JJJJ h; . J J ,I'.'|U‘I,I":y' w, ”,th , ; JJJJJ J JJJJ JJJJJJJJ JJJJJJ JJJJJ JJJJJJJJJJ‘J’JJ: w: . JJJ JJJJJJJJJJJJJJ J . JJJ J JJJJJJJJJJJJJJJJJJ JJ JJ JJJJJJJJJJJ J J ~ ~ , JJ'J'JJJJJJ JJJJJJ “J JJJJJJJ JJ' JJ JJJ ; JJJJJJ JJJJJ JJJ JJJJ‘JJJJJJJJJJJIJJJ JJJJJ’ J J” J J IIIIJ I I . ' JJ' ’J JJJJ JJJJJJ J J‘JJ‘JJ II, JJJJJJJJ, JJJJ J] J J . JJJIIJJJJ.JJJJ'JJJJ‘JJJJJ .JJJJJ J JJJ J J JJJJJJJJ “:N“ 1‘... FM“- i‘m‘ - - - ‘ ‘ V w . - ‘ . m -- . - -V ml-—-_ - ‘ -m._ f—g—D- . -_ - ‘ “ «W m J “W.” _‘ w“ r A .- -—7 —fi_ _ fi- _—- . E I ~35 .~ -~—.._ -- _‘—'H‘_-_-‘.__ .— JJJIJJJJIJ . JJJ ’JJJJ' JJJJJ IJJJJJI JJJIJJ:IJ:JJJIJJJI IIJJJJ'JIJJIJ,IJJIJJ [J'JIII JJJJJ. JJJJJJJJJJ 1}" J'JJJ' 'IJJ JJJJJ IJ JJJ'JJJ JJJ JJJJJJJJJ JJI J J‘ :JJJJJJJ‘J JJJJJJJJJJ'I J ‘J JJ "J‘ JJJ J JJJ EJJ! JJJ J' ”J JJJJJ JJ g'IJJ‘v J J‘ . JJ‘ JJ JJJ JJ J'JJJJJJJ JJJJIJJJJJJ JJJJJJJ JJJ: J JJJJJJ' JIJIIJJJ JI J‘JJJ'JJ JJJJIJJJJJ JJJJ {I JJJJJJ' III JJJ‘JJJJIIJIJJI JJJ JJJIIHJJJJJJ JJJ‘JJ' IHJJ JJJ JJJ JJ JJJJJJJ JJJJ JJIJ JJJJJ J'J ' J JJ ,, J Jl‘i‘ I. J' . I .. ~ JJJJJI JJ’J JJJJJJJJ J J JJJJ JJJJJJ th JJJJJJ’ JJJ J This is to certify that the thesis entitled l. REGIOSPECIFICITY IN THE DI-fl-METHANE REARRANGEMENT 2. THE IONIZATION OF TWO TETRACYCLIC GEM-DIFLUORIDES UNDER STABLE ION CONDITIONS 3 . MISCELLANEOUS presented by Daryl L. Stein has been accepted towards fulfillment of the requirements for Ph . D . degree in Chemis trv .d a d lief Major professor DateDAPC, 2: ‘qu7 0—7 639 PART I REGIOSPECIFICITY IN THE DI-fl-METHANE REARRANGEMENT PART II THE IONIzATION OF TWO TETRACYCLIC egg—DIELUORIDES UNDER STABLE ION CONDITIONS PART III MISCELLANEOUS By Daryl L. Stein A THESIS Submitted tO Michigan State University in partial fulfillment Of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1978 ABSTRACT PART I REGIOSPECIFICITY IN THE DI-w-METHANE REARRANGEMENT PART II THE IONIZATION OF TWO TETRACYCLIC figfiéDIFLUORIDES UNDER STABLE ION CONDITIONS PART III MISCELLANEOUS BY Daryl L. Stein In Part I of this thesis the acetone sensitized di-n—methane res arrangement of 9,9-dimethy1-1,4-dihydro-1,4-ethanonaphthalene a was studied. 0f the two possible photoproducts, 6,6-dimethyl-3,4- benzotricyclo[3.3.0.02'8]oct-3-ene $3 is the major (662) one. The structure of the major iSOmer was determined by synthesis. This result is in agreement with previous work in which other substituents on the 9 position caused a similar regiospecific rearrangement. In each case the major photoproduct arises from a diradical species in which the unpaired electrons are close to rather than remote from the R groups on the saturated bridge. In addition the quantum yields of formation for the corresponding photoproducts of compounds %I, 1, lg, Daryl L. Stein RR1=R2=CH3 33 3’2 ZRl-H1,R2=OH g g IRRl-OH,R2=H 3;?» II %2 R1 ' H, R2 ' CH3 k2 ti IIQRIBCHTRZ-H ii (I; I}, and *2 were determined in benzene with various sensitizers. In each case acetone was the most efficient sensitizer. In addition the triplet energy levels for these compounds were found to be between 68 and 72 Kcal/mole. In Part II when either fluorocarbon 22,0r gz'was treated with SbF5 at -95° C the corresponding apical substituted fluoro pyramidal ion was not observed. F | .9. R1 ’I'o “x --_= --)(--> g g R gQRl-CH3,R2-H $2 5 R =3 a «X 1 H, R2 CH3 Instead 22 ionized to the bicyclic[3.2.l] ion 23. The structure Daryl L. Stein 53—» /*%*&§”% 72 of 62 was confirmed by quenching in sodium methoxide-methanol to give the corresponding ketone 69. The structure of ketone 69 was mm mm proved by treatment with methyllithium to give exoumethylene compound 2% which when treated with F8038 at -95° C, gave the symmetric allyl ion 1%. In contrast to 3%, fluorocarbon 22 ionized to the homotropylium “ 1 .. .. 5° 1‘ -* [mm] 7' é" é ion 22. iA methanol quench at -78° C gave a mixture of methoxyether quench products. Attempted VPC separation of this mixture resulted in ring contraction and loss of methanol to give fluorostyrene 2% as the major product. The structure of 2% was confirmed by converting it to the known fluorotoluic acid 2%. mechanisms for the formation of ions 22 and 22 are also proposed. In Part III hexamethylcyclopentadiene 26 was treated with various Daryl L. Stein dienophiles and subsequent exploratory reactions were performed on the various Dials-Alder adducts. m + arc-cm, --—> R, 96 104 WWW m R1 - c113, R2 - c0211 $92 R1 - ca3, R2 - cozca3 LR R1 ' R2 " “’3 In addition, 26 also undergoes a Diels-Alder reaction with J? I R2 ' C02CH3 singlet oxygen to form the corresponding endo-peroxide Ill. The I - PRO 0 . 0 ....,_..g_.. o -~g 111 122 123 structure of III was confirmed by reduction of III to dial lgg which upon treatment with acid gave the known triene NEE“ ACKNOWLEDGEMENTS I am indebted to Professor Harold Hart for his guidance and support throughout the course of this study. I am also indebted to the National Science Foundation, the National Institute of Health and Michigan State University for their financial support in the form of research and teaching assistantships. I am especially indebted to my wife Susan for her patience, support and constant encouragement. ii TABLE OF CONTENTS PART I REGIOSPECIFICITY IN THE DI-N-METHANE REARRANGEMENT INDUCTION O O O I O O O O O O O O O O O 0 O O O O O O O O O 0 RESULTS DISCUSS ION O O O O O O O O O O O O O O O O O O O O O O O O O O O mum“ O O O O O O O O O O I O O O O I O O O O O O O O O 0 General . . . . . . . . . . . . . . . . . . . . . . . . Quantum Yields . . . . . . . . . . . . . . . . . . . materials (for quantum yield studies) . . . . . . . . . Procedure (for quantum yield studies) . . . . . . . . . Preparation of 9,9-dimethyl-l,4-dihydro-1,4- ethanonaphthalene (El). . . . . . . . . . . . . . . . Preparation of 9,9-dimethyl-l,2,3,4-tetrahydro-1,4- ethanonaphthalene (g2). . . . . . . . . . . . . . . Preparation of 7,7-dimethyl-2,3-benzobicyclo[3.2.l]- OCten-O-One (a) o o o o o o o o o o o o o o o o o o 0 Preparation of 7,7-dimethyl-2,3-benzobicyclol3.2.1]- OCtane (%) o o o o o o o o o o o o o o o o o o o o o Hydrogenation of 6,6-dimethyl-3,4-benzotricyclo- [3.3.0.02’810Ct-3-ene (a). o o o o o o o o o o o o o o Preparative photolysis of 9,9edimethy1-l,4-dihydro-1,4- ethanonaphthalene (gl). . . . . . . . . . . . . . . . . iii Page 16 26 26 27 27 28 28 30 3O 31 32 32 TABLE OF CONTENTS (continued) PART II THE IONIZATION OF TWO TETRACYCLIC GEMPDIFLUORIDES UNDER STABLE ION CONDITIONS INTRODUCTION . . . . ..... . ........ . . . . RESULTS 0 O O O O ..... O O O O O O O O O ...... A. The Structure of Ion QR ..... . . ..... B. The Structure of Ion 66’. . . . . . . . . . . . DISCUSSION Page . . . 35 . . . 42 . . . 44 . . . 53 A. Mechanism for the Ionization of Fluorocarbon QR'. . . 65 B. Mechanism for the Ionization of Fluorocarbon 2a . . . 67 EHERMNTAII O O O O O O O O O O O O O ...... O 0 O . . . 73 A I General 0 O I C O O C O O C O O O O I O O O O O O O O 73 B. Preparation of bornadiene (2Q). . . . . . . . . . . . 73 C. Preparatiog of 4,4-difluoro-6,7,7-trimethy1 tetracyclo- (3.3.0.02, trimethyl tetracyclo[3.3. .O2 9 ]octane (2g) 03:6]octane ( ang g, -difluoro-7,7,8- O I O 74 D. Preparation of 2-f1uoro-4,8,8-trimethylbicyclo[3.2.1]- octa-2,6-dienylium ion (6;) and 4—f1uoro-3,8,8- trimethylcyclooctatrienylium ion (66) . . . . . E. Preparation of 2-methy1ene-4,8,8-trimethy1- O O O 74 bicyclo[3.2.1]octatriene (1}) . . . . . . . . . . . . 77 F. Preparation of 2,4,8,8-tetramethy1bicyclo[3.2.1]Octa- 2,6“dieny11lnn 1011 (a). o o o o o o o o o o o o G. Preparation of fluorostyrene (QT) . . . . . . . H. Oxidation of fluorostyrene (gl) . . . . . . . . iv 0 O O 78 . . . 78 . . . 79 TABLE OF CONTENTS (continued) PART III MISCELLANEOUS INTRODUCTION ....... . ............ . . . . . RESULTS EDHMNTAL O O O O O I O O O I ..... O O O O O O O O O O A. B. General 0 O O O O ..... O I O O C O O O O C O O 0 Preparation of 1,2,3,4,5,5-hexamethylcyclopentadiene (fi) 0 ..... O I O O O O O O O O O O O O O O O O I Attempted preparation of 1,2,3,4,5,6,7,7-octamethyl— b1CYC10[2 o 2 o 1 ] hepta.2 ’ 5’diene (22) o o o o o o o o o o Attempted preparation of 1,2,3,4,7,7-hexamethyl-5,6- diphenylbicyclo[2.1.1]hepta-2,5-diene (12g) . . . . . Preparation of 5,6-di(carbomethoxy)-1,2,3,4,7,7- hexamethylbicyclo[2.2.1]hepta-2,4-diene (62$) . . . . Saponification of diester 104 . . . . . . . . . . . . Nmm Attempted reduction of diester 124 .......... Reaction of tetrolic acid with 26 . . . . . . . . . . Preparation of 2-(carbomethoxy)-1,3,4,S,6,7,7- heptamethylbicyclolz.2.1]hepta-2,5-diene (12$). . . . Reduction of 2-(carbomethoxy)-1,3,4,5,6,7,7-hepta- methylbicyclolZ.2.1]hepta-2,5-diene (12$) . . . . . . Attempted reduction of ester IDS to hydrocarbon 3%. . Preparation of endo-cis-S,6-di(carbomethoxy)- l,2,3,4,7,7-hexamethy1bicyclo[2.2.1]hepta-2-ene (110) -~ Epimerization of cis-S,6-di(carbomethoxy)- 1 , 2 , 3, 4 , 7 , 7-hexamethylbicyclo [2 . 2 . l]hepta-2-ene (1,1,0) Photooxidation of 1,2,3,4,5,S-hexamethylcyclopenta- diene (22). o o o o o o o o o o o o o o o o o o o o o Page 82 84 100 100 100 . 101 O 101 . 102 102 103 . 103 104 . 104 . 105 106 106 107 TABLE OF CONTENTS (continued) Page 0. Preparation of cis-1,2,3,4-di(epoxy)-l,2,3,4,5,5- hexamethylcyclopentane (124). . . . . . . . . . . . . . 107 P. Preparation Of cis-l,2,3,4,4,5-hexamethylcyclopent-1- ens—3,5 diol (122). . . . . . . . . . . . . . . . . . . 108 Q. Preparation of 1,2,4,4—tetramethy1—3,5-dimethylene- cyclopentene (12;). . . . . . . . . . . . . . . . . . . 108 R. Preparation of 1,2,3,4,7,7-hexamethyl-5,6-di(tri- f1uoromethy1)bicyclo[2.2.1]hepta-2,5-diene (112). . . . 109 8. Preparation of 7,7-dimethyl-2,3-di(trif1uoromethyl)- bicyclo[Z.2.1]hepta-2,5-diene (121) . . . . . . . . . . 109 T. Preparation of 2,3-di(trifluoromethyl)bicyclo[2.2.1]- hept8-2,5-diene (1.3) o o o o o o O o o o o o o o o o o 110 U. Photolysis of 1,2,3,4,7,7—hexamethyl-5,6-di(tri- fluoromethy1)bicyclo[2.2.1]hepta-2,5—diene (121). . . . 110 V. Photolysis of 7,7-dimethyl-2,3-di(trif1uoromethyl)- b1CYC10I2 o 2 o llhepta-z ’ 5-diene (W) o o o o o o o o o o 111 W. Attempted carbene addition to 122’. . . . . . . . . . . 111 X. Preparation of endo—5,6-epoxy-1,4,5,6,7,7-hexamethy1- 2,3-di(trifluoromethy1)bicyclo[2.2.1]hepta-2-éne (1%1). 112 Y. Preparation Of exo-S,6-epoxy-2,3-di(trif1uoromethy1)- bicyclo[2.21]hepta-2-ene (A§§)' . . . . . . . . . . . . 113 Z. Preparation Of l-acetyl-l,4,5,6,6—pentamethyl-2,3- di(trifluoromethyl)cyclohexa-2,4-diene (Iii). . . . . . 114 AA. Bromination of diene I33; . . . . . . . . . . . . . . . 114 REFERENCES 0 o o o o o o o o o o o o o o o o o o o o o o o o o o 1'16 vi TABLE LIST OF TABLES Quantum Yields of Formation with Various Sensitizers. Carbon-13 Chemical Shifts of Compounds 32, 22,“ 64 . . Carbon-13 Spectra of Fluorocarbonium Ions 63 and 66 Compared with Those of Mbdel Ions . . . . . . . . Proton and Carbon-13 Data for Ketone 62 and Mbdel compo und m. 0 O O O O O O O O O O O O O O O O 9 Proton NMR Spectra of 11 and the Corresponding Carbonium Ion 72. . . . . . . . . . . . . . . . . . . Proton NMR Spectrum of Ion 66 Compared with That of Madel Ion O O O O O O O O O O O O O O O O O O O O O O 0) vii Page 15 43 45 48 51 55 10. 11. 12. 13. LIST OF FIGURES Singly occupied MD'S of donor-acceptor substituted benzene after interaction with an ethylene unit . . . Two possible diradical intermediates. . . . . . . . . . The di-fl-methane system as a substituted benzene and e thyl ene O O O O O O O O O I O O O O O O O O Q 0 O O O 0 The 2-norbornyl cation represented as an equilibrating pair of classical ions and as a nonclassical ion. . . . The pyramidal nonclassical (CH)5+ ion . . . . . . . . . Examples of known pyramidal cations . . . . . . . . . . Proton NMR spectrum of ion 62 . . . . . . . . . . . . . Europium shift slopes for the protons in ketone 22. . . Carbon-13 chemical shifts of protonated fluorobenzene, 77, with carbon-fluorine coupling constants in ParenthBBEB o o o o o o o o o o o o o o o o o Q o o o o The proton chemical shift (6) assignments for 122 . . . . Chemical shift (6) assignments for 110. . . . . . . . . The proton and carbon-13 chemical shift (5) assignments for 111 in parts per million from TMS. The carbon-l3 shifggmare in parentheses . . . . . . . . . . . . . . . The proton chemical shifts (5) of IBS in ppm from.TMS. Europium shifts are in parentheses. . . . . . . . . . . viii Page 23 24 24 35 36 37 47 49 58 89 90 92 97 PART I REGIOSPECIFICITY IN THE DI-fl-METHANE REARRANGEMENT INTRODUCTION Zimmerman first recognized the importance of the di-w-methane rearrangement and has extensively characterized it.1 Basically, this photochemical reaction involves the formation of a vinyl cyclopropane from molecules having two vinyl groups attached to the same methylene carbon (eq 1). The first step Of the reaction is usually depicted as initial bonding between 0-2 and C-4, which results in a diradical with a cyclopropane ring. Cleavage of either the C-2, C-3 bond or the C-3, C-4 bond of the cyclopropane ring leads to a new diradical which 4 <1) 3 3 2fr’/~\\Irt hI' 2[¢::::>14. 2 '_'> -—> 1 5 1o .5 1 then closes to the final product. The diradical species are not necessarily discrete intermediates. For example, Zimmerman has suggested on the basis of stereochemical studies that reaction from the singlet excited state may be concerted. In this event, the di- radical would simply represent one point along the reaction coordinate. Nevertheless, these qualitative valence bond pictures are useful in describing the gross mechanism and in making predictions when one u system will react in preference to another. Thus in an unsymmetrical molecule (eq 2), the major product can be predicted by examining which .reaction leads to the more stable diradical intermediate.2 Cleavage of bond A is preferred because the odd electron adjacent to the phenyl groups is stabilized to a greater extent (through resonance) than is the Odd electron adjacent A» I .’___&¢ A B 2 | I J31» - (2) ¢ ° 1» 1 Lfiifi ¢ ¢ ¢ ¢ 3 Equation 3 represents another type of regiospecificity in which to the methyl groups. initial bond formation determines the final product. Although there are formally four di-n-methane units in the molecule, only two products are observed. Both of these products can be rationalized by an initial bonding which forms an allyl radical. Initial bonding at other positions would form a less stable primary or second diradical species or it would disrupt the aryl ring. Once the allyl radical is formed, the rearrangement then proceeds so as to maintain the odd electron at this position for a long as possible which explains the preponderance Of 5 over 6. Since this reaction proceeds via both the singlet and triplet excited states, multiplicity does not seem to be a controlling factor in determining regiospecificity. (3) 6 (4%) Although all of the above products can be rationalized by through bond stabilization of radical sites, there are examples of a less well understood through-space control of the di-fl-methane reaction. In the reactions represented by eqs 5 and 6, the hydroxyl function is not directly attached to the di-fl-methane system, yet somehow it controls the course of the reaction, favoring one diradical intermediate over. the otherfm"b Proposed explanations for these phenomena are the following: 1) internal hydrogen bonding in which the hydroxyl group stabilizes one of the two possible diradicals, 2) a charge transfer complex involving the lone pairs of oxygen with a radical site, or 3) some type of asymmetric energy transfer from an excited donor molecule to the acceptor molecule. 11 lam (5) 0' ’ "'°" 12 hex! The Observation that the acetates corresponding to Z and 12 displayed a similar regiospecificity eliminated hydrogen bonding as the exclusive explanation but did not rule out lOne pair interaction. However, doubt also was recently cast on this rationalization by the observation that substitution of a methyl group for a hydroxyl as in R and 12 does not affect the regiospecificityS (eq 6 and 7). Since a methyl group has no unshared electron pairs, there must be another explanation . for the observed effect (vide infra). 18 man In order to further understand the factors that control this remote substituent effect of the di-n-methane reaction, several new experiments were carried out. First, in order to determine whether a gem-dimethyl group would exhibit the same regiospecificity as the single methyl group in compounds 13 and '10?) the analoggrl‘.’ was synthesized and irradiated. Second, the quantum yields for the di-n-methane rearrangement of z, IQ, 13, 16, and 21 were determined with various triplet sensitizers in order to ascertain and compare the triplet energy levels of the excited states, and also to better understand the partitioning of the excited state among the various products. These studies constitute the first part of this thesis. RESULTS The synthesis of 9,9—dimethy1—1,4-dihydro-l,4-ethanonaphthalene (21) was accomplished by adding benzyne to the known6 5,5-dimethyl- cyclohexa-1,3-diene 22 (eq 8). Unfortunately, the reaction mixture contained not only 21 but also %% and 2%, in the ratio of 3:2:1. The '39—- 19 (8) O + o + 21 21 formation of this mixture was not too surprising, since benzyne is known to react with cyclic olefins to give not only 4n + 21 additions, but also ene reactions and 2n + 2n products.7 .The separation of 21 from this product mixture, although difficult, could be accomplished by silica gel chromatography, followed by prepara- tive VPC of the 2lrenriched fractions in order to remove the last traces of 22 and 23. Compound 22 proved to be an effective quencher of the di-n-methane rearrangement of 21, and unless it was completely removed photolysis of gl'was slow. The structure proof of the products rests on spectral data, and on subsequent chemical transformations (vide infra). The NMR (Nuclear Magnetic Resonance) and IR (Infrared) spectra of fig were identical with those of an authentic sample.8 Comparison of the NMR spectrum of El with those of the closely related *3 and £2 was instructive. In compound $2, the methyl group is syn. to the aromatic ring, and appears as a shielded doublet at 6 0.60 ppm. In *2, the methyl group is anti to the ring and appears as a downfield doublet at 6 0.93 ppm. Compound gl, with methyl groups both syn and anti to the ring, exhibits two singlets, one at 5 0.60 ppm and the other at 6 1.03 ppm, consistent with expectation. Furthermore, the aromatic, vinyl, bridgehead and methylene protons for all three compounds have nearly identical chemical shifts. Their mass spectra are also similar, with a base peak at m/e 128, corresponding to bridge cleavage to naphthalene in each case. The structural assignment of £3 rests mainly on the premise that benzyne should add to the less hindered double bond (i.e., the double bond which is remote from the gem-dimethyl group). Furthermore the chemical shifts and coupling constants for the vinyl protons [5 5.45 ppm (1 H, d, J = 10 Hz) and 5 5.84 (1 H, dd, J I 10, J = 3 Hz)] are con- sistent with the assigned structure. The splitting pattern of the vinyl protons would have been different had cycloaddition occurred at the other double bond. The preparative scale photolysis of %l was carried out in acetone and was monitored by VPC. In addition to a small amount of starting material, three peaks with an area ratio of 1:1:2 were obtained by 10 preparative VPC. The products were collected and identified as $3, $2 and gg (eq 9). Photoproduct gfi is the result of photoreduction. In addition to literature precedent for such reactions,9 a similar reaction occurred with kg and $3.5 The NMR spectrum of $3 is consistent with its structure, and in its mass spectrum the parent ion was 2 atomic mass units greater than the parent ion of the precursor, gl. The structure of g& was confirmed by catalytically hydrogenating gl over Pd/C. The NMR and IR spectra of the hydrogenation product were identical to those of %g obtained from the photolysis. Photoproducts g; and g3 arise from the expected di-w-methane rearrangement of gl. In each case the parent ion in the mass spectrum (9) 11 indicated that these products were isomeric with the starting material. The absence of vinyl protons in the NMR spectra of £2 and £8 indicated the formation of another ring. The two di-n—methane photoproducts can be distinguished by examining the chemical shifts of the ggmfdimethyl groups. Examination of models reveals that these photoproducts are cup-shaped. Thus for %2 the gynfmethyl group is tucked down toward the phenyl ring, causing it to be shielded by the ring current. This methyl group appears at high field in the NMR spectrum, at 6 0.48 ppm which is in agreement with the chemical shift of the gygfmethyl group of 18 which is also at 6 0.48 ppm. The gygfmethyl for £2, although it also points toward the phenyl ring, is not nearly as close to it. This methyl group therefore appears at somewhat lower field, at 6 0.76 ppm which is close to the 6 0.74 ppm shift observed for the gzgfmethyl group of 11. The antirmethyl groups in $2 and £2 appeared at 6 1.12 and 1.19 ppm, respectively which is in good agreement with the chemical shift at 6 1.08 ppm observed for the .gggifmethyl in 13. Thus it is clear that the expected di-fl-methane rearrangement did, in fact, occur upon sensitized irradiation of a1. However, since proton NMR spectra for the rearrangement products are complex, it was desirable to have a more rigorous criteria than just the chemical shifts of the methyl groups for assigning the product structures. In order to put these assigned structures on more firm ground, the following independent synthetic correlation was carried out (Scheme 1). 12 Scheme 1 By analogy with previous work,5 the three-membered ring in 22 was catalytically hydrogenated to give 24. Although_a priori any one of the three cyclopropyl bonds could have been cleaved, one would expect either the 1,2 or 2,8 bond to be preferentially hydrogenated since they are activated by the aryl ring. Furthermore, molecular models indicate 13 that the 1,2 bond of 22 can lie nearly flat upon a table top (catalytic surface). The geometry of the molecule prevents the 2,8 and 1,8 bonds from approaching the catalyst surface as easily as the 1,2 bond. Consequently a: (from 1,2 bond cleavage) is expected to be the major hydrogenation product. Analysis of the reaction mixture on an analytical gas chromatograph with an SE-30 column gave only a single peak. The mass spectrum indicated an increase of two mass units. The NMR spectrum indicated 4 aromatic proton signals at 6 6.80 ppm and two methyl group signals at 6 1.10 and 6 0.63 ppm. The remaining proton signals appeared as a complex multiplet. It remained to synthesize 24 by an independent route. The bicyclic ketone 2Zfb’ 10 was photolyzed to the known tricyclic ketone gg‘by established procedures.11 This ketone was then methylated, reduced and hydrogenated according to a procedure developed by P. Lavrik5 to give alcohol 21. The ketone must be reduced to the alcohol prior to hydrogenolysis because the carbonyl function in 22 activates the 2,8 cyclopropyl bond. Hydrogenation at this stage would exclusively cleave the 2,8 bond. Alcohol zk‘was then oxidized to ketone 22. This ketone, although hindered, was alkylated by using the very reactive base potassium hydride.13 Wolff-Kishner reduction12 of the dialkylated ketone 23 gave %3 in low yield. The NMR, IR and mass spectra were identical to that of 24 produced from catalytic hydrogenation of 26. This then proves that the structural assignments of 22 and 26 based on NMR data were correct, and that 22 is the major di-n-methane rearrangement product of 21 which is in agreement with all the other previous work. 14 The quantum yield for the sensitized formation of the photoproducts of hydrocarbons lg, IQ, and 21 as well as alcohols Z and 12 were determined in degassed benzene solutions using valerophenone actinometry. The conversionIx>photoproducts proceeded to less than 202 in all cases, while the conversion of valerophenone was less than 3%. Under these conditions the photoreduction reaction was suppressed. As can be seen in Table 1, acetone was in all cases the most effective sensitizer, whereas acetophenone was the poorest. Benzophenone does not sensitize the reactions. Hence the energies for the triplet states of z, 19, 1%, 16, and 2% lie somewhere between 72 and 68 Kcal/mol. 15 Table 1. Quantum Yields of Formation with Various Sensitizers No. Substrate Quantum Yields x 10 Sensitizer £2 R1 13 R1 k2 R1 1 R1 R2 ' CH 1.3 0.70 0.25 CH3, R - H 1.5 2 0.87 0.3 H, R = CH 1.2 0.83 0.22 OH, R = H 0.64 0.49 0.23 H, R2 = OH 0.41 0.29 0.1 0.67 0.34 0.12 0.28 0.19 A = Acetone; B a Diethyl Ketone; C = Acetophenone DISCUSSION Since the compounds under consideration do not have a stabilizing group directly attached to the di-n-methane system, it is difficult to understand why there should be any regiospecificity. In general, there are two ways in which this effect might occur. Either the R groups (R - 0H, OAc, Me) attached to the bridge carbon are exerting some type of sterically-directed energy transfer (vide infra) or the bridge R group, by some type of long range interaction, preferentially stabilizes one of the two diradical intermediates after energy transfer. Either theory could rationalize the observed regiospecificity but there are certain objections to each. Steric effects of energy transfer are known in the literature. However there is conflicting evidence as to their importance. Asymmetric induction in racemic acceptors of optically active sensitizers demonstrates the close contact made by the donor and acceptor during energy transfer.14 In agreement with this observation there is also an example of the photoequilibrium of £127 and transfstilbenes changing as more hindered 15 In addition there are also several ketone sensitizers are used. examples of hindered ketone and azo quenchers which quench the fluorescence and phosphorescence of various sensitizers at a lower rate than that of unhindered quenchers.16 In contrast several workers have determined the excited state life times of tert-aryl and alkyl ketones by standard Stern-Volmer quenching experiments without noting any steric effects to 16 17 energy transfer.17 Furthermore,(1gx-dimethylvalerophenone 33, a tertiary aryl ketone, underwent Norrish Type-II photoprocesses without any steric effects. Even though there is conflicting evidence on the importance of steric effects to energy transfer, an argument still could be formulated in which the excited sensitizer due to steric effects preferentially transfers its energy to one side of the molecule. For example, the bridge R group may hinder the approach of the sensitizer on the side of the molecule in which the R group is located. As a result of this steric hindrance the excited sensitizer transfers its energy to some less hindered location, a site on the molecule which is remote from the R group. If this preferential energy transfer in fact happened an explanation for the regiospecificity in these di-u-methane system could be made by postulating that the site of energy transfer favored one diradical intermediate over the other one. However, hindered energy transfer seems.unlikely in view of the fact that polymethylated (hindered) alcohols 36 and 21 show nearly the same degree of regio- specificity as do the unhindered alcohols 1 and 12.4 If steric hindrance to energy transfer was important varying the steric bulk of the acceptor 18 (or sensitizer) should affect the product ratios. The effect, however, was not observed. Another possible rationalization of the results also based on steric effects, involves a Schenck type mechanism.19 According to Schenck's scheme (eq 10) an electronically excited sensitizer (in this case acetone) forms a covalent bond with the acceptor. This initiates a rearrangement in which the sensitizer is eventually excised unharmed. OH H“ * —- o 3 a HSCJCLC"3 o 13 We, 1 (10) " 3 '30. 4—— 0 + 2.. H30 CH3 Although Schenck's original proposals have been disproved, there are some recent studies of reactions where this mechanism may be involved. In the Paterno-Buchi reaction, the photoaddition of an excited state carbonyl compound to an olefin, Yang proposed a biradical intermediate (eq 11) which may either ring close to an oxetane or cleave to regenerate the ground state ketone and isomerized olefin.20 This latter cleavage process is a Schenck type mechanism. A second example is found in the 19 R O Ir---u» R’ RI + | —-> ‘4 (11) :31. + “1. photoepoxidation of olefins (eq 12).21 Here Bartlett proposed that an electronically excited ketone attacks an oxygen molecule to form a diradical intermediate. This intermediate in turn attacks the olefin which is present, eventually leading to a ground state ketone and an epoxide. (Ego: In»: a ink+fi+ V20? (12) Despite these two examples there are several objections to a Schenck-type mechanism initiating a di-n-methane rearrangement. The first is the lack of literature precedent. The second objection is evident from examining eq 10. In the first step the excited state acetone molecule attacks the acceptor resulting in bond formation. 20 In order to explain the regiospecificity of the rearrangement, the attack must occur preferentially on one side of the molecule. For the sake of discussion, attack might be depicted to occur on the least hindered side, the side remote from the bridge R group (methyl).22 If this hypothesis were correct, attack would consistently occur on the remote side in all cases, in order to rationalize the observed regio- specificity. This would be regardless of the orientation (£12 or anti to the aromatic ring) of the bridge R group and regardless of the degree of methyl substitution on the bicyclic ring system. Consequently this explanation does not seem reasonable. A third objection is that different sensitizers with different steric requirements ought to give different isomer ratios of products. However, as is evident in Table 1, the opposite result occurred. Even though the quantum yields vary with different sensitizers, the ratios between major and minor products are invariant. In fact the decrease in quantum yield matches the decrease in triplet energy levels of the sensitizers. This result agrees with the accepted collisional transfer mechanism in which quantum yields decrease when the sensitizers' triplet energy levels fall below that of the acceptor. Thus since steric effects on energy transfer are unlikely, the only alternative is that some inherent feature in the bicyclic ring system itself is responsible for the regiospecificity after energy transfer. The interaction of lone pairs in the R groups (R = OH, OAc) may be part of the answer, especially in light of Marata's work23 (eq 13). Here the bridge R group (R = OH) is two carbon units removed from the di-w-methane system, yet there was only one photOproduct. When the hydroxyl group was replaced by an exo—methylene group the 21 regiospecificity was destroyed (eq 14). This suggested that a through space interaction with lone pairs on oxygen was an important effect. However, lone pair interaction cannot be the only answer since compounds 13, 16, and 21 which have only a methyl group(s) to destroy the symmetry, still react regiospecifically. This fact implies that other effects beside lone pair interactions are important. 22 The only type of interaction that has not been discussed is an inductive effect by the bridge R group. But even with this hypothesis there are difficulties. First, a methyl group is inductively electron-donating whereas the hydroxyl and acetate groups are inductively electron-withdrawing. ‘It is not obvious how two opposing interactions could favor the same diradical intermediate in the Zimmerman formalism. Second, any inductive effect would have to be operating over two or more carbons but the strength of inductive effects is known to decrease rapidly with distance. Recent work by Houk on a related system may serve as a model for explaining the observed regiospecificity. In Houk's case (eq 15) an electron donating group in the meta position favored initial meta bridging (meta to the donor group) while 3 meta acceptor group favored (15) initial para bridging. When the R group was in the ortho position (eq 16) initial ortho bridging was favored in all cases whether R was an electron donor or an electron acceptor. The explanation for these. effects comes from frontier orbital theory. 23 ‘50... .10 \ R ' (16) The compounds that Houk studied were approximated by allowing an ethylene unit and the substituted benzene orbitals to interact. Figure 1 shows the resulting HOMO and LUMO orbitals with their co- efficients for an ortho substituted aryl ring. In the case of an acceptor substituted benzene, the large coefficient on the lobe ortho JR LthflC) I) 3 05 4 .31 .52 1.32 .43 0.40 .37 -4 ~34 -37 .4 o .44 135 .05 HOMO D .19 .50 .19. .45 .38 4 ' .50 .21 o .38 .51 .15 Figure l. Singly occupied MD's of donor-acceptor substituted benzene after interaction with an ethylene unit. 24 to the acceptor group in the LUMO favors initial ortho bridging. In the donor substituted case, the large coefficient ortho to the donor group in the HOMO again favors ortho bridging. Applying Houk's theory to the compounds under investigation could provide an explanation for the observed regiospecificity, especially forty: and 16 which have no lone pairs. Although no calculations were done a qualitative description may be useful. The problem essentially is to explain why initial bridging occurs at the site remote from the bridge R group (Fig 2). This is most easily done by considering the Figure 2. Two possible diradical intermediates. compounds studied as a pair of substituted benzenes and olefins (Fig 3). R' and R" represent the points where the aromatic ring and olefin unit .1 FT I . .3 Figure 3. The di—w-methane system as a substituted benzene and ethylene. 25 are attached to the remainder of the bicyclic ring system. R' is the point of attachment nearer the bridge R group (R - OH, OAc, CH3) while R" is the point of attachment farther from the bridge R group. If R' is either slightly more negative or positive than R" due to its proximity to the bridge R group, the HOMO and LUMO orbitals may be polarized in a similar manner as in Figure 1. Although the coefficients of the various lobes probably will not be the same as in Figure 1, there is still the possibility of a larger coefficient ortho to R' in either the HOMO or LUMO (depending on whether R' is an electron donor or acceptor) in analogy to the ortho substituted molecules (Fig 1). Thus initial bridging would be favored ortho to R', that is, away from the bridge R group. An alternate way to describe this bridging is to say that the unpaired electrons of the diradical intermediate are on the same side of the molecule as the bridge R group (which was how the di-w—methane rearrangement was described in all previous work). So instead of a through space stabilization of a formal diradical intermediate, the bridge R groups by a through bond inductive effect favor initial ortho bridging (diradical A, Fig 2) which then leads to the observed products. Thus this proposal rationalizes the observed regiospecificity in these systems. In order to test this proposal, however, frontier orbital calculations would have to be employed. EXPERIMENTAL A. General NMR spectra were measured on a Varian model T-60 with TMS (tetramethylsilane) as an internal standard. Infrared spectra were recorded on either a Perkin Elmer 167 or a Unicam SP-200 grating spectrophotometer, calibrated with a polystyrene film. Mass spectra were obtained on a Hitachi-Perkin Elmer RMU-6 at 70 ev operated by Mrs. Ralph Guile. Elemental analyses were performed at Spang Microanalytical Laboratories, Ann Arbor, MI. All preparative VPC (vapor phase chromatography) separations were carried out with a VarianrAerograph MOdel 90—P chromatograph. A11 analytical VPC separations were carried out with a Varian-Aerograph Series 1400 chromatograph with a flame ionization detector. Integrations of analytical VPC traces were accomplished with a Disc Instruments, Inc., Disc Integrator. The following VPC columns were used: A) 5 ft x .125 in. column of 20% FFAP (Analabs, Inc.) on 80/100 mesh Chromasorb W, B) 10 ft x .25 in. column of 20% FFAP on 80/100 mesh Chromasorb W, C) 5 ft x .125:h1.column of 5% SE-30 on 60/80 mesh Chromasorb G, D) 10 ft x .125 in. column of 10% FFAP on 80/100 mesh Chromasorb W, E) 10 ft x .25 in. column of 20% SE—30 on 80/100 mesh Chromasorb W. 26 27 B. Quantum Yields Quantum yields were determined with valerophenone actinometry25 and a standard merry-go-round apparatus fitted with a Hanovia 450 W mercury lamp in a water cooled quartz immersion well. The 300-313 nm band was isolated with Pyrex and a 1-cm path of .002 M potassium chromate in 5% aqueous potassium carbonate. The conversion of valerophenone was limited to less than 3% while the conversion of the substrates was limited to less than 20%. C. Materials (for quantumgyield studies) Thiophene-free benzene was stirred over reagent grade sulfuric acid for three days. After the benzene was washed with water and was dried over magnesium sulfate, it was distilled from phosphorus pentoxide through a 40-cm vacuum-jacketed column packed with glass helices. Spectrograde acetone (Mallinckrodt) was used without further purification. Diethyl ketone, valerophenone (Aldrich), and acetophenone were passed through alumina and twice distilled under reduced pressure through a 10-cm vacuum-jacketed vigreux column. Benzophenone was recrystallized 3 times from ethanol. Compounds 1, lg, 13, 16, and 21 were twice collected with VPC using column 3. Analytical VPC of the collected materials, using column D, gaveaisingle peak in all cases. 28 D. Procedure (for quantum yield studies) The materials to be photolyzed were weighed out in volumetric flasks. An internal standard (for valerophenone, 13, 16, and 21, mm mm mm pentadecane; for Z and 12, docosane) and sensitizer were added and the contents diluted to the mark with benzene. The solutions were transferred with a 9-in. pipet into separate, constricted, 13 x 100 mm Pyrex test tubes. The samples were degassed by 3 freeze-pump-thaw cycles and sealed under vacuum at < .005 Torr. E. Preparation of 9,9-dimethyl-l.4-dihydro-1,4-ethanonaphthalene($1) To 70 mL of 1,2-dichloroethane was added 3.24 g (30 mmol) of 5,S-dimethylcyclohexa-l,3—diene, 5.87 g (32 mmol) of benzenediazonium— 2—carboxylate-HC126 and 4.64 g (80 mmol) of propylene oxide. The magnetically stirred solution was then refluxed for 2 h. After the mixture was cooled, and solvent removed, the dark red oil was eluted through a short alumina column with pentane. The colorless oil thus obtained (3.9 g, 71%) gave three peaks with VPC using column A at 130° C, in an area ratio of 3:2:1. Chromatography with silica gel (Woelm, > 230 mesh) using hexane as the eluent followed by preparative VPC with column B at 1400 C, (He flow rate 50 mL/min) of the glfenriched fractions yielded pure 21 as the major product: NMR (CC14) 5 6.83 (s, 4, aromatic), 6.38-5.93 (m, 2, olefinic), 3.76-3.54 (m, l, bridgehead), 3.30-3.15 (m, l, bridgehead), 1.12 (m, 2, CH2), 1.30 (s, 3, EEEEfCHB): 0.60 (s, 3, gyE¢CH3); IR (neat) 3050 (m), 3020 (m), 2950 (s), 2865 (s), 1600 (w), 1470 (m), 1457 (m), 29 1382 (m), 1361 (m), 1343 (m), several weak bands from 1240 to 760, 750 (s), 720 (s), 680 cm'1 (5); UV (cyclohexane)_ Amax 226 nm (e = 1700), 252 (450), 257 (540), 264 (660), 271 (630); mass spectrum, m/e (rel intensity) 184 (l), 128 (100). Anal. Calcd for C14H16: C, 91.25; H, 8.75. Found: C, 91.08; H, 8.78. The second component (22) collected with column E at 135° C (He flow rate 50 mL/min) had spectral data identical to known8 l-phenyl- 4,4-dimethylcyclohexa-2,5—diene: NMR (CC14) 6 7.03 (s, 4, aromatic), 6.48 (s, 4, olefinic), 3.73 (s, l, benzyl), 1.13 (s, 3, CH3), 1.08 (s, 3, CH3); IR (neat) 3060 (w), 3020 (m), 2950 (s), 2860 (m), 1600 (w), 1490 (m), 1465 (m), 1450 (m), 1370 (w), 1355 (m), weak absorptions from 1170 to 910, 853 (s), 735 (m), 760 (m), 740 (m), 730 (s), 695 curl (3); mass spectrum m/e (rel intensity) 184 (42), 169 (100). The third component (2%) collected with column E at 135° C (He flow rate 50 mL/min) had the following spectral data: NMR (CC14) 156.90 (bs, 4, aromatic), 5.84 (dd, 1, J = 10 Hz, J - 3 Hz, olefinic), 5.45 (d, l, J = 10 Hz, olefinic), 3.78 (m, 2, bridgehead), 1.70 (m, 2, CH2), 1.02 (s, 3, CH3), 0.78 (s, 3, CH3); IR (neat) 3064 (m), 3016 (s), 2956 (s), 2920 (s), 2865 (s), 1600 (w), 1489 (w), 1468 (m), 1453 (s), 1395 (w), 1373 (m), 1360 (m), many weak bands from 1280 to 780, 748 (m, sh), 740 (s), 720 (s), 697 cm."1 (m); UV (cyclohexane) Amax 223 nm (e - 1600), 247 (1600), 252 (2100), 266 (2300), 272 (1900); mass spectrum m/e (rel intensity) 184 (22), 128 (100). ‘Anal. Calcd for C14H16: C, 91.25; H, 8.75. Found: C, 91.03; H, 8.93. 30 F. Preparation of 9,9-dimethyl-1,2,3,4-tetrahydro-l,4-ethanonaphthalene (54’) To 10 mL of methanol was added 105 mg of 21 and 10 mg 10% Pd-C. The mixture was stirred under one atmosphere of H2 until one equivalent had been consumed. After the solution was filtered and the solvent was removed, preparative TLC (alumina with hexane eluent) of the residue yielded 56 mg of product (24): NMR (CC14) 6 6.90 (s, 4, aromatic), 2.88 (m, l, bridgehead), 2.38 (m, l, bridgehead), 2.13-1.30 (m, 6), 1.16 (s, 3, a3517CH3), 0.56 (s, 3, gyngH3); IR (neat) 3070 (m), 3040 (m), 3020 (m), 2940 (s), 2860 (s), 1485 (s), 1460 (m), 1382 (m), 1364 (m), several weak bands from 1330 to 1250, 1183 (m), 1168 (m), 1130 (m), 1120 (m), 1100 (m), 1042 (w), 1025 (m), 932 (w), 910 (w), 750 cm."1 (3); mass spectrum m/e (rel intensity) 186 (13), 130 (100). ‘Anal. Calcd for C14H18: C, 90.26; H, 9.74. Found: C, 90.17; H, 9.82. G. Preparation of 7,7-dimethyl-2,3—benzobicyc1o[3.2.1]octen-6-one (33) In 3 mL of dry THF was dissolved 140 mg (0.81 mmol) of EEEEF7' methyl-2,3-benzobicyclo[3.2.1]octen-6-one (22).5 The solution was slowly injected under nitrogen at room temperature into 4 mL of THF containing 200 mg of 27.6% KB. The magnetically-stirred mixture slowly evolved hydrogen for approximately 0.5 h. Upon the addition of 570 mg (4.0 mmol) methyl iodide, a white precipitate immediately formed. After the mixture was stirred an additional 0.5 h, the reaction was carefully quenched with water. The solution was extracted with ether, and dried 31 over magnesium sulfate. Evaporation of the solvent and TLC of the resultant oil (silica gel with 10% ether/hexane eluent) gave 112 mg, 69%, of product (3}): NMR (CC14) 5 7.07-6.66 (m, 4, aromatic), 3.28-1.73 (m, 6, aliphatic), 1.13 (s, 3, antifCH3), 0.59 (s, 3, syn-CH3); IR (neat) 2950 (s), 1725 (s), 1495 (m), 1480 (m), 1420 (m), 1385 (m), 1365 (w), 1345 (w), 1290 (m), 1160 (m), 1085 (m), 1020 (m), 970 (w), 895 (w), 885 (w), 850 (w), 775 (s). H- Preparation of 7,7-dimethyl-2,3-benzobicyclol3.2.1]octane (31) To 2 mL of triethylene glycol was added 112 mg (0.56 mmol) 33, 100 mg KOH and 64 mg anhydrous hydrazine. The mixture was heated at 2000 C for 5 H. The solution was cooled, poured into 6N HCl and extracted with ether. After the ether extract was dried over magnesium sulfate and the solvent removed, TLC (alumina with hexane eluent) yielded 16 mg of product (34): NMR (C014) 6 6.97-6.63 (m, 4, aromatic), 3.23-1.16 (m, 8, aliphatic), 1.10 (s, 3, anti-CH3), 0.63 (s, 3, syn-CH3); IR (neat) 2960 (s), 1500 (m), 1480 (m), 1468 (s), 1426 (m), 1398 (m), 1376 (m), many weak bands from 1340 to 840, 832 (m), 776 (s), 758 (s), 734 cm.1 (s); mass spectrum m/e (rel intesity) 186 (16), 129 (100). ‘Agal. Calcd for C14H18: C, 90.26; H, 9.74. Found: C, 90.24; H, 9.73. 32 2,8 1. Hydrogenation of 6,6-dimethyl-3,4-benzotricyclo[3.3.0.0 ]oct-3-ene (g2), Catalytic hydrogenation of 60 mg of 22 in methanol with 10% Pd-C yielded, after 1 equivalent of hydrogen was absorbed, 40 mg of product. Analytical VPC with column C at 1250 C indicated only one peak.27 The spectral data were identical to that of 34. J. Prep arativephotolys is of 9J 9-dimethyl-1, 4-dihydro-lL4-ethano- naphthalene (21), Compound 21 (224 mg) was dissolved in 140 mL of spectrograde acetone and the solution was purged with nitrogen. Photolysis with a 450 W Hanovia mercury lamp with a Pyrex filter was monitored using analytical VPC with column D. After most of the starting material was consumed, the photolysis was stopped and the solvent removed. Preparative VPC with column B at 1550 C (He flow rate of 50 mL/min) gave in addition to a trace of starting material 21.4 mg of 24, 25.6 mg of 25, and 51.7 mg of 26. Compound 24 had the same spectral characteristics as an independently prepared sample (vide supra). Compound 25 had the following spectral data: NMR (CC14) 6 7.03-6.33 (m, 4, aromatic), 3.60-3.33 (m, 1), 3.02-2.67 (m, 1), 2.33-1.95 (m, 2), 1.67-1.33 (m, 2), 1.12 (s, 3, anti-CH3), 0.48 (s, 3, syn-CH3); IR (neat) 3070 (m), 3040 (m), 3020 (m), 2950 (s), 2920 (a), 2860 (s), 1585 (s), 1520 (s), 1382 (m), 1362 (m), 1330 (m), several weak bands from 1312 to 1220, 1183 (m), 1164 (m), 1130 (m), 1120 (m), 1100 (m), several weak 33 bands from 1035 to 900, 750 cm"1 (3); mass spectrum m/e (rel intensity) 184 (41), 128 (100). Compound 26,had the following spectral data: NMR (CC14) 5 7.03 (m, 4, aromatic), 2.93-2.70 (m, 2), 2.36-1.50 (m, 4), 1.19 (s, 3, anti- CH3), 0.76 (s, 3, syn-CH3); IR (neat) 3040 (m), 3020 (m), 2950 (s), 2860 (m), 1470 (m), 1380 (m), 1360 (m), 1290 (m), several weak bands from 1370 to 875, 780 (m), 755 (m), 740 (s), 725 cmfl (m); mass spectra m/e (rel intensity) 184 (36), 128 (100). PART II THE IONIZATION OF TWO TETRACYCLIC GEMFDIFLUORIDES UNDER STABLE ION CONDITIONS INTRODUCTION The study of carbocations is one of the most intensely in- vestigated fields in organic chemistry. This is not only because such species are intermediates in numerous organic reactions but also because they are a subject of considerable theoretical interest. One area of interest and controversy deals with the existence of nonclassical carbonium ions.28 The crux of this problem is illustrated in Figure 4. Here the 2-norbornyl cation 42 may be viewed either as a set of rapidly equilibrating localized (classical) secondary cations or it may be viewed as a static delocalized (non-classical) two-electron, three- center cation in which the charge is distributed through the o-bonds. WL‘EENWL Figure 4. The 2-norborny1 cation represented as an equilibrating pair of classical ions and as a nonclassical ion. In addition to the 2-norbornyl cation, many other types of nonclassical ions have been suggested. For example, in 1972 based on Extended HUckel Calculations, Stohrer and Hoffmann proposed that the pyramidal 35 36 structure :4 may be a true energy minimum for the (CH)5+cation.29 Formally this pyramidal cation may be viewed as a positively charged methine carbon bonding to the face of a cyclobutadiene molecule. The alternate view, of course, is that the cation is actually a set of rapidly equilibrating classical ions. Since then a number of +1 11 " I :9, ’0“ +1 I C: -—~Ior .==|==e«~. 4h! -1 Figure 5. The pyramidal nonclassical (CH)5+ ion. examples of pyramidal ions, 45,30 46,31 47,32 48,33 49,34 50,35 51,36 52.36 mm mm mm mm mm mm mm mm have appeared in the literature (Figure 6). Evidence for the existence of these ions is as follows: 1) deuterium label is scrambled equally among the basal carbons, 2) the NMR (Nuclear Magnetic Resonance) spectra of these ions indicate a symmetric structure, 3) the apical carbons appear at an unusually high field in the carbon-l3 NMR spectra; for ions 45, 47A, 51 and 52 their resonances occur above TMS (tetramethyl- 'V\: 'VV'U 'VM ’Vb silane). This latter observation is in agreement with Stohrer and Hoffmann's prediction that the apical carbon possesses little if any positive charge. It further suggests that these ions indeed have a nonclassical structure since it is difficult to rationalize this high field NMR resonance with a set of equilibrating classical ions. The work of Hart and Kuzuya provides additional insight into the 37 46 49 51 Examples of known pyramidal cations. Figure 6. 38 factors which govern the stability of pyramidal ions.35’ 37 Ion 46, which was prepared by treating alcohol 53 with strong acid (eq 17), is a reasonably stable carbonium ion which can be warmed to -40° C before irreversible rearrangements occur. .éipriori one would predict that methyl or phenyl substituents would stabilize this ion by dissipating the positive charge. However, the opposite result occurred. Ion 20 is stable only below —1000 C and ion 52 proved impossible to prepare, even at -1300 C. Studies of a model indicate that the phenyl groups in 50, are twisted out of the basal plane due to steric crowding. Consequently any stabilizing resonance interaction with the basal carbons is hindered and the ion is destabilized by the inductive effect of the phenyl groups. The inability of ion 52 to form is somewhat more puzzling since ions 45 and 41% do possess apical methyl groups. One possible explanation is that upon ionization of alcohol 52 the methyl group by electron release lessens the demand for cyclopropyl participation (delocalization) and other rearrangements then become feasible. (Another possibility is that since the apical carbon may in fact be slightly negative, methyl substitution at this position may actually be destabilizing. Since ion 45 is not polymethyl substituted and since ion 47% is doubly charged, perhaps these ions can better tolerate apical methyl substitution than ion 56 because the apical position, in these two cases, feels a greater effective positive charge from the basal carbons. The fact that the substitution of an apical methyl group for a more electron donating isopropyl group in ion 47% destabilized the system supports this hypothesis.32d If there is indeed a slight negative charge at the apical position, perhaps in contrast to an electron donor, an electron 39 withdrawing substituent like fluorine would stabilize the ion. Not only would this be the first example of a pyramidal cation with an apical fluorine but it would also provide an explanation for the instability of cation 22. The problem then is to prepare suitable pyramidal ion precursors which can be ionized under stable ion conditions. Recently Jefford discovered that when the 7-position of the 40 norbornadiene skeleton is blocked by di-alkyl substitution the normal exo-l,2 addition of a carbene to the double bond is sterically hindered. As a result of this steric hindrance, the carbene is forced to attack in an unusual endo-1,4 manner (eq 18).38 Compounds 58 and 52, which are the products of 1,4 endo attack by difluorocarbene on bornadiene 57, are potential precursors to a fluoro substituted pyramidal carbonium ion (eq 19). 'A'priori there are three possible reactions which could occur after the abstraction of a fluorine by SbFS from 58 (or 52). One possibility is that nonclassical delocalization with the cyclopropyl group would lead directly to the tris-homocyclopropenyl cation 62. Another possibility is that a fully delocalized pyramidal cation 22 would be formed, or finally, as in the case of alcohol 55, the molecule could rearrange to form a non- pyramidal cation. In addition to being synthetically available there are several other reasons for selecting these compounds for study. First, both 52 and 52 are expected to form the same pyramidal ion. Second, since a pyramidal ion like 63 has the same skeleton as the ions studied by Hart and Kuzuya, further insight may be gained into the charge distribu- tion and other factors that govern the stability of this ring system. 41 (18) RESULTS Compounds 58 and 52 were synthesized by the method outlined by Jefford.38 The mixture could be separated by VPC (vapor phase chromato- graphy) with column A. The proton NMR spectra of these compounds agree 38 In Table 2 the carbon-l3 chemical with those published by Jefford. shifts of 58 and 59 are compared with those of 64, the adduct of ’Vb ’V'V ’Vh difluorocarbene to norbornadiene. In 22 the chemical shift and carbon- fluorine coupling constants for the difluoro substituted carbon agree with those found in adduct 64. Furthermore the chemical shifts of the 'Vb 1,2 positions in 22 are in fairly close agreement with the chemical shifts of the 1,2 positions in adduct 64. Unfortunately the difluoro substituted carbon in 58 was not observed because its signal is inherently weak due to the lack of Nuclear Overhauser Enhancement and the fluorine coupling. For compounds 22 and 64 the difluoro carbon signal was barely observable. However the adjacent carbons 3 and 5 in 58 have nearly the same chemical shifts and fluorine carbon coupling constants as the corresponding 3 and 5 carbons in adduct 64. The remaining carbon positions in both 28 and 22 are affected by the a and B deshielding effects of the methyl groups and have different chenical shifts from the corresponding positions in the adduct 64. W The carbonium ions were generated by adding under a nitrogen atmosphere at ~95° C (acetone slush) a solution of SbFS in FSOZCl to an 42 43 .momosucoumo a« mum .Nuuom cw .AmIUhv honeymoon wcfiaosouu .uoo>aom occuoom ououoov ca .mZH aoum nowaawa use momma ow mum mumwsm malnooumon .pmm .momm 3.83 an m AH.ev Am.mv Aav Aav Am.mav Ae.emuv Am.mav Aa.HHv Aa.HHV w.eH e.ms e.wm o.me o.~e m.mm m.m~a m.mm e.m e.a on Am.HV Am.mv Am.ev Am.ev Ao.HHv Am.o~V Am.o~v Ao.mv Ae.mv N.OH e.aa m.ee ~.me m.ee w.me . m.me m.HN o.am.a~ 3 AuNNV a a Aomv AoomV Aomv « . ea.m~ an.Hm wa.mm mH.me H.w~H oa.me H¢.~H mam.~a a a: we a: so a: Nu m A o m e m N H easemaoo umoasz sonumo «w .ww .WN monsooaoo mo mumwnm Hmoflaono maloonumo N «Haas 44 NMR tube containing either 58 or 5% frozen at -130° C (pentane slush). After addition of the acid solution, the NMR tube was then warmed to -95° C in order to complete ionization. A clear red-orange solution resulted in each case. The NMR spectra were recorded between -850 C to -950 C. This procedure generally afforded clean ion solutions with minimal decomposition. It is obvious from the carbon-13 spectra (Table 3) of ions 65 and 66 that the same ion was not formed from 58 and 52. Furthermore since both the pyramidal ion 62 and the tricyclopropenyl ion 62 have a plane of symmetry and only 8 peaks are expected in their carbon-13 spectra neither 6; nor 66 can be either of these ions since they both have 11 peaks in their spectra. A mixture of 58 or {a ionized to give a mixture whose spectrum was a composite of the spectra for 65 and 66; Thus, it is clear that 58 and 52 ionize in different ways to give two different ions. The structures of the ions were assigned from their spectral data and from an analysis of their quenching products. The evidence for each structure will be discussed separately. Later, the reasons for the different ionization behavior of 58 and 52 will be discussed. A. The Structure of Ion 65 For ion 62 it is obvious from the carbon-l3 chemical shifts that the positive charge resides primarily on two carbons. This deshielding is reminiscent of an allyl type ion where the charge is primarily on the terminal carbons. However, the exact ring structure of ion Qz‘was not obvious from the carbon-13 spectrum. Furthermore as can be seen in 45 .oq .mmmo .mwmonuamumo aw mum .Nuuom aw .Amlohv musmumaoo wawaosooo .zumaawomo Hmsuouxm :H mZH aoum .sowHHHa use muumo cw mum mumwnm Hmowsonoo .am .womm 8.8 C.o$£.em:m.~w38.2V 8.3V 8 9mm 5.: 93 as New oi: «.32 as: meme TE .23 a u an. a n 98 Name ~62 «.9: he: Ne: ~53 e~.~2 .6. an 8.3 8.8 :88 3.3: 3.28 3.3 w.o~ a.e~ o.em o.om ~.eea m.wme m.Ha a.oe~ e.HNH H.mH~ a.ae.ee so u v - e.oe e.aa e.aH m.om m.NH~ H.eme m.~e~ am.om azwmum as now as us a: on m a e m e m N H eaaoeaoo Hmoasz :onnmm chH Home: mo omocH cows woummaou ww mam WW msOH anaconumoouosam mo muuomom manconuoo m edema 46 Figure 7 the proton NMR spectrum was surprisingly complex and essentially uninterpretable. The clue to this ring structure was obtained by quenching the ion solution in either a saturated solution of aqueous sodium bicarbonate at 00 C or in a solution of sodium.methoxide and methanol at -78° C (eq 20). In either case ketone 92 was obtained. The methanol quench gave a cleaner reaction mixture (by NMR) and a higher yield of ketone 23. Some care was taken to prove the structure of 63 since it provides chemical evidence for the structure of ion 65. The structure proof rests on spectral data and subsequent chemical transformations. The IR (infrared) spectrum of 22 had a carbonyl absorption at l 1670 emf which is in agreement with the 1680 cm“1 absorption for the known parent ketone 12. Also the mass spectrum of 22 gave the expected m/e 162 value for parent ion. The carbon-13 NMR spectrum assignments of 22 (Table 4) were made by comparing them with the chemical shifts of the parent ketone 7241 and by off resonance decoupling. Similarly, the corresponding proton NMR spectrum assignments (Table 4) were made by comparison with chemical shifts of the parent ketone 12, by the spin multiplicities of the peaks and by a europium shift study (Figure 8). 47 5 l ACETONE—d 5‘0 ppm Figure 7. Proton NMR spectrum of ion 55. ‘48 ode anode u.sac .uouv .snououoanoououaov ow mxu ou voonououou .souanul.uoe ounce one savan- «sou-onus .oouonuaoutm a“ out nuuom a“ any sue-unsou usuaesoo can sowuauwnewuaaxo .mzu ou oousoueuou .sowaafil use sauna nu Head-onus " 1 7 an «42 «.3: 13 3.5 0&2 ~62 24.... So a 2 o NOoMQOofiU AO.M.Oo0 AocfinooOV AO.O.O¢°H.V E 3.3 3.3 3.3 3.3 3.3.3 3 ! en.~ n~.a 33.3 n~.n NM.“ ee.n en~.n n a n.- .o.n~ «o.- n.dn m.~m~ o.mc« a.wn o.¢o~ n.o- o.¢aa um.oo mac 0 an.n.n.nu na.n.n.nv 3 3 3.13 3.3. 3.3 3 3.15 _ 3 on; 84 as; 3.“ «.3 3d 2.“ .38.“ a; Ammuvmw Ammuwmm 1Anmuv¢ 1 m s o muonrnz ammuwo c m N H Aww1vssomsoo Hove: can «a osouox now some n~1oooueu use nououm c manna 49 Figure 8. Europium shift slopes for the protons in ketone m2. Comparison of the assigned chemical shifts of $2 with those of 79 shows that there is reasonably good agreement between the two systems. However, in the proton NMR spectrum the chemical shift differences between the bridgehead protons H-1 and H-5 in the two systems are rather large. The chemical shifts of these protons for ketone 32 are 0.57 and 0.78 ppm higher field than those for the corresponding protons in 72. Since other bicyclic[3.2.1] ketones42 show bridgehead proton chemical shifts which are approximately the same as those for 72, the high field absorptions found for ketone 62 are indeed puzzling. Perhaps these high field bridgehead absorptions are the results of shielding by the C-8 methyl groups. These chemical shift differences made it desirable to perform some chemical transformations on 22 in order to put its structure on firmer ground. These chemical transformations are‘outlined in Scheme 2. In the NMR spectrum the proton absorption at 6 5.13 and the methyl group Scheme 2 x 00.00 / "—‘aa / . 30m D 0 o 69 59' 1)MsLi 2)H30 1r - -H20 Fso,H _ I I 95'C a I O _ .J 71 12 absorption at 5 1.88 disappeared upon base-catalyzed deuterium exchange. This clearly indicates that these protons are attached to the 5,3-unsaturated carbonyl fragment of the molecule. Further proof for the structure of 22 was obtained by adding methyllithium to 62 followed by elimination of water to give the 252-methylene compound 71. Evidence for the structure of 11 is found in the absence of a carbonyl absorption in the IR, in the mass spectrum with an expected m/e 160 value for the parent ion, and in two new olefinic proton NMR signals at 6 4.60 and 6 4.40 The chemical shift assignments for olefin 11 were made by comparison with the proton shifts of ketone 62. 51 .Naomno ea anemone ma sown: Assn mm.mv Naoaom cu noosouowomo .mmmonuaoumo ca mum mono: cw va musmuwooo onHosoo new mofiuaowaofiuaozn .Nmu ow mSH ou pooaouowou .soHHHHa poo momma ow omega Hmoaaosum 3 3 3 E 3 E 3 E 3 an o~.H @q.H «m.c am.o no.m mm.~ mH.n mm.~ unm.m \ uh 88.3 88.3 83 83 c3 c3 83 88.3 88.3 23 A oo.¢ 08.8 no.6 mo.H om.m ~m.m oH.N mn.H wN.m n.Mm¢.N 83.33365 8 we a a m e m 8 8 2888 ww GOH anacooumu wswoaoowouuoo sou new ww mo muuooom mzz sououm m memH 52 When 11 was added to fluorosulfonic acid at -95° C ion 3% was formed. The symmetry of ion Z2 is clearly indicated by the proton NMR spectra, which contained only six signals. The chemical shifts for the allyl methyl groups, 5 2.88 ppm and the allylic methine proton, 6 7.17 ppm, are comparable to those of other allyl systems.43 These results clearly establish the structure of $2, the quench product of ion Q2» and consequently provide additional chemical support for the ring structure assigned to QQ. Once the ring structure for Q; was established the assignment of the corresponding carbon-13 chemical shifts is fairly straightforward (Table 3). That C—2 (6 218.1) held a fluorine substituent was clear from the C—F coupling. To C-4, the other terminus of the allyl system, was assigned the 6 240.9 peak. The shifts of C-2 and C-4 also compare favorably with those of other cyclic allyl cation systems.43 Furthermore, that more of the charge is located on C-4 than on C-2 (6 240.9 vs 6 2.18.1) is consistent with the methyl group being somewhat more effective than the fluorine in stabilizing the positive charge. An off resonance decoupling experiment did not further split either of these peaks showing that these carbons have no protons. By examining the magnitude of the long range fluorine-carbon coupling constants the chemical shifts for the other positions could be assigned.44 Due to its proximity to fluorine, C-3 was assigned the peak at 6 121.6 (ZJC—F = 14.8 Hz), because its fluorine—carbon coupling constant is larger than that of the other peaks in the olefinic region of the spectrum. The chemical shift for C—3 is approximately 15 ppm higher (upfield) than for the corresponding center carbons of 53 other allyl ions.43 This is probably the result of the B-shielding effect of f1uorine.453’ 46 With similar reasoning (vide supra) C-7, which is in closer proximity to fluorine than C-6, was assigned the split 6 144.2 peak (3JC-F = 6.0 Hz) whereas the unsplit 5 138.8 peak was assigned to the more remote C-6 position. Off resonance proton decoupling did not further split the 6 90.0 (3JC-F = 5.0 Hz) peak so this absorption corresponds to C-8. Since the 5 64.6 (ZJC-F = 5.4 Hz) peak is split by fluorine this absorption was assigned to C-1 whereas the unsplit 6 71.8 peak was assigned to the more remote C-S. The absorption at 5 34.0 is typical for a methyl group on the terminus of an allyl system.and is therefore the C-4 CH3. Since gyErC-B CH3 is over the allyl ion system it corresponds to the lower field 6 27.9 peak, while the anti-C-8 CH3 which points over the W system corresponds to the 5 20.8 peak. B. The Structure of Ion $6 The clue to the homotropylium structure for 22 came mainly from the high field proton NMR peak of the endo methyl group at 6 -0.10 and the peak from the exo—methyl group at 6 2.43 (Table 6). The high field endo group absorption along with the large endo-exo chemical shift difference is characteristic of the homotropylium system,388’ f and has traditionally been used as evidence for a ring current. This is evident in the parent homotropylium ion 68 where the gndgfproton resonance occurs at 5 -0.73 ppm while the exo—proton resonance is much farther down field at 5 5.13. Additional support for a homotropylium structure is found 54 47 Here the in the spectrum of the 8,8-dimethy1 homotropylium ion. endgfmethyl group shift is at 5 -0.48 ppm.while the gngmethyl group shift is at 5 2.36 ppm in close agreement with the shifts observed for 66. mm Furthermore, it is evident from the carbon-13 spectrum of 22 (Table 2) that the charge in this ion is much more delocalized than the charge in the allyl ion 22. The most deshielded carbon in 66 is the one bearing the fluorine with a chemical shift of 6 174.6, which is 43.5 to 66.3 ppm higher field than the charge-bearing allyl carbons (6 218.8 and 6 240.9) of 62. Another indication of the charge delocalization in 62 can be seen by noting that the fluorine substituted carbon in 66 is only 10.8 to 13.5 ppm more deshielded than the corresponding carbons in fluorobenzene,44 (5 163-8) and 2-fluoropropene (6 161.1).48 Thus both the proton and carbon-l3 spectra indicate an extensively delocalized system, like the homotropylium ion, for cation 66. The problem then is to locate the positions of the methyl and fluoro substituents on the ring. This was accomplished by examining the proton NMR spectrum (Table 6). Since there are two proton absorptions at 5 5.01 and 5 4.80, these two peaks probably correspond to the H-1 and H-7 positions on the homotropylium ring. Therefore the fluorine and methyl group can only be on either the H—2, H—3, H-4, H-5, or the H—6 positions. Since the methyl group is split by fluorine (the methyl doublet does not collapse when the other proton frequencies are de- coupled), the methyl group is probably adjacent to the fluorine. With these restrictions there are only four possible isomers for the homotropylium ion. A further restriction is placed on the system by 55 .mumaawnmo Hmsuouxo cm a“ couscoumom ca uaomoum ma Loans Aaem me.~ my mvlosoumom ou bouzouomomo .momonuaouma ca .Nuuom ca .mucmumsoo wswamsoo onu suaa sowaafia non manna ca mum mumwzm Hmueswnon .maq .momm CK .m.$ 36H .~.o.3 CK 8.53 8 A3 A3 :6 .3 c3 3 .3 86 .3 A3 3 .3 m mc.~ oa.on Ho.m mc.m oo.m ow.~ m¢.w oow.< 8.8 Aw.a .~.nV AH.OHV AH.OH .o.mv Aq.m .o.mv A~.n .¢.NV ma.m mn.0l wq.o mm.w we.» -.w m¢.w am.w n.mwQ.e No w as w m o m e m N H umnasz :ououm :oH Hobo: m mo umne sows woumaaoo WW coH mo asuuownm mzz aououm och o manna 56 noting that the 6 8.00 peak is a doublet of doublets. The larger coupling constant J - 18.6 Hz is probably due to hydrogen-fluorine splitting. Support for this hypothesis comes from the fact that the largest J value for the parent homotropylium system is only 10.1 Hz (Table 6). Therefore if the 6 8.00 peak is indeed split by a fluorine and an adjacent proton, only structures 66, 73, and 74 are possible. mm mm mm However, once the 6 8.00 ppm peak is assigned to a proton adjacent to the fluorine, it becomes very difficult to rationalize the observed coupling constants and chemical shifts with either ion 73 or 14. H 73 14 Structures 39 andlze were conclusively eliminated by the following decoupling results. When the frequencies at H-l or H—7 were irradiated the 6 8.00 doublet of doublets peak does not collapse. Conversely decoupling at 6 8.00 does not collapse either the H-1 or the H—7 peaks. Therefore since the 6 8.00 proton can not be coupled to either of these protons structure 33 is eliminated. When the 2 proton multiplet at 6 8.45 was irradiated both H-1 and H—7 collapsed to doublets (J - 4.3 Hz) which implies both of these protons are coupled to other ring protons which is inconsistent with structureizé. These decoupling results are however consistent with structure 66. The fact that Bel and H-7 collapse only to doublets upon irradiation at 6 8.45 indicates that either H-1 and H-7 are coupled to each other or H-1 and H—7 are coupled to the fluorine. S7 Coupling between the H-1 and H-7 protons of ion 66 is somewhat unusual since there is considerable experimental and theoretical evidence,50’ 283’ f that the parent ion 62 exists in an open form (682) as opposed to closed form (622). Therefore if coupling does occur between H-1 and H—7 in ion 22, a canonical structure like 62% may be an important resonance contributor (vide infra). 4 C! “A “I 0“ However the possibility of long range (5 bonds) fluorine, proton coupling for H-1 and H-7 can not be completely discounted in view of the fact that the methylene protons in para protonated fluorobenzene (refer to Figure 9) have a long range (5 bonds) coupling of 12 Hz with the fluorine.51 Additionally in protonated fluoromesitylene 12 H H F F 75 10 and in 12 the long range (4 bonds) couplings between the fluorine and meta ring protons are 4 and 5.5 Hz, respectively.52 58 Once the structure for 62 was deduced, the remaining carbon-13 shifts could be assigned in the following way (Table 3). The peak at 5 138.8‘was assigned to C-3, whereas the peak at 5 124.8 was assigned to C-5. There are two reasons for making these assignments. The first reason is that in other carbonium ions,46 the fluorine shields the ortho position. Thus in para protonated fluorobenzene 17 (Figure 9) the carbons ortho to the fluorine are actually at a higher H 47.7 130.1 (m) m2 144.: I167) m Figure 9. Carbon-l3 chemical shifts of protonated fluorobenzene, 77, with carbon-fluorine coupling constants in parentheses. ”m field than the meta carbons. A similar chemical shift pattern is proposed for 26. The second reason is that off-resonance decoupling does not further split the 5 138.2 peak which is consistent with the 3 position being methyl substituted. The 13.4 ppm shift difference between C-3 and C-5 positions can then be rationalized on the basis of the well known deshielding effect of a a-methyl group.449 53 Although it is difficult to distinguish between the chemical shifts for the C-2 and C-6 positions the 6 164.7 peak probably corresponds to C-2 because of an expected deshielding B—effect from the methyl group. Therefore the 5 162.0 peak was assigned to C-6. Similarly the higher field 5 92.4 peak was assigned to C-1 since this position is expected to be more shielded than C-7 (6 95.2) due to the shielding Y-effect 59 from the methyl substituent. The fact that the peaks at C-7 and C—1 are 27 ppm and 29.8 ppm higher field than for the corresponding peaks in the parent ion 28 may either be another indication of partial bonding (26%) between these positions which would be expected to shift these carbon peaks to higher field in the NMR spectrum or it may be due to the charge delocalization by the methyl and fluorine substituents. Position C-8 corresponds to the peak at 6 47.1 since it remains a singlet in off-resonance decoupling. The peak at 6 28.6 is assigned to the exo-C—8 methyl group while high field 6 11.7 peak is assigned the endo-C-8 methyl. The peak at 6 19.0 should be from the C-3 methyl group since it is split by fluorine. When ion 62 was carefully quenched in a sodium methoxide-methanol solution at -780 C a methoxyether could be isolated after preparative TLC chromatography with 50% ether/hexane eluent. Similar quenching of ion 22 in saturated aqueous sodium bicarbonate at 0° C yielded an intractable mixture. The absence of hydroxyl and carbonyl absorptions in the IR spectrum along with a sharp 3—proton singlet at 6 3.13 in the NMR spectrum is evidence for methoxyether formation. Unfortunately both the proton and carbon-13 NMR spectra had more peaks than can adequately be accounted for by a single compound. However, since an acceptable carbon, hydrogen analysis was obtained, the quench product must be a mixture of isomers. Assuming the methoxide ion would attack only the most positive carbons in 22, a priori three sets of epimers lg, 12, and 80 are possible. However the major isomer has not yet been identified due to the complexity of the proton NMR spectrum. 6O CH30 we 0‘ F F F 78 79 80 Attempted VPC separation of the quench products with column C resulted in ring contraction and loss of methanol to give fluoro- styrene 81" along with two minor products (not identified) in a 6:1:1 ratio. The proposed structure for 81 is based on spectral data and subsequent chemical transformations (Scheme 3). In the proton NMR the two vinyl methyl doublets at 6 1.83 (J = 1.6 Hz) and 6 1.88 (J = 1.6 Hz) collapse to singlets upon irradiation of the proton at 6 6.20. This result is consistent with methyl groups on an ethylene fragment being split by a vinyl proton. The 3 proton doublet at 6 2.27 ppm (J - 2.1 Hz) has a chemical shift typical for an aromatic methyl group. Since the doublet does not collapse upon irradiation of the other proton positions, this methyl group is probably split by an adjacent fluorine. Unfortunately even in a 180 MHz NMR spectrum the three aromatic protons appear as a complex multiplet centered at 6 6.98 with overlapping peaks. So it is not possible by NMR to determine exactly the position of the ethylene substituent on the aromatic ring. The chemical structure proof for 81 is outlined in Scheme 3. Periodate oxidation converted 8% to the corresponding 61 Scheme 3 KMnO. u 82 HMO; 0 1mm. ) O znhfliqafii NH2 F 33 84 fluorotoluic acid 82.53 Several fluorotoluic acids are known and were prepared by oxidation of the corresponding f1uoroxylenes.54 The oxidation of 2,4-dimethy1f1uorobenzene 84 gave the known carboxylic acid 82 which had NMR and IR spectra identical to the carboxylic acid obtained from periodate oxidation. Furthermore since the thermal rearrangement of bicyclo[5.1.0]octa- 2,4-diene 82 to l-vinyl-cyclohexa-Z,4-diene 86 is a known reaction,55 there is at least some chemical support for the proposal that either 78, 79, or 80 is indeed the quench product(s) of ion 66 which thermally mm mm mm mm 62 (21) rearranges to 81. mm Unfortunately the results for the pyrolysis of the quench product(s) do not give any additional clues about its structure. The reason for this is illustrated in Scheme 4. During the pyrolysis and rearrangement a molecule of methanol is lost which is depicted as a two step process which involves the loss of a methoxy group followed by the loss of a proton. Thus the quench product(s) of ion 22 should lose the methoxy group to give the corresponding carbonium ion 62'- Even though a free carbonium ion never may be formed during the pyrolysis, the intermediates of the pyrolitic rearrangement will be depicted, for the sake of simplicity, as carrying a formal positive charge. Ion 66' could then undergo ring expansion in one of two mmm directions to give either ion 66" or 66'". These ions after ring mmm mmmm opening and proton loss, would yield the corresponding fluorostyrenes 88 and 89. However, neither 88 nor 89 was the observed pyrolysis 'VV 'V'h ’Vb ’VL product. The observed pyrolysis product,le%,could be obtained via this mechanism by invoking ion 12'- Ion 73;, in analogy with the formation of ion (62;, could be obtained from the quench product(s) of ion 73. W 63 13’ 73" 81 Quench Products ———~ —-~ / F F F 73, 73!!! .1 6 6 ’ “I” 89 64 However, the NMR evidence clearly favors ion 66 as the ion from which the quench products are derived and not ion 1%. In other words, the most likely sequence of reactions is fluorocarbon 52 + ion 66 + quench product(s) + pyrolysis ion Qéé + fluorostyrene 81, and not the other sequence, fluorocarbon 22 +ion 1% + quench product(s) + pyrolysis ion £3; + fluorostyrene 81. In order to obtain fluorostyrene 81 via the pyrolysis of ion 66', the mechanism outlined in Scheme 4 has to be modified. One way to modify this mechanism is to invoke an additional rearrangement step in which the cyclopropyl ring shifts before ring expansion. ‘ F F F fi;~‘_‘ / 66 73 81 It is obvious that there are now two possibilities: l) the general mechanism outlined in Scheme 4 is incorrect, or 2) the pyrolysis of the quench product(s) entails extenSive rearrangement. In either event it is difficult to draw conclusions about the structure of the quench product(s) from the structure of the pyrolysis product 81. DISCUSSION These carbocations are of interest not only because ion 65 is a relatively rare example of a fluoro substituted allyl ion56 and ion 22 is the first example of a fluorine substituted homotropylium ion but also because there are several intriguing questions concerning the ionizations and rearrangements of the precursor fluorocarbons 28 and 22. The first question is, what are the mechanisms of their rearrange- ments? The second question is, why do these compounds ionize to different types of carbonium ions? The third question is, why is the pyramidal ion not formed? First, mechanisms for the ionization of 58 and 52 will be discussed. Later, some reasons why the pyramidal ion is not formed will be suggested. A. Mechanism for the ionization of Fluorocarbon 28 Outlined in Scheme 5 is a proposed mechanism for the ionization of 58 to cation 65. The Lewis acid SbFS first pulls off a fluoride ion from 58 to form fluorocation A. Cation A then ring contracts via a 1,2 carbon shift to form the cyclopropylcarbinyl cation B which ring opens to cation C. The steps leading from ions A to C are analogous to the steps in the ionization mechanism proposed for alcohol 22 which also forms an allyl cation with a bicyclic[3.2.l] ring system.57 However it is not obvious how the methyl group at the C—5 65 66 position in ion C shifts to the C-4 position in the observed ion 65. A straight forward 1,2 methyl shift (from C-5 to C-4) is unlikely since this would place a positive charge at the bridgehead position. One possible way around this difficulty is to have a 1,2 bridge shift to form ion D. Since ion D has a methyl substituted allyl system s~s=a= F F F 58 A B / L a. 3:"— 2 C D C F 67 as opposed to the fluorine substituted one in ion C, ion D is probably the more stable species (recall that more charge is located on methyl substituted C—4 than on fluorine substituted C-2 in ion 65. The greater stability of ion D thus provides a driving force for the 1,2 bridge shift. Once ion D is formed it can, Eli a series of cyclopropylcarbonyl rearrangements, proceed to ion 65. It is interesting to note that the steps from ion D to ion 65, constitute one cycle in a process known as circumambulation.57 The net effect of circumambulation is that C-1, C-8, and C-5 positions remain fixed in space while the C-2, C-3, C-4, C-6, and C-7 positions "rotate" around them. In principle, ion 62 could be a composite of five rapidly equilibrating ions. There is, however, no NMR evidence for such rapid equilibration. Once ion 22 is formed further rearrangement probably ceases because the two stabilizing groups are at the termini of the allyl system. Ion 65 is in effect an energy sink for the bicyclic[3.2.l] system. If there is any type of degenerate rearrangement in this system, it is "unseen" by NMR. One way to test for this possibility is to incorporate a label into the system which would be scrambled if the ion was rearranging. B. Mechanism for the Ionization of Fluorocarbon 22 Scheme 6 depicts a proposed mechanism for the formation of ion 66. The first three steps of the mechanism are similar to the mechanism for 69 fluorocarbon 22 (Scheme 5). It is after the formation of ion J where these reaction paths diverge. A 1,2 bridge shift in ion J would produce ion K. However ion K is not expected to be any more stable than ion J because of the destabilizing inductive effect on the allyl system by the adjacent fluorine. In fact none of the other possible isomers in the bicyclic[3.2.l] system in which the methyl and fluorine substituents are adjacent to each other seem to possess any special stability like that found for ion 22 (Scheme 5). Therefore ions J and K, which are in either rapid equilibrium or are canonical structures of ion L, have little driving force to rearrange further along the bicyclic[3.2.1] pathway, This allows other reactions to become competitive. Ion L (or J, or K) can now ring open to ion M. There are two ways in which ion M could reclose. However closure to ion 22,18 the preferred course because ion 3% has two stabilizing influences as opposed to only one for ion 73. In both 66 and 73 the methyl group may stabilize mm mm mm charge by its inductive effect. However since resonance places the 7O charge on the C-2, C-4, and C-6 positions in the homotrOpylium system, ion 22 has an extra canonical form, 33%, in which fluorine stabilizes charge by back bonding. This implies that QQ is the more stable of the two ions which is in agreement with the experimental results. It is also noteworthy that in principle ion J could also, via a series of circumambulations and 1,2 bridge shifts, rearrange to ion 33 as outlined in Scheme 7. It is not obvious why this type rearrangement does not occur. One possibility is that one of the ions produced in going from ion J to ion Q via circumambulation is ion P. Ion P is expected to be especially unstable since the positive charge is placed adjacent to the fluorine substituted carbon. Therefore any Scheme 7 / m“ CIRCUM- a” 1,2 SHIFT AMBULATION ' (3U‘3UMA- AmfiBlflJNTKNN / In, 71 circumambulatory process that "separates" the fluorine and methyl substituents on ion J (or K) is unlikely. A second possibility is that since any rearrangement from ion J to ion QR entails many steps there are simply more opportunities for intervening side reactions. These results indicate the subtle effects that influence the course of these ion rearrangements. But an explanation is still necessary to understand why the pyramidal ion is not formed. Scheme 8 which depicts the tetracyclic skeleton under consideration suggests one possible explanation. When the R group is a hydrogen, there is Scheme 8 I. ‘\ I \ I f ‘ \ a demand for stabilization by the incipient carbonium ion. The cyclopropyl part of the molecule is able to meet this demand by non-classical delocalization of the charge into the cyclopropyl ring. As a result of this initial delocalization of charge, the ion then 72 proceeds to the fully delocalized pyramidal ion. However when the R group is a fluorine or a methyl, there is less demand for non- classical delocalization because the R group is now able to stabilize the charge. As a result, the initial charge is now more localized and other rearrangements, like 1,2 carbon shifts, become competitive with pyramidal ion formation. EXPERIMENTAL A. General In addition to the equipment described in Part I of the thesis the following instruments were also used. The carbon-l3 NMR spectra were recorded with a Varian Mbdel OFT-20 instrument. The probe temperature was calibrated with a Doric Trendicator 400 type T/OC temperature meter. Proton NMR spectra were recorded on a Brucker WHl8O instrument. The following VPC columns were used: A) 10 ft x .375 in. column of 20% carbowax ZO-M on 80/100 mesh chromasorb W, B) 10 ft x .125 in. column of 10% FFAP on 80/100 mesh chromasorb W, C) 5 ft x .375 in. column of 20% carbowax 20-M on 80/100 mesh chromasorb W. B. Preparation of bornadiene (51) Compound 22,wa3 distilled from a solution of 2,6-dichlorobornane583 and potassium octyloxide in octyl alcohol by the method by Willcott.58c The product can be conveniently separated from the co-distilled octyl alcohol by chromatography with basic alumina (Baker) and pentane eluent. A 2 g sample of crude product on a 10 x 1.5 cm column of alumina was completely eluted by the second 20 mL fraction as indicated by analytical VPC at 900 C with column B. 73 74 C. Preparation of 4,4-difluoro-6,7,7-trimethy1tetracyclo- [3. 3.0.02’803’6loctane (5.2) and 4,4—difluoro-7,7,8—trimethyl- 2’803’6loctane (5g). tetracyclo[3.3.0.0 These compounds were essentially prepared by the method of Jefford.38’ 59 The reaction could be followed by withdrawing a small aliquot and adding it to 1 mL of ether. The ether solution was then analyzed by VPC at 90° C with column B. Products 22 and 52 were separated by preparative VPC at 120° C using column A with a helium flow rate of 50 mL/min and retention times of 40 and 50 min, respectively. Compound 52 is a white solid mp 44-450 C whereas 5% is a colorless liquid. D. Preparation of 2-f1uoro-4,8,8-trimethylbicyclo[3.2.1]octa-2,6- dienylium ion (Q2) and 4-f1uoro-3,8,8-trimethylcyclooctatrienylium ion (fig) . Carbon-13 Spectra The following procedure is typical for preparing carbonium ions 65 and 66. Into a 8 mm NMR tube was added 259 mg (1.39 mmol) of 58. mm mm mm The tube was flushed with nitrogen and placed in a Dewar containing a pentane-liquid nitrogen slush. The Dewar was then placed in a nitrogen filled glove bag. To compound as was then added via a disposable pipet 1.5 mL of a 3.0 M acid solution (0.45 mmol SbFS in 1.6 mL FSOZCl) which was cooled in a Dewar of acetone-liquid nitrogen slush in a 15 mL graduated distilling tube. After the acid solution 75 was added, the NMR tube and its contents were transferred to the Dewar of acetone slush. The mixture was then stirred occasionally with a supermixer in order to dissolve the frozen starting material. After approximately 1 h a clear dark red-orange solution of Q; was obtained. Next a glass insert containing deutero-acetone and TMS was frozen in liquid nitrogen and added in the glove bag to the cold acid solution. The spectrum was then recorded at -85° to -90° C. Proton Spectra. The following procedure is typical for proton NMR spectra. Into a 5 mm NMR tube fitted with Teflon spacers was added 76.8 mg of 52. The 5 mm tube was inserted into a 10 mm NMR tube containing acetone-d6 (99.52) and the tubes were frozen in pentane-liquid nitrogen slush. To 22 was added via disposable pipet approximately 0.5 mL of a 3.5 M acid solution (0.5 mL SbFS to 1.5 mL FSOZC1) which was cooled in acetone slush. After the addition of the acid solution the NMR tubes were then transferred to the acetone slush. The tubes were occasionally shaken with the supermixer in order to dissolve the frozen starting material. The spectrum was recorded at -85° C. The proton signal from the 0.51 acetone-d5 present (2.49 ppm) was used as an external reference. QuenchingiProcedure The following procedure for QQ is representative. To 40 mL of methanol is a 125 mL Erlenmeyer flask was carefully added 3.0 g of NaH. After the addition was complete the sodium methoxide solution 76 was cooled to -780 C. To this solution with vigorous stirring were quickly injected via a 9 in. disposable pipet small aliquots of the ion solution (3 mL containing 6.9 mmol SbF5 and 1.48 mmol 62 cooled in acetone slush. After the addition was complete, the quench solution was warmed to room temperature and added to 150 mL of cold water. If necessary the solution was adjusted to a basic pH with saturated aqueous sodium bicarbonate. The aqueous solution was extracted with ether and the combined ether extracts dried over magnesium sulfate. After filtering and removal of the solvent, a yellow oil was obtained. Silica gel TLC with 502 ether/pentane eluent afforded 96 mg, 33%, of quench product. The product could be microdistilled at 620 C under 1 mm pressure. NMR (DCC13) 6 (The ratios of the peaks are based on the assumption that the 6 3.23 peak is 3 protons), 6.97 (m, 0.60), 6.11 (m, 2.11), 5.36 (m, 1.90), 3.23 (s, 3.00), 2.26 (d, 0.64), 2.03 (d, 1.72), 1.98 (d, 0.94), 1.89 (d, 0.72), 1.83 (d, 0.67), 1.72 (m, 1.42), 1.28 (s, 5.60), (The multiplicities are a description of the peaks as they appear in the spectrum); UV (methanol) Amax 214 (e = 3400), 248 (4100); IR (CC14) 3300 (m), 2970 (s), 2930 (s), 2820 (m), 1680 (m), 1636 (m), 1613 (m), 1500 (m), 1454 (m), 1400 (m), 1380 (s), 1377 (s), 1310 (m), 1250 (m), 1233 (m), 1198 (m, sh), 1185 (s), 1177 (8, sh), 1158 (m), 1152 (s), 1142 (s), 1113 (s), 1076 (s), 1050 (m, sh), 943 (m), 911 (m), 889 (m), 710 (m), 696 (m), 688 (m), 602 (m); mass spectrum, m/e (rel intensity) 216 (1), 205 (2), 196 (3). 73 (100). ‘Agal. Calcd for C12H17F0: C, 73.43; H, 8.73. Found: C, 73.42; H, 8.71. 77 A similar quenching of 6; yielded ketone 62 in 50% yield after silica gel chromatography with 50% ether/pentane eluent: NMR, see Table 3; IR (CC14) 3070 (m), 3030 (m), 2980 (m), 2920 (s), 2860 (m), 1670 (s), 1625 (m), 1585 (m), 1472 (m), 1440 (s), 1390 (m), 1375 (m), 1369 (s), 1323 (s), 1309 (m), 1294 (m), 1232 (m), 1228 (m), 1211 (m), 1183 (w), 1166 (m), 1115 (w), 1090 (w), 1029 (w), 1000 (m), 950 (m), 940 (m), 910 (m), 877 (m), 7700 (m, hr), 680 cm’1 (8); UV (methanol) Amax 209 nm.(€ - 2700), 237 (5200), 274 (2100); mass spectrum m/e (rel intensity) 162 (55), 147 (100). E. Preparation of 2-methylene-4,8,8-trimethy1bicyclo[3.2.1] octatriene (1%), To a 10 mL pear shaped flask with a sidearm was added 21.2 mg of ketone QR'in 5 mL ether. After the flask was purged with nitrogen and cooled in an ice-salt bath, 0.5 mL of 1.77 M methyllithium.was injected through the serum cap of the sidearm. The solution was magnetically stirred at room temperature overnight. After again cooling the solution in an ice-salt bath, 0.3 mL of water was carefully injected. Enough 6 N HCl was added to make the solution acidic to litmus paper. As much of the aqueous layer as possible was removed via syringe and the acidic ethereal solution was stirred at 00 C for 1 h. The solution was then neutralized with 52 sodium carbonate solution. After an additional 3 mL of ether was added to the flask, the aqueous layer was again removed via syringe. The ether was dried over MgSOA and filtered. Removal of the ether solvent afforded a yellow oil. Chromatography of the oil through a short alumina column 78 with pentane eluent yielded 20.2 mg of 71. The product can also be purified by VPC at 1400 C employing column A with a helium flow rate of 50 mL/min and a retention time of 8 min. NMR, see Table 5; IR (CClA) 3060 (m), 2980 (m), 2915 (s), 2865 (m), 1638 (m), 1610 (m), 1590 (w), 1540 (w), 1472 (m), 1445 (m), 1436 (m), 1420 (w), 1385 (m), 1376 (m), 1365 (m), 1330 (m), 1315 (m), 1297 (m), 1128 (w), 955 (m), 908 (m), 881 (s), 871 (s), 860 (s), 710 (m), 637 (3); UV (methanol) Amax 216 (e - 6900), 241 (8700), 261 (5500); mass spectrum, m/e (rel intensity) 160 (39), 145 (100). ‘A351. Calcd for C12H16: C, 89.94; H, 10.06. Found: C, 89.53, H, 9.92. F. Preparation of 2,4,8,8-tetramethy1bicyclo[3.2.l]octa-2,6-dienylium ion (7%). Into a 10 mm NMR tube was added approximately 1/4 mL of FSO3H. The NMR tube was placed in an acetone-liquid nitrogen slush and 2.5 mL of FSOZC1 were condensed on top of the acid. Next 20.2 mg of 11 in 1/2 mL CD2C12 was carefully layered on top of the FSOZC1 phase. After an additional 5 min of cooling the contents of the tube were stirred together with the supermixer. The spectrum was recorded at --850 C. NMR, see Table 5. G. Preparation of fluorostyrene (81). The methoxyether quench product obtained for ion gg‘was dissolved in C014 and injected onto column C at 1700 C with a He flow rate of 79 50 mL/min. The injector and detector ports were at a temperature greater than 2000 C. The product was trapped in 60% yield in a collection tube cooled in dry ice-acetone. NMR (DCC13)15 6.98 (3, m), 6.20 (1, s, broad), 2.27 (3, d, J = 2.1 Hz), 1.88 (3, d, J = 1.6 Hz), 1.83 (3, d, J = 1.6 Hz); IR (CC14) 2965 (m), 2920 (m), 2850 (m), 1500 (s), 1450 (m), 1378 (m), 1248 (m), 1212 (s), 1178 (m), 1142 (m), 1112 (s), 903 (m), 888 (m); mass spectrum, m/e (rel intensity) 164 (75), 149 (100); UV (methanol) Amax 226 (e = 11,000), 243 (12,000). ‘Anal. Calcd for C11H14F: C, 80.45; H, 7.98. Found: C, 80.29; H, 7.85. H. Oxidation of fluorostyrene (81). To 100 mL of water was added 2.1 g of NaI04 and 300 mg of KMn04.12 A 25 mL aliquot of this stock solution was added to a 50 mL aqueous solution containing 61 mg of 81 and 200 mg sodium carbonate. After 2 h the excess oxidizing agent was destroyed by the addition of a sodium bisulfite solution. The mixture was acidified with 10 mL of 6 N HC1 and it was then extracted with four 50wm1 portions of ether. The combined ether extracts were dried over magnesium sulfate and filtered. Removal of the ether afforded 37 mg of the crude acid. Vacuum sublimation yielded 8.1 mg of pure acid 8%. The NMR and IR spectra of this acid were identical to the spectra of the acid prepared from oxidation of the corresponding fluoroxylene 84.54 NMR (DCCl3) 5 9.83 (1, s, broad), 7.77 (2, m), 6.93 (1, m), 2.28 (3, d, J - 2 Hz); IR (DCCl 3520 (w), 2920 (a, broad), 2600 (m, broad, sh), 1688 (s), 3) 1612 (m, sh), 1592 (s), 1475 (m, broad, sh), 1423 (s), 1492 (m, broad, sh), 80 1292 (s), 1264 (s), 1248 (s, eh), 1190 (s), 1158 (m), 1170 (s), 1160 (3, sh), 835 (m) cm-l. PART III MISCELLANEOUS INTRODUCTION One of the original goals of this thesis work was to find a useful synthetic route to octamethylnorbornadiene 23. This known compound can be produced by prolonged photolysis of 22:60 Unfortunately, the overall yield of 93 by this route is poor, thus prompting a search for a better synthetic method. The reason for wanting to prepare 93 mm \ .11"... .mL. $1? 3 90 91 I} _A_. 92 93 is that it is an attractive precursor to 94 which might possibly ionize mm (22) y. to the pyramidal carbocation 25. We wanted to compare the stability of 25 with that of the known analog in which the apical position has a hydrogen rather than a fluoro substituent. Since the Dials-Alder reaction provides the quickest entry into the bicyclo[2.2.1] system, the original synthetic strategy was to treat hexamethylcyclopentadiene 96 with an appropriate dienophile to produce ’\/\1 82 83 either 92 directly or a compound which could be chemically transformed into 2; (eq 24). Although a new synthesis of 92 was not achieved, the experiments in this section may be viewed as an exploratory study of the reactions of 96 with various dienophiles, mm as well as an investigation of the chemistry of the resultant Diels- Alder adducts. Ii -* 93 (23) R (24) m +R-ClC-R—-—> /‘ -——->93 R . RESULTS Hexamethylcyclopentadiene was prepared by treating the pentamethyl derivative with methyllithium in THF, followed by alkylation with methyl iodide. This method is superior to methylation with methyl iodide and sodamide in liquid ammonia, a procedure used by DeVries.61 1)IAsLi .________. (25) 2)Nkn 97 96 (95%) Scheme 9 depicts several dienophiles which did not add to 21. It is unfortunate that 2-butyne 100 does not add, since 88 would be produced directly. Compound 22 was treated with an excess of 2-butyne in a sealed tube for days at room temperature or at temperatures up to 2500 C without any NMR evidence for the formation of 23. At the higher temperatures 22 slowly decomposed. Similarly diphenylacetylene 121 and dimethylmaleic anhydride 28 also did not react with 26 under a variety of conditions. Since most Diels-Alder reactions involve addition of an electron deficient dienophile to an electron rich diene, 84 85 Scheme 9 9‘ H3C-CIC-CH3 \ 100 R n a» / U P'C'C‘P . 101 93 R SCHa 102 R 8 0 it is understandable why the electron rich 122 and 101 have little propensity for addition to 26, which is also electron rich. Although 28 has an electron deficient double bond, it is sterically hindered and therefore unable to react with 22, which has methyl substituents at both termini of the diene moiety. As expected, however, 26 readily reacts with less hindered electron deficient dienophiles (Scheme 10). Dimethyl acetylenedicarboxylate 123 reacts exothermically with one equivalent of 26 without solvent to form 123 in 1002 yield. Compound 12: was saponified in alcoholic KOH to the known dicarboxylic acid 114.62 86 Scheme 10 cuaogc-c-c-oozma 103 ’ RT cascac-oozn 105 155°C ensue-002cm CF3—Cac-CF3 112 104 00.3 87 Unfortunately, all attempts to reduce either 124 or 114 (with a variety of reducing agents) to the corresponding dialcohol 115 were unsuccess~ ful.63 In general the starting material was consumed (by NMR) and an intractable mixture was obtained. Since the mass spectrum of the crude reduction products had a peak 2 mass units higher than starting material, at least one of the reactions was reduction of the electron deficient double bond, a known side reaction in the reduction of a,B—unsaturated esters.12 Therefore, this particular route to 23 was abandoned. 02°": 1) Kai/80:1» ‘ 002H / ’ 2)H* / (26) c02H 104 114 Tetrolic acid 105 and methyl tetrolate 107 which contain only one mmm mmm electron withdrawing group, are less reactive dienophiles than 123 towards 26. Thus it is necessary to heat these reagents in order to \ cnzou 1M0r114-fl-—> It —>93 '20" us 88 initiate the Diels-Alder reaction. Under these conditions 122’ first added then cyclized to give the lactone 122; This reaction is reminiscent of the addition of maleic acid 116 to cyclopentadiene 117, W W eq 27.64 The structural assignment of 122 is based on spectral data. The IR spectrum showed no hydroxyl (OH) absorption that would be expected for a carboxylic acid. There was, however, a strong band fiwfinmi, o... (27) 116 117 118 119 at 1776 cm-1, characteristic of a y-lactone.65 The proton NMR spectrum had seven methyl group absorptions, and the most deshielded methyl group was at only 6 1.25 ppm, too far upfield for the methyl to be attached to a double bonded carbon. The compound therefore has no double bonds which implies that 122 is not the expected Diels- Alder adduct. This conclusion was verified by a carbon-13 NMR spectrum which contained no olefinic carbon absorptions. Furthermore, the 6 177.3 carbonyl carbon peak was consistent with the carbonyl chemical shift observed in other 1actones.44 The reaction of 22 with 107 did produce the expected Diels-Alder MA; adduct 108 along with a side product. VPC indicated that the ratio of “AA; 108 to side product was 3:2. Since the side product had vinyl proton absorptions in its NMR spectrum and had the same molecular weight as 89 1283;1tis probably some type of ene reaction product.66 However, its exact structure was not established. The structure of 122 rests on spectral data. In addition to a parent ion at m/e 248 in the mass spectrum, 128 also had a strong ester carbonyl absorption at 1752 cm-1. This indicated that 128 was indeed a cyclo adduct of 22,3nd £21. The NMR spectral assignments were (Figure 10) also consistent with the structure. 0.75 0'75 1L!) / / 1.53 “3 COQCH3 3.53 1.53 ‘193 Figure 10. The proton chemical shift (5) assignments for 10%. Compound 128 could be reduced with either HA1(iBu)2 or A1H3 to what is believed to be the corresponding alcohol 122. The absence of a carbonyl absorption in the IR spectrum in addition to a hydroxyl absorption at 3610 cm"1 indicated that 128 was reduced to an alcohol. However attempts to reduce this alcohol to the hydrocarbon by reduction 67 or by using Corey's pyridine sulfur trioxide method68 of the tosylate resulted in an elimination reaction as indicated by the appearance of vinyl protons in the NMR spectrum. Furthermore a comparison of the NMR spectrum of 23 with the corresponding reaction mixtures clearly ruled out the presence of any 22. 9O / / 120 CH2OH In contrast to the lack of reaction of 26 with anhydride 28, compound 26 did react smoothly with dimethyl maleate 122 to form 110. The structure of 112 was evident from the spectral data. Figure 11 shows the proton NMR spectral assignments. ‘1‘") (170’ 2.93 / C32013 - H ‘57 1.03 3 3.45 Figure 11. Chemical shift (6) assignments for 112; An attempt to methylate 112 by treatment with potassium hydride followed by the addition of methyl iodide resulted in only epimerization to the_£5§n§-diester 121; At this stage attempts to prepare the permethylated norbornadiene a} were abandoned. The remainder of the experiments in this section are exploratory Diels-Alder reactions with diene 96; 91 C02CH3 / 0'“ H 121 COZCH3 The reaction of 26 with singlet oxygen produced the corresponding endo—peroxide 1 Although the reaction can be carried out at room 11. mmm temperature, photolysis at 00 C forms a cleaner reaction mixture which is important since:§he product decomposed upon attempted purification by silica gel chromatography. The structure of the endoneroxide is based on spectral data and on subsequent chemical transformations (Scheme 11). Scheme 11 9 LiAIH. )_ H 9 122 111 0.0 Q 124 123 92 The symmetry of 111 was clear from both the proton and carbon—13 NMR spectra (assignments in Figure 12). Furthermore the proton chemical shifts were as expected for structure 111. 0.70 ( “3 1.15 Figure 12. The proton and carbon-13 chemical shift (6) assignments for 111 in parts per million from TMS. The carbon-13 shifts are in parentheses. The structure of 111 was confirmed first by heating it in CC14, to form the diepoxide 124. This type of rearrangement is a known thermal reaction of endo—peroxides.69 Epoxide formation was indicated in the NMR spectrum by a shift of the vinyl methyl protons (at 6 1.63 in 111) upfield to 6 1.32. Unfortunately 124 decomposed upon attempted purification by silica gel chromatogrpahy. Reduction of 111 with lithium aluminum hydride gave the dialcohol 122. Compound 12%, without further purification, was treated with a mmm trace of acid in ether to form the known triene 123,61b thus confirming the structure of 111. mmm Hexafluoro-Z-butyne 11% reacted exothermically with diene 26 to produce cycloadduct 113. Since it was also of interest to compare the reactions of adduct 113 with those of less methylated adducts, 125 and mmm mmm 127 were prepared by adding hexafluoro-Z-butyne to cyclopentadiene 117 mmm mmm 93 and 5,5-dimethylcyclopentadiene 126, respectively. Adducts 112 and 127 like 12570 underwent photolysis to the corresponding quadricyclic compounds 128 and 122 (Scheme 12). In general the structure assignments for 128 and 129 rest on the absence mmm mmm of the vinyl protons (or vinyl methyl groups) in the proton NMR spectra and on the analogy with the known photochemical reaction of 122. When 113 was irradiated in the presence of l M piperlyene 128 was still mmm mmm formed without any apparent quenching. Thus the 2n + 2% addition proceeds via the singlet state (or a very short lived triplet). Scheme 12 .+112—-—>/ *3 CF: 117 125 \ @ +112-—> /QCF3 or, 126 127 In addition to photolysis other exploratory reactions were performed on these adducts. For example, these compounds were treated with various electrophiles. Curiously, despite repeated attempts, adducts 113 and 125 would not mmm mmm 94 Scheme 13 react with dihalocarbenes :CXZ (X = F, Cl, Br). In every instance only starting material was recovered. In contrast, however, treatment of 113 or lzé'with'm-chloro— perbenzoic acid formed thecorresponding epoxides 121, 132, and 13%, Scheme 14). VPC analysis with column A indicated that epoxides 131 and 122 are formed in a 10:1 ratio. The structure assignments are based on the chemical shifts of the m— and anti—methyl groups at C-8. For 1,3,1, the gm: and _a_1_1t_i-methyls appeared at 6 0.85 and 6 0.78, respectively, whereas for 132 they were at 6 0.93 and 6 0.80. In epoxide 132, the 1 methyl group syn to the epoxide ring is expected to be farther down field (deshielded) than the gynfmethyl group of epoxide 131, due to the 95 Scheme 14 M 113 I ‘III, fi3____q.. I! (:53 125 133 deshielding effect of the oxygen atom. This deshielding by the epoxide oxygen has been observed in other systems71 and is further illustrated by the larger chemical shift difference (6 0.13) between thegggy-and‘gngirmethyls of epoxide 111 as opposed to the smaller chemical shift difference (5 0.07) between them-w and anti-methyls of epoxide 111. Furthermore the endo attack (caused by the steric hindrance of the bridge methyl groups) observed in 112, is also found in other systems when dimethyl substitution on the bridge position prevents the normal exo attack,72 so the assignment of the endo con- figuration to the major product, 111, seems reasonable. In contrast to adduct 113, unhindered adduct 125 is attacked on the mmm mmm exo (normal) side. Exo addition is indicated in the proton NMR spectrum 96 by the epoxide protons (6 3.50) appearing as a singlet which implies there is little (if any) coupling between the bridgehead and gnggf- protons. This small coupling constant (JQQ;O), is characteristic for the couplings observed between the bridgehead and gndgrprotons in other bicyclic[2.2.1] systems73 and agrees with the epoxide protons of 124’ also appearing as a singlet. 'V Compound 121 proved to be a fairly stable epoxide in contrast to epoxide 124:74 For example, it did not rearrange upon VPC and it was unaffected upon refluxing with trifluoroacetic acid in tetra- hydrofuran. However, upon treatment with BF3 etherate in benzene, 111 smoothly rearranged to a diene assigned structure 111. The mechanism of the rearrangement is depicted in Scheme 15. The spectral evidence for 112 is as follows: 1) the proton NMR spectrum indicated six methyl groups; from the chemical shifts, two of the methyls are attached to vinyl carbons while one methyl group is attached to a carbonyl carbon; 2) the presence of a carbonyl group was clear from the strong 1717 cm-1 absorption in the IR spectrum; 3) the trifluoromethyl groups appeared in the fluorine NMR spectrum of 122 as two quartets (long range fluorine- fluorine splitting);tfiuu;the molecule is not symmetric with respect to these groups; 4) the carbon—l3 spectrum confirms the presence of two methyl substituted vinyl carbons; 5) the mass spectrum had the expected 97 Scheme 15 135 m/e 328 value for the parent ion. Although the spectral data could possible fit two other structures 136 and 137, structure 135 is favored on mechanistic grounds. ’VW’VVD WV Figure 13. The proton chemical shifts (6) of 135 in ppm from TMS. mmm Europium shifts are in parentheses. 98 0 F3 0% <3F§ (:Fb 136 137 Finally treatment of diene 125 with one equivalent of bromine QM gave only one major addition product which is assigned structure 123. 9 u I 3 The assignment is based on spectral data. The bridge, bridgehead, 33513- and gig-protons have chemical shifts which are consistent with the structure. In 132 the e§27proton appeared as a doublet of doublets (J = 3 Hz, J = 4 Hz) whereas the endgfproton which has a smaller coupling with the bridgehead proton appeared as a broadened singlet. Furthermore the magnitudescmfthe coupling constants found for the gxgrproton were consistent with the exo-proton coupling constants in other bicyclic[2.2.1] systems.73b This ruled out the possible structure 132. In compound 122 one 99 of the protons attached to a bromine substituted carbon is in the endo position and would be expected to have different coupling constants than those found in 13%. Further evidence for 13% is that the trifluoro— methyl groups appear as two quartets in the fluorine NMR spectrum, which indicates that the molecule is not symmetric with respect to these groups. The results of bromine addition to diene 1&5 contrasts dramatically with those of bromine addition to norbornadiene 1&9 in which three rearrangement products are observed.75 Thus the trifluoromethyl groups so deactivate the double bond to which they are attached that the double bond cannot readily stabilize the intermediate carbonium ion 3.92 BA A ‘a = Br gBr 140 141 142 143 formed in the bromine addition reaction. As a result carbonium ion rearrangements, like those found in 140, are not observed in 125. mmm mmm EXPERIMENTAL A. General The fluorine NMR spectrum were obtained on a Varian model A-56/6O NMR instrument with FCC13 as an internal standard. The fluorine chemical shifts (¢) are in parts per million. A minus sign indicates a signal upfield from FCC13. In general the Diels-Alder reactions were performed by weighing out the dieneophile into a heavy walled Pyrex tube, then adding the appropriate amount of diene 33. The mixture was cooled in dry ice- acetone bath and the tube sealed with a torch. After warming to room temperature the tube was heated to the desired temperature in an oven. The following VPC columns were used: A) 10 ft x .125 in. column of 10% FFAP on 80/100 mesh chromasorb W, B) 10 ft x .250 in. column of 10% FFAP on 80/100 chromasorb W, C) 10 ft x .375 in. column of carbowax 20 M on 80/100 mesh chromasorb W. B. Preparation of l,2,3,4,5,5-hexamethylcyclopentadiene (2Q) To a 3-necked 250-mL flask was added 5.0 g (33.3 mmol) of 1,2,3,4,5-pentamethy1cyclopentadiene76 along with 100 mL of freshly distilled THF (tetrahydrofuran). The flask was swept with nitrogen and cooled in an ice bath. Upon the careful addition of 20 mL of 2 M 100 101 methyllithium, a white precipitate formed. The mixture was magnetically stirred for an additional 0.5 h whereupon 2 mL of methyl iodide was added. After the mixture was stirred overnight at room temperature, a clear yellow solution remained. The reaction mixture was carefully quenched with water at 0° C, extracted with methylene chloride and dried over magnesium sulfate. Removal of the solvent afforded 5.2 g (95%) of a light yellow 011. NMR6la (c014)5 1.70 (12, s), 0.83 (6, s). C. Attempted preparation of 1,2,3,4,5,6,7,7-octamethy1bicyclo[2.2.l]- hepta-2,5-diene (3%), The following procedure is representative. Into a 7 mm x 8 in. Pyrex tube was added ZOO-300 mg of 2-butyne 182 followed by 100 mg of diene 22; The tubes prepared in this manner were then heated to the following temperatures: 2000 C for 15 h, 2400 C for 15 h, 2000 C for 96 h. In each case the reaction mixture darkened. However, an NMR spectrum indicated only starting material. D. Attempted preparation of l,2,3,4,7,7-hexamethy1-5,6—dipheny1- bicyclo[2.2.1]hepta-2,5—diene (%Qg). The following procedure is representative. Into a 7 mm x 8 in. Pyrex tube was added 264 mg of 12% followed by 225 mg diene 2g. After adding 20 mg hydroquinone (radical trap), enough benzene was added to dissolve the diphenylacetylene 121. The tubes prepared in this manner were then heated to the following temperatures: 140° c for 15 h, 200° c for 24 h, 210° c for 72 h, room 102 temperature for approximately one year. In each case the NMR spectrum indicated starting material. E. Preparation of 5,6-di(carbomethoxy)-l,2,3,4,7,7-hexamethyl- bicyclo[2.2.1]hepta-2,4-diene (&Q£)’ Into a 10 mL flask were weighed out 0.5 g (3.3 mmol) of 96 and mm 0.47 g (3.3 mmol) of dimethyl acetylenedicarboxylate 122: The mixture became warm upon stirring. After 3 h the reaction was complete. NMR (CC14) 6 3.60 (6, s, O-methyl), 1.60 (6, 8, vinyl methyl), 1.13 (6, s, bridgehead methyl), 0.93 (3, s, anti—methyl),0.77 (3, s, syn. methyl). F. Saponification of diester (kgg). The diester (1.0 g) along with 0.5 g KOH was refluxed in 40 mL of 90% ethanol for 4 h. The solution was cooled and most of the ethanol removed under reduced pressure. The solution was added to 50 mL water and the mixture acidified with 10% HC1. The aqueous solution was then extracted with ether. The ether was dried over magnesium sulfate, filtered, and evaporated under reduced pressure to yield 0.6 g of 11g as yellowish crystals. Compound llfi was recrystallized once from methylene chloride to yield 0.54 g of white crystals (60%). Even though the melting point was low 202-2060 C (lit62b 216-217) the NMR spectrum was consistent with the structure. The product was not further purified: NMR (DCC13) 6 1.63 (6, s), 1.23 (6, s), 0.93 (3, s), 0.83 (3, s). 103 G. Attempted reduction of diester (18k). To a 10 mL sidearm flask was added 50 mg of LiAlH followed by 4. 4 mL dry THF. The flask was swept with nitrogen and cooled to 00 C. After the addition of 105 mg of diester 124 in 1 mL THF, the mixture was stirred magnetically at room temperature for 4 h. The solution was again cooled to 00 C and quenched by the careful addition of a saturated aqueous ammonium chloride solution. The precipitate was filtered and washed with ether. The combined ether-THF solutions were dried over magnesium sulfate, filtered, and evaporated under reduced pressure to yield 80 mg of an oil. An NMR spectrum of the oil indicated that starting material was consumed (the ester methyl groups disappeared). However the spectrum was extremely complex and uninterpretable. An IR spectrum contained absorptions at 3400 cm"1 (0H) and 1720 cm"1 (carbonyl ?). Since separation of the mixture could not be accomplished with silica gel TLC with a variety of eluents the reaction was not further investigated. H. Reaction of tetrolic acid with (23). Diene 96 (341 mg, 2.28 mmol) and tetrolic acid lRR’(236 mg, 1.81 mmol) were heated at 155° C in approximately 2 mL benzene for 24 h. TLC chromatography (silica gel) with 10% ether/hexane eluent gave 394 mg (74%) of product: Proton NMR (0014) 6 2.25 (1, s), 1.25 (3, s), 1.05 (3, s), 0.95 (3, s), 0.87 (3, s), 0.83 (3, s), 0.75 (3, s); carbon-13 NMR (DCCl3) 6 177.3 (carbonyl), 97.6, 59.9 (methine), 51.9, 47.0, 43.0, 36.2, 31.9, 20.7 (methyl), 19.2 (methyl), 14.4 (methyl), 8.5 (methyl), 6.3 (methyl), 5.8 (methyl), 4.2 (methyl); IR (CC14) 2985 (s), 2875 (s), 1776 (s), 1673 (m). 104 1452 (m), 1388 (s), 1380 (s), 1303 (m), 1276 (m), 1229 (m), 1180 (m), 1150-1140 (m, br), 937 (m), 915 cm"1 (m); mass spectrum m/e (rel intensity) 234 (33), 175 (100). Anal. Calcd for C15H2202: C, 76.88; H, 9.46. Found: C, 76.42; H, 9.14. 1. Preparation of 2-(carbomethoxy)-l,3,4,5,6,7,7-heptamethyl- bicyclo[2.2.1]hepta-2,5-diene (198): Compound 26 (5.0 g, 33.3 mmol) and methyl tetrolate £21 (3-6 8, 36.7 mmol) were heated at 150° C in 5 mL benzene for 48 h. Distillation at 980 C under 2.5 mm Torr provided 5.6 g (68%) of reasonably pure 128. Adduct 122 was also purified by preparative VPC with column B at 150° C: Proton NMR (CC14) see Figure 10; carbon-13 NMR (D0013) 6 169.4 (carbonyl), 144.1 (vinyl), 140.2 (vinyl), 139.2 (vinyl), 102.3 (vinyl), 77.8, 66.6, 63.5, 50.0 (O-methyl), 18.2 (2, methyls ?), 14.4 (methyl), 11.4 (methyl), 11.2 (methyl), 9.00 (methyl), 7.2 (methyl); IR (C014) 3030 (m), 1700 (s), 1612 (m), 1433 (s), 1380 (s), 1314 (s), 1290 (s), 1198 (s), 1170 (m), 1075 (m), 1050 cm"1 (8); mass spectrum m/e (rel intensity) 248 (45), 233 (100). ‘5231. Calcd for C16H2402: C, 77.38; H, 9.74. Found: C, 77.63; H, 9.74. J. Reduction of 2-(carbomethoxy)-1,3,4,5,6,7,7-heptamethylbicyclo[2.2.1]- hepta-2,5-diene (108). Into a 25 mL sidearm pear-shaped flask fitted with a serum cap, was added a 7 mL hexane solution containing 300 mg (1.21 mmol) of 108. 105 The solution was purged with nitrogen and cooled in an ice-salt bath. A 202 hexane solution (6 mL) of cold di-isobutylaluminum hydride was slowly injected into the flask. The magnetically stirred solution was allowed to warm to room temperature. After 15 h, the solution was again cooled in an ice-salt bath and carefully quenched with 10% water—methanol. The white precipitate which formed was filtered and washed with 10 mL methanol and 75 mL ether. The combined ether-methanol-hexane solutions were washed twice with brine and dried over sodium sulfate. Removal of the solvent afforded 220 mg (86%) of 122 as a colorless oil which yellowed on standing: NMR (0014) 6 3.90 (2, s), 1.63 (3, s), 1.53 (6, s), 1.37 (3, s),1,05 (3, s), 0.96 (3, s), 0.72 (3, 3); IR (neat) OR at 3610 cm’l; mass spectrum m/e (rel intensity) 220 (12), 187 (100). K. Attempted reduction of ester 108 to hydrocarbon 93. mmm mm The ester (509 mg) was reduced with di-isobutylaluminum hydride by the previously described method. The alcohol in turn was treated with three equivalents of p-toluenesulfonyl chloride for 5 days by the method outlined by Fieser.12 After work up 456 mg of a light yellow oil was obtained. The IR spectrum had no hydroxyl absorption, but it did have a sharp doublet at 1193 cm"1 and 1180 cm"1 characteristic of a tosylate. The tosylate was dissolved in dry THF and treated with lithium triethylborohydride by the method outlined by Brown.77 After work up an oil was obtained which had a complex NMR spectrum with many new vinyl proton peaks between 4-5 ppm. Comparison of this spectrum with an authentic spectrum for 2% (NMR 6 1.55 (12, s), 0.97 (6, s), 0.70 (6, 8)) clearly indicated that the desired product was not formed. 106 L. Preparation of endo-cis-S,6-di(carbomethoxy)-1,2,3,4,7,7- hexamethylbicyclo[2.2.1]hepta-2-ene.(lkg), Diene 26 (3.0 g, 20 mmol) and dimethyl maleate 128 (2.88 g, 20 mmol) were heated in 5 mL benzene at 1500 C for 24 h. Removal of the solvent afforded 112 which was used without further purification. Proton NMR (CC14) 5 3.45 (6, s, O-methyl), 2.93 (2, s), 1.57 (6, 5, vinyl methyl), 1.03 (6, s, bridgehead methyl), 0.70 (3, s, anti—methyl), 0.60 (3, s, synumethyl). M. Epimerization of cis—S,6-di(carbomethoxy)-1,2,3,4,7,7- hexamethylbicyclo[2.2.1]hepta-2-ene (1&8), To a dry 50 mL 3-necked flask was added 0.68 g (3.4 mmol) of 20% KH in oil. The KH was washed twice under nitrogen with 3 mL portions of pentane. The pentane was removed with a syringe. To the KH was added 10 mL of freshly distilled THF followed by a 2 mL THF solution of 1.0 g (3.4 mmol) of 112. A gas bubbler attached to the flask indicated that hydrogen was slowly evolved. After 0.5 h, 0.22 mL (3.5 mmol) of methyl iodide was added to the solution. The solution was magnetically stirred overnight during which time a white precipitate formed. The reaction was carefully quenched with water and the aqueous phase extracted with ether. The combined ether-THF phases were dried over magnesium sulfate. Removal of the solvent afforded 0.75 g (752) of 131: NMR (CC14)6 3.57 (3, s, O-methyl), 3.53 (3, s, 0- methyl), 3.12 (1, d, J a 6 Hz), 2.48 (1, d, J = 6 Hz), 1.53 (3, s broad, vinyl methyl), 1.37 (3, m, vinyl methyl), 1.10 (3, s bridgehead methyl), 107 1.02 (3, s bridgehead methyl), 0.73 (3, s, anti—methyl), 0.60 (3, s, s n“ methyl); mass spectrum m/e (rel intensity) 294 (12), 150 (100). N. Photooxidation of 1,2,3,4,5,5-hexamethy1cyclopentadiene (26). Into a 200 mL round-bottomed flask was added 1.0 g of 92, 300 mg of methylene blue and 125 mL of methylene chloride. The flask was cooled in an ice—salt bath and a steady stream of oxygen was bubbled through the solution. The solution was then irradiated by light from a Kodak 800 carousel slide projector. The reaction was monitored by VPC with column A at 900 C by observing the disappearance of starting material. After 2 h the solvent was removed and the residue was extracted with ether. The ether solution was filtered through a bed of Celite filter aid to remove the last traces of methylene blue. Removal of the ether afforded 0.97 g (80%) of 111 as a reddish oil. Proton NMR (CC14) and carbon-l3 NMR (DCC13) see Figure 12; mass spectrum m/e (rel intensity) 182 (5), 151 (100). 0. Preparation of cis-l,2,3,4-di(epoxy)-1,2,3,4,5,5-hexamethyl- cyclopentane<124). mmm Into an NMR tube was added 120 mg ofllll in CC14. The tube was then heated in an oil bath at 80° C for 2 h: Praton NMR (0014) 6 1.32 (6, s, epoxide methyl), 1.07 (6, s, epoxide methyl), 0.95 (3, 8, EDGE: methyl), 0.88 (3, s, antirmethyl). Chromatography with silica gel resulted in decomposition of the product. 108 P. Preparation of cis-1,2,3,4,4,5-hexamethylcyclopent-l-ene—3,5 diol(122). Into a 25 mL sidearm flask with a serum cap attached was added 220 mg LiAlH4 (5.5 mmol) and 10 mL dry ether. The flask was cooled in an ice—salt bath and 520 mg (2.85 mmol) of 111 in 5 mL dry ether were slowly injected into the solution. The solution was magnetically stirred under nitrogen at room temperature overnight whereupon the solution was again cooled in an ice-salt bath and the reaction quenched by the careful addition of 0.2 mL water, 0.2 mL 15% NaOH solution, and 0.7 mL water. The solution was filtered and the precipitate washed with ether. The combined ether solutions were dried over magnesium sulfate and filtered. Removal of the ether afforded 319 mg (61%) of 12%: NMR (C014)152.70 (2, s broad, hydroxyl), 1.57 (6, 8, vinyl methyl), 1.02 (6, 3, hydroxyl methyls), 0.88 (3, s, gzgemethyl), 0.67 (3, 8, anti? methyl); IR (CCl4) hydroxyl absorptions at 3616, 3540 (shoulder), 3430 cm'l. Q. Preparation of 1,2,4,4-tetramethy1-3,S-dimethylenecyclopentene(12m). Dialcohol 122 (319 mg, 1.73 mmol) was dissolved in 15 mL of ether. Two drops of 6 N HC1 solution were added and the mixture was stirred for 3 h. The ether solution was washed with 102 sodium bicarbonate and was dried over magnesium sulfate. Removal of the ether afforded 49 mg (19%) of 23 as a yellow oil: Proton NMR (CC14) 6 4.62 (2, S,_§9_-methy1ene), 4.50 (2, meg-methylene), 1.80 (6, a, vinyl methyl), 1.05 (6, s, gem- dimethyl). Lit..61b NMR (neat) 4.75 (2, s), 4.65 (2, s), 1.81 (6, s), 1.09 (6, s). 109 R. Preparation of 1,2,3,4,7,7-hexamethy1—5,6-di(trif1uoromethyl)- bicyclo[2.2.1]hepta-2,5—diene(1&é)~ Diene 22 (19.4 g, 129 mmol) and hexafluoro-Z-butyne (21.5 g, 133 mmol) in 20 mL benzene solution were allowed to stand in a sealed tube overnight at room temperature. Distillation afforded 37.7 g (93%) of 113 as a colorless oil: bp 47-480 C (1.5 mm). Proton NMR (CC14) 5 1.60 (6, 3, vinyl methyls), 1.17 (6, s, bridgehead methyls), 0.96 (3, s, anti-methyl), 0.78 (3, s, gynemethyl); carbon-13 NMR (acetone-d6) 5 157.2 (m, vinyl carbons), 143.8 (vinyl carbons), 132.5 (q, J = 272 Hz, C-F trifluoromethyl), 81.3 (bridge), 67.8 (bridgehead), 18.4 (methyl), 11.5 (methyl), 8.70 (methyl); fluorine NMR (CCla) 6 -60.0 (6, 3); IR (CC14) 2990 (m), 1658 (m), 1460 (m), 1440 (m), 1400 (w), 1390 (m), 1370 (m), 1322 (s), 1307 (s), 1245 (s), 1180 (s), 1143 (8, sh), 1120 (s), 1020 (w), 925 (m), 702 (w), 675 cm‘1 (m); UV (cyclohexane) Amax 223 (E = 1500); mass spectrum m/e (rel intensity) 312 (11), 41 (100). Anal. Calcd for C15H C, 57.69; H, 5.81. Found: C, 57.63; 18F6‘ a, 5.73. S. Preparation of 7,7-dimethyl-2,3-di(trifluoromethy1)bicyclo[2.2.1]- hepta-2,5-diene(ng). Diene 12678 (2.1 g, 22.4 mmol) and hexafluoro-Z-butyne 112 (4.1 g, mmm mmm 25.3 mmol) in 10 mL of benzene were allowed to stand in a sealed tube at room temperature for 24 h. Distillation at atmospheric pressure afforded 3.5 g (612) of 127 as a colorless liquid: bp 145-1470 C. Proton NMR (C014) 5 6.72 (2, m, vinyl), 3.40 (2, m, bridgehead), 1.23 110 (3, s, antirmethyl), 1.17 (3, s, gynfmethyl); fluorine NMR (CC14) 0 -63.6 (6, s); IR (C014) 3000 (m), 2960 (m), 1683 (m), 1472 (w), 1449 (w), 1428 (w), 1393 (w), 1378 (m), 1352 (s), 1300 (s), 1250 (m), 1232 (m), 1200 (m, sh), 1182 (s), 1097 (m, sh), 1002 (w), 960 em‘l (m); mass spectrum m/e (rel intensity) 256 (2), 241 (100). ‘Angl. Calcd for C11H10F6: C, 51.57; H, 3.93. Found: C, 51.57; H, 4.07. T. Preparation of 2,3-di(trifluoromethy1)bicyclo[2.2.1]hepta-2,5- diene(&%2), Cyclopentadiene 117 (6.8 g, 103 mmol) and hexafluoro-Z-butyne 112 mmm mmm (17.0 g, 105 mmol) were allowed to stand in a sealed tube at room temperature for 24 h. Distillation at atmospheric pressure afforded 16.8 g (722) of 122 as a colorless liquid79: bp 1250 C. U. Photolysis of 1,2,3,4,7,7-hexamethy1-5,6-di(trif1uoromethyl)- bicyclo[2.2.1]hepta-2,5-diene(11%). A 350—mg (1.12 mmol) sample of 112 in 75 mL pentane was added to a 100 mL Vycor test tube which was sealed with a serum cap. The sample was irradiated with a 450 W Hanovia medium pressure mercury lamp in a quartz immersion well fitted with a Vycor filter. After approximately 3 h the photolysis was stopped and the solvent removed. A light yellow solid 80 The product was purified by VPC with column C (345 mg) was obtained. at 130° C to yield 128 as a waxy white solid: mp 138-1400 C. Proton W1: NMR (001,) 5 1.28 (6, s), 1.08 (6, s), 1.02 (3, s), 0.87 (3, s); 111 fluorine NMR (CCla) 0 -57.5 (6, 8); IR (C014) 2970 (m), 2930 (m), 2875 (w), 1470 (w), 1425 (m), 1386 (w), 1368 (w), 1343 (m), 1310 (m), 1185 (s), 1175 (8, sh), 1150 (s), 1115 (m, sh), 1050 and (w); mass spectrum m/e (rel intensity) 312 (26), 41 (100). Anal. Calcd for 015H18F6: C, 57.69; H, 5.81. Found: C, 57.49; H, 5.68. V. Photolysis of 7,7-dimethyl-2,3-di(trifluoromethy1)bicyclo[2.2.1]- hepta—2,5-diene ($61)“ A 500 mg (1.95 mmol) sample of 117 in 75 mL pentane was photolysed by the previously described method. Purification by VPC with column C at 1100 C afforded 122 as a clear oil. Proton NMR (C014) 6 2.25 (2, d, J = 5 Hz), 1.62 (2, d, J - 5 Hz), 1.31 (3, s), 1.27 (3, s); fluorine NMR 0 -64.2 (6, 3); IR (CC14) 2960 (m), 2920 (m), 2860 (m), 1463 (s), 1448 (s), 1400 (w, sh), 1386 (m), 1368 (m), 1335 (s), 1232 (m, sh), 1210 (s), 1185 (s), 1150 (s), 1081 (s), 1073 (s), 1010 (w), 998 (m), 925 cm‘1 (w); mass spectrum m/e (rel intensity) 256 (4), 241 (100). Anal. Calcd for C11H10F6: C, 51.57; H, 3.93. Found: C, 51.35; H, 3.76. W. Attempted carbene addition to {62" To 3.1 g (26 mmol) of chloroform was added 2.0 g (8.8 mmol) of diene 125. To this solution was added 50 mg of cetyltrimethylammonium bromide (phase transfer catalyst)81 and 3.5 g of 50% aqueous sodium hydroxide. The mixture was vigorously stirred and warmed in an oil bath 112 at 500 C. The reaction was followed by NMR by taking small aliquots and adding them to CCla. After 24 h the chloroform signal disappeared with the starting material signals still present and no indication of any product signals. X. Preparation of endo-5,6-epoxy-1,4,5,6,7,7-hexamethy1-2,3- di(trifluoromethy1)bicyclo[2.2.1]hepta-2-ene (lgi), To a solution of 1.31 g (6.45 mmol) of 85% m-chlorOperbenzoic acid in 50 mL methylene chloride was added 2.00 g (6.41 mmol) of kwa. The solution was magnetically stirred overnight at room.temperature. The precipitated m-chlorobenzoic acid was filtered and the methylene chloride solution was washed twice with 10% sodium bisulfite, twice with 10% sodium hydroxide and once with water. The solution was then dried over sodium sulfate and filtered. Removal of the solvent afforded 1.71 g (812) of a colorless oil which solidified on standing. Prepara— tive VPC with column B at 130° C gave 121 and 12% in a 10:1 ratio: For 121 proton NMR (D0013) 6 1.35 (6, s, epoxide methyls), 1.17 (6, s broad, bridgehead methyls), 0.85 (3, s, maethyl, to the epoxide ring), 0.78 (3, s, angiamethyl); carbon-13 NMR82 (acetone-d6) 6 123 (q, J 8 274 Hz, trifluoromethyl), 74.3, 63.8, 61.1, 18.9 (methyl), 16.6 (methyl), 12.4 (methyl), 9.3 (methyl); IR (C014) 2900 (m), 1660 (m), 1471 (m), 1455 (m), 1400 (m), 1391 (m), 1386 (m), 1379 (m), 1362 (m), 1340 (m), 1311 (s), 1239 (m), 1172 (s), 1145 (s), 959 (s), 700 (m), 672 curl (111); UV (cyclohexane) Amax 215 nm (e = 900); mass spectrum m/e (rel intensity) 328 (6), 41 (100). Anal. Calcd for C15H18F60: C, 54.88; H, 5.53. Found: C, 54.82; H. 5.61. 113 For 13% proton NMR (DCC13) 6 1.38 (6, s, epoxy methyls), 1.27 (6, s broad, bridgehead methyls), 0.93 (3, s, gygfimethy1,to epoxide ring), 0.80 (3, s, anti-methyl). Y. Preparation of exo-5,6-epoxy-2,3-di(trifluoromethy1)bicyclo[2.2.l]— hepta-Z-ene {133). MC: To a 40 mL solution of methylene chloride containing 4.5 g (22.2 mmol) of 85% m-chloroperbenzoic acid was added 5.0 g (20.5 mmol) of 125. After being stirred magnetically at room temperature for 24 h, the methylene chloride solution was filtered, washed twice with 10% sodium bisulfite, twice with 10% sodium hydroxide and once with brine. Then the solution was dried over magnesium sulfate, filtered and distilled. There was obtained 3.08 g (582) of 122 as a colorless liquid, bp 150-1520 C: Proton NMR (CC14) 6 3.53 (2, s, epoxide protons), 3.28 (2, s broad, bridgehead protons), 1.88 (1, d, J = 9 Hz 8 neproton), 1.61 (1, d, J = 9 Hz .antieproton); fluorine NMR (C014) 0 -61.8 (6, s); carbon-13 NMR82 (DCCl3) 6 121.3 (q, JC-F - 270 Hz, trifluoro- methyl), 57.7 (epoxide carbons), 46.3 (bridgehead), 40.2 (bridge); IR (CC14) 3150 (w), 3120 (w), 1723 (w), 1673 (w), 1450 (w), 1386 (m), 1362 (m), 1353 (m), 1301 (s), 1292 (8, sh), 1268 (m), 1242 (m), 1185 (s), 1167 (s), 1152 (s), 1042 (m), 1014 (m), 934 (m), 660 (w), 610 cm.“1 (w); UV (cyclohexane) Amax 233 nm (e - 1600); mass spectrum m/e (rel intensity) 244 (35), 204 (100). Anal: Calcd for C9H60F6: C, 44.28; H, 2.48. Found: C, 44.32; H, 2.51. 114 2. Preparation of 1-acetyl-1,4,5,6,6-pentamethy1-2,3-di(trifluoro- methyl)cyclohexa-2,4-diene (135). W To 50 mL benzene containing 1.05 g (3.2 mmol) of 121 was added 1.1 mL of BF3 etherate. The solution was stirred at room temperature for 3 h before quenching with saturated sodium bicarbonate. The benzene layer was dried over magnesium sulfate and filtered. Removal of the solvent afforded 968 mg of crude 135. Purification was accomplished by silica gel chromatogrpahy with 30% ethyl acetate/hexane eluent: carbon-13 NMR82 (DCCl3) 6 218 (carbonyl), 143.6 (vinyl), 121.5 (vinyl), 57.7, 40.8, 29.8 (methyl), 21.8 (methyl), 15.2 (methyl), 15.1 (methyl), 14.3 (methyl); fluorine NMR (CC14) 0 -57.4 (3, q, J = 7 Hz), -59.4 (3, q, J . 7 Hz); IR (CC14) 2995 (m), 2960 (m), 1717 (s), 1560 (m, broad), 1468 (m), 1404 (w), 1382 (m), 1372 (m), 1360 (m), 1323 (m), 1290 (s), 1272 (s), 1170 (s), 1152 (s), 1140 (s), 1090 (m), 1060 (w), 960 (w), 932 (w), 903 (w), 702 (w), 680 (w), 672 (w) cm"1; mass spectrum m/e (rel intensity) 328 (6), 41 (100). ‘Aggl. Calcd for C15H18F60: C, 54.88; H, 5.33. Found: C, 54.89; H, 5.47. AA. Bromination of diene 125 . mmm Diene 125 (0.50 g, 2.2 mmol) was dissolved in 10 mL methylene chloride. The solution was cooled to -50 C and one equivalent of bromine in 1.5 mL methylene chloride was slowly added. After allowing the solution to warm to room temperature, a VPC of the reaction mixture using column A at 1500 C indicated three peaks in the ratio of 20:1:1. 115 Removal of the solvent afforded 0.85 g (100%) of a light yellow oil. Microdistillation under 0.2 mm pressure afforded 0.75 g (88%) of 128 as a colorless oil: bp 430 C. Proton NMR (C014) 0 4.37 (1, d, d, J = 4 Hz, J = 3 Hz), 3.87 (1, s broad), 3.53 (1, m), 3.40 (1, m), 2.23 (2, m); fluorine NMR (CC14) 6 -65.1 (3, q, J = 4 Hz), -66.9 (3, q, J = 4 Hz); IR (CC14) 1422 (m), 1457 (s), 1387 (m), 1324 (m), 1296 (s), 1260 (s), 1050 (m), 975 (w), 948 (m). Anal. Calcd for C9H6F6Br2: C, 27.86; H, 1.56; Br, 41.19. Found: C, 27.76; H, 1.54; Br, 41.22. REFERENCES l. 10. 11. 12. 13. 14. 15. REFERENCES S.S. Hixson, P.S. Mariano and H.E. Zimmerman, Chem. Rev., 73, 531 (1973) . ’V” (a) H.E. Zimmerman and A.C. Pratt, J. Am. Chem. Soc., 92, 6259 (1970). (b) H.E. Zimmerman and A.A. Baum, J. Am. Chem. Soc., 91, 3646 (1971). Z. Goldschmidt and A. Kende, Tetrahedron Lett., 4625 (1971). (a) H. Hart and R.R. Murray, J. Am. Chem. Soc., 91’ 2183 (1969). (b) H. Hart and G.M. Love, J. Am. Chem. Soc., 95, 4592 (1973). P. Lavrik, Ph.D. Thesis, (Mich. St. Univ., 1974). C.W. Spangler and D.L. Bales, J. Org. Chem., 37, 1020 (1972). Philip Crews and John Beard, J. Org. Chem., 38, 522 (1973). (a) M. Jones, Jr., Angew. Chem.Int. Ed. Engl., 8, 76 (1969). (b) M. Jones, Jr., Private communication. m (a) Paul J. Kropp, J. Am. Chem. Soc., 91, 5783 (1969). (b) R.R. Sauers, W. Schinski, and MgM. Mason, Tetrahedron Lett., 47, 4763 (1967). (a) H.E. Zimmerman, R.S. Givens, R.M. Pagni, J. Am. Chem. Soc., 2% 6096 (1968). (b) R.A. Snow, C. Degenhart and L. Paquette, Tetrahe ron Lett., 49, 4447 (1976). (c) 0.8. Warren and J.J. Bloomfield, J. Org. Chem., 38, 4011 (1973). R.S. Givens, W.F. Oettle, R.L. Coffin, and R.L. Carlson, J. Am. Chem. Soc., 93, 3957 (1971). L.F. Fieser and M. Fieser, "Reagents for Organic Synthesis" Vol. 1, 435 (1967). References therein. C.A. Brown, J. Qrg. Chem., 26, 3913 (1974). (a) 0.8. Hammond and R.S. Cole, J. Am. Chem. Soc., 87, 3256 (1965). (b) 0.3. Drucker, V.G. Toscano, R.G. Weiss, J. Am. Chem. Soc., 25, 6483 (1973). (c) C. Ovannes, R. Beugelmans and G. Roussi, J. Am. Chem. Soc., 92’ 8472 (1973). W.B. Herkstrotter, L.B. Jones and 0.8. Hammond, J. Am, Chem. Soc., 88, 4777 (1966). 116 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 117 (a) A.J. Frey, R.S.H. Liu and 0.8. Hammond, J. Am. Chem. Soc., 88 4781 (1966). (b) C.C. Wamser and P.L. Chang, J. Am. Chem. Soc.,gé 2044 (1973). (c) K. Janda and F.S. Wettack, J. Am. Chem. Soc., §¥ 305 (1972). (d) A. Yeta and N.J. Turro, Chem. Phys. Lett., 17, (1972). (e) C.C. Wamser, R.T. Medary, I.E. Kochevar, N.J. Turro, P.L. Chang, J. Am. Chem. Soc., 97, 4864 (1975). (a) N.C. Yang and E.G. Feit, J. Am. Chem. Soc., , 504 (1968). (b) F.D. Lewis, Tetrahedron Lett., 16, 1373 (197 ). (c) F.D. Lewis and T.A. Hilliard, J. Am. Chem. Soc., 22, 6672 (1970). (d) N.C. Yang, E.D. Feit, M.H. Hui, N.J. Turro, and J.C. Dalton, J. Am. Chem. Soc., 92, 6974 (1970). P.J. Wagner, J.M. McGrath and Richard G. Zepp, J. Am. Chem. Soc., 94, 6883 (1972). For a discussion of the Schenck mechanism, see: A.A. Lamola in P.A. Leermakers and A. Weissberger, editors, "Technique of Organic Chemistry", Interscience, New York, N.Y., 1969. p. 80 and references cited therein. N.C. Yang, J.I. Cohen and A. Shani, J. Am. Chem. Soc., 90, 3264 (1968). N. Shimizu and P.D. Bartlett, J. Am. Chem. Soc., 28, 4193 (1976). The attack by excited acetone could just as well occur on the most hindered side of the molecule and a mechanism could be written to explain the regiospecificity. I. Marata and Y. Sugihara, Tetrahedron Lett., 36, 3785 (1972). (a) K.N. Houk and Cielo Santiago, J. Am. Chem. Soc., 28, 3380 (1976). (b) C. Santiago, K.N. Houk, R.A. Snow and L.A. Paquette, J. Am. Chem. '§23., 28, 7443 (1976). (c) L.A. Paquette, D.M. Cottrell and R.A. Snow, J. Am. Chem. Soc., 22, 3723, 3734 (1977). (a) P.J. Wagner, I.E. Kochevar and A.E. Kemppainen, J. Am. Chem. Soc., 24, 7489 (1972). (b) P.J. Wagner, J. Am. Chem. Soc., 89, 5898 (1967). mm L. Friedman and F.M. Logullo, J. Org, Chem., 34, 3089 (1969). When the catalytic hydrogenation was repeated, two products were obtained (observed in the NMR spectrum and by VPC with column D at 120° C) in a 1:1 ratio. One of the products was 4 (by NMR). The other the product is probably the result of 2,8 bond cleavage (Scheme 1). The reason for this discrepancy is not clear. (a) C.A. Olah and P. vR. Schleyer, "Carbonium Ions" Vol. 3, Wiley-Interscience, New York, 1972. (b) H.C. Brown, Tetrahedron, 2, 179 (1976). (c) H.C. Brown, Accounts Chem. Res., 6, 337 (1973). ) C.A. Olah, Accounts Chem. Res., , 41 (1976). (d) C.A. Olah, Agnew. Chem., Int. Ed. Engl. 1,2,, 173 1973). (f) S. Winstein, Q. Rev., Chem. Soc., , 1411 (1969). (g) P.D. Bartlett, "Nonclassical Ions," .A. Benjamin, New York, 1965. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 118 (a) W.D. Stohrer and R. Hoffmann, J. Am. Chem. Soc., , 1661 (1972). for other theoretical calculations see (b) . Yoneda and Z. Yoshida, Chem. Lett., 607 (1972). (c) H. Kollmar, H.O. Smith and P. vR. Schleyer, J. Am. Chem. Soc., 99, 5834 (1973). (d) M.J.S. Dewar and R.C. Hadden, J. Am. Chem. Soc., 95, 5837 (1973). for calculations concerning a homopyramidal catiofimsee (e) K. Morio and S. Masamune, Chem. Lett., 1107 (1974). (a) S. Masamune, M. Sakai and H. Ona, J. Am. Chem. Soc., 94, 8955 (1972). (b) S. Masamune, M. Sakai, H. Ona, and Alanmfi. Jones, J. Am. Chem. Soc., 94, 8956 (1972). (c) S. Masamune, Pure and 515151. Chem., 33, 861“(1975). (a) H. Hart and M. Kuzuya, J. Am. Chem. Soc., 4, 8959 (1972). (b) H. Hart and M. Kuzuya, J. Am. Chem. Soc., mm, 6436 (1974). (a) H. Hogeveen and P.W. Kwant, Tetrahedron Lett., 1665 (1973). (b) H. Hogeveen and P.W. Kwant, J. Am. Chem. Soc., 99, 2208 (1974). (c) H. Hogeveen and P.W. Kwant, Accounts Chem. Res., 8, 413 (1975). (d) C. Giordano, R.F. Heldeweg and H. Hogeveen, J. Am? Chem. Soc., 99, 5181 (1977). (a) S. Masamune, M. Sakai, A.V. Kemp-Jones, H. Ona, A. Venot and T. Nakashima, Agnew. Chem. Int. Ed. Engl., , 769 (1974). (b) R.K. Lustgarten, J. Am. Chem. Soc., 93, 602 (1972). A.V. Kemp-Jones, N. Nakamura and S. Masamune, Chem. Comm., 109 (1974). H. Hart and M. Kuzuya, J. Am. Chem. Soc., 97, 2450 (1975). mm R.M. Coates and E.R. Fretz, Tetrahedron Lett., 1955 (1977). H. Hart and M. Kuzuya, J. Am. Chem. Soc., 99, 2459 (1975). (a) C.W. Jefford, W.D. Graham and U. Burger, Tetrahedron Lett., 4717 (1975). (b) C.W. Jefford, J. Mareda, J.C.E. Gehret, N. Kabengle, W.C. Graham.and U. Burger, J. Am. Chem. Soc., 98, 2585 (1976) and references cited therein. (c) Y. Jean, Tetrzhedron Lett., 2689 (1977). G.A. Olah and G. Liang, J. Am. Chem. S°c°’43€’ 7026 (1976). L. Paquette, M.J. Broadhurst, P. Warner, G.A. Olah, G. Liang, Tetrahedron Lett., 3759 (1973). (a) P.K. Freeman and D.G. Kuper, Chem. and Ind., 424 (1965). (b) J.B. Stothers, J.R. Swenson, and C.T. Tan, Can. J. Chem., , 581 (1975). (c) D. Aue and N.J. Meshishnek, J. Am. Chem. Soc., mm, 223 (1977). Z. Goldschmidt, U. Gutman, Y. Bakal, A. Worchel, Tetrahedron Lett., 3759 (1973). 43. 44. 45. 46. 47. 48. 49. 50 51. 52. 53. 54. 55. 56. 57. 58. 59. 119 G.A. Olah and G. Liang, J. Am. Chem. Soc., 6434 (1972) J.B. Stothers, "Carbon-13 Spectroscopy", Academic Press, New York, 1972. (a) G.A. Olah, G. Liang and Y.K. Mo, J. Org, Chem. 99} 2394 (1974). (b) J.C. Tatlow, R.D. Peacock and H.H. Hymen, "Advances in Fluorine Chemistry," Vol. 7, Butterworth & Co., London, 1973, p. 69. (c) R.D. Chambers, "Fluorine in Organic Chemistry," Wiley- Interscience, New York, 1973, p. 73. R.J. Spear, D.A. Forsyth, and G.A. Olah, J. Am. Chem. Soc., 9%, 2493 (1976). R.F. Childs and C.V. Rogerson, J. Am. Chem. Soc. In Press. G.A. Olah, Y.K. Mo, Y. Halpern, J. Am. Chem. Soc., 94, 3551 (1972). (a) P. Warner, D.L. Harris, C.H. Bradley and S. Winstein, Tetrahedron Lett., 4013 (1970). (b) C.H. Keller and R. Pettit, J. Am. Chem. Soc., ng 606 (1966). R.C. Haddon, Tetrahedron Lett., 863 (1975). G.A. Olah and T.E. Kiovsky, J. Am. Chem. Soc., 29} 5692 (1967). (a) D.M. Brouwer, Rec. Trav. Chim. Pays-Bas, (4), 335 (1968). (b) D.M. Brouwer, Rec. Trav. Chim. Pays-Bas, (4), 342 (1968). G.A. Olah and Daniel J. Donovan, J. Am. Chem. Soc., 99, 5026 (1977). E.W. Crandall, R. Beasley, L.L. Lamburg and R. Mbriconi, J. Org. Chem., 9%, 134 (1967). (a) W. vE. Doering and W.R. Roth, Tetrahedron, 19, 715 (1963). (b) W.R. Roth and B. Peltzer, Ann., 22,2: 56 (1933). R.D. Chambers, A. Parkin and R.S. Matthews, J. Chem. Soc. Perkin I 2107 (1976). (a) H. Hart and M. Kuzuya, J. Am. Chem. Soc., 98, 1545 (1976). (b) H. Hart and M. Kuzuya, J. Am. Chem. Soc., 9%, 1551 (1976). (a) H. Kwart and G. Null, J. Am. Chem. 925., 78, 5943 (1956). (b) M. Hanack, H. Eggensperger and R. Hhhnle,thn., , 96 (1962). (c) M.R. Willcott and C. Boriack, J. Am. Chem. Soc.,ééé, 3288 (1968). The triglyme solvent was distilled once from sodium followed by a second distillation from either LiA1H4 or sodium. The distillations were conducted under reduced pressure so that triglyme distilled at approximately 1300 C. The solvent must be pure or the carbene addition reaction will not occur. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 120 H. Hart and M. Kuzuya. Unpublished results. (a) L. deVries, J. Org; Chem., , 1838 (1960). (b) L. deVries, J. Am. Chem. Soc., 99, 2392 (19 1). (a) E.N. Prilezhaeva, V.A. Asovskaya, G.U. Stepanyanz, D. Mondeshla and R.J. Shekhtman, Tetrahedron Lett., 4909 (1969). (b) V.A. Asovskaya, A.U. Stepanyants, D. Mondeshla, R.J. Shekhtman, I. Koshel'shaya and E.N. Prilezhaeva, J. Org;gChem. (USSR), 9, 2470 (1970). Examples of reducing agents used include: LiA1H4, A1H3, LiBH sodium bis(2-methoxyethoxy)aluminum hydride (Vitride) and di-isobutylaluminum hydride (DiBal). 49 (a) K. Alder and G. Stein, Ann., 999, 1 (1934). (b) For an example of the ring system see Re erence 26 in H. Hart and M. Kuzuya, J. Am. Chem. Soc., 99, 2450 (1975). R.M. Silverstein and C.C. Bassler, "Spectrometric Identification of Organic Compound," John Wiley & Sons, Inc., New York, 1967. H.M.R. Hoffmann, Angew. Chem. Int. Ed. Engl., 9, 556 (1969). S. Krishnamurthy and H.C. Brown, J. Org. Chemu,z%, 3064 (1976) and references cited therein. E.J. Corey and K. Achiwa, J. Org. Chem., 99, 3667 (1969). (a) J. Boche and O. Runquist, J. Org, Chem., 99, 4285 (1968). (b) H. Takeshite, H. Kanamori and T. Hatsui, Tetrahedron Lett., 3139 (1973). (c) H.H. Wasserman and R. Kitzing, Tetrahedron Lett., 5315 (1969). P.G. Gassman and T.H. Johnson, J. Am. Chem. Soc., 99, 861 (1976). K. Tori, K. Aono, K. Kitahonoki, R. Muneyuki, Y. Takano, H. Tanida, and T. Tsuji, Tetrahedron Lett., 2921 (1966). H.C. Brown, J.H. Kawakami, and K.T. Liu, J. Am. Chem. Soc., 95, 2209 (1973). ”W (a) J. Meinwald, S.S. Labana, L.L. Labana, and C.H. Wahl, Tetrahedron Lett., 1789 (1965). (b) F.A.L. Anet, Can. J. Chem., 39, 789 (1961). mm (a) J. Meinwald, S.S. Labana, and M.S. Chadha, J. Am. Chem. Soc., 99, 582 (1963). (b) J.T. Lumb and C.H. Whitham, J. Chem. Soc., 1189 (1964). S. Winstein, J. Am. Chem. Soc., 99, 1516 (1961). 76. 77. 78. 79. 80. 81. 82. 121 (a) G. Whitesides and D. Feitler, J. Inorg. Chem., lé, 466 (1976)., (b) U. Burger, A. Delay, and F. Mazenod, Helv. Chim. Acta, 57, 2106 (1974) (c) R.B. King and A. Erfraty, J. Am. Chem. Soc?? 99, 3773 (1972). S. Krishnamurthy and H.C. Brown, J. Org. Chem., 2%, 3064 (1976). R.S. Rouse and W.B. Tyler, J. Org! Chem., 99, 3525 (1961). A.R.L. Bursics, E. Bursics-Szekeres, M. Murray and F.G.A. Stone, J. Fluorine Chem., 9, 619 (1976). Similar results were obtained when the photolysis was run in the presence of 1 M piperlyene. (a) M. Makosza and M. Fedorynski, Synthetic Comm., 9, 305 (1973). (b) M. Makosza and M. Wawrzyniewiz, Tetrahedron Lett., 4659 (1969). (c) C.C. Joshi, N. Singh and L.M. Pande, Tetrahedron Lett., 1461 (1972). In some cases the trifluoromethyl groups and/or the vinyl carbons to which they are attached were not observed in the carbon-l3 NMR. In general the peaks from these carbons are weak because of a lack of a Nuclear Overhauser Effect and because of splitting by fluorine.