THE svnmssm AND momenmwm OF 2,7,7-TRMHHYL-2, 4-CYCLOHEPTADIENONE AND 2.6,6,?-TETRAMETHYL-2. aucvemnamoemone Thes‘ss for the fiegsee of Ph. D‘ MECMGAN S’s-'ATE {INEVERSETY ANTHQN‘! F. NAPLES 1973 123 RA :2. : Mitrhigare SF. University x“ 'J’ \\ This is to pertify that the thesis entitled The Synthesis and Photochemistry of 2,7,7—Trimethyl—2,4- Cycloheptadienone and 2,6,6,7-Tetramethy1—2,4-Cyclohepta- dienone presented by Anthony F. Naples has been accepted towards fulfillment of the requirements for Ph . D . Chemistry degree in Major professor r-" l "Lu ABSTRACT THE SYNTHESIS AND PHOTOCHEMISTRY OF 2,7,7-TRIMETHYL-2,4-CYCLOHEPTADIENONE AND 2,6,6,7-TETRAMETHYL-2,4-CYCLOHEPTADIENONE By Anthony F. Naples This thesis describes the synthesis and photochemistry in various media of 2,7,7-trimethyl-2,4-cycloheptadienone (4%) and 2,6,6,7-tetra- methyl-2,4-cycloheptadienone (&3)‘ Compound lg was synthesized from cycloheptanone in approximately 10% overall yield. Irradiation in cyclohexane of the nn* band (350 nm) gave a trace amount of 1,3,3-trimethy1bicyclo[3.2.0]hept-6-en-2-one Lbaa) as the only volatile product. Irradiation through Pyrex in cyclohexane or TFE solution gave, in addition to kéa and its photoisomer 3,3,6-trimethy1bicyclo[3.2.l]hept-6-en-2-one ($ER)’ 2,5,5-trimethyl- 4-vinyl-2-cyclopentenone (kg) as the major product. Compound $8 arises from a rearrangement previously unobserved with any other cyclohepta- dienone. The excited state of $& from which this major product arises is undoubtedly nn*, since irradiation (>330 nm) of protonated lg (lgflkg) Ell . L. «h '0‘.“ Quit 1 Ad the [3. How tit): e A. Anthony F. Naples ’7 L/ G 8 b / 2 in FSO3H solution gave kRRRK as the exclusive photoproduct. Compound %3 was synthesized from eucarvone (4) in about 35% yield by reacting a with methyl iodide in THF using lithium bis(trimethyl- silyl) amide (LiHMDS) as the base. Preferential excitation of the nn* band (350 nm) of kg gave l,3,4,4-tetramethylbicyclo[3.2.0]hept-6— en-Z-one as the exclusive product. Photolysis of kg in cyclohexane using 3000 X light resulted in the formation of gi§_and trans - 2,5- dimethyl-4-(2-methy1—l-propenyl)-2-cyclopentenones %& and g3 respectively in addition to ZA; Compounds Z} and g; are clearly analogous to &8 and are considered to arise via a un* excited state of kg. When the solvent was changed to the highly polar 1,1,l-trif1uoroethanol (TFE), the photolysis of‘kg through Pyrex gave 2,4,7,7-tetramethylbicyclo- [4.1.0]hept-4-en—3-one (gg) as the major product, in addition to EAR zg, éé and secondary photolysis products. That gé also arises from a un* excited state of‘Lg was demonstrated by irradiation of.Lgd@: (in F803H), which can only react via a flfl* state. Under these conditions m: was the exclusive product. Compound %&_undergoes the familiar photochemical 1,3-acy1 shift when irradiated through Pyrex in TFE to give 3,4,4,6-tetramethylbicyclo- [3.2.0]hept-6-en-2-one Céfi) apparently via an nn* singlet excited state. However, irradiation of'gg in cyclohexane resulted in a triplet reac- 2,7 tion to 1,4,S,S-tetramethyltricyclo[4.l.0.0 ]heptan-3-one (3Q). THE SYNTHESIS AND PHOTOCHEMISTRY 0F 2,7,7-TRIMETHYL-2,4-CYCLOHEPTADIENONE AND 2,6,6,7-TETRAMETHYL-2,4-CYCLOHEPTADIENONE BY Anthony F? Naples A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1973 H; Na Jill ACKNOWLEDGMENT The author wishes to express his sincere appreciation to Professor Harold Hart for his enthusiasm, encouragement and guidance throughout the course of this study. Appreciation is extended to Michigan State University for a Graduate Teaching Assistantship from September, 1968 to June, 1971, and to the National Institutes of Health for financial support from June, 1971 to June, 1972 and for the June-September periods of years 1969, 1970, 1971, and 1973. ii EXPl TABLE OF CONTENTS THE SYNTHESIS AND PHOTOCHEMISTRY OF 2,7,7-TRIMETHYL-2,4-CYCLOHEPTADIENONE AND 2,6,6,7-TETRAMETHYL-2,4-CYCLOHEPTADIENONE Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 1 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . . . 8 A. The Synthesis and Irradiation of 2,7,7-Trimethyl- 2,4-cycloheptadienone (IA). . . . . . . . . . . . . . 8 B. The Synthesis and Irradiation of 2,6,6,7-Tetra- methyl-2,4-cycloheptadienone (AR) . . . . . . . . . . 17 C. The Structural Determination and Photochemistry of l, 3, 4, 4- Tetramethylbicyclo[3. 2. 0]hept- -6- ~en- 2- -one (71).. . . . . . 28 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . . . . . 36 General Procedures. . . . . . . . . . . . . . . . . . 36 General Photolysis Procedures . . . . . . . . . . . . 37 Synthesis of 2-Carbomethoxycycloheptanone . . . . . . 37 Synthesis of 2- -Carbomethoxy- -2- -methylcyclo- heptanone . . . . . . . . . . . . . 37 Synthesis of 2, 7, 7- -Trimethy1- -2- -carbomethoxy- cycloheptanone. . . . . . . . . . . . . . . . . 38 Synthesis of 2,7,7-Trimethylcycloheptanone. . . . . . 39 Synthesis of 2,7,7-Trimethyl-2-cycloheptenone . . . . 40 Synthesis of 2, 7, 7- -Trimethyl- -2, 4- -cyclohept- adienone (14). . . . . . . . . . . . . 40 Synthesis of 2, 6, 6, 7- -Tetramethy1- -2, 4- -cyclo- heptadienone (i9). . . . . . . . . . . . . . . . 41 Irradiation of Eucarvone (4) with >330 nm Light . . . . . . . . . . . . . . . . . . . . . . . 42 iii TABLE TABLE OF CONTENTS (Continued) Page Irradiation of 14 with >330 nm Light. . . . . . . . . 43 Photolysis of 14 in TFE . . . . . . . . . . . . . . . 43 Photolysis of IA in Cyclohexane . . . . . . . . . . . 44 Characterization of 1,3,3-Trimethylbicyclo- [3.2.0]hept-6-en-2-one (ARR)' . . . . . . . . . . . . 44 Characterization of 2,5,S-Trimethy1-4-vinyl- 2-cyclopentenone (1R) . . . . . . . . . . . . . . . . 45 Reaction of 46 with CH30Na in Methanol. . . . . . . . 46 Reaction of IR with CH 3ONa in Refluxing Methanol. . . . . . . . . . . . . . . . . . . . . . 47 Characterization of 3,3,6-Trimethylbicyclo- [3.2.0]hept-6-en-2-one (ARR). . . . . . . . . . . . . 48 Interconversion of 4R3 and 43R; . . . . . . . . . . . 48 Photolysis of AARNA.‘ . . . . . . . . . . . . . . . . 49 Protonation of 44 and TR in FSO H . . . . . . . . . . SO 3 Irradiation of 49 with >330 nm Light. . . . . . . . . 51 Reaction of 4A with Alkoxide Base . . . . . . . . . . 52 Reaction of 44 with LiHMDS in THF . . . . . . . . . . 52 Reaction of ZA."ith CH ONa in Methanol-d. . . . . . . 52 3 Synthesis of 44 . . . . . . . . . . . . . . . . . . . S3 Photolysis of 44 in Cyclohexane Using 3000 A Light . . . . . . . . . . . . . . . S4 Photolysis of 49 in Cyclohexane Using a 450 Watt Lamp . . . . . . . . . . . . . . . . . . 54 Characterization of Trans-2, S- -dimethy1-4- (2-methy1-l-—propenyl)-2-cyclopentenone (4%) . . . . . 55 Characterization of Cis- 2, S- -dimethy1- -4- (2- methy1-l-propeny1)-:2 cyclopentenone (44). . . . . . . 55 Reaction of £4 with CHSONa in Methanol .. . . . . . . 56 iv TABLE F3 FEP£ TABLE OF CONTENTS (Continued) Page Photolysis of RA in Cyclohexane . . . . . . . . . . . S7 Photolysis of (A in TFE . . . . . . . . . . . . . . . 57 Thermal Rearrangement of l,4,S,S-Tetra- methyltricyclo[4.1.0.02:7]heptan-3-one (36) . . . . . 58 Compound QR’. . . . . . . . . . . . . . . . . . . . . 59 Photolysis of A2 in TFE . . . . . . . . . . . . . . . 59 Photolysis of ARENA.“ . . . . . . . . . . . . . . . . 61 Quenched Photolysis of 44 . . . . . . . . . . . . . . 61 Methylation of £4 . . . . . . . . . . . . . . . - . . 63 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . 64 FIGURE 1. LIST OF FIGURES A comparison of observed (a) and Computer simulated (b) 100 MHz nmr resonances of the C4 and CS protons in 4E4 . . . . . An illustration of the chemical and nmr spectral evidence in support of the assignment of endo-C3 stereochemistry in {Ar . . Equations (with nmr data) illustrating the photo- chemical conversion of to 44 and to 44 and the thermal rearrangement o 44 to 44. . . . An illustration of the thermal conversion of 44 to via consecutive (025+023) pericyclic and con- rotatory (4q) electrocyclic reactions. vi Page II 29 32 34 te nu ca $ 4 . “if n ‘1’]! INTRODUCTION Highly polar media alter the spectra and normal photoisomerization path of 2,4-cyclohexadienones 431 Most conjugated cyclohexadienones show weak nn* absorption at about 350 nm and a much more intense nn* band at about 290-310 nm; when irradiated in ether, hexane, methanol or other common solvents, they produce ketenes,2’3 probably from the nw* singlet state.3 Depending on the solvent and the substitution pat- tern, the ketene may recyclize to the starting dienone l) react with a nucleophile to give unsaturated acids or their derivatives, or in the case of highly substituted ketenes (e.g. R2 —-R6 = CH3) cyclize via a . 4 . (fl4a+n2a) reaction to b1cyc10[3.1.0]hexenones 4. 0 R R R2 6 ,/ 0 6 unsat . hv, nfl* R2 Nu-H acid °r .' R . ’ ac1d deriv. A R R4 4 1 m 0 R6 1 hv, nn* s R6 R5 R R4 hv E 2 t -H+ chroz l I and Vers the flag fine H? on R . net: Dissolution of the dienones in trifluoroethanol (TFE) or adsorption on silica gel causes a large (10-40 nm, depending on substituents) batho- chromic shift of the fin* band. The band also broadens considerably, and because of its greater intensity it usually completely obscures the nn* absorption band. When irradiated under these conditions, the dienones isomerize directly to bicyclo[3.1.0]hexenones via the nu* state,1 rather than to ketenes. At least two explanations for this phenomenon are plausible. Since the nn* state is more polar than the ground state, whereas the reverse is true for the nn* state, polar solvents lower the energy of the former and raise that of the latter.5 Exceedingly polar solvents such as trifluoroethanol or the silica gel surface may reverse the relative energies of the two states. Whereas ring opening to a ketene occurs from the nfl* state,3 isomerization to the bicyclo[3.1.0]hexenone may be a flfl* singlet reaction of the dienone. If the nn* and nn* states of a molecule are not too far apart in energy (say 50 nm), such re- versals by highly polar media may be general, and useful in altering the customary photochemistry of such molecules. Alternatively, tri— fluoroethanol or silica gel may be functioning as proton donors and one may be witnessing a nu* photoisomerization of the protonated dienone dig; (a hydroxybenzenonium ion). Indeed, protonated dienones (e.g. R2-— R5 = H, R6 = CH3 or R2- R6 = CH3) in HSO3F solution which of necessity react by a flfl* state give only gang upon photolysis.6 The counterpart of the photoisomerization *7? could conceivably be realized from 2,4-cycloheptadienones, but in fact no example of the reaction gig is known. 2,4-Cycloheptadienones exhibit great é Bx variety in their photochemical behavior depending on the presence and location of substituents, the solvent and the particular excited state from which the reaction occurs. Eucarvone,4 was the first and most studied compound of this type, because of its accessibility from carvone.7 Its uv spectrum in cyclo- hexane has a fin* maximum at 298 nm (65500) and an nn* maximum at 350 nm (840). In polar solvents, the nn* band shifts to longer wavelengths as expected (303 nm in EtOH, 310 nm in TFE, and 318 nm for eucarvone 8 adsorbed on silica gel suspended in cyclohexane. In sulfuric acid, where the dienone is fully protonated, this maximum is shifted to 400 nm.9 Therefore it is not surprising that the product structure depends upon the solvent polarity. The photoisomerization 4144 studied initially by Buchi and Burgess10 is historically important, since it was the first example of a photolytic 1,3-acyl migration in B,Y-unsatu- rated ketones Q§§;§b). The photoisomerization to 44 is a relatively inefficient process (¢benzene .0025) which is thought to occur from O hv ———-. 99% EtOH as 'l 3..» both isomu .(7‘ 3‘1th- and both the singlet and the triplet states11 of 4. This type of photo- isomerization has been observed for several 2,4-cycloheptadienones: 0 O hv —lb OMe O 0 Me . o hv ‘Td-’ R f. 9 O methanol ( e l ) 0 R 0 hv . - ,, (Refs. 14, 19) CHZCI2 1 Q R=(H, CH3) A second type of photoisomerization, to the trimethyl 7-norbor- nenone 8’ was observed when eucarvone was irradiated in aqueous acid12 and in other acidic medias’g’ls. 0 0 hv ‘—t- + H it Although this type of product is formed in relatively minor amounts from eucarvone, it is the exclusive photoproduct of protonated 2,4- cycloheptadienone or its Z-methyl derivative QRHAII4 A th solv en-4 from able Err F) J‘ /§ 4. folmz CYClc ing s Cal-ho My b Prom. 81> ‘OH r hv, HSOKF -60° A third type of photoisomerism of eucarvone occurs in highly polar solvents, where a major primary product is the bicyclo[4.1.0]hept-2- en-4-one 44.8 Dehydrocamphor 4413 and the cyclobutanone 4g? arise from the photolysis of,LQ and have also been detected among the isol- able products of the photolysis of eucarvone. O I“) hv + TFE 5 or 4“ adsorbed on silica gel ’LQ Mechanisms Quenching and sensitization experimentsll’15 indicate that the formation of bicyclo[3.2.0]hex-6-en-2-ones such as 44 and 1 from 2,4- cycloheptadienones in neutral solvents probably arises from the lowest nn* singlet and triplet states. Orbital symmetry-allowed disrotatory ring closures could account for the products. The nfl* process 4jkg (and its protonated analog) is more interest- ing since it involves cleavage of a carbon-carbon bond beta to the carbonyl group. Since this pathway is favored by polar solvents, it may best be represented by dipolar intermediates (or in the case of the protonated ketones, carbonium ions). This interpretation accounts for the alt 1‘83 PIC Chi Pro The rec rin 5% Suc the the observed substituent effects more satisfactorily than does the alternative view of the reaction as an allowed [02a+n2a] concerted 16 reaction. If the nu* state is represented as the dipolar structure C, then an electr0cyclic ring Opening to R can lead to the observed 0- (m)- 0- (”U 0' (if) t} / -——¢- 1 -———Cu IIIIII r-—4- 40 ‘-- + Q R 1% product either directly or through the intermediate 4. Since Hine and Childs,9 in their irradiation of 4QU+, found the phenol 44 among the OH on OH teem—am t 56” R products after quenching, the stepwise path via E seems the more likely.17 The ion g'partitions between the paths which lead to 44 and‘ké. We recognized early17 that the fbrmation of 40 (or 4805:) through the ring opening of 4 might be aided by the presence of the gem dimethyl substituents on the carbon bearing positive charge in intermediates such as‘Q and E. It seemed possible thatlkg and related products were therefore not necessarily characteristic of all cycloheptadienones, but only of those with such substituents at C6. In fact, it has recently been observed14 that the absence of these substituents causes the reaction to take an alternate path. OH OH an: ‘3 The protonated dienone, reacting from a nu* state, undergoes a photo- chemically allowed ring closure20 of the pentadienyl cation Q4H+ to the allyl cation H; a thermal 1,2 alkyl shift leads to the observed protonated 7-norbornenone 9. In the case of eucarvone this path (lead- ing to Q) competes rather poorly with the ring opening (to Q). It also seemed to us that the ring opening reaction should occur equally well whether the gem dimethyl group is at C6 or C7 in the original dienone since the methyl groups can stabilize the positive charge by being at either end of the heptatrienyl cation. However, since the conversion of Q to E should be strongly disfavored if methyls are absent from C6, the ion corresponding to R, but having the methyls at the other end of the carbonium ion system might be expected to react in a different manner. To test these ideas, we synthesized and irradiated‘b417 and 4%.21 Our results are the subject of this thesis. 0 O "550 Eu 71% (4CYI ma; 301-p- VaVe RESULTS AND DISCUSSION A. The Synthesis and Irradiation of 2,7,7-Trimethyl-2,4-cyclohepta- dienone, 44. Compound.44 was synthesized from cycloheptanone in about 10°o over— all yield by the sequence shown. The dienone displayed an nn* band 0 EtO2 C. EtO 2c 5102 c [CH 51 NLi’ ——-> U 3); 12 Et 23,co CH31,THF, 25° 77. 5% 92. 5% 63.2% 1.92 1.07 (d, J: 2). Br /CCl NBS, C H 2 25 367gajggaajfi.0.:Collidile,.2'C°111dine’ 6.33. ~(d J_ 5) 150 9 76‘5° m centered at 5.97 Lt (Acyclohexane max 350 nm, 675) as a shoulder of the more intense nn* ab- sorption at 290 nm (65650). In TFE the fin* band was shifted to longer wavelength (302 nm, E4650), completely obscuring the nn* band. Irradiation of a .01 M solution of 44 in TFB through Pyrex with a Hanovia 450 watt mercury lamp gave a fast and clean (90%) conversion to three photoproducts, 444 (8.0%), 44 (82.5%), and 44R (9.5%). In cyclohexane, under identical photolysis conditions, the reaction gave 27% 444, 66% 44 and 7%,bfib. It is noteworthy that no a cleavage occurs with 44 despite the fact that one can envisage several plausible reaction 8 0r . Pro doll} The products from such a reaction path. The photoproducts and starting ° r' @hvh Inn Pyrex 15% (8. 0%) W(9 5°o) TFE O .4. -- o ,LQ (82.5%) dienone were stable to heat. Therefore, vapor phase chromatography (vPc) was used to monitor the reaction and to separate the products for analysis. 1,3,3-Trimethylbicyclo[3.2.0]hept-6-en-2-one, 15% had the shortest retention time of the four compounds and its mass spectrum (M=150) showed that it was an isomer of 1%. The ir spectrum (CC14) showed carbonyl absorption at 1722 cm-l, characteristic of a cyclopentanone. The nmr spectrum revealed three 3H singlets (61.03, 1.22, 1.27) having high and approximately equal europium shift numbers (methyls a to the carbonyl group). The two vinyl proton doublets (gf6.0).at 66.38 and 6.12 were characteristic of C6 and C7 protons respectively in a bi- cyclo[3.2.0]hept-6-en-2-one system.22 The C6 proton doublet had the smallest shift number in the spectrum and showed additional coupling Qlél) with the C5 proton. The CS proton appeared at 62.95 as a quartet or doublet of doublets due to coupling Q1=6.S,.g'=3.5) with the geminal protons at C4. Each peak of this quartet was split further into a doublet due to a small (£31) interaction with the C6 vinyl proton. The C4 geminal protons gave rise to a complex pattern of S apparent 10 lines between 61.60 and 62.10, which was not satisfactorily simplified by spin decoupling with the C5 proton. However, a computer simulation (Figure lb) of the ABX system formed by the C4 and C5 protons gave a close approximation of the observed (Figure 1a) 100 MHz spectrum. The simulated spectrum showed 8 lines for the C4 geminal protons. Two of these lines overlapped to give the peak of highest intensity. Further details are included in the experimental section of this thesis. 3,3,6-Trimethylbicyclo[3.2.0]hept-6-en-2—one, 15R had the longest retention time of the photoproducts. The mass spectrum (M=150) showed that it was an isomer of 1%. The ir spectrum (CC14) displayed a carbon- yl band at 1722 cm.1 (cyclopentanone). The nmr spectrum showed two 3H singlets of a gem dimethyl group (61.00 and 1.15) having high and approximately equal shift numbers (alpha to a carbonyl function). The third methyl appeared as a 3H multiplet at 61.75 (allylic) having the lowest shift number in the spectrum, which suggested its location at C6. The C1 alpha proton (63.15, m) had the largest shift number in the spectrum. Also observed was a 1H vinyl multiplet at 65.80, a 1H C5 proton signal (zq, £35) at 63.13, and a sharp doublet for the C3 protons which were unexpectedly equivalent and were coupled to the C5 proton (ng). Spin decoupling of a 100 MHz spectrum verified the interaction between the CS and C3 protons. Compound 15% is undoubtedly formed from 1% via the well known photoisomerization first reported for eucarvone, (4‘)10 and later for other 2,4-cycloheptadienones.14’18’19 In addition to the spectral data for 15R;and‘&ék further evidence for their assigned structures was obtained from their facile interconver- sion via a photochemical 1,3-acyl shift. Thus, irradiation of 15a in cyclohexane or TFE gave mixtures of 15a and 15k whose compositions (b) Figure l became of this (47% b: for a s a stead out by this re the wax stabilj not exa Has net 12 became constant after the times indicated in the experimental portion of this thesis. The photostationary state for the reaction in TFE (47% kék’ 53% lék) was established in about 1/3 the time necessary for a similar state to be reached in cyclohexane. In the latter case, a steady state ratio of 68% 15a to 32% lab was established. As pointed out by Bfichi, the relative amounts of the two photoisomers obtained in this reaction reflect merely the absorbances of the two compounds at the wavelength of incident light and not their relative thermodynamic stabilities. Therefore, although the uv spectra of lag and lab were not examined, the observed change in product ratio with solvent polarity was not unexpected. The exact nature of the excited state(s) of 1% from which 13a arises is still unclear. Attempts to isolate the nn* band by irradiat- ing 1& in cyclohexane through a Corning #3718 uranium glass filter (>330 nm) have thus far been only partially successful. Irradiation at such wavelengths would be expected to cause preferential excitation of the nu* band at 350 nm while eliminating reactions that result from un* excitation. The main reaction, however, appeared to be the formation of nonvolatile material. A single volatile product having the same retention time as lag did appear in the vpc chromatograms of the photolysis mixture. Its formation accounted for about 1% of the converted starting material. Similar experiments with grand its methyl analog kg were much more successful. Thus, only Ea andlél were obtained from a and 19 respectively when each was irradiated in cyclohexane with wavelengths longer than 330 nm (vide infra). These reactions were considerably slower than the corresponding reactions 13 o 1 0 R2 nfl* R1 --. R3 R2 R:5 R3. R3 15 R1=R2=CH3, R3=H m R1=R2=C113, R3=H 4’ R1=R2=H, R3=CH3 gm R1=R2=H, R3=CH3 1,2 121:“, R2=R3=CH3 a R1=H, R2=R3=CH3 through Pyrex due to the decreased light intensity. The photoproduct of 14 which deserves the greatest interest was 2,S,S-trimethyl-4-vinylcyclopent-Z-enone (16) which resulted from a rearrangement previously unobserved with any other cycloheptadienone. It had the second longest retention time of the photoproducts (appearing between.lé& and lab) and its mass spectrum (M=150) showed that it was also an isomer of k&. The ir spectrum (CC14) showed absorption at 1705, 1640 (cyclopentenone) and 930 cm.1 (terminal methylene). The uv spectrum (cyclohexane) with maxima at 344 nm (845), 330 (60), 317 (50) and 225 (9,370) was indicative of an a,B-unsaturated ketone. In the nmr spectrum, the gem dimethyl singlets appeared at 6.87 and 1.07 while the third methyl group was seen as a doublet of doublets at 61.75 due to coupling with the C3 (£31.8) and C4 (g;2.5) protons. The C4 proton resonance at 63.03 was a doublet (gf8.0) due to coupling with the adjacent proton of the C4 vinyl group. Each peak of this doublet was split further into a quintet as a result of interaction with the C3 vinyl proton (gf2.5) and with the protons of the methyl group at C2 (g;2.5). The low field position of the C3 vinyl proton 14 (66.98) suggested its location on the 8 carbon of an a,B-unsaturated ketone system. The remaining three vinyl protons of the C4 vinyl group gave rise to a complex group of signals between 64.60 and 66.00 corres- ponding to an ABC system further perturbed by the adjacent C4 proton. The splittings given above were verified by decoupling. Treatment of kg with base (a 5-fold excess of NaOCH in methanol) 3 caused isomerization to the fully conjugated dienones 123 and 17R; After three hours at room temperature, only AXE was formed but 25 hours at reflux gave a mixture of the two isomers with 11R predominating (52:48). 0 O 1.88 , 1.3 Cd» 93-1-5) ‘ \1'05‘5) (d, ._J_=1.5) ‘ )I-INS) ‘ \ 192 d J=7.0). 7.63 H H548 H ‘('— 1.87 ' — ' (m) 5.53 (q, zI_---7.0) (d, ._J_==7.0) W. m Geometry is tentatively assigned on the basis of observations that (a) the chemical shifts of the C3 vinyl protons differ as expected,23 and (b) interaction between the gem dimethyls at C5 with the ethylidene methyl should make big the thermodynamically more stable isomer. The mass spectra of the isomers (M=150) showed that they were isomeric with 16. ‘The ir spectra of the neat liquids (1698, 1610 cm'l) and uv spectra [Acyclohexane max 277 (22,600), 268 (16,400) for,L%% and 369 (60), 350 (94), 336 (100), 355 nm (860), 340 (90), 328 (90), 288 (14,300), 298 (8,100), 285 (12,800), 276 (11,550) for,Lzb] show that the exocyclic double bond in kg has moved into full conjugation in,yz. Vie V61: Prod I‘ema HOHe k110w] 15 Since its formation is favored in polar solvents, the vinylcyclo- pentenone lk'is thought to arise from a HH* state of AA: To test this possibility, Aflflfixn which can only react via a flfi* state, was irradiated in F803H at -78° using a 450 watt Hanovia mercury lamp with a Corning #3718 filter. Under these conditions, the long wavelength band of lflaflx,(lggio4 399 nm 64730) was preferentially excited, resulting in . . . + a clean isomerization to one product ARRRAK . 2. FsoxH, 78 1 hv, 450 w Hanovia, >330 nm br m 6.92 * H H 4‘ - + W M br s, 5.52 CH30H-l FSOSH CHSOH-l 1 F803}! ch03 -78° 1<2co3 -78° -73' -78° 44 19 The photoproduct was identified by its nmr spectrum (-46°) in the F803H photolysis mixture and by quenching the cation product in a methanol-carbonate suspension at -78°, which gave lR,in about 50% isolated yield (based on the amount of lA’used in the experiment). A 90% con- version of’lg'was observed upon quenching of the photolysis mixture after 6 hours of irradiation and subsequent examination of the crude product by nmr. The relative yield of,y§ was approximately 75%. The remaining 25% of the product mixture consisted of 17%, &@R and 18. However, the relative amounts of these minor products is not precisely known and further investigation of this matter is in progress. 1: C0 151 an l6 Reprotonation of 16 by the method described gave back lg&g;. The structure of 18 is yet unknown. However, 18 (like 17% and,b@b) is undoubtedly formed by an acid rearrangement of the photoproduct 16. To demonstrate this, 16 was protonated by the method described and kept at -78° for 7 hr. Quenching of this mixture, followed by nmr analysis of the recovered material, revealed a 70% conversion of $9 to 55.5% 17b, 25% ilfi and 19.3%,L8. It is noteworthy that this experiment, in contrast to the isomerization of 16 in base, gave,b@b as the predominant isomer of the 17& -,L36 pair. The results can be rationalized using an intermediate such as k (analogous to R). In contrast with Q, I o (H‘) o‘ (11*) 0(H*) / h + _T?:—:*# * \/ J.\_. _+ 00*) ! + will have the greatest positive charge density at the other end of the unsaturated system because of the different location of the gem dimethyl group. Cyclization to g,(analog of Rafi) is highly unfavorable because it would lead to a primary cation, whereas cyclopentenone formation encounters no difficulties. One can generalize these results thus far as follows. In reac- tions from the NH* state of 2,4-cycloheptadienones (or their protonated forms), substituents which stabilize a positive charge will favor pro The- s11, Has : Using obser 17 products of the type,lg if located at C6, and of the type.@g if located at C7. To explore the generality of this proposal, 7—methyleucarvone (2,6,6,7-tetramethyl-2,4-cycloheptadienone) was synthesized and ir- radiated. B. The Synthesis and Irradiation of 2,6,6,7-Tetramethyl-2,4-cyclo- heptadienone (L9). Compound 19 was prepared from eucarvone by the method shown below. 0‘ 0’ 4 LiHMDS —-u- CH 3 I 4eh- -————.. N THE" Hexane ‘— Reflux 0 1.03 (d, J=7.0) 0 1.87 - d, J=l.5 ( - ) H 2.47 (q,._J_=7.0) 6.33 ./I 1'02 4 (m) 1.08 5.75 9. 47% (35") (d. 1:50) 2,9 (65 ) The result of this methylation procedure, in which lithium bis(trimethyl- silyl)amide (henceforth referred to as LiHMDS) was utilized as the base, was in marked contrast to the results reported by Corey and Burke.7 Using a different base and solvent system, these workers failed to observe the formation of 19 from eucarvone. While the present work was in progress, a paper describing an alternate synthesis of 19 from euc dis MOI tow. tio: 1am; 0511 .2). b) C) 18 eucarvone appeared in the literature.24 cyclohexane displayed an nn* band ( x Like its analogs gland AA) AR. 350 nm €100) as a shoulder of its more intense 00* band (315 sh (84,500), 300 (6,200), 216 sh (5,900)). In TFE the 00* band was also shifted to longer wavelength (312 (5,600)) and completely obscured the nfi* band. In a nonpolar solvent such as cyclohexane, the photochemistry of,1€ displayed much the same behavior toward changes of incident light wavelength as did 14. Thus, irradia- tion of la in cyclohexane through Pyrex using a 450 watt Hanovia mercury lamp gave 91% 21, 3% %& and 6% 23, whereas a similar photolysis of 12 using a Rayonet reactor with 3000A lamps gave 38% 21) 8.3% ZK,and 48.7% 23. In each case, the reaction was stopped at 50-60% conversion of 19, 1%% hv CH3 0 H" CH3 4%. 91% 38% 8.3% 100% -- a) 450 w Hanovia, Pyrex O b) 3000 A, Pyrex c) 450 w Hanovia, >330 nm 48.7% 5 at which time the vpc fraction corresponding to the remaining 19 in the photolysis mixture was collected. The nmr spectrum of this fraction from hove (vid ing 30001 ph0t< comp< 19 from the 450 watt lamp photolysis showed that the material was pure kg. A similar spectrum of the fraction from the 3000A photolysis, however, showed that the sample contained a small amount (<5%) of 24 (vide infra). In addition, traces of the secondary photoproducts aris- ing from 24 were detected in the analytical vpc chromatogram of the 3000A photolysis mixture. In neither the 450 watt lamp nor 3000A photolysis did this starting material fraction show the presence of compound 35, a secondary photoproduct arising from 21 (vide infra). Since the relative yield of 21 was nearly doubled when the incident light was polychromatic above the Pyrex cutoff (2280 nm), an nn* reac- tion was implicated in the formation of this product. This supposition was verified upon irradiation of a cyclohexane solution of 12 with a 450 watt lamp through a Corning #3718 uranium glass filter, whence the exclusive product was 21 after 95% conversion of the dienone. By analogy, 21 is presumed to be formed via the same mechanism that gives rise to 15% and ARR? Only one of the two possible epimers of 21 was obtained from the photolyses of 19 or from direct synthesis of 21 from 5a. The tentatively assigned structure(s) for compound 21 shown on page 18 was arrived at by a circuitous route and resulted in some chemistry that has little to do with the main topic of this thesis. For this reason, the evidence for structure 21 and the related work are presented in Part C of this portion of the thesis. ‘1522372,5-dimethy1-4-(Z-methyl-l-propenyl)-2-cyclopentenone (2%) and gi§72,S-dimethyl-4-(2-methyl-l-propenyl)-2-cyclopentenone (23) are clearly analogous to %Q in structure and in the mechanism by which they are formed. The mass spectra of the two isomers showed (M=l64) that they were both isomeric with 19. Compound 22 displayed bands in 20 the infrared (neat) at 1700, 1665 (w) and 1640 (w) cm-1 and uv (cyclo- hexane) maxima at 356 nm (£16), 342 (40), 328 (52), 316 (48) and 219 (9,800) indicating a conjugated cyclopentenone with an a-alkyl substit- uent. The calculated wavelength for the most intense maximum was 217 nm with solvent correction. In the nmr spectrum, the C5 methyl doublet (g;7.5) at 61.18 had the highest shift number of the four methyl signals present. The three remaining methyls were allylic (chemical shift). The C2 methyl with a slightly smaller europium shift number appeared at 61.77 as a triplet (gf2) apparently through coupling with the C3 and C4 protons, while the butylidene methyl doublets at 61.72 had the lowest shift numbers in the spectrum and were coupled equally (g;1.5) with the butylidene proton. The C5 proton at 61.90 had the largest shift number in the spectrum and appeared as a quartet of doublets due to coupling with the protons of the C5 methyl (gf7.5) and with the C4 proton (g;3.0). The C4 proton at 63.12 had the appearance of a doublet (large coupling (g;9.5) with the butylidene proton) which was broadened by interactions with the C3, C5 and butylidene protons. Finally, the spectrum showed two vinyl protons, i.e. the butylidene proton doublet Q1;9.5) at 64.93 and the C3 vinyl proton multiplet at 66.98. The spectra of‘éé were quite similar to the above with ir (CC14) absorptions at 1700, 1665 (w) and 1640 (w) cm-1 and uv (cyclohexane) maxima at 358 (E16), 342 (42), 327 (68), 315 (116), 281 (200) and 219 (9000). As in 22, these data are consistent with a conjugated cyclo- pentenone with an a methyl substituent. The nmr spectrum showed a C5 methyl doublet having the highest shift number of the methyls. The remaining three methyls were allylic. The C2 methyl had a slightly smaller shift number than the one at C5 and appeared at 61.78 as a 21 triplet (g;1.5) due to coupling with the C3 and C4 protons, while the butylidene methyl doublets at 61.73 had the lowest shift numbers in the spectrum and were coupled equally (g;1.5) with the butylidene proton. The C5 proton, having the largest shift number in the spectrum, appeared at 62.43 as a quintet (gs7.5), with apparently equal interactions with the C4 and C5 methyl protons. The C4 proton signal had the appearance of a broad triplet (1:7-10). The spectrum showed two vinyl protons, i.e. the butylidene proton doublet (g;10.0) with each peak split into a septet (g;1.5) due to further coupling with the butylidene methyls and, lastly, the multiplet (£Fl.5) for the C3 proton at 66.92. The splittings for ,3 were verified by decoupling of a 100 M12 spectrum. Spectral and chemical evidence were used to assign the C5 geometry in the two epimers, 22 and 23. The C4H-C5H coupling constants (7.5 Hz for 23 and 3.0 for 22) are in agreement with the dihedral angles (20° and =120° respectively) observed in models of the two epimers. CH30Na°CH30H eib- In addition, 23 was quantitatively converted to 22 in sodium methoxide- methanol at 0°, using conditions which did not bring the double bonds into conjugation. This result is quite reasonable considering the ex- pected greater thermodynamic stability of 2%, Given the observations on eucarvone, it was not surprising that is Sta time 22 the methyl analog AR also displayed a marked alteration in photochemistry with increasing solvent polarity. Thus, when a .06 M solution of A2. in TFE was irradiated through Pyrex, the formation of 2,4,7,7-Tetra- methylbicyclo[4.1.0]hept-4-ene-3-one (24) which had been obtained only in trace amounts in cyclohexane now accounted for nearly 60% of the reaction products. In addition, the rate of disappearance of the start- ing material was greatly accelerated (relative to the cyclohexane re- action), giving an 85% conversion in 1 hour. The bicyclic enone Q4 450 w Hanovia h 1)- ’ + Others 4/2 a + a + a + [2,4, '4' CH3 P re , TFE o y x (28%) (5.7%) (7.5%) (40%) 36 0 2; (10.8%) is the methyl analog of 10. All photoproducts were separable from the starting material by preparative vpc except 23; Identical retention times were recorded for 19 and 24; however, pure samples of 24 were obtained for analysis from another route (vide infra). The ir spectrum of 24'showed carbonyl absorption at 1660 cm'1 and the uv (cyclohexane) spectrum had maxima at 346 nm (688), 317 sh (270), 268 (3,700) and 244 (4,660), consistent with an a,3-unsaturated ketone having additional conjugation with a cyclopropane ring. The mass spectrum (M=l64) showed that 24 was an isomer 0f.&3- The nmr spectrum revealed a single vinyl protoriat low field (66.72), probably at the end of an a,B-unsaturated ketone system. This signal appeared as a doublet Ql§5.5) of quartets (g;1.5) due to coupling with the C6 proton and allylic methyl protons 23 respectively. 0f the gem dimethyl singlets, the one at 61.22 had the lower shift number (lowest in the spectrum) and was assigned 353 stereochemistry. The gnd9_methyl singlet appeared at high field (6.80), being directly over the 0 system. The spectrum showed two additional methyl signals, one allylic (61.45, d, g;1.5) and the other at C2 (61.15, d, g;7.5) with gxg_stereochemistry. The C2 proton had the largest shift number and appeared as a quartet Q1;7.5) broadened by further slight coupling with the Cl proton. The signals of the two cyclopropyl protons, which could not be adequately shifted from the methyls, appeared in the 6.90-1.50 region of the spectrum. The egg geometry assigned to the C2 methyl group has not been verified by compari- son with the corresponding gnd9_isomer, since an attempt to epimerize as in base (ZS-fold excess of NaOCH in methanol)at 25° resulted in 3 rapid decomposition to tarry material. Examination of a model of‘24 in a preferred conformation (0 system approximately planar) reveals a dihedral angle of about 90° between the Cl and C2 protons. A model of the gndg_isomer of 24 shows this angle to be about 60° which would also give rise to a very small coupling constant for the C1-C2 proton interaction. However, in the gngg_model, the C2 and C7 gndg methyl groups experience serious crowding that would make this epimer an un- likely choice for the structure of 24; Although quantitative measurements were not performed, it is clear that 24 is photolytically labile with its maximum concentration occurring at about 60% conversion of 19 in the preceding photolysis. At 85% conversion of 19, 24 was itself about half converted to 229271232727‘ tetramethylbicyclo[2.2.1]hept-5-en-2-one géiand two minor products of shorter retention time. The minor products, formed in an approximately 1'1 ob he r85 MEI Prc C)’c Has 24 1:1 molar ratio, accounted for a total of 8% of the product yield of the TFE photolysis of 12 and were left unidentified. That 25 is a secondary product arising from 2435 (an example of the photochemical vinylcyclopropane-cyclopentene rearrangement)26 was confirmed by the observation that 25 was not formed in the photolysis of 19 in cyclo- hexane.27 Also, irradiation of a solution of 24 in TFE through Pyrex O H l 0 "°' CH h 3 v 11,, ‘aIIIII" 4- " TFE H A CH3 CH3 3X (80%) M. resulted in a rapid conversion to 25 (80%) and two minor products (210% each) having retention times identical to the minor products mentioned above in the TFE photolysis of 19. One of these unidentified products could be 26, resulting from the alternative vinylcyclopropane- cyclopentene rearrangement pathway. The stereochemistry at C3 in 25 was assigned with the assumption that this center (C2 in 24) is not altered during the transformation of 24 to 25. Although the ir (neat) carbonyl frequency of 25 (1730 cm'l) differed somewhat from the carbonyl absorption of its analog (lacking the C3 methyl group) reported by 13 cyclohexane 319 nm (8179), 307 max Schuster, et al., the uv spectrum (1 (256), 296 (230), 280 sh (180) and 214 (2,410)) was in good agreement. The structural assignment is based on the nmr spectrum and Eu(fod)3 shift data which appear in the experimental part of this thesis. 2.18 Que) of 1 Rep1 in; 25 It remained to be shown that Zfi,is formed via a 00* excited state of 19. This was demonstrated in the photolysis of protonated kg (lgxflz) which, like lgfluz_and éaflle can only react from its 00* state. Irradiation of the long wavelength band of'lggfl; [:3:04 405 nm (65,450)] resulted in a clean transformation to Zfiflkfi} the only species observable in the nmr spectrum (-46°) after 6 hours. 0H 0H . , J=7.5 _ 2.18(s) H 2 87 (q -— ) 2.08(S) H 3.35 (q, gr7.0) CH3 1.47 (d, £37.5) ,. ' 1.45 d, J=7.0 7.85(b 1.32 _, CH3 ( _ ) 8.58(d) 1.23 6.6S(br)(2H) + 4 + .87 kfififlm Zhflflm 1.53 CH30H-. FSOSH CH30H F503H O - 0 K2003 -78 xzco3 78 -78° -78° Quenching of the product cation gave pure 2% in good yield. The purity of the product was verified by its nmr spectrum and vpc chromatogram. Reprotonation of 24 gave zémflln Experiments showed $3 to be stable in FSOSH in the absence of irradiation for at least 48 hours at -78°. As noted previously, only one epimer (exo) of 24 was produced from 26 photolysis of 19 in TFE or cyclohexane. The same result was observed for the reaction in acid, where one would suspect epimerization of the ketone to be more likely. Using the mechanism (invoking intermediates M and Q) proposed for the formation of 24 (or EARNA) it must be concluded that path a in the Figure below is energetically favored over path b, and that structure 24 is indeed the exo form. H —* 8M: (exo) H3C on M + H3C —* 8803: (___end03 H In summary, 2,4-cycloheptadienones react photochemically by two paths depending on whether an n0* or 00* excited state is produced. The n0* state leads to bicyclo[3.2.0]hex-6-en-2-ones 28, possibly from both singlet and lowest triplet states. The 00* state Q (pos- sibly singlet) undergoes one of two electrocyclic processes, depend- ing on the substitution pattern. When neither R6 nor R7 are carbonium ion stabilizing groups (e.g. R6=R7=H), ring closure followed by an alkyl shift occurs, leading via P to a 7-norbornenone 29. Protonation may be required since such reactions have been observed only in acidic media. When either R6 or R7 can stabilize a positive charge, electro- cyclic ring opening competes favorably with this path to give 5%, which C}’( the 27 andégated prod ucls cyclizes to § if the stabilizing substituents are at C6 or to 31 if they are at C7. When substituents are present at both positions, the cyclization of Q is partitioned along both paths apparently according to the degree of substitution at each site. Compounds such as $4 with two methyls at C7 and one methyl, one hydrogen at C6 or one methyl, one hydrogen at both C7 and C6 might be expected to react accordingly, but this prediction remains to be tested experimentally. The reasons for the significant solvent effect on the partitioning of Q in the case of the cycloheptadienone $3 remain unclear. This question awaits further information on the relative energetics of the ('7 and 0f ' 0f cor £01 of in F1 of th ca a1 dL‘ 31 28 the competing cyclization processes. C. The Structural Determination and Photochemistry of 21 One expects a pair of epimeric structures for compound 21’(g§9_ and gndg_C3 methyl) to be formed from the electrocyclic ring closure of the butadiene moiety in A2? In fact, only one epimer was observed. Of all the photoproducts of 19, 21 had the shortest retention time. The mass spectrum (M=l64) showed that 21 was isomeric with La and the ir spectrum (1725 cm'1 for the neat liquid) and uv spectrum [Agiilohexane 320 nm (£108), 309 (156), 299 (133), 290 (94) and 210 (1,700)] were consistent with a bicyclo[3.2.0]hept-6-en-2-one structure.10 Except for the expected absence of n0* band fine structure, the uv spectrum of'gl in TFE 300 nm (£230), 219 (1,100) was similar to the spectrum in cyclohexane. Salient features of the nmr spectrum are shown in Figures 2 and 3 and the complete data appear in the experimental part of this thesis. Nmr data and the chemical evidence that follow are the basis for the assignment of £232 stereochemistry about the C3 carbon. It is noted that the two C3 proton signals in 5% (61.77 and 2.85) are over 1 ppm apart with the gngg_proton presumably at higher field due to its position over the 0 system of the cyclobutene moiety. Similar values would be expected for the C3 protons in the two epimeric ferns of'ak, The observed chemical shift (62.90) tends to favor an 252 orientation fOr the C3 proton and thus an gngg_methyl group. Although 21 underwent rapid exchange of the C3 proton in CH30D- CHSOHa at 25° (to give M), it was impossible to epimerize a in 29 refluxing NaOEt-ethanol or in LiHMDS-THF. In the latter case, enolate formation was assured. Only 21 was recovered, in quantitative yield, after quenching such solutions with cold water or dilute HCl. Figure 2 The greater thermodynamic stability of 21 relative to its epimer (in which a serious steric interaction between the Cl and C3 methyls is obvious) could be the determining factor in these observations. Arguments invoking a favored direction of protonation of enolate 3% seem unconvincing. From that standpoint, direction g_seems to be favored resulting in the formation of the presumably less stable, g§g_epimer. If the less stable epimer is formed first, then there must follow a rapid and complete isomerization to the observed product. In a related experiment, 21,was the only monomethylated product of the reaction of 5% with methyl iodide in base. (This constitutes an independent syn- thesis of'gl’and confirms its gross structure.) W1 ‘1 tC de ‘T pr 10 of sh 58 fr th fr« “'3 Va. tr: l7( of We Sis Intat; 30 When a .l M solution of 21 in TFE was irradiated through Pyrex, with a 450 watt Hanovia Lamp, an expected photostationary state between 21 and its photoisomer 34 was reached after 4 hours. At this time, Vpc and nmr analysis of the isomer mixture showed a 7:3 ratio 0f.%{ to 34. Compound 34 was epimerized to 3; (partially or completely, depending on column temperature) during all attempts to purify it by vpc. Thus, when the peak corresponding to 34 was collected from a preparative column and re-analyzed using an analytical column at a lower temperature, two peaks of very similar retention time (shoulders of the same peak) were observed. Nmr analysis of the collected fraction showed a mixture of 34 and 35. It was also observed that both 34 and 35'have approximately the same retention time as 19. For this reason, recovered starting material from the TFE and cyclohexane photolyses ofllg'was examined by nmr for the presence of 34 and 35, No evidence for these compounds was obtained from the spectrum, however. This indicates that 68 (and gg'in TFE) was absorbing most of the incident light. Irradiation of’gl in cyclohexane through Pyrex using either a 450 watt lamp of Rayonet 3000 A lamps resulted in its rapid and complete transfbrmation to 1,4,S,S-tetramethyltricyclo[4.1.0.02’7]heptan-3-one (36) (Figure 3). The ir Spectrum (0014) displayed a carbonyl band at 1705 cm'1 and the mass spectrum (M=l64) showed that it was an isomer of'gl. In the nmr spectrum, one of the four methyl signals present ap- peared at a considerably lower field (61.68, 5) than the others, con- sistent with a bridgehead methyl in a bicyclobutane system.28 This methyl signal also had the lowest shift number in the spectrum. The 31 remaining three methyl signals were the gem dimethyl singlets (6.73, 1.05) and the C4 methyl (6.80, d), the latter being coupled (gf7.3) with the C4 proton (61.85, q). Of the bicyclobutyl protons the one at C2 (62.47, d of d, gf4.3, gf=2.0) had the highest shift number (high- est in the spectrum) indicating a position a to the carbonyl function. The C7 proton appeared at 61.82 as a doublet of doublets due to coupling with both the C2 (gf2.0) and C6 (552.7) protons. Lastly, the C6 proton (62.19, d of d) was coupled to both C7 (gf2.7) and C2 (g;4.3) protons. The coupling through four bonds between the C6 and C2 protons is characteristic of nuclei arranged in a W-shaped configuration within a strained ring system such as bicyclobutane.29 Compound 36 may be considered a product of a di-0-methane rear- rangement of 21, since examples of such rearrangements are abundant in the photochemistry of B,Y-unsaturated ketones. A preliminary review of the literature30’31'32'33 indicates that this reaction probably proceeds through a triplet excited state (n0*) of the ketone, whereas the alternative pathway, a 1,3-acyl shift is probably a singlet reac- tion. The observation that gl’when irradiated through Pyrex gives only Qg'in cyclohexane but only QA'in TFE suggests that n0* triplet and n0* singlet states respectively are operative. A qualitative quenching study, in which 21 was irradiated in cyclohexane solutions containing increasing amounts of 21371.3—pentadiene, showed that the rate of formation of 36 was substantially decreased with increasing triplet quencher concentration. On the other hand, the formation of 33 in the TFE photolysis of'gl'was apparently unaffected by the presence of gisfl,3-pentadiene. Attempts to purify ég‘by preparative vpc resulted in its complete rearrangement to 3&,and 3§'(Figure 3) within A U 1 \ll’ ‘0. I AQ‘QV 1:4 [11. .4 J s 5" ‘1 a 1‘ .J I 0f f0 C01 32 the column (FFAP or Carbowax 20M on Chromosorb W at a minimum tempera- ture of 140°). Higher temperatures (l60-170°) gave only 35. Collection 1.23[2.4] .85[2.8](d,£fi7.5)CH3 0 CH3 [l.1](d,£f3) .1e—331———- 2.90[4.6] H cycl hexane (q, J=7,5) . H 6.40[1.0](d,g?3) 30° -' H 2.63 .73 1.7] [1.6] 1.07[1.1] hv :39 TFE %& Pyrex A .82[4.4] 1.87[4.9] 3-12 (d'i77'5) 2 92[s 7](m) , J=8 [4.81010 CH O . . ' (q ,__ ) H H s 87[2.1](mq) 3 H H 5.97[2.2]o»q) catalyzed? . . :” H" 1 15[§3f] 83 1 0 2.77[6.2] (é, J=é)\L\- : CH3 1' [ ' 1(m) (q’ g;7.5) H CH3 l.87[l.0](m) " H 2.77[2.0](m) .7 [2.2] . 98[1.5] 1.17[1.5]2'83[2 01‘“) 1.12[1.2] 35 ’VD % Figure 3 of the single fraction corresponding to the mixture of’ég and 35 was followed by nmr analysis which verified the presence of the two isomers. Compound 36 is apparently isomerized first to 34, which then undergoes .2... 33 (possibly catalyzed) epimerization to 35; This was demonstrated by heating 3R'at 145° in CCl4 solution in a sealed nmr tube. Nmr spectra taken at 15 min intervals showed a rapid transformation to 33, the only observable compound in the spectrum after 1 hour. The spectrum was identical to that of éflifOrmed (but not isolated) in the TFE photolysis of 2); Continued heating of the CCl4 solution at 145° resulted in complete conversion of 34,to 35 after 8 additional hours. At this time only 35 was observed (in addition to impurity peaks) in the nmr spectrum of the sample. Compound 35 was stable to further heating and to the vpc conditions mentioned above. Additional structural data for 3&,and‘3§ appear in the experimental part of this thesis. The preceding observations may be rationalized by the scheme illustrated below in Figure 4. However, although there is ample prece- 34 and uncatalyzed35 bicyclobutane rearrangements dent for catalyzed of this type, little can be said concerning the mechanism of the preceding example in the absence of a careful gas phase pyrolysis of 36. The relatively low temperature required above for its rearrange- ment suggests a catalyzed reaction. Compound 36 rapidly decomposed (or rearranged) in the presence of a small amount of Eu(fod)3 shift reagent in CCl4 to products with olefinic protons (signals similar in shape and chemical shift to those of'ég or éé)' If one considers the thermolysis of 36 as a concerted (025+oza) process, then a contradiction with the stereochemical assignment in 34 (and therefore in 2*) becomes apparent. Since 34 and 35 (the ob- served thermolysis products) both contain C6 (instead of C7) methyls, it was assumed that only path g_(illustrated below) of the two pos- sible (025+02a) processes is operative. Then only path d_of the two 34 + enantiomer + enantiomer Figure 4 possible conrotatory ring closures is operative (since the product formed, 34, has an nmr spectrum identical to the Bfichi photoisomer of 21 obtained in the TFE photolysis). Compound 34, arising from the 3S conrotatory closure, must then undergo epimerization (since that is what is observed above upon heating 34) to 35, hitherto presumed to be the less thermodynamically stable epimer of the 34735 pair. The limitations of nmr spectra in the assignment of C3 stereo- chemistry in gl'are well realized. It is possible that the compound thus far presumed to have structure 21 is actually the gxg_epimer illustrated below as 39. In that case, changes in structures 34 (to '40), 35 (to 41) and 36 (to 42) are necessary. EXPERIMENTAL General Procedures Except where otherwise noted, all nmr spectra were measured in CCl4 solutions with TMS as an internal standard. The 60 MHz spectra were recorded on a Varian T-60 or A56/60 spectrometer and the 100 MHz spectra were recorded on a Varian HAIOO spectrometer operated by Mr. Erich Roach. Simulated nmr spectra were obtained from a Nicolet In- strument Corporation 1082 Computer operated by Mr. Erich Roach. Spectra are recorded in units of delta. Numbers placed next to protons in struc- tures in the discussion section refer to chemical shifts of those protons. Numbers in brackets beside chemical shifts in the discussion and experi- mental sections are "europium shift numbers" obtained by adding small increments of tris(l,l,1,2,2,3,3,-heptaf1uoro-7,7-dimethy1-4,6-octane- dione)Eu(III) to the CCl4 solution being investigated. After each addition the nmr spectrum was scanned and the new frequency of each absorption was recorded. The shift for each absorption is the dif- ference between the frequency of the shifted absorption and the original one. Shift numbers are ratios obtained by dividing the shift of each signal in the spectrum by the shift of the least shifted signal. Infra- red spectra were recorded on a Unicam SP-200 spectrophotometer in units of cm'l. Ultraviolet spectra were recorded on a Unicam SP-800 spectro- photometer in cyclohexane, unless otherwise noted. Mass spectra were Obtained from a Hitachi-Perkin Elmer EMU-6 operated by Mrs. Ralph Guile. Elemental analyses were performed by Spang Microanalytical Laboratories, Ann Arbor, Michigan. 36 “M‘s“evflgnhfifir I 37 General Photolysis Procedures Solutions of the compounds to be irradiated were placed in septum- capped Pyrex tubes and purged of oxygen by bubbling dry, oxygen-free nitrogen through them for 30 minutes prior to photolysis. Irradiations were carried out with a 450 watt Hanovia Type L medium pressure mer- cury vapor lamp with the appropriate filter. The tubes were fastened to an immersion well apparatus which was immersed in water at ambient temperature. Alternatively, a Rayonet Photochemical Chamber Reactor or Type RS Preparative Photochemical Reactor was used. Photolyses were monitored by withdrawing small (330 nm Light. A solution of .4 g of‘g in 50 m1 of cyclohexane (.005 M) was purged with nitrogen and irradiated using a 450 watt lamp fitted with a Corning #3718 uranium 43 glass filter. After 48 hr, vpc analysis (5' x 1/8" 3% FFAP on Chromo- sorb W at 120°) revealed a 60% conversion of the starting material to a single photoproduct having a shorter retention time than 4. The product was separated for analysis by preparative vpc (10' x l/4" 20% FFAP on Chromosorb W at 140°) and identified as l,4,4-trimethylbicyclo[3.2.0]- hept—6-en-2—one. Ir (0014) 1725 cm'l; nmr (cc14) 6.97, 1.10, 1.24 (3H each, s, CH3), 1.78 (1H, d, £316, each peak split into a doublet, £31, endo-methylene proton), 2.85 (1H, d, £316, each peak split into a doublet, 5 ‘£31, exo-methylene proton), 2.60 (1H, s, C2 proton), 6.13 (1H, d, £33, C6 proton), 6.35 (1H, d, £33, C7 proton). Irradiation of 2,7,7-Trimethyl-2,4-cycloheptadienone (14) with >330 nm Light. A solution of .084 g of 14 in 22 ml of cyclohexane (.025 M) was purged with nitrogen in a 2 cm x 15 cm Pyrex test tube, which was then sealed and placed inside a 2.5 cm x 16 cm Corning #3718 uranium glass tube. This apparatus was suspended in a Rayonet Photo- chemical Reactor with 3500 A lamps and irradiated. The progress of the reaction was monitored by Vpc (5' x 1/8" 10% FFAP on DMCS Chromo- sorb W at 112°) by withdrawing lul samples at 12 hr intervals. After 48 hr the area of the starting material peak was halved. Only one volatile product was seen in the chromatogram and had a retention time identical with that of an authentic sample of,b%a. The volatile product peak was <5% of the peak corresponding to unconverted starting material. A similar result was obtained when a cyclohexane solution of'lg was irradiated with a uranium glass filtered 450 watt lamp. Photolysis of 2,7,7-Trimethylcyclohepta-2,4-dienone (6%) in TFE. A solution of .01 g of the dienone in 4 ml of TFE (1.4 x 10-2M) in a 44 10 x 100 mm Pyrex test tube was degassed and irradiated using a 450 watt lamp with a Pyrex filter. Reaction progress was monitored by vpc (5' x 1/8" 10% FFAP on Chromosorb W DMCS 80/100 at 110°). The N2 carrier gas flow rate was 30 ml/min. Samples of lul were injected in- to the vpc at 15 minute intervals. Approximately 90% conversion of starting material was observed at 45 minutes of irradiation. The vpc chromatogram consisted of four peaks: 4287(5'2 min), 68 (9.2 min), 15k'(11.8 min), 14 (15.6 min, corresponding to the starting material). Relative yields at t=lS min photolysis time, 50% conversion: ARE 14.8%, kg 86.0%, 13R 0.0%; t=30 min, 70% conversion: ARR;14'8%’ AR 82.3%, 15R,2.9%; t=45 min, 90% conversion: 15a 8.05%, 16 82.5%, 15R 9.45%. Evaporation of the solvent gave a colorless oil, which was separated into its components by preparative vpc (10' x 3/8" 20% FFAP on Chromo- sorb W 30/60 at 155°). The helium carrier gas flow rate was 90 mI/min. Retention times for the photoproducts were: 68% 6.7 min,,y6 12.0 min, 1,5); 19.8 min. Photolysis of 14 in Cyclohexane. The preceding experiment was repeated using cyclohexane as solvent. Approximately 60% conversion of starting material was observed after 5 hr of irradiation, at which time the relative yields of the photoproducts were 15g 27%, lg 66% and 1,5}; 7%. Characterization of 1,3,3-Trimethylbicyclo[3.2.0]hept-6-en-2- one (15%). The product was collected as a colorless liquid. Ir (CC14) 1722 cm'1 ; nmr (CC14) 61.03 [2.38], 1.22 [2.8] and 1.27 [2.18] (3H each, 5, Cl and C3 methyls), 2.95 ([2.8], 1H, d of d, £36.5, £j=3.5, each of the four peaks split into a doublet, £31, C5 proton), 6.12 45 ([1.5], 1H, d, £36.0, C7 vinyl proton), 6.38 ([l.0], 1H, d, £36.0, each peak split into a doublet, £31, C6 vinyl proton). In both 60 and 100 MHz spectra, the C4 geminal protons appeared as a complex 2H multi- plet (5 lines) between 61.60 and 62.10. Since the spacings between the peaks of this multiplet did not change significantly with succes- sive additions of Eu(fod)3 shift reagent, it was possible to assign a shift number (1.6) to the C4 protons. In decoupling experiments, ir- radiation at the C5 proton absorption frequency resulted in a new complex multiplet for the C4 protons instead of the expected doublet of doublets. Therefore, this ABX system (where A is the C4 endo-proton, presumably at highest field, B is the C4 exo-proton and X is the C5 proton) was simulated by computer using a trial-and-error method in which approximate values for J J J v(A) and v(B) were varied AB' AX’ BX’ until a close approximation of the observed 100 MHz spectrum was obtained. The small coupling between the C5 and C6 protons was neglected in this simulation. A good approximation of the observed spectrum (Figure la) was obtained using the values J =14 Hz, J =3Hz and J =7 Hz for the AB Ax BX coupling constants and v(A)=6l.76, v(B)=6l.82 and v(X)=62.92 as the chemical shifts. All lines were assigned an arbitrary linewidth of 1 Hz. No significant change in the simulated spectrum above was observed when JAB=14 Hz was replaced by JAB=-14 Hz. Mass spectrum (70 eV) n/_e 150 (M+), 107 (base). Anni. Calcd for C10H14O: C, 79.95; H, 9.32. Found: C, 79.78; H, 9.47. Characterization of 2,5,5-trimethy1-4-vinylcyclopent-2-enone, (18). The product was collected as a colorless liquid. Ir (CC14) 1705, 1640 em'l; uv (cyclohexane) 344 nm (£945), 330 (60), 317 (50), 46 225 (9,370); nmr (CC14) 60.87, 1.07 (s, 3H each, gem dimethyls at C5), 1.75 (3H, d of d, £31.8, £f=2.5, allylic methyl at C2 split by protons at C3 and C4 respectively), 3.03 (1H, d split into quintets, £38.0, £f=2.5 Hz, methine proton at C4 split by the adjacent proton of the C4 vinyl group and by the C2 methyl and C3 vinyl protons respectively), 4.6-6.0 (3H, m, protons of the C4 vinyl group), 6.98 (1H, m, C3 vinyl proton). Splittings were verified by decoupling of a 60 MHz spectrum. Thus, irradiation at the C2 methyl absorption frequency resulted in the appearance of the C4 proton signal as a doublet of doublets (£38.0, £f= 2.5) and the appearance of the C3 proton signal as a doublet (£32.5). Irradiation at the C3 proton absorption frequency resulted in the ap- pearance of the C2 methyl signal as a doublet (£32.5). Mass spectrum (70 eV) _m_/£ 150 (M+), 135 (base). .5211- Calcd for C H 0: C, 79.95; H, 9.32. Found: C, 79.89; 10 14 H, 9.36. Reaction of 16 with CH3ONa in Methanol. The vinylcyclopentenone 4 16 (.052 g, 3.4 x 10’ mol) in 5 m1 of methanol containing 5 equivalents of CHSONa was stirred at room temperature for three hours. Five m1 of water and 5 m1 of hexane were added, and the organic layer was ex- tracted with water until the washings were neutral to pH paper. The hexane layer was dried over MgSO4 and concentrated under reduced pres- sure to a colorless oil, which was then examined by vpc (5' x 1/8" 10% FFAP on Chromosorb W DMCS 80/100 at 110°). The NZ carrier gas flow rate was 30 ml/min. The chromatogram revealed the conversion of the starting material principally to gynfz,5,S-trimethyl-4-ethy1idene-Z- cyclopentenone (12%) having a retention time of 19.2 min. A trace amount (83%) of anti-2,5,5-trimethyl-4-ethylidene-2-cyclopentenone 47 CLLR) was also visible with a retention time of 17.4 min. Here nyn_ and nnii_refer to the orientation of the ethylidene methyl group with respect to the gem dimethyl group at C5. The main product was isolated for analysis using a 10' x 3/8" column of 20% FFAP on Chromosorb W at 170°. With the flow rate of helium carrier gas at 90 ml/min, its retention time was 34.0 min. Ir (neat) 1698, 1610 cm'l; uv (cyclohex- ane) 355 nm (860), 340 (90), 328 (90), 288 (14,300), 227 (22,600), 268 (16,400); nmr (CC14) 61.18 (6H, s, gem dimethyls at C5), 1.83 (3H, d, £3l.5 Hz, allylic methyl at C2), 1.92 (3H, d, £37.0, ethyli- dene methyl group), 5.58 (1H, q, £37.0, ethylidene vinyl proton), 7.10 (1H, m, vinyl proton at C3); mass spectrum (70 eV) nin 150 (M+), 107 (base). Anni, Calcd for C10H14O: C, 79.95; H, 9.32. Found: C, 79.92; H, 9.31. Reaction of kg with CH ONa in Refluxing Methanol. The previous 3 experiment was repeated, but at reflux temperature and for 25 hours total. Progress of the reaction was monitored by withdrawing aliquots of 50 pl, quenching in water, and extracting with hexane. The ex- tracts were then examined by analytical vpc under the conditions mentioned previously. A gradual increase in ilk relative to AXR;was observed, until after 15 hours the ratio was 41:69, and after 25 hours 52:48. There was an increasing loss of material with increasing length of reflux time and amount of base added. The refluxing mixtures were generally yellow due to an unknown impurity which was extracted into the aqueous layer during workup. Compound iZk was purified for analysis using the same conditions described for,Lbn. Due to the small dif- ference in retention times, it was necessary to pass 66R through the 48 column a second time after it was collected. The spectral data on 465% are as follows: Ir(neat) 1698, 1610 cm'l; uv (cyclohexane) 369 nm (5:60), 350 (94), 336 (100), 298 (8100), 285 (12,800), 276 (11,550); nmr (CC14) 61.05 (6H, s, gem dimethyls at C5), 1.88 (3H, d, £31.5, allylic methyl at C2), 1.87 (3H, d, £37.0, ethylidene methyl group), 5.48 (1H, q, £37.0, ethylidene vinyl proton), 7.63 (1H, m, vinyl proton at C3); mass spectrum (70 eV) n/n 150 (M+), 107 (base). Anni, Calcd for C H O: C, 79.95; H, 9.32. Found: C, 80.08; 10 14 H, 9.50. _ Q.- Emml“ AKJI' ". ‘ I Characterization of 3,3,6-Trimethylbicyclo[3.2.0]hept-6-en-2-one (15R). The product was isolated as a colorless liquid. Ir (CC14) l 1722, 1635 cm' ; nmr (CC14) 61.00 [3.07], 1.15 [3.0] (3H each, s, gem dimethyl at C3), 1.75 ([l.0], 3H, m, C7 methyl), 1.78 ([2.23], 2H, d, £35.0, C4 protons, which appear to be equivalent), 3.13 ([1.54], 1H, ”q, £35, each peak split into a doublet, £32, C5 proton), 3.15 ([4.77], 1H, m, Cl proton), 5.80 ([l.85], 1H, m, C6 proton). A spin decoupling experiment was performed to demonstrate that the C5 proton was coupled to the geminal C4 protons and that the two C4 protons were indeed equivalent and therefbre not coupled to each other. Thus, irradiation at the absorption frequency of the C5 proton in a Eu(fod)3 shifted spectrum (100 or 60 M12), in which the C4 proton resonance was shifted clear of the methyl resonances, resulted in the collapse of the sharp CS proton doublet into a broad singlet. Mass spectrum (70 eV) m/e 150 (M’), 107 (base). Anal. Calcd. for 010H140: C, 79.95; H, 9.32. Found: C, 79.88; H, 9.33. Interconversion of ién and iék. A solution of .002 g of 18% in 2 .4 ml of TFE (3.3 x 10' M) in a 5 x 50 mm Pyrex tube was degassed and 49 irradiated using a 450 watt lamp with a Pyrex filter. Reaction progress was monitored by vpc (5' x 1/8" 10% FFAP on Chromosorb W DMCS 80/100 at 110°). A gradual reduction in the area of the peak corresponding to ign, and a simultaneous increase in the peak corresponding to the sole photoproduct 66R was observed. Photoequilibrium was apparently established after about 2 hr, giving a mixture of 47% ikfi;and 53% iék. The solvent was removed under reduced pressure yielding a clear liquid residue. An nmr spectrum confirmed the presence of the two compounds in the indicated ratio. Photolysis of a solution of .05 g of ién in 2 ml of cyclohexane using a Rayonet Photochemical Reactor with 3000 A lamps gave a mixture of 68% ién and 32% 68R after 7 hours. The isomers were easily separated using a 5' x 3/8" column of 20% FFAP on 30/60 Chromosorb W at 160°. With the helium carrier gas flow rate at 60 ml/min the retention times were 5.5 and 10.7 min respectively. Photolysis of AARNXK Under a nitrogen atmosphere, 5 m1 of spectro- scopic grade HSOSF was placed in a 10 ml round-bottomed flask, cooled 3 to -78°. With vigorous magnetic stirring, .4 g (2.6 x 10' mol) of kg was then added through a septum top via a 100 pl syringe. The bright yellow solution was quickly transferred to five smooth-walled nmr tubes which had been previously flushed with nitrogen and cooled 4 as an internal standard and the photolysis was monitored by nmr. Spectra to -78°. About .2 g of (CH3)4N+BF was placed in one of the tubes at -46° were taken at 2 hr intervals using a Varian AS6/6O spectrometer. Cation igRKQLdisplayed signals at 61.43 (6H, s, gem dimethyl) 2.28 (3H, 5, C2 methyl), 2.82 (2H, d, £35.5, C6 protons), 6.92 (2H, m, C4 and C5 protons), 7.90 (1H, d, £37.0, C3 proton). The tubes were fasten- ed to a quartz immersion photolysis well which was then immersed in a 50 Dewar jar containing a dry ice-isopropyl alcohol mixture. Absolute ethanol at ~78° was circulated as a coolant. The irradiation was carried out using a 450 watt lamp with a Corning #3718 uranium glass filter. The signals corresponding to 135”; gradually decreased in intensity during the photolysis, to be replaced by those of ARKRAH 61.28 and 1.45 (3H each, s, gem dimethyl), 2.13 (3H, 5, C2 methyl), 3.67 (1H, broad, C4 proton), 5.52 (3H, broad s, ethylidene group protons), 8.84 (1H, broad 5, C3 proton). Conversion of lgfifig'was complete in h- 6 hours. The products were recovered by transferring the acid solu— E tion to a jacketed dropping funnel, previously cooled to -78° and under a nitrogen atmosphere. The solution was then added dropwise to a vigorously stirred suspension of 10 g of powdered K CO in 75 ml of 2 3 methanol at -78°. The mixture was then brought to room temperature, 25 ml of hexane was added and the organic layer was extracted with saturated NaCl solution until the washings were neutral. The organic layer was dried over MgSO4 and concentrated to a colorless oil under reduced pressure. The crude product was examined by nmr which indi- cated a 75% yield of $6. The remaining 25% of the product mixture con- sisted of lXR, lXR'and 18, Since each of the products displays a vinyl proton resonance in the region between 66.98 and 67.63, the yield of the major product lR'was readily determined relative to the minor products. However, the relative yields of each of the three minor products are not yet precisely known and further investigation of this matter is in progress. The Protonation of lfl’and 18.1“ FSO H. A .1 g sample of each of 3 the compounds lflland lR'was dissolved in 1 ml of spectral grade FSOSH using the method described above, then transferred to nmr tubes (-78°) SI 4 as internal nmr standard. The sample of $4 in FSO3H displayed an nmr spectrum (-46°) identical containing a small amount of (CH3)4N+BF to the spectrum given above fer AARUL' This spectrum was unchanged after 48 hr, during which time the sample was kept at -78°. Subsequent quenching of the cation by the procedure given above gave approximately 75 mg of recovered iA, The sample of $6 in FSOSH displayed an nmr spectrum identical to the spectrum observed fer ARE”; formed in the photolysis 0f.k$ in FSOSH. The sample of $6 in FSOSH was kept at 1.45.; 2.190.!" -78° for 7 hr, after which time it was quenched according to the pro- cedure given above. Nmr and vpc analysis of the recovered material showed a 70% conversion of kg to 55.5% igb, 25% ién and 19.3% $8. Irradiation of 2,6,6,7-Tetramethyl-2,4-cycloheptadienone (18) Using >330 nm Ligni, A solution of .l g of the dienone in 10 ml of cyclohexane (6 x 10'2 M) in a 4" Pyrex teSt tube was degassed and ir- radiated using a 450 watt lamp with a Corning #3718 "uranium glass" filter. Reaction progress was monitored by vpc (5' x 1/8" 3% FFAP on Chromosorb W 80/100 at 120°). The nitrogen carrier gas flow rate was 30 ml/min. Vpc chromatograms at regular intervals revealed a gradual decrease of the peak corresponding to the starting material (3.4 min) and a simultaneous increase in a single photoproduct 2A (7.6 min). Conversion of the starting material was about 90% after 48 hours and the yield of 2A was 100% and no secondary photoproducts. The product was purified for analysis using a 10' x 1/4" 20% FFAP on 30/60 Chromosorb W column at 180°. Ir (neat) 1725 cm-1 ; nmr (CC14) 6.73 [1.7], 1.07 [1.1] (3H each, s, gem dimethyl at C4), .85 ([2.8], 3H d, £37.5, C3 methyl), 1.23 ([2.4], 3H, 5, Cl methyl), 2.63 ([l.6], 1H, s, C5 proton), 2.90 ([4.6], 1H, q, £37.5 Hz, C3 proton), 6.23 52 [1.1] (1H, d, £33 Hz, C6 proton), 6.40 [1.0] (1H, d, £33 Hz, C7 proton); mass spectrum (70 eV) 311/3 164 (M+), 122 (base), 66 (99%). Anni, Calcd for C H 0: C, 80.43; H, 9.82. Found: C, 80.52; ll 16 H, 9.88. Reaction of 2A with Alkoxide Base. A mixture of .1 g of 2A, 4 (6 x 10' mol) and 5 equivalents of sodium methoxide in 5 ml of meth- anol was stirred at room temperature for 5 hours. The reaction was a: then quenched by adding 5 ml of water and 5 m1 of ether. The ether layer was washed with water until the washings were neutral, dried s over MgSO4, and evaporated to an oil which was found to be pure 2k unchanged. An identical result was obtained when the reaction was carried out in refluxing methanol or refluxing ethanol-sodium ethoxide. Reaction of 2A with LiHMDS in THF. One ml of 1.6 M n-butyllithium was added to .5 ml of HMDS at 0°, under a nitrogen atmosphere. Five ml of THF was then added and the solution was allowed to come to room 3 temperature. Compound 2} (.2 g, 1.2 x 10' mol) was then added through a septum top dropwise with stirring. After 3 hours, the reaction was worked up by adding 10 ml of cold, dilute HCl and 5 ml of ether. The ether layer was washed three times with water, once with NaHCO3 solu- tion, then dried over MgSO4. Evaporation of the ether yielded pure unchan ed. 2,}, 3 Reaction of 2* with CHSONa in CH 0D. A mixture of .l g of %i and 3 5 equivalents of sodium methoxide in 5 ml of methanol-d was stirred at room temperature for 2 hours.' The reaction was worked up by adding S3 2 ml of D20 and 5 m1 of hexane, then shaking vigorously. The aqueous layer was discarded and the organic layer was washed with ice water until the washings were neutral. The hexane layer was dried over MgSO4 and evaporated to give pure ZARA: The product had an identical vpc retention time and carbonyl ir absorption frequency as 2*, The nmr spectrum of 2A5R>showed a 3H singlet at 6.85 instead of a doublet, and the absence of the quartet at 62.90. In all other respects, it was identical to the spectrum of'ZA, Mass spectrum (70 eV) nin 165 (M+), 66 (base), 122 (99.4% of base). Synthesis of 2A, The LiHMDS base was prepared as outlined above 6 using 1.2 m1 of n-butyllithium, and 1 ml of HMDS in 5 m1 of THF. To this solution, .19 g (1.3 x 10'3 mol) of 1,4,4-trimethylbicyclo[3.2.0] hept-6-en-2-one (5R) was added through a septum cap dropwise with stirring. This solution was allowed to stir at room temperature under a nitrogen atmosphere for 15 minutes, after which 1 m1 of methyl iodide was added. The mixture was then refluxed for 30 minutes, cooled and worked up by the procedure outlined above. The product was a color- less oil which was examined by vpc (S' x 1/8" 3% FFAP on Chromosorb W 80/100 at 112°. The nitrogen carrier gas flow rate was 30 ml/min. The chromatogram revealed a 95% conversion of the starting material to compounds 2i (87.5%) and 33 (12.5%) at retention times of 4.8 min and 6.8 min respectively. The fermation of'gi'was verified by compari- son of spectra with the sample obtained previously from the photolysis of 68' Separation of the products for analysis was accomplished easily using a 10' x 1/4" 20% Carbowax 20M on 30/60 Chromosorb W column at 150°. Compound 33 was isolated as a colorless oil which solidified on standing (mp 37-39°). Ir (0014) 1723 cm'l; nmr (c014) 6.83 ([3.01, 54 3H, 5, C3 nn£n_methyl), .88 ([1.7], 3H, 5, C4 nn£n_methyl), 1.00 ([l.3], 3H, 5, C4 gnn_methyl), 1.18 [2.4] and 1.27 [2.9] (3H each, 5, Cl and C3 nng_methyls), 2.62 ([1.7], 1H, s, C5 proton), 6.17 ([1.7], 1H, d, £33 Hz, C7 proton), 6.45 ([1.0], 1H, d, £33 Hz, C6 proton); mass spectrum (70 eV) M9. 178 (0*), 66 (base). Photolysis of kg in Cyclohexane Using 3000 A Light. A solution of .2 g of AR in 40 m1 of cyclohexane (.028 M) in a 13 mm x 600 mm Py- rex tube was purged with nitrogen and irradiated in a Rayonet Type RS preparative photochemical reactor fitted with 3000 A lamps. Reaction progress was monitored by vpc (5' x 1/8" 15% FFAP on 80/100 Chromosorb W DMCS at 110°). The NZ carrier gas flow rate was 30 ml/min. Three peaks were observed, corresponding to 2A (8.2 min), 22 (24.2 min), and 23 (31.8 min) in addition to the starting material peak (19.8 min). Relative yields after 10 hours of photolysis were 54.8%, 12.1% and 33.0% respectively. After 15 hours the yields were 51%, 10.2% and 38.8% respectively. A 60% conversion of the starting material was observed after 20 hours, at which time the yields were 50%, 9.4% and 40.6% respectively. The products were separated for analysis using a preparative vpc column (10' x 3/8" 20% FFAP on 30/60 Chromosorb W at 170°). With the helium carrier gas flow rate at 90 ml/min the reten- tion times were 9.1 min for 2k, 22 min for zg'and 26.8 min for 23. The starting material had a retention time of 19.8 min. Compound 2i was identified by comparison with samples obtained previously. Photolysis of 63 in Cyclohexane Using a 450 Watt Lamp. A solu- tion of .026 g 0f.L2 in'7 ml of cyclohexane (.02 M) in a 12 mm x 200 mm Pyrex tube was purged with nitrogen and irradiated using a 450 55 watt lamp fitted with a Pyrex filter. A 50% conversion of the start- ing material was observed after 3 hours, at which time the relative yields of the photoproducts were 91% 2k, 3% 22 and 6%‘23. The products and remaining starting material were collected by preparative vpc. Examination of the starting material fraction by nmr showed that it contained only,L9. Characterization of 22, The product was isolated as a colorless s1 liquid. Ir (neat) 1700, 1665 (w), 1640 (w) cm'l; uv (cyclohexane) 356 nm (€=16) 342 (40), 328 (52), 316 (48), 219 (9800); nmr (CC14) E 61.18 ([5.9], 3H, d, £37.5, C5 methyl), 1.72 ([1.0], 3H, d, £31.5, butylidene methyl), 1.72 ([1.5], 3H, d, £31.5, butylidene methyl), 1.77 ([6.2], 3H, t, £32, C2 allylic methyl), 1.90 ([10.7], 1H, q, £3 7.5, each peak split into a doublet, £f=3.0, C5 proton), 3.12 ([5.1], 1H, d, £=9.S each peak split into a broad multiplet, C4 proton), 4.93 ([2.9], 1H, d, £39.5, each peak split into a broad multiplet, butyli- dene proton), 6.98 ([3.9], 1H, m, C3 proton); mass spectrum (70 eV) nin_l64 (M+), 122 (base). Anni, Calcd for C H O: C, 80.43; H, 9.82. Found: C, 80.22; 11 16 H, 9.89. Characterization of 23, The product was isolated as a colorless liquid. Ir (0014) 1700, 1665 (w), 1640 (w) cm'l; uv (cyclohexane) 358 (£316), 342 (42), 327 (68), 315 (116), 281 (200), 219 (9000); nmr (C01,) 6.93 ([5.8], 3H, d, g;7.5, cs methyl), 1.73 ([1.2], 3H, d, 331.5, butylidene methyl), 1.73 ([1.0], 3H, d, £31.S, butylidene methyl), 1.78 ([5.6], 3H, 5,‘£3l.5, C2 allylic methyl), 2.43 ([10.0], 1H, quintet, £37.S, C5 proton), 3.68 ([4.3], 1H, broad triplet, £37-10, C4 proton), 5.23 ((3.29], 1H, d, £310.0, each peak split into a septet, £f=l.5, 56 butylidene proton), 6.92 ([3.94], 1H, m, £31.5, C3 proton). Spin de- coupling of a 100 M12 spectrum (irradiation at the frequency of the three allylic methyls) confirmed the following interactions: 1. The C3 proton signal appeared as a doublet, £33.0. Therefore, the C3 pro- ton is coupled to the C4 proton, £33.0 and to the C2 methyl, £31.5. 2. The butylidene proton signal appeared as a sharp doublet, £310.0. Therefore the butylidene proton is coupled to the C4 proton, £310.0 and to the allylic gem methyls, £31.5. 3. The C4 proton signal ap- peared as a pair of quartets (pair of doublets of doublets) showing the C4 proton to be coupled to the butylidene proton, £310.0, to the C5 proton, £f=7.5 and to the C3 proton, £f=3.0. In the absence of de- coupling, the C4 proton is also coupled to the allylic gem methyls which gives rise to its broad triplet appearance mentioned above. Mass spectrum (70 eV) 164 (M+), 122 (base). Anni, Calcd for C H O: C, 80.43; H, 9.82. Found: C, 80.38; 11 16 H, 9.93. Reaction of 23 with CH ONa in CH OH. Compound 23 (.05 g) was 3 3 added to a solution of five equivalents of sodium methoxide in 5 ml of methanol under a nitrogen atmosphere at 0°. After stirring at 0° for 15 min the reaction was quenched by addition of 5 ml of water and 5 ml of hexane. The organic layer was washed with water until the washings were neutral, then dried over MgSO4 and concentrated to a colorless oil. Examination of the mixture by vpc revealed a 90% con- version of 23 to a single product having the same retention time as 22. The product was isolated for analysis using a preparative 5' x 3/8" FFAP column at 150°. Nmr and ir analysis of the product showed it to be identical to 22 obtained from the photolysis of’ig. 57 Photolysis of 4A.i" Cyclohexane. A solution of .l g of 4A.in 6 ml of cyclohexane (.lM) in a 6 mm x 40 mm Pyrex tube was purged with nitrogen and irradiated in a Rayonet Type RS preparative photochemical reactor fitted with 3000 A lamps. Reaction progress was monitored by analytical vpc under conditions identical to those used in the photolysis of i2 except for an injector temperature of 140°. The chromatograms revealed the complete conversion of &i to a single photoproduct (QR) after about 4 hours. The retention time of 36 was then compared to that of kg and found to be identical. Removal of the solvent under reduced pressure yielded the clear oily photoproduct which required no purification. Ir (0014) 1705 cm’l; nmr (0014) 6.73 [2.3] and 1.05 [1.7] (3H each, s, gem dimethyl), .80 ([4.4], 3H, d, £37.3, C4 methyl), 1.68 ([1.0], 3H, 5, Cl methyl), 1.82 ([2.3], 1H, doublet of doublets, £32.7, £f=2.0, C7 proton), 1.85 ([5.7], 1H, q, £37.3, C4 proton), 2.13 ([l.9], 1H, doublet of doublets, £34.3, £f=2.7, C6 proton), 2.47 ([6.3], 1H, doublet of doublets, £34.3, £f=2.0). The coupling constants were obtained by careful analysis of a 100 MHz spectrum. The compound re- arranged rapidly in the presence of Eu(fod)3 shift reagent with a half life of approximately four minutes. However, reasonably good results were obtained by working quickly. Mass spectrum (70 eV) nin 164 (M+), 65 (base), 122 (76% of base), 123 (54% of base). Anal. Calcd for C O: C, 80.43; H, 9.82. Found: C, 80.37; ll“16 H, 9.92. Photolysis of 2) in TFE. A solution of .05 g of 2) in 3 m1 of TFE in a 6 mm x 12 mm Pyrex tube was purged with nitrogen and irradiat- ed using a 450 watt lamp with a Pyrex filter. A similar sample with cyclohexane as the solvent was irradiated simultaneously for compari- son. Reaction progress was monitored by analytical vPc under the same 58 conditions used in the preceding experiment. A chromatogram of each reaction mixture showed the formation of a single product in each case. Retention times for the photoproducts were identical to each other and, fortuitously, to the retention time of *2. After three hours the TFE sample contained 78% 21 and 22% 34. Irradiation for an additional hour resulted in a photoequilibrium mixture of 70% 21 to 30% 34. This ratio did not change upon further irradiation. The cyclohexane sample con- ’I-‘HF A—l tained a mixture of 28% thand 72%‘36 after 3 hours of irradiation. Examination of the photolysis mixtures by nmr confirmed the ratios 1 given above. Compound 34 was not isolated by vpc, since it underwent (partial to complete) epimerization to 35 under all conditions that were tried. Instead, it was prepared for spectral (nmr and ir) analysis by thermal rearrangement of 36, Thermal Rearrangement of 36. A solution of .05 g of 36 in .3 m1 of CCl4 was sealed in an nmr tube and placed in an oil bath maintained at 145°. Reaction progress was monitored by nmr at 15 min intervals. A gradual decrease in the signals corresponding to 36 was accompanied by a simultaneous increase in the signals of 34. A complete conver- sion of the starting material to 34, was observed after 1 hour. No other products were detected in the spectrum. The spectrum of the product was identical to the spectrum of the photoisomer of 21 obtained from the irradiation in TFE described above. Compound 34 had a car- bonyl stretching frequency in the infrared (CC14) of 1725,1635 (w) cm'l;. Nmr (CC14) 6.98 [1.5] and 1.12 [1.2] (3H each, s, gem dimethyl), 1.15 ([3.1], 3H, d, £78.0, C3 methyl), 1.83 ([1.0], 3H, m, C7 methyl), 1~87([4.9], 1H, q, gf8.0, C3 proton), 2.77 ([2.0], 1H, m, C5 proton), 3.12 ([4.8], 1H, m, C1 proton), 5.87 ([2.1], 1H, zq, g§1.3, C6 methyl). 59 In a decoupling experiment, the doublet at 51.15 collapsed into a sing- let upon irradiation at the frequency of the quartet at 51.87. Compound 3;; The previous experiment was repeated. Continued heating beyond 1 hour led to the complete conversion of Qfi’to its epimer 3§,after an additional 90 min. Compound 3R‘was stable to further heat- ing and was purified for analysis using a 5' x 3/8" column of 20% FFAP on 30/60 Chromosorb w at 160°. Ir (cc14) 1725, 1635 (w) cm'l; nmr ,_ (cc14) 6.72 [2.2] and 1.17 [1.5] (3H each, s, gem dimethyl), .82 ([4.4], 3H, d, g;7.5, C3 methyl), 1.87 ([1.0], 3H, m, C7 methyl), 2.77 ([6.2], 1H, q, gf7.5, C3 proton), 2.83 ([2.0], 1H, m, C5 proton), 2.92 ([5.7], 1H, m, C1 proton), 5.97 ([2.2], 1H, ‘q, £31.3, C6 proton); mass spectrum (70 eV) gig 164 (M+), 121 (base). Mixtures of’é4 and 35 were obtained in early attempts to purify 36 by preparative vpc. Relative amounts of’34 and 35 varied with column conditions and type. Injection of 36 resulted in a single peak. Collection and subsequent analysis by nmr revealed the complete conversion of'ég to the two isomers. Compounds 34 and‘gé had nearly the same retention time of about 19.5 min (single assymetrical peak) on a 5' x 1/8" analytical column of 15% FFAP on 80/100 DMCS Chromosorb W at 110°. This is similar to the retention time of 7 £2. Anal. Calcd. for C11H160: C, 80.43; H, 9.82. Found: C, 80.33; H, 9.81. Photolysis of’lg in TFE. A solution of .1 g 0f.L% in 10 m1 of TFE (.06 M) in a 1.5 cm x 12.5 cm Pyrex tube was purged with nitrogen and irradiated for 1 hour using a 450 watt lamp with Pyrex filter. Reaction progress was monitored by Vpc under the same conditions used in the photolysis 0f.L§ in cyclohexane. An 85% conversion 0f.L€ to photoproducts was observed. Compounds 21 (28%), 2% (5.7%) andtgé (7.5%) had the previously observed retention times. Also observed were 60 compounds ZA’(43%) and 25 (15.8%) at retention times of 19.8 and 11.2 min respectively. The chromatogram also showed two other peaks (6.0 and 8.2 min) totalling about 8%. These were shown to be secondary photoproducts arising from 24 and were not investigated further. Com- pound ZA'had the same retention time as IR and was rapidly converted to secondary photoproducts under the reaction conditions. The peak corresponding to IR'and gg'was therefore collected and analyzed by nmr to determine the relative amounts of each isomer in the photolysis: mixture. Preparative vpc conditions were the same as those used in the separation of the cyclohexane photolysis mixture. Pure afl’was obtained by another method (see below) and isolated as a colorless liquid. Ir (neat) 1660 cm"1 ; uv (cyclohexane) 346 nm (e=88), 317 (sh 270), 268 (3700), 244 (4660); nmr (CC14) 6.80 [2.2] and 1.22 [1.0], (3H each, s, gem dimethyl), 1.15 ([1.7], 3H, d, g;7.s, C2 gxg_methyl), 1.45 ([2.7], 3H, d, g?1.5, C4 allylic methyl), 2.50([7.6], 1H, q, g;7.5, each peak broadened by further coupling, C2 gndg_protonL 6.72 ([2.4], 1H, d, g;s.s, each peak split into a q, gf1.5, C5 vinyl proton. The C1 and C6 protons appeared as multiplets in the 6.90-1.50 region which could not be satisfactorily shifted away from the methyl signals. Mass spectrum (70 eV) mg 164 (M+), 122 (base). A231, Calcd for C11H160: C, 80.43; H, 9.82. Found: C, 80.30; H, 9.92. Compound{%§ was identified as the main product of the photolysis of 2% in TFE. Irradiation of a .05 g sample of pure 2% under the con- ditions of the preceding experiment led to a 75% conversion. In addi- tion to an 80% yield of 25, two minor products (observed above in the TFE photolysis of‘Lg) were formed in about 20% total yield. These were 61 not investigated further. Compound 25 had a carbonyl stretching fre- quency of 1730 cm"1 in the infrared (neat). Uv (cyclohexane) 319 nm (68179), 307 (256), 296 (230), 280 (sh 18), 214 (2410); nmr (CC14) 6.92 [1.4] and 1.07 [1.0] (3H each, s, gem dimethyl), .97 ([2.3], 3H, 5, C1 methyl), 1.02 ([2.3], 3H, d, g;7.0, C3 gndg_methyl), 2.25 ([3.9], 1H, q, gf7.0, each peak split into a doublet, gf=3.5, C3 3x9 proton), 2.50 ([1.4], 1H, broad t, C4 proton), 5.55 ([1.8], 1H, d, g§6.0, C6 vinyl proton), 6.38 ([l.3], 1H, d of d, g§6.0, gf-3.0, C5 vinyl proton); mass spectrum (70 eV) Me 164 (M‘), 107 (base), 93 (999. of base). Photolysis of ARRRXR The procedure for the preparation and ir- radiation of ARRR;.and the recovery of its photoproduct was identical to that used for Aflaflix The cation displayed nmr signals at 51.23 and 1.32 (3H each, s, gem dimethyl), 1.47 (3H, d, g;7.5, C7 methyl), 2.18 (3H, 5, C2 methyl), 2.87 (1H, q, g;7.5, C7 proton), 6.65 (2H, broad 5, C4 and C5 protons), 7.85 (1H, broad d, gf6, C3 proton). These gradually decreased in intensity during the photolysis, and were replaced by the signals corresponding to (ARRK; 6.87 and 1.53 (3H a- piece, 5, gem dimethyl), 1.45 (3H, d, g;7.0, C2 methyl), 2.08 (3H, 5, C4 methyl), 3.35 (1H, q, g;7.o, C2 proton), 8.58 (1H, d, gf6.5, C5 proton), C1 and C6 protons appeared as multiplets at about 1.90 and 2.60, partially obscured by the methyl signals. Conversion of (RRH;,"35 apparently complete in 6 hours, but irradiation was continued for an additional 3 hours to insure purity of the product. Cation 24a“; was stable and unaffected by further irradiation. Upon quenching, the sole product, 2%, was obtained as a pure, pale yellow oil. The Quenched Photolysis of 2*; Into each of five 5 mm x 50 mm Pyrex tubes was placed 2.5 mg of 21. To one of the tubes was added 62 .3 m1 of cyclohexane. Into each of the remaining tubes was placed .3 m1 of a gi§_1,3-pentadiene solution in cyclohexane. The concentrations of quencher in these four solutions were 1M, 2M, 3M, and 4M. The five tubes were then fastened to a quartz immersion photolysis well and ir- radiated for 45 min using a 450 watt lamp with a Pyrex filter. A lul aliquot of each of the five samples was then injected into an analytical vpc column (5' x 1/8" 10% FFAP on DMCS Chromosorb W at 120°) and the ratio of the areas of the two observed peaks (starting material at shorter retention time and photoproduct at longer retention time) was recorded for each sample. The following ratios of ‘6 to photoprod- uct were observed: .09 (no quencher), 2.3 (1M quencher), 3.3 (2M quencher), 4.3 (3M quencher), 3.8 (4M quencher). Since the conver- sion of gA'was greater than 90% in the absence of quencher and decreas- ed with increasing quencher concentration, 66 was assumed to arise (at least in part) via a triplet excited state of 2*, At higher quencher concentrations (above 3M) the photoproduct peak undoubtedly contained 66, the alternative photoproduct of 2A; This experiment was complicated by the fact that 66 is rapidly converted to 66 in a Vpc column (see above) so that the amount and identity of the photoproduct(s) could not be determined by inspection of a vpc chromatogram. In an attempt to identify the photoproduct(s) formed in the presence of the triplet quencher, a preparative photolysis (.025 g of £6 in 3.5 m1 of the 3M quencher solution irradiated for 90 min) was carried out. Evaporation of the solvent gave a colorless oil which was examined by nmr and ir. The nmr spectrum of this crude material was complicated by the presence of polymerized quencher. However, peaks corresponding to 66 (65.87, 3.12, 2.77, 1.83) were visible in the spectrum of the crude product. 63 Also visible was a peak at 61.68 corresponding to the bicyclobutyl methyl of 36. All other peaks of the photoproducts 66 and 66 were hidden by peaks corresponding to the starting material 26 and to polymerized quencher. The nmr spectrum suggested an approximately 1:1 ratio of 66 to 66. The ir spectrum (CC14) of the crude mixture showed strong carbonyl absorption at 1722 cm-1 corresponding to 66 and 66. A shoulder (1705 cm'l) of this band was perhaps due to the presence of 66 in the mixture. A qualitative quenching experiment was then performed using a 3M gi§_1,3-pentadiene in TFE solution as the solvent. Irradiation of 2.5 mg of’ZA'in .3 m1 of this quencher solution for 90 min resulted in a starting material to photoproduct ratio of 1.4. An identical ratio was obtained from the photolysis of 2.5 mg of gA'in .3 ml of TFE. This result is perhaps an indication that 66 is formed via an excited singlet state of 66; However, no preparative photolysis was carried out in TFE-quencher solution, and so the identity of the photoproduct formed under these conditions remains unconfirmed. Also unknown is the effect of the change in solvent polarity (TFE-quencher vs pure TFE) on the type of photoproduct formed. Methylation of’g6, To a solution of 1 m1 of HMDS, 1 ml of 1.6M n-butyllithium in hexane and S m] of THF was added .2 g (.0012 mol) of'g6, Addition of 1 ml of methyl iodide was then followed by 30 minutes of refluxing. The mixture was worked up by adding 5 ml of dilute HCl and extracting the mixture with ether. The ether layer was dried and concentrated to an oil. A vpc chromatogram showed a near quantitative transformation of 66 to 66. Compound 66 prepared in this manner was identical (ir, nmr) to the minor product of the methylation of,§a. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. REFERENCES J. Griffiths and H. Hart, J. Amer. Chem. Soc., 22, 5296 (1968). G. Quinkert, Photochem. Photobiol., Z, 783 (1968); O. L. Chapman and J. D. Lassila, J. Amer. Chem. Soc., 22, 2449 (1968). J. Griffiths and H. Hart, J. Amer. Chem. Soc., 22, 3297 (1968). R. B. Woodward and R. Hoffman, The Conservation of Orbital Symmetry, pp 81-82. Academic Press; New York (1970). H. H. Jaffe' and M. Orchin, Theory and Applications of Ultraviolet Spectroscopy, pp 186-195. Wiley; New York (1962). B. Parrington and R. F. Childs, Chem. Comm., 1581 (1970). E. J. Corey and H. J. Burke, J. Amer. Chem. Soc., 12, 174 (1956). H. Hart and T. Takino, J. Amer. Chem. Soc., 22, 720 (1971)._ K. E. Hine and R. F. Childs, J. Amer. Chem. Soc., 22, 2323 (1971). G. Bfichi and E. M. Burgess, J. Amer. Chem. Soc., 22, 433 (1960). D. I. Schuster and D. H. Sussman, Tetrahedron Lett., 1657 (1970). J. J. Hurst and G. H. Whitham, J. Chem. Soc., 710 (1963). D. I. Shuster, M. J. Nash, and M. L. Kantor, Tetrahedron Lett., 1375 (1964). K. E. Hine and R. F. Childs, J. Chem. Soc. D, 145 (1972). D. I. Schuster and M. A. Tainsky, Molec. Photochem., 2, 437 (1972). R. B. Woodward and R. Hoffman, The Conservation of 0rb2ta12§ym- metrz, pp 65-73, 89-100. Academic Press; New York (1970). H. Hart and A. F. Naples, J. Amer. Chem. Soc., 22, 3256 (1972). 0. L. Chapman and G. W. Borden, J. Org. Chem., 22, 4185 (1961); 0. L. Chapman, D. J. Pasto, G. W. Borden, and A. A. Griswold, 22_ Amer. Chem. Soc., g5, 1220 (1962). L. A. Paquette and 0. Cox, J. Amer. Chem. Soc., 22, 5633 (1967). R. B. Woodward and R. Hoffman, The Conservgtion of Orbital Sym- metry, p 58. Academic Press; New York (1970). H. Hart and A. F. Naples, Pure Appl. Chem., 22, 247 (1973). 64 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 65 See Ref. 21 for leading references. Also see the spectra for 666, 66, 66, and 26 given on pages 44, 43, 53, and 51, respectively in the xperimental portion of this thesis. See L. M. Jackman and S. Sternhell, Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry, 2nd Edition, Permagon Press, Oxford, 1969, p 225 and references therein. A. J. Bellamy and W. Crilly, J. Chem. Soc. Perkin II, 395 (1972). An analogous compound is formed in the photolysis of eucarvone (see Ref. 13). For leading references, see Ref. 2_and P. H. Mazzochi and R. C. Ladenson, Chem. Commun, 469 (1970). Possibly was formed, but only in trace amounts. Vpc chroma- tograms o the photolysis of’6 in cyclohexane show a small peak (