The crude bis(ketoester) was then dried and used without further purification for the hydrolysis and decarboxylation. The crude ketoester 55 weighed 109.5g (86% crude) and had a melting point of 91-98°. The entire amount of crude ketoester 55 was added to 300-“00 m1 of 6M hydrochloric acid in a 2% erlenmeyer flask. Several carborundum boiling chips were added and the mixture was warmed on the steam bath. The temperature was kept below 70° to control foaming. When gas evolution had ceased (one-two hours), the solution was cooled and filtered. The filtrate was extracted six times with 200 ml portions of chloroform. The chloroform extracts were com- bined and washed twice with 25 ml of 5M aqueous sodium hy- droxide and concentrated to EE- 500 ml. The chloroform solution was then filtered through a cone of Drierite. Re— moval of the solvent on the rotary evaporator gave a golden yellow oil which slowly solidified. The crude dione 55 was purified by sublimation (35-M0°/0.02 mm) onto a cold finger kept at 0°. The sublimed product was in the form of white blocky crystals, mp 95-h6°, with the following spectral 61 prOperties: ir (Nujol) 17U5 cm-l; nmr (6 CClu); 2.90 (2H, broad s); 2.20 (8H, broad s). 2,6-Bis(p-methoxyphenyl)-2,6-dihydroxybicyclo- [3.3.0]octane 55. Addition of p-methoxyphenylmagnesium bromide to dione 55. A 500 m1 three-necked round bottom flask was equipped with a mechanical stirrer, an addition funnel, a condenser, and a nitrogen inlet. The flask was flame dried and flushed with nitrogen. Magnesium ribbon (5.0g, 0.21 mole), cleaned with carborundum paper, was cut into small pieces, introduced into the flask, and covered with a 1:1 mixture of dry ether-benzene. A small portion of p-bormoanisole (27g, 0.1““ mole in #0 ml 1:1 ether-ben- zene) was added and stirring started. Once the reagent started to form, the remaining p-bromoanisole was added slowly. The reaction was allowed to stir one hour after addition was completed. Dione 55 (10g, 0.072 mole), dis- solved in 80 ml of 1:1 ether-benzene, was added slowly and one hour later 33. 100 ml of dry benzene was added and the reaction refluxed overnight. The reaction was then cooled and hydrolyzed with EE- 50 ml of saturated ammonium chlor- ide solution (five hours). The crude product was collected on a Buchner funnel, dried, and placed in a Soxhlet extrac- tion apparatus. Extracting for four days with ethyl ace- tate yielded 20.Ng (80%) of a white crystalline solid, mp 187-190° (decomposes): ir (KBr) 3150, 1600, 1500, 1240, 1 1030, 825 cm“ ; nmr (6 CDC13) 7.u5-6.80 (8H, AA'BB'), 3.81 62 (6H, s), 3.6 (2H, s, washed out with D20), 2.85—2.75 (2H, m), 2.35—2.10 (8H, m); m/e 35M (M+, weak) 336, 318, 18u, 173. 2,6—Bis(p-methoxyphenyl)bicyclo[3.3.0]octa-l,5- diene 31. Dehydration of diol $0 to conjugated diene 31. Diol 30 (2.12g, 6.0 x 10'3 mole) was placed in a 250 ml round bottom flask with ca. “0 m1 of benzene (0.15fl) and heated to the reflux point. Toluenesulfonic acid (0.05g, A%) was added and the reaction continued at reflux with stirring. The solution turned deep green after thirty minutes and was seeded at this point with a small amount of diene 31 obtained from previous dehydrations. After an additional two hours reflux, the reaction was cooled and ca. 10 ml of saturated sodium bicarbonate solution was added; the reaction was then stirred, cooled, and the yel- low-green solid collected. The solid product (l.h7g, 77%) had an mp 233-235°: ir (NuJol) 1610, 1h80, 1258, 1030, 833 cm'l; nmr (5, 00013) 7.50—6.90 (8H, AA'BB'), 3.9 (6H, s), 3.45-3.15 (AH, m), 3.10-2.70 (UH, m); uv (ethanol) Amax 350 (log 2 H.35u), 33A, 371 nm; m/3 310 (M+), 303, 159, 121, 78. 2,6-Bis(p-methoxyphenyl)bicyclo[3.3.0]octa-2,6- diene. Dehydration of diol g0 to nonconjugated diene $6. D101 30 (2.12g, 6.0 x 10"3 mole) was dissolved in 22- 300 ml of benzene (0.02M) in a 500 ml round bottom flask by heating to the reflux point. A crystal of toluenesulfonic 63 acid was added and the reflux continued overnight with stirring. The reaction became orange colored and contained some solid precipitate. The reaction was cooled, 3a. 10 ml of saturated sodium bicarbonate solution was added, and stirring continued. Further cooling promoted the formation of a solid material (diene 31) which was removed by filtra- tion. The benzene solution was separated from the bicarbo- nate solution, dried over sodium sulfate, and the benzene removed on the rotary evaporator. The product (1.5g, 78%) was a salmon colored solid material, mp l90-l93°: ir (NuJol) 1612, 1250, 1176, 826 cm‘l; nmr (5, 00013) 7.35- 6.65 (8H, AA'BB'), 5.85-5.75 (2H, m), 4.1-3.9 (2H, m), 3.8 (6H, s), 2.9-2.4 (m); uv (ethanol) Am 268 nm (log a 3X 4.53); m/e 318