THE i-MPORTANCE or cmaoumaaou BOND , HYPERCONJUGATTON IN THE SOLVOLYSIS OF endo—Z-NORBORNYL BROSYLATE. ‘_ Thesis for the Degree of Ph, D. MECHIGAN STATE UNIVERSITY JAMES DAVID RICHARDSON 1972 ‘. - 5' Michigan .330 Umversrty This is to certify that the thesis entitled THE IMPORTANCE OF CARBON-CARBON BOND HYPERCONJUGATION IN THE SOLVOLYSIS 0F endo -2-NORBONYL BROSYLATE presented by James David Richardson has been accepted towards fulfillment of the requirements for Ph.D. Chemistry degree in éf/lC/uzm Major professor Date January 8, 1973 0-7639 smor- av ’" HUAG & SUNS' 800K BlNUERY INC. LIBRARY BINDERS 'Tfi} 1%.4‘H‘i“i. “A “ -J. E NT‘ HYPPRLU! .. -'-' ‘\ ." “f 3': - IN“ .l: 1'.”- .e i”““ ~~ A? 17. Hrpr-r::‘n.i,£‘fim c: 4' : ,, s gut-or. udrbcn. u l-cvnds ‘18 been suggested 5:. .i major ’rwz‘mr or" tistiiizzzikcn fox" 0 olc¢tron-dctltinnr .CSlels p’Uu46nJ during certain type: nation. ’10 trts’ 1731-: suggcm (LP. an: have determined “Secondary Y-denterinv Motors #th -.‘.' for the smug-IV“! ” :00!“ 2vncrbornu-1 7 7»-'2 lrzvnlate, .' lt¢ Che Ottfitittli ’Clhliiith Flnfivéifi “we! {no ‘_' II «A, 4 .‘ 'G'C".J~ '1‘ “d “1) u‘ .X‘i‘fli'i ’1.'l.‘:y' lies the the Of; ABSTRACT THE IMPORTANCE OF CARBON-CARBON BOND HYPERCONJUGATION IN THE SOLVOLYSIS 0F endo-Z-NORBORNYL BROSYLATE BY James David Richardson Hyperconjugation of strained carbon-carbon 0 bonds has been suggested as a major source of stabilization for the electron-deficient centers produced during certain types of reactions. To test this suggestion, we have determined the secondary y—deuterium isotope effect for the solvolysis of endo-Z—norborny1-7,7-d2 brosylate, I. D D 085 The isotope effect (kH/kD = 1.015 t .009) was found to be essentially constant over a 15° temperature range. A comparison of the activation parameters revealed that the observed isotope effect was due to enthalpy differences between the isotopically related compounds. James David Richardson ‘. Although other possible explanations were examined, it _,u_:was concluded that the observed effect is best explained in . terms of C1-C7 bond hyperconjugation in the transition state .during the solvolysis of I. The importance of this result lies in its application to other systems. It now appears that much of the previous work involving the norbornyl system can be explained without resorting to the nonclassical carbonium ion concept. REFERENCES F. R. Jensen and B. E. Smart, J. Amer. Chem. $00., 2;, 5686 (1969). T. G. Traylor, W. Hanstein, H. J. Berwin, Nye A. Clinton, and R. S. Brown, ibid., 2;, 5715 (1971). THE IMPORTANCE OF CARBON-CARBON BOND HYPERCONJUGATION IN THE SOLVOLYSIS OF endo-Z-NORBORNYL BROSYLATE By James David Richardson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Ai,'. Department of Chemistry 2:. I j a 1972 .0 " - ‘ ., 1 I I l o : 1 ‘ . 3'.” W ' ill.:\tc m: u_ . 3;..2‘5. 9;. ~. "1 -:_'L I’l'.‘.“. if.» ’I'GCOSPC-‘i 1;. 'IL 1M "-. :11- 3:".‘wiz11'31's; and ' 0' i- - o 1 Y - - . -. ,f '. -: . COQ.LI.I.a-. 1"» 1"" .1, ' '* Ix." \.P‘"l"‘ ". \t-z ‘ ' l I - "r. a .5. inve111k1;;;n. 7” 1| ‘ ' I'M " mu... gun. ,. :2: ‘v 1th Kay-Ha - . "' Stienr. dune 1i1:r «in a , Tera rc.a.nt or Chemist.v. ' Mlchigsn “'3'“ V ‘t- , ; ‘ , iuhnuh rfigel if . T.“ Fl{;a1._;, 4.3“. A .1 :,~;1,.& 2-13- 5;.mg1y, "h' '15! The patience itii“. Ugh-aha“ "’. l.'-.' :-- hh. “235‘. . ‘ ’3“: Janet, anl the y wi'fu: envhr;an;n OT 2:» sons, sch“ and iqry, were a "In:;xn: smarts-oi “"ful 2nd thysrca; vtlajuvinltlon which Ha: uttcn hccuasufy during the _.”‘iybplraticn of It. sort A «p.c-a: u.kuoaieflgneml Cs :Jfio extended tn rae anther ’s Vethe:.v hlahaq! hat ACKNOWLEDGMENTS The author wishes to express his appreciation to ’Professor Gerasimos J. Karabatsos for his guidance and continued interest throughout the course of this investigation. The financial support provided by the National Science Foundation and the Department of Chemistry, Michigan State University, is also acknowledged. Finally, the author wishes to thank his family. The patience and understanding offered by his wife, Janet, and the youthful exuberance of his sons, John and .. Jeffery, were a constant source of mental and physical rejuvination which was often necessary during the . preparation of this work. A special acknowledgment is also extended to the author's Mother. Without her continued support and encouragement the author's current level of education would have been impossible. fl. P‘. e'. '/ ’A&:%EZ¢2. 9 ’AV‘ 9‘~-> an! “ I. 5.3.5:). lulI. j INTRODUCTION. . Controversy. A. a-Effects B. B-Effects C. y- and 6-H EXPERIMENTAL. . . . L Apparatus. . . II. Synthesis. Preparation52 norbornanol Preparation52 Preparation52 norbornanol Preparation53 PreparationS3 TABLE OF CONTENTS I. Origin of the Nonclassical Carbonium Ion II. Origin of Secondary Kinetic Deuterium Isotope Effects. ffects. IV. Carbon-Carbon Bond Hyperconjugation. of exo-Z—Methyl-endo-Z- of 1- -Methyl- exo- 2- norbornyl Acetate . . . . of l-Methyl-exo-Z- of a Chronic Acid Oxidizing Reagent . . . . . . . of l-Methylnorcamphor . iv III. Secondary Kinetic Deuterium Isotope Effects in the Norbornyl System. . . . . . 14 14 16 24 29 37 37 37 37 39 39 40 41 TABLE OF CONTENTS (Continued) Preparation54 of 1- -Methyl- 2- norbornanone- 3, 3-d2. . . . . . . . . . Preparation of 1- -Methyl- -endo- 2- norbornanol- 3 ,3-d2. . . . . . . . . . Preparation55 of Z-Methylenenorbornane- 7,7-d2- I a n o o o o o o I o o I o a I I Preparation56 of Norcamphor-7,7-d2. . . Preparation of Ethggolic Sodium Bisulfite Solution . . . . . . . . . . . Preparation of Sodium endo-Z-Hydroxy—exo- 2-Norborny1-7,7—d2 Sulfite. . . . . . . Hydrolysis of Sodium endo- 2 Hydroxy- exo- 2- ~norborny1- 7, 7— d2 Sulfite. . . . . . . Preparation of endo-Z-Norbornanol-7,7-d2. Preparation of endo-Z-Norborny1-7,7-d2 Brosylate . . . . . . . . . . . . . . Selective Solvolysis58 of a Mixture of endo- and exo-Z-Norborny1-7,7-d2 Brosylate. Kinetics . . . . . . . . . . . . . . . . . Preparation of Solvents . . . . Water. . . . . . . . . . . . . Ethanol. . . . . . . . . . . . . , Mixed Solvent. . . . Kinetic Apparatus . . . . . . . . . . . . Page 42 43 44 45 46 47 47 48 49 51 51 51 51 52 52 .i“) ‘ "Hv \. _ 7E4“ ”A“ Y QUACL ‘1 1i. ‘ ' “1,. 1'.!al ~51 ¢ 1 '3, RESULTS . '5 I. K . II. T 11. .,§;Dlscussron. Ef TABLE OF CONTENTS (Continued) COnstant Temperature Bath Temperature Measurement Time Measurements . . . . . . . . . . . Rate Determinations . . . . . . . . . . Synthesis. . . . . . . . . . . . . . . . inetics . . . . . . . . . . . . . . . . emperature Dependence of the Isotope feet 0 I O I I O C I C O O I C D D O 0 Interpretation of Isotope Effect . . . . ;§lBLIOGRAPHY; . . . . . . . . . . . . . . . . . . 54;,1 . ”fig; Page 52 . . 54 . . 54 . . 54 . . 56 . . 56 . . 60 . . 71 . . 71 . . 72 . . 79 . . ab 43.23: . -‘§?T "2:117:23; », TABLE III IV VI VII VIII LIST OF TABLES Page Secondary a-Deuterium Isotope Effects for Some Norbornyl Substrates . . . . . . . 15 8- Secondary Deuterium Isotope Effects in the Norbornyl System . . . . . . . . 17 B— d Isotope Effects for Solvolysis of 3, 3-Dimethy1- 2-buty1 Brosylate and 2- -Methyl- 2- chlorobutane . . . . . . . 22 Secondary y- and 6- Isotope Effects in the Norbornyl System. . . . . . . . 26 Rate Constants for Solvolysis of endo- 2- Norbornyl- 7, 7-d2Brosy1ate in 60% (v/v) Aqueous Ethanol2 . . . 66 Isotope Effects for Solvolysis of endo— 2- Norbornyl- 7, 7— d2 Brosylate at Different Temperatures. . . . . . 67 Activation Parameters for Solvolysis of endo- 2-Norborny1 Brosylates in 60% (v/v) Aqueous Ethanol . . . 69 Enthalpy and Entropy Differences Between Isotopically Related endo- Z— Norbornyl Brosylates. . . . . . . 70 LIST OF FIGURES Page Potential energy curve for decreasing force constant in transition state. . . . . A 9 . Potential energy curve for increasing force constant in transition state. . . . . 10 Beer's law plot DR IV-29. . . . . . . . .' . 63 " Plot of 1n (A -Am) vs time for solvolysis of endo-Z-norfiornyl brosylate . . . . . . . -65 PREFACE The primary objectives of this dissertation are two fold: (l) to ascertain the extent of carbon-carbon ,:;hond hyperconjugation in the solvolysis of endo-Z- r,‘norborny1 brosylate, and (2) to determine the temperature dependence of the secondary y-deuterium isotope effect g,in the solvolysis of endo-Lnorbornyl—7,7-d2 brosylate. ggAlso included are a brief synopsis of the origin of the a:nonc1assical carbonium ion controversy and a more ,=;Lthorough review of the past attempts to settle this controversy by the use of secondary kinetic deuterium isotope effects. INTRODUCTION 1. Origin of the Nonclassical Carbonium Ion Controversy In 1939, Wilson, et a1. studied the isomerization of camphene hydrochloride to isobornyl chloride in the presence of deuterium radio-chloride.1 Their results led them to make the following statement regarding the nature of the cationic intermediate formed in the reaction: "Although the intermediate ion is represented as having the camphene structure, it is possible that it is mesomeric between this and the corresponding isobornyl structure." Mesomeric Ion Although this statement had little impact at the time, it was the genesis of a concept which has elicited one of the most exhaustive research efforts2 in the recent history of organic chemistry, i.e. the nonclassical carbonium ion concept. 7n~.-—- < After lying dormant for nearly ten years, the nonclassical carbonium ion was revived by Winstein and Trifan. In a thorough study?)-6 of the solvolysis of norbornyl brosylates, they obtained the following results: (i) The rate of solvolysis of exo-Z-Norbornyl brosylate exceeded the rate for endo-Z- norbornyl brosylate by a factor of 350. (ii) The products formed in the solvolysis of both endo- and emo-Z-norbornyl brosylates were of exo configuration (>99.98%). (iii) Solvolysis of optically active endo-Z- norbornyl brosylate in acetic acid resulted in 93% racemization, whereas acetolysis of optically active exo-Z-norbornyl brosylate results in >99.95% racemization. (iv) The rate of racemization, ka’ of the endo compound was essentially equal to its rate of solvolysis, k for the era isomer, t’ ka/kt was 3.46 in acetic acid, 2.94 in ethanol, and 1.40 in 75% aqueous acetone. The high exo/endo rate ratio, coupled with the difference in the ka/kt ratios, indicated to them that the exo- and endo-Z-norbornyl brosylates solvolyzed via different pathways. To explain what they considered to be an abnormally high rate for the solvolysis of exo-Z-norbornyl brosylate, they suggested that this isomer experienced anchimeric 5A: 1': ‘ ?"”" A k assistance from the electrons in the Cl-C6 0 bond. The resulting bridged intermediate prevented solvent attack from the endo side of the molecule, thus leading exclusively to exo products. Due to a lack of the proper geometry, the endo-Z-norbornyl brosylate does not experience anchimeric assistance during solvolysis and, therefore, ionizes at a normal rate to a classical, unbridged intermediate. This intermediate can either react with solvent to give product, or it can collapse to the more stable nonclassical ion which leads to racemic products upon attack by solvent. The over- all mechanism for solvolysis of the exo-isomer is shown in Scheme I, while the mechanism for solvolysis of the endo epimer is shown in Scheme II. Scheme 1. Mechanism for Solvolysis of exo-Z-Norbornyl Brosylate Bs ’ggbg,437 __El;> if, k2 -1 "' 39% 2 Bs where k_1 > R2 Scheme 11. Mechanism for Solvolysis of endo-Z-Norbornyl Brosylate 1 HOAC ___A Eb VIE—1‘ + 2 55/0“ OBs -OBs H I HOAC HOAC > Acgl//P::7 E k4 ,-}:;] k4 OAC H :o‘ Bs where k3 > k2 Support for the nonclassical carbonium ion has come from semiemperical methods for calculating solvolysis rates,7—9 from substituent effects,1 11-12 0 from molecular orbital calculations, 13-14 and from spectroscopic data. Perhaps the most conclusive evidence for the nonclassical ion was obtained by Olah and his coworkers from their studies of the spectroscopic properties of the ion.14 When generated in SbFS-SO2 solution at -78°, carbonium ions are sufficiently stable to allow spectroscopic studies to be made. Using this technique, Olah studied the norbornyl cation by proton nmr, LEW—'— carbon-13 nmr, and Raman spectrosc0py. He determined that the actual structure of the norbornyl cation, under these conditions, was that of a corner-protonated nortricyclene, III. He estimated that a minimum of 5.8 kcal/mole of III stabilization energy is obtained by 0 bond delocalization. Although the concept of the nonclassical carbonium ion has gained widespread acceptance, there are still those who feel that the evidence for such an intermediate is not quite so compelling. Foremost among these "dissenters" is H. C. Brown.”-16 It is Brown's contention that most, if not all, of the evidence for the nonclassical ion can be explained equally as well by assuming the cation to exist as a pair of rapidly equilibrating classical ions, IV. IV Brown's argument is predicated on the existance of rather severe steric interactions on the endo side of the norbornyl A substrate. In the solvolysis of II, these interactions are increased as the leaving group departs and, consequently, the reaction rate is decelerated substantially. These same steric interactions, along with the "windshield wiper" effect of the rapidly equilibrating cations, are also proposed to account for the stereospecificity of product formation. Brown has used the following experimental data to support his hypothesis: (1) Comparison with model compounds shows that the solvolysis rate for I is nearly what would be expected, and that the rate for II is abnormally low. (ii) The use of substituent effects indicates that a high exo/endo rate ratio is obtained even in cases where nonclassical ions are not expected to intervene. (iii) Data from other reactions indicate that attack from the era side of the molecule is preferred even in reactions which do not involve carbonium ions. Thus, it appears that Olah's spectroscopic data are the only seemingly irrefutable evidence for the existance of the nonclassical ion. Although this evidence is substantial, it should be kept in mind that the behavior of a cation in SbFS-SO2 solution may be quite different from the behavior of the same cation in solvents which are normally used for solvolysis reactions. The SbFS-SO2 is a very poor medium for solvating a positive charge. The lack of any appreciable solvent stabilization may cause the norbornyl cation to look elsewhere in an effort to lower its energy. The most likely alternative is the pair of electrons in the Cl-C6 bond, which can be delocalized to distribute the positive charge between C1 and C2. In the foregoing discussion, most of the evidence arose from experiments in which either the substrate or the solvent was changed, or in which the behavior of model compounds was compared with the behavior of the norbornyl substrates in solvolysis reactions. A much more subtle method for mechanistic studies is the use of secondary kinetic deuterium isotope effects. This method has been used extensively to study the mechanism of the solvolysis of norbornyl brosylates. The efficacy of kinetic isotope effects for determining the nature of the norbornyl cation will be discussed in the following sections. 11. Origin of Secondary Kinetic Deuterium Isotope Effects Substitution of deuterium for hydrogen on a C—H bond a, B, or occasionally y to the reactive center of an organic molecule often causes a change in the rate at which that molecule undergoes a given reaction. The difference in rates between the isotopically substituted and unsubstituted compound is called the kinetic deuterium isotope effect and the magnitude of this effect is believed to have definite mechanistic implications. The use of secondary kinetic isotope effects for mechanistic interpretations presupposes that the exact origin of these effects is known. The theory behind kinetic isotope effects has been subjected to the scrutiny of several investigators.17 In general, secondary isotope effects are believed to arise from force constant changes in the C-H (D) bond as the molecule proceeds from the ground state to the transition state during a reaction. Although the potentaal energy surface, and therefore the force constant, is the same for both C-H and C-D bonds, there is a difference in the zero-point vibrational energy (AZPE) between the two bonds. If the force constant of the bonds changes during the activation process, then the AZPE will also change. it is this difference in AZPE between the ground state and the transition state that causes the difference in reaction rate between the isotopically labeled compounds. Thus, Figure I shows that a weaker force constant in the transition state results in a decrease in AZPE. Consequently, AG; < AG; and kH/kD > 1.00 (normal isotope effect). Conversely, Figure II shows that a stronger force constant in the transition state causes an increase in AZPE. In this case, AG; > AG; and kH/kD < 1.00 (inverse isotope effect). A great deal of research has been carried out in an effort to determine exactly what causes these force constant Potential Energy 9 t t AGH < AGD f kH/kD > 1 q. AGD ¢ AGH Y / Figure 1. Reaction coordinate Potential energy curve for decreasing force constant in transition state. 10 4 ___--_A---_ _ _ Reaction coordinate ' : ~::;;r 'garocpotential energy curve for increasing- - .='f¥" , Vforce constant in transition state. - UN”... ~ ' ‘F'— '-.- . }— 57 Q ~ r... n - ‘ . VI :4 .1? .. , \ I " 1" T .. V , 4 1': . . l; I. . v‘» gr... .- t 5‘9” l ‘ "‘ 11 changes. For a-effects, it is generally agreed that the isotope effect is due to changes in the CHX bending force 18 Thus, the spz-sp2 constant during the activation process. rehybridization that occurs during an SNl type reaction causes a loosening of the out—of—plane bending force constant and, consequently, leads to a normal isotope effect. For SNZ type reactions, the decrease in force constant due to departure of the leaving group is counteracted by the formation of a new force constant with the entering nucleophile. Therefore, the isotope effect for an SNZ reaction is expected to be near 1.00. Experimentally, the a-isotope effect for SNZ reactions has been found to vary from slightly inverse, to unity, to slightly normal. Usually, however, the effect is very close to 1.00. The origin of secondary B-isotope effects is somewhat more complicated. The force constant change has been attributed to inductive effects,19 hyperconjugation,20 and/or nonbonded interactions.21 Although all of these factors may contribute to the B-effect, hyperconjugation has been shown to be the most important.22 In any event, it is agreed that B-effects are a function of the magnitude of the positive charge at the reaction center. For normal SNl type reactions an isotope effect of ~1.15 per deuterium is expected, whereas for SNZ type reactions a very small B-isotope effect is expected. 12 There are a few examples of y-isotope effects in solvolysis reactions.23 These have been interpreted as arising from inductive effects, nonbonded interactions, and charge delocalization as a result of o-bond participation. It should be evident that no single straight-forward explanation exists for all secondary deuterium isotope effects. One thing does seem clear, however; the magnitude of a-, 8-, and y-isotope effects is sensitive to the amount of positive charge which is localized at the reaction center. When a considerable positive charge is formed during the reaction, both a- and 8- effects are expected to be in the range 1.10-1.20. When little positive charge is centered at the a—carbon, the isotope effects are expected to be close to 1.00. It is this sensitivity to a localized positive charge that makes secondary kinetic deuterium isotope effects useful for studying anchimeric assistance. This can be seen clearly by examining the two possible transition states for the solvolysis of exo-Z-norbornyl brosylate. The transition state, V, in which there is no anchimeric assistance, has a r ”‘1‘ ‘5. 13 localized positive charge on C2. Therefore, isotopic substitution at either C2 or C3 should result in an isotope effect similar to that observed in most SNl type reactions, i.e. 1.12-1.20. Transition state VI, in which anchimeric assistance is present, has the positive charge distributed between C2 and C1. Also, since a partial bond is formed between C6 and C2, another force constant is being introduced. The delocalized positive charge, along with the formation of an additional force constant at C2, should result in a much smaller a- and B- isotope effect. Of course, in order to properly interpret the isotope effects observed for the solvolysis of exo-Z-norbornyl brosylate, it is necessary to know what the effect would be if the transition state were classical. It is for this reason that the isotope effects for solvolysis of endo-Z-norbornyl brosylate are also determined. It is generally agreed that the endo isomer ionizes to the classical cation. Therefore, the isotope effect for this process is used as the standard to which the effects for the era isomer are compared. Kinetic isotope effects have been used with varying degrees of success to ascertain the amount of anchimeric assistance involved in the solvolysis of several substrates.24 In the following section, only those studies involving the norbornyl system will be discussed. ——v— 14 III. Secondary Kinetic Deuterium Isotope Effects in the Norbornyl System. A. a—Effects The norbornyl compounds on which secondary o-isotope effects have been measured are shown in Table I. The initial study of the a-effect in solvolysis of norbornyl brosylate was carried out by Lee and Wong26 (entries 3 and 4 in Table I). Since the a-effect for the era-brosylate was considerably smaller than that for the endo-brosylate, they concluded that participation did occur in the solvolysis of the exo isomer. However, they acknowledged that scrambling of the deuterium, due to internal return, may also be a factor in the lower a-effect for the exo compound. This 27 Internal possibility was tested by Murr and Conkling. return is known to occur to the extent of 78% in acetic acid, but only 35% in 80% ethanol. Therefore, if scrambling due to internal return is responsible for the low a-effect in acetic acid, then the a-effect should be higher in 80% ethanol. Examples 5 and 6 in Table I indicate that the a-isotope effect is essentially solvent independent. Therefore, they concluded that the lower a—effect for the era isomer, combined with the lack of solvent dependence of this effect, definitely shows that participation does occur in the solvolysis of the era-brosylate. 15 Table I. Secondary a-Deuterium Isotope Effects for Some Norbornyl Substrates \ Substrate Solvent kH/kD Reference / 1. HOAC 1.21 25 0B5 // 2. OBs HOAC 1.21 25 D 3. 1::E::Z6D HOAC 1.20 26 Bs / 4. CBS HOAC 1.07 26 5. 80% EtOH 1.19 27 D HOAC 1.20 Bs 6. B 80% EtOH 1.124 27 5 HOAC 1.118 D 7. 5&1) HOAC 1.20 28 CBS 8. [::E::Z.OB5 HOAC 1.20 28 D 16 28 have shown In a more recent study, Sunko and Borcic that the a-effect for both endo- and exo-Z-norbornyl brosylate is approximately the same. They found that by measuring initial rates for the era epimer, the a-isotope effect was 1.208, while measurements over a more extended period (35-50% of the reaction) resulted in an a—effect of only 1.106. They attributed this discrepancy to internal 25 that the a-effect for both 1- return. They also found methyl-endo—norbornyl and exo-Z-norbornyl brosylate was 1.21. They interpreted their data as showing that "participation of the Cl—C6 bond occurs from a distance which does not allow interference with the C-H (D) bending motions at the reacting center.” Thus far, all of the experimental data from a-isotope effects have been interpreted as showing the existance of o-bond participation in the solvolysis of exo-Z-norbornyl brosylate. However, due to the lack of consistency in these data, it would seem that any conclusions derived from them would be purely speculative. B. B-Effects Secondary B-isotope effects have been used extensively to study the mechanism of the solvolysis of exo- and endo-Z- norbornyl derivatives. The results of these studies are given in Table II. As was mentioned in a previous section, anchimeric assistance by the Cl-C6 a bond during ionization of the ewe isomer would cause delocalization of the incipient 17 Table II. B-Secondary Deuterium Isotope Effects in the Norbornyl System Compound Solvent kH/kD Reference D D 9 CBS HOAC 1.014 29 D D 10. HOAC 1.26 29 0B5 D H 11. 80% EtOH 1.11 30 0B5 H D 12. B5 80% EtOH 1.02 30 D D 13. B5 80% EtOH 1.11 30 H 14. 80% EtOH 1.19 30 0B5 H z’ D 15. 80% EtOH 1.12 30 85 D / D 16. 80% EtOH 1.31 30 B S w 18 Table II. (Continued) Compound Solvent kH/kD Reference D / D 17. Br HOAC 1.09 31 D D 18. 50% EtOH 1.23 31 Br D 19. Ts HOAC 1.16 32 ” D 20. HOAC 1.29 32 OTs / D 21. 70% Acetone 1.33 32 OPNB D D 22. 70% Acetone 1.31 32 PNB D 23 D 60% D' l 18 33 . OPNB 1oxane . ¢ D D 24. 60% Dioxane 1.15 33 ¢ OPNB 19 positive charge at C2. If this delocalization occurs, the B-isotOpe effect for the era substrate would be expected to be somewhat smaller than the B-effect for the endo isomer. Examples 9-10 and 17-18 in Table II indicate that this is exactly what was observed. The conclusion reached in both studieszg’31 was that participation does occur during ionization of exo-Z-norbornyl derivatives. This conclusion was supported by the work of Murr and Conkling.30 They determined the contribution of each of the B-Cz-H(D) bonds (i.e. exo—3-d and endo-3-d) to the total B-isotope effect for both endo- and exo-Z-norbornyl brosylates. The results of their work appear as examples 11-16 in Table II. For both endo- and exo-Z-norbornyl brosylate, the exo-C3-D bond is the major contributor to the overall B-effect. For the exo-brosylate, the total B-effect is due almost entirely to the exo-C3-D bond. They attributed this to the fact that only the exo-C3-H bond is properly alligned for hyper- conjugation with the developing p—orbital of the nonclassical ion. If one considers Newman projections along the Cz-C3 bond, the stereochemistry which results from bridging is shown in (a) and from classical ion formation is shown in (b). In (a) the era-Cs-H is coplanar with the p-orbital, while the Hexo H exo H H H endo Hendo (a) (b) 20 endo—C3-H is at an angle of 60° to the p-orbital. In (b) both exo— and endo-Cs-H bonds are displaced 30° from coplanarity with the developing p—orbital. Secondary B- isotope effects are known to be dependent upon the dihedral angle34 involving the developing p-orbital and the C-H bond. For a dihedral angle of 60° the isotope effect expected is 1.01—l.03. Thus, the substantial effect for the exo-C -D 3 and the small effect for the endo-CS-D can be explained by a transition state similar to the nonclassical ion inter- mediate. Olah's spectroscopic studies14 showed that substitution of a methyl or phenyl group at C2 caused the amount of charge delocalization by o-bond participation to be greatly reduced. Hence, one might expect that the B-effect for tertiary norbornyl substrates would be larger than the effect for the corresponding secondary substrate. Examples 21-24 in Table II indicate that this is generally true.32’33 Although the B-effect for the methyl-substituted compound is considerably higher than the effect for the phenyl-substituted compound, the significant factor is that, for both compounds, the effect for the exo isomer is essentially the same as the effect for its endo epimer. Therefore, the transition states for the solvolysis of the two isomers must be structurally similar. The B-effects for exo-Z-norbornyl derivatives have also been shown to be sensitive to methyl substitution at €1.32 21 Although the effects are not as dramatic as with substitution at C2, they do increase by four per cent upon such substitution. According to Sunko, ionization of exo-Z- norbornyl substates may be accompanied by three separate processes: (1) ionization of the C—X bond and development of a positive charge at the reaction center, (ii) charge delocalization, and (iii) bridging. He proposes that all three occur simultaneously for unsubstituted exo-Z-norbornyl isomers, while only (1) occurs for the Z-methyl-Z-norbornyl substrates. For the 1—methyl-emo-2-norbornyl derivatives he proposes that ionization is accompanied by some charge delocalization, but that bridging has not as yet advanced enough to cause complete elimination of the B-effect. Scheppele35 has recently pointed out what he regards as a "major inconsistency" in the B-effects observed in the norbornyl system. As mentioned previously, bridging is not present in tertiary norbornyl substrates. Apparently the Z—methyl or Z-phenyl groups offer more stabilization of the incipient cation than can be obtained by delocalization due to Cl-C6 o-bond participation. Scheppele feels that this added stabilization should effectively decrease the B-effect rather than increase it. As we have already seen, this is not the case. While this argument is not unreasonable, existing data indicate that it is not necessarily correct. For instance, the solvolysis of 3,3—dimethyl-2-buty1 brosylate is believed to proceed via a rate determining ionization of 22 36 followed by rearrangement. Like the norbornyl the brosylate, substrates, it is a secondary brosylate which ionizes with little or no nucleophilic solvent participation. The B-d3 isotope effect for this reaction is shown in Table III along with the B—d3 effect for solvolysis of Z-methyl-Z-chloro- butane.37 Addition of a methyl group at C2 in 3,3-dimethy1-2- Table III. B-d3 Isotope Effects for Solvolysis of 3,3-Dimethyl-2-butyl Brosylate and Z-Methyl-Z-chlorobutane Compound Solvent kH/kD B-d3 3,3-Dimethyl-2-butyl Brosylate 50% EtOH 1.205 Z-Methyl-Z-chlorobutane 80% EtOH 1.34 butyl brosylate should offer added stabilization of the positive charge at that position. According to Scheppele's argument, this added stabilization should decrease the demand for hyperconjugative stabilization by the B-CD3 group. Consequently, the B-d3 isotope effect should decrease. However, the B-d3 effect for a similar tertiary system (i.e. Z-methyl—Z-chlorobutane) has actually been shown to be b.” 23 larger than that for the secondary system. A more appropriate test of Scheppele's hypothesis might be a comparison of the B-d3 effects for compounds VII and VIII. Compound VII is believed to ionize with phenyl participation, whereas, participation should not occur in VIII. Scheppele's hypothesis would predict a lower B-isotope effect for VIII than for VII. Unfortunately, the isotope effects necessary CH3 CH3—CH-fH-CH3(d3) CHs-CH-CH-CH3 OPNB PNB VII VIII for this comparison have never been determined. Like the a-isotope effects, all of the secondary B- isotope effects in the norbornyl system have been interpreted as supporting the nonclassical ion concept. Although there is some variation in the value of the B-effect for identical systems (i.e. compare 9 with 13, and 10 with 16), the same general trend was always found. However, if one compares the interpretations given for the B-effects with those for the a-effects, some inconsistencies are apparent. For instance, Sunko has suggested that the small B-effect for exo-Z- norbornyl brosylate is due to the simultaneous occurrence 24 of ionization, charge delocalization, and bridging during solvolysis of this compound. On the other hand, this same author suggested that bridging and charge delocalization lag behind ionization sufficiently so as not to cause a decrease in the a-effect for this compound. Obviously, one of these explanations is incorrect. This inconsistency may be circumvented if one assumes that Sunko's data for the a-effect are incorrect and that the correct value is actually 1.12 as found by Murr. If this is the case, then the proposal that ionization is accompanied by bridging and delocalization seems to explain both a- and B-effects. However, there may also be a discrepancy in this explanation. Murr determined the a-effect to be 1.12 and the B-effect to be 1.11 for the exo-CS-D bond. It is somewhat surprising that the B-effect is as large as the a-effect, especially when the B-C-D is cis-periplanar to the incipient p-orbital as opposed to the preferred trans-periplanar arrangement. Under these circumstances, an a-effect of 1.20 might seem more reasonable. Therefore, regardless of which value one chooses for the a-isotOpe effect, it is still impossible to make any irrefutable statements regarding the mechanism of the solvolysis of exo-Z-norbornyl brosylates. C. y- and G-Effects Secondary isotope effects in which the deuterium is more than two bonds removed from the reaction center have received comparatively little attention. However, the 25 y-isotope effects are perhaps the most significant of all the isotope effect studies in the norbornyl system.38’39 The 7- and G-effects which have been determined are given in Table IV. The y-effects, like the a- and B-effects, have been interpreted as proof of o—bond participation in the ionization of exo-Z-norbornyl brosylate. This interpretation is based on the fact that a substantial y-effect is found for the era-brosylate, while the y-effect for the endo epimer is essentially nil. Participation by the Cl-C6 bond would cause a change in the electron density at C6' This in turn, would cause a change in the force constants for the two C6-H bonds. Since the nonclassical norbornonium ion is symmetrical around C both the exo- and endo—C6-H bonds 6’ should be effected to the same extent. Examples 25 and 26 show that this is essentially what is observed. For endo-Z-norbornyl brosylate the y-isotope effect is nearly 1.00. According to Brown's theory, the nonbonded interactions between the endo-leaving group and the endo-C6-H ‘are increased during ionization. On this basis, Brown predicted15 that a substantial inverse isotope effect should be observed for endo-Z-norbornyl-endo—é-d brosylate, but very little effect when the deuterium is in the era-position. Examples 27 and 28 show that this is not the case. Although there is a slight inverse effect, it is not nearly as large as one might expect if severe nonbonded interactions are built up during ionization. Table IV. Secondary y- and G-Isotope Effects in the Norbornyl System 26 Compound Solvent kH/kD Reference 25. 06b 085 HOAC 1.09 38 80: EtOH 1.09 39 26. 085 HOAC 1.11 38 80: EtOH 1.11 39 27. ”\fib HOAC 0.98 38 085 80% EtOH 1.00 39 28. HOAC 0.99 38 80% EtOH 0.97 39 Bs 29. fight)?“ 90% Acetone 1.02 40 n 30. 085 HOAC 1.09 41 31. bib, HOAC 1.00 41 Bs D 32. D 0135 HOAC 0.99 42 n o 33. Ago“ HOAC 0.998 43 27 Further support for the nonclassical ion is shown by 29. The y-effect for the tertiary substrate, which ionizes to a classical ion, is much smaller than the y-effect for the secondary substrate. Therefore, the transition state for the two compounds must be different. The secondary G-effect is shown by 30, 31, and 32. As might be expected, deuteration at C5 has little effect on the rate of ionization of either endo- or exo-Z-norbornyl brosylates. The y—isotope effect has also been determined for deuteration at C7. Example 33 shows the C7- y-effect for l-methyl-exo-Z-norbornyl tosylate. This effect was interpreted as showing that bridging was not very far advanced at the transition state for this compound. However, this result could be explained equally well by proposing that the compound ionizes directly to a classical ion, or that any positive charge which is delocalized to C1 is stabilized by the 1-methyl substituent to the extent that only a slight inductive effect arises from the two C7—D bonds. Schaefer attempted to measure the C7 y-isotope effect for the unsubstituted norbornyl bromide.31 He took advantage of the fact that, due to internal return, C3 and C7 become equivalent during the course of the reaction. Thus, by starting with exo-Z-norbornyl bromide-3,3—d2 and reacting 28 D D D /, D > D - ____;s B Br <:___— ’ ’,—-Br < r under conditions which promote internal return, he was able to approximate the y-effect at C7. The effect he observed was approximately 1.00. He concluded that the molecular reorganization necessary to produce an isotope effect at C7 was insufficient in the transition state to be detected by his method. Although the y-effects at C6 are fairly convincing evidence for the existence of the nonclassical ion, they are not necessarily conclusive. For example, the near equivalence of the y-effect for both endo- and exo-C6—D was deemed normal because, in the nonclassical ion, both of the hydrogens at C6 are equivalent. This is certainly the case for the nonclassical ion intermediate, IX. However, H); _: IX bridging is not necessarily complete in the transition state and, therefore, these hydrogens may not be equivalent at 29 this point. If this is true, one might expect different y-effects from the two epimeric hydrogens. Another point which should be mentioned is the lack of any temperature studies on these isotope effects. It has recently been shown44 that many secondary isotope effects, especially those involving nonbonded interactions, are much more temperature dependent than was previously believed. Most of the isotope effects in the norbornyl system have been determined at one or, at most, two temperatures. Therefore, any temperature dependence which may exist has gone unnoticed. So far, all of the isotope effect data have been interpreted as supporting the nonclassical ion concept. Another possible explanation for this and other data will be discussed in the next section. IV. Carbon-Carbon Bond Hyperconjugation In a study of the relative rates of aluminum chloride catalyzed benzoylation of para-alkylbenzenes,45 Jensen and Smart found that all of the isomeric norbornyl benzenes reacted faster than did toluene. This is contrary to the order which would be predicted on the basis of the Baker- Nathan effect. They concluded that, in the norbornyl system, carbon-carbon bond hyperconjugation is more effective at stabilizing a positive charge than is carbon-hydrogen bond hyperconjugation. Furthermore, they suggested that this 30 might be true for all systems in which the carbon-carbon 0 bond contains more than the normal 75% p character (i.e. strained ring compounds). Jensen and Smart also suggested that carbon-carbon (hereafter referred to as C-C) bond hyperconjugation might be a factor in the high exo/endo rate ratio in the solvolysis of norbornyl derivatives.46 Hyperconjugation is known to be dependent upon the stereochemical relationship between the leaving group and the hyperconjugating bond. The optimum allignment is when the hyperconjugating group is trans- periplanar47 to the leaving group. Newman projections of the Cl-C2 bond illustrate the stereochemistry around C for both 2 exo- and endo-Z-norbornyl brosylate. The era isomer has the OBS C7 H C3 C7 H 3 C C6 0% 6 exo-Z-Norbornyl Brosylate endo-Z-Norbornyl Brosylate Cl—C6 bond in the trans-periplanar allignment, while the endo isomer has only the Cl-C7 bond in the trans position and this bond is displaced several degrees from the ideal allignment. Therefore, hyperconjugative stabilization should be greater for the era-isomer than for the endo—isomer. 31 The stereoelectronic dependence of hyperconjugative stabilization has been dramatically illustrated by Schleyer 8 They devised a method for correlating the and Bingham.4 carbonium ion reactivities of bridgehead systems by quantitative conformational analysis. The method worked well for sixteen bridgehead systems over a reactivity range 18 of 10. The only significant deviation was with 10-tricyclo E.2.l.04’10]decyl tosylate, X. This compound solvolyzed 109 slower than the predicted rate. The only difference between X and the other fifteen bridgehead substrates was in the stereochemistry of the bonds 8 to the leaving group. Thus, for X the torsional arrangement around the leaving group is shown in (a), while the arrangement around the leaving group C(H) (a) (b) 32 in the other fifteen substrates is depicted in (b). Therefore, they concluded that the cis—periplanar allignment of the B-C-H bond caused a decrease in the amount of hyper- conjugative stabilization offered by this bond to the developing cation. Traylor has also advocated C-C bond hyperconjugation as 49 He a major stabilizing factor in solvolysis reactions. used data from charge transfer processes to try to distinguish between bridging and what he refers to as o-w conjugation (hyperconjugation). According to the Frank-Condon principle, movement of atoms is not allowed during charge transfer reactions. Therefore, any stabilization by alkyl groups of the positive charge formed in these processes must be the result of hyperconjugation rather than bridging. Traylor measured the charge transfer frequencies for (l), where R was a series of Strained-ring alkyl groups. He obtained an RO _._, R4: TCNE TCNE" excellent correlation by plotting the observed frequencies vs the relative rate constants for the solvolysis of the substituted alkyl groups. Apparently the same stabilizing phenomenon which is operative in the solvolysis of these compounds, also exists in the charge transfer processes. He concluded that, particularly in strained-ring compounds, 33 C-C bond hyperconjugation is very effective at stabilizing neighboring cations. The effectiveness of C-C bond hyperconjugation has also been shown by theoretical calculations. Pople, et al. used ab initio calculations to determine the preferred 50 conformation of the propyl cation. They found that conformation a, in which the B-C-C bond is coplanar with the CH3 CH3 empty p orbital, is favored by 2.3 kcal/mole over 6. Also, Danen used semiemperical INDO calculations to show that the bisected conformation, g, for the cyclopropylcarbinyl cation is favored over the perpendicular conformation, d, by 34 41.15 kcal/mole. Both authors concluded that C-C bond hyperconjugation was more effective than was C—H bond hyperconjugation. Upon close examination, it appears that most of the experimental data which have been used to ”prove" the existence of the nonclassical norbornonium ion can be interpreted equally well in terms of C-C bond hyperconjugation. For instance, the hyperconjugation model proposes that the transition state formed from exo-Z-norbornyl brosylate is stabilized very well by Cl-C6 bond hyperconjugation, while the transition state from the endo—brosylate does not receive this type of stabilization. Consequently, the demand for hyperconjugative stabilization from the B-CS-H bonds should be much greater for the endo-brosylate than for its exo epimer. This would result in a lower B-isotope effect for exo-Z-norbornyl brosylate, which agrees with the experimental observations. The y-effect may similarly be rationalized as follows: Hyperconjugation, involving the Cl-C6 bond, which may be pictorially represented by the two resonance structures shown in XI, decreases the electron density at C6 during the 5b, <———>/ XI 35 process. This should decrease the force constants of the two C6-H bonds, thus leading to a normal isotope effect on deuteration at C6' Since this type of resonance does not exist in solvolysis of the endo-brosylate, no y-isotope effect would be expected for this isomer. Again, this is what was observed. Since C-C bond hyperconjugation and nonclassical ion formation have essentially the same effect on the solvolysis of exo-Z-norbornyl substrates, it is difficult to assess the relative importance of either. Such assessment can be made only from an experiment for which the two concepts would have different effects. One possibility is in the solvolysis of endo-Z-norbornyl substrates. Nonclassical ion formation does not occur in the rate determining step of the solvolysis of these compounds. Therefore, this concept can not be used to explain any type of kinetic data obtained for this process. Due to the unfavorable allignment of the Cl-C7 bond, C-C bond hyperconjugation is also expected to be greatly reduced in the solvolysis of endo isomers. However, Shiner has predicted that some C-H bond hyperconjugation will occur even when the dihedral angle between the C-H and C—X bonds is 60°. There- fore, we would not necessarily expect hyperconjugation of the Cl-C7 bond to be totally nonexistent. To test this hypothesis, we plan to measure the secondary y-deuterium isotope effect in the solvolysis of endo-2- norbornyl-7,7-d2 brosylate, XII. If Cl-C7 bond hyper- r . I 36 0B5 XII conjugation occurs during the solvolysis of XII, then a small normal isotope effect should be observed. We will also study the temperature dependence of the isotope effect on XII. Secondary deuterium isotope effects are normally attributed to AAH¢ (i.e. AHE-AHE), with AAS* (i.e. ASE-ASE) assumed to be nearly zero. However, recent studies have shown that this is not always true. In fact, for some systems TAAS* actually contributes more to the isotope effect than does AAH*. For systems in which AAS* # 0, determination of the isotope effect at only one temperature can lead to conclusions which may be incorrect. '. A I .“-.' _ EXPERIMENTAL 1. Apparatus The instruments used for the kinetic study are descrbied elsewhere. All nmr spectra were recorded on a Varian T-60 NMR Spectrometer. Both deuterochloroform and carbon tetra- chloride were used as solvents and TMS was the standard. Gas chromatographic analysis were performed on an F 8 M Scientific 700 Laboratory Chromatograph. The chromatograph was equipped with a thermal conductivity detector and a Sargent, Model SR recorder. The carrier gas was helium, and the column used in all cases was a 15' x 1/4" stainless steel coil, packed with 25% SE-30 on Chromasorb W. All melting points were taken in a Hoover Capillary Melting Point Apparatus, manufactured by A. H. Thomas Company. II. Synthesis Preparation52 of exo-Z—Methyl—endo-Z—norbornanol A 1000 ml three-necked round-bottomed flask was equipped with a mechanical stirrer, a reflux condenser and dropping funnel. To the flask was added 12.0 g (0.494 mol) magnesium metal (powder, 40 mesh, Mallinckrodt). The entire apparatus was thoroughly flamed to dryness, followed by purging with 37 38 dry nitrogen. To the dry flask was added, dropwise, a solution of 63.6 g (28 ml, 0.45 mol) methyl iodide (Aldrich) in 150 m1 of anhydrous ether (dried by distillation over lithium aluminum hydride). After the addition was complete, the mixture was refluxed for 1 hr. The flask containing the Grignard reagent was then cooled in an ice-bath. To the cold mixture was added, dropwise, a solution of 44.0 g (0.40 mol) norcamphor (Columbia Organic) in 200 m1 of anhydrous ether. Following addition, the mixture was refluxed gently for three hours. The flask and contents were cooled to room temperature and the reaction was quenched by adding saturated ammonium chloride solution. The resulting mixture was filtered and the precipitate was washed thoroughly with ether. The combined filtrate and washings was washed with two 50 ml- portions of saturated sodium chloride solution, then dried over molecular sieve (8-12 mesh, 'Baker Analyzed' Reagent). The dried solution was poured into a round-bottomed flask and the ether was removed by distillation at atmospheric pressure. The residue was distilled at reduced pressure. A total of 46.25 g (0.367 mol, 91.6% theoretical amount) exo-Z- methyl-endo-Z-norbornanol (light yellow oil, b.p. 78° at 14 mm; reported b.p. 70.5-71.0 at 10 mm) was obtained. The product appeared homogeneous to vpc analysis. 39 Preparation52 of l-Methyl-exo-Z-norbornyl Acetate To a 500 ml round-bottomed flask equipped with reflux condenser and magnetic stirring bar were added 45 g (0.357 mol) exo-Z-methyl-endo-Z-norbornanol, 129 ml (135.3 g, 2.24 mol) glacial acetic acid (Fisher Scientific) and 33 drops of 75% sulfuric acid. This solution was heated, with stirring, at loo-110° for 60 minutes. The flask and contents were then cooled in the refrigerator. The cold solution was poured into a large beaker containing 400 ml distilled water and the solution was neutralized with sodium bicarbonate. The neutral solution was extracted with eight 50 ml-portions of ether. The combined ether extracts were washed four times with 50 ml- portions of saturated sodium chloride solution, then dried over anhydrous magnesium sulfate ('Baker Analyzed' Reagent). The product was not isolated, but the ether solution was concentrated to approximately 200 ml by distillation at atmospheric pressure. Analysis of the concentrated solution by vpc showed the presence of 8% unreacted alcohol. Preparation52 of l-Methyl-exo-Z-norbornanol A 1000 m1 three-necked round-bottomed flask was equipped with a reflux condenser, a mechanical stirrer and a dropping funnel. The entire apparatus was dried and purged with dry nitrogen. To the dry flask were added 200 ml anhydrous ether (dried by distillation over lithium aluminum hydride) and 10.0 g (0.262 mol) lithium aluminum hydride (J. T. Baker 40 Chemical Co., 95%). This mixture was stirred at room temperature for thirty minutes. To the stirred mixture was then added, dropwise, all 200 m1 of the solution containing l-methyl-exo-Z-norbornyl acetate prepared previously. After addition was complete, the mixture was stirred at room temperature for three hours, followed by gently refluxing for three hours. The reaction was quenched by cautiousaddition of 20 ml distilled water followed by 20 ml 5% sodium hydroxide solution. The resulting mixture was filtered and the white precipitate was washed thoroughly with ether. The combined filtrate and ether washings was then dried over molecular sieve (8-12 mesh, 'Baker Analyzed' Reagent). The dried solution was poured into a 1000 m1 round-bottomed flask and the ether was removed by distillation. A residue weighing 47.1 g was obtained. Analysis of this residue showed the presence of a trace amount of ether, approximately 20% unreacted acetate, a trace of a higher-boiling unknown, and approximately 75% 1-methy1-exo-Z-norbornanol. 53 Preparation of a Chromic Acid Oxidizing Reagent To a 500 ml volumetric flask were added 100g (0.336 mol) sodium dichromate dihydrate (Matheson Coleman and Bell) and 300 ml distilled water. After dissolution of the dichromate and addition of 74 m1 (136 g, 1.39 mol) of 97.7% sulfuric acid ('Baker Analyzed' Reagent) the solution was diluted with water to 500 ml. The resulting red solution was stored at 0°. 41 53 Preparation of l-Methylnorcamphor To a 1000 ml three-necked round-bottomed flask equipped with reflux condenser, mechanical stirrer and dropping funnel was added a solution of approximately 36 g (0.28 mol) l-methyl~exo-2-norbornanol in 175 ml anhydorus ether. This solution was stirred for 20 minutes while the flask was immersed in an ice-water bath. To the cold solution was then added a total of 344 m1 of the chromic acid_oxidizing reagent. The addition was done dropwise, in 100 ml portions. During the addition of each 100 ml portion the remaining reagent was cooled in an ice bath. After addition was complete, the dark brown/black mixture was stirred for 5 minutes. The mixture was then poured into a separatory funnel and extracted with four 50 ml-portions of ether. The combined ether was washed with seven 40 ml-portions 5% sodium carbonate solution and three 40 ml-portions saturated sodium chloride solution. The ether solution was then dried over molecular sieve (Activated "Linde", Type 5A, MC/B). The dried solution was poured into a flask and the ether was removed by distillation at atmospheric pressure. The residue was fractionally distilled at reduced pressure. The product cut, b.p. 60-63° at 15 mm (reported b.p. 61.5- 63 at 15 mm), weighed 26.2 g (0.211 mol, 75.4% of theoretical). Analysis by VpC showed this mixture to be mainly the ketone, with a trace of alcohol or acetate as an impurity. The nmr 42 spectrum of the product is characterized as follows: broad singlet, T = 7.41 (one proton, bridgehead hydrogen); multiplet, r = 7.91-8.15 (two protons, a-methylene hydrogens); broad multiplet at r = 8.33-8.66 (six protons, a- and B- methylene hydrogens); and sharp singlet, r = 8.90 (three protons, l-methyl hydrogens). 54 Preparation of 1-Methyl-2-norbornanone-3,3-d2 To a large, thick-walled glass ampoule were added 15 g (0.121 mol) 1-methy1-2-norbornanone, 38.4 g (0.182 mol) trifluoroacetic anhydride (J. T. Baker Chem. Co.), and 43 g deuterium oxide (K G K Laboratories). The ampoule was purged with dry nitrogen and sealed with a gas torch. The ampoule and contents were then heated at loo-125° for six days. The ampoule was cooled and opened. The contents were cautiously poured into a solution of 100 g deuterium oxide and 100 g potassium carbonate ('Baker Analyzed' Reagent). This mixture was poured into a separatory funnel and extracted with five 20 ml-portions of pentane. The combined pentane was washed with two 25 ml-portions of saturated sodium chloride solution and dried over molecular sieve. The dried solution was poured into a round-bottomed flask and the pentane was removed by distillation at atmospheric pressure. The residue was distilled at reduced pressure. A total of 13.9 g (0.11 mol, 91.3% of theoretical) of 1-methyl-2-norbornanone-3,3-d2 was obtained. 43 The exchange was then repeated. All 13.9 g of the ketone was used while the amounts of the other reagents remained the same as in the previous reaction. After work-up and removal of the pentane, the final residue was distilled under vacuum. A total of 11.42 g (0.0906 mol, 75% of theoretical) of 1-methy1-2-norbornanone-3,3-d2, b.p. 57-58° (5 mm), was obtained. The nmr spectrum of the product was exactly like that of the undeuderated compound except that the two-proton multiplet at r = 7.91-8.15 was essentially nonexistent. Preparation of 1-Methyl-endo-Z-norbornanol-3,3-d2 All of the glassware used in this reaction was dried overnight in an oven prior to being used for the reaction. To a 250 m1 three-necked round-bottomed flask equipped with a reflux condenser, mechanical stirrer, and dropping funnel were added 1.0 g (0.0264 mol) of lithium aluminum hydride and 60 m1 of anhydrous ether (dried by distillation over lithium aluminum hydride). This mixture was stirred for ten minutes. To the stirred suspension was then added, drOpwise, a solution of 6.0 g (0.0485 mol) of l-methyl-Z- norbornanone-3,3-d2 in 30 m1 of anhydrous ether. After addition was complete, the mixture was heated to a gentle reflux overnight. The flask and contents were then cooled to room temperature and the reaction was quenched by adding 2.0 m1 of distilled water, followed by 2.0 m1 of a 5% solution of 44 sodium hydroxide. After stirring for thirty minutes, the mixture was filtered. The white precipitate was washed thoroughly with ether and the filtrate and combined ether washings were dried over molecular sieve. The dried solution was poured into a round-bottomed flask and the ether was removed on the rotary evaporator. The residue was a white solid weighing 5.85 g (0.0457 mol, 95% of theoretical). Preparation55 of LMethylenenorbornane-7,7-dz To a 10 m1 flask were added 2.0 g (0.0156 mol) of 1- methyl-endo-2-norbornanol-3,3~d2, 0.4 g (.00292 mol) of potassium hydrogen sulfate (freshly fused, Merck and Co.), and a small magnetic stirring bar. The flask was attached to a fractional distillation head and the mixture was heated. The oil bath temperature was maintained at 150-180° for fifteen minutes. The temperature was then raised until the product began to distill over. When all of the olefin was over, the unreacted alcohol began to distill. At this time the heat was removed and the flask and contents were cooled. Another 0.4 g-portion of potassium hydrogen sulfate was added and the heat was applied again. The mixture was heated until all of the olefin was distilled over. This reaction was repeated with another 2.0 g sample of the alcohol. The olefin from this reaction was combined with that from the previous reaction and the mixture was dried over anhydrous magnesium sulfate. After drying and 45 filtering, a total of 2.0 g of product mixture was isolated. Analysis by VpC showed the presence of three components. One of the components, approximately 30% of the total, was unreacted 1-methyl-endo-2-norbornanol-3,3-d2. The other two components were 2-methy1enenorbornane-7,7-d2 (approximately 60% of total), and Z-methyl-Z-norbornene- 7,7-d2 (approximately 5-10% of total). The nmr spectrum of the Z-methylenenorbornane was characterized as follows: broad singlets at r 5.16 and 5.46 (one proton each, vinyl hydrogens); singlets at r 7.3 and 7.62 (one proton each, bridgehead hydrogens); multiplet at r 7.8-8.1 (two protons, a-methylene hydrogens); complex multiplet at T 8.3-9.23 (four pootons, B- and y-methylene hydrogens). This data is in good agreement with the reported64 values for this compound. The two bridge hydrogens would ordinarily appear as part of the multiplet at T 8.3-9.23. The vinyl hydrogen for the 2-methy1-2-norbornene-7,7-d2 appeared as a small singlet at r 4.5 and the methyl group appeared as a sharp doublet at T 8.28. Preparation56 of Norcamphor-7,7-d2 To a 1000 m1 round—bottomed flask equipped with a magnetic stirring bar and a reflux condenser were added 3.5 g of a mixture of Z-methylenenorbornane, Z-methyl-Z- norbornene, and l-methyl-endo-2-norbornanol, 25.2 g (0.11 mol) of potassium periodate (Reagent, A.C.S., Crystals, 46 Matheson Coleman 5 Bell), 0.98 g (0.0062 mol) of potassium permanganate ('Baker Analyzed' Reagent), 36.0 g (0.361 mol) of potassium carbonate, and 350 m1 of distilled water. This mixture was then stirred vigorously at room temperature for twenty-two hours. After stirring, the mixture was poured into a separatory funnel and extracted with 6 x 40 m1 of ether. The combined ether was washed with water and dried over anhydrous magnesium sulfate. The dried solution was poured into a flask and concentrated by distilling the ether at atmospheric pressure. Analysis of the residue by Vpc showed the presence of ether, unreacted olefin, l-methyl-endo-2-norbornanol-7,7-d2, and norcamphor-7,7-dz. This mixture was poured into a sublimator and the crude product was collected by sublimation. A total of 0.72 g of material was collected. This product still contained trace amounts of both unreacted olefin and l-methyl-endo-Z-norbornanol. Preparation of Ethanolic Sodium Bisulfite Solution57 A 40% solution of sodium bisulfite in water was prepared by dissolving 20 g of sodium bisulfite (Fisher Scientific Co.) in 30 ml of distilled water. To this solution was added 9 ml of absolute ethanol and the mixture was stirred. The mixture was then filtered to remove any sodium bisulfite which had precipitated upon addition of the ethanol. The clear filtrate 47 will be used to form the bisulfite addition product of norcamphor-7,7-dz. Preparation of Sodium endo-Z-Hydroxy-exo-2-Norbornyl- 7,7—d Sulfite 2 To a small Erlynmeyer flask were added 1.72 g of a mixture which contained mainly norcamphor-7,7-d2, along with some 1-methyl-endo-2-norbornanol-3,3-dz and 2- methylene-norbornane-7,7-d2 and 6 m1 of the ethanolic sodium bisulfite solution. The flask was stoppered and the mixture was shaken vigorously until a thick white paste formed. This mixture was then filtered and the residue was washed thoroughly with pentane. The residue was then dried under vacuum. The dried product was a snow-white solid which weighed 1.95 g (.0091 mol). Hydrolysis of Sodium endo-Z-Hydroxy-exo-Z-norbornyl- 7,7-d Sulfite 2 To a small 25 m1 Erlynmeyer flask were added 1.95 g of sodium endo-2-hydroxy-exo—Z-norbornyl sulfite and 10 m1 of a 10% aqueous sodium carbonate solution. The mixture was stirred until all of the material had dissolved, then the flask was placed on a hot plate. The solution was heated gently until two distinct layers had formed and the odor of norcamphor was very noticeable. This mixture was then poured into a separatory funnel and extracted with 4 x 10 m1 of pentane. The pentane extract was dried over anhydrous 48 magnesium sulfate, then concentrated by distilling the pentane. The concentrated solution was poured into a sublimator and the remainder of the pentane was removed under vacuum. The product was then collected by sublimation. A total of 0.55 g of norcamphor-7,7-d2, m.p. 91-91.5° (lit. 90-91°) was collected. The nmr spectrum is characterized as follows: broad singlet centered at r 7.41 (two protons, bridgehead hydrogens); complex multiplet at r 7.9-8.66 (six protons, a- and B-methylene hydrogens). Preparation of endo-LNorbornanol-7,7-d2 A 25 m1 three-necked flask was equipped with a reflux condenser, dropping funnel, and magnetic stirring bar. The entire apparatus was flamed dry, then purged with dry nitrogen gas. To the flask were then added 0.07 g (0.0018 mol) of lithium aluminum hydride (J. T. Baker Co., 95%) and 7 m1 of anhydrous ether (dried by distillation over lithium aluminym hydride). This mixture was stirred for fifteen minutes. To the stirred mixture was added, dropwise, a solution of 0.4 g (0.00358 mol) of norcamphor—7,7-d2 in 5 m1 of anhydrous ether. After addition was complete, the mixture was refluxed gently for two hours. The flask and contents were cooled to room temperature and the reaction was quenched by the cautious addition of 0.14 ml of distilled water, followed by 0.14 ml of a 5% aqueous sodium hydroxide solution. The mixture was filtered 49 and the residue was washed thoroughly with ether. The combined filtrate and ether washings was then dried over molecular sieve. The dried solution was poured into a round-bottomed flask and the ether was removed by distillation. The concentrated residue was poured onto a watch glass and the remaining ether was removed under vacuum. The residue was a white solid weighing 0.32 g (.00309 mol, 86.4% of the theoretical). Preparation of endo—2-Norbornyl-7,7-d2 Brosylate To a small dry Erlynmeyer flask were added 0.300 g (0.00264 mol) of endo-2-norbornanol-7,7-d and 10 ml of dry 2 pyridine ('Baker Analyzed' Reagent, stored over molecular sieve). The flask was purged with dry nitrogen, stOppered, and placed in an ice-bath to cool. After thirty minutes, 0.82 g (.0032 mol) of p-bromobenzenesulfonyl chloride (Eastman Organic Chemicals, recrystallized from hexane) was added to the cold solution. The flask was left in the ice- bath for an additional hour. It was then transferred to a refrigerator. After sitting in the refrigerator for seven days, the flask was removed and the reaction was quenched by pouring the pyridine solution into a mixture of ether and ice water. This mixture was poured into a separatory funnel and the layers were separated. The aqueous layer was extracted with 5 x 10 m1 of ether. The ether was washed with cold 10% 50 hydrochloric acid until the washings were acidic. The ether solution was then washed with 2 x 15 m1 of saturated sodium bicarbonate solution and 2 x 15 m1 of saturated sodium chlorice solution; it was then dried over anhydrous magnesium sulfate. The dried solution was filtered and the filtrate was poured into a round-bottomed flask. The ether was removed by evaporation on a rotary evaporator. The residue was a white solid weighing 1.087 g. Since the theoretical yield for the reaction was only 0.878 g, the product was obviously still wet. 58 Selective Solvolysis of a Mixture of endo- and exo-Z-Norbornyl-7,7-d2 Brosylate To a 35 ml round-bottomed flask equipped with a reflux condenser and a magnetic stirring bar were added 1.087 g of the crude brosylate mixture and 14.5 g of a 75% aqueous acetone solution. The solution was stirred at 45° for thirty minutes. It was then cooled and neutralized to the phenolphthalien end-point by the drop-wise addition of a 10% sodium hydroxide solution. The neutral solution was poured into a separatory funnel and extracted with 5 x 5 m1 of ether. The ether solution was then washed with distilled water and dried over anhydrous magnesium sulfate. The dried solution was filtered and the filtrate was concentrated by evaporating the ether on the rotary evaporator. Pentane was added to the 51 concentrated solution and the mixture was placed in the refrigerator until a precipitate had formed. After two hours, the mixture was removed from the refrigerator and filtered. The crude residue, dried under vacuum, weighed 0.524 g, m.p. 57-60°. This material was recrystallized twice from an ether-pentane mixture. The final product weighed 0.2933 g, m;p. 61.5-62.0° (literature value for the hydrogen analog is m.p. 62-63°). III. Kinetics Preparation of Solvents Water. -- House-supplied distilled water was refluxed for four hours and then fractionally distilled through a 30 cm Vigreaux column. Ethanol. - Absolute ethanol was prepared by the 59 A 1000 m1 three-necked procedure described by Vogel. round—bottomed flask was equipped with a mechanical stirrer, a 30 cm Vigreaux column and distillation head, and a ground glass stOpper. The entire apparatus was thoroughly flamed to ensure dryness; a stream of dry nitrogen gas was then passed through the system for fifteen minutes. To the dry flask were added 1000 m1 of commercial absolute ethanol and 7.0 g of clean, dry sodium metal. After all of the sodium had reacted, 25 g of pure diethyl succinate (Eastman Organic Chem.) was added and the solution was refluxed for five hours. The 52 solution was then fractionally distilled and the first 25 m1 fraction was discarded. Mixed Solvent. -- The solvent used for the kinetic runs was 60/40 ethanol-water (V/V). It was prepared by measuring the corresponding volumes with a 50 m1 volumetric pipette. Kinetic Apparatus A Unicam SP .800 ultraviolet spectrOphotometer equipped with an SP .850 scale eXpansion accessory, an SP .825 program controller, an SP .874 constant temperature cell housing and an SP .820 constant wavelength scanner was used for all absorbance measurements. An external recorder (Sargent Recorder, Model SR) was attached to the spectrophotometer and all absorbance readings were taken from the expanded scale of this recorder. The SP .874 constant temperature cell housing was attached by means of insulated rubber tubing to an external constant temperature bath. The cell compartment temperature was maintained by pumping a continuous flow of the bath medium through the cell compartment. All kinetic runs were followed continuously by using the SP .825 automatic program controller. Constant Temperature Bath A 14.5" x 14.5" x 13.25" wooden box was assembled from 1/2" plywood sheets. The bottom of the box on the inside was covered with a sheet of polystyrene, one inch thick, which S3 in turn was covered with 1/4" of vermiculite packing. On this base was centered a 16 liter glass jar. The area between the glass jar and the sides of the box (at least one inch on all sides) was packed with more vermiculite up to within one inch of the top of the jar. On top of this, another sheet of polystyrene, out to fit tightly around the outside of the jar, was placed so that it was flush with the t0p of the jar. The top of the bath was covered with a sheet of 1/4" asbestos. The bath was equipped with a mechanical stirrer (Talboys Instrument Corp., Model No. 104, 1/18 H.P.), a Teel pump (Dayton Electrical Mfg. Co., Model No. 1P731), a Beckmann differential thermometer, a 125W heating blade (Cenco), an electronic relay (Precision Scientific Co., Model No. 62690), a mercury micro-set thermoregulator (-35° to 500°F, Precision Scientific Co., Model No. 62541), and a 500W heating coil. The voltage supplied to both the 500W heating coil and 125W heating blade was controlled by means of a Variac. Water was used as the bath medium for the temperature range 50° to 70° while HTF-100 Ucon Fluid (a mixture of polyalkylene glycols available from Matheson Scientific) was used for temperatures > 70°. Excellent temperature control was achieved by heating the bath medium to within one degree of the desired temperature by means of the 500W continuous heating coil. The intermittant heating knife, which was connected to the electronic relay and the thermoregulator, was then used to achieve the correct 54 temperature. The voltage supplied to the intermittant heater was adjusted so that heating and non-heating periods were about the same. Using this method, the bath temperature could be controlled to within i.005° and the cell compartment temperature was constant to t.015°. Temperature Measurement The absolute temperatures of the bath and cell compart- ment were measured by using a Hewlett Packard quartz thermometer (Model No. 2801A). Time Measurements Time measurements for the kinetic runs were made by an electrical digital timer (Precision Scientific, Model No. 69235) which was accurate to 0.01 minute. Rate Determinations The kinetic runs on the hydrogen and deuterium compounds were made consecutively in a random order. The same quartz uv cell and cell compartment were used for each urn. Prior to each run the external recorder was adjusted so that a full scale reading on the Sargent recorder corresponded to an absorbance reading of 0.5 absorbance units on the Unicam recorder. The SP .820 constant wavelength control was set for 265 mu. The uv cell was then filled with 2.6 m1 of solvent and allowed to thermally equilibrate in the cell compartment for 1.5 hours. The cell was then removed and 8 ul of a 0.278M solution of brosylate in absolute ethanol 55 was injected. The cell was capped, shaken for 10 seconds, and placed back in the cell compartment; readings began immediately. The SP .825 program controller was set to allow 30-50 absorbance readings to be taken over the first 2.5—3.5 half-lives of the reaction. The infinity reading was taken after approximately 10 half-lives. The total absorbance range was 0.5-0.1 absorbance units. RESULTS 1. Synthesis The sequence of reactions used to synthesize endo-2- norbornyl-7,7-d2 brosylate is illustrated in Scheme III. Scheme III 1 MeM 1 H2504 g; 1::fE:1LOA° Hm 4c1 Me HOAC o XIV 0H xv 1) 1.1111114 2) NaOH, H O 2 <(c1:3 c0) 20 1) Na 2Cr “71::1E:1L%0 XVIII XVII 1) LiAlH 4 2) NaOH, H20 D D D D KMnO4 D 4;; H H XXI . XIX OH XX H l) L1All|4 2) NaOll , HZO D D D D ,/ BsCl H \‘ PYT- H XXIII OBS XXII )u 56 S7 The reactions leading to XVI were carried out exactly as 52 described by Berson, et a1. Oxidation of XVI according to S3 Brown's procedure gave XVII in 75% yield. The ketone was deuterated by heating with trifluoroacetic anhydride and 54 for five days at 125°. deuterium oxide in a sealed tube After two exchanges, nmr analysis indicated that deuterium incorporation was >95% complete. Several different methods were tested for conversion of the endo-alcohol, XIX, to the exocyclic olefin, XX. Brown has shown60 that exocyclic olefins can be obtained from 1- chloro-l-methylcycloalkanes by elimination with a bulky base such as potassium triehtylmethoxide. Therefore, our initial efforts involved converting XIX to the tertiary chloride exo-Z-chloro-endo-Z-methylnorbornane. Only unreacted starting material was recovered from a mixture of XIX with anhydrous63 hydrochloric acid. The tertiary chloride was produced by reacting exo-2-hydroxy—endo-Z-methylnorbornane with anhydrous HCl; however, conversion of XIX to the tertiary alcohol via tosylation and subsequent solvolysis62 resulted in yields too low to be practical. The most efficient method for preparing the tertiary chloride proved to be the reaction of XIX with thionyl chloride61 in tetrahydrofuran. By this method, a mixture of secondary and tertiary chlorides was produced, approximately 70% of which was the tertiary chloride. Unfortunately, when this mixture was reacted with potassium triethylmethoxide, the yield of olefin was less than 50%. We considered this unacceptably 10w. 58 KriegerSS has prepared 2-methy1enenorbornane by dehydration of endo-2-hydroxy-exo-Z-methylnorbornane with several different acids. More significantly, he has also prepared this olefin by dehydration72 of 1-methyl-exo-2- hydroxynorbornane with these same acids. In this latter dehydration, the molecule undergoes the same transformations necessary for conversion of XIX to XX, i.e., rearrangement and elimination. Of all the acids tested by Krieger, potassium bisulfate seemed to work the best. The dehydration of XIX with this acid salt is described in the Experimental section. By this method, XX was obtained in a yield of m60%. Two different methods were also tested for the oxidation of XX to norcamphor-7,7-dz, XXI. Although a fairly good yield of ketone was obtained by ozonolysis6S of XX, the presence of a trace of an unknown byproduct unfortunately prevented the isolation of pure XXI from the product mixture of this reaction. To circumvent this problem, the permanganate-periodate method was tried. Again, norcamphor was produced in reasonable yield. The ketone was isolated via its bisulfite addition complex and purified by sublimation. The nmr spectrum of XXI was analyzed to affirm the presence of the deuterium atoms at carbon seven. In deuterochloroform, the two bridgehead protons appeared as a broad singlet centered at T 7.4, while the a-methylene protons at C3 appeared as a two-proton multiplet at r 7.9- 8.15. For norcamphor, the B- and y-methylene hydrogens 59 appear as a six-proton multiplet at r 8.15-8.67. The spectrum of XXI did exhibit this multiplet at r 8.15-8.67; however, it corresponded to only four protons. Although it is impossible to specify from these data whether the deuteriums are at C5’ C6’ or C7, it is clear that two deuterium atoms are present and that they are either 8 or y to the carbonyl. When this fact is combined with a knowledge of the reactions leading to XXI, it is quite certain that the deuterium atoms must be at C7. The last two steps of the reaction sequence are standard reactions which were carried out in the usual way. Reduction of the ketone with lithium aluminum hydride yields primarily endo-alcohol, with some (%5%) exo-alcohol also formed. Since these isomers were not separated at this stage, the final brosylate, XXIII, also contained some axe-brosylate, which was removed by selective solvolysis66 of the mixture. The resulting endo—Z-norbornyl-7,7-d2 brosylate was recrystallized from an ether-pentane mixture to a constant melting point of 61.5-62.0°. 67-71 which was also considered for An alternative route the synthesis of XXIII is shown in Scheme IV . This route has the advantage that it is not necessary for the molecule to undergo a rearrangement in order to get the deuterium atoms at C7. However, it has the disadvantage that the oxidation of XXV to XXVI and the solvolysis of XXVII, with 60 Scheme IV H Ac 0“ 1) LiA1H4 A1(t-Bu0)3 is. -——-—-—-—€> / 2) NaOH,H26 0, o=©=o XXIV xxv XXVI 1) L1A184 2) NaOH, H 0 H 2SO4 2 D 1) NaBH4 l) TsCl, PyrD BF3 -Et 20 .z’ [2) NaBD4 2) N32 Cr HO éDiglyme, D 20 XXI XXVIII XXVII subsequent trapping of the carbonium ion intermediate with NaBD4, to give XXVIII are both very low yield reactions. We attempted to carry out the oxidation of XXV with CrO3 in methylene chloride but no perceptible increase in yield was obtained. Consequently, we decided to abandon this scheme in favor of that shown in Scheme III. 11. Kinetics The solvolysis of endo-Z-norbornyl brosylate aqueous ethanol is characterized by the formation bromobenzene sulfonic acid. Both titrametric and photometric methods have been used in the past to in 60% of p- spectro- follow 61 the kinetics of this reaction. Although the spectre- photometric method is usually considered the more accurate, there is one disadvantage associated with it, i.e., the brosylate ester and the aryl sulfonic acid both absorb in the same region of the uv spectrum. Consequently, the use of this method often entails extracting aliquots to separate the two components, and then measuring either the decrease in absorbance of the ester or the increase in absorbance of the acid. Fortunately, a method has been developed74 by which the decrease in absorbance of the ester can be followed without recourse to extracting aliquots. This method is based on the fact that the molar absorptivities for the two species are different. The rate equation may be developed as follows: 6 = molar absorptivity of ester at 265 nm e e = molar absorptivity of acid at 265 nm absorbance of sample at time t A A = absorbance of sample at t = O C = concentration of ester at t = O C = concentration of ester at time t C = concentration of acid at time t The absorbance of the sample at time t is given by At = Ceee I Caea (1) or, Since Ce = Co-Ca At = (Co-Ca) 6e + C853 (2) thus A -C c _ t o a Ce _ e -e (3) e a 62 The rate equation for a first-order reaction is C o kt = 11'). T— (4) e By substituting for Ce we have Co (Se-ea) At-Coea kt = 1n (5) S1nce Cote = A0 and Cosa = A the final rate equation becomes kt = 1n (Ac-Am) -ln (At-Am) (6) As is evident from equation (5), this method is applicable 75 only if 8e f ca. Swartz has shown that, in 80% (v/v) aqueous ethanol at 265 nm, 6e = 640, while the corresponding value for Ea is 330. Therefore, by using this equation it should be possible to determine the rate constant for the solvolysis of endo-Z-norbornyl brosylate simply by following the change in absorbance of the solution at 265 nm over a specified time period. Before beginning the actual kinetic study, we prepared a Beer's Law plot for the brosylate ester in 60% (v/v) aqueous ethanol. This plot is shown in Figure III. Although there is some deviation at the higher concentrations, the absorbance is linear over the concentration range in which 4 we are interested (2-10 x 10' M). Absorbance 63 c = 566 ’ / / / O / / / / / / C / O O l 1 J. 1 1 l I l J 1 4 6 8 10 12 14 18 20 22 24 Concentration x 104 Figure 3. Beer's law plot DR IV—29. 64 The actual technique used for the kinetic runs is described in the Experimental section. The individual rate constants were determined by fitting the absorbance data to equation (6) by a least-squares curvefitting computer program, kt = 1n (AC-Am) -1n (At-Am) (6) Kinfit.76 The linearity of the line described by this equation was checked by plotting 1n (At-Am) vs time for a sample kinetic run (Figure 4). The slope of the line in this plot remains constant throughout the kinetic run. The calculated rate constants for both the labeled and unlabeled compounds are shown in Table V. The average rate constant for each compound is the mean average of the individual rate constants determined for that compound at the specified temperature. Any rate constant which deviated from the average by more than four times the mean deviation of the other values was deleted. The deviations of the individual rate constants calculated for a given temperature increased as the temperature increased. This was due to a slightly larger temperature fluctuation inside the cell compartment at the higher temperatures. To compensate for this a minimum of four runs were made for each compound at each temperature. Therefore, the average rate constants at the higher temperatures were determined from 4-5 individual rate determinations, as opposed to only 3-4 at the lower temperatures. 65 1 1 .1 .1 I 1 l 1 1, 1 1 1 2 3 4 5 6 7 8 9 10 11 time (minutes) Figure 4. Plot of ln (At-Am) vs time for solvolysis of endo-Z-norbornyl brosylate. 66 Table V. Rate Constants for Solvolysis of endo-Z-Norbornyl- 7,7-d2 Brosylate in 60% (v/v) Aqueous Ethanol Substrate Temperature K x 104 sec'1 kavg x 104 sec-1 d0 53.533 2.958 i .021 d0 2.949 1 .023 2.944 t .023 do 2.926 t .021 d2 2.986 t .018 d2 2.947 t .021 2.967 t .024 d2 2.967 t .034 dO 58.680 5.359 t .046 do 5.349 t .040 5.355 1 .041 do 5.356 t .037 d2 5.295 f .044 d2 5.195 5 .040 _5.281 1 .042 (I2 5.266 1 .045 d2 5.369 t .039 (10 64.501 10.087 1 .065 (10 9.941 t .071 10.060 1 .062 d0 10.044 1 .059 do 10.167 1 .054 d2 9.895 1 .053 d2 9.998 1 .068 dz 9.745 t .071 9.908 1 .066 d2 10.163 i .068 d2 9.738 t .071 do 70.791 18.542 1 .154 do 18.425 1 .135 18.354 i .148 do 18.039 1 .154 (10 18.408 t .148 d2 17.765 t .161 d2 18.293 1 .122 d2 18.344 1 .095 18.106 1 .141 d2 17.815 1 .199 d2 18.312 1 .280 67 The isotope effects determined at the various temperatures are shown in Table VI. These were calculated Table VI. Isotope Effects for Solvolysis of endo-2- Norbornyl-7,7-d2 Brosylate at Different Temperatures Temperature, °C IsotOpe Effect 53.533 0.992 1 .011 58.680 1.014 t .011 64.501 1.015 t .009 70.791 1.015 t .011 from the average rate constants for the two compounds at each temperature . The error associated with each isotope effect was determined from the following equation: 0 = kH/kD (UHZ/kH2 + ODZ/kD2)1/2 where RH and kD are the average rate constants for the unlabeled and labeled conpounds, and OH and CD are the errors for the respective constants. Although there is a substantial change in the isotope effect as the temperature is raised from S3.S33° to 58.680°, the effect remains fairly constant over the temperature range 58.680-70.791°. 68 The thermodynamic activation parameters were calculated for the solvolysis of both labeled and unlabeled compounds by fitting the average rate constants at each temperature to the following equation: # Il- ln (kavg/T) = 1n (K/h) - AH /RT + AS /R where kav is the average rage constant, T is the absolute 8 temperature, R is the ideal gas constant, K is Boltzman's constant, and h is Plank's constant. The computations were made by a linear least-squares computer program, ACTIV.77 Two different trends are indicated by the data in Table VI: (1) an increase in the isotOpe effect as the temperature increases, and (2) a constant isotope effect over the temperature range 58.680-70.791°. Therefore, we calculated the activation parameters by two different methods. In Method I all four temperatures and their corresponding rate constants were included in the calculations, whereas for Method II, only the three higher temperatures were used, along . . + 4 4: With their rate constants, to calculate AHH, AHD, ASH, 3. The results are shown in Table VII. The AH* values for and AS each compound are essentially the same regardless of which method is used for the calculation. However, the A8* value differs considerably for the two methods, especially for the unlabeled compound. The final calculation involved disecting the observed isotOpe effects into the contribution due to enthalpy 69 differences (AAHt) and that due to entropy differences (005*). This was done by fitting the temperature-isotope Table VII. Activation Parameters for Solvolysis of endo-Z-Norbornyl Brosylates in 60% (v/v) Aqueous Ethanol AH; kcal as; kcal AH; kcal ASE kcal Method moIe moIe-deg moIe moIe-deg I 23.04 t .68 —4.3 t 2.0 22.79 t .50 -S.l i 1.5 II 22.39 H- .69 -6.2 i 2.0 22.40 H- .65 -6.2 1 1.9 effect data to the following equation: AH+ - AH* 05* - 08* _ H D H D 111 (kH/kD) - 'fif + R The calculations were made by the least-squares computer program, Kinfit. As in the calculations of AH* and AS*, both Method I and Method II were used to calculate AAH¢ and 008*. The results of these calculations appear in Table VIII. In these calculations, the differences between the two methods become very apparent. By Method I, the AAS* term is quite large and in favor of the unlabeled compound, while the AAHt is also large and in favor of the deuteratrd compound. By Method II, the AAS* term is essentially nil, 70 Table VIII. Enthalpy and Entropy Differences Between Isotopically Related endo-Z-Norbornyl Brosylates Method AAH* kcal/mole AAS* kcal/mole-deg I 251.3 f 156 .768 t .464 II - 0.292 t 22.7 .027 i .067 4. while the AAH shows a slight preference for the unlabeled compound. DISCUSSION 1. Temperature Dependence of the Isotope Effect Secondary deuterium isotope effects are usually believed to arise from enthalpy differences (AAH*) between the deuterium and hydrogen compounds. However, recent studies44 have shown that, for some systems, the entropy difference (AASt) is the major contributor to the observed effect. If the secondary y-deuterium isotope effect is to be an effective probe for C-C bond hyperconjugation in the solvolysis of endo-Z-norbornyl brosylate, it is first necessary to establish the origin of the observed effect. It is for this reason that we have studied the temperature dependence of the isotope effect obtained in the solvolysis of endo-2-norbornyl-7,7-d2 brosylate. As can be seen from Table VIII, the conclusion reached from this study depends considerably on which of the data we use for calculating AAH* and AAS*. Thus, if we use the isotope effects observed at all four temperatures, it appears that AAS* is a major contributor to the observed effect. Conversely, if only those isotOpe effects observed at the three higher temperatures are considered, AAS* appears to be essentially nil, and the observed isotope effect is due to a small difference in AAH*. 71 72 The calculations determined from all four temperatures not , 4: a: only show a Sizeable AAS*, they also show AHD < AHH by 251 cal/mole. According to this calculation, the isotope effect at 70°, based strictly on AAH*, is kH/k = 0.70. D If correct, this would indeed be a rather startling result. However, there appears to be no way of explaining this result within the confines of the present theories of secondary deuterium isotope effects. On the other hand, the results obtained by using only the data for the three higher temperatures seems much more reasonable. Consequently, we have concluded that the secondary y-deuterium isotope effect for the solvolysis of endo-2-norborny1—7,7-d2 brosylate shows no anomalous temperature dependence, and that the observed isotOpe effect arises from enthalpy differences between the deuterated and undeuterated compounds. II. Interpretation of Isotope Effect The observed secondary y-deuterium isotope effect on the solvolysis of endo-2-norbornyl-7,7-d2 brosylate is 1.015. Although this would seem to support the existence of Cl-C7 bond hyperconjugation in the transition state for this reaction, we should also examine the system for other possible factors which might cause such an isotope effect. As mentioned earlier, nonclassical ion formation (i.e. bridging) is not believed to exist during the solvolysis of endo-Z-norbornyl substrates. However, one might question 73 whether bridging is any less likely to occur than is Cl-C7 bond hyperconjugation. This question can perhaps best be answered by comparing the effects of bridging and hyper- conjugation on the remainder of the molecule. Thus, both bridging and hyperconjugation involve a partial delocalization of the electrons in the Cl-C7 0 bond. Bridging also requires that a spatial reorganization of the carbon skeleton coincide with this electron delocalization. The transition state leading to the bridged ion is shown in A. If one examines this structure closely, it can be seen that the new partial 0 bond formed between C and C2 is part of a four-membered 7 ring including carbons C C C4, and C 2’ 3’ 7‘ spatial reorganization required for bridging would be Thus, the Opposed by the formation of unfavorable interactions (e.g. bond angle strain) inherent in four-membered rings. However, hyperconjugation does not require, and indeed does not allow, any skeletal reorganizations. Therefore, Cl-C7 bond hyper- conjugation would seem to be more feasible than bridging in the solvolysis of endo-Z-norbornyl substrates. 74 Since the positive charge is three bonds removed from the deuterium atoms, inductive effects should not be significant in our system. However, there is the possibility of nonbonded interactions involving the hydrogens (deuteriums) at carbon seven during the solvolysis of endo-Z-norbornyl 44c found that a substantial non-bonded brosylate. Smith interaction does occur between the exo-6-CH3 (d3) and the C7-H syn to it during the solvolysis of 2,6,6-trimethyl-endo- Z—norbornyl p-nitrobenzoate. Although it does not seem likely, a similar interaction may occur between the exo-6-H and the syn-7-H (D) in the solvolysis of our compound. However, if such an interaction were significant, the observed isotope effect would be expected to be inverse rather than normal. Similarly, one might envision a slight non-bonded interaction in the ground state between the exo-Z-H and the syn-7-H (D) in endo-Z-norbornyl brosylate. Relief of this interaction in the transition state may result in a normal isotope effect. The distance between these two hydrogens 26 has been calculated to be ca. 2.68 3. Since the sum of the van der Waals radii of two hydrogens is only 2.4 R, this interaction should not be significant. Let us now consider C1-C7 bond hyperconjugation. Although the concept of hyperconjugation is usually discussed in terms of C-H 0 bonds, the eXperimental work of Jensen45’46 49 50 and Traylor, and the theoretical calculations of Pople 75 and Danen51 indicate that C-C bond hyperconjugation is also effective for stabilizing a positive charge. In this regard, 78 that "there are no obvious theoretical Mulliken has stated reasons for expecting radical differences in stabilization energy or in magnitude of electron release between C-C and C-H bond hyperconjugation." Therefore, there are no apparent theoretical reasons for dismissing Cl-C7 bond hyperconjugation in the solvolysis of endo-Z-norbornyl brosylate. We have pointed out previously that there is no atomic movement involved in hyperconjugation. Consequently, this method of stabilizing an incipient positive charge is not restricted by any adverse non-bonded interactions. Apparently, the only factor which is unfavorable for Cl-C7 bond hyperconjugation in the solvolysis of endo-2— norbornyl brosylate is the stereochemical alignment of the Cl-C7 bond with respect to the leaving group. As described earlier, this bond is displaced by several degrees from the ideal position, i.e., trans-periplanar to the leaving group. However, Shiner34 has shown that some C-H bond hyper- conjugation occurs even when the dihedral angle between the C-H bond and the leaving group is 60°. In view of this, one would expect the extent of Cl-C7 bond hyperconjugation to be small, but not necessarily nonexistent. Therefore, on the basis of the previous discussion, Cl-C7 bond hyper- conjugation would seem to offer a very reasonable explanation for the secondary y-deuterium isotope effect of 1.015 observed for the solvolysis of endo-Z-norbornyl-7,7-d2 brosylate. 76 Interpretations of isotope effect data are usually supported by citing previous studies on similar systems. Unfortunately, there are no previous studies in which secondary isotope effects have been interpreted in terms of C-C bond hyperconjugation. However, there have been studies in which C-C bond hyperconjugation could have been used to interpret the results. The secondary y-deuterium effect on the solvolysis of exo-Z-norbornyl-6-d brosylate38’39 was discussed in the Introduction and that discussion will not be reiterated here. Another y—deuterium isotope effect which could have been explained in terms of C-C bond hyperconjugation is the isotOpe effect on the solvolysis of cyclobutyl mesylate, XXIX. The y-effect found for the solvolysis of XXIX was D >K<::>FOMS D XXIX 1.08.79 This effect was offered in support of bridging in the transition state for the reaction. However, the puckering of the cyclobutane ring to relieve eclipsing interactions brings the B-C-C bond into reasonable position for C-C bond hyperconjugation. Consequently, both bridging and hyper- conjugation would give a reasonable explanation for the effect observed in this system. 77 It would be interesting to test the C-C bond hyper— conjugation concept on another system in which bridging is not expected to be important. One possibility might be the 80 bicyclo[2.l.l]hexyl system. Wiberg has shown that XXX 6 solvolyzes 3 x 10 faster than does XXXI. He attributes Ts xxx 0T5 XXXI this to bridging of the Cl-C6 bond in the transition state of XXX. However, he states that the acetolysis of XXXI is still 105 faster than acetolysis of 7-norbornyl tosylate. For this reason, he suggests the possibility of some bridging in the solvolysis of XXXI. If one considers the Newman projections for XXXI, it appears that the Cl-CS bond is in C5 H excellent position for C-C bond hyperconjugation. On the other hand, bridging in this sytem is complicated by the 78 same type of interactions described for bridging in endo-2- norbornyl brosylate. Thus, determination of the secondary y-deuterium isotope effect for XXXII might lend additional D ,‘I'I"// 0T5 XXXII support for the C-C bond hyperconjugation concept. CONCLUSION The secondary y-deuterium isotope effect on the solvolysis of endo-2-norbornyl-7,7-d2 brosylate appears to support the existence of Cl-C7 bond hyperconjugation in the transition state for this reaction. Application of this concept to other systems has also been discussed. It appears that much of the previous isotope effect data, which were interpreted in terms of nonclassical ion formation (i.e. bridging), can be eXplained equally well by carbon- carbon bond hyperconjugation. Further support for this type of carbonium ion stabilization must come from systems in which the possibility of bridging is minimal., We have suggested one such system. 79 BIBLIOGRAPHY 11. 12. 13. 14. BIBLIOGRAPHY T. P. Nevell, E. de Salas, and C. L. Wilson, J. Chem. 500., 1232, 1188. (a) P. D. Bartlett, "Nonclassical Ions," Frontiers in Chemistry Series, W. A. Benjamin, Inc., New York, New York. (b) C. Dann Sargent, Quart. Rev. (London), 20, 301. S. Winstein and D. Trifan, J. Amer. Chem. 500., 71, 2953 (1949). S. Winstein and D. Trifan, ibid., 74, 1147 (1952). S. Winstein and D. Trifan, ibid., 74, 1154 (1952). S. Winstein, E. Clippinger, R. Howe, and E. Vogelfanger, ibid., 87, 376 (1965). C. S. Foote, ibid., 86, 1853 (1964). P. von R. Schleyer, ibid., 86, 1854 (1964). P. von R. Schleyer, ibid., 86, 1856 (1964). P. von R. Schleyer, M. M. Donaldson, and W. E. Watts, ibid., 87, 375 (1965). G. KlOpman, ibid., 21, 89 (1969). M. Simonetta and S. Winstein, ibid., 16, 18 (1954). P. von Schleyer, et al., ibid., 86, 5679 (1964). G. A. Olah, et al., ibid., 92, 4627 (1970). 80 15. 16. 17. 18. 19. 20. 21. 81 H. C. Brown, Chem. Soc. (London), Spec. PubZ., 16, 140 (1962). H. C. Brown, Chem. Brit., 1966, 199. (a) W. Alexander Van Hook in "Isotope Effects in Chemical Reactions," ACS Monograph, Clair J. Collins and Newell S. Bowman, Ed., Van Nostrand Reinhold Company, New York, New York, and references cited therein. (b) E. K. Thornton and E. R. Thornton, ibid., and references cited therein. (a) A. Streitwieser, R. H. Jagow, R. C. Fahey, and S. Suzuki, J. Amer. Chem. Soc., 80, 2326 (1958). (b) M. Wolfsberg and M. J. Stern, Pure Appl. Chem., 8, 225 (1964). (c) M. Wolfsberg and M.JJ. Stern, Pure Appl. Chem., 8, 325 (1964). E. A. Halevi, "Secondary Isotope Effects," Progress in Physical Organic Chemistry, Vol. 1, S. G. Cohen, A. Streitwieser, and R. W. Taft, eds., John Wiley and Sons, New York, 1963. V. J. Shiner, Jr., "Isotope Effects in Chemical Reactions", ACS Monograph, Clair J. Collins and Newell S. Bowman, eds., Van Nostrand Reinhold Company, New York, 1970 and references cited therein. L. S. Bartel], J. Amer. Chem. Soc., 83, 3567 (1961). 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 82 G. J. Karabatsos, G. C. Sonnichsen, C. G. Papaioannou, S. E. Scheppele, and R. L. Shone, ibid., 89, 463 (1967). V. J. Shiner, Jr., "Isotope Effects in Chemical Reactions," ACS Monograph, Clair J. Collins and Newell S. Bowman, eds., Van Nostrand Reinhold Company, New York, 1970, pp. 151-155. D. E. Sunko and S. Borcic, ibid., pp. 160-209. D. E. Sunko and S. Borcic, ibid., pp. 201-204. C. C. Lee and Ed. W. C. Wong, J. Amer. Chem. 500., 86, 2752 (1964). B. L. Murr and J. A. Conkling, ibid., 92, 3462 (1970). D. E. Sunko and S. Borcic, "Isotope Effects in Chemical Reactions," ACS Monograph, Clair S. Collins and Newell S. Bowman, Eds., Van Nostrand Reinhold Company, New York, 1970, pp. 201-203. J. M. Jerkunica, S. Borcic, and D. E. Sunko, Chem. Commun., 1961, 1302. B. L. Murr and J. A. Conkling, J. Amer. Chem. 500., 92, 3464 (1970). J. P. Schaefer and D. S. Weinberg, Tet. Lett., 29, 2491 (1965). D. E. Sunko and S. Borcic, "Isotope Effects in Chemical Reactions," ACS Monograph, Clair J. Collins and Newell S. Bowman, eds., Van Nostrand Reinhold Company, New York, 1970, pp. 204-207. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 83 J. P. Schaefer, J. P. Foster, M. J. Dagani, and L. M. Honig, J. Amer. Chem. Soc., 99, 4497 (1968). V. J. Shiner, Jr. and J. S. Humphrey, Jr., ibid., 89, 2416 (1963). Stuart B. Scheppele, Chem. Commun., 8998, 592. V. J. Shiner, Jr., R. D. Fisher, and W. Dowo, J. Amer. Chem. 500., 99, 7748 (1969). V. J. Shiner, Jr., ibid., 88, 240 (1961). J. M. Jerkunica, S. Borcic, and D. E. Sunko, ibid., 89, 1732 (1967). B. L. Murr, A. Nickon, T. D. Swartz, and N. H. Werstiuk, ibid., 89, 1730 (1967). Harlan L. Goering and Kresimir Humski, ibid., 98, 4594 (1969). S. Borcic, V. Belanic-Lipovac, and D. E. Sunko, Croat. Chem. Acta, 89, 3S (1961). N. H. Werstiuk, et aZ., Can. J. Chem., 89, 618 (1972). D. E. Sunko and S. Borcic, "Isotope Effects in Chemical Reactions," ACS Monograph, Clair J. Collins and Newell S. Bowman, eds., Van Nostrand Reinhold Company, New York, 1970, p. 205. (a) T. A. Evans, Ph.D. Thesis, Michigan State University, 1968. (b) A. A. Effio Leon, Ph.D. Thesis, Michigan State University, 1971. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 84 (c) V. F. Smith, Jr., Ph.D. Thesis, Michigan State University, 1972. (d) R. L. Turner, Ph.D. Thesis, Michigan State University, 1972. (e) U. G. Kang, Ph.D. Thesis, Michigan State University, 1972. F. R. Jensen and B. E. Smart, J. Amer. Chem. 500., 99, 5686 (1969). F. R. Jensen and B. E. Smart, ibid., 99, S688 (1969). R. Hoffmann, J. Chem. Phys., 98, 2480 (1964). R. C. Bingham and P. von R. Schleyer, J. Amer. Chem. Soc., 98, 3189 (1971). T. G. Traylor, W. Hanstein, H. J. Berwin, Nye A. Clinton, and R. S. Brown, ibid., 98, S715 (1971); and references cited therein. L. Ransom, J. A. Pople, U. Boss, and P. von R. Schleyer, ibid., 98, 6380 (1970). Wayne C. Danen, ibid., 99, 4835 (1972). Jerome A. Berson, J. S. Walia, A. Remanick, S. Suzuki, P. Reynolds-Warnhoff and D. Willner, ibid., 88, 3986 (1961). H. C. Brown, C. P. Garg, and K.-T. Liu, J. Org. Chem., 36, 387 (1971). J. P. Schaefer, M. J. Dagani, and D. S. Weinberg, J. Amer. Chem. Soc., 89, 6938 (1967). SS. 56. 47. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 85 H. Krieger, Suomen KemistiZehti, B38, 260 (1965). Jerrold Meinwald and Paul G. Gassman, J. Amer. Chem. Soc., 88, 2857 (1960). R. L. Shriner, R. C. Fuson, and D. Y. Curtin, "The Systematic Identifications of Organic Compounds," 5th Edition, John Wiley and Sons, Inc., New York, 1964, pp. 163-4. J. D. Roberts, C. C. Lee, and W. H. Saunders, Jr., J. Amer. Chem. Soc., 88, 4501 (1954). Arthur I. Vogel, "Practical Organic Chemistry," 3rd Edition, Spottiswoode, Ballantyne and Co., Ltd., London, 1962, pp. 166-7. S. P. Acharya and H. C. Brown, Chem. Commun., 8988, 305. J. K. Stille and Fred Sonnenberg, J. Amer. Chem. Soc., 88, 4915 (1966). H. C. Brown and H. M. Bell, ibid., 5006 (1964). H. C. Brown and Min-Hon Rei, J. Org. Chem., 88, 1090 (1966). S. Bank, C. A. Rowe, Jr., A. Schriesheim, and L. A. Naslund, J. Amer. Chem. Soc., 89, 6897 (1967). Carl Nebel, Chem. Commun., 8988, 101. J. D. Roberts, C. C. Lee, and W. H. Saunders, Jr., J. Amer. Chem. Soc., 88, 4501 (1954). Boris Franzus and Eugene I. Snyder, ibid., 88, 3423 (1963). 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 86 Ruta K. Bly and Robert S. Bly, J. Org. Chem., 88, 3165 (1963). Alan P. Marchland and Joseph E. Rose, J. Amer. Chem. Soc., 99, 3724 (1968). H. C. Brown and Chandra P. Garg, ibid., 88, 2951 (1961). H. C. Brown, C. P. Garg and Kwang-Ting Liu, J. Org. Chem., 88, 387 (1971). H. Krieger, Suomen Kemistilehti, 888, 136 (1962). Donald G. Farnum and G. Mehta, J. Amer. Chem. Soc., 98, 3256 (1969). R. S. Roy, Anal. Chem., 88, 1724 (1968). T. D. Swartz, Ph.D. Thesis, Johns Hopkins University, 1966. J. L. Dye and V. A. Nicely, J. Chem. Ed., 88, 443 (1971). Program written by Georg C. Sonnichsen. R. S. Mulliken, Tetrahedron, 8989, 253. D. E. Sunko and S. Borcic, "Isotope Effects in Chemical Reactions," ACS Monograph, Clair J. Collins and Newell S. Bowman, eds., Van Nostrand Reinhold Company, New York, 1970, p. 185. K. B. Wiberg and R. Fenoglio, Tet. Lett., 8988, 1273. 4 196 631 93 03 3 '70410414"